$person_any seems to be messing with region extraction when used in a query
Example:
If I run pattern3_entity_in_favour_of_concept with the first slot being:
@0:[::] [pos_ark = "[DA@ONZ#]"]+ ([lemma="and|\&|,"]? [pos_ark = "[DA@ONZ#^]"]+)* @1:[::]
(i.e. pos_ark of the first token can be N; no use of the macro),
the fillers of the 0 slot can contain longer NPs:
1) my Twitter account
2) @janboehm and his tweet
If I remove N and do something like this instead:
@0:[::] [pos_ark = "[DA@OZ#]" | lemma=$person_any]+ ([lemma="and|\&|,"]? [pos_ark = "[DA@ONZ#^]"]+)* @1:[::]
all results of this kind disappear and it only returns hits with a single token in BOTH 0 and 1 (despite not changing anything in region 1, which goes from @2 to @3)
$person_any seems to be messing with region extraction when used in a query
Example: If I run pattern3_entity_in_favour_of_concept with the first slot being:
@0:[::] [pos_ark = "[DA@ONZ#]"]+ ([lemma="and|\&|,"]? [pos_ark = "[DA@ONZ#^]"]+)* @1:[::] (i.e. pos_ark of the first token can be N; no use of the macro), the fillers of the 0 slot can contain longer NPs:
1) my Twitter account 2) @janboehm and his tweet
If I remove N and do something like this instead: @0:[::] [pos_ark = "[DA@OZ#]" | lemma=$person_any]+ ([lemma="and|\&|,"]? [pos_ark = "[DA@ONZ#^]"]+)* @1:[::]
all results of this kind disappear and it only returns hits with a single token in BOTH 0 and 1 (despite not changing anything in region 1, which goes from @2 to @3)