Merge d893535d7b62439186353dad0da4fddb251a6a69 into 2b4fbfce25902d557b86b003cf48f738129efce4

thx seproDev, bashonly: yt-dlp/yt-dlp#12748

test/test_youtube_signature.py

@@ -84,6 +84,21 @@ _SIG_TESTS = [
         '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
         '0QJ8wRAIgXmPlOPSBkkUs1bYFYlJCfe29xxAj7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJ2OySqa0q',
     ),
+    (
+        'https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js',
+        '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
+        'AAOAOq0QJ8wRAIgXmPlOPSBkkUs1bYFYlJCfe29xx8j7vgpDL0QwbdV06sCIEzpWqMGkFR20CFOS21Tp-7vj_EMu-m37KtXJoOy1',
+    ),
+    (
+        'https://www.youtube.com/s/player/363db69b/player_ias.vflset/en_US/base.js',
+        '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
+        '0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpz2ICs6EVdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
+    ),
+    (
+        'https://www.youtube.com/s/player/4fcd6e4a/player_ias.vflset/en_US/base.js',
+        '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
+        'wAOAOq0QJ8ARAIgXmPlOPSBkkUs1bYFYlJCfe29xx8q7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0',
+    ),
 ]
 
 _NSIG_TESTS = [
@@ -153,7 +168,7 @@ _NSIG_TESTS = [
     ),
     (
         'https://www.youtube.com/s/player/c57c113c/player_ias.vflset/en_US/base.js',
-        '-Txvy6bT5R6LqgnQNx', 'dcklJCnRUHbgSg',
+        'M92UUMHa8PdvPd3wyM', '3hPqLJsiNZx7yA',
     ),
     (
         'https://www.youtube.com/s/player/5a3b6271/player_ias.vflset/en_US/base.js',
@@ -173,7 +188,7 @@ _NSIG_TESTS = [
     ),
     (
         'https://www.youtube.com/s/player/cfa9e7cb/player_ias.vflset/en_US/base.js',
-        'qO0NiMtYQ7TeJnfFG2', 'k9cuJDHNS5O7kQ',
+        'aCi3iElgd2kq0bxVbQ', 'QX1y8jGb2IbZ0w',
     ),
     (
         'https://www.youtube.com/s/player/8c7583ff/player_ias.vflset/en_US/base.js',
@@ -231,10 +246,6 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/f6e09c70/player_ias_tce.vflset/en_US/base.js',
         'W9HJZKktxuYoDTqW', 'jHbbkcaxm54',
     ),
-    (
-        'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js',
-        'W9HJZKktxuYoDTqW', 'larxUlagTRAcSw',
-    ),
     (
         'https://www.youtube.com/s/player/e7567ecf/player_ias_tce.vflset/en_US/base.js',
         'Sy4aDGc0VpYRR9ew_', '5UPOT1VhoZxNLQ',
@@ -259,6 +270,22 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js',
         'ir9-V6cdbCiyKxhr', '2PL7ZDYAALMfmA',
     ),
+    (
+        'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js',
+        'W9HJZKktxuYoDTqW', 'larxUlagTRAcSw',
+    ),
+    (
+        'https://www.youtube.com/s/player/363db69b/player_ias.vflset/en_US/base.js',
+        'eWYu5d5YeY_4LyEDc', 'XJQqf-N7Xra3gg',
+    ),
+    (
+        'https://www.youtube.com/s/player/4fcd6e4a/player_ias.vflset/en_US/base.js',
+        'o_L251jm8yhZkWtBW', 'lXoxI3XvToqn6A',
+    ),
+    (
+        'https://www.youtube.com/s/player/4fcd6e4a/tv-player-ias.vflset/tv-player-ias.js',
+        'o_L251jm8yhZkWtBW', 'lXoxI3XvToqn6A',
+    ),
 ]
 
