X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=distancecommand.cpp;h=67b87f10e21ab5524a090241e442b388c03306aa;hb=bede5752eaeedb5a7e339439b6ba2f6670bca2b9;hp=e5878b3170fcb256d2a4d8655e32b8c6bb97a795;hpb=c35f02a218ce8f430a75850b4d9fabb96b3a022b;p=mothur.git diff --git a/distancecommand.cpp b/distancecommand.cpp index e5878b3..67b87f1 100644 --- a/distancecommand.cpp +++ b/distancecommand.cpp @@ -20,12 +20,19 @@ DistanceCommand::DistanceCommand(){ try { globaldata = GlobalData::getInstance(); validCalculator = new ValidCalculators(); - ends = globaldata->getEnds(); + countends = globaldata->getCountEnds(); convert(globaldata->getProcessors(), processors); convert(globaldata->getCutOff(), cutoff); + phylip = globaldata->getPhylipFile(); + + //open file + string filename = globaldata->getFastaFile(); + openInputFile(filename, in); + + int i; - if (ends != "T") { + if (isTrue(countends) == true) { for (i=0; iEstimators.size(); i++) { if (validCalculator->isValidCalculator("distance", globaldata->Estimators[i]) == true) { if (globaldata->Estimators[i] == "nogaps") { @@ -33,7 +40,7 @@ DistanceCommand::DistanceCommand(){ }else if (globaldata->Estimators[i] == "eachgap") { distCalculator = new eachGapDist(); }else if (globaldata->Estimators[i] == "onegap") { - distCalculator = new oneGapDist(); } + distCalculator = new oneGapDist(); } } } }else { @@ -49,7 +56,7 @@ DistanceCommand::DistanceCommand(){ } } } - + //reset calc for next command globaldata->setCalc(""); } @@ -67,113 +74,70 @@ DistanceCommand::DistanceCommand(){ int DistanceCommand::execute(){ try { - //read file - string filename = globaldata->inputFileName; - - if(globaldata->getFastaFile() != "") { - readSeqs = new ReadFasta(filename); } - else if(globaldata->getNexusFile() != "") { - readSeqs = new ReadNexus(filename); } - else if(globaldata->getClustalFile() != "") { - readSeqs = new ReadClustal(filename); } - else if(globaldata->getPhylipFile() != "") { - readSeqs = new ReadPhylip(filename); } - - readSeqs->read(); - seqDB = readSeqs->getDB(); - + //reads fasta file and fills sequenceDB + if(globaldata->getFastaFile() != "") { seqDB = new SequenceDB(in); } + else { cout << "Error no fasta file." << endl; return 0; } + int numSeqs = seqDB->getNumSeqs(); + cutoff += 0.005; - string distFile = getRootName(globaldata->getFastaFile()) + "dist"; - - remove(distFile.c_str()); + string outputFile; - // # if defined (WIN_VERSION) - // driver(distCalculator, seqDB, 0, numSeqs, distFile, cutoff); - // # endif - - # if defined (__APPLE__) || (__MACH__) - if(processors == 1){ - driver(distCalculator, seqDB, 0, numSeqs, distFile, cutoff); - } - else if(processors == 2){ + //doses the user want the phylip formatted file as well + if (isTrue(phylip) == true) { + outputFile = getRootName(globaldata->getFastaFile()) + "phylip.dist"; + remove(outputFile.c_str()); + + //output numSeqs to phylip formatted dist file + openOutputFile(outputFile, outFile); + outFile << numSeqs << endl; + outFile.close(); + }else { //user wants column format + outputFile = getRootName(globaldata->getFastaFile()) + "dist"; + remove(outputFile.c_str()); + } + + //# if defined (_WIN32) + //figure out how to implement the fork and wait commands in windows + // driver(distCalculator, seqDB, 0, numSeqs, distFile, phylipFile, cutoff); + //# endif - int pid = fork(); - if(pid > 0){ - driver(distCalculator, seqDB, 0, (numSeqs/sqrt(2)), distFile + "tempa", cutoff); - appendFiles((distFile+"tempa"), distFile); - remove((distFile + "tempa").c_str()); - } - else{ - driver(distCalculator, seqDB, (numSeqs/sqrt(2)), numSeqs, distFile + "tempb", cutoff); - appendFiles((distFile+"tempb"), distFile); - remove((distFile + "tempb").c_str()); - } - wait(NULL); - - } - else if(processors == 3){ - int pid1 = fork(); - if(pid1 > 0){ - int pid2 = fork(); - if(pid2 > 0){ - driver(distCalculator, seqDB, 0, sqrt(3) * numSeqs / 3, distFile + "tempa", cutoff); - appendFiles(distFile+"tempa", distFile); - remove((distFile + "tempa").c_str()); - } - else{ - driver(distCalculator, seqDB, sqrt(3) * numSeqs / 3, sqrt(6) * numSeqs / 3, distFile + "tempb", cutoff); - appendFiles(distFile+"tempb", distFile); - remove((distFile + "tempb").c_str()); - } - wait(NULL); - } - else{ - driver(distCalculator, seqDB, sqrt(6) * numSeqs / 3, numSeqs, distFile + "tempc", cutoff); - appendFiles(distFile+"tempc", distFile); - remove((distFile + "tempc").c_str()); - } - wait(NULL); + +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + //if you don't need to fork anything + if(processors == 1){ + driver(distCalculator, seqDB, 0, numSeqs, outputFile + ".temp", cutoff); + appendFiles((outputFile + ".temp"), outputFile); + remove((outputFile + ".temp").c_str()); + }else{ //you have multiple processors + + for (int i = 0; i < processors; i++) { + lines.push_back(new linePair()); + lines[i]->start = int (sqrt(float(i)/float(processors)) * numSeqs); + lines[i]->end = int (sqrt(float(i+1)/float(processors)) * numSeqs); } - else if(processors == 4){ - int pid1 = fork(); - if(pid1 > 0){ - int pid2 = fork(); - if(pid2 > 0){ - driver(distCalculator, seqDB, 0, numSeqs / 