*/
#include "filterseqscommand.h"
+#include "sequence.hpp"
/**************************************************************************************/
-FilterSeqsCommand::FilterSeqsCommand(){
- globaldata = GlobalData::getInstance();
-
- if(globaldata->getFastaFile() == "") { cout << "You must enter a fasta formatted file" << endl; }
- trump = globaldata->getTrump()[0];
- vertical =
-// readSeqs->read();
-// db = readSeqs->getDB();
-// numSeqs = db->size();
-//
-// alignmentLength = db->get(0).getAlignLength();
-//
-// filter = string(alignmentLength, '1');
+FilterSeqsCommand::FilterSeqsCommand(string option) {
+ try {
+ abort = false;
+ filterFileName = "";
+
+ //allow user to run help
+ if(option == "help") { help(); abort = true; }
+
+ else {
+ //valid paramters for this command
+ string Array[] = {"fasta", "trump", "soft", "hard", "vertical", "outputdir","inputdir", "processors"};
+ vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
+
+ OptionParser parser(option);
+ map<string,string> parameters = parser.getParameters();
+
+ ValidParameters validParameter;
+ map<string,string>::iterator it;
+
+ //check to make sure all parameters are valid for command
+ for (it = parameters.begin(); it != parameters.end(); it++) {
+ if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
+ }
+
+ //if the user changes the input directory command factory will send this info to us in the output parameter
+ string inputDir = validParameter.validFile(parameters, "inputdir", false);
+ if (inputDir == "not found"){ inputDir = ""; }
+ else {
+ string path;
+ it = parameters.find("fasta");
+ //user has given a template file
+ if(it != parameters.end()){
+ path = hasPath(it->second);
+ //if the user has not given a path then, add inputdir. else leave path alone.
+ if (path == "") { parameters["fasta"] = inputDir + it->second; }
+ }
+
+ it = parameters.find("hard");
+ //user has given a template file
+ if(it != parameters.end()){
+ path = hasPath(it->second);
+ //if the user has not given a path then, add inputdir. else leave path alone.
+ if (path == "") { parameters["hard"] = inputDir + it->second; }
+ }
+ }
+
+ //check for required parameters
+ fasta = validParameter.validFile(parameters, "fasta", false);
+ if (fasta == "not found") { m->mothurOut("fasta is a required parameter for the filter.seqs command."); m->mothurOutEndLine(); abort = true; }
+ else {
+ splitAtDash(fasta, fastafileNames);
+
+ //go through files and make sure they are good, if not, then disregard them
+ for (int i = 0; i < fastafileNames.size(); i++) {
+ if (inputDir != "") {
+ string path = hasPath(fastafileNames[i]);
+ //if the user has not given a path then, add inputdir. else leave path alone.
+ if (path == "") { fastafileNames[i] = inputDir + fastafileNames[i]; }
+ }
+
+ int ableToOpen;
+ ifstream in;
+ ableToOpen = openInputFile(fastafileNames[i], in);
+ if (ableToOpen == 1) {
+ m->mothurOut(fastafileNames[i] + " will be disregarded."); m->mothurOutEndLine();
+ //erase from file list
+ fastafileNames.erase(fastafileNames.begin()+i);
+ i--;
+ }else{
+ string simpleName = getSimpleName(fastafileNames[i]);
+ filterFileName += simpleName.substr(0, simpleName.find_first_of('.'));
+ }
+ in.close();
+ }
+
+ //make sure there is at least one valid file left
+ if (fastafileNames.size() == 0) { m->mothurOut("no valid files."); m->mothurOutEndLine(); abort = true; }
+ }
+
+ if (!abort) {
+ //if the user changes the output directory command factory will send this info to us in the output parameter
+ outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){
+ outputDir = "";
+ outputDir += hasPath(fastafileNames[0]); //if user entered a file with a path then preserve it
+ }
+ }
+ //check for optional parameter and set defaults
+ // ...at some point should added some additional type checking...
