//**********************************************************************************************************************
vector<string> AlignCommand::setParameters(){
try {
- CommandParameter ptemplate("reference", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(ptemplate);
- CommandParameter pcandidate("fasta", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(pcandidate);
- CommandParameter psearch("search", "Multiple", "kmer-blast-suffix", "kmer", "", "", "",false,false); parameters.push_back(psearch);
- CommandParameter pksize("ksize", "Number", "", "8", "", "", "",false,false); parameters.push_back(pksize);
- CommandParameter pmatch("match", "Number", "", "1.0", "", "", "",false,false); parameters.push_back(pmatch);
- CommandParameter palign("align", "Multiple", "needleman-gotoh-blast-noalign", "needleman", "", "", "",false,false); parameters.push_back(palign);
- CommandParameter pmismatch("mismatch", "Number", "", "-1.0", "", "", "",false,false); parameters.push_back(pmismatch);
- CommandParameter pgapopen("gapopen", "Number", "", "-2.0", "", "", "",false,false); parameters.push_back(pgapopen);
- CommandParameter pgapextend("gapextend", "Number", "", "-1.0", "", "", "",false,false); parameters.push_back(pgapextend);
- CommandParameter pprocessors("processors", "Number", "", "1", "", "", "",false,false); parameters.push_back(pprocessors);
- CommandParameter pflip("flip", "Boolean", "", "F", "", "", "",false,false); parameters.push_back(pflip);
- CommandParameter psave("save", "Boolean", "", "F", "", "", "",false,false); parameters.push_back(psave);
- CommandParameter pthreshold("threshold", "Number", "", "0.50", "", "", "",false,false); parameters.push_back(pthreshold);
- CommandParameter pinputdir("inputdir", "String", "", "", "", "", "",false,false); parameters.push_back(pinputdir);
- CommandParameter poutputdir("outputdir", "String", "", "", "", "", "",false,false); parameters.push_back(poutputdir);
+ CommandParameter ptemplate("reference", "InputTypes", "", "", "none", "none", "none","",false,true,true); parameters.push_back(ptemplate);
+ CommandParameter pcandidate("fasta", "InputTypes", "", "", "none", "none", "none","fasta-alignreport-accnos",false,true,true); parameters.push_back(pcandidate);
+ CommandParameter psearch("search", "Multiple", "kmer-blast-suffix", "kmer", "", "", "","",false,false,true); parameters.push_back(psearch);
+ CommandParameter pksize("ksize", "Number", "", "8", "", "", "","",false,false); parameters.push_back(pksize);
+ CommandParameter pmatch("match", "Number", "", "1.0", "", "", "","",false,false); parameters.push_back(pmatch);
+ CommandParameter palign("align", "Multiple", "needleman-gotoh-blast-noalign", "needleman", "", "", "","",false,false,true); parameters.push_back(palign);
+ CommandParameter pmismatch("mismatch", "Number", "", "-1.0", "", "", "","",false,false); parameters.push_back(pmismatch);
+ CommandParameter pgapopen("gapopen", "Number", "", "-5.0", "", "", "","",false,false); parameters.push_back(pgapopen);
+ CommandParameter pgapextend("gapextend", "Number", "", "-2.0", "", "", "","",false,false); parameters.push_back(pgapextend);
+ CommandParameter pprocessors("processors", "Number", "", "1", "", "", "","",false,false,true); parameters.push_back(pprocessors);
+ CommandParameter pflip("flip", "Boolean", "", "F", "", "", "","",false,false); parameters.push_back(pflip);
+ CommandParameter psave("save", "Boolean", "", "F", "", "", "","",false,false); parameters.push_back(psave);
+ CommandParameter pthreshold("threshold", "Number", "", "0.50", "", "", "","",false,false); parameters.push_back(pthreshold);
+ CommandParameter pinputdir("inputdir", "String", "", "", "", "", "","",false,false); parameters.push_back(pinputdir);
+ CommandParameter poutputdir("outputdir", "String", "", "", "", "", "","",false,false); parameters.push_back(poutputdir);
vector<string> myArray;
for (int i = 0; i < parameters.size(); i++) { myArray.push_back(parameters[i].name); }
helpString += "The ksize parameter allows you to specify the kmer size for finding most similar template to candidate. The default is 8.";
helpString += "The match parameter allows you to specify the bonus for having the same base. The default is 1.0.";
helpString += "The mistmatch parameter allows you to specify the penalty for having different bases. The default is -1.0.";
