2 * treegroupscommand.cpp
5 * Created by Sarah Westcott on 4/8/09.
6 * Copyright 2009 Schloss Lab UMASS Amherst. All rights reserved.
10 #include "treegroupscommand.h"
11 #include "subsample.h"
12 #include "consensus.h"
14 //**********************************************************************************************************************
15 vector<string> TreeGroupCommand::setParameters(){
17 CommandParameter pshared("shared", "InputTypes", "", "", "PhylipColumnShared", "PhylipColumnShared", "none","tree",false,false,true); parameters.push_back(pshared);
18 CommandParameter pphylip("phylip", "InputTypes", "", "", "PhylipColumnShared", "PhylipColumnShared", "none","tree",false,false); parameters.push_back(pphylip);
19 CommandParameter pname("name", "InputTypes", "", "", "NameCount", "none", "ColumnName","",false,false); parameters.push_back(pname);
20 CommandParameter pcount("count", "InputTypes", "", "", "NameCount", "none", "countcolumn","",false,false); parameters.push_back(pcount);
21 CommandParameter pcolumn("column", "InputTypes", "", "", "PhylipColumnShared", "PhylipColumnShared", "ColumnName-countcolumn","tree",false,false); parameters.push_back(pcolumn);
22 CommandParameter piters("iters", "Number", "", "1000", "", "", "","",false,false); parameters.push_back(piters);
23 CommandParameter psubsample("subsample", "String", "", "", "", "", "","",false,false); parameters.push_back(psubsample);
24 CommandParameter pcutoff("cutoff", "Number", "", "10", "", "", "","",false,false); parameters.push_back(pcutoff);
25 CommandParameter pprecision("precision", "Number", "", "100", "", "", "","",false,false); parameters.push_back(pprecision);
26 CommandParameter plabel("label", "String", "", "", "", "", "","",false,false); parameters.push_back(plabel);
27 CommandParameter pgroups("groups", "String", "", "", "", "", "","",false,false); parameters.push_back(pgroups);
28 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);
30 CommandParameter pprocessors("processors", "Number", "", "1", "", "", "","",false,false,true); parameters.push_back(pprocessors);
31 //CommandParameter poutput("output", "Multiple", "lt-square", "lt", "", "", "",false,false); parameters.push_back(poutput);
32 CommandParameter pinputdir("inputdir", "String", "", "", "", "", "","",false,false); parameters.push_back(pinputdir);
33 CommandParameter poutputdir("outputdir", "String", "", "", "", "", "","",false,false); parameters.push_back(poutputdir);
35 vector<string> myArray;
36 for (int i = 0; i < parameters.size(); i++) { myArray.push_back(parameters[i].name); }
40 m->errorOut(e, "TreeGroupCommand", "setParameters");
44 //**********************************************************************************************************************
45 string TreeGroupCommand::getHelpString(){
47 string helpString = "";
48 ValidCalculators validCalculator;
49 helpString += "The tree.shared command creates a .tre to represent the similiarity between groups or sequences.\n";
50 helpString += "The tree.shared command parameters are shared, groups, calc, phylip, column, name, cutoff, precision, processors, subsample, iters and label.\n";
51 helpString += "The groups parameter allows you to specify which of the groups in your groupfile you would like included used.\n";
52 helpString += "The group names are separated by dashes. The label allow you to select what distance levels you would like trees created for, and are also separated by dashes.\n";
53 helpString += "The phylip or column parameter are required if you do not provide a sharedfile, and only one may be used. If you use a column file the name filename is required. \n";
54 helpString += "If you do not provide a cutoff value 10.00 is assumed. If you do not provide a precision value then 100 is assumed.\n";
55 helpString += "The tree.shared command should be in the following format: tree.shared(groups=yourGroups, calc=yourCalcs, label=yourLabels).\n";
56 helpString += "The iters parameter allows you to choose the number of times you would like to run the subsample.\n";
57 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. The subsample parameter may only be used with a shared file.\n";
58 helpString += "Example tree.shared(groups=A-B-C, calc=jabund-sorabund).\n";
59 helpString += "The default value for groups is all the groups in your groupfile.\n";
60 helpString += "The default value for calc is jclass-thetayc.\n";
61 helpString += "The tree.shared command outputs a .tre file for each calculator you specify at each distance you choose.\n";
62 helpString += validCalculator.printCalc("treegroup");
63 helpString += "Or the tree.shared command can be in the following format: tree.shared(phylip=yourPhylipFile).\n";
64 helpString += "Example tree.shared(phylip=abrecovery.dist).\n";
65 helpString += "Note: No spaces between parameter labels (i.e. groups), '=' and parameters (i.e.yourGroups).\n";
69 m->errorOut(e, "TreeGroupCommand", "getHelpString");
73 //**********************************************************************************************************************
74 string TreeGroupCommand::getOutputPattern(string type) {
78 if (type == "tree") { pattern = "[filename],[calc],[distance],[tag],tre-[filename],tre"; }
79 else { m->mothurOut("[ERROR]: No definition for type " + type + " output pattern.\n"); m->control_pressed = true; }
84 m->errorOut(e, "TreeGroupCommand", "getOutputPattern");
88 //**********************************************************************************************************************
89 TreeGroupCommand::TreeGroupCommand(){
91 abort = true; calledHelp = true;
93 //initialize outputTypes
94 vector<string> tempOutNames;
95 outputTypes["tree"] = tempOutNames;
98 m->errorOut(e, "TreeGroupCommand", "TreeGroupCommand");
102 //**********************************************************************************************************************
104 TreeGroupCommand::TreeGroupCommand(string option) {
106 abort = false; calledHelp = false;
109 //allow user to run help
110 if(option == "help") { help(); abort = true; calledHelp = true; }
111 else if(option == "citation") { citation(); abort = true; calledHelp = true;}
114 vector<string> myArray = setParameters();
116 OptionParser parser(option);
117 map<string, string> parameters = parser. getParameters();
119 ValidParameters validParameter;
120 map<string, string>::iterator it;
122 //check to make sure all parameters are valid for command
123 for (it = parameters.begin(); it != parameters.end(); it++) {
124 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
127 //initialize outputTypes
128 vector<string> tempOutNames;
129 outputTypes["tree"] = tempOutNames;
131 //if the user changes the input directory command factory will send this info to us in the output parameter
132 string inputDir = validParameter.validFile(parameters, "inputdir", false);
133 if (inputDir == "not found"){ inputDir = ""; }
136 it = parameters.find("phylip");
137 //user has given a template file
138 if(it != parameters.end()){
139 path = m->hasPath(it->second);
140 //if the user has not given a path then, add inputdir. else leave path alone.
