X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=filterseqscommand.cpp;h=8b68988884acb408d0e72c5c3fa407a7b9f241b6;hp=173227f372f4d4cbf7445e5f3fea440220e70b4a;hb=b206f634aae1b4ce13978d203247fb64757d5482;hpb=19fcbbdba99658f5eca244803280f9ee7f9f6607 diff --git a/filterseqscommand.cpp b/filterseqscommand.cpp index 173227f..8b68988 100644 --- a/filterseqscommand.cpp +++ b/filterseqscommand.cpp @@ -14,14 +14,14 @@ //********************************************************************************************************************** vector FilterSeqsCommand::setParameters(){ try { - CommandParameter pfasta("fasta", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(pfasta); - CommandParameter phard("hard", "InputTypes", "", "", "none", "none", "none",false,false); parameters.push_back(phard); - CommandParameter ptrump("trump", "String", "", "*", "", "", "",false,false); parameters.push_back(ptrump); - CommandParameter psoft("soft", "Number", "", "0", "", "", "",false,false); parameters.push_back(psoft); - CommandParameter pvertical("vertical", "Boolean", "", "T", "", "", "",false,false); parameters.push_back(pvertical); - CommandParameter pprocessors("processors", "Number", "", "1", "", "", "",false,false); parameters.push_back(pprocessors); - CommandParameter pinputdir("inputdir", "String", "", "", "", "", "",false,false); parameters.push_back(pinputdir); - CommandParameter poutputdir("outputdir", "String", "", "", "", "", "",false,false); parameters.push_back(poutputdir); + CommandParameter pfasta("fasta", "InputTypes", "", "", "none", "none", "none","fasta-filter",false,true, true); parameters.push_back(pfasta); + CommandParameter phard("hard", "InputTypes", "", "", "none", "none", "none","",false,false); parameters.push_back(phard); + CommandParameter ptrump("trump", "String", "", "*", "", "", "","",false,false, true); parameters.push_back(ptrump); + CommandParameter psoft("soft", "Number", "", "0", "", "", "","",false,false); parameters.push_back(psoft); + CommandParameter pvertical("vertical", "Boolean", "", "T", "", "", "","",false,false, true); parameters.push_back(pvertical); + CommandParameter pprocessors("processors", "Number", "", "1", "", "", "","",false,false, true); parameters.push_back(pprocessors); + 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); } @@ -57,6 +57,22 @@ string FilterSeqsCommand::getHelpString(){ } } //********************************************************************************************************************** +string FilterSeqsCommand::getOutputPattern(string type) { + try { + string pattern = ""; + + if (type == "fasta") { pattern = "[filename],filter.fasta"; } + else if (type == "filter") { pattern = "[filename],filter"; } + else { m->mothurOut("[ERROR]: No definition for type " + type + " output pattern.\n"); m->control_pressed = true; } + + return pattern; + } + catch(exception& e) { + m->errorOut(e, "FilterSeqsCommand", "getOutputPattern"); + exit(1); + } +} +//********************************************************************************************************************** FilterSeqsCommand::FilterSeqsCommand(){ try { abort = true; calledHelp = true; @@ -125,7 +141,12 @@ 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 { @@ -222,7 +243,7 @@ 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") { @@ -272,9 +293,10 @@ int FilterSeqsCommand::execute() { ofstream outFilter; //prevent giantic file name - string filterFile; - if (fastafileNames.size() > 3) { filterFile = outputDir + "merge.filter"; } - else { filterFile = outputDir + filterFileName + ".filter"; } + map variables; + variables["[filename]"] = outputDir + filterFileName + "."; + if (fastafileNames.size() > 3) { variables["[filename]"] = outputDir + "merge."; } + string filterFile = getOutputFileName("filter", variables); m->openOutputFile(filterFile, outFilter); outFilter << filter << endl; @@ -337,11 +359,13 @@ 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"; + map variables; + variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(fastafileNames[s])); + string filteredFasta = getOutputFileName("fasta", variables); #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 @@ -420,32 +444,56 @@ 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); - m->mothurRemove((fastafileNames[s] + toString(processIDS[i]) + ".temp")); - } } 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 @@ -462,7 +510,7 @@ 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; @@ -510,7 +558,7 @@ int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_Fil } - if((i+1) % 100 == 0){ cout << (i+1) << endl; m->mothurOutJustToLog(toString(i+1) + "\n"); } + if((i+1) % 100 == 0){ cout << (i+1) << endl; } } if(outputString != ""){ //output to file @@ -524,7 +572,7 @@ int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_Fil delete buf; } - if((num) % 100 != 0){ cout << (num) << endl; m->mothurOutJustToLog(toString(num) + "\n"); } + if((num) % 100 != 0){ cout << (num) << endl; } return 0; } @@ -567,18 +615,18 @@ 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; } #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"); } out.close(); @@ -593,12 +641,15 @@ 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) { @@ -626,8 +677,10 @@ 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(); 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; + 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; + } + 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"); @@ -676,7 +777,7 @@ string FilterSeqsCommand::createFilter() { #ifdef USE_MPI int pid, numSeqsPerProcessor, num; int tag = 2001; - vector MPIPos; + vector MPIPos; MPI_Status status; MPI_File inMPI; @@ -736,12 +837,14 @@ 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; @@ -749,13 +852,32 @@ 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 } @@ -842,7 +964,7 @@ string FilterSeqsCommand::createFilter() { MPI_Barrier(MPI_COMM_WORLD); #endif - + return filterString; } catch(exception& e) { @@ -876,18 +998,18 @@ 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; } #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"); } in.close(); return count; @@ -899,7 +1021,7 @@ 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; @@ -930,11 +1052,11 @@ int FilterSeqsCommand::MPICreateFilter(int start, int num, Filters& F, MPI_File& cout.flush(); //report progress - if((i+1) % 100 == 0){ cout << (i+1) << endl; m->mothurOutJustToLog(toString(i+1) + "\n"); } + if((i+1) % 100 == 0){ cout << (i+1) << endl; } } //report progress - if((num) % 100 != 0){ cout << num << endl; m->mothurOutJustToLog(toString(num) + "\n"); } + if((num) % 100 != 0){ cout << num << endl; } return 0; } @@ -948,11 +1070,12 @@ 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(); @@ -1027,8 +1150,53 @@ int FilterSeqsCommand::createProcessesCreateFilter(Filters& F, string filename) m->mothurRemove(tempFilename); } - return num; -#endif + +#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; + 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; + } + 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]; + } + +#endif + return num; + } catch(exception& e) { m->errorOut(e, "FilterSeqsCommand", "createProcessesCreateFilter");