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);
81 for (int j = 1; j < processors; j++) {
82 MPI_Send(&ableToOpen, 1, MPI_INT, j, 2001, MPI_COMM_WORLD);
86 MPI_Recv(&ableToOpen, 1, MPI_INT, 0, 2001, MPI_COMM_WORLD, &status);
91 if (ableToOpen == 1) {
92 m->mothurOut(fastaFileNames[i] + " will be disregarded."); m->mothurOutEndLine();
93 //erase from file list
94 fastaFileNames.erase(fastaFileNames.begin()+i);
99 //make sure there is at least one valid file left
100 if (fastaFileNames.size() == 0) { m->mothurOut("no valid files."); m->mothurOutEndLine(); abort = true; }
103 //if the user changes the output directory command factory will send this info to us in the output parameter
104 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){
106 outputDir += hasPath(fastafile); //if user entered a file with a path then preserve it
109 templatefile = validParameter.validFile(parameters, "template", true);
110 if (templatefile == "not open") { abort = true; }
111 else if (templatefile == "not found") { templatefile = ""; m->mothurOut("template is a required parameter for the chimera.ccode command."); m->mothurOutEndLine(); abort = true; }
113 maskfile = validParameter.validFile(parameters, "mask", false);
114 if (maskfile == "not found") { maskfile = ""; }
115 else if (maskfile != "default") {
116 if (inputDir != "") {
117 string path = hasPath(maskfile);
118 //if the user has not given a path then, add inputdir. else leave path alone.
119 if (path == "") { maskfile = inputDir + maskfile; }
123 int ableToOpen = openInputFile(maskfile, in);
124 if (ableToOpen == 1) { abort = true; }
129 temp = validParameter.validFile(parameters, "filter", false); if (temp == "not found") { temp = "F"; }
130 filter = isTrue(temp);
132 temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found") { temp = "1"; }
133 convert(temp, processors);
135 temp = validParameter.validFile(parameters, "window", false); if (temp == "not found") { temp = "0"; }
136 convert(temp, window);
138 temp = validParameter.validFile(parameters, "numwanted", false); if (temp == "not found") { temp = "20"; }
139 convert(temp, numwanted);
143 catch(exception& e) {
144 m->errorOut(e, "ChimeraCcodeCommand", "ChimeraCcodeCommand");
148 //**********************************************************************************************************************
150 void ChimeraCcodeCommand::help(){
153 m->mothurOut("The chimera.ccode command reads a fastafile and templatefile and outputs potentially chimeric sequences.\n");
154 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");
155 m->mothurOut("The chimera.ccode command parameters are fasta, template, filter, mask, processors, window and numwanted.\n");
156 m->mothurOut("The fasta parameter allows you to enter the fasta file containing your potentially chimeric sequences, and is required. \n");
157 m->mothurOut("You may enter multiple fasta files by separating their names with dashes. ie. fasta=abrecovery.fasta-amzon.fasta \n");
158 m->mothurOut("The template parameter allows you to enter a template file containing known non-chimeric sequences, and is required. \n");
159 m->mothurOut("The filter parameter allows you to specify if you would like to apply a vertical and 50% soft filter. \n");
160 m->mothurOut("The processors parameter allows you to specify how many processors you would like to use. The default is 1. \n");
162 m->mothurOut("When using MPI, the processors parameter is set to the number of MPI processes running. \n");
164 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");
165 m->mothurOut("The window parameter allows you to specify the window size for searching for chimeras. \n");
166 m->mothurOut("The numwanted parameter allows you to specify how many sequences you would each query sequence compared with.\n");
167 m->mothurOut("The chimera.ccode command should be in the following format: \n");
168 m->mothurOut("chimera.ccode(fasta=yourFastaFile, template=yourTemplate) \n");
169 m->mothurOut("Example: chimera.ccode(fasta=AD.align, template=core_set_aligned.imputed.fasta) \n");
170 m->mothurOut("Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFastaFile).\n\n");
172 catch(exception& e) {
173 m->errorOut(e, "ChimeraCcodeCommand", "help");
178 //***************************************************************************************************************
180 ChimeraCcodeCommand::~ChimeraCcodeCommand(){ /* do nothing */ }
182 //***************************************************************************************************************
184 int ChimeraCcodeCommand::execute(){
187 if (abort == true) { return 0; }
189 for (int s = 0; s < fastaFileNames.size(); s++) {
191 m->mothurOut("Checking sequences from " + fastaFileNames[s] + " ..." ); m->mothurOutEndLine();
193 int start = time(NULL);
196 if (maskfile == "default") { m->mothurOut("I am using the default 236627 EU009184.1 Shigella dysenteriae str. FBD013."); m->mothurOutEndLine(); }
198 chimera = new Ccode(fastaFileNames[s], templatefile, filter, maskfile, window, numwanted, outputDir);
200 //is your template aligned?