141 if (path == "") { parameters["phylip"] = inputDir + it->second; }
144 it = parameters.find("column");
145 //user has given a template file
146 if(it != parameters.end()){
147 path = m->hasPath(it->second);
148 //if the user has not given a path then, add inputdir. else leave path alone.
149 if (path == "") { parameters["column"] = inputDir + it->second; }
152 it = parameters.find("name");
153 //user has given a template file
154 if(it != parameters.end()){
155 path = m->hasPath(it->second);
156 //if the user has not given a path then, add inputdir. else leave path alone.
157 if (path == "") { parameters["name"] = inputDir + it->second; }
160 it = parameters.find("count");
161 //user has given a template file
162 if(it != parameters.end()){
163 path = m->hasPath(it->second);
164 //if the user has not given a path then, add inputdir. else leave path alone.
165 if (path == "") { parameters["count"] = inputDir + it->second; }
169 //check for required parameters
170 phylipfile = validParameter.validFile(parameters, "phylip", true);
171 if (phylipfile == "not open") { phylipfile = ""; abort = true; }
172 else if (phylipfile == "not found") { phylipfile = ""; }
173 else { inputfile = phylipfile; format = "phylip"; m->setPhylipFile(phylipfile); }
175 columnfile = validParameter.validFile(parameters, "column", true);
176 if (columnfile == "not open") { columnfile = ""; abort = true; }
177 else if (columnfile == "not found") { columnfile = ""; }
178 else { inputfile = columnfile; format = "column"; m->setColumnFile(columnfile); }
180 sharedfile = validParameter.validFile(parameters, "shared", true);
181 if (sharedfile == "not open") { sharedfile = ""; abort = true; }
182 else if (sharedfile == "not found") { sharedfile = ""; }
183 else { inputfile = sharedfile; format = "sharedfile"; m->setSharedFile(sharedfile); }
185 namefile = validParameter.validFile(parameters, "name", true);
186 if (namefile == "not open") { abort = true; }
187 else if (namefile == "not found") { namefile = ""; }
188 else { m->setNameFile(namefile); }
190 countfile = validParameter.validFile(parameters, "count", true);
191 if (countfile == "not open") { abort = true; countfile = ""; }
192 else if (countfile == "not found") { countfile = ""; }
193 else { m->setCountTableFile(countfile); }
195 if ((phylipfile == "") && (columnfile == "") && (sharedfile == "")) {
196 //is there are current file available for either of these?
197 //give priority to shared, then column, then phylip
198 sharedfile = m->getSharedFile();
199 if (sharedfile != "") { inputfile = sharedfile; format = "sharedfile"; m->mothurOut("Using " + sharedfile + " as input file for the shared parameter."); m->mothurOutEndLine(); }
201 columnfile = m->getColumnFile();
202 if (columnfile != "") { inputfile = columnfile; format = "column"; m->mothurOut("Using " + columnfile + " as input file for the column parameter."); m->mothurOutEndLine(); }
204 phylipfile = m->getPhylipFile();
205 if (phylipfile != "") { inputfile = phylipfile; format = "phylip"; m->mothurOut("Using " + phylipfile + " as input file for the phylip parameter."); m->mothurOutEndLine(); }
207 m->mothurOut("No valid current files. You must provide a shared, phylip or column file."); m->mothurOutEndLine();
213 else if ((phylipfile != "") && (columnfile != "")) { m->mothurOut("When running the tree.shared command with a distance file you may not use both the column and the phylip parameters."); m->mothurOutEndLine(); abort = true; }
215 if (columnfile != "") {
216 if ((namefile == "") && (countfile == "")){
217 namefile = m->getNameFile();
218 if (namefile != "") { m->mothurOut("Using " + namefile + " as input file for the name parameter."); m->mothurOutEndLine(); }
220 countfile = m->getCountTableFile();
221 if (countfile != "") { m->mothurOut("Using " + countfile + " as input file for the count parameter."); m->mothurOutEndLine(); }
223 m->mothurOut("You need to provide a namefile or countfile if you are going to use the column format."); m->mothurOutEndLine();
231 //check for optional parameter and set defaults
232 // ...at some point should added some additional type checking...
