2 * matrixoutputcommand.cpp
5 * Created by Sarah Westcott on 5/20/09.
6 * Copyright 2009 Schloss Lab UMASS Amherst. All rights reserved.
10 #include "matrixoutputcommand.h"
11 #include "subsample.h"
13 //**********************************************************************************************************************
14 vector<string> MatrixOutputCommand::setParameters(){
16 CommandParameter pshared("shared", "InputTypes", "", "", "none", "none", "none","phylip",false,true,true); parameters.push_back(pshared);
17 CommandParameter plabel("label", "String", "", "", "", "", "","",false,false); parameters.push_back(plabel);
18 CommandParameter psubsample("subsample", "String", "", "", "", "", "","",false,false); parameters.push_back(psubsample);
19 CommandParameter pgroups("groups", "String", "", "", "", "", "","",false,false); parameters.push_back(pgroups);
20 CommandParameter pcalc("calc", "Multiple", "sharedsobs-sharedchao-sharedace-jabund-sorabund-jclass-sorclass-jest-sorest-thetayc-thetan-kstest-sharednseqs-ochiai-anderberg-kulczynski-kulczynskicody-lennon-morisitahorn-braycurtis-whittaker-odum-canberra-structeuclidean-structchord-hellinger-manhattan-structpearson-soergel-spearman-structkulczynski-speciesprofile-hamming-structchi2-gower-memchi2-memchord-memeuclidean-mempearson-jsd-rjsd", "jclass-thetayc", "", "", "","",true,false,true); parameters.push_back(pcalc);
21 CommandParameter poutput("output", "Multiple", "lt-square-column", "lt", "", "", "","",false,false); parameters.push_back(poutput);
22 CommandParameter pmode("mode", "Multiple", "average-median", "average", "", "", "","",false,false); parameters.push_back(pmode);
23 CommandParameter pprocessors("processors", "Number", "", "1", "", "", "","",false,false,true); parameters.push_back(pprocessors);
24 CommandParameter piters("iters", "Number", "", "1000", "", "", "","",false,false); parameters.push_back(piters);
25 CommandParameter pinputdir("inputdir", "String", "", "", "", "", "","",false,false); parameters.push_back(pinputdir);
26 CommandParameter poutputdir("outputdir", "String", "", "", "", "", "","",false,false); parameters.push_back(poutputdir);
28 vector<string> myArray;
29 for (int i = 0; i < parameters.size(); i++) { myArray.push_back(parameters[i].name); }
33 m->errorOut(e, "MatrixOutputCommand", "setParameters");
37 //**********************************************************************************************************************
38 string MatrixOutputCommand::getHelpString(){
40 string helpString = "";
41 ValidCalculators validCalculator;
42 helpString += "The dist.shared command parameters are shared, groups, calc, output, processors, subsample, iters, mode, and label. shared is a required, unless you have a valid current file.\n";
43 helpString += "The groups parameter allows you to specify which of the groups in your groupfile you would like included used.\n";
44 helpString += "The group names are separated by dashes. The label parameter allows you to select what distance levels you would like distance matrices created for, and is also separated by dashes.\n";
45 helpString += "The iters parameter allows you to choose the number of times you would like to run the subsample.\n";
46 helpString += "The subsample parameter allows you to enter the size pergroup of the sample or you can set subsample=T and mothur will use the size of your smallest group.\n";
47 helpString += "The dist.shared command should be in the following format: dist.shared(groups=yourGroups, calc=yourCalcs, label=yourLabels).\n";
48 helpString += "The output parameter allows you to specify format of your distance matrix. Options are lt, column and square. The default is lt.\n";
49 helpString += "The mode parameter allows you to specify if you want the average or the median values reported when subsampling. Options are average, and median. The default is average.\n";
