X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=treegroupscommand.cpp;h=753b5cbe3cf7f67ed623bb8d866b848c6c96e5be;hp=9c56d2228092a25779cd0431c5a3a5ef06f81563;hb=cf9987b67aa49777a4c91c2d21f96e58bf17aa82;hpb=0cc0c01eb5127ef2b09b894e1f224ccc1d70bef0 diff --git a/treegroupscommand.cpp b/treegroupscommand.cpp index 9c56d22..753b5cb 100644 --- a/treegroupscommand.cpp +++ b/treegroupscommand.cpp @@ -8,63 +8,29 @@ */ #include "treegroupscommand.h" -#include "sharedsobscollectsummary.h" -#include "sharedchao1.h" -#include "sharedace.h" -#include "sharednseqs.h" -#include "sharedjabund.h" -#include "sharedsorabund.h" -#include "sharedjclass.h" -#include "sharedsorclass.h" -#include "sharedjest.h" -#include "sharedsorest.h" -#include "sharedthetayc.h" -#include "sharedthetan.h" -#include "sharedkstest.h" -#include "whittaker.h" -#include "sharedochiai.h" -#include "sharedanderbergs.h" -#include "sharedkulczynski.h" -#include "sharedkulczynskicody.h" -#include "sharedlennon.h" -#include "sharedmorisitahorn.h" -#include "sharedbraycurtis.h" -#include "sharedjackknife.h" -#include "whittaker.h" -#include "odum.h" -#include "canberra.h" -#include "structeuclidean.h" -#include "structchord.h" -#include "hellinger.h" -#include "manhattan.h" -#include "structpearson.h" -#include "soergel.h" -#include "spearman.h" -#include "structkulczynski.h" -#include "structchi2.h" -#include "speciesprofile.h" -#include "hamming.h" -#include "gower.h" -#include "memchi2.h" -#include "memchord.h" -#include "memeuclidean.h" -#include "mempearson.h" +#include "subsample.h" +#include "consensus.h" //********************************************************************************************************************** vector TreeGroupCommand::setParameters(){ try { - CommandParameter pshared("shared", "InputTypes", "", "", "PhylipColumnShared", "PhylipColumnShared", "none",false,false); parameters.push_back(pshared); - CommandParameter pphylip("phylip", "InputTypes", "", "", "PhylipColumnShared", "PhylipColumnShared", "none",false,false); parameters.push_back(pphylip); - CommandParameter pname("name", "InputTypes", "", "", "none", "none", "ColumnName",false,false); parameters.push_back(pname); - CommandParameter pcolumn("column", "InputTypes", "", "", "PhylipColumnShared", "PhylipColumnShared", "ColumnName",false,false); parameters.push_back(pcolumn); - CommandParameter pcutoff("cutoff", "Number", "", "10", "", "", "",false,false); parameters.push_back(pcutoff); - CommandParameter pprecision("precision", "Number", "", "100", "", "", "",false,false); parameters.push_back(pprecision); - CommandParameter plabel("label", "String", "", "", "", "", "",false,false); parameters.push_back(plabel); - CommandParameter pgroups("groups", "String", "", "", "", "", "",false,false); parameters.push_back(pgroups); - CommandParameter pcalc("calc", "Multiple", "sharedsobs-sharedchao-sharedace-jabund-sorabund-jclass-sorclass-jest-sorest-thetayc-thetan-kstest-sharednseqs-ochiai-anderberg-kulczynski-kulczynskicody-lennon-morisitahorn-braycurtis-whittaker-odum-canberra-structeuclidean-structchord-hellinger-manhattan-structpearson-soergel-spearman-structkulczynski-speciesprofile-hamming-structchi2-gower-memchi2-memchord-memeuclidean-mempearson", "jclass-thetayc", "", "", "",true,false); parameters.push_back(pcalc); - CommandParameter poutput("output", "Multiple", "lt-square", "lt", "", "", "",false,false); parameters.push_back(poutput); - CommandParameter pinputdir("inputdir", "String", "", "", "", "", "",false,false); parameters.push_back(pinputdir); - CommandParameter poutputdir("outputdir", "String", "", "", "", "", "",false,false); parameters.push_back(poutputdir); + CommandParameter pshared("shared", "InputTypes", "", "", "PhylipColumnShared", "PhylipColumnShared", "none","tree",false,false,true); parameters.push_back(pshared); + CommandParameter pphylip("phylip", "InputTypes", "", "", "PhylipColumnShared", "PhylipColumnShared", "none","tree",false,false); parameters.push_back(pphylip); + CommandParameter pname("name", "InputTypes", "", "", "NameCount", "none", "ColumnName","",false,false); parameters.push_back(pname); + CommandParameter pcount("count", "InputTypes", "", "", "NameCount", "none", "countcolumn","",false,false); parameters.push_back(pcount); + CommandParameter pcolumn("column", "InputTypes", "", "", "PhylipColumnShared", "PhylipColumnShared", "ColumnName-countcolumn","tree",false,false); parameters.push_back(pcolumn); + CommandParameter piters("iters", "Number", "", "1000", "", "", "","",false,false); parameters.push_back(piters); + CommandParameter psubsample("subsample", "String", "", "", "", "", "","",false,false); parameters.push_back(psubsample); + CommandParameter