Clearly define behaviour: leading and trailing whitespace in gazetteer entries #26
Description
The ANNIE DefaultGazetteer accepts trailing (and leading?) spaces as part of a "word".
I think we simply trim leading and trailing whitespace.
This should be clearly defined and justified:
- would we ever want to match a space token after/before a token as part of an entry?
More generally, the matching between space tokens and whitespace should be more clearly defined.