50 helpString += "Example dist.shared(groups=A-B-C, calc=jabund-sorabund).\n";
51 helpString += "The default value for groups is all the groups in your groupfile.\n";
52 helpString += "The default value for calc is jclass and thetayc.\n";
53 helpString += validCalculator.printCalc("matrix");
54 helpString += "The dist.shared command outputs a .dist file for each calculator you specify at each distance you choose.\n";
55 helpString += "Note: No spaces between parameter labels (i.e. groups), '=' and parameters (i.e.yourGroups).\n";
59 m->errorOut(e, "MatrixOutputCommand", "getHelpString");
63 //**********************************************************************************************************************
64 string MatrixOutputCommand::getOutputPattern(string type) {
68 if (type == "phylip") { pattern = "[filename],[calc],[distance],[outputtag],dist-[filename],[calc],[distance],[outputtag],[tag2],dist"; }
69 else { m->mothurOut("[ERROR]: No definition for type " + type + " output pattern.\n"); m->control_pressed = true; }
74 m->errorOut(e, "MatrixOutputCommand", "getOutputPattern");
78 //**********************************************************************************************************************
79 MatrixOutputCommand::MatrixOutputCommand(){
81 abort = true; calledHelp = true;
83 vector<string> tempOutNames;
84 outputTypes["phylip"] = tempOutNames;
87 m->errorOut(e, "MatrixOutputCommand", "MatrixOutputCommand");
91 //**********************************************************************************************************************
93 MatrixOutputCommand::MatrixOutputCommand(string option) {
95 abort = false; calledHelp = false;
98 //allow user to run help
99 if(option == "help") { help(); abort = true; calledHelp = true; }
100 else if(option == "citation") { citation(); abort = true; calledHelp = true;}
103 vector<string> myArray = setParameters();
105 OptionParser parser(option);
106 map<string,string> parameters = parser.getParameters();
107 map<string,string>::iterator it;
109 ValidParameters validParameter;
111 //check to make sure all parameters are valid for command
112 for (it = parameters.begin(); it != parameters.end(); it++) {
113 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
116 //initialize outputTypes
117 vector<string> tempOutNames;
118 outputTypes["phylip"] = tempOutNames;
120 //if the user changes the input directory command factory will send this info to us in the output parameter
121 string inputDir = validParameter.validFile(parameters, "inputdir", false);
122 if (inputDir == "not found"){ inputDir = ""; }
125 it = parameters.find("shared");
126 //user has given a template file
127 if(it != parameters.end()){
128 path = m->hasPath(it->second);
129 //if the user has not given a path then, add inputdir. else leave path alone.
130 if (path == "") { parameters["shared"] = inputDir + it->second; }
134 sharedfile = validParameter.validFile(parameters, "shared", true);
135 if (sharedfile == "not found") {
136 //if there is a current shared file, use it
137 sharedfile = m->getSharedFile();
138 if (sharedfile != "") { m->mothurOut("Using " + sharedfile + " as input file for the shared parameter."); m->mothurOutEndLine(); }
139 else { m->mothurOut("You have no current sharedfile and the shared parameter is required."); m->mothurOutEndLine(); abort = true; }
140 }else if (sharedfile == "not open") { sharedfile = ""; abort = true; }
141 else { m->setSharedFile(sharedfile); }
143 //if the user changes the output directory command factory will send this info to us in the output parameter
144 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){
146 outputDir += m->hasPath(sharedfile); //if user entered a file with a path then preserve it
149 //check for optional parameter and set defaults
150 // ...at some point should added some additional type checking...
