unifracweightedcommand.cpp

   1 /*
   2  *  unifracweightedcommand.cpp
   3  *  Mothur
   4  *
   5  *  Created by Sarah Westcott on 2/9/09.
   6  *  Copyright 2009 Schloss Lab UMASS Amherst. All rights reserved.
   7  *
   8  */
   9
  10 #include "unifracweightedcommand.h"
  11
  12 //**********************************************************************************************************************
  13 vector<string> UnifracWeightedCommand::getValidParameters(){
  14         try {
  15                 string Array[] =  {"groups","iters","distance","random","processors","outputdir","inputdir"};
  16                 vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
  17                 return myArray;
  18         }
  19         catch(exception& e) {
  20                 m->errorOut(e, "UnifracWeightedCommand", "getValidParameters");
  21                 exit(1);
  22         }
  23 }
  24 //**********************************************************************************************************************
  25 UnifracWeightedCommand::UnifracWeightedCommand(){
  26         try {
  27                 abort = true; calledHelp = true;
  28                 vector<string> tempOutNames;
  29                 outputTypes["weighted"] = tempOutNames;
  30                 outputTypes["wsummary"] = tempOutNames;
  31                 outputTypes["phylip"] = tempOutNames;
  32                 outputTypes["column"] = tempOutNames;
  33         }
  34         catch(exception& e) {
  35                 m->errorOut(e, "UnifracWeightedCommand", "UnifracWeightedCommand");
  36                 exit(1);
  37         }
  38 }
  39 //**********************************************************************************************************************
  40 vector<string> UnifracWeightedCommand::getRequiredParameters(){
  41         try {
  42                 vector<string> myArray;
  43                 return myArray;
  44         }
  45         catch(exception& e) {
  46                 m->errorOut(e, "UnifracWeightedCommand", "getRequiredParameters");
  47                 exit(1);
  48         }
  49 }
  50 //**********************************************************************************************************************
  51 vector<string> UnifracWeightedCommand::getRequiredFiles(){
  52         try {
  53                 string Array[] =  {"tree","group"};
  54                 vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
  55
  56                 return myArray;
  57         }
  58         catch(exception& e) {
  59                 m->errorOut(e, "UnifracWeightedCommand", "getRequiredFiles");
  60                 exit(1);
  61         }
  62 }
  63 /***********************************************************/
  64 UnifracWeightedCommand::UnifracWeightedCommand(string option) {
  65         try {
  66                 globaldata = GlobalData::getInstance();
  67                 abort = false; calledHelp = false;
  68                 Groups.clear();
  69
  70                 //allow user to run help
  71                 if(option == "help") { help(); abort = true; calledHelp = true; }
  72
  73                 else {
  74                         //valid paramters for this command
  75                         string Array[] =  {"groups","iters","distance","random","processors","outputdir","inputdir"};
  76                         vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
  77
  78                         OptionParser parser(option);
  79                         map<string,string> parameters=parser.getParameters();
  80
  81                         ValidParameters validParameter;
  82
  83                         //check to make sure all parameters are valid for command
  84                         for (map<string,string>::iterator it = parameters.begin(); it != parameters.end(); it++) {
  85                                 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) {  abort = true;  }
  86                         }
  87
  88                         //initialize outputTypes
  89                         vector<string> tempOutNames;
  90                         outputTypes["weighted"] = tempOutNames;
  91                         outputTypes["wsummary"] = tempOutNames;
  92                         outputTypes["phylip"] = tempOutNames;
  93                         outputTypes["column"] = tempOutNames;
  94
  95                         if (globaldata->gTree.size() == 0) {//no trees were read
  96                                 m->mothurOut("You must execute the read.tree command, before you may execute the unifrac.weighted command."); m->mothurOutEndLine(); abort = true;  }
  97
  98                         //if the user changes the output directory command factory will send this info to us in the output parameter
  99                         outputDir = validParameter.validFile(parameters, "outputdir", false);           if (outputDir == "not found"){
 100                                 outputDir = "";
 101                                 outputDir += m->hasPath(globaldata->inputFileName); //if user entered a file with a path then preserve it
 102                         }
 103
 104                         //check for optional parameter and set defaults
 105                         // ...at some point should added some additional type checking...
