archive
/
yt-dlp
mirror of https://github.com/yt-dlp/yt-dlp
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

[ie/NHKRadiru] Fix extractor (#10106)

Browse Source 
Closes #10105
Authored by: garret1317
pull/10097/head
garret1317 5 months ago committed by GitHub
parent e53e56b735
commit b8e2a5e0e1
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 172 additions and 70 deletions
Show Stats Download Patch File Download Diff File

238
yt_dlp/extractor/nhk.py
Unescape Escape View File

@ -4,6 +4,7 @@ from .common import InfoExtractor
from ..utils import ( from ..utils import (
ExtractorError, ExtractorError,
clean_html, clean_html,
filter_dict,
get_element_by_class, get_element_by_class,
int_or_none, int_or_none,
join_nonempty, join_nonempty,
@ -590,21 +591,22 @@ class NhkRadiruIE(InfoExtractor):
IE_DESC = 'NHK らじる (Radiru/Rajiru)' 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]+))?' _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 = [{ _TESTS = [{
'url': 'https://www.nhk.or.jp/radio/player/ondemand.html?p=0449_01_3926210', 'url': 'https://www.nhk.or.jp/radio/player/ondemand.html?p=0449_01_4003239',
'skip': 'Episode expired on 2024-02-24', 'skip': 'Episode expired on 2024-06-09',
'info_dict': { 'info_dict': {
'title': 'ジャズ・トゥナイト　シリーズＪＡＺＺジャイアンツ　５６　ジョニー・ホッジス', 'title': 'ジャズ・トゥナイト　ジャズ「Ｎｉｇｈｔ　ａｎｄ　Ｄａｙ」特集',
'id': '0449_01_3926210', 'id': '0449_01_4003239',
'ext': 'm4a', 'ext': 'm4a',
'uploader': 'NHK FM 東京',
'description': 'md5:ad05f3c3f3f6e99b2e69f9b5e49551dc',
'series': 'ジャズ・トゥナイト', 'series': 'ジャズ・トゥナイト',
'uploader': 'NHK-FM', 'channel': 'NHK FM 東京',
'channel': 'NHK-FM',
'thumbnail': 'https://www.nhk.or.jp/prog/img/449/g449.jpg', 'thumbnail': 'https://www.nhk.or.jp/prog/img/449/g449.jpg',
'release_date': '20240217', 'upload_date': '20240601',
'description': 'md5:a456ee8e5e59e6dd2a7d32e62386e811', 'series_id': '0449_01',
'timestamp': 1708185600, 'release_date': '20240601',
'release_timestamp': 1708178400, 'timestamp': 1717257600,
'upload_date': '20240217', 'release_timestamp': 1717250400,
}, },
}, { }, {
# playlist, airs every weekday so it should _hopefully_ be okay forever # playlist, airs every weekday so it should _hopefully_ be okay forever
@ -613,71 +615,145 @@ class NhkRadiruIE(InfoExtractor):
'id': '0458_01', 'id': '0458_01',
'title': 'ベストオブクラシック', '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。', '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',
'uploader': 'NHK-FM',
'thumbnail': 'https://www.nhk.or.jp/prog/img/458/g458.jpg', 'thumbnail': 'https://www.nhk.or.jp/prog/img/458/g458.jpg',
'series_id': '0458_01',
'uploader': 'NHK FM',
'channel': 'NHK FM',
'series': 'ベストオブクラシック',
}, },
'playlist_mincount': 3, 'playlist_mincount': 3,
}, { }, {
# one with letters in the id # one with letters in the id
'url': 'https://www.nhk.or.jp/radio/player/ondemand.html?p=F300_06_3738470', 'url': 'https://www.nhk.or.jp/radio/player/ondemand.html?p=F683_01_3910688',
'note': 'Expires on 2024-03-31', 'note': 'Expires on 2025-03-31',
'info_dict': { 'info_dict': {
'id': 'F300_06_3738470', 'id': 'F683_01_3910688',
'ext': 'm4a', 'ext': 'm4a',
'title': '有島武郎「一房のぶどう」', 'title': '夏目漱石「文鳥」第1回',
'description': '朗読：川野一宇（ラジオ深夜便アンカー）\r\n\r\n（2016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）', 'series': '【らじる文庫】夏目漱石「文鳥」（全4回）',
'channel': 'NHKラジオ第1、NHK-FM', 'series_id': 'F683_01',
'uploader': 'NHKラジオ第1、NHK-FM', 'description': '朗読：浅井理アナウンサー',
