OptionParser parser(option);
map<string, string> parameters = parser.getParameters();
- ValidParameters validParameter;
+ ValidParameters validParameter("classify.seqs");
map<string, string>::iterator it;
//check to make sure all parameters are valid for command
int ableToOpen;
- #ifdef USE_MPI
- int pid;
- 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
-
- if (pid == 0) {
- #endif
-
ifstream in;
- ableToOpen = openInputFile(fastaFileNames[i], in);
- in.close();
-
- #ifdef USE_MPI
- for (int j = 1; j < processors; j++) {
- MPI_Send(&ableToOpen, 1, MPI_INT, j, 2001, MPI_COMM_WORLD);
- }
- }else{
- MPI_Status status;
- MPI_Recv(&ableToOpen, 1, MPI_INT, 0, 2001, MPI_COMM_WORLD, &status);
+ 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;
}
-
- #endif
+ }
+ 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--;
}
int ableToOpen;
- #ifdef USE_MPI
- int pid;
- 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
-
- if (pid == 0) {
- #endif
-
ifstream in;
- ableToOpen = openInputFile(namefileNames[i], in);
- in.close();
-
- #ifdef USE_MPI
- for (int j = 1; j < processors; j++) {
- MPI_Send(&ableToOpen, 1, MPI_INT, j, 2001, MPI_COMM_WORLD);
- }
- }else{
- MPI_Status status;
- MPI_Recv(&ableToOpen, 1, MPI_INT, 0, 2001, MPI_COMM_WORLD, &status);
+ ableToOpen = openInputFile(namefileNames[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(namefileNames[i]);
+ m->mothurOut("Unable to open " + namefileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine();
+ ableToOpen = openInputFile(tryPath, in, "noerror");
+ namefileNames[i] = tryPath;
}
-
- #endif
- if (ableToOpen == 1) { m->mothurOut("Unable to match name file with fasta file."); m->mothurOutEndLine(); abort = true; }
+ }
+ in.close();
+ if (ableToOpen == 1) {
+ m->mothurOut("Unable to open " + namefileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); abort = true;
+ //erase from file list
+ namefileNames.erase(namefileNames.begin()+i);
+ i--;
+ }
+
}
}
}
int ableToOpen;
- #ifdef USE_MPI
- int pid;
- 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
-
- if (pid == 0) {
- #endif
-
ifstream in;
- ableToOpen = openInputFile(groupfileNames[i], in);
- in.close();
-
- #ifdef USE_MPI
- for (int j = 1; j < processors; j++) {
- MPI_Send(&ableToOpen, 1, MPI_INT, j, 2001, MPI_COMM_WORLD);
- }
- }else{
- MPI_Status status;
- MPI_Recv(&ableToOpen, 1, MPI_INT, 0, 2001, MPI_COMM_WORLD, &status);
+ ableToOpen = openInputFile(groupfileNames[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(groupfileNames[i]);
+ m->mothurOut("Unable to open " + groupfileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine();
+ ableToOpen = openInputFile(tryPath, in, "noerror");
+ groupfileNames[i] = tryPath;
}
-
- #endif
- if (ableToOpen == 1) { m->mothurOut("Unable to match group file with fasta file, not using " + groupfileNames[i] + "."); m->mothurOutEndLine(); groupfileNames[i] = ""; }
+ }
+ in.close();
+ if (ableToOpen == 1) {
+ m->mothurOut("Unable to open " + groupfileNames[i] + ". It will be disregarded."); m->mothurOutEndLine(); groupfileNames[i] = "";
+ //erase from file list
+ groupfileNames.erase(groupfileNames.begin()+i);
+ i--;
+ }
}
}
m->mothurOut("The gapextend parameter allows you to specify the penalty for extending a gap in an alignment. The default is -1.0.\n");
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");
m->mothurOut("The cutoff parameter allows you to specify a bootstrap confidence threshold for your taxonomy. The default is 0.\n");
- m->mothurOut("The probs parameter shut off the bootstrapping results for the bayesian method. The default is true, meaning you want the bootstrapping to be run.\n");
+ 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");
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");
