#!/usr/bin/python

# Feed Normalizer
# http://www.myelin.co.nz/feed_normalizer/

# Copyright (C) 2003-04 Phillip Pearson

# Permission is hereby granted, free of charge, to any person
# obtaining a copy of this software and associated documentation files
# (the "Software"), to deal in the Software without restriction,
# including without limitation the rights to use, copy, modify, merge,
# publish, distribute, sublicense, and/or sell copies of the Software,
# and to permit persons to whom the Software is furnished to do so,
# subject to the following conditions:

# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.

# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
# EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
# NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
# BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
# ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
# CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
# SOFTWARE.

import feedparser, cgi, os, types, sys, httplib2, os.path, stat, time

def esc(s):
    # convert to utf-8, and escape
    if type(s) == types.UnicodeType:
        # if unicode, just convert
        s = s.encode('utf-8')
    else:
        # otherwise see if it's already utf-8
        try:
            s.decode('utf-8')
        except UnicodeError:
            # it's not - cross fingers and assume us-ascii
            s = s.decode('iso-8859-1').encode('utf-8')
    # s is utf-8 by now
    return cgi.escape(s, 1)

def prepare(x, keys):
    # turn some possibly dodgy data into escaped xml without leading or trailing space
    return tuple([esc(x.get(k, '').strip()) for k in keys])

def munge(s):
    r = ""
    for c in s:
        if c.isalnum() or c in "_(),.":
            r += c
        else:
            r += "%%%02d" % ord(c)
    return r

def norm(url, stripdesc, noisy=0):
    # do some normalization.  not too hard when someone else does the parser for you!

    # make sure we're not being asked to read a local file
    assert url.startswith('http://')

    # parse it
    for d in ("cache", "cache/httplib2", "cache/normalize"):
        if not os.path.exists(d):
            os.mkdir(d)
    cache_fn = os.path.join("cache", "normalize", munge(url))
    if not os.path.exists(cache_fn) or os.stat(cache_fn)[stat.ST_MTIME] < (time.time() - 3600):
        http = httplib2.Http("cache/httplib2")
        response, content = http.request(url)
        open(cache_fn, "wt").write(content)
    f = feedparser.parse(cache_fn)
    if noisy:
        print f

    # and print out some rss 2.0
    print """Content-Type: text/xml

<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0">
<channel>
 <title>%s</title>
 <link>%s</link>
 <description>%s</description>""" % prepare(f['channel'], ('title', 'link', 'description'))

    for item in f['items']:
        title, guid, link = prepare(item, ('title', 'guid', 'link'))
        # there's probably a better way, but for the moment, if the guid looks like a url,
        # assume it's a permalink.
        if guid.startswith('http://'):
            isperm = ""
        else:
            isperm = ' ispermalink="false"'
        print """ <item>
  <title>%s</title>
  <guid%s>%s</guid>
  <link>%s</link>""" % (title, isperm, guid, link)
        #from pprint import pprint
        #pprint(item)
        if stripdesc:
            # strip description - this is good for sharpreader as it lets you
            # see the linked article in the preview pane instead of the description
            # from the feed.
            content = ''
        elif item.has_key('content_encoded'):
            # prefer content_encoded over description, as some people put full
            # content in <content:encoded> and only a summary in <description>
            content = item['content_encoded']
        elif item.has_key('content'):
            # failing that, <content type="..."> is often good
            content = item['content'][0]['value']
        elif item.has_key('description'):
            # couldn't find anything - just use the description
            content = item['description']
        else:
            # no content??
            content = ''
        print """  <description>%s</description>""" % esc(content.strip())
        print """ </item>"""

    print """</channel>
</rss>"""

def usage():
    # we weren't called properly
    print """Location: %(url)s
Status: 301
Content-Type: text/html

Go to <a href="%(url)s">%(url)s</a> ...
""" % {'url': 'http://www.myelin.co.nz/feed_normalizer/',
       }

def cgi_main():
    # read cgi params, then do stuff
    fs = cgi.FieldStorage()
    if fs.getfirst('showsource'):
        print "Content-Type: text/plain\n"
        print open('normalize.py').read()
        return
    stripdesc = fs.getfirst('stripdesc') and 1 or 0
    url = fs.getfirst('url')
    if url is None:
        usage()
        return
    norm(url, stripdesc)

def main():
    # some tests
    import sys
    if len(sys.argv) > 1:
        norm(sys.argv[1])
        return
    for url in ('http://www.simplegeek.com/blogxbrowsing.asmx/GetRss?',
                'http://alex.halavais.net/news/index.xml',
                'http://boingboing.net/rss.xml',
                'http://www.simplegeek.com/blogxbrowsing.asmx/GetRss?',
                'http://www.myelin.co.nz/post/rss.xml',
                ):
        print "--- trying to normalise %s ---" % url
        norm(url, 0, noisy=1)

if __name__ == '__main__':
    # are we running under cgi?
    if os.environ.has_key('SERVER_NAME'):
        cgi_main()
    else:
        # nope - run tests instead
        main()