fixed seq. error align=f issue. added std and ave calcs to mothurout. working on...

[mothur.git] / mothurout.cpp

diff --git a/mothurout.cpp b/mothurout.cpp
index 37c0916d4c49e1f8ed8ca979c7ff61b7e1a13ef3..240e7ec2e5ac24ec3e637b78ae28dd7608a0389d 100644 (file)
--- a/mothurout.cpp
+++ b/mothurout.cpp
@@ -1183,7 +1183,7 @@ string MothurOut::sortFile(string distFile, string outputDir){
 
                        string firstName, secondName;
                        float dist;
-                       while (input) {
+                       while (!input.eof()) {
                                input >> firstName >> secondName >> dist;
                                output << dist << '\t' << firstName << '\t' << secondName << endl;
                                gobble(input);
@@ -1199,16 +1199,17 @@ string MothurOut::sortFile(string distFile, string outputDir){
                
                        //read in sorted file and put distance at end again
                        ifstream input2;
+            ofstream output2;
                        openInputFile(tempOutfile, input2);
-                       openOutputFile(outfile, output);
+                       openOutputFile(outfile, output2);
                
-                       while (input2) {
+            while (!input2.eof()) {
                                input2 >> dist >> firstName >> secondName;
-                               output << firstName << '\t' << secondName << '\t' << dist << endl;
+                               output2 << firstName << '\t' << secondName << '\t' << dist << endl;
                                gobble(input2);
                        }
                        input2.close();
-                       output.close();
+                       output2.close();
                
