0
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-11-18 01:35:12 +00:00
yt-dlp/yt_dlp/extractor/aitube.py
HobbyistDev a1d9aca338
[extractor/aitube] Add extractor (#5946)
Closes #5627
Authored by: HobbyistDev
2023-01-04 17:03:36 +05:30

60 lines
2.9 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

from .common import InfoExtractor
from ..utils import int_or_none, merge_dicts
class AitubeKZVideoIE(InfoExtractor):
_VALID_URL = r'https?://aitube\.kz/(?:video|embed/)\?(?:[^\?]+)?id=(?P<id>[\w-]+)'
_TESTS = [{
# id paramater as first parameter
'url': 'https://aitube.kz/video?id=9291d29b-c038-49a1-ad42-3da2051d353c&playlistId=d55b1f5f-ef2a-4f23-b646-2a86275b86b7&season=1',
'info_dict': {
'id': '9291d29b-c038-49a1-ad42-3da2051d353c',
'ext': 'mp4',
'duration': 2174.0,
'channel_id': '94962f73-013b-432c-8853-1bd78ca860fe',
'like_count': int,
'channel': 'ASTANA TV',
'comment_count': int,
'view_count': int,
'description': 'Смотреть любимые сериалы и видео, поделиться видео и сериалами с друзьями и близкими',
'thumbnail': 'https://cdn.static02.aitube.kz/kz.aitudala.aitube.staticaccess/files/ddf2a2ff-bee3-409b-b5f2-2a8202bba75b',
'upload_date': '20221102',
'timestamp': 1667370519,
'title': 'Ангел хранитель 1 серия',
'channel_follower_count': int,
}
}, {
# embed url
'url': 'https://aitube.kz/embed/?id=9291d29b-c038-49a1-ad42-3da2051d353c',
'only_matching': True,
}, {
# id parameter is not as first paramater
'url': 'https://aitube.kz/video?season=1&id=9291d29b-c038-49a1-ad42-3da2051d353c&playlistId=d55b1f5f-ef2a-4f23-b646-2a86275b86b7',
'only_matching': True,
}]
def _real_extract(self, url):
video_id = self._match_id(url)
webpage = self._download_webpage(url, video_id)
nextjs_data = self._search_nextjs_data(webpage, video_id)['props']['pageProps']['videoInfo']
json_ld_data = self._search_json_ld(webpage, video_id)
formats, subtitles = self._extract_m3u8_formats_and_subtitles(
f'https://api-http.aitube.kz/kz.aitudala.aitube.staticaccess/video/{video_id}/video', video_id)
return merge_dicts({
'id': video_id,
'title': nextjs_data.get('title') or self._html_search_meta(['name', 'og:title'], webpage),
'description': nextjs_data.get('description'),
'formats': formats,
'subtitles': subtitles,
'view_count': (nextjs_data.get('viewCount')
or int_or_none(self._html_search_meta('ya:ovs:views_total', webpage))),
'like_count': nextjs_data.get('likeCount'),
'channel': nextjs_data.get('channelTitle'),
'channel_id': nextjs_data.get('channelId'),
'thumbnail': nextjs_data.get('coverUrl'),
'comment_count': nextjs_data.get('commentCount'),
'channel_follower_count': int_or_none(nextjs_data.get('channelSubscriberCount')),
}, json_ld_data)