clustercommand.cpp

   1 /*
   2  *  clustercommand.cpp
   3  *  Dotur
   4  *
   5  *  Created by Sarah Westcott on 1/2/09.
   6  *  Copyright 2009 Schloss Lab UMASS Amherst. All rights reserved.
   7  *
   8  */
   9
  10 #include "clustercommand.h"
  11
  12 //**********************************************************************************************************************
  13 //This function checks to make sure the cluster command has no errors and then clusters based on the method chosen.
  14 ClusterCommand::ClusterCommand(string option)  {
  15         try{
  16                 globaldata = GlobalData::getInstance();
  17
  18                 abort = false;
  19
  20                 //allow user to run help
  21                 if(option == "help") { help(); abort = true; }
  22
  23                 else {
  24                         //valid paramters for this command
  25                         string Array[] =  {"cutoff","precision","method","showabund","timing","hard","outputdir","inputdir"};
  26                         vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
  27
  28                         OptionParser parser(option);
  29                         map<string,string> parameters = parser.getParameters();
  30
  31                         ValidParameters validParameter;
  32
  33                         //check to make sure all parameters are valid for command
  34                         for (map<string,string>::iterator it = parameters.begin(); it != parameters.end(); it++) {
  35                                 if (validParameter.isValidParameter(it->first, myArray, it->second) != true) {
  36                                         abort = true;
  37                                 }
  38                         }
  39
  40                         //if the user changes the output directory command factory will send this info to us in the output parameter
  41                         outputDir = validParameter.validFile(parameters, "outputdir", false);           if (outputDir == "not found"){  outputDir = "";         }
  42
  43                         //error checking to make sure they read a distance file
  44                         if ((globaldata->gSparseMatrix == NULL) || (globaldata->gListVector == NULL)) {
  45                                 m->mothurOut("Before you use the cluster command, you first need to read in a distance matrix."); m->mothurOutEndLine();
  46                                 abort = true;
  47                         }
  48
  49                         //check for optional parameter and set defaults
  50                         // ...at some point should added some additional type checking...
  51                         //get user cutoff and precision or use defaults
  52                         string temp;
  53                         temp = validParameter.validFile(parameters, "precision", false);
  54                         if (temp == "not found") { temp = "100"; }
  55                         //saves precision legnth for formatting below
  56                         length = temp.length();
  57                         convert(temp, precision);
  58
  59                         temp = validParameter.validFile(parameters, "hard", false);                     if (temp == "not found") { temp = "F"; }
  60                         hard = isTrue(temp);
  61
  62                         temp = validParameter.validFile(parameters, "cutoff", false);
  63                         if (temp == "not found") { temp = "10"; }
  64                         convert(temp, cutoff);
  65                         cutoff += (5 / (precision * 10.0));
  66
  67                         method = validParameter.validFile(parameters, "method", false);
  68                         if (method == "not found") { method = "furthest"; }
  69
  70                         if ((method == "furthest") || (method == "nearest") || (method == "average") || (method == "weighted")) { }
  71                         else { m->mothurOut("Not a valid clustering method.  Valid clustering algorithms are furthest, nearest, average, and weighted."); m->mothurOutEndLine(); abort = true; }
  72
  73                         showabund = validParameter.validFile(parameters, "showabund", false);
  74                         if (showabund == "not found") { showabund = "T"; }
  75
  76                         timing = validParameter.validFile(parameters, "timing", false);
  77                         if (timing == "not found") { timing = "F"; }
  78
  79                         if (abort == false) {
  80
  81
  82                                                         //get matrix, list and rabund for execute
  83                                 if(globaldata->gSparseMatrix != NULL)   {       matrix = globaldata->gSparseMatrix;             }
  84
  85                                 if(globaldata->gListVector != NULL){
  86                                         list = globaldata->gListVector;
  87                                         rabund = new RAbundVector(list->getRAbundVector());
  88                                 }
  89
  90                                 //create cluster
