[viki] Fix description extraction (closes #6339)

[youtube-dl] / youtube_dl / extractor / viki.py

diff --git a/youtube_dl/extractor/viki.py b/youtube_dl/extractor/viki.py
index 51cdc6b65143aaf4a0d2823ffa8c859c96e25972..e987badbde608770ce8449c9c06158c8570e3506 100644 (file)
--- a/youtube_dl/extractor/viki.py
+++ b/youtube_dl/extractor/viki.py
@@ -173,6 +173,19 @@ class VikiIE(VikiBaseIE):
     }, {
         'url': 'http://www.viki.com/player/44699v',
         'only_matching': True,
+    }, {
+        # non-English description
+        'url': 'http://www.viki.com/videos/158036v-love-in-magic',
+        'md5': '1713ae35df5a521b31f6dc40730e7c9c',
+        'info_dict': {
+            'id': '158036v',
+            'ext': 'mp4',
+            'uploader': 'I Planet Entertainment',
+            'upload_date': '20111122',
+            'timestamp': 1321985454,
+            'description': 'md5:44b1e46619df3a072294645c770cef36',
+            'title': 'Love In Magic',
+        },
     }]
 
     def _real_extract(self, url):
@@ -192,8 +205,12 @@ class VikiIE(VikiBaseIE):
                 container_title = container_titles.get('en') or container_titles[container_titles.keys()[0]]
                 title = '%s - %s' % (container_title, title)
 
-        descriptions = video.get('descriptions')
-        description = descriptions.get('en') or descriptions[titles.keys()[0]] if descriptions else None
+        descriptions = video.get('descriptions', {})
+        description = descriptions.get('en')
+        if description is None:
+            filtered_descriptions = list(filter(None, [descriptions.get(k) for k in titles.keys()]))
+            if filtered_descriptions:
+                description = filtered_descriptions[0]
 
         duration = int_or_none(video.get('duration'))
         timestamp = parse_iso8601(video.get('created_at'))
@@ -242,8 +259,8 @@ class VikiIE(VikiBaseIE):
 
         formats = []
         for format_id, stream_dict in streams.items():
-            height = self._search_regex(
-                r'^(\d+)[pP]$', format_id, 'height', default=None)
+            height = int_or_none(self._search_regex(
+                r'^(\d+)[pP]$', format_id, 'height', default=None))
             for protocol, format_dict in stream_dict.items():
                 if format_id == 'm3u8':
                     formats = self._extract_m3u8_formats(