2 * chimeraccodecommand.cpp
5 * Created by westcott on 3/30/10.
6 * Copyright 2010 Schloss Lab. All rights reserved.
10 #include "chimeraccodecommand.h"
13 //***************************************************************************************************************
15 ChimeraCcodeCommand::ChimeraCcodeCommand(string option) {
19 //allow user to run help
20 if(option == "help") { help(); abort = true; }
23 //valid paramters for this command
24 string Array[] = {"fasta", "filter", "processors", "window", "template", "mask", "numwanted", "outputdir","inputdir", };
25 vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
27 OptionParser parser(option);
28 map<string,string> parameters = parser.getParameters();
30 ValidParameters validParameter("chimera.ccode");
31 map<string,string>::iterator it;
33 //check to make sure all parameters are valid for command
34 for (it = parameters.begin(); it != parameters.end(); it++) {
35 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
38 //if the user changes the input directory command factory will send this info to us in the output parameter
39 string inputDir = validParameter.validFile(parameters, "inputdir", false);
40 if (inputDir == "not found"){ inputDir = ""; }
43 it = parameters.find("template");
44 //user has given a template file
45 if(it != parameters.end()){
46 path = m->hasPath(it->second);
47 //if the user has not given a path then, add inputdir. else leave path alone.
48 if (path == "") { parameters["template"] = inputDir + it->second; }
52 //check for required parameters
53 fastafile = validParameter.validFile(parameters, "fasta", false);
54 if (fastafile == "not found") { fastafile = ""; m->mothurOut("fasta is a required parameter for the chimera.ccode command."); m->mothurOutEndLine(); abort = true; }
56 m->splitAtDash(fastafile, fastaFileNames);
58 //go through files and make sure they are good, if not, then disregard them
59 for (int i = 0; i < fastaFileNames.size(); i++) {
61 string path = m->hasPath(fastaFileNames[i]);
62 //if the user has not given a path then, add inputdir. else leave path alone.
63 if (path == "") { fastaFileNames[i] = inputDir + fastaFileNames[i]; }
69 ableToOpen = m->openInputFile(fastaFileNames[i], in, "noerror");
71 //if you can't open it, try default location
72 if (ableToOpen == 1) {
73 if (m->getDefaultPath() != "") { //default path is set
74 string tryPath = m->getDefaultPath() + m->getSimpleName(fastaFileNames[i]);
75 m->mothurOut("Unable to open " + fastaFileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine();
76 ableToOpen = m->openInputFile(tryPath, in, "noerror");
77 fastaFileNames[i] = tryPath;
82 if (ableToOpen == 1) {
83 m->mothurOut("Unable to open " + fastaFileNames[i] + ". It will be disregarded."); m->mothurOutEndLine();
84 //erase from file list
85 fastaFileNames.erase(fastaFileNames.begin()+i);
90 //make sure there is at least one valid file left
91 if (fastaFileNames.size() == 0) { m->mothurOut("no valid files."); m->mothurOutEndLine(); abort = true; }
94 //if the user changes the output directory command factory will send this info to us in the output parameter
95 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){ outputDir = ""; }
97 templatefile = validParameter.validFile(parameters, "template", true);
98 if (templatefile == "not open") { abort = true; }
99 else if (templatefile == "not found") { templatefile = ""; m->mothurOut("template is a required parameter for the chimera.ccode command."); m->mothurOutEndLine(); abort = true; }
101 maskfile = validParameter.validFile(parameters, "mask", false);
102 if (maskfile == "not found") { maskfile = ""; }
103 else if (maskfile != "default") {
104 if (inputDir != "") {
105 string path = m->hasPath(maskfile);
106 //if the user has not given a path then, add inputdir. else leave path alone.
