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"
18 //**********************************************************************************************************************
19 vector<string> ClassifySeqsCommand::getValidParameters(){
21 string AlignArray[] = {"template","fasta","name","group","search","ksize","method","processors","taxonomy","match","mismatch","gapopen","gapextend","numwanted","cutoff","probs","iters", "outputdir","inputdir"};
22 vector<string> myArray (AlignArray, AlignArray+(sizeof(AlignArray)/sizeof(string)));
26 m->errorOut(e, "ClassifySeqsCommand", "getValidParameters");
30 //**********************************************************************************************************************
31 ClassifySeqsCommand::ClassifySeqsCommand(){
33 //initialize outputTypes
34 vector<string> tempOutNames;
35 outputTypes["taxonomy"] = tempOutNames;
36 outputTypes["taxsummary"] = tempOutNames;
37 outputTypes["matchdist"] = tempOutNames;
40 m->errorOut(e, "ClassifySeqsCommand", "ClassifySeqsCommand");
44 //**********************************************************************************************************************
45 vector<string> ClassifySeqsCommand::getRequiredParameters(){
47 string Array[] = {"fasta","template","taxonomy"};
48 vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
52 m->errorOut(e, "ClassifySeqsCommand", "getRequiredParameters");
56 //**********************************************************************************************************************
57 vector<string> ClassifySeqsCommand::getRequiredFiles(){
59 vector<string> myArray;
63 m->errorOut(e, "ClassifySeqsCommand", "getRequiredFiles");
67 //**********************************************************************************************************************
68 ClassifySeqsCommand::ClassifySeqsCommand(string option) {
72 //allow user to run help
73 if(option == "help") { help(); abort = true; }
77 //valid paramters for this command
78 string AlignArray[] = {"template","fasta","name","group","search","ksize","method","processors","taxonomy","match","mismatch","gapopen","gapextend","numwanted","cutoff","probs","iters", "outputdir","inputdir"};
79 vector<string> myArray (AlignArray, AlignArray+(sizeof(AlignArray)/sizeof(string)));
81 OptionParser parser(option);
82 map<string, string> parameters = parser.getParameters();
84 ValidParameters validParameter("classify.seqs");
85 map<string, string>::iterator it;
87 //check to make sure all parameters are valid for command
88 for (it = parameters.begin(); it != parameters.end(); it++) {
89 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
92 //initialize outputTypes
93 vector<string> tempOutNames;
94 outputTypes["taxonomy"] = tempOutNames;
95 outputTypes["taxsummary"] = tempOutNames;
96 outputTypes["matchdist"] = tempOutNames;
98 //if the user changes the output directory command factory will send this info to us in the output parameter
99 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){ outputDir = ""; }
101 //if the user changes the input directory command factory will send this info to us in the output parameter
102 string inputDir = validParameter.validFile(parameters, "inputdir", false);
103 if (inputDir == "not found"){ inputDir = ""; }
106 it = parameters.find("template");
107 //user has given a template file
108 if(it != parameters.end()){
109 path = m->hasPath(it->second);
110 //if the user has not given a path then, add inputdir. else leave path alone.
111 if (path == "") { parameters["template"] = inputDir + it->second; }
114 it = parameters.find("taxonomy");
115 //user has given a template file
116 if(it != parameters.end()){
117 path = m->hasPath(it->second);
118 //if the user has not given a path then, add inputdir. else leave path alone.
119 if (path == "") { parameters["taxonomy"] = inputDir + it->second; }
122 it = parameters.find("group");
123 //user has given a template file
124 if(it != parameters.end()){
125 path = m->hasPath(it->second);
126 //if the user has not given a path then, add inputdir. else leave path alone.
