Rchemcpp/0000755000175000017500000000000012116123101012727 5ustar klambaueklambaueRchemcpp/source_prep/0000755000175000017500000000000012116123101015255 5ustar klambaueklambaueRchemcpp/source_prep/strip_chemcpp_1.0.2.sh0000755000175000017500000000102712050404203021174 0ustar klambaueklambaue#!/bin/sh # Removes everything from chemcpp which is not necessary for compiling the R-package if [ -d ../src/chemcpp ]; then rm ../src/chemcpp/* 2> /dev/null rm -rf ../src/chemcpp/bin rm -rf ../src/chemcpp/data rm -rf ../src/chemcpp/debug rm -rf ../src/chemcpp/doc rm -rf ../src/chemcpp/examples rm -rf ../src/chemcpp/html rm -rf ../src/chemcpp/templates rm -rf ../src/chemcpp/tools rm -rf ../src/chemcpp/src/Makefile.am rm -rf ../src/chemcpp/src/Makefile.in rm -rf ../src/chemcpp/src/libchemcpp.so echo "stripping done" fi Rchemcpp/source_prep/Makefile0000644000175000017500000000071712050404203016724 0ustar klambaueklambaue CPPFLAGS = -w -fPIC -I../src/ objects := constant.o node.o atom.o bond.o cerror.o chemcpp.o datacontainer.o descriptor.o elements.o jlpioutils.o kcfmolecule.o molecule.o moleculeset.o moleculeutils.o ring.o stringutils.o all: ../../libchemcpp.dll ../../libchemcpp.so ../../libchemcpp.dll: $(objects) $(CXX) -shared -o ../../libchemcpp.dll $(CFLAGS) $(objects) ../../libchemcpp.so: $(objects) $(CXX) -shared -o ../../libchemcpp.so $(CFLAGS) $(objects) Rchemcpp/source_prep/del_chemcpp.sh0000755000175000017500000000007712050404203020065 0ustar klambaueklambaue#!/bin/sh rm -rf ../src/chemcpp rm -f ../inst/chemcpp/data/* Rchemcpp/source_prep/prep_chemcpp_1.0.2.sh0000755000175000017500000000225212050404203021002 0ustar klambaueklambaue#!/bin/sh # This file downloads the chemcpp sources, # 1) It checks wheter the directory "src/chemcpp" exists # if NOT: # 2) It downloads and extracts them into the chemcpp_1.0.2 folder # 3) Patches them (a bug) # 4) It moves the sources to the "src" folder #So: if you want to try out a new/different version of chemcpp, just put it unter src/chemcpp #If you are on windows, put the directory there yourself (and patch it)... if [ ! -d ../src/chemcpp ]; then rm -rf chemcpp_1.0.2 wget http://downloads.sourceforge.net/project/chemcpp/chemcpp/1.0.2/chemcpp_1.0.2.tar.gz tar xfvz chemcpp_1.0.2.tar.gz rm chemcpp_1.0.2.tar.gz patch -p0 < chemcpp_1.0.2_uintbug.diff cp ./Makefile ./chemcpp_1.0.2/src/ cp ./constant.cpp ./chemcpp_1.0.2/src/ grep -v "#define CHEMCPPPATH" ./chemcpp_1.0.2/src/constant.h > ./chemcpp_1.0.2/src/constant.h.tmp mv ./chemcpp_1.0.2/src/constant.h.tmp ./chemcpp_1.0.2/src/constant.h cat ./constant.h.append >> ./chemcpp_1.0.2/src/constant.h mv ./chemcpp_1.0.2 ../src/chemcpp #data-directory #(no symbolic links on windows...) rm -f ../inst/chemcpp/data/* cp ../src/chemcpp/data/*.csv ../inst/chemcpp/data/ fi #you might want to strip chemcpp afterwards... Rchemcpp/source_prep/constant.h.append0000644000175000017500000000015712050404203020532 0ustar klambaueklambaue #include #include std::string GETChemcpppath(); #define CHEMCPPPATH GETChemcpppath() + Rchemcpp/source_prep/chemcpp_1.0.2_uintbug.diff0000644000175000017500000000041212050404203022000 0ustar klambaueklambaue--- ./chemcpp_1.0.2/src/stringutils.h 2007-11-28 12:18:05.000000000 +0100 +++ ./chemcpp_1.0.2/src/stringutils.hnew 2012-01-08 08:44:00.094888849 +0100 @@ -32,6 +32,8 @@ #include #include +typedef unsigned int uint; + using namespace std; Rchemcpp/source_prep/constant.cpp0000644000175000017500000000265212050404203017621 0ustar klambaueklambaue/**************************************************************************************** constants.cpp --------------- copyright : (C) 2012 Mahr Michael ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #include "constant.h" std::string GETChemcpppath() { return getenv("RCHEMCPPPATH"); } Rchemcpp/DESCRIPTION0000644000175000017500000000162712116123101014443 0ustar klambaueklambauePackage: Rchemcpp Type: Package Title: An R package for computing the similarity of molecules Version: 1.1.1 Date: 2013-07-03 Author: Michael Mahr, Guenter Klambauer Maintainer: Guenter Klambauer Description: The Rchemcpp package implements the marginalized graph kernel and extensions, Tanimoto kernels, graph kernels, pharmacophore and 3D kernels suggested for measuring the similarity of molecules. License: GPL (>= 2.1) URL: http://www.bioinf.jku.at/software/Rchemcpp Depends: Rcpp (>= 0.9.13), R (>= 2.15.0), methods Suggests: apcluster LinkingTo: Rcpp RcppModules: Rmolecule, Rmoleculeset, Relements, spectrumhelper, spectrum3Dhelper, subtreehelper Collate: 'getMoleculeNamesFromSDF.R' 'sd2gram3Dpharma.R' 'sd2gram3Dspectrum.R' 'sd2gram.R' 'sd2gramSpectrum.R' 'sd2gramSubtree.R' 'utility.R' 'zzz.R' 'roxygen.R' Packaged: 2013-03-07 14:37:53 UTC; klambaue Rchemcpp/R/0000755000175000017500000000000012116123101013130 5ustar klambaueklambaueRchemcpp/R/sd2gramSpectrum.R0000644000175000017500000002454512116121315016354 0ustar klambaueklambaue#' @title sd2gramSpectrum - Similarity of molecules by walk-based graph kernels #' #' @description This function computes several walk-based graph kernel functions #' based on finite length walks and a fast implementation for input sd file(s). #' #' @usage sd2gramSpectrum(sdFileName, sdFileName2 = "", #' kernelType = c("spectrum", "tanimoto", "minmaxTanimoto","marginalized","lambda"), #' margKernelEndProbability = 0.1, lambdaKernelLambda = 1.0, #' depthMax = as.integer(0), onlyDepthMax = FALSE , flagRemoveH = FALSE, #' morganOrder = as.integer(0), fileType = c("sd","genericsd","kcf"), #' silentMode = FALSE, returnNormalized = FALSE, moleculeNameProperty = "", #' moleculeNameProperty2 = "") #' #' @param sdFileName File containing the molecules. Must be in MDL file format #' (MOL and SDF files). For more information on the file format see #' http://en.wikipedia.org/wiki/Chemical_table_file. Default = "missing" #' @param sdFileName2 A second file containing molecules. Must also be in SDF. #' If specified the molecules of the first file will be compared with the #' molecules of this second file. Default = "missing". #' @param kernelType Sets which kernel is to be used. Options are "spectrum (Spectrum #' kernel) , "tanimoto" (Tanimoto kernel), "minmaxTanimoto" (MinMax Tanimoto kernel), #' "marginalized (Marginalized kernel approximation) and #' "lambda" (LambdaK kernel). Default = "spectrum". #' @param margKernelEndProbability The ending probability for the marginalized #' kernel. Default = 0.1. #' @param lambdaKernelLambda The lambda parameter of the LambdaK kernel. #' Default = 1.0. #' @param depthMax The maximal length of the molecular fragments. #' @param onlyDepthMax Whether fragments up to the given length should be #' used or only fragments of the given length. #' @param flagRemoveH A logical that indicates whether H-atoms should be #' removed or not. #' @param morganOrder The order of the DeMorgan Indices to be used. If set to #' zero no DeMorgan Indices are used. The higher the order the more different #' types of atoms exist and consequently the more dissimilar will be the molecules. #' @param fileType Which filetype was submitted. #' @param silentMode Whether or not the program should print progress reports #' to the standart output. #' @param returnNormalized A logical specifying whether a normalized kernel #' matrix should be returned. Default = TRUE. #' @param moleculeNameProperty A string which specifies the name of the property #' of the molecules in the sdFile from which the row names (and column names) #' are read from. Default = "". #' @param moleculeNameProperty2 A string which specifies the name of the property #' of the molecules in the sdFile 2 from which the column names are read from. #' Default = "". #' @examples #' sdfolder <- system.file("sample_data",package="Rchemcpp") #' sdf <- list.files(sdfolder,full.names=TRUE,pattern="small") #' K <- sd2gramSpectrum(sdf, moleculeNameProperty="Compound Name") #' @return A numeric matrix containing the similarity values between the #' molecules. #' @author Michael Mahr #' c++ function written by Jean-Luc Perret and Pierre Mahe #' #' @examples #' sdfolder <- system.file("sample_data",package="Rchemcpp") #' sdf <- list.files(sdfolder,full.names=TRUE,pattern="tiny") #' moleculeNames <- sd2gramSpectrum(sdf) #' #' @export sd2gramSpectrum = function(sdFileName, sdFileName2 = "", kernelType = c("spectrum", "tanimoto", "minmaxTanimoto","marginalized","lambda"), margKernelEndProbability = 0.1, lambdaKernelLambda = 1.0, depthMax = as.integer(0), onlyDepthMax = FALSE , flagRemoveH = FALSE, morganOrder = as.integer(0), fileType = c("sd","genericsd","kcf"), silentMode = FALSE, returnNormalized = FALSE, moleculeNameProperty = "", moleculeNameProperty2 = "") { margKernelConvgce = 10000; margKernelSkipSkeleton = FALSE; if(!is.character(sdFileName)) stop("sdFileName must be string") if(!is.character(sdFileName2)) stop("sdFileName2 must be string") if(!is.character(kernelType)) stop("kernelType must be a string") if(!is.numeric(margKernelEndProbability)) stop("margKernelEndProbability must be numeric") if(!is.numeric(lambdaKernelLambda)) stop("lambdaKernelLambda must be numeric") if(!is.numeric(depthMax)) stop("depthMax must be integer") depthMax <- as.integer(depthMax) if(!is.logical(onlyDepthMax)) stop("onlyDepthMax must be logical") if(!is.logical(flagRemoveH)) stop("flagRemoveH must be logical") if(!is.numeric(morganOrder)) stop("morganOrder must be integer") morganOrder <- as.integer(morganOrder) if(!is.character(fileType)) stop("fileType must be string") if(!is.logical(silentMode)) stop("silentMode must be logical") if(!is.logical(returnNormalized)) stop("returnNormalized must be logical") if (missing(fileType)) {fileType = fileType[1]} fileType = match.arg(fileType) if (missing(kernelType)) {kernelType = kernelType[1]} kernelType = match.arg(kernelType) #For passing... if (kernelType == "marginalized"){ kernelParam = margKernelEndProbability; }else if (kernelType == "lambda"){ kernelParam = lambdaKernelLambda; }else{ kernelParam = 0.0; } kernelTypeIndex = as.integer(which(c("spectrum", "tanimoto", "minmaxTanimoto","marginalized","lambda") == kernelType)-1) aSet = new (Rchemcpp::Rmoleculeset); aSet2 = new (Rchemcpp::Rmoleculeset); # if sflag = 1 --> SEPARATE TEST SET if( sdFileName2 != "" ){ # 1 - data initialization # ----------------------- # read the set of molecules if( fileType == "sd" ){ aSet$addSD( sdFileName, FALSE ); aSet2$addSD( sdFileName2, FALSE ); }else if( fileType == "genericsd" ){ aSet$addSD( sdFileName, TRUE ); aSet2$addSD( sdFileName2, TRUE ); }else if( fileType == "kcf" ){ aSet$addKCF( sdFileName ); aSet2$addKCF( sdFileName2 ); } if( !silentMode ){ print(paste("*** sd file added ; number of molecules in the set 1 = ", aSet$numMolecules() )); print(paste("*** sd file added ; number of molecules in the set 2 = ", aSet2$numMolecules() )); } # remove H when specified on command line if( flagRemoveH == TRUE ){ if( !silentMode ){ print("removing hydrogens"); } aSet$hideHydrogens(); aSet2$hideHydrogens(); } # compute Morgan labels if( !silentMode ){ print(paste("setting morgan labels ", morganOrder)); } aSet$setMorganLabels( morganOrder ); aSet2$setMorganLabels( morganOrder ); # set kashima probabilities if kernelType = marginalized if(kernelType == "marginalized"){ aSet$setKashimaKernelParam( kernelParam, margKernelConvgce, margKernelSkipSkeleton ); aSet2$setKashimaKernelParam( kernelParam, margKernelConvgce, margKernelSkipSkeleton ); } # initialize gram matrices aSet2$initializeSelfKernel( 0.0); aSet$initializeSelfKernel( 0.0); aSet$setComparisonSetCopy( aSet2 ); aSet$initializeGram( 0.0 ); if( !silentMode){ print("#### initialization Gram OK"); } # 3 - compute the gram matrix # ---------------------------- gramSpectrum_test( aSet, depthMax, kernelTypeIndex, kernelParam, onlyDepthMax, silentMode); if( !silentMode ){ print("gramComputeSpectrum (test) OK"); } # normalize gram if (kernelType == "tanimoto") { aSet$normalizeTanimoto(); } else if (kernelType == "minmaxTanimoto") { aSet$normalizeTanimotoMinMax(); } else { aSet$normalizeGram(); } if( !silentMode ){ print("normalize gram (test) OK"); } if (returnNormalized == FALSE) { K <- do.call(rbind,aSet$getGram() ) } else { K <- do.call(rbind,aSet$getGramNormal() ) } #aSet$writeSelfKernelList( outputDir + baseName + "_test", silentMode ); #aSet$getComparisonSet$writeSelfKernelList( outputDir + baseName + "_train", silentMode ); !!! }else{ # --> SELF KERNEL # 1 - data initialization # ----------------------- # read the set of molecules if( fileType == "sd" ){ aSet$addSD( sdFileName, FALSE ); }else if( fileType == "genericsd" ){ aSet$addSD( sdFileName, TRUE ); }else if( fileType == "kcf" ){ aSet$addKCF( sdFileName ); } if( !silentMode ){ print (paste("*** sd file added ; number of molecules = ", aSet$numMolecules() )); } # remove H when specified on command line if( flagRemoveH == TRUE ){ if( !silentMode ){ print("removing hydrogens"); } aSet$hideHydrogens(); } # compute Morgan labels if( !silentMode ){ print(paste("setting morgan labels ", morganOrder)); } aSet$setMorganLabels( morganOrder ); # set kashima probabilities if kernelType = marginalized if(kernelType == "marginalized"){ aSet$setKashimaKernelParam( kernelParam, margKernelConvgce, margKernelSkipSkeleton ); } # initialize gram matrices aSet$setComparisonSetSelf(); aSet$initializeGram( 0.0 ); aSet$initializeSelfKernel( 0.0); if( !silentMode) { print("#### initialization Gram OK"); } # 3 - compute the gram matrix # --------------------------- gramSpectrum_self( aSet, depthMax, kernelTypeIndex, kernelParam, onlyDepthMax, silentMode); if( !silentMode ){ print("gramComputeSpectrum (self) OK"); } # normalize gram if (kernelType == "tanimoto") { aSet$normalizeTanimoto(); } else if (kernelType == "minmaxTanimoto") { aSet$normalizeTanimotoMinMax(); } else { aSet$normalizeGram(); } if( !silentMode ){ print("normalize gram (self) OK"); } if (returnNormalized == FALSE) { K <- do.call(rbind,aSet$getGram() ) } else { K <- do.call(rbind,aSet$getGramNormal() ) } #aSet$writeSelfKernelList( outputDir + baseName + "_train", false); } #name the molecules if (moleculeNameProperty != "") { molnames = c() for (i in 0:(aSet$numMolecules() -1)) { mol = aSet$getMolByIndex(i); if( moleculeNameProperty %in% mol$listStringDescriptors() ) { molnames = c(molnames, mol$getStringDescriptorValue(moleculeNameProperty)) } else { molnames = c(molnames, i) } } rownames(K) <- molnames if (sdFileName2==""){ colnames(K) <- molnames } else { molnames2 = c() for (i in 0:(aSet2$numMolecules() -1)) { mol2 = aSet2$getMolByIndex(i); if( moleculeNameProperty2 %in% mol2$listStringDescriptors() ) { molnames2 = c(molnames2, mol2$getStringDescriptorValue(moleculeNameProperty2)) } else { molnames2 = c(molnames2, i) } } colnames(K) <- molnames2 } } return ( K ) } Rchemcpp/R/zzz.R0000644000175000017500000000110012057042710014112 0ustar klambaueklambaue #Cannot work with NAMESPACE, because environment-variable has to be set first... .onLoad <- function(lib, pkg, ...) { Sys.setenv(RCHEMCPPPATH=file.path(lib,pkg,"chemcpp")) library.dynam(chname="Rchemcpp",package=pkg,lib.loc=lib) } #loadRcppModules(direct=TRUE) #loadModule("Ratom", TRUE) #loadModule("Rbond", TRUE) loadModule("Rmolecule", TRUE) loadModule("Rmoleculeset", TRUE) #loadModule("Rnode", TRUE) #loadModule("Rring", TRUE) loadModule("Relements", TRUE) loadModule("subtreehelper",TRUE) loadModule("spectrumhelper",TRUE) loadModule("spectrum3Dhelper",TRUE) Rchemcpp/R/sd2gram3Dpharma.R0000644000175000017500000002755112116121726016217 0ustar klambaueklambaue #' @title sd2gram3Dpharma - Similarity of molecules by the exact pharmacophore #' kernel. #' #' @description This tool implements the (exact version of) pharmacophore kernel for 3D #' structures of molecules (\cite{Mahe, 2006}). #' #' @usage sd2gram3Dpharma(sdFileName, sdFileName2 = "", chargesFileName = "", #' chargesFileName2 = "", edgeKernelType = c("RBF", "triangular"), #' edgeKernelParameter = 1, atomKernelMatrix = "", flagRemoveH = FALSE, #' morganOrder = as.integer(0), morganChargesThreshold = 0, #' fileType = c("sd","genericsd","kcf"), silentMode = FALSE, #' returnNormalized = FALSE, moleculeNameProperty = "", moleculeNameProperty2 = "") #' #' #' @param sdFileName File containing the molecules. Must be in MDL file format #' (MOL and SDF files). For more information on the file format see #' http://en.wikipedia.org/wiki/Chemical_table_file. Default = "missing". #' @param sdFileName2 A second file containing molecules. Must also be in SDF. #' If specified the molecules of the first file will be compared with the #' molecules of this second file. Default = "missing". #' @param chargesFileName A character with the name of the file containing #' the atom charges. Default = missing. #' @param chargesFileName2 A character with the name of the file containing #' the atom charges. Default = missing. #' @param edgeKernelType Options to specify the kernel function comparing #' distances between atoms. Choices are "RBF" or "triangular". Default = "RBF". #' @param edgeKernelParameter Specifies the parameter associated to these #' kernels. Either the bandwith of the RBF kernel or the cut-off of the #' triangular kernel. Default = 1. #' @param atomKernelMatrix A string that sets the similarity measure between #' atoms that should be used. Dfault = "missing". #' @param flagRemoveH A logical that indicates whether H-atoms should be #' removed or not. #' @param morganOrder The order of the DeMorgan Indices to be used. If set to #' zero no DeMorgan Indices are used. The higher the order the more different #' types of atoms exist and consequently the more dissimilar will be the molecules. #' @param morganChargesThreshold specifies a threshold above which partial #' Morgan charges are considered as positive/negative. #' @param fileType Which filetype was submitted. #' @param silentMode Whether or not the program should print progress reports #' to the standart output. #' @param returnNormalized A logical specifying whether a normalized kernel #' matrix should be returned. Default = TRUE. #' @param moleculeNameProperty A string which specifies the name of the property #' of the molecules in the sdFile from which the row names (and column names) #' are read from. Default = "". #' @param moleculeNameProperty2 A string which specifies the name of the property #' of the molecules in the sdFile 2 from which the column names are read from. #' Default = "". #' @examples #' sdfolder <- system.file("sample_data",package="Rchemcpp") #' sdf <- list.files(sdfolder,full.names=TRUE,pattern="small") #' K <- sd2gram(sdf, moleculeNameProperty="Compound Name") #' @return A numeric matrix containing the similarity values between the #' molecules. #' @author Michael Mahr #' c++ function written by Jean-Luc Perret and Pierre Mahe #' @references (Mahe, 2006) -- P. Mahe, L. Ralaivola, V. Stoven, and J.-P. Vert. #' The pharmacophore kernel for virtual screening #' with support vector machines. Technical Report, HAL:ccsd-00020066, Ecole des #' Mines de Paris, March 2006. #' #' @examples #' sdfolder <- system.file("sample_data",package="Rchemcpp") #' sdf <- list.files(sdfolder,full.names=TRUE,pattern="tiny") #' moleculeNames <- sd2gram3Dpharma(sdf) #' #' @export sd2gram3Dpharma = function(sdFileName, sdFileName2 = "", chargesFileName = "", chargesFileName2 = "", edgeKernelType = c("RBF", "triangular"), edgeKernelParameter = 1, atomKernelMatrix = "", flagRemoveH = FALSE, morganOrder = as.integer(0), morganChargesThreshold = 0, fileType = c("sd","genericsd","kcf"), silentMode = FALSE, returnNormalized = FALSE, moleculeNameProperty = "", moleculeNameProperty2 = "") { if(!is.character(sdFileName)) stop("sdFileName must be string") if(!is.character(sdFileName2)) stop("sdFileName2 must be string") if(!is.character(chargesFileName)) stop("chargesFileName must be string") if(!is.character(chargesFileName2)) stop("chargesFileName2 must be string") if(!is.character(edgeKernelType)) stop("edgeKernelType must be a string") if(!is.numeric(edgeKernelParameter)) stop("edgeKernelParameter must be numeric") if(!is.character(atomKernelMatrix)) stop("atomKernelMatrix must be string") if(!is.logical(flagRemoveH)) stop("flagRemoveH must be logical") if(!is.numeric(morganOrder)) stop("morganOrder must be integer") morganOrder <- as.integer(morganOrder) if(!is.numeric(morganChargesThreshold)) stop("morganchargesThreshold must be numeric") if(!is.character(fileType)) stop("fileType must be string") if(!is.logical(silentMode)) stop("silentMode must be logical") if(!is.logical(returnNormalized)) stop("returnNormalized must be logical") if (missing(fileType)) {fileType = fileType[1]} fileType = match.arg(fileType) if (missing(edgeKernelType)) {edgeKernelType = edgeKernelType[1]} edgeKernelType = match.arg(edgeKernelType) #-- if((sdFileName2 == "") && (chargesFileName2 != "")) print("chargesFilename2 is useless") if((sdFileName2 != "") && (chargesFileName == "") && (chargesFileName2 != "")) stop("both chargesFileNames have to be specified") if((sdFileName2 != "") && (chargesFileName != "") && (chargesFileName2 == "")) stop("both chargesFileNames have to be specified") if( atomKernelMatrix != "" ){ loadGramAtoms( atomKernelMatrix ); # atomKernelMatrix is set by the } if( sdFileName2 != "" ){ # i.e., test set version. aSet = new (Rchemcpp::Rmoleculeset); aSet2 = new (Rchemcpp::Rmoleculeset); if( fileType == "sd" ){ aSet$addSD( sdFileName, FALSE ); aSet2$addSD( sdFileName2, FALSE ); }else if( fileType == "genericsd" ){ aSet$addSD( sdFileName, TRUE ); aSet2$addSD( sdFileName2, TRUE ); }else if( fileType == "kcf" ){ aSet$addKCF( sdFileName ); aSet2$addKCF( sdFileName2 ); } # read partial charges if specified on command line if(chargesFileName != ""){ if( !silentMode ){ print("reading partial charges"); } aSet$readPartialCharges(chargesFileName); aSet2$readPartialCharges(chargesFileName2); } # remove hydrogens if( flagRemoveH == TRUE ){ if( !silentMode ){ print("removing hydrogens"); } aSet$hideHydrogens(); aSet2$hideHydrogens(); } # set morgan labels if( !silentMode ) print(paste("setting morgan labels ", morganOrder)); aSet$setMorganLabels( morganOrder ); aSet2$setMorganLabels( morganOrder ); # introduce partial charges in Morgan labels if specified on command line if(chargesFileName != ""){ if( !silentMode ){ print("setting partial charges"); } aSet$setMorganChargesLabels(morganChargesThreshold); aSet2$setMorganChargesLabels(morganChargesThreshold); } # compute gram matrix aSet$setComparisonSetCopy( aSet2 ); if( atomKernelMatrix != "" ){ if( !silentMode ){ print("using external atomKernel"); } if (edgeKernelType == "RBF"){ # --> RBF edge kernel aSet$gramCompute3D (edgeKernelParameter, silentMode, TRUE, TRUE) #external,rbf }else if (edgeKernelType == "triangular"){ # --> triangular edge kernel aSet$gramCompute3D (edgeKernelParameter, silentMode, TRUE, FALSE) #external,triangular } } else{ if (edgeKernelType == "RBF"){ # --> RBF edge kernel aSet$gramCompute3D (edgeKernelParameter, silentMode, FALSE, TRUE) #morgan,rbf }else if (edgeKernelType == "triangular"){ # --> triangular edge kernel aSet$gramCompute3D (edgeKernelParameter, silentMode, FALSE, FALSE) #morgan,triangular } } if( !silentMode ){ print("gram matrix computation OK"); } # write the gram matrix to a file (raw and normalized matrices) #aSet$writeGramMatrix( outputDir + baseName, false, false, silentMode); #aSet$writeGramMatrix( outputDir + baseName, true, false, silentMode); # --> write self kernel values if (returnNormalized == FALSE) { K <- do.call(rbind,aSet$getGram() ) } else { K <- do.call(rbind,aSet$getGramNormal() ) } #aSet$writeSelfKernelList( outputDir + baseName + "_test", silentMode ); #aSet2$writeSelfKernelList( outputDir + baseName + "_train", silentMode ); # and exit #aSet$deleteAll(); should be done by destructor #aSet2$deleteAll(); should be done by destructor } else{ # self-set version # create a new moleculeSet aSet = new (Rchemcpp::Rmoleculeset); if( fileType == "sd" ){ aSet$addSD( sdFileName, FALSE ); }else if( fileType == "genericsd" ){ aSet$addSD( sdFileName, TRUE ); }else if( fileType == "kcf" ){ aSet$addKCF( sdFileName ); } # read partial charges if specified on command line if(chargesFileName != ""){ if( !silentMode ) print("reading partial charges"); aSet$readPartialCharges(chargesFileName); } # remove hydrogens if( flagRemoveH == TRUE ) { if( !silentMode ){ print("removing Hydrogens"); } aSet$hideHydrogens(); } # set morgan labels if( !silentMode ){ print(paste("setting morgan labels ", morganOrder)); } aSet$setMorganLabels(morganOrder); if( !silentMode ){ print("setting morgan labels OK"); } # introduce partial charges in Morgan labels if specified on command line if(chargesFileName != ""){ if( !silentMode ){ print("setting partial charges"); } aSet$setMorganChargesLabels(morganChargesThreshold); } # compute gram matrix aSet$setComparisonSetSelf(); if( atomKernelMatrix != "" ){ if( !silentMode ) { print("using external atomKernel"); } if (edgeKernelType == "RBF"){ # --> RBF edge kernel aSet$gramCompute3D (edgeKernelParameter, silentMode, TRUE, TRUE) #external,rbf }else if (edgeKernelType == "triangular"){ # --> triangular edge kernel aSet$gramCompute3D (edgeKernelParameter, silentMode, TRUE, FALSE) #external,triangular } } else{ if (edgeKernelType == "RBF"){ # --> RBF edge kernel aSet$gramCompute3D (edgeKernelParameter, silentMode, FALSE, TRUE) #morgan,rbf }else if (edgeKernelType == "triangular"){ # --> triangular edge kernel aSet$gramCompute3D (edgeKernelParameter, silentMode, FALSE, FALSE) #morgan,triangular } } if( !silentMode ) { print("gram matrix computation OK"); } if (returnNormalized == FALSE) { K <- do.call(rbind,aSet$getGram() ) } else { K <- do.call(rbind,aSet$getGramNormal() ) } # write the gram matrix to a file (raw and normalized matrices) #aSet$writeGramMatrix( outputDir + baseName, false, false, silentMode); #aSet$writeGramMatrix( outputDir + baseName, true, false, silentMode); # --> write self kernel values #aSet$writeSelfKernelList( outputDir + baseName + "_train", false); # and exit #aSet$deleteAll(); should be done by destructor } #name the molecules if (moleculeNameProperty != "") { molnames = c() for (i in 0:(aSet$numMolecules() -1)) { mol = aSet$getMolByIndex(i); if( moleculeNameProperty %in% mol$listStringDescriptors() ) { molnames = c(molnames, mol$getStringDescriptorValue(moleculeNameProperty)) } else { molnames = c(molnames, i) } } rownames(K) <- molnames if (sdFileName2==""){ colnames(K) <- molnames } else { molnames2 = c() for (i in 0:(aSet2$numMolecules() -1)) { mol2 = aSet2$getMolByIndex(i); if( moleculeNameProperty2 %in% mol2$listStringDescriptors() ) { molnames2 = c(molnames2, mol2$getStringDescriptorValue(moleculeNameProperty2)) } else { molnames2 = c(molnames2, i) } } colnames(K) <- molnames2 } } return ( K ) } Rchemcpp/R/utility.R0000644000175000017500000000255412116120670014774 0ustar klambaueklambaue#' @title createRMolecule #' #' @description Creates an Rchemcpp::Rmolecule from an atom-vector and a bond-matrix #' #' @usage createRMolecule(atoms, bonds) #' #' @param atoms A vector containing the symbol names of all atoms in the molecule #' @param bonds A matrix with the same number of rows and columns as the atoms-vector #' containing the type of bonds between the atoms #' @return an instance of "molecule" #' @author Michael Mahr #' #' @examples #' m <- createRMolecule(c("C","C"),matrix(c(0,3,3,0),nrow=2)) #' @export createRMolecule = function(atoms, bonds) { if(!is.vector(atoms)) stop("atoms must be vector") if(!is.matrix(bonds)) stop("bonds must be matrix") bonds = apply(bonds,1,as.integer) if (length(atoms) != nrow(bonds)) stop("matrix must have as many rows as vector") if (length(atoms) != ncol(bonds)) stop("matrix must have as many cols as vector has rows") if (! isSymmetric(bonds) ) stop("matrix must be symetric") if (! all(diag(bonds) == 0)) stop("cannot connect atoms with themselves") if ( (max(bonds) > 3) || (min(bonds) < 0 ) ) stop("Bond types can only range from 1 to 3") m = new(Rchemcpp::Rmolecule) for (k in atoms) { m$addAtom(k) } for (i in 1:length(atoms)){ for (j in 1:length(atoms)){ if ((i < j) && (bonds[i,j] != 0)){ m$linkAtoms(i-1,j-1,bonds[i,j]); } } } return (m) } Rchemcpp/R/getMoleculeNamesFromSDF.R0000644000175000017500000000533112116121314017673 0ustar klambaueklambaue#' @title getMoleculeNamesFromSDF - a helper function #' #' @description This function helps to extract a certain property from an SDF file. Usually #' the molecule class, like "active/non-active" or a property of the molecule, #' like "biological activity", is also stored in the SDF file. These values #' often serve as targets for a prediction task. This function is a small #' wrapper that extracts the information. #' #' @usage #' getMoleculeNamesFromSDF(sdfile) #' #' @param sdfile A character containing the name of the SDF file. #' @return A character vector with one name per molecule. #' @author Guenter Klambauer #' #' @examples #' sdfolder <- system.file("sample_data",package="Rchemcpp") #' sdf <- list.files(sdfolder,full.names=TRUE,pattern="small") #' moleculeNames <- getMoleculeNamesFromSDF(sdf) #' #' @export getMoleculeNamesFromSDF <- function(sdfile){ rl <- readLines(sdfile) mnames <- rl[grep("V[2|3]000",rl)-3] if (length(unique(mnames))!=length(mnames)){ message("Molecule names not unique. Adding ID number.") mnames <- paste(mnames,1:length(mnames),sep="_") } #check if number of molecules is ok if (length(grep("\\$\\$\\$\\$",rl))!=length(mnames)){ stop("Inconsistency in SDF/Mol file. Could not read names.") } return(mnames) } #' @title getMoleculePropertyFromSDF - a helper function #' #' @description This function helps to extract a certain property from an SDF #' file. Usually #' the molecule class, like "active/non-active" or a property of the molecule, #' like "biological activity", is also stored in the SDF file. These values #' often serve as targets for a prediction task. This function is a small #' wrapper that extracts the information. #' #' @usage #' getMoleculePropertyFromSDF(sdfile,property) #' #' #' @param sdfile A character containing the name of the SDF file. #' @param property The name of the slot in the SDF. #' @return A character vector with one value per molecule. #' @author Guenter Klambauer #' #' @examples #' sdfolder <- system.file("sample_data",package="Rchemcpp") #' sdf <- list.files(sdfolder,full.names=TRUE,pattern="small") #' moleculeNames <- getMoleculePropertyFromSDF(sdf,"Activity") #' #' @export getMoleculePropertyFromSDF <- function(sdfile,property){ if (missing(sdfile)){ stop("You must submit an SDF file containing the molecules.") } if (missing(property)){ stop("You must submit the name of the property to be extracted.") } rl <- readLines(sdfile) property <- rl[grep(paste("<",property,">",sep=""),rl)+1] #check if number of molecules is ok if (length(grep("\\$\\$\\$\\$",rl))!=length(property)){ stop("Inconsistency in SDF/Mol file. Could not read properties.") } return(property) } Rchemcpp/R/roxygen.R0000644000175000017500000000013112050404203014743 0ustar klambaueklambaue#' #' @exportPattern ^\\Rmolecule #' @exportPattern ^\\Rmoleculeset #' @import Rcpp NULL Rchemcpp/R/sd2gram.R0000644000175000017500000002165512116121314014627 0ustar klambaueklambaue#' @title sd2gram - Similarity of molecules by the marginalized kernel and #' proposed extensions. #' #' @description This tools compute the marginalized kernel (\cite{Kashima, 2004}) #' and its proposed extensions (\cite{Mahe, 2005)}. #' #' @usage sd2gram(sdFileName, sdFileName2 = "", stopP = 0.1, #' filterTottering = FALSE, converg = as.integer(1000), #' atomKernelMatrix = "", flagRemoveH = FALSE, morganOrder = as.integer(0), #' fileType = c("sd","genericsd","kcf"), silentMode = FALSE, #' returnNormalized = TRUE, moleculeNameProperty = "", #' moleculeNameProperty2 = "") #' #' #' #' @param sdFileName File containing the molecules. Must be in MDL file format #' (MOL and SDF files). For more information on the file format see #' http://en.wikipedia.org/wiki/Chemical_table_file. Default = "missing". #' @param sdFileName2 A second file containing molecules. Must also be in SDF. #' If specified the molecules of the first file will be compared with the #' molecules of this second file. Default = "missing". #' @param stopP ... . Default = 0.1. #' @param filterTottering A logical specifying whether tottering paths should #' be removed. Default = FALSE. #' @param converg A numeric value specifying when convergence is reached. The #' algorithm stops when the kernel value does not change by more #' than 1/c, where c is the value specified by the converg option. #' Default = 1000. #' @param atomKernelMatrix A string that sets the similarity measure between #' atoms that should be used. Dfault = "missing". #' @param flagRemoveH A logical that indicates whether H-atoms should be #' removed or not. #' @param morganOrder The order of the DeMorgan Indices to be used. If set to #' zero no DeMorgan Indices are used. The higher the order the more different #' types of atoms exist and consequently the more dissimilar will be the molecules. #' @param fileType Which filetype was submitted. #' @param silentMode Whether or not the program should print progress reports #' to the standart output. Default = FALSE. #' @param returnNormalized A logical specifying whether a normalized kernel #' matrix should be returned. Default = TRUE. #' @param moleculeNameProperty A string which specifies the name of the property #' of the molecules in the sdFile from which the row names (and column names) #' are read from. Default = "". #' @param moleculeNameProperty2 A string which specifies the name of the property #' of the molecules in the sdFile 2 from which the column names are read from. #' Default = "". #' @examples #' sdfolder <- system.file("sample_data",package="Rchemcpp") #' sdf <- list.files(sdfolder,full.names=TRUE,pattern="small") #' K <- sd2gram(sdf, moleculeNameProperty="Compound Name") #' @return A numeric matrix containing the similarity values between the #' molecules. #' @author Michael Mahr #' c++ function written by Jean-Luc Perret and Pierre Mahe #' @references #' (Kashima, 2004) -- H. Kashima, K. Tsuda, and A. Inokuchi. Kernels for graphs. #' In B. Schoelkopf, K. Tsuda, and J.P. #' Vert, editors, Kernel Methods in Computational Biology, pages 155-170. #' MIT Press, 2004. #' #' (Mahe, 2005) -- P. Mahe, N. Ueda, T. Akutsu, J.-L. Perret, and J.-P. Vert. #' Graph kernels for molecular structure- #' activity relationship analysis with support vector machines. #' J Chem Inf Model, 45(4):939-51, 2005. #' #' @examples #' sdfolder <- system.file("sample_data",package="Rchemcpp") #' sdf <- list.files(sdfolder,full.names=TRUE,pattern="small") #' moleculeNames <- sd2gram(sdf) #' #' @export sd2gram = function(sdFileName, sdFileName2 = "", stopP = 0.1, filterTottering = FALSE, converg = as.integer(1000), atomKernelMatrix = "", flagRemoveH = FALSE, morganOrder = as.integer(0), fileType = c("sd","genericsd","kcf"), silentMode = FALSE, returnNormalized = TRUE, moleculeNameProperty = "", moleculeNameProperty2 = "") { nbThreadsWanted = as.integer(1) fromN = as.integer(1) if(!is.numeric(stopP)) stop("stopP must be numeric") if(!is.logical(filterTottering)) stop("filterTottering must be logical") if(!is.numeric(converg)) stop("converg must be integer") converg <- as.integer(converg) if(!is.character(sdFileName)) stop("sdFileName must be string") if(!is.character(sdFileName2)) stop("sdFileName2 must be string") if(!is.character(atomKernelMatrix)) stop("atomKernelMatrix must be string") if(!is.logical(flagRemoveH)) stop("flagRemoveH must be logical") #if(!is.numeric(nbThreadsWanted)) stop("nbThreadsWanted must be integer") #nbThreadsWanted <- as.integer(nbThreadsWanted) if(!is.numeric(morganOrder)) stop("morganOrder must be integer") morganOrder <- as.integer(morganOrder) #if(!is.numeric(fromN)) stop("fromN must be integer") #fromN <- as.integer(fromN) if(!is.character(fileType)) stop("fileType must be string") if(!is.logical(silentMode)) stop("silentMode must be logical") if(!is.logical(returnNormalized)) stop("returnNormalized must be logical") if (missing(fileType)) {fileType = fileType[1]} fileType = match.arg(fileType) aSet = new (Rchemcpp::Rmoleculeset); aSet2 = new (Rchemcpp::Rmoleculeset); if( atomKernelMatrix != "" ){ loadGramAtoms( atomKernelMatrix ); # atomKernelMatrix is set by the # -k argument on the command line } if( !silentMode ){ print("reading file"); } if( fileType == "sd" ){ aSet$addSD( sdFileName, FALSE); if( flagRemoveH == TRUE ) { if( !silentMode ){ print("removing Hydrogens"); } aSet$hideHydrogens(); } }else if( fileType == "genericsd" ){ aSet$addSD( sdFileName, TRUE); }else if( fileType == "kcf" ){ aSet$addKCF( sdFileName ); if( flagRemoveH == TRUE ) { if( !silentMode ){ print("removing Hydrogens"); } aSet$hideHydrogens(); } } if( !silentMode ){ print("reading file done"); } if(sdFileName2 == ""){ # if the gram matrix is to be computed with all mol files in a single directory if( !silentMode ){ print("setting morgan labels"); } aSet$setMorganLabels(morganOrder); if( atomKernelMatrix != "" ){ if( !silentMode ){ print("using external atomKernel"); } aSet$gramCompute(stopP, converg, fromN, nbThreadsWanted, silentMode, filterTottering, TRUE); #external }else{ if( !silentMode ){ print("using moleculeKernel Kashima"); } aSet$gramCompute(stopP, converg, fromN, nbThreadsWanted, silentMode, filterTottering, FALSE); #morgan } # write the gram matrix to a file # raw matrix and normalised matrix #aSet$deleteAll(); #Should be done by destructor!!! }else{ if( !silentMode ){ print("test set mode"); } if( !silentMode ){ print("reading second file"); } if( fileType == "sd" ){ aSet2$addSD( sdFileName2, FALSE); if( flagRemoveH == TRUE ) { if( !silentMode ){ print("removing Hydrogens"); } aSet2$hideHydrogens(); } }else if( fileType == "genericsd" ){ aSet2$addSD( sdFileName2, TRUE ); }else if( fileType == "kcf" ){ aSet2$addKCF( sdFileName2 ); if( flagRemoveH == TRUE ) { if( !silentMode ){ print("removing Hydrogens"); } aSet2$hideHydrogens(); } } if( !silentMode ){ print("reading second file done"); } # if the gram matrix is to be computed between two datasets of mol files # in two separate directories #deleted if( !silentMode ){ print ("setting morgan labels"); } aSet$setMorganLabels( morganOrder ); aSet2$setMorganLabels( morganOrder ); if( atomKernelMatrix != "" ){ if( !silentMode ){ print("using external atomKernel"); } aSet$gramCompute(stopP, converg, fromN, nbThreadsWanted, silentMode, filterTottering, TRUE); #external }else{ if( !silentMode ){ print("using moleculeKernel Kashima"); } aSet$gramCompute(stopP, converg, fromN, nbThreadsWanted, silentMode, filterTottering, FALSE); #morgan } #aSet$deleteAll(); #Should be done by destructor!!! #aSet2$deleteAll(); #Should be done by destructor!!! } if( !silentMode ){ print("end"); } if (returnNormalized == FALSE) { K <- do.call(rbind,aSet$getGram() ) } else { K <- do.call(rbind,aSet$getGramNormal() ) } #name the molecules if (moleculeNameProperty != "") { molnames = c() for (i in 0:(aSet$numMolecules() -1)) { mol = aSet$getMolByIndex(i); if( moleculeNameProperty %in% mol$listStringDescriptors() ) { molnames = c(molnames, mol$getStringDescriptorValue(moleculeNameProperty)) } else { molnames = c(molnames, i) } } rownames(K) <- molnames if (sdFileName2==""){ colnames(K) <- molnames } else { molnames2 = c() for (i in 0:(aSet2$numMolecules() -1)) { mol2 = aSet2$getMolByIndex(i); if( moleculeNameProperty2 %in% mol2$listStringDescriptors() ) { molnames2 = c(molnames2, mol2$getStringDescriptorValue(moleculeNameProperty2)) } else { molnames2 = c(molnames2, i) } } colnames(K) <- molnames2 } } return ( K ) } Rchemcpp/R/sd2gram3Dspectrum.R0000644000175000017500000003235012116121314016573 0ustar klambaueklambaue #' @title sd2gram3Dspectrum - Similarity of molecules by fast #' approximations of the pharmacophore kernel #' #' @description This tool implements the six discrete approximations of the pharmacophore #' kernel presented in "The pharmacophore kernel for virtual screening #' with support vector machines" (\cite{Mahe, 2006}). #' #' @usage sd2gram3Dspectrum(sdFileName, sdFileName2 = "", #' chargesFileName = "", chargesFileName2 = "", #' kernelType = c("3Pspectrum", "3Pbinary", "3Ptanimoto", #' "2Pspectrum", "2Pbinary", "2Ptanimoto" ), #' depthMax = as.integer(3), nBins = as.integer(20), distMin = 0, #' distMax = 20, flagRemoveH = FALSE, morganOrder = as.integer(0), #' chargesThreshold = 0, fileType = c("sd", "genericsd","kcf"), #' silentMode = FALSE, returnNormalized = FALSE, #' moleculeNameProperty = "", moleculeNameProperty2 = "") #' #' @param sdFileName File containing the molecules. Must be in MDL file format #' (MOL and SDF files). For more information on the file format see #' http://en.wikipedia.org/wiki/Chemical_table_file. Default = "missing". #' @param sdFileName2 A second file containing molecules. Must also be in SDF. #' If specified the molecules of the first file will be compared with the #' molecules of this second file. Default = "missing". #' @param chargesFileName A character with the name of the file containing #' the atom charges. Default = missing. #' @param chargesFileName2 A character with the name of the file containing #' the atom charges. Default = missing. #' @param kernelType Type of kernel to be used. Possible choices are #' 3-points spectrum kernel ("3Pspectrum"), #' 3-points binary kernel ("3Pbinary"), #' 3-points Tanimoto kernel ("3Ptanimoto"), #' 2-points spectrum kernel ("2Pspectrum"), #' 2-points binary kernel ("2Pbinary"), #' 2-points Tanimoto kernel ("2Ptanimoto"). Default = "3Pspectrum". #' @param depthMax The maximal length of the molecular fragments. Default = 3. #' @param nBins number of bins used to discretize the inter-atomic lengths. #' An adequate value for the number of bins is between 20 and 30. #' Default = 20. #' @param distMin minimum distance for inter-atomic distance range. #' Default = 0. #' @param distMax maximum distance in angstrom for inter-atomic distance range. #' Default = 20. #' @param chargesThreshold specifies a threshold above which partial charges #' are considered as positive/negative. #' By default this threshold is zero, and every positive (resp. negative) #' partial charge is seen as a positive (resp. negative) charge. #' However, it might be interesting to consider a threshold of 0.2 for #' example, in which case only partial charges greater than 0.2 #' (resp. smaller than -0.2) would be seen as positive (resp. negative). #' Default = 0. #' @param flagRemoveH A logical that indicates whether H-atoms should be #' removed or not. #' @param morganOrder The order of the DeMorgan Indices to be used. If set to #' zero no DeMorgan Indices are used. The higher the order the more different #' types of atoms exist and consequently the more dissimilar will be the molecules. #' @param fileType Which filetype was submitted. #' @param silentMode Whether or not the program should print progress reports #' to the standart output. #' @param returnNormalized A logical specifying whether a normalized kernel #' matrix should be returned. Default = TRUE. #' @param moleculeNameProperty A string which specifies the name of the property #' of the molecules in the sdFile from which the row names (and column names) #' are read from. Default = "". #' @param moleculeNameProperty2 A string which specifies the name of the property #' of the molecules in the sdFile 2 from which the column names are read from. #' Default = "". #' @examples #' sdfolder <- system.file("sample_data",package="Rchemcpp") #' sdf <- list.files(sdfolder,full.names=TRUE,pattern="small") #' K <- sd2gram3Dspectrum(sdf, moleculeNameProperty="Compound Name") #' @return A numeric matrix containing the similarity values between the #' molecules. #' @author Michael Mahr #' c++ function written by Jean-Luc Perret and Pierre Mahe #' @references (Mahe, 2006) -- P. Mahe, L. Ralaivola, V. Stoven, and J.-P. Vert. #' The pharmacophore kernel for virtual screening #' with support vector machines. Technical Report, HAL:ccsd-00020066, Ecole des #' Mines de Paris, March 2006. #' #' #' @examples #' sdfolder <- system.file("sample_data",package="Rchemcpp") #' sdf <- list.files(sdfolder,full.names=TRUE,pattern="tiny") #' moleculeNames <- sd2gram3Dspectrum(sdf) #' #' #' @export sd2gram3Dspectrum = function(sdFileName, sdFileName2 = "", chargesFileName = "", chargesFileName2 = "", kernelType = c("3Pspectrum", "3Pbinary", "3Ptanimoto", "2Pspectrum", "2Pbinary", "2Ptanimoto" ), depthMax = as.integer(3), nBins = as.integer(20), distMin = 0, distMax = 20, flagRemoveH = FALSE, morganOrder = as.integer(0), chargesThreshold = 0, fileType = c("sd", "genericsd","kcf"), silentMode = FALSE, returnNormalized = FALSE, moleculeNameProperty = "", moleculeNameProperty2 = "") { if(!is.character(sdFileName)) stop("sdFileName must be string") if(!is.character(sdFileName2)) stop("sdFileName2 must be string") if(!is.character(chargesFileName)) stop("chargesFileName must be string") if(!is.character(chargesFileName2)) stop("chargesFileName2 must be string") if(!is.character(kernelType)) stop("kernelType must be a string") if(!is.integer(depthMax)) stop("depthMax must be integer") if(!is.logical(flagRemoveH)) stop("flagRemoveH must be logical") if(!is.numeric(morganOrder)) stop("morganOrder must be integer") morganOrder <- as.integer(morganOrder) if(!is.numeric(chargesThreshold)) stop("chargesThreshold must be numeric") if(!is.numeric(nBins)) stop("nBins must be integer") nBins <- as.integer(nBins) if(!is.numeric(distMin)) stop("distMin must be numeric") if(!is.numeric(distMax)) stop("distMax must be numeric") if(!is.character(fileType)) stop("fileType must be string") if(!is.logical(silentMode)) stop("silentMode must be logical") if(!is.logical(returnNormalized)) stop("returnNormalized must be logical") if((sdFileName2 == "") && (chargesFileName2 != "")) print("chargesFilename2 is useless") if((sdFileName2 != "") && (chargesFileName == "") && (chargesFileName2 != "")) stop("both chargesFileNames have to be specified") if((sdFileName2 != "") && (chargesFileName != "") && (chargesFileName2 == "")) stop("both chargesFileNames have to be specified") if (missing(fileType)) {fileType = fileType[1]} fileType = match.arg(fileType) if (missing(kernelType)) {kernelType = kernelType[1]} kernelType = match.arg(kernelType) #for passing kernelTypeIndex = as.integer(which(c("3Pspectrum", "3Pbinary", "3Ptanimoto", "2Pspectrum", "2Pbinary", "2Ptanimoto" ) == kernelType)-1) aSet = new (Rchemcpp::Rmoleculeset); aSet2 = new (Rchemcpp::Rmoleculeset); # if sflag = 1 --> SEPARATE TEST SET if( sdFileName2 != "" ){ # 1 - data initialization # ----------------------- # read the set of molecules if( fileType == "sd" ){ aSet$addSD( sdFileName, FALSE ); aSet2$addSD( sdFileName2, FALSE ); }else if( fileType == "genericsd" ){ aSet$addSD( sdFileName, TRUE ); aSet2$addSD( sdFileName2, TRUE ); }else if( fileType == "kcf" ){ aSet$addKCF( sdFileName ); aSet2$addKCF( sdFileName2 ); } if( !silentMode ){ print(paste("*** sd file added ; number of molecules in the set 1 = ", aSet$numMolecules() )); print(paste("*** sd file added ; number of molecules in the set 2 = ", aSet2$numMolecules() )); } # read partial charges if specified on command line if(chargesFileName != ""){ if( !silentMode ){ print("reading partial charges"); } aSet$readPartialCharges(chargesFileName); aSet2$readPartialCharges(chargesFileName2); } # remove H when specified on command line if( flagRemoveH == TRUE ){ if( !silentMode ){ print("removing hydrogens"); } aSet$hideHydrogens(); aSet2$hideHydrogens(); } # compute Morgan labels if( !silentMode ){ print(paste("setting morgan labels ", morganOrder)); } aSet$setMorganLabels( morganOrder ); aSet2$setMorganLabels( morganOrder ); # introduce partial charges in Morgan labels if specified on command line if(chargesFileName != ""){ if( !silentMode ){ print("setting partial charges"); } aSet$setMorganChargesLabels(chargesThreshold); aSet2$setMorganChargesLabels(chargesThreshold); } # set kashima probabilities if kernelType = marginalized #if(kernelType == "marginalized"){ # aSet$setKashimaKernelParam( kernelParam, convgce, skipSkeleton ); # aSet2$setKashimaKernelParam( kernelParam, convgce, skipSkeleton ); #} # initialize gram matrices aSet$initializeSelfKernel( 0.0); aSet2$initializeSelfKernel(0.0); aSet$setComparisonSetCopy( aSet2 ); aSet$initializeGram( 0.0 ); if( !silentMode){ print("#### initialization Gram OK"); } # 3 - compute the gram matrix # ---------------------------- gramSpectrum3D_test( aSet, depthMax, kernelTypeIndex, nBins, distMin, distMax, silentMode); if( !silentMode ){ print("gramComputeSpectrum (test) OK"); } if (kernelType == "3Pspectrum"){ aSet$normalizeGram(); } else if (kernelType == "3Pbinary"){ aSet$normalizeGram(); } else if (kernelType == "3Ptanimoto"){ aSet$normalizeTanimoto(); } else if (kernelType == "2Pspectrum"){ aSet$normalizeGram(); } else if (kernelType == "2Pbinary"){ aSet$normalizeGram(); } else if (kernelType == "2Ptanimoto"){ aSet$normalizeTanimoto(); } if( !silentMode ){ print("normalize gram (test) OK"); } if (returnNormalized == FALSE) { K <- do.call(rbind,aSet$getGram() ) } else { K <- do.call(rbind,aSet$getGramNormal() ) } #aSet$writeSelfKernelList( outputDir + baseName + "_test", silentMode ); #aSet$getComparisonSet$writeSelfKernelList( outputDir + baseName + "_train", silentMode ); !!! }else{ # --> SELF KERNEL # 1 - data initialization # ----------------------- # read the set of molecules if( fileType == "sd" ){ aSet$addSD( sdFileName, FALSE ); }else if( fileType == "genericsd" ){ aSet$addSD( sdFileName, TRUE ); }else if( fileType == "kcf" ){ aSet$addKCF( sdFileName ); } if( !silentMode ){ print (paste("*** sd file added ; number of molecules = ", aSet$numMolecules() )); } # read partial charges if specified on command line if(chargesFileName != ""){ if( !silentMode ){ print("reading partial charges"); } aSet$readPartialCharges(chargesFileName); } # remove H when specified on command line if( flagRemoveH == TRUE ){ if( !silentMode ){ print("removing hydrogens"); } aSet$hideHydrogens(); } # compute Morgan labels if( !silentMode ){ print(paste("setting morgan labels ", morganOrder)); } aSet$setMorganLabels( morganOrder ); # introduce partial charges in Morgan labels if specified on command line if(chargesFileName != ""){ if( !silentMode ){ print("setting partial charges"); } aSet$setMorganChargesLabels(chargesThreshold); } # set kashima probabilities if kernelType = marginalized #if(kernelType == "marginalized"){ # aSet$setKashimaKernelParam( kernelParam, convgce, skipSkeleton ); #} # initialize gram matrices aSet$setComparisonSetSelf(); aSet$initializeGram( 0.0 ); aSet$initializeSelfKernel( 0.0); if( !silentMode) { print("#### initialization Gram OK"); } # 3 - compute the gram matrix # --------------------------- gramSpectrum3D_self( aSet, depthMax, kernelTypeIndex, nBins, distMin, distMax, silentMode); if( !silentMode ){ print("gramComputeSpectrum (self) OK"); } #normalize Gram if (kernelType == "3Pspectrum"){ aSet$normalizeGram(); } else if (kernelType == "3Pbinary"){ aSet$normalizeGram(); } else if (kernelType == "3Ptanimoto"){ aSet$normalizeTanimoto(); } else if (kernelType == "2Pspectrum"){ aSet$normalizeGram(); } else if (kernelType == "2Pbinary"){ aSet$normalizeGram(); } else if (kernelType == "2Ptanimoto"){ aSet$normalizeTanimoto(); } if( !silentMode ){ print("normalize gram (self) OK"); } if (returnNormalized == FALSE) { K <- do.call(rbind,aSet$getGram() ) } else { K <- do.call(rbind,aSet$getGramNormal() ) } #aSet$writeSelfKernelList( outputDir + baseName + "_train", false); } #name the molecules if (moleculeNameProperty != "") { molnames = c() for (i in 0:(aSet$numMolecules() -1)) { mol = aSet$getMolByIndex(i); if( moleculeNameProperty %in% mol$listStringDescriptors() ) { molnames = c(molnames, mol$getStringDescriptorValue(moleculeNameProperty)) } else { molnames = c(molnames, i) } } rownames(K) <- molnames if (sdFileName2==""){ colnames(K) <- molnames } else { molnames2 = c() for (i in 0:(aSet2$numMolecules() -1)) { mol2 = aSet2$getMolByIndex(i); if( moleculeNameProperty2 %in% mol2$listStringDescriptors() ) { molnames2 = c(molnames2, mol2$getStringDescriptorValue(moleculeNameProperty2)) } else { molnames2 = c(molnames2, i) } } colnames(K) <- molnames2 } } return ( K ) } Rchemcpp/R/sd2gramSubtree.R0000644000175000017500000002402412116122117016154 0ustar klambaueklambaue#' @title sd2gramSubtree - Similarity of molecules by several graph kernels #' based on the count of common subtrees #' #' @description This tools computes several graph kernels based on the detection of common #' subtrees: the so-called #' tree-pattern graph kernels, originally introduced in (\cite{Ramon, 2003}), #' and revisited in (\cite{Mahe, 2006}). #' #' @usage sd2gramSubtree(sdFileName, sdFileName2 = "", #' kernelType = c("sizebased","branchingbased") , #' branchKernelUntilN = FALSE, lambda = 1, depthMax = as.integer(3), #' flagRemoveH = FALSE, filterTottering = FALSE, morganOrder = as.integer(0), #' fileType = c("sd", "genericsd","kcf"), silentMode = FALSE, #' returnNormalized = FALSE, moleculeNameProperty = "", #' moleculeNameProperty2 = "") #' #' #' @param sdFileName File containing the molecules. Must be in MDL file format #' (MOL and SDF files). For more information on the file format see #' http://en.wikipedia.org/wiki/Chemical_table_file. Default = "missing" #' @param sdFileName2 A second file containing molecules. Must also be in SDF. #' If specified the molecules of the first file will be compared with the #' molecules of this second file. Default = "missing". #' @param kernelType Determines whether subtrees of the molecule are penalized #' size-based or branching-based. Default = "sizebased". #' @param branchKernelUntilN Logical whether tree patterns of until N should be #' considered. Default = FALSE. #' @param lambda Weighted contribution of tree-patterns depending on their sizes #' Default = 1. #' @param depthMax tree-patterns of depth. Default = 3. #' @param flagRemoveH A logical that indicates whether H-atoms should be #' removed or not. #' @param filterTottering A logical that indicates whether tottering paths #' should be removed. Default = FALSE. #' @param morganOrder The order of the DeMorgan Indices to be used. If set to #' zero no DeMorgan Indices are used. The higher the order the more different #' types of atoms exist and consequently the more dissimilar will be the molecules. #' @param fileType Which filetype was submitted. #' @param silentMode Whether the program should print progress reports #' to the standart output. #' @param returnNormalized A logical specifying whether a normalized kernel #' matrix should be returned. Default = TRUE. #' @param moleculeNameProperty A string which specifies the name of the property #' of the molecules in the sdFile from which the row names (and column names) #' are read from. Default = "". #' @param moleculeNameProperty2 A string which specifies the name of the property #' of the molecules in the sdFile 2 from which the column names are read from. #' Default = "". #' @examples #' sdfolder <- system.file("sample_data",package="Rchemcpp") #' sdf <- list.files(sdfolder,full.names=TRUE,pattern="small") #' K <- sd2gram(sdf, moleculeNameProperty="Compound Name") #' @return A numeric matrix containing the similarity values between the #' molecules. #' @author Michael Mahr #' c++ function written by Jean-Luc Perret and Pierre Mahe #' @references #' (Mahe, 2006) -- P. Mahe and J.-P. Vert. Graph kernels based on tree patterns #' for molecules. Technical Report, HAL:ccsd-00095488, Ecoles des Mines de Paris, #' September 2006. #' (Ramon, 2003) -- J. Ramon and T. Gaertner. Expressivity versus efficiency of #' graph kernels. In T. Washio and L. De Raedt, editors, #' Proceedings of the First International Workshop on Mining Graphs, #' Trees and Sequences, pages 65-74, 2003. #' #' @examples #' sdfolder <- system.file("sample_data",package="Rchemcpp") #' sdf <- list.files(sdfolder,full.names=TRUE,pattern="tiny") #' moleculeNames <- sd2gramSubtree(sdf) #' #' @export sd2gramSubtree = function(sdFileName, sdFileName2 = "", kernelType = c("sizebased","branchingbased") , branchKernelUntilN = FALSE, lambda = 1, depthMax = as.integer(3), flagRemoveH = FALSE, filterTottering = FALSE, morganOrder = as.integer(0), fileType = c("sd", "genericsd","kcf"), silentMode = FALSE, returnNormalized = FALSE, moleculeNameProperty = "", moleculeNameProperty2 = "") { branchFlag = FALSE; if(!is.character(sdFileName)) stop("sdFileName must be string") if(!is.character(sdFileName2)) stop("sdFileName2 must be string") if(!is.logical(branchKernelUntilN)) stop("branchKernelUntilN must be logical") if(!is.numeric(lambda)) stop("lambda must be numeric") if(!is.numeric(depthMax)) stop("depthMax must be integer") depthMax <- as.integer(depthMax) if(!is.logical(flagRemoveH)) stop("flagRemoveH must be logical") if(!is.logical(filterTottering)) stop("filterTottering must be logical") if(!is.numeric(morganOrder)) stop("morganOrder must be integer") morganOrder <- as.integer(morganOrder) if(!is.character(fileType)) stop("fileType must be string") if(!is.logical(silentMode)) stop("silentMode must be logical") if(!is.logical(returnNormalized)) stop("returnNormalized must be logical") if (missing(fileType)) {fileType = fileType[1]} fileType = match.arg(fileType) if (missing(kernelType)) {kernelType = kernelType[1]} kernelType = match.arg(kernelType) aSet = new (Rchemcpp::Rmoleculeset); aSet2 = new (Rchemcpp::Rmoleculeset); if(kernelType == 1) branchFlag = TRUE; # JP: For the fast version we need a static variable to store the list of subsets of p indeices among n choices # The static variable tuples is initialized once with the initialize_tuples( int nmax) function # After initialization, tuples[n][p] is a vector of vector that contains all different p-tuples of {1,..,n}. initialize_tuples(4); if( sdFileName2 != "" ){ # 1 - data initialization # ----------------------- # read the set of molecules if( fileType == "sd" ){ aSet$addSD( sdFileName, FALSE ); aSet2$addSD( sdFileName2, FALSE ); }else if( fileType == "genericsd" ){ aSet$addSD( sdFileName, TRUE ); aSet2$addSD( sdFileName2, TRUE ); }else if( fileType == "kcf" ){ aSet$addKCF( sdFileName ); aSet2$addKCF( sdFileName2 ); } if( !silentMode ){ print(paste("*** sd file added ; number of molecules in the set 1 = ", aSet$numMolecules() ) ); print(paste("*** sd file added ; number of molecules in the set 2 = ", aSet2$numMolecules() ) ); } # remove H when specified on command line if( flagRemoveH == TRUE ){ if( !silentMode ){ print( "removing hydrogens" ); } aSet$hideHydrogens(); aSet2$hideHydrogens(); } # compute Morgan labels if( !silentMode ){ print(paste("setting morgan labels ", morganOrder)); } aSet$setMorganLabels( morganOrder ); aSet2$setMorganLabels( morganOrder ); # no-totters transformation if(filterTottering){ aSet$noTottersTransform(); aSet2$noTottersTransform(); } # initialize gram matrices aSet$initializeSelfKernel( 0.0); aSet2$initializeSelfKernel( 0.0); aSet$setComparisonSetCopy( aSet2 ); aSet$initializeGram( 0.0 ); if( !silentMode) print("initialization Gram OK"); # Initialize the indexes #aSet$initialize_extended(); #moved to function #aSet2$initialize_extended(); #moved to function gramSubtree_test( aSet, lambda, depthMax, filterTottering, branchFlag, branchKernelUntilN, silentMode); if( !silentMode ){ print("compute gram OK"); } # normalize gram aSet$normalizeGram(); if( !silentMode ){ print("normalize gram (test) OK"); } if (returnNormalized == FALSE) { K <- do.call(rbind,aSet$getGram() ) } else { K <- do.call(rbind,aSet$getGramNormal() ) } #aSet$writeSelfKernelList( outputDir + baseName + "_test", silentMode ); #aSet2$writeSelfKernelList( outputDir + baseName + "_train", silentMode ); }else{ # SELF KERNEL # 1 - data initialization # ----------------------- # read the set of molecules if( fileType == "sd" ){ aSet$addSD( sdFileName, FALSE ); }else if( fileType == "genericsd" ){ aSet$addSD( sdFileName, TRUE ); }else if( fileType == "kcf" ){ aSet$addKCF( sdFileName ); } if( !silentMode ) print(paste("*** sd file added ; number of molecules in the set 1 = ", aSet$numMolecules() ) ); # remove H when specified on command line if( flagRemoveH == TRUE ){ if( !silentMode ){ print("removing hydrogens"); } aSet$hideHydrogens(); } # compute Morgan labels if( !silentMode ){ print(paste("setting morgan labels ", morganOrder)); } aSet$setMorganLabels( morganOrder ); # no-totters transformation if(filterTottering){ aSet$noTottersTransform(); } # initialize gram matrices aSet$setComparisonSetSelf(); aSet$initializeGram( 0.0 ); aSet$initializeSelfKernel( 0.0 ); if( !silentMode){ print("initialization Gram OK"); } #aSet$initialize_extended(); #moved to function gramSubtree_self(aSet, lambda, depthMax, filterTottering, branchFlag, branchKernelUntilN, silentMode); if( !silentMode ) print("compute gram OK"); # normalize gram aSet$normalizeGram(); if( !silentMode ) print("normalize gram (self) OK"); # write gram matrices and self kernel values if (returnNormalized == FALSE) { K <- do.call(rbind,aSet$getGram() ) } else { K <- do.call(rbind,aSet$getGramNormal() ) } #aSet$writeSelfKernelList( outputDir + baseName + "_train", false); } #name the molecules if (moleculeNameProperty != "") { molnames = c() for (i in 0:(aSet$numMolecules() -1)) { mol = aSet$getMolByIndex(i); if( moleculeNameProperty %in% mol$listStringDescriptors() ) { molnames = c(molnames, mol$getStringDescriptorValue(moleculeNameProperty)) } else { molnames = c(molnames, i) } } rownames(K) <- molnames if (sdFileName2==""){ colnames(K) <- molnames } else { molnames2 = c() for (i in 0:(aSet2$numMolecules() -1)) { mol2 = aSet2$getMolByIndex(i); if( moleculeNameProperty2 %in% mol2$listStringDescriptors() ) { molnames2 = c(molnames2, mol2$getStringDescriptorValue(moleculeNameProperty2)) } else { molnames2 = c(molnames2, i) } } colnames(K) <- molnames2 } } return ( K ) } Rchemcpp/scripts/0000755000175000017500000000000012116123101014416 5ustar klambaueklambaueRchemcpp/scripts/testing.R0000644000175000017500000000501712050404203016223 0ustar klambaueklambauelibrary(Rchemcpp) library(apcluster) sdfolder <- system.file("sample_data",package="Rchemcpp") sdf <- list.files(sdfolder,full.names=TRUE,pattern="small") K <- sd2gram(sdf,returnNormalized=TRUE) depth <- 10 K <- sd2gramSpectrum(sdf,kernelType="spectrum",depthMax=as.integer(depth),returnNormalized=TRUE) K <- sd2gramSpectrum(sdf,kernelType="tanimoto",depthMax=as.integer(depth),returnNormalized=TRUE) K <- sd2gramSpectrum(sdf,kernelType="minmaxTanimoto",depthMax=as.integer(depth),returnNormalized=TRUE) K <- sd2gramSpectrum(sdf,kernelType="marginalized",depthMax=as.integer(depth),returnNormalized=TRUE) K1 <- sd2gramSpectrum(sdf,kernelType="lambda",depthMax=as.integer(5),lambdaKernelLambda=4,returnNormalized=TRUE) K2 <- sd2gramSpectrum(sdf,kernelType="lambda",depthMax=as.integer(depth),lambdaKernelLambda=4,returnNormalized=TRUE) all(K1==K2) K1 <- sd2gramSubtree(sdf,kernelType="sizebased",depthMax=as.integer(10),branchKernelUntilN=TRUE) K2 <- sd2gramSubtree(sdf,kernelType="branchingbased",depthMax=as.integer(10),branchKernelUntilN=TRUE) all(K1==K2) K <- sd2gram3Dspectrum(sdf,returnNormalized=TRUE) K <- sd2gram3Dpharma(sdf,returnNormalized=TRUE) r <- apcluster(K) plot(r,K) #getMoleculeNamesFromSDF(sdf) getMoleculePropertyFromSDF(sdf,property="Activity") #http://www.clab.kwansei.ac.jp/mining/datasets/PAKDD2000/okd.htm #[Debnath 91] A. K. Debnath, R. L. Lopez de Compadre, G. Debnath, A. J. Shusterman & C. Hansch: Structure-Activity Relationship of Mutagenic Aromatic and Heteroaromatic Nitro Compounds. Correlation with Molecular Orbital Energies and Hydrophobicity, Journal of Medicinal Chemistry, Vol.34, pp.786-797 (1991). #perl one-liner to remove non-ASCII Characters: perl -i.bk -pe 's/[^[:ascii:]]//g;' file ## svm prediction ################################################################ library(kernlab) library(Rchemcpp) sdfolder <- system.file("sample_data",package="Rchemcpp") sdf <- list.files(sdfolder,full.names=TRUE,pattern="mutag.sdf") K <- sd2gramSpectrum(sdf,kernelType="lambda",depthMax=as.integer(15),lambdaKernelLambda=4,returnNormalized=TRUE) y <- as.numeric(getMoleculePropertyFromSDF(sdf,"Activity")) idxRm <- which(y==-99) K <- K[-idxRm,-idxRm] y <- y[-idxRm] prediction <- vector("numeric",length(y)) for (idx in 1:length(y)){ trainK <- as.kernelMatrix(K)[-idx,-idx] model <- ksvm(trainK,y=y[-idx],kernel="matrix",type="nu-svr") testK <- as.kernelMatrix(K[idx,-idx,drop=FALSE][ ,SVindex(model),drop=FALSE]) prediction[idx] <- as.numeric(predict(model, testK)) } plot(prediction,y) cvError <- y - prediction Rchemcpp/NAMESPACE0000644000175000017500000000043012116120732014152 0ustar klambaueklambaueexport(createRMolecule) export(getMoleculeNamesFromSDF) export(getMoleculePropertyFromSDF) exportPattern("^\\Rmolecule") exportPattern("^\\Rmoleculeset") export(sd2gram) export(sd2gram3Dpharma) export(sd2gram3Dspectrum) export(sd2gramSpectrum) export(sd2gramSubtree) import(Rcpp) Rchemcpp/src/0000755000175000017500000000000012116123101013516 5ustar klambaueklambaueRchemcpp/src/Rmoleculeset.h0000644000175000017500000002201512116123101016332 0ustar klambaueklambaue #ifndef RMOLECULESET_H #define RMOLECULESET_H #include #include #include "Rmolecule.h" #include #include class Rmoleculeset : public MoleculeSet { public: Rmoleculeset() { //superclass constructor is called... comparisonSet = NULL; } ~Rmoleculeset() { MoleculeSet::deleteAll(); //It is assumed that all molecules are a part of the set! if (comparisonSet != NULL) if (comparisonSet != this) delete comparisonSet; //Also the comparisonset is a part of the set! } Rmoleculeset * getComparisonSet2() //no R accessor {return (Rmoleculeset*)comparisonSet;} vector< vector > * getGram2() //no R accessor {return (gram);} vector< vector > * getGramNormal2() //no R accessor {return (gramNormal);} Rmoleculeset(Rmoleculeset * aSet) //copy constructor (more or less) { gramCalculated = (*aSet).gramCalculated; pq = (*aSet).pq; convergenceCondition = (*aSet).convergenceCondition; subsetStart = (*aSet).subsetStart; subsetSize = (*aSet).subsetSize; activitySet = (*aSet).activitySet; if (aSet->getComparisonSet2() != NULL) comparisonSet = new Rmoleculeset(aSet->getComparisonSet2()); //deep copy (recursive) else comparisonSet = NULL; if (aSet->getGram2() != NULL) (*gram) = (*aSet->getGram2()); //deep copy else gram = NULL; if (aSet->getGramNormal2() != NULL) (*gramNormal) = (*aSet->getGramNormal2()); //deep copy else gramNormal = NULL; //deep copy all contained molecules this->clear(); for( vector::iterator m1 = aSet->begin(); m1 != aSet->end(); m1++ ) { this->push_back(new Molecule(*(*m1),false)); } } //RcppModules in Rcpp 0.9.13 does not allow exposing base class method, so they need to be redefined here int addSD2( string aFileName, bool genericAtomTypeFlag = false, long beginMolecule = -1, long endMolecule = -1 ) {return MoleculeSet::addSD(aFileName, genericAtomTypeFlag, beginMolecule, endMolecule);} //R or RcppModules also seems not to allow overloaded methods in Rcpp 0.9.13 //It also does not seem to allow "Formal argument specification" (default-values) with ".method" member-functions. int addSD( string aFileName, bool genericAtomTypeFlag) {return MoleculeSet::addSD(aFileName, genericAtomTypeFlag);} //-- int addKCF2( string aFileName, long beginMolecule = -1, long endMolecule = -1 ) {return MoleculeSet::addKCF(aFileName, beginMolecule, endMolecule);} int addKCF( string aFileName) {return MoleculeSet::addKCF(aFileName);} //-- uint numMolecules() {return MoleculeSet::numMolecules();} void hideHydrogens() {MoleculeSet::hideHydrogens();} void setMorganLabels(int order) {MoleculeSet::setMorganLabels(order);} void setKashimaKernelParam( double aPq, int aConvergenceCondition, bool skipSkeleton = false ) {MoleculeSet::setKashimaKernelParam( aPq, aConvergenceCondition, skipSkeleton );} //avoid a segfault... void initializeGram( double value ) { if (comparisonSet == NULL) { CError e = CError(MOLECULENOTFOUND, "Comparisonset is not yet set"); e.describe(); throw(e); } MoleculeSet::initializeGram(value); } void initializeSelfKernel( double value ) {MoleculeSet::initializeSelfKernel(value);} void setComparisonSetCopy(SEXP s) { //if (comparisonSet != NULL) delete comparisonSet; std::string rtypename("Rcpp_Rmoleculeset"); Rcpp::S4 s4obj( s ); if ( !s4obj.is( rtypename.c_str() ) ) { Rf_error( (std::string("object is not of the type ")+rtypename).c_str() ); } Rcpp::Environment env( s4obj ); Rcpp::XPtr xptr( env.get(".pointer") ); Rmoleculeset *o = static_cast (R_ExternalPtrAddr(xptr)); //Note: it could already be assigned here - however it is cloned to avoid scope/garbage-collection issues //MoleculeSet::setComparisonSet(o); MoleculeSet::setComparisonSet(new Rmoleculeset(o)); //invoke copy-constructor to clone object } void setComparisonSetSelf() { comparisonSet = this; } SEXP getComparisonSet() { Rcpp::XPtr xp((Rmoleculeset*)comparisonSet, false); //do NOT mark as finalizable! Rcpp::Function maker=Rcpp::Environment::Rcpp_namespace()[ "cpp_object_maker"]; return maker ( typeid(Rmoleculeset).name() , xp ); } void normalizeTanimoto() {MoleculeSet::normalizeTanimoto();} void normalizeTanimotoMinMax() {MoleculeSet::normalizeTanimotoMinMax();} void normalizeGram() {MoleculeSet::normalizeGram();} void noTottersTransform() {MoleculeSet::noTottersTransform();} void readPartialCharges(string fileName) {MoleculeSet::readPartialCharges(fileName);} void setMorganChargesLabels(double threshold) {MoleculeSet::setMorganChargesLabels(threshold);} //R cannot pass C++ function pointers - therfore it is wrapped here bluntly //the function now is a litte more intelligent, because it recognizes whether there is a comparison-set void gramCompute3D (float edgeKernelParameter, bool silentMode, bool external, bool rbf) { double(*atomkernel)(Atom*, Atom*); double(*bondkernel)(float, float, float); if (external) atomkernel = MoleculeUtils::atomKernelExternalMatrix; else atomkernel = MoleculeUtils::atomKernelMorganLabel; if (rbf) bondkernel = MoleculeUtils::threeDedgeKernelRBF; else bondkernel = MoleculeUtils::threeDedgeKernelTriangle; if (comparisonSet == NULL) { MoleculeSet::gramCompute3D( MoleculeUtils::threeDkernel, atomkernel, bondkernel, edgeKernelParameter, silentMode); //sorts out a very ugly hack in chemcpp... comparisonSet = NULL; } else if(comparisonSet == this) { MoleculeSet::gramCompute3D( MoleculeUtils::threeDkernel, atomkernel, bondkernel, edgeKernelParameter, silentMode); } else { MoleculeSet::gramCompute3D(comparisonSet, MoleculeUtils::threeDkernel, atomkernel, bondkernel, edgeKernelParameter, silentMode); } } void gramCompute (double stopP, int converg, int fromN, int nbThreadsWanted, bool silentMode, bool filterTotters, bool external) { double(*atomkernel)(Atom*, Atom*); if (external) atomkernel = MoleculeUtils::atomKernelExternalMatrix; else atomkernel = MoleculeUtils::atomKernelMorganLabel; if (comparisonSet == NULL) { MoleculeSet::gramCompute( stopP, MoleculeUtils::moleculeKernel, atomkernel, MoleculeUtils::bondKernelType, converg, fromN, "", nbThreadsWanted, silentMode, filterTotters ); //sorts out a very ugly hack in chemcpp... comparisonSet = NULL; } else if(comparisonSet == this) { MoleculeSet::gramCompute( stopP, MoleculeUtils::moleculeKernel, atomkernel, MoleculeUtils::bondKernelType, converg, fromN, "", nbThreadsWanted, silentMode, filterTotters ); } else { MoleculeSet::gramCompute(comparisonSet, stopP, MoleculeUtils::moleculeKernel, atomkernel, MoleculeUtils::bondKernelType, converg, fromN, "", nbThreadsWanted, silentMode, filterTotters ); } } void addMoleculeCopy(SEXP s) { std::string rtypename("Rcpp_Rmolecule"); Rcpp::S4 s4obj( s ); if ( !s4obj.is( rtypename.c_str() ) ) { Rf_error( (std::string("object is not of the type ")+rtypename).c_str() ); } Rcpp::Environment env( s4obj ); Rcpp::XPtr xptr( env.get(".pointer") ); Rmolecule *o = static_cast (R_ExternalPtrAddr(xptr)); MoleculeSet::addMoleculeCopy(o); } vector< vector > getGram() {return (*gram);} vector< vector > getGramNormal() {return (*gramNormal);} //currently not expicitly necessary std::vector atomsLabelsListing() {return MoleculeSet::atomsLabelsListing();} std::vector bondsListing() {return MoleculeSet::bondsListing();} void normalizeTanimoto_raw() {MoleculeSet::normalizeTanimoto_raw();} void normalizeGram_raw() {MoleculeSet::normalizeGram_raw();} void writeGramMatrix(string aFileName, bool normal, bool self, bool silentMode) {MoleculeSet::writeGramMatrix( aFileName, normal, self, silentMode );} void writeSelfKernelList( string aFileName, bool silentMode ) {MoleculeSet::writeSelfKernelList(aFileName,silentMode);} //Extension 1.0.7 SEXP getMolByIndex(int anInd) { //Unfortunately the RTLD_GLOBAL is not used by Rs library.dynam() internally, so catching exceptions from the stdcpp library does not work... //So we have to check ourselves... (we are zero-based here) if (anInd < 0 || anInd >= MoleculeSet::numMolecules()) { CError e = CError(MOLECULENOTFOUND, "Index out of range"); e.describe(); throw(e); } Molecule *res = MoleculeSet::getMolByIndex(anInd); //Not nice, but works, because the derived class has no additional members: // cast from base to derived class void* res2 = static_cast(res); Rmolecule* res3 = static_cast(res2); Rcpp::XPtr xp((Rmolecule*)res3, false); //do NOT mark as finalizable! Rcpp::Function maker=Rcpp::Environment::Rcpp_namespace()[ "cpp_object_maker"]; return maker ( typeid(Rmolecule).name() , xp ); } private: }; #endif Rchemcpp/src/chemcpp/0000755000175000017500000000000012116123101015135 5ustar klambaueklambaueRchemcpp/src/chemcpp/src/0000755000175000017500000000000012116123101015724 5ustar klambaueklambaueRchemcpp/src/chemcpp/src/kcfmolecule.cpp0000644000175000017500000000671112116123101020726 0ustar klambaueklambaue/**************************************************************************************** kcfmolecule.cpp ------------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #include "kcfmolecule.h" #include "moleculeutils.h" extern Elements KEGGelements; KCFMolecule::KCFMolecule(){ //cout << "KCFMolecule constructor" << endl; } KCFMolecule::~KCFMolecule(){ } /** reads the molecule description from a KCF file. The internal representation of the molecule is the Atom KEGG atoms, so the function Elements:loadDefinition should have previously loaded the KEGG atom definition types). It is not possible to mix KCF and MOL files. A function readKCFconverted should be written to read KCF file with an internal representation of elements. */ void KCFMolecule::readKCF( string aFileName ) throw( CError ){ // erase all eventually existing atoms in the molecule erase(); // open molfile ifstream inFile; inFile.open( aFileName.c_str(), ios::in ); if( !inFile.good() ){ CError e = CError(FILENOTFOUND, aFileName + " file not found"); e.describe(); throw(e); } MoleculeUtils::readKCFMolecule( *this, inFile ); inFile.close(); } /** write a kcf file */ void KCFMolecule::writeKCF( string aFileName ){ // open kcf file ofstream outFile; outFile.open( aFileName.c_str(), ios::out ); if( !outFile.good() ){ CError e = CError(FILENOTFOUND, aFileName + " could not be created"); e.describe(); throw(e); } MoleculeUtils::writeKCF( *this, outFile ); outFile.close(); } Atom* KCFMolecule::addAtom(string aSymbol) throw( CError ){ /*ToUpper up(std::locale::classic()); ToLower down(std::locale::classic()); std::transform (aSymbol.begin(), aSymbol.begin()+1, aSymbol.begin(), up); if( aSymbol.length() > 1 ){ std::transform (aSymbol.begin()+1, aSymbol.end(), aSymbol.begin()+1, down); } */ string h = aSymbol.substr(0, 1); string r = ""; if( aSymbol.length() > 1 ){ r = aSymbol.substr( 1, aSymbol.length()-1 ); } h = StringUtils::toUpper( h ); r = StringUtils::toLower( r ); aSymbol = h + r; Atom* anAtom = new Atom( *KEGGelements[aSymbol] ); atoms.push_back(anAtom); #ifdef DEBUG //cout << "created atom " << anAtom->toString() << endl; #endif return( anAtom ); } Rchemcpp/src/chemcpp/src/moleculeutils.cpp0000644000175000017500000016742112116123101021331 0ustar klambaueklambaue/**************************************************************************************** moleculeutils.cpp --------------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ //#define DEBUG 1 //#define VERBOSE 1 #include "moleculeutils.h" #include //MoleculeUtils::MoleculeUtils(){ //} //MoleculeUtils::~MoleculeUtils(){ //} /** read the ctab block of a stream (connection table), and add atoms and bonds to aMolecule */ void MoleculeUtils::readMDLCtabBlock( Molecule& aMolecule, ifstream& inFile, bool genericAtomTypeFlag ) throw( CError ) { //cout << "reading connection table " << endl; int linesize = 256; char *line = new char[linesize]; string stringLine; int nbAtoms = 0; int nbBonds = 0; int nbAtomlist = 0; int nbStext = 0; int chiral = 0; int nbProperties = 0; string versionString = ""; int version = NAVALUE; int bondType = 0; Atom* anAtom = NULL; int lineNumber = 0; int j = 0; //read block line if( inFile.eof() ){ stringstream out; out << "MoleculeUtils::readMDLCtabBlock: eof" << endl; CError e(EOFERROR, out.str() ); //e.describe(); throw(e); } inFile.getline( line, linesize-1,'\n' ); stringLine = line; //cout << "# " << stringLine << endl; nbAtoms = atoi( stringLine.substr(0,3).c_str() ); nbBonds = atoi( stringLine.substr(3,3).c_str() ); nbAtomlist = atoi( stringLine.substr(6,3).c_str() ); nbStext = atoi( stringLine.substr(15,3).c_str() ); chiral = atoi( stringLine.substr(9,3).c_str() ); nbProperties = atoi( stringLine.substr(30,3).c_str() ); versionString = StringUtils::rmSpace( stringLine.substr( 33, stringLine.length() - 33 ) ); #ifdef DEBUG cout << "FOUND version " << versionString << endl; #endif if( versionString == "V2000" || versionString == "v2000" ){ aMolecule.setOriginalFormat( MDL2000 ); version = MDL2000; }else if( versionString == "V3000" || versionString == "v3000" ){ aMolecule.setOriginalFormat( MDL3000 ); version = MDL3000; }else if( versionString == "V2000JLP" || versionString == "v2000JLP" ){ aMolecule.setOriginalFormat( MDL2000JLP ); version = MDL2000JLP; }else{ /*stringstream out; out << "MoleculeUtils::readMDLCtabBlock: Bad file, cannot read " << versionString << " files" << endl; CError e(BADFILE, out.str() ); e.describe();*/ cout << "could not find MDL version number, infering this is a MDL 2000 file and continuing" << endl; //throw(e); aMolecule.setOriginalFormat( MDL2000 ); version = MDL2000; } int bondStereo = 0; int bondNotUsed = 0; int bondTopology = 0; int bondReactionCenter = 0; int numRings = 0; if(nbProperties == 999){ nbProperties = 0; }else{ nbProperties--; // nbProperties usually includes the M END line } //version = stringLine.substr(33,6); aMolecule.addStringDescriptor( "formatVersion", versionString, "", "mol file version" ); //cout << nbAtoms << " atoms (56567)" << endl; aMolecule.setHasSSSR(); //while( !inFile.eof() ){ int atomC = 0; bool finished = false; while( lineNumber < nbAtoms + nbBonds + nbAtomlist + nbStext + nbProperties && finished == false ){ lineNumber++; //read line if( inFile.eof() ){ stringstream out; out << "MoleculeUtils::readMDLCtabBlock: eof" << endl; CError e(EOFERROR, out.str() ); //e.describe(); throw(e); } inFile.getline( line, linesize-1,'\n' ); stringLine = line; //cout << " " << stringLine << endl; // data if( nbAtoms > 0 ){ if( j < nbAtoms ){ // atoms //cout << "atom " << lineNumber << ": " << stringLine << endl; atomC++; if( genericAtomTypeFlag == false ){ try{ //cout << "MoleculeUtils::readMDLCtabBlock: " << stringLine.substr(30,3) << endl; anAtom = aMolecule.addAtom( StringUtils::rmSpace( stringLine.substr(30,3) ), false ); }catch( CError E ){ cerr << "->REPLACED WITH H" << endl; anAtom = aMolecule.addAtom( "H", false ); } }else{ //cout << "MoleculeUtils::readMDLCtabBlock: " << stringLine.substr(30,3) << endl; anAtom = new Atom( stringLine.substr(30,3) ); aMolecule.addAtom( anAtom, false , true); } anAtom->setIdInMolecule( atomC ); anAtom->setCoordinates( atof( stringLine.substr(0,10).c_str() ), atof( stringLine.substr(10,10).c_str() ), atof( stringLine.substr(20,10).c_str() ) ); //j++; } if( j >= nbAtoms && j < nbAtoms + nbBonds ) { // bonds //cout << "bond " << lineNumber << ": " << stringLine << endl; bondType = atoi( stringLine.substr(6,3).c_str() ); bondStereo = atoi( stringLine.substr(9,3).c_str() ); bondNotUsed = atoi( stringLine.substr(12,3).c_str() ); bondTopology = atoi( stringLine.substr(15,3).c_str() ); bondReactionCenter = atoi( stringLine.substr(18,3).c_str() ); Bond* newForward = aMolecule.linkAtoms( atoi( stringLine.substr(0,3).c_str() ) - 1, atoi( stringLine.substr(3,3).c_str() ) - 1, bondType, bondStereo, bondNotUsed, bondTopology, bondReactionCenter, false ); Bond* newBackward = newForward->getReverse(); if( bondTopology == 1 ){ // check if the number of rings and the ring ids are appended at the end of line if( version == MDL2000JLP && stringLine.size() > 24 ){ // read the number of rings it is member of numRings = atoi( stringLine.substr( 21, 3 ).c_str() ); // read each ring id, adding the bond membership for( int j = 0; j != numRings; j++ ){ if( stringLine.size() >= 24+j*3 ){ int newID = atoi( stringLine.substr( 24+(j*3),3 ).c_str() ); Ring* aRing = aMolecule.getRingWithID( newID, true ); aRing->addBond( newForward, true ); aRing->addBond( newBackward, true ); aRing->addAtom( newForward->getSource(), true ); aRing->addAtom( newForward->getTarget(), true ); newForward->addRing( aRing ); newBackward->addRing( aRing ); }else{ #ifdef DEBUG cout << "error reading ring membership, will recompute sssr" << endl; #endif aMolecule.resetSSSR(); } } }else{ #ifdef DEBUG cout << "no ring membership info, will recompute sssr" << endl; #endif aMolecule.resetSSSR(); } }else if( bondTopology == 0){ // sssr needs to be recomputed #ifdef DEBUG cout << "one bond's topology is 0, will recompute sssr" << endl; #endif aMolecule.resetSSSR(); } //j++; } if( j >= nbAtoms + nbBonds && j < nbAtoms + nbBonds + nbAtomlist + nbStext ){ // atomlist block and stext block (SKIPPING) } if( j >= nbAtoms + nbBonds + nbAtomlist + nbStext && j < nbAtoms + nbBonds + nbAtomlist + nbStext + nbProperties ){ // properties block (SKIPPING) } //if( j >= nbAtoms + nbBonds + nbAtomlist + nbStext + nbProperties && j < nbAtoms + nbBonds + nbAtomlist + nbStext + nbProperties + 1){ if( j >= nbAtoms + nbBonds + nbAtomlist + nbStext + nbProperties ){ if(stringLine.substr(0,6) != "M END" ){ //cout << "MoleculeUtils::readConnectionTable: WARNING END LINE OF CONNECTION TABLE IS \"" << stringLine.substr(0,6) << "\" WHILE IT SHOULD BE: \"M END\" " << endl; cout << "skiping " << stringLine.substr(0,6) << endl; }else{ //cout << "found M END" << endl; finished = true; } } j++; } } // there can be properties even when nbProperties is == 0 read until meeting M END if( finished == false ){ do{ inFile.getline( line, linesize-1,'\n' ); stringLine = line; if(stringLine.substr(0,6) == "M END" ){ //cout << "Prop finished " << endl; finished = true; }else{ #ifdef VERBOSE cout << " skipping: " << stringLine << endl; #endif } }while( finished == false ); } delete[] line; if( nbAtoms < 1 ){ // this molecule has no structure information, emit error stringstream out; out << "MoleculeUtils::readMDLCtabBlock: no structure information" << endl; CError e(NOSTRUCTURE, out.str() ); //e.describe(); throw(e); } aMolecule.compute(); } void MoleculeUtils::skipMDLEntry( Molecule& aMolecule, ifstream& inFile ) throw( CError ) { //cout << "MoleculeUtils::skipMDLEntry: skiping entry" << endl; int linesize = 256; char *line = new char[linesize]; string stringLine; int lineNumber = 0; int j = 0; bool finished = false; //read block line if( inFile.eof() ){ stringstream out; out << "MoleculeUtils::readMDLCtabBlock: eof" << endl; CError e(EOFERROR, out.str() ); //e.describe(); throw(e); } while( finished == false ){ lineNumber++; //read line if( inFile.eof() ){ stringstream out; out << "MoleculeUtils::readMDLCtabBlock: eof" << endl; CError e(EOFERROR, out.str() ); //e.describe(); throw(e); } inFile.getline( line, linesize-1,'\n' ); stringLine = line; //cout << " " << stringLine << endl; if( stringLine.length() > 0){ if( StringUtils::rmSpace( stringLine.substr(0,4) ) == "$$$$"){ finished = true; } } } delete[] line; } /** read the header block of a stream, and set molecule name to the comment content or to aName if specified */ void MoleculeUtils::readMDLHeaderBlock( Molecule& aMolecule, ifstream& inFile, string aName ) throw( CError ){ int linesize = 256; char *line = new char[linesize]; string stringLine; Descriptor* ds; //Descriptor* di; //Descriptor* df; // read comment block if( inFile.eof() ){ stringstream out; out << "MoleculeUtils::readMDLHeaderBlock: eof" << endl; CError e(EOFERROR, out.str() ); //e.describe(); throw(e); } inFile.getline( line, linesize-1,'\n' ); stringLine = line; #ifdef DEBUG cout << " adding comment descriptor " << endl; #endif ds = aMolecule.setStringDescriptor( "comment", stringLine, "", "", true, true ); //cout << "comment: " << ds->getValue() << endl; //if( stringLine == "" ){ // cout << "comment is empty " << endl; // ds->setEmpty(); //} #ifdef DEBUG cout << "done, setting mol name" << endl; #endif if( aName != "COMMENT" ){ #ifdef DEBUG cout << "filename: " << StringUtils::getFileName( aName ) << endl; //cout << "setting name to " << StringUtils::getNoExtension( StringUtils::getFileName( aName ) ) << endl; #endif aMolecule.setName( StringUtils::getNoExtension( StringUtils::getFileName( aName ) ) ); //aMolecule.setName( StringUtils::getNoExtension( aName ) ); }else{ //cout << "setting name to " << stringLine << endl; aMolecule.setName( stringLine ); } #ifdef DEBUG cout << " adding comment2 descriptor " << endl; #endif if( inFile.eof() ){ stringstream out; out << "MoleculeUtils::readMDLHeaderBlock: eof" << endl; CError e(EOFERROR, out.str() ); //e.describe(); throw(e); } inFile.getline( line, linesize-1,'\n' ); stringLine = line; //cout << "comment2: " << stringLine << endl; ds = aMolecule.setStringDescriptor( "comment2", stringLine, "", "", true, true ); //if( stringLine == "" ){ //cout << "comment2 is empty " << endl; //ds->setEmpty(); //} #ifdef DEBUG cout << "done, adding comment 3 descriptor" << endl; #endif if( inFile.eof() ){ stringstream out; out << "MoleculeUtils::readMDLHeaderBlock: eof" << endl; CError e(EOFERROR, out.str() ); //e.describe(); throw(e); } inFile.getline( line, linesize-1,'\n' ); stringLine = line; if( stringLine.substr( 0, 21 ) == "Copyright by the U.S." ){ aMolecule.setStringDescriptor( "comment3", "", "", "", true, true ); //cout << "comment3 is empty " << endl; //ds->setEmpty(); }else{ aMolecule.setStringDescriptor( "comment3", stringLine, "", "", true, true ); //if( stringLine == "" ){ // ds->setEmpty(); // cout << "comment3 is empty " << endl; //} } // aMolecule.describe(); #ifdef DEBUG cout << "finished reading header " << endl; #endif // comment (skipping line 3) delete[] line; } void MoleculeUtils::writeMDLCtabBlock( Molecule& aMolecule, ofstream& outFile ){ outFile << StringUtils::preFill( StringUtils::toString( aMolecule.numAtoms() ), 3, " " ); outFile << StringUtils::preFill( StringUtils::toString( aMolecule.numBonds() ), 3, " " ); outFile << StringUtils::preFill( "0", 3, " " ); // atom lists outFile << StringUtils::preFill( " ", 3, " " ); // obsolete if( aMolecule.isChiral() ){ outFile << StringUtils::preFill( "1", 3, " " ); // chiral }else{ outFile << StringUtils::preFill( "0", 3, " " ); // not chiral } outFile << StringUtils::preFill( "0", 3, " " ); // stext outFile << StringUtils::preFill( "1", 15, " " ); // stext outFile << " V2000JLP" << endl; // write atom block vector::iterator ait; int id = 1; for( ait = aMolecule.beginAtom(); ait != aMolecule.endAtom(); ait++ ){ (*ait)->setIdInMolecule( id ); outFile << setw(10) << setprecision(4) << setiosflags(ios::fixed | ios::showpoint) << (*ait)->getX(); outFile << setw(10) << setprecision(4) << setiosflags(ios::fixed | ios::showpoint) << (*ait)->getY(); outFile << setw(10) << setprecision(4) << setiosflags(ios::fixed | ios::showpoint) << (*ait)->getZ(); outFile << " "; outFile << StringUtils::fill( (*ait)->getElementSymbol(), 3, " "); // symbol //outFile << StringUtils::fill( (*ait)->getSymbol(), 3, " "); // symbol outFile << StringUtils::preFill( "0", 2, " "); // mass diff outFile << StringUtils::preFill( "0", 3, " "); // charge diff outFile << StringUtils::preFill( "0", 3, " "); // stereo parity outFile << StringUtils::preFill( "0", 3, " "); // hydrogen counts outFile << StringUtils::preFill( "0", 3, " "); // stereocare box outFile << StringUtils::preFill( "0", 3, " "); // valence outFile << StringUtils::preFill( " ", 3, " "); // H0 designator outFile << StringUtils::preFill( " ", 3, " "); // reaction component type outFile << StringUtils::preFill( " ", 3, " "); // reaction component number outFile << StringUtils::preFill( "0", 3, " "); // atom-atom mapping number outFile << StringUtils::preFill( "0", 3, " "); // inversion/retention flag outFile << StringUtils::preFill( "0", 3, " "); // exact change flag outFile << endl; id++; } // write bond block map::iterator bit; for( ait = aMolecule.beginAtom(); ait != aMolecule.endAtom(); ait++ ){ for( bit = (*ait)->beginBond(); bit != (*ait)->endBond(); bit++ ){ if( (*bit).second->hasFlagOriginal() ){ outFile << StringUtils::preFill( (*bit).second->getSource()->getIdInMolecule(), 3, " " ); outFile << StringUtils::preFill( (*bit).first->getIdInMolecule(), 3, " " ); outFile << StringUtils::preFill( (*bit).second->getLabel(), 3, " " ); outFile << StringUtils::preFill( (*bit).second->getStereo(), 3, " "); // bond stereo outFile << StringUtils::preFill( (*bit).second->getNotUsed(), 3, " "); // notUsed field if( aMolecule.hasSSSRDetected() ){ // the molecule has ring information, save it if( (*bit).second->hasRing() ){ outFile << " 1"; // bond topology: 1: ring }else{ outFile << " 2"; // bond topology: 2: chain } }else{ outFile << " 0"; // bond topology: 0 either bond or chain } outFile << StringUtils::preFill( (*bit).second->getReactionCenter(), 3, " "); // reaction center // add the ring information at the end of the line outFile << MoleculeUtils::getRingString( (*bit).second ); outFile << endl; } } } outFile << "M END" << endl; } string MoleculeUtils::getRingString( Bond* aBond ) throw( CError ){ // NNNRRR... where NNN is the number of rings this bond is part of, RRR is the first ring id. other rings membership are appended totalising NNN*RRR columns. Ring id should start with 1. stringstream out; //cout << "OUTPUT " << aBond->numRings() << " RINGS " << endl; try{ out << StringUtils::preFill( aBond->numRings(), 3, " " ); if( aBond->numRings() > 0 ){ vector::iterator ri; for( ri = aBond->beginRing(); ri != aBond->endRing(); ri++ ){ out << StringUtils::preFill( (*ri)->getID(), 3, " " ); } } //cout << out.str() << endl; return( out.str() ); }catch( CError e ){ e.describe(); cout << "MoleculeUtils::getRingString: Error, sssr was not detected, please call Molecule::detectSSSR() before calling me" << endl; throw( e ); } } /** read the non structural data block of a compound in a MDL SDfile WARNING only data entry with a structure > will be considered. entries missing the entry will be ignored */ void MoleculeUtils::readMDLNSDBlock( Molecule& aMolecule, ifstream& inFile ) throw( CError ){ int linesize = 256; char *line = new char[linesize]; string stringLine = ""; string aLabel = ""; string aValue = ""; int aType = STRING; bool nameSet = false; int lc = 0; while( stringLine != "$$$$" ){ lc++; if( inFile.eof() ){ delete[] line; stringstream out; out << "MoleculeUtils::readMDLNSDBlock: eof" << endl; CError e(EOFERROR, out.str() ); //e.describe(); throw(e); } inFile.getline( line, linesize-1,'\n' ); stringLine = line; //cout << "1: " << stringLine << endl; if( stringLine.substr( 0, 1 ) != ">" ){ delete[] line; if( stringLine.substr( 0, 4 ) == "$$$$" ){ return; } // bad formated file or 2 values for one descriptor stringstream out; out << "MoleculeUtils::readMDLNSDBlock: > not found in descriptor " << aLabel; out << " in molecule " << aMolecule.getName() << "; line: "; out << " (the sd import function does not support descriptors with multiple values)" << endl; CError e(BADFILE, out.str() ); e.describe(); throw(e); }else{ // look for <> content uint dls = 0; uint dle = 0; dls = stringLine.find("<", 1); dle = stringLine.find(">", 1); if( dls == string::npos || dls == string::npos ){ stringstream out; out << "MoleculeUtils::readMDLNSDBlock: not found in line " << stringLine << endl; delete[] line; CError e(BADFILE, out.str() ); e.describe(); throw(e); }else{ aLabel = stringLine.substr( dls + 1, (dle - dls) - 1 ); //cout << "found descriptor " << aLabel << endl; if( aLabel.substr( aLabel.length() - min( (int) aLabel.length(), 7), 7 ) == "integer" || aLabel.substr( aLabel.length() - min( (int) aLabel.length(), 4), 4 ) == ".int"){ aLabel = aLabel.substr( 0, aLabel.length() - 8 ); aType = INTEGER; }else if( aLabel.substr( aLabel.length() - min( (int) aLabel.length(), 5), 5 ) == "float" || aLabel.substr( aLabel.length() - min( (int) aLabel.length(), 4), 4 ) == ".flo"){ aLabel = aLabel.substr( 0, aLabel.length() - 6 ); aType = FLOAT; }else if( aLabel.substr( aLabel.length() - min( (int) aLabel.length(),6), 6 ) == "string" || aLabel.substr( aLabel.length() - min( (int) aLabel.length(),4), 4 ) == ".str" ){ aLabel = aLabel.substr( 0, aLabel.length() - 7 ); aType = STRING; }else{ aType = STRING; } } // look for () content uint pls = 0; uint ple = 0; pls = stringLine.find("(", 1); ple = stringLine.find(")", 1); // there is a registry number in parenthesis. Use this name for the compound name if( aType == STRING && pls != string::npos && ple != string::npos && pls != ple ){ aMolecule.setName( stringLine.substr( pls + 1, (ple - pls) - 1 ) ); nameSet = true; } } // read value lines if( inFile.eof() ){ delete[] line; stringstream out; out << "MoleculeUtils::readMDLNSDBlock: eof" << endl; CError e(EOFERROR, out.str() ); //e.describe(); throw(e); } inFile.getline( line, linesize-1,'\n' ); stringLine = line; //cout << "2: " << stringLine << endl; if( aType == STRING ){ Descriptor* d = NULL; if( aLabel == "name" ){ if( nameSet == false ){ aMolecule.setName( stringLine ); nameSet = true; } }else if( aLabel == ACTIVITY ){ aMolecule.setActivity( stringLine ); }else{ d = aMolecule.setStringDescriptor( aLabel, stringLine, "", "", true, true ); if( stringLine == "" ){ d->setEmpty(); } } }else if( aType == INTEGER ) { Descriptor* d = NULL; if( aLabel == ACTIVITY ){ aMolecule.setActivity( stringLine ); }else{ //cout << "adding " << aLabel << endl; d = aMolecule.setIntDescriptor( aLabel, atoi( stringLine.c_str() ), "", "", true, true ); if( stringLine == "" ){ d->setEmpty(); } } }else if( aType == FLOAT ) { Descriptor* d = NULL; if( aLabel == ACTIVITY ){ aMolecule.setActivity( stringLine ); }else{ d = aMolecule.setFloatDescriptor( aLabel, atof( stringLine.c_str() ), "", "", true, true ); if( stringLine == "" ){ d->setEmpty(); } } } // read blank line if( inFile.eof() ){ delete[] line; stringstream out; out << "MoleculeUtils::readMDLNSDBlock: eof" << endl; CError e(EOFERROR, out.str() ); //e.describe(); throw(e); } inFile.getline( line, linesize-1,'\n' ); stringLine = line; //cout << "3: " << stringLine << endl; } delete[] line; } void MoleculeUtils::writeMDLHeaderBlock( Molecule& aMolecule, ofstream& outFile ){ //int flagNameWritten = 0; string s = ""; //aMolecule.describe(); try{ s = aMolecule.getStringDescriptor( "comment" )->getValue( true ); }catch( CError e ){ //e.describe(); //cerr << "error accessing comment" << endl; // comment is empty //try{ // // try get the name // s = aMolecule.getStringDescriptor( "name" )->getValue(); // cout << " writing name instead " << endl; // flagNameWritten++; //}catch( CError e ){ // // name is empty or does not exist // cout << " writing nothing instead " << endl; s = ""; //} } outFile << s << endl; try{ s = aMolecule.getStringDescriptor( "comment2" )->getValue( true ); }catch( CError e ){ //e.describe(); //cerr << "error accessing comment2" << endl; // comment is empty //if( flagNameWritten == 0){ // try{ // // try get the name // s = aMolecule.getStringDescriptor( "name" )->getValue(); // cout << " writing name instead " << endl; // flagNameWritten++; // }catch( CError e ){ // // name is empty or does not exist // s = ""; // cout << " writing nothing instead " << endl; // } //}else{ // cout << " writing nothing instead " << endl; s = ""; //} } outFile << s << endl; try{ s = aMolecule.getStringDescriptor( "comment3" )->getValue( true ); }catch( CError e ){ //e.describe(); //cerr << "error accessing comment3" << endl; // comment is empty //if( flagNameWritten == 0){ // try{ // // try get the name // s = aMolecule.getStringDescriptor( "name" )->getValue(); // cout << " writing name instead " << endl; // flagNameWritten++; // }catch( CError e ){ // // name is empty or does not exist // cout << " writing nothing instead " << endl; // s = ""; // } //}else{ // cout << " writing nothing instead " << endl; s = ""; //} } outFile << s << endl; //if( aMolecule.hasStringDescriptor( "comment" ) ){ // outFile << aMolecule.getStringDescriptor( "comment" )->getValue() << endl; //}else{ // outFile << aMolecule.getStringDescriptor( "name" )->getValue() << endl; // flagNameWritten++; //} //if( aMolecule.hasStringDescriptor( "comment2" ) ){ // outFile << aMolecule.getStringDescriptor( "comment2" )->getValue() << endl; //}else{ // if( flagNameWritten == 0){ // outFile << aMolecule.getStringDescriptor( "name" )->getValue() << endl; // flagNameWritten++; // }else{ // outFile << endl; // } //} //if( aMolecule.hasStringDescriptor( "comment3" ) ){ // outFile << aMolecule.getStringDescriptor( "comment3" )->getValue() << endl; //}else{ // if( flagNameWritten == 0){ // outFile << aMolecule.getStringDescriptor( "name" )->getValue() << endl; // flagNameWritten++; // }else{ // outFile << endl; // } //} //outFile << endl << endl; } /** writes the non structural data block to a stream, for aMolecule this function writes all stringDescriptors, floatDescriptors and IntDescriptors of a molecule in appropriate format */ void MoleculeUtils::writeMDLNSDBlock( Molecule& aMolecule, ofstream& outFile ){ map< string, Descriptor< int >* >::iterator iit; map< string, Descriptor< float >* >::iterator fit; map< string, Descriptor< string >* >::iterator sit; Descriptor< int >* di; Descriptor< float >* df; Descriptor< string >* ds; // if the molecule has activity write the activity descriptor if( aMolecule.hasActivity() ){ outFile << "> <" << "activity>" << endl << aMolecule.getActivity() << endl << endl; } int i; float f; string s; for( iit = aMolecule.beginIntDescriptor(); iit != aMolecule.endIntDescriptor(); iit++ ){ di = (*iit).second; try{ i = di->getValue( true ); // silentError outFile << "> <" << di->getLabel(); outFile << ".integer"; outFile << ">" << endl; outFile << i << endl; outFile << endl; }catch( CError e ){ // descriptor is empty, do nothing } } for( fit = aMolecule.beginFloatDescriptor(); fit != aMolecule.endFloatDescriptor(); fit++ ){ df = (*fit).second; try{ f = df->getValue( true ); // silentError outFile << "> <" << df->getLabel(); outFile << ".float"; outFile << ">" << endl; outFile << f << endl; outFile << endl; }catch( CError e ){ // descriptor is empty, do nothing } } for( sit = aMolecule.beginStringDescriptor(); sit != aMolecule.endStringDescriptor(); sit++ ){ ds = (*sit).second; if( ds->getLabel() != "comment" && ds->getLabel() != "comment2" && ds->getLabel() != "comment3" && ds->getLabel() != "formatVersion" ){ try{ s = ds->getValue( true ); // silentError outFile << "> <" << ds->getLabel(); outFile << ".string"; outFile << ">" << endl; outFile << s << endl; outFile << endl; }catch( CError e ){ // descriptor is empty, do nothing } } } outFile << "$$$$" << endl; } void MoleculeUtils::writeKCF( Molecule& aMolecule, ofstream& outFile ){ outFile << StringUtils::fill( "ENTRY", 12 ) << StringUtils::fill( aMolecule.getName(), 28 ) << "Compound #chiral" << endl; // write ATOMS outFile << StringUtils::fill( "ATOM", 12) << aMolecule.numAtoms() << endl; vector::iterator ait; int i = 1; stringstream istream; string istring = ""; string xstring = ""; string ystring = ""; int minId = (*aMolecule.beginAtom())->getId()-1; for( ait = aMolecule.beginAtom(); ait != aMolecule.endAtom(); ait++ ){ istream << (*ait)->getId() - minId; istream >> istring; istream.clear(); istream << (*ait)->getX(); istream >> xstring; istream.clear(); istream << (*ait)->getY(); istream >> ystring; istream.clear(); outFile << StringUtils::fill(" ", 12) << StringUtils::fill( istring , 4) << StringUtils::fill( (*ait)->getSymbol() , 4) << StringUtils::fill( (*ait)->getStringDescriptor("ElementSymbol")->getValue() , 5) << StringUtils::preFill( xstring , 7) << StringUtils::preFill( ystring , 10) << endl; i++; } // write BONDS outFile << StringUtils::fill( "BOND", 12) << aMolecule.numBonds() << endl; map::iterator aib; i = 1; string source = ""; string target = ""; string label = ""; for( ait = aMolecule.beginAtom(); ait != aMolecule.endAtom(); ait++ ){ for( aib = (*ait)->beginBond(); aib != (*ait)->endBond(); aib++ ){ istream << i; istream >> istring; istream.clear(); istream << (*aib).second->getSource()->getId() - minId; istream >> source; istream.clear(); istream << (*aib).first->getId() - minId; istream >> target; istream.clear(); istream << (*aib).second->getLabel(); istream >> label; istream.clear(); // set flag on back bond if( !(*aib).second->hasFlagOriginal() ){ outFile << StringUtils::fill(" ", 12) << StringUtils::fill( istring , 3) << StringUtils::preFill( source , 4) << StringUtils::preFill( target , 4) << StringUtils::preFill( label , 2) << endl; i++; } } } // write other descriptors MoleculeUtils::writeKCFNSDBlock( aMolecule, outFile ); outFile << "///" << endl; } /** writes the non structural data block to a stream, for aMolecule in KCF format this function writes all stringDescriptors, floatDescriptors and IntDescriptors of a molecule in appropriate format */ void MoleculeUtils::writeKCFNSDBlock( Molecule& aMolecule, ofstream& outFile ){ map< string, Descriptor< int >* >::iterator iit; map< string, Descriptor< float >* >::iterator fit; map< string, Descriptor< string >* >::iterator sit; Descriptor< int >* di = NULL; Descriptor< float >* df = NULL; Descriptor< string >* ds = NULL; int i; float f; string s; string outLabel; for( iit = aMolecule.beginIntDescriptor(); iit != aMolecule.endIntDescriptor(); iit++ ){ di = (*iit).second; try{ i = di->getValue(); outLabel = di->getLabel().substr( 0, 8 ) + ".int"; outFile << StringUtils::fill( outLabel, 12, " " ); outFile << i << endl; }catch( CError e ){ // descriptor is empty, do nothing } } for( fit = aMolecule.beginFloatDescriptor(); fit != aMolecule.endFloatDescriptor(); fit++ ){ df = (*fit).second; try{ f = df->getValue(); outLabel = df->getLabel().substr( 0, 8 ) + ".flo"; outFile << StringUtils::fill( outLabel, 12, " " ); outFile << f << endl; }catch( CError e ){ // descriptor is empty, do nothing } } for( sit = aMolecule.beginStringDescriptor(); sit != aMolecule.endStringDescriptor(); sit++ ){ ds = (*sit).second; if( ds->getLabel() != "comment" && ds->getLabel() != "comment2" && ds->getLabel() != "comment3" && ds->getLabel() != "formatVersion" ){ try{ s = ds->getValue(); outLabel = ds->getLabel().substr( 0, 8 ) + ".str"; outFile << StringUtils::fill( outLabel, 12, " " ); outFile << s << endl; }catch( CError e ){ // descriptor is empty, do nothing } } } //outFile << "$$$$" << endl; } /** read one molecule from a kcf file. returns true if a valid molecule was found (containing at least one atom. */ bool MoleculeUtils::readKCFMolecule( KCFMolecule& m, ifstream& inFile ) throw( CError ){ // COMPLETE *** // erase all eventually existing atoms in the molecule m.erase(); bool valid = false; int linesize = 256; char *line = new char[linesize]; string stringLine; string lineLabel; string lineContent; int readerStatus = FINDENTRY; int nbAtoms = 0; int nbBonds = 0; string KCFType = ""; m.unsetBondFlagsOriginal(); Atom* anAtom = NULL; int lineNumber = 0; // for each line in file int atomC = 0; while( !inFile.eof()){ lineNumber++; //cout << lineNumber << endl; //read line inFile.getline( line, linesize-1,'\n' ); stringLine = line; //cout << stringLine.length() << " " << stringLine << endl; if(stringLine.length()>11){ // parse line lineLabel = StringUtils::rmSpace( stringLine.substr( 0, 11 ) ); lineContent = stringLine.substr( 12, stringLine.length() - 12 ); //cout << "line: " << stringLine << endl; //cout << "label :" << lineLabel << "." << readerStatus << endl; //cout << "content:" << lineContent << "." << readerStatus << endl; if(lineLabel == "ENTRY"){ //cout << "FOUND ENTRY:" << StringUtils::rmTailSpace( lineContent ) << endl; //cout << " setting name to " << StringUtils::rmSpace( StringUtils::field( lineContent, 0, 18 ) ) << endl; m.setName( StringUtils::rmSpace( StringUtils::field( lineContent, 0, 18 ) ) ); KCFType = StringUtils::field( stringLine, 30, 8 ); } //cout << "checking node" << endl; if(lineLabel == "NODE" || lineLabel == "ATOM"){ //cout << "FOUND NODE"<< endl; valid = true; readerStatus = READINGNODES; nbAtoms = atoi( stringLine.substr(12,stringLine.length()-12).c_str() ); //cout << "this molecule has " << nbAtoms << " atoms" << endl; }else if(lineLabel == "EDGE" || lineLabel == "BOND"){ //cout << "FOUND EDGE" << endl; readerStatus = READINGEDGES; nbBonds = atoi( stringLine.substr(12,stringLine.length()-12).c_str() ); //cout << "this molecule has " << nbBonds << " bonds" << endl; }else if( lineLabel == "" ){ //cout << "label is empty" << endl; if( readerStatus == READINGNODES ){ //cout << "p1" << endl; if(lineLabel!=""){ readerStatus = FINDENTRY; }else{ // read node //cout << "MoleculeUtils::readKCFMolecule: adding atom " << StringUtils::rmSpace( stringLine.substr(16,3) ) << endl; atomC++; anAtom = m.addAtom( StringUtils::rmSpace( stringLine.substr(16,3) ) ); anAtom->setIdInMolecule( atomC ); //cout << " coordinates: " << atof( stringLine.substr(23,9).c_str() ) << "; " << atof( stringLine.substr(33,9).c_str() ) << endl; anAtom->setCoordinates( atof( stringLine.substr(23,9).c_str() ), atof( stringLine.substr(33,9).c_str() ), 0 ); anAtom->setElementSymbol( stringLine.substr(20,2) ); //cout << anAtom->getX() << " " << anAtom->getY() << endl; } }else if( readerStatus == READINGEDGES ){ //cout << "p2" << endl; if(lineLabel!=""){ readerStatus = FINDENTRY; }else{ // read edge //cout << "linking " << atoi( stringLine.substr(15,3).c_str() ) - 1 << " -" << atoi( stringLine.substr(23,1).c_str() ) << "- " << atoi( stringLine.substr(19,3).c_str() ) - 1 << endl; m.linkAtoms( atoi( stringLine.substr(16,3).c_str() ) - 1, atoi( stringLine.substr(20,3).c_str() ) - 1, atoi( stringLine.substr(24,1).c_str() ) ); //cout << "done" << endl; } }else{ //cout << "p3" << endl; //cout << "doing nothing" << endl; } //cout << "p4" << endl; }else{ // unknown tag met if( lineLabel != "" && lineLabel != "ENTRY" ){ //cout << "unknown tag: " << lineLabel << endl; //cout << "adding descriptor " << lineLabel << " with value " << lineContent << endl; m.addUnknownTypeDescriptor( lineLabel, lineContent ); readerStatus = FINDENTRY; }else if( lineLabel != "ENTRY" ){ cout << "skipping unlabeled descriptor with value " << lineContent << endl; } } }else{ if(stringLine.substr(0,3) == "///"){ //cout << "END OF MOLECULE " << endl; if( valid ){ int i = m.hideHydrogens(); if( i > 0 ){ cout << i << " H atoms hidden" << endl; } } return( valid ); }else if( stringLine.substr(0,2) == "//" ){ }else if( stringLine.substr(0,4) == "$$$$" ){ }else{ if( stringLine!="" ){ cout << "skipping short line: " << stringLine << endl; } } } } if( valid ){ cout << m.hideHydrogens() << " H atoms hidden" << endl; } delete[] line; m.compute(); return( valid ); } double MoleculeUtils::moleculeKernel( Molecule* mol1, Molecule* mol2, double (*pt2AtomKernel)(Atom*, Atom*), double (*pt2BondKernel)(Bond*, Bond*), int convergenceCondition, int parameter2 ){ // we need an array containing values calculated for each atom pairs among the two molecules #ifdef DEBUG cout << "MoleculeUtils::moleculeKernel " << convergenceCondition << endl; #endif if( convergenceCondition <= 0 ){ stringstream out; out << "MoleculeUtils::moleculeKernel: Invalid convergence condition: should be > 0 and it is = " << convergenceCondition << endl; CError e( VALUENOTALLOWED, out.str() ); e.describe(); throw(e); } vector< vector >* r = new vector< vector >; vector< vector >* rwork = new vector< vector >; vector< vector >* rstart = new vector< vector >; // contains the values at length parameter2 //vector< vector >* rFinal; vector::iterator atom; vector::iterator otherAtom; // initial Array filled with the probability to die at the first step int i = 0; int j = 0; //cout << "initializing matrix" << endl; for( atom = mol1->beginAtom(); atom != mol1->endAtom(); atom++ ){ // add each line //cout << "MoleculeUtils::moleculeKernel: i = " << i << endl; r->push_back( vector() ); rwork->push_back( vector() ); rstart->push_back( vector() ); j = 0; for( otherAtom = mol2->beginAtom(); otherAtom != mol2->endAtom(); otherAtom++ ){ //cout << "MoleculeUtils::moleculeKernel: j = " << j << endl; // fill line with the values for each column try{ //if( parameter2 < 2 ){ (*r)[i].push_back( (*atom)->getKashimaPQ() * (*otherAtom)->getKashimaPQ() ); //}else{ //cout << "gghgh 0 " << endl; //(*r)[i].push_back( 0 ); //} }catch(CError e){ cout << "MoleculeUtils::graphKernel(): random walk probability was not set" <setRPosition( j ); //cout << "MoleculeUtils::moleculeKernel: F3" << endl; j++; } (*atom)->setRPosition( i ); i++; } //cout << "MoleculeUtils::moleculeKernel: start rlk" << endl; // calcul de la somme des probabilites de tous les chemins possibles communs aux deux molecules partant de tous les noeuds MoleculeUtils::rlk( r, rwork, rstart, mol1, mol2, convergenceCondition, parameter2, pt2AtomKernel, pt2BondKernel, 1 ); //cout << "MoleculeUtils::moleculeKernel: rlk done" << endl; double s = 0.0; double k = 0.0; i = 0; //int localAN = 0; double localPS = 0; for( atom = mol1->beginAtom(); atom != mol1->endAtom(); atom++ ){ j = 0; //localAN = (*atom)->getAN()-1; //localAN = (*atom)->getType(); localPS = (*atom)->getKashimaPS(); for( otherAtom = mol2->beginAtom(); otherAtom != mol2->endAtom(); otherAtom++ ){ s = pt2AtomKernel( (*atom), (*otherAtom) ); if( s != 0 ){ if( parameter2 < 2 ){ s = localPS * (*otherAtom)->getKashimaPS() * s; k = k + ( s * (*rwork)[i][j] ); }else{ k = k + ( (*rwork)[i][j] ); } } j++; } i++; } delete r; delete rstart; delete rwork; return(k); } vector< vector >* MoleculeUtils::rlk( vector< vector >* r, vector< vector >* rwork, vector< vector >* rstart, Molecule* mol1, Molecule* mol2, int convergenceCondition, int parameter2, double (*pt2AtomKernel)(Atom*, Atom*), double (*pt2BondKernel)(Bond*, Bond*), int depth ){ //cout << "MOLECULE RLK" << endl; vector::iterator atom; vector::iterator otherAtom; map::iterator bond; map::iterator otherBond; double q = 0; double t = 0; int i = 0; int j = 0; Bond* bondSecond = NULL; Atom* bondFirst = NULL; double bondSecondPT = 0; int bondFirstRP = 0; #ifdef DEBUG cout << "MoleculeUtils::rlk( depth = " << depth << " )" << endl; #endif //if( depth >= parameter2 ){ //cout << "MoleculeUtils::rlk 1" << endl; //cout << "molecules have " << mol1->numAtoms() << " and " << mol2->numAtoms() << " atoms " << endl; // pour tous les atomes deux a deux for( atom = mol1->beginAtom(); atom != mol1->endAtom(); atom++ ){ j = 0; for( otherAtom = mol2->beginAtom(); otherAtom != mol2->endAtom(); otherAtom++ ){ //if( parameter2 < 2 ){ q = (*atom)->getKashimaPQ() * (*otherAtom)->getKashimaPQ(); //}else{ //if( depth >= parameter2 ){ // q = (*atom)->getKashimaPQ() * (*otherAtom)->getKashimaPQ(); //}else{ // q = 0; //} //} //cout << "A" << endl; t = 0; // pour toutes les liaisons des atomes deux a deux for( bond = (*atom)->beginBond(); bond != (*atom)->endBond(); bond++ ){ //cout << "B" << endl; bondSecond = (*bond).second; bondFirst = (*bond).first; //cout << bondSecond << endl; //cout << "C" << endl; bondSecondPT = bondSecond->getKashimaPT(); //cout << "D" << endl; bondFirstRP = bondFirst->getRPosition(); //cout << "E" << endl; for( otherBond = (*otherAtom)->beginBond(); otherBond != (*otherAtom)->endBond(); otherBond++ ){ t = t + bondSecondPT * (*otherBond).second->getKashimaPT() * pt2BondKernel( bondSecond, (*otherBond).second ) * pt2AtomKernel( bondFirst, (*otherBond).first ) * (*r)[ bondFirstRP ][ (*otherBond).first->getRPosition() ]; } } (*rwork)[i][j] = q + t; j++; } i++; } if( parameter2 > 1 && depth == parameter2-1 ){ //cout << "SUBSTR RLK at depth " << depth << endl; i = 0; for( atom = mol1->beginAtom(); atom != mol1->endAtom(); atom++ ){ j = 0; for( otherAtom = mol2->beginAtom(); otherAtom != mol2->endAtom(); otherAtom++ ){ (*rstart)[i][j] = (*rwork)[i][j]; j++; } i++; } } //cout << "MoleculeUtils::rlk 2" << endl; // if r and rwork did not converge yet, make another recursion //if( depth >= parameter2 ){ if( MoleculeUtils::converge( r, rwork, mol1, mol2, convergenceCondition ) || depth >= 50 ) { }else{ //cout << "MoleculeUtils::rlk 3" << endl; //if( t > 0 ){ MoleculeUtils::rlk( rwork, r, rstart, mol1, mol2, convergenceCondition, parameter2, pt2AtomKernel, pt2BondKernel, depth + 1 ); //}else{ // cout << "i stopped at depth " << depth << endl; //} }//else{ // cout << "stopped at depth " << depth << endl; //} //}else{ // MoleculeUtils::rlk( rwork, r, mol1, mol2, convergenceCondition, parameter2, pt2AtomKernel, pt2BondKernel, depth + 1 ); //} //}else{ // MoleculeUtils::rlk( rwork, r, mol1, mol2, convergenceCondition, parameter2, pt2AtomKernel, pt2BondKernel, depth + 1 ); //} if( parameter2 > 1 && depth == 1){ i = 0; for( atom = mol1->beginAtom(); atom != mol1->endAtom(); atom++ ){ j = 0; for( otherAtom = mol2->beginAtom(); otherAtom != mol2->endAtom(); otherAtom++ ){ //cout << (*rwork)[i][j] << " - " << (*rstart)[i][j] << endl; (*rwork)[i][j] = (*rwork)[i][j] - (*rstart)[i][j]; j++; } i++; } } return( rwork ); } inline bool MoleculeUtils::converge( vector< vector >* r1, vector< vector >* r2, Molecule* mol1, Molecule* mol2, int convergenceCondition ) { vector::iterator atom; vector::iterator otherAtom; int i = 0; int j = 0; for( atom = mol1->beginAtom(); atom != mol1->endAtom(); atom++ ){ j = 0; for( otherAtom = mol2->beginAtom(); otherAtom != mol2->endAtom(); otherAtom++ ){ if( fabs ( (*r2)[i][j] - (*r1)[i][j] ) > (*r1)[i][j] / convergenceCondition ){ #ifdef DEBUG cout << " MoleculeUtils::converge: " << fabs ( (*r2)[i][j] - (*r1)[i][j] ) << " >? " << (*r1)[i][j] / convergenceCondition << endl; #endif return( false ); } j++; } i++; } return( true ); } double MoleculeUtils::atomKernelSymbol( Atom* a1, Atom* a2 ){ if( a1->getAN() == a2->getAN() ) { return 1.0; }else{ return 0.0; } } double MoleculeUtils::atomKernelMorganLabel( Atom* a1, Atom* a2 ){ if( a1->getMorganLabel() == a2->getMorganLabel() ) { return 1.0; }else{ return 0.0; } } double MoleculeUtils::atomKernelPerretLabel( Atom* a1, Atom* a2 ){ //#ifdef DEBUG // cout << "MoleculeUtils::atomKernelPerretLabel" << endl; //#endif if( a1->getPerretLabel() == a2->getPerretLabel() ) { return 1.0; }else{ return 0.0; } } double MoleculeUtils::atomKernelPerretLabelExternalMatrix( Atom* a1, Atom* a2 ){ string l1 = a1->getPerretLabel(); string l2 = a2->getPerretLabel(); if( l1 == l2 ) { return 1.0; }else{ if( l1 == "CJ" || l1 == "CK" || l2 == "CJ" || l2 == "CK" ){ return 0.0; }else{ return( elements.gramAtom[ a1->getType() ][ a2->getType() ] ); } } } double MoleculeUtils::atomKernelExternalMatrix( Atom* a1, Atom* a2 ){ return( elements.gramAtom[ a1->getType() ][ a2->getType() ] ); } /*! \fn MoleculeUtils::atomKernelLabel( Atom* a1, Atom*, a2 ) returns 1 if two atoms have the same label, 0 othwerwise */ double MoleculeUtils::atomKernelLabel( Atom* a1, Atom* a2 ){ if( a1->getLabel() == a2->getLabel() ) { return 1.0; }else{ return 0.0; } } double MoleculeUtils::bondKernelType( Bond* b1, Bond* b2 ){ if( b1->getLabel() == b2->getLabel() ) { return 1.0; }else{ return 0.0; } } double MoleculeUtils::bondKernelRotable( Bond* b1, Bond* b2 ){ if( b1->getLabel() == SINGLEBOND || b2->getLabel() == SINGLEBOND ){ if( b1->getLabel() == b2->getLabel() ) { return 1.0; }else{ return 0.0; } }else{ return 1.0; } } double MoleculeUtils::bondKernelPerretLabelStrict( Bond* b1, Bond* b2 ){ if( b1->getPerretLabel() == b2->getPerretLabel() ) { return 1.0; }else{ return 0.0; } } double MoleculeUtils::bondKernelPerretLabel( Bond* b1, Bond* b2 ){ //#ifdef DEBUG // cout << "MoleculeUtils::bondKernelPerretLabel(...)" << endl; //#endif bool b1IsCycle = b1->hasRing(); bool b2IsCycle = b2->hasRing(); int b1Label = b1->getPerretLabel(); int b2Label = b2->getPerretLabel(); //cout << "rrh comparing bond " << b1->toStringShort() << " with " << b2->toStringShort() << endl; double result = -1; if( b1IsCycle != b2IsCycle ){ // one bond is part of a cycle, the other not, return 0 //return( 0 ); result = 0; }else{ if( b1IsCycle == true ){ // both bonds are member of a cycle switch( b1Label ){ case AROMATICBOND: switch( b2Label ){ case AROMATICBOND: result = 1.0; break; case SINGLECYCLEBOND: result = 0.99; break; case DOUBLECYCLEBOND: result = 0.25; break; case TRIPLECYCLEBOND: result = 0.125; break; } break; case SINGLECYCLEBOND: switch( b2Label ){ case AROMATICBOND: result = 0.99; break; case SINGLECYCLEBOND: result = 1.0; break; case DOUBLECYCLEBOND: result = 0.5; break; case TRIPLECYCLEBOND: result = 0.25; break; } break; case DOUBLECYCLEBOND: switch( b2Label ){ case AROMATICBOND: result = 0.25; break; case SINGLECYCLEBOND: result = 0.5; break; case DOUBLECYCLEBOND: result = 1.0; break; case TRIPLECYCLEBOND: result = 0.5; break; } break; case TRIPLECYCLEBOND: switch( b2Label ){ case AROMATICBOND: result = 0.125; break; case SINGLECYCLEBOND: result = 0.25; break; case DOUBLECYCLEBOND: result = 0.5; break; case TRIPLECYCLEBOND: result = 1.0; break; } break; } }else{ // both bonds are aliphatic switch( b1Label ){ case SINGLEBOND: switch( b2Label ){ case SINGLEBOND: result = 1.0; break; case DOUBLEBOND: result = 0.5; break; case TRIPLEBOND: result = 0.25; break; } break; case DOUBLEBOND: switch( b2Label ){ case SINGLEBOND: result = 0.5; break; case DOUBLEBOND: result = 1.0; break; case TRIPLEBOND: result = 0.5; break; } break; case TRIPLEBOND: switch( b2Label ){ case SINGLEBOND: result = 0.25; break; case DOUBLEBOND: result = 0.5; break; case TRIPLEBOND: result = 1.0; break; } break; } } } //cout << " = " << result << endl; /*if( result == -1 ){ cout << "rrh comparing bond " << b1->toStringShort() << " with " << b2->toStringShort() << endl; cout << b1IsCycle << " " << b2IsCycle << endl; exit(1); }*/ return( result ); /*if( b1->getPerretLabel() == b2->getPerretLabel() ) { return 1.0; }else{ return 0.0; }*/ } void MoleculeUtils::writeDOTGraph( Molecule& aMolecule, ofstream& outFile, bool perretLabels ){ //cout << "MoleculeUtils::writeDOTGraph" << endl; //cout << "digraph " << aMolecule.getName() << " {" << endl; //cout << " size = \"4,4\";" << endl; outFile << "digraph \"" << aMolecule.getName() << "\" {" << endl; outFile << " size = \"4,4\";" << endl; Atom* source; Bond* bond; Atom* target; // write atoms vector::iterator ai; map::iterator bi; for( ai = aMolecule.beginAtom(); ai != aMolecule.endAtom(); ai++ ){ source = (*ai); outFile << " " << source->getElementSymbol() << "_" << source->getId(); //outFile << " [ label = \"" << source->getElementSymbol(); outFile << " [ label = \"" << source->getMorganLabel(); if( perretLabels == true ){ outFile << "/" << source->getPerretLabel(); } //try{ //outFile << "/" << source->getKashimaPS(); //outFile << "/" << source->getMorganLabel(); //} catch( CError e){ // continue; //} //try{ // outFile << "/" << source->getKashimaPQ(); //} catch( CError e){ // continue; //} outFile << "\"]" << endl; } // write bonds for( ai = aMolecule.beginAtom(); ai != aMolecule.endAtom(); ai++ ){ source = (*ai); for( bi = source->beginBond(); bi != source->endBond(); bi++ ){ target = (*bi).first; bond = (*bi).second; outFile << " " << source->getElementSymbol() << "_" << source->getId(); outFile << "->" << target->getElementSymbol() << "_" << target->getId(); outFile << " [label = \"" << bond->getLabel(); //try{ // outFile << "/" << bond->getKashimaPT(); //} catch( CError e){ // continue; //} if( perretLabels == true ){ outFile << "/" << bond->getPerretLabel(); } outFile << "\"]" << endl; } } outFile << "}" << endl; } /** returns the graph kernel value between two molecules using the Ecole des Mine approach, based on the fused graph sum of matrix power. The kernel computes the sum of te probabilities of finite length paths, as opposed to infinite length path as in Kashima's original paper */ double MoleculeUtils::powerKernelUntilN( Molecule* mol1, Molecule* mol2, double (*pt2AtomKernel)( Atom*, Atom* ), double (*pt2BondKernel)(Bond*, Bond*), int maxPower, int minLength ) throw( CError ){ //cout << "MoleculeUtils::powerKernelUntilN " << minLength << " -> " << maxPower << endl; if( maxPower < 0 ){ stringstream out; out << "MoleculeUtils::powerKernelUntilN: bad number of itterations: " << maxPower << " should be integer > 0 " << endl; CError e(BADVALUE, out.str() ); e.describe(); throw(e); } // make the fused graph Molecule* fused = new Molecule( *mol1, *mol2, pt2AtomKernel, pt2BondKernel ); // compute the sum of the probabilities for the first maxPower powers of the transition matrix double kernel = 0; if( minLength < 2 ){ kernel = fused->sumPQPSFast(); //cout << "START PROB" << endl; } if( maxPower > 0 ){ int i = 0; //cout << "rrr " << endl; //cout << "---" << endl; i++; while( i < maxPower ){ //cout << "power " << i << endl; if( i >= minLength ){ //cout << "USING LENGTH " << i << endl; kernel += fused->sumProbabilitiesFast(); } fused->raisePowerFast(); i++; } kernel += fused->sumProbabilitiesFast(); //cout << "USING FINAL LENGTH " << i << endl; } delete fused; return( kernel ); } double MoleculeUtils::powerKernelOrderN( Molecule* mol1, Molecule* mol2, double (*pt2AtomKernel)( Atom*, Atom* ), double (*pt2BondKernel)(Bond*, Bond*), int anOrder, int parameter2 ) throw( CError ) { //cout << "MoleculeUtils::powerKernelOrderN " << anOrder << endl; if( anOrder < 0 ){ stringstream out; out << "MoleculeUtils::powerKernelOrderN: bad order: " << anOrder << " should be integer > 0 " << endl; CError e(BADVALUE, out.str() ); e.describe(); throw(e); } double kernel = 0; // make the fused graph Molecule* fused = new Molecule( *mol1, *mol2, pt2AtomKernel, pt2BondKernel ); if( anOrder == 0 ){ kernel = fused->sumPQPSFast(); }else{ // compute the transition probabilities for paths of order anOrder int i = 0; //fused->saveAllBonds(); //kernel = fused->sumProbabilities(); //kernel = fused->sumPQPS(); i++; while( i < anOrder ){ //cout << "rrr" << endl; fused->raisePowerFast(); //cout << "---" << endl; i++; } kernel += fused->sumProbabilitiesFast(); } delete fused; return( kernel ); } double MoleculeUtils::powerKernelConverge( Molecule* mol1, Molecule* mol2, double (*pt2AtomKernel)( Atom*, Atom* ), double (*pt2BondKernel)(Bond*, Bond*), int converge, int minLength ) throw( CError ){ //cout << "MoleculeUtils::powerKernelConverge " << converge << endl; //<< mol1->toStringShort() << " / " << mol2->toStringShort() << endl; if( converge <= 0 ){ stringstream out; out << "MoleculeUtils::powerKernelUntilN: bad number of itterations: " << converge << " should be > 0 " << endl; CError e(VALUENOTALLOWED, out.str() ); e.describe(); throw(e); } double okernel = 0; double condition = 1.0/converge; // make the fused graph Molecule* fused = new Molecule( *mol1, *mol2, pt2AtomKernel, pt2BondKernel ); //Molecule* fused = new Molecule( *mol1 ); //fused->describeLong(); // compute the sum of the probabilities for the first maxPower powers of the transition matrix //fused->saveAllBonds(); double kernel = 0; if( minLength < 2 ){ kernel = fused->sumPQPSFast(); } int i = 0; //cout << "rrr " << endl; //cout << "---" << endl; i++; while( kernel - okernel > condition || i < minLength+1 ){ //cout << "power " << i << endl; okernel = kernel; if( i >= minLength ){ //cout << "USING LENGTH " << i << endl; kernel += fused->sumProbabilitiesFast(); } fused->raisePowerFast(); i++; } //cout << "stopped at " << i << " " << minLength+1 << endl; kernel += fused->sumProbabilitiesFast(); //cout << "USING FINAL LENGTH " << i << endl; delete fused; return( kernel ); } /** No descriptions */ void MoleculeUtils::describeMap( map* aMap ){ map::iterator i; for( i = aMap->begin(); i != aMap->end(); i++ ){ cout << (*i).first->toStringShort() << " " << (*i).second << endl; } } /** helper function for detectSSSR. returns in result a vector containing atom pointer for atoms present in full but not in exclude */ void MoleculeUtils::substractSet( vector* full, vector* exclude, vector* result){ result->clear(); vector::iterator ai; for ( ai = full->begin(); ai != full->end(); ai++ ){ vector::iterator aj; bool found = false; for ( aj = exclude->begin(); aj != exclude->end(); aj++ ){ if( (*aj) == (*ai) ){ #ifdef DEBUG cout << "a10h excluding " << (*ai)->toStringShort() << endl; #endif found = true; break; } } if( found == false ){ result->push_back( (*ai) ); } } } void MoleculeUtils::mergeSet( vector* v1, vector* v2, vector* result){ vector::iterator ai; for ( ai = v1->begin(); ai != v1->end(); ai++ ){ result->push_back( (*ai) ); } for ( ai = v2->begin(); ai != v2->end(); ai++ ){ vector::iterator aj; bool found = false; for ( aj = result->begin(); aj != result->end(); aj++ ){ if( (*aj) == (*ai) ){ found = true; break; } } if( found == false ){ result->push_back( (*ai) ); } } } void MoleculeUtils::selectRingMemberBonds( vector* v1, vector* a, vector* v2 ){ vector::iterator bi; for ( bi = v1->begin(); bi != v1->end(); bi++ ){ #ifdef DEBUG cout << "ffh verifying bond " << (*bi)->toStringShort() << endl; #endif Atom* source = (*bi)->getSource(); Atom* target = (*bi)->getTarget(); bool foundSource = false; bool foundTarget = false; vector::iterator ai; for( ai = a->begin(); ai != a->end(); ai++ ){ #ifdef DEBUG cout << "ffh with " << (*ai)->toStringShort(); #endif if( (*ai) == source ){ //cout << " found source "; foundSource = true; } if( (*ai) == target ){ //cout << " found target "; foundTarget = true; } if( foundSource == true && foundTarget == true ){ v2->push_back( (*bi) ); #ifdef DEBUG cout << " member " << endl; #endif break; } //cout << endl; } #ifdef DEBUG if( foundSource == false || foundTarget == false ){ cout << " non member" << endl; } #endif } } void MoleculeUtils::mergeBondSet( vector* v1, vector* v2, vector* result){ #ifdef DEBUG cout << "MoleculeUtils::mergeBondSet" << endl; #endif vector::iterator ai; for ( ai = v1->begin(); ai != v1->end(); ai++ ){ #ifdef DEBUG cout << (*ai)->toStringShort() << endl; #endif vector::iterator aj; bool found = false; for ( aj = result->begin(); aj != result->end(); aj++ ){ if( (*aj)->getSource() == (*ai)->getSource() && (*aj)->getTarget() == (*ai)->getTarget() ){ found = true; break; } } if( found == false ){ result->push_back( (*ai) ); #ifdef DEBUG cout << " added 1" << endl; #endif } #ifdef DEBUG else{ cout << " skip" << endl; } #endif } for ( ai = v2->begin(); ai != v2->end(); ai++ ){ vector::iterator aj; bool found = false; #ifdef DEBUG cout << (*ai)->toStringShort() << endl; #endif for ( aj = result->begin(); aj != result->end(); aj++ ){ if( (*aj)->getSource() == (*ai)->getSource() && (*aj)->getTarget() == (*ai)->getTarget() ){ found = true; break; } } if( found == false ){ result->push_back( (*ai) ); #ifdef DEBUG cout << " added 2" << endl; #endif } #ifdef DEBUG else{ cout << " skip" << endl; } #endif } } bool MoleculeUtils::atomVectorHas( vector* atomVector, Atom* anAtom ){ vector::iterator ai; for( ai = atomVector->begin(); ai != atomVector->end(); ai++ ){ if( (*ai) == anAtom ){ return( true ); } } return( false ); } // THREE-D KERNELS // //-----------------// double MoleculeUtils::threeDkernel( Molecule* mol1, Molecule* mol2, double (*pt2AtomKernel)( Atom*, Atom*), double (*pt2BondKernel)(float, float, float), float edgeKernelParameter ) { double kernel = 0; // make the fused graph Molecule* fused = new Molecule( *mol1, *mol2, pt2AtomKernel, pt2BondKernel, edgeKernelParameter); #ifdef DEBUG cout << "fused->numAtoms() : " << fused->numAtoms() << endl; cout << "ADJACENCY MATRIX"; for(int i = 0 ; i < fused->numAtoms() ; i++){ cout << endl; for(int j = 0 ; j < fused->numAtoms() ; j++){ cout << fused->getAdjacency(i,j) << "\t"; } } cout << endl << endl; #endif // 1st possibility : compute Ax^3 and get its trace // --> NOT OPTIMUM : we don't need all the cubic matrix //fused->raisePowerAdjacency(); // --> fused->raisePowerAdjacency_sparse() //fused->raisePowerAdjacency(); // --> fused->raisePowerAdjacency_sparse() //kernel = fused->traceWalks(); // --> fused->traceWalks_sparse() // 2nd possibility : compute Ax^2 and only the diagonal of Ax^3 fused->raisePowerAdjacency(); kernel = fused->traceDiagWalks(); #ifdef DEBUG cout << "CUBE MATRIX"; for(int i = 0 ; i < fused->numAtoms() ; i++){ cout << endl; for(int j = 0 ; j < fused->numAtoms() ; j++){ cout << fused->getWalks(i,j) << "\t"; } } cout << endl << endl << endl; #endif fused->eraseAdjacency(); fused->eraseWalks(); delete fused; return(kernel); } double MoleculeUtils::threeDedgeKernelRBF( float dist1, float dist2 , float param ){ // RBF kernel based on the difference between the edges distances : exp( -(d1-d2)^2 / 2*sigma^2) // param = sigma (RBF bandwidth) double res; res = exp( -0.5 * (dist1-dist2) * (dist1-dist2) * (1/(param*param)) ); return res; } double MoleculeUtils::threeDedgeKernelTriangle( float dist1, float dist2 , float param ){ // triangular kernel : k(d1,d2) = max( 0 , (C - |d1-d2|)/C ) // Rq : 0 <= k(d1,d2) <= 1 // param = C = maximum difference tolerated double tmp; tmp = (param - fabs(dist1 - dist2)) / param; if (tmp > 0.0) return tmp; else return 0.0; } Rchemcpp/src/chemcpp/src/moleculeset.cpp0000644000175000017500000032452712116123101020766 0ustar klambaueklambaue/**************************************************************************************** moleculeset.cpp ------------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ //#define DEBUG 1 #include "moleculeset.h" #include "constant.h" #include #include #include //#include // old #include namespace std { template charT toupper (charT c, const locale& loc) ; template charT tolower (charT c, const locale& loc) ; } struct AscendingOrder { bool operator()(Molecule* const & m1, Molecule* const & m2) { string aName = m1->getSortDescriptorName(); int aType = m1->getSortDescriptorType(); if( aType == DEFAULT ){ return( m1->getId() < m2->getId() ); }else if( aType == INTEGER ){ int v1, v2; try{ v1 = m1->getIntDescriptor( aName )->getValue(); }catch( CError e ){ return( false ); } try{ v2 = m2->getIntDescriptor( aName )->getValue(); }catch( CError e ){ return( true ); } return( v1 < v2 ); //return( m1->getIntDescriptor( aName )->getValue() < m2->getIntDescriptor( aName )->getValue() ); }else if( aType == FLOAT ){ float v1, v2; try{ v1 = m1->getFloatDescriptor( aName )->getValue(); }catch( CError e ){ return( false ); } try{ v2 = m2->getFloatDescriptor( aName )->getValue(); }catch( CError e ){ return( true ); } return( v1 < v2 ); //return( m1->getFloatDescriptor( aName )->getValue() < m2->getFloatDescriptor( aName )->getValue() ); }else if( aType == STRING ){ string v1, v2; try{ v1 = m1->getStringDescriptor( aName )->getValue(); }catch( CError e ){ return( false ); } try{ v2 = m2->getStringDescriptor( aName )->getValue(); }catch( CError e ){ return( true ); } return( v1 < v2 ); //return( m1->getStringDescriptor( aName )->getValue() < m2->getStringDescriptor( aName )->getValue() ); } return( false ); } }; struct DescendingOrder { bool operator()(Molecule* const & m2, Molecule* const & m1) { string aName = m1->getSortDescriptorName(); int aType = m1->getSortDescriptorType(); if( aType == DEFAULT ){ return( m1->getId() < m2->getId() ); }else if( aType == INTEGER ){ int v1, v2; try{ v1 = m1->getIntDescriptor( aName )->getValue(); }catch( CError e ){ return( false ); } try{ v2 = m2->getIntDescriptor( aName )->getValue(); }catch( CError e ){ return( true ); } return( v1 < v2 ); //return( m1->getIntDescriptor( aName )->getValue() < m2->getIntDescriptor( aName )->getValue() ); }else if( aType == FLOAT ){ float v1, v2; try{ v1 = m1->getFloatDescriptor( aName )->getValue(); }catch( CError e ){ return( false ); } try{ v2 = m2->getFloatDescriptor( aName )->getValue(); }catch( CError e ){ return( true ); } return( v1 < v2 ); //return( m1->getFloatDescriptor( aName )->getValue() < m2->getFloatDescriptor( aName )->getValue() ); }else if( aType == STRING ){ string v1, v2; try{ v1 = m1->getStringDescriptor( aName )->getValue(); }catch( CError e ){ return( false ); } try{ v2 = m2->getStringDescriptor( aName )->getValue(); }catch( CError e ){ return( true ); } return( v1 < v2 ); //return( m1->getStringDescriptor( aName )->getValue() < m2->getStringDescriptor( aName )->getValue() ); } return( false ); } }; struct AscendingMW { bool operator()(Molecule* const & m1, Molecule* const & m2){ return( m1->getMW() < m2->getMW() ); } }; struct AscendingNumAtoms { bool operator()(Molecule* const & m1, Molecule* const & m2){ int n1 = m1->getNumAtoms(); int n2 = m2->getNumAtoms(); if( n1 == n2){ return( m1->getMW() < m2->getMW() ); } return( n1 < n2 ); } }; MoleculeSet::MoleculeSet(){ activitySet = false; gram = new vector< vector >; gramNormal = new vector< vector >; gramCalculated = false; setKashimaKernelParam( 0.1, 1000, 1 ); } /** Copy constructor (TODO: check carrefully!) */ MoleculeSet::MoleculeSet(const MoleculeSet& aSet){ activitySet = false; gramCalculated = false; pq = aSet.pq; gram = new vector< vector >; gramNormal = new vector< vector >; /* vector::iterator m; for( m = aSet.begin(); m != aSet.end(); m++ ){ addMoleculeCopy( (*m) ); } */ } MoleculeSet::~MoleculeSet(){ delete gram; delete gramNormal; } //void MoleculeSet::addMolecule( Molecule* aMolecule, bool updateGram ){ void MoleculeSet::addMolecule( Molecule* aMolecule ){ push_back( aMolecule ); aMolecule->setKashimaKernelProb( pq ); //if(gramCalculated == true && updateGram == true){ // completeGramMatrix(); //}//else{ // gramCalculated = false; //} } /** creates a new molecule in the dataset and reads its definition from a Mol file */ Molecule* MoleculeSet::addSingleMOL( string aMolFile, bool genericAtomType ) { Molecule* molecule = new Molecule(); molecule->readMOL( aMolFile, genericAtomType ); this->addMolecule( molecule ); return( molecule ); } /** creates a new molecule in the dataset and reads its definition from a Kcf file 14 June 2004: RENAMED addKcfFile to addSingleKCF */ Molecule* MoleculeSet::addSingleKCF( string aKcfFile ) { //cout << "adding kcf file: " << aKcfFile << endl; KCFMolecule* molecule = new KCFMolecule(); molecule->readKCF( aKcfFile ); this->addMolecule( molecule ); return( molecule ); } //void MoleculeSet::addMutag(string aFileName, string rFileName, int numMolToRead, bool updateGram ){ void MoleculeSet::addMutag(string aFileName, string rFileName, uint numMolToRead ){ #ifdef DEBUG //cout << "reading atoms and bonds from mutag in " << aFileName << endl; #endif ifstream inFile; inFile.open( aFileName.c_str(), ios::in ); if(!inFile.good()){ CError e = CError(FILENOTFOUND, aFileName + " file not found"); e.describe(); throw(e); } int linesize=512; int lineNumber = 0; char *line = new char[linesize]; vector typeWords; vector valueWords; string stringLine; string capSymbol; Atom* atom; Molecule* molecule; int atomC = 0; uint i=0; while( !inFile.eof() && i < numMolToRead + 1 ){ //while( lineNumber < 55 ){ lineNumber++; inFile.getline(line, linesize-1,'\n'); stringLine = line; // if line is empty or starts with // or # then skip this line if(stringLine.size() == 0){ continue; } if(stringLine.substr(0,1) == "#" || stringLine.substr(0,2) == "//"){ continue; } StringUtils::Split(line, "(", typeWords); StringUtils::Split(typeWords[1], ",", valueWords); if(typeWords[0] == "atm"){ try{ // try to find the molecule in the moleculeSet molecule = (*this)[ valueWords[0] ]; // IMPLEMENTER L'OPERATEUR NA } catch(CError e){ // if molecule not found in the dataset then add it //#ifdef DEBUG //cout << "adding molecule " << valueWords[0] << " to the molecule set"<< endl; //#endif molecule = new Molecule(); molecule->setStringDescriptor( "name", valueWords[0], "", "", true, true ); molecule->setKashimaKernelProb( pq, 1 ); atomC = 0; addMolecule( molecule ); i++; } //ToUpper up(std::locale::classic()); //ToLower down(std::locale::classic()); capSymbol = valueWords[2]; //std::transform (capSymbol.begin(), capSymbol.begin()+1, capSymbol.begin(), up); #ifdef DEBUG //cout << "adding atom " << valueWords[1] << " of type " << capSymbol << " to molecule " << valueWords[0] << endl; #endif atomC++; atom = molecule->addAtom( capSymbol ); atom->setStringDescriptor( "name", valueWords[1], "", "", true, true ); atom->setIdInMolecule( atomC ); //molecule->describeShort(); }else if(typeWords[0] == "bond"){ // check if molecule exists, eventually create it #ifdef DEBUG //cout << "adding bond of type " << valueWords[3][0] << " between atoms " << valueWords[1] << " and " << valueWords[2] << " in molecule " << valueWords[0] << endl; #endif try{ // try to find the molecule in the moleculeSet molecule = (*this)[ valueWords[0] ]; } catch(CError e){ cout << "ERROR in MoleculeSet::addMutag reading file " << aFileName << endl; cout << "CANNOT ADD A BOND TO A NON EXISTING MOLECULE " << endl; exit(1); } try{ if(atoi(&valueWords[3][0]) == 7 ){ molecule->linkAtoms( valueWords[1], valueWords[2], 4 ); }else{ molecule->linkAtoms( valueWords[1], valueWords[2], atoi(&valueWords[3][0]) ); } } catch( CError e ) { cout << "Error in line " << lineNumber << " of file " << aFileName << ": "; exit(1); } }else{ delete[] line; stringstream out; out << "MoleculeSet::addMutag: Expecting atm() or bond() at line " << lineNumber << " in file " << aFileName; CError e(BADFILE, out.str() ); e.describe(); throw(e); } #ifdef DEBUG //cout << "found " << typeWords.size() << " words in line " << lineNumber << ": " << typeWords[0]<< " " << typeWords[1] << endl; #endif typeWords.clear(); valueWords.clear(); } inFile.close(); if( numMolecules() > numMolToRead ){ cout << "poping back " << endl; //(*end()-1)->getName() << endl; pop_back(); } subsetStart = 0; subsetSize = numMolecules(); // #ifdef DEBUG //cout << "reading mutag biological activity in " << rFileName << endl; // #endif if( rFileName != "" ){ ifstream inFile2; inFile2.open( rFileName.c_str(), ios::in ); if(!inFile2.good()){ delete[] line; CError e = CError(FILENOTFOUND, rFileName + " file not found"); e.describe(); throw(e); } i=0; lineNumber = 0; while( !inFile2.eof() ){ lineNumber++; inFile2.getline(line, linesize-1,'\n'); stringLine = line; //cout << line << endl; // if line is empty or starts with // or # then skip this line if(stringLine.size() == 0){ continue; } if(stringLine.substr(0,1) == "#" || stringLine.substr(0,2) == "//"){ continue; } StringUtils::Split(line, "(", typeWords); StringUtils::Split(typeWords[1], ")", valueWords); //cout << lineNumber << " " << typeWords[0] << ": " << valueWords[0] << endl; if( typeWords[0] == "active" ){ //cout << "active" << endl; try{ //setIntDescriptor(ACTIVITY, valueWords[0], 1); setActivity( valueWords[0], 1.0 ); } catch(CError e){ //e.describe(); } }else{ //cout << "inactive" << endl; try{ //setIntDescriptor(ACTIVITY, valueWords[0], -1); setActivity( valueWords[0], -1.0 ); } catch(CError e){ //e.describe(); } } typeWords.clear(); valueWords.clear(); i++; } inFile2.close(); } delete[] line; // if(gramCalculated == true && updateGram == true){ // completeGramMatrix(); // }//else{ // gramCalculated = false; //} } bool MoleculeSet::nameExists(string aName){ #ifdef DEBUG //cout << "looking for " << aName << endl; #endif vector::iterator molecule; // for each molecule for( molecule = begin(); molecule != end(); molecule++ ){ if( (*molecule)->getStringDescriptor("name")->getValue() == aName){ return(true); } } return(false); } Molecule* MoleculeSet::operator[](int anInd) throw( CError ){ return( getMolByIndex( anInd ) ); } Molecule* MoleculeSet::getMolByIndex(int anInd) throw( CError ){ Molecule *res; //res = this[anInd]; res = this->at(anInd); return( res ); } Molecule* MoleculeSet::operator[](string aName) throw( CError ){ return( getMolByName( aName ) ); } Molecule* MoleculeSet::getMolByName(string aName) throw( CError ){ vector::iterator molecule; int numRes = 0; Molecule* res; // for each molecule for( molecule = begin(); molecule != end(); molecule++ ){ if( (*molecule)->getStringDescriptor("name")->getValue() == aName){ res = *molecule; numRes++; } } if( numRes == 0 ){ stringstream out; out << "MoleculeSet::[]: Error in [] operator for class MoleculeSet: No molecule with name " << aName << " found in the moleculeSet"; CError e(NOTFOUND, out.str() ); //e.describe(); throw(e); }else if( numRes > 1 ){ stringstream out; out << "MoleculeSet::[]: Error in [] operator for class MoleculeSet: More than one molecule with name " << aName << " were found in the moleculeSet (" << numRes << " molecules found)"; CError e(DUPLICATEENTRIES, out.str() ); //e.describe(); throw(e); } return( res ); } void MoleculeSet::setKashimaKernelParam( double aPq, int aConvergenceCondition, bool skipSkeleton ){ //cout << "########### " << aPq << " " << pq << " " << aConvergenceCondition << " " << getConvergenceCondition() << endl; //if( aPq != pq || aConvergenceCondition != getConvergenceCondition() ){ resetGramMatrix(); resetSelfKernels(); vector::iterator molecule; // for each molecule for( molecule = begin(); molecule != end(); molecule++ ){ //cout << "SETTING K PARAM OF " << (*molecule)->getName() << endl; //aPq = 1/(*molecule)->numAtoms(); (*molecule)->setKashimaKernelProb( aPq, skipSkeleton ); } pq = aPq; convergenceCondition = aConvergenceCondition; //} } void MoleculeSet::gramCompute( double aPq, double(*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), int, int ), double(*pt2AtomKernel)(Atom*, Atom*) = MoleculeUtils::atomKernelExternalMatrix, double(*pt2BondKernel)(Bond*, Bond*) = MoleculeUtils::bondKernelType, int aParameter1, string aReportFileName, int nbThreadsWanted, bool silentMode , bool filterTotters ){ gramCompute( this, aPq, pt2GraphKernel, pt2AtomKernel, pt2BondKernel , aParameter1, 1, aReportFileName, nbThreadsWanted, silentMode, filterTotters ); } void MoleculeSet::gramCompute( double aPq, double(*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), int, int ), double(*pt2AtomKernel)(Atom*, Atom*) = MoleculeUtils::atomKernelExternalMatrix, double(*pt2BondKernel)(Bond*, Bond*) = MoleculeUtils::bondKernelType, int aParameter1, int aParameter2, string aReportFileName, int nbThreadsWanted, bool silentMode , bool filterTotters ){ gramCompute( this, aPq, pt2GraphKernel, pt2AtomKernel, pt2BondKernel , aParameter1, aParameter2, aReportFileName, nbThreadsWanted, silentMode, filterTotters ); } void MoleculeSet::gramCompute( MoleculeSet* anotherSet, double aPq, double(*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), int, int ), double(*pt2AtomKernel)(Atom*, Atom*) = MoleculeUtils::atomKernelExternalMatrix, double(*pt2BondKernel)(Bond*, Bond*) = MoleculeUtils::bondKernelType, int aParameter1, string aReportFileName, int nbThreadsWanted, bool silentMode , bool filterTotters ){ gramCompute( anotherSet, aPq, pt2GraphKernel, pt2AtomKernel, pt2BondKernel , aParameter1, 1, aReportFileName, nbThreadsWanted ,silentMode, filterTotters); } void MoleculeSet::gramCompute( MoleculeSet* anotherSet, double aPq, double(*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), int, int ), double(*pt2AtomKernel)(Atom*, Atom*) = MoleculeUtils::atomKernelExternalMatrix, double(*pt2BondKernel)(Bond*, Bond*) = MoleculeUtils::bondKernelType, int aParameter1, int aParameter2, string aReportFileName, int nbThreadsWanted, bool silentMode, bool filterTotters){ this->comparisonSet = anotherSet; subsetStart = 0; subsetSize = numMolecules(); bool noActivity = false; // set to true uf one of the molecule has no activity if( !silentMode ){ cout << "calculating the Kashima gram matrix for " << numMolecules() << " x " << anotherSet->numMolecules() << " molecules" << endl; } // clear the gram matrix gram->clear(); gramNormal->clear(); // set the probabilities according to the article by Kashima et al. setKashimaKernelParam( aPq, aParameter1 ); if( anotherSet != this ){ anotherSet->setKashimaKernelParam( aPq, aParameter1 ); } if(filterTotters) { // DEBUG (PM) // vector::iterator mtest,mtest2; // int i = 0; // for( mtest = begin(); mtest != end(); mtest++ ) // { // i++; // cout << "molecule no before transfo " << i << " ; nbr atoms = " << (*mtest)->numAtoms() << " ; nbr edges = " <<(*mtest)->numBonds() << endl; // } this->noTottersTransform(); if( anotherSet != this){ anotherSet->noTottersTransform(); } // DEBUG (PM) // i = 0; // for( mtest2 = begin(); mtest2 != end(); mtest2++ ) // { // i++; // cout << "molecule no after transfo " << i << " ; nbr atoms = " << (*mtest2)->numAtoms() << " ; nbr edges = " <<(*mtest2)->numBonds() << endl; // } // exit(1); } // calculate Gram Matrix vector::iterator m1; vector::iterator m2; int i = 0; int j = 0; // counter // initialize the gram matrix and calculate all Selfkernels i = 0; for( m1 = begin(); m1 != end(); m1++ ){ gram->push_back( vector() ); gramNormal->push_back( vector() ); j = 0; for( m2 = anotherSet->begin(); m2 != anotherSet->end(); m2++ ){ (*gram)[i].push_back( -1 ); (*gramNormal)[i].push_back( -1 ); j++; } i++; } // for each pair of molecules calculate the Kashima Kernel if(nbThreadsWanted > 1) { // MULTITHREAD CODE COMES HERE stringstream out; out << "Atom::atomKernel function is not implemented yet"; CError e( NOTIMPLEMENTED, out.str() ); e.describe(); throw(e); }else{ // no threads if( !silentMode ){ cout << "Using a single thread" << endl; } double sk1 = -1000.0; // self kernel mol1 double sk2 = -1000.0; // self kernel mol2 double k = 0; // kernel double nk = 0; // normalised kernel int nbUndistinct = 0; // number of indistinct molecules int nbUndistinctDiffAct = 0; // number of indistinct molecules with different activity status int nbUndistinctUnknownAct = 0; // number of indistinct molecules with unknown activity stringstream confoundedMolecules; // to report the confounded molecules int nbOrthogonal = 0; // number of orthogonal molecules int nbOrthogonalDiffAct = 0; // number of orthogonal molecules with different activity status int nbOrthogonalUnknownAct = 0; // number of orthogonal molecules with unknown activity stringstream orthogonalMolecules; // to report the orthogonal molecules if( !silentMode ){ cout << "calculating " << this->numMolecules() << " x " << anotherSet->numMolecules() << endl; } if(this == anotherSet){ // this is a symetric gram Matrix therefore only one triangular matrix needs to be computed i = 0; for( m1 = this->begin(); m1 != this->end(); m1++ ){ if( !silentMode ){ cout << i << " / " << subsetSize << endl; } #ifdef DEBUG cout << " sk1" << endl; #endif sk1 = (*m1)->getSelfKernel( pt2GraphKernel, pt2AtomKernel, pt2BondKernel, aParameter1, aParameter2 ); j = 0; for( m2 = anotherSet->begin(); m2 != anotherSet->end() - ( anotherSet->numMolecules() - i - 1) ; m2++ ){ #ifdef DEBUG cout << " " << j << endl; #endif //cout << i << "/" << j << ": " << (*m1)->getName() << " " << (*m2)->getName() << endl; if( (*m1) == (*m2) ){ #ifdef DEBUG cout << " sk = k" << endl; #endif k = sk1; //(*m1)->getSelfKernel( pt2GraphKernel, pt2AtomKernel, pt2BondKernel, aParameter1, aParameter2 ); sk2 = k; }else{ #ifdef DEBUG cout << " k" << endl; #endif k = (*m1)->computeKernel(*m2, pt2GraphKernel, pt2AtomKernel, pt2BondKernel, aParameter1, aParameter2 ); #ifdef DEBUG cout << " sk2" << endl; #endif sk2 = (*m2)->getSelfKernel( pt2GraphKernel, pt2AtomKernel, pt2BondKernel, aParameter1, aParameter2 ); } double denom = sqrt( sk1 * sk2 ); nk = k / denom; //sqrt( sk1 * sk2 ); if( !