Replacing absolute URLs with relative URLs in Python

From , 3 Years ago, written in Python, viewed 86 times.

#!/usr/bin/env python
#-*- coding:utf-8 -*-
#
#
# author  : cold night
# email   : wh_linux@126.com
#
 
import pymongo
import re
from StringIO import StringIO
 
conn = pymongo.Connection()
db = conn.test
 
def replace_url():
    regex = re.compile(r'([href¦src])=["¦\']http://.*?(/m/getimg\?.*?)["¦\']')
    results = db['test'].find()
    db_coll = db['test']
    def replace(r):
        content = r.get('content')
        if not content: return
        content = StringIO(content)
        content.seek(0)
        result = StringIO()
        for line in content.readlines():
            t = regex.sub(r'\1="\2"', line)
            result.write(t)
 
        result.seek(0)
        content = result.read()
        if content:
            r['content'] = content
        _id = r.get('_id')
        db_coll.update({'_id':_id}, r)
 
    results = [replace(i) for i in results]
 
if __name__=="__main__":replace_url()
#//python/5323

Reply to "Replacing absolute URLs with relative URLs in Python"

Here you can reply to the paste above

Author What's your name?

Title Give your paste a title.

Language What language is your paste written in?

Your paste Paste your paste here

Create Shorturl Create a shorter url that redirects to your paste?

Private Private paste aren't shown in recent listings.

Delete After When should we delete your paste?

Spam protection Type in the letters

https://burned.cc - Burn After Reading Website

Pastebin.vip

Replacing absolute URLs with relative URLs in Python

Reply to "Replacing absolute URLs with relative URLs in Python"