151 label = validParameter.validFile(parameters, "label", false);
152 if (label == "not found") { label = ""; }
154 if(label != "all") { m->splitAtDash(label, labels); allLines = 0; }
155 else { allLines = 1; }
158 output = validParameter.validFile(parameters, "output", false); if(output == "not found"){ output = "lt"; }
159 if ((output != "lt") && (output != "square") && (output != "column")) { m->mothurOut(output + " is not a valid output form. Options are lt, column and square. I will use lt."); m->mothurOutEndLine(); output = "lt"; }
161 mode = validParameter.validFile(parameters, "mode", false); if(mode == "not found"){ mode = "average"; }
162 if ((mode != "average") && (mode != "median")) { m->mothurOut(mode + " is not a valid mode. Options are average and medina. I will use average."); m->mothurOutEndLine(); output = "average"; }
164 groups = validParameter.validFile(parameters, "groups", false);
165 if (groups == "not found") { groups = ""; }
167 m->splitAtDash(groups, Groups);
168 m->setGroups(Groups);
171 string temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); }
172 m->setProcessors(temp);
173 m->mothurConvert(temp, processors);
175 calc = validParameter.validFile(parameters, "calc", false);
176 if (calc == "not found") { calc = "jclass-thetayc"; }
178 if (calc == "default") { calc = "jclass-thetayc"; }
180 m->splitAtDash(calc, Estimators);
181 if (m->inUsersGroups("citation", Estimators)) {
182 ValidCalculators validCalc; validCalc.printCitations(Estimators);
183 //remove citation from list of calcs
184 for (int i = 0; i < Estimators.size(); i++) { if (Estimators[i] == "citation") { Estimators.erase(Estimators.begin()+i); break; } }
187 temp = validParameter.validFile(parameters, "iters", false); if (temp == "not found") { temp = "1000"; }
188 m->mothurConvert(temp, iters);
190 temp = validParameter.validFile(parameters, "subsample", false); if (temp == "not found") { temp = "F"; }
191 if (m->isNumeric1(temp)) { m->mothurConvert(temp, subsampleSize); subsample = true; }
193 if (m->isTrue(temp)) { subsample = true; subsampleSize = -1; } //we will set it to smallest group later
194 else { subsample = false; }
197 if (subsample == false) { iters = 0; }
199 if (abort == false) {
201 ValidCalculators validCalculator;
204 for (i=0; i<Estimators.size(); i++) {
205 if (validCalculator.isValidCalculator("matrix", Estimators[i]) == true) {
206 if (Estimators[i] == "sharedsobs") {
207 matrixCalculators.push_back(new SharedSobsCS());
208 }else if (Estimators[i] == "sharedchao") {
209 matrixCalculators.push_back(new SharedChao1());
210 }else if (Estimators[i] == "sharedace") {
211 matrixCalculators.push_back(new SharedAce());
212 }else if (Estimators[i] == "jabund") {
213 matrixCalculators.push_back(new JAbund());
214 }else if (Estimators[i] == "sorabund") {
215 matrixCalculators.push_back(new SorAbund());
216 }else if (Estimators[i] == "jclass") {
217 matrixCalculators.push_back(new Jclass());
218 }else if (Estimators[i] == "sorclass") {
219 matrixCalculators.push_back(new SorClass());
220 }else if (Estimators[i] == "jest") {
221 matrixCalculators.push_back(new Jest());
222 }else if (Estimators[i] == "sorest") {
223 matrixCalculators.push_back(new SorEst());
224 }else if (Estimators[i] == "thetayc") {
225 matrixCalculators.push_back(new ThetaYC());
226 }else if (Estimators[i] == "thetan") {
227 matrixCalculators.push_back(new ThetaN());
228 }else if (Estimators[i] == "kstest") {
229 matrixCalculators.push_back(new KSTest());
230 }else if (Estimators[i] == "sharednseqs") {
231 matrixCalculators.push_back(new SharedNSeqs());
232 }else if (Estimators[i] == "ochiai") {
233 matrixCalculators.push_back(new Ochiai());
234 }else if (Estimators[i] == "anderberg") {
235 matrixCalculators.push_back(new Anderberg());
236 }else if (Estimators[i] == "kulczynski") {
237 matrixCalculators.push_back(new Kulczynski());
238 }else if (Estimators[i] == "kulczynskicody") {
239 matrixCalculators.push_back(new KulczynskiCody());
240 }else if (Estimators[i] == "lennon") {
241 matrixCalculators.push_back(new Lennon());
242 }else if (Estimators[i] == "morisitahorn") {
243 matrixCalculators.push_back(new MorHorn());
244 }else if (Estimators[i] == "braycurtis") {
245 matrixCalculators.push_back(new BrayCurtis());
246 }else if (Estimators[i] == "whittaker") {
247 matrixCalculators.push_back(new Whittaker());
248 }else if (Estimators[i] == "odum") {
249 matrixCalculators.push_back(new Odum());
250 }else if (Estimators[i] == "canberra") {
251 matrixCalculators.push_back(new Canberra());
252 }else if (Estimators[i] == "structeuclidean") {
253 matrixCalculators.push_back(new StructEuclidean());
254 }else if (Estimators[i] == "structchord") {
255 matrixCalculators.push_back(new StructChord());
256 }else if (Estimators[i] == "hellinger") {
257 matrixCalculators.push_back(new Hellinger());
258 }else if (Estimators[i] == "manhattan") {
259 matrixCalculators.push_back(new Manhattan());
260 }else if (Estimators[i] == "structpearson") {
261 matrixCalculators.push_back(new StructPearson());
262 }else if (Estimators[i] == "soergel") {
263 matrixCalculators.push_back(new Soergel());
264 }else if (Estimators[i] == "spearman") {
265 matrixCalculators.push_back(new Spearman());
266 }else if (Estimators[i] == "structkulczynski") {
267 matrixCalculators.push_back(new StructKulczynski());
268 }else if (Estimators[i] == "speciesprofile") {
269 matrixCalculators.push_back(new SpeciesProfile());
270 }else if (Estimators[i] == "hamming") {
271 matrixCalculators.push_back(new Hamming());
272 }else if (Estimators[i] == "structchi2") {
273 matrixCalculators.push_back(new StructChi2());
274 }else if (Estimators[i] == "gower") {
275 matrixCalculators.push_back(new Gower());
276 }else if (Estimators[i] == "memchi2") {
277 matrixCalculators.push_back(new MemChi2());
278 }else if (Estimators[i] == "memchord") {
279 matrixCalculators.push_back(new MemChord());
280 }else if (Estimators[i] == "memeuclidean") {
281 matrixCalculators.push_back(new MemEuclidean());
282 }else if (Estimators[i] == "mempearson") {
283 matrixCalculators.push_back(new MemPearson());
284 }else if (Estimators[i] == "jsd") {
285 matrixCalculators.push_back(new JSD());
286 }else if (Estimators[i] == "rjsd") {
287 matrixCalculators.push_back(new RJSD());
296 catch(exception& e) {
297 m->errorOut(e, "MatrixOutputCommand", "MatrixOutputCommand");
302 //**********************************************************************************************************************
304 MatrixOutputCommand::~MatrixOutputCommand(){}
306 //**********************************************************************************************************************
308 int MatrixOutputCommand::execute(){
311 if (abort == true) { if (calledHelp) { return 0; } return 2; }
313 //if the users entered no valid calculators don't execute command
314 if (matrixCalculators.size() == 0) { m->mothurOut("No valid calculators."); m->mothurOutEndLine(); return 0; }
316 input = new InputData(sharedfile, "sharedfile");
317 lookup = input->getSharedRAbundVectors();
318 string lastLabel = lookup[0]->getLabel();
320 //if the users enters label "0.06" and there is no "0.06" in their file use the next lowest label.
321 set<string> processedLabels;
322 set<string> userLabels = labels;
324 if (lookup.size() < 2) { m->mothurOut("You have not provided enough valid groups. I cannot run the command."); m->mothurOutEndLine(); delete input; for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } return 0;}
327 if (subsampleSize == -1) { //user has not set size, set size = smallest samples size
328 subsampleSize = lookup[0]->getNumSeqs();
329 for (int i = 1; i < lookup.size(); i++) {
330 int thisSize = lookup[i]->getNumSeqs();
332 if (thisSize < subsampleSize) { subsampleSize = thisSize; }
337 vector<SharedRAbundVector*> temp;
338 for (int i = 0; i < lookup.size(); i++) {
339 if (lookup[i]->getNumSeqs() < subsampleSize) {
340 m->mothurOut(lookup[i]->getGroup() + " contains " + toString(lookup[i]->getNumSeqs()) + ". Eliminating."); m->mothurOutEndLine();
343 Groups.push_back(lookup[i]->getGroup());
344 temp.push_back(lookup[i]);
348 m->setGroups(Groups);
351 if (lookup.size() < 2) { m->mothurOut("You have not provided enough valid groups. I cannot run the command."); m->mothurOutEndLine(); m->control_pressed = true; delete input; return 0; }
