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; }
130 m->namesOfGroups.clear();
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; }
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 = ""; }
154 namefile = validParameter.validFile(parameters, "name", true);
155 if (namefile == "not open") { abort = true; }
156 else if (namefile == "not found") { namefile = ""; }
159 //if the user changes the output directory command factory will send this info to us in the output parameter
160 string outputDir = validParameter.validFile(parameters, "outputdir", false); if (outputDir == "not found"){ outputDir = ""; if (randomtree == "") { outputDir += m->hasPath(treefile); } }
162 //check for optional parameter and set defaults
163 // ...at some point should added some additional type checking...
164 groups = validParameter.validFile(parameters, "groups", false);
165 if (groups == "not found") { groups = ""; m->Groups.clear(); }
167 m->splitAtDash(groups, Groups);
171 itersString = validParameter.validFile(parameters, "iters", false); if (itersString == "not found") { itersString = "1000"; }
172 convert(itersString, iters);
174 string temp = validParameter.validFile(parameters, "processors", false); if (temp == "not found"){ temp = m->getProcessors(); }
175 m->setProcessors(temp);
176 convert(temp, processors);
181 catch(exception& e) {
182 m->errorOut(e, "ParsimonyCommand", "ParsimonyCommand");
186 /***********************************************************/
187 int ParsimonyCommand::execute() {
190 if (abort == true) { if (calledHelp) { return 0; } return 2; }
193 //randomtree will tell us if user had their own treefile or if they just want the random distribution
194 //user has entered their own tree
195 if (randomtree == "") {
197 m->setTreeFile(treefile);
199 if (groupfile != "") {
200 //read in group map info.
201 tmap = new TreeMap(groupfile);
203 }else{ //fake out by putting everyone in one group
204 Tree* tree = new Tree(treefile); delete tree; //extracts names from tree to make faked out groupmap
205 tmap = new TreeMap();
207 for (int i = 0; i < m->Treenames.size(); i++) { tmap->addSeq(m->Treenames[i], "Group1"); }
210 if (namefile != "") { readNamesFile(); }
212 read = new ReadNewickTree(treefile);
213 int readOk = read->read(tmap);
215 if (readOk != 0) { m->mothurOut("Read Terminated."); m->mothurOutEndLine(); delete tmap; delete read; return 0; }
217 read->AssembleTrees();
218 T = read->getTrees();
221 //make sure all files match
222 //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.
224 if (namefile != "") {
225 if (numUniquesInName == m->Treenames.size()) { numNamesInTree = nameMap.size(); }
226 else { numNamesInTree = m->Treenames.size(); }
227 }else { numNamesInTree = m->Treenames.size(); }
230 //output any names that are in group file but not in tree
231 if (numNamesInTree < tmap->getNumSeqs()) {
232 for (int i = 0; i < tmap->namesOfSeqs.size(); i++) {
233 //is that name in the tree?
235 for (int j = 0; j < m->Treenames.size(); j++) {
236 if (tmap->namesOfSeqs[i] == m->Treenames[j]) { break; } //found it
240 if (m->control_pressed) {
241 delete tmap; for (int i = 0; i < T.size(); i++) { delete T[i]; }
242 for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } outputTypes.clear();
247 //then you did not find it so report it
248 if (count == m->Treenames.size()) {
249 //if it is in your namefile then don't remove
250 map<string, string>::iterator it = nameMap.find(tmap->namesOfSeqs[i]);
252 if (it == nameMap.end()) {
253 m->mothurOut(tmap->namesOfSeqs[i] + " is in your groupfile and not in your tree. It will be disregarded."); m->mothurOutEndLine();
254 tmap->removeSeq(tmap->namesOfSeqs[i]);
255 i--; //need this because removeSeq removes name from namesOfSeqs
261 if(outputDir == "") { outputDir += m->hasPath(treefile); }
262 output = new ColumnFile(outputDir + m->getSimpleName(treefile) + ".parsimony", itersString);
263 outputNames.push_back(outputDir + m->getSimpleName(treefile) + ".parsimony");
264 outputTypes["parsimony"].push_back(outputDir + m->getSimpleName(treefile) + ".parsimony");
266 sumFile = outputDir + m->getSimpleName(treefile) + ".psummary";
267 m->openOutputFile(sumFile, outSum);
268 outputNames.push_back(sumFile);
269 outputTypes["psummary"].push_back(sumFile);
270 }else { //user wants random distribution
273 if(outputDir == "") { outputDir += m->hasPath(randomtree); }
274 output = new ColumnFile(outputDir+ m->getSimpleName(randomtree), itersString);
