added get.dists and remove.dists commands. fixed bug in trim.seqs windows paralelliza...

[mothur.git] / chimeraslayercommand.h

diff --git a/chimeraslayercommand.h b/chimeraslayercommand.h
index 6f3455debc4b6f161102a7bd2a75e797d3c639f0..ee43bcbf72a2ca00af3e2270768f34ac1357bb71 100644 (file)
--- a/chimeraslayercommand.h
+++ b/chimeraslayercommand.h
@@ -15,6 +15,7 @@
 #include "chimera.h"
 #include "chimeraslayer.h"
 #include "sequenceparser.h"
+#include "sequencecountparser.h"
 
 /***********************************************************/
 
@@ -27,8 +28,10 @@ public:
        vector<string> setParameters();
        string getCommandName()                 { return "chimera.slayer";              }
        string getCommandCategory()             { return "Sequence Processing"; }
+       
        string getHelpString(); 
-       string getCitation() { return "Haas BJ, Gevers D, Earl A, Feldgarden M, Ward DV, Giannokous G, Ciulla D, Tabbaa D, Highlander SK, Sodergren E, Methe B, Desantis TZ, Petrosino JF, Knight R, Birren BW (2011). Chimeric 16S rRNA sequence formation and detection in Sanger and 454-pyrosequenced PCR amplicons. Genome Res. \nhttp://www.mothur.org/wiki/Chimera.slayer"; }
+    string getOutputPattern(string);   
+       string getCitation() { return "Haas BJ, Gevers D, Earl A, Feldgarden M, Ward DV, Giannokous G, Ciulla D, Tabbaa D, Highlander SK, Sodergren E, Methe B, Desantis TZ, Petrosino JF, Knight R, Birren BW (2011). Chimeric 16S rRNA sequence formation and detection in Sanger and 454-pyrosequenced PCR amplicons. Genome Res  21:494.\nhttp://www.mothur.org/wiki/Chimera.slayer"; }
        string getDescription()         { return "detect chimeric sequences"; }
        
        int execute(); 
@@ -50,12 +53,14 @@ private:
        int divideInHalf(Sequence, string&, string&);
        map<string, int> sortFastaFile(string, string);
        map<string, int> sortFastaFile(vector<Sequence>&, map<string, string>&, string newFile);
+    int sortFastaFile(vector<Sequence>&, map<string, int>&, string newFile);
        string getNamesFile(string&);
        //int setupChimera(string,);
        int MPIExecute(string, string, string, string, map<string, int>&);
-       int deconvoluteResults(SequenceParser*, string, string, string);
+       int deconvoluteResults(map<string, string>&, string, string, string);
        map<string, int> priority;
        int setUpForSelfReference(SequenceParser*&, map<string, string>&, map<string, map<string, int> >&, int);
+    int setUpForSelfReference(SequenceCountParser*&, map<string, string>&, map<string, map<string, int> >&, int);
        int driverGroups(string, string, string, map<string, map<string, int> >&, map<string, string>&);
        int createProcessesGroups(string, string, string, map<string, map<string, int> >&, map<string, string>&);
        int MPIExecuteGroups(string, string, string, map<string, map<string, int> >&, map<string, string>&);
@@ -65,8 +70,8 @@ private:
        int driverMPI(int, int, MPI_File&, MPI_File&, MPI_File&, MPI_File&, vector<unsigned long long>&, string, map<string, int>&, bool);
        #endif
 
-       bool abort, realign, trim, trimera, save;
-       string fastafile, groupfile, templatefile, outputDir, search, namefile, blastlocation;
+       bool abort, realign, trim, trimera, save, hasName, hasCount, dups;
+       string fastafile, groupfile, templatefile, outputDir, search, namefile, countfile, blastlocation;
        int processors, window, iters, increment, numwanted, ksize, match, mismatch, parents, minSimilarity, minCoverage, minBS, minSNP, numSeqs, templateSeqsLength;
        float divR;
        
@@ -172,7 +177,7 @@ struct slayerData {
 };
 
 /**************************************************************************************************/
-#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
+#if defined (__APPLE__) || (__MACH__) || (linux) || (__linux) || (__linux__) || (__unix__) || (__unix)
 #else
 static DWORD WINAPI MySlayerThreadFunction(LPVOID lpParam){ 
        slayerData* pDataArray;
@@ -208,8 +213,6 @@ static DWORD WINAPI MySlayerThreadFunction(LPVOID lpParam){
                        inFASTA.seekg(pDataArray->start-1); pDataArray->m->gobble(inFASTA); 
                }
                
-               pDataArray->count = pDataArray->end;
-               
                if (pDataArray->m->control_pressed) { out.close(); out2.close(); if (pDataArray->trim) { out3.close(); } inFASTA.close(); delete chimera;  return 0;    }
                
                if (chimera->getUnaligned()) { 
@@ -222,7 +225,7 @@ static DWORD WINAPI MySlayerThreadFunction(LPVOID lpParam){
                
                if (pDataArray->start == 0) { chimera->printHeader(out); }
                
-               int count = 0;
+               pDataArray->count = 0;
                for(int i = 0; i < pDataArray->end; i++){
                        
                        if (pDataArray->m->control_pressed) {   out.close(); out2.close(); if (pDataArray->trim) { out3.close(); } inFASTA.close(); delete chimera; return 1;   }
@@ -313,18 +316,18 @@ static DWORD WINAPI MySlayerThreadFunction(LPVOID lpParam){
                                        
                                        
                                }
-                               count++;
+                               pDataArray->count++;
                        }
                        
                        delete candidateSeq;
                        //report progress
-                       if((count) % 100 == 0){ pDataArray->m->mothurOut("Processing sequence: " + toString(count)); pDataArray->m->mothurOutEndLine();         }
+                       if((pDataArray->count) % 100 == 0){     pDataArray->m->mothurOut("Processing sequence: " + toString(pDataArray->count)); pDataArray->m->mothurOutEndLine();             }
                }
                //report progress
-               if((count) % 100 != 0){ pDataArray->m->mothurOut("Processing sequence: " + toString(count)); pDataArray->m->mothurOutEndLine();         }
+               if((pDataArray->count) % 100 != 0){     pDataArray->m->mothurOut("Processing sequence: " + toString(pDataArray->count)); pDataArray->m->mothurOutEndLine();             }
                
                pDataArray->numNoParents = chimera->getNumNoParents();
-               if (pDataArray->numNoParents == count) {        pDataArray->m->mothurOut("[WARNING]: megablast returned 0 potential parents for all your sequences. This could be due to formatdb.exe not being setup properly, please check formatdb.log for errors.\n"); }
+               if (pDataArray->numNoParents == pDataArray->count) {    pDataArray->m->mothurOut("[WARNING]: megablast returned 0 potential parents for all your sequences. This could be due to formatdb.exe not being setup properly, please check formatdb.log for errors.\n"); }
 
                out.close();
                out2.close();