201 if (chimera->getUnaligned()) { m->mothurOut("Your template sequences are different lengths, please correct."); m->mothurOutEndLine(); delete chimera; return 0; }
202 templateSeqsLength = chimera->getLength();
204 string outputFileName, accnosFileName;
205 if (maskfile != "") {
206 outputFileName = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + maskfile + ".ccode.chimeras";
207 accnosFileName = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + maskfile + ".ccode.accnos";
209 outputFileName = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + "ccode.chimeras";
210 accnosFileName = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + "ccode.accnos";
213 string mapInfo = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + "mapinfo";
214 bool hasAccnos = true;
216 if (m->control_pressed) { delete chimera; for (int j = 0; j < outputNames.size(); j++) { remove(outputNames[j].c_str()); } return 0; }
220 int pid, end, numSeqsPerProcessor;
223 MPIWroteAccnos = false;
226 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
227 MPI_Comm_size(MPI_COMM_WORLD, &processors);
231 MPI_File outMPIAccnos;
233 int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
234 int inMode=MPI_MODE_RDONLY;
236 char outFilename[1024];
237 strcpy(outFilename, outputFileName.c_str());
239 char outAccnosFilename[1024];
240 strcpy(outAccnosFilename, accnosFileName.c_str());
242 char inFileName[1024];
243 strcpy(inFileName, fastaFileNames[s].c_str());
245 MPI_File_open(MPI_COMM_WORLD, inFileName, inMode, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer
246 MPI_File_open(MPI_COMM_WORLD, outFilename, outMode, MPI_INFO_NULL, &outMPI);
247 MPI_File_open(MPI_COMM_WORLD, outAccnosFilename, outMode, MPI_INFO_NULL, &outMPIAccnos);
249 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; }
251 if (pid == 0) { //you are the root process
252 string outTemp = "For full window mapping info refer to " + mapInfo + "\n\n";
255 int length = outTemp.length();
256 char* buf2 = new char[length];
257 memcpy(buf2, outTemp.c_str(), length);
259 MPI_File_write_shared(outMPI, buf2, length, MPI_CHAR, &status);
262 MPIPos = setFilePosFasta(fastaFileNames[s], numSeqs); //fills MPIPos, returns numSeqs
264 //send file positions to all processes
265 for(int i = 1; i < processors; i++) {
266 MPI_Send(&numSeqs, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
267 MPI_Send(&MPIPos[0], (numSeqs+1), MPI_LONG, i, tag, MPI_COMM_WORLD);
270 //figure out how many sequences you have to align
271 numSeqsPerProcessor = numSeqs / processors;
272 int startIndex = pid * numSeqsPerProcessor;
273 if(pid == (processors - 1)){ numSeqsPerProcessor = numSeqs - pid * numSeqsPerProcessor; }
277 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPI, outMPIAccnos, MPIPos);
279 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; }
281 for (int i = 1; i < processors; i++) {
283 MPI_Recv(&tempResult, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status);
284 if (tempResult != 0) { MPIWroteAccnos = true; }
286 }else{ //you are a child process
287 MPI_Recv(&numSeqs, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
288 MPIPos.resize(numSeqs+1);
289 MPI_Recv(&MPIPos[0], (numSeqs+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status);
291 //figure out how many sequences you have to align
292 numSeqsPerProcessor = numSeqs / processors;
293 int startIndex = pid * numSeqsPerProcessor;
294 if(pid == (processors - 1)){ numSeqsPerProcessor = numSeqs - pid * numSeqsPerProcessor; }
298 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPI, outMPIAccnos, MPIPos);
300 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; }
302 MPI_Send(&MPIWroteAccnos, 1, MPI_INT, 0, tag, MPI_COMM_WORLD);
306 MPI_File_close(&inMPI);
307 MPI_File_close(&outMPI);
308 MPI_File_close(&outMPIAccnos);
310 MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
312 //delete accnos file if blank
314 if (!MPIWroteAccnos) {
316 remove(accnosFileName.c_str());