( denom > 0 ) ){ if( (*m1) == (*m2) ){ nk = 1; }else{ nk = 0; } } //cout << sk1 << " " << sk2 << " " << nk << " " << endl; (*gram)[i][j] = k; (*gram)[j][i] = k; (*gramNormal)[i][j] = nk; (*gramNormal)[j][i] = nk; j++; if( nk > MAXSIMILARITYWARNING && m1 != m2 ){ nbUndistinct++; try{ //if( (*m1)->getIntDescriptor(ACTIVITY)->getValue() == (*m2)->getIntDescriptor(ACTIVITY)->getValue() ){ float a1 = 0; float a2 = 0; //cout << "c1h b" << endl; a1 = (*m1)->getActivity( true ); a2 = (*m2)->getActivity( true ); //cout << "c2h c" << endl; if( a1 == a2 ){ if( !silentMode ){ cout << "WARNING: molecule " << (*m1)->getName() << " and " << (*m2)->getName() << " cannot be distinguished. Normalized kernel = " << nk << endl; } }else{ nbUndistinctDiffAct++; if( !silentMode ){ cout << "WARNING: molecule " << (*m1)->getName() << " and " << (*m2)->getName() << " cannot be distinguished and they have DIFFERENT BIOLOGICAL ACTIVITY. Normalized kernel = " << nk << endl; } confoundedMolecules << (*m1)->getName() << " - " << (*m2)->getName() << endl; } } catch( CError e ){ noActivity = true; nbUndistinctUnknownAct++; continue; } } if( nk < MINSIMILARITYWARNING ){ nbOrthogonal++; try{ float a1 = 0; float a2 = 0; //cout << "c2h b" << endl; a1 = (*m1)->getActivity( true ); a2 = (*m2)->getActivity( true ); //cout << "c2h c" << endl; //if( (*m1)->getIntDescriptor(ACTIVITY)->getValue() == (*m2)->getIntDescriptor(ACTIVITY)->getValue() ){ if( a1 == a2 ){ if( !silentMode ){ cout << "WARNING: molecule " << (*m1)->getName() << " and " << (*m2)->getName() << " are orthogonal. Normalized kernel = " << nk << endl; } }else{ nbOrthogonalDiffAct++; if( !silentMode ){ cout << "WARNING: molecule " << (*m1)->getName() << " and " << (*m2)->getName() << " are orthogonal but have DIFFERENT BIOLOGICAL ACTIVITY. Normalized kernel = " << nk << endl; } orthogonalMolecules << (*m1)->getName() << " - " << (*m2)->getName() << endl; } } catch( CError e ){ nbOrthogonalUnknownAct++; noActivity = true; continue; } } } i++; } }else{ // this is not a symetric matrix, compute all i = 0; for( m1 = this->begin(); m1 != this->end(); m1++ ){ if( !silentMode ){ cout << i << " / " << subsetSize << endl; } sk1 = (*m1)->getSelfKernel( pt2GraphKernel, pt2AtomKernel, pt2BondKernel, getConvergenceCondition(), aParameter2 ); j = 0; for( m2 = anotherSet->begin(); m2 != anotherSet->end() ; m2++ ){ if( (*m1) == (*m2) ){ k = (*m1)->getSelfKernel( pt2GraphKernel, pt2AtomKernel, pt2BondKernel, getConvergenceCondition(), aParameter2 ); }else{ k = (*m1)->computeKernel(*m2, pt2GraphKernel, pt2AtomKernel, pt2BondKernel, getConvergenceCondition(), aParameter2 ); } sk2 = (*m2)->getSelfKernel( pt2GraphKernel, pt2AtomKernel, pt2BondKernel, getConvergenceCondition(), aParameter2 ); nk = k / sqrt(sk1 * sk2); (*gram)[i][j] = k; (*gramNormal)[i][j] = nk; j++; if( nk > MAXSIMILARITYWARNING && m1 != m2 ){ nbUndistinct++; try{ float a1 = 0; float a2 = 0; a1 = (*m1)->getActivity( true ); a2 = (*m2)->getActivity( true ); if( a1 == a2 ){ //if( (*m1)->getIntDescriptor(ACTIVITY)->getValue() == (*m2)->getIntDescriptor(ACTIVITY)->getValue() ){ if( !silentMode ){ cout << "WARNING: molecule " << (*m1)->getName() << " and " << (*m2)->getName() << " cannot be distinguished. Normalized kernel = " << nk << endl; } }else{ nbUndistinctDiffAct++; if( !silentMode ){ cout << "WARNING: molecule " << (*m1)->getName() << " and " << (*m2)->getName() << " cannot be distinguished and they have DIFFERENT BIOLOGICAL ACTIVITY. Normalized kernel = " << nk << endl; } confoundedMolecules << (*m1)->getName() << " - " << (*m2)->getName() << endl; } } catch( CError e ){ nbUndistinctUnknownAct++; noActivity = true; continue; } } if( nk < MINSIMILARITYWARNING ){ nbOrthogonal++; try{ //if( (*m1)->getIntDescriptor(ACTIVITY)->getValue() == (*m2)->getIntDescriptor(ACTIVITY)->getValue() ){ if( (*m1)->getActivity( true ) == (*m2)->getActivity( true ) ){ if( !silentMode ){ cout << "WARNING: molecule " << (*m1)->getName() << " and " << (*m2)->getName() << " are orthogonal. Normalized kernel = " << nk << endl; } }else{ nbOrthogonalDiffAct++; if( !silentMode ){ cout << "WARNING: molecule " << (*m1)->getName() << " and " << (*m2)->getName() << " are orthogonal but have DIFFERENT BIOLOGICAL ACTIVITY. Normalized kernel = " << nk << endl; } orthogonalMolecules << (*m1)->getName() << " - " << (*m2)->getName() << endl; } } catch( CError e ){ nbOrthogonalUnknownAct++; noActivity = true; continue; } } } i++; } } if( !silentMode ){ cout << endl<< "Report:" << endl; cout << nbUndistinct << " molecules pairs could not be distinguished using the graph kernel" << endl; cout << nbUndistinctDiffAct << " of them had a different biological activity" << endl; cout << nbUndistinctUnknownAct << " of them had unknown biological activity" << endl<0 && !silentMode ){ cout << "here is a list of the undistinguished molecules with different activities:" << endl; cout << confoundedMolecules.str(); } if( nbOrthogonalDiffAct>0 && !silentMode ){ cout << "here is a list of the orthogonal molecules with different activities:" << endl; cout << orthogonalMolecules.str(); } if( aReportFileName != "" ){ // write a report in aReportFile fstream oFile; oFile.open( aReportFileName.c_str(), ios::out ); if(!oFile.good()){ CError e = CError(COULDNOTOPENFILE, "MoleculeSet::kashimaGramMatrix: could not write file " + aReportFileName); e.describe(); throw(e); } fstream oFileShort; string aReportFileNameShort = aReportFileName + ".short"; oFileShort.open( aReportFileNameShort.c_str(), ios::out ); if(!oFileShort.good()){ CError e2 = CError(COULDNOTOPENFILE, "MoleculeSet::kashimaGramMatrix: could not write file " + aReportFileName + ".short"); e2.describe(); throw(e2); } oFile << "Kashima Gram Kernel Matrix with pq = " << aPq << endl; oFile << "convergence until rlk does not change by more than 1/" << aParameter1 << endl; oFile << "atom kernel matrix loaded from " << elements.getAtomKernelName() << endl; oFile << nbUndistinct << " molecule pairs could not be distinguished using the graph kernel " << endl; oFile << nbUndistinctUnknownAct << " of them had unknown biological activity" << endl; if( hasActivity() ){ oFileShort << nbUndistinct << ";" << nbUndistinctDiffAct << ";" << nbUndistinctUnknownAct << ";" << nbOrthogonal << ";" << nbOrthogonalDiffAct << ";" << nbOrthogonalUnknownAct << endl; oFile << nbUndistinctDiffAct << " of them had a different biological activity" << endl; if( nbUndistinctDiffAct>0 ){ oFile << "here is a list of the undistinguished molecules with different activities:" << endl; oFile << confoundedMolecules.str(); } }else{ oFileShort << nbUndistinct << ";NA;" << nbUndistinctUnknownAct << ";" << nbOrthogonal << ";NA;" << nbOrthogonalUnknownAct << endl; } oFile << nbOrthogonal << " molecule pairs were orthogonal using the graph kernel" << endl; oFile << nbOrthogonalUnknownAct << " of them had unknown biological activity" << endl; if( hasActivity() ){ oFile << nbOrthogonalDiffAct << " of them had a different biological activity" << endl; } oFile.close(); oFileShort.close(); } } gramCalculated = true; } void MoleculeSet::gramCompute3D( double(*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(float,float,float), float ), double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(float, float, float), float edgeKernelparameter, bool silentMode ){ gramCompute3D(this, pt2GraphKernel, pt2AtomKernel, pt2BondKernel, edgeKernelparameter, silentMode); } void MoleculeSet::gramCompute3D( MoleculeSet* anotherSet, double(*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(float,float,float), float ), double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(float, float, float), float edgeKernelParameter, bool silentMode ){ this->comparisonSet = anotherSet; // clear the gram matrix gram->clear(); gramNormal->clear(); // WARNING : you MUST set these parameters // otherwise, problems reading the gram matrix . WHY ?? subsetStart = 0; subsetSize = numMolecules(); //// set the probabilities according to the article by Kashima et al. //setKashimaKernelParam( aPq, aParameter1 ); //if( anotherSet != this ){ // anotherSet->setKashimaKernelParam( aPq, aParameter1 ); //} // calculate Gram Matrix // // ----------------------// vector::iterator m1; vector::iterator m2; int i = 0; int j = 0; // counter // initialize the gram matrix and calculate all Selfkernels i = 0; for( m1 = begin(); m1 != end(); m1++ ){ gram->push_back( vector() ); gramNormal->push_back( vector() ); j = 0; for( m2 = anotherSet->begin(); m2 != anotherSet->end(); m2++ ){ (*gram)[i].push_back( -1 ); (*gramNormal)[i].push_back( -1 ); j++; } i++; } // for each pair of molecules calculate the3D Kernel double sk1 = -1000.0; // self kernel mol1 double sk2 = -1000.0; // self kernel mol2 double k = 0; // kernel double nk = 0; // normalised kernel if( !silentMode ){ cout << "calculating " << this->numMolecules() << " x " << anotherSet->numMolecules() << endl; } if(this == anotherSet){ // this is a symetric gram Matrix therefore only one triangular matrix needs to be computed i = 0; for( m1 = this->begin(); m1 != this->end(); m1++ ){ if( !silentMode ){ cout << " i = " << i+1 << " / " << this->numMolecules() << endl; } #ifdef DEBUG cout << "molecule " << i+1 << " , self kernel" << endl; cout << "--------------------------"<< endl; #endif sk1 = pt2GraphKernel( *m1, *m1, pt2AtomKernel, pt2BondKernel, edgeKernelParameter ); (*m1)->setSelfKernel(sk1); j = 0; for( m2 = this->begin(); m2 != this->end() - ( this->numMolecules() - i - 1) ; m2++ ){ //cout <<"\t j = " << j+1 << endl; if( (*m1) == (*m2) ){ k = sk1; sk2 = k; }else{ #ifdef DEBUG cout << "molecule " << i+1 << " X molecule " << j+1 << endl; cout << "--------------------------"<< endl; #endif k = pt2GraphKernel( *m1, *m2, pt2AtomKernel, pt2BondKernel, edgeKernelParameter ); // NB : core modification to avoid REPETED CALCULATIONS OF SELF KERNELS //#ifdef DEBUG //cout << "molecule " << j+1 << " , self kernel" << endl; //cout << "--------------------------"<< endl; //sk2 = pt2GraphKernel( *m2, *m2, pt2AtomKernel, pt2BondKernel, atomKernelParameter, edgeKernelParameter ); //#endif sk2 = (*m2)->getSelfKernel(); } double denom = sqrt( sk1 * sk2 ); nk = k / denom; if( !( denom > 0 ) ){ if( (*m1) == (*m2) ){ nk = 1; }else{ nk = 0; } } (*gram)[i][j] = k; (*gram)[j][i] = k; (*gramNormal)[i][j] = nk; (*gramNormal)[j][i] = nk; j++; } i++; } }else{ // this is not a symetric matrix, compute all i = 0; for( m1 = this->begin(); m1 != this->end(); m1++ ){ //sk1 = (*m1)->getSelfKernel( pt2GraphKernel, pt2AtomKernel, pt2BondKernel, getConvergenceCondition(), aParameter2 ); sk1 = pt2GraphKernel( *m1, *m1, pt2AtomKernel, pt2BondKernel, edgeKernelParameter ); (*m1)->setSelfKernel(sk1); if( !silentMode ){ cout << " i = " << i+1 << " / " << this->numMolecules() << endl; } j = 0; for( m2 = anotherSet->begin(); m2 != anotherSet->end() ; m2++ ){ //cout <<"\t j = " << j+1 << endl; //k = (*m1)->computeKernel(*m2, pt2GraphKernel, pt2AtomKernel, pt2BondKernel, getConvergenceCondition(), aParameter2 ); k = pt2GraphKernel( *m1, *m2, pt2AtomKernel, pt2BondKernel, edgeKernelParameter ); //sk2 = (*m2)->getSelfKernel( pt2GraphKernel, pt2AtomKernel, pt2BondKernel, getConvergenceCondition(), aParameter2 ); sk2 = pt2GraphKernel( *m2, *m2, pt2AtomKernel, pt2BondKernel, edgeKernelParameter ); (*m2)->setSelfKernel(sk2); nk = k / sqrt(sk1 * sk2); (*gram)[i][j] = k; (*gramNormal)[i][j] = nk; j++; } i++; } } gramCalculated = true; } void MoleculeSet::kernelCompute( Molecule* m2, double(*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), int, int ), double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), vector* resultsRaw, vector* resultsNormal, int convergenceCondition, int aParameter2, bool silentMode ){ vector::iterator m1; double sk1 = -1000.0; // self kernel mol1 double sk2 = -1000.0; // self kernel mol2 double k = 0; // kernel double nk = 0; // normalised kernel int nbUndistinct = 0; // number of indistinct molecules int nbUndistinctDiffAct = 0; // number of indistinct molecules with different activity status int nbUndistinctUnknownAct = 0; // number of indistinct molecules with unknown activity stringstream confoundedMolecules; // to report the confounded molecules int nbOrthogonal = 0; // number of orthogonal molecules int nbOrthogonalDiffAct = 0; // number of orthogonal molecules with different activity status int nbOrthogonalUnknownAct = 0; // number of orthogonal molecules with unknown activity stringstream orthogonalMolecules; // to report the orthogonal molecules for( m1 = this->begin(); m1 != this->end(); m1++ ){ sk1 = (*m1)->getSelfKernel( pt2GraphKernel, pt2AtomKernel, pt2BondKernel, convergenceCondition, aParameter2 ); sk2 = m2->getSelfKernel( pt2GraphKernel, pt2AtomKernel, pt2BondKernel, convergenceCondition, aParameter2 ); k = (*m1)->computeKernel(m2, pt2GraphKernel, pt2AtomKernel, pt2BondKernel, convergenceCondition ); nk = k / sqrt(sk1 * sk2); resultsRaw->push_back( k ); resultsNormal->push_back( nk ); if( nk > MAXSIMILARITYWARNING ){ nbUndistinct++; try{ if( (*m1)->getActivity( true ) == m2->getActivity( true ) ){ //if( (*m1)->getIntDescriptor(ACTIVITY)->getValue() == m2->getIntDescriptor(ACTIVITY)->getValue() ){ if( !silentMode ){ cout << "WARNING: molecule " << (*m1)->getName() << " and " << m2->getName() << " cannot be distinguished. Normalized kernel = " << nk << endl; } }else{ nbUndistinctDiffAct++; if( !silentMode ){ cout << "WARNING: molecule " << (*m1)->getName() << " and " << m2->getName() << " cannot be distinguished and they have DIFFERENT BIOLOGICAL ACTIVITY. Normalized kernel = " << nk << endl; } confoundedMolecules << (*m1)->getName() << " - " << m2->getName() << endl; } } catch( CError e ){ nbUndistinctUnknownAct++; continue; } } if( nk < MINSIMILARITYWARNING ){ nbOrthogonal++; try{ if( (*m1)->getActivity( true ) == m2->getActivity( true ) ){ //if( (*m1)->getIntDescriptor(ACTIVITY)->getValue() == m2->getIntDescriptor(ACTIVITY)->getValue() ){ if( !silentMode ){ cout << "WARNING: molecule " << (*m1)->getName() << " and " << m2->getName() << " are orthogonal. Normalized kernel = " << nk << " but have identical biological activity" << endl; } }else{ nbOrthogonalDiffAct++; //cout << "WARNING: molecule " << (*m1)->getName() << " and " << m2->getName() << " are orthogonal but have DIFFERENT BIOLOGICAL ACTIVITY. Normalized kernel = " << nk << endl; orthogonalMolecules << (*m1)->getName() << " - " << m2->getName() << endl; } } catch( CError e ){ nbOrthogonalUnknownAct++; continue; } } } } void MoleculeSet::writeGramMatrix( string aFileName, bool normal, bool self, bool silentMode ){ // open output file ofile if(gramCalculated == true) { //cout << "completing matrix" << endl; // completeGramMatrix(); }else{ //cout << "calculating matrix" << endl; //cout << "with param " << getPq() << " " << getConvergenceCondition() << endl; // kashimaGramMatrix( getPq(), getConvergenceCondition() ); } string baseName = aFileName; if(normal == false){ aFileName = aFileName + ".gramRaw"; }else{ aFileName = aFileName + ".gramNormal"; } fstream oFile; oFile.open( aFileName.c_str(), ios::out ); if(!oFile.good()){ CError e = CError(COULDNOTOPENFILE, "MoleculeSet::writeGramMatrix: could not write file " + aFileName); e.describe(); throw(e); } oFile.precision(GRAMDECIMALPRECISION); int i = 0; int j = 0; vector::iterator m; vector< vector >::iterator c1; vector::iterator c2; //cout << "writing " << subsetStart << " + " << subsetSize << endl; oFile << "name\t"; //for( m = comparisonSet->begin() + subsetStart; m != comparisonSet->begin() + subsetStart + subsetSize - 1; m++ ){ bool first=true; //cout << " MoleculeSet::writeGramMatrix ; comparisonSet.size() = " << comparisonSet->numMolecules() << endl; //cout << " MoleculeSet::writeGramMatrix ; size() = " << numMolecules() << endl; //cout << " MoleculeSet::writeGramMatrix ; subsetstart = " << subsetStart << endl; for( m = comparisonSet->begin(); m != comparisonSet->end(); m++ ){ //cout << (*m)->getName() << "\t"; if( first ){ oFile << (*m)->getName(); //cout << "\t - getName = " << (*m)->getName() << endl; first = false; }else{ oFile << "\t" << (*m)->getName(); //cout << "\t - getName = " << (*m)->getName() << endl; } } oFile << endl; //cout << endl; // oFile << (*m)->getName() << endl; if(normal == false){ // write the raw gram matrix for( c1 = gram->begin(), m = begin() + subsetStart; c1 != gram->end(); c1++, m++ ){ oFile << (*m)->getName() << "\t"; j = 0; for( c2 = (*c1).begin(); c2 != (*c1).end()-1; c2++ ){ oFile << *c2 << "\t"; //oFile << (*gram)[i][j]; j++; } oFile << *c2 << endl; i++; } }else{ // write the normalised gram matrix for( c1 = gramNormal->begin(), m = begin() + subsetStart; c1 != gramNormal->end(); c1++, m++ ){ oFile << (*m)->getName() << "\t"; j = 0; for( c2 = (*c1).begin(); c2 != (*c1).end()-1; c2++ ){ oFile << *c2 << "\t"; j++; } oFile << *c2 << endl; i++; } } if( self == true ){ writeSelfKernelList( baseName, silentMode ); } } void MoleculeSet::resetGramMatrix(){ gram->clear(); gramCalculated = false; } /** hides all hydrogens in all molecules of the set */ void MoleculeSet::hideHydrogens(){ cout << "HIDING HYDROGENS in the " << numMolecules() << " molecules of the molecule set" << endl; vector::iterator m; int i = 1; for( m = begin(); m != end(); m++ ){ (*m)->hideHydrogens(); i++; } } /** hides all but the largest connected graph in all molecules of the set */ void MoleculeSet::hideSalts( string aReportFileName ){ stringstream out; cout << "HIDING SALTS COUNTERIONS in the " << numMolecules() << " molecules of the molecule set" << endl; out<<"name;graphRemoved;biggestComponent;secondBiggestComponent;difference;numHiddenAtoms;warning" << endl; int numRemoved = 0; vector::iterator m; for( m = begin(); m != end(); m++ ){ //cout << "MoleculeSet: calling hideSalts() " << endl; numRemoved = (*m)->hideSalts( &out ); } if( aReportFileName != "" ){ ofstream outFile; outFile.open( aReportFileName.c_str(), ios::out ); if( !outFile.good() ){ CError e = CError(COULDNOTOPENFILE, aReportFileName + " could not open file"); e.describe(); throw(e); } outFile << out.str(); outFile.close(); } } /** restores the hidden atoms for all molecules */ void MoleculeSet::restoreHiddenAtoms(){ vector::iterator m; for( m = begin(); m != end(); m++ ){ (*m)->restoreHiddenAtoms(); } } /** sets the value of intDescriptor aLabel of molecule aLabel to aValue */ void MoleculeSet::setIntDescriptor( string aLabel, string aMolecule, int aValue ){ (*this)[aMolecule]->setIntDescriptor( aLabel, aValue, "", "", true, true ); } /** sets the avtivity of aMolecule to aValue */ void MoleculeSet::setActivity( string aMolecule, float aValue ){ (*this)[aMolecule]->setActivity( aValue ); } /** writes a file with the biological activity of molecules in a format compatible with GIST */ void MoleculeSet::writeActivityFile(string aFileName, bool addActivityExtension, string activityDescriptor ){ if( addActivityExtension ){ aFileName = aFileName + ".activity"; } //cout << "writing activity file " << aFileName << " with descriptor " << activityDescriptor << endl; fstream oFile; oFile.open( aFileName.c_str(), ios::out ); if(!oFile.good()){ CError e = CError(COULDNOTOPENFILE, "MoleculeSet::writeActivityFile: could not write file " + aFileName); e.describe(); throw(e); } //if(subsetSize < 1){ // subsetSize = numMolecules() - subsetStart; //} //cout << "for molecules " << subsetStart << " to " << subsetSize << endl; oFile << "label" << "\t" << "class" << endl; vector::iterator m; //for( m = begin() + subsetStart; m != begin() + subsetStart + subsetSize; m++ ){ //cout << "NOW" << endl; //cout << toString() << endl; for( m = begin(); m != end(); m++ ){ //cout << (*m)->toStringLong() << endl; //Descriptor *D = (*m)->getIntDescriptor( activityDescriptor, false ); //cout << "\t" << D->getValue() << endl; try{ if( activityDescriptor == "activity"){ oFile << (*m)->getName() << "\t" << (*m)->getActivity() << endl; }else{ oFile << (*m)->getName() << "\t" << (*m)->getIntDescriptor( activityDescriptor, false )->getValue() << endl; } }catch( CError e ){ continue; } } oFile.close(); } /** reads descriptors for the molecules from a ; separated file the first line in the file indicates the Descriptor comment text the second line in the file indicates the Descriptor units the third line in the file indicates the Descriptor types the forth line indicates the Descriptor names data start at fifth line. the first column should contain the same molecule names as those obtained by getName() for the molecules in the dataset example: Full name;Log octanol partition coefficient;boiling point;special comment NA;NA;K;NA string;float;float;string name;logP;Bp;acomment ethane;10;200;junk data WARNING: does not check for duplicate name entries in the descriptor file WARNING: does not add anything to the molecules not contained in the descriptor file */ void MoleculeSet::readDescriptorFile( string aFileName, string separator ){ ifstream inFile; inFile.open( aFileName.c_str(), ios::in ); if(!inFile.good()){ CError e = CError(FILENOTFOUND, aFileName + " file not found"); e.describe(); throw(e); } int linesize = 512; int lineNumber = 0; char *line = new char[linesize]; string stringLine; vector words; vector comments; vector units; vector types; vector names; vector modMol; int numD = 0; while( !inFile.eof() ){ lineNumber++; //cout << lineNumber << endl; // read line inFile.getline(line, linesize-1,'\n'); stringLine = line; //cout << stringLine << endl; if( lineNumber == 1 ){ // comment line numD = StringUtils::Split(line, separator, comments); cout << "FOUND " << numD << " columns in first line" << endl; }else if( lineNumber == 2 ) { // unit line int a = StringUtils::Split(line, separator, units); if( a != numD ){ delete[] line; stringstream out; out << "MoleculeSet::readDescriptorFile: error in line " << lineNumber << " of file " << aFileName << " found " << a << " columns while " << numD << " are required"; CError e(BADFILE, out.str() ); e.describe(); throw(e); } }else if( lineNumber == 3 ) { // type line int a = StringUtils::Split(line, separator, types); if( a != numD ){ delete[] line; stringstream out; out << "MoleculeSet::readDescriptorFile: error in line " << lineNumber << " of file " << aFileName << " found " << a << " columns while " << numD << " are required"; CError e(BADFILE, out.str() ); e.describe(); throw(e); } }else if( lineNumber == 4 ) { // names line int a = StringUtils::Split(line, separator, names); if( a != numD ){ delete[] line; stringstream out; out << "MoleculeSet::readDescriptorFile: error in line " << lineNumber << " of file " << aFileName << " found " << a << " columns while " << numD << " are required"; CError e(BADFILE, out.str() ); e.describe(); throw(e); } }else if( lineNumber > 4) { // data line if(stringLine!=""){ //cout << "spliting" << endl; int a = StringUtils::Split(line, separator, words); if( a != numD ){ delete[] line; stringstream out; out << "MoleculeSet::readDescriptorFile: error in line " << lineNumber << " of file " << aFileName << " found " << a << " columns while " << numD << " are required"; CError e(BADFILE, out.str() ); e.describe(); throw(e); } string molName = (*words.begin()); vector::iterator itw = words.begin()+1; vector::iterator itc = comments.begin()+1; vector::iterator itu = units.begin()+1; vector::iterator itt = types.begin()+1; vector::iterator itn = names.begin()+1; //cout << "looking for molecule " << molName << endl; try{ Molecule* m = findFirstMoleculeWithName( molName ); Descriptor* di = NULL; Descriptor* df = NULL; Descriptor* ds = NULL; //cout << "found" << endl; modMol.push_back( m ); for(; itw!= words.end(); itw++, itc++, itu++, itt++, itn++ ){ //cout << "creating descriptor " << *itn << " for molecule " << molName << endl; // look for the molecule // set descriptor for the molecule if( *itt == "integer" ){ if( *itw == "NA" || *itw == "" ){ //cout << "setting empty INT" << endl; di = m->setIntDescriptor(*itn, 0, *itu, *itw, true, true ); di->setEmpty(); }else{ //cout << "setting INT" << endl; m->setIntDescriptor(*itn, StringUtils::toInt( *itw ), *itu, *itw, true, true ); } }else if( *itt == "float" ){ if( *itw == "NA" || *itw == "" ){ //cout << "setting empty FLOAT" << endl; df = m->setFloatDescriptor(*itn, 0, *itu, *itw, true, true ); df->setEmpty(); }else{ //cout << "setting FLOAT" << endl; m->setFloatDescriptor(*itn, StringUtils::toFloat( *itw ), *itu, *itw, true, true ); } }else if( *itt == "string" ){ if( *itw == "NA" || *itw == "" ){ //cout << "setting empty STRING" << endl; ds = m->setStringDescriptor(*itn, "NA", *itu, *itw, true, true ); ds->setEmpty(); }else{ //cout << "setting STRING" << endl; m->setStringDescriptor(*itn, *itw, *itu, *itw, true, true ); } }else{ delete[] line; //unknown data type emit error stringstream out; out << "MoleculeSet::readDescriptorFile: Invalid type " << *itt << " found in descriptor in line 3 of file " << aFileName << " (expecting string, integer, or float) "; CError e(BADFILE, out.str() ); e.describe(); throw(e); } } //m->describe(); } catch( CError ){ // molecule not found cerr << "WARNING MoleculeSet::readDescriptorFile: molecule " << molName << " present in " << aFileName << ", but is not present in the dataset " << endl; } //cout << "clear" << endl; words.clear(); } } } //cout << "finish" << endl; //cout << modMol.size() << ", " << numMolecules() << endl; if( modMol.size() < numMolecules() ){ cerr << "WARNING MoleculeSet::readDescriptorFile: descriptor file " << aFileName << " only contains descriptors for " << modMol.size() << " while dataset contains " << numMolecules() << " molecules " << endl; } modMol.clear(); comments.clear(); units.clear(); names.clear(); words.clear(); delete[] line; inFile.close(); } /** reads the activity of molecules from an activity file */ void MoleculeSet::readActivityFile( string aFileName ){ ifstream inFile2; inFile2.open( aFileName.c_str(), ios::in ); if(!inFile2.good()){ CError e = CError(FILENOTFOUND, aFileName + " file not found"); e.describe(); throw(e); } int linesize = 512; int lineNumber = 0; char *line = new char[linesize]; vector words; string stringLine; int i = 0; while( !inFile2.eof() ){ lineNumber++; inFile2.getline(line, linesize-1,'\n'); stringLine = line; //cout << line << endl; // if line is empty or starts with // or # then skip this line if(stringLine.size() == 0){ continue; } if(stringLine.substr(0,1) == "#" || stringLine.substr(0,2) == "//"){ continue; } StringUtils::Split(line, ";", words); if( words[1] == "1" ){ //cout << "active" << endl; try{ //setIntDescriptor(ACTIVITY, words[0], 1); setActivity( words[0], 1 ); } catch(CError e){ //e.describe(); } }else{ //cout << "inactive" << endl; try{ //setIntDescriptor(ACTIVITY, words[0], -1); setActivity( words[0], -1 ); } catch(CError e){ //e.describe(); } } words.clear(); i++; } inFile2.close(); delete[] line; activitySet = true; } /** loads all .mol files in a directory and adds the molecules to the set */ void MoleculeSet::readMolDirectory(string dataDir, bool genericAtomType, long beginMolecule, long endMolecule){ // add all files in the directory vector fileVector; vector pathVector; DIR* pDir = opendir ( dataDir.c_str() ); if ( !pDir ){ cerr << "could not read directory " << dataDir << endl; exit (2); } dirent* pEntry; long j = 0; while ( ( pEntry = readdir ( pDir ) ) ) { string toto = pEntry->d_name; //cout << "found " << toto; if( toto != "." && toto != ".." && toto.length()>2){ if( toto.substr(toto.length()-3,3) == "mol" ){ //cout << " added" << endl; if( ( j >= beginMolecule || beginMolecule < 0 ) && ( j <= endMolecule || endMolecule < 0 )){ string sFound = dataDir + string ( "/" ) + string ( pEntry->d_name ); pathVector.push_back ( sFound ); fileVector.push_back ( toto.substr(0,toto.length()-4) ); } j++; }else{ //cout << " not added" << endl; } }else{ //cout << " not added" << endl; } } cout << "adding mol files" << endl; vector::iterator pit; vector::iterator fit; Molecule* aMolecule; for( pit = pathVector.begin(), fit = fileVector.begin(); pit != pathVector.end(); pit++, fit++ ){ cout << (*pit) << endl; aMolecule = this->addSingleMOL( (*pit), genericAtomType ); aMolecule->setName( (*fit) ); } } /** loads all .kcf files in a directory and adds the molecules to the set */ void MoleculeSet::readKcfDirectory(string dataDir, long beginMolecule, long endMolecule){ // add all files in the directory vector fileVector; vector pathVector; DIR* pDir = opendir ( dataDir.c_str() ); if ( !pDir ){ cerr << "could not read directory " << dataDir << endl; exit (2); } dirent* pEntry; int j = 0; while ( ( pEntry = readdir ( pDir ) ) ) { string toto = pEntry->d_name; //cout << "found " << toto; if( toto != "." && toto != ".." && toto.length()>2){ if( toto.substr(toto.length()-3,3) == "kcf" ){ if( ( j >= beginMolecule || beginMolecule < 0 ) && ( j <= endMolecule || endMolecule < 0 )){ //cout << " added" << endl; string sFound = dataDir + string ( "/" ) + string ( pEntry->d_name ); pathVector.push_back ( sFound ); fileVector.push_back ( toto.substr(0,toto.length()-4) ); } j++; }else{ //cout << " not added" << endl; } }else{ //cout << " not added" << endl; } } cout << "adding kcf files" << endl; vector::iterator pit; vector::iterator fit; Molecule* aMolecule; for( pit = pathVector.begin(), fit = fileVector.begin(); pit != pathVector.end(); pit++, fit++ ){ cout << (*pit) << endl; aMolecule = this->addSingleKCF( (*pit) ); aMolecule->setName( (*fit) ); } } /** Writes all self kernel values in a file */ void MoleculeSet::writeSelfKernelList( string aFileName, bool silentMode ){ aFileName = aFileName + ".self"; //cout << silentMode << endl; if( !silentMode ){ cout << "writing self kernel file " << aFileName << endl; } fstream oFile; oFile.open( aFileName.c_str(), ios::out ); if(!oFile.good()){ CError e = CError(COULDNOTOPENFILE, "MoleculeSet::writeSelfKernelList: could not write file " + aFileName); e.describe(); throw(e); } oFile << "name" << "\t" << "selfKernel" << endl; vector::iterator m; for( m = begin(); m != end(); m++ ){ oFile << (*m)->getName() << "\t" << (*m)->getSelfKernel() << endl; } oFile.close(); } /** reset all calculated selfkernels for the molecules in the dataset */ void MoleculeSet::resetSelfKernels(){ vector::iterator m; for( m = begin(); m != end(); m++ ){ (*m)->resetSelfKernel(); } } void MoleculeSet::deleteAll(){ // write all molecules vector::iterator m; for( m = begin(); m != end(); m++ ){ delete (*m); } clear(); } /** write a MDL structure data (SD) file with the whole moleculeSet */ void MoleculeSet::writeSD( string aFileName, bool selectedOnly ){ ofstream outFile; outFile.open( aFileName.c_str(), ios::out ); if( !outFile.good() ){ CError e = CError(COULDNOTOPENFILE, aFileName + " could not open file"); e.describe(); throw(e); } if( selectedOnly == true ){ // write only selected molecules vector::iterator m; for( m = begin(); m != end(); m++ ){ if( (*m)->isSelected() ){ MoleculeUtils::writeMDLHeaderBlock( *(*m), outFile ); MoleculeUtils::writeMDLCtabBlock( *(*m), outFile ); MoleculeUtils::writeMDLNSDBlock( *(*m), outFile ); } } }else{ // write all molecules vector::iterator m; for( m = begin(); m != end(); m++ ){ //cout << "writing header" << endl; MoleculeUtils::writeMDLHeaderBlock( *(*m), outFile ); //cout << "writing ctab" << endl; MoleculeUtils::writeMDLCtabBlock( *(*m), outFile ); //cout << "writing NSD" << endl; MoleculeUtils::writeMDLNSDBlock( *(*m), outFile ); } } outFile.close(); } void MoleculeSet::writeSubsetSD( string aFileName, vector* anOrder ){ ofstream outFile; outFile.open( aFileName.c_str(), ios::out ); if( !outFile.good() ){ CError e = CError(COULDNOTOPENFILE, aFileName + " could not open file"); e.describe(); throw(e); } // write all molecules vector::iterator m; Molecule* aMol = NULL; for( m = anOrder->begin(); m != anOrder->end(); m++ ){ try{ aMol = getMolByName( (*m) ); MoleculeUtils::writeMDLHeaderBlock( *aMol, outFile ); MoleculeUtils::writeMDLCtabBlock( *aMol, outFile ); MoleculeUtils::writeMDLNSDBlock( *aMol, outFile ); }catch( CError e ){ continue; } } outFile.close(); } void MoleculeSet::writeSubsetKCF( string aFileName, vector* anOrder ){ ofstream outFile; outFile.open( aFileName.c_str(), ios::out ); if( !outFile.good() ){ CError e = CError(COULDNOTOPENFILE, aFileName + " could not open file"); e.describe(); throw(e); } // write all molecules vector::iterator m; Molecule* aMol = NULL; for( m = anOrder->begin(); m != anOrder->end(); m++ ){ try{ aMol = getMolByName( (*m) ); MoleculeUtils::writeKCF( *aMol, outFile ); }catch( CError e ){ continue; } } outFile.close(); } /** write a KCF file with the whole moleculeSet */ void MoleculeSet::writeKCF( string aFileName, bool selectedOnly ){ ofstream outFile; outFile.open( aFileName.c_str(), ios::out ); if( !outFile.good() ){ CError e = CError(COULDNOTOPENFILE, aFileName + " could not open file"); e.describe(); throw(e); } if( selectedOnly == true ){ // write only selected molecules vector::iterator m; for( m = begin(); m != end(); m++ ){ if( (*m)->isSelected() ){ MoleculeUtils::writeKCF( *(*m), outFile ); } } }else{ // write all molecules vector::iterator m; for( m = begin(); m != end(); m++ ){ MoleculeUtils::writeKCF( *(*m), outFile ); } } outFile.close(); } /** write a description of the moleculeSet to cout */ void MoleculeSet::describeShort(){ cout << toStringShort() << endl; //DataContainer::describe(); } void MoleculeSet::describe( bool selectedOnly ){ cout << toString() << endl; //DataContainer::describe(); } void MoleculeSet::describeLong(){ cout << toStringLong() << endl; //DataContainer::describe(); } /** returns a short string description of the MoleculeSet (number of molecules in the set and short description of each molecule) */ string MoleculeSet::toStringShort(){ stringstream out; //cout << "MoleculeSet has " << numMolecules() << " molecules"<< endl; out << "MoleculeSet has " << numMolecules() << " molecules"<< endl; vector::iterator m; //int i = 1; for( m = begin(); m != end(); m++ ){ //cout << i << endl; out << (*m)->toStringShort() << endl; //cout << (*m)->toStringShort() << endl; //i++; } return( out.str() ); } string MoleculeSet::toString( bool selectedOnly ){ stringstream out; if( selectedOnly = false ){ out << "MoleculeSet has " << numMolecules() << " molecules"<< endl; vector::iterator m; for( m = begin(); m != end(); m++ ){ out << (*m)->toString() << endl; } }else{ out << "MoleculeSet has " << numMolecules() << " molecules out of which: "<< endl; vector::iterator m; int c = 0; for( m = begin(); m != end(); m++ ){ if( (*m)->isSelected() ){ out << (*m)->toString() << endl; c++; } } out << c << " are selected" << endl; } return( out.str() ); } string MoleculeSet::toStringLong(){ stringstream out; out << "MoleculeSet has " << numMolecules() << " molecules"<< endl; vector::iterator m; for( m = begin(); m != end(); m++ ){ out << (*m)->toStringLong() << endl; } return( out.str() ); } /** unselect all molecules in the dataset */ void MoleculeSet::unSelectAll(){ vector::iterator m; for( m = begin(); m != end(); m++ ){ (*m)->unSelect(); } } /** select all molecules in the dataset */ void MoleculeSet::selectAll(){ vector::iterator m; for( m = begin(); m != end(); m++ ){ (*m)->select(); } } void MoleculeSet::sortByMW(){ sort( begin(), end(), AscendingMW() ); } void MoleculeSet::sortByNumAtoms(){ sort( begin(), end(), AscendingNumAtoms() ); } /** sorts the molecule collection according to the molecule Descriptor descriptorName of type descriptorType */ void MoleculeSet::sortByDescriptor( string aDescriptorName, int aDescriptorType, bool reverse ){ setSortDescriptor( aDescriptorName, aDescriptorType ); //cout << "MoleculeSet: sorting by " << aDescriptorName << endl; if( reverse ){ cout << "MoleculeSet: sorting by " << getSortDescriptorName() << " descending " << endl; sort( begin(), end(), DescendingOrder() ); }else{ cout << "MoleculeSet: sorting by " << getSortDescriptorName() << " ascending" << endl; sort( begin(), end(), AscendingOrder() ); } } /** sorts the molecule collection according to the molecule Descriptor descriptorName. descriptor type will be read from descriptor name if name is of type ******.integer, or *******.float, set to string otherwise*/ void MoleculeSet::sortByDescriptor( string aDescriptorName, bool reverse ){ string ext = StringUtils::getExtension( aDescriptorName ); //cout << "TYPE: " << ext << endl; if( ext == "integer" ){ aDescriptorName = aDescriptorName.substr( 0, aDescriptorName.size() - 8 ); //cout << "sorting by " << aDescriptorName << endl; sortByDescriptor( aDescriptorName, INTEGER, reverse ); }else if( ext == "float" ){ aDescriptorName = aDescriptorName.substr( 0, aDescriptorName.size() - 6 ); //cout << "sorting by " << aDescriptorName << endl; sortByDescriptor( aDescriptorName, FLOAT, reverse ); }else if( ext == "string" ){ aDescriptorName = aDescriptorName.substr( 0, aDescriptorName.size() - 7 ); //cout << "sorting by " << aDescriptorName << endl; sortByDescriptor( aDescriptorName, STRING, reverse ); }else{ //cout << "sorting by " << aDescriptorName << endl; sortByDescriptor( aDescriptorName, STRING, reverse ); } } /** sets the type and name of descriptor to be used when sorting molecules */ void MoleculeSet::setSortDescriptor( string aName, int aType, bool reverse ) throw( CError ){ // if aType == default, then molecules will be sorted by id whatever aName is if( aType == INTEGER || aType == FLOAT || aType == STRING || aType == DEFAULT ){ vector::iterator m; for( m = begin(); m != end(); m++ ){ (*m)->setSortDescriptor( aName, aType ); } }else{ stringstream out; out << "MoleculeSet::setSortDescriptor: aType is none of integer, float, string or default"; CError e( UNKNOWNDATATYPE, out.str() ); e.describe(); throw(e); } } string MoleculeSet::getSortDescriptorName(){ vector::iterator m = begin(); return( (*m)->getSortDescriptorName() ); } /** returns a pointer to the molecule with name aName in the MoleculeSet */ Molecule* MoleculeSet::findFirstMoleculeWithName( string aName ) throw( CError ){ vector::iterator m; for( m = begin(); m != end(); m++ ){ if( (*m)->getName() == aName ){ return( (*m) ); } } CError e( MOLECULENOTFOUND, "No molecule with name " + aName + " in dataset " + toStringShort() ); throw( e ); } /** writes a ; separated file containing all descriptors for all molecules */ void MoleculeSet::writeDescriptors( string aFileName, bool selectedOnly ) throw( CError ) { vector::iterator m; map< string, Descriptor< string >* >::iterator ds; map< string, Descriptor< int >* >::iterator di; map< string, Descriptor< float >* >::iterator df; // get list of all existing descriptors with their type map< string, int > vs; // string descriptors map< string, int >::iterator v; bool found = false; for( m = begin(); m != end(); m++ ){ for( ds = (*m)->beginStringDescriptor(); ds != (*m)->endStringDescriptor(); ds++ ){ found = false; for( v = vs.begin(); v != vs.begin(); v++ ){ if( (*v).first == (*ds).first ){ found = true; break; } } if( found == false ){ vs[ (*ds).first ] = STRING; } } for( di = (*m)->beginIntDescriptor(); di != (*m)->endIntDescriptor(); di++ ){ found = false; for( v = vs.begin(); v != vs.begin(); v++ ){ if( (*v).first == (*di).first ){ found = true; break; } } if( found == false ){ vs[ (*di).first ] = INTEGER; } } for( df = (*m)->beginFloatDescriptor(); df != (*m)->endFloatDescriptor(); df++ ){ found = false; for( v = vs.begin(); v != vs.begin(); v++ ){ if( (*v).first == (*df).first ){ found = true; break; } } if( found == false ){ vs[ (*df).first ] = FLOAT; } } } std::ostream* out; if( aFileName == "cout" ){ out = &std::cout; }else{ out = new std::ofstream( aFileName.c_str() ); if( !out->good() ){ delete out; CError e = CError(COULDNOTOPENFILE, aFileName + " could not open file"); e.describe(); throw(e); } } //cout << "found " << vs.size() << " descriptors" << endl; // write list of descriptors: *out << "name;"; for( v = vs.begin(); v != vs.end(); v++ ){ //cout << "@" << endl; *out << (*v).first << "." << (*v).second << ";"; } *out << endl; // write descriptors for( m = begin(); m != end(); m++ ){ if( ( selectedOnly && (*m)->isSelected() ) || !selectedOnly ){ *out << (*m)->getName() << ";"; for( v = vs.begin(); v != vs.end(); v++ ){ if( (*v).second == INTEGER ){ try{ if( (*m)->getIntDescriptor( (*v).first )->isEmpty() ){ *out << "NA;"; }else{ *out << (*m)->getIntDescriptor( (*v).first )->getValue() << ";"; } }catch( CError e ){ if( e.getType() == MISSINGDESCRIPTOR ){ *out << "NA;"; }else{ throw( e ); } } }else if( (*v).second == FLOAT ){ try{ if( (*m)->getFloatDescriptor( (*v).first )->isEmpty() ){ *out << "NA;"; }else{ *out << (*m)->getFloatDescriptor( (*v).first )->getValue() << ";"; } }catch( CError e ){ if( e.getType() == MISSINGDESCRIPTOR ){ *out << "NA;"; }else{ throw( e ); } } }else{ try{ if( (*m)->getStringDescriptor( (*v).first )->isEmpty() ){ *out << "NA;"; }else{ *out << (*m)->getStringDescriptor( (*v).first )->getValue() << ";"; } }catch( CError e ){ if( e.getType() == MISSINGDESCRIPTOR ){ *out << "NA;"; }else{ throw( e ); } } } } *out << endl; } } if (out != &std::cout){ //(*out).close(); delete out; } } /** selects the molecules with the names provided as arguments in a vector< string >. Unselect all others. Returns the number of selected molecules. */ int MoleculeSet::select( vector< string >* aSubset ){ int nbselected = 0; unSelectAll(); vector< string >::iterator it; vector< Molecule* >::iterator m; for( m = begin(); m != end(); m++ ){ for( it = aSubset->begin(); it != aSubset->end(); it++ ){ //cout << (*m)->getName() << "?=" << (*it) << endl; if( (*m)->getName() == (*it) ){ //cout << " MATCH" << endl; (*m)->select(); nbselected++; } } } return( nbselected ); } /** unselects the molecules with the names provided as arguments in a vector< string >. Returns the number of selected molecules. */ int MoleculeSet::unSelect( vector< string >* aSubset ){ int nbselected = 0; vector< string >::iterator it; vector< Molecule* >::iterator m; for( m = begin(); m != end(); m++ ){ for( it = aSubset->begin(); it != aSubset->end(); it++ ){ //cout << (*m)->getName() << "?=" << (*it) << endl; if( (*m)->getName() == (*it) ){ //cout << " MATCH" << endl; (*m)->unSelect(); nbselected++; } } } return( nbselected ); } /** write a mol file for each molecule in the set to aDirName. if selectedOnly == true then only selected molecules are written */ long MoleculeSet::writeMolToDir( string aDirName, bool selectedOnly ){ vector::iterator m; int i = 0; string fileName = ""; if( selectedOnly == false ){ for( m = begin(); m != end(); m++ ){ //cout << "next" << endl; //cout << (*m)->getName() << endl; if( StringUtils::right( (*m)->getName(), 4 ) == ".mol" ){ fileName = aDirName + "/" + StringUtils::slashToUnderscore( (*m)->getName() ); }else{ fileName = aDirName + "/" + StringUtils::slashToUnderscore( (*m)->getName() ) + ".mol"; } //cout << "writing " << fileName << endl; (*m)->writeMOL( fileName ); //cout << "done" << endl; i++; } }else{ for( m = begin(); m != end(); m++ ){ if( (*m)->isSelected() ){ if( StringUtils::right( (*m)->getName(), 4 ) == ".mol" ){ fileName = aDirName + "/" + StringUtils::slashToUnderscore( (*m)->getName() ); }else{ fileName = aDirName + "/" + StringUtils::slashToUnderscore( (*m)->getName() ) + ".mol"; } //cout << "writing " << fileName << endl; (*m)->writeMOL( fileName ); //cout << "done" << endl; i++; } } } return i; } /** reads an sd file and returns the number of created molecules beginMolecule and endMolecule specify the index of the first and last molecule to include in the training set (starting count from 0). values of -1 (default) means no limit. */ int MoleculeSet::addSD( string aFileName, bool genericAtomTypeFlag, long beginMolecule, long endMolecule ){ ifstream inFile; inFile.open( aFileName.c_str(), ios::in ); if( !inFile.good() ){ CError e = CError(FILENOTFOUND, aFileName + " file not found"); e.describe(); throw(e); } Molecule* m; int j=0; while ( !inFile.eof() ){ if( j >= beginMolecule && ( j <= endMolecule || endMolecule == -1 ) ){ //#ifdef DEBUG cout << "reading molecule " << j+1 << endl; //#endif m = new Molecule; #ifdef DEBUG cout << "reading Header" << endl; #endif try{ MoleculeUtils::readMDLHeaderBlock( *m, inFile ); //cout << m->getStringDescriptor("comment")->isEmpty() << endl; }catch( CError e ) { delete m; if( e.getType() == EOFERROR ){ //e.describe(); inFile.close(); return(j); }else{ cout << "MoleculeSet::addSD: CAUGHT UNKNOWN ERROR:" << endl; e.describe(); throw(e); } } #ifdef DEBUG cout << "reading CTAB" << endl; #endif bool hasStructure = true; try{ MoleculeUtils::readMDLCtabBlock( *m, inFile, genericAtomTypeFlag ); }catch( CError e ) { if( e.getType() == EOFERROR ){ //e.describe(); delete m; inFile.close(); return(j); }else if( e.getType() == NOSTRUCTURE ){ hasStructure = false; }else{ delete m; cout << "MoleculeSet::addSD: CAUGHT UNKNOWN ERROR:" << endl; e.describe(); throw(e); } } #ifdef DEBUG cout << "reading SD block" << endl; #endif try{ MoleculeUtils::readMDLNSDBlock( *m, inFile ); }catch( CError e ) { if( e.getType() == EOFERROR ){ delete m; //e.describe(); inFile.close(); return(j); }else{ delete m; cout << "MoleculeSet::addSD: CAUGHT UNKNOWN ERROR:" << endl; e.describe(); throw(e); } } //m->describe(); #ifdef DEBUG cout << m->getName() << endl; #endif if( hasStructure == true ){ addMolecule( m ); //cout << " MW: " << m->getMW() << endl; }else{ cerr << "WARNING: molecule " << m->getName() << " was not loaded from " << aFileName << " because it has no structure information " << endl; m->describe(); delete m; } }else{ // skip entry try{ #ifdef DEBUG cout << "skipping molecule " << j+1 << endl; #endif MoleculeUtils::skipMDLEntry( *m, inFile ); }catch( CError e ) { if( e.getType() == EOFERROR ){ inFile.close(); return(j); }else{ cout << "MoleculeSet::addSD: CAUGHT UNKNOWN ERROR:" << endl; e.describe(); throw(e); } } } if( j > endMolecule && endMolecule > 0 ){ inFile.close(); return(j); } j++; } inFile.close(); return( j ); } /** reads an KCF file and returns the number of created molecules */ int MoleculeSet::addKCF( string aFileName, long beginMolecule, long endMolecule ){ ifstream inFile; inFile.open( aFileName.c_str(), ios::in ); if( !inFile.good() ){ CError e = CError(FILENOTFOUND, aFileName + " file not found"); e.describe(); throw(e); } KCFMolecule* m; bool valid = false; int j=0; while ( !inFile.eof() ){ if( j >= beginMolecule && ( j <= endMolecule || endMolecule == -1 ) ){ m = new KCFMolecule; //cout << " READING MOLECULE " << j << endl; try{ valid = MoleculeUtils::readKCFMolecule( *m, inFile ); }catch( CError e ) { if( e.getType() == EOFERROR ){ delete m; //e.describe(); return(j); }else{ delete m; cout << "MoleculeSet::addKCF: CAUGHT UNKNOWN ERROR:" << endl; e.describe(); throw(e); } } if( valid) { //m->describe(); addMolecule( m ); } } if( j > endMolecule && endMolecule > 0 ){ inFile.close(); return(j); } j++; } inFile.close(); return( j ); } /** add all molecules in aSet to the current set */ int MoleculeSet::add( MoleculeSet* aSet ){ vector::iterator m; int i = 0; for( m = aSet->begin(); m != aSet->end(); m++ ){ addMoleculeCopy( (*m) ); i++; } return( i ); } /** add a copy of the molecule in argument. Used by add to merge to datasets */ Molecule* MoleculeSet::addMoleculeCopy( Molecule* aMolecule ){ Molecule* aNewMolecule = new Molecule( *aMolecule ); //aNewMolecule = aMolecule; push_back( aNewMolecule ); aNewMolecule->setKashimaKernelProb( pq ); return( aNewMolecule ); } /** set integer aName to aValue for all compounds in the dataset */ void MoleculeSet::setIntDescriptor( string aName, int aValue ){ vector::iterator m; for( m = begin(); m != end(); m++ ){ (*m)->setIntDescriptor( aName, aValue, "", "", true, true ); } } /** returns the mean distance of all molecules to the moleculeSet barycenter */ double MoleculeSet::diversityBaryMean(){ //cout << "MoleculeSet::diversityBaryMean: " << endl; if(!gramCalculated){ // gram matrix was not calculated, emit error CError e = CError(MISSINGDATA, "MoleculeSet::diversityBaryMean: no gram matrix"); e.describe(); throw(e); } double sumSelf = 0; // we use normal kernel matrix so average of self kernels is 1 double sumNonSelf = 0; vector< vector >::iterator l; vector::iterator c; long lc = 0; long cc = 0; long numMol = gramNormal->size(); //numMol = ((numMol * numMol) - numMol)/2; double kernelNormalValue = 0; for( l = gramNormal->begin(); l != gramNormal->end(); l++ ){ //cout << lc << ": "; cc = lc; for( c = (*l).begin()+lc; c != (*l).end() ; c++ ){ kernelNormalValue = (*c); //cout << lc << ";" << cc << ";" << kernelNormalValue << endl; if( cc == lc ){ sumSelf += kernelNormalValue; } else{ sumNonSelf += kernelNormalValue; } cc++; } lc++; } sumNonSelf = ( sumNonSelf * 2 ) + sumSelf; sumSelf = sumSelf / numMol; sumNonSelf = sumNonSelf / (numMol*numMol); //sumNonSelf = sumNonSelf/ ((( numMol*numMol) - numMol) / 2); //sumNonSelf = sumNonSelf / ( ( ( ( numMol * numMol) - numMol) / 2 ) + numMol ); //cout << numMol << " " << sumNonSelf << " " << sumSelf << " " << endl; return( sumSelf - sumNonSelf ); } /** this function reads a gram matric matching the dataset. Emits an error if the dimension of the read gram matrix read does not match the number of compounds in the dataset. */ void MoleculeSet::readGramNormal( string aFileName ){ readGram( aFileName, gramNormal ); } void MoleculeSet::readGramRaw( string aFileName ){ readGram( aFileName, gram ); } void MoleculeSet::readGram( string aFileName, vector< vector >* aGram ){ // clear potentially existing matrix bool readingGramNormal = true; if( aGram == gram){ readingGramNormal = false; } delete gram; gram = new vector< vector >; delete gramNormal; gramNormal = new vector< vector >; if( readingGramNormal == true ){ aGram = gramNormal; }else{ aGram = gram; } // clear MoleculeSet clear(); ifstream inFile; inFile.open( aFileName.c_str(), ios::in ); if(!inFile.good()){ CError e = CError(FILENOTFOUND, aFileName + " file not found"); e.describe(); throw(e); } int linesize = 30000; char *line = new char[linesize]; vector words; string stringLine; int lineNumber = 1; inFile.getline(line, linesize-1,'\n'); lineNumber++; //cout << "MoleculeSet::readGram 2 " << endl; //cout << "MoleculeSet::readGram reading names" << endl; // add molecules (no structure) with their names StringUtils::Split(line, " ", words); vector< string >::iterator wi2; for( wi2 = words.begin(); wi2 != words.end(); wi2++ ){ Molecule* aMol = new Molecule(); aMol->setName( (*wi2) ); push_back( aMol ); } //cout << "MoleculeSet::readGram 3 " << endl; //cout << "MoleculeSet::readGram reading values" << endl; // read matrix values while( !inFile.eof() ){ //cout << lineNumber << endl; inFile.getline(line, linesize-1,'\n'); stringLine = line; if(stringLine.size() == 0){ continue; } if(stringLine.substr(0,1) == "#" || stringLine.substr(0,2) == "//"){ continue; } words.clear(); //cout << "splitting line " << lineNumber << ": " << line << " into words" << endl; StringUtils::Split(line, " ", words); //cout << "splitted in " << words.size() << endl; if( words.size() < 2 ){ // bad line emit error stringstream out; out << aFileName << " error at line " << lineNumber; CError e = CError( BADFILE, out.str() ); e.describe(); throw(e); }else{ vector< string >::iterator wi; aGram->push_back( vector() ); vector< vector >::iterator li = gramNormal->begin(); vector::iterator ci = (*li).begin(); for( wi = words.begin()+1, ci = (*li).begin() ; wi != words.end(); wi++, ci++){ //cout << atof( (*wi).c_str() ) << " " << lineNumber-2 << endl; (*aGram)[lineNumber-2].push_back( atof( (*wi).c_str() ) ); } li++; } words.clear(); lineNumber++; } delete[] line; inFile.close(); gramCalculated = true; //cout << "gramNormal has now size " << gramNormal->size() << ";" << aGram->size() << endl; } /** sets the morganLabels of each molecule to the anOrder iteration of the Morgan index calculation process */ void MoleculeSet::setMorganLabels( int anOrder ){ vector::iterator m; for( m = begin(); m != end(); m++ ){ (*m)->setMorganLabels( anOrder ); } } /** sets the uniqueMorganIndex of each atom to the Morgan index having the maximum of different connectivity values for the molecule */ void MoleculeSet::setUniqueMorganIndices(){ vector::iterator m; for( m = begin(); m != end(); m++ ){ (*m)->setUniqueMorganIndices(); } } /*! \fn MoleculeSet::selectByMW( float minmw, float maxmw, bool addDescriptor ) selects all molecules in the dataset with mw >= minmw and < maxmw if maxmw = -1 then there is no maximum limit */ int MoleculeSet::selectByMW( float minmw, float maxmw, bool addD){ int i = 0; vector::iterator m; for( m = begin(); m != end(); m++ ){ float mw = (*m)->getMW(); if( mw >= minmw ){ if( mw <= maxmw || maxmw == -1 ){ (*m)->select(); i++; if( addD == true ){ (*m)->setFloatDescriptor( "mw", mw, "", "Molecular weight", true, true ); } } } } return( i ); } int MoleculeSet::selectByNumAtoms( float minAtoms, float maxAtoms, bool addD){ int i = 0; vector::iterator m; for( m = begin(); m != end(); m++ ){ long numAtoms = (*m)->getNumAtoms(); if( numAtoms >= minAtoms ){ if( numAtoms <= maxAtoms || maxAtoms == -1 ){ (*m)->select(); i++; if( addD == true ){ (*m)->setIntDescriptor( "numAtoms", numAtoms, "", "Number of atoms", true, true ); } } } } return( i ); } void MoleculeSet::addFragmentsToSet( Molecule* aMol, int minAtoms ){ string molName = aMol->getName(); string fragName = ""; aMol->unsetBondFlags(); vector::iterator ai; vector* atomList = new vector; // a vector containing all atoms // to be successively examined for( ai = aMol->beginAtom(); ai != aMol->endAtom(); ai++ ){ //cout << " adding atom " << (*ai)->toStringShort() << endl; atomList->push_back( (*ai) ); } for( ai = atomList->begin(); ai!= atomList->end(); ai++){ //cout << "####################################" << endl; //(*ai)->describeShort(); map::iterator bi; vector bondList; // a vector containing all bonds to be successively removed for( bi = (*ai)->beginBond(); bi != (*ai)->endBond(); bi++ ){ //cout << " # examining bond " << (*bi).second->toStringShort(); if( (*bi).second->getLabel() != AROMATICBOND && !(*bi).second->hasFlag() ){ //if( (*bi).second->getLabel() != AROMATICBOND ){ //cout << " stored " << endl; bondList.push_back( (*bi).second ); try{ Bond* oppositeBond = (*bi).first->getBondWithTarget( (*ai) ); oppositeBond->setFlag(); }catch( CError ){ } }//else{ //cout << " ignored " << endl; //} } vector::iterator bli; for( bli = bondList.begin(); bli != bondList.end(); bli++ ){ //cout << "##" << " treating bond " << (*bli)->toStringShort() << endl; aMol->setName( molName + "." + (*bli)->toStringShort() ); //describe(); //(*bli)->describeShort(); (*bli)->getSource()->hideBond( (*bli) ); (*bli)->getTarget()->hideBond( (*bli)->getSource() ); aMol->markFragments(); pushFragments( aMol, minAtoms ); aMol->unmarkFragments(); (*bli)->getSource()->restoreHiddenBonds(); (*bli)->getTarget()->restoreHiddenBonds(); //describe(); } } aMol->setName( molName ); aMol->unsetBondFlags(); atomList->clear(); delete atomList; //removeDuplicates(); } void MoleculeSet::pushFragments( Molecule* aMol, int minAtoms ){ map< int, int >::iterator j; int i = 1; for( j = aMol->beginComponentSizes(); j != aMol->endComponentSizes(); j++ ){ //describe(); aMol->hideAllFragmentsBut( (*j).first ); if( aMol->getNumAtoms() >= minAtoms ){ string molName = aMol->getName(); stringstream newName; newName << molName << "." << i; aMol->setName( newName.str() ); addMoleculeCopy( aMol ); aMol->setName( molName ); i++; } aMol->restoreHiddenAtoms( false ); } } void MoleculeSet::removeDuplicates() { // for each molecule filter based on // 1) number of atoms // 2) molecular weight // 3) graph kernel vector::iterator m; vector::iterator n; int na1 = 0; int na2 = 0; float mw1 = 0; float mw2 = 0; for( m = begin(); m != end(); m++ ){ na1 = (*m)->getNumAtoms(); for( n = begin(); n != end(); n++ ){ // do not compare a molecule with itself if( (*m) != (*n) ){ // same number of atoms? na2 =(*n)->getNumAtoms(); //cout << "Num atoms: " << na1-na2 << endl; if( na1 == na2 ){ // was n not already marked as a duplicate? if( !(*n)->hasIntDescriptor( "markForDelete" ) && !(*n)->hasIntDescriptor( "markForKeep" ) ){ // if it has not the same sum of bond orders long bo1 = (*m)->bondSum(); long bo2 = (*n)->bondSum(); //cout << "Bond Orders: " << bo1 << " " << bo2 << endl; if( bo1 == bo2 ){ //cout << "MW" << endl; mw1 = (*m)->getMW(); mw2 = (*n)->getMW(); //cout << "MW: " << mw1-mw2 << endl; // same molecular weight? if( fabs( mw1 - mw2 ) < DUPLICATEMWIDENTITY ){ // same graphKernel? //cout << "calculating kernel bewteen " << (*n)->getName() << " and " << (*m)->getName() << endl; (*m)->setKashimaKernelProb( 0.01 ); (*n)->setKashimaKernelProb( 0.01 ); double sk1 = (*m)->getSelfKernel( MoleculeUtils::moleculeKernel, MoleculeUtils::atomKernelSymbol, MoleculeUtils::bondKernelType, 1000 ); double sk2 = (*n)->getSelfKernel( MoleculeUtils::moleculeKernel, MoleculeUtils::atomKernelSymbol, MoleculeUtils::bondKernelType, 1000 ); double k = (*m)->computeKernel( (*n), MoleculeUtils::moleculeKernel, MoleculeUtils::atomKernelSymbol, MoleculeUtils::bondKernelType, 1000 ); double nk = k / sqrt(sk1 * sk2); //cout << "kernel: " << nk << endl; if( nk >= 1){ //(*m) and (*n) are duplicates, mark (*n) for delete (*n)->setIntDescriptor( "markForDelete", 1, "", "", true, true ); (*m)->setIntDescriptor( "markForKeep", 1, "", "", true, true ); //cout << "DUPLICATES" << endl; } } } } } } } } // delete marked compounds for( m = begin(); m != end(); m++ ){ if( (*m)->hasIntDescriptor( "markForDelete" ) ){ delete (*m); erase( m ); m--; }else{ (*m)->deleteDescriptor( "markForKeep" ); } } } void MoleculeSet::deleteHiddenAtoms() { vector::iterator m; for( m = begin(); m != end(); m++ ){ (*m)->deleteHiddenAtoms(); } } void MoleculeSet::readGistClassifyFile( string aFileName ) { ifstream inFile; inFile.open( aFileName.c_str(), ios::in ); if(!inFile.good()){ CError e = CError(FILENOTFOUND, aFileName + " file not found"); e.describe(); throw(e); } int linesize = 512; int lineNumber = 0; char *line = new char[linesize]; string stringLine; vector words; vector modMol; bool readingComments = true; while( !inFile.eof() ){ lineNumber++; // read line inFile.getline(line, linesize-1,'\n'); stringLine = line; // data line if( stringLine != "" && stringLine.substr( 0, 1 ) != "#" ){ if( readingComments == true ){ if( stringLine.substr(0, 4) == "name" ){ readingComments = false; } }else{ StringUtils::Split(line, "\t", words); string molName = (*words.begin()); //cout << "looking for molecule " << molName << endl; try{ Molecule* m = findFirstMoleculeWithName( molName ); modMol.push_back( m ); m->setStringDescriptor( "gistname", words[0], "", "name in gist classify file", true, true ); m->setIntDescriptor( "gistclass", StringUtils::toInt( words[1] ),"", "activity class", true, true ); m->setFloatDescriptor( "gistdiscr", StringUtils::toFloat( words[2] ),"", "activity class", true, true ); //m->describe(); } catch( CError e ){ // molecule not found cerr << "WARNING MoleculeSet::readDescriptorFile: molecule " << molName << " present in " << aFileName << ", but is not present in the dataset " << endl; } //cout << "clear" << endl; words.clear(); } } } //cout << "finish" << endl; //cout << modMol.size() << ", " << numMolecules() << endl; if( modMol.size() < numMolecules() ){ cerr << "WARNING MoleculeSet::readDescriptorFile: descriptor file " << aFileName << " only contains descriptors for " << modMol.size() << " while dataset contains " << numMolecules() << " molecules " << endl; } delete[] line; modMol.clear(); words.clear(); inFile.close(); } void MoleculeSet::readGistActivityFile( string aFileName, string aDescriptor ){ ifstream inFile; inFile.open( aFileName.c_str(), ios::in ); if(!inFile.good()){ CError e = CError(FILENOTFOUND, aFileName + " file not found"); e.describe(); throw(e); } int linesize = 512; int lineNumber = 0; char *line = new char[linesize]; string stringLine; vector words; vector modMol; bool readingFirstLine = true; while( !inFile.eof() ){ lineNumber++; // read line inFile.getline(line, linesize-1,'\n'); stringLine = line; // data line if( stringLine != "" && stringLine.substr( 0, 1 ) != "#" ){ if( readingFirstLine == true ){ //if( stringLine.substr(0, 4) == "name" ){ readingFirstLine = false; //} }else{ StringUtils::Split(line, "\t", words); if( words.size() < 2 ){ // bad file emit error delete[] line; stringstream s; s << aFileName << " error reading line " << lineNumber << endl; CError e = CError(BADFILE, s.str() ); e.describe(); throw(e); } string molName = ( *words.begin() ); string activity = ( words[1] ); //cout << "looking for molecule " << molName << endl; try{ Molecule* m = findFirstMoleculeWithName( molName ); modMol.push_back( m ); m->setIntDescriptor( aDescriptor, StringUtils::toInt( activity ),"", "activity class", true, true ); //m->describe(); }catch( CError e ){ // molecule not found cerr << "WARNING MoleculeSet::readDescriptorFile: molecule " << molName << " present in " << aFileName << ", but is not present in the dataset " << endl; } //cout << "clear" << endl; words.clear(); } } } //cout << "finish" << endl; //cout << modMol.size() << ", " << numMolecules() << endl; if( modMol.size() < numMolecules() ){ cerr << "WARNING MoleculeSet::readDescriptorFile: descriptor file " << aFileName << " only contains descriptors for " << modMol.size() << " while dataset contains " << numMolecules() << " molecules " << endl; } delete[] line; modMol.clear(); words.clear(); inFile.close(); } long MoleculeSet::selectByFloatDescriptor( string aName, float aValue ) { vector< Molecule* >::iterator m; long c = 0; for( m = begin(); m != end(); m++ ){ Descriptor< float >* d = NULL; try{ d = (*m)->getFloatDescriptor( aName ); }catch( CError e ){ d = NULL; } if( d != NULL ){ if( d->getValue( false ) == aValue ){ (*m)->select(); c++; }else{ (*m)->unSelect(); } }else{ (*m)->unSelect(); } } return( c ); } long MoleculeSet::selectByIntDescriptor( string aName, int aValue ) { if( aName.substr( aName.length() - min( (int) aName.length(),8), 8 ) == ".integer"){ aName = aName.substr( 0, aName.size() - 8 ); } vector< Molecule* >::iterator m; long c = 0; for( m = begin(); m != end(); m++ ){ Descriptor< int >* d = NULL; try{ d = (*m)->getIntDescriptor( aName ); }catch( CError e ){ d = NULL; } if( d != NULL ){ if( d->getValue( false ) == aValue ){ (*m)->select(); c++; }else{ (*m)->unSelect(); } }else{ (*m)->unSelect(); } } return( c ); } long MoleculeSet::selectByActivity( float aValue ) { vector< Molecule* >::iterator m; long c = 0; for( m = begin(); m != end(); m++ ){ float activityValue = -1; try{ if( (*m)->getActivity() == aValue ){ (*m)->select(); c++; }else{ (*m)->unSelect(); } }catch( CError e ){ (*m)->unSelect(); } } return( c ); } long MoleculeSet::selectHasActivity() { vector< Molecule* >::iterator m; long c = 0; for( m = begin(); m != end(); m++ ){ if( (*m)->hasActivity() ){ (*m)->select(); }//else{ //(*m)->unSelect(); //} } return( c ); } long MoleculeSet::getPossibleValuesInIntDescriptor( string aDescriptorName, vector< int >* p ){ long res = 0; if( aDescriptorName.substr( aDescriptorName.length() - min( (int) aDescriptorName.length(),8), 8 ) == ".integer"){ aDescriptorName = aDescriptorName.substr( 0, aDescriptorName.size() - 8 ); } vector< Molecule* >::iterator m; for( m = begin(); m != end(); m++ ){ int newValue = -1; try{ if( aDescriptorName == "activity" ){ newValue = (int) (*m)->getActivity(); }else{ newValue = (*m)->getIntDescriptor( aDescriptorName )->getValue(); } vector::iterator vi; long nfound = 0; for( vi = p->begin(); vi != p->end(); vi++){ if( (*vi) == newValue ){ nfound += 1; break; } } if( nfound == 0 ){ p->push_back( newValue ); res += 1; } }catch( CError e ){ continue; } } return( res ); } void MoleculeSet::binClassifyFromDescriptor( string descriptorName, float value, bool smallerOrEqual ){ vector< Molecule* >::iterator m; for( m = begin(); m != end(); m++ ){ (*m)->binClassifyFromDescriptor( descriptorName, value, smallerOrEqual ); } } // initialize every element gram matrix to the given value void MoleculeSet::initializeGram( double value ){ gram->clear(); gramNormal->clear(); vector::iterator m1; vector::iterator m2; int i = 0; for( m1 = begin(); m1 != end(); m1++ ){ gram->push_back( vector() ); gramNormal->push_back( vector() ); for( m2 = comparisonSet->begin(); m2 != comparisonSet->end(); m2++ ){ (*gram)[i].push_back( value ); (*gramNormal)[i].push_back( value ); } i++; } } // initialize every self-kernelto the given value void MoleculeSet::initializeSelfKernel( double value ){ vector::iterator m; for( m = begin(); m != end(); m++ ){ (*m)->resetSelfKernel(); (*m)->setSelfKernel(0.0); // (*m)->addToSelfKernel(0.0); // --> bug fixed by PM (november 2007) } } /* // normalize the gram matrix void MoleculeSet::normalizeGram(){ vector::iterator mol1; vector::iterator mol2; int i = 0; int j = 0; for(mol1 = begin(); mol1 != end() ; mol1++){ j = 0; for(mol2 = comparisonSet->begin() ; mol2 != comparisonSet->end() ; mol2++){ //(*gramNormal)[i][j] = (*gram)[i][j]/sqrt( (*mol1)->getSelfKernel() * (*mol2)->getSelfKernel() ); (*gramNormal)[i][j] = (*gram)[i][j]/sqrt( (*gram)[i][i] * (*gram)[j][j] ); j++; } i++; } } */ // normalize the gram matrix void MoleculeSet::normalizeGram(){ vector::iterator mol1; vector::iterator mol2; int i = 0; int j = 0; for(mol1 = begin(); mol1 != end() ; mol1++){ j = 0; for(mol2 = comparisonSet->begin() ; mol2 != comparisonSet->end() ; mol2++){ if( ( (*mol1)->getSelfKernel() == 0 ) || ( (*mol2)->getSelfKernel() == 0 ) ){ (*gramNormal)[i][j] = 0; } else{ (*gramNormal)[i][j] = (*gram)[i][j]/sqrt( (*mol1)->getSelfKernel() * (*mol2)->getSelfKernel() ); } j++; } i++; } } // normalize the gram matrix BASED ON THE RAW GRAM MATRIX (instead of self-kernels) // WARNING: can only work in "train set mode", i.e., with a square gram matrix void MoleculeSet::normalizeGram_raw(){ vector::iterator mol1; vector::iterator mol2; int i = 0; int j = 0; for(mol1 = begin(); mol1 != end() ; mol1++){ j = 0; for(mol2 = comparisonSet->begin() ; mol2 != comparisonSet->end() ; mol2++){ //if( ( (*mol1)->getSelfKernel() == 0 ) || ( (*mol2)->getSelfKernel() == 0 ) ){ if( ((*gram)[i][i] == 0) || ((*gram)[j][j] == 0) ){ (*gramNormal)[i][j] = 0; } else{ //(*gramNormal)[i][j] = (*gram)[i][j]/sqrt( (*mol1)->getSelfKernel() * (*mol2)->getSelfKernel() ); (*gramNormal)[i][j] = (*gram)[i][j]/sqrt( (*gram)[i][i] * (*gram)[j][j] ); } j++; } i++; } } // normalize the tanimoto coefficient void MoleculeSet::normalizeTanimoto(){ vector::iterator mol1; vector::iterator mol2; int i = 0; int j = 0; for(mol1 = begin() ; mol1 != end() ; mol1++){ j = 0; for(mol2 = comparisonSet->begin(); mol2 != comparisonSet->end() ; mol2++){ if( ( (*mol1)->getSelfKernel() == 0 ) || ( (*mol2)->getSelfKernel() == 0 ) ){ (*gramNormal)[i][j] = 0; } else{ (*gramNormal)[i][j] = (*gram)[i][j]/ ( (*mol1)->getSelfKernel() + (*mol2)->getSelfKernel() - (*gram)[i][j] ); } j++; } i++; } } // normalize the tanimoto coefficient BASED ON THE RAW GRAM MATRIX (instead of self-kernels) // WARNING: can only work in "train set mode", i.e., with a square gram matrix void MoleculeSet::normalizeTanimoto_raw(){ vector::iterator mol1; vector::iterator mol2; int i = 0; int j = 0; for(mol1 = begin(); mol1 != end() ; mol1++){ j = 0; for(mol2 = comparisonSet->begin() ; mol2 != comparisonSet->end() ; mol2++){ //if( ( (*mol1)->getSelfKernel() == 0 ) || ( (*mol2)->getSelfKernel() == 0 ) ){ if( ( (*gram)[i][i] == 0 ) || ( (*gram)[j][j] == 0 ) ){ (*gramNormal)[i][j] = 0; } else{ //(*gramNormal)[i][j] = (*gram)[i][j]/( (*mol1)->getSelfKernel() + (*mol2)->getSelfKernel() - (*gram)[i][j]); (*gramNormal)[i][j] = (*gram)[i][j]/( (*gram)[i][i] + (*gram)[j][j] - (*gram)[i][j]); } j++; } i++; } } // normalize the Min/Max tanimoto coefficient void MoleculeSet::normalizeTanimotoMinMax(){ vector::iterator mol1; vector::iterator mol2; int i = 0; int j = 0; for(mol1 = begin() ; mol1 != end() ; mol1++){ j = 0; for(mol2 = comparisonSet->begin(); mol2 != comparisonSet->end() ; mol2++){ if( (*gramNormal)[i][j] != 0){ (*gramNormal)[i][j] = (*gram)[i][j]/ (*gramNormal)[i][j]; } j++; } i++; } } // add a value to the gram matrix at the given coordinates void MoleculeSet::addToGram( int row, int col, double value ){ (*gram)[row][col] += value; } // add a value to the normal gram matrix at the given coordinates void MoleculeSet::addToGramNormal( int row, int col, double value ){ (*gramNormal)[row][col] += value; } // substract a value to the gram at the given coordinates void MoleculeSet::substractToGram( int row, int col, double value ){ (*gram)[row][col] -= value; } // return a gram matrix value double MoleculeSet::getGramValue( int row, int col ){ return (*gram)[row][col]; } // set parameters to use only training data.... void MoleculeSet::setComparisonSet(MoleculeSet *aSet){ comparisonSet = aSet; subsetStart = 0; } // list the different kinds of atoms present in the set vector MoleculeSet::atomsLabelsListing(){ vector atomLabels; vector::iterator molecule; // for each molecule : update atomLabels for( molecule = begin(); molecule != end(); molecule++ ){ (*molecule)->atomsLabelsListing( &atomLabels ); } for(molecule = comparisonSet->begin(); molecule != comparisonSet->end() ; molecule++){ (*molecule)->atomsLabelsListing( &atomLabels ); } return atomLabels; } // list the different kinds of atoms present in the set vector MoleculeSet::atomsSymbolsListing(){ vector atomSymbols; vector::iterator molecule; // for each molecule : update atomSymbols for( molecule = begin(); molecule != end(); molecule++ ){ (*molecule)->atomsSymbolsListing( &atomSymbols ); } for(molecule = comparisonSet->begin(); molecule != comparisonSet->end() ; molecule++){ (*molecule)->atomsSymbolsListing( &atomSymbols ); } return atomSymbols; } // list the different kinds of bonds present in the set vector MoleculeSet::bondsListing(){ vector bondTypes; vector::iterator molecule; // for each molecule : update bondTypes for( molecule = begin(); molecule != end(); molecule++ ){ (*molecule)->bondsListing( &bondTypes ); } for(molecule = comparisonSet->begin(); molecule != comparisonSet->end() ; molecule++){ (*molecule)->bondsListing( &bondTypes ); } return bondTypes; } long MoleculeSet::writeDotsToDir( string aDirName, bool selectedOnly, bool perretLabels ) { vector::iterator m; int i = 0; string fileName = ""; if( selectedOnly == false ){ for( m = begin(); m != end(); m++ ){ //cout << "next" << endl; //cout << (*m)->getName() << endl; if( StringUtils::right( (*m)->getName(), 4 ) == ".dot" ){ fileName = aDirName + "/" + StringUtils::slashToUnderscore( (*m)->getName() ); }else{ fileName = aDirName + "/" + StringUtils::slashToUnderscore( (*m)->getName() ) + ".dot"; } //cout << "writing " << fileName << endl; (*m)->writeDOT( fileName, perretLabels ); //cout << "done" << endl; i++; } }else{ for( m = begin(); m != end(); m++ ){ if( (*m)->isSelected() ){ if( StringUtils::right( (*m)->getName(), 4 ) == ".dot" ){ fileName = aDirName + "/" + StringUtils::slashToUnderscore( (*m)->getName() ); }else{ fileName = aDirName + "/" + StringUtils::slashToUnderscore( (*m)->getName() ) + ".dot"; } //cout << "writing " << fileName << endl; (*m)->writeDOT( fileName ); //cout << "done" << endl; i++; } } } return i; } void MoleculeSet::noTottersTransform(){ vector::iterator molecule; //for each molecule for( molecule = begin(); molecule != end(); molecule++){ (*molecule)->noTottersTransform(); } /* // DEBUG (PM): int i = 0; int j = 0; for( molecule = begin(); molecule != end(); molecule++){ i++; cout << "Molecule no : " << i << " ; " << (*molecule)->numAtoms() << " atom.s" << endl; vector::iterator anAtom; j = 0; for( anAtom = (*molecule)->beginAtom(); anAtom != (*molecule)->endAtom(); anAtom++ ) { j++; cout << "\t atom no " << j << " ; morgan label = " << (*anAtom)->getMorganLabel() << " ; ps = " << (*anAtom)->getKashimaPS() << " ; pq = " << (*anAtom)->getKashimaPQ() << " ; AN = " << (*anAtom)->getAN() <::iterator aBond; // for all the atom's bonds --> create a node for( aBond = (*anAtom)->beginBond(); aBond != (*anAtom)->endBond(); aBond++) { cout << "\t\t bond with prob :" << aBond->second->getKashimaPT() << endl; } } } exit(142); */ } // find smallest and largest distances between atoms void MoleculeSet::minMaxDistances(double *distMin, double* distMax){ // step1 : in moleculeSet vector::iterator aMol; for( aMol = begin(); aMol != end(); aMol++){ vector::iterator anAtom1, anAtom2; for(anAtom1 = (*aMol)->beginAtom() ; anAtom1 != (*aMol)->endAtom() - 1 ; anAtom1++ ){ for(anAtom2 = anAtom1 + 1 ; anAtom2 != (*aMol)->endAtom() ; anAtom2++ ){ float atomDist; atomDist = (*aMol)->atomicDistance(*anAtom1,*anAtom2); if(atomDist < *distMin) *distMin = atomDist; if(atomDist > *distMax) *distMax = atomDist; } } } // step2 : in comparisonSet if(this != comparisonSet){ for( aMol = comparisonSet->begin(); aMol != comparisonSet->end(); aMol++){ vector::iterator anAtom1, anAtom2; for(anAtom1 = (*aMol)->beginAtom() ; anAtom1 != (*aMol)->endAtom() - 1 ; anAtom1++ ){ for(anAtom2 = anAtom1 + 1 ; anAtom2 != (*aMol)->endAtom() ; anAtom2++ ){ float atomDist; atomDist = (*aMol)->atomicDistance(*anAtom1,*anAtom2); if(atomDist < *distMin) *distMin = atomDist; if(atomDist > *distMax) *distMax = atomDist; } } } } } // transform the 2D molecule set into a 3D molecule set // i.e., completely connected graphs, where edges labels are taken in {1,2,...,nBins} // where labels correspond to a discretization of the distance range into nBins bins void MoleculeSet::threeDtransform(int nBins, double distMin, double distMax){ vector::iterator aMol; // transform all molecules for( aMol = begin(); aMol != end(); aMol++){ (*aMol)->threeDtransform(nBins, distMin, distMax); } } void MoleculeSet::readPartialCharges(string fileName){ vector molCharges; ifstream fileInCharges; fileInCharges.open(fileName.c_str(), ios::in); if(!fileInCharges.good()){ CError e = CError(FILENOTFOUND, fileName + " file not found"); e.describe(); throw(e); } int linesize = 1024; char *line = new char[linesize]; while( !fileInCharges.eof() ){ fileInCharges.getline(line, linesize-1,'\n'); molCharges.push_back(line); } if( molCharges[molCharges.size()-1] == ""){ //cout << "EMPTY LINE!!!"<< endl; molCharges.pop_back(); } if(molCharges.size() != numMolecules()){ cout << "ERROR : MoleculeSet::setPartialCharges" << endl; cout << " --> number of lines read in charges file != numMolecules" << endl; cout << " - numMolecules = " << numMolecules() << " ; number of lines = " << molCharges.size() << endl; cout << " last entry : " << molCharges[molCharges.size()-1] << endl; exit(12); } int i = 0; vector::iterator m; for( m = begin(); m != end(); m++ ){ (*m)->readPartialCharges(molCharges[i]); i++; } fileInCharges.close(); } void MoleculeSet::setMorganChargesLabels(double threshold){ vector::iterator m; for( m = begin(); m != end(); m++ ){ (*m)->setMorganChargesLabels(threshold); } } // ********************************* // // ***** DEPRECATED FUNCTIONS ***** // // ******************************** // //normalize the gram matrix, i.e. fill in gramNormal //void MoleculeSet::normalizeGram_self(){ // for( int i = 0 ; i < numMolecules() ; i++ ){ // for( int j = i ; j < numMolecules() ; j++ ){ // (*gramNormal)[i][j] = (*gram)[i][j] / sqrt( (*gram)[i][i] * (*gram)[j][j] ); // (*gramNormal)[j][i] = (*gram)[i][j] / sqrt( (*gram)[i][i] * (*gram)[j][j] ); // } // } //} // // normalize the gram matrix, i.e. fill in gramNormal WITH RBF NORMALIZATION // void MoleculeSet::normalizeGram_self(){ // double dist2, sigma2, ker; // sigma2 = 1.0; // for( int i = 0 ; i < numMolecules() ; i++ ){ // for( int j = i ; j < numMolecules() ; j++ ){ // dist2 = (*gram)[i][i] + (*gram)[j][j] - 2 * (*gram)[i][j]; // ker = exp( -dist2/sigma2 ); // (*gramNormal)[i][j] = ker; // (*gramNormal)[j][i] = ker; // } // } // } // normalize the gram matrix, i.e. fill in gramNormal //void MoleculeSet::normalizeGram_test(){ //vector::iterator mol1; // vector::iterator mol2; // int i = 0; // int j = 0; // for(mol1 = begin() ; mol1 != end() ; mol1++){ // j = 0; // for(mol2 = comparisonSet->begin(); mol2 != comparisonSet->end() ; mol2++){ // (*gramNormal)[i][j] = (*gram)[i][j]/sqrt( (*mol1)->getSelfKernel() * (*mol2)->getSelfKernel() ); // j++; // } // i++; // } //} // // normalize the gram matrix, i.e. fill in gramNormal WITH RBF NORMALIZATION // void MoleculeSet::normalizeGram_test(){ // vector::iterator mol1; // vector::iterator mol2; // int i = 0; // int j = 0; // double dist2, sigma2, ker; // sigma2 = 1.0; // for(mol1 = begin() ; mol1 != end() ; mol1++){ // j = 0; // for(mol2 = comparisonSet->begin(); mol2 != comparisonSet->end() ; mol2++){ // dist2 = (*mol1)->getSelfKernel() + (*mol2)->getSelfKernel() - 2 * (*gram)[i][j]; // ker = exp( -dist2/sigma2 ); // (*gramNormal)[i][j] = ker; // j++; // } // i++; // } // } /*struct ToLower { char operator() (char c) const { return std::tolower(c); } } struct ToUpper { char operator() (char c) const { return std::toupper(c); } }*/ /*void* kashimaKernelThread(void* arg){ kashimaThreadStruct *str = (kashimaThreadStruct *) arg; *str->kernel = (str->molecule1)->newKernel( str->molecule2, str->convergence); float sk1 = (str->molecule1)->getSelfNewKernel(); float sk2 = (str->molecule2)->getSelfNewKernel(); *str->normalKernel = *str->kernel / sqrt(sk1 * sk2); // cout << " comparing " << (str->molecule1)->getName() << " with " << (str->molecule2)->getName() << " " << "K = " << *str->kernel; // cout << " storing result in " << str << endl; str->noErrors = true; pthread_exit(0); } */ Rchemcpp/src/chemcpp/src/cerror.h0000644000175000017500000001012012116123101017363 0ustar klambaueklambaue /**************************************************************************************** cerror.h ------------ copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef CERROR_H #define CERROR_H #include #include using std::string; using std::endl; using std::cout; using std::cerr; // Error codes #define BADFILE 1 #define MISSINGDESCRIPTOR 2 #define ATOMALREADYEXISTS 3 #define NOTENOUGHATOMSINMOLECULE 4 #define FILENOTFOUND 5 #define ERRORNA 6 #define UNKNOWNDATATYPE 7 #define ERRORATOMNOTFOUND 8 #define VALUENOTALLOWED 9 #define NOTIMPLEMENTED 10 #define NOTFOUND 11 #define COULDNOTOPENFILE 12 #define DUPLICATEENTRIES 13 #define DEPRECATED 14 #define THREADCREATIONERROR 15 #define EOFERROR 16 #define MISSINGDATA 17 #define NOTCALCULATED 18 #define BADVALUE 19 #define BONDALREADYEXISTS 20 #define IOERROR 21 #define NOSTRUCTURE 22 #define ATOMNOTFOUND 23 #define SSSRNOTDETECTED 24 #define MISSINGRING 25 #define ATOMNOTINRING 26 #define ALLNEIGHBOURSVISITED 27 #define BONDNOTFOUND 28 #define MOLECULENOTFOUND 29 /** CError class thrown when errors occur in chemcpp @author Dr Jean-Luc Perret (luc@kuicr.kyoto-u.ac.jp), Kyoto University, Japan @version 0.3 @date 17 Jan 2004 CLASS NAME: CError FOR: SNSF SPONSORED PROJECT PURPOSE: implements error message thrown by the classes in the chemcpp project. When an error occurs an instance of this class is created and thrown. The class has two attributes: a type (integer why) and a comment (string comment). These two attributes can be retrieved using the getType() and getComment() functions. An error message corresponding to the error can also be printed on cerr using the describe() function. The following error types are defined: - BADFILE 1 - MISSINGDESCRIPTOR 2 - ATOMALREADYEXISTS 3 - NOTENOUGHATOMSINMOLECULE 4 - FILENOTFOUND 5 - ERRORNA 6 - UNKNOWNDATATYPE 7 - ERRORATOMNOTFOUND 8 - VALUENOTALLOWED 9 - NOTIMPLEMENTED 10 - NOTFOUND 11 - COULDNOTOPENFILE 12 - DUPLICATEENTRIES 13 - DEPRECATED 14 - THREADCREATIONERROR 15 - EOFERROR 16 - MISSINGDATA 17 - NOTCALCULATED 18 - BADVALUE 19 - BONDALREADYEXISTS 20 - IOERROR 21 - NOSTRUCTURE 22 - ATOMNOTFOUND 23 - SSSRNOTDETECTED 24 - MISSINGRING 25 - ATOMNOTINRING 26 - ALLNEIGHBOURSVISITED 27 - BONDNOTFOUND 28 - MOLECULENOTFOUND 29 */ class CError { /** \example cerror_example.cpp */ public: /** class constructor. */ CError( int anError, string aComment ); /** class destructor. */ ~CError(); /** returns the error type. */ int getType() { return why; } /** returns the comment associated with the error. */ string getComment() { return comment; } /** prints a description of the error to stderr. */ void describe(); private: /** error type. */ int why; /** error comment. */ string comment; }; #endif Rchemcpp/src/chemcpp/src/molecule.cpp0000644000175000017500000026376612116123101020261 0ustar klambaueklambaue/**************************************************************************************** molecule.cpp ---------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #include "molecule.h" #include "moleculeutils.h" //#define DEBUGK 1 //#define DEBUG 1 //#define DEBUGPERRETLABEL 1 //#define DEBUGSSSR 1 int Molecule::counter = 0; Molecule::Molecule() : DataContainer(){ fastPT = new map* >; fastPTSave = new map* >; fastPTNext = NULL; Molecule::counter++; id = Molecule::counter; selfKernelCalculated = false; addStringDescriptor( "name" ,"", "", "molecule name" ); addStringDescriptor( "comment" ,"", "", "comment" ); addStringDescriptor( "comment2" ,"", "", "comment2" ); addStringDescriptor( "comment3" ,"", "", "comment3" ); //setIntDescriptor( ACTIVITY ,98, "", "biological activity" ); selectedFlag = false; chiral=false; location = ""; flagHasSSSRDetected = false; moleculeChanged(true,true); activity = 0; flagActivity = false; originalFormat = NAVALUE; } Molecule::Molecule( Molecule& aMolecule , bool bool_resetMorganIndex ) : DataContainer( (DataContainer&) aMolecule ){ fastPT = new map* >; fastPTSave = new map* >; fastPTNext = NULL; #ifdef DEBUG cout << "Molecule COPY operator" << endl; #endif //setStringDescriptor( "name" ,"", "", "molecule name" ); //setStringDescriptor( "comment" ,"", "", "comment" ); addStringDescriptor( "name" ,"", "", "molecule name" ); addStringDescriptor( "comment" ,"", "", "comment" ); addStringDescriptor( "comment2" ,"", "", "comment2" ); addStringDescriptor( "comment3" ,"", "", "comment3" ); this -> selectedFlag = aMolecule.selectedFlag; this -> chiral = aMolecule.chiral; this -> maxMorganIteration = aMolecule.maxMorganIteration; this -> selfKernel = aMolecule.selfKernel; this -> selfKernelCalculated = aMolecule.selfKernelCalculated; this -> sortDescriptorName = aMolecule.sortDescriptorName; this -> sortDescriptorType = aMolecule.sortDescriptorType; this -> activity = aMolecule.activity; this -> flagActivity = aMolecule.flagActivity; this->location = aMolecule.location; this->originalFormat = aMolecule.originalFormat; this-> flagHasSSSRDetected = aMolecule.flagHasSSSRDetected; // increment the molecule id and set the id for this molecule Molecule::counter++; id = Molecule::counter; selfKernelCalculated = false; // add atoms and make a map of corresponding atoms in original molecule map match; // old and new atom vector::iterator iit; for( iit = aMolecule.beginAtom(); iit != aMolecule.endAtom(); iit++ ){ Atom* newAtom = new Atom( *(*iit) ); atoms.push_back( newAtom ); match[(*iit)] = newAtom; newAtom->setCoordinates( (*iit)->getX(), (*iit)->getY(), (*iit)->getZ() ); newAtom->setAN( (*iit)->getAN() ); newAtom->setType( (*iit)->getType() ); newAtom->setMorganLabel( (*iit)->getMorganLabel() ); } // add bonds // THIS SHOULD BE IMPLEMENTED YET map< Atom*, Bond*>::iterator bi; for( iit = aMolecule.beginAtom(); iit != aMolecule.endAtom(); iit++ ){ for( bi = (*iit)->beginBond(); bi != (*iit)->endBond(); bi++ ){ Bond* newBond = new Bond( match[(*iit)], match[(*bi).first], (*bi).second->getLabel() ); if( (*bi).second->hasFlagOriginal() ){ newBond->setFlagOriginal(); }else{ newBond->unsetFlagOriginal(); } match[(*iit)]->addBond( newBond ); } //match[(*iit)]->describeShort(); } // copy int descriptors map* >::iterator dii; for( dii = aMolecule.beginIntDescriptor(); dii != aMolecule.endIntDescriptor(); dii++ ){ if( (*dii).second->isEmpty() ){ Descriptor* d = addIntDescriptor( (*dii).second->getLabel(), -1, (*dii).second->getUnit(), (*dii).second->getComment() ); d->setEmpty(); }else{ addIntDescriptor( (*dii).second->getLabel(), (*dii).second->getValue(), (*dii).second->getUnit(), (*dii).second->getComment() ); } } // copy float descriptors map* >::iterator dif; for( dif = aMolecule.beginFloatDescriptor(); dif != aMolecule.endFloatDescriptor(); dif++ ){ if( (*dif).second->isEmpty() ){ Descriptor* d = addFloatDescriptor( (*dif).second->getLabel(), -1.0, (*dif).second->getUnit(), (*dif).second->getComment()); d->setEmpty(); }else{ addFloatDescriptor( (*dif).second->getLabel(), (*dif).second->getValue(), (*dif).second->getUnit(), (*dif).second->getComment()); } } // copy string descriptors map* >::iterator dis; for( dis = aMolecule.beginStringDescriptor(); dis != aMolecule.endStringDescriptor(); dis++ ){ if( (*dis).second->isEmpty() ){ Descriptor* d = addStringDescriptor( (*dis).second->getLabel(), "", (*dis).second->getUnit(), (*dis).second->getComment() ); d->setEmpty(); }else{ addStringDescriptor( (*dis).second->getLabel(), (*dis).second->getValue(), (*dis).second->getUnit(), (*dis).second->getComment() ); } } //cout << "WARNING: the molecule copy operator does not copy fastPQ, fastPS and fastPT!!!" << endl; if(bool_resetMorganIndex == true) { resetMorganIndex(); } } Molecule::Molecule( Molecule& m1, Molecule& m2, double (*pt2AtomKernel)( Atom*, Atom* ), double (*pt2BondKernel)( Bond*, Bond* ) ) : DataContainer(){ atoms.clear(); fastPT = new map* >; fastPTSave = new map* >; fastPTNext = NULL; Molecule::counter++; id = Molecule::counter; selfKernelCalculated = false; addStringDescriptor( "name" ,"", "", "molecule name" ); addStringDescriptor( "comment" ,"", "", "comment" ); addStringDescriptor( "comment2" ,"", "", "comment2" ); addStringDescriptor( "comment3" ,"", "", "comment3" ); //setIntDescriptor( ACTIVITY ,98, "", "biological activity" ); selectedFlag = false; chiral=false; activity = 0; flagActivity = false; flagHasSSSRDetected = false; vector::iterator iit; vector::iterator jit; vector< Atom*> source1; vector< Atom*> source2; vector< Atom*> newAtoms; // get a list of all matching atoms in both graphs // and add all atoms matching in both graphs #ifdef DEBUG cout << "adding atoms" << endl; #endif int ac = 0; for ( iit = m1.beginAtom(); iit != m1.endAtom(); iit++ ){ for ( jit = m2.beginAtom(); jit != m2.endAtom(); jit++ ){ if ( pt2AtomKernel( *iit, *jit ) == 1 ){ #ifdef DEBUG cout << " " << ac << endl; cout << " " << (*iit)->toStringShort() << endl; #endif Atom* newAtom = new Atom( *(*iit) ); this->addAtom( newAtom , true, true); newAtoms.push_back( newAtom ); ac++; source1.push_back( *iit ); source2.push_back( *jit ); try{ double apq = (*iit)->getKashimaPQ( true ) * (*jit)->getKashimaPQ( true ); newAtom->setKashimaPQ( apq ); fastPQ[newAtom] = apq; } catch( CError e ){ cout << "WARNING: please set PQ before computing fused graph" << endl; } try{ double aps = (*iit)->getKashimaPS( true ) * (*jit)->getKashimaPS( true ); newAtom->setKashimaPS( aps ); fastPS[newAtom] = aps; } catch( CError e ){ cout << "WARNING: please set PS before computing fused graph" << endl; } } } } #ifdef DEBUG cout << "adding atoms DONE" << endl; cout << " " << m1.toStringShort() << " - " << m2.toStringShort() << ": product graph has " << ac << " atoms" << endl; #endif Bond* bond1 = NULL; Bond* bond2 = NULL; Bond* bond1r = NULL; Bond* bond2r = NULL; for( int i = 0; i < numAtoms()-1; i++ ){ //cout << " " << i << endl; for( int j = i+1; j < numAtoms(); j++ ){ //cout << " " << j << endl; //cout << source1[i]->toStringShort() << " " << source2[i]->toStringShort() << endl; //cout << source1[j]->toStringShort() << " " << source2[j]->toStringShort() << endl; try{ bond1 = source1[i]->getBondWithTarget( source1[j] ); bond2 = source2[i]->getBondWithTarget( source2[j] ); bond1r = source1[j]->getBondWithTarget( source1[i] ); bond2r = source2[j]->getBondWithTarget( source2[i] ); //cout << "checking " << source1[i]->toStringShort() << " with " << source1[j]->toStringShort() << " -> " << bond1->getLabel() << endl; //cout << "checking " << source2[i]->toStringShort() << " with " << source2[j]->toStringShort() << " -> " << bond2->getLabel() << endl; //if( bondType1 == bondType2 ){ if( pt2BondKernel( bond1, bond2 ) ){ //cout << "Linking " << endl; Atom* aSource = newAtoms[i]; Atom* aTarget = newAtoms[j]; Bond* forwardBond = new Bond( aSource, aTarget, 1 ); Bond* backwardBond = new Bond( aTarget, aSource, 1 ); double fpt = bond1->getKashimaPT() * bond2->getKashimaPT(); double bpt = bond1r->getKashimaPT() * bond2r->getKashimaPT(); forwardBond->setKashimaPT( fpt ); backwardBond->setKashimaPT( bpt ); backwardBond->unsetFlagOriginal(); forwardBond->setFlagOriginal(); aSource->addBond( forwardBond ); aTarget->addBond( backwardBond ); map* aMap = (*fastPT)[aSource]; if( aMap == NULL ){ aMap = new map; (*aMap)[aTarget] = fpt; (*fastPT)[aSource] = aMap; }else{ (*aMap)[aTarget] = fpt; } aMap = (*fastPT)[aTarget]; if( aMap == NULL ){ aMap = new map; (*aMap)[aSource] = bpt; (*fastPT)[aTarget] = aMap; }else{ (*aMap)[aSource] = bpt; } aMap = (*fastPTSave)[aSource]; if( aMap == NULL ){ aMap = new map; (*aMap)[aTarget] = fpt; (*fastPTSave)[aSource] = aMap; }else{ (*aMap)[aTarget] = fpt; } aMap = (*fastPTSave)[aTarget]; if( aMap == NULL ){ aMap = new map; (*aMap)[aSource] = bpt; (*fastPTSave)[aTarget] = aMap; }else{ (*aMap)[aSource] = bpt; } //cout << aSource->toStringShort() << " - " << aTarget->toStringShort() << ": " << fpt << " " << bpt << endl; //cout << (*fastPTSave[aSource])[aTarget] << " " <<(*fastPTSave[aTarget])[aSource]<< endl; //cout << (*fastPT[aSource])[aTarget] << " " <<(*fastPT[aTarget])[aSource]<< endl; //linkAtoms( newAtoms[i], newAtoms[j], bondType1 ); } }catch( CError ){ } //cout << "####" << endl; } } source1.clear(); source2.clear(); newAtoms.clear(); location = ""; originalFormat = NAVALUE; //resetMorganIndex(); //cout << "DONE" << endl; } // 3D product graph constructor // //------------------------------// Molecule::Molecule( Molecule& m1, Molecule& m2, double (*pt2AtomKernel)( Atom*, Atom*), double (*pt2BondKernel)( float, float, float ), float edgeKernelParameter ) : DataContainer(){ // step 0 : initialization // ----------------------- atoms.clear(); Molecule::counter++; id = Molecule::counter; addStringDescriptor( "name" ,"", "", "molecule name" ); addStringDescriptor( "comment" ,"", "", "comment" ); addStringDescriptor( "comment2" ,"", "", "comment2" ); addStringDescriptor( "comment3" ,"", "", "comment3" ); //setIntDescriptor( ACTIVITY ,98, "", "biological activity" ); selfKernelCalculated = false; selectedFlag = false; chiral=false; activity = 0; flagActivity = false; flagHasSSSRDetected = false; // NOT USED YET, but need to be declared when deleting molecule fastPT = new map* >; fastPTSave = new map* >; fastPTNext = NULL; // step 1 : build atoms // -------------------- vector::iterator iit; vector::iterator jit; vector< Atom*> source1; vector< Atom*> source2; vector< Atom*> newAtoms; // get a list of all matching atoms in both graphs // and add all atoms matching in both graphs int ac = 0; double k_at1_at2 = 0.0; for ( iit = m1.beginAtom(); iit != m1.endAtom(); iit++ ){ for ( jit = m2.beginAtom(); jit != m2.endAtom(); jit++ ){ k_at1_at2 = pt2AtomKernel( *iit, *jit); if ( k_at1_at2 != 0 ){ Atom* newAtom = new Atom( *(*iit) ); this->addAtom( newAtom , true, true); newAtoms.push_back( newAtom ); ac++; source1.push_back( *iit ); source2.push_back( *jit ); } } } // step 2 : build edges // -------------------- // --> compute bonds + adjacency matrix double k_e1_e2 = 0; double k_n1_n2 = 0; double k_n2_n1 = 0; float d1 = 0; float d2 = 0; //init adjacency matrix + walks matrix adjacency = new vector< vector >; walks = new vector< vector >; for(int i = 0 ; i < numAtoms() ; i++){ adjacency->push_back(vector() ); walks->push_back(vector() ); for(int j = 0 ; j < numAtoms() ; j++){ (*adjacency)[i].push_back(0.0); (*walks)[i].push_back(0.0); } } for( int i = 0; i < numAtoms()-1; i++ ){ for( int j = i+1; j < numAtoms(); j++ ){ d1 = atomicDistance(source1[i],source1[j]); d2 = atomicDistance(source2[i],source2[j]); k_e1_e2 = pt2BondKernel(d1,d2,edgeKernelParameter); // WARNING : new test to be sure to introduce a correct edge if( source1[i] == source1[j] || source2[i] == source2[j] ){ k_e1_e2 = 0; } //else{ // cout << "distance 1 : " << d1 << endl; // cout << "distance 2 : " << d2 << endl; //} k_n1_n2 = pt2AtomKernel(source1[i],source2[i]); k_n2_n1 = pt2AtomKernel(source1[j],source2[j]); //cout << "#### atom kernel value : " << k_n2_n1 << endl; //cout << "#### edge kernel value : " << k_e1_e2 << endl; // 1 - set adjacency matrix + walks matrix setAdjacency(i, j, k_n1_n2 * k_e1_e2); setAdjacency(j, i, k_n2_n1 * k_e1_e2); setWalks(i, j, k_n1_n2 * k_e1_e2); setWalks(j, i, k_n2_n1 * k_e1_e2); // 2 - introdce atomic bonds : a pair of bonds : forward (i->j) + backward (j->i) if(k_e1_e2 != 0){ Atom* aSource = newAtoms[i]; Atom* aTarget = newAtoms[j]; Bond* forwardBond = new Bond( aSource, aTarget, 1 ); Bond* backwardBond = new Bond( aTarget, aSource, 1 ); forwardBond->setKashimaPT( k_n1_n2 * k_e1_e2 ); backwardBond->setKashimaPT( k_n2_n1 * k_e1_e2 ); //backwardBond->unsetFlagOriginal(); //forwardBond->setFlagOriginal(); aSource->addBond( forwardBond ); aTarget->addBond( backwardBond ); } } } } // function to compute the Euclidian distance between a pair of atoms float Molecule::atomicDistance(Atom* atom1, Atom* atom2){ float distance; distance = (atom1->getX() - atom2->getX())*(atom1->getX() - atom2->getX()) + (atom1->getY() - atom2->getY())*(atom1->getY() - atom2->getY()) + (atom1->getZ() - atom2->getZ())*(atom1->getZ() - atom2->getZ()); return( sqrt(distance) ); } /** this function should be called everytime a molecule is modified */ void Molecule::compute(){ #ifdef DEBUG cout << "NOW IN MOLECULE::COMPUTE " << endl; #endif if( !hasSSSRDetected() ){ // detectSSSR(); } setMorganLabels(0); try{ setPerretLabels(); }catch( CError e ){ cout << "Warning: SSSR was not computed in molecule " << toStringShort() << " so cannot set Perret Labels, skippink!" << endl; } } void Molecule::moleculeChanged( bool resetSSSR, bool bool_resetMorganIndex ){ if( bool_resetMorganIndex ){ resetMorganIndex(); } if( resetSSSR ){ flagHasSSSRDetected = false; } } Molecule& Molecule::operator=( const Molecule& aMolecule ) { #ifdef DEBUG cout << "Molecule = operator" << endl; #endif if( this != &aMolecule ){ this->kindStringDescriptors = aMolecule.kindStringDescriptors; this->kindIntDescriptors = aMolecule.kindIntDescriptors; this->kindFloatDescriptors = aMolecule.kindFloatDescriptors; this -> selectedFlag = aMolecule.selectedFlag; this -> location = aMolecule.location; // Copy all non kind descriptors // first copy int descriptors map* >::iterator iti; for( iti = (intDescriptors).begin(); iti != (intDescriptors).end(); iti++ ){ addIntDescriptor( (*iti).second->getLabel() , (*iti).second->getValue(), (*iti).second->getUnit(), (*iti).second->getComment() ); } map* >::iterator itf; for( itf = (floatDescriptors).begin(); itf != (floatDescriptors).end(); itf++ ){ addFloatDescriptor( (*itf).second->getLabel() , (*itf).second->getValue(), (*itf).second->getUnit(), (*itf).second->getComment() ); } map* >::iterator its; for( its = (stringDescriptors).begin(); its != (stringDescriptors).end(); its++ ){ addStringDescriptor( (*its).second->getLabel() , (*its).second->getValue(), (*its).second->getUnit(), (*its).second->getComment() ); } // increment the atom id and set the id for this atom Molecule::counter++; id = Molecule::counter; // selfKernelCalculated = false; } //resetMorganIndex(); ***??? return( *this ); } Molecule::~Molecule(){ // delete all atoms erase(); delete fastPT; delete fastPTSave; delete fastPTNext; } Atom* Molecule::addAtom( string aSymbol, bool resetSSSR ) throw( CError ){ /*ToUpper up(std::locale::classic()); ToLower down(std::locale::classic()); std::transform (aSymbol.begin(), aSymbol.begin()+1, aSymbol.begin(), up); if( aSymbol.length() > 1 ){ std::transform (aSymbol.begin()+1, aSymbol.end(), aSymbol.begin()+1, down); } */ string h = aSymbol.substr(0, 1); string r = ""; if( aSymbol.length() > 1 ){ r = aSymbol.substr( 1, aSymbol.length()-1 ); } h = StringUtils::toUpper( h ); r = StringUtils::toLower( r ); aSymbol = h + r; Atom* anAtom = NULL; try{ anAtom = new Atom( *elements[aSymbol] ); }catch( CError e ){ //cerr << "WARNING could not add element " << aSymbol << endl; throw( e ); } atoms.push_back(anAtom); moleculeChanged( resetSSSR , true); #ifdef DEBUG //cout << "created atom " << anAtom->toString() << endl; #endif return( anAtom ); } Atom* Molecule::addAtom( Atom* anAtom, bool resetSSSR, bool resetMorganIndex ) throw( CError ){ // if atom does not already exist, then add the new atom if( atomExists( anAtom ) ){ //cout << "ERROR ATOM EXISTS!!!!" << endl; stringstream out; out << "Atom " << anAtom->toString() << " already exists "; CError e( ATOMALREADYEXISTS, out.str() ); e.describe(); throw(e); //return( anAtom ); }else{ atoms.push_back( anAtom ); //cout << "adding new atom " << anAtom->toString() << endl; moleculeChanged( resetSSSR, resetMorganIndex ); return( anAtom ); } } /** check whether Atom exists in the atom vector */ bool Molecule::atomExists(Atom* anAtom){ vector< Atom* >::iterator i; for( i = atoms.begin(); i != atoms.end(); i++ ){ if( *i == anAtom ){ return( true ); } } return( false ); } Atom* Molecule::getAtom(int anId) throw( CError ){ vector::iterator atom; // for each atom for( atom = beginAtom(); atom != endAtom(); atom++ ){ if( (*atom)->getId() == anId ){ return( *atom ); } } stringstream out; out << "No atom with id " << anId << " in molecule " << getId(); CError e(ERRORATOMNOTFOUND, out.str() ); e.describe(); describeLong(); throw(e); } Atom* Molecule::getAtomByIndex(int ind){ return atoms[ind]; } Bond* Molecule::linkAtoms( Atom* aSource, Atom* aTarget, int aBondLabel, int aBondStereo, int aBondNotUsed, int aBondTopology, int aBondReactionCenter, bool resetSSSR ){ if( !atomExists( aSource ) ){ addAtom( aSource , true, true); } if( !atomExists( aTarget ) ){ addAtom( aTarget , true , true ); } Bond* forwardBond = new Bond(aSource, aTarget, aBondLabel, NAVALUE, aBondStereo, aBondNotUsed, aBondTopology, aBondReactionCenter ); Bond* backwardBond = new Bond(aTarget, aSource, aBondLabel, NAVALUE, aBondStereo, aBondNotUsed, aBondTopology, aBondReactionCenter); backwardBond->unsetFlagOriginal(); forwardBond->setFlagOriginal(); aSource->addBond(forwardBond); aTarget->addBond(backwardBond); moleculeChanged( resetSSSR, true ); return( forwardBond ); } Bond* Molecule::linkAtomsNoReturn( Atom* aSource, Atom* aTarget, int aBondLabel, int aBondStereo, int aBondNotUsed, int aBondTopology, int aBondReactionCenter ){ //addAtom( aSource ); //addAtom( aTarget ); Bond* forwardBond = new Bond( aSource, aTarget, aBondLabel, NAVALUE, aBondStereo, aBondNotUsed, aBondTopology, aBondReactionCenter ); //forwardBond->setFlagOriginal(); forwardBond = aSource->addBond( forwardBond ); //resetMorganIndex(); // WARNING !!!! // --> modified by PM ?? // --> PM needs this modification in function Molecule::threeDtransform // in order not to loose the MI related to te 2D structure when switching to 3D structure (i.e., completely connected graph) return( forwardBond ); } Bond* Molecule::linkAtoms( int aSource, int aTarget, int aBondLabel, int aBondStereo, int aBondNotUsed, int aBondTopology, int aBondReactionCenter, bool resetSSSR ) throw( CError ){ // atoms must exists otherwise cannot link if( max(aSource, aTarget) > numAtoms()-1 ){ stringstream out; out << "molecule " << toString() << " has only " << numAtoms() << " atoms, so cannot link atom " << aSource << " to " << aTarget ; CError e = CError(NOTENOUGHATOMSINMOLECULE, out.str() ); e.describe(); throw(e); }else{ try{ Bond* forwardBond = linkAtoms( atoms[aSource], atoms[aTarget], aBondLabel, aBondStereo, aBondNotUsed, aBondTopology, aBondReactionCenter, resetSSSR); return( forwardBond ); }catch( CError e){ throw( e ); } } } Bond* Molecule::linkAtoms(string aSource, string aTarget, int aBondLabel, int aBondStereo, int aBondNotUsed, int aBondTopology, int aBondReactionCenter, bool resetSSSR ) throw( CError ){ // atoms must exists otherwise cannot link int a1 = 0; int a2 = 0; int i = 0; bool foundSource = false; bool foundTarget = false; // find aSource and aTarget vector::iterator atom; for( atom = beginAtom(); atom != endAtom(); atom++ ){ if( (*atom)->getName() == aSource) { a1 = i; foundSource = true; } if( (*atom)->getName() == aTarget) { a2 = i; foundTarget = true; } i++; } if( foundSource == false ){ stringstream out; out << "Molecule::linkAtoms: atom " << aSource << " has not been found in molecule " << getName(); CError e = CError(NOTFOUND, out.str() ); e.describe(); throw(e); } if( foundTarget == false ){ stringstream out; out << "Molecule::linkAtoms: atom " << aTarget << " has not been found in molecule " << getName(); CError e = CError(NOTFOUND, out.str() ); e.describe(); throw(e); } #ifdef DEBUG //cout << "linking " << atoms[a1]->getName() << " to " << atoms[a2]->getName() << endl; #endif Bond* forwardBond = linkAtoms( atoms[a1], atoms[a2], aBondLabel, aBondStereo, aBondNotUsed, aBondTopology, aBondReactionCenter, resetSSSR); return( forwardBond ); } /** returns a string description of the molecule */ string Molecule::toString(){ stringstream out; //out << "Molecule " << getName() << " (id = " << getId() << ")"; //if( isSelected() ){ // out << " SELECTED"; //} out << toStringShort(); try{ out << " at " << this << ", has " << numAtoms() << " atoms ( + " << numHiddenAtoms() << " hidden atoms ) and " << numBonds() << " bonds ( + " << numHiddenBonds() << " hidden bonds ), activity: " << getActivity(); } catch( CError ){ out << " at " << this << ", has " << numAtoms() << " atoms ( + " << numHiddenAtoms() << " hidden atoms ) and " << numBonds() << " bonds ( + " << numHiddenBonds() << " hidden bonds ), no activity"; } out << " / " << sssr.size() << " rings, size: "; vector::iterator ri; for( ri = sssr.begin(); ri != sssr.end(); ri++ ){ out << (*ri)->size() << " "; } return( out.str() ); } string Molecule::toStringShort(){ stringstream out; out << "Molecule " << getName() << " (id = " << getId() << ") "; if( isSelected() ){ out << " SELECTED"; } return( out.str() ); } string Molecule::toStringLong(){ stringstream out; vector::iterator atom; map::iterator bond; out << toString() << endl; try{ if( numRings() == 0 ) { out << " no rings "; } }catch( CError ){ out << " sssr not detected "; } // for each atom out << "ATOMS: " << endl; for( atom = beginAtom(); atom != endAtom(); atom++ ){ out << (*atom)->toString(); try{ out << " " << (*atom)->getMorganLabel(true); }catch( CError e ){ out << " morganLabel NOT SET"; } try{ out << " (ps=" << (*atom)->getKashimaPS(true) << ", pq=" << (*atom)->getKashimaPQ(true) << ")"; }catch( CError e ){ out << " (ps=NOT SET, pq=NOT SET)"; } /*try{ if( hasRing() ){ if( (*atom)->numRings() > 0 ){ out << " " << (*atom)->numRings(); }else{ out << " no rings " << endl; } } }catch( CError e ){ out << " rings not detected" << endl; }*/ out << endl; // and for each bond // for( bond = (*atom)->beginBond(); bond != (*atom)->endBond(); bond++ ){ // cout << " " << (*bond).second->toString() << " (pt=" << (*bond).second->getKashimaPT() << ")" <beginBond(); bond != (*atom)->endBond(); bond++ ){ out << (*bond).second->toString() << endl; } } out << "Number of Distinct Morgan Indices: " << getNumberOfDistinctMorganIndices( 1 ) << " " << getNumberOfDistinctMorganIndices( 2 ) << " " << getNumberOfDistinctMorganIndices( 3 ) << " " << endl; out << "Maximum number of distinct Morgan Indices reached after " << getMaxMorganIteration() << " iterations with " << getNumberOfDistinctMorganIndices( getMaxMorganIteration() ) << " distinct connectivity values" << endl; return( out.str() ); } /** convenience function to retrieve the id as a string */ string Molecule::getIdString(){ stringstream out; out << getId(); return( out.str() ); } float Molecule::getActivity( bool silentMode ) throw( CError ){ if( hasActivity() ){ return( activity ); }else{ stringstream out; out << toStringShort() << " has no activity"; CError e( ERRORNA, out.str() ); if( silentMode != true ){ e.describe(); } throw(e); } } /** write a description of the molecule to stdout */ void Molecule::describe(){ cout << toString() << endl; //DataContainer::describe(); } void Molecule::describeShort(){ cout << toStringShort() << endl; } void Molecule::describeLong(){ cout << toStringLong() << endl; } void Molecule::describeEachAtom(){ vector::iterator atom; map::iterator bond; cout << "Molecule " << getId() << " has " << numAtoms() << " atoms: " << endl; // for each atom for( atom = beginAtom(); atom != endAtom(); atom++ ){ (*atom)->describe(); // describe bonds cout << " " << "this atom has " << (*atom)->numBonds() << " bonds:" << endl; for( bond = (*atom)->beginBond(); bond != (*atom)->endBond(); bond++ ){ (*bond).second->describe(); } } } /** Sets the start, stop and transition probabilities for the calculation of the Kashima kernel as published */ void Molecule::setKashimaKernelProb( double aPq, bool skipSkeleton ){ // aPq must be a probability between 0 and 1 if not, this is an error if(aPq > 1 || aPq < 0){ stringstream out; out << aPq << " is not a valid stop probability (should be between 0 and 1) "; CError e(VALUENOTALLOWED, out.str() ); e.describe(); throw(e); } if( skipSkeleton == false ){ int numatoms = numAtoms(); if(numatoms>0){ vector::iterator atom; map::iterator bond; int numbonds = 0; for( atom = beginAtom(); atom != endAtom(); atom++ ){ // for each atom set the start probability to 1/numAtoms (*atom)->setKashimaPS( 1/(double) numatoms ); // for each bond set the transition probability to 1/numbonds numbonds = (*atom)->numBonds(); for( bond = (*atom)->beginBond(); bond != (*atom)->endBond(); bond++ ){ // For each bond set the transition probability (*bond).second->setKashimaPT( (1.0 - aPq) / numbonds ); } // for each atom set the stop probability to (1-aPq)/numBonds (*atom)->setKashimaPQ( aPq ); } } }else{ cout << "setting rdmwk parameters for molecule " << getName() << endl; int numatoms = numAtomsNonCSkeleton(); cout << "found " << numatoms << " non skeleton atoms " << endl; if(numatoms>0){ vector::iterator atom; map::iterator bond; int numbonds = 0; for( atom = beginAtom(); atom != endAtom(); atom++ ){ // for each atom set the start probability to 1/numAtoms if( (*atom)->isCSkeleton() ){ (*atom)->setKashimaPS( 0.0 ); }else{ (*atom)->setKashimaPS( 1/(double) numatoms ); } // for each bond set the transition probability to 1/numbonds numbonds = (*atom)->numBonds(); for( bond = (*atom)->beginBond(); bond != (*atom)->endBond(); bond++ ){ // For each bond set the transition probability (*bond).second->setKashimaPT( (1.0 - aPq) / numbonds ); } // for each atom set the stop probability to (1-aPq)/numBonds (*atom)->setKashimaPQ( aPq ); } } } } double Molecule::computeKernel( Molecule* anotherMolecule, double (*pt2GraphKernel) ( Molecule* mol1, Molecule* mol2, double (*pt2AtomKernel)( Atom*, Atom* ), double (*pt2BondKernel)( Bond*, Bond* ), int parameter1, int parameter2 ), double (*pt2AtomKernel)( Atom*, Atom* ), double (*pt2BondKernel)( Bond*, Bond* ), int parameter1, int parameter2 ){ //cout << "START" << endl; double k = pt2GraphKernel( this, anotherMolecule, pt2AtomKernel, pt2BondKernel, parameter1, parameter2 ); //cout << "END" << endl; return ( k ); } double Molecule::calculateSelfKernel( double (*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double (*pt2BondKernel)(Bond*, Bond*), int, int ), double (*pt2AtomKernel)( Atom*, Atom* ), double (*pt2BondKernel)( Bond*, Bond* ), int parameter1, int parameter2 ){ //selfKernel = this->moleculeKernel(this, pt2AtomKernel, pt2BondKernel, convergenceCondition); //cout << "NOW computing self kernel of molecule " << getName() << endl; selfKernel = pt2GraphKernel( this, this, pt2AtomKernel, pt2BondKernel, parameter1, parameter2 ); // moleculeKernel( this, anotherMolecule, pt2AtomKernel, pt2BondKernel, convergenceCondition ) ); selfKernelCalculated = true; return( selfKernel ); } /** move the hydrogen atoms to the hiddenAtoms container, so that further algorithms will run without hydrogens. Returns the number of hydrogens hidden */ int Molecule::hideHydrogens(){ int i = 0; vector::iterator atom; for( atom = beginAtom(); atom != endAtom(); atom++ ){ // hide hidrogen atoms if( (*atom)->getAN() == 1 ){ //cout << "Hiding H atom" << endl; hiddenAtoms.push_back( (*atom) ); //cout << "erasing atom" << endl; atoms.erase( atom ); //cout << "done" << endl; if( atom != beginAtom() ){ atom--; } i++; } //cout << "hiding bonds" << endl; // hide bonds to hydrogens (*atom)->hideHydrogenBonds(); //cout << "done" << endl; } return(i); } /** restores the hidden atoms (for example hydrogens) */ int Molecule::restoreHiddenAtoms( bool flagRestoreBonds ){ //cout << "Molecule::restoreHiddenAtoms: restoring hidden atoms" << endl; int i = 0; vector::iterator atom; for( atom = hiddenAtoms.begin(); atom != hiddenAtoms.end(); atom++ ){ //cout << " " << i << endl; // restore hydrogen atoms atoms.push_back( (*atom) ); //hiddenAtoms.erase(atom); i++; } hiddenAtoms.clear(); if( flagRestoreBonds == true ){ for( atom = atoms.begin(); atom != atoms.end(); atom++ ){ (*atom)->restoreHiddenBonds(); } } return(i); } int Molecule::restoreHiddenBonds(){ //cout << "Molecule::restoreHiddenAtoms: restoring hidden atoms" << endl; int i = 0; vector::iterator atom; for( atom = atoms.begin(); atom != atoms.end(); atom++ ){ i += (*atom)->restoreHiddenBonds(); } return(i); } /** This function reads the molecule description from a Mol file */ void Molecule::readMOL( string aFilename, bool genericAtomType ) throw( CError ) { // erase all eventually existing atoms in the molecule erase(); // open molfile ifstream inFile; inFile.open( aFilename.c_str(), ios::in ); if( !inFile.good() ){ CError e = CError(FILENOTFOUND, aFilename + " file not found"); e.describe(); throw(e); } #ifdef DEBUG cout << " reading readMDLHeaderBlock " << endl; #endif MoleculeUtils::readMDLHeaderBlock( *this, inFile, aFilename ); #ifdef DEBUG cout << " reading readMDLCtabBlock " << endl; #endif MoleculeUtils::readMDLCtabBlock( *this, inFile, genericAtomType ); inFile.close(); location = aFilename; compute(); } /** This function reads the molecule description from a Mol file */ void Molecule::readMOLOld( string aFilename ) throw( CError ) { stringstream out; out << "Molecule::readMOLOld is deprecated use Molecule::readMol instead"; CError e(NOTIMPLEMENTED, out.str() ); e.describe(); throw(e); // erase all eventually existing atoms in the molecule erase(); // open molfile ifstream inFile; inFile.open( aFilename.c_str(), ios::in ); if( !inFile.good() ){ CError e = CError(FILENOTFOUND, aFilename + " file not found"); e.describe(); throw(e); } int linesize = 256; char *line = new char[linesize]; string stringLine; int nbAtoms = 0; int nbBonds = 0; int nbAtomlist = 0; int nbStext = 0; int chiral = 0; int nbProperties = 0; string version = ""; int bondType = 0; Atom* anAtom = NULL; int lineNumber = 0; int j = 0; // for each line in file while( !inFile.eof() ){ lineNumber++; //read line inFile.getline( line, linesize-1,'\n' ); stringLine = line; if( lineNumber == 1 ){ // name setStringDescriptor( "comment", stringLine, "", "", true, true ); //vector< string > pathWords; //StringUtils::Split(aFilename, "/", pathWords); //char fileNameOnly[256] = ""; //strcpy( pathWords[0].c_str(), fileNameOnly ); //cout << "MMMM setting name to " << fileNameOnly << endl; //setName( fileNameOnly ); setName( aFilename ); }else if( lineNumber == 2 ){ // comment (skipping line 3) }else if( lineNumber == 4 ){ // counts line //cout << lineNumber << ": " << stringLine << endl; nbAtoms = atoi( stringLine.substr(0,3).c_str() ); nbBonds = atoi( stringLine.substr(3,3).c_str() ); nbAtomlist = atoi( stringLine.substr(6,3).c_str() ); nbStext = atoi( stringLine.substr(15,3).c_str() ); chiral = atoi( stringLine.substr(9,3).c_str() ); nbProperties = atoi( stringLine.substr(30,3).c_str() ); version = stringLine.substr(33,6); //cout << "found " << nbAtoms << " atoms" << " and " << nbBonds << " bonds "<< endl; }else if( lineNumber > 4 ){ // data if( nbAtoms > 0 && j < nbAtoms ){ // atoms // cout << "atom " << lineNumber << ": " << stringLine << endl; anAtom = addAtom( StringUtils::rmSpace( stringLine.substr(30,3) ) ); anAtom->setCoordinates( atof( stringLine.substr(0,10).c_str() ), atof( stringLine.substr(10,10).c_str() ), atof( stringLine.substr(20,10).c_str() ) ); j++; }else if( j >= nbAtoms && j < nbAtoms + nbBonds ) { // bonds //cout << "bond " << lineNumber << ": " << stringLine << endl; bondType = atoi( stringLine.substr(6,3).c_str() ); //bondTopology = atoi( stringLine.substr(15,3) ) linkAtoms( atoi( stringLine.substr(0,3).c_str() ) - 1, atoi( stringLine.substr(3,3).c_str() ) - 1, bondType ); j++; }else if( j >= nbAtoms + nbBonds && j < nbAtoms + nbBonds + nbAtomlist + nbStext ){ // atomlist block and stext block (SKIPPING) //cout << "skipping line " << lineNumber << " in file " << aFilename << ": " << stringLine << endl; j++; }else if( j >= nbAtoms + nbBonds + nbAtomlist + nbStext && j < nbAtoms + nbBonds + nbAtomlist + nbStext + nbProperties ){ // properties block (SKIPPING) //cout << "skipping line " << lineNumber << " in file " << aFilename << ": " << stringLine << endl; j++; }else{ //cout << "skipping line " << lineNumber << " in file " << aFilename << ": " << stringLine << endl; j++; } }else{ //cout << "skipping line " << lineNumber << " in file " << aFilename << ": " << stringLine << endl; } } delete[] line; } /** Erases all atoms in the current molecule. Returns the molecule in the state of construction by the default constructor. */ void Molecule::erase(){ #ifdef DEBUG cout << "Molecule::erase()" << endl; #endif #ifdef DEBUG cout << " deleting fastPQ" << endl; #endif fastPQ.clear(); #ifdef DEBUG cout << " deleting fastPQ DONE" << endl; cout << " deleting fastPS" << endl; #endif fastPS.clear(); #ifdef DEBUG cout << " deleting fastPS DONE" << endl; cout << " deleting fastPT" << endl; #endif map* >::iterator j; for( j = fastPT->begin(); j != fastPT->end(); j++ ){ delete (*j).second; // delete map } fastPT->clear(); #ifdef DEBUG cout << " deleted fastPT DONE" << endl; #endif for( j = fastPTSave->begin(); j != fastPTSave->end(); j++ ){ delete (*j).second; // delete map } fastPTSave->clear(); if( fastPTNext != NULL){ for( j = fastPTNext->begin(); j != fastPTNext->end(); j++ ){ delete (*j).second; // delete map } fastPTNext->clear(); } #ifdef DEBUG cout << " deleted fastPTSave DONE" << endl; #endif // delete all atoms #ifdef DEBUG cout << "DELETING " << atoms.size() << " ATOMS" << endl; #endif vector::iterator i; for( i = atoms.begin(); i != atoms.end(); i++ ){ #ifdef DEBUG cout << "deleting " << (*i)->toStringShort() << endl; #endif delete (*i); // delete atom } atoms.clear(); eraseHiddenAtoms(); // delete hidden atoms eraseRings(); selfKernelCalculated = false; #ifdef DEBUG cout << "Molecule::erase() done" << endl; #endif } void Molecule::eraseAdjacency(){ for(int i = 0 ; i < numAtoms() ; i++){ // NB : dimension(adjacency) = numAtoms() (*adjacency)[i].clear(); } delete adjacency; } void Molecule::eraseWalks(){ for(int i = 0 ; i < numAtoms() ; i++){ (*walks)[i].clear(); } delete walks; } void Molecule::eraseHiddenAtoms(){ vector::iterator i; // delete hidden atoms for( i = hiddenAtoms.begin(); i != hiddenAtoms.end(); i++ ){ #ifdef DEBUG cout << "deleting " << (*i)->toStringShort() << endl; #endif delete (*i); // delete atom } hiddenAtoms.clear(); } void Molecule::eraseRings(){ // erase rings; vector::iterator ri; for( ri = sssr.begin(); ri != sssr.end(); ri++ ){ delete (*ri); } sssr.clear(); } /** unset all bonds flags */ void Molecule::unsetBondFlags(){ vector::iterator ait; for( ait = beginAtom(); ait != endAtom(); ait++ ){ (*ait)->unsetBondFlags(); } } /** unset all bonds flags */ void Molecule::unsetBondFlagsOriginal(){ vector::iterator ait; for( ait = beginAtom(); ait != endAtom(); ait++ ){ (*ait)->unsetBondFlagsOriginal(); } } /** returns the number of bonds in the molecule */ int Molecule::numBonds(){ int res = 0; if( numAtoms() > 1 ){ vector::iterator ait; for( ait = beginAtom(); ait != endAtom(); ait++ ){ res = res + (*ait)->numBonds(); } return( res/2 ); }else{ return( 0 ); } } long Molecule::bondSum(){ long res = 0; if( numAtoms() > 1 ){ vector::iterator ait; for( ait = beginAtom(); ait != endAtom(); ait++ ){ res += (*ait)->bondSum(); } return( res ); }else{ return( 0 ); } } int Molecule::numHiddenBonds(){ int res = 0; if( numAtoms()>1 ){ vector::iterator ait; for( ait = atoms.begin(); ait != atoms.end(); ait++ ){ res = res + (*ait)->numHiddenBonds(); } for( ait = hiddenAtoms.begin(); ait != hiddenAtoms.end(); ait++ ){ res = res + (*ait)->numBonds(); } return( res/2 ); }else{ return( 0 ); } } /** write the molecule definition in a mdl mol file */ void Molecule::writeMOL( string aFilename ){ ofstream outFile; outFile.open( aFilename.c_str(), ios::out ); if( !outFile.good() ){ CError e = CError(COULDNOTOPENFILE, aFilename + " could not open file"); e.describe(); throw(e); } MoleculeUtils::writeMDLHeaderBlock( *this, outFile ); MoleculeUtils::writeMDLCtabBlock( *this, outFile ); outFile.close(); } /** writes a dot file representing the molecule. The molecule can be converted to graphioc representationj using Graphviz's dot program */ void Molecule::writeDOT( string aFilename, bool perretLabels ){ ofstream outFile; outFile.open( aFilename.c_str(), ios::out ); if( !outFile.good() ){ CError e = CError(COULDNOTOPENFILE, aFilename + " could not open file"); e.describe(); throw(e); } MoleculeUtils::writeDOTGraph( *this, outFile, perretLabels ); outFile.close(); } /** write a MDL structure data (SD) file with a single molecule */ void Molecule::writeSD( string aFilename ){ ofstream outFile; outFile.open( aFilename.c_str(), ios::out ); if( !outFile.good() ){ CError e = CError(COULDNOTOPENFILE, aFilename + " could not open file"); e.describe(); throw(e); } MoleculeUtils::writeMDLHeaderBlock( *this, outFile ); MoleculeUtils::writeMDLCtabBlock( *this, outFile ); MoleculeUtils::writeMDLNSDBlock( *this, outFile ); outFile.close(); } int Molecule::markFragments(){ // add componentIndex intDescriptor to all atoms //cout << "Molecule::markFragments: adding descriptor COMPONENTINDEX" << endl; Descriptor< int >* aDescriptor; vector< Atom* >::iterator i; for( i = atoms.begin(); i != atoms.end(); i++ ){ aDescriptor = (*i)->setIntDescriptor( COMPONENTINDEX, -1, "NA", "temporary variable", true, true ); aDescriptor->setEmpty(); (*i)->unsetVisited(); } // search for components using Depth First Search algorithm on graphs // (marking componentIndex) and save the number of atoms in each component in componentSizes //cout << "Molecule::markFragments: search for components with DFS" << endl; bool flagAllMarked = false; int componentIndex = 0; // as long as not all atoms where marked: while( flagAllMarked == false ){ Atom* startAtom = NULL; // find start atom flagAllMarked = true; vector< Atom* >::iterator i; for( i = atoms.begin(); i != atoms.end(); i++ ){ if( (*i)->getIntDescriptor( COMPONENTINDEX )->isEmpty() ){ startAtom = (*i); //cout << " --> atom " << startAtom->toStringShort() << " was not marked by DFS" << endl; flagAllMarked = false; } } if( flagAllMarked == false){ int componentSize = 0; // contains the number of atoms in the component at the end of DFS componentSize = DFS( (Atom*) startAtom, COMPONENTINDEX, componentIndex ); // start ,desriptor for marking, number to mark //cout << "Molecule::markFragments: componentSize :" << componentSize << endl; //cout << componentSizes[0] << endl; componentSizes[ componentIndex ] = componentSize; componentIndex ++; } } //cout << "Molecule::markFragments: done" << endl; return( componentIndex ); } void Molecule::unmarkFragments(){ // remove componentIndex intDescriptor of all visible and hidden atoms vector< Atom* >::iterator i; for( i = atoms.begin(); i != atoms.end(); i++ ){ (*i)->deleteDescriptor( COMPONENTINDEX, false ); } for( i = hiddenAtoms.begin(); i != hiddenAtoms.end(); i++ ){ (*i)->deleteDescriptor( COMPONENTINDEX, false ); } componentSizes.clear(); } void Molecule::hideAllFragmentsBut( int aFragmentNumber ){ //cout << "Molecule::hideAllFragmentsBut fragment " << aFragmentNumber << endl; //cout << " this molecule has " << componentSizes.size() << " fragments" << endl; map< int, int >::iterator j; for( j = componentSizes.begin(); j != componentSizes.end(); j++ ){ if( (*j).first != aFragmentNumber ){ hideAtomsByIntDescriptor( COMPONENTINDEX, (*j).first, false ); //componentSizes.erase( j ); //j--; } } } void Molecule::writeFragments( ofstream* outStream ){ //cout << "Molecule::writeFragments" << endl; map< int, int >::iterator j; int i = 1; for( j = componentSizes.begin(); j != componentSizes.end(); j++ ){ //describe(); hideAllFragmentsBut( (*j).first ); string molName = getName(); stringstream newName; newName << molName << "." << i; setName( newName.str() ); MoleculeUtils::writeMDLHeaderBlock( *this, *outStream ); MoleculeUtils::writeMDLCtabBlock( *this, *outStream ); MoleculeUtils::writeMDLNSDBlock( *this, *outStream ); setName( molName ); restoreHiddenAtoms( false ); i++; } } /** hide all but the biggest (in atom numbers) connected atoms in the molecule. Returns the number of components removed WARNING: if two compounds have identical maximum size. it keep the compounds with the highest number of Carbons + Nitrogens WARNING: deletes all hidden atoms */ int Molecule::hideSalts( stringstream* out ){ eraseHiddenAtoms(); // delete hidden atoms. markFragments(); // mark fragments // find the two biggest components map< int, int >::iterator j; int biggestComponent = 0; int secondBiggestComponent = 0; int biggestComponentValue = 0; int secondBiggestComponentValue = 0; for( j = componentSizes.begin(); j != componentSizes.end(); j++ ){ //cout << (*j).first << ", " << (*j).second << endl; if( (*j).second >= biggestComponentValue ){ secondBiggestComponent = biggestComponent; secondBiggestComponentValue = biggestComponentValue; biggestComponent = (*j).first; biggestComponentValue = (*j).second; //cout << " this one"; } } // if the two biggest compounds have an identical number of atoms // keep the one with the highest number of carbons + nitrogens (hydrogens might be missing) if( biggestComponentValue == secondBiggestComponentValue ){ if( getNumCarbonsOfComponent( COMPONENTINDEX, biggestComponent ) + getNumNitrogensOfComponent( COMPONENTINDEX, biggestComponent ) > getNumCarbonsOfComponent( COMPONENTINDEX, secondBiggestComponent ) + getNumNitrogensOfComponent( COMPONENTINDEX, secondBiggestComponent ) ){ }else{ int toto = biggestComponent; int totoValue = biggestComponentValue; biggestComponent = secondBiggestComponent; biggestComponentValue = secondBiggestComponentValue; secondBiggestComponent = toto; secondBiggestComponentValue = totoValue; } } // hide all other components and associated bonds hideAllFragmentsBut( biggestComponent ); if( componentSizes.size() - 1 > 0 ){ if( biggestComponentValue - secondBiggestComponentValue > 7 || secondBiggestComponentValue == 1 ){ *out << getName() << ";" << componentSizes.size() - 1 << ";" << biggestComponentValue << ";" << secondBiggestComponentValue << ";" << biggestComponentValue - secondBiggestComponentValue << ";" << numHiddenAtoms() << ";" << endl; setStringDescriptor( "saltFilterWarning", "FILTERED", "NA", "", true, true ); }else{ *out << getName() << ";" << componentSizes.size() - 1 << ";" << biggestComponentValue << ";" << secondBiggestComponentValue << ";" << biggestComponentValue - secondBiggestComponentValue << ";" << numHiddenAtoms() << ";WARNING" << endl; setStringDescriptor( "saltFilterWarning", "WARNING", "NA", "", true, true ); } }else{ setStringDescriptor( "saltFilterWarning", "UNFILTERED", "NA", "", true, true ); } // document actions setIntDescriptor( "removedCompounds", componentSizes.size() - 1, "NA", "", true, true ); unmarkFragments(); return( componentSizes.size() - 1 ); } /** This function performs a DFS search on the graph starting at startAtom, marking atoms with markValue in descriptor intDescriptorName, and returning the number of nodes in the connected component. */ int Molecule::DFS( Atom* startAtom, string intDescriptorName, int markValue ){ int markedAtoms = 0; // mark node as visited startAtom->setVisited(); //cout << " now at " << startAtom->toStringShort(); // find next unvisited Atom int end = 0; Atom* nextAtom; while( end == 0 ){ try{ nextAtom = startAtom->nextUnvisitedAtom(); markedAtoms = markedAtoms + DFS( nextAtom, intDescriptorName, markValue ); //nextAtom = startAtom->nextUnvisitedAtom(); }catch( CError ){ end = 1; break; } } //cout << " all atoms of this branch visited " << endl; startAtom->setIntDescriptor(intDescriptorName, markValue, "", "", true, true); return( 1 + markedAtoms ); } /** hide all atoms (and associated edges) with intDescriptor aDescriptorName == aValue Returns the number of hidden atoms */ int Molecule::hideAtomsByIntDescriptor( string aDescriptorName, int aValue, bool flagRefreshBonds ){ //cout << "Molecule::hideAtomsByIntDescriptor: hiding all atoms with " << aDescriptorName << " = " << aValue << endl; vector< Atom* >::iterator i; int c = 0; for( i = atoms.begin(); i != atoms.end(); i++ ){ if( (*i)->getIntDescriptor( aDescriptorName )->getValue() == aValue ){ //cout << " found atom " << (*i)->toStringShort() << endl; hideAtom( i ); //hiddenAtoms.push_back( *i ); //atoms.erase( i ); c++; i--; } } if( flagRefreshBonds == true ){ //cout << "REFRESHING BONDS" << endl; refreshBonds(); } return( c ); } /** returns the number of carbons in a connected component graph defined with intDescriptor aDescriptorName having value aValue */ int Molecule::getNumCarbonsOfComponent( string aDescriptorName, int aValue ){ vector< Atom* >::iterator i; int c = 0; for( i = atoms.begin(); i != atoms.end(); i++ ){ if( (*i)->getIntDescriptor( aDescriptorName )->getValue() == aValue ){ if( (*i)->getElementSymbol() == "C" ){ c++; } } } return( c ); } /** returns the number of nitrogens in a connected component graph defined with intDescriptor aDescriptorName having value aValue */ int Molecule::getNumNitrogensOfComponent( string aDescriptorName, int aValue ){ vector< Atom* >::iterator i; int c = 0; for( i = atoms.begin(); i != atoms.end(); i++ ){ if( (*i)->getIntDescriptor( aDescriptorName )->getValue() == aValue ){ if( (*i)->getElementSymbol() == "N" ){ c++; } } } return( c ); } /** hides an atom (but not the edges to this node). call refreshBonds after hiding all atoms. */ void Molecule::hideAtom( vector< Atom* >::iterator anAtomI ){ // cout << "Molecule::hideAtom " << (*anAtomI)->toStringShort() << endl; hiddenAtoms.push_back( *anAtomI ); atoms.erase( anAtomI ); } void Molecule::hideAtomAndToFromBonds( vector< Atom* >::iterator anAtomI ){ cout << "Molecule::hideAtomAndToFromBonds (iterator), hiding " << (*anAtomI)->toStringShort() << endl; cout << "1" << endl; (*anAtomI)->hideAllToFromBonds(); cout << "2" << endl; hiddenAtoms.push_back( *anAtomI ); cout << "3" << endl; //atoms.erase( anAtomI ); cout << "4" << endl; } void Molecule::hideAtomAndToFromBonds( Atom* anAtom ){ // cout << "Molecule::hideAtom " << (*anAtomI)->toStringShort() << endl; anAtom->hideAllToFromBonds(); hiddenAtoms.push_back( anAtom ); eraseAtom( anAtom ); } void Molecule::eraseAtom( Atom* anAtom ) throw( CError ){ vector::iterator ai; for( ai = atoms.begin(); ai != atoms.end(); ai++){ if( (*ai) == anAtom ){ atoms.erase( ai ); return; } } stringstream out; out << "Atom " << anAtom->toString() << " does not exist in molecule " << toString(); CError e( ATOMNOTFOUND, out.str() ); e.describe(); throw(e); } /** hides all bonds to and from hidden atoms. Call this function after hiding all atoms in a molecule. Returns the number of edges hidden */ int Molecule::refreshBonds(){ // iterate over all bonds of all atoms. If the destination atom is hidden then remove this bond int c = 0; vector< Atom* >::iterator n; map< Atom*, Bond* >::iterator e; for( n = atoms.begin(); n != atoms.end(); n++ ){ for( e = (*n)->beginBond(); e != (*n)->endBond(); e++){ if( isHiddenAtom( (*e).second->getTarget() ) && ! isHiddenAtom( (*e).second->getSource() ) ){ // hide the edges only from non hidden atoms (so in case of restore the edges can be restored (*e).first->hideBond( e ); e--; c++; } } } return( c ); } /** returns true if an atom is hidden */ bool Molecule::isHiddenAtom( Atom* anAtom ){ vector< Atom* >::iterator n; for( n = atoms.begin(); n != atoms.end(); n++ ){ if( anAtom == *n ){ return( false ); } } return( true ); } /** resets the calculations of the morgan indices. CALL THIS FUNCTION EACH TIME THE MOLECULE IS MODIFIED! */ void Molecule::resetMorganIndex(){ maxMorganIteration = -1; vector< Atom* >::iterator n; for( n = atoms.begin(); n != atoms.end(); n++ ){ (*n)->resetMorganIndex(); } } /** returns the number of distinct morgan indices (of order anOrder) present in the molecule */ int Molecule::getNumberOfDistinctMorganIndices( int anOrder ){ map< int, int > morganIs; vector< Atom* >::iterator n; for( n = atoms.begin(); n != atoms.end(); n++ ){ morganIs[ (*n)->getMorganIndex( anOrder ) ] += 1; } return( morganIs.size() ); } /** returns the iteration of the Morgan index at which the different connectivity values have reached a maximum */ int Molecule::getMaxMorganIteration(){ if( maxMorganIteration == -1 ){ int i = 1; bool found = false; int lastN = 0; int N = 0; while( found == false ){ N = getNumberOfDistinctMorganIndices( i ); if( N == lastN ){ found = true; return( i-1 ); }else{ lastN = N; } i++; } }else{ return( maxMorganIteration ); } } /** sets the morganLabel of atoms to the concatenation of the atomic symbol and the Morgan index of anOrder iteration */ void Molecule::setMorganLabels( int anOrder ){ vector< Atom* >::iterator n; for( n = atoms.begin(); n != atoms.end(); n++ ){ (*n)->setMorganLabel( anOrder ); } } void Molecule::setPerretLabels(){ #ifdef DEBUGPERRETLABEL cout << "Molecule::setPerretLabels for " << toStringShort() << endl; cout << " molecule has " << sssr.size() << " rings" << endl; #endif vector< Atom* >::iterator n; for( n = atoms.begin(); n != atoms.end(); n++ ){ (*n)->setPerretLabel(); } } /** sets the uniqueMorganIndex of each atom to the Morgan index having the maximum of different connectivity values for the molecule */ int Molecule::setUniqueMorganIndices(){ int anOrder = getMaxMorganIteration(); vector< Atom* >::iterator n; for( n = atoms.begin(); n != atoms.end(); n++ ){ (*n)->setUniqueMorganIndex( anOrder ); } return( anOrder ); } double Molecule::getSelfKernel() throw( CError ){ if( selfKernelCalculated == false ){ // kernel was not computed, emit error stringstream out; out << "Molecule::getSelfKernel::self kernel was not calculated in " << toString(); CError e = CError(NOTCALCULATED, out.str() ); e.describe(); throw(e); } return( selfKernel ); } /** returns the sum of transition probabilities. Should be used with fused graph. */ double Molecule::sumPT(){ // add the PT of each bond of each atom vector::iterator ai; map::iterator bi; double total = 0; for( ai = beginAtom(); ai != endAtom(); ai++ ){ // source atom is (*ai) for( bi = (*ai)->beginBond(); bi != (*ai)->endBond(); bi++){ // target atom is (*bi).first total += (*bi).second->getKashimaPT(); } } return( total ); } /** function used to compute the graph kernel using the fused graph / sum of the powers of the transition matrix approach. See MoleculeUtils::powerKernel */ double Molecule::sumProbabilities(){ // add the PT of each bond of each atom vector::iterator ai; map::iterator bi; double total = 0; for( ai = beginAtom(); ai != endAtom(); ai++ ){ // source atom is (*ai) for( bi = (*ai)->beginBond(); bi != (*ai)->endBond(); bi++){ // target atom is (*bi).first total += ( (*ai)->getKashimaPS() * (*bi).first->getKashimaPQ() * (*bi).second->getKashimaPT() ); } } return( total ); } double Molecule::sumProbabilitiesFast(){ // add the PT of each bond of each atom //cout << "Molecule::sumProbabilitiesFast" << endl; map* >::iterator ptli; map< Atom*, double>::iterator pti; double total = 0; //int i = 0; //int j = 0; for( ptli = fastPT->begin(); ptli != fastPT->end(); ptli++ ){ //cout << i << endl; //i++; // source atom is (*ai) //j = 0; for( pti = (*ptli).second->begin(); pti != (*ptli).second->end(); pti++ ){ //cout << " " << j << endl; //cout << " ps: " << fastPS[ (*ptli).first ] << endl; //cout << " pq: " << fastPQ[ (*pti).first ] << endl; //cout << " pt: " << (*pti).second << endl; //j++; // target atom is (*bi).first total += ( fastPS[ (*ptli).first ] * fastPQ[ (*pti).first ] * (*pti).second ); //cout << " total: " << total << endl; } } return( total ); } double Molecule::sumPQPS(){ // add the PT of each bond of each atom vector::iterator ai; map::iterator bi; double total = 0; for( ai = beginAtom(); ai != endAtom(); ai++ ){ // source atom is (*ai) total += ( (*ai)->getKashimaPS() * (*ai)->getKashimaPQ() ); } return( total ); } double Molecule::sumPQPSFast(){ // add the PT of each bond of each atom vector::iterator ai; double total = 0; for( ai = beginAtom(); ai != endAtom(); ai++ ){ // source atom is (*ai) total += ( fastPQ[ *ai ] * fastPS[ *ai ] ); } return( total ); } void Molecule::raisePowerFast(){ //cout << "Molecule::raisePowerFast" << endl; vector::iterator i; vector::iterator j; vector::iterator ci; double aPt = 0; map* sources; map* savedSources; double a = 0; double b = 0; fastPTNext = new map* >; for( i = beginAtom(); i != endAtom(); i++ ){ sources = (*fastPT)[*i]; if( sources != NULL ){ //MoleculeUtils::describeMap( sources ); //cout << "Calculating line of " << (*i)->toStringShort() << " " << sources->size() << " bonds " << endl; for( j = beginAtom(); j != endAtom(); j++ ){ aPt = 0; for( ci = beginAtom(); ci != endAtom(); ci++){ savedSources = (*fastPTSave)[*ci]; if( savedSources != NULL ){ a = 0; a = (*sources)[*ci]; if( a != 0 ){ b = 0; b = (*savedSources)[*j]; aPt += a*b; //cout << (*i)->toStringShort() << " - " << (*ci)->toStringShort() << " * " << (*ci)->toStringShort() << " - " << (*j)->toStringShort() << ": " << a << " * " << b << " = " << a*b << " " << aPt <* aMap = (*fastPTNext)[*i]; if( aMap == NULL ){ aMap = new map; (*aMap)[*j] = aPt; (*fastPTNext)[*i] = aMap; }else{ (*aMap)[*j] = aPt; } } } } } } // now the sparse matrix fastPT sould contain the data stored in fastPTNext // erase the content of fastPT map* >::iterator k; for( k = fastPT->begin(); k != fastPT->end(); k++ ){ delete (*k).second; // delete bonds } fastPT->clear(); delete fastPT; // fastPT is the newly computed version fastPT = fastPTNext; fastPTNext = NULL; } /** delete all bonds in all atoms */ void Molecule::deleteBonds(){ vector::iterator ai; for( ai = beginAtom(); ai != endAtom(); ai++ ){ (*ai)->deleteBonds(); } } float Molecule::getMW( bool silentError ) throw( CError ){ float mw = 0; vector::iterator ai; for( ai = beginAtom(); ai != endAtom(); ai++ ){ try{ mw += (*ai)->getFloatDescriptor( AM )->getValue( true ); }catch( CError e ){ if( e.getType() == MISSINGDESCRIPTOR || e.getType() == ERRORNA ){ if( silentError == false) { cerr << "Cannot compute molecular weight in " << getName() << endl; //cerr << "Molecule::getMW() : WARNING atom " << (*ai)->toStringShort() << " has no descriptor for atomic mass" << endl; throw( e ); } } } } return( mw ); } /*! \fn Molecule::deleteHiddenAtoms() */ void Molecule::deleteHiddenAtoms() { // delete all hiddenbonds first vector::iterator ai; for( ai = atoms.begin(); ai != atoms.end(); ai++ ){ (*ai)->deleteHiddenBonds(); } for( ai = hiddenAtoms.begin(); ai != hiddenAtoms.end(); ai++ ){ delete (*ai); } hiddenAtoms.clear(); } // fill in vector with new kind atoms void Molecule::atomsLabelsListing(vector *atomLabels ){ vector::iterator anAtom; // for each atom : for( anAtom = atoms.begin(); anAtom != atoms.end(); anAtom++ ){ bool present = false; for(int i = 0 ; i < atomLabels->size() ; i++){ if( (*anAtom)->getMorganLabel( true ) == (*atomLabels)[i] ){ present = true; } } if( present == false ){ (*atomLabels).push_back( (*anAtom)->getMorganLabel(true) ); } } } // fill in vector with new kind atoms void Molecule::atomsSymbolsListing(vector *atomSymbols ){ vector::iterator anAtom; // for each atom : for( anAtom = atoms.begin(); anAtom != atoms.end(); anAtom++ ){ bool present = false; for(int i = 0 ; i < atomSymbols->size() ; i++){ if( (*anAtom)->getSymbol() == (*atomSymbols)[i] ){ present = true; } } if( present == false ){ (*atomSymbols).push_back( (*anAtom)->getSymbol() ); } } } // fill in vector with new kind of bonds void Molecule::bondsListing(vector *bondTypes ){ vector::iterator anAtom; // for each atom for( anAtom = atoms.begin(); anAtom != atoms.end(); anAtom++ ) { map::iterator aBond; // for each bond of the atom : for( aBond = (*anAtom)->beginBond() ; aBond != (*anAtom)->endBond() ; aBond++ ){ bool present = false; for( int i = 0 ; i < bondTypes->size() ; i++ ){ if ( aBond->second->getLabel() == (*bondTypes)[i] ) present = true; } if( present == false ) (*bondTypes).push_back( aBond->second->getLabel() ); } } } // add a value to the self kernel void Molecule::addToSelfKernel(double value){ selfKernel += value; selfKernelCalculated = true; } // substract a value to the self kernel void Molecule::substractToSelfKernel(double value){ selfKernel -= value; selfKernelCalculated = true; } /*! \fn Molecule::numAtomsNonCSkeleton() */ int Molecule::numAtomsNonCSkeleton() { int n = 0; vector::iterator ai; for( ai = atoms.begin(); ai != atoms.end(); ai++ ){ if( !(*ai)->isCSkeleton() ){ n++; } } return( n ); } void Molecule::binClassifyFromDescriptor( string descriptorName, float value, bool smallerOrEqual){ float dvalue = 0; bool biggerthan = false; bool smallerthan = false; try{ string svalue = getStringDescriptor( descriptorName, true )->getValue(); //check if the first character of the descriptor value is > or < which is enough to set the threshold if( svalue.substr(0,1) == ">" ){ biggerthan = true; dvalue = atof( svalue.substr(1,svalue.length()-1).c_str() ); }else if( svalue.substr(0,1) == "<" ){ smallerthan = true; dvalue = atof( svalue.substr(1,svalue.length()-1).c_str() ); }else{ dvalue = atof( svalue.c_str() ); } }catch( CError e ){ try{ dvalue = getFloatDescriptor( descriptorName, true )->getValue(); }catch( CError e ){ try{ dvalue = getIntDescriptor( descriptorName, true )->getValue(); }catch( CError e ){ return; } } } cout << "comparing "; if( biggerthan == true){ cout << ">"; } if( smallerthan == true){ cout << "<"; } cout << dvalue << " with " << value << " setting activity to "; if( smallerOrEqual == true ){ if( dvalue <= value ){ if( biggerthan == false ){ cout << "true" << endl; setActivity( 1 ); }else{ if( dvalue == value ){ cout << "false" << endl; setActivity( 0 ); }else{ cout << "nothing" << endl; unsetActivity(); } } }else{ cout << "false" << endl; setActivity( 0 ); } }else{ if(dvalue >= value ){ if( smallerthan == false ){ cout << "true" << endl; setActivity( 1 ); }else{ if( dvalue == value ){ cout << "false" << endl; setActivity( 0 ); }else{ cout << "nothing" << endl; unsetActivity(); } } }else{ cout << "false" << endl; setActivity( 0 ); } } return; } void Molecule::setSortDescriptor( string aName, int aType ){ if( aType == INTEGER ){ if( !hasIntDescriptor( aName ) ){ stringstream out; out << "Molecule has no INT descriptor called " << aName ; CError e( MISSINGDESCRIPTOR, out.str() ); e.describe(); throw(e); } }else if( aType == FLOAT ){ if( !hasFloatDescriptor( aName ) ){ stringstream out; out << "Molecule has no FLOAT descriptor called " << aName ; CError e( MISSINGDESCRIPTOR, out.str() ); e.describe(); throw(e); } }else{ if( !hasFloatDescriptor( aName ) ){ stringstream out; out << "Molecule has no STRING descriptor called " << aName ; CError e( MISSINGDESCRIPTOR, out.str() ); e.describe(); throw(e); } } sortDescriptorType = aType; sortDescriptorName = aName; } int Molecule::detectSSSR(){ if( hasSSSRDetected() ){ return( sssr.size() ); }else{ cout << toStringShort() << "::detectSSSR()" << endl; #ifdef DEBUGSSSR cout << "a3h detecting SSSR" << endl; #endif deleteHiddenAtoms(); // fullset is atoms vector trimset; // store broken bonds Bond* brokenBond = NULL; #ifdef DEBUGSSSR cout << "a3h 1" << endl; #endif // add nodes of degree 0 to trimset vector substractSet; vector::iterator ai; for ( ai = beginAtom(); ai != endAtom(); ai++ ){ substractSet.push_back( (*ai) ); } long finishedAt = atoms.size(); while( trimset.size() != finishedAt ){ // while trimset != fullset #ifdef DEBUGSSSR cout << "a3h trimming nodes with degree 0" << endl; #endif int i = 0; //for ( ai = beginAtom(); ai != endAtom(); ai++ ){ for ( ai = substractSet.begin(); ai != substractSet.end(); ai++ ){ if( (*ai)->degree() == 0 ){ #ifdef DEBUGSSSR cout << "a3hb " << (*ai)->toStringShort() << " has degree 0, adding to trimset" << endl; #endif trimset.push_back( (*ai) ); //hideAtomAndToFromBonds( (*ai) ); i++; } } #ifdef DEBUGSSSR cout << "a3h trimmed " << i << " nodes " << endl; #endif // add nodes of degree N2 in fullset but not in trimset to N2 vector nodesN2; vector nodesNLarge; MoleculeUtils::substractSet( &atoms, &trimset, &substractSet ); #ifdef DEBUGSSSR cout << "a3hb atoms is " << atoms.size() << " nodes " << endl; cout << "a3hb trimset is " << trimset.size() << " nodes " << endl; for ( ai = trimset.begin(); ai != trimset.end(); ai++ ){ cout << "a3hb " << (*ai)->toStringShort() << endl; } cout << "a3hb substractSet is " << substractSet.size() << " nodes " << endl; for ( ai = substractSet.begin(); ai != substractSet.end(); ai++ ){ cout << "a3hb " << (*ai)->toStringShort() << endl; } #endif if( substractSet.size() == 0 ){ #ifdef DEBUGSSSR cout << "a3hb BREAKING" << endl; #endif break; } int minDegree = 10000; for ( ai = substractSet.begin(); ai != substractSet.end(); ai++ ){ int degree = (*ai)->degree(); if( degree == 2 ){ nodesN2.push_back((*ai)); }else if( degree > 2 ){ nodesNLarge.push_back((*ai)); } if( degree < minDegree ){ minDegree = degree; } } #ifdef DEBUGSSSR cout << "a3h found " << nodesN2.size() << " N2 nodes, mindegree is " << minDegree << endl; #endif // find a node init in fullset but not in trimset having minimum degree Atom* init; for ( ai = substractSet.begin(); ai != substractSet.end(); ai++ ){ if( (*ai)->degree() == minDegree ){ init = (*ai); break; } } #ifdef DEBUGSSSR cout << "a3h initialising search with " << init->toStringShort() << endl; #endif if( minDegree == 0 ){ setIntDescriptor( "sssrSize", sssr.size(), "NA", "", true, true ); return( sssr.size() ); }else if( minDegree == 1 ){ /// if minimum degree == 1 trim(init) and add to trimset #ifdef DEBUGSSSR cout << "trimming atom " << init->toStringShort() << endl; #endif init->hideAllToFromBonds(); //trimset.push_back( init ); // not needed, will be added as node of degree 0. }else if( minDegree == 2){ // elsif minimum degree == 2 //vector::iterator ci; //int gg = 0; //for( ci = nodesNLarge.begin(); ci != nodesNLarge.end(); ci++){ //(*ci)->hideAllToFromBonds(); //#ifdef DEBUGSSSR // cout << "rrrty hiding all to from bonds in atom " << (*ci)->toStringShort() << endl; // cout << gg << endl; // gg++; //#endif vector dBonds; bool dring = false; do{ for ( ai = nodesN2.begin(); ai != nodesN2.end(); ai++ ){ //if( (*ai) != (*ci) ){ // for each node i in nodesN2 #ifdef DEBUGSSSR cout << "a3hb examining N2 atom " << (*ai)-> toStringShort() << endl; #endif if( (*ai)->getBFSVectorSize() == 0 ){ #ifdef DEBUGSSSR cout << "a3h starting getRingBFS" << endl; #endif Ring* ring; vector* toVisit = new vector; vector* toVisitBond = new vector; try{ ring = (*ai)->getRingBFS( toVisit, toVisitBond ); // ring = getRing(i) delete toVisit; delete toVisitBond; // no error, so this atom is part of a ring // if ring.size > 0 then // check sssr for a duplicate, if no duplicate add ring to sssr #ifdef DEBUGSSSR cout << "a3h -> found ring of size " << ring->size() << endl; #endif if( sssr.size() == 0 ){ ring->setID( 1 ); sssr.push_back( ring ); //rememberNodes.push_back( (*ai) ); #ifdef DEBUGSSSR cout << "a3hc describing bonds in new ring: (" << ring->getBonds()->size() << ")" << endl; vector::iterator bb; for( bb = ring->getBonds()->begin(); bb != ring->getBonds()->end(); bb++ ){ cout << "a3hc > " << (*bb)->toStringShort() << endl; } cout << "a3hb found " << ring->toString() << ", added to sssr set " << endl; #endif }else{ if( !hasRing( ring, true ) ){ ring->setID( sssr.size() + 1 ); sssr.push_back( ring ); //rememberNodes.push_back( (*ai) ); #ifdef DEBUGSSSR cout << "a3hc describing bonds in new ring: " << endl; vector::iterator bb; for( bb = ring->getBonds()->begin(); bb != ring->getBonds()->end(); bb++ ){ cout << " > " << (*bb)->toString() << endl; } cout << "a3hb found " << ring->toString() << ", added to sssr set " << endl; #endif }else{ #ifdef DEBUGSSSR cout << "a3hb found " << ring->toString() << ", but this ring already exists in sssr set" << endl; #endif delete ring; } } }catch( CError e ){ // This N2 atom is part of no ring delete toVisit; delete toVisitBond; if( e.getType() == ATOMNOTINRING ){ // this atom is not member of a ring #ifdef DEBUGSSSR cout << "a3h getRing returned no ring" << endl; #endif }else{ throw( e ); } } // resetBFSSearch vector::iterator ai; for ( ai = beginAtom(); ai != endAtom(); ai++ ){ (*ai)->resetBFSVector(); } }else{ #ifdef DEBUGSSSR cout << "a3h this node was already visited, skipping" << endl; #endif }//} } if( dring == false ){ //if( dBonds.size() == 0 && dring == false){ #ifdef DEBUGSSSR cout << "DBOND: first visit " << dring << endl; #endif // just found the easy rings, search for the difficult now // find one bond for each N2 chain and store it in dBonds // for simplicity we will hide one bond of all N2 nodes in all easy rings // for all rings vector::iterator dri; for( dri = sssr.begin(); dri != sssr.end(); dri++ ){ // for all n2 atoms vector* aRingBonds = (*dri)->getBonds(); vector::iterator dbi; for( dbi = aRingBonds->begin(); dbi != aRingBonds->end(); dbi++ ){ if( (*dbi)->getSource()->numBonds() == 2 || (*dbi)->getTarget()->numBonds() == 2 ){ if( !