X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=aligncommand.cpp;h=f19810e131b687ea372ae1914ecf3a02d03ef8d8;hb=a3d9877d9c62058dbe167ceca159f4b76744f1cf;hp=708a85472dc53fecec1f022ce86512a82d620c30;hpb=5a1e62397b91f57d0d3aff635891df04b8999a88;p=mothur.git diff --git a/aligncommand.cpp b/aligncommand.cpp index 708a854..f19810e 100644 --- a/aligncommand.cpp +++ b/aligncommand.cpp @@ -28,11 +28,11 @@ //********************************************************************************************************************** -AlignCommand::AlignCommand(string option){ +AlignCommand::AlignCommand(string option) { try { - // globaldata = GlobalData::getInstance(); - abort = false; + abort = false; + //allow user to run help if(option == "help") { help(); abort = true; } @@ -45,7 +45,7 @@ AlignCommand::AlignCommand(string option){ OptionParser parser(option); map parameters = parser.getParameters(); - ValidParameters validParameter; + ValidParameters validParameter("align.seqs"); map::iterator it; //check to make sure all parameters are valid for command @@ -78,13 +78,13 @@ AlignCommand::AlignCommand(string option){ templateFileName = validParameter.validFile(parameters, "template", true); if (templateFileName == "not found") { - mothurOut("template is a required parameter for the align.seqs command."); - mothurOutEndLine(); + 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") { mothurOut("candidate is a required parameter for the align.seqs command."); mothurOutEndLine(); abort = true; } + if (candidateFileName == "not found") { m->mothurOut("candidate is a required parameter for the align.seqs command."); m->mothurOutEndLine(); abort = true; } else { splitAtDash(candidateFileName, candidateFileNames); @@ -95,23 +95,54 @@ AlignCommand::AlignCommand(string option){ //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); + + #ifdef USE_MPI + int pid; + MPI_Comm_size(MPI_COMM_WORLD, &processors); //set processors to the number of mpi processes running + MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are + + if (pid == 0) { + #endif + + 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; + } + } + in.close(); + #ifdef USE_MPI + for (int j = 1; j < processors; j++) { + MPI_Send(&ableToOpen, 1, MPI_INT, j, 2001, MPI_COMM_WORLD); + } + }else{ + MPI_Status status; + MPI_Recv(&ableToOpen, 1, MPI_INT, 0, 2001, MPI_COMM_WORLD, &status); + } + + #endif + if (ableToOpen == 1) { - mothurOut(candidateFileNames[i] + " will be disregarded."); mothurOutEndLine(); + m->mothurOut("Unable to open " + candidateFileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); //erase from file list candidateFileNames.erase(candidateFileNames.begin()+i); i--; } - in.close(); + } //make sure there is at least one valid file left - if (candidateFileNames.size() == 0) { mothurOut("no valid files."); mothurOutEndLine(); abort = true; } + if (candidateFileNames.size() == 0) { m->mothurOut("no valid files."); m->mothurOutEndLine(); abort = true; } } - + //check for optional parameter and set defaults // ...at some point should added some additional type checking... string temp; @@ -146,7 +177,7 @@ AlignCommand::AlignCommand(string option){ } catch(exception& e) { - errorOut(e, "AlignCommand", "AlignCommand"); + m->errorOut(e, "AlignCommand", "AlignCommand"); exit(1); } } @@ -166,27 +197,27 @@ AlignCommand::~AlignCommand(){ void AlignCommand::help(){ try { - mothurOut("The align.seqs command reads a file containing sequences and creates an alignment file and a report file.\n"); - mothurOut("The align.seqs command parameters are template, candidate, search, ksize, align, match, mismatch, gapopen and gapextend.\n"); - 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"); - 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"); - 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"); - mothurOut("The ksize parameter allows you to specify the kmer size for finding most similar template to candidate. The default is 8.