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("fasta");
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["fasta"] = inputDir + it->second; }
51 it = parameters.find("template");
52 //user has given a template file
53 if(it != parameters.end()){
54 path = hasPath(it->second);
55 //if the user has not given a path then, add inputdir. else leave path alone.
56 if (path == "") { parameters["template"] = inputDir + it->second; }
61 //check for required parameters
62 fastafile = validParameter.validFile(parameters, "fasta", true);
63 if (fastafile == "not open") { abort = true; }
64 else if (fastafile == "not found") { fastafile = ""; m->mothurOut("fasta is a required parameter for the chimera.ccode command."); m->mothurOutEndLine(); abort = true; }
66 //if the user changes the output directory command factory will send this info to us in the output parameter
67 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){
69 outputDir += hasPath(fastafile); //if user entered a file with a path then preserve it
72 templatefile = validParameter.validFile(parameters, "template", true);
73 if (templatefile == "not open") { abort = true; }
74 else if (templatefile == "not found") { templatefile = ""; m->mothurOut("template is a required parameter for the chimera.ccode command."); m->mothurOutEndLine(); abort = true; }
76 maskfile = validParameter.validFile(parameters, "mask", false);
77 if (maskfile == "not found") { maskfile = ""; }
78 else if (maskfile != "default") {
80 string path = hasPath(maskfile);
81 //if the user has not given a path then, add inputdir. else leave path alone.
82 if (path == "") { maskfile = inputDir + maskfile; }
86 int ableToOpen = openInputFile(maskfile, in);
87 if (ableToOpen == 1) { abort = true; }
92 temp = validParameter.validFile(parameters, "filter", false); if (temp == "not found") { temp = "F"; }
93 filter = isTrue(temp);
95 temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found") { temp = "1"; }
96 convert(temp, processors);
98 temp = validParameter.validFile(parameters, "window", false); if (temp == "not found") { temp = "0"; }
99 convert(temp, window);
101 temp = validParameter.validFile(parameters, "numwanted", false); if (temp == "not found") { temp = "20"; }
102 convert(temp, numwanted);
106 catch(exception& e) {
107 m->errorOut(e, "ChimeraCcodeCommand", "ChimeraCcodeCommand");
111 //**********************************************************************************************************************
113 void ChimeraCcodeCommand::help(){
116 m->mothurOut("The chimera.ccode command reads a fastafile and templatefile and outputs potentially chimeric sequences.\n");
117 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");
118 m->mothurOut("The chimera.ccode command parameters are fasta, template, filter, mask, processors, window and numwanted.\n");
119 m->mothurOut("The fasta parameter allows you to enter the fasta file containing your potentially chimeric sequences, and is required. \n");
120 m->mothurOut("The template parameter allows you to enter a template file containing known non-chimeric sequences, and is required. \n");
121 m->mothurOut("The filter parameter allows you to specify if you would like to apply a vertical and 50% soft filter. \n");
122 m->mothurOut("The processors parameter allows you to specify how many processors you would like to use. The default is 1. \n");
124 m->mothurOut("When using MPI, the processors parameter is set to the number of MPI processes running. \n");
126 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");
127 m->mothurOut("The window parameter allows you to specify the window size for searching for chimeras. \n");
128 m->mothurOut("The numwanted parameter allows you to specify how many sequences you would each query sequence compared with.\n");
129 m->mothurOut("The chimera.ccode command should be in the following format: \n");
130 m->mothurOut("chimera.ccode(fasta=yourFastaFile, template=yourTemplate) \n");
131 m->mothurOut("Example: chimera.seqs(fasta=AD.align, template=core_set_aligned.imputed.fasta) \n");
132 m->mothurOut("Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFastaFile).\n\n");
134 catch(exception& e) {
135 m->errorOut(e, "ChimeraCcodeCommand", "help");
140 //***************************************************************************************************************
142 ChimeraCcodeCommand::~ChimeraCcodeCommand(){ /* do nothing */ }
144 //***************************************************************************************************************
146 int ChimeraCcodeCommand::execute(){
149 if (abort == true) { return 0; }
151 int start = time(NULL);
154 if (maskfile == "default") { m->mothurOut("I am using the default 236627 EU009184.1 Shigella dysenteriae str. FBD013."); m->mothurOutEndLine(); }
156 chimera = new Ccode(fastafile, templatefile, filter, maskfile, window, numwanted, outputDir);
158 //is your template aligned?
