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", "", "", "none", "none", "ColumnName",false,false); parameters.push_back(pname);
20 CommandParameter pcolumn("column", "InputTypes", "", "", "PhylipColumnShared", "PhylipColumnShared", "ColumnName",false,false); parameters.push_back(pcolumn);
21 CommandParameter piters("iters", "Number", "", "1000", "", "", "",false,false); parameters.push_back(piters);
22 CommandParameter psubsample("subsample", "String", "", "", "", "", "",false,false); parameters.push_back(psubsample);
23 CommandParameter pcutoff("cutoff", "Number", "", "10", "", "", "",false,false); parameters.push_back(pcutoff);
24 CommandParameter pprecision("precision", "Number", "", "100", "", "", "",false,false); parameters.push_back(pprecision);
25 CommandParameter plabel("label", "String", "", "", "", "", "",false,false); parameters.push_back(plabel);
26 CommandParameter pgroups("groups", "String", "", "", "", "", "",false,false); parameters.push_back(pgroups);
27 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);
29 CommandParameter pprocessors("processors", "Number", "", "1", "", "", "",false,false); parameters.push_back(pprocessors);
30 //CommandParameter poutput("output", "Multiple", "lt-square", "lt", "", "", "",false,false); parameters.push_back(poutput);
31 CommandParameter pinputdir("inputdir", "String", "", "", "", "", "",false,false); parameters.push_back(pinputdir);
32 CommandParameter poutputdir("outputdir", "String", "", "", "", "", "",false,false); parameters.push_back(poutputdir);
34 vector<string> myArray;
35 for (int i = 0; i < parameters.size(); i++) { myArray.push_back(parameters[i].name); }
39 m->errorOut(e, "TreeGroupCommand", "setParameters");
43 //**********************************************************************************************************************
44 string TreeGroupCommand::getHelpString(){
46 string helpString = "";
47 ValidCalculators validCalculator;
48 helpString += "The tree.shared command creates a .tre to represent the similiarity between groups or sequences.\n";
49 helpString += "The tree.shared command parameters are shared, groups, calc, phylip, column, name, cutoff, precision, processors, subsample, iters and label.\n";
50 helpString += "The groups parameter allows you to specify which of the groups in your groupfile you would like included used.\n";
51 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";
52 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";
53 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";
54 helpString += "The tree.shared command should be in the following format: tree.shared(groups=yourGroups, calc=yourCalcs, label=yourLabels).\n";
55 helpString += "The iters parameter allows you to choose the number of times you would like to run the subsample.\n";
56 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";
57 helpString += "Example tree.shared(groups=A-B-C, calc=jabund-sorabund).\n";
58 helpString += "The default value for groups is all the groups in your groupfile.\n";
59 helpString += "The default value for calc is jclass-thetayc.\n";
60 helpString += "The tree.shared command outputs a .tre file for each calculator you specify at each distance you choose.\n";
61 helpString += validCalculator.printCalc("treegroup");
62 helpString += "Or the tree.shared command can be in the following format: tree.shared(phylip=yourPhylipFile).\n";
63 helpString += "Example tree.shared(phylip=abrecovery.dist).\n";
64 helpString += "Note: No spaces between parameter labels (i.e. groups), '=' and parameters (i.e.yourGroups).\n";
68 m->errorOut(e, "TreeGroupCommand", "getHelpString");
72 //**********************************************************************************************************************
73 string TreeGroupCommand::getOutputFileNameTag(string type, string inputName=""){
75 string outputFileName = "";
76 map<string, vector<string> >::iterator it;
78 //is this a type this command creates
79 it = outputTypes.find(type);
80 if (it == outputTypes.end()) { m->mothurOut("[ERROR]: this command doesn't create a " + type + " output file.\n"); }
82 if (type == "tree") { outputFileName = "tre"; }
83 else { m->mothurOut("[ERROR]: No definition for type " + type + " output file tag.\n"); m->control_pressed = true; }
85 return outputFileName;
88 m->errorOut(e, "TreeGroupCommand", "getOutputFileNameTag");
92 //**********************************************************************************************************************
93 TreeGroupCommand::TreeGroupCommand(){
95 abort = true; calledHelp = true;
97 //initialize outputTypes
98 vector<string> tempOutNames;
99 outputTypes["tree"] = tempOutNames;
101 catch(exception& e) {
102 m->errorOut(e, "TreeGroupCommand", "TreeGroupCommand");
106 //**********************************************************************************************************************
108 TreeGroupCommand::TreeGroupCommand(string option) {
110 abort = false; calledHelp = false;
113 //allow user to run help
114 if(option == "help") { help(); abort = true; calledHelp = true; }
115 else if(option == "citation") { citation(); abort = true; calledHelp = true;}
118 vector<string> myArray = setParameters();
120 OptionParser parser(option);
121 map<string, string> parameters = parser. getParameters();
123 ValidParameters validParameter;
124 map<string, string>::iterator it;
126 //check to make sure all parameters are valid for command
127 for (it = parameters.begin(); it != parameters.end(); it++) {
128 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
131 //initialize outputTypes
132 vector<string> tempOutNames;
133 outputTypes["tree"] = tempOutNames;
135 //if the user changes the input directory command factory will send this info to us in the output parameter
136 string inputDir = validParameter.validFile(parameters, "inputdir", false);
137 if (inputDir == "not found"){ inputDir = ""; }
140 it = parameters.find("phylip");
141 //user has given a template file
142 if(it != parameters.end()){
143 path = m->hasPath(it->second);
144 //if the user has not given a path then, add inputdir. else leave path alone.
