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",false,false); parameters.push_back(pshared);
18 CommandParameter pphylip("phylip", "InputTypes", "", "", "PhylipColumnShared", "PhylipColumnShared", "none",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",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", "jclass-thetayc", "", "", "",true,false); parameters.push_back(pcalc);
30 CommandParameter pprocessors("processors", "Number", "", "1", "", "", "",false,false); 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::getOutputFileNameTag(string type, string inputName=""){
76 string outputFileName = "";
77 map<string, vector<string> >::iterator it;
79 //is this a type this command creates
80 it = outputTypes.find(type);
81 if (it == outputTypes.end()) { m->mothurOut("[ERROR]: this command doesn't create a " + type + " output file.\n"); }
83 if (type == "tree") { outputFileName = "tre"; }
84 else { m->mothurOut("[ERROR]: No definition for type " + type + " output file tag.\n"); m->control_pressed = true; }
86 return outputFileName;
89 m->errorOut(e, "TreeGroupCommand", "getOutputFileNameTag");
93 //**********************************************************************************************************************
94 TreeGroupCommand::TreeGroupCommand(){
96 abort = true; calledHelp = true;
98 //initialize outputTypes
99 vector<string> tempOutNames;
100 outputTypes["tree"] = tempOutNames;
102 catch(exception& e) {
103 m->errorOut(e, "TreeGroupCommand", "TreeGroupCommand");
107 //**********************************************************************************************************************
109 TreeGroupCommand::TreeGroupCommand(string option) {
111 abort = false; calledHelp = false;
114 //allow user to run help
115 if(option == "help") { help(); abort = true; calledHelp = true; }
116 else if(option == "citation") { citation(); abort = true; calledHelp = true;}
119 vector<string> myArray = setParameters();
121 OptionParser parser(option);
122 map<string, string> parameters = parser. getParameters();
124 ValidParameters validParameter;
125 map<string, string>::iterator it;
127 //check to make sure all parameters are valid for command
128 for (it = parameters.begin(); it != parameters.end(); it++) {
129 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
132 //initialize outputTypes
133 vector<string> tempOutNames;
134 outputTypes["tree"] = tempOutNames;
136 //if the user changes the input directory command factory will send this info to us in the output parameter
137 string inputDir = validParameter.validFile(parameters, "inputdir", false);
138 if (inputDir == "not found"){ inputDir = ""; }
141 it = parameters.find("phylip");
142 //user has given a template file
143 if(it != parameters.end()){
144 path = m->hasPath(it->second);
145 //if the user has not given a path then, add inputdir. else leave path alone.
146 if (path == "") { parameters["phylip"] = inputDir + it->second; }
149 it = parameters.find("column");
150 //user has given a template file
151 if(it != parameters.end()){
152 path = m->hasPath(it->second);
153 //if the user has not given a path then, add inputdir. else leave path alone.
154 if (path == "") { parameters["column"] = inputDir + it->second; }
157 it = parameters.find("name");
158 //user has given a template file
159 if(it != parameters.end()){
160 path = m->hasPath(it->second);
161 //if the user has not given a path then, add inputdir. else leave path alone.
162 if (path == "") { parameters["name"] = inputDir + it->second; }
165 it = parameters.find("count");
166 //user has given a template file
167 if(it != parameters.end()){
168 path = m->hasPath(it->second);
169 //if the user has not given a path then, add inputdir. else leave path alone.
170 if (path == "") { parameters["count"] = inputDir + it->second; }
174 //check for required parameters
175 phylipfile = validParameter.validFile(parameters, "phylip", true);
176 if (phylipfile == "not open") { phylipfile = ""; abort = true; }
177 else if (phylipfile == "not found") { phylipfile = ""; }
178 else { inputfile = phylipfile; format = "phylip"; m->setPhylipFile(phylipfile); }
180 columnfile = validParameter.validFile(parameters, "column", true);
181 if (columnfile == "not open") { columnfile = ""; abort = true; }
182 else if (columnfile == "not found") { columnfile = ""; }
183 else { inputfile = columnfile; format = "column"; m->setColumnFile(columnfile); }
185 sharedfile = validParameter.validFile(parameters, "shared", true);
186 if (sharedfile == "not open") { sharedfile = ""; abort = true; }
187 else if (sharedfile == "not found") { sharedfile = ""; }
188 else { inputfile = sharedfile; format = "sharedfile"; m->setSharedFile(sharedfile); }
190 namefile = validParameter.validFile(parameters, "name", true);
191 if (namefile == "not open") { abort = true; }
192 else if (namefile == "not found") { namefile = ""; }
193 else { m->setNameFile(namefile); }
195 countfile = validParameter.validFile(parameters, "count", true);
196 if (countfile == "not open") { abort = true; countfile = ""; }
197 else if (countfile == "not found") { countfile = ""; }
198 else { m->setCountTableFile(countfile); }
200 if ((phylipfile == "") && (columnfile == "") && (sharedfile == "")) {
201 //is there are current file available for either of these?
