added topdown parameter to pre.cluster. added more debugging output to bayesian...

[mothur.git] / preclustercommand.h

diff --git a/preclustercommand.h b/preclustercommand.h
index 56bf1f36deaaea02b1357b1d229d87ad96bcafaa..e9568860e30bdc7fbf77b946838c99447e0eec5f 100644 (file)
--- a/preclustercommand.h
+++ b/preclustercommand.h
@@ -29,13 +29,21 @@ struct seqPNode {
        ~seqPNode() {}
 };
 /************************************************************/
-inline bool comparePriority(seqPNode first, seqPNode second) {  
+inline bool comparePriorityTopDown(seqPNode first, seqPNode second) {  
     if (first.numIdentical > second.numIdentical) { return true;  }
     else if (first.numIdentical == second.numIdentical) { 
         if (first.seq.getName() > second.seq.getName()) { return true; }
     }
     return false; 
 }
+/************************************************************/
+inline bool comparePriorityDownTop(seqPNode first, seqPNode second) {  
+    if (first.numIdentical < second.numIdentical) { return true;  }
+    else if (first.numIdentical == second.numIdentical) { 
+        if (first.seq.getName() > second.seq.getName()) { return true; }
+    }
+    return false; 
+}
 //************************************************************/
 
 class PreClusterCommand : public Command {
@@ -71,7 +79,7 @@ private:
     CountTable ct;
     
        int diffs, length, processors;
-       bool abort, bygroup;
+       bool abort, bygroup, topdown;
        string fastafile, namefile, outputDir, groupfile, countfile;
        vector<seqPNode> alignSeqs; //maps the number of identical seqs to a sequence
        map<string, string> names; //represents the names file first column maps to second column
@@ -107,9 +115,10 @@ struct preClusterData {
        int diffs, threadID;
        vector<string> groups;
        vector<string> mapFileNames;
+    bool topdown;
        
        preClusterData(){}
-       preClusterData(string f, string n, string g, string c, string nff,  string nnf, string nmf, vector<string> gr, MothurOut* mout, int st, int en, int d, int tid) {
+       preClusterData(string f, string n, string g, string c, string nff,  string nnf, string nmf, vector<string> gr, MothurOut* mout, int st, int en, int d, bool td, int tid) {
                fastafile = f;
                namefile = n;
                groupfile = g;
@@ -123,6 +132,7 @@ struct preClusterData {
                threadID = tid;
                groups = gr;
         countfile = c;
+        topdown = td;
        }
 };
 
@@ -235,9 +245,10 @@ static DWORD WINAPI MyPreclusterThreadFunction(LPVOID lpParam){
                        pDataArray->m->openOutputFile(pDataArray->newMName+pDataArray->groups[k]+".map", out);
                        pDataArray->mapFileNames.push_back(pDataArray->newMName+pDataArray->groups[k]+".map");
                        
-                       //sort seqs by number of identical seqs
-                       sort(alignSeqs.begin(), alignSeqs.end(), comparePriority);
-                       
+            //sort seqs by number of identical seqs
+            if (pDataArray->topdown) { sort(alignSeqs.begin(), alignSeqs.end(), comparePriorityTopDown);  }
+            else {  sort(alignSeqs.begin(), alignSeqs.end(), comparePriorityDownTop);  }
+            
                        int count = 0;
                        
                        //think about running through twice...