distancecommand.cpp

   1 /*
   2  *  distancecommand.cpp
   3  *  Mothur
   4  *
   5  *  Created by Sarah Westcott on 5/7/09.
   6  *  Copyright 2009 Schloss Lab UMASS Amherst. All rights reserved.
   7  *
   8  */
   9
  10 #include "distancecommand.h"
  11 #include "ignoregaps.h"
  12 #include "eachgapdist.h"
  13 #include "eachgapignore.h"
  14 #include "onegapdist.h"
  15 #include "onegapignore.h"
  16
  17 //**********************************************************************************************************************
  18
  19 DistanceCommand::DistanceCommand(string option) {
  20         try {
  21                 abort = false;
  22                 Estimators.clear();
  23
  24                 //allow user to run help
  25                 if(option == "help") { help(); abort = true; }
  26
  27                 else {
  28                         //valid paramters for this command
  29                         string Array[] =  {"fasta", "output", "calc", "countends", "cutoff", "processors", "outputdir","inputdir"};
  30                         vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
  31
  32                         OptionParser parser(option);
  33                         map<string, string> parameters = parser.getParameters();
  34
  35                         ValidParameters validParameter;
  36                         map<string, string>::iterator it2;
  37
  38                         //check to make sure all parameters are valid for command
  39                         for (it2 = parameters.begin(); it2 != parameters.end(); it2++) {
  40                                 if (validParameter.isValidParameter(it2->first, myArray, it2->second) != true) {  abort = true;  }
  41                         }
  42
  43                         //if the user changes the input directory command factory will send this info to us in the output parameter
  44                         string inputDir = validParameter.validFile(parameters, "inputdir", false);
  45                         if (inputDir == "not found"){   inputDir = "";          }
  46                         else {
  47                                 string path;
  48                                 it2 = parameters.find("fasta");
  49                                 //user has given a template file
  50                                 if(it2 != parameters.end()){
  51                                         path = hasPath(it2->second);
  52                                         //if the user has not given a path then, add inputdir. else leave path alone.
  53                                         if (path == "") {       parameters["fasta"] = inputDir + it2->second;           }
  54                                 }
  55                         }
  56
  57                         //check for required parameters
  58                         fastafile = validParameter.validFile(parameters, "fasta", true);
  59                         if (fastafile == "not found") { m->mothurOut("fasta is a required parameter for the dist.seqs command."); m->mothurOutEndLine(); abort = true; }
  60                         else if (fastafile == "not open") { abort = true; }
  61                         else{
  62                                 ifstream inFASTA;
  63                                 openInputFile(fastafile, inFASTA);
  64                                 alignDB = SequenceDB(inFASTA);
  65                                 inFASTA.close();
  66                         }
  67
  68                         //if the user changes the output directory command factory will send this info to us in the output parameter
  69                         outputDir = validParameter.validFile(parameters, "outputdir", false);           if (outputDir == "not found"){
  70                                 outputDir = "";
  71                                 outputDir += hasPath(fastafile); //if user entered a file with a path then preserve it
  72                         }
  73
  74                         //check for optional parameter and set defaults
  75                         // ...at some point should added some additional type checking...