((*dbi)->getSource()->numBonds() == 2 && (*dbi)->getTarget()->numBonds() == 2 )){ if( !MoleculeUtils::atomVectorHas( &trimset, (*dbi)->getSource() ) && !MoleculeUtils::atomVectorHas( &trimset, (*dbi)->getTarget() ) ){ dBonds.push_back(*dbi); } } } } } #ifdef DEBUGSSSR cout << "DBOND: dBonds now has size " << dBonds.size() << endl; #endif if( dBonds.size() != 0 ){ #ifdef DEBUGSSSR cout << "DBOND: hiding bond " << (*dBonds.begin())->toStringShort() << endl; #endif // hide the first dBond and repeat the ring search (*dBonds.begin())->hideToFrom(); } dring = true; }else{ // restore the hidden dBond #ifdef DEBUGSSSR cout << "DBOND: restoring " << (*dBonds.begin())->toStringShort() << endl; #endif (*dBonds.begin())->restoreToFrom(); dBonds.erase( dBonds.begin() ); // hide the next dBond and repeat the ring search (*dBonds.begin())->hideToFrom(); #ifdef DEBUGSSSR cout << "DBOND: dBonds now has size " << dBonds.size() << endl; #endif } }while( dBonds.size() > 0 ); #ifdef DEBUGSSSR cout << "DBOND: FINISHED" << endl; #endif //if( rememberNodes.size() == 0 ){ // rememberNodes.push_back( (*nodesN2.begin()) ); //} /*vector::iterator rai; for( rai = rememberNodes.begin(); rai != rememberNodes.end(); rai++ ){ #ifdef DEBUGSSSR cout << "hidingToFromBonds in atom " << (*rai)->toStringShort(); #endif (*rai)->hideAllToFromBonds(); }*/ /* // resetBFSSearch for ( ai = beginAtom(); ai != endAtom(); ai++ ){ (*ai)->resetBFSVector(); } #ifdef DEBUGSSSR (*ci)->describe(); cout << "rrrty: restoring broken bond " << (*ci)->toStringShort() << endl; #endif (*ci)->restoreHiddenBonds(); #ifdef DEBUGSSSR cout << "rrrty: done" << endl; #endif*/ //} if( brokenBond != NULL ){ #ifdef DEBUGSSSR cout << "rrrtz: restoring broken bond " << brokenBond->toStringShort() << " "; #endif brokenBond->getSource()->restoreHiddenBond( brokenBond->getTarget() ); cout << "2" << endl; brokenBond->getTarget()->restoreHiddenBond( brokenBond->getSource() ); brokenBond = NULL; } // all N2 atoms were tested for their ring membership, now break // one bond for each N2 chain. // for each chain of N2 nodes isolate one N2 node and break one bond #ifdef DEBUGSSSR cout << "a3h now checking borders " << endl; #endif for ( ai = nodesN2.begin(); ai != nodesN2.end(); ai++ ){ #ifdef DEBUGSSSR cout << "a3h removing " << (*ai)->toStringShort() << endl; #endif hideAtomAndToFromBonds( (*ai) ); //reduceDegreeToOne( (*ai) ); } }else if( minDegree > 2 ){ // elsif minimum degree == 3 vector* toVisit = new vector; vector* toVisitBond = new vector; try{ Ring* ring = (*ai)->getRingBFS( toVisit, toVisitBond ); // ring = getRing(i) delete toVisit; delete toVisitBond; if( !hasRing( ring, true ) ){ ring->setID( sssr.size() + 1 ); sssr.push_back( ring ); } // resetBFSSearch vector::iterator ai; for ( ai = beginAtom(); ai != endAtom(); ai++ ){ (*ai)->resetBFSVector(); } //checkedges( init, ringSet) trimset.push_back( init ); brokenBond = checkEdges( init ); }catch( CError e ){ delete toVisit; delete toVisitBond; // resetBFSSearch vector::iterator ai; for ( ai = beginAtom(); ai != endAtom(); ai++ ){ (*ai)->resetBFSVector(); } if( e.getType() == ATOMNOTINRING ){ // this atom is not member of a ring #ifdef DEBUGSSSR cout << "a3h getRing returned no ring" << endl; #endif }else{ throw( e ); } } } } #ifdef DEBUGSSSR cout << "a3h restoring hidden atoms and bonds " << endl; #endif restoreHiddenAtoms( true ); // restore hidden atoms and bonds // add the ring information to member atoms and member bonds vector::iterator ri; for( ri = sssr.begin(); ri != sssr.end(); ri++){ vector::iterator mi; for( mi = (*ri)->begin(); mi != (*ri)->end(); mi++ ){ (*mi)->addRing( (*ri) ); } vector::iterator bi; for( bi = (*ri)->getBonds()->begin(); bi != (*ri)->getBonds()->end(); bi++ ){ (*bi)->addRing( (*ri) ); (*bi)->getReverse()->addRing( (*ri) ); } } // resetBFSSearch for ( ai = beginAtom(); ai != endAtom(); ai++ ){ (*ai)->resetBFSVector(); } flagHasSSSRDetected = true; //#ifdef DEBUGSSSR cout << "detectSSSR() found " << sssr.size() << " rings" << endl; //#endif } setIntDescriptor( "sssrSize", sssr.size(), "NA","", true, true ); return( sssr.size() ); } Bond* Molecule::checkEdges( Atom* anAtom ){ // /todo implement Molecule::checkEdges! used at the end of detectSSSR. // this varies from the original algorithm in Figueras 1996 // the first bond of the atom is hidden. return( anAtom->hideToFromFirstBond() ); } bool Molecule::hasRing( Ring* aRing, bool detectingRing ) throw( CError ) { if( !flagHasSSSRDetected && detectingRing == false ){ //detectSSSR(); CError e( SSSRNOTDETECTED, "Smallest Set of Smallest Rings was not detected before calling Molecule::hasRing( Ring* )" ); e.describe(); throw(e); } vector::iterator i; for( i = sssr.begin(); i != sssr.end(); i++ ){ if( (*i)->equals( aRing ) ){ return( true ); } } return( false ); } bool Molecule::hasRing() throw( CError ){ if( !flagHasSSSRDetected ){ //detectSSSR(); CError e( SSSRNOTDETECTED, "Smallest Set of Smallest Rings was not detected before calling Molecule::hasRing()" ); e.describe(); throw(e); } if( sssr.size() > 0 ){ return( true ); } return( false ); } Ring* Molecule::getRingWithID( int anID, bool createIfMissing) throw( CError ){ if( !flagHasSSSRDetected ){ //detectSSSR(); CError e( SSSRNOTDETECTED, "Smallest Set of Smallest Rings was not detected before calling Molecule::hasRing()" ); e.describe(); throw(e); } if( sssr.size() > 0 ){ // check if this ring already exist vector::iterator ri; for( ri = sssr.begin(); ri != sssr.end(); ri++ ){ if( (*ri)->getID() == anID ){ return( (*ri) ); } } if( createIfMissing ){ Ring* newRing = new Ring(); newRing->setID( anID ); sssr.push_back( newRing ); return( newRing ); }else{ stringstream out; out << "There is no ring with id " << anID << " in molecule " << toStringShort(); CError e( MISSINGRING, out.str() ); e.describe(); throw(e); } }else{ if( createIfMissing ){ Ring* newRing = new Ring(); newRing->setID( anID ); sssr.push_back( newRing ); return( newRing ); }else{ stringstream out; out << "There is no ring with id " << anID << " in molecule " << toStringShort(); CError e( MISSINGRING, out.str() ); e.describe(); throw(e); } } } // no totters transformation void Molecule::noTottersTransform(){ vector< vector > coord; vector::iterator anAtom; map::iterator aBond; int i = -1; Molecule oldMol(*this, false); this->erase(); // DEBUG : //cout << "Nbre Atoms initiaux: " << oldMol.numAtoms() << endl; //cout << "Nbre laisons initiales : " << oldMol.numBonds() << endl; // 1st step : create new atoms : 1 per original atom + 1 per original bond // for each atom in the moldeule --> create an atom for( anAtom = oldMol.atoms.begin(); anAtom != oldMol.atoms.end(); anAtom++ ) { Atom* newAtom = new Atom( *(*anAtom) ); newAtom->setMorganLabel((*anAtom)->getMorganLabel() ); this->addAtom(newAtom, false, false); i++; coord.push_back(vector()); coord[i].push_back(0); // coord of the 1st atom of the bond (=0 in the case of an atom) coord[i].push_back((*anAtom)->getId()); // coord of the second atom of the bond coord[i].push_back(0); // bond label // for all the atom's bonds --> create a node for( aBond = (*anAtom)->beginBond(); aBond != (*anAtom)->endBond(); aBond++) { Atom* newAtom2 = new Atom(aBond->first->getLabel()); newAtom2->setAN(aBond->first->getAN()); newAtom2->setType(aBond->first->getType()); newAtom2->setMorganLabel(aBond->first->getMorganLabel() ); newAtom2->setKashimaPS(0.0); if( aBond->first->numBonds() == 1) { newAtom2->setKashimaPQ(1.0); // cout <<" --> set a PQ to 1.0" << endl; } else { newAtom2->setKashimaPQ(aBond->first->getKashimaPQ()); } this->addAtom(newAtom2,false,false); i++; coord.push_back(vector()); coord[i].push_back((*anAtom)->getId()); coord[i].push_back(aBond->first->getId()); coord[i].push_back(aBond->second->getLabel()); } } // DEBUG //cout << "Nbr atoms created : " << numAtoms() << endl; //for( anAtom = atoms.begin(); anAtom != atoms.end(); anAtom++ ) // { // cout << "\t\t (morgan label after atoms creation = "<< (*anAtom)->getMorganLabel() << " )"<< endl; // } // 2nd step : create the bonds from the coordinates vector vector coord1, coord2; int nbLinks = 0; for(i = 0 ; i < coord.size() ; i++) { coord1 = coord[i]; for(int j = 0 ; j < coord.size() ; j++) { coord2 = coord[j]; if(coord1[1] == coord2[0] && coord2[1] != coord1[0]) { linkAtomsNoReturn(i,j,coord2[2]); nbLinks++; } } } // cout << "Nbr bonds created : " << nbLinks << endl; // 3rd step : compute the transition probabilities for( anAtom = atoms.begin(); anAtom != atoms.end(); anAtom++ ) { //cout << "numbre of bonds per atom = " << (*anAtom)->numBonds() << endl; //cout << "\t\t (morgan label = "<< (*anAtom)->getMorganLabel() << " )"<< endl; for( aBond = (*anAtom)->beginBond(); aBond != (*anAtom)->endBond(); aBond++) { aBond->second->setKashimaPT( (1 - (*anAtom)->getKashimaPQ()) / (*anAtom)->numBonds() ); } } // cout << "\t in MOLECULE.cpp : nb atoms = " << numAtoms() << endl; // cout << "\t in MOLECULE.cpp : nb bonds = " << numBonds() << endl; // exit(0); } void Molecule::linkAtomsNoReturn( int aSource, int aTarget, int aBondLabel ) throw( CError ){ // atoms must exists otherwise cannot link if( max(aSource, aTarget) > numAtoms()-1 ){ stringstream out; out << "molecule " << toString() << " has only " << numAtoms() << " atoms, so cannot link atom " << aSource << " to " << aTarget ; CError e = CError(NOTENOUGHATOMSINMOLECULE, out.str() ); e.describe(); throw(e); }else{ linkAtomsNoReturn( atoms[aSource], atoms[aTarget], aBondLabel); } // resetMorganIndex(); // WARNING !!!! // --> modified by PM ?? // NB : consistant with linkAtomsNoReturn( Atom* aSource, Atom* aTarget, int aBondLabel) } // 3D KERNEL UTILITIES FUNCTIONS // --> matrix manipulations void Molecule::setAdjacency(int i, int j, double value){ (*adjacency)[i][j] = value; } double Molecule::getAdjacency(int i, int j){ return ( (*adjacency)[i][j] ); } void Molecule::setWalks(int i, int j, double value){ (*walks)[i][j] = value; } double Molecule::getWalks(int i, int j){ return ( (*walks)[i][j] ); } void Molecule::raisePowerAdjacency(){ //copy old walks matrix in a temp matrix vector< vector >* oldWalks; oldWalks = new vector< vector >; for(int i = 0 ; i < numAtoms() ; i++){ oldWalks->push_back(vector() ); for(int j = 0 ; j < numAtoms() ; j++){ (*oldWalks)[i].push_back( getWalks(i,j) ); } } // matrix product for(int i = 0 ; i < numAtoms() ; i++){ for(int j = 0 ; j < numAtoms() ; j++){ double temp= 0.0; for(int k = 0 ; k < numAtoms() ; k++){ temp += (*oldWalks)[i][k] * getAdjacency(k,j); } setWalks(i,j,temp); } } // free temp matrix for(int i = 0 ; i < numAtoms() ; i++){ (*oldWalks)[i].clear(); } //oldWalks->clear(); delete oldWalks; } double Molecule::traceWalks(){ double res = 0.0; for(int i = 0 ; i < numAtoms() ; i++){ res += getWalks(i,i); } return(res); } void Molecule::setSelfKernel(double value){ selfKernel = value; selfKernelCalculated = true; } double Molecule::traceDiagWalks(){ double res = 0.0; for(int i = 0 ; i < numAtoms() ; i++){ double tmp = 0.0; for(int j = 0 ; j < numAtoms() ; j++){ tmp += getWalks(i,j) * getAdjacency(j,i); } res += tmp; } return res; } // transform the 2D molecule into a 3D molecule // i.e., completely connected graphs, where edges labels are taken in {1,2,...,nBins} // where labels correspond to a discretization of the distance range into nBins bins void Molecule::threeDtransform(int nBins, double distMin, double distMax){ vector::iterator anAtom1, anAtom2; double atomDist; int bondLabel; double binSize; deleteBonds(); binSize = (1.0001*distMax - distMin) / nBins; // NB : trick to ensure that atomDist = distMax falls into the last bin. // this is equivalent to use bins of the form ]A,B], where A and B are the bounds of the bins // (except for the 1st bin : if atomDist = distMin, atomDist falls in the 1st bin) for(anAtom1 = beginAtom() ; anAtom1 != endAtom() - 1 ; anAtom1 ++){ for(anAtom2 = anAtom1 + 1 ; anAtom2 != endAtom() ; anAtom2 ++){ atomDist = atomicDistance(*anAtom1, *anAtom2); bondLabel = (int)( (atomDist - distMin)/binSize ) + 1; // nb : x = (int)y --> x is the truncation the float y //cout << "bondLabel = "<< bondLabel << endl; // OLD VERSION: based on distMin and distMax computed from input datasets //if(bondLabel <= 0 || bondLabel > nBins){ // cout << "ERROR : Molecule::threeDtransform : invalid bond label : " << endl; // cout << "aBondLabel = " << bondLabel << " (must be > 0 and <= " << nBins << ")" << endl; //} //linkAtomsNoReturn(*anAtom1, *anAtom2, bondLabel); //linkAtomsNoReturn(*anAtom2, *anAtom1, bondLabel); // // NB : linkAtoms = link forward + backward / linkAtomsNoReturn = just link forward // // HOWEVER, linAtomsNoReturn does not reset the Morgan Indices // // WARNING : need to check if this fact is due to a (previous) modification of PM (--> where?) // NEW VERSION: distMin and distMax defined a priori (i.e., parameters) // (NB: here, not all atoms are linked but only those having a distance within the distance range) if(bondLabel >= 0 && bondLabel <= nBins){ linkAtomsNoReturn(*anAtom1, *anAtom2, bondLabel); linkAtomsNoReturn(*anAtom2, *anAtom1, bondLabel); // NB : linkAtoms = link forward + backward / linkAtomsNoReturn = just link forward // HOWEVER, linAtomsNoReturn does not reset the Morgan Indices // WARNING : need to check if this fact is due to a (previous) modification of PM (--> where?) } } } } void Molecule::readPartialCharges(string charges){ vector atomCharges; StringUtils::Split(charges, ";", atomCharges); if(atomCharges.size() != numAtoms()){ cout << "ERROR : Molecule.setPartialCharges" << endl; cout << " --> number of charges != number of atoms" << endl; cout << " - numAtoms = " << numAtoms() << " ; numCharges = " << atomCharges.size() << endl; } int i = 0; vector< Atom* >::iterator anAtom; for( anAtom = atoms.begin(); anAtom != atoms.end(); anAtom++ ){ (*anAtom)->setPartialCharge(atof(atomCharges[i].c_str())); i++; } } void Molecule::setMorganChargesLabels(double threshold){ vector< Atom* >::iterator anAtom; for( anAtom = atoms.begin(); anAtom != atoms.end(); anAtom++ ){ (*anAtom)->setMorganChargeLabel(threshold); } } // ******************************* // // **** DEPRECATED FUNCTIONS **** // // ****************************** // /** function used to compute the graph kernel using the fused graph / sum of the powers of the transition matrix approach. See MoleculeUtils::powerKernel this function multiplies the adjacency matrix (containing the transition probabilities instead of 0 and 1) whith itself */ /*void Molecule::raisePower(){ vector::iterator i; vector::iterator j; vector::iterator ci; Bond* newBond; float aPt = 0; hideAllBonds(); // multiplier la matrice d'adjacence par elle-meme. L'ennui c'est qu'on a // pas de matrice d'adjacence. Faisons le avec les listes de pointeurs // remplissons les lignes de la pseudo matrice resultats //cout << endl; //int aci = 0; float toto = 0; for( i = beginAtom(); i != endAtom(); i++ ){ //cout << aci << endl; //aci++; // remplissons les colonnes de la pseudo matrice resultats if( (*i)->numHiddenBonds() > 0 ){ // les noeuds sans liaisons n'ont pas de pt... //int acj = 0; for( j = beginAtom(); j != endAtom(); j++ ){ //cout << " " << acj << endl; //acj++; // chaque entree de la nouvelle pseudo matrice est: if( (*j)->numHiddenBonds() > 0 ){ // les noeuds sans liaisons n'ont pas de pt... aPt = 0; for( ci = beginAtom(); ci != endAtom(); ci++){ toto = (*i)->getHiddenKashimaPT( (*ci) ); if( toto != 0 ){ aPt += toto * (*ci)->getSavedKashimaPT( (*j) ); } } if( aPt > 0 ){ newBond = linkAtomsNoReturn( (*i), (*j), 1 ); newBond->setKashimaPT( aPt ); //cout << " " << aPt; } } } } //cout << endl; } }*/ /** hide all bonds in the molecule */ /*void Molecule::hideAllBonds(){ vector::iterator ai; for( ai = beginAtom(); ai != endAtom(); ai++ ){ (*ai)->clearHiddenBonds(); (*ai)->hiddeAllBonds(); } } void Molecule::saveAllBonds(){ vector::iterator ai; for( ai = beginAtom(); ai != endAtom(); ai++ ){ (*ai)->saveAllBonds(); } } */ /*void Molecule::exportFragments( string sdFileName, int minAtoms ) { MoleculeSet* fragSet = new MoleculeSet(); fragSet->addFragmentsToSet( this, minAtoms ); fragSet->writeSD( sdFileName, false ); delete fragSet; }*/ Rchemcpp/src/chemcpp/src/atom.cpp0000644000175000017500000006742312116123101017404 0ustar klambaueklambaue/**************************************************************************************** atom.cpp ------------ copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #include "atom.h" #include #include //#define DEBUG 1 //#define DEBUGPERRETLABEL 1 int Atom::counter = 0; /** atom class constructor */ Atom::Atom() : Node() { #ifdef DEBUG cout << "Atom::Atom()" << endl; #endif Atom::counter++; id = Atom::counter; x = 0; y = 0; z = 0; flagHasCoordinates = false; an = 0; pq = 0; flagHasPq = false; ps = 0; flagHasPs = false; genericAtomType = false; #ifdef DEBUG cout << "Atom::Atom() DONE" << endl; #endif } /* the following function should be called to create an atom outside the elements it can be used to define a particular atom type. The only property therefore available from it is its label and its unique id. NOTE: atoms created this way are deleted when the molecule they are part of is deleted */ Atom::Atom( string aLabel ) : Node() { #ifdef DEBUG cout << "Atom::Atom( string aLabel )" << endl; #endif Atom::counter++; id = Atom::counter; x = 0; y = 0; z = 0; flagHasCoordinates = false; an = -1; pq = 0; flagHasPq = false; ps = 0; flagHasPs = false; label = aLabel; genericAtomType = true; #ifdef DEBUG cout << "Atom::Atom() DONE" << endl; #endif } Atom::Atom(const Atom& anAtom) : Node( (DataContainer&) anAtom ) { //: DataContainer( (DataContainer&) anAtom ){ //Atom::Atom(const Atom& anAtom) : DataContainer( anAtom ){ #ifdef DEBUG cout << "Atom::Atom(const Atom* anAtom)" << endl; #endif Atom::counter++; id = Atom::counter; x = 0; y = 0; z = 0; flagHasCoordinates = false; this->setAN( anAtom.getAN() ); this->setType( anAtom.getType() ); this->setLabel( anAtom.label ); // MODIFIED JLP BUG #ifdef DEBUG cout << "created atom " << getSymbol() << getId() << endl; #endif ps = anAtom.ps; flagHasPs = anAtom.flagHasPs; pq = anAtom.pq; flagHasPq = anAtom.flagHasPq; //copy bonds //map::iterator bit; //for ( bit = anAtom.bonds.begin(); bit != anAtom.bonds.end(); bit++ ){ //Bond* newBond = new Bond( *(*bit) ); //Atom* newAtom = new Atom( *elements["H"] ); //atoms.push_back( newAtom ); //this->addBond( newBond ); //} genericAtomType = false; resetMorganIndex(); #ifdef DEBUG //cout << "created atom " << toStringShort() << endl; cout << "Atom::Atom(const Atom* anAtom) DONE" << endl; #endif } Atom& Atom::operator=(const Atom& anAtom) { #ifdef DEBUG cout << "COPYING ATOM" << endl; #endif if(this != &anAtom){ /*this->kindStringDescriptors = anAtom.kindStringDescriptors; this->kindIntDescriptors = anAtom.kindIntDescriptors; this->kindFloatDescriptors = anAtom.kindFloatDescriptors;*/ // Copy all non kind descriptors // first copy int descriptors //cout << "copying An: " << anAtom.getAN() << endl; this->setAN( anAtom.getAN() ); this->setType( anAtom.getType() ); x = anAtom.x; y = anAtom.y; z = anAtom.z; flagHasCoordinates = anAtom.flagHasCoordinates; ps = anAtom.ps; flagHasPs = anAtom.flagHasPs; pq = anAtom.pq; flagHasPq = anAtom.flagHasPq; // increment the atom id and set the id for this atom Atom::counter++; id = Atom::counter; } resetMorganIndex(); return( *this ); } void Atom::describe() throw( CError ) { cout << getLabel() << endl; cout << getMorganLabel() << endl; cout << getPerretLabel() << endl; cout << numBonds() << " bonds " << endl; cout << numHiddenBonds() << " hidden bonds" << endl; DataContainer::describe(); cout << toString() << endl; cout << "coordinates: " << x << ", " << y << ", " << z << endl; cout << "An = " << getAN() << endl; cout << "Type = " << getType() << endl; } void Atom::describeShort() { cout << this->toString() << endl; } Bond* Atom::addBond(Bond *aBond) throw ( CError ){ #ifdef DEBUG cout << "adding bond [ " << aBond->toStringShort() << "]" << endl; #endif if( bondExists( aBond ) ){ //cout << "bond exists" << endl; stringstream out; out << "Bond " << aBond->toString() << " already exists "; CError e( BONDALREADYEXISTS, out.str() ); e.describe(); throw(e); return( bonds[aBond->getTarget()] ); }else{ bonds[aBond->getTarget()] = aBond; return( aBond ); } } Atom::~Atom(){ // delete all bonds #ifdef DEBUG cout << "Atom::~Atom()" << endl; #endif deleteBonds(); #ifdef DEBUG cout << "Atom::~Atom() DONE" << endl; #endif } /** returns a string describing the atom */ string Atom::toString(){ stringstream out; out << this->getStringDescriptor(SYMBOLNAME)->getValue() << this->getIdString() << " "; out << "PerretLabel: " << getPerretLabel() << " "; out << "Morgan: " << getMorganIndex(1) << " " << getMorganIndex(2) << " " << getMorganIndex(3) << " "; if( hasCoordinates() ){ out << " (" << x << ", " << y << ", " << z << ") "; }else{ //out << this->getStringDescriptor(SYMBOLNAME)->getValue() << this->getIdString() << " Morgan: " << getMorganIndex(1) << " " << getMorganIndex(2) << " " << getMorganIndex(3) << " " << " at " << this << " (" << numBonds() << " bonds)"; } out << " at " << this << " (" << numBonds() << " bonds)"; try{ if( hasRing() ){ out << ", in " << numRings() << " ring of size:"; vector::iterator ri; for( ri = beginRing(); ri != endRing(); ri++ ){ out << " " << (*ri)->size(); } out << " "; }else{ out << ", no rings "; } }catch( CError e ){ out << " rings not detected " << endl; } return( out.str() ); } /** returns a short string describing the atom */ string Atom::toStringShort(){ stringstream out; out << this->getStringDescriptor(SYMBOLNAME)->getValue() << this->getIdString(); return( out.str() ); } /** convenience function to retrieve the id as a string */ string Atom::getIdString(){ stringstream out; out << getId(); return( out.str() ); } ostream& operator << (ostream& os, const Atom& anAtom) { os << "ATOM " << anAtom.getId(); return os; } /** hide all bonds whose target is an hydrogen */ int Atom::hideHydrogenBonds(){ //deleteHiddenBonds(); //cout << "entering ATOM::hideHydrogenBonds" << endl; int i = 0; map::iterator bond; for( bond = beginBond(); bond != endBond(); ){ //cout << "checking " << (*bond).second->toString() << endl; if( (*bond).first->getAN() == 1 ){ //cout << "ERASING" << endl; //hiddenBonds.push_back( (*bond) ); hiddenBonds[(*bond).first] = (*bond).second; i++; map::iterator aBond = bond; bond++; bonds.erase((*aBond).first); if( bond != beginBond() ){ bond--; } //cout << "DONE" << endl; }else{ bond++; } } //cout << "leaving Atom::hideHydrogenBonds" << endl; return(i); } /** restore all hidden bonds (for example to hydrogens) */ int Atom::restoreHiddenBonds(){ #ifdef DEBUG cout << "tttty Atom::restoreHiddenBonds()" << endl; #endif int i = 0; map::iterator bond; for( bond = hiddenBonds.begin(); bond != hiddenBonds.end(); bond++ ){ // restore bonds bonds[(*bond).first] = (*bond).second; i++; } hiddenBonds.clear(); //hiddenBonds.erase( hiddenBonds.begin(), hiddenBonds.end() ); //cout << " there are now " << hiddenBonds.size() << " hidden bonds" << endl; //cout << " there are now " << bonds.size() << " bonds" << endl; return(i); } void Atom::restoreHiddenBond( Bond* aBond ) throw( CError ){ map::iterator bi; for( bi = hiddenBonds.begin(); bi != hiddenBonds.end(); bi++ ){ if( (*bi).second == aBond ){ bonds[(*bi).first] = (*bi).second; hiddenBonds.erase( bi ); return; } } stringstream out; out << "Bond " << aBond->toString() << " not found among hidden bounds"; CError e( BONDNOTFOUND, out.str() ); e.describe(); throw(e); } void Atom::restoreHiddenBond( Atom* aTarget ) throw( CError ){ Bond* aBond = hiddenBonds[aTarget]; if( hiddenBonds.find( aTarget ) == hiddenBonds.end() ){ stringstream out; out << "Bond " << aBond->toString() << " not found among hidden bounds"; CError e( BONDNOTFOUND, out.str() ); e.describe(); throw(e); }else{ #ifdef DEBUG cout << "Atom::restoreHiddenBond( Atom* ): " << aBond->toStringShort() << endl; #endif bonds[aBond->getTarget()] = aBond; } } /** sets the x, y, and z coordinate of the atom */ void Atom::setCoordinates( float aX, float aY, float aZ ){ //cout << " Atom::setCoordinates: setting coordinates to : " << aX << ", " << aY << ", " << aZ << endl; x = aX; y = aY; z = aZ; flagHasCoordinates = true; } /** erases the atoms coordinates (set flagHasCoordinates to false) */ void Atom::eraseCoordinates(){ flagHasCoordinates = false; } /** returns true if the atom has valid coordinates */ bool Atom::hasCoordinates(){ return( flagHasCoordinates ); } /** unset all bond flags */ void Atom::unsetBondFlags(){ map::iterator bond; for( bond = beginBond(); bond != endBond(); bond++ ){ (*bond).second->unsetFlag(); } } /** unset all bond flags */ void Atom::unsetBondFlagsOriginal(){ map::iterator bond; for( bond = beginBond(); bond != endBond(); bond++ ){ (*bond).second->unsetFlagOriginal(); } } bool Atom::bondExists(Bond* aBond){ map< Atom*, Bond* >::iterator i; for( i = bonds.begin(); i != bonds.end(); i++ ){ if( (*i).first == aBond->getTarget() ){ return( true ); } } return( false ); } /** sets the id of this atom in the molecule */ void Atom::setIdInMolecule( int anId ){ idInMolecule = anId; } /** sets the value of the ElementSymbol descriptor */ void Atom::setElementSymbol( string anElement ){ setStringDescriptor("ElementSymbol", anElement, "", "", true, true); } /** returns a pointer to the next unvisited Node throws an exception if all atoms were visited */ Atom* Atom::nextUnvisitedAtom() throw( CError ){ map::iterator i; for( i = bonds.begin(); i != bonds.end(); i++ ){ //if(!(*i).second->getSource()->wasVisited()){ // return( (*i).second->getSource() ); //} //cout << " was " << (*i).first->toStringShort() << " visited ? "; if( !(*i).first->wasVisited() ){ //cout << "no" << endl; return( (*i).first ); }//else{ //cout << "yes" << endl; //} } CError e( ALLNEIGHBOURSVISITED, "all neighbours to atom " + toStringShort() + " were already visited" ); throw(e); } /** hides a bond (but not reverse bond) */ void Atom::hideBond( map< Atom*, Bond* >::iterator aBondI ){ hiddenBonds[(*aBondI).first] = (*aBondI).second; bonds.erase( aBondI ); } void Atom::hideBond( Bond* aBond ){ hiddenBonds[ aBond->getTarget() ] = aBond; bonds.erase( aBond->getTarget() ); } Bond* Atom::hideBond( Atom* aTarget ){ Bond* result = bonds[ aTarget ]; #ifdef DEBUG cout << "Atom::hideBond( Atom* ): hiding bond " << result->toStringShort() << endl; #endif hiddenBonds[ aTarget ] = result; bonds.erase( aTarget ); return( result ); } /** hides all bonds to and from this Atom */ void Atom::hideAllToFromBonds(){ #ifdef DEBUG cout << "Atom::hideAllToFromBonds: hiding bonds from " << toStringShort() << " to " << bonds.size() << " atoms "<< endl; #endif int i = 0; map::iterator bond; for( bond = beginBond(); bond != endBond(); ){ i++; map::iterator aBond = bond; bond++; hideToFromBonds( (*aBond).first ); if( bond != beginBond() ){ bond--; } } #ifdef DEBUG cout << "Atom::hideAllToFromBonds done" << endl; #endif } // hides the first bond in bonds (both to and from bonds) // Bond* Atom::hideToFromFirstBond(){ return( hideToFromBonds( ( *bonds.begin() ).first ) ); } /** hides all bonds to and from to aTarget Atom */ Bond* Atom::hideToFromBonds( Atom* aTarget ){ // hide from bond #ifdef DEBUG cout << "hiding bond from " << aTarget->toStringShort() << " to this" << endl; #endif aTarget->hideBond( this ); // hide to bond #ifdef DEBUG cout << "hiding bond from this to " << aTarget->toStringShort() << endl; #endif Bond* result = hideBond( aTarget ); #ifdef DEBUG cout << "hiding bond to done" << endl; #endif return( result ); } /** returns the morgan index of order anOrder. Emits an error if it was not calculated before */ int Atom::getMorganIndex( int order ){ //order = order - 1; if( order < 1 ){ stringstream out; out << "Atom::getMorganIndex Morgan index cannot be calculated for an order < 1 (here : " << order << ")"; CError e(VALUENOTALLOWED, out.str() ); e.describe(); throw(e); } if( morganIndex.count( order ) == 1 ){ // the morgan index was already calculated return( morganIndex[ order ] ); }else{ // morgan as not been computed yet int res = 0; if( order > 1 ){ // the morgan index is the sum of the connectivity values of the neighboors res = getSumOfNeighboorMorganIndex( order - 1 ); morganIndex[ order ] = res; return( res ); }else{ // the first order morgan index is just the number of neighboors res = numBonds(); morganIndex[ 1 ] = res; return( res ); } } } /** returns the sum of the morgan indices of order anOrder of all neighboors. */ int Atom::getSumOfNeighboorMorganIndex( int anOrder ){ // for each neighboor add the morgan index of order anOrder int sum = 0; map::iterator i; for( i = bonds.begin(); i != bonds.end(); i++ ){ sum += (*i).first->getMorganIndex( anOrder ); } return( sum ); } long Atom::bondSum(){ // for each neighboor add the morgan index of order anOrder long sum = 0; map::iterator i; for( i = bonds.begin(); i != bonds.end(); i++ ){ sum += (*i).second->getLabel(); } return( sum ); } /** reset the morgan index map. Call this function whenever the molecule is modified! */ void Atom::resetMorganIndex(){ morganLabel = ""; uniqueMorganIndex = -1; morganIndex.clear(); } /** sets the morgan label (used in the Molecule::morganKernel function) to the concatenation of the atomSymbol and the morgan index of iteration anOrder */ void Atom::setMorganLabel( int anOrder ){ stringstream out; if( anOrder > 0 ){ out << getLabel() << getMorganIndex( anOrder ); }else{ out << getLabel(); } morganLabel = out.str(); } void Atom::setPerretLabel() throw( CError ){ stringstream out; out << getLabel(); try{ int numR = numRings(); #ifdef DEBUGPERRETLABEL cout << "Atom::setPerretLabel for atom: " << toStringShort() << endl; cout << " rings: " << numR << endl; #endif if( numR > 1 ){ if( numR > 2 ){ out << "K"; }else{ out << "J"; } } perretLabel = out.str(); #ifdef DEBUGPERRETLABEL cout << " -> " << perretLabel << endl; #endif //set the perretLabel for all bonds in this atom map::iterator bi; for( bi = bonds.begin(); bi != bonds.end(); bi++ ){ (*bi).second->setPerretLabel(); } }catch( CError e ){ CError e(NOTCALCULATED, "Atom::setPerretLabel: ring membership was not calculated, please use Molecule::setectSSSR before calling Molecule::setPerretLabel"); e.describe(); throw(e); } } int Atom::getNumAromaticBonds(){ map::iterator i; int numAromaticBonds = 0; for( i = bonds.begin(); i != bonds.end(); i++ ){ if( (*i).second->getLabel() == AROMATICBOND ){ numAromaticBonds++; } } return( numAromaticBonds ); } /** sets the unique morgan index of this atom to the value of anOrder iteration of the Morgan index calculation */ void Atom::setUniqueMorganIndex( int anOrder ){ uniqueMorganIndex = getMorganIndex( anOrder ); } string Atom::getMorganLabel( bool silentError ) throw( CError ){ /*if(morganLabel == ""){ // morgan label was not set, emit error CError e(NOTCALCULATED, "Atom::getMorganLabel: morganLabel was not calculated, use Molecule::setMorganLabels( int )"); if( silentError == false){ e.describe(); } throw(e); } */ return( morganLabel ); } string Atom::getPerretLabel( bool silentError ) throw( CError ){ return( perretLabel ); } int Atom::getUniqueMorganIndex( bool silentError ) throw( CError ){ if(uniqueMorganIndex == -1){ // unique morgan index was not set, emit error CError e(NOTCALCULATED, "Atom::getUniqueMorganIndex: uniqueMorganIndex was not calculated, use Molecule::setUniqueMorganIndices()"); if( silentError == false){ e.describe(); } throw(e); } return( uniqueMorganIndex ); } /** returns the label of the bond connecting this atom with otherAtom. Returns NOBOND if there are no bonds connecting both atoms */ Bond* Atom::getBondWithTarget( Atom* otherAtom ) throw( CError ){ if( bonds.find( otherAtom ) == bonds.end() ){ CError e( BONDNOTFOUND, "No bond with target " + otherAtom->toStringShort() + " in atom " + toStringShort() ); throw( e ); } return( bonds[ otherAtom ] ); } map::iterator Atom::getBondIteratorWithTarget( Atom* otherAtom ){ return( bonds.find( otherAtom ) ); } /** returns the transition probability to anAtom. Returns 0 if no atoms exist with anAtom */ double Atom::getKashimaPT( Atom* anAtom ){ try{ Bond* ab = getBondWithTarget( anAtom ); return( ab->getKashimaPT() ); }catch( CError ){ return( 0 ); } } /** delete all bonds in of that atom (called from the destructor of molecule) */ void Atom::deleteBonds(){ //cout << "Atom::deleteBonds - deleting bonds for atom " << toStringShort() << endl; map::iterator i; for( i = bonds.begin(); i != bonds.end(); i++ ){ #ifdef DEBUG cout << "deleting bond " << (*i).second->toStringShort() << endl; #endif delete (*i).second; // delete bonds } bonds.clear(); } void Atom::deleteHiddenBonds(){ //cout << "Atom::deleteBonds - deleting bonds for atom " << toStringShort() << endl; map::iterator i; for( i = hiddenBonds.begin(); i != hiddenBonds.end(); i++ ){ #ifdef DEBUG cout << "deleting bond " << (*i).second->toStringShort() << endl; #endif delete (*i).second; // delete bonds } hiddenBonds.clear(); } bool Atom::isCSkeleton() { // check if this atom is a carbon int numNonH = 0; if( getSymbol() == "C" ){ //cout << " got a carbon " << endl; // if true, check the neighbours if there are more than one non H than this is a // non terminal carbon map::iterator i; for( i = bonds.begin(); i != bonds.end(); i++ ){ //cout << " linked to " << (*i).first->getSymbol() << endl; if( (*i).first->getSymbol() != "H" ){ //cout << "."; numNonH++; //if( numNonH > 1){ // cout << " NOT A SKELETON" << endl; // return( true ); //} } } if( numNonH > 1 ){ //cout << " SKELETON" << endl; return( true ); }else{ //cout << " NOT SKELETON" << endl; return( false ); } }else{ return( false ); } } bool Atom::hasRing( Ring* aRing ){ vector::iterator ri; for( ri = rings.begin(); ri != rings.end(); ri++ ){ if( (*ri) == aRing ){ return( true ); } } return( false ); } Ring* Atom::getRingBFS( vector* toVisit, vector* toVisitBond ) throw( CError ){ #ifdef DEBUG cout << "a5h ATOM::GETRINGBFS now at " << toStringShort() << endl; #endif if( toVisit->size() == 0 ){ #ifdef DEBUG cout << "a5h ROOT" << endl; #endif toVisit = new vector; BFSVector.push_back( this ); // update the path for this }else{ BFSVector.push_back( (*toVisit->begin()) ); // update the path for this atom toVisit->erase( toVisit->begin() ); // erase the bond we just used from toVisit } if( toVisitBond->size() == 0 ){ toVisitBond = new vector; }else{ BFSBondVector.push_back( (*toVisitBond->begin()) ); // update the path for this atom toVisitBond->erase( toVisitBond->begin() ); // erase the bond we just used from toVisit } #ifdef DEBUG cout << "a5h with path " << toStringBFSVector() << endl; cout << " (length = " << BFSVector.size() << ")" << endl; cout << " (bond length = " << BFSBondVector.size() << ")" << endl; cout << "a5h ---->" << endl; #endif // add neighbour bonds with unvisited atoms to toVisit map::iterator i; for( i = bonds.begin(); i != bonds.end(); i++ ){ #ifdef DEBUG cout << "a5h checking " << (*i).second->toStringShort() << endl; cout << " with target " << (*i).first->toStringShort() << endl; #endif //cout << " cannot go to " << (*(BFSVector.end()-2))->toStringShort() << endl; if( BFSVector.size() > 1 && (*i).first == (*(BFSVector.end()-2)) ){ #ifdef DEBUG cout << " cannot go to " << (*(BFSVector.end()-2))->toStringShort() << endl; cout << "a5h return bond, skipping " << endl; #endif }else{ //if( (*i).first->degree() > 1 ){ //cout << "a5h has degree >1 " << endl; // check if target atom was visited if( (*i).first->getBFSVectorSize() == 0 ){ // target atom was not visited, add the bond to toVisit and set the path of the target #ifdef DEBUG cout << "a5h adding Bond " << (*i).second->toStringShort() << endl; #endif toVisit->push_back( (*i).first ); // add this atom to toVisit toVisitBond->push_back( (*i).second ); // add this bond to the list of bonds to follow. BFSBondVector.push_back( (*i).second ); (*i).second->getTarget()->pushBFSVector( &BFSVector, &BFSBondVector ); // set the target's path }else{ #ifdef DEBUG cout << "a5h TARGET WAS VISITED BEFORE" << endl; #endif // set the target's path // target atom was visited. Check if it is a valid ring // (if the paths of this and target atom // form an intersection with only one atom) vector vectorIntersect; getVectorIntersect( getBFSVector(), (*i).first->getBFSVector(), &vectorIntersect ); #ifdef DEBUG cout << "a5h ring has intersect of size " << vectorIntersect.size() << endl; #endif if( vectorIntersect.size() == 1 ){ // if yes create the ring and return a pointer to it BFSVector.push_back( (*i).second->getTarget() ); #ifdef DEBUG cout << "a5h ADDING LAST BOND: " << (*i).second->toStringShort() << endl; #endif BFSBondVector.push_back( (*i).second ); #ifdef DEBUG cout << "a5h NOW CREATING RING" << endl; cout << "a5h merging atoms: " << getBFSVector()->size() << " " << (*i).first->getBFSVector()->size() << endl; #endif vector vectorUnion; //getVectorUnion( getBFSVector(), (*i).first->getBFSVector(), &vectorUnion, &vectorIntersect ); MoleculeUtils::mergeSet( getBFSVector(), (*i).first->getBFSVector(), &vectorUnion ); #ifdef DEBUG cout << "a5h union vector has size " << vectorUnion.size() << endl; cout << "a5h merging bonds: " << getBFSBondVector()->size() << " " << (*i).first->getBFSBondVector()->size() << endl; #endif vector bondVectorUnion; MoleculeUtils::mergeBondSet( getBFSBondVector(), (*i).first->getBFSBondVector(), &bondVectorUnion ); // check that all bond source and target atoms are member of the ring #ifdef DEBUG cout << "a5h bond union vector has size " << bondVectorUnion.size() << endl; #endif vector bondVectorMembers; MoleculeUtils::selectRingMemberBonds( &bondVectorUnion, &vectorUnion, &bondVectorMembers ); #ifdef DEBUG cout << "a5h among which " << bondVectorMembers.size() << " are ring members " << endl; #endif // add reverse bonds to the set of rings vector::iterator bi; vector reverseBonds; for( bi = bondVectorMembers.begin(); bi != bondVectorMembers.end(); bi++ ){ try{ Bond* reverse = (*bi)->getReverse(); #ifdef DEBUG cout << "a5h added reverse bond " << reverse->toStringShort() << endl; #endif reverseBonds.push_back( reverse ); }catch( CError ){ } } for( bi = reverseBonds.begin(); bi != reverseBonds.end(); bi++ ){ bondVectorMembers.push_back( (*bi) ); } delete toVisit; delete toVisitBond; return( new Ring( &vectorUnion, &bondVectorMembers ) ); }else{ // it is not a valid ring, skip this bond } } //}else{ // cout << "a5h skipping Bond " << (*i).second->toStringShort() << " because " << (*i).second->getTarget()->toStringShort() << " has degree 1, cannot be member of a cycle" << endl; //} } } #ifdef DEBUG cout << "a5h " << toVisit->size() << " in toVisit" << endl; #endif if( toVisit->size() > 0 ){ // proceed to the nex toVisit Bond target atom return( (*toVisit->begin())->getRingBFS( toVisit, toVisitBond ) ); }else{ #ifdef DEBUG cout << "a5h NO RING THROWING ATOMNOTINRING" << endl; #endif delete toVisit; delete toVisitBond; CError e = CError( ATOMNOTINRING, "Atom " + toStringShort() + " is not a ring member " ); throw( e ); } } void Atom::pushBFSVector( vector* aPath, vector* aBondPath ){ #ifdef DEBUG cout << "b1h atom " << toStringShort() << " now has bfsVector:" << endl; #endif vector::iterator ai; for( ai = aPath->begin(); ai != aPath->end(); ai++ ){ BFSVector.push_back(*ai); #ifdef DEBUG (*ai)->describeShort(); #endif } #ifdef DEBUG cout << " and bfsBondVector " << endl; #endif vector::iterator bi; for( bi = aBondPath->begin(); bi != aBondPath->end(); bi++ ){ BFSBondVector.push_back(*bi); #ifdef DEBUG (*bi)->describeShort(); #endif } } void Atom::getVectorIntersect( vector* v1, vector* v2, vector* result){ vector::iterator ai; for ( ai = v1->begin(); ai != v1->end(); ai++ ){ vector::iterator aj; for ( aj = v2->begin(); aj != v2->end(); aj++ ){ if( (*aj) == (*ai) ){ result->push_back( (*ai) ); } } } } /*void Atom::getVectorUnion( vector* v1, vector* v2, vector* result, vector* intersect ){ cout << "a9h UNION" << endl; vector all; vector::iterator ai; for ( ai = v1->begin(); ai != v1->end(); ai++ ){ all.push_back( *ai ); cout << (*ai)->toStringShort() << ","; } for ( ai = v2->begin(); ai != v2->end(); ai++ ){ all.push_back( *ai ); cout << (*ai)->toStringShort() << ","; } cout << endl; MoleculeUtils::mergeSet( &all, intersect, result ); }*/ string Atom::toStringBFSVector(){ stringstream out; vector::iterator i; for( i = BFSVector.begin(); i!= BFSVector.end(); i++){ out << (*i)->toStringShort() << ", "; } return( out.str() ); } /** set the Morgan label of the instance to aLabel */ void Atom::setMorganLabel( string aLabel ){ morganLabel = aLabel; } /** set the value of the partial charge of the instance to aValue */ void Atom::setPartialCharge(double aValue){ partialCharge = aValue; } /** include the sign of the partial charge in the Morgan label of the instance. If partial charge > threshold: sign = + ; otherwise, sign = - */ void Atom::setMorganChargeLabel(double threshold){ if(partialCharge > threshold){ setMorganLabel(getMorganLabel() + "+"); } if(partialCharge < -1.0*threshold){ setMorganLabel(getMorganLabel() + "-"); } } Rchemcpp/src/chemcpp/src/jlpioutils.cpp0000644000175000017500000001003312116123101020623 0ustar klambaueklambaue/**************************************************************************************** jlpioutils.cpp ------------------ copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ //#define DEBUG 1 #include "jlpioutils.h" #include "cerror.h" #include JLPIOUtils::JLPIOUtils(){ } JLPIOUtils::~JLPIOUtils(){ } string JLPIOUtils::vectorToString( vector* a, string separator ){ stringstream result; vector::iterator i; bool first = true; for( i = a->begin(); i!= a->end(); i++ ){ if( first ){ result << (*i); first = false; }else{ result << separator << (*i); } } return ( result.str() ); } string JLPIOUtils::vectorToString( vector* a, string separator ){ stringstream result; vector::iterator i; bool first = true; for( i = a->begin(); i!= a->end(); i++ ){ if( first ){ result << (*i); first = false; }else{ result << separator << (*i); } } return ( result.str() ); } /** list files in a directory */ void JLPIOUtils::readDirectory( const string directoryLocation, vector< string > *result, string extension, long start, long end ){ #ifdef DEBUG cout << "JLPIOUtils::readDirectory " << start << " " << end << endl; #endif DIR* pDir = opendir ( directoryLocation.c_str() ); if ( !pDir ){ // return false; } dirent* pEntry; struct stat buf; long i = 0; while ( ( pEntry = readdir ( pDir ) ) ) { stat( pEntry->d_name, &buf ); string aFileName = pEntry->d_name; #ifdef DEBUG cout << "found " << aFileName << endl; #endif //#ifdef SGI //if( S_ISDIR( buf.st_mode ) ){ //for SGI, if it is not a directory //#else if( S_ISDIR( buf.st_mode ) ){ //for LINUX, if it is not a directory //#endif #ifdef DEBUG cout << " -> it is not a directory" << endl; #endif if( aFileName != "." && aFileName != ".." ){ if( extension == "" ){ string sFound = directoryLocation + "/" + aFileName; result->push_back ( sFound ); }else{ if( aFileName.length() > extension.length()-1 ) { #ifdef DEBUG cout << "comparing " << aFileName.substr( aFileName.length()-3,3 ).c_str() << " and " << extension.c_str() << endl; #endif if( aFileName.substr( aFileName.length()-3,3 ) == extension ){ #ifdef DEBUG cout << "ADDED " << directoryLocation + "/" + aFileName << endl; #endif string sFound = directoryLocation + "/" + aFileName; //cout << i << " " << (i >= start || start < 0 ) << " " << ( i <= end || end < 0 ) << endl; if( (i >= start || start < 0 ) && ( i <= end || end < 0 )){ result->push_back ( sFound ); } i++; } } } } #ifdef DEBUG }else{ cout << " -> it is a directory" << endl; #endif } } closedir( pDir ); } Rchemcpp/src/chemcpp/src/datacontainer.cpp0000644000175000017500000006266712116123101021265 0ustar klambaueklambaue /**************************************************************************************** datacontainer.cpp -------------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #include "datacontainer.h" #include "stringutils.h" //#define DEBUG 1 DataContainer::DataContainer(){ #ifdef DEBUG cout << "CREATING KIND DESCRIPTOR MAPS" << endl; #endif kindStringDescriptors = new map< string, Descriptor< string >* >; kindIntDescriptors = new map< string, Descriptor< int >* >; kindFloatDescriptors = new map< string, Descriptor< float >* >; #ifdef DEBUG cout << "string map has " << kindStringDescriptors->size() << " elements" << endl; #endif //cout << "ELEMENT ####" << endl; flagElement = true; // this datacontainer is a new datacontainer. Useful for destruction //cout << "Int map at :" << kindIntDescriptors << endl; } DataContainer::~DataContainer(){ #ifdef DEBUG cout << "DataContainer::~DataContainer()" << endl; #endif deleteAllDescriptors(); #ifdef DEBUG cout << "shall I delete kind descriptors? "; #endif if( flagElement == true ){ #ifdef DEBUG cout << "yes" << endl; #endif deleteAllKindDescriptors(); } #ifdef DEBUG else{ cout << "no" << endl; } #endif #ifdef DEBUG cout << "DataContainer::~DataContainer() DONE" << endl; #endif } DataContainer::DataContainer( DataContainer& aDataContainer ){ #ifdef DEBUG cout << "DataContainer::DataContainer( DataContainer& aDataContainer )" << endl; #endif //cout << "DataContainer copy constructor " << endl; // set kindDescritors pointer to the same as the source kindStringDescriptors = aDataContainer.kindStringDescriptors; kindIntDescriptors = aDataContainer.kindIntDescriptors; kindFloatDescriptors = aDataContainer.kindFloatDescriptors; //cout << kindStringDescriptors << endl; // copy all non kind descriptors #ifdef DEBUG cout << "copying all non kind descriptors" << endl; #endif map* >::iterator iti; int i = 0; //cout << " int" << endl; //cout << (aDataContainer).intDescriptors.size() << endl; for( iti = (aDataContainer).intDescriptors.begin(); iti != (aDataContainer).intDescriptors.end(); iti++ ){ //cout << i << endl; i++; int aValue = 0; bool empty = false; try{ aValue = (*iti).second->getValue(true); }catch( CError e ){ if( e.getType() == ERRORNA ){ empty = true; continue; }else{ e.describe(); throw( e ); } } Descriptor* toto = addIntDescriptor( (*iti).second->getLabel() , aValue, (*iti).second->getUnit(), (*iti).second->getComment() ); if( empty == true ){ toto->setEmpty(); } } #ifdef DEBUG cout << " int DONE" << endl; #endif #ifdef DEBUG cout << " float" << endl; cout << (aDataContainer).floatDescriptors.size() << endl; #endif map* >::iterator itf; for( itf = (aDataContainer).floatDescriptors.begin(); itf != (aDataContainer).floatDescriptors.end(); itf++ ){ //cout << i << endl; i++; //cout << (*itf).second->getLabel() << endl; float aValue = 0.0; bool empty = false; try{ aValue = (*itf).second->getValue(true); }catch( CError e ){ if( e.getType() == ERRORNA ){ empty = true; continue; }else{ e.describe(); throw( e ); } } Descriptor* toto = addFloatDescriptor( (*itf).second->getLabel() , aValue, (*itf).second->getUnit(), (*itf).second->getComment() ); if( empty == true ){ toto->setEmpty(); } } #ifdef DEBUG cout << "float done" << endl; #endif map* >::iterator its; for( its = aDataContainer.stringDescriptors.begin(); its != aDataContainer.stringDescriptors.end(); its++ ){ //cout << i << endl; i++; string aValue = ""; bool empty = false; try{ aValue = (*its).second->getValue(true); }catch( CError e ){ if( e.getType() == ERRORNA ){ empty = true; continue; }else{ e.describe(); throw( e ); } } Descriptor* toto = addStringDescriptor( (*its).second->getLabel() , aValue, (*its).second->getUnit(), (*its).second->getComment() ); if( empty == true ){ toto->setEmpty(); } } //cout << "copied " << i << " non kind descriptors" << endl; #ifdef DEBUG cout << "string done" << endl; #endif #ifdef DEBUG cout << "copying all non kind descriptors DONE" << endl; #endif flagElement = false; // this datacontainer is a copy of an existing datacontainer. Useful for destruction #ifdef DEBUG cout << "DataContainer::DataContainer( DataContainer& aDataContainer ) DONE" << endl; #endif } /*DataContainer& DataContainer::operator= (const DataContainer& aDataContainer){ // copy kind descriptors // first string descriptors map* >::iterator its; // for( its = (aDataContainer).kindStringDescriptors.begin(); its != (aDataContainer).kindStringDescriptors.end(); its++ ){ // addKindIntDescriptor( (*its).second -> getLabel(), (*its).second ); // } // then integer descriptors // then float descriptors }*/ /** Adds an integer kind descriptor with a label, value, unit and comment */ Descriptor< int >* DataContainer::addKindIntDescriptor(string aLabel, int aValue, string aUnit, string aComment){ //cout << "setting " << (*kindIntDescriptors)[aLabel] << " to " << atoi(aValue.c_str()) << endl; //(*kindIntDescriptors)[aLabel] = new Descriptor( aLabel, atoi(aValue.c_str()), aUnit, aComment ); //cout << "map now contains " << endl; //(*kindIntDescriptors)[aLabel]->describe(); (*kindIntDescriptors)[aLabel] = new Descriptor( aLabel, aValue, aUnit, aComment ); return( (*kindIntDescriptors)[aLabel] ); } Descriptor< int >* DataContainer::addKindIntDescriptor(Descriptor< int >* aDescriptor){ (*kindIntDescriptors)[aDescriptor->getLabel()] = aDescriptor; return( aDescriptor ); } /** Adds a float kind descriptor with a label, value, unit and comment */ Descriptor< float >* DataContainer::addKindFloatDescriptor(string aLabel, float aValue, string aUnit, string aComment){ (*kindFloatDescriptors)[aLabel] = new Descriptor( aLabel, aValue, aUnit, aComment ); return( (*kindFloatDescriptors)[aLabel] ); } Descriptor< float >* DataContainer::addKindFloatDescriptor(Descriptor< float >* aDescriptor){ (*kindFloatDescriptors)[aDescriptor->getLabel()] = aDescriptor; return( aDescriptor ); } /** Adds a string kind descriptor with a label, value, unit and comment */ Descriptor< string >* DataContainer::addKindStringDescriptor(string aLabel, string aValue, string aUnit, string aComment){ (*kindStringDescriptors)[aLabel] = new Descriptor( aLabel, aValue, aUnit, aComment ); return( (*kindStringDescriptors)[aLabel] ); } Descriptor< string >* DataContainer::addKindStringDescriptor(Descriptor< string >* aDescriptor){ (*kindStringDescriptors)[aDescriptor->getLabel()] = aDescriptor; return( aDescriptor ); } /** Adds an integer descriptor with a label, value, unit and comment */ Descriptor< int >* DataContainer::addIntDescriptor(string aLabel, int aValue, string aUnit, string aComment){ Descriptor< int >* d = new Descriptor( aLabel, aValue, aUnit, aComment ); intDescriptors[aLabel] = d; return( d ); } /** Adds an float descriptor with a label, value, unit and comment */ Descriptor< float >* DataContainer::addFloatDescriptor(string aLabel, float aValue, string aUnit, string aComment){ Descriptor< float >* d = new Descriptor( aLabel, aValue, aUnit, aComment ); floatDescriptors[aLabel] = d; return( d ); } /** Adds an integer descriptor with a label, value, unit and comment */ Descriptor< string >* DataContainer::addStringDescriptor(string aLabel, string aValue, string aUnit, string aComment){ //cout << "DataContainer::addStringDescriptor"<* d = new Descriptor( aLabel, aValue, aUnit, aComment ); stringDescriptors[aLabel] = d; //cout << "DataContainer::addStringDescriptor DONE"<* DataContainer::setIntDescriptor(string aLabel, int aValue, string aUnit, string aComment, bool addIfMissing, bool silentError ){ // check if string descriptor exists. If not create it Descriptor< int >* d; if( hasIntDescriptor( aLabel ) ){ d = intDescriptors[aLabel]; d -> setValue(aValue); }else{ if( addIfMissing ){ d = addIntDescriptor( aLabel, aValue, aUnit, aComment ); }else{ CError e( MISSINGDESCRIPTOR, "DataContainer::setIntDescriptor: no descriptor " + aLabel ); if( !silentError ){ e.describe(); } throw(e); } } return( d ); } Descriptor< float >* DataContainer::setFloatDescriptor(string aLabel, float aValue, string aUnit, string aComment, bool addIfMissing, bool silentError ){ // check if string descriptor exists. If not create it Descriptor< float >* d; if( hasFloatDescriptor( aLabel ) ){ d = floatDescriptors[aLabel]; //cout << "found float descriptor " << d->toString() << endl; d -> setValue( aValue ); }else{ if( addIfMissing ){ d = addFloatDescriptor( aLabel, aValue, aUnit, aComment ); }else{ CError e( MISSINGDESCRIPTOR, "DataContainer::setFloatDescriptor: no descriptor " + aLabel ); if( !silentError ){ e.describe(); } throw(e); } } return ( d ); } Descriptor< string >* DataContainer::setStringDescriptor(string aLabel, string aValue, string aUnit, string aComment, bool addIfMissing, bool silentError ){ // check if string descriptor exists. If not create it Descriptor< string >* d; if( hasStringDescriptor( aLabel ) ){ d = stringDescriptors[aLabel]; d -> setValue(aValue); }else{ if( addIfMissing ){ d = addStringDescriptor( aLabel, aValue, aUnit, aComment ); }else{ CError e( MISSINGDESCRIPTOR, "DataContainer::setStringDescriptor: no descriptor " + aLabel ); if( !silentError ){ e.describe(); } throw(e); } } return( d ); } /** write a only descriptors of the data container to cout (not Kind descriptors) */ void DataContainer::describeShort() throw( CError ) { map* >::iterator its; map* >::iterator iti; map* >::iterator itf; // first write string descriptors for( its = stringDescriptors.begin(); its != (stringDescriptors).end(); its++ ){ (*its).second->describeShort(); } // then write integer descriptors for( iti = intDescriptors.begin(); iti != (intDescriptors).end(); iti++ ){ (*iti).second->describeShort(); } // finally write float descriptors for( itf = floatDescriptors.begin(); itf != (floatDescriptors).end(); itf++ ){ (*itf).second->describeShort(); } cout << "-------------------------" << endl; } /** write a description of the data container to cout */ void DataContainer::describe() throw( CError ) { string valueS; cout << "KIND DESCRIPTORS:" << endl; cout << "string kindDescriptors:" << endl; // first write string kindDescriptors map* >::iterator its; for( its = (*kindStringDescriptors).begin(); its != (*kindStringDescriptors).end(); its++ ){ (*its).second->describe(); } cout << "int kindDescriptors:" << endl; // then write integer kindDescriptors map* >::iterator iti; for( iti = (*kindIntDescriptors).begin(); iti != (*kindIntDescriptors).end(); iti++ ){ (*iti).second->describe(); } cout << "float kindDescriptors" << endl; // finally write float kindDescriptors map* >::iterator itf; for( itf = (*kindFloatDescriptors).begin(); itf != (*kindFloatDescriptors).end(); itf++ ){ (*itf).second->describe(); } cout << "DESCRIPTORS:" << endl; // first write string descriptors for( its = stringDescriptors.begin(); its != (stringDescriptors).end(); its++ ){ (*its).second->describe(); } // then write integer descriptors for( iti = intDescriptors.begin(); iti != (intDescriptors).end(); iti++ ){ (*iti).second->describe(); } // finally write float descriptors for( itf = floatDescriptors.begin(); itf != (floatDescriptors).end(); itf++ ){ (*itf).second->describe(); } cout << "-------------------------" << endl; } Descriptor< string >* DataContainer::getStringDescriptor( string aLabel, bool silentError ) throw( CError ){ /*bool found; // look among kind descriptors first map* >::iterator its; for( its = (*kindStringDescriptors).begin(); its != (*kindStringDescriptors).end(); its++ ){ if( (*its).first == aLabel ){ found = true; return ( (*its).second ); } } // look among descriptors defined for this instance for( its = stringDescriptors.begin(); its != stringDescriptors.end(); its++ ){ if( (*its).first == aLabel ){ found = true; return ( (*its).second ); } } // if no corresponding descriptors found, throw exception CError e( MISSINGDESCRIPTOR, "no descriptor " + aLabel ); e.describe(); throw(e); */ //cout << "looking among " << kindStringDescriptors->size() << " string kind descriptors" << endl; map< string, Descriptor* >::iterator answer = kindStringDescriptors->find(aLabel); //cout << "and?" << endl; if( answer != kindStringDescriptors->end() ){ //cout << "found" << endl; return(answer->second); }else{ //cout << "looking in non kind descriptors" << endl; answer = stringDescriptors.find(aLabel); if( answer != stringDescriptors.end() ){ //cout << "found" << endl; return(answer->second); }else{ //cout << "not found" << endl; // if no corresponding descriptors found, throw exception CError e( MISSINGDESCRIPTOR, "no descriptor " + aLabel ); if( silentError == false ){ e.describe(); } throw(e); } } } long DataContainer::getPossibleValuesInIntDescriptor( string aDescriptorName, vector< int >* p ){ long res = 0; if( aDescriptorName.substr( aDescriptorName.length() - min( (int) aDescriptorName.length(),8), 8 ) == ".integer"){ aDescriptorName = aDescriptorName.substr( 0, aDescriptorName.size() - 8 ); } map< const string, Descriptor* >::iterator iti; int i = 0; for( iti = intDescriptors.begin(); iti != intDescriptors.end(); iti++ ){ if( (*iti).second->getLabel() == aDescriptorName ){ // check if this value was already found, if not add it to the vector of possible values int newValue = (*iti).second->getValue(); vector::iterator vi; long nfound = 0; for( vi = p->begin(); vi != p->end(); vi++){ if( (*vi) == newValue ){ nfound += 1; break; } } if( nfound == 0 ){ p->push_back( newValue ); res += 1; } } } return( res ); } Descriptor< int >* DataContainer::getIntDescriptor( string aLabel, bool silentError ) throw( CError ){ /* bool found; // look among kind descriptors first map* >::iterator its; for( its = (*kindIntDescriptors).begin(); its != (*kindIntDescriptors).end(); its++ ){ if( (*its).first == aLabel ){ found = true; return ( (*its).second ); } } // look among descriptors defined for this instance for( its = intDescriptors.begin(); its != intDescriptors.end(); its++ ){ if( (*its).first == aLabel ){ found = true; return ( (*its).second ); } } // if no corresponding descriptors found, throw exception CError e( MISSINGDESCRIPTOR, "no descriptor " + aLabel ); e.describe(); throw(e); */ map< string, Descriptor* >::iterator answer = kindIntDescriptors->find(aLabel); if( answer != kindIntDescriptors->end() ){ return(answer->second); }else{ answer = intDescriptors.find(aLabel); if( answer != intDescriptors.end() ){ return(answer->second); }else{ // if no corresponding descriptors found, throw exception CError e( MISSINGDESCRIPTOR, "no descriptor " + aLabel ); if( silentError == false ){ e.describe(); } throw(e); } } } Descriptor< float >* DataContainer::getFloatDescriptor( string aLabel, bool silentError ) throw( CError ){ //bool found; map< string, Descriptor* >::iterator answer; // look among kind descriptors first /*map* >::iterator its; for( its = (*kindFloatDescriptors).begin(); its != (*kindFloatDescriptors).end(); its++ ){ if( (*its).first == aLabel ){ found = true; return ( (*its).second ); } } // look among descriptors defined for this instance for( its = floatDescriptors.begin(); its != floatDescriptors.end(); its++ ){ if( (*its).first == aLabel ){ found = true; return ( (*its).second ); } } */ answer = kindFloatDescriptors->find(aLabel); if( answer != kindFloatDescriptors->end() ){ return(answer->second); }else{ answer = floatDescriptors.find(aLabel); if( answer != floatDescriptors.end() ){ return(answer->second); }else{ // if no corresponding descriptors found, throw exception CError e( MISSINGDESCRIPTOR, "no descriptor " + aLabel ); if( silentError == false ){ e.describe(); } throw(e); } } } /** deletes the int descriptor aString from intDescriptors (deletes the object and removes pointer from the Hash) */ bool DataContainer::deleteDescriptor( string aLabel, bool found ){ // first check for the descriptor among the int descriptors map* >::iterator iti; for( iti = (intDescriptors).begin(); iti != (intDescriptors).end(); iti++ ){ if( (*iti).first == aLabel ){ #ifdef DEBUG cout << "deleting int descriptor [ " << (*iti).second->toString() << "]" << endl; #endif delete (*iti).second; (intDescriptors).erase(iti); found = true; return( true ); } } // then check for the descriptor among the float descriptors map* >::iterator itf; for( itf = (floatDescriptors).begin(); itf != (floatDescriptors).end(); itf++ ){ if( (*itf).first == aLabel ){ #ifdef DEBUG cout << "deleting float descriptor [ " << (*itf).second->toString() << "]" << endl; #endif delete (*itf).second; (floatDescriptors).erase(itf); found = true; return( true ); } } // finally check for the descriptor among the string descriptors map* >::iterator its; for( its = (stringDescriptors).begin(); its != (stringDescriptors).end(); its++ ){ if( (*its).first == aLabel ){ #ifdef DEBUG cout << "deleting string descriptor [ " << (*its).second->toString() << "]" << endl; #endif delete (*its).second; (stringDescriptors).erase(its); found = true; return( true ); } } return( found ); } /** deletes all descriptors from ***Descriptors (deletes the object and removes pointer from the Hash) */ void DataContainer::deleteAllDescriptors(){ // first delete int descriptors #ifdef DEBUG int i = 0; #endif map* >::iterator iti; for( iti = intDescriptors.begin(); iti != intDescriptors.end(); iti++ ){ #ifdef DEBUG cout << "deleting int descriptor [ " << (*iti).second->toString() << "]" << endl; i++; #endif delete (*iti).second; //(intDescriptors).erase(iti); } intDescriptors.clear(); // then delete float descriptors map* >::iterator itf; for( itf = floatDescriptors.begin(); itf != floatDescriptors.end(); itf++ ){ #ifdef DEBUG cout << "deleting float descriptor [ " << (*itf).second->toString() << "]" << endl; i++; #endif delete (*itf).second; //(floatDescriptors).erase(itf); } floatDescriptors.clear(); // finally delete string descriptors map* >::iterator its; for( its = stringDescriptors.begin(); its != stringDescriptors.end(); its++ ){ #ifdef DEBUG cout << "deleting string descriptor [ " << (*its).second->toString() << "]" << endl; i++; #endif delete (*its).second; //(stringDescriptors).erase(its); } stringDescriptors.clear(); #ifdef DEBUG cout << " DELETED " << i << " DESCRIPTORS" << endl; #endif } /** deletes all kindDescriptors from ***Descriptors (deletes the object and removes pointer from the Hash) */ void DataContainer::deleteAllKindDescriptors(){ // first delete int descriptors //cout << "REMOVING ALLKINDDESCRIPTORS ####" << endl; #ifdef DEBUG cout << "DataContainer::deleteAllKindDescriptors" << endl; #endif int i = 0; map* >::iterator iti; for( iti = (kindIntDescriptors)->begin(); iti != (kindIntDescriptors)->end(); iti++ ){ #ifdef DEBUG cout << "deleting kindInt descriptor [ " << (*iti).second->toString() << "]" << endl; #endif delete (*iti).second; i++; //(kindIntDescriptors).erase(iti); } kindIntDescriptors->clear(); // then delete float descriptors map* >::iterator itf; for( itf = (kindFloatDescriptors)->begin(); itf != (kindFloatDescriptors)->end(); itf++ ){ #ifdef DEBUG cout << "deleting kindFloat descriptor [ " << (*itf).second->toString() << "]" << endl; #endif delete (*itf).second; i++;; //(kindFloatDescriptors).erase(itf); } kindFloatDescriptors->clear(); // finally delete string descriptors map* >::iterator its; for( its = (kindStringDescriptors)->begin(); its != (kindStringDescriptors)->end(); its++ ){ #ifdef DEBUG cout << "deleting kindString descriptor [ " << (*its).second->toString() << "]" << endl; #endif delete (*its).second; i++; //(kindStringDescriptors).erase(its); } kindStringDescriptors->clear(); delete kindIntDescriptors; delete kindFloatDescriptors; delete kindStringDescriptors; #ifdef DEBUG cout << " DELETED " << i << " KIND DESCRIPTORS" << endl; #endif } /** returns true if the datacontainer has a string descriptor with label aLabel */ bool DataContainer::hasStringDescriptor( string aLabel ){ if( stringDescriptors.find( aLabel ) == stringDescriptors.end() ){ return( false ); }else{ return( true ); } } /** returns true if the datacontainer has a float descriptor with label aLabel */ bool DataContainer::hasFloatDescriptor( string aLabel ){ if( floatDescriptors.find( aLabel ) == floatDescriptors.end() ){ return( false ); }else{ return( true ); } } /** returns true if the datacontainer has a int descriptor with label aLabel */ bool DataContainer::hasIntDescriptor( string aLabel ){ if( intDescriptors.find( aLabel ) == intDescriptors.end() ){ return( false ); }else{ return( true ); } } /** adds a descriptor to the datacontainer. If aName terminates with .string, a string descriptor will be added. If the aName terminates with a .integer, an integer descriptor will be added. If aName terminates with .float, a float descriptor will be added. */ void DataContainer::addUnknownTypeDescriptor( string aDescriptorName, string aValue ){ if( aDescriptorName.substr( aDescriptorName.length() - min( (int) aDescriptorName.length(),8), 8 ) == ".integer"){ aDescriptorName = aDescriptorName.substr( 0, aDescriptorName.size() - 8 ); setIntDescriptor( aDescriptorName, StringUtils::toInt( aValue ), "", "", true, true ); }else if( aDescriptorName.substr( aDescriptorName.length() - min( (int) aDescriptorName.length(),4), 4 ) == ".int" ){ aDescriptorName = aDescriptorName.substr( 0, aDescriptorName.size() - 4); setIntDescriptor( aDescriptorName, StringUtils::toInt( aValue ), "", "", true, true ); }else if( aDescriptorName.substr( aDescriptorName.length() - min( (int) aDescriptorName.length(),6), 6 ) == ".float"){ aDescriptorName = aDescriptorName.substr( 0, aDescriptorName.size() - 6 ); setFloatDescriptor( aDescriptorName, StringUtils::toFloat( aValue ), "", "", true, true ); }else if( aDescriptorName.substr( aDescriptorName.length() - min( (int) aDescriptorName.length(),4), 4 ) == ".flo" ){ aDescriptorName = aDescriptorName.substr( 0, aDescriptorName.size() - 4 ); setFloatDescriptor( aDescriptorName, StringUtils::toFloat( aValue ), "", "", true, true ); }else if( aDescriptorName.substr( aDescriptorName.length() - min( (int) aDescriptorName.length(),7), 7 ) == ".string"){ aDescriptorName = aDescriptorName.substr( 0, aDescriptorName.size() - 7 ); setStringDescriptor( aDescriptorName, aValue, "", "", true, true ); }else if( aDescriptorName.substr( aDescriptorName.length() - min( (int) aDescriptorName.length(),4), 4 ) == ".str" ){ aDescriptorName = aDescriptorName.substr( 0, aDescriptorName.size() - 4 ); setStringDescriptor( aDescriptorName, aValue, "", "", true, true ); }else{ setStringDescriptor( aDescriptorName, aValue, "", "", true, true ); } } Rchemcpp/src/chemcpp/src/cerror.cpp0000644000175000017500000000657312116123101017737 0ustar klambaueklambaue/**************************************************************************************** cerror.cpp ------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #include "cerror.h" CError::CError( int anError, string aComment ) { why = anError; comment = aComment; } CError::~CError(){ } void CError::describe(){ switch(why){ case BADFILE: cerr << "BAD FILE error:"; break; case MISSINGDESCRIPTOR: cerr << "MISSING DESCRIPTOR error:"; break; case NOTENOUGHATOMSINMOLECULE: cerr << "NOT ENOUGH ATOMS IN MOLECULE error: "; break; case FILENOTFOUND: cerr << "FILE NOT FOUND error: "; break; case ERRORNA: cerr << "EMPTY VALUE ERROR: "; break; case UNKNOWNDATATYPE: cerr << "UNKNOWN DATA TYPE ERROR: "; break; case ERRORATOMNOTFOUND: cerr << "ATOM NOT FOUND: "; break; case VALUENOTALLOWED: cerr << "VALUE NOT ALLOWED: "; break; case NOTIMPLEMENTED: cerr << "FEATURE NOT YET IMPLEMENTED: "; break; case NOTFOUND: cerr << "NOT FOUND: "; break; case COULDNOTOPENFILE: cerr << "COULD NOT OPEN FILE: "; break; case DUPLICATEENTRIES: cerr << "DUPLICATE ENTRIES: "; break; case DEPRECATED: cerr << "DEPRECATED: "; break; case THREADCREATIONERROR: cerr << "THREADCREATIONERROR: "; break; case EOFERROR: cerr << "EOFERROR: "; break; case MISSINGDATA: cerr << "MISSING DATA: "; break; case NOTCALCULATED: cerr << "NOTCALCULATED: "; break; case BADVALUE: cerr << "BADVALUE: "; break; case BONDALREADYEXISTS: cerr << "BONDALREADYEXISTS: "; break; case IOERROR: cerr << "IOERROR: "; break; case NOSTRUCTURE: cerr << "NOSTRUCTURE: "; break; case ATOMNOTFOUND: cerr << "ATOMNOTFOUND: "; break; case SSSRNOTDETECTED: cerr << "SSSRNOTDETECTED: "; break; case MISSINGRING: cerr << "MISSINGRING: "; break; case ATOMNOTINRING: cerr << "ATOMNOTINRING: "; break; case ALLNEIGHBOURSVISITED: cerr << "ALLNEIGHBOURSVISITED: "; break; case BONDNOTFOUND: cerr << "BONDNOTFOUND: "; break; case MOLECULENOTFOUND: cerr << "MOLECULENOTFOUND: "; break; default: cerr << "ERROR type:" << getType(); break; } cerr << " " << getComment() << endl; } Rchemcpp/src/chemcpp/src/ring.cpp0000644000175000017500000000734412116123101017377 0ustar klambaueklambaue/**************************************************************************************** ring.cpp ------------ copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #include "ring.h" #include #include #include #include Ring::Ring( vector* atomsList, vector* bondsList ){ vector::iterator ai; for( ai = atomsList->begin(); ai != atomsList->end(); ai++ ){ push_back( *ai ); } vector::iterator bi; for( bi = bondsList->begin(); bi != bondsList->end(); bi++ ){ bonds.push_back( *bi ); } #ifdef DEBUG cout << "a7h created ring with " << size() << " atoms and " << bonds.size() << " bonds" << endl; #endif } Ring::Ring(){ } Ring::~Ring() { } bool Ring::equals( Ring* anotherRing ){ // two rings are equal if they are the same size and contain the same atoms if( size() != anotherRing->size() ){ return( false ); } // they are the same size so let's check if they contain the same atoms vector::iterator ai; for( ai = begin(); ai != end(); ai++ ){ if( anotherRing->hasAtom( (*ai) ) ){ }else{ return( false ); } } return( true ); } bool Ring::hasAtom( Atom* anAtom ){ vector< Atom* >::iterator ai; for( ai = begin(); ai != end(); ai++ ){ if( (*ai) == anAtom ){ return( true ); } } return( false ); } bool Ring::hasBond( Bond* aBond ){ vector< Bond* >::iterator ai; for( ai = bonds.begin(); ai != bonds.end(); ai++ ){ if( (*ai) == aBond ){ return( true ); } } return( false ); } string Ring::toStringShort(){ stringstream out; out << "Ring of size " << size(); return( out.str() ); } string Ring::toString(){ stringstream out; out << toStringShort() << " whith " << bonds.size() << " bonds"; return( out.str() ); } void Ring::describeShort(){ cout << toStringShort() << endl; } void Ring::describe(){ cout << toString() << endl; } void Ring::addBond( Bond* aBond, bool silentError ) throw( CError ){ if( !hasBond( aBond ) ){ // add the bond to the vector of bonds bonds.push_back( aBond ); }else{ if( !silentError ){ CError e = CError( BONDALREADYEXISTS, "Ring::addBond: cannot add bond because bond already exist in ring" ); throw( e ); } } } void Ring::addAtom( Atom* anAtom, bool silentError ) throw( CError ){ if( !hasAtom( anAtom ) ){ // add the bond to the vector of bonds push_back( anAtom ); }else{ if( !silentError ){ CError e = CError( ATOMALREADYEXISTS, "cannot add atom Ring::addAtom: because atom already exist in ring" ); throw( e ); } } } Rchemcpp/src/chemcpp/src/datacontainer.h0000644000175000017500000003150512116123101020715 0ustar klambaueklambaue /**************************************************************************************** datacontainer.h ------------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef DATACONTAINER_H #define DATACONTAINER_H #include #include #include #include using namespace std; using std::map; using std::ostream; #include "descriptor.h" #include "cerror.h" /** Container for scientific values of type int, float and string @author Dr Jean-Luc Perret (luc@kuicr.kyoto-u.ac.jp), Kyoto University, Japan @version 0.1 @date 8 Jan 2004 CLASS NAME: DataContainer FOR: SNSF SPONSORED PROJECT PURPOSE: Base Class implementing a collection of Descriptors and a collection of kind Descriptors ( of types int, float, and string ) DataContainer is a base class to be used to derive classs representing concepts which need to be described with real values (properties) We distinguish two kind of properties: unique descriptors and kind descriptors. Unique descriptors are properties whose values are unique to an instance of the concept. Kind descriptors are properties whose values are shared among a group of concept instances. For example the class Atom is derived from DataContainer. Atoms have unique descriptors like a name, a start and stop probability. On the other hand atoms have many kind descriptors which are shared among atoms belonging to the same chemical element. The atomic number of all hydrogens is the same, but is different for carbons. Kind descriptors allow for this. Kind desriptors are stored only once in memory and are then referenced by all class instances of the same kind can therefore be added at any time but should be removed with caution since many class instances may point to it. In chemcpp generic instances of Atoms are instanciated globally in the elements.h file by creation of an Elements instance and stored in the elements map. Kind descriptors are created at that time and will only be deleted with the deletion of the unique instance of Elements. \todo deletion of kindDescriptors by elements is not yet implemented Memory allocation for unique descriptors on the other hand can be managed by the DataContainer class itself since only one instance of the derived class points to them. So deletion of the DataContainer causes the deletion of the descriptors. \warning unique descriptors should not be referenced by something else than this container class. */ class DataContainer { public: ///@name DataContainer construction functions //@{ /** class constructor. */ DataContainer(); /** class constructor. */ DataContainer( DataContainer& aDataContainer ); //DataContainer& operator= (const DataContainer& aDataContainer); /** (virtual) class desctructor. */ virtual ~DataContainer(); //@} ///@name Accessor functions //@{ /** returns true if the datacontainer has a string descriptor with label aLabel. */ bool hasStringDescriptor( string aLabel ); /** returns true if the datacontainer has a float descriptor with label aLabel. */ bool hasFloatDescriptor( string aLabel ); /** returns true if the datacontainer has an int descriptor with label aLabel. */ bool hasIntDescriptor( string aLabel ); //@} ///@name DataContainer manipulation functions //@{ /** adds an integer kind descriptor with a label, value, unit and comment to the int kind descriptor container. WARNING this descriptor will be deleted with the destruction of the Elements class. */ Descriptor< int >* addKindIntDescriptor( string aLabel, int aValue, string aUnit, string aComment ); /** adds an existing int descriptor to the int kind descriptor container. WARNING this descriptor will be deleted with the destruction of the Elements class. */ Descriptor< int >* addKindIntDescriptor(Descriptor< int>* aDescriptor); /** adds an float kind descriptor with a label, value, unit and comment to the float kind descriptor container. WARNING this descriptor will be deleted with the destruction of the Elements class. */ Descriptor< float >* addKindFloatDescriptor( string aLabel, float aValue, string aUnit, string aComment ); /** adds an existing float descriptor to the float kind descriptor container. WARNING this descriptor will be deleted with the destruction of the Elements class. */ Descriptor< float >* addKindFloatDescriptor(Descriptor< float>* aDescriptor); /** adds a string kind descriptor with a label, value, unit and comment to the string kind descriptor container. WARNING this descriptor will be deleted with the destruction of the Elements class. */ Descriptor< string >* addKindStringDescriptor( string aLabel, string aValue, string aUnit, string aComment ); /** adds an existing string descriptor to the string kind descriptor container. WARNING this descriptor will be deleted with the destruction of the Elements class. */ Descriptor< string >* addKindStringDescriptor(Descriptor< string>* aDescriptor); /** adds an integer unique descriptor with a label, value, unit and comment. this descriptor will be deleted with the destruction of this class. */ Descriptor< int >* addIntDescriptor( string aLabel, int aValue, string aUnit, string aComment ); /** adds a float unique descriptor with a label, value, unit and comment. this descriptor will be deleted with the destruction of this class. */ Descriptor< float >* addFloatDescriptor( string aLabel, float aValue, string aUnit, string aComment ); /** adds a string unique descriptor with a label, value, unit and comment. this descriptor will be deleted with the destruction of this class. */ Descriptor< string >* addStringDescriptor( string aLabel, string aValue, string aUnit, string aComment ); /** sets the value of an int label (unique or kind descriptor). */ Descriptor< int >* setIntDescriptor(string aLabel, int aValue, string aUnit, string aComment, bool addIfMissing, bool silentError ); /** sets the value of float label (unique or kind descriptor). */ Descriptor< float >* setFloatDescriptor(string aLabel, float aValue, string aUnit, string aComment, bool addIfMissing, bool silentError ); /** sets the value of a string label (unique or kind descriptor). */ Descriptor< string >* setStringDescriptor(string aLabel, string aValue, string aUnit, string aComment, bool addIfMissing, bool silentError); /** returns a int property (unique or kind descriptor). */ Descriptor< int >* getIntDescriptor( string aLabel, bool silentError = true ) throw( CError ); /** fills a vector with the possible values of a int descriptor and returns the number of possible values. */ long getPossibleValuesInIntDescriptor( string aDescriptorName, vector< int >* ); /** returns a float property (unique or kind descriptor). */ Descriptor< float >* getFloatDescriptor( string aLabel, bool silentError = true ) throw( CError ); /** returns a string property (unique or kind descriptor). */ Descriptor< string >* getStringDescriptor( string aLabel, bool silentError = true ) throw( CError ); /** deletes the unique descriptor aString from Descriptors. deletes the object and removes pointer from the Hash, does not remove kind descriptors. */ virtual bool deleteDescriptor( string aString, bool found = false ); /** deletes all unique descriptors from ***Descriptors. deletes the objects and removes pointer from the Hash, does not remove kind descriptors. */ void deleteAllDescriptors(); /** adds a descriptor to the datacontainer. If aName terminates with .string, a string descriptor will be added. If the aName terminates with a .integer, an integer descriptor will be added. If aName terminates with .float, a float descriptor will be added. */ void addUnknownTypeDescriptor( string aName, string aValue ); //@} ///@name Iterators //@{ /** start iterator for int descriptors. */ map< string, Descriptor< int >* >::iterator beginIntDescriptor(){ return( intDescriptors.begin() ); } /** end iterator for int descriptors. */ map< string, Descriptor< int >* >::iterator endIntDescriptor(){ return( intDescriptors.end() ); } /** start iterator for float descriptors. */ map< string, Descriptor< float >* >::iterator beginFloatDescriptor(){ return( floatDescriptors.begin() ); } /** end iterator for float descriptors. */ map< string, Descriptor< float >* >::iterator endFloatDescriptor(){ return( floatDescriptors.end() ); } /** start iterator for string descriptors. */ map< string, Descriptor< string >* >::iterator beginStringDescriptor(){ return( stringDescriptors.begin() ); } /** end iterator for string descriptors. */ map< string, Descriptor< string >* >::iterator endStringDescriptor(){ return( stringDescriptors.end() ); } //@} ///@name Output functions //@{ /** writes a description of all labels to cout (unique and kind descriptors). */ void describe() throw( CError ); /** writes a description of unique descriptors only to cout (no kind descriptors). */ void describeShort() throw( CError ); //@} protected: /** false if this datacontainer was constructed by copy of an existingdatacontainer. Useful for destruction. */ bool flagElement; /** called by the destructor if flagElement == true. */ void deleteAllKindDescriptors(); // Data container provides two different Descriptors: // - kind descriptors which are shared by a group of instances of DataContainers // - descriptors which are unique to an instance of this class. // Kind descriptors can be created on the heap by this class, but are never deleted. // on the other hand descriptors are fully managed by this class. WARNING: // Descriptors are deleted by the destructor of this class so they // should not be used by reference from elsewhere. // DESCRIPTORS // Hash of pointers on Descriptor<> to store properties of an instance of // this class // Objects in the hash are created by the add***Descriptor function // Objects are deleted in the deleteDescriptor or the // deleteAllDescriptors functions (also called by the destructor) /** hash of pointers on Integer descriptors. */ map< string, Descriptor< int >* > intDescriptors; /** hash of pointers on Float descriptors. */ map< string, Descriptor< float >* > floatDescriptors; /** hash of pointers on String descriptors. */ map< string, Descriptor< string >* > stringDescriptors; // KIND DESCRIPTORS // Hashes of pointers on Descriptor<> to store properties shared by a group of // instances of this class. The values of these descriptors are shared among the group // of instances. Example: the kind descriptor symbol with value H // is shared among all hydrogen instances of the derived class atom. // // Objects in these maps can be created by the function // addKind***Descriptor( string aLabel, float aValue, string aUnit, string aComment ) // which returns a pointer to the created descriptor. // An already created descriptor can also be included in the map by the function // addKind***Descriptor( descriptor& ) /** hash of pointers on Integer kind descriptors. */ map< string, Descriptor< int >* >* kindIntDescriptors; /** hash of pointers on Float kind descriptors. */ map< string, Descriptor< float >* >* kindFloatDescriptors; /** hash of pointers on String kind descriptors. */ map< string, Descriptor< string >* >* kindStringDescriptors; // Hash of Desriptor // Hashes of pointers on Descriptor // use when datacontainer Class instances should only contain a reference to a Descriptor // map< string, Descriptor< int >* > intPDescriptors; // map< string, Descriptor< float >* > floatPDescriptors; // map< string, Descriptor< string >* > stringPDescriptors; }; #endif Rchemcpp/src/chemcpp/src/jlpioutils.h0000644000175000017500000000511012116123101020270 0ustar klambaueklambaue/**************************************************************************************** jlpioutils.h ---------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef JLPIOUTILS_H #define JLPIOUTILS_H #include #include #include #include #include #include using namespace std; /**Provides utilities for io operations *@author Jean-Luc Perret */ class JLPIOUtils { public: /** class constructor. */ JLPIOUtils(); /** class desctructor. */ ~JLPIOUtils(); /** list files in a directory. usage: vector\ fileVector; readDirectory ( "/home/anydir", fileVector); start and end allow to limit the number of file returned: limiting from start to end (included) starting count from 0 the defauld (-1) indicates no limits. */ static void readDirectory( const string directoryLocation, vector< string > *result, string extension = "", long start = -1, long end = -1 ); /** concatenates the entries of a vector of float into a string, separated by a given separator. */ static string vectorToString(vector* a, string separator = "\t"); /** concatenates the entries of a vector of double into a string, separated by a given separator. */ static string vectorToString(vector* a, string separator = "\t"); }; #endif Rchemcpp/src/chemcpp/src/descriptor.h0000644000175000017500000001505112116123101020255 0ustar klambaueklambaue/**************************************************************************************** descriptor.h ---------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef DESCRIPTOR_H #define DESCRIPTOR_H #include #include using std::string; using std::endl; using std::stringstream; #include #include /** Template class for storing scientific values @author Dr Jean-Luc Perret (luc@kuicr.kyoto-u.ac.jp), Kyoto University, Japan @version 0.3 @date 17 Jan 2004 CLASS NAME: Descriptor FOR: SNSF SPONSORED PROJECT PURPOSE: Storage type for a general scientific descriptor with label, value, unit and comment This template class implements the storage of a general scientific measurment It not only stores a value with its label, but also the units used, and a comment. the label, unit and comment are all strings. The value however can be of any type since its type is a tamplate T. Creating a new descriptor can be done in that way: (in this example we are interested in a descriptor for body temperature and we initialize its value to 37.5) Descriptor myDescriptor = new Descriptor("temperature", 37.5, "degrees C", "body temperature"); If we don't want the descriptor to contain any value we can give it any value and call the setEmpty() function. After that any call to getValue() will throw an exception, until a value is assigned with setValue(T) Descriptors of type int, float and string are used by the DataContainer class. */ template class Descriptor { public: ///@name Descriptor construction functions //@{ /** class constructor. */ Descriptor(); /** class constructor. */ Descriptor(T avalue); /** class constructor. */ Descriptor(string aLabel, T aValue, string aUnit, string aComment); //~Descriptor(){} //@} ///@name Accessor functions //@{ /** sets the value of the descriptor. */ void setValue( T ); /** sets the label of the descriptor. */ void setLabel( string aLabel ){ label = aLabel; } /** sets the unit of the descriptor. */ void setUnit( string aUnit ){ unit = aUnit; } /** sets the comment of the descriptor. */ void setComment( string aComment){ comment = aComment; } /** empties the value of the Descriptor. isEmpty() function will then return true. */ void setEmpty(){ empty = true; } /** returns the value. Throws a CError exception if the descriptor has been setEmpty(). */ T getValue( bool silentError = false ) throw( CError ); /** returns the label string. */ string getLabel(){ return( label ); } /** returns the unit string. */ string getUnit(){ return( unit ); } /** returns the comment string. */ string getComment(){ return( comment ); } /** returns true is the value was not set, false otherwise. */ bool isEmpty(){ return(empty); } //@} ///@name Output functions //@{ /** returns a string description of the descriptor. */ string toString(); /** returns a string description of the descriptor omiting the comment. */ string toStringShort(); /** prints a description of the descriptor to cout. */ void describe(); /** prints a description of the descriptor to cout omiting the comment. */ void describeShort(); //@} private: /** string representing the descriptor name, example: temperature. */ string label; /** variable representing the value of the descriptor, exemple: 37.5. */ T value; /** string representing the descriptor unit, example: degree C. */ string unit; /** string representing a comment for the descriptor, example: body temperature. */ string comment; /** flag telling if the Descriptor is empty. */ bool empty; }; template Descriptor::Descriptor(){ setLabel( "no name" ); setUnit( "no unit" ); setComment( "no comment" ); empty = true; } template Descriptor::Descriptor( T avalue ){ //Descriptor(); setLabel( "no name" ); setUnit( "no unit" ); setComment( "no comment" ); empty = true; setValue( avalue ); } template Descriptor::Descriptor( string aLabel, T aValue, string aUnit, string aComment ){ setLabel( aLabel ); setUnit( aUnit ); setComment( aComment ); setValue( aValue ); } template void Descriptor::setValue( T avalue ){ value = avalue; empty = false; } template T Descriptor::getValue( bool silentError ) throw( CError ){ if( !isEmpty() ){ return value; }else{ // value is emtpy, so ERROR_NA exception is raised CError e( ERRORNA, getLabel() + " is empty" ); if( silentError == false){ e.describe(); } throw( e ); } } template void Descriptor::describe(){ cout << toString() << endl; } template void Descriptor::describeShort(){ cout << toStringShort() << endl; } template string Descriptor::toString(){ stringstream out; if( empty == false ){ out << getComment() << ": " << getLabel() << " = " << getValue() << " (" << getUnit() << ") "; }else{ out << getComment() << ": " << getLabel() << " = " << "NA" << " (" << getUnit() << ") "; } return( out.str() ); } template string Descriptor::toStringShort(){ stringstream out; if( empty == false ){ out << getLabel() << " = " << getValue() << " (" << getUnit() << ") "; }else{ out << getLabel() << " = " << "NA" << " (" << getUnit() << ") "; } return( out.str() ); } #endif Rchemcpp/src/chemcpp/src/chemcpp.cpp0000644000175000017500000000315312116123101020051 0ustar klambaueklambaue/*************************************************************************** * Copyright (C) 2005 by Jean-Luc Perret / Pierre Mahé * * jean-luc.perret@net2000.ch / pierre.mahe@ensmp.fr * * * * This program is free software; you can redistribute it and/or modify * * it under the terms of the GNU General Public License as published by * * the Free Software Foundation; either version 2 of the License, or * * (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * * GNU General Public License for more details. * * * * You should have received a copy of the GNU General Public License * * along with this program; if not, write to the * * Free Software Foundation, Inc., * * 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. * ***************************************************************************/ #ifdef HAVE_CONFIG_H #include #endif #include #include using namespace std; int main(int argc, char *argv[]) { cout << "Hello, world!" << endl; return EXIT_SUCCESS; } Rchemcpp/src/chemcpp/src/moleculeset.h0000644000175000017500000006174412116123101020432 0ustar klambaueklambaue/**************************************************************************************** moleculeset.h ----------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef MOLECULESET_H #define MOLECULESET_H #include #include #include #include #include #include #include //#include #include #include #include #define VERBOSECALC 1 /**Set of molecules on which virtual experiments may be performed @author Dr Jean-Luc Perret (luc@kuicr.kyoto-u.ac.jp), Kyoto University, Japan @version 0.3 @date 17 Jan 2004 CLASS NAME: MoleculeSet FOR: SNSF SPONSORED PROJECT PURPOSE: This class implements the notion of molecule set It is a set in the mathematical way in the sense that no two molecule should have the same name. However a MoleculeSet can contain two identical graphs. WARNING: No checks are made for throwing error in case there are two molecules with the same name, except when calling the [] operator. */ class MoleculeSet : public std::vector { /** \example moleculeset_example.cpp */ public: ///@name MoleculeSet construction functions //@{ /** class constructor. */ MoleculeSet(); MoleculeSet(const MoleculeSet& aSet); /** destructor for the MoleculeSet. deletes gram and gramNormal. */ ~MoleculeSet(); /** adds all molecules in aSet to the current set. */ int add( MoleculeSet* aSet ); /** adds a molecule to the set. molecules added to the molecule set are not deleted when the set is deleted. use the deleteAllMolecule() function to do so (CAUTION if you reference these molecules from elsewhere). */ void addMolecule( Molecule* aMolecule ); /** adds a copy of the molecule in argument. used by add() to merge two datasets. */ Molecule* addMoleculeCopy( Molecule* aMolecule ); /** deletes all molecules in dataset (memory deallocation) and clears the vector containing the pointers. */ void deleteAll(); //@} ///@name Input functions //@{ /** adds an sd file content and returns the number of created molecules. if genericAtomType is true, then the atoms are not read from the periodic table, but are created based only on the label provided. beginMolecule and endMolecule specify the index of the first and last molecule to include in the training set (starting count from 0). values of -1 (default) means no limit. */ int addSD( string aFileName, bool genericAtomType = false, long beginMolecule = -1, long endMolecule = -1 ); /** reads an KCF file and returns the number of created molecules. */ int addKCF( string aFileName, long beginMolecule = -1, long endMolecule = -1 ); /** creates a new molecule in the dataset and reads its definition from a MDL MOL file. */ Molecule* addSingleMOL( string aMolFile, bool genericAtomType = false ); /** creates a new molecule in the dataset and reads its definition from a Kcf file. */ Molecule* addSingleKCF( string aMolFile ); /** loads all .mol files in a directory and adds the molecules to the set. */ void readMolDirectory( string aPath, bool genericAtomType = false, long beginMolecule = -1, long endMolecule = -1 ); /** loads all .kcf files in a directory and adds the molecules to the set. */ void readKcfDirectory( string dataDir, long beginMolecule = -1, long endMolecule = -1 ); /** reads the Mutag dataset. Atoms and bonds are read from aFilename while the biological activity of the molecule is read from file rFilename. numMolToRead allows to specify the number of first molecules to read. */ void addMutag( string aFileName, string rFileName="", uint numMolToRead = 500 ); /** reads the activity of molecules from an activity file. expects a file containing : label [tab] class molname [tab] 1 for active molecules molname [tab] -1 for inactive molecules molname is the fileName of the Mol file. */ void readActivityFile( string aFileName ); /** reads a descriptor file. the first line in the file indicates the Descriptor comment text. the second line in the file indicates the Descriptor units. the third line in the file indicates the Descriptor types. the forth line indicates the Descriptor names. data start at fifth line. the first column should contain the same molecule names as those obtained by getName() for the molecules in the dataset. example: Full name;Log octanol partition coefficient;boiling point;special comment NA;NA;K;NA string;float;float;string name;logP;Bp;acomment ethane;10;200;junk data WARNING: does not check for duplicate name entries in the descriptor file. WARNING: does not add anything to the molecules not contained in the descriptor file. */ void readDescriptorFile( string aFileName, string separator = ";" ); /** reads the content of a file produced by gist-classify and adds the information matching molecule names in the set as descriptors gistname, gistclass, and gistdiscr. */ void readGistClassifyFile( string aFileName ); /** reads the content of a gist activity file and adds the information to descriptor aDescriptor for each molecule matching molecule names. */ void readGistActivityFile( string aFileName, string aDescriptor ); /** reads a gram matrix matching the dataset. emits an error if the dimension of the read gram matrix read does not match the number of compounds in the dataset. */ void readGram( string aFileName, vector< vector >* gram ); /** reads a normalized gram matrix matching the dataset. */ void readGramNormal( string aFileName ); /** reads a raw gram matrix matching the dataset. */ void readGramRaw( string aFileName ); /** reads the partial charges associated to the molecule set from an input file. NOTE : the input file has one line per molecule of the molecule set, and within each line, the values of the partial charges are separated by ';'. */ void readPartialCharges(string fileName); //@} ///@name Accessor functions //@{ /** returns the number of molecules in the MoleculeSet. */ uint numMolecules(){ return( size() ); } /** returns a pointer to the first molecule with name aName in the set. Throws a CError exception if no molecule with that name exists in the set. WARNING: if more than one molecule have the same name than a CError is also thrown. */ Molecule* operator[]( string aName ) throw( CError ); /** returns a pointer to the first molecule with name aName in the set. Throws a CError exception if no molecule with that name exists in the set. WARNING: if more than one molecule have the same name than a CError is also thrown. */ Molecule* getMolByName( string aName ) throw( CError ); /** returns a pointer to the molecule of index anInd in the set. */ Molecule* operator[]( int anInd ) throw( CError ); /** returns a pointer to the molecule of index anInd in the set. */ Molecule* getMolByIndex( int anInd ) throw( CError ); /** fills a vector of int with the values an int descriptor can take among all molecules and returns the number of such values. */ long getPossibleValuesInIntDescriptor( string aDescriptorName, vector< int >* p ); /** returns the stop probability currently in use in the MoleculeSet for the calculation of random walk graph Kernel. */ double getPq(){ return( pq ); } /** returns the convergence condition currently in use in the MoleculeSet for the calculation of random walk graph Kernel. */ int getConvergenceCondition(){ return( convergenceCondition ); } /** returns true if a molecule with name aName exists in the set, false otherwise. */ bool nameExists( string aName ); /** returns true if the activity of molecules in the set was set using readActivityFile. WARNING: no check is made to verify if all molecules were set. */ bool hasActivity(){ return activitySet; } /** sets integer aName to aValue for all compounds in the dataset. */ void setIntDescriptor( string aName, int aValue ); /** sets the uniqueMorganIndex of each atom to the Morgan index having the maximum of different connectivity values for the molecule. */ void setUniqueMorganIndices(); /** sets the morganLabels of each molecule to the anOrder iteration of the Morgan index calculation process. */ void setMorganLabels( int anOrder ); /** sets the value of intDescriptor aLabel of molecule aLabel to aValue. */ void setIntDescriptor( string aLabel, string aMolecule, int aValue ); /** sets the avtivity of aMolecule to aValue. */ void setActivity( string aMolecule, float aValue ); /** sets the start, stop (aPq) and transition probabilities. sets the start, stop (aPq) and transition probabilities for all molecules according to the article by Kashima et al. using the setKashimaKernelProb(aFloat) function of the Molecule class. WARNING: a call to this function erases the gram Matrix and it will be recalculated. */ void setKashimaKernelParam( double aPq, int aConvergenceCondition, bool skipSkeleton = false ); /** sets the comparison set of the moleculeSet. */ void setComparisonSet( MoleculeSet* ); /** sets the 'morgan charges' labels of the atoms, i.e., the concatenation of the Morgan labels of the atoms and the (+/-) sign of their partial charges. */ void setMorganChargesLabels(double threshold); //@} ///@name MoleculeSet manipulation functions //@{ /** selects all molecules in the dataset. WARNING the selection status is stored in the Molecule class. Therefore molecules in other datasets will be selected too... */ void selectAll(); /** unselects all molecules in the dataset. WARNING the selection status is stored in the Molecule class. Therefore molecules in other datasets will be unselected too... */ void unSelectAll(); /** selects the molecules with the names provided as arguments in a vector of string. unselect all others. returns the number of selected molecules. */ int select( vector< string >* aSubset ); /** unselects the molecules with the names provided as arguments in a vector of string. returns the number of selected molecules. */ int unSelect( vector< string >* aSubset ); /** selects all molecules having float descriptor aName with value aValue, unselects all others. */ long selectByFloatDescriptor( string aName, float aValue ); /** selects all molecules having int descriptor aName with value aValue, unselects all others. */ long selectByIntDescriptor( string aName, int aValue ); /** selects all molecules having activity equal to aValue. */ long selectByActivity( float aValue ); /** selects all molecules which have a defined activity status. */ long selectHasActivity(); /** selects all molecules in the dataset with mw >= minmw and <= maxmw. if maxmw = -1 then there is no maximum limit. if addMolecularDescriptor is true then a floatDescriptor with label mw is added to the molecule. */ int selectByMW( float minmw, float maxmw = -1 , bool addMolecularDescriptor = false ); /** selects all molecules in the dataset with number of atoms >= numAtoms and <= numAtoms. if maxNumAtoms = -1 then there is no maximum limit. if addMolecularDescriptor is true then a intDescriptor with label numAtoms is added to the molecule. */ int selectByNumAtoms( float minNumAtoms, float maxNumAtoms = -1 , bool addMolecularDescriptor = false ); /** sorts the molecule collection according to the molecule Descriptor descriptorName of type descriptorType. if reverse = true then the sorting is in decreasing order. */ void sortByDescriptor( string aDescriptorName, int aDescriptorType, bool reverse = false ); /** sorts the molecule collection according to the molecule Descriptor descriptorName. descriptor type will be read from descriptor name if name is of type ******.integer, or *******.float, set to string otherwise. if reverse = true then the sorting is in decreasing order. */ void sortByDescriptor( string aDescriptorName, bool reverse = false ); /** sorts all compounds in the set by Molecular weight. */ void sortByMW(); /** sorts all compounds in the set by their number of atoms. */ void sortByNumAtoms(); /** sets the activity of molecules in the set according to the value of a descriptor and a value. if true is passed as third argument (default) then descriptorName <= value are considered positive. if false is passed as third argument then molecules with descriptorName >= value are considered positive. if a descriptor is missing then the molecule is left without activity. */ void binClassifyFromDescriptor( string descriptorName, float value, bool smallerOrEqual = true ); /** returns a pointer to the molecule with name aName in the MoleculeSet. */ Molecule* findFirstMoleculeWithName( string aName ) throw( CError ); /** removes duplicate entries in the set. */ void removeDuplicates(); /** deletes all hidden atoms in all molecules of the set. */ void deleteHiddenAtoms(); /** hides all hydrogens in all molecules of the set. */ void hideHydrogens(); /** hides all but the largest connected graph in all molecules of the set. */ void hideSalts( string aReportFileName = "" ); /** restores the hidden atoms for all molecules. */ void restoreHiddenAtoms(); /** NOT DOCUMENTED */ void addFragmentsToSet( Molecule* aMol, int minAtoms = 1 ); /** NOT DOCUMENTED */ void pushFragments( Molecule* aMol, int minAtoms = 1 ); /** returns the mean distance of all molecules to the moleculeSet barycenter. */ double diversityBaryMean(); /** lists the different kinds of atoms present in the set based on their Morgan labels. */ vector atomsLabelsListing(); /** lists the different kinds of atoms present in the set based on their symbols. */ vector atomsSymbolsListing(); /** lists the different kinds of bonds present in the set. */ vector bondsListing(); /** transforms the molecular graphs into graph preventing tottering paths (see (Mahe et al., 2004)). */ void noTottersTransform(); /** transforms the molecular graphs into '3D complete graphs' with edges labeled by inter atomic distances in order to compute the pharmacophore kernel (see (Mahe et al., 2006)). */ void threeDtransform(int nBins, double distMin, double distMax); /** computes the Euclidian distance between the XYZ coordinates of two atoms. */ void minMaxDistances(double *distMin, double* distMax); //@} ///@name Kernels and Gram matrices related functions //@{ /** calculates the gram matrix of similarity using the marginalized graph kernel for all molecules in the MoleculeSet using a specified graph, atom and bond kernel. ff aReportFilename is specified a report is saved in that file. */ void gramCompute( double aPq, double(*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), int, int ), double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), int aParameter = 1000, string aReportFileName = "", int nbThreadsWanted = 1, bool silentMode = false, bool filterTotters = false); /** calculates the gram matrix of similarity using the marginalized graph kernel for all molecules in the MoleculeSet using a specified graph, atom and bond kernel. ff aReportFilename is specified a report is saved in that file. */ void gramCompute( MoleculeSet* anotherSet, double aPq, double(*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), int, int ), double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), int aParameter = 1000, string aReportFileName = "", int nbThreadsWanted = 1, bool silentMode = false, bool filterTotters = false); /** calculates the gram matrix of similarity using the marginalized graph kernel for all molecules in the MoleculeSet using a specified graph, atom and bond kernel. ff aReportFilename is specified a report is saved in that file. */ void gramCompute( double aPq, double(*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), int, int ), double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), int parameter1 = 1000, int parameter2 = 1, string aReportFileName = "", int nbThreadsWanted = 1, bool silentMode = false, bool filterTotters = false); /** "TRUE" gramCompute() function, i.e., the one called by EVERY OTHER GRAMCOMPUTE FUNCTION. */ void gramCompute( MoleculeSet* anotherSet, double aPq, double(*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), int, int ), double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), int parameter1 = 1000, int parameter2 = 1, string aReportFileName = "", int nbThreadsWanted = 1, bool silentMode = false, bool filterTotters = false); /** 3D kernel computation. NOTE : AtomKernel/BondKernel prototyes different from gramCompute. */ void gramCompute3D( double(*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(float,float,float), float), double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(float, float, float), float edgeKernelparameter, bool silentMode ); /** "TRUE" gramCompute3D function, i.e., the one called by EVERY OTHER GRAMCOMPUTE3D FUNCTION. */ void gramCompute3D( MoleculeSet* anotherSet, double(*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(float,float,float), float), double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(float, float, float), float edgeKernelparameter, bool silentMode ); /** computes all kernel values between a molecule aMol and all compounds in the set. */ void kernelCompute( Molecule* aMol, double(*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), int, int ), double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), vector* resultsRaw, vector* resultsNormal, int convergenceCondition = 1000, int parameter2 = 1, bool silentMode = false ); /** erases the Gram Matrix and sets the gramCalculated flag to false. */ void resetGramMatrix(); /** resets all calculated selfkernels for the molecules in the dataset. */ void resetSelfKernels(); /** initializes every element gram matrix to the given value. */ void initializeGram( double value ); /** initializes the self kernels to the given value. */ void initializeSelfKernel( double value ); /** normalizes the gram matrix, i.e. compute gramNormal. */ void normalizeGram(); /** normalizes the gram matrix, i.e. compute gramNormal. WARNING: NORMALIZATION BASED ON THE RAW GRAM MATRIX (instead of self-kernels) */ void normalizeGram_raw(); /** normalizes the Gram matrix according to the Tanimoto kernel definition (Ralaivola et al. 2005). */ void normalizeTanimoto(); /** normalizes the Gram matrix according to the Tanimoto kernel definition, BASED ON THE RAW GRAM MATRIX */ void normalizeTanimoto_raw(); /** normalizes the Gram matrix according to the 'min-max' Tanimoto kernel definition (Ralaivola et al. 2005). */ void normalizeTanimotoMinMax(); /** adds a value to a Gram matrix entry. */ void addToGram( int row, int col, double value ); /** adds a value to a Gram normal matrix entry. */ void addToGramNormal( int row, int col, double value ); /** substracts a value to a Grammatrix entry. */ void substractToGram( int row, int col, double value ); /** returns a Gram matrix entry. */ double getGramValue( int row, int col ); //@} ///@name Output functions //@{ /** writes a file with the biological activity of molecules in a format compatible with GIST. */ void writeActivityFile( string aFilename, bool addActivityExtension = true, string activityDescriptor = ACTIVITY ); /** writes the gram matrix. */ void writeGramMatrix( string aFileName, bool normal = false, bool self = false, bool silentMode = false ); /** writes all self kernel values in a file. */ void writeSelfKernelList( string aFilename, bool silentMode = false ); /** writes a MDL structure data (SD) file with the molecules in the moleculeSet setting selectedOnly to true outputs only the selectwed compounds. a pointer to a vector of strings containing the molecule names for ordered output can be specified. */ void writeSD( string aFileName, bool selectedOnly = false ); /** writes a MDL structure data (SD) file with the molecules in the moleculeSet matching the names given as argument. */ void writeSubsetSD( string aFileName, vector* anOrder ); /** writes a KCF file with the molecules in the moleculeSet matching the names given as argument. */ void writeSubsetKCF( string aFileName, vector* anOrder ); /** writes a KCF file with the whole moleculeSet. */ void writeKCF( string aFileName, bool selectedOnly = false ); /** returns a string description of the MoleculeSet. */ string toString( bool selectedOnly = false ); /** returns a short string description of the MoleculeSet (number of molecules in the set and short description of each molecule). */ string toStringShort(); /** returns a long string description of the MoleculeSet. */ string toStringLong(); /** writes a description of the moleculeSet to cout. */ void describe( bool selectedOnly = false ); /** writes a short description of the moleculeSet to cout. */ void describeShort(); /** writes a long description of the moleculeSet to cout. */ void describeLong(); /** writes a ';' separated file containing all descriptors for all molecules. */ void writeDescriptors( string aFileName, bool selectedOnly = false ) throw( CError ) ; /** writes a mol file for each molecule in the set to aDirName. if selectedOnly == true then only selected molecules are written. */ long writeMolToDir( string aDirName, bool selectedOnly = false ); /** writes a dot file for all molecule in the set to aDirectory. */ long writeDotsToDir( string aDirectory, bool selectedOnly = false, bool perretLabels = false ); //@} // DEPRECATED FUNCTIONS : //void addMolecule( Molecule* aMolecule, bool updateGram = true ); /** removes the last molecule from the set */ //void pop_back(); //void normalizeGram_self(); //void normalizeGram_test(); //void setSelfGram(); protected: // Protected attributes /** comparison set of the molecule set. In test set mode, the comparison set is another set of compounds. In self set mode, the comparison set is the molecule set itself. */ MoleculeSet* comparisonSet; /** Gram matrix. */ vector< vector >* gram; /** normal Gram matrix. */ vector< vector >* gramNormal; /** contains true if the gram matrix was evaluated for the current MoleculeSet, false otherwise. note that setKashimaKernelProb sets this flag to false. {Addition of a new molecule to the set when the flag is set to true induces the calculation of a new line to the gram matrix DEPRECATED} */ bool gramCalculated; /** kashima Stop probability for the moleculeSet used to set kashimaProb in added molecules. */ double pq; /** convergence condition used in the calculation of the Kashima Kernel. */ int convergenceCondition; /** subset start. */ int subsetStart; /** subset size. */ int subsetSize; /** stores if the activity of the molecules in the set was specified. */ bool activitySet; /** sets the type and name of descriptor to be used when sorting molecules. if reverse == true the sorting will be in descending order. */ void setSortDescriptor( string aName, int aType, bool reverse = false ) throw( CError ); /** returns the name of the sorting descriptor. */ string getSortDescriptorName(); /** vector of molecules */ //vector molecules; }; #endif Rchemcpp/src/chemcpp/src/node.h0000644000175000017500000000714612116123101017032 0ustar klambaueklambaue/**************************************************************************************** node.h ---------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef NODE_H #define NODE_H #include #include /** @author Jean-Luc Perret */ class Node : public DataContainer{ public: ///@name Node construction functions //@{ /** class constructor. */ Node(); /** class constructor. */ Node( const DataContainer& aDataContainer ); /** class destructor. */ ~Node(); //@} ///@name Accessor functions //@{ /** returns the Label of the node. */ string getLabel(){ return label; } /** sets the Label of the node. */ void setLabel( string aString ); /** sets the Start probability of the atom to aPs (used in the Kashima Kernel). */ void setKashimaPS(double aPs){ ps = aPs; flagHasPs = true; } /** returns the Start probability of the atom (used in the Kashima Kernel). */ double getKashimaPS( bool silentError = false ); //{ return ( getFloatDescriptor(PS)->getValue( silentError ) ); } /** sets the Stop probability of the atom to aPq (used in the Kashima Kernel). */ void setKashimaPQ( double aPq ){ pq = aPq; flagHasPq = true; } /** returns the Stop probability of the atom ( used in the Kashima Kernel ). */ double getKashimaPQ( bool silentError = false ); /** helper function for the calculation of the Kashima Kernel. */ int getRPosition(){ return( rPosition ); } /** helper function for the calculation of the Kashima Kernel. */ void setRPosition( int aR ){ rPosition = aR; } //@} ///@name Output functions //@{ /** gives a short (string) description of the node. */ string toStringShort(); //@} // ****************************** // // **** DEPRECATED FUNCTIONS **** // // ****************************** // //void resetFlagHasPQ(); //void setFlagHasPQ( bool ); //bool hasPQ(){ return flagHasPq; } protected: /** Label of the node. */ string label; /** stop probability for random walks (graph kernel). */ double pq; /** flag indicating wether the stop probability variable is initialized or not. */ bool flagHasPq; /** start probability for random walks (graph kernel). */ double ps; /** flag indicating wether the start probability variable is initialized or not. */ bool flagHasPs; /** helper variable for the calculation of the Kashima Kernel. */ int rPosition; private: }; #endif Rchemcpp/src/chemcpp/src/atom.h0000644000175000017500000004064512116123101017046 0ustar klambaueklambaue/**************************************************************************************** atom.h ---------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef ATOM_H #define ATOM_H #include #include #include #include #include #include #include #include #include using namespace std; /** Atom class @author Dr Jean-Luc Perret ("luc@kuicr.kyoto-u.ac.jp"), Kyoto University, Japan @version 0.3 @date 17 Jan 2004 CLASS NAME: Atom FOR: SNSF SPONSORED PROJECT PURPOSE: This class implements the notion of atom. Atoms have - a map of bonds - properties, 2 kinds: - descriptors - kind descriptors - Some descriptors are unique to an atom instance and are called descriptors. - Some descriptors are common to a chemical element and are called kind descriptors. These descriptors are implemented using the DataContainer class which takes care of memory allocation for these descriptors. Therefore descriptors can be added and removed at runtime using the DataContainer functions. Some atom descriptors need fast access and are thus directly implemented in the atom class using native data types (int or float). These is the case for id and should be done for ps and pq. Atoms should always be created using the copy operator using the atoms stored in the global elements vector as template (\#include \). for example instanciating a new hydrogen atom should be done in this way: Atom* myatom = new Atom( elements["H"] ); */ class Atom : public Node { friend class Elements; friend class Bond; friend class Molecule; /** \example atom_example.cpp */ public : ///@name Atom construction functions //@{ /** class constructor. */ Atom(); /** class constructor. */ Atom( const Atom& anAtom ); /** class constructor. */ Atom( string aLabel ); /** class destructor. */ ~Atom(); /** assignment operator. */ Atom& operator=( const Atom& anAtom ); /** adds a bond to the bond hash. */ Bond* addBond( Bond *aBond ) throw ( CError ); //@} ///@name Accessor functions //@{ /** returns the atom unique Id. */ int getId() const { return( id ); } /** returns the atom unique Id as a string. */ string getIdString(); /** returns the atom Id in the molecule. */ int getIdInMolecule() const { return( idInMolecule ); } /** sets the Id of this atom in the molecule. */ void setIdInMolecule( int anId ); /** returns the element type. */ int getType() const { return( type ); } /** returns the value of the name descriptor in case atoms where names (used for reading the Mutag dataset). */ string getName(){ return( getStringDescriptor( "name" )->getValue() ); } /** returns the value of the symbol descriptor in case atoms where names (used for reading the Mutag dataset). */ string getSymbol(){ return( getStringDescriptor( "Symbol" )->getValue() ); } /** returns the element symbol. In case of mol files it is the same as getSymbol() but in case of KCF files, getSymbol() returns the Kegg atom type while getElementSymbol() returns the element atom type. */ string getElementSymbol(){ return( getStringDescriptor( "ElementSymbol" )->getValue() ); } /** sets the value of the ElementSymbol descriptor. */ void setElementSymbol( string anElement ); /** returns the atomic number of the atom (ex: hydrogens should return 1). */ int getAN() const { return ( an ); } /** returns true if the atom is 'generic'. */ bool isGeneric(){ return genericAtomType; } /** returns true if the atom is a non terminal Carbon (a terminal carbon is a carbon atom attached to only one non hydrogen atom). */ bool isCSkeleton(); /** returns true if the atom has valid coordinates. */ bool hasCoordinates(); /** returns the x coordinate of the atom. */ float getX(){ return( x ); } /** returns the y coordinate of the atom. */ float getY(){ return( y ); } /** returns the z coordinate of the atom. */ float getZ(){ return( z ); } /** sets the x, y, and z coordinates of the atom. */ void setCoordinates( float aX, float aY, float aZ ); /** returns the number of Bonds. */ int numBonds(){ return ( bonds.size() ); } /** returns the number of hidden Bonds. */ int numHiddenBonds(){ return ( hiddenBonds.size() ); } /** returns the number of aromatic Bonds for this atom. */ int getNumAromaticBonds(); /** returns the number of Bonds (synonymous to numBonds). */ int degree(){ return ( bonds.size() ); } /** returns true if the node was already visited. */ bool wasVisited(){ return visited;} /** sets the node as visited. */ void setVisited(){ visited = true;} /** sets the node as unvisited. */ void unsetVisited(){ visited = false;} /** returns the Morgan index of order anOrder. Emits an error if it was not calculated before. */ int getMorganIndex( int order ); /** returns the value of the Morgan index for which the diversity of Morgan indices is maximum in the current molecule. set by setUniqueMorganIndex(int). */ int getUniqueMorganIndex( bool silentError = false ) throw( CError ); /** set the Morgan label of the atom. */ void setMorganLabel(string aLabel); /** sets the Morgan label (used in the Molecule::morganKernel function) to the concatenation of the atomSymbol and the Morgan index of iteration anOrder. */ void setMorganLabel( int anOrder ); /** returns the value of the Morgan label (set by setMorganLabel()). */ string getMorganLabel( bool silentError = false ) throw( CError ); /** sets the Perret label of the atom (if a carbon has more than 2 aromatic bonds it is renamed CJ). */ void setPerretLabel() throw( CError ); /** returns the value of the perret label (set by setPerretLabel(), called everytime a molecule is modified). */ string getPerretLabel( bool silentError = false ) throw( CError ); /** adds a ring to the set of rings the atom is member of. */ void addRing( Ring* aRing ){ if( !hasRing( aRing ) ){ rings.push_back( aRing ); } } /** returns true if the atom is member of the given ring. */ bool hasRing( Ring* aRing ); /** returns true if the atom is member of any ring. */ bool hasRing(){ if( numRings() > 0 ){ return( true ); }else{ return( false );};} /** returns the number of rings the atom is member of. */ int numRings(){ return( rings.size() ); } /** returns the transition probability to anAtom. returns 0 if no atoms exist with anAtom. */ double getKashimaPT( Atom* anAtom ); /** returns the BFS vector. */ vector* getBFSVector(){ return( &BFSVector ); } /** returns the BFS Bond vector. */ vector* getBFSBondVector(){ return( &BFSBondVector ); } /** returns the BFSVector size. */ int getBFSVectorSize(){ return( BFSVector.size() ); } /** resets the BFSVector. */ void resetBFSVector(){ BFSVector.clear(); BFSBondVector.clear(); } //@} ///@name Atom manipulation functions //@{ /** erases the atoms coordinates (set flagHasCoordinates to false). */ void eraseCoordinates(); /** hides a bond (but not reverse bond). */ void hideBond( map< Atom*, Bond* >::iterator aBondI ); /** hides a bond (but not reverse bond). */ void hideBond( Bond* aBond ); /** hides a bond (but not reverse bond). */ Bond* hideBond( Atom* aTarget ); /** hides all bonds whose target is an hydrogen. */ int hideHydrogenBonds(); /** hides all bonds to and from this Atom. */ void hideAllToFromBonds(); /** hides all bonds to and from to aTarget Atom and returns a poiter to the 'to Bond' of aTarget atom. */ Bond* hideToFromBonds( Atom* aTarget ); /** hides the first bond in bonds (both to and from bonds) and returns a pointer to the 'to Bond'. */ Bond* hideToFromFirstBond(); /** restores all hidden bonds (for example, bonds to hydrogens). */ int restoreHiddenBonds(); /** restores hidden bond aBond. */ void restoreHiddenBond( Bond* aBond ) throw( CError ); /** restores hidden bond to atom aTarget. Throws a CError if no hidden bond to aTarget is found. */ void restoreHiddenBond( Atom* aTarget ) throw( CError ); /** deletes all bonds of that atom (called from the destructor of molecule). */ void deleteBonds(); /** deletes all hidden bonds of that atom (called from the destructor of molecule). */ void deleteHiddenBonds(); /** returns true if a bond (of any type) exists with the target atom. */ bool bondExists( Bond* aBond ); /** returns the sum of the bond types. */ long bondSum(); /** unsets all bond flags. */ void unsetBondFlags(); /** unsets all bond flags original. */ void unsetBondFlagsOriginal(); /** sets the unique Morgan index of this atom to the value of anOrder iteration of the Morgan index calculation. */ void setUniqueMorganIndex( int anOrder ); /** returns the sum of the Morgan indices of order anOrder of all neighboors. */ int getSumOfNeighboorMorganIndex( int anOrder ); /** resets the Morgan index map. call this function whenever the molecule is modified! */ void resetMorganIndex(); /** set the value of the partial charge of the instance to aValue */ void setPartialCharge(double aValue); /** include the sign of the partial charge in the Morgan label of the instance. If partial charge > threshold: sign = + ; otherwise, sign = - */ void setMorganChargeLabel(double threshold); /** returns the label of the bond connecting this atom with otherAtom. returns NOBOND if there are no bonds connecting both atoms. */ Bond* getBondWithTarget( Atom* otherAtom ) throw( CError ); /** returns a pointer to the next unvisited Node. returns NULL pointer if all atoms were visited. */ Atom* nextUnvisitedAtom() throw( CError ); /** BFS procedure for smallest ring detection. returns a pointer to the smallest ring containing the atom or NULL if the atom is not member of any ring. note you should check ring membership with hasRing(). */ Ring* getRingBFS( vector* toVisit, vector* toVisitBond ) throw( CError ); /** NO DOCUMENTATION */ void pushBFSVector( vector* aPath, vector* aBondPath ); //@} ///@name Iterators //@{ /** returns a pointer to the bonds map. */ map& getBonds(){ return( bonds ); } /** returns an iterator to the first Bond. */ map::iterator beginBond(){ return( bonds.begin() ); } /** returns a pointer to the last Bond. */ map::iterator endBond(){ return( bonds.end() ); } /** returns an iterator to the first ring the atom is member of. */ vector::iterator beginRing(){ return( rings.begin() ); } /** returns an iterator to the last ring the atom is member of. */ vector::iterator endRing(){ return( rings.end() ); } /** NOT DOCUMENTED */ map::iterator getBondIteratorWithTarget( Atom* otherAtom ); //@} ///@name Output functions //@{ /** returns a string describing the atom (atomic symbol + unique Id + memory location + number of bonds). */ string toString(); /** returns a short string describing the atom (atomic symbol + unique Id). */ string toStringShort(); /** prints a description of the atom to cout (describeShort + descriptors, but not the kind descriptors). */ void describe() throw( CError ); /** prints a short description of the atom to cout (atomic symbol + unique Id). */ void describeShort(); /** NO DOCUMENTATION */ string toStringBFSVector(); //@} /** stream operator for Atom. */ friend ostream& operator<<(ostream& os, const Atom& anAtom); /** NO DOCUMENTATION. */ static void getVectorIntersect( vector* v1, vector* v2, vector* result); // ******************************* // // **** DEPRECATED FUNCTIONS **** // // ****************************** // //Atom(const Atom& anAtom); //Atom(const Atom* anAtom) : DataContainer( (DataContainer&) anAtom ); /** returns an iterator to the first Hiden Bond */ //map::iterator beginHiddenBond(){ return( hiddenBonds.begin() ); } /** returns an iterator to the last Hiden Bond */ //map::iterator endHiddenBond(){ return( hiddenBonds.end() ) /** same as getBondWithTarget but among hidden bonds */ //Bond* getHiddenBondWithTarget( Atom* otherAtom ); /** same as getBondWithTarget but among saved bonds */ //Bond* getSavedBondWithTarget( Atom* otherAtom ); /** returns the transition probability to anAtom, but looks among hidden bonds Returns 0 if no atoms exist with anAtom */ //float getHiddenKashimaPT( Atom* anAtom ); /** returns the transition probability to anAtom, but looks among saved bonds Returns 0 if no atoms exist with anAtom */ //float getSavedKashimaPT( Atom* anAtom ); //static void getVectorUnion( vector* v1, vector* v2, vector* result, vector* intersect ); private : /** atom unique Id. */ int id; /** atom unique Id in the molecule it belongs. */ int idInMolecule; /** atom type: start from Hydrogen at 0, incremented at the instanciation of elements. used by MoleculeUtils::atomKernelExternalMatrix(). */ int type; /** atomic number. to enhance retrival speed this property is hard coded rather than using the DataContainer mechanism. */ int an; /** atom x coodinate. */ float x; /** atom y coodinate. */ float y; /** atom z coodinate. */ float z; /** flag indicating if the atom has valid coordinates. */ bool flagHasCoordinates; /** counter of the number of atoms which have been instanciated. */ static int counter; /** hash of bonds indexed by atom. */ map bonds; /** vector of hidden bonds indexed by atom. */ map hiddenBonds; /** vector of saved bonds indexed by atom. filled by calling saveAllBonds(). */ //map savedBonds; /** smallest rings of which this atom is member (set by Molecule::detectSSSR()). */ vector rings; /** flag to indicate if the atom was visited. Used in DFS graph algorithm. */ bool visited; /** indicates if this atom is part of the elements or if it is generic (created with constructor: Atom::Atom(string)). */ bool genericAtomType; protected : // Protected methods /** this function is used when instanciating the atom for setting the atomic number. */ void setAN( int a ) { an = a; } /** resets the atom counter. Only the elements should call this function. */ static void resetCounter() { counter = 0; } /** sets the atom type. */ void setType( int aType ) { type = aType; } /** clear hidden bonds. */ //void clearHiddenBonds(); /** hide all bonds leaving this atom. */ //void hiddeAllBonds(); /** save all bonds leaving this atom. */ //void saveAllBonds(); /** map containing the morgan indices. */ map< int, int > morganIndex; /** Morgan label of the atom set by the Atom::setMorganLabel() function. */ string morganLabel; /** Perret label of the atom.set by the Atom::setPerretLabel() function. */ string perretLabel; /** Morgan index value for the iteration with the biggest diversity for a molecule (set by Molecule::setUniqueMorganIndices()). */ int uniqueMorganIndex; /** this vector is used to store paths in a BFS search (stores atoms). */ vector BFSVector; /** this vector is used to store paths in a BFS search (stores bonds). */ vector BFSBondVector; /** partial charge of the atom. */ double partialCharge; }; #endif Rchemcpp/src/chemcpp/src/moleculeutils.h0000644000175000017500000002403612116123101020770 0ustar klambaueklambaue /**************************************************************************************** moleculeutils.h ------------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef MOLECULEUTILS_H #define MOLECULEUTILS_H //#include #include #include #include #include //#include //#include #include //#include #define FINDENTRY 0 #define READINGNODES 1 #define READINGEDGES 2 #define READINGUNKNOWN 3 using namespace std; /**static functions to be used to process molecules *@author Jean-Luc Perret */ class MoleculeUtils { public: ///@name Input/Ouput utility functions //@{ /** reads the ctab block of a stream (connection table), and adds atoms and bonds to aMolecule. if genericAtomTypeFlag is true then the atoms are not taken from the periodic table of elements, but are created from the label provided. */ static void readMDLCtabBlock( Molecule& aMolecule, ifstream& inFile, bool genericAtomTypeFlag = false ) throw( CError ); /** reads the header block of a stream, and sets molecule name to the comment content or to aName if specified. */ static void readMDLHeaderBlock( Molecule& aMolecule, ifstream& inFile, string aName = "COMMENT" ) throw( CError ); /** reads the non structural data block of a compound in a MDL SDfile. WARNING only data entry with a structure aLabel will be considered. entries missing the aLabel entry will be ignored. */ static void readMDLNSDBlock( Molecule& aMolecule, ifstream& inFile ) throw( CError ); /** skips all lines until next entry. */ static void skipMDLEntry( Molecule& aMolecule, ifstream& inFile ) throw( CError ); /** writes the header block to a stream, for aMolecule. */ static void writeMDLHeaderBlock( Molecule& aMolecule, ofstream& outFile ); /** writes the ctab block to a stream (connection table), for aMolecule. */ static void writeMDLCtabBlock( Molecule& aMolecule, ofstream& outFile ); /** writes the non structural data block to a stream, for aMolecule. this function writes all stringDescriptors, floatDescriptors and IntDescriptors of a molecule in appropriate MDL SD format. */ static void writeMDLNSDBlock( Molecule& aMolecule, ofstream& outFile ); /** writes a molecule under the KCF format in the outFile stream */ static void writeKCF( Molecule& aMolecule, ofstream& outFile ); /** writes the non structural data block to a stream, for aMolecule. this function writes all stringDescriptors, floatDescriptors and IntDescriptors of a molecule in appropriate KEGG KCF format. */ static void writeKCFNSDBlock( Molecule& aMolecule, ofstream& outFile ); /** reads one molecule from a kcf file. */ static bool readKCFMolecule( KCFMolecule& aMolecule, ifstream& inFile ) throw( CError ); /** writes a graph in Graphviz's dot format. */ static void writeDOTGraph( Molecule& aMolecule, ofstream& outFile, bool perretabels = false ); //@} ///@name Graph kernels related functions //@{ /** computes the marginalized graph kernel between two molecules. takes two molecules, a function pointer to the atom and the bond kernels and the convergence condition as parameter. WARNING: random walk parameters must first be set. */ static double moleculeKernel( Molecule* mol1, Molecule* mol2, double (*pt2AtomKernel)(Atom*, Atom*) = &atomKernelSymbol, double (*pt2BondKernel)(Bond*, Bond*) = &bondKernelType, int convergenceCondition = 1000, int parameter2 = 1 ); /** returns the marginalized graph kernel value between two molecules using the product graph approach and the sum of matrix powers. The kernel computes the sum of the probabilities of finite length paths, until order maxPower. */ static double powerKernelUntilN( Molecule* mol1, Molecule* mol2, double (*pt2AtomKernel)( Atom*, Atom* ), double (*pt2BondKernel)(Bond*, Bond*), int maxPower, int minLength = 1 ) throw( CError ); /** returns the marginalized graph kernel value between two molecules using the product graph approach and the sum of matrix powers. The kernel computes the probabilities of finite length of length anOrder. */ static double powerKernelOrderN( Molecule* mol1, Molecule* mol2, double (*pt2AtomKernel)( Atom*, Atom* ), double (*pt2BondKernel)(Bond*, Bond*), int anOrder, int parameter2 = 1 ) throw( CError ); /** returns the marginalized graph kernel value between two molecules using the product graph approach and the sum of matrix powers. The kernel computes the sum of the probabilities of infinite length paths, until the kernel value does not change by more than 1/converge. */ static double powerKernelConverge( Molecule* mol1, Molecule* mol2, double (*pt2AtomKernel)( Atom*, Atom* ), double (*pt2BondKernel)(Bond*, Bond*), int converge, int parameter2 = 1 ) throw( CError ); /** computes the "rlk" term involved in the marginalized kernel computation. */ static vector< vector >* rlk( vector< vector >* r, vector< vector >* rwork, vector< vector >* rstart, Molecule* mol1, Molecule* mol2, int convergenceCondition, int parameter2, double (*pt2AtomKernel)(Atom*, Atom*), double (*pt2BondKernel)(Bond*, Bond*), int depth ); /** returns true if the marginalized kernel computation has converged according to a convergence condition. */ static bool converge( vector< vector >* r1, vector< vector >* r2, Molecule* mol1, Molecule* mol2, int convergenceCondition ); /** atom kernel based on atomic number. */ static double atomKernelSymbol( Atom* a1, Atom* a2 ); /** atom kernel based on Morgan label. WARNING: labels must first be set. */ static double atomKernelMorganLabel( Atom* a1, Atom* a2 ); /** atom kernel based on Perret label. */ static double atomKernelPerretLabel( Atom* a1, Atom* a2 ); /** atom kernel based on Perret label and external gram matrix. */ static double atomKernelPerretLabelExternalMatrix( Atom* a1, Atom* a2 ); /** atom kernel based on an external gram matrix. WARNING: the gram matrix must first be loaded using Elements::loadGramAtoms(). */ static double atomKernelExternalMatrix( Atom* a1, Atom* a2 ); /** generic atom kernel label. */ static double atomKernelLabel( Atom* a1, Atom* a2 ); /** bond kernel. returns 1 if two bonds have the same type, 0 otherwise. */ static double bondKernelType( Bond* b1, Bond* b2 ); /** bond kernel. returns 1 if two bonds have the same perretLabel, 0 otherwise. */ static double bondKernelPerretLabelStrict( Bond* b1, Bond* b2 ); /** bond kernel. returns 1 if two bonds have the same perretLabel. returns 0 if one bond is a cycle while the other is not. returns a weight is both bonds are cycles or non cycles. */ static double bondKernelPerretLabel( Bond* b1, Bond* b2 ); /** bond kernel. Single bonds are rotable, others are not. Returns 1 if two bonds. are rotable or two bonds are non rotable, 0 if one of the bonds. is rotable and the other is not. */ static double bondKernelRotable( Bond* b1, Bond* b2 ); //@} ///@name 3D kernels related functions //@{ /** (3D) generic pharmacophore kernel between two molecules. defined for a choice of atom an edge kernels, and corresponding parameters. */ static double threeDkernel( Molecule* mol1, Molecule* mol2, double (*pt2AtomKernel)( Atom*, Atom*), double (*pt2BondKernel)(float, float, float), float edgeKernelParameter ); /** (3D) RBF edge kernel. param = bandwith. */ static double threeDedgeKernelRBF( float dist1, float dist2 , float param ); /** (3D) triangular edge kernel. param = cut-off. */ static double threeDedgeKernelTriangle( float dist1, float dist2 , float param ); //@} ///@name Miscallenous utility functions //@{ /** returns a string with the ring membership of aBond. */ static string getRingString( Bond* aBond ) throw( CError ); /** NO DOCUMENTATION */ static void describeMap( map* aMap ); /** helper function for Molecule::detectSSSR. returns in result a vector containing atom pointer for atoms present in full but not in exclude. */ static void substractSet( vector* full, vector* exclude, vector* result); /** returns true if anAtom is present into the atomVector vector. */ static bool atomVectorHas( vector* atomVector, Atom* anAtom ); /** helper function for Molecule::detextSSSR. the result vector is the concatenation of v1 and v2 without duplicates. */ static void mergeSet( vector* v1, vector* v2, vector* result); /** merges two bonds sets. */ static void mergeBondSet( vector* v1, vector* v2, vector* result); /** NO DOCUMENTATION */ static void selectRingMemberBonds( vector* v1, vector* a, vector* v2 ); //@} //MoleculeUtils(); //~MoleculeUtils(); }; #endif Rchemcpp/src/chemcpp/src/ring.h0000644000175000017500000000672312116123101017044 0ustar klambaueklambaue/**************************************************************************************** ring.h ---------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef RING_H #define RING_H #include #include using namespace std; class Atom; class Bond; /** This class implements the notion of ring. It contains a list of atoms. @author Jean-Luc Perret */ class Ring : public std::vector{ public: ///@name Ring construction functions //@{ /** class constructor. */ Ring(); /** class constructor. */ Ring( vector*, vector* ); /** class desctructor. */ ~Ring(); /** adds a bond to the ring. this function should only be called during ring creation from a file. Not for ring detection. */ void addBond( Bond* aBond, bool silentError = false ) throw ( CError ); /** adds an atom to the ring. */ void addAtom( Atom* anAtom, bool silentError = false ) throw ( CError ); //@} ///@name Accessor functions //@{ /** returns the id of the ring. */ int getID(){ return( id ); } /** sest the id of the ring. */ void setID( int a ){ id = a; } /** returns the vector of bonds the ring is made of. */ vector* getBonds(){ return( &bonds ); } //@} ///@name Ring manipulation functions //@{ /** returns true if the ring contains anAtom. */ bool hasAtom( Atom* anAtom ); /** returns true if the ring contains aBond. */ bool hasBond( Bond* aBond ); /** comparison operator. */ bool equals( Ring* anotherRing ); //@} ///@name Output functions //@{ /** returns a string shortly describing of the ring. */ string toStringShort(); /** returns a string describing of the ring. */ string toString(); /** gives a short description of the ring. */ void describeShort(); /** gives a description of the ring. */ void describe(); //@} // ****************************** // // **** DEPRECATED FUNCTIONS **** // // ****************************** // //Ring( vector* ); //vector::iterator beginBond(){ return( bonds.begin() ) ; } //vector::iterator endBond(){ return( bonds.end() ) ; } protected: /** vector of bonds the ring is made of. */ vector bonds; /** stores a unique id of the ring in the molecule. */ int id; }; #endif Rchemcpp/src/chemcpp/src/descriptor.cpp0000644000175000017500000000216112116123101020606 0ustar klambaueklambaue/*************************************************************************** descriptor.cpp - description ------------------- begin : Wed Dec 10 2003 copyright : (C) 2003 by Jean-Luc Perret email : luc@kuicr.kyoto-u.ac.jp ***************************************************************************/ /*************************************************************************** * * * This program is free software; you can redistribute it and/or modify * * it under the terms of the GNU General Public License as published by * * the Free Software Foundation; either version 2 of the License, or * * (at your option) any later version. * * * ***************************************************************************/ #include "descriptor.h" // Descriptor is a template class, so all functions are defined in the header file Rchemcpp/src/chemcpp/src/molecule.h0000644000175000017500000007736112116123101017720 0ustar klambaueklambaue/**************************************************************************************** molecule.h -------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef MOLECULE_H #define MOLECULE_H #include #include #include #include #include #include #include #include #include #include //#include extern Elements elements; /** Molecule class which can be compared using Graph Kernels @author Dr Jean-Luc Perret (luc@kuicr.kyoto-u.ac.jp), Kyoto University, Japan @version 0.3 @date 17 Jan 2004 CLASS NAME: Molecule FOR: SNSF SPONSORED PROJECT PURPOSE: This class implements the notion of molecule. Molecules have - a vector of atoms - properties Properties are implemented by deriving the Molecule from the DataContainer class which takes care of memory allocation for these descriptors. Therefore descriptors can be added and removed at runtime using the DataContainer functions. The Molecule class can be seen as a graph containing a list of vertex (Atom), each vertex containing a list of edges (Bond) to the neighbour vertex (Atom). Edges are directed and have a label. Chemical bonds are not directed so bonds are always double, one in each direction. So a Molecule can be seen as a directed labeled graph with bidirectional edges. ... many other changes */ class Molecule : public DataContainer { /** \example molecule_example.cpp */ public: ///@name Molecule construction functions //@{ /** class constructor. */ Molecule(); /** class destructor. deletes all atoms in the molecule. */ virtual ~Molecule(); /** copy constructor. */ Molecule( Molecule& aMolecule , bool bool_resetMorganIndex = true ); /** product graph constructor. this constructor constructs a fusion molecule from two molecules using a product graph operation. two atoms which are connected in both molecules have an edge connecting them in the fused graph. */ Molecule( Molecule& molecule1, Molecule& molecule2, double (*pt2AtomKernel)( Atom*, Atom* ), double (*pt2BondKernel)( Bond*, Bond* ) ); /** "3D" constructor. Note : BondKernel prototype is different of those used on the "product graph constructor". */ Molecule( Molecule& m1, Molecule& m2, double (*pt2AtomKernel)( Atom*, Atom*), double (*pt2BondKernel)( float, float, float ), float edgeKernelParameter ); /** assignment operator. */ Molecule& operator=( const Molecule& aMolecule ); /** adds an Atom with aSymbol chemical symbol to the molecule. aSymbol can be in any case. The first letter of aSymbol will be automatically changed to capital and the second letter to normal caps. Returns a pointer to the created Atom. WARNING this atom will be deleted when the molecule is deleted so don't use this pointer! if resetSSSR = true, then this action invalidates previous sssr computations. */ virtual Atom* addAtom( string aSymbol, bool resetSSSR = true ) throw( CError ); /** adds an Atom to the atoms vector. Returns a pointer to the created Atom. WARNING this atom will be deleted when the molecule is deleted so don't use this pointer! if resetSSSR = true, then this action invalidates previous sssr computations. */ Atom* addAtom( Atom*, bool resetSSSR = true, bool resetMorganIndex = true ) throw( CError); /** links two atoms poited by aSource and aTarget using aBondLabel bond type. If the two atoms are not in the atoms vector they are first added to the molecule. returns a pointer to the forward bond just created (the return bond can be retrived with forwardBond->getReverse()). */ Bond* linkAtoms( Atom* aSource, Atom* aTarget, int aBondLabel, int aBondStereo = 0, int aBondNotUsed = 0, int aBondTopology = 0, int aBondReactionCenter = 0, bool resetSSSR = true ); /** links two atoms using their name. If no atom with name aSource or aTarget are found a CError error is thrown. returns a pointer to the forward bond just created (the return bond can be retrived with forwardBond->getReverse()). */ Bond* linkAtoms( string aSource, string aTarget, int aBondLabel, int aBondStereo = 0, int aBondNotUsed = 0, int aBondTopology = 0, int aBondReactionCenter = 0, bool resetSSSR = true ) throw( CError ); /** links two atoms using their position in the atom container. if firstAtom or secondAtom are not between 0 and numAtoms()-1 a CError error is thrown. returns a pointer to the forward bond just created (the return bond can be retrived with forwardBond->getReverse()). */ Bond* linkAtoms( int firstAtom, int secondAtom, int aBondLabel, int aBondStereo = 0, int aBondNotUsed = 0, int aBondTopology = 0, int aBondReactionCenter = 0, bool resetSSSR = true ) throw( CError ); /** reads the molecule description from a Mol file. MUST STILL BE COMPLETED \todo atomlist block is not taken into account \todo stext block is not taken into account \todo properties block is not taken into account \todo chirality is not taken into account (chiral flag is set by reading mol file) \todo bondTopology is not taken into account */ void readMOL( string aFilename, bool genericAtomType = false ) throw(CError); /** OLD version of readMOL(). */ void readMOLOld( string aFilename ) throw(CError); /** erases all atoms in the current molecule. returns the molecule in the state of construction by the default constructor. */ void erase(); /** erases hidden atoms of the molecule. */ void eraseHiddenAtoms(); /** erases the rings of the molecule. */ void eraseRings(); /** erases the adjacency matrix of the molecule. */ void eraseAdjacency(); /** erases the 'walks' matrix of the molecule (used to compute the pharmacophore kernel, see (Mahe, 2006)). */ void eraseWalks(); //@} ///@name Accessor functions //@{ /** returns the unique id of the molecule. */ int getId() const { return( id ); } /** returns the unique id of the molecule as a string. */ string getIdString(); /** returns the molecule name, i.e., the name descriptor value (Nb :not necessarily unique). */ string getName(){ return( getStringDescriptor( "name" )->getValue() ); } /** sets the molecule name. */ void setName( string aName ) { setStringDescriptor( "name", aName, "", "", true, true ); } /** sets the activity of the molecule (a descriptor in float native format for faster retrieval). */ void setActivity( float aNumber ){ activity = aNumber; flagActivity = true; } /** sets the activity of the molecule from a string activity. provided for convenience. */ void setActivity( string aNumber ){ activity = atof( aNumber.c_str() ); flagActivity = true; }; /** returns the activity of the molecule. */ float getActivity( bool silentMode = false ) throw( CError ); /** returns true if the activity of the molecule was initialized. */ float hasActivity(){ return( flagActivity ); } /** turns off the activity flag. */ void unsetActivity(){ flagActivity = false; } /** returns a pointer to the atom's vector. */ vector& getAtoms(){ return( atoms ); } /** returns a pointer to an atom designated by its unique Id. raises a CError exception if no atom with this Id is present in the molecule. */ Atom* getAtom(int anId) throw( CError ); Atom* getAtomByIndex(int ind); /** checks if an atom is part of the molecule. returns true if anAtom is included in the atom vector, false otherwise. */ bool atomExists(Atom* anAtom); /** returns the number of atoms of the molecule. */ int numAtoms(){ return( atoms.size() ); } /** returns the number of hidden atoms of the molecule. */ int numHiddenAtoms(){ return( hiddenAtoms.size() ); } /** returns the number of bonds of the molecule. */ int numBonds(); /** returns the number of hidden bonds of the molecule. */ int numHiddenBonds(); /** returns the sum of the bond types. */ long bondSum(); /** unsets all bonds flags. */ void unsetBondFlags(); /** unsets all bonds flags original. */ void unsetBondFlagsOriginal(); /** returns true if the molecule has chiral information. */ bool isChiral(){ return( chiral ); } /** sets selectedFlag. */ void select(){ selectedFlag = true; } /** unset selectedFlag. */ void unSelect(){ selectedFlag = false; } /** returns selectedFlag. */ bool isSelected(){ return( selectedFlag ); } /** sets the descriptor type and name used by the < operator to compare molecules. if reverse == true then the sorting will be in descending order. */ void setSortDescriptor( string aName, int aType ); /** returns the type of the descriptor used to sort molecules. */ int getSortDescriptorType(){ return( sortDescriptorType ); } /** returns the name of the descriptor used to sort molecules. */ string getSortDescriptorName(){ return( sortDescriptorName ); } /** returns the ring of the corresponding Id. */ Ring* getRingWithID( int anID, bool createIfMissing) throw( CError ); /** returns true if the molecule already has newRing in sssr. If sssr was not detected before this call, detectSSSR() is called. */ bool hasRing( Ring* newRing, bool detectingRing = false ) throw( CError ); /** returns true if the molecule has a ring. If sssr was not detected before this call, detectSSSR() is called. */ bool hasRing() throw( CError ); /** returns the number of rings present in the molecule. */ int numRings() throw( CError ){ if( !hasSSSRDetected() ){ //detectSSSR(); CError e( SSSRNOTDETECTED, "Smallest Set of Smallest Rings was not detected before calling Molecule::numRings()" ); //e.describe(); throw(e); } return( sssr.size() ); } /** returns the value of the 'hasSSRDetected' flag. */ bool hasSSSRDetected(){ return( flagHasSSSRDetected ); } /** returns the molecular weight. */ float getMW( bool silentError = false ) throw( CError ); /** returns the number of atoms in the molecule (ignoring hidden atoms). */ int getNumAtoms(){ return( atoms.size() ); } /** returns the location where the molecule is stored. */ string getLocation(){ return( location ); } /** returns the format of the original file. one of MDL2000, MDL3000, MDL2000JLP. */ int getOriginalFormat(){ return( originalFormat ); } /** sets the molecule's original format. one of MDL2000, MDL3000, MDL2000JLP. */ void setOriginalFormat( int a ){ originalFormat = a; } //@} ///@name Molecule manipulation functions //@{ /** hides all atoms (and associated edges) with intDescriptor aDescriptorName == aValue. returns the number of hidden atoms */ int hideAtomsByIntDescriptor( string aDescriptorName, int aValue, bool refreshBonds = true ); /** hides an atom (but not the bonds to this atom). call refreshBonds() after hiding all atoms. */ void hideAtom( vector< Atom* >::iterator anAtomI ); /** hides anAtom and all to / from bonds for this atom. */ void hideAtomAndToFromBonds( vector< Atom* >::iterator anAtomI ); /** hides anAtom and all to / from bonds for this atom. */ void hideAtomAndToFromBonds( Atom* anAtom ); /** moves the hydrogen atoms to the hiddenAtoms container, so that further algorithms will run without hydrogens. returns the number of hydrogens hidden. hidden hydrogens can be restored using restoreHiddenAtoms(). */ int hideHydrogens(); /** returns true if an atom is hidden. */ bool isHiddenAtom( Atom* anAtom ); /** restores the hidden atoms (for example hydrogens). */ int restoreHiddenAtoms( bool flagRestoreBonds = true ); /** restores the hidden bonds among atoms (does not restore bonds in hidden atoms). */ int restoreHiddenBonds(); /** hides all bonds to and from hidden atoms. call this function after hiding all atoms in a molecule. returns the number of edges hidden. */ int refreshBonds(); /** erases anAtom from the list. Does not delete the atom though! */ void eraseAtom( Atom* anAtom ) throw( CError ); /** deletes all bonds in all atoms. */ void deleteBonds(); /** deletes all hidden atoms. */ void deleteHiddenAtoms(); /** sets the morganLabel of atoms to the concatenation of the atomic symbol and the Morgan index of anOrder iteration. */ void setMorganLabels( int anOrder ); /** sets the perretLabel of atoms: the atomic symbol. Only carbons get an additional J if they have more than two aromatic bonds (aromatic rings joining carbons). */ void setPerretLabels(); /** sets the uniqueMorganIndex of each atom to the Morgan index having the maximum of different connectivity values for the molecule. returns the iteration of the smallest Morgan Index with the maximum number of connectivity values. */ int setUniqueMorganIndices(); /** resets the calculations of the morgan indices. CALL THIS FUNCTION EACH TIME THE MOLECULE IS MODIFIED! */ void resetMorganIndex(); /** returns the number of distinct morgan indices (of order anOrder) present in the molecule. */ int getNumberOfDistinctMorganIndices( int anOrder ); /** returns the iteration of the Morgan index at which the different connectivity values have reached a maximum. */ int getMaxMorganIteration(); /** returns the number of carbons in a connected component graph defined with an Int descriptor named aDescriptorName having value aValue. */ int getNumCarbonsOfComponent( string aDescriptorName, int aValue ); /** returns the number of nitrogens in a connected component graph defined with an Int descriptor named aDescriptorName having value aValue. */ int getNumNitrogensOfComponent( string aDescriptorName, int aValue ); /** returns the number of terminal atoms of the molecule. */ int numAtomsNonCSkeleton(); /** computes the Euclidian distance between two atoms. */ float atomicDistance(Atom* atom1, Atom* atom2); /** this function should be called everytime a molecule is created / modified. automatically called at the end of Molecule::readMol() and MoleculeUtils::readMDLCtabBlock(). */ void compute(); /** sets the activity according to the value of a descriptor and a value. If true is given as third argument (default) then descriptorName <= value are considered positive. If false is given as third argument then molecules with descriptorName >= value are considered positive. if a descriptor is missing then the molecule is left without activity. */ void binClassifyFromDescriptor( string descriptorName, float value, bool smallerOrEqual = true ); /** fills in a vector of String with the distinct labels of the atoms of the molecule. */ void atomsLabelsListing( vector* ); /** fills in vector of String with the distinct symbols of the atoms of the molecule. */ void atomsSymbolsListing( vector* ); /** fills in vector of Int with the distinct types of bonds of the molecule. */ void bondsListing( vector* ); /** transforms the molecule into a graph preventing 'totters' (see (Mahe 2004)). */ void noTottersTransform(); /** transforms the molecule into a 3D graph (see (Mahe 2006)). */ void threeDtransform(int nBins, double distMin, double distMax); /** reads partial charges of the atoms of the molecule from a ';' separated string. */ void readPartialCharges(string charges); /** introduces the sign of the partial charge into the label of the atoms of the molecule. */ void setMorganChargesLabels(double threshold); /** hides all but the biggest (in atom numbers) connected atoms in the molecule. returns the number of components removed. WARNING: does not test if two compounds have identical maximum size. in that case, it will hide all but the first one. */ int hideSalts( stringstream* out ); /** NOT DOCUMENTED */ int markFragments(); /** NOT DOCUMENTED */ void unmarkFragments(); /** NOT DOCUMENTED */ void hideAllFragmentsBut( int aFragmentNumber ); /** NOT DOCUMENTED */ void writeFragments( ofstream* outStream ); /** performs a DFS search on the graph starting at startAtom, marking atoms with markValue in descriptor intDescriptorName, and returning the number of nodes in the connected component. */ int DFS( Atom* startAtom, string intDescriptorName, int markValue ); /** detects the smallest set of smallest rings and returns the number of rings. this procedure is automatically called when a molecule is loaded from a mol or sd file through the function compute. call compute() explicitly after modifying atom or bond composition. this function also sets the ring membership of atoms and bonds. according to: Ring Perception Using Breadth-First Search, John Figueras 399 Baker s Pond Road, Orleans, Massachusetts 02653, J. Chem. Inf. Comput. Sci. 1996, 36, 986-991. */ int detectSSSR(); /** sets the flagHasSSSRDetected to false. */ void resetSSSR(){ flagHasSSSRDetected = false; } /** sets the flagHasSSSRDetected to true. */ void setHasSSSR(){ flagHasSSSRDetected = true; } /** helper function for detectSSSR. selects an optimum edge for elimination in structures without N2 nodes. */ Bond* checkEdges( Atom* ); //@} ///@name Iterators //@{ /** returns an iterator to the first Atom. */ vector::iterator beginAtom(){ return( atoms.begin() ); } /** returns an iterator to the last Atom. */ vector::iterator endAtom(){ return( atoms.end() ); } /** returns an iterator to the first Bond of an Atom. */ map::iterator beginBond( Atom* anAtom ){ return( anAtom->beginBond() ); } /** returns an iterator to the first Bond of an Atom. */ map::iterator endBond( Atom* anAtom ){ return( anAtom->endBond() ); } /** returns an iterator to the first Bond of an Atom designated by its id. */ map::iterator beginBond( int anId ){ return( getAtom( anId )->beginBond() ); } /** returns an iterator to the first Bond of an Atom designated by its id. */ map::iterator endBond( int anId ){ return( getAtom( anId )->endBond() ); } /** NOT DOCUMENTED */ map::iterator beginComponentSizes(){ return( componentSizes.begin() ); } /** NOT DOCUMENTED */ map::iterator endComponentSizes(){ return( componentSizes.end() ); } //@} ///@name Marginalized and general kernel related functions //@{ /** sets the start, stop and transition probabilities for the calculation of the Kashima kernel as published in (Kashima et al. 2003). the start probability is set to an identical value of 1/numAtoms() for all atoms. the stop probability can be choosen between 0 and 1. the transition probability of each bond of an Atom is set to (1.0 - stop probability) / numbonds. */ void setKashimaKernelProb( double aPq, bool skipSkeleton = false ); /** returns the sum of transition probabilities. should be used with fused graph. */ double sumPT(); /** function used to compute the graph kernel using the fused graph + sum the powers of the transition matrix approach. see MoleculeUtils::powerKernel(). */ double sumProbabilities(); /** fast version of sumProbabilities(). */ double sumProbabilitiesFast(); /** function used to compute the graph kernel using the fused graph + sum of the powers of the transition matrix approach. see MoleculeUtils::powerKernel() */ double sumPQPS(); /** fast version of sumPQPS(). */ double sumPQPSFast(); /** function used to compute the graph kernel using the fused graph + sum of the powers of the transition matrix approach. see MoleculeUtils::powerKernel(). same as raisePower() but instead of using the Bond class the transition probabilities are read from fastPt and fastPTSave. */ void raisePowerFast(); /** compares this molecule with anotherMolecule and returns the value defined by the function pt2GraphKernel, using the function pt2AtomKernel to compare atoms and pt2BondKernel to compare bonds. */ double computeKernel( Molecule* anotherMolecule, double (*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double (*pt2AtomKernel) ( Atom*, Atom* ), double (*pt2BondKernel)( Bond*, Bond* ), int parameter1, int parameter2 ), double (*pt2AtomKernel)( Atom*, Atom* ), double (*pt2BondKernel)( Bond*, Bond* ), int parameter1, int parameter2 = 1 ); /** returns the kernel value for the molecule with itself. */ double getSelfKernel( double(*pt2GraphKernel)( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)( Bond*, Bond* ), int, int ), double(*pt2AtomKernel)( Atom*, Atom* ), double(*pt2BondKernel)( Bond*, Bond* ) , int parameter1, int parameter2 = 1 ){ #ifdef DEBUG cout << "Molecule::getSelfKernel(...) " << endl; #endif if(selfKernelCalculated == false){ #ifdef DEBUG cout << " sk not computed, computing" << endl; #endif calculateSelfKernel( pt2GraphKernel, pt2AtomKernel, pt2BondKernel, parameter1, parameter2 ); } #ifdef DEBUG cout << " returning " << selfKernel << endl; #endif return( selfKernel ); } /** calculates the molecule's self Kashima kernel and stores the result in the private selfKernel variable. */ double calculateSelfKernel ( double(*pt2GraphKernel) ( Molecule* mol1, Molecule* mol2, double(*pt2AtomKernel)(Atom*, Atom*), double(*pt2BondKernel)(Bond*, Bond*), int, int ), double(*pt2AtomKernel)( Atom*, Atom* ), double(*pt2BondKernel)( Bond*, Bond* ), int paramter1, int parameter2 ); /** emits an error if the kernel was not calculated before. */ double getSelfKernel() throw( CError ); /** sets the set kernel to the given value. */ void setSelfKernel(double value); /** adds a value to the self kernel. */ void addToSelfKernel( double ); /** substracts a value to the self kernel. */ void substractToSelfKernel( double ); /** resets the self kernel. */ void resetSelfKernel(){ selfKernelCalculated = false; } //@} ///@name Matrix utilities functions related to the 3D pharmacophore kernel (Mahe 2006) //@{ /** sets an entry of the adjacency matrix of the molecule to a given value. */ void setAdjacency(int i, int j, double value); /** returns the value of an entry of the adjacency matrix of the molecule. */ double getAdjacency(int i, int j); /** sets an entry of the 'walks' matrix (i.e., a weighted adjacency matrix) of the molecule to a given value. */ void setWalks(int i, int j, double value); /** returns the value of an entry of the walks matrix of the molecule. */ double getWalks(int i, int j); /** multiplies the walks matrix by the adjacency matrix. */ void raisePowerAdjacency(); /** computes the trace of the walks matrix. */ double traceWalks(); /** computes the diagonal elements of the walk matrix multiplied by the adjacency matrix, and sums to get the trace. */ double traceDiagWalks(); //@} ///@name Output functions //@{ /** returns a string description of the molecule (name, unique Id, memory Adress, number of atoms). */ string toString(); /** returns a short string description of the molecule. (name and unique Id). */ string toStringShort(); /** returns a string description of the molecule (name, unique Id, memory Adress, number of atoms and atom descriptions). */ string toStringLong(); /** prints a description of the molecule to stdout using toString() and DataContainer describe(). */ void describe(); /** prints a description of the molecule to stdout using toString(). */ void describeShort(); /** prints a description of the molecule to stdout using the molecule toString(), all atoms toStringShort(), and all bonds toStringShort(). */ void describeLong(); /** prints a description of each atom (using describe()) and each bond (using describe())to stdout. */ void describeEachAtom(); /** write the molecule definition in a MDL MOL file. */ void writeMOL( string aFileName ); /** writes a dot file representing the molecule. the molecule can be converted to graphic representation using Graphviz's dot program. */ void writeDOT( string aFilename, bool perretLabels = false ); /** write a MDL structure data (SD) file with a single molecule. */ void writeSD( string aFileName ); //@} // ******************************* // // **** DEPRECATED FUNCTIONS **** // // ****************************** // /** save all bonds in the molecule */ //void saveAllBonds(); /** returns the graph kernel value between two molecules by default the atom kernel is: MoleculeUtils::atomKernelExternalMatrix the values used as kernel for atoms are read in a table. by default this table returns the same value as the Kashima kernel. However the kernel used to compare atoms can be changed by loading a gram matrix for elements using elements.loadGramAtoms(...), or by specifying another atom kernel function pointer as the pt2AtomKernel argument and the bond kernel is: MoleculeUtils::bondKernelType */ //virtual float moleculeKernel( Molecule* anotherMolecule, float (*pt2AtomKernel)( Atom*, Atom* ), float (*pt2BondKernel)( Bond*, Bond* ), int convergenceCondition = 1000 ); /** returns the graph kernel value between two molecules using the Ecole des Mine approach, based on the fused graph sum of matrix power. The kernel computes the sum of te probabilities of finite length paths, as opposed to infinite length path as in Kashima's original paper */ //virtual float powerKernelUntilN( Molecule* anotherMolecule, float (*pt2AtomKernel)( Atom*, Atom* ), int maxPower = 4 ); /** returns the graph kernel value between two molecules using the Ecole des Mine approach, based on the fused graph sum of matrix power. The kernel computes an approximation of the sum of the probabilities of infinite finite length paths (until the kernel value does not change by more than 1/converg */ //virtual float powerKernelConverge( Molecule* anotherMolecule, float (*pt2AtomKernel)( Atom*, Atom* ), int converg = 1000 ); /** returns the graph kernel value between two molecules using the Ecole des Mine approach, based on the fused graph sum of matrix power. The kernel computes the sum of the probabilities of paths of length N */ //virtual float powerKernelOrderN( Molecule* anotherMolecule, float (*pt2AtomKernel)( Atom*, Atom* ), int length ); /** thread version of the Kashima kernel. Takes a kashimaThreadStruct argument as input (using the molecule1, molecule2, and convergence members to calculate the kernel). returns the result of the kernel calculation in the member "result" of the kashimaThreadStruct. If calculation occurred without errors than the noErrors member of the kashimaThreadStruct is set to true. to create a new threads with this function use: #define MAXTHREADS 100 pthread_t thread; kashimaThreadStruct args; if(error=pthread_create( &thread, NULL, (void*)&kashimaKernelThread, (void*) &args) !