 
@@ -271,6 +298,8 @@ class TestPlayerInfo(unittest.TestCase):
             ('https://www.youtube.com/s/player/64dddad9/player-plasma-ias-phone-en_US.vflset/base.js', '64dddad9'),
             ('https://www.youtube.com/s/player/64dddad9/player-plasma-ias-phone-de_DE.vflset/base.js', '64dddad9'),
             ('https://www.youtube.com/s/player/64dddad9/player-plasma-ias-tablet-en_US.vflset/base.js', '64dddad9'),
+            ('https://www.youtube.com/s/player/e7567ecf/player_ias_tce.vflset/en_US/base.js', 'e7567ecf'),
+            ('https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js', '643afba4'),
             # obsolete
             ('https://www.youtube.com/yts/jsbin/player_ias-vfle4-e03/en_US/base.js', 'vfle4-e03'),
             ('https://www.youtube.com/yts/jsbin/player_ias-vfl49f_g4/en_US/base.js', 'vfl49f_g4'),
@@ -280,8 +309,9 @@ class TestPlayerInfo(unittest.TestCase):
             ('https://s.ytimg.com/yts/jsbin/html5player-en_US-vflXGBaUN.js', 'vflXGBaUN'),
             ('https://s.ytimg.com/yts/jsbin/html5player-en_US-vflKjOTVq/html5player.js', 'vflKjOTVq'),
         )
+        ie = YoutubeIE(FakeYDL({'cachedir': False}))
         for player_url, expected_player_id in PLAYER_URLS:
-            player_id = YoutubeIE._extract_player_info(player_url)
+            player_id = ie._extract_player_info(player_url)
             self.assertEqual(player_id, expected_player_id)
 
 
@@ -301,8 +331,8 @@ class TestSignature(unittest.TestCase):
 def t_factory(name, sig_func, url_pattern):
     def make_tfunc(url, sig_input, expected_sig):
         m = url_pattern.match(url)
-        assert m, '%r should follow URL format' % url
-        test_id = m.group('id')
+        assert m, '{0!r} should follow URL format'.format(url)
+        test_id = re.sub(r'[/.-]', '_', m.group('id') or m.group('compat_id'))
 
         def test_func(self):
             basename = 'player-{0}-{1}.js'.format(name, test_id)
@@ -335,12 +365,16 @@ def n_sig(jscode, sig_input):
 
 
 make_sig_test = t_factory(
-    'signature', signature, re.compile(r'.*(?:-|/player/)(?P<id>[a-zA-Z0-9_-]+)(?:/.+\.js|(?:/watch_as3|/html5player)?\.[a-z]+)$'))
+    'signature', signature,
+    re.compile(r'''(?x)
+        .+/(?P<h5>html5)?player(?(h5)(?:-en_US)?-|/)(?P<id>[a-zA-Z0-9/._-]+)
+        (?(h5)/(?:watch_as3|html5player))?\.js$
+    '''))
 for test_spec in _SIG_TESTS:
     make_sig_test(*test_spec)
 
 make_nsig_test = t_factory(
-    'nsig', n_sig, re.compile(r'.+/player/(?P<id>[a-zA-Z0-9_-]+)/.+.js$'))
+    'nsig', n_sig, re.compile(r'.+/player/(?P<id>[a-zA-Z0-9_/.-]+)\.js$'))
 for test_spec in _NSIG_TESTS:
     make_nsig_test(*test_spec)
 

youtube_dl/extractor/extractors.py

@@ -1482,7 +1482,7 @@ from .vidme import (
     VidmeUserIE,
     VidmeUserLikesIE,
 )
-from .vier import VierIE, VierVideosIE
+from .vier import VierIE, VierVideosIE, VierVijfKijkOnlineIE
 from .viewlift import (
     ViewLiftIE,
     ViewLiftEmbedIE,

youtube_dl/extractor/vier.py

@@ -2,8 +2,10 @@
 from __future__ import unicode_literals
 
 import re
+import json
 import itertools
 
+from .vier_auth_aws import AwsIdp
 from .common import InfoExtractor
 from ..utils import (
     urlencode_postdata,
@@ -12,16 +14,116 @@ from ..utils import (
 )
 
