X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=matrixoutputcommand.cpp;h=4b92c577fa3f5ef319fb8805eaef432e793c8d2e;hb=2ecee16fec29d4c525f740ec19b27962ca09c050;hp=0bfb0098eba2de2600ab4939da8b3e3197b1b4fa;hpb=bb8d942036365d2a66330944269f17841ed2388a;p=mothur.git diff --git a/matrixoutputcommand.cpp b/matrixoutputcommand.cpp index 0bfb009..4b92c57 100644 --- a/matrixoutputcommand.cpp +++ b/matrixoutputcommand.cpp @@ -19,6 +19,7 @@ vector MatrixOutputCommand::setParameters(){ 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 pmode("mode", "Multiple", "average-median", "average", "", "", "",false,false); parameters.push_back(pmode); CommandParameter pprocessors("processors", "Number", "", "1", "", "", "",false,false); 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); @@ -38,13 +39,14 @@ string MatrixOutputCommand::getHelpString(){ try { string helpString = ""; ValidCalculators validCalculator; - helpString += "The dist.shared command parameters are shared, groups, calc, output, processors, subsample, iters 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 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"; @@ -59,6 +61,26 @@ string MatrixOutputCommand::getHelpString(){ } } //********************************************************************************************************************** +string MatrixOutputCommand::getOutputFileNameTag(string type, string inputName=""){ + try { + string outputFileName = ""; + map >::iterator it; + + //is this a type this command creates + it = outputTypes.find(type); + if (it == outputTypes.end()) { m->mothurOut("[ERROR]: this command doesn't create a " + type + " output file.\n"); } + else { + if (type == "phylip") { outputFileName = "dist"; } + else { m->mothurOut("[ERROR]: No definition for type " + type + " output file tag.\n"); m->control_pressed = true; } + } + return outputFileName; + } + catch(exception& e) { + m->errorOut(e, "MatrixOutputCommand", "getOutputFileNameTag"); + exit(1); + } +} +//********************************************************************************************************************** MatrixOutputCommand::MatrixOutputCommand(){ try { abort = true; calledHelp = true; @@ -140,6 +162,9 @@ 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"; } + + 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 = ""; } @@ -298,13 +323,6 @@ 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;} - - numGroups = lookup.size(); - 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); - } if (subsample) { if (subsampleSize == -1) { //user has not set size, set size = smallest samples size @@ -330,8 +348,17 @@ int MatrixOutputCommand::execute(){ 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); + 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); + } + 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 @@ -432,18 +459,18 @@ void MatrixOutputCommand::printSims(ostream& out, vector< vector >& simM 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'; + 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{ - 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'; + for (int b = 0; b < simMatrix.size(); m++) { + out << lookup[b]->getGroup() << '\t'; + for (int n = 0; n < simMatrix[b].size(); n++) { + out << simMatrix[b][n] << '\t'; } out << endl; } @@ -460,11 +487,11 @@ int MatrixOutputCommand::process(vector thisLookup){ 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; thisIter++) { + for (int thisIter = 0; thisIter < iters+1; thisIter++) { vector thisItersLookup = thisLookup; - if (subsample) { + if (subsample && (thisIter != 0)) { SubSample sample; vector tempLabels; //dont need since we arent printing the sampled sharedRabunds @@ -617,15 +644,47 @@ int MatrixOutputCommand::process(vector thisLookup){ #endif } - calcDistsTotals.push_back(calcDists); - - if (subsample) { - + 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(); - for (int i = 0; i < calcDists.size(); i++) { calcDists[i].