'timestamp': 1635757200, 'thumbnail': 'https://www.nhk.or.jp/radioondemand/json/F683/img/roudoku_05_rod_640.jpg',
'thumbnail': 'https://www.nhk.or.jp/radioondemand/json/F300/img/corner/box_109_thumbnail.jpg', 'upload_date': '20240106',
'release_date': '20161207', 'release_date': '20240106',
'series': 'らじる文庫 by ラジオ深夜便 ', 'uploader': 'NHK R1',
'release_timestamp': 1481126700, 'release_timestamp': 1704511800,
'upload_date': '20211101', 'channel': 'NHK R1',
'timestamp': 1704512700,
}, },
'expected_warnings': ['Unable to download JSON metadata', 'Failed to get extended description'], 'expected_warnings': ['Unable to download JSON metadata',
'Failed to get extended metadata. API returned Error 1: Invalid parameters'],
}, { }, {
# news # news
'url': 'https://www.nhk.or.jp/radio/player/ondemand.html?p=F261_01_3855109', 'url': 'https://www.nhk.or.jp/radio/player/ondemand.html?p=F261_01_4012173',
'skip': 'Expires on 2023-04-17',
'info_dict': { 'info_dict': {
'id': 'F261_01_3855109', 'id': 'F261_01_4012173',
'ext': 'm4a', 'ext': 'm4a',
'channel': 'NHKラジオ第1', 'channel': 'NHKラジオ第1',
'uploader': 'NHKラジオ第1', 'uploader': 'NHKラジオ第1',
'timestamp': 1681635900,
'release_date': '20230416',
'series': 'NHKラジオニュース', 'series': 'NHKラジオニュース',
'title': '午後６時のNHKニュース', 'title': '午前０時のNHKニュース',
'thumbnail': 'https://www.nhk.or.jp/radioondemand/json/F261/img/RADIONEWS_640.jpg', 'thumbnail': 'https://www.nhk.or.jp/radioondemand/json/F261/img/RADIONEWS_640.jpg',
'upload_date': '20230416', 'release_timestamp': 1718290800,
'release_timestamp': 1681635600, 'release_date': '20240613',
'timestamp': 1718291400,
'upload_date': '20240613',
}, },
}, {
# fallback when extended metadata fails
'url': 'https://www.nhk.or.jp/radio/player/ondemand.html?p=2834_01_4009298',
'skip': 'Expires on 2024-06-07',
'info_dict': {
'id': '2834_01_4009298',
'title': 'まち☆キラ！開成町特集',
'ext': 'm4a',
'release_date': '20240531',
'upload_date': '20240531',
'series': 'はま☆キラ！',
'thumbnail': 'https://www.nhk.or.jp/prog/img/2834/g2834.jpg',
'channel': 'NHK R1,FM',
'description': '',
'timestamp': 1717123800,
'uploader': 'NHK R1,FM',
'release_timestamp': 1717120800,
'series_id': '2834_01',
},
'expected_warnings': ['Failed to get extended metadata. API returned empty list.'],
}] }]
_API_URL_TMPL = None _API_URL_TMPL = None
def _extract_extended_description(self, episode_id, episode): def _extract_extended_metadata(self, episode_id, aa_vinfo):
service, _, area = traverse_obj(episode, ('aa_vinfo2', {str}, {lambda x: (x or '').partition(',')})) service, _, area = traverse_obj(aa_vinfo, (2, {str}, {lambda x: (x or '').partition(',')}))
aa_vinfo3 = traverse_obj(episode, ('aa_vinfo3', {str}))
detail_url = try_call( detail_url = try_call(
lambda: self._API_URL_TMPL.format(service=service, area=area, dateid=aa_vinfo3)) lambda: self._API_URL_TMPL.format(area=area, service=service, dateid=aa_vinfo[3]))
if not detail_url: if not detail_url:
return return {}
response = self._download_json(
detail_url, episode_id, 'Downloading extended metadata',
'Failed to download extended metadata', fatal=False, expected_status=400)
if not response:
return {}
if error := traverse_obj(response, ('error', {dict})):
self.report_warning(
'Failed to get extended metadata. API returned '
f'Error {join_nonempty("code", "message", from_dict=error, delim=": ")}')
return {}
full_meta = traverse_obj(response, ('list', service, 0, {dict}))
if not full_meta:
self.report_warning('Failed to get extended metadata. API returned empty list.')
