w/youtube-dl
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

[thescene] Fix extraction and improve style (Closes #8978)

Browse source
This commit is contained in:
Sergey M․ 2016-03-25 20:14:34 +06:00
parent 4db441de72
commit 2156f16ca7
1 changed files with 11 additions and 7 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

18
youtube_dl/extractor/thescene.py
View file

@ -2,7 +2,7 @@ from __future__ import unicode_literals
from .common import InfoExtractor from .common import InfoExtractor
from ..compat import compat_urllib_parse from ..compat import compat_urlparse
from ..utils import qualities from ..utils import qualities
@ -21,17 +21,21 @@ class TheSceneIE(InfoExtractor):
def _real_extract(self, url): def _real_extract(self, url):
display_id = self._match_id(url) display_id = self._match_id(url)
webpage = self._download_webpage(url, display_id) webpage = self._download_webpage(url, display_id)
player_url = compat_urllib_parse.urljoin(
player_url = compat_urlparse.urljoin(
url, url,
self._html_search_regex( self._html_search_regex(
r'id=\'js-player-script\'[^>]+src=\'(.+?)\'', webpage, 'player url')) r'id=\'js-player-script\'[^>]+src=\'(.+?)\'', webpage, 'player url'))
self.to_screen(player_url) player = self._download_webpage(player_url, display_id)
player = self._download_webpage(player_url, player_url) info = self._parse_json(
info = self._parse_json(self._search_regex(r'(?m)var\s+video\s+=\s+({.+?});$', player, 'info json'), display_id) self._search_regex(
r'(?m)var\s+video\s+=\s+({.+?});$', player, 'info json'),
display_id)
qualities_order = qualities(['low', 'high']) qualities_order = qualities(('low', 'high'))
formats = [{ formats = [{
'format_id': '{0}-{1}'.format(f['type'].split('/')[0], f['quality']), 'format_id': '{0}-{1}'.format(f['type'].split('/')[0], f['quality']),
'url': f['src'], 'url': f['src'],
@ -41,8 +45,8 @@ class TheSceneIE(InfoExtractor):
return { return {
'id': info['id'], 'id': info['id'],
'display_id': display_id,
'title': info['title'], 'title': info['title'],
'formats': formats, 'formats': formats,
'thumbnail': info.get('poster_frame'), 'thumbnail': info.get('poster_frame'),
'display_id': display_id,
} }