# Author: Jonathan Ellis

import datetime, re, time, logging, traceback
from google.appengine.api.urlfetch import fetch
from google.appengine.api import memcache
import PyRSS2Gen
from BeautifulSoup import BeautifulSoup


logging.getLogger().setLevel(logging.DEBUG)
url = 'http://www.python.org/community/jobs/'


def items():
    response = fetch(url)
    soup = BeautifulSoup(response.content)

    L = soup.fetch('div', {'class': 'section'})
    for div in L:
        if not div.fetch('h3', recursive=False):
            continue
        title = 'no title'
        href = url
        for a in div.fetch('a'):
            try:
                title = a.renderContents()
                href = a['href']
            except:
                pass
            else:
                break
        logging.debug('title is ' + title)

        # date handling... bleh
        try:
            when_raw = div.first('em').renderContents()
        except:
            # crap html...
            m = re.search('''<em>(.*?)<''', div.renderContents())
            if m:
                when_raw = m.group(1)
            else:
                when_raw = None # some don't even give a date
        try:
            # could throw either b/c when_raw is None, or no date was found
            m = re.search('''(\d+)-(\w\w\w)-(\d+)''', when_raw)
            (d, m, y) = m.groups()
            t = time.strptime('%s %s %s' % (d, m.capitalize(), y), '%d %b %Y')
            # generates GMT, but hell, maybe that's what pyjobs means :P
            when_t = time.localtime(time.mktime(t))
        except:
            # "right now" should be close enough for most purposes
            when_t = time.localtime(0)
        when_d = datetime.datetime(*when_t[:-3])

        desc = '\n'.join([str(p) for p in div.fetch('p')[1:]])
        item = PyRSS2Gen.RSSItem(
            title = title,
            link = href,
            description = desc,
            guid = PyRSS2Gen.Guid(href),
            pubDate = when_d)
        yield item

def xml():
    xml = memcache.get('xml')
    if not xml:
        rss = PyRSS2Gen.RSS2(
            title = 'Unofficial Python Job Board Feed',
            link = url,
            description = """
                This RSS feed is autogenerated from the handwritten html that
                makes up the python.org jobs page.  This page is... irregular in
                places.  Notably, I don't always get all or even most of the
                description, but at least you'll know when a new job goes up.
                """,
            lastBuildDate = datetime.datetime.now(),
            items = items())
        xml = rss.to_xml()
        memcache.add('xml', xml, 3600)
    return xml

def main():
    print "Content-Type: text/xml\n\n" + xml()


if __name__ == '__main__':
    main()