 
+class VierVijfKijkOnlineIE(InfoExtractor):
+    IE_NAME = 'viervijfkijkonline'
+    IE_DESC = 'vier.be and vijf.be - Kijk Online'
+    _VALID_URL = r'https?://(?:www\.)?(?P<site>vier|vijf|goplay)\.be/video/(?P<series>(?!v3)[^/]+)/(?P<season>[^/]+)(/(?P<episode>[^/]+)|)'
+    _NETRC_MACHINE = 'vier'
+    _TESTS = [{
+        'url': 'https://www.vier.be/video/hotel-romantiek/2017/hotel-romantiek-aflevering-1',
+        'info_dict': {
+            'id': 'ebcd3c39-10a2-4730-b137-b0e7aaed247c',
+            'ext': 'mp4',
+            'title': 'Hotel Römantiek - Seizoen 1 - Aflevering 1',
+            'series': 'Hotel Römantiek',
+            'season_number': 1,
+            'episode_number': 1,
+        },
+        'skip': 'This video is only available for registered users'
+    }, {
+        'url': 'https://www.vier.be/video/blockbusters/in-juli-en-augustus-summer-classics',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.vier.be/video/achter-de-rug/2017/achter-de-rug-seizoen-1-aflevering-6',
+        'only_matching': True,
+    }]
+
+    def _real_initialize(self):
+        self._logged_in = False
+        self.id_token = ''
+
+    def _login(self):
+
+        username, password = self._get_login_info()
+        if username is None or password is None:
+            self.raise_login_required()
+
+        aws = AwsIdp(pool_id='eu-west-1_dViSsKM5Y', client_id='6s1h851s8uplco5h6mqh1jac8m')
+        self.id_token, _ = aws.authenticate(username=username, password=password)
+        self._logged_in = True
+
+    def _real_extract(self, url):
+        if "#" in url:
+            url = url.split("#")[0]
+
+        if not self._logged_in:
+            self._login()
+
+        webpage = self._download_webpage(url, None)
+
+        title = self._html_search_regex(
+            r'<meta\s*property="og:title"\s*content="(.+?)"\s*/>',
+            webpage, 'title')
+
+        title_split = title.split(' - ')
+        series = title_split[0].strip()
+        if len(title_split) == 3:
+            if 'Seizoen' in title_split[1]:
+                season = title_split[1].split('Seizoen')[1].strip()
+            else:
+                season = title_split[1].strip()
+            episode = title_split[2].split('Aflevering')[1].strip()
+        elif len(title_split) == 1:
+            season = None
+            episode = None
+        else:
+            season = None
+            episode = title_split[1].split('Aflevering')[1].strip()
+
+        video_data = self._html_search_regex(
+            r'<div data-hero="([^"]*)"',
+            webpage, 'video_data')
+
+        playlists = json.loads(video_data.replace('&quot;', '"'))['data']['playlists']
+        wanted_playlist = [x for x in playlists if x['pageInfo']['url'] in url][0]
+        wanted_episode = [x for x in wanted_playlist['episodes'] if x['pageInfo']['url'] == url][0] or [x for x in wanted_playlist['episodes'] if x['pageInfo']['url'] in url][0]
+        video_id = wanted_episode['videoUuid']
+
+        api_url = 'https://api.viervijfzes.be/content/%s' % (video_id)
+        api_headers = {
+            'authorization': self.id_token,
+        }
+        api = self._download_json(
+            api_url,
+            None, note='Peforming API Call', errnote='API Call Failed',
+            headers=api_headers,
+        )
+
+        formats = []
+        formats.extend(self._extract_m3u8_formats(
+            api['video']['S'], video_id, 'mp4', entry_protocol='m3u8_native',
+            m3u8_id='HLS', fatal=False))
+
+        self._sort_formats(formats)
+        return {
+            'id': video_id,
+            'title': title,
+            'series': series,
+            'season_number': int_or_none(season),
+            'episode_number': int_or_none(episode),
+            'formats': formats,
+        }
+
+
 class VierIE(InfoExtractor):
     IE_NAME = 'vier'
     IE_DESC = 'vier.be and vijf.be'
     _VALID_URL = r'''(?x)
                     https?://
-                        (?:www\.)?(?P<site>vier|vijf)\.be/
+                        (?:www\.)?(?P<site>vier|vijf|goplay)\.be/
                         (?:
                             (?:
-                                [^/]+/videos|
-                                video(?:/[^/]+)*
+                                [^/]+/videos
                             )/
                             (?P<display_id>[^/]+)(?:/(?P<id>\d+))?|
                             (?:
@@ -100,12 +202,6 @@ class VierIE(InfoExtractor):
     }, {
         'url': 'https://www.vijf.be/embed/video/public/4093',
         'only_matching': True,
-    }, {
-        'url': 'https://www.vier.be/video/blockbusters/in-juli-en-augustus-summer-classics',
-        'only_matching': True,
-    }, {
-        'url': 'https://www.vier.be/video/achter-de-rug/2017/achter-de-rug-seizoen-1-aflevering-6',
-        'only_matching': True,
     }]
 
     def _real_initialize(self):

