yt_dlp/extractor/cbc.py

   1 import base64
   2 import json
   3 import re
   4 import time
   5 import urllib.parse
   6 import xml.etree.ElementTree
   7
   8 from .common import InfoExtractor
   9 from ..compat import (
  10     compat_str,
  11 )
  12 from ..utils import (
  13     ExtractorError,
  14     int_or_none,
  15     join_nonempty,
  16     js_to_json,
  17     orderedSet,
  18     parse_iso8601,
  19     smuggle_url,
  20     strip_or_none,
  21     traverse_obj,
  22     try_get,
  23 )
  24
  25
  26 class CBCIE(InfoExtractor):
  27     IE_NAME = 'cbc.ca'
  28     _VALID_URL = r'https?://(?:www\.)?cbc\.ca/(?!player/)(?:[^/]+/)+(?P<id>[^/?#]+)'
  29     _TESTS = [{
  30         # with mediaId
  31         'url': 'http://www.cbc.ca/22minutes/videos/clips-season-23/don-cherry-play-offs',
  32         'md5': '97e24d09672fc4cf56256d6faa6c25bc',
  33         'info_dict': {
  34             'id': '2682904050',
  35             'ext': 'mp4',
  36             'title': 'Don Cherry – All-Stars',
  37             'description': 'Don Cherry has a bee in his bonnet about AHL player John Scott because that guy’s got heart.',
  38             'timestamp': 1454463000,
  39             'upload_date': '20160203',
  40             'uploader': 'CBCC-NEW',
  41         },
  42         'skip': 'Geo-restricted to Canada',
  43     }, {
  44         # with clipId, feed available via tpfeed.cbc.ca and feed.theplatform.com
  45         'url': 'http://www.cbc.ca/22minutes/videos/22-minutes-update/22-minutes-update-episode-4',
  46         'md5': '162adfa070274b144f4fdc3c3b8207db',
  47         'info_dict': {
  48             'id': '2414435309',
  49             'ext': 'mp4',
  50             'title': '22 Minutes Update: What Not To Wear Quebec',
  51             'description': "This week's latest Canadian top political story is What Not To Wear Quebec.",
  52             'upload_date': '20131025',
  53             'uploader': 'CBCC-NEW',
  54             'timestamp': 1382717907,
  55         },
  56         'skip': 'No longer available',
  57     }, {
  58         # with clipId, feed only available via tpfeed.cbc.ca
  59         'url': 'http://www.cbc.ca/archives/entry/1978-robin-williams-freestyles-on-90-minutes-live',
  60         'md5': '0274a90b51a9b4971fe005c63f592f12',
  61         'info_dict': {
  62             'id': '2487345465',
  63             'ext': 'mp4',
  64             'title': 'Robin Williams freestyles on 90 Minutes Live',
  65             'description': 'Wacky American comedian Robin Williams shows off his infamous "freestyle" comedic talents while being interviewed on CBC\'s 90 Minutes Live.',
  66             'upload_date': '19780210',
  67             'uploader': 'CBCC-NEW',
  68             'timestamp': 255977160,
  69         },
  70         'skip': '404 Not Found',
  71     }, {
  72         # multiple iframes
  73         'url': 'http://www.cbc.ca/natureofthings/blog/birds-eye-view-from-vancouvers-burrard-street-bridge-how-we-got-the-shot',
  74         'playlist': [{
  75             'md5': '377572d0b49c4ce0c9ad77470e0b96b4',
  76             'info_dict': {
  77                 'id': '2680832926',
  78                 'ext': 'mp4',
  79                 'title': 'An Eagle\'s-Eye View Off Burrard Bridge',
  80                 'description': 'Hercules the eagle flies from Vancouver\'s Burrard Bridge down to a nearby park with a mini-camera strapped to his back.',
  81                 'upload_date': '20160201',
  82                 'timestamp': 1454342820,
  83                 'uploader': 'CBCC-NEW',
  84             },
  85         }, {
  86             'md5': '415a0e3f586113894174dfb31aa5bb1a',
  87             'info_dict': {
  88                 'id': '2658915080',
  89                 'ext': 'mp4',
  90                 'title': 'Fly like an eagle!',
  91                 'description': 'Eagle equipped with a mini camera flies from the world\'s tallest tower',
  92                 'upload_date': '20150315',
  93                 'timestamp': 1426443984,
  94                 'uploader': 'CBCC-NEW',
  95             },
  96         }],
  97         'skip': 'Geo-restricted to Canada',
  98     }, {
  99         # multiple CBC.APP.Caffeine.initInstance(...)
