def get_wordnet_sim(self, m1, m2, all_sim=True):
a1, a2 = [], []
for m in m1:
a1.append(Word(m).lemmatize("v"))
for m in m2:
a2.append(Word(m).lemmatize("v"))
syn1 = Word("_".join(a1)).synsets
syn2 = Word("_".join(a2)).synsets
if len(syn1) < 1:
for m in m1:
syn1.extend(Word(m).synsets)
if len(syn2) < 1:
for m in m2:
syn2.extend(Word(m).synsets)
sim_max, max_syn1, max_syn2 = self.get_max_synsets(syn1, syn2)
hyper_sim = None
verb_convert_sim = None
if all_sim:
if sim_max != None:
hyp1, _ = zip(*max_syn1.hypernym_distances())
hyp2, _ = zip(*max_syn2.hypernym_distances())
hyper_sim, _, _ = self.get_max_synsets(list(hyp1), list(hyp2))
# new similarity measure by converting everything to a verb
a1, a2 = self.convert_to_verb(max_syn1, m1), self.convert_to_verb(
max_syn2, m2)
verb_convert_sim, _, _ = self.get_wordnet_sim(a1, a2, False)
return sim_max, hyper_sim, verb_convert_sim