2, distFile + "tempa", cutoff); - appendFiles(distFile+"tempa", distFile); - remove((distFile + "tempa").c_str()); - } - else{ - driver(distCalculator, seqDB, numSeqs / 2, (numSeqs/sqrt(2)), distFile + "tempb", cutoff); - appendFiles(distFile+"tempb", distFile); - remove((distFile + "tempb").c_str()); - } - wait(NULL); - } - else{ - int pid3 = fork(); - if(pid3 > 0){ - driver(distCalculator, seqDB, (numSeqs/sqrt(2)), (sqrt(3) * numSeqs / 2), distFile + "tempc", cutoff); - appendFiles(distFile+"tempc", distFile); - remove((distFile + "tempc").c_str()); - } - else{ - driver(distCalculator, seqDB, (sqrt(3) * numSeqs / 2), numSeqs, distFile + "tempd", cutoff); - appendFiles(distFile+"tempd", distFile); - remove((distFile + "tempd").c_str()); - } - wait(NULL); - } - wait(NULL); + + cout << lines[0]->start << '\t' << lines[0]->end << endl; + cout << lines[1]->start << '\t' << lines[1]->end << endl; + + createProcesses(outputFile); + + //append and remove temp files + for (it = processIDS.begin(); it != processIDS.end(); it++) { + appendFiles((outputFile + toString(it->second) + ".temp"), outputFile); + remove((outputFile + toString(it->second) + ".temp").c_str()); } - wait(NULL); - # endif - + } +#else + driver(distCalculator, seqDB, 0, numSeqs, outputFile + ".temp", cutoff); + appendFiles((outputFile + ".temp"), outputFile); + remove((outputFile + ".temp").c_str()); +#endif + delete distCalculator; - + return 0; - + } catch(exception& e) { cout << "Standard Error: " << e.what() << " has occurred in the DistanceCommand class Function execute. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n"; @@ -184,39 +148,79 @@ int DistanceCommand::execute(){ exit(1); } } +/**************************************************************************************************/ +void DistanceCommand::createProcesses(string filename) { + try { + int process = 0; + processIDS.clear(); + + //loop through and create all the processes you want + while (process != processors) { + int pid = fork(); + + if (pid > 0) { + processIDS[lines[process]->end] = pid; //create map from line number to pid so you can append files in correct order later + process++; + }else if (pid == 0){ + driver(distCalculator, seqDB, lines[process]->start, lines[process]->end, filename + toString(getpid()) + ".temp", cutoff); + exit(0); + }else { cout << "unable to spawn the necessary processes." << endl; exit(0); } + } + + //force parent to wait until all the processes are done + for (it = processIDS.begin(); it != processIDS.end(); it++) { + int temp = it->second; + wait(&temp); + } + + } + catch(exception& e) { + cout << "Standard Error: " << e.what() << " has occurred in the DistanceCommand class Function createProcesses. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n"; + exit(1); + } + catch(...) { + cout << "An unknown error has occurred in the DistanceCommand class function createProcesses. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n"; + exit(1); + } +} /**************************************************************************************************/ /////// need to fix to work with calcs and sequencedb int DistanceCommand::driver(Dist* distCalculator, SequenceDB* align, int startLine, int endLine, string dFileName, float cutoff){ try { + int startTime = time(NULL); - - ofstream distFile(dFileName.c_str(), ios::trunc); - distFile.setf(ios::fixed, ios::showpoint); - distFile << setprecision(4); - - for(int i=startLine;iget(i).getUnaligned() << " " << align->get(j).getUnaligned() << endl; -//cout << "aligned" << endl; -//cout << align->get(i).getAligned() << " " << align->get(j).getAligned() << endl; - distCalculator->calcDist(align->get(i), align->get(j)); + distCalculator->calcDist(*(align->get(i)), *(align->get(j))); double dist = distCalculator->getDist(); - + if(dist <= cutoff){ - distFile << align->get(i).getName() << ' ' << align->get(j).getName() << ' ' << dist << endl; + if (isTrue(phylip) != true) { outFile << align->get(i)->getName() << ' ' << align->get(j)->getName() << ' ' << dist << endl; } } - + if (isTrue(phylip) == true) { outFile << dist << '\t'; } + } + + if (isTrue(phylip) == true) { outFile << endl; } + if(i % 100 == 0){ cout << i << '\t' << time(NULL) - startTime << endl; } - + } cout << endLine-1 << '\t' << time(NULL) - startTime << endl; - + + //philFile.close(); + //distFile.close(); + return 1; } catch(exception& e) { @@ -227,7 +231,7 @@ int DistanceCommand::driver(Dist* distCalculator, SequenceDB* align, int startLi cout << "An unknown error has occurred in the DistanceCommand class function driver. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n"; exit(1); } - + } /**************************************************************************************************/ @@ -235,7 +239,7 @@ void DistanceCommand::appendFiles(string temp, string filename) { try{ ofstream output; ifstream input; - + //open output file in append mode openOutputFileAppend(filename, output); @@ -250,7 +254,7 @@ void DistanceCommand::appendFiles(string temp, string filename) { output << line << endl; // Appending back newline char } } - + input.close(); output.close(); }