return {}
station = ' '.join(traverse_obj(full_meta, (('service', 'area'), 'name', {str}))) or None
thumbnails = [{
'id': str(id_),
'preference': 1 if id_.startswith('thumbnail') else -2 if id_.startswith('logo') else -1,
**traverse_obj(thumb, {
'url': 'url',
'width': ('width', {int_or_none}),
'height': ('height', {int_or_none}),
}),
} for id_, thumb in traverse_obj(full_meta, ('images', {dict.items}, lambda _, v: v[1]['url']))]
full_meta = traverse_obj( return filter_dict({
self._download_json(detail_url, episode_id, 'Downloading extended metadata', fatal=False), 'channel': station,
('list', service, 0, {dict})) or {} 'uploader': station,
return join_nonempty('subtitle', 'content', 'act', 'music', delim='\n\n', from_dict=full_meta) 'description': join_nonempty(
'subtitle', 'content', 'act', 'music', delim='\n\n', from_dict=full_meta),
'thumbnails': thumbnails,
**traverse_obj(full_meta, {
'title': ('title', {str}),
'timestamp': ('end_time', {unified_timestamp}),
'release_timestamp': ('start_time', {unified_timestamp}),
}),
})
def _extract_episode_info(self, episode, programme_id, series_meta):
episode_id = f'{programme_id}_{episode["id"]}'
aa_vinfo = traverse_obj(episode, ('aa_contents_id', {lambda x: x.split(';')}))
extended_metadata = self._extract_extended_metadata(episode_id, aa_vinfo)
fallback_start_time, _, fallback_end_time = traverse_obj(
aa_vinfo, (4, {str}, {lambda x: (x or '').partition('_')}))
return {
**series_meta,
'id': episode_id,
'formats': self._extract_m3u8_formats(episode.get('stream_url'), episode_id, fatal=False),
'container': 'm4a_dash', # force fixup, AAC-only HLS
'was_live': True,
'title': episode.get('program_title'),
'description': episode.get('program_sub_title'), # fallback
'timestamp': unified_timestamp(fallback_end_time),
'release_timestamp': unified_timestamp(fallback_start_time),
**extended_metadata,
}
def _extract_episode_info(self, headline, programme_id, series_meta): def _extract_news_info(self, headline, programme_id, series_meta):
episode_id = f'{programme_id}_{headline["headline_id"]}' episode_id = f'{programme_id}_{headline["headline_id"]}'
episode = traverse_obj(headline, ('file_list', 0, {dict})) episode = traverse_obj(headline, ('file_list', 0, {dict}))
description = self._extract_extended_description(episode_id, episode)
if not description:
self.report_warning('Failed to get extended description, falling back to summary')
description = traverse_obj(episode, ('file_title_sub', {str}))
return { return {
**series_meta, **series_meta,
@ -687,9 +763,9 @@ class NhkRadiruIE(InfoExtractor):
'was_live': True, 'was_live': True,
'series': series_meta.get('title'), 'series': series_meta.get('title'),
'thumbnail': url_or_none(headline.get('headline_image')) or series_meta.get('thumbnail'), 'thumbnail': url_or_none(headline.get('headline_image')) or series_meta.get('thumbnail'),
'description': description,
**traverse_obj(episode, { **traverse_obj(episode, {
'title': 'file_title', 'title': ('file_title', {str}),
'description': ('file_title_sub', {str}),
'timestamp': ('open_time', {unified_timestamp}), 'timestamp': ('open_time', {unified_timestamp}),
'release_timestamp': ('aa_vinfo4', {lambda x: x.split('_')[0]}, {unified_timestamp}), 'release_timestamp': ('aa_vinfo4', {lambda x: x.split('_')[0]}, {unified_timestamp}),
}), }),
@ -706,32 +782,58 @@ class NhkRadiruIE(InfoExtractor):
site_id, corner_id, headline_id = self._match_valid_url(url).group('site', 'corner', 'headline') site_id, corner_id, headline_id = self._match_valid_url(url).group('site', 'corner', 'headline')
programme_id = f'{site_id}_{corner_id}' programme_id = f'{site_id}_{corner_id}'
if site_id == 'F261': if site_id == 'F261': # XXX: News programmes use old API (for now?)
