X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=classifyseqscommand.cpp;h=6f8bc7ee1ae7436dc395291bf078336aaff1889f;hb=ca9ac1d80c62f57270b0dcd49410ebe08a8aecd6;hp=ba854e98b40403c8d963fb65c9c4bd6178950947;hpb=aa9238c0a9e6e7aa0ed8b8b606b08ad4fd7dcfe3;p=mothur.git diff --git a/classifyseqscommand.cpp b/classifyseqscommand.cpp index ba854e9..6f8bc7e 100644 --- a/classifyseqscommand.cpp +++ b/classifyseqscommand.cpp @@ -11,27 +11,109 @@ #include "sequence.hpp" #include "bayesian.h" #include "phylotree.h" +#include "phylosummary.h" #include "knn.h" -//********************************************************************************************************************** +//********************************************************************************************************************** +vector ClassifySeqsCommand::setParameters(){ + try { + CommandParameter ptaxonomy("taxonomy", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(ptaxonomy); + CommandParameter ptemplate("reference", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(ptemplate); + CommandParameter pfasta("fasta", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(pfasta); + CommandParameter pname("name", "InputTypes", "", "", "none", "none", "none",false,false); parameters.push_back(pname); + CommandParameter pgroup("group", "InputTypes", "", "", "none", "none", "none",false,false); parameters.push_back(pgroup); + CommandParameter psearch("search", "Multiple", "kmer-blast-suffix-distance", "kmer", "", "", "",false,false); parameters.push_back(psearch); + CommandParameter pksize("ksize", "Number", "", "8", "", "", "",false,false); parameters.push_back(pksize); + CommandParameter pmethod("method", "Multiple", "bayesian-knn", "bayesian", "", "", "",false,false); parameters.push_back(pmethod); + CommandParameter pprocessors("processors", "Number", "", "1", "", "", "",false,false); parameters.push_back(pprocessors); + CommandParameter pmatch("match", "Number", "", "1.0", "", "", "",false,false); parameters.push_back(pmatch); + CommandParameter pmismatch("mismatch", "Number", "", "-1.0", "", "", "",false,false); parameters.push_back(pmismatch); + CommandParameter pgapopen("gapopen", "Number", "", "-2.0", "", "", "",false,false); parameters.push_back(pgapopen); + CommandParameter pgapextend("gapextend", "Number", "", "-1.0", "", "", "",false,false); parameters.push_back(pgapextend); + CommandParameter pcutoff("cutoff", "Number", "", "0", "", "", "",false,true); parameters.push_back(pcutoff); + CommandParameter pprobs("probs", "Boolean", "", "T", "", "", "",false,false); parameters.push_back(pprobs); + CommandParameter piters("iters", "Number", "", "100", "", "", "",false,true); parameters.push_back(piters); + CommandParameter pnumwanted("numwanted", "Number", "", "10", "", "", "",false,true); parameters.push_back(pnumwanted); + CommandParameter pinputdir("inputdir", "String", "", "", "", "", "",false,false); parameters.push_back(pinputdir); + CommandParameter poutputdir("outputdir", "String", "", "", "", "", "",false,false); parameters.push_back(poutputdir); + + vector myArray; + for (int i = 0; i < parameters.size(); i++) { myArray.push_back(parameters[i].name); } + return myArray; + } + catch(exception& e) { + m->errorOut(e, "ClassifySeqsCommand", "setParameters"); + exit(1); + } +} +//********************************************************************************************************************** +string ClassifySeqsCommand::getHelpString(){ + try { + string helpString = ""; + helpString += "The classify.seqs command reads a fasta file containing sequences and creates a .taxonomy file and a .tax.summary file.\n"; + helpString += "The classify.seqs command parameters are reference, fasta, name, search, ksize, method, taxonomy, processors, match, mismatch, gapopen, gapextend, numwanted and probs.\n"; + helpString += "The reference, 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"; + helpString += "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"; + helpString += "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"; + helpString += "The group parameter allows you add a group file so you can have the summary totals broken up by group.\n"; + helpString += "The method parameter allows you to specify classification method to use. Your options are: bayesian and knn. The default is bayesian.\n"; + helpString += "The ksize parameter allows you to specify the kmer size for finding most similar template to candidate. The default is 8.\n"; + helpString += "The processors parameter allows you to specify the number of processors to use. The default is 1.\n"; +#ifdef USE_MPI + helpString += "When using MPI, the processors parameter is set to the number of MPI processes running. \n"; +#endif + helpString += "The match parameter allows you to specify the bonus for having the same base. The default is 1.0.\n"; + helpString += "The mistmatch parameter allows you to specify the penalty for having different bases. The default is -1.0.\n"; + helpString += "The gapopen parameter allows you to specify the penalty for opening a gap in an alignment. The default is -2.0.\n"; + helpString += "The gapextend parameter allows you to specify the penalty for extending a gap in an alignment. The default is -1.0.