.*? class="format"\s+title="([^"]+)"
          .*? /by_format/([^"]+)
      """, html, re.X|re.S)
      
      # Assemble
      for homepage, title, listeners, playing, tags, url, bits, fmt in ls:
          r.append(dict(
              genre = clean(tags),
              title = clean(title),
              genre = unhtml(tags),
              title = unhtml(title),
              homepage = ahttp.fix_url(homepage),
              playing = clean(playing),
              playing = unhtml(playing),
              url = "http://dir.xiph.org{}".format(url),
              listformat = "xspf",
              listeners = int(listeners),
              bitrate = bitrate(bits),
              format = self.mime_fmt(guess_format(fmt)),
          ))
      return r

        return 0


# Extract mime type from text
rx_fmt = re.compile("ogg|mp3|mp4|theora|nsv|webm|opus|mpeg")
def guess_format(str):
    return rx_fmt.findall(str.lower() + "mpeg")[0]

# Clean up HTML text snippets
def clean(str):
    return nl(entity_decode(strip_tags(str)))