2 * classifyseqscommand.cpp
5 * Created by westcott on 11/2/09.
6 * Copyright 2009 Schloss Lab. All rights reserved.
10 #include "classifyseqscommand.h"
11 #include "sequence.hpp"
13 #include "phylotree.h"
14 #include "phylosummary.h"
17 //**********************************************************************************************************************
19 ClassifySeqsCommand::ClassifySeqsCommand(string option) {
23 //allow user to run help
24 if(option == "help") { help(); abort = true; }
28 //valid paramters for this command
29 string AlignArray[] = {"template","fasta","name","group","search","ksize","method","processors","taxonomy","match","mismatch","gapopen","gapextend","numwanted","cutoff","probs","iters", "outputdir","inputdir"};
30 vector<string> myArray (AlignArray, AlignArray+(sizeof(AlignArray)/sizeof(string)));
32 OptionParser parser(option);
33 map<string, string> parameters = parser.getParameters();
35 ValidParameters validParameter;
36 map<string, string>::iterator it;
38 //check to make sure all parameters are valid for command
39 for (it = parameters.begin(); it != parameters.end(); it++) {
40 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
43 //if the user changes the output directory command factory will send this info to us in the output parameter
44 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){ outputDir = ""; }
46 //if the user changes the input directory command factory will send this info to us in the output parameter
47 string inputDir = validParameter.validFile(parameters, "inputdir", false);
48 if (inputDir == "not found"){ inputDir = ""; }
51 it = parameters.find("template");
52 //user has given a template file
53 if(it != parameters.end()){
54 path = hasPath(it->second);
55 //if the user has not given a path then, add inputdir. else leave path alone.
56 if (path == "") { parameters["template"] = inputDir + it->second; }
59 it = parameters.find("taxonomy");
60 //user has given a template file
61 if(it != parameters.end()){
62 path = hasPath(it->second);
63 //if the user has not given a path then, add inputdir. else leave path alone.
64 if (path == "") { parameters["taxonomy"] = inputDir + it->second; }
67 it = parameters.find("group");
68 //user has given a template file
69 if(it != parameters.end()){
70 path = hasPath(it->second);
71 //if the user has not given a path then, add inputdir. else leave path alone.
72 if (path == "") { parameters["group"] = inputDir + it->second; }
76 //check for required parameters
77 templateFileName = validParameter.validFile(parameters, "template", true);
78 if (templateFileName == "not found") {
79 m->mothurOut("template is a required parameter for the classify.seqs command.");
80 m->mothurOutEndLine();
83 else if (templateFileName == "not open") { abort = true; }
86 fastaFileName = validParameter.validFile(parameters, "fasta", false);
87 if (fastaFileName == "not found") { m->mothurOut("fasta is a required parameter for the classify.seqs command."); m->mothurOutEndLine(); abort = true; }
89 splitAtDash(fastaFileName, fastaFileNames);
91 //go through files and make sure they are good, if not, then disregard them
92 for (int i = 0; i < fastaFileNames.size(); i++) {
94 string path = hasPath(fastaFileNames[i]);
95 //if the user has not given a path then, add inputdir. else leave path alone.
96 if (path == "") { fastaFileNames[i] = inputDir + fastaFileNames[i]; }
103 MPI_Comm_size(MPI_COMM_WORLD, &processors); //set processors to the number of mpi processes running
104 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
110 ableToOpen = openInputFile(fastaFileNames[i], in);
114 for (int j = 1; j < processors; j++) {
115 MPI_Send(&ableToOpen, 1, MPI_INT, j, 2001, MPI_COMM_WORLD);
119 MPI_Recv(&ableToOpen, 1, MPI_INT, 0, 2001, MPI_COMM_WORLD, &status);
124 if (ableToOpen == 1) {
125 m->mothurOut(fastaFileNames[i] + " will be disregarded."); m->mothurOutEndLine();
126 //erase from file list
127 fastaFileNames.erase(fastaFileNames.begin()+i);
133 //make sure there is at least one valid file left
134 if (fastaFileNames.size() == 0) { m->mothurOut("no valid files."); m->mothurOutEndLine(); abort = true; }
138 taxonomyFileName = validParameter.validFile(parameters, "taxonomy", true);
139 if (taxonomyFileName == "not found") {
140 m->mothurOut("taxonomy is a required parameter for the classify.seqs command.");
141 m->mothurOutEndLine();
144 else if (taxonomyFileName == "not open") { abort = true; }
147 namefile = validParameter.validFile(parameters, "name", false);
148 if (namefile == "not found") { namefile = ""; }
151 splitAtDash(namefile, namefileNames);
153 //go through files and make sure they are good, if not, then disregard them
154 for (int i = 0; i < namefileNames.size(); i++) {
155 if (inputDir != "") {
156 string path = hasPath(namefileNames[i]);
157 //if the user has not given a path then, add inputdir. else leave path alone.
