X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=clustersplitcommand.cpp;h=95693cc420a7e55d5c9a96a102e3d9e55ad0221c;hp=f9d38227f2d59709235bfe69f87385067317b1ab;hb=050a3ff02473a3d4c0980964e1a9ebe52e55d6b8;hpb=348de0f8b17d84ede77081dcf67bd6ef43496677 diff --git a/clustersplitcommand.cpp b/clustersplitcommand.cpp index f9d3822..95693cc 100644 --- a/clustersplitcommand.cpp +++ b/clustersplitcommand.cpp @@ -8,61 +8,109 @@ */ #include "clustersplitcommand.h" -#include "readcluster.h" -#include "splitmatrix.h" -#include "readphylip.h" -#include "readcolumn.h" -#include "readmatrix.hpp" -#include "inputdata.h" //********************************************************************************************************************** -vector ClusterSplitCommand::getValidParameters(){ +vector ClusterSplitCommand::setParameters(){ try { - string AlignArray[] = {"fasta","phylip","column","name","cutoff","precision","method","splitmethod","taxonomy","taxlevel","large","showabund","timing","hard","processors","outputdir","inputdir"}; - vector myArray (AlignArray, AlignArray+(sizeof(AlignArray)/sizeof(string))); + CommandParameter ptaxonomy("taxonomy", "InputTypes", "", "", "none", "none", "FastaTaxName","",false,false,true); parameters.push_back(ptaxonomy); + CommandParameter pphylip("phylip", "InputTypes", "", "", "PhylipColumnFasta", "PhylipColumnFasta", "none","list",false,false,true); parameters.push_back(pphylip); + CommandParameter pfasta("fasta", "InputTypes", "", "", "PhylipColumnFasta", "PhylipColumnFasta", "FastaTaxName","list",false,false,true); parameters.push_back(pfasta); + CommandParameter pname("name", "InputTypes", "", "", "NameCount", "none", "ColumnName-FastaTaxName","rabund-sabund",false,false,true); parameters.push_back(pname); + CommandParameter pcount("count", "InputTypes", "", "", "NameCount", "none", "","",false,false,true); parameters.push_back(pcount); + CommandParameter pcolumn("column", "InputTypes", "", "", "PhylipColumnFasta", "PhylipColumnFasta", "ColumnName","list",false,false,true); parameters.push_back(pcolumn); + CommandParameter ptaxlevel("taxlevel", "Number", "", "3", "", "", "","",false,false,true); parameters.push_back(ptaxlevel); + CommandParameter psplitmethod("splitmethod", "Multiple", "classify-fasta-distance", "distance", "", "", "","",false,false,true); parameters.push_back(psplitmethod); + CommandParameter plarge("large", "Boolean", "", "F", "", "", "","",false,false); parameters.push_back(plarge); + CommandParameter pshowabund("showabund", "Boolean", "", "T", "", "", "","",false,false); parameters.push_back(pshowabund); + CommandParameter pcluster("cluster", "Boolean", "", "T", "", "", "","",false,false); parameters.push_back(pcluster); + CommandParameter ptiming("timing", "Boolean", "", "F", "", "", "","",false,false); parameters.push_back(ptiming); + CommandParameter pprocessors("processors", "Number", "", "1", "", "", "","",false,false,true); parameters.push_back(pprocessors); + CommandParameter pcutoff("cutoff", "Number", "", "0.25", "", "", "","",false,false,true); parameters.push_back(pcutoff); + CommandParameter pprecision("precision", "Number", "", "100", "", "", "","",false,false); parameters.push_back(pprecision); + CommandParameter pmethod("method", "Multiple", "furthest-nearest-average-weighted", "average", "", "", "","",false,false); parameters.push_back(pmethod); + CommandParameter phard("hard", "Boolean", "", "T", "", "", "","",false,false); parameters.push_back(phard); + CommandParameter pclassic("classic", "Boolean", "", "F", "", "", "","",false,false); parameters.push_back(pclassic); + 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, "ClusterSplitCommand", "getValidParameters"); + m->errorOut(e, "ClusterSplitCommand", "setParameters"); exit(1); } } //********************************************************************************************************************** -ClusterSplitCommand::ClusterSplitCommand(){ +string ClusterSplitCommand::getHelpString(){ try { - abort = true; calledHelp = true; - vector tempOutNames; - outputTypes["list"] = tempOutNames; - outputTypes["rabund"] = tempOutNames; - outputTypes["sabund"] = tempOutNames; - outputTypes["column"] = tempOutNames; + string helpString = ""; + helpString += "The cluster.split command parameter options are fasta, phylip, column, name, count, cutoff, precision, method, splitmethod, taxonomy, taxlevel, showabund, timing, hard, large, cluster, processors. Fasta or Phylip or column and name are required.\n"; + helpString += "The cluster.split command can split your files in 3 ways. Splitting by distance file, by classification, or by classification also using a fasta file. \n"; + helpString += "For the distance file method, you need only provide your distance file and mothur will split the file into distinct groups. \n"; + helpString += "For the classification method, you need to provide your distance file and taxonomy file, and set the splitmethod to classify. \n"; + helpString += "You will also need to set the taxlevel you want to split by. mothur will split the sequences into distinct taxonomy groups, and split the distance file based on those groups. \n"; + helpString += "For the classification method using a fasta file, you need to provide your fasta file, names file and taxonomy file. \n"; + helpString += "You will also need to set the taxlevel you want to split by. mothur will split the sequence into distinct taxonomy groups, and create distance files for each grouping. \n"; + helpString += "The phylip and column parameter allow you to enter your distance file. \n"; + helpString += "The fasta parameter allows you to enter your aligned fasta file. \n"; + helpString += "The name parameter allows you to enter your name file. \n"; + helpString += "The count parameter allows you to enter your count file. \n A count or name file is required if your distance file is in column format"; + helpString += "The cluster parameter allows you to indicate whether you want to run the clustering or just split the distance matrix, default=t"; + helpString += "The cutoff parameter allow you to set the distance you want to cluster to, default is 0.25. \n"; + helpString += "The precision parameter allows you specify the precision of the precision of the distances outputted, default=100, meaning 2 decimal places. \n"; + helpString += "The method allows you to specify what clustering algorithm you want to use, default=average, option furthest, nearest, or average. \n"; + helpString += "The splitmethod parameter allows you to specify how you want to split your distance file before you cluster, default=distance, options distance, classify or fasta. \n"; + helpString += "The taxonomy parameter allows you to enter the taxonomy file for your sequences, this is only valid if you are using splitmethod=classify. Be sure your taxonomy file does not include the probability scores. \n"; + helpString += "The taxlevel parameter allows you to specify the taxonomy level you want to use to split the distance file, default=3, meaning use the first taxon in each list. \n"; + helpString += "The large parameter allows you to indicate that your distance matrix is too large to fit in RAM. The default value is false.