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","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, "cutoff", false);
  60                         if (temp == "not found") { temp = "10"; }
  61                         convert(temp, cutoff);
  62                         cutoff += (5 / (precision * 10.0));
  63
  64                         method = validParameter.validFile(parameters, "method", false);
  65                         if (method == "not found") { method = "furthest"; }
  66
  67                         if ((method == "furthest") || (method == "nearest") || (method == "average")) { }
  68                         else { m->mothurOut("Not a valid clustering method.  Valid clustering algorithms are furthest, nearest or average."); m->mothurOutEndLine(); abort = true; }
  69
  70                         showabund = validParameter.validFile(parameters, "showabund", false);
  71                         if (showabund == "not found") { showabund = "T"; }
  72
  73                         timing = validParameter.validFile(parameters, "timing", false);
  74                         if (timing == "not found") { timing = "F"; }
  75
  76                         if (abort == false) {
  77
  78
  79                                                         //get matrix, list and rabund for execute
  80                                 if(globaldata->gSparseMatrix != NULL)   {       matrix = globaldata->gSparseMatrix;             }
  81
  82                                 if(globaldata->gListVector != NULL){
  83                                         list = globaldata->gListVector;
  84                                         rabund = new RAbundVector(list->getRAbundVector());
  85                                 }
  86
  87                                 //create cluster
  88                                 if (method == "furthest")       {       cluster = new CompleteLinkage(rabund, list, matrix, cutoff, method); }
  89                                 else if(method == "nearest"){   cluster = new SingleLinkage(rabund, list, matrix, cutoff, method); }
  90                                 else if(method == "average"){   cluster = new AverageLinkage(rabund, list, matrix, cutoff, method);     }
  91                                 tag = cluster->getTag();
  92
  93                                 if (outputDir == "") { outputDir += hasPath(globaldata->inputFileName); }
  94                                 fileroot = outputDir + getRootName(getSimpleName(globaldata->inputFileName));
  95
  96                                 openOutputFile(fileroot+ tag + ".sabund",       sabundFile);
  97                                 openOutputFile(fileroot+ tag + ".rabund",       rabundFile);
  98                                 openOutputFile(fileroot+ tag + ".list",         listFile);
  99
 100                                 outputNames.push_back(fileroot+ tag + ".sabund");
 101                                 outputNames.push_back(fileroot+ tag + ".rabund");
 102                                 outputNames.push_back(fileroot+ tag + ".list");
 103                         }
 104                 }
 105         }
 106         catch(exception& e) {
 107                 m->errorOut(e, "ClusterCommand", "ClusterCommand");
 108                 exit(1);
 109         }
 110 }
 111
 112 //**********************************************************************************************************************
 113
 114 void ClusterCommand::help(){
 115         try {
 116                 m->mothurOut("The cluster command can only be executed after a successful read.dist command.\n");
 117                 m->mothurOut("The cluster command parameter options are method, cuttoff, precision, showabund and timing. No parameters are required.\n");
 118                 m->mothurOut("The cluster command should be in the following format: \n");
 119                 m->mothurOut("cluster(method=yourMethod, cutoff=yourCutoff, precision=yourPrecision) \n");
 120                 m->mothurOut("The acceptable cluster methods are furthest, nearest and average.  If no method is provided then furthest is assumed.\n\n");
 121         }
 122         catch(exception& e) {
 123                 m->errorOut(e, "ClusterCommand", "help");
 124                 exit(1);
 125         }
 126 }
 127
 128 //**********************************************************************************************************************
 129
 130 ClusterCommand::~ClusterCommand(){
 131         if (abort == false) {
 132                 delete cluster;
 133                 delete rabund;
 134         }
 135 }
 136
 137 //**********************************************************************************************************************
 138
 139 int ClusterCommand::execute(){
 140         try {
 141
 142                 if (abort == true) {    return 0;       }
 143
 144                 time_t estart = time(NULL);
 145                 //int ndist = matrix->getNNodes();
 146                 float previousDist = 0.00000;
 147                 float rndPreviousDist = 0.00000;
 148                 oldRAbund = *rabund;
 149                 oldList = *list;
 150
 151                 print_start = true;
 152                 start = time(NULL);
 153                 loops = 0;
 154                 double saveCutoff = cutoff;