  91                                 if (method == "furthest")       {       cluster = new CompleteLinkage(rabund, list, matrix, cutoff, method); }
  92                                 else if(method == "nearest"){   cluster = new SingleLinkage(rabund, list, matrix, cutoff, method); }
  93                                 else if(method == "average"){   cluster = new AverageLinkage(rabund, list, matrix, cutoff, method);     }
  94                                 else if(method == "weighted"){  cluster = new WeightedLinkage(rabund, list, matrix, cutoff, method);    }
  95                                 tag = cluster->getTag();
  96
  97                                 if (outputDir == "") { outputDir += hasPath(globaldata->inputFileName); }
  98                                 fileroot = outputDir + getRootName(getSimpleName(globaldata->inputFileName));
  99
 100                                 openOutputFile(fileroot+ tag + ".sabund",       sabundFile);
 101                                 openOutputFile(fileroot+ tag + ".rabund",       rabundFile);
 102                                 openOutputFile(fileroot+ tag + ".list",         listFile);
 103
 104                                 outputNames.push_back(fileroot+ tag + ".sabund");
 105                                 outputNames.push_back(fileroot+ tag + ".rabund");
 106                                 outputNames.push_back(fileroot+ tag + ".list");
 107                         }
 108                 }
 109         }
 110         catch(exception& e) {
 111                 m->errorOut(e, "ClusterCommand", "ClusterCommand");
 112                 exit(1);
 113         }
 114 }
 115
 116 //**********************************************************************************************************************
 117
 118 void ClusterCommand::help(){
 119         try {
 120                 m->mothurOut("The cluster command can only be executed after a successful read.dist command.\n");
 121                 m->mothurOut("The cluster command parameter options are method, cuttoff, hard, precision, showabund and timing. No parameters are required.\n");
 122                 m->mothurOut("The cluster command should be in the following format: \n");
 123                 m->mothurOut("cluster(method=yourMethod, cutoff=yourCutoff, precision=yourPrecision) \n");
 124                 m->mothurOut("The acceptable cluster methods are furthest, nearest and average.  If no method is provided then furthest is assumed.\n\n");
 125         }
 126         catch(exception& e) {
 127                 m->errorOut(e, "ClusterCommand", "help");
 128                 exit(1);
 129         }
 130 }
 131
 132 //**********************************************************************************************************************
 133
 134 ClusterCommand::~ClusterCommand(){
 135         if (abort == false) {
 136                 delete cluster;
 137                 delete rabund;
 138         }
 139 }
 140
 141 //**********************************************************************************************************************
 142
 143 int ClusterCommand::execute(){
 144         try {
 145
 146                 if (abort == true) {    return 0;       }
 147
 148                 time_t estart = time(NULL);
 149                 //int ndist = matrix->getNNodes();
 150                 float previousDist = 0.00000;
 151                 float rndPreviousDist = 0.00000;
 152                 oldRAbund = *rabund;
 153                 oldList = *list;
 154
 155                 print_start = true;
 156                 start = time(NULL);
 157                 loops = 0;
 158                 double saveCutoff = cutoff;
 159
 160                 while (matrix->getSmallDist() < cutoff && matrix->getNNodes() > 0){
 161
 162                         if (m->control_pressed) { //clean up
 163                                 delete globaldata->gSparseMatrix;  globaldata->gSparseMatrix = NULL;
 164                                 delete globaldata->gListVector;  globaldata->gListVector = NULL;
 165                                 if (globaldata->getFormat() == "phylip") { globaldata->setPhylipFile(""); }
 166                                 else if (globaldata->getFormat() == "column") { globaldata->setColumnFile(""); }
 167                                 sabundFile.close();rabundFile.close();listFile.close();
 168                                 for (int i = 0; i < outputNames.size(); i++) {  remove(outputNames[i].c_str());         }
 169                                 return 0;
 170                         }
 171
 172                         if (print_start && isTrue(timing)) {
 173                                 m->mothurOut("Clustering (" + tag + ") dist " + toString(matrix->getSmallDist()) + "/"
 174                                         + toString(roundDist(matrix->getSmallDist(), precision))
 175                                         + "\t(precision: " + toString(precision) + ", Nodes: " + toString(matrix->getNNodes()) + ")");
 176                                 cout.flush();
 177                                 print_start = false;
 178                         }
