sync.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-

"""Kenneth Reitz's GitHub Syncer

This script uses the GitHub API to get a list of all forked, mirrored, public, and 
private repos in your GitHub account. If the repo already exists locally, it will 
update it via git-pull. Otherwise, it will properly clone the repo.

It will organize your repos into the following directory structure:

+ repos
  ├── forks    (public fork repos)
  ├── mirrors  (public mirror repos)
  ├── private  (private repos)
  ├── public   (public repos)
  ├── watched  (public watched repos)
  └── sync.py  (this script)

Requires Ask Solem's github2 (http://pypi.python.org/pypi/github2).

Inspired by Gisty (http://github.com/swdyh/gisty). 
"""

import errno
import os
from commands import getoutput as cmd

from github2.client import Github
try:
    from gistapi import Gists
except ImportError:
    gist = False
else:
    gist = True


def makedir(directory):
    try:
        os.makedirs(directory)
    except OSError, exc:
        if exc.errno != errno.EEXIST:
            raise


def have(executable):
    for path in os.environ['PATH'].split(':'):
        if os.path.exists(os.path.join(path, executable)):
            return True
    return False

def clone(url, dest):
    if have('git-cached-clone'):
        cmd = 'git cached-clone'
    else:
        cmd = 'git clone'
    cmd += ' %s %s' % (url, dest)
    os.system(cmd)

# GitHub configurations
GITHUB_USER = cmd('git config github.user')
GITHUB_TOKEN = cmd('git config github.token')

# API Object
github = Github(username=GITHUB_USER, api_token=GITHUB_TOKEN)


# repo slots
repos = {}
repos['watched'] = [r for r in github.repos.watching(GITHUB_USER)]
repos['private'] = []
repos['mirrors'] = []
repos['public'] = []
repos['forks'] = []

# Collect GitHub repos via API
for repo in github.repos.list():
    if repo.private:
        repos['private'].append(repo)
    elif repo.fork:
        repos['forks'].append(repo)
    elif 'mirror' in repo.description.lower():
        # mirrors owned by self if mirror in description...
        repos['mirrors'].append(repo)
    else:
        repos['public'].append(repo)

topdir = os.getcwd()
for org, repos in repos.iteritems():
    for repo in repos:
        
        makedir(org)
        if org == 'watched':
            localdir = os.path.join(repo.owner, repo.name)
            if repo.owner == GITHUB_USER:
                # Don't clone any of my repositories into the watched area
                continue
        else:
            localdir = repo.name
            
        # I own the repo
        private = org in ('private', 'fork', 'mirror')

        # enter org dir
        os.chdir(org)
        
        # just `git pull` if it's already there
        if os.path.exists(localdir):
            os.chdir(localdir)
            print('Updating repo: %s' % (localdir))
            os.system('git pull')
        else:
            print('Cloning repo: %s/%s' % (org, localdir))
            clone('git://github.com/%s/%s.git' % (repo.owner, repo.name),
                  localdir)
        
        # return to base
        os.chdir(topdir)
        print

if gist:
    makedir('gist')
    for gist in Gists.fetch_by_user(GITHUB_USER):
        destdir = os.path.join('gist', gist.repo)
        if os.path.exists(destdir):
            os.system('cd gist/%s && git pull' % gist.repo)
        else:
            clone('git://gist.github.com/%s.git' % gist.repo,
                  'gist/%s' % gist.repo)