X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=aligncommand.cpp;h=f757a7920bb337883a2a7773f35ed533d2186c80;hp=5da0ee8d0f091b8ce07acdedd48d2069e53ec1e9;hb=050a3ff02473a3d4c0980964e1a9ebe52e55d6b8;hpb=220dc345e493cddc569521111ce32ac4d965ab7f diff --git a/aligncommand.cpp b/aligncommand.cpp index 5da0ee8..f757a79 100644 --- a/aligncommand.cpp +++ b/aligncommand.cpp @@ -15,33 +15,26 @@ */ #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); parameters.push_back(ptemplate); - CommandParameter pcandidate("fasta", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(pcandidate); - CommandParameter psearch("search", "Multiple", "kmer-blast-suffix", "kmer", "", "", "",false,false); 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); parameters.push_back(palign); - 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 pprocessors("processors", "Number", "", "1", "", "", "",false,false); parameters.push_back(pprocessors); - CommandParameter pflip("flip", "Boolean", "", "F", "", "", "",false,false); parameters.push_back(pflip); - 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); + 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); } @@ -64,11 +57,12 @@ string AlignCommand::getHelpString(){ 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 -2.0."; - helpString += "The gapextend parameter allows you to specify the penalty for extending a gap in an alignment. 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)"; @@ -82,6 +76,23 @@ string AlignCommand::getHelpString(){ } } //********************************************************************************************************************** +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; @@ -99,8 +110,9 @@ AlignCommand::AlignCommand(){ //********************************************************************************************************************** AlignCommand::AlignCommand(string option) { try { - abort = false; calledHelp = false; - + abort = false; calledHelp = false; + ReferenceDB* rdb = ReferenceDB::getInstance(); + //allow user to run help if(option == "help") { help(); abort = true; calledHelp = true;} else if(option == "citation") { citation(); abort = true; calledHelp = true;} @@ -146,15 +158,6 @@ AlignCommand::AlignCommand(string option) { } } - //check for required parameters - templateFileName = validParameter.validFile(parameters, "reference", true); - - if (templateFileName == "not found") { - m->mothurOut("reference is a required parameter for the align.seqs command."); - m->mothurOutEndLine(); - abort = true; - }else if (templateFileName == "not open") { abort = true; } - candidateFileName = validParameter.validFile(parameters, "fasta", false); if (candidateFileName == "not found") { //if there is a current fasta file, use it @@ -238,29 +241,50 @@ 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 = m->getProcessors(); } m->setProcessors(temp); - convert(temp, processors); + m->mothurConvert(temp, processors); temp = validParameter.validFile(parameters, "flip", false); if (temp == "not found"){ temp = "f"; } 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; } @@ -282,7 +306,6 @@ AlignCommand::~AlignCommand(){ if (abort == false) { for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear(); delete templateDB; - delete alignment; } } //********************************************************************************************************************** @@ -291,18 +314,7 @@ int AlignCommand::execute(){ try { 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); - } + templateDB = new AlignmentDB(templateFileName, search, kmerSize, gapOpen, gapExtend, match, misMatch, rand()); for (int s = 0; s < candidateFileNames.size(); s++) { if (m->control_pressed) { outputTypes.clear(); return 0; } @@ -310,9 +322,11 @@ int AlignCommand::execute(){ m->mothurOut("Aligning sequences from " + candidateFileNames[s] + " ..." ); m->mothurOutEndLine(); if (outputDir == "") { outputDir += m->hasPath(candidateFileNames[s]); } - string alignFileName = outputDir + m->getRootName(m->getSimpleName(candidateFileNames[s])) + "align"; - string reportFileName = outputDir + m->getRootName(m->getSimpleName(candidateFileNames[s])) + "align.report"; - string accnosFileName = outputDir + m->getRootName(m->getSimpleName(candidateFileNames[s])) + "flip.accnos"; + 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; @@ -322,7 +336,7 @@ int AlignCommand::execute(){ #ifdef USE_MPI int pid, numSeqsPerProcessor; int tag = 2001; - vector MPIPos; + vector MPIPos; MPIWroteAccnos = false; MPI_Status