try {
if (abort == true) { return 0; }
- vector<string> outputNames;
ifstream inFASTA;
openInputFile(fastafileNames[0], inFASTA);
filter = createFilter();
+ if (m->control_pressed) { return 0; }
+
+ #ifdef USE_MPI
+ int pid;
+ MPI_Comm_rank(MPI_COMM_WORLD, &pid);
+
+ if (pid == 0) { //only one process should output the filter
+ #endif
+
ofstream outFilter;
string filterFile = outputDir + filterFileName + ".filter";
outFilter.close();
outputNames.push_back(filterFile);
+ #ifdef USE_MPI
+ }
+ #endif
////////////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 {
+
numSeqs = 0;
- for (int i = 0; i < fastafileNames.size(); i++) {
- ifstream in;
- openInputFile(fastafileNames[i], in);
- string filteredFasta = outputDir + getRootName(getSimpleName(fastafileNames[i])) + "filter.fasta";
- ofstream outFASTA;
- openOutputFile(filteredFasta, outFASTA);
+
+ 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]);
+
+ char bufF[alignmentLength];
+ strcpy(bufF, filter.c_str());
+
+ 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
+ MPI_Send(bufF, alignmentLength, MPI_CHAR, j, tag, MPI_COMM_WORLD);
+ }
+ }
+
+ //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, bufferSize;
+ char bufF[alignmentLength];
+ MPI_Recv(&startPos, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&bufferSize, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(bufF, alignmentLength, MPI_CHAR, 0, tag, MPI_COMM_WORLD, &status);
+
+ filter = bufF; //filter was made by process 0 so other processes need to get it
+
+ //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++){
+ appendFiles((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]);
+
+ if (m->control_pressed) { return 1; }
+ #endif
+#endif
outputNames.push_back(filteredFasta);
+ }
+
+ 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);
- while(!in.eof()){
- if (m->control_pressed) { in.close(); outFASTA.close(); for(int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; }
+ 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 = "";
+ }
+
+ }
+ }
+
+ 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);
+ }
+}
+/**************************************************************************************/
+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->num;i++){
+
+ if (m->control_pressed) { in.close(); out.close(); return 0; }
Sequence seq(in);
if (seq.getName() != "") {
}
}
- outFASTA << '>' << seq.getName() << endl << filterSeq << endl;
- numSeqs++;
+ out << '>' << seq.getName() << endl << filterSeq << endl;
}
gobble(in);
- }
- outFASTA.close();
- in.close();
- }
-
- int filteredLength = 0;
- for(int i=0;i<alignmentLength;i++){
- if(filter[i] == '1'){ filteredLength++; }
}
+ out.close();
+ in.close();
- if (m->control_pressed) { for(int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } return 0; }
+ 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();
- 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();
+ //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); }
+ }
- return 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", "execute");
+ m->errorOut(e, "FilterSeqsCommand", "createProcessesRunFilter");
exit(1);
}
}
/**************************************************************************************/
string FilterSeqsCommand::createFilter() {
try {
- string filterString = "";
-
+ string filterString = "";
Filters F;
if (soft != 0) { F.setSoft(soft); }
else { F.setFilter(string(alignmentLength, '1')); }
numSeqs = 0;
-
if(trump != '*' || isTrue(vertical) || soft != 0){
for (int s = 0; s < fastafileNames.size(); s++) {
for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
#ifdef USE_MPI
- int pid, rc, ierr;
+ int pid;
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);
-
+ MPI_File inMPI;
+ MPI_Comm_size(MPI_COMM_WORLD, &processors);
+ MPI_Comm_rank(MPI_COMM_WORLD, &pid);
char* tempFileName = new char(fastafileNames[s].length());
tempFileName = &(fastafileNames[s][0]);
-
- MPI_File_open(MPI_COMM_WORLD, tempFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &in); //comm, filename, mode, info, filepointer
+
+ MPI_File_open(MPI_COMM_WORLD, tempFileName, MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer
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(&lines[j]->numSeqs, 1, MPI_INT, j, tag, MPI_COMM_WORLD); //how many sequences we are sending
+ MPI_Send(&numSeqs, 1, MPI_INT, j, tag, MPI_COMM_WORLD);
MPI_Send(&bufferSizes[j], 1, MPI_INT, j, tag, MPI_COMM_WORLD); //how bytes for the read
}
}
- //cout << "done sending" << endl;