youtube_dl/extractor/vier_auth_aws.py

@@ -0,0 +1,320 @@
+# coding: utf-8
+from __future__ import absolute_import, division, unicode_literals
+
+# Taken from https://github.com/add-ons/plugin.video.viervijfzes/blob/master/resources/lib/viervijfzes/auth_awsidp.py
+# with some small modifications (replacement of six and requests, removal of unused function, reformatting)
+# This file is licensed as GPL-3 from https://github.com/add-ons/plugin.video.viervijfzes:
+# https://github.com/add-ons/plugin.video.viervijfzes/blob/master/LICENSE
+
+# Amazon Cognito implementation without external dependencies
+# Based on https://github.com/retrospect-addon/plugin.video.retrospect/blob/master/channels/channel.be/vier/awsidp.py
+
+import base64
+import binascii
+import datetime
+import hashlib
+import hmac
+import json
+import os
+
+from ..compat import compat_basestring, compat_urllib_request
+
+
+class InvalidLoginException(Exception):
+    """ The login credentials are invalid """
+
+
+class AuthenticationException(Exception):
+    """ Something went wrong while logging in """
+
+
+class AwsIdp:
+    """ AWS Identity Provider """
+
+    def __init__(self, pool_id, client_id):
+        """
+        :param str pool_id:     The AWS user pool to connect to (format: <region>_<poolid>).
+                                E.g.: eu-west-1_aLkOfYN3T
+        :param str client_id:   The client application ID (the ID of the application connecting)
+        """
+
+        self.pool_id = pool_id
+        if "_" not in self.pool_id:
+            raise ValueError("Invalid pool_id format. Should be <region>_<poolid>.")
+
+        self.client_id = client_id
+        self.region = self.pool_id.split("_")[0]
+        self.url = "https://cognito-idp.%s.amazonaws.com/" % (self.region,)
+
+        # Initialize the values
+        # https://github.com/aws/amazon-cognito-identity-js/blob/master/src/AuthenticationHelper.js#L22
+        self.n_hex = 'FFFFFFFFFFFFFFFFC90FDAA22168C234C4C6628B80DC1CD1' + \
+                     '29024E088A67CC74020BBEA63B139B22514A08798E3404DD' + \
+                     'EF9519B3CD3A431B302B0A6DF25F14374FE1356D6D51C245' + \
+                     'E485B576625E7EC6F44C42E9A637ED6B0BFF5CB6F406B7ED' + \
+                     'EE386BFB5A899FA5AE9F24117C4B1FE649286651ECE45B3D' + \
+                     'C2007CB8A163BF0598DA48361C55D39A69163FA8FD24CF5F' + \
+                     '83655D23DCA3AD961C62F356208552BB9ED529077096966D' + \
+                     '670C354E4ABC9804F1746C08CA18217C32905E462E36CE3B' + \
+                     'E39E772C180E86039B2783A2EC07A28FB5C55DF06F4C52C9' + \
+                     'DE2BCBF6955817183995497CEA956AE515D2261898FA0510' + \
+                     '15728E5A8AAAC42DAD33170D04507A33A85521ABDF1CBA64' + \
+                     'ECFB850458DBEF0A8AEA71575D060C7DB3970F85A6E1E4C7' + \
+                     'ABF5AE8CDB0933D71E8C94E04A25619DCEE3D2261AD2EE6B' + \
+                     'F12FFA06D98A0864D87602733EC86A64521F2B18177B200C' + \
+                     'BBE117577A615D6C770988C0BAD946E208E24FA074E5AB31' + \
+                     '43DB5BFCE0FD108E4B82D120A93AD2CAFFFFFFFFFFFFFFFF'
+
+        # https://github.com/aws/amazon-cognito-identity-js/blob/master/src/AuthenticationHelper.js#L49
+        self.g_hex = '2'
+        self.info_bits = bytearray('Caldera Derived Key', 'utf-8')
+
+        self.big_n = self.__hex_to_long(self.n_hex)
+        self.g = self.__hex_to_long(self.g_hex)  # pylint: disable=invalid-name
+        self.k = self.__hex_to_long(self.__hex_hash('00' + self.n_hex + '0' + self.g_hex))  # pylint: disable=invalid-name
+        self.small_a_value = self.__generate_random_small_a()
+        self.large_a_value = self.__calculate_a()
+
+    def authenticate(self, username, password):
+        """ Authenticate with a username and password. """
+        # Step 1: First initiate an authentication request
+        auth_data_dict = self.__get_authentication_request(username)
+        auth_data = json.dumps(auth_data_dict).encode("utf-8")
+        auth_headers = {
+            "X-Amz-Target": "AWSCognitoIdentityProviderService.InitiateAuth",
+            "Accept-Encoding": "identity",
+            "Content-Type": "application/x-amz-json-1.1"
+        }
+        auth_request = compat_urllib_request.Request(url=self.url, data=auth_data, headers=auth_headers)
+        auth_response = compat_urllib_request.urlopen(auth_request).read()
+        auth_response_json = json.loads(auth_response.decode("utf-8"))
