X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=matrixoutputcommand.cpp;h=310e32bf23e048c73ae37579ac2e8e2d1cb8b0ff;hp=9c7a83681bb8208d1317a629947a3078084a351e;hb=1a20e24ee786195ab0e1cccd4f5aede7a88f3f4e;hpb=ec1b5bc7460ac8ad40f54f0729900d9ed98e7292 diff --git a/matrixoutputcommand.cpp b/matrixoutputcommand.cpp index 9c7a836..310e32b 100644 --- a/matrixoutputcommand.cpp +++ b/matrixoutputcommand.cpp @@ -8,59 +8,22 @@ */ #include "matrixoutputcommand.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" //********************************************************************************************************************** vector MatrixOutputCommand::setParameters(){ try { - CommandParameter pshared("shared", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(pshared); - 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 pprocessors("processors", "Number", "", "1", "", "", "",false,false); parameters.push_back(pprocessors); - CommandParameter pinputdir("inputdir", "String", "", "", "", "", "",false,false); parameters.push_back(pinputdir); - CommandParameter poutputdir("outputdir", "String", "", "", "", "", "",false,false); parameters.push_back(poutputdir); + CommandParameter pshared("shared", "InputTypes", "", "", "none", "none", "none","phylip",false,true,true); parameters.push_back(pshared); + CommandParameter plabel("label", "String", "", "", "", "", "","",false,false); parameters.push_back(plabel); + CommandParameter psubsample("subsample", "String", "", "", "", "", "","",false,false); parameters.push_back(psubsample); + 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 poutput("output", "Multiple", "lt-square-column", "lt", "", "", "","",false,false); parameters.push_back(poutput); + CommandParameter pmode("mode", "Multiple", "average-median", "average", "", "", "","",false,false); parameters.push_back(pmode); + CommandParameter pprocessors("processors", "Number", "", "1", "", "", "","",false,false,true); parameters.push_back(pprocessors); + CommandParameter piters("iters", "Number", "", "1000", "", "", "","",false,false); parameters.push_back(piters); + 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); } @@ -76,11 +39,14 @@ string MatrixOutputCommand::getHelpString(){ try { string helpString = ""; ValidCalculators validCalculator; - helpString += "The dist.shared command parameters are shared, groups, calc, output, processors and label. shared is a required, unless you have a valid current file.\n"; + helpString += "The dist.shared command parameters are shared, groups, calc, output, processors, subsample, iters, mode, and label. shared is a required, unless you have a valid current file.\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 parameter allows you to select what distance levels you would like distance matrices created for, and is also separated by dashes.\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.\n"; helpString += "The dist.shared command should be in the following format: dist.shared(groups=yourGroups, calc=yourCalcs, label=yourLabels).\n"; - helpString += "The output parameter allows you to specify format of your distance matrix. Options are lt, and square. The default is lt.\n"; + helpString += "The output parameter allows you to specify format of your distance matrix. Options are lt, column and square. The default is lt.\n"; + helpString += "The mode parameter allows you to specify if you want the average or the median values reported when subsampling. Options are average, and median. The default is average.\n"; helpString += "Example dist.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 and thetayc.\n"; @@ -95,6 +61,21 @@ string MatrixOutputCommand::getHelpString(){ } } //********************************************************************************************************************** +string MatrixOutputCommand::getOutputPattern(string type) { + try { + string pattern = ""; + + if (type == "phylip") { pattern = "[filename],[calc],[distance],[outputtag],dist-[filename],[calc],[distance],[outputtag],[tag2],dist"; } + else { m->mothurOut("[ERROR]: No definition for type " + type + " output pattern.