mirror of
https://github.com/yt-dlp/yt-dlp.git
synced 2024-12-25 11:36:28 -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
98 lines
3.3 KiB
Python
98 lines
3.3 KiB
Python
from .common import InfoExtractor
|
||
from ..utils import (
|
||
clean_html,
|
||
traverse_obj,
|
||
unescapeHTML,
|
||
)
|
||
|
||
import itertools
|
||
from urllib.parse import urlencode
|
||
|
||
|
||
class RadioKapitalBaseIE(InfoExtractor):
|
||
def _call_api(self, resource, video_id, note='Downloading JSON metadata', qs={}):
|
||
return self._download_json(
|
||
f'https://www.radiokapital.pl/wp-json/kapital/v1/{resource}?{urlencode(qs)}',
|
||
video_id, note=note)
|
||
|
||
def _parse_episode(self, data):
|
||
release = '%s%s%s' % (data['published'][6:11], data['published'][3:6], data['published'][:3])
|
||
return {
|
||
'_type': 'url_transparent',
|
||
'url': data['mixcloud_url'],
|
||
'ie_key': 'Mixcloud',
|
||
'title': unescapeHTML(data['title']),
|
||
'description': clean_html(data.get('content')),
|
||
'tags': traverse_obj(data, ('tags', ..., 'name')),
|
||
'release_date': release,
|
||
'series': traverse_obj(data, ('show', 'title')),
|
||
}
|
||
|
||
|
||
class RadioKapitalIE(RadioKapitalBaseIE):
|
||
IE_NAME = 'radiokapital'
|
||
_VALID_URL = r'https?://(?:www\.)?radiokapital\.pl/shows/[a-z\d-]+/(?P<id>[a-z\d-]+)'
|
||
|
||
_TESTS = [{
|
||
'url': 'https://radiokapital.pl/shows/tutaj-sa-smoki/5-its-okay-to-be-immaterial',
|
||
'info_dict': {
|
||
'id': 'radiokapital_radio-kapitał-tutaj-są-smoki-5-its-okay-to-be-immaterial-2021-05-20',
|
||
'ext': 'm4a',
|
||
'title': '#5: It’s okay to\xa0be\xa0immaterial',
|
||
'description': 'md5:2499da5fbfb0e88333b7d37ec8e9e4c4',
|
||
'uploader': 'Radio Kapitał',
|
||
'uploader_id': 'radiokapital',
|
||
'timestamp': 1621640164,
|
||
'upload_date': '20210521',
|
||
},
|
||
}]
|
||
|
||
def _real_extract(self, url):
|
||
video_id = self._match_id(url)
|
||
|
||
episode = self._call_api('episodes/%s' % video_id, video_id)
|
||
return self._parse_episode(episode)
|
||
|
||
|
||
class RadioKapitalShowIE(RadioKapitalBaseIE):
|
||
IE_NAME = 'radiokapital:show'
|
||
_VALID_URL = r'https?://(?:www\.)?radiokapital\.pl/shows/(?P<id>[a-z\d-]+)/?(?:$|[?#])'
|
||
|
||
_TESTS = [{
|
||
'url': 'https://radiokapital.pl/shows/wesz',
|
||
'info_dict': {
|
||
'id': '100',
|
||
'title': 'WĘSZ',
|
||
'description': 'md5:3a557a1e0f31af612b0dcc85b1e0ca5c',
|
||
},
|
||
'playlist_mincount': 17,
|
||
}]
|
||
|
||
def _get_episode_list(self, series_id, page_no):
|
||
return self._call_api(
|
||
'episodes', series_id,
|
||
f'Downloading episode list page #{page_no}', qs={
|
||
'show': series_id,
|
||
'page': page_no,
|
||
})
|
||
|
||
def _entries(self, series_id):
|
||
for page_no in itertools.count(1):
|
||
episode_list = self._get_episode_list(series_id, page_no)
|
||
yield from (self._parse_episode(ep) for ep in episode_list['items'])
|
||
if episode_list['next'] is None:
|
||
break
|
||
|
||
def _real_extract(self, url):
|
||
series_id = self._match_id(url)
|
||
|
||
show = self._call_api(f'shows/{series_id}', series_id, 'Downloading show metadata')
|
||
entries = self._entries(series_id)
|
||
return {
|
||
'_type': 'playlist',
|
||
'entries': entries,
|
||
'id': str(show['id']),
|
||
'title': show.get('title'),
|
||
'description': clean_html(show.get('content')),
|
||
}
|