Cool_Tools/yt-dlp
1
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-11-24 15:50:32 +01:00
Code Issues Packages Projects Releases Wiki Activity

[patreon] Ignore erroneous media attachments (#4638)

Browse Source 
Fixes https://github.com/yt-dlp/yt-dlp/issues/4608
Authored by: coletdjnz
This commit is contained in:
coletdjnz 2022-08-13 00:25:20 +00:00 committed by GitHub
parent ffcd62c289
commit cea4b857f0
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 30 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

33
yt_dlp/extractor/patreon.py
View File

@ -154,6 +154,28 @@ class PatreonIE(PatreonBaseIE):
'channel_url': 'https://www.patreon.com/loish', 'channel_url': 'https://www.patreon.com/loish',
'channel_follower_count': int, 'channel_follower_count': int,
} }
}, {
# bad videos under media (if media is included). Real one is under post_file
'url': 'https://www.patreon.com/posts/premium-access-70282931',
'info_dict': {
'id': '70282931',
'ext': 'mp4',
'title': '[Premium Access + Uncut] The Office - 2x6 The Fight - Group Reaction',
'channel_url': 'https://www.patreon.com/thenormies',
'channel_id': '573397',
'uploader_id': '2929435',
'uploader': 'The Normies',
'description': 'md5:79c9fd8778e2cef84049a94c058a5e23',
'comment_count': int,
'upload_date': '20220809',
'thumbnail': r're:^https?://.*$',
'channel_follower_count': int,
'like_count': int,
'timestamp': 1660052820,
'tags': ['The Office', 'early access', 'uncut'],
'uploader_url': 'https://www.patreon.com/thenormies',
},
'skip': 'Patron-only content',
}] }]
def _real_extract(self, url): def _real_extract(self, url):
@ -166,7 +188,7 @@ def _real_extract(self, url):
'fields[post_tag]': 'value', 'fields[post_tag]': 'value',
'fields[campaign]': 'url,name,patron_count', 'fields[campaign]': 'url,name,patron_count',
'json-api-use-default-includes': 'false', 'json-api-use-default-includes': 'false',
'include': 'media,user,user_defined_tags,campaign', 'include': 'audio,user,user_defined_tags,campaign,attachments_media',
}) })
attributes = post['data']['attributes'] attributes = post['data']['attributes']
title = attributes['title'].strip() title = attributes['title'].strip()
@ -190,11 +212,16 @@ def _real_extract(self, url):
media_attributes = i.get('attributes') or {} media_attributes = i.get('attributes') or {}
download_url = media_attributes.get('download_url') download_url = media_attributes.get('download_url')
ext = mimetype2ext(media_attributes.get('mimetype')) ext = mimetype2ext(media_attributes.get('mimetype'))
if download_url and ext in KNOWN_EXTENSIONS:
# if size_bytes is None, this media file is likely unavailable
# See: https://github.com/yt-dlp/yt-dlp/issues/4608
size_bytes = int_or_none(media_attributes.get('size_bytes'))
if download_url and ext in KNOWN_EXTENSIONS and size_bytes is not None:
# XXX: what happens if there are multiple attachments?
return { return {
**info, **info,
'ext': ext, 'ext': ext,
'filesize': int_or_none(media_attributes.get('size_bytes')), 'filesize': size_bytes,
'url': download_url, 'url': download_url,
} }
elif i_type == 'user': elif i_type == 'user':