158 if (path == "") { namefileNames[i] = inputDir + namefileNames[i]; }
164 MPI_Comm_size(MPI_COMM_WORLD, &processors); //set processors to the number of mpi processes running
165 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
171 ableToOpen = openInputFile(namefileNames[i], in);
175 for (int j = 1; j < processors; j++) {
176 MPI_Send(&ableToOpen, 1, MPI_INT, j, 2001, MPI_COMM_WORLD);
180 MPI_Recv(&ableToOpen, 1, MPI_INT, 0, 2001, MPI_COMM_WORLD, &status);
184 if (ableToOpen == 1) { m->mothurOut("Unable to match name file with fasta file."); m->mothurOutEndLine(); abort = true; }
189 if (namefile != "") {
190 if (namefileNames.size() != fastaFileNames.size()) { abort = true; m->mothurOut("If you provide a name file, you must have one for each fasta file."); m->mothurOutEndLine(); }
193 groupfile = validParameter.validFile(parameters, "group", false);
194 if (groupfile == "not found") { groupfile = ""; }
196 splitAtDash(groupfile, groupfileNames);
198 //go through files and make sure they are good, if not, then disregard them
199 for (int i = 0; i < groupfileNames.size(); i++) {
200 if (inputDir != "") {
201 string path = hasPath(groupfileNames[i]);
202 //if the user has not given a path then, add inputdir. else leave path alone.
203 if (path == "") { groupfileNames[i] = inputDir + groupfileNames[i]; }
209 MPI_Comm_size(MPI_COMM_WORLD, &processors); //set processors to the number of mpi processes running
210 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
216 ableToOpen = openInputFile(groupfileNames[i], in);
220 for (int j = 1; j < processors; j++) {
221 MPI_Send(&ableToOpen, 1, MPI_INT, j, 2001, MPI_COMM_WORLD);
225 MPI_Recv(&ableToOpen, 1, MPI_INT, 0, 2001, MPI_COMM_WORLD, &status);
229 if (ableToOpen == 1) { m->mothurOut("Unable to match group file with fasta file, not using " + groupfileNames[i] + "."); m->mothurOutEndLine(); groupfileNames[i] = ""; }
234 if (groupfile != "") {
235 if (groupfileNames.size() != fastaFileNames.size()) { abort = true; m->mothurOut("If you provide a group file, you must have one for each fasta file."); m->mothurOutEndLine(); }
237 for (int i = 0; i < fastaFileNames.size(); i++) { groupfileNames.push_back(""); }
240 //check for optional parameter and set defaults
241 // ...at some point should added some additional type checking...