 179
 180                         loops++;
 181
 182                         cluster->update(cutoff);
 183
 184                         float dist = matrix->getSmallDist();
 185                         float rndDist;
 186                         if (hard) {
 187                                 rndDist = ceilDist(dist, precision);
 188                         }else{
 189                                 rndDist = roundDist(dist, precision);
 190                         }
 191
 192                         if(previousDist <= 0.0000 && dist != previousDist){
 193                                 printData("unique");
 194                         }
 195                         else if(rndDist != rndPreviousDist){
 196                                 printData(toString(rndPreviousDist,  length-1));
 197                         }
 198
 199                         previousDist = dist;
 200                         rndPreviousDist = rndDist;
 201                         oldRAbund = *rabund;
 202                         oldList = *list;
 203                 }
 204
 205                 if (print_start && isTrue(timing)) {
 206                         m->mothurOut("Clustering (" + tag + ") for distance " + toString(previousDist) + "/" + toString(rndPreviousDist)
 207                                          + "\t(precision: " + toString(precision) + ", Nodes: " + toString(matrix->getNNodes()) + ")");
 208                         cout.flush();
 209                         print_start = false;
 210                 }
 211
 212                 if(previousDist <= 0.0000){
 213                         printData("unique");
 214                 }
 215                 else if(rndPreviousDist<cutoff){
 216                         printData(toString(rndPreviousDist, length-1));
 217                 }
 218
 219                 //delete globaldata's copy of the sparsematrix and listvector to free up memory
 220                 delete globaldata->gSparseMatrix;  globaldata->gSparseMatrix = NULL;
 221                 delete globaldata->gListVector;  globaldata->gListVector = NULL;
 222
 223                 //saves .list file so you can do the collect, rarefaction and summary commands without doing a read.list
 224                 if (globaldata->getFormat() == "phylip") { globaldata->setPhylipFile(""); }
 225                 else if (globaldata->getFormat() == "column") { globaldata->setColumnFile(""); }
 226
 227                 globaldata->setListFile(fileroot+ tag + ".list");
 228                 globaldata->setNameFile("");
 229                 globaldata->setFormat("list");
 230
 231                 sabundFile.close();
 232                 rabundFile.close();
 233                 listFile.close();
 234
 235                 if (saveCutoff != cutoff) {
 236                         if (hard)       {  saveCutoff = ceilDist(saveCutoff, precision);        }
 237                         else            {       saveCutoff = roundDist(saveCutoff, precision);  }
 238
 239                         m->mothurOut("changed cutoff to " + toString(cutoff)); m->mothurOutEndLine();
 240                 }
 241
 242                 m->mothurOutEndLine();
 243                 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
 244                 for (int i = 0; i < outputNames.size(); i++) {  m->mothurOut(outputNames[i]); m->mothurOutEndLine();    }
 245                 m->mothurOutEndLine();
 246
 247
 248                 //if (isTrue(timing)) {
 249                         m->mothurOut("It took " + toString(time(NULL) - estart) + " seconds to cluster"); m->mothurOutEndLine();
 250                 //}
 251
 252
 253                 return 0;
 254         }
 255         catch(exception& e) {
 256                 m->errorOut(e, "ClusterCommand", "execute");
 257                 exit(1);
 258         }
 259 }
 260
 261 //**********************************************************************************************************************
 262
 263 void ClusterCommand::printData(string label){
 264         try {
 265                 if (isTrue(timing)) {
 266                         m->mothurOut("\tTime: " + toString(time(NULL) - start) + "\tsecs for " + toString(oldRAbund.getNumBins())
 267                      + "\tclusters. Updates: " + toString(loops)); m->mothurOutEndLine();
 268                 }
 269                 print_start = true;
 270                 loops = 0;
 271                 start = time(NULL);
 272
 273                 oldRAbund.setLabel(label);
 274                 if (isTrue(showabund)) {
 275                         oldRAbund.getSAbundVector().print(cout);
 276                 }
 277                 oldRAbund.print(rabundFile);
 278                 oldRAbund.getSAbundVector().print(sabundFile);
 279
 280                 oldList.setLabel(label);
 281                 oldList.print(listFile);
 282         }
 283         catch(exception& e) {
 284                 m->errorOut(e, "ClusterCommand", "printData");
 285                 exit(1);
 286         }
 287
 288
 289 }
 290 //**********************************************************************************************************************