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; + } + + string distFileName = outputDir + m->getRootName(m->getSimpleName(sharedfile)) + matrixCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + "." + output + "." + getOutputFileNameTag("phylip"); + 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(); } } if (iters != 1) { @@ -636,34 +695,46 @@ int MatrixOutputCommand::process(vector thisLookup){ calcAverages[i].resize(calcDistsTotals[0][i].size()); for (int j = 0; j < calcAverages[i].size(); j++) { - calcAverages[i][j].seq1 = calcDists[i][j].seq1; - calcAverages[i][j].seq2 = calcDists[i][j].seq2; + calcAverages[i][j].seq1 = calcDistsTotals[0][i][j].seq1; + calcAverages[i][j].seq2 = calcDistsTotals[0][i][j].seq2; calcAverages[i][j].dist = 0.0; } } - - for (int thisIter = 0; thisIter < iters; thisIter++) { //sum all groups dists for each calculator - for (int i = 0; i < calcAverages.size(); i++) { //initialize sums to zero. + if (mode == "average") { + for (int thisIter = 0; thisIter < iters; thisIter++) { //sum all groups dists for each calculator + for (int i = 0; i < calcAverages.size(); i++) { //initialize sums to zero. + for (int j = 0; j < calcAverages[i].size(); j++) { + calcAverages[i][j].dist += calcDistsTotals[thisIter][i][j].dist; + if (m->debug) { m->mothurOut("[DEBUG]: Totaling for average calc: iter = " + toString(thisIter) + ", " + thisLookup[calcDistsTotals[thisIter][i][j].seq1]->getGroup() + " - " + thisLookup[calcDistsTotals[thisIter][i][j].seq2]->getGroup() + " distance = " + toString(calcDistsTotals[thisIter][i][j].dist) + ". New total = " + toString(calcAverages[i][j].dist) + ".\n"); } + } + } + } + + for (int i = 0; i < calcAverages.size(); i++) { //finds average. for (int j = 0; j < calcAverages[i].size(); j++) { - calcAverages[i][j].dist += calcDistsTotals[thisIter][i][j].dist; + calcAverages[i][j].dist /= (float) iters; } } - } - - for (int i = 0; i < calcAverages.size(); i++) { //finds average. - for (int j = 0; j < calcAverages[i].size(); j++) { - calcAverages[i][j].dist /= (float) iters; + }else { //find median + for (int i = 0; i < calcAverages.size(); i++) { //for each calc + for (int j = 0; j < calcAverages[i].size(); j++) { //for each comparison + vector dists; + for (int thisIter = 0; thisIter < iters; thisIter++) { //for each subsample + dists.push_back(calcDistsTotals[thisIter][i][j].dist); + } + sort(dists.begin(), dists.end()); + calcAverages[i][j].dist = dists[(iters/2)]; + } } } - //find standard deviation vector< vector > stdDev; stdDev.resize(matrixCalculators.size()); for (int i = 0; i < stdDev.size(); i++) { //initialize sums to zero. stdDev[i].resize(calcDistsTotals[0][i].size()); for (int j = 0; j < stdDev[i].size(); j++) { - stdDev[i][j].seq1 = calcDists[i][j].seq1; - stdDev[i][j].seq2 = calcDists[i][j].seq2; + stdDev[i][j].seq1 = calcDistsTotals[0][i][j].seq1; + stdDev[i][j].seq2 = calcDistsTotals[0][i][j].seq2; stdDev[i][j].dist = 0.0; } } @@ -706,7 +777,7 @@ int MatrixOutputCommand::process(vector thisLookup){ stdmatrix[column][row] = stdDist; } - string distFileName = outputDir + m->getRootName(m->getSimpleName(sharedfile)) + matrixCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + "." + output + ".ave.dist"; + string distFileName = outputDir + m->getRootName(m->getSimpleName(sharedfile)) + matrixCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + "." + output + ".ave." + getOutputFileNameTag("phylip"); outputNames.push_back(distFileName); outputTypes["phylip"].push_back(distFileName); ofstream outAve; m->openOutputFile(distFileName, outAve); @@ -716,7 +787,7 @@ int MatrixOutputCommand::process(vector thisLookup){ outAve.close(); - distFileName = outputDir + m->getRootName(m->getSimpleName(sharedfile)) + matrixCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + "." + output + ".std.dist"; + distFileName = outputDir + m->getRootName(m->getSimpleName(sharedfile)) + matrixCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + "." + output + ".std." + getOutputFileNameTag("phylip"); outputNames.push_back(distFileName); outputTypes["phylip"].push_back(distFileName); ofstream outSTD; m->openOutputFile(distFileName, outSTD); @@ -727,35 +798,6 @@ int MatrixOutputCommand::process(vector thisLookup){ outSTD.close(); } - }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; - } - - 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(); - } } return 0; @@ -769,6 +811,7 @@ 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++) {