[vivo] Fix extraction (closes #18906)

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

diff --git a/youtube_dl/extractor/shared.py b/youtube_dl/extractor/shared.py
index 89e19e9277f42b69ac6f01b03c78ced00f3ec990..eade8fd9e0148b5906c5ad8d89c1a8ab5fe17cda 100644 (file)
--- a/youtube_dl/extractor/shared.py
+++ b/youtube_dl/extractor/shared.py
@@ -1,12 +1,15 @@
 from __future__ import unicode_literals
 
-import base64
+import re
 
 from .common import InfoExtractor
+from ..compat import compat_b64decode
 from ..utils import (
     ExtractorError,
     int_or_none,
+    url_or_none,
     urlencode_postdata,
+    unescapeHTML,
 )
 
 
@@ -22,8 +25,7 @@ class SharedBaseIE(InfoExtractor):
 
         video_url = self._extract_video_url(webpage, video_id, url)
 
-        title = base64.b64decode(self._html_search_meta(
-            'full:title', webpage, 'title').encode('utf-8')).decode('utf-8')
+        title = self._extract_title(webpage)
         filesize = int_or_none(self._html_search_meta(
             'full:size', webpage, 'file size', fatal=False))
 
@@ -35,6 +37,10 @@ class SharedBaseIE(InfoExtractor):
             'title': title,
         }
 
+    def _extract_title(self, webpage):
+        return compat_b64decode(self._html_search_meta(
+            'full:title', webpage, 'title')).decode('utf-8')
+
 
 class SharedIE(SharedBaseIE):
     IE_DESC = 'shared.sx'
@@ -86,11 +92,25 @@ class VivoIE(SharedBaseIE):
         },
     }
 
+    def _extract_title(self, webpage):
+        data_title = self._search_regex(
+            r'data-name\s*=\s*(["\'])(?P<title>(?:(?!\1).)+)\1', webpage,
+            'title', default=None, group='title')
+        if data_title:
+            return unescapeHTML(re.sub(r"\.[a-z0-9]{3,4}$", "", data_title))
+        return self._og_search_title(webpage)
+
     def _extract_video_url(self, webpage, video_id, *args):
+        def decode_url(encoded_url):
+            return compat_b64decode(encoded_url).decode('utf-8')
+
+        stream_url = url_or_none(decode_url(self._search_regex(
+            r'data-stream\s*=\s*(["\'])(?P<url>(?:(?!\1).)+)\1', webpage,
+            'stream url', default=None, group='url')))
+        if stream_url:
+            return stream_url
         return self._parse_json(
             self._search_regex(
                 r'InitializeStream\s*\(\s*(["\'])(?P<url>(?:(?!\1).)+)\1',
                 webpage, 'stream', group='url'),
-            video_id,
-            transform_source=lambda x: base64.b64decode(
-                x.encode('ascii')).decode('utf-8'))[0]
+            video_id, transform_source=decode_url)[0]