\n"); m->control_pressed = true; } + + return pattern; + } + catch(exception& e) { + m->errorOut(e, "MatrixOutputCommand", "getOutputPattern"); + exit(1); + } +} +//********************************************************************************************************************** MatrixOutputCommand::MatrixOutputCommand(){ try { abort = true; calledHelp = true; @@ -157,6 +138,7 @@ MatrixOutputCommand::MatrixOutputCommand(string option) { if (sharedfile != "") { m->mothurOut("Using " + sharedfile + " as input file for the shared parameter."); m->mothurOutEndLine(); } else { m->mothurOut("You have no current sharedfile and the shared parameter is required."); m->mothurOutEndLine(); abort = true; } }else if (sharedfile == "not open") { sharedfile = ""; abort = true; } + else { m->setSharedFile(sharedfile); } //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"){ @@ -174,18 +156,21 @@ MatrixOutputCommand::MatrixOutputCommand(string option) { } output = validParameter.validFile(parameters, "output", false); if(output == "not found"){ output = "lt"; } - if ((output != "lt") && (output != "square")) { m->mothurOut(output + " is not a valid output form. Options are lt and square. I will use lt."); m->mothurOutEndLine(); output = "lt"; } + if ((output != "lt") && (output != "square") && (output != "column")) { m->mothurOut(output + " is not a valid output form. Options are lt, column and square. I will use lt."); m->mothurOutEndLine(); output = "lt"; } + + mode = validParameter.validFile(parameters, "mode", false); if(mode == "not found"){ mode = "average"; } + if ((mode != "average") && (mode != "median")) { m->mothurOut(mode + " is not a valid mode. Options are average and medina. I will use average."); m->mothurOutEndLine(); output = "average"; } groups = validParameter.validFile(parameters, "groups", false); if (groups == "not found") { groups = ""; } else { m->splitAtDash(groups, Groups); - m->Groups = Groups; + m->setGroups(Groups); } string temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); } m->setProcessors(temp); - convert(temp, processors); + m->mothurConvert(temp, processors); calc = validParameter.validFile(parameters, "calc", false); if (calc == "not found") { calc = "jclass-thetayc"; } @@ -198,7 +183,19 @@ MatrixOutputCommand::MatrixOutputCommand(string option) { //remove citation from list of calcs for (int i = 0; i < Estimators.size(); i++) { if (Estimators[i] == "citation") { Estimators.erase(Estimators.begin()+i); break; } } } - + + 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 = 0; } + if (abort == false) { ValidCalculators validCalculator; @@ -321,21 +318,48 @@ int MatrixOutputCommand::execute(){ set userLabels = labels; if (lookup.size() < 2) { m->mothurOut("You have not provided enough valid groups. I cannot run the command."); m->mothurOutEndLine(); delete input; for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } return 0;} - + + 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(); + 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]); + } + } + 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; delete input; return 0; } + } + numGroups = lookup.size(); - lines.resize(processors); + 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); - cout << i << '\t' << lines[i].start << '\t' << lines[i].end << endl; } - - if (m->control_pressed) { delete input; for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } m->Groups.clear(); return 0; } + + if (m->control_pressed) { delete input; for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } m->clearGroups(); return 0; } //as long as you are not at the end of the file or done wih the lines you want while((lookup[0] != NULL) && ((allLines == 1) || (userLabels.size() != 0))) { - if (m->control_pressed) { outputTypes.clear(); delete input; for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } m->Groups.clear(); return 0; } + if (m->control_pressed) { outputTypes.clear(); delete input; for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } m->clearGroups(); return 0; } if(allLines == 1 || labels.count(lookup[0]->getLabel()) == 1){ m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine(); @@ -368,7 +392,7 @@ int MatrixOutputCommand::execute(){ lookup = input->getSharedRAbundVectors(); } - if (m->control_pressed) { outputTypes.clear(); delete