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;
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 = 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 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 = 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]; }
71 MPI_Comm_size(MPI_COMM_WORLD, &processors); //set processors to the number of mpi processes running
72 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
77 ableToOpen = openInputFile(fastaFileNames[i], in, "noerror");
79 //if you can't open it, try default location
80 if (ableToOpen == 1) {
81 if (m->getDefaultPath() != "") { //default path is set
82 string tryPath = m->getDefaultPath() + getSimpleName(fastaFileNames[i]);
83 m->mothurOut("Unable to open " + fastaFileNames[i] + ". Trying default " + tryPath); m->mothurOutEndLine();
84 ableToOpen = openInputFile(tryPath, in, "noerror");
85 fastaFileNames[i] = tryPath;
91 for (int j = 1; j < processors; j++) {
92 MPI_Send(&ableToOpen, 1, MPI_INT, j, 2001, MPI_COMM_WORLD);
96 MPI_Recv(&ableToOpen, 1, MPI_INT, 0, 2001, MPI_COMM_WORLD, &status);
101 if (ableToOpen == 1) {
102 m->mothurOut("Unable to open " + fastaFileNames[i] + ". It will be disregarded."); m->mothurOutEndLine();
103 //erase from file list
104 fastaFileNames.erase(fastaFileNames.begin()+i);
109 //make sure there is at least one valid file left
110 if (fastaFileNames.size() == 0) { m->mothurOut("no valid files."); m->mothurOutEndLine(); abort = true; }
113 //if the user changes the output directory command factory will send this info to us in the output parameter
114 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){
116 outputDir += hasPath(fastafile); //if user entered a file with a path then preserve it
119 templatefile = validParameter.validFile(parameters, "template", true);
120 if (templatefile == "not open") { abort = true; }
121 else if (templatefile == "not found") { templatefile = ""; m->mothurOut("template is a required parameter for the chimera.ccode command."); m->mothurOutEndLine(); abort = true; }
123 maskfile = validParameter.validFile(parameters, "mask", false);
124 if (maskfile == "not found") { maskfile = ""; }
125 else if (maskfile != "default") {
126 if (inputDir != "") {
127 string path = hasPath(maskfile);
128 //if the user has not given a path then, add inputdir. else leave path alone.
129 if (path == "") { maskfile = inputDir + maskfile; }
133 int ableToOpen = openInputFile(maskfile, in);
134 if (ableToOpen == 1) { abort = true; }
139 temp = validParameter.validFile(parameters, "filter", false); if (temp == "not found") { temp = "F"; }
140 filter = isTrue(temp);
142 temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found") { temp = "1"; }
143 convert(temp, processors);
145 temp = validParameter.validFile(parameters, "window", false); if (temp == "not found") { temp = "0"; }
146 convert(temp, window);
148 temp = validParameter.validFile(parameters, "numwanted", false); if (temp == "not found") { temp = "20"; }
149 convert(temp, numwanted);
153 catch(exception& e) {
154 m->errorOut(e, "ChimeraCcodeCommand", "ChimeraCcodeCommand");
158 //**********************************************************************************************************************
160 void ChimeraCcodeCommand::help(){
163 m->mothurOut("The chimera.ccode command reads a fastafile and templatefile and outputs potentially chimeric sequences.\n");
164 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");
165 m->mothurOut("The chimera.ccode command parameters are fasta, template, filter, mask, processors, window and numwanted.\n");
166 m->mothurOut("The fasta parameter allows you to enter the fasta file containing your potentially chimeric sequences, and is required. \n");
167 m->mothurOut("You may enter multiple fasta files by separating their names with dashes. ie. fasta=abrecovery.fasta-amzon.fasta \n");
168 m->mothurOut("The template parameter allows you to enter a template file containing known non-chimeric sequences, and is required. \n");
169 m->mothurOut("The filter parameter allows you to specify if you would like to apply a vertical and 50% soft filter. \n");
170 m->mothurOut("The processors parameter allows you to specify how many processors you would like to use. The default is 1. \n");
