2 * screenseqscommand.cpp
5 * Created by Pat Schloss on 6/3/09.
6 * Copyright 2009 Patrick D. Schloss. All rights reserved.
10 #include "screenseqscommand.h"
11 #include "sequence.hpp"
13 //***************************************************************************************************************
15 ScreenSeqsCommand::ScreenSeqsCommand(string option) {
19 //allow user to run help
20 if(option == "help") { help(); abort = true; }
23 //valid paramters for this command
24 string AlignArray[] = {"fasta", "start", "end", "maxambig", "maxhomop", "minlength", "maxlength",
25 "name", "group", "alignreport","processors","outputdir","inputdir"};
26 vector<string> myArray (AlignArray, AlignArray+(sizeof(AlignArray)/sizeof(string)));
28 OptionParser parser(option);
29 map<string,string> parameters = parser.getParameters();
31 ValidParameters validParameter("screen.seqs");
32 map<string,string>::iterator it;
34 //check to make sure all parameters are valid for command
35 for (it = parameters.begin(); it != parameters.end(); it++) {
36 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
39 //if the user changes the input directory command factory will send this info to us in the output parameter
40 string inputDir = validParameter.validFile(parameters, "inputdir", false);
41 if (inputDir == "not found"){ inputDir = ""; }
44 it = parameters.find("fasta");
45 //user has given a template file
46 if(it != parameters.end()){
47 path = m->hasPath(it->second);
48 //if the user has not given a path then, add inputdir. else leave path alone.
49 if (path == "") { parameters["fasta"] = inputDir + it->second; }
52 it = parameters.find("group");
53 //user has given a template file
54 if(it != parameters.end()){
55 path = m->hasPath(it->second);
56 //if the user has not given a path then, add inputdir. else leave path alone.
57 if (path == "") { parameters["group"] = inputDir + it->second; }
60 it = parameters.find("name");
61 //user has given a template file
62 if(it != parameters.end()){
63 path = m->hasPath(it->second);
64 //if the user has not given a path then, add inputdir. else leave path alone.
65 if (path == "") { parameters["name"] = inputDir + it->second; }
68 it = parameters.find("alignreport");
69 //user has given a template file
70 if(it != parameters.end()){
71 path = m->hasPath(it->second);
72 //if the user has not given a path then, add inputdir. else leave path alone.
73 if (path == "") { parameters["alignreport"] = inputDir + it->second; }
77 //check for required parameters
78 fastafile = validParameter.validFile(parameters, "fasta", true);
79 if (fastafile == "not found") { m->mothurOut("fasta is a required parameter for the screen.seqs command."); m->mothurOutEndLine(); abort = true; }
80 else if (fastafile == "not open") { abort = true; }
82 groupfile = validParameter.validFile(parameters, "group", true);
83 if (groupfile == "not open") { abort = true; }
84 else if (groupfile == "not found") { groupfile = ""; }
86 namefile = validParameter.validFile(parameters, "name", true);
87 if (namefile == "not open") { abort = true; }
88 else if (namefile == "not found") { namefile = ""; }
90 alignreport = validParameter.validFile(parameters, "alignreport", true);
91 if (alignreport == "not open") { abort = true; }
92 else if (alignreport == "not found") { alignreport = ""; }
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"){
97 outputDir += m->hasPath(fastafile); //if user entered a file with a path then preserve it
100 //check for optional parameter and set defaults
101 // ...at some point should added some additional type checking...