- helpString += "The gapopen parameter allows you to specify the penalty for opening a gap in an alignment. The default is -2.0.";
- helpString += "The gapextend parameter allows you to specify the penalty for extending a gap in an alignment. The default is -1.0.";
+ helpString += "The gapopen parameter allows you to specify the penalty for opening a gap in an alignment. The default is -5.0.";
+ helpString += "The gapextend parameter allows you to specify the penalty for extending a gap in an alignment. The default is -2.0.";
helpString += "The flip parameter is used to specify whether or not you want mothur to try the reverse complement if a sequence falls below the threshold. The default is false.";
helpString += "The threshold is used to specify a cutoff at which an alignment is deemed 'bad' and the reverse complement may be tried. The default threshold is 0.50, meaning 50% of the bases are removed in the alignment.";
helpString += "If the flip parameter is set to true the reverse complement of the sequence is aligned and the better alignment is reported.";
}
}
//**********************************************************************************************************************
+string AlignCommand::getOutputPattern(string type) {
+ try {
+ string pattern = "";
+
+ if (type == "fasta") { pattern = "[filename],align"; } //makes file like: amazon.align
+ else if (type == "alignreport") { pattern = "[filename],align.report"; }
+ else if (type == "accnos") { pattern = "[filename],flip.accnos"; }
+ else { m->mothurOut("[ERROR]: No definition for type " + type + " output pattern.\n"); m->control_pressed = true; }
+
+ return pattern;
+ }
+ catch(exception& e) {
+ m->errorOut(e, "AlignCommand", "getOutputPattern");
+ exit(1);
+ }
+}
+//**********************************************************************************************************************
AlignCommand::AlignCommand(){
try {
abort = true; calledHelp = true;
//**********************************************************************************************************************
AlignCommand::AlignCommand(string option) {
try {
- abort = false; calledHelp = false;
+ abort = false; calledHelp = false;
ReferenceDB* rdb = ReferenceDB::getInstance();
//allow user to run help
// ...at some point should added some additional type checking...
string temp;
temp = validParameter.validFile(parameters, "ksize", false); if (temp == "not found"){ temp = "8"; }
- convert(temp, kmerSize);
+ m->mothurConvert(temp, kmerSize);
temp = validParameter.validFile(parameters, "match", false); if (temp == "not found"){ temp = "1.0"; }
- convert(temp, match);
+ m->mothurConvert(temp, match);
temp = validParameter.validFile(parameters, "mismatch", false); if (temp == "not found"){ temp = "-1.0"; }
- convert(temp, misMatch);
+ m->mothurConvert(temp, misMatch);
- temp = validParameter.validFile(parameters, "gapopen", false); if (temp == "not found"){ temp = "-2.0"; }
- convert(temp, gapOpen);
+ temp = validParameter.validFile(parameters, "gapopen", false); if (temp == "not found"){ temp = "-5.0"; }
+ m->mothurConvert(temp, gapOpen);
- temp = validParameter.validFile(parameters, "gapextend", false); if (temp == "not found"){ temp = "-1.0"; }
- convert(temp, gapExtend);
+ temp = validParameter.validFile(parameters, "gapextend", false); if (temp == "not found"){ temp = "-2.0"; }
+ m->mothurConvert(temp, gapExtend);
temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); }
m->setProcessors(temp);
- convert(temp, processors);
+ m->mothurConvert(temp, processors);
temp = validParameter.validFile(parameters, "flip", false); if (temp == "not found"){ temp = "f"; }
flip = m->isTrue(temp);
temp = validParameter.validFile(parameters, "save", false); if (temp == "not found"){ temp = "f"; }
save = m->isTrue(temp);
- //this is so the threads can quickly load the reference data
- #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
- #else
- if (processors != 1) { save = true; }
- #endif
rdb->save = save;
if (save) { //clear out old references
rdb->clearMemory();
else { if (save) { rdb->setSavedReference(templateFileName); } }
temp = validParameter.validFile(parameters, "threshold", false); if (temp == "not found"){ temp = "0.50"; }
- convert(temp, threshold);
+ m->mothurConvert(temp, threshold);
search = validParameter.validFile(parameters, "search", false); if (search == "not found"){ search = "kmer"; }
