X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=aligncommand.cpp;h=f757a7920bb337883a2a7773f35ed533d2186c80;hp=613f9b7d6f13b04dec679a162ba71faf25326c65;hb=050a3ff02473a3d4c0980964e1a9ebe52e55d6b8;hpb=d84bb41d7dadcfa2c67ce5edb9b94060e5659fa7 diff --git a/aligncommand.cpp b/aligncommand.cpp index 613f9b7..f757a79 100644 --- a/aligncommand.cpp +++ b/aligncommand.cpp @@ -15,32 +15,110 @@ */ #include "aligncommand.h" -#include "sequence.hpp" - -#include "gotohoverlap.hpp" -#include "needlemanoverlap.hpp" -#include "blastalign.hpp" -#include "noalign.hpp" - -#include "nast.hpp" -#include "nastreport.hpp" - +#include "referencedb.h" //********************************************************************************************************************** +vector AlignCommand::setParameters(){ + try { + CommandParameter ptemplate("reference", "InputTypes", "", "", "none", "none", "none","",false,true,true); parameters.push_back(ptemplate); + CommandParameter pcandidate("fasta", "InputTypes", "", "", "none", "none", "none","fasta-alignreport-accnos",false,true,true); parameters.push_back(pcandidate); + CommandParameter psearch("search", "Multiple", "kmer-blast-suffix", "kmer", "", "", "","",false,false,true); parameters.push_back(psearch); + CommandParameter pksize("ksize", "Number", "", "8", "", "", "","",false,false); parameters.push_back(pksize); + CommandParameter pmatch("match", "Number", "", "1.0", "", "", "","",false,false); parameters.push_back(pmatch); + CommandParameter palign("align", "Multiple", "needleman-gotoh-blast-noalign", "needleman", "", "", "","",false,false,true); parameters.push_back(palign); + CommandParameter pmismatch("mismatch", "Number", "", "-1.0", "", "", "","",false,false); parameters.push_back(pmismatch); + CommandParameter pgapopen("gapopen", "Number", "", "-5.0", "", "", "","",false,false); parameters.push_back(pgapopen); + CommandParameter pgapextend("gapextend", "Number", "", "-2.0", "", "", "","",false,false); parameters.push_back(pgapextend); + CommandParameter pprocessors("processors", "Number", "", "1", "", "", "","",false,false,true); parameters.push_back(pprocessors); + CommandParameter pflip("flip", "Boolean", "", "F", "", "", "","",false,false); parameters.push_back(pflip); + CommandParameter psave("save", "Boolean", "", "F", "", "", "","",false,false); parameters.push_back(psave); + CommandParameter pthreshold("threshold", "Number", "", "0.50", "", "", "","",false,false); parameters.push_back(pthreshold); + 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, "AlignCommand", "setParameters"); + exit(1); + } +} +//********************************************************************************************************************** +string AlignCommand::getHelpString(){ + try { + string helpString = ""; + helpString += "The align.seqs command reads a file containing sequences and creates an alignment file and a report file."; + helpString += "The align.seqs command parameters are reference, fasta, search, ksize, align, match, mismatch, gapopen, gapextend and processors."; + helpString += "The reference and fasta parameters are required. You may leave fasta blank if you have a valid fasta file. You may enter multiple fasta files by separating their names with dashes. ie. fasta=abrecovery.fasta-amzon.fasta."; + helpString += "The search parameter allows you to specify the method to find most similar template. Your options are: suffix, kmer and blast. The default is kmer."; + helpString += "The align parameter allows you to specify the alignment method to use. Your options are: gotoh, needleman, blast and noalign. The default is needleman."; + helpString += "The ksize parameter allows you to specify the kmer size for finding most similar template to candidate. The default is 8."; + helpString += "The match parameter allows you to specify the bonus for having the same base. The default is 1.0."; + helpString += "The mistmatch parameter allows you to specify the penalty for having different bases. The default is -1.0."; + helpString += "The gapopen parameter allows you to specify the penalty for opening a gap in an alignment. The default is -5.0."; + helpString += "The gapextend parameter allows you to specify the penalty for extending a gap in an alignment. The default is -2.0."; + helpString += "The flip parameter is used to specify whether or not you want mothur to try the reverse complement if a sequence falls below the threshold. The default is false."; + helpString += "The threshold is used to specify a cutoff at which an alignment is deemed 'bad' and the reverse complement may be tried. The default threshold is 0.50, meaning 50% of the bases are removed in the alignment."; + helpString += "If the flip parameter is set to true the reverse complement of the sequence is aligned and the better alignment is reported."; + helpString += "If the save parameter is set to true the reference sequences will be saved in memory, to clear them later you can use the clear.memory command. Default=f."; + helpString += "The default for the threshold parameter is 0.50, meaning at least 50% of the bases must remain or the sequence is reported as potentially