From 0d32e124c69c5d53eb9dd286aa6147ec4cf52e54 Mon Sep 17 00:00:00 2001 From: makeworld <25111343+makeworld-the-better-one@users.noreply.github.com> Date: Sun, 19 Sep 2021 18:13:26 -0400 Subject: [PATCH] [CBC] Fix CBC Gem extractors (#1013) Closes #936 Authored by: makeworld-the-better-one --- yt_dlp/extractor/cbc.py | 486 ++++++++++++++------------------- yt_dlp/extractor/extractors.py | 6 +- 2 files changed, 212 insertions(+), 280 deletions(-) diff --git a/yt_dlp/extractor/cbc.py b/yt_dlp/extractor/cbc.py index fd5ec6033..061b09908 100644 --- a/yt_dlp/extractor/cbc.py +++ b/yt_dlp/extractor/cbc.py @@ -1,30 +1,18 @@ # coding: utf-8 from __future__ import unicode_literals -import hashlib -import json import re -from xml.sax.saxutils import escape from .common import InfoExtractor from ..compat import ( compat_str, - compat_HTTPError, ) from ..utils import ( js_to_json, smuggle_url, try_get, - xpath_text, - xpath_element, - xpath_with_ns, - find_xpath_attr, orderedSet, - parse_duration, - parse_iso8601, - parse_age_limit, strip_or_none, - int_or_none, ExtractorError, ) @@ -59,6 +47,7 @@ class CBCIE(InfoExtractor): 'uploader': 'CBCC-NEW', 'timestamp': 1382717907, }, + 'skip': 'No longer available', }, { # with clipId, feed only available via tpfeed.cbc.ca 'url': 'http://www.cbc.ca/archives/entry/1978-robin-williams-freestyles-on-90-minutes-live', @@ -209,289 +198,232 @@ def _real_extract(self, url): } -class CBCWatchBaseIE(InfoExtractor): - _device_id = None - _device_token = None - _API_BASE_URL = 'https://api-cbc.cloud.clearleap.com/cloffice/client/' - _NS_MAP = { - 'media': 'http://search.yahoo.com/mrss/', - 'clearleap': 'http://www.clearleap.com/namespace/clearleap/1.0/', - } - _GEO_COUNTRIES = ['CA'] - _LOGIN_URL = 'https://api.loginradius.com/identity/v2/auth/login' - _TOKEN_URL = 'https://cloud-api.loginradius.com/sso/jwt/api/token' - _API_KEY = '3f4beddd-2061-49b0-ae80-6f1f2ed65b37' - _NETRC_MACHINE = 'cbcwatch' - - def _signature(self, email, password): - data = json.dumps({ - 'email': email, - 'password': password, - }).encode() - headers = {'content-type': 'application/json'} - query = {'apikey': self._API_KEY} - resp = self._download_json(self._LOGIN_URL, None, data=data, headers=headers, query=query) - access_token = resp['access_token'] - - # token - query = { - 'access_token': access_token, - 'apikey': self._API_KEY, - 'jwtapp': 'jwt', - } - resp = self._download_json(self._TOKEN_URL, None, headers=headers, query=query) - return resp['signature'] - - def _call_api(self, path, video_id): - url = path if path.startswith('http') else self._API_BASE_URL + path - for _ in range(2): - try: - result = self._download_xml(url, video_id, headers={ - 'X-Clearleap-DeviceId': self._device_id, - 'X-Clearleap-DeviceToken': self._device_token, - }) - except ExtractorError as e: - if isinstance(e.cause, compat_HTTPError) and e.cause.code == 401: - # Device token has expired, re-acquiring device token - self._register_device() - continue - raise - error_message = xpath_text(result, 'userMessage') or xpath_text(result, 'systemMessage') - if error_message: - raise ExtractorError('%s said: %s' % (self.IE_NAME, error_message)) - return result - - def _real_initialize(self): - if self._valid_device_token(): - return - device = self._downloader.cache.load( - 'cbcwatch', self._cache_device_key()) or {} - self._device_id, self._device_token = device.get('id'), device.get('token') - if self._valid_device_token(): - return - self._register_device() - - def _valid_device_token(self): - return self._device_id and self._device_token - - def _cache_device_key(self): - email, _ = self._get_login_info() - return '%s_device' % hashlib.sha256(email.encode()).hexdigest() if email else 'device' - - def _register_device(self): - result = self._download_xml( - self._API_BASE_URL + 'device/register', - None, 'Acquiring device token', - data=b'web') - self._device_id = xpath_text(result, 'deviceId', fatal=True) - email, password = self._get_login_info() - if email and password: - signature = self._signature(email, password) - data = '{0}{1}web'.format( - escape(signature), escape(self._device_id)).encode() - url = self._API_BASE_URL + 'device/login' - result = self._download_xml( - url, None, data=data, - headers={'content-type': 'application/xml'}) - self._device_token = xpath_text(result, 'token', fatal=True) - else: - self._device_token = xpath_text(result, 'deviceToken', fatal=True) - self._downloader.cache.store( - 'cbcwatch', self._cache_device_key(), { - 'id': self._device_id, - 'token': self._device_token, - }) - - def _parse_rss_feed(self, rss): - channel = xpath_element(rss, 'channel', fatal=True) - - def _add_ns(path): - return xpath_with_ns(path, self._NS_MAP) - - entries = [] - for item in channel.findall('item'): - guid = xpath_text(item, 'guid', fatal=True) - title = xpath_text(item, 'title', fatal=True) - - media_group = xpath_element(item, _add_ns('media:group'), fatal=True) - content = xpath_element(media_group, _add_ns('media:content'), fatal=True) - content_url = content.attrib['url'] - - thumbnails = [] - for thumbnail in media_group.findall(_add_ns('media:thumbnail')): - thumbnail_url = thumbnail.get('url') - if not thumbnail_url: - continue - thumbnails.append({ - 'id': thumbnail.get('profile'), - 'url': thumbnail_url, - 'width': int_or_none(thumbnail.get('width')), - 'height': int_or_none(thumbnail.get('height')), - }) - - timestamp = None - release_date = find_xpath_attr( - item, _add_ns('media:credit'), 'role', 'releaseDate') - if release_date is not None: - timestamp = parse_iso8601(release_date.text) - - entries.append({ - '_type': 'url_transparent', - 'url': content_url, - 'id': guid, - 'title': title, - 'description': xpath_text(item, 'description'), - 'timestamp': timestamp, - 'duration': int_or_none(content.get('duration')), - 'age_limit': parse_age_limit(xpath_text(item, _add_ns('media:rating'))), - 'episode': xpath_text(item, _add_ns('clearleap:episode')), - 'episode_number': int_or_none(xpath_text(item, _add_ns('clearleap:episodeInSeason'))), - 'series': xpath_text(item, _add_ns('clearleap:series')), - 'season_number': int_or_none(xpath_text(item, _add_ns('clearleap:season'))), - 'thumbnails': thumbnails, - 'ie_key': 'CBCWatchVideo', - }) - - return self.playlist_result( - entries, xpath_text(channel, 'guid'), - xpath_text(channel, 'title'), - xpath_text(channel, 'description')) - - -class CBCWatchVideoIE(CBCWatchBaseIE): - IE_NAME = 'cbc.ca:watch:video' - _VALID_URL = r'https?://api-cbc\.cloud\.clearleap\.com/cloffice/client/web/play/?\?.*?\bcontentId=(?P[\da-f]{8}-[\da-f]{4}-[\da-f]{4}-[\da-f]{4}-[\da-f]{12})' - _TEST = { - # geo-restricted to Canada, bypassable - 'url': 'https://api-cbc.cloud.clearleap.com/cloffice/client/web/play/?contentId=3c84472a-1eea-4dee-9267-2655d5055dcf&categoryId=ebc258f5-ee40-4cca-b66b-ba6bd55b7235', - 'only_matching': True, - } - - def _real_extract(self, url): - video_id = self._match_id(url) - result = self._call_api(url, video_id) - - m3u8_url = xpath_text(result, 'url', fatal=True) - formats = self._extract_m3u8_formats(re.sub(r'/([^/]+)/[^/?]