added zap method to classify.seqs and changed bayesian method name to wang.

[mothur.git] / classify.cpp

diff --git a/classify.cpp b/classify.cpp
index 2d01183479a504c798a18071df045f723643f850..8aa3cdb381ed7a389667ce61962d47cefac15ddd 100644 (file)
--- a/classify.cpp
+++ b/classify.cpp
@@ -61,7 +61,8 @@ void Classify::generateDatabaseAndNames(string tfile, string tempFile, string me
                                        names.push_back(temp.getName());
                                        database->addSequence(temp);    
                                }
-                               database->generateDB();
+                               if ((method == "kmer") && (!shortcuts)) {;} //don't print
+                else {database->generateDB(); }
                        }else if ((method == "kmer") && (!needToGenerate)) {    
                                ifstream kmerFileTest(kmerDBName.c_str());
                                database->readKmerDB(kmerFileTest);     
@@ -200,7 +201,8 @@ void Classify::generateDatabaseAndNames(string tfile, string tempFile, string me
                                }
                                fastaFile.close();
 
-                               database->generateDB();
+                if ((method == "kmer") && (!shortcuts)) {;} //don't print
+                else {database->generateDB(); } 
                                
                        }else if ((method == "kmer") && (!needToGenerate)) {    
                                ifstream kmerFileTest(kmerDBName.c_str());
@@ -249,7 +251,8 @@ int Classify::readTaxonomy(string file) {
                
                m->mothurOutEndLine();
                m->mothurOut("Reading in the " + file + " taxonomy...\t");      cout.flush();
-
+        if (m->debug) { m->mothurOut("[DEBUG]: Taxonomies read in...\n"); }
+        
 #ifdef USE_MPI 
                int pid, num, processors;
                vector<unsigned long long> positions;
@@ -259,9 +262,6 @@ int Classify::readTaxonomy(string file) {
                MPI_File inMPI;
                MPI_Comm_rank(MPI_COMM_WORLD, &pid); //find out who we are
                MPI_Comm_size(MPI_COMM_WORLD, &processors);
-
-               //char* inFileName = new char[file.length()];
-               //memcpy(inFileName, file.c_str(), file.length());
                
                char inFileName[1024];
                strcpy(inFileName, file.c_str());
@@ -298,30 +298,19 @@ int Classify::readTaxonomy(string file) {
                        istringstream iss (tempBuf,istringstream::in);
                        iss >> name; m->gobble(iss);
             iss >> taxInfo;
+            if (m->debug) { m->mothurOut("[DEBUG]: name = " + name + " tax = " + taxInfo + "\n"); }
                        taxonomy[name] = taxInfo;
                        phyloTree->addSeqToTree(name, taxInfo);
                }
                
                MPI_File_close(&inMPI);
                MPI_Barrier(MPI_COMM_WORLD); //make everyone wait - just in case
-#else                          
-               ifstream inTax;
-               m->openInputFile(file, inTax);
-       
-               //read template seqs and save
-               while (!inTax.eof()) {
-                       inTax >> name; m->gobble(inTax);
-            inTax >> taxInfo;
-               
-                       taxonomy[name] = taxInfo;
-                       
-                       phyloTree->addSeqToTree(name, taxInfo);
-               
-                       m->gobble(inTax);
-               }
-               inTax.close();
+#else  
+        
+        taxonomy.clear(); 
+        m->readTax(file, taxonomy);
+        for (map<string, string>::iterator itTax = taxonomy.begin(); itTax != taxonomy.end(); itTax++) {  phyloTree->addSeqToTree(itTax->first, itTax->second);  }
 #endif 
-       
                phyloTree->assignHeirarchyIDs(0);
                
                phyloTree->setUp(file);
@@ -365,3 +354,37 @@ vector<string> Classify::parseTax(string tax) {
 }
 /**************************************************************************************************/
 
+double Classify::getLogExpSum(vector<double> probabilities, int& maxIndex){
+       try {
+        //     http://jblevins.org/notes/log-sum-exp
+        
+        double maxProb = probabilities[0];
+        maxIndex = 0;
+        
+        int numProbs = (int)probabilities.size();
+        
+        for(int i=1;i<numProbs;i++){
+            if(probabilities[i] >= maxProb){
+                maxProb = probabilities[i];
+                maxIndex = i;
+            }
+        }
+        
+        double probSum = 0.0000;
+        
+        for(int i=0;i<numProbs;i++){
+            probSum += exp(probabilities[i] - maxProb);                
+        }
+        
+        probSum = log(probSum) + maxProb;
+        
+        return probSum;
+       }
+       catch(exception& e) {
+               m->errorOut(e, "Classify", "getLogExpSum");
+               exit(1);
+       }
+}
+
+/**************************************************************************************************/
+