5 * Created by Sarah Westcott on 5/7/09.
6 * Copyright 2009 Schloss Lab UMASS Amherst. All rights reserved.
10 #include "distancecommand.h"
11 #include "ignoregaps.h"
12 #include "eachgapdist.h"
13 #include "eachgapignore.h"
14 #include "onegapdist.h"
15 #include "onegapignore.h"
17 //**********************************************************************************************************************
18 vector<string> DistanceCommand::getValidParameters(){
20 string Array[] = {"fasta","oldfasta","column", "output", "calc", "countends", "cutoff", "processors", "outputdir","inputdir","compress"};
21 vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
25 m->errorOut(e, "DistanceCommand", "getValidParameters");
29 //**********************************************************************************************************************
30 DistanceCommand::DistanceCommand(){
33 //initialize outputTypes
34 vector<string> tempOutNames;
35 outputTypes["phylip"] = tempOutNames;
36 outputTypes["column"] = tempOutNames;
39 m->errorOut(e, "DistanceCommand", "DistanceCommand");
43 //**********************************************************************************************************************
44 vector<string> DistanceCommand::getRequiredParameters(){
46 string Array[] = {"fasta"};
47 vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
51 m->errorOut(e, "ChopSeqsCommand", "getRequiredParameters");
55 //**********************************************************************************************************************
56 vector<string> DistanceCommand::getRequiredFiles(){
58 vector<string> myArray;
62 m->errorOut(e, "DistanceCommand", "getRequiredFiles");
66 //**********************************************************************************************************************
67 DistanceCommand::DistanceCommand(string option) {
72 //allow user to run help
73 if(option == "help") { help(); abort = true; }
76 //valid paramters for this command
77 string Array[] = {"fasta","oldfasta","column", "output", "calc", "countends", "cutoff", "processors", "outputdir","inputdir","compress"};
79 vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
81 OptionParser parser(option);
82 map<string, string> parameters = parser.getParameters();
84 ValidParameters validParameter("dist.seqs");
85 map<string, string>::iterator it2;
87 //check to make sure all parameters are valid for command
88 for (it2 = parameters.begin(); it2 != parameters.end(); it2++) {
89 if (validParameter.isValidParameter(it2->first, myArray, it2->second) != true) { abort = true; }
92 //initialize outputTypes
93 vector<string> tempOutNames;
94 outputTypes["phylip"] = tempOutNames;
95 outputTypes["column"] = tempOutNames;
97 //if the user changes the input directory command factory will send this info to us in the output parameter
98 string inputDir = validParameter.validFile(parameters, "inputdir", false);
99 if (inputDir == "not found"){ inputDir = ""; }
102 it2 = parameters.find("fasta");
103 //user has given a template file
104 if(it2 != parameters.end()){
105 path = m->hasPath(it2->second);
106 //if the user has not given a path then, add inputdir. else leave path alone.
107 if (path == "") { parameters["fasta"] = inputDir + it2->second; }
110 it2 = parameters.find("oldfasta");
111 //user has given a template file
112 if(it2 != parameters.end()){
113 path = m->hasPath(it2->second);
114 //if the user has not given a path then, add inputdir. else leave path alone.
115 if (path == "") { parameters["oldfasta"] = inputDir + it2->second; }
118 it2 = parameters.find("column");
119 //user has given a template file
120 if(it2 != parameters.end()){
121 path = m->hasPath(it2->second);
122 //if the user has not given a path then, add inputdir. else leave path alone.
123 if (path == "") { parameters["column"] = inputDir + it2->second; }
127 //check for required parameters
128 fastafile = validParameter.validFile(parameters, "fasta", true);
129 if (fastafile == "not found") { m->mothurOut("fasta is a required parameter for the dist.seqs command."); m->mothurOutEndLine(); abort = true; }
130 else if (fastafile == "not open") { abort = true; }
133 m->openInputFile(fastafile, inFASTA);
134 alignDB = SequenceDB(inFASTA);
138 oldfastafile = validParameter.validFile(parameters, "oldfasta", true);
139 if (oldfastafile == "not found") { oldfastafile = ""; }
140 else if (oldfastafile == "not open") { abort = true; }
142 column = validParameter.validFile(parameters, "column", true);
143 if (column == "not found") { column = ""; }
144 else if (column == "not open") { abort = true; }
146 //if the user changes the output directory command factory will send this info to us in the output parameter
147 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){
149 outputDir += m->hasPath(fastafile); //if user entered a file with a path then preserve it
152 //check for optional parameter and set defaults
153 // ...at some point should added some additional type checking...
