X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=hclustercommand.cpp;h=b991ccd6289b9afd44628c12bf451eb86529fa44;hp=38ba4b6cd9839205aef8ff01887cdcf7f2b2ecef;hb=d1c97b8c04bb75faca1e76ffad60b37a4d789d3d;hpb=72ebb6fa35b45b149812d47c2b1cb8acaca64659 diff --git a/hclustercommand.cpp b/hclustercommand.cpp index 38ba4b6..b991ccd 100644 --- a/hclustercommand.cpp +++ b/hclustercommand.cpp @@ -9,55 +9,187 @@ #include "hclustercommand.h" +//********************************************************************************************************************** +vector HClusterCommand::setParameters(){ + try { + CommandParameter pphylip("phylip", "InputTypes", "", "", "PhylipColumn", "PhylipColumn", "none","list-rabund-sabund",false,false,true); parameters.push_back(pphylip); + CommandParameter pname("name", "InputTypes", "", "", "none", "none", "ColumnName","",false,false,true); parameters.push_back(pname); + CommandParameter pcolumn("column", "InputTypes", "", "", "PhylipColumn", "PhylipColumn", "ColumnName","list-rabund-sabund",false,false,true); parameters.push_back(pcolumn); + CommandParameter pcutoff("cutoff", "Number", "", "10", "", "", "","",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 psorted("sorted", "Boolean", "", "F", "", "", "","",false,false); parameters.push_back(psorted); + CommandParameter pshowabund("showabund", "Boolean", "", "T", "", "", "","",false,false); parameters.push_back(pshowabund); + CommandParameter ptiming("timing", "Boolean", "", "F", "", "", "","",false,false); parameters.push_back(ptiming); + 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, "HClusterCommand", "setParameters"); + exit(1); + } +} +//********************************************************************************************************************** +string HClusterCommand::getHelpString(){ + try { + string helpString = ""; + helpString += "The hcluster command parameter options are cutoff, precision, method, phylip, column, name, showabund, timing and sorted. Phylip or column and name are required, unless you have valid current files.\n"; + helpString += "The phylip and column parameter allow you to enter your distance file, and sorted indicates whether your column distance file is already sorted. \n"; + helpString += "The name parameter allows you to enter your name file and is required if your distance file is in column format. \n"; + helpString += "The hcluster command should be in the following format: \n"; + helpString += "hcluster(column=youDistanceFile, name=yourNameFile, method=yourMethod, cutoff=yourCutoff, precision=yourPrecision) \n"; + helpString += "The acceptable hcluster methods are furthest, nearest, weighted and average.\n"; + return helpString; + } + catch(exception& e) { + m->errorOut(e, "HClusterCommand", "getHelpString"); + exit(1); + } +} +//********************************************************************************************************************** +string HClusterCommand::getOutputPattern(string type) { + try { + string pattern = ""; + + if (type == "list") { pattern = "[filename],[clustertag],list"; } + else if (type == "rabund") { pattern = "[filename],[clustertag],rabund"; } + else if (type == "sabund") { pattern = "[filename],[clustertag],sabund"; } + else { m->mothurOut("[ERROR]: No definition for type " + type + " output pattern.\n"); m->control_pressed = true; } + + return pattern; + } + catch(exception& e) { + m->errorOut(e, "HClusterCommand", "getOutputPattern"); + exit(1); + } +} + +//********************************************************************************************************************** +HClusterCommand::HClusterCommand(){ + try { + abort = true; calledHelp = true; + setParameters(); + vector tempOutNames; + outputTypes["list"] = tempOutNames; + outputTypes["rabund"] = tempOutNames; + outputTypes["sabund"] = tempOutNames; + } + catch(exception& e) { + m->errorOut(e, "HClusterCommand", "HClusterCommand"); + exit(1); + } +} //********************************************************************************************************************** //This function checks to make sure the cluster command has no errors and then clusters based on the method chosen. -HClusterCommand::HClusterCommand(string option){ +HClusterCommand::HClusterCommand(string option) { try{ - globaldata = GlobalData::getInstance(); - abort = false; + abort = false; calledHelp = false; //allow user to run help - if(option == "help") { help(); abort = true; } + if(option == "help") { help(); abort = true; calledHelp = true; } + else if(option == "citation") { citation(); abort = true; calledHelp = true;} else { - //valid paramters for this command - string Array[] = {"cutoff","precision","method","showabund","timing","phylip","column","name","sorted"}; - vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); + vector myArray = setParameters(); OptionParser parser(option); map parameters = parser.getParameters(); ValidParameters validParameter; + map::iterator it; //check to make sure all parameters are valid for command - for (map::iterator it = parameters.begin(); it != parameters.end(); it++) { + for (it = parameters.begin(); it != parameters.end(); it++) { if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; } } - globaldata->newRead(); + //initialize outputTypes + vector tempOutNames; + outputTypes["list"] = tempOutNames; + outputTypes["rabund"] = tempOutNames; + outputTypes["sabund"] = tempOutNames; + + //if the user changes the input directory command factory will send this info to us in the output parameter + string inputDir = validParameter.validFile(parameters, "inputdir", false); + if (inputDir == "not found"){ inputDir = ""; } + else { + string path; + it = parameters.find("phylip"); + //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["phylip"] = inputDir + it->second; } + } + + it = parameters.find("column"); + //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["column"] = inputDir + it->second; } + } + + it = parameters.find("name"); + //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["name"] = inputDir + it->second; } + } + } + + //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 = ""; } //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 = ""; } + else { m->setNameFile(namefile); } - if ((phylipfile == "") && (columnfile == "")) { mothurOut("When executing a cluster command you must enter a phylip or a column."); mothurOutEndLine(); abort = true; } - else if ((phylipfile != "") && (columnfile != "")) { mothurOut("When executing a cluster command you must enter ONLY ONE of the following: phylip or column."); mothurOutEndLine(); abort = true; } + if ((phylipfile == "") && (columnfile == "")) { + //is there are current file available for either of these? + //give priority to column, then phylip + 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 { + m->mothurOut("No valid current files. You must provide a phylip or column file before you can use the hcluster command."); m->mothurOutEndLine(); + abort = true; + } + } + } + else if ((phylipfile != "") && (columnfile != "")) { m->mothurOut("When executing a hcluster command you must enter ONLY ONE of the following: phylip or column."); m->mothurOutEndLine(); abort = true; } if (columnfile != "") { - if (namefile == "") { cout << "You need to provide a namefile if you are going to use the column format." << endl; abort = true; } + if (namefile == "") { + namefile = m->getNameFile(); + if (namefile != "") { m->mothurOut("Using " + namefile + " as input file for the name parameter."); m->mothurOutEndLine(); } + else { + m->mothurOut("You need to provide a namefile if you are going to use the column format."); m->mothurOutEndLine(); + abort = true; + } + } } //check for optional parameter and set defaults @@ -68,103 +200,122 @@ HClusterCommand::HClusterCommand(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 = "T"; } + hard = m->isTrue(temp); temp = validParameter.validFile(parameters, "cutoff", false); if (temp == "not found") { temp = "10"; } - convert(temp, cutoff); - cutoff += (5 / (precision * 10.0)); + m->mothurConvert(temp, cutoff); + cutoff += (5 / (precision * 10.0)); method = validParameter.validFile(parameters, "method", false); - if (method == "not found") { method = "nearest"; } + if (method == "not found") { method = "average"; } - 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; } + if ((method == "furthest") || (method == "nearest") || (method == "average") || (method == "weighted")) { } + else { m->mothurOut("Not a valid clustering method. Valid clustering algorithms are furthest, nearest, average or weighted."); m->mothurOutEndLine(); abort = true; } showabund = validParameter.validFile(parameters, "showabund", false); if (showabund == "not found") { showabund = "T"; } sort = validParameter.validFile(parameters, "sorted", false); if (sort == "not found") { sort = "F"; } - sorted = isTrue(sort); + sorted = m->isTrue(sort); timing = validParameter.validFile(parameters, "timing", false); if (timing == "not found") { timing = "F"; } if (abort == false) { - - fileroot = getRootName(distfile); - tag = "fn"; //until we figure out average and nearest methods + if (outputDir == "") { outputDir += m->hasPath(distfile); } + fileroot = outputDir + m->getRootName(m->getSimpleName(distfile)); + + if (method == "furthest") { tag = "fn"; } + else if (method == "nearest") { tag = "nn"; } + else if (method == "weighted") { tag = "wn"; } + else { tag = "an"; } + + map variables; + variables["[filename]"] = fileroot; + variables["[clustertag]"] = tag; - openOutputFile(fileroot+ tag + ".sabund", sabundFile); - openOutputFile(fileroot+ tag + ".rabund", rabundFile); - openOutputFile(fileroot+ tag + ".list", listFile); + string sabundFileName = getOutputFileName("sabund",variables); + string rabundFileName = getOutputFileName("rabund",variables); + string listFileName = getOutputFileName("list", variables); + + m->openOutputFile(sabundFileName, sabundFile); + m->openOutputFile(rabundFileName, rabundFile); + m->openOutputFile(listFileName, listFile); + + outputNames.push_back(sabundFileName); outputTypes["sabund"].push_back(sabundFileName); + outputNames.push_back(rabundFileName); outputTypes["rabund"].push_back(rabundFileName); + outputNames.push_back(listFileName); outputTypes["list"].push_back(listFileName); } } } catch(exception& e) { - errorOut(e, "HClusterCommand", "HClusterCommand"); + m->errorOut(e, "HClusterCommand", "HClusterCommand"); exit(1); } } //********************************************************************************************************************** -void HClusterCommand::help(){ - try { - mothurOut("The cluster command can only be executed after a successful read.dist command.\n"); - mothurOut("The cluster command parameter options are method, cuttoff, precision, showabund and timing. No parameters are required.\n"); - mothurOut("The cluster command should be in the following format: \n"); - mothurOut("cluster(method=yourMethod, cutoff=yourCutoff, precision=yourPrecision) \n"); - mothurOut("The acceptable cluster methods are furthest, nearest and average. If no method is provided then furthest is assumed.\n\n"); - } - catch(exception& e) { - errorOut(e, "HClusterCommand", "help"); - exit(1); - } -} - -//********************************************************************************************************************** - -HClusterCommand::~HClusterCommand(){} - -//********************************************************************************************************************** - int HClusterCommand::execute(){ try { - if (abort == true) { return 0; } + if (abort == true) { if (calledHelp) { return 0; } return 2; } + NameAssignment* nameMap = NULL; if(namefile != ""){ - globaldata->nameMap = new NameAssignment(namefile); - globaldata->nameMap->readMap(); - }else{ - globaldata->nameMap = NULL; - } + nameMap = new NameAssignment(namefile); + nameMap->readMap(); + } + + time_t estart = time(NULL); if (!sorted) { - read = new ReadCluster(distfile, cutoff); + read = new ReadCluster(distfile, cutoff, outputDir, true); read->setFormat(format); - read->read(globaldata->nameMap); + read->read(nameMap); + + if (m->control_pressed) { + delete read; + sabundFile.close(); + rabundFile.close(); + listFile.close(); + for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } outputTypes.clear(); + return 0; + } + distfile = read->getOutputFile(); list = read->getListVector(); delete read; }else { - list = new ListVector(globaldata->nameMap->getListVector()); + list = new ListVector(nameMap->getListVector()); + } + + if (m->control_pressed) { + sabundFile.close(); + rabundFile.close(); + listFile.close(); + for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } outputTypes.clear(); + return 0; } + + m->mothurOut("It took " + toString(time(NULL) - estart) + " seconds to sort. "); m->mothurOutEndLine(); + estart = time(NULL); //list vector made by read contains