275 outputNames.push_back(outputDir+ m->getSimpleName(randomtree));
276 outputTypes["parsimony"].push_back(outputDir+ m->getSimpleName(randomtree));
279 //set users groups to analyze
280 util = new SharedUtil();
281 util->setGroups(m->Groups, tmap->namesOfGroups, allGroups, numGroups, "parsimony"); //sets the groups the user wants to analyze
282 util->getCombos(groupComb, m->Groups, numComp);
285 if (numGroups == 1) { numComp++; groupComb.push_back(allGroups); }
287 pars = new Parsimony(tmap);
291 reading = new Progress("Comparing to random:", iters);
293 if (m->control_pressed) {
294 delete reading; delete pars; delete output;
295 delete tmap; for (int i = 0; i < T.size(); i++) { delete T[i]; }
296 if (randomtree == "") { outSum.close(); }
297 for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } outputTypes.clear();
303 //get pscore for users tree
304 userData.resize(numComp,0); //data = AB, AC, BC, ABC.
305 randomData.resize(numComp,0); //data = AB, AC, BC, ABC.
306 rscoreFreq.resize(numComp);
307 uscoreFreq.resize(numComp);
308 rCumul.resize(numComp);
309 uCumul.resize(numComp);
310 userTreeScores.resize(numComp);
311 UScoreSig.resize(numComp);
313 if (randomtree == "") {
314 //get pscores for users trees
315 for (int i = 0; i < T.size(); i++) {
316 userData = pars->getValues(T[i], processors, outputDir); //data = AB, AC, BC, ABC.
318 if (m->control_pressed) {
319 delete reading; delete pars; delete output;
320 delete tmap; for (int i = 0; i < T.size(); i++) { delete T[i]; }
321 if (randomtree == "") { outSum.close(); }
322 for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } outputTypes.clear();
328 //output scores for each combination
329 for(int k = 0; k < numComp; k++) {
332 map<int,double>::iterator it = uscoreFreq[k].find(userData[k]);
333 if (it == uscoreFreq[k].end()) {//new score
334 uscoreFreq[k][userData[k]] = 1;
335 }else{ uscoreFreq[k][userData[k]]++; }
337 //add users score to valid scores
338 validScores[userData[k]] = userData[k];
340 //save score for summary file
341 userTreeScores[k].push_back(userData[k]);
345 //get pscores for random trees
346 for (int j = 0; j < iters; j++) {
348 //create new tree with same num nodes and leaves as users
349 randT = new Tree(tmap);
351 //create random relationships between nodes
352 randT->assembleRandomTree();
354 //get pscore of random tree
355 randomData = pars->getValues(randT, processors, outputDir);
357 if (m->control_pressed) {
358 delete reading; delete pars; delete output; delete randT;
359 if (randomtree == "") { outSum.close(); }
360 for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } outputTypes.clear();
361 delete tmap; for (int i = 0; i < T.size(); i++) { delete T[i]; }
366 for(int r = 0; r < numComp; r++) {
367 //add trees pscore to map of scores
368 map<int,double>::iterator it = rscoreFreq[r].find(randomData[r]);
369 if (it != rscoreFreq[r].end()) {//already have that score
370 rscoreFreq[r][randomData[r]]++;
371 }else{//first time we have seen this score
372 rscoreFreq[r][randomData[r]] = 1;
375 //add randoms score to validscores
376 validScores[randomData[r]] = randomData[r];
379 //update progress bar
386 //get pscores for random trees
387 for (int j = 0; j < iters; j++) {
389 //create new tree with same num nodes and leaves as users
390 randT = new Tree(tmap);
391 //create random relationships between nodes
393 randT->assembleRandomTree();
395 if (m->control_pressed) {
396 delete reading; delete pars; delete output; delete randT;
398 for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } outputTypes.clear();
404 //get pscore of random tree
405 randomData = pars->getValues(randT, processors, outputDir);
407 if (m->control_pressed) {
408 delete reading; delete pars; delete output; delete randT;
410 for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } outputTypes.clear();
415 for(int r = 0; r < numComp; r++) {
416 //add trees pscore to map of scores
417 map<int,double>::iterator it = rscoreFreq[r].find(randomData[r]);
418 if (it != rscoreFreq[r].end()) {//already have that score
419 rscoreFreq[r][randomData[r]]++;
420 }else{//first time we have seen this score
421 rscoreFreq[r][randomData[r]] = 1;
424 //add randoms score to validscores
425 validScores[randomData[r]] = randomData[r];
428 //update progress bar
435 for(int a = 0; a < numComp; a++) {
436 float rcumul = 0.0000;
437 float ucumul = 0.0000;
438 //this loop fills the cumulative maps and put 0.0000 in the score freq map to make it easier to print.
