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 //**********************************************************************************************************************
19 DistanceCommand::DistanceCommand(string option) {
24 //allow user to run help
25 if(option == "help") { help(); abort = true; }
28 //valid paramters for this command
29 string Array[] = {"fasta", "output", "calc", "countends", "cutoff", "processors", "outputdir","inputdir"};
30 vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
32 OptionParser parser(option);
33 map<string, string> parameters = parser.getParameters();
35 ValidParameters validParameter;
36 map<string, string>::iterator it2;
38 //check to make sure all parameters are valid for command
39 for (it2 = parameters.begin(); it2 != parameters.end(); it2++) {
40 if (validParameter.isValidParameter(it2->first, myArray, it2->second) != true) { abort = true; }
43 //if the user changes the input directory command factory will send this info to us in the output parameter
44 string inputDir = validParameter.validFile(parameters, "inputdir", false);
45 if (inputDir == "not found"){ inputDir = ""; }
48 it2 = parameters.find("fasta");
49 //user has given a template file
50 if(it2 != parameters.end()){
51 path = hasPath(it2->second);
52 //if the user has not given a path then, add inputdir. else leave path alone.
53 if (path == "") { parameters["fasta"] = inputDir + it2->second; }
57 //check for required parameters
58 fastafile = validParameter.validFile(parameters, "fasta", true);
59 if (fastafile == "not found") { m->mothurOut("fasta is a required parameter for the dist.seqs command."); m->mothurOutEndLine(); abort = true; }
60 else if (fastafile == "not open") { abort = true; }
63 openInputFile(fastafile, inFASTA);
64 alignDB = SequenceDB(inFASTA);
68 //if the user changes the output directory command factory will send this info to us in the output parameter
69 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){
71 outputDir += hasPath(fastafile); //if user entered a file with a path then preserve it
74 //check for optional parameter and set defaults
75 // ...at some point should added some additional type checking...
76 calc = validParameter.validFile(parameters, "calc", false);
77 if (calc == "not found") { calc = "onegap"; }
79 if (calc == "default") { calc = "onegap"; }
81 splitAtDash(calc, Estimators);
84 temp = validParameter.validFile(parameters, "countends", false); if(temp == "not found"){ temp = "T"; }
85 convert(temp, countends);
87 temp = validParameter.validFile(parameters, "cutoff", false); if(temp == "not found"){ temp = "1.0"; }
88 convert(temp, cutoff);
90 temp = validParameter.validFile(parameters, "processors", false); if(temp == "not found"){ temp = "1"; }
91 convert(temp, processors);
93 output = validParameter.validFile(parameters, "output", false); if(output == "not found"){ output = "column"; }
95 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"; }
97 ValidCalculators validCalculator;
99 if (isTrue(countends) == true) {
100 for (int i=0; i<Estimators.size(); i++) {
101 if (validCalculator.isValidCalculator("distance", Estimators[i]) == true) {
102 if (Estimators[i] == "nogaps") { distCalculator = new ignoreGaps(); }
103 else if (Estimators[i] == "eachgap") { distCalculator = new eachGapDist(); }
104 else if (Estimators[i] == "onegap") { distCalculator = new oneGapDist(); }
108 for (int i=0; i<Estimators.size(); i++) {
109 if (validCalculator.isValidCalculator("distance", Estimators[i]) == true) {
110 if (Estimators[i] == "nogaps") { distCalculator = new ignoreGaps(); }
111 else if (Estimators[i] == "eachgap"){ distCalculator = new eachGapIgnoreTermGapDist(); }
112 else if (Estimators[i] == "onegap") { distCalculator = new oneGapIgnoreTermGapDist(); }
120 catch(exception& e) {
121 m->errorOut(e, "DistanceCommand", "DistanceCommand");
126 //**********************************************************************************************************************
128 DistanceCommand::~DistanceCommand(){
130 for(int i=0;i<lines.size();i++){
136 //**********************************************************************************************************************
138 void DistanceCommand::help(){
140 m->mothurOut("The dist.seqs command reads a file containing sequences and creates a distance file.\n");
141 m->mothurOut("The dist.seqs command parameters are fasta, calc, countends, output, cutoff and processors. \n");
142 m->mothurOut("The fasta parameter is required.\n");
143 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");
144 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");
145 m->mothurOut("The cutoff parameter allows you to specify maximum distance to keep. The default is 1.0.\n");
146 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");
147 m->mothurOut("The processors parameter allows you to specify number of processors to use. The default is 1.\n");
148 m->mothurOut("The dist.seqs command should be in the following format: \n");