 106                         groups = validParameter.validFile(parameters, "groups", false);
 107                         if (groups == "not found") { groups = ""; }
 108                         else {
 109                                 m->splitAtDash(groups, Groups);
 110                                 globaldata->Groups = Groups;
 111                         }
 112
 113                         itersString = validParameter.validFile(parameters, "iters", false);                     if (itersString == "not found") { itersString = "1000"; }
 114                         convert(itersString, iters);
 115
 116                         string temp = validParameter.validFile(parameters, "distance", false);
 117                         if (temp == "not found") { phylip = false; outputForm = ""; }
 118                         else{
 119                                 if ((temp == "lt") || (temp == "column") || (temp == "square")) {  phylip = true;  outputForm = temp; }
 120                                 else { m->mothurOut("Options for distance are: lt, square, or column. Using lt."); m->mothurOutEndLine(); phylip = true; outputForm = "lt"; }
 121                         }
 122
 123                         temp = validParameter.validFile(parameters, "random", false);                                   if (temp == "not found") { temp = "F"; }
 124                         random = m->isTrue(temp);
 125
 126                         temp = validParameter.validFile(parameters, "processors", false);       if (temp == "not found"){       temp = "1";                             }
 127                         convert(temp, processors);
 128
 129                         if (!random) {  iters = 0;  } //turn off random calcs
 130
 131
 132                         if (abort == false) {
 133                                 T = globaldata->gTree;
 134                                 tmap = globaldata->gTreemap;
 135                                 sumFile = outputDir + m->getSimpleName(globaldata->getTreeFile()) + ".wsummary";
 136                                 m->openOutputFile(sumFile, outSum);
 137                                 outputNames.push_back(sumFile);  outputTypes["wsummary"].push_back(sumFile);
 138
 139                                 util = new SharedUtil();
 140                                 string s; //to make work with setgroups
 141                                 util->setGroups(globaldata->Groups, tmap->namesOfGroups, s, numGroups, "weighted");     //sets the groups the user wants to analyze
 142                                 util->getCombos(groupComb, globaldata->Groups, numComp);
 143
 144                                 weighted = new Weighted(tmap);
 145
 146                         }
 147                 }
 148
 149
 150         }
 151         catch(exception& e) {
 152                 m->errorOut(e, "UnifracWeightedCommand", "UnifracWeightedCommand");
 153                 exit(1);
 154         }
 155 }
 156 //**********************************************************************************************************************
 157
 158 void UnifracWeightedCommand::help(){
 159         try {
 160                 m->mothurOut("The unifrac.weighted command can only be executed after a successful read.tree command.\n");
 161                 m->mothurOut("The unifrac.weighted command parameters are groups, iters, distance, processors and random.  No parameters are required.\n");
 162                 m->mothurOut("The groups parameter allows you to specify which of the groups in your groupfile you would like analyzed.  You must enter at least 2 valid groups.\n");
 163                 m->mothurOut("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");
 164                 m->mothurOut("The distance parameter allows you to create a distance file from the results. The default is false.\n");
 165                 m->mothurOut("The random parameter allows you to shut off the comparison to random trees. The default is false, meaning don't compare your trees with randomly generated trees.\n");
 166                 m->mothurOut("The processors parameter allows you to specify the number of processors to use. The default is 1.\n");
 167                 m->mothurOut("The unifrac.weighted command should be in the following format: unifrac.weighted(groups=yourGroups, iters=yourIters).\n");
 168                 m->mothurOut("Example unifrac.weighted(groups=A-B-C, iters=500).\n");
 169                 m->mothurOut("The default value for groups is all the groups in your groupfile, and iters is 1000.\n");
 170                 m->mothurOut("The unifrac.weighted command output two files: .weighted and .wsummary their descriptions are in the manual.\n");
 171                 m->mothurOut("Note: No spaces between parameter labels (i.e. groups), '=' and parameters (i.e.yourGroups).\n\n");
 172         }
 173         catch(exception& e) {
 174                 m->errorOut(e, "UnifracWeightedCommand", "help");
 175                 exit(1);
 176         }
 177 }
 178
 179 /***********************************************************/
 180 int UnifracWeightedCommand::execute() {
 181         try {
 182
 183                 if (abort == true) { if (calledHelp) { return 0; }  return 2;   }
 184
 185                 int start = time(NULL);
 186
 187                 //get weighted for users tree
 188                 userData.resize(numComp,0);  //data[0] = weightedscore AB, data[1] = weightedscore AC...
