5 * Created by Sarah Westcott on 1/26/09.
6 * Copyright 2009 Schloss Lab UMASS Amherst. All rights reserved.
10 #include "parsimonycommand.h"
12 //**********************************************************************************************************************
13 vector<string> ParsimonyCommand::setParameters(){
15 CommandParameter ptree("tree", "InputTypes", "", "", "none", "none", "none",false,true); parameters.push_back(ptree);
16 CommandParameter pgroup("group", "InputTypes", "", "", "none", "none", "none",false,false); parameters.push_back(pgroup);
17 CommandParameter pname("name", "InputTypes", "", "", "none", "none", "none",false,false); parameters.push_back(pname);
18 CommandParameter pgroups("groups", "String", "", "", "", "", "",false,false); parameters.push_back(pgroups);
19 CommandParameter prandom("random", "String", "", "", "", "", "",false,false); parameters.push_back(prandom);
20 CommandParameter piters("iters", "Number", "", "1000", "", "", "",false,false); parameters.push_back(piters);
21 CommandParameter pprocessors("processors", "Number", "", "1", "", "", "",false,false); parameters.push_back(pprocessors);
22 CommandParameter pinputdir("inputdir", "String", "", "", "", "", "",false,false); parameters.push_back(pinputdir);
23 CommandParameter poutputdir("outputdir", "String", "", "", "", "", "",false,false); parameters.push_back(poutputdir);
25 vector<string> myArray;
26 for (int i = 0; i < parameters.size(); i++) { myArray.push_back(parameters[i].name); }
30 m->errorOut(e, "ParsimonyCommand", "setParameters");
34 //**********************************************************************************************************************
35 string ParsimonyCommand::getHelpString(){
37 string helpString = "";
38 helpString += "The parsimony command parameters are tree, group, name, random, groups, processors and iters. tree parameter is required unless you have valid current tree file or are using random.\n";
39 helpString += "The groups parameter allows you to specify which of the groups in your groupfile you would like analyzed. You must enter at least 1 valid group.\n";
40 helpString += "The group names are separated by dashes. The iters parameter allows you to specify how many random trees you would like compared to your tree.\n";
41 helpString += "The parsimony command should be in the following format: parsimony(random=yourOutputFilename, groups=yourGroups, iters=yourIters).\n";
42 helpString += "The processors parameter allows you to specify the number of processors to use. The default is 1.\n";
43 helpString += "Example parsimony(random=out, iters=500).\n";
44 helpString += "The default value for random is "" (meaning you want to use the trees in your inputfile, randomtree=out means you just want the random distribution of trees outputted to out.rd_parsimony),\n";
45 helpString += "and iters is 1000. The parsimony command output two files: .parsimony and .psummary their descriptions are in the manual.\n";
46 helpString += "Note: No spaces between parameter labels (i.e. random), '=' and parameters (i.e.yourOutputFilename).\n";
50 m->errorOut(e, "ParsimonyCommand", "getHelpString");
55 //**********************************************************************************************************************
56 ParsimonyCommand::ParsimonyCommand(){
58 abort = true; calledHelp = true;
60 vector<string> tempOutNames;
61 outputTypes["parsimony"] = tempOutNames;
62 outputTypes["psummary"] = tempOutNames;
65 m->errorOut(e, "ParsimonyCommand", "ParsimonyCommand");
69 /***********************************************************/
70 ParsimonyCommand::ParsimonyCommand(string option) {
72 abort = false; calledHelp = false;
75 //allow user to run help
76 if(option == "help") { help(); abort = true; calledHelp = true; }
77 else if(option == "citation") { citation(); abort = true; calledHelp = true;}
80 vector<string> myArray = setParameters();
82 OptionParser parser(option);
83 map<string, string> parameters = parser.getParameters();
84 map<string,string>::iterator it;
86 ValidParameters validParameter;
88 //check to make sure all parameters are valid for command
89 for (it = parameters.begin(); it != parameters.end(); it++) {
90 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) { abort = true; }
93 //initialize outputTypes
94 vector<string> tempOutNames;
95 outputTypes["parsimony"] = tempOutNames;
96 outputTypes["psummary"] = tempOutNames;
98 //if the user changes the input directory command factory will send this info to us in the output parameter
99 string inputDir = validParameter.validFile(parameters, "inputdir", false);
100 if (inputDir == "not found"){ inputDir = ""; }
103 it = parameters.find("tree");
104 //user has given a template file
105 if(it != parameters.end()){
106 path = m->hasPath(it->second);
107 //if the user has not given a path then, add inputdir. else leave path alone.
