nojhan/weboob-devel
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

on mosaic, parse preview_url, duration and rating

Browse source
This commit is contained in:
Romain Bignon 2010-04-17 11:21:03 +02:00
commit baa2fd937a
1 changed files with 42 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

50
weboob/backends/youporn/pages/index.py
View file

@ -23,13 +23,45 @@ from weboob.capabilities.video import Video
class IndexPage(PornPage):
def iter_videos(self):
for h1 in self.document.getiterator('h1'):
a = h1.find('a')
if a is None:
continue
uls = self.document.getroot().cssselect("ul[class=clearfix]")
if not uls:
return
url = a.attrib['href']
_id = url[len('/watch/'):]
_id = _id[:_id.find('/')]
title = a.text
yield Video(int(_id), title)
for ul in uls:
for li in ul.findall('li'):
a = li.find('a')
if a is None or a.find('img') is None:
continue
preview_url = a.find('img').attrib['src']
h1 = li.find('h1')
a = h1.find('a')
if a is None:
continue
url = a.attrib['href']
_id = url[len('/watch/'):]
_id = _id[:_id.find('/')]
title = a.text
duration = 0
div = li.cssselect('div[class=duration_views]')
if div:
h2 = div[0].find('h2')
duration = 60 * int(h2.text)
duration += int(h2.find('span').tail)
rating = 0
rating_max = 0
div = li.cssselect('div[class=rating]')
if div:
p = div[0].find('p')
rating = float(p.text.strip())
rating_max = float(p.find('span').text.strip()[2:])
yield Video(int(_id),
title,
rating=rating,
rating_max=rating_max,
duration=duration,
preview_url=preview_url)