You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
yt-dlp/yt_dlp/extractor/nhk.py

620 lines
25 KiB
Python

This file contains ambiguous Unicode characters!

This file contains ambiguous Unicode characters that may be confused with others in your current locale. If your use case is intentional and legitimate, you can safely ignore this warning. Use the Escape button to highlight these characters.

import re
from .common import InfoExtractor
from ..utils import (
ExtractorError,
int_or_none,
join_nonempty,
parse_duration,
traverse_obj,
unescapeHTML,
unified_timestamp,
url_or_none,
urljoin,
)
class NhkBaseIE(InfoExtractor):
_API_URL_TEMPLATE = 'https://nwapi.nhk.jp/nhkworld/%sod%slist/v7b/%s/%s/%s/all%s.json'
_BASE_URL_REGEX = r'https?://www3\.nhk\.or\.jp/nhkworld/(?P<lang>[a-z]{2})/ondemand'
_TYPE_REGEX = r'/(?P<type>video|audio)/'
def _call_api(self, m_id, lang, is_video, is_episode, is_clip):
return self._download_json(
self._API_URL_TEMPLATE % (
'v' if is_video else 'r',
'clip' if is_clip else 'esd',
'episode' if is_episode else 'program',
m_id, lang, '/all' if is_video else ''),
m_id, query={'apikey': 'EJfK8jdS57GqlupFgAfAAwr573q01y6k'})['data']['episodes'] or []
def _get_api_info(self, refresh=True):
if not refresh:
return self.cache.load('nhk', 'api_info')
self.cache.store('nhk', 'api_info', {})
movie_player_js = self._download_webpage(
'https://movie-a.nhk.or.jp/world/player/js/movie-player.js', None,
note='Downloading stream API information')
api_info = {
'url': self._search_regex(
r'prod:[^;]+\bapiUrl:\s*[\'"]([^\'"]+)[\'"]', movie_player_js, None, 'stream API url'),
'token': self._search_regex(
r'prod:[^;]+\btoken:\s*[\'"]([^\'"]+)[\'"]', movie_player_js, None, 'stream API token'),
}
self.cache.store('nhk', 'api_info', api_info)
return api_info
def _extract_formats_and_subtitles(self, vod_id):
for refresh in (False, True):
api_info = self._get_api_info(refresh)
if not api_info:
continue
api_url = api_info.pop('url')
stream_url = traverse_obj(
self._download_json(
api_url, vod_id, 'Downloading stream url info', fatal=False, query={
**api_info,
'type': 'json',
'optional_id': vod_id,
'active_flg': 1,
}),
('meta', 0, 'movie_url', ('mb_auto', 'auto_sp', 'auto_pc'), {url_or_none}), get_all=False)
if stream_url:
return self._extract_m3u8_formats_and_subtitles(stream_url, vod_id)
raise ExtractorError('Unable to extract stream url')
def _extract_episode_info(self, url, episode=None):
fetch_episode = episode is None
lang, m_type, episode_id = NhkVodIE._match_valid_url(url).group('lang', 'type', 'id')
is_video = m_type == 'video'
if is_video:
episode_id = episode_id[:4] + '-' + episode_id[4:]
if fetch_episode:
episode = self._call_api(
episode_id, lang, is_video, True, episode_id[:4] == '9999')[0]
title = episode.get('sub_title_clean') or episode['sub_title']
def get_clean_field(key):
return episode.get(key + '_clean') or episode.get(key)
series = get_clean_field('title')
thumbnails = []
for s, w, h in [('', 640, 360), ('_l', 1280, 720)]:
img_path = episode.get('image' + s)
if not img_path:
continue
thumbnails.append({
'id': '%dp' % h,
'height': h,
'width': w,
'url': 'https://www3.nhk.or.jp' + img_path,
})
info = {
'id': episode_id + '-' + lang,
'title': '%s - %s' % (series, title) if series and title else title,
'description': get_clean_field('description'),
'thumbnails': thumbnails,
'series': series,
'episode': title,
}
if is_video:
vod_id = episode['vod_id']
formats, subs = self._extract_formats_and_subtitles(vod_id)
info.update({
'id': vod_id,
'formats': formats,
'subtitles': subs,
})
else:
if fetch_episode:
audio_path = episode['audio']['audio']
info['formats'] = self._extract_m3u8_formats(
'https://nhkworld-vh.akamaihd.net/i%s/master.m3u8' % audio_path,
episode_id, 'm4a', entry_protocol='m3u8_native',
m3u8_id='hls', fatal=False)
for f in info['formats']:
f['language'] = lang
else:
info.update({
'_type': 'url_transparent',
'ie_key': NhkVodIE.ie_key(),
'url': url,
})
return info
class NhkVodIE(NhkBaseIE):
# the 7-character IDs can have alphabetic chars too: assume [a-z] rather than just [a-f], eg
_VALID_URL = [rf'{NhkBaseIE._BASE_URL_REGEX}/(?P<type>video)/(?P<id>[0-9a-z]+)',
rf'{NhkBaseIE._BASE_URL_REGEX}/(?P<type>audio)/(?P<id>[^/?#]+?-\d{{8}}-[0-9a-z]+)']
