ner_similarity.py 文件源码

def extract_keywords(text):
    tokens = [i.lower() for i in nltk.word_tokenize(text) if i not in stop_words ]
    pos_tagged_tokens = nltk.pos_tag(tokens)
    result=[]
    for token in pos_tagged_tokens:
        # print token
        if token[1] in  POS_KEYS:
            result.append(token[0])

    return [ ps.stem(w) for w in result]