X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=distancecommand.cpp;h=4720df3cf0d968a84ba49c4af449b226a304c1cd;hb=aba5f8811829037b0a3004ef33f0ad4ed5e5fcf8;hp=c615b8068ae1576b07f2e64a2742bfcb36e3e9b3;hpb=9c23307c583d4e8595f75278c13e708788f2f058;p=mothur.git diff --git a/distancecommand.cpp b/distancecommand.cpp index c615b80..4720df3 100644 --- a/distancecommand.cpp +++ b/distancecommand.cpp @@ -16,11 +16,11 @@ //********************************************************************************************************************** -DistanceCommand::DistanceCommand(string option){ +DistanceCommand::DistanceCommand(string option) { try { abort = false; Estimators.clear(); - + //allow user to run help if(option == "help") { help(); abort = true; } @@ -56,7 +56,7 @@ DistanceCommand::DistanceCommand(string option){ //check for required parameters fastafile = validParameter.validFile(parameters, "fasta", true); - if (fastafile == "not found") { mothurOut("fasta is a required parameter for the dist.seqs command."); mothurOutEndLine(); abort = true; } + if (fastafile == "not found") { m->mothurOut("fasta is a required parameter for the dist.seqs command."); m->mothurOutEndLine(); abort = true; } else if (fastafile == "not open") { abort = true; } else{ ifstream inFASTA; @@ -92,7 +92,7 @@ DistanceCommand::DistanceCommand(string option){ output = validParameter.validFile(parameters, "output", false); if(output == "not found"){ output = "column"; } - if ((output != "column") && (output != "lt") && (output != "square")) { mothurOut(output + " is not a valid output form. Options are column, lt and square. I will use column."); mothurOutEndLine(); output = "column"; } + if ((output != "column") && (output != "lt") && (output != "square")) { m->mothurOut(output + " is not a valid output form. Options are column, lt and square. I will use column."); m->mothurOutEndLine(); output = "column"; } ValidCalculators validCalculator; @@ -118,7 +118,7 @@ DistanceCommand::DistanceCommand(string option){ } catch(exception& e) { - errorOut(e, "DistanceCommand", "DistanceCommand"); + m->errorOut(e, "DistanceCommand", "DistanceCommand"); exit(1); } } @@ -137,21 +137,21 @@ DistanceCommand::~DistanceCommand(){ void DistanceCommand::help(){ try { - mothurOut("The dist.seqs command reads a file containing sequences and creates a distance file.\n"); - mothurOut("The dist.seqs command parameters are fasta, calc, countends, output, cutoff and processors. \n"); - mothurOut("The fasta parameter is required.\n"); - mothurOut("The calc parameter allows you to specify the method of calculating the distances. Your options are: nogaps, onegap or eachgap. The default is onegap.\n"); - mothurOut("The countends parameter allows you to specify whether to include terminal gaps in distance. Your options are: T or F. The default is T.\n"); - mothurOut("The cutoff parameter allows you to specify maximum distance to keep. The default is 1.0.\n"); - mothurOut("The output parameter allows you to specify format of your distance matrix. Options are column, lt, and square. The default is column.\n"); - mothurOut("The processors parameter allows you to specify number of processors to use. The default is 1.\n"); - mothurOut("The dist.seqs command should be in the following format: \n"); - mothurOut("dist.seqs(fasta=yourFastaFile, calc=yourCalc, countends=yourEnds, cutoff= yourCutOff, processors=yourProcessors) \n"); - mothurOut("Example dist.seqs(fasta=amazon.fasta, calc=eachgap, countends=F, cutoff= 2.0, processors=3).\n"); - mothurOut("Note: No spaces between parameter labels (i.e. calc), '=' and parameters (i.e.yourCalc).\n\n"); + m->mothurOut("The dist.seqs command reads a file containing sequences and creates a distance file.\n"); + m->mothurOut("The dist.seqs command parameters are fasta, calc, countends, output, cutoff and processors. \n"); + m->mothurOut("The fasta parameter is required.\n"); + m->mothurOut("The calc parameter allows you to specify the method of calculating the distances. Your options are: nogaps, onegap or eachgap. The default is onegap.\n"); + m->mothurOut("The countends parameter allows you to specify whether to include terminal gaps in distance. Your options are: T or F. The default is T.\n"); + m->mothurOut("The cutoff parameter allows you to specify maximum distance to keep. The default is 1.0.\n"); + m->mothurOut("The output parameter allows you to specify format of your distance matrix. Options are column, lt, and square. The default is column.\n"); + m->mothurOut("The processors parameter allows you to specify number of processors to use. The default is 1.\n"); + m->mothurOut("The dist.seqs command should be in the following format: \n"); + m->mothurOut("dist.seqs(fasta=yourFastaFile, calc=yourCalc, countends=yourEnds, cutoff= yourCutOff, processors=yourProcessors) \n"); + m->mothurOut("Example dist.seqs(fasta=amazon.fasta, calc=eachgap, countends=F, cutoff= 2.0, processors=3).