X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=mgclustercommand.cpp;h=eef54ef23b290a930d5560e8dbf444a20da46073;hb=260ae19c36cb11a53ddc5a75b5e507f8dd8b31d6;hp=07c51479f35ff00db1694dcef2bc135df32976da;hpb=315e38cf393c82be238da5b32574f225a020d25c;p=mothur.git diff --git a/mgclustercommand.cpp b/mgclustercommand.cpp index 07c5147..eef54ef 100644 --- a/mgclustercommand.cpp +++ b/mgclustercommand.cpp @@ -10,7 +10,7 @@ #include "mgclustercommand.h" //********************************************************************************************************************** -MGClusterCommand::MGClusterCommand(string option){ +MGClusterCommand::MGClusterCommand(string option) { try { globaldata = GlobalData::getInstance(); abort = false; @@ -20,7 +20,7 @@ MGClusterCommand::MGClusterCommand(string option){ else { //valid paramters for this command - string Array[] = {"blast", "method", "name", "cutoff", "precision", "length", "min", "penalty", "hcluster","merge","outputdir","inputdir"}; + string Array[] = {"blast", "method", "name", "hard", "cutoff", "precision", "length", "min", "penalty", "hcluster","merge","outputdir","inputdir"}; vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); OptionParser parser(option); @@ -42,7 +42,7 @@ MGClusterCommand::MGClusterCommand(string option){ it = parameters.find("blast"); //user has given a template file if(it != parameters.end()){ - path = hasPath(it->second); + path = m->hasPath(it->second); //if the user has not given a path then, add inputdir. else leave path alone. if (path == "") { parameters["blast"] = inputDir + it->second; } } @@ -50,7 +50,7 @@ MGClusterCommand::MGClusterCommand(string option){ it = parameters.find("name"); //user has given a template file if(it != parameters.end()){ - path = hasPath(it->second); + path = m->hasPath(it->second); //if the user has not given a path then, add inputdir. else leave path alone. if (path == "") { parameters["name"] = inputDir + it->second; } } @@ -65,14 +65,14 @@ MGClusterCommand::MGClusterCommand(string option){ //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(blastfile); //if user entered a file with a path then preserve it + outputDir += m->hasPath(blastfile); //if user entered a file with a path then preserve it } namefile = validParameter.validFile(parameters, "name", true); if (namefile == "not open") { abort = true; } else if (namefile == "not found") { namefile = ""; } - if ((blastfile == "")) { mothurOut("When executing a mgcluster command you must provide a blastfile."); mothurOutEndLine(); abort = true; } + if ((blastfile == "")) { m->mothurOut("When executing a mgcluster command you must provide a blastfile."); m->mothurOutEndLine(); abort = true; } //check for optional parameter and set defaults string temp; @@ -88,7 +88,7 @@ MGClusterCommand::MGClusterCommand(string option){ if (method == "not found") { method = "furthest"; } if ((method == "furthest") || (method == "nearest") || (method == "average")) { } - else { mothurOut("Not a valid clustering method. Valid clustering algorithms are furthest, nearest or average."); mothurOutEndLine(); abort = true; } + else { m->mothurOut("Not a valid clustering method. Valid clustering algorithms are furthest, nearest or average."); m->mothurOutEndLine(); abort = true; } temp = validParameter.validFile(parameters, "length", false); if (temp == "not found") { temp = "5"; } convert(temp, length); @@ -97,18 +97,21 @@ MGClusterCommand::MGClusterCommand(string option){ convert(temp, penalty); temp = validParameter.validFile(parameters, "min", false); if (temp == "not found") { temp = "true"; } - minWanted = isTrue(temp); + minWanted = m->isTrue(temp); temp = validParameter.validFile(parameters, "merge", false); if (temp == "not found") { temp = "true"; } - merge = isTrue(temp); + merge = m->isTrue(temp); temp = validParameter.validFile(parameters, "hcluster", false); if (temp == "not found") { temp = "false"; } - hclusterWanted = isTrue(temp); + hclusterWanted = m->isTrue(temp); + + temp = validParameter.validFile(parameters, "hard", false); if (temp == "not found") { temp = "F"; } + hard = m->isTrue(temp); } } catch(exception& e) { - errorOut(e, "MGClusterCommand", "MGClusterCommand"); + m->errorOut(e, "MGClusterCommand", "MGClusterCommand"); exit(1); } } @@ -116,23 +119,23 @@ MGClusterCommand::MGClusterCommand(string option){ void MGClusterCommand::help(){ try { - mothurOut("The mgcluster command parameter options are blast, name, cutoff, precision, method, merge, min, length, penalty and hcluster. The blast parameter is required.\n"); - mothurOut("The mgcluster command reads a blast and name file and clusters the sequences into OPF units similiar to the OTUs.\n"); - mothurOut("This command outputs a .list, .rabund and .sabund file that can be used with mothur other commands to estimate richness.\n"); - mothurOut("The cutoff parameter is used to specify the maximum distance you would like to cluster to. The default is 0.70.\n"); - mothurOut("The precision parameter's default value is 100. \n"); - mothurOut("The acceptable mgcluster methods are furthest, nearest and average. If no method is provided then furthest is assumed.\n\n"); - mothurOut("The min parameter allows you to specify is you want the minimum or maximum blast score ratio used in calculating the distance. The default is true, meaning you want the minimum.\n"); - mothurOut("The length parameter is used to specify the minimum overlap required. The default is 5.\n"); - mothurOut("The penalty parameter is used to adjust the error rate. The default is 0.10.\n"); - mothurOut("The merge parameter allows you to shut off merging based on overlaps and just cluster. By default merge is true, meaning you want to merge.\n"); - mothurOut("The hcluster parameter allows you to use the hcluster algorithm when clustering. This may be neccessary if your file is too large to fit into RAM. The default is false.\n"); - mothurOut("The mgcluster command should be in the following format: \n"); - mothurOut("mgcluster(blast=yourBlastfile, name=yourNameFile, cutoff=yourCutOff).\n"); - mothurOut("Note: No spaces between parameter labels (i.e. balst), '=' and parameters (i.e.yourBlastfile).\n\n"); + m->mothurOut("The mgcluster command parameter options are blast, name, cutoff, precision, method, merge, min, length, penalty and hcluster. The blast parameter is required.\n"); + m->mothurOut("The mgcluster command reads a blast and name file and clusters the sequences into OPF units similiar to the OTUs.\n"); + m->mothurOut("This command outputs a .list, .rabund and .sabund file that can be used with mothur other commands to estimate richness.\n"); + m->mothurOut("The cutoff parameter is used to specify the maximum distance you would like to cluster to. The default is 0.70.\n"); + m->mothurOut("The precision parameter's default value is 100. \n"); + m->mothurOut("The acceptable mgcluster methods are furthest, nearest and average. If no method is provided then furthest is assumed.\n\n"); + m->mothurOut("The min parameter allows you to specify is you want the minimum or maximum blast score ratio used in calculating the distance. The default is true, meaning you want the minimum.\n"); + m->mothurOut("The length parameter is used to specify the minimum overlap required. The default is 5.\n"); + m->mothurOut("The penalty parameter is used to adjust the error rate. The default is 0.10.\n"); + m->mothurOut("The merge parameter allows you to shut off merging based on overlaps and just cluster. By default merge is true, meaning you want to merge.