sync gists to local hg repos

syncgiststohg.py

#!/usr/bin/env python
import os
import subprocess
import sys
import urlparse

from lxml import html

BASE_URL = 'http://gist.github.com/%s'

def get_gists_for_user(username):
    url = BASE_URL % username
    doc = html.parse(url)
    doc = doc.getroot()
    doc.make_links_absolute(url)
    gists = doc.cssselect('#files .file .meta .info span a')
    for gist in gists:
        yield gist.attrib['href']

def hg_pull(repodir, remoteurl):
    p = subprocess.Popen(['hg', '-R', repodir, 'pull', '-u', remoteurl],
        stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
    stdout, stderr = p.communicate()
    if p.wait() != 0:
        print >> sys.stderr, 'error updating %s from %s' % (repodir, remoteurl)
        print >> sys.stderr, stdout
        sys.exit(1)

def hg_clone(targetdir, remoteurl):
    p = subprocess.Popen(['hg', 'clone', remoteurl, targetdir],
        stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
    stdout, stderr = p.communicate()
    if p.wait() != 0:
        print >> sys.stderr, 'error cloning %s to %s' % (remoteurl, targetdir)
        print >> sys.stderr, stdout
        sys.exit(1)

def sync_gist_to_dir(targetdir, gisturl):
    gistnum = urlparse.urlparse(gisturl).path.lstrip('/')
    gistdir = os.path.join(targetdir, gistnum)
    doc = html.parse(gisturl)
    doc = doc.getroot()
    giturl = doc.xpath('//a[@rel="#git-clone"]/@href')[0]

    hg_pull(gistdir, giturl) if os.path.exists(gistdir)\
        else hg_clone(gistdir, giturl)

if __name__ == '__main__':
    try:
        username = sys.argv[1]
        targetdir = sys.argv[2]
    except IndexError, e:
        print >> sys.stderr, 'Usage: %s username repo_target_dir' % sys.argv[0]
        sys.exit(1)

    targetdir = os.path.abspath(targetdir)
    for gist in get_gists_for_user(username):
        sync_gist_to_dir(targetdir, gist)