 189                 randomData.resize(numComp,0); //data[0] = weightedscore AB, data[1] = weightedscore AC...
 190
 191                 if (numComp < processors) { processors = numComp; }
 192
 193                 //get weighted scores for users trees
 194                 for (int i = 0; i < T.size(); i++) {
 195
 196                         if (m->control_pressed) { outSum.close(); for (int i = 0; i < outputNames.size(); i++) {        remove(outputNames[i].c_str());  } return 0; }
 197
 198                         counter = 0;
 199                         rScores.resize(numComp);  //data[0] = weightedscore AB, data[1] = weightedscore AC...
 200                         uScores.resize(numComp);  //data[0] = weightedscore AB, data[1] = weightedscore AC...
 201
 202                         if (random) {
 203                                 output = new ColumnFile(outputDir + m->getSimpleName(globaldata->getTreeFile())  + toString(i+1) + ".weighted", itersString);
 204                                 outputNames.push_back(outputDir + m->getSimpleName(globaldata->getTreeFile())  + toString(i+1) + ".weighted");
 205                                 outputTypes["weighted"].push_back(outputDir + m->getSimpleName(globaldata->getTreeFile())  + toString(i+1) + ".weighted");
 206                         }
 207
 208                         userData = weighted->getValues(T[i], processors, outputDir);  //userData[0] = weightedscore
 209
 210                         if (m->control_pressed) { if (random) { delete output; } outSum.close(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str());  } return 0; }
 211
 212                         //save users score
 213                         for (int s=0; s<numComp; s++) {
 214                                 //add users score to vector of user scores
 215                                 uScores[s].push_back(userData[s]);
 216
 217                                 //save users tree score for summary file
 218                                 utreeScores.push_back(userData[s]);
 219                         }
 220
 221                         if (random) {
 222
 223                                 //calculate number of comparisons i.e. with groups A,B,C = AB, AC, BC = 3;
 224                                 vector< vector<string> > namesOfGroupCombos;
 225                                 for (int a=0; a<numGroups; a++) {
 226                                         for (int l = 0; l < a; l++) {
 227                                                 vector<string> groups; groups.push_back(globaldata->Groups[a]); groups.push_back(globaldata->Groups[l]);
 228                                                 namesOfGroupCombos.push_back(groups);
 229                                         }
 230                                 }
 231
 232                                 lines.clear();
 233
 234                                 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
 235                                         if(processors != 1){
 236                                                 int numPairs = namesOfGroupCombos.size();
 237                                                 int numPairsPerProcessor = numPairs / processors;
 238
 239                                                 for (int i = 0; i < processors; i++) {
 240                                                         int startPos = i * numPairsPerProcessor;
 241                                                         if(i == processors - 1){
 242                                                                 numPairsPerProcessor = numPairs - i * numPairsPerProcessor;
 243                                                         }
 244                                                         lines.push_back(linePair(startPos, numPairsPerProcessor));
 245                                                 }
 246                                         }
 247                                 #endif
 248
 249
 250                                 //get scores for random trees
 251                                 for (int j = 0; j < iters; j++) {
 252
 253                                         #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
 254                                                 if(processors == 1){
 255                                                         driver(T[i],  namesOfGroupCombos, 0, namesOfGroupCombos.size(),  rScores);
 256                                                 }else{
 257                                                         createProcesses(T[i],  namesOfGroupCombos, rScores);
 258                                                 }
 259                                         #else
 260                                                 driver(T[i], namesOfGroupCombos, 0, namesOfGroupCombos.size(), rScores);
 261                                         #endif
 262
 263                                         if (m->control_pressed) { delete output; outSum.close(); for (int i = 0; i < outputNames.size(); i++) { remove(outputNames[i].c_str());  } return 0; }
 264
 265                                         //report progress
 266 //                                      m->mothurOut("Iter: " + toString(j+1)); m->mothurOutEndLine();
 267                                 }
 268                                 lines.clear();
 269
 270                                 //find the signifigance of the score for summary file
 271                                 for (int f = 0; f < numComp; f++) {
 272                                         //sort random scores
 273                                         sort(rScores[f].begin(), rScores[f].end());
 274
 275                                         //the index of the score higher than yours is returned
 276                                         //so if you have 1000 random trees the index returned is 100
 277                                         //then there are 900 trees with a score greater then you.