+        challenge_parameters = auth_response_json.get("ChallengeParameters")
+
+        challenge_name = auth_response_json.get("ChallengeName")
+        if not challenge_name == "PASSWORD_VERIFIER":
+            raise AuthenticationException(auth_response_json.get("message"))
+
+        # Step 2: Respond to the Challenge with a valid ChallengeResponse
+        challenge_request = self.__get_challenge_response_request(challenge_parameters, password)
+        challenge_data = json.dumps(challenge_request).encode("utf-8")
+        challenge_headers = {
+            "X-Amz-Target": "AWSCognitoIdentityProviderService.RespondToAuthChallenge",
+            "Content-Type": "application/x-amz-json-1.1"
+        }
+        auth_request = compat_urllib_request.Request(url=self.url, data=challenge_data, headers=challenge_headers)
+        auth_response = compat_urllib_request.urlopen(auth_request).read()
+        auth_response_json = json.loads(auth_response.decode("utf-8"))
+
+        if "message" in auth_response_json:
+            raise InvalidLoginException(auth_response_json.get("message"))
+
+        id_token = auth_response_json.get("AuthenticationResult", {}).get("IdToken")
+        refresh_token = auth_response_json.get("AuthenticationResult", {}).get("RefreshToken")
+        return id_token, refresh_token
+
+    def __get_authentication_request(self, username):
+        """
+
+        :param str username:    The username to use
+
+        :return: A full Authorization request.
+        :rtype: dict
+        """
+        auth_request = {
+            "AuthParameters": {
+                "USERNAME": username,
+                "SRP_A": self.__long_to_hex(self.large_a_value)
+            },
+            "AuthFlow": "USER_SRP_AUTH",
+            "ClientId": self.client_id
+        }
+        return auth_request
+
+    def __get_challenge_response_request(self, challenge_parameters, password):
+        """ Create a Challenge Response Request object.
+
+        :param dict[str,str|imt] challenge_parameters:  The parameters for the challenge.
+        :param str password:                            The password.
+
+        :return: A valid and full request data object to use as a response for a challenge.
+        :rtype: dict
+        """
+        user_id = challenge_parameters["USERNAME"]
+        user_id_for_srp = challenge_parameters["USER_ID_FOR_SRP"]
+        srp_b = challenge_parameters["SRP_B"]
+        salt = challenge_parameters["SALT"]
+        secret_block = challenge_parameters["SECRET_BLOCK"]
+
+        timestamp = self.__get_current_timestamp()
+
+        # Get a HKDF key for the password, SrpB and the Salt
+        hkdf = self.__get_hkdf_key_for_password(
+            user_id_for_srp,
+            password,
+            self.__hex_to_long(srp_b),
+            salt
+        )
+        secret_block_bytes = base64.standard_b64decode(secret_block)
+
+        # the message is a combo of the pool_id, provided SRP userId, the Secret and Timestamp
+        msg = \
+            bytearray(self.pool_id.split('_')[1], 'utf-8') + \
+            bytearray(user_id_for_srp, 'utf-8') + \
+            bytearray(secret_block_bytes) + \
+            bytearray(timestamp, 'utf-8')
+        hmac_obj = hmac.new(hkdf, msg, digestmod=hashlib.sha256)
+        signature_string = base64.standard_b64encode(hmac_obj.digest()).decode('utf-8')
+        challenge_request = {
+            "ChallengeResponses": {
+                "USERNAME": user_id,
+                "TIMESTAMP": timestamp,
+                "PASSWORD_CLAIM_SECRET_BLOCK": secret_block,
+                "PASSWORD_CLAIM_SIGNATURE": signature_string
+            },
+            "ChallengeName": "PASSWORD_VERIFIER",
+            "ClientId": self.client_id
+        }
+        return challenge_request
+
+    def __get_hkdf_key_for_password(self, username, password, server_b_value, salt):
+        """ Calculates the final hkdf based on computed S value, and computed U value and the key.
+
+        :param str username:        Username.
+        :param str password:        Password.
+        :param int server_b_value:  Server B value.
+        :param int salt:            Generated salt.
+
+        :return Computed HKDF value.
+        :rtype: object
+        """
+
+        u_value = self.__calculate_u(self.large_a_value, server_b_value)
+        if u_value == 0:
+            raise ValueError('U cannot be zero.')