354 numGroups = lookup.size();
355 lines.resize(processors);
356 for (int i = 0; i < processors; i++) {
357 lines[i].start = int (sqrt(float(i)/float(processors)) * numGroups);
358 lines[i].end = int (sqrt(float(i+1)/float(processors)) * numGroups);
361 if (m->control_pressed) { delete input; for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } m->clearGroups(); return 0; }
363 //as long as you are not at the end of the file or done wih the lines you want
364 while((lookup[0] != NULL) && ((allLines == 1) || (userLabels.size() != 0))) {
366 if (m->control_pressed) { outputTypes.clear(); delete input; for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } m->clearGroups(); return 0; }
368 if(allLines == 1 || labels.count(lookup[0]->getLabel()) == 1){
369 m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
372 processedLabels.insert(lookup[0]->getLabel());
373 userLabels.erase(lookup[0]->getLabel());
376 if ((m->anyLabelsToProcess(lookup[0]->getLabel(), userLabels, "") == true) && (processedLabels.count(lastLabel) != 1)) {
377 string saveLabel = lookup[0]->getLabel();
379 for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
380 lookup = input->getSharedRAbundVectors(lastLabel);
382 m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
385 processedLabels.insert(lookup[0]->getLabel());
386 userLabels.erase(lookup[0]->getLabel());
388 //restore real lastlabel to save below
389 lookup[0]->setLabel(saveLabel);
392 lastLabel = lookup[0]->getLabel();
394 //get next line to process
395 for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
396 lookup = input->getSharedRAbundVectors();
399 if (m->control_pressed) { outputTypes.clear(); delete input; for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } m->clearGroups(); return 0; }
401 //output error messages about any remaining user labels
402 set<string>::iterator it;
403 bool needToRun = false;
404 for (it = userLabels.begin(); it != userLabels.end(); it++) {
405 m->mothurOut("Your file does not include the label " + *it);
406 if (processedLabels.count(lastLabel) != 1) {
407 m->mothurOut(". I will use " + lastLabel + "."); m->mothurOutEndLine();
410 m->mothurOut(". Please refer to " + lastLabel + "."); m->mothurOutEndLine();
414 if (m->control_pressed) { outputTypes.clear(); delete input; for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } m->clearGroups(); return 0; }
416 //run last label if you need to
417 if (needToRun == true) {
418 for (int i = 0; i < lookup.size(); i++) { if (lookup[i] != NULL) { delete lookup[i]; } }
419 lookup = input->getSharedRAbundVectors(lastLabel);
421 m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
423 for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
426 if (m->control_pressed) { outputTypes.clear(); delete input; for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } m->clearGroups(); return 0; }
428 //reset groups parameter
431 //set phylip file as new current phylipfile
433 itTypes = outputTypes.find("phylip");
434 if (itTypes != outputTypes.end()) {
435 if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; if (!subsample) { m->setPhylipFile(current); } }
438 m->mothurOutEndLine();
439 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
440 for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
441 m->mothurOutEndLine();
446 catch(exception& e) {
447 m->errorOut(e, "MatrixOutputCommand", "execute");
451 /***********************************************************/
452 void MatrixOutputCommand::printSims(ostream& out, vector< vector<double> >& simMatrix) {
455 out.setf(ios::fixed, ios::floatfield); out.setf(ios::showpoint);
457 if (output == "lt") {
458 out << simMatrix.size() << endl;
459 for (int b = 0; b < simMatrix.size(); b++) {
460 out << lookup[b]->getGroup() << '\t';
461 for (int n = 0; n < b; n++) {
462 out << simMatrix[b][n] << '\t';
466 }else if (output == "column") {
467 for (int b = 0; b < simMatrix.size(); b++) {
468 for (int n = 0; n < b; n++) {
469 out << lookup[b]->getGroup() << '\t' << lookup[n]->getGroup() << '\t' << simMatrix[b][n] << endl;
473 out << simMatrix.size() << endl;
