[Python爬虫]新闻网页爬虫+jieba分词+关键词搜索排序

1,373 阅读2分钟

前言

最近做了一个python3作业题目,涉及到:

  • 网页爬虫
  • 网页中文文字提取
  • 建立文字索引
  • 关键词搜索

涉及到的库有:

  • 爬虫库:requests
  • 解析库:xpath
  • 正则:re
  • 分词库:jieba
  • ...

放出代码方便大家快速参考,实现一个小demo。

题目描述

搜索引擎的设计与实现

  1. 输入:腾讯体育的页面链接,以列表的方式作为输入,数量不定,例如:
["http://fiba.qq.com/a/20190420/001968.htm","http://sports.qq.com/a/20190424/000181.htm","http://sports.qq.com/a/20190423/007933.htm","http://new.qq.com/omn/SPO2019042400075107"]
  1. 过程:网络爬虫,页面分析、中文提取分析、建立索引,要求应用教材中的第三方库,中间过程在内存中完成,输出该过程的运行时间;

  2. 检索:提示输入一个关键词进行检索;

  3. 输出:输入的链接列表的按照关键词的出现频率由高到低排序输出,并以JSON格式输出词频信息等辅助信息;未出现关键词的文档链接不输出,最后输出检索时间,例如:

1 "http:xxxxxx.htm" 3
2 "https:xxxx.htm" 2
3 "https:xxxxx.htm" 1

代码

代码实现的主要步骤是:

  • 网页爬虫:crawler函数
  • 网页文本元素清洗:清理掉多余的英文字符和标签,bs4_page_clean函数
  • 用正则提取中文:re_chinese函数
  • 使用dict保存每个网页的中文字和词,做索引:jieba_create_index函数
  • 输入关键词进行搜索:search函数
import requests
from bs4 import BeautifulSoup
import json
import re
import jieba
import time

USER_AGENT = {'user-agent': 'Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/536.6 (KHTML, like Gecko) '
                            'Chrome/20.0.1092.0 Safari/536.6'}
URL_TIMEOUT = 10
SLEEP_TIME = 2

# dict_result格式:{"1":
#                       {"url": "xxxxx", "word": {"word1": x, "word2": x, "word3": x}}
#                  "2":
#                       {"url": "xxxxx", "word": {"word1": x, "word2": x, "word3": x}}
#                 }
dict_result = {}

# dict_search格式:[#                   [url, count]
#                   [url, count]
#                 ]
list_search_result = []


def crawler(list_URL):
    for i, url in enumerate(list_URL):
        print("网页爬取:", url, "...")
        page = requests.get(url, headers=USER_AGENT, timeout=URL_TIMEOUT)
        page.encoding = page.apparent_encoding  # 防止编码解析错误
        result_clean_page = bs4_page_clean(page)
        result_chinese = re_chinese(result_clean_page)
        # print("网页中文内容:", result_chinese)
        dict_result[i + 1] = {"url": url, "word": jieba_create_index(result_chinese)}
        print("爬虫休眠中...")
        time.sleep(SLEEP_TIME)


def bs4_page_clean(page):
    print("正则表达式:清除网页标签等无关信息...")
    soup = BeautifulSoup(page.text, "html.parser")
    [script.extract() for script in soup.findAll('script')]
    [style.extract() for style in soup.findAll('style')]
    reg1 = re.compile("<[^>]*>")
    content = reg1.sub('', soup.prettify())
    return str(content)


def re_chinese(content):
    print("正则表达式:提取中文...")
    pattern = re.compile(u'[\u1100-\uFFFD]+?')
    result = pattern.findall(content)
    return ''.join(result)


def jieba_create_index(string):
    list_word = jieba.lcut_for_search(string)
    dict_word_temp = {}
    for word in list_word:
        if word in dict_word_temp:
            dict_word_temp[word] += 1
        else:
            dict_word_temp[word] = 1
    return dict_word_temp


def search(string):
    for k, v in dict_result.items():
        if string in v["word"]:
            list_search_result.append([v["url"], v["word"][string]])
    # 使用词频对列表进行排序
    list_search_result.sort(key=lambda x: x[1], reverse=True)

if __name__ == "__main__":

    list_URL_sport = input("请输入网址列表:")
    list_URL_sport = list_URL_sport.split(",")
    print(list_URL_sport)
    # 删除输入的网页双引号
    for i in range(len(list_URL_sport)):
        list_URL_sport[i] = list_URL_sport[i][1:-1]
    print(list_URL_sport)
    # list_URL_sport = ["http://fiba.qq.com/a/20190420/001968.htm",
    #                   "http://sports.qq.com/a/20190424/000181.htm",
    #                   "http://sports.qq.com/a/20190423/007933.htm",
    #                   "http://new.qq.com/omn/SPO2019042400075107"]
    time_start_crawler = time.time()
    crawler(list_URL_sport)
    time_end_crawler = time.time()
    print("网页爬取和分析时间:", time_end_crawler - time_start_crawler)
    word = input("请输入查询的关键词:")
    time_start_search = time.time()
    search(word)
    time_end_search = time.time()
    print("检索时间:", time_end_search - time_start_search)
    for i, row in enumerate(list_search_result):
        print(i+1, row[0], row[1])
    print("词频信息:")
    print(json.dumps(dict_result, ensure_ascii=False))

运行结果

关注我

个人博客: