import base64 import binascii import functools import hashlib import hmac import io import json import re import struct import time import urllib.parse import urllib.request import urllib.response import uuid from .common import InfoExtractor from ..aes import aes_ecb_decrypt from ..utils import ( ExtractorError, bytes_to_intlist, decode_base_n, int_or_none, intlist_to_bytes, OnDemandPagedList, request_to_url, time_seconds, traverse_obj, update_url_query, ) # NOTE: network handler related code is temporary thing until network stack overhaul PRs are merged (#2861/#2862) def add_opener(ydl, handler): ''' Add a handler for opening URLs, like _download_webpage ''' # https://github.com/python/cpython/blob/main/Lib/urllib/request.py#L426 # https://github.com/python/cpython/blob/main/Lib/urllib/request.py#L605 assert isinstance(ydl._opener, urllib.request.OpenerDirector) ydl._opener.add_handler(handler) def remove_opener(ydl, handler): ''' Remove handler(s) for opening URLs @param handler Either handler object itself or handler type. Specifying handler type will remove all handler which isinstance returns True. ''' # https://github.com/python/cpython/blob/main/Lib/urllib/request.py#L426 # https://github.com/python/cpython/blob/main/Lib/urllib/request.py#L605 opener = ydl._opener assert isinstance(ydl._opener, urllib.request.OpenerDirector) if isinstance(handler, (type, tuple)): find_cp = lambda x: isinstance(x, handler) else: find_cp = lambda x: x is handler removed = [] for meth in dir(handler): if meth in ["redirect_request", "do_open", "proxy_open"]: # oops, coincidental match continue i = meth.find("_") protocol = meth[:i] condition = meth[i + 1:] if condition.startswith("error"): j = condition.find("_") + i + 1 kind = meth[j + 1:] try: kind = int(kind) except ValueError: pass lookup = opener.handle_error.get(protocol, {}) opener.handle_error[protocol] = lookup elif condition == "open": kind = protocol lookup = opener.handle_open elif condition == "response": kind = protocol lookup = opener.process_response elif condition == "request": kind = protocol lookup = opener.process_request else: continue handlers = lookup.setdefault(kind, []) if handlers: handlers[:] = [x for x in handlers if not find_cp(x)] removed.append(x for x in handlers if find_cp(x)) if removed: for x in opener.handlers: if find_cp(x): x.add_parent(None) opener.handlers[:] = [x for x in opener.handlers if not find_cp(x)] class AbemaLicenseHandler(urllib.request.BaseHandler): handler_order = 499 STRTABLE = '123456789ABCDEFGHJKLMNPQRSTUVWXYZabcdefghijkmnopqrstuvwxyz' HKEY = b'3AF0298C219469522A313570E8583005A642E73EDD58E3EA2FB7339D3DF1597E' def __init__(self, ie: 'AbemaTVIE'): # the protocol that this should really handle is 'abematv-license://' # abematv_license_open is just a placeholder for development purposes # ref. https://github.com/python/cpython/blob/f4c03484da59049eb62a9bf7777b963e2267d187/Lib/urllib/request.py#L510 setattr(self, 'abematv-license_open', getattr(self, 'abematv_license_open')) self.ie = ie def _get_videokey_from_ticket(self, ticket): to_show = self.ie.get_param('verbose', False) media_token = self.ie._get_media_token(to_show=to_show) license_response = self.ie._download_json( 'https://license.abema.io/abematv-hls', None, note='Requesting playback license' if to_show else False, query={'t': media_token}, data=json.dumps({ 'kv': 'a', 'lt': ticket }).encode('utf-8'), headers={ 'Content-Type': 'application/json', }) res = decode_base_n(license_response['k'], table=self.STRTABLE) encvideokey = bytes_to_intlist(struct.pack('>QQ', res >> 64, res & 0xffffffffffffffff)) h = hmac.new( binascii.unhexlify(self.HKEY), (license_response['cid'] + self.ie._DEVICE_ID).encode('utf-8'), digestmod=hashlib.sha256) enckey = bytes_to_intlist(h.digest()) return intlist_to_bytes(aes_ecb_decrypt(encvideokey, enