Fixed issue with nameMap in optimatrix read

source/commands/clustercommand.cpp

@@ -11,9 +11,9 @@
 #include "readphylip.h"
 #include "readcolumn.h"
 #include "readmatrix.hpp"
-
 #include "sequence.hpp"
 #include "systemcommand.h"
+#include "sensspeccommand.h"
 
 //**********************************************************************************************************************
 vector<string> ClusterCommand::setParameters(){	
@@ -889,7 +889,7 @@ int ClusterCommand::runOptiCluster(){
         long long numBins = cluster.getNumBins();
         m->mothurOut("\n\niter\ttime\tlabel\tnum_otus\tcutoff\ttp\ttn\tfp\tfn\tsensitivity\tspecificity\tppv\tnpv\tfdr\taccuracy\tmcc\tf1score\n");
         outStep << "iter\ttime\tlabel\tnum_otus\tcutoff\ttp\ttn\tfp\tfn\tsensitivity\tspecificity\tppv\tnpv\tfdr\taccuracy\tmcc\tf1score\n";
-         long long tp, tn, fp, fn;
+        long long tp, tn, fp, fn;
         vector<double> results = cluster.getStats(tp, tn, fp, fn);
         m->mothurOut("0\t0\t" + toString(cutoff) + "\t" + toString(numBins) + "\t"+ toString(cutoff) + "\t" + toString(tp) + "\t" + toString(tn) + "\t" + toString(fp) + "\t" + toString(fn) + "\t");
         outStep << "0\t0\t" + toString(cutoff) + "\t" + toString(numBins) + "\t" + toString(cutoff) + "\t" << tp << '\t' << tn << '\t' << fp << '\t' << fn << '\t';

source/commands/clustersplitcommand.cpp

@@ -654,16 +654,18 @@ map<double, int> ClusterSplitCommand::completeListFile(vector<string> listNames,
             if (countfile != "") { m->getline(in); m->gobble(in); }
             
 			while (!in.eof()) {
-				in >> firstCol >> secondCol; m->getline(in); m->gobble(in);
+				in >> firstCol >> secondCol;
+                m->getline(in);
 				if (countfile == "") { listSingle->push_back(secondCol); }
                 else { listSingle->push_back(firstCol); }
+                m->gobble(in);
 			}
             
 			in.close();
 			m->mothurRemove(singleton);
             
 			numSingleBins = listSingle->getNumBins();
-		}else{  listSingle = NULL; numSingleBins = 0;  }
+        }else{  listSingle = NULL; numSingleBins = 0;  }
 		
         //go through users set and make them floats so we can sort them
         double tcutoff = cutoff * 1000; tcutoff = ceil(tcutoff);

source/commands/clustersplitcommand.h

@@ -51,11 +51,12 @@ class ClusterSplitCommand : public Command {
 	vector<string> outputNames;
 	
 	string file, method, fileroot, tag, outputDir, phylipfile, columnfile, namefile, countfile, distfile, format, showabund, timing, splitmethod, taxFile, fastafile, inputDir, vsearchLocation, metric, initialize;
-	double cutoff, splitcutoff, stableMetric, numSingletons;
+	double cutoff, splitcutoff, stableMetric;
 	int precision, length, processors, taxLevelCutoff, maxIters;
 	bool print_start, abort, large, classic, runCluster, deleteFiles, isList, cutoffNotSet, makeDist;
 	time_t start;
 	ofstream outList, outRabund, outSabund;
+    long long numSingletons;
 	
 	void printData(ListVector*);
 	vector<string> createProcesses(vector< map<string, string> >, set<string>&);

source/commands/distancecommand.cpp

@@ -240,7 +240,7 @@ int DistanceCommand::execute(){
 		if (m->control_pressed) { return 0; }
 		
 		int numSeqs = alignDB.getNumSeqs();
-		cutoff += 0.005;
+		//cutoff += 0.005;
 		
 		if (!alignDB.sameLength()) {  m->mothurOut("[ERROR]: your sequences are not the same length, aborting."); m->mothurOutEndLine(); return 0; }
 		

source/datastructures/optimatrix.cpp

@@ -215,12 +215,21 @@ int OptiMatrix::readPhylip(){
 
         closeness.resize(nonSingletonCount);
         
+        map<string, string> names;
+        if (namefile != "") {  m->readNames(namefile, names); }
+        
         Progress* reading;
         ifstream in;
         
         m->openInputFile(distFile, in);
         in >> nseqs >> name;
         
+        int newA = singletonIndexSwap[0];
+        if (namefile != "") {
+            name = names[name];  //redundant names
+        }
+        nameMap[newA] = name;
+
         int fivepercent = (int)(0.05 * nseqs);
         
         string line = "";
@@ -246,6 +255,11 @@ int OptiMatrix::readPhylip(){
                         int newA = singletonIndexSwap[i];
                         closeness[newA].insert(newB);
                         closeness[newB].insert(newA);
+                        
+                        if (namefile != "") {
+                            name = names[name];  //redundant names
+                        }
+                        nameMap[newA] = name;
                     }
                     index++; reading->update(index);
                 }
@@ -278,6 +292,11 @@ int OptiMatrix::readPhylip(){
                         int newA = singletonIndexSwap[i];
                         closeness[newA].insert(newB);
                         closeness[newB].insert(newA);
+                        
+                        if (namefile != "") {
+                            name = names[name];  //redundant names
+                        }
+                        nameMap[newA] = name;
                     }
                     index++; reading->update(index);
                 }
@@ -297,23 +316,6 @@ int OptiMatrix::readPhylip(){
         if (m->debug) { unsigned long long ramUsed = m->getRAMUsed(); unsigned long long total = m->getTotalRAM();
             m->mothurOut("\nCurrent RAM usage: " + toString(ramUsed/(double)GIG) + " Gigabytes. Total Ram: " + toString(total/(double)GIG) + " Gigabytes.\n"); }
 
-        if (namefile != "") {
-            map<string, string> names;
-            m->readNames(namefile, names);
-            //update nameMap
-            for (int i = 0; i < nameMap.size(); i++) {
-                map<string, string>::iterator it = names.find(nameMap[i]);
-                nameMap[i] = it->second;  //we know its there because we read it above
-            }
-            names.clear();
-        }
-        
-        for (int i = 0; i < closeness.size(); i++) {
-            string newName = nameMap[i];
-            int newIndex = singletonIndexSwap[i];
-            nameMap[newIndex] = newName;
-        }
-        
         return 0;
         
     }
@@ -389,6 +391,9 @@ int OptiMatrix::readColumn(){
         
         closeness.resize(nonSingletonCount);
         
+        map<string, string> names;
+        if (namefile != "") {  m->readNames(namefile, names); }
+        
         while(in){  //let's assume it's a triangular matrix...
             
             in >> firstName; m->gobble(in);
@@ -416,29 +421,19 @@ int OptiMatrix::readColumn(){
                 int newA = singletonIndexSwap[indexA];
                 closeness[newA].insert(newB);
                 closeness[newB].insert(newA);
+                
+                if (namefile != "") {
+                    firstName = names[firstName];  //redundant names
+                    secondName = names[secondName]; //redundant names
+                }
+                
+                nameMap[newA] = firstName;
+                nameMap[newB] = secondName;
             }
         }
         in.close();
         nameAssignment.clear();
         
-        if (namefile != "") {
-            map<string, string> names;
-            m->readNames(namefile, names);
-            //update nameMap
-            for (int i = 0; i < nameMap.size(); i++) {
-                map<string, string>::iterator it = names.find(nameMap[i]);
-                nameMap[i] = it->second;  //we know its there because we read it above
-            }
-            names.clear();
-        }
-        
-        
-        for (int i = 0; i < closeness.size(); i++) {
-            string newName = nameMap[i];
-            int newIndex = singletonIndexSwap[i];
-            nameMap[newIndex] = newName;
-        }
-        
         return 1;
         
     }

source/opticluster.cpp

@@ -171,10 +171,7 @@ bool OptiCluster::update(double& listMetric) {
                     bins[binNumber].erase(remove(bins[binNumber].begin(), bins[binNumber].end(), seqNumber), bins[binNumber].end()); //remove from old bin i
                 }
                 
-                if (usedInsert) {
-                    if (bins[binNumber].size() == 0) { insertLocation = binNumber;  } //set flag if old bin is empty.
-                    insertLocation = findInsert();
-                }
+                if (usedInsert) { insertLocation = findInsert(); }
                 
                 //update seqBins
                 seqBin[seqNumber] = bestBin; //set new OTU location

source/opticluster.h

@@ -23,7 +23,7 @@ class OptiCluster : public Cluster {
 #endif
     
 public:
-    OptiCluster(OptiMatrix* mt, string met, double ns) : Cluster() {
+    OptiCluster(OptiMatrix* mt, string met, long long ns) : Cluster() {
         m = MothurOut::getInstance(); matrix = mt; metric = met; truePositives = 0; trueNegatives = 0; falseNegatives = 0; falsePositives = 0; numSingletons = ns;
     }
     ~OptiCluster() {}

source/read/splitmatrix.h

@@ -26,7 +26,7 @@ class SplitMatrix  {
 		int split();
 		vector< map<string, string> > getDistanceFiles();  //returns map of distance files -> namefile sorted by distance file size
 		string getSingletonNames() { return singleton; } //returns namesfile containing singletons
-        int getNumSingleton() { return numSingleton; } //returns namesfile containing singletons
+        long long getNumSingleton() { return numSingleton; } //returns namesfile containing singletons
 	
 	private:
 		MothurOut* m;
@@ -35,7 +35,8 @@ class SplitMatrix  {
 		vector< map< string, string> > dists;
 		float cutoff, distCutoff;
 		bool large, classic;
-		int processors, numSingleton;
+        int processors;
+        long long numSingleton;
 				
 		int splitDistance();
 		int splitClassify();