X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=preclustercommand.cpp;h=c2564c1f769a835fd85559f10811672eb6a34e0d;hp=67b2f31bebe6185fef4137d9c5e839a35b23f0b8;hb=1a20e24ee786195ab0e1cccd4f5aede7a88f3f4e;hpb=ae57e166b2ed7b475ec3f466106bd76fabadd063 diff --git a/preclustercommand.cpp b/preclustercommand.cpp index 67b2f31..c2564c1 100644 --- a/preclustercommand.cpp +++ b/preclustercommand.cpp @@ -8,22 +8,21 @@ */ #include "preclustercommand.h" -#include "sequenceparser.h" #include "deconvolutecommand.h" -//********************************************************************************************************************** -inline bool comparePriority(seqPNode first, seqPNode second) { return (first.numIdentical > second.numIdentical); } - //********************************************************************************************************************** vector PreClusterCommand::setParameters(){ try { - CommandParameter pfasta("fasta", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(pfasta); - CommandParameter pname("name", "InputTypes", "", "", "none", "none", "none",false,false); parameters.push_back(pname); - CommandParameter pgroup("group", "InputTypes", "", "", "none", "none", "none",false,false); parameters.push_back(pgroup); - CommandParameter pdiffs("diffs", "Number", "", "0", "", "", "",false,false); parameters.push_back(pdiffs); - CommandParameter pbygroup("bygroup", "Boolean", "", "T", "", "", "",false,false); parameters.push_back(pbygroup); - CommandParameter pinputdir("inputdir", "String", "", "", "", "", "",false,false); parameters.push_back(pinputdir); - CommandParameter poutputdir("outputdir", "String", "", "", "", "", "",false,false); parameters.push_back(poutputdir); + CommandParameter pfasta("fasta", "InputTypes", "", "", "none", "none", "none","fasta-name",false,true,true); parameters.push_back(pfasta); + CommandParameter pname("name", "InputTypes", "", "", "NameCount", "none", "none","name",false,false,true); parameters.push_back(pname); + CommandParameter pcount("count", "InputTypes", "", "", "NameCount-CountGroup", "none", "none","count",false,false,true); parameters.push_back(pcount); + CommandParameter pgroup("group", "InputTypes", "", "", "CountGroup", "none", "none","",false,false,true); parameters.push_back(pgroup); + CommandParameter pdiffs("diffs", "Number", "", "1", "", "", "","",false,false,true); parameters.push_back(pdiffs); + CommandParameter pprocessors("processors", "Number", "", "1", "", "", "","",false,false,true); parameters.push_back(pprocessors); + CommandParameter ptopdown("topdown", "Boolean", "", "T", "", "", "","",false,false); parameters.push_back(ptopdown); + + 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); } @@ -40,11 +39,12 @@ string PreClusterCommand::getHelpString(){ string helpString = ""; helpString += "The pre.cluster command groups sequences that are within a given number of base mismatches.\n"; helpString += "The pre.cluster command outputs a new fasta and name file.\n"; - helpString += "The pre.cluster command parameters are fasta, names and diffs. The fasta parameter is required. \n"; - helpString += "The names parameter allows you to give a list of seqs that are identical. This file is 2 columns, first column is name or representative sequence, second column is a list of its identical sequences separated by commas.\n"; - helpString += "The group parameter allows you to provide a group file. \n"; + helpString += "The pre.cluster command parameters are fasta, name, group, count, topdown, processors and diffs. The fasta parameter is required. \n"; + helpString += "The name parameter allows you to give a list of seqs that are identical. This file is 2 columns, first column is name or representative sequence, second column is a list of its identical sequences separated by commas.\n"; + helpString += "The group parameter allows you to provide a group file so you can cluster by group. \n"; + helpString += "The count parameter allows you to provide a count file so you can cluster by group. \n"; helpString += "The diffs parameter allows you to specify maximum number of mismatched bases allowed between sequences in a grouping. The default is 1.\n"; - helpString += "The bygroup parameter allows you to indicate you whether you want to cluster sequences only within each group, default=T, when a groupfile is given. \n"; + helpString += "The topdown parameter allows you to specify whether to cluster from largest abundance to smallest or smallest to largest. Default=T, meaning largest to smallest.