154 calc = validParameter.validFile(parameters, "calc", false);
155 if (calc == "not found") { calc = "onegap"; }
157 if (calc == "default") { calc = "onegap"; }
159 m->splitAtDash(calc, Estimators);
162 temp = validParameter.validFile(parameters, "countends", false); if(temp == "not found"){ temp = "T"; }
163 convert(temp, countends);
165 temp = validParameter.validFile(parameters, "cutoff", false); if(temp == "not found"){ temp = "1.0"; }
166 convert(temp, cutoff);
168 temp = validParameter.validFile(parameters, "processors", false); if(temp == "not found"){ temp = "1"; }
169 convert(temp, processors);
171 temp = validParameter.validFile(parameters, "compress", false); if(temp == "not found"){ temp = "F"; }
172 convert(temp, compress);
174 output = validParameter.validFile(parameters, "output", false); if(output == "not found"){ output = "column"; }
176 if (((column != "") && (oldfastafile == "")) || ((column == "") && (oldfastafile != ""))) { m->mothurOut("If you provide column or oldfasta, you must provide both."); m->mothurOutEndLine(); abort=true; }
178 if ((column != "") && (oldfastafile != "") && (output != "column")) { m->mothurOut("You have provided column and oldfasta, indicating you want to append distances to your column file. Your output must be in column format to do so."); m->mothurOutEndLine(); abort=true; }
180 if ((output != "column") && (output != "lt") && (output != "square")) { m->mothurOut(output + " is not a valid output form. Options are column, lt and square. I will use column."); m->mothurOutEndLine(); output = "column"; }
182 ValidCalculators validCalculator;
184 if (m->isTrue(countends) == true) {
185 for (int i=0; i<Estimators.size(); i++) {
186 if (validCalculator.isValidCalculator("distance", Estimators[i]) == true) {
187 if (Estimators[i] == "nogaps") { distCalculator = new ignoreGaps(); }
188 else if (Estimators[i] == "eachgap") { distCalculator = new eachGapDist(); }
189 else if (Estimators[i] == "onegap") { distCalculator = new oneGapDist(); }
193 for (int i=0; i<Estimators.size(); i++) {
194 if (validCalculator.isValidCalculator("distance", Estimators[i]) == true) {
195 if (Estimators[i] == "nogaps") { distCalculator = new ignoreGaps(); }
196 else if (Estimators[i] == "eachgap"){ distCalculator = new eachGapIgnoreTermGapDist(); }
197 else if (Estimators[i] == "onegap") { distCalculator = new oneGapIgnoreTermGapDist(); }
205 catch(exception& e) {
206 m->errorOut(e, "DistanceCommand", "DistanceCommand");
211 //**********************************************************************************************************************
213 DistanceCommand::~DistanceCommand(){
215 for(int i=0;i<lines.size();i++){
221 //**********************************************************************************************************************
223 void DistanceCommand::help(){
225 m->mothurOut("The dist.seqs command reads a file containing sequences and creates a distance file.\n");
226 m->mothurOut("The dist.seqs command parameters are fasta, oldfasta, column, calc, countends, output, compress, cutoff and processors. \n");
227 m->mothurOut("The fasta parameter is required.\n");
228 m->mothurOut("The oldfasta and column parameters allow you to append the distances calculated to the column file.\n");
229 m->mothurOut("The calc parameter allows you to specify the method of calculating the distances. Your options are: nogaps, onegap or eachgap. The default is onegap.\n");
230 m->mothurOut("The countends parameter allows you to specify whether to include terminal gaps in distance. Your options are: T or F. The default is T.\n");
231 m->mothurOut("The cutoff parameter allows you to specify maximum distance to keep. The default is 1.0.\n");