\n"); + m->mothurOut("The hcluster parameter allows you to use the hcluster algorithm when clustering. This may be neccessary if your file is too large to fit into RAM. The default is false.\n"); + m->mothurOut("The mgcluster command should be in the following format: \n"); + m->mothurOut("mgcluster(blast=yourBlastfile, name=yourNameFile, cutoff=yourCutOff).\n"); + m->mothurOut("Note: No spaces between parameter labels (i.e. balst), '=' and parameters (i.e.yourBlastfile).\n\n"); } catch(exception& e) { - errorOut(e, "MGClusterCommand", "help"); + m->errorOut(e, "MGClusterCommand", "help"); exit(1); } } @@ -150,7 +153,7 @@ int MGClusterCommand::execute(){ nameMap->readMap(); }else{ nameMap= new NameAssignment(); } - string fileroot = outputDir + getRootName(getSimpleName(blastfile)); + string fileroot = outputDir + m->getRootName(m->getSimpleName(blastfile)); string tag = ""; time_t start; float previousDist = 0.00000; @@ -164,17 +167,27 @@ int MGClusterCommand::execute(){ list = new ListVector(nameMap->getListVector()); RAbundVector* rabund = new RAbundVector(list->getRAbundVector()); + if (m->control_pressed) { delete nameMap; delete read; delete list; delete rabund; return 0; } + start = time(NULL); oldList = *list; + map Seq2Bin; + map oldSeq2Bin; if (method == "furthest") { tag = "fn"; } else if (method == "nearest") { tag = "nn"; } else { tag = "an"; } //open output files - openOutputFile(fileroot+ tag + ".list", listFile); - openOutputFile(fileroot+ tag + ".rabund", rabundFile); - openOutputFile(fileroot+ tag + ".sabund", sabundFile); + m->openOutputFile(fileroot+ tag + ".list", listFile); + m->openOutputFile(fileroot+ tag + ".rabund", rabundFile); + m->openOutputFile(fileroot+ tag + ".sabund", sabundFile); + + if (m->control_pressed) { + delete nameMap; delete read; delete list; delete rabund; + listFile.close(); rabundFile.close(); sabundFile.close(); remove((fileroot+ tag + ".list").c_str()); remove((fileroot+ tag + ".rabund").c_str()); remove((fileroot+ tag + ".sabund").c_str()); + return 0; + } if (!hclusterWanted) { //get distmatrix and overlap @@ -183,17 +196,37 @@ int MGClusterCommand::execute(){ delete read; //create cluster - if (method == "furthest") { cluster = new CompleteLinkage(rabund, list, distMatrix, cutoff); } - else if(method == "nearest"){ cluster = new SingleLinkage(rabund, list, distMatrix, cutoff); } - else if(method == "average"){ cluster = new AverageLinkage(rabund, list, distMatrix, cutoff); } + if (method == "furthest") { cluster = new CompleteLinkage(rabund, list, distMatrix, cutoff, method); } + else if(method == "nearest"){ cluster = new SingleLinkage(rabund, list, distMatrix, cutoff, method); } + else if(method == "average"){ cluster = new AverageLinkage(rabund, list, distMatrix, cutoff, method); } cluster->setMapWanted(true); + Seq2Bin = cluster->getSeqtoBin(); + oldSeq2Bin = Seq2Bin; + if (m->control_pressed) { + delete nameMap; delete distMatrix; delete list; delete rabund; delete cluster; + listFile.close(); rabundFile.close(); sabundFile.close(); remove((fileroot+ tag + ".list").c_str()); remove((fileroot+ tag + ".rabund").c_str()); remove((fileroot+ tag + ".sabund").c_str()); + return 0; + } + //cluster using cluster classes while (distMatrix->getSmallDist() < cutoff && distMatrix->getNNodes() > 0){ - cluster->update(); + cluster->update(cutoff); + + if (m->control_pressed) { + delete nameMap; delete distMatrix; delete list; delete rabund; delete cluster; + listFile.close(); rabundFile.close(); sabundFile.close(); remove((fileroot+ tag + ".list").c_str()); remove((fileroot+ tag + ".rabund").c_str()); remove((fileroot+ tag + ".sabund").c_str()); + return 0; + } + float dist = distMatrix->getSmallDist(); - float