+
+ string temp;
+ temp = validParameter.validFile(parameters, "trump", false); if (temp == "not found") { temp = "*"; }
+ trump = temp[0];
+
+ temp = validParameter.validFile(parameters, "soft", false); if (temp == "not found") { soft = 0; }
+ else { soft = (float)atoi(temp.c_str()) / 100.0; }
+
+ 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"; }
+
+ numSeqs = 0;
+
+ }
+
+ }
+ catch(exception& e) {
+ m->errorOut(e, "FilterSeqsCommand", "FilterSeqsCommand");
+ exit(1);
+ }
+}
+
+//**********************************************************************************************************************
+
+void FilterSeqsCommand::help(){
+ try {
+
+ m->mothurOut("The filter.seqs command reads a file containing sequences and creates a .filter and .filter.fasta file.\n");
+ m->mothurOut("The filter.seqs command parameters are fasta, trump, soft, hard and vertical. \n");
+ m->mothurOut("The fasta parameter is required. You may enter several fasta files to build the filter from and filter, by separating their names with -'s.\n");
+ 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 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("Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFasta).\n\n");
+
+ }
+ catch(exception& e) {
+ m->errorOut(e, "FilterSeqsCommand", "help");
+ exit(1);
+ }
}
/**************************************************************************************/
-void FilterSeqsCommand::doHard() {
+int FilterSeqsCommand::execute() {
+ try {
-// string hardName = globaldata->getHard();
-// string hardFilter = "";
-//
-// ifstream fileHandle;
-// openInputFile(hardName, fileHandle);
-//
-// fileHandle >> hardFilter;
-//
-// if(hardFilter.length() != filter.length()){
-// cout << "The hard filter is not the same length as the alignment: Hard filter will not be applied." << endl;
-// }
-// else{
-// filter = hardFilter;
-// }
+ if (abort == true) { return 0; }
+
+ ifstream inFASTA;
+ openInputFile(fastafileNames[0], inFASTA);
+
+ Sequence testSeq(inFASTA);
+ alignmentLength = testSeq.getAlignLength();
+ inFASTA.close();
+
+ ////////////create filter/////////////////
+
+ filter = createFilter();
+
+ if (m->control_pressed) { return 0; }
+
+ ofstream outFilter;
+
+ string filterFile = outputDir + filterFileName + ".filter";
+ openOutputFile(filterFile, outFilter);
+ outFilter << filter << endl;
+ outFilter.close();
+ outputNames.push_back(filterFile);
+
+
+ ////////////run filter/////////////////
+
+ filterSequences();
+
+ int filteredLength = 0;
+ for(int i=0;i<alignmentLength;i++){
+ if(filter[i] == '1'){ filteredLength++; }
+ }
+
+ if (m->control_pressed) { for(int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; }
+
+ m->mothurOutEndLine();
+ m->mothurOut("Length of filtered alignment: " + toString(filteredLength)); m->mothurOutEndLine();
+ m->mothurOut("Number of columns removed: " + toString((alignmentLength-filteredLength))); m->mothurOutEndLine();
+ m->mothurOut("Length of the original alignment: " + toString(alignmentLength)); m->mothurOutEndLine();
+ m->mothurOut("Number of sequences used to construct filter: " + toString(numSeqs)); m->mothurOutEndLine();
+
+
+ m->mothurOutEndLine();
+ m->mothurOut("Output File Names: "); m->mothurOutEndLine();
+ for(int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
+ m->mothurOutEndLine();
+
+ return 0;
+
+ }
+ catch(exception& e) {
+ m->errorOut(e, "FilterSeqsCommand", "execute");
+ exit(1);
+ }
}
-
/**************************************************************************************/
+int FilterSeqsCommand::filterSequences() {
+ try {
+
+ for (int s = 0; s < fastafileNames.size(); s++) {
+
+ for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
+
+ string filteredFasta = outputDir + getRootName(getSimpleName(fastafileNames[s])) + "filter.fasta";
+#ifdef USE_MPI
+ int pid, start, end;
+ int tag = 2001;
+
+ 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
+
+ MPI_File outMPI;
+ MPI_File inMPI;
+ int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
+ int inMode=MPI_MODE_RDONLY;
+
+ char outFilename[filteredFasta.length()];
+ strcpy(outFilename, filteredFasta.c_str());
+
+ char inFileName[fastafileNames[s].length()];
+ 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);
+
+ if (pid == 0) { //you are the root process
+
+ setLines(fastafileNames[s]);
+
+ for (int j = 0; j < lines.size(); j++) { //each process
+ if (j != 0) { //don't send to yourself
+ MPI_Send(&lines[j]->start, 1, MPI_INT, j, tag, MPI_COMM_WORLD); //start position in file
+ MPI_Send(&bufferSizes[j], 1, MPI_INT, j, tag, MPI_COMM_WORLD); //how bytes for the read
+ }
+ }
+
+ //read your peice of file
+ char buf[bufferSizes[0]];
+ MPI_File_read_at(inMPI, lines[0]->start, buf, bufferSizes[0], MPI_CHAR, &status);
+ istringstream iss (buf,istringstream::in);
+
+ //do your part
+ driverMPIRun(iss, outMPI);
+
+ //wait on chidren
+ for(int i = 1; i < processors; i++) {
+ char buf[4];
+ MPI_Recv(buf, 4, MPI_CHAR, i, tag, MPI_COMM_WORLD, &status);
+ }
+
+ }else { //you are a child process
+ //receive your section of file
+ int startPos, numLines, bufferSize;
+ MPI_Recv(&startPos, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&bufferSize, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+
+ //read your peice of file
+ char buf2[bufferSize];
+ MPI_File_read_at(inMPI, startPos, buf2, bufferSize, MPI_CHAR, &status);
+ istringstream iss (buf2,istringstream::in);
+
+ //do your part
+ driverMPIRun(iss, outMPI);
+
+ char buf[4];
+ strcpy(buf, "done");
+
+ //tell parent you are done.