108 if (path == "") { parameters["tree"] = inputDir + it->second; }
111 it = parameters.find("group");
112 //user has given a template file
113 if(it != parameters.end()){
114 path = m->hasPath(it->second);
115 //if the user has not given a path then, add inputdir. else leave path alone.
116 if (path == "") { parameters["group"] = inputDir + it->second; }
119 it = parameters.find("name");
120 //user has given a template file
121 if(it != parameters.end()){
122 path = m->hasPath(it->second);
123 //if the user has not given a path then, add inputdir. else leave path alone.
124 if (path == "") { parameters["name"] = inputDir + it->second; }
131 m->Treenames.clear();
134 outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){ outputDir = ""; }
136 randomtree = validParameter.validFile(parameters, "random", false); if (randomtree == "not found") { randomtree = ""; }
138 //are you trying to use parsimony without reading a tree or saying you want random distribution
139 if (randomtree == "") {
140 //check for required parameters
141 treefile = validParameter.validFile(parameters, "tree", true);
142 if (treefile == "not open") { abort = true; }
143 else if (treefile == "not found") { //if there is a current design file, use it
144 treefile = m->getTreeFile();
145 if (treefile != "") { m->mothurOut("Using " + treefile + " as input file for the tree parameter."); m->mothurOutEndLine(); }
146 else { m->mothurOut("You have no current tree file and the tree parameter is required."); m->mothurOutEndLine(); abort = true; }
147 }else { m->setTreeFile(treefile); }
149 //check for required parameters
150 groupfile = validParameter.validFile(parameters, "group", true);
151 if (groupfile == "not open") { abort = true; }
152 else if (groupfile == "not found") { groupfile = ""; }
153 else { m->setGroupFile(groupfile); }
155 namefile = validParameter.validFile(parameters, "name", true);
156 if (namefile == "not open") { abort = true; }
157 else if (namefile == "not found") { namefile = ""; }
158 else { m->setNameFile(namefile); }
161 //if the user changes the output directory command factory will send this info to us in the output parameter
162 string outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){ outputDir = ""; if (randomtree == "") { outputDir += m->hasPath(treefile); } }
164 //check for optional parameter and set defaults
165 // ...at some point should added some additional type checking...
166 groups = validParameter.validFile(parameters, "groups", false);
167 if (groups == "not found") { groups = ""; m->clearGroups(); }
169 m->splitAtDash(groups, Groups);
170 m->setGroups(Groups);
173 itersString = validParameter.validFile(parameters, "iters", false); if (itersString == "not found") { itersString = "1000"; }
174 convert(itersString, iters);
176 string temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); }
177 m->setProcessors(temp);
178 convert(temp, processors);
183 catch(exception& e) {
184 m->errorOut(e, "ParsimonyCommand", "ParsimonyCommand");
188 /***********************************************************/
189 int ParsimonyCommand::execute() {
192 if (abort == true) { if (calledHelp) { return 0; } return 2; }
195 //randomtree will tell us if user had their own treefile or if they just want the random distribution
196 //user has entered their own tree
197 if (randomtree == "") {
199 m->setTreeFile(treefile);
201 if (groupfile != "") {
202 //read in group map info.
203 tmap = new TreeMap(groupfile);
205 }else{ //fake out by putting everyone in one group
206 Tree* tree = new Tree(treefile); delete tree; //extracts names from tree to make faked out groupmap
207 tmap = new TreeMap();
209 for (int i = 0; i < m->Treenames.size(); i++) { tmap->addSeq(m->Treenames[i], "Group1"); }
212 if (namefile != "") { readNamesFile(); }
214 read = new ReadNewickTree(treefile);
215 int readOk = read->read(tmap);
217 if (readOk != 0) { m->mothurOut("Read Terminated."); m->mothurOutEndLine(); delete tmap; delete read; return 0; }
219 read->AssembleTrees();
220 T = read->getTrees();
223 //make sure all files match
224 //if you provide a namefile we will use the numNames in the namefile as long as the number of unique match the tree names size.
226 if (namefile != "") {
227 if (numUniquesInName == m->Treenames.size()) { numNamesInTree = nameMap.size(); }
228 else { numNamesInTree = m->Treenames.size(); }
229 }else { numNamesInTree = m->Treenames.size(); }
232 //output any names that are in group file but not in tree
233 if (numNamesInTree < tmap->getNumSeqs()) {
234 for (int i = 0; i < tmap->namesOfSeqs.size(); i++) {
235 //is that name in the tree?