=0 ) { switch(error){ case EAGAIN: cout<<"EAGAIN\n";break; case EINVAL: cout<<"EINVAL\n";break; case ENOMEM: cout<<"ENOMEM\n";break; } exit(1); } // ensure thread is finished before comtinuing pthread_join( thread, NULL ); */ // void* kashimaKernelThread(void* arg); /** function used to compute the graph kernel using the fused graph / sum of the powers of the transition matrix approach. See MoleculeUtils::powerKernel */ //void raisePower(); /** this function outputs fragments of the molecule into a new SDFile the fragments are generated by removing one bound. Aromatic bounds are not broken. */ //void exportFragments( string sdFileName, int minAtoms = 1 ); protected: /** this function is called internally every time a molecule is changed (atom or bonding), reset flags. */ void moleculeChanged( bool resetSSSR = true, bool resetMorganIndex = true ); /** container for atoms. */ vector atoms; /** container for hidden atoms (for example the removed hydrogens are moved to this container). */ vector hiddenAtoms; /** container for rings (smallest set of smallest rings). */ vector sssr; /** flag indicating if smallest set of smallest rings was detected. */ bool flagHasSSSRDetected; /** counter of the number of molecules which have been instanciated. */ static int counter; /** molecule unique Id. */ int id; /** flag to select subsets of molecules. */ bool selectedFlag; /** molecule's self Kernel. */ double selfKernel; /** true if selfKernel was calculated, false otherwise. */ bool selfKernelCalculated; /** true if the molecule has chiral information. */ bool chiral; /** true if the molecule has activity information. */ bool flagActivity; /** type and name of descriptor to be used to sort molecules. use setSortDescriptor( aType, aName ) getSortDescriptorType(), setSortDescriptorName() to access. */ int sortDescriptorType; /** name of the sorting descriptor. */ string sortDescriptorName; /** contains the iteration number at which the maximum number of distinct connectivity values is reached. */ int maxMorganIteration; private: // Private methods /** hide all bonds in the molecule warning clears all previously hidden bonds */ //void hideAllBonds(); /** links two atoms poited by aSource and aTarget using aBondLabel bond type. If the two atoms are not in the atoms vector they are first added to the molecule WARNING this function adds a single directed bond, only to be used with in Molecule::raisePower(). */ Bond* linkAtomsNoReturn( Atom* aSource, Atom* aTarget, int aBondLabel, int aBondStereo = 0, int aBondNotUsed = 0, int aBondTopology = 0, int aBondReactionCenter = 0 ); /** links a pair of atoms in a single direction only : firstAtom -> secondAtom. */ void linkAtomsNoReturn( int firstAtom, int secondAtom, int aBondLabel ) throw( CError ); /** hash used in the fast computations of the powers of the product graph adjacency matrix. */ map* >* fastPT; /** hash used in the fast computations of the powers of the product graph adjacency matrix. */ map* >* fastPTNext; /** hash used in the fast computations of the powers of the product graph adjacency matrix. */ map* >* fastPTSave; /** hash used in the fast computations of the powers of the product graph adjacency matrix. */ map fastPQ; /** hash used in the fast computations of the powers of the product graph adjacency matrix. */ map fastPS; /** NO DOCUMENTATION */ map< int, int > componentSizes; /** activity value of the molecule. */ float activity; /** location where the molecule is stored. automatically set by readMOL. */ string location; /** original format of the molecule. */ int originalFormat; /** adjacency matrix of the molecule. used to compute the 3D pharmacophore kernel. */ vector< vector >* adjacency; /** 'walks' matrix of the molecule. used to compute the 3D pharmacophore kernel. */ vector< vector >* walks; }; #endif Rchemcpp/src/chemcpp/src/kcfmolecule.h0000644000175000017500000000524412116123101020373 0ustar klambaueklambaue/**************************************************************************************** kcfmolecule.h ----------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef KCFMOLECULE_H #define KCFMOLECULE_H #include /**this is a subclass of molecules providing special functions for molecules represented in KCF format *@author Jean-Luc Perret */ class KCFMolecule : public Molecule { public: /** class constructor. */ KCFMolecule(); /** class desctructor. */ virtual ~KCFMolecule(); /** reads the molecule description from a KCF file. The internal representation of the molecule is the Atom KEGG atoms, so the function Elements:loadDefinition should have previously loaded the KEGG atom definition types). It is not possible to mix KCF and MOL files. A function readKCFconverted should be written to read KCF file with an internal representation of elements. WARNING this function only reads the first entry in the kcf file. */ void readKCF( string aFileName ) throw( CError ); /** writes a kcf file. */ void writeKCF( string aFileName ); /** adds an atom to the kcf molecule. */ virtual Atom* addAtom(string aSymbol) throw( CError ); //virtual float newKernel( Molecule* anotherMolecule, int convergenceCondition); //virtual vector< vector >* rlk( vector< vector >* r, vector< vector >* rwork, Molecule* anotherMolecule, int convergenceCondition ); }; #endif Rchemcpp/src/chemcpp/src/elements.cpp0000644000175000017500000002304712116123101020252 0ustar klambaueklambaue/**************************************************************************************** elements.cpp --------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #include "elements.h" #include "constant.h" //#define ELEMENTS_DEBUG 1 // instanciate Elements (Atoms) Elements elements( ELEMENTSFILENAME, GRAMATOMSIDENTITYFILENAME ); Elements KEGGelements( KEGGATOMS, GRAMKEGGATOMSIDENTITYFILENAME ); // load gram matrices for comparing atoms void Elements::loadGramAtoms( string aFileName ){ //periodicTable.clear(); #ifdef ELEMENTS_DEBUG cout << "loading Atom Kernel matrix from file: " << aFileName << endl; #endif ifstream inFile; inFile.open( aFileName.c_str(), ios::in ); if(!inFile.good()){ CError e = CError( FILENOTFOUND, aFileName + " file not found" ); e.describe(); throw(e); } int linesize=2048; int lineNumber = 0; string delimiter = ";"; char *line = new char[linesize]; vector value; // read and instanciate each new element string stringLine; while( !inFile.eof() ){ lineNumber++; inFile.getline( line, linesize-1,'\n' ); stringLine = line; // if line is empty or starts with // or # then skip this line if( stringLine.size() == 0 ){ continue; } if( stringLine.substr( 0, 1 ) == "#" || stringLine.substr(0,2) == "//" ){ continue; } StringUtils::Split( line, delimiter, value ); if( value.size() < numElements() ){ stringstream out; out << aFileName << " line " << lineNumber << ": has " << value.size() << " values while " << numElements() << " are required "; CError e = CError( BADFILE, out.str() ); e.describe(); cout << "found:" << endl; cout << line << endl; for( uint c = 0; c < value.size(); c++ ){ cout << c+1 << " " << value[c] << endl; } throw(e); } // PROCESS LINE for( uint i = 0; i < numElements(); i++ ){ gramAtom[lineNumber - 1][i] = atof( value[i].c_str() ); } value.clear(); } delete[] line; gramAtomName = aFileName; #ifdef ELEMENTS_DEBUG cout << "loading Atom Kernel matrix from file: " << aFileName << " DONE " << endl; #endif } /** loads the elements from a file */ Elements::Elements( string aFileName, string aGramAtomFileName ){ //cout << "loading element from file " << aFileName << "... " << endl; loadDefinition( aFileName ); loadGramAtoms( aGramAtomFileName ); } Elements::~Elements(){ map::iterator i; for( i = periodicTable.begin(); i != periodicTable.end(); i++ ){ delete (*i).second; // delete bonds } periodicTable.clear(); } /** returns an atom pointer to an element designed using its chemical symbol */ Atom* Elements::operator[]( string aSymbol ){ // return( *periodicTable[aSymbol] ); return( getElement( aSymbol ) ); } /** returns a pointer to the element designed by aSymbol */ Atom* Elements::getElement( string aSymbol ){ map::iterator anElement = periodicTable.find( aSymbol ); if( anElement == periodicTable.end() ) { // element does not exist, emit exception stringstream out; out << "Error in Elements::getElement: element " << aSymbol << " not found in the Elements Set"; CError e( NOTFOUND, out.str() ); e.describe(); throw( e ); } return( (*anElement).second ); } /** returns the filename of the atom kernel matrix loaded. returns "default" if no kernel matrix was loaded. */ string Elements::getAtomKernelName(){ return gramAtomName; } /** loads the definition of elements */ void Elements::loadDefinition( string aFileName ){ #ifdef ELEMENTS_DEBUG cout << "Elements::loadDefinition loading element from file " << aFileName << "... " << endl; #endif // empty periodic table periodicTable.clear(); Atom::resetCounter(); Atom* atomP; gramAtomName = "default"; ifstream inFile; inFile.open( aFileName.c_str(), ios::in ); if(!inFile.good()){ CError e = CError( FILENOTFOUND, aFileName + " file not found" ); e.describe(); throw(e); } int linesize=512; int lineNumber = 0; string delimiter( ";" ); char *line = new char[linesize]; vector header; vector fullName; vector unit; vector type; vector value; Descriptor* newDs; Descriptor* newDi; Descriptor* newDf; //cout << "found " << SYMBOLNAME << " at position " << symbolPosition << endl; // read fullName line and split fields lineNumber++; inFile.getline( line, linesize-1, '\n' ); //cout << "read fullname line, found: " << line << endl; StringUtils::Split( line, delimiter, fullName ); // LINE 1: full name // read units line and split fields lineNumber++; inFile.getline( line, linesize-1, '\n' ); //cout << "read units line, found: " << line << endl; StringUtils::Split( line, delimiter, unit ); // LINE 2: unit // read descriptor types line and split fields lineNumber++; inFile.getline( line, linesize-1, '\n' ); //cout << "read types line, found: " << line << endl; StringUtils::Split( line, delimiter, type ); // LINE 3: type // read header line and split fields lineNumber++; inFile.getline( line, linesize-1, '\n' ); StringUtils::Split( line, delimiter, header ); // LINE 4: header // find symbol position int symbolPosition = 0; bool found = false; uint i; for( i = 0; i < header.size(); i++ ){ // cout << "C:"<setType( typeC ); typeC++; #ifdef DEBUG cout << value[symbolPosition] << ", "; #endif // add all descriptors to the new Element for( i = 0; i < header.size(); i++ ){ if( type[i] == "String" ) { if( value[i] == "Nil" || value[i] == "NA" ){ newDs = atomP->addKindStringDescriptor( header[i], "", unit[i], fullName[i] ); newDs->setEmpty(); }else{ newDs = atomP->addKindStringDescriptor( header[i], value[i], unit[i], fullName[i] ); } }else if( type[i] == "Integer" ){ if(header[i] == "AN" || header[i] == "An" ){ atomP->setAN( atoi(value[i].c_str()) ); }else{ if( value[i] == "Nil" || value[i] == "NA" ){ newDi = atomP->addKindIntDescriptor( header[i], 0, unit[i], fullName[i] ); newDi->setEmpty(); }else{ newDi = atomP->addKindIntDescriptor( header[i], atoi(value[i].c_str()), unit[i], fullName[i] ); } } }else if( type[i] == "Float" ){ if( value[i] == "Nil" || value[i] == "NA" ){ newDf = atomP->addKindFloatDescriptor( header[i], 0.0, unit[i], fullName[i] ); newDf->setEmpty(); }else{ atomP->addKindFloatDescriptor( header[i], atof( value[i].c_str() ), unit[i], fullName[i] ); } }else{ // unknown data type stringstream out; out << "type " << type[i] << " in line " << lineNumber-1 << " of file " << aFileName << " is not allowed. String, Float and Integer only are allowed "; CError e( UNKNOWNDATATYPE, out.str() ); e.describe(); throw( e ); } } atomP->setLabel( atomP->getSymbol() ); //cout << "atom " << atomP->getSymbol() << " now has label " << atomP->getLabel() << endl; //atomP->describe(); value.clear(); } #ifdef DEBUG cout << endl; #endif delete[] line; inFile.close(); // by default define an identity matrix to compare atoms uint j = 0; for( i = 0 ; i < numElements() ; i++ ){ for( j = 0 ; j < numElements() ; j++ ){ if(i == j){ gramAtom[i][j] = 1; }else{ gramAtom[i][j] = 0; } } } #ifdef ELEMENTS_DEBUG cout << "Elements::loadDefinition done" << endl; #endif } /** write a description of the elements to cout */ //void Elements::describe(){ // map::iterator i; //for(i = periodicTable.begin(); i!=periodicTable.end(); i++){ //cout << i->second << endl; // } //} // Rchemcpp/src/chemcpp/src/Makefile0000644000175000017500000000124412116123101017365 0ustar klambaueklambaue CXXFLAGS += -w -fPIC -I../src/ objects := constant.o node.o atom.o bond.o cerror.o chemcpp.o datacontainer.o descriptor.o elements.o jlpioutils.o kcfmolecule.o molecule.o moleculeset.o moleculeutils.o ring.o stringutils.o all: ../../libchemcpp.dll ../../libchemcpp.so ../../libchemcpp.dll: $(objects) $(CXX) -shared -o ../../libchemcpp.dll $(CXXFLAGS) $(objects) UNAME := $(shell uname -s) ifeq ($(UNAME), Darwin) ../../libchemcpp.so: $(objects) $(CXX) -dynamiclib -o ../../libchemcpp.so $(CXXFLAGS) $(objects) -Wl,-install_name -Wl,@loader_path/libchemcpp.so else ../../libchemcpp.so: $(objects) $(CXX) -shared -o ../../libchemcpp.so $(CXXFLAGS) $(objects) endif Rchemcpp/src/chemcpp/src/constant.h0000644000175000017500000000664212116123101017736 0ustar klambaueklambaue#ifndef CONSTANT_H #define CONSTANT_H #define MDL2000 2000 #define MDL2000JLP 2001 #define MDL3000 3000 #define NAVALUE -9999 #define INTEGER 0 #define FLOAT 1 #define STRING 3 #define DEFAULT 4 #define MULTITHREAD 1 #define MAXTHREADS 100 #ifndef CHEMCPPPATH #endif #define ELEMENTSFILENAME CHEMCPPPATH "/data/elements.csv" // semicolumn (;) separated file with each line describing an element #define KEGGATOMS CHEMCPPPATH "/data/keggatoms.csv" #define NUMELEMENTS 109 // maximum number of elements #define GRAMATOMSIDENTITYFILENAME CHEMCPPPATH "/data/gramAtoms.binary.csv" // KASHIMA ORIGINAL #define GRAMATOMSDOTPRODUCT1FILENAME CHEMCPPPATH "/data/gramAtoms.dotproduct.csv" // WARNING WRONG #define GRAMATOMSCRENDOTPRODUCTFILENAME CHEMCPPPATH "/data/gramAtoms.cren.dotproduct.csv" #define GRAMATOMSCRENPOLYNOMIALFILENAME CHEMCPPPATH "/data/gramAtoms.cren.polynomial.csv" #define GRAMATOMSCRENGAUSSIANFILENAME CHEMCPPPATH "/data/gramAtoms.cren.gaussian.csv" #define GRAMKEGGATOMSIDENTITYFILENAME CHEMCPPPATH "/data/gramKeggAtoms.kashima.csv" #define PS "ps" // name used to designate the Kashima Start Probability #define PT "pt" // name used to designate the Kashima Transition Probability #define PQ "pq" // name used to designate the Kashima Stop Probability // ATOM PROPERTIES #define FULLNAME "Name" // name used to designate the Full name #define SYMBOLNAME "Symbol" // name used to designate the symbol letter in the elements file #define AN "An" // name used to designate the atomic number #define AM "Am" // name used to designate the atomic mass #define VWR "Vwr" // name used to designate the van der Waals radius #define CR "Cr" // Covalent radius #define AV "Av" // Atomic volume #define MP "Mp" // Melting point #define BP "Bp" // Boiling point #define VALENCE "Valence" // Valence #define EA1 "Ea1" // First electron affinity #define IE1 "Ie1" // First ionization energy #define IE2 "Ie2" // second ionization energy #define IE3 "Ie3" // third ionization energy #define IE4 "Ie4" // fourth ionization energy #define IE5 "Ie5" // fifth ionization energy #define IE6 "Ie6" // sixth ionization energy #define IE7 "Ie7" // seventh ionization energy #define IE8 "Ie8" // eigth ionization energy #define EN "En" // Electronegativity // MOLECULE PROPERTIES #define ACTIVITY "activity" // stores the biological activity // BOND LABELS #define NOBOND 0 #define SINGLEBOND 1 #define DOUBLEBOND 2 #define TRIPLEBOND 3 #define AROMATICBOND 4 #define SINGLECYCLEBOND 5 #define DOUBLECYCLEBOND 6 #define TRIPLECYCLEBOND 7 // CALCULATIONS #define GRAMDECIMALPRECISION 9 // number of decimals to print in the gram matrix #define MAXSIMILARITYWARNING 0.999999 // number between 0 and 1. If the normalised kernel // exceeds this value between two molecules a warning // is printed #define MINSIMILARITYWARNING 1-MAXSIMILARITYWARNING #define DUPLICATEMWIDENTITY 0.0001 // in MoleculeSet::removeDuplicates, if mw1-mw2 < DUPLICATEMWIDENTITY then // two compounds are considered to have identical MW #define COMPONENTINDEX "componentIndex" // name of the descriptor used for marking fragments in a // molecule. #endif #include #include std::string GETChemcpppath(); #define CHEMCPPPATH GETChemcpppath() + Rchemcpp/src/chemcpp/src/stringutils.cpp0000644000175000017500000002451012116123101021021 0ustar klambaueklambaue/**************************************************************************************** stringutils.cpp ------------------ copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ //#define DEBUG 1 #include "stringutils.h" #include //StringUtils::StringUtils(){ //} //StringUtils::~StringUtils(){ //} int StringUtils::toInt( const string input ){ std::stringstream oss ( input ); int rv; oss >> rv; return rv; } float StringUtils::toFloat( const string input ){ std::stringstream oss ( input ); float rv; oss >> rv; return rv; } string StringUtils::toString( const int input ){ std::stringstream oss; string rv; oss << input; oss >> rv; return rv; } string StringUtils::toString( const float input ){ std::stringstream oss; string rv; oss << input; oss >> rv; return rv; } int StringUtils::Split(const string input, const string delimiter, vector& results){ int iPos = -1; int newPos = -1; int sizeS2 = delimiter.size(); int isize = input.size(); vector positions; newPos = input.find( delimiter, 0 ); #ifdef DEBUG //cout << newPos << endl; #endif // only one word in the input if( newPos < 0 ){ results.push_back( input ); return 0; } // store delimiters positions int numFound = 0; while( newPos > iPos ){ numFound++; positions.push_back(newPos); iPos = newPos; newPos = input.find (delimiter, iPos+sizeS2+1); } // push words for( uint i=0; i <= positions.size(); i++ ){ string s; int offset = 0; if( i == 0 ){ s = input.substr( i, positions[i] ); offset = sizeS2; }else{ offset = positions[i-1] + sizeS2; } if( offset < isize ){ if( i == positions.size() ){ s = input.substr(offset); }else if( i > 0 ){ s = input.substr( positions[i-1] + sizeS2, positions[i] - positions[i-1] - sizeS2 ); } } if( s.size() > 0 ){ results.push_back(s); } } return numFound+1; } string StringUtils::mergeWords( vector& words, const string separator ){ stringstream result; vector< string >::iterator i; bool first = true; for( i = words.begin(); i != words.end(); i++ ){ if( first == true ){ result << (*i); first = false; }else{ result << separator << (*i); } } return( result.str() ); } string StringUtils::rmSpace( const string input ){ stringstream output; string letter; for( uint i=0; i=0; i--){ //cout << input[i] << "#"; letter = input[i]; //cout << dec << letter << "#" << endl; if(status<2){ if( letter != " "){ output = letter + output; status++; // finished removing spaces //cout << "."; }else{ //cout << "#"; } }else{ output = letter + output; //cout << "."; } } //cout << ":" << output << ";" << endl; return( output ); } string StringUtils::chomp(string inString) { while ((inString.substr(inString.length(),1) == "\n") || (inString.substr(inString.length(),1) == "\r") ) { inString = inString.substr(1,inString.length()-1); } return inString; } string StringUtils::field( string aString, int start, int length ){ int sl = aString.length(); if(sl >= start + length){ return(aString.substr(start, length)); }else{ if(start < sl){ return(aString.substr(start, sl-start)); }else{ return(""); } } } /** append aChar to aString so that the total length is exactly equal to aLength. If aString is longer, it gets cropped. */ string StringUtils::fill( string aString, int aLength, string aChar ){ int sl = aString.length(); if( sl == aLength){ return( aString ); }else if ( sl > aLength ){ return( aString.substr( 0, aLength ) ); }else{ string cs = ""; for( int i = 0; i < aLength - sl; i++ ){ cs = cs + aChar; } return( aString + cs ); } } /** prepend aChar to aString so that the total length is exactly equal to aLength. If aString is longer, it gets cropped. */ string StringUtils::preFill( string aString, int aLength, string aChar ){ //cout << "preFill: " << aString << endl; int sl = aString.length(); if( sl == aLength){ return( aString ); }else if ( sl > aLength ){ return( aString.substr( 0, aLength ) ); }else{ string cs = ""; for( int i = 0; i < aLength - sl; i++ ){ cs = cs + aChar; } return( cs + aString ); } } string StringUtils::preFill( int anInt, int aLength, string aChar ){ return( StringUtils::preFill( StringUtils::toString( anInt ), aLength, aChar ) ); } string StringUtils::preFill( float aFloat, int aLength, string aChar ){ //cout << "Float prefill" << aFloat << " " << StringUtils::toString( aFloat ) << endl; return( StringUtils::preFill( StringUtils::toString( aFloat ), aLength, aChar ) ); } /** returns the directory portion of a full path */ string StringUtils::getPath( string aLocation ){ vector words; StringUtils::Split( aLocation, "/", words ); std::stringstream res; if( words.size() > 1){ vector::iterator itw = words.begin(); for(; itw!= words.end()-1; itw++ ){ res << (*itw) << "/"; } }else{ return( "./" ); } return( res.str() ); } /** returns the directory portion of a full path */ string StringUtils::getFileName( string aLocation ){ //#ifdef DEBUG //cout << "StringUtils::getFileName: getting file name from " << aLocation << endl; //#endif vector words; // replace all // by / // replace all // with / /*bool end = false; while( end == false ){ StringUtils::Split( aLocation, "//", words ); if( words.size() > 1 ){ aLocation = mergeWords( words , "/" ); #ifdef DEBUG cout << " -> " << aLocation << endl; #endif words.clear(); }else{ #ifdef DEBUG cout << "end" << endl; #endif end = true; } } words.clear();*/ #ifdef DEBUG cout << "now splitting with delimiter / " << endl; //aLocation = "/scratch/fhfjdk"; cout << aLocation << endl; #endif StringUtils::Split( aLocation, "/", words ); #ifdef DEBUG cout << "splitting done: " << words.size() << " words: " << words[0] << endl; #endif std::stringstream res; if( words.size() > 1){ vector::iterator itw = words.end()-1; //res << (*itw) << "/"; res << (*itw); }else{ //cout << " ---> " << aLocation << endl; return( aLocation ); } //cout << " ---> " << res.str() << endl; return( res.str() ); } /** returns the directory portion of a full path */ string StringUtils::getExtension( string aLocation ){ string aFileName = getFileName( aLocation ); vector words; StringUtils::Split( aLocation, ".", words ); std::stringstream res; if( words.size() > 1){ vector::iterator itw = words.end()-1; res << (*itw); }else{ return( "" ); } return( res.str() ); } /** returns the path without extension */ string StringUtils::getNoExtension( string aLocation ){ string aFileName = getFileName( aLocation ); #ifdef DEBUG cout << "StringUtils::getNoExtension: removing extension from " << aFileName << endl; #endif vector words; StringUtils::Split( aLocation, ".", words ); std::stringstream res; if( words.size() > 1){ vector::iterator itw; for( itw = words.begin(); itw != words.end()-1; itw++ ){ res << (*itw) << "."; } }else{ return( aLocation ); } return( getPath( aLocation ) + res.str().substr( 0, res.str().size()-1 ) ); } /** returns aLength characters at the right end of aString */ string StringUtils::right( string aString, uint aLength ){ if( aString.length() >= aLength){ return( aString.substr( aString.length() - aLength, aLength ) ); }else{ return( aString ); } } #include string StringUtils::toUpper( string aString ){ for( unsigned i=0; i #include // for back_inserter #include #include //#include // old #include struct ToUpper { ToUpper(std::locale const& l) : loc(l) {;} //char operator() (char c) const { return std::toupper(c,loc); } char operator() (char c) const { return toupper(c); } private: std::locale const& loc; }; struct ToLower { ToLower(std::locale const& l) : loc(l) {;} //char operator() (char c) const { return std::tolower(c,loc); } char operator() (char c) const { return tolower(c); } private: std::locale const& loc; }; */ string StringUtils::getFirstNonSpace( string aString ) { for( long i = 0; i < aString.size() ; i++ ){ string c = aString.substr(i,1); if( c != " " ){ return( c ); } } } Rchemcpp/src/chemcpp/src/bond.cpp0000644000175000017500000002251412116123101017356 0ustar klambaueklambaue/**************************************************************************************** bond.cpp ------------ copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #include "bond.h" Bond::Bond(){ pt = 0; label = NAVALUE; perretLabel = NAVALUE; stereo = 0; notUsed = 0; topology = 0; reactionCenter = 0; } Bond::Bond(Atom* aSource, Atom* aTarget, int aLabel, int aPerretLabel, int aBondStereo, int aBondNotUsed, int aBondTopology, int aBondReactionCenter ){ source = aSource; target = aTarget; label = aLabel; perretLabel = aPerretLabel; pt = 0; stereo = aBondStereo; notUsed = aBondNotUsed; topology = aBondTopology; reactionCenter = aBondReactionCenter; //addFloatDescriptor("pt" ,0.0, "probability", "transition probability for calculation of the Kashima Kernel"); } Bond::~Bond(){ //cout << "DELETING BOND" << endl; // delete atom descriptors /*map* >::iterator iti; for( iti = atomDescriptors.begin(); iti != atomDescriptors.end(); iti++ ){ #ifdef DEBUG cout << "deleting atom descriptor [ " << (*iti).second << "]" << endl; #endif delete (*iti).second; (atomDescriptors).erase(iti); }*/ } /** reverses the bond direction */ void Bond::reverse(){ Atom* toto; toto = target; target = source; source = toto; } /** returns a string description of the bond */ string Bond::toString(){ stringstream out; out << getSource()->toStringShort() << " -" << getLabel() << "/" << getPerretLabel() << "- " << getTarget()->toStringShort(); if( rings.size() > 0 ){ out << " in " << rings.size() << " ring of size"; vector::iterator ri; for( ri = rings.begin(); ri != rings.end(); ri++ ){ out << " " << (*ri)->size(); } }else{ out << " in no ring"; } return( out.str() ); } string Bond::toStringShort(){ stringstream out; out << getSource()->toStringShort() << " -" << getLabel() << "- " << getTarget()->toStringShort(); return( out.str() ); } void Bond::describe(){ cout << " " << this->toString() << endl; //DataContainer::describeShort(); } void Bond::describeShort(){ cout << " " << this->toStringShort() << endl; } Bond* Bond::getReverse() throw( CError ){ try{ Bond* result = getTarget()->getBondWithTarget( getSource() ); return( result ); }catch( CError e ){ throw( e ); } } bool Bond::hasRing( Ring* aRing ){ vector::iterator ri; for( ri = rings.begin(); ri != rings.end(); ri++ ){ if( (*ri) == aRing ){ return( true ); } } return( false ); } void Bond::setPerretLabel(){ perretLabel = label; if( hasRing() ){ // only labels for bonds member of a ring a changed if( label != AROMATICBOND ){ // only non aromatic cycles need renaming switch( label ){ case SINGLEBOND: perretLabel = SINGLECYCLEBOND; break; case DOUBLEBOND: perretLabel = DOUBLECYCLEBOND; break; case TRIPLEBOND: perretLabel = TRIPLECYCLEBOND; break; } } } } Atom* Bond::getTarget() throw( CError ){ if( !hasTarget() ){ stringstream out; out << "Bond::getTarget: Bond " << toStringShort() << " has no target"; CError e( ATOMNOTFOUND , out.str() ); e.describe(); throw(e); }else{ return( target ); } } Atom* Bond::getSource() throw( CError ){ if( !hasSource() ){ stringstream out; out << "Bond::getSource: Bond " << toStringShort() << " has no source"; CError e( ATOMNOTFOUND , out.str() ); e.describe(); throw(e); }else{ return( source ); } } void Bond::hideToFrom(){ getSource()->hideBond( getTarget() ); getTarget()->hideBond( getSource() ); } void Bond::restoreToFrom(){ getSource()->restoreHiddenBond( getTarget() ); getTarget()->restoreHiddenBond( getSource() ); } /** Adds an atom descriptor intended to store stoms, for example source ant target are automatically added to a bond */ //void Bond::addAtomDescriptor(string aLabel, Atom* aValue, string aUnit, string aComment){ // atomDescriptors[aLabel] = new Descriptor( aLabel, *aValue, aUnit, aComment ); //} /*Descriptor< Atom >* Bond::getAtomDescriptor(string aLabel){ //bool found; // look among atom descriptors map* >::iterator its; for( its = atomDescriptors.begin(); its != atomDescriptors.end(); its++ ){ if( (*its).first == aLabel ){ //found = true; return ( (*its).second ); } } // if no corresponding descriptors found, throw exception CError e( MISSINGDESCRIPTOR, "no Atom descriptor " + aLabel ); e.describe(); throw(e); }*/ /*bool Bond::deleteDescriptor( string aLabel, bool found ){ // first check for the descriptor among the atom descriptors map* >::iterator iti; for( iti = (atomDescriptors).begin(); iti != (atomDescriptors).end(); iti++ ){ if( (*iti).first == aLabel ){ #ifdef DEBUG cout << "deleting Atom descriptor "; #endif (*iti).second->describe(); delete (*iti).second; (atomDescriptors).erase(iti); found = true; } } // then look in the remaining descriptors found = DataContainer::deleteDescriptor( aLabel, found ); return(found); }*/ /** returns 1 if two bonds have the same label, 0 otherwise */ /*float Bond::binKernel(Bond* aBond){ if( aBond->getLabel() == this->getLabel() ) { return 1; }else{ return 0; } } */ /*string Bond::toStringBFSVector(){ stringstream out; vector::iterator i; for( i = BFSVector.begin(); i!= BFSVector.end(); i++){ out << (*i)->toStringShort() << ", "; } return( out.str() ); }*/ /*Ring* Bond::getRingBFS( vector* toVisit, Bond* init ){ // set BFSVector for this atom //if( (*toVisit->begin()) == this ){ toVisit->erase( toVisit->begin() ); //} // update BFSPath // vector* ppath = getSource()->getBFSVector(); vector::iterator ai; for( ai = ppath->begin(); ai != ppath->end(); ai++ ){ BFSVector.push_back( (*ai) ); } BFSVector.push_back( this ); cout << "a4h IN GETRINGBFS for atom " << toStringShort() << endl; cout << "a4h with BFSVector of size " << getBFSVectorSize() << ": " << toStringBFSVector() << endl; map::iterator i; cout << "a4h checking target neighbour bonds" << endl; for( i = getTarget()->bonds.begin(); i != getTarget()->bonds.end(); i++ ){ // for all neighbours except where we just come from check if they were already visited cout << "a4h checking " << (*i).second->toStringShort() << endl; if( (*i).first != getSource() ){ // if the bond is not a bond returning were we just come from cout << "a4h is not a return bond" << endl; if( (*i).second->getBFSVectorSize() > 0){ // if the bond was already visited cout << "a4h " << (*i).first->toStringShort() << " was already visited" << endl; // check if the ring is valid vector vectorIntersect; getVectorIntersect( getBFSVector(), (*i).second->getBFSVector(), &vectorIntersect ); cout << "a4h ring has intersect of size " << vectorIntersect.size() << endl; if( vectorIntersect.size() == 1 ){ // if yes create the ring and return a pointer to it vector vectorUnion; getVectorUnion( getBFSVector(), (*i).second->getBFSVector(), &vectorUnion, &vectorIntersect ); return( new Ring( &vectorUnion ) ); }else{ // if no abandon this branch } }else{ // it the bond was not yet visited, add it to the list of bonds to visit and set BFSVector of target atom cout << "a4h " << (*i).first->toStringShort() << " not yet visited, will continue this branch" << endl; for( ai = BFSVector.begin(); ai != BFSVector.end(); ai++ ){ getTarget()->BFSVector.push_back( (*ai) ); } cout << "a4h target now has BFSVector" << endl; getTarget()->describe(); // if no continue with this branch toVisit->push_back((*i).second); } }else{ cout << "a4h this bond returns to where we come from skipping" << endl; } } // call getRingBFS for the next bond in the queue if( toVisit->size() > 0 ){ vector::iterator nextI = toVisit->begin(); Bond* next = *nextI; toVisit->erase( nextI ); Ring* ring = next->getRingBFS( toVisit, this ); }else{ // did not find a ring return( NULL ); } // if no ring was found return NULL return( NULL ); }*/ Rchemcpp/src/chemcpp/src/elements.h0000644000175000017500000001326712116123101017722 0ustar klambaueklambaue/**************************************************************************************** elements.h ------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef ELEMENTS_H #define ELEMENTS_H #include #include #include #include #include //using namespace std; #include #include #include /** Instanciates chemical elements @author Dr Jean-Luc Perret (luc@kuicr.kyoto-u.ac.jp), Kyoto University, Japan @version 0.1 @date 8 Jan 2004 CLASS NAME: Elements FOR: SNSF SPONSORED PROJECT PURPOSE: Instanciates the chemical elements (each of class Atom) and load their physico-chemical properties Physico-chemical properties are loaded from file ELEMENTSFILENAME in the chemcpp distribution. \warning Elements currently only instanciates the following elements: H, He, Li, Be, B, C, N, O, F, Ne, Na, Mg, Al, Si, P, S, Cl, Ar, K, Ca with the full set of physico-chemical properties, and Br, I with only full name, atomic number and symbol. An Atom pointer to an Element can be obtained by using the elements global instance of class Elements, using the following syntax: \code Atom* myPointer = elements["H"]; \endcode \warning note that to create a new Atom the copy operator should be used as described in "atom_example.cpp". The properties of the elements can be retrieve as pointers to a Descriptor by using the getIntDescriptor(), getFloatDescriptor() or getStringDescriptor() functions of the DataContainer class from which the class Atom is derived. To designate the physico-chemical properties use the macros defined in constants.h Here is a list of these macros: - FULLNAME: "Name" name used to designate the Full name - SYMBOLNAME: "Symbol" name used to designate the symbol letter in the elements file - AN: name used to designate the atomic number - AM: name used to designate the atomic mass - VWR: name used to designate the van der Waals radius - CR: Covalent radius - AV: Atomic volume - MP: Melting point - BP: Boiling point - VALENCE: Valence - EA1: First electron affinity - IE1: First ionization energy - IE2: second ionization energy - IE3: third ionization energy - IE4: fourth ionization energy - IE5: fifth ionization energy - IE6: sixth ionization energy - IE7: seventh ionization energy - IE8: eigth ionization energy - EN: Electronegativity for example to get the covalent radius value of an atom use: \code float myCR = myPointer->getFloatDescriptor(CR)->getValue(); \endcode \todo Implement deletion of KindDescriptors in destructor of class Elements */ class Elements : public DataContainer { public: ///@name General functions for building and manipulating the set of elements //@{ /** loads the elements from a file. */ Elements( string aFileName, string aGramAtomFileName ); /** elements destructor. at the moment only a default desctructor, but could eventually delete all chemical elements this class created. */ ~Elements(); /** returns an atom pointer to an element designed using its chemical symbol. */ Atom* operator[](string aSymbol); /** returns a pointer to the element designed by aSymbol. */ Atom* getElement(string aSymbol); /** function to load a gram matrix to compare atoms. this matrix will then be used by the kashima kernel. by default without calling this function, the identity matrix is loaded by the elements constructor. */ void loadGramAtoms(string filename); /** returns the filename of the atom kernel matrix loaded. returns "default" if no kernel matrix was loaded. */ string getAtomKernelName(); /** loads the definition of elements. */ void loadDefinition( string aFilename ); /** returns the number of elements in periodicTable. */ uint numElements(){ return( periodicTable.size() ); } //@} // ****************************** // // **** DEPRECATED FUNCTIONS **** // // ****************************** // /** write a description of the elements to cout */ //void describe(); //Elements(); // SECOND TIMES PUBLIC ?????? //public: /** gram matrix used to compare atoms. can be set with loadGramAtom(). this matrix is initialized to the identity matrix by the constructor. */ float gramAtom[NUMELEMENTS][NUMELEMENTS]; /** filename of the atom kernel matrix loaded. */ string gramAtomName; private : /** hash of atoms indexed by their symbolic names (ex: H, Na,...). */ map periodicTable; }; #endif Rchemcpp/src/chemcpp/src/bond.h0000644000175000017500000002122512116123101017021 0ustar klambaueklambaue/**************************************************************************************** bond.h ---------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef BOND_H #define BOND_H //#include #include "constant.h" #include "datacontainer.h" #include "atom.h" #include class Atom; /** Bond class @author Dr Jean-Luc Perret (luc@kuicr.kyoto-u.ac.jp), Kyoto University, Japan @version 0.3 @date 17 Jan 2004 CLASS NAME: Bond FOR: SNSF SPONSORED PROJECT PURPOSE: this class implements the notion of bond between two atoms A bond is directed: it has a source atom and a target atom, so a chemical bond should be represented by two instances of class Bond, one for each direction. A bond as implemented here contains a source and a target atom, a label describing the bond type (see constants.h). Since bond is a daughter of DataContainer other Descriptors may be added at runtime. A bond is derived from DataContainer, so it has all the associated descriptors, but in addition bonds inplement an atomDescriptors map allowing to store atoms as descriptors to the bond. (in addition to source ant target atom). For faster access, some descriptors are directly implemented in the class bond using native data type (double and int). This is the case for pt and label. */ class Bond { //: public DataContainer { // bonds do not need to be a DataContainer /** \example bond_example.cpp */ public: ///@name Bond construction functions //@{ /** class constructor. */ Bond(); /** class constructor. */ Bond(Atom* aSource, Atom* aTarget, int aLabel, int aPerretLabel = NAVALUE, int aBondStereo = 0, int aBondNotUsed = 0, int aBondTopology = 0, int aBondReactionCenter = 0 ); /** class destructor. */ ~Bond(); //@} ///@name Accessor functions //@{ /** returns a pointer to the target atom. */ Atom* getTarget() throw( CError ); /** returns true if target atom is set. */ bool hasTarget(){ if( target == NULL ){ return( false ); }else{ return( true ); }} /** returns a pointer to the source atom. */ Atom* getSource() throw( CError ); /** returns true if source atom is set. */ bool hasSource(){ if( source == NULL ){ return( false ); }else{ return( true ); }} /** returns the bond label. */ int getLabel(){ return(label);} /** returns the value of the Perret label (set by setPerretLabel, called everytime a molecule is modified). */ int getPerretLabel(){ return( perretLabel ); } /** sets the Perret label of the bond. */ void setPerretLabel(); /** sets the transition probability used in the Kashima kernel. note that for performance reasons this function does not use the descriptors of DataContainer to store its values but rather uses a native double variable in the bond class. */ void setKashimaPT(double aPt){ pt = aPt; } /** returns the transition probability used in the Kashima kernel. */ double getKashimaPT(){ return( pt ); } /** returns true if the bond has been marked with setFlag(). flag can be removed with unsetFlag(). */ bool hasFlag(){ return flag; } /** sets 'flag' to true */ void setFlag(){ flag = true; } /** sets 'flag' to false */ void unsetFlag(){ flag = false; } /** returns true if the bond has been marked with setFlagOriginal(). flag can be removed with unsetFlagOriginal(). */ bool hasFlagOriginal(){ return flagOriginal; } /** sets 'flagOriginal' to true */ void setFlagOriginal(){ flagOriginal = true; } /** sets 'flagOriginal' to false */ void unsetFlagOriginal(){ flagOriginal = false; } /** adds a ring to the set of rings the bond is member of. */ void addRing( Ring* aRing ){ if( !hasRing( aRing ) ){ rings.push_back( aRing ); } } /** returns true if the bond is member of the given ring. */ bool hasRing( Ring* aRing ); /** returns true if the bond is member of any ring. */ bool hasRing(){ if( numRings() > 0 ){ return( true ); }else{ return( false );};} /** returns the number of rings the bond is member of. */ int numRings(){ return( rings.size() ); } /** sets the stereo value. */ void setStereo( int a ){ stereo = a; } /** returns the stereo value. */ int getStereo(){ return( stereo ); } /** sets the topology value. */ void setTopology( int a ){ topology = a; } /** returns the topology value. */ int getTopology(){ return( topology ); } /** sets the reaction center value. */ void setReactionCenter( int a ){ reactionCenter = a; } /** returns the reaction center value. */ int getReactionCenter(){ return( reactionCenter ); } /** sets the 'not used' value. */ void setNotUsed( int a ){ notUsed = a; } /** returns the 'not used' value. */ int getNotUsed(){ return( notUsed ); } //@} ///@name Bond manipulation functions //@{ /** reverses the bond direction. */ void reverse(); /** since bonds are represented directional, returns the reciprocal bond. */ Bond* getReverse() throw( CError ); /** NO DOCUMENTATION */ void hideToFrom(); /** NO DOCUMENTATION */ void restoreToFrom(); //@} ///@name Output functions //@{ /** returns a string description of the bond ( = string description of source and target atoms and bond label). */ string toString(); /** returns a short string description of the bond ( = short string description of source and target atoms and bond label). */ string toStringShort(); /** writes a string description of the bond to cout. */ void describe(); /** writes a short string description of the bond to cout. */ void describeShort(); //@} ///@name Iterators //@{ /** iterator on the first element of the set of rings the bond is member of. */ vector::iterator beginRing(){ return( rings.begin() );} /** iterator on the last element of the set of rings the bond is member of. */ vector::iterator endRing(){ return( rings.end() );} //@} // ****************************** // // **** DEPRECATED FUNCTIONS **** // // ****************************** // //string toStringBFSVector(); //Ring* getRingBFS( vector* toVisit, Bond* init = NULL ); private: // Private attributes /** Source atom of this bond. */ Atom *source; /** Target atom of this bond. */ Atom *target; /** Label of the bond. Possible values are defined in constants.h */ int label; /** Perret labeling. */ int perretLabel; /** stereo information. one of: if single bond: STEREONOT, STEREOUP, STEREOEITHER, STERODOWN if double bond: STEROCISTRANSNOT, STEROCISTRANS */ int stereo; /** not used field (xxx in bond mol files). */ int notUsed; /** topology (rrr in bond mol files). DO NOT USE, just for storing informations contained in mol files. */ int topology; /** reaction center (ccc in mol files). */ int reactionCenter; /** transition probability (used in Kashima Kernel). */ double pt; /** flag which can be set using setFlag() and unset using unsetFlag(). */ bool flag; /** flag which can be set using setFlagOriginal() and unset using unsetFlagOriginal(). this flag is set if the bond is present in the original molecule file (backward bonds don't). */ bool flagOriginal; /** atom Descriptors. Intended to store source and target atom of the bond. may include other descriptors via the addAtomDescriptor() function. */ //map< string, Descriptor< Atom >* > atomDescriptors; /** this vector is used to store paths in a BFS search */ //vector BFSVector; /** vector of rings the bond is member of. */ vector rings; }; #endif Rchemcpp/src/chemcpp/src/node.cpp0000644000175000017500000000503412116123101017357 0ustar klambaueklambaue/**************************************************************************************** node.cpp ------------ copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #include "node.h" Node::Node() : DataContainer() { } Node::Node( const DataContainer& aDataContainer ) : DataContainer( (DataContainer&) aDataContainer ) { } Node::~Node() { } void Node::setLabel( string aString ){ label = aString; } double Node::getKashimaPQ( bool silentError ){ if( !flagHasPq ){ stringstream out; out << "No value was defined for pq in atom " << toStringShort() << " "; CError e( ERRORNA, out.str() ); if( !silentError ){ e.describe(); } throw(e); } return( pq ); } double Node::getKashimaPS( bool silentError ){ if( !flagHasPs ){ stringstream out; out << "No value was defined for ps in atom " << toStringShort() << " "; CError e( ERRORNA, out.str() ); if( !silentError ){ e.describe(); } throw(e); } return( ps ); } string Node::toStringShort(){ stringstream out; out << "Node " << getLabel() << endl; return( out.str() ); } /*! \fn Node::resetFlagHasPQ() the node has no pq value */ //void Node::resetFlagHasPQ() //{ // flagHasPq = false; //} /*! \fn Node::setFlagHasPQ( bool ) */ //void Node::setFlagHasPQ( bool aBool ) //{ // flagHasPQ = aBool //} Rchemcpp/src/chemcpp/src/constant.cpp0000644000175000017500000000265212116123101020266 0ustar klambaueklambaue/**************************************************************************************** constants.cpp --------------- copyright : (C) 2012 Mahr Michael ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #include "constant.h" std::string GETChemcpppath() { return getenv("RCHEMCPPPATH"); } Rchemcpp/src/chemcpp/src/stringutils.h0000644000175000017500000001153512116123101020471 0ustar klambaueklambaue/**************************************************************************************** stringutils.h ----------------- copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef STRINGUTILS_H #define STRINGUTILS_H #include #include #include typedef unsigned int uint; using namespace std; /**Static functions to be used to process strings @author Dr Jean-Luc Perret (luc@kuicr.kyoto-u.ac.jp), Kyoto University, Japan @version 0.3 @date 17 Jan 2004 */ class StringUtils { public: /** splits a string into a vector using a string delimiter. */ static int Split( const string input, const string delimiter, vector& results ); /** merges words in a container< string >, inserting separator. */ static string mergeWords( vector& words, const string separator ); /** transforms a string into an int using a stringstream. */ static int toInt( const string input ); /** transforms a string into a float using a stringstream. */ static float toFloat( const string input ); /** transforms a int into an string using a stringstream. */ static string toString( const int input ); /** transforms a float into an string using a stringstream. */ static string toString( const float input ); /** removes all spaces in a string. \todo StringUtils::rmSpace can be improved for more efficiency !!! */ static string rmSpace( const string input ); /** removes all tail spaces in a string. */ static string rmTailSpace( const string input ); /** "chomp" the string, i.e., removes the last character if it is 'end of line'. */ static string chomp(string inString); /** replaces slash by underscore in the string. */ static string slashToUnderscore( const string input ); /** returns the content of a field starting at position start and of length maxlength in string aString. */ static string field(string aString, int start, int maxlength); /** completes aString with aChar so that the total length is exactly equal to aLength. If aString is longer, it gets cropped. */ static string fill( string aString, int aLength, string aChar = " " ); /** prepends aChar to aString so that the total length is exactly equal to aLength. If aString is longer, it gets cropped. */ static string preFill( string aString, int aLength, string aChar = " " ); /** prepends aChar to anInt so that the total length is exactly equal to aLength. If aString is longer, it gets cropped. */ static string preFill( int anInt, int aLength, string aChar = " " ); /** prepends aChar to aFloat so that the total length is exactly equal to aLength. If aString is longer, it gets cropped. */ static string preFill( float aFloat, int aLength, string aChar = " " ); /** returns the directory portion of a full path. */ static string getPath( string aLocation ); /** returns the extension of a file given a full path. */ static string getExtension( string aLocation ); /** returns the path without extension. */ static string getNoExtension( string aLocation ); /** returns the filename portion of a full path (including extension). */ static string getFileName( string aLocation ); /** returns aLength characters at the right end of aString. */ static string right( string aString, uint aLength ); /** converts aString to Upper case. */ static string toUpper( string aString ); /** converts aString to Lower case. */ static string toLower( string aString ); /** returns the string starting at the 1st non space character of aString. */ static string getFirstNonSpace( string aString ); //StringUtils(); //~StringUtils(); }; #endif Rchemcpp/src/subtreehelper.h0000644000175000017500000000561112116123101016543 0ustar klambaueklambaue/**************************************************************************************** subtreehelper.cpp ----------------- copyright : (C) 2012 Mahr Michael email : michael.mahr@gmail.com rearranged from: copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef SUBTREEHELPER_H #define SUBTREEHELPER_H #include #include #include "Rmoleculeset.h" #include //-------- //external //-------- void gramSubtree_self(SEXP s, double lambda, int depthMax, bool filterTotters, bool branchFlag, bool branchKernelUntilN, bool silentMode); void gramSubtree_test(SEXP s, double lambda, int depthMax, bool filterTotters, bool branchFlag, bool branchKernelUntilN, bool silentMode); void initialize_tuples(int nmax); //-------- //internal //-------- static vector< vector< vector< vector > > > tuples; class Nextatom { // a class to describe the list of neighbors of a given atoms that share the same bond and atom type public: int b; // the bond type string a; // the atom type vector i; // the list of atom indices }; // We also need a basic ordering for this class in order to sort and have fast matching bool operator<(const Nextatom& a, const Nextatom& b) { if (a.a < b.a) return true; if (a.a > b.a) return false; return (a.b < b.b); } void initialize_extended(MoleculeSet* aSet, vector< vector< vector< Nextatom> > > * edges); double subTreeKernel( Molecule* mol1, Molecule* mol2, vector< vector< Nextatom> >* edges1, vector< vector< Nextatom> >* edges2, int depthMax, double lambda, bool noTotters, bool untilFlag, bool branchFlag); #endif Rchemcpp/src/Makevars0000644000175000017500000000053012055067220015223 0ustar klambaueklambaue## Use the R_HOME indirection to support installations of multiple R version PKG_LIBS = `$(R_HOME)/bin/Rscript -e "Rcpp:::LdFlags()"` -L./ -lchemcpp -Wl,-rpath,$(R_PACKAGE_DIR)/libs PKG_CPPFLAGS = -I./chemcpp/src all: sublib sublib: @(cd chemcpp/src && CXX="$(CXX)" CXXFLAGS="$(CXXFLAGS)" MkInclude="$(MkInclude)" $(MAKE) all) || exit 1; Rchemcpp/src/spectrum3Dhelper.h0000644000175000017500000000566312116123101017132 0ustar klambaueklambaue/**************************************************************************************** spectrumhelper.h ---------------- copyright : (C) 2012 Mahr Michael email : michael.mahr@gmail.com rearranged from: copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef SPECTRUM3DHELPER_H #define SPECTRUM3DHELPER_H #include #include #include "Rmoleculeset.h" //--------- // external //--------- void gramSpectrum3D_self(SEXP s, int depthMax, int kernelType, int nBins, double distMin, double distMax, bool silentMode); void gramSpectrum3D_test(SEXP s, int depthMax, int kernelType, int nBins, double distMin, double distMax, bool silentMode); //--------- // internal //--------- typedef struct{ Atom* endAtom; Atom* startAtom; Atom* previousAtom; double weight; }path3D; typedef struct{ vector path_pointers; string molName; int molInd; }pathsInMol3D; void gramComputeSpectrum3D_self( MoleculeSet*, int, int, int, vector< pathsInMol3D >*, vector*, vector*, bool silentMode = false); void gramComputeSpectrum3D_test( MoleculeSet*, MoleculeSet*, int, int, int, vector< pathsInMol3D >*, vector< pathsInMol3D >*, vector*, vector*, bool silentMode = false); void init_path_pointers3D( MoleculeSet*, vector< pathsInMol3D >*, string); void updatePaths3D( MoleculeSet*, string, int, vector< pathsInMol3D >*, vector< pathsInMol3D >*, int, int); void updateGram3D_self( MoleculeSet*, vector< pathsInMol3D >*, int); void updateGram3D_test( MoleculeSet*, MoleculeSet*, vector< pathsInMol3D >*, vector< pathsInMol3D >*, int); void updateSelfKernel3D(MoleculeSet* , vector< pathsInMol3D >*, int); #endif Rchemcpp/src/Rmoleculeset.cpp0000644000175000017500000000640612116123101016673 0ustar klambaueklambaue #include "Rmoleculeset.h" #include #include #include RCPP_MODULE(Rmoleculeset){ Rcpp::class_( "Rmoleculeset") .constructor() // .constructor() .method( "addSD", &Rmoleculeset::addSD, "read molecule" ) .method( "addSD2", &Rmoleculeset::addSD2, "read molecule" ) //Rcpp::List::create( Rcpp::_["aFileName"] Rcpp::_["genericAtomTypeFlag"] = 0, Rcpp::_["beginMolecule"] = -1, Rcpp::_["endMolecule"] = -1 ), .method( "addKCF", &Rmoleculeset::addKCF, "read molecule" ) .method( "addKCF2", &Rmoleculeset::addKCF2, "read molecule" ) .method( "numMolecules", &Rmoleculeset::numMolecules, "return the number of molecules in the set") .method( "hideHydrogens", &Rmoleculeset::hideHydrogens, "hide hydrogens") .method( "setMorganLabels", &Rmoleculeset::setMorganLabels, "set Morgan labels") .method( "setKashimaKernelParam", &Rmoleculeset::setKashimaKernelParam, "set Kashima Kernel parameter") .method( "initializeGram", &Rmoleculeset::initializeGram, "initialize the Gram matrix") .method( "initializeSelfKernel", &Rmoleculeset::initializeSelfKernel, "initialize the Self Kernel") .method( "setComparisonSetSelf", &Rmoleculeset::setComparisonSetSelf, "set comparison set to be the moleculeset itself;") .method( "setComparisonSetCopy", &Rmoleculeset::setComparisonSetCopy, "set comparison set; argument-object is cloned! (and disposed of at destruction)") .method( "getComparisonSet", &Rmoleculeset::getComparisonSet, "Pointer to the current comparison set; managed by chemcpp; Object becomes invalid, when the moleculeset runs ouf of scope.") //new method .method( "normalizeTanimoto", &Rmoleculeset::normalizeTanimoto, "normalize Tanimoto") .method( "normalizeTanimotoMinMax", &Rmoleculeset::normalizeTanimotoMinMax, "normalize TanimotoMinMax") .method( "normalizeGram", &Rmoleculeset::normalizeGram, "normalize Gram") .method( "noTottersTransform", &Rmoleculeset::noTottersTransform, "transform to avoid totters") .method( "readPartialCharges", &Rmoleculeset::readPartialCharges, "add partial charges from file") .method( "setMorganChargesLabels", &Rmoleculeset::setMorganChargesLabels, "set Morgan charges threshold") .method( "gramCompute3D", &Rmoleculeset::gramCompute3D, "compute 3d gram") .method( "gramCompute", &Rmoleculeset::gramCompute, "compute gram") .method( "addMoleculeCopy", &Rmoleculeset::addMoleculeCopy, "Clone Molecule and add it") .method( "getGram", &Rmoleculeset::getGram, "Clone Molecule and add it") .method( "getGramNormal", &Rmoleculeset::getGramNormal, "Clone Molecule and add it") //currently not expicitly necessary .method( "atomsLabelsListing", &Rmoleculeset::atomsLabelsListing, "get atom label listings") .method( "bondsListing", &Rmoleculeset::bondsListing, "get bond listings") .method( "normalizeTanimoto_raw", &Rmoleculeset::normalizeTanimoto_raw, "normalize Tanimoto") .method( "normalizeGram_raw", &Rmoleculeset::normalizeGram_raw, "normalize Gram") .method( "writeGramMatrix", &Rmoleculeset::writeGramMatrix, "write Gram matrix to file") .method( "writeSelfKernelList", &Rmoleculeset::writeSelfKernelList, "write self kernel list") //Extension 1.0.7 .method( "getMolByIndex", &Rmoleculeset::getMolByIndex, "get Molecule by Index (zero-based)") ; } Rchemcpp/src/Relements.h0000644000175000017500000000031112116123101015620 0ustar klambaueklambaue #ifndef RELEMENTS_H #define RELEMENTS_H #include #include extern Elements elements; //--------- // external //--------- void loadGramAtoms( std::string aFileName ); #endif Rchemcpp/src/Relements.cpp0000644000175000017500000000027612116123101016165 0ustar klambaueklambaue #include "Relements.h" void loadGramAtoms( std::string aFileName ){ elements.loadGramAtoms( aFileName ); } RCPP_MODULE(Relements) { Rcpp::function( "loadGramAtoms", &loadGramAtoms ); } Rchemcpp/src/spectrumhelper.h0000644000175000017500000000563512116123101016742 0ustar klambaueklambaue/**************************************************************************************** spectrumhelper.h ---------------- copyright : (C) 2012 Mahr Michael email : michael.mahr@gmail.com rearranged from: copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #ifndef SPECTRUMHELPER_H #define SPECTRUMHELPER_H #include #include #include "Rmoleculeset.h" //--------- // external //--------- void gramSpectrum_self(SEXP s, int depthMax, int kernelType, double kernelParam, bool onlyN, bool silentMode); void gramSpectrum_test(SEXP s, int depthMax, int kernelType, double kernelParam, bool onlyN, bool silentMode); //--------- // internal //--------- typedef struct{ Atom* endAtom; Atom* previousAtom; double weight; }path; typedef struct{ vector path_pointers; string molName; int molInd; }pathsInMol; void gramComputeSpectrum_self( MoleculeSet*, int, int, int, double, vector< pathsInMol >*, vector*, vector*, bool, bool silentMode = false); void gramComputeSpectrum_test( MoleculeSet*, MoleculeSet*, int, int, int, double, vector< pathsInMol >*, vector< pathsInMol >*, vector*, vector*, bool, bool silentMode = false); void init_path_pointers( MoleculeSet*, vector< pathsInMol >*, string, int); void updatePaths( MoleculeSet*, string, int, vector< pathsInMol >*, vector< pathsInMol >*, int, int); void updateGram_self( MoleculeSet*, vector< pathsInMol >*, int, double, int); void updateGram_test( MoleculeSet*, MoleculeSet*, vector< pathsInMol >*, vector< pathsInMol >*, int, double, int); void updateSelfKernel(MoleculeSet* , vector< pathsInMol >*, int, double, int); #endif Rchemcpp/src/spectrumhelper.cpp0000644000175000017500000006047212116123101017275 0ustar klambaueklambaue/**************************************************************************************** spectrumhelper.cpp ------------------ copyright : (C) 2012 Mahr Michael email : michael.mahr@gmail.com rearranged from: copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #include "spectrumhelper.h" // ----------------------------------------------- // ------------- UTILITY FUNCTIONS --------------- // ----------------------------------------------- //-------- //External //-------- void gramSpectrum_self(SEXP s, int depthMax, int kernelType, double kernelParam, bool onlyN, bool silentMode){ std::string rtypename("Rcpp_Rmoleculeset"); Rcpp::S4 s4obj( s ); if ( !s4obj.is( rtypename.c_str() ) ) { Rf_error( (std::string("object is not of the type ")+rtypename).c_str() ); } Rcpp::Environment env( s4obj ); Rcpp::XPtr xptr( env.get(".pointer") ); Rmoleculeset *aSet = static_cast (R_ExternalPtrAddr(xptr)); //--------- //--------- //--------- int depth = -1; vector< pathsInMol > molPaths; // 2 - get the database description // -------------------------------- // i.e. fill in atom and bond types vectors vector atomLabels; vector bondTypes; atomLabels = aSet->atomsLabelsListing(); bondTypes = aSet->bondsListing(); if( !silentMode ){ for( int i = 0 ; i < atomLabels.size() ; i++ ){ Rcpp::Rcout << "atom type no " << i+1 << " ; atomic number = " << atomLabels[i] << endl; } for( int i = 0 ; i < bondTypes.size() ; i++ ){ Rcpp::Rcout << "bond type no " << i+1 << " ; bond type = " << bondTypes[i] << endl; } } gramComputeSpectrum_self( aSet, depth, depthMax, kernelType, kernelParam, &molPaths, &atomLabels, &bondTypes, onlyN, silentMode); if( !silentMode ){ Rcpp::Rcout << "gramComputeSpectrum (self) OK" << endl; } } void gramSpectrum_test(SEXP s, int depthMax, int kernelType, double kernelParam, bool onlyN, bool silentMode){ std::string rtypename("Rcpp_Rmoleculeset"); Rcpp::S4 s4obj( s ); if ( !s4obj.is( rtypename.c_str() ) ) { Rf_error( (std::string("object is not of the type ")+rtypename).c_str() ); } Rcpp::Environment env( s4obj ); Rcpp::XPtr xptr( env.get(".pointer") ); Rmoleculeset *aSet = static_cast (R_ExternalPtrAddr(xptr)); /* std::string rtypename2("Rcpp_Rmoleculeset"); Rcpp::S4 s4obj2( s2 ); if ( !s4obj2.is( rtypename2.c_str() ) ) { Rf_error( (std::string("object is not of the type ")+rtypename2).c_str() ); } Rcpp::Environment env2( s4obj2 ); Rcpp::XPtr xptr2( env2.get(".pointer") ); Rmoleculeset *aSet2 = static_cast (R_ExternalPtrAddr(xptr2)); */ Rmoleculeset *aSet2 = aSet->getComparisonSet2(); //--------- //--------- //--------- int depth = -1; vector< pathsInMol > molPaths; vector< pathsInMol > molPaths_test; // 2 - get the database description // -------------------------------- // i.e. fill in atom and bond types vectors vector atomLabels; vector bondTypes; atomLabels = aSet->atomsLabelsListing(); bondTypes = aSet->bondsListing(); if( !silentMode ){ for( int i = 0 ; i < atomLabels.size() ; i++ ){ Rcpp::Rcout << "atom type no " << i+1 << " ; atomic number = " << atomLabels[i] << endl; } for( int i = 0 ; i < bondTypes.size() ; i++ ){ Rcpp::Rcout << "bond type no " << i+1 << " ; bond type = " << bondTypes[i] << endl; } } // 3 - compute the gram matrix // ---------------------------- gramComputeSpectrum_test( aSet, aSet2, depth, depthMax, kernelType, kernelParam, &molPaths, &molPaths_test, &atomLabels, &bondTypes, onlyN, silentMode); if( !silentMode ){ Rcpp::Rcout << "gramComputeSpectrum (test) OK" << endl; } } //-------- //internal //-------- //---------------------------------------------------------------------------------------------------------------------------------------------------------------- void gramComputeSpectrum_self(MoleculeSet* aSet, int depth, int depthMax, int kernelType, double kernelParam, vector< pathsInMol >* molPaths, vector* atomLabels, vector* bondTypes, bool onlyN, bool silentMode){ //---------------------------------------------------------------------------------------------------------------------------------------------------------------- vector< pathsInMol> molPaths_NEW; depth = depth + 1; if (depth == 0){ // depth = 0 --> initialization step // for each kind of atoms, find paths starting points and launch the recursion for(int i = 0 ; i < atomLabels->size() ; i++){ molPaths->clear(); if( !silentMode ){ Rcpp::Rcout << " \t finding paths starting from atoms labeled = " << (*atomLabels)[i] << endl; } init_path_pointers(aSet, molPaths, (*atomLabels)[i], kernelType); if( !onlyN ){ updateGram_self(aSet,molPaths, kernelType, kernelParam, depth); updateSelfKernel(aSet,molPaths, kernelType, kernelParam, depth); } if(depthMax == 0){ //...then, just include the weights in the gram matrix if(onlyN){ updateGram_self(aSet,molPaths, kernelType, kernelParam, depth); updateSelfKernel(aSet,molPaths, kernelType, kernelParam, depth); } continue; }else{ gramComputeSpectrum_self(aSet, depth, depthMax, kernelType, kernelParam, molPaths, atomLabels, bondTypes, onlyN, silentMode); } } } // end "if(depth == 0)" else{ // if depth >= 1 : we can try to extend the paths stored in path_pointers for(int j = 0 ; j < atomLabels->size() ; j++){ for(int k = 0 ; k < bondTypes->size() ; k++){ // for each atom type/bond type : extend the paths updatePaths(aSet, (*atomLabels)[j], (*bondTypes)[k], molPaths, &molPaths_NEW, kernelType, depth); int nonEmpty; nonEmpty = molPaths_NEW.size(); if( nonEmpty > 0 ){ // i.e : if at least 1 vector of paths is non-empty, we go further in the recursion if( !onlyN ){ updateGram_self(aSet, &molPaths_NEW, kernelType, kernelParam, depth); updateSelfKernel(aSet,&molPaths_NEW, kernelType, kernelParam, depth); } if(depth == depthMax){ if( onlyN ){ updateGram_self(aSet, &molPaths_NEW, kernelType, kernelParam, depth); updateSelfKernel(aSet,&molPaths_NEW, kernelType, kernelParam, depth); } continue; }else{ gramComputeSpectrum_self(aSet, depth, depthMax, kernelType, kernelParam, &molPaths_NEW, atomLabels, bondTypes, onlyN, silentMode); } }// end "if(nonEmpty > 0)" } } } // end "else if(depth == 0)" } //---------------------------------------------------------------------------------------------------------------------------------------------------------------- void gramComputeSpectrum_test(MoleculeSet* aSet, MoleculeSet* aSet2, int depth, int depthMax, int kernelType, double kernelParam, vector< pathsInMol >* molPaths, vector< pathsInMol >* molPaths_test, vector* atomLabels, vector* bondTypes, bool onlyN, bool silentMode){ //---------------------------------------------------------------------------------------------------------------------------------------------------------------- vector< pathsInMol > molPaths_NEW; vector< pathsInMol > molPaths_test_NEW; depth = depth + 1; if (depth == 0){ // depth = 0 --> initialization step // for each kind of atoms, find paths starting points and launch the recursion for(int i = 0 ; i < atomLabels->size() ; i++){ molPaths->clear(); molPaths_test->clear(); if( !silentMode ){ Rcpp::Rcout << "\t - finding paths starting from atoms labeled = " << (*atomLabels)[i] << endl; } init_path_pointers(aSet, molPaths, (*atomLabels)[i], kernelType); init_path_pointers(aSet2, molPaths_test, (*atomLabels)[i], kernelType); if( !onlyN ){ updateGram_test(aSet, aSet2, molPaths, molPaths_test, kernelType, kernelParam, depth); updateSelfKernel(aSet,molPaths, kernelType, kernelParam, depth); updateSelfKernel(aSet2,molPaths_test, kernelType, kernelParam, depth); } if(depthMax == 0){ if( onlyN ){ updateGram_test(aSet, aSet2, molPaths, molPaths_test, kernelType, kernelParam, depth); updateSelfKernel(aSet,molPaths, kernelType, kernelParam, depth); updateSelfKernel(aSet2,molPaths_test, kernelType, kernelParam, depth); } continue; }else{ gramComputeSpectrum_test(aSet, aSet2, depth, depthMax, kernelType, kernelParam, molPaths, molPaths_test, atomLabels, bondTypes, onlyN, silentMode); } } } // end "if (depth == 0)" else{ // if depth >= 1, we can try to extend the paths previously stored in path_pointers for(int j = 0 ; j < atomLabels->size() ; j++){ for(int k = 0 ; k < bondTypes->size() ; k++){ // for each atom type/bond type : extend the paths updatePaths(aSet, (*atomLabels)[j], (*bondTypes)[k], molPaths, &molPaths_NEW, kernelType, depth); updatePaths(aSet2, (*atomLabels)[j], (*bondTypes)[k], molPaths_test, &molPaths_test_NEW, kernelType, depth); int nonEmpty1, nonEmpty2; nonEmpty1 = molPaths_NEW.size(); nonEmpty2 = molPaths_test_NEW.size(); if(nonEmpty1 > 0 || nonEmpty2 > 0){ // i.e : if at least 1 vector of paths is non-empty IN THE TWO SETS, we go further in the recursion if( !onlyN ){ updateGram_test(aSet, aSet2, &molPaths_NEW, &molPaths_test_NEW, kernelType, kernelParam, depth); updateSelfKernel(aSet,&molPaths_NEW, kernelType, kernelParam, depth); updateSelfKernel(aSet2,&molPaths_test_NEW, kernelType, kernelParam, depth); } if(depth == depthMax){ if( onlyN ){ updateGram_test(aSet, aSet2, &molPaths_NEW, &molPaths_test_NEW, kernelType, kernelParam, depth); updateSelfKernel(aSet,&molPaths_NEW, kernelType, kernelParam, depth); updateSelfKernel(aSet2,&molPaths_test_NEW, kernelType, kernelParam, depth); } continue; }else{ gramComputeSpectrum_test(aSet, aSet2, depth, depthMax, kernelType, kernelParam, &molPaths_NEW, &molPaths_test_NEW, atomLabels, bondTypes, onlyN, silentMode); } } // end if(nonEmpty1 > 0 || nonEmpty2 > 0) } } } // end "else if(depth == 0)" } //************************************************************************************************ // ----------------------------------------------------------------------------------------------- void init_path_pointers(MoleculeSet* aSet, vector< pathsInMol >* molPaths, string atomLabel,int kernelType ){ // ----------------------------------------------------------------------------------------------- // --> fill in "path_pointers" for starting atoms labeled as "atomLabel" path aPath; pathsInMol aPathInMol; int ind = 0; vector::iterator aMol; // for each molecule of the set : for(aMol = (*aSet).begin() ; aMol != (*aSet).end() ; aMol++ ){ bool first = true; aPathInMol.path_pointers.clear(); vector::iterator anAtom; // for each atom of the molecule for(anAtom = (*aMol)->beginAtom() ; anAtom != (*aMol)->endAtom() ; anAtom++){ // if the atomic number corresponds to the one needed (i.e. "atomNumb") if((*anAtom)->getMorganLabel(true) == atomLabel){ if(first == true){ if(kernelType == 3){ // i.e., marginalized kernel approximation aPath.weight = (*anAtom)->getKashimaPS(); } else{ aPath.weight = 1; } if( aPath.weight > 0){ aPath.endAtom = *anAtom; aPathInMol.path_pointers.push_back(aPath); aPathInMol.molName = (*aMol)->getName(); aPathInMol.molInd = ind; first = false; } } // end "if(first == true)" else{ if(kernelType == 3){ aPath.weight = (*anAtom)->getKashimaPS(); } else{ aPath.weight = 1; } if( aPath.weight > 0){ aPath.endAtom = *anAtom; aPathInMol.path_pointers.push_back(aPath); } } // end "else" } // end "if((*anAtom)->getMorganLabel(true) == atomLabel)" } if(aPathInMol.path_pointers.size() > 0 ){ (*molPaths).push_back(aPathInMol); } ind++; } } // ************************************************************************************************************************************************ // ------------------------------------------------------------------------------------------------------------------------------------------------ void updatePaths(MoleculeSet *aSet, string atomLabel, int bondType, vector< pathsInMol >* molPaths, vector< pathsInMol >* molPaths_NEW, int kernelType, int depth){ // ------------------------------------------------------------------------------------------------------------------------------------------------ // --> extends path stored in "path_pointers" to atom of type "atomNumb". Stores the results in "path_pointers_NEW" path aPath; pathsInMol aPathInMol; // object initialized for each molecule present in the input molPaths, possibly inserted in the output molPaths_NEW molPaths_NEW->clear(); // for each molecule in molPaths for(int i = 0 ; i < molPaths->size() ; i++){ bool first = true; aPathInMol.path_pointers.clear(); //for each previous path pointer for(int j = 0 ; j < (*molPaths)[i].path_pointers.size() ; j++ ){ map::iterator aBond; // along each bond of the path's ending atom : for(aBond = (*molPaths)[i].path_pointers[j].endAtom->beginBond() ; aBond != (*molPaths)[i].path_pointers[j].endAtom->endBond() ; aBond++ ){ if( ( ( depth > 1) && ( aBond->second->getLabel() == bondType ) && ( aBond->first->getMorganLabel(true) == atomLabel ) && ( aBond->first->getId() != (*molPaths)[i].path_pointers[j].previousAtom->getId() ) ) // if depth > 1 : we have to prevent tottering paths || ( (depth == 1) && ( aBond->second->getLabel() == bondType ) && ( aBond->first->getMorganLabel(true) == atomLabel ) ) // if depth = 1 : path can't totter ) { if(first == true){ if(kernelType == 3){ // i.e., marginalized kernel approximation if(depth == 1) { aPath.weight = (*molPaths)[i].path_pointers[j].weight * aBond->second->getKashimaPT(); } // if depth = 1 : paths can't totter else{ aPath.weight = (*molPaths)[i].path_pointers[j].weight * ( 1.0 - (*molPaths)[i].path_pointers[j].endAtom->getKashimaPQ() ) / ( (*molPaths)[i].path_pointers[j].endAtom->numBonds() - 1 ); } // if depth > 1 : modify weight for the non-tottering paths } else{ aPath.weight = 1; } if(aPath.weight > 0 ){ aPath.endAtom = aBond->first; aPath.previousAtom = (*molPaths)[i].path_pointers[j].endAtom; aPathInMol.path_pointers.push_back(aPath); aPathInMol.molName = (*molPaths)[i].molName; aPathInMol.molInd = (*molPaths)[i].molInd; first = false; } } // end "if(first == true)" else{ if(kernelType == 3){ if(depth == 1){ aPath.weight = (*molPaths)[i].path_pointers[j].weight * aBond->second->getKashimaPT(); } else{ aPath.weight = (*molPaths)[i].path_pointers[j].weight * ( 1.0 - (*molPaths)[i].path_pointers[j].endAtom->getKashimaPQ() ) / ( (*molPaths)[i].path_pointers[j].endAtom->numBonds() - 1 ); } } else{ aPath.weight = 1; } if(aPath.weight > 0 ){ aPath.endAtom = aBond->first; aPath.previousAtom = (*molPaths)[i].path_pointers[j].endAtom; aPathInMol.path_pointers.push_back(aPath); } } } } } if(aPathInMol.path_pointers.size() > 0){ (*molPaths_NEW).push_back(aPathInMol); } } } // *********************************************************************************************************************** // ----------------------------------------------------------------------------------------------------------------------- void updateGram_self(MoleculeSet *aSet, vector< pathsInMol >* molPaths, int kernelType, double kernelParam, int depth){ // ----------------------------------------------------------------------------------------------------------------------- vector sumWeight; if(kernelType == 3){ // i.e., marginalized approximation --> fill in a vector summing the weight of the paths for each molecule for(int i = 0 ; i < molPaths->size() ; i++){ sumWeight.push_back(0.0); for(int j = 0 ; j < (*molPaths)[i].path_pointers.size() ; j++ ){ if( depth >= 1 && (*molPaths)[i].path_pointers[j].endAtom->numBonds() == 1 ) sumWeight[i] += (*molPaths)[i].path_pointers[j].weight; else sumWeight[i] += (*molPaths)[i].path_pointers[j].weight * (*molPaths)[i].path_pointers[j].endAtom->getKashimaPQ(); } } } // update the gram matrix double update, update2; for(int i = 0 ; i < molPaths->size() ; i++){ for(int j = i ; j < molPaths->size() ; j++){ // compute the update factor depending on the kernel type switch(kernelType){ case 0 : // 'spectrum' update = (*molPaths)[i].path_pointers.size() * (*molPaths)[j].path_pointers.size(); break; case 1 : // tanimoto update = 1.0; break; case 2 : // min/max tanimoto update = min( (*molPaths)[i].path_pointers.size(), (*molPaths)[j].path_pointers.size() ); // NB : for the min/max kernel we have to keep the maximum values too, for further normalization update2 = max( (*molPaths)[i].path_pointers.size(), (*molPaths)[j].path_pointers.size() ); (*aSet).addToGramNormal( (*molPaths)[i].molInd , (*molPaths)[j].molInd, update2); if( i!