145 if (path == "") { parameters["phylip"] = inputDir + it->second; }
148 it = parameters.find("column");
149 //user has given a template file
150 if(it != parameters.end()){
151 path = m->hasPath(it->second);
152 //if the user has not given a path then, add inputdir. else leave path alone.
153 if (path == "") { parameters["column"] = inputDir + it->second; }
156 it = parameters.find("name");
157 //user has given a template file
158 if(it != parameters.end()){
159 path = m->hasPath(it->second);
160 //if the user has not given a path then, add inputdir. else leave path alone.
161 if (path == "") { parameters["name"] = inputDir + it->second; }
165 //check for required parameters
166 phylipfile = validParameter.validFile(parameters, "phylip", true);
167 if (phylipfile == "not open") { phylipfile = ""; abort = true; }
168 else if (phylipfile == "not found") { phylipfile = ""; }
169 else { inputfile = phylipfile; format = "phylip"; m->setPhylipFile(phylipfile); }
171 columnfile = validParameter.validFile(parameters, "column", true);
172 if (columnfile == "not open") { columnfile = ""; abort = true; }
173 else if (columnfile == "not found") { columnfile = ""; }
174 else { inputfile = columnfile; format = "column"; m->setColumnFile(columnfile); }
176 sharedfile = validParameter.validFile(parameters, "shared", true);
177 if (sharedfile == "not open") { sharedfile = ""; abort = true; }
178 else if (sharedfile == "not found") { sharedfile = ""; }
179 else { inputfile = sharedfile; format = "sharedfile"; m->setSharedFile(sharedfile); }
181 namefile = validParameter.validFile(parameters, "name", true);
182 if (namefile == "not open") { abort = true; }
183 else if (namefile == "not found") { namefile = ""; }
184 else { m->setNameFile(namefile); }
186 if ((phylipfile == "") && (columnfile == "") && (sharedfile == "")) {
187 //is there are current file available for either of these?
188 //give priority to shared, then column, then phylip
189 sharedfile = m->getSharedFile();
190 if (sharedfile != "") { inputfile = sharedfile; format = "sharedfile"; m->mothurOut("Using " + sharedfile + " as input file for the shared parameter."); m->mothurOutEndLine(); }
192 columnfile = m->getColumnFile();
193 if (columnfile != "") { inputfile = columnfile; format = "column"; m->mothurOut("Using " + columnfile + " as input file for the column parameter."); m->mothurOutEndLine(); }
195 phylipfile = m->getPhylipFile();
196 if (phylipfile != "") { inputfile = phylipfile; format = "phylip"; m->mothurOut("Using " + phylipfile + " as input file for the phylip parameter."); m->mothurOutEndLine(); }
198 m->mothurOut("No valid current files. You must provide a shared, phylip or column file."); m->mothurOutEndLine();
204 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; }
206 if (columnfile != "") {
207 if (namefile == "") {
208 namefile = m->getNameFile();
209 if (namefile != "") { m->mothurOut("Using " + namefile + " as input file for the name parameter."); m->mothurOutEndLine(); }
211 m->mothurOut("You need to provide a namefile if you are going to use the column format."); m->mothurOutEndLine();
217 //check for optional parameter and set defaults
218 // ...at some point should added some additional type checking...