+\.m3u8', r'/\1/\1.m3u8', m3u8_url), video_id, 'mp4', fatal=False) - if len(formats) < 2: - formats = self._extract_m3u8_formats(m3u8_url, video_id, 'mp4') - for f in formats: - format_id = f.get('format_id') - if format_id.startswith('AAC'): - f['acodec'] = 'aac' - elif format_id.startswith('AC3'): - f['acodec'] = 'ac-3' - self._sort_formats(formats) - - info = { - 'id': video_id, - 'title': video_id, - 'formats': formats, - } - - rss = xpath_element(result, 'rss') - if rss: - info.update(self._parse_rss_feed(rss)['entries'][0]) - del info['url'] - del info['_type'] - del info['ie_key'] - return info - - -class CBCWatchIE(CBCWatchBaseIE): - IE_NAME = 'cbc.ca:watch' - _VALID_URL = r'https?://(?:gem|watch)\.cbc\.ca/(?:[^/]+/)+(?P[0-9a-f-]+)' +class CBCGemIE(InfoExtractor): + IE_NAME = 'gem.cbc.ca' + _VALID_URL = r'https?://gem\.cbc\.ca/media/(?P[0-9a-z-]+/s[0-9]+[a-z][0-9]+)' _TESTS = [{ # geo-restricted to Canada, bypassable - 'url': 'http://watch.cbc.ca/doc-zone/season-6/customer-disservice/38e815a-009e3ab12e4', + # This is a normal, public, TV show video + 'url': 'https://gem.cbc.ca/media/schitts-creek/s06e01', + 'md5': '93dbb31c74a8e45b378cf13bd3f6f11e', 'info_dict': { - 'id': '9673749a-5e77-484c-8b62-a1092a6b5168', + 'id': 'schitts-creek/s06e01', 'ext': 'mp4', - 'title': 'Customer (Dis)Service', - 'description': 'md5:8bdd6913a0fe03d4b2a17ebe169c7c87', - 'upload_date': '20160219', - 'timestamp': 1455840000, - }, - 'params': { - # m3u8 download - 'skip_download': True, - 'format': 'bestvideo', + 'title': 'Smoke Signals', + 'description': 'md5:929868d20021c924020641769eb3e7f1', + 'thumbnail': 'https://images.radio-canada.ca/v1/synps-cbc/episode/perso/cbc_schitts_creek_season_06e01_thumbnail_v01.jpg?im=Resize=(Size)', + 'duration': 1314, + 'categories': ['comedy'], + 'series': 'Schitt\'s Creek', + 'season': 'Season 6', + 'season_number': 6, + 'episode': 'Smoke Signals', + 'episode_number': 1, + 'episode_id': 'schitts-creek/s06e01', }, + 'params': {'format': 'bv'}, + 'skip': 'Geo-restricted to Canada', }, { # geo-restricted to Canada, bypassable - 'url': 'http://watch.cbc.ca/arthur/all/1ed4b385-cd84-49cf-95f0-80f004680057', + # This video requires an account in the browser, but works fine in yt-dlp + 'url': 'https://gem.cbc.ca/media/schitts-creek/s01e01', + 'md5': '297a9600f554f2258aed01514226a697', 'info_dict': { - 'id': '1ed4b385-cd84-49cf-95f0-80f004680057', - 'title': 'Arthur', - 'description': 'Arthur, the sweetest 8-year-old aardvark, and his pals solve all kinds of problems with humour, kindness and teamwork.', + 'id': 'schitts-creek/s01e01', + 'ext': 'mp4', + 'title': 'The Cup Runneth Over', + 'description': 'md5:9bca14ea49ab808097530eb05a29e797', + 'thumbnail': 'https://images.radio-canada.ca/v1/synps-cbc/episode/perso/cbc_schitts_creek_season_01e01_thumbnail_v01.jpg?im=Resize=(Size)', + 'series': 'Schitt\'s Creek', + 'season_number': 1, + 'season': 'Season 1', + 'episode_number': 1, + 'episode': 'The Cup Runneth Over', + 'episode_id': 'schitts-creek/s01e01', + 'duration': 1309, + 'categories': ['comedy'], }, - 'playlist_mincount': 30, - }, { - 'url': 'https://gem.cbc.ca/media/this-hour-has-22-minutes/season-26/episode-20/38e815a-0108c6c6a42', - 'only_matching': True, + 'params': {'format': 'bv'}, + 'skip': 'Geo-restricted to Canada', }] + _API_BASE = 'https://services.radio-canada.ca/ott/cbc-api/v2/assets/' def _real_extract(self, url): video_id = self._match_id(url) - rss = self._call_api('web/browse/' + video_id, video_id) - return self._parse_rss_feed(rss) + video_info = self._download_json(self._API_BASE + video_id, video_id) - -class CBCOlympicsIE(InfoExtractor): - IE_NAME = 'cbc.ca:olympics' - _VALID_URL = r'https?://olympics\.cbc\.ca/video/[^/]+/(?P[^/?