X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=classifyseqscommand.cpp;h=0a1ef8823e35f687529c70de5ef29e7d33eb91fe;hb=45b880733289694b2a23c9dc8a774aa626458f81;hp=26df57ffbb56d6ddd0cc3532013c2be65fe5e3f8;hpb=e189982e0a9b7352ad57cc38ccee675f128be22e;p=mothur.git diff --git a/classifyseqscommand.cpp b/classifyseqscommand.cpp index 26df57f..0a1ef88 100644 --- a/classifyseqscommand.cpp +++ b/classifyseqscommand.cpp @@ -11,10 +11,61 @@ #include "sequence.hpp" #include "bayesian.h" #include "phylotree.h" +#include "phylosummary.h" #include "knn.h" -//********************************************************************************************************************** +//********************************************************************************************************************** +vector ClassifySeqsCommand::getValidParameters(){ + try { + string AlignArray[] = {"template","fasta","name","group","search","ksize","method","processors","taxonomy","match","mismatch","gapopen","gapextend","numwanted","cutoff","probs","iters", "outputdir","inputdir"}; + vector myArray (AlignArray, AlignArray+(sizeof(AlignArray)/sizeof(string))); + return myArray; + } + catch(exception& e) { + m->errorOut(e, "ClassifySeqsCommand", "getValidParameters"); + exit(1); + } +} +//********************************************************************************************************************** +ClassifySeqsCommand::ClassifySeqsCommand(){ + try { + abort = true; + //initialize outputTypes + vector tempOutNames; + outputTypes["taxonomy"] = tempOutNames; + outputTypes["taxsummary"] = tempOutNames; + outputTypes["matchdist"] = tempOutNames; + } + catch(exception& e) { + m->errorOut(e, "ClassifySeqsCommand", "ClassifySeqsCommand"); + exit(1); + } +} +//********************************************************************************************************************** +vector ClassifySeqsCommand::getRequiredParameters(){ + try { + string Array[] = {"fasta","template","taxonomy"}; + vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); + return myArray; + } + catch(exception& e) { + m->errorOut(e, "ClassifySeqsCommand", "getRequiredParameters"); + exit(1); + } +} +//********************************************************************************************************************** +vector ClassifySeqsCommand::getRequiredFiles(){ + try { + vector myArray; + return myArray; + } + catch(exception& e) { + m->errorOut(e, "ClassifySeqsCommand", "getRequiredFiles"); + exit(1); + } +} +//********************************************************************************************************************** ClassifySeqsCommand::ClassifySeqsCommand(string option) { try { abort = false; @@ -25,13 +76,13 @@ ClassifySeqsCommand::ClassifySeqsCommand(string option) { else { //valid paramters for this command - string AlignArray[] = {"template","fasta","name","search","ksize","method","processors","taxonomy","match","mismatch","gapopen","gapextend","numwanted","cutoff","probs","iters", "outputdir","inputdir"}; + string AlignArray[] = {"template","fasta","name","group","search","ksize","method","processors","taxonomy","match","mismatch","gapopen","gapextend","numwanted","cutoff","probs","iters", "outputdir","inputdir"}; vector myArray (AlignArray, AlignArray+(sizeof(AlignArray)/sizeof(string))); OptionParser parser(option); map parameters = parser.getParameters(); - ValidParameters validParameter; + ValidParameters validParameter("classify.seqs"); map::iterator it; //check to make sure all parameters are valid for command @@ -39,6 +90,12 @@ ClassifySeqsCommand::ClassifySeqsCommand(string option) { if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; } } + //initialize outputTypes + vector tempOutNames; + outputTypes["taxonomy"] = tempOutNames; + outputTypes["taxsummary"] = tempOutNames; + outputTypes["matchdist"] = tempOutNames; + //if the user changes the output directory command factory will send this info to us in the output parameter outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){ outputDir = ""; } @@ -50,7 +107,7 @@ ClassifySeqsCommand::ClassifySeqsCommand(string option) { it = parameters.find("template"); //user has given a template file if(it != parameters.end()){ - path = hasPath(it->second); + path = m->hasPath(it->second); //if the user has not given a path then, add inputdir. else leave path alone. if (path == "") { parameters["template"] = inputDir + it->second; } } @@ -58,10 +115,18 @@ ClassifySeqsCommand::ClassifySeqsCommand(string option) { it = parameters.find("taxonomy"); //user