m->mothurOut("The classify.seqs command should be in the following format: \n");
m->mothurOut("classify.seqs(template=yourTemplateFile, fasta=yourFastaFile, method=yourClassificationMethod, search=yourSearchmethod, ksize=yourKmerSize, taxonomy=yourTaxonomyFile, processors=yourProcessors) \n");
m->mothurOut("Classifying sequences from " + fastaFileNames[s] + " ..." ); m->mothurOutEndLine();
+ string RippedTaxName = getRootName(getSimpleName(taxonomyFileName));
+ RippedTaxName = getExtension(RippedTaxName.substr(0, RippedTaxName.length()-1));
+ if (RippedTaxName[0] == '.') { RippedTaxName = RippedTaxName.substr(1, RippedTaxName.length()); }
+ RippedTaxName += ".";
+
if (outputDir == "") { outputDir += hasPath(fastaFileNames[s]); }
- string newTaxonomyFile = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + getRootName(getSimpleName(taxonomyFileName)) + "taxonomy";
+ string newTaxonomyFile = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + RippedTaxName + "taxonomy";
string tempTaxonomyFile = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + "taxonomy.temp";
- string taxSummary = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + getRootName(getSimpleName(taxonomyFileName)) + "tax.summary";
+ string taxSummary = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + RippedTaxName + "tax.summary";
+
+ if ((method == "knn") && (search == "distance")) {
+ string DistName = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + "match.dist";
+ classify->setDistName(DistName); outputNames.push_back(DistName);
+ }
outputNames.push_back(newTaxonomyFile);
outputNames.push_back(taxSummary);
#ifdef USE_MPI
int pid, end, numSeqsPerProcessor;
int tag = 2001;
- vector<long> MPIPos;
+ vector<unsigned long int> MPIPos;
MPI_Status status;
MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
//delete inFileName;
if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPINewTax); MPI_File_close(&outMPITempTax); delete classify; return 0; }
-
- if(namefile != "") { MPIReadNamesFile(namefileNames[s]); }
if (pid == 0) { //you are the root process
MPIPos = setFilePosFasta(fastaFileNames[s], numFastaSeqs); //fills MPIPos, returns numSeqs
//send file positions to all processes
- MPI_Bcast(&numFastaSeqs, 1, MPI_INT, 0, MPI_COMM_WORLD); //send numSeqs
- MPI_Bcast(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, 0, MPI_COMM_WORLD); //send file pos
+ for(int i = 1; i < processors; i++) {
+ MPI_Send(&numFastaSeqs, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
+ MPI_Send(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, i, tag, MPI_COMM_WORLD);
+ }
//figure out how many sequences you have to align
numSeqsPerProcessor = numFastaSeqs / processors;
MPI_Recv(&done, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status);
}
}else{ //you are a child process
- MPI_Bcast(&numFastaSeqs, 1, MPI_INT, 0, MPI_COMM_WORLD); //get numSeqs
+ MPI_Recv(&numFastaSeqs, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
MPIPos.resize(numFastaSeqs+1);
- MPI_Bcast(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, 0, MPI_COMM_WORLD); //get file positions
+ MPI_Recv(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status);
//figure out how many sequences you have to align
numSeqsPerProcessor = numFastaSeqs / processors;
MPI_File_close(&inMPI);
MPI_File_close(&outMPINewTax);
MPI_File_close(&outMPITempTax);
+ MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
#else
- //read namefile
- if(namefile != "") {
- nameMap.clear(); //remove old names
-
- ifstream inNames;
- openInputFile(namefileNames[s], inNames);
+
+ vector<unsigned long int> positions = divideFile(fastaFileNames[s], processors);
- string firstCol, secondCol;
- while(!inNames.eof()) {
- inNames >> firstCol >> secondCol; gobble(inNames);
- nameMap[firstCol] = getNumNames(secondCol); //ex. seq1 seq1,seq3,seq5 -> seq1 = 3.
