yt_dlp/extractor/nhk.py

   1 import re
   2
   3 from .common import InfoExtractor
   4 from ..utils import (
   5     parse_duration,
   6     traverse_obj,
   7     unescapeHTML,
   8     unified_timestamp,
   9     urljoin
  10 )
  11
  12
  13 class NhkBaseIE(InfoExtractor):
  14     _API_URL_TEMPLATE = 'https://nwapi.nhk.jp/nhkworld/%sod%slist/v7b/%s/%s/%s/all%s.json'
  15     _BASE_URL_REGEX = r'https?://www3\.nhk\.or\.jp/nhkworld/(?P<lang>[a-z]{2})/ondemand'
  16     _TYPE_REGEX = r'/(?P<type>video|audio)/'
  17
  18     def _call_api(self, m_id, lang, is_video, is_episode, is_clip):
  19         return self._download_json(
  20             self._API_URL_TEMPLATE % (
  21                 'v' if is_video else 'r',
  22                 'clip' if is_clip else 'esd',
  23                 'episode' if is_episode else 'program',
  24                 m_id, lang, '/all' if is_video else ''),
  25             m_id, query={'apikey': 'EJfK8jdS57GqlupFgAfAAwr573q01y6k'})['data']['episodes'] or []
  26
  27     def _extract_episode_info(self, url, episode=None):
  28         fetch_episode = episode is None
  29         lang, m_type, episode_id = NhkVodIE._match_valid_url(url).groups()
  30         if len(episode_id) == 7:
  31             episode_id = episode_id[:4] + '-' + episode_id[4:]
  32
  33         is_video = m_type == 'video'
  34         if fetch_episode:
  35             episode = self._call_api(
  36                 episode_id, lang, is_video, True, episode_id[:4] == '9999')[0]
  37         title = episode.get('sub_title_clean') or episode['sub_title']
  38
  39         def get_clean_field(key):
  40             return episode.get(key + '_clean') or episode.get(key)
  41
  42         series = get_clean_field('title')
  43
  44         thumbnails = []
  45         for s, w, h in [('', 640, 360), ('_l', 1280, 720)]:
  46             img_path = episode.get('image' + s)
  47             if not img_path:
  48                 continue
  49             thumbnails.append({
  50                 'id': '%dp' % h,
  51                 'height': h,
  52                 'width': w,
  53                 'url': 'https://www3.nhk.or.jp' + img_path,
  54             })
  55
  56         info = {
  57             'id': episode_id + '-' + lang,
  58             'title': '%s - %s' % (series, title) if series and title else title,
  59             'description': get_clean_field('description'),
  60             'thumbnails': thumbnails,
  61             'series': series,
  62             'episode': title,
  63         }
  64         if is_video:
  65             vod_id = episode['vod_id']
  66             info.update({
  67                 '_type': 'url_transparent',
  68                 'ie_key': 'Piksel',
  69                 'url': 'https://player.piksel.com/v/refid/nhkworld/prefid/' + vod_id,
  70                 'id': vod_id,
  71             })
  72         else:
  73             if fetch_episode:
  74                 audio_path = episode['audio']['audio']
  75                 info['formats'] = self._extract_m3u8_formats(
  76                     'https://nhkworld-vh.akamaihd.net/i%s/master.m3u8' % audio_path,
  77                     episode_id, 'm4a', entry_protocol='m3u8_native',
  78                     m3u8_id='hls', fatal=False)
  79                 for f in info['formats']:
  80                     f['language'] = lang
  81                 self._sort_formats(info['formats'])
  82             else:
  83                 info.update({
  84                     '_type': 'url_transparent',
  85                     'ie_key': NhkVodIE.ie_key(),
  86                     'url': url,
  87                 })
  88         return info
  89
  90
  91 class NhkVodIE(NhkBaseIE):
  92     # the 7-character IDs can have alphabetic chars too: assume [a-z] rather than just [a-f], eg
  93     _VALID_URL = r'%s%s(?P<id>[0-9a-z]{7}|[^/]+?-\d{8}-[0-9a-z]+)' % (NhkBaseIE._BASE_URL_REGEX, NhkBaseIE._TYPE_REGEX)
  94     # Content available only for a limited period of time. Visit
  95     # https://www3.nhk.or.jp/nhkworld/en/ondemand/ for working samples.
