etc/yt-dlp
0
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-11-16 01:25:06 +00:00
Code Issues Projects Releases Wiki Activity

[ted] Use unicode_literals

Browse source
This commit is contained in:
Philipp Hagemeister 2014-01-17 03:52:17 +01:00
parent 8307aa73fb
commit f853f8594d
1 changed files with 10 additions and 7 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

17
youtube_dl/extractor/ted.py
View file

@ -1,3 +1,5 @@
from __future__ import unicode_literals
import json import json
import re import re
@ -7,6 +9,7 @@
RegexNotFoundError, RegexNotFoundError,
) )
class TEDIE(SubtitlesInfoExtractor): class TEDIE(SubtitlesInfoExtractor):
_VALID_URL=r'''http://www\.ted\.com/ _VALID_URL=r'''http://www\.ted\.com/
( (
@ -18,12 +21,12 @@ class TEDIE(SubtitlesInfoExtractor):
/(?P<name>\w+) # Here goes the name and then ".html" /(?P<name>\w+) # Here goes the name and then ".html"
''' '''
_TEST = { _TEST = {
u'url': u'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html', 'url': 'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html',
u'file': u'102.mp4', 'file': '102.mp4',
u'md5': u'2d76ee1576672e0bd8f187513267adf6', 'md5': '2d76ee1576672e0bd8f187513267adf6',
u'info_dict': { 'info_dict': {
u"description": u"md5:c6fa72e6eedbd938c9caf6b2702f5922", "description": "md5:c6fa72e6eedbd938c9caf6b2702f5922",
u"title": u"Dan Dennett: The illusion of consciousness" "title": "Dan Dennett: The illusion of consciousness"
} }
} }
@ -47,7 +50,7 @@ def _playlist_videos_info(self, url, name, playlist_id):
'''Returns the videos of the playlist''' '''Returns the videos of the playlist'''
webpage = self._download_webpage( webpage = self._download_webpage(
url, playlist_id, u'Downloading playlist webpage') url, playlist_id, 'Downloading playlist webpage')
matches = re.finditer( matches = re.finditer(
r'<p\s+class="talk-title[^"]*"><a\s+href="(?P<talk_url>/talks/[^"]+\.html)">[^<]*</a></p>', r'<p\s+class="talk-title[^"]*"><a\s+href="(?P<talk_url>/talks/[^"]+\.html)">[^<]*</a></p>',
webpage) webpage)