reversed."; + helpString += "The align.seqs command should be in the following format:"; + helpString += "align.seqs(reference=yourTemplateFile, fasta=yourCandidateFile, align=yourAlignmentMethod, search=yourSearchmethod, ksize=yourKmerSize, match=yourMatchBonus, mismatch=yourMismatchpenalty, gapopen=yourGapopenPenalty, gapextend=yourGapExtendPenalty)"; + helpString += "Example align.seqs(candidate=candidate.fasta, template=core.filtered, align=kmer, search=gotoh, ksize=8, match=2.0, mismatch=3.0, gapopen=-2.0, gapextend=-1.0)"; + helpString += "Note: No spaces between parameter labels (i.e. candidate), '=' and parameters (i.e.yourFastaFile)."; + return helpString; + } + catch(exception& e) { + m->errorOut(e, "AlignCommand", "getHelpString"); + exit(1); + } +} +//********************************************************************************************************************** +string AlignCommand::getOutputPattern(string type) { + try { + string pattern = ""; + + if (type == "fasta") { pattern = "[filename],align"; } //makes file like: amazon.align + else if (type == "alignreport") { pattern = "[filename],align.report"; } + else if (type == "accnos") { pattern = "[filename],flip.accnos"; } + else { m->mothurOut("[ERROR]: No definition for type " + type + " output pattern.\n"); m->control_pressed = true; } + + return pattern; + } + catch(exception& e) { + m->errorOut(e, "AlignCommand", "getOutputPattern"); + exit(1); + } +} +//********************************************************************************************************************** +AlignCommand::AlignCommand(){ + try { + abort = true; calledHelp = true; + setParameters(); + vector tempOutNames; + outputTypes["fasta"] = tempOutNames; + outputTypes["alignreport"] = tempOutNames; + outputTypes["accnos"] = tempOutNames; + } + catch(exception& e) { + m->errorOut(e, "AlignCommand", "AlignCommand"); + exit(1); + } +} +//********************************************************************************************************************** AlignCommand::AlignCommand(string option) { try { + abort = false; calledHelp = false; + ReferenceDB* rdb = ReferenceDB::getInstance(); - abort = false; - //allow user to run help - if(option == "help") { help(); abort = true; } + if(option == "help") { help(); abort = true; calledHelp = true;} + else if(option == "citation") { citation(); abort = true; calledHelp = true;} else { - - //valid paramters for this command - string AlignArray[] = {"template","candidate","search","ksize","align","match","mismatch","gapopen","gapextend", "processors","flip","threshold","outputdir","inputdir"}; - vector myArray (AlignArray, AlignArray+(sizeof(AlignArray)/sizeof(string))); + vector myArray = setParameters(); OptionParser parser(option); map parameters = parser.getParameters(); @@ -52,7 +130,13 @@ AlignCommand::AlignCommand(string option) { for (it = parameters.begin(); it != parameters.end(); it++) { if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; } } - + + //initialize outputTypes + vector tempOutNames; + outputTypes["fasta"] = tempOutNames; + outputTypes["alignreport"] = tempOutNames; + outputTypes["accnos"] = 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 = ""; } @@ -64,61 +148,89 @@ AlignCommand::AlignCommand(string option) { 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; } } } - //check for required parameters - templateFileName = validParameter.validFile(parameters, "template", true); - - if (templateFileName == "not found") { - m->mothurOut("template is a required parameter for the align.seqs command."); - m->mothurOutEndLine(); - abort = true; - }else if (templateFileName == "not open") { abort = true; } - - candidateFileName = validParameter.validFile(parameters, "candidate", false); - if (candidateFileName == "not found") { m->mothurOut("candidate is a required parameter for the align.seqs command."); m->mothurOutEndLine(); abort = true; } - else { - splitAtDash(candidateFileName, candidateFileNames); + candidateFileName = validParameter.validFile(parameters, "fasta", false); + if (candidateFileName == "not found") { + //if there is a current fasta file, use it + string filename = m->getFastaFile(); + if (filename != "") { candidateFileNames.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 candidate parameter is required."); m->mothurOutEndLine(); abort = true; } + }else { + m->splitAtDash(candidateFileName, candidateFileNames); //go through files and make sure they are good, if not, then disregard them for (int i = 0; i < candidateFileNames.size(); i++) { - if (inputDir != "") { - string path = hasPath(candidateFileNames[i]); - //if the user has not given a path then, add inputdir. else leave path alone. - if (path == "") { candidateFileNames[i] = inputDir + candidateFileNames[i]; } - } - - int ableToOpen; - ifstream in; - - ableToOpen = openInputFile(candidateFileNames[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() + getSimpleName(candidateFileNames[i]); - m->mothurOut("Unable to open " + candidateFileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine(); - ableToOpen = openInputFile(tryPath, in, "noerror"); - candidateFileNames[i] = tryPath; + //candidateFileNames[i] = m->getFullPathName(candidateFileNames[i]); + + bool ignore = false; + if (candidateFileNames[i] == "current") { + candidateFileNames[i] = m->getFastaFile(); + if (candidateFileNames[i] != "") { m->mothurOut("Using " + candidateFileNames[i] + " as input file for the fasta parameter where you had given current."); m->mothurOutEndLine(); } + else { + m->mothurOut("You have no current fastafile, ignoring current."); m->mothurOutEndLine(); ignore=true; + //erase from file list + candidateFileNames.erase(candidateFileNames.begin()+i); + i--; } } - in.close(); + + if (!ignore) { + + if (inputDir != "") { + string path = m->hasPath(candidateFileNames[i]); + //if the user has not given a path then, add inputdir. else leave path alone. + if (path == "") { candidateFileNames[i] = inputDir + candidateFileNames[i]; } + } + + int ableToOpen; + ifstream in; + ableToOpen = m->openInputFile(candidateFileNames[i], in, "noerror"); + in.close(); + + //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(candidateFileNames[i]); + m->mothurOut("Unable to open " + candidateFileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine(); + ifstream in2; + ableToOpen = m->openInputFile(tryPath, in2, "noerror"); + in2.close(); + candidateFileNames[i] = tryPath; + } + } + + //if you can't open it, try output location + if (ableToOpen == 1) { + if (m->getOutputDir() != "") { //default path is set + string tryPath = m->getOutputDir() + m->getSimpleName(candidateFileNames[i]); + m->mothurOut("Unable to open " + candidateFileNames[i] + ". Trying output directory " + tryPath); m->mothurOutEndLine(); + ifstream in2; + ableToOpen = m->openInputFile(tryPath, in2, "noerror"); + in2.close(); + candidateFileNames[i] = tryPath; + } + } + + - if (ableToOpen == 1) { - m->mothurOut("Unable to open " + candidateFileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); - //erase from file list - candidateFileNames.erase(candidateFileNames.begin()+i); - i--; + if (ableToOpen == 1) { + m->mothurOut("Unable to open " + candidateFileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); + //erase from file list + candidateFileNames.erase(candidateFileNames.begin()+i); + i--; + }else { + m->setFastaFile(candidateFileNames[i]); + } } - } //make sure there is at least one valid file left @@ -129,32 +241,57 @@ AlignCommand::AlignCommand(string option) { // ...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); + m->mothurConvert(temp, kmerSize); temp = validParameter.validFile(parameters, "match", false); if (temp == "not found"){ temp = "1.0"; } - convert(temp, match); + m->mothurConvert(temp, match); temp = validParameter.validFile(parameters, "mismatch", false); if (temp == "not found"){ temp = "-1.0"; } - convert(temp, misMatch); + m->mothurConvert(temp, misMatch); - temp = validParameter.validFile(parameters, "gapopen", false); if (temp == "not found"){ temp = "-2.0"; } - convert(temp, gapOpen); + temp = validParameter.validFile(parameters, "gapopen", false); if (temp == "not found"){ temp = "-5.0"; } + m->mothurConvert(temp, gapOpen); - temp = validParameter.validFile(parameters, "gapextend", false); if (temp == "not found"){ temp = "-1.0"; } - convert(temp, gapExtend); + temp = validParameter.validFile(parameters, "gapextend", false); if (temp == "not found"){ temp = "-2.0"; } + m->mothurConvert(temp, gapExtend); - temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = "1"; } - convert(temp, processors); + temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); } + m->setProcessors(temp); + m->mothurConvert(temp, processors); temp = validParameter.validFile(parameters, "flip", false); if (temp == "not found"){ temp = "f"; } - flip = isTrue(temp); + flip = m->isTrue(temp); + + temp = validParameter.validFile(parameters, "save", false); if (temp == "not found"){ temp = "f"; } + save = m->isTrue(temp); + rdb->save = save; + if (save) { //clear out old references + rdb->clearMemory(); + } + + //this has to go after save so that if the user sets save=t and provides no reference we abort + templateFileName = validParameter.validFile(parameters, "reference", true); + if (templateFileName == "not found") { + //check for saved reference sequences + if (rdb->referenceSeqs.size() != 0) { + templateFileName = "saved"; + }else { + m->mothurOut("[ERROR]: You don't have any saved reference sequences and the reference parameter is a required for the align.seqs command."); + m->mothurOutEndLine(); + abort = true; + } + }else if (templateFileName == "not open") { abort = true; } + else { if (save) { rdb->setSavedReference(templateFileName); } } temp = validParameter.validFile(parameters, "threshold", false); if (temp == "not found"){ temp = "0.50"; } - convert(temp, threshold); + m->mothurConvert(temp, threshold); search = validParameter.validFile(parameters, "search", false); if (search == "not found"){ search = "kmer"; } + if ((search != "suffix") && (search != "kmer") && (search != "blast")) { m->mothurOut("invalid search option: choices are kmer, suffix or blast."); m->mothurOutEndLine(); abort=true; } align = validParameter.validFile(parameters, "align", false); if (align == "not found"){ align = "needleman"; } + if ((align != "needleman") && (align != "gotoh") && (align != "blast") && (align != "noalign")) { m->mothurOut("invalid align option: choices are needleman, gotoh, blast or noalign."); m->mothurOutEndLine(); abort=true; } + } } @@ -163,77 +300,33 @@ AlignCommand::AlignCommand(string option) { exit(1); } } - //********************************************************************************************************************** - AlignCommand::~AlignCommand(){ if (abort == false) { for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear(); delete templateDB; - delete alignment; - } -} - -//********************************************************************************************************************** - -void AlignCommand::help(){ - try { - m->mothurOut("The align.seqs command reads a file containing sequences and creates an alignment file and a report file.\n"); - m->mothurOut("The align.seqs command parameters are template, candidate, search, ksize, align, match, mismatch, gapopen and gapextend.\n"); - m->mothurOut("The template and candidate 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 and blast. The default is kmer.\n"); - m->mothurOut("The align parameter allows you to specify the alignment method to use. Your options are: gotoh, needleman, blast and noalign. The default is needleman.\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 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 flip parameter is used to specify whether or not you want mothur to try the reverse complement if a sequence falls below the threshold. The default is false.\n"); - m->mothurOut("The threshold is used to specify a cutoff at which an alignment is deemed 'bad' and the reverse complement may be tried. The default threshold is 0.50, meaning 50% of the bases are removed in the alignment.\n"); - m->mothurOut("If the flip parameter is set to true the reverse complement of the sequence is aligned and the better alignment is reported.\n"); - m->mothurOut("The default for the threshold parameter is 0.50, meaning at least 50% of the bases must remain or the sequence is reported as potentially reversed.\n"); - m->mothurOut("The align.seqs command should be in the following format: \n"); - m->mothurOut("align.seqs(template=yourTemplateFile, candidate=yourCandidateFile, align=yourAlignmentMethod, search=yourSearchmethod, ksize=yourKmerSize, match=yourMatchBonus, mismatch=yourMismatchpenalty, gapopen=yourGapopenPenalty, gapextend=yourGapExtendPenalty) \n"); - m->mothurOut("Example align.seqs(candidate=candidate.fasta, template=core.filtered, align=kmer, search=gotoh, ksize=8, match=2.0, mismatch=3.0, gapopen=-2.0, gapextend=-1.0)\n"); - m->mothurOut("Note: No spaces between parameter labels (i.e. candidate), '=' and parameters (i.e.yourFastaFile).\n\n"); - } - catch(exception& e) { - m->errorOut(e, "AlignCommand", "help"); - exit(1); } } - - //********************************************************************************************************************** int AlignCommand::execute(){ try { - if (abort == true) { return 0; } + if (abort == true) { if (calledHelp) { return 0; } return 2; } - templateDB = new AlignmentDB(templateFileName, search, kmerSize, gapOpen, gapExtend, match, misMatch); - int longestBase = templateDB->getLongestBase(); - - if(align == "gotoh") { alignment = new GotohOverlap(gapOpen, gapExtend, match, misMatch, longestBase); } - else if(align == "needleman") { alignment = new NeedlemanOverlap(gapOpen, match, misMatch, longestBase); } - else if(align == "blast") { alignment = new BlastAlignment(gapOpen, gapExtend, match, misMatch); } - else if(align == "noalign") { alignment = new NoAlign(); } - else { - m->mothurOut(align + " is not a valid alignment option. I will run the command using needleman."); - m->mothurOutEndLine(); - alignment = new NeedlemanOverlap(gapOpen, match, misMatch, longestBase); - } - vector outputNames; + templateDB = new AlignmentDB(templateFileName, search, kmerSize, gapOpen, gapExtend, match, misMatch, rand()); for (int s = 0; s < candidateFileNames.size(); s++) { - if (m->control_pressed) { return 0; } + if (m->control_pressed) { outputTypes.clear(); return 0; } m->mothurOut("Aligning sequences from " + candidateFileNames[s] + " ..." ); m->mothurOutEndLine(); - if (outputDir == "") { outputDir += hasPath(candidateFileNames[s]); } - string alignFileName = outputDir + getRootName(getSimpleName(candidateFileNames[s])) + "align"; - string reportFileName = outputDir + getRootName(getSimpleName(candidateFileNames[s])) + "align.report"; - string accnosFileName = outputDir + getRootName(getSimpleName(candidateFileNames[s])) + "flip.accnos"; + if (outputDir == "") { outputDir += m->hasPath(candidateFileNames[s]); } + map variables; variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(candidateFileNames[s])); + string alignFileName = getOutputFileName("fasta", variables); + string reportFileName = getOutputFileName("alignreport", variables); + string accnosFileName = getOutputFileName("accnos", variables); + bool hasAccnos = true; int numFastaSeqs = 0; @@ -241,9 +334,9 @@ int AlignCommand::execute(){ int start = time(NULL); #ifdef USE_MPI - int pid, end, numSeqsPerProcessor; + int pid, numSeqsPerProcessor; int tag = 2001; - vector MPIPos; + vector MPIPos; MPIWroteAccnos = false; MPI_Status status; @@ -275,11 +368,11 @@ int AlignCommand::execute(){ MPI_File_open(MPI_COMM_WORLD, outReportFilename, outMode, MPI_INFO_NULL, &outMPIReport); MPI_File_open(MPI_COMM_WORLD, outAccnosFilename, outMode, MPI_INFO_NULL, &outMPIAccnos); - if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPIAlign); MPI_File_close(&outMPIReport); MPI_File_close(&outMPIAccnos); return 0; } + if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPIAlign); MPI_File_close(&outMPIReport); MPI_File_close(&outMPIAccnos); outputTypes.clear(); return 0; } if (pid == 0) { //you are the root process - MPIPos = setFilePosFasta(candidateFileNames[s], numFastaSeqs); //fills MPIPos, returns numSeqs + MPIPos = m->setFilePosFasta(candidateFileNames[s], numFastaSeqs); //fills MPIPos, returns numSeqs //send file positions to all processes for(int i = 1; i < processors; i++) { @@ -295,7 +388,7 @@ int AlignCommand::execute(){ //align your part driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPIAlign, outMPIReport, outMPIAccnos, MPIPos); - if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPIAlign); MPI_File_close(&outMPIReport); MPI_File_close(&outMPIAccnos); return 0; } + if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPIAlign); MPI_File_close(&outMPIReport); MPI_File_close(&outMPIAccnos); outputTypes.clear(); return 0; } for (int i = 1; i < processors; i++) { bool tempResult; @@ -317,7 +410,7 @@ int AlignCommand::execute(){ //align your part driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPIAlign, outMPIReport, outMPIAccnos, MPIPos); - if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPIAlign); MPI_File_close(&outMPIReport); MPI_File_close(&outMPIAccnos); return 0; } + if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPIAlign); MPI_File_close(&outMPIReport); MPI_File_close(&outMPIAccnos); outputTypes.clear(); return 0; } MPI_Send(&MPIWroteAccnos, 1, MPI_INT, 0, tag, MPI_COMM_WORLD); } @@ -341,80 +434,43 @@ int AlignCommand::execute(){ //MPI_Info info; //MPI_File_delete(outAccnosFilename, info); hasAccnos = false; - remove(accnosFileName.c_str()); + m->mothurRemove(accnosFileName); } } #else - vector positions = divideFile(candidateFileNames[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) + + vector positions; + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + positions = m->divideFile(candidateFileNames[s], processors); + for (int i = 0; i < (positions.size()-1); i++) { lines.push_back(new linePair(positions[i], positions[(i+1)])); } + #else + if (processors == 1) { + lines.push_back(new linePair(0, 1000)); + }else { + positions = m->setFilePosFasta(candidateFileNames[s], numFastaSeqs); + if (positions.size() < processors) { processors = positions.size(); } + + //figure out how many sequences you have to process + int numSeqsPerProcessor = numFastaSeqs / processors; + for (int i = 0; i < processors; i++) { + int startIndex = i * numSeqsPerProcessor; + if(i == (processors - 1)){ numSeqsPerProcessor = numFastaSeqs - i * numSeqsPerProcessor; } + lines.push_back(new linePair(positions[startIndex], numSeqsPerProcessor)); + } + } + #endif + if(processors == 1){ numFastaSeqs = driver(lines[0], alignFileName, reportFileName, accnosFileName, candidateFileNames[s]); - - if (m->control_pressed) { remove(accnosFileName.c_str()); remove(alignFileName.c_str()); remove(reportFileName.c_str()); return 0; } - - //delete accnos file if its blank else report to user - if (isBlank(accnosFileName)) { remove(accnosFileName.c_str()); hasAccnos = false; } - else { - m->mothurOut("Some of you sequences generated alignments that eliminated too many bases, a list is provided in " + accnosFileName + "."); - if (!flip) { - m->mothurOut(" If you set the flip parameter to true mothur will try aligning the reverse compliment as well."); - }else{ m->mothurOut(" If the reverse compliment proved to be better it was reported."); } - m->mothurOutEndLine(); - } }else{ - processIDS.resize(0); - numFastaSeqs = createProcesses(alignFileName, reportFileName, accnosFileName, candidateFileNames[s]); - - rename((alignFileName + toString(processIDS[0]) + ".temp").c_str(), alignFileName.c_str()); - rename((reportFileName + toString(processIDS[0]) + ".temp").c_str(), reportFileName.c_str()); - - //append alignment and report files - for(int i=1;i nonBlankAccnosFiles; - //delete blank accnos files generated with multiple processes - for(int i=0;imothurOut("Some of you sequences generated alignments that eliminated too many bases, a list is provided in " + accnosFileName + "."); - if (!flip) { - m->mothurOut(" If you set the flip parameter to true mothur will try aligning the reverse compliment as well."); - }else{ m->mothurOut(" If the reverse compliment proved to be better it was reported."); } - m->mothurOutEndLine(); - }else{ hasAccnos = false; } - - if (m->control_pressed) { remove(accnosFileName.c_str()); remove(alignFileName.c_str()); remove(reportFileName.c_str()); return 0; } } - #else - numFastaSeqs = driver(lines[0], alignFileName, reportFileName, accnosFileName, candidateFileNames[s]); - - if (m->control_pressed) { remove(accnosFileName.c_str()); remove(alignFileName.c_str()); remove(reportFileName.c_str()); return 0; } + + if (m->control_pressed) { m->mothurRemove(accnosFileName); m->mothurRemove(alignFileName); m->mothurRemove(reportFileName); outputTypes.clear(); return 0; } //delete accnos file if its blank else report to user - if (isBlank(accnosFileName)) { remove(accnosFileName.c_str()); hasAccnos = false; } + if (m->isBlank(accnosFileName)) { m->mothurRemove(accnosFileName); hasAccnos = false; } else { m->mothurOut("Some of you sequences generated alignments that eliminated too many bases, a list is provided in " + accnosFileName + "."); if (!flip) { @@ -422,8 +478,6 @@ int AlignCommand::execute(){ }else{ m->mothurOut(" If the reverse compliment proved to be better it was reported."); } m->mothurOutEndLine(); } - - #endif #endif @@ -434,9 +488,9 @@ int AlignCommand::execute(){ if (pid == 0) { //only one process should output to screen #endif - outputNames.push_back(alignFileName); - outputNames.push_back(reportFileName); - if (hasAccnos) { outputNames.push_back(accnosFileName); } + outputNames.push_back(alignFileName); outputTypes["fasta"].push_back(alignFileName); + outputNames.push_back(reportFileName); outputTypes["alignreport"].push_back(reportFileName); + if (hasAccnos) { outputNames.push_back(accnosFileName); outputTypes["accnos"].push_back(accnosFileName); } #ifdef USE_MPI } @@ -447,6 +501,12 @@ int AlignCommand::execute(){ m->mothurOutEndLine(); } + //set align file as new current fastafile + string currentFasta = ""; + itTypes = outputTypes.find("fasta"); + if (itTypes != outputTypes.end()) { + if ((itTypes->second).size() != 0) { currentFasta = (itTypes->second)[0]; m->setFastaFile(currentFasta); } + } m->mothurOutEndLine(); m->mothurOut("Output File Names: "); m->mothurOutEndLine(); @@ -462,31 +522,45 @@ int AlignCommand::execute(){ } //********************************************************************************************************************** - int AlignCommand::driver(linePair* filePos, string alignFName, string reportFName, string accnosFName, string filename){ try { ofstream alignmentFile; - openOutputFile(alignFName, alignmentFile); + m->openOutputFile(alignFName, alignmentFile); ofstream accnosFile; - openOutputFile(accnosFName, accnosFile); + m->openOutputFile(accnosFName, accnosFile); NastReport report(reportFName); ifstream inFASTA; - openInputFile(filename, inFASTA); + m->openInputFile(filename, inFASTA); inFASTA.seekg(filePos->start); bool done = false; int count = 0; + + //moved this into driver to avoid deep copies in windows paralellized version + Alignment* alignment; + int longestBase = templateDB->getLongestBase(); + if(align == "gotoh") { alignment = new GotohOverlap(gapOpen, gapExtend, match, misMatch, longestBase); } + else if(align == "needleman") { alignment = new NeedlemanOverlap(gapOpen, match, misMatch, longestBase); } + else if(align == "blast") { alignment = new BlastAlignment(gapOpen, gapExtend, match, misMatch); } + else if(align == "noalign") { alignment = new NoAlign(); } + else { + m->mothurOut(align + " is not a valid alignment option. I will run the command using needleman."); + m->mothurOutEndLine(); + alignment = new NeedlemanOverlap(gapOpen, match, misMatch, longestBase); + } while (!done) { - if (m->control_pressed) { return 0; } + if (m->control_pressed) { break; } - Sequence* candidateSeq = new Sequence(inFASTA); gobble(inFASTA); - + Sequence* candidateSeq = new Sequence(inFASTA); m->gobble(inFASTA); + cout << candidateSeq->getAligned() << endl; + report.setCandidate(candidateSeq); + int origNumBases = candidateSeq->getNumBases(); string originalUnaligned = candidateSeq->getUnaligned(); int numBasesNeeded = origNumBases * threshold; @@ -495,13 +569,13 @@ int AlignCommand::driver(linePair* filePos, string alignFName, string reportFNam if (candidateSeq->getUnaligned().length() > alignment->getnRows()) { alignment->resize(candidateSeq->getUnaligned().length()+1); } - Sequence temp = templateDB->findClosestSequence(candidateSeq); Sequence* templateSeq = &temp; float searchScore = templateDB->getSearchScore(); Nast* nast = new Nast(alignment, candidateSeq, templateSeq); + Sequence* copy; Nast* nast2; @@ -516,6 +590,7 @@ int AlignCommand::driver(linePair* filePos, string alignFName, string reportFNam string wasBetter = ""; //if the user wants you to try the reverse if (flip) { + //get reverse compliment copy = new Sequence(candidateSeq->getName(), originalUnaligned); copy->reverseComplement(); @@ -547,7 +622,6 @@ int AlignCommand::driver(linePair* filePos, string alignFName, string reportFNam accnosFile << candidateSeq->getName() << wasBetter << endl; } - report.setCandidate(candidateSeq); report.setTemplate(templateSeq); report.setSearchParameters(search, searchScore); report.setAlignmentParameters(align, alignment); @@ -563,16 +637,21 @@ int AlignCommand::driver(linePair* filePos, string alignFName, string reportFNam } delete candidateSeq; - unsigned long int pos = inFASTA.tellg(); - if ((pos == -1) || (pos >= filePos->end)) { break; } + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + unsigned long long pos = inFASTA.tellg(); + if ((pos == -1) || (pos >= filePos->end)) { break; } + #else + if (inFASTA.eof()) { break; } + #endif //report progress - if((count) % 100 == 0){ m->mothurOut(toString(count)); m->mothurOutEndLine(); } + if((count) % 100 == 0){ m->mothurOutJustToScreen(toString(count) + "\n"); } } //report progress - if((count) % 100 != 0){ m->mothurOut(toString(count)); m->mothurOutEndLine(); } + if((count) % 100 != 0){ m->mothurOutJustToScreen(toString(count) + "\n"); } + delete alignment; alignmentFile.close(); inFASTA.close(); accnosFile.close(); @@ -586,7 +665,7 @@ int AlignCommand::driver(linePair* filePos, string alignFName, string reportFNam } //********************************************************************************************************************** #ifdef USE_MPI -int AlignCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& alignFile, MPI_File& reportFile, MPI_File& accnosFile, vector& MPIPos){ +int AlignCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& alignFile, MPI_File& reportFile, MPI_File& accnosFile, vector& MPIPos){ try { string outputString = ""; MPI_Status statusReport; @@ -610,9 +689,22 @@ int AlignCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& align delete buf; } + Alignment* alignment; + int longestBase = templateDB->getLongestBase(); + if(align == "gotoh") { alignment = new GotohOverlap(gapOpen, gapExtend, match, misMatch, longestBase); } + else if(align == "needleman") { alignment = new NeedlemanOverlap(gapOpen, match, misMatch, longestBase); } + else if(align == "blast") { alignment = new BlastAlignment(gapOpen, gapExtend, match, misMatch); } + else if(align == "noalign") { alignment = new NoAlign(); } + else { + m->mothurOut(align + " is not a valid alignment option. I will run the command using needleman."); + m->mothurOutEndLine(); + alignment = new NeedlemanOverlap(gapOpen, match, misMatch, longestBase); + } + + for(int i=0;icontrol_pressed) { return 0; } + if (m->control_pressed) { delete alignment; return 0; } //read next sequence int length = MPIPos[start+i+1] - MPIPos[start+i]; @@ -631,7 +723,8 @@ int AlignCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& align istringstream iss (tempBuf,istringstream::in); Sequence* candidateSeq = new Sequence(iss); - + report.setCandidate(candidateSeq); + int origNumBases = candidateSeq->getNumBases(); string originalUnaligned = candidateSeq->getUnaligned(); int numBasesNeeded = origNumBases * threshold; @@ -680,6 +773,7 @@ int AlignCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& align delete nast; nast = nast2; needToDeleteCopy = true; + wasBetter = "\treverse complement produced a better alignment, so mothur used the reverse complement."; }else{ wasBetter = "\treverse complement did NOT produce a better alignment, please check sequence."; delete nast2; @@ -701,7 +795,6 @@ int AlignCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& align MPIWroteAccnos = true; } - report.setCandidate(candidateSeq); report.setTemplate(templateSeq); report.setSearchParameters(search, searchScore); report.setAlignmentParameters(align, alignment); @@ -751,10 +844,10 @@ int AlignCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& align int AlignCommand::createProcesses(string alignFileName, string reportFileName, string accnosFName, string filename) { try { -#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - int process = 0; int num = 0; - // processIDS.resize(0); + processIDS.resize(0); +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + int process = 1; //loop through and create all the processes you want while (process != processors) { @@ -769,56 +862,153 @@ int AlignCommand::createProcesses(string alignFileName, string reportFileName, s //pass numSeqs to parent ofstream out; string tempFile = alignFileName + toString(getpid()) + ".num.temp"; - openOutputFile(tempFile, out); + 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); + } } + //do my part + num = driver(lines[0], alignFileName, reportFileName, accnosFName, filename); + //force parent to wait until all the processes are done - for (int i=0;i nonBlankAccnosFiles; + if (!(m->isBlank(accnosFName))) { nonBlankAccnosFiles.push_back(accnosFName); } + else { m->mothurRemove(accnosFName); } //remove so other files can be renamed to it + for (int i = 0; i < processIDS.size(); i++) { ifstream in; string tempFile = alignFileName + toString(processIDS[i]) + ".num.temp"; - openInputFile(tempFile, in); + m->openInputFile(tempFile, in); if (!in.eof()) { int tempNum = 0; in >> tempNum; num += tempNum; } - in.close(); remove(tempFile.c_str()); + in.close(); m->mothurRemove(tempFile); + + m->appendFiles((alignFileName + toString(processIDS[i]) + ".temp"), alignFileName); + m->mothurRemove((alignFileName + toString(processIDS[i]) + ".temp")); + + appendReportFiles((reportFileName + toString(processIDS[i]) + ".temp"), reportFileName); + m->mothurRemove((reportFileName + toString(processIDS[i]) + ".temp")); + + if (!