474 for (int b = 0; b < simMatrix.size(); b++) {
475 out << lookup[b]->getGroup() << '\t';
476 for (int n = 0; n < simMatrix[b].size(); n++) {
477 out << simMatrix[b][n] << '\t';
483 catch(exception& e) {
484 m->errorOut(e, "MatrixOutputCommand", "printSims");
488 /***********************************************************/
489 int MatrixOutputCommand::process(vector<SharedRAbundVector*> thisLookup){
491 vector< vector< vector<seqDist> > > calcDistsTotals; //each iter, one for each calc, then each groupCombos dists. this will be used to make .dist files
492 vector< vector<seqDist> > calcDists; calcDists.resize(matrixCalculators.size());
494 for (int thisIter = 0; thisIter < iters+1; thisIter++) {
495 map<string, string> variables;
496 variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(sharedfile));
497 variables["[distance]"] = thisLookup[0]->getLabel();
498 variables["[tag2]"] = "";
500 vector<SharedRAbundVector*> thisItersLookup = thisLookup;
502 if (subsample && (thisIter != 0)) {
504 vector<string> tempLabels; //dont need since we arent printing the sampled sharedRabunds
506 //make copy of lookup so we don't get access violations
507 vector<SharedRAbundVector*> newLookup;
508 for (int k = 0; k < thisItersLookup.size(); k++) {
509 SharedRAbundVector* temp = new SharedRAbundVector();
510 temp->setLabel(thisItersLookup[k]->getLabel());
511 temp->setGroup(thisItersLookup[k]->getGroup());
512 newLookup.push_back(temp);
516 for (int k = 0; k < thisItersLookup[0]->getNumBins(); k++) {
517 if (m->control_pressed) { for (int j = 0; j < newLookup.size(); j++) { delete newLookup[j]; } return 0; }
518 for (int j = 0; j < thisItersLookup.size(); j++) { newLookup[j]->push_back(thisItersLookup[j]->getAbundance(k), thisItersLookup[j]->getGroup()); }
521 tempLabels = sample.getSample(newLookup, subsampleSize);
522 thisItersLookup = newLookup;
526 driver(thisItersLookup, 0, numGroups, calcDists);
529 vector<int> processIDS;
531 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
532 //loop through and create all the processes you want
533 while (process != processors) {
537 processIDS.push_back(pid);
541 driver(thisItersLookup, lines[process].start, lines[process].end, calcDists);
543 string tempdistFileName = m->getRootName(m->getSimpleName(sharedfile)) + m->mothurGetpid(process) + ".dist";
545 m->openOutputFile(tempdistFileName, outtemp);
547 for (int i = 0; i < calcDists.size(); i++) {
548 outtemp << calcDists[i].size() << endl;
550 for (int j = 0; j < calcDists[i].size(); j++) {
551 outtemp << calcDists[i][j].seq1 << '\t' << calcDists[i][j].seq2 << '\t' << calcDists[i][j].dist << endl;
558 m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine();
559 for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); }
564 //parent do your part
565 driver(thisItersLookup, lines[0].start, lines[0].end, calcDists);
567 //force parent to wait until all the processes are done
568 for (int i = 0; i < processIDS.size(); i++) {
569 int temp = processIDS[i];
573 for (int i = 0; i < processIDS.size(); i++) {
574 string tempdistFileName = m->getRootName(m->getSimpleName(sharedfile)) + toString(processIDS[i]) + ".dist";
576 m->openInputFile(tempdistFileName, intemp);
578 for (int k = 0; k < calcDists.size(); k++) {
580 intemp >> size; m->gobble(intemp);
582 for (int j = 0; j < size; j++) {
587 intemp >> seq1 >> seq2 >> dist; m->gobble(intemp);
589 seqDist tempDist(seq1, seq2, dist);
590 calcDists[k].push_back(tempDist);
594 m->mothurRemove(tempdistFileName);
597 //////////////////////////////////////////////////////////////////////////////////////////////////////
598 //Windows version shared memory, so be careful when passing variables through the distSharedData struct.
599 //Above fork() will clone, so memory is separate, but that's not the case with windows,
600 //Taking advantage of shared memory to pass results vectors.
601 //////////////////////////////////////////////////////////////////////////////////////////////////////
603 vector<distSharedData*> pDataArray;
604 DWORD dwThreadIdArray[processors-1];
605 HANDLE hThreadArray[processors-1];
607 //Create processor worker threads.
