X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=distancecommand.cpp;h=9243df3873a9687d43da9da940b32da1ce9a5091;hp=23f2834648bb905ff2139c4bc05de59760e4729b;hb=cf9987b67aa49777a4c91c2d21f96e58bf17aa82;hpb=30c1fd8c45b6f0d66c17f2714dbb58b8ddccdce2 diff --git a/distancecommand.cpp b/distancecommand.cpp index 23f2834..9243df3 100644 --- a/distancecommand.cpp +++ b/distancecommand.cpp @@ -8,31 +8,103 @@ */ #include "distancecommand.h" -#include "ignoregaps.h" -#include "eachgapdist.h" -#include "eachgapignore.h" -#include "onegapdist.h" -#include "onegapignore.h" //********************************************************************************************************************** - +vector DistanceCommand::setParameters(){ + try { + CommandParameter pcolumn("column", "InputTypes", "", "", "none", "none", "OldFastaColumn","column",false,false); parameters.push_back(pcolumn); + CommandParameter poldfasta("oldfasta", "InputTypes", "", "", "none", "none", "OldFastaColumn","",false,false); parameters.push_back(poldfasta); + CommandParameter pfasta("fasta", "InputTypes", "", "", "none", "none", "none","phylip-column",false,true, true); parameters.push_back(pfasta); + CommandParameter poutput("output", "Multiple", "column-lt-square-phylip", "column", "", "", "","phylip-column",false,false, true); parameters.push_back(poutput); + CommandParameter pcalc("calc", "Multiple", "nogaps-eachgap-onegap", "onegap", "", "", "","",false,false); parameters.push_back(pcalc); + CommandParameter pcountends("countends", "Boolean", "", "T", "", "", "","",false,false); parameters.push_back(pcountends); + CommandParameter pcompress("compress", "Boolean", "", "F", "", "", "","",false,false); parameters.push_back(pcompress); + CommandParameter pprocessors("processors", "Number", "", "1", "", "", "","",false,false, true); parameters.push_back(pprocessors); + CommandParameter pcutoff("cutoff", "Number", "", "1.0", "", "", "","",false,false, true); parameters.push_back(pcutoff); + 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, "DistanceCommand", "setParameters"); + exit(1); + } +} +//********************************************************************************************************************** +string DistanceCommand::getHelpString(){ + try { + string helpString = ""; + helpString += "The dist.seqs command reads a file containing sequences and creates a distance file.\n"; + helpString += "The dist.seqs command parameters are fasta, oldfasta, column, calc, countends, output, compress, cutoff and processors. \n"; + helpString += "The fasta parameter is required, unless you have a valid current fasta file.\n"; + helpString += "The oldfasta and column parameters allow you to append the distances calculated to the column file.\n"; + helpString += "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"; + helpString += "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"; + helpString += "The cutoff parameter allows you to specify maximum distance to keep. The default is 1.0.\n"; + helpString += "The output parameter allows you to specify format of your distance matrix. Options are column, lt, and square. The default is column.\n"; + helpString += "The processors parameter allows you to specify number of processors to use. The default is 1.\n"; + helpString += "The compress parameter allows you to indicate that you want the resulting distance file compressed. The default is false.\n"; + helpString += "The dist.seqs command should be in the following format: \n"; + helpString += "dist.seqs(fasta=yourFastaFile, calc=yourCalc, countends=yourEnds, cutoff= yourCutOff, processors=yourProcessors) \n"; + helpString += "Example dist.seqs(fasta=amazon.fasta, calc=eachgap, countends=F, cutoff= 2.0, processors=3).\n"; + helpString += "Note: No spaces between parameter labels (i.e. calc), '=' and parameters (i.e.yourCalc).\n"; + return helpString; + } + catch(exception& e) { + m->errorOut(e, "DistanceCommand", "getHelpString"); + exit(1); + } +} +//********************************************************************************************************************** +string DistanceCommand::getOutputPattern(string type) { + try { + string pattern = ""; + + if (type == "phylip") { pattern = "[filename],[outputtag],dist"; } + else if (type == "column") { pattern = "[filename],dist"; } + else { m->mothurOut("[ERROR]: No definition for type " + type + " output pattern.