439 for (map<int,double>::iterator it = validScores.begin(); it != validScores.end(); it++) {
440 if (randomtree == "") {
441 map<int,double>::iterator it2 = uscoreFreq[a].find(it->first);
442 //user data has that score
443 if (it2 != uscoreFreq[a].end()) { uscoreFreq[a][it->first] /= T.size(); ucumul+= it2->second; }
444 else { uscoreFreq[a][it->first] = 0.0000; } //no user trees with that score
446 uCumul[a][it->first] = ucumul;
449 //make rscoreFreq map and rCumul
450 map<int,double>::iterator it2 = rscoreFreq[a].find(it->first);
451 //get percentage of random trees with that info
452 if (it2 != rscoreFreq[a].end()) { rscoreFreq[a][it->first] /= iters; rcumul+= it2->second; }
453 else { rscoreFreq[a][it->first] = 0.0000; } //no random trees with that score
454 rCumul[a][it->first] = rcumul;
457 //find the signifigance of each user trees score when compared to the random trees and save for printing the summary file
458 for (int h = 0; h < userTreeScores[a].size(); h++) {
459 UScoreSig[a].push_back(rCumul[a][userTreeScores[a][h]]);
463 if (m->control_pressed) {
464 delete reading; delete pars; delete output;
465 delete tmap; for (int i = 0; i < T.size(); i++) { delete T[i]; }
466 if (randomtree == "") { outSum.close(); }
467 for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } outputTypes.clear();
472 //finish progress bar
477 printParsimonyFile();
478 if (randomtree == "") { printUSummaryFile(); }
480 //reset groups parameter
483 delete pars; delete output;
484 delete tmap; for (int i = 0; i < T.size(); i++) { delete T[i]; }
486 if (m->control_pressed) { for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str()); } outputTypes.clear(); return 0;}
488 m->mothurOutEndLine();
489 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
490 for (int i = 0; i < outputNames.size(); i++) { m->mothurOut(outputNames[i]); m->mothurOutEndLine(); }
491 m->mothurOutEndLine();
497 catch(exception& e) {
498 m->errorOut(e, "ParsimonyCommand", "execute");
503 /***********************************************************/
504 void ParsimonyCommand::printParsimonyFile() {
509 if (randomtree == "") {
510 tags.push_back("Score"); tags.push_back("UserFreq"); tags.push_back("UserCumul"); tags.push_back("RandFreq"); tags.push_back("RandCumul");
512 tags.push_back("Score"); tags.push_back("RandFreq"); tags.push_back("RandCumul");
515 for(int a = 0; a < numComp; a++) {
516 output->initFile(groupComb[a], tags);
518 for (map<int,double>::iterator it = validScores.begin(); it != validScores.end(); it++) {
519 if (randomtree == "") {
520 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]);
522 data.push_back(it->first); data.push_back(rscoreFreq[a][it->first]); data.push_back(rCumul[a][it->first]);
524 output->output(data);
530 catch(exception& e) {
531 m->errorOut(e, "ParsimonyCommand", "printParsimonyFile");
535 /***********************************************************/
536 int ParsimonyCommand::printUSummaryFile() {
539 outSum << "Tree#" << '\t' << "Groups" << '\t' << "ParsScore" << '\t' << "ParsSig" << endl;
540 m->mothurOut("Tree#\tGroups\tParsScore\tParsSig"); m->mothurOutEndLine();
543 outSum.setf(ios::fixed, ios::floatfield); outSum.setf(ios::showpoint);
547 for (int i = 0; i< T.size(); i++) {
548 for(int a = 0; a < numComp; a++) {