608 for( int i=1; i<processors; i++ ){
610 //make copy of lookup so we don't get access violations
611 vector<SharedRAbundVector*> newLookup;
612 for (int k = 0; k < thisItersLookup.size(); k++) {
613 SharedRAbundVector* temp = new SharedRAbundVector();
614 temp->setLabel(thisItersLookup[k]->getLabel());
615 temp->setGroup(thisItersLookup[k]->getGroup());
616 newLookup.push_back(temp);
620 for (int k = 0; k < thisItersLookup[0]->getNumBins(); k++) {
621 if (m->control_pressed) { for (int j = 0; j < newLookup.size(); j++) { delete newLookup[j]; } return 0; }
622 for (int j = 0; j < thisItersLookup.size(); j++) { newLookup[j]->push_back(thisItersLookup[j]->getAbundance(k), thisItersLookup[j]->getGroup()); }
625 // Allocate memory for thread data.
626 distSharedData* tempSum = new distSharedData(m, lines[i].start, lines[i].end, Estimators, newLookup);
627 pDataArray.push_back(tempSum);
628 processIDS.push_back(i);
630 hThreadArray[i-1] = CreateThread(NULL, 0, MyDistSharedThreadFunction, pDataArray[i-1], 0, &dwThreadIdArray[i-1]);
633 //parent do your part
634 driver(thisItersLookup, lines[0].start, lines[0].end, calcDists);
636 //Wait until all threads have terminated.
637 WaitForMultipleObjects(processors-1, hThreadArray, TRUE, INFINITE);
639 //Close all thread handles and free memory allocations.
640 for(int i=0; i < pDataArray.size(); i++){
641 if (pDataArray[i]->count != (pDataArray[i]->end-pDataArray[i]->start)) {
642 m->mothurOut("[ERROR]: process " + toString(i) + " only processed " + toString(pDataArray[i]->count) + " of " + toString(pDataArray[i]->end-pDataArray[i]->start) + " groups assigned to it, quitting. \n"); m->control_pressed = true;
644 for (int j = 0; j < pDataArray[i]->thisLookup.size(); j++) { delete pDataArray[i]->thisLookup[j]; }
646 for (int k = 0; k < calcDists.size(); k++) {
647 int size = pDataArray[i]->calcDists[k].size();
648 for (int j = 0; j < size; j++) { calcDists[k].push_back(pDataArray[i]->calcDists[k][j]); }
651 CloseHandle(hThreadArray[i]);
652 delete pDataArray[i];
658 if (subsample && (thisIter != 0)) {
659 if((thisIter) % 100 == 0){ m->mothurOutJustToScreen(toString(thisIter)+"\n"); }
660 calcDistsTotals.push_back(calcDists);
661 for (int i = 0; i < calcDists.size(); i++) {
662 for (int j = 0; j < calcDists[i].size(); j++) {
663 if (m->debug) { m->mothurOut("[DEBUG]: Results: iter = " + toString(thisIter) + ", " + thisLookup[calcDists[i][j].seq1]->getGroup() + " - " + thisLookup[calcDists[i][j].seq2]->getGroup() + " distance = " + toString(calcDists[i][j].dist) + ".\n"); }
667 for (int i = 0; i < thisItersLookup.size(); i++) { delete thisItersLookup[i]; }
668 thisItersLookup.clear();
669 }else { //print results for whole dataset
670 for (int i = 0; i < calcDists.size(); i++) {
671 if (m->control_pressed) { break; }
674 vector< vector<double> > matrix; //square matrix to represent the distance
675 matrix.resize(thisLookup.size());
676 for (int k = 0; k < thisLookup.size(); k++) { matrix[k].resize(thisLookup.size(), 0.0); }
678 for (int j = 0; j < calcDists[i].size(); j++) {
679 int row = calcDists[i][j].seq1;
680 int column = calcDists[i][j].seq2;
681 double dist = calcDists[i][j].dist;
683 matrix[row][column] = dist;
684 matrix[column][row] = dist;
687 variables["[outputtag]"] = output;
688 variables["[calc]"] = matrixCalculators[i]->getName();
689 string distFileName = getOutputFileName("phylip",variables);
690 outputNames.push_back(distFileName); outputTypes["phylip"].push_back(distFileName);
693 m->openOutputFile(distFileName, outDist);
694 outDist.setf(ios::fixed, ios::floatfield); outDist.setf(ios::showpoint);
696 printSims(outDist, matrix);