\n"); m->control_pressed = true; } + + return pattern; + } + catch(exception& e) { + m->errorOut(e, "DistanceCommand", "getOutputPattern"); + exit(1); + } +} +//********************************************************************************************************************** +DistanceCommand::DistanceCommand(){ + try { + abort = true; calledHelp = true; + setParameters(); + vector tempOutNames; + outputTypes["phylip"] = tempOutNames; + outputTypes["column"] = tempOutNames; + } + catch(exception& e) { + m->errorOut(e, "DistanceCommand", "DistanceCommand"); + exit(1); + } +} +//********************************************************************************************************************** DistanceCommand::DistanceCommand(string option) { try { - abort = false; + abort = false; calledHelp = false; Estimators.clear(); //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", "output", "calc", "countends", "cutoff", "processors", "outputdir","inputdir"}; - vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); + vector myArray = setParameters(); OptionParser parser(option); map parameters = parser.getParameters(); - ValidParameters validParameter; + ValidParameters validParameter("dist.seqs"); map::iterator it2; //check to make sure all parameters are valid for command @@ -40,6 +112,11 @@ DistanceCommand::DistanceCommand(string option) { if (validParameter.isValidParameter(it2->first, myArray, it2->second) != true) { abort = true; } } + //initialize outputTypes + vector tempOutNames; + outputTypes["phylip"] = tempOutNames; + outputTypes["column"] = 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 = ""; } @@ -48,27 +125,60 @@ DistanceCommand::DistanceCommand(string option) { it2 = parameters.find("fasta"); //user has given a template file if(it2 != parameters.end()){ - path = hasPath(it2->second); + path = m->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; } } + + it2 = parameters.find("oldfasta"); + //user has given a template file + if(it2 != parameters.end()){ + path = m->hasPath(it2->second); + //if the user has not given a path then, add inputdir. else leave path alone. + if (path == "") { parameters["oldfasta"] = inputDir + it2->second; } + } + + it2 = parameters.find("column"); + //user has given a template file + if(it2 != parameters.end()){ + path = m->hasPath(it2->second); + //if the user has not given a path then, add inputdir. else leave path alone. + if (path == "") { parameters["column"] = inputDir + it2->second; } + } } //check for required parameters fastafile = validParameter.validFile(parameters, "fasta", 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; } + if (fastafile == "not found") { + fastafile = m->getFastaFile(); + if (fastafile != "") { m->mothurOut("Using " + fastafile + " as input file for the fasta parameter."); m->mothurOutEndLine(); + ifstream inFASTA; + m->openInputFile(fastafile, inFASTA); + alignDB = SequenceDB(inFASTA); + inFASTA.close(); + }else { m->mothurOut("You have no current fastafile and the fasta parameter is required."); m->mothurOutEndLine(); abort = true; } + }else if (fastafile == "not open") { abort = true; } else{ ifstream inFASTA; - openInputFile(fastafile, inFASTA); + m->openInputFile(fastafile, inFASTA); alignDB = SequenceDB(inFASTA); inFASTA.close(); + m->setFastaFile(fastafile); } + oldfastafile = validParameter.validFile(parameters, "oldfasta", true); + if (oldfastafile == "not found") { oldfastafile = ""; } + else if (oldfastafile == "not open") { abort = true; } + + column = validParameter.validFile(parameters, "column", true); + if (column == "not found") { column = ""; } + else if (column == "not open") { abort = true; } + else { m->setColumnFile(column); } + //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 + outputDir += m->hasPath(fastafile); //if user entered a file with a path then preserve it } //check for optional parameter and set defaults @@ -78,41 +188,30 @@ DistanceCommand::DistanceCommand(string option) { else { if (calc == "default") { calc = "onegap"; } } - splitAtDash(calc, Estimators); + m->splitAtDash(calc, Estimators); string temp; temp = validParameter.validFile(parameters, "countends", false); if(temp == "not found"){ temp = "T"; } convert(temp, countends); temp = validParameter.validFile(parameters, "cutoff", false); if(temp == "not found"){ temp = "1.0"; } - convert(temp, cutoff); + m->mothurConvert(temp, cutoff); - 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); + temp = validParameter.validFile(parameters, "compress", false); if(temp == "not found"){ temp = "F"; } + convert(temp, compress); + output = validParameter.validFile(parameters, "output", false); if(output == "not found"){ output = "column"; } + if (output == "phylip") { output = "lt"; } - 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"; } + if (((column != "") && (oldfastafile == "")) || ((column == "") && (oldfastafile != ""))) { m->mothurOut("If you provide column or oldfasta, you must provide both."); m->mothurOutEndLine(); abort=true; } - ValidCalculators validCalculator; + if ((column != "") && (oldfastafile != "") && (output != "column")) { m->mothurOut("You have provided column and oldfasta, indicating you want to append distances to your column file. Your output must be in column format to do so."); m->mothurOutEndLine(); abort=true; } - if (isTrue(countends) == true) { - for (int i=0; imothurOut(output + " is not a valid output form. Options are column, lt and square. I will use column."); m->mothurOutEndLine(); output = "column"; } } @@ -122,62 +221,54 @@ DistanceCommand::DistanceCommand(string option) { exit(1); } } - -//********************************************************************************************************************** - -DistanceCommand::~DistanceCommand(){ - - for(int i=0;imothurOut("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) { - m->errorOut(e, "DistanceCommand", "help"); - exit(1); - } -} //********************************************************************************************************************** int DistanceCommand::execute(){ try { - if (abort == true) { return 0; } + if (abort == true) { if (calledHelp) { return 0; } return 2; } + + int startTime = time(NULL); + + //save number of new sequence + numNewFasta = alignDB.getNumSeqs(); + + //sanity check the oldfasta and column file as well as add oldfasta sequences to alignDB + if ((oldfastafile != "") && (column != "")) { if (!(sanityCheck())) { return 0; } } + + if (m->control_pressed) { return 0; } int numSeqs = alignDB.getNumSeqs(); cutoff += 0.005; + if (!alignDB.sameLength()) { m->mothurOut("[ERROR]: your sequences are not the same length, aborting."); m->mothurOutEndLine(); return 0; } + string outputFile; - + + map variables; + variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(fastafile)); if (output == "lt") { //does the user want lower triangle phylip formatted file - outputFile = outputDir + getRootName(getSimpleName(fastafile)) + "phylip.dist"; - remove(outputFile.c_str()); + variables["[outputtag]"] = "phylip"; + outputFile = getOutputFileName("phylip", variables); + m->mothurRemove(outputFile); outputTypes["phylip"].push_back(outputFile); //output numSeqs to phylip formatted dist file }else if (output == "column") { //user wants column format - outputFile = outputDir + getRootName(getSimpleName(fastafile)) + "dist"; - remove(outputFile.c_str()); + outputFile = getOutputFileName("column", variables); + outputTypes["column"].push_back(outputFile); + + //so we don't accidentally overwrite + if (outputFile == column) { + string tempcolumn = column + ".old"; + rename(column.c_str(), tempcolumn.c_str()); + } + + m->mothurRemove(outputFile); }else { //assume square - outputFile = outputDir + getRootName(getSimpleName(fastafile)) + "square.dist"; - remove(outputFile.c_str()); + variables["[outputtag]"] = "square"; + outputFile = getOutputFileName("phylip", variables); + m->mothurRemove(outputFile); + outputTypes["phylip"].push_back(outputFile); } @@ -191,82 +282,170 @@ int DistanceCommand::execute(){ 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); + if (output != "square") { + start = int (sqrt(float(pid)/float(processors)) * numSeqs); + end = int (sqrt(float(pid+1)/float(processors)) * numSeqs); + }else{ + start = int ((float(pid)/float(processors)) * numSeqs); + end = int ((float(pid+1)/float(processors)) * numSeqs); + } - if (pid == 0) { //you are the root process - //do your part - string outputMyPart; - driverMPI(start, end, outputMyPart, cutoff); - - ofstream out; - openOutputFile(outputFile, out); + if (output == "column") { + MPI_File outMPI; + int amode=MPI_MODE_CREATE|MPI_MODE_WRONLY; + + //char* filename = new char[outputFile.length()]; + //memcpy(filename, outputFile.c_str(), outputFile.length()); - out << outputMyPart; + char filename[1024]; + strcpy(filename, outputFile.c_str()); - //get the childrens parts - for(int i = 1; i < processors; i++) { - int length; - MPI_Recv(&length, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status); + MPI_File_open(MPI_COMM_WORLD, filename, amode, MPI_INFO_NULL, &outMPI); + //delete filename; + + if (pid == 0) { //you are the root process + + //do your part + string