import glob, libxml2, os, time, sys, sgmllib, urllib
from xml.sax.saxutils import escape

now = time.time()
week = 7 * 86400
week_ago = now - week

cache = os.path.join(sys.argv[1], '*')

all_links = {}

for name in glob.glob(cache):
  # ensure that this is within the past week
  if os.path.isdir(name): continue
  mtime = os.stat(name).st_mtime
  if mtime < week_ago: continue

  # parse the file
  doc = libxml2.parseFile(name)
  xp = doc.xpathNewContext()
  xp.xpathRegisterNs("atom", "http://www.w3.org/2005/Atom")

  # determine the entry
  entry = xp.xpathEval("//atom:link[@rel='alternate']")
  if not entry: continue
  entry = entry[0].prop("href")

  # determine the feed
  feed = xp.xpathEval("//atom:source/atom:id")
  if feed:
    feed = feed[0].content
  else:
    feed = xp.xpathEval("//atom:source/atom:link[@href and @rel='self']")
    if not feed: feed = xp.xpathEval("//atom:source/atom:link[@href]")
    if not feed: continue
    feed = feed[0].prop('href')

  # identify the unique links
  entry_links = []
  for node in doc.xpathEval("//*[@href and not(@rel='source')]"):
      if node.parent.name == 'source': continue
      if not node.prop('href') in entry_links:
          entry_links.append(str(node.prop('href')))

  # add the votes
  vote = [(1.0 - (now - mtime)**2 / week**2, str(entry), str(feed))]
  for link in entry_links:
    all_links[link] = all_links.get(link,list()) + vote

  # free the entry
  doc.freeDoc()

# tally the votes
weighted_links = []
for link, votes in all_links.items():
  site = {}
  for weight, entry, feed in votes:
    site[feed] = min(site.get(feed,1), weight) 
  weighted_links.append((sum(site.values()), link))
weighted_links.sort()
weighted_links.reverse()

# determine the title for a given url
class html(sgmllib.SGMLParser):
  def __init__(self, url):
    sgmllib.SGMLParser.__init__(self)
    self.title = ""
    self.intitle = False
    try:
      self.feed(urllib.urlopen(url).read())
    except:
      pass
    if not self.title: self.title = url.split('/')[-1]
  def start_title(self, attributes):
    if not self.title: self.intitle = True
  def end_title(self):
    self.intitle = False
  def handle_data(self, text):
    if self.intitle: self.title += text

# output the results
print "<ul>"
for weight, link in weighted_links[:10]:
  print "<li><a href='%s'>%s</a></li>" % (escape(link), html(link).title)
print "</ul>"