# Content available only for a limited period of time. Visit
# https://www3.nhk.or.jp/nhkworld/en/ondemand/ for working samples.
_TESTS = [{
'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/video/2049126/',
'info_dict': {
'id': 'nw_vod_v_en_2049_126_20230413233000_01_1681398302',
'ext': 'mp4',
'title': 'Japan Railway Journal - The Tohoku Shinkansen: Full Speed Ahead',
'description': 'md5:49f7c5b206e03868a2fdf0d0814b92f6',
'thumbnail': 'md5:51bcef4a21936e7fea1ff4e06353f463',
'episode': 'The Tohoku Shinkansen: Full Speed Ahead',
'series': 'Japan Railway Journal',
},
}, {
# video clip
'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/video/9999011/',
'md5': '153c3016dfd252ba09726588149cf0e7',
'info_dict': {
'id': 'lpZXIwaDE6_Z-976CPsFdxyICyWUzlT5',
'ext': 'mp4',
'title': 'Dining with the Chef - Chef Saito\'s Family recipe: MENCHI-KATSU',
'description': 'md5:5aee4a9f9d81c26281862382103b0ea5',
'thumbnail': 'md5:d6a4d9b6e9be90aaadda0bcce89631ed',
'series': 'Dining with the Chef',
'episode': 'Chef Saito\'s Family recipe: MENCHI-KATSU',
},
}, {
# radio
'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/audio/livinginjapan-20231001-1/',
'info_dict': {
'id': 'livinginjapan-20231001-1-en',
'ext': 'm4a',
'title': 'Living in Japan - Tips for Travelers to Japan / Ramen Vending Machines',
'series': 'Living in Japan',
'description': 'md5:850611969932874b4a3309e0cae06c2f',
'thumbnail': 'md5:960622fb6e06054a4a1a0c97ea752545',
'episode': 'Tips for Travelers to Japan / Ramen Vending Machines'
},
}, {
'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/video/2015173/',
'only_matching': True,
}, {
'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/audio/plugin-20190404-1/',
'only_matching': True,
}, {
'url': 'https://www3.nhk.or.jp/nhkworld/fr/ondemand/audio/plugin-20190404-1/',
'only_matching': True,
}, {
'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/audio/j_art-20150903-1/',
'only_matching': True,
}, {
# video, alphabetic character in ID #29670
'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/video/9999a34/',
'info_dict': {
'id': 'qfjay6cg',
'ext': 'mp4',
'title': 'DESIGN TALKS plus - Fishermens Finery',
'description': 'md5:8a8f958aaafb0d7cb59d38de53f1e448',
'thumbnail': r're:^https?:/(/[a-z0-9.-]+)+\.jpg\?w=1920&h=1080$',
'upload_date': '20210615',
'timestamp': 1623722008,
},
'skip': '404 Not Found',
}, {
# japanese-language, longer id than english
'url': 'https://www3.nhk.or.jp/nhkworld/ja/ondemand/video/0020271111/',
'info_dict': {
'id': 'nw_ja_v_jvod_ohayou_20231008',
'ext': 'mp4',
'title': 'おはよう日本7時台 - 10月8日放送',
'series': 'おはよう日本7時台',
'episode': '10月8日放送',
'thumbnail': 'md5:d733b1c8e965ab68fb02b2d347d0e9b4',
'description': 'md5:9c1d6cbeadb827b955b20e99ab920ff0',
},
'skip': 'expires 2023-10-15',
}]
def _real_extract(self, url):
return self._extract_episode_info(url)
class NhkVodProgramIE(NhkBaseIE):
_VALID_URL = rf'{NhkBaseIE._BASE_URL_REGEX}/program{NhkBaseIE._TYPE_REGEX}(?P<id>\w+)(?:.+?\btype=(?P<episode_type>clip|(?:radio|tv)Episode))?'
