X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=filterseqscommand.cpp;h=8b68988884acb408d0e72c5c3fa407a7b9f241b6;hp=addc5f338c9202811a7d6e341167c8a097e03ea0;hb=b206f634aae1b4ce13978d203247fb64757d5482;hpb=8c616f2509abd2fb9485a38607c1b439d243b85c diff --git a/filterseqscommand.cpp b/filterseqscommand.cpp index addc5f3..8b68988 100644 --- a/filterseqscommand.cpp +++ b/filterseqscommand.cpp @@ -10,20 +10,94 @@ #include "filterseqscommand.h" #include "sequence.hpp" -/**************************************************************************************/ +//********************************************************************************************************************** +vector FilterSeqsCommand::setParameters(){ + try { + 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); } + return myArray; + } + catch(exception& e) { + m->errorOut(e, "FilterSeqsCommand", "setParameters"); + exit(1); + } +} +//********************************************************************************************************************** +string FilterSeqsCommand::getHelpString(){ + try { + string helpString = ""; + helpString += "The filter.seqs command reads a file containing sequences and creates a .filter and .filter.fasta file.\n"; + helpString += "The filter.seqs command parameters are fasta, trump, soft, hard, processors and vertical. \n"; + helpString += "The fasta parameter is required, unless you have a valid current fasta file. You may enter several fasta files to build the filter from and filter, by separating their names with -'s.\n"; + helpString += "For example: fasta=abrecovery.fasta-amazon.fasta \n"; + helpString += "The trump option will remove a column if the trump character is found at that position in any sequence of the alignment. Default=*, meaning no trump. \n"; + helpString += "A soft mask removes any column where the dominant base (i.e. A, T, G, C, or U) does not occur in at least a designated percentage of sequences. Default=0.\n"; + helpString += "The hard parameter allows you to enter a file containing the filter you want to use.\n"; + helpString += "The vertical parameter removes columns where all sequences contain a gap character. The default is T.\n"; + helpString += "The processors parameter allows you to specify the number of processors to use. The default is 1.\n"; + helpString += "The filter.seqs command should be in the following format: \n"; + helpString += "filter.seqs(fasta=yourFastaFile, trump=yourTrump) \n"; + helpString += "Example filter.seqs(fasta=abrecovery.fasta, trump=.).\n"; + helpString += "Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFasta).\n"; + return helpString; + } + catch(exception& e) { + m->errorOut(e, "FilterSeqsCommand", "getHelpString"); + exit(1); + } +} +//********************************************************************************************************************** +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; + setParameters(); + vector tempOutNames; + outputTypes["fasta"] = tempOutNames; + outputTypes["filter"] = tempOutNames; + } + catch(exception& e) { + m->errorOut(e, "FilterSeqsCommand", "FilterSeqsCommand"); + exit(1); + } +} +/**************************************************************************************/ FilterSeqsCommand::FilterSeqsCommand(string option) { try { - abort = false; + abort = false; calledHelp = false; filterFileName = ""; //allow user to run help - if(option == "help") { help(); abort = true; } + if(option == "help") { help(); abort = true; calledHelp = true; } + else if(option == "citation") { citation(); abort = true; calledHelp = true;} else { - //valid paramters for this command - string Array[] = {"fasta", "trump", "soft", "hard", "vertical", "outputdir","inputdir", "processors"}; - vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); + vector myArray = setParameters(); OptionParser parser(option); map parameters = parser.getParameters(); @@ -36,6 +110,11 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; } } + //initialize outputTypes + vector tempOutNames; + outputTypes["fasta"] = tempOutNames; + outputTypes["filter"] = tempOutNames; + //if the user changes the input directory command factory will send this info to us in the output parameter string inputDir = validParameter.validFile(parameters, "inputdir", false); if (inputDir == "not found"){ inputDir = ""; } @@ -60,42 +139,82 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { //check for required parameters 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; } + 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(); + 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 { 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 = 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(); - ableToOpen = m->openInputFile(tryPath, in, "noerror"); - 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 (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 (!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 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 @@ -122,8 +241,9 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { temp = validParameter.validFile(parameters, "soft", false); if (temp == "not found") { soft = 0; } else { soft = (float)atoi(temp.c_str()) / 100.0; } - temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = "1"; } - convert(temp, processors); + temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); } + m->setProcessors(temp); + m->mothurConvert(temp, processors); vertical = validParameter.validFile(parameters, "vertical", false); if (vertical == "not found") { @@ -140,38 +260,12 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { exit(1); } } - -//********************************************************************************************************************** - -void FilterSeqsCommand::help(){ - try { - - m->mothurOut("The filter.seqs command reads a file containing sequences and creates a .filter and .filter.fasta file.\n"); - m->mothurOut("The filter.seqs command parameters are fasta, trump, soft, hard and vertical. \n"); - m->mothurOut("The fasta parameter is required. You may enter several fasta files to build the filter from and filter, by separating their names with -'s.