hclustercommand.cpp

   1 /*
   2  *  hclustercommand.cpp
   3  *  Mothur
   4  *
   5  *  Created by westcott on 10/13/09.
   6  *  Copyright 2009 Schloss Lab. All rights reserved.
   7  *
   8  */
   9
  10 #include "hclustercommand.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 HClusterCommand::HClusterCommand(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","phylip","column","name","sorted","showabund","timing"};
  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) {
  35                                         abort = true;
  36                                 }
  37                         }
  38
  39                         globaldata->newRead();
  40
  41                         //check for required parameters
  42                         phylipfile = validParameter.validFile(parameters, "phylip", true);
  43                         if (phylipfile == "not open") { abort = true; }
  44                         else if (phylipfile == "not found") { phylipfile = ""; }
  45                         else {  distfile = phylipfile;  format = "phylip";      }
  46
  47                         columnfile = validParameter.validFile(parameters, "column", true);
  48                         if (columnfile == "not open") { abort = true; }
  49                         else if (columnfile == "not found") { columnfile = ""; }
  50                         else {  distfile = columnfile; format = "column";       }
  51
  52                         namefile = validParameter.validFile(parameters, "name", true);
  53                         if (namefile == "not open") { abort = true; }
  54                         else if (namefile == "not found") { namefile = ""; }
  55
  56                         if ((phylipfile == "") && (columnfile == "")) { mothurOut("When executing a hcluster command you must enter a phylip or a column."); mothurOutEndLine(); abort = true; }
  57                         else if ((phylipfile != "") && (columnfile != "")) { mothurOut("When executing a hcluster command you must enter ONLY ONE of the following: phylip or column."); mothurOutEndLine(); abort = true; }
  58
  59                         if (columnfile != "") {
  60                                 if (namefile == "") {  cout << "You need to provide a namefile if you are going to use the column format." << endl; abort = true; }
  61                         }
  62
  63                         //check for optional parameter and set defaults
  64                         // ...at some point should added some additional type checking...
  65                         //get user cutoff and precision or use defaults
  66                         string temp;
  67                         temp = validParameter.validFile(parameters, "precision", false);
  68                         if (temp == "not found") { temp = "100"; }
  69                         //saves precision legnth for formatting below
  70                         length = temp.length();
  71                         convert(temp, precision);
  72
  73                         temp = validParameter.validFile(parameters, "cutoff", false);
  74                         if (temp == "not found") { temp = "10"; }
  75                         convert(temp, cutoff);
  76                         cutoff += (5 / (precision * 10.0));
  77
  78                         method = validParameter.validFile(parameters, "method", false);
  79                         if (method == "not found") { method = "furthest"; }
  80
  81                         if ((method == "furthest") || (method == "nearest") || (method == "average")) { }
  82                         else { mothurOut("Not a valid clustering method.  Valid clustering algorithms are furthest, nearest or average."); mothurOutEndLine(); abort = true; }
  83
  84                         showabund = validParameter.validFile(parameters, "showabund", false);
  85                         if (showabund == "not found") { showabund = "T"; }
  86
  87                         sort = validParameter.validFile(parameters, "sorted", false);
  88                         if (sort == "not found") { sort = "F"; }
  89                         sorted = isTrue(sort);
  90
  91                         timing = validParameter.validFile(parameters, "timing", false);
  92                         if (timing == "not found") { timing = "F"; }
  93
  94
  95                         if (abort == false) {
  96
  97                                 fileroot = getRootName(distfile);
  98
  99                                 if (method == "furthest")               { tag = "fn";  }
 100                                 else if (method == "nearest")   { tag = "nn";  }
 101                                 else                                                    { tag = "an";  }
 102
 103                                 openOutputFile(fileroot+ tag + ".sabund",       sabundFile);
 104                                 openOutputFile(fileroot+ tag + ".rabund",       rabundFile);
 105                                 openOutputFile(fileroot+ tag + ".list",         listFile);
 106                         }
 107                 }
 108         }
 109         catch(exception& e) {
 110                 errorOut(e, "HClusterCommand", "HClusterCommand");
 111                 exit(1);
 112         }
 113 }
 114
 115 //**********************************************************************************************************************
 116
 117 void HClusterCommand::help(){
 118         try {
 119                 mothurOut("The hcluster command parameter options are cutoff, precision, method, phylip, column, name, showabund, timing and sorted. Phylip or column and name are required.\n");
 120                 mothurOut("The phylip and column parameter allow you to enter your distance file, and sorted indicates whether your column distance file is already sorted. \n");
 121                 mothurOut("The name parameter allows you to enter your name file and is required if your distance file is in column format. \n");
 122                 mothurOut("The hcluster command should be in the following format: \n");
 123                 mothurOut("hcluster(column=youDistanceFile, name=yourNameFile, method=yourMethod, cutoff=yourCutoff, precision=yourPrecision) \n");
 124                 mothurOut("The acceptable hcluster methods are furthest and nearest, but we hope to add average in the future.\n\n");
 125         }
 126         catch(exception& e) {
 127                 errorOut(e, "HClusterCommand", "help");
 128                 exit(1);
 129         }
 130 }
 131
 132 //**********************************************************************************************************************
 133
 134 HClusterCommand::~HClusterCommand(){}
 135
 136 //**********************************************************************************************************************
 137
 138 int HClusterCommand::execute(){
 139         try {
 140
 141                 if (abort == true) {    return 0;       }
 142
 143                 if(namefile != ""){
 144                         globaldata->nameMap = new NameAssignment(namefile);