127 if (path == "") { parameters["group"] = inputDir + it->second; }
131 //check for required parameters
132 templateFileName = validParameter.validFile(parameters, "template", true);
133 if (templateFileName == "not found") {
134 m->mothurOut("template is a required parameter for the classify.seqs command.");
135 m->mothurOutEndLine();
138 else if (templateFileName == "not open") { abort = true; }
141 fastaFileName = validParameter.validFile(parameters, "fasta", false);
142 if (fastaFileName == "not found") { m->mothurOut("fasta is a required parameter for the classify.seqs command."); m->mothurOutEndLine(); abort = true; }
144 m->splitAtDash(fastaFileName, fastaFileNames);
146 //go through files and make sure they are good, if not, then disregard them
147 for (int i = 0; i < fastaFileNames.size(); i++) {
148 if (inputDir != "") {
149 string path = m->hasPath(fastaFileNames[i]);
150 //if the user has not given a path then, add inputdir. else leave path alone.
151 if (path == "") { fastaFileNames[i] = inputDir + fastaFileNames[i]; }
157 ableToOpen = m->openInputFile(fastaFileNames[i], in, "noerror");
159 //if you can't open it, try default location
160 if (ableToOpen == 1) {
161 if (m->getDefaultPath() != "") { //default path is set
162 string tryPath = m->getDefaultPath() + m->getSimpleName(fastaFileNames[i]);
163 m->mothurOut("Unable to open " + fastaFileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine();
165 ableToOpen = m->openInputFile(tryPath, in2, "noerror");
167 fastaFileNames[i] = tryPath;
171 if (ableToOpen == 1) {
172 if (m->getOutputDir() != "") { //default path is set
173 string tryPath = m->getOutputDir() + m->getSimpleName(fastaFileNames[i]);
174 m->mothurOut("Unable to open " + fastaFileNames[i] + ". Trying output directory " + tryPath); m->mothurOutEndLine();
176 ableToOpen = m->openInputFile(tryPath, in2, "noerror");
178 fastaFileNames[i] = tryPath;
184 if (ableToOpen == 1) {
185 m->mothurOut("Unable to open " + fastaFileNames[i] + ". It will be disregarded."); m->mothurOutEndLine();
186 //erase from file list
187 fastaFileNames.erase(fastaFileNames.begin()+i);
193 //make sure there is at least one valid file left
194 if (fastaFileNames.size() == 0) { m->mothurOut("no valid files."); m->mothurOutEndLine(); abort = true; }
198 taxonomyFileName = validParameter.validFile(parameters, "taxonomy", true);
199 if (taxonomyFileName == "not found") {
200 m->mothurOut("taxonomy is a required parameter for the classify.seqs command.");
201 m->mothurOutEndLine();
204 else if (taxonomyFileName == "not open") { abort = true; }
207 namefile = validParameter.validFile(parameters, "name", false);
208 if (namefile == "not found") { namefile = ""; }
211 m->splitAtDash(namefile, namefileNames);
213 //go through files and make sure they are good, if not, then disregard them
214 for (int i = 0; i < namefileNames.size(); i++) {
215 if (inputDir != "") {
216 string path = m->hasPath(namefileNames[i]);
217 //if the user has not given a path then, add inputdir. else leave path alone.
218 if (path == "") { namefileNames[i] = inputDir + namefileNames[i]; }
223 ableToOpen = m->openInputFile(namefileNames[i], in, "noerror");
225 //if you can't open it, try default location
226 if (ableToOpen == 1) {
227 if (m->getDefaultPath() != "") { //default path is set
228 string tryPath = m->getDefaultPath() + m->getSimpleName(namefileNames[i]);
229 m->mothurOut("Unable to open " + namefileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine();
231 ableToOpen = m->openInputFile(tryPath, in2, "noerror");
233 namefileNames[i] = tryPath;
237 if (ableToOpen == 1) {
238 if (m->getOutputDir() != "") { //default path is set
239 string tryPath = m->getOutputDir() + m->getSimpleName(namefileNames[i]);
240 m->mothurOut("Unable to open " + namefileNames[i] + ". Trying output directory " + tryPath); m->mothurOutEndLine();
242 ableToOpen = m->openInputFile(tryPath, in2, "noerror");
244 namefileNames[i] = tryPath;
249 if (ableToOpen == 1) {
250 m->mothurOut("Unable to open " + namefileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); abort = true;
251 //erase from file list
252 namefileNames.erase(namefileNames.begin()+i);
259 if (namefile != "") {
260 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(); }
263 groupfile = validParameter.validFile(parameters, "group", false);
264 if (groupfile == "not found") { groupfile = ""; }
266 m->splitAtDash(groupfile, groupfileNames);
268 //go through files and make sure they are good, if not, then disregard them
269 for (int i = 0; i < groupfileNames.size(); i++) {
270 if (inputDir != "") {
271 string path = m->hasPath(groupfileNames[i]);
272 //if the user has not given a path then, add inputdir. else leave path alone.