                        //remove temp files
                        mothurRemove(tempDistFile);
@@ -1607,8 +1608,17 @@ int MothurOut::readTax(string namefile, map<string, string>& taxMap) {
                 if (pairDone) { 
                     //are there confidence scores, if so remove them
                     if (secondCol.find_first_of('(') != -1) {  removeConfidences(secondCol);   }
-                    taxMap[firstCol] = secondCol;
-                    if (debug) {  mothurOut("[DEBUG]: name = '" + firstCol + "' tax = '" + secondCol + "'\n");  }
+                    map<string, string>::iterator itTax = taxMap.find(firstCol);
+                    
+                    if(itTax == taxMap.end()) {
+                        bool ignore = false;
+                        if (secondCol != "") { if (secondCol[secondCol.length()-1] != ';') { mothurOut("[ERROR]: " + firstCol + " is missing the final ';', ignoring.\n"); ignore=true; }
+                        }
+                        if (!ignore) { taxMap[firstCol] = secondCol; }
+                        if (debug) {  mothurOut("[DEBUG]: name = '" + firstCol + "' tax = '" + secondCol + "'\n");  }
+                    }else {
+                        mothurOut("[ERROR]: " + firstCol + " is already in your taxonomy file, names must be unique./n"); control_pressed = true;
+                    }
                     pairDone = false; 
                 }
             }
@@ -1625,8 +1635,18 @@ int MothurOut::readTax(string namefile, map<string, string>& taxMap) {
                 if (pairDone) { 
                     //are there confidence scores, if so remove them
                     if (secondCol.find_first_of('(') != -1) {  removeConfidences(secondCol);   }
-                    taxMap[firstCol] = secondCol;
-                    if (debug) {  mothurOut("[DEBUG]: name = '" + firstCol + "' tax = '" + secondCol + "'\n");  }
+                    map<string, string>::iterator itTax = taxMap.find(firstCol);
+                    
+                    if(itTax == taxMap.end()) {
+                        bool ignore = false;
+                        if (secondCol != "") { if (secondCol[secondCol.length()-1] != ';') { mothurOut("[ERROR]: " + firstCol + " is missing the final ';', ignoring.\n"); ignore=true; }
+                        }
+                        if (!ignore) { taxMap[firstCol] = secondCol; }
+                        if (debug) {  mothurOut("[DEBUG]: name = '" + firstCol + "' tax = '" + secondCol + "'\n");  }
+                    }else {
+                        mothurOut("[ERROR]: " + firstCol + " is already in your taxonomy file, names must be unique./n"); control_pressed = true;
+                    }
+
                     pairDone = false; 
                 }
             } 
@@ -2470,6 +2490,9 @@ void MothurOut::getNumSeqs(ifstream& file, int& numSeqs){
 //This function parses the estimator options and puts them in a vector
 void MothurOut::splitAtChar(string& estim, vector<string>& container, char symbol) {
        try {
+        
+        if (symbol == '-') { splitAtDash(estim, container); return; }
+        
                string individual = "";
                int estimLength = estim.size();
                for(int i=0;i<estimLength;i++){
@@ -2902,6 +2925,79 @@ bool MothurOut::checkReleaseVersion(ifstream& file, string version) {
        }
 }
 /**************************************************************************************************/
+vector<double> MothurOut::getAverages(vector< vector<double> >& dists) {
+       try{
+        vector<double> averages; //averages.resize(numComp, 0.0);
+        for (int i = 0; i < dists[0].size(); i++) { averages.push_back(0.0); }
+      
+        for (int thisIter = 0; thisIter < dists.size(); thisIter++) {
+            for (int i = 0; i < dists[thisIter].size(); i++) {  
+                averages[i] += dists[thisIter][i];
+            }
+        }
+        
+        //finds average.
+        for (int i = 0; i < averages.size(); i++) {  averages[i] /= (double) dists.size(); }
+        
+        return averages;
+    }
+       catch(exception& e) {
+               errorOut(e, "MothurOut", "getAverages");                
+               exit(1);
+       }
+}
+/**************************************************************************************************/
+vector<double> MothurOut::getStandardDeviation(vector< vector<double> >& dists) {
+       try{
+        
+        vector<double> averages = getAverages(dists);
+        
+        //find standard deviation
+        vector<double> stdDev; //stdDev.resize(numComp, 0.0);
+        for (int i = 0; i < dists[0].size(); i++) { stdDev.push_back(0.0); }
+        
+        for (int thisIter = 0; thisIter < dists.size(); thisIter++) { //compute the difference of each dist from the mean, and square the result of each
+            for (int j = 0; j < dists[thisIter].size(); j++) {
+                stdDev[j] += ((dists[thisIter][j] - averages[j]) * (dists[thisIter][j] - averages[j]));
+            }
+        }
+        for (int i = 0; i < stdDev.size(); i++) {  
+            stdDev[i] /= (double) dists.size(); 
+            stdDev[i] = sqrt(stdDev[i]);
+        }
+        
+        return stdDev;
+    }
+       catch(exception& e) {
+               errorOut(e, "MothurOut", "getAverages");                
+               exit(1);
+       }
+}
+/**************************************************************************************************/
+vector<double> MothurOut::getStandardDeviation(vector< vector<double> >& dists, vector<double>& averages) {
+       try{
+        //find standard deviation
+        vector<double> stdDev; //stdDev.resize(numComp, 0.0);
+        for (int i = 0; i < dists[0].size(); i++) { stdDev.push_back(0.0); }
+        
+        for (int thisIter = 0; thisIter < dists.size(); thisIter++) { //compute the difference of each dist from the mean, and square the result of each
+            for (int j = 0; j < dists[thisIter].size(); j++) {
+                stdDev[j] += ((dists[thisIter][j] - averages[j]) * (dists[thisIter][j] - averages[j]));
+            }
+        }
+        for (int i = 0; i < stdDev.size(); i++) {  
+            stdDev[i] /= (double) dists.size(); 
+            stdDev[i] = sqrt(stdDev[i]);
+        }
+        
+        return stdDev;
+    }
+       catch(exception& e) {
+               errorOut(e, "MothurOut", "getAverages");                
+               exit(1);
+       }
+}
+/**************************************************************************************************/
 bool MothurOut::isContainingOnlyDigits(string input) {
        try{