summary refs log tree commit diff
diff options
context:
space:
mode:
authorSergey M․ <dstftw@gmail.com>2018-05-01 03:09:04 +0700
committerSergey M․ <dstftw@gmail.com>2018-05-01 03:09:04 +0700
commitc21692fa94df49ef925c06c00e5db1d8bb0f770d (patch)
treeb65514fe175bbc1741a8e717f95f56671e919149
parent851396346803f77ab9573af56cae056aa904cf93 (diff)
downloadyoutube-dl-c21692fa94df49ef925c06c00e5db1d8bb0f770d.tar.gz
youtube-dl-c21692fa94df49ef925c06c00e5db1d8bb0f770d.tar.xz
youtube-dl-c21692fa94df49ef925c06c00e5db1d8bb0f770d.zip
[kaltura] Improve iframe embeds detection (closes #16337)
-rw-r--r--youtube_dl/extractor/generic.py17
-rw-r--r--youtube_dl/extractor/kaltura.py3
2 files changed, 19 insertions, 1 deletions
diff --git a/youtube_dl/extractor/generic.py b/youtube_dl/extractor/generic.py
index 252f97c26..73980601c 100644
--- a/youtube_dl/extractor/generic.py
+++ b/youtube_dl/extractor/generic.py
@@ -1283,6 +1283,23 @@ class GenericIE(InfoExtractor):
             'add_ie': ['Kaltura'],
         },
         {
+            # Kaltura iframe embed, more sophisticated
+            'url': 'http://www.cns.nyu.edu/~eero/math-tools/Videos/lecture-05sep2017.html',
+            'info_dict': {
+                'id': '1_9gzouybz',
+                'ext': 'mp4',
+                'title': 'lecture-05sep2017',
+                'description': 'md5:40f347d91fd4ba047e511c5321064b49',
+                'upload_date': '20170913',
+                'uploader_id': 'eps2',
+                'timestamp': 1505340777,
+            },
+            'params': {
+                'skip_download': True,
+            },
+            'add_ie': ['Kaltura'],
+        },
+        {
             # meta twitter:player
             'url': 'http://thechive.com/2017/12/08/all-i-want-for-christmas-is-more-twerk/',
             'info_dict': {
diff --git a/youtube_dl/extractor/kaltura.py b/youtube_dl/extractor/kaltura.py
index 0ea89e4d6..04f68fce4 100644
--- a/youtube_dl/extractor/kaltura.py
+++ b/youtube_dl/extractor/kaltura.py
@@ -136,9 +136,10 @@ class KalturaIE(InfoExtractor):
             re.search(
                 r'''(?xs)
                     <(?:iframe[^>]+src|meta[^>]+\bcontent)=(?P<q1>["'])
-                      (?:https?:)?//(?:(?:www|cdnapi)\.)?kaltura\.com/(?:(?!(?P=q1)).)*\b(?:p|partner_id)/(?P<partner_id>\d+)
+                      (?:https?:)?//(?:(?:www|cdnapi(?:sec)?)\.)?kaltura\.com/(?:(?!(?P=q1)).)*\b(?:p|partner_id)/(?P<partner_id>\d+)
                       (?:(?!(?P=q1)).)*
                       [?&;]entry_id=(?P<id>(?:(?!(?P=q1))[^&])+)
+                      (?:(?!(?P=q1)).)*
                     (?P=q1)
                 ''', webpage)
         )