pcutoff("cutoff", "Number", "", "10", "", "", "","",false,false); parameters.push_back(pcutoff); + CommandParameter pprecision("precision", "Number", "", "100", "", "", "","",false,false); parameters.push_back(pprecision); + CommandParameter plabel("label", "String", "", "", "", "", "","",false,false); parameters.push_back(plabel); + CommandParameter pgroups("groups", "String", "", "", "", "", "","",false,false); parameters.push_back(pgroups); + CommandParameter pcalc("calc", "Multiple", "sharedsobs-sharedchao-sharedace-jabund-sorabund-jclass-sorclass-jest-sorest-thetayc-thetan-kstest-sharednseqs-ochiai-anderberg-kulczynski-kulczynskicody-lennon-morisitahorn-braycurtis-whittaker-odum-canberra-structeuclidean-structchord-hellinger-manhattan-structpearson-soergel-spearman-structkulczynski-speciesprofile-hamming-structchi2-gower-memchi2-memchord-memeuclidean-mempearson", "jclass-thetayc", "", "", "","",true,false,true); parameters.push_back(pcalc); + + CommandParameter pprocessors("processors", "Number", "", "1", "", "", "","",false,false,true); parameters.push_back(pprocessors); +//CommandParameter poutput("output", "Multiple", "lt-square", "lt", "", "", "",false,false); parameters.push_back(poutput); + 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); } @@ -81,12 +47,14 @@ string TreeGroupCommand::getHelpString(){ string helpString = ""; ValidCalculators validCalculator; helpString += "The tree.shared command creates a .tre to represent the similiarity between groups or sequences.\n"; - helpString += "The tree.shared command parameters are shared, groups, calc, phylip, column, name, cutoff, precision and label.\n"; + helpString += "The tree.shared command parameters are shared, groups, calc, phylip, column, name, cutoff, precision, processors, subsample, iters and label.\n"; helpString += "The groups parameter allows you to specify which of the groups in your groupfile you would like included used.\n"; helpString += "The group names are separated by dashes. The label allow you to select what distance levels you would like trees created for, and are also separated by dashes.\n"; helpString += "The phylip or column parameter are required if you do not provide a sharedfile, and only one may be used. If you use a column file the name filename is required. \n"; helpString += "If you do not provide a cutoff value 10.00 is assumed. If you do not provide a precision value then 100 is assumed.\n"; helpString += "The tree.shared command should be in the following format: tree.shared(groups=yourGroups, calc=yourCalcs, label=yourLabels).\n"; + helpString += "The iters parameter allows you to choose the number of times you would like to run the subsample.\n"; + helpString += "The subsample parameter allows you to enter the size pergroup of the sample or you can set subsample=T and mothur will use the size of your smallest group. The subsample parameter may only be used with a shared file.\n"; helpString += "Example tree.shared(groups=A-B-C, calc=jabund-sorabund).\n"; helpString += "The default value for groups is all the groups in your groupfile.\n"; helpString += "The default value for calc is jclass-thetayc.\n"; @@ -103,6 +71,21 @@ string TreeGroupCommand::getHelpString(){ } } //********************************************************************************************************************** +string TreeGroupCommand::getOutputPattern(string type) { + try { + string pattern = ""; + + if (type == "tree") { pattern = "[filename],[calc],[distance],[tag],tre-[filename],tre"; } + else { m->mothurOut("[ERROR]: No definition for type " + type + " output pattern.\n"); m->control_pressed = true; } + + return pattern; + } + catch(exception& e) { + m->errorOut(e, "TreeGroupCommand", "getOutputPattern"); + exit(1); + } +} +//********************************************************************************************************************** TreeGroupCommand::TreeGroupCommand(){ try { abort = true; calledHelp = true; @@ -173,27 +156,41 @@ TreeGroupCommand::TreeGroupCommand(string option) { //if the user has not given a path then, add inputdir. else leave path alone. if (path == "") { parameters["name"] = 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") { phylipfile = ""; abort = true; } else if (phylipfile == "not found") { phylipfile = ""; } - else { inputfile = phylipfile; format = "phylip"; } + else { inputfile = phylipfile; format = "phylip"; m->setPhylipFile(phylipfile); } columnfile = validParameter.validFile(parameters, "column", true); if (columnfile == "not open") { columnfile = ""; abort = true; } else if (columnfile == "not found") { columnfile = ""; } - else { inputfile = columnfile; format = "column"; } + else { inputfile = columnfile; format = "column"; m->setColumnFile(columnfile); } sharedfile = validParameter.validFile(parameters, "shared", true); if (sharedfile == "not open") { sharedfile = ""; abort = true; } else if (sharedfile == "not found") { sharedfile = ""; } - else { inputfile = sharedfile; format = "sharedfile"; } + else { inputfile = sharedfile; format = "sharedfile"; m->setSharedFile(sharedfile); } namefile = validParameter.validFile(parameters, "name", true); if (namefile == "not open") { abort = true; } 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); } if ((phylipfile == "") && (columnfile == "") && (sharedfile == "")) { //is there are current file available for either of these? @@ -216,15 +213,20 @@ TreeGroupCommand::TreeGroupCommand(string option) { else if ((phylipfile != "") && (columnfile != "")) { m->mothurOut("When running the tree.shared command with a distance file you may not use both the column and the phylip parameters."); m->mothurOutEndLine(); abort = true; } if (columnfile != "") { - if (namefile == "") { + if ((namefile == "") && (countfile == "")){ 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; + 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; + } } } } + //check for optional parameter and set defaults // ...at some point should added some additional type checking... @@ -239,7 +241,7 @@ TreeGroupCommand::TreeGroupCommand(string option) { if (groups == "not found") { groups = ""; } else { m->splitAtDash(groups, Groups); - m->Groups = Groups; + m->setGroups(Groups); } calc = validParameter.validFile(parameters, "calc", false); @@ -256,12 +258,30 @@ TreeGroupCommand::TreeGroupCommand(string option) { string temp; temp = validParameter.validFile(parameters, "precision", false); if (temp == "not found") { temp = "100"; } - convert(temp, precision); + m->mothurConvert(temp, precision); temp = validParameter.validFile(parameters, "cutoff", false); if (temp == "not found") { temp = "10"; } - convert(temp, cutoff); + m->mothurConvert(temp, cutoff); cutoff += (5 / (precision * 10.0)); + temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); } + m->setProcessors(temp); + m->mothurConvert(temp, processors); + + temp = validParameter.validFile(parameters, "iters", false); if (temp == "not found") { temp = "1000"; } + m->mothurConvert(temp, iters); + + temp = validParameter.validFile(parameters, "subsample", false); if (temp == "not found") { temp = "F"; } + if (m->isNumeric1(temp)) { m->mothurConvert(temp, subsampleSize); subsample = true; } + else { + if (m->isTrue(temp)) { subsample = true; subsampleSize = -1; } //we will set it to smallest group later + else { subsample = false; } + } + + if (subsample == false) { iters = 1; } + + if (subsample && (format != "sharedfile")) { m->mothurOut("[ERROR]: the subsample parameter can only be used with a shared file.\n"); abort=true; } + //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 = ""; @@ -280,8 +300,8 @@ TreeGroupCommand::TreeGroupCommand(string option) { TreeGroupCommand::~TreeGroupCommand(){ if (abort == false) { if (format == "sharedfile") { delete input; } - else { delete readMatrix; delete matrix; delete list; } - delete tmap; + else { delete list; } + delete ct; } } @@ -394,71 +414,100 @@ int TreeGroupCommand::execute(){ m->runParse = false; //create treemap class from groupmap for tree class to use - tmap = new TreeMap(); - tmap->makeSim(m->namesOfGroups); + ct = new CountTable(); + set nameMap; + map groupMap; + set gps; + for (int i = 0; i < m->getAllGroups().size(); i++) { + nameMap.insert(m->getAllGroups()[i]); + gps.insert(m->getAllGroups()[i]); + groupMap[m->getAllGroups()[i]] = m->getAllGroups()[i]; + } + ct->createTable(nameMap, groupMap, gps); //clear globaldatas old tree names if any m->Treenames.clear(); //fills globaldatas tree names - m->Treenames = m->Groups; + //m->Treenames = m->getGroups(); + for (int k = 0; k < lookup.size(); k++) { + m->Treenames.push_back(lookup[k]->getGroup()); + } if (m->control_pressed) { return 0; } //create tree file makeSimsShared(); - 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; } }else{ //read in dist file filename = inputfile; - + + ReadMatrix* readMatrix; if (format == "column") { readMatrix = new ReadColumnMatrix(filename); } else if (format == "phylip") { readMatrix = new ReadPhylipMatrix(filename); } readMatrix->setCutoff(cutoff); - if(namefile != ""){ - nameMap = new NameAssignment(namefile); - nameMap->readMap(); - } - else{ - nameMap = NULL; - } - - readMatrix->read(nameMap); - list = readMatrix->getListVector(); - matrix = readMatrix->getMatrix(); + ct = NULL; + nameMap = NULL; + if(namefile != ""){ + nameMap = new