149 m->mothurOut("dist.seqs(fasta=yourFastaFile, calc=yourCalc, countends=yourEnds, cutoff= yourCutOff, processors=yourProcessors) \n");
150 m->mothurOut("Example dist.seqs(fasta=amazon.fasta, calc=eachgap, countends=F, cutoff= 2.0, processors=3).\n");
151 m->mothurOut("Note: No spaces between parameter labels (i.e. calc), '=' and parameters (i.e.yourCalc).\n\n");
153 catch(exception& e) {
154 m->errorOut(e, "DistanceCommand", "help");
158 //**********************************************************************************************************************
160 int DistanceCommand::execute(){
163 if (abort == true) { return 0; }
165 int numSeqs = alignDB.getNumSeqs();
170 if (output == "lt") { //does the user want lower triangle phylip formatted file
171 outputFile = outputDir + getRootName(getSimpleName(fastafile)) + "phylip.dist";
172 remove(outputFile.c_str());
174 //output numSeqs to phylip formatted dist file
175 }else if (output == "column") { //user wants column format
176 outputFile = outputDir + getRootName(getSimpleName(fastafile)) + "dist";
177 remove(outputFile.c_str());
178 }else { //assume square
179 outputFile = outputDir + getRootName(getSimpleName(fastafile)) + "square.dist";
180 remove(outputFile.c_str());
190 MPI_Comm_size(MPI_COMM_WORLD, &processors); //set processors to the number of mpi processes running
191 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
193 //each process gets where it should start and stop in the file
194 start = int (sqrt(float(pid)/float(processors)) * numSeqs);
195 end = int (sqrt(float(pid+1)/float(processors)) * numSeqs);
197 if (pid == 0) { //you are the root process
200 driverMPI(start, end, outputMyPart, cutoff);
203 openOutputFile(outputFile, out);
207 //get the childrens parts
208 for(int i = 1; i < processors; i++) {
210 MPI_Recv(&length, 1, MPI_INT, i, tag, MPI_COMM_WORLD, &status);
214 MPI_Recv(buf, length, MPI_CHAR, i, tag, MPI_COMM_WORLD, &status);
222 }else { //you are a child process
225 driverMPI(start, end, outputMyPart, cutoff);
227 //send results to parent
228 int length = outputMyPart.length();
230 strcpy(buf, outputMyPart.c_str());
232 MPI_Send( &length, 1, MPI_INT, 0, tag, MPI_COMM_WORLD);
233 MPI_Send(buf, length, MPI_CHAR, 0, tag, MPI_COMM_WORLD);
239 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
240 //if you don't need to fork anything
242 driver(0, numSeqs, outputFile, cutoff);
243 }else{ //you have multiple processors
245 for (int i = 0; i < processors; i++) {
246 lines.push_back(new linePair());
247 lines[i]->start = int (sqrt(float(i)/float(processors)) * numSeqs);
248 lines[i]->end = int (sqrt(float(i+1)/float(processors)) * numSeqs);
251 createProcesses(outputFile);
253 map<int, int>::iterator it = processIDS.begin();
254 rename((outputFile + toString(it->second) + ".temp").c_str(), outputFile.c_str());
257 //append and remove temp files
258 for (; it != processIDS.end(); it++) {
259 appendFiles((outputFile + toString(it->second) + ".temp"), outputFile);
260 remove((outputFile + toString(it->second) + ".temp").c_str());
265 driver(0, numSeqs, outputFile, cutoff);
269 if (m->control_pressed) { delete distCalculator; remove(outputFile.c_str()); return 0; }
272 MPI_Comm_rank(MPI_COMM_WORLD, &pid);
274 if (pid == 0) { //only one process should output to screen
277 if (output == "square") { convertMatrix(outputFile); }
283 if (m->control_pressed) { delete distCalculator; remove(outputFile.c_str()); return 0; }
285 delete distCalculator;
287 m->mothurOutEndLine();
288 m->mothurOut("Output File Name: "); m->mothurOutEndLine();
289 m->mothurOut(outputFile); m->mothurOutEndLine();
290 m->mothurOutEndLine();
295 catch(exception& e) {
296 m->errorOut(e, "DistanceCommand", "execute");
300 /**************************************************************************************************/
301 void DistanceCommand::createProcesses(string filename) {
303 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
307 //loop through and create all the processes you want
308 while (process != processors) {
312 processIDS[lines[process]->end] = pid; //create map from line number to pid so you can append files in correct order later
315 driver(lines[process]->start, lines[process]->end, filename + toString(getpid()) + ".temp", cutoff);
317 }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
320 //force parent to wait until all the processes are done
321 for (map<int, int>::iterator it = processIDS.begin(); it != processIDS.end(); it++) {
322 int temp = it->second;
327 catch(exception& e) {
328 m->errorOut(e, "DistanceCommand", "createProcesses");
333 /**************************************************************************************************/
334 /////// need to fix to work with calcs and sequencedb
335 int DistanceCommand::driver(int startLine, int endLine, string dFileName, float cutoff){
338 int startTime = time(NULL);
341 ofstream outFile(dFileName.c_str(), ios::trunc);
342 outFile.setf(ios::fixed, ios::showpoint);
343 outFile << setprecision(4);
345 if((output == "lt") && startLine == 0){ outFile << alignDB.getNumSeqs() << endl; }
