diff --git a/Makefile b/Makefile index 6cb9e2f57e..19a377002b 100644 --- a/Makefile +++ b/Makefile @@ -74,8 +74,7 @@ offlinetest: codetest $(PYTHON) -m pytest -k "not download" # XXX: This is hard to maintain -CODE_FOLDERS = yt_dlp yt_dlp/downloader yt_dlp/extractor yt_dlp/postprocessor yt_dlp/compat \ - yt_dlp/extractor/anvato_token_generator +CODE_FOLDERS = yt_dlp yt_dlp/downloader yt_dlp/extractor yt_dlp/postprocessor yt_dlp/compat yt-dlp: yt_dlp/*.py yt_dlp/*/*.py mkdir -p zip for d in $(CODE_FOLDERS) ; do \ diff --git a/setup.py b/setup.py index e376a694a3..3641dfae95 100644 --- a/setup.py +++ b/setup.py @@ -32,7 +32,6 @@ def packages(): return [ 'yt_dlp', 'yt_dlp.extractor', 'yt_dlp.downloader', 'yt_dlp.postprocessor', 'yt_dlp.compat', - 'yt_dlp.extractor.anvato_token_generator', ] diff --git a/yt_dlp/extractor/anvato.py b/yt_dlp/extractor/anvato.py index cb94835693..5d03070852 100644 --- a/yt_dlp/extractor/anvato.py +++ b/yt_dlp/extractor/anvato.py @@ -5,10 +5,8 @@ import re import time -from .anvato_token_generator import NFLTokenGenerator from .common import InfoExtractor from ..aes import aes_encrypt -from ..compat import compat_str from ..utils import ( bytes_to_intlist, determine_ext, @@ -16,20 +14,61 @@ int_or_none, join_nonempty, strip_jsonp, + smuggle_url, + traverse_obj, unescapeHTML, unsmuggle_url, ) def md5_text(s): - if not isinstance(s, compat_str): - s = compat_str(s) - return hashlib.md5(s.encode('utf-8')).hexdigest() + return hashlib.md5(str(s).encode()).hexdigest() class AnvatoIE(InfoExtractor): _VALID_URL = r'anvato:(?P[^:]+):(?P\d+)' + _API_BASE_URL = 'https://tkx.mp.lura.live/rest/v2' + _ANVP_RE = r']+\bdata-anvp\s*=\s*(["\'])(?P(?:(?!\1).)+)\1' + _AUTH_KEY = b'\x31\xc2\x42\x84\x9e\x73\xa0\xce' # from anvplayer.min.js + + _TESTS = [{ + # from https://www.nfl.com/videos/baker-mayfield-s-game-changing-plays-from-3-td-game-week-14 + 'url': 'anvato:GXvEgwyJeWem8KCYXfeoHWknwP48Mboj:899441', + 'md5': '921919dab3cd0b849ff3d624831ae3e2', + 'info_dict': { + 'id': '899441', + 'ext': 'mp4', + 'title': 'Baker Mayfield\'s game-changing plays from 3-TD game Week 14', + 'description': 'md5:85e05a3cc163f8c344340f220521136d', + 'upload_date': '20201215', + 'timestamp': 1608009755, + 'thumbnail': r're:^https?://.*\.jpg', + 'uploader': 'NFL', + 'tags': ['Baltimore Ravens at Cleveland Browns (2020-REG-14)', 'Baker Mayfield', 'Game Highlights', + 'Player Highlights', 'Cleveland Browns', 'league'], + 'duration': 157, + 'categories': ['Entertainment', 'Game', 'Highlights'], + }, + }, { + # from https://ktla.com/news/99-year-old-woman-learns-to-fly-in-torrance-checks-off-bucket-list-dream/ + 'url': 'anvato:X8POa4zpGZMmeiq0wqiO8IP5rMqQM9VN:8032455', + 'md5': '837718bcfb3a7778d022f857f7a9b19e', + 'info_dict': { + 'id': '8032455', + 'ext': 'mp4', + 'title': '99-year-old woman learns to fly plane in Torrance, checks off bucket list dream', + 'description': 'md5:0a12bab8159445e78f52a297a35c6609', + 'upload_date': '20220928', + 'timestamp': 1664408881, + 'thumbnail': r're:^https?://.