\n"; + helpString += "The numwanted parameter allows you to specify the number of sequence matches you want with the knn method. The default is 10.\n"; + helpString += "The cutoff parameter allows you to specify a bootstrap confidence threshold for your taxonomy. The default is 0.\n"; + helpString += "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"; + helpString += "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"; + helpString += "The classify.seqs command should be in the following format: \n"; + helpString += "classify.seqs(reference=yourTemplateFile, fasta=yourFastaFile, method=yourClassificationMethod, search=yourSearchmethod, ksize=yourKmerSize, taxonomy=yourTaxonomyFile, processors=yourProcessors) \n"; + helpString += "Example classify.seqs(fasta=amazon.fasta, reference=core.filtered, method=knn, search=gotoh, ksize=8, processors=2)\n"; + helpString += "The .taxonomy file consists of 2 columns: 1 = your sequence name, 2 = the taxonomy for your sequence. \n"; + helpString += "The .tax.summary is a summary of the different taxonomies represented in your fasta file. \n"; + helpString += "Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFastaFile).\n\n"; + return helpString; + } + catch(exception& e) { + m->errorOut(e, "ClassifySeqsCommand", "getHelpString"); + exit(1); + } +} +//********************************************************************************************************************** +ClassifySeqsCommand::ClassifySeqsCommand(){ + try { + abort = true; calledHelp = true; + setParameters(); + vector tempOutNames; + outputTypes["taxonomy"] = tempOutNames; + outputTypes["taxsummary"] = tempOutNames; + outputTypes["matchdist"] = tempOutNames; + } + catch(exception& e) { + m->errorOut(e, "ClassifySeqsCommand", "ClassifySeqsCommand"); + exit(1); + } +} +//********************************************************************************************************************** ClassifySeqsCommand::ClassifySeqsCommand(string option) { try { - abort = false; + abort = false; calledHelp = false; //allow user to run help - if(option == "help") { help(); abort = true; } + if(option == "help") { help(); abort = true; calledHelp = true; } 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"}; - vector myArray (AlignArray, AlignArray+(sizeof(AlignArray)/sizeof(string))); + vector myArray = setParameters(); 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 +121,12 @@ 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 = ""; } @@ -47,55 +135,96 @@ ClassifySeqsCommand::ClassifySeqsCommand(string option) { if (inputDir == "not found"){ inputDir = ""; } else { string path; - it = parameters.find("template"); + it = parameters.find("reference"); //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; } + if (path == "") { parameters["reference"] = inputDir + it->second; } } 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 - templateFileName = validParameter.validFile(parameters, "template", true); + templateFileName = validParameter.validFile(parameters, "reference", true); if (templateFileName == "not found") { - m->mothurOut("template is a required parameter for the classify.seqs command."); + m->mothurOut("reference is a required parameter for the classify.seqs command."); m->mothurOutEndLine(); abort = true; } 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; } + if (fastaFileName == "not found") { + //if there is a current fasta file, use it + string filename = m->getFastaFile(); + if (filename != "") { fastaFileNames.push_back(filename); m->mothurOut("Using " + filename + " as input file for the fasta parameter."); m->mothurOutEndLine(); } + else { m->mothurOut("You have no current fastafile and the fasta parameter is required."); 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; + ifstream in; - ableToOpen = openInputFile(fastaFileNames[i], in); + 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; + } + } + + 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; + } + } + + 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--; } - in.close(); + } //make sure there is at least one valid file left @@ -116,21 +245,51 @@ 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; + ifstream in; - ableToOpen = openInputFile(namefileNames[i], in); - if (ableToOpen == 1) { m->mothurOut("Unable to match name file with fasta file."); m->mothurOutEndLine(); abort = true; } + 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; + } + } + + 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; + } + } 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--; + } + } } @@ -138,13 +297,71 @@ 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; temp = validParameter.validFile(parameters, "ksize", false); if (temp == "not found"){ temp = "8"; } convert(temp, kmerSize); - temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = "1"; } + temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); } + m->setProcessors(temp); convert(temp, processors); search = validParameter.validFile(parameters, "search", false); if (search == "not found"){ search = "kmer"; } @@ -170,7 +387,7 @@ 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); @@ -191,53 +408,16 @@ ClassifySeqsCommand::ClassifySeqsCommand(string option) { } //********************************************************************************************************************** - ClassifySeqsCommand::~ClassifySeqsCommand(){ - if (abort == false) { for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear(); } } - -//********************************************************************************************************************** - -void ClassifySeqsCommand::help(){ - try { - m->mothurOut("The classify.seqs command reads a fasta file containing sequences and creates a .taxonomy file and a .tax.summary file.