- }
- inNames.close();
- }
-
- #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+ for (int i = 0; i < (positions.size()-1); i++) {
+ lines.push_back(new linePair(positions[i], positions[(i+1)]));
+ }
+
+ #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
if(processors == 1){
- ifstream inFASTA;
- openInputFile(fastaFileNames[s], inFASTA);
- numFastaSeqs=count(istreambuf_iterator<char>(inFASTA),istreambuf_iterator<char>(), '>');
- inFASTA.close();
-
- lines.push_back(new linePair(0, numFastaSeqs));
-
- driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]);
+ numFastaSeqs = driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]);
}
else{
- vector<int> positions;
processIDS.resize(0);
- ifstream inFASTA;
- openInputFile(fastaFileNames[s], inFASTA);
-
- string input;
- while(!inFASTA.eof()){
- input = getline(inFASTA);
- if (input.length() != 0) {
- if(input[0] == '>'){ int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); }
- }
- }
- inFASTA.close();
-
- numFastaSeqs = positions.size();
-
- int numSeqsPerProcessor = numFastaSeqs / processors;
-
- for (int i = 0; i < processors; i++) {
- int startPos = positions[ i * numSeqsPerProcessor ];
- if(i == processors - 1){
- numSeqsPerProcessor = numFastaSeqs - i * numSeqsPerProcessor;
- }
- lines.push_back(new linePair(startPos, numSeqsPerProcessor));
- }
- createProcesses(newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]);
+ numFastaSeqs = createProcesses(newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]);
rename((newTaxonomyFile + toString(processIDS[0]) + ".temp").c_str(), newTaxonomyFile.c_str());
rename((tempTaxonomyFile + toString(processIDS[0]) + ".temp").c_str(), tempTaxonomyFile.c_str());
}
#else
- ifstream inFASTA;
- openInputFile(fastaFileNames[s], inFASTA);
- numFastaSeqs=count(istreambuf_iterator<char>(inFASTA),istreambuf_iterator<char>(), '>');
- inFASTA.close();
-
- lines.push_back(new linePair(0, numFastaSeqs));
-
- driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]);
+ numFastaSeqs = driver(lines[0], newTaxonomyFile, tempTaxonomyFile, fastaFileNames[s]);
#endif
#endif
+ m->mothurOutEndLine();
+ m->mothurOut("It took " + toString(time(NULL) - start) + " secs to classify " + toString(numFastaSeqs) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine();
+ start = time(NULL);
+
+
#ifdef USE_MPI
if (pid == 0) { //this part does not need to be paralellized
+
+ if(namefile != "") { m->mothurOut("Reading " + namefileNames[s] + "..."); cout.flush(); MPIReadNamesFile(namefileNames[s]); m->mothurOut(" Done."); m->mothurOutEndLine(); }
+ #else
+ //read namefile
+ if(namefile != "") {
+
+ m->mothurOut("Reading " + namefileNames[s] + "..."); cout.flush();
+
+ nameMap.clear(); //remove old names
+
+ ifstream inNames;
+ openInputFile(namefileNames[s], inNames);
+
+ string firstCol, secondCol;
+ while(!inNames.eof()) {
+ inNames >> firstCol >> secondCol; gobble(inNames);
+
+ vector<string> temp;
+ splitAtComma(secondCol, temp);
+
+ nameMap[firstCol] = temp;
+ }
+ inNames.close();
+
+ m->mothurOut(" Done."); m->mothurOutEndLine();
+ }
#endif
- m->mothurOutEndLine();
- m->mothurOut("It took " + toString(time(NULL) - start) + " secs to classify " + toString(numFastaSeqs) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine();
- start = time(NULL);
-
string group = "";
if (groupfile != "") { group = groupfileNames[s]; }
PhyloSummary taxaSum(taxonomyFileName, group);
if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } delete classify; return 0; }
-
+
if (namefile == "") { taxaSum.summarize(tempTaxonomyFile); }
else {
ifstream in;
//read in users taxonomy file and add sequences to tree
string name, taxon;
+
while(!in.eof()){
in >> name >> taxon; gobble(in);
if (itNames == nameMap.end()) {
m->mothurOut(name + " is not in your name file please correct."); m->mothurOutEndLine(); exit(1);
}else{
- for (int i = 0; i < itNames->second; i++) {
- taxaSum.addSeqToTree(name, taxon); //add it as many times as there are identical seqs
+ for (int i = 0; i < itNames->second.size(); i++) {
+ taxaSum.addSeqToTree(itNames->second[i], taxon); //add it as many times as there are identical seqs
}
+ itNames->second.clear();
+ nameMap.erase(itNames->first);
}
}
in.close();
rename(unclass.c_str(), newTaxonomyFile.c_str());