233 label = validParameter.validFile(parameters, "label", false);
234 if (label == "not found") { label = ""; }
236 if(label != "all") { m->splitAtDash(label, labels); allLines = 0; }
237 else { allLines = 1; }
240 groups = validParameter.validFile(parameters, "groups", false);
241 if (groups == "not found") { groups = ""; }
243 m->splitAtDash(groups, Groups);
244 m->setGroups(Groups);
247 calc = validParameter.validFile(parameters, "calc", false);
248 if (calc == "not found") { calc = "jclass-thetayc"; }
250 if (calc == "default") { calc = "jclass-thetayc"; }
252 m->splitAtDash(calc, Estimators);
253 if (m->inUsersGroups("citation", Estimators)) {
254 ValidCalculators validCalc; validCalc.printCitations(Estimators);
255 //remove citation from list of calcs
256 for (int i = 0; i < Estimators.size(); i++) { if (Estimators[i] == "citation") { Estimators.erase(Estimators.begin()+i); break; } }
260 temp = validParameter.validFile(parameters, "precision", false); if (temp == "not found") { temp = "100"; }
261 m->mothurConvert(temp, precision);
263 temp = validParameter.validFile(parameters, "cutoff", false); if (temp == "not found") { temp = "10"; }
264 m->mothurConvert(temp, cutoff);
265 cutoff += (5 / (precision * 10.0));
267 temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); }
268 m->setProcessors(temp);
269 m->mothurConvert(temp, processors);
271 temp = validParameter.validFile(parameters, "iters", false); if (temp == "not found") { temp = "1000"; }
272 m->mothurConvert(temp, iters);
274 temp = validParameter.validFile(parameters, "subsample", false); if (temp == "not found") { temp = "F"; }
275 if (m->isNumeric1(temp)) { m->mothurConvert(temp, subsampleSize); subsample = true; }
277 if (m->isTrue(temp)) { subsample = true; subsampleSize = -1; } //we will set it to smallest group later
278 else { subsample = false; }
281 if (subsample == false) { iters = 1; }
283 if (subsample && (format != "sharedfile")) { m->mothurOut("[ERROR]: the subsample parameter can only be used with a shared file.\n"); abort=true; }
285 //if the user changes the output directory command factory will send this info to us in the output parameter
286 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){
288 outputDir += m->hasPath(inputfile); //if user entered a file with a path then preserve it
293 catch(exception& e) {
294 m->errorOut(e, "TreeGroupCommand", "TreeGroupCommand");
298 //**********************************************************************************************************************
300 TreeGroupCommand::~TreeGroupCommand(){
301 if (abort == false) {
302 if (format == "sharedfile") { delete input; }
303 else { delete list; }
309 //**********************************************************************************************************************
311 int TreeGroupCommand::execute(){
314 if (abort == true) { if (calledHelp) { return 0; } return 2; }
316 if (format == "sharedfile") {
318 ValidCalculators validCalculator;
320 for (int i=0; i<Estimators.size(); i++) {
321 if (validCalculator.isValidCalculator("treegroup", Estimators[i]) == true) {
322 if (Estimators[i] == "sharedsobs") {
323 treeCalculators.push_back(new SharedSobsCS());
324 }else if (Estimators[i] == "sharedchao") {
325 treeCalculators.push_back(new SharedChao1());
326 }else if (Estimators[i] == "sharedace") {
327 treeCalculators.push_back(new SharedAce());
328 }else if (Estimators[i] == "jabund") {
329 treeCalculators.push_back(new JAbund());
330 }else if (Estimators[i] == "sorabund") {
331 treeCalculators.push_back(new SorAbund());
332 }else if (Estimators[i] == "jclass") {
333 treeCalculators.push_back(new Jclass());
334 }else if (Estimators[i] == "sorclass") {
335 treeCalculators.push_back(new SorClass());
336 }else if (Estimators[i] == "jest") {
337 treeCalculators.push_back(new Jest());
338 }else if (Estimators[i] == "sorest") {
339 treeCalculators.push_back(new SorEst());
340 }else if (Estimators[i] == "thetayc") {
341 treeCalculators.push_back(new ThetaYC());
342 }else if (Estimators[i] == "thetan") {
343 treeCalculators.push_back(new ThetaN());
344 }else if (Estimators[i] == "kstest") {
345 treeCalculators.push_back(new KSTest());
346 }else if (Estimators[i] == "sharednseqs") {
347 treeCalculators.push_back(new SharedNSeqs());
348 }else if (Estimators[i] == "ochiai") {
349 treeCalculators.push_back(new Ochiai());
350 }else if (Estimators[i] == "anderberg") {
351 treeCalculators.push_back(new Anderberg());
352 }else if (Estimators[i] == "kulczynski") {
353 treeCalculators.push_back(new Kulczynski());
354 }else if (Estimators[i] == "kulczynskicody") {
355 treeCalculators.push_back(new KulczynskiCody());
356 }else if (Estimators[i] == "lennon") {
357 treeCalculators.push_back(new Lennon());
358 }else if (Estimators[i] == "morisitahorn") {
359 treeCalculators.push_back(new MorHorn());
360 }else if (Estimators[i] == "braycurtis") {
