midgard/youtube-dl
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

[extractor/common] keep support for non standard JSON-LD VideoObject author values

Browse source
This commit is contained in:
Remita Amine 2021-04-04 19:16:17 +01:00
parent 3ae9c0f410
commit 6beb1ac65b
1 changed files with 6 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

8
youtube_dl/extractor/common.py
View file

@ -70,7 +70,6 @@ from ..utils import (
str_or_none, str_or_none,
str_to_int, str_to_int,
strip_or_none, strip_or_none,
try_get,
unescapeHTML, unescapeHTML,
unified_strdate, unified_strdate,
unified_timestamp, unified_timestamp,
@ -1276,6 +1275,7 @@ class InfoExtractor(object):
def extract_video_object(e): def extract_video_object(e):
assert e['@type'] == 'VideoObject' assert e['@type'] == 'VideoObject'
author = e.get('author')
info.update({ info.update({
'url': url_or_none(e.get('contentUrl')), 'url': url_or_none(e.get('contentUrl')),
'title': unescapeHTML(e.get('name')), 'title': unescapeHTML(e.get('name')),
@ -1283,7 +1283,11 @@ class InfoExtractor(object):
'thumbnail': url_or_none(e.get('thumbnailUrl') or e.get('thumbnailURL')), 'thumbnail': url_or_none(e.get('thumbnailUrl') or e.get('thumbnailURL')),
'duration': parse_duration(e.get('duration')), 'duration': parse_duration(e.get('duration')),
'timestamp': unified_timestamp(e.get('uploadDate')), 'timestamp': unified_timestamp(e.get('uploadDate')),
'uploader': try_get(e, lambda x: x['author']['name'], compat_str), # author can be an instance of 'Organization' or 'Person' types.
# both types can have 'name' property(inherited from 'Thing' type). [1]
# however some websites are using 'Text' type instead.
# 1. https://schema.org/VideoObject
'uploader': author.get('name') if isinstance(author, dict) else author if isinstance(author, compat_str) else None,
'filesize': float_or_none(e.get('contentSize')), 'filesize': float_or_none(e.get('contentSize')),
'tbr': int_or_none(e.get('bitrate')), 'tbr': int_or_none(e.get('bitrate')),
'width': int_or_none(e.get('width')), 'width': int_or_none(e.get('width')),