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                 abort = false;
  18
  19                 //allow user to run help
  20                 if(option == "help") { help(); abort = true; }
  21
  22                 else {
  23                         //valid paramters for this command
  24                         string Array[] =  {"cutoff","precision","method"};
  25                         vector<string> myArray (Array, Array+(sizeof(Array)/sizeof(string)));
  26
  27                         parser = new OptionParser();
  28                         parser->parse(option, parameters);  delete parser;
  29
  30                         ValidParameters* validParameter = new ValidParameters();
  31
  32                         //check to make sure all parameters are valid for command
  33                         for (it = parameters.begin(); it != parameters.end(); it++) {
  34                                 if (validParameter->isValidParameter(it->first, myArray, it->second) != true) {  abort = true;  }
  35                         }
  36
  37                         //error checking to make sure they read a distance file
  38                         if ((globaldata->gSparseMatrix == NULL) || (globaldata->gListVector == NULL)) {
  39                                 cout << "Before you use the cluster command, you first need to read in a distance matrix." << endl;  abort = true;
  40                         }
  41
  42                         //check for optional parameter and set defaults
  43                         // ...at some point should added some additional type checking...
  44                         //get user cutoff and precision or use defaults
  45                         string temp;
  46                         temp = validParameter->validFile(parameters, "precision", false);               if (temp == "not found") { temp = "100"; }
  47                         //saves precision legnth for formatting below
  48                         length = temp.length();
  49                         convert(temp, precision);
  50
  51                         temp = validParameter->validFile(parameters, "cutoff", false);                  if (temp == "not found") { temp = "10"; }
  52                         convert(temp, cutoff);
  53                         cutoff += (5 / (precision * 10.0));
  54
  55                         method = validParameter->validFile(parameters, "method", false);                        if (method == "not found") { method = "furthest"; }
  56
  57                         delete validParameter;
  58
  59                         if ((method == "furthest") || (method == "nearest") || (method == "average")) { }
  60                         else {cout << "Not a valid clustering method.  Valid clustering algorithms are furthest, nearest or average." << endl; abort = true; }
  61
  62
  63                         if (abort == false) {
  64
  65                                 //get matrix, list and rabund for execute
  66                                 if(globaldata->gSparseMatrix != NULL)   {       matrix = new SparseMatrix(*globaldata->gSparseMatrix);          }
  67
  68                                 if(globaldata->gListVector != NULL){
  69                                         list = new ListVector(*globaldata->gListVector);
  70                                         rabund = new RAbundVector(list->getRAbundVector());
  71                                 }
  72
  73                                 //create cluster
  74                                 if(method == "furthest")        {       cluster = new CompleteLinkage(rabund, list, matrix);    tag = "fn";     }
  75                                 else if(method == "nearest"){   cluster = new SingleLinkage(rabund, list, matrix);              tag = "nn";     }
  76                                 else if(method == "average"){   cluster = new AverageLinkage(rabund, list, matrix);             tag = "an";     }
  77                                 else                                            {       cout << "error - not recognized method" << endl;        abort = true;   }
  78
  79                                 fileroot = getRootName(globaldata->inputFileName);
  80
  81                                 openOutputFile(fileroot+ tag + ".sabund",       sabundFile);
  82                                 openOutputFile(fileroot+ tag + ".rabund",       rabundFile);
  83                                 openOutputFile(fileroot+ tag + ".list",         listFile);
  84
  85
  86                         }
  87
  88                 }
  89
  90         }
  91         catch(exception& e) {
  92                 cout << "Standard Error: " << e.what() << " has occurred in the ClusterCommand class Function ClusterCommand. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
  93                 exit(1);
  94         }
  95         catch(...) {
  96                 cout << "An unknown error has occurred in the ClusterCommand class function ClusterCommand. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
  97                 exit(1);
  98         }
  99 }
 100
 101 //**********************************************************************************************************************
 102
 103 void ClusterCommand::help(){
 104         try {
 105                 cout << "The cluster command can only be executed after a successful read.dist command." << "\n";