107 if (path == "") { maskfile = inputDir + maskfile; }
111 int ableToOpen = m->openInputFile(maskfile, in);
112 if (ableToOpen == 1) { abort = true; }
117 temp = validParameter.validFile(parameters, "filter", false); if (temp == "not found") { temp = "F"; }
118 filter = m->isTrue(temp);
120 temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found") { temp = "1"; }
121 convert(temp, processors);
123 temp = validParameter.validFile(parameters, "window", false); if (temp == "not found") { temp = "0"; }
124 convert(temp, window);
126 temp = validParameter.validFile(parameters, "numwanted", false); if (temp == "not found") { temp = "20"; }
127 convert(temp, numwanted);
131 catch(exception& e) {
132 m->errorOut(e, "ChimeraCcodeCommand", "ChimeraCcodeCommand");
136 //**********************************************************************************************************************
138 void ChimeraCcodeCommand::help(){
141 m->mothurOut("The chimera.ccode command reads a fastafile and templatefile and outputs potentially chimeric sequences.\n");
142 m->mothurOut("This command was created using the algorythms described in the 'Evaluating putative chimeric sequences from PCR-amplified products' paper by Juan M. Gonzalez, Johannes Zimmerman and Cesareo Saiz-Jimenez.\n");
143 m->mothurOut("The chimera.ccode command parameters are fasta, template, filter, mask, processors, window and numwanted.\n");
144 m->mothurOut("The fasta parameter allows you to enter the fasta file containing your potentially chimeric sequences, and is required. \n");
145 m->mothurOut("You may enter multiple fasta files by separating their names with dashes. ie. fasta=abrecovery.fasta-amzon.fasta \n");
146 m->mothurOut("The template parameter allows you to enter a template file containing known non-chimeric sequences, and is required. \n");
147 m->mothurOut("The filter parameter allows you to specify if you would like to apply a vertical and 50% soft filter. \n");
148 m->mothurOut("The processors parameter allows you to specify how many processors you would like to use. The default is 1. \n");
150 m->mothurOut("When using MPI, the processors parameter is set to the number of MPI processes running. \n");
152 m->mothurOut("The mask parameter allows you to specify a file containing one sequence you wish to use as a mask for the your sequences. \n");
153 m->mothurOut("The window parameter allows you to specify the window size for searching for chimeras. \n");
154 m->mothurOut("The numwanted parameter allows you to specify how many sequences you would each query sequence compared with.\n");
155 m->mothurOut("The chimera.ccode command should be in the following format: \n");
156 m->mothurOut("chimera.ccode(fasta=yourFastaFile, template=yourTemplate) \n");
157 m->mothurOut("Example: chimera.ccode(fasta=AD.align, template=core_set_aligned.imputed.fasta) \n");
158 m->mothurOut("Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFastaFile).\n\n");
160 catch(exception& e) {
161 m->errorOut(e, "ChimeraCcodeCommand", "help");
166 //***************************************************************************************************************
168 ChimeraCcodeCommand::~ChimeraCcodeCommand(){ /* do nothing */ }
170 //***************************************************************************************************************
172 int ChimeraCcodeCommand::execute(){
175 if (abort == true) { return 0; }
177 for (int s = 0; s < fastaFileNames.size(); s++) {
179 m->mothurOut("Checking sequences from " + fastaFileNames[s] + " ..." ); m->mothurOutEndLine();
181 int start = time(NULL);
184 if (maskfile == "default") { m->mothurOut("I am using the default 236627 EU009184.1 Shigella dysenteriae str. FBD013."); m->mothurOutEndLine(); }
186 chimera = new Ccode(fastaFileNames[s], templatefile, filter, maskfile, window, numwanted, outputDir);
188 //is your template aligned?
189 if (chimera->getUnaligned()) { m->mothurOut("Your template sequences are different lengths, please correct."); m->mothurOutEndLine(); delete chimera; return 0; }
190 templateSeqsLength = chimera->getLength();
192 if (outputDir == "") { outputDir = m->hasPath(fastaFileNames[s]); }//if user entered a file with a path then preserve it
193 string outputFileName, accnosFileName;
194 if (maskfile != "") {
195 outputFileName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + maskfile + ".ccode.chimeras";
196 accnosFileName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + maskfile + ".ccode.accnos";
198 outputFileName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + "ccode.chimeras";
199 accnosFileName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + "ccode.accnos";
202 string mapInfo = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + "mapinfo";
204 if (m->control_pressed) { delete chimera; for (int j = 0; j < outputNames.size(); j++) { remove(outputNames[j].c_str()); } return 0; }
208 int pid, end, numSeqsPerProcessor;
210 vector<unsigned long int> MPIPos;
213 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
214 MPI_Comm_size(MPI_COMM_WORLD, &processors);
218 MPI_File outMPIAccnos;
220 int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
