added parameter to chimera.uchime

diff --git a/addtargets2.cpp b/addtargets2.cpp
index 4e0dbd1fea24d32885e00b3e83b62690d6c8065b..57a4ef58b1b77b586ff104c5f62c25f0bd901941 100644 (file)
--- a/addtargets2.cpp
+++ b/addtargets2.cpp
@@ -1,5 +1,5 @@
 //#if  UCHIMES\r
-\r
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
 #include "myutils.h"\r
 #include "chime.h"\r
 #include "ultra.h"\r

diff --git a/alignchime.cpp b/alignchime.cpp
index d7b05a877f2e2e09746d4ed02864fb0884349613..c97c333ce42ad6d50fcdca96cb2758c30dc8d213 100644 (file)
--- a/alignchime.cpp
+++ b/alignchime.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #include "myutils.h"\r
 #include "seq.h"\r
 #include "chime.h"\r

diff --git a/alignchimel.cpp b/alignchimel.cpp
index ae152af8dea0d02ff298b367756971468d0ca078..8f4cbd57bd1cda8e29e8f1cbdde32292643aeb5d 100644 (file)
--- a/alignchimel.cpp
+++ b/alignchimel.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #include "myutils.h"\r
 #include "seq.h"\r
 #include "chime.h"\r

diff --git a/alnheuristics.h b/alnheuristics.h
index 9a8d283ed4f710b8951a2d72faf7ce9f39450e83..cd346cf1201d5ac6e265da145f99476a1168b82d 100644 (file)
--- a/alnheuristics.h
+++ b/alnheuristics.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef alnheuristics_h\r
 #define alnheuristics_h\r
 \r

diff --git a/alnparams.cpp b/alnparams.cpp
index d1b90364212dc4aa41bd275166e22cf37afc2038..41ee738027921b1a217544a67a449ff0dc5fb761 100644 (file)
--- a/alnparams.cpp
+++ b/alnparams.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #include "myutils.h"\r
 #include <float.h>     // for FLT_MAX\r
 #include "mx.h"\r
@@ -6,12 +8,12 @@
 \r
 #define TEST   0\r
 \r
-void SetBLOSUM62();
+void SetBLOSUM62();\r
 void SetNucSubstMx(double Match, double Mismatch);\r
 void ReadSubstMx(const string &FileName, Mx<float> &Mxf);\r
-
-extern Mx<float> g_SubstMxf;
-extern float **g_SubstMx;
+\r
+extern Mx<float> g_SubstMxf;\r
+extern float **g_SubstMx;\r
 \r
 void AlnParams::Clear()\r
        {\r
@@ -313,14 +315,14 @@ void AlnParams::SetPenalties(const string &OpenStr, const string &ExtStr)
 void AlnParams::SetMxFromCmdLine(bool IsNucleo)\r
        {\r
        if (IsNucleo)\r
-               SetNucSubstMx(opt_match, opt_mismatch);
+               SetNucSubstMx(opt_match, opt_mismatch);\r
        else\r
                {\r
                if (opt_matrix == "")\r
                        {\r
                        SubstMxName = "BLOSUM62";\r
-                       SetBLOSUM62();
-                       }
+                       SetBLOSUM62();\r
+                       }\r
                else\r
                        {\r
                        ReadSubstMx(opt_matrix, g_SubstMxf);\r
@@ -361,9 +363,9 @@ void AlnParams::InitFromCmdLine(bool IsNucleo)
 \r
 // Global\r
        if (IsNucleo)\r
-               Init4(g_SubstMx, -10.0, -1.0, -0.5, -0.5);
+               Init4(g_SubstMx, -10.0, -1.0, -0.5, -0.5);\r
        else\r
-               Init4(g_SubstMx, -17.0, -1.0, -0.5, -0.5);
+               Init4(g_SubstMx, -17.0, -1.0, -0.5, -0.5);\r
        SetPenalties(opt_gapopen, opt_gapext);\r
        }\r
 \r

diff --git a/alnparams.h b/alnparams.h
index 4037912566144aff0ce326609029004395a15a6f..b159179098426ae1eaac9ab11c20251b15c9e0dc 100644 (file)
--- a/alnparams.h
+++ b/alnparams.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef alnparams_h\r
 #define alnparams_h\r
 \r

diff --git a/alpha.cpp b/alpha.cpp
index 0efca3ba0bc09a61c0ad980948a7f9c49761a3f7..9d82f350eb7d8045d59f1005af13057a5ec1a3ac 100644 (file)
--- a/alpha.cpp
+++ b/alpha.cpp
@@ -1,4 +1,5 @@
 // Generated by /p/py/alphac.py
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.
 #include "alpha.h"
 
 unsigned g_CharToLetterAminoStop[256] =

diff --git a/alpha.h b/alpha.h
index e021b7fbc307ea1277a26d4a86275415cc167754..e491c9c602ffe0186f0d3b2aec6d514d87832b05 100644 (file)
--- a/alpha.h
+++ b/alpha.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef alpha_h\r
 #define alpha_h\r
 \r

diff --git a/alpha2.cpp b/alpha2.cpp
index 26bc1c6bac30ae7c3c024d74eb0fb97d55477397..08f6e21cefab25906fa8b54a0afa330099a6079c 100644 (file)
--- a/alpha2.cpp
+++ b/alpha2.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #include "myutils.h"\r
 #include "alpha.h"\r
 #include "timing.h"\r

diff --git a/blastdb.cpp b/blastdb.cpp
index 70349e50f0264b68fba76c4f04ebbddd0c8c96f6..b1a7b48d1108d747c1423eb0f8bad19e3ceede42 100644 (file)
--- a/blastdb.cpp
+++ b/blastdb.cpp
@@ -83,8 +83,10 @@ vector<int> BlastDB::findClosestSequences(Sequence* seq, int n) {
                        blastCommand = path + "blast/bin/blastall -p blastn -d " + dbFileName + " -m 8 -W 28 -v " + toString(n) + " -b " + toString(n);;
                        blastCommand += (" -i " + (queryFileName+seq->getName()) + " -o " + blastFileName+seq->getName());
                #else
-                       blastCommand =  "\"" + path + "blast\\bin\\blastall\" -p blastn -d " + dbFileName + " -m 8 -W 28 -v " + toString(n) + " -b " + toString(n);;
+                       blastCommand =  "\"" + path + "blast\\bin\\blastall\" -p blastn -d " + "\"" + dbFileName + "\"" + " -m 8 -W 28 -v " + toString(n) + " -b " + toString(n);
                        blastCommand += (" -i " + (queryFileName+seq->getName()) + " -o " + blastFileName+seq->getName());
+                       //wrap entire string in ""
+                       blastCommand = "\"" + blastCommand + "\"";
                #endif
                system(blastCommand.c_str());
                