237 for (int j = 0; j < m->Treenames.size(); j++) {
238 if (tmap->namesOfSeqs[i] == m->Treenames[j]) { break; } //found it
242 if (m->control_pressed) {
243 delete tmap; for (int i = 0; i < T.size(); i++) { delete T[i]; }
244 for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } outputTypes.clear();
249 //then you did not find it so report it
250 if (count == m->Treenames.size()) {
251 //if it is in your namefile then don't remove
252 map<string, string>::iterator it = nameMap.find(tmap->namesOfSeqs[i]);
254 if (it == nameMap.end()) {
255 m->mothurOut(tmap->namesOfSeqs[i] + " is in your groupfile and not in your tree. It will be disregarded."); m->mothurOutEndLine();
256 tmap->removeSeq(tmap->namesOfSeqs[i]);
257 i--; //need this because removeSeq removes name from namesOfSeqs
263 if(outputDir == "") { outputDir += m->hasPath(treefile); }
264 output = new ColumnFile(outputDir + m->getSimpleName(treefile) + ".parsimony", itersString);
265 outputNames.push_back(outputDir + m->getSimpleName(treefile) + ".parsimony");
266 outputTypes["parsimony"].push_back(outputDir + m->getSimpleName(treefile) + ".parsimony");
268 sumFile = outputDir + m->getSimpleName(treefile) + ".psummary";
269 m->openOutputFile(sumFile, outSum);
270 outputNames.push_back(sumFile);
271 outputTypes["psummary"].push_back(sumFile);
272 }else { //user wants random distribution
275 if(outputDir == "") { outputDir += m->hasPath(randomtree); }
276 output = new ColumnFile(outputDir+ m->getSimpleName(randomtree), itersString);
277 outputNames.push_back(outputDir+ m->getSimpleName(randomtree));
278 outputTypes["parsimony"].push_back(outputDir+ m->getSimpleName(randomtree));
281 //set users groups to analyze
282 util = new SharedUtil();
283 vector<string> mGroups = m->getGroups();
284 vector<string> tGroups = tmap->getNamesOfGroups();
285 util->setGroups(mGroups, tGroups, allGroups, numGroups, "parsimony"); //sets the groups the user wants to analyze
286 util->getCombos(groupComb, mGroups, numComp);
287 m->setGroups(mGroups);
290 if (numGroups == 1) { numComp++; groupComb.push_back(allGroups); }
292 pars = new Parsimony(tmap);
296 reading = new Progress("Comparing to random:", iters);
298 if (m->control_pressed) {
299 delete reading; delete pars; delete output;
300 delete tmap; for (int i = 0; i < T.size(); i++) { delete T[i]; }
301 if (randomtree == "") { outSum.close(); }
302 for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } outputTypes.clear();
308 //get pscore for users tree
309 userData.resize(numComp,0); //data = AB, AC, BC, ABC.
310 randomData.resize(numComp,0); //data = AB, AC, BC, ABC.
311 rscoreFreq.resize(numComp);
312 uscoreFreq.resize(numComp);
313 rCumul.resize(numComp);
314 uCumul.resize(numComp);
315 userTreeScores.resize(numComp);
316 UScoreSig.resize(numComp);
318 if (randomtree == "") {
319 //get pscores for users trees
320 for (int i = 0; i < T.size(); i++) {
321 userData = pars->getValues(T[i], processors, outputDir); //data = AB, AC, BC, ABC.
323 if (m->control_pressed) {
324 delete reading; delete pars; delete output;
325 delete tmap; for (int i = 0; i < T.size(); i++) { delete T[i]; }
326 if (randomtree == "") { outSum.close(); }
327 for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } outputTypes.clear();
333 //output scores for each combination
334 for(int k = 0; k < numComp; k++) {
337 map<int,double>::iterator it = uscoreFreq[k].find(userData[k]);
338 if (it == uscoreFreq[k].end()) {//new score
339 uscoreFreq[k][userData[k]] = 1;
340 }else{ uscoreFreq[k][userData[k]]++; }
342 //add users score to valid scores
343 validScores[userData[k]] = userData[k];
345 //save score for summary file
346 userTreeScores[k].push_back(userData[k]);
350 //get pscores for random trees
351 for (int j = 0; j < iters; j++) {
353 //create new tree with same num nodes and leaves as users
354 randT = new Tree(tmap);
356 //create random relationships between nodes
357 randT->assembleRandomTree();
359 //get pscore of random tree
360 randomData = pars->getValues(randT, processors, outputDir);
362 if (m->control_pressed) {
363 delete reading; delete pars; delete output; delete randT;
364 if (randomtree == "") { outSum.close(); }
365 for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } outputTypes.clear();