_TESTS = [{
# video program episodes
'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/program/video/sumo',
'info_dict': {
'id': 'sumo',
'title': 'GRAND SUMO Highlights',
},
'playlist_mincount': 12,
}, {
'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/program/video/japanrailway',
'info_dict': {
'id': 'japanrailway',
'title': 'Japan Railway Journal',
},
'playlist_mincount': 12,
}, {
# video program clips
'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/program/video/japanrailway/?type=clip',
'info_dict': {
'id': 'japanrailway',
'title': 'Japan Railway Journal',
},
'playlist_mincount': 5,
}, {
'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/program/video/10yearshayaomiyazaki/',
'only_matching': True,
}, {
# audio program
'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/program/audio/listener/',
'only_matching': True,
}]
def _real_extract(self, url):
lang, m_type, program_id, episode_type = self._match_valid_url(url).group('lang', 'type', 'id', 'episode_type')
episodes = self._call_api(
program_id, lang, m_type == 'video', False, episode_type == 'clip')
entries = []
for episode in episodes:
episode_path = episode.get('url')
if not episode_path:
continue
entries.append(self._extract_episode_info(
urljoin(url, episode_path), episode))
program_title = None
if entries:
program_title = entries[0].get('series')
return self.playlist_result(entries, program_id, program_title)
class NhkForSchoolBangumiIE(InfoExtractor):
_VALID_URL = r'https?://www2\.nhk\.or\.jp/school/movie/(?P<type>bangumi|clip)\.cgi\?das_id=(?P<id>[a-zA-Z0-9_-]+)'
_TESTS = [{
'url': 'https://www2.nhk.or.jp/school/movie/bangumi.cgi?das_id=D0005150191_00000',
'info_dict': {
'id': 'D0005150191_00003',
'title': 'にている かな',
'duration': 599.999,
'timestamp': 1396414800,
'upload_date': '20140402',
'ext': 'mp4',
'chapters': 'count:12'
},
'params': {
# m3u8 download
'skip_download': True,
},
}]
def _real_extract(self, url):
program_type, video_id = self._match_valid_url(url).groups()
webpage = self._download_webpage(
f'https://www2.nhk.or.jp/school/movie/{program_type}.cgi?das_id={video_id}', video_id)
# searches all variables
base_values = {g.group(1): g.group(2) for g in re.finditer(r'var\s+([a-zA-Z_]+)\s*=\s*"([^"]+?)";', webpage)}
# and programObj values too
program_values = {g.group(1): g.group(3) for g in re.finditer(r'(?:program|clip)Obj\.([a-zA-Z_]+)\s*=\s*(["\'])([^"]+?)\2;', webpage)}
# extract all chapters
chapter_durations = [parse_duration(g.group(1)) for g in re.finditer(r'chapterTime\.push\(\'([0-9:]+?)\'\);', webpage)]
chapter_titles = [' '.join([g.group(1) or '', unescapeHTML(g.group(2))]).strip() for g in re.finditer(r'<div class="cpTitle"><span>(scene\s*\d+)?</span>([^<]+?)</div>', webpage)]