status; @@ -420,30 +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 = m->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]); }else{ numFastaSeqs = createProcesses(alignFileName, reportFileName, accnosFileName, candidateFileNames[s]); } - #else - numFastaSeqs = driver(lines[0], alignFileName, reportFileName, accnosFileName, candidateFileNames[s]); - #endif - - if (m->control_pressed) { remove(accnosFileName.c_str()); remove(alignFileName.c_str()); remove(reportFileName.c_str()); outputTypes.clear(); 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 (m->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) { @@ -495,7 +522,6 @@ int AlignCommand::execute(){ } //********************************************************************************************************************** - int AlignCommand::driver(linePair* filePos, string alignFName, string reportFName, string accnosFName, string filename){ try { ofstream alignmentFile; @@ -513,12 +539,26 @@ int AlignCommand::driver(linePair* filePos, string alignFName, string reportFNam 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); m->gobble(inFASTA); + cout << candidateSeq->getAligned() << endl; report.setCandidate(candidateSeq); int origNumBases = candidateSeq->getNumBases(); @@ -529,10 +569,9 @@ 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); @@ -598,20 +637,21 @@ int AlignCommand::driver(linePair* filePos, string alignFName, string reportFNam } delete candidateSeq; - #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - unsigned long int pos = inFASTA.tellg(); + #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(); @@ -625,7 +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; @@ -649,9 +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]; @@ -791,11 +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 num = 0; processIDS.resize(0); +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) int process = 1; - int num = 0; - // processIDS.resize(0); //loop through and create all the processes you want while (process != processors) { @@ -833,24 +885,24 @@ int AlignCommand::createProcesses(string alignFileName, string reportFileName, s vector nonBlankAccnosFiles; if (!(m->isBlank(accnosFName))) { nonBlankAccnosFiles.push_back(accnosFName); } - else { remove(accnosFName.c_str()); } //remove so other files can be renamed to it + 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"; 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); - appendAlignFiles((alignFileName + toString(processIDS[i]) + ".temp"), alignFileName); - remove((alignFileName + toString(processIDS[i]) + ".temp").c_str()); + m->appendFiles((alignFileName + toString(processIDS[i]) + ".temp"), alignFileName); + m->mothurRemove((alignFileName + toString(processIDS[i]) + ".temp")); appendReportFiles((reportFileName + toString(processIDS[i]) + ".temp"), reportFileName); - remove((reportFileName + toString(processIDS[i]) + ".temp").c_str()); + m->mothurRemove((reportFileName + toString(processIDS[i]) + ".temp")); if (!(m->isBlank(accnosFName + toString(processIDS[i]) + ".temp"))) { nonBlankAccnosFiles.push_back(accnosFName + toString(processIDS[i]) + ".temp"); - }else { remove((accnosFName + toString(processIDS[i]) + ".temp").c_str()); } + }else { m->mothurRemove((accnosFName + toString(processIDS[i]) + ".temp")); } } @@ -859,43 +911,104 @@ int AlignCommand::createProcesses(string alignFileName, string reportFileName, s rename(nonBlankAccnosFiles[0].c_str(), accnosFName.c_str()); for (int h=1; h < nonBlankAccnosFiles.size(); h++) { - appendAlignFiles(nonBlankAccnosFiles[h], accnosFName); - remove(nonBlankAccnosFiles[h].c_str()); + 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, + ////////////////////////////////////////////////////////////////////////////////////////////////////// - return num; -#endif - } - catch(exception& e) { - m->errorOut(e, "AlignCommand", "createProcesses"); - exit(1); - } -} -/**************************************************************************************************/ - -void AlignCommand::appendAlignFiles(string temp, string filename) { - try{ + vector pDataArray; + DWORD dwThreadIdArray[processors-1]; + HANDLE hThreadArray[processors-1]; - ofstream output; - ifstream input; - m->openOutputFileAppend(filename, output); - m->openInputFile(temp, input); + //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); } } @@ -911,10 +1024,11 @@ void AlignCommand::appendReportFiles(string temp, string filename) { 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();