 278                                         //giving you a signifigance of 0.900
 279                                         int index = findIndex(userData[f], f);    if (index == -1) { m->mothurOut("error in UnifracWeightedCommand"); m->mothurOutEndLine(); exit(1); } //error code
 280
 281                                         //the signifigance is the number of trees with the users score or higher
 282                                         WScoreSig.push_back((iters-index)/(float)iters);
 283                                 }
 284
 285                                 //out << "Tree# " << i << endl;
 286                                 calculateFreqsCumuls();
 287                                 printWeightedFile();
 288
 289                                 delete output;
 290
 291                         }
 292
 293                         //clear data
 294                         rScores.clear();
 295                         uScores.clear();
 296                         validScores.clear();
 297                 }
 298
 299
 300                 if (m->control_pressed) { outSum.close(); for (int i = 0; i < outputNames.size(); i++) {        remove(outputNames[i].c_str());  } return 0;  }
 301
 302                 printWSummaryFile();
 303
 304                 if (phylip) {   createPhylipFile();             }
 305
 306                 //clear out users groups
 307                 globaldata->Groups.clear();
 308
 309
 310                 if (m->control_pressed) {
 311                         for (int i = 0; i < outputNames.size(); i++) {  remove(outputNames[i].c_str());  }
 312                         return 0;
 313                 }
 314
 315                 m->mothurOut("It took " + toString(time(NULL) - start) + " secs to run unifrac.weighted."); m->mothurOutEndLine();
 316
 317                 m->mothurOutEndLine();
 318                 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
 319                 for (int i = 0; i < outputNames.size(); i++) {  m->mothurOut(outputNames[i]); m->mothurOutEndLine();    }
 320                 m->mothurOutEndLine();
 321
 322                 return 0;
 323
 324         }
 325         catch(exception& e) {
 326                 m->errorOut(e, "UnifracWeightedCommand", "execute");
 327                 exit(1);
 328         }
 329 }
 330 /**************************************************************************************************/
 331
 332 int UnifracWeightedCommand::createProcesses(Tree* t, vector< vector<string> > namesOfGroupCombos, vector< vector<double> >& scores) {
 333         try {
 334 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
 335                 int process = 1;
 336                 vector<int> processIDS;
 337
 338                 EstOutput results;
 339
 340                 //loop through and create all the processes you want
 341                 while (process != processors) {
 342                         int pid = fork();
 343
 344                         if (pid > 0) {
 345                                 processIDS.push_back(pid);  //create map from line number to pid so you can append files in correct order later
 346                                 process++;
 347                         }else if (pid == 0){
 348                                 driver(t, namesOfGroupCombos, lines[process].start, lines[process].num, scores);
 349
 350                                 //pass numSeqs to parent
 351                                 ofstream out;
 352                                 string tempFile = outputDir + toString(getpid()) + ".weightedcommand.results.temp";
 353                                 m->openOutputFile(tempFile, out);
 354                                 for (int i = lines[process].start; i < (lines[process].start + lines[process].num); i++) { out << scores[i][(scores[i].size()-1)] << '\t';  } out << endl;
 355                                 out.close();
 356
 357                                 exit(0);
 358                         }else {
 359                                 m->mothurOut("[ERROR]: unable to spawn the necessary processes."); m->mothurOutEndLine();
 360                                 for (int i = 0; i < processIDS.size(); i++) { kill (processIDS[i], SIGINT); }
 361                                 exit(0);
 362                         }
 363                 }
 364
 365                 driver(t, namesOfGroupCombos, lines[0].start, lines[0].num, scores);
 366
 367                 //force parent to wait until all the processes are done
 368                 for (int i=0;i<(processors-1);i++) {
 369                         int temp = processIDS[i];
 370                         wait(&temp);
 371                 }
 372
