Python crawler crawls American drama websites

I have always had the habit of watching American TV series. On the one hand, I can practice my English listening skills, and on the other hand, I can pass the time. It used to be possible to watch online on video websites, but since the restriction order imposed by the State Administration of Radio, Film and Television, it seems that imported American and British dramas are no longer updated simultaneously as before. However, as a nerd, how can I be willing to not follow any dramas, so I checked online and found an American drama download website [Tiantian American Dramas] that can be downloaded using Thunder. I can download various resources at will. Recently, I am obsessed with BBC’s High-definition documentary, nature is so beautiful.

Although I have found a resource website that can be downloaded, I have to open the browser every time, enter the URL, find the American drama, and then click the link to download. After a long time, the process becomes very cumbersome, and sometimes the website link cannot be opened, which is a bit troublesome. I happen to have been learning Python crawlers, so today I wrote a crawler on a whim to grab all the American drama links on the website and save them in a text document. If you want any drama, just open and copy the link to Xunlei to download it.

In fact, I originally planned to write something that finds a URL, uses requests to open it, grabs the download link, and crawls the entire site starting from the homepage. However, there are a lot of repeated links, and the URL of the website is not as regular as I thought. After writing for a long time, I still haven't written the kind of divergent crawler I want. Maybe I am not mature enough, so keep working hard. . .

Later I discovered that the links to the TV series were all in the article, and there was a number behind the article URL, like this http://cn163.net/archives/24016/, so I was smart and used Based on the crawler experience I wrote before, the solution is to automatically generate the URL. Can’t the number behind it be changed? Moreover, each drama is unique, so I tried to find out how many articles there are, and then use the range function to directly Continuously generate numbers to construct the url.

But many URLs do not exist, so they will hang up directly. Don’t worry, we are using requests, and its built-in status_code is used to determine the status returned by the request, so as long as it is the returned status We skip all those with code 404, and crawl the other links, which solves the URL problem.

The following is the implementation code of the above steps.

def get_urls(self):
  try:
    for i in range(2015,25000):
      base_url='http://cn163.net/archives/'
      url=base_url+str(i)+'/'
      if requests.get(url).status_code == 404:
        continue
      else:
        self.save_links(url)
  except Exception,e:
    pass

The rest went very smoothly. I found a similar crawler written by someone before on the Internet, but it only crawled one article, so I learned from it. Take a look at its regular expression. I used BeautifulSoup but the effect was not as good as the regular method, so I gave up decisively. There is no limit to my learning. However, the effect is not so ideal. About half of the links cannot be crawled correctly and need to continue to be optimized.

# -*- coding:utf-8 -*-
import requests 
import re
import sys
import threading
import time
reload(sys)
sys.setdefaultencoding('utf-8')
class Archives(object):

  def save_links(self,url):
    try:

      data=requests.get(url,timeout=3)
      content=data.text
      link_pat='"(ed2k://\|file\|[^"]+?\.(S\d+)(E\d+)[^"]+?1024X\d{3}[^"]+?)"'
      name_pat=re.compile(r'<h2 id="">(.*?)</h2>',re.S)
      links = set(re.findall(link_pat,content))
      name=re.findall(name_pat,content)
      links_dict = {}
      count=len(links)
    except Exception,e:
      pass
    for i in links:
      links_dict[int(i[1][1:3]) * 100 + int(i[2][1:3])] = i#把剧集按s和e提取编号
    try:
      with open(name[0].replace('/',' ')+'.txt','w') as f:
        print name[0]
        for i in sorted(list(links_dict.keys())):#按季数+集数排序顺序写入
          f.write(links_dict[i][0] + '\n')
      print "Get links ... ", name[0], count
    except Exception,e:
      pass

  def get_urls(self):
    try:
      for i in range(2015,25000):
        base_url='http://cn163.net/archives/'
        url=base_url+str(i)+'/'
        if requests.get(url).status_code == 404:
          continue
        else:
          self.save_links(url)
    except Exception,e:
      pass
  def main(self):
    thread1=threading.Thread(target=self.get_urls())
    thread1.start()
    thread1.join()
  if __name__ == '__main__':
  start=time.time()
  a=Archives()
  a.main()
  end=time.time()
  print end-start

The complete version of the code also uses multi-threading, but it feels useless. It seems to be more than 20,000 parts because of Python's GIL. I thought it would take a long time to complete the crawling, but excluding the URL errors and unmatched URLs, the total crawling time was less than 20 minutes. I originally wanted to use Redis to crawl on two Linux machines, but after a lot of fussing, I felt it was unnecessary, so I left it at that and will do it later when I need more data.

Another problem that tortured me during the process was the saving of file names. I must complain here. File names in txt text format can have spaces, but they cannot have slashes, backslashes, brackets etc. This is the problem. I spent the whole morning on this. At first I thought it was an error in crawling the data. After checking for a long time, I found out that the crawled drama title had a slash in it. This made me miserable. .

For more articles related to Python crawlers crawling American drama websites, please pay attention to the PHP Chinese website!