outputMyPart; + + driverMPI(start, end, outMPI, cutoff); - char buf[length]; + if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&outMPI); return 0; } + + //wait on chidren + for(int i = 1; i < processors; i++) { + if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&outMPI); return 0; } - MPI_Recv(buf, length, 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 + //do your part + driverMPI(start, end, outMPI, cutoff); - outputMyPart = buf; - out << outputMyPart; + if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&outMPI); return 0; } + + char buf[5]; + strcpy(buf, "done"); + //tell parent you are done. + MPI_Send(buf, 5, MPI_CHAR, 0, tag, MPI_COMM_WORLD); } - out.close(); + MPI_File_close(&outMPI); - }else { //you are a child process - //do your part - string outputMyPart; - driverMPI(start, end, outputMyPart, cutoff); - - //send results to parent - int length = outputMyPart.length(); - char buf[length]; - strcpy(buf, outputMyPart.c_str()); + }else { //lower triangle format + if (pid == 0) { //you are the root process - MPI_Send( &length, 1, MPI_INT, 0, tag, MPI_COMM_WORLD); - MPI_Send(buf, length, MPI_CHAR, 0, tag, MPI_COMM_WORLD); - } - + //do your part + string outputMyPart; + unsigned long long mySize; + + if (output != "square"){ driverMPI(start, end, outputFile, mySize); } + else { driverMPI(start, end, outputFile, mySize, output); } + + if (m->control_pressed) { outputTypes.clear(); return 0; } + + int amode=MPI_MODE_APPEND|MPI_MODE_WRONLY|MPI_MODE_CREATE; // + MPI_File outMPI; + MPI_File inMPI; + + //char* filename = new char[outputFile.length()]; + //memcpy(filename, outputFile.c_str(), outputFile.length()); + + char filename[1024]; + strcpy(filename, outputFile.c_str()); + + MPI_File_open(MPI_COMM_SELF, filename, amode, MPI_INFO_NULL, &outMPI); + //delete filename; + + //wait on chidren + for(int b = 1; b < processors; b++) { + unsigned long long fileSize; + + if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&outMPI); return 0; } + + MPI_Recv(&fileSize, 1, MPI_LONG, b, tag, MPI_COMM_WORLD, &status); + + string outTemp = outputFile + toString(b) + ".temp"; + char* buf = new char[outTemp.length()]; + memcpy(buf, outTemp.c_str(), outTemp.length()); + + MPI_File_open(MPI_COMM_SELF, buf, MPI_MODE_DELETE_ON_CLOSE|MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI); + delete buf; + + int count = 0; + while (count < fileSize) { + char buf2[1]; + MPI_File_read(inMPI, buf2, 1, MPI_CHAR, &status); + MPI_File_write(outMPI, buf2, 1, MPI_CHAR, &status); + count += 1; + } + + MPI_File_close(&inMPI); //deleted on close + } + + MPI_File_close(&outMPI); + }else { //you are a child process + //do your part + unsigned long long size; + if (output != "square"){ driverMPI(start, end, (outputFile + toString(pid) + ".temp"), size); } + else { driverMPI(start, end, (outputFile + toString(pid) + ".temp"), size, output); } + + if (m->control_pressed) { return 0; } + + //tell parent you are done. + MPI_Send(&size, 1, MPI_LONG, 0, tag, MPI_COMM_WORLD); + } + } + MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case #else - #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + //#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) //if you don't need to fork anything if(processors == 1){ - driver(0, numSeqs, outputFile, cutoff); + if (output != "square") { driver(0, numSeqs, outputFile, cutoff); } + else { driver(0, numSeqs, outputFile, "square"); } }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); + unsigned long long numDists = 0; + + if (output == "square") { + numDists = numSeqs * numSeqs; + }else { + for(int i=0;i processors) { break; } + } + } } - - createProcesses(outputFile); - - map::iterator it = processIDS.begin(); - rename((outputFile + toString(it->second) + ".temp").c_str(), outputFile.c_str()); - it++; - //append and remove temp files - for (; it != processIDS.end(); it++) { - appendFiles((outputFile + toString(it->second) + ".temp"), outputFile); - remove((outputFile + toString(it->second) + ".temp").c_str()); + if (numDists < processors) { processors = numDists; } + + for (int i = 0; i < processors; i++) { + distlinePair tempLine; + lines.push_back(tempLine); + if (output != "square") { + lines[i].start = int (sqrt(float(i)/float(processors)) * numSeqs); + lines[i].end = int (sqrt(float(i+1)/float(processors)) * numSeqs); + }else{ + lines[i].start = int ((float(i)/float(processors)) * numSeqs); + lines[i].end = int ((float(i+1)/float(processors)) * numSeqs); + } + } + + createProcesses(outputFile); } - #else - ifstream