Love the tool! Super helpful. However, it bugs out if you try to run the maxsim disambiguation on a sentence where the wn.sysnet pos doesn't match the NLTK tagged pos.
Try running
sen = 'these potato chips are great'
disambiguate(sen, algorithm=maxsim)
and you get an index out of range error because result in max_similarity in similarity.py is [], because wn.synsets(ambiguous_word, pos=pos) is nothing as NLTK has (incorrectly) decided the part of speach of 'Potato' is an adjective, and there's no synset for that.
A very simple fix- change line 114 from
for i in wn.synsets(ambiguous_word, pos=pos):
to
for i in wn.synsets(ambiguous_word, pos=pos) or wn.synsets(ambiguous_word):
to provide a fallback option
Love the tool! Super helpful. However, it bugs out if you try to run the maxsim disambiguation on a sentence where the wn.sysnet pos doesn't match the NLTK tagged pos.
Try running
and you get an index out of range error because
resultinmax_similarityin similarity.py is[], becausewn.synsets(ambiguous_word, pos=pos)is nothing as NLTK has (incorrectly) decided the part of speach of 'Potato' is an adjective, and there's no synset for that.A very simple fix- change line 114 from
for i in wn.synsets(ambiguous_word, pos=pos):to
for i in wn.synsets(ambiguous_word, pos=pos) or wn.synsets(ambiguous_word):to provide a fallback option