273 if (path == "") { groupfileNames[i] = inputDir + groupfileNames[i]; }
278 ableToOpen = m->openInputFile(groupfileNames[i], in, "noerror");
280 //if you can't open it, try default location
281 if (ableToOpen == 1) {
282 if (m->getDefaultPath() != "") { //default path is set
283 string tryPath = m->getDefaultPath() + m->getSimpleName(groupfileNames[i]);
284 m->mothurOut("Unable to open " + groupfileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine();
286 ableToOpen = m->openInputFile(tryPath, in2, "noerror");
288 groupfileNames[i] = tryPath;
292 if (ableToOpen == 1) {
293 if (m->getOutputDir() != "") { //default path is set
294 string tryPath = m->getOutputDir() + m->getSimpleName(groupfileNames[i]);
295 m->mothurOut("Unable to open " + groupfileNames[i] + ". Trying output directory " + tryPath); m->mothurOutEndLine();
297 ableToOpen = m->openInputFile(tryPath, in2, "noerror");
299 groupfileNames[i] = tryPath;
305 if (ableToOpen == 1) {
306 m->mothurOut("Unable to open " + groupfileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); groupfileNames[i] = "";
307 //erase from file list
308 groupfileNames.erase(groupfileNames.begin()+i);
314 if (groupfile != "") {
315 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(); }
317 for (int i = 0; i < fastaFileNames.size(); i++) { groupfileNames.push_back(""); }
320 //check for optional parameter and set defaults
321 // ...at some point should added some additional type checking...
323 temp = validParameter.validFile(parameters, "ksize", false); if (temp == "not found"){ temp = "8"; }
324 convert(temp, kmerSize);
326 temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = "1"; }
327 convert(temp, processors);
329 search = validParameter.validFile(parameters, "search", false); if (search == "not found"){ search = "kmer"; }
331 method = validParameter.validFile(parameters, "method", false); if (method == "not found"){ method = "bayesian"; }
333 temp = validParameter.validFile(parameters, "match", false); if (temp == "not found"){ temp = "1.0"; }
334 convert(temp, match);
336 temp = validParameter.validFile(parameters, "mismatch", false); if (temp == "not found"){ temp = "-1.0"; }
337 convert(temp, misMatch);
339 temp = validParameter.validFile(parameters, "gapopen", false); if (temp == "not found"){ temp = "-2.0"; }
340 convert(temp, gapOpen);
342 temp = validParameter.validFile(parameters, "gapextend", false); if (temp == "not found"){ temp = "-1.0"; }
343 convert(temp, gapExtend);
345 temp = validParameter.validFile(parameters, "numwanted", false); if (temp == "not found"){ temp = "10"; }
346 convert(temp, numWanted);
348 temp = validParameter.validFile(parameters, "cutoff", false); if (temp == "not found"){ temp = "0"; }
349 convert(temp, cutoff);
351 temp = validParameter.validFile(parameters, "probs", false); if (temp == "not found"){ temp = "true"; }
352 probs = m->isTrue(temp);
354 temp = validParameter.validFile(parameters, "iters", false); if (temp == "not found") { temp = "100"; }
355 convert(temp, iters);
359 if ((method == "bayesian") && (search != "kmer")) {
360 m->mothurOut("The bayesian method requires the kmer search." + search + "will be disregarded." ); m->mothurOutEndLine();
366 catch(exception& e) {
367 m->errorOut(e, "ClassifySeqsCommand", "ClassifySeqsCommand");
372 //**********************************************************************************************************************
374 ClassifySeqsCommand::~ClassifySeqsCommand(){
376 if (abort == false) {
377 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
381 //**********************************************************************************************************************
383 void ClassifySeqsCommand::help(){
