X-Git-Url: https://git.donarmstrong.com/?p=mothur.git;a=blobdiff_plain;f=chimera.cpp;h=53e77322f52ba40eb478e0b385a84982de752d6b;hp=83d40fe8342558eacb598525862364dba0f7e5ca;hb=cf9987b67aa49777a4c91c2d21f96e58bf17aa82;hpb=813f0516e9c7533859f57ebe6d7b8854896bf009 diff --git a/chimera.cpp b/chimera.cpp index 83d40fe..53e7732 100644 --- a/chimera.cpp +++ b/chimera.cpp @@ -8,6 +8,7 @@ */ #include "chimera.h" +#include "referencedb.h" //*************************************************************************************************************** //this is a vertical soft filter @@ -31,6 +32,8 @@ string Chimera::createFilter(vector seqs, float t) { if (m->control_pressed) { return filterString; } string seqAligned = seqs[i]->getAligned(); + + if (seqAligned.length() != filterString.length()) { m->mothurOut(seqs[i]->getName() + " is not the same length as the template sequences. Aborting!\n"); exit(1); } for (int j = 0; j < seqAligned.length(); j++) { //if this spot is a gap @@ -92,89 +95,163 @@ map Chimera::runFilter(Sequence* seq) { //*************************************************************************************************************** vector Chimera::readSeqs(string file) { try { - + vector container; int count = 0; length = 0; unaligned = false; - - m->mothurOut("Reading sequences from " + file + "..."); cout.flush(); + ReferenceDB* rdb = ReferenceDB::getInstance(); - #ifdef USE_MPI - int pid; - vector positions; - int numSeqs; - - MPI_Status status; - MPI_File inMPI; - MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are - - char* inFileName = new char[file.length()]; - memcpy(inFileName, file.c_str(), file.length()); - - 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 = setFilePosFasta(file, numSeqs); //fills MPIPos, returns numSeqs - - //send file positions to all processes - MPI_Bcast(&numSeqs, 1, MPI_INT, 0, MPI_COMM_WORLD); //send numSeqs - MPI_Bcast(&positions[0], (numSeqs+1), MPI_LONG, 0, MPI_COMM_WORLD); //send file pos - }else{ - MPI_Bcast(&numSeqs, 1, MPI_INT, 0, MPI_COMM_WORLD); //get numSeqs - positions.resize(numSeqs+1); - MPI_Bcast(&positions[0], (numSeqs+1), MPI_LONG, 0, MPI_COMM_WORLD); //get file positions - } + if (file == "saved") { - //read file - for(int i=0;icontrol_pressed) { MPI_File_close(&inMPI); return container; } - - //read next sequence - int seqlength = positions[i+1] - positions[i]; - char* buf4 = new char[seqlength]; - - MPI_File_read_at(inMPI, positions[i], buf4, seqlength, MPI_CHAR, &status); + m->mothurOutEndLine(); m->mothurOut("Using sequences from " + rdb->getSavedReference() + " that are saved in memory."); m->mothurOutEndLine(); + + for (int i = 0; i < rdb->referenceSeqs.size(); i++) { + Sequence* temp = new Sequence(rdb->referenceSeqs[i].getName(), rdb->referenceSeqs[i].getAligned()); + + if (count == 0) { length = temp->getAligned().length(); count++; } //gets first seqs length + else if (length != temp->getAligned().length()) { unaligned = true; } - string tempBuf = buf4; - if (tempBuf.length() > seqlength) { tempBuf = tempBuf.substr(0, seqlength); } - delete buf4; + if (temp->getName() != "") { container.push_back(temp); } + } + + templateFileName = rdb->getSavedReference(); + + }else { + + m->mothurOut("Reading sequences from " + file + "..."); cout.flush(); + + #ifdef USE_MPI + int pid, processors; + vector positions; + int numSeqs; + int tag = 2001; + MPI_File inMPI; + MPI_Status status; + + if (byGroup) { + char inFileName[1024]; + strcpy(inFileName, file.c_str()); + + MPI_File_open(MPI_COMM_SELF, inFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer + + positions = m->setFilePosFasta(file, numSeqs); //fills MPIPos, returns numSeqs + + //read file + for(int i=0;icontrol_pressed) { MPI_File_close(&inMPI); return container; } + + //read next sequence + int seqlength = positions[i+1] - positions[i]; + char* buf4 = new char[seqlength]; + + MPI_File_read_at(inMPI, positions[i], buf4, seqlength, MPI_CHAR, &status); + + string tempBuf = buf4; + if (tempBuf.length() > seqlength) { tempBuf = tempBuf.substr(0, seqlength); } + delete buf4; + + istringstream iss (tempBuf,istringstream::in); + + Sequence* current = new Sequence(iss); + if (current->getName() != "") { + if (count == 0) { length = current->getAligned().length(); count++; } //gets first seqs length + else if (length != current->getAligned().length()) { unaligned = true; } + + container.push_back(current); + if (rdb->save) { rdb->referenceSeqs.push_back(*current); } + } + } + + MPI_File_close(&inMPI); + + }else { + + 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->setFilePosFasta(file, 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); + } + + //read file + for(int i=0;icontrol_pressed) { MPI_File_close(&inMPI); return container; } + + //read next sequence + int seqlength = positions[i+1] - positions[i]; + char* buf4 = new char[seqlength]; + + MPI_File_read_at(inMPI, positions[i], buf4, seqlength, MPI_CHAR, &status); + + string tempBuf = buf4; + if (tempBuf.length() > seqlength) { tempBuf = tempBuf.substr(0, seqlength); } + delete buf4; + + istringstream iss (tempBuf,istringstream::in); + + Sequence* current = new Sequence(iss); + if (current->getName() != "") { + if (count == 0) { length = current->getAligned().length(); count++; } //gets first seqs length + else if (length != current->getAligned().length()) { unaligned = true; } + + container.push_back(current); + if (rdb->save) { rdb->referenceSeqs.push_back(*current); } + } + } + + MPI_File_close(&inMPI); + MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case + } + #else - istringstream iss (tempBuf,istringstream::in); - - Sequence* current = new Sequence(iss); - if (current->getName() != "") { - if (count == 0) { length = current->getAligned().length(); count++; } //gets first seqs length - else if (length != current->getAligned().length()) { unaligned = true; } + ifstream in; + m->openInputFile(file, in); + //read in seqs and store in vector + while(!in.eof()){ + + if (m->control_pressed) { return container; } + + Sequence* current = new Sequence(in); m->gobble(in); + + if (count == 0) { length = current->getAligned().length(); count++; } //gets first seqs length + else if (length != current->getAligned().length()) { unaligned = true; } + + if (current->getName() != "") { container.push_back(current); + if (rdb->save) { rdb->referenceSeqs.push_back(*current); } } } - - MPI_File_close(&inMPI); - #else - - ifstream in; - openInputFile(file, in); + in.close(); + #endif - //read in seqs and store in vector - while(!in.eof()){ + m->mothurOut("Done."); m->mothurOutEndLine(); - if (m->control_pressed) { return container; } - - Sequence* current = new Sequence(in); gobble(in); - - if (count == 0) { length = current->getAligned().length(); count++; } //gets first seqs length - else if (length != current->getAligned().length()) { unaligned = true; } - - if (current->getName() != "") { container.push_back(current); } + filterString = (string(container[0]->getAligned().length(), '1')); } - in.close(); - #endif - - m->mothurOut("Done."); m->mothurOutEndLine(); return container; } @@ -199,13 +276,16 @@ void Chimera::setMask(string filename) { MPI_Offset size; MPI_Status status; - char* inFileName = new char[filename.length()]; - memcpy(inFileName, filename.c_str(), filename.length()); + //char* inFileName = new char[filename.length()]; + //memcpy(inFileName, filename.c_str(), filename.length()); + + char inFileName[1024]; + strcpy(inFileName, filename.c_str()); MPI_File_open(MPI_COMM_WORLD, inFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer MPI_File_get_size(inMPI, &size); - delete inFileName; + //delete inFileName; char* buffer = new char[size]; MPI_File_read(inMPI, buffer, size, MPI_CHAR, &status); @@ -228,7 +308,7 @@ void Chimera::setMask(string filename) { #else ifstream infile; - openInputFile(filename, infile); + m->openInputFile(filename, infile); if (!infile.eof()) { Sequence temp(infile);