103 temp = validParameter.validFile(parameters, "start", false); if (temp == "not found") { temp = "-1"; }
104 convert(temp, startPos);
106 temp = validParameter.validFile(parameters, "end", false); if (temp == "not found") { temp = "-1"; }
107 convert(temp, endPos);
109 temp = validParameter.validFile(parameters, "maxambig", false); if (temp == "not found") { temp = "-1"; }
110 convert(temp, maxAmbig);
112 temp = validParameter.validFile(parameters, "maxhomop", false); if (temp == "not found") { temp = "-1"; }
113 convert(temp, maxHomoP);
115 temp = validParameter.validFile(parameters, "minlength", false); if (temp == "not found") { temp = "-1"; }
116 convert(temp, minLength);
118 temp = validParameter.validFile(parameters, "maxlength", false); if (temp == "not found") { temp = "-1"; }
119 convert(temp, maxLength);
121 temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = "1"; }
122 convert(temp, processors);
127 catch(exception& e) {
128 m->errorOut(e, "ScreenSeqsCommand", "ScreenSeqsCommand");
132 //**********************************************************************************************************************
134 void ScreenSeqsCommand::help(){
136 m->mothurOut("The screen.seqs command reads a fastafile and creates .....\n");
137 m->mothurOut("The screen.seqs command parameters are fasta, start, end, maxambig, maxhomop, minlength, maxlength, name, group and processors.\n");
138 m->mothurOut("The fasta parameter is required.\n");
139 m->mothurOut("The start parameter .... The default is -1.\n");
140 m->mothurOut("The end parameter .... The default is -1.\n");
141 m->mothurOut("The maxambig parameter .... The default is -1.\n");
142 m->mothurOut("The maxhomop parameter .... The default is -1.\n");
143 m->mothurOut("The minlength parameter .... The default is -1.\n");
144 m->mothurOut("The maxlength parameter .... The default is -1.\n");
145 m->mothurOut("The processors parameter allows you to specify the number of processors to use while running the command. The default is 1.\n");
146 m->mothurOut("The name parameter allows you to provide a namesfile, and the group parameter allows you to provide a groupfile.\n");
147 m->mothurOut("The screen.seqs command should be in the following format: \n");
148 m->mothurOut("screen.seqs(fasta=yourFastaFile, name=youNameFile, group=yourGroupFIle, start=yourStart, end=yourEnd, maxambig=yourMaxambig, \n");
149 m->mothurOut("maxhomop=yourMaxhomop, minlength=youMinlength, maxlength=yourMaxlength) \n");
150 m->mothurOut("Example screen.seqs(fasta=abrecovery.fasta, name=abrecovery.names, group=abrecovery.groups, start=..., end=..., maxambig=..., maxhomop=..., minlength=..., maxlength=...).\n");
151 m->mothurOut("Note: No spaces between parameter labels (i.e. fasta), '=' and parameters (i.e.yourFasta).\n\n");
154 catch(exception& e) {
155 m->errorOut(e, "ScreenSeqsCommand", "help");
160 //***************************************************************************************************************
162 ScreenSeqsCommand::~ScreenSeqsCommand(){ /* do nothing */ }
164 //***************************************************************************************************************
166 int ScreenSeqsCommand::execute(){
169 if (abort == true) { return 0; }
171 string goodSeqFile = outputDir + m->getRootName(m->getSimpleName(fastafile)) + "good" + m->getExtension(fastafile);
172 string badAccnosFile = outputDir + m->getRootName(m->getSimpleName(fastafile)) + "bad.accnos";
174 int numFastaSeqs = 0;
175 set<string> badSeqNames;
176 int start = time(NULL);
179 int pid, end, numSeqsPerProcessor;
181 vector<unsigned long int> MPIPos;
184 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
185 MPI_Comm_size(MPI_COMM_WORLD, &processors);
189 MPI_File outMPIBadAccnos;
191 int outMode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
192 int inMode=MPI_MODE_RDONLY;
194 char outGoodFilename[1024];
195 strcpy(outGoodFilename, goodSeqFile.c_str());
197 char outBadAccnosFilename[1024];
198 strcpy(outBadAccnosFilename, badAccnosFile.c_str());
200 char inFileName[1024];
201 strcpy(inFileName, fastafile.c_str());
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, outGoodFilename, outMode, MPI_INFO_NULL, &outMPIGood);
205 MPI_File_open(MPI_COMM_WORLD, outBadAccnosFilename, outMode, MPI_INFO_NULL, &outMPIBadAccnos);
207 if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPIGood); MPI_File_close(&outMPIBadAccnos); return 0; }
209 if (pid == 0) { //you are the root process
211 MPIPos = m->setFilePosFasta(fastafile, numFastaSeqs); //fills MPIPos, returns numSeqs
213 //send file positions to all processes
214 for(int i = 1; i < processors; i++) {
215 MPI_Send(&numFastaSeqs, 1, MPI_INT, i, tag, MPI_COMM_WORLD);
216 MPI_Send(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, i, tag, MPI_COMM_WORLD);
219 //figure out how many sequences you have to align
220 numSeqsPerProcessor = numFastaSeqs / processors;
221 int startIndex = pid * numSeqsPerProcessor;
222 if(pid == (processors - 1)){ numSeqsPerProcessor = numFastaSeqs - pid * numSeqsPerProcessor; }
223 // cout << pid << '\t' << numSeqsPerProcessor << '\t' << startIndex << endl;
225 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPIGood, outMPIBadAccnos, MPIPos, badSeqNames);
226 //cout << pid << " done" << endl;
227 if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPIGood); MPI_File_close(&outMPIBadAccnos); return 0; }
229 for (int i = 1; i < processors; i++) {
233 MPI_Recv(&badSize, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status);
234 /*for (int j = 0; j < badSize; j++) {
236 MPI_Recv(&length, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status); //recv the length of the name
237 char* buf2 = new char[length]; //make space to recieve it
238 MPI_Recv(buf2, length, MPI_CHAR, i, tag, MPI_COMM_WORLD, &status); //get name
240 string tempBuf = buf2;
241 if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); }
244 badSeqNames.insert(tempBuf);
247 }else{ //you are a child process
248 MPI_Recv(&numFastaSeqs, 1, MPI_INT, 0, tag, MPI_COMM_WORLD, &status);
249 MPIPos.resize(numFastaSeqs+1);
250 MPI_Recv(&MPIPos[0], (numFastaSeqs+1), MPI_LONG, 0, tag, MPI_COMM_WORLD, &status);
252 //figure out how many sequences you have to align
253 numSeqsPerProcessor = numFastaSeqs / processors;
254 int startIndex = pid * numSeqsPerProcessor;
255 if(pid == (processors - 1)){ numSeqsPerProcessor = numFastaSeqs - pid * numSeqsPerProcessor; }
256 //cout << pid << '\t' << numSeqsPerProcessor << '\t' << startIndex << endl;
258 driverMPI(startIndex, numSeqsPerProcessor, inMPI, outMPIGood, outMPIBadAccnos, MPIPos, badSeqNames);
259 //cout << pid << " done" << endl;
260 if (m->control_pressed) { MPI_File_close(&inMPI); MPI_File_close(&outMPIGood); MPI_File_close(&outMPIBadAccnos); return 0; }
263 int badSize = badSeqNames.size();
264 MPI_Send(&badSize, 1, MPI_INT, 0, tag, MPI_COMM_WORLD);
267 set<string>::iterator it;
268 for (it = badSeqNames.begin(); it != badSeqNames.end(); it++) {
270 int length = name.length();
271 char* buf2 = new char[length];
272 memcpy(buf2, name.c_str(), length);
274 MPI_Send(&length, 1, MPI_INT, 0, tag, MPI_COMM_WORLD);
275 MPI_Send(buf2, length, MPI_CHAR, 0, tag, MPI_COMM_WORLD);
280 MPI_File_close(&inMPI);
281 MPI_File_close(&outMPIGood);
282 MPI_File_close(&outMPIBadAccnos);
283 MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
286 vector<unsigned long int> positions = m->divideFile(fastafile, processors);
288 for (int i = 0; i < (positions.size()-1); i++) {
289 lines.push_back(new linePair(positions[i], positions[(i+1)]));
292 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
294 numFastaSeqs = driver(lines[0], goodSeqFile, badAccnosFile, fastafile, badSeqNames);
296 if (m->control_pressed) { remove(goodSeqFile.c_str()); return 0; }
299 processIDS.resize(0);
301 numFastaSeqs = createProcesses(goodSeqFile, badAccnosFile, fastafile, badSeqNames);
303 rename((goodSeqFile + toString(processIDS[0]) + ".temp").c_str(), goodSeqFile.c_str());
304 rename((badAccnosFile + toString(processIDS[0]) + ".temp").c_str(), badAccnosFile.c_str());
306 //append alignment and report files
307 for(int i=1;i<processors;i++){
308 m->appendFiles((goodSeqFile + toString(processIDS[i]) + ".temp"), goodSeqFile);
309 remove((goodSeqFile + toString(processIDS[i]) + ".temp").c_str());
311 m->appendFiles((badAccnosFile + toString(processIDS[i]) + ".temp"), badAccnosFile);
312 remove((badAccnosFile + toString(processIDS[i]) + ".temp").c_str());
315 if (m->control_pressed) { remove(goodSeqFile.c_str()); return 0; }
317 //read badSeqs in because root process doesnt know what other "bad" seqs the children found
319 int ableToOpen = m->openInputFile(badAccnosFile, inBad, "no error");
321 if (ableToOpen == 0) {
324 while (!inBad.eof()) {
325 inBad >> tempName; m->gobble(inBad);
326 badSeqNames.insert(tempName);
332 numFastaSeqs = driver(lines[0], goodSeqFile, badAccnosFile, fastafile, badSeqNames);
334 if (m->control_pressed) { remove(goodSeqFile.c_str()); return 0; }
341 MPI_Comm_rank(MPI_COMM_WORLD, &pid);
343 if (pid == 0) { //only one process should fix files
345 //read accnos file with all names in it, process 0 just has its names
346 MPI_File inMPIAccnos;
349 char inFileName[1024];
350 strcpy(inFileName, badAccnosFile.c_str());
352 MPI_File_open(MPI_COMM_SELF, inFileName, inMode, MPI_INFO_NULL, &inMPIAccnos); //comm, filename, mode, info, filepointer
353 MPI_File_get_size(inMPIAccnos, &size);
355 char* buffer = new char[size];
356 MPI_File_read(inMPIAccnos, buffer, size, MPI_CHAR, &status);
358 string tempBuf = buffer;
359 if (tempBuf.length() > size) { tempBuf = tempBuf.substr(0, size); }
360 istringstream iss (tempBuf,istringstream::in);
363 MPI_File_close(&inMPIAccnos);
368 iss >> tempName; m->gobble(iss);
369 badSeqNames.insert(tempName);
373 if(namefile != "" && groupfile != "") {
374 screenNameGroupFile(badSeqNames);
375 if (m->control_pressed) { remove(goodSeqFile.c_str()); return 0; }
376 }else if(namefile != "") {
377 screenNameGroupFile(badSeqNames);
378 if (m->control_pressed) { remove(goodSeqFile.c_str()); return 0; }
379 }else if(groupfile != "") { screenGroupFile(badSeqNames); } // this screens just the group
381 if (m->control_pressed) { remove(goodSeqFile.c_str()); return 0; }
383 if(alignreport != "") { screenAlignReport(badSeqNames); }
385 if (m->control_pressed) { remove(goodSeqFile.c_str()); return 0; }
391 m->mothurOutEndLine();
392 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
393 m->mothurOut(goodSeqFile); m->mothurOutEndLine();
394 m->mothurOut(badAccnosFile); m->mothurOutEndLine();
395 for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
396 m->mothurOutEndLine();
397 m->mothurOutEndLine();
399 m->mothurOut("It took " + toString(time(NULL) - start) + " secs to screen " + toString(numFastaSeqs) + " sequences.");
400 m->mothurOutEndLine();
404 catch(exception& e) {
405 m->errorOut(e, "ScreenSeqsCommand", "execute");
410 //***************************************************************************************************************
412 int ScreenSeqsCommand::screenNameGroupFile(set<string> badSeqNames){
415 m->openInputFile(namefile, inputNames);
416 set<string> badSeqGroups;
417 string seqName, seqList, group;
418 set<string>::iterator it;
420 string goodNameFile = outputDir + m->getRootName(m->getSimpleName(namefile)) + "good" + m->getExtension(namefile);
421 outputNames.push_back(goodNameFile);
423 ofstream goodNameOut; m->openOutputFile(goodNameFile, goodNameOut);
425 while(!inputNames.eof()){
426 if (m->control_pressed) { goodNameOut.close(); inputNames.close(); remove(goodNameFile.c_str()); return 0; }
428 inputNames >> seqName >> seqList;
429 it = badSeqNames.find(seqName);
431 if(it != badSeqNames.end()){
432 badSeqNames.erase(it);
436 for(int i=0;i<seqList.length();i++){
437 if(seqList[i] == ','){
438 badSeqGroups.insert(seqList.substr(start,i-start));
442 badSeqGroups.insert(seqList.substr(start,seqList.length()-start));
446 goodNameOut << seqName << '\t' << seqList << endl;
448 m->gobble(inputNames);
453 //we were unable to remove some of the bad sequences
454 if (badSeqNames.size() != 0) {
455 for (it = badSeqNames.begin(); it != badSeqNames.end(); it++) {
456 m->mothurOut("Your namefile does not include the sequence " + *it + " please correct.");
457 m->mothurOutEndLine();
463 ifstream inputGroups;
464 m->openInputFile(groupfile, inputGroups);
466 string goodGroupFile = outputDir + m->getRootName(m->getSimpleName(groupfile)) + "good" + m->getExtension(groupfile);
467 outputNames.push_back(goodGroupFile);
469 ofstream goodGroupOut; m->openOutputFile(goodGroupFile, goodGroupOut);
471 while(!inputGroups.eof()){
472 if (m->control_pressed) { goodGroupOut.close(); inputGroups.close(); remove(goodNameFile.c_str()); remove(goodGroupFile.c_str()); return 0; }
474 inputGroups >> seqName >> group;
476 it = badSeqGroups.find(seqName);
478 if(it != badSeqGroups.end()){
479 badSeqGroups.erase(it);
482 goodGroupOut << seqName << '\t' << group << endl;
484 m->gobble(inputGroups);
487 goodGroupOut.close();
489 //we were unable to remove some of the bad sequences
490 if (badSeqGroups.size() != 0) {
491 for (it = badSeqGroups.begin(); it != badSeqGroups.end(); it++) {
492 m->mothurOut("Your groupfile does not include the sequence " + *it + " please correct.");
493 m->mothurOutEndLine();
501 catch(exception& e) {
502 m->errorOut(e, "ScreenSeqsCommand", "screenNameGroupFile");
507 //***************************************************************************************************************
509 int ScreenSeqsCommand::screenGroupFile(set<string> badSeqNames){
511 ifstream inputGroups;
512 m->openInputFile(groupfile, inputGroups);
513 string seqName, group;
514 set<string>::iterator it;
516 string goodGroupFile = outputDir + m->getRootName(m->getSimpleName(groupfile)) + "good" + m->getExtension(groupfile);
517 outputNames.push_back(goodGroupFile);
518 ofstream goodGroupOut; m->openOutputFile(goodGroupFile, goodGroupOut);
520 while(!inputGroups.eof()){
521 if (m->control_pressed) { goodGroupOut.close(); inputGroups.close(); remove(goodGroupFile.c_str()); return 0; }
523 inputGroups >> seqName >> group;
524 it = badSeqNames.find(seqName);
526 if(it != badSeqNames.end()){
527 badSeqNames.erase(it);
530 goodGroupOut << seqName << '\t' << group << endl;
532 m->gobble(inputGroups);
535 if (m->control_pressed) { goodGroupOut.close(); inputGroups.close(); remove(goodGroupFile.c_str()); return 0; }
537 //we were unable to remove some of the bad sequences
538 if (badSeqNames.size() != 0) {
539 for (it = badSeqNames.begin(); it != badSeqNames.end(); it++) {
540 m->mothurOut("Your groupfile does not include the sequence " + *it + " please correct.");
541 m->mothurOutEndLine();
546 goodGroupOut.close();
548 if (m->control_pressed) { remove(goodGroupFile.c_str()); }
553 catch(exception& e) {