\n"; helpString += "The pre.cluster command should be in the following format: \n"; helpString += "pre.cluster(fasta=yourFastaFile, names=yourNamesFile, diffs=yourMaxDiffs) \n"; helpString += "Example pre.cluster(fasta=amazon.fasta, diffs=2).\n"; @@ -56,7 +56,24 @@ string PreClusterCommand::getHelpString(){ exit(1); } } - +//********************************************************************************************************************** +string PreClusterCommand::getOutputPattern(string type) { + try { + string pattern = ""; + + if (type == "fasta") { pattern = "[filename],precluster,[extension]"; } + else if (type == "name") { pattern = "[filename],precluster.names"; } + else if (type == "count") { pattern = "[filename],precluster.count_table"; } + else if (type == "map") { pattern = "[filename],precluster.map"; } + else { m->mothurOut("[ERROR]: No definition for type " + type + " output pattern.\n"); m->control_pressed = true; } + + return pattern; + } + catch(exception& e) { + m->errorOut(e, "PreClusterCommand", "getOutputPattern"); + exit(1); + } +} //********************************************************************************************************************** PreClusterCommand::PreClusterCommand(){ try { @@ -65,6 +82,8 @@ PreClusterCommand::PreClusterCommand(){ vector tempOutNames; outputTypes["fasta"] = tempOutNames; outputTypes["name"] = tempOutNames; + outputTypes["count"] = tempOutNames; + outputTypes["map"] = tempOutNames; } catch(exception& e) { m->errorOut(e, "PreClusterCommand", "PreClusterCommand"); @@ -99,6 +118,8 @@ PreClusterCommand::PreClusterCommand(string option) { vector tempOutNames; outputTypes["fasta"] = tempOutNames; outputTypes["name"] = tempOutNames; + outputTypes["map"] = tempOutNames; + outputTypes["count"] = 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); @@ -128,6 +149,14 @@ PreClusterCommand::PreClusterCommand(string option) { //if the user has not given a path then, add inputdir. else leave path alone. if (path == "") { parameters["group"] = inputDir + it->second; } } + + it = parameters.find("count"); + //user has given a template file + if(it != parameters.end()){ + path = m->hasPath(it->second); + //if the user has not given a path then, add inputdir. else leave path alone. + if (path == "") { parameters["count"] = inputDir + it->second; } + } } //check for required parameters @@ -150,26 +179,49 @@ PreClusterCommand::PreClusterCommand(string option) { // ...at some point should added some additional type checking... namefile = validParameter.validFile(parameters, "name", true); if (namefile == "not found") { namefile = ""; } - else if (namefile == "not open") { abort = true; } + else if (namefile == "not open") { namefile = ""; abort = true; } else { m->setNameFile(namefile); } groupfile = validParameter.validFile(parameters, "group", true); - if (groupfile == "not found") { groupfile = ""; } + if (groupfile == "not found") { groupfile = ""; bygroup = false; } else if (groupfile == "not open") { abort = true; groupfile = ""; } - else { m->setGroupFile(groupfile); } + else { m->setGroupFile(groupfile); bygroup = true; } + + countfile = validParameter.validFile(parameters, "count", true); + if (countfile == "not found") { countfile = ""; } + else if (countfile == "not open") { abort = true; countfile = ""; } + else { + m->setCountTableFile(countfile); + ct.readTable(countfile); + if (ct.hasGroupInfo()) { bygroup = true; } + else { bygroup = false; } + } + + if ((namefile != "") && (countfile != "")) { + m->mothurOut("[ERROR]: you may only use one of the following: name or count."); m->mothurOutEndLine(); abort = true; + } + + if ((groupfile != "") && (countfile != "")) { + m->mothurOut("[ERROR]: you may only use one of the following: group or count."); m->mothurOutEndLine(); abort=true; + } + string temp = validParameter.validFile(parameters, "diffs", false); if(temp == "not found"){ temp = "1"; } - convert(temp, diffs); + m->mothurConvert(temp, diffs); - temp = validParameter.validFile(parameters, "bygroup", false); - if (temp == "not found") { - if (groupfile != "") { temp = "T"; } - else { temp = "F"; } - } - bygroup = m->isTrue(temp); + temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); } + m->setProcessors(temp); + m->mothurConvert(temp, processors); - if (bygroup && (groupfile == "")) { m->mothurOut("You cannot set bygroup=T, unless you provide a groupfile."); m->mothurOutEndLine(); abort=true; } - + temp = validParameter.validFile(parameters, "topdown", false); if(temp == "not found"){ temp = "T"; } + topdown = m->isTrue(temp); + + if (countfile == "") { + if (namefile == "") { + vector files; files.push_back(fastafile); + parser.getNameFile(files); + } + } } } @@ -188,100 +240,94 @@ int PreClusterCommand::execute(){ int start = time(NULL); string fileroot = outputDir + m->getRootName(m->getSimpleName(fastafile)); - string newFastaFile = fileroot + "precluster" + m->getExtension(fastafile); - string newNamesFile = fileroot + "precluster.names"; + map variables; + variables["[filename]"] = fileroot; + string newNamesFile = getOutputFileName("name",variables); + string newCountFile = getOutputFileName("count",variables); + string newMapFile = getOutputFileName("map",variables); //add group name if by group + variables["[extension]"] = m->getExtension(fastafile); + string newFastaFile = getOutputFileName("fasta", variables); + outputNames.push_back(newFastaFile); outputTypes["fasta"].push_back(newFastaFile); + if (countfile == "") { outputNames.push_back(newNamesFile); outputTypes["name"].push_back(newNamesFile); } + else { outputNames.push_back(newCountFile); outputTypes["count"].push_back(newCountFile); } if (bygroup) { //clear out old files ofstream outFasta; m->openOutputFile(newFastaFile, outFasta); outFasta.close(); ofstream outNames; m->openOutputFile(newNamesFile, outNames); outNames.close(); + newMapFile = fileroot + "precluster."; //parse fasta and name file by group - SequenceParser* parser; - if (namefile != "") { parser = new SequenceParser(groupfile, fastafile, namefile); } - else { parser = new SequenceParser(groupfile, fastafile); } - - vector groups = parser->getNamesOfGroups(); - - //precluster each group - for (int i = 0; i < groups.size(); i++) { - - start = time(NULL); - - if (m->control_pressed) { delete parser; m->mothurRemove(newFastaFile); m->mothurRemove(newNamesFile); return 0; } - - m->mothurOutEndLine(); m->mothurOut("Processing group " + groups[i] + ":"); m->mothurOutEndLine(); - - map thisNameMap; - if (namefile != "") { thisNameMap = parser->getNameMap(groups[i]); } - vector thisSeqs = parser->getSeqs(groups[i]); - - //fill alignSeqs with this groups info. - int numSeqs = loadSeqs(thisNameMap, thisSeqs); - - if (m->control_pressed) { delete parser; m->mothurRemove(newFastaFile); m->mothurRemove(newNamesFile); return 0; } - - if (diffs > length) { m->mothurOut("Error: diffs is greater than your sequence length."); m->mothurOutEndLine(); return 0; } - - int count = process(); - - if (m->control_pressed) { delete parser; m->mothurRemove(newFastaFile); m->mothurRemove(newNamesFile); return 0; } - - m->mothurOut("Total number of sequences before precluster was " + toString(alignSeqs.size()) + "."); m->mothurOutEndLine(); - m->mothurOut("pre.cluster removed " + toString(count) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine(); - printData(newFastaFile, newNamesFile); - - m->mothurOut("It took " + toString(time(NULL) - start) + " secs to cluster " + toString(numSeqs) + " sequences."); m->mothurOutEndLine(); - + vector groups; + if (countfile != "") { + cparser = new SequenceCountParser(countfile, fastafile); + groups = cparser->getNamesOfGroups(); + }else { + if (namefile != "") { parser = new SequenceParser(groupfile, fastafile, namefile); } + else { parser = new SequenceParser(groupfile, fastafile); } + groups = parser->getNamesOfGroups(); } + + if(processors == 1) { driverGroups(newFastaFile, newNamesFile, newMapFile, 0, groups.size(), groups); } + else { createProcessesGroups(newFastaFile, newNamesFile, newMapFile, groups); } - //run unique.seqs for deconvolute results - string inputString = "fasta=" + newFastaFile; - if (namefile != "") { inputString += ", name=" + newNamesFile; } - m->mothurOutEndLine(); - m->mothurOut("/******************************************/"); m->mothurOutEndLine(); - m->mothurOut("Running command: unique.seqs(" + inputString + ")"); m->mothurOutEndLine(); - - Command* uniqueCommand = new DeconvoluteCommand(inputString); - uniqueCommand->execute(); - - map > filenames = uniqueCommand->getOutputFiles(); - - delete uniqueCommand; - - m->mothurOut("/******************************************/"); m->mothurOutEndLine(); - - newNamesFile = filenames["name"][0]; - newFastaFile = filenames["fasta"][0]; - + if (countfile != "") { + mergeGroupCounts(newCountFile, newNamesFile, newFastaFile); + delete cparser; + }else { + delete parser; + //run unique.seqs for deconvolute results + string inputString = "fasta=" + newFastaFile; + if (namefile != "") { inputString += ", name=" + newNamesFile; } + m->mothurOutEndLine(); + m->mothurOut("/******************************************/"); m->mothurOutEndLine(); + m->mothurOut("Running command: unique.seqs(" + inputString + ")"); m->mothurOutEndLine(); + m->mothurCalling = true; + + Command* uniqueCommand = new DeconvoluteCommand(inputString); + uniqueCommand->execute(); + + map > filenames = uniqueCommand->getOutputFiles(); + + delete uniqueCommand; + m->mothurCalling = false; + m->mothurOut("/******************************************/"); m->mothurOutEndLine(); + + m->renameFile(filenames["fasta"][0], newFastaFile); + m->renameFile(filenames["name"][0], newNamesFile); + } + if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } return 0; } + m->mothurOut("It took " + toString(time(NULL) - start) + " secs to run pre.cluster."); m->mothurOutEndLine(); + }else { if (namefile != "") { readNameFile(); } //reads fasta file and return number of seqs int numSeqs = readFASTA(); //fills alignSeqs and makes all seqs active - if (m->control_pressed) { return 0; } + if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } return 0; } if (numSeqs == 0) { m->mothurOut("Error reading fasta file...please correct."); m->mothurOutEndLine(); return 0; } if (diffs > length) { m->mothurOut("Error: diffs is greater than your sequence length."); m->mothurOutEndLine(); return 0; } - int count = process(); + int count = process(newMapFile); + outputNames.push_back(newMapFile); outputTypes["map"].push_back(newMapFile); - if (m->control_pressed) { return 0; } + if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } return 0; } m->mothurOut("Total number of sequences before precluster was " + toString(alignSeqs.size()) + "."); m->mothurOutEndLine(); m->mothurOut("pre.cluster removed " + toString(count) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine(); - printData(newFastaFile, newNamesFile); - + if (countfile != "") { newNamesFile = newCountFile; } + printData(newFastaFile, newNamesFile, ""); + m->mothurOut("It took " + toString(time(NULL) - start) + " secs to cluster " + toString(numSeqs) + " sequences."); m->mothurOutEndLine(); } - - if (m->control_pressed) { m->mothurRemove(newFastaFile); m->mothurRemove(newNamesFile); return 0; } + + if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } return 0; } m->mothurOutEndLine(); m->mothurOut("Output File Names: "); m->mothurOutEndLine(); - m->mothurOut(newFastaFile); m->mothurOutEndLine(); outputNames.push_back(newFastaFile); outputTypes["fasta"].push_back(newFastaFile); - m->mothurOut(newNamesFile); m->mothurOutEndLine(); outputNames.push_back(newNamesFile); outputTypes["name"].push_back(newNamesFile); + for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); } m->mothurOutEndLine(); //set fasta file as new current fastafile @@ -295,6 +341,11 @@ int PreClusterCommand::execute(){ if (itTypes != outputTypes.end()) { if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setNameFile(current); } } + + itTypes = outputTypes.find("count"); + if (itTypes != outputTypes.end()) { + if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setCountTableFile(current); } + } return 0; @@ -305,51 +356,294 @@ int PreClusterCommand::execute(){ } } /**************************************************************************************************/ -int PreClusterCommand::process(){ +int PreClusterCommand::createProcessesGroups(string newFName, string newNName, string newMFile, vector groups) { try { - //sort seqs by number of identical seqs - sort(alignSeqs.begin(), alignSeqs.end(), comparePriority); + vector processIDS; + int process = 1; + int num = 0; - int count = 0; - int numSeqs = alignSeqs.size(); + //sanity check + if (groups.size() < processors) { processors = groups.size(); } + + //divide the groups between the processors + vector lines; + int numGroupsPerProcessor = groups.size() / processors; + for (int i = 0; i < processors; i++) { + int startIndex = i * numGroupsPerProcessor; + int endIndex = (i+1) * numGroupsPerProcessor; + if(i == (processors - 1)){ endIndex = groups.size(); } + lines.push_back(linePair(startIndex, endIndex)); + } + +#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix) - //think about running through twice... - for (int i = 0; i < numSeqs; i++) { + //loop through and create all the processes you want + while (process != processors) { + int pid = fork(); - //are you active - // itActive = active.find(alignSeqs[i].seq.getName()); + 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){ + outputNames.clear(); + num = driverGroups(newFName + toString(getpid()) + ".temp", newNName + toString(getpid()) + ".temp", newMFile, lines[process].start, lines[process].end, groups); + + string tempFile = toString(getpid()) + ".outputNames.temp"; + ofstream outTemp; + m->openOutputFile(tempFile, outTemp); + + outTemp << outputNames.size(); + for (int i = 0; i < outputNames.size(); i++) { outTemp << outputNames[i] << endl; } + outTemp.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); + } + } + + //do my part + num = driverGroups(newFName, newNName, newMFile, lines[0].start, lines[0].end, groups); + + //force parent to wait until all the processes are done + for (int i=0;iopenInputFile(tempFile, intemp); + + int num; + intemp >> num; + for (int k = 0; k < num; k++) { + string name = ""; + intemp >> name; m->gobble(intemp); + + outputNames.push_back(name); outputTypes["map"].push_back(name); + } + intemp.close(); + m->mothurRemove(tempFile); + } +#else + + ////////////////////////////////////////////////////////////////////////////////////////////////////// + //Windows version shared memory, so be careful when passing variables through the preClusterData struct. + //Above fork() will clone, so memory is separate, but that's not the case with windows, + ////////////////////////////////////////////////////////////////////////////////////////////////////// + + vector pDataArray; + DWORD dwThreadIdArray[processors-1]; + HANDLE hThreadArray[processors-1]; + + //Create processor worker threads. + for( int i=1; icontrol_pressed) { return 0; } - - if (alignSeqs[j].active) { //this sequence has not been merged yet - //are you within "diff" bases - int mismatch = calcMisMatches(alignSeqs[i].seq.getAligned(), alignSeqs[j].seq.getAligned()); - - if (mismatch <= diffs) { - //merge - alignSeqs[i].names += ',' + alignSeqs[j].names; - alignSeqs[i].numIdentical += alignSeqs[j].numIdentical; - - alignSeqs[j].active = 0; - alignSeqs[j].numIdentical = 0; - count++; - } - }//end if j active - }//end if i != j - - //remove from active list - alignSeqs[i].active = 0; + preClusterData* tempPreCluster = new preClusterData(fastafile, namefile, groupfile, countfile, (newFName+extension), (newNName+extension), newMFile, groups, m, lines[i].start, lines[i].end, diffs, topdown, i); + pDataArray.push_back(tempPreCluster); + processIDS.push_back(i); + + //MySeqSumThreadFunction is in header. It must be global or static to work with the threads. + //default security attributes, thread function name, argument to thread function, use default creation flags, returns the thread identifier + hThreadArray[i-1] = CreateThread(NULL, 0, MyPreclusterThreadFunction, pDataArray[i-1], 0, &dwThreadIdArray[i-1]); + } + - }//end if active i - if(i % 100 == 0) { m->mothurOut(toString(i) + "\t" + toString(numSeqs - count) + "\t" + toString(count)); m->mothurOutEndLine(); } + //using the main process as a worker saves time and memory + num = driverGroups(newFName, newNName, newMFile, lines[0].start, lines[0].end, groups); + + //Wait until all threads have terminated. + WaitForMultipleObjects(processors-1, hThreadArray, TRUE, INFINITE); + + //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]->mapFileNames.size(); j++) { + outputNames.push_back(pDataArray[i]->mapFileNames[j]); outputTypes["map"].push_back(pDataArray[i]->mapFileNames[j]); + } + CloseHandle(hThreadArray[i]); + delete pDataArray[i]; } +#endif + + //append output files + for(int i=0;igetFullPathName(".\\" + newFName); + //newNName = m->getFullPathName(".