243 temp = validParameter.validFile(parameters, "ksize", false); if (temp == "not found"){ temp = "8"; }
244 convert(temp, kmerSize);
246 temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = "1"; }
247 convert(temp, processors);
249 search = validParameter.validFile(parameters, "search", false); if (search == "not found"){ search = "kmer"; }
251 method = validParameter.validFile(parameters, "method", false); if (method == "not found"){ method = "bayesian"; }
253 temp = validParameter.validFile(parameters, "match", false); if (temp == "not found"){ temp = "1.0"; }
254 convert(temp, match);
256 temp = validParameter.validFile(parameters, "mismatch", false); if (temp == "not found"){ temp = "-1.0"; }
257 convert(temp, misMatch);
259 temp = validParameter.validFile(parameters, "gapopen", false); if (temp == "not found"){ temp = "-2.0"; }
260 convert(temp, gapOpen);
262 temp = validParameter.validFile(parameters, "gapextend", false); if (temp == "not found"){ temp = "-1.0"; }
263 convert(temp, gapExtend);
265 temp = validParameter.validFile(parameters, "numwanted", false); if (temp == "not found"){ temp = "10"; }
266 convert(temp, numWanted);
268 temp = validParameter.validFile(parameters, "cutoff", false); if (temp == "not found"){ temp = "0"; }
269 convert(temp, cutoff);
271 temp = validParameter.validFile(parameters, "probs", false); if (temp == "not found"){ temp = "true"; }
272 probs = isTrue(temp);
274 temp = validParameter.validFile(parameters, "iters", false); if (temp == "not found") { temp = "100"; }
275 convert(temp, iters);
279 if ((method == "bayesian") && (search != "kmer")) {
280 m->mothurOut("The bayesian method requires the kmer search." + search + "will be disregarded." ); m->mothurOutEndLine();
286 catch(exception& e) {
287 m->errorOut(e, "ClassifySeqsCommand", "ClassifySeqsCommand");
292 //**********************************************************************************************************************
294 ClassifySeqsCommand::~ClassifySeqsCommand(){
296 if (abort == false) {
297 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
301 //**********************************************************************************************************************
303 void ClassifySeqsCommand::help(){
305 m->mothurOut("The classify.seqs command reads a fasta file containing sequences and creates a .taxonomy file and a .tax.summary file.\n");
306 m->mothurOut("The classify.seqs command parameters are template, fasta, name, search, ksize, method, taxonomy, processors, match, mismatch, gapopen, gapextend, numwanted and probs.\n");
307 m->mothurOut("The template, fasta and taxonomy parameters are required. You may enter multiple fasta files by separating their names with dashes. ie. fasta=abrecovery.fasta-amzon.fasta \n");
308 m->mothurOut("The search parameter allows you to specify the method to find most similar template. Your options are: suffix, kmer, blast and distance. The default is kmer.\n");
309 m->mothurOut("The name parameter allows you add a names file with your fasta file, if you enter multiple fasta files, you must enter matching names files for them.\n");
310 m->mothurOut("The group parameter allows you add a group file so you can have the summary totals broken up by group.\n");
311 m->mothurOut("The method parameter allows you to specify classification method to use. Your options are: bayesian and knn. The default is bayesian.\n");
312 m->mothurOut("The ksize parameter allows you to specify the kmer size for finding most similar template to candidate. The default is 8.\n");
313 m->mothurOut("The processors parameter allows you to specify the number of processors to use. The default is 1.\n");
315 m->mothurOut("When using MPI, the processors parameter is set to the number of MPI processes running. \n");
317 m->mothurOut("The match parameter allows you to specify the bonus for having the same base. The default is 1.0.\n");
318 m->mothurOut("The mistmatch parameter allows you to specify the penalty for having different bases. The default is -1.0.\n");
319 m->mothurOut("The gapopen parameter allows you to specify the penalty for opening a gap in an alignment. The default is -2.0.\n");
320 m->mothurOut("The gapextend parameter allows you to specify the penalty for extending a gap in an alignment. The default is -1.0.\n");
