switch to version 3.0.1b4

primer would be missed

python/obitools3/commands/alignpairedend.pyx

@@ -205,10 +205,16 @@ def run(config):
     if type(entries) == list:
         forward = entries[0]
         reverse = entries[1]
+        if len(forward) == 0 or len(reverse) == 0:
+            aligner = None
+        else:
             aligner = Kmer_similarity(forward, \
                                       view2=reverse, \
                                       kmer_size=config['alignpairedend']['kmersize'], \
                                       reversed_column=None)
+    else:
+        if len(entries) == 0:
+            aligner = None
         else:
             aligner = Kmer_similarity(entries, \
                                       column2=entries[REVERSE_SEQUENCE_COLUMN], \
@@ -251,7 +257,7 @@ def run(config):
         pb(i, force=True)
         print("", file=sys.stderr)
 
-    if kmer_ali :
+    if kmer_ali and aligner is not None:
         aligner.free()
 
     # Save command config in View and DMS comments

python/obitools3/commands/ngsfilter.pyx

@@ -394,7 +394,7 @@ cdef tuple annotate(sequences, infos, no_tags, verbose=False):
             seq_to_match = sequences[0]
         reversematch = []
         # Compute begin
-        begin=directmatch[1][2]+1  # end of match + 1 on the same sequence
+        #begin=directmatch[1][2]+1  # end of match + 1 on the same sequence -- No, already cut out forward primer        
         # Try reverse matching on the other sequence:
         new_seq = True
         pattern = 0
@@ -408,7 +408,7 @@ cdef tuple annotate(sequences, infos, no_tags, verbose=False):
                 primer=p
             # Saving original primer as 4th member of the tuple to serve as correct key in infos dict even if it might have been reversed complemented
             # (3rd member already used by directmatch)
-            reversematch.append((primer, primer(seq_to_match, same_sequence=not new_seq, pattern=pattern, begin=begin), None, p))
+            reversematch.append((primer, primer(seq_to_match, same_sequence=not new_seq, pattern=pattern, begin=0), None, p))
             new_seq = False
             pattern+=1
         # Choose match closer to the end of the sequence
@@ -645,6 +645,7 @@ def run(config):
     
     g = 0
     u = 0
+    i = 0
     no_tags = config['ngsfilter']['notags']
     try:
         for i in range(entries_len):

python/obitools3/commands/uniq.pyx

@@ -354,8 +354,8 @@ cdef uniq_sequences(View_NUC_SEQS view, View_NUC_SEQS o_view, ProgressBar pb, di
         key = mergedKeys[k]
         merged_col_name = mergedKeys_m[k]
         
-        if merged_infos[merged_col_name]['nb_elts'] == 1:
+#        if merged_infos[merged_col_name]['nb_elts'] == 1:
-            raise Exception("Can't merge information from a tag with only one element (e.g. one sample ; don't use -m option)")
+#            raise Exception("Can't merge information from a tag with only one element (e.g. one sample ; don't use -m option)")
         
         if merged_col_name in view:
             i_col = view[merged_col_name]

python/obitools3/parsers/tab.pyx

@@ -8,7 +8,7 @@ Created on feb 20th 2018
 
 import types
 from obitools3.utils cimport __etag__
-
+from obitools3.utils cimport str2bytes
 
 def tabIterator(lineiterator, 
                 bint header = False,
@@ -75,7 +75,7 @@ def tabIterator(lineiterator,
                 continue
             else:
                 # TODO ??? default column names? like R?
-                keys = [i for i in range(len(line.split(sep)))]
+                keys = [str2bytes(str(i)) for i in range(len(line.split(sep)))]
                 
         while skipped < skip :
             line = next(iterator)

python/obitools3/parsers/universal.pyx

@@ -53,7 +53,11 @@ def entryIteratorFactory(lineiterator,
 
     i = iterator
         
+    try:
         first=next(i)
+    except StopIteration:
+        first=""
+        pass
 
     format=b"tabular"
     

python/obitools3/version.py

@@ -1,5 +1,5 @@
 major = 3
 minor = 0
-serial= '0b43'
+serial= '1b4'
 
 version ="%d.%d.%s" % (major,minor,serial)

src/obi_clean.c

@@ -229,6 +229,8 @@ int obi_clean(const char* dms_name,
 		return -1;
 	}
 
+	seq_count = (i_view->infos)->line_count;
+
 	// Open the sequence column
 	if (strcmp((i_view->infos)->view_type, VIEW_TYPE_NUC_SEQS) == 0)
 		iseq_column = obi_view_get_column(i_view, NUC_SEQUENCE_COLUMN);
@@ -245,7 +247,7 @@ int obi_clean(const char* dms_name,
 	}
 
 	// Open the sample column if there is one
-	if ((strcmp(sample_column_name, "") == 0) || (sample_column_name == NULL))
+	if ((strcmp(sample_column_name, "") == 0) || (sample_column_name == NULL) || (seq_count == 0))
 	{
 		fprintf(stderr, "Info: No sample information provided, assuming one sample.\n");
 		sample_column = obi_view_get_column(i_view, COUNT_COLUMN);
@@ -340,6 +342,8 @@ int obi_clean(const char* dms_name,
 		return -1;
 	}
 
+	if (seq_count > 0)
+	{
 		// Build kmer tables
 		ktable = hash_seq_column(i_view, iseq_column, 0);
 		if (ktable == NULL)
@@ -349,8 +353,6 @@ int obi_clean(const char* dms_name,
 			return -1;
 		}
 
-	seq_count = (i_view->infos)->line_count;
-
 		// Allocate arrays for sample counts otherwise reading in mapped files takes longer
 		complete_sample_count_array = (int*) malloc(seq_count * sample_count * sizeof(int));
 		if (complete_sample_count_array == NULL)
@@ -404,6 +406,7 @@ int obi_clean(const char* dms_name,
 				}
 			}
 		}
+	}
 
 	for (i=0; i< (seq_count-1); i++)
 	{
@@ -551,17 +554,20 @@ int obi_clean(const char* dms_name,
 		}
 	}
 
+	if (seq_count > 0)
+	{
 		free_kmer_tables(ktable, seq_count);
 		free(complete_sample_count_array);
 		free(blob_array);
 		free(alignment_result_array);
+	}
 
 	fprintf(stderr, "\n");
 
 	if (stop)
 		return -1;
 
-	if (heads_only)
+	if (heads_only && (seq_count > 0))
 	{
 		line_selection = malloc((((o_view->infos)->line_count) + 1) * sizeof(index_t));
 		if (line_selection == NULL)
@@ -635,7 +641,7 @@ int obi_clean(const char* dms_name,
 	}
 
 	// Flag the end of the line selection
-	if (heads_only)
+	if (heads_only && (seq_count > 0))
 		line_selection[l] = -1;
 
 	// Create new view with line selection if heads only