202 //give priority to shared, then column, then phylip
203 sharedfile = m->getSharedFile();
204 if (sharedfile != "") { inputfile = sharedfile; format = "sharedfile"; m->mothurOut("Using " + sharedfile + " as input file for the shared parameter."); m->mothurOutEndLine(); }
206 columnfile = m->getColumnFile();
207 if (columnfile != "") { inputfile = columnfile; format = "column"; m->mothurOut("Using " + columnfile + " as input file for the column parameter."); m->mothurOutEndLine(); }
209 phylipfile = m->getPhylipFile();
210 if (phylipfile != "") { inputfile = phylipfile; format = "phylip"; m->mothurOut("Using " + phylipfile + " as input file for the phylip parameter."); m->mothurOutEndLine(); }
212 m->mothurOut("No valid current files. You must provide a shared, phylip or column file."); m->mothurOutEndLine();
218 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; }
220 if (columnfile != "") {
221 if ((namefile == "") && (countfile == "")){
222 namefile = m->getNameFile();
223 if (namefile != "") { m->mothurOut("Using " + namefile + " as input file for the name parameter."); m->mothurOutEndLine(); }
225 countfile = m->getCountTableFile();
226 if (countfile != "") { m->mothurOut("Using " + countfile + " as input file for the count parameter."); m->mothurOutEndLine(); }
228 m->mothurOut("You need to provide a namefile or countfile if you are going to use the column format."); m->mothurOutEndLine();
236 //check for optional parameter and set defaults
237 // ...at some point should added some additional type checking...
238 label = validParameter.validFile(parameters, "label", false);
239 if (label == "not found") { label = ""; }
241 if(label != "all") { m->splitAtDash(label, labels); allLines = 0; }
242 else { allLines = 1; }
245 groups = validParameter.validFile(parameters, "groups", false);
246 if (groups == "not found") { groups = ""; }
248 m->splitAtDash(groups, Groups);
249 m->setGroups(Groups);
252 calc = validParameter.validFile(parameters, "calc", false);
253 if (calc == "not found") { calc = "jclass-thetayc"; }
255 if (calc == "default") { calc = "jclass-thetayc"; }
257 m->splitAtDash(calc, Estimators);
258 if (m->inUsersGroups("citation", Estimators)) {
259 ValidCalculators validCalc; validCalc.printCitations(Estimators);
260 //remove citation from list of calcs
261 for (int i = 0; i < Estimators.size(); i++) { if (Estimators[i] == "citation") { Estimators.erase(Estimators.begin()+i); break; } }
265 temp = validParameter.validFile(parameters, "precision", false); if (temp == "not found") { temp = "100"; }
266 m->mothurConvert(temp, precision);
268 temp = validParameter.validFile(parameters, "cutoff", false); if (temp == "not found") { temp = "10"; }
269 m->mothurConvert(temp, cutoff);
270 cutoff += (5 / (precision * 10.0));
272 temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); }
273 m->setProcessors(temp);
274 m->mothurConvert(temp, processors);
276 temp = validParameter.validFile(parameters, "iters", false); if (temp == "not found") { temp = "1000"; }
277 m->mothurConvert(temp, iters);
279 temp = validParameter.validFile(parameters, "subsample", false); if (temp == "not found") { temp = "F"; }
280 if (m->isNumeric1(temp)) { m->mothurConvert(temp, subsampleSize); subsample = true; }
282 if (m->isTrue(temp)) { subsample = true; subsampleSize = -1; } //we will set it to smallest group later
283 else { subsample = false; }
286 if (subsample == false) { iters = 1; }
288 if (subsample && (format != "sharedfile")) { m->mothurOut("[ERROR]: the subsample parameter can only be used with a shared file.\n"); abort=true; }
290 //if the user changes the output directory command factory will send this info to us in the output parameter
291 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){
293 outputDir += m->hasPath(inputfile); //if user entered a file with a path then preserve it
298 catch(exception& e) {
299 m->errorOut(e, "TreeGroupCommand", "TreeGroupCommand");
303 //**********************************************************************************************************************