219 label = validParameter.validFile(parameters, "label", false);
220 if (label == "not found") { label = ""; }
222 if(label != "all") { m->splitAtDash(label, labels); allLines = 0; }
223 else { allLines = 1; }
226 groups = validParameter.validFile(parameters, "groups", false);
227 if (groups == "not found") { groups = ""; }
229 m->splitAtDash(groups, Groups);
230 m->setGroups(Groups);
233 calc = validParameter.validFile(parameters, "calc", false);
234 if (calc == "not found") { calc = "jclass-thetayc"; }
236 if (calc == "default") { calc = "jclass-thetayc"; }
238 m->splitAtDash(calc, Estimators);
239 if (m->inUsersGroups("citation", Estimators)) {
240 ValidCalculators validCalc; validCalc.printCitations(Estimators);
241 //remove citation from list of calcs
242 for (int i = 0; i < Estimators.size(); i++) { if (Estimators[i] == "citation") { Estimators.erase(Estimators.begin()+i); break; } }
246 temp = validParameter.validFile(parameters, "precision", false); if (temp == "not found") { temp = "100"; }
247 m->mothurConvert(temp, precision);
249 temp = validParameter.validFile(parameters, "cutoff", false); if (temp == "not found") { temp = "10"; }
250 m->mothurConvert(temp, cutoff);
251 cutoff += (5 / (precision * 10.0));
253 temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); }
254 m->setProcessors(temp);
255 m->mothurConvert(temp, processors);
257 temp = validParameter.validFile(parameters, "iters", false); if (temp == "not found") { temp = "1000"; }
258 m->mothurConvert(temp, iters);
260 temp = validParameter.validFile(parameters, "subsample", false); if (temp == "not found") { temp = "F"; }
261 if (m->isNumeric1(temp)) { m->mothurConvert(temp, subsampleSize); subsample = true; }
263 if (m->isTrue(temp)) { subsample = true; subsampleSize = -1; } //we will set it to smallest group later
264 else { subsample = false; }
267 if (subsample == false) { iters = 1; }
269 if (subsample && (format != "sharedfile")) { m->mothurOut("[ERROR]: the subsample parameter can only be used with a shared file.\n"); abort=true; }
271 //if the user changes the output directory command factory will send this info to us in the output parameter
272 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){
274 outputDir += m->hasPath(inputfile); //if user entered a file with a path then preserve it
279 catch(exception& e) {
280 m->errorOut(e, "TreeGroupCommand", "TreeGroupCommand");
284 //**********************************************************************************************************************
286 TreeGroupCommand::~TreeGroupCommand(){
287 if (abort == false) {
288 if (format == "sharedfile") { delete input; }
289 else { delete list; }
295 //**********************************************************************************************************************
297 int TreeGroupCommand::execute(){
300 if (abort == true) { if (calledHelp) { return 0; } return 2; }
302 if (format == "sharedfile") {
304 ValidCalculators validCalculator;
306 for (int i=0; i<Estimators.size(); i++) {
307 if (validCalculator.isValidCalculator("treegroup", Estimators[i]) == true) {
308 if (Estimators[i] == "sharedsobs") {
309 treeCalculators.push_back(new SharedSobsCS());
310 }else if (Estimators[i] == "sharedchao") {
311 treeCalculators.push_back(new SharedChao1());
312 }else if (Estimators[i] == "sharedace") {
313 treeCalculators.push_back(new SharedAce());
314 }else if (Estimators[i] == "jabund") {
315 treeCalculators.push_back(new JAbund());
316 }else if (Estimators[i] == "sorabund") {
317 treeCalculators.push_back(new SorAbund());
318 }else if (Estimators[i] == "jclass") {
319 treeCalculators.push_back(new Jclass());
320 }else if (Estimators[i] == "sorclass") {
321 treeCalculators.push_back(new SorClass());
322 }else if (Estimators[i] == "jest") {
323 treeCalculators.push_back(new Jest());
324 }else if (Estimators[i] == "sorest") {
325 treeCalculators.push_back(new SorEst());
326 }else if (Estimators[i] == "thetayc") {
327 treeCalculators.push_back(new ThetaYC());
328 }else if (Estimators[i] == "thetan") {
329 treeCalculators.push_back(new ThetaN());
330 }else if (Estimators[i] == "kstest") {
331 treeCalculators.push_back(new KSTest());
332 }else if (Estimators[i] == "sharednseqs") {
333 treeCalculators.push_back(new SharedNSeqs());
334 }else if (Estimators[i] == "ochiai") {
335 treeCalculators.push_back(new Ochiai());
336 }else if (Estimators[i] == "anderberg") {
337 treeCalculators.push_back(new Anderberg());
338 }else if (Estimators[i] == "kulczynski") {
339 treeCalculators.push_back(new Kulczynski());
340 }else if (Estimators[i] == "kulczynskicody") {
341 treeCalculators.push_back(new KulczynskiCody());