 106                 cout << "The cluster command parameter options are method, cuttoff and precision. No parameters are required." << "\n";
 107                 cout << "The cluster command should be in the following format: " << "\n";
 108                 cout << "cluster(method=yourMethod, cutoff=yourCutoff, precision=yourPrecision) " << "\n";
 109                 cout << "The acceptable cluster methods are furthest, nearest and average.  If no method is provided then furthest is assumed." << "\n" << "\n";
 110         }
 111         catch(exception& e) {
 112                 cout << "Standard Error: " << e.what() << " has occurred in the ClusterCommand class Function help. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
 113                 exit(1);
 114         }
 115         catch(...) {
 116                 cout << "An unknown error has occurred in the ClusterCommand class function help. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
 117                 exit(1);
 118         }
 119 }
 120
 121 //**********************************************************************************************************************
 122
 123 ClusterCommand::~ClusterCommand(){
 124         delete cluster;
 125         delete matrix;
 126         delete list;
 127         delete rabund;
 128 }
 129
 130 //**********************************************************************************************************************
 131
 132 int ClusterCommand::execute(){
 133         try {
 134
 135                 if (abort == true) {    return 0;       }
 136
 137                 float previousDist = 0.00000;
 138                 float rndPreviousDist = 0.00000;
 139                 oldRAbund = *rabund;
 140                 oldList = *list;
 141
 142                 float x;
 143                 x=0.1;
 144                 toString(x, 2);
 145
 146                 while(matrix->getSmallDist() < cutoff && matrix->getNNodes() > 0){
 147                         cluster->update();
 148                         float dist = matrix->getSmallDist();
 149                         float rndDist = roundDist(dist, precision);
 150
 151                         if(previousDist <= 0.0000 && dist != previousDist){
 152                                 printData("unique");
 153                         }
 154                         else if(rndDist != rndPreviousDist){
 155                                 printData(toString(rndPreviousDist,  length-1));
 156                         }
 157
 158                         previousDist = dist;
 159                         rndPreviousDist = rndDist;
 160                         oldRAbund = *rabund;
 161                         oldList = *list;
 162                 }
 163
 164                 if(previousDist <= 0.0000){
 165                         printData("unique");
 166                 }
 167                 else if(rndPreviousDist<cutoff){
 168                         printData(toString(rndPreviousDist, length-1));
 169                 }
 170
 171                 //delete globaldata's copy of the sparsematrix and listvector to free up memory
 172                 delete globaldata->gSparseMatrix;  globaldata->gSparseMatrix = NULL;
 173                 delete globaldata->gListVector;  globaldata->gListVector = NULL;
 174
 175                 //saves .list file so you can do the collect, rarefaction and summary commands without doing a read.list
 176                 if (globaldata->getFormat() == "phylip") { globaldata->setPhylipFile(""); }
 177                 else if (globaldata->getFormat() == "column") { globaldata->setColumnFile(""); }
 178
 179                 globaldata->setListFile(fileroot+ tag + ".list");
 180                 globaldata->setNameFile("");
 181                 globaldata->setFormat("list");
 182
 183                 return 0;
 184         }
 185         catch(exception& e) {
 186                 cout << "Standard Error: " << e.what() << " has occurred in the ClusterCommand class Function execute. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
 187                 exit(1);
 188         }
 189         catch(...) {
 190                 cout << "An unknown error has occurred in the ClusterCommand class function execute. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
 191                 exit(1);
 192         }
 193
 194 }
 195
 196 //**********************************************************************************************************************
 197
 198 void ClusterCommand::printData(string label){
 199         try {
 200                 oldRAbund.setLabel(label);
 201                 oldRAbund.getSAbundVector().print(cout);
 202                 oldRAbund.print(rabundFile);
 203                 oldRAbund.getSAbundVector().print(sabundFile);
 204
 205                 oldList.setLabel(label);
 206                 oldList.print(listFile);
 207         }
 208         catch(exception& e) {
 209                 cout << "Standard Error: " << e.what() << " has occurred in the ClusterCommand class Function printData. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
 210                 exit(1);
 211         }
 212         catch(...) {
 213                 cout << "An unknown error has occurred in the ClusterCommand class function printData. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
 214                 exit(1);
 215         }
 216
 217 }
 218 //**********************************************************************************************************************