#]+)' - _TESTS = [{ - 'url': 'https://olympics.cbc.ca/video/whats-on-tv/olympic-morning-featuring-the-opening-ceremony/', - 'only_matching': True, - }] - - def _real_extract(self, url): - display_id = self._match_id(url) - webpage = self._download_webpage(url, display_id) - video_id = self._hidden_inputs(webpage)['videoId'] - video_doc = self._download_xml( - 'https://olympics.cbc.ca/videodata/%s.xml' % video_id, video_id) - title = xpath_text(video_doc, 'title', fatal=True) - is_live = xpath_text(video_doc, 'kind') == 'Live' - if is_live: - title = self._live_title(title) - - formats = [] - for video_source in video_doc.findall('videoSources/videoSource'): - uri = xpath_text(video_source, 'uri') - if not uri: - continue - tokenize = self._download_json( - 'https://olympics.cbc.ca/api/api-akamai/tokenize', - video_id, data=json.dumps({ - 'VideoSource': uri, - }).encode(), headers={ - 'Content-Type': 'application/json', - 'Referer': url, - # d3.VideoPlayer._init in https://olympics.cbc.ca/components/script/base.js - 'Cookie': '_dvp=TK:C0ObxjerU', # AKAMAI CDN cookie - }, fatal=False) - if not tokenize: - continue - content_url = tokenize['ContentUrl'] - video_source_format = video_source.get('format') - if video_source_format == 'IIS': - formats.extend(self._extract_ism_formats( - content_url, video_id, ism_id=video_source_format, fatal=False)) + last_error = None + attempt = -1 + retries = self.get_param('extractor_retries', 15) + while attempt < retries: + attempt += 1 + if last_error: + self.report_warning('%s. Retrying ...' % last_error) + m3u8_info = self._download_json( + video_info['playSession']['url'], video_id, + note='Downloading JSON metadata%s' % f' (attempt {attempt})') + m3u8_url = m3u8_info.get('url') + if m3u8_url: + break + elif m3u8_info.get('errorCode') == 1: + self.raise_geo_restricted(countries=['CA']) else: - formats.extend(self._extract_m3u8_formats( - content_url, video_id, 'mp4', - 'm3u8' if is_live else 'm3u8_native', - m3u8_id=video_source_format, fatal=False)) + last_error = f'{self.IE_NAME} said: {m3u8_info.get("errorCode")} - {m3u8_info.get("message")}' + # 35 means media unavailable, but retries work + if m3u8_info.get('errorCode') != 35 or attempt >= retries: + raise ExtractorError(last_error) + + formats = self._extract_m3u8_formats(m3u8_url, video_id, m3u8_id='hls') + self._remove_duplicate_formats(formats) + + for i, format in enumerate(formats): + if format.get('vcodec') == 'none': + if format.get('ext') is None: + format['ext'] = 'm4a' + if format.get('acodec') is None: + format['acodec'] = 'mp4a.40.2' + + # Put described audio at the beginning of the list, so that it + # isn't chosen by default, as most people won't want it. + if 'descriptive' in format['format_id'].lower(): + format['preference'] = -2 + self._sort_formats(formats) return { 'id': video_id, - 'display_id': display_id, - 'title': title, - 'description': xpath_text(video_doc, 'description'), - 'thumbnail': xpath_text(video_doc, 'thumbnailUrl'), - 'duration': parse_duration(xpath_text(video_doc, 'duration')), + 'title': video_info['title'], + 'description': video_info.get('description'), + 'thumbnail': video_info.get('image'), + 'series': video_info.get('series'), + 'season_number': video_info.get('season'), + 'season': f'Season {video_info.get("season")}', + 'episode_number': video_info.get('episode'), + 'episode': video_info.get('title'), + 'episode_id': video_id, + 'duration': video_info.get('duration'), + 'categories': [video_info.get('category')], 'formats': formats, - 'is_live': is_live, + 'release_timestamp': video_info.get('airDate'), + 'timestamp': video_info.get('availableDate'), + } + + +class CBCGemPlaylistIE(InfoExtractor): + IE_NAME = 'gem.cbc.ca:playlist' + _VALID_URL = r'https?://gem\.cbc\.ca/media/(?P(?P[0-9a-z-]+)/s(?P[0-9]+))/?(?:[?