305 TreeGroupCommand::~TreeGroupCommand(){
306 if (abort == false) {
307 if (format == "sharedfile") { delete input; }
308 else { delete list; }
314 //**********************************************************************************************************************
316 int TreeGroupCommand::execute(){
319 if (abort == true) { if (calledHelp) { return 0; } return 2; }
321 if (format == "sharedfile") {
323 ValidCalculators validCalculator;
325 for (int i=0; i<Estimators.size(); i++) {
326 if (validCalculator.isValidCalculator("treegroup", Estimators[i]) == true) {
327 if (Estimators[i] == "sharedsobs") {
328 treeCalculators.push_back(new SharedSobsCS());
329 }else if (Estimators[i] == "sharedchao") {
330 treeCalculators.push_back(new SharedChao1());
331 }else if (Estimators[i] == "sharedace") {
332 treeCalculators.push_back(new SharedAce());
333 }else if (Estimators[i] == "jabund") {
334 treeCalculators.push_back(new JAbund());
335 }else if (Estimators[i] == "sorabund") {
336 treeCalculators.push_back(new SorAbund());
337 }else if (Estimators[i] == "jclass") {
338 treeCalculators.push_back(new Jclass());
339 }else if (Estimators[i] == "sorclass") {
340 treeCalculators.push_back(new SorClass());
341 }else if (Estimators[i] == "jest") {
342 treeCalculators.push_back(new Jest());
343 }else if (Estimators[i] == "sorest") {
344 treeCalculators.push_back(new SorEst());
345 }else if (Estimators[i] == "thetayc") {
346 treeCalculators.push_back(new ThetaYC());
347 }else if (Estimators[i] == "thetan") {
348 treeCalculators.push_back(new ThetaN());
349 }else if (Estimators[i] == "kstest") {
350 treeCalculators.push_back(new KSTest());
351 }else if (Estimators[i] == "sharednseqs") {
352 treeCalculators.push_back(new SharedNSeqs());
353 }else if (Estimators[i] == "ochiai") {
354 treeCalculators.push_back(new Ochiai());
355 }else if (Estimators[i] == "anderberg") {
356 treeCalculators.push_back(new Anderberg());
357 }else if (Estimators[i] == "kulczynski") {
358 treeCalculators.push_back(new Kulczynski());
359 }else if (Estimators[i] == "kulczynskicody") {
360 treeCalculators.push_back(new KulczynskiCody());
361 }else if (Estimators[i] == "lennon") {
362 treeCalculators.push_back(new Lennon());
363 }else if (Estimators[i] == "morisitahorn") {
364 treeCalculators.push_back(new MorHorn());
365 }else if (Estimators[i] == "braycurtis") {
366 treeCalculators.push_back(new BrayCurtis());
367 }else if (Estimators[i] == "whittaker") {
368 treeCalculators.push_back(new Whittaker());
369 }else if (Estimators[i] == "odum") {
370 treeCalculators.push_back(new Odum());
371 }else if (Estimators[i] == "canberra") {
372 treeCalculators.push_back(new Canberra());
373 }else if (Estimators[i] == "structeuclidean") {
374 treeCalculators.push_back(new StructEuclidean());
375 }else if (Estimators[i] == "structchord") {
376 treeCalculators.push_back(new StructChord());
377 }else if (Estimators[i] == "hellinger") {
378 treeCalculators.push_back(new Hellinger());
379 }else if (Estimators[i] == "manhattan") {
380 treeCalculators.push_back(new Manhattan());
381 }else if (Estimators[i] == "structpearson") {
382 treeCalculators.push_back(new StructPearson());
383 }else if (Estimators[i] == "soergel") {
384 treeCalculators.push_back(new Soergel());
385 }else if (Estimators[i] == "spearman") {
386 treeCalculators.push_back(new Spearman());
387 }else if (Estimators[i] == "structkulczynski") {
388 treeCalculators.push_back(new StructKulczynski());
389 }else if (Estimators[i] == "speciesprofile") {
390 treeCalculators.push_back(new SpeciesProfile());
391 }else if (Estimators[i] == "hamming") {
392 treeCalculators.push_back(new Hamming());
393 }else if (Estimators[i] == "structchi2") {
394 treeCalculators.push_back(new StructChi2());
395 }else if (Estimators[i] == "gower") {
396 treeCalculators.push_back(new Gower());
397 }else if (Estimators[i] == "memchi2") {
398 treeCalculators.push_back(new MemChi2());
399 }else if (Estimators[i] == "memchord") {
400 treeCalculators.push_back(new MemChord());
401 }else if (Estimators[i] == "memeuclidean") {