rndDist = roundDist(dist, precision); + float rndDist; + if (hard) { + rndDist = m->ceilDist(dist, precision); + }else{ + rndDist = m->roundDist(dist, precision); + } if(previousDist <= 0.0000 && dist != previousDist){ oldList.setLabel("unique"); @@ -201,8 +234,14 @@ int MGClusterCommand::execute(){ } else if(rndDist != rndPreviousDist){ if (merge) { - map seq2Bin = cluster->getSeqtoBin(); - ListVector* temp = mergeOPFs(seq2Bin, rndPreviousDist); + ListVector* temp = mergeOPFs(oldSeq2Bin, rndPreviousDist); + + if (m->control_pressed) { + delete nameMap; delete distMatrix; delete list; delete rabund; delete cluster; delete temp; + listFile.close(); rabundFile.close(); sabundFile.close(); remove((fileroot+ tag + ".list").c_str()); remove((fileroot+ tag + ".rabund").c_str()); remove((fileroot+ tag + ".sabund").c_str()); + return 0; + } + temp->setLabel(toString(rndPreviousDist, precisionLength-1)); printData(temp); delete temp; @@ -211,10 +250,12 @@ int MGClusterCommand::execute(){ printData(&oldList); } } - + previousDist = dist; rndPreviousDist = rndDist; oldList = *list; + Seq2Bin = cluster->getSeqtoBin(); + oldSeq2Bin = Seq2Bin; } if(previousDist <= 0.0000){ @@ -223,8 +264,14 @@ int MGClusterCommand::execute(){ } else if(rndPreviousDist seq2Bin = cluster->getSeqtoBin(); - ListVector* temp = mergeOPFs(seq2Bin, rndPreviousDist); + ListVector* temp = mergeOPFs(oldSeq2Bin, rndPreviousDist); + + if (m->control_pressed) { + delete nameMap; delete distMatrix; delete list; delete rabund; delete cluster; delete temp; + listFile.close(); rabundFile.close(); sabundFile.close(); remove((fileroot+ tag + ".list").c_str()); remove((fileroot+ tag + ".rabund").c_str()); remove((fileroot+ tag + ".sabund").c_str()); + return 0; + } + temp->setLabel(toString(rndPreviousDist, precisionLength-1)); printData(temp); delete temp; @@ -247,26 +294,61 @@ int MGClusterCommand::execute(){ //sort the distance and overlap files sortHclusterFiles(distFile, overlapFile); + + if (m->control_pressed) { + delete nameMap; delete list; delete rabund; + listFile.close(); rabundFile.close(); sabundFile.close(); remove((fileroot+ tag + ".list").c_str()); remove((fileroot+ tag + ".rabund").c_str()); remove((fileroot+ tag + ".sabund").c_str()); + return 0; + } //create cluster hcluster = new HCluster(rabund, list, method, distFile, nameMap, cutoff); hcluster->setMapWanted(true); + Seq2Bin = cluster->getSeqtoBin(); + oldSeq2Bin = Seq2Bin; vector seqs; seqs.resize(1); // to start loop //ifstream inHcluster; - //openInputFile(distFile, inHcluster); + //m->openInputFile(distFile, inHcluster); + + if (m->control_pressed) { + delete nameMap; delete list; delete rabund; delete hcluster; + listFile.close(); rabundFile.close(); sabundFile.close(); remove((fileroot+ tag + ".list").c_str()); remove((fileroot+ tag + ".rabund").c_str()); remove((fileroot+ tag + ".sabund").c_str()); + return 0; + } while (seqs.size() != 0){ seqs = hcluster->getSeqs(); + if (m->control_pressed) { + delete nameMap; delete list; delete rabund; delete hcluster; + listFile.close(); rabundFile.close(); sabundFile.close(); remove((fileroot+ tag + ".list").c_str()); remove((fileroot+ tag + ".rabund").c_str()); remove((fileroot+ tag + ".sabund").c_str()); + remove(distFile.c_str()); + remove(overlapFile.c_str()); + return 0; + } + for (int i = 0; i < seqs.size(); i++) { //-1 means skip me if (seqs[i].seq1 != seqs[i].seq2) { hcluster->update(seqs[i].seq1, seqs[i].seq2, seqs[i].dist); + + if (m->control_pressed) { + delete nameMap; delete list; delete rabund; delete hcluster; + listFile.close(); rabundFile.close(); sabundFile.close(); remove((fileroot+ tag + ".list").c_str()); remove((fileroot+ tag + ".rabund").c_str()); remove((fileroot+ tag + ".sabund").c_str()); + remove(distFile.c_str()); + remove(overlapFile.c_str()); + return 0; + } - float rndDist = roundDist(seqs[i].dist, precision); + float rndDist; + if (hard) { + rndDist = m->ceilDist(seqs[i].dist, precision); + }else{ + rndDist = m->roundDist(seqs[i].dist, precision); + } if((previousDist <= 0.0000) && (seqs[i].dist != previousDist)){ oldList.setLabel("unique"); @@ -274,8 +356,16 @@ int MGClusterCommand::execute(){ } else if((rndDist != rndPreviousDist)){ if (merge) { - map seq2Bin = hcluster->getSeqtoBin(); - ListVector* temp = mergeOPFs(seq2Bin, rndPreviousDist); + ListVector* temp = mergeOPFs(oldSeq2Bin, rndPreviousDist); + + if (m->control_pressed) { + delete nameMap; delete list; delete rabund; delete hcluster; delete temp; + listFile.close(); rabundFile.close(); sabundFile.close(); remove((fileroot+ tag + ".list").c_str()); remove((fileroot+ tag + ".rabund").c_str()); remove((fileroot+ tag + ".sabund").c_str()); + remove(distFile.c_str()); + remove(overlapFile.c_str()); + return 0; + } + temp->setLabel(toString(rndPreviousDist, precisionLength-1)); printData(temp); delete temp; @@ -288,6 +378,8 @@ int MGClusterCommand::execute(){ previousDist = seqs[i].dist; rndPreviousDist = rndDist; oldList = *list; + Seq2Bin = cluster->getSeqtoBin(); + oldSeq2Bin = Seq2Bin; } } } @@ -299,8 +391,16 @@ int MGClusterCommand::execute(){ } else if(rndPreviousDist seq2Bin = hcluster->getSeqtoBin(); - ListVector* temp = mergeOPFs(seq2Bin, rndPreviousDist); + ListVector* temp = mergeOPFs(oldSeq2Bin, rndPreviousDist); + + if (m->control_pressed) { + delete nameMap; delete list; delete rabund; delete hcluster; delete temp; + listFile.close(); rabundFile.close(); sabundFile.close(); remove((fileroot+ tag + ".list").c_str()); remove((fileroot+ tag + ".rabund").c_str()); remove((fileroot+ tag + ".sabund").c_str()); + remove(distFile.c_str()); + remove(overlapFile.c_str()); + return 0; + } + temp->setLabel(toString(rndPreviousDist, precisionLength-1)); printData(temp); delete temp; @@ -323,13 +423,28 @@ int MGClusterCommand::execute(){ globaldata->setListFile(fileroot+ tag + ".list"); globaldata->setFormat("list"); - - mothurOut("It took " + toString(time(NULL) - start) + " seconds to cluster."); mothurOutEndLine(); + + if (m->control_pressed) { + delete nameMap; + listFile.close(); rabundFile.close(); sabundFile.close(); remove((fileroot+ tag + ".list").c_str()); remove((fileroot+ tag + ".rabund").c_str()); remove((fileroot+ tag + ".sabund").c_str()); + globaldata->setListFile(""); + globaldata->setFormat(""); + return 0; + } + + m->mothurOutEndLine(); + m->mothurOut("Output File Names: "); m->mothurOutEndLine(); + m->mothurOut(fileroot+ tag + ".list"); m->mothurOutEndLine(); + m->mothurOut(fileroot+ tag + ".rabund"); m->mothurOutEndLine(); + m->mothurOut(fileroot+ tag + ".sabund"); m->mothurOutEndLine(); + m->mothurOutEndLine(); + + m->mothurOut("It took " + toString(time(NULL) - start) + " seconds to cluster."); m->mothurOutEndLine(); return 0; } catch(exception& e) { - errorOut(e, "MGClusterCommand", "execute"); + m->errorOut(e, "MGClusterCommand", "execute"); exit(1); } } @@ -345,7 +460,7 @@ void MGClusterCommand::printData(ListVector* mergedList){ sabund.print(sabundFile); } catch(exception& e) { - errorOut(e, "MGClusterCommand", "printData"); + m->errorOut(e, "MGClusterCommand", "printData"); exit(1); } } @@ -356,16 +471,21 @@ ListVector* MGClusterCommand::mergeOPFs(map binInfo, float dist){ try { //create new listvector so you don't overwrite the clustering ListVector* newList = new ListVector(oldList); + bool done = false; ifstream inOverlap; int count = 0; if (hclusterWanted) { - openInputFile(overlapFile, inOverlap); + m->openInputFile(overlapFile, inOverlap); if (inOverlap.eof()) { done = true; } }else { if (overlapMatrix.size() == 0) { done = true; } } while (!done) { + if (m->control_pressed) { + if (hclusterWanted) { inOverlap.close(); } + return newList; + } //get next overlap seqDist overlapNode; @@ -378,15 +498,18 @@ ListVector* MGClusterCommand::mergeOPFs(map binInfo, float dist){ if (!inOverlap.eof()) { string firstName, secondName; float overlapDistance; - inOverlap >> firstName >> secondName >> overlapDistance; gobble(inOverlap); + inOverlap >> firstName >> secondName >> overlapDistance; m->gobble(inOverlap); - map::iterator itA = nameMap->find(firstName); - map::iterator itB = nameMap->find(secondName); - if(itA == nameMap->end()){ cerr << "AAError: Sequence '" << firstName << "' was not found in the names file, please correct\n"; exit(1); } - if(itB == nameMap->end()){ cerr << "ABError: Sequence '" << secondName << "' was not found in the names file, please correct\n"; exit(1); } + //commented out because we check this in readblast already + //map::iterator itA = nameMap->find(firstName); + //map::iterator itB = nameMap->find(secondName); + //if(itA == nameMap->end()){ cerr << "AAError: Sequence '" << firstName << "' was not found in the names file, please correct\n"; exit(1); } + //if(itB == nameMap->end()){ cerr << "ABError: Sequence '" << secondName << "' was not found in the names file, please correct\n"; exit(1); } - overlapNode.seq1 = itA->second; - overlapNode.seq2 = itB->second; + //overlapNode.seq1 = itA->second; + //overlapNode.seq2 = itB->second; + overlapNode.seq1 = nameMap->get(firstName); + overlapNode.seq2 = nameMap->get(secondName); overlapNode.dist = overlapDistance; }else { inOverlap.close(); break; } } @@ -395,16 +518,26 @@ ListVector* MGClusterCommand::mergeOPFs(map binInfo, float dist){ //get names of seqs that overlap string name1 = nameMap->get(overlapNode.seq1); string name2 = nameMap->get(overlapNode.seq2); - + //use binInfo to find out if they are already in the same bin + //map::iterator itBin1 = binInfo.find(name1); + //map::iterator itBin2 = binInfo.find(name2); + + //if(itBin1 == binInfo.end()){ cerr << "AAError: Sequence '" << name1 << "' does not have any bin info.\n"; exit(1); } + //if(itBin2 == binInfo.end()){ cerr << "ABError: Sequence '" << name2 << "' does not have any bin info.\n"; exit(1); } + + //int binKeep = itBin1->second; + //int binRemove = itBin2->second; + int binKeep = binInfo[name1]; int binRemove = binInfo[name2]; - + //if not merge bins and update binInfo if(binKeep != binRemove) { + //save names in old bin - string names = list->get(binRemove); - + string names = newList->get(binRemove); + //merge bins into name1s bin newList->set(binKeep, newList->get(binRemove)+','+newList->get(binKeep)); newList->set(binRemove, ""); @@ -430,7 +563,7 @@ ListVector* MGClusterCommand::mergeOPFs(map binInfo, float dist){ } catch(exception& e) { - errorOut(e, "MGClusterCommand", "mergeOPFs"); + m->errorOut(e, "MGClusterCommand", "mergeOPFs"); exit(1); } } @@ -438,17 +571,17 @@ ListVector* MGClusterCommand::mergeOPFs(map binInfo, float dist){ void MGClusterCommand::sortHclusterFiles(string unsortedDist, string unsortedOverlap) { try { //sort distFile - string sortedDistFile = sortFile(unsortedDist); + string sortedDistFile = m->sortFile(unsortedDist, outputDir); remove(unsortedDist.c_str()); //delete unsorted file distFile = sortedDistFile; //sort overlap file - string sortedOverlapFile = sortFile(unsortedOverlap); + string sortedOverlapFile = m->sortFile(unsortedOverlap, outputDir); remove(unsortedOverlap.c_str()); //delete unsorted file overlapFile = sortedOverlapFile; } catch(exception& e) { - errorOut(e, "MGClusterCommand", "sortHclusterFiles"); + m->errorOut(e, "MGClusterCommand", "sortHclusterFiles"); exit(1); } }