\n"); + m->mothurOut("Note: No spaces between parameter labels (i.e. calc), '=' and parameters (i.e.yourCalc).\n\n"); } catch(exception& e) { - errorOut(e, "DistanceCommand", "help"); + m->errorOut(e, "DistanceCommand", "help"); exit(1); } } @@ -166,7 +166,7 @@ int DistanceCommand::execute(){ cutoff += 0.005; string outputFile; - + if (output == "lt") { //does the user want lower triangle phylip formatted file outputFile = outputDir + getRootName(getSimpleName(fastafile)) + "phylip.dist"; remove(outputFile.c_str()); @@ -179,8 +179,60 @@ int DistanceCommand::execute(){ outputFile = outputDir + getRootName(getSimpleName(fastafile)) + "square.dist"; remove(outputFile.c_str()); } + + +#ifdef USE_MPI + + int pid, start, end; + int tag = 2001; -#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + 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 + + //each process gets where it should start and stop in the file + start = int (sqrt(float(pid)/float(processors)) * numSeqs); + end = int (sqrt(float(pid+1)/float(processors)) * numSeqs); + + MPI_File outMPI; + int amode=MPI_MODE_CREATE|MPI_MODE_WRONLY; + + char filename[outputFile.length()]; + strcpy(filename, outputFile.c_str()); + + MPI_File_open(MPI_COMM_WORLD, filename, amode, MPI_INFO_NULL, &outMPI); + + if (pid == 0) { //you are the root process + + //do your part + string outputMyPart; + driverMPI(start, end, outMPI, cutoff); + + //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); + } + + if (output == "lt") { + convertToLowerTriangle(outputFile); + } + + }else { //you are a child process + //do your part + driverMPI(start, end, outMPI, cutoff); + + char buf[4]; + strcpy(buf, "done"); + + //tell parent you are done. + MPI_Send(buf, 4, MPI_CHAR, 0, tag, MPI_COMM_WORLD); + } + + MPI_File_close(&outMPI); +#else + + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) //if you don't need to fork anything if(processors == 1){ driver(0, numSeqs, outputFile, cutoff); @@ -204,20 +256,40 @@ int DistanceCommand::execute(){ remove((outputFile + toString(it->second) + ".temp").c_str()); } } -#else + #else ifstream inFASTA; driver(0, numSeqs, outputFile, cutoff); + #endif + #endif + if (m->control_pressed) { delete distCalculator; remove(outputFile.c_str()); return 0; } + + #ifdef USE_MPI + MPI_Comm_rank(MPI_COMM_WORLD, &pid); + + if (pid == 0) { //only one process should output to screen + #endif if (output == "square") { convertMatrix(outputFile); } + #ifdef USE_MPI + } + #endif + + if (m->control_pressed) { delete distCalculator; remove(outputFile.c_str()); return 0; } + delete distCalculator; + m->mothurOutEndLine(); + m->mothurOut("Output File Name: "); m->mothurOutEndLine(); + m->mothurOut(outputFile); m->mothurOutEndLine(); + m->mothurOutEndLine(); + return 0; } catch(exception& e) { - errorOut(e, "DistanceCommand", "execute"); + m->errorOut(e, "DistanceCommand", "execute"); exit(1); } } @@ -238,7 +310,7 @@ void DistanceCommand::createProcesses(string filename) { }else if (pid == 0){ driver(lines[process]->start, lines[process]->end, filename + toString(getpid()) + ".temp", cutoff); exit(0); - }else { mothurOut("unable to spawn the necessary processes."); mothurOutEndLine(); exit(0); } + }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); } } //force parent to wait until all the processes are done @@ -249,7 +321,7 @@ void DistanceCommand::createProcesses(string filename) { #endif } catch(exception& e) { - errorOut(e, "DistanceCommand", "createProcesses"); + m->errorOut(e, "DistanceCommand", "createProcesses"); exit(1); } } @@ -277,6 +349,9 @@ int DistanceCommand::driver(int startLine, int endLine, string dFileName, float outFile << name << '\t'; } for(int j=0;jcontrol_pressed) { outFile.close(); return 0; } + distCalculator->calcDist(alignDB.get(i), alignDB.get(j)); double dist = distCalculator->getDist(); @@ -295,23 +370,88 @@ int DistanceCommand::driver(int startLine, int endLine, string dFileName, float if (output == "lt") { outFile << endl; } if(i % 100 == 0){ - mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); mothurOutEndLine(); + m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine(); } } - mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); mothurOutEndLine(); + m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine(); outFile.close(); return 1; } catch(exception& e) { - errorOut(e, "DistanceCommand", "driver"); + m->errorOut(e, "DistanceCommand", "driver"); exit(1); } } +#ifdef USE_MPI /**************************************************************************************************/ -void DistanceCommand::convertMatrix(string outputFile) { +/////// need to fix to work with calcs and sequencedb +int DistanceCommand::driverMPI(int startLine, int