\n"); - m->mothurOut("For example: fasta=abrecovery.fasta-amazon.fasta \n"); - m->mothurOut("The trump parameter .... The default is ...\n"); - m->mothurOut("The soft parameter .... The default is ....\n"); - m->mothurOut("The hard parameter allows you to enter a file containing the filter you want to use.\n"); - m->mothurOut("The vertical parameter removes columns where all sequences contain a gap character. The default is T.\n"); - m->mothurOut("The filter.seqs command should be in the following format: \n"); - m->mothurOut("filter.seqs(fasta=yourFastaFile, trump=yourTrump) \n"); - m->mothurOut("Example filter.seqs(fasta=abrecovery.fasta, trump=.).\n"); - m->mothurOut("Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFasta).\n\n"); - - } - catch(exception& e) { - m->errorOut(e, "FilterSeqsCommand", "help"); - exit(1); - } -} - /**************************************************************************************/ int FilterSeqsCommand::execute() { try { - if (abort == true) { return 0; } + if (abort == true) { if (calledHelp) { return 0; } return 2; } ifstream inFASTA; m->openInputFile(fastafileNames[0], inFASTA); @@ -187,7 +281,7 @@ int FilterSeqsCommand::execute() { m->mothurOutEndLine(); m->mothurOutEndLine(); - if (m->control_pressed) { return 0; } + if (m->control_pressed) { outputTypes.clear(); return 0; } #ifdef USE_MPI int pid; @@ -198,11 +292,16 @@ int FilterSeqsCommand::execute() { ofstream outFilter; - string filterFile = outputDir + filterFileName + ".filter"; + //prevent giantic file name + 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; outFilter.close(); - outputNames.push_back(filterFile); + outputNames.push_back(filterFile); outputTypes["filter"].push_back(filterFile); #ifdef USE_MPI } @@ -221,7 +320,7 @@ int FilterSeqsCommand::execute() { if(filter[i] == '1'){ filteredLength++; } } - if (m->control_pressed) { 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(); @@ -230,6 +329,12 @@ int FilterSeqsCommand::execute() { m->mothurOut("Length of the original alignment: " + toString(alignmentLength)); m->mothurOutEndLine(); m->mothurOut("Number of sequences used to construct filter: " + toString(numSeqs)); m->mothurOutEndLine(); + //set fasta file as new current fastafile + string current = ""; + itTypes = outputTypes.find("fasta"); + if (itTypes != outputTypes.end()) { + if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setFastaFile(current); } + } m->mothurOutEndLine(); m->mothurOut("Output File Names: "); m->mothurOutEndLine(); @@ -254,18 +359,19 @@ 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, start, end, numSeqsPerProcessor, num; + 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 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are MPI_File outMPI; - MPI_File tempMPI; MPI_File inMPI; int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY; int inMode=MPI_MODE_RDONLY; @@ -305,8 +411,8 @@ int FilterSeqsCommand::filterSequences() { //wait on chidren for(int i = 1; i < processors; i++) { - char buf[4]; - MPI_Recv(buf, 4, MPI_CHAR, i, tag, MPI_COMM_WORLD, &status); + char buf[5]; + MPI_Recv(buf, 5, MPI_CHAR, i, tag, MPI_COMM_WORLD, &status); } }else { //you are a child process @@ -326,11 +432,11 @@ int FilterSeqsCommand::filterSequences() { if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPI); return 0; } - char buf[4]; + char buf[5]; strcpy(buf, "done"); //tell parent you are done. - MPI_Send(buf, 4, MPI_CHAR, 0, tag, MPI_COMM_WORLD); + MPI_Send(buf, 5, MPI_CHAR, 0, tag, MPI_COMM_WORLD); } MPI_File_close(&outMPI); @@ -338,37 +444,61 @@ 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 #endif - outputNames.push_back(filteredFasta); + outputNames.push_back(filteredFasta); outputTypes["fasta"].push_back(filteredFasta); } return 0; @@ -380,7 +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; @@ -428,7 +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 @@ -442,7 +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; } @@ -485,18 +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(); @@ -511,12 +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) { @@ -537,11 +670,17 @@ int FilterSeqsCommand::createProcessesRunFilter(string F, string filename) { out.close(); exit(0); - }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); } + }else { + m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine(); + for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); } + exit(0); + } } + 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. + ////////////////////////////////////////////////////////////////////////////////////////////////////// - return num; -#endif + 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); + + //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"); @@ -590,7 +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; @@ -650,12 +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; @@ -663,13 +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 } @@ -756,7 +964,7 @@ string FilterSeqsCommand::createFilter() { MPI_Barrier(MPI_COMM_WORLD); #endif - + return filterString; } catch(exception& e) { @@ -790,18 +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; @@ -813,7 +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; @@ -844,11 +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; } @@ -862,11 +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 = 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) { int pid = fork(); @@ -901,16 +1110,23 @@ int FilterSeqsCommand::createProcessesCreateFilter(Filters& F, string filename) out.close(); exit(0); - }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); } + }else { + m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine(); + for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); } + exit(0); + } } + //parent do your part + num = driverCreateFilter(F, filename, lines[0]); + //force parent to wait until all the processes are done - for (int i=0;i> 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; + 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]; } - return num; -#endif +#endif + return num; + } catch(exception& e) { m->errorOut(e, "FilterSeqsCommand", "createProcessesCreateFilter");