X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=filterseqscommand.cpp;h=7a8fbc8cdfe6236601dbc0b369f6ef415a5c7c48;hb=c7e8c2d15bd7cedcfdf18675cb0ea1a0dcd0e3c0;hp=a677eef9729ef5f1035db03f0a5daff9f7280e36;hpb=1d898dc6edaf9e9f287fab53bf1f21fb29757a17;p=mothur.git diff --git a/filterseqscommand.cpp b/filterseqscommand.cpp index a677eef..7a8fbc8 100644 --- a/filterseqscommand.cpp +++ b/filterseqscommand.cpp @@ -57,6 +57,27 @@ string FilterSeqsCommand::getHelpString(){ } } //********************************************************************************************************************** +string FilterSeqsCommand::getOutputFileNameTag(string type, string inputName=""){ + try { + string outputFileName = ""; + map >::iterator it; + + //is this a type this command creates + it = outputTypes.find(type); + if (it == outputTypes.end()) { m->mothurOut("[ERROR]: this command doesn't create a " + type + " output file.\n"); } + else { + if (type == "fasta") { outputFileName = "filter.fasta"; } + else if (type == "filter") { outputFileName = "filter"; } + else { m->mothurOut("[ERROR]: No definition for type " + type + " output file tag.\n"); m->control_pressed = true; } + } + return outputFileName; + } + catch(exception& e) { + m->errorOut(e, "FilterSeqsCommand", "getOutputFileNameTag"); + exit(1); + } +} +//********************************************************************************************************************** FilterSeqsCommand::FilterSeqsCommand(){ try { abort = true; calledHelp = true; @@ -78,6 +99,7 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { //allow user to run help if(option == "help") { help(); abort = true; calledHelp = true; } + else if(option == "citation") { citation(); abort = true; calledHelp = true;} else { vector myArray = setParameters(); @@ -124,7 +146,12 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { fasta = validParameter.validFile(parameters, "fasta", false); if (fasta == "not found") { fasta = m->getFastaFile(); - if (fasta != "") { fastafileNames.push_back(fasta); m->mothurOut("Using " + fasta + " as input file for the fasta parameter."); m->mothurOutEndLine(); } + if (fasta != "") { + fastafileNames.push_back(fasta); + m->mothurOut("Using " + fasta + " as input file for the fasta parameter."); m->mothurOutEndLine(); + string simpleName = m->getSimpleName(fasta); + filterFileName += simpleName.substr(0, simpleName.find_first_of('.')); + } else { m->mothurOut("You have no current fastafile and the fasta parameter is required."); m->mothurOutEndLine(); abort = true; } } else { @@ -132,51 +159,67 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { //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 = 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 = m->openInputFile(fastafileNames[i], in, "noerror"); - - //if you can't open it, try default location - if (ableToOpen == 1) { - if (m->getDefaultPath() != "") { //default path is set - string tryPath = m->getDefaultPath() + m->getSimpleName(fastafileNames[i]); - m->mothurOut("Unable to open " + fastafileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine(); - ifstream in2; - ableToOpen = m->openInputFile(tryPath, in2, "noerror"); - in2.close(); - fastafileNames[i] = tryPath; - } - } - //if you can't open it, try default location - if (ableToOpen == 1) { - if (m->getOutputDir() != "") { //default path is set - string tryPath = m->getOutputDir() + m->getSimpleName(fastafileNames[i]); - m->mothurOut("Unable to open " + fastafileNames[i] + ". Trying output directory " + tryPath); m->mothurOutEndLine(); - ifstream in2; - ableToOpen = m->openInputFile(tryPath, in2, "noerror"); - in2.close(); - fastafileNames[i] = tryPath; + bool ignore = false; + if (fastafileNames[i] == "current") { + fastafileNames[i] = m->getFastaFile(); + if (fastafileNames[i] != "") { m->mothurOut("Using " + fastafileNames[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 + fastafileNames.erase(fastafileNames.begin()+i); + i--; } } - in.close(); + if (!ignore) { + if (inputDir != "") { + 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 = m->openInputFile(fastafileNames[i], in, "noerror"); - if (ableToOpen == 1) { - m->mothurOut("Unable to open " + fastafileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); - //erase