  76                         calc = validParameter.validFile(parameters, "calc", false);
  77                         if (calc == "not found") { calc = "onegap";  }
  78                         else {
  79                                  if (calc == "default")  {  calc = "onegap";  }
  80                         }
  81                         splitAtDash(calc, Estimators);
  82
  83                         string temp;
  84                         temp = validParameter.validFile(parameters, "countends", false);        if(temp == "not found"){        temp = "T";     }
  85                         convert(temp, countends);
  86
  87                         temp = validParameter.validFile(parameters, "cutoff", false);           if(temp == "not found"){        temp = "1.0"; }
  88                         convert(temp, cutoff);
  89
  90                         temp = validParameter.validFile(parameters, "processors", false);       if(temp == "not found"){        temp = "1"; }
  91                         convert(temp, processors);
  92
  93                         output = validParameter.validFile(parameters, "output", false);         if(output == "not found"){      output = "column"; }
  94
  95                         if ((output != "column") && (output != "lt") && (output != "square")) { m->mothurOut(output + " is not a valid output form. Options are column, lt and square. I will use column."); m->mothurOutEndLine(); output = "column"; }
  96
  97                         ValidCalculators validCalculator;
  98
  99                         if (isTrue(countends) == true) {
 100                                 for (int i=0; i<Estimators.size(); i++) {
 101                                         if (validCalculator.isValidCalculator("distance", Estimators[i]) == true) {
 102                                                 if (Estimators[i] == "nogaps")                  {       distCalculator = new ignoreGaps();      }
 103                                                 else if (Estimators[i] == "eachgap")    {       distCalculator = new eachGapDist();     }
 104                                                 else if (Estimators[i] == "onegap")             {       distCalculator = new oneGapDist();      }
 105                                         }
 106                                 }
 107                         }else {
 108                                 for (int i=0; i<Estimators.size(); i++) {
 109                                         if (validCalculator.isValidCalculator("distance", Estimators[i]) == true) {
 110                                                 if (Estimators[i] == "nogaps")          {       distCalculator = new ignoreGaps();                                      }
 111                                                 else if (Estimators[i] == "eachgap"){   distCalculator = new eachGapIgnoreTermGapDist();        }
 112                                                 else if (Estimators[i] == "onegap")     {       distCalculator = new oneGapIgnoreTermGapDist();         }
 113                                         }
 114                                 }
 115                         }
 116
 117                 }
 118
 119         }
 120         catch(exception& e) {
 121                 m->errorOut(e, "DistanceCommand", "DistanceCommand");
 122                 exit(1);
 123         }
 124 }
 125
 126 //**********************************************************************************************************************
 127
 128 DistanceCommand::~DistanceCommand(){
 129
 130         for(int i=0;i<lines.size();i++){
 131                 delete lines[i];
 132         }
 133
 134 }
 135
 136 //**********************************************************************************************************************
 137
 138 void DistanceCommand::help(){
 139         try {
 140                 m->mothurOut("The dist.seqs command reads a file containing sequences and creates a distance file.\n");
 141                 m->mothurOut("The dist.seqs command parameters are fasta, calc, countends, output, cutoff and processors.  \n");
 142                 m->mothurOut("The fasta parameter is required.\n");
 143                 m->mothurOut("The calc parameter allows you to specify the method of calculating the distances.  Your options are: nogaps, onegap or eachgap. The default is onegap.\n");
 144                 m->mothurOut("The countends parameter allows you to specify whether to include terminal gaps in distance.  Your options are: T or F. The default is T.\n");
 145                 m->mothurOut("The cutoff parameter allows you to specify maximum distance to keep. The default is 1.0.\n");
