Expand URLs for Brave News RSS feeds

resolve-urls.py

#!/usr/bin/python3

import csv
import requests
from urllib.parse import urlparse, urlunparse

def feedburner_upgrade(url):
    parsed = urlparse(url)
    if parsed.netloc in ('feeds2.feedburner.com', 'feeds.feedburner.com', 'feedproxy.google.com'):
        return parsed._replace(scheme='https').geturl()
    return url

def resolve_url(url):
    try:
        r = requests.get(url, timeout=30)
        return r.url
    except:
        print("ERROR: %s" % url)
        return url

def main():
    with open('sources2.csv', 'w') as outfile:
        writer = csv.writer(outfile)
        with open('sources.csv', newline='') as infile:
            reader = csv.reader(infile)
            seen_first = False
            for row in reader:
                if seen_first:
                    url = row[1]
                    r_url = resolve_url(url)
                    final_url = feedburner_upgrade(r_url)
                    if final_url != url:
                        print("%s -> %s" % (url, final_url))
                        row[1] = final_url

                writer.writerow(row)
                seen_first = True

main()