+ MPI_Send(buf, 4, MPI_CHAR, 0, tag, MPI_COMM_WORLD);
+ }
+
+ MPI_File_close(&outMPI);
+ MPI_File_close(&inMPI);
+
+#else
+ #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+ if(processors == 1){
+ ifstream inFASTA;
+ openInputFile(fastafileNames[s], inFASTA);
+ int numFastaSeqs=count(istreambuf_iterator<char>(inFASTA),istreambuf_iterator<char>(), '>');
+ inFASTA.close();
+
+ lines.push_back(new linePair(0, numFastaSeqs));
+
+ driverRunFilter(filter, filteredFasta, fastafileNames[s], lines[0]);
+ }else{
+ setLines(fastafileNames[s]);
+ createProcessesRunFilter(filter, fastafileNames[s]);
+
+ rename((fastafileNames[s] + toString(processIDS[0]) + ".temp").c_str(), filteredFasta.c_str());
+
+ //append fasta files
+ for(int i=1;i<processors;i++){
+ appendAlignFiles((fastafileNames[s] + toString(processIDS[i]) + ".temp"), filteredFasta);
+ remove((fastafileNames[s] + toString(processIDS[i]) + ".temp").c_str());
+ }
+ }
+
+ if (m->control_pressed) { return 1; }
+ #else
+ ifstream inFASTA;
+ openInputFile(fastafileNames[s], inFASTA);
+ int numFastaSeqs=count(istreambuf_iterator<char>(inFASTA),istreambuf_iterator<char>(), '>');
+ inFASTA.close();
+
+ lines.push_back(new linePair(0, numFastaSeqs));
+
+ driverRunFilter(filter, filteredFasta, fastafileNames[s], lines[0]);
-void FilterSeqsCommand::doTrump() {
+ if (m->control_pressed) { return 1; }
+ #endif
+#endif
+ outputNames.push_back(filteredFasta);
+ }
-
- for(int i = 0; i < numSeqs; i++) {
- string curAligned = db->get(i).getAligned();;
+ return 0;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "FilterSeqsCommand", "filterSequences");
+ exit(1);
+ }
+}
+/**************************************************************************************/
+int FilterSeqsCommand::driverMPIRun(istringstream& in, MPI_File& outMPI) {
+ try {
+ string outputString = "";
+ int count = 0;
+ MPI_Status status;
+
+ while (!in.eof()) {
+
+ Sequence seq(in); gobble(in);
+
+ if (seq.getName() != "") {
+ string align = seq.getAligned();
+ string filterSeq = "";
+
+ for(int j=0;j<alignmentLength;j++){
+ if(filter[j] == '1'){
+ filterSeq += align[j];
+ }
+ }
+
+ count++;
+ outputString += ">" + seq.getName() + "\n" + filterSeq + "\n";
+
+ if(count % 10 == 0){ //output to file
+ //send results to parent
+ int length = outputString.length();
+ char buf[length];
+ strcpy(buf, outputString.c_str());
+
+ MPI_File_write_shared(outMPI, buf, length, MPI_CHAR, &status);
+ outputString = "";
+ }
- for(int j = 0; j < alignmentLength; j++) {
- if(curAligned[j] == trump){
- filter[j] = '0';
}
}
- }
+
+ if(outputString != ""){ //output to file
+ //send results to parent
+ int length = outputString.length();
+ char buf[length];
+ strcpy(buf, outputString.c_str());
+
+ MPI_File_write_shared(outMPI, buf, length, MPI_CHAR, &status);
+ outputString = "";
+ }
+
+ return 0;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "FilterSeqsCommand", "driverRunFilter");
+ exit(1);
+ }
}
-
/**************************************************************************************/
-
-void FilterSeqsCommand::doVertical() {
-
-// vector<int> counts(alignmentLength, 0);
-//
-// for(int i = 0; i < numSeqs; i++) {
-// string curAligned = db->get(i).getAligned();;
-//
-// for(int j = 0; j < alignmentLength; j++) {
-// if(curAligned[j] == '-' || curAligned[j] == '.'){
-// counts[j]++;
-// }
-// }
-// }
-// for(int i=0;i<alignmentLength;i++){
-// if(counts[i] == numSeqs) { filter[i] = '0'; }
-// }
+int FilterSeqsCommand::driverRunFilter(string F, string outputFilename, string inputFilename, linePair* line) {
