From d948e09b6174179adb43ac10cdb95dffeb4854dd Mon Sep 17 00:00:00 2001
From: Yen Chi Hsuan <yan12125@gmail.com>
Date: Sat, 2 May 2015 01:19:06 +0800
Subject: [PATCH] [viki] Extract m3u8 videos (#4855)

---
 youtube_dl/extractor/viki.py | 34 +++++++++++++++++++++++++++++++---
 1 file changed, 31 insertions(+), 3 deletions(-)

diff --git a/youtube_dl/extractor/viki.py b/youtube_dl/extractor/viki.py
index 0fc1ceb19..f85e43042 100644
--- a/youtube_dl/extractor/viki.py
+++ b/youtube_dl/extractor/viki.py
@@ -12,6 +12,8 @@
     unified_strdate,
     US_RATINGS,
     clean_html,
+    determine_ext,
+    mimetype2ext,
 )
 from .common import InfoExtractor
 
@@ -45,6 +47,19 @@ class VikiIE(InfoExtractor):
             'upload_date': '20150430',
             'title': '\'The Avengers: Age of Ultron\' Press Conference',
         }
+    }, {
+        'url': 'http://www.viki.com/videos/1048879v-ankhon-dekhi',
+        'info_dict': {
+            'id': '1048879v',
+            'ext': 'mp4',
+            'upload_date': '20140820',
+            'description': 'md5:54ff56d51bdfc7a30441ec967394e91c',
+            'title': 'Ankhon Dekhi',
+        },
+        'params': {
+            # requires ffmpeg
+            'skip_download': True,
+        }
     }]
 
     def _real_extract(self, url):
@@ -81,8 +96,21 @@ def _real_extract(self, url):
                     expected=True)
             else:
                 raise ExtractorError('Viki said: ' + err_msg)
-        video_url = self._html_search_regex(
-            r'<source[^>]+src="([^"]+)"', info_webpage, 'video URL')
+        mobj = re.search(
+            r'<source[^>]+type="(?P<mime_type>[^"]+)"[^>]+src="(?P<url>[^"]+)"', info_webpage)
+        if not mobj:
+            raise ExtractorError('Unable to find video URL')
+        video_url = unescapeHTML(mobj.group('url'))
+        video_ext = mimetype2ext(mobj.group('mime_type'))
+
+        if determine_ext(video_url) == 'm3u8':
+            formats = self._extract_m3u8_formats(
+                video_url, video_id, ext=video_ext)
+        else:
+            formats = [{
+                'url': video_url,
+                'ext': video_ext,
+            }]
 
         upload_date_str = self._html_search_regex(
             r'"created_at":"([^"]+)"', info_webpage, 'upload date')
@@ -98,7 +126,7 @@ def _real_extract(self, url):
         return {
             'id': video_id,
             'title': title,
-            'url': video_url,
+            'formats': formats,
             'description': description,
             'thumbnail': thumbnail,
             'age_limit': age_limit,