X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=unifracweightedcommand.cpp;h=ad012b542b82a99ad0ac40782688c325360268b0;hb=bd93b1a6f9fe9a6a4a7ac2e9f106e5c83a438856;hp=cd7fec12320e3ee50e415a348df4a5b503e35763;hpb=aa9238c0a9e6e7aa0ed8b8b606b08ad4fd7dcfe3;p=mothur.git diff --git a/unifracweightedcommand.cpp b/unifracweightedcommand.cpp index cd7fec1..ad012b5 100644 --- a/unifracweightedcommand.cpp +++ b/unifracweightedcommand.cpp @@ -21,7 +21,7 @@ UnifracWeightedCommand::UnifracWeightedCommand(string option) { 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); @@ -40,7 +40,7 @@ UnifracWeightedCommand::UnifracWeightedCommand(string option) { //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 @@ -48,7 +48,7 @@ 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; } @@ -56,10 +56,13 @@ UnifracWeightedCommand::UnifracWeightedCommand(string option) { convert(itersString, iters); string temp = validParameter.validFile(parameters, "distance", false); if (temp == "not found") { temp = "false"; } - phylip = isTrue(temp); + phylip = m->isTrue(temp); - temp = validParameter.validFile(parameters, "random", false); if (temp == "not found") { temp = "true"; } - random = isTrue(temp); + 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 @@ -67,8 +70,8 @@ UnifracWeightedCommand::UnifracWeightedCommand(string option) { if (abort == false) { T = globaldata->gTree; tmap = globaldata->gTreemap; - sumFile = outputDir + getSimpleName(globaldata->getTreeFile()) + ".wsummary"; - openOutputFile(sumFile, outSum); + sumFile = outputDir + m->getSimpleName(globaldata->getTreeFile()) + ".wsummary"; + m->openOutputFile(sumFile, outSum); outputNames.push_back(sumFile); util = new SharedUtil(); @@ -97,7 +100,7 @@ void UnifracWeightedCommand::help(){ 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 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"); @@ -116,8 +119,7 @@ int UnifracWeightedCommand::execute() { if (abort == true) { return 0; } - 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... @@ -129,28 +131,22 @@ int UnifracWeightedCommand::execute() { //get weighted scores for users trees for (int i = 0; i < T.size(); i++) { - if (m->control_pressed) { - delete randT; - if (random) { delete reading; } - outSum.close(); - for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } - return 0; - } + if (m->control_pressed) { delete randT; 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"); } - userData = weighted->getValues(T[i]); //userData[0] = weightedscore + userData = weighted->getValues(T[i], processors, outputDir); //userData[0] = weightedscore if (m->control_pressed) { delete randT; - if (random) { delete reading; delete output; } + if (random) { delete output; } outSum.close(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; @@ -166,51 +162,54 @@ int UnifracWeightedCommand::execute() { utreeScores.push_back(userData[s]); } - //get scores for random trees - for (int j = 0; j < iters; j++) { - int count = 0; - for (int r=0; rgetCopy(T[i]); - - //create a random tree with same topology as T[i], but different labels - randT->assembleRandomUnifracTree(globaldata->Groups[r], globaldata->Groups[l]); - - if (m->control_pressed) { - delete randT; - if (random) { delete reading; delete output; } - outSum.close(); - for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } - return 0; + if (random) { + vector sums = weighted->getBranchLengthSums(T[i]); + + //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); + } + } + + #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(new linePair(startPos, numPairsPerProcessor)); } + } + #endif - - //get wscore of random tree - randomData = weighted->getValues(randT, globaldata->Groups[r], globaldata->Groups[l]); - - if (m->control_pressed) { - delete randT; - if (random) { delete reading; delete output; } - outSum.close(); - for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } - return 0; + + //get scores for random trees + for (int j = 0; j < iters; j++) { + int count = 0; + + #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + if(processors == 1){ + driver(T[i], randT, namesOfGroupCombos, 0, namesOfGroupCombos.size(), sums, rScores); + }else{ + createProcesses(T[i], randT, namesOfGroupCombos, sums, rScores); + for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear(); } - - //save scores - rScores[count].push_back(randomData[0]); - count++; - } + #else + driver(T[i], randT, namesOfGroupCombos, 0, namesOfGroupCombos.size(), sums, 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; } } - //update progress bar - reading->update(j); - - } - - //removeValidScoresDuplicates(); - //find the signifigance of the score for summary file - if (random) { + + //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()); @@ -230,6 +229,7 @@ int UnifracWeightedCommand::execute() { printWeightedFile(); delete output; + } //clear data @@ -239,16 +239,7 @@ int UnifracWeightedCommand::execute() { } - if (m->control_pressed) { - delete randT; - if (random) { delete reading; } - outSum.close(); - for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } - return 0; - } - - //finish progress bar - if (random) { reading->finish(); delete reading; } + if (m->control_pressed) { delete randT; outSum.close(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; } printWSummaryFile(); @@ -264,6 +255,8 @@ int UnifracWeightedCommand::execute() { 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(); for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); } @@ -277,6 +270,118 @@ int UnifracWeightedCommand::execute() { exit(1); } } +/**************************************************************************************************/ + +int UnifracWeightedCommand::createProcesses(Tree* t, Tree* randT, vector< vector > namesOfGroupCombos, vector& sums, vector< vector >& scores) { + try { +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) + int process = 1; + int num = 0; + 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, randT, namesOfGroupCombos, lines[process]->start, lines[process]->num, sums, scores); + + m->mothurOut("Merging results."); m->mothurOutEndLine(); + + //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][0] << '\t'; } out << endl; + out.close(); + + exit(0); + }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); } + } + + driver(t, randT, namesOfGroupCombos, lines[0]->start, lines[0]->num, sums, 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); + + for (int i = lines[process]->start; i < (lines[process]->start + lines[process]->num); i++) { in >> scores[i][0]; } + in.close(); + remove(s.c_str()); + } + + m->mothurOut("DONE."); m->mothurOutEndLine(); m->mothurOutEndLine(); + + return 0; +#endif + } + catch(exception& e) { + m->errorOut(e, "UnifracWeightedCommand", "createProcesses"); + exit(1); + } +} + +/**************************************************************************************************/ +int UnifracWeightedCommand::driver(Tree* t, Tree* randT, vector< vector > namesOfGroupCombos, int start, int num, vector& sums, vector< vector >& scores) { + try { + int count = 0; + int total = num; + int twentyPercent = (total * 0.20); + + 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, sums); + + if (m->control_pressed) { delete randT; return 0; } + + //save scores + scores[h].push_back(randomData[0]); + + count++; + + //report progress + if((count) % twentyPercent == 0){ m->mothurOut("Random comparison percentage complete: " + toString(int((count / (float)total) * 100.0))); m->mothurOutEndLine(); } + } + + m->mothurOut("Random comparison percentage complete: 100"); m->mothurOutEndLine(); + + return 0; + + } + catch(exception& e) { + m->errorOut(e, "UnifracWeightedCommand", "driver"); + exit(1); + } +} /***********************************************************/ void UnifracWeightedCommand::printWeightedFile() { try { @@ -348,10 +453,10 @@ 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"; + string phylipFileName = outputDir + m->getSimpleName(globaldata->getTreeFile()) + toString(i+1) + ".weighted.dist"; outputNames.push_back(phylipFileName); ofstream out; - openOutputFile(phylipFileName, out); + m->openOutputFile(phylipFileName, out); //output numSeqs out << globaldata->Groups.size() << endl; @@ -365,8 +470,8 @@ 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]); + dists[r][l] = utreeScores[count]; + dists[l][r] = utreeScores[count]; count++; } }