402 treeCalculators.push_back(new MemEuclidean());
403 }else if (Estimators[i] == "mempearson") {
404 treeCalculators.push_back(new MemPearson());
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();
439 if (m->control_pressed) { return 0; }
444 if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } return 0; }
447 filename = inputfile;
449 ReadMatrix* readMatrix;
450 if (format == "column") { readMatrix = new ReadColumnMatrix(filename); }
451 else if (format == "phylip") { readMatrix = new ReadPhylipMatrix(filename); }
453 readMatrix->setCutoff(cutoff);
457 nameMap = new NameAssignment(namefile);
459 readMatrix->read(nameMap);
460 }else if (countfile != "") {
461 ct = new CountTable();
462 ct->readTable(countfile);
463 readMatrix->read(ct);
466 list = readMatrix->getListVector();
467 SparseDistanceMatrix* dMatrix = readMatrix->getDMatrix();
470 if (ct != NULL) { delete ct; }
471 ct = new CountTable();
473 map<string, string> groupMap;
475 for (int i = 0; i < list->getNumBins(); i++) {
476 string bin = list->get(i);
481 ct->createTable(nameMap, groupMap, gps);
483 vector<string> namesGroups = ct->getNamesOfGroups();
484 m->setGroups(namesGroups);
486 //clear globaldatas old tree names if any
487 m->Treenames.clear();
489 //fills globaldatas tree names
490 m->Treenames = m->getGroups();
492 //used in tree constructor
495 if (m->control_pressed) { return 0; }
497 vector< vector<double> > matrix = makeSimsDist(dMatrix);
501 if (m->control_pressed) { return 0; }
503 //create a new filename
504 string outputFile = outputDir + m->getRootName(m->getSimpleName(inputfile)) + getOutputFileNameTag("tree");
505 outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile);
507 Tree* newTree = createTree(matrix);
509 if (newTree != NULL) { writeTree(outputFile, newTree); delete newTree; }
511 if (m->control_pressed) { return 0; }
513 m->mothurOut("Tree complete. "); m->mothurOutEndLine();
517 //reset groups parameter
520 //set tree file as new current treefile
522 itTypes = outputTypes.find("tree");
523 if (itTypes != outputTypes.end()) {
524 if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setTreeFile(current); }
527 m->mothurOutEndLine();
528 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
529 for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
530 m->mothurOutEndLine();
534 catch(exception& e) {
535 m->errorOut(e, "TreeGroupCommand", "execute");
539 //**********************************************************************************************************************
541 Tree* TreeGroupCommand::createTree(vector< vector<double> >& simMatrix){
544 t = new Tree(ct, simMatrix);
546 if (m->control_pressed) { delete t; t = NULL; return t; }
553 catch(exception& e) {
554 m->errorOut(e, "TreeGroupCommand", "createTree");
558 /***********************************************************/
559 int TreeGroupCommand::writeTree(string out, Tree* T) {
563 t->createNewickFile(out);
565 if (m->control_pressed) { m->mothurRemove(out); outputNames.pop_back(); return 1; }
570 catch(exception& e) {
571 m->errorOut(e, "TreeGroupCommand", "printSims");
576 /***********************************************************/
577 void TreeGroupCommand::printSims(ostream& out, vector< vector<double> >& simMatrix) {
580 for (int m = 0; m < simMatrix.size(); m++) {
581 //out << lookup[m]->getGroup() << '\t';
582 for (int n = 0; n < simMatrix.size(); n++) {
583 out << simMatrix[m][n] << '\t';
589 catch(exception& e) {
590 m->errorOut(e, "TreeGroupCommand", "printSims");
594 /***********************************************************/
595 vector< vector<double> > TreeGroupCommand::makeSimsDist(SparseDistanceMatrix* matrix) {
597 numGroups = list->size();
599 //initialize simMatrix
600 vector< vector<double> > simMatrix;
601 simMatrix.resize(numGroups);
602 for (int k = 0; k < simMatrix.size(); k++) {
603 for (int j = 0; j < simMatrix.size(); j++) {
604 simMatrix[k].push_back(0.0);
608 //go through sparse matrix and fill sims
609 //go through each cell in the sparsematrix
610 for (int i = 0; i < matrix->seqVec.size(); i++) {