 100         'url': 'http://www.cbc.ca/news/canada/calgary/dog-indoor-exercise-winter-1.3928238',
 101         'info_dict': {
 102             'title': 'Keep Rover active during the deep freeze with doggie pushups and other fun indoor tasks',  # FIXME
 103             'id': 'dog-indoor-exercise-winter-1.3928238',
 104             'description': 'md5:c18552e41726ee95bd75210d1ca9194c',
 105         },
 106         'playlist_mincount': 6,
 107     }]
 108
 109     @classmethod
 110     def suitable(cls, url):
 111         return False if CBCPlayerIE.suitable(url) else super(CBCIE, cls).suitable(url)
 112
 113     def _extract_player_init(self, player_init, display_id):
 114         player_info = self._parse_json(player_init, display_id, js_to_json)
 115         media_id = player_info.get('mediaId')
 116         if not media_id:
 117             clip_id = player_info['clipId']
 118             feed = self._download_json(
 119                 'http://tpfeed.cbc.ca/f/ExhSPC/vms_5akSXx4Ng_Zn?byCustomValue={:mpsReleases}{%s}' % clip_id,
 120                 clip_id, fatal=False)
 121             if feed:
 122                 media_id = try_get(feed, lambda x: x['entries'][0]['guid'], compat_str)
 123             if not media_id:
 124                 media_id = self._download_json(
 125                     'http://feed.theplatform.com/f/h9dtGB/punlNGjMlc1F?fields=id&byContent=byReleases%3DbyId%253D' + clip_id,
 126                     clip_id)['entries'][0]['id'].split('/')[-1]
 127         return self.url_result('cbcplayer:%s' % media_id, 'CBCPlayer', media_id)
 128
 129     def _real_extract(self, url):
 130         display_id = self._match_id(url)
 131         webpage = self._download_webpage(url, display_id)
 132         title = (self._og_search_title(webpage, default=None)
 133                  or self._html_search_meta('twitter:title', webpage, 'title', default=None)
 134                  or self._html_extract_title(webpage))
 135         entries = [
 136             self._extract_player_init(player_init, display_id)
 137             for player_init in re.findall(r'CBC\.APP\.Caffeine\.initInstance\(({.+?})\);', webpage)]
 138         media_ids = []
 139         for media_id_re in (
 140                 r'<iframe[^>]+src="[^"]+?mediaId=(\d+)"',
 141                 r'<div[^>]+\bid=["\']player-(\d+)',
 142                 r'guid["\']\s*:\s*["\'](\d+)'):
 143             media_ids.extend(re.findall(media_id_re, webpage))
 144         entries.extend([
 145             self.url_result('cbcplayer:%s' % media_id, 'CBCPlayer', media_id)
 146             for media_id in orderedSet(media_ids)])
 147         return self.playlist_result(
 148             entries, display_id, strip_or_none(title),
 149             self._og_search_description(webpage))
 150
 151
 152 class CBCPlayerIE(InfoExtractor):
 153     IE_NAME = 'cbc.ca:player'
 154     _VALID_URL = r'(?:cbcplayer:|https?://(?:www\.)?cbc\.ca/(?:player/play/|i/caffeine/syndicate/\?mediaId=))(?P<id>\d+)'
 155     _TESTS = [{
 156         'url': 'http://www.cbc.ca/player/play/2683190193',
 157         'md5': '64d25f841ddf4ddb28a235338af32e2c',
 158         'info_dict': {
 159             'id': '2683190193',
 160             'ext': 'mp4',
 161             'title': 'Gerry Runs a Sweat Shop',
 162             'description': 'md5:b457e1c01e8ff408d9d801c1c2cd29b0',
 163             'timestamp': 1455071400,
 164             'upload_date': '20160210',
 165             'uploader': 'CBCC-NEW',
 166         },
 167         'skip': 'Geo-restricted to Canada and no longer available',
 168     }, {
 169         # Redirected from http://www.cbc.ca/player/AudioMobile/All%20in%20a%20Weekend%20Montreal/ID/2657632011/
 170         'url': 'http://www.cbc.ca/player/play/2657631896',
 171         'md5': 'e5e708c34ae6fca156aafe17c43e8b75',
 172         'info_dict': {
 173             'id': '2657631896',
 174             'ext': 'mp3',
 175             'title': 'CBC Montreal is organizing its first ever community hackathon!',
 176             'description': 'The modern technology we tend to depend on so heavily, is never without it\'s share of hiccups and headaches. Next weekend - CBC Montreal will be getting members of the public for its first Hackathon.',
 177             'timestamp': 1425704400,
 178             'upload_date': '20150307',