361 treeCalculators.push_back(new BrayCurtis());
362 }else if (Estimators[i] == "whittaker") {
363 treeCalculators.push_back(new Whittaker());
364 }else if (Estimators[i] == "odum") {
365 treeCalculators.push_back(new Odum());
366 }else if (Estimators[i] == "canberra") {
367 treeCalculators.push_back(new Canberra());
368 }else if (Estimators[i] == "structeuclidean") {
369 treeCalculators.push_back(new StructEuclidean());
370 }else if (Estimators[i] == "structchord") {
371 treeCalculators.push_back(new StructChord());
372 }else if (Estimators[i] == "hellinger") {
373 treeCalculators.push_back(new Hellinger());
374 }else if (Estimators[i] == "manhattan") {
375 treeCalculators.push_back(new Manhattan());
376 }else if (Estimators[i] == "structpearson") {
377 treeCalculators.push_back(new StructPearson());
378 }else if (Estimators[i] == "soergel") {
379 treeCalculators.push_back(new Soergel());
380 }else if (Estimators[i] == "spearman") {
381 treeCalculators.push_back(new Spearman());
382 }else if (Estimators[i] == "structkulczynski") {
383 treeCalculators.push_back(new StructKulczynski());
384 }else if (Estimators[i] == "speciesprofile") {
385 treeCalculators.push_back(new SpeciesProfile());
386 }else if (Estimators[i] == "hamming") {
387 treeCalculators.push_back(new Hamming());
388 }else if (Estimators[i] == "structchi2") {
389 treeCalculators.push_back(new StructChi2());
390 }else if (Estimators[i] == "gower") {
391 treeCalculators.push_back(new Gower());
392 }else if (Estimators[i] == "memchi2") {
393 treeCalculators.push_back(new MemChi2());
394 }else if (Estimators[i] == "memchord") {
395 treeCalculators.push_back(new MemChord());
396 }else if (Estimators[i] == "memeuclidean") {
397 treeCalculators.push_back(new MemEuclidean());
398 }else if (Estimators[i] == "mempearson") {
399 treeCalculators.push_back(new MemPearson());
400 }else if (Estimators[i] == "jsd") {
401 treeCalculators.push_back(new JSD());
402 }else if (Estimators[i] == "rjsd") {
403 treeCalculators.push_back(new RJSD());
409 //if the users entered no valid calculators don't execute command
410 if (treeCalculators.size() == 0) { m->mothurOut("You have given no valid calculators."); m->mothurOutEndLine(); return 0; }
412 input = new InputData(sharedfile, "sharedfile");
413 lookup = input->getSharedRAbundVectors();
414 lastLabel = lookup[0]->getLabel();
416 if (lookup.size() < 2) { m->mothurOut("You have not provided enough valid groups. I cannot run the command."); m->mothurOutEndLine(); return 0; }
418 //used in tree constructor
421 //create treemap class from groupmap for tree class to use
422 ct = new CountTable();
424 map<string, string> groupMap;
426 for (int i = 0; i < m->getAllGroups().size(); i++) {
427 nameMap.insert(m->getAllGroups()[i]);
428 gps.insert(m->getAllGroups()[i]);
429 groupMap[m->getAllGroups()[i]] = m->getAllGroups()[i];
431 ct->createTable(nameMap, groupMap, gps);
433 //clear globaldatas old tree names if any
434 m->Treenames.clear();
436 //fills globaldatas tree names
437 //m->Treenames = m->getGroups();
438 for (int k = 0; k < lookup.size(); k++) {
439 m->Treenames.push_back(lookup[k]->getGroup());
442 if (m->control_pressed) { return 0; }
447 if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } return 0; }
450 filename = inputfile;
452 ReadMatrix* readMatrix;
453 if (format == "column") { readMatrix = new ReadColumnMatrix(filename); }
454 else if (format == "phylip") { readMatrix = new ReadPhylipMatrix(filename); }
456 readMatrix->setCutoff(cutoff);
461 nameMap = new NameAssignment(namefile);
463 readMatrix->read(nameMap);
464 }else if (countfile != "") {
465 ct = new CountTable();
466 ct->readTable(countfile, true, false);
467 readMatrix->read(ct);
469 readMatrix->read(nameMap);
472 list = readMatrix->getListVector();
473 SparseDistanceMatrix* dMatrix = readMatrix->getDMatrix();
475 //clear globaldatas old tree names if any
476 m->Treenames.clear();
479 if (ct != NULL) { delete ct; }
480 ct = new CountTable();
482 map<string, string> groupMap;
484 for (int i = 0; i < list->getNumBins(); i++) {
485 string bin = list->get(i);
489 m->Treenames.push_back(bin);
491 ct->createTable(nameMap, groupMap, gps);
493 vector<string> namesGroups = ct->getNamesOfGroups();
494 m->setGroups(namesGroups);
496 //used in tree constructor
499 if (m->control_pressed) { return 0; }
501 vector< vector<double> > matrix = makeSimsDist(dMatrix);
505 if (m->control_pressed) { return 0; }
507 //create a new filename
508 map<string, string> variables;
509 variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(inputfile));
510 string outputFile = getOutputFileName("tree",variables);
511 outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile);
513 Tree* newTree = createTree(matrix);
515 if (newTree != NULL) { writeTree(outputFile, newTree); delete newTree; }
517 if (m->control_pressed) { return 0; }
519 m->mothurOut("Tree complete. "); m->mothurOutEndLine();