+ try {
+ ofstream out;
+ openOutputFile(outputFilename, out);
+
+ ifstream in;
+ openInputFile(inputFilename, in);
+
+ in.seekg(line->start);
+
+ for(int i=0;i<line->numSeqs;i++){
+
+ if (m->control_pressed) { in.close(); out.close(); return 0; }
+
+ Sequence seq(in);
+ if (seq.getName() != "") {
+ string align = seq.getAligned();
+ string filterSeq = "";
+
+ for(int j=0;j<alignmentLength;j++){
+ if(filter[j] == '1'){
+ filterSeq += align[j];
+ }
+ }
+
+ out << '>' << seq.getName() << endl << filterSeq << endl;
+ }
+ gobble(in);
+ }
+ out.close();
+ in.close();
+
+ return 0;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "FilterSeqsCommand", "driverRunFilter");
+ exit(1);
+ }
}
+/**************************************************************************************************/
+int FilterSeqsCommand::createProcessesRunFilter(string F, string filename) {
+ try {
+#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+ int process = 0;
+ int exitCommand = 1;
+ processIDS.clear();
+
+ //loop through and create all the processes you want
+ while (process != processors) {
+ int pid = fork();
+
+ if (pid > 0) {
+ 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){
+ string filteredFasta = filename + toString(getpid()) + ".temp";
+ driverRunFilter(F, filteredFasta, filename, lines[process]);
+ exit(0);
+ }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
+ }
+
+ //force parent to wait until all the processes are done
+ for (int i=0;i<processors;i++) {
+ int temp = processIDS[i];
+ wait(&temp);
+ }
+
+ return exitCommand;
+#endif
+ }
+ catch(exception& e) {
+ m->errorOut(e, "FilterSeqsCommand", "createProcessesRunFilter");
+ exit(1);
+ }
+}
/**************************************************************************************/
+string FilterSeqsCommand::createFilter() {
+ try {
+ string filterString = "";
+
+ Filters F;
+
+ if (soft != 0) { F.setSoft(soft); }
+ if (trump != '*') { F.setTrump(trump); }
+
+ F.setLength(alignmentLength);
+
+ if(soft != 0 || isTrue(vertical)){
+ F.initialize();
+ }
+
+ if(hard.compare("") != 0) { F.doHard(hard); }
+ else { F.setFilter(string(alignmentLength, '1')); }
+
+ numSeqs = 0;
+
+ if(trump != '*' || isTrue(vertical) || soft != 0){
+ for (int s = 0; s < fastafileNames.size(); s++) {
+
+#ifdef USE_MPI
+ int pid, rc, ierr;
+ int Atag = 1; int Ttag = 2; int Ctag = 3; int Gtag = 4; int Gaptag = 5;
+ int tag = 2001;
+
+ MPI_Status status;
+ MPI_File in;
+ rc = MPI_Comm_size(MPI_COMM_WORLD, &processors);
+ rc = MPI_Comm_rank(MPI_COMM_WORLD, &pid);
+
+ char tempFileName[fastafileNames[s].length()];
+ strcpy(tempFileName, fastafileNames[s].c_str());
+ cout << pid << " tempFileName " << tempFileName << endl;
+ MPI_File_open(MPI_COMM_WORLD, tempFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &in); //comm, filename, mode, info, filepointer
+ cout << pid << " here" << endl;
+ if (pid == 0) { //you are the root process
+ setLines(fastafileNames[s]);
+ cout << pid << " after setlines" << endl;
+ for (int j = 0; j < lines.size(); j++) { //each process
+ if (j != 0) { //don't send to yourself
+ MPI_Send(&lines[j]->start, 1, MPI_INT, j, tag, MPI_COMM_WORLD); //start position in file
+ MPI_Send(&bufferSizes[j], 1, MPI_INT, j, tag, MPI_COMM_WORLD); //how bytes for the read
+ }
+ }
+ cout << pid << " done sending" << endl;
+ char buf[bufferSizes[0]];
+ MPI_File_read_at(in, 0, buf, bufferSizes[0], MPI_CHAR, &status);
+ cout << pid << " done reading" << endl;
+ string tempBuf = buf;