 179             'uploader': 'CBCC-NEW',
 180             'thumbnail': 'http://thumbnails.cbc.ca/maven_legacy/thumbnails/sonali-karnick-220.jpg',
 181             'chapters': [],
 182             'duration': 494.811,
 183         },
 184     }, {
 185         'url': 'http://www.cbc.ca/player/play/2164402062',
 186         'md5': '33fcd8f6719b9dd60a5e73adcb83b9f6',
 187         'info_dict': {
 188             'id': '2164402062',
 189             'ext': 'mp4',
 190             'title': 'Cancer survivor four times over',
 191             'description': 'Tim Mayer has beaten three different forms of cancer four times in five years.',
 192             'timestamp': 1320410746,
 193             'upload_date': '20111104',
 194             'uploader': 'CBCC-NEW',
 195             'thumbnail': 'https://thumbnails.cbc.ca/maven_legacy/thumbnails/277/67/cancer_852x480_2164412612.jpg',
 196             'chapters': [],
 197             'duration': 186.867,
 198         },
 199     }, {
 200         # Has subtitles
 201         # These broadcasts expire after ~1 month, can find new test URL here:
 202         # https://www.cbc.ca/player/news/TV%20Shows/The%20National/Latest%20Broadcast
 203         'url': 'http://www.cbc.ca/player/play/2249992771553',
 204         'md5': '2f2fb675dd4f0f8a5bb7588d1b13bacd',
 205         'info_dict': {
 206             'id': '2249992771553',
 207             'ext': 'mp4',
 208             'title': 'The National | Women’s soccer pay, Florida seawater, Swift quake',
 209             'description': 'md5:adba28011a56cfa47a080ff198dad27a',
 210             'timestamp': 1690596000,
 211             'duration': 2716.333,
 212             'subtitles': {'eng': [{'ext': 'vtt', 'protocol': 'm3u8_native'}]},
 213             'thumbnail': 'https://thumbnails.cbc.ca/maven_legacy/thumbnails/481/326/thumbnail.jpeg',
 214             'uploader': 'CBCC-NEW',
 215             'chapters': 'count:5',
 216             'upload_date': '20230729',
 217         },
 218     }]
 219
 220     def _real_extract(self, url):
 221         video_id = self._match_id(url)
 222         return {
 223             '_type': 'url_transparent',
 224             'ie_key': 'ThePlatform',
 225             'url': smuggle_url(
 226                 'http://link.theplatform.com/s/ExhSPC/media/guid/2655402169/%s?mbr=true&formats=MPEG4,FLV,MP3' % video_id, {
 227                     'force_smil_url': True
 228                 }),
 229             'id': video_id,
 230             '_format_sort_fields': ('res', 'proto')  # Prioritize direct http formats over HLS
 231         }
 232
 233
 234 class CBCPlayerPlaylistIE(InfoExtractor):
 235     IE_NAME = 'cbc.ca:player:playlist'
 236     _VALID_URL = r'https?://(?:www\.)?cbc\.ca/(?:player/)(?!play/)(?P<id>[^?#]+)'
 237     _TESTS = [{
 238         'url': 'https://www.cbc.ca/player/news/TV%20Shows/The%20National/Latest%20Broadcast',
 239         'playlist_mincount': 25,
 240         'info_dict': {
 241             'id': 'news/tv shows/the national/latest broadcast',
 242         }
 243     }, {
 244         'url': 'https://www.cbc.ca/player/news/Canada/North',
 245         'playlist_mincount': 25,
 246         'info_dict': {
 247             'id': 'news/canada/north',
 248         }
 249     }]
 250
 251     def _real_extract(self, url):
 252         playlist_id = urllib.parse.unquote(self._match_id(url)).lower()
 253         webpage = self._download_webpage(url, playlist_id)
 254         json_content = self._search_json(
 255             r'window\.__INITIAL_STATE__\s*=', webpage, 'initial state', playlist_id)
 256
 257         def entries():
 258             for video_id in traverse_obj(json_content, (
 259                 'video', 'clipsByCategory', lambda k, _: k.lower() == playlist_id, 'items', ..., 'id'
 260             )):
 261                 yield self.url_result(f'https://www.cbc.ca/player/play/{video_id}', CBCPlayerIE)
 262
 263         return self.playlist_result(entries(), playlist_id)
 264
 265
 266 class CBCGemIE(InfoExtractor):
 267     IE_NAME = 'gem.cbc.ca'
 268     _VALID_URL = r'https?://gem\.cbc\.ca/(?:media/)?(?P<id>[0-9a-z-]+/s[0-9]+[a-z][0-9]+)'
 269     _TESTS = [{
 270         # This is a normal, public, TV show video
 271         'url': 'https://gem.cbc.ca/media/schitts-creek/s06e01',
 272         'md5': '93dbb31c74a8e45b378cf13bd3f6f11e',
 273         'info_dict': {
 274             'id': 'schitts-creek/s06e01',
 275             'ext': 'mp4',
 276             'title': 'Smoke Signals',
 277             'description': 'md5:929868d20021c924020641769eb3e7f1',
 278             'thumbnail': 'https://images.radio-canada.ca/v1/synps-cbc/episode/perso/cbc_schitts_creek_season_06e01_thumbnail_v01.jpg?im=Resize=(Size)',
 279             'duration': 1314,
 280             'categories': ['comedy'],
 281             'series': 'Schitt\'s Creek',
 282             'season': 'Season 6',
 283             'season_number': 6,
 284             'episode': 'Smoke Signals',
 285             'episode_number': 1,
 286             'episode_id': 'schitts-creek/s06e01',
 287         },
 288         'params': {'format': 'bv'},
 289         'skip': 'Geo-restricted to Canada',
 290     }, {
 291         # This video requires an account in the browser, but works fine in yt-dlp
 292         'url': 'https://gem.cbc.ca/media/schitts-creek/s01e01',
 293         'md5': '297a9600f554f2258aed01514226a697',
 294         'info_dict': {
 295             'id': 'schitts-creek/s01e01',
 296             'ext': 'mp4',
 297             'title': 'The Cup Runneth Over',
 298             'description': 'md5:9bca14ea49ab808097530eb05a29e797',
 299             'thumbnail': 'https://images.radio-canada.ca/v1/synps-cbc/episode/perso/cbc_schitts_creek_season_01e01_thumbnail_v01.jpg?im=Resize=(Size)',
 300             'series': 'Schitt\'s Creek',
 301             'season_number': 1,
 302             'season': 'Season 1',
 303             'episode_number': 1,
 304             'episode': 'The Cup Runneth Over',
 305             'episode_id': 'schitts-creek/s01e01',
 306             'duration': 1309,
 307             'categories': ['comedy'],
 308         },
 309         'params': {'format': 'bv'},
 310         'skip': 'Geo-restricted to Canada',
 311     }, {
 312         'url': 'https://gem.cbc.ca/nadiyas-family-favourites/s01e01',
 313         'only_matching': True,
 314     }]
 315
 316     _GEO_COUNTRIES = ['CA']
 317     _TOKEN_API_KEY = '3f4beddd-2061-49b0-ae80-6f1f2ed65b37'
 318     _NETRC_MACHINE = 'cbcgem'
 319     _claims_token = None
 320
 321     def _new_claims_token(self, email, password):
 322         data = json.dumps({
 323             'email': email,
 324             'password': password,
 325         }).encode()
 326         headers = {'content-type': 'application/json'}
 327         query = {'apikey': self._TOKEN_API_KEY}
 328         resp = self._download_json('https://api.loginradius.com/identity/v2/auth/login',
 329                                    None, data=data, headers=headers, query=query)
 330         access_token = resp['access_token']
 331
 332         query = {
 333             'access_token': access_token,
 334             'apikey': self._TOKEN_API_KEY,
 335             'jwtapp': 'jwt',
 336         }
 337         resp = self._download_json('https://cloud-api.loginradius.com/sso/jwt/api/token',
 338                                    None, headers=headers, query=query)
 339         sig = resp['signature']
 340
 341         data = json.dumps({'jwt': sig}).encode()
 342         headers = {'content-type': 'application/json', 'ott-device-type': 'web'}
 343         resp = self._download_json('https://services.radio-canada.ca/ott/cbc-api/v2/token',
 344                                    None, data=data, headers=headers, expected_status=426)
 345         cbc_access_token = resp['accessToken']
 346
 347         headers = {'content-type': 'application/json', 'ott-device-type': 'web', 'ott-access-token': cbc_access_token}
 348         resp = self._download_json('https://services.radio-canada.ca/ott/cbc-api/v2/profile',
 349                                    None, headers=headers, expected_status=426)
 350         return resp['claimsToken']
 351
 352     def _get_claims_token_expiry(self):
 353         # Token is a JWT
 354         # JWT is decoded here and 'exp' field is extracted
 355         # It is a Unix timestamp for when the token expires
 356         b64_data = self._claims_token.split('.')[1]
 357         data = base64.urlsafe_b64decode(b64_data + "==")
 358         return json.loads(data)['exp']
 359
 360     def claims_token_expired(self):