\\" + newNName); + + m->appendFiles((newFName + toString(processIDS[i]) + ".temp"), newFName); + m->mothurRemove((newFName + toString(processIDS[i]) + ".temp")); + + m->appendFiles((newNName + toString(processIDS[i]) + ".temp"), newNName); + m->mothurRemove((newNName + toString(processIDS[i]) + ".temp")); + } + + return num; + + } + catch(exception& e) { + m->errorOut(e, "PreClusterCommand", "createProcessesGroups"); + exit(1); + } +} +/**************************************************************************************************/ +int PreClusterCommand::driverGroups(string newFFile, string newNFile, string newMFile, int start, int end, vector groups){ + try { + + int numSeqs = 0; + + //precluster each group + for (int i = start; i < end; i++) { + + start = time(NULL); + + if (m->control_pressed) { return 0; } + + m->mothurOutEndLine(); m->mothurOut("Processing group " + groups[i] + ":"); m->mothurOutEndLine(); + + map thisNameMap; + vector thisSeqs; + if (groupfile != "") { + thisSeqs = parser->getSeqs(groups[i]); + }else if (countfile != "") { + thisSeqs = cparser->getSeqs(groups[i]); + } + if (namefile != "") { thisNameMap = parser->getNameMap(groups[i]); } + + //fill alignSeqs with this groups info. + numSeqs = loadSeqs(thisNameMap, thisSeqs, groups[i]); + + if (m->control_pressed) { return 0; } + + if (diffs > length) { m->mothurOut("Error: diffs is greater than your sequence length."); m->mothurOutEndLine(); m->control_pressed = true; return 0; } + + int count= process(newMFile+groups[i]+".map"); + outputNames.push_back(newMFile+groups[i]+".map"); outputTypes["map"].push_back(newMFile+groups[i]+".map"); + + if (m->control_pressed) { return 0; } + + m->mothurOut("Total number of sequences before pre.cluster was " + toString(alignSeqs.size()) + "."); m->mothurOutEndLine(); + m->mothurOut("pre.cluster removed " + toString(count) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine(); + printData(newFFile, newNFile, groups[i]); + + m->mothurOut("It took " + toString(time(NULL) - start) + " secs to cluster " + toString(numSeqs) + " sequences."); m->mothurOutEndLine(); + + } + + return numSeqs; + } + catch(exception& e) { + m->errorOut(e, "PreClusterCommand", "driverGroups"); + exit(1); + } +} +/**************************************************************************************************/ +int PreClusterCommand::process(string newMapFile){ + try { + ofstream out; + m->openOutputFile(newMapFile, out); + + //sort seqs by number of identical seqs + if (topdown) { sort(alignSeqs.begin(), alignSeqs.end(), comparePriorityTopDown); } + else { sort(alignSeqs.begin(), alignSeqs.end(), comparePriorityDownTop); } + + int count = 0; + int numSeqs = alignSeqs.size(); + + if (topdown) { + //think about running through twice... + for (int i = 0; i < numSeqs; i++) { + + if (alignSeqs[i].active) { //this sequence has not been merged yet + + string chunk = alignSeqs[i].seq.getName() + "\t" + toString(alignSeqs[i].numIdentical) + "\t" + toString(0) + "\t" + alignSeqs[i].seq.getAligned() + "\n"; + + //try to merge it with all smaller seqs + for (int j = i+1; j < numSeqs; j++) { + + if (m->control_pressed) { out.close(); return 0; } + + if (alignSeqs[j].active) { //this sequence has not been merged yet + //are you within "diff" bases + int mismatch = calcMisMatches(alignSeqs[i].seq.getAligned(), alignSeqs[j].seq.getAligned()); + + if (mismatch <= diffs) { + //merge + alignSeqs[i].names += ',' + alignSeqs[j].names; + alignSeqs[i].numIdentical += alignSeqs[j].numIdentical; + + chunk += alignSeqs[j].seq.getName() + "\t" + toString(alignSeqs[j].numIdentical) + "\t" + toString(mismatch) + "\t" + alignSeqs[j].seq.getAligned() + "\n"; + + alignSeqs[j].active = 0; + alignSeqs[j].numIdentical = 0; + count++; + } + }//end if j active + }//end for loop j + + //remove from active list + alignSeqs[i].active = 0; + + out << "ideal_seq_" << (i+1) << '\t' << alignSeqs[i].numIdentical << endl << chunk << endl;; + + }//end if active i + if(i % 100 == 0) { m->mothurOut(toString(i) + "\t" + toString(numSeqs - count) + "\t" + toString(count)); m->mothurOutEndLine(); } + } + }else { + map mapFile; + map originalCount; + map::iterator itCount; + for (int i = 0; i < numSeqs; i++) { mapFile[i] = ""; originalCount[i] = alignSeqs[i].numIdentical; } + + //think about running through twice... + for (int i = 0; i < numSeqs; i++) { + + //try to merge it into larger seqs + for (int j = i+1; j < numSeqs; j++) { + + if (m->control_pressed) { out.close(); return 0; } + + if (originalCount[j] > originalCount[i]) { //this sequence is more