+ if (tempBuf.length() > bufferSizes[0]) { tempBuf = tempBuf.substr(0, bufferSizes[0]); }
-void FilterSeqsCommand::doSoft() {
-
-// int softThreshold = numSeqs * (float)atoi(globaldata->getSoft().c_str()) / 100.0;
-//
-// vector<int> a(alignmentLength, 0);
-// vector<int> t(alignmentLength, 0);
-// vector<int> g(alignmentLength, 0);
-// vector<int> c(alignmentLength, 0);
-// vector<int> x(alignmentLength, 0);
-//
-// for(int i=0;i<numSeqs;i++){
-// string curAligned = db->get(i).getAligned();;
-//
-// for(int j=0;j<alignmentLength;j++){
-// if(toupper(curAligned[j]) == 'A') { a[j]++; }
-// else if(toupper(curAligned[j]) == 'T' || toupper(curAligned[i]) == 'U') { t[j]++; }
-// else if(toupper(curAligned[j]) == 'G') { g[j]++; }
-// else if(toupper(curAligned[j]) == 'C') { c[j]++; }
-// }
-// }
-//
-// for(int i=0;i<alignmentLength;i++){
-// if(a[i] < softThreshold && t[i] < softThreshold && g[i] < softThreshold && c[i] < softThreshold){
-// filter[i] = '0';
-// }
-// }
-}
+ MPICreateFilter(F, tempBuf);
+ cout << pid << "done with mpi create filter " << endl;
+ if (m->control_pressed) { MPI_File_close(&in); return filterString; }
+
+ vector<int> 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, tag, 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 { //i am the child process
+ cout << pid << endl;
+ int startPos, bufferSize;
+ ierr = MPI_Recv(&startPos, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ ierr = MPI_Recv(&bufferSize, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ cout << pid << '\t' << startPos << '\t' << bufferSize << endl;
+ //send freqs
+ char buf2[bufferSize];
+ MPI_File_read_at( in, startPos, buf2, bufferSize, MPI_CHAR, &status);
+
+ string tempBuf = buf2;
+ if (tempBuf.length() > bufferSize) { tempBuf = tempBuf.substr(0, bufferSize); }
+
+ MPICreateFilter(F, tempBuf);
+ cout << pid << "done with mpi create filter " << endl;
+ if (m->control_pressed) { MPI_File_close(&in); return filterString; }
+
+ //send my fequency counts
+ F.a.push_back(Atag);
+ int ierr = MPI_Send(&(F.a[0]), (alignmentLength+1), MPI_INT, 0, tag, MPI_COMM_WORLD);
+ F.t.push_back(Ttag);
+ ierr = MPI_Send (&(F.t[0]), (alignmentLength+1), MPI_INT, 0, tag, MPI_COMM_WORLD);
+ F.c.push_back(Ctag);
+ ierr = MPI_Send(&(F.c[0]), (alignmentLength+1), MPI_INT, 0, tag, MPI_COMM_WORLD);
+ F.g.push_back(Gtag);
+ ierr = MPI_Send(&(F.g[0]), (alignmentLength+1), MPI_INT, 0, tag, MPI_COMM_WORLD);
+ F.gap.push_back(Gaptag);
+ ierr = MPI_Send(&(F.gap[0]), (alignmentLength+1), MPI_INT, 0, tag, MPI_COMM_WORLD);
+ }
+
+ MPI_Barrier(MPI_COMM_WORLD);
+ MPI_File_close(&in);
+
+#else
+ #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+ if(processors == 1){
+ ifstream inFASTA;
+ openInputFile(fastafileNames[s], inFASTA);
+ int numFastaSeqs=count(istreambuf_iterator<char>(inFASTA),istreambuf_iterator<char>(), '>');
+ inFASTA.close();
+
+ numSeqs += numFastaSeqs;
+
+ for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
+ lines.push_back(new linePair(0, numFastaSeqs));
+
+ driverCreateFilter(F, fastafileNames[s], lines[0]);
+ }else{
+ setLines(fastafileNames[s]);
+ createProcessesCreateFilter(F, fastafileNames[s]);
+ }
+
+ if (m->control_pressed) { return filterString; }
+ #else
+ ifstream inFASTA;
+ openInputFile(fastafileNames[s], inFASTA);
+ int numFastaSeqs=count(istreambuf_iterator<char>(inFASTA),istreambuf_iterator<char>(), '>');
+ inFASTA.close();
+
+ numSeqs += numFastaSeqs;
+
+ for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
+ lines.push_back(new linePair(0, numFastaSeqs));
+