if ((search != "suffix") && (search != "kmer") && (search != "blast")) { m->mothurOut("invalid search option: choices are kmer, suffix or blast."); m->mothurOutEndLine(); abort=true; }
m->mothurOut("Aligning sequences from " + candidateFileNames[s] + " ..." ); m->mothurOutEndLine();
if (outputDir == "") { outputDir += m->hasPath(candidateFileNames[s]); }
- string alignFileName = outputDir + m->getRootName(m->getSimpleName(candidateFileNames[s])) + "align";
- string reportFileName = outputDir + m->getRootName(m->getSimpleName(candidateFileNames[s])) + "align.report";
- string accnosFileName = outputDir + m->getRootName(m->getSimpleName(candidateFileNames[s])) + "flip.accnos";
+ map<string, string> variables; variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(candidateFileNames[s]));
+ string alignFileName = getOutputFileName("fasta", variables);
+ string reportFileName = getOutputFileName("alignreport", variables);
+ string accnosFileName = getOutputFileName("accnos", variables);
+
bool hasAccnos = true;
int numFastaSeqs = 0;
#ifdef USE_MPI
int pid, numSeqsPerProcessor;
int tag = 2001;
- vector<unsigned long int> MPIPos;
+ vector<unsigned long long> MPIPos;
MPIWroteAccnos = false;
MPI_Status status;
#else
- vector<unsigned long int> positions;
- #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+ vector<unsigned long long> positions;
+ #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
positions = m->divideFile(candidateFileNames[s], processors);
for (int i = 0; i < (positions.size()-1); i++) { lines.push_back(new linePair(positions[i], positions[(i+1)])); }
#else
lines.push_back(new linePair(0, 1000));
}else {
positions = m->setFilePosFasta(candidateFileNames[s], numFastaSeqs);
-
+ if (positions.size() < processors) { processors = positions.size(); }
+
//figure out how many sequences you have to process
int numSeqsPerProcessor = numFastaSeqs / processors;
for (int i = 0; i < processors; i++) {
//moved this into driver to avoid deep copies in windows paralellized version
Alignment* alignment;
int longestBase = templateDB->getLongestBase();
+ if (m->debug) { m->mothurOut("[DEBUG]: template longest base = " + toString(templateDB->getLongestBase()) + " \n"); }
if(align == "gotoh") { alignment = new GotohOverlap(gapOpen, gapExtend, match, misMatch, longestBase); }
else if(align == "needleman") { alignment = new NeedlemanOverlap(gapOpen, match, misMatch, longestBase); }
else if(align == "blast") { alignment = new BlastAlignment(gapOpen, gapExtend, match, misMatch); }
int numBasesNeeded = origNumBases * threshold;
if (candidateSeq->getName() != "") { //incase there is a commented sequence at the end of a file
- if (candidateSeq->getUnaligned().length() > alignment->getnRows()) {
- alignment->resize(candidateSeq->getUnaligned().length()+1);
+ if (candidateSeq->getUnaligned().length()+1 > alignment->getnRows()) {
+ if (m->debug) { m->mothurOut("[DEBUG]: " + candidateSeq->getName() + " " + toString(candidateSeq->getUnaligned().length()) + " " + toString(alignment->getnRows()) + " \n"); }
+ alignment->resize(candidateSeq->getUnaligned().length()+2);
}
-
Sequence temp = templateDB->findClosestSequence(candidateSeq);
- Sequence* templateSeq = &temp;
+ Sequence* templateSeq = new Sequence(temp.getName(), temp.getAligned());
float searchScore = templateDB->getSearchScore();
//get reverse compliment
copy = new Sequence(candidateSeq->getName(), originalUnaligned);
copy->reverseComplement();
+
+ if (m->debug) { m->mothurOut("[DEBUG]: flipping " + candidateSeq->getName() + " \n"); }
//rerun alignment
Sequence temp2 = templateDB->findClosestSequence(copy);
- Sequence* templateSeq2 = &temp2;
+ Sequence* templateSeq2 = new Sequence(temp2.getName(), temp2.getAligned());
+
+ if (m->debug) { m->mothurOut("[DEBUG]: closest template " + temp2.getName() + " \n"); }
searchScore = templateDB->getSearchScore();
nast2 = new Nast(alignment, copy, templateSeq2);
+
+ if (m->debug) { m->mothurOut("[DEBUG]: completed Nast2 " + candidateSeq->getName() + " flipped numBases = " + toString(copy->getNumBases()) + " old numbases = " + toString(candidateSeq->getNumBases()) +" \n"); }
//check if any better
if (copy->getNumBases() > candidateSeq->getNumBases()) {
candidateSeq->setAligned(copy->getAligned()); //use reverse compliments alignment since its better
- templateSeq = templateSeq2;
+ delete templateSeq;
+ templateSeq = templateSeq2;
delete nast;
nast = nast2;
needToDeleteCopy = true;
}else{