+        username_password = '%s%s:%s' % (self.pool_id.split('_')[1], username, password)
+        username_password_hash = self.__hash_sha256(username_password.encode('utf-8'))
+
+        x_value = self.__hex_to_long(self.__hex_hash(self.__pad_hex(salt) + username_password_hash))
+        g_mod_pow_xn = pow(self.g, x_value, self.big_n)
+        int_value2 = server_b_value - self.k * g_mod_pow_xn
+        s_value = pow(int_value2, self.small_a_value + u_value * x_value, self.big_n)
+        hkdf = self.__compute_hkdf(
+            bytearray.fromhex(self.__pad_hex(s_value)),
+            bytearray.fromhex(self.__pad_hex(self.__long_to_hex(u_value)))
+        )
+        return hkdf
+
+    def __compute_hkdf(self, ikm, salt):
+        """ Standard hkdf algorithm
+
+        :param {Buffer} ikm Input key material.
+        :param {Buffer} salt Salt value.
+        :return {Buffer} Strong key material.
+        """
+
+        prk = hmac.new(salt, ikm, hashlib.sha256).digest()
+        info_bits_update = self.info_bits + bytearray(chr(1), 'utf-8')
+        hmac_hash = hmac.new(prk, info_bits_update, hashlib.sha256).digest()
+        return hmac_hash[:16]
+
+    def __calculate_u(self, big_a, big_b):
+        """ Calculate the client's value U which is the hash of A and B
+
+        :param int big_a:   Large A value.
+        :param int big_b:   Server B value.
+
+        :return Computed U value.
+        :rtype: int
+        """
+
+        u_hex_hash = self.__hex_hash(self.__pad_hex(big_a) + self.__pad_hex(big_b))
+        return self.__hex_to_long(u_hex_hash)
+
+    def __generate_random_small_a(self):
+        """ Helper function to generate a random big integer
+
+        :return a random value.
+        :rtype: int
+        """
+        random_long_int = self.__get_random(128)
+        return random_long_int % self.big_n
+
+    def __calculate_a(self):
+        """ Calculate the client's public value A = g^a%N with the generated random number a
+
+        :return Computed large A.
+        :rtype: int
+        """
+
+        big_a = pow(self.g, self.small_a_value, self.big_n)
+        # safety check
+        if (big_a % self.big_n) == 0:
+            raise ValueError('Safety check for A failed')
+        return big_a
+
+    @staticmethod
+    def __long_to_hex(long_num):
+        return '%x' % long_num
+
+    @staticmethod
+    def __hex_to_long(hex_string):
+        return int(hex_string, 16)
+
+    @staticmethod
+    def __hex_hash(hex_string):
+        return AwsIdp.__hash_sha256(bytearray.fromhex(hex_string))
+
+    @staticmethod
+    def __hash_sha256(buf):
+        """AuthenticationHelper.hash"""
+        digest = hashlib.sha256(buf).hexdigest()
+        return (64 - len(digest)) * '0' + digest
+
+    @staticmethod
+    def __pad_hex(long_int):
+        """ Converts a Long integer (or hex string) to hex format padded with zeroes for hashing
+
+        :param int|str long_int:    Number or string to pad.
+
+        :return Padded hex string.
+        :rtype: str
+        """
+
+        # noinspection PyTypeChecker
+        if not isinstance(long_int, compat_basestring):
+            hash_str = AwsIdp.__long_to_hex(long_int)
+        else:
+            hash_str = long_int
+        if len(hash_str) % 2 == 1:
+            hash_str = '0%s' % hash_str
+        elif hash_str[0] in '89ABCDEFabcdef':
+            hash_str = '00%s' % hash_str
+        return hash_str
+
+    @staticmethod
+    def __get_random(nbytes):
+        random_hex = binascii.hexlify(os.urandom(nbytes))
+        return AwsIdp.__hex_to_long(random_hex)
+
+    @staticmethod
+    def __get_current_timestamp():
+        """ Creates a timestamp with the correct English format.
+
+        :return: timestamp in format 'Sun Jan 27 19:00:04 UTC 2019'
+        :rtype: str
+        """
+
+        # We need US only data, so we cannot just do a strftime:
+        # Sun Jan 27 19:00:04 UTC 2019
+        months = [None, 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec']
+        days = ['Mon', 'Tue', 'Wed', 'Thu', 'Fri', 'Sat', 'Sun']
+
+        time_now = datetime.datetime.utcnow()
+        format_string = "{} {} {} %H:%M:%S UTC %Y".format(days[time_now.weekday()], months[time_now.month], time_now.day)
+        time_string = datetime.datetime.utcnow().strftime(format_string)
+        return time_string
+
+    def __str__(self):
+        return "AWS IDP Client for:\nRegion: %s\nPoolId: %s\nAppId:  %s" % (
+            self.region, self.pool_id.split("_")[1], self.client_id
+        )

