Python script to map old to new URLs for migrating Disqus threads.

migrate_disqus.py

# coding: utf-8
import pandas as pd

def fix(url):
    # remove www subdomain
    url = url.replace('://www.', '://')
    # add slash if URL doesn't end with .html
    if not url.endswith('html'):
        url = url.rstrip('/') + '/'
    return url

df = pd.io.parsers.read_csv('ramiro.org.disqus.csv', header=None)
df['X.2'] = df['X.1'].map(fix)
df.to_csv('fixed.csv', header=False, index=False)