- //cout << "parent = " << pid << " lines = " << lines[pid]->start << '\t' << lines[pid]->numSeqs << " size = " << lines.size() << endl;
-
- cout << "parent = " << pid << " address of Filter " << &F << " address of FilterString " << &filterString << " address of numSeqs = " << &numSeqs << " address of soft = " << &soft << endl;
-
- char* buf = new char(bufferSizes[0]);
- //cout << pid << '\t' << bufferSizes[0] << " line 1 start pos = " << lines[1]->start << " buffer size 0 " << bufferSizes[0] << " buffer size 1 " << bufferSizes[1] << endl;
- MPI_File_read_at(in, 0, buf, bufferSizes[0], MPI_CHAR, &status);
-
- cout << pid << " done reading " << &buf << endl;
- string tempBuf = buf;
- delete buf;
- //cout << pid << '\t' << (tempBuf.substr(0, 10)) << endl;
-
- //parse buffer
- istringstream iss (tempBuf,istringstream::in);
- string name, seqstring;
- vector<string> seqs;
-
- while (iss) {
-
- if (m->control_pressed) { return filterString; }
- cout << "here" << endl;
- Sequence seq(iss);
- cout << "here1" << endl;
- gobble(iss);
- cout << seq.getName() << endl;
- if (seq.getName() != "") {
- seqs.push_back(seq.getAligned());
- }
-
- }
-
- for(int i=0;i<seqs.size();i++){
-
- if (m->control_pressed) { return filterString; }
- Sequence seq("", seqs[i]);
+ char buf[bufferSizes[0]];
+ MPI_File_read_at(inMPI, 0, buf, bufferSizes[0], MPI_CHAR, &status);
- 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((seqs.size()) % 100 != 0){ m->mothurOut(toString(seqs.size())); m->mothurOutEndLine(); }
+ string tempBuf = buf;
+ if (tempBuf.length() > bufferSizes[0]) { tempBuf = tempBuf.substr(0, bufferSizes[0]); }
- //do your part
- //MPICreateFilter(F, seqs);
+ MPICreateFilter(F, tempBuf);
- vector<int> temp; temp.resize(alignmentLength);
+ if (m->control_pressed) { MPI_File_close(&inMPI); 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++) {
- cout << "i = " << i << endl;
- int ierr = MPI_Recv(&temp, alignmentLength, MPI_INT, MPI_ANY_SOURCE, tag, MPI_COMM_WORLD, &status);
-
+ 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
-
- int sender = status.MPI_SOURCE;
-
+
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
}else if (receiveTag == Gaptag) { //you are recieveing the gap frequencies
for (int k = 0; k < alignmentLength; k++) { F.gap[k] += temp[k]; }
}
-
- m->mothurOut("receive tag = " + toString(receiveTag) + " " + toString(sender) + " is complete."); m->mothurOutEndLine();
}
}else { //i am the child process
- int startPos, numLines, bufferSize;
- cout << "child = " << pid << " address of Filter " << &F << " address of FilterString " << &filterString << " address of numSeqs = " << &numSeqs << " address of soft = " << &soft<< endl;
- ierr = MPI_Recv(&startPos, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
- ierr = MPI_Recv(&numLines, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
- ierr = MPI_Recv(&bufferSize, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
- //cout << "child = " << pid << " done recv messages startpos = " << startPos << " numLines = " << numLines << " buffersize = " << bufferSize << endl;
-
-
+
+ int startPos, bufferSize;
+ MPI_Recv(&startPos, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&numSeqs, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+ MPI_Recv(&bufferSize, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
+
//send freqs
- char* buf2 = new char(bufferSize);
- MPI_File_read_at( in, startPos, buf2, bufferSize, MPI_CHAR, &status);
- cout << pid << " done reading " << &buf2 << endl;
-
- string tempBuf = buf2;
- delete buf2;
- // cout << pid << '\t' << (tempBuf.substr(0, 10)) << endl;
- istringstream iss (tempBuf,istringstream::in);
-
- string name, seqstring;
- vector<string> seqs;
-
- while (iss) {
-
- if (m->control_pressed) { return filterString; }
- cout << "here" << endl;
- Sequence seq(iss);
- cout << "here1" << endl;
- gobble(iss);
- cout << seq.getName() << endl;
-
- if (seq.getName() != "") {
- seqs.push_back(seq.getAligned());
- }
- }
-
- for(int i=0;i<seqs.size();i++){
-
- if (m->control_pressed) { return filterString; }
+ char buf2[bufferSize];
+ MPI_File_read_at(inMPI, startPos, buf2, bufferSize, MPI_CHAR, &status);
- Sequence seq("", seqs[i]);
+ string tempBuf = buf2;
+ if (tempBuf.length() > bufferSize) { tempBuf = tempBuf.substr(0, bufferSize); }
- 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((seqs.size()) % 100 != 0){ m->mothurOut(toString(seqs.size())); m->mothurOutEndLine(); }
-
- //MPICreateFilter(F, seqs);
+ MPICreateFilter(F, tempBuf);
+ if (m->control_pressed) { MPI_File_close(&inMPI); return filterString; }
+
//send my fequency counts
F.a.push_back(Atag);
