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();
164 ableToOpen = m->openInputFile(tryPath, in, "noerror");
165 fastaFileNames[i] = tryPath;
169 if (ableToOpen == 1) {
170 if (m->getOutputDir() != "") { //default path is set
171 string tryPath = m->getOutputDir() + m->getSimpleName(fastaFileNames[i]);
172 m->mothurOut("Unable to open " + fastaFileNames[i] + ". Trying output directory " + tryPath); m->mothurOutEndLine();
173 ableToOpen = m->openInputFile(tryPath, in, "noerror");
174 fastaFileNames[i] = tryPath;
180 if (ableToOpen == 1) {
181 m->mothurOut("Unable to open " + fastaFileNames[i] + ". It will be disregarded."); m->mothurOutEndLine();
182 //erase from file list
183 fastaFileNames.erase(fastaFileNames.begin()+i);
189 //make sure there is at least one valid file left
190 if (fastaFileNames.size() == 0) { m->mothurOut("no valid files."); m->mothurOutEndLine(); abort = true; }
194 taxonomyFileName = validParameter.validFile(parameters, "taxonomy", true);
195 if (taxonomyFileName == "not found") {
196 m->mothurOut("taxonomy is a required parameter for the classify.seqs command.");
197 m->mothurOutEndLine();
200 else if (taxonomyFileName == "not open") { abort = true; }
203 namefile = validParameter.validFile(parameters, "name", false);
204 if (namefile == "not found") { namefile = ""; }
207 m->splitAtDash(namefile, namefileNames);
209 //go through files and make sure they are good, if not, then disregard them
210 for (int i = 0; i < namefileNames.size(); i++) {
211 if (inputDir != "") {
212 string path = m->hasPath(namefileNames[i]);
213 //if the user has not given a path then, add inputdir. else leave path alone.
214 if (path == "") { namefileNames[i] = inputDir + namefileNames[i]; }
219 ableToOpen = m->openInputFile(namefileNames[i], in, "noerror");
221 //if you can't open it, try default location
222 if (ableToOpen == 1) {
223 if (m->getDefaultPath() != "") { //default path is set
224 string tryPath = m->getDefaultPath() + m->getSimpleName(namefileNames[i]);
225 m->mothurOut("Unable to open " + namefileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine();
226 ableToOpen = m->openInputFile(tryPath, in, "noerror");
227 namefileNames[i] = tryPath;
231 if (ableToOpen == 1) {
232 if (m->getOutputDir() != "") { //default path is set
233 string tryPath = m->getOutputDir() + m->getSimpleName(namefileNames[i]);
234 m->mothurOut("Unable to open " + namefileNames[i] + ". Trying output directory " + tryPath); m->mothurOutEndLine();
235 ableToOpen = m->openInputFile(tryPath, in, "noerror");
236 namefileNames[i] = tryPath;
241 if (ableToOpen == 1) {
242 m->mothurOut("Unable to open " + namefileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); abort = true;
243 //erase from file list
244 namefileNames.erase(namefileNames.begin()+i);
251 if (namefile != "") {
252 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(); }
255 groupfile = validParameter.validFile(parameters, "group", false);
256 if (groupfile == "not found") { groupfile = ""; }
258 m->splitAtDash(groupfile, groupfileNames);
260 //go through files and make sure they are good, if not, then disregard them
261 for (int i = 0; i < groupfileNames.size(); i++) {
262 if (inputDir != "") {
263 string path = m->hasPath(groupfileNames[i]);
264 //if the user has not given a path then, add inputdir. else leave path alone.
265 if (path == "") { groupfileNames[i] = inputDir + groupfileNames[i]; }
270 ableToOpen = m->openInputFile(groupfileNames[i], in, "noerror");
272 //if you can't open it, try default location
273 if (ableToOpen == 1) {
274 if (m->getDefaultPath() != "") { //default path is set
275 string tryPath = m->getDefaultPath() + m->getSimpleName(groupfileNames[i]);
276 m->mothurOut("Unable to open " + groupfileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine();
277 ableToOpen = m->openInputFile(tryPath, in, "noerror");
278 groupfileNames[i] = tryPath;
282 if (ableToOpen == 1) {
283 if (m->getOutputDir() != "") { //default path is set
284 string tryPath = m->getOutputDir() + m->getSimpleName(groupfileNames[i]);
285 m->mothurOut("Unable to open " + groupfileNames[i] + ". Trying output directory " + tryPath); m->mothurOutEndLine();
286 ableToOpen = m->openInputFile(tryPath, in, "noerror");
287 groupfileNames[i] = tryPath;
293 if (ableToOpen == 1) {
294 m->mothurOut("Unable to open " + groupfileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); groupfileNames[i] = "";
295 //erase from file list
296 groupfileNames.erase(groupfileNames.begin()+i);
302 if (groupfile != "") {
303 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(); }
305 for (int i = 0; i < fastaFileNames.size(); i++) { groupfileNames.push_back(""); }
308 //check for optional parameter and set defaults
309 // ...at some point should added some additional type checking...