 146                 m->mothurOut("The output parameter allows you to specify format of your distance matrix. Options are column, lt, and square. The default is column.\n");
 147                 m->mothurOut("The processors parameter allows you to specify number of processors to use.  The default is 1.\n");
 148                 m->mothurOut("The dist.seqs command should be in the following format: \n");
 149                 m->mothurOut("dist.seqs(fasta=yourFastaFile, calc=yourCalc, countends=yourEnds, cutoff= yourCutOff, processors=yourProcessors) \n");
 150                 m->mothurOut("Example dist.seqs(fasta=amazon.fasta, calc=eachgap, countends=F, cutoff= 2.0, processors=3).\n");
 151                 m->mothurOut("Note: No spaces between parameter labels (i.e. calc), '=' and parameters (i.e.yourCalc).\n\n");
 152         }
 153         catch(exception& e) {
 154                 m->errorOut(e, "DistanceCommand", "help");
 155                 exit(1);
 156         }
 157 }
 158 //**********************************************************************************************************************
 159
 160 int DistanceCommand::execute(){
 161         try {
 162
 163                 if (abort == true) { return 0; }
 164
 165                 int startTime = time(NULL);
 166
 167                 int numSeqs = alignDB.getNumSeqs();
 168                 cutoff += 0.005;
 169
 170                 string outputFile;
 171
 172                 if (output == "lt") { //does the user want lower triangle phylip formatted file
 173                         outputFile = outputDir + getRootName(getSimpleName(fastafile)) + "phylip.dist";
 174                         remove(outputFile.c_str());
 175
 176                         //output numSeqs to phylip formatted dist file
 177                 }else if (output == "column") { //user wants column format
 178                         outputFile = outputDir + getRootName(getSimpleName(fastafile)) + "dist";
 179                         remove(outputFile.c_str());
 180                 }else { //assume square
 181                         outputFile = outputDir + getRootName(getSimpleName(fastafile)) + "square.dist";
 182                         remove(outputFile.c_str());
 183                 }
 184
 185
 186 #ifdef USE_MPI
 187
 188                 int pid, start, end;
 189                 int tag = 2001;
 190
 191                 MPI_Status status;
 192                 MPI_Comm_size(MPI_COMM_WORLD, &processors); //set processors to the number of mpi processes running
 193                 MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
 194
 195                 //each process gets where it should start and stop in the file
 196                 start = int (sqrt(float(pid)/float(processors)) * numSeqs);
 197                 end = int (sqrt(float(pid+1)/float(processors)) * numSeqs);
 198
 199                 if (output != "lt") {
 200                         MPI_File outMPI;
 201                         int amode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
 202
 203                         char filename[outputFile.length()];
 204                         strcpy(filename, outputFile.c_str());
 205
 206                         MPI_File_open(MPI_COMM_WORLD, filename, amode, MPI_INFO_NULL, &outMPI);
 207
 208                         if (m->control_pressed) {   MPI_File_close(&outMPI);  delete distCalculator;  return 0;  }
 209
 210                         if (pid == 0) { //you are the root process
 211
 212                                 //do your part
 213                                 string outputMyPart;
 214                                 driverMPI(start, end, outMPI, cutoff);
 215
 216                                 if (m->control_pressed) { MPI_File_close(&outMPI);  delete distCalculator;  return 0; }
 217
 218                                 //wait on chidren
 219                                 for(int i = 1; i < processors; i++) {
 220                                         if (m->control_pressed) { MPI_File_close(&outMPI);  delete distCalculator;  return 0; }
 221
 222                                         char buf[4];
 223                                         MPI_Recv(buf, 4, MPI_CHAR, i, tag, MPI_COMM_WORLD, &status);
 224                                 }
 225                         }else { //you are a child process
 226                                 //do your part
 227                                 driverMPI(start, end, outMPI, cutoff);
 228
 229                                 if (m->control_pressed) { MPI_File_close(&outMPI);  delete distCalculator;  return 0; }
 230
 231                                 char buf[4];
 232                                 strcpy(buf, "done");
 233                                 //tell parent you are done.