has given a template file if(it != parameters.end()){ - path = hasPath(it->second); + path = m->hasPath(it->second); //if the user has not given a path then, add inputdir. else leave path alone. if (path == "") { parameters["taxonomy"] = inputDir + it->second; } } + + it = parameters.find("group"); + //user has given a template file + if(it != parameters.end()){ + path = m->hasPath(it->second); + //if the user has not given a path then, add inputdir. else leave path alone. + if (path == "") { parameters["group"] = inputDir + it->second; } + } } //check for required parameters @@ -73,46 +138,52 @@ ClassifySeqsCommand::ClassifySeqsCommand(string option) { } else if (templateFileName == "not open") { abort = true; } + fastaFileName = validParameter.validFile(parameters, "fasta", false); if (fastaFileName == "not found") { m->mothurOut("fasta is a required parameter for the classify.seqs command."); m->mothurOutEndLine(); abort = true; } else { - splitAtDash(fastaFileName, fastaFileNames); + m->splitAtDash(fastaFileName, fastaFileNames); //go through files and make sure they are good, if not, then disregard them for (int i = 0; i < fastaFileNames.size(); i++) { if (inputDir != "") { - string path = hasPath(fastaFileNames[i]); + string path = m->hasPath(fastaFileNames[i]); //if the user has not given a path then, add inputdir. else leave path alone. if (path == "") { fastaFileNames[i] = inputDir + fastaFileNames[i]; } } int ableToOpen; - #ifdef USE_MPI - int pid; - MPI_Comm_size(MPI_COMM_WORLD, &processors); //set processors to the number of mpi processes running - MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are - - if (pid == 0) { - #endif - ifstream in; - ableToOpen = openInputFile(fastaFileNames[i], in); - in.close(); + ableToOpen = m->openInputFile(fastaFileNames[i], in, "noerror"); + + //if you can't open it, try default location + if (ableToOpen == 1) { + if (m->getDefaultPath() != "") { //default path is set + string tryPath = m->getDefaultPath() + m->getSimpleName(fastaFileNames[i]); + m->mothurOut("Unable to open " + fastaFileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine(); + ifstream in2; + ableToOpen = m->openInputFile(tryPath, in2, "noerror"); + in2.close(); + fastaFileNames[i] = tryPath; + } + } - #ifdef USE_MPI - for (int j = 1; j < processors; j++) { - MPI_Send(&ableToOpen, 1, MPI_INT, j, 2001, MPI_COMM_WORLD); - } - }else{ - MPI_Status status; - MPI_Recv(&ableToOpen, 1, MPI_INT, 0, 2001, MPI_COMM_WORLD, &status); + if (ableToOpen == 1) { + if (m->getOutputDir() != "") { //default path is set + string tryPath = m->getOutputDir() + m->getSimpleName(fastaFileNames[i]); + m->mothurOut("Unable to open " + fastaFileNames[i] + ". Trying output directory " + tryPath); m->mothurOutEndLine(); + ifstream in2; + ableToOpen = m->openInputFile(tryPath, in2, "noerror"); + in2.close(); + fastaFileNames[i] = tryPath; } - - #endif + } + + in.close(); if (ableToOpen == 1) { - m->mothurOut(fastaFileNames[i] + " will be disregarded."); m->mothurOutEndLine(); + m->mothurOut("Unable to open " + fastaFileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); //erase from file list fastaFileNames.erase(fastaFileNames.begin()+i); i--; @@ -138,41 +209,51 @@ ClassifySeqsCommand::ClassifySeqsCommand(string option) { if (namefile == "not found") { namefile = ""; } else { - splitAtDash(namefile, namefileNames); + m->splitAtDash(namefile, namefileNames); //go through files and make sure they are good, if not, then disregard them for (int i = 0; i < namefileNames.size(); i++) { if (inputDir != "") { - string path = hasPath(namefileNames[i]); + string path = m->hasPath(namefileNames[i]); //if the user has not given a path then, add inputdir. else leave path alone. if (path == "") { namefileNames[i] = inputDir + namefileNames[i]; } } int ableToOpen; - #ifdef USE_MPI - int pid; - MPI_Comm_size(MPI_COMM_WORLD, &processors); //set processors to the number of mpi processes running - MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are - - if (pid == 0) { - #endif - ifstream in; - ableToOpen = openInputFile(namefileNames[i], in); - in.close(); + ableToOpen = m->openInputFile(namefileNames[i], in, "noerror"); + + //if you can't open it, try default location + if (ableToOpen == 1) { + if (m->getDefaultPath() != "") { //default path is set + string tryPath = m->getDefaultPath() + m->getSimpleName(namefileNames[i]); + m->mothurOut("Unable to open " + namefileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine(); + ifstream in2; + ableToOpen = m->openInputFile(tryPath, in2, "noerror"); + in2.close(); + namefileNames[i] = tryPath; + } + } - #ifdef USE_MPI - for (int j = 1; j < processors; j++) { - MPI_Send(&ableToOpen, 1, MPI_INT, j, 2001, MPI_COMM_WORLD); - } - }else{ - MPI_Status status; - MPI_Recv(&ableToOpen, 1, MPI_INT, 0, 2001, MPI_COMM_WORLD, &status); + if (ableToOpen == 1) { + if (m->getOutputDir() != "") { //default path is set + string tryPath = m->getOutputDir() + m->getSimpleName(namefileNames[i]); + m->mothurOut("Unable to open " + namefileNames[i] + ". Trying output directory " + tryPath); m->mothurOutEndLine(); + ifstream in2; + ableToOpen = m->openInputFile(tryPath, in2, "noerror"); + in2.close(); + namefileNames[i] = tryPath; } - - #endif - if (ableToOpen == 1) { m->mothurOut("Unable to match name file with fasta file."); m->mothurOutEndLine(); abort = true; } + } + in.close(); + if (ableToOpen == 1) { + m->mothurOut("Unable to open " + namefileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); abort = true; + //erase from file list + namefileNames.erase(namefileNames.begin()+i); + i--; + } + } } @@ -180,6 +261,63 @@ ClassifySeqsCommand::ClassifySeqsCommand(string option) { if (namefileNames.size() != fastaFileNames.size()) { abort = true; m->mothurOut("If you provide a name file, you must have one for each fasta file."); m->mothurOutEndLine(); } } + groupfile = validParameter.validFile(parameters, "group", false); + if (groupfile == "not found") { groupfile = ""; } + else { + m->splitAtDash(groupfile, groupfileNames); + + //go through files and make sure they are good, if not, then disregard them + for (int i = 0; i < groupfileNames.size(); i++) { + if (inputDir != "") { + string path = m->hasPath(groupfileNames[i]); + //if the user has not given a path then, add inputdir. else leave path alone. + if (path == "") { groupfileNames[i] = inputDir + groupfileNames[i]; } + } + int ableToOpen; + + ifstream in; + ableToOpen = m->openInputFile(groupfileNames[i], in, "noerror"); + + //if you can't open it, try default location + if (ableToOpen == 1) { + if (m->getDefaultPath() != "") { //default path is set + string tryPath = m->getDefaultPath() + m->getSimpleName(groupfileNames[i]); + m->mothurOut("Unable to open " + groupfileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine(); + ifstream in2; + ableToOpen = m->openInputFile(tryPath, in2, "noerror"); + in2.close(); + groupfileNames[i] = tryPath; + } + } + + if (ableToOpen == 1) { + if (m->getOutputDir() != "") { //default path is set + string tryPath = m->getOutputDir() + m->getSimpleName(groupfileNames[i]); + m->mothurOut("Unable to open " + groupfileNames[i] + ". Trying output directory " + tryPath); m->mothurOutEndLine(); + ifstream in2; + ableToOpen = m->openInputFile(tryPath, in2, "noerror"); + in2.close(); + groupfileNames[i] = tryPath; + } + } + + in.close(); + + if (ableToOpen == 1) { + m->mothurOut("Unable to open " + groupfileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); groupfileNames[i] = ""; + //erase from file list + groupfileNames.erase(groupfileNames.begin()+i); + i--; + } + } + } + + if (groupfile != "") { + if (groupfileNames.size() != fastaFileNames.size()) { abort = true; m->mothurOut("If you provide a group file, you must have one for each fasta file."); m->mothurOutEndLine(); } + }else { + for (int i = 0; i < fastaFileNames.size(); i++) { groupfileNames.push_back(""); } + } + //check for optional parameter and set defaults // ...at some point should added some additional type checking... string temp; @@ -212,7 +350,7 @@ ClassifySeqsCommand::ClassifySeqsCommand(string option) { convert(temp, cutoff); temp = validParameter.validFile(parameters, "probs", false); if (temp == "not found"){ temp = "true"; } - probs = isTrue(temp); + probs = m->isTrue(temp); temp = validParameter.validFile(parameters, "iters", false); if (temp == "not found") { temp = "100"; } convert(temp, iters); @@ -250,6 +388,7 @@ void ClassifySeqsCommand::help(){ m->mothurOut("The template, fasta and taxonomy parameters are required. You may enter multiple fasta files by separating their names with dashes. ie. fasta=abrecovery.fasta-amzon.fasta \n"); m->mothurOut("The search parameter allows you to specify the method to find most similar template. Your options are: suffix, kmer, blast and distance. The default is kmer.\n"); m->mothurOut("The name parameter allows you add a names file with your fasta file, if you enter multiple fasta files, you must enter matching names files for them.\n"); + m->mothurOut("The group parameter allows you add a group file so you can have the summary totals broken up by group.