#!/usr/bin/env python
# -*- coding: utf-8 -*-

import re, urllib, os.path, cPickle
import feedparser

################################################################################
#
# configure HISTORYPATH, HISTORYMAX and DSTXMLPATH as you like
#
HISTORYPATH = '/path/to/history.p'
HISTORYMAX = 50
DSTXMLPATH = '/path/to/voaspecialenglish.xml'
#
################################################################################

ORGXMLPATH = ('http://www.voanews.com/specialenglish/customCF/' 
              'RecentStoriesRSS.cfm?keyword=TopStories')
XML = '''<?xml version="1.0" encoding="UTF-8" ?>
<rss xmlns:itunes="http://www.itunes.com/dtds/podcast-1.0.dtd" version="2.0">
 <channel>
  <title>VOA News: Top Stories</title>
  <link>http://voaspecialenglish.com/</link>
  <pubDate>%(pubdate)s</pubDate>
  <language>en-us</language>
  <copyright>VOICE OF AMERICA</copyright>
  <itunes:subtitle>VOA News: Top Stories</itunes:subtitle>
  <itunes:author>VOA News: Top Stories</itunes:author>
  <itunes:summary>Up to the minute news from Voice of America</itunes:summary>
  <description>Up to the minute news from Voice of America</description>
  <itunes:category text="News &amp; Politics" />
  <itunes:owner>
   <itunes:name>VOA Podcasts -  Voice Of America - English - Special English</itunes:name>
   <itunes:email>voanews@voanews.com</itunes:email>
  </itunes:owner>
  <itunes:image href="http://www.voanews.com/voanews_shared/images/widgetBarIcon.png" />
%(items)s
 </channel>
</rss>
'''

ITEM = '''\
  <item>
   <title>%(title)s</title>
   <link>%(link)s</link>
   <pubDate>%(pubdate)s</pubDate>
   <enclosure url="%(mp3)s" length="%(mp3length)d" type="audio/mpeg" />
   <guid isPermaLink="false">%(link)s</guid>
   <itunes:author>VOICE OF AMERICA</itunes:author>
   <itunes:subtitle>%(title)s</itunes:subtitle>
   <itunes:summary>%(summary)s</itunes:summary>
   <itunes:explicit>no</itunes:explicit>
  </item> 
'''

##
def load_history(p):
    '''
    Load history list.  Each element is map.
    '''
    if os.path.exists(p):
        f = file(p)
        h = cPickle.load(f)
        f.close()
    else:
        h = []

    return h

##
def dump_history(h, p):
    '''
    Dump history list.  Adjust the length up to HISTORYMAX.
    '''
    h.sort(cmp = lambda x, y: cmp(y['pubdate_tuple'], x['pubdate_tuple']))
    while len(h) > HISTORYMAX:
        h.pop(-1)

    f = file(p, 'w')
    cPickle.dump(h, f)
    f.close()

##
def save_xml(xml, p):
    '''
    If there are any changes in XML, save it.
    '''
    if os.path.exists(p):
        f = file(p)
        if xml == f.read():
            resave = False
        else:
            resave = True
        f.close()
    else:
        resave = True

    if resave:
        f = file(p, 'w')
        f.write(xml)
        f.close()

##
def semiplain(s):
    '''
    Remove HTML tags and entity reference which are not allowed in 
    Podcast feed.
    '''
    tagpat = re.compile('<[a-z/].*?>', re.DOTALL|re.IGNORECASE)
    entpat = re.compile('&.*?;')
    spcpat = re.compile('\s\s+')

    s = tagpat.sub('', s)
    for ent in set(entpat.findall(s)):
        if ent not in ('&amp;', '&lt;', '&gt;', '&apos;', '&quot;'):
            s = s.replace(ent, '')
    s = spcpat.sub(' ', s)
    s = s.strip()

    if len(s) < 4000:
        return s[:4000]
    else:
        return s[:3996] + ' ...'

##
def get_from_article(link):
    '''
    Extract body of article, URL of MP3 and file size of MP3
    '''
    bodypat = re.compile('<span class="body">(.*?)</span>\s', re.DOTALL)
    mp3pat = re.compile('<a href="(http://www.voanews.com/mediaassets/.*?)"\s')

    f = urllib.urlopen(link)
    p = f.read()
    f.close()

    searched = bodypat.search(p)
    if searched:
        body = searched.group(1)
        summary = semiplain(body)
    else:
        summary = ''

    searched = mp3pat.search(p)
    if searched:
        mp3 = searched.group(1)

        f = urllib.urlopen(mp3)
        mp3length = int(f.info()['Content-Length'])
        f.close()
    else:
        mp3 = ''
        mp3length = 0

    return summary, mp3.replace(' ', '%20'), mp3length

##
def remake_item(history, e):
    '''
    Remake RSS item of VOA News to Podcast item.
    If the entry was processed before, use it.
    '''
    title = xml.sax.saxutils.escape(e.title.encode('utf-8'))
    for h in history:
        if h['title'] == title:
            m = h
            break
    else:
        m = {}
        m['title'] = xml.sax.saxutils.escape(e.title.encode('utf-8'))
        m['link'] = e.link.encode('utf-8')
        m['pubdate'] =  e.updated.encode('utf-8')
        m['pubdate_tuple'] =  e.updated_parsed
        m['summary'], m['mp3'], m['mp3length'] = get_from_article(m['link'])

        history.append(m)

    link, summary, pubdate = m['link'], m['summary'], m['pubdate']
    mp3, mp3length = m['mp3'], m['mp3length']

    return history, ITEM % locals()

##
def get_xml():
    '''
    Main process to make Podcast XML.
    '''
    history = load_history(HISTORYPATH)
    items = []
    d = feedparser.parse(ORGXMLPATH)
    for e in d.entries:
        history, item = remake_item(history, e)
        items.append(item)

    m = {'pubdate': d.entries[0].updated.encode('utf-8'),
         'items': ''.join(items)}

    dump_history(history, HISTORYPATH)

    return XML % m

##
def index(req):
    '''
    This function is for mod_python.
    If Podcast client access to
      http://your.domain/voaspecialenglishpodcast.py
    XML will be returned.
    '''
    req.content_type = 'text/xml'

    return get_xml()

##
if __name__ == '__main__':
    xml = get_xml()
    save_xml(xml, DSTXMLPATH)