youtube_dl/extractor/youtube.py

@@ -692,9 +692,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         'invidious': '|'.join(_INVIDIOUS_SITES),
     }
     _PLAYER_INFO_RE = (
-        r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})//(?:tv-)?player',
-        r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$',
-        r'\b(?P<id>vfl[a-zA-Z0-9_-]+)\b.*?\.js$',
+        r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})/(?:tv-)?player',
+        r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias(?:_tce)?\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$',
+        r'\b(?P<id>vfl[a-zA-Z0-9_-]{6,})\b.*?\.js$',
     )
     _SUBTITLE_FORMATS = ('json3', 'srv1', 'srv2', 'srv3', 'ttml', 'vtt')
 
@@ -1626,15 +1626,13 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         """ Return a string representation of a signature """
         return '.'.join(compat_str(len(part)) for part in example_sig.split('.'))
 
-    @classmethod
-    def _extract_player_info(cls, player_url):
-        for player_re in cls._PLAYER_INFO_RE:
-            id_m = re.search(player_re, player_url)
-            if id_m:
-                break
-        else:
-            raise ExtractorError('Cannot identify player %r' % player_url)
-        return id_m.group('id')
+    def _extract_player_info(self, player_url):
+        try:
+            return self._search_regex(
+                self._PLAYER_INFO_RE, player_url, 'player info', group='id')
+        except ExtractorError as e:
+            raise ExtractorError(
+                'Cannot identify player %r' % (player_url,), cause=e)
 
     def _load_player(self, video_id, player_url, fatal=True, player_id=None):
         if not player_id:
@@ -1711,6 +1709,23 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 '    return %s\n') % (signature_id_tuple, expr_code)
         self.to_screen('Extracted signature function:\n' + code)
 
+    def _extract_sig_fn(self, jsi, funcname):
+        var_ay = self._search_regex(
+            r'''(?x)
+                (?:\*/|\{|\n|^)\s*(?:'[^']+'\s*;\s*)
+                    (var\s*[\w$]+\s*=\s*(?:
+                        ('|")(?:\\\2|(?!\2).)+\2\s*\.\s*split\(\s*('|")\W+\3\s*\)|
+                        \[\s*(?:('|")(?:\\\4|(?!\4).)*\4\s*(?:(?=\])|,\s*))+\]
+                    ))(?=\s*[,;])
+            ''', jsi.code, 'useful values', default='')
+
+        sig_fn = jsi.extract_function_code(funcname)
+
+        if var_ay:
+            sig_fn = (sig_fn[0], ';\n'.join((var_ay, sig_fn[1])))
+
+        return sig_fn
+
     def _parse_sig_js(self, jscode):
         # Examples where `sig` is funcname:
         # sig=function(a){a=a.split(""); ... ;return a.join("")};
@@ -1736,8 +1751,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             jscode, 'Initial JS player signature function name', group='sig')
 
         jsi = JSInterpreter(jscode)
-        initial_function = jsi.extract_function(funcname)
-        return lambda s: initial_function([s])
+
+        initial_function = self._extract_sig_fn(jsi, funcname)
+
+        func = jsi.extract_function_from_code(*initial_function)
+
+        return lambda s: func([s])
 
     def _cached(self, func, *cache_id):
         def inner(*args, **kwargs):
@@ -1856,15 +1875,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
 
     def _extract_n_function_code_jsi(self, video_id, jsi, player_id=None):
 
-        var_ay = self._search_regex(
-            r'(?:[;\s]|^)\s*(var\s*[\w$]+\s*=\s*"(?:\\"|[^"])+"\s*\.\s*split\("\W+"\))(?=\s*[,;])',
-            jsi.code, 'useful values', default='')
-
         func_name = self._extract_n_function_name(jsi.code)
 
-        func_code = jsi.extract_function_code(func_name)
-        if var_ay:
-            func_code = (func_code[0], ';\n'.join((var_ay, func_code[1])))
+        func_code = self._extract_sig_fn(jsi, func_name)
 
         if player_id:
             self.cache.store('youtube-nsig', player_id, func_code)
@@ -2136,7 +2149,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     video_details = merge_dicts(*traverse_obj(
                         (player_response, api_player_response),
                         (Ellipsis, 'videoDetails', T(dict))))
-                    player_response.update(api_player_response or {})
+                    player_response.update(filter_dict(
+                        api_player_response or {}, cndn=lambda k, _: k != 'captions'))
                     player_response['videoDetails'] = video_details
 
         def is_agegated(playability):
@@ -2566,8 +2580,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         }
 