\n"; + helpString += "The classic parameter allows you to indicate that you want to run your files with cluster.classic. It is only valid with splitmethod=fasta. Default=f.\n"; +#ifdef USE_MPI + helpString += "When using MPI, the processors parameter is set to the number of MPI processes running. \n"; +#endif + helpString += "The cluster.split command should be in the following format: \n"; + helpString += "cluster.split(column=youDistanceFile, name=yourNameFile, method=yourMethod, cutoff=yourCutoff, precision=yourPrecision, splitmethod=yourSplitmethod, taxonomy=yourTaxonomyfile, taxlevel=yourtaxlevel) \n"; + helpString += "Example: cluster.split(column=abrecovery.dist, name=abrecovery.names, method=furthest, cutoff=0.10, precision=1000, splitmethod=classify, taxonomy=abrecovery.silva.slv.taxonomy, taxlevel=5) \n"; + return helpString; } catch(exception& e) { - m->errorOut(e, "ClusterSplitCommand", "ClusterSplitCommand"); + m->errorOut(e, "ClusterSplitCommand", "getHelpString"); exit(1); } } //********************************************************************************************************************** -vector ClusterSplitCommand::getRequiredParameters(){ - try { - string Array[] = {"fasta","phylip","column","or"}; - vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); - return myArray; - } - catch(exception& e) { - m->errorOut(e, "ClusterSplitCommand", "getRequiredParameters"); - exit(1); - } +string ClusterSplitCommand::getOutputPattern(string type) { + try { + string pattern = ""; + + if (type == "list") { pattern = "[filename],[clustertag],list-[filename],[clustertag],[tag2],list"; } + else if (type == "rabund") { pattern = "[filename],[clustertag],rabund"; } + else if (type == "sabund") { pattern = "[filename],[clustertag],sabund"; } + 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, "ClusterSplitCommand", "getOutputPattern"); + exit(1); + } } //********************************************************************************************************************** -vector ClusterSplitCommand::getRequiredFiles(){ +ClusterSplitCommand::ClusterSplitCommand(){ try { - vector myArray; - return myArray; + abort = true; calledHelp = true; + setParameters(); + vector tempOutNames; + outputTypes["list"] = tempOutNames; + outputTypes["rabund"] = tempOutNames; + outputTypes["sabund"] = tempOutNames; + outputTypes["column"] = tempOutNames; } catch(exception& e) { - m->errorOut(e, "ClusterSplitCommand", "getRequiredFiles"); + m->errorOut(e, "ClusterSplitCommand", "ClusterSplitCommand"); exit(1); } } @@ -70,17 +118,15 @@ vector ClusterSplitCommand::getRequiredFiles(){ //This function checks to make sure the cluster command has no errors and then clusters based on the method chosen. ClusterSplitCommand::ClusterSplitCommand(string option) { try{ - globaldata = GlobalData::getInstance(); abort = false; calledHelp = false; format = ""; //allow user to run help 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","phylip","column","name","cutoff","precision","method","splitmethod","taxonomy","taxlevel","large","showabund","timing","hard","processors","outputdir","inputdir"}; - vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); + vector myArray = setParameters(); OptionParser parser(option); map parameters = parser.getParameters(); @@ -102,8 +148,6 @@ ClusterSplitCommand::ClusterSplitCommand(string option) { outputTypes["sabund"] = tempOutNames; outputTypes["column"] = tempOutNames; - globaldata->newRead(); - //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 = ""; } @@ -151,42 +195,106 @@ ClusterSplitCommand::ClusterSplitCommand(string option) { //if the user has not given a path then, add inputdir. else leave path alone. if (path == "") { parameters["fasta"] = inputDir + it->second; } } + + it = parameters.find("count"); + //user has given a template file + if(it != parameters.end()){ + path = m->hasPath(it->second); + //if the user has not given a path then, add inputdir. else leave path alone. + if (path == "") { parameters["count"] = inputDir + it->second; } + } } //check for required parameters phylipfile = validParameter.validFile(parameters, "phylip", true); if (phylipfile == "not open") { abort = true; } else if (phylipfile == "not found") { phylipfile = ""; } - else { distfile = phylipfile; format = "phylip"; } + else { distfile = phylipfile; format = "phylip"; m->setPhylipFile(phylipfile); } columnfile = validParameter.validFile(parameters, "column", true); if (columnfile == "not open") { abort = true; } else if (columnfile == "not found") { columnfile = ""; } - else { distfile = columnfile; format = "column"; } + else { distfile = columnfile; format = "column"; m->setColumnFile(columnfile); } namefile = validParameter.validFile(parameters, "name", true); - if (namefile == "not open") { abort = true; } - else if (namefile == "not found") { namefile = ""; } + if (namefile == "not open") { abort = true; namefile = "";} + else if (namefile == "not found") { namefile = ""; } + else { m->setNameFile(namefile); } + + countfile = validParameter.validFile(parameters, "count", true); + if (countfile == "not open") { abort = true; countfile = "";} + else if (countfile == "not found") { countfile = ""; } + else { m->setCountTableFile(countfile); } fastafile = validParameter.validFile(parameters, "fasta", true); if (fastafile == "not open") { abort = true; } else if (fastafile == "not found") { fastafile = ""; } - else { distfile = fastafile; splitmethod = "fasta"; } + else { distfile = fastafile; splitmethod = "fasta"; m->setFastaFile(fastafile); } taxFile = validParameter.validFile(parameters, "taxonomy", true); - if (taxFile == "not open") { abort = true; } + if (taxFile == "not open") { taxFile = ""; abort = true; } else if (taxFile == "not found") { taxFile = ""; } - - if ((phylipfile == "") && (columnfile == "") && (fastafile == "")) { m->mothurOut("When executing a cluster.split command you must enter a phylip or a column or fastafile."); m->mothurOutEndLine(); abort = true; } + else { m->setTaxonomyFile(taxFile); if (splitmethod != "fasta") { splitmethod = "classify"; } } + + if ((phylipfile == "") && (columnfile == "") && (fastafile == "")) { + //is there are current file available for either of these? + //give priority to column, then phylip, then fasta + columnfile = m->getColumnFile(); + if (columnfile != "") { m->mothurOut("Using " + columnfile + " as input file for the column parameter."); m->mothurOutEndLine(); } + else { + phylipfile = m->getPhylipFile(); + if (phylipfile != "") { m->mothurOut("Using " + phylipfile + " as input file for the phylip parameter."); m->mothurOutEndLine(); } + else { + fastafile = m->getFastaFile(); + if (fastafile != "") { m->mothurOut("Using " + fastafile + " as input file for the fasta parameter."); m->mothurOutEndLine(); } + else { + m->mothurOut("No valid current files. When executing a cluster.split command you must enter a phylip or a column or fastafile."); m->mothurOutEndLine(); + abort = true; + } + } + } + } else if ((phylipfile != "") && (columnfile != "") && (fastafile != "")) { m->mothurOut("When executing a cluster.split command you must enter ONLY ONE of the following: fasta, phylip or column."); m->mothurOutEndLine(); abort = true; } - + + if ((countfile != "") && (namefile != "")) { m->mothurOut("When executing a cluster.split command you must enter ONLY ONE of the following: count or name."); m->mothurOutEndLine(); abort = true; } + if (columnfile != "") { - if (namefile == "") { m->mothurOut("You need to provide a namefile if you are going to use the column format."); m->mothurOutEndLine(); abort = true; } + if ((namefile == "") && (countfile == "")) { + namefile = m->getNameFile(); + if (namefile != "") { m->mothurOut("Using " + namefile + " as input file for the name parameter."); m->mothurOutEndLine(); } + else { + countfile = m->getCountTableFile(); + if (countfile != "") { m->mothurOut("Using " + countfile + " as input file for the count parameter."); m->mothurOutEndLine(); } + else { + m->mothurOut("You need to provide a namefile or countfile if you are going to use the column format."); m->mothurOutEndLine(); + abort = true; + } + } + } } if (fastafile != "") { - if (taxFile == "") { m->mothurOut("You need to provide a taxonomy file if you are using a fasta file to generate the split."); m->mothurOutEndLine(); abort = true; } - if (namefile == "") { m->mothurOut("You need to provide a names file if you are using a fasta file to generate the split."); m->mothurOutEndLine(); abort = true; } + if (taxFile == "") { + taxFile = m->getTaxonomyFile(); + if (taxFile != "") { m->mothurOut("Using " + taxFile + " as input file for the taxonomy parameter."); m->mothurOutEndLine(); } + else { + m->mothurOut("You need to provide a taxonomy file if you are if you are using a fasta file to generate the split."); m->mothurOutEndLine(); + abort = true; + } + } + + if ((namefile == "") && (countfile == "")) { + namefile = m->getNameFile(); + if (namefile != "") { m->mothurOut("Using " + namefile + " as input file for the name parameter."); m->mothurOutEndLine(); } + else { + countfile = m->getCountTableFile(); + if (countfile != "") { m->mothurOut("Using " + countfile + " as input file for the count parameter."); m->mothurOutEndLine(); } + else { + m->mothurOut("You need to provide a namefile or countfile if you are going to use the fasta file to generate the split."); m->mothurOutEndLine(); + abort = true; + } + } + } } //check for optional parameter and set defaults @@ -197,33 +305,39 @@ ClusterSplitCommand::ClusterSplitCommand(string option) { if (temp == "not found") { temp = "100"; } //saves precision legnth for formatting below length = temp.length(); - convert(temp, precision); + m->mothurConvert(temp, precision); - temp = validParameter.validFile(parameters, "hard", false); if (temp == "not found") { temp = "F"; } + temp = validParameter.validFile(parameters, "hard", false); if (temp == "not found") { temp = "T"; } hard = m->isTrue(temp); temp = validParameter.validFile(parameters, "large", false); if (temp == "not found") { temp = "F"; } large = m->isTrue(temp); - - 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, "splitmethod", false); - if (splitmethod != "fasta") { + if ((splitmethod != "fasta") && (splitmethod != "classify")) { if (temp == "not found") { splitmethod = "distance"; } else { splitmethod = temp; } } - temp = validParameter.validFile(parameters, "cutoff", false); if (temp == "not found") { temp = "10"; } - convert(temp, cutoff); + temp = validParameter.validFile(parameters, "classic", false); if (temp == "not found") { temp = "F"; } + classic = m->isTrue(temp); + + if ((splitmethod != "fasta") && classic) { m->mothurOut("splitmethod must be fasta to use cluster.classic.\n"); abort=true; } + + temp = validParameter.validFile(parameters, "cutoff", false); if (temp == "not found") { temp = "0.25"; } + m->mothurConvert(temp, cutoff); cutoff += (5 / (precision * 10.0)); - temp = validParameter.validFile(parameters, "taxlevel", false); if (temp == "not found") { temp = "1"; } - convert(temp, taxLevelCutoff); + temp = validParameter.validFile(parameters, "taxlevel", false); if (temp == "not found") { temp = "3"; } + m->mothurConvert(temp, taxLevelCutoff); - method = validParameter.validFile(parameters, "method", false); if (method == "not found") { method = "furthest"; } + method = validParameter.validFile(parameters, "method", false); if (method == "not found") { method = "average"; } - if ((method == "furthest") || (method == "nearest") || (method == "average")) { } + if ((method == "furthest") || (method == "nearest") || (method == "average")) { m->mothurOut("Using splitmethod " + splitmethod + ".\n"); } else { m->mothurOut("Not a valid clustering method. Valid clustering algorithms are furthest, nearest or average."); m->mothurOutEndLine(); abort = true; } if ((splitmethod == "distance") || (splitmethod == "classify") || (splitmethod == "fasta")) { } @@ -233,6 +347,9 @@ ClusterSplitCommand::ClusterSplitCommand(string option) { showabund = validParameter.validFile(parameters, "showabund", false); if (showabund == "not found") { showabund = "T"; } + + temp = validParameter.validFile(parameters, "cluster", false); if (temp == "not found") { temp = "T"; } + runCluster = m->isTrue(temp); timing = validParameter.validFile(parameters, "timing", false); if (timing == "not found") { timing = "F"; } @@ -247,45 +364,6 @@ ClusterSplitCommand::ClusterSplitCommand(string option) { //********************************************************************************************************************** -void ClusterSplitCommand::help(){ - try { - m->mothurOut("The cluster.split command parameter options are fasta, phylip, column, name, cutoff, precision, method, splitmethod, taxonomy, taxlevel, showabund, timing, hard, large, processors. Fasta or Phylip or column and name are required.