all sequence names if(list != NULL){ rabund = new RAbundVector(list->getRAbundVector()); }else{ - mothurOut("Error: no list vector!"); mothurOutEndLine(); return 0; + m->mothurOut("Error: no list vector!"); m->mothurOutEndLine(); return 0; } - - float previousDist = 0.00000; float rndPreviousDist = 0.00000; oldRAbund = *rabund; @@ -172,42 +323,61 @@ int HClusterCommand::execute(){ print_start = true; start = time(NULL); - -//cout << "here" << endl; - ifstream in; - openInputFile(distfile, in); - string firstName, secondName; - float distance; - - cluster = new HCluster(rabund, list); - bool clusteredSomething; + + cluster = new HCluster(rabund, list, method, distfile, nameMap, cutoff); vector seqs; seqs.resize(1); // to start loop - exitedBreak = false; //lets you know if there is a distance stored in next + + if (m->control_pressed) { + delete cluster; + sabundFile.close(); + rabundFile.close(); + listFile.close(); + for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } outputTypes.clear(); + return 0; + } + + float saveCutoff = cutoff; while (seqs.size() != 0){ - seqs = getSeqs(in); - random_shuffle(seqs.begin(), seqs.end()); + seqs = cluster->getSeqs(); - if (seqs.size() == 0) { break; } //there are no more distances - - for (int i = 0; i < seqs.size(); i++) { //-1 means skip me + //to account for cutoff change in average neighbor + if (seqs.size() != 0) { + if (seqs[0].dist > cutoff) { break; } + } + + if (m->control_pressed) { + delete cluster; + sabundFile.close(); + rabundFile.close(); + listFile.close(); + for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } outputTypes.clear(); + return 0; + } - if (print_start && isTrue(timing)) { - mothurOut("Clustering (" + tag + ") dist " + toString(distance) + "/" - + toString(roundDist(distance, precision)) - + "\t(precision: " + toString(precision) + ")"); - cout.flush(); - print_start = false; - } + for (int i = 0; i < seqs.size(); i++) { //-1 means skip me - ///cout << "before cluster update" << endl; if (seqs[i].seq1 != seqs[i].seq2) { - clusteredSomething = cluster->update(seqs[i].seq1, seqs[i].seq2, seqs[i].dist); - - float rndDist = roundDist(seqs[i].dist, precision); - //cout << "after cluster update clusterSomething = " << clusteredSomething << " rndDist = " << rndDist << " rndPreviousDist = " << rndPreviousDist << endl; + cutoff = cluster->update(seqs[i].seq1, seqs[i].seq2, seqs[i].dist); + if (m->control_pressed) { + delete cluster; + sabundFile.close(); + rabundFile.close(); + listFile.close(); + for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } outputTypes.clear(); + return 0; + } + + + 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)){ printData("unique"); @@ -215,7 +385,7 @@ int HClusterCommand::execute(){ else if((rndDist != rndPreviousDist)){ printData(toString(rndPreviousDist, length-1)); } - + previousDist = seqs[i].dist; rndPreviousDist = rndDist; oldRAbund = *rabund; @@ -223,44 +393,72 @@ int HClusterCommand::execute(){ } } } - - in.close(); - if (print_start && isTrue(timing)) { - //mothurOut("Clustering (" + tag + ") for distance " + toString(previousDist) + "/" + toString(rndPreviousDist) - //+ "\t(precision: " + toString(precision) + ", Nodes: " + toString(matrix->getNNodes()) + ")"); - cout.flush(); - print_start = false; + if (m->control_pressed) { + delete cluster; + sabundFile.close(); + rabundFile.close(); + listFile.close(); + for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } outputTypes.clear(); + return 0; } - + if(previousDist <= 0.0000){ printData("unique"); } else if(rndPreviousDistcontrol_pressed) { + for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } outputTypes.clear(); + return 0; + } - //delete globaldata's copy of the sparsematrix and listvector to free up memory - delete globaldata->gListVector; globaldata->gListVector = NULL; - //saves .list file so you can do the collect, rarefaction and summary commands without doing a read.list - if (globaldata->getFormat() == "phylip") { globaldata->setPhylipFile(""); } - else if (globaldata->getFormat() == "column") { globaldata->setColumnFile(""); } + if (saveCutoff != cutoff) { + if (hard) { saveCutoff = m->ceilDist(saveCutoff, precision); } + else { saveCutoff = m->roundDist(saveCutoff, precision); } + + m->mothurOut("changed cutoff to " + toString(cutoff)); m->mothurOutEndLine(); + } - globaldata->setListFile(fileroot+ tag + ".list"); - globaldata->setNameFile(""); - globaldata->setFormat("list"); + //set list file as new current listfile + string current = ""; + itTypes = outputTypes.find("list"); + if (itTypes != outputTypes.end()) { + if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setListFile(current); } + } - sabundFile.close(); - rabundFile.close(); - listFile.close(); - if (isTrue(timing)) { - //mothurOut("It took " + toString(time(NULL) - estart) + " seconds to cluster " + toString(ndist) + " distances"); mothurOutEndLine(); + //set rabund file as new current rabundfile + itTypes = outputTypes.find("rabund"); + if (itTypes != outputTypes.end()) { + if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setRabundFile(current); } + } + + //set sabund file as new current sabundfile + itTypes = outputTypes.find("sabund"); + if (itTypes != outputTypes.end()) { + if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setSabundFile(current); } } + + + m->mothurOutEndLine(); + m->mothurOut("Output File Names: "); m->mothurOutEndLine(); + for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); } + m->mothurOutEndLine(); + + m->mothurOut("It took " + toString(time(NULL) - estart) + " seconds to cluster. "); m->mothurOutEndLine(); + return 0; } catch(exception& e) { - errorOut(e, "HClusterCommand", "execute"); + m->errorOut(e, "HClusterCommand", "execute"); exit(1); } } @@ -269,16 +467,16 @@ int HClusterCommand::execute(){ void HClusterCommand::printData(string label){ try { - if (isTrue(timing)) { - mothurOut("\tTime: " + toString(time(NULL) - start) + "\tsecs for " + toString(oldRAbund.getNumBins()) - + "\tclusters. Updates: " + toString(loops)); mothurOutEndLine(); + if (m->isTrue(timing)) { + m->mothurOut("\tTime: " + toString(time(NULL) - start) + "\tsecs for " + toString(oldRAbund.getNumBins()) + + "\tclusters. Updates: " + toString(loops)); m->mothurOutEndLine(); } print_start = true; loops = 0; start = time(NULL); oldRAbund.setLabel(label); - if (isTrue(showabund)) { + if (m->isTrue(showabund)) { oldRAbund.getSAbundVector().print(cout); } oldRAbund.print(rabundFile); @@ -288,7 +486,7 @@ void HClusterCommand::printData(string label){ oldList.print(listFile); } catch(exception& e) { - errorOut(e, "HClusterCommand", "printData"); + m->errorOut(e, "HClusterCommand", "printData"); exit(1); } @@ -296,75 +494,3 @@ void HClusterCommand::printData(string label){ } //********************************************************************************************************************** -vector HClusterCommand::getSeqs(ifstream& filehandle){ - try { - string firstName, secondName; - float distance, prevDistance; - vector sameSeqs; - prevDistance = -1; - - //if you are not at the beginning of the file - if (exitedBreak) { - sameSeqs.push_back(next); - prevDistance = next.dist; - exitedBreak = false; - } - - //get entry - while (filehandle) { - - filehandle >> firstName >> secondName >> distance; -//cout << firstName << '\t' << secondName << '\t' << distance << endl; - gobble(filehandle); - - //save first one - if (prevDistance == -1) { prevDistance = distance; } - - map::iterator itA = globaldata->nameMap->find(firstName); - map::iterator itB = globaldata->nameMap->find(secondName); - - if(itA == globaldata->nameMap->end()){ - cerr << "AAError: Sequence '" << firstName << "' was not found in the names file, please correct\n"; - } - if(itB == globaldata->nameMap->end()){ - cerr << "ABError: Sequence '" << secondName << "' was not found in the names file, please correct\n"; - } - - //using cutoff - if (distance > cutoff) { break; } - - if (distance != -1) { //-1 means skip me - - //are the distances the same - if (distance == prevDistance) { //save in vector - seqDist temp; - temp.seq1 = itA->second; - temp.seq2 = itB->second; - temp.dist = distance; - sameSeqs.push_back(temp); - exitedBreak = false; - //what about precision?? - - }else{ - next.seq1 = itA->second; - next.seq2 = itB->second; - next.dist = distance; - exitedBreak = true; - break; - } - - } - } - - return sameSeqs; - } - catch(exception& e) { - errorOut(e, "HClusterCommand", "getSeqs"); - exit(1); - } - - -} - -//********************************************************************************************************************** -