\n"); - mothurOut("The match parameter allows you to specify the bonus for having the same base. The default is 1.0.\n"); - mothurOut("The mistmatch parameter allows you to specify the penalty for having different bases. The default is -1.0.\n"); - mothurOut("The gapopen parameter allows you to specify the penalty for opening a gap in an alignment. The default is -2.0.\n"); - mothurOut("The gapextend parameter allows you to specify the penalty for extending a gap in an alignment. The default is -1.0.\n"); - 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"); - 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"); - mothurOut("If the flip parameter is set to true the reverse complement of the sequence is aligned and the better alignment is reported.\n"); - 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"); - mothurOut("The align.seqs command should be in the following format: \n"); - mothurOut("align.seqs(template=yourTemplateFile, candidate=yourCandidateFile, align=yourAlignmentMethod, search=yourSearchmethod, ksize=yourKmerSize, match=yourMatchBonus, mismatch=yourMismatchpenalty, gapopen=yourGapopenPenalty, gapextend=yourGapExtendPenalty) \n"); - 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"); - mothurOut("Note: No spaces between parameter labels (i.e. candidate), '=' and parameters (i.e.yourFastaFile).\n\n"); + 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) { - errorOut(e, "AlignCommand", "help"); + m->errorOut(e, "AlignCommand", "help"); exit(1); } } @@ -197,22 +228,25 @@ void AlignCommand::help(){ int AlignCommand::execute(){ try { if (abort == true) { return 0; } - + 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 { - mothurOut(align + " is not a valid alignment option. I will run the command using needleman."); - mothurOutEndLine(); + 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; for (int s = 0; s < candidateFileNames.size(); s++) { - mothurOut("Aligning sequences from " + candidateFileNames[s] + " ..." ); mothurOutEndLine(); + if (m->control_pressed) { 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"; @@ -223,30 +257,145 @@ int AlignCommand::execute(){ int numFastaSeqs = 0; for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear(); int start = time(NULL); + +#ifdef USE_MPI + int pid, end, numSeqsPerProcessor; + int tag = 2001; + vector MPIPos; + MPIWroteAccnos = false; + + 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 outMPIAlign; + MPI_File outMPIReport; + MPI_File outMPIAccnos; + + int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY; + int inMode=MPI_MODE_RDONLY; + + char outAlignFilename[1024]; + strcpy(outAlignFilename, alignFileName.c_str()); + + char outReportFilename[1024]; + strcpy(outReportFilename, reportFileName.c_str()); + + char outAccnosFilename[1024]; + strcpy(outAccnosFilename, accnosFileName.c_str()); + + char inFileName[1024]; + strcpy(inFileName, candidateFileNames[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, outAlignFilename, outMode, MPI_INFO_NULL, &outMPIAlign); + 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 (pid == 0) { //you are the root process + + MPIPos = setFilePosFasta(candidateFileNames[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; } + + + //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; } + + for (int i = 1; i < processors; i++) { + bool tempResult; + MPI_Recv(&tempResult, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status); + if (tempResult != 0) { MPIWroteAccnos = true; } + } + }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, 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; } + + MPI_Send(&MPIWroteAccnos, 1, MPI_INT, 0, tag, MPI_COMM_WORLD); + } + + //close files + MPI_File_close(&inMPI); + MPI_File_close(&outMPIAlign); + MPI_File_close(&outMPIReport); + MPI_File_close(&outMPIAccnos); + + //delete accnos file if blank + if (pid == 0) { + //delete accnos file if its blank else report to user + if (MPIWroteAccnos) { + 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 { + //MPI_Info