\n"); m->mothurOut("The method parameter allows you to specify classification method to use. Your options are: bayesian and knn. The default is bayesian.\n"); m->mothurOut("The ksize parameter allows you to specify the kmer size for finding most similar template to candidate. The default is 8.\n"); m->mothurOut("The processors parameter allows you to specify the number of processors to use. The default is 1.\n"); @@ -262,7 +401,7 @@ void ClassifySeqsCommand::help(){ m->mothurOut("The gapextend parameter allows you to specify the penalty for extending a gap in an alignment. The default is -1.0.\n"); m->mothurOut("The numwanted parameter allows you to specify the number of sequence matches you want with the knn method. The default is 10.\n"); m->mothurOut("The cutoff parameter allows you to specify a bootstrap confidence threshold for your taxonomy. The default is 0.\n"); - m->mothurOut("The probs parameter shut off the bootstrapping results for the bayesian method. The default is true, meaning you want the bootstrapping to be run.\n"); + m->mothurOut("The probs parameter shuts off the bootstrapping results for the bayesian method. The default is true, meaning you want the bootstrapping to be shown.\n"); m->mothurOut("The iters parameter allows you to specify how many iterations to do when calculating the bootstrap confidence score for your taxonomy with the bayesian method. The default is 100.\n"); m->mothurOut("The classify.seqs command should be in the following format: \n"); m->mothurOut("classify.seqs(template=yourTemplateFile, fasta=yourFastaFile, method=yourClassificationMethod, search=yourSearchmethod, ksize=yourKmerSize, taxonomy=yourTaxonomyFile, processors=yourProcessors) \n"); @@ -294,29 +433,37 @@ int ClassifySeqsCommand::execute(){ if (m->control_pressed) { delete classify; return 0; } - vector outputNames; for (int s = 0; s < fastaFileNames.size(); s++) { m->mothurOut("Classifying sequences from " + fastaFileNames[s] + " ..." ); m->mothurOutEndLine(); - if (outputDir == "") { outputDir += hasPath(fastaFileNames[s]); } - string newTaxonomyFile = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + getRootName(getSimpleName(taxonomyFileName)) + "taxonomy"; - string tempTaxonomyFile = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + "taxonomy.temp"; - string taxSummary = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + getRootName(getSimpleName(taxonomyFileName)) + "tax.summary"; + string RippedTaxName = m->getRootName(m->getSimpleName(taxonomyFileName)); + RippedTaxName = m->getExtension(RippedTaxName.substr(0, RippedTaxName.length()-1)); + if (RippedTaxName[0] == '.') { RippedTaxName = RippedTaxName.substr(1, RippedTaxName.length()); } + RippedTaxName += "."; + + if (outputDir == "") { outputDir += m->hasPath(fastaFileNames[s]); } + string newTaxonomyFile = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + RippedTaxName + "taxonomy"; + string tempTaxonomyFile = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + "taxonomy.temp"; + string taxSummary = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + RippedTaxName + "tax.summary"; + + if ((method == "knn") && (search == "distance")) { + string DistName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + "match.dist"; + classify->setDistName(DistName); outputNames.push_back(DistName); outputTypes["matchdist"].push_back(DistName); + } - outputNames.push_back(newTaxonomyFile); - outputNames.push_back(taxSummary); + outputNames.push_back(newTaxonomyFile); outputTypes["taxonomy"].push_back(newTaxonomyFile); + outputNames.push_back(taxSummary); outputTypes["taxsummary"].push_back(taxSummary); int start = time(NULL); int numFastaSeqs = 0; for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear(); #ifdef USE_MPI - - int pid, end, numSeqsPerProcessor; + int pid, numSeqsPerProcessor; int tag = 2001; - vector MPIPos; + vector MPIPos; MPI_Status status; MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are @@ -328,60 +475,62 @@ int ClassifySeqsCommand::execute(){ int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY; int inMode=MPI_MODE_RDONLY; - - char outNewTax[newTaxonomyFile.length()]; + + char outNewTax[1024]; strcpy(outNewTax, newTaxonomyFile.c_str()); - char outTempTax[tempTaxonomyFile.length()]; + char outTempTax[1024]; strcpy(outTempTax, tempTaxonomyFile.c_str()); - char inFileName[fastaFileNames[s].length()]; + char inFileName[1024]; strcpy(inFileName, fastaFileNames[s].c_str()); MPI_File_open(MPI_COMM_WORLD, inFileName, inMode, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer MPI_File_open(MPI_COMM_WORLD, outNewTax, outMode, MPI_INFO_NULL, &outMPINewTax); MPI_File_open(MPI_COMM_WORLD, outTempTax, outMode, MPI_INFO_NULL, &outMPITempTax); - if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); delete classify; return 0; } - - if(namefile != "") { MPIReadNamesFile(namefileNames[s]); } + if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); delete classify; return 0; } if (pid == 0) { //you are the root process - MPIPos = setFilePosFasta(fastaFileNames[s], numFastaSeqs); //fills MPIPos, returns numSeqs + MPIPos = m->setFilePosFasta(fastaFileNames[s], numFastaSeqs); //fills MPIPos, returns numSeqs //send file positions to all processes - MPI_Bcast(&numFastaSeqs, 1, MPI_INT, 0, MPI_COMM_WORLD); //send numSeqs - MPI_Bcast(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, 0, MPI_COMM_WORLD); //send file pos + for(int i = 1; i < processors; i++) { + MPI_Send(&numFastaSeqs, 1, MPI_INT, i, tag, MPI_COMM_WORLD); + MPI_Send(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, i, tag, MPI_COMM_WORLD); + } //figure out how many sequences you have to align numSeqsPerProcessor = numFastaSeqs / processors; - if(pid == (processors - 1)){ numSeqsPerProcessor = numFastaSeqs - pid * numSeqsPerProcessor; } int startIndex = pid * numSeqsPerProcessor; + if(pid == (processors - 1)){ numSeqsPerProcessor = numFastaSeqs - pid * numSeqsPerProcessor; } + //align your part driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPINewTax, outMPITempTax, MPIPos); - if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete classify; return 0; } + if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete classify; return 0; } for (int i = 1; i < processors; i++) { int done; MPI_Recv(&done, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status); } }else{ //you are a child process - MPI_Bcast(&numFastaSeqs, 1, MPI_INT, 0, MPI_COMM_WORLD); //get numSeqs + MPI_Recv(&numFastaSeqs, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status); MPIPos.resize(numFastaSeqs+1); - MPI_Bcast(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, 0, MPI_COMM_WORLD); //get file positions + MPI_Recv(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status); //figure out how many sequences you have to align numSeqsPerProcessor = numFastaSeqs / processors; - if(pid == (processors - 1)){ numSeqsPerProcessor = numFastaSeqs - pid * numSeqsPerProcessor; } int startIndex = pid * numSeqsPerProcessor; + if(pid == (processors - 1)){ numSeqsPerProcessor = numFastaSeqs - pid * numSeqsPerProcessor; } + //align your part driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPINewTax, outMPITempTax, MPIPos); - if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); delete classify; return 0; } + if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); delete classify; return 0; } int done = 0; MPI_Send(&done, 1, MPI_INT, 0, tag, MPI_COMM_WORLD); @@ -391,156 +540,125 @@ int ClassifySeqsCommand::execute(){ MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); + MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case #else - - //read namefile - if(namefile != "") { - nameMap.clear(); //remove old names - - ifstream inNames; - openInputFile(namefileNames[s], inNames); + + vector positions = m->divideFile(fastaFileNames[s], processors); - string firstCol, secondCol; - while(!inNames.eof()) { - inNames >> firstCol >> secondCol; gobble(inNames); - nameMap[firstCol] = getNumNames(secondCol); //ex. seq1 seq1,seq3,seq5 -> seq1 = 3. - } - inNames.close(); - } - - #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + for (int i = 0; i < (positions.size()-1); i++) { + lines.push_back(new linePair(positions[i], positions[(i+1)])); + } + + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) if(processors == 1){ - ifstream inFASTA; - openInputFile(fastaFileNames[s], inFASTA); - numFastaSeqs=count(istreambuf_iterator(inFASTA),istreambuf_iterator(), '>'); - inFASTA.close(); - - lines.push_back(new linePair(0, numFastaSeqs)); - - driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]); + numFastaSeqs = driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]); } else{ - vector positions; processIDS.resize(0); - ifstream inFASTA; - openInputFile(fastaFileNames[s], inFASTA); - - string input; - while(!inFASTA.eof()){ - input = getline(inFASTA); - if (input.length() != 0) { - if(input[0] == '>'){ int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); } - } - } - inFASTA.close(); - - numFastaSeqs = positions.size(); - - int