  96     _TESTS = [{
  97         # video clip
  98         'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/video/9999011/',
  99         'md5': '7a90abcfe610ec22a6bfe15bd46b30ca',
 100         'info_dict': {
 101             'id': 'a95j5iza',
 102             'ext': 'mp4',
 103             'title': "Dining with the Chef - Chef Saito's Family recipe: MENCHI-KATSU",
 104             'description': 'md5:5aee4a9f9d81c26281862382103b0ea5',
 105             'timestamp': 1565965194,
 106             'upload_date': '20190816',
 107         },
 108     }, {
 109         # audio clip
 110         'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/audio/r_inventions-20201104-1/',
 111         'info_dict': {
 112             'id': 'r_inventions-20201104-1-en',
 113             'ext': 'm4a',
 114             'title': "Japan's Top Inventions - Miniature Video Cameras",
 115             'description': 'md5:07ea722bdbbb4936fdd360b6a480c25b',
 116         },
 117         'params': {
 118             # m3u8 download
 119             'skip_download': True,
 120         },
 121     }, {
 122         'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/video/2015173/',
 123         'only_matching': True,
 124     }, {
 125         'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/audio/plugin-20190404-1/',
 126         'only_matching': True,
 127     }, {
 128         'url': 'https://www3.nhk.or.jp/nhkworld/fr/ondemand/audio/plugin-20190404-1/',
 129         'only_matching': True,
 130     }, {
 131         'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/audio/j_art-20150903-1/',
 132         'only_matching': True,
 133     }, {
 134         # video, alphabetic character in ID #29670
 135         'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/video/9999a34/',
 136         'only_matching': True,
 137         'info_dict': {
 138             'id': 'qfjay6cg',
 139             'ext': 'mp4',
 140             'title': 'DESIGN TALKS plus - Fishermen’s Finery',
 141             'description': 'md5:8a8f958aaafb0d7cb59d38de53f1e448',
 142             'thumbnail': r're:^https?:/(/[a-z0-9.-]+)+\.jpg\?w=1920&h=1080$',
 143             'upload_date': '20210615',
 144             'timestamp': 1623722008,
 145         }
 146     }]
 147
 148     def _real_extract(self, url):
 149         return self._extract_episode_info(url)
 150
 151
 152 class NhkVodProgramIE(NhkBaseIE):
 153     _VALID_URL = r'%s/program%s(?P<id>[0-9a-z]+)(?:.+?\btype=(?P<episode_type>clip|(?:radio|tv)Episode))?' % (NhkBaseIE._BASE_URL_REGEX, NhkBaseIE._TYPE_REGEX)
 154     _TESTS = [{
 155         # video program episodes
 156         'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/program/video/japanrailway',
 157         'info_dict': {
 158             'id': 'japanrailway',
 159             'title': 'Japan Railway Journal',
 160         },
 161         'playlist_mincount': 1,
 162     }, {
 163         # video program clips
 164         'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/program/video/japanrailway/?type=clip',
 165         'info_dict': {
 166             'id': 'japanrailway',
 167             'title': 'Japan Railway Journal',
 168         },
 169         'playlist_mincount': 5,
 170     }, {
 171         'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/program/video/10yearshayaomiyazaki/',
 172         'only_matching': True,
 173     }, {
 174         # audio program
 175         'url': 'https://www3.nhk.or.jp/nhkworld/en/ondemand/program/audio/listener/',
 176         'only_matching': True,
 177     }]
 178
 179     def _real_extract(self, url):
 180         lang, m_type, program_id, episode_type = self._match_valid_url(url).groups()
 181
 182         episodes = self._call_api(
 183             program_id, lang, m_type == 'video', False, episode_type == 'clip')
 184
 185         entries = []
 186         for episode in episodes:
 187             episode_path = episode.get('url')
 188             if not episode_path:
 189                 continue
 190             entries.append(self._extract_episode_info(
 191                 urljoin(url, episode_path), episode))
 192
 193         program_title = None
 194         if entries:
 195             program_title = entries[0].get('series')
 196
 197         return self.playlist_result(entries, program_id, program_title)
 198
 199
 200 class NhkForSchoolBangumiIE(InfoExtractor):
 201     _VALID_URL = r'https?://www2\.nhk\.or\.jp/school/movie/(?P<type>bangumi|clip)\.cgi\?das_id=(?P<id>[a-zA-Z0-9_-]+)'
 202     _TESTS = [{
 203         'url': 'https://www2.nhk.or.jp/school/movie/bangumi.cgi?das_id=D0005150191_00000',
 204         'info_dict': {
 205             'id': 'D0005150191_00003',
 206             'title': 'にている かな',
 207             'duration': 599.999,
 208             'timestamp': 1396414800,
 209
 210             'upload_date': '20140402',
 211             'ext': 'mp4',
 212
 213             'chapters': 'count:12'
 214         },
 215         'params': {
 216             # m3u8 download
 217             'skip_download': True,
 218         },
 219     }]
 220
 221     def _real_extract(self, url):
 222         program_type, video_id = self._match_valid_url(url).groups()
 223
 224         webpage = self._download_webpage(
 225             f'https://www2.nhk.or.jp/school/movie/{program_type}.cgi?das_id={video_id}', video_id)
 226
 227         # searches all variables
 228         base_values = {g.group(1): g.group(2) for g in re.finditer(r'var\s+([a-zA-Z_]+)\s*=\s*"([^"]+?)";', webpage)}
 229         # and programObj values too
 230         program_values = {g.group(1): g.group(3) for g in re.finditer(r'(?:program|clip)Obj\.([a-zA-Z_]+)\s*=\s*(["\'])([^"]+?)\2;', webpage)}
 231         # extract all chapters
 232         chapter_durations = [parse_duration(g.group(1)) for g in re.finditer(r'chapterTime\.push\(\'([0-9:]+?)\'\);', webpage)]
 233         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)]
 234
 235         # this is how player_core.js is actually doing (!)