NameAssignment(namefile); + nameMap->readMap(); + readMatrix->read(nameMap); + }else if (countfile != "") { + ct = new CountTable(); + ct->readTable(countfile, true, false); + readMatrix->read(ct); + }else { + readMatrix->read(nameMap); + } + list = readMatrix->getListVector(); + SparseDistanceMatrix* dMatrix = readMatrix->getDMatrix(); + + //clear globaldatas old tree names if any + m->Treenames.clear(); + //make treemap - tmap = new TreeMap(); - - if (m->control_pressed) { return 0; } + if (ct != NULL) { delete ct; } + ct = new CountTable(); + set nameMap; + map groupMap; + set gps; + for (int i = 0; i < list->getNumBins(); i++) { + string bin = list->get(i); + nameMap.insert(bin); + gps.insert(bin); + groupMap[bin] = bin; + m->Treenames.push_back(bin); + } + ct->createTable(nameMap, groupMap, gps); - tmap->makeSim(list); - - m->Groups = tmap->namesOfGroups; - - //clear globaldatas old tree names if any - m->Treenames.clear(); - - //fills globaldatas tree names - m->Treenames = m->Groups; + vector namesGroups = ct->getNamesOfGroups(); + m->setGroups(namesGroups); //used in tree constructor m->runParse = false; if (m->control_pressed) { return 0; } - makeSimsDist(); + vector< vector > matrix = makeSimsDist(dMatrix); + delete readMatrix; + delete dMatrix; if (m->control_pressed) { return 0; } //create a new filename - outputFile = outputDir + m->getRootName(m->getSimpleName(inputfile)) + "tre"; + map variables; + variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(inputfile)); + string outputFile = getOutputFileName("tree",variables); outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile); - createTree(); + Tree* newTree = createTree(matrix); + + if (newTree != NULL) { writeTree(outputFile, newTree); delete newTree; } if (m->control_pressed) { return 0; } @@ -467,7 +516,7 @@ int TreeGroupCommand::execute(){ } //reset groups parameter - m->Groups.clear(); + m->clearGroups(); //set tree file as new current treefile string current = ""; @@ -490,83 +539,17 @@ int TreeGroupCommand::execute(){ } //********************************************************************************************************************** -int TreeGroupCommand::createTree(){ +Tree* TreeGroupCommand::createTree(vector< vector >& simMatrix){ try { //create tree - t = new Tree(tmap); - - //do merges and create tree structure by setting parents and children - //there are numGroups - 1 merges to do - for (int i = 0; i < (numGroups - 1); i++) { - float largest = -1000.0; - - if (m->control_pressed) { delete t; return 1; } - - int row, column; - //find largest value in sims matrix by searching lower triangle - for (int j = 1; j < simMatrix.size(); j++) { - for (int k = 0; k < j; k++) { - if (simMatrix[j][k] > largest) { largest = simMatrix[j][k]; row = j; column = k; } - } - } - - //set non-leaf node info and update leaves to know their parents - //non-leaf - t->tree[numGroups + i].setChildren(index[row], index[column]); - - //parents - t->tree[index[row]].setParent(numGroups + i); - t->tree[index[column]].setParent(numGroups + i); - - //blength = distance / 2; - float blength = ((1.0 - largest) / 2); - - //branchlengths - t->tree[index[row]].setBranchLength(blength - t->tree[index[row]].getLengthToLeaves()); - t->tree[index[column]].setBranchLength(blength - t->tree[index[column]].getLengthToLeaves()); - - //set your length to leaves to your childs length plus branchlength - t->tree[numGroups + i].setLengthToLeaves(t->tree[index[row]].getLengthToLeaves() + t->tree[index[row]].getBranchLength()); - - - //update index - index[row] = numGroups+i; - index[column] = numGroups+i; - - //remove highest value that caused the merge. - simMatrix[row][column] = -1000.0; - simMatrix[column][row] = -1000.0; - - //merge values in simsMatrix - for (int n = 0; n < simMatrix.size(); n++) { - //row becomes merge of 2 groups - simMatrix[row][n] = (simMatrix[row][n] + simMatrix[column][n]) / 2; - simMatrix[n][row] = simMatrix[row][n]; - //delete column - simMatrix[column][n] = -1000.0; - simMatrix[n][column] = -1000.0; - } - } - - //adjust tree to make sure root to tip length is .5 - int root = t->findRoot(); - t->tree[root].setBranchLength((0.5 - t->tree[root].getLengthToLeaves())); + t = new Tree(ct, simMatrix); + + if (m->control_pressed) { delete t; t = NULL; return t; } - //assemble tree + //assemble tree t->assembleTree(); - - if (m->control_pressed) { delete t; return 1; } - - //print newick file - t->createNewickFile(outputFile); - - //delete tree - delete t; - - if (m->control_pressed) { remove(outputFile.c_str()); outputNames.pop_back(); return 1; } - - return 0; - + + return t; } catch(exception& e) { m->errorOut(e, "TreeGroupCommand", "createTree"); @@ -574,16 +557,28 @@ int TreeGroupCommand::createTree(){ } } /***********************************************************/ -void TreeGroupCommand::printSims(ostream& out) { +int TreeGroupCommand::writeTree(string out, Tree* T) { try { - //output column headers - //out << '\t'; - //for (int i = 0; i < lookup.size(); i++) { out << lookup[i]->getGroup() << '\t'; } - //out << endl; + //print newick file + t->createNewickFile(out); + if (m->control_pressed) { m->mothurRemove(out); outputNames.pop_back(); return 1; } + + return 0; + + } + catch(exception& e) { + m->errorOut(e, "TreeGroupCommand", "printSims"); + exit(1); + } +} + +/***********************************************************/ +void TreeGroupCommand::printSims(ostream& out, vector< vector >& simMatrix) { + try { - for (int m = 0; m < simMatrix.size(); m++) { + for (int m = 0; m < simMatrix.size(); m++) { //out << lookup[m]->getGroup() << '\t'; for (int n = 0; n < simMatrix.size(); n++) { out << simMatrix[m][n] << '\t'; @@ -598,16 +593,12 @@ void TreeGroupCommand::printSims(ostream& out) { } } /***********************************************************/ -int TreeGroupCommand::makeSimsDist() { +vector< vector > TreeGroupCommand::makeSimsDist(SparseDistanceMatrix* matrix) { try { numGroups = list->size(); - //initialize index - index.clear(); - for (int g = 0; g < numGroups; g++) { index[g] = g; } - //initialize simMatrix - simMatrix.clear(); + vector< vector > simMatrix; simMatrix.resize(numGroups); for (int k = 0; k < simMatrix.size(); k++) { for (int j = 0; j < simMatrix.size(); j++) { @@ -617,16 +608,20 @@ int TreeGroupCommand::makeSimsDist() { //go through sparse matrix and fill sims //go through each cell in the sparsematrix - for(MatData currentCell = matrix->begin(); currentCell != matrix->end(); currentCell++){ - //similairity = -(distance-1) - simMatrix[currentCell->row][currentCell->column] = -(currentCell->dist -1.0); - simMatrix[currentCell->column][currentCell->row] = -(currentCell->dist -1.0); - - if (m->control_pressed) { return 1; } + for (int i = 0; i < matrix->seqVec.size(); i++) { + for (int j = 0; j < matrix->seqVec[i].size(); j++) { + + //already checked everyone else in row + if (i < matrix->seqVec[i][j].index) { + simMatrix[i][matrix->seqVec[i][j].index] = -(matrix->seqVec[i][j].dist -1.0); + simMatrix[matrix->seqVec[i][j].index][i] = -(matrix->seqVec[i][j].dist -1.0); + if (m->control_pressed) { return simMatrix; } + } + } } - return 0; + return simMatrix; } catch(exception& e) { m->errorOut(e, "TreeGroupCommand", "makeSimsDist"); @@ -637,6 +632,54 @@ int TreeGroupCommand::makeSimsDist() { /***********************************************************/ int TreeGroupCommand::makeSimsShared() { try { + + if (subsample) { + if (subsampleSize == -1) { //user has not set size, set size = smallest samples size + subsampleSize = lookup[0]->getNumSeqs(); + for (int i = 1; i < lookup.size(); i++) { + int thisSize = lookup[i]->getNumSeqs(); + + if (thisSize < subsampleSize) { subsampleSize = thisSize; } + } + }else { + m->clearGroups(); + Groups.clear(); + m->Treenames.clear(); + vector temp; + for (int i = 0; i < lookup.size(); i++) { + if (lookup[i]->getNumSeqs() < subsampleSize) { + m->mothurOut(lookup[i]->getGroup() + " contains " + toString(lookup[i]->getNumSeqs()) + ". Eliminating."); m->mothurOutEndLine(); + delete lookup[i]; + }else { + Groups.push_back(lookup[i]->getGroup()); + temp.push_back(lookup[i]); + m->Treenames.push_back(lookup[i]->getGroup()); + } + } + lookup = temp; + m->setGroups(Groups); + } + + if (lookup.size() < 2) { m->mothurOut("You have not provided enough valid groups. I cannot run the command."); m->mothurOutEndLine(); m->control_pressed = true; return 0; } + } + numGroups = lookup.size(); + + //sanity check to make sure processors < numComparisions + int numDists = 0; + for(int i=0;i processors) { break; } + } + } + if (numDists < processors) { processors = numDists; } + + lines.resize(processors); + for (int i = 0; i < processors; i++) { + lines[i].start = int (sqrt(float(i)/float(processors)) * numGroups); + lines[i].end = int (sqrt(float(i+1)/float(processors)) * numGroups); + } + set processedLabels; set userLabels = labels; @@ -713,84 +756,365 @@ int TreeGroupCommand::makeSimsShared() { /***********************************************************/ int TreeGroupCommand::process(vector thisLookup) { try{ - EstOutput data; - vector subset; - numGroups = thisLookup.size(); - - //for each calculator - for(int i = 0 ; i < treeCalculators.size(); i++) { - //initialize simMatrix - simMatrix.clear(); - simMatrix.resize(numGroups); - for (int k = 0; k < simMatrix.size(); k++) { - for (int j = 0; j < simMatrix.size(); j++) { - simMatrix[k].push_back(0.0); - } - } - - //initialize