 145                         globaldata->nameMap->readMap();
 146                 }else{
 147                         globaldata->nameMap = NULL;
 148                 }
 149
 150                 time_t estart = time(NULL);
 151
 152                 if (!sorted) {
 153                         read = new ReadCluster(distfile, cutoff);
 154                         read->setFormat(format);
 155                         read->read(globaldata->nameMap);
 156                         distfile = read->getOutputFile();
 157
 158                         list = read->getListVector();
 159                         delete read;
 160                 }else {
 161                         list = new ListVector(globaldata->nameMap->getListVector());
 162                 }
 163
 164                 mothurOut("It took " + toString(time(NULL) - estart) + " seconds to sort. "); mothurOutEndLine();
 165                 estart = time(NULL);
 166
 167                 //list vector made by read contains all sequence names
 168                 if(list != NULL){
 169                         rabund = new RAbundVector(list->getRAbundVector());
 170                 }else{
 171                         mothurOut("Error: no list vector!"); mothurOutEndLine(); return 0;
 172                 }
 173
 174                 float previousDist = 0.00000;
 175                 float rndPreviousDist = 0.00000;
 176                 oldRAbund = *rabund;
 177                 oldList = *list;
 178
 179                 print_start = true;
 180                 start = time(NULL);
 181
 182                 ifstream in;
 183                 openInputFile(distfile, in);
 184                 string firstName, secondName;
 185                 float distance;
 186
 187                 cluster = new HCluster(rabund, list, method);
 188                 vector<seqDist> seqs; seqs.resize(1); // to start loop
 189
 190                 while (seqs.size() != 0){
 191
 192                         seqs = cluster->getSeqs(in, globaldata->nameMap, cutoff);
 193
 194                         for (int i = 0; i < seqs.size(); i++) {  //-1 means skip me
 195
 196                                 if (print_start && isTrue(timing)) {
 197                                         mothurOut("Clustering (" + tag + ") dist " + toString(distance) + "/"
 198                                                           + toString(roundDist(distance, precision))
 199                                                           + "\t(precision: " + toString(precision) + ")");
 200                                         cout.flush();
 201                                         print_start = false;
 202                                 }
 203
 204
 205                                 if (seqs[i].seq1 != seqs[i].seq2) {
 206                                         cluster->update(seqs[i].seq1, seqs[i].seq2, seqs[i].dist);
 207
 208                                         float rndDist = roundDist(seqs[i].dist, precision);
 209
 210                                         if((previousDist <= 0.0000) && (seqs[i].dist != previousDist)){
 211                                                 printData("unique");
 212                                         }
 213                                         else if((rndDist != rndPreviousDist)){
 214                                                 printData(toString(rndPreviousDist,  length-1));
 215                                         }
 216
 217                                         previousDist = seqs[i].dist;
 218                                         rndPreviousDist = rndDist;
 219                                         oldRAbund = *rabund;
 220                                         oldList = *list;
 221                                 }
 222                         }
 223                 }
 224
 225                 in.close();
 226
 227                 if (print_start && isTrue(timing)) {
 228                         //mothurOut("Clustering (" + tag + ") for distance " + toString(previousDist) + "/" + toString(rndPreviousDist)
 229                                          //+ "\t(precision: " + toString(precision) + ", Nodes: " + toString(matrix->getNNodes()) + ")");
 230                         cout.flush();
 231                         print_start = false;
 232                 }
 233
 234                 if(previousDist <= 0.0000){
 235                         printData("unique");
 236                 }
 237                 else if(rndPreviousDist<cutoff){
 238                         printData(toString(rndPreviousDist, length-1));
 239                 }
 240
 241                 //delete globaldata's copy of the sparsematrix and listvector to free up memory
 242                 delete globaldata->gListVector;  globaldata->gListVector = NULL;
 243
 244                 //saves .list file so you can do the collect, rarefaction and summary commands without doing a read.list
 245                 if (globaldata->getFormat() == "phylip") { globaldata->setPhylipFile(""); }
 246                 else if (globaldata->getFormat() == "column") { globaldata->setColumnFile(""); }
 247
 248                 globaldata->setListFile(fileroot+ tag + ".list");
 249                 globaldata->setNameFile("");
 250                 globaldata->setFormat("list");
 251
 252                 sabundFile.close();
 253                 rabundFile.close();
 254                 listFile.close();
 255                 delete cluster;
 256
 257                 mothurOut("It took " + toString(time(NULL) - estart) + " seconds to cluster. "); mothurOutEndLine();
 258
 259                 return 0;
 260         }
 261         catch(exception& e) {
 262                 errorOut(e, "HClusterCommand", "execute");
 263                 exit(1);
 264         }
 265 }
 266
 267 //**********************************************************************************************************************
 268
 269 void HClusterCommand::printData(string label){
 270         try {
 271                 if (isTrue(timing)) {
 272                         mothurOut("\tTime: " + toString(time(NULL) - start) + "\tsecs for " + toString(oldRAbund.getNumBins())
 273                      + "\tclusters. Updates: " + toString(loops)); mothurOutEndLine();
 274                 }
 275                 print_start = true;
 276                 loops = 0;
 277                 start = time(NULL);
 278
 279                 oldRAbund.setLabel(label);
 280                 if (isTrue(showabund)) {
 281                         oldRAbund.getSAbundVector().print(cout);
 282                 }
 283                 oldRAbund.print(rabundFile);
 284                 oldRAbund.getSAbundVector().print(sabundFile);
 285
 286                 oldList.setLabel(label);
 287                 oldList.print(listFile);
 288         }
 289         catch(exception& e) {
 290                 errorOut(e, "HClusterCommand", "printData");
 291                 exit(1);
 292         }
 293
 294
 295 }
 296 //**********************************************************************************************************************
 297