= j){ (*aSet).addToGramNormal( (*molPaths)[j].molInd , (*molPaths)[i].molInd, update2); } break; case 3 : // marginalized approximation update = sumWeight[i] * sumWeight[j]; break; case 4 : // lambda^k update = (*molPaths)[i].path_pointers.size() * (*molPaths)[j].path_pointers.size() * pow(kernelParam, depth+1); // NB : we put depth+1 because 0 edge <-> 1 atom => x lambda // Rq : equivalent to case 0 with lambda = 1 break; } // add the update factor to the gram matrix (*aSet).addToGram( (*molPaths)[i].molInd , (*molPaths)[j].molInd, update); if( i!= j){ (*aSet).addToGram( (*molPaths)[j].molInd , (*molPaths)[i].molInd, update); } } } } //************************************************************************** //-------------------------------------------------------------------------- void updateGram_test(MoleculeSet *aSet, MoleculeSet *aSet2, vector< pathsInMol >* molPaths, vector< pathsInMol >* molPaths_test, int kernelType, double kernelParam, int depth){ //-------------------------------------------------------------------------- vector sumWeight, sumWeight_test; if(kernelType == 3){ // i.e., marginalized approximation --> fill in a vector summing the weight of the paths for each molecule (train + test) for(int i = 0 ; i < molPaths->size() ; i++){ sumWeight.push_back(0.0); for(int k = 0 ; k < (*molPaths)[i].path_pointers.size() ; k++){ if( (*molPaths)[i].path_pointers[k].endAtom->numBonds() ==1 && depth >= 1) // NB : no-tottering implementation sumWeight[i] += (*molPaths)[i].path_pointers[k].weight; else sumWeight[i] += (*molPaths)[i].path_pointers[k].weight * (*molPaths)[i].path_pointers[k].endAtom->getKashimaPQ(); } } for(int i = 0 ; i < molPaths_test->size() ; i++){ sumWeight_test.push_back(0.0); for(int k = 0 ; k < (*molPaths_test)[i].path_pointers.size() ; k++){ if( (*molPaths_test)[i].path_pointers[k].endAtom->numBonds() == 1 && depth >= 1) // NB : no-tottering implementation sumWeight_test[i] += (*molPaths_test)[i].path_pointers[k].weight; else sumWeight_test[i] += (*molPaths_test)[i].path_pointers[k].weight * (*molPaths_test)[i].path_pointers[k].endAtom->getKashimaPQ(); } } } // end "if(kernelType == 3)" // update the gram matrix double update, update2; for(int i = 0 ; i < molPaths->size() ; i++){ for(int j = 0 ; j < molPaths_test->size() ; j++){ // compute the update factor depending on the kernel type switch(kernelType){ case 0 : // spectrum update = (*molPaths)[i].path_pointers.size() * (*molPaths_test)[j].path_pointers.size(); break; case 1 : // tanimoto update = 1.0; break; case 2 : // min/max tanimoto update = min( (*molPaths)[i].path_pointers.size(), (*molPaths_test)[j].path_pointers.size() ); // NB : for the min/max kernel we have to keep the maximum values too, for further normalization update2 = max( (*molPaths)[i].path_pointers.size(), (*molPaths_test)[j].path_pointers.size() ); (*aSet).addToGramNormal( (*molPaths)[i].molInd , (*molPaths_test)[j].molInd, update2); break; case 3 : // marginalized approximation update = sumWeight[i] * sumWeight_test[j]; break; case 4 : // lambda^k update = (*molPaths)[i].path_pointers.size() * (*molPaths_test)[j].path_pointers.size() * pow(kernelParam, depth+1); // NB : we put depth+1 because 0 edge <-> 1 atom => x lambda // Rq : equivalent to case 0 with lambda = 1 break; } // add the update factor to the gram matrix (*aSet).addToGram( (*molPaths)[i].molInd , (*molPaths_test)[j].molInd, update); } } } // ************************************************************************** // -------------------------------------------------------------------------- void updateSelfKernel(MoleculeSet* aSet, vector< pathsInMol >* molPaths, int kernelType, double kernelParam, int depth){ // -------------------------------------------------------------------------- vector sumWeight; Molecule *aMol; if(kernelType == 3){ // i.e., marginalized approximation --> fill in a vector summing the weight of the paths for each molecule for(int i = 0 ; i < molPaths->size() ; i++){ sumWeight.push_back(0.0); for(int j = 0 ; j < (*molPaths)[i].path_pointers.size() ; j++ ){ if( (*molPaths)[i].path_pointers[j].endAtom->numBonds() == 1 && depth >= 1) // NB : no-tottering implementation sumWeight[i] += (*molPaths)[i].path_pointers[j].weight; else sumWeight[i] += (*molPaths)[i].path_pointers[j].weight * (*molPaths)[i].path_pointers[j].endAtom->getKashimaPQ(); } } } // update the self kernel values double update; for(int i = 0 ; i < molPaths->size() ; i++){ //aMol = (*aSet)[(*molPaths)[i].molName]; //cout << "### MolInd = " << (*molPaths)[i].molInd << endl; aMol = (*aSet)[(*molPaths)[i].molInd]; // compute the update facor depending on the kernel type switch(kernelType){ case 0 : // spectrum update = (*molPaths)[i].path_pointers.size() * (*molPaths)[i].path_pointers.size(); break; case 1 : // tanimoto update = 1.0; break; case 2 : // min/max tanimoto update = (*molPaths)[i].path_pointers.size(); break; case 3 : // marginalized approximation update = sumWeight[i]*sumWeight[i]; break; case 4 : // lambda^k update = (*molPaths)[i].path_pointers.size() * (*molPaths)[i].path_pointers.size() * pow(kernelParam, depth+1); break; } // add the update factor to the self kernel values aMol->addToSelfKernel(update); } } RCPP_MODULE(spectrumhelper) { Rcpp::function( "gramSpectrum_self", &gramSpectrum_self ); Rcpp::function( "gramSpectrum_test", &gramSpectrum_test ); } Rchemcpp/src/spectrum3Dhelper.cpp0000644000175000017500000005026012116123101017456 0ustar klambaueklambaue/**************************************************************************************** spectrumhelper.cpp ------------------ copyright : (C) 2012 Mahr Michael email : michael.mahr@gmail.com rearranged from: copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #include "spectrum3Dhelper.h" // ----------------------------------------------- // ------------- UTILITY FUNCTIONS --------------- // ----------------------------------------------- //-------- //External //-------- void gramSpectrum3D_self(SEXP s, int depthMax, int kernelType, int nBins, double distMin, double distMax, bool silentMode){ std::string rtypename("Rcpp_Rmoleculeset"); Rcpp::S4 s4obj( s ); if ( !s4obj.is( rtypename.c_str() ) ) { Rf_error( (std::string("object is not of the type ")+rtypename).c_str() ); } Rcpp::Environment env( s4obj ); Rcpp::XPtr xptr( env.get(".pointer") ); Rmoleculeset *aSet = static_cast (R_ExternalPtrAddr(xptr)); //--------- //--------- //--------- int depth = -1; vector< pathsInMol3D > molPaths; // 2 - get the database description // -------------------------------- // i.e. fill in atom and bond types vectors vector atomLabels; vector bondTypes; atomLabels = aSet->atomsLabelsListing(); for(int i = 1 ; i <= nBins; i++){ bondTypes.push_back(i); } if( !silentMode ){ for( int i = 0 ; i < atomLabels.size() ; i++ ){ Rcpp::Rcout << "atom type no " << i+1 << " ; atomic number = " << atomLabels[i] << endl; } //for( int i = 0 ; i < bondTypes.size() ; i++ ){ //cout << "bond type no " << i+1 << " ; bond type = " << bondTypes[i] << endl; //} } // 3 - transform the dataset from 2D to 3D // --------------------------------------- // OLD VERSION: distMin and distMin were computed from the data // --> now: specified as parameters //double distMin = 100000.0; //double distMax = 0.0; //aSet.minMaxDistances(&distMin, &distMax); if( !silentMode ){ Rcpp::Rcout << " - distMin = " << distMin << endl; Rcpp::Rcout << " - distMax = " << distMax << endl; Rcpp::Rcout << " - nBins = " << nBins << endl; Rcpp::Rcout << " --> binSize = " << (1.0001*distMax - distMin)/nBins << endl; } aSet->threeDtransform(nBins, distMin, distMax); // vector::iterator aMol; // for(aMol = aSet.begin() ; aMol != aSet.end() ; aMol++ ){ // (*aMol)->describe(); // vector::iterator anAtom; // //vector::iterator aBond; // map::iterator aBond; // for(anAtom = (*aMol)->beginAtom(); anAtom != (*aMol)->endAtom() ; anAtom++ ){ // //for( aBond = (*anAtom)->beginBond(); aBond != (*anAtom)->endBond(); aBond++) // // cout << "bondLabel = " << aBond->second->getLabel() << endl; // cout << "Atom Label after 3D-transfo ; " << (*anAtom)->getMorganLabel() << endl; // } // } // exit(12); // 4 - compute the gram matrix // --------------------------- gramComputeSpectrum3D_self( aSet, depth, depthMax, kernelType, &molPaths, &atomLabels, &bondTypes, silentMode); if( !silentMode ){ Rcpp::Rcout << "gramComputeSpectrum (self) OK" << endl; } } void gramSpectrum3D_test(SEXP s, int depthMax, int kernelType, int nBins, double distMin, double distMax, bool silentMode){ std::string rtypename("Rcpp_Rmoleculeset"); Rcpp::S4 s4obj( s ); if ( !s4obj.is( rtypename.c_str() ) ) { Rf_error( (std::string("object is not of the type ")+rtypename).c_str() ); } Rcpp::Environment env( s4obj ); Rcpp::XPtr xptr( env.get(".pointer") ); Rmoleculeset *aSet = static_cast (R_ExternalPtrAddr(xptr)); /* std::string rtypename2("Rcpp_Rmoleculeset"); Rcpp::S4 s4obj2( s2 ); if ( !s4obj2.is( rtypename2.c_str() ) ) { Rf_error( (std::string("object is not of the type ")+rtypename2).c_str() ); } Rcpp::Environment env2( s4obj2 ); Rcpp::XPtr xptr2( env2.get(".pointer") ); Rmoleculeset *aSet2 = static_cast (R_ExternalPtrAddr(xptr2)); */ Rmoleculeset *aSet2 = aSet->getComparisonSet2(); //--------- //--------- //--------- int depth = -1; vector< pathsInMol3D > molPaths; vector< pathsInMol3D > molPaths_test; // 2 - get the database description // -------------------------------- // i.e. fill in atomLabels vectors vector atomLabels; vector bondTypes; atomLabels = aSet->atomsLabelsListing(); for(int i = 1 ; i <= nBins; i++){ bondTypes.push_back(i); } if( !silentMode ){ for( int i = 0 ; i < atomLabels.size() ; i++ ){ Rcpp::Rcout << "atom type no " << i+1 << " ; atomic number = " << atomLabels[i] << endl; } //for( int i = 0 ; i < bondTypes.size() ; i++ ){ //cout << "bond type no " << i+1 << " ; bond type = " << bondTypes[i] << endl; //} } // 3 - transform the dataset from 2D to 3D // --------------------------------------- // OLD VERSION: distMin and distMin were computed from the data // --> now: specified as parameters //double distMin = 100000.0; //double distMax = 0.0; //aSet.minMaxDistances(&distMin, &distMax); if( !silentMode ){ Rcpp::Rcout << " - distMin = " << distMin << endl; Rcpp::Rcout << " - distMax = " << distMax << endl; Rcpp::Rcout << " - nBins = " << nBins << endl; Rcpp::Rcout << " --> binSize = " << (1.0001*distMax - distMin)/nBins << endl; } aSet->threeDtransform(nBins, distMin, distMax); aSet2->threeDtransform(nBins, distMin, distMax); // 4 - compute the gram matrix // ---------------------------- gramComputeSpectrum3D_test( aSet, aSet2, depth, depthMax, kernelType, &molPaths, &molPaths_test, &atomLabels, &bondTypes, silentMode); if( !silentMode ){ Rcpp::Rcout << "gramComputeSpectrum (test) OK" << endl; } } //-------- //internal //-------- //---------------------------------------------------------------------------------------------------------------------------------------------------------------- void gramComputeSpectrum3D_self(MoleculeSet* aSet, int depth, int depthMax, int kernelType, vector< pathsInMol3D >* molPaths, vector* atomLabels, vector* bondTypes, bool silentMode){ //---------------------------------------------------------------------------------------------------------------------------------------------------------------- vector< pathsInMol3D> molPaths_NEW; depth = depth + 1; if (depth == 0){ // depth = 0 --> initialization step // for each kind of atoms, find paths starting points and launch the recursion for(int i = 0 ; i < atomLabels->size() ; i++){ molPaths->clear(); if( !silentMode ){ Rcpp::Rcout << " \t finding paths starting from atoms labeled = " << (*atomLabels)[i] << endl; } init_path_pointers3D(aSet, molPaths, (*atomLabels)[i]); gramComputeSpectrum3D_self(aSet, depth, depthMax, kernelType, molPaths, atomLabels, bondTypes, silentMode); } // end "for(int i = 0 ; i < atomLabels->size() ; i++)" } // end "if(depth == 0)" else{ // if depth >= 1 : we can try to extend the paths stored in path_pointers for(int j = 0 ; j < atomLabels->size() ; j++){ for(int k = 0 ; k < bondTypes->size() ; k++){ // for each atom type/bond type : extend the paths updatePaths3D(aSet, (*atomLabels)[j], (*bondTypes)[k], molPaths, &molPaths_NEW, kernelType, depth); int nonEmpty; nonEmpty = molPaths_NEW.size(); if(nonEmpty > 0){ // i.e : if at least 1 vector of paths is non-empty we update the gram matrix or launch the recursion once again if(depth == depthMax){ updateGram3D_self(aSet, &molPaths_NEW, kernelType); updateSelfKernel3D(aSet,&molPaths_NEW, kernelType); continue; }else{ gramComputeSpectrum3D_self(aSet, depth, depthMax, kernelType, &molPaths_NEW, atomLabels, bondTypes, silentMode); } } // end "if(nonEmpty > 0)" } } } // end "else if(depth == 0)" } //---------------------------------------------------------------------------------------------------------------------------------------------------------------- void gramComputeSpectrum3D_test(MoleculeSet* aSet, MoleculeSet* aSet2, int depth, int depthMax, int kernelType, vector< pathsInMol3D >* molPaths, vector< pathsInMol3D >* molPaths_test, vector* atomLabels, vector* bondTypes, bool silentMode){ //---------------------------------------------------------------------------------------------------------------------------------------------------------------- vector< pathsInMol3D > molPaths_NEW; vector< pathsInMol3D > molPaths_test_NEW; depth = depth + 1; if (depth == 0){ // depth = 0 --> initialization step // for each kind of atoms, find paths starting points and launch the recursion for(int i = 0 ; i < atomLabels->size() ; i++){ molPaths->clear(); molPaths_test->clear(); if( !silentMode ){ Rcpp::Rcout << "\t - finding paths starting from atoms labeled = " << (*atomLabels)[i] << endl; } init_path_pointers3D(aSet, molPaths, (*atomLabels)[i]); init_path_pointers3D(aSet2, molPaths_test, (*atomLabels)[i]); gramComputeSpectrum3D_test(aSet, aSet2, depth, depthMax, kernelType, molPaths, molPaths_test, atomLabels, bondTypes, silentMode); } // end "for(int i = 0 ; i < atomLabels->size() ; i++)" } // end "if (depth == 0)" else{ // if depth >= 1, we can try to extend the paths previously stored in path_pointers for(int j = 0 ; j < atomLabels->size() ; j++){ for(int k = 0 ; k < bondTypes->size() ; k++){ // for each atom type/bond type : extend the paths updatePaths3D(aSet, (*atomLabels)[j], (*bondTypes)[k], molPaths, &molPaths_NEW, kernelType, depth); updatePaths3D(aSet2, (*atomLabels)[j], (*bondTypes)[k], molPaths_test, &molPaths_test_NEW, kernelType, depth); int nonEmpty1, nonEmpty2; nonEmpty1 = molPaths_NEW.size(); nonEmpty2 = molPaths_test_NEW.size(); if(nonEmpty1 > 0 || nonEmpty2 > 0){ // i.e : if at least 1 vector of paths is non-empty IN THE TWO SETS, we can try to extend the paths (otherwise, we do nothing) if(depth == depthMax){ updateGram3D_test(aSet, aSet2, &molPaths_NEW, &molPaths_test_NEW, kernelType); updateSelfKernel3D(aSet,&molPaths_NEW, kernelType); updateSelfKernel3D(aSet2,&molPaths_test_NEW, kernelType); continue; }else{ gramComputeSpectrum3D_test(aSet, aSet2, depth, depthMax, kernelType, &molPaths_NEW, &molPaths_test_NEW, atomLabels, bondTypes, silentMode); } } // end if(nonEmpty1 > 0 || nonEmpty2 > 0) } } } // end "else if(depth == 0)" } void init_path_pointers3D(MoleculeSet* aSet, vector< pathsInMol3D >* molPaths, string atomLabel){ // ----------------------------------------------------------------------------------------------- // --> fill in "path_pointers" for starting atoms labeled as "atomLabel" path3D aPath; pathsInMol3D aPathInMol; int ind = 0; vector::iterator aMol; // for each molecule of the set : for(aMol = (*aSet).begin() ; aMol != (*aSet).end() ; aMol++ ){ bool first = true; aPathInMol.path_pointers.clear(); vector::iterator anAtom; // for each atom of the molecule for(anAtom = (*aMol)->beginAtom() ; anAtom != (*aMol)->endAtom() ; anAtom++){ // if the atomic number corresponds to the one needed (i.e. "atomNumb") if((*anAtom)->getMorganLabel(true) == atomLabel){ if(first == true){ aPath.weight = 1; aPath.endAtom = *anAtom; aPath.startAtom = *anAtom; aPathInMol.path_pointers.push_back(aPath); aPathInMol.molName = (*aMol)->getName(); aPathInMol.molInd = ind; first = false; } // end "if(first == true)" else{ aPath.weight = 1; aPath.endAtom = *anAtom; aPath.startAtom = *anAtom; aPathInMol.path_pointers.push_back(aPath); } // end "else" } // end "if((*anAtom)->getMorganLabel(true) == atomLabel)" } if(aPathInMol.path_pointers.size() > 0 ){ (*molPaths).push_back(aPathInMol); } ind++; } } // ************************************************************************************************************************************************ // ------------------------------------------------------------------------------------------------------------------------------------------------ void updatePaths3D(MoleculeSet *aSet, string atomLabel, int bondType, vector< pathsInMol3D >* molPaths, vector< pathsInMol3D >* molPaths_NEW, int kernelType, int depth){ // ------------------------------------------------------------------------------------------------------------------------------------------------ // --> extends path stored in "path_pointers" to atom of type "atomNumb". Stores the results in "path_pointers_NEW" path3D aPath; pathsInMol3D aPathInMol; // object initialized for each molecule present in the input molPaths, possibly inserted in the output molPaths_NEW molPaths_NEW->clear(); // for each molecule in molPaths for(int i = 0 ; i < molPaths->size() ; i++){ bool first = true; aPathInMol.path_pointers.clear(); //for each previous path pointer for(int j = 0 ; j < (*molPaths)[i].path_pointers.size() ; j++ ){ map::iterator aBond; // along each bond of the path's ending atom : for(aBond = (*molPaths)[i].path_pointers[j].endAtom->beginBond() ; aBond != (*molPaths)[i].path_pointers[j].endAtom->endBond() ; aBond++ ){ if( ( depth < 3 && ( aBond->second->getLabel() == bondType ) && ( aBond->first->getMorganLabel(true) == atomLabel ) ) || ( depth == 3 // (NB : implicit here : kernelType = 0, 1 or 2. kernelType parameter useless for the moment, but may be usefull later...) && ( aBond->second->getLabel() == bondType ) && ( aBond->first->getMorganLabel(true) == atomLabel ) && ( aBond->first->getId() == (*molPaths)[i].path_pointers[j].startAtom->getId() ) ) ) { if(first == true){ aPath.weight = 1; aPath.endAtom = aBond->first; aPath.previousAtom = (*molPaths)[i].path_pointers[j].endAtom; aPath.startAtom = (*molPaths)[i].path_pointers[j].startAtom; aPathInMol.path_pointers.push_back(aPath); aPathInMol.molName = (*molPaths)[i].molName; aPathInMol.molInd = (*molPaths)[i].molInd; first = false; } // end "if(first == true)" else{ aPath.weight = 1; aPath.endAtom = aBond->first; aPath.previousAtom = (*molPaths)[i].path_pointers[j].endAtom; aPath.startAtom = (*molPaths)[i].path_pointers[j].startAtom; aPathInMol.path_pointers.push_back(aPath); } } // end "if (depth < 3 .....)" } } if(aPathInMol.path_pointers.size() > 0){ (*molPaths_NEW).push_back(aPathInMol); } } } // *********************************************************************************************************************** // ----------------------------------------------------------------------------------------------------------------------- void updateGram3D_self(MoleculeSet *aSet, vector< pathsInMol3D >* molPaths, int kernelType){ // ----------------------------------------------------------------------------------------------------------------------- double update; for(int i = 0 ; i < molPaths->size() ; i++){ for(int j = i ; j < molPaths->size() ; j++){ // 1 - compute the update factor depending on the kernel type switch(kernelType){ case 0 : // 3-points spectrum update = (*molPaths)[i].path_pointers.size() * (*molPaths)[j].path_pointers.size(); break; case 1 : // 3-points binary update = 1.0; break; case 2 : // 3-points Tanimoto update = 1.0; break; case 3 : // 2-points spectrum update = (*molPaths)[i].path_pointers.size() * (*molPaths)[j].path_pointers.size(); break; case 4 : // 2-points binary update = 1.0; break; case 5 : // 2-points Tanimoto update = 1.0; break; } // 2 - add the update factor to the gram matrix (*aSet).addToGram( (*molPaths)[i].molInd , (*molPaths)[j].molInd, update); if( i!= j){ (*aSet).addToGram( (*molPaths)[j].molInd , (*molPaths)[i].molInd, update); } } } } //************************************************************************** //-------------------------------------------------------------------------- void updateGram3D_test(MoleculeSet *aSet, MoleculeSet *aSet2, vector< pathsInMol3D >* molPaths, vector< pathsInMol3D >* molPaths_test, int kernelType){ //-------------------------------------------------------------------------- double update; for(int i = 0 ; i < molPaths->size() ; i++){ for(int j = 0 ; j < molPaths_test->size() ; j++){ // 1 - compute the update factor depending on the kernel type switch(kernelType){ case 0 : // 3-points spectrum update = (*molPaths)[i].path_pointers.size() * (*molPaths_test)[j].path_pointers.size(); break; case 1 : // 3-points binary update = 1.0; break; case 2 : // 3-points Tanimoto update = 1.0; break; case 3 : // 2-points spectrum update = (*molPaths)[i].path_pointers.size() * (*molPaths_test)[j].path_pointers.size(); break; case 4 : // 2-points binary update = 1.0; break; case 5 : // 2-points Tanimoto update = 1.0; break; } // 2 - add the update factor to the gram matrix (*aSet).addToGram( (*molPaths)[i].molInd , (*molPaths_test)[j].molInd, update); } } } // ************************************************************************** // -------------------------------------------------------------------------- void updateSelfKernel3D(MoleculeSet* aSet, vector< pathsInMol3D >* molPaths, int kernelType){ // -------------------------------------------------------------------------- Molecule *aMol; double update; for(int i = 0 ; i < molPaths->size() ; i++){ //aMol = (*aSet)[(*molPaths)[i].molName]; //cout << "### MolInd = " << (*molPaths)[i].molInd << endl; aMol = (*aSet)[(*molPaths)[i].molInd]; // 1 - compute the update facor depending on the kernel type switch(kernelType){ case 0 : // 3-points spectrum update = (*molPaths)[i].path_pointers.size() * (*molPaths)[i].path_pointers.size(); break; case 1 : // 3-points binary update = 1.0; break; case 2 : // 3-points Tanimoto update = 1.0; break; case 3 : // 2-points spectrum update = (*molPaths)[i].path_pointers.size() * (*molPaths)[i].path_pointers.size(); break; case 4 : // 2-points binary update = 1.0; break; case 5 : // 2-points Tanimoto update = 1.0; break; } // 2 - add the update factor to the self kernel values aMol->addToSelfKernel(update); } } RCPP_MODULE(spectrum3Dhelper) { Rcpp::function( "gramSpectrum3D_self", &gramSpectrum3D_self ); Rcpp::function( "gramSpectrum3D_test", &gramSpectrum3D_test ); } Rchemcpp/src/Rmolecule.cpp0000644000175000017500000000366612116123101016164 0ustar klambaueklambaue #include "Rmolecule.h" RCPP_MODULE(Rmolecule){ Rcpp::class_( "Rmolecule") .constructor() .method( "addAtom", &Rmolecule::addAtom, "Add an atom" ) .method( "linkAtoms", &Rmolecule::linkAtoms, "link two atoms (zero-indexed)" ) .method( "writeSD", &Rmolecule::writeSD, "write an SD file") .method( "listStringDescriptors", &Rmolecule::listStringDescriptors, "list the string-descriptors") //uncommon for Molecules ... therefore commented out... //.method( "listFloatDescriptors", &Rmolecule::listFloatDescriptors, "list the float-descriptors") //.method( "listIntDescriptors", &Rmolecule::listIntDescriptors, "list the int-descriptors") .method( "getStringDescriptorValue", &Rmolecule::getStringDescriptorValue, "get the value of a string-descriptor") .method( "getStringDescriptorUnit", &Rmolecule::getStringDescriptorUnit, "get the unit of a string-descriptor") .method( "getStringDescriptorComment", &Rmolecule::getStringDescriptorComment, "get the comment of a string-descriptor") //uncommon for Molecules ... therefore commented out... //.method( "getFloatDescriptorValue", &Rmolecule::getFloatDescriptorValue, "get the value of a float-descriptor") //.method( "getFloatDescriptorUnit", &Rmolecule::getFloatDescriptorUnit, "get the unit of a float-descriptor") //.method( "getFloatDescriptorComment", &Rmolecule::getFloatDescriptorComment, "get the comment of a float-descriptor") //.method( "getIntDescriptorValue", &Rmolecule::getIntDescriptorValue, "get the value of a int-descriptor") //.method( "getIntDescriptorUnit", &Rmolecule::getIntDescriptorUnit, "get the unit of a int-descriptor") //.method( "getIntDescriptorComment", &Rmolecule::getIntDescriptorComment, "get the comment of a int-descriptor") .method( "setStringDescriptor", &Rmolecule::setStringDescriptor, "set the value of a string-descriptor") .method( "deleteStringDescriptor", &Rmolecule::deleteStringDescriptor, "delete a string-descriptor") ; } Rchemcpp/src/Rmolecule.h0000644000175000017500000000621612116123101015623 0ustar klambaueklambaue #ifndef RMOLECULE_H #define RMOLECULE_H #include #include class Rmolecule : public Molecule { public: Rmolecule(){}; void addAtom(string aSymbol) {Molecule::addAtom(aSymbol);} void linkAtoms( int firstAtom, int secondAtom, int aBondLabel) {Molecule::linkAtoms(firstAtom, secondAtom, aBondLabel);} void writeSD( string aFileName ) {Molecule::writeSD(aFileName);} //Extension 1.0.7 std::vector listStringDescriptors() { std::vector vec; for (map< string, Descriptor< string >* >::iterator i = beginStringDescriptor(); i != endStringDescriptor(); i++) { if( !(*i).second->isEmpty() ) { vec.push_back((*i).second->getLabel()); } } return vec; } //uncommon for molecules /* std::vector listFloatDescriptors() { std::vector vec; for (map< string, Descriptor< float >* >::iterator i = beginFloatDescriptor(); i != endFloatDescriptor(); i++) { if( !(*i).second->isEmpty() ) { vec.push_back((*i).second->getLabel()); } } return vec; } std::vector listIntDescriptors() { std::vector vec; for (map< string, Descriptor< int >* >::iterator i = beginIntDescriptor(); i != endIntDescriptor(); i++) { if( !(*i).second->isEmpty() ) { vec.push_back((*i).second->getLabel()); } } return vec; } */ string getStringDescriptorValue(string label){ return getStringDescriptor(label)->getValue(); } string getStringDescriptorUnit(string label){ return getStringDescriptor(label)->getUnit(); } string getStringDescriptorComment(string label){ return getStringDescriptor(label)->getComment(); } //uncommon for molecules -> commented out //float getFloatDescriptorValue(string label){ return getFloatDescriptor(label)->getValue(); } //float getFloatDescriptorUnit(string label){ return getFloatDescriptor(label)->getUnit(); } //float getFloatDescriptorComment(string label){ return getFloatDescriptor(label)->getComment(); } //int getIntDescriptorValue(string label){ return getIntDescriptor(label)->getValue(); } //int getIntDescriptorUnit(string label){ return getIntDescriptor(label)->getUnit(); } //int getIntDescriptorComment(string label){ return getIntDescriptor(label)->getComment(); } void setStringDescriptor(string aLabel, string aValue, string aUnit, string aComment) { //Code taken over from datacontainer.cpp (corrected bug: only value was set if existing) Descriptor< string >* d; if( hasStringDescriptor( aLabel ) ){ d = stringDescriptors[aLabel]; d -> setValue(aValue); d -> setUnit(aUnit); d -> setComment(aComment); }else{ d = addStringDescriptor( aLabel, aValue, aUnit, aComment ); } } void deleteStringDescriptor(string aLabel) { //Code taken from datacontainer.cpp (corrected bug: Original function deletes all types of descriptors with that label) map* >::iterator its; for( its = (stringDescriptors).begin(); its != (stringDescriptors).end(); its++ ){ if( (*its).first == aLabel ){ delete (*its).second; (stringDescriptors).erase(its); //found = true; //return( true ); } } } //... //private: }; #endif Rchemcpp/src/subtreehelper.cpp0000644000175000017500000003512412116123101017100 0ustar klambaueklambaue/**************************************************************************************** subtreehelper.cpp ----------------- copyright : (C) 2012 Mahr Michael email : michael.mahr@gmail.com rearranged from: copyright : (C) 2006 Jean-Luc Perret - Pierre Mahé email : jean-luc.perret@unine.ch - pierre.mahe@ensmp.fr ***************************************************************************************/ /**************************************************************************************** * * * This program is free software; you can redistribute it and/or * * modify it under the terms of the GNU Lesser General Public * * License as published by the Free Software Foundation; either * * version 2.1 of the License, or (at your option) any later version. * * * * This program is distributed in the hope that it will be useful, * * but WITHOUT ANY WARRANTY; without even the implied warranty of * * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * * Lesser General Public License for more details. * * * * You should have received a copy of the GNU Lesser General Public * * License along with this library; if not, write to the Free Software * * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA * * * ****************************************************************************************/ #include "subtreehelper.h" //-------- //external //-------- void gramSubtree_self(SEXP s, double lambda, int depthMax, bool filterTotters, bool branchFlag, bool branchKernelUntilN, bool silentMode){ std::string rtypename("Rcpp_Rmoleculeset"); Rcpp::S4 s4obj( s ); if ( !s4obj.is( rtypename.c_str() ) ) { Rf_error( (std::string("object is not of the type ")+rtypename).c_str() ); } Rcpp::Environment env( s4obj ); Rcpp::XPtr xptr( env.get(".pointer") ); Rmoleculeset *aSet = static_cast (R_ExternalPtrAddr(xptr)); //--------- //--------- //--------- vector< vector< vector< Nextatom> > > edges; // the efficient indexation of the edges initialize_extended( aSet, &edges); // 2 - compute the gram matrix // --------------------------- int depth = -1; if (!silentMode) { Rcpp::Rcout << "Subtree-kernel computation:" << endl; Rcpp::Rcout << "\t- depthMax = " << depthMax << endl; Rcpp::Rcout << "\t- lambda = " << lambda << endl; if(filterTotters) Rcpp::Rcout << "\t- no-totters" << endl; else Rcpp::Rcout << "\t- with-totters" << endl; if(branchKernelUntilN) Rcpp::Rcout << "\t- until-N patterns" << endl; Rcpp::Rcout << endl; } vector::iterator mol1, mol2; int i = 0; int j = 0; double kernel; for( mol1 = aSet->begin(); mol1 != aSet->end(); mol1++ ){ if( !silentMode) Rcpp::Rcout << "\t\t-molecule no " << i+1 << "/" << aSet->numMolecules() << endl; j = 0; for( mol2 = aSet->begin() + i ; mol2!= aSet->end() ; mol2++){ kernel = subTreeKernel(*mol1, *mol2, &(edges[i]) , &(edges[i+j]) , depthMax, lambda, filterTotters, branchKernelUntilN, branchFlag); aSet->addToGram(i, i+j, kernel); if(j == 0){ (*mol1)->addToSelfKernel(kernel); } else{ aSet->addToGram(i+j, i, kernel); } j++; } i++; } } void gramSubtree_test(SEXP s, double lambda, int depthMax, bool filterTotters, bool branchFlag, bool branchKernelUntilN, bool silentMode){ std::string rtypename("Rcpp_Rmoleculeset"); Rcpp::S4 s4obj( s ); if ( !s4obj.is( rtypename.c_str() ) ) { Rf_error( (std::string("object is not of the type ")+rtypename).c_str() ); } Rcpp::Environment env( s4obj ); Rcpp::XPtr xptr( env.get(".pointer") ); Rmoleculeset *aSet = static_cast (R_ExternalPtrAddr(xptr)); /* std::string rtypename2("Rcpp_Rmoleculeset"); Rcpp::S4 s4obj2( s2 ); if ( !s4obj2.is( rtypename2.c_str() ) ) { Rf_error( (std::string("object is not of the type ")+rtypename2).c_str() ); } Rcpp::Environment env2( s4obj2 ); Rcpp::XPtr xptr2( env2.get(".pointer") ); Rmoleculeset *aSet2 = static_cast (R_ExternalPtrAddr(xptr2)); */ Rmoleculeset *aSet2 = aSet->getComparisonSet2(); //--------- //--------- //--------- vector< vector< vector< Nextatom> > > edges; // the efficient indexation of the edges initialize_extended( aSet, &edges); vector< vector< vector< Nextatom> > > edges2; // the efficient indexation of the edges initialize_extended( aSet2, &edges2); // 2 - compute the gram matrix // --------------------------- int depth = -1; if (!silentMode) { Rcpp::Rcout << "Subtree-kernel computation:" << endl; Rcpp::Rcout << "\t- depthMax = " << depthMax << endl; Rcpp::Rcout << "\t- lambda = " << lambda << endl; if(filterTotters) Rcpp::Rcout << "\t- no-totters" << endl; else Rcpp::Rcout << "\t- with-totters" << endl; if(branchKernelUntilN) Rcpp::Rcout << "\t- until-N patterns" << endl; Rcpp::Rcout << endl; } vector::iterator mol1, mol2; int i = 0; int j = 0; double kernel; for( mol1 = aSet->begin(); mol1 != aSet->end(); mol1++ ){ if( !silentMode) Rcpp::Rcout << "\t\t-molecule no " << i+1 << "/" << aSet->numMolecules() << endl; j = 0; for( mol2 = aSet2->begin(); mol2!= aSet2->end() ; mol2++){ kernel = subTreeKernel(*mol1, *mol2, &(edges[i]), &(edges2[j]), depthMax, lambda, filterTotters, branchKernelUntilN, branchFlag); aSet->addToGram(i, j, kernel); // self kernel: only once!! if(i == 0){ kernel = subTreeKernel(*mol2, *mol2, &(edges2[j]) , &(edges2[j]) , depthMax, lambda, filterTotters, branchKernelUntilN, branchFlag); (*mol2)->addToSelfKernel(kernel); } j++; } // self kernel kernel = subTreeKernel(*mol1, *mol1, &(edges[i]) , &(edges[i]) , depthMax, lambda, filterTotters, branchKernelUntilN, branchFlag); (*mol1)->addToSelfKernel(kernel); i++; } } //-------- //internal //-------- void initialize_extended(MoleculeSet* aSet, vector< vector< vector< Nextatom> > > * edges) { vector::iterator mol; map::iterator aBond; // For each molecule in the set: int n=0; for( mol = aSet->begin(); mol != aSet->end(); mol++ ){ edges->push_back(vector< vector< Nextatom> >()); // Make a map between the atoms and their index map< Atom*, int> indAtoms; for(int i = 0 ; i < (*mol)->numAtoms() ; i++){ indAtoms[ (*mol)->getAtomByIndex(i) ] = i; } // For each atom in the molecule: for (int i=0 ; i<(*mol)->numAtoms() ; i++){ Atom *atom = (*mol)->getAtomByIndex(i); edges->at(n).push_back( vector< Nextatom>() ); // For each bond leaving this atom for(aBond = atom->beginBond() ; aBond != atom->endBond() ; aBond++ ){ Atom *atom2 = aBond->first; int isIn = 0; int bondtype = atom->getBondWithTarget(atom2)->getLabel(); string atomtype = atom2->getMorganLabel(); int atomindex = indAtoms[atom2]; // Check whether a neighbor with the same bound/atom type already exists for (int k=0; k < edges->at(n)[i].size() ; k++){ if ( (atomtype == edges->at(n)[i][k].a) && (bondtype == edges->at(n)[i][k].b )){ // If a similar link is found, add the current atom to the corresponding list edges->at(n)[i][k].i.push_back(atomindex); isIn = 1; } } if (isIn == 0){ // If no similar link is found, create a nex list and add the current atom to it edges->at(n)[i].push_back(Nextatom()); int k=edges->at(n)[i].size()-1; edges->at(n)[i][k].b = bondtype; edges->at(n)[i][k].a = atomtype; edges->at(n)[i][k].i.push_back(atomindex); } } // All neighbours of *atom have been processed // Sort the list of neighbor links for faster matching later sort(edges->at(n)[i].begin(),edges->at(n)[i].end()); } // The current molecule is done. n++; } } double subTreeKernel( Molecule* mol1, Molecule* mol2, vector< vector< Nextatom> >* edges1, vector< vector< Nextatom> >* edges2, int depthMax, double lambda, bool filterTotters, bool untilFlag, bool branchFlag){ // ------------------------------------------------------------------ // 1 - create matrices of k_1(r,s)...k_h(r,s) // --> subMatrices[0],...,subMatrices[h-1] vector< vector< vector > > subMatrices; for(int i = 0; i < depthMax ; i++){ subMatrices.push_back( vector< vector >() ); for(int j = 0 ; j < mol1->numAtoms() ; j++){ subMatrices[i].push_back( vector() ); for (int k = 0 ; k < mol2->numAtoms() ; k++){ subMatrices[i][j].push_back(0.0); } } } // 2 - initialize subMatrices[0] with binary kernels based on Morgan Labels // NB: MorganLabelKernel can be changed to a different kernel (--> SEE SD2GRAM, OPTION -K) vector toupdate1,toupdate2; for(int i = 0 ; i < mol1->numAtoms() ; i++){ for(int j = 0 ; j < mol2->numAtoms() ; j++){ if(branchFlag){ subMatrices[0][i][j] = MoleculeUtils::atomKernelMorganLabel( mol1->getAtomByIndex(i) , mol2->getAtomByIndex(j) ); } else{ subMatrices[0][i][j] = lambda * MoleculeUtils::atomKernelMorganLabel( mol1->getAtomByIndex(i) , mol2->getAtomByIndex(j) ); } if (subMatrices[0][i][j] != 0.0 ) { toupdate1.push_back(i); toupdate2.push_back(j); } } } // 3 - compute subMatrices[1]...subMatrices[depthMax] double update, temp, ker; for(int depth = 1 ; depth < depthMax ; depth++){ // FOR EACH PAIR OF ATOMS THAT MATCH: for (int istep=0 ; istep < toupdate1.size() ; istep++) { int atom1 = toupdate1[istep]; int atom2 = toupdate2[istep]; vector tempKernel; // Detect common neighbor labels int n1=0,n2=0; while ( (n1 < (*edges1)[atom1].size()) && (n2 < (*edges2)[atom2].size() ) ) { if ((*edges1)[atom1][n1] < (*edges2)[atom2][n2]) { n1++; continue; } if ((*edges2)[atom2][n2] < (*edges1)[atom1][n1]) { n2++; continue; } // A matching pair has been found int natoms1 = (*edges1)[atom1][n1].i.size(); int natoms2 = (*edges2)[atom2][n2].i.size(); int maxCard = min( natoms1, natoms2); temp = 0.0; for(int d = 1; d <= maxCard; d++){ for(int l = 0 ; l < tuples[natoms1-1][d-1].size() ; l++){ for(int m = 0 ; m < tuples[natoms2-1][d-1].size() ; m++){ update = 1.0; for (int toto = 0 ; totonumAtoms() ; i++){ for(int j = 0 ; j < mol2->numAtoms() ; j++){ if( !(filterTotters && (mol1->getAtomByIndex(i)->getKashimaPS() * mol2->getAtomByIndex(j)->getKashimaPS() == 0.0) ) ){ kernel += subMatrices[depthMax-1][i][j]; } } } if(!branchFlag){ kernel = kernel / pow(lambda,depthMax); } /* // DEBUG: // CHECK CORRECTNESS --> print DP matrices on screen for(int k = 0 ; k < depthMax ; k++){ Rcpp::Rcout << "####--- subMatrices[" << k << "] ---####"<< endl; Rcpp::Rcout << "corner"; for(int i = 0 ; i < mol2->numAtoms() ; i++){ if( mol2->getAtomByIndex(i)->getKashimaPS() != 0.0 ){ Rcpp::Rcout << "\tno" << i+1; } } Rcpp::Rcout << endl; for(int i = 0 ; i < mol1->numAtoms() ; i++){ if( mol1->getAtomByIndex(i)->getKashimaPS() != 0.0 ){ Rcpp::Rcout << "no" << i+1; for(int j = 0 ; j < mol2->numAtoms() ; j++){ if(mol2->getAtomByIndex(j)->getKashimaPS() != 0.0 ){ Rcpp::Rcout << "\t" << subMatrices[k][i][j]; } } Rcpp::Rcout << endl; } } Rcpp::Rcout << "#####################" << endl ; } Rcpp::Rcout << endl << endl; */ return kernel; } // Initialize the static variable the contains the list of all p-tuples among n indices for various p and n Rchemcpp: Deleted MoleculeSetExtended; converted Member-function to this void initialize_tuples( int nmax){ tuples.clear(); //MAHRMIC: Remove all of them first - otherwise the algorithms segfault for (int n=0; n > >() ); for (int p=0; p<=n ; p++) { // (p+1) is the number of points that we choose tuples[n].push_back( vector< vector >() ); if (p==0) { // If we choose only one point, then we directly list all 1-tuples (the n singletons) for (int i=0; i<=n ; i++) { tuples[n][p].push_back( vector() ); tuples[n][p][i].push_back(i); } } else { // If we choose (p+1)>1 points, we deduce the (p+1)-tuples from the p-uples for (int i=0 ; i\z@ \thesection\quad \fi ##1}{ \ifnum \c@secnumdepth >\z@ \thesection\quad \fi ##1}}% \def\subsectionmark##1{}} \renewenvironment{abstract}{\section*{\abstractname}% \@mkboth{\abstractname}{\abstractname}}{} \newenvironment{acknowledgments}{\section*{Acknowledgments}% \@mkboth{Acknowledgments}{Acknowledgments}}{} \renewcommand\tableofcontents{% \section*{\contentsname \@mkboth{\contentsname}{\contentsname}}% \@starttoc{toc}% } \renewcommand\listoffigures{% \section*{\listfigurename}% \@mkboth{\listfigurename}% {\listfigurename}% \@starttoc{lof}% } \renewcommand\listoftables{% \section*{\listtablename}% \@mkboth{\listtablename}% {\listtablename}% \@starttoc{lot}% } \renewenvironment{thebibliography}[1] {\section*{\refname}% \@mkboth{\refname}{\refname}% \list{\@biblabel{\@arabic\c@enumiv}}% {\settowidth\labelwidth{\@biblabel{#1}}% \leftmargin\labelwidth \advance\leftmargin\labelsep \@openbib@code \usecounter{enumiv}% \let\p@enumiv\@empty \renewcommand\theenumiv{\@arabic\c@enumiv}}% \sloppy \clubpenalty4000 \@clubpenalty \clubpenalty \widowpenalty4000% \sfcode`\.\@m} {\def\@noitemerr {\@latex@warning{Empty `thebibliography' environment}}% \endlist} \endinput Rchemcpp/vignettes/bioinf.cls0000644000175000017500000001717112050404207016725 0ustar klambaueklambaue% % % BIOINF document class % % last revision 2010-02-15 by Ulrich Bodenhofer % % \NeedsTeXFormat{LaTeX2e}[1995/12/01] \ProvidesClass{bioinf}[2010/02/15] \def\bioinf@basisclass{article} \newif\ifbasisclassdefined \basisclassdefinedfalse % % declare options for three basis classes % \DeclareOption{article}{\ifbasisclassdefined \ClassError{bioinf}{You may only supply one basis class.}{Remove the `article' option} \fi \def\bioinf@basisclass{article}\basisclassdefinedtrue} \DeclareOption{report}{\ifbasisclassdefined \ClassError{bioinf}{You may only supply one basis class.}{Remove the `report' option} \fi \def\bioinf@basisclass{report}\basisclassdefinedtrue} \DeclareOption{book}{\ifbasisclassdefined \ClassError{bioinf}{You may only supply one basis class.}{Remove the `book' option} \fi \def\bioinf@basisclass{book}\basisclassdefinedtrue} % % forward all other options to document class article % \DeclareOption*{{\ClassError{bioinf}{Option `\CurrentOption' not supported for BIOINF document class.}{Remove the `\CurrentOption' option}}} % % process options % \ProcessOptions\relax % % load document class article (always with options a4paper and titlepage) % \LoadClass[a4paper,11pt,twoside]{\bioinf@basisclass} \typeout{=============================================================}% \typeout{BIOINF document class loaded with basis class '\bioinf@basisclass'}% \typeout{=============================================================}% % % font customizations % \renewcommand{\encodingdefault}{T1} \renewcommand{\rmdefault}{ptm} \renewcommand{\sfdefault}{phv} \renewcommand{\ttdefault}{cmtt} \fontencoding{T1}\selectfont % % customization of item marks % \renewcommand{\labelitemi}{\rule[0.2ex]{0.8ex}{0.8ex}} \renewcommand{\labelitemii}{$\m@th\bullet$} % % include of additional packages % \RequirePackage{color} \RequirePackage{graphicx} \RequirePackage{ifthen} % % setup of page format % \setlength{\evensidemargin}{-1in} \setlength{\oddsidemargin}{-1in} \setlength{\topmargin}{-1in} \addtolength\evensidemargin{30mm} \addtolength\oddsidemargin{30mm} \addtolength\topmargin{28mm} \setlength{\textwidth}{150mm} \setlength{\headheight}{12pt} \setlength{\headsep}{8mm} \setlength{\textheight}{222mm} \ifx\pdfoutput\undefined \relax \else \setlength{\pdfpagewidth}{21cm} \setlength{\pdfpageheight}{29.7cm} \fi % % setup of customized headings style and listings % \input{bioinf-\bioinf@basisclass.txi} % % set page style % \pagestyle{headings} % % adjust parskip % \setlength{\parskip}{5pt plus 2pt minus 1pt} % % define colors and color commands % \definecolor{biowh}{rgb}{1.00,1.00,1.00} \definecolor{biobl}{rgb}{0.00,0.00,0.00} \definecolor{bioaz}{rgb}{0.204,0.325,0.631} \definecolor{biove}{rgb}{0.24,0.63,0.37} \definecolor{bioli}{rgb}{0.667,0.9333,0.667} \newcommand{\blue}{\color{bioaz}} \newcommand{\green}{\color{biove}} % % define BIOINF title page commands % \newcommand{\customtitlepage}[2]{% \thispagestyle{empty} #2 \vfill \noindent\makebox[0pt][l]{% \setlength{\unitlength}{1cm} \begin{picture}(0,0)(2,2.4) %\put(0,25.5){\makebox(0,0)[lb]{\color{bioli}\rule{19cm}{2.2cm}}} \put(19,25.5){\makebox(0,0)[rb]{\includegraphics[width=19cm]{bioinf-bar}}} \put(0.3,26.0){\makebox(0,0)[l]{\color{bioaz}\fontfamily{phv}\bfseries\large% Institute of Bioinformatics, Johannes Kepler University Linz}} \put(0.3,26.9){\makebox(0,0)[l]{\color{bioaz}\fontfamily{phv}\bfseries\Huge #1}} \put(0,1.6){\makebox(0,0)[lb]{\color{bioli}\rule{19cm}{1mm}}} \put(0,0){\makebox(0,0)[lb]{\fontfamily{phv}\normalsize\begin{tabular}[b]{l} Institute of Bioinformatics\\ Johannes Kepler University Linz\\ A-4040 Linz, Austria \end{tabular}}} \put(19,0){\makebox(0,0)[rb]{\fontfamily{phv}\normalsize\begin{tabular}[b]{r} Tel. +43 732 2468 8880\\ Fax +43 732 2468 9511\\ http://www.bioinf.jku.at \end{tabular}}} %\put(0,24.5){\makebox(0,0)[lt]{\noindent\begin{minipage}{19cm} %#2 %\end{minipage}}} \end{picture}}\newpage} \def\bioinf@number{} \def\bioinf@keywords{} \def\bioinf@abstract{} \def\bioinf@citation{} \def\bioinf@affiliation{} \newcommand{\affiliation}[1]{\def\bioinf@affiliation{#1}} \newcommand{\pubnumber}[1]{\def\bioinf@number{#1}} \newcommand{\setkeywords}[1]{\def\bioinf@keywords{#1}} \newcommand{\setabstract}[1]{\def\bioinf@abstract{#1}} \newcommand{\setcitation}[1]{\def\bioinf@citation{#1}} \newcommand{\affilmark}[1]{\textsuperscript{\scriptsize #1}} \newcommand{\doi}[1]{% \ifx\href\undefined doi:#1 \else \href{http://dx.doi.org/#1}{doi:#1} \fi} \newcommand{\email}[1]{\ifx\href\undefined {\em #1} \else \href{mailto:#1}{{\em #1}} \fi} \newcommand{\reporttitlepage}[1]{\customtitlepage{#1}{% \mbox{ } \vfill \begin{center} \sffamily\bfseries\LARGE\@title \end{center} \begin{center} {\sffamily\bfseries\large\@author}\\[2mm] \ifx\@empty\bioinf@affiliation \relax \else {\sffamily\normalsize\bioinf@affiliation} \end{center} \fi \vspace{1.5cm} \ifthenelse{\equal{}{\bioinf@abstract}}{% \relax }{% \centerline{\rule{15cm}{1pt}}\vspace{2mm} \centerline{\begin{minipage}[b]{15cm} \noindent{\bf Abstract --- } \bioinf@abstract \end{minipage}} \ifx\@empty\bioinf@keywords \vspace{-1.5mm}\centerline{\rule{15cm}{1pt}} \fi } \ifthenelse{\equal{}{\bioinf@keywords}}{% \relax }{% \ifx\@empty\bioinf@abstract \centerline{\rule{15cm}{1pt}}\vspace{2mm} \else \vspace{5mm} \fi \centerline{\begin{minipage}[b]{15cm} \noindent{\bf Key words --- } \bioinf@keywords \end{minipage}} \vspace{-1.5mm}\centerline{\rule{15cm}{1pt}} } \ifthenelse{\equal{}{\bioinf@citation}}{% \relax }{% \vspace{2cm} \centerline{\begin{minipage}[b]{15cm} \noindent{\bf Citation: } \bioinf@citation \end{minipage}} } \vfill}} \newcommand{\researchreporttitlepage}{\reporttitlepage{Research Report \bioinf@number}} \newcommand{\techreporttitlepage}{\reporttitlepage{Technical Report \bioinf@number}} \newcommand{\preprinttitlepage}{\reporttitlepage{Publication Preprint}{}} \newcommand{\reprinttitlepage}{\reporttitlepage{Publication Reprint}{}} \newcommand{\manualtitlepage}[1][]{\customtitlepage{Software Manual}{% \mbox{ } \vfill \begin{center} \sffamily\bfseries\LARGE\@title \end{center} \ifthenelse{\equal{\@author}{}}{}{\begin{center} {\sffamily\bfseries\large\@author}\\[2mm] \ifx\@empty\bioinf@affiliation \relax \else {\sffamily\normalsize\bioinf@affiliation} \end{center} \fi} \vspace{1cm} \begin{center} {\Large\sffamily\bfseries #1} \end{center} \vfill \mbox{ }}} \newcommand{\lecturenotestitlepage}[1][]{\customtitlepage{Lecture Notes}{% \ifthenelse{\equal{#1}{}}{% \vspace*{2cm} \begin{center} \sffamily\bfseries\huge \@title \end{center} \vspace{2cm} \begin{center} \sffamily\bfseries\Large \@author \end{center} \vfill \mbox{ }}{% {\LARGE\sffamily\bfseries\mbox{ } \vfill \begin{center} #1 \end{center} \vfill \vfill \mbox{ }}}}} \newcommand{\copyrightpage}[1][]{% \clearpage \mbox{ } \vfill \vfill \noindent{\large \textcircled{c} \ifthenelse{\equal{#1}{}}{% \@author}{% #1}}\\[7mm] \noindent \BIOINFCopyrightNotice\thispagestyle{empty} \vfill \mbox{ } \clearpage} \newcommand{\BIOINFCopyrightNotice}{This material, no matter whether in printed or electronic form, may be used for personal and educational use only. Any reproduction of this manuscript, no matter whether as a whole or in parts, no matter whether in printed or in electronic form, requires explicit prior acceptance of the author.} \endinput % % end of scch-tr.cls % Rchemcpp/vignettes/literature.bib0000644000175000017500000004445612050404207017620 0ustar klambaueklambaue@ARTICLE{Mahr2012, author = {Mahr, M. and Klambauer, G. }, title = {{Rchemcpp - An R package for similarity measure of small molecules}}, journal = {Unpublished}, year = {2012}, volume = {xx}, pages = {xx}, URL = {http://www.bioinf.jku.at/software/Rchemcpp/index.html}, eprint = {http://www.bioinf.jku.at/software/Rchemcpp/index.html}, number = {1}, } @ARTICLE{Hochreiter2006, author = {Sepp Hochreiter and Klaus Obermayer}, title = {Support vector machines for dyadic data.}, journal = {Neural Comput}, year = {2006}, volume = {18}, pages = {1472--1510}, number = {6}, month = {Jun}, abstract = {We describe a new technique for the analysis of dyadic data, where two sets of objects (row and column objects) are characterized by a matrix of numerical values that describe their mutual relationships. The new technique, called potential support vector machine (P-SVM), is a large-margin method for the construction of classifiers and regression functions for the column objects. Contrary to standard support vector machine approaches, the P-SVM minimizes a scale-invariant capacity measure and requires a new set of constraints. As a result, the P-SVM method leads to a usually sparse expansion of the classification and regression functions in terms of the row rather than the column objects and can handle data and kernel matrices that are neither positive definite nor square. We then describe two complementary regularization schemes. The first scheme improves generalization performance for classification and regression tasks; the second scheme leads to the selection of a small, informative set of row support objects and can be applied to feature selection. Benchmarks for classification, regression, and feature selection tasks are performed with toy data as well as with several real-world data sets. The results show that the new method is at least competitive with but often performs better than the benchmarked standard methods for standard vectorial as well as true dyadic data sets. In addition, a theoretical justification is provided for the new approach.}, doi = {10.1162/neco.2006.18.6.1472}, institution = {Department of Electrical Engineering and Computer Science, Technische Universität, Berlin, Germany. hochreit@cs.tu-berlin.de}, keywords = {Artificial Intelligence; Databases, Genetic; Databases, Protein; Models, Theoretical; Oligonucleotide Array Sequence Analysis}, language = {eng}, medline-pst = {ppublish}, owner = {klambaue}, pmid = {16764511}, timestamp = {2011.04.29}, url = {http://dx.doi.org/10.1162/neco.2006.18.6.1472} } @article{Kashima2004, title={Kernels for graphs}, author={Kashima, H. and Tsuda, K. and Inokuchi, A.}, journal={Kernel methods in computational biology}, volume={39}, number={1}, pages={101--113}, year={2004}, publisher={MIT Press: Cambridge, MA} } @INPROCEEDINGS{Mahe2004, author = {Mah{\'e}, P. and Ueda, N. and Akutsu, T. and Perret, J.-L. and Vert, J.-P.}, title = {Extensions of marginalized graph kernels}, booktitle = {Proc of the 21st {ICML}}, year = {2004}, editor = {Greiner, R. and Schuurmans, D.}, pages = {552-559}, publisher = {ACM Press}, abstract = {Positive definite kernels between labeled graphs have recently been proposed.{T}hey enable the application of kernel methods, such as support vectormachines, to the analysis and classification of graphs, for example, chemicalcompounds. {T}hese graph kernels are obtained by marginalizing a kernel betweenpaths with respect to a random walk model on the graph vertices along theedges. {W}e propose two extensions of these graph kernels, with the double goal toreduce their computation time and increase their relevance as measure ofsimilarity between graphs. {F}irst, we propose to modify the label of eachvertex by automatically adding information about its environment with the useof the {M}organ algorithm. {S}econd, we suggest a modification of the random walkmodel to prevent the walk from coming back to a vertex that was just visited.{T}hese extensions are then tested on benchmark experiments of chemicalcompounds classification, with promising results.}, } @ARTICLE{Mahe2005, author = {Mah{\'e}, P. and Ueda, N. and Akutsu, T. and Perret, J.-L. and Vert, J.-P.}, title = {Graph kernels for molecular structure-activity relationship analysis with support vector machines.}, journal = {J. {C}hem. {I}nf. {M}odel.}, year = {2005}, volume = {45}, pages = {939-51}, number = {4}, abstract = {The support vector machine algorithm together with graph kernel functions has recently been introduced to model structure-activity relationships ({SAR}) of molecules from their 2{D} structure, without the need for explicit molecular descriptor computation. {W}e propose two extensions to this approach with the double goal to reduce the computational burden associated with the model and to enhance its predictive accuracy: description of the molecules by a {M}organ index process and definition of a second-order {M}arkov model for random walks on 2{D} structures. {E}xperiments on two mutagenicity data sets validate the proposed extensions, making this approach a possible complementary alternative to other modeling strategies.}, doi = {10.1021/ci050039t}, url = {http://dx.doi.org/10.1021/ci050039t}, } @ARTICLE{Mahe2006, author = {Pierre Mah{\'e} and Liva Ralaivola and V{\'e}ronique Stoven and Jean-Philippe Vert}, title = {The Pharmacophore Kernel for Virtual Screening with Support Vector Machines}, journal = {J. {C}hem. {I}nf. {M}odel.}, year = {2006}, volume = {46}, pages = {2003-2014}, number = {5}, abstract = {We introduce a family of positive definite kernels specifically optimized for the manipulation of 3D structures of molecules with kernel methods. The kernels are based on the comparison of the three-point pharmacophores present in the 3D structures of molecules, a set of molecular features known to be particularly relevant for virtual screening applications. We present a computationally demanding exact implementation of these kernels, as well as fast approximations related to the classical fingerprint-based approaches. Experimental results suggest that this new approach is competitive with state-of-the-art algorithms based on the 2D structure of molecules for the detection of inhibitors of several drug targets.}, doi = {10.1021/ci060138m}, url = {http://dx.doi.org/10.1021/ci060138m}, } @article{Mahe2009a, author = {Mah{\'e}, Pierre and Vert, Jean-Philippe}, title = {Graph kernels based on tree patterns for molecules}, journal = {Mach. Learn.}, issue_date = {April 2009}, volume = {75}, number = {1}, month = apr, year = {2009}, issn = {0885-6125}, pages = {3--35}, numpages = {33}, url = {http://dx.doi.org/10.1007/s10994-008-5086-2}, doi = {10.1007/s10994-008-5086-2}, acmid = {1519209}, publisher = {Kluwer Academic Publishers}, address = {Hingham, MA, USA}, keywords = {Chemoinformatics, Graph kernels, Support vector machines}, } @ARTICLE{Mahe2009b, author = {Mah{\'e}, Pierre and Vert, Jean-Philippe}, title = {Virtual screening with support vector machines and structure kernels.}, journal = {Comb Chem High Throughput Screen}, year = {2009}, volume = {12}, pages = {409--423}, number = {4}, month = {May}, abstract = {Support vector machines and kernel methods have recently gained considerable attention in chemoinformatics. They offer generally good performance for problems of supervised classification or regression, and provide a flexible and computationally efficient framework to include relevant information and prior knowledge about the data and problems to be handled. In particular, with kernel methods molecules do not need to be represented and stored explicitly as vectors or fingerprints, but only to be compared to each other through a comparison function technically called a kernel. While classical kernels can be used to compare vector or fingerprint representations of molecules, completely new kernels were developed in the recent years to directly compare the 2D or 3D structures of molecules, without the need for an explicit vectorization step through the extraction of molecular descriptors. While still in their infancy, these approaches have already demonstrated their relevance on several toxicity prediction and structure-activity relationship problems.}, file = {Mahe2009.pdf:Mahe2009.pdf:PDF;Mahe2009.pdf:Mahe2009.pdf:PDF}, institution = {Xerox Research Center Europe, 6 Chemin de Maupertuis, Meylan, France.}, keywords = {Algorithms; Artificial Intelligence; Computer Simulation; Databases, Factual; Drug Evaluation, Preclinical, methods; Ligands; Pharmaceutical Preparations, chemistry}, language = {eng}, medline-pst = {ppublish}, owner = {Günter}, pmid = {19442068}, timestamp = {2009.10.02} } @ARTICLE{Mohr2010, author = {Johannes Mohr and Brijnesh Jain and Andreas Sutter and Antonius Ter Laak and Thomas Steger-Hartmann and Nikolaus Heinrich and Klaus Obermayer}, title = {A maximum common subgraph kernel method for predicting the chromosome aberration test.}, journal = {J Chem Inf Model}, year = {2010}, volume = {50}, pages = {1821--1838}, number = {10}, month = {Oct}, abstract = {The chromosome aberration test is frequently used for the assessment of the potential of chemicals and drugs to elicit genetic damage in mammalian cells in vitro. Due to the limitations of experimental genotoxicity testing in early drug discovery phases, a model to predict the chromosome aberration test yielding high accuracy and providing guidance for structure optimization is urgently needed. In this paper, we describe a machine learning approach for predicting the outcome of this assay based on the structure of the investigated compound. The novelty of the proposed method consists in combining a maximum common subgraph kernel for measuring the similarity of two chemical graphs with the potential support vector machine for classification. In contrast to standard support vector machine classifiers, the proposed approach does not provide a black box model but rather allows to visualize structural elements with high positive or negative contribution to the class decision. In order to compare the performance of different methods for predicting the outcome of the chromosome aberration test, we compiled a large data set exhibiting high quality, reliability, and consistency from public sources and configured a fixed cross-validation protocol, which we make publicly available. In a comparison to standard methods currently used in pharmaceutical industry as well as to other graph kernel approaches, the proposed method achieved significantly better performance.}, doi = {10.1021/ci900367j}, institution = {School for Electrical Engineering and Computer Science, Berlin Institute of Technology, Berlin, Germany. johann@cs.tu-berlin.de}, language = {eng}, medline-pst = {ppublish}, owner = {klambaue}, pmid = {20883013}, timestamp = {2010.12.14}, url = {http://dx.doi.org/10.1021/ci900367j} } @ARTICLE{Ralaivola2005, author = {Liva Ralaivola and Sanjay J Swamidass and Hiroto Saigo and Pierre Baldi}, title = {Graph kernels for chemical informatics.}, journal = {Neural Netw}, year = {2005}, volume = {18}, pages = {1093--1110}, number = {8}, month = {Oct}, abstract = {Increased availability of large repositories of chemical compounds is creating new challenges and opportunities for the application of machine learning methods to problems in computational chemistry and chemical informatics. Because chemical compounds are often represented by the graph of their covalent bonds, machine learning methods in this domain must be capable of processing graphical structures with variable size. Here, we first briefly review the literature on graph kernels and then introduce three new kernels (Tanimoto, MinMax, Hybrid) based on the idea of molecular fingerprints and counting labeled paths of depth up to d using depth-first search from each possible vertex. The kernels are applied to three classification problems to predict mutagenicity, toxicity, and anti-cancer activity on three publicly available data sets. The kernels achieve performances at least comparable, and most often superior, to those previously reported in the literature reaching accuracies of 91.5\% on the Mutag dataset, 65-67\% on the PTC (Predictive Toxicology Challenge) dataset, and 72\% on the NCI (National Cancer Institute) dataset. Properties and tradeoffs of these kernels, as well as other proposed kernels that leverage 1D or 3D representations of molecules, are briefly discussed.}, doi = {10.1016/j.neunet.2005.07.009}, file = {Ralaivola2005.pdf:Ralaivola2005.pdf:PDF;Ralaivola2005.pdf:Ralaivola2005.pdf:PDF}, institution = {School of Information and Computer Sciences, University of California, Irvine, CA 92697-3425, USA.}, keywords = {Anticarcinogenic Agents; Artificial Intelligence; Computer Graphics; Databases, Genetic; Drug Toxicity; Information Storage and Retrieval; National Practitioner Data Bank; Pattern Recognition, Automated; Sequence Analysis, Protein; Structure-Activity Relationship; United States; emical}, language = {eng}, medline-pst = {ppublish}, owner = {Günter}, pii = {S0893-6080(05)00169-3}, pmid = {16157471}, timestamp = {2009.10.02}, url = {http://dx.doi.org/10.1016/j.neunet.2005.07.009} } @ARTICLE{Swamidass2005, author = {S. Joshua Swamidass and Jonathan Chen and Jocelyne Bruand and Peter Phung and Liva Ralaivola and Pierre Baldi}, title = {Kernels for small molecules and the prediction of mutagenicity, toxicity and anti-cancer activity.}, journal = {Bioinformatics}, year = {2005}, volume = {21 Suppl 1}, pages = {i359--i368}, month = {Jun}, abstract = {MOTIVATION: Small molecules play a fundamental role in organic chemistry and biology. They can be used to probe biological systems and to discover new drugs and other useful compounds. As increasing numbers of large datasets of small molecules become available, it is necessary to develop computational methods that can deal with molecules of variable size and structure and predict their physical, chemical and biological properties. RESULTS: Here we develop several new classes of kernels for small molecules using their 1D, 2D and 3D representations. In 1D, we consider string kernels based on SMILES strings. In 2D, we introduce several similarity kernels based on conventional or generalized fingerprints. Generalized fingerprints are derived by counting in different ways subpaths contained in the graph of bonds, using depth-first searches. In 3D, we consider similarity measures between histograms of pairwise distances between atom classes. These kernels can be computed efficiently and are applied to problems of classification and prediction of mutagenicity, toxicity and anti-cancer activity on three publicly available datasets. The results derived using cross-validation methods are state-of-the-art. Tradeoffs between various kernels are briefly discussed. AVAILABILITY: Datasets available from http://www.igb.uci.edu/servers/servers.html}, doi = {10.1093/bioinformatics/bti1055}, file = {Swamidass2005.pdf:Swamidass2005.pdf:PDF;Swamidass2005.pdf:Swamidass2005.pdf:PDF}, institution = {Institute for Genomics and Bioinformatics, School of Information and Computer Sciences, University of California Irvine, CA, USA.}, keywords = {Animals; Antineoplastic Agents, pharmacology; Computational Biology, methods; Computer Simulation; Databases, Factual; Drug Toxicity; Female; Male; Mice; Models, Molecular; Models, Statistical; Mutagens; Neoplasms, pathology; Pattern Recognition, Automated; ROC Curve; Rats}, language = {eng}, medline-pst = {ppublish}, owner = {Günter}, pii = {21/suppl_1/i359}, pmid = {15961479}, timestamp = {2009.10.02}, url = {http://dx.doi.org/10.1093/bioinformatics/bti1055} } @Article{Bodenhofer2011, title = {{APCluster: an R package for affinity propagation clustering}}, author = {Ulrich Bodenhofer and Andreas Kothmeier and Sepp Hochreiter}, journal = {Bioinformatics}, year = {2011}, volume = {27}, pages = {2463--2464}, doi = {10.1093/bioinformatics/btr406}, } @Article{Frey2007, title = {Clustering by passing messages between data points}, author = {Brendan J. Frey and Delbert Dueck}, journal = {Science}, volume = {315}, pages = {972--977}, year = {2007}, doi = {10.1126/science.1136800}, } Rchemcpp/vignettes/Rchemcpp.Rnw0000644000175000017500000001214412114111703017175 0ustar klambaueklambaue\documentclass[article]{bioinf} \usepackage[noae]{Sweave} \usepackage{amsmath,amssymb} \usepackage{hyperref} \usepackage{float} \usepackage[authoryear]{natbib} \hypersetup{colorlinks=false, pdfborder=0 0 0, pdftitle={Rchemcpp - Kernels for molecules}, pdfauthor={G\"unter Klambauer}} \title{Rchemcpp - Kernels for molecules} \author{Michael Mahr and G\"unter Klambauer} \affiliation{Institute of Bioinformatics, Johannes Kepler University Linz\\Altenberger Str. 69, 4040 Linz, Austria\\ \email{michael.mahr@gmail.com}} %\VignetteIndexEntry{Rchemcpp} %\VignetteKeywords{Kernel, kernels, graph, graphs, similarity, molecules, compounds} \newcommand{\Rchemcpp}{\texttt{Rchemcpp}} \newcommand{\method}[1]{{\fontfamily{phv}\fontshape{rm}\selectfont #1}} \newcommand{\R}{R} \newcommand{\Real}{\mathbb{R}} \renewcommand{\vec}[1]{\mathbf{#1}} \setkeys{Gin}{width=0.55\textwidth} \SweaveOpts{eps=FALSE} \begin{document} <>= options(width=75) set.seed(0) library(Rchemcpp) RchemcppVersion <- packageDescription("Rchemcpp")$Version @ \newcommand{\RchemcppVersion}{\Sexpr{RchemcppVersion}} \manualtitlepage[Version \RchemcppVersion, \today] %\section*{Scope and Purpose of this Document} % %This document is a user manual for the \R\ package \cnmops. %It is only meant as a gentle introduction into how to use the basic %functions implemented in this package. Not all features of the \R\ %package are described in full detail. Such details can be obtained %from the documentation enclosed in the \R\ package. Further note %the following: (1) this is neither an introduction to CNV detection from NGS %data; (2) this is not an introduction to \R. %If you lack the background for understanding this manual, you first %have to read introductory literature on these subjects. % \vspace{1cm} \newlength{\auxparskip} \setlength{\auxparskip}{\parskip} \setlength{\parskip}{0pt} \tableofcontents \clearpage \setlength{\parskip}{\auxparskip} \newlength{\Nboxwidth} \setlength{\Nboxwidth}{\textwidth} \addtolength{\Nboxwidth}{-2\fboxrule} \addtolength{\Nboxwidth}{-2\fboxsep} \newcommand{\notebox}[1]{% \begin{center} \fbox{\begin{minipage}{\Nboxwidth} \noindent{\sffamily\bfseries Note:} #1 \end{minipage}} \end{center}} \section{Introduction} The \Rchemcpp\ package is part of the CRAN project. The functionality of {\tt Chemcpp} (\url{http://chemcpp.sourceforge.net/html/index.html}) is provided in \R, that is the computation of similarities between molecules by kernel functions. The following kernels are implemented: \begin{itemize} \item the marginalized graph kernel between labeled graphs \citep{Kashima2004}. \item extensions of the marginalized kernel \citep{Mahe2004}. \item Tanimoto kernels \citep{Ralaivola2005}. \item graph kernels based on tree patterns \citep{Mahe2009a}. \item kernels based on pharmacophores for 3D structure of molecules \citep{Mahe2006}. \end{itemize} %See \url{http://www.bioinf.jku.at/software/Rchemcpp/index.html} for additional %information.\par \section{Getting started and quick start} To load the package, enter the following in your \R\ session: <<>>= library(Rchemcpp) @ We enter the filename of and SDF file to the function {\tt sd2gram}. This function computes the similarity of the molecules with the marginalized kernel \citep{Kashima2004} approach.\\ <>= sdfolder <- system.file("sample_data",package="Rchemcpp") sdf <- list.files(sdfolder,full.names=TRUE,pattern="small") K <- sd2gram(sdf,returnNormalized=TRUE,moleculeNameProperty="Compound Name") @ The similarity values are now stored in {\tt K}. We visualize this matrix as a heatmap. <>= heatmap(K,Rowv=NA,Colv=NA,scale="none") @ <>= pdf("001.pdf") par(oma=c(6,0,0,2)) heatmap(K,Rowv=NA,Colv=NA,scale="none") dev.off() @ \begin{figure}[H] \begin{center} \includegraphics[angle=0,width= 0.9\columnwidth]{001.pdf} \end{center} \end{figure} \newpage \section{Molecular similarity for clustering} Based on the similarity measure we can run clustering algorithms on the data in order to find groups among the molecules. We use Affinity Propagation Clustering \citep{Frey2007} as implemented by \citet{Bodenhofer2011} for this task, because the cluster centers are real molecules. <<>>= library(apcluster) r <- apcluster(K) @ <>= plot(r,K) @ <>= pdf("002.pdf") par(oma=c(6,0,0,2)) plot(r,K) dev.off() @ \begin{figure}[H] \begin{center} \includegraphics[angle=0,width= 0.9\columnwidth]{002.pdf} \end{center} \end{figure} %\section{Molecular similarity for prediction tasks using SVMs} %\section{How to cite this package} %If you use this package for research that is published later, you are kindly %asked to cite it as follows: \citep{Mahr:12}. %To obtain Bib\TeX\ entries of the reference, you can enter the following %into your R session: %<>= %toBibtex(citation("Rchemcpp")) %@ \bibliographystyle{natbib} \bibliography{literature} \end{document} Rchemcpp/vignettes/natbib.sty0000644000175000017500000010266312050404207016755 0ustar klambaueklambaue%% %% This is file `natbib.sty', %% generated with the docstrip utility. %% %% The original source files were: %% %% natbib.dtx (with options: `package,all') %% ============================================= %% IMPORTANT NOTICE: %% %% This program can be redistributed and/or modified under the terms %% of the LaTeX Project Public License Distributed from CTAN %% archives in directory macros/latex/base/lppl.txt; either %% version 1 of the License, or any later version. %% %% This is a generated file. %% It may not be distributed without the original source file natbib.dtx. %% %% Full documentation can be obtained by LaTeXing that original file. %% Only a few abbreviated comments remain here to describe the usage. %% ============================================= %% Copyright 1993-2000 Patrick W Daly %% Max-Planck-Institut f\"ur Aeronomie %% Max-Planck-Str. 2 %% D-37191 Katlenburg-Lindau %% Germany %% E-mail: daly@linmpi.mpg.de \NeedsTeXFormat{LaTeX2e}[1995/06/01] \ProvidesPackage{natbib} [2000/07/24 7.0a (PWD)] % This package reimplements the LaTeX \cite command to be used for various % citation styles, both author-year and numerical. It accepts BibTeX % output intended for many other packages, and therefore acts as a % general, all-purpose citation-style interface. % % With standard numerical .bst files, only numerical citations are % possible. With an author-year .bst file, both numerical and % author-year citations are possible. % % If author-year citations are selected, \bibitem must have one of the % following forms: % \bibitem[Jones et al.(1990)]{key}... % \bibitem[Jones et al.(1990)Jones, Baker, and Williams]{key}... % \bibitem[Jones et al., 1990]{key}... % \bibitem[\protect\citeauthoryear{Jones, Baker, and Williams}{Jones % et al.}{1990}]{key}... % \bibitem[\protect\citeauthoryear{Jones et al.}{1990}]{key}... % \bibitem[\protect\astroncite{Jones et al.}{1990}]{key}... % \bibitem[\protect\citename{Jones et al., }1990]{key}... % \harvarditem[Jones et al.]{Jones, Baker, and Williams}{1990}{key}... % % This is either to be made up manually, or to be generated by an % appropriate .bst file with BibTeX. % Author-year mode || Numerical mode % Then, \citet{key} ==>> Jones et al. (1990) || Jones et al. [21] % \citep{key} ==>> (Jones et al., 1990) || [21] % Multiple citations as normal: % \citep{key1,key2} ==>> (Jones et al., 1990; Smith, 1989) || [21,24] % or (Jones et al., 1990, 1991) || [21,24] % or (Jones et al., 1990a,b) || [21,24] % \cite{key} is the equivalent of \citet{key} in author-year mode % and of \citep{key} in numerical mode % Full author lists may be forced with \citet* or \citep*, e.g. % \citep*{key} ==>> (Jones, Baker, and Williams, 1990) % Optional notes as: % \citep[chap. 2]{key} ==>> (Jones et al., 1990, chap. 2) % \citep[e.g.,][]{key} ==>> (e.g., Jones et al., 1990) % \citep[see][pg. 34]{key}==>> (see Jones et al., 1990, pg. 34) % (Note: in standard LaTeX, only one note is allowed, after the ref. % Here, one note is like the standard, two make pre- and post-notes.) % \citealt{key} ==>> Jones et al. 1990 % \citealt*{key} ==>> Jones, Baker, and Williams 1990 % \citealp{key} ==>> Jones et al., 1990 % \citealp*{key} ==>> Jones, Baker, and Williams, 1990 % Additional citation possibilities (both author-year and numerical modes) % \citeauthor{key} ==>> Jones et al. % \citeauthor*{key} ==>> Jones, Baker, and Williams % \citeyear{key} ==>> 1990 % \citeyearpar{key} ==>> (1990) % \citetext{priv. comm.} ==>> (priv. comm.) % Note: full author lists depends on whether the bib style supports them; % if not, the abbreviated list is printed even when full requested. % % For names like della Robbia at the start of a sentence, use % \Citet{dRob98} ==>> Della Robbia (1998) % \Citep{dRob98} ==>> (Della Robbia, 1998) % \Citeauthor{dRob98} ==>> Della Robbia % % % Citation aliasing is achieved with % \defcitealias{key}{text} % \citetalias{key} ==>> text % \citepalias{key} ==>> (text) % % Defining the citation style of a given bib style: % Use \bibpunct (in the preamble only) with 6 mandatory arguments: % 1. opening bracket for citation % 2. closing bracket % 3. citation separator (for multiple citations in one \cite) % 4. the letter n for numerical styles, s for superscripts % else anything for author-year % 5. punctuation between authors and date % 6. punctuation between years (or numbers) when common authors missing % One optional argument is the character coming before post-notes. It % appears in square braces before all other arguments. May be left off. % Example (and default) \bibpunct[, ]{(}{)}{;}{a}{,}{,} % % To make this automatic for a given bib style, named newbib, say, make % a local configuration file, natbib.cfg, with the definition % \newcommand{\bibstyle@newbib}{\bibpunct...} % Then the \bibliographystyle{newbib} will cause \bibstyle@newbib to % be called on THE NEXT LATEX RUN (via the aux file). % % Such preprogrammed definitions may be invoked in the text (preamble only) % by calling \citestyle{newbib}. This is only useful if the style specified % differs from that in \bibliographystyle. % % With \citeindextrue and \citeindexfalse, one can control whether the % \cite commands make an automatic entry of the citation in the .idx % indexing file. For this, \makeindex must also be given in the preamble. % % LaTeX2e Options: (for selecting punctuation) % round - round parentheses are used (default) % square - square brackets are used [option] % curly - curly braces are used {option} % angle - angle brackets are used