311 temp = validParameter.validFile(parameters, "ksize", false); if (temp == "not found"){ temp = "8"; }
312 convert(temp, kmerSize);
314 temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = "1"; }
315 convert(temp, processors);
317 search = validParameter.validFile(parameters, "search", false); if (search == "not found"){ search = "kmer"; }
319 method = validParameter.validFile(parameters, "method", false); if (method == "not found"){ method = "bayesian"; }
321 temp = validParameter.validFile(parameters, "match", false); if (temp == "not found"){ temp = "1.0"; }
322 convert(temp, match);
324 temp = validParameter.validFile(parameters, "mismatch", false); if (temp == "not found"){ temp = "-1.0"; }
325 convert(temp, misMatch);
327 temp = validParameter.validFile(parameters, "gapopen", false); if (temp == "not found"){ temp = "-2.0"; }
328 convert(temp, gapOpen);
330 temp = validParameter.validFile(parameters, "gapextend", false); if (temp == "not found"){ temp = "-1.0"; }
331 convert(temp, gapExtend);
333 temp = validParameter.validFile(parameters, "numwanted", false); if (temp == "not found"){ temp = "10"; }
334 convert(temp, numWanted);
336 temp = validParameter.validFile(parameters, "cutoff", false); if (temp == "not found"){ temp = "0"; }
337 convert(temp, cutoff);
339 temp = validParameter.validFile(parameters, "probs", false); if (temp == "not found"){ temp = "true"; }
340 probs = m->isTrue(temp);
342 temp = validParameter.validFile(parameters, "iters", false); if (temp == "not found") { temp = "100"; }
343 convert(temp, iters);
347 if ((method == "bayesian") && (search != "kmer")) {
348 m->mothurOut("The bayesian method requires the kmer search." + search + "will be disregarded." ); m->mothurOutEndLine();
354 catch(exception& e) {
355 m->errorOut(e, "ClassifySeqsCommand", "ClassifySeqsCommand");
360 //**********************************************************************************************************************
362 ClassifySeqsCommand::~ClassifySeqsCommand(){
364 if (abort == false) {
365 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
369 //**********************************************************************************************************************
371 void ClassifySeqsCommand::help(){
373 m->mothurOut("The classify.seqs command reads a fasta file containing sequences and creates a .taxonomy file and a .tax.summary file.\n");
374 m->mothurOut("The classify.seqs command parameters are template, fasta, name, search, ksize, method, taxonomy, processors, match, mismatch, gapopen, gapextend, numwanted and probs.\n");
375 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");
376 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");
377 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");
378 m->mothurOut("The group parameter allows you add a group file so you can have the summary totals broken up by group.\n");
379 m->mothurOut("The method parameter allows you to specify classification method to use. Your options are: bayesian and knn. The default is bayesian.\n");
380 m->mothurOut("The ksize parameter allows you to specify the kmer size for finding most similar template to candidate. The default is 8.\n");
381 m->mothurOut("The processors parameter allows you to specify the number of processors to use. The default is 1.\n");
383 m->mothurOut("When using MPI, the processors parameter is set to the number of MPI processes running. \n");
385 m->mothurOut("The match parameter allows you to specify the bonus for having the same base. The default is 1.0.\n");
386 m->mothurOut("The mistmatch parameter allows you to specify the penalty for having different bases. The default is -1.0.\n");
387 m->mothurOut("The gapopen parameter allows you to specify the penalty for opening a gap in an alignment. The default is -2.0.\n");
388 m->mothurOut("The gapextend parameter allows you to specify the penalty for extending a gap in an alignment. The default is -1.0.\n");
389 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");
390 m->mothurOut("The cutoff parameter allows you to specify a bootstrap confidence threshold for your taxonomy. The default is 0.\n");
391 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");
392 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");
393 m->mothurOut("The classify.seqs command should be in the following format: \n");
394 m->mothurOut("classify.seqs(template=yourTemplateFile, fasta=yourFastaFile, method=yourClassificationMethod, search=yourSearchmethod, ksize=yourKmerSize, taxonomy=yourTaxonomyFile, processors=yourProcessors) \n");