342 }else if (Estimators[i] == "lennon") {
343 treeCalculators.push_back(new Lennon());
344 }else if (Estimators[i] == "morisitahorn") {
345 treeCalculators.push_back(new MorHorn());
346 }else if (Estimators[i] == "braycurtis") {
347 treeCalculators.push_back(new BrayCurtis());
348 }else if (Estimators[i] == "whittaker") {
349 treeCalculators.push_back(new Whittaker());
350 }else if (Estimators[i] == "odum") {
351 treeCalculators.push_back(new Odum());
352 }else if (Estimators[i] == "canberra") {
353 treeCalculators.push_back(new Canberra());
354 }else if (Estimators[i] == "structeuclidean") {
355 treeCalculators.push_back(new StructEuclidean());
356 }else if (Estimators[i] == "structchord") {
357 treeCalculators.push_back(new StructChord());
358 }else if (Estimators[i] == "hellinger") {
359 treeCalculators.push_back(new Hellinger());
360 }else if (Estimators[i] == "manhattan") {
361 treeCalculators.push_back(new Manhattan());
362 }else if (Estimators[i] == "structpearson") {
363 treeCalculators.push_back(new StructPearson());
364 }else if (Estimators[i] == "soergel") {
365 treeCalculators.push_back(new Soergel());
366 }else if (Estimators[i] == "spearman") {
367 treeCalculators.push_back(new Spearman());
368 }else if (Estimators[i] == "structkulczynski") {
369 treeCalculators.push_back(new StructKulczynski());
370 }else if (Estimators[i] == "speciesprofile") {
371 treeCalculators.push_back(new SpeciesProfile());
372 }else if (Estimators[i] == "hamming") {
373 treeCalculators.push_back(new Hamming());
374 }else if (Estimators[i] == "structchi2") {
375 treeCalculators.push_back(new StructChi2());
376 }else if (Estimators[i] == "gower") {
377 treeCalculators.push_back(new Gower());
378 }else if (Estimators[i] == "memchi2") {
379 treeCalculators.push_back(new MemChi2());
380 }else if (Estimators[i] == "memchord") {
381 treeCalculators.push_back(new MemChord());
382 }else if (Estimators[i] == "memeuclidean") {
383 treeCalculators.push_back(new MemEuclidean());
384 }else if (Estimators[i] == "mempearson") {
385 treeCalculators.push_back(new MemPearson());
390 //if the users entered no valid calculators don't execute command
391 if (treeCalculators.size() == 0) { m->mothurOut("You have given no valid calculators."); m->mothurOutEndLine(); return 0; }
393 input = new InputData(sharedfile, "sharedfile");
394 lookup = input->getSharedRAbundVectors();
395 lastLabel = lookup[0]->getLabel();
397 if (lookup.size() < 2) { m->mothurOut("You have not provided enough valid groups. I cannot run the command."); m->mothurOutEndLine(); return 0; }
399 //used in tree constructor
402 //create treemap class from groupmap for tree class to use
403 ct = new CountTable();
405 map<string, string> groupMap;
407 for (int i = 0; i < m->getAllGroups().size(); i++) {
408 nameMap.insert(m->getAllGroups()[i]);
409 gps.insert(m->getAllGroups()[i]);
410 groupMap[m->getAllGroups()[i]] = m->getAllGroups()[i];
412 ct->createTable(nameMap, groupMap, gps);
414 //clear globaldatas old tree names if any
415 m->Treenames.clear();
417 //fills globaldatas tree names
418 m->Treenames = m->getGroups();
420 if (m->control_pressed) { return 0; }
425 if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } return 0; }
428 filename = inputfile;
430 ReadMatrix* readMatrix;
431 if (format == "column") { readMatrix = new ReadColumnMatrix(filename); }
432 else if (format == "phylip") { readMatrix = new ReadPhylipMatrix(filename); }
434 readMatrix->setCutoff(cutoff);
437 nameMap = new NameAssignment(namefile);
440 else{ nameMap = NULL; }
442 readMatrix->read(nameMap);
443 list = readMatrix->getListVector();
444 SparseDistanceMatrix* dMatrix = readMatrix->getDMatrix();
447 ct = new CountTable();
449 map<string, string> groupMap;
451 for (int i = 0; i < list->getNumBins(); i++) {
452 string bin = list->get(i);
457 ct->createTable(nameMap, groupMap, gps);
459 vector<string> namesGroups = ct->getNamesOfGroups();
460 m->setGroups(namesGroups);
462 //clear globaldatas old tree names if any
463 m->Treenames.clear();
465 //fills globaldatas tree names
466 m->Treenames = m->getGroups();
468 //used in tree constructor
471 if (m->control_pressed) { return 0; }
473 vector< vector<double> > matrix = makeSimsDist(dMatrix);
477 if (m->control_pressed) { return 0; }
479 //create a new filename
480 string outputFile = outputDir + m->getRootName(m->getSimpleName(inputfile)) + getOutputFileNameTag("tree");
481 outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile);
483 Tree* newTree = createTree(matrix);
485 if (newTree != NULL) { writeTree(outputFile, newTree); delete newTree; }
487 if (m->control_pressed) { return 0; }
489 m->mothurOut("Tree complete. "); m->mothurOutEndLine();
