@@ -56,7 +56,11 @@ def convert_json_to_csv(jsonfile):
5656 # expand the target name and skip the next relation
5757 start_target = int (r ['target_ids' ][0 ].split ('_' )[1 ])
5858 end_target = int (r ['target_ids' ][0 ].split ('_' )[2 ])
59+ # These are arrays, but for auto-annotations,
60+ # they will only ever have one item
5961 targ_name = r ['target_names' ][0 ]
62+ cont_name = r ['cont_names' ][0 ]
63+
6064 if start_target in skip_inds :
6165 continue
6266 next_rels = [r2 for r2 in rels if
@@ -76,14 +80,19 @@ def convert_json_to_csv(jsonfile):
7680 next_rel ['target_names' ][0 ]))
7781 targ_name += ' ' + next_rel ['target_names' ][0 ]
7882 skip_inds .append (start_next_target )
79-
83+
84+ # If cont_name is something like Fe-rich or Mg_sulfate,
85+ # only keep the first bit.
86+ if '-' in cont_name :
87+ cont_name = cont_name [:cont_name .find ('-' )]
88+ elif '_' in cont_name :
89+ cont_name = cont_name [:cont_name .find ('_' )]
90+
8091 outf .write (',%s,%s,%s,"%s"\n ' %
8192 (docid ,
82- # These are arrays, but for auto-annotations,
83- # they will only ever have one item
8493 #r['target_names'][0],
8594 targ_name ,
86- r [ 'cont_names' ][ 0 ] ,
95+ cont_name ,
8796 r ['sentence' ]))
8897 # build URL manually?
8998 pbar .update (i )
0 commit comments