395 m->mothurOut("Example classify.seqs(fasta=amazon.fasta, template=core.filtered, method=knn, search=gotoh, ksize=8, processors=2)\n");
396 m->mothurOut("The .taxonomy file consists of 2 columns: 1 = your sequence name, 2 = the taxonomy for your sequence. \n");
397 m->mothurOut("The .tax.summary is a summary of the different taxonomies represented in your fasta file. \n");
398 m->mothurOut("Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFastaFile).\n\n");
400 catch(exception& e) {
401 m->errorOut(e, "ClassifySeqsCommand", "help");
407 //**********************************************************************************************************************
409 int ClassifySeqsCommand::execute(){
411 if (abort == true) { return 0; }
413 if(method == "bayesian"){ classify = new Bayesian(taxonomyFileName, templateFileName, search, kmerSize, cutoff, iters); }
414 else if(method == "knn"){ classify = new Knn(taxonomyFileName, templateFileName, search, kmerSize, gapOpen, gapExtend, match, misMatch, numWanted); }
416 m->mothurOut(search + " is not a valid method option. I will run the command using bayesian.");
417 m->mothurOutEndLine();
418 classify = new Bayesian(taxonomyFileName, templateFileName, search, kmerSize, cutoff, iters);
421 if (m->control_pressed) { delete classify; return 0; }
424 for (int s = 0; s < fastaFileNames.size(); s++) {
426 m->mothurOut("Classifying sequences from " + fastaFileNames[s] + " ..." ); m->mothurOutEndLine();
428 string RippedTaxName = m->getRootName(m->getSimpleName(taxonomyFileName));
429 RippedTaxName = m->getExtension(RippedTaxName.substr(0, RippedTaxName.length()-1));
430 if (RippedTaxName[0] == '.') { RippedTaxName = RippedTaxName.substr(1, RippedTaxName.length()); }
431 RippedTaxName += ".";
433 if (outputDir == "") { outputDir += m->hasPath(fastaFileNames[s]); }
434 string newTaxonomyFile = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + RippedTaxName + "taxonomy";
435 string tempTaxonomyFile = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + "taxonomy.temp";
436 string taxSummary = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + RippedTaxName + "tax.summary";
438 if ((method == "knn") && (search == "distance")) {
439 string DistName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + "match.dist";
440 classify->setDistName(DistName); outputNames.push_back(DistName); outputTypes["matchdist"].push_back(DistName);
443 outputNames.push_back(newTaxonomyFile); outputTypes["taxonomy"].push_back(newTaxonomyFile);
444 outputNames.push_back(taxSummary); outputTypes["taxsummary"].push_back(taxSummary);
446 int start = time(NULL);
447 int numFastaSeqs = 0;
448 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
451 int pid, end, numSeqsPerProcessor;
453 vector<unsigned long int> MPIPos;
456 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
457 MPI_Comm_size(MPI_COMM_WORLD, &processors);
460 MPI_File outMPINewTax;
461 MPI_File outMPITempTax;
463 int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
464 int inMode=MPI_MODE_RDONLY;
466 char outNewTax[1024];
467 strcpy(outNewTax, newTaxonomyFile.c_str());
469 char outTempTax[1024];
470 strcpy(outTempTax, tempTaxonomyFile.c_str());
472 char inFileName[1024];
473 strcpy(inFileName, fastaFileNames[s].c_str());
475 MPI_File_open(MPI_COMM_WORLD, inFileName, inMode, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer
476 MPI_File_open(MPI_COMM_WORLD, outNewTax, outMode, MPI_INFO_NULL, &outMPINewTax);
477 MPI_File_open(MPI_COMM_WORLD, outTempTax, outMode, MPI_INFO_NULL, &outMPITempTax);
479 if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); delete classify; return 0; }
481 if (pid == 0) { //you are the root process
483 MPIPos = m->setFilePosFasta(fastaFileNames[s], numFastaSeqs); //fills MPIPos, returns numSeqs
485 //send file positions to all processes
486 for(int i = 1; i < processors; i++) {
487 MPI_Send(&numFastaSeqs, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
488 MPI_Send(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, i, tag, MPI_COMM_WORLD);
491 //figure out how many sequences you have to align
492 numSeqsPerProcessor = numFastaSeqs / processors;
493 int startIndex = pid * numSeqsPerProcessor;
494 if(pid == (processors - 1)){ numSeqsPerProcessor = numFastaSeqs - pid * numSeqsPerProcessor; }
498 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPINewTax, outMPITempTax, MPIPos);
