We had a bug report that 12 main st, boston ma does not parse correctly, however 12 main st, boston massachusetts does.
The issue seems to be caused by the ma token being interpreted as a place by a WordClassifier, once it's tagged as a place there is additional logic preventing it from also being considered a locality.
I think the original intention was to boost venue parsing, but it's not correct that it doesn't consider ma as either magistratsabteilung|ma or massachusetts.
This PR is a simple fix, albeit somewhat cosmetic, removing the ma token from the place token index.
We had a bug report that
12 main st, boston ma
does not parse correctly, however12 main st, boston massachusetts
does.The issue seems to be caused by the
ma
token being interpreted as aplace
by aWordClassifier
, once it's tagged as aplace
there is additional logic preventing it from also being considered alocality
.I think the original intention was to boost venue parsing, but it's not correct that it doesn't consider
ma
as eithermagistratsabteilung|ma
ormassachusetts
.This PR is a simple fix, albeit somewhat cosmetic, removing the
ma
token from theplace
token index.