info; + //MPI_File_delete(outAccnosFilename, info); + hasAccnos = false; + remove(accnosFileName.c_str()); + } + } + +#else -#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) if(processors == 1){ ifstream inFASTA; openInputFile(candidateFileNames[s], inFASTA); - numFastaSeqs=count(istreambuf_iterator(inFASTA),istreambuf_iterator(), '>'); + getNumSeqs(inFASTA, numFastaSeqs); inFASTA.close(); lines.push_back(new linePair(0, 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()); } + if (isBlank(accnosFileName)) { remove(accnosFileName.c_str()); hasAccnos = false; } else { - mothurOut("Some of you sequences generated alignments that eliminated too many bases, a list is provided in " + accnosFileName + "."); + m->mothurOut("Some of you sequences generated alignments that eliminated too many bases, a list is provided in " + accnosFileName + "."); if (!flip) { - mothurOut(" If you set the flip parameter to true mothur will try aligning the reverse compliment as well."); - }else{ mothurOut(" If the reverse compliment proved to be better it was reported."); } - mothurOutEndLine(); + 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{ - vector positions; + vector positions; processIDS.resize(0); ifstream inFASTA; @@ -256,7 +405,7 @@ int AlignCommand::execute(){ while(!inFASTA.eof()){ input = getline(inFASTA); if (input.length() != 0) { - if(input[0] == '>'){ long int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); } + if(input[0] == '>'){ unsigned long int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); } } } inFASTA.close(); @@ -266,7 +415,7 @@ int AlignCommand::execute(){ int numSeqsPerProcessor = numFastaSeqs / processors; for (int i = 0; i < processors; i++) { - long int startPos = positions[ i * numSeqsPerProcessor ]; + unsigned long int startPos = positions[ i * numSeqsPerProcessor ]; if(i == processors - 1){ numSeqsPerProcessor = numFastaSeqs - i * numSeqsPerProcessor; } @@ -303,49 +452,81 @@ int AlignCommand::execute(){ appendAlignFiles(nonBlankAccnosFiles[h], accnosFileName); remove(nonBlankAccnosFiles[h].c_str()); } - mothurOut("Some of you sequences generated alignments that eliminated too many bases, a list is provided in " + accnosFileName + "."); + m->mothurOut("Some of you sequences generated alignments that eliminated too many bases, a list is provided in " + accnosFileName + "."); if (!flip) { - mothurOut(" If you set the flip parameter to true mothur will try aligning the reverse compliment as well."); - }else{ mothurOut(" If the reverse compliment proved to be better it was reported."); } - mothurOutEndLine(); + 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 + #else ifstream inFASTA; openInputFile(candidateFileNames[s], inFASTA); - numFastaSeqs=count(istreambuf_iterator(inFASTA),istreambuf_iterator(), '>'); + getNumSeqs(inFASTA, numFastaSeqs); inFASTA.close(); lines.push_back(new linePair(0, 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 { - mothurOut("Some of you sequences generated alignments that eliminated too many bases, a list is provided in " + accnosFileName + "."); + m->mothurOut("Some of you sequences generated alignments that eliminated too many bases, a list is provided in " + accnosFileName + "."); if (!flip) { - mothurOut(" If you set the flip parameter to true mothur will try aligning the reverse compliment as well."); - }else{ mothurOut(" If the reverse compliment proved to be better it was reported."); } - mothurOutEndLine(); + 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(); } -#endif - - mothurOut("Output File Names: " + alignFileName + ", " + reportFileName); - if (hasAccnos) { mothurOut(", " + accnosFileName + "."); } - else { mothurOut("."); } - mothurOutEndLine(); + #endif + +#endif + + + #ifdef USE_MPI + MPI_Comm_rank(MPI_COMM_WORLD, &pid); + + 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); } - mothurOut("It took " + toString(time(NULL) - start) + " secs to align " + toString(numFastaSeqs) + " sequences."); - mothurOutEndLine(); - mothurOutEndLine(); + #ifdef