@@ -140,8 +142,11 @@ vector<int> BlastDB::findClosestMegaBlast(Sequence* seq, int n, int minPerID) {
                        blastCommand = path + "blast/bin/megablast -e 1e-10 -d " + dbFileName + " -m 8 -b " + toString(n) + " -v " + toString(n); //-W 28 -p blastn
                        blastCommand += (" -i " + (queryFileName+seq->getName()) + " -o " + blastFileName+seq->getName());
                #else
-                       blastCommand =  "\"" + path + "blast\\bin\\megablast\" -e 1e-10 -d " + dbFileName + " -m 8 -b " + toString(n) + " -v " + toString(n); //-W 28 -p blastn
+                       blastCommand =  "\"" + path + "blast\\bin\\megablast\" -e 1e-10 -d " + "\"" + dbFileName + "\"" + " -m 8 -b " + toString(n) + " -v " + toString(n); //-W 28 -p blastn
                        blastCommand += (" -i " + (queryFileName+seq->getName()) + " -o " + blastFileName+seq->getName());
+                       //wrap entire string in ""
+                       blastCommand = "\"" + blastCommand + "\"";
+
                #endif
                
                system(blastCommand.c_str());
@@ -213,7 +218,9 @@ void BlastDB::generateDB() {
                #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
                        formatdbCommand = path + "blast/bin/formatdb -p F -o T -i " + dbFileName;       //      format the database, -o option gives us the ability
                #else
-                       formatdbCommand = "\"" + path + "blast\\bin\\formatdb\" -p F -o T -i " + dbFileName;
+                       formatdbCommand = "\"" + path + "blast\\bin\\formatdb\" -p F -o T -i " + "\"" +  dbFileName + "\"";
+                       //wrap entire string in ""
+                       formatdbCommand = "\"" + formatdbCommand + "\"";
                #endif
                system(formatdbCommand.c_str());                                                                //      to get the right sequence names, i think. -p F
                                                                                                                                        //      option tells formatdb that seqs are DNA, not prot

diff --git a/chainer.h b/chainer.h
index a954dc04b326b54d5ce2f83f027f492b3f4dd73a..7f541311e721e6778fd3b5312266fd3578fa5d42 100644 (file)
--- a/chainer.h
+++ b/chainer.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef chainer_h\r
 #define chainer_h\r
 \r

diff --git a/chime.h b/chime.h
index 1b0662a6ef39fe1eadd98bdf46fc4fa1abb8b3ce..f143ea10795df3ed995660b0862bb0142f2db91b 100644 (file)
--- a/chime.h
+++ b/chime.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef chime_h\r
 #define chime_h\r
 \r

diff --git a/chimerauchimecommand.cpp b/chimerauchimecommand.cpp
index 891b77b63a082c43002d337986ba258e8f88ee0f..a47aed0ca840197331cdf251ef48bcb68cf9df56 100644 (file)
--- a/chimerauchimecommand.cpp
+++ b/chimerauchimecommand.cpp
@@ -22,6 +22,24 @@ vector<string> ChimeraUchimeCommand::setParameters(){
                CommandParameter pprocessors("processors", "Number", "", "1", "", "", "",false,false); parameters.push_back(pprocessors);
                CommandParameter pinputdir("inputdir", "String", "", "", "", "", "",false,false); parameters.push_back(pinputdir);
                CommandParameter poutputdir("outputdir", "String", "", "", "", "", "",false,false); parameters.push_back(poutputdir);
+               CommandParameter pabskew("abskew", "Number", "", "1.9", "", "", "",false,false); parameters.push_back(pabskew);
+               CommandParameter pchimealns("chimealns", "Boolean", "", "F", "", "", "",false,false); parameters.push_back(pchimealns);
+               CommandParameter pminh("minh", "Number", "", "0.3", "", "", "",false,false); parameters.push_back(pminh);
+               CommandParameter pmindiv("mindiv", "Number", "", "0.5", "", "", "",false,false); parameters.push_back(pmindiv);
+               CommandParameter pxn("xn", "Number", "", "8.0", "", "", "",false,false); parameters.push_back(pxn);
+               CommandParameter pdn("dn", "Number", "", "1.4", "", "", "",false,false); parameters.push_back(pdn);
+               CommandParameter pxa("xa", "Number", "", "1", "", "", "",false,false); parameters.push_back(pxa);
+               CommandParameter pchunks("chunks", "Number", "", "4", "", "", "",false,false); parameters.push_back(pchunks);
+               CommandParameter pminchunk("minchunk", "Number", "", "64", "", "", "",false,false); parameters.push_back(pminchunk);
+               CommandParameter pidsmoothwindow("idsmoothwindow", "Number", "", "32", "", "", "",false,false); parameters.push_back(pidsmoothwindow);
+               CommandParameter pminsmoothid("minsmoothid", "Number", "", "0.95", "", "", "",false,false); parameters.push_back(pminsmoothid);
+               CommandParameter pmaxp("maxp", "Number", "", "2", "", "", "",false,false); parameters.push_back(pmaxp);
+               CommandParameter pskipgaps("skipgaps", "Boolean", "", "T", "", "", "",false,false); parameters.push_back(pskipgaps);
+               CommandParameter pskipgaps2("skipgaps2", "Boolean", "", "T", "", "", "",false,false); parameters.push_back(pskipgaps2);
+               CommandParameter pminlen("minlen", "Number", "", "10", "", "", "",false,false); parameters.push_back(pminlen);
+               CommandParameter pmaxlen("maxlen", "Number", "", "10000", "", "", "",false,false); parameters.push_back(pmaxlen);
+               CommandParameter pucl("ucl", "Boolean", "", "F", "", "", "",false,false); parameters.push_back(pucl);
+               CommandParameter pqueryfract("queryfract", "Number", "", "0.5", "", "", "",false,false); parameters.push_back(pqueryfract);
                