(m->isBlank(accnosFName + toString(processIDS[i]) + ".temp"))) { + nonBlankAccnosFiles.push_back(accnosFName + toString(processIDS[i]) + ".temp"); + }else { m->mothurRemove((accnosFName + toString(processIDS[i]) + ".temp")); } + } - return num; -#endif - } - catch(exception& e) { - m->errorOut(e, "AlignCommand", "createProcesses"); - exit(1); - } -} -/**************************************************************************************************/ - -void AlignCommand::appendAlignFiles(string temp, string filename) { - try{ + //append accnos files + if (nonBlankAccnosFiles.size() != 0) { + rename(nonBlankAccnosFiles[0].c_str(), accnosFName.c_str()); + + for (int h=1; h < nonBlankAccnosFiles.size(); h++) { + m->appendFiles(nonBlankAccnosFiles[h], accnosFName); + m->mothurRemove(nonBlankAccnosFiles[h]); + } + }else { //recreate the accnosfile if needed + ofstream out; + m->openOutputFile(accnosFName, out); + out.close(); + } +#else + ////////////////////////////////////////////////////////////////////////////////////////////////////// + //Windows version shared memory, so be careful when passing variables through the alignData struct. + //Above fork() will clone, so memory is separate, but that's not the case with windows, + ////////////////////////////////////////////////////////////////////////////////////////////////////// - ofstream output; - ifstream input; - openOutputFileAppend(filename, output); - openInputFile(temp, input); + vector pDataArray; + DWORD dwThreadIdArray[processors-1]; + HANDLE hThreadArray[processors-1]; + + //Create processor worker threads. + for( int i=0; istart, lines[i]->end, flip, match, misMatch, gapOpen, gapExtend, threshold, i); + pDataArray.push_back(tempalign); + processIDS.push_back(i); + + //MySeqSumThreadFunction is in header. It must be global or static to work with the threads. + //default security attributes, thread function name, argument to thread function, use default creation flags, returns the thread identifier + hThreadArray[i] = CreateThread(NULL, 0, MyAlignThreadFunction, pDataArray[i], 0, &dwThreadIdArray[i]); + } + + //need to check for line ending error + ifstream inFASTA; + m->openInputFile(filename, inFASTA); + inFASTA.seekg(lines[processors-1]->start-1); + char c = inFASTA.peek(); - while(char c = input.get()){ - if(input.eof()) { break; } - else { output << c; } + if (c != '>') { //we need to move back + lines[processors-1]->start--; } - input.close(); - output.close(); + //using the main process as a worker saves time and memory + //do my part - do last piece because windows is looking for eof + num = driver(lines[processors-1], (alignFileName + toString(processors-1) + ".temp"), (reportFileName + toString(processors-1) + ".temp"), (accnosFName + toString(processors-1) + ".temp"), filename); + + //Wait until all threads have terminated. + WaitForMultipleObjects(processors-1, hThreadArray, TRUE, INFINITE); + + //Close all thread handles and free memory allocations. + for(int i=0; i < pDataArray.size(); i++){ + if (pDataArray[i]->count != pDataArray[i]->end) { + m->mothurOut("[ERROR]: process " + toString(i) + " only processed " + toString(pDataArray[i]->count) + " of " + toString(pDataArray[i]->end) + " sequences assigned to it, quitting. \n"); m->control_pressed = true; + } + num += pDataArray[i]->count; + CloseHandle(hThreadArray[i]); + delete pDataArray[i]; + } + + vector nonBlankAccnosFiles; + if (!(m->isBlank(accnosFName))) { nonBlankAccnosFiles.push_back(accnosFName); } + else { m->mothurRemove(accnosFName); } //remove so other files can be renamed to it + + for (int i = 1; i < processors; i++) { + m->appendFiles((alignFileName + toString(i) + ".temp"), alignFileName); + m->mothurRemove((alignFileName + toString(i) + ".temp")); + + appendReportFiles((reportFileName + toString(i) + ".temp"), reportFileName); + m->mothurRemove((reportFileName + toString(i) + ".temp")); + + if (!(m->isBlank(accnosFName + toString(i) + ".temp"))) { + nonBlankAccnosFiles.push_back(accnosFName + toString(i) + ".temp"); + }else { m->mothurRemove((accnosFName + toString(i) + ".temp")); } + } + + //append accnos files + if (nonBlankAccnosFiles.size() != 0) { + rename(nonBlankAccnosFiles[0].c_str(), accnosFName.c_str()); + + for (int h=1; h < nonBlankAccnosFiles.size(); h++) { + m->appendFiles(nonBlankAccnosFiles[h], accnosFName); + m->mothurRemove(nonBlankAccnosFiles[h]); + } + }else { //recreate the accnosfile if needed + ofstream out; + m->openOutputFile(accnosFName, out); + out.close(); + } +#endif + + return num; } catch(exception& e) { - m->errorOut(e, "AlignCommand", "appendAlignFiles"); + m->errorOut(e, "AlignCommand", "createProcesses"); exit(1); } } @@ -829,15 +1019,16 @@ void AlignCommand::appendReportFiles(string temp, string filename) { ofstream output; ifstream input; - openOutputFileAppend(filename, output); - openInputFile(temp, input); + m->openOutputFileAppend(filename, output); + m->openInputFile(temp, input); while (!input.eof()) { char c = input.get(); if (c == 10 || c == 13){ break; } } // get header line - while(char c = input.get()){ - if(input.eof()) { break; } - else { output << c; } - } + char buffer[4096]; + while (!input.eof()) { + input.read(buffer, 4096); + output.write(buffer, input.gcount()); + } input.close(); output.close();