554 m->errorOut(e, "ScreenSeqsCommand", "screenGroupFile");
559 //***************************************************************************************************************
561 int ScreenSeqsCommand::screenAlignReport(set<string> badSeqNames){
563 ifstream inputAlignReport;
564 m->openInputFile(alignreport, inputAlignReport);
565 string seqName, group;
566 set<string>::iterator it;
568 string goodAlignReportFile = outputDir + m->getRootName(m->getSimpleName(alignreport)) + "good" + m->getExtension(alignreport);
569 outputNames.push_back(goodAlignReportFile);
570 ofstream goodAlignReportOut; m->openOutputFile(goodAlignReportFile, goodAlignReportOut);
572 while (!inputAlignReport.eof()) { // need to copy header
573 char c = inputAlignReport.get();
574 goodAlignReportOut << c;
575 if (c == 10 || c == 13){ break; }
578 while(!inputAlignReport.eof()){
579 if (m->control_pressed) { goodAlignReportOut.close(); inputAlignReport.close(); remove(goodAlignReportFile.c_str()); return 0; }
581 inputAlignReport >> seqName;
582 it = badSeqNames.find(seqName);
584 while (!inputAlignReport.eof()) { // need to copy header
585 char c = inputAlignReport.get();
587 if (c == 10 || c == 13){ break; }
590 if(it != badSeqNames.end()){
591 badSeqNames.erase(it);
594 goodAlignReportOut << seqName << '\t' << line;
596 m->gobble(inputAlignReport);
599 if (m->control_pressed) { goodAlignReportOut.close(); inputAlignReport.close(); remove(goodAlignReportFile.c_str()); return 0; }
601 //we were unable to remove some of the bad sequences
602 if (badSeqNames.size() != 0) {
603 for (it = badSeqNames.begin(); it != badSeqNames.end(); it++) {
604 m->mothurOut("Your alignreport file does not include the sequence " + *it + " please correct.");
605 m->mothurOutEndLine();
609 inputAlignReport.close();
610 goodAlignReportOut.close();
612 if (m->control_pressed) { remove(goodAlignReportFile.c_str()); return 0; }
617 catch(exception& e) {
618 m->errorOut(e, "ScreenSeqsCommand", "screenAlignReport");
623 //**********************************************************************************************************************
625 int ScreenSeqsCommand::driver(linePair* filePos, string goodFName, string badAccnosFName, string filename, set<string>& badSeqNames){
628 m->openOutputFile(goodFName, goodFile);
630 ofstream badAccnosFile;
631 m->openOutputFile(badAccnosFName, badAccnosFile);
634 m->openInputFile(filename, inFASTA);
636 inFASTA.seekg(filePos->start);
643 if (m->control_pressed) { return 0; }
645 Sequence currSeq(inFASTA); m->gobble(inFASTA);
646 if (currSeq.getName() != "") {
647 bool goodSeq = 1; // innocent until proven guilty
648 if(goodSeq == 1 && startPos != -1 && startPos < currSeq.getStartPos()) { goodSeq = 0; }
649 if(goodSeq == 1 && endPos != -1 && endPos > currSeq.getEndPos()) { goodSeq = 0; }
650 if(goodSeq == 1 && maxAmbig != -1 && maxAmbig < currSeq.getAmbigBases()) { goodSeq = 0; }
651 if(goodSeq == 1 && maxHomoP != -1 && maxHomoP < currSeq.getLongHomoPolymer()) { goodSeq = 0; }
652 if(goodSeq == 1 && minLength != -1 && minLength > currSeq.getNumBases()) { goodSeq = 0; }
653 if(goodSeq == 1 && maxLength != -1 && maxLength < currSeq.getNumBases()) { goodSeq = 0; }
656 currSeq.printSequence(goodFile);
659 badAccnosFile << currSeq.getName() << endl;
660 badSeqNames.insert(currSeq.getName());
665 unsigned long int pos = inFASTA.tellg();
666 if ((pos == -1) || (pos >= filePos->end)) { break; }
669 if((count) % 100 == 0){ m->mothurOut("Processing sequence: " + toString(count)); m->mothurOutEndLine(); }
672 if((count) % 100 != 0){ m->mothurOut("Processing sequence: " + toString(count)); m->mothurOutEndLine(); }
677 badAccnosFile.close();
681 catch(exception& e) {