 236         version = base_values.get('r_version') or program_values.get('version')
 237         if version:
 238             video_id = f'{video_id.split("_")[0]}_{version}'
 239
 240         formats = self._extract_m3u8_formats(
 241             f'https://nhks-vh.akamaihd.net/i/das/{video_id[0:8]}/{video_id}_V_000.f4v/master.m3u8',
 242             video_id, ext='mp4', m3u8_id='hls')
 243         self._sort_formats(formats)
 244
 245         duration = parse_duration(base_values.get('r_duration'))
 246
 247         chapters = None
 248         if chapter_durations and chapter_titles and len(chapter_durations) == len(chapter_titles):
 249             start_time = chapter_durations
 250             end_time = chapter_durations[1:] + [duration]
 251             chapters = [{
 252                 'start_time': s,
 253                 'end_time': e,
 254                 'title': t,
 255             } for s, e, t in zip(start_time, end_time, chapter_titles)]
 256
 257         return {
 258             'id': video_id,
 259             'title': program_values.get('name'),
 260             'duration': parse_duration(base_values.get('r_duration')),
 261             'timestamp': unified_timestamp(base_values['r_upload']),
 262             'formats': formats,
 263             'chapters': chapters,
 264         }
 265
 266
 267 class NhkForSchoolSubjectIE(InfoExtractor):
 268     IE_DESC = 'Portal page for each school subjects, like Japanese (kokugo, 国語) or math (sansuu/suugaku or 算数・数学)'
 269     KNOWN_SUBJECTS = (
 270         'rika', 'syakai', 'kokugo',
 271         'sansuu', 'seikatsu', 'doutoku',
 272         'ongaku', 'taiiku', 'zukou',
 273         'gijutsu', 'katei', 'sougou',
 274         'eigo', 'tokkatsu',
 275         'tokushi', 'sonota',
 276     )
 277     _VALID_URL = r'https?://www\.nhk\.or\.jp/school/(?P<id>%s)/?(?:[\?#].*)?$' % '|'.join(re.escape(s) for s in KNOWN_SUBJECTS)
 278
 279     _TESTS = [{
 280         'url': 'https://www.nhk.or.jp/school/sougou/',
 281         'info_dict': {
 282             'id': 'sougou',
 283             'title': '総合的な学習の時間',
 284         },
 285         'playlist_mincount': 16,
 286     }, {
 287         'url': 'https://www.nhk.or.jp/school/rika/',
 288         'info_dict': {
 289             'id': 'rika',
 290             'title': '理科',
 291         },
 292         'playlist_mincount': 15,
 293     }]
 294
 295     def _real_extract(self, url):
 296         subject_id = self._match_id(url)
 297         webpage = self._download_webpage(url, subject_id)
 298
 299         return self.playlist_from_matches(
 300             re.finditer(rf'href="((?:https?://www\.nhk\.or\.jp)?/school/{re.escape(subject_id)}/[^/]+/)"', webpage),
 301             subject_id,
 302             self._html_search_regex(r'(?s)<span\s+class="subjectName">\s*<img\s*[^<]+>\s*([^<]+?)</span>', webpage, 'title', fatal=False),
 303             lambda g: urljoin(url, g.group(1)))
 304
 305
 306 class NhkForSchoolProgramListIE(InfoExtractor):
 307     _VALID_URL = r'https?://www\.nhk\.or\.jp/school/(?P<id>(?:%s)/[a-zA-Z0-9_-]+)' % (
 308         '|'.join(re.escape(s) for s in NhkForSchoolSubjectIE.KNOWN_SUBJECTS)
 309     )
 310     _TESTS = [{
 311         'url': 'https://www.nhk.or.jp/school/sougou/q/',
 312         'info_dict': {
 313             'id': 'sougou/q',
 314             'title': 'Ｑ～こどものための哲学',
 315         },
 316         'playlist_mincount': 20,
 317     }]
 318
 319     def _real_extract(self, url):
 320         program_id = self._match_id(url)
 321
 322         webpage = self._download_webpage(f'https://www.nhk.or.jp/school/{program_id}/', program_id)
 323
 324         title = (self._og_search_title(webpage)
 325                  or self._html_extract_title(webpage)
 326                  or self._html_search_regex(r'<h3>([^<]+?)とは？\s*</h3>', webpage, 'title', fatal=False))
 327         title = re.sub(r'\s*\|\s*NHK\s+for\s+School\s*$', '', title) if title else None
 328         description = self._html_search_regex(
 329             r'(?s)<div\s+class="programDetail\s*">\s*<p>[^<]+</p>',
 330             webpage, 'description', fatal=False, group=0)
 331
 332         bangumi_list = self._download_json(
 333             f'https://www.nhk.or.jp/school/{program_id}/meta/program.json', program_id)
 334         # they're always bangumi
 335         bangumis = [
 336             self.url_result(f'https://www2.nhk.or.jp/school/movie/bangumi.cgi?das_id={x}')
 337             for x in traverse_obj(bangumi_list, ('part', ..., 'part-video-dasid')) or []]
 338
 339         return self.playlist_result(bangumis, program_id, title, description)