mirrors/yt-dlp
1
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2025-11-13 13:05:13 +00:00
Code Issues Projects Releases Wiki Activity

[ie/GoogleDrive] Fix subtitles extraction (#14809)

Browse Source 
Closes #14806
Authored by: seproDev
This commit is contained in:
sepro
2025-10-28 16:17:16 +01:00
committed by GitHub
parent 1c2ad94353
commit 6d05cee4df
1 changed files with 18 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
yt_dlp/extractor/googledrive.py
View File

@@ -4,6 +4,7 @@ from .common import InfoExtractor
from ..utils import ( from ..utils import (
determine_ext, determine_ext,
extract_attributes, extract_attributes,
filter_dict,
get_element_by_class, get_element_by_class,
get_element_html_by_id, get_element_html_by_id,
int_or_none, int_or_none,
@@ -60,6 +61,17 @@ class GoogleDriveIE(InfoExtractor):
'duration': 9.999, 'duration': 9.999,
'thumbnail': r're:https://lh3\.googleusercontent\.com/drive-storage/', 'thumbnail': r're:https://lh3\.googleusercontent\.com/drive-storage/',
}, },
}, {
# Has subtitle track with kind 'asr'
'url': 'https://drive.google.com/file/d/1Prvv9-mtDDfN_gkJgtt1OFvIULK8c3Ev/view',
'md5': 'ccae12d07f18b5988900b2c8b92801fc',
'info_dict': {
'id': '1Prvv9-mtDDfN_gkJgtt1OFvIULK8c3Ev',
'ext': 'mp4',
'title': 'LEE NA GYUNG-3410-VOICE_MESSAGE.mp4',
'duration': 8.766,
'thumbnail': r're:https://lh3\.googleusercontent\.com/drive-storage/',
},
}, { }, {
# 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)
@@ -89,15 +101,16 @@ class GoogleDriveIE(InfoExtractor):
yield 'https://drive.google.com/file/d/{}'.format(mobj.group('id')) yield 'https://drive.google.com/file/d/{}'.format(mobj.group('id'))
@staticmethod @staticmethod
def _construct_subtitle_url(base_url, video_id, language, fmt): def _construct_subtitle_url(base_url, video_id, language, fmt, kind):
return update_url_query( return update_url_query(
base_url, { base_url, filter_dict({
'hl': 'en-US', 'hl': 'en-US',
'v': video_id, 'v': video_id,
'type': 'track', 'type': 'track',
'lang': language, 'lang': language,
'fmt': fmt, 'fmt': fmt,
}) 'kind': kind,
}))
def _get_subtitles(self, video_id, video_info): def _get_subtitles(self, video_id, video_info):
subtitles = {} subtitles = {}
@@ -116,7 +129,8 @@ class GoogleDriveIE(InfoExtractor):
for track in traverse_obj(subtitle_data, (lambda _, v: v.tag == 'track' and v.get('lang_code'))): for track in traverse_obj(subtitle_data, (lambda _, v: v.tag == 'track' and v.get('lang_code'))):
language = track.get('lang_code') language = track.get('lang_code')
subtitles.setdefault(language, []).extend([{ subtitles.setdefault(language, []).extend([{
'url': self._construct_subtitle_url(timed_text_base_url, video_id, language, sub_fmt), 'url': self._construct_subtitle_url(
timed_text_base_url, video_id, language, sub_fmt, track.get('kind')),
'name': track.get('lang_original'), 'name': track.get('lang_original'),
'ext': sub_fmt, 'ext': sub_fmt,
} for sub_fmt in subtitle_formats]) } for sub_fmt in subtitle_formats])