493 //reset groups parameter
496 //set tree file as new current treefile
498 itTypes = outputTypes.find("tree");
499 if (itTypes != outputTypes.end()) {
500 if ((itTypes->second).size() != 0) { current = (itTypes->second)[0]; m->setTreeFile(current); }
503 m->mothurOutEndLine();
504 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
505 for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
506 m->mothurOutEndLine();
510 catch(exception& e) {
511 m->errorOut(e, "TreeGroupCommand", "execute");
515 //**********************************************************************************************************************
517 Tree* TreeGroupCommand::createTree(vector< vector<double> >& simMatrix){
520 t = new Tree(ct, simMatrix);
522 if (m->control_pressed) { delete t; t = NULL; return t; }
529 catch(exception& e) {
530 m->errorOut(e, "TreeGroupCommand", "createTree");
534 /***********************************************************/
535 int TreeGroupCommand::writeTree(string out, Tree* T) {
539 t->createNewickFile(out);
541 if (m->control_pressed) { m->mothurRemove(out); outputNames.pop_back(); return 1; }
546 catch(exception& e) {
547 m->errorOut(e, "TreeGroupCommand", "printSims");
552 /***********************************************************/
553 void TreeGroupCommand::printSims(ostream& out, vector< vector<double> >& simMatrix) {
556 for (int m = 0; m < simMatrix.size(); m++) {
557 //out << lookup[m]->getGroup() << '\t';
558 for (int n = 0; n < simMatrix.size(); n++) {
559 out << simMatrix[m][n] << '\t';
565 catch(exception& e) {
566 m->errorOut(e, "TreeGroupCommand", "printSims");
570 /***********************************************************/
571 vector< vector<double> > TreeGroupCommand::makeSimsDist(SparseDistanceMatrix* matrix) {
573 numGroups = list->size();
575 //initialize simMatrix
576 vector< vector<double> > simMatrix;
577 simMatrix.resize(numGroups);
578 for (int k = 0; k < simMatrix.size(); k++) {
579 for (int j = 0; j < simMatrix.size(); j++) {
580 simMatrix[k].push_back(0.0);
584 //go through sparse matrix and fill sims
585 //go through each cell in the sparsematrix
586 for (int i = 0; i < matrix->seqVec.size(); i++) {
587 for (int j = 0; j < matrix->seqVec[i].size(); j++) {
589 //already checked everyone else in row
590 if (i < matrix->seqVec[i][j].index) {
591 simMatrix[i][matrix->seqVec[i][j].index] = -(matrix->seqVec[i][j].dist -1.0);
592 simMatrix[matrix->seqVec[i][j].index][i] = -(matrix->seqVec[i][j].dist -1.0);
594 if (m->control_pressed) { return simMatrix; }
601 catch(exception& e) {
602 m->errorOut(e, "TreeGroupCommand", "makeSimsDist");
607 /***********************************************************/
608 int TreeGroupCommand::makeSimsShared() {
612 if (subsampleSize == -1) { //user has not set size, set size = smallest samples size
613 subsampleSize = lookup[0]->getNumSeqs();
614 for (int i = 1; i < lookup.size(); i++) {
615 int thisSize = lookup[i]->getNumSeqs();
617 if (thisSize < subsampleSize) { subsampleSize = thisSize; }
622 vector<SharedRAbundVector*> temp;
623 for (int i = 0; i < lookup.size(); i++) {
624 if (lookup[i]->getNumSeqs() < subsampleSize) {
625 m->mothurOut(lookup[i]->getGroup() + " contains " + toString(lookup[i]->getNumSeqs()) + ". Eliminating."); m->mothurOutEndLine();
628 Groups.push_back(lookup[i]->getGroup());
629 temp.push_back(lookup[i]);
633 m->setGroups(Groups);
636 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; }
639 numGroups = lookup.size();
640 lines.resize(processors);
641 for (int i = 0; i < processors; i++) {
642 lines[i].start = int (sqrt(float(i)/float(processors)) * numGroups);
643 lines[i].end = int (sqrt(float(i+1)/float(processors)) * numGroups);
646 set<string> processedLabels;
647 set<string> userLabels = labels;
649 //as long as you are not at the end of the file or done wih the lines you want
650 while((lookup[0] != NULL) && ((allLines == 1) || (userLabels.size() != 0))) {
651 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; }
653 if(allLines == 1 || labels.count(lookup[0]->getLabel()) == 1){
654 m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
657 processedLabels.insert(lookup[0]->getLabel());
658 userLabels.erase(lookup[0]->getLabel());
661 if ((m->anyLabelsToProcess(lookup[0]->getLabel(), userLabels, "") == true) && (processedLabels.count(lastLabel) != 1)) {
662 string saveLabel = lookup[0]->getLabel();
664 for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
665 lookup = input->getSharedRAbundVectors(lastLabel);
667 m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
670 processedLabels.insert(lookup[0]->getLabel());
671 userLabels.erase(lookup[0]->getLabel());