611 for (int j = 0; j < matrix->seqVec[i].size(); j++) {
613 //already checked everyone else in row
614 if (i < matrix->seqVec[i][j].index) {
615 simMatrix[i][matrix->seqVec[i][j].index] = -(matrix->seqVec[i][j].dist -1.0);
616 simMatrix[matrix->seqVec[i][j].index][i] = -(matrix->seqVec[i][j].dist -1.0);
618 if (m->control_pressed) { return simMatrix; }
625 catch(exception& e) {
626 m->errorOut(e, "TreeGroupCommand", "makeSimsDist");
631 /***********************************************************/
632 int TreeGroupCommand::makeSimsShared() {
636 if (subsampleSize == -1) { //user has not set size, set size = smallest samples size
637 subsampleSize = lookup[0]->getNumSeqs();
638 for (int i = 1; i < lookup.size(); i++) {
639 int thisSize = lookup[i]->getNumSeqs();
641 if (thisSize < subsampleSize) { subsampleSize = thisSize; }
646 vector<SharedRAbundVector*> temp;
647 for (int i = 0; i < lookup.size(); i++) {
648 if (lookup[i]->getNumSeqs() < subsampleSize) {
649 m->mothurOut(lookup[i]->getGroup() + " contains " + toString(lookup[i]->getNumSeqs()) + ". Eliminating."); m->mothurOutEndLine();
652 Groups.push_back(lookup[i]->getGroup());
653 temp.push_back(lookup[i]);
657 m->setGroups(Groups);
660 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; }
663 numGroups = lookup.size();
664 lines.resize(processors);
665 for (int i = 0; i < processors; i++) {
666 lines[i].start = int (sqrt(float(i)/float(processors)) * numGroups);
667 lines[i].end = int (sqrt(float(i+1)/float(processors)) * numGroups);
670 set<string> processedLabels;
671 set<string> userLabels = labels;
673 //as long as you are not at the end of the file or done wih the lines you want
674 while((lookup[0] != NULL) && ((allLines == 1) || (userLabels.size() != 0))) {
675 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; }
677 if(allLines == 1 || labels.count(lookup[0]->getLabel()) == 1){
678 m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
681 processedLabels.insert(lookup[0]->getLabel());
682 userLabels.erase(lookup[0]->getLabel());
685 if ((m->anyLabelsToProcess(lookup[0]->getLabel(), userLabels, "") == true) && (processedLabels.count(lastLabel) != 1)) {
686 string saveLabel = lookup[0]->getLabel();
688 for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
689 lookup = input->getSharedRAbundVectors(lastLabel);
691 m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
694 processedLabels.insert(lookup[0]->getLabel());
695 userLabels.erase(lookup[0]->getLabel());
697 //restore real lastlabel to save below
698 lookup[0]->setLabel(saveLabel);
701 lastLabel = lookup[0]->getLabel();
703 //get next line to process
704 for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
705 lookup = input->getSharedRAbundVectors();
708 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; }
710 //output error messages about any remaining user labels
711 set<string>::iterator it;
712 bool needToRun = false;
713 for (it = userLabels.begin(); it != userLabels.end(); it++) {
714 m->mothurOut("Your file does not include the label " + *it);
715 if (processedLabels.count(lastLabel) != 1) {
716 m->mothurOut(". I will use " + lastLabel + "."); m->mothurOutEndLine();
719 m->mothurOut(". Please refer to " + lastLabel + "."); m->mothurOutEndLine();
723 //run last label if you need to
724 if (needToRun == true) {
725 for (int i = 0; i < lookup.size(); i++) { if (lookup[i] != NULL) { delete lookup[i]; } }
726 lookup = input->getSharedRAbundVectors(lastLabel);
728 m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
730 for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
733 for(int i = 0 ; i < treeCalculators.size(); i++) { delete treeCalculators[i]; }
737 catch(exception& e) {
738 m->errorOut(e, "TreeGroupCommand", "makeSimsShared");
743 /***********************************************************/
744 int TreeGroupCommand::process(vector<SharedRAbundVector*> thisLookup) {
746 vector< vector< vector<seqDist> > > calcDistsTotals; //each iter, one for each calc, then each groupCombos dists. this will be used to make .dist files