inFASTA; - driver(0, numSeqs, outputFile, cutoff); - #endif + //#else + //ifstream inFASTA; + //if (output != "square") { driver(0, numSeqs, outputFile, cutoff); } + //else { driver(0, numSeqs, outputFile, "square"); } + //#endif #endif - if (m->control_pressed) { delete distCalculator; remove(outputFile.c_str()); return 0; } + if (m->control_pressed) { outputTypes.clear(); m->mothurRemove(outputFile); return 0; } #ifdef USE_MPI MPI_Comm_rank(MPI_COMM_WORLD, &pid); @@ -274,20 +453,69 @@ int DistanceCommand::execute(){ if (pid == 0) { //only one process should output to screen #endif - if (output == "square") { convertMatrix(outputFile); } + //if (output == "square") { convertMatrix(outputFile); } + + ifstream fileHandle; + fileHandle.open(outputFile.c_str()); + if(fileHandle) { + m->gobble(fileHandle); + if (fileHandle.eof()) { m->mothurOut(outputFile + " is blank. This can result if there are no distances below your cutoff."); m->mothurOutEndLine(); } + } + + //append the old column file to the new one + if ((oldfastafile != "") && (column != "")) { + //we had to rename the column file so we didnt overwrite above, but we want to keep old name + if (outputFile == column) { + string tempcolumn = column + ".old"; + m->appendFiles(tempcolumn, outputFile); + m->mothurRemove(tempcolumn); + }else{ + m->appendFiles(outputFile, column); + m->mothurRemove(outputFile); + outputFile = column; + } + + if (outputDir != "") { + string newOutputName = outputDir + m->getSimpleName(outputFile); + rename(outputFile.c_str(), newOutputName.c_str()); + m->mothurRemove(outputFile); + outputFile = newOutputName; + } + } + #ifdef USE_MPI } #endif - if (m->control_pressed) { delete distCalculator; remove(outputFile.c_str()); return 0; } + if (m->control_pressed) { outputTypes.clear(); m->mothurRemove(outputFile); return 0; } - delete distCalculator; + //set phylip file as new current phylipfile + string current = ""; + itTypes = outputTypes.find("phylip"); + if (itTypes != outputTypes.end()) { + if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setPhylipFile(current); } + } + + //set column file as new current columnfile + itTypes = outputTypes.find("column"); + if (itTypes != outputTypes.end()) { + if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setColumnFile(current); } + } m->mothurOutEndLine(); - m->mothurOut("Output File Name: "); m->mothurOutEndLine(); + m->mothurOut("Output File Names: "); m->mothurOutEndLine(); m->mothurOut(outputFile); m->mothurOutEndLine(); m->mothurOutEndLine(); + m->mothurOut("It took " + toString(time(NULL) - startTime) + " to calculate the distances for " + toString(numSeqs) + " sequences."); m->mothurOutEndLine(); + + + if (m->isTrue(compress)) { + m->mothurOut("Compressing..."); m->mothurOutEndLine(); + m->mothurOut("(Replacing " + outputFile + " with " + outputFile + ".gz)"); m->mothurOutEndLine(); + system(("gzip -v " + outputFile).c_str()); + outputNames.push_back(outputFile + ".gz"); + }else { outputNames.push_back(outputFile); } return 0; @@ -300,8 +528,8 @@ int DistanceCommand::execute(){ /**************************************************************************************************/ void DistanceCommand::createProcesses(string filename) { try { -#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - int process = 0; +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + int process = 1; processIDS.clear(); //loop through and create all the processes you want @@ -309,32 +537,107 @@ void DistanceCommand::createProcesses(string filename) { 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 + processIDS.push_back(pid); //create map from line number to pid so you can append files in correct order later process++; }else if (pid == 0){ - driver(lines[process]->start, lines[process]->end, filename + toString(getpid()) + ".temp", cutoff); + if (output != "square") { driver(lines[process].start, lines[process].end, filename + toString(getpid()) + ".temp", cutoff); } + else { driver(lines[process].start, lines[process].end, filename + toString(getpid()) + ".temp", "square"); } + exit(0); + }else { + m->mothurOut("[ERROR]: unable to spawn the necessary processes. Error code: " + toString(pid)); m->mothurOutEndLine(); + perror(" : "); + for (int i=0;imothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); } + } } - + + //parent does its part + if (output != "square") { driver(lines[0].start, lines[0].end, filename, cutoff); } + else { driver(lines[0].start, lines[0].end, filename, "square"); } + + //force parent to wait until all the