爬豆瓣影评，记下解决maximum recursion depth exceeded in cmp

#主要是爬取后给别人做自然语言分析，没其他意思。

#coding=utf8import requests,re
from lxml import etree

import sys
reload(sys)
sys.setdefaultencoding('utf8')

sys.setrecursionlimit(10000000)  #解决maximum recursion depth exceeded in cmp
def craw(url):
    headerx={
'Cookie':'bid=OIBtzThxxA; ct=y; _pk_ref.100001.4cf6=%5B%22%22%2C%22%22%2C1502186407%2C%22http%3A%2F%2Fqianxun.baidu.com%2Fmovie%2Fcard_2162.html%22%5D; __utmt=1; ps=y; dbcl2="165xxx93:UV/wbzXasBQ"; ck=d-ep; _pk_id.100001.4cf6=7bff167cf6dxxxxxx.10.1502186411; __utmc=30149280; __utmz=30149280.1501649654.7.7.utmcsr=baidu|utmccn=(organic)|utmcmd=organic; __utma=22369xxxxx59553.1502186411.2; __utmb=223695111.48.10.1502186411; __utmc=223695111; __utmz=223695111.1500959553.1.1.utmcsr=qianxun.baidu.com|utmccn=(referral)|utmcxxxxx.html; push_noty_num=0; push_doumail_num=0; ap=1'
    }
while 1:
try:
            resp=requests.get(url,headers=headerx)
if resp.status_code==200:
                flag=1
break
else:
                pass
        except Exception,e:
            print e

    selector=etree.HTML(resp.content.decode('utf8'))


    #print resp.content.decode('utf8')
    all_comment=selector.xpath('//div[@class="comment-item"]')
for comment in all_comment:
        #print  etree.tounicode(comment),'************************'
        star_class=comment.xpath('.//span[contains(@class,"allstar")]/@class')
if star_class:
            starx=re.findall('tar(.*?)0',star_class[0])[0]
else:
            starx=0   #有的评论没有打星
        textx=comment.xpath('.//div[@class="comment"]/p/text()')[0]
        print starx,'星  ',textx

        f.write('%s星 %s\r\n'%(starx,textx))


    next_start=re.search(u'前页[\s\S]*?<a href="\?start=(.*?)&amp[\s\S]*?后页',resp.content.decode('utf8')).group(1)
    next_url='https://movie.douban.com/subject/25662329/comments?start=%s&limit=20&sort=new_score&status=P'%next_start
    print '爬取下一页：',next_url
    craw(next_url)                            ####这行调用自己

if __name__=="__main__":
  f = open('pinlun.txt', 'a')
  craw('https://movie.douban.com/subject/25662329/comments?start=71726&limit=20&sort=new_score&status=P')

 

豆瓣的影评，每一页的页面链接不是可以预知的，url中start从0开始，第二页是20，第三页是40，这是理想状态,但真实不是这样。所以要从网页中提取下一页的链接，单线程，自己调用自己，每爬了几十分钟后报错maximum recursion depth exceeded in cmp，以为是个偶然，反复把脚本停了再重启了好几次都是这样，就找下答案了，上面是函数里面有个调用函数本身的代码，反复的调用自己，这种次数不能超过900，设置了sys.setrecursionlimit=100000果然好了。