232 m->mothurOut("The output parameter allows you to specify format of your distance matrix. Options are column, lt, and square. The default is column.\n");
233 m->mothurOut("The processors parameter allows you to specify number of processors to use. The default is 1.\n");
234 m->mothurOut("The compress parameter allows you to indicate that you want the resulting distance file compressed. The default is false.\n");
235 m->mothurOut("The dist.seqs command should be in the following format: \n");
236 m->mothurOut("dist.seqs(fasta=yourFastaFile, calc=yourCalc, countends=yourEnds, cutoff= yourCutOff, processors=yourProcessors) \n");
237 m->mothurOut("Example dist.seqs(fasta=amazon.fasta, calc=eachgap, countends=F, cutoff= 2.0, processors=3).\n");
238 m->mothurOut("Note: No spaces between parameter labels (i.e. calc), '=' and parameters (i.e.yourCalc).\n\n");
240 catch(exception& e) {
241 m->errorOut(e, "DistanceCommand", "help");
245 //**********************************************************************************************************************
247 int DistanceCommand::execute(){
250 if (abort == true) { return 0; }
252 int startTime = time(NULL);
254 //save number of new sequence
255 numNewFasta = alignDB.getNumSeqs();
257 //sanity check the oldfasta and column file as well as add oldfasta sequences to alignDB
258 if ((oldfastafile != "") && (column != "")) { if (!(sanityCheck())) { return 0; } }
260 if (m->control_pressed) { return 0; }
262 int numSeqs = alignDB.getNumSeqs();
267 if (output == "lt") { //does the user want lower triangle phylip formatted file
268 outputFile = outputDir + m->getRootName(m->getSimpleName(fastafile)) + "phylip.dist";
269 remove(outputFile.c_str()); outputTypes["phylip"].push_back(outputFile);
271 //output numSeqs to phylip formatted dist file
272 }else if (output == "column") { //user wants column format
273 outputFile = outputDir + m->getRootName(m->getSimpleName(fastafile)) + "dist";
274 outputTypes["column"].push_back(outputFile);
276 //so we don't accidentally overwrite
277 if (outputFile == column) {
278 string tempcolumn = column + ".old";
279 rename(column.c_str(), tempcolumn.c_str());
282 remove(outputFile.c_str());
283 }else { //assume square
284 outputFile = outputDir + m->getRootName(m->getSimpleName(fastafile)) + "square.dist";
285 remove(outputFile.c_str());
286 outputTypes["phylip"].push_back(outputFile);
296 MPI_Comm_size(MPI_COMM_WORLD, &processors); //set processors to the number of mpi processes running
297 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
299 //each process gets where it should start and stop in the file
300 if (output != "square") {
301 start = int (sqrt(float(pid)/float(processors)) * numSeqs);
302 end = int (sqrt(float(pid+1)/float(processors)) * numSeqs);
304 start = int ((float(pid)/float(processors)) * numSeqs);
305 end = int ((float(pid+1)/float(processors)) * numSeqs);
308 if (output == "column") {
310 int amode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
312 //char* filename = new char[outputFile.length()];
313 //memcpy(filename, outputFile.c_str(), outputFile.length());
316 strcpy(filename, outputFile.c_str());
318 MPI_File_open(MPI_COMM_WORLD, filename, amode, MPI_INFO_NULL, &outMPI);
321 if (pid == 0) { //you are the root process
326 driverMPI(start, end, outMPI, cutoff);
328 if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&outMPI); delete distCalculator; return 0; }
331 for(int i = 1; i < processors; i++) {
332 if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&outMPI); delete distCalculator; return 0; }
335 MPI_Recv(buf, 5, MPI_CHAR, i, tag, MPI_COMM_WORLD, &status);
337 }else { //you are a child process
339 driverMPI(start, end, outMPI, cutoff);
341 if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&outMPI); delete distCalculator; return 0; }
345 //tell parent you are done.
