cluster.cpp

   1 /*
   2  *  cluster.cpp
   3  *
   4  *
   5  *  Created by Pat Schloss on 8/14/08.
   6  *  Copyright 2008 Patrick D. Schloss. All rights reserved.
   7  *
   8  */
   9
  10 #include "cluster.hpp"
  11 #include "rabundvector.hpp"
  12 #include "listvector.hpp"
  13 #include "sparsematrix.hpp"
  14
  15 /***********************************************************************/
  16
  17 Cluster::Cluster(RAbundVector* rav, ListVector* lv, SparseMatrix* dm, float c, string f) :
  18 rabund(rav), list(lv), dMatrix(dm), method(f)
  19 {
  20 /*
  21         cout << "sizeof(MatData): " << sizeof(MatData) << endl;
  22         cout << "sizeof(PCell*): " << sizeof(PCell*) << endl;
  23
  24         int nCells = dMatrix->getNNodes();
  25         time_t start = time(NULL);
  26
  27         MatVec matvec = MatVec(nCells);
  28         int i = 0;
  29         for (MatData currentCell = dMatrix->begin(); currentCell != dMatrix->end(); currentCell++) {
  30                 matvec[i++] = currentCell;
  31         }
  32         for (i= matvec.size();i>0;i--) {
  33                 dMatrix->rmCell(matvec[i-1]);
  34         }
  35         MatData it = dMatrix->begin();
  36         while (it != dMatrix->end()) {
  37                 it = dMatrix->rmCell(it);
  38         }
  39         cout << "Time to remove " << nCells << " cells: " << time(NULL) - start << " seconds" << endl;
  40     exit(0);
  41         MatData it = dMatrix->begin();
  42         cout << it->row << "/" << it->column << "/" << it->dist << endl;
  43         dMatrix->rmCell(dMatrix->begin());
  44         cout << it->row << "/" << it->column << "/" << it->dist << endl;
  45         exit(0);
  46 */
  47
  48         // Create a data structure to quickly access the PCell information
  49         // for a certain sequence. It consists of a vector of lists, where
  50         // a list contains pointers (iterators) to the all distances related
  51         // to a certain sequence. The Vector is accessed via the index of a
  52         // sequence in the distance matrix.
  53 //ofstream outtemp;
  54 //string temp = "temp";
  55 //openOutputFile(temp, outtemp);
  56 //cout << lv->size() << endl;
  57         seqVec = vector<MatVec>(lv->size());
  58         for (MatData currentCell = dMatrix->begin(); currentCell != dMatrix->end(); currentCell++) {
  59 //outtemp << currentCell->row << '\t' << currentCell->column  << '\t' << currentCell->dist << endl;
  60                 seqVec[currentCell->row].push_back(currentCell);
  61                 seqVec[currentCell->column].push_back(currentCell);
  62         }
  63 //outtemp.close();
  64         mapWanted = false;  //set to true by mgcluster to speed up overlap merge
  65
  66         //save so you can modify as it changes in average neighbor
  67         cutoff = c;
  68         m = MothurOut::getInstance();
  69 }
  70
  71 /***********************************************************************/
  72
  73 void Cluster::getRowColCells() {
  74         try {
  75                 PCell* smallCell = dMatrix->getSmallestCell();  //find the smallest cell - this routine should probably not be in the SpMat class
  76
  77                 smallRow = smallCell->row;              // get its row
  78                 smallCol = smallCell->column;   // get its column
  79                 smallDist = smallCell->dist;    // get the smallest distance
  80         //cout << "small row = " << smallRow << "small col = " << smallCol << "small dist = " << smallDist << endl;
  81
  82                 rowCells = seqVec[smallRow];    // all distances related to the row index
  83                 colCells = seqVec[smallCol];    // all distances related to the column index
  84                 nRowCells = rowCells.size();
  85                 nColCells = colCells.size();
  86 //cout << "num rows = " << nRowCells << "num col = " << nColCells << endl;
  87
  88                 //for (int i = 0; i < nColCells; i++) { cout << colCells[i]->row << '\t' << colCells[i]->column << endl;  }
  89                 //for (int i = 0; i < nRowCells; i++) { cout << rowCells[i]->row << '\t' << rowCells[i]->column << endl;  }
  90         }
  91         catch(exception& e) {
  92                 m->errorOut(e, "Cluster", "getRowColCells");
  93                 exit(1);
  94         }
  95
  96 }
  97 /***********************************************************************/
  98 // Remove the specified cell from the seqVec and from the sparse
  99 // matrix