wasBetter = "\treverse complement did NOT produce a better alignment so it was not used, please check sequence.";
delete nast2;
+ delete templateSeq2;
delete copy;
}
+ if (m->debug) { m->mothurOut("[DEBUG]: done.\n"); }
}
//create accnos file with names
report.print();
delete nast;
+ delete templateSeq;
if (needToDeleteCopy) { delete copy; }
count++;
}
delete candidateSeq;
- #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
- unsigned long int pos = inFASTA.tellg();
+ #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
+ unsigned long long pos = inFASTA.tellg();
if ((pos == -1) || (pos >= filePos->end)) { break; }
#else
if (inFASTA.eof()) { break; }
#endif
//report progress
- if((count) % 100 == 0){ m->mothurOut(toString(count)); m->mothurOutEndLine(); }
+ if((count) % 100 == 0){ m->mothurOutJustToScreen(toString(count) + "\n"); }
}
//report progress
- if((count) % 100 != 0){ m->mothurOut(toString(count)); m->mothurOutEndLine(); }
+ if((count) % 100 != 0){ m->mothurOutJustToScreen(toString(count) + "\n"); }
delete alignment;
alignmentFile.close();
}
//**********************************************************************************************************************
#ifdef USE_MPI
-int AlignCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& alignFile, MPI_File& reportFile, MPI_File& accnosFile, vector<unsigned long int>& MPIPos){
+int AlignCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& alignFile, MPI_File& reportFile, MPI_File& accnosFile, vector<unsigned long long>& MPIPos){
try {
string outputString = "";
MPI_Status statusReport;
}
Sequence temp = templateDB->findClosestSequence(candidateSeq);
- Sequence* templateSeq = &temp;
+ Sequence* templateSeq = new Sequence(temp.getName(), temp.getAligned());
float searchScore = templateDB->getSearchScore();
//rerun alignment
Sequence temp2 = templateDB->findClosestSequence(copy);
- Sequence* templateSeq2 = &temp2;
+ Sequence* templateSeq2 = new Sequence(temp2.getName(), temp2.getAligned());
searchScore = templateDB->getSearchScore();
//check if any better
if (copy->getNumBases() > candidateSeq->getNumBases()) {
candidateSeq->setAligned(copy->getAligned()); //use reverse compliments alignment since its better
- templateSeq = templateSeq2;
+ delete templateSeq;
+ templateSeq = templateSeq2;
delete nast;
nast = nast2;
needToDeleteCopy = true;
}else{
wasBetter = "\treverse complement did NOT produce a better alignment, please check sequence.";
delete nast2;
+ delete templateSeq2;
delete copy;
}
}
delete buf3;
delete nast;
+ delete templateSeq;
if (needToDeleteCopy) { delete copy; }
}
delete candidateSeq;
try {
int num = 0;
processIDS.resize(0);
-#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+ bool recalc = false;
+
+#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
int process = 1;
//loop through and create all the processes you want
while (process != processors) {
- int pid = fork();
+ pid_t 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){
- num = driver(lines[process], alignFileName + toString(getpid()) + ".temp", reportFileName + toString(getpid()) + ".temp", accnosFName + toString(getpid()) + ".temp", filename);
+ num = driver(lines[process], alignFileName + toString(m->mothurGetpid(process)) + ".temp", reportFileName + toString(m->mothurGetpid(process)) + ".temp", accnosFName + m->mothurGetpid(process) + ".temp", filename);
//pass numSeqs to parent
ofstream out;
- string tempFile = alignFileName + toString(getpid()) + ".num.temp";
+ string tempFile = alignFileName + toString(m->mothurGetpid(process)) + ".num.temp";
m->openOutputFile(tempFile, out);
out << num << endl;
out.close();
exit(0);
}else {
- m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine();
- for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); }
- exit(0);
+ m->mothurOut("[ERROR]: unable to spawn the number of processes you requested, reducing number to " + toString(process) + "\n"); processors = process;
+ for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); }
+ recalc = true;
+ break;
}
}
+ if (recalc) {
+ for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
+ vector<unsigned long long> positions;
+ positions = m->divideFile(filename, processors);
+ for (int i = 0; i < (positions.size()-1); i++) { lines.push_back(new linePair(positions[i], positions[(i+1)])); }
+
+ num = 0;
+ processIDS.resize(0);
+ process = 1;
+
+ while (process != processors) {
