[ie/googledrive] Fix formats extraction (#9908)

Closes #8281
Authored by: WyohKnott
This commit is contained in:
WyohKnott 2024-05-13 01:05:47 +02:00 committed by GitHub
parent b207d26f83
commit 85ec2a337a
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 20 additions and 18 deletions

View File

@ -1,9 +1,11 @@
import re import re
from .common import InfoExtractor from .common import InfoExtractor
from .youtube import YoutubeIE
from ..compat import compat_parse_qs from ..compat import compat_parse_qs
from ..utils import ( from ..utils import (
ExtractorError, ExtractorError,
bug_reports_message,
determine_ext, determine_ext,
extract_attributes, extract_attributes,
get_element_by_class, get_element_by_class,
@ -38,6 +40,17 @@ class GoogleDriveIE(InfoExtractor):
'duration': 45, 'duration': 45,
'thumbnail': 'https://drive.google.com/thumbnail?id=0ByeS4oOUV-49Zzh4R1J6R09zazQ', 'thumbnail': 'https://drive.google.com/thumbnail?id=0ByeS4oOUV-49Zzh4R1J6R09zazQ',
} }
}, {
# has itag 50 which is not in YoutubeIE._formats (royalty Free music from 1922)
'url': 'https://drive.google.com/uc?id=1IP0o8dHcQrIHGgVyp0Ofvx2cGfLzyO1x',
'md5': '322db8d63dd19788c04050a4bba67073',
'info_dict': {
'id': '1IP0o8dHcQrIHGgVyp0Ofvx2cGfLzyO1x',
'ext': 'mp3',
'title': 'My Buddy - Henry Burr - Gus Kahn - Walter Donaldson.mp3',
'duration': 184,
'thumbnail': 'https://drive.google.com/thumbnail?id=1IP0o8dHcQrIHGgVyp0Ofvx2cGfLzyO1x',
},
}, { }, {
# video can't be watched anonymously due to view count limit reached, # video can't be watched anonymously due to view count limit reached,
# but can be downloaded (see https://github.com/ytdl-org/youtube-dl/issues/14046) # but can be downloaded (see https://github.com/ytdl-org/youtube-dl/issues/14046)
@ -58,22 +71,8 @@ class GoogleDriveIE(InfoExtractor):
'only_matching': True, 'only_matching': True,
}] }]
_FORMATS_EXT = { _FORMATS_EXT = {
'5': 'flv', **{k: v['ext'] for k, v in YoutubeIE._formats.items() if v.get('ext')},
'6': 'flv', '50': 'm4a',
'13': '3gp',
'17': '3gp',
'18': 'mp4',
'22': 'mp4',
'34': 'flv',
'35': 'flv',
'36': '3gp',
'37': 'mp4',
'38': 'mp4',
'43': 'webm',
'44': 'webm',
'45': 'webm',
'46': 'webm',
'59': 'mp4',
} }
_BASE_URL_CAPTIONS = 'https://drive.google.com/timedtext' _BASE_URL_CAPTIONS = 'https://drive.google.com/timedtext'
_CAPTIONS_ENTRY_TAG = { _CAPTIONS_ENTRY_TAG = {
@ -194,10 +193,13 @@ def get_value(key):
if len(fmt_stream_split) < 2: if len(fmt_stream_split) < 2:
continue continue
format_id, format_url = fmt_stream_split[:2] format_id, format_url = fmt_stream_split[:2]
ext = self._FORMATS_EXT.get(format_id)
if not ext:
self.report_warning(f'Unknown format {format_id}{bug_reports_message()}')
f = { f = {
'url': lowercase_escape(format_url), 'url': lowercase_escape(format_url),
'format_id': format_id, 'format_id': format_id,
'ext': self._FORMATS_EXT[format_id], 'ext': ext,
} }
resolution = resolutions.get(format_id) resolution = resolutions.get(format_id)
if resolution: if resolution:

View File

@ -1181,7 +1181,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$', r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$',
r'\b(?P<id>vfl[a-zA-Z0-9_-]+)\b.*?\.js$', r'\b(?P<id>vfl[a-zA-Z0-9_-]+)\b.*?\.js$',
) )
_formats = { _formats = { # NB: Used in YoutubeWebArchiveIE and GoogleDriveIE
'5': {'ext': 'flv', 'width': 400, 'height': 240, 'acodec': 'mp3', 'abr': 64, 'vcodec': 'h263'}, '5': {'ext': 'flv', 'width': 400, 'height': 240, 'acodec': 'mp3', 'abr': 64, 'vcodec': 'h263'},
'6': {'ext': 'flv', 'width': 450, 'height': 270, 'acodec': 'mp3', 'abr': 64, 'vcodec': 'h263'}, '6': {'ext': 'flv', 'width': 450, 'height': 270, 'acodec': 'mp3', 'abr': 64, 'vcodec': 'h263'},
'13': {'ext': '3gp', 'acodec': 'aac', 'vcodec': 'mp4v'}, '13': {'ext': '3gp', 'acodec': 'aac', 'vcodec': 'mp4v'},