#!/usr/bin/env python3

import os
import re
import requests

urls = set()

def check_url(url):
    try:
        r = requests.get(url, timeout = 5)
        print(r)
        return True
    except requests.exceptions.ConnectionError as e:
        print(e)
        return False

for root, dirs, files in os.walk('.'):
    for f in files:
        if f.endswith('.html'):
            full = os.path.join(root, f)
            for l in open(full).readlines():
                m = re.search(r'"(http:\/\/[^"]*)"', l)
                if m:
                    urls.add(m.group(1))

urls = list(sorted(list(urls)))
to_replace = []

for i, url in enumerate(urls):
    https = url.replace('http://', 'https://')
    print('%d/%d: %s' % (i, len(urls), https))
    if (check_url(https)):
        to_replace.append(url)

print()
print('Can use HTTPS for %d of links:' % (len(to_replace)))
for url in to_replace:
    print(url)