346 MPI_Send(buf, 5, MPI_CHAR, 0, tag, MPI_COMM_WORLD);
349 MPI_File_close(&outMPI);
351 }else { //lower triangle format
352 if (pid == 0) { //you are the root process
356 unsigned long int mySize;
358 if (output != "square"){ driverMPI(start, end, outputFile, mySize); }
359 else { driverMPI(start, end, outputFile, mySize, output); }
361 if (m->control_pressed) { outputTypes.clear(); delete distCalculator; return 0; }
363 int amode=MPI_MODE_APPEND|MPI_MODE_WRONLY|MPI_MODE_CREATE; //
367 //char* filename = new char[outputFile.length()];
368 //memcpy(filename, outputFile.c_str(), outputFile.length());
371 strcpy(filename, outputFile.c_str());
373 MPI_File_open(MPI_COMM_SELF, filename, amode, MPI_INFO_NULL, &outMPI);
377 for(int b = 1; b < processors; b++) {
378 unsigned long int fileSize;
380 if (m->control_pressed) { outputTypes.clear(); MPI_File_close(&outMPI); delete distCalculator; return 0; }
382 MPI_Recv(&fileSize, 1, MPI_LONG, b, tag, MPI_COMM_WORLD, &status);
384 string outTemp = outputFile + toString(b) + ".temp";
386 char* buf = new char[outTemp.length()];
387 memcpy(buf, outTemp.c_str(), outTemp.length());
389 MPI_File_open(MPI_COMM_SELF, buf, MPI_MODE_DELETE_ON_CLOSE|MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI);
393 while (count < fileSize) {
395 MPI_File_read(inMPI, buf2, 1, MPI_CHAR, &status);
396 MPI_File_write(outMPI, buf2, 1, MPI_CHAR, &status);
400 MPI_File_close(&inMPI); //deleted on close
403 MPI_File_close(&outMPI);
404 }else { //you are a child process
406 unsigned long int size;
407 if (output != "square"){ driverMPI(start, end, (outputFile + toString(pid) + ".temp"), size); }
408 else { driverMPI(start, end, (outputFile + toString(pid) + ".temp"), size, output); }
410 if (m->control_pressed) { delete distCalculator; return 0; }
412 //tell parent you are done.
413 MPI_Send(&size, 1, MPI_LONG, 0, tag, MPI_COMM_WORLD);
416 MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
419 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
420 //if you don't need to fork anything
422 if (output != "square") { driver(0, numSeqs, outputFile, cutoff); }
423 else { driver(0, numSeqs, outputFile, "square"); }
424 }else{ //you have multiple processors
426 for (int i = 0; i < processors; i++) {
427 lines.push_back(new linePair());
428 if (output != "square") {
429 lines[i]->start = int (sqrt(float(i)/float(processors)) * numSeqs);
430 lines[i]->end = int (sqrt(float(i+1)/float(processors)) * numSeqs);
432 lines[i]->start = int ((float(i)/float(processors)) * numSeqs);
433 lines[i]->end = int ((float(i+1)/float(processors)) * numSeqs);
437 createProcesses(outputFile);
439 map<int, int>::iterator it = processIDS.begin();
440 rename((outputFile + toString(it->second) + ".temp").c_str(), outputFile.c_str());
443 //append and remove temp files
444 for (; it != processIDS.end(); it++) {
445 m->appendFiles((outputFile + toString(it->second) + ".temp"), outputFile);
446 remove((outputFile + toString(it->second) + ".temp").c_str());
451 if (output != "square") { driver(0, numSeqs, outputFile, cutoff); }
452 else { driver(0, numSeqs, outputFile, "square"); }
456 if (m->control_pressed) { outputTypes.clear(); delete distCalculator; remove(outputFile.c_str()); return 0; }
459 MPI_Comm_rank(MPI_COMM_WORLD, &pid);
461 if (pid == 0) { //only one process should output to screen
464 //if (output == "square") { convertMatrix(outputFile); }
467 fileHandle.open(outputFile.c_str());
469 m->gobble(fileHandle);
470 if (fileHandle.eof()) { m->mothurOut(outputFile + " is blank. This can result if there are no distances below your cutoff."); m->mothurOutEndLine(); }
473 //append the old column file to the new one
474 if ((oldfastafile != "") && (column != "")) {
475 //we had to rename the column file so we didnt overwrite above, but we want to keep old name
476 if (outputFile == column) {
477 string tempcolumn = column + ".old";
478 m->appendFiles(tempcolumn, outputFile);
479 remove(tempcolumn.c_str());
481 m->appendFiles(outputFile, column);
482 remove(outputFile.c_str());
486 if (outputDir != "") {