 373                 //get data created by processes
 374                 for (int i=0;i<(processors-1);i++) {
 375
 376                         ifstream in;
 377                         string s = outputDir + toString(processIDS[i]) + ".weightedcommand.results.temp";
 378                         m->openInputFile(s, in);
 379
 380                         double tempScore;
 381                         for (int j = lines[(i+1)].start; j < (lines[(i+1)].start + lines[(i+1)].num); j++) { in >> tempScore; scores[j].push_back(tempScore); }
 382                         in.close();
 383                         remove(s.c_str());
 384                 }
 385
 386                 return 0;
 387 #endif
 388         }
 389         catch(exception& e) {
 390                 m->errorOut(e, "UnifracWeightedCommand", "createProcesses");
 391                 exit(1);
 392         }
 393 }
 394
 395 /**************************************************************************************************/
 396 int UnifracWeightedCommand::driver(Tree* t, vector< vector<string> > namesOfGroupCombos, int start, int num, vector< vector<double> >& scores) {
 397  try {
 398                 Tree* randT = new Tree();
 399
 400                 for (int h = start; h < (start+num); h++) {
 401
 402                         if (m->control_pressed) { return 0; }
 403
 404                         //initialize weighted score
 405                         string groupA = namesOfGroupCombos[h][0];
 406                         string groupB = namesOfGroupCombos[h][1];
 407
 408                         //copy T[i]'s info.
 409                         randT->getCopy(t);
 410
 411                         //create a random tree with same topology as T[i], but different labels
 412                         randT->assembleRandomUnifracTree(groupA, groupB);
 413
 414                         if (m->control_pressed) { delete randT;  return 0;  }
 415
 416                         //get wscore of random tree
 417                         EstOutput randomData = weighted->getValues(randT, groupA, groupB);
 418
 419                         if (m->control_pressed) { delete randT;  return 0;  }
 420
 421                         //save scores
 422                         scores[h].push_back(randomData[0]);
 423                 }
 424
 425                 delete randT;
 426
 427                 return 0;
 428
 429         }
 430         catch(exception& e) {
 431                 m->errorOut(e, "UnifracWeightedCommand", "driver");
 432                 exit(1);
 433         }
 434 }
 435 /***********************************************************/
 436 void UnifracWeightedCommand::printWeightedFile() {
 437         try {
 438                 vector<double> data;
 439                 vector<string> tags;
 440                 tags.push_back("Score"); tags.push_back("RandFreq"); tags.push_back("RandCumul");
 441
 442                 for(int a = 0; a < numComp; a++) {
 443                         output->initFile(groupComb[a], tags);
 444                         //print each line
 445                         for (map<float,float>::iterator it = validScores.begin(); it != validScores.end(); it++) {
 446                                 data.push_back(it->first);  data.push_back(rScoreFreq[a][it->first]); data.push_back(rCumul[a][it->first]);
 447                                 output->output(data);
 448                                 data.clear();
 449                         }
 450                         output->resetFile();
 451                 }
 452         }
 453         catch(exception& e) {
 454                 m->errorOut(e, "UnifracWeightedCommand", "printWeightedFile");
 455                 exit(1);
 456         }
 457 }
 458
 459
 460 /***********************************************************/
 461 void UnifracWeightedCommand::printWSummaryFile() {
 462         try {
 463                 //column headers
 464                 outSum << "Tree#" << '\t' << "Groups" << '\t' << "WScore" << '\t' << "WSig" <<  endl;
 465                 m->mothurOut("Tree#\tGroups\tWScore\tWSig"); m->mothurOutEndLine();
 466
 467                 //format output
 468                 outSum.setf(ios::fixed, ios::floatfield); outSum.setf(ios::showpoint);
 469
 470                 //print each line
 471                 int count = 0;
 472                 for (int i = 0; i < T.size(); i++) {
 473                         for (int j = 0; j < numComp; j++) {
 474                                 if (random) {
 475                                         if (WScoreSig[count] > (1/(float)iters)) {
 476                                                 outSum << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << WScoreSig[count] << endl;
