better-github-rss

.gitignore

/bin
/include
/lib

README.rst

better-github-rss

GitHub's user activity RSS feed is plain annoying. Too little context to really grasp what the tickets, comments and commits are about.

This script is going to fix that. Stay tuned.

ghrss.py

import logging
import lxml.etree
import lxml.html
import sys
import urllib2


logging.basicConfig(level=logging.DEBUG, format=logging.BASIC_FORMAT)
logger = logging.getLogger('ghrss')


def modify_rss(url):
    """Process an XML document of an RSS feed

    Returns the processed XML document as a pretty-printed XML string

    Removes entries whose links are duplicates, i.e. point to the same webpage.

    Replaces contents of GitHub issue entries with the actual webpage of the
    issue.
    """
    if url.startswith(('http://', 'https://')):
        feed = urllib2.urlopen(url)
    else:
        feed = open(url)
    root = lxml.etree.parse(feed).getroot()
    atom_ns = root.nsmap[None]

    def tag(tagname):
        return '{{{0}}}{1}'.format(atom_ns, tagname)

    pages = set()
    for entry in root.findall(tag('entry')):
        link = entry.find(tag('link')).attrib['href']
        page = link.split('#')[0]
        if page in pages:
            logger.info('remove %s', link)
            root.remove(entry)
        elif page.split('/')[-2] == 'issues':
            logger.info('fix %s', link)
            content = entry.find(tag('content'))
            content.text = modify_ticket(page)
        pages.add(page)

    return lxml.etree.tostring(root, pretty_print=True)


def modify_ticket(url):
    page = urllib2.urlopen(url)
    dom = lxml.html.parse(page, parser=lxml.html.HTMLParser())
    root = dom.getroot()
    comments = root.find('.//div[@class="new-comments"]')
    style = '<link href="github.css" media="screen" rel="stylesheet" type="text/css"/>'
    comments = lxml.etree.tostring(comments, pretty_print=True)
    return style + comments


print modify_rss(sys.argv[1])