523 //reset groups parameter
526 //set tree file as new current treefile
528 itTypes = outputTypes.find("tree");
529 if (itTypes != outputTypes.end()) {
530 if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setTreeFile(current); }
533 m->mothurOutEndLine();
534 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
535 for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
536 m->mothurOutEndLine();
540 catch(exception& e) {
541 m->errorOut(e, "TreeGroupCommand", "execute");
545 //**********************************************************************************************************************
547 Tree* TreeGroupCommand::createTree(vector< vector<double> >& simMatrix){
550 t = new Tree(ct, simMatrix);
552 if (m->control_pressed) { delete t; t = NULL; return t; }
559 catch(exception& e) {
560 m->errorOut(e, "TreeGroupCommand", "createTree");
564 /***********************************************************/
565 int TreeGroupCommand::writeTree(string out, Tree* T) {
569 t->createNewickFile(out);
571 if (m->control_pressed) { m->mothurRemove(out); outputNames.pop_back(); return 1; }
576 catch(exception& e) {
577 m->errorOut(e, "TreeGroupCommand", "printSims");
582 /***********************************************************/
583 void TreeGroupCommand::printSims(ostream& out, vector< vector<double> >& simMatrix) {
586 for (int m = 0; m < simMatrix.size(); m++) {
587 //out << lookup[m]->getGroup() << '\t';
588 for (int n = 0; n < simMatrix.size(); n++) {
589 out << simMatrix[m][n] << '\t';
595 catch(exception& e) {
596 m->errorOut(e, "TreeGroupCommand", "printSims");
600 /***********************************************************/
601 vector< vector<double> > TreeGroupCommand::makeSimsDist(SparseDistanceMatrix* matrix) {
603 numGroups = list->size();
605 //initialize simMatrix
606 vector< vector<double> > simMatrix;
607 simMatrix.resize(numGroups);
608 for (int k = 0; k < simMatrix.size(); k++) {
609 for (int j = 0; j < simMatrix.size(); j++) {
610 simMatrix[k].push_back(0.0);
614 //go through sparse matrix and fill sims
615 //go through each cell in the sparsematrix
616 for (int i = 0; i < matrix->seqVec.size(); i++) {
617 for (int j = 0; j < matrix->seqVec[i].size(); j++) {
619 //already checked everyone else in row
620 if (i < matrix->seqVec[i][j].index) {
621 simMatrix[i][matrix->seqVec[i][j].index] = -(matrix->seqVec[i][j].dist -1.0);
622 simMatrix[matrix->seqVec[i][j].index][i] = -(matrix->seqVec[i][j].dist -1.0);
624 if (m->control_pressed) { return simMatrix; }
631 catch(exception& e) {
632 m->errorOut(e, "TreeGroupCommand", "makeSimsDist");
637 /***********************************************************/
638 int TreeGroupCommand::makeSimsShared() {
642 if (subsampleSize == -1) { //user has not set size, set size = smallest samples size
643 subsampleSize = lookup[0]->getNumSeqs();
644 for (int i = 1; i < lookup.size(); i++) {
645 int thisSize = lookup[i]->getNumSeqs();
647 if (thisSize < subsampleSize) { subsampleSize = thisSize; }
652 m->Treenames.clear();
653 vector<SharedRAbundVector*> temp;
654 for (int i = 0; i < lookup.size(); i++) {
655 if (lookup[i]->getNumSeqs() < subsampleSize) {
656 m->mothurOut(lookup[i]->getGroup() + " contains " + toString(lookup[i]->getNumSeqs()) + ". Eliminating."); m->mothurOutEndLine();
659 Groups.push_back(lookup[i]->getGroup());
660 temp.push_back(lookup[i]);
661 m->Treenames.push_back(lookup[i]->getGroup());
665 m->setGroups(Groups);
668 if (lookup.size() < 2) { m->mothurOut("You have not provided enough valid groups. I cannot run the command."); m->mothurOutEndLine(); m->control_pressed = true; return 0; }
670 numGroups = lookup.size();
672 //sanity check to make sure processors < numComparisions
674 for(int i=0;i<numGroups;i++){
675 for(int j=0;j<i;j++){
677 if (numDists > processors) { break; }
680 if (numDists < processors) { processors = numDists; }
682 lines.resize(processors);
683 for (int i = 0; i < processors; i++) {
684 lines[i].start = int (sqrt(float(i)/float(processors)) * numGroups);
685 lines[i].end = int (sqrt(float(i+1)/float(processors)) * numGroups);
688 set<string> processedLabels;
689 set<string> userLabels = labels;
691 //as long as you are not at the end of the file or done wih the lines you want
692 while((lookup[0] != NULL) && ((allLines == 1) || (userLabels.size() != 0))) {
693 if (m->control_pressed) { for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } for(int i = 0 ; i < treeCalculators.size(); i++) { delete treeCalculators[i]; } return 1; }
695 if(allLines == 1 || labels.count(lookup[0]->getLabel()) == 1){
696 m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
699 processedLabels.insert(lookup[0]->getLabel());
700 userLabels.erase(lookup[0]->getLabel());
703 if ((m->anyLabelsToProcess(lookup[0]->getLabel(), userLabels, "") == true) && (processedLabels.count(lastLabel) != 1)) {
704 string saveLabel = lookup[0]->getLabel();