USE_MPI + } + #endif + + m->mothurOut("It took " + toString(time(NULL) - start) + " secs to align " + toString(numFastaSeqs) + " sequences."); + m->mothurOutEndLine(); + m->mothurOutEndLine(); } + + 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(); + return 0; } catch(exception& e) { - errorOut(e, "AlignCommand", "execute"); + m->errorOut(e, "AlignCommand", "execute"); exit(1); } } @@ -366,10 +547,13 @@ int AlignCommand::driver(linePair* line, string alignFName, string reportFName, openInputFile(filename, inFASTA); inFASTA.seekg(line->start); - + for(int i=0;inumSeqs;i++){ - + + if (m->control_pressed) { return 0; } + Sequence* candidateSeq = new Sequence(inFASTA); gobble(inFASTA); + int origNumBases = candidateSeq->getNumBases(); string originalUnaligned = candidateSeq->getUnaligned(); int numBasesNeeded = origNumBases * threshold; @@ -444,10 +628,10 @@ int AlignCommand::driver(linePair* line, string alignFName, string reportFName, delete candidateSeq; //report progress - if((i+1) % 100 == 0){ mothurOut(toString(i+1)); mothurOutEndLine(); } + if((i+1) % 100 == 0){ m->mothurOut(toString(i+1)); m->mothurOutEndLine(); } } //report progress - if((line->numSeqs) % 100 != 0){ mothurOut(toString(line->numSeqs)); mothurOutEndLine(); } + if((line->numSeqs) % 100 != 0){ m->mothurOut(toString(line->numSeqs)); m->mothurOutEndLine(); } alignmentFile.close(); inFASTA.close(); @@ -456,17 +640,178 @@ int AlignCommand::driver(linePair* line, string alignFName, string reportFName, return 1; } catch(exception& e) { - errorOut(e, "AlignCommand", "driver"); + m->errorOut(e, "AlignCommand", "driver"); exit(1); } } +//********************************************************************************************************************** +#ifdef USE_MPI +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; + MPI_Status statusAlign; + MPI_Status statusAccnos; + MPI_Status status; + int pid; + MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are + + NastReport report; + + if (pid == 0) { + outputString = report.getHeaders(); + int length = outputString.length(); + + char* buf = new char[length]; + memcpy(buf, outputString.c_str(), length); + + MPI_File_write_shared(reportFile, buf, length, MPI_CHAR, &statusReport); + + delete buf; + } + + 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]; + memcpy(buf4, outputString.c_str(), length); + + MPI_File_read_at(inMPI, MPIPos[start+i], buf4, length, MPI_CHAR, &status); + + string tempBuf = buf4; + delete buf4; + + if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); } + istringstream iss (tempBuf,istringstream::in); + + Sequence* candidateSeq = new Sequence(iss); + int origNumBases = candidateSeq->getNumBases(); + string originalUnaligned = candidateSeq->getUnaligned(); + int numBasesNeeded = origNumBases * threshold; + + if (candidateSeq->getName() != "") { //incase there is a commented sequence at the end of a file + 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; + bool needToDeleteCopy = false; //this is needed in case you have you enter the ifs below + //since nast does not make a copy of hte sequence passed, and it is used by the reporter below + //you can't delete the copy sequence til after you report, but you may choose not to create it in the first place + //so this bool tells you if you need to delete it + + //if there is a possibility that this sequence should be reversed + if (candidateSeq->getNumBases() < numBasesNeeded) { + + string wasBetter = ""; + //if the user wants you to try the reverse + if (flip) { + //get reverse compliment + copy = new Sequence(candidateSeq->getName(), originalUnaligned); + copy->reverseComplement(); + + //rerun alignment + Sequence temp2 = templateDB->findClosestSequence(copy); + Sequence* templateSeq2 = &temp2; + + searchScore = templateDB->getSearchScore(); + + nast2 = new Nast(alignment, copy, templateSeq2); + + //check if any better + if (copy->getNumBases() > candidateSeq->getNumBases()) { + candidateSeq->setAligned(copy->getAligned()); //use