youtube-dl

commit edc2a1f68b267abc6b4c94991da4ad83fd8374bb
parent 1ced222120c00854865c5b16e89838235ed549ee
Author: Sergey M․ <dstftw@gmail.com>
Date:   Wed, 27 Nov 2019 02:28:06 +0700

[vivo] Fix extraction (closes #22328, closes #22279)

Diffstat:
M
youtube_dl/extractor/shared.py
 | 
27
+++++++++++++++++++--------

1 file changed, 19 insertions(+), 8 deletions(-)
diff --git a/youtube_dl/extractor/shared.py b/youtube_dl/extractor/shared.py
@@ -1,13 +1,18 @@
 from __future__ import unicode_literals
 
 from .common import InfoExtractor
-from ..compat import compat_b64decode
+from ..compat import (
+    compat_b64decode,
+    compat_urllib_parse_unquote_plus,
+)
 from ..utils import (
     determine_ext,
     ExtractorError,
     int_or_none,
+    js_to_json,
     KNOWN_EXTENSIONS,
     parse_filesize,
+    rot47,
     url_or_none,
     urlencode_postdata,
 )
@@ -112,16 +117,22 @@ class VivoIE(SharedBaseIE):
             webpage, 'filesize', fatal=False))
 
     def _extract_video_url(self, webpage, video_id, url):
-        def decode_url(encoded_url):
+        def decode_url_old(encoded_url):
             return compat_b64decode(encoded_url).decode('utf-8')
 
-        stream_url = url_or_none(decode_url(self._search_regex(
+        stream_url = self._search_regex(
             r'data-stream\s*=\s*(["\'])(?P<url>(?:(?!\1).)+)\1', webpage,
-            'stream url', default=None, group='url')))
+            'stream url', default=None, group='url')
+        if stream_url:
+            stream_url = url_or_none(decode_url_old(stream_url))
         if stream_url:
             return stream_url
-        return self._parse_json(
+
+        def decode_url(encoded_url):
+            return rot47(compat_urllib_parse_unquote_plus(encoded_url))
+
+        return decode_url(self._parse_json(
             self._search_regex(
-                r'InitializeStream\s*\(\s*(["\'])(?P<url>(?:(?!\1).)+)\1',
-                webpage, 'stream', group='url'),
-            video_id, transform_source=decode_url)[0]
+                r'(?s)InitializeStream\s*\(\s*({.+?})\s*\)\s*;', webpage,
+                'stream'),
+            video_id, transform_source=js_to_json)['source'])