500 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; }
502 for (int i = 1; i < processors; i++) {
504 MPI_Recv(&done, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status);
506 }else{ //you are a child process
507 MPI_Recv(&numFastaSeqs, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
508 MPIPos.resize(numFastaSeqs+1);
509 MPI_Recv(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status);
511 //figure out how many sequences you have to align
512 numSeqsPerProcessor = numFastaSeqs / processors;
513 int startIndex = pid * numSeqsPerProcessor;
514 if(pid == (processors - 1)){ numSeqsPerProcessor = numFastaSeqs - pid * numSeqsPerProcessor; }
518 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPINewTax, outMPITempTax, MPIPos);
520 if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); delete classify; return 0; }
523 MPI_Send(&done, 1, MPI_INT, 0, tag, MPI_COMM_WORLD);
527 MPI_File_close(&inMPI);
528 MPI_File_close(&outMPINewTax);
529 MPI_File_close(&outMPITempTax);
530 MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
534 vector<unsigned long int> positions = m->divideFile(fastaFileNames[s], processors);
536 for (int i = 0; i < (positions.size()-1); i++) {
537 lines.push_back(new linePair(positions[i], positions[(i+1)]));
540 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
542 numFastaSeqs = driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]);
545 processIDS.resize(0);
547 numFastaSeqs = createProcesses(newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]);
549 rename((newTaxonomyFile + toString(processIDS[0]) + ".temp").c_str(), newTaxonomyFile.c_str());
550 rename((tempTaxonomyFile + toString(processIDS[0]) + ".temp").c_str(), tempTaxonomyFile.c_str());
552 for(int i=1;i<processors;i++){
553 appendTaxFiles((newTaxonomyFile + toString(processIDS[i]) + ".temp"), newTaxonomyFile);
554 appendTaxFiles((tempTaxonomyFile + toString(processIDS[i]) + ".temp"), tempTaxonomyFile);
555 remove((newTaxonomyFile + toString(processIDS[i]) + ".temp").c_str());
556 remove((tempTaxonomyFile + toString(processIDS[i]) + ".temp").c_str());
561 numFastaSeqs = driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]);
565 m->mothurOutEndLine();
566 m->mothurOut("It took " + toString(time(NULL) - start) + " secs to classify " + toString(numFastaSeqs) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine();
571 if (pid == 0) { //this part does not need to be paralellized
573 if(namefile != "") { m->mothurOut("Reading " + namefileNames[s] + "..."); cout.flush(); MPIReadNamesFile(namefileNames[s]); m->mothurOut(" Done."); m->mothurOutEndLine(); }
578 m->mothurOut("Reading " + namefileNames[s] + "..."); cout.flush();
580 nameMap.clear(); //remove old names
583 m->openInputFile(namefileNames[s], inNames);
585 string firstCol, secondCol;
586 while(!inNames.eof()) {
587 inNames >> firstCol >> secondCol; m->gobble(inNames);
590 m->splitAtComma(secondCol, temp);
592 nameMap[firstCol] = temp;
596 m->mothurOut(" Done."); m->mothurOutEndLine();
601 if (groupfile != "") { group = groupfileNames[s]; }
603 PhyloSummary taxaSum(taxonomyFileName, group);
605 if (m->control_pressed) { outputTypes.clear(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete classify; return 0; }
607 if (namefile == "") { taxaSum.summarize(tempTaxonomyFile); }
610 m->openInputFile(tempTaxonomyFile, in);
612 //read in users taxonomy file and add sequences to tree
616 in >> name >> taxon; m->gobble(in);
618 itNames = nameMap.find(name);
620 if (itNames == nameMap.end()) {
621 m->mothurOut(name + " is not in your name file please correct."); m->mothurOutEndLine(); exit(1);
623 for (int i = 0; i < itNames->second.size(); i++) {
624 taxaSum.addSeqToTree(itNames->second[i], taxon); //add it as many times as there are identical seqs
626 itNames->second.clear();
627 nameMap.erase(itNames->first);
632 remove(tempTaxonomyFile.c_str());
634 if (m->control_pressed) { outputTypes.clear(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete classify; return 0; }
638 m->openOutputFile(taxSummary, outTaxTree);
639 taxaSum.print(outTaxTree);
642 //output taxonomy with the unclassified bins added
644 m->openInputFile(newTaxonomyFile, inTax);
647 string unclass = newTaxonomyFile + ".unclass.temp";
648 m->openOutputFile(unclass, outTax);
650 //get maxLevel from phylotree so you know how many 'unclassified's to add
651 int maxLevel = taxaSum.getMaxLevel();
653 //read taxfile - this reading and rewriting is done to preserve the confidence scores.