\n"); - m->mothurOut("The cluster.split command can split your files in 3 ways. Splitting by distance file, by classification, or by classification also using a fasta file. \n"); - m->mothurOut("For the distance file method, you need only provide your distance file and mothur will split the file into distinct groups. \n"); - m->mothurOut("For the classification method, you need to provide your distance file and taxonomy file, and set the splitmethod to classify. \n"); - m->mothurOut("You will also need to set the taxlevel you want to split by. mothur will split the sequences into distinct taxonomy groups, and split the distance file based on those groups. \n"); - m->mothurOut("For the classification method using a fasta file, you need to provide your fasta file, names file and taxonomy file. \n"); - m->mothurOut("You will also need to set the taxlevel you want to split by. mothur will split the sequence into distinct taxonomy groups, and create distance files for each grouping. \n"); - m->mothurOut("The phylip and column parameter allow you to enter your distance file. \n"); - m->mothurOut("The fasta parameter allows you to enter your aligned fasta file. \n"); - m->mothurOut("The name parameter allows you to enter your name file and is required if your distance file is in column format. \n"); - m->mothurOut("The cutoff parameter allow you to set the distance you want to cluster to, default is 10.0. \n"); - m->mothurOut("The precision parameter allows you specify the precision of the precision of the distances outputted, default=100, meaning 2 decimal places. \n"); - m->mothurOut("The method allows you to specify what clustering algorythm you want to use, default=furthest, option furthest, nearest, or average. \n"); - m->mothurOut("The splitmethod parameter allows you to specify how you want to split your distance file before you cluster, default=distance, options distance, classify or fasta. \n"); - m->mothurOut("The taxonomy parameter allows you to enter the taxonomy file for your sequences, this is only valid if you are using splitmethod=classify. Be sure your taxonomy file does not include the probability scores. \n"); - m->mothurOut("The taxlevel parameter allows you to specify the taxonomy level you want to use to split the distance file, default=1, meaning use the first taxon in each list. \n"); - m->mothurOut("The large parameter allows you to indicate that your distance matrix is too large to fit in RAM. The default value is false.\n"); - #ifdef USE_MPI - m->mothurOut("When using MPI, the processors parameter is set to the number of MPI processes running. \n"); - #endif - m->mothurOut("The cluster.split command should be in the following format: \n"); - m->mothurOut("cluster.split(column=youDistanceFile, name=yourNameFile, method=yourMethod, cutoff=yourCutoff, precision=yourPrecision, splitmethod=yourSplitmethod, taxonomy=yourTaxonomyfile, taxlevel=yourtaxlevel) \n"); - m->mothurOut("Example: cluster.split(column=abrecovery.dist, name=abrecovery.names, method=furthest, cutoff=0.10, precision=1000, splitmethod=classify, taxonomy=abrecovery.silva.slv.taxonomy, taxlevel=5) \n"); - - } - catch(exception& e) { - m->errorOut(e, "ClusterSplitCommand", "help"); - exit(1); - } -} - -//********************************************************************************************************************** - -ClusterSplitCommand::~ClusterSplitCommand(){} - -//********************************************************************************************************************** - int ClusterSplitCommand::execute(){ try { @@ -326,7 +404,7 @@ int ClusterSplitCommand::execute(){ //if no names file given with phylip file, create it ListVector* listToMakeNameFile = convert->getListVector(); - if (namefile == "") { //you need to make a namefile for split matrix + if ((namefile == "") && (countfile == "")) { //you need to make a namefile for split matrix ofstream out; namefile = phylipfile + ".names"; m->openOutputFile(namefile, out); @@ -348,9 +426,9 @@ int ClusterSplitCommand::execute(){ //split matrix into non-overlapping groups SplitMatrix* split; - if (splitmethod == "distance") { split = new SplitMatrix(distfile, namefile, taxFile, cutoff, splitmethod, large); } - else if (splitmethod == "classify") { split = new SplitMatrix(distfile, namefile, taxFile, taxLevelCutoff, splitmethod, large); } - else if (splitmethod == "fasta") { split = new SplitMatrix(fastafile, namefile, taxFile, taxLevelCutoff, cutoff, splitmethod, processors, outputDir); } + if (splitmethod == "distance") { split = new SplitMatrix(distfile, namefile, countfile, taxFile, cutoff, splitmethod, large); } + else if (splitmethod == "classify") { split = new SplitMatrix(distfile, namefile, countfile, taxFile, taxLevelCutoff, splitmethod, large); } + else if (splitmethod == "fasta") { split = new SplitMatrix(fastafile, namefile, countfile, taxFile, taxLevelCutoff, cutoff, splitmethod, processors, classic, outputDir); } else { m->mothurOut("Not a valid splitting method. Valid splitting algorithms are distance, classify or fasta."); m->mothurOutEndLine(); return 0; } split->split(); @@ -361,15 +439,26 @@ int ClusterSplitCommand::execute(){ vector< map > distName = split->getDistanceFiles(); //returns map of distance files -> namefile sorted by distance file size delete split; + if (m->debug) { m->mothurOut("[DEBUG]: distName.size() = " + toString(distName.size()) + ".\n"); } + //output a merged distance file - if (splitmethod == "fasta") { createMergedDistanceFile(distName); } - + //if (splitmethod == "fasta") { createMergedDistanceFile(distName); } if (m->control_pressed) { return 0; } m->mothurOut("It took " + toString(time(NULL) - estart) + " seconds to split the distance file."); m->mothurOutEndLine(); estart = time(NULL); - + + if (!runCluster) { + + m->mothurOutEndLine(); + m->mothurOut("Output File Names: "); m->mothurOutEndLine(); + for (int i = 0; i < distName.size(); i++) { m->mothurOut(distName[i].begin()->first); m->mothurOutEndLine(); m->mothurOut(distName[i].begin()->second); m->mothurOutEndLine(); } + m->mothurOutEndLine(); + return 0; + + } + //****************** break up files between processes and cluster each file set ******************************// #ifdef USE_MPI ////you are process 0 from above//// @@ -526,72 +615,21 @@ int ClusterSplitCommand::execute(){ MPI_Barrier(MPI_COMM_WORLD); #else - - #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + ///////////////////// WINDOWS CAN ONLY USE 1 PROCESSORS ACCESS VIOLATION UNRESOLVED /////////////////////// + //sanity check + if (processors > distName.size()) { processors = distName.size(); } + + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) if(processors == 1){ listFileNames = cluster(distName, labels); //clusters individual files and returns names of list files }else{ - vector < vector < map > > dividedNames; //distNames[1] = vector of filenames for process 1... - dividedNames.resize(processors); - - //for each file group figure out which process will complete it - //want to divide the load intelligently so the big files are spread between processes - for (int i = 0; i < distName.size(); i++) { - int processToAssign = (i+1) % processors; - if (processToAssign == 0) { processToAssign = processors; } - - dividedNames[(processToAssign-1)].push_back(distName[i]); - } - - //not lets reverse the order of ever other process, so we balance big files running with little ones - for (int i = 0; i < processors; i++) { - int remainder = ((i+1) % processors); - if (remainder) { reverse(dividedNames[i].begin(), dividedNames[i].end()); } - } - - createProcesses(dividedNames); - - if (m->control_pressed) { return 0; } - - //get list of list file names from each process - for(int i=0;iopenInputFile(filename, in); - - in >> tag; m->gobble(in); - - while(!in.eof()) { - string tempName; - in >> tempName; m->gobble(in); - listFileNames.push_back(tempName); - } - in.close(); - remove((toString(processIDS[i]) + ".temp").c_str()); - - //get labels - filename = toString(processIDS[i]) + ".temp.labels"; - ifstream in2; - m->openInputFile(filename, in2); - - float tempCutoff; - in2 >> tempCutoff; m->gobble(in2); - if (tempCutoff < cutoff) { cutoff = tempCutoff; } - - while(!in2.eof()) { - string tempName; - in2 >> tempName; m->gobble(in2); - if (labels.count(tempName) == 0) { labels.insert(tempName); } - } - in2.close(); - remove((toString(processIDS[i]) + ".temp.labels").c_str()); - } - } + listFileNames = createProcesses(distName, labels); + } #else listFileNames = cluster(distName, labels); //clusters individual files and returns names of list files #endif #endif - if (m->control_pressed) { for (int i = 0; i < listFileNames.size(); i++) { remove(listFileNames[i].c_str()); } return 0; } + if (m->control_pressed) { for (int i = 0; i < listFileNames.size(); i++) { m->mothurRemove(listFileNames[i]); } return 0; } if (saveCutoff != cutoff) { m->mothurOut("Cutoff was " + toString(saveCutoff) + " changed cutoff to " + toString(cutoff)); m->mothurOutEndLine(); } @@ -608,11 +646,11 @@ int ClusterSplitCommand::execute(){ ListVector* listSingle; map labelBins = completeListFile(listFileNames, singletonName, labels, listSingle); //returns map of label to numBins - if (m->control_pressed) { if (listSingle != NULL) { delete listSingle; } for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; } + if (m->control_pressed) { if (listSingle != NULL) { delete listSingle; } for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } return 0; } mergeLists(listFileNames, labelBins, listSingle); - if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; } + if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } return 0; } m->mothurOut("It took " + toString(time(NULL) - estart) + " seconds to merge."); m->mothurOutEndLine(); @@ -664,17 +702,23 @@ map ClusterSplitCommand::completeListFile(vector listNames, //read in singletons if (singleton != "none") { - ifstream in; - m->openInputFile(singleton, in); + + ifstream in; + m->openInputFile(singleton, in); string firstCol, secondCol; listSingle = new ListVector(); + + if (countfile != "") { m->getline(in); m->gobble(in); } + while (!in.eof()) { - in >> firstCol >> secondCol; m->gobble(in); - listSingle->push_back(secondCol); + in >> firstCol >> secondCol; m->getline(in); m->gobble(in); + if (countfile == "") { listSingle->push_back(secondCol); } + else { listSingle->push_back(firstCol); } } + in.close(); - remove(singleton.c_str()); + m->mothurRemove(singleton); numSingleBins = listSingle->getNumBins(); }else{ listSingle = NULL; numSingleBins = 0; } @@ -700,8 +744,8 @@ map ClusterSplitCommand::completeListFile(vector listNames, for (int k = 0; k < listNames.size(); k++) { if (m->control_pressed) { - if (listSingle != NULL) { delete listSingle; listSingle = NULL; remove(singleton.c_str()); } - for (int i = 0; i < listNames.size(); i++) { remove(listNames[i].c_str()); } + if (listSingle != NULL) { delete listSingle; listSingle = NULL; m->mothurRemove(singleton); } + for (int i = 0; i < listNames.size(); i++) { m->mothurRemove(listNames[i]); } return labelBin; } @@ -754,7 +798,7 @@ map ClusterSplitCommand::completeListFile(vector listNames, delete input; outFilled.close(); - remove(listNames[k].c_str()); + m->mothurRemove(listNames[k]); rename(filledInList.c_str(), listNames[k].c_str()); } @@ -771,13 +815,24 @@ int ClusterSplitCommand::mergeLists(vector listNames, map us if (outputDir == "") { outputDir += m->hasPath(distfile); } fileroot = outputDir + m->getRootName(m->getSimpleName(distfile)); - m->openOutputFile(fileroot+ tag + ".sabund", outSabund); - m->openOutputFile(fileroot+ tag + ".rabund", outRabund); - m->openOutputFile(fileroot+ tag + ".list", outList); - - outputNames.push_back(fileroot+ tag + ".sabund"); outputTypes["list"].push_back(fileroot+ tag + ".list"); - outputNames.push_back(fileroot+ tag + ".rabund"); outputTypes["rabund"].push_back(fileroot+ tag + ".rabund"); - outputNames.push_back(fileroot+ tag + ".list"); outputTypes["sabund"].push_back(fileroot+ tag + ".sabund"); + map variables; + variables["[filename]"] = fileroot; + variables["[clustertag]"] = tag; + string sabundFileName = getOutputFileName("sabund", variables); + string rabundFileName = getOutputFileName("rabund", variables); + if (countfile != "") { variables["[tag2]"] = "unique_list"; } + string listFileName = getOutputFileName("list", variables); + + if (countfile == "") { + m->openOutputFile(sabundFileName, outSabund); + m->openOutputFile(rabundFileName, outRabund); + outputNames.push_back(sabundFileName); outputTypes["sabund"].push_back(sabundFileName); + outputNames.push_back(rabundFileName); outputTypes["rabund"].push_back(rabundFileName); + + } + m->openOutputFile(listFileName, outList); + outputNames.push_back(listFileName); outputTypes["list"].push_back(listFileName); + map::iterator itLabel; @@ -788,23 +843,30 @@ int ClusterSplitCommand::mergeLists(vector listNames, map us if (itLabel->first == -1) { thisLabel = "unique"; } else { thisLabel = toString(itLabel->first, length-1); } - outList << thisLabel << '\t' << itLabel->second << '\t'; - - RAbundVector* rabund = new RAbundVector(); - rabund->setLabel(thisLabel); + //outList << thisLabel << '\t' << itLabel->second << '\t'; + + RAbundVector* rabund = NULL; + ListVector