Method get_tfidf_vec

build_graph.py:139–175

学习获得tfidf矩阵，及其对应的单词序列 :param content_lst: :return:

(self)

        print_graph_detail(self.g)

    def get_tfidf_vec(self):
        """
        学习获得tfidf矩阵，及其对应的单词序列
        :param content_lst:
        :return:
        """
        start = time()
        text_tfidf = Pipeline([
            ("vect", CountVectorizer(min_df=1,
                                     max_df=1.0,
                                     token_pattern=r"\S+",
                                     )),
            ("tfidf", TfidfTransformer(norm=None,
                                       use_idf=True,
                                       smooth_idf=False,
                                       sublinear_tf=False
                                       ))
        ])

        tfidf_vec = text_tfidf.fit_transform(open(self.content, "r"))

        self.tfidf_time = time() - start
        print("tfidf time:", self.tfidf_time)
        print("tfidf_vec shape:", tfidf_vec.shape)
        print("tfidf_vec type:", type(tfidf_vec))

        self.node_num = tfidf_vec.shape[0]

        # 映射单词
        vocab_lst = text_tfidf["vect"].get_feature_names()
        print("vocab_lst len:", len(vocab_lst))
        for ind, word in enumerate(vocab_lst):
            self.word2id[word] = ind

        self.vocab_lst = vocab_lst

        return tfidf_vec

    def save(self):
        print("total time:", self.pmi_time + self.tfidf_time)