747 vector< vector<seqDist> > calcDists; calcDists.resize(treeCalculators.size());
749 for (int thisIter = 0; thisIter < iters; thisIter++) {
751 vector<SharedRAbundVector*> thisItersLookup = thisLookup;
755 vector<string> tempLabels; //dont need since we arent printing the sampled sharedRabunds
757 //make copy of lookup so we don't get access violations
758 vector<SharedRAbundVector*> newLookup;
759 for (int k = 0; k < thisItersLookup.size(); k++) {
760 SharedRAbundVector* temp = new SharedRAbundVector();
761 temp->setLabel(thisItersLookup[k]->getLabel());
762 temp->setGroup(thisItersLookup[k]->getGroup());
763 newLookup.push_back(temp);
767 for (int k = 0; k < thisItersLookup[0]->getNumBins(); k++) {
768 if (m->control_pressed) { for (int j = 0; j < newLookup.size(); j++) { delete newLookup[j]; } return 0; }
769 for (int j = 0; j < thisItersLookup.size(); j++) { newLookup[j]->push_back(thisItersLookup[j]->getAbundance(k), thisItersLookup[j]->getGroup()); }
772 tempLabels = sample.getSample(newLookup, subsampleSize);
773 thisItersLookup = newLookup;
777 driver(thisItersLookup, 0, numGroups, calcDists);
780 vector<int> processIDS;
782 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
783 //loop through and create all the processes you want
784 while (process != processors) {
788 processIDS.push_back(pid);
792 driver(thisItersLookup, lines[process].start, lines[process].end, calcDists);
794 string tempdistFileName = m->getRootName(m->getSimpleName(sharedfile)) + toString(getpid()) + ".dist";
796 m->openOutputFile(tempdistFileName, outtemp);
798 for (int i = 0; i < calcDists.size(); i++) {
799 outtemp << calcDists[i].size() << endl;
801 for (int j = 0; j < calcDists[i].size(); j++) {
802 outtemp << calcDists[i][j].seq1 << '\t' << calcDists[i][j].seq2 << '\t' << calcDists[i][j].dist << endl;
809 m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine();
810 for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); }
815 //parent do your part
816 driver(thisItersLookup, lines[0].start, lines[0].end, calcDists);
818 //force parent to wait until all the processes are done
819 for (int i = 0; i < processIDS.size(); i++) {
820 int temp = processIDS[i];
824 for (int i = 0; i < processIDS.size(); i++) {
825 string tempdistFileName = m->getRootName(m->getSimpleName(sharedfile)) + toString(processIDS[i]) + ".dist";
827 m->openInputFile(tempdistFileName, intemp);
829 for (int k = 0; k < calcDists.size(); k++) {
831 intemp >> size; m->gobble(intemp);
833 for (int j = 0; j < size; j++) {
838 intemp >> seq1 >> seq2 >> dist; m->gobble(intemp);
840 seqDist tempDist(seq1, seq2, dist);
841 calcDists[k].push_back(tempDist);
845 m->mothurRemove(tempdistFileName);
848 //////////////////////////////////////////////////////////////////////////////////////////////////////
849 //Windows version shared memory, so be careful when passing variables through the treeSharedData struct.
850 //Above fork() will clone, so memory is separate, but that's not the case with windows,
851 //Taking advantage of shared memory to pass results vectors.
852 //////////////////////////////////////////////////////////////////////////////////////////////////////
854 vector<treeSharedData*> pDataArray;
855 DWORD dwThreadIdArray[processors-1];
856 HANDLE hThreadArray[processors-1];
858 //Create processor worker threads.
859 for( int i=1; i<processors; i++ ){
861 //make copy of lookup so we don't get access violations
862 vector<SharedRAbundVector*> newLookup;
863 for (int k = 0; k < thisItersLookup.size(); k++) {
864 SharedRAbundVector* temp = new SharedRAbundVector();
865 temp->setLabel(thisItersLookup[k]->getLabel());
866 temp->setGroup(thisItersLookup[k]->getGroup());
867 newLookup.push_back(temp);
871 for (int k = 0; k < thisItersLookup[0]->getNumBins(); k++) {
872 if (m->control_pressed) { for (int j = 0; j < newLookup.size(); j++) { delete newLookup[j]; } return 0; }
873 for (int j = 0; j < thisItersLookup.size(); j++) { newLookup[j]->push_back(thisItersLookup[j]->getAbundance(k), thisItersLookup[j]->getGroup()); }
876 // Allocate memory for thread data.