 361         exp = self._get_claims_token_expiry()
 362         if exp - time.time() < 10:
 363             # It will expire in less than 10 seconds, or has already expired
 364             return True
 365         return False
 366
 367     def claims_token_valid(self):
 368         return self._claims_token is not None and not self.claims_token_expired()
 369
 370     def _get_claims_token(self, email, password):
 371         if not self.claims_token_valid():
 372             self._claims_token = self._new_claims_token(email, password)
 373             self.cache.store(self._NETRC_MACHINE, 'claims_token', self._claims_token)
 374         return self._claims_token
 375
 376     def _real_initialize(self):
 377         if self.claims_token_valid():
 378             return
 379         self._claims_token = self.cache.load(self._NETRC_MACHINE, 'claims_token')
 380
 381     def _find_secret_formats(self, formats, video_id):
 382         """ Find a valid video url and convert it to the secret variant """
 383         base_format = next((f for f in formats if f.get('vcodec') != 'none'), None)
 384         if not base_format:
 385             return
 386
 387         base_url = re.sub(r'(Manifest\(.*?),filter=[\w-]+(.*?\))', r'\1\2', base_format['url'])
 388         url = re.sub(r'(Manifest\(.*?),format=[\w-]+(.*?\))', r'\1\2', base_url)
 389
 390         secret_xml = self._download_xml(url, video_id, note='Downloading secret XML', fatal=False)
 391         if not isinstance(secret_xml, xml.etree.ElementTree.Element):
 392             return
 393
 394         for child in secret_xml:
 395             if child.attrib.get('Type') != 'video':
 396                 continue
 397             for video_quality in child:
 398                 bitrate = int_or_none(video_quality.attrib.get('Bitrate'))
 399                 if not bitrate or 'Index' not in video_quality.attrib:
 400                     continue
 401                 height = int_or_none(video_quality.attrib.get('MaxHeight'))
 402
 403                 yield {
 404                     **base_format,
 405                     'format_id': join_nonempty('sec', height),
 406                     # Note: \g<1> is necessary instead of \1 since bitrate is a number
 407                     'url': re.sub(r'(QualityLevels\()\d+(\))', fr'\g<1>{bitrate}\2', base_url),
 408                     'width': int_or_none(video_quality.attrib.get('MaxWidth')),
 409                     'tbr': bitrate / 1000.0,
 410                     'height': height,
 411                 }
 412
 413     def _real_extract(self, url):
 414         video_id = self._match_id(url)
 415         video_info = self._download_json(
 416             f'https://services.radio-canada.ca/ott/cbc-api/v2/assets/{video_id}',
 417             video_id, expected_status=426)
 418
 419         email, password = self._get_login_info()
 420         if email and password:
 421             claims_token = self._get_claims_token(email, password)
 422             headers = {'x-claims-token': claims_token}
 423         else:
 424             headers = {}
 425         m3u8_info = self._download_json(video_info['playSession']['url'], video_id, headers=headers)
 426         m3u8_url = m3u8_info.get('url')
 427
 428         if m3u8_info.get('errorCode') == 1:
 429             self.raise_geo_restricted(countries=['CA'])
 430         elif m3u8_info.get('errorCode') == 35:
 431             self.raise_login_required(method='password')
 432         elif m3u8_info.get('errorCode') != 0:
 433             raise ExtractorError(f'{self.IE_NAME} said: {m3u8_info.get("errorCode")} - {m3u8_info.get("message")}')
 434
 435         formats = self._extract_m3u8_formats(m3u8_url, video_id, m3u8_id='hls')
 436         self._remove_duplicate_formats(formats)
 437         formats.extend(self._find_secret_formats(formats, video_id))
 438
 439         for format in formats:
 440             if format.get('vcodec') == 'none':
 441                 if format.get('ext') is None:
 442                     format['ext'] = 'm4a'
 443                 if format.get('acodec') is None:
 444                     format['acodec'] = 'mp4a.40.2'
 445
 446                 # Put described audio at the beginning of the list, so that it
 447                 # isn't chosen by default, as most people won't want it.
