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                         OptionParser parser(option);
  28                         map<string,string> parameters = parser.getParameters();
  29
  30                         ValidParameters validParameter;
  31
  32                         //check to make sure all parameters are valid for command
  33                         for (map<string,string>::iterator 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
  58                         if ((method == "furthest") || (method == "nearest") || (method == "average")) { }
  59                         else {cout << "Not a valid clustering method.  Valid clustering algorithms are furthest, nearest or average." << endl; abort = true; }
  60
  61
  62                         if (abort == false) {
  63
  64                                 //get matrix, list and rabund for execute
  65                                 if(globaldata->gSparseMatrix != NULL)   {       matrix = globaldata->gSparseMatrix;             }
  66
  67                                 if(globaldata->gListVector != NULL){
  68                                         list = globaldata->gListVector;
  69                                         rabund = new RAbundVector(list->getRAbundVector());
  70                                 }
  71
  72                                 //create cluster
  73                                 if(method == "furthest")        {       cluster = new CompleteLinkage(rabund, list, matrix);    tag = "fn";     }
  74                                 else if(method == "nearest"){   cluster = new SingleLinkage(rabund, list, matrix);              tag = "nn";     }
  75                                 else if(method == "average"){   cluster = new AverageLinkage(rabund, list, matrix);             tag = "an";     }
  76                                 else                                            {       cout << "error - not recognized method" << endl;        abort = true;   }
  77
  78                                 fileroot = getRootName(globaldata->inputFileName);
  79
  80                                 openOutputFile(fileroot+ tag + ".sabund",       sabundFile);
  81                                 openOutputFile(fileroot+ tag + ".rabund",       rabundFile);
  82                                 openOutputFile(fileroot+ tag + ".list",         listFile);
  83
  84
  85                         }
  86
  87                 }
  88
  89         }
  90         catch(exception& e) {
  91                 cout << "Standard Error: " << e.what() << " has occurred in the ClusterCommand class Function ClusterCommand. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
  92                 exit(1);
  93         }
  94         catch(...) {
  95                 cout << "An unknown error has occurred in the ClusterCommand class function ClusterCommand. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
  96                 exit(1);
  97         }
  98 }
  99
 100 //**********************************************************************************************************************
 101
 102 void ClusterCommand::help(){
 103         try {
 104                 cout << "The cluster command can only be executed after a successful read.dist command." << "\n";
 105                 cout << "The cluster command parameter options are method, cuttoff and precision. No parameters are required." << "\n";
 106                 cout << "The cluster command should be in the following format: " << "\n";
 107                 cout << "cluster(method=yourMethod, cutoff=yourCutoff, precision=yourPrecision) " << "\n";
 108                 cout << "The acceptable cluster methods are furthest, nearest and average.  If no method is provided then furthest is assumed." << "\n" << "\n";
 109         }
 110         catch(exception& e) {
 111                 cout << "Standard Error: " << e.what() << " has occurred in the ClusterCommand class Function help. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
 112                 exit(1);
 113         }
 114         catch(...) {
 115                 cout << "An unknown error has occurred in the ClusterCommand class function help. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
 116                 exit(1);
 117         }
 118 }
 119
 120 //**********************************************************************************************************************
 121
 122 ClusterCommand::~ClusterCommand(){
 123         if (abort == false) {
 124                 delete cluster;
 125                 delete rabund;
 126         }
 127 }
 128
 129 //**********************************************************************************************************************
 130
 131 int ClusterCommand::execute(){
 132         try {
 133
 134                 if (abort == true) {    return 0;       }
 135
 136                 float previousDist = 0.00000;
 137                 float rndPreviousDist = 0.00000;
 138                 oldRAbund = *rabund;
 139                 oldList = *list;
 140
 141                 float x;
 142                 x=0.1;
 143                 toString(x, 2);
 144
 145                 while(matrix->getSmallDist() < cutoff && matrix->getNNodes() > 0){
 146                         cluster->update();
 147                         float dist = matrix->getSmallDist();
 148                         float rndDist = roundDist(dist, precision);
 149
 150                         if(previousDist <= 0.0000 && dist != previousDist){
 151                                 printData("unique");
 152                         }
 153                         else if(rndDist != rndPreviousDist){
 154                                 printData(toString(rndPreviousDist,  length-1));
 155                         }
 156
 157                         previousDist = dist;
 158                         rndPreviousDist = rndDist;
 159                         oldRAbund = *rabund;
 160                         oldList = *list;
 161                 }
 162
 163                 if(previousDist <= 0.0000){
 164                         printData("unique");
 165                 }
 166                 else if(rndPreviousDist<cutoff){
 167                         printData(toString(rndPreviousDist, length-1));
 168                 }
 169
 170                 //delete globaldata's copy of the sparsematrix and listvector to free up memory
 171                 delete globaldata->gSparseMatrix;  globaldata->gSparseMatrix = NULL;
 172                 delete globaldata->gListVector;  globaldata->gListVector = NULL;
 173
 174                 //saves .list file so you can do the collect, rarefaction and summary commands without doing a read.list
 175                 if (globaldata->getFormat() == "phylip") { globaldata->setPhylipFile(""); }
 176                 else if (globaldata->getFormat() == "column") { globaldata->setColumnFile(""); }
 177
 178                 globaldata->setListFile(fileroot+ tag + ".list");
 179                 globaldata->setNameFile("");
 180                 globaldata->setFormat("list");
 181
 182                 sabundFile.close();
 183                 rabundFile.close();
 184                 listFile.close();
 185
 186                 return 0;
 187         }
 188         catch(exception& e) {
 189                 cout << "Standard Error: " << e.what() << " has occurred in the ClusterCommand class Function execute. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
 190                 exit(1);
 191         }
 192         catch(...) {
 193                 cout << "An unknown error has occurred in the ClusterCommand class function execute. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
 194                 exit(1);
 195         }
 196
 197 }
 198
 199 //**********************************************************************************************************************
 200
 201 void ClusterCommand::printData(string label){
 202         try {
 203                 oldRAbund.setLabel(label);
 204                 oldRAbund.getSAbundVector().print(cout);
 205                 oldRAbund.print(rabundFile);
 206                 oldRAbund.getSAbundVector().print(sabundFile);
 207
 208                 oldList.setLabel(label);
 209                 oldList.print(listFile);
 210         }
 211         catch(exception& e) {
 212                 cout << "Standard Error: " << e.what() << " has occurred in the ClusterCommand class Function printData. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
 213                 exit(1);
 214         }
 215         catch(...) {
 216                 cout << "An unknown error has occurred in the ClusterCommand class function printData. Please contact Pat Schloss at pschloss@microbio.umass.edu." << "\n";
 217                 exit(1);
 218         }
 219
 220 }
 221 //**********************************************************************************************************************