# this is how player_core.js is actually doing (!)
version = base_values.get('r_version') or program_values.get('version')
if version:
video_id = f'{video_id.split("_")[0]}_{version}'
formats = self._extract_m3u8_formats(
f'https://nhks-vh.akamaihd.net/i/das/{video_id[0:8]}/{video_id}_V_000.f4v/master.m3u8',
video_id, ext='mp4', m3u8_id='hls')
duration = parse_duration(base_values.get('r_duration'))
chapters = None
if chapter_durations and chapter_titles and len(chapter_durations) == len(chapter_titles):
start_time = chapter_durations
end_time = chapter_durations[1:] + [duration]
chapters = [{
'start_time': s,
'end_time': e,
'title': t,
} for s, e, t in zip(start_time, end_time, chapter_titles)]
return {
'id': video_id,
'title': program_values.get('name'),
'duration': parse_duration(base_values.get('r_duration')),
'timestamp': unified_timestamp(base_values['r_upload']),
'formats': formats,
'chapters': chapters,
}
class NhkForSchoolSubjectIE(InfoExtractor):
IE_DESC = 'Portal page for each school subjects, like Japanese (kokugo, 国語) or math (sansuu/suugaku or 算数・数学)'
KNOWN_SUBJECTS = (
'rika', 'syakai', 'kokugo',
'sansuu', 'seikatsu', 'doutoku',
'ongaku', 'taiiku', 'zukou',
'gijutsu', 'katei', 'sougou',
'eigo', 'tokkatsu',
'tokushi', 'sonota',
)
_VALID_URL = r'https?://www\.nhk\.or\.jp/school/(?P<id>%s)/?(?:[\?#].*)?$' % '|'.join(re.escape(s) for s in KNOWN_SUBJECTS)
_TESTS = [{
'url': 'https://www.nhk.or.jp/school/sougou/',
'info_dict': {
'id': 'sougou',
'title': '総合的な学習の時間',
},
'playlist_mincount': 16,
}, {
'url': 'https://www.nhk.or.jp/school/rika/',
'info_dict': {
'id': 'rika',
'title': '理科',
},
'playlist_mincount': 15,
}]
def _real_extract(self, url):
subject_id = self._match_id(url)
webpage = self._download_webpage(url, subject_id)
return self.playlist_from_matches(
re.finditer(rf'href="((?:https?://www\.nhk\.or\.jp)?/school/{re.escape(subject_id)}/[^/]+/)"', webpage),
subject_id,
self._html_search_regex(r'(?s)<span\s+class="subjectName">\s*<img\s*[^<]+>\s*([^<]+?)</span>', webpage, 'title', fatal=False),
lambda g: urljoin(url, g.group(1)))
class NhkForSchoolProgramListIE(InfoExtractor):
_VALID_URL = r'https?://www\.nhk\.or\.jp/school/(?P<id>(?:%s)/[a-zA-Z0-9_-]+)' % (
'|'.join(re.escape(s) for s in NhkForSchoolSubjectIE.KNOWN_SUBJECTS)
)
_TESTS = [{
'url': 'https://www.nhk.or.jp/school/sougou/q/',
'info_dict': {
'id': 'sougou/q',
'title': 'Q~こどものための哲学',
},
'playlist_mincount': 20,
}]
def _real_extract(self, url):
program_id = self._match_id(url)
webpage = self._download_webpage(f'https://www.nhk.or.jp/school/{program_id}/', program_id)
title = (self._generic_title('', webpage)
or self._html_search_regex(r'<h3>([^<]+?)とは?\s*</h3>', webpage, 'title', fatal=False))
title = re.sub(r'\s*\|\s*NHK\s+for\s+School\s*$', '', title) if title else None
description = self._html_search_regex(
r'(?s)<div\s+class="programDetail\s*">\s*<p>[^<]+</p>',
webpage, 'description', fatal=False, group=0)
bangumi_list = self._download_json(
f'https://www.nhk.or.jp/school/{program_id}/meta/program.json', program_id)
# they're always bangumi
bangumis = [
self.url_result(f'https://www2.nhk.or.jp/school/movie/bangumi.cgi?das_id={x}')
for x in traverse_obj(bangumi_list, ('part', ..., 'part-video-dasid')) or []]
return self.playlist_result(bangumis, program_id, title, description)
class NhkRadiruIE(InfoExtractor):
_GEO_COUNTRIES = ['JP']
IE_DESC = 'NHK らじる (Radiru/Rajiru)'
_VALID_URL = r'https?://www\.nhk\.or\.jp/radio/(?:player/ondemand|ondemand/detail)\.html\?p=(?P<site>[\da-zA-Z]+)_(?P<corner>[\da-zA-Z]+)(?:_(?P<headline>[\da-zA-Z]+))?'