159 if (chimera->getUnaligned()) { m->mothurOut("Your template sequences are different lengths, please correct."); m->mothurOutEndLine(); delete chimera; return 0; }
160 templateSeqsLength = chimera->getLength();
162 string outputFileName, accnosFileName;
163 if (maskfile != "") {
164 outputFileName = outputDir + getRootName(getSimpleName(fastafile)) + maskfile + ".ccode.chimeras";
165 accnosFileName = outputDir + getRootName(getSimpleName(fastafile)) + maskfile + ".ccode.accnos";
167 outputFileName = outputDir + getRootName(getSimpleName(fastafile)) + "ccode.chimeras";
168 accnosFileName = outputDir + getRootName(getSimpleName(fastafile)) + "ccode.accnos";
171 string mapInfo = outputDir + getRootName(getSimpleName(fastafile)) + "mapinfo";
172 bool hasAccnos = true;
174 if (m->control_pressed) { delete chimera; return 0; }
178 int pid, end, numSeqsPerProcessor;
181 MPIWroteAccnos = false;
184 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
185 MPI_Comm_size(MPI_COMM_WORLD, &processors);
189 MPI_File outMPIAccnos;
191 int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
192 int inMode=MPI_MODE_RDONLY;
194 char* outFilename = new char[outputFileName.length()];
\r
195 memcpy(outFilename, outputFileName.c_str(), outputFileName.length());
197 char* outAccnosFilename = new char[accnosFileName.length()];
\r
198 memcpy(outAccnosFilename, accnosFileName.c_str(), accnosFileName.length());
200 char* inFileName = new char[fastafile.length()];
\r
201 memcpy(inFileName, fastafile.c_str(), fastafile.length());
203 MPI_File_open(MPI_COMM_WORLD, inFileName, inMode, MPI_INFO_NULL, &inMPI); //comm, filename, mode, info, filepointer
204 MPI_File_open(MPI_COMM_WORLD, outFilename, outMode, MPI_INFO_NULL, &outMPI);
205 MPI_File_open(MPI_COMM_WORLD, outAccnosFilename, outMode, MPI_INFO_NULL, &outMPIAccnos);
209 delete outAccnosFilename;
211 if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPI); MPI_File_close(&outMPIAccnos); delete chimera; return 0; }
213 if (pid == 0) { //you are the root process
214 string outTemp = "For full window mapping info refer to " + mapInfo + "\n\n";
217 int length = outTemp.length();
218 char* buf2 = new char[length];
\r
219 memcpy(buf2, outTemp.c_str(), length);
221 MPI_File_write_shared(outMPI, buf2, length, MPI_CHAR, &status);
224 MPIPos = setFilePosFasta(fastafile, numSeqs); //fills MPIPos, returns numSeqs
226 //send file positions to all processes
227 MPI_Bcast(&numSeqs, 1, MPI_INT, 0, MPI_COMM_WORLD); //send numSeqs
228 MPI_Bcast(&MPIPos[0], (numSeqs+1), MPI_LONG, 0, MPI_COMM_WORLD); //send file pos
230 //figure out how many sequences you have to align
231 numSeqsPerProcessor = numSeqs / processors;
232 if(pid == (processors - 1)){ numSeqsPerProcessor = numSeqs - pid * numSeqsPerProcessor; }
233 int startIndex = pid * numSeqsPerProcessor;
236 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPI, outMPIAccnos, MPIPos);
238 if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPI); MPI_File_close(&outMPIAccnos); remove(outputFileName.c_str()); remove(accnosFileName.c_str()); delete chimera; return 0; }
240 for (int i = 1; i < processors; i++) {
242 MPI_Recv(&tempResult, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status);
243 if (tempResult != 0) { MPIWroteAccnos = true; }
245 }else{ //you are a child process
246 MPI_Bcast(&numSeqs, 1, MPI_INT, 0, MPI_COMM_WORLD); //get numSeqs
247 MPIPos.resize(numSeqs+1);
248 MPI_Bcast(&MPIPos[0], (numSeqs+1), MPI_LONG, 0, MPI_COMM_WORLD); //get file positions
250 //figure out how many sequences you have to align
251 numSeqsPerProcessor = numSeqs / processors;
252 if(pid == (processors - 1)){ numSeqsPerProcessor = numSeqs - pid * numSeqsPerProcessor; }
253 int startIndex = pid * numSeqsPerProcessor;
256 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPI, outMPIAccnos, MPIPos);
258 if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPI); MPI_File_close(&outMPIAccnos); delete chimera; return 0; }
260 MPI_Send(&MPIWroteAccnos, 1, MPI_INT, 0, tag, MPI_COMM_WORLD);
264 MPI_File_close(&inMPI);
265 MPI_File_close(&outMPI);
266 MPI_File_close(&outMPIAccnos);
268 //delete accnos file if blank
270 if (!MPIWroteAccnos) {