321 m->mothurOut("The numwanted parameter allows you to specify the number of sequence matches you want with the knn method. The default is 10.\n");
322 m->mothurOut("The cutoff parameter allows you to specify a bootstrap confidence threshold for your taxonomy. The default is 0.\n");
323 m->mothurOut("The probs parameter shut off the bootstrapping results for the bayesian method. The default is true, meaning you want the bootstrapping to be run.\n");
324 m->mothurOut("The iters parameter allows you to specify how many iterations to do when calculating the bootstrap confidence score for your taxonomy with the bayesian method. The default is 100.\n");
325 m->mothurOut("The classify.seqs command should be in the following format: \n");
326 m->mothurOut("classify.seqs(template=yourTemplateFile, fasta=yourFastaFile, method=yourClassificationMethod, search=yourSearchmethod, ksize=yourKmerSize, taxonomy=yourTaxonomyFile, processors=yourProcessors) \n");
327 m->mothurOut("Example classify.seqs(fasta=amazon.fasta, template=core.filtered, method=knn, search=gotoh, ksize=8, processors=2)\n");
328 m->mothurOut("The .taxonomy file consists of 2 columns: 1 = your sequence name, 2 = the taxonomy for your sequence. \n");
329 m->mothurOut("The .tax.summary is a summary of the different taxonomies represented in your fasta file. \n");
330 m->mothurOut("Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFastaFile).\n\n");
332 catch(exception& e) {
333 m->errorOut(e, "ClassifySeqsCommand", "help");
339 //**********************************************************************************************************************
341 int ClassifySeqsCommand::execute(){
343 if (abort == true) { return 0; }
345 if(method == "bayesian"){ classify = new Bayesian(taxonomyFileName, templateFileName, search, kmerSize, cutoff, iters); }
346 else if(method == "knn"){ classify = new Knn(taxonomyFileName, templateFileName, search, kmerSize, gapOpen, gapExtend, match, misMatch, numWanted); }
348 m->mothurOut(search + " is not a valid method option. I will run the command using bayesian.");
349 m->mothurOutEndLine();
350 classify = new Bayesian(taxonomyFileName, templateFileName, search, kmerSize, cutoff, iters);
353 if (m->control_pressed) { delete classify; return 0; }
355 vector<string> outputNames;
357 for (int s = 0; s < fastaFileNames.size(); s++) {
359 m->mothurOut("Classifying sequences from " + fastaFileNames[s] + " ..." ); m->mothurOutEndLine();
361 if (outputDir == "") { outputDir += hasPath(fastaFileNames[s]); }
362 string newTaxonomyFile = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + getRootName(getSimpleName(taxonomyFileName)) + "taxonomy";
363 string tempTaxonomyFile = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + "taxonomy.temp";
364 string taxSummary = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + getRootName(getSimpleName(taxonomyFileName)) + "tax.summary";
366 outputNames.push_back(newTaxonomyFile);
367 outputNames.push_back(taxSummary);
369 int start = time(NULL);
370 int numFastaSeqs = 0;
371 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
374 int pid, end, numSeqsPerProcessor;
379 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
380 MPI_Comm_size(MPI_COMM_WORLD, &processors);
383 MPI_File outMPINewTax;
384 MPI_File outMPITempTax;
386 int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
387 int inMode=MPI_MODE_RDONLY;
389 //char* outNewTax = new char[newTaxonomyFile.length()];
390 //memcpy(outNewTax, newTaxonomyFile.c_str(), newTaxonomyFile.length());
392 char outNewTax[1024];
393 strcpy(outNewTax, newTaxonomyFile.c_str());
395 //char* outTempTax = new char[tempTaxonomyFile.length()];
396 //memcpy(outTempTax, tempTaxonomyFile.c_str(), tempTaxonomyFile.length());
398 char outTempTax[1024];
399 strcpy(outTempTax, tempTaxonomyFile.c_str());
401 //char* inFileName = new char[fastaFileNames[s].length()];
402 //memcpy(inFileName, fastaFileNames[s].c_str(), fastaFileNames[s].length());
404 char inFileName[1024];
405 strcpy(inFileName, fastaFileNames[s].c_str());
407 MPI_File_open(MPI_COMM_WORLD, inFileName, inMode, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer
408 MPI_File_open(MPI_COMM_WORLD, outNewTax, outMode, MPI_INFO_NULL, &outMPINewTax);
409 MPI_File_open(MPI_COMM_WORLD, outTempTax, outMode, MPI_INFO_NULL, &outMPITempTax);
