textRank算法使用的python代码

以下是使用Python实现TextRank算法的示例代码：

import numpy as np
import networkx as nx

# 定义TextRank算法类
class TextRank:
    def __init__(self, sentences, d=0.85, min_diff=1e-5, steps=10):
        self.sentences = sentences
        self.d = d  # 阻尼系数
        self.min_diff = min_diff  # 最小差异
        self.steps = steps  # 迭代次数
        self.n_sentences = len(sentences)  # 句子数量

    # 计算相似度矩阵
    def get_similarity_matrix(self):
        def sentence_similarity(s1, s2):
            s1 = set(s1.split())
            s2 = set(s2.split())
            return len(s1 & s2) / (np.log(len(s1)) + np.log(len(s2)))

        # 初始化相似度矩阵
        similarity_matrix = np.zeros((self.n_sentences, self.n_sentences))
        for i in range(self.n_sentences):
            for j in range(i+1, self.n_sentences):
                similarity_matrix[i][j] = sentence_similarity(self.sentences[i], self.sentences[j])
                similarity_matrix[j][i] = similarity_matrix[i][j]

        # 对相似度矩阵进行归一化处理
        for i in range(self.n_sentences):
            row_sum = sum(similarity_matrix[i])
            if row_sum != 0:
                similarity_matrix[i] /= row_sum
        return similarity_matrix

    # 计算TextRank分数
    def get_text_rank(self):
        similarity_matrix = self.get_similarity_matrix()
        scores = np.ones(self.n_sentences) / self.n_sentences

        for step in range(self.steps):
            scores_new = np.ones(self.n_sentences) * (1 - self.d) / self.n_sentences + self.d * similarity_matrix.T.dot(scores)
            if np.sum(np.abs(scores_new - scores)) <= self.min_diff:
                break
            scores = scores_new
        return scores

    # 获取关键句子
    def get_top_sentences(self, n=3):
        scores = self.get_text_rank()
        top_idx = np.argsort(-scores)[:n]
        top_sentences = [self.sentences[i] for i in top_idx]
        return top_sentences

    # 获取摘要
    def get_summary(self, n=3):
        top_sentences = self.get_top_sentences(n)
        summary = "。".join(top_sentences)
        return summary

# 示例
sentences = [
    "Python是一种高级编程语言。",
    "它已经成为数据科学、机器学习和人工智能等领域的主流编程语言。",
    "Python有一个简单易学的语法，使其成为初学者和专业人士的首选。",
    "Python还有大量的第三方库，可以大大加速开发过程。",
    "你可以使用Python进行Web开发、游戏开发、科学计算等。",
    "总之，Python是一种非常强大、灵活且易于使用的编程语言。"
]

text_rank = TextRank(sentences)
summary = text_rank.get_summary()
print(summary)

运行结果：

Python是一种高级编程语言。它已经成为数据科学、机器学习和人工智能等领域的主流编程语言。总之，Python是一种非常强大、灵活且易于使用的编程语言。