浏览代码

[JWPlatform] Support iframes

Support content.jwplatform... src attributes inside <iframe> tags in
addition to <script> tags. Just a regexp change.

Add a test (currently Generic_72).
John Hawkinson 7 年之前
父节点
当前提交
939be9adfe
共有 2 个文件被更改,包括 13 次插入1 次删除
  1. 12 0
      youtube_dl/extractor/generic.py
  2. 1 1
      youtube_dl/extractor/jwplatform.py

+ 12 - 0
youtube_dl/extractor/generic.py

@@ -1136,6 +1136,18 @@ class GenericIE(InfoExtractor):
                 'skip_download': True,
             }
         },
+        {
+            # JWPlatform iframe
+            'url': 'https://www.mediaite.com/tv/dem-senator-claims-gary-cohn-faked-a-bad-connection-during-trump-call-to-get-him-off-the-phone/',
+            'md5': 'ca00a040364b5b439230e7ebfd02c4e9',
+            'info_dict': {
+                'id': 'O0c5JcKT',
+                'ext': 'mp4',
+                'upload_date': '20171122',
+                'timestamp': 1511366290,
+                'title': 'Dem Senator Claims Gary Cohn Faked a Bad Connection During Trump Call to Get Him Off the Phone',
+            },
+        },
         {
             # Video.js embed, multiple formats
             'url': 'http://ortcam.com/solidworks-урок-6-настройка-чертежа_33f9b7351.html',

+ 1 - 1
youtube_dl/extractor/jwplatform.py

@@ -24,7 +24,7 @@ class JWPlatformIE(InfoExtractor):
     @staticmethod
     def _extract_url(webpage):
         mobj = re.search(
-            r'<script[^>]+?src=["\'](?P<url>(?:https?:)?//content.jwplatform.com/players/[a-zA-Z0-9]{8})',
+            r'<(script|iframe)[^>]+?src=["\'](?P<url>(?:https?:)?//content.jwplatform.com/players/[a-zA-Z0-9]{8})',
             webpage)
         if mobj:
             return mobj.group('url')