         pctr = traverse_obj(
-            player_response,
-            ('captions', 'playerCaptionsTracklistRenderer', T(dict)))
+            (player_response, api_player_response),
+            (Ellipsis, 'captions', 'playerCaptionsTracklistRenderer', T(dict)))
         if pctr:
             def process_language(container, base_url, lang_code, query):
                 lang_subs = []
@@ -2584,20 +2598,21 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             def process_subtitles():
                 subtitles = {}
                 for caption_track in traverse_obj(pctr, (
-                        'captionTracks', lambda _, v: v.get('baseUrl'))):
+                        Ellipsis, 'captionTracks', lambda _, v: (
+                            v.get('baseUrl') and v.get('languageCode')))):
                     base_url = self._yt_urljoin(caption_track['baseUrl'])
                     if not base_url:
                         continue
+                    lang_code = caption_track['languageCode']
                     if caption_track.get('kind') != 'asr':
-                        lang_code = caption_track.get('languageCode')
-                        if not lang_code:
-                            continue
                         process_language(
                             subtitles, base_url, lang_code, {})
                         continue
                     automatic_captions = {}
+                    process_language(
+                        automatic_captions, base_url, lang_code, {})
                     for translation_language in traverse_obj(pctr, (
-                            'translationLanguages', lambda _, v: v.get('languageCode'))):
+                            Ellipsis, 'translationLanguages', lambda _, v: v.get('languageCode'))):
                         translation_language_code = translation_language['languageCode']
                         process_language(
                             automatic_captions, base_url, translation_language_code,

youtube_dl/jsinterp.py

@@ -678,7 +678,7 @@ class JSInterpreter(object):
             return len(obj)
         try:
             return obj[int(idx)] if isinstance(obj, list) else obj[compat_str(idx)]
-        except (TypeError, KeyError, IndexError) as e:
+        except (TypeError, KeyError, IndexError, ValueError) as e:
             # allow_undefined is None gives correct behaviour
             if allow_undefined or (
                     allow_undefined is None and not isinstance(e, TypeError)):
@@ -1038,6 +1038,10 @@ class JSInterpreter(object):
                     left_val = self._index(left_val, idx)
             if isinstance(idx, float):
                 idx = int(idx)
+            if isinstance(left_val, list) and len(left_val) <= int_or_none(idx, default=-1):
+                # JS Array is a sparsely assignable list
+                # TODO: handle extreme sparsity without memory bloat, eg using auxiliary dict
+                left_val.extend((idx - len(left_val) + 1) * [JS_Undefined])
             left_val[idx] = self._operator(
                 m.group('op'), self._index(left_val, idx) if m.group('op') else None,
                 m.group('expr'), expr, local_vars, allow_recursion)
@@ -1204,9 +1208,10 @@ class JSInterpreter(object):
                 elif member == 'join':
                     assertion(isinstance(obj, list), 'must be applied on a list')
                     assertion(len(argvals) <= 1, 'takes at most one argument')
-                    return (',' if len(argvals) == 0 else argvals[0]).join(
-                        ('' if x in (None, JS_Undefined) else _js_toString(x))
-                        for x in obj)
+                    return (',' if len(argvals) == 0 or argvals[0] in (None, JS_Undefined)
+                            else argvals[0]).join(
+                                ('' if x in (None, JS_Undefined) else _js_toString(x))
+                                for x in obj)
                 elif member == 'reverse':
                     assertion(not argvals, 'does not take any arguments')
                     obj.reverse()
@@ -1364,19 +1369,21 @@ class JSInterpreter(object):
         code, _ = self._separate_at_paren(func_m.group('code'))  # refine the match
         return self.build_arglist(func_m.group('args')), code
 
-    def extract_function(self, funcname):
+    def extract_function(self, funcname, *global_stack):
         return function_with_repr(
-            self.extract_function_from_code(*self.extract_function_code(funcname)),
+            self.extract_function_from_code(*itertools.chain(
+                self.extract_function_code(funcname), global_stack)),
             'F<%s>' % (funcname,))
 
     def extract_function_from_code(self, argnames, code, *global_stack):
         local_vars = {}
 
+        start = None
         while True:
-            mobj = re.search(r'function\((?P<args>[^)]*)\)\s*{', code)
+            mobj = re.search(r'function\((?P<args>[^)]*)\)\s*{', code[start:])
             if mobj is None:
                 break
-            start, body_start = mobj.span()
+            start, body_start = ((start or 0) + x for x in mobj.span())
             body, remaining = self._separate_at_paren(code[body_start - 1:])
             name = self._named_object(local_vars, self.extract_function_from_code(
                 [x.strip() for x in mobj.group('args').split(',')],