 234                                 MPI_Send(buf, 4, MPI_CHAR, 0, tag, MPI_COMM_WORLD);
 235                         }
 236
 237                         MPI_File_close(&outMPI);
 238
 239                 }else { //lower triangle format
 240                         if (pid == 0) { //you are the root process
 241
 242                                 //do your part
 243                                 string outputMyPart;
 244                                 long mySize;
 245                                 driverMPI(start, end, outputFile, mySize);
 246
 247                                 if (m->control_pressed) {  delete distCalculator;  return 0; }
 248
 249                                 int amode=MPI_MODE_APPEND|MPI_MODE_WRONLY|MPI_MODE_CREATE; //
 250                                 MPI_File outMPI;
 251                                 MPI_File inMPI;
 252
 253                                 char filename[outputFile.length()];
 254                                 strcpy(filename, outputFile.c_str());
 255
 256                                 MPI_File_open(MPI_COMM_SELF, filename, amode, MPI_INFO_NULL, &outMPI);
 257
 258                                 //wait on chidren
 259                                 for(int b = 1; b < processors; b++) {
 260                                         long fileSize;
 261
 262                                         if (m->control_pressed) { MPI_File_close(&outMPI);  delete distCalculator;  return 0; }
 263
 264                                         MPI_Recv(&fileSize, 1, MPI_LONG, b, tag, MPI_COMM_WORLD, &status);
 265
 266                                         string outTemp = outputFile + toString(b) + ".temp";
 267                                         char buf[outTemp.length()];
 268                                         strcpy(buf, outTemp.c_str());
 269
 270                                         MPI_File_open(MPI_COMM_SELF, buf, MPI_MODE_DELETE_ON_CLOSE|MPI_MODE_RDONLY, MPI_INFO_NULL, &inMPI);
 271
 272                                         int count = 0;
 273                                         while (count < fileSize) { //read 1000 characters at a time
 274                                                 //send freqs
 275                                                 char buf2[1];
 276                                                 MPI_File_read(inMPI, buf2, 1, MPI_CHAR, &status);
 277                                                 MPI_File_write(outMPI, buf2, 1, MPI_CHAR, &status);
 278                                                 count += 1;
 279                                         }
 280
 281                                         MPI_File_close(&inMPI); //deleted on close
 282                                 }
 283
 284                                 MPI_File_close(&outMPI);
 285                         }else { //you are a child process
 286                                 //do your part
 287                                 long size;
 288                                 driverMPI(start, end, (outputFile + toString(pid) + ".temp"), size);
 289
 290                                 if (m->control_pressed) { delete distCalculator;  return 0; }
 291
 292                                 //tell parent you are done.
 293                                 MPI_Send(&size, 1, MPI_LONG, 0, tag, MPI_COMM_WORLD);
 294                         }
 295                 }
 296 #else
 297
 298         #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
 299                 //if you don't need to fork anything
 300                 if(processors == 1){
 301                         driver(0, numSeqs, outputFile, cutoff);
 302                 }else{ //you have multiple processors
 303
 304                         for (int i = 0; i < processors; i++) {
 305                                 lines.push_back(new linePair());
 306                                 lines[i]->start = int (sqrt(float(i)/float(processors)) * numSeqs);
 307                                 lines[i]->end = int (sqrt(float(i+1)/float(processors)) * numSeqs);
 308                         }
 309
 310                         createProcesses(outputFile);
 311
 312                         map<int, int>::iterator it = processIDS.begin();
 313                         rename((outputFile + toString(it->second) + ".temp").c_str(), outputFile.c_str());
 314                         it++;
 315
 316                         //append and remove temp files
 317                         for (; it != processIDS.end(); it++) {
 318                                 appendFiles((outputFile + toString(it->second) + ".temp"), outputFile);
 319                                 remove((outputFile + toString(it->second) + ".temp").c_str());
 320                         }
 321                 }
 322         #else
 323                 ifstream inFASTA;
 324                 driver(0, numSeqs, outputFile, cutoff);
 325         #endif
 326
 327 #endif
 328                 if (m->control_pressed) { delete distCalculator; remove(outputFile.c_str()); return 0; }
 329
 330                 #ifdef USE_MPI