                vector<string> myArray;
                for (int i = 0; i < parameters.size(); i++) {   myArray.push_back(parameters[i].name);          }
@@ -38,12 +56,30 @@ string ChimeraUchimeCommand::getHelpString(){
                string helpString = "";
                helpString += "The chimera.uchime command reads a fastafile and referencefile and outputs potentially chimeric sequences.\n";
                helpString += "This command is a wrapper for uchime written by Robert C. Edgar.\n";
-               helpString += "The chimera.uchime command parameters are fasta, name, reference and processors.\n";
+               helpString += "The chimera.uchime command parameters are fasta, name, reference, processors, abskew, chimealns, minh, mindiv, xn, dn, xa, chunks, minchunk, idsmoothwindow, minsmoothid, maxp, skipgaps, skipgaps2, minlen, maxlen, ucl and queryfact.\n";
                helpString += "The fasta parameter allows you to enter the fasta file containing your potentially chimeric sequences, and is required, unless you have a valid current fasta file. \n";
                helpString += "The name parameter allows you to provide a name file, if you are using template=self. \n";
                helpString += "You may enter multiple fasta files by separating their names with dashes. ie. fasta=abrecovery.fasta-amazon.fasta \n";
                helpString += "The reference parameter allows you to enter a reference file containing known non-chimeric sequences, and is required. You may also set template=self, in this case the abundant sequences will be used as potential parents. \n";
                helpString += "The processors parameter allows you to specify how many processors you would like to use.  The default is 1. \n";
+               helpString += "The abskew parameter can only be used with template=self. Minimum abundance skew. Default 1.9. Abundance skew is: min [ abund(parent1), abund(parent2) ] / abund(query).\n";
+               helpString += "The chimealns parameter allows you to indicate you would like a file containing multiple alignments of query sequences to parents in human readable format. Alignments show columns with differences that support or contradict a chimeric model.\n";
+               helpString += "The minh parameter - mininum score to report chimera. Default 0.3. Values from 0.1 to 5 might be reasonable. Lower values increase sensitivity but may report more false positives. If you decrease xn you may need to increase minh, and vice versa.\n";
+               helpString += "The mindiv parameter - minimum divergence ratio, default 0.5. Div ratio is 100%% - %%identity between query sequence and the closest candidate for being a parent. If you don't care about very close chimeras, then you could increase mindiv to, say, 1.0 or 2.0, and also decrease --min h, say to 0.1, to increase sensitivity. How well this works will depend on your data. Best is to tune parameters on a good benchmark.\n";
+               helpString += "The xn parameter - weight of a no vote. Default 8.0. Decreasing this weight to around 3 or 4 may give better performance on denoised data.\n";
+               helpString += "The dn parameter - pseudo-count prior on number of no votes. Default 1.4. Probably no good reason to change this unless you can retune to a good benchmark for your data. Reasonable values are probably in the range from 0.2 to 2.\n";
+               helpString += "The xa parameter - weight of an abstain vote. Default 1. So far, results do not seem to be very sensitive to this parameter, but if you have a good training set might be worth trying. Reasonable values might range from 0.1 to 2.\n";
+               helpString += "The chunks parameter is the number of chunks to extract from the query sequence when searching for parents. Default 4.\n";
+               helpString += "The minchunk parameter is the minimum length of a chunk. Default 64.\n";
+               helpString += "The idsmoothwindow parameter is the length of id smoothing window. Default 32.\n";
+               helpString += "The minsmoothid parameter - minimum factional identity over smoothed window of candidate parent. Default 0.95.\n";
+               helpString += "The maxp parameter - maximum number of candidate parents to consider. Default 2. In tests so far, increasing maxp gives only a very small improvement in sensivity but tends to increase the error rate quite a bit.\n";
+               helpString += "The skipgaps parameter controls how gapped columns affect counting of diffs. If skipgaps is set to T, columns containing gaps do not found as diffs. Default = T.\n";
+               helpString += "The skipgaps2 parameter controls how gapped columns affect counting of diffs. If skipgaps2 is set to T, if column is immediately adjacent to a column containing a gap, it is not counted as a diff. Default = T.\n";
+               helpString += "The minlen parameter is the minimum unaligned sequence length. Defaults 10. Applies to both query and reference sequences.\n";
+               helpString += "The maxlen parameter is the maximum unaligned sequence length. Defaults 10000. Applies to both query and reference sequences.\n";
+               helpString += "The ucl parameter - use local-X alignments. Default is global-X or false. On tests so far, global-X is always better; this option is retained because it just might work well on some future type of data.\n";
+               helpString += "The queryfract parameter - minimum fraction of the query sequence that must be covered by a local-X alignment. Default 0.5. Applies only when ucl is true.\n";
 #ifdef USE_MPI
                helpString += "When using MPI, the processors parameter is set to the number of MPI processes running. \n";
 #endif
@@ -66,6 +102,7 @@ ChimeraUchimeCommand::ChimeraUchimeCommand(){
                vector<string> tempOutNames;
                outputTypes["chimera"] = tempOutNames;
                outputTypes["accnos"] = tempOutNames;
+               outputTypes["alns"] = tempOutNames;
        }
        catch(exception& e) {
                m->errorOut(e, "ChimeraUchimeCommand", "ChimeraUchimeCommand");
@@ -98,6 +135,7 @@ ChimeraUchimeCommand::ChimeraUchimeCommand(string option)  {
                        vector<string> tempOutNames;
                        outputTypes["chimera"] = tempOutNames;
                        outputTypes["accnos"] = tempOutNames;
+                       outputTypes["alns"] = tempOutNames;
                        
                        //if the user changes the input directory command factory will send this info to us in the output parameter 
                        string inputDir = validParameter.validFile(parameters, "inputdir", false);              
@@ -278,6 +316,38 @@ ChimeraUchimeCommand::ChimeraUchimeCommand(string option)  {
                        string temp = validParameter.validFile(parameters, "processors", false);        if (temp == "not found"){       temp = m->getProcessors();      }
                        m->setProcessors(temp);
                        convert(temp, processors);
+                       
+                       abskew = validParameter.validFile(parameters, "abskew", false); if (abskew == "not found"){     useAbskew = false;  abskew = "1.9";     }else{  useAbskew = true;  }
+                       if (useAbskew && templatefile != "self") { m->mothurOut("The abskew parameter is only valid with template=self, ignoring."); m->mothurOutEndLine(); useAbskew = false; }
+                       
+                       temp = validParameter.validFile(parameters, "chimealns", false);                        if (temp == "not found") { temp = "f"; }
+                       chimealns = m->isTrue(temp); 
+                       
+                       minh = validParameter.validFile(parameters, "minh", false);                                             if (minh == "not found")                        { useMinH = false; minh = "0.3";                                        }       else{ useMinH = true;                   }
+                       mindiv = validParameter.validFile(parameters, "mindiv", false);                                 if (mindiv == "not found")                      { useMindiv = false; mindiv = "0.5";                            }       else{ useMindiv = true;                 }
+                       xn = validParameter.validFile(parameters, "xn", false);                                                 if (xn == "not found")                          { useXn = false; xn = "8.0";                                            }       else{ useXn = true;                             }
+                       dn = validParameter.validFile(parameters, "dn", false);                                                 if (dn == "not found")                          { useDn = false; dn = "1.4";                                            }       else{ useDn = true;                             }
+                       xa = validParameter.validFile(parameters, "xa", false);                                                 if (xa == "not found")                          { useXa = false; xa = "1";                                                      }       else{ useXa = true;                             }
+                       chunks = validParameter.validFile(parameters, "chunks", false);                                 if (chunks == "not found")                      { useChunks = false; chunks = "4";                                      }       else{ useChunks = true;                 }
+                       minchunk = validParameter.validFile(parameters, "minchunk", false);                             if (minchunk == "not found")            { useMinchunk = false; minchunk = "64";                         }       else{ useMinchunk = true;               }
+                       idsmoothwindow = validParameter.validFile(parameters, "idsmoothwindow", false); if (idsmoothwindow == "not found")      { useIdsmoothwindow = false; idsmoothwindow = "32";     }       else{ useIdsmoothwindow = true; }
+                       minsmoothid = validParameter.validFile(parameters, "minsmoothid", false);               if (minsmoothid == "not found")         { useMinsmoothid = false; minsmoothid = "0.95";         }       else{ useMinsmoothid = true;    }
+                       maxp = validParameter.validFile(parameters, "maxp", false);                                             if (maxp == "not found")                        { useMaxp = false; maxp = "2";                                          }       else{ useMaxp = true;                   }
+                       minlen = validParameter.validFile(parameters, "minlen", false);                                 if (minlen == "not found")                      { useMinlen = false; minlen = "10";                                     }       else{ useMinlen = true;                 }
+                       maxlen = validParameter.validFile(parameters, "maxlen", false);                                 if (maxlen == "not found")                      { useMaxlen = false; maxlen = "10000";                          }       else{ useMaxlen = true;                 }
+                       
+                       temp = validParameter.validFile(parameters, "ucl", false);                                              if (temp == "not found") { temp = "f"; }
+                       ucl = m->isTrue(temp);
+                       
+                       queryfract = validParameter.validFile(parameters, "queryfract", false);                 if (queryfract == "not found")          { useQueryfract = false; queryfract = "0.5";            }       else{ useQueryfract = true;             }
+                       if (!ucl && useQueryfract) { m->mothurOut("queryfact may only be used when ucl=t, ignoring."); m->mothurOutEndLine(); useQueryfract = false; }
+                       
+                       temp = validParameter.validFile(parameters, "skipgaps", false);                                 if (temp == "not found") { temp = "t"; }
+                       skipgaps = m->isTrue(temp); 
+
+                       temp = validParameter.validFile(parameters, "skipgaps2", false);                                if (temp == "not found") { temp = "t"; }
+                       skipgaps2 = m->isTrue(temp); 
+
                }
        }
        catch(exception& e) {
@@ -291,6 +361,8 @@ int ChimeraUchimeCommand::execute(){
        try{
                if (abort == true) { if (calledHelp) { return 0; }  return 2;   }
                