input; for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } m->Groups.clear(); return 0; } + if (m->control_pressed) { outputTypes.clear(); delete input; for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } m->clearGroups(); return 0; } //output error messages about any remaining user labels set::iterator it; @@ -383,7 +407,7 @@ int MatrixOutputCommand::execute(){ } } - if (m->control_pressed) { outputTypes.clear(); delete input; for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } m->Groups.clear(); return 0; } + if (m->control_pressed) { outputTypes.clear(); delete input; for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } m->clearGroups(); return 0; } //run last label if you need to if (needToRun == true) { @@ -395,10 +419,10 @@ int MatrixOutputCommand::execute(){ for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } } - if (m->control_pressed) { outputTypes.clear(); delete input; for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } m->Groups.clear(); return 0; } + if (m->control_pressed) { outputTypes.clear(); delete input; for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } m->clearGroups(); return 0; } //reset groups parameter - m->Groups.clear(); + m->clearGroups(); //set phylip file as new current phylipfile string current = ""; @@ -421,27 +445,32 @@ int MatrixOutputCommand::execute(){ } } /***********************************************************/ -void MatrixOutputCommand::printSims(ostream& out, vector< vector >& simMatrix) { +void MatrixOutputCommand::printSims(ostream& out, vector< vector >& simMatrix) { try { out.setf(ios::fixed, ios::floatfield); out.setf(ios::showpoint); - - //output num seqs - out << simMatrix.size() << endl; - + if (output == "lt") { - for (int m = 0; m < simMatrix.size(); m++) { - out << lookup[m]->getGroup() << '\t'; - for (int n = 0; n < m; n++) { - out << simMatrix[m][n] << '\t'; + out << simMatrix.size() << endl; + for (int b = 0; b < simMatrix.size(); b++) { + out << lookup[b]->getGroup() << '\t'; + for (int n = 0; n < b; n++) { + out << simMatrix[b][n] << '\t'; } out << endl; } + }else if (output == "column") { + for (int b = 0; b < simMatrix.size(); b++) { + for (int n = 0; n < b; n++) { + out << lookup[b]->getGroup() << '\t' << lookup[n]->getGroup() << '\t' << simMatrix[b][n] << endl; + } + } }else{ - for (int m = 0; m < simMatrix.size(); m++) { - out << lookup[m]->getGroup() << '\t'; - for (int n = 0; n < simMatrix[m].size(); n++) { - out << simMatrix[m][n] << '\t'; + out << simMatrix.size() << endl; + for (int b = 0; b < simMatrix.size(); b++) { + out << lookup[b]->getGroup() << '\t'; + for (int n = 0; n < simMatrix[b].size(); n++) { + out << simMatrix[b][n] << '\t'; } out << endl; } @@ -455,113 +484,278 @@ void MatrixOutputCommand::printSims(ostream& out, vector< vector >& simMa /***********************************************************/ int MatrixOutputCommand::process(vector thisLookup){ try { - EstOutput data; - vector subset; - vector< vector > calcDists; calcDists.resize(matrixCalculators.size()); //one for each calc, this will be used to make .dist files - - #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) - if(processors == 1){ - driver(thisLookup, 0, numGroups, calcDists); - }else{ - int process = 1; - vector processIDS; - - //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(thisLookup, 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(thisLookup, 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(); - remove(tempdistFileName.c_str()); - } - + 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(matrixCalculators.size()); + + for (int thisIter = 0; thisIter < iters+1; thisIter++) { + map variables; + variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(sharedfile)); + variables["[distance]"] = thisLookup[0]->getLabel(); + variables["[tag2]"] = ""; + + vector thisItersLookup = thisLookup; + + if (subsample && (thisIter != 0)) { + 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 distSharedData 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. + distSharedData* tempSum = new distSharedData(m, lines[i].start, lines[i].end, Estimators, newLookup); + pDataArray.push_back(tempSum); + processIDS.push_back(i); + + hThreadArray[i-1] = CreateThread(NULL, 0, MyDistSharedThreadFunction, 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 + } + + if (subsample && (thisIter != 0)) { + if((thisIter) % 100 == 0){ m->mothurOut(toString(thisIter)); m->mothurOutEndLine(); } + calcDistsTotals.push_back(calcDists); + for (int i = 0; i < calcDists.size(); i++) { + for (int j = 0; j < calcDists[i].size(); j++) { + if (m->debug) { m->mothurOut("[DEBUG]: Results: iter = " + toString(thisIter) + ", " + thisLookup[calcDists[i][j].seq1]->getGroup() + " - " + thisLookup[calcDists[i][j].seq2]->getGroup() + " distance = " + toString(calcDists[i][j].dist) + ".