+ driverCreateFilter(F, fastafileNames[s], lines[0]);
+ if (m->control_pressed) { return filterString; }
+ #endif
+#endif
+
+ }
+ }
-int FilterSeqsCommand::execute() {
+ F.setNumSeqs(numSeqs);
+
+ if(isTrue(vertical) == 1) { F.doVertical(); }
+ if(soft != 0) { F.doSoft(); }
+
+ filterString = F.getFilter();
+
+ return filterString;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "FilterSeqsCommand", "createFilter");
+ exit(1);
+ }
+}
+/**************************************************************************************/
+int FilterSeqsCommand::driverCreateFilter(Filters& F, string filename, linePair* line) {
try {
- ifstream inFASTA;
- openInputFile(globaldata->getFastaFile(), inFASTA);
+ ifstream in;
+ openInputFile(filename, in);
+
+ in.seekg(line->start);
- Sequence currSequence(inFASTA);
- alignmentLength = currSequence.getAlignLength();
+ for(int i=0;i<line->numSeqs;i++){
+
+ if (m->control_pressed) { in.close(); return 1; }
+
+ Sequence seq(in);
+ if (seq.getName() != "") {
+ if (seq.getAligned().length() != alignmentLength) { m->mothurOut("Sequences are not all the same length, please correct."); m->mothurOutEndLine(); m->control_pressed = true; }
+
+ if(trump != '*'){ F.doTrump(seq); }
+ if(isTrue(vertical) || soft != 0){ F.getFreqs(seq); }
+ cout.flush();
+ }
+
+ //report progress
+ if((i+1) % 100 == 0){ m->mothurOut(toString(i+1)); m->mothurOutEndLine(); }
+ }
- //while
+ //report progress
+ if((line->numSeqs) % 100 != 0){ m->mothurOut(toString(line->numSeqs)); m->mothurOutEndLine(); }
+ in.close();
- if(globaldata->getHard().compare("") != 0) { doHard(); } // has to be applied first!
- if(globaldata->getTrump().compare("") != 0) { doTrump(); }
- if(isTrue(globaldata->getVertical())) { doVertical(); }
- if(globaldata->getSoft().compare("") != 0) { doSoft(); }
-
+ return 0;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "FilterSeqsCommand", "driverCreateFilter");
+ exit(1);
+ }
+}
+/**************************************************************************************/
+int FilterSeqsCommand::MPICreateFilter(Filters& F, string input) {
+ try {
+ vector<string> seqStrings;
+ parseBuffer(input, seqStrings);
+ for(int i=0;i<seqStrings.size();i++){
+
+ if (seqStrings[i].length() != alignmentLength) { cout << i << '\t' << seqStrings[i].length() << "Sequences are not all the same length, please correct." << endl; m->control_pressed = true; }
+
+ if (m->control_pressed) { return 1; }
+
+ Sequence seq("", seqStrings[i]);
+
+ if(trump != '*'){ F.doTrump(seq); }
+ if(isTrue(vertical) || soft != 0){ F.getFreqs(seq); }
+ cout.flush();
+
+ //report progress
+ if((i+1) % 100 == 0){ m->mothurOut(toString(i+1)); m->mothurOutEndLine(); }
+ }
+ //report progress
+ if((seqStrings.size()) % 100 != 0){ m->mothurOut(toString(seqStrings.size())); m->mothurOutEndLine(); }
+ return 0;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "FilterSeqsCommand", "MPICreateFilter");
+ exit(1);
+ }
+}
+
+/**************************************************************************************************/
+
+int FilterSeqsCommand::createProcessesCreateFilter(Filters& F, string filename) {
+ try {
+#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+ int process = 0;
+ int exitCommand = 1;
+ processIDS.clear();
+ //loop through and create all the processes you want
+ while (process != processors) {
+ int pid = vfork();
+
+ if (pid > 0) {
+ 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){
+ driverCreateFilter(F, filename, lines[process]);
+ exit(0);
+ }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