487 string newOutputName = outputDir + m->getSimpleName(outputFile);
488 rename(outputFile.c_str(), newOutputName.c_str());
489 remove(outputFile.c_str());
490 outputFile = newOutputName;
499 if (m->control_pressed) { outputTypes.clear(); delete distCalculator; remove(outputFile.c_str()); return 0; }
501 delete distCalculator;
503 m->mothurOutEndLine();
504 m->mothurOut("Output File Name: "); m->mothurOutEndLine();
505 m->mothurOut(outputFile); m->mothurOutEndLine();
506 m->mothurOutEndLine();
507 m->mothurOut("It took " + toString(time(NULL) - startTime) + " to calculate the distances for " + toString(numSeqs) + " sequences."); m->mothurOutEndLine();
510 if (m->isTrue(compress)) {
511 m->mothurOut("Compressing..."); m->mothurOutEndLine();
512 m->mothurOut("(Replacing " + outputFile + " with " + outputFile + ".gz)"); m->mothurOutEndLine();
513 system(("gzip -v " + outputFile).c_str());
514 outputNames.push_back(outputFile + ".gz");
515 }else { outputNames.push_back(outputFile); }
520 catch(exception& e) {
521 m->errorOut(e, "DistanceCommand", "execute");
525 /**************************************************************************************************/
526 void DistanceCommand::createProcesses(string filename) {
528 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
532 //loop through and create all the processes you want
533 while (process != processors) {
537 processIDS[lines[process]->end] = pid; //create map from line number to pid so you can append files in correct order later
540 if (output != "square") { driver(lines[process]->start, lines[process]->end, filename + toString(getpid()) + ".temp", cutoff); }
541 else { driver(lines[process]->start, lines[process]->end, filename + toString(getpid()) + ".temp", "square"); }
543 }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
546 //force parent to wait until all the processes are done
547 for (map<int, int>::iterator it = processIDS.begin(); it != processIDS.end(); it++) {
548 int temp = it->second;
553 catch(exception& e) {
554 m->errorOut(e, "DistanceCommand", "createProcesses");
559 /**************************************************************************************************/
560 /////// need to fix to work with calcs and sequencedb
561 int DistanceCommand::driver(int startLine, int endLine, string dFileName, float cutoff){
564 int startTime = time(NULL);
567 ofstream outFile(dFileName.c_str(), ios::trunc);
568 outFile.setf(ios::fixed, ios::showpoint);
569 outFile << setprecision(4);
571 if((output == "lt") && startLine == 0){ outFile << alignDB.getNumSeqs() << endl; }
573 for(int i=startLine;i<endLine;i++){
575 string name = alignDB.get(i).getName();
576 if (name.length() < 10) { //pad with spaces to make compatible
577 while (name.length() < 10) { name += " "; }
579 outFile << name << '\t';
581 for(int j=0;j<i;j++){
583 if (m->control_pressed) { outFile.close(); return 0; }
585 //if there was a column file given and we are appending, we don't want to calculate the distances that are already in the column file
586 //the alignDB contains the new sequences and then the old, so if i an oldsequence and j is an old sequence then break out of this loop
587 if ((i >= numNewFasta) && (j >= numNewFasta)) { break; }
589 distCalculator->calcDist(alignDB.get(i), alignDB.get(j));
590 double dist = distCalculator->getDist();
593 if (output == "column") { outFile << alignDB.get(i).getName() << ' ' << alignDB.get(j).getName() << ' ' << dist << endl; }
595 if (output == "lt") { outFile << dist << '\t'; }
598 if (output == "lt") { outFile << endl; }
601 m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
605 m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
611 catch(exception& e) {
612 m->errorOut(e, "DistanceCommand", "driver");
616 /**************************************************************************************************/
617 /////// need to fix to work with calcs and sequencedb
618 int DistanceCommand::driver(int startLine, int endLine, string dFileName, string square){
621 int startTime = time(NULL);