385 m->mothurOut("The classify.seqs command reads a fasta file containing sequences and creates a .taxonomy file and a .tax.summary file.\n");
386 m->mothurOut("The classify.seqs command parameters are template, fasta, name, search, ksize, method, taxonomy, processors, match, mismatch, gapopen, gapextend, numwanted and probs.\n");
387 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");
388 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");
389 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");
390 m->mothurOut("The group parameter allows you add a group file so you can have the summary totals broken up by group.\n");
391 m->mothurOut("The method parameter allows you to specify classification method to use. Your options are: bayesian and knn. The default is bayesian.\n");
392 m->mothurOut("The ksize parameter allows you to specify the kmer size for finding most similar template to candidate. The default is 8.\n");
393 m->mothurOut("The processors parameter allows you to specify the number of processors to use. The default is 1.\n");
395 m->mothurOut("When using MPI, the processors parameter is set to the number of MPI processes running. \n");
397 m->mothurOut("The match parameter allows you to specify the bonus for having the same base. The default is 1.0.\n");
398 m->mothurOut("The mistmatch parameter allows you to specify the penalty for having different bases. The default is -1.0.\n");
399 m->mothurOut("The gapopen parameter allows you to specify the penalty for opening a gap in an alignment. The default is -2.0.\n");
400 m->mothurOut("The gapextend parameter allows you to specify the penalty for extending a gap in an alignment. The default is -1.0.\n");
401 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");
402 m->mothurOut("The cutoff parameter allows you to specify a bootstrap confidence threshold for your taxonomy. The default is 0.\n");
403 m->mothurOut("The probs parameter shuts off the bootstrapping results for the bayesian method. The default is true, meaning you want the bootstrapping to be shown.\n");
404 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");
405 m->mothurOut("The classify.seqs command should be in the following format: \n");
406 m->mothurOut("classify.seqs(template=yourTemplateFile, fasta=yourFastaFile, method=yourClassificationMethod, search=yourSearchmethod, ksize=yourKmerSize, taxonomy=yourTaxonomyFile, processors=yourProcessors) \n");
407 m->mothurOut("Example classify.seqs(fasta=amazon.fasta, template=core.filtered, method=knn, search=gotoh, ksize=8, processors=2)\n");
408 m->mothurOut("The .taxonomy file consists of 2 columns: 1 = your sequence name, 2 = the taxonomy for your sequence. \n");
409 m->mothurOut("The .tax.summary is a summary of the different taxonomies represented in your fasta file. \n");
410 m->mothurOut("Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFastaFile).\n\n");
412 catch(exception& e) {
413 m->errorOut(e, "ClassifySeqsCommand", "help");
419 //**********************************************************************************************************************
421 int ClassifySeqsCommand::execute(){
423 if (abort == true) { return 0; }
425 if(method == "bayesian"){ classify = new Bayesian(taxonomyFileName, templateFileName, search, kmerSize, cutoff, iters); }
426 else if(method == "knn"){ classify = new Knn(taxonomyFileName, templateFileName, search, kmerSize, gapOpen, gapExtend, match, misMatch, numWanted); }
428 m->mothurOut(search + " is not a valid method option. I will run the command using bayesian.");
429 m->mothurOutEndLine();
430 classify = new Bayesian(taxonomyFileName, templateFileName, search, kmerSize, cutoff, iters);
433 if (m->control_pressed) { delete classify; return 0; }
436 for (int s = 0; s < fastaFileNames.size(); s++) {
438 m->mothurOut("Classifying sequences from " + fastaFileNames[s] + " ..." ); m->mothurOutEndLine();
440 string RippedTaxName = m->getRootName(m->getSimpleName(taxonomyFileName));
441 RippedTaxName = m->getExtension(RippedTaxName.substr(0, RippedTaxName.length()-1));
442 if (RippedTaxName[0] == '.') { RippedTaxName = RippedTaxName.substr(1, RippedTaxName.length()); }
443 RippedTaxName += ".";
445 if (outputDir == "") { outputDir += m->hasPath(fastaFileNames[s]); }
446 string newTaxonomyFile = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + RippedTaxName + "taxonomy";
447 string tempTaxonomyFile = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + "taxonomy.temp";
448 string taxSummary = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + RippedTaxName + "tax.summary";
450 if ((method == "knn") && (search == "distance")) {
451 string DistName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + "match.dist";
452 classify->setDistName(DistName); outputNames.push_back(DistName); outputTypes["matchdist"].push_back(DistName);
455 outputNames.push_back(newTaxonomyFile); outputTypes["taxonomy"].push_back(newTaxonomyFile);
456 outputNames.push_back(taxSummary); outputTypes["taxsummary"].push_back(taxSummary);
458 int start = time(NULL);
459 int numFastaSeqs = 0;
460 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
463 int pid, end, numSeqsPerProcessor;
465 vector<unsigned long int> MPIPos;
468 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
469 MPI_Comm_size(MPI_COMM_WORLD, &processors);
472 MPI_File outMPINewTax;
473 MPI_File outMPITempTax;
475 int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
476 int inMode=MPI_MODE_RDONLY;
478 char outNewTax[1024];
479 strcpy(outNewTax, newTaxonomyFile.c_str());
481 char outTempTax[1024];
482 strcpy(outTempTax, tempTaxonomyFile.c_str());
484 char inFileName[1024];
485 strcpy(inFileName, fastaFileNames[s].c_str());
487 MPI_File_open(MPI_COMM_WORLD, inFileName, inMode, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer
488 MPI_File_open(MPI_COMM_WORLD, outNewTax, outMode, MPI_INFO_NULL, &outMPINewTax);
489 MPI_File_open(MPI_COMM_WORLD, outTempTax, outMode, MPI_INFO_NULL, &outMPITempTax);
491 if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); delete classify; return 0; }
493 if (pid == 0) { //you are the root process
495 MPIPos = m->setFilePosFasta(fastaFileNames[s], numFastaSeqs); //fills MPIPos, returns numSeqs
497 //send file positions to all processes
498 for(int i = 1; i < processors; i++) {
499 MPI_Send(&numFastaSeqs, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
500 MPI_Send(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, i, tag, MPI_COMM_WORLD);
503 //figure out how many sequences you have to align
504 numSeqsPerProcessor = numFastaSeqs / processors;
505 int startIndex = pid * numSeqsPerProcessor;
506 if(pid == (processors - 1)){ numSeqsPerProcessor = numFastaSeqs - pid * numSeqsPerProcessor; }
510 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPINewTax, outMPITempTax, MPIPos);
512 if (m->control_pressed) { outputTypes.clear(); 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; }
514 for (int i = 1; i < processors; i++) {
516 MPI_Recv(&done, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status);
518 }else{ //you are a child process
519 MPI_Recv(&numFastaSeqs, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
520 MPIPos.resize(numFastaSeqs+1);
521 MPI_Recv(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status);
523 //figure out how many sequences you have to align
524 numSeqsPerProcessor = numFastaSeqs / processors;
525 int startIndex = pid * numSeqsPerProcessor;
526 if(pid == (processors - 1)){ numSeqsPerProcessor = numFastaSeqs - pid * numSeqsPerProcessor; }
530 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPINewTax, outMPITempTax, MPIPos);