415 if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); delete classify; return 0; }
417 if (pid == 0) { //you are the root process
419 MPIPos = setFilePosFasta(fastaFileNames[s], numFastaSeqs); //fills MPIPos, returns numSeqs
421 //send file positions to all processes
422 MPI_Bcast(&numFastaSeqs, 1, MPI_INT, 0, MPI_COMM_WORLD); //send numSeqs
423 MPI_Bcast(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, 0, MPI_COMM_WORLD); //send file pos
425 //figure out how many sequences you have to align
426 numSeqsPerProcessor = numFastaSeqs / processors;
427 int startIndex = pid * numSeqsPerProcessor;
428 if(pid == (processors - 1)){ numSeqsPerProcessor = numFastaSeqs - pid * numSeqsPerProcessor; }
432 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPINewTax, outMPITempTax, MPIPos);
434 if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete classify; return 0; }
436 for (int i = 1; i < processors; i++) {
438 MPI_Recv(&done, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status);
440 }else{ //you are a child process
441 MPI_Bcast(&numFastaSeqs, 1, MPI_INT, 0, MPI_COMM_WORLD); //get numSeqs
442 MPIPos.resize(numFastaSeqs+1);
443 MPI_Bcast(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, 0, MPI_COMM_WORLD); //get file positions
445 //figure out how many sequences you have to align
446 numSeqsPerProcessor = numFastaSeqs / processors;
447 int startIndex = pid * numSeqsPerProcessor;
448 if(pid == (processors - 1)){ numSeqsPerProcessor = numFastaSeqs - pid * numSeqsPerProcessor; }
452 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPINewTax, outMPITempTax, MPIPos);
454 if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); delete classify; return 0; }
457 MPI_Send(&done, 1, MPI_INT, 0, tag, MPI_COMM_WORLD);
461 MPI_File_close(&inMPI);
462 MPI_File_close(&outMPINewTax);
463 MPI_File_close(&outMPITempTax);
466 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
469 openInputFile(fastaFileNames[s], inFASTA);
470 numFastaSeqs=count(istreambuf_iterator<char>(inFASTA),istreambuf_iterator<char>(), '>');
473 lines.push_back(new linePair(0, numFastaSeqs));
475 driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]);
478 vector<int> positions;
479 processIDS.resize(0);
482 openInputFile(fastaFileNames[s], inFASTA);
485 while(!inFASTA.eof()){
486 input = getline(inFASTA);
487 if (input.length() != 0) {
488 if(input[0] == '>'){ int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); }
493 numFastaSeqs = positions.size();
495 int numSeqsPerProcessor = numFastaSeqs / processors;
497 for (int i = 0; i < processors; i++) {
498 int startPos = positions[ i * numSeqsPerProcessor ];
499 if(i == processors - 1){
500 numSeqsPerProcessor = numFastaSeqs - i * numSeqsPerProcessor;
502 lines.push_back(new linePair(startPos, numSeqsPerProcessor));
504 createProcesses(newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]);
506 rename((newTaxonomyFile + toString(processIDS[0]) + ".temp").c_str(), newTaxonomyFile.c_str());
507 rename((tempTaxonomyFile + toString(processIDS[0]) + ".temp").c_str(), tempTaxonomyFile.c_str());
509 for(int i=1;i<processors;i++){
510 appendTaxFiles((newTaxonomyFile + toString(processIDS[i]) + ".temp"), newTaxonomyFile);
511 appendTaxFiles((tempTaxonomyFile + toString(processIDS[i]) + ".temp"), tempTaxonomyFile);
512 remove((newTaxonomyFile + toString(processIDS[i]) + ".temp").c_str());
513 remove((tempTaxonomyFile + toString(processIDS[i]) + ".temp").c_str());
519 openInputFile(fastaFileNames[s], inFASTA);
520 numFastaSeqs=count(istreambuf_iterator<char>(inFASTA),istreambuf_iterator<char>(), '>');
523 lines.push_back(new linePair(0, numFastaSeqs));
525 driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]);
529 m->mothurOutEndLine();
530 m->mothurOut("It took " + toString(time(NULL) - start) + " secs to classify " + toString(numFastaSeqs) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine();
535 if (pid == 0) { //this part does not need to be paralellized
537 if(namefile != "") { m->mothurOut("Reading " + namefileNames[s] + "..."); cout.flush(); MPIReadNamesFile(namefileNames[s]); m->mothurOut(" Done."); m->mothurOutEndLine(); }
542 m->mothurOut("Reading " + namefileNames[s] + "..."); cout.flush();
544 nameMap.clear(); //remove old names
547 openInputFile(namefileNames[s], inNames);
