mirror of
https://github.com/yt-dlp/yt-dlp.git
synced 2024-12-23 10:37:18 -05:00
86e5f3ed2e
Using https://github.com/asottile/pyupgrade 1. `__future__` imports and `coding: utf-8` were removed 2. Files were rewritten with `pyupgrade --py36-plus --keep-percent-format` 3. f-strings were cherry-picked from `pyupgrade --py36-plus` Extractors are left untouched (except removing header) to avoid unnecessary merge conflicts
82 lines
3.4 KiB
Python
82 lines
3.4 KiB
Python
from .common import InfoExtractor
|
|
from ..utils import int_or_none
|
|
|
|
|
|
class KelbyOneIE(InfoExtractor):
|
|
_VALID_URL = r'https?://members\.kelbyone\.com/course/(?P<id>[^$&?#/]+)'
|
|
|
|
_TESTS = [{
|
|
'url': 'https://members.kelbyone.com/course/glyn-dewis-mastering-selections/',
|
|
'playlist_mincount': 1,
|
|
'info_dict': {
|
|
'id': 'glyn-dewis-mastering-selections',
|
|
'title': 'Trailer - Mastering Selections in Photoshop',
|
|
},
|
|
'playlist': [{
|
|
'info_dict': {
|
|
'id': 'MkiOnLqK',
|
|
'ext': 'mp4',
|
|
'title': 'Trailer - Mastering Selections in Photoshop',
|
|
'description': 'md5:d41d8cd98f00b204e9800998ecf8427e',
|
|
'thumbnail': 'https://content.jwplatform.com/v2/media/MkiOnLqK/poster.jpg?width=720',
|
|
'timestamp': 1601568639,
|
|
'duration': 90,
|
|
'upload_date': '20201001',
|
|
},
|
|
}]
|
|
}]
|
|
|
|
def _entries(self, playlist):
|
|
for item in playlist:
|
|
video_id = item['mediaid']
|
|
thumbnails = [{
|
|
'url': image.get('src'),
|
|
'width': int_or_none(image.get('width')),
|
|
} for image in item.get('images') or []]
|
|
formats, subtitles = [], {}
|
|
for source in item.get('sources') or []:
|
|
if not source.get('file'):
|
|
continue
|
|
if source.get('type') == 'application/vnd.apple.mpegurl':
|
|
fmts, subs = self._extract_m3u8_formats_and_subtitles(source['file'], video_id)
|
|
formats.extend(fmts)
|
|
subtitles = self._merge_subtitles(subs, subtitles)
|
|
elif source.get('type') == 'audio/mp4':
|
|
formats.append({
|
|
'format_id': source.get('label'),
|
|
'url': source['file'],
|
|
'vcodec': 'none',
|
|
})
|
|
else:
|
|
formats.append({
|
|
'format_id': source.get('label'),
|
|
'height': source.get('height'),
|
|
'width': source.get('width'),
|
|
'url': source['file'],
|
|
})
|
|
for track in item.get('tracks'):
|
|
if track.get('kind') == 'captions' and track.get('file'):
|
|
subtitles.setdefault('en', []).append({
|
|
'url': track['file'],
|
|
})
|
|
self._sort_formats(formats)
|
|
yield {
|
|
'id': video_id,
|
|
'title': item['title'],
|
|
'description': item.get('description'),
|
|
'thumbnails': thumbnails,
|
|
'thumbnail': item.get('image'),
|
|
'timestamp': item.get('pubdate'),
|
|
'duration': item.get('duration'),
|
|
'formats': formats,
|
|
'subtitles': subtitles,
|
|
}
|
|
|
|
def _real_extract(self, url):
|
|
item_id = self._match_id(url)
|
|
webpage = self._download_webpage(url, item_id)
|
|
playlist_url = self._html_search_regex(r'playlist"\:"(https.*content\.jwplatform\.com.*json)"', webpage, 'playlist url').replace('\\', '')
|
|
course_data = self._download_json(playlist_url, item_id)
|
|
return self.playlist_result(self._entries(course_data['playlist']), item_id,
|
|
course_data.get('title'), course_data.get('description'))
|