532 if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); delete classify; return 0; }
535 MPI_Send(&done, 1, MPI_INT, 0, tag, MPI_COMM_WORLD);
539 MPI_File_close(&inMPI);
540 MPI_File_close(&outMPINewTax);
541 MPI_File_close(&outMPITempTax);
542 MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
546 vector<unsigned long int> positions = m->divideFile(fastaFileNames[s], processors);
548 for (int i = 0; i < (positions.size()-1); i++) {
549 lines.push_back(new linePair(positions[i], positions[(i+1)]));
552 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
554 numFastaSeqs = driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]);
557 processIDS.resize(0);
559 numFastaSeqs = createProcesses(newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]);
561 rename((newTaxonomyFile + toString(processIDS[0]) + ".temp").c_str(), newTaxonomyFile.c_str());
562 rename((tempTaxonomyFile + toString(processIDS[0]) + ".temp").c_str(), tempTaxonomyFile.c_str());
564 for(int i=1;i<processors;i++){
565 appendTaxFiles((newTaxonomyFile + toString(processIDS[i]) + ".temp"), newTaxonomyFile);
566 appendTaxFiles((tempTaxonomyFile + toString(processIDS[i]) + ".temp"), tempTaxonomyFile);
567 remove((newTaxonomyFile + toString(processIDS[i]) + ".temp").c_str());
568 remove((tempTaxonomyFile + toString(processIDS[i]) + ".temp").c_str());
573 numFastaSeqs = driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]);
577 m->mothurOutEndLine();
578 m->mothurOut("It took " + toString(time(NULL) - start) + " secs to classify " + toString(numFastaSeqs) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine();
583 if (pid == 0) { //this part does not need to be paralellized
585 if(namefile != "") { m->mothurOut("Reading " + namefileNames[s] + "..."); cout.flush(); MPIReadNamesFile(namefileNames[s]); m->mothurOut(" Done."); m->mothurOutEndLine(); }
590 m->mothurOut("Reading " + namefileNames[s] + "..."); cout.flush();
592 nameMap.clear(); //remove old names
595 m->openInputFile(namefileNames[s], inNames);
597 string firstCol, secondCol;
598 while(!inNames.eof()) {
599 inNames >> firstCol >> secondCol; m->gobble(inNames);
602 m->splitAtComma(secondCol, temp);
604 nameMap[firstCol] = temp;
608 m->mothurOut(" Done."); m->mothurOutEndLine();
613 if (groupfile != "") { group = groupfileNames[s]; }
615 PhyloSummary taxaSum(taxonomyFileName, group);
617 if (m->control_pressed) { outputTypes.clear(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete classify; return 0; }
619 if (namefile == "") { taxaSum.summarize(tempTaxonomyFile); }
622 m->openInputFile(tempTaxonomyFile, in);
624 //read in users taxonomy file and add sequences to tree
628 in >> name >> taxon; m->gobble(in);
630 itNames = nameMap.find(name);
632 if (itNames == nameMap.end()) {
633 m->mothurOut(name + " is not in your name file please correct."); m->mothurOutEndLine(); exit(1);
635 for (int i = 0; i < itNames->second.size(); i++) {
636 taxaSum.addSeqToTree(itNames->second[i], taxon); //add it as many times as there are identical seqs
638 itNames->second.clear();
639 nameMap.erase(itNames->first);
644 remove(tempTaxonomyFile.c_str());
646 if (m->control_pressed) { outputTypes.clear(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete classify; return 0; }
650 m->openOutputFile(taxSummary, outTaxTree);
651 taxaSum.print(outTaxTree);
654 //output taxonomy with the unclassified bins added
656 m->openInputFile(newTaxonomyFile, inTax);
659 string unclass = newTaxonomyFile + ".unclass.temp";
660 m->openOutputFile(unclass, outTax);
662 //get maxLevel from phylotree so you know how many 'unclassified's to add
663 int maxLevel = taxaSum.getMaxLevel();
665 //read taxfile - this reading and rewriting is done to preserve the confidence scores.
