use Regexp filter
This commit is contained in:
parent
06cee506f9
commit
b830781128
1 changed files with 2 additions and 8 deletions
|
|
@ -18,11 +18,9 @@
|
||||||
# along with weboob. If not, see <http://www.gnu.org/licenses/>.
|
# along with weboob. If not, see <http://www.gnu.org/licenses/>.
|
||||||
|
|
||||||
|
|
||||||
import re
|
|
||||||
|
|
||||||
from weboob.tools.browser2 import HTMLPage
|
from weboob.tools.browser2 import HTMLPage
|
||||||
from weboob.tools.browser2.page import ListElement, method, ItemElement
|
from weboob.tools.browser2.page import ListElement, method, ItemElement
|
||||||
from weboob.tools.browser2.filters import Filter, Link, CleanText, Duration
|
from weboob.tools.browser2.filters import Link, CleanText, Duration, Regexp
|
||||||
from weboob.capabilities.base import NotAvailable
|
from weboob.capabilities.base import NotAvailable
|
||||||
from weboob.capabilities.image import BaseImage
|
from weboob.capabilities.image import BaseImage
|
||||||
from weboob.capabilities.video import BaseVideo
|
from weboob.capabilities.video import BaseVideo
|
||||||
|
|
@ -41,11 +39,7 @@ class IndexPage(HTMLPage):
|
||||||
class item(ItemElement):
|
class item(ItemElement):
|
||||||
klass = BaseVideo
|
klass = BaseVideo
|
||||||
|
|
||||||
class Id(Filter):
|
obj_id = Regexp(Link('.//a'), r'/videos/(.+)\.html')
|
||||||
def filter(self, link):
|
|
||||||
return re.sub(r'/videos/(.+)\.html', r'\1', link)
|
|
||||||
|
|
||||||
obj_id = Id(Link('.//a'))
|
|
||||||
obj_title = CleanText('.//span[@id="title1"]')
|
obj_title = CleanText('.//span[@id="title1"]')
|
||||||
obj_duration = Duration(CleanText('.//span[@class="thumbtime"]//span'), default=NotAvailable)
|
obj_duration = Duration(CleanText('.//span[@class="thumbtime"]//span'), default=NotAvailable)
|
||||||
obj_nsfw = True
|
obj_nsfw = True
|
||||||
|
|
|
||||||
Loading…
Add table
Add a link
Reference in a new issue