#]|$)' + _TESTS = [{ + # geo-restricted to Canada, bypassable + # TV show playlist, all public videos + 'url': 'https://gem.cbc.ca/media/schitts-creek/s06', + 'playlist_count': 16, + 'info_dict': { + 'id': 'schitts-creek/s06', + 'title': 'Season 6', + 'description': 'md5:6a92104a56cbeb5818cc47884d4326a2', + }, + 'skip': 'Geo-restricted to Canada', + }] + _API_BASE = 'https://services.radio-canada.ca/ott/cbc-api/v2/shows/' + + def _real_extract(self, url): + match = self._match_valid_url(url) + season_id = match.group('id') + show = match.group('show') + show_info = self._download_json(self._API_BASE + show, season_id) + season = int(match.group('season')) + season_info = try_get(show_info, lambda x: x['seasons'][season - 1]) + + if season_info is None: + raise ExtractorError(f'Couldn\'t find season {season} of {show}') + + episodes = [] + for episode in season_info['assets']: + episodes.append({ + '_type': 'url_transparent', + 'ie_key': 'CBCGem', + 'url': 'https://gem.cbc.ca/media/' + episode['id'], + 'id': episode['id'], + 'title': episode.get('title'), + 'description': episode.get('description'), + 'thumbnail': episode.get('image'), + 'series': episode.get('series'), + 'season_number': episode.get('season'), + 'season': season_info['title'], + 'season_id': season_info.get('id'), + 'episode_number': episode.get('episode'), + 'episode': episode.get('title'), + 'episode_id': episode['id'], + 'duration': episode.get('duration'), + 'categories': [episode.get('category')], + }) + + thumbnail = None + tn_uri = season_info.get('image') + # the-national was observed to use a "data:image/png;base64" + # URI for their 'image' value. The image was 1x1, and is + # probably just a placeholder, so it is ignored. + if tn_uri is not None and not tn_uri.startswith('data:'): + thumbnail = tn_uri + + return { + '_type': 'playlist', + 'entries': episodes, + 'id': season_id, + 'title': season_info['title'], + 'description': season_info.get('description'), + 'thumbnail': thumbnail, + 'series': show_info.get('title'), + 'season_number': season_info.get('season'), + 'season': season_info['title'], + } + + +class CBCGemLiveIE(InfoExtractor): + IE_NAME = 'gem.cbc.ca:live' + _VALID_URL = r'https?://gem\.cbc\.ca/live/(?P[0-9]{12})' + _TEST = { + 'url': 'https://gem.cbc.ca/live/920604739687', + 'info_dict': { + 'title': 'Ottawa', + 'description': 'The live TV channel and local programming from Ottawa', + 'thumbnail': 'https://thumbnails.cbc.ca/maven_legacy/thumbnails/CBC_OTT_VMS/Live_Channel_Static_Images/Ottawa_2880x1620.jpg', + 'is_live': True, + 'id': 'AyqZwxRqh8EH', + 'ext': 'mp4', + 'timestamp': 1492106160, + 'upload_date': '20170413', + 'uploader': 'CBCC-NEW', + }, + 'skip': 'Live might have ended', + } + + # It's unclear where the chars at the end come from, but they appear to be + # constant. Might need updating in the future. + _API = 'https://tpfeed.cbc.ca/f/ExhSPC/t_t3UKJR6MAT' + + def _real_extract(self, url): + video_id = self._match_id(url) + live_info = self._download_json(self._API, video_id)['entries'] + + video_info = None + for stream in live_info: + if stream.get('guid') == video_id: + video_info = stream + + if video_info is None: + raise ExtractorError( + 'Couldn\'t find video metadata, maybe this livestream is now offline', + expected=True) + + return { + '_type': 'url_transparent', + 'ie_key': 'ThePlatform', + 'url': video_info['content'][0]['url'], + 'id': video_id, + 'title': video_info.get('title'), + 'description': video_info.get('description'), + 'tags': try_get(video_info, lambda x: x['keywords'].split(', ')), + 'thumbnail': video_info.get('cbc$staticImage'), + 'is_live': True, } diff --git a/yt_dlp/extractor/extractors.py b/yt_dlp/extractor/extractors.py index 6cafa82a2..10581a71b 100644 --- a/yt_dlp/extractor/extractors.py +++ b/yt_dlp/extractor/extractors.py @@ -208,9 +208,9 @@ from .cbc import ( CBCIE, CBCPlayerIE, - CBCWatchVideoIE, - CBCWatchIE, - CBCOlympicsIE, + CBCGemIE, + CBCGemPlaylistIE, + CBCGemLiveIE, ) from .cbs import CBSIE from .cbslocal import (