624 ofstream outFile(dFileName.c_str(), ios::trunc);
625 outFile.setf(ios::fixed, ios::showpoint);
626 outFile << setprecision(4);
628 if(startLine == 0){ outFile << alignDB.getNumSeqs() << endl; }
630 for(int i=startLine;i<endLine;i++){
632 string name = alignDB.get(i).getName();
633 //pad with spaces to make compatible
634 if (name.length() < 10) { while (name.length() < 10) { name += " "; } }
636 outFile << name << '\t';
638 for(int j=0;j<alignDB.getNumSeqs();j++){
640 if (m->control_pressed) { outFile.close(); return 0; }
642 distCalculator->calcDist(alignDB.get(i), alignDB.get(j));
643 double dist = distCalculator->getDist();
645 outFile << dist << '\t';
651 m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
655 m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
661 catch(exception& e) {
662 m->errorOut(e, "DistanceCommand", "driver");
667 /**************************************************************************************************/
668 /////// need to fix to work with calcs and sequencedb
669 int DistanceCommand::driverMPI(int startLine, int endLine, MPI_File& outMPI, float cutoff){
672 int startTime = time(NULL);
674 string outputString = "";
676 for(int i=startLine;i<endLine;i++){
678 for(int j=0;j<i;j++){
680 if (m->control_pressed) { return 0; }
682 //if there was a column file given and we are appending, we don't want to calculate the distances that are already in the column file
683 //the alignDB contains the new sequences and then the old, so if i an oldsequence and j is an old sequence then break out of this loop
684 if ((i >= numNewFasta) && (j >= numNewFasta)) { break; }
686 distCalculator->calcDist(alignDB.get(i), alignDB.get(j));
687 double dist = distCalculator->getDist();
690 outputString += (alignDB.get(i).getName() + ' ' + alignDB.get(j).getName() + ' ' + toString(dist) + '\n');
695 //m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
696 cout << i << '\t' << (time(NULL) - startTime) << endl;
700 //send results to parent
701 int length = outputString.length();
703 char* buf = new char[length];
704 memcpy(buf, outputString.c_str(), length);
706 MPI_File_write_shared(outMPI, buf, length, MPI_CHAR, &status);
712 //m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
713 cout << (endLine-1) << '\t' << (time(NULL) - startTime) << endl;
716 catch(exception& e) {
717 m->errorOut(e, "DistanceCommand", "driverMPI");
721 /**************************************************************************************************/
722 /////// need to fix to work with calcs and sequencedb
723 int DistanceCommand::driverMPI(int startLine, int endLine, string file, unsigned long int& size){
728 int amode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
730 //char* filename = new char[file.length()];
731 //memcpy(filename, file.c_str(), file.length());
734 strcpy(filename, file.c_str());
736 MPI_File_open(MPI_COMM_SELF, filename, amode, MPI_INFO_NULL, &outMPI);
739 int startTime = time(NULL);
741 string outputString = "";
744 if(startLine == 0){ outputString += toString(alignDB.getNumSeqs()) + "\n"; }
746 for(int i=startLine;i<endLine;i++){
748 string name = alignDB.get(i).getName();
749 if (name.length() < 10) { //pad with spaces to make compatible
750 while (name.length() < 10) { name += " "; }
752 outputString += name + "\t";
754 for(int j=0;j<i;j++){
756 if (m->control_pressed) { return 0; }
758 distCalculator->calcDist(alignDB.get(i), alignDB.get(j));
759 double dist = distCalculator->getDist();
761 outputString += toString(dist) + "\t";
764 outputString += "\n";
768 //m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
769 cout << i << '\t' << (time(NULL) - startTime) << endl;
773 //send results to parent
774 int length = outputString.length();
775 char* buf = new char[length];
776 memcpy(buf, outputString.c_str(), length);
778 MPI_File_write(outMPI, buf, length, MPI_CHAR, &status);
779 size += outputString.length();
784 //m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
785 cout << (endLine-1) << '\t' << (time(NULL) - startTime) << endl;