 331                         MPI_Comm_rank(MPI_COMM_WORLD, &pid);
 332
 333                         if (pid == 0) { //only one process should output to screen
 334                 #endif
 335
 336                 if (output == "square") {  convertMatrix(outputFile); }
 337
 338                 #ifdef USE_MPI
 339                         }
 340                 #endif
 341
 342                 if (m->control_pressed) { delete distCalculator; remove(outputFile.c_str()); return 0; }
 343
 344                 delete distCalculator;
 345
 346                 m->mothurOutEndLine();
 347                 m->mothurOut("Output File Name: "); m->mothurOutEndLine();
 348                 m->mothurOut(outputFile); m->mothurOutEndLine();
 349                 m->mothurOutEndLine();
 350                 m->mothurOut("It took " + toString(time(NULL) - startTime) + " to calculate the distances for " + toString(numSeqs) + " sequences."); m->mothurOutEndLine();
 351                 return 0;
 352
 353         }
 354         catch(exception& e) {
 355                 m->errorOut(e, "DistanceCommand", "execute");
 356                 exit(1);
 357         }
 358 }
 359 /**************************************************************************************************/
 360 void DistanceCommand::createProcesses(string filename) {
 361         try {
 362 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
 363                 int process = 0;
 364                 processIDS.clear();
 365
 366                 //loop through and create all the processes you want
 367                 while (process != processors) {
 368                         int pid = fork();
 369
 370                         if (pid > 0) {
 371                                 processIDS[lines[process]->end] = pid;  //create map from line number to pid so you can append files in correct order later
 372                                 process++;
 373                         }else if (pid == 0){
 374                                 driver(lines[process]->start, lines[process]->end, filename + toString(getpid()) + ".temp", cutoff);
 375                                 exit(0);
 376                         }else { m->mothurOut("unable to spawn the necessary processes."); m->mothurOutEndLine(); exit(0); }
 377                 }
 378
 379                 //force parent to wait until all the processes are done
 380                 for (map<int, int>::iterator it = processIDS.begin(); it != processIDS.end(); it++) {
 381                         int temp = it->second;
 382                         wait(&temp);
 383                 }
 384 #endif
 385         }
 386         catch(exception& e) {
 387                 m->errorOut(e, "DistanceCommand", "createProcesses");
 388                 exit(1);
 389         }
 390 }
 391
 392 /**************************************************************************************************/
 393 /////// need to fix to work with calcs and sequencedb
 394 int DistanceCommand::driver(int startLine, int endLine, string dFileName, float cutoff){
 395         try {
 396
 397                 int startTime = time(NULL);
 398
 399                 //column file
 400                 ofstream outFile(dFileName.c_str(), ios::trunc);
 401                 outFile.setf(ios::fixed, ios::showpoint);
 402                 outFile << setprecision(4);
 403
 404                 if((output == "lt") && startLine == 0){ outFile << alignDB.getNumSeqs() << endl;        }
 405
 406                 for(int i=startLine;i<endLine;i++){
 407                         if(output == "lt")      {
 408                                 string name = alignDB.get(i).getName();
 409                                 if (name.length() < 10) { //pad with spaces to make compatible
 410                                         while (name.length() < 10) {  name += " ";  }
 411                                 }
 412                                 outFile << name << '\t';
 413                         }
 414                         for(int j=0;j<i;j++){
 415
 416                                 if (m->control_pressed) { outFile.close(); return 0;  }
 417
 418                                 distCalculator->calcDist(alignDB.get(i), alignDB.get(j));
 419                                 double dist = distCalculator->getDist();
 420
 421                                 if(dist <= cutoff){
 422                                         if (output == "column") { outFile << alignDB.get(i).getName() << ' ' << alignDB.get(j).getName() << ' ' << dist << endl; }
 423                                 }
 424                                 if (output == "lt") {  outFile << dist << '\t'; }
 425
 426                                 if (output == "square") { //make a square column you can convert to square phylip
 427                                         outFile << alignDB.get(i).getName() << '\t' << alignDB.get(j).getName() << '\t' << dist << endl;
 428                                         outFile << alignDB.get(j).getName() << '\t' << alignDB.get(i).getName() << '\t' << dist << endl;
 429                                 }