549 if (m->control_pressed) { outSum.close(); return 0; }
550 if (UScoreSig[a][i] > (1/(float)iters)) {
551 outSum << setprecision(6) << i+1 << '\t' << groupComb[a] << '\t' << userTreeScores[a][i] << setprecision(itersString.length()) << '\t' << UScoreSig[a][i] << endl;
552 cout << setprecision(6) << i+1 << '\t' << groupComb[a] << '\t' << userTreeScores[a][i] << setprecision(itersString.length()) << '\t' << UScoreSig[a][i] << endl;
553 m->mothurOutJustToLog(toString(i+1) + "\t" + groupComb[a] + "\t" + toString(userTreeScores[a][i]) + "\t" + toString(UScoreSig[a][i])); m->mothurOutEndLine();
555 outSum << setprecision(6) << i+1 << '\t' << groupComb[a] << '\t' << userTreeScores[a][i] << setprecision(itersString.length()) << '\t' << "<" << (1/float(iters)) << endl;
556 cout << setprecision(6) << i+1 << '\t' << groupComb[a] << '\t' << userTreeScores[a][i] << setprecision(itersString.length()) << '\t' << "<" << (1/float(iters)) << endl;
557 m->mothurOutJustToLog(toString(i+1) + "\t" + groupComb[a] + "\t" + toString(userTreeScores[a][i]) + "\t" + toString((1/float(iters)))); m->mothurOutEndLine();
565 catch(exception& e) {
566 m->errorOut(e, "ParsimonyCommand", "printUSummaryFile");
571 /***********************************************************/
572 void ParsimonyCommand::getUserInput() {
576 tmap = new TreeMap();
578 m->mothurOut("Please enter the number of groups you would like to analyze: ");
580 m->mothurOutJustToLog(toString(numGroups)); m->mothurOutEndLine();
584 numEachGroup.resize(numGroups, 0);
586 for (int i = 1; i <= numGroups; i++) {
587 m->mothurOut("Please enter the number of sequences in group " + toString(i) + ": ");
589 m->mothurOutJustToLog(toString(num)); m->mothurOutEndLine();
591 //set tmaps seqsPerGroup
592 tmap->seqsPerGroup[toString(i)] = num;
593 tmap->namesOfGroups.push_back(toString(i));
595 //set tmaps namesOfSeqs
596 for (int j = 0; j < num; j++) {
597 tmap->namesOfSeqs.push_back(toString(count));
598 tmap->treemap[toString(count)].groupname = toString(i);
603 //clears buffer so next command doesn't have error
607 m->Treenames = tmap->namesOfSeqs;
610 catch(exception& e) {
611 m->errorOut(e, "ParsimonyCommand", "getUserInput");
615 /*****************************************************************/
616 int ParsimonyCommand::readNamesFile() {
619 numUniquesInName = 0;
622 m->openInputFile(namefile, in);
624 string first, second;
625 map<string, string>::iterator itNames;
628 in >> first >> second; m->gobble(in);
632 itNames = m->names.find(first);
633 if (itNames == m->names.end()) {
634 m->names[first] = second;
636 //we need a list of names in your namefile to use above when removing extra seqs above so we don't remove them
637 vector<string> dupNames;
638 m->splitAtComma(second, dupNames);
640 for (int i = 0; i < dupNames.size(); i++) {
641 nameMap[dupNames[i]] = dupNames[i];
642 if ((groupfile == "") && (i != 0)) { tmap->addSeq(dupNames[i], "Group1"); }
644 }else { m->mothurOut(first + " has already been seen in namefile, disregarding names file."); m->mothurOutEndLine(); in.close(); m->names.clear(); namefile = ""; return 1; }
650 catch(exception& e) {
651 m->errorOut(e, "ParsimonyCommand", "readNamesFile");
655 /***********************************************************/