549 string firstCol, secondCol;
550 while(!inNames.eof()) {
551 inNames >> firstCol >> secondCol; gobble(inNames);
554 splitAtComma(secondCol, temp);
556 nameMap[firstCol] = temp;
560 m->mothurOut(" Done."); m->mothurOutEndLine();
565 if (groupfile != "") { group = groupfileNames[s]; }
567 PhyloSummary taxaSum(taxonomyFileName, group);
569 if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete classify; return 0; }
571 if (namefile == "") { taxaSum.summarize(tempTaxonomyFile); }
574 openInputFile(tempTaxonomyFile, in);
576 //read in users taxonomy file and add sequences to tree
580 in >> name >> taxon; gobble(in);
582 itNames = nameMap.find(name);
584 if (itNames == nameMap.end()) {
585 m->mothurOut(name + " is not in your name file please correct."); m->mothurOutEndLine(); exit(1);
587 for (int i = 0; i < itNames->second.size(); i++) {
588 taxaSum.addSeqToTree(itNames->second[i], taxon); //add it as many times as there are identical seqs
590 itNames->second.clear();
591 nameMap.erase(itNames->first);
596 remove(tempTaxonomyFile.c_str());
598 if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete classify; return 0; }
602 openOutputFile(taxSummary, outTaxTree);
603 taxaSum.print(outTaxTree);
606 //output taxonomy with the unclassified bins added
608 openInputFile(newTaxonomyFile, inTax);
611 string unclass = newTaxonomyFile + ".unclass.temp";
612 openOutputFile(unclass, outTax);
614 //get maxLevel from phylotree so you know how many 'unclassified's to add
615 int maxLevel = taxaSum.getMaxLevel();
617 //read taxfile - this reading and rewriting is done to preserve the confidence scores.
619 while (!inTax.eof()) {
620 if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } remove(unclass.c_str()); delete classify; return 0; }
622 inTax >> name >> taxon; gobble(inTax);
624 string newTax = addUnclassifieds(taxon, maxLevel);
626 outTax << name << '\t' << newTax << endl;
631 remove(newTaxonomyFile.c_str());
632 rename(unclass.c_str(), newTaxonomyFile.c_str());
634 m->mothurOutEndLine();
635 m->mothurOut("It took " + toString(time(NULL) - start) + " secs to create the summary file for " + toString(numFastaSeqs) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine();
641 m->mothurOutEndLine();
642 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
643 for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
644 m->mothurOutEndLine();
650 catch(exception& e) {
651 m->errorOut(e, "ClassifySeqsCommand", "execute");
656 /**************************************************************************************************/
657 string ClassifySeqsCommand::addUnclassifieds(string tax, int maxlevel) {
659 string newTax, taxon;
662 //keep what you have counting the levels
663 while (tax.find_first_of(';') != -1) {
665 taxon = tax.substr(0,tax.find_first_of(';'))+';';
666 tax = tax.substr(tax.find_first_of(';')+1, tax.length());
671 //add "unclassified" until you reach maxLevel
672 while (level < maxlevel) {
673 newTax += "unclassified;";
679 catch(exception& e) {
680 m->errorOut(e, "ClassifySeqsCommand", "addUnclassifieds");
685 /**************************************************************************************************/
687 void ClassifySeqsCommand::createProcesses(string taxFileName, string tempTaxFile, string filename) {
689 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
691 // processIDS.resize(0);
693 //loop through and create all the processes you want
694 while (process != processors) {
698 processIDS.push_back(pid); //create map from line number to pid so you can append files in correct order later
701 driver(lines[process], taxFileName + toString(getpid()) + ".temp", tempTaxFile + toString(getpid()) + ".temp", filename);
703 }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
706 //force parent to wait until all the processes are done
707 for (int i=0;i<processors;i++) {
708 int temp = processIDS[i];
713 catch(exception& e) {
714 m->errorOut(e, "ClassifySeqsCommand", "createProcesses");
718 /**************************************************************************************************/
720 void ClassifySeqsCommand::appendTaxFiles(string temp, string filename) {
725 openOutputFileAppend(filename, output);
726 openInputFile(temp, input);
728 while(char c = input.get()){
729 if(input.eof()) { break; }