172 m->mothurOut("When using MPI, the processors parameter is set to the number of MPI processes running. \n");
174 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");
175 m->mothurOut("The window parameter allows you to specify the window size for searching for chimeras. \n");
176 m->mothurOut("The numwanted parameter allows you to specify how many sequences you would each query sequence compared with.\n");
177 m->mothurOut("The chimera.ccode command should be in the following format: \n");
178 m->mothurOut("chimera.ccode(fasta=yourFastaFile, template=yourTemplate) \n");
179 m->mothurOut("Example: chimera.ccode(fasta=AD.align, template=core_set_aligned.imputed.fasta) \n");
180 m->mothurOut("Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFastaFile).\n\n");
182 catch(exception& e) {
183 m->errorOut(e, "ChimeraCcodeCommand", "help");
188 //***************************************************************************************************************
190 ChimeraCcodeCommand::~ChimeraCcodeCommand(){ /* do nothing */ }
192 //***************************************************************************************************************
194 int ChimeraCcodeCommand::execute(){
197 if (abort == true) { return 0; }
199 for (int s = 0; s < fastaFileNames.size(); s++) {
201 m->mothurOut("Checking sequences from " + fastaFileNames[s] + " ..." ); m->mothurOutEndLine();
203 int start = time(NULL);
206 if (maskfile == "default") { m->mothurOut("I am using the default 236627 EU009184.1 Shigella dysenteriae str. FBD013."); m->mothurOutEndLine(); }
208 chimera = new Ccode(fastaFileNames[s], templatefile, filter, maskfile, window, numwanted, outputDir);
210 //is your template aligned?
211 if (chimera->getUnaligned()) { m->mothurOut("Your template sequences are different lengths, please correct."); m->mothurOutEndLine(); delete chimera; return 0; }
212 templateSeqsLength = chimera->getLength();
214 string outputFileName, accnosFileName;
215 if (maskfile != "") {
216 outputFileName = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + maskfile + ".ccode.chimeras";
217 accnosFileName = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + maskfile + ".ccode.accnos";
219 outputFileName = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + "ccode.chimeras";
220 accnosFileName = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + "ccode.accnos";
223 string mapInfo = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + "mapinfo";
225 if (m->control_pressed) { delete chimera; for (int j = 0; j < outputNames.size(); j++) { remove(outputNames[j].c_str()); } return 0; }
229 int pid, end, numSeqsPerProcessor;
234 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
235 MPI_Comm_size(MPI_COMM_WORLD, &processors);
239 MPI_File outMPIAccnos;
241 int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
242 int inMode=MPI_MODE_RDONLY;
244 char outFilename[1024];
245 strcpy(outFilename, outputFileName.c_str());
247 char outAccnosFilename[1024];
248 strcpy(outAccnosFilename, accnosFileName.c_str());
250 char inFileName[1024];
251 strcpy(inFileName, fastaFileNames[s].c_str());
253 MPI_File_open(MPI_COMM_WORLD, inFileName, inMode, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer
254 MPI_File_open(MPI_COMM_WORLD, outFilename, outMode, MPI_INFO_NULL, &outMPI);
255 MPI_File_open(MPI_COMM_WORLD, outAccnosFilename, outMode, MPI_INFO_NULL, &outMPIAccnos);
257 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; }
259 if (pid == 0) { //you are the root process
260 string outTemp = "For full window mapping info refer to " + mapInfo + "\n\n";
263 int length = outTemp.length();
264 char* buf2 = new char[length];
265 memcpy(buf2, outTemp.c_str(), length);
267 MPI_File_write_shared(outMPI, buf2, length, MPI_CHAR, &status);
270 MPIPos = setFilePosFasta(fastaFileNames[s], numSeqs); //fills MPIPos, returns numSeqs
272 //send file positions to all processes
273 for(int i = 1; i < processors; i++) {
274 MPI_Send(&numSeqs, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
275 MPI_Send(&MPIPos[0], (numSeqs+1), MPI_LONG, i, tag, MPI_COMM_WORLD);
278 //figure out how many sequences you have to align
279 numSeqsPerProcessor = numSeqs / processors;