877 treeSharedData* tempSum = new treeSharedData(m, lines[i].start, lines[i].end, Estimators, newLookup);
878 pDataArray.push_back(tempSum);
879 processIDS.push_back(i);
881 hThreadArray[i-1] = CreateThread(NULL, 0, MyTreeSharedThreadFunction, pDataArray[i-1], 0, &dwThreadIdArray[i-1]);
884 //parent do your part
885 driver(thisItersLookup, lines[0].start, lines[0].end, calcDists);
887 //Wait until all threads have terminated.
888 WaitForMultipleObjects(processors-1, hThreadArray, TRUE, INFINITE);
890 //Close all thread handles and free memory allocations.
891 for(int i=0; i < pDataArray.size(); i++){
892 for (int j = 0; j < pDataArray[i]->thisLookup.size(); j++) { delete pDataArray[i]->thisLookup[j]; }
894 for (int k = 0; k < calcDists.size(); k++) {
895 int size = pDataArray[i]->calcDists[k].size();
896 for (int j = 0; j < size; j++) { calcDists[k].push_back(pDataArray[i]->calcDists[k][j]); }
899 CloseHandle(hThreadArray[i]);
900 delete pDataArray[i];
906 calcDistsTotals.push_back(calcDists);
911 for (int i = 0; i < thisItersLookup.size(); i++) { delete thisItersLookup[i]; }
912 thisItersLookup.clear();
913 for (int i = 0; i < calcDists.size(); i++) { calcDists[i].clear(); }
918 //we need to find the average distance and standard deviation for each groups distance
920 vector< vector<seqDist> > calcAverages; calcAverages.resize(treeCalculators.size());
921 for (int i = 0; i < calcAverages.size(); i++) { //initialize sums to zero.
922 calcAverages[i].resize(calcDistsTotals[0][i].size());
924 for (int j = 0; j < calcAverages[i].size(); j++) {
925 calcAverages[i][j].seq1 = calcDists[i][j].seq1;
926 calcAverages[i][j].seq2 = calcDists[i][j].seq2;
927 calcAverages[i][j].dist = 0.0;
931 for (int thisIter = 0; thisIter < iters; thisIter++) { //sum all groups dists for each calculator
932 for (int i = 0; i < calcAverages.size(); i++) { //initialize sums to zero.
933 for (int j = 0; j < calcAverages[i].size(); j++) {
934 calcAverages[i][j].dist += calcDistsTotals[thisIter][i][j].dist;
939 for (int i = 0; i < calcAverages.size(); i++) { //finds average.
940 for (int j = 0; j < calcAverages[i].size(); j++) {
941 calcAverages[i][j].dist /= (float) iters;
945 //create average tree for each calc
946 for (int i = 0; i < calcDists.size(); i++) {
947 vector< vector<double> > matrix; //square matrix to represent the distance
948 matrix.resize(thisLookup.size());
949 for (int k = 0; k < thisLookup.size(); k++) { matrix[k].resize(thisLookup.size(), 0.0); }
951 for (int j = 0; j < calcAverages[i].size(); j++) {
952 int row = calcAverages[i][j].seq1;
953 int column = calcAverages[i][j].seq2;
954 float dist = calcAverages[i][j].dist;
956 matrix[row][column] = dist;
957 matrix[column][row] = dist;
960 //create a new filename
961 string outputFile = outputDir + m->getRootName(m->getSimpleName(inputfile)) + treeCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + ".ave." + getOutputFileNameTag("tree");
962 outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile);
964 //creates tree from similarity matrix and write out file
965 Tree* newTree = createTree(matrix);
966 if (newTree != NULL) { writeTree(outputFile, newTree); }
969 //create all trees for each calc and find their consensus tree
970 for (int i = 0; i < calcDists.size(); i++) {
971 if (m->control_pressed) { break; }
973 //create a new filename
974 string outputFile = outputDir + m->getRootName(m->getSimpleName(inputfile)) + treeCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + ".all." + getOutputFileNameTag("tree");
975 outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile);
978 m->openOutputFile(outputFile, outAll);
981 for (int myIter = 0; myIter < iters; myIter++) {
983 if(m->control_pressed) { break; }
986 vector< vector<double> > matrix; //square matrix to represent the distance
