X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=filterseqscommand.cpp;h=1b3b3cfd415cd51605fe3c09a8b3b5a9f11b160d;hb=c6e5f84783b39a558119835963dec3c4facb0154;hp=3784c3deab49a0c4302101372247f48f5ec3e89b;hpb=80c3fe03a8fd1635d3ad7918ff00beb973449f50;p=mothur.git diff --git a/filterseqscommand.cpp b/filterseqscommand.cpp index 3784c3d..1b3b3cf 100644 --- a/filterseqscommand.cpp +++ b/filterseqscommand.cpp @@ -28,7 +28,7 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { OptionParser parser(option); map parameters = parser.getParameters(); - ValidParameters validParameter; + ValidParameters validParameter("filter.seqs"); map::iterator it; //check to make sure all parameters are valid for command @@ -44,7 +44,7 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { it = parameters.find("fasta"); //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["fasta"] = inputDir + it->second; } } @@ -52,7 +52,7 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { it = parameters.find("hard"); //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["hard"] = inputDir + it->second; } } @@ -62,25 +62,25 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { fasta = validParameter.validFile(parameters, "fasta", false); if (fasta == "not found") { m->mothurOut("fasta is a required parameter for the filter.seqs command."); m->mothurOutEndLine(); abort = true; } else { - splitAtDash(fasta, fastafileNames); + m->splitAtDash(fasta, fastafileNames); //go through files and make sure they are good, if not, then disregard them for (int i = 0; i < fastafileNames.size(); i++) { if (inputDir != "") { - string path = hasPath(fastafileNames[i]); + string path = m->hasPath(fastafileNames[i]); //if the user has not given a path then, add inputdir. else leave path alone. if (path == "") { fastafileNames[i] = inputDir + fastafileNames[i]; } } ifstream in; - int ableToOpen = openInputFile(fastafileNames[i], in, "noerror"); + int ableToOpen = m->openInputFile(fastafileNames[i], in, "noerror"); //if you can't open it, try default location if (ableToOpen == 1) { if (m->getDefaultPath() != "") { //default path is set - string tryPath = m->getDefaultPath() + getSimpleName(fastafileNames[i]); + string tryPath = m->getDefaultPath() + m->getSimpleName(fastafileNames[i]); m->mothurOut("Unable to open " + fastafileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine(); - ableToOpen = openInputFile(tryPath, in, "noerror"); + ableToOpen = m->openInputFile(tryPath, in, "noerror"); fastafileNames[i] = tryPath; } } @@ -92,7 +92,7 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { fastafileNames.erase(fastafileNames.begin()+i); i--; }else{ - string simpleName = getSimpleName(fastafileNames[i]); + string simpleName = m->getSimpleName(fastafileNames[i]); filterFileName += simpleName.substr(0, simpleName.find_first_of('.')); } in.close(); @@ -106,7 +106,7 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { //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 = ""; - outputDir += hasPath(fastafileNames[0]); //if user entered a file with a path then preserve it + outputDir += m->hasPath(fastafileNames[0]); //if user entered a file with a path then preserve it } } //check for optional parameter and set defaults @@ -174,7 +174,7 @@ int FilterSeqsCommand::execute() { if (abort == true) { return 0; } ifstream inFASTA; - openInputFile(fastafileNames[0], inFASTA); + m->openInputFile(fastafileNames[0], inFASTA); Sequence testSeq(inFASTA); alignmentLength = testSeq.getAlignLength(); @@ -199,7 +199,7 @@ int FilterSeqsCommand::execute() { ofstream outFilter; string filterFile = outputDir + filterFileName + ".filter"; - openOutputFile(filterFile, outFilter); + m->openOutputFile(filterFile, outFilter); outFilter << filter << endl; outFilter.close(); outputNames.push_back(filterFile); @@ -254,11 +254,11 @@ int FilterSeqsCommand::filterSequences() { for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear(); - string filteredFasta = outputDir + getRootName(getSimpleName(fastafileNames[s])) + "filter.fasta"; + string filteredFasta = outputDir + m->getRootName(m->getSimpleName(fastafileNames[s])) + "filter.fasta"; #ifdef USE_MPI int pid, start, end, numSeqsPerProcessor, num; int tag = 2001; - vectorMPIPos; + vectorMPIPos; MPI_Status status; MPI_Comm_size(MPI_COMM_WORLD, &processors); //set processors to the number of mpi processes running @@ -283,7 +283,7 @@ int FilterSeqsCommand::filterSequences() { if (pid == 0) { //you are the root process - MPIPos = setFilePosFasta(fastafileNames[s], num); //fills MPIPos, returns numSeqs + MPIPos = m->setFilePosFasta(fastafileNames[s], num); //fills MPIPos, returns numSeqs numSeqs += num; //send file positions to all processes @@ -338,45 +338,32 @@ int FilterSeqsCommand::filterSequences() { MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case #else + vector positions = m->divideFile(fastafileNames[s], processors); + + for (int i = 0; i < (positions.size()-1); i++) { + lines.push_back(new linePair(positions[i], positions[(i+1)])); + } #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) if(processors == 1){ - ifstream inFASTA; - int numFastaSeqs; - openInputFile(fastafileNames[s], inFASTA); - getNumSeqs(inFASTA, numFastaSeqs); - inFASTA.close(); - - lines.push_back(new linePair(0, numFastaSeqs)); - + int numFastaSeqs = driverRunFilter(filter, filteredFasta, fastafileNames[s], lines[0]); numSeqs += numFastaSeqs; - - driverRunFilter(filter, filteredFasta, fastafileNames[s], lines[0]); }else{ - setLines(fastafileNames[s]); - createProcessesRunFilter(filter, fastafileNames[s]); + int numFastaSeqs = createProcessesRunFilter(filter, fastafileNames[s]); + numSeqs += numFastaSeqs; rename((fastafileNames[s] + toString(processIDS[0]) + ".temp").c_str(), filteredFasta.c_str()); //append fasta files for(int i=1;iappendFiles((fastafileNames[s] + toString(processIDS[i]) + ".temp"), filteredFasta); remove((fastafileNames[s] + toString(processIDS[i]) + ".temp").c_str()); } } if (m->control_pressed) { return 1; } #else - ifstream inFASTA; - int numFastaSeqs; - openInputFile(fastafileNames[s], inFASTA); - getNumSeqs(inFASTA, numFastaSeqs); - inFASTA.close(); - - lines.push_back(new linePair(0, numFastaSeqs)); - + int numFastaSeqs = driverRunFilter(filter, filteredFasta, fastafileNames[s], lines[0]); numSeqs += numFastaSeqs; - - driverRunFilter(filter, filteredFasta, fastafileNames[s], lines[0]); if (m->control_pressed) { return 1; } #endif @@ -393,7 +380,7 @@ int FilterSeqsCommand::filterSequences() { } #ifdef USE_MPI /**************************************************************************************/ -int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_File& outMPI, vector& MPIPos) { +int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_File& outMPI, vector& MPIPos) { try { string outputString = ""; int count = 0; @@ -413,7 +400,7 @@ int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_Fil istringstream iss (tempBuf,istringstream::in); delete buf4; - Sequence seq(iss); gobble(iss); + Sequence seq(iss); m->gobble(iss); if (seq.getName() != "") { string align = seq.getAligned(); @@ -466,21 +453,24 @@ int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_Fil } #endif /**************************************************************************************/ -int FilterSeqsCommand::driverRunFilter(string F, string outputFilename, string inputFilename, linePair* line) { +int FilterSeqsCommand::driverRunFilter(string F, string outputFilename, string inputFilename, linePair* filePos) { try { ofstream out; - openOutputFile(outputFilename, out); + m->openOutputFile(outputFilename, out); ifstream in; - openInputFile(inputFilename, in); + m->openInputFile(inputFilename, in); - in.seekg(line->start); - - for(int i=0;inum;i++){ + in.seekg(filePos->start); + + bool done = false; + int count = 0; + + while (!done) { if (m->control_pressed) { in.close(); out.close(); return 0; } - Sequence seq(in); + Sequence seq(in); m->gobble(in); if (seq.getName() != "") { string align = seq.getAligned(); string filterSeq = ""; @@ -492,13 +482,23 @@ int FilterSeqsCommand::driverRunFilter(string F, string outputFilename, string i } out << '>' << seq.getName() << endl << filterSeq << endl; - } - gobble(in); + count++; + } + + unsigned long int pos = in.tellg(); + if ((pos == -1) || (pos >= filePos->end)) { break; } + + //report progress + if((count) % 100 == 0){ m->mothurOut(toString(count)); m->mothurOutEndLine(); } } + //report progress + if((count) % 100 != 0){ m->mothurOut(toString(count)); m->mothurOutEndLine(); } + + out.close(); in.close(); - return 0; + return count; } catch(exception& e) { m->errorOut(e, "FilterSeqsCommand", "driverRunFilter"); @@ -511,7 +511,7 @@ int FilterSeqsCommand::createProcessesRunFilter(string F, string filename) { try { #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) int process = 0; - int exitCommand = 1; + int num = 0; processIDS.clear(); //loop through and create all the processes you want @@ -523,7 +523,15 @@ int FilterSeqsCommand::createProcessesRunFilter(string F, string filename) { process++; }else if (pid == 0){ string filteredFasta = filename + toString(getpid()) + ".temp"; - driverRunFilter(F, filteredFasta, filename, lines[process]); + num = driverRunFilter(F, filteredFasta, filename, lines[process]); + + //pass numSeqs to parent + ofstream out; + string tempFile = filename + toString(getpid()) + ".num.temp"; + m->openOutputFile(tempFile, out); + out << num << endl; + out.close(); + exit(0); }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); } } @@ -532,9 +540,18 @@ int FilterSeqsCommand::createProcessesRunFilter(string F, string filename) { for (int i=0;iopenInputFile(tempFile, in); + if (!in.eof()) { int tempNum = 0; in >> tempNum; num += tempNum; } + in.close(); remove(tempFile.c_str()); } + - return exitCommand; + return num; #endif } catch(exception& e) { @@ -553,7 +570,7 @@ string FilterSeqsCommand::createFilter() { F.setLength(alignmentLength); - if(trump != '*' || isTrue(vertical) || soft != 0){ + if(trump != '*' || m->isTrue(vertical) || soft != 0){ F.initialize(); } @@ -561,7 +578,7 @@ string FilterSeqsCommand::createFilter() { else { F.setFilter(string(alignmentLength, '1')); } numSeqs = 0; - if(trump != '*' || isTrue(vertical) || soft != 0){ + if(trump != '*' || m->isTrue(vertical) || soft != 0){ for (int s = 0; s < fastafileNames.size(); s++) { for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear(); @@ -569,7 +586,7 @@ string FilterSeqsCommand::createFilter() { #ifdef USE_MPI int pid, numSeqsPerProcessor, num; int tag = 2001; - vector MPIPos; + vector MPIPos; MPI_Status status; MPI_File inMPI; @@ -587,7 +604,7 @@ string FilterSeqsCommand::createFilter() { if (m->control_pressed) { MPI_File_close(&inMPI); return 0; } if (pid == 0) { //you are the root process - MPIPos = setFilePosFasta(fastafileNames[s], num); //fills MPIPos, returns numSeqs + MPIPos = m->setFilePosFasta(fastafileNames[s], num); //fills MPIPos, returns numSeqs numSeqs += num; //send file positions to all processes @@ -629,37 +646,24 @@ string FilterSeqsCommand::createFilter() { MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case #else + vector positions = m->divideFile(fastafileNames[s], processors); + + for (int i = 0; i < (positions.size()-1); i++) { + lines.push_back(new linePair(positions[i], positions[(i+1)])); + } #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) if(processors == 1){ - ifstream inFASTA; - int numFastaSeqs; - openInputFile(fastafileNames[s], inFASTA); - getNumSeqs(inFASTA, numFastaSeqs); - inFASTA.close(); - + int numFastaSeqs = driverCreateFilter(F, fastafileNames[s], lines[0]); numSeqs += numFastaSeqs; - - lines.push_back(new linePair(0, numFastaSeqs)); - - driverCreateFilter(F, fastafileNames[s], lines[0]); }else{ - setLines(fastafileNames[s]); - createProcessesCreateFilter(F, fastafileNames[s]); + int numFastaSeqs = createProcessesCreateFilter(F, fastafileNames[s]); + numSeqs += numFastaSeqs; } if (m->control_pressed) { return filterString; } #else - ifstream inFASTA; - int numFastaSeqs; - openInputFile(fastafileNames[s], inFASTA); - getNumSeqs(inFASTA, numFastaSeqs); - inFASTA.close(); - + int numFastaSeqs = driverCreateFilter(F, fastafileNames[s], lines[0]); numSeqs += numFastaSeqs; - - lines.push_back(new linePair(0, numFastaSeqs)); - - driverCreateFilter(F, fastafileNames[s], lines[0]); if (m->control_pressed) { return filterString; } #endif #endif @@ -675,7 +679,7 @@ string FilterSeqsCommand::createFilter() { MPI_Comm_rank(MPI_COMM_WORLD, &pid); - if(trump != '*' || isTrue(vertical) || soft != 0){ + if(trump != '*' || m->isTrue(vertical) || soft != 0){ if (pid == 0) { //only one process should output the filter @@ -723,11 +727,10 @@ string FilterSeqsCommand::createFilter() { if (pid == 0) { //only one process should output the filter #endif + F.setNumSeqs(numSeqs); - - if(isTrue(vertical) == 1) { F.doVertical(); } + if(m->isTrue(vertical) == 1) { F.doVertical(); } if(soft != 0) { F.doSoft(); } - filterString = F.getFilter(); #ifdef USE_MPI @@ -758,37 +761,42 @@ string FilterSeqsCommand::createFilter() { } } /**************************************************************************************/ -int FilterSeqsCommand::driverCreateFilter(Filters& F, string filename, linePair* line) { +int FilterSeqsCommand::driverCreateFilter(Filters& F, string filename, linePair* filePos) { try { ifstream in; - openInputFile(filename, in); + m->openInputFile(filename, in); - in.seekg(line->start); - - for(int i=0;inum;i++){ + in.seekg(filePos->start); + + bool done = false; + int count = 0; + + while (!done) { if (m->control_pressed) { in.close(); return 1; } - Sequence seq(in); + Sequence seq(in); m->gobble(in); if (seq.getName() != "") { if (seq.getAligned().length() != alignmentLength) { m->mothurOut("Sequences are not all the same length, please correct."); m->mothurOutEndLine(); m->control_pressed = true; } - if(trump != '*'){ F.doTrump(seq); } - if(isTrue(vertical) || soft != 0){ F.getFreqs(seq); } + if(trump != '*') { F.doTrump(seq); } + if(m->isTrue(vertical) || soft != 0) { F.getFreqs(seq); } cout.flush(); + count++; } + unsigned long int pos = in.tellg(); + if ((pos == -1) || (pos >= filePos->end)) { break; } + //report progress - if((i+1) % 100 == 0){ m->mothurOut(toString(i+1)); m->mothurOutEndLine(); } + if((count) % 100 == 0){ m->mothurOut(toString(count)); m->mothurOutEndLine(); } } - //report progress - if((line->num) % 100 != 0){ m->mothurOut(toString(line->num)); m->mothurOutEndLine(); } - + if((count) % 100 != 0){ m->mothurOut(toString(count)); m->mothurOutEndLine(); } in.close(); - return 0; + return count; } catch(exception& e) { m->errorOut(e, "FilterSeqsCommand", "driverCreateFilter"); @@ -797,7 +805,7 @@ int FilterSeqsCommand::driverCreateFilter(Filters& F, string filename, linePair* } #ifdef USE_MPI /**************************************************************************************/ -int FilterSeqsCommand::MPICreateFilter(int start, int num, Filters& F, MPI_File& inMPI, vector& MPIPos) { +int FilterSeqsCommand::MPICreateFilter(int start, int num, Filters& F, MPI_File& inMPI, vector& MPIPos) { try { MPI_Status status; @@ -824,7 +832,7 @@ int FilterSeqsCommand::MPICreateFilter(int start, int num, Filters& F, MPI_File& if (seq.getAligned().length() != alignmentLength) { cout << "Alignment length is " << alignmentLength << " and sequence " << seq.getName() << " has length " << seq.getAligned().length() << ", please correct." << endl; exit(1); } if(trump != '*'){ F.doTrump(seq); } - if(isTrue(vertical) || soft != 0){ F.getFreqs(seq); } + if(m->isTrue(vertical) || soft != 0){ F.getFreqs(seq); } cout.flush(); //report progress @@ -848,18 +856,42 @@ int FilterSeqsCommand::createProcessesCreateFilter(Filters& F, string filename) try { #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) int process = 0; - int exitCommand = 1; + int num = 0; processIDS.clear(); //loop