completeList; + completeList.setLabel(thisLabel); + + if (countfile == "") { + rabund = new RAbundVector(); + rabund->setLabel(thisLabel); + } //add in singletons if (listSingle != NULL) { for (int j = 0; j < listSingle->getNumBins(); j++) { - outList << listSingle->get(j) << '\t'; - rabund->push_back(m->getNumNames(listSingle->get(j))); + //outList << listSingle->get(j) << '\t'; + completeList.push_back(listSingle->get(j)); + if (countfile == "") { rabund->push_back(m->getNumNames(listSingle->get(j))); } } } //get the list info from each file for (int k = 0; k < listNames.size(); k++) { - if (m->control_pressed) { if (listSingle != NULL) { delete listSingle; } for (int i = 0; i < listNames.size(); i++) { remove(listNames[i].c_str()); } delete rabund; return 0; } + if (m->control_pressed) { if (listSingle != NULL) { delete listSingle; } for (int i = 0; i < listNames.size(); i++) { m->mothurRemove(listNames[i]); } if (rabund != NULL) { delete rabund; } return 0; } InputData* input = new InputData(listNames[k], "list"); ListVector* list = input->getListVector(thisLabel); @@ -813,30 +875,34 @@ int ClusterSplitCommand::mergeLists(vector listNames, map us if (list == NULL) { m->mothurOut("Error merging listvectors in file " + listNames[k]); m->mothurOutEndLine(); } else { for (int j = 0; j < list->getNumBins(); j++) { - outList << list->get(j) << '\t'; - rabund->push_back(m->getNumNames(list->get(j))); + //outList << list->get(j) << '\t'; + completeList.push_back(list->get(j)); + if (countfile == "") { rabund->push_back(m->getNumNames(list->get(j))); } } delete list; } delete input; } - SAbundVector sabund = rabund->getSAbundVector(); + if (countfile == "") { + SAbundVector sabund = rabund->getSAbundVector(); + sabund.print(outSabund); + rabund->print(outRabund); + } + //outList << endl; + completeList.print(outList); - sabund.print(outSabund); - rabund->print(outRabund); - outList << endl; - - delete rabund; + if (rabund != NULL) { delete rabund; } } outList.close(); - outRabund.close(); - outSabund.close(); - + if (countfile == "") { + outRabund.close(); + outSabund.close(); + } if (listSingle != NULL) { delete listSingle; } - for (int i = 0; i < listNames.size(); i++) { remove(listNames[i].c_str()); } + for (int i = 0; i < listNames.size(); i++) { m->mothurRemove(listNames[i]); } return 0; } @@ -868,12 +934,35 @@ void ClusterSplitCommand::printData(ListVector* oldList){ } } //********************************************************************************************************************** -int ClusterSplitCommand::createProcesses(vector < vector < map > > dividedNames){ +vector ClusterSplitCommand::createProcesses(vector< map > distName, set& labels){ try { + + vector listFiles; + vector < vector < map > > dividedNames; //distNames[1] = vector of filenames for process 1... + dividedNames.resize(processors); + + //for each file group figure out which process will complete it + //want to divide the load intelligently so the big files are spread between processes + for (int i = 0; i < distName.size(); i++) { + //cout << i << endl; + int processToAssign = (i+1) % processors; + if (processToAssign == 0) { processToAssign = processors; } + + dividedNames[(processToAssign-1)].push_back(distName[i]); + if ((processToAssign-1) == 1) { m->mothurOut(distName[i].begin()->first + "\n"); } + } + + //now lets reverse the order of ever other process, so we balance big files running with little ones + for (int i = 0; i < processors; i++) { + //cout << i << endl; + int remainder = ((i+1) % processors); + if (remainder) { reverse(dividedNames[i].begin(), dividedNames[i].end()); } + } + + if (m->control_pressed) { return listFiles; } - #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - int process = 0; - int exitCommand = 1; + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + int process = 1; processIDS.clear(); //loop through and create all the processes you want @@ -914,14 +1003,99 @@ int ClusterSplitCommand::createProcesses(vector < vector < map > } } + //do your part + listFiles = cluster(dividedNames[0], labels); + //force parent to wait until all the processes are done - for (int i=0;iopenInputFile(filename, in); + + in >> tag; m->gobble(in); + + while(!in.eof()) { + string tempName; + in >> tempName; m->gobble(in); + listFiles.push_back(tempName); + } + in.close(); + m->mothurRemove((toString(processIDS[i]) + ".temp")); + + //get labels + filename = toString(processIDS[i]) + ".temp.labels"; + ifstream in2; + m->openInputFile(filename, in2); + + float tempCutoff; + in2 >> tempCutoff; m->gobble(in2); + if (tempCutoff < cutoff) { cutoff = tempCutoff; } + + while(!in2.eof()) { + string tempName; + in2 >> tempName; m->gobble(in2); + if (labels.count(tempName) == 0) { labels.insert(tempName); } + } + in2.close(); + m->mothurRemove((toString(processIDS[i]) + ".temp.labels")); + } + + + #else + + ////////////////////////////////////////////////////////////////////////////////////////////////////// + //Windows version shared memory, so be careful when passing variables through the clusterData 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 labels. + ////////////////////////////////////////////////////////////////////////////////////////////////////// + /* + vector pDataArray; + DWORD dwThreadIdArray[processors-1]; + HANDLE hThreadArray[processors-1]; + + //Create processor worker threads. + for( int i=1; itag; + //get listfiles created + for(int j=0; j < pDataArray[i]->listFiles.size(); j++){ listFiles.push_back(pDataArray[i]->listFiles[j]); } + //get labels + set::iterator it; + for(it = pDataArray[i]->labels.begin(); it != pDataArray[i]->labels.end(); it++){ labels.insert(*it); } + //check cutoff + if (pDataArray[i]->cutoff < cutoff) { cutoff = pDataArray[i]->cutoff; } + CloseHandle(hThreadArray[i]); + delete pDataArray[i]; + } +*/ #endif + + return listFiles; } catch(exception& e) { @@ -933,169 +1107,327 @@ int ClusterSplitCommand::createProcesses(vector < vector < map > vector ClusterSplitCommand::cluster(vector< map > distNames, set& labels){ try { - Cluster* cluster; - SparseMatrix* matrix; - ListVector* list; - ListVector oldList; - RAbundVector* rabund; vector listFileNames; - double smallestCutoff = cutoff; //cluster each distance file for (int i = 0; i < distNames.size(); i++) { - if (m->control_pressed) { return listFileNames; } - + string thisNamefile = distNames[i].begin()->second; string thisDistFile = distNames[i].begin()->first; - //read in distance file - globaldata->setNameFile(thisNamefile); - globaldata->setColumnFile(thisDistFile); globaldata->setFormat("column"); - - #ifdef USE_MPI - int pid; - MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are - - //output your