_TESTS = [{
'url': 'https://www.nhk.or.jp/radio/player/ondemand.html?p=0449_01_3853544',
'skip': 'Episode expired on 2023-04-16',
'info_dict': {
'channel': 'NHK-FM',
'description': 'md5:94b08bdeadde81a97df4ec882acce3e9',
'ext': 'm4a',
'id': '0449_01_3853544',
'series': 'ジャズ・トゥナイト',
'thumbnail': 'https://www.nhk.or.jp/prog/img/449/g449.jpg',
'timestamp': 1680969600,
'title': 'ジャズ・トゥナイト NEWジャズ特集',
'upload_date': '20230408',
'release_timestamp': 1680962400,
'release_date': '20230408',
'was_live': True,
},
}, {
# playlist, airs every weekday so it should _hopefully_ be okay forever
'url': 'https://www.nhk.or.jp/radio/ondemand/detail.html?p=0458_01',
'info_dict': {
'id': '0458_01',
'title': 'ベストオブクラシック',
'description': '世界中の上質な演奏会をじっくり堪能する本格派クラシック番組。',
'channel': 'NHK-FM',
'thumbnail': 'https://www.nhk.or.jp/prog/img/458/g458.jpg',
},
'playlist_mincount': 3,
}, {
# one with letters in the id
'url': 'https://www.nhk.or.jp/radio/player/ondemand.html?p=F300_06_3738470',
'note': 'Expires on 2024-03-31',
'info_dict': {
'id': 'F300_06_3738470',
'ext': 'm4a',
'title': '有島武郎「一房のぶどう」',
'description': '朗読:川野一宇(ラジオ深夜便アンカー)\r\n\r\n2016年12月8日放送「ラジオ深夜便『アンカー朗読シリーズ』」より',
'channel': 'NHKラジオ第1、NHK-FM',
'timestamp': 1635757200,
'thumbnail': 'https://www.nhk.or.jp/radioondemand/json/F300/img/corner/box_109_thumbnail.jpg',
'release_date': '20161207',
'series': 'らじる文庫 by ラジオ深夜便 ',
'release_timestamp': 1481126700,
'upload_date': '20211101',
}
}, {
# news
'url': 'https://www.nhk.or.jp/radio/player/ondemand.html?p=F261_01_3855109',
'skip': 'Expires on 2023-04-17',
'info_dict': {
'id': 'F261_01_3855109',
'ext': 'm4a',
'channel': 'NHKラジオ第1',
'timestamp': 1681635900,
'release_date': '20230416',
'series': 'NHKラジオニュース',
'title': '午後時のNHKニュース',
'thumbnail': 'https://www.nhk.or.jp/radioondemand/json/F261/img/RADIONEWS_640.jpg',
'upload_date': '20230416',
'release_timestamp': 1681635600,
},
}]
def _extract_episode_info(self, headline, programme_id, series_meta):
episode_id = f'{programme_id}_{headline["headline_id"]}'
episode = traverse_obj(headline, ('file_list', 0, {dict}))
return {
**series_meta,
'id': episode_id,
'formats': self._extract_m3u8_formats(episode.get('file_name'), episode_id, fatal=False),
'container': 'm4a_dash', # force fixup, AAC-only HLS
'was_live': True,
'series': series_meta.get('title'),
'thumbnail': url_or_none(headline.get('headline_image')) or series_meta.get('thumbnail'),
**traverse_obj(episode, {
'title': 'file_title',
'description': 'file_title_sub',
'timestamp': ('open_time', {unified_timestamp}),
'release_timestamp': ('aa_vinfo4', {lambda x: x.split('_')[0]}, {unified_timestamp}),
}),
}
def _real_extract(self, url):
site_id, corner_id, headline_id = self._match_valid_url(url).group('site', 'corner', 'headline')
programme_id = f'{site_id}_{corner_id}'
if site_id == 'F261':
json_url = 'https://www.nhk.or.jp/s-media/news/news-site/list/v1/all.json'
else:
json_url = f'https://www.nhk.or.jp/radioondemand/json/{site_id}/bangumi_{programme_id}.json'
meta = self._download_json(json_url, programme_id)['main']
series_meta = traverse_obj(meta, {
'title': 'program_name',
'channel': 'media_name',
'thumbnail': (('thumbnail_c', 'thumbnail_p'), {url_or_none}),
}, get_all=False)
if headline_id:
return self._extract_episode_info(