+               m->mothurOut("\nuchime by Robert C. Edgar\nhttp://drive5.com/uchime\nThis code is donated to the public domain.\n\n");
+               
                for (int s = 0; s < fastaFileNames.size(); s++) {
                        
                        m->mothurOut("Checking sequences from " + fastaFileNames[s] + " ..." ); m->mothurOutEndLine();
@@ -375,23 +447,27 @@ int ChimeraUchimeCommand::execute(){
                        }
                        
                        if (outputDir == "") { outputDir = m->hasPath(fastaFileNames[s]);  }//if user entered a file with a path then preserve it                               
-                       string outputFileName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + "slayer.chimera";
-                       string accnosFileName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s]))  + "slayer.accnos";
+                       string outputFileName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s])) + "uchime.chimera";
+                       string accnosFileName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s]))  + "uchime.accnos";
+                       string alnsFileName = outputDir + m->getRootName(m->getSimpleName(fastaFileNames[s]))  + "uchime.alns";
                        
                        if (m->control_pressed) {  for (int j = 0; j < outputNames.size(); j++) {       remove(outputNames[j].c_str()); }  return 0;    }
                        
                        int numSeqs = 0;
 #if defined (__APPLE__) || (__MACH__) || (linux) || (__linux)
-                       if(processors == 1){ numSeqs = driver(outputFileName, fastaFileNames[s], accnosFileName); }
-                       else{   numSeqs = createProcesses(outputFileName, fastaFileNames[s], accnosFileName); }
+                       if(processors == 1){ numSeqs = driver(outputFileName, fastaFileNames[s], accnosFileName, alnsFileName); }
+                       else{   numSeqs = createProcesses(outputFileName, fastaFileNames[s], accnosFileName, alnsFileName); }
 #else
-                       numSeqs = driver(outputFileName, fastaFileNames[s], accnosFileName);
+                       numSeqs = driver(outputFileName, fastaFileNames[s], accnosFileName, alnsFileName);
 #endif
                        if (m->control_pressed) { for (int j = 0; j < outputNames.size(); j++) {        remove(outputNames[j].c_str()); } return 0; }
-
+                       
+                       //remove file made for uchime
+                       if (templatefile == "self") {  remove(fastaFileNames[s].c_str()); }
                        
                        outputNames.push_back(outputFileName); outputTypes["chimera"].push_back(outputFileName);
                        outputNames.push_back(accnosFileName); outputTypes["accnos"].push_back(accnosFileName);
+                       if (chimealns) { outputNames.push_back(alnsFileName); outputTypes["alns"].push_back(alnsFileName); }
                        
                        m->mothurOutEndLine(); m->mothurOut("It took " + toString(time(NULL) - start) + " secs to check " + toString(numSeqs) + " sequences."); m->mothurOutEndLine();
                }
@@ -418,7 +494,7 @@ int ChimeraUchimeCommand::execute(){
 }
 //**********************************************************************************************************************
 
-int ChimeraUchimeCommand::driver(string outputFName, string filename, string accnos){
+int ChimeraUchimeCommand::driver(string outputFName, string filename, string accnos, string alns){
        try {
                
                vector<char*> cPara;
@@ -453,6 +529,160 @@ int ChimeraUchimeCommand::driver(string outputFName, string filename, string acc
                strcpy(tempout, outputFName.c_str());
                cPara.push_back(tempout);
                