366 delete tmap; for (int i = 0; i < T.size(); i++) { delete T[i]; }
371 for(int r = 0; r < numComp; r++) {
372 //add trees pscore to map of scores
373 map<int,double>::iterator it = rscoreFreq[r].find(randomData[r]);
374 if (it != rscoreFreq[r].end()) {//already have that score
375 rscoreFreq[r][randomData[r]]++;
376 }else{//first time we have seen this score
377 rscoreFreq[r][randomData[r]] = 1;
380 //add randoms score to validscores
381 validScores[randomData[r]] = randomData[r];
384 //update progress bar
391 //get pscores for random trees
392 for (int j = 0; j < iters; j++) {
394 //create new tree with same num nodes and leaves as users
395 randT = new Tree(tmap);
396 //create random relationships between nodes
398 randT->assembleRandomTree();
400 if (m->control_pressed) {
401 delete reading; delete pars; delete output; delete randT;
403 for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } outputTypes.clear();
409 //get pscore of random tree
410 randomData = pars->getValues(randT, processors, outputDir);
412 if (m->control_pressed) {
413 delete reading; delete pars; delete output; delete randT;
415 for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } outputTypes.clear();
420 for(int r = 0; r < numComp; r++) {
421 //add trees pscore to map of scores
422 map<int,double>::iterator it = rscoreFreq[r].find(randomData[r]);
423 if (it != rscoreFreq[r].end()) {//already have that score
424 rscoreFreq[r][randomData[r]]++;
425 }else{//first time we have seen this score
426 rscoreFreq[r][randomData[r]] = 1;
429 //add randoms score to validscores
430 validScores[randomData[r]] = randomData[r];
433 //update progress bar
440 for(int a = 0; a < numComp; a++) {
441 float rcumul = 0.0000;
442 float ucumul = 0.0000;
443 //this loop fills the cumulative maps and put 0.0000 in the score freq map to make it easier to print.
444 for (map<int,double>::iterator it = validScores.begin(); it != validScores.end(); it++) {
445 if (randomtree == "") {
446 map<int,double>::iterator it2 = uscoreFreq[a].find(it->first);
447 //user data has that score
448 if (it2 != uscoreFreq[a].end()) { uscoreFreq[a][it->first] /= T.size(); ucumul+= it2->second; }
449 else { uscoreFreq[a][it->first] = 0.0000; } //no user trees with that score
451 uCumul[a][it->first] = ucumul;
454 //make rscoreFreq map and rCumul
455 map<int,double>::iterator it2 = rscoreFreq[a].find(it->first);
456 //get percentage of random trees with that info
457 if (it2 != rscoreFreq[a].end()) { rscoreFreq[a][it->first] /= iters; rcumul+= it2->second; }
458 else { rscoreFreq[a][it->first] = 0.0000; } //no random trees with that score
459 rCumul[a][it->first] = rcumul;
462 //find the signifigance of each user trees score when compared to the random trees and save for printing the summary file
463 for (int h = 0; h < userTreeScores[a].size(); h++) {
464 UScoreSig[a].push_back(rCumul[a][userTreeScores[a][h]]);
468 if (m->control_pressed) {
469 delete reading; delete pars; delete output;
470 delete tmap; for (int i = 0; i < T.size(); i++) { delete T[i]; }
471 if (randomtree == "") { outSum.close(); }
472 for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } outputTypes.clear();
477 //finish progress bar
482 printParsimonyFile();
483 if (randomtree == "") { printUSummaryFile(); }
485 //reset groups parameter
488 delete pars; delete output;
489 delete tmap; for (int i = 0; i < T.size(); i++) { delete T[i]; }
491 if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { m->mothurRemove(outputNames[i]); } outputTypes.clear(); return 0;}
493 m->mothurOutEndLine();
494 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
495 for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
496 m->mothurOutEndLine();
502 catch(exception& e) {
503 m->errorOut(e, "ParsimonyCommand", "execute");
508 /***********************************************************/
509 void ParsimonyCommand::printParsimonyFile() {
514 if (randomtree == "") {
515 tags.push_back("Score"); tags.push_back("UserFreq"); tags.push_back("UserCumul"); tags.push_back("RandFreq"); tags.push_back("RandCumul");
517 tags.push_back("Score"); tags.push_back("RandFreq"); tags.push_back("RandCumul");
520 for(int a = 0; a < numComp; a++) {
521 output->initFile(groupComb[a], tags);