 430
 431                         }
 432
 433                         if (output == "lt") { outFile << endl; }
 434
 435                         if(i % 100 == 0){
 436                                 m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
 437                         }
 438
 439                 }
 440                 m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
 441
 442                 outFile.close();
 443
 444                 return 1;
 445         }
 446         catch(exception& e) {
 447                 m->errorOut(e, "DistanceCommand", "driver");
 448                 exit(1);
 449         }
 450 }
 451 #ifdef USE_MPI
 452 /**************************************************************************************************/
 453 /////// need to fix to work with calcs and sequencedb
 454 int DistanceCommand::driverMPI(int startLine, int endLine, MPI_File& outMPI, float cutoff){
 455         try {
 456                 MPI_Status status;
 457                 int startTime = time(NULL);
 458
 459                 string outputString = "";
 460
 461                 for(int i=startLine;i<endLine;i++){
 462
 463                         for(int j=0;j<i;j++){
 464
 465                                 if (m->control_pressed) {  return 0;  }
 466
 467                                 distCalculator->calcDist(alignDB.get(i), alignDB.get(j));
 468                                 double dist = distCalculator->getDist();
 469
 470                                 if(dist <= cutoff){
 471                                         if (output == "column") { outputString += (alignDB.get(i).getName() + ' ' + alignDB.get(j).getName() + ' ' + toString(dist) + '\n'); }
 472                                 }
 473
 474                                 if (output == "square") { //make a square column you can convert to square phylip
 475                                         outputString += (alignDB.get(i).getName() + ' ' + alignDB.get(j).getName() + ' ' + toString(dist) + '\n');
 476                                         outputString += (alignDB.get(j).getName() + ' ' + alignDB.get(i).getName() + ' ' + toString(dist) + '\n');
 477                                 }
 478                         }
 479
 480                         if(i % 100 == 0){
 481                                 //m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
 482                                 cout << i << '\t' << (time(NULL) - startTime) << endl;
 483                         }
 484
 485
 486                         //send results to parent
 487                         int length = outputString.length();
 488                         char buf[length];
 489                         strcpy(buf, outputString.c_str());
 490
 491                         MPI_File_write_shared(outMPI, buf, length, MPI_CHAR, &status);
 492                         outputString = "";
 493
 494                 }
 495
 496                 //m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
 497                 cout << (endLine-1) << '\t' << (time(NULL) - startTime) << endl;
 498                 return 1;
 499         }
 500         catch(exception& e) {
 501                 m->errorOut(e, "DistanceCommand", "driverMPI");
 502                 exit(1);
 503         }
 504 }
 505 /**************************************************************************************************/
 506 /////// need to fix to work with calcs and sequencedb
 507 int DistanceCommand::driverMPI(int startLine, int endLine, string file, long& size){
 508         try {
 509                 MPI_Status status;
 510
 511                 MPI_File outMPI;
 512                 int amode=MPI_MODE_CREATE|MPI_MODE_WRONLY;
 513
 514                 char filename[file.length()];
 515                 strcpy(filename, file.c_str());
 516
 517                 MPI_File_open(MPI_COMM_SELF, filename, amode, MPI_INFO_NULL, &outMPI);
 518
 519                 int startTime = time(NULL);
 520
 521                 string outputString = "";
 522                 size = 0;
 523
 524                 if((output == "lt") && startLine == 0){ outputString += toString(alignDB.getNumSeqs()) + "\n";  }
 525
 526                 for(int i=startLine;i<endLine;i++){
 527                         if(output == "lt")      {
 528                                 string name = alignDB.get(i).getName();
 529                                 if (name.length() < 10) { //pad with spaces to make compatible
 530                                         while (name.length() < 10) {  name += " ";  }
 531                                 }
 532                                 outputString += name + "\t";
 533                         }
 534                         for(int j=0;j<i;j++){
 535
 536                                 if (m->control_pressed) {  return 0;  }
 537
 538                                 distCalculator->calcDist(alignDB.get(i), alignDB.get(j));
 539                                 double dist = distCalculator->getDist();
 540
 541                                 if (output == "lt") {  outputString += toString(dist) + "\t"; }
