def lemmatize(keywords,add_synsets=False):
# lemmatizer = WordNetLemmatizer()
new_keywords = []
for word in keywords:
new_keywords.append(lemmatizer.lemmatize(word))
if add_synsets:
synsets = wn.synsets(word)
for syn in synsets:
for lemma in syn.lemmas:
new_keywords += lemma.name.lower().split("_")
new_keywords = remove_duplicates_list(new_keywords)
return new_keywords
def get_wordnet_def_entity_keywords(entity):
synsets = wn.synsets(entity)
if len(synsets) > 0:
keywords = []
for synset in synsets:
for lemma in synset.lemmas:
keywords.extend( lemma.name.lower().split("_") )
definition = synset.definition.lower()
tokenized_def = nltk.word_tokenize(definition)
def_pos = nltk.pos_tag(tokenized_def)
for (word, pos) in def_pos:
if (pos in ['NN','NNP','NNPS','NNS']) and word != '(': # for some reason nltk tags ( as NN
keywords.append(word)
return remove_duplicates_list(keywords)
else:
return None