 477                                                 cout << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << WScoreSig[count] << endl;
 478                                                 m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t" +  toString(WScoreSig[count]) + "\n");
 479                                         }else{
 480                                                 outSum << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << "<" << (1/float(iters)) << endl;
 481                                                 cout << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << setprecision(itersString.length()) << "<" << (1/float(iters)) << endl;
 482                                                 m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t<" +  toString((1/float(iters))) + "\n");
 483                                         }
 484                                 }else{
 485                                         outSum << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << "0.00" << endl;
 486                                         cout << setprecision(6) << i+1 << '\t' << groupComb[j] << '\t' << utreeScores[count] << '\t' << "0.00" << endl;
 487                                         m->mothurOutJustToLog(toString(i+1) +"\t" + groupComb[j] +"\t" + toString(utreeScores[count]) +"\t0.00\n");
 488                                 }
 489                                 count++;
 490                         }
 491                 }
 492                 outSum.close();
 493         }
 494         catch(exception& e) {
 495                 m->errorOut(e, "UnifracWeightedCommand", "printWSummaryFile");
 496                 exit(1);
 497         }
 498 }
 499 /***********************************************************/
 500 void UnifracWeightedCommand::createPhylipFile() {
 501         try {
 502                 int count = 0;
 503                 //for each tree
 504                 for (int i = 0; i < T.size(); i++) {
 505
 506                         string phylipFileName;
 507                         if ((outputForm == "lt") || (outputForm == "square")) {
 508                                 phylipFileName = outputDir + m->getSimpleName(globaldata->getTreeFile())  + toString(i+1) + ".weighted.phylip.dist";
 509                                 outputNames.push_back(phylipFileName); outputTypes["phylip"].push_back(phylipFileName);
 510                         }else { //column
 511                                 phylipFileName = outputDir + m->getSimpleName(globaldata->getTreeFile())  + toString(i+1) + ".weighted.column.dist";
 512                                 outputNames.push_back(phylipFileName); outputTypes["column"].push_back(phylipFileName);
 513                         }
 514
 515                         ofstream out;
 516                         m->openOutputFile(phylipFileName, out);
 517
 518                         if ((outputForm == "lt") || (outputForm == "square")) {
 519                                 //output numSeqs
 520                                 out << globaldata->Groups.size() << endl;
 521                         }
 522
 523                         //make matrix with scores in it
 524                         vector< vector<float> > dists;  dists.resize(globaldata->Groups.size());
 525                         for (int i = 0; i < globaldata->Groups.size(); i++) {
 526                                 dists[i].resize(globaldata->Groups.size(), 0.0);
 527                         }
 528
 529                         //flip it so you can print it
 530                         for (int r=0; r<globaldata->Groups.size(); r++) {
 531                                 for (int l = 0; l < r; l++) {
 532                                         dists[r][l] = utreeScores[count];
 533                                         dists[l][r] = utreeScores[count];
 534                                         count++;
 535                                 }
 536                         }
 537
 538                         //output to file
 539                         for (int r=0; r<globaldata->Groups.size(); r++) {
 540                                 //output name
 541                                 string name = globaldata->Groups[r];
 542                                 if (name.length() < 10) { //pad with spaces to make compatible
 543                                         while (name.length() < 10) {  name += " ";  }
 544                                 }
 545
 546                                 if (outputForm == "lt") {
 547                                         out << name << '\t';
 548
 549                                         //output distances