from file list - fastafileNames.erase(fastafileNames.begin()+i); - i--; - }else{ - string simpleName = m->getSimpleName(fastafileNames[i]); - filterFileName += simpleName.substr(0, simpleName.find_first_of('.')); + //if you can't open it, try default location + if (ableToOpen == 1) { + if (m->getDefaultPath() != "") { //default path is set + string tryPath = m->getDefaultPath() + m->getSimpleName(fastafileNames[i]); + m->mothurOut("Unable to open " + fastafileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine(); + ifstream in2; + ableToOpen = m->openInputFile(tryPath, in2, "noerror"); + in2.close(); + fastafileNames[i] = tryPath; + } + } + + //if you can't open it, try default location + if (ableToOpen == 1) { + if (m->getOutputDir() != "") { //default path is set + string tryPath = m->getOutputDir() + m->getSimpleName(fastafileNames[i]); + m->mothurOut("Unable to open " + fastafileNames[i] + ". Trying output directory " + tryPath); m->mothurOutEndLine(); + ifstream in2; + ableToOpen = m->openInputFile(tryPath, in2, "noerror"); + in2.close(); + fastafileNames[i] = tryPath; + } + } + + in.close(); + + if (ableToOpen == 1) { + m->mothurOut("Unable to open " + fastafileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); + //erase from file list + fastafileNames.erase(fastafileNames.begin()+i); + i--; + }else{ + string simpleName = m->getSimpleName(fastafileNames[i]); + filterFileName += simpleName.substr(0, simpleName.find_first_of('.')); + m->setFastaFile(fastafileNames[i]); + } + in.close(); } - in.close(); } //make sure there is at least one valid file left @@ -205,7 +248,7 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); } m->setProcessors(temp); - convert(temp, processors); + m->mothurConvert(temp, processors); vertical = validParameter.validFile(parameters, "vertical", false); if (vertical == "not found") { @@ -256,8 +299,8 @@ int FilterSeqsCommand::execute() { //prevent giantic file name string filterFile; - if (fastafileNames.size() > 3) { filterFile = outputDir + "merge.filter"; } - else { filterFile = outputDir + filterFileName + ".filter"; } + if (fastafileNames.size() > 3) { filterFile = outputDir + "merge." + getOutputFileNameTag("filter"); } + else { filterFile = outputDir + filterFileName + "." + getOutputFileNameTag("filter"); } m->openOutputFile(filterFile, outFilter); outFilter << filter << endl; @@ -281,7 +324,7 @@ int FilterSeqsCommand::execute() { if(filter[i] == '1'){ filteredLength++; } } - if (m->control_pressed) { outputTypes.clear(); for(int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; } + if (m->control_pressed) { outputTypes.clear(); for(int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } return 0; } m->mothurOutEndLine(); @@ -320,11 +363,11 @@ int FilterSeqsCommand::filterSequences() { for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear(); - string filteredFasta = outputDir + m->getRootName(m->getSimpleName(fastafileNames[s])) + "filter.fasta"; + string filteredFasta = outputDir + m->getRootName(m->getSimpleName(fastafileNames[s])) + getOutputFileNameTag("fasta"); #ifdef USE_MPI int pid, 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 @@ -403,32 +446,56 @@ int FilterSeqsCommand::filterSequences() { MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case #else - vector positions = m->divideFile(fastafileNames[s], processors); - + + vector positions; + if (savedPositions.size() != 0) { positions = savedPositions[s]; } + else { +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + positions = m->divideFile(fastafileNames[s], processors); +#else + if(processors != 1){ + int numFastaSeqs = 0; + positions = m->setFilePosFasta(fastafileNames[s], numFastaSeqs); + if (positions.size() < processors) { processors = positions.size(); } + } +#endif + } + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + //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){ int numFastaSeqs = driverRunFilter(filter, filteredFasta, fastafileNames[s], lines[0]); numSeqs += numFastaSeqs; }else{ - int numFastaSeqs = createProcessesRunFilter(filter, fastafileNames[s]); + int numFastaSeqs = createProcessesRunFilter(filter, fastafileNames[s], filteredFasta); 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 + if(processors == 1){ + lines.push_back(new linePair(0, 1000)); int numFastaSeqs = driverRunFilter(filter, filteredFasta, fastafileNames[s], lines[0]); numSeqs += numFastaSeqs; + }else { + int numFastaSeqs = positions.size()-1; + //positions = m->setFilePosFasta(fastafileNames[s], numFastaSeqs); + + //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)); + } + + numFastaSeqs = createProcessesRunFilter(filter, fastafileNames[s], filteredFasta); + numSeqs += numFastaSeqs; + } if (m->control_pressed) { return 1; } #endif @@ -445,7 +512,7 @@ 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; @@ -550,8 +617,8 @@ int FilterSeqsCommand::driverRunFilter(string F, string outputFilename, string i count++; } - #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - unsigned long int pos = in.tellg(); + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + unsigned long long pos = in.tellg(); if ((pos == -1) || (pos >= filePos->end)) { break; } #else if (in.eof()) { break; } @@ -576,12 +643,15 @@ int FilterSeqsCommand::driverRunFilter(string F, string outputFilename, string i } /**************************************************************************************************/ -int FilterSeqsCommand::createProcessesRunFilter(string F, string filename) { +int FilterSeqsCommand::createProcessesRunFilter(string F, string filename, string filteredFastaName) { try { -#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - int process = 0; + + int process = 1; int num = 0; processIDS.clear(); + +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + //loop through and create all the processes you want while (process != processors) { @@ -609,8 +679,10 @@ int FilterSeqsCommand::createProcessesRunFilter(string F, string filename) { } } + num = driverRunFilter(F, filteredFastaName, filename, lines[0]); + //force parent to wait until all the processes are done - for (int i=0;iopenInputFile(tempFile, in); if (!in.eof()) { int tempNum = 0; in >> tempNum; num += tempNum; } - in.close(); remove(tempFile.c_str()); + in.close(); m->mothurRemove(tempFile); + + m->appendFiles((filename + toString(processIDS[i]) + ".temp"), filteredFastaName); + m->mothurRemove((filename + toString(processIDS[i]) + ".temp")); } - + +#else + + ////////////////////////////////////////////////////////////////////////////////////////////////////// + //Windows version shared memory, so be careful when passing variables through the filterData struct. + //Above fork() will clone, so memory is separate, but that's not the case with windows, + //Taking advantage of shared memory to allow both threads to add info to F. + ////////////////////////////////////////////////////////////////////////////////////////////////////// + + vector pDataArray; + DWORD dwThreadIdArray[processors-1]; + HANDLE hThreadArray[processors-1]; + + //Create processor worker threads. + for( int i=0; istart, lines[i]->end, alignmentLength, i); + pDataArray.push_back(tempFilter); + processIDS.push_back(i); + + hThreadArray[i] = CreateThread(NULL, 0, MyRunFilterThreadFunction, pDataArray[i], 0, &dwThreadIdArray[i]); + } + + num = driverRunFilter(F, (filteredFastaName + toString(processors-1) + ".temp"), filename, lines[processors-1]); + + //Wait until all threads have terminated. + WaitForMultipleObjects(processors-1, hThreadArray, TRUE, INFINITE); - return num; -#endif + //Close all thread handles and free memory allocations. + for(int i=0; i < pDataArray.size(); i++){ + num += pDataArray[i]->count; + CloseHandle(hThreadArray[i]); + delete pDataArray[i]; + } + + for (int i = 1; i < processors; i++) { + m->appendFiles((filteredFastaName + toString(i) + ".temp"), filteredFastaName); + m->mothurRemove((filteredFastaName + toString(i) + ".temp")); + } +#endif + + return num; + } catch(exception& e) { m->errorOut(e, "FilterSeqsCommand", "createProcessesRunFilter"); @@ -659,7 +776,7 @@ string FilterSeqsCommand::createFilter() { #ifdef USE_MPI int pid, numSeqsPerProcessor, num; int tag = 2001; - vector MPIPos; + vector MPIPos; MPI_Status status; MPI_File inMPI; @@ -719,12 +836,14 @@ 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) + vector positions; + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + 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(processors == 1){ int numFastaSeqs = driverCreateFilter(F, fastafileNames[s], lines[0]); numSeqs += numFastaSeqs; @@ -732,13 +851,32 @@ string FilterSeqsCommand::createFilter() { int numFastaSeqs = createProcessesCreateFilter(F, fastafileNames[s]); numSeqs += numFastaSeqs; } - - if (m->control_pressed) { return filterString; } #else - int numFastaSeqs = driverCreateFilter(F, fastafileNames[s], lines[0]); - numSeqs += numFastaSeqs; - if (m->control_pressed) { return filterString; } + if(processors == 1){ + lines.push_back(new linePair(0, 1000)); + int numFastaSeqs = driverCreateFilter(F, fastafileNames[s], lines[0]); + numSeqs += numFastaSeqs; + }else { + int numFastaSeqs = 0; + positions = m->setFilePosFasta(fastafileNames[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)); + } + + numFastaSeqs = createProcessesCreateFilter(F, fastafileNames[s]); + numSeqs += numFastaSeqs; + } #endif + //save the file positions so we can reuse them in the runFilter function + savedPositions[s] = positions; + + if (m->control_pressed) { return filterString; } #endif } @@ -825,7 +963,7 @@ string FilterSeqsCommand::createFilter() { MPI_Barrier(MPI_COMM_WORLD); #endif - + return filterString; } catch(exception& e) { @@ -859,8 +997,8 @@ int FilterSeqsCommand::driverCreateFilter(Filters& F, string filename, linePair* count++; } - #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - unsigned long int pos = in.tellg(); + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + unsigned long long pos = in.tellg(); if ((pos == -1) || (pos >= filePos->end)) { break; } #else if (in.eof()) { break; } @@ -882,7 +1020,7 @@ 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; @@ -931,11 +1069,12 @@ int FilterSeqsCommand::MPICreateFilter(int start, int num, Filters& F, MPI_File& int FilterSeqsCommand::createProcessesCreateFilter(Filters& F, string filename) { try { -#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - int process = 1; + int process = 1; int num = 0; processIDS.clear(); - + +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + //loop through and create all the processes you want while (process != processors) { int pid = fork(); @@ -1007,11 +1146,53 @@ int FilterSeqsCommand::createProcessesCreateFilter(Filters& F, string filename) for (int k = 0; k < alignmentLength; k++) { in >> temp; F.gap[k] += temp; } m->gobble(in); in.close(); - remove(tempFilename.c_str()); + m->mothurRemove(tempFilename); + } + + +#else + + ////////////////////////////////////////////////////////////////////////////////////////////////////// + //Windows version shared memory, so be careful when passing variables through the filterData struct. + //Above fork() will clone, so memory is separate, but that's not the case with windows, + //Taking advantage of shared memory to allow both threads to add info to F. + ////////////////////////////////////////////////////////////////////////////////////////////////////// + + vector pDataArray; + DWORD dwThreadIdArray[processors]; + HANDLE hThreadArray[processors]; + + //Create processor worker threads. + for( int i=0; istart, lines[i]->end, alignmentLength, trump, vertical, soft, hard, i); + pDataArray.push_back(tempFilter); + processIDS.push_back(i); + + hThreadArray[i] = CreateThread(NULL, 0, MyCreateFilterThreadFunction, pDataArray[i], 0, &dwThreadIdArray[i]); + } + + //Wait until all threads have terminated. + WaitForMultipleObjects(processors, hThreadArray, TRUE, INFINITE); + + //Close all thread handles and free memory allocations. + for(int i=0; i < pDataArray.size(); i++){ + num += pDataArray[i]->count; + F.mergeFilter(pDataArray[i]->F.getFilter()); + + for (int k = 0; k < alignmentLength; k++) { F.a[k] += pDataArray[i]->F.a[k]; } + for (int k = 0; k < alignmentLength; k++) { F.t[k] += pDataArray[i]->F.t[k]; } + for (int k = 0; k < alignmentLength; k++) { F.g[k] += pDataArray[i]->F.g[k]; } + for (int k = 0; k < alignmentLength; k++) { F.c[k] += pDataArray[i]->F.c[k]; } + for (int k = 0; k < alignmentLength; k++) { F.gap[k] += pDataArray[i]->F.gap[k]; } + + CloseHandle(hThreadArray[i]); + delete pDataArray[i]; } - return num; -#endif +#endif + return num; + } catch(exception& e) { m->errorOut(e, "FilterSeqsCommand", "createProcessesCreateFilter");