272 //MPI_File_delete(outAccnosFilename, info);
274 remove(accnosFileName.c_str());
280 string tempHeader = outputDir + getRootName(getSimpleName(fastafile)) + maskfile + "ccode.chimeras.tempHeader";
281 openOutputFile(tempHeader, outHeader);
283 outHeader << "For full window mapping info refer to " << mapInfo << endl << endl;
288 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
291 openInputFile(fastafile, inFASTA);
292 numSeqs=count(istreambuf_iterator<char>(inFASTA),istreambuf_iterator<char>(), '>');
295 lines.push_back(new linePair(0, numSeqs));
297 driver(lines[0], outputFileName, fastafile, accnosFileName);
299 if (m->control_pressed) {
300 remove(outputFileName.c_str());
301 remove(tempHeader.c_str());
302 remove(accnosFileName.c_str());
303 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
308 //delete accnos file if its blank
309 if (isBlank(accnosFileName)) { remove(accnosFileName.c_str()); hasAccnos = false; }
312 vector<int> positions;
313 processIDS.resize(0);
316 openInputFile(fastafile, inFASTA);
319 while(!inFASTA.eof()){
320 input = getline(inFASTA);
321 if (input.length() != 0) {
322 if(input[0] == '>'){ long int pos = inFASTA.tellg(); positions.push_back(pos - input.length() - 1); }
327 numSeqs = positions.size();
329 int numSeqsPerProcessor = numSeqs / processors;
331 for (int i = 0; i < processors; i++) {
332 long int startPos = positions[ i * numSeqsPerProcessor ];
333 if(i == processors - 1){
334 numSeqsPerProcessor = numSeqs - i * numSeqsPerProcessor;
336 lines.push_back(new linePair(startPos, numSeqsPerProcessor));
340 createProcesses(outputFileName, fastafile, accnosFileName);
342 rename((outputFileName + toString(processIDS[0]) + ".temp").c_str(), outputFileName.c_str());
344 //append output files
345 for(int i=1;i<processors;i++){
346 appendFiles((outputFileName + toString(processIDS[i]) + ".temp"), outputFileName);
347 remove((outputFileName + toString(processIDS[i]) + ".temp").c_str());
350 vector<string> nonBlankAccnosFiles;
351 //delete blank accnos files generated with multiple processes
352 for(int i=0;i<processors;i++){
353 if (!(isBlank(accnosFileName + toString(processIDS[i]) + ".temp"))) {
354 nonBlankAccnosFiles.push_back(accnosFileName + toString(processIDS[i]) + ".temp");
355 }else { remove((accnosFileName + toString(processIDS[i]) + ".temp").c_str()); }
358 //append accnos files
359 if (nonBlankAccnosFiles.size() != 0) {
360 rename(nonBlankAccnosFiles[0].c_str(), accnosFileName.c_str());
362 for (int h=1; h < nonBlankAccnosFiles.size(); h++) {
363 appendFiles(nonBlankAccnosFiles[h], accnosFileName);
364 remove(nonBlankAccnosFiles[h].c_str());
366 }else{ hasAccnos = false; }
368 if (m->control_pressed) {
369 remove(outputFileName.c_str());
370 remove(accnosFileName.c_str());
371 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
380 openInputFile(fastafile, inFASTA);
381 numSeqs=count(istreambuf_iterator<char>(inFASTA),istreambuf_iterator<char>(), '>');
383 lines.push_back(new linePair(0, numSeqs));
385 driver(lines[0], outputFileName, fastafile, accnosFileName);
387 if (m->control_pressed) {
388 remove(outputFileName.c_str());
389 remove(tempHeader.c_str());
390 remove(accnosFileName.c_str());
391 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
396 //delete accnos file if its blank
397 if (isBlank(accnosFileName)) { remove(accnosFileName.c_str()); hasAccnos = false; }
400 //m->mothurOut("Output File Names: ");
401 //if ((filter) && (method == "bellerophon")) { m->mothurOut(
402 //if (outputDir == "") { fastafile = getRootName(fastafile) + "filter.fasta"; }
403 // else { fastafile = outputDir + getRootName(getSimpleName(fastafile)) + "filter.fasta"; }
405 appendFiles(outputFileName, tempHeader);
407 remove(outputFileName.c_str());
408 rename(tempHeader.c_str(), outputFileName.c_str());
414 m->mothurOutEndLine();
415 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
416 m->mothurOut(outputFileName); m->mothurOutEndLine();
417 if (hasAccnos) { m->mothurOut(accnosFileName); m->mothurOutEndLine(); }
418 m->mothurOutEndLine();
420 for (int i = 0; i < lines.size(); i++) { delete lines[i]; } lines.clear();