\n"); } + } + } + //clean up memory + for (int i = 0; i < thisItersLookup.size(); i++) { delete thisItersLookup[i]; } + thisItersLookup.clear(); + }else { //print results for whole dataset + 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; + } + + variables["[outputtag]"] = output; + variables["[calc]"] = matrixCalculators[i]->getName(); + string distFileName = getOutputFileName("phylip",variables); + outputNames.push_back(distFileName); outputTypes["phylip"].push_back(distFileName); + + ofstream outDist; + m->openOutputFile(distFileName, outDist); + outDist.setf(ios::fixed, ios::floatfield); outDist.setf(ios::showpoint); + + printSims(outDist, matrix); + + outDist.close(); + } + } + for (int i = 0; i < calcDists.size(); i++) { calcDists[i].clear(); } } -#else - driver(thisLookup, 0, numGroups, calcDists); -#endif - 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; - float dist = calcDists[i][j].dist; - - matrix[row][column] = dist; - matrix[column][row] = dist; - } - - string distFileName = outputDir + m->getRootName(m->getSimpleName(sharedfile)) + matrixCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + "." + output + ".dist"; - outputNames.push_back(distFileName); outputTypes["phylip"].push_back(distFileName); - ofstream outDist; - m->openOutputFile(distFileName, outDist); - outDist.setf(ios::fixed, ios::floatfield); outDist.setf(ios::showpoint); - - printSims(outDist, matrix); - - outDist.close(); - } + if (iters != 0) { + //we need to find the average distance and standard deviation for each groups distance + vector< vector > calcAverages = m->getAverages(calcDistsTotals, mode); + + //find standard deviation + vector< vector > stdDev = m->getStandardDeviation(calcDistsTotals, calcAverages); + + //print results + 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); } + + vector< vector > stdmatrix; //square matrix to represent the stdDev + stdmatrix.resize(thisLookup.size()); + for (int k = 0; k < thisLookup.size(); k++) { stdmatrix[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; + float stdDist = stdDev[i][j].dist; + + matrix[row][column] = dist; + matrix[column][row] = dist; + stdmatrix[row][column] = stdDist; + stdmatrix[column][row] = stdDist; + } + + map variables; + variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(sharedfile)); + variables["[distance]"] = thisLookup[0]->getLabel(); + variables["[outputtag]"] = output; + variables["[tag2]"] = "ave"; + variables["[calc]"] = matrixCalculators[i]->getName(); + string distFileName = getOutputFileName("phylip",variables); + outputNames.push_back(distFileName); outputTypes["phylip"].push_back(distFileName); + ofstream outAve; + m->openOutputFile(distFileName, outAve); + outAve.setf(ios::fixed, ios::floatfield); outAve.setf(ios::showpoint); + + printSims(outAve, matrix); + + outAve.close(); + + variables["[tag2]"] = "std"; + distFileName = getOutputFileName("phylip",variables); + outputNames.push_back(distFileName); outputTypes["phylip"].push_back(distFileName); + ofstream outSTD; + m->openOutputFile(distFileName, outSTD); + outSTD.setf(ios::fixed, ios::floatfield); outSTD.setf(ios::showpoint); + + printSims(outSTD, stdmatrix); + + outSTD.close(); + + } + } return 0; } @@ -573,8 +767,8 @@ int MatrixOutputCommand::process(vector thisLookup){ /**************************************************************************************************/ int MatrixOutputCommand::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++) { @@ -598,7 +792,7 @@ int MatrixOutputCommand::driver(vector thisLookup, int star if (m->control_pressed) { return 1; } - seqDist temp(l, k, (1.0 - tempdata[0])); + seqDist temp(l, k, tempdata[0]); calcDists[i].push_back(temp); } }