221 int inMode=MPI_MODE_RDONLY;
223 char outFilename[1024];
224 strcpy(outFilename, outputFileName.c_str());
226 char outAccnosFilename[1024];
227 strcpy(outAccnosFilename, accnosFileName.c_str());
229 char inFileName[1024];
230 strcpy(inFileName, fastaFileNames[s].c_str());
232 MPI_File_open(MPI_COMM_WORLD, inFileName, inMode, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer
233 MPI_File_open(MPI_COMM_WORLD, outFilename, outMode, MPI_INFO_NULL, &outMPI);
234 MPI_File_open(MPI_COMM_WORLD, outAccnosFilename, outMode, MPI_INFO_NULL, &outMPIAccnos);
236 if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPI); MPI_File_close(&outMPIAccnos); for (int j = 0; j < outputNames.size(); j++) { remove(outputNames[j].c_str()); } delete chimera; return 0; }
238 if (pid == 0) { //you are the root process
239 string outTemp = "For full window mapping info refer to " + mapInfo + "\n\n";
242 int length = outTemp.length();
243 char* buf2 = new char[length];
244 memcpy(buf2, outTemp.c_str(), length);
246 MPI_File_write_shared(outMPI, buf2, length, MPI_CHAR, &status);
249 MPIPos = m->setFilePosFasta(fastaFileNames[s], numSeqs); //fills MPIPos, returns numSeqs
251 //send file positions to all processes
252 for(int i = 1; i < processors; i++) {
253 MPI_Send(&numSeqs, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
254 MPI_Send(&MPIPos[0], (numSeqs+1), MPI_LONG, i, tag, MPI_COMM_WORLD);
257 //figure out how many sequences you have to align
258 numSeqsPerProcessor = numSeqs / processors;
259 int startIndex = pid * numSeqsPerProcessor;
260 if(pid == (processors - 1)){ numSeqsPerProcessor = numSeqs - pid * numSeqsPerProcessor; }
264 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPI, outMPIAccnos, MPIPos);
266 if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPI); MPI_File_close(&outMPIAccnos); remove(outputFileName.c_str()); remove(accnosFileName.c_str()); for (int j = 0; j < outputNames.size(); j++) { remove(outputNames[j].c_str()); } delete chimera; return 0; }
268 }else{ //you are a child process
269 MPI_Recv(&numSeqs, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
270 MPIPos.resize(numSeqs+1);
271 MPI_Recv(&MPIPos[0], (numSeqs+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status);
273 //figure out how many sequences you have to align
274 numSeqsPerProcessor = numSeqs / processors;
275 int startIndex = pid * numSeqsPerProcessor;
276 if(pid == (processors - 1)){ numSeqsPerProcessor = numSeqs - pid * numSeqsPerProcessor; }
280 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPI, outMPIAccnos, MPIPos);
282 if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPI); MPI_File_close(&outMPIAccnos); for (int j = 0; j < outputNames.size(); j++) { remove(outputNames[j].c_str()); } delete chimera; return 0; }
286 MPI_File_close(&inMPI);
287 MPI_File_close(&outMPI);
288 MPI_File_close(&outMPIAccnos);
290 MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
294 string tempHeader = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + maskfile + "ccode.chimeras.tempHeader";
295 m->openOutputFile(tempHeader, outHeader);
297 outHeader << "For full window mapping info refer to " << mapInfo << endl << endl;
301 vector<unsigned long int> positions = m->divideFile(fastaFileNames[s], processors);
303 for (int i = 0; i < (positions.size()-1); i++) {
304 lines.push_back(new linePair(positions[i], positions[(i+1)]));
308 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
311 numSeqs = driver(lines[0], outputFileName, fastaFileNames[s], accnosFileName);
313 if (m->control_pressed) { remove(outputFileName.c_str()); remove(tempHeader.c_str()); remove(accnosFileName.c_str()); for (int j = 0; j < outputNames.size(); j++) { remove(outputNames[j].c_str()); } for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear(); delete chimera; return 0; }
316 processIDS.resize(0);
318 numSeqs = createProcesses(outputFileName, fastaFileNames[s], accnosFileName);
320 rename((outputFileName + toString(processIDS[0]) + ".temp").c_str(), outputFileName.c_str());
321 rename((accnosFileName + toString(processIDS[0]) + ".temp").c_str(), accnosFileName.c_str());
323 //append output files
324 for(int i=1;i<processors;i++){
325 m->appendFiles((outputFileName + toString(processIDS[i]) + ".temp"), outputFileName);
326 remove((outputFileName + toString(processIDS[i]) + ".temp").c_str());
329 //append output files
330 for(int i=1;i<processors;i++){
331 m->appendFiles((accnosFileName + toString(processIDS[i]) + ".temp"), accnosFileName);
332 remove((accnosFileName + toString(processIDS[i]) + ".temp").c_str());
335 if (m->control_pressed) {
336 remove(outputFileName.c_str());
337 remove(accnosFileName.c_str());
338 for (int j = 0; j < outputNames.size(); j++) { remove(outputNames[j].c_str()); }
339 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
347 numSeqs = driver(lines[0], outputFileName, fastaFileNames[s], accnosFileName);
349 if (m->control_pressed) { remove(outputFileName.c_str()); remove(tempHeader.c_str()); remove(accnosFileName.c_str()); for (int j = 0; j < outputNames.size(); j++) { remove(outputNames[j].c_str()); } for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear(); delete chimera; return 0; }