 155
 156                 while (matrix->getSmallDist() < cutoff && matrix->getNNodes() > 0){
 157                         if (print_start && isTrue(timing)) {
 158                                 m->mothurOut("Clustering (" + tag + ") dist " + toString(matrix->getSmallDist()) + "/"
 159                                         + toString(roundDist(matrix->getSmallDist(), precision))
 160                                         + "\t(precision: " + toString(precision) + ", Nodes: " + toString(matrix->getNNodes()) + ")");
 161                                 cout.flush();
 162                                 print_start = false;
 163                         }
 164
 165                         loops++;
 166
 167                         cluster->update(cutoff);
 168                         float dist = matrix->getSmallDist();
 169                         float rndDist = roundDist(dist, precision);
 170
 171                         if(previousDist <= 0.0000 && dist != previousDist){
 172                                 printData("unique");
 173                         }
 174                         else if(rndDist != rndPreviousDist){
 175                                 printData(toString(rndPreviousDist,  length-1));
 176                         }
 177
 178                         previousDist = dist;
 179                         rndPreviousDist = rndDist;
 180                         oldRAbund = *rabund;
 181                         oldList = *list;
 182                 }
 183
 184                 if (print_start && isTrue(timing)) {
 185                         m->mothurOut("Clustering (" + tag + ") for distance " + toString(previousDist) + "/" + toString(rndPreviousDist)
 186                                          + "\t(precision: " + toString(precision) + ", Nodes: " + toString(matrix->getNNodes()) + ")");
 187                         cout.flush();
 188                         print_start = false;
 189                 }
 190
 191                 if(previousDist <= 0.0000){
 192                         printData("unique");
 193                 }
 194                 else if(rndPreviousDist<cutoff){
 195                         printData(toString(rndPreviousDist, length-1));
 196                 }
 197
 198                 //delete globaldata's copy of the sparsematrix and listvector to free up memory
 199                 delete globaldata->gSparseMatrix;  globaldata->gSparseMatrix = NULL;
 200                 delete globaldata->gListVector;  globaldata->gListVector = NULL;
 201
 202                 //saves .list file so you can do the collect, rarefaction and summary commands without doing a read.list
 203                 if (globaldata->getFormat() == "phylip") { globaldata->setPhylipFile(""); }
 204                 else if (globaldata->getFormat() == "column") { globaldata->setColumnFile(""); }
 205
 206                 globaldata->setListFile(fileroot+ tag + ".list");
 207                 globaldata->setNameFile("");
 208                 globaldata->setFormat("list");
 209
 210                 sabundFile.close();
 211                 rabundFile.close();
 212                 listFile.close();
 213
 214                 if (saveCutoff != cutoff) { m->mothurOut("changed cutoff to " + toString(cutoff)); m->mothurOutEndLine();  }
 215
 216                 m->mothurOutEndLine();
 217                 m->mothurOut("Output File Names: "); m->mothurOutEndLine();
 218                 for (int i = 0; i < outputNames.size(); i++) {  m->mothurOut(outputNames[i]); m->mothurOutEndLine();    }
 219                 m->mothurOutEndLine();
 220
 221
 222                 //if (isTrue(timing)) {
 223                         m->mothurOut("It took " + toString(time(NULL) - estart) + " seconds to cluster"); m->mothurOutEndLine();
 224                 //}
 225
 226
 227                 return 0;
 228         }
 229         catch(exception& e) {
 230                 m->errorOut(e, "ClusterCommand", "execute");
 231                 exit(1);
 232         }
 233 }
 234
 235 //**********************************************************************************************************************
 236
 237 void ClusterCommand::printData(string label){
 238         try {
 239                 if (isTrue(timing)) {
 240                         m->mothurOut("\tTime: " + toString(time(NULL) - start) + "\tsecs for " + toString(oldRAbund.getNumBins())
 241                      + "\tclusters. Updates: " + toString(loops)); m->mothurOutEndLine();
 242                 }
 243                 print_start = true;
 244                 loops = 0;
 245                 start = time(NULL);
 246
 247                 oldRAbund.setLabel(label);
 248                 if (isTrue(showabund)) {
 249                         oldRAbund.getSAbundVector().print(cout);
 250                 }
 251                 oldRAbund.print(rabundFile);
 252                 oldRAbund.getSAbundVector().print(sabundFile);
 253
 254                 oldList.setLabel(label);
 255                 oldList.print(listFile);
 256         }
 257         catch(exception& e) {
 258                 m->errorOut(e, "ClusterCommand", "printData");
 259                 exit(1);
 260         }
 261
 262
 263 }
 264 //**********************************************************************************************************************