 448                 if 'descriptive' in format['format_id'].lower():
 449                     format['preference'] = -2
 450
 451         return {
 452             'id': video_id,
 453             'title': video_info['title'],
 454             'description': video_info.get('description'),
 455             'thumbnail': video_info.get('image'),
 456             'series': video_info.get('series'),
 457             'season_number': video_info.get('season'),
 458             'season': f'Season {video_info.get("season")}',
 459             'episode_number': video_info.get('episode'),
 460             'episode': video_info.get('title'),
 461             'episode_id': video_id,
 462             'duration': video_info.get('duration'),
 463             'categories': [video_info.get('category')],
 464             'formats': formats,
 465             'release_timestamp': video_info.get('airDate'),
 466             'timestamp': video_info.get('availableDate'),
 467         }
 468
 469
 470 class CBCGemPlaylistIE(InfoExtractor):
 471     IE_NAME = 'gem.cbc.ca:playlist'
 472     _VALID_URL = r'https?://gem\.cbc\.ca/(?:media/)?(?P<id>(?P<show>[0-9a-z-]+)/s(?P<season>[0-9]+))/?(?:[?#]|$)'
 473     _TESTS = [{
 474         # TV show playlist, all public videos
 475         'url': 'https://gem.cbc.ca/media/schitts-creek/s06',
 476         'playlist_count': 16,
 477         'info_dict': {
 478             'id': 'schitts-creek/s06',
 479             'title': 'Season 6',
 480             'description': 'md5:6a92104a56cbeb5818cc47884d4326a2',
 481             'series': 'Schitt\'s Creek',
 482             'season_number': 6,
 483             'season': 'Season 6',
 484             'thumbnail': 'https://images.radio-canada.ca/v1/synps-cbc/season/perso/cbc_schitts_creek_season_06_carousel_v03.jpg?impolicy=ott&im=Resize=(_Size_)&quality=75',
 485         },
 486     }, {
 487         'url': 'https://gem.cbc.ca/schitts-creek/s06',
 488         'only_matching': True,
 489     }]
 490     _API_BASE = 'https://services.radio-canada.ca/ott/cbc-api/v2/shows/'
 491
 492     def _real_extract(self, url):
 493         match = self._match_valid_url(url)
 494         season_id = match.group('id')
 495         show = match.group('show')
 496         show_info = self._download_json(self._API_BASE + show, season_id, expected_status=426)
 497         season = int(match.group('season'))
 498
 499         season_info = next((s for s in show_info['seasons'] if s.get('season') == season), None)
 500
 501         if season_info is None:
 502             raise ExtractorError(f'Couldn\'t find season {season} of {show}')
 503
 504         episodes = []
 505         for episode in season_info['assets']:
 506             episodes.append({
 507                 '_type': 'url_transparent',
 508                 'ie_key': 'CBCGem',
 509                 'url': 'https://gem.cbc.ca/media/' + episode['id'],
 510                 'id': episode['id'],
 511                 'title': episode.get('title'),
 512                 'description': episode.get('description'),
 513                 'thumbnail': episode.get('image'),
 514                 'series': episode.get('series'),
 515                 'season_number': episode.get('season'),
 516                 'season': season_info['title'],
 517                 'season_id': season_info.get('id'),
 518                 'episode_number': episode.get('episode'),
 519                 'episode': episode.get('title'),
 520                 'episode_id': episode['id'],
 521                 'duration': episode.get('duration'),
 522                 'categories': [episode.get('category')],
 523             })
 524
 525         thumbnail = None
 526         tn_uri = season_info.get('image')
 527         # the-national was observed to use a "data:image/png;base64"
 528         # URI for their 'image' value. The image was 1x1, and is
 529         # probably just a placeholder, so it is ignored.