 542                         }
 543
 544                         if (output == "lt") { outputString += "\n"; }
 545
 546
 547                         if(i % 100 == 0){
 548                                 //m->mothurOut(toString(i) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
 549                                 cout << i << '\t' << (time(NULL) - startTime) << endl;
 550                         }
 551
 552
 553                         //send results to parent
 554                         int length = outputString.length();
 555                         char buf[length];
 556                         strcpy(buf, outputString.c_str());
 557
 558                         MPI_File_write(outMPI, buf, length, MPI_CHAR, &status);
 559                         size += outputString.length();
 560                         outputString = "";
 561
 562
 563                 }
 564
 565                 //m->mothurOut(toString(endLine-1) + "\t" + toString(time(NULL) - startTime)); m->mothurOutEndLine();
 566                 cout << (endLine-1) << '\t' << (time(NULL) - startTime) << endl;
 567                 MPI_File_close(&outMPI);
 568
 569                 return 1;
 570         }
 571         catch(exception& e) {
 572                 m->errorOut(e, "DistanceCommand", "driverMPI");
 573                 exit(1);
 574         }
 575 }
 576 #endif
 577 /**************************************************************************************************/
 578 int DistanceCommand::convertMatrix(string outputFile) {
 579         try{
 580
 581                 //sort file by first column so the distances for each row are together
 582                 string outfile = getRootName(outputFile) + "sorted.dist.temp";
 583
 584                 //use the unix sort
 585                 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
 586                         string command = "sort -n " + outputFile + " -o " + outfile;
 587                         system(command.c_str());
 588                 #else //sort using windows sort
 589                         string command = "sort " + outputFile + " /O " + outfile;
 590                         system(command.c_str());
 591                 #endif
 592
 593
 594                 //output to new file distance for each row and save positions in file where new row begins
 595                 ifstream in;
 596                 openInputFile(outfile, in);
 597
 598                 ofstream out;
 599                 openOutputFile(outputFile, out);
 600
 601                 out.setf(ios::fixed, ios::floatfield); out.setf(ios::showpoint);
 602
 603                 out << alignDB.getNumSeqs() << endl;
 604
 605                 //get first currentRow
 606                 string first, currentRow, second;
 607                 float dist;
 608                 map<string, float> rowDists; //take advantage of the fact that maps are already sorted by key
 609                 map<string, float>::iterator it;
 610
 611                 in >> first;
 612                 currentRow = first;
 613
 614                 rowDists[first] = 0.00; //distance to yourself is 0.0
 615
 616                 in.seekg(0);
 617                 //openInputFile(outfile, in);
 618
 619                 while(!in.eof()) {
 620                         if (m->control_pressed) { in.close(); remove(outfile.c_str()); out.close(); return 0; }
 621
 622                         in >> first >> second >> dist; gobble(in);
 623
 624                         if (first != currentRow) {
 625                                 //print out last row
 626                                 out << currentRow << '\t'; //print name
 627
 628                                 //print dists
 629                                 for (it = rowDists.begin(); it != rowDists.end(); it++) {
 630                                         out << it->second << '\t';
 631                                 }
 632                                 out << endl;
 633
 634                                 //start new row
 635                                 currentRow = first;
 636                                 rowDists.clear();
 637                                 rowDists[first] = 0.00;
 638                                 rowDists[second] = dist;
 639                         }else{
 640                                 rowDists[second] = dist;
 641                         }
 642                 }
 643                 //print out last row
 644                 out << currentRow << '\t'; //print name
 645
 646                 //print dists
 647                 for (it = rowDists.begin(); it != rowDists.end(); it++) {
 648                         out << it->second << '\t';
 649                 }
 650                 out << endl;
 651
 652                 in.close();
 653                 out.close();
 654
 655                 remove(outfile.c_str());
 656
 657                 return 1;
 658
 659         }
 660         catch(exception& e) {
 661                 m->errorOut(e, "DistanceCommand", "convertMatrix");
 662                 exit(1);