655 while (!inTax.eof()) {
656 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; }
658 inTax >> name >> taxon; m->gobble(inTax);
660 string newTax = addUnclassifieds(taxon, maxLevel);
662 outTax << name << '\t' << newTax << endl;
667 remove(newTaxonomyFile.c_str());
668 rename(unclass.c_str(), newTaxonomyFile.c_str());
670 m->mothurOutEndLine();
671 m->mothurOut("It took " + toString(time(NULL) - start) + " secs to create the summary file for " + toString(numFastaSeqs) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine();
677 m->mothurOutEndLine();
678 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
679 for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
680 m->mothurOutEndLine();
686 catch(exception& e) {
687 m->errorOut(e, "ClassifySeqsCommand", "execute");
692 /**************************************************************************************************/
693 string ClassifySeqsCommand::addUnclassifieds(string tax, int maxlevel) {
695 string newTax, taxon;
698 //keep what you have counting the levels
699 while (tax.find_first_of(';') != -1) {
701 taxon = tax.substr(0,tax.find_first_of(';'))+';';
702 tax = tax.substr(tax.find_first_of(';')+1, tax.length());
707 //add "unclassified" until you reach maxLevel
708 while (level < maxlevel) {
709 newTax += "unclassified;";
715 catch(exception& e) {
716 m->errorOut(e, "ClassifySeqsCommand", "addUnclassifieds");
721 /**************************************************************************************************/
723 int ClassifySeqsCommand::createProcesses(string taxFileName, string tempTaxFile, string filename) {
725 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
729 //loop through and create all the processes you want
730 while (process != processors) {
734 processIDS.push_back(pid); //create map from line number to pid so you can append files in correct order later
737 num = driver(lines[process], taxFileName + toString(getpid()) + ".temp", tempTaxFile + toString(getpid()) + ".temp", filename);
739 //pass numSeqs to parent
741 string tempFile = filename + toString(getpid()) + ".num.temp";
742 m->openOutputFile(tempFile, out);
747 }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
750 //force parent to wait until all the processes are done
751 for (int i=0;i<processors;i++) {
752 int temp = processIDS[i];
756 for (int i = 0; i < processIDS.size(); i++) {
758 string tempFile = filename + toString(processIDS[i]) + ".num.temp";
759 m->openInputFile(tempFile, in);
760 if (!in.eof()) { int tempNum = 0; in >> tempNum; num += tempNum; }
761 in.close(); remove(tempFile.c_str());
767 catch(exception& e) {
768 m->errorOut(e, "ClassifySeqsCommand", "createProcesses");
772 /**************************************************************************************************/
774 void ClassifySeqsCommand::appendTaxFiles(string temp, string filename) {
779 m->openOutputFileAppend(filename, output);
780 m->openInputFile(temp, input);
782 while(char c = input.get()){
783 if(input.eof()) { break; }
784 else { output << c; }
790 catch(exception& e) {
791 m->errorOut(e, "ClassifySeqsCommand", "appendTaxFiles");
796 //**********************************************************************************************************************
798 int ClassifySeqsCommand::driver(linePair* filePos, string taxFName, string tempTFName, string filename){
801 m->openOutputFile(taxFName, outTax);
803 ofstream outTaxSimple;
804 m->openOutputFile(tempTFName, outTaxSimple);
807 m->openInputFile(filename, inFASTA);
811 inFASTA.seekg(filePos->start);
817 if (m->control_pressed) { return 0; }
819 Sequence* candidateSeq = new Sequence(inFASTA); m->gobble(inFASTA);
821 if (candidateSeq->getName() != "") {
822 taxonomy = classify->getTaxonomy(candidateSeq);
824 if (m->control_pressed) { delete candidateSeq; return 0; }
