X-Git-Url: https://git.donarmstrong.com/?a=blobdiff_plain;f=filterseqscommand.cpp;h=7ab00a7822f7a7fc7ddd1ef0de00d2d24b8c51de;hb=9013e13ecfb2fda3c2664a76f76cc99b8c7fa74c;hp=b64e2f530ff91af78cc69dea71b9d2f53f3f981b;hpb=a61980e36d61937acccadc41be260165a1e389f4;p=mothur.git diff --git a/filterseqscommand.cpp b/filterseqscommand.cpp index b64e2f5..7ab00a7 100644 --- a/filterseqscommand.cpp +++ b/filterseqscommand.cpp @@ -28,7 +28,7 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { OptionParser parser(option); map parameters = parser.getParameters(); - ValidParameters validParameter; + ValidParameters validParameter("filter.seqs"); map::iterator it; //check to make sure all parameters are valid for command @@ -72,11 +72,22 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { if (path == "") { fastafileNames[i] = inputDir + fastafileNames[i]; } } - int ableToOpen; ifstream in; - ableToOpen = openInputFile(fastafileNames[i], in); + int ableToOpen = openInputFile(fastafileNames[i], in, "noerror"); + + //if you can't open it, try default location + if (ableToOpen == 1) { + if (m->getDefaultPath() != "") { //default path is set + string tryPath = m->getDefaultPath() + getSimpleName(fastafileNames[i]); + m->mothurOut("Unable to open " + fastafileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine(); + ableToOpen = openInputFile(tryPath, in, "noerror"); + fastafileNames[i] = tryPath; + } + } + in.close(); + if (ableToOpen == 1) { - m->mothurOut(fastafileNames[i] + " will be disregarded."); m->mothurOutEndLine(); + m->mothurOut("Unable to open " + fastafileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); //erase from file list fastafileNames.erase(fastafileNames.begin()+i); i--; @@ -102,6 +113,9 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { // ...at some point should added some additional type checking... string temp; + hard = validParameter.validFile(parameters, "hard", true); if (hard == "not found") { hard = ""; } + else if (hard == "not open") { hard = ""; abort = true; } + temp = validParameter.validFile(parameters, "trump", false); if (temp == "not found") { temp = "*"; } trump = temp[0]; @@ -111,13 +125,13 @@ FilterSeqsCommand::FilterSeqsCommand(string option) { temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = "1"; } convert(temp, processors); - hard = validParameter.validFile(parameters, "hard", true); if (hard == "not found") { hard = ""; } - else if (hard == "not open") { abort = true; } - - vertical = validParameter.validFile(parameters, "vertical", false); if (vertical == "not found") { vertical = "T"; } + vertical = validParameter.validFile(parameters, "vertical", false); + if (vertical == "not found") { + if ((hard == "") && (trump == '*') && (soft == 0)) { vertical = "T"; } //you have not given a hard file or set the trump char. + else { vertical = "F"; } + } numSeqs = 0; - } } @@ -138,11 +152,11 @@ void FilterSeqsCommand::help(){ m->mothurOut("For example: fasta=abrecovery.fasta-amazon.fasta \n"); m->mothurOut("The trump parameter .... The default is ...\n"); m->mothurOut("The soft parameter .... The default is ....\n"); - m->mothurOut("The hard parameter .... The default is ....\n"); - m->mothurOut("The vertical parameter .... The default is T.\n"); + m->mothurOut("The hard parameter allows you to enter a file containing the filter you want to use.\n"); + m->mothurOut("The vertical parameter removes columns where all sequences contain a gap character. The default is T.\n"); m->mothurOut("The filter.seqs command should be in the following format: \n"); - m->mothurOut("filter.seqs(fasta=yourFastaFile, trump=yourTrump, soft=yourSoft, hard=yourHard, vertical=yourVertical) \n"); - m->mothurOut("Example filter.seqs(fasta=abrecovery.fasta, trump=..., soft=..., hard=..., vertical=T).\n"); + m->mothurOut("filter.seqs(fasta=yourFastaFile, trump=yourTrump) \n"); + m->mothurOut("Example filter.seqs(fasta=abrecovery.fasta, trump=.).\n"); m->mothurOut("Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFasta).