X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=matrixoutputcommand.cpp;h=310e32bf23e048c73ae37579ac2e8e2d1cb8b0ff;hp=cada2f16035f6bfea109aa4112908b9bb8c128f3;hb=1a20e24ee786195ab0e1cccd4f5aede7a88f3f4e;hpb=53171f07cc0c0e560e2b4ba2946f690d59fc2dc4 diff --git a/matrixoutputcommand.cpp b/matrixoutputcommand.cpp index cada2f1..310e32b 100644 --- a/matrixoutputcommand.cpp +++ b/matrixoutputcommand.cpp @@ -13,16 +13,17 @@ //********************************************************************************************************************** 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 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); 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 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); + 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); } @@ -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 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"; @@ -59,13 +61,27 @@ 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; setParameters(); vector tempOutNames; outputTypes["phylip"] = tempOutNames; - outputTypes["subsample"] = tempOutNames; } catch(exception& e) { m->errorOut(e, "MatrixOutputCommand", "MatrixOutputCommand"); @@ -100,7 +116,6 @@ MatrixOutputCommand::MatrixOutputCommand(string option) { //initialize outputTypes vector tempOutNames; outputTypes["phylip"] = tempOutNames; - outputTypes["subsample"] = 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); @@ -141,7 +156,10 @@ 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 = ""; } @@ -176,7 +194,7 @@ MatrixOutputCommand::MatrixOutputCommand(string option) { else { subsample = false; } } - if (subsample == false) { iters = 1; } + if (subsample == false) { iters = 0; } if (abort == false) { @@ -300,13 +318,6 @@ 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 @@ -332,8 +343,17 @@ 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 @@ -425,27 +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; } @@ -459,22 +484,40 @@ void MatrixOutputCommand::printSims(ostream& out, vector< vector >& simMa /***********************************************************/ int MatrixOutputCommand::process(vector thisLookup){ try { - EstOutput data; - vector subset; 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++) { + map variables; + variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(sharedfile)); + variables["[distance]"] = thisLookup[0]->getLabel(); + variables["[tag2]"] = ""; vector thisItersLookup = thisLookup; - if (subsample) { + if (subsample && (thisIter != 0)) { SubSample sample; vector tempLabels; //dont need since we arent printing the sampled sharedRabunds - thisItersLookup = sample.getSamplePreserve(thisLookup, tempLabels, subsampleSize); + + //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; } - cout << thisIter << endl; + if(processors == 1){ driver(thisItersLookup, 0, numGroups, calcDists); }else{ @@ -591,6 +634,9 @@ int MatrixOutputCommand::process(vector thisLookup){ //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++) { @@ -605,77 +651,66 @@ 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 < 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(); } } - if (iters != 1) { + if (iters != 0) { //we need to find the average distance and standard deviation for each groups distance - - vector< vector > calcAverages; calcAverages.resize(matrixCalculators.size()); - for (int i = 0; i < calcAverages.size(); i++) { //initialize sums to zero. - calcAverages[i].resize(calcDists[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].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. - for (int j = 0; j < calcAverages[i].size(); j++) { - calcAverages[i][j].dist += calcDistsTotals[thisIter][i][j].dist; - } - } - } - - 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; - } - } + vector< vector > calcAverages = m->getAverages(calcDistsTotals, mode); //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(calcDists[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].dist = 0.0; - } - } - - for (int thisIter = 0; thisIter < iters; thisIter++) { //compute the difference of each dist from the mean, and square the result of each - for (int i = 0; i < stdDev.size(); i++) { - for (int j = 0; j < stdDev[i].size(); j++) { - stdDev[i][j].dist += ((calcDistsTotals[thisIter][i][j].dist - calcAverages[i][j].dist) * (calcDistsTotals[thisIter][i][j].dist - calcAverages[i][j].dist)); - } - } - } - - for (int i = 0; i < stdDev.size(); i++) { //finds average. - for (int j = 0; j < stdDev[i].size(); j++) { - stdDev[i][j].dist /= (float) iters; - stdDev[i][j].dist = sqrt(stdDev[i][j].dist); - } - } + 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 + 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 + 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); } @@ -691,55 +726,36 @@ int MatrixOutputCommand::process(vector thisLookup){ stdmatrix[row][column] = stdDist; stdmatrix[column][row] = stdDist; } - - string distFileName = outputDir + m->getRootName(m->getSimpleName(sharedfile)) + matrixCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + ".results"; - outputNames.push_back(distFileName); outputTypes["subsample"].push_back(distFileName); - ofstream outDist; - m->openOutputFile(distFileName, outDist); - outDist.setf(ios::fixed, ios::floatfield); outDist.setf(ios::showpoint); - outDist << "Group1\tGroup2\tAverageDist\tStdDev\n"; - for (int m = 0; m < matrix.size(); m++) { - for (int n = 0; n < m; n++) { - outDist << lookup[m]->getGroup() << '\t' << lookup[n]->getGroup() << '\t'; - outDist << matrix[m][n] << '\t' << stdmatrix[m][n] << endl; - } - } - outDist.close(); - } - - //output averages as distance matrix - calcDists = calcAverages; - } - - 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; + 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); - matrix[row][column] = dist; - matrix[column][row] = dist; + 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(); + } - - 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; } @@ -751,8 +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++) {