*\.jpg', + 'uploader': 'LIN', + 'tags': ['video', 'news', '5live'], + 'duration': 155, + 'categories': ['News'], + }, + }] + # Copied from anvplayer.min.js _ANVACK_TABLE = { 'nbcu_nbcd_desktop_web_prod_93d8ead38ce2024f8f544b78306fbd15895ae5e6': 'NNemUkySjxLyPTKvZRiGntBIjEyK8uqicjMakIaQ', @@ -202,86 +241,74 @@ class AnvatoIE(InfoExtractor): 'telemundo': 'anvato_mcp_telemundo_web_prod_c5278d51ad46fda4b6ca3d0ea44a7846a054f582' } + def _generate_nfl_token(self, anvack, mcp_id): + reroute = self._download_json( + 'https://api.nfl.com/v1/reroute', mcp_id, data=b'grant_type=client_credentials', + headers={'X-Domain-Id': 100}, note='Fetching token info') + token_type = reroute.get('token_type') or 'Bearer' + auth_token = f'{token_type} {reroute["access_token"]}' + response = self._download_json( + 'https://api.nfl.com/v3/shield/', mcp_id, data=json.dumps({ + 'query': '''{ + viewer { + mediaToken(anvack: "%s", id: %s) { + token + } + } +}''' % (anvack, mcp_id), + }).encode(), headers={ + 'Authorization': auth_token, + 'Content-Type': 'application/json', + }, note='Fetching NFL API token') + return traverse_obj(response, ('data', 'viewer', 'mediaToken', 'token')) + _TOKEN_GENERATORS = { - 'GXvEgwyJeWem8KCYXfeoHWknwP48Mboj': NFLTokenGenerator, + 'GXvEgwyJeWem8KCYXfeoHWknwP48Mboj': _generate_nfl_token, } - _API_KEY = '3hwbSuqqT690uxjNYBktSQpa5ZrpYYR0Iofx7NcJHyA' - - _ANVP_RE = r']+\bdata-anvp\s*=\s*(["\'])(?P(?:(?!\1).)+)\1' - _AUTH_KEY = b'\x31\xc2\x42\x84\x9e\x73\xa0\xce' - - _TESTS = [{ - # from https://www.boston25news.com/news/watch-humpback-whale-breaches-right-next-to-fishing-boat-near-nh/817484874 - 'url': 'anvato:8v9BEynrwx8EFLYpgfOWcG1qJqyXKlRM:4465496', - 'info_dict': { - 'id': '4465496', - 'ext': 'mp4', - 'title': 'VIDEO: Humpback whale breaches right next to NH boat', - 'description': 'VIDEO: Humpback whale breaches right next to NH boat. Footage courtesy: Zach Fahey.', - 'duration': 22, - 'timestamp': 1534855680, - 'upload_date': '20180821', - 'uploader': 'ANV', - }, - 'params': { - 'skip_download': True, - }, - }, { - # from https://sanfrancisco.cbslocal.com/2016/06/17/source-oakland-cop-on-leave-for-having-girlfriend-help-with-police-reports/ - 'url': 'anvato:DVzl9QRzox3ZZsP9bNu5Li3X7obQOnqP:3417601', - 'only_matching': True, - }] - - def __init__(self, *args, **kwargs): - super(AnvatoIE, self).__init__(*args, **kwargs) - self.__server_time = None - def _server_time(self, access_key, video_id): - if self.__server_time is not None: - return self.__server_time + return int_or_none(traverse_obj(self._download_json( + f'{self._API_BASE_URL}/server_time', video_id, query={'anvack': access_key}, + note='Fetching server time', fatal=False), 'server_time')) or int(time.time()) - self.__server_time = int(self._download_json( - self._api_prefix(access_key) + 'server_time?anvack=' + access_key, video_id, - note='Fetching server time')['server_time']) - - return self.__server_time - - def _api_prefix(self, access_key): - return 'https://tkx2-%s.anvato.net/rest/v2/' % ('prod' if 'prod' in access_key else 'stage') - - def _get_video_json(self, access_key, video_id): + def _get_video_json(self, access_key, video_id, extracted_token): # See et() in anvplayer.min.js, which is an alias of getVideoJSON() - video_data_url = self._api_prefix(access_key) + 'mcp/video/%s?anvack=%s' % (video_id, access_key) + video_data_url = f'{self._API_BASE_URL}/mcp/video/{video_id}?anvack={access_key}' server_time = self._server_time(access_key, video_id) - input_data = '%d~%s~%s' % (server_time, md5_text(video_data_url), md5_text(server_time)) + input_data = f'{server_time}~{md5_text(video_data_url)}~{md5_text(server_time)}' auth_secret = intlist_to_bytes(aes_encrypt( bytes_to_intlist(input_data[:64]), bytes_to_intlist(self._AUTH_KEY))) - - video_data_url += '&X-Anvato-Adst-Auth=' + base64.b64encode(auth_secret).decode('ascii') + query = { + 'X-Anvato-Adst-Auth': base64.b64encode(auth_secret).decode('ascii'), + 'rtyp': 'fp', + } anvrid = md5_text(time.time() * 1000 * random.random())[:30] api = { 'anvrid': anvrid, 'anvts': server_time, } - if self._TOKEN_GENERATORS.get(access_key) is not None: - api['anvstk2'] = self._TOKEN_GENERATORS[access_key].generate(self, access_key, video_id) + if extracted_token is not None: + api['anvstk2'] = extracted_token + elif self._TOKEN_GENERATORS.get(access_key) is not None: + api['anvstk2'] = self._TOKEN_GENERATORS[access_key](self, access_key, video_id) + elif self._ANVACK_TABLE.get(access_key) is not None: + api['anvstk'] = md5_text(f'{access_key}|{anvrid}|{server_time}|{self._ANVACK_TABLE[access_key]}') else: - api['anvstk'] = md5_text('%s|%s|%d|%s' % ( - access_key, anvrid, server_time, - self._ANVACK_TABLE.get(access_key, self._API_KEY))) + api['anvstk2'] = 'default' return self._download_json( - video_data_url, video_id, transform_source=strip_jsonp, - data=json.dumps({'api': api}).encode('utf-8')) + video_data_url, video_id, transform_source=strip_jsonp, query=query, + data=json.dumps({'api': api}, separators=(',', ':')).encode('utf-8')) - def _get_anvato_videos(self, access_key, video_id): - video_data = self._get_video_json(access_key, video_id) + def _get_anvato_videos(self, access_key, video_id, token): + video_data = self._get_video_json(access_key, video_id, token) formats = [] for published_url in video_data['published_urls']: - video_url = published_url['embed_url'] + video_url = published_url.get('embed_url') + if not video_url: + continue media_format = published_url.get('format') ext = determine_ext(video_url) @@ -296,15 +323,27 @@ def _get_anvato_videos(self, access_key, video_id): 'tbr': tbr or None, } - if media_format == 'm3u8' and tbr is not None: + vtt_subs, hls_subs = {}, {} + if media_format == 'vtt': + _, vtt_subs = self._extract_m3u8_formats_and_subtitles( + video_url, video_id, m3u8_id='vtt', fatal=False) + continue + elif media_format == 'm3u8' and tbr is not None: a_format.update({ 'format_id': join_nonempty('hls', tbr), 'ext': 'mp4', }) elif media_format == 'm3u8-variant' or ext == 'm3u8': - formats.extend(self._extract_m3u8_formats( - video_url, video_id, 'mp4', entry_protocol='m3u8_native', - m3u8_id='hls', fatal=False)) + # For some videos the initial m3u8 URL returns JSON instead + manifest_json = self._download_json( + video_url, video_id, note='Downloading manifest JSON', errnote=False) + if manifest_json: + video_url = manifest_json.get('master_m3u8') + if not video_url: + continue + hls_fmts, hls_subs = self._extract_m3u8_formats_and_subtitles( + video_url, video_id, ext='mp4', m3u8_id='hls', fatal=False) + formats.extend(hls_fmts) continue elif ext == 'mp3' or media_format == 'mp3': a_format['vcodec'] = 'none' @@ -324,6 +363,7 @@ def _get_anvato_videos(self, access_key, video_id): 'ext': 'tt' if caption.get('format') == 'SMPTE-TT' else None } subtitles.setdefault(caption['language'], []).append(a_caption) + subtitles = self._merge_subtitles(subtitles, hls_subs, vtt_subs) return { 'id': video_id, @@ -349,7 +389,10 @@ def _extract_from_webpage(cls, url, webpage): access_key = cls._MCP_TO_ACCESS_KEY_TABLE.get((anvplayer_data.get('mcp') or '').lower()) if not (video_id or '').isdigit() or not access_key: continue - yield cls.url_result(f'anvato:{access_key}:{video_id}', AnvatoIE, video_id) + url = f'anvato:{access_key}:{video_id}' + if anvplayer_data.get('token'): + url = smuggle_url(url, {'token': anvplayer_data['token']}) + yield cls.url_result(url, AnvatoIE, video_id) def _extract_anvato_videos(self, webpage, video_id): anvplayer_data = self._parse_json( @@ -357,7 +400,7 @@ def _extract_anvato_videos(self, webpage, video_id): self._ANVP_RE, webpage, 'Anvato player data', group='anvp'), video_id) return self._get_anvato_videos( - anvplayer_data['accessKey'], anvplayer_data['video']) + anvplayer_data['accessKey'], anvplayer_data['video'], 'default') # cbslocal token = 'default' def _real_extract(self, url): url, smuggled_data = unsmuggle_url(url, {}) @@ -365,9 +408,7 @@ def _real_extract(self, url): 'countries': smuggled_data.get('geo_countries'), }) - mobj = self._match_valid_url(url) - access_key, video_id = mobj.group('access_key_or_mcp', 'id') + access_key, video_id = self._match_valid_url(url).group('access_key_or_mcp', 'id') if access_key not in self._ANVACK_TABLE: - access_key = self._MCP_TO_ACCESS_KEY_TABLE.get( - access_key) or access_key - return self._get_anvato_videos(access_key, video_id) + access_key = self._MCP_TO_ACCESS_KEY_TABLE.get(access_key) or access_key + return self._get_anvato_videos(access_key, video_id, smuggled_data.get('token')) diff --git a/yt_dlp/extractor/anvato_token_generator/__init__.py b/yt_dlp/extractor/anvato_token_generator/__init__.py deleted file mode 100644 index 6530caf530..0000000000 --- a/yt_dlp/extractor/anvato_token_generator/__init__.py +++ /dev/null @@ -1,5 +0,0 @@ -from .nfl import NFLTokenGenerator - -__all__ = [ - 'NFLTokenGenerator', -] diff --git a/yt_dlp/extractor/anvato_token_generator/common.py b/yt_dlp/extractor/anvato_token_generator/common.py deleted file mode 100644 index 3800b5808e..0000000000 --- a/yt_dlp/extractor/anvato_token_generator/common.py +++ /dev/null @@ -1,3 +0,0 @@ -class TokenGenerator: - def generate(self, anvack, mcp_id): - raise NotImplementedError('This method must be implemented by subclasses') diff --git a/yt_dlp/extractor/anvato_token_generator/nfl.py b/yt_dlp/extractor/anvato_token_generator/nfl.py deleted file mode 100644 index 9ee4aa002e..0000000000 --- a/yt_dlp/extractor/anvato_token_generator/nfl.py +++ /dev/null @@ -1,28 +0,0 @@ -import json - -from .common import TokenGenerator - - -class NFLTokenGenerator(TokenGenerator): - _AUTHORIZATION = None - - def generate(ie, anvack, mcp_id): - if not NFLTokenGenerator._AUTHORIZATION: - reroute = ie._download_json( - 'https://api.nfl.com/v1/reroute', mcp_id, - data=b'grant_type=client_credentials', - headers={'X-Domain-Id': 100}) - NFLTokenGenerator._AUTHORIZATION = '%s %s' % (reroute.get('token_type') or 'Bearer', reroute['access_token']) - return ie._download_json( - 'https://api.nfl.com/v3/shield/', mcp_id, data=json.dumps({ - 'query': '''{ - viewer { - mediaToken(anvack: "%s", id: %s) { - token - } - } -}''' % (anvack, mcp_id), - }).encode(), headers={ - 'Authorization': NFLTokenGenerator._AUTHORIZATION, - 'Content-Type': 'application/json', - })['data']['viewer']['mediaToken']['token']