+ }
+
+ //force parent to wait until all the processes are done
+ for (int i=0;i<processors;i++) {
+ int temp = processIDS[i];
+ wait(&temp);
+ }
- ofstream outfile;
- string filterFile = getRootName(globaldata->inputFileName) + "filter";
- openOutputFile(filterFile, outfile);
+ return exitCommand;
+#endif
+ }
+ catch(exception& e) {
+ m->errorOut(e, "FilterSeqsCommand", "createProcessesCreateFilter");
+ exit(1);
+ }
+}
+/**************************************************************************************************/
- outfile << filter << endl;
- outfile.close();
+int FilterSeqsCommand::setLines(string filename) {
+ try {
+ for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
+ vector<long int> positions;
+ bufferSizes.clear();
- string filteredFasta = getRootName(globaldata->inputFileName) + "filter.fasta";
- openOutputFile(filteredFasta, outfile);
+ ifstream inFASTA;
+ openInputFile(filename, inFASTA);
+
+ string input;
+ while(!inFASTA.eof()){
+ input = getline(inFASTA);
- for(int i=0;i<numSeqs;i++){
- string curAligned = db->get(i).getAligned();
- outfile << '>' << db->get(i).getName() << endl;
- for(int j=0;j<alignmentLength;j++){
- if(filter[j] == '1'){
- outfile << curAligned[j];
- }
+ if (input.length() != 0) {
+ if(input[0] == '>'){ long int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); }
}
- outfile << endl;
}
- outfile.close();
+ inFASTA.close();
- int filteredLength = 0;
- for(int i=0;i<alignmentLength;i++){
- if(filter[i] == '1'){ filteredLength++; }
+ int numFastaSeqs = positions.size();
+
+ FILE * pFile;
+ long size;
+
+ //get num bytes in file
+ pFile = fopen (filename.c_str(),"rb");
+ if (pFile==NULL) perror ("Error opening file");
+ else{
+ fseek (pFile, 0, SEEK_END);
+ size=ftell (pFile);
+ fclose (pFile);
}
+
+ numSeqs += numFastaSeqs;
- cout << endl;
- cout << "Length of filtered alignment: " << filteredLength << endl;
- cout << "Number of columns removed: " << alignmentLength-filteredLength << endl;
- cout << "Length of the original alignment: " << alignmentLength << endl;
- cout << "Number of sequences used to construct filter: " << numSeqs << endl;
+ int numSeqsPerProcessor = numFastaSeqs / processors;
- globaldata->clear();
+ for (int i = 0; i < processors; i++) {
+
+ 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 ];
+ bufferSizes.push_back(myEnd-startPos);
+ }
+ lines.push_back(new linePair(startPos, numSeqsPerProcessor));
+ }
return 0;
-
}
catch(exception& e) {
- cout << "Standard Error: " << e.what() << " has occurred in the FilterSeqsCommand class Function execute. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
+ m->errorOut(e, "FilterSeqsCommand", "setLines");
exit(1);
}
- catch(...) {
- cout << "An unknown error has occurred in the FilterSeqsCommand class function execute. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
+}
+/**************************************************************************************************/
+int FilterSeqsCommand::parseBuffer(string file, vector<string>& seqs) {
+ try {
+ istringstream iss (file); //,istringstream::in
+ string name, seqstring;
+
+ while (!iss.eof()) {
+
+ if (m->control_pressed) { return 0; }
+
+ Sequence seq(iss); gobble(iss);
+
+ if (seq.getName() != "") {
+ seqs.push_back(seq.getAligned());
+ }
+ }
+
+ return 0;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "FilterSeqsCommand", "parseBuffer");
exit(1);
}
}
-
/**************************************************************************************/
projects / mothur.git / blobdiff