\n"); - m->mothurOut("The classify.seqs command parameters are template, fasta, name, search, ksize, method, taxonomy, processors, match, mismatch, gapopen, gapextend, numwanted and probs.\n"); - 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 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"); - m->mothurOut("The match parameter allows you to specify the bonus for having the same base. The default is 1.0.\n"); - m->mothurOut("The mistmatch parameter allows you to specify the penalty for having different bases. The default is -1.0.\n"); - m->mothurOut("The gapopen parameter allows you to specify the penalty for opening a gap in an alignment. The default is -2.0.\n"); - 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 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"); - m->mothurOut("Example classify.seqs(fasta=amazon.fasta, template=core.filtered, method=knn, search=gotoh, ksize=8, processors=2)\n"); - m->mothurOut("The .taxonomy file consists of 2 columns: 1 = your sequence name, 2 = the taxonomy for your sequence. \n"); - m->mothurOut("The .tax.summary is a summary of the different taxonomies represented in your fasta file. \n"); - m->mothurOut("Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFastaFile).\n\n"); - } - catch(exception& e) { - m->errorOut(e, "ClassifySeqsCommand", "help"); - exit(1); - } -} - - //********************************************************************************************************************** int ClassifySeqsCommand::execute(){ try { - if (abort == true) { return 0; } + if (abort == true) { if (calledHelp) { return 0; } return 2; } if(method == "bayesian"){ classify = new Bayesian(taxonomyFileName, templateFileName, search, kmerSize, cutoff, iters); } else if(method == "knn"){ classify = new Knn(taxonomyFileName, templateFileName, search, kmerSize, gapOpen, gapExtend, match, misMatch, numWanted); } @@ -249,162 +429,232 @@ int ClassifySeqsCommand::execute(){ if (m->control_pressed) { delete classify; return 0; } - vector outputNames; for (int s = 0; s < fastaFileNames.size(); s++) { - //read namefile - if(namefile != "") { - nameMap.clear(); //remove old names - - ifstream inNames; - openInputFile(namefileNames[s], inNames); - - string firstCol, secondCol; - while(!inNames.eof()) { - inNames >> firstCol >> secondCol; gobble(inNames); - nameMap[firstCol] = getNumNames(secondCol); //ex. seq1 seq1,seq3,seq5 -> seq1 = 3. - } - inNames.close(); - } - 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(); -#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - if(processors == 1){ - ifstream inFASTA; - openInputFile(fastaFileNames[s], inFASTA); - numFastaSeqs=count(istreambuf_iterator(inFASTA),istreambuf_iterator(), '>'); - inFASTA.close(); +#ifdef USE_MPI + int pid, numSeqsPerProcessor; + int tag = 2001; + vector MPIPos; - lines.push_back(new linePair(0, numFastaSeqs)); + MPI_Status status; + MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are + MPI_Comm_size(MPI_COMM_WORLD, &processors); + + MPI_File inMPI; + MPI_File outMPINewTax; + MPI_File outMPITempTax; + + int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY; + int inMode=MPI_MODE_RDONLY; - driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]); - } - else{ - vector positions; - processIDS.resize(0); + char outNewTax[1024]; + strcpy(outNewTax, newTaxonomyFile.c_str()); - ifstream inFASTA; - openInputFile(fastaFileNames[s], inFASTA); + char outTempTax[1024]; + strcpy(outTempTax, tempTaxonomyFile.c_str()); - 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(); + 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); - numFastaSeqs = positions.size(); + if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); delete classify; return 0; } - int numSeqsPerProcessor = numFastaSeqs / processors; + if (pid == 0) { //you are the root process + + MPIPos = m->setFilePosFasta(fastaFileNames[s], numFastaSeqs); //fills MPIPos, returns numSeqs + + //send file positions to all processes + 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; + int startIndex = pid * numSeqsPerProcessor; + if(pid == (processors - 1)){ numSeqsPerProcessor = numFastaSeqs - pid * numSeqsPerProcessor; } + - for (int i = 0; i < processors; i++) { - int startPos = positions[ i * numSeqsPerProcessor ]; - if(i == processors - 1){ - numSeqsPerProcessor = numFastaSeqs - i * numSeqsPerProcessor; + //align your part + driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPINewTax, outMPITempTax, MPIPos); + + 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); } - lines.push_back(new