 100 void Cluster::removeCell(const MatData& cell, int vrow, int vcol, bool rmMatrix){
 101         try {
 102
 103                 ull drow = cell->row;
 104                         ull dcol = cell->column;
 105                         if (((vrow >=0) && (drow != smallRow)) ||
 106                                 ((vcol >=0) && (dcol != smallCol))) {
 107                                 ull dtemp = drow;
 108                                 drow = dcol;
 109                                 dcol = dtemp;
 110                         }
 111
 112                         ull crow;
 113                         ull ccol;
 114                         int nCells;
 115                         if (vrow < 0) {
 116                                 nCells = seqVec[drow].size();
 117                                 for (vrow=0; vrow<nCells;vrow++) {
 118                                         crow = seqVec[drow][vrow]->row;
 119                                         ccol = seqVec[drow][vrow]->column;
 120                                         if (((crow == drow) && (ccol == dcol)) ||
 121                                                 ((ccol == drow) && (crow == dcol))) {
 122                                                 break;
 123                                         }
 124                                 }
 125                         }
 126
 127                         seqVec[drow].erase(seqVec[drow].begin()+vrow);
 128                         if (vcol < 0) {
 129                                 nCells = seqVec[dcol].size();
 130                                 for (vcol=0; vcol<nCells;vcol++) {
 131                                         crow = seqVec[dcol][vcol]->row;
 132                                         ccol = seqVec[dcol][vcol]->column;
 133                                         if (((crow == drow) && (ccol == dcol)) ||
 134                                                 ((ccol == drow) && (crow == dcol))) {
 135                                                 break;
 136                                         }
 137                                 }
 138                         }
 139
 140                         seqVec[dcol].erase(seqVec[dcol].begin()+vcol);
 141
 142                         if (rmMatrix) {
 143                         //cout << " removing = " << cell->row << '\t' << cell->column  << '\t' << cell->dist << endl;
 144                                 dMatrix->rmCell(cell);
 145                 //      cout << "done" << endl;
 146                         }
 147
 148         }
 149         catch(exception& e) {
 150                 m->errorOut(e, "Cluster", "removeCell");
 151                 exit(1);
 152         }
 153 }
 154 /***********************************************************************/
 155
 156 void Cluster::clusterBins(){
 157         try {
 158         //      cout << smallCol << '\t' << smallRow << '\t' << smallDist << '\t' << rabund->get(smallRow) << '\t' << rabund->get(smallCol);
 159
 160                 rabund->set(smallCol, rabund->get(smallRow)+rabund->get(smallCol));
 161                 rabund->set(smallRow, 0);
 162                 rabund->setLabel(toString(smallDist));
 163
 164         //      cout << '\t' << rabund->get(smallRow) << '\t' << rabund->get(smallCol) << endl;
 165         }
 166         catch(exception& e) {
 167                 m->errorOut(e, "Cluster", "clusterBins");
 168                 exit(1);
 169         }
 170
 171
 172 }
 173
 174 /***********************************************************************/
 175
 176 void Cluster::clusterNames(){
 177         try {
 178         //      cout << smallCol << '\t' << smallRow << '\t' << smallDist << '\t' << list->get(smallRow) << '\t' << list->get(smallCol);
 179                 if (mapWanted) {  updateMap();  }
 180
 181                 list->set(smallCol, list->get(smallRow)+','+list->get(smallCol));
 182                 list->set(smallRow, "");
 183                 list->setLabel(toString(smallDist));
 184
 185         //      cout << '\t' << list->get(smallRow) << '\t' << list->get(smallCol) << endl;
 186     }
 187         catch(exception& e) {
 188                 m->errorOut(e, "Cluster", "clusterNames");
 189                 exit(1);
 190         }
 191
 192 }
 193
 194 /***********************************************************************/
 195 //This function clusters based on the method of the derived class
 196 //At the moment only average and complete linkage are covered, because
 197 //single linkage uses a different approach.
