A github RSS feed sanitizer

parser.py

#!/usr/bin/env python
# encoding: utf-8
"""
parser.py

A github RSS feed sanitizer, whereby we mean "sane" - i.e., that groups updates
per project and not as a mess of completely unrelated events.

Created by Rui Carmo on 2011-01-06. Still incomplete.

Distributed under the MIT license
"""

import os, re, sys, feedparser, htmlentitydefs
from xml.etree.cElementTree import XMLParser, tostring
try:
  import json
except:
  import simplejson as json

patterns = [re.compile(x) for x in ['.+(the) (\w+)\/(\w+) wiki$', '.+(on|at) (\w+)\/(\w+)$']]
projects = {}

# Use inline styles to ensure this has the proper look in RSS

css = {
  'thumbnail': "width:30px; height:30px; background-color: white; border: 1px solid #D0D0D0; float: left; margin-right: 0.7em; padding: 2px; background: url(%(thumbnail)s) no-repeat;"
}

template = u"""<entry><div><div class="thumbnail"> </div>%(title)s</div>%(description)s</entry>"""

# method for injection in feedparser
def _start_thumbnail(self, attrsD):
  context = self._getContext()
  context.setdefault('thumbnail', attrsD)

def collate(entry):
  for p in patterns:
    m = p.match(entry.title)
    if m:
      (dummy, owner, project) = m.groups()
      break
    project = 'General'
  try:
    p = projects[project]
  except KeyError:
    p = projects[project] = []
  buffer = template % {'title': entry.title, 'description': entry.content[0]['value']}
  tree = XMLParser()
  for i in htmlentitydefs.name2codepoint.keys():
    buffer = buffer.replace("&%s;" % i,"&#%d;" % htmlentitydefs.name2codepoint[i])
  tree.feed(buffer.encode('utf-8'))
  t = tree.close()
  for e in t.findall('entry'):
    try:
      e.attrib['style'] = css[e.attrib['class']] % locals()
    except:
      pass 
  p.append(tostring(t))
  

def main():
  #feedparser._debug=1  
  feedparser._FeedParserMixin._start_thumbnail = _start_thumbnail
  d = feedparser.parse('https://github.com/rcarmo.private.atom?token=foobar', agent = 'Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; .NET CLR 1.1.4322)') # be annoying
  for i in d.entries:
    collate(i)
  for p in projects.keys():
    print "<h2>%s</h2>" % p
    for e in projects[p]:
      print e


if __name__ == '__main__':
  main()