linePair(startPos, numSeqsPerProcessor)); + }else{ //you are a child process + MPI_Recv(&numFastaSeqs, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status); + MPIPos.resize(numFastaSeqs+1); + 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; + 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) { 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); } - 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()); + //close files + MPI_File_close(&inMPI); + MPI_File_close(&outMPINewTax); + MPI_File_close(&outMPITempTax); + MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case - for(int i=1;i positions = m->divideFile(fastaFileNames[s], processors); + 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){ + numFastaSeqs = driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]); } -#else - ifstream inFASTA; - openInputFile(fastaFileNames[s], inFASTA); - numFastaSeqs=count(istreambuf_iterator(inFASTA),istreambuf_iterator(), '>'); - inFASTA.close(); + else{ + processIDS.resize(0); + + numFastaSeqs = createProcesses(newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]); + + } + #else + numFastaSeqs = driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]); + #endif +#endif + + 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 - lines.push_back(new linePair(0, numFastaSeqs)); + 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); - driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]); -#endif - //make taxonomy tree from new taxonomy file - PhyloTree taxaBrowser; + nameMap[firstCol] = temp; + } + inNames.close(); + + m->mothurOut(" Done."); m->mothurOutEndLine(); + } + #endif + + 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); @@ -416,14 +666,24 @@ 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 + m->mothurOutEndLine(); 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(); + } + + //set taxonomy file as new current taxonomyfile + string current = ""; + itTypes = outputTypes.find("taxonomy"); + if (itTypes != outputTypes.end()) { + if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setTaxonomyFile(current); } } delete classify; @@ -466,11 +726,11 @@ 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) { @@ -480,16 +740,48 @@ 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("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine(); + for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); } exit(0); - }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); 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; } @@ -523,27 +815,31 @@ 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); + + Sequence* candidateSeq = new Sequence(inFASTA); m->gobble(inFASTA); if (candidateSeq->getName() != "") { + taxonomy = classify->getTaxonomy(candidateSeq); if (m->control_pressed) { delete candidateSeq; return 0; } @@ -558,24 +854,154 @@ int ClassifySeqsCommand::driver(linePair* line, string taxFName, string tempTFNa outTaxSimple << candidateSeq->getName() << '\t' << classify->getSimpleTax() << endl; } - } + 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"); exit(1); } } +//********************************************************************************************************************** +#ifdef USE_MPI +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; + MPI_Status status; + + int pid; + MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are + + string taxonomy; + string outputString; + + for(int i=0;icontrol_pressed) { return 0; } + + //read next sequence + int length = MPIPos[start+i+1] - MPIPos[start+i]; + 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") { + //output confidence scores or not + if (probs) { + outputString = candidateSeq->getName() + "\t" + taxonomy + "\n"; + }else{ + outputString = candidateSeq->getName() + "\t" + classify->getSimpleTax() + "\n"; + } + + int length = outputString.length(); + 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 = new char[length]; + memcpy(buf, outputString.c_str(), length); + + MPI_File_write_shared(tempFile, buf, length, MPI_CHAR, &statusTemp); + delete buf; + } + } + delete candidateSeq; + + if((i+1) % 100 == 0){ cout << "Classifying sequence " << (i+1) << endl; } + } + + if(num % 100 != 0){ cout << "Classifying sequence " << (num) << endl; } + + + return 1; + } + catch(exception& e) { + m->errorOut(e, "ClassifySeqsCommand", "driverMPI"); + exit(1); + } +} + +//********************************************************************************************************************** +int ClassifySeqsCommand::MPIReadNamesFile(string nameFilename){ + try { + + nameMap.clear(); //remove old names + + MPI_File inMPI; + MPI_Offset size; + MPI_Status status; + + //char* inFileName = new char[nameFilename.length()]; + //memcpy(inFileName, nameFilename.c_str(), 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 = 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; m->gobble(iss); + + vector temp; + m->splitAtComma(secondCol, temp); + + nameMap[firstCol] = temp; + } + + MPI_File_close(&inMPI); + + return 1; + } + catch(exception& e) { + m->errorOut(e, "ClassifySeqsCommand", "MPIReadNamesFile"); + exit(1); + } +} +#endif /**************************************************************************************************/