667 while (!inTax.eof()) {
668 if (m->control_pressed) { outputTypes.clear(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } remove(unclass.c_str()); delete classify; return 0; }
670 inTax >> name >> taxon; m->gobble(inTax);
672 string newTax = addUnclassifieds(taxon, maxLevel);
674 outTax << name << '\t' << newTax << endl;
679 remove(newTaxonomyFile.c_str());
680 rename(unclass.c_str(), newTaxonomyFile.c_str());
682 m->mothurOutEndLine();
683 m->mothurOut("It took " + toString(time(NULL) - start) + " secs to create the summary file for " + toString(numFastaSeqs) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine();
689 m->mothurOutEndLine();
690 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
691 for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
692 m->mothurOutEndLine();
698 catch(exception& e) {
699 m->errorOut(e, "ClassifySeqsCommand", "execute");
704 /**************************************************************************************************/
705 string ClassifySeqsCommand::addUnclassifieds(string tax, int maxlevel) {
707 string newTax, taxon;
710 //keep what you have counting the levels
711 while (tax.find_first_of(';') != -1) {
713 taxon = tax.substr(0,tax.find_first_of(';'))+';';
714 tax = tax.substr(tax.find_first_of(';')+1, tax.length());
719 //add "unclassified" until you reach maxLevel
720 while (level < maxlevel) {
721 newTax += "unclassified;";
727 catch(exception& e) {
728 m->errorOut(e, "ClassifySeqsCommand", "addUnclassifieds");
733 /**************************************************************************************************/
735 int ClassifySeqsCommand::createProcesses(string taxFileName, string tempTaxFile, string filename) {
737 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
741 //loop through and create all the processes you want
742 while (process != processors) {
746 processIDS.push_back(pid); //create map from line number to pid so you can append files in correct order later
749 num = driver(lines[process], taxFileName + toString(getpid()) + ".temp", tempTaxFile + toString(getpid()) + ".temp", filename);
751 //pass numSeqs to parent
753 string tempFile = filename + toString(getpid()) + ".num.temp";
754 m->openOutputFile(tempFile, out);
759 }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
762 //force parent to wait until all the processes are done
763 for (int i=0;i<processors;i++) {
764 int temp = processIDS[i];
768 for (int i = 0; i < processIDS.size(); i++) {
770 string tempFile = filename + toString(processIDS[i]) + ".num.temp";
771 m->openInputFile(tempFile, in);
772 if (!in.eof()) { int tempNum = 0; in >> tempNum; num += tempNum; }
773 in.close(); remove(tempFile.c_str());
779 catch(exception& e) {
780 m->errorOut(e, "ClassifySeqsCommand", "createProcesses");
784 /**************************************************************************************************/
786 void ClassifySeqsCommand::appendTaxFiles(string temp, string filename) {
791 m->openOutputFileAppend(filename, output);
792 m->openInputFile(temp, input);
794 while(char c = input.get()){
795 if(input.eof()) { break; }
796 else { output << c; }
802 catch(exception& e) {
803 m->errorOut(e, "ClassifySeqsCommand", "appendTaxFiles");
808 //**********************************************************************************************************************
810 int ClassifySeqsCommand::driver(linePair* filePos, string taxFName, string tempTFName, string filename){
813 m->openOutputFile(taxFName, outTax);
815 ofstream outTaxSimple;
816 m->openOutputFile(tempTFName, outTaxSimple);
819 m->openInputFile(filename, inFASTA);
823 inFASTA.seekg(filePos->start);
829 if (m->control_pressed) { return 0; }
831 Sequence* candidateSeq = new Sequence(inFASTA); m->gobble(inFASTA);
833 if (candidateSeq->getName() != "") {
834 taxonomy = classify->getTaxonomy(candidateSeq);