abundant than I am + //are you within "diff" bases + int mismatch = calcMisMatches(alignSeqs[i].seq.getAligned(), alignSeqs[j].seq.getAligned()); + + if (mismatch <= diffs) { + //merge + alignSeqs[j].names += ',' + alignSeqs[i].names; + alignSeqs[j].numIdentical += alignSeqs[i].numIdentical; + + mapFile[j] = alignSeqs[i].seq.getName() + "\t" + toString(alignSeqs[i].numIdentical) + "\t" + toString(mismatch) + "\t" + alignSeqs[i].seq.getAligned() + "\n" + mapFile[i]; + alignSeqs[i].numIdentical = 0; + originalCount.erase(i); + mapFile[i] = ""; + count++; + j+=numSeqs; //exit search, we merged this one in. + } + }//end abundance check + }//end for loop j + + if(i % 100 == 0) { m->mothurOut(toString(i) + "\t" + toString(numSeqs - count) + "\t" + toString(count)); m->mothurOutEndLine(); } + } + + for (int i = 0; i < numSeqs; i++) { + if (alignSeqs[i].numIdentical != 0) { + out << "ideal_seq_" << (i+1) << '\t' << alignSeqs[i].numIdentical << endl << alignSeqs[i].seq.getName() + "\t" + toString(alignSeqs[i].numIdentical) + "\t" + toString(0) + "\t" + alignSeqs[i].seq.getAligned() + "\n" << mapFile[i] << endl; + } + } + + } + out.close(); + if(numSeqs % 100 != 0) { m->mothurOut(toString(numSeqs) + "\t" + toString(numSeqs - count) + "\t" + toString(count)); m->mothurOutEndLine(); } return count; @@ -366,26 +660,13 @@ int PreClusterCommand::readFASTA(){ //ifstream inNames; ifstream inFasta; - //m->openInputFile(namefile, inNames); m->openInputFile(fastafile, inFasta); - - //string firstCol, secondCol, nameString; - length = 0; + set lengths; while (!inFasta.eof()) { if (m->control_pressed) { inFasta.close(); return 0; } - - //inNames >> firstCol >> secondCol; - //nameString = secondCol; - - //m->gobble(inNames); - //int size = 1; - //while (secondCol.find_first_of(',') != -1) { - // size++; - // secondCol = secondCol.substr(secondCol.find_first_of(',')+1, secondCol.length()); - //} - + Sequence seq(inFasta); m->gobble(inFasta); if (seq.getName() != "") { //can get "" if commented line is at end of fasta file @@ -396,17 +677,22 @@ int PreClusterCommand::readFASTA(){ else{ seqPNode tempNode(itSize->second, seq, names[seq.getName()]); alignSeqs.push_back(tempNode); - if (seq.getAligned().length() > length) { length = seq.getAligned().length(); } + lengths.insert(seq.getAligned().length()); } }else { //no names file, you are identical to yourself - seqPNode tempNode(1, seq, seq.getName()); + int numRep = 1; + if (countfile != "") { numRep = ct.getNumSeqs(seq.getName()); } + seqPNode tempNode(numRep, seq, seq.getName()); alignSeqs.push_back(tempNode); - if (seq.getAligned().length() > length) { length = seq.getAligned().length(); } + lengths.insert(seq.getAligned().length()); } } } inFasta.close(); - //inNames.close(); + + if (lengths.size() > 1) { m->control_pressed = true; m->mothurOut("[ERROR]: your sequences are not all the same length. pre.cluster requires sequences to be aligned."); m->mothurOutEndLine(); } + else if (lengths.size() == 1) { length = *(lengths.begin()); } + return alignSeqs.size(); } @@ -416,13 +702,15 @@ int PreClusterCommand::readFASTA(){ } } /**************************************************************************************************/ -int PreClusterCommand::loadSeqs(map& thisName, vector& thisSeqs){ +int PreClusterCommand::loadSeqs(map& thisName, vector& thisSeqs, string group){ try { - length = 0; + set lengths; alignSeqs.clear(); map::iterator it; bool error = false; - + map thisCount; + if (countfile != "") { thisCount = cparser->getCountTable(group); } + for (int i = 0; i < thisSeqs.size(); i++) { if (m->control_pressed) { return 0; } @@ -434,22 +722,33 @@ int PreClusterCommand::loadSeqs(map& thisName, vector& if (it == thisName.end()) { m->mothurOut(thisSeqs[i].getName() + " is not in your names file, please correct."); m->mothurOutEndLine(); error = true; } else{ //get number of reps - int numReps = 0; + int numReps = 1; for(int j=0;j<(it->second).length();j++){ if((it->second)[j] == ','){ numReps++; } } seqPNode tempNode(numReps, thisSeqs[i], it->second); alignSeqs.push_back(tempNode); - if (thisSeqs[i].getAligned().length() > length) { length = thisSeqs[i].getAligned().length(); } + lengths.insert(thisSeqs[i].getAligned().length()); } }else { //no names file, you are identical to yourself - seqPNode tempNode(1, thisSeqs[i], thisSeqs[i].getName()); + int numRep = 