endLine, MPI_File& outMPI, float cutoff){ + try { + MPI_Status status; + int startTime = time(NULL); + + string outputString = ""; + + for(int i=startLine;icontrol_pressed) { return 0; } + + distCalculator->calcDist(alignDB.get(i), alignDB.get(j)); + double dist = distCalculator->getDist(); + + if(dist <= cutoff){ + if (output == "column") { outputString += (alignDB.get(i).getName() + ' ' + alignDB.get(j).getName() + ' ' + toString(dist) + '\n'); } + } + + if ((output == "square") || (output == "lt")){ //make a square column you can convert to square phylip + outputString += (alignDB.get(i).getName() + ' ' + alignDB.get(j).getName() + ' ' + toString(dist) + '\n'); + outputString += (alignDB.get(j).getName() + ' ' + alignDB.get(i).getName() + ' ' + toString(dist) + '\n'); + } + + } + + if(i % 100 == 0){ + m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine(); + } + + if(i % 10 == 0){ //output to file + //send results to parent + int length = outputString.length(); + char buf[length]; + strcpy(buf, outputString.c_str()); + + MPI_File_write_shared(outMPI, buf, length, MPI_CHAR, &status); + outputString = ""; + } + + } + + m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine(); + if(outputString != ""){ //output to file + //send results to parent + int length = outputString.length(); + char buf[length]; + strcpy(buf, outputString.c_str()); + + MPI_File_write_shared(outMPI, buf, length, MPI_CHAR, &status); + outputString = ""; + } + + return 1; + } + catch(exception& e) { + m->errorOut(e, "DistanceCommand", "driverMPI"); + exit(1); + } +} +#endif +/**************************************************************************************************/ +int DistanceCommand::convertMatrix(string outputFile) { try{ //sort file by first column so the distances for each row are together @@ -353,6 +493,8 @@ void DistanceCommand::convertMatrix(string outputFile) { //openInputFile(outfile, in); while(!in.eof()) { + if (m->control_pressed) { in.close(); remove(outfile.c_str()); out.close(); return 0; } + in >> first >> second >> dist; gobble(in); if (first != currentRow) { @@ -388,9 +530,105 @@ void DistanceCommand::convertMatrix(string outputFile) { remove(outfile.c_str()); + return 1; + + } + catch(exception& e) { + m->errorOut(e, "DistanceCommand", "convertMatrix"); + exit(1); + } +} +/**************************************************************************************************/ +int DistanceCommand::convertToLowerTriangle(string outputFile) { + try{ + + //sort file by first column so the distances for each row are together + string outfile = getRootName(outputFile) + "sorted.dist.temp"; + + //use the unix sort + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + string command = "sort -n " + outputFile + " -o " + outfile; + system(command.c_str()); + #else //sort using windows sort + string command = "sort " + outputFile + " /O " + outfile; + system(command.c_str()); + #endif + + + //output to new file distance for each row and save positions in file where new row begins + ifstream in; + openInputFile(outfile, in); + + ofstream out; + openOutputFile(outputFile, out); + + out.setf(ios::fixed, ios::floatfield); out.setf(ios::showpoint); + + out << alignDB.getNumSeqs() << endl; + + //get first currentRow + string first, currentRow, second; + float dist; + int i, j; + i = 0; j = 0; + map rowDists; //take advantage of the fact that maps are already sorted by key + map::iterator it; + + in >> first; + currentRow = first; + + rowDists[first] = 0.00; //distance to yourself is 0.0 + + in.seekg(0); + //openInputFile(outfile, in); + + while(!in.eof()) { + if (m->control_pressed) { in.close(); remove(outfile.c_str()); out.close(); return 0; } + + in >> first >> second >> dist; gobble(in); + + if (first != currentRow) { + //print out last row + out << currentRow << '\t'; //print name + + //print dists + for (it = rowDists.begin(); it != rowDists.end(); it++) { + if (j >= i) { break; } + out << it->second << '\t'; + j++; + } + out << endl; + + //start new row + currentRow = first; + rowDists.clear(); + rowDists[first] = 0.00; + rowDists[second] = dist; + j = 0; + i++; + }else{ + rowDists[second] = dist; + } + } + //print out last row + out << currentRow << '\t'; //print name + + //print dists + for (it = rowDists.begin(); it != rowDists.end(); it++) { + out << it->second << '\t'; + } + out << endl; + + in.close(); + out.close(); + + remove(outfile.c_str()); + + return 1; + } catch(exception& e) { - errorOut(e, "DistanceCommand", "convertMatrix"); + m->errorOut(e, "DistanceCommand", "convertToLowerTriangle"); exit(1); } } @@ -413,7 +651,7 @@ void DistanceCommand::appendFiles(string temp, string filename) { output.close(); } catch(exception& e) { - errorOut(e, "DistanceCommand", "appendFiles"); + m->errorOut(e, "DistanceCommand", "appendFiles"); exit(1); } }