706 for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
707 lookup = input->getSharedRAbundVectors(lastLabel);
709 m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
712 processedLabels.insert(lookup[0]->getLabel());
713 userLabels.erase(lookup[0]->getLabel());
715 //restore real lastlabel to save below
716 lookup[0]->setLabel(saveLabel);
719 lastLabel = lookup[0]->getLabel();
721 //get next line to process
722 for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
723 lookup = input->getSharedRAbundVectors();
726 if (m->control_pressed) { for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; } for(int i = 0 ; i < treeCalculators.size(); i++) { delete treeCalculators[i]; } return 1; }
728 //output error messages about any remaining user labels
729 set<string>::iterator it;
730 bool needToRun = false;
731 for (it = userLabels.begin(); it != userLabels.end(); it++) {
732 m->mothurOut("Your file does not include the label " + *it);
733 if (processedLabels.count(lastLabel) != 1) {
734 m->mothurOut(". I will use " + lastLabel + "."); m->mothurOutEndLine();
737 m->mothurOut(". Please refer to " + lastLabel + "."); m->mothurOutEndLine();
741 //run last label if you need to
742 if (needToRun == true) {
743 for (int i = 0; i < lookup.size(); i++) { if (lookup[i] != NULL) { delete lookup[i]; } }
744 lookup = input->getSharedRAbundVectors(lastLabel);
746 m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
748 for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
751 for(int i = 0 ; i < treeCalculators.size(); i++) { delete treeCalculators[i]; }
755 catch(exception& e) {
756 m->errorOut(e, "TreeGroupCommand", "makeSimsShared");
761 /***********************************************************/
762 int TreeGroupCommand::process(vector<SharedRAbundVector*> thisLookup) {
764 vector< vector< vector<seqDist> > > calcDistsTotals; //each iter, one for each calc, then each groupCombos dists. this will be used to make .dist files
765 vector< vector<seqDist> > calcDists; calcDists.resize(treeCalculators.size());
767 for (int thisIter = 0; thisIter < iters; thisIter++) {
769 vector<SharedRAbundVector*> thisItersLookup = thisLookup;
773 vector<string> tempLabels; //dont need since we arent printing the sampled sharedRabunds
775 //make copy of lookup so we don't get access violations
776 vector<SharedRAbundVector*> newLookup;
777 for (int k = 0; k < thisItersLookup.size(); k++) {
778 SharedRAbundVector* temp = new SharedRAbundVector();
779 temp->setLabel(thisItersLookup[k]->getLabel());
780 temp->setGroup(thisItersLookup[k]->getGroup());
781 newLookup.push_back(temp);
785 for (int k = 0; k < thisItersLookup[0]->getNumBins(); k++) {
786 if (m->control_pressed) { for (int j = 0; j < newLookup.size(); j++) { delete newLookup[j]; } return 0; }
787 for (int j = 0; j < thisItersLookup.size(); j++) { newLookup[j]->push_back(thisItersLookup[j]->getAbundance(k), thisItersLookup[j]->getGroup()); }
790 tempLabels = sample.getSample(newLookup, subsampleSize);
791 thisItersLookup = newLookup;
795 driver(thisItersLookup, 0, numGroups, calcDists);
798 vector<int> processIDS;
800 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
801 //loop through and create all the processes you want
802 while (process != processors) {
806 processIDS.push_back(pid);
810 driver(thisItersLookup, lines[process].start, lines[process].end, calcDists);
812 string tempdistFileName = m->getRootName(m->getSimpleName(sharedfile)) + m->mothurGetpid(process) + ".dist";
814 m->openOutputFile(tempdistFileName, outtemp);
816 for (int i = 0; i < calcDists.size(); i++) {
817 outtemp << calcDists[i].size() << endl;
819 for (int j = 0; j < calcDists[i].size(); j++) {
820 outtemp << calcDists[i][j].seq1 << '\t' << calcDists[i][j].seq2 << '\t' << calcDists[i][j].dist << endl;
827 m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine();
828 for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); }
833 //parent do your part
834 driver(thisItersLookup, lines[0].start, lines[0].end, calcDists);
836 //force parent to wait until all the processes are done
837 for (int i = 0; i < processIDS.size(); i++) {
838 int temp = processIDS[i];
842 for (int i = 0; i < processIDS.size(); i++) {
843 string tempdistFileName = m->getRootName(m->getSimpleName(sharedfile)) + toString(processIDS[i]) + ".dist";
845 m->openInputFile(tempdistFileName, intemp);
847 for (int k = 0; k < calcDists.size(); k++) {
849 intemp >> size; m->gobble(intemp);
851 for (int j = 0; j < size; j++) {
856 intemp >> seq1 >> seq2 >> dist; m->gobble(intemp);
858 seqDist tempDist(seq1, seq2, dist);
859 calcDists[k].push_back(tempDist);
863 m->mothurRemove(tempdistFileName);
866 //////////////////////////////////////////////////////////////////////////////////////////////////////
867 //Windows version shared memory, so be careful when passing variables through the treeSharedData struct.
