X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=unifracweightedcommand.cpp;h=ad012b542b82a99ad0ac40782688c325360268b0;hb=bd93b1a6f9fe9a6a4a7ac2e9f106e5c83a438856;hp=f9cdd5a86458c6f3039e4392e6e8b27f7d5eaa08;hpb=d5bf2c1354d0811a33394d918b15620606560d58;p=mothur.git diff --git a/unifracweightedcommand.cpp b/unifracweightedcommand.cpp index f9cdd5a..ad012b5 100644 --- a/unifracweightedcommand.cpp +++ b/unifracweightedcommand.cpp @@ -15,13 +15,13 @@ UnifracWeightedCommand::UnifracWeightedCommand(string option) { globaldata = GlobalData::getInstance(); abort = false; Groups.clear(); - + //allow user to run help if(option == "help") { help(); abort = true; } else { //valid paramters for this command - string Array[] = {"groups","iters","distance","random"}; + string Array[] = {"groups","iters","distance","random","processors","outputdir","inputdir"}; vector myArray (Array, Array+(sizeof(Array)/sizeof(string))); OptionParser parser(option); @@ -35,14 +35,20 @@ UnifracWeightedCommand::UnifracWeightedCommand(string option) { } if (globaldata->gTree.size() == 0) {//no trees were read - mothurOut("You must execute the read.tree command, before you may execute the unifrac.weighted command."); mothurOutEndLine(); abort = true; } - + 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 += m->hasPath(globaldata->inputFileName); //if user entered a file with a path then preserve it + } + //check for optional parameter and set defaults // ...at some point should added some additional type checking... groups = validParameter.validFile(parameters, "groups", false); if (groups == "not found") { groups = ""; } else { - splitAtDash(groups, Groups); + m->splitAtDash(groups, Groups); globaldata->Groups = Groups; } @@ -50,10 +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 @@ -61,8 +70,9 @@ UnifracWeightedCommand::UnifracWeightedCommand(string option) { if (abort == false) { T = globaldata->gTree; tmap = globaldata->gTreemap; - sumFile = globaldata->getTreeFile() + ".wsummary"; - openOutputFile(sumFile, outSum); + sumFile = outputDir + m->getSimpleName(globaldata->getTreeFile()) + ".wsummary"; + m->openOutputFile(sumFile, outSum); + outputNames.push_back(sumFile); util = new SharedUtil(); string s; //to make work with setgroups @@ -77,7 +87,7 @@ UnifracWeightedCommand::UnifracWeightedCommand(string option) { } catch(exception& e) { - errorOut(e, "UnifracWeightedCommand", "UnifracWeightedCommand"); + m->errorOut(e, "UnifracWeightedCommand", "UnifracWeightedCommand"); exit(1); } } @@ -85,20 +95,20 @@ UnifracWeightedCommand::UnifracWeightedCommand(string option) { void UnifracWeightedCommand::help(){ try { - mothurOut("The unifrac.weighted command can only be executed after a successful read.tree command.\n"); - mothurOut("The unifrac.weighted command parameters are groups, iters, distance and random. No parameters are required.\n"); - 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"); - 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"); - mothurOut("The distance parameter allows you to create a distance file from the results. The default is false.\n"); - 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"); - mothurOut("The unifrac.weighted command should be in the following format: unifrac.weighted(groups=yourGroups, iters=yourIters).\n"); - mothurOut("Example unifrac.weighted(groups=A-B-C, iters=500).\n"); - mothurOut("The default value for groups is all the groups in your groupfile, and iters is 1000.\n"); - mothurOut("The unifrac.weighted command output two files: .weighted and .wsummary their descriptions are in the manual.\n"); - mothurOut("Note: No spaces between parameter labels (i.e. groups), '=' and parameters (i.e.yourGroups).\n\n"); + 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 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 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"); + m->mothurOut("The unifrac.weighted command output two files: .weighted and .wsummary their descriptions are in the manual.\n"); + m->mothurOut("Note: No spaces between parameter labels (i.e. groups), '=' and parameters (i.e.yourGroups).\n\n"); } catch(exception& e) { - errorOut(e, "UnifracWeightedCommand", "help"); + m->errorOut(e, "UnifracWeightedCommand", "help"); exit(1); } } @@ -109,8 +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... @@ -121,13 +130,28 @@ int UnifracWeightedCommand::execute() { //get weighted scores for users trees for (int i = 0; i < T.size(); i++) { + + 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(globaldata->getTreeFile() + toString(i+1) + ".weighted", itersString); } + if (random) { + 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], processors, outputDir); //userData[0] = weightedscore + + if (m->control_pressed) { + delete randT; + if (random) { delete output; } + outSum.close(); + for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } + return 0; + } - userData = weighted->getValues(T[i]); //userData[0] = weightedscore //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) { + 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); } } - - //update progress bar - reading->update(j); - - } + + #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 - //removeValidScoresDuplicates(); - //find the signifigance of the score for summary file - if (random) { + + //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(); + } + #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; } + } + + + //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()); @@ -173,7 +218,7 @@ int UnifracWeightedCommand::execute() { //so if you have 1000 random trees the index returned is 100 //then there are 900 trees with a score greater then you. //giving you a signifigance of 0.900 - int index = findIndex(userData[f], f); if (index == -1) { mothurOut("error in UnifracWeightedCommand"); mothurOutEndLine(); exit(1); } //error code + int index = findIndex(userData[f], f); if (index == -1) { m->mothurOut("error in UnifracWeightedCommand"); m->mothurOutEndLine(); exit(1); } //error code //the signifigance is the number of trees with the users score or higher WScoreSig.push_back((iters-index)/(float)iters); @@ -184,6 +229,7 @@ int UnifracWeightedCommand::execute() { printWeightedFile(); delete output; + } //clear data @@ -192,8 +238,8 @@ int UnifracWeightedCommand::execute() { validScores.clear(); } - //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(); @@ -204,11 +250,135 @@ int UnifracWeightedCommand::execute() { 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(); + for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); } + m->mothurOutEndLine(); + + return 0; + + } + catch(exception& e) { + m->errorOut(e, "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) { - errorOut(e, "UnifracWeightedCommand", "execute"); + m->errorOut(e, "UnifracWeightedCommand", "driver"); exit(1); } } @@ -231,7 +401,7 @@ void UnifracWeightedCommand::printWeightedFile() { } } catch(exception& e) { - errorOut(e, "UnifracWeightedCommand", "printWeightedFile"); + m->errorOut(e, "UnifracWeightedCommand", "printWeightedFile"); exit(1); } } @@ -242,7 +412,7 @@ void UnifracWeightedCommand::printWSummaryFile() { try { //column headers outSum << "Tree#" << '\t' << "Groups" << '\t' << "WScore" << '\t' << "WSig" << endl; - mothurOut("Tree#\tGroups\tWScore\tWSig"); mothurOutEndLine(); + m->mothurOut("Tree#\tGroups\tWScore\tWSig"); m->mothurOutEndLine(); //format output outSum.setf(ios::fixed, ios::floatfield); outSum.setf(ios::showpoint); @@ -255,16 +425,16 @@ 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; - mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t" + toString(WScoreSig[count])); mothurOutEndLine(); + m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t" + toString(WScoreSig[count])); m->mothurOutEndLine(); }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; - mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t<" + toString((1/float(iters)))); mothurOutEndLine(); + m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t<" + toString((1/float(iters)))); m->mothurOutEndLine(); } }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; - mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t0.00"); mothurOutEndLine(); + m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t0.00"); m->mothurOutEndLine(); } count++; } @@ -272,7 +442,7 @@ void UnifracWeightedCommand::printWSummaryFile() { outSum.close(); } catch(exception& e) { - errorOut(e, "UnifracWeightedCommand", "printWSummaryFile"); + m->errorOut(e, "UnifracWeightedCommand", "printWSummaryFile"); exit(1); } } @@ -283,9 +453,10 @@ void UnifracWeightedCommand::createPhylipFile() { //for each tree for (int i = 0; i < T.size(); i++) { - string phylipFileName = 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; @@ -299,8 +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++; } } @@ -308,7 +479,11 @@ void UnifracWeightedCommand::createPhylipFile() { //output to file for (int r=0; rGroups.size(); r++) { //output name - out << globaldata->Groups[r] << '\t'; + string name = globaldata->Groups[r]; + 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'; } @@ -318,7 +493,7 @@ void UnifracWeightedCommand::createPhylipFile() { } } catch(exception& e) { - errorOut(e, "UnifracWeightedCommand", "createPhylipFile"); + m->errorOut(e, "UnifracWeightedCommand", "createPhylipFile"); exit(1); } } @@ -331,7 +506,7 @@ int UnifracWeightedCommand::findIndex(float score, int index) { return rScores[index].size(); } catch(exception& e) { - errorOut(e, "UnifracWeightedCommand", "findIndex"); + m->errorOut(e, "UnifracWeightedCommand", "findIndex"); exit(1); } } @@ -376,7 +551,7 @@ void UnifracWeightedCommand::calculateFreqsCumuls() { } catch(exception& e) { - errorOut(e, "UnifracWeightedCommand", "calculateFreqsCums"); + m->errorOut(e, "UnifracWeightedCommand", "calculateFreqsCums"); exit(1); } }