+               if (chimealns) {
+                       char* tempA = new char[12]; 
+                       strcpy(tempA, "--uchimealns"); 
+                       cPara.push_back(tempA);
+                       char* tempa = new char[alns.length()];
+                       strcpy(tempa, alns.c_str());
+                       cPara.push_back(tempa);
+               }
+               
+               if (useAbskew) {
+                       char* tempskew = new char[8]; 
+                       strcpy(tempskew, "--abskew"); 
+                       cPara.push_back(tempskew);
+                       char* tempSkew = new char[abskew.length()];
+                       strcpy(tempSkew, abskew.c_str());
+                       cPara.push_back(tempSkew);
+               }
+               
+               if (useMinH) {
+                       char* tempminh = new char[6]; 
+                       strcpy(tempminh, "--minh"); 
+                       cPara.push_back(tempminh);
+                       char* tempMinH = new char[minh.length()];
+                       strcpy(tempMinH, minh.c_str());
+                       cPara.push_back(tempMinH);
+               }
+               
+               if (useMindiv) {
+                       char* tempmindiv = new char[8]; 
+                       strcpy(tempmindiv, "--mindiv"); 
+                       cPara.push_back(tempmindiv);
+                       char* tempMindiv = new char[mindiv.length()];
+                       strcpy(tempMindiv, mindiv.c_str());
+                       cPara.push_back(tempMindiv);
+               }
+               
+               if (useXn) {
+                       char* tempxn = new char[4]; 
+                       strcpy(tempxn, "--xn"); 
+                       cPara.push_back(tempxn);
+                       char* tempXn = new char[xn.length()];
+                       strcpy(tempXn, xn.c_str());
+                       cPara.push_back(tempXn);
+               }
+               
+               if (useDn) {
+                       char* tempdn = new char[4]; 
+                       strcpy(tempdn, "--dn"); 
+                       cPara.push_back(tempdn);
+                       char* tempDn = new char[dn.length()];
+                       strcpy(tempDn, dn.c_str());
+                       cPara.push_back(tempDn);
+               }
+               
+               if (useXa) {
+                       char* tempxa = new char[4]; 
+                       strcpy(tempxa, "--xa"); 
+                       cPara.push_back(tempxa);
+                       char* tempXa = new char[xa.length()];
+                       strcpy(tempXa, xa.c_str());
+                       cPara.push_back(tempXa);
+               }
+               
+               if (useChunks) {
+                       char* tempchunks = new char[8]; 
+                       strcpy(tempchunks, "--chunks"); 
+                       cPara.push_back(tempchunks);
+                       char* tempChunks = new char[chunks.length()];
+                       strcpy(tempChunks, chunks.c_str());
+                       cPara.push_back(tempChunks);
+               }
+               
+               if (useMinchunk) {
+                       char* tempminchunk = new char[10]; 
+                       strcpy(tempminchunk, "--minchunk"); 
+                       cPara.push_back(tempminchunk);
+                       char* tempMinchunk = new char[minchunk.length()];
+                       strcpy(tempMinchunk, minchunk.c_str());
+                       cPara.push_back(tempMinchunk);
+               }
+               
+               if (useIdsmoothwindow) {
+                       char* tempidsmoothwindow = new char[16]; 
+                       strcpy(tempidsmoothwindow, "--idsmoothwindow"); 
+                       cPara.push_back(tempidsmoothwindow);
+                       char* tempIdsmoothwindow = new char[idsmoothwindow.length()];
+                       strcpy(tempIdsmoothwindow, idsmoothwindow.c_str());
+                       cPara.push_back(tempIdsmoothwindow);
+               }
+               
+               if (useMinsmoothid) {
+                       char* tempminsmoothid = new char[13]; 
+                       strcpy(tempminsmoothid, "--minsmoothid"); 
+                       cPara.push_back(tempminsmoothid);
+                       char* tempMinsmoothid = new char[minsmoothid.length()];
+                       strcpy(tempMinsmoothid, minsmoothid.c_str());
+                       cPara.push_back(tempMinsmoothid);
+               }
+               
+               if (useMaxp) {
+                       char* tempmaxp = new char[6]; 
+                       strcpy(tempmaxp, "--maxp"); 
+                       cPara.push_back(tempmaxp);
+                       char* tempMaxp = new char[maxp.length()];
+                       strcpy(tempMaxp, maxp.c_str());
+                       cPara.push_back(tempMaxp);
+               }
+               
+               if (!skipgaps) {
+                       char* tempskipgaps = new char[14]; 
+                       strcpy(tempskipgaps, "--[no]skipgaps"); 
+                       cPara.push_back(tempskipgaps);
+               }
+               
+               if (!skipgaps2) {
+                       char* tempskipgaps2 = new char[15]; 
+                       strcpy(tempskipgaps2, "--[no]skipgaps2"); 
+                       cPara.push_back(tempskipgaps2);
+               }
+               
+               if (useMinlen) {
+                       char* tempminlen = new char[8]; 
+                       strcpy(tempminlen, "--minlen"); 
+                       cPara.push_back(tempminlen);
+                       char* tempMinlen = new char[minlen.length()];
+                       strcpy(tempMinlen, minlen.c_str());
+                       cPara.push_back(tempMinlen);
+               }
+               
+               if (useMaxlen) {
+                       char* tempmaxlen = new char[8]; 
+                       strcpy(tempmaxlen, "--maxlen"); 
+                       cPara.push_back(tempmaxlen);
+                       char* tempMaxlen = new char[maxlen.length()];
+                       strcpy(tempMaxlen, maxlen.c_str());
+                       cPara.push_back(tempMaxlen);
+               }
+               
+               if (ucl) {
+                       char* tempucl = new char[5]; 
+                       strcpy(tempucl, "--ucl"); 
+                       cPara.push_back(tempucl);
+               }
+               
+               if (useQueryfract) {
+                       char* tempqueryfract = new char[12]; 
+                       strcpy(tempqueryfract, "--queryfract"); 
+                       cPara.push_back(tempqueryfract);
+                       char* tempQueryfract = new char[queryfract.length()];
+                       strcpy(tempQueryfract, queryfract.c_str());
+                       cPara.push_back(tempQueryfract);
+               }
+               
+               
                char** uchimeParameters;
                uchimeParameters = new char*[cPara.size()];
                for (int i = 0; i < cPara.size(); i++) {  uchimeParameters[i] = cPara[i];  } 
@@ -504,7 +734,7 @@ int ChimeraUchimeCommand::driver(string outputFName, string filename, string acc
 }
 /**************************************************************************************************/
 
-int ChimeraUchimeCommand::createProcesses(string outputFileName, string filename, string accnos) {
+int ChimeraUchimeCommand::createProcesses(string outputFileName, string filename, string accnos, string alns) {
        try {
                
                processIDS.clear();
@@ -526,7 +756,7 @@ int ChimeraUchimeCommand::createProcesses(string outputFileName, string filename
                MPI_Comm_size(MPI_COMM_WORLD, &processors); 
                                
                if (pid == 0) { //you are the root process 
-                       num = driver(outputFileName, files[0], accnos);
+                       num = driver(outputFileName, files[0], accnos, alns);
                        
                        if (templatefile != "self") {
                                //wait on chidren
@@ -540,11 +770,16 @@ int ChimeraUchimeCommand::createProcesses(string outputFileName, string filename
                                        
                                        m->appendFiles((accnos + toString(j) + ".temp"), accnos);
                                        remove((accnos + toString(j) + ".temp").c_str());
+                                       
+                                       if (chimealns) {
+                                               m->appendFiles((alns + toString(j) + ".temp"), alns);
+                                               remove((alns + toString(j) + ".temp").c_str());
+                                       }
                                }
                        }
                }else{ //you are a child process
                        if (templatefile != "self") { //if template=self we can only use 1 processor
-                               num = driver(outputFileName+toString(pid) + ".temp", files[pid], accnos+toString(pid) + ".temp");       
+                               num = driver(outputFileName+toString(pid) + ".temp", files[pid], accnos+toString(pid) + ".temp", alns+toString(pid) + ".temp"); 
                                
                                //send numSeqs to parent
                                MPI_Send(&num, 1, MPI_INT, 0, tag, MPI_COMM_WORLD);
@@ -562,7 +797,7 @@ int ChimeraUchimeCommand::createProcesses(string outputFileName, string filename
                                processIDS.push_back(pid);  //create map from line number to pid so you can append files in correct order later
                                process++;
                        }else if (pid == 0){
-                               num = driver(outputFileName + toString(getpid()) + ".temp", files[process], accnos + toString(getpid()) + ".temp");
+                               num = driver(outputFileName + toString(getpid()) + ".temp", files[process], accnos + toString(getpid()) + ".temp", alns + toString(getpid()) + ".temp");
                                
                                //pass numSeqs to parent
                                ofstream out;
@@ -580,7 +815,7 @@ int ChimeraUchimeCommand::createProcesses(string outputFileName, string filename
                }
                
                //do my part
-               num = driver(outputFileName, files[0], accnos);
+               num = driver(outputFileName, files[0], accnos, alns);
                