987 matrix.resize(thisLookup.size());
988 for (int k = 0; k < thisLookup.size(); k++) { matrix[k].resize(thisLookup.size(), 0.0); }
990 for (int j = 0; j < calcDistsTotals[myIter][i].size(); j++) {
991 int row = calcDistsTotals[myIter][i][j].seq1;
992 int column = calcDistsTotals[myIter][i][j].seq2;
993 double dist = calcDistsTotals[myIter][i][j].dist;
995 matrix[row][column] = dist;
996 matrix[column][row] = dist;
999 //creates tree from similarity matrix and write out file
1000 Tree* newTree = createTree(matrix);
1001 if (newTree != NULL) {
1002 newTree->print(outAll);
1003 trees.push_back(newTree);
1007 if (m->control_pressed) { for (int k = 0; k < trees.size(); k++) { delete trees[k]; } }
1009 Consensus consensus;
1010 //clear old tree names if any
1011 m->Treenames.clear(); m->Treenames = m->getGroups(); //may have changed if subsample eliminated groups
1012 Tree* conTree = consensus.getTree(trees);
1014 //create a new filename
1015 string conFile = outputDir + m->getRootName(m->getSimpleName(inputfile)) + treeCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + ".cons." + getOutputFileNameTag("tree");
1016 outputNames.push_back(conFile); outputTypes["tree"].push_back(conFile);
1018 m->openOutputFile(conFile, outTree);
1020 if (conTree != NULL) { conTree->print(outTree, "boot"); delete conTree; }
1025 for (int i = 0; i < calcDists.size(); i++) {
1026 if (m->control_pressed) { break; }
1029 vector< vector<double> > matrix; //square matrix to represent the distance
1030 matrix.resize(thisLookup.size());
1031 for (int k = 0; k < thisLookup.size(); k++) { matrix[k].resize(thisLookup.size(), 0.0); }
1033 for (int j = 0; j < calcDists[i].size(); j++) {
1034 int row = calcDists[i][j].seq1;
1035 int column = calcDists[i][j].seq2;
1036 double dist = calcDists[i][j].dist;
1038 matrix[row][column] = dist;
1039 matrix[column][row] = dist;
1042 //create a new filename
1043 string outputFile = outputDir + m->getRootName(m->getSimpleName(inputfile)) + treeCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + "." + getOutputFileNameTag("tree");
1044 outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile);
1046 //creates tree from similarity matrix and write out file
1047 Tree* newTree = createTree(matrix);
1048 if (newTree != NULL) { writeTree(outputFile, newTree); delete newTree; }
1054 catch(exception& e) {
1055 m->errorOut(e, "TreeGroupCommand", "process");
1059 /**************************************************************************************************/
1060 int TreeGroupCommand::driver(vector<SharedRAbundVector*> thisLookup, int start, int end, vector< vector<seqDist> >& calcDists) {
1062 vector<SharedRAbundVector*> subset;
1063 for (int k = start; k < end; k++) { // pass cdd each set of groups to compare
1065 for (int l = 0; l < k; l++) {
1067 if (k != l) { //we dont need to similiarity of a groups to itself
1068 subset.clear(); //clear out old pair of sharedrabunds
1069 //add new pair of sharedrabunds
1070 subset.push_back(thisLookup[k]); subset.push_back(thisLookup[l]);
1072 for(int i=0;i<treeCalculators.size();i++) {
1074 //if this calc needs all groups to calculate the pair load all groups
1075 if (treeCalculators[i]->getNeedsAll()) {
1076 //load subset with rest of lookup for those calcs that need everyone to calc for a pair
1077 for (int w = 0; w < thisLookup.size(); w++) {
1078 if ((w != k) && (w != l)) { subset.push_back(thisLookup[w]); }
1082 vector<double> tempdata = treeCalculators[i]->getValues(subset); //saves the calculator outputs
1084 if (m->control_pressed) { return 1; }
1086 seqDist temp(l, k, -(tempdata[0]-1.0));
1087 calcDists[i].push_back(temp);
1095 catch(exception& e) {
1096 m->errorOut(e, "TreeGroupCommand", "driver");
1100 /***********************************************************/