little-things/github-list-repos

#!/usr/bin/env python3
import logging
import re
import requests
import sys
import time


GIT_URLS_OPTION = '--git-urls'
GITGUD_COMPLETE_ITEMS_OPTION = '--gitgud-complete-items'
MODES = (GIT_URLS_OPTION, GITGUD_COMPLETE_ITEMS_OPTION)


mode = None
users = sys.argv[1:]
if users and users[0] in MODES:
	mode = users[0]
	users = users[1:]
assert users and (mode is None or mode in MODES) and not users[0].startswith('--'), f'Usage: github-list-repos [{" | ".join(MODES)}] USER [USER...]'


def get(url):
	while True:
		logging.info(f'Fetching {url}')
		r = requests.get(url, headers = {'User-Agent': 'Mozilla/5.0 (Windows NT 6.1; rv:60.0) Gecko/20100101 Firefox/60.0'})
		if r.status_code == 429:
			logging.warning(f'Got 429, sleeping and retrying')
			time.sleep(5)
		else:
			break
	return r


def p(repoName):
	if mode is None:
		print(f'https://github.com/{repoName}')
	elif mode == GIT_URLS_OPTION:
		print(f'https://github.com/{repoName}.git')
		print(f'https://github.com/{repoName}.wiki.git')
	elif mode == GITGUD_COMPLETE_ITEMS_OPTION:
		print(f'web:complete:{repoName}')


for user in users:
	r = get(f'https://github.com/{user}')
	if '<div id="org-repositories"' in r.text:
		# Organisation, complete list under /orgs/ with ?page=2 pagination
		r = get(f'https://github.com/orgs/{user}/repositories')
		page = 1
		while True:
			for m in re.finditer(r'<a itemprop="name codeRepository"\s(?:[^>]*\s)?data-hovercard-url="/([^/>"]+/[^/>"]+)/hovercard"', r.text):
				p(m.group(1))
			if '<a class="next_page"' not in r.text:
				# End of pagination
				break
			page += 1
			r = get(f'https://github.com/orgs/{user}/repositories?page={page}')
	else:
		# User, ?tab=repositories + cursor pagination
		r = get(f'https://github.com/{user}?tab=repositories')
		while True:
			for m in re.finditer(r'<a href="/([^/>"]+/[^/>"]+)" itemprop="name codeRepository"(\s[^>]*)?>', r.text):
				p(m.group(1))
			if not (m := re.search(r'<a\s(?=(?:[^>]*\s)?class="next_page"(?:\s[^>]*)?>)(?:[^>]*\s)?href="/[^/?"]+\?page=([^&]+)&amp;tab=repositories"(?:\s[^>]*)?>', r.text)):
				# End of pagination
				break
			r = get(f'https://github.com/{user}?page={m.group(1)}&tab=repositories')