about summary refs log tree commit diff
path: root/youtube_dl/extractor/thesun.py
diff options
context:
space:
mode:
authorFilippo Valsorda <hi@filippo.io>2018-01-07 15:03:28 +0100
committerGitHub <noreply@github.com>2018-01-07 15:03:28 +0100
commit97bc05116e15df3cf1b7b7216c60fad7ee0acc9f (patch)
tree453059413184d22a44241d251a82196c5d2461eb /youtube_dl/extractor/thesun.py
parent7608a91ee7b40c74a35c30b44731606382702592 (diff)
parent0a5b1295b7c1aa6395b65ee137087c540b37b32b (diff)
downloadyoutube-dl-97bc05116e15df3cf1b7b7216c60fad7ee0acc9f.tar.gz
youtube-dl-97bc05116e15df3cf1b7b7216c60fad7ee0acc9f.tar.xz
youtube-dl-97bc05116e15df3cf1b7b7216c60fad7ee0acc9f.zip
Merge branch 'master' into totalwebcasting totalwebcasting
Diffstat (limited to 'youtube_dl/extractor/thesun.py')
-rw-r--r--youtube_dl/extractor/thesun.py32
1 files changed, 32 insertions, 0 deletions
diff --git a/youtube_dl/extractor/thesun.py b/youtube_dl/extractor/thesun.py
new file mode 100644
index 000000000..22d003776
--- /dev/null
+++ b/youtube_dl/extractor/thesun.py
@@ -0,0 +1,32 @@
+from __future__ import unicode_literals
+
+import re
+
+from .common import InfoExtractor
+from .ooyala import OoyalaIE
+
+
+class TheSunIE(InfoExtractor):
+    _VALID_URL = r'https://(?:www\.)?thesun\.co\.uk/[^/]+/(?P<id>\d+)'
+    _TEST = {
+        'url': 'https://www.thesun.co.uk/tvandshowbiz/2261604/orlando-bloom-and-katy-perry-post-adorable-instagram-video-together-celebrating-thanksgiving-after-split-rumours/',
+        'info_dict': {
+            'id': '2261604',
+            'title': 'md5:cba22f48bad9218b64d5bbe0e16afddf',
+        },
+        'playlist_count': 2,
+    }
+
+    def _real_extract(self, url):
+        article_id = self._match_id(url)
+
+        webpage = self._download_webpage(url, article_id)
+
+        entries = []
+        for ooyala_id in re.findall(
+                r'<[^>]+\b(?:id\s*=\s*"thesun-ooyala-player-|data-content-id\s*=\s*")([^"]+)',
+                webpage):
+            entries.append(OoyalaIE._build_url_result(ooyala_id))
+
+        return self.playlist_result(
+            entries, article_id, self._og_search_title(webpage, fatal=False))