files too - if (pid != 0) { - cout << endl << "Reading " << thisDistFile << endl; - } - #endif - - m->mothurOutEndLine(); m->mothurOut("Reading " + thisDistFile); m->mothurOutEndLine(); - - ReadMatrix* read = new ReadColumnMatrix(thisDistFile); - read->setCutoff(cutoff); + string listFileName = ""; + if (classic) { listFileName = clusterClassicFile(thisDistFile, thisNamefile, labels, smallestCutoff); } + else { listFileName = clusterFile(thisDistFile, thisNamefile, labels, smallestCutoff); } - NameAssignment* nameMap = new NameAssignment(thisNamefile); - nameMap->readMap(); - read->read(nameMap); - - if (m->control_pressed) { delete read; delete nameMap; return listFileNames; } - - list = read->getListVector(); - oldList = *list; - matrix = read->getMatrix(); - - delete read; - delete nameMap; - - - #ifdef USE_MPI - //output your files too - if (pid != 0) { - cout << endl << "Clustering " << thisDistFile << endl; - } - #endif - - m->mothurOutEndLine(); m->mothurOut("Clustering " + thisDistFile); m->mothurOutEndLine(); - - rabund = new RAbundVector(list->getRAbundVector()); - - //create cluster - if (method == "furthest") { cluster = new CompleteLinkage(rabund, list, matrix, cutoff, method); } - else if(method == "nearest"){ cluster = new SingleLinkage(rabund, list, matrix, cutoff, method); } - else if(method == "average"){ cluster = new AverageLinkage(rabund, list, matrix, cutoff, method); } - tag = cluster->getTag(); - - if (outputDir == "") { outputDir += m->hasPath(thisDistFile); } - fileroot = outputDir + m->getRootName(m->getSimpleName(thisDistFile)); - - ofstream listFile; - m->openOutputFile(fileroot+ tag + ".list", listFile); + if (m->control_pressed) { //clean up + for (int i = 0; i < listFileNames.size(); i++) { m->mothurRemove(listFileNames[i]); } + listFileNames.clear(); return listFileNames; + } + + listFileNames.push_back(listFileName); + } - listFileNames.push_back(fileroot+ tag + ".list"); - - float previousDist = 0.00000; - float rndPreviousDist = 0.00000; - - oldList = *list; + cutoff = smallestCutoff; + + return listFileNames; + + } + catch(exception& e) { + m->errorOut(e, "ClusterSplitCommand", "cluster"); + exit(1); + } + - print_start = true; - start = time(NULL); - double saveCutoff = cutoff; +} +//********************************************************************************************************************** +string ClusterSplitCommand::clusterClassicFile(string thisDistFile, string thisNamefile, set& labels, double& smallestCutoff){ + try { + string listFileName = ""; + + ListVector* list = NULL; + ListVector oldList; + RAbundVector* rabund = NULL; + +#ifdef USE_MPI + int pid; + MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are + + //output your files too + if (pid != 0) { + cout << endl << "Reading " << thisDistFile << endl; + } +#endif + + m->mothurOutEndLine(); m->mothurOut("Reading " + thisDistFile); m->mothurOutEndLine(); + + //reads phylip file storing data in 2D vector, also fills list and rabund + bool sim = false; + ClusterClassic* cluster = new ClusterClassic(cutoff, method, sim); + + NameAssignment* nameMap = NULL; + CountTable* ct = NULL; + if(namefile != ""){ + nameMap = new NameAssignment(thisNamefile); + nameMap->readMap(); + cluster->readPhylipFile(thisDistFile, nameMap); + }else if (countfile != "") { + ct = new CountTable(); + ct->readTable(thisNamefile, false); + cluster->readPhylipFile(thisDistFile, ct); + } + tag = cluster->getTag(); + + if (m->control_pressed) { if(namefile != ""){ delete nameMap; } + else { delete ct; } delete cluster; return 0; } - while (matrix->getSmallDist() < cutoff && matrix->getNNodes() > 0){ + list = cluster->getListVector(); + rabund = cluster->getRAbundVector(); + + if (outputDir == "") { outputDir += m->hasPath(thisDistFile); } + fileroot = outputDir + m->getRootName(m->getSimpleName(thisDistFile)); + listFileName = fileroot+ tag + ".list"; + + ofstream listFile; + m->openOutputFile(fileroot+ tag + ".list", listFile); - if (m->control_pressed) { //clean up - delete matrix; delete list; delete cluster; delete rabund; - listFile.close(); - for (int i = 0; i < listFileNames.size(); i++) { remove(listFileNames[i].c_str()); } - listFileNames.clear(); return listFileNames; - } + float previousDist = 0.00000; + float rndPreviousDist = 0.00000; + oldList = *list; - cluster->update(saveCutoff); - - float dist = matrix->getSmallDist(); - float rndDist; - if (hard) { - rndDist = m->ceilDist(dist, precision); - }else{ - rndDist = m->roundDist(dist, precision); - } +#ifdef USE_MPI + //output your files too + if (pid != 0) { + cout << endl << "Clustering " << thisDistFile << endl; + } +#endif - if(previousDist <= 0.0000 && dist != previousDist){ - oldList.setLabel("unique"); - oldList.print(listFile); - if (labels.count("unique") == 0) { labels.insert("unique"); } - } - else if(rndDist != rndPreviousDist){ - oldList.setLabel(toString(rndPreviousDist, length-1)); - oldList.print(listFile); - if (labels.count(toString(rndPreviousDist, length-1)) == 0) { labels.insert(toString(rndPreviousDist, length-1)); } - } - - previousDist = dist; - rndPreviousDist = rndDist; - oldList = *list; + m->mothurOutEndLine(); m->mothurOut("Clustering " + thisDistFile); m->mothurOutEndLine(); + + while ((cluster->getSmallDist() < cutoff) && (cluster->getNSeqs() > 1)){ + if (m->control_pressed) { delete cluster; delete list; delete rabund; listFile.close(); if(namefile != ""){ delete nameMap; } + else { delete ct; } return listFileName; } + + cluster->update(cutoff); + + float dist = cluster->getSmallDist(); + float rndDist; + if (hard) { + rndDist = m->ceilDist(dist, precision); + }else{ + rndDist = m->roundDist(dist, precision); } + + if(previousDist <= 0.0000 && dist != previousDist){ + oldList.setLabel("unique"); + oldList.print(listFile); + if (labels.count("unique") == 0) { labels.insert("unique"); } + } + else if(rndDist != rndPreviousDist){ + oldList.setLabel(toString(rndPreviousDist, length-1)); + oldList.print(listFile); + if (labels.count(toString(rndPreviousDist, length-1)) == 0) { labels.insert(toString(rndPreviousDist, length-1)); } + } - - if(previousDist <= 0.0000){ - oldList.setLabel("unique"); - oldList.print(listFile); - if (labels.count("unique") == 0) { labels.insert("unique"); } - } - else if(rndPreviousDistcontrol_pressed) { //clean up - for (int i = 0; i < listFileNames.size(); i++) { remove(listFileNames[i].c_str()); } - listFileNames.clear(); return listFileNames; - } - - remove(thisDistFile.c_str()); - remove(thisNamefile.c_str()); - - if (saveCutoff != cutoff) { - if (hard) { saveCutoff = m->ceilDist(saveCutoff, precision); } - else { saveCutoff = m->roundDist(saveCutoff, precision); } - - m->mothurOut("Cutoff