processes are done - for (map::iterator it = processIDS.begin(); it != processIDS.end(); it++) { - int temp = it->second; + for (int i=0;i pDataArray; //[processors-1]; + DWORD dwThreadIdArray[processors-1]; + HANDLE hThreadArray[processors-1]; + + //Create processor-1 worker threads. + for( int i=0; icount != (pDataArray[i]->endLine-pDataArray[i]->startLine)) { + m->mothurOut("[ERROR]: process " + toString(i) + " only processed " + toString(pDataArray[i]->count) + " of " + toString(pDataArray[i]->endLine-pDataArray[i]->startLine) + " sequences assigned to it, quitting. \n"); m->control_pressed = true; + } + CloseHandle(hThreadArray[i]); + delete pDataArray[i]; + } #endif + + //append and remove temp files + for (int i=0;iappendFiles((filename + toString(processIDS[i]) + ".temp"), filename); + m->mothurRemove((filename + toString(processIDS[i]) + ".temp")); + } + } catch(exception& e) { m->errorOut(e, "DistanceCommand", "createProcesses"); exit(1); } } - /**************************************************************************************************/ /////// need to fix to work with calcs and sequencedb int DistanceCommand::driver(int startLine, int endLine, string dFileName, float cutoff){ try { - + ValidCalculators validCalculator; + Dist* distCalculator; + if (m->isTrue(countends) == true) { + for (int i=0; icontrol_pressed) { outFile.close(); return 0; } + if (m->control_pressed) { delete distCalculator; outFile.close(); return 0; } + + //if there was a column file given and we are appending, we don't want to calculate the distances that are already in the column file + //the alignDB contains the new sequences and then the old, so if i an oldsequence and j is an old sequence then break out of this loop + if ((i >= numNewFasta) && (j >= numNewFasta)) { break; } distCalculator->calcDist(alignDB.get(i), alignDB.get(j)); double dist = distCalculator->getDist(); @@ -363,24 +670,19 @@ int DistanceCommand::driver(int startLine, int endLine, string dFileName, float if (output == "column") { outFile << alignDB.get(i).getName() << ' ' << alignDB.get(j).getName() << ' ' << dist << endl; } } 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 (output == "lt") { outFile << endl; } if(i % 100 == 0){ - m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine(); + m->mothurOutJustToScreen(toString(i) + "\t" + toString(time(NULL) - startTime)+"\n"); } } - m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine(); + m->mothurOutJustToScreen(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)+"\n"); outFile.close(); + delete distCalculator; return 1; } @@ -391,69 +693,347 @@ int DistanceCommand::driver(int startLine, int endLine, string dFileName, float } /**************************************************************************************************/ /////// need to fix to work with calcs and sequencedb -int DistanceCommand::driverMPI(int startLine, int endLine, string& outputString, float cutoff){ +int DistanceCommand::driver(int startLine, int endLine, string dFileName, string square){ try { - + ValidCalculators validCalculator; + Dist* distCalculator; + if (m->isTrue(countends) == true) { + for (int i=0; icontrol_pressed) { delete distCalculator; outFile.close(); return 0; } + + distCalculator->calcDist(alignDB.get(i), alignDB.get(j)); + double dist = distCalculator->getDist(); + + outFile << dist << '\t'; + } + + outFile << endl; + + if(i % 100 == 0){ + m->mothurOutJustToScreen(toString(i) + "\t" + toString(time(NULL) - startTime)+"\n"); + } + + } + m->mothurOutJustToScreen(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)+"\n"); + + outFile.close(); + delete distCalculator; + + return 1; + } + catch(exception& e) { + m->errorOut(e, "DistanceCommand", "driver"); + 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 { + + ValidCalculators validCalculator; + Dist* distCalculator; + if (m->isTrue(countends) == true) { + for (int i=0; icontrol_pressed) { return 0; } + if (m->control_pressed) { delete distCalculator; return 0; } + + //if there was a column file given and we are appending, we don't want to calculate the distances that are already in the column file + //the alignDB contains the new sequences and then the old, so if i an oldsequence and j is an old sequence then break out of this loop + if ((i >= numNewFasta) && (j >= numNewFasta)) { break; } 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'); } + outputString += (alignDB.get(i).getName() + ' ' + alignDB.get(j).getName() + ' ' + toString(dist) + '\n'); } - if (output == "lt") { outputString += (toString(dist) + '\t'); } + } + + if(i % 100 == 0){ + m->mothurOutJustToScreen(toString(i) + "\t" + toString(time(NULL) - startTime)+"\n"); + } + + + //send results to parent + int length = outputString.length(); + + char* buf = new char[length]; + memcpy(buf, outputString.c_str(), length); + + MPI_File_write_shared(outMPI, buf, length, MPI_CHAR, &status); + outputString = ""; + delete buf; + + } + + m->mothurOutJustToScreen(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)+"\n"); + delete distCalculator; + return 1; + } + catch(exception& e) { + m->errorOut(e, "DistanceCommand", "driverMPI"); + exit(1); + } +} +/**************************************************************************************************/ +/////// need to fix to work with calcs and sequencedb +int DistanceCommand::driverMPI(int startLine, int endLine, string file, unsigned long long& size){ + try { + ValidCalculators validCalculator; + Dist* distCalculator; + if (m->isTrue(countends) == true) { + for (int i=0; icontrol_pressed) { delete distCalculator; return 0; } + + distCalculator->calcDist(alignDB.get(i), alignDB.get(j)); + double dist = distCalculator->getDist(); + + outputString += toString(dist) + "\t"; + } + + outputString += "\n"; + + + if(i % 100 == 0){ + m->mothurOutJustToScreen(toString(i) + "\t" + toString(time(NULL) - startTime)+"\n"); } + + + //send results to parent + int length = outputString.length(); + char* buf = new char[length]; + memcpy(buf, outputString.c_str(), length); + + MPI_File_write(outMPI, buf, length, MPI_CHAR, &status); + size += outputString.length(); + outputString = ""; + delete buf; + } + + m->mothurOutJustToScreen(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)+"\n"); + + MPI_File_close(&outMPI); + delete distCalculator; + + return 1; + } + catch(exception& e) { + m->errorOut(e, "DistanceCommand", "driverMPI"); + exit(1); + } +} +/**************************************************************************************************/ +/////// need to fix to work with calcs and sequencedb +int DistanceCommand::driverMPI(int startLine, int endLine, string file, unsigned long long& size, string square){ + try { + ValidCalculators validCalculator; + Dist* distCalculator; + if (m->isTrue(countends) == true) { + for (int i=0; icontrol_pressed) { delete distCalculator; return 0; } + + distCalculator->calcDist(alignDB.get(i), alignDB.get(j)); + double dist = distCalculator->getDist(); + + outputString += toString(dist) + "\t"; + } + outputString += "\n"; + + if(i % 100 == 0){ - m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine(); + m->mothurOutJustToScreen(toString(i) + "\t" + toString(time(NULL) - startTime)+"\n"); } + + //send results to parent + int length = outputString.length(); + char* buf = new char[length]; + memcpy(buf, outputString.c_str(), length); + + MPI_File_write(outMPI, buf, length, MPI_CHAR, &status); + size += outputString.length(); + outputString = ""; + delete buf; } - m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine(); + m->mothurOutJustToScreen(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)+"\n"); + + MPI_File_close(&outMPI); + delete distCalculator; return 1; } catch(exception& e) { - m->errorOut(e, "DistanceCommand", "driver"); + 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"; + string outfile = m->getRootName(outputFile) + "sorted.dist.temp"; //use the unix sort - #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) string command = "sort -n " + outputFile + " -o " + outfile; system(command.c_str()); #else //sort using windows sort @@ -464,10 +1044,10 @@ int DistanceCommand::convertMatrix(string outputFile) { //output to new file distance for each row and save positions in file where new row begins ifstream in; - openInputFile(outfile, in); + m->openInputFile(outfile, in); ofstream out; - openOutputFile(outputFile, out); + m->openOutputFile(outputFile, out); out.setf(ios::fixed, ios::floatfield); out.setf(ios::showpoint); @@ -485,12 +1065,12 @@ int DistanceCommand::convertMatrix(string outputFile) { rowDists[first] = 0.00; //distance to yourself is 0.0 in.seekg(0); - //openInputFile(outfile, in); + //m->openInputFile(outfile, in); while(!in.eof()) { - if (m->control_pressed) { in.close(); remove(outfile.c_str()); out.close(); return 0; } + if (m->control_pressed) { in.close(); m->mothurRemove(outfile); out.close(); return 0; } - in >> first >> second >> dist; gobble(in); + in >> first >> second >> dist; m->gobble(in); if (first != currentRow) { //print out last row @@ -523,7 +1103,7 @@ int DistanceCommand::convertMatrix(string outputFile) { in.close(); out.close(); - remove(outfile.c_str()); + m->mothurRemove(outfile); return 1; @@ -533,29 +1113,194 @@ int DistanceCommand::convertMatrix(string outputFile) { exit(1); } } -/************************************************************************************************** -void DistanceCommand::appendFiles(string temp, string filename) { +************************************************************************************************** +int DistanceCommand::convertToLowerTriangle(string outputFile) { try{ - ofstream output; - ifstream input; - - //open output file in append mode - openOutputFileAppend(filename, output); - openInputFile(temp, input); + + //sort file by first column so the distances for each row are together + string outfile = m->getRootName(outputFile) + "sorted.dist.temp"; + + //use the unix sort + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + 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; + m->openInputFile(outfile, in); + + ofstream out; + m->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; - while(char c = input.get()){ - if(input.eof()) { break; } - else { output << c; } + in >> first; + currentRow = first; + + rowDists[first] = 0.00; //distance to yourself is 0.0 + + in.seekg(0); + //m->openInputFile(outfile, in); + + while(!in.eof()) { + if (m->control_pressed) { in.close(); m->mothurRemove(outfile); out.close(); return 0; } + + in >> first >> second >> dist; m->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(); + + m->mothurRemove(outfile); + + return 1; + + } + catch(exception& e) { + m->errorOut(e, "DistanceCommand", "convertToLowerTriangle"); + exit(1); + } +} +**************************************************************************************************/ +//its okay if the column file does not contain all the names in the fasta file, since some distance may have been above a cutoff, +//but no sequences can be in the column file that are not in oldfasta. also, if a distance is above the cutoff given then remove it. +//also check to make sure the 2 files have the same alignment length. +bool DistanceCommand::sanityCheck() { + try{ + bool good = true; + + //make sure the 2 fasta files have the same alignment length + ifstream in; + m->openInputFile(fastafile, in); + int fastaAlignLength = 0; + if (in) { + Sequence tempIn(in); + fastaAlignLength = tempIn.getAligned().length(); + } + in.close(); + + ifstream in2; + m->openInputFile(oldfastafile, in2); + int oldfastaAlignLength = 0; + if (in2) { + Sequence tempIn2(in2); + oldfastaAlignLength = tempIn2.getAligned().length(); + } + in2.close(); + + if (fastaAlignLength != oldfastaAlignLength) { m->mothurOut("fasta files do not have the same alignment length."); m->mothurOutEndLine(); return false; } + + //read fasta file and save names as well as adding them to the alignDB + set namesOldFasta; + + ifstream inFasta; + m->openInputFile(oldfastafile, inFasta); + + while (!inFasta.eof()) { + if (m->control_pressed) { inFasta.close(); return good; } + + Sequence temp(inFasta); + + if (temp.getName() != "") { + namesOldFasta.insert(temp.getName()); //save name + alignDB.push_back(temp); //add to DB + } + + m->gobble(inFasta); } - input.close(); - output.close(); + inFasta.close(); + + //read through the column file checking names and removing distances above the cutoff + ifstream inDist; + m->openInputFile(column, inDist); + + ofstream outDist; + string outputFile = column + ".temp"; + m->openOutputFile(outputFile, outDist); + + string name1, name2; + float dist; + while (!inDist.eof()) { + if (m->control_pressed) { inDist.close(); outDist.close(); m->mothurRemove(outputFile); return good; } + + inDist >> name1 >> name2 >> dist; m->gobble(inDist); + + //both names are in fasta file and distance is below cutoff + if ((namesOldFasta.count(name1) == 0) || (namesOldFasta.count(name2) == 0)) { good = false; break; } + else{ + if (dist <= cutoff) { + outDist << name1 << '\t' << name2 << '\t' << dist << endl; + } + } + } + + inDist.close(); + outDist.close(); + + if (good) { + m->mothurRemove(column); + rename(outputFile.c_str(), column.c_str()); + }else{ + m->mothurRemove(outputFile); //temp file is bad because file mismatch above + } + + return good; + } catch(exception& e) { - m->errorOut(e, "DistanceCommand", "appendFiles"); + m->errorOut(e, "DistanceCommand", "sanityCheck"); exit(1); } } /**************************************************************************************************/ + +