ngsfilter: made more robust and practical to use with empty tags

python/obitools3/apps/optiongroups/__init__.py

@@ -55,7 +55,7 @@ def __addImportInputOption(optionManager):
                      action="store_const", dest="obi:inputformat",
                      default=None,
                      const=b'ngsfilter',
-                     help="Input file is an ngsfilter file")
+                     help="Input file is an ngsfilter file. If not using tags, use ':' or 'None:None' or '-:-' or any combination")
 
     group.add_argument('--ecopcr-result-input',
                      action="store_const", dest="obi:inputformat",

python/obitools3/commands/ngsfilter.pyx

@@ -57,6 +57,11 @@ def addOptions(parser):
                      default=None,
                      help="URI to the view used to store the sequences unassigned to any sample")
 
+    group.add_argument('--no-tags',
+                     action="store_true", dest="ngsfilter:notags",
+                     default=False,
+                     help="Use this option if your experiment does not use tags to identify samples")
+    
     group.add_argument('-e','--error',
                      action="store", dest="ngsfilter:error",
                      metavar="###",
@@ -167,7 +172,7 @@ cdef read_info_view(info_view, max_errors=2, verbose=False, not_aligned=False):
     i=0
     for p in info_view:
         forward=Primer(p[b'forward_primer'],
-                       len(p[b'forward_tag']) if p[b'forward_tag']!=b'-' else None,
+                       len(p[b'forward_tag']) if (b'forward_tag' in p and p[b'forward_tag']!=None) else None,
                        True,
                        max_errors=max_errors,
                        verbose=verbose,
@@ -178,7 +183,7 @@ cdef read_info_view(info_view, max_errors=2, verbose=False, not_aligned=False):
         infos[forward]=fp
         
         reverse=Primer(p[b'reverse_primer'],
-                       len(p[b'reverse_tag']) if p[b'reverse_tag']!=b'-' else None,
+                       len(p[b'reverse_tag']) if (b'reverse_tag' in p and p[b'reverse_tag']!=None) else None,
                        False,
                        max_errors=max_errors,
                        verbose=verbose,
@@ -213,9 +218,10 @@ cdef read_info_view(info_view, max_errors=2, verbose=False, not_aligned=False):
             rpp=rp.get(cf,{})
             rp[cf]=rpp
 
-        tags = (p[b'forward_tag'] if p[b'forward_tag']!=b'-' else None,
-                p[b'reverse_tag'] if p[b'reverse_tag']!=b'-' else None)
+        tags = (p[b'forward_tag'] if (b'forward_tag' in p and p[b'forward_tag']!=None) else None,
+                p[b'reverse_tag'] if (b'reverse_tag' in p and p[b'reverse_tag']!=None) else None)
         
+        if tags != (None, None):
             assert tags not in dpp, \
                "Tag pair %s is already used with primer pairs: (%s,%s)" % (str(tags),forward,reverse)
         
@@ -234,7 +240,7 @@ cdef read_info_view(info_view, max_errors=2, verbose=False, not_aligned=False):
     return infos, primer_list
     
 
-cdef tuple annotate(sequences, infos, verbose=False):
+cdef tuple annotate(sequences, infos, no_tags, verbose=False):
         
     def sortMatch(match):
         if match[1] is None:
@@ -430,7 +436,7 @@ cdef tuple annotate(sequences, infos, verbose=False):
         final_sequence[b'reversed'] = True   # used by the alignpairedend tool (in kmer_similarity.c)
 
     sample=None
-
+    if not no_tags:
         if tags[0] is not None:                                    # Direct  tag known
             if tags[1] is not None:                                # Reverse tag known
                 sample = samples.get(tags, None)             
@@ -572,6 +578,7 @@ def run(config):
     
     g = 0
     u = 0
+    no_tags = config['ngsfilter']['notags']
     try:
         for i in range(entries_len):
             PyErr_CheckSignals()
@@ -580,7 +587,7 @@ def run(config):
                 modseq = [Nuc_Seq.new_from_stored(forward[i]), Nuc_Seq.new_from_stored(reverse[i])]
             else:
                 modseq = [Nuc_Seq.new_from_stored(entries[i])]
-            good, oseq = annotate(modseq, infos)
+            good, oseq = annotate(modseq, infos, no_tags)
             if good:
                 o_view[g].set(oseq.id, oseq.seq, definition=oseq.definition, quality=oseq.quality, tags=oseq)
                 g+=1
@@ -596,6 +603,7 @@ def run(config):
     # Save command config in View and DMS comments
     command_line = " ".join(sys.argv[1:])
     o_view.write_config(config, "ngsfilter", command_line, input_dms_name=input_dms_name, input_view_name=input_view_name)
+    if unidentified is not None:
         unidentified.write_config(config, "ngsfilter", command_line, input_dms_name=input_dms_name, input_view_name=input_view_name)
         # Add comment about unidentified seqs
         unidentified.comments["info"] = "View containing sequences categorized as unidentified by the ngsfilter command"
@@ -607,6 +615,7 @@ def run(config):
     input[0].close()
     output[0].close()
     info_input[0].close()
+    if unidentified is not None:
         unidentified_input[0].close()
     aligner.free()
 

python/obitools3/parsers/ngsfilter.pyx

@@ -57,6 +57,9 @@ def ngsfilterIterator(lineiterator,
         split_line = line.split()
         tags = split_line.pop(2)
         tags = tags.split(b":")
+        for t_idx in range(2):
+            if tags[t_idx]==b"-" or tags[t_idx]==b"None" or tags[t_idx]==b"":
+                tags[t_idx] = nastring
         if len(tags) == 1:          # Forward and reverse tags are the same
             tags.append(tags[0])
         split_line.insert(2, tags[0])