                //force parent to wait until all the processes are done
                for (int i=0;i<processIDS.size();i++) { 
@@ -604,6 +839,11 @@ int ChimeraUchimeCommand::createProcesses(string outputFileName, string filename
                        
                        m->appendFiles((accnos + toString(processIDS[i]) + ".temp"), accnos);
                        remove((accnos + toString(processIDS[i]) + ".temp").c_str());
+                       
+                       if (chimealns) {
+                               m->appendFiles((alns + toString(processIDS[i]) + ".temp"), alns);
+                               remove((alns + toString(processIDS[i]) + ".temp").c_str());
+                       }
                }
 #endif         
                //get rid of the file pieces.

diff --git a/chimerauchimecommand.h b/chimerauchimecommand.h
index 36e4a39aae24570e6f6cb5965f19261f0b3c9b63..a7d5ad60435f753d088b441aa5b509bce1999808 100644 (file)
--- a/chimerauchimecommand.h
+++ b/chimerauchimecommand.h
@@ -26,7 +26,7 @@ public:
        string getCommandName()                 { return "chimera.uchime";              }
        string getCommandCategory()             { return "Sequence Processing"; }
        string getHelpString(); 
-       string getCitation() { return "http://drive5.com/uchime/ \nhttp://www.mothur.org/wiki/Chimera.uchime"; }
+       string getCitation() { return "uchime by Robert C. Edgar\nhttp://drive5.com/uchime\nThis code is donated to the public domain.\nhttp://www.mothur.org/wiki/Chimera.uchime"; }
        
        
        int execute(); 
@@ -34,15 +34,11 @@ public:
        
 private:
        vector<int> processIDS;   //processid
-       int driver(string, string, string);
-       int createProcesses(string, string, string);
-       
-#ifdef USE_MPI
-       int driverMPI(int, int, MPI_File&, MPI_File&, MPI_File&, MPI_File&, vector<unsigned long int>&);
-#endif
-       
-       bool abort;
-       string fastafile, templatefile, outputDir, namefile;
+       int driver(string, string, string, string);
+       int createProcesses(string, string, string, string);
+               
+       bool abort, useAbskew, chimealns, useMinH, useMindiv, useXn, useDn, useXa, useChunks, useMinchunk, useIdsmoothwindow, useMinsmoothid, useMaxp, skipgaps, skipgaps2, useMinlen, useMaxlen, ucl, useQueryfract;
+       string fastafile, templatefile, outputDir, namefile, abskew, minh, mindiv, xn, dn, xa, chunks, minchunk, idsmoothwindow, minsmoothid, maxp, minlen, maxlen, queryfract;
        int processors;
        
        vector<string> outputNames;

diff --git a/diagbox.h b/diagbox.h
index 0c5846ce28e510509266238eb0ac4ce7151e483f..2491a58accce9c691c21f471af5eccda75354c81 100644 (file)
--- a/diagbox.h
+++ b/diagbox.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef diagbox_h\r
 #define diagbox_h\r
 \r
@@ -184,10 +186,10 @@ Don't have proof..
                }\r
        };\r
 \r
-typedef const char *(*NWDIAG)(const byte *A, unsigned LA, const byte *B, unsigned LB,
-  unsigned DiagLo, unsigned DiagHi, bool LeftTerm, bool RightTerm);
-
-const char *NWBandWrap(NWDIAG NW, const byte *A, unsigned LA, const byte *B, unsigned LB,
-  unsigned DiagLo, unsigned DiagHi, bool LeftTerm, bool RightTerm);
+typedef const char *(*NWDIAG)(const byte *A, unsigned LA, const byte *B, unsigned LB,\r
+  unsigned DiagLo, unsigned DiagHi, bool LeftTerm, bool RightTerm);\r
+\r
+const char *NWBandWrap(NWDIAG NW, const byte *A, unsigned LA, const byte *B, unsigned LB,\r
+  unsigned DiagLo, unsigned DiagHi, bool LeftTerm, bool RightTerm);\r
 \r
 #endif // diagbox_h\r

diff --git a/dp.h b/dp.h
index c771538a32c648f8bc6f195a368fb58a14385124..d8581fbc6fbc6e91be56acb3b59dd52af1df64d2 100644 (file)
--- a/dp.h
+++ b/dp.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef dp_h\r
 #define dp_h\r
 \r

diff --git a/evalue.h b/evalue.h
index c9308dbf4a1f7f89db5036f7e0e36bbd25f852cb..fe324f840f1da9b3886186068d26375bb53873a9 100644 (file)
--- a/evalue.h
+++ b/evalue.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef evalue_h\r
 #define evalue_h\r
 \r

diff --git a/fractid.cpp b/fractid.cpp
index f2988771cac56bbaeda7e68d8be866a3e7f666b2..ad4c1579c43c1b75d41583e700fdf4ac85cc49e9 100644 (file)
--- a/fractid.cpp
+++ b/fractid.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #include "myutils.h"\r
 #include "alpha.h"\r
 \r

diff --git a/getparents.cpp b/getparents.cpp
index d82f9028b25f1ff9c37f8ec100970df5d687c511..26a2cd1099cf8dc8b6a51891a59a99ea9a916fd7 100644 (file)
--- a/getparents.cpp
+++ b/getparents.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #include "myutils.h"\r
 #include "chime.h"\r
 #include "ultra.h"\r

diff --git a/globalalign2.cpp b/globalalign2.cpp
index 6bb35a910737c2511b950625f750ef2dd363ded0..10f8032f02deb7b300eba638ccf3e155dc21d2b2 100644 (file)
--- a/globalalign2.cpp
+++ b/globalalign2.cpp
@@ -1,5 +1,5 @@
 //#if  UCHIMES\r
-\r
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
 #include "dp.h"\r
 #include "seq.h"\r
 \r

diff --git a/help.h b/help.h
index 9d7a89f114d00d1e68fc74c75a6f50a585c141a2..1534cd171551736a145bbe33da893ac6bdd91fc5 100644 (file)
--- a/help.h
+++ b/help.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.
+
 "\n"
 "Usage\n"
 "-----\n"

diff --git a/hsp.h b/hsp.h
index 339256f20efc395b50d3f28f31a5f5302337eaf6..7f003bf1a085733d35fe9ac251e4c350a6af7ee4 100644 (file)
--- a/hsp.h
+++ b/hsp.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef hsp_h\r
 #define hsp_h  1\r
 \r

diff --git a/hspfinder.h b/hspfinder.h
index 2b8e9d84cac56d05d891beb9db9bdbf984a8bef5..dc287e991502f16c83fa7d678011eeaf4d9426d1 100644 (file)
--- a/hspfinder.h
+++ b/hspfinder.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.
+
 #ifndef hspfinder_h
 #define hspfinder_h
 

diff --git a/make3way.cpp b/make3way.cpp
index ce88f8616bd21f2362e3706d58c9b7917928bf00..dbadd1ab26269b10ee98229c638582ea8774b30b 100644 (file)
--- a/make3way.cpp
+++ b/make3way.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #include "myutils.h"\r
 #include "sfasta.h"\r
 #include "path.h"\r

diff --git a/mx.cpp b/mx.cpp
index 48c347e5832c461750cf72bd7b0a483013f7d57c..9e61d1beb368b75eb447f6f3159cfa4cecb74c25 100644 (file)
--- a/mx.cpp
+++ b/mx.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #include "myutils.h"\r
 #include "mx.h"\r
 #include "seqdb.h"\r

diff --git a/mx.h b/mx.h
index 143890086d59bb8d0f1d7f35d62cdd7dae3da1d0..24629377ed662bb9291704087243cbd95e9eef2e 100644 (file)
--- a/mx.h
+++ b/mx.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef mx_h\r
 #define mx_h\r
 \r

diff --git a/myopts.h b/myopts.h
index ba901ea48e38ad5b9f59b85dc7be43e208e70a4b..812dd8baed5386990ba4160048ac98d2e783af1a 100644 (file)
--- a/myopts.h
+++ b/myopts.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef MY_VERSION\r
 #define MY_VERSION     "4.2"\r
 #endif\r

diff --git a/myutils.cpp b/myutils.cpp
index ea983eb9e229a8549e1231aaaf75625d12135274..e9d52c2c936eb2d131109f7b46fe365d41931a9d 100755 (executable)
--- a/myutils.cpp
+++ b/myutils.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #include <time.h>\r
 #include <stdarg.h>\r
 #include <sys/stat.h>\r

diff --git a/myutils.h b/myutils.h
index b63ad3cff91af09741a5d53865916f9781a53973..61220546b81894283bb6350f5c2c51ecc8800cb2 100644 (file)
--- a/myutils.h
+++ b/myutils.h
@@ -1,8 +1,10 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef myutils_h\r
 #define myutils_h\r
 \r
 #define RCE_MALLOC     0\r
-
+\r
 #include <stdio.h>\r
 #include <sys/types.h>\r
 #include <string>\r
@@ -267,8 +269,8 @@ void GetCmdLine(string &s);
 \r
 extern const char *SVN_VERSION;\r
 extern const char *SVN_MODS;\r
-extern bool opt_quiet;
-extern bool opt_version;
-extern FILE *g_fLog;
+extern bool opt_quiet;\r
+extern bool opt_version;\r
+extern FILE *g_fLog;\r
 \r
 #endif // myutils_h\r

diff --git a/orf.h b/orf.h
index 90b29d1441589b41e7f64525134a80e0364e820b..06f73a1f9a733d3c0a6b0f3cdea5fc414f9b3cd4 100644 (file)
--- a/orf.h
+++ b/orf.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef orf_h\r
 #define orf_h\r
 \r

diff --git a/out.h b/out.h
index 4ca50c7c3d3d0937a0b94b63b1fe0ea93082b6d0..595e08c889849aba8ad25501b3bff742ff14f855 100644 (file)
--- a/out.h
+++ b/out.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef out_h\r
 #define out_h\r
 \r

diff --git a/path.cpp b/path.cpp
index 93403444ce9c0ce9780d896682ddde00ccc0bfdc..555830a9397eefa47fc52a9b1c590d7d50676ab4 100644 (file)
--- a/path.cpp
+++ b/path.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #include "myutils.h"\r
 #include "path.h"\r
 #include "timing.h"\r

diff --git a/path.h b/path.h
index f63be7ee109391164451d4397c16ab4715749ad3..b04ab74424691407f6a46a48588384ee630404a3 100644 (file)
--- a/path.h
+++ b/path.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef path_h\r
 #define path_h\r
 \r

diff --git a/searchchime.cpp b/searchchime.cpp
index c00a9c43b9e08cd6fd9e3cc997af5908861ba995..a1c981a5dc32f7b267c8c0a11e727870a426e4fe 100644 (file)
--- a/searchchime.cpp
+++ b/searchchime.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #include "myutils.h"\r
 #include "ultra.h"\r
 #include "chime.h"\r

diff --git a/seq.h b/seq.h
index 90146415cea631766f8d9ac8bd4c746e781eb8b5..83ba30fbafc3aa1317a6fe1f65e13907b7308069 100644 (file)
--- a/seq.h
+++ b/seq.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef seq_h\r
 #define seq_h\r
 \r

diff --git a/seqdb.cpp b/seqdb.cpp
index 03de189b8780bb0a898677a6afd520cb6bccd705..be3373a692b33097ef5677738b518cd549cfd914 100644 (file)
--- a/seqdb.cpp
+++ b/seqdb.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #include "myutils.h"\r
 #include "seqdb.h"\r
 #include "alpha.h"\r

diff --git a/seqdb.h b/seqdb.h
index fafbdd93c547d50f29104f450bb5ec7ff49c4e5d..6907015095f7f004ff9b1cd0d07450b0c4a1bb41 100644 (file)
--- a/seqdb.h
+++ b/seqdb.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef seqdb_h\r
 #define seqdb_h\r
 \r

diff --git a/setnucmx.cpp b/setnucmx.cpp
index 030ff5ab4dbb821cebaf539e0896792b7e0b65ba..123791d2b8dfafcb4313fc6d0fe45e092dafc573 100644 (file)
--- a/setnucmx.cpp
+++ b/setnucmx.cpp
@@ -1,11 +1,13 @@
-#include "myutils.h"
-#include "mx.h"
-
-Mx<float> g_SubstMxf;
-float **g_SubstMx;
-
-static const char Alphabet[] = "ACGTU";
-
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
+#include "myutils.h"\r
+#include "mx.h"\r
+\r
+Mx<float> g_SubstMxf;\r
+float **g_SubstMx;\r
+\r
+static const char Alphabet[] = "ACGTU";\r
+\r
 void SetNucSubstMx(double Match, double Mismatch)\r
        {\r
        static bool Done = false;\r

diff --git a/sfasta.cpp b/sfasta.cpp
index 5e794c67d5f5369df91f347e0a45b9365cb34ca0..95590404d57f43b94ed1c18a9d714cfed2bb11f7 100644 (file)
--- a/sfasta.cpp
+++ b/sfasta.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #include "sfasta.h"\r
 #include "orf.h"\r
 #include "alpha.h"\r
@@ -225,8 +227,9 @@ const byte *SFasta::GetNextSeqLo()
                        if (!WarningDone)\r
                                {\r
                                if (isgap(c))\r
-                                       Warning("Ignoring gaps in FASTA file '%s'",\r
-                                               m_FileName.c_str());\r
+                                       //Warning("Ignoring gaps in FASTA file '%s'",\r
+                                               //m_FileName.c_str());\r
+                                       ;\r
                                else if (isprint(c))\r
                                        Warning("Invalid FASTA file '%s', non-letter '%c' in sequence >%s",\r
                                          m_FileName.c_str(), c, Label);\r

diff --git a/sfasta.h b/sfasta.h
index ed2f2ffb56afea966a61b241e5b54645eb76f75c..79ab9617527efa014fad336156d8dd7064865748 100644 (file)
--- a/sfasta.h
+++ b/sfasta.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef sfasta_h\r
 #define sfasta_h\r
 \r

diff --git a/timing.h b/timing.h
index b566e1b801a932374736991076457f0efd9dda4c..10bd1f7705bed69363a08c9a2b3d2dd99cf3c054 100644 (file)
--- a/timing.h
+++ b/timing.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.
+
 #define TIMING 0
 #ifndef timing_h
 #define timing_h

diff --git a/tracebackbit.cpp b/tracebackbit.cpp
index 94159cd24aa99ab7584df620d6b576164b11531e..d58b9002612aa2f0e3fab29d7ea26441af40e6f9 100644 (file)
--- a/tracebackbit.cpp
+++ b/tracebackbit.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.
+
 #include "dp.h"
 