868 //Above fork() will clone, so memory is separate, but that's not the case with windows,
869 //Taking advantage of shared memory to pass results vectors.
870 //////////////////////////////////////////////////////////////////////////////////////////////////////
872 vector<treeSharedData*> pDataArray;
873 DWORD dwThreadIdArray[processors-1];
874 HANDLE hThreadArray[processors-1];
876 //Create processor worker threads.
877 for( int i=1; i<processors; i++ ){
879 //make copy of lookup so we don't get access violations
880 vector<SharedRAbundVector*> newLookup;
881 for (int k = 0; k < thisItersLookup.size(); k++) {
882 SharedRAbundVector* temp = new SharedRAbundVector();
883 temp->setLabel(thisItersLookup[k]->getLabel());
884 temp->setGroup(thisItersLookup[k]->getGroup());
885 newLookup.push_back(temp);
889 for (int k = 0; k < thisItersLookup[0]->getNumBins(); k++) {
890 if (m->control_pressed) { for (int j = 0; j < newLookup.size(); j++) { delete newLookup[j]; } return 0; }
891 for (int j = 0; j < thisItersLookup.size(); j++) { newLookup[j]->push_back(thisItersLookup[j]->getAbundance(k), thisItersLookup[j]->getGroup()); }
894 // Allocate memory for thread data.
895 treeSharedData* tempSum = new treeSharedData(m, lines[i].start, lines[i].end, Estimators, newLookup);
896 pDataArray.push_back(tempSum);
897 processIDS.push_back(i);
899 hThreadArray[i-1] = CreateThread(NULL, 0, MyTreeSharedThreadFunction, pDataArray[i-1], 0, &dwThreadIdArray[i-1]);
902 //parent do your part
903 driver(thisItersLookup, lines[0].start, lines[0].end, calcDists);
905 //Wait until all threads have terminated.
906 WaitForMultipleObjects(processors-1, hThreadArray, TRUE, INFINITE);
908 //Close all thread handles and free memory allocations.
909 for(int i=0; i < pDataArray.size(); i++){
910 if (pDataArray[i]->count != (pDataArray[i]->end-pDataArray[i]->start)) {
911 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;
913 for (int j = 0; j < pDataArray[i]->thisLookup.size(); j++) { delete pDataArray[i]->thisLookup[j]; }
915 for (int k = 0; k < calcDists.size(); k++) {
916 int size = pDataArray[i]->calcDists[k].size();
917 for (int j = 0; j < size; j++) { calcDists[k].push_back(pDataArray[i]->calcDists[k][j]); }
920 CloseHandle(hThreadArray[i]);
921 delete pDataArray[i];
927 calcDistsTotals.push_back(calcDists);
932 for (int i = 0; i < thisItersLookup.size(); i++) { delete thisItersLookup[i]; }
933 thisItersLookup.clear();
934 for (int i = 0; i < calcDists.size(); i++) { calcDists[i].clear(); }
937 if (m->debug) { m->mothurOut("[DEBUG]: iter = " + toString(thisIter) + ".\n"); }
940 if (m->debug) { m->mothurOut("[DEBUG]: done with iters.\n"); }
943 //we need to find the average distance and standard deviation for each groups distance
944 vector< vector<seqDist> > calcAverages = m->getAverages(calcDistsTotals);
946 if (m->debug) { m->mothurOut("[DEBUG]: found averages.\n"); }
948 //create average tree for each calc
949 for (int i = 0; i < calcDists.size(); i++) {
950 vector< vector<double> > matrix; //square matrix to represent the distance
951 matrix.resize(thisLookup.size());
952 for (int k = 0; k < thisLookup.size(); k++) { matrix[k].resize(thisLookup.size(), 0.0); }
954 for (int j = 0; j < calcAverages[i].size(); j++) {
955 int row = calcAverages[i][j].seq1;
956 int column = calcAverages[i][j].seq2;
957 float dist = calcAverages[i][j].dist;
959 matrix[row][column] = dist;
960 matrix[column][row] = dist;
963 //create a new filename
964 map<string, string> variables;
965 variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(inputfile));
966 variables["[calc]"] = treeCalculators[i]->getName();
967 variables["[distance]"] = thisLookup[0]->getLabel();
968 variables["[tag]"] = "ave";
969 string outputFile = getOutputFileName("tree",variables);
970 outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile);
972 //creates tree from similarity matrix and write out file
973 Tree* newTree = createTree(matrix);
974 if (newTree != NULL) { writeTree(outputFile, newTree); }
977 if (m->debug) { m->mothurOut("[DEBUG]: done averages trees.\n"); }
979 //create all trees for each calc and find their consensus tree
980 for (int i = 0; i < calcDists.size(); i++) {
981 if (m->control_pressed) { break; }
983 //create a new filename
984 //create a new filename
985 map<string, string> variables;
986 variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(inputfile));
987 variables["[calc]"] = treeCalculators[i]->getName();
