[Python爬虫]高并发cnblogs博客备份工具（可扩展成并行）

并发爬虫小练习。

直接粘贴到本地，命名为.py文件即可运行，运行时的参数为你想要爬取的用户。默认是本博客。

输出是以用户名命名的目录，目录内便是博客内容。

仅供学习python的多线程编程方法，后续会重写成并行爬虫。

爬虫代码如下：

  1 # -*- coding:utf-8 -*-
  2 from multiprocessing.managers import BaseManager
  3 from pyquery import PyQuery
  4 import os, sys, urllib
  5 import re, random, logging, time
  6 import Queue, threading, multiprocessing, threadpool
  7
  8 USER_NAME = ‘kirai‘
  9 TOTAL_PAGE_NUMBER = 0
 10 INT_REGEXP = re.compile(‘([\d]+)‘)
 11 BASE_URL = ‘http://www.cnblogs.com/‘+USER_NAME+‘/p/?page=‘
 12 ARTICLE_REGEXP = re.compile(‘href=\"(http://www.cnblogs.com/‘+USER_NAME+‘/p/[\d]+.html)\"‘)
 13 THREAD_NUMBER = multiprocessing.cpu_count() * 2
 14 ARTICLE_URLS_MUTEX = threading.Lock()
 15 ARTICLE_URLS = []
 16
 17 class ListWithLinkExtend(list):
 18     def extend(self, value):
 19         super(ListWithLinkExtend, self).extend(value)
 20         return self
 21
 22 def get_total_page_number():
 23     doc = PyQuery(url=BASE_URL)
 24     return int(INT_REGEXP.findall(
 25         doc.find(‘.pager .Pager‘).text())[0].encode(‘ascii‘))
 26
 27 def get_page_url():
 28     global TOTAL_PAGE_NUMBER
 29     return map(lambda page: BASE_URL+str(page),
 30                          [i for i in range(1, TOTAL_PAGE_NUMBER+1)])
 31
 32 def get_article_url(idx):
 33     url = PAGE_URLS[idx]
 34     doc = PyQuery(url=url)
 35     article_urls = ARTICLE_REGEXP.findall(str(doc.find(‘.PostList .postTitl2‘)))
 36     return article_urls
 37
 38 def handle_result(request, result):
 39     global ARTICLE_URLS_MUTEX, ARTICLE_URLS
 40     try:
 41         ARTICLE_URLS_MUTEX.acquire()
 42         ARTICLE_URLS.append(result)
 43     finally:
 44         ARTICLE_URLS_MUTEX.release()
 45
 46 def cluster_process():
 47     global ARTICLE_URLS
 48     # list : urls
 49     task_queue = Queue.Queue()
 50     # str : path
 51     result_queue = Queue.Queue()
 52     KiraiManager.register(‘get_task_queue‘, callable=lambda: task_queue)
 53     KiraiManager.register(‘get_result_queue‘, callable=lambda: result_queue)
 54     manager = KiraiManager(address=(‘‘, 6969), authkey=‘whosyourdaddy‘)
 55     manager.start()
 56     manager.shutdown()
 57     # article_flag, article_urls = get_article_url()
 58
 59 # a simple way.
 60 def get_article(url):
 61     html = urllib.urlopen(url).read()
 62     return html, INT_REGEXP.findall(url)[0]
 63
 64 def save_article(request, result):
 65     content = result[0]
 66     file_name = result[1]
 67     path = ‘./‘ + USER_NAME + ‘/‘ + file_name + ‘.html‘
 68     try:
 69         fp = file(path, ‘w‘)
 70         fp.writelines(content)
 71     finally:
 72         fp.close()
 73
 74 def thread_process():
 75     global ARTICLE_URLS
 76     os.mkdir(USER_NAME)
 77     thread_pool = threadpool.ThreadPool(THREAD_NUMBER)
 78     requests = threadpool.makeRequests(get_article, ARTICLE_URLS, save_article)
 79     [thread_pool.putRequest(req) for req in requests]
 80     thread_pool.wait()
 81
 82 def __main__(argv):
 83     global ARTICLE_URLS, TOTAL_PAGE_NUMBER, USER_NAME, BASE_URL, ARTICLE_REGEXP, PAGE_URLS, TOTAL_PAGE_NUMBER
 84     if len(argv) == 2:
 85         USER_NAME = argv[1]
 86     BASE_URL = ‘http://www.cnblogs.com/‘+USER_NAME+‘/p/?page=‘
 87     ARTICLE_REGEXP = re.compile(‘href=\"(http://www.cnblogs.com/‘+USER_NAME+‘/p/[\d]+.html)\"‘)
 88     TOTAL_PAGE_NUMBER = get_total_page_number()
 89     PAGE_URLS = get_page_url()
 90     thread_pool = threadpool.ThreadPool(THREAD_NUMBER)
 91     requests = threadpool.makeRequests(
 92         get_article_url,
 93         [i for i in range(0, TOTAL_PAGE_NUMBER)],
 94         handle_result)
 95     [thread_pool.putRequest(req) for req in requests]
 96     thread_pool.wait()
 97     ARTICLE_URLS = list(reduce(lambda a, b: ListWithLinkExtend(a).extend(ListWithLinkExtend(b)),
 98                                                          ARTICLE_URLS))
 99     thread_process()
100
101 if __name__ == ‘__main__‘:
102     __main__(sys.argv)

简单介绍下全局变量的意义：

USER_NAME：希望爬取的用户名，默认为kirai。

TOTAL_PAGE_NUMBER：会被更新成博客随笔的总页数。

INT_REGEXP：为了匹配数字的正则。
BASE_URL：随笔页的初始URL。

ARTICLE_REGEXP：在经过pyquery处理过后的每个随笔目录页中提取出博客文章页面的正则。

THREAD_NUMBER：线程数，默认设置是本机cpu核数的2倍。

ARTICLE_URLS_MUTEX：ARTICLE_URLS的锁，保证线程唯一占用。

ARTICLE_URLS：用于存放所有的文章url。