reverse compliments alignment since its better + templateSeq = templateSeq2; + delete nast; + nast = nast2; + needToDeleteCopy = true; + }else{ + wasBetter = "\treverse complement did NOT produce a better alignment, please check sequence."; + delete nast2; + delete copy; + } + } + + //create accnos file with names + outputString = candidateSeq->getName() + wasBetter + "\n"; + + //send results to parent + int length = outputString.length(); + + char* buf = new char[length]; + memcpy(buf, outputString.c_str(), length); + + MPI_File_write_shared(accnosFile, buf, length, MPI_CHAR, &statusAccnos); + delete buf; + MPIWroteAccnos = true; + } + + report.setCandidate(candidateSeq); + report.setTemplate(templateSeq); + report.setSearchParameters(search, searchScore); + report.setAlignmentParameters(align, alignment); + report.setNastParameters(*nast); + + outputString = ">" + candidateSeq->getName() + "\n" + candidateSeq->getAligned() + "\n"; + + //send results to parent + int length = outputString.length(); + char* buf2 = new char[length]; + memcpy(buf2, outputString.c_str(), length); + + MPI_File_write_shared(alignFile, buf2, length, MPI_CHAR, &statusAlign); + + delete buf2; + + outputString = report.getReport(); + + //send results to parent + length = outputString.length(); + char* buf3 = new char[length]; + memcpy(buf3, outputString.c_str(), length); + + MPI_File_write_shared(reportFile, buf3, length, MPI_CHAR, &statusReport); + + delete buf3; + delete nast; + if (needToDeleteCopy) { delete copy; } + } + delete candidateSeq; + + //report progress + if((i+1) % 100 == 0){ cout << (toString(i+1)) << endl; } + } + //report progress + if((num) % 100 != 0){ cout << (toString(num)) << endl; } + + return 1; + } + catch(exception& e) { + m->errorOut(e, "AlignCommand", "driverMPI"); + exit(1); + } +} +#endif /**************************************************************************************************/ -void AlignCommand::createProcesses(string alignFileName, string reportFileName, string accnosFName, string filename) { +int AlignCommand::createProcesses(string alignFileName, string reportFileName, string accnosFName, string filename) { try { #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) int process = 0; + int exitCommand = 1; // processIDS.resize(0); //loop through and create all the processes you want @@ -477,9 +822,9 @@ void AlignCommand::createProcesses(string alignFileName, string reportFileName, 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], alignFileName + toString(getpid()) + ".temp", reportFileName + toString(getpid()) + ".temp", accnosFName + toString(getpid()) + ".temp", filename); + exitCommand = driver(lines[process], alignFileName + toString(getpid()) + ".temp", reportFileName + toString(getpid()) + ".temp", accnosFName + toString(getpid()) + ".temp", filename); exit(0); - }else { mothurOut("unable to spawn the necessary processes."); mothurOutEndLine(); exit(0); } + }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); } } //force parent to wait until all the processes are done @@ -487,10 +832,12 @@ void AlignCommand::createProcesses(string alignFileName, string reportFileName, int temp = processIDS[i]; wait(&temp); } + + return exitCommand; #endif } catch(exception& e) { - errorOut(e, "AlignCommand", "createProcesses"); + m->errorOut(e, "AlignCommand", "createProcesses"); exit(1); } } @@ -514,7 +861,7 @@ void AlignCommand::appendAlignFiles(string temp, string filename) { output.close(); } catch(exception& e) { - errorOut(e, "AlignCommand", "appendAlignFiles"); + m->errorOut(e, "AlignCommand", "appendAlignFiles"); exit(1); } } @@ -539,9 +886,8 @@ void AlignCommand::appendReportFiles(string temp, string filename) { output.close(); } catch(exception& e) { - errorOut(e, "AlignCommand", "appendReportFiles"); + m->errorOut(e, "AlignCommand", "appendReportFiles"); exit(1); } } - //**********************************************************************************************************************