[youtube]: add YoutubeShowIE (closes #14)

It just extracts the playlists urls for each season
pull/932/merge
Jaime Marquínez Ferrándiz 2013-07-01 17:59:28 +02:00
parent d828f3a550
commit 75dff0eef7
3 changed files with 22 additions and 2 deletions

View File

@ -8,7 +8,7 @@ import json
import os
sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
from youtube_dl.extractor import YoutubeUserIE, YoutubePlaylistIE, YoutubeIE, YoutubeChannelIE
from youtube_dl.extractor import YoutubeUserIE, YoutubePlaylistIE, YoutubeIE, YoutubeChannelIE, YoutubeShowIE
from youtube_dl.utils import *
from helper import FakeYDL
@ -88,5 +88,11 @@ class TestYoutubeLists(unittest.TestCase):
result = ie.extract('PLtPgu7CB4gbY9oDN3drwC3cMbJggS7dKl')[0]
self.assertEqual(len(result['entries']), 2)
def test_youtube_show(self):
dl = FakeYDL()
ie = YoutubeShowIE(dl)
result = ie.extract('http://www.youtube.com/show/airdisasters')
self.assertTrue(len(result) >= 4)
if __name__ == '__main__':
unittest.main()

View File

@ -67,7 +67,7 @@ from .yahoo import YahooIE, YahooSearchIE
from .youjizz import YouJizzIE
from .youku import YoukuIE
from .youporn import YouPornIE
from .youtube import YoutubeIE, YoutubePlaylistIE, YoutubeSearchIE, YoutubeUserIE, YoutubeChannelIE
from .youtube import YoutubeIE, YoutubePlaylistIE, YoutubeSearchIE, YoutubeUserIE, YoutubeChannelIE, YoutubeShowIE
from .zdf import ZDFIE

View File

@ -853,3 +853,17 @@ class YoutubeSearchIE(SearchInfoExtractor):
video_ids = video_ids[:n]
videos = [self.url_result('http://www.youtube.com/watch?v=%s' % id, 'Youtube') for id in video_ids]
return self.playlist_result(videos, query)
class YoutubeShowIE(InfoExtractor):
_VALID_URL = r'https?://www\.youtube\.com/show/(.*)'
IE_NAME = u'youtube:show'
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url)
show_name = mobj.group(1)
webpage = self._download_webpage(url, show_name, u'Downloading show webpage')
# There's one playlist for each season of the show
m_seasons = list(re.finditer(r'href="(/playlist\?list=.*?)"', webpage))
self.to_screen(u'%s: Found %s seasons' % (show_name, len(m_seasons)))
return [self.url_result('https://www.youtube.com' + season.group(1), 'YoutubePlaylist') for season in m_seasons]