traverse_obj(meta, (
'detail_list', lambda _, v: v['headline_id'] == headline_id), get_all=False),
programme_id, series_meta)
def entries():
for headline in traverse_obj(meta, ('detail_list', ..., {dict})):
yield self._extract_episode_info(headline, programme_id, series_meta)
return self.playlist_result(
entries(), programme_id, playlist_description=meta.get('site_detail'), **series_meta)
class NhkRadioNewsPageIE(InfoExtractor):
_VALID_URL = r'https?://www\.nhk\.or\.jp/radionews/?(?:$|[?#])'
_TESTS = [{
# airs daily, on-the-hour most hours
'url': 'https://www.nhk.or.jp/radionews/',
'playlist_mincount': 5,
'info_dict': {
'id': 'F261_01',
'thumbnail': 'https://www.nhk.or.jp/radioondemand/json/F261/img/RADIONEWS_640.jpg',
'description': 'md5:bf2c5b397e44bc7eb26de98d8f15d79d',
'channel': 'NHKラジオ第1',
'title': 'NHKラジオニュース',
}
}]
def _real_extract(self, url):
return self.url_result('https://www.nhk.or.jp/radio/ondemand/detail.html?p=F261_01', NhkRadiruIE)
class NhkRadiruLiveIE(InfoExtractor):
_GEO_COUNTRIES = ['JP']
_VALID_URL = r'https?://www\.nhk\.or\.jp/radio/player/\?ch=(?P<id>r[12]|fm)'
_TESTS = [{
# radio 1, no area specified
'url': 'https://www.nhk.or.jp/radio/player/?ch=r1',
'info_dict': {
'id': 'r1-tokyo',
'title': 're:^ネットラジオ第1 東京.+$',
'ext': 'm4a',
'thumbnail': 'https://www.nhk.or.jp/common/img/media/r1-200x200.png',
'live_status': 'is_live',
},
}, {
# radio 2, area specified
# (the area doesnt actually matter, r2 is national)
'url': 'https://www.nhk.or.jp/radio/player/?ch=r2',
'params': {'extractor_args': {'nhkradirulive': {'area': ['fukuoka']}}},
'info_dict': {
'id': 'r2-fukuoka',
'title': 're:^ネットラジオ第2 福岡.+$',
'ext': 'm4a',
'thumbnail': 'https://www.nhk.or.jp/common/img/media/r2-200x200.png',
'live_status': 'is_live',
},
}, {
# fm, area specified
'url': 'https://www.nhk.or.jp/radio/player/?ch=fm',
'params': {'extractor_args': {'nhkradirulive': {'area': ['sapporo']}}},
'info_dict': {
'id': 'fm-sapporo',
'title': 're:^NHKネットラジオFM 札幌.+$',
'ext': 'm4a',
'thumbnail': 'https://www.nhk.or.jp/common/img/media/fm-200x200.png',
'live_status': 'is_live',
}
}]
_NOA_STATION_IDS = {'r1': 'n1', 'r2': 'n2', 'fm': 'n3'}
def _real_extract(self, url):
station = self._match_id(url)
area = self._configuration_arg('area', ['tokyo'])[0]
config = self._download_xml(
'https://www.nhk.or.jp/radio/config/config_web.xml', station, 'Downloading area information')
data = config.find(f'.//data//area[.="{area}"]/..')
if not data:
raise ExtractorError('Invalid area. Valid areas are: %s' % ', '.join(
[i.text for i in config.findall('.//data//area')]), expected=True)
noa_info = self._download_json(
f'https:{config.find(".//url_program_noa").text}'.format(area=data.find('areakey').text),
station, note=f'Downloading {area} station metadata')
present_info = traverse_obj(noa_info, ('nowonair_list', self._NOA_STATION_IDS.get(station), 'present'))
return {
'title': ' '.join(traverse_obj(present_info, (('service', 'area',), 'name', {str}))),
'id': join_nonempty(station, area),
'thumbnails': traverse_obj(present_info, ('service', 'images', ..., {
'url': 'url',
'width': ('width', {int_or_none}),
'height': ('height', {int_or_none}),
})),
'formats': self._extract_m3u8_formats(data.find(f'{station}hls').text, station),
'is_live': True,
}