347 for(int i=startLine;i<endLine;i++){
349 string name = alignDB.get(i).getName();
350 if (name.length() < 10) { //pad with spaces to make compatible
351 while (name.length() < 10) { name += " "; }
353 outFile << name << '\t';
355 for(int j=0;j<i;j++){
357 if (m->control_pressed) { outFile.close(); return 0; }
359 distCalculator->calcDist(alignDB.get(i), alignDB.get(j));
360 double dist = distCalculator->getDist();
363 if (output == "column") { outFile << alignDB.get(i).getName() << ' ' << alignDB.get(j).getName() << ' ' << dist << endl; }
365 if (output == "lt") { outFile << dist << '\t'; }
367 if (output == "square") { //make a square column you can convert to square phylip
368 outFile << alignDB.get(i).getName() << '\t' << alignDB.get(j).getName() << '\t' << dist << endl;
369 outFile << alignDB.get(j).getName() << '\t' << alignDB.get(i).getName() << '\t' << dist << endl;
374 if (output == "lt") { outFile << endl; }
377 m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
381 m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
387 catch(exception& e) {
388 m->errorOut(e, "DistanceCommand", "driver");
392 /**************************************************************************************************/
393 /////// need to fix to work with calcs and sequencedb
394 int DistanceCommand::driverMPI(int startLine, int endLine, string& outputString, float cutoff){
397 int startTime = time(NULL);
401 if((output == "lt") && startLine == 0){ outputString += (toString(alignDB.getNumSeqs()) + '\n'); }
403 for(int i=startLine;i<endLine;i++){
406 string name = alignDB.get(i).getName();
407 if (name.length() < 10) { //pad with spaces to make compatible
408 while (name.length() < 10) { name += " "; }
410 outputString += (name + '\t');
412 for(int j=0;j<i;j++){
414 if (m->control_pressed) { return 0; }
416 distCalculator->calcDist(alignDB.get(i), alignDB.get(j));
417 double dist = distCalculator->getDist();
420 if (output == "column") { outputString += (alignDB.get(i).getName() + ' ' + alignDB.get(j).getName() + ' ' + toString(dist) + '\n'); }
422 if (output == "lt") { outputString += (toString(dist) + '\t'); }
424 if (output == "square") { //make a square column you can convert to square phylip
425 outputString += (alignDB.get(i).getName() + ' ' + alignDB.get(j).getName() + ' ' + toString(dist) + '\n');
426 outputString += (alignDB.get(j).getName() + ' ' + alignDB.get(i).getName() + ' ' + toString(dist) + '\n');
431 if (output == "lt") { outputString += '\n'; }
434 m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
438 m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
442 catch(exception& e) {
443 m->errorOut(e, "DistanceCommand", "driver");
448 /**************************************************************************************************/
449 int DistanceCommand::convertMatrix(string outputFile) {
452 //sort file by first column so the distances for each row are together
453 string outfile = getRootName(outputFile) + "sorted.dist.temp";
456 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
457 string command = "sort -n " + outputFile + " -o " + outfile;
458 system(command.c_str());
459 #else //sort using windows sort
460 string command = "sort " + outputFile + " /O " + outfile;
461 system(command.c_str());
465 //output to new file distance for each row and save positions in file where new row begins
467 openInputFile(outfile, in);
470 openOutputFile(outputFile, out);
472 out.setf(ios::fixed, ios::floatfield); out.setf(ios::showpoint);
474 out << alignDB.getNumSeqs() << endl;
476 //get first currentRow
477 string first, currentRow, second;
479 map<string, float> rowDists; //take advantage of the fact that maps are already sorted by key
480 map<string, float>::iterator it;
485 rowDists[first] = 0.00; //distance to yourself is 0.0
488 //openInputFile(outfile, in);
491 if (m->control_pressed) { in.close(); remove(outfile.c_str()); out.close(); return 0; }
493 in >> first >> second >> dist; gobble(in);
495 if (first != currentRow) {
497 out << currentRow << '\t'; //print name
500 for (it = rowDists.begin(); it != rowDists.end(); it++) {
501 out << it->second << '\t';
508 rowDists[first] = 0.00;
509 rowDists[second] = dist;
511 rowDists[second] = dist;
515 out << currentRow << '\t'; //print name
518 for (it = rowDists.begin(); it != rowDists.end(); it++) {
519 out << it->second << '\t';
526 remove(outfile.c_str());
531 catch(exception& e) {
532 m->errorOut(e, "DistanceCommand", "convertMatrix");
536 /**************************************************************************************************
537 void DistanceCommand::appendFiles(string temp, string filename) {
542 //open output file in append mode
543 openOutputFileAppend(filename, output);
544 openInputFile(temp, input);
546 while(char c = input.get()){
547 if(input.eof()) { break; }
548 else { output << c; }
554 catch(exception& e) {
555 m->errorOut(e, "DistanceCommand", "appendFiles");
559 /**************************************************************************************************/