index - index.clear(); - for (int g = 0; g < numGroups; g++) { index[g] = g; } + vector< vector< vector > > calcDistsTotals; //each iter, one for each calc, then each groupCombos dists. this will be used to make .dist files + vector< vector > calcDists; calcDists.resize(treeCalculators.size()); + + for (int thisIter = 0; thisIter < iters; thisIter++) { + + vector thisItersLookup = thisLookup; + + if (subsample) { + SubSample sample; + vector tempLabels; //dont need since we arent printing the sampled sharedRabunds + + //make copy of lookup so we don't get access violations + vector newLookup; + for (int k = 0; k < thisItersLookup.size(); k++) { + SharedRAbundVector* temp = new SharedRAbundVector(); + temp->setLabel(thisItersLookup[k]->getLabel()); + temp->setGroup(thisItersLookup[k]->getGroup()); + newLookup.push_back(temp); + } + + //for each bin + for (int k = 0; k < thisItersLookup[0]->getNumBins(); k++) { + if (m->control_pressed) { for (int j = 0; j < newLookup.size(); j++) { delete newLookup[j]; } return 0; } + for (int j = 0; j < thisItersLookup.size(); j++) { newLookup[j]->push_back(thisItersLookup[j]->getAbundance(k), thisItersLookup[j]->getGroup()); } + } + + tempLabels = sample.getSample(newLookup, subsampleSize); + thisItersLookup = newLookup; + } + + if(processors == 1){ + driver(thisItersLookup, 0, numGroups, calcDists); + }else{ + int process = 1; + vector processIDS; + +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) + //loop through and create all the processes you want + while (process != processors) { + int pid = fork(); + + if (pid > 0) { + processIDS.push_back(pid); + process++; + }else if (pid == 0){ + + driver(thisItersLookup, lines[process].start, lines[process].end, calcDists); + + string tempdistFileName = m->getRootName(m->getSimpleName(sharedfile)) + toString(getpid()) + ".dist"; + ofstream outtemp; + m->openOutputFile(tempdistFileName, outtemp); + + for (int i = 0; i < calcDists.size(); i++) { + outtemp << calcDists[i].size() << endl; + + for (int j = 0; j < calcDists[i].size(); j++) { + outtemp << calcDists[i][j].seq1 << '\t' << calcDists[i][j].seq2 << '\t' << calcDists[i][j].dist << endl; + } + } + outtemp.close(); + + exit(0); + }else { + m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine(); + for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); } + exit(0); + } + } + + //parent do your part + driver(thisItersLookup, lines[0].start, lines[0].end, calcDists); + + //force parent to wait until all the processes are done + for (int i = 0; i < processIDS.size(); i++) { + int temp = processIDS[i]; + wait(&temp); + } + + for (int i = 0; i < processIDS.size(); i++) { + string tempdistFileName = m->getRootName(m->getSimpleName(sharedfile)) + toString(processIDS[i]) + ".dist"; + ifstream intemp; + m->openInputFile(tempdistFileName, intemp); + + for (int k = 0; k < calcDists.size(); k++) { + int size = 0; + intemp >> size; m->gobble(intemp); + + for (int j = 0; j < size; j++) { + int seq1 = 0; + int seq2 = 0; + float dist = 1.0; + + intemp >> seq1 >> seq2 >> dist; m->gobble(intemp); + + seqDist tempDist(seq1, seq2, dist); + calcDists[k].push_back(tempDist); + } + } + intemp.close(); + m->mothurRemove(tempdistFileName); + } +#else + ////////////////////////////////////////////////////////////////////////////////////////////////////// + //Windows version shared memory, so be careful when passing variables through the treeSharedData struct. + //Above fork() will clone, so memory is separate, but that's not the case with windows, + //Taking advantage of shared memory to pass results vectors. + ////////////////////////////////////////////////////////////////////////////////////////////////////// + + vector pDataArray; + DWORD dwThreadIdArray[processors-1]; + HANDLE hThreadArray[processors-1]; + + //Create processor worker threads. + for( int i=1; i newLookup; + for (int k = 0; k < thisItersLookup.size(); k++) { + SharedRAbundVector* temp = new SharedRAbundVector(); + temp->setLabel(thisItersLookup[k]->getLabel()); + temp->setGroup(thisItersLookup[k]->getGroup()); + newLookup.push_back(temp); + } + + //for each bin + for (int k = 0; k < thisItersLookup[0]->getNumBins(); k++) { + if (m->control_pressed) { for (int j = 0; j < newLookup.size(); j++) { delete newLookup[j]; } return 0; } + for (int j = 0; j < thisItersLookup.size(); j++) { newLookup[j]->push_back(thisItersLookup[j]->getAbundance(k), thisItersLookup[j]->getGroup()); } + } + + // Allocate memory for thread data. + treeSharedData* tempSum = new treeSharedData(m, lines[i].start, lines[i].end, Estimators, newLookup); + pDataArray.push_back(tempSum); + processIDS.push_back(i); + + hThreadArray[i-1] = CreateThread(NULL, 