730 else { output << c; }
736 catch(exception& e) {
737 m->errorOut(e, "ClassifySeqsCommand", "appendTaxFiles");
742 //**********************************************************************************************************************
744 int ClassifySeqsCommand::driver(linePair* line, string taxFName, string tempTFName, string filename){
747 openOutputFile(taxFName, outTax);
749 ofstream outTaxSimple;
750 openOutputFile(tempTFName, outTaxSimple);
753 openInputFile(filename, inFASTA);
755 inFASTA.seekg(line->start);
759 for(int i=0;i<line->numSeqs;i++){
760 if (m->control_pressed) { return 0; }
762 Sequence* candidateSeq = new Sequence(inFASTA);
764 if (candidateSeq->getName() != "") {
765 taxonomy = classify->getTaxonomy(candidateSeq);
767 if (m->control_pressed) { delete candidateSeq; return 0; }
769 if (taxonomy != "bad seq") {
770 //output confidence scores or not
772 outTax << candidateSeq->getName() << '\t' << taxonomy << endl;
774 outTax << candidateSeq->getName() << '\t' << classify->getSimpleTax() << endl;
777 outTaxSimple << candidateSeq->getName() << '\t' << classify->getSimpleTax() << endl;
782 if((i+1) % 100 == 0){
783 m->mothurOut("Classifying sequence " + toString(i+1)); m->mothurOutEndLine();
789 outTaxSimple.close();
793 catch(exception& e) {
794 m->errorOut(e, "ClassifySeqsCommand", "driver");
798 //**********************************************************************************************************************
800 int ClassifySeqsCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& newFile, MPI_File& tempFile, vector<long>& MPIPos){
802 MPI_Status statusNew;
803 MPI_Status statusTemp;
807 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
812 for(int i=0;i<num;i++){
814 if (m->control_pressed) { return 0; }
817 int length = MPIPos[start+i+1] - MPIPos[start+i];
818 char* buf4 = new char[length];
819 MPI_File_read_at(inMPI, MPIPos[start+i], buf4, length, MPI_CHAR, &status);
821 string tempBuf = buf4;
822 if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); }
823 istringstream iss (tempBuf,istringstream::in);
826 Sequence* candidateSeq = new Sequence(iss);
828 if (candidateSeq->getName() != "") {
829 taxonomy = classify->getTaxonomy(candidateSeq);
831 if (taxonomy != "bad seq") {
832 //output confidence scores or not
834 outputString = candidateSeq->getName() + "\t" + taxonomy + "\n";
836 outputString = candidateSeq->getName() + "\t" + classify->getSimpleTax() + "\n";
839 int length = outputString.length();
840 char* buf2 = new char[length];
841 memcpy(buf2, outputString.c_str(), length);
843 MPI_File_write_shared(newFile, buf2, length, MPI_CHAR, &statusNew);
846 outputString = candidateSeq->getName() + "\t" + classify->getSimpleTax() + "\n";
847 length = outputString.length();
848 char* buf = new char[length];
849 memcpy(buf, outputString.c_str(), length);
851 MPI_File_write_shared(tempFile, buf, length, MPI_CHAR, &statusTemp);
857 if((i+1) % 100 == 0){ cout << "Classifying sequence " << (i+1) << endl; }
860 if(num % 100 != 0){ cout << "Classifying sequence " << (num) << endl; }
865 catch(exception& e) {
866 m->errorOut(e, "ClassifySeqsCommand", "driverMPI");
871 //**********************************************************************************************************************
872 int ClassifySeqsCommand::MPIReadNamesFile(string nameFilename){
875 nameMap.clear(); //remove old names
881 //char* inFileName = new char[nameFilename.length()];
882 //memcpy(inFileName, nameFilename.c_str(), nameFilename.length());
884 char inFileName[1024];
885 strcpy(inFileName, nameFilename.c_str());
887 MPI_File_open(MPI_COMM_WORLD, inFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI);
888 MPI_File_get_size(inMPI, &size);
891 char* buffer = new char[size];
892 MPI_File_read(inMPI, buffer, size, MPI_CHAR, &status);
894 string tempBuf = buffer;
895 if (tempBuf.length() > size) { tempBuf = tempBuf.substr(0, size); }
896 istringstream iss (tempBuf,istringstream::in);
899 string firstCol, secondCol;
901 iss >> firstCol >> secondCol; gobble(iss);
904 splitAtComma(secondCol, temp);
906 nameMap[firstCol] = temp;
909 MPI_File_close(&inMPI);
913 catch(exception& e) {
914 m->errorOut(e, "ClassifySeqsCommand", "MPIReadNamesFile");
919 /**************************************************************************************************/