json_url = 'https://www.nhk.or.jp/s-media/news/news-site/list/v1/all.json' meta = self._download_json(
else: 'https://www.nhk.or.jp/s-media/news/news-site/list/v1/all.json', programme_id)['main']
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, { series_meta = traverse_obj(meta, {
'title': 'program_name', 'title': ('program_name', {str}),
'channel': 'media_name', 'channel': ('media_name', {str}),
'uploader': 'media_name', 'uploader': ('media_name', {str}),
'thumbnail': (('thumbnail_c', 'thumbnail_p'), {url_or_none}), 'thumbnail': (('thumbnail_c', 'thumbnail_p'), {url_or_none}),
}, get_all=False) }, get_all=False)
if headline_id: if headline_id:
return self._extract_episode_info( headline = traverse_obj(
traverse_obj(meta, ( meta, ('detail_list', lambda _, v: v['headline_id'] == headline_id, any))
'detail_list', lambda _, v: v['headline_id'] == headline_id), get_all=False), if not headline:
programme_id, series_meta) raise ExtractorError('Content not found; it has most likely expired', expected=True)
return self._extract_news_info(headline, programme_id, series_meta)
def entries(): def news_entries():
for headline in traverse_obj(meta, ('detail_list', ..., {dict})): for headline in traverse_obj(meta, ('detail_list', ..., {dict})):
yield self._extract_episode_info(headline, programme_id, series_meta) yield self._extract_news_info(headline, programme_id, series_meta)
return self.playlist_result(
news_entries(), programme_id, description=meta.get('site_detail'), **series_meta)
meta = self._download_json(
'https://www.nhk.or.jp/radio-api/app/v1/web/ondemand/series', programme_id, query={
'site_id': site_id,
'corner_site_id': corner_id,
})
fallback_station = join_nonempty('NHK', traverse_obj(meta, ('radio_broadcast', {str})), delim=' ')
series_meta = {
'series': join_nonempty('title', 'corner_name', delim=' ', from_dict=meta),
'series_id': programme_id,
'thumbnail': traverse_obj(meta, ('thumbnail_url', {url_or_none})),
'channel': fallback_station,
'uploader': fallback_station,
}
if headline_id:
episode = traverse_obj(meta, ('episodes', lambda _, v: v['id'] == int(headline_id), any))
if not episode:
raise ExtractorError('Content not found; it has most likely expired', expected=True)
return self._extract_episode_info(episode, programme_id, series_meta)
def entries():
for episode in traverse_obj(meta, ('episodes', ..., {dict})):
yield self._extract_episode_info(episode, programme_id, series_meta)
return self.playlist_result( return self.playlist_result(
entries(), programme_id, playlist_description=meta.get('site_detail'), **series_meta) entries(), programme_id, title=series_meta.get('series'),
description=meta.get('series_description'), **series_meta)
class NhkRadioNewsPageIE(InfoExtractor): class NhkRadioNewsPageIE(InfoExtractor):

Write Preview
Loading…
Cancel
Save