 #define TRACE  0

diff --git a/uc.h b/uc.h
index 631ea36a3024606f0d1635b70034305111aae818..8edef832487b8756f108ef1e1323d5ab4076dcee 100644 (file)
--- a/uc.h
+++ b/uc.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef uc_h\r
 #define uc_h\r
 \r

diff --git a/uchime_main.cpp b/uchime_main.cpp
index 40e7f44546668bb7d22d1a0b4e8800f29fc89e56..0465c720a6645c1ecd5a1d29155cbb0855748436 100644 (file)
--- a/uchime_main.cpp
+++ b/uchime_main.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #include "myutils.h"\r
 #include "chime.h"\r
 #include "seqdb.h"\r
@@ -107,11 +109,7 @@ int uchime_main(int argc, char *argv[])
                return 0;\r
                }\r
 \r
-       //printf("uchime v" MY_VERSION ".%s\n", SVN_VERSION);\r
-       //printf("by Robert C. Edgar\n");\r
-       //printf("http://drive5.com/uchime\n");\r
-       //printf("This code is donated to the public domain.\n");\r
-       //printf("\n");\r
+               \r
        if (!optset_w)\r
                opt_w = 8;\r
        \r
@@ -203,10 +201,16 @@ int uchime_main(int argc, char *argv[])
 \r
                WriteChimeHit(g_fUChime, Hit);\r
 \r
-               ProgressStep(i, QuerySeqCount, "%u/%u chimeras found (%.1f%%)", HitCount, i, Pct(HitCount, i+1));\r
-               \r
+               //ProgressStep(i, QuerySeqCount, "%u/%u chimeras found (%.1f%%)", HitCount, i, Pct(HitCount, i+1));\r
+                       //report progress\r
+                       if((i+1) % 100 == 0){   m->mothurOut("Processing sequence: " + toString(i+1) + ", " + toString(HitCount) + " chimeras found."); m->mothurOutEndLine();          }\r
+               }\r
+               if (!m->control_pressed) { \r
+                       //report progress\r
+                       if((QuerySeqCount) % 100 != 0){ m->mothurOut("Processing sequence: " + toString(QuerySeqCount) + ", " + toString(HitCount) + " chimeras found."); m->mothurOutEndLine();                }\r
                }\r
 \r
+\r
        Log("\n");\r
        Log("%s: %u/%u chimeras found (%.1f%%)\n",\r
          opt_input.c_str(), HitCount, QuerySeqCount, Pct(HitCount, QuerySeqCount));\r

diff --git a/usort.cpp b/usort.cpp
index 7afbf4279bb23aafdffdd474cdf7185ab226b6af..33f102e9363bc7ac39c3bf25f6b24059e647afd2 100644 (file)
--- a/usort.cpp
+++ b/usort.cpp
@@ -1,5 +1,5 @@
 //#if  UCHIMES\r
-\r
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
 #include "myutils.h"\r
 #include "seqdb.h"\r
 #include "seq.h"\r

diff --git a/viterbifast.cpp b/viterbifast.cpp
index 2b201741448e5bad79aa93b7c9887b4cdf2bda42..3f3fc970a2075a248bdace6b599773e04a84db01 100644 (file)
--- a/viterbifast.cpp
+++ b/viterbifast.cpp
@@ -1,9 +1,11 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.
+
 #include "dp.h"
 #include "out.h"
 #include "evalue.h"
 
 #define CMP_SIMPLE     0
-\r
+
 #if    SAVE_FAST
 static Mx<float> g_MxDPM;
 static Mx<float> g_MxDPD;
@@ -66,21 +68,21 @@ static void AllocSave(unsigned LA, unsigned LB)
        g_DPDSimple = g_DPDSimpleMx->GetData();
        g_DPISimple = g_DPISimpleMx->GetData();
 #endif
-       g_MxDPM.Alloc("FastM", LA+1, LB+1);\r
-       g_MxDPD.Alloc("FastD", LA+1, LB+1);\r
-       g_MxDPI.Alloc("FastI", LA+1, LB+1);\r
-\r
-       g_MxTBM.Alloc("FastTBM", LA+1, LB+1);\r
-       g_MxTBD.Alloc("FastTBD", LA+1, LB+1);\r
-       g_MxTBI.Alloc("FastTBI", LA+1, LB+1);\r
-\r
-       g_DPM = g_MxDPM.GetData();\r
-       g_DPD = g_MxDPD.GetData();\r
-       g_DPI = g_MxDPI.GetData();\r
-\r
-       g_TBM = g_MxTBM.GetData();\r
-       g_TBD = g_MxTBD.GetData();\r
-       g_TBI = g_MxTBI.GetData();\r
+       g_MxDPM.Alloc("FastM", LA+1, LB+1);
+       g_MxDPD.Alloc("FastD", LA+1, LB+1);
+       g_MxDPI.Alloc("FastI", LA+1, LB+1);
+
+       g_MxTBM.Alloc("FastTBM", LA+1, LB+1);
+       g_MxTBD.Alloc("FastTBD", LA+1, LB+1);
+       g_MxTBI.Alloc("FastTBI", LA+1, LB+1);
+
+       g_DPM = g_MxDPM.GetData();
+       g_DPD = g_MxDPD.GetData();
+       g_DPI = g_MxDPI.GetData();
+
+       g_TBM = g_MxTBM.GetData();
+       g_TBD = g_MxTBD.GetData();
+       g_TBI = g_MxTBI.GetData();
        }
 
 static void SAVE_DPM(unsigned i, unsigned j, float x)

diff --git a/windex.h b/windex.h
index 0b324ca9f7b720859cb54c34253dc6f1ba67f6ff..b1e002db276469c0ad73a17c96b1416556340923 100644 (file)
--- a/windex.h
+++ b/windex.h
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #ifndef windex_h\r
 #define windex_h\r
 \r

diff --git a/writechhit.cpp b/writechhit.cpp
index ea67061fe049908012cd7ac5b6fb8c1195c00de9..85ec26d7ed31ad1fda9349b0fb38e862e91d3ae5 100644 (file)
--- a/writechhit.cpp
+++ b/writechhit.cpp
@@ -1,3 +1,5 @@
+//uchime by Robert C. Edgar http://drive5.com/uchime This code is donated to the public domain.\r
+\r
 #include "myutils.h"\r
 #include "chime.h"\r
 \r