+ pid_t 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){
+ num = driver(lines[process], alignFileName + toString(m->mothurGetpid(process)) + ".temp", reportFileName + toString(m->mothurGetpid(process)) + ".temp", accnosFName + m->mothurGetpid(process) + ".temp", filename);
+
+ //pass numSeqs to parent
+ ofstream out;
+ string tempFile = alignFileName + toString(m->mothurGetpid(process)) + ".num.temp";
+ m->openOutputFile(tempFile, out);
+ out << num << endl;
+ out.close();
+
+ exit(0);
+ }else {
+ m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine();
+ for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); }
+ exit(0);
+ }
+ }
+ }
+
//do my part
num = driver(lines[0], alignFileName, reportFileName, accnosFName, filename);
if (!in.eof()) { int tempNum = 0; in >> tempNum; num += tempNum; }
in.close(); m->mothurRemove(tempFile);
- appendAlignFiles((alignFileName + toString(processIDS[i]) + ".temp"), alignFileName);
+ m->appendFiles((alignFileName + toString(processIDS[i]) + ".temp"), alignFileName);
m->mothurRemove((alignFileName + toString(processIDS[i]) + ".temp"));
appendReportFiles((reportFileName + toString(processIDS[i]) + ".temp"), reportFileName);
rename(nonBlankAccnosFiles[0].c_str(), accnosFName.c_str());
for (int h=1; h < nonBlankAccnosFiles.size(); h++) {
- appendAlignFiles(nonBlankAccnosFiles[h], accnosFName);
+ m->appendFiles(nonBlankAccnosFiles[h], accnosFName);
m->mothurRemove(nonBlankAccnosFiles[h]);
}
}else { //recreate the accnosfile if needed
string extension = "";
if (i != 0) { extension = toString(i) + ".temp"; }
- alignData* tempalign = new alignData((alignFileName + extension), (reportFileName + extension), (accnosFName + extension), filename, align, search, kmerSize, m, lines[i]->start, lines[i]->end, flip, match, misMatch, gapOpen, gapExtend, threshold, i);
+ alignData* tempalign = new alignData(templateFileName, (alignFileName + extension), (reportFileName + extension), (accnosFName + extension), filename, align, search, kmerSize, m, lines[i]->start, lines[i]->end, flip, match, misMatch, gapOpen, gapExtend, threshold, i);
pDataArray.push_back(tempalign);
processIDS.push_back(i);
//Close all thread handles and free memory allocations.
for(int i=0; i < pDataArray.size(); i++){
+ if (pDataArray[i]->count != pDataArray[i]->end) {
+ m->mothurOut("[ERROR]: process " + toString(i) + " only processed " + toString(pDataArray[i]->count) + " of " + toString(pDataArray[i]->end) + " sequences assigned to it, quitting. \n"); m->control_pressed = true;
+ }
num += pDataArray[i]->count;
CloseHandle(hThreadArray[i]);
delete pDataArray[i];
else { m->mothurRemove(accnosFName); } //remove so other files can be renamed to it
for (int i = 1; i < processors; i++) {
- appendAlignFiles((alignFileName + toString(i) + ".temp"), alignFileName);
+ m->appendFiles((alignFileName + toString(i) + ".temp"), alignFileName);
m->mothurRemove((alignFileName + toString(i) + ".temp"));
appendReportFiles((reportFileName + toString(i) + ".temp"), reportFileName);
rename(nonBlankAccnosFiles[0].c_str(), accnosFName.c_str());
for (int h=1; h < nonBlankAccnosFiles.size(); h++) {
- appendAlignFiles(nonBlankAccnosFiles[h], accnosFName);
+ m->appendFiles(nonBlankAccnosFiles[h], accnosFName);
m->mothurRemove(nonBlankAccnosFiles[h]);
}
}else { //recreate the accnosfile if needed
exit(1);
}
}
-/**************************************************************************************************/
-
-void AlignCommand::appendAlignFiles(string temp, string filename) {
- try{
-
- ofstream output;
- ifstream input;
- m->openOutputFileAppend(filename, output);
- m->openInputFile(temp, input);
-
- while(char c = input.get()){
- if(input.eof()) { break; }
- else { output << c; }
- }
-
- input.close();
- output.close();
- }
- catch(exception& e) {
- m->errorOut(e, "AlignCommand", "appendAlignFiles");
- exit(1);
- }
-}
//**********************************************************************************************************************
void AlignCommand::appendReportFiles(string temp, string filename) {
while (!input.eof()) { char c = input.get(); if (c == 10 || c == 13){ break; } } // get header line
- while(char c = input.get()){
- if(input.eof()) { break; }
- else { output << c; }
- }
+ char buffer[4096];
+ while (!input.eof()) {
+ input.read(buffer, 4096);
+ output.write(buffer, input.gcount());
+ }
input.close();
output.close();
projects / mothur.git / blobdiff