 198 void Cluster::update(double& cutOFF){
 199         try {
 200                 getRowColCells();
 201 //cout << "got rowcells" << endl;
 202
 203                 vector<int> foundCol(nColCells, 0);
 204
 205                 int search;
 206                 bool changed;
 207
 208                 // The vector has to be traversed in reverse order to preserve the index
 209                 // for faster removal in removeCell()
 210                 for (int i=nRowCells-1;i>=0;i--) {
 211                         if (!((rowCells[i]->row == smallRow) && (rowCells[i]->column == smallCol))) {
 212                                 if (rowCells[i]->row == smallRow) {
 213                                         search = rowCells[i]->column;
 214                                 } else {
 215                                         search = rowCells[i]->row;
 216                                 }
 217
 218                                 bool merged = false;
 219                                 for (int j=0;j<nColCells;j++) {
 220                                         if (!((colCells[j]->row == smallRow) && (colCells[j]->column == smallCol))) { //if you are not hte smallest distance
 221                                                 if (colCells[j]->row == search || colCells[j]->column == search) {
 222                                                         foundCol[j] = 1;
 223                                                         merged = true;
 224                                                         changed = updateDistance(colCells[j], rowCells[i]);
 225                                                         // If the cell's distance changed and it had the same distance as
 226                                                         // the smallest distance, invalidate the mins vector in SparseMatrix
 227                                                         if (changed) {
 228                                                                 if (colCells[j]->vectorMap != NULL) {
 229                                                                         *(colCells[j]->vectorMap) = NULL;
 230                                                                         colCells[j]->vectorMap = NULL;
 231                                                                 }
 232                                                         }
 233                                                         break;
 234                                                 }
 235                                         }
 236                                 }
 237                                 //if not merged it you need it for warning
 238                                 if ((!merged) && (method == "average")) {
 239                                         //m->mothurOut("Warning: trying to merge cell " + toString(rowCells[i]->row+1) + " " + toString(rowCells[i]->column+1) + " distance " + toString(rowCells[i]->dist) + " with value above cutoff. Results may vary from using cutoff at cluster command instead of read.dist."); m->mothurOutEndLine();
 240                                         if (cutOFF > rowCells[i]->dist) {
 241                                                 cutOFF = rowCells[i]->dist;
 242                                                 //m->mothurOut("changing cutoff to " + toString(cutOFF));  m->mothurOutEndLine();
 243                                         }
 244
 245                                 }
 246                                 removeCell(rowCells[i], i , -1);
 247
 248                         }
 249                 }
 250                 clusterBins();
 251                 clusterNames();
 252
 253                 // Special handling for singlelinkage case, not sure whether this
 254                 // could be avoided
 255                 for (int i=nColCells-1;i>=0;i--) {
 256                         if (foundCol[i] == 0) {
 257                                 if (method == "average") {
 258                                         if (!((colCells[i]->row == smallRow) && (colCells[i]->column == smallCol))) {
 259                                                 //m->mothurOut("Warning: merging cell " + toString(colCells[i]->row+1) + " " + toString(colCells[i]->column+1) + " distance " + toString(colCells[i]->dist) + " value above cutoff. Results may vary from using cutoff at cluster command instead of read.dist."); m->mothurOutEndLine();
 260                                                 if (cutOFF > colCells[i]->dist) {
 261                                                         cutOFF = colCells[i]->dist;
 262                                                         //m->mothurOut("changing cutoff to " + toString(cutOFF));  m->mothurOutEndLine();
 263                                                 }
 264                                         }
 265                                 }
 266                                 removeCell(colCells[i], -1, i);
 267                         }
 268                 }
 269         }
 270         catch(exception& e) {
 271                 m->errorOut(e, "Cluster", "update");
 272                 exit(1);
 273         }
 274 }
 275 /***********************************************************************/
 276 void Cluster::setMapWanted(bool f)  {
 277         try {
 278                 mapWanted = f;
 279
 280                 //initialize map
 281                 for (int i = 0; i < list->getNumBins(); i++) {
 282
 283                         //parse bin
 284                         string names = list->get(i);
 285                         while (names.find_first_of(',') != -1) {
 286                                 //get name from bin
 287                                 string name = names.substr(0,names.find_first_of(','));
 288                                 //save name and bin number
 289                                 seq2Bin[name] = i;
 290                                 names = names.substr(names.find_first_of(',')+1, names.length());
 291                         }
 292
 293                         //get last name
 294                         seq2Bin[names] = i;
 295                 }
 296
 297         }
 298         catch(exception& e) {
 299                 m->errorOut(e, "Cluster", "setMapWanted");
 300                 exit(1);
 301         }
 302 }
 303 /***********************************************************************/
 304 void Cluster::updateMap() {
 305 try {
 306                 //update location of seqs in smallRow since they move to smallCol now
 307                 string names = list->get(smallRow);
 308                 while (names.find_first_of(',') != -1) {
 309                         //get name from bin
 310                         string name = names.substr(0,names.find_first_of(','));
 311                         //save name and bin number
 312                         seq2Bin[name] = smallCol;
 313                         names = names.substr(names.find_first_of(',')+1, names.length());
 314                 }
 315
 316                 //get last name
 317                 seq2Bin[names] = smallCol;
 318
 319         }
 320         catch(exception& e) {
 321                 m->errorOut(e, "Cluster", "updateMap");
 322                 exit(1);
 323         }
 324 }
 325 /***********************************************************************/
 326
 327
 328