X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=classify.cpp;h=eb0865c2744acfe120505a5a533cd482af8b9b19;hb=ae57e166b2ed7b475ec3f466106bd76fabadd063;hp=2db19735a8726e78269fbcce2a58521aa2205b19;hpb=74844a60d80c6dd06e3fb02ee9b928424f9019b0;p=mothur.git diff --git a/classify.cpp b/classify.cpp index 2db1973..eb0865c 100644 --- a/classify.cpp +++ b/classify.cpp @@ -13,100 +13,300 @@ #include "suffixdb.hpp" #include "blastdb.hpp" #include "distancedb.hpp" +#include "referencedb.h" /**************************************************************************************************/ -Classify::Classify(string tfile, string tempFile, string method, int kmerSize, float gapOpen, float gapExtend, float match, float misMatch) : taxFile(tfile), templateFile(tempFile) { +void Classify::generateDatabaseAndNames(string tfile, string tempFile, string method, int kmerSize, float gapOpen, float gapExtend, float match, float misMatch) { try { - m = MothurOut::getInstance(); - readTaxonomy(taxFile); + ReferenceDB* rdb = ReferenceDB::getInstance(); + + if (tfile == "saved") { tfile = rdb->getSavedTaxonomy(); } - int start = time(NULL); + taxFile = tfile; + readTaxonomy(taxFile); int numSeqs = 0; - //need to know number of template seqs for suffixdb - if (method == "suffix") { - ifstream inFASTA; - openInputFile(tempFile, inFASTA); - numSeqs = count(istreambuf_iterator(inFASTA),istreambuf_iterator(), '>'); - inFASTA.close(); - } + + if (tempFile == "saved") { + int start = time(NULL); + m->mothurOutEndLine(); m->mothurOut("Using sequences from " + rdb->getSavedReference() + " that are saved in memory."); m->mothurOutEndLine(); + + numSeqs = rdb->referenceSeqs.size(); + templateFile = rdb->getSavedReference(); + tempFile = rdb->getSavedReference(); + + bool needToGenerate = true; + string kmerDBName; + if(method == "kmer") { + database = new KmerDB(tempFile, kmerSize); + + kmerDBName = tempFile.substr(0,tempFile.find_last_of(".")+1) + char('0'+ kmerSize) + "mer"; + ifstream kmerFileTest(kmerDBName.c_str()); + if(kmerFileTest){ + bool GoodFile = m->checkReleaseVersion(kmerFileTest, m->getVersion()); + if (GoodFile) { needToGenerate = false; } + } + } + else if(method == "suffix") { database = new SuffixDB(numSeqs); } + else if(method == "blast") { database = new BlastDB(tempFile.substr(0,tempFile.find_last_of(".")+1), gapOpen, gapExtend, match, misMatch, "", threadID); } + else if(method == "distance") { database = new DistanceDB(); } + else { + m->mothurOut(method + " is not a valid search option. I will run the command using kmer, ksize=8."); + m->mothurOutEndLine(); + database = new KmerDB(tempFile, 8); + } + + if (needToGenerate) { + for (int k = 0; k < rdb->referenceSeqs.size(); k++) { + Sequence temp(rdb->referenceSeqs[k].getName(), rdb->referenceSeqs[k].getAligned()); + names.push_back(temp.getName()); + database->addSequence(temp); + } + database->generateDB(); + }else if ((method == "kmer") && (!needToGenerate)) { + ifstream kmerFileTest(kmerDBName.c_str()); + database->readKmerDB(kmerFileTest); + + for (int k = 0; k < rdb->referenceSeqs.size(); k++) { + names.push_back(rdb->referenceSeqs[k].getName()); + } + } + + database->setNumSeqs(numSeqs); + + //sanity check + bool okay = phyloTree->ErrorCheck(names); + + if (!okay) { m->control_pressed = true; } + + m->mothurOut("It took " + toString(time(NULL) - start) + " to load " + toString(rdb->referenceSeqs.size()) + " sequences and generate the search databases.");m->mothurOutEndLine(); + + }else { + + templateFile = tempFile; + + int start = time(NULL); + + m->mothurOut("Generating search database... "); cout.flush(); + #ifdef USE_MPI + int pid, processors; + vector positions; + int tag = 2001; + + MPI_Status status; + MPI_File inMPI; + MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are + MPI_Comm_size(MPI_COMM_WORLD, &processors); - m->mothurOut("Generating search database... "); cout.flush(); + //char* inFileName = new char[tempFile.length()]; + //memcpy(inFileName, tempFile.c_str(), tempFile.length()); - bool needToGenerate = true; - string kmerDBName; - if(method == "kmer") { - database = new KmerDB(tempFile, kmerSize); - - kmerDBName = tempFile.substr(0,tempFile.find_last_of(".")+1) + char('0'+ kmerSize) + "mer"; - ifstream kmerFileTest(kmerDBName.c_str()); - if(kmerFileTest){ needToGenerate = false; } - } - else if(method == "suffix") { database = new SuffixDB(numSeqs); } - else if(method == "blast") { database = new BlastDB(gapOpen, gapExtend, match, misMatch); } - else if(method == "distance") { database = new DistanceDB(); } - else { - m->mothurOut(method + " is not a valid search option. I will run the command using kmer, ksize=8."); - m->mothurOutEndLine(); - database = new KmerDB(tempFile, 8); - } - - if (needToGenerate) { - ifstream fastaFile; - openInputFile(tempFile, fastaFile); + char inFileName[1024]; + strcpy(inFileName, tempFile.c_str()); + + MPI_File_open(MPI_COMM_WORLD, inFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer + //delete inFileName; + + if (pid == 0) { //only one process needs to scan file + positions = m->setFilePosFasta(tempFile, numSeqs); //fills MPIPos, returns numSeqs + + //send file positions to all processes + for(int i = 1; i < processors; i++) { + MPI_Send(&numSeqs, 1, MPI_INT, i, tag, MPI_COMM_WORLD); + MPI_Send(&positions[0], (numSeqs+1), MPI_LONG, i, tag, MPI_COMM_WORLD); + } + }else{ + MPI_Recv(&numSeqs, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status); + positions.resize(numSeqs+1); + MPI_Recv(&positions[0], (numSeqs+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status); + } + + //create database + if(method == "kmer") { database = new KmerDB(tempFile, kmerSize); } + else if(method == "suffix") { database = new SuffixDB(numSeqs); } + else if(method == "blast") { database = new BlastDB(tempFile.substr(0,tempFile.find_last_of(".")+1), gapOpen, gapExtend, match, misMatch, ""); } + else if(method == "distance") { database = new DistanceDB(); } + else { + m->mothurOut(method + " is not a valid search option. I will run the command using kmer, ksize=8."); m->mothurOutEndLine(); + database = new KmerDB(tempFile, 8); + } + + //read file + for(int i=0;i length) { tempBuf = tempBuf.substr(0, length); } + delete buf4; + istringstream iss (tempBuf,istringstream::in); + + Sequence temp(iss); + if (temp.getName() != "") { + if (rdb->save) { rdb->referenceSeqs.push_back(temp); } + names.push_back(temp.getName()); + database->addSequence(temp); + } + } + + database->generateDB(); + MPI_File_close(&inMPI); + MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case + #else - while (!fastaFile.eof()) { - Sequence temp(fastaFile); - gobble(fastaFile); + //need to know number of template seqs for suffixdb + if (method == "suffix") { + ifstream inFASTA; + m->openInputFile(tempFile, inFASTA); + m->getNumSeqs(inFASTA, numSeqs); + inFASTA.close(); + } + + bool needToGenerate = true; + string kmerDBName; + if(method == "kmer") { + database = new KmerDB(tempFile, kmerSize); + + kmerDBName = tempFile.substr(0,tempFile.find_last_of(".")+1) + char('0'+ kmerSize) + "mer"; + ifstream kmerFileTest(kmerDBName.c_str()); + if(kmerFileTest){ + bool GoodFile = m->checkReleaseVersion(kmerFileTest, m->getVersion()); + if (GoodFile) { needToGenerate = false; } + } + } + else if(method == "suffix") { database = new SuffixDB(numSeqs); } + else if(method == "blast") { database = new BlastDB(tempFile.substr(0,tempFile.find_last_of(".")+1), gapOpen, gapExtend, match, misMatch, "", threadID); } + else if(method == "distance") { database = new DistanceDB(); } + else { + m->mothurOut(method + " is not a valid search option. I will run the command using kmer, ksize=8."); + m->mothurOutEndLine(); + database = new KmerDB(tempFile, 8); + } - names.push_back(temp.getName()); + if (needToGenerate) { + ifstream fastaFile; + m->openInputFile(tempFile, fastaFile); + + while (!fastaFile.eof()) { + Sequence temp(fastaFile); + m->gobble(fastaFile); + + if (rdb->save) { rdb->referenceSeqs.push_back(temp); } + + names.push_back(temp.getName()); - database->addSequence(temp); - } - fastaFile.close(); + database->addSequence(temp); + } + fastaFile.close(); - database->generateDB(); + database->generateDB(); + + }else if ((method == "kmer") && (!needToGenerate)) { + ifstream kmerFileTest(kmerDBName.c_str()); + database->readKmerDB(kmerFileTest); - }else if ((method == "kmer") && (!needToGenerate)) { - ifstream kmerFileTest(kmerDBName.c_str()); - database->readKmerDB(kmerFileTest); + ifstream fastaFile; + m->openInputFile(tempFile, fastaFile); + + while (!fastaFile.eof()) { + Sequence temp(fastaFile); + m->gobble(fastaFile); + + if (rdb->save) { rdb->referenceSeqs.push_back(temp); } + names.push_back(temp.getName()); + } + fastaFile.close(); + } + #endif + + database->setNumSeqs(names.size()); - ifstream fastaFile; - openInputFile(tempFile, fastaFile); + //sanity check + bool okay = phyloTree->ErrorCheck(names); - while (!fastaFile.eof()) { - Sequence temp(fastaFile); - gobble(fastaFile); + if (!okay) { m->control_pressed = true; } - names.push_back(temp.getName()); - } - fastaFile.close(); + m->mothurOut("DONE."); m->mothurOutEndLine(); + m->mothurOut("It took " + toString(time(NULL) - start) + " seconds generate search database. "); m->mothurOutEndLine(); } - - database->setNumSeqs(names.size()); - - m->mothurOut("DONE."); m->mothurOutEndLine(); - m->mothurOut("It took " + toString(time(NULL) - start) + " seconds generate search database. "); m->mothurOutEndLine(); } catch(exception& e) { - m->errorOut(e, "Classify", "Classify"); + m->errorOut(e, "Classify", "generateDatabaseAndNames"); exit(1); } } /**************************************************************************************************/ +Classify::Classify() { m = MothurOut::getInstance(); database = NULL; } +/**************************************************************************************************/ -void Classify::readTaxonomy(string file) { +int Classify::readTaxonomy(string file) { try { phyloTree = new PhyloTree(); + string name, taxInfo; - ifstream inTax; - openInputFile(file, inTax); - m->mothurOutEndLine(); m->mothurOut("Reading in the " + file + " taxonomy...\t"); cout.flush(); + +#ifdef USE_MPI + int pid, num, processors; + vector positions; + int tag = 2001; - string name, taxInfo; + MPI_Status status; + MPI_File inMPI; + MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are + MPI_Comm_size(MPI_COMM_WORLD, &processors); + + //char* inFileName = new char[file.length()]; + //memcpy(inFileName, file.c_str(), file.length()); + + char inFileName[1024]; + strcpy(inFileName, file.c_str()); + + MPI_File_open(MPI_COMM_WORLD, inFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer + //delete inFileName; + + if (pid == 0) { + positions = m->setFilePosEachLine(file, num); + + //send file positions to all processes + for(int i = 1; i < processors; i++) { + MPI_Send(&num, 1, MPI_INT, i, tag, MPI_COMM_WORLD); + MPI_Send(&positions[0], (num+1), MPI_LONG, i, tag, MPI_COMM_WORLD); + } + }else{ + MPI_Recv(&num, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status); + positions.resize(num+1); + MPI_Recv(&positions[0], (num+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status); + } + + //read file + for(int i=0;i length) { tempBuf = tempBuf.substr(0, length); } + delete buf4; + + istringstream iss (tempBuf,istringstream::in); + iss >> name >> taxInfo; + taxonomy[name] = taxInfo; + phyloTree->addSeqToTree(name, taxInfo); + } + + MPI_File_close(&inMPI); + MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case +#else + ifstream inTax; + m->openInputFile(file, inTax); + //read template seqs and save while (!inTax.eof()) { inTax >> name >> taxInfo; @@ -115,14 +315,19 @@ void Classify::readTaxonomy(string file) { phyloTree->addSeqToTree(name, taxInfo); - gobble(inTax); + m->gobble(inTax); } - - phyloTree->assignHeirarchyIDs(0); inTax.close(); +#endif + + phyloTree->assignHeirarchyIDs(0); + + phyloTree->setUp(file); m->mothurOut("DONE."); m->mothurOutEndLine(); cout.flush(); + + return phyloTree->getNumSeqs(); } catch(exception& e) {