[tmz:article] Fix extraction (closes #11052)
authorYen Chi Hsuan <yan12125@gmail.com>
Tue, 8 Nov 2016 13:53:41 +0000 (21:53 +0800)
committerYen Chi Hsuan <yan12125@gmail.com>
Tue, 8 Nov 2016 13:53:41 +0000 (21:53 +0800)
ChangeLog
youtube_dl/extractor/tmz.py

index 7ca72b3d7de2de7022312bb54ecc3ee80a75a2c0..78c78afcdbbb3699695ab0d89c84abf089080c48 100644 (file)
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,6 +1,7 @@
 version <unreleased>
 
 Extractors
+* [tmz:article] Fix extraction (#11052)
 * [mitele] Fix extraction after website redesign (#10824)
 
 
index 979856e9a6663332a5eed7be8c2f9a30a9495d40..419f9d92eea375e4aeabc0c39f736955b818c921 100644 (file)
@@ -32,12 +32,15 @@ class TMZArticleIE(InfoExtractor):
     _VALID_URL = r'https?://(?:www\.)?tmz\.com/\d{4}/\d{2}/\d{2}/(?P<id>[^/]+)/?'
     _TEST = {
         'url': 'http://www.tmz.com/2015/04/19/bobby-brown-bobbi-kristina-awake-video-concert',
-        'md5': 'e482a414a38db73087450e3a6ce69d00',
+        'md5': '3316ff838ae5bb7f642537825e1e90d2',
         'info_dict': {
             'id': '0_6snoelag',
-            'ext': 'mp4',
+            'ext': 'mov',
             'title': 'Bobby Brown Tells Crowd ... Bobbi Kristina is Awake',
             'description': 'Bobby Brown stunned his audience during a concert Saturday night, when he told the crowd, "Bobbi is awake.  She\'s watching me."',
+            'timestamp': 1429467813,
+            'upload_date': '20150419',
+            'uploader_id': 'batchUser',
         }
     }
 
@@ -45,12 +48,9 @@ class TMZArticleIE(InfoExtractor):
         video_id = self._match_id(url)
 
         webpage = self._download_webpage(url, video_id)
-        embedded_video_info_str = self._html_search_regex(
-            r'tmzVideoEmbedV2\("([^)]+)"\);', webpage, 'embedded video info')
-
-        embedded_video_info = self._parse_json(
-            embedded_video_info_str, video_id,
-            transform_source=lambda s: s.replace('\\', ''))
+        embedded_video_info = self._parse_json(self._html_search_regex(
+            r'tmzVideoEmbed\(({.+?})\);', webpage, 'embedded video info'),
+            video_id)
 
         return self.url_result(
             'http://www.tmz.com/videos/%s/' % embedded_video_info['id'])