 663         }
 664 }
 665 /**************************************************************************************************/
 666 int DistanceCommand::convertToLowerTriangle(string outputFile) {
 667         try{
 668
 669                 //sort file by first column so the distances for each row are together
 670                 string outfile = getRootName(outputFile) + "sorted.dist.temp";
 671
 672                 //use the unix sort
 673                 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
 674                         string command = "sort -n " + outputFile + " -o " + outfile;
 675                         system(command.c_str());
 676                 #else //sort using windows sort
 677                         string command = "sort " + outputFile + " /O " + outfile;
 678                         system(command.c_str());
 679                 #endif
 680
 681
 682                 //output to new file distance for each row and save positions in file where new row begins
 683                 ifstream in;
 684                 openInputFile(outfile, in);
 685
 686                 ofstream out;
 687                 openOutputFile(outputFile, out);
 688
 689                 out.setf(ios::fixed, ios::floatfield); out.setf(ios::showpoint);
 690
 691                 out << alignDB.getNumSeqs() << endl;
 692
 693                 //get first currentRow
 694                 string first, currentRow, second;
 695                 float dist;
 696                 int i, j;
 697                 i = 0; j = 0;
 698                 map<string, float> rowDists; //take advantage of the fact that maps are already sorted by key
 699                 map<string, float>::iterator it;
 700
 701                 in >> first;
 702                 currentRow = first;
 703
 704                 rowDists[first] = 0.00; //distance to yourself is 0.0
 705
 706                 in.seekg(0);
 707                 //openInputFile(outfile, in);
 708
 709                 while(!in.eof()) {
 710                         if (m->control_pressed) { in.close(); remove(outfile.c_str()); out.close(); return 0; }
 711
 712                         in >> first >> second >> dist; gobble(in);
 713
 714                         if (first != currentRow) {
 715                                 //print out last row
 716                                 out << currentRow << '\t'; //print name
 717
 718                                 //print dists
 719                                 for (it = rowDists.begin(); it != rowDists.end(); it++) {
 720                                         if (j >= i) { break; }
 721                                         out << it->second << '\t';
 722                                         j++;
 723                                 }
 724                                 out << endl;
 725
 726                                 //start new row
 727                                 currentRow = first;
 728                                 rowDists.clear();
 729                                 rowDists[first] = 0.00;
 730                                 rowDists[second] = dist;
 731                                 j = 0;
 732                                 i++;
 733                         }else{
 734                                 rowDists[second] = dist;
 735                         }
 736                 }
 737                 //print out last row
 738                 out << currentRow << '\t'; //print name
 739
 740                 //print dists
 741                 for (it = rowDists.begin(); it != rowDists.end(); it++) {
 742                         out << it->second << '\t';
 743                 }
 744                 out << endl;
 745
 746                 in.close();
 747                 out.close();
 748
 749                 remove(outfile.c_str());
 750
 751                 return 1;
 752
 753         }
 754         catch(exception& e) {
 755                 m->errorOut(e, "DistanceCommand", "convertToLowerTriangle");
 756                 exit(1);
 757         }
 758 }
 759 /**************************************************************************************************
 760 void DistanceCommand::appendFiles(string temp, string filename) {
 761         try{
 762                 ofstream output;
 763                 ifstream input;
 764
 765                 //open output file in append mode
 766                 openOutputFileAppend(filename, output);
 767                 openInputFile(temp, input);
 768
 769                 while(char c = input.get()){
 770                         if(input.eof())         {       break;                  }
 771                         else                            {       output << c;    }
 772                 }
 773
 774                 input.close();
 775                 output.close();
 776         }
 777         catch(exception& e) {
 778                 m->errorOut(e, "DistanceCommand", "appendFiles");
 779                 exit(1);
 780         }
 781 }
 782 /**************************************************************************************************/
 783
 784