#!/usr/bin/env python
"""
persian.py - Phenny Persian Module
Copyright 2012, Sean B. Palmer, inamidst.com
Licensed under the Eiffel Forum License 2.

http://inamidst.com/phenny/
"""

import re
import web

link = "http://dsal.uchicago.edu/cgi-bin/philologic/search3dsal"
query = "dbname=steingass&query=%s&matchtype=exact&display=utf8"

r_paragraph = re.compile("(?is)<p>(.*?)(?=</?p>)")
r_tag = re.compile("(?i)<[^>]+>")
r_whitespace = re.compile(r"(?s)[ \t\r\n]+")

def parse(definition):
   definition = definition.rsplit("</pa>", 1).pop()
   definition = definition.rsplit("</i>", 1).pop()
   definition = definition.strip()
   definition = definition.replace("-\n", "")
   definition = r_tag.sub("", definition)
   definition = r_whitespace.sub(" ", definition)
   return definition

def lookup(search, limit=3):
   assert isinstance(search, unicode)
   search = web.urllib.quote(search.encode("utf-8"))
   bytes = web.get(link + "?" + (query % search))

   definitions = set()
   for paragraph in r_paragraph.findall(bytes):
      if not ("<i>" in paragraph): continue

      definition = parse(paragraph)
      if not definition: continue
      if definition.endswith("q.v."): continue

      if not (definition in definitions):
         definitions.add(definition)
         yield definition

      if len(definitions) >= limit: break

def persian(phenny, input):
   search = input.group(2) # unicode text

   results = list(lookup(search))
   if not results:
      return phenny.reply("Sorry, no results found.")

   result = " / ".join(results)
   if len(result) > 256:
      result = result[:251] + "[...]"
   phenny.reply(result)
persian.commands = ["persian"]

if __name__ == "__main__":
   print __doc__