numSeqsPerProcessor = numFastaSeqs / processors; - - for (int i = 0; i < processors; i++) { - int startPos = positions[ i * numSeqsPerProcessor ]; - if(i == processors - 1){ - numSeqsPerProcessor = numFastaSeqs - i * numSeqsPerProcessor; - } - lines.push_back(new linePair(startPos, numSeqsPerProcessor)); - } - createProcesses(newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]); - - rename((newTaxonomyFile + toString(processIDS[0]) + ".temp").c_str(), newTaxonomyFile.c_str()); - rename((tempTaxonomyFile + toString(processIDS[0]) + ".temp").c_str(), tempTaxonomyFile.c_str()); - - for(int i=1;i(inFASTA),istreambuf_iterator(), '>'); - inFASTA.close(); - - lines.push_back(new linePair(0, numFastaSeqs)); - - driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]); + numFastaSeqs = driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]); #endif #endif - - delete classify; - + + m->mothurOutEndLine(); + m->mothurOut("It took " + toString(time(NULL) - start) + " secs to classify " + toString(numFastaSeqs) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine(); + start = time(NULL); + + #ifdef USE_MPI if (pid == 0) { //this part does not need to be paralellized + + if(namefile != "") { m->mothurOut("Reading " + namefileNames[s] + "..."); cout.flush(); MPIReadNamesFile(namefileNames[s]); m->mothurOut(" Done."); m->mothurOutEndLine(); } + #else + //read namefile + if(namefile != "") { + + m->mothurOut("Reading " + namefileNames[s] + "..."); cout.flush(); + + nameMap.clear(); //remove old names + + ifstream inNames; + m->openInputFile(namefileNames[s], inNames); + + string firstCol, secondCol; + while(!inNames.eof()) { + inNames >> firstCol >> secondCol; m->gobble(inNames); + + vector temp; + m->splitAtComma(secondCol, temp); + + nameMap[firstCol] = temp; + } + inNames.close(); + + m->mothurOut(" Done."); m->mothurOutEndLine(); + } #endif - //make taxonomy tree from new taxonomy file - PhyloTree taxaBrowser; + string group = ""; + if (groupfile != "") { group = groupfileNames[s]; } - if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete classify; return 0; } + PhyloSummary taxaSum(taxonomyFileName, group); - ifstream in; - openInputFile(tempTaxonomyFile, in); + if (m->control_pressed) { outputTypes.clear(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete classify; return 0; } - //read in users taxonomy file and add sequences to tree - string name, taxon; - - while(!in.eof()){ - in >> name >> taxon; gobble(in); + if (namefile == "") { taxaSum.summarize(tempTaxonomyFile); } + else { + ifstream in; + m->openInputFile(tempTaxonomyFile, in); - if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } remove(tempTaxonomyFile.c_str()); delete classify; return 0; } + //read in users taxonomy file and add sequences to tree + string name, taxon; - if (namefile != "") { + while(!in.eof()){ + in >> name >> taxon; m->gobble(in); + itNames = nameMap.find(name); if (itNames == nameMap.end()) { m->mothurOut(name + " is not in your name file please correct."); m->mothurOutEndLine(); exit(1); }else{ - for (int i = 0; i < itNames->second; i++) { - taxaBrowser.addSeqToTree(name+toString(i), taxon); //add it as many times as there are identical seqs + for (int i = 0; i < itNames->second.size(); i++) { + taxaSum.addSeqToTree(itNames->second[i], taxon); //add it as many times as there are identical seqs } + itNames->second.clear(); + nameMap.erase(itNames->first); } - }else { taxaBrowser.addSeqToTree(name, taxon); } //add it once + } + in.close(); } - in.close(); - - taxaBrowser.assignHeirarchyIDs(0); - - if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } remove(tempTaxonomyFile.c_str()); delete classify; return 0; } - - taxaBrowser.binUnclassified(); - remove(tempTaxonomyFile.c_str()); - if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete classify; return 0; } - + if (m->control_pressed) { outputTypes.clear(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete classify; return 0; } //print summary file ofstream outTaxTree; - openOutputFile(taxSummary, outTaxTree); - taxaBrowser.print(outTaxTree); + m->openOutputFile(taxSummary, outTaxTree); + taxaSum.print(outTaxTree); outTaxTree.close(); //output taxonomy with the unclassified bins added ifstream inTax; - openInputFile(newTaxonomyFile, inTax); + m->openInputFile(newTaxonomyFile, inTax); ofstream outTax; string unclass = newTaxonomyFile + ".unclass.temp"; - openOutputFile(unclass, outTax); + m->openOutputFile(unclass, outTax); //get maxLevel from phylotree so you know how many 'unclassified's to add - int