701 for (int i = 0; i < calcDists.size(); i++) { calcDists[i].clear(); }
705 //we need to find the average distance and standard deviation for each groups distance
706 vector< vector<seqDist> > calcAverages = m->getAverages(calcDistsTotals, mode);
708 //find standard deviation
709 vector< vector<seqDist> > stdDev = m->getStandardDeviation(calcDistsTotals, calcAverages);
712 for (int i = 0; i < calcDists.size(); i++) {
713 vector< vector<double> > matrix; //square matrix to represent the distance
714 matrix.resize(thisLookup.size());
715 for (int k = 0; k < thisLookup.size(); k++) { matrix[k].resize(thisLookup.size(), 0.0); }
717 vector< vector<double> > stdmatrix; //square matrix to represent the stdDev
718 stdmatrix.resize(thisLookup.size());
719 for (int k = 0; k < thisLookup.size(); k++) { stdmatrix[k].resize(thisLookup.size(), 0.0); }
722 for (int j = 0; j < calcAverages[i].size(); j++) {
723 int row = calcAverages[i][j].seq1;
724 int column = calcAverages[i][j].seq2;
725 float dist = calcAverages[i][j].dist;
726 float stdDist = stdDev[i][j].dist;
728 matrix[row][column] = dist;
729 matrix[column][row] = dist;
730 stdmatrix[row][column] = stdDist;
731 stdmatrix[column][row] = stdDist;
734 map<string, string> variables;
735 variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(sharedfile));
736 variables["[distance]"] = thisLookup[0]->getLabel();
737 variables["[outputtag]"] = output;
738 variables["[tag2]"] = "ave";
739 variables["[calc]"] = matrixCalculators[i]->getName();
740 string distFileName = getOutputFileName("phylip",variables);
741 outputNames.push_back(distFileName); outputTypes["phylip"].push_back(distFileName);
742 //set current phylip file to average distance matrix
743 m->setPhylipFile(distFileName);
745 m->openOutputFile(distFileName, outAve);
746 outAve.setf(ios::fixed, ios::floatfield); outAve.setf(ios::showpoint);
748 printSims(outAve, matrix);
752 variables["[tag2]"] = "std";
753 distFileName = getOutputFileName("phylip",variables);
754 outputNames.push_back(distFileName); outputTypes["phylip"].push_back(distFileName);
756 m->openOutputFile(distFileName, outSTD);
757 outSTD.setf(ios::fixed, ios::floatfield); outSTD.setf(ios::showpoint);
759 printSims(outSTD, stdmatrix);
768 catch(exception& e) {
769 m->errorOut(e, "MatrixOutputCommand", "process");
773 /**************************************************************************************************/
774 int MatrixOutputCommand::driver(vector<SharedRAbundVector*> thisLookup, int start, int end, vector< vector<seqDist> >& calcDists) {
776 vector<SharedRAbundVector*> subset;
778 for (int k = start; k < end; k++) { // pass cdd each set of groups to compare
780 for (int l = 0; l < k; l++) {
782 if (k != l) { //we dont need to similiarity of a groups to itself
783 subset.clear(); //clear out old pair of sharedrabunds
784 //add new pair of sharedrabunds
785 subset.push_back(thisLookup[k]); subset.push_back(thisLookup[l]);
787 for(int i=0;i<matrixCalculators.size();i++) {
789 //if this calc needs all groups to calculate the pair load all groups
790 if (matrixCalculators[i]->getNeedsAll()) {
791 //load subset with rest of lookup for those calcs that need everyone to calc for a pair
792 for (int w = 0; w < thisLookup.size(); w++) {
793 if ((w != k) && (w != l)) { subset.push_back(thisLookup[w]); }
797 vector<double> tempdata = matrixCalculators[i]->getValues(subset); //saves the calculator outputs
799 if (m->control_pressed) { return 1; }
801 seqDist temp(l, k, tempdata[0]);
802 calcDists[i].push_back(temp);
810 catch(exception& e) {
811 m->errorOut(e, "MatrixOutputCommand", "driver");
815 /***********************************************************/