682 m->errorOut(e, "ScreenSeqsCommand", "driver");
686 //**********************************************************************************************************************
688 int ScreenSeqsCommand::driverMPI(int start, int num, MPI_File& inMPI, MPI_File& goodFile, MPI_File& badAccnosFile, vector<unsigned long int>& MPIPos, set<string>& badSeqNames){
690 string outputString = "";
691 MPI_Status statusGood;
692 MPI_Status statusBadAccnos;
695 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
697 for(int i=0;i<num;i++){
699 if (m->control_pressed) { return 0; }
702 int length = MPIPos[start+i+1] - MPIPos[start+i];
704 char* buf4 = new char[length];
705 memcpy(buf4, outputString.c_str(), length);
707 MPI_File_read_at(inMPI, MPIPos[start+i], buf4, length, MPI_CHAR, &status);
709 string tempBuf = buf4; delete buf4;
710 if (tempBuf.length() > length) { tempBuf = tempBuf.substr(0, length); }
711 istringstream iss (tempBuf,istringstream::in);
713 Sequence currSeq(iss);
716 if (currSeq.getName() != "") {
717 bool goodSeq = 1; // innocent until proven guilty
718 if(goodSeq == 1 && startPos != -1 && startPos < currSeq.getStartPos()) { goodSeq = 0; }
719 if(goodSeq == 1 && endPos != -1 && endPos > currSeq.getEndPos()) { goodSeq = 0; }
720 if(goodSeq == 1 && maxAmbig != -1 && maxAmbig < currSeq.getAmbigBases()) { goodSeq = 0; }
721 if(goodSeq == 1 && maxHomoP != -1 && maxHomoP < currSeq.getLongHomoPolymer()) { goodSeq = 0; }
722 if(goodSeq == 1 && minLength != -1 && minLength > currSeq.getNumBases()) { goodSeq = 0; }
723 if(goodSeq == 1 && maxLength != -1 && maxLength < currSeq.getNumBases()) { goodSeq = 0; }
726 outputString = ">" + currSeq.getName() + "\n" + currSeq.getAligned() + "\n";
729 length = outputString.length();
730 char* buf2 = new char[length];
731 memcpy(buf2, outputString.c_str(), length);
733 MPI_File_write_shared(goodFile, buf2, length, MPI_CHAR, &statusGood);
738 badSeqNames.insert(currSeq.getName());
740 //write to bad accnos file
741 outputString = currSeq.getName() + "\n";
743 length = outputString.length();
744 char* buf3 = new char[length];
745 memcpy(buf3, outputString.c_str(), length);
747 MPI_File_write_shared(badAccnosFile, buf3, length, MPI_CHAR, &statusBadAccnos);
755 catch(exception& e) {
756 m->errorOut(e, "ScreenSeqsCommand", "driverMPI");
761 /**************************************************************************************************/
763 int ScreenSeqsCommand::createProcesses(string goodFileName, string badAccnos, string filename, set<string>& badSeqNames) {
765 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
769 //loop through and create all the processes you want
770 while (process != processors) {
774 processIDS.push_back(pid); //create map from line number to pid so you can append files in correct order later
777 num = driver(lines[process], goodFileName + toString(getpid()) + ".temp", badAccnos + toString(getpid()) + ".temp", filename, badSeqNames);
779 //pass numSeqs to parent
781 string tempFile = filename + toString(getpid()) + ".num.temp";
782 m->openOutputFile(tempFile, out);
787 }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
790 //force parent to wait until all the processes are done
791 for (int i=0;i<processors;i++) {
792 int temp = processIDS[i];
796 for (int i = 0; i < processIDS.size(); i++) {
798 string tempFile = filename + toString(processIDS[i]) + ".num.temp";
799 m->openInputFile(tempFile, in);
800 if (!in.eof()) { int tempNum = 0; in >> tempNum; num += tempNum; }
801 in.close(); remove(tempFile.c_str());
807 catch(exception& e) {
808 m->errorOut(e, "ScreenSeqsCommand", "createProcesses");
813 //***************************************************************************************************************