0, MyTreeSharedThreadFunction, pDataArray[i-1], 0, &dwThreadIdArray[i-1]); + } + + //parent do your part + driver(thisItersLookup, lines[0].start, lines[0].end, calcDists); + + //Wait until all threads have terminated. + WaitForMultipleObjects(processors-1, hThreadArray, TRUE, INFINITE); + + //Close all thread handles and free memory allocations. + for(int i=0; i < pDataArray.size(); i++){ + if (pDataArray[i]->count != (pDataArray[i]->end-pDataArray[i]->start)) { + m->mothurOut("[ERROR]: process " + toString(i) + " only processed " + toString(pDataArray[i]->count) + " of " + toString(pDataArray[i]->end-pDataArray[i]->start) + " groups assigned to it, quitting. \n"); m->control_pressed = true; + } + for (int j = 0; j < pDataArray[i]->thisLookup.size(); j++) { delete pDataArray[i]->thisLookup[j]; } + + for (int k = 0; k < calcDists.size(); k++) { + int size = pDataArray[i]->calcDists[k].size(); + for (int j = 0; j < size; j++) { calcDists[k].push_back(pDataArray[i]->calcDists[k][j]); } + } + + CloseHandle(hThreadArray[i]); + delete pDataArray[i]; + } + +#endif + } + + calcDistsTotals.push_back(calcDists); + + if (subsample) { + + //clean up memory + for (int i = 0; i < thisItersLookup.size(); i++) { delete thisItersLookup[i]; } + thisItersLookup.clear(); + for (int i = 0; i < calcDists.size(); i++) { calcDists[i].clear(); } + } + + if (m->debug) { m->mothurOut("[DEBUG]: iter = " + toString(thisIter) + ".\n"); } + } + + if (m->debug) { m->mothurOut("[DEBUG]: done with iters.\n"); } + + if (iters != 1) { + //we need to find the average distance and standard deviation for each groups distance + vector< vector > calcAverages = m->getAverages(calcDistsTotals); + + if (m->debug) { m->mothurOut("[DEBUG]: found averages.\n"); } + + //create average tree for each calc + for (int i = 0; i < calcDists.size(); i++) { + vector< vector > matrix; //square matrix to represent the distance + matrix.resize(thisLookup.size()); + for (int k = 0; k < thisLookup.size(); k++) { matrix[k].resize(thisLookup.size(), 0.0); } + + for (int j = 0; j < calcAverages[i].size(); j++) { + int row = calcAverages[i][j].seq1; + int column = calcAverages[i][j].seq2; + float dist = calcAverages[i][j].dist; + + matrix[row][column] = dist; + matrix[column][row] = dist; + } + + //create a new filename + map variables; + variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(inputfile)); + variables["[calc]"] = treeCalculators[i]->getName(); + variables["[distance]"] = thisLookup[0]->getLabel(); + variables["[tag]"] = "ave"; + string outputFile = getOutputFileName("tree",variables); + outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile); + + //creates tree from similarity matrix and write out file + Tree* newTree = createTree(matrix); + if (newTree != NULL) { writeTree(outputFile, newTree); } + } + + if (m->debug) { m->mothurOut("[DEBUG]: done averages trees.\n"); } + + //create all trees for each calc and find their consensus tree + for (int i = 0; i < calcDists.size(); i++) { + if (m->control_pressed) { break; } + + //create a new filename + //create a new filename + map variables; + variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(inputfile)); + variables["[calc]"] = treeCalculators[i]->getName(); + variables["[distance]"] = thisLookup[0]->getLabel(); + variables["[tag]"] = "all"; + string outputFile = getOutputFileName("tree",variables); + outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile); + + ofstream outAll; + m->openOutputFile(outputFile, outAll); + + vector trees; + for (int myIter = 0; myIter < iters; myIter++) { + + if(m->control_pressed) { break; } + + //initialize matrix + vector< vector > matrix; //square matrix to represent the distance + matrix.resize(thisLookup.size()); + for (int k = 0; k < thisLookup.size(); k++) { matrix[k].resize(thisLookup.size(), 0.0); } + + for (int j = 0; j < calcDistsTotals[myIter][i].size(); j++) { + int row = calcDistsTotals[myIter][i][j].seq1; + int column = calcDistsTotals[myIter][i][j].seq2; + double dist = calcDistsTotals[myIter][i][j].dist; + + matrix[row][column] = dist; + matrix[column][row] = dist; + } + + //creates tree from similarity matrix and write out file + Tree* newTree = createTree(matrix); + if (newTree != NULL) { + newTree->print(outAll); + trees.push_back(newTree); + } + } + outAll.close(); + if (m->control_pressed) { for (int k = 0; k < trees.size(); k++) { delete trees[k]; } } + + if (m->debug) { m->mothurOut("[DEBUG]: done all trees.\n"); } + + Consensus consensus; + //clear old tree names if any + m->Treenames.clear(); m->Treenames = m->getGroups(); //may have changed if subsample eliminated groups + Tree* conTree = consensus.getTree(trees); + + if (m->debug) { m->mothurOut("[DEBUG]: done cons tree.