X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=clustersplitcommand.cpp;h=2eee59f6165e28c50f86fd6ab9b1db0920135040;hb=15b252e317fde72ce0a5de8776cb88112e0ed2b6;hp=9cdd7b065b857e5ddb7eaacbc27e4c3af84a43cb;hpb=4c302368ef34f0d897afefc7853edf86fb45b9f3;p=mothur.git diff --git a/clustersplitcommand.cpp b/clustersplitcommand.cpp index 9cdd7b0..2eee59f 100644 --- a/clustersplitcommand.cpp +++ b/clustersplitcommand.cpp @@ -10,7 +10,6 @@ #include "clustersplitcommand.h" - //********************************************************************************************************************** vector ClusterSplitCommand::setParameters(){ try { @@ -29,6 +28,7 @@ vector ClusterSplitCommand::setParameters(){ 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); @@ -62,6 +62,7 @@ string ClusterSplitCommand::getHelpString(){ 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 @@ -266,7 +267,7 @@ ClusterSplitCommand::ClusterSplitCommand(string option) { 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 = m->getProcessors(); } m->setProcessors(temp); m->mothurConvert(temp, processors); @@ -277,6 +278,11 @@ ClusterSplitCommand::ClusterSplitCommand(string option) { else { splitmethod = temp; } } + 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)); @@ -374,7 +380,7 @@ int ClusterSplitCommand::execute(){ 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); } + else if (splitmethod == "fasta") { split = new SplitMatrix(fastafile, namefile, 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(); @@ -1021,148 +1027,20 @@ vector ClusterSplitCommand::cluster(vector< map > distNa //cluster each distance file for (int i = 0; i < distNames.size(); i++) { - Cluster* cluster = NULL; - SparseMatrix* matrix = NULL; - ListVector* list = NULL; - ListVector oldList; - RAbundVector* rabund = NULL; - - if (m->control_pressed) { return listFileNames; } - string thisNamefile = distNames[i].begin()->second; string thisDistFile = distNames[i].begin()->first; - - #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 = 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; read = NULL; - 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(); - - 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); - - listFileNames.push_back(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(); - for (int i = 0; i < listFileNames.size(); i++) { m->mothurRemove(listFileNames[i]); } - listFileNames.clear(); return listFileNames; - } - - 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; - } + string listFileName = ""; + if (classic) { listFileName = clusterClassicFile(thisDistFile, thisNamefile, labels, smallestCutoff); } + else { listFileName = clusterFile(thisDistFile, thisNamefile, labels, smallestCutoff); } - - 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++) { m->mothurRemove(listFileNames[i]); } listFileNames.clear(); return listFileNames; } - - 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; } - } + + listFileNames.push_back(listFileName); + } cutoff = smallestCutoff; @@ -1177,6 +1055,269 @@ vector ClusterSplitCommand::cluster(vector< map > distNa } //********************************************************************************************************************** +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(); + + NameAssignment* nameMap = new NameAssignment(thisNamefile); + nameMap->readMap(); + + //reads phylip file storing data in 2D vector, also fills list and rabund + bool sim = false; + ClusterClassic* cluster = new ClusterClassic(cutoff, method, sim); + cluster->readPhylipFile(thisDistFile, nameMap); + tag = cluster->getTag(); + + if (m->control_pressed) { delete cluster; return 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); + + float previousDist = 0.00000; + float rndPreviousDist = 0.00000; + oldList = *list; + +#ifdef USE_MPI + //output your files too + if (pid != 0) { + cout << endl << "Clustering " << thisDistFile << endl; + } +#endif + + 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(); 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)); } + } + + + 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(rndPreviousDisterrorOut(e, "ClusterSplitCommand", "clusterClassicFile"); + exit(1); + } +} + +//********************************************************************************************************************** +string ClusterSplitCommand::clusterFile(string thisDistFile, string thisNamefile, set& labels, double& smallestCutoff){ + try { + string listFileName = ""; + + Cluster* cluster = NULL; + SparseMatrix* 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 = new NameAssignment(thisNamefile); + nameMap->readMap(); + read->read(nameMap); + + if (m->control_pressed) { delete read; delete nameMap; return listFileName; } + + list = read->getListVector(); + oldList = *list; + matrix = read->getMatrix(); + + delete read; read = NULL; + 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(); + + 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); + + 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); + } +} +//********************************************************************************************************************** int ClusterSplitCommand::createMergedDistanceFile(vector< map > distNames) { try{