673 //restore real lastlabel to save below
674 lookup[0]->setLabel(saveLabel);
677 lastLabel = lookup[0]->getLabel();
679 //get next line to process
680 for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
681 lookup = input->getSharedRAbundVectors();
684 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; }
686 //output error messages about any remaining user labels
687 set<string>::iterator it;
688 bool needToRun = false;
689 for (it = userLabels.begin(); it != userLabels.end(); it++) {
690 m->mothurOut("Your file does not include the label " + *it);
691 if (processedLabels.count(lastLabel) != 1) {
692 m->mothurOut(". I will use " + lastLabel + "."); m->mothurOutEndLine();
695 m->mothurOut(". Please refer to " + lastLabel + "."); m->mothurOutEndLine();
699 //run last label if you need to
700 if (needToRun == true) {
701 for (int i = 0; i < lookup.size(); i++) { if (lookup[i] != NULL) { delete lookup[i]; } }
702 lookup = input->getSharedRAbundVectors(lastLabel);
704 m->mothurOut(lookup[0]->getLabel()); m->mothurOutEndLine();
706 for (int i = 0; i < lookup.size(); i++) { delete lookup[i]; }
709 for(int i = 0 ; i < treeCalculators.size(); i++) { delete treeCalculators[i]; }
713 catch(exception& e) {
714 m->errorOut(e, "TreeGroupCommand", "makeSimsShared");
719 /***********************************************************/
720 int TreeGroupCommand::process(vector<SharedRAbundVector*> thisLookup) {
722 vector< vector< vector<seqDist> > > calcDistsTotals; //each iter, one for each calc, then each groupCombos dists. this will be used to make .dist files
723 vector< vector<seqDist> > calcDists; calcDists.resize(treeCalculators.size());
725 for (int thisIter = 0; thisIter < iters; thisIter++) {
727 vector<SharedRAbundVector*> thisItersLookup = thisLookup;
731 vector<string> tempLabels; //dont need since we arent printing the sampled sharedRabunds
733 //make copy of lookup so we don't get access violations
734 vector<SharedRAbundVector*> newLookup;
735 for (int k = 0; k < thisItersLookup.size(); k++) {
736 SharedRAbundVector* temp = new SharedRAbundVector();
737 temp->setLabel(thisItersLookup[k]->getLabel());
738 temp->setGroup(thisItersLookup[k]->getGroup());
739 newLookup.push_back(temp);
743 for (int k = 0; k < thisItersLookup[0]->getNumBins(); k++) {
744 if (m->control_pressed) { for (int j = 0; j < newLookup.size(); j++) { delete newLookup[j]; } return 0; }
745 for (int j = 0; j < thisItersLookup.size(); j++) { newLookup[j]->push_back(thisItersLookup[j]->getAbundance(k), thisItersLookup[j]->getGroup()); }
748 tempLabels = sample.getSample(newLookup, subsampleSize);
749 thisItersLookup = newLookup;
753 driver(thisItersLookup, 0, numGroups, calcDists);
756 vector<int> processIDS;
758 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
759 //loop through and create all the processes you want
760 while (process != processors) {
764 processIDS.push_back(pid);
768 driver(thisItersLookup, lines[process].start, lines[process].end, calcDists);
770 string tempdistFileName = m->getRootName(m->getSimpleName(sharedfile)) + toString(getpid()) + ".dist";
772 m->openOutputFile(tempdistFileName, outtemp);
774 for (int i = 0; i < calcDists.size(); i++) {
775 outtemp << calcDists[i].size() << endl;
777 for (int j = 0; j < calcDists[i].size(); j++) {
778 outtemp << calcDists[i][j].seq1 << '\t' << calcDists[i][j].seq2 << '\t' << calcDists[i][j].dist << endl;
785 m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine();
786 for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); }
791 //parent do your part
792 driver(thisItersLookup, lines[0].start, lines[0].end, calcDists);
794 //force parent to wait until all the processes are done
795 for (int i = 0; i < processIDS.size(); i++) {
796 int temp = processIDS[i];
800 for (int i = 0; i < processIDS.size(); i++) {
801 string tempdistFileName = m->getRootName(m->getSimpleName(sharedfile)) + toString(processIDS[i]) + ".dist";
803 m->openInputFile(tempdistFileName, intemp);
805 for (int k = 0; k < calcDists.size(); k++) {
807 intemp >> size; m->gobble(intemp);
809 for (int j = 0; j < size; j++) {
814 intemp >> seq1 >> seq2 >> dist; m->gobble(intemp);
816 seqDist tempDist(seq1, seq2, dist);
817 calcDists[k].push_back(tempDist);
821 m->mothurRemove(tempdistFileName);
824 //////////////////////////////////////////////////////////////////////////////////////////////////////
825 //Windows version shared memory, so be careful when passing variables through the treeSharedData struct.
826 //Above fork() will clone, so memory is separate, but that's not the case with windows,
827 //Taking advantage of shared memory to pass results vectors.
