Backup GitHub with dumb scripts.

build_clone_script.py

"A dumb script builder to clone scraped GitHub repo listings."

# 1. Scrape <ul id="repo_listing" class="repo_list"> into listing.html
# 2. python build_clone_script.py | sort -u > clone.sh
# 3. time sh clone.sh

from __future__ import print_function

import lxml.html


data = open('listing.html', 'r')
dom = lxml.html.fromstring(data.read())

for element in dom.cssselect('.source'):
    owner = element.cssselect('.owner')[0].text.strip()
    repo = element.cssselect('.repo')[0].text.strip()
    cmd = 'git clone --mirror git@github.com:{0}/{1}.git {0}-{1}.git'\
        .format(owner, repo)
    print(cmd)

update.bash

#!/bin/bash
# Updates all mirrored .git directories in the current directory.

for repo in *.git; do
    pushd $repo
    git fetch
    popd > /dev/null
done