\n\n"); } @@ -244,35 +258,26 @@ int FilterSeqsCommand::filterSequences() { #ifdef USE_MPI int pid, start, end, numSeqsPerProcessor, num; int tag = 2001; - vectorMPIPos; + vectorMPIPos; MPI_Status status; MPI_Comm_size(MPI_COMM_WORLD, &processors); //set processors to the number of mpi processes running MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are - + cout << pid << "is in create filter " << endl; MPI_File outMPI; MPI_File tempMPI; MPI_File inMPI; int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY; int inMode=MPI_MODE_RDONLY; - //char* outFilename = new char[filteredFasta.length()]; - //memcpy(outFilename, filteredFasta.c_str(), filteredFasta.length()); - char outFilename[1024]; strcpy(outFilename, filteredFasta.c_str()); - - //char* inFileName = new char[fastafileNames[s].length()]; - //memcpy(inFileName, fastafileNames[s].c_str(), fastafileNames[s].length()); - + char inFileName[1024]; strcpy(inFileName, fastafileNames[s].c_str()); MPI_File_open(MPI_COMM_WORLD, inFileName, inMode, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer MPI_File_open(MPI_COMM_WORLD, outFilename, outMode, MPI_INFO_NULL, &outMPI); - - //delete inFileName; - //delete outFilename; if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPI); return 0; } @@ -282,13 +287,16 @@ int FilterSeqsCommand::filterSequences() { numSeqs += num; //send file positions to all processes - MPI_Bcast(&num, 1, MPI_INT, 0, MPI_COMM_WORLD); //send numSeqs - MPI_Bcast(&MPIPos[0], (num+1), MPI_LONG, 0, MPI_COMM_WORLD); //send file pos + for(int i = 1; i < processors; i++) { + MPI_Send(&num, 1, MPI_INT, i, tag, MPI_COMM_WORLD); + MPI_Send(&MPIPos[0], (num+1), MPI_LONG, i, tag, MPI_COMM_WORLD); + } //figure out how many sequences you have to do numSeqsPerProcessor = num / processors; - if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } int startIndex = pid * numSeqsPerProcessor; + if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } + //do your part driverMPIRun(startIndex, numSeqsPerProcessor, inMPI, outMPI, MPIPos); @@ -302,15 +310,16 @@ int FilterSeqsCommand::filterSequences() { } }else { //you are a child process - MPI_Bcast(&num, 1, MPI_INT, 0, MPI_COMM_WORLD); //get numSeqs - numSeqs += num; + MPI_Recv(&num, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status); MPIPos.resize(num+1); - MPI_Bcast(&MPIPos[0], (num+1), MPI_LONG, 0, MPI_COMM_WORLD); //get file positions + numSeqs += num; + MPI_Recv(&MPIPos[0], (num+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status); //figure out how many sequences you have to align numSeqsPerProcessor = num / processors; - if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } int startIndex = pid * numSeqsPerProcessor; + if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } + //align your part driverMPIRun(startIndex, numSeqsPerProcessor, inMPI, outMPI, MPIPos); @@ -326,17 +335,21 @@ int FilterSeqsCommand::filterSequences() { MPI_File_close(&outMPI); MPI_File_close(&inMPI); + MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case #else #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) if(processors == 1){ ifstream inFASTA; + int numFastaSeqs; openInputFile(fastafileNames[s], inFASTA); - int numFastaSeqs=count(istreambuf_iterator(inFASTA),istreambuf_iterator(), '>'); + getNumSeqs(inFASTA, numFastaSeqs); inFASTA.close(); lines.push_back(new linePair(0, numFastaSeqs)); + numSeqs += numFastaSeqs; + driverRunFilter(filter, filteredFasta, fastafileNames[s], lines[0]); }else{ setLines(fastafileNames[s]); @@ -354,12 +367,15 @@ int FilterSeqsCommand::filterSequences() { if (m->control_pressed) { return 1; } #else ifstream inFASTA; + int numFastaSeqs; openInputFile(fastafileNames[s], inFASTA); - int numFastaSeqs=count(istreambuf_iterator(inFASTA),istreambuf_iterator(), '>'); + getNumSeqs(inFASTA, numFastaSeqs); inFASTA.close(); lines.push_back(new linePair(0, numFastaSeqs)); + numSeqs += numFastaSeqs; + driverRunFilter(filter, filteredFasta, fastafileNames[s], lines[0]); if (m->control_pressed) { return 1; } @@ -377,7 +393,7 @@ int FilterSeqsCommand::filterSequences() { } #ifdef USE_MPI /**************************************************************************************/ -int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_File& outMPI, vector& MPIPos) { +int FilterSeqsCommand::driverMPIRun(int start, int num, MPI_File& inMPI, MPI_File& outMPI, vector& MPIPos) { try { string outputString = ""; int count = 0; @@ -537,7 +553,7 @@ string FilterSeqsCommand::createFilter() { F.setLength(alignmentLength); - if(soft != 0 || isTrue(vertical)){ + if(trump != '*' || isTrue(vertical) || soft != 0){ F.initialize(); } @@ -553,7 +569,7 @@ string FilterSeqsCommand::createFilter() { #ifdef USE_MPI int pid, numSeqsPerProcessor, num; int tag = 2001; - vector MPIPos; + vector MPIPos; MPI_Status status; MPI_File inMPI; @@ -575,13 +591,16 @@ string FilterSeqsCommand::createFilter() { numSeqs += num; //send file positions to all processes - MPI_Bcast(&num, 1, MPI_INT, 0, MPI_COMM_WORLD); //send numSeqs - MPI_Bcast(&MPIPos[0], (num+1), MPI_LONG, 0, MPI_COMM_WORLD); //send file pos + for(int i = 1; i < processors; i++) { + MPI_Send(&num, 1, MPI_INT, i, tag, MPI_COMM_WORLD); + MPI_Send(&MPIPos[0], (num+1), MPI_LONG, i, tag, MPI_COMM_WORLD); + } //figure out how many sequences you have to do numSeqsPerProcessor = num / processors; - if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } int startIndex = pid * numSeqsPerProcessor; + if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } + //do your part MPICreateFilter(startIndex, numSeqsPerProcessor, F, inMPI, MPIPos); @@ -589,16 +608,16 @@ string FilterSeqsCommand::createFilter() { if (m->control_pressed) { MPI_File_close(&inMPI); return 0; } }else { //i am the child process - - MPI_Bcast(&num, 1, MPI_INT, 0, MPI_COMM_WORLD); //get numSeqs + MPI_Recv(&num, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status); MPIPos.resize(num+1); numSeqs += num; - MPI_Bcast(&MPIPos[0], (num+1), MPI_LONG, 0, MPI_COMM_WORLD); //get file positions + MPI_Recv(&MPIPos[0], (num+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status); //figure out how many sequences you have to align numSeqsPerProcessor = num / processors; - if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } int startIndex = pid * numSeqsPerProcessor; + if(pid == (processors - 1)){ numSeqsPerProcessor = num - pid * numSeqsPerProcessor; } + //do your part MPICreateFilter(startIndex, numSeqsPerProcessor, F, inMPI, MPIPos); @@ -607,13 +626,15 @@ string FilterSeqsCommand::createFilter() { } MPI_File_close(&inMPI); + MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case #else #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) if(processors == 1){ ifstream inFASTA; + int numFastaSeqs; openInputFile(fastafileNames[s], inFASTA); - int numFastaSeqs=count(istreambuf_iterator(inFASTA),istreambuf_iterator(), '>'); + getNumSeqs(inFASTA, numFastaSeqs); inFASTA.close(); numSeqs += numFastaSeqs; @@ -629,8 +650,9 @@ string FilterSeqsCommand::createFilter() { if (m->control_pressed) { return filterString; } #else ifstream inFASTA; + int numFastaSeqs; openInputFile(fastafileNames[s], inFASTA); - int numFastaSeqs=count(istreambuf_iterator(inFASTA),istreambuf_iterator(), '>'); + getNumSeqs(inFASTA, numFastaSeqs); inFASTA.close(); numSeqs += numFastaSeqs; @@ -660,20 +682,24 @@ string FilterSeqsCommand::createFilter() { vector temp; temp.resize(alignmentLength+1); //get the frequencies from the child processes - for(int i = 0; i < ((processors-1)*5); i++) { - MPI_Recv(&temp[0], (alignmentLength+1), MPI_INT, MPI_ANY_SOURCE, 2001, MPI_COMM_WORLD, &status); - int receiveTag = temp[temp.size()-1]; //child process added a int to the end to indicate what letter count this is for - - if (receiveTag == Atag) { //you are recieveing the A frequencies - for (int k = 0; k < alignmentLength; k++) { F.a[k] += temp[k]; } - }else if (receiveTag == Ttag) { //you are recieveing the T frequencies - for (int k = 0; k < alignmentLength; k++) { F.t[k] += temp[k]; } - }else if (receiveTag == Ctag) { //you are recieveing the C frequencies - for (int k = 0; k < alignmentLength; k++) { F.c[k] += temp[k]; } - }else if (receiveTag == Gtag) { //you are recieveing the G frequencies - for (int k = 0; k < alignmentLength; k++) { F.g[k] += temp[k]; } - }else if (receiveTag == Gaptag) { //you are recieveing the gap frequencies - for (int k = 0; k < alignmentLength; k++) { F.gap[k] += temp[k]; } + for(int i = 1; i < processors; i++) { + + for (int j = 0; j < 5; j++) { + + MPI_Recv(&temp[0], (alignmentLength+1), MPI_INT, i, 2001, MPI_COMM_WORLD, &status); + int receiveTag = temp[temp.size()-1]; //child process added a int to the end to indicate what letter count this is for + + if (receiveTag == Atag) { //you are recieveing the A frequencies + for (int k = 0; k < alignmentLength; k++) { F.a[k] += temp[k]; } + }else if (receiveTag == Ttag) { //you are recieveing the T frequencies + for (int k = 0; k < alignmentLength; k++) { F.t[k] += temp[k]; } + }else if (receiveTag == Ctag) { //you are recieveing the C frequencies + for (int k = 0; k < alignmentLength; k++) { F.c[k] += temp[k]; } + }else if (receiveTag == Gtag) { //you are recieveing the G frequencies + for (int k = 0; k < alignmentLength; k++) { F.g[k] += temp[k]; } + }else if (receiveTag == Gaptag) { //you are recieveing the gap frequencies + for (int k = 0; k < alignmentLength; k++) { F.gap[k] += temp[k]; } + } } } }else{ @@ -693,6 +719,8 @@ string FilterSeqsCommand::createFilter() { } + MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case + if (pid == 0) { //only one process should output the filter #endif F.setNumSeqs(numSeqs); @@ -704,10 +732,14 @@ string FilterSeqsCommand::createFilter() { #ifdef USE_MPI //send filter string to kids - MPI_Bcast(&filterString[0], alignmentLength, MPI_CHAR, 0, MPI_COMM_WORLD); + //for(int i = 1; i < processors; i++) { + // MPI_Send(&filterString[0], alignmentLength, MPI_CHAR, i, 2001, MPI_COMM_WORLD); + //} + MPI_Bcast(&filterString[0], alignmentLength, MPI_CHAR, 0, MPI_COMM_WORLD); }else{ //recieve filterString char* tempBuf = new char[alignmentLength]; + //MPI_Recv(&tempBuf[0], alignmentLength, MPI_CHAR, 0, 2001, MPI_COMM_WORLD, &status); MPI_Bcast(tempBuf, alignmentLength, MPI_CHAR, 0, MPI_COMM_WORLD); filterString = tempBuf; @@ -717,8 +749,7 @@ string FilterSeqsCommand::createFilter() { MPI_Barrier(MPI_COMM_WORLD); #endif - - + return filterString; } catch(exception& e) { @@ -766,7 +797,7 @@ int FilterSeqsCommand::driverCreateFilter(Filters& F, string filename, linePair* } #ifdef USE_MPI /**************************************************************************************/ -int FilterSeqsCommand::MPICreateFilter(int start, int num, Filters& F, MPI_File& inMPI, vector& MPIPos) { +int FilterSeqsCommand::MPICreateFilter(int start, int num, Filters& F, MPI_File& inMPI, vector& MPIPos) { try { MPI_Status status; @@ -852,7 +883,7 @@ int FilterSeqsCommand::createProcessesCreateFilter(Filters& F, string filename) int FilterSeqsCommand::setLines(string filename) { try { - vector positions; + vector positions; bufferSizes.clear(); ifstream inFASTA; @@ -863,7 +894,7 @@ int FilterSeqsCommand::setLines(string filename) { input = getline(inFASTA); if (input.length() != 0) { - if(input[0] == '>'){ long int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); } + if(input[0] == '>'){ unsigned long int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); } } } inFASTA.close(); @@ -871,7 +902,7 @@ int FilterSeqsCommand::setLines(string filename) { int numFastaSeqs = positions.size(); FILE * pFile; - long size; + unsigned long int size; //get num bytes in file pFile = fopen (filename.c_str(),"rb"); @@ -888,12 +919,12 @@ int FilterSeqsCommand::setLines(string filename) { for (int i = 0; i < processors; i++) { - long int startPos = positions[ i * numSeqsPerProcessor ]; + unsigned long int startPos = positions[ i * numSeqsPerProcessor ]; if(i == processors - 1){ numSeqsPerProcessor = numFastaSeqs - i * numSeqsPerProcessor; bufferSizes.push_back(size - startPos); }else{ - long int myEnd = positions[ (i+1) * numSeqsPerProcessor ]; + unsigned long int myEnd = positions[ (i+1) * numSeqsPerProcessor ]; bufferSizes.push_back(myEnd-startPos); } lines.push_back(new linePair(startPos, numSeqsPerProcessor));