786 MPI_File_close(&outMPI);
790 catch(exception& e) {
791 m->errorOut(e, "DistanceCommand", "driverMPI");
795 /**************************************************************************************************/
796 /////// need to fix to work with calcs and sequencedb
797 int DistanceCommand::driverMPI(int startLine, int endLine, string file, unsigned long int& size, string square){
802 int amode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
804 //char* filename = new char[file.length()];
805 //memcpy(filename, file.c_str(), file.length());
808 strcpy(filename, file.c_str());
810 MPI_File_open(MPI_COMM_SELF, filename, amode, MPI_INFO_NULL, &outMPI);
813 int startTime = time(NULL);
815 string outputString = "";
818 if(startLine == 0){ outputString += toString(alignDB.getNumSeqs()) + "\n"; }
820 for(int i=startLine;i<endLine;i++){
822 string name = alignDB.get(i).getName();
823 if (name.length() < 10) { //pad with spaces to make compatible
824 while (name.length() < 10) { name += " "; }
826 outputString += name + "\t";
828 for(int j=0;j<alignDB.getNumSeqs();j++){
830 if (m->control_pressed) { return 0; }
832 distCalculator->calcDist(alignDB.get(i), alignDB.get(j));
833 double dist = distCalculator->getDist();
835 outputString += toString(dist) + "\t";
838 outputString += "\n";
842 //m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
843 cout << i << '\t' << (time(NULL) - startTime) << endl;
847 //send results to parent
848 int length = outputString.length();
849 char* buf = new char[length];
850 memcpy(buf, outputString.c_str(), length);
852 MPI_File_write(outMPI, buf, length, MPI_CHAR, &status);
853 size += outputString.length();
858 //m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
859 cout << (endLine-1) << '\t' << (time(NULL) - startTime) << endl;
860 MPI_File_close(&outMPI);
864 catch(exception& e) {
865 m->errorOut(e, "DistanceCommand", "driverMPI");
870 /**************************************************************************************************
871 int DistanceCommand::convertMatrix(string outputFile) {
874 //sort file by first column so the distances for each row are together
875 string outfile = m->getRootName(outputFile) + "sorted.dist.temp";
878 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
879 string command = "sort -n " + outputFile + " -o " + outfile;
880 system(command.c_str());
881 #else //sort using windows sort
882 string command = "sort " + outputFile + " /O " + outfile;
883 system(command.c_str());
887 //output to new file distance for each row and save positions in file where new row begins
889 m->openInputFile(outfile, in);
892 m->openOutputFile(outputFile, out);
894 out.setf(ios::fixed, ios::floatfield); out.setf(ios::showpoint);
896 out << alignDB.getNumSeqs() << endl;
898 //get first currentRow
899 string first, currentRow, second;
901 map<string, float> rowDists; //take advantage of the fact that maps are already sorted by key
902 map<string, float>::iterator it;
907 rowDists[first] = 0.00; //distance to yourself is 0.0
910 //m->openInputFile(outfile, in);
913 if (m->control_pressed) { in.close(); remove(outfile.c_str()); out.close(); return 0; }
915 in >> first >> second >> dist; m->gobble(in);
917 if (first != currentRow) {
919 out << currentRow << '\t'; //print name
922 for (it = rowDists.begin(); it != rowDists.end(); it++) {
923 out << it->second << '\t';
930 rowDists[first] = 0.00;
931 rowDists[second] = dist;
933 rowDists[second] = dist;
937 out << currentRow << '\t'; //print name
940 for (it = rowDists.begin(); it != rowDists.end(); it++) {
941 out << it->second << '\t';
948 remove(outfile.c_str());
953 catch(exception& e) {
954 m->errorOut(e, "DistanceCommand", "convertMatrix");
958 /**************************************************************************************************
959 int DistanceCommand::convertToLowerTriangle(string outputFile) {
962 //sort file by first column so the distances for each row are together
963 string outfile = m->getRootName(outputFile) + "sorted.dist.temp";
966 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