322 string tempHeader = outputDir + getRootName(getSimpleName(fastaFileNames[s])) + maskfile + "ccode.chimeras.tempHeader";
323 openOutputFile(tempHeader, outHeader);
325 outHeader << "For full window mapping info refer to " << mapInfo << endl << endl;
330 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
333 openInputFile(fastaFileNames[s], inFASTA);
334 getNumSeqs(inFASTA, numSeqs);
337 lines.push_back(new linePair(0, numSeqs));
339 driver(lines[0], outputFileName, fastaFileNames[s], accnosFileName);
341 if (m->control_pressed) {
342 remove(outputFileName.c_str());
343 remove(tempHeader.c_str());
344 remove(accnosFileName.c_str());
345 for (int j = 0; j < outputNames.size(); j++) { remove(outputNames[j].c_str()); }
346 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
351 //delete accnos file if its blank
352 if (isBlank(accnosFileName)) { remove(accnosFileName.c_str()); hasAccnos = false; }
355 vector<int> positions;
356 processIDS.resize(0);
359 openInputFile(fastaFileNames[s], inFASTA);
362 while(!inFASTA.eof()){
363 input = getline(inFASTA);
364 if (input.length() != 0) {
365 if(input[0] == '>'){ long int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); }
370 numSeqs = positions.size();
372 int numSeqsPerProcessor = numSeqs / processors;
374 for (int i = 0; i < processors; i++) {
375 long int startPos = positions[ i * numSeqsPerProcessor ];
376 if(i == processors - 1){
377 numSeqsPerProcessor = numSeqs - i * numSeqsPerProcessor;
379 lines.push_back(new linePair(startPos, numSeqsPerProcessor));
383 createProcesses(outputFileName, fastaFileNames[s], accnosFileName);
385 rename((outputFileName + toString(processIDS[0]) + ".temp").c_str(), outputFileName.c_str());
387 //append output files
388 for(int i=1;i<processors;i++){
389 appendFiles((outputFileName + toString(processIDS[i]) + ".temp"), outputFileName);
390 remove((outputFileName + toString(processIDS[i]) + ".temp").c_str());
393 vector<string> nonBlankAccnosFiles;
394 //delete blank accnos files generated with multiple processes
395 for(int i=0;i<processors;i++){
396 if (!(isBlank(accnosFileName + toString(processIDS[i]) + ".temp"))) {
397 nonBlankAccnosFiles.push_back(accnosFileName + toString(processIDS[i]) + ".temp");
398 }else { remove((accnosFileName + toString(processIDS[i]) + ".temp").c_str()); }
401 //append accnos files
402 if (nonBlankAccnosFiles.size() != 0) {
403 rename(nonBlankAccnosFiles[0].c_str(), accnosFileName.c_str());
405 for (int h=1; h < nonBlankAccnosFiles.size(); h++) {
406 appendFiles(nonBlankAccnosFiles[h], accnosFileName);
407 remove(nonBlankAccnosFiles[h].c_str());
409 }else{ hasAccnos = false; }
411 if (m->control_pressed) {
412 remove(outputFileName.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();
424 openInputFile(fastaFileNames[s], inFASTA);
425 getNumSeqs(inFASTA, numSeqs);
427 lines.push_back(new linePair(0, numSeqs));
429 driver(lines[0], outputFileName, fastaFileNames[s], accnosFileName);
431 if (m->control_pressed) {
432 remove(outputFileName.c_str());
433 remove(tempHeader.c_str());
434 remove(accnosFileName.c_str());
435 for (int j = 0; j < outputNames.size(); j++) { remove(outputNames[j].c_str()); }
436 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
441 //delete accnos file if its blank
442 if (isBlank(accnosFileName)) { remove(accnosFileName.c_str()); hasAccnos = false; }
445 appendFiles(outputFileName, tempHeader);
447 remove(outputFileName.c_str());
448 rename(tempHeader.c_str(), outputFileName.c_str());
453 outputNames.push_back(outputFileName);
454 outputNames.push_back(mapInfo);
455 if (hasAccnos) { outputNames.push_back(accnosFileName); }
457 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
459 m->mothurOutEndLine(); m->mothurOut("It took " + toString(time(NULL) - start) + " secs to check " + toString(numSeqs) + " sequences."); m->mothurOutEndLine();
462 m->mothurOutEndLine();
463 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
464 for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
465 m->mothurOutEndLine();