523 for (map<int,double>::iterator it = validScores.begin(); it != validScores.end(); it++) {
524 if (randomtree == "") {
525 data.push_back(it->first); data.push_back(uscoreFreq[a][it->first]); data.push_back(uCumul[a][it->first]); data.push_back(rscoreFreq[a][it->first]); data.push_back(rCumul[a][it->first]);
527 data.push_back(it->first); data.push_back(rscoreFreq[a][it->first]); data.push_back(rCumul[a][it->first]);
529 output->output(data);
535 catch(exception& e) {
536 m->errorOut(e, "ParsimonyCommand", "printParsimonyFile");
540 /***********************************************************/
541 int ParsimonyCommand::printUSummaryFile() {
544 outSum << "Tree#" << '\t' << "Groups" << '\t' << "ParsScore" << '\t' << "ParsSig" << endl;
545 m->mothurOut("Tree#\tGroups\tParsScore\tParsSig"); m->mothurOutEndLine();
548 outSum.setf(ios::fixed, ios::floatfield); outSum.setf(ios::showpoint);
552 for (int i = 0; i< T.size(); i++) {
553 for(int a = 0; a < numComp; a++) {
554 if (m->control_pressed) { outSum.close(); return 0; }
555 if (UScoreSig[a][i] > (1/(float)iters)) {
556 outSum << setprecision(6) << i+1 << '\t' << groupComb[a] << '\t' << userTreeScores[a][i] << setprecision(itersString.length()) << '\t' << UScoreSig[a][i] << endl;
557 cout << setprecision(6) << i+1 << '\t' << groupComb[a] << '\t' << userTreeScores[a][i] << setprecision(itersString.length()) << '\t' << UScoreSig[a][i] << endl;
558 m->mothurOutJustToLog(toString(i+1) + "\t" + groupComb[a] + "\t" + toString(userTreeScores[a][i]) + "\t" + toString(UScoreSig[a][i])); m->mothurOutEndLine();
560 outSum << setprecision(6) << i+1 << '\t' << groupComb[a] << '\t' << userTreeScores[a][i] << setprecision(itersString.length()) << '\t' << "<" << (1/float(iters)) << endl;
561 cout << setprecision(6) << i+1 << '\t' << groupComb[a] << '\t' << userTreeScores[a][i] << setprecision(itersString.length()) << '\t' << "<" << (1/float(iters)) << endl;
562 m->mothurOutJustToLog(toString(i+1) + "\t" + groupComb[a] + "\t" + toString(userTreeScores[a][i]) + "\t" + toString((1/float(iters)))); m->mothurOutEndLine();
570 catch(exception& e) {
571 m->errorOut(e, "ParsimonyCommand", "printUSummaryFile");
576 /***********************************************************/
577 void ParsimonyCommand::getUserInput() {
581 tmap = new TreeMap();
583 m->mothurOut("Please enter the number of groups you would like to analyze: ");
585 m->mothurOutJustToLog(toString(numGroups)); m->mothurOutEndLine();
589 numEachGroup.resize(numGroups, 0);
592 for (int i = 1; i <= numGroups; i++) {
593 m->mothurOut("Please enter the number of sequences in group " + toString(i) + ": ");
595 m->mothurOutJustToLog(toString(num)); m->mothurOutEndLine();
597 //set tmaps seqsPerGroup
598 tmap->seqsPerGroup[toString(i)] = num;
599 tmap->addGroup(toString(i));
601 //set tmaps namesOfSeqs
602 for (int j = 0; j < num; j++) {
603 tmap->namesOfSeqs.push_back(toString(count));
604 tmap->treemap[toString(count)].groupname = toString(i);
609 //clears buffer so next command doesn't have error
613 m->Treenames = tmap->namesOfSeqs;
616 catch(exception& e) {
617 m->errorOut(e, "ParsimonyCommand", "getUserInput");
621 /*****************************************************************/
622 int ParsimonyCommand::readNamesFile() {
625 numUniquesInName = 0;
628 m->openInputFile(namefile, in);
630 string first, second;
631 map<string, string>::iterator itNames;
634 in >> first >> second; m->gobble(in);
638 itNames = m->names.find(first);
639 if (itNames == m->names.end()) {
640 m->names[first] = second;
642 //we need a list of names in your namefile to use above when removing extra seqs above so we don't remove them
643 vector<string> dupNames;
644 m->splitAtComma(second, dupNames);
646 for (int i = 0; i < dupNames.size(); i++) {
647 nameMap[dupNames[i]] = dupNames[i];
648 if ((groupfile == "") && (i != 0)) { tmap->addSeq(dupNames[i], "Group1"); }
650 }else { m->mothurOut(first + " has already been seen in namefile, disregarding names file."); m->mothurOutEndLine(); in.close(); m->names.clear(); namefile = ""; return 1; }
656 catch(exception& e) {
657 m->errorOut(e, "ParsimonyCommand", "readNamesFile");
661 /***********************************************************/