828 //////////////////////////////////////////////////////////////////////////////////////////////////////
830 vector<treeSharedData*> pDataArray;
831 DWORD dwThreadIdArray[processors-1];
832 HANDLE hThreadArray[processors-1];
834 //Create processor worker threads.
835 for( int i=1; i<processors; i++ ){
837 //make copy of lookup so we don't get access violations
838 vector<SharedRAbundVector*> newLookup;
839 for (int k = 0; k < thisItersLookup.size(); k++) {
840 SharedRAbundVector* temp = new SharedRAbundVector();
841 temp->setLabel(thisItersLookup[k]->getLabel());
842 temp->setGroup(thisItersLookup[k]->getGroup());
843 newLookup.push_back(temp);
847 for (int k = 0; k < thisItersLookup[0]->getNumBins(); k++) {
848 if (m->control_pressed) { for (int j = 0; j < newLookup.size(); j++) { delete newLookup[j]; } return 0; }
849 for (int j = 0; j < thisItersLookup.size(); j++) { newLookup[j]->push_back(thisItersLookup[j]->getAbundance(k), thisItersLookup[j]->getGroup()); }
852 // Allocate memory for thread data.
853 treeSharedData* tempSum = new treeSharedData(m, lines[i].start, lines[i].end, Estimators, newLookup);
854 pDataArray.push_back(tempSum);
855 processIDS.push_back(i);
857 hThreadArray[i-1] = CreateThread(NULL, 0, MyTreeSharedThreadFunction, pDataArray[i-1], 0, &dwThreadIdArray[i-1]);
860 //parent do your part
861 driver(thisItersLookup, lines[0].start, lines[0].end, calcDists);
863 //Wait until all threads have terminated.
864 WaitForMultipleObjects(processors-1, hThreadArray, TRUE, INFINITE);
866 //Close all thread handles and free memory allocations.
867 for(int i=0; i < pDataArray.size(); i++){
868 for (int j = 0; j < pDataArray[i]->thisLookup.size(); j++) { delete pDataArray[i]->thisLookup[j]; }
870 for (int k = 0; k < calcDists.size(); k++) {
871 int size = pDataArray[i]->calcDists[k].size();
872 for (int j = 0; j < size; j++) { calcDists[k].push_back(pDataArray[i]->calcDists[k][j]); }
875 CloseHandle(hThreadArray[i]);
876 delete pDataArray[i];
882 calcDistsTotals.push_back(calcDists);
887 for (int i = 0; i < thisItersLookup.size(); i++) { delete thisItersLookup[i]; }
888 thisItersLookup.clear();
889 for (int i = 0; i < calcDists.size(); i++) { calcDists[i].clear(); }
894 //we need to find the average distance and standard deviation for each groups distance
896 vector< vector<seqDist> > calcAverages; calcAverages.resize(treeCalculators.size());
897 for (int i = 0; i < calcAverages.size(); i++) { //initialize sums to zero.
898 calcAverages[i].resize(calcDistsTotals[0][i].size());
900 for (int j = 0; j < calcAverages[i].size(); j++) {
901 calcAverages[i][j].seq1 = calcDists[i][j].seq1;
902 calcAverages[i][j].seq2 = calcDists[i][j].seq2;
903 calcAverages[i][j].dist = 0.0;
907 for (int thisIter = 0; thisIter < iters; thisIter++) { //sum all groups dists for each calculator
908 for (int i = 0; i < calcAverages.size(); i++) { //initialize sums to zero.
909 for (int j = 0; j < calcAverages[i].size(); j++) {
910 calcAverages[i][j].dist += calcDistsTotals[thisIter][i][j].dist;
915 for (int i = 0; i < calcAverages.size(); i++) { //finds average.