988 variables["[distance]"] = thisLookup[0]->getLabel();
989 variables["[tag]"] = "all";
990 string outputFile = getOutputFileName("tree",variables);
991 outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile);
994 m->openOutputFile(outputFile, outAll);
997 for (int myIter = 0; myIter < iters; myIter++) {
999 if(m->control_pressed) { break; }
1002 vector< vector<double> > matrix; //square matrix to represent the distance
1003 matrix.resize(thisLookup.size());
1004 for (int k = 0; k < thisLookup.size(); k++) { matrix[k].resize(thisLookup.size(), 0.0); }
1006 for (int j = 0; j < calcDistsTotals[myIter][i].size(); j++) {
1007 int row = calcDistsTotals[myIter][i][j].seq1;
1008 int column = calcDistsTotals[myIter][i][j].seq2;
1009 double dist = calcDistsTotals[myIter][i][j].dist;
1011 matrix[row][column] = dist;
1012 matrix[column][row] = dist;
1015 //creates tree from similarity matrix and write out file
1016 Tree* newTree = createTree(matrix);
1017 if (newTree != NULL) {
1018 newTree->print(outAll);
1019 trees.push_back(newTree);
1023 if (m->control_pressed) { for (int k = 0; k < trees.size(); k++) { delete trees[k]; } }
1025 if (m->debug) { m->mothurOut("[DEBUG]: done all trees.\n"); }
1027 Consensus consensus;
1028 //clear old tree names if any
1029 m->Treenames.clear(); m->Treenames = m->getGroups(); //may have changed if subsample eliminated groups
1030 Tree* conTree = consensus.getTree(trees);
1032 if (m->debug) { m->mothurOut("[DEBUG]: done cons tree.\n"); }
1034 //create a new filename
1035 variables["[tag]"] = "cons";
1036 string conFile = getOutputFileName("tree",variables);
1037 outputNames.push_back(conFile); outputTypes["tree"].push_back(conFile);
1039 m->openOutputFile(conFile, outTree);
1041 if (conTree != NULL) { conTree->print(outTree, "boot"); delete conTree; }
1046 for (int i = 0; i < calcDists.size(); i++) {
1047 if (m->control_pressed) { break; }
1050 vector< vector<double> > matrix; //square matrix to represent the distance
1051 matrix.resize(thisLookup.size());
1052 for (int k = 0; k < thisLookup.size(); k++) { matrix[k].resize(thisLookup.size(), 0.0); }
1054 for (int j = 0; j < calcDists[i].size(); j++) {
1055 int row = calcDists[i][j].seq1;
1056 int column = calcDists[i][j].seq2;
1057 double dist = calcDists[i][j].dist;
1059 matrix[row][column] = dist;
1060 matrix[column][row] = dist;
1063 //create a new filename
1064 map<string, string> variables;
1065 variables["[filename]"] = outputDir + m->getRootName(m->getSimpleName(inputfile));
1066 variables["[calc]"] = treeCalculators[i]->getName();
1067 variables["[distance]"] = thisLookup[0]->getLabel();
1068 variables["[tag]"] = "";
1069 string outputFile = getOutputFileName("tree",variables);
1070 outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile);
1072 //creates tree from similarity matrix and write out file
1073 Tree* newTree = createTree(matrix);
1074 if (newTree != NULL) { writeTree(outputFile, newTree); delete newTree; }
1080 catch(exception& e) {
1081 m->errorOut(e, "TreeGroupCommand", "process");
1085 /**************************************************************************************************/
1086 int TreeGroupCommand::driver(vector<SharedRAbundVector*> thisLookup, int start, int end, vector< vector<seqDist> >& calcDists) {
1088 vector<SharedRAbundVector*> subset;
1089 for (int k = start; k < end; k++) { // pass cdd each set of groups to compare
1091 for (int l = 0; l < k; l++) {
1093 if (k != l) { //we dont need to similiarity of a groups to itself
1094 subset.clear(); //clear out old pair of sharedrabunds
1095 //add new pair of sharedrabunds
1096 subset.push_back(thisLookup[k]); subset.push_back(thisLookup[l]);
1098 for(int i=0;i<treeCalculators.size();i++) {
1100 //if this calc needs all groups to calculate the pair load all groups
1101 if (treeCalculators[i]->getNeedsAll()) {
1102 //load subset with rest of lookup for those calcs that need everyone to calc for a pair
1103 for (int w = 0; w < thisLookup.size(); w++) {
1104 if ((w != k) && (w != l)) { subset.push_back(thisLookup[w]); }
1108 vector<double> tempdata = treeCalculators[i]->getValues(subset); //saves the calculator outputs
1110 if (m->control_pressed) { return 1; }
1112 seqDist temp(l, k, -(tempdata[0]-1.0));
1113 calcDists[i].push_back(temp);
1121 catch(exception& e) {
1122 m->errorOut(e, "TreeGroupCommand", "driver");
1126 /***********************************************************/