aboutsummaryrefslogtreecommitdiff
path: root/youtube_dl/extractor
diff options
context:
space:
mode:
authorYen Chi Hsuan <yan12125@gmail.com>2016-09-12 02:55:15 +0800
committerYen Chi Hsuan <yan12125@gmail.com>2016-09-12 02:55:15 +0800
commitbe457302267b456412fb9848bcb8ce36874d8d7e (patch)
tree301abaec5bfd1306dcc489c23243c45863d76676 /youtube_dl/extractor
parentee7e672eb0eca7a916845b359511262935f9ef1b (diff)
[nbc] Add new extractor for NBC Olympics (#10295, #10361)
Diffstat (limited to 'youtube_dl/extractor')
-rw-r--r--youtube_dl/extractor/extractors.py1
-rw-r--r--youtube_dl/extractor/nbc.py40
2 files changed, 41 insertions, 0 deletions
diff --git a/youtube_dl/extractor/extractors.py b/youtube_dl/extractor/extractors.py
index a3cd9c289..522691de1 100644
--- a/youtube_dl/extractor/extractors.py
+++ b/youtube_dl/extractor/extractors.py
@@ -534,6 +534,7 @@ from .nbc import (
CSNNEIE,
NBCIE,
NBCNewsIE,
+ NBCOlympicsIE,
NBCSportsIE,
NBCSportsVPlayerIE,
)
diff --git a/youtube_dl/extractor/nbc.py b/youtube_dl/extractor/nbc.py
index f694e210b..f37bf2f30 100644
--- a/youtube_dl/extractor/nbc.py
+++ b/youtube_dl/extractor/nbc.py
@@ -335,3 +335,43 @@ class NBCNewsIE(ThePlatformIE):
'url': 'http://feed.theplatform.com/f/2E2eJC/nnd_NBCNews?byId=%s' % video_id,
'ie_key': 'ThePlatformFeed',
}
+
+
+class NBCOlympicsIE(InfoExtractor):
+ _VALID_URL = r'https?://www\.nbcolympics\.com/video/(?P<id>[a-z-]+)'
+
+ _TEST = {
+ # Geo-restricted to US
+ 'url': 'http://www.nbcolympics.com/video/justin-roses-son-leo-was-tears-after-his-dad-won-gold',
+ 'md5': '54fecf846d05429fbaa18af557ee523a',
+ 'info_dict': {
+ 'id': 'WjTBzDXx5AUq',
+ 'display_id': 'justin-roses-son-leo-was-tears-after-his-dad-won-gold',
+ 'ext': 'mp4',
+ 'title': 'Rose\'s son Leo was in tears after his dad won gold',
+ 'description': 'Olympic gold medalist Justin Rose gets emotional talking to the impact his win in men\'s golf has already had on his children.',
+ 'timestamp': 1471274964,
+ 'upload_date': '20160815',
+ 'uploader': 'NBCU-SPORTS',
+ },
+ }
+
+ def _real_extract(self, url):
+ display_id = self._match_id(url)
+
+ webpage = self._download_webpage(url, display_id)
+
+ drupal_settings = self._parse_json(self._search_regex(
+ r'jQuery\.extend\(Drupal\.settings\s*,\s*({.+?})\);',
+ webpage, 'drupal settings'), display_id)
+
+ iframe_url = drupal_settings['vod']['iframe_url']
+ theplatform_url = iframe_url.replace(
+ 'vplayer.nbcolympics.com', 'player.theplatform.com')
+
+ return {
+ '_type': 'url_transparent',
+ 'url': theplatform_url,
+ 'ie_key': ThePlatformIE.ie_key(),
+ 'display_id': display_id,
+ }