X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=distancecommand.cpp;h=4720df3cf0d968a84ba49c4af449b226a304c1cd;hb=aba5f8811829037b0a3004ef33f0ad4ed5e5fcf8;hp=0c9d011503370886d2f802d2903c72f985cbc40e;hpb=02909d6cae9963ba00dc746969a370fa8ca934fc;p=mothur.git diff --git a/distancecommand.cpp b/distancecommand.cpp index 0c9d011..4720df3 100644 --- a/distancecommand.cpp +++ b/distancecommand.cpp @@ -16,32 +16,47 @@ //********************************************************************************************************************** -DistanceCommand::DistanceCommand(string option){ +DistanceCommand::DistanceCommand(string option) { try { abort = false; Estimators.clear(); - + //allow user to run help if(option == "help") { help(); abort = true; } else { //valid paramters for this command - string Array[] = {"fasta", "phylip", "calc", "countends", "cutoff", "processors"}; + string Array[] = {"fasta", "output", "calc", "countends", "cutoff", "processors", "outputdir","inputdir"}; vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); OptionParser parser(option); map parameters = parser.getParameters(); ValidParameters validParameter; + map::iterator it2; //check to make sure all parameters are valid for command - for (map::iterator it2 = parameters.begin(); it2 != parameters.end(); it2++) { + for (it2 = parameters.begin(); it2 != parameters.end(); it2++) { if (validParameter.isValidParameter(it2->first, myArray, it2->second) != true) { abort = true; } } + //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 = ""; } + else { + string path; + it2 = parameters.find("fasta"); + //user has given a template file + if(it2 != parameters.end()){ + path = hasPath(it2->second); + //if the user has not given a path then, add inputdir. else leave path alone. + if (path == "") { parameters["fasta"] = inputDir + it2->second; } + } + } + //check for required parameters fastafile = validParameter.validFile(parameters, "fasta", true); - if (fastafile == "not found") { cout << "fasta is a required parameter for the dist.seqs command." << endl; 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; @@ -49,6 +64,12 @@ DistanceCommand::DistanceCommand(string option){ alignDB = SequenceDB(inFASTA); inFASTA.close(); } + + //if the user changes the output directory command factory will send this info to us in the output parameter + outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){ + outputDir = ""; + outputDir += hasPath(fastafile); //if user entered a file with a path then preserve it + } //check for optional parameter and set defaults // ...at some point should added some additional type checking... @@ -69,8 +90,9 @@ DistanceCommand::DistanceCommand(string option){ temp = validParameter.validFile(parameters, "processors", false); if(temp == "not found"){ temp = "1"; } convert(temp, processors); - phylip = validParameter.validFile(parameters, "phylip", false); if(phylip == "not found"){ phylip = "F"; } - + output = validParameter.validFile(parameters, "output", false); if(output == "not found"){ 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; @@ -96,13 +118,9 @@ DistanceCommand::DistanceCommand(string option){ } catch(exception& e) { - cout << "Standard Error: " << e.what() << " has occurred in the DistanceCommand class Function DistanceCommand. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n"; + m->errorOut(e, "DistanceCommand", "DistanceCommand"); exit(1); } - catch(...) { - cout << "An unknown error has occurred in the DistanceCommand class function DistanceCommand. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n"; - exit(1); - } } //********************************************************************************************************************** @@ -119,29 +137,24 @@ DistanceCommand::~DistanceCommand(){ void DistanceCommand::help(){ try { - cout << "The dist.seqs command reads a file containing sequences and creates a distance file." << "\n"; - cout << "The dist.seqs command parameters are fasta, calc, countends, cutoff and processors. " << "\n"; - cout << "The fasta parameter is required." << "\n"; - cout << "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"; - cout << "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"; - cout << "The cutoff parameter allows you to specify maximum distance to keep. The default is 1.0." << "\n"; - cout << "The processors parameter allows you to specify number of processors to use. The default is 1." << "\n"; - cout << "The dist.seqs command should be in the following format: " << "\n"; - cout << "dist.seqs(fasta=yourFastaFile, calc=yourCalc, countends=yourEnds, cutoff= yourCutOff, processors=yourProcessors) " << "\n"; - cout << "Example dist.seqs(fasta=amazon.fasta, calc=eachgap, countends=F, cutoff= 2.0, processors=3)." << "\n"; - cout << "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) { - cout << "Standard Error: " << e.what() << " has occurred in the DistanceCommand class Function help. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n"; + m->errorOut(e, "DistanceCommand", "help"); exit(1); } - catch(...) { - cout << "An unknown error has occurred in the DistanceCommand class function help. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n"; - exit(1); - } } - - //********************************************************************************************************************** int DistanceCommand::execute(){ @@ -153,25 +166,73 @@ int DistanceCommand::execute(){ cutoff += 0.005; string outputFile; - - //doses the user want the phylip formatted file as well - if (isTrue(phylip) == true) { - outputFile = getRootName(fastafile) + "phylip.dist"; + + if (output == "lt") { //does the user want lower triangle phylip formatted file + outputFile = outputDir + getRootName(getSimpleName(fastafile)) + "phylip.dist"; remove(outputFile.c_str()); //output numSeqs to phylip formatted dist file - }else { //user wants column format - outputFile = getRootName(fastafile) + "dist"; + }else if (output == "column") { //user wants column format + outputFile = outputDir + getRootName(getSimpleName(fastafile)) + "dist"; + remove(outputFile.c_str()); + }else { //assume square + outputFile = outputDir + getRootName(getSimpleName(fastafile)) + "square.dist"; remove(outputFile.c_str()); } + + +#ifdef USE_MPI + + int pid, start, end; + int tag = 2001; - //# if defined (_WIN32) - //figure out how to implement the fork and wait commands in windows - // driver(distCalculator, seqDB, 0, numSeqs, distFile, phylipFile, cutoff); - //# endif + 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 defined (__APPLE__) || (__MACH__) || (linux) || (__linux) //if you don't need to fork anything if(processors == 1){ driver(0, numSeqs, outputFile, cutoff); @@ -195,24 +256,42 @@ 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) { - cout << "Standard Error: " << e.what() << " has occurred in the DistanceCommand class Function execute. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n"; + m->errorOut(e, "DistanceCommand", "execute"); exit(1); } - catch(...) { - cout << "An unknown error has occurred in the DistanceCommand class function execute. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n"; - exit(1); - } } /**************************************************************************************************/ void DistanceCommand::createProcesses(string filename) { @@ -231,7 +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 { cout << "unable to spawn the necessary processes." << endl; 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 @@ -242,13 +321,9 @@ void DistanceCommand::createProcesses(string filename) { #endif } 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"; + m->errorOut(e, "DistanceCommand", "createProcesses"); 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); - } } /**************************************************************************************************/ @@ -263,45 +338,301 @@ int DistanceCommand::driver(int startLine, int endLine, string dFileName, float outFile.setf(ios::fixed, ios::showpoint); outFile << setprecision(4); - if(isTrue(phylip) && startLine == 0){ outFile << alignDB.getNumSeqs() << endl; } + if((output == "lt") && startLine == 0){ outFile << alignDB.getNumSeqs() << endl; } + for(int i=startLine;icontrol_pressed) { outFile.close(); return 0; } + distCalculator->calcDist(alignDB.get(i), alignDB.get(j)); double dist = distCalculator->getDist(); if(dist <= cutoff){ - if (!isTrue(phylip)) { outFile << alignDB.get(i).getName() << ' ' << alignDB.get(j).getName() << ' ' << dist << endl; } + if (output == "column") { outFile << alignDB.get(i).getName() << ' ' << alignDB.get(j).getName() << ' ' << dist << endl; } } - if (isTrue(phylip)) { outFile << dist << '\t'; } + if (output == "lt") { outFile << dist << '\t'; } + if (output == "square") { //make a square column you can convert to square phylip + outFile << alignDB.get(i).getName() << '\t' << alignDB.get(j).getName() << '\t' << dist << endl; + outFile << alignDB.get(j).getName() << '\t' << alignDB.get(i).getName() << '\t' << dist << endl; + } + } - if (isTrue(phylip) == true) { outFile << endl; } + if (output == "lt") { outFile << endl; } if(i % 100 == 0){ - cout << i << '\t' << time(NULL) - startTime << endl; + m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine(); } } - cout << endLine-1 << '\t' << time(NULL) - startTime << endl; + m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine(); outFile.close(); return 1; } catch(exception& e) { - cout << "Standard Error: " << e.what() << " has occurred in the DistanceCommand class Function driver. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n"; + m->errorOut(e, "DistanceCommand", "driver"); exit(1); } - catch(...) { - cout << "An unknown error has occurred in the DistanceCommand class function driver. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n"; - exit(1); - } +} +#ifdef USE_MPI +/**************************************************************************************************/ +/////// 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 + 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; + 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++) { + out << it->second << '\t'; + } + out << endl; + + //start new row + currentRow = first; + rowDists.clear(); + rowDists[first] = 0.00; + rowDists[second] = dist; + }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) { + 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) { + m->errorOut(e, "DistanceCommand", "convertToLowerTriangle"); + exit(1); + } +} +/************************************************************************************************** void DistanceCommand::appendFiles(string temp, string filename) { try{ ofstream output; @@ -309,29 +640,21 @@ void DistanceCommand::appendFiles(string temp, string filename) { //open output file in append mode openOutputFileAppend(filename, output); - - //open temp file for reading openInputFile(temp, input); - string line; - //read input file and write to output file - while(input.eof() != true) { - getline(input, line); //getline removes the newline char - if (line != "") { - output << line << endl; // Appending back newline char - } - } + while(char c = input.get()){ + if(input.eof()) { break; } + else { output << c; } + } input.close(); output.close(); } catch(exception& e) { - cout << "Standard Error: " << e.what() << " has occurred in the DistanceCommand class Function appendFiles. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n"; + m->errorOut(e, "DistanceCommand", "appendFiles"); exit(1); } - catch(...) { - cout << "An unknown error has occurred in the DistanceCommand class function appendFiles. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n"; - exit(1); - } } -/**************************************************************************************************/ \ No newline at end of file +/**************************************************************************************************/ + +