#!/usr/bin/env python
# Take Chuck Taggart's "Looka" blog page, and produce RSS from it.
# Adam Sampson <ats@offog.org>

import sys, os, time, re, htmlentitydefs

# {{{ feedwriter
# This section's a slightly cut-down version of feedwriter.py:
# http://offog.org/darcs/misccode/feedwriter.py

class FeedError(Exception): pass

def escape_xml(s):
	"""Escape special characters in XML."""
	chars = { "<" : "lt",
	          ">" : "gt",
	          "&" : "amp",
	          '"' : "quot",
	          "'" : "apos" }
	cs = []
	for c in s:
		if chars.has_key(c):
			cs.append("&" + chars[c] + ";")
		else:
			cs.append(c)
	return "".join(cs)

def wrap(name, value):
	if value is None:
		return ""
	else:
		return "<" + name + ">" + escape_xml(value).encode("UTF-8") + "</" + name + ">\n"

class Channel:
	"""A summary of a syndicated site."""

	def __init__(self, title, link, description):
		self.title = title
		self.link = link
		self.description = description
		self.items = []

	def add_item(self, *args, **keywords):
		"""Add an item to the feed. The arguments to this method
		   are the same as for the Item constructor."""
		self.items.append(Item(*args, **keywords))

	def rss2(self):
		"""Return the RSS 2.0 representation of this feed."""
		bits = []
		bits.append('<?xml version="1.0" encoding="UTF-8"?>\n')
		bits.append('<rss version="2.0">\n')
		bits.append('<channel>\n')
		bits.append(wrap("title", self.title))
		bits.append(wrap("link", self.link))
		bits.append(wrap("description", self.description))
		for item in self.items:
			bits.append('<item>\n')
			bits.append(wrap("title", item.title))
			bits.append(wrap("link", item.link))
			bits.append(wrap("description", item.description))
			if item.pubDate is not None:
				d = item.pubDate
				if type(d) is not tuple and type(d) is not time.struct_time:
					d = time.gmtime(d)
				s = time.strftime("%a, %d %b %Y %H:%M:%S +0000",
				                  d)
				bits.append(wrap("pubDate", s))
			bits.append('</item>\n')
		bits.append('</channel>\n</rss>\n')
		return "".join(bits)

class Item:
	"""An item in a syndication feed."""

	def __init__(self, title = None, link = None, description = None, pubDate = None):
		if title is None and description is None:
			raise FeedError("Item must have either title or description")
		self.title = title
		self.link = link
		self.description = description
		self.pubDate = pubDate

# }}}

def strip_html(s):
	"""Remove HTML tags and entities from a string."""
	s = re.sub(r'<[^>]*>', '', s)
	def decode(m):
		return htmlentitydefs.entitydefs.get(m.group(1), "").decode("ISO-8859-1")
	s = re.sub(r'&([a-z]*);', decode, s)
	return s.strip()

def convert(html_fn, rss_fn):
	"""Convert Looka's HTML to RSS."""

	# Read in the HTML
	f = open(html_fn)
	html = f.read().decode("ISO-8859-1")
	f.close()

	current_date = None

	main_url = "http://www.gumbopages.com/looka/"
	channel = Channel(title = "Looka!", link = main_url, description = "All that and a Creole's Stuffed Bread.")

	# Find all the articles, which are bracketed by <span class="RSS:item">...</span>,
	# and all the date markers which look like <a name="number">...</a>.
	for (article, daynumber, date) in re.findall(r'(?ms)(?:<span class="RSS:item">(.*?)</span>|<a name="(\d+)">(.*?)</a>)', html):
		if article != "":
			# It's an article.
			# The title is the first line after the <span>; the rest is the article proper.
			m = re.match(r'(?ms)([^\n]*)\n(.*)$', article)
			if m is None:
				print "Can't find title in article that starts " + repr(article[:80]) + "; ignoring it"
				continue
			title = strip_html(m.group(1))
			body = m.group(2)

			# If we managed to parse a date header before this article, use it.
			if current_date is not None:
				(date, link) = current_date
			else:
				date, link = None, main_url

			# If there's a Javascript link for comments, replace it
			# with a link back to the original article.
			body = re.sub(r'(?ms)<a href="javascript:HaloScan.*?</a>', '<a href="' + escape_xml(link) + '">Comments</a>', body)

			# Add the item to the channel.
			channel.add_item(title = title, link = link, description = body, pubDate = date)

		else:
			# It's a date marker.
			# The date's in the format "Tuesday, April 3, 2007". Try to parse it.
			s = re.sub(r'(,|\s+)', ' ', strip_html(date))
			try:
				parsed = time.strptime(s, "%A %B %d %Y")
			except:
				print "Can't parse date " + repr(s)
				current_date = None
				continue

			link = main_url + "#" + daynumber
			current_date = (parsed, link)

	# Write the RSS to a temporary file...
	f = open(rss_fn + ".new", "w")
	f.write(channel.rss2())
	f.close()

	# ... and rename that temporary file over the output file.
	os.rename(rss_fn + ".new", rss_fn)

if __name__ == "__main__":
	args = sys.argv[1:]
	if len(args) != 2:
		print "Usage: lookafeed.py input.html output.rss"
		sys.exit(1)
	else:
		convert(args[0], args[1])