836 if (m->control_pressed) { delete candidateSeq; return 0; }
838 if (taxonomy != "bad seq") {
839 //output confidence scores or not
841 outTax << candidateSeq->getName() << '\t' << taxonomy << endl;
843 outTax << candidateSeq->getName() << '\t' << classify->getSimpleTax() << endl;
846 outTaxSimple << candidateSeq->getName() << '\t' << classify->getSimpleTax() << endl;
852 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
853 unsigned long int pos = inFASTA.tellg();
854 if ((pos == -1) || (pos >= filePos->end)) { break; }
856 if (inFASTA.eof()) { break; }
860 if((count) % 100 == 0){ m->mothurOut("Processing sequence: " + toString(count)); m->mothurOutEndLine(); }
863 if((count) % 100 != 0){ m->mothurOut("Processing sequence: " + toString(count)); m->mothurOutEndLine(); }
867 outTaxSimple.close();
871 catch(exception& e) {
872 m->errorOut(e, "ClassifySeqsCommand", "driver");
876 //**********************************************************************************************************************
878 int ClassifySeqsCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& newFile, MPI_File& tempFile, vector<unsigned long int>& MPIPos){
880 MPI_Status statusNew;
881 MPI_Status statusTemp;
885 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
890 for(int i=0;i<num;i++){
892 if (m->control_pressed) { return 0; }
895 int length = MPIPos[start+i+1] - MPIPos[start+i];
896 char* buf4 = new char[length];
897 MPI_File_read_at(inMPI, MPIPos[start+i], buf4, length, MPI_CHAR, &status);
899 string tempBuf = buf4;
900 if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); }
901 istringstream iss (tempBuf,istringstream::in);
904 Sequence* candidateSeq = new Sequence(iss);
906 if (candidateSeq->getName() != "") {
907 taxonomy = classify->getTaxonomy(candidateSeq);
909 if (taxonomy != "bad seq") {
910 //output confidence scores or not
912 outputString = candidateSeq->getName() + "\t" + taxonomy + "\n";
914 outputString = candidateSeq->getName() + "\t" + classify->getSimpleTax() + "\n";
917 int length = outputString.length();
918 char* buf2 = new char[length];
919 memcpy(buf2, outputString.c_str(), length);
921 MPI_File_write_shared(newFile, buf2, length, MPI_CHAR, &statusNew);
924 outputString = candidateSeq->getName() + "\t" + classify->getSimpleTax() + "\n";
925 length = outputString.length();
926 char* buf = new char[length];
927 memcpy(buf, outputString.c_str(), length);
929 MPI_File_write_shared(tempFile, buf, length, MPI_CHAR, &statusTemp);
935 if((i+1) % 100 == 0){ cout << "Classifying sequence " << (i+1) << endl; }
938 if(num % 100 != 0){ cout << "Classifying sequence " << (num) << endl; }
943 catch(exception& e) {
944 m->errorOut(e, "ClassifySeqsCommand", "driverMPI");
949 //**********************************************************************************************************************
950 int ClassifySeqsCommand::MPIReadNamesFile(string nameFilename){
953 nameMap.clear(); //remove old names
959 //char* inFileName = new char[nameFilename.length()];
960 //memcpy(inFileName, nameFilename.c_str(), nameFilename.length());
962 char inFileName[1024];
963 strcpy(inFileName, nameFilename.c_str());
965 MPI_File_open(MPI_COMM_WORLD, inFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI);
966 MPI_File_get_size(inMPI, &size);
969 char* buffer = new char[size];
970 MPI_File_read(inMPI, buffer, size, MPI_CHAR, &status);
972 string tempBuf = buffer;
973 if (tempBuf.length() > size) { tempBuf = tempBuf.substr(0, size); }
974 istringstream iss (tempBuf,istringstream::in);
977 string firstCol, secondCol;
979 iss >> firstCol >> secondCol; m->gobble(iss);
982 m->splitAtComma(secondCol, temp);
984 nameMap[firstCol] = temp;
987 MPI_File_close(&inMPI);
991 catch(exception& e) {
992 m->errorOut(e, "ClassifySeqsCommand", "MPIReadNamesFile");
997 /**************************************************************************************************/