 530         if tn_uri is not None and not tn_uri.startswith('data:'):
 531             thumbnail = tn_uri
 532
 533         return {
 534             '_type': 'playlist',
 535             'entries': episodes,
 536             'id': season_id,
 537             'title': season_info['title'],
 538             'description': season_info.get('description'),
 539             'thumbnail': thumbnail,
 540             'series': show_info.get('title'),
 541             'season_number': season_info.get('season'),
 542             'season': season_info['title'],
 543         }
 544
 545
 546 class CBCGemLiveIE(InfoExtractor):
 547     IE_NAME = 'gem.cbc.ca:live'
 548     _VALID_URL = r'https?://gem\.cbc\.ca/live(?:-event)?/(?P<id>\d+)'
 549     _TESTS = [
 550         {
 551             'url': 'https://gem.cbc.ca/live/920604739687',
 552             'info_dict': {
 553                 'title': 'Ottawa',
 554                 'description': 'The live TV channel and local programming from Ottawa',
 555                 'thumbnail': 'https://thumbnails.cbc.ca/maven_legacy/thumbnails/CBC_OTT_VMS/Live_Channel_Static_Images/Ottawa_2880x1620.jpg',
 556                 'is_live': True,
 557                 'id': 'AyqZwxRqh8EH',
 558                 'ext': 'mp4',
 559                 'timestamp': 1492106160,
 560                 'upload_date': '20170413',
 561                 'uploader': 'CBCC-NEW',
 562             },
 563             'skip': 'Live might have ended',
 564         },
 565         {
 566             'url': 'https://gem.cbc.ca/live/44',
 567             'info_dict': {
 568                 'id': '44',
 569                 'ext': 'mp4',
 570                 'is_live': True,
 571                 'title': r're:^Ottawa [0-9\-: ]+',
 572                 'description': 'The live TV channel and local programming from Ottawa',
 573                 'live_status': 'is_live',
 574                 'thumbnail': r're:https://images.gem.cbc.ca/v1/cbc-gem/live/.*'
 575             },
 576             'params': {'skip_download': True},
 577             'skip': 'Live might have ended',
 578         },
 579         {
 580             'url': 'https://gem.cbc.ca/live-event/10835',
 581             'info_dict': {
 582                 'id': '10835',
 583                 'ext': 'mp4',
 584                 'is_live': True,
 585                 'title': r're:^The National \| Biden’s trip wraps up, Paltrow testifies, Bird flu [0-9\-: ]+',
 586                 'description': 'March 24, 2023 | President Biden’s Ottawa visit ends with big pledges from both countries. Plus, Gwyneth Paltrow testifies in her ski collision trial.',
 587                 'live_status': 'is_live',
 588                 'thumbnail': r're:https://images.gem.cbc.ca/v1/cbc-gem/live/.*',
 589                 'timestamp': 1679706000,
 590                 'upload_date': '20230325',
 591             },
 592             'params': {'skip_download': True},
 593             'skip': 'Live might have ended',
 594         }
 595     ]
 596
 597     def _real_extract(self, url):
 598         video_id = self._match_id(url)
 599         webpage = self._download_webpage(url, video_id)
 600         video_info = self._search_nextjs_data(webpage, video_id)['props']['pageProps']['data']
 601
 602         # Two types of metadata JSON
 603         if not video_info.get('formattedIdMedia'):
 604             video_info = traverse_obj(
 605                 video_info, (('freeTv', ('streams', ...)), 'items', lambda _, v: v['key'] == video_id, {dict}),
 606                 get_all=False, default={})
 607
 608         video_stream_id = video_info.get('formattedIdMedia')
 609         if not video_stream_id:
 610             raise ExtractorError('Couldn\'t find video metadata, maybe this livestream is now offline', expected=True)
 611
 612         stream_data = self._download_json(
 613             'https://services.radio-canada.ca/media/validation/v2/', video_id, query={
 614                 'appCode': 'mpx',
 615                 'connectionType': 'hd',
 616                 'deviceType': 'ipad',
 617                 'idMedia': video_stream_id,
 618                 'multibitrate': 'true',
 619                 'output': 'json',
 620                 'tech': 'hls',
 621                 'manifestType': 'desktop',
 622             })
 623
 624         return {
 625             'id': video_id,
 626             'formats': self._extract_m3u8_formats(stream_data['url'], video_id, 'mp4', live=True),
 627             'is_live': True,
 628             **traverse_obj(video_info, {
 629                 'title': 'title',
 630                 'description': 'description',
 631                 'thumbnail': ('images', 'card', 'url'),
 632                 'timestamp': ('airDate', {parse_iso8601}),
 633             })
 634         }