aboutsummaryrefslogtreecommitdiff
path: root/youtube_dl/extractor/youtube.py
diff options
context:
space:
mode:
Diffstat (limited to 'youtube_dl/extractor/youtube.py')
-rw-r--r--youtube_dl/extractor/youtube.py26
1 files changed, 18 insertions, 8 deletions
diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index 2496d27f1..eb5a58807 100644
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -2478,24 +2478,37 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
headers = {
'x-youtube-client-name': '1',
'x-youtube-client-version': '2.20201112.04.01',
+ 'content-type': 'application/json',
}
if identity_token:
headers['x-youtube-identity-token'] = identity_token
+ data = {
+ 'context': {
+ 'client': {
+ 'clientName': 'WEB',
+ 'clientVersion': '2.20201021.03.00',
+ }
+ },
+ }
+
for page_num in itertools.count(1):
if not continuation:
break
+ data['continuation'] = continuation['continuation']
+ data['clickTracking'] = {
+ 'clickTrackingParams': continuation['itct']
+ }
count = 0
retries = 3
while count <= retries:
try:
# Downloading page may result in intermittent 5xx HTTP error
# that is usually worked around with a retry
- browse = self._download_json(
- 'https://www.youtube.com/browse_ajax', None,
- 'Downloading page %d%s'
- % (page_num, ' (retry #%d)' % count if count else ''),
- headers=headers, query=continuation)
+ response = self._download_json(
+ 'https://www.youtube.com/youtubei/v1/browse?key=AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8',
+ None, 'Downloading page %d%s' % (page_num, ' (retry #%d)' % count if count else ''),
+ headers=headers, data=json.dumps(data).encode('utf8'))
break
except ExtractorError as e:
if isinstance(e.cause, compat_HTTPError) and e.cause.code in (500, 503):
@@ -2503,9 +2516,6 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
if count <= retries:
continue
raise
- if not browse:
- break
- response = try_get(browse, lambda x: x[1]['response'], dict)
if not response:
break