- int ierr = MPI_Send( &F.a[0], alignmentLength, MPI_INT, 0, tag, MPI_COMM_WORLD);
+ 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, MPI_INT, 0, tag, MPI_COMM_WORLD);
+ 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, MPI_INT, 0, tag, MPI_COMM_WORLD);
+ 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, MPI_INT, 0, tag, MPI_COMM_WORLD);
+ 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, MPI_INT, 0, tag, MPI_COMM_WORLD);
-
- cout << "child " << pid << " done sending counts" << endl;
+ ierr = MPI_Send(&(F.gap[0]), (alignmentLength+1), MPI_INT, 0, tag, MPI_COMM_WORLD);
}
MPI_Barrier(MPI_COMM_WORLD);
+ MPI_File_close(&inMPI);
#else
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
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);
lines.push_back(new linePair(0, numFastaSeqs));
- driverCreateFilter(F, lines[0], fastafileNames[s]);
+ driverCreateFilter(F, fastafileNames[s], lines[0]);
+ if (m->control_pressed) { return filterString; }
#endif
#endif
}
}
-#ifdef USE_MPI
-
-//merge all frequency data and create filter string
- //int pid;
- //MPI_Comm_rank(MPI_COMM_WORLD, &pid);
-
- //if (pid == 0) { //only one process should output to screen
-#endif
-
- cout << "made it here" << endl;
F.setNumSeqs(numSeqs);
if(isTrue(vertical) == 1) { F.doVertical(); }
if(soft != 0) { F.doSoft(); }
-//cout << "Filter String = " << F.getFilter() << endl;
+
filterString = F.getFilter();
return filterString;
in.seekg(line->start);
- for(int i=0;i<line->numSeqs;i++){
+ for(int i=0;i<line->num;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((line->numSeqs) % 100 != 0){ m->mothurOut(toString(line->numSeqs)); m->mothurOutEndLine(); }
+ if((line->num) % 100 != 0){ m->mothurOut(toString(line->num)); m->mothurOutEndLine(); }
in.close();
}
}
/**************************************************************************************/
-int FilterSeqsCommand::MPICreateFilter(Filters& F, vector<string>& seqStrings) {
+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[0]);
+ Sequence seq("", seqStrings[i]);
if(trump != '*'){ F.doTrump(seq); }
if(isTrue(vertical) || soft != 0){ F.getFreqs(seq); }
#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
int process = 0;
int exitCommand = 1;
- vector<int> processIDS;
+ processIDS.clear();
//loop through and create all the processes you want
while (process != processors) {
int FilterSeqsCommand::setLines(string filename) {
try {
- vector<int> positions;
- map<int, int> buf;
+
+ vector<long int> positions;
bufferSizes.clear();
- int pid;
- MPI_Comm_rank(MPI_COMM_WORLD, &pid);
-
ifstream inFASTA;
openInputFile(filename, inFASTA);
string input;
- int numbuf = 0;
while(!inFASTA.eof()){
input = getline(inFASTA);
if (input.length() != 0) {
- numbuf += input.length();
- if(input[0] == '>'){ long int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); buf[(pos - input.length() - 1)] = numbuf; }
+ if(input[0] == '>'){ long int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); }
}
}
-
inFASTA.close();
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;
int numSeqsPerProcessor = numFastaSeqs / processors;
long int startPos = positions[ i * numSeqsPerProcessor ];
if(i == processors - 1){
numSeqsPerProcessor = numFastaSeqs - i * numSeqsPerProcessor;
- bufferSizes.push_back(numbuf-buf[startPos]);
+ bufferSizes.push_back(size - startPos);
}else{
- int myEnd = positions[ (i+1) * numSeqsPerProcessor ];
- bufferSizes.push_back(buf[myEnd]-buf[startPos]);
+ long int myEnd = positions[ (i+1) * numSeqsPerProcessor ];
+ bufferSizes.push_back(myEnd-startPos);
}
lines.push_back(new linePair(startPos, numSeqsPerProcessor));
}
}
/**************************************************************************************************/
int FilterSeqsCommand::parseBuffer(string file, vector<string>& seqs) {
- try {
-
- istringstream iss (file,istringstream::in);
+ try {
+ istringstream iss (file); //,istringstream::in
string name, seqstring;
-int pid;
-MPI_Comm_rank(MPI_COMM_WORLD, &pid);
- Sequence* seq34 = new Sequence();
-cout << "address of new sequence " << pid << '\t' << seq34 << endl;
-cout << "address of seqStrings " << pid << '\t' << &seqs << endl;
-
- while (iss) {
+
+ while (!iss.eof()) {
if (m->control_pressed) { return 0; }
- cout << "here" << endl;
- Sequence* seq = new Sequence(iss);
- cout << "here1" << endl;
- gobble(iss);
- cout << seq->getName() << endl;
- if (seq->getName() != "") {
- seqs.push_back(seq->getAligned());
+
+ Sequence seq(iss); gobble(iss);
+
+ if (seq.getName() != "") {
+ seqs.push_back(seq.getAligned());
}
- delete seq;
}
return 0;