1; + if (countfile != "") { + map::iterator it2 = thisCount.find(thisSeqs[i].getName()); + + //should never be true since parser checks for this + if (it2 == thisCount.end()) { m->mothurOut(thisSeqs[i].getName() + " is not in your count file, please correct."); m->mothurOutEndLine(); error = true; } + else { numRep = it2->second; } + } + seqPNode tempNode(numRep, thisSeqs[i], thisSeqs[i].getName()); alignSeqs.push_back(tempNode); - if (thisSeqs[i].getAligned().length() > length) { length = thisSeqs[i].getAligned().length(); } + lengths.insert(thisSeqs[i].getAligned().length()); } } - + + if (lengths.size() > 1) { error = true; m->mothurOut("[ERROR]: your sequences are not all the same length. pre.cluster requires sequences to be aligned."); m->mothurOutEndLine(); } + else if (lengths.size() == 1) { length = *(lengths.begin()); } + //sanity check if (error) { m->control_pressed = true; } @@ -483,10 +782,84 @@ int PreClusterCommand::calcMisMatches(string seq1, string seq2){ exit(1); } } +/**************************************************************************************************/ + +int PreClusterCommand::mergeGroupCounts(string newcount, string newname, string newfasta){ + try { + ifstream inNames; + m->openInputFile(newname, inNames); + + string group, first, second; + set uniqueNames; + while (!inNames.eof()) { + if (m->control_pressed) { break; } + inNames >> group; m->gobble(inNames); + inNames >> first; m->gobble(inNames); + inNames >> second; m->gobble(inNames); + + vector names; + m->splitAtComma(second, names); + + uniqueNames.insert(first); + + int total = ct.getGroupCount(first, group); + for (int i = 1; i < names.size(); i++) { + total += ct.getGroupCount(names[i], group); + ct.setAbund(names[i], group, 0); + } + ct.setAbund(first, group, total); + } + inNames.close(); + + vector namesOfSeqs = ct.getNamesOfSeqs(); + for (int i = 0; i < namesOfSeqs.size(); i++) { + if (ct.getNumSeqs(namesOfSeqs[i]) == 0) { + ct.remove(namesOfSeqs[i]); + } + } + + ct.printTable(newcount); + m->mothurRemove(newname); + + if (bygroup) { //if by group, must remove the duplicate seqs that are named the same + ifstream in; + m->openInputFile(newfasta, in); + + ofstream out; + m->openOutputFile(newfasta+"temp", out); + + int count = 0; + set already; + while(!in.eof()) { + if (m->control_pressed) { break; } + + Sequence seq(in); m->gobble(in); + + if (seq.getName() != "") { + count++; + if (already.count(seq.getName()) == 0) { + seq.printSequence(out); + already.insert(seq.getName()); + } + } + } + in.close(); + out.close(); + m->mothurRemove(newfasta); + m->renameFile(newfasta+"temp", newfasta); + } + return 0; + + } + catch(exception& e) { + m->errorOut(e, "PreClusterCommand", "mergeGroupCounts"); + exit(1); + } +} /**************************************************************************************************/ -void PreClusterCommand::printData(string newfasta, string newname){ +void PreClusterCommand::printData(string newfasta, string newname, string group){ try { ofstream outFasta; ofstream outNames; @@ -498,11 +871,15 @@ void PreClusterCommand::printData(string newfasta, string newname){ m->openOutputFile(newfasta, outFasta); m->openOutputFile(newname, outNames); } - + + if ((countfile != "") && (group == "")) { outNames << "Representative_Sequence\ttotal\n"; } for (int i = 0; i < alignSeqs.size(); i++) { if (alignSeqs[i].numIdentical != 0) { alignSeqs[i].seq.printSequence(outFasta); - outNames << alignSeqs[i].seq.getName() << '\t' << alignSeqs[i].names << endl; + if (countfile != "") { + if (group != "") { outNames << group << '\t' << alignSeqs[i].seq.getName() << '\t' << alignSeqs[i].names << endl; } + else { outNames << alignSeqs[i].seq.getName() << '\t' << alignSeqs[i].numIdentical << endl; } + }else { outNames << alignSeqs[i].seq.getName() << '\t' << alignSeqs[i].names << endl; } } } @@ -525,13 +902,19 @@ void PreClusterCommand::readNameFile(){ while (!in.eof()) { in >> firstCol >> secondCol; m->gobble(in); + + for (int i = 0; i < firstCol.length(); i++) { + if (firstCol[i] == ':') { firstCol[i] = '_'; m->changedSeqNames = true; } + } + + int size = 1; + for (int i = 0; i < secondCol.length(); i++) { + if (secondCol[i] == ':') { secondCol[i] = '_'; m->changedSeqNames = true; } + else if(secondCol[i] == ','){ size++; } + } + names[firstCol] = secondCol; - int size = 1; - - for(int i=0;i