353 m->appendFiles(outputFileName, tempHeader);
355 remove(outputFileName.c_str());
356 rename(tempHeader.c_str(), outputFileName.c_str());
361 outputNames.push_back(outputFileName);
362 outputNames.push_back(mapInfo);
363 outputNames.push_back(accnosFileName);
365 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
367 m->mothurOutEndLine(); m->mothurOut("It took " + toString(time(NULL) - start) + " secs to check " + toString(numSeqs) + " sequences."); m->mothurOutEndLine();
370 m->mothurOutEndLine();
371 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
372 for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
373 m->mothurOutEndLine();
378 catch(exception& e) {
379 m->errorOut(e, "ChimeraCcodeCommand", "execute");
383 //**********************************************************************************************************************
385 int ChimeraCcodeCommand::driver(linePair* filePos, string outputFName, string filename, string accnos){
388 m->openOutputFile(outputFName, out);
391 m->openOutputFile(accnos, out2);
394 m->openInputFile(filename, inFASTA);
396 inFASTA.seekg(filePos->start);
403 if (m->control_pressed) { return 1; }
405 Sequence* candidateSeq = new Sequence(inFASTA); m->gobble(inFASTA);
407 if (candidateSeq->getName() != "") { //incase there is a commented sequence at the end of a file
409 if (candidateSeq->getAligned().length() != templateSeqsLength) {
410 m->mothurOut(candidateSeq->getName() + " is not the same length as the template sequences. Skipping."); m->mothurOutEndLine();
413 chimera->getChimeras(candidateSeq);
415 if (m->control_pressed) { delete candidateSeq; return 1; }
418 chimera->print(out, out2);
424 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
425 unsigned long int pos = inFASTA.tellg();
426 if ((pos == -1) || (pos >= filePos->end)) { break; }
428 if (inFASTA.eof()) { break; }
432 if((count) % 100 == 0){ m->mothurOut("Processing sequence: " + toString(count)); m->mothurOutEndLine(); }
435 if((count) % 100 != 0){ m->mothurOut("Processing sequence: " + toString(count)); m->mothurOutEndLine(); }
443 catch(exception& e) {
444 m->errorOut(e, "ChimeraCcodeCommand", "driver");
448 //**********************************************************************************************************************
450 int ChimeraCcodeCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& outMPI, MPI_File& outAccMPI, vector<unsigned long int>& MPIPos){
455 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
457 for(int i=0;i<num;i++){
459 if (m->control_pressed) { return 0; }
462 int length = MPIPos[start+i+1] - MPIPos[start+i];
464 char* buf4 = new char[length];
466 MPI_File_read_at(inMPI, MPIPos[start+i], buf4, length, MPI_CHAR, &status);
468 string tempBuf = buf4;
469 if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); }
470 istringstream iss (tempBuf,istringstream::in);
473 Sequence* candidateSeq = new Sequence(iss); m->gobble(iss);
475 if (candidateSeq->getName() != "") { //incase there is a commented sequence at the end of a file
477 if (candidateSeq->getAligned().length() != templateSeqsLength) {
478 m->mothurOut(candidateSeq->getName() + " is not the same length as the template sequences. Skipping."); m->mothurOutEndLine();
481 chimera->getChimeras(candidateSeq);
483 if (m->control_pressed) { delete candidateSeq; return 1; }
486 bool isChimeric = chimera->print(outMPI, outAccMPI);
492 if((i+1) % 100 == 0){ cout << "Processing sequence: " << (i+1) << endl; m->mothurOutJustToLog("Processing sequence: " + toString(i+1) + "\n"); }
495 if(num % 100 != 0){ cout << "Processing sequence: " << num << endl; m->mothurOutJustToLog("Processing sequence: " + toString(num) + "\n"); }
500 catch(exception& e) {
501 m->errorOut(e, "ChimeraCcodeCommand", "driverMPI");
507 /**************************************************************************************************/
509 int ChimeraCcodeCommand::createProcesses(string outputFileName, string filename, string accnos) {
511 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
515 //loop through and create all the processes you want
516 while (process != processors) {
520 processIDS.push_back(pid); //create map from line number to pid so you can append files in correct order later
523 num = driver(lines[process], outputFileName + toString(getpid()) + ".temp", filename, accnos + toString(getpid()) + ".temp");
525 //pass numSeqs to parent
527 string tempFile = outputFileName + toString(getpid()) + ".num.temp";
528 m->openOutputFile(tempFile, out);
533 }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
536 //force parent to wait until all the processes are done
537 for (int i=0;i<processors;i++) {
538 int temp = processIDS[i];
542 for (int i = 0; i < processIDS.size(); i++) {
544 string tempFile = outputFileName + toString(processIDS[i]) + ".num.temp";
545 m->openInputFile(tempFile, in);
546 if (!in.eof()) { int tempNum = 0; in >> tempNum; num += tempNum; }
547 in.close(); remove(tempFile.c_str());
553 catch(exception& e) {
554 m->errorOut(e, "ChimeraCcodeCommand", "createProcesses");
558 //**********************************************************************************************************************