280 int startIndex = pid * numSeqsPerProcessor;
281 if(pid == (processors - 1)){ numSeqsPerProcessor = numSeqs - pid * numSeqsPerProcessor; }
285 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPI, outMPIAccnos, MPIPos);
287 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; }
289 }else{ //you are a child process
290 MPI_Recv(&numSeqs, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
291 MPIPos.resize(numSeqs+1);
292 MPI_Recv(&MPIPos[0], (numSeqs+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status);
294 //figure out how many sequences you have to align
295 numSeqsPerProcessor = numSeqs / processors;
296 int startIndex = pid * numSeqsPerProcessor;
297 if(pid == (processors - 1)){ numSeqsPerProcessor = numSeqs - pid * numSeqsPerProcessor; }
301 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPI, outMPIAccnos, MPIPos);
303 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; }
307 MPI_File_close(&inMPI);
308 MPI_File_close(&outMPI);
309 MPI_File_close(&outMPIAccnos);
311 MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
315 string tempHeader = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + maskfile + "ccode.chimeras.tempHeader";
316 openOutputFile(tempHeader, outHeader);
318 outHeader << "For full window mapping info refer to " << mapInfo << endl << endl;
323 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
326 openInputFile(fastaFileNames[s], inFASTA);
327 getNumSeqs(inFASTA, numSeqs);
330 lines.push_back(new linePair(0, numSeqs));
332 driver(lines[0], outputFileName, fastaFileNames[s], accnosFileName);
334 if (m->control_pressed) {
335 remove(outputFileName.c_str());
336 remove(tempHeader.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();
345 vector<unsigned long int> positions;
346 processIDS.resize(0);
349 openInputFile(fastaFileNames[s], inFASTA);
352 while(!inFASTA.eof()){
353 input = getline(inFASTA);
354 if (input.length() != 0) {
355 if(input[0] == '>'){ unsigned long int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); }
360 numSeqs = positions.size();
362 int numSeqsPerProcessor = numSeqs / processors;
364 for (int i = 0; i < processors; i++) {
365 unsigned long int startPos = positions[ i * numSeqsPerProcessor ];
366 if(i == processors - 1){
367 numSeqsPerProcessor = numSeqs - i * numSeqsPerProcessor;
369 lines.push_back(new linePair(startPos, numSeqsPerProcessor));
373 createProcesses(outputFileName, fastaFileNames[s], accnosFileName);
375 rename((outputFileName + toString(processIDS[0]) + ".temp").c_str(), outputFileName.c_str());
376 rename((accnosFileName + toString(processIDS[0]) + ".temp").c_str(), accnosFileName.c_str());
378 //append output files
379 for(int i=1;i<processors;i++){
380 appendFiles((outputFileName + toString(processIDS[i]) + ".temp"), outputFileName);
381 remove((outputFileName + toString(processIDS[i]) + ".temp").c_str());
384 //append output files
385 for(int i=1;i<processors;i++){
386 appendFiles((accnosFileName + toString(processIDS[i]) + ".temp"), accnosFileName);
387 remove((accnosFileName + toString(processIDS[i]) + ".temp").c_str());
390 if (m->control_pressed) {
391 remove(outputFileName.c_str());
392 remove(accnosFileName.c_str());
393 for (int j = 0; j < outputNames.size(); j++) { remove(outputNames[j].c_str()); }
394 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
403 openInputFile(fastaFileNames[s], inFASTA);
404 getNumSeqs(inFASTA, numSeqs);
406 lines.push_back(new linePair(0, numSeqs));
408 driver(lines[0], outputFileName, fastaFileNames[s], accnosFileName);
410 if (m->control_pressed) {
411 remove(outputFileName.c_str());
412 remove(tempHeader.c_str());
413 remove(accnosFileName.c_str());
414 for (int j = 0; j < outputNames.size(); j++) { remove(outputNames[j].c_str()); }
415 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
422 appendFiles(outputFileName, tempHeader);
424 remove(outputFileName.c_str());
425 rename(tempHeader.c_str(), outputFileName.c_str());
430 outputNames.push_back(outputFileName);
431 outputNames.push_back(mapInfo);
432 outputNames.push_back(accnosFileName);
434 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
436 m->mothurOutEndLine(); m->mothurOut("It took " + toString(time(NULL) - start) + " secs to check " + toString(numSeqs) + " sequences."); m->mothurOutEndLine();
439 m->mothurOutEndLine();
440 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
441 for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
442 m->mothurOutEndLine();
447 catch(exception& e) {
448 m->errorOut(e, "ChimeraCcodeCommand", "execute");
452 //**********************************************************************************************************************
454 int ChimeraCcodeCommand::driver(linePair* line, string outputFName, string filename, string accnos){
457 openOutputFile(outputFName, out);
460 openOutputFile(accnos, out2);
463 openInputFile(filename, inFASTA);
465 inFASTA.seekg(line->start);
467 for(int i=0;i<line->numSeqs;i++){
469 if (m->control_pressed) { return 1; }
471 Sequence* candidateSeq = new Sequence(inFASTA); gobble(inFASTA);
473 if (candidateSeq->getName() != "") { //incase there is a commented sequence at the end of a file
475 if (candidateSeq->getAligned().length() != templateSeqsLength) {
476 m->mothurOut(candidateSeq->getName() + " is not the same length as the template sequences. Skipping."); m->mothurOutEndLine();
479 chimera->getChimeras(candidateSeq);
481 if (m->control_pressed) { delete candidateSeq; return 1; }
484 chimera->print(out, out2);
490 if((i+1) % 100 == 0){ m->mothurOut("Processing sequence: " + toString(i+1)); m->mothurOutEndLine(); }
493 if((line->numSeqs) % 100 != 0){ m->mothurOut("Processing sequence: " + toString(line->numSeqs)); m->mothurOutEndLine(); }
501 catch(exception& e) {
502 m->errorOut(e, "ChimeraCcodeCommand", "driver");
506 //**********************************************************************************************************************
508 int ChimeraCcodeCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& outMPI, MPI_File& outAccMPI, vector<long>& MPIPos){
513 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
515 for(int i=0;i<num;i++){
517 if (m->control_pressed) { return 0; }
520 int length = MPIPos[start+i+1] - MPIPos[start+i];
522 char* buf4 = new char[length];
524 MPI_File_read_at(inMPI, MPIPos[start+i], buf4, length, MPI_CHAR, &status);
526 string tempBuf = buf4;
527 if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); }
528 istringstream iss (tempBuf,istringstream::in);
531 Sequence* candidateSeq = new Sequence(iss); gobble(iss);
533 if (candidateSeq->getName() != "") { //incase there is a commented sequence at the end of a file
535 if (candidateSeq->getAligned().length() != templateSeqsLength) {
536 m->mothurOut(candidateSeq->getName() + " is not the same length as the template sequences. Skipping."); m->mothurOutEndLine();
539 chimera->getChimeras(candidateSeq);
541 if (m->control_pressed) { delete candidateSeq; return 1; }
544 bool isChimeric = chimera->print(outMPI, outAccMPI);
550 if((i+1) % 100 == 0){ cout << "Processing sequence: " << (i+1) << endl; m->mothurOutJustToLog("Processing sequence: " + toString(i+1) + "\n"); }
553 if(num % 100 != 0){ cout << "Processing sequence: " << num << endl; m->mothurOutJustToLog("Processing sequence: " + toString(num) + "\n"); }
558 catch(exception& e) {
559 m->errorOut(e, "ChimeraCcodeCommand", "driverMPI");
565 /**************************************************************************************************/
567 int ChimeraCcodeCommand::createProcesses(string outputFileName, string filename, string accnos) {
569 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
571 // processIDS.resize(0);
573 //loop through and create all the processes you want
574 while (process != processors) {
578 processIDS.push_back(pid); //create map from line number to pid so you can append files in correct order later
581 driver(lines[process], outputFileName + toString(getpid()) + ".temp", filename, accnos + toString(getpid()) + ".temp");
583 }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
586 //force parent to wait until all the processes are done
587 for (int i=0;i<processors;i++) {
588 int temp = processIDS[i];
595 catch(exception& e) {
596 m->errorOut(e, "ChimeraCcodeCommand", "createProcesses");
600 //**********************************************************************************************************************