forced rarefaction.single to output ending line for all groups. added subsample...

[mothur.git] / cluster.cpp

diff --git a/cluster.cpp b/cluster.cpp
index e2d307e1bfbce54162b55a7368b58180205f87e5..ac9f4482da12110796c54f15c4137681a3d29038 100644 (file)
--- a/cluster.cpp
+++ b/cluster.cpp
@@ -17,6 +17,7 @@
 Cluster::Cluster(RAbundVector* rav, ListVector* lv, SparseMatrix* dm, float c, string f) :
 rabund(rav), list(lv), dMatrix(dm), method(f)
 {
+       try {
 /*
        cout << "sizeof(MatData): " << sizeof(MatData) << endl;
        cout << "sizeof(PCell*): " << sizeof(PCell*) << endl;
@@ -50,19 +51,28 @@ rabund(rav), list(lv), dMatrix(dm), method(f)
        // a list contains pointers (iterators) to the all distances related
        // to a certain sequence. The Vector is accessed via the index of a 
        // sequence in the distance matrix.
-       
-
+//ofstream outtemp;
+//string temp = "temp";
+//m->openOutputFile(temp, outtemp);    
+//cout << lv->size() << endl;
        seqVec = vector<MatVec>(lv->size());
        for (MatData currentCell = dMatrix->begin(); currentCell != dMatrix->end(); currentCell++) {
+//outtemp << currentCell->row << '\t' << currentCell->column  << '\t' << currentCell->dist << endl;
                seqVec[currentCell->row].push_back(currentCell);
                seqVec[currentCell->column].push_back(currentCell);
        }
-
+//outtemp.close();
        mapWanted = false;  //set to true by mgcluster to speed up overlap merge
        
        //save so you can modify as it changes in average neighbor
        cutoff = c;
        m = MothurOut::getInstance();
+       
+       }
+       catch(exception& e) {
+               m->errorOut(e, "Cluster", "Cluster");
+               exit(1);
+       }
 }
 
 /***********************************************************************/
@@ -74,11 +84,16 @@ void Cluster::getRowColCells() {
                smallRow = smallCell->row;              // get its row
                smallCol = smallCell->column;   // get its column
                smallDist = smallCell->dist;    // get the smallest distance
-       
+       //cout << "small row = " << smallRow << "small col = " << smallCol << "small dist = " << smallDist << endl;
+        
                rowCells = seqVec[smallRow];    // all distances related to the row index
                colCells = seqVec[smallCol];    // all distances related to the column index
                nRowCells = rowCells.size();
                nColCells = colCells.size();
+//cout << "num rows = " << nRowCells << "num col = " << nColCells << endl;
+               
+               //for (int i = 0; i < nColCells; i++) { cout << colCells[i]->row << '\t' << colCells[i]->column << endl;  }
+               //for (int i = 0; i < nRowCells; i++) { cout << rowCells[i]->row << '\t' << rowCells[i]->column << endl;  }
        }
        catch(exception& e) {
                m->errorOut(e, "Cluster", "getRowColCells");
@@ -132,7 +147,9 @@ void Cluster::removeCell(const MatData& cell, int vrow, int vcol, bool rmMatrix)
                        seqVec[dcol].erase(seqVec[dcol].begin()+vcol);
                
                        if (rmMatrix) {
+                       //cout << " removing = " << cell->row << '\t' << cell->column  << '\t' << cell->dist << endl;
                                dMatrix->rmCell(cell);
+               //      cout << "done" << endl;
                        }
                
        }
@@ -197,6 +214,7 @@ void Cluster::update(double& cutOFF){
                // The vector has to be traversed in reverse order to preserve the index
                // for faster removal in removeCell()
                for (int i=nRowCells-1;i>=0;i--) {
+                       //if you are not the smallCell
                        if (!((rowCells[i]->row == smallRow) && (rowCells[i]->column == smallCol))) {
                                if (rowCells[i]->row == smallRow) {
                                        search = rowCells[i]->column;
@@ -224,7 +242,7 @@ void Cluster::update(double& cutOFF){
                                        }               
                                }
                                //if not merged it you need it for warning 
-                               if ((!merged) && (method == "average")) {  
+                               if ((!merged) && (method == "average" || method == "weighted")) {  
                                        //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(); 
                                        if (cutOFF > rowCells[i]->dist) {  
                                                cutOFF = rowCells[i]->dist;  
@@ -243,7 +261,7 @@ void Cluster::update(double& cutOFF){
                // could be avoided
                for (int i=nColCells-1;i>=0;i--) {
                        if (foundCol[i] == 0) {
-                               if (method == "average") {
+                               if (method == "average" || method == "weighted") {
                                        if (!((colCells[i]->row == smallRow) && (colCells[i]->column == smallCol))) {
                                                //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();
                                                if (cutOFF > colCells[i]->dist) {