826 if (taxonomy != "bad seq") {
827 //output confidence scores or not
829 outTax << candidateSeq->getName() << '\t' << taxonomy << endl;
831 outTax << candidateSeq->getName() << '\t' << classify->getSimpleTax() << endl;
834 outTaxSimple << candidateSeq->getName() << '\t' << classify->getSimpleTax() << endl;
840 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
841 unsigned long int pos = inFASTA.tellg();
842 if ((pos == -1) || (pos >= filePos->end)) { break; }
844 if (inFASTA.eof()) { break; }
848 if((count) % 100 == 0){ m->mothurOut("Processing sequence: " + toString(count)); m->mothurOutEndLine(); }
851 if((count) % 100 != 0){ m->mothurOut("Processing sequence: " + toString(count)); m->mothurOutEndLine(); }
855 outTaxSimple.close();
859 catch(exception& e) {
860 m->errorOut(e, "ClassifySeqsCommand", "driver");
864 //**********************************************************************************************************************
866 int ClassifySeqsCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& newFile, MPI_File& tempFile, vector<unsigned long int>& MPIPos){
868 MPI_Status statusNew;
869 MPI_Status statusTemp;
873 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
878 for(int i=0;i<num;i++){
880 if (m->control_pressed) { return 0; }
883 int length = MPIPos[start+i+1] - MPIPos[start+i];
884 char* buf4 = new char[length];
885 MPI_File_read_at(inMPI, MPIPos[start+i], buf4, length, MPI_CHAR, &status);
887 string tempBuf = buf4;
888 if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); }
889 istringstream iss (tempBuf,istringstream::in);
892 Sequence* candidateSeq = new Sequence(iss);
894 if (candidateSeq->getName() != "") {
895 taxonomy = classify->getTaxonomy(candidateSeq);
897 if (taxonomy != "bad seq") {
898 //output confidence scores or not
900 outputString = candidateSeq->getName() + "\t" + taxonomy + "\n";
902 outputString = candidateSeq->getName() + "\t" + classify->getSimpleTax() + "\n";
905 int length = outputString.length();
906 char* buf2 = new char[length];
907 memcpy(buf2, outputString.c_str(), length);
909 MPI_File_write_shared(newFile, buf2, length, MPI_CHAR, &statusNew);
912 outputString = candidateSeq->getName() + "\t" + classify->getSimpleTax() + "\n";
913 length = outputString.length();
914 char* buf = new char[length];
915 memcpy(buf, outputString.c_str(), length);
917 MPI_File_write_shared(tempFile, buf, length, MPI_CHAR, &statusTemp);
923 if((i+1) % 100 == 0){ cout << "Classifying sequence " << (i+1) << endl; }
926 if(num % 100 != 0){ cout << "Classifying sequence " << (num) << endl; }
931 catch(exception& e) {
932 m->errorOut(e, "ClassifySeqsCommand", "driverMPI");
937 //**********************************************************************************************************************
938 int ClassifySeqsCommand::MPIReadNamesFile(string nameFilename){
941 nameMap.clear(); //remove old names
947 //char* inFileName = new char[nameFilename.length()];
948 //memcpy(inFileName, nameFilename.c_str(), nameFilename.length());
950 char inFileName[1024];
951 strcpy(inFileName, nameFilename.c_str());
953 MPI_File_open(MPI_COMM_WORLD, inFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI);
954 MPI_File_get_size(inMPI, &size);
957 char* buffer = new char[size];
958 MPI_File_read(inMPI, buffer, size, MPI_CHAR, &status);
960 string tempBuf = buffer;
961 if (tempBuf.length() > size) { tempBuf = tempBuf.substr(0, size); }
962 istringstream iss (tempBuf,istringstream::in);
965 string firstCol, secondCol;
967 iss >> firstCol >> secondCol; m->gobble(iss);
970 m->splitAtComma(secondCol, temp);
972 nameMap[firstCol] = temp;
975 MPI_File_close(&inMPI);
979 catch(exception& e) {
980 m->errorOut(e, "ClassifySeqsCommand", "MPIReadNamesFile");
985 /**************************************************************************************************/