422 m->mothurOutEndLine(); m->mothurOut("It took " + toString(time(NULL) - start) + " secs to check " + toString(numSeqs) + " sequences."); m->mothurOutEndLine();
427 catch(exception& e) {
428 m->errorOut(e, "ChimeraCcodeCommand", "execute");
432 //**********************************************************************************************************************
434 int ChimeraCcodeCommand::driver(linePair* line, string outputFName, string filename, string accnos){
437 openOutputFile(outputFName, out);
440 openOutputFile(accnos, out2);
443 openInputFile(filename, inFASTA);
445 inFASTA.seekg(line->start);
447 for(int i=0;i<line->numSeqs;i++){
449 if (m->control_pressed) { return 1; }
451 Sequence* candidateSeq = new Sequence(inFASTA); gobble(inFASTA);
453 if (candidateSeq->getName() != "") { //incase there is a commented sequence at the end of a file
455 if (candidateSeq->getAligned().length() != templateSeqsLength) {
456 m->mothurOut(candidateSeq->getName() + " is not the same length as the template sequences. Skipping."); m->mothurOutEndLine();
459 chimera->getChimeras(candidateSeq);
461 if (m->control_pressed) { delete candidateSeq; return 1; }
464 chimera->print(out, out2);
470 if((i+1) % 100 == 0){ m->mothurOut("Processing sequence: " + toString(i+1)); m->mothurOutEndLine(); }
473 if((line->numSeqs) % 100 != 0){ m->mothurOut("Processing sequence: " + toString(line->numSeqs)); m->mothurOutEndLine(); }
481 catch(exception& e) {
482 m->errorOut(e, "ChimeraCcodeCommand", "driver");
486 //**********************************************************************************************************************
488 int ChimeraCcodeCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& outMPI, MPI_File& outAccMPI, vector<long>& MPIPos){
493 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
495 for(int i=0;i<num;i++){
497 if (m->control_pressed) { return 0; }
500 int length = MPIPos[start+i+1] - MPIPos[start+i];
502 char* buf4 = new char[length];
\r
504 MPI_File_read_at(inMPI, MPIPos[start+i], buf4, length, MPI_CHAR, &status);
506 string tempBuf = buf4;
507 if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); }
508 istringstream iss (tempBuf,istringstream::in);
511 Sequence* candidateSeq = new Sequence(iss); gobble(iss);
513 if (candidateSeq->getName() != "") { //incase there is a commented sequence at the end of a file
515 if (candidateSeq->getAligned().length() != templateSeqsLength) {
516 m->mothurOut(candidateSeq->getName() + " is not the same length as the template sequences. Skipping."); m->mothurOutEndLine();
519 chimera->getChimeras(candidateSeq);
521 if (m->control_pressed) { delete candidateSeq; return 1; }
524 bool isChimeric = chimera->print(outMPI, outAccMPI);
525 if (isChimeric) { MPIWroteAccnos = true; }
531 if((i+1) % 100 == 0){ cout << "Processing sequence: " << (i+1) << endl; m->mothurOutJustToLog("Processing sequence: " + toString(i+1) + "\n"); }
534 if(num % 100 != 0){ cout << "Processing sequence: " << num << endl; m->mothurOutJustToLog("Processing sequence: " + toString(num) + "\n"); }
539 catch(exception& e) {
540 m->errorOut(e, "ChimeraCcodeCommand", "driverMPI");
546 /**************************************************************************************************/
548 int ChimeraCcodeCommand::createProcesses(string outputFileName, string filename, string accnos) {
550 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
552 // processIDS.resize(0);
554 //loop through and create all the processes you want
555 while (process != processors) {
559 processIDS.push_back(pid); //create map from line number to pid so you can append files in correct order later
562 driver(lines[process], outputFileName + toString(getpid()) + ".temp", filename, accnos + toString(getpid()) + ".temp");
564 }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
567 //force parent to wait until all the processes are done
568 for (int i=0;i<processors;i++) {
569 int temp = processIDS[i];
576 catch(exception& e) {
577 m->errorOut(e, "ChimeraCcodeCommand", "createProcesses");
581 //**********************************************************************************************************************