through and create all the processes you want while (process != processors) { - int pid = vfork(); + int pid = fork(); if (pid > 0) { 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){ - driverCreateFilter(F, filename, lines[process]); + //reset child's filter counts to 0; + F.a.clear(); F.a.resize(alignmentLength, 0); + F.t.clear(); F.t.resize(alignmentLength, 0); + F.g.clear(); F.g.resize(alignmentLength, 0); + F.c.clear(); F.c.resize(alignmentLength, 0); + F.gap.clear(); F.gap.resize(alignmentLength, 0); + + num = driverCreateFilter(F, filename, lines[process]); + + //write out filter counts to file + filename += toString(getpid()) + "filterValues.temp"; + ofstream out; + m->openOutputFile(filename, out); + + out << num << endl; + out << F.getFilter() << endl; + for (int k = 0; k < alignmentLength; k++) { out << F.a[k] << '\t'; } out << endl; + for (int k = 0; k < alignmentLength; k++) { out << F.t[k] << '\t'; } out << endl; + for (int k = 0; k < alignmentLength; k++) { out << F.g[k] << '\t'; } out << endl; + for (int k = 0; k < alignmentLength; k++) { out << F.c[k] << '\t'; } out << endl; + for (int k = 0; k < alignmentLength; k++) { out << F.gap[k] << '\t'; } out << endl; + + //cout << F.getFilter() << endl; + out.close(); + exit(0); }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); } } @@ -870,70 +902,35 @@ int FilterSeqsCommand::createProcessesCreateFilter(Filters& F, string filename) wait(&temp); } - return exitCommand; -#endif - } - catch(exception& e) { - m->errorOut(e, "FilterSeqsCommand", "createProcessesCreateFilter"); - exit(1); - } -} -/**************************************************************************************************/ - -int FilterSeqsCommand::setLines(string filename) { - try { - - vector positions; - bufferSizes.clear(); - - ifstream inFASTA; - openInputFile(filename, inFASTA); + //parent reads in and combine Filter info + for (int i = 0; i < processIDS.size(); i++) { + string tempFilename = filename + toString(processIDS[i]) + "filterValues.temp"; + ifstream in; + m->openInputFile(tempFilename, in); - string input; - while(!inFASTA.eof()){ - input = getline(inFASTA); + int temp, tempNum; + string tempFilterString; - if (input.length() != 0) { - if(input[0] == '>'){ unsigned long int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); } - } - } - inFASTA.close(); - - int numFastaSeqs = positions.size(); - - FILE * pFile; - unsigned long int size; - - //get num bytes in file - pFile = fopen (filename.c_str(),"rb"); - if (pFile==NULL) perror ("Error opening file"); - else{ - fseek (pFile, 0, SEEK_END); - size=ftell (pFile); - fclose (pFile); - } - - numSeqs += numFastaSeqs; - - int numSeqsPerProcessor = numFastaSeqs / processors; - - for (int i = 0; i < processors; i++) { + in >> tempNum; m->gobble(in); num += tempNum; - unsigned long int startPos = positions[ i * numSeqsPerProcessor ]; - if(i == processors - 1){ - numSeqsPerProcessor = numFastaSeqs - i * numSeqsPerProcessor; - bufferSizes.push_back(size - startPos); - }else{ - unsigned long int myEnd = positions[ (i+1) * numSeqsPerProcessor ]; - bufferSizes.push_back(myEnd-startPos); - } - lines.push_back(new linePair(startPos, numSeqsPerProcessor)); + in >> tempFilterString; + F.mergeFilter(tempFilterString); + + for (int k = 0; k < alignmentLength; k++) { in >> temp; F.a[k] += temp; } m->gobble(in); + for (int k = 0; k < alignmentLength; k++) { in >> temp; F.t[k] += temp; } m->gobble(in); + for (int k = 0; k < alignmentLength; k++) { in >> temp; F.g[k] += temp; } m->gobble(in); + for (int k = 0; k < alignmentLength; k++) { in >> temp; F.c[k] += temp; } m->gobble(in); + for (int k = 0; k < alignmentLength; k++) { in >> temp; F.gap[k] += temp; } m->gobble(in); + + in.close(); + remove(tempFilename.c_str()); } - return 0; + return num; +#endif } catch(exception& e) { - m->errorOut(e, "FilterSeqsCommand", "setLines"); + m->errorOut(e, "FilterSeqsCommand", "createProcessesCreateFilter"); exit(1); } }