\n"); } + + //create a new filename + variables["[tag]"] = "cons"; + string conFile = getOutputFileName("tree",variables); + outputNames.push_back(conFile); outputTypes["tree"].push_back(conFile); + ofstream outTree; + m->openOutputFile(conFile, outTree); + + if (conTree != NULL) { conTree->print(outTree, "boot"); delete conTree; } + } + + }else { + + for (int i = 0; i < calcDists.size(); i++) { + if (m->control_pressed) { break; } + + //initialize matrix + vector< vector > matrix; //square matrix to represent the distance + matrix.resize(thisLookup.size()); + for (int k = 0; k < thisLookup.size(); k++) { matrix[k].resize(thisLookup.size(), 0.0); } + + for (int j = 0; j < calcDists[i].size(); j++) { + int row = calcDists[i][j].seq1; + int column = calcDists[i][j].seq2; + double dist = calcDists[i][j].dist; + + matrix[row][column] = dist; + matrix[column][row] = dist; + } + + //create a new filename + map variables; + variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(inputfile)); + variables["[calc]"] = treeCalculators[i]->getName(); + variables["[distance]"] = thisLookup[0]->getLabel(); + variables["[tag]"] = ""; + string outputFile = getOutputFileName("tree",variables); + outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile); + + //creates tree from similarity matrix and write out file + Tree* newTree = createTree(matrix); + if (newTree != NULL) { writeTree(outputFile, newTree); delete newTree; } + } + } - //create a new filename - outputFile = outputDir + m->getRootName(m->getSimpleName(inputfile)) + treeCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + ".tre"; - outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile); - - for (int k = 0; k < thisLookup.size(); k++) { - for (int l = k; l < thisLookup.size(); l++) { - if (k != l) { //we dont need to similiarity of a groups to itself - //get estimated similarity between 2 groups - - subset.clear(); //clear out old pair of sharedrabunds - //add new pair of sharedrabunds - subset.push_back(thisLookup[k]); subset.push_back(thisLookup[l]); - - //if this calc needs all groups to calculate the pair load all groups - if (treeCalculators[i]->getNeedsAll()) { - //load subset with rest of lookup for those calcs that need everyone to calc for a pair - for (int w = 0; w < thisLookup.size(); w++) { - if ((w != k) && (w != l)) { subset.push_back(thisLookup[w]); } - } - } - - data = treeCalculators[i]->getValues(subset); //saves the calculator outputs - //cout << thisLookup[k]->getGroup() << '\t' << thisLookup[l]->getGroup() << '\t' << (1.0 - data[0]) << endl; - if (m->control_pressed) { return 1; } - - //save values in similarity matrix - simMatrix[k][l] = -(data[0]-1.0); - simMatrix[l][k] = -(data[0]-1.0); - } - } - } + return 0; + } + catch(exception& e) { + m->errorOut(e, "TreeGroupCommand", "process"); + exit(1); + } +} +/**************************************************************************************************/ +int TreeGroupCommand::driver(vector thisLookup, int start, int end, vector< vector >& calcDists) { + try { + vector subset; + for (int k = start; k < end; k++) { // pass cdd each set of groups to compare + + for (int l = 0; l < k; l++) { + + if (k != l) { //we dont need to similiarity of a groups to itself + subset.clear(); //clear out old pair of sharedrabunds + //add new pair of sharedrabunds + subset.push_back(thisLookup[k]); subset.push_back(thisLookup[l]); - //createdistance file from simMatrix - /*string o = outputDir + m->getRootName(m->getSimpleName(globaldata->inputFileName)) + treeCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + ".dist"; - ofstream outDist; - m->openOutputFile(o, outDist); - outDist << simMatrix.size() << endl; - for (int k = 0; k < simMatrix.size(); k++) { - outDist << thisLookup[k]->getGroup() << '\t'; - for (int l = 0; l < k; l++) { - outDist << (1.0-simMatrix[k][l]) << '\t'; + for(int i=0;igetNeedsAll()) { + //load subset with rest of lookup for those calcs that need everyone to calc for a pair + for (int w = 0; w < thisLookup.size(); w++) { + if ((w != k) && (w != l)) { subset.push_back(thisLookup[w]); } + } } - outDist << endl; + + vector tempdata = treeCalculators[i]->getValues(subset); //saves the calculator outputs + + if (m->control_pressed) { return 1; } + + seqDist temp(l, k, -(tempdata[0]-1.0)); + calcDists[i].push_back(temp); } - outDist.close();*/ - - - if (m->control_pressed) { return 1; } - //creates tree from similarity matrix and write out file - createTree(); - - if (m->control_pressed) { return 1; } } - - return 0; - + } + } + + return 0; } catch(exception& e) { - m->errorOut(e, "TreeGroupCommand", "process"); + m->errorOut(e, "TreeGroupCommand", "driver"); exit(1); } }