916 for (int j = 0; j < calcAverages[i].size(); j++) {
917 calcAverages[i][j].dist /= (float) iters;
921 //create average tree for each calc
922 for (int i = 0; i < calcDists.size(); i++) {
923 vector< vector<double> > matrix; //square matrix to represent the distance
924 matrix.resize(thisLookup.size());
925 for (int k = 0; k < thisLookup.size(); k++) { matrix[k].resize(thisLookup.size(), 0.0); }
927 for (int j = 0; j < calcAverages[i].size(); j++) {
928 int row = calcAverages[i][j].seq1;
929 int column = calcAverages[i][j].seq2;
930 float dist = calcAverages[i][j].dist;
932 matrix[row][column] = dist;
933 matrix[column][row] = dist;
936 //create a new filename
937 string outputFile = outputDir + m->getRootName(m->getSimpleName(inputfile)) + treeCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + ".ave." + getOutputFileNameTag("tree");
938 outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile);
940 //creates tree from similarity matrix and write out file
941 Tree* newTree = createTree(matrix);
942 if (newTree != NULL) { writeTree(outputFile, newTree); }
945 //create all trees for each calc and find their consensus tree
946 for (int i = 0; i < calcDists.size(); i++) {
947 if (m->control_pressed) { break; }
949 //create a new filename
950 string outputFile = outputDir + m->getRootName(m->getSimpleName(inputfile)) + treeCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + ".all." + getOutputFileNameTag("tree");
951 outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile);
954 m->openOutputFile(outputFile, outAll);
957 for (int myIter = 0; myIter < iters; myIter++) {
959 if(m->control_pressed) { break; }
962 vector< vector<double> > matrix; //square matrix to represent the distance
963 matrix.resize(thisLookup.size());
964 for (int k = 0; k < thisLookup.size(); k++) { matrix[k].resize(thisLookup.size(), 0.0); }
966 for (int j = 0; j < calcDistsTotals[myIter][i].size(); j++) {
967 int row = calcDistsTotals[myIter][i][j].seq1;
968 int column = calcDistsTotals[myIter][i][j].seq2;
969 double dist = calcDistsTotals[myIter][i][j].dist;
971 matrix[row][column] = dist;
972 matrix[column][row] = dist;
975 //creates tree from similarity matrix and write out file
976 Tree* newTree = createTree(matrix);
977 if (newTree != NULL) {
978 newTree->print(outAll);
979 trees.push_back(newTree);
983 if (m->control_pressed) { for (int k = 0; k < trees.size(); k++) { delete trees[k]; } }
986 //clear old tree names if any
987 m->Treenames.clear(); m->Treenames = m->getGroups(); //may have changed if subsample eliminated groups
988 Tree* conTree = consensus.getTree(trees);
990 //create a new filename
991 string conFile = outputDir + m->getRootName(m->getSimpleName(inputfile)) + treeCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + ".cons." + getOutputFileNameTag("tree");
992 outputNames.push_back(conFile); outputTypes["tree"].push_back(conFile);
994 m->openOutputFile(conFile, outTree);
996 if (conTree != NULL) { conTree->print(outTree, "boot"); delete conTree; }
1001 for (int i = 0; i < calcDists.size(); i++) {
1002 if (m->control_pressed) { break; }
1005 vector< vector<double> > matrix; //square matrix to represent the distance
1006 matrix.resize(thisLookup.size());
1007 for (int k = 0; k < thisLookup.size(); k++) { matrix[k].resize(thisLookup.size(), 0.0); }
1009 for (int j = 0; j < calcDists[i].size(); j++) {
1010 int row = calcDists[i][j].seq1;
1011 int column = calcDists[i][j].seq2;
1012 double dist = calcDists[i][j].dist;
1014 matrix[row][column] = dist;
1015 matrix[column][row] = dist;
1018 //create a new filename
1019 string outputFile = outputDir + m->getRootName(m->getSimpleName(inputfile)) + treeCalculators[i]->getName() + "." + thisLookup[0]->getLabel() + "." + getOutputFileNameTag("tree");
1020 outputNames.push_back(outputFile); outputTypes["tree"].push_back(outputFile);
1022 //creates tree from similarity matrix and write out file
1023 Tree* newTree = createTree(matrix);
1024 if (newTree != NULL) { writeTree(outputFile, newTree); delete newTree; }
1030 catch(exception& e) {
1031 m->errorOut(e, "TreeGroupCommand", "process");
1035 /**************************************************************************************************/
1036 int TreeGroupCommand::driver(vector<SharedRAbundVector*> thisLookup, int start, int end, vector< vector<seqDist> >& calcDists) {
1038 vector<SharedRAbundVector*> subset;
1039 for (int k = start; k < end; k++) { // pass cdd each set of groups to compare
1041 for (int l = 0; l < k; l++) {
1043 if (k != l) { //we dont need to similiarity of a groups to itself
1044 subset.clear(); //clear out old pair of sharedrabunds
1045 //add new pair of sharedrabunds
1046 subset.push_back(thisLookup[k]); subset.push_back(thisLookup[l]);
1048 for(int i=0;i<treeCalculators.size();i++) {
1050 //if this calc needs all groups to calculate the pair load all groups
1051 if (treeCalculators[i]->getNeedsAll()) {
1052 //load subset with rest of lookup for those calcs that need everyone to calc for a pair
1053 for (int w = 0; w < thisLookup.size(); w++) {
1054 if ((w != k) && (w != l)) { subset.push_back(thisLookup[w]); }
1058 vector<double> tempdata = treeCalculators[i]->getValues(subset); //saves the calculator outputs
1060 if (m->control_pressed) { return 1; }
1062 seqDist temp(l, k, -(tempdata[0]-1.0));
1063 calcDists[i].push_back(temp);
1071 catch(exception& e) {
1072 m->errorOut(e, "TreeGroupCommand", "driver");
1076 /***********************************************************/