load intersect is now much quicker. however there might be unexpected...

load intersect is now much quicker. however there might be unexpected behaviour in detect_gene_copies (or maybe elsewhere) if a segment is present several times in a feature (duplication within the feature). this case was not encountered in rice and human data, but it is technically possible.

GrAnnoT/Graph_gff.py

@@ -79,8 +79,10 @@ def add_child(feat,new_child):
 
 # add a segment to an existing feature
 def add_seg(feat_id,new_seg_oriented):
-    if new_seg_oriented not in Features[feat_id].segments_list_source:
-        Features[feat_id].segments_list_source.append(new_seg_oriented)
+    # if new_seg_oriented in Features[feat_id].segments_list_source:
+    #     print("seg already in feat, duplicate segment ! feature:",feat_id,"segment:",new_seg_oriented)
+    #     print("list for now :",Features[feat_id].segments_list_source)
+    Features[feat_id].segments_list_source.append(new_seg_oriented)
 
 # add the position of the feature on its first and last segment
 def add_pos(feature_id):
@@ -110,9 +112,9 @@ def add_segment_to_discontinuous_feature(line,feature,seg_oriented,current_start
         if (other_part.start==current_start) and (other_part.stop==current_stop): # look for the current part of the feature to add the segment
             add_seg(other_part_id,seg_oriented)
             added=True
+            break
     if added==False : # if the right part of the feature was not found, create a new part for the feature
         new_part_name=feature.id+''.join(random.choice(string.ascii_uppercase + string.digits) for _ in range(7))
-        feature.other_parts_list.append(new_part_name)
         [chr,start,stop]=[line[4],int(line[7]),int(line[8])]
         segments_list=[seg_oriented]
         Features[new_part_name]=Feature(feature.id,feature.type,chr,start,stop,feature.annot,feature.childs,feature.parent,segments_list,feature.strand,True,True)