maxLevel = taxaBrowser.getMaxLevel(); - + int maxLevel = taxaSum.getMaxLevel(); + //read taxfile - this reading and rewriting is done to preserve the confidence scores. + string name, taxon; while (!inTax.eof()) { - if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } remove(unclass.c_str()); delete classify; return 0; } + if (m->control_pressed) { outputTypes.clear(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } remove(unclass.c_str()); delete classify; return 0; } - inTax >> name >> taxon; gobble(inTax); + inTax >> name >> taxon; m->gobble(inTax); string newTax = addUnclassifieds(taxon, maxLevel); @@ -552,6 +670,9 @@ int ClassifySeqsCommand::execute(){ remove(newTaxonomyFile.c_str()); rename(unclass.c_str(), newTaxonomyFile.c_str()); + m->mothurOutEndLine(); + m->mothurOut("It took " + toString(time(NULL) - start) + " secs to create the summary file for " + toString(numFastaSeqs) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine(); + #ifdef USE_MPI } #endif @@ -560,12 +681,9 @@ int ClassifySeqsCommand::execute(){ m->mothurOut("Output File Names: "); m->mothurOutEndLine(); for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); } m->mothurOutEndLine(); - - - m->mothurOutEndLine(); - m->mothurOut("It took " + toString(time(NULL) - start) + " secs to classify " + toString(numFastaSeqs) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine(); } + delete classify; return 0; } catch(exception& e) { @@ -605,11 +723,11 @@ string ClassifySeqsCommand::addUnclassifieds(string tax, int maxlevel) { /**************************************************************************************************/ -void ClassifySeqsCommand::createProcesses(string taxFileName, string tempTaxFile, string filename) { +int ClassifySeqsCommand::createProcesses(string taxFileName, string tempTaxFile, string filename) { try { #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - int process = 0; - // processIDS.resize(0); + int process = 1; + int num = 0; //loop through and create all the processes you want while (process != processors) { @@ -619,16 +737,48 @@ void ClassifySeqsCommand::createProcesses(string taxFileName, string tempTaxFile processIDS.push_back(pid); //create map from line number to pid so you can append files in correct order later process++; }else if (pid == 0){ - driver(lines[process], taxFileName + toString(getpid()) + ".temp", tempTaxFile + toString(getpid()) + ".temp", filename); + num = driver(lines[process], taxFileName + toString(getpid()) + ".temp", tempTaxFile + toString(getpid()) + ".temp", filename); + + //pass numSeqs to parent + ofstream out; + string tempFile = filename + toString(getpid()) + ".num.temp"; + m->openOutputFile(tempFile, out); + out << num << endl; + out.close(); + exit(0); - }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); } + }else { + m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine(); + for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); } + exit(0); + } } + //parent does its part + num = driver(lines[0], taxFileName, tempTaxFile, filename); + //force parent to wait until all the processes are done - for (int i=0;iopenInputFile(tempFile, in); + if (!in.eof()) { int tempNum = 0; in >> tempNum; num += tempNum; } + in.close(); remove(tempFile.c_str()); + } + + for(int i=0;iopenOutputFileAppend(filename, output); + m->openInputFile(temp, input); while(char c = input.get()){ if(input.eof()) { break; } @@ -662,32 +812,36 @@ void ClassifySeqsCommand::appendTaxFiles(string temp, string filename) { //********************************************************************************************************************** -int ClassifySeqsCommand::driver(linePair* line, string taxFName, string tempTFName, string filename){ +int ClassifySeqsCommand::driver(linePair* filePos, string taxFName, string tempTFName, string filename){ try { ofstream outTax; - openOutputFile(taxFName, outTax); + m->openOutputFile(taxFName, outTax); ofstream outTaxSimple; - openOutputFile(tempTFName, outTaxSimple); + m->openOutputFile(tempTFName, outTaxSimple); ifstream inFASTA; - openInputFile(filename, inFASTA); - - inFASTA.seekg(line->start); + m->openInputFile(filename, inFASTA); string taxonomy; - for(int i=0;inumSeqs;i++){ + inFASTA.seekg(filePos->start); + + bool done = false; + int count = 0; + + while (!done) { if (m->control_pressed) { return 0; } + + Sequence* candidateSeq = new Sequence(inFASTA); m->gobble(inFASTA); - Sequence* candidateSeq = new Sequence(inFASTA); - if (candidateSeq->getName() != "") { + taxonomy = classify->getTaxonomy(candidateSeq); if (m->control_pressed) { delete candidateSeq; return 0; } - if ((taxonomy != "bad seq") && (taxonomy != "")) { + if (taxonomy != "bad seq") { //output confidence scores or not if (probs) { outTax << candidateSeq->getName() << '\t' << taxonomy << endl; @@ -696,20 +850,30 @@ int ClassifySeqsCommand::driver(linePair* line, string taxFName, string tempTFNa } outTaxSimple << candidateSeq->getName() << '\t' << classify->getSimpleTax() << endl; - }else{ m->mothurOut("Sequence: " + candidateSeq->getName() + " is bad."); m->mothurOutEndLine(); } - } + } + count++; + } delete candidateSeq; - if((i+1) % 100 == 0){ - m->mothurOut("Classifying sequence " + toString(i+1)); m->mothurOutEndLine(); - } + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + unsigned long int pos = inFASTA.tellg(); + if ((pos == -1) || (pos >= filePos->end)) { break; } + #else + if (inFASTA.eof()) { break; } + #endif + + //report progress + if((count) % 100 == 0){ m->mothurOut("Processing sequence: " + toString(count)); m->mothurOutEndLine(); } + } - + //report progress + if((count) % 100 != 0){ m->mothurOut("Processing sequence: " + toString(count)); m->mothurOutEndLine(); } + inFASTA.close(); outTax.close(); outTaxSimple.close(); - return 1; + return count; } catch(exception& e) { m->errorOut(e, "ClassifySeqsCommand", "driver"); @@ -718,7 +882,7 @@ int ClassifySeqsCommand::driver(linePair* line, string taxFName, string tempTFNa } //********************************************************************************************************************** #ifdef USE_MPI -int ClassifySeqsCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& newFile, MPI_File& tempFile, vector& MPIPos){ +int ClassifySeqsCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& newFile, MPI_File& tempFile, vector& MPIPos){ try { MPI_Status statusNew; MPI_Status statusTemp; @@ -736,19 +900,20 @@ int ClassifySeqsCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File //read next sequence int length = MPIPos[start+i+1] - MPIPos[start+i]; - char buf4[length]; + char* buf4 = new char[length]; MPI_File_read_at(inMPI, MPIPos[start+i], buf4, length, MPI_CHAR, &status); string tempBuf = buf4; if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); } istringstream iss (tempBuf,istringstream::in); + delete buf4; Sequence* candidateSeq = new Sequence(iss); if (candidateSeq->getName() != "") { taxonomy = classify->getTaxonomy(candidateSeq); - if ((taxonomy != "bad seq") && (taxonomy != "")) { + if (taxonomy != "bad seq") { //output confidence scores or not if (probs) { outputString = candidateSeq->getName() + "\t" + taxonomy + "\n"; @@ -757,18 +922,20 @@ int ClassifySeqsCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File } int length = outputString.length(); - char buf2[length]; - strcpy(buf2, outputString.c_str()); + char* buf2 = new char[length]; + memcpy(buf2, outputString.c_str(), length); MPI_File_write_shared(newFile, buf2, length, MPI_CHAR, &statusNew); - + delete buf2; + outputString = candidateSeq->getName() + "\t" + classify->getSimpleTax() + "\n"; length = outputString.length(); - char buf[length]; - strcpy(buf, outputString.c_str()); + char* buf = new char[length]; + memcpy(buf, outputString.c_str(), length); MPI_File_write_shared(tempFile, buf, length, MPI_CHAR, &statusTemp); - }else{ cout << "Sequence: " << candidateSeq->getName() << " is bad." << endl; } + delete buf; + } } delete candidateSeq; @@ -795,24 +962,33 @@ int ClassifySeqsCommand::MPIReadNamesFile(string nameFilename){ MPI_File inMPI; MPI_Offset size; MPI_Status status; + + //char* inFileName = new char[nameFilename.length()]; + //memcpy(inFileName, nameFilename.c_str(), nameFilename.length()); - char inFileName[nameFilename.length()]; + char inFileName[1024]; strcpy(inFileName, nameFilename.c_str()); MPI_File_open(MPI_COMM_WORLD, inFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI); MPI_File_get_size(inMPI, &size); + //delete inFileName; - char buffer[size]; + char* buffer = new char[size]; MPI_File_read(inMPI, buffer, size, MPI_CHAR, &status); string tempBuf = buffer; if (tempBuf.length() > size) { tempBuf = tempBuf.substr(0, size); } istringstream iss (tempBuf,istringstream::in); + delete buffer; string firstCol, secondCol; while(!iss.eof()) { - iss >> firstCol >> secondCol; gobble(iss); - nameMap[firstCol] = getNumNames(secondCol); //ex. seq1 seq1,seq3,seq5 -> seq1 = 3. + iss >> firstCol >> secondCol; m->gobble(iss); + + vector temp; + m->splitAtComma(secondCol, temp); + + nameMap[firstCol] = temp; } MPI_File_close(&inMPI);