m->mothurOutEndLine();
- m->mothurOut("It took " + toString(time(NULL) - start) + " secs to create the summary file for " + toString(numFastaSeqs) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine();
+ m->mothurOut("It took " + toString(time(NULL) - start) + " secs to create the summary file for " + toString(numFastaSeqs) + " sequences."); m->mothurOutEndLine(); m->mothurOutEndLine();
#ifdef USE_MPI
}
/**************************************************************************************************/
-void ClassifySeqsCommand::createProcesses(string taxFileName, string tempTaxFile, string filename) {
+int ClassifySeqsCommand::createProcesses(string taxFileName, string tempTaxFile, string filename) {
try {
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
int process = 0;
- // processIDS.resize(0);
+ int num = 0;
//loop through and create all the processes you want
while (process != processors) {
processIDS.push_back(pid); //create map from line number to pid so you can append files in correct order later
process++;
}else if (pid == 0){
- driver(lines[process], taxFileName + toString(getpid()) + ".temp", tempTaxFile + toString(getpid()) + ".temp", filename);
+ num = driver(lines[process], taxFileName + toString(getpid()) + ".temp", tempTaxFile + toString(getpid()) + ".temp", filename);
+
+ //pass numSeqs to parent
+ ofstream out;
+ string tempFile = toString(getpid()) + ".temp";
+ openOutputFile(tempFile, out);
+ out << num << endl;
+ out.close();
+
exit(0);
}else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
}
int temp = processIDS[i];
wait(&temp);
}
+
+ for (int i = 0; i < processIDS.size(); i++) {
+ ifstream in;
+ string tempFile = toString(processIDS[i]) + ".temp";
+ openInputFile(tempFile, in);
+ if (!in.eof()) { int tempNum = 0; in >> tempNum; num += tempNum; }
+ in.close(); remove(tempFile.c_str());
+ }
+
+ return num;
#endif
}
catch(exception& e) {
//**********************************************************************************************************************
-int ClassifySeqsCommand::driver(linePair* line, string taxFName, string tempTFName, string filename){
+int ClassifySeqsCommand::driver(linePair* filePos, string taxFName, string tempTFName, string filename){
try {
ofstream outTax;
openOutputFile(taxFName, outTax);
ifstream inFASTA;
openInputFile(filename, inFASTA);
-
- inFASTA.seekg(line->start);
string taxonomy;
- for(int i=0;i<line->numSeqs;i++){
+ inFASTA.seekg(filePos->start);
+
+ bool done = false;
+ int count = 0;
+
+ while (!done) {
if (m->control_pressed) { return 0; }
-
- Sequence* candidateSeq = new Sequence(inFASTA);
-
+
+ Sequence* candidateSeq = new Sequence(inFASTA); gobble(inFASTA);
+
if (candidateSeq->getName() != "") {
taxonomy = classify->getTaxonomy(candidateSeq);
outTaxSimple << candidateSeq->getName() << '\t' << classify->getSimpleTax() << endl;
}
- }
+ count++;
+ }
delete candidateSeq;
- if((i+1) % 100 == 0){
- m->mothurOut("Classifying sequence " + toString(i+1)); m->mothurOutEndLine();
- }
+ unsigned long int pos = inFASTA.tellg();
+ if ((pos == -1) || (pos >= filePos->end)) { break; }
+
+ //report progress
+ if((count) % 100 == 0){ m->mothurOut("Processing sequence: " + toString(count)); m->mothurOutEndLine(); }
}
-
+ //report progress
+ if((count) % 100 != 0){ m->mothurOut("Processing sequence: " + toString(count)); m->mothurOutEndLine(); }
+
inFASTA.close();
outTax.close();
outTaxSimple.close();
- return 1;
+ return count;
}
catch(exception& e) {
m->errorOut(e, "ClassifySeqsCommand", "driver");
}
//**********************************************************************************************************************
#ifdef USE_MPI
-int ClassifySeqsCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& newFile, MPI_File& tempFile, vector<long>& MPIPos){
+int ClassifySeqsCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& newFile, MPI_File& tempFile, vector<unsigned long int>& MPIPos){
try {
MPI_Status statusNew;
MPI_Status statusTemp;
string firstCol, secondCol;
while(!iss.eof()) {
iss >> firstCol >> secondCol; gobble(iss);
- nameMap[firstCol] = getNumNames(secondCol); //ex. seq1 seq1,seq3,seq5 -> seq1 = 3.
+
+ vector<string> temp;
+ splitAtComma(secondCol, temp);
+
+ nameMap[firstCol] = temp;
}
MPI_File_close(&inMPI);