was " + toString(cutoff) + " changed cutoff to " + toString(saveCutoff)); m->mothurOutEndLine(); - } - - if (saveCutoff < smallestCutoff) { smallestCutoff = saveCutoff; } + + previousDist = dist; + rndPreviousDist = rndDist; + oldList = *list; } + + if(previousDist <= 0.0000){ + oldList.setLabel("unique"); + oldList.print(listFile); + if (labels.count("unique") == 0) { labels.insert("unique"); } + } + else if(rndPreviousDistmothurRemove(thisDistFile); + m->mothurRemove(thisNamefile); + + return listFileName; + } catch(exception& e) { - m->errorOut(e, "ClusterSplitCommand", "cluster"); + m->errorOut(e, "ClusterSplitCommand", "clusterClassicFile"); exit(1); } +} +//********************************************************************************************************************** +string ClusterSplitCommand::clusterFile(string thisDistFile, string thisNamefile, set& labels, double& smallestCutoff){ + try { + string listFileName = ""; + + Cluster* cluster = NULL; + SparseDistanceMatrix* matrix = NULL; + ListVector* list = NULL; + ListVector oldList; + RAbundVector* rabund = NULL; + + if (m->control_pressed) { return listFileName; } + +#ifdef USE_MPI + int pid; + MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are + + //output your files too + if (pid != 0) { + cout << endl << "Reading " << thisDistFile << endl; + } +#endif + + m->mothurOutEndLine(); m->mothurOut("Reading " + thisDistFile); m->mothurOutEndLine(); + + ReadMatrix* read = new ReadColumnMatrix(thisDistFile); + read->setCutoff(cutoff); + + NameAssignment* nameMap = NULL; + CountTable* ct = NULL; + if(namefile != ""){ + nameMap = new NameAssignment(thisNamefile); + nameMap->readMap(); + read->read(nameMap); + }else if (countfile != "") { + ct = new CountTable(); + ct->readTable(thisNamefile, false); + read->read(ct); + }else { read->read(nameMap); } + + list = read->getListVector(); + oldList = *list; + matrix = read->getDMatrix(); + + if(countfile != "") { + rabund = new RAbundVector(); + createRabund(ct, list, rabund); //creates an rabund that includes the counts for the unique list + delete ct; + }else { rabund = new RAbundVector(list->getRAbundVector()); } + delete read; read = NULL; + if (namefile != "") { delete nameMap; nameMap = NULL; } + + +#ifdef USE_MPI + //output your files too + if (pid != 0) { + cout << endl << "Clustering " << thisDistFile << endl; + } +#endif + + m->mothurOutEndLine(); m->mothurOut("Clustering " + thisDistFile); m->mothurOutEndLine(); + + //create cluster + float adjust = -1.0; + if (method == "furthest") { cluster = new CompleteLinkage(rabund, list, matrix, cutoff, method, adjust); } + else if(method == "nearest"){ cluster = new SingleLinkage(rabund, list, matrix, cutoff, method, adjust); } + else if(method == "average"){ cluster = new AverageLinkage(rabund, list, matrix, cutoff, method, adjust); } + tag = cluster->getTag(); + + if (outputDir == "") { outputDir += m->hasPath(thisDistFile); } + fileroot = outputDir + m->getRootName(m->getSimpleName(thisDistFile)); + + ofstream listFile; + m->openOutputFile(fileroot+ tag + ".list", listFile); + + listFileName = fileroot+ tag + ".list"; + + float previousDist = 0.00000; + float rndPreviousDist = 0.00000; + + oldList = *list; + + print_start = true; + start = time(NULL); + double saveCutoff = cutoff; + + while (matrix->getSmallDist() < cutoff && matrix->getNNodes() > 0){ + + if (m->control_pressed) { //clean up + delete matrix; delete list; delete cluster; delete rabund; + listFile.close(); + m->mothurRemove(listFileName); + return listFileName; + } + + cluster->update(saveCutoff); + + float dist = matrix->getSmallDist(); + float rndDist; + if (hard) { + rndDist = m->ceilDist(dist, precision); + }else{ + rndDist = m->roundDist(dist, precision); + } + + if(previousDist <= 0.0000 && dist != previousDist){ + oldList.setLabel("unique"); + oldList.print(listFile); + if (labels.count("unique") == 0) { labels.insert("unique"); } + } + else if(rndDist != rndPreviousDist){ + oldList.setLabel(toString(rndPreviousDist, length-1)); + oldList.print(listFile); + if (labels.count(toString(rndPreviousDist, length-1)) == 0) { labels.insert(toString(rndPreviousDist, length-1)); } + } + + previousDist = dist; + rndPreviousDist = rndDist; + oldList = *list; + } + + + if(previousDist <= 0.0000){ + oldList.setLabel("unique"); + oldList.print(listFile); + if (labels.count("unique") == 0) { labels.insert("unique"); } + } + else if(rndPreviousDistcontrol_pressed) { //clean up + m->mothurRemove(listFileName); + return listFileName; + } + + m->mothurRemove(thisDistFile); + m->mothurRemove(thisNamefile); + + if (saveCutoff != cutoff) { + if (hard) { saveCutoff = m->ceilDist(saveCutoff, precision); } + else { saveCutoff = m->roundDist(saveCutoff, precision); } + + m->mothurOut("Cutoff was " + toString(cutoff) + " changed cutoff to " + toString(saveCutoff)); m->mothurOutEndLine(); + } + + if (saveCutoff < smallestCutoff) { smallestCutoff = saveCutoff; } + + return listFileName; + + } + catch(exception& e) { + m->errorOut(e, "ClusterSplitCommand", "clusterFile"); + exit(1); + } } //********************************************************************************************************************** @@ -1111,8 +1443,10 @@ int ClusterSplitCommand::createMergedDistanceFile(vector< map > string thisOutputDir = outputDir; if (outputDir == "") { thisOutputDir = m->hasPath(fastafile); } - string outputFileName = thisOutputDir + m->getRootName(m->getSimpleName(fastafile)) + "dist"; - remove(outputFileName.c_str()); + map variables; + variables["[filename]"] = thisOutputDir + m->getRootName(m->getSimpleName(fastafile)); + string outputFileName = getOutputFileName("column", variables); + m->mothurRemove(outputFileName); for (int i = 0; i < distNames.size(); i++) { @@ -1139,3 +1473,24 @@ int ClusterSplitCommand::createMergedDistanceFile(vector< map > } } //********************************************************************************************************************** +int ClusterSplitCommand::createRabund(CountTable*& ct, ListVector*& list, RAbundVector*& rabund){ + try { + rabund->setLabel(list->getLabel()); + for(int i = 0; i < list->getNumBins(); i++) { + if (m->control_pressed) { break; } + vector binNames; + string bin = list->get(i); + m->splitAtComma(bin, binNames); + int total = 0; + for (int j = 0; j < binNames.size(); j++) { total += ct->getNumSeqs(binNames[j]); } + rabund->push_back(total); + } + return 0; + } + catch(exception& e) { + m->errorOut(e, "ClusterCommand", "createRabund"); + exit(1); + } + +} +//**********************************************************************************************************************