967 string command = "sort -n " + outputFile + " -o " + outfile;
968 system(command.c_str());
969 #else //sort using windows sort
970 string command = "sort " + outputFile + " /O " + outfile;
971 system(command.c_str());
975 //output to new file distance for each row and save positions in file where new row begins
977 m->openInputFile(outfile, in);
980 m->openOutputFile(outputFile, out);
982 out.setf(ios::fixed, ios::floatfield); out.setf(ios::showpoint);
984 out << alignDB.getNumSeqs() << endl;
986 //get first currentRow
987 string first, currentRow, second;
991 map<string, float> rowDists; //take advantage of the fact that maps are already sorted by key
992 map<string, float>::iterator it;
997 rowDists[first] = 0.00; //distance to yourself is 0.0
1000 //m->openInputFile(outfile, in);
1003 if (m->control_pressed) { in.close(); remove(outfile.c_str()); out.close(); return 0; }
1005 in >> first >> second >> dist; m->gobble(in);
1007 if (first != currentRow) {
1008 //print out last row
1009 out << currentRow << '\t'; //print name
1012 for (it = rowDists.begin(); it != rowDists.end(); it++) {
1013 if (j >= i) { break; }
1014 out << it->second << '\t';
1022 rowDists[first] = 0.00;
1023 rowDists[second] = dist;
1027 rowDists[second] = dist;
1030 //print out last row
1031 out << currentRow << '\t'; //print name
1034 for (it = rowDists.begin(); it != rowDists.end(); it++) {
1035 out << it->second << '\t';
1042 remove(outfile.c_str());
1047 catch(exception& e) {
1048 m->errorOut(e, "DistanceCommand", "convertToLowerTriangle");
1052 /**************************************************************************************************/
1053 //its okay if the column file does not contain all the names in the fasta file, since some distance may have been above a cutoff,
1054 //but no sequences can be in the column file that are not in oldfasta. also, if a distance is above the cutoff given then remove it.
1055 //also check to make sure the 2 files have the same alignment length.
1056 bool DistanceCommand::sanityCheck() {
1060 //make sure the 2 fasta files have the same alignment length
1062 m->openInputFile(fastafile, in);
1063 int fastaAlignLength = 0;
1065 Sequence tempIn(in);
1066 fastaAlignLength = tempIn.getAligned().length();
1071 m->openInputFile(oldfastafile, in2);
1072 int oldfastaAlignLength = 0;
1074 Sequence tempIn2(in2);
1075 oldfastaAlignLength = tempIn2.getAligned().length();
1079 if (fastaAlignLength != oldfastaAlignLength) { m->mothurOut("fasta files do not have the same alignment length."); m->mothurOutEndLine(); return false; }
1081 //read fasta file and save names as well as adding them to the alignDB
1082 set<string> namesOldFasta;
1085 m->openInputFile(oldfastafile, inFasta);
1087 while (!inFasta.eof()) {
1088 if (m->control_pressed) { inFasta.close(); return good; }
1090 Sequence temp(inFasta);
1092 if (temp.getName() != "") {
1093 namesOldFasta.insert(temp.getName()); //save name
1094 alignDB.push_back(temp); //add to DB
1102 //read through the column file checking names and removing distances above the cutoff
1104 m->openInputFile(column, inDist);
1107 string outputFile = column + ".temp";
1108 m->openOutputFile(outputFile, outDist);
1110 string name1, name2;
1112 while (!inDist.eof()) {
1113 if (m->control_pressed) { inDist.close(); outDist.close(); remove(outputFile.c_str()); return good; }
1115 inDist >> name1 >> name2 >> dist; m->gobble(inDist);
1117 //both names are in fasta file and distance is below cutoff
1118 if ((namesOldFasta.count(name1) == 0) || (namesOldFasta.count(name2) == 0)) { good = false; break; }
1120 if (dist <= cutoff) {
1121 outDist << name1 << '\t' << name2 << '\t' << dist << endl;
1130 remove(column.c_str());
1131 rename(outputFile.c_str(), column.c_str());
1133 remove(outputFile.c_str()); //temp file is bad because file mismatch above
1137 catch(exception& e) {
1138 m->errorOut(e, "DistanceCommand", "m->appendFiles");
1142 /**************************************************************************************************/