470 catch(exception& e) {
471 m->errorOut(e, "ChimeraCcodeCommand", "execute");
475 //**********************************************************************************************************************
477 int ChimeraCcodeCommand::driver(linePair* line, string outputFName, string filename, string accnos){
480 openOutputFile(outputFName, out);
483 openOutputFile(accnos, out2);
486 openInputFile(filename, inFASTA);
488 inFASTA.seekg(line->start);
490 for(int i=0;i<line->numSeqs;i++){
492 if (m->control_pressed) { return 1; }
494 Sequence* candidateSeq = new Sequence(inFASTA); gobble(inFASTA);
496 if (candidateSeq->getName() != "") { //incase there is a commented sequence at the end of a file
498 if (candidateSeq->getAligned().length() != templateSeqsLength) {
499 m->mothurOut(candidateSeq->getName() + " is not the same length as the template sequences. Skipping."); m->mothurOutEndLine();
502 chimera->getChimeras(candidateSeq);
504 if (m->control_pressed) { delete candidateSeq; return 1; }
507 chimera->print(out, out2);
513 if((i+1) % 100 == 0){ m->mothurOut("Processing sequence: " + toString(i+1)); m->mothurOutEndLine(); }
516 if((line->numSeqs) % 100 != 0){ m->mothurOut("Processing sequence: " + toString(line->numSeqs)); m->mothurOutEndLine(); }
524 catch(exception& e) {
525 m->errorOut(e, "ChimeraCcodeCommand", "driver");
529 //**********************************************************************************************************************
531 int ChimeraCcodeCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& outMPI, MPI_File& outAccMPI, vector<long>& MPIPos){
536 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
538 for(int i=0;i<num;i++){
540 if (m->control_pressed) { return 0; }
543 int length = MPIPos[start+i+1] - MPIPos[start+i];
545 char* buf4 = new char[length];
547 MPI_File_read_at(inMPI, MPIPos[start+i], buf4, length, MPI_CHAR, &status);
549 string tempBuf = buf4;
550 if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); }
551 istringstream iss (tempBuf,istringstream::in);
554 Sequence* candidateSeq = new Sequence(iss); gobble(iss);
556 if (candidateSeq->getName() != "") { //incase there is a commented sequence at the end of a file
558 if (candidateSeq->getAligned().length() != templateSeqsLength) {
559 m->mothurOut(candidateSeq->getName() + " is not the same length as the template sequences. Skipping."); m->mothurOutEndLine();
562 chimera->getChimeras(candidateSeq);
564 if (m->control_pressed) { delete candidateSeq; return 1; }
567 bool isChimeric = chimera->print(outMPI, outAccMPI);
568 if (isChimeric) { MPIWroteAccnos = true; }
574 if((i+1) % 100 == 0){ cout << "Processing sequence: " << (i+1) << endl; m->mothurOutJustToLog("Processing sequence: " + toString(i+1) + "\n"); }
577 if(num % 100 != 0){ cout << "Processing sequence: " << num << endl; m->mothurOutJustToLog("Processing sequence: " + toString(num) + "\n"); }
582 catch(exception& e) {
583 m->errorOut(e, "ChimeraCcodeCommand", "driverMPI");
589 /**************************************************************************************************/
591 int ChimeraCcodeCommand::createProcesses(string outputFileName, string filename, string accnos) {
593 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
595 // processIDS.resize(0);
597 //loop through and create all the processes you want
598 while (process != processors) {
602 processIDS.push_back(pid); //create map from line number to pid so you can append files in correct order later
605 driver(lines[process], outputFileName + toString(getpid()) + ".temp", filename, accnos + toString(getpid()) + ".temp");
607 }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
610 //force parent to wait until all the processes are done
611 for (int i=0;i<processors;i++) {
612 int temp = processIDS[i];
619 catch(exception& e) {
620 m->errorOut(e, "ChimeraCcodeCommand", "createProcesses");
624 //**********************************************************************************************************************