 550                                         for (int l = 0; l < r; l++) {   out  << dists[r][l] << '\t';  }
 551                                         out << endl;
 552                                 }else if (outputForm == "square") {
 553                                         out << name << '\t';
 554
 555                                         //output distances
 556                                         for (int l = 0; l < globaldata->Groups.size(); l++) {   out  << dists[r][l] << '\t';  }
 557                                         out << endl;
 558                                 }else{
 559                                         //output distances
 560                                         for (int l = 0; l < r; l++) {
 561                                                 string otherName = globaldata->Groups[l];
 562                                                 if (otherName.length() < 10) { //pad with spaces to make compatible
 563                                                         while (otherName.length() < 10) {  otherName += " ";  }
 564                                                 }
 565
 566                                                 out  << name << '\t' << otherName << dists[r][l] << endl;
 567                                         }
 568                                 }
 569                         }
 570                         out.close();
 571                 }
 572         }
 573         catch(exception& e) {
 574                 m->errorOut(e, "UnifracWeightedCommand", "createPhylipFile");
 575                 exit(1);
 576         }
 577 }
 578 /***********************************************************/
 579 int UnifracWeightedCommand::findIndex(float score, int index) {
 580         try{
 581                 for (int i = 0; i < rScores[index].size(); i++) {
 582                         if (rScores[index][i] >= score) {       return i;       }
 583                 }
 584                 return rScores[index].size();
 585         }
 586         catch(exception& e) {
 587                 m->errorOut(e, "UnifracWeightedCommand", "findIndex");
 588                 exit(1);
 589         }
 590 }
 591
 592 /***********************************************************/
 593
 594 void UnifracWeightedCommand::calculateFreqsCumuls() {
 595         try {
 596                 //clear out old tree values
 597                 rScoreFreq.clear();
 598                 rScoreFreq.resize(numComp);
 599                 rCumul.clear();
 600                 rCumul.resize(numComp);
 601                 validScores.clear();
 602
 603                 //calculate frequency
 604                 for (int f = 0; f < numComp; f++) {
 605                         for (int i = 0; i < rScores[f].size(); i++) { //looks like 0,0,1,1,1,2,4,7...  you want to make a map that say rScoreFreq[0] = 2, rScoreFreq[1] = 3...
 606                                 validScores[rScores[f][i]] = rScores[f][i];
 607                                 map<float,float>::iterator it = rScoreFreq[f].find(rScores[f][i]);
 608                                 if (it != rScoreFreq[f].end()) {
 609                                         rScoreFreq[f][rScores[f][i]]++;
 610                                 }else{
 611                                         rScoreFreq[f][rScores[f][i]] = 1;
 612                                 }
 613                         }
 614                 }
 615
 616                 //calculate rcumul
 617                 for(int a = 0; a < numComp; a++) {
 618                         float rcumul = 1.0000;
 619                         //this loop fills the cumulative maps and put 0.0000 in the score freq map to make it easier to print.
 620                         for (map<float,float>::iterator it = validScores.begin(); it != validScores.end(); it++) {
 621                                 //make rscoreFreq map and rCumul
 622                                 map<float,float>::iterator it2 = rScoreFreq[a].find(it->first);
 623                                 rCumul[a][it->first] = rcumul;
 624                                 //get percentage of random trees with that info
 625                                 if (it2 != rScoreFreq[a].end()) {  rScoreFreq[a][it->first] /= iters; rcumul-= it2->second;  }
 626                                 else { rScoreFreq[a][it->first] = 0.0000; } //no random trees with that score
 627                         }
 628                 }
 629
 630         }
 631         catch(exception& e) {
 632                 m->errorOut(e, "UnifracWeightedCommand", "calculateFreqsCums");
 633                 exit(1);
 634         }
 635 }
 636
 637 /***********************************************************/
 638
 639
 640
 641
 642