X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=unifracweightedcommand.cpp;h=4c387da9332c6ca280a9359105cf8dfa837cd0a9;hb=0486bc2eed084ac387d2f59b6d23d13b2382daf7;hp=3c6c037147c3fb0a1a3da670dfaed6b63c2767e7;hpb=74844a60d80c6dd06e3fb02ee9b928424f9019b0;p=mothur.git diff --git a/unifracweightedcommand.cpp b/unifracweightedcommand.cpp index 3c6c037..4c387da 100644 --- a/unifracweightedcommand.cpp +++ b/unifracweightedcommand.cpp @@ -9,19 +9,70 @@ #include "unifracweightedcommand.h" +//********************************************************************************************************************** +vector UnifracWeightedCommand::getValidParameters(){ + try { + string Array[] = {"groups","iters","distance","random","processors","outputdir","inputdir"}; + vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); + return myArray; + } + catch(exception& e) { + m->errorOut(e, "UnifracWeightedCommand", "getValidParameters"); + exit(1); + } +} +//********************************************************************************************************************** +UnifracWeightedCommand::UnifracWeightedCommand(){ + try { + abort = true; calledHelp = true; + vector tempOutNames; + outputTypes["weighted"] = tempOutNames; + outputTypes["wsummary"] = tempOutNames; + outputTypes["phylip"] = tempOutNames; + outputTypes["column"] = tempOutNames; + } + catch(exception& e) { + m->errorOut(e, "UnifracWeightedCommand", "UnifracWeightedCommand"); + exit(1); + } +} +//********************************************************************************************************************** +vector UnifracWeightedCommand::getRequiredParameters(){ + try { + vector myArray; + return myArray; + } + catch(exception& e) { + m->errorOut(e, "UnifracWeightedCommand", "getRequiredParameters"); + exit(1); + } +} +//********************************************************************************************************************** +vector UnifracWeightedCommand::getRequiredFiles(){ + try { + string Array[] = {"tree","group"}; + vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); + + return myArray; + } + catch(exception& e) { + m->errorOut(e, "UnifracWeightedCommand", "getRequiredFiles"); + exit(1); + } +} /***********************************************************/ UnifracWeightedCommand::UnifracWeightedCommand(string option) { try { globaldata = GlobalData::getInstance(); - abort = false; + abort = false; calledHelp = false; Groups.clear(); //allow user to run help - if(option == "help") { help(); abort = true; } + if(option == "help") { help(); abort = true; calledHelp = true; } else { //valid paramters for this command - string Array[] = {"groups","iters","distance","random","outputdir","inputdir"}; + string Array[] = {"groups","iters","distance","random","processors","outputdir","inputdir"}; vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); OptionParser parser(option); @@ -34,13 +85,20 @@ UnifracWeightedCommand::UnifracWeightedCommand(string option) { if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; } } + //initialize outputTypes + vector tempOutNames; + outputTypes["weighted"] = tempOutNames; + outputTypes["wsummary"] = tempOutNames; + outputTypes["phylip"] = tempOutNames; + outputTypes["column"] = tempOutNames; + if (globaldata->gTree.size() == 0) {//no trees were read m->mothurOut("You must execute the read.tree command, before you may execute the unifrac.weighted command."); m->mothurOutEndLine(); 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 = ""; - outputDir += hasPath(globaldata->inputFileName); //if user entered a file with a path then preserve it + outputDir += m->hasPath(globaldata->inputFileName); //if user entered a file with a path then preserve it } //check for optional parameter and set defaults @@ -48,18 +106,25 @@ UnifracWeightedCommand::UnifracWeightedCommand(string option) { groups = validParameter.validFile(parameters, "groups", false); if (groups == "not found") { groups = ""; } else { - splitAtDash(groups, Groups); + m->splitAtDash(groups, Groups); globaldata->Groups = Groups; } itersString = validParameter.validFile(parameters, "iters", false); if (itersString == "not found") { itersString = "1000"; } convert(itersString, iters); - string temp = validParameter.validFile(parameters, "distance", false); if (temp == "not found") { temp = "false"; } - phylip = isTrue(temp); - - temp = validParameter.validFile(parameters, "random", false); if (temp == "not found") { temp = "true"; } - random = isTrue(temp); + string temp = validParameter.validFile(parameters, "distance", false); + if (temp == "not found") { phylip = false; outputForm = ""; } + else{ + if ((temp == "lt") || (temp == "column") || (temp == "square")) { phylip = true; outputForm = temp; } + else { m->mothurOut("Options for distance are: lt, square, or column. Using lt."); m->mothurOutEndLine(); phylip = true; outputForm = "lt"; } + } + + temp = validParameter.validFile(parameters, "random", false); if (temp == "not found") { temp = "F"; } + random = m->isTrue(temp); + + temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = "1"; } + convert(temp, processors); if (!random) { iters = 0; } //turn off random calcs @@ -67,9 +132,9 @@ UnifracWeightedCommand::UnifracWeightedCommand(string option) { if (abort == false) { T = globaldata->gTree; tmap = globaldata->gTreemap; - sumFile = outputDir + getSimpleName(globaldata->getTreeFile()) + ".wsummary"; - openOutputFile(sumFile, outSum); - outputNames.push_back(sumFile); + sumFile = outputDir + m->getSimpleName(globaldata->getTreeFile()) + ".wsummary"; + m->openOutputFile(sumFile, outSum); + outputNames.push_back(sumFile); outputTypes["wsummary"].push_back(sumFile); util = new SharedUtil(); string s; //to make work with setgroups @@ -93,11 +158,12 @@ UnifracWeightedCommand::UnifracWeightedCommand(string option) { void UnifracWeightedCommand::help(){ try { m->mothurOut("The unifrac.weighted command can only be executed after a successful read.tree command.\n"); - m->mothurOut("The unifrac.weighted command parameters are groups, iters, distance and random. No parameters are required.\n"); + m->mothurOut("The unifrac.weighted command parameters are groups, iters, distance, processors and random. No parameters are required.\n"); m->mothurOut("The groups parameter allows you to specify which of the groups in your groupfile you would like analyzed. You must enter at least 2 valid groups.\n"); m->mothurOut("The group names are separated by dashes. The iters parameter allows you to specify how many random trees you would like compared to your tree.\n"); m->mothurOut("The distance parameter allows you to create a distance file from the results. The default is false.\n"); - m->mothurOut("The random parameter allows you to shut off the comparison to random trees. The default is true, meaning compare your trees with randomly generated trees.\n"); + m->mothurOut("The random parameter allows you to shut off the comparison to random trees. The default is false, meaning don't compare your trees with randomly generated trees.\n"); + m->mothurOut("The processors parameter allows you to specify the number of processors to use. The default is 1.\n"); m->mothurOut("The unifrac.weighted command should be in the following format: unifrac.weighted(groups=yourGroups, iters=yourIters).\n"); m->mothurOut("Example unifrac.weighted(groups=A-B-C, iters=500).\n"); m->mothurOut("The default value for groups is all the groups in your groupfile, and iters is 1000.\n"); @@ -114,30 +180,32 @@ void UnifracWeightedCommand::help(){ int UnifracWeightedCommand::execute() { try { - if (abort == true) { return 0; } + if (abort == true) { if (calledHelp) { return 0; } return 2; } - Progress* reading; - if (random) { reading = new Progress("Comparing to random:", iters); } + int start = time(NULL); //get weighted for users tree userData.resize(numComp,0); //data[0] = weightedscore AB, data[1] = weightedscore AC... randomData.resize(numComp,0); //data[0] = weightedscore AB, data[1] = weightedscore AC... - //create new tree with same num nodes and leaves as users - randT = new Tree(); - //get weighted scores for users trees for (int i = 0; i < T.size(); i++) { + + if (m->control_pressed) { outSum.close(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; } + counter = 0; rScores.resize(numComp); //data[0] = weightedscore AB, data[1] = weightedscore AC... uScores.resize(numComp); //data[0] = weightedscore AB, data[1] = weightedscore AC... if (random) { - output = new ColumnFile(outputDir + getSimpleName(globaldata->getTreeFile()) + toString(i+1) + ".weighted", itersString); - outputNames.push_back(outputDir + getSimpleName(globaldata->getTreeFile()) + toString(i+1) + ".weighted"); + output = new ColumnFile(outputDir + m->getSimpleName(globaldata->getTreeFile()) + toString(i+1) + ".weighted", itersString); + outputNames.push_back(outputDir + m->getSimpleName(globaldata->getTreeFile()) + toString(i+1) + ".weighted"); + outputTypes["weighted"].push_back(outputDir + m->getSimpleName(globaldata->getTreeFile()) + toString(i+1) + ".weighted"); } - userData = weighted->getValues(T[i]); //userData[0] = weightedscore + userData = weighted->getValues(T[i], processors, outputDir); //userData[0] = weightedscore + + if (m->control_pressed) { if (random) { delete output; } outSum.close(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; } //save users score for (int s=0; sgetCopy(T[i]); - - //create a random tree with same topology as T[i], but different labels - randT->assembleRandomUnifracTree(globaldata->Groups[r], globaldata->Groups[l]); - //get wscore of random tree - randomData = weighted->getValues(randT, globaldata->Groups[r], globaldata->Groups[l]); - - //save scores - rScores[count].push_back(randomData[0]); - count++; + if (random) { + + //calculate number of comparisons i.e. with groups A,B,C = AB, AC, BC = 3; + vector< vector > namesOfGroupCombos; + for (int a=0; a groups; groups.push_back(globaldata->Groups[a]); groups.push_back(globaldata->Groups[l]); + namesOfGroupCombos.push_back(groups); } } - //update progress bar - reading->update(j); - - } + lines.clear(); + + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + if(processors != 1){ + int numPairs = namesOfGroupCombos.size(); + int numPairsPerProcessor = numPairs / processors; + + for (int i = 0; i < processors; i++) { + int startPos = i * numPairsPerProcessor; + if(i == processors - 1){ + numPairsPerProcessor = numPairs - i * numPairsPerProcessor; + } + lines.push_back(linePair(startPos, numPairsPerProcessor)); + } + } + #endif - //removeValidScoresDuplicates(); - //find the signifigance of the score for summary file - if (random) { + + //get scores for random trees + for (int j = 0; j < iters; j++) { + + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + if(processors == 1){ + driver(T[i], namesOfGroupCombos, 0, namesOfGroupCombos.size(), rScores); + }else{ + createProcesses(T[i], namesOfGroupCombos, rScores); + } + #else + driver(T[i], namesOfGroupCombos, 0, namesOfGroupCombos.size(), rScores); + #endif + + if (m->control_pressed) { delete output; outSum.close(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; } + + //report progress +// m->mothurOut("Iter: " + toString(j+1)); m->mothurOutEndLine(); + } + lines.clear(); + + //find the signifigance of the score for summary file for (int f = 0; f < numComp; f++) { //sort random scores sort(rScores[f].begin(), rScores[f].end()); @@ -194,6 +285,7 @@ int UnifracWeightedCommand::execute() { printWeightedFile(); delete output; + } //clear data @@ -202,8 +294,8 @@ int UnifracWeightedCommand::execute() { validScores.clear(); } - //finish progress bar - if (random) { reading->finish(); delete reading; } + + if (m->control_pressed) { outSum.close(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; } printWSummaryFile(); @@ -212,7 +304,13 @@ int UnifracWeightedCommand::execute() { //clear out users groups globaldata->Groups.clear(); - delete randT; + + if (m->control_pressed) { + for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } + return 0; + } + + m->mothurOut("It took " + toString(time(NULL) - start) + " secs to run unifrac.weighted."); m->mothurOutEndLine(); m->mothurOutEndLine(); m->mothurOut("Output File Names: "); m->mothurOutEndLine(); @@ -227,6 +325,111 @@ int UnifracWeightedCommand::execute() { exit(1); } } +/**************************************************************************************************/ + +int UnifracWeightedCommand::createProcesses(Tree* t, vector< vector > namesOfGroupCombos, vector< vector >& scores) { + try { +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + int process = 1; + vector processIDS; + + EstOutput results; + + //loop through and create all the processes you want + while (process != processors) { + int pid = fork(); + + if (pid > 0) { + processIDS.push_back(pid); //create map from line number to pid so you can append files in correct order later + process++; + }else if (pid == 0){ + driver(t, namesOfGroupCombos, lines[process].start, lines[process].num, scores); + + //pass numSeqs to parent + ofstream out; + string tempFile = outputDir + toString(getpid()) + ".weightedcommand.results.temp"; + m->openOutputFile(tempFile, out); + for (int i = lines[process].start; i < (lines[process].start + lines[process].num); i++) { out << scores[i][(scores[i].size()-1)] << '\t'; } out << endl; + out.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); + } + } + + driver(t, namesOfGroupCombos, lines[0].start, lines[0].num, scores); + + //force parent to wait until all the processes are done + for (int i=0;i<(processors-1);i++) { + int temp = processIDS[i]; + wait(&temp); + } + + //get data created by processes + for (int i=0;i<(processors-1);i++) { + + ifstream in; + string s = outputDir + toString(processIDS[i]) + ".weightedcommand.results.temp"; + m->openInputFile(s, in); + + double tempScore; + for (int j = lines[(i+1)].start; j < (lines[(i+1)].start + lines[(i+1)].num); j++) { in >> tempScore; scores[j].push_back(tempScore); } + in.close(); + remove(s.c_str()); + } + + return 0; +#endif + } + catch(exception& e) { + m->errorOut(e, "UnifracWeightedCommand", "createProcesses"); + exit(1); + } +} + +/**************************************************************************************************/ +int UnifracWeightedCommand::driver(Tree* t, vector< vector > namesOfGroupCombos, int start, int num, vector< vector >& scores) { + try { + Tree* randT = new Tree(); + + for (int h = start; h < (start+num); h++) { + + if (m->control_pressed) { return 0; } + + //initialize weighted score + string groupA = namesOfGroupCombos[h][0]; + string groupB = namesOfGroupCombos[h][1]; + + //copy T[i]'s info. + randT->getCopy(t); + + //create a random tree with same topology as T[i], but different labels + randT->assembleRandomUnifracTree(groupA, groupB); + + if (m->control_pressed) { delete randT; return 0; } + + //get wscore of random tree + EstOutput randomData = weighted->getValues(randT, groupA, groupB); + + if (m->control_pressed) { delete randT; return 0; } + + //save scores + scores[h].push_back(randomData[0]); + } + + delete randT; + + return 0; + + } + catch(exception& e) { + m->errorOut(e, "UnifracWeightedCommand", "driver"); + exit(1); + } +} /***********************************************************/ void UnifracWeightedCommand::printWeightedFile() { try { @@ -270,16 +473,16 @@ void UnifracWeightedCommand::printWSummaryFile() { if (WScoreSig[count] > (1/(float)iters)) { outSum << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << WScoreSig[count] << endl; cout << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << WScoreSig[count] << endl; - m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t" + toString(WScoreSig[count])); m->mothurOutEndLine(); + m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t" + toString(WScoreSig[count]) + "\n"); }else{ outSum << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << "<" << (1/float(iters)) << endl; cout << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << "<" << (1/float(iters)) << endl; - m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t<" + toString((1/float(iters)))); m->mothurOutEndLine(); + m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t<" + toString((1/float(iters))) + "\n"); } }else{ outSum << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << "0.00" << endl; cout << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << "0.00" << endl; - m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t0.00"); m->mothurOutEndLine(); + m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t0.00\n"); } count++; } @@ -298,14 +501,23 @@ void UnifracWeightedCommand::createPhylipFile() { //for each tree for (int i = 0; i < T.size(); i++) { - string phylipFileName = outputDir + getSimpleName(globaldata->getTreeFile()) + toString(i+1) + ".weighted.dist"; - outputNames.push_back(phylipFileName); - ofstream out; - openOutputFile(phylipFileName, out); + string phylipFileName; + if ((outputForm == "lt") || (outputForm == "square")) { + phylipFileName = outputDir + m->getSimpleName(globaldata->getTreeFile()) + toString(i+1) + ".weighted.phylip.dist"; + outputNames.push_back(phylipFileName); outputTypes["phylip"].push_back(phylipFileName); + }else { //column + phylipFileName = outputDir + m->getSimpleName(globaldata->getTreeFile()) + toString(i+1) + ".weighted.column.dist"; + outputNames.push_back(phylipFileName); outputTypes["column"].push_back(phylipFileName); + } - //output numSeqs - out << globaldata->Groups.size() << endl; + ofstream out; + m->openOutputFile(phylipFileName, out); + if ((outputForm == "lt") || (outputForm == "square")) { + //output numSeqs + out << globaldata->Groups.size() << endl; + } + //make matrix with scores in it vector< vector > dists; dists.resize(globaldata->Groups.size()); for (int i = 0; i < globaldata->Groups.size(); i++) { @@ -314,9 +526,9 @@ void UnifracWeightedCommand::createPhylipFile() { //flip it so you can print it for (int r=0; rGroups.size(); r++) { - for (int l = r+1; l < globaldata->Groups.size(); l++) { - dists[r][l] = (1.0 - utreeScores[count]); - dists[l][r] = (1.0 - utreeScores[count]); + for (int l = 0; l < r; l++) { + dists[r][l] = utreeScores[count]; + dists[l][r] = utreeScores[count]; count++; } } @@ -328,11 +540,30 @@ void UnifracWeightedCommand::createPhylipFile() { if (name.length() < 10) { //pad with spaces to make compatible while (name.length() < 10) { name += " "; } } - out << name << '\t'; - //output distances - for (int l = 0; l < r; l++) { out << dists[r][l] << '\t'; } - out << endl; + if (outputForm == "lt") { + out << name << '\t'; + + //output distances + for (int l = 0; l < r; l++) { out << dists[r][l] << '\t'; } + out << endl; + }else if (outputForm == "square") { + out << name << '\t'; + + //output distances + for (int l = 0; l < globaldata->Groups.size(); l++) { out << dists[r][l] << '\t'; } + out << endl; + }else{ + //output distances + for (int l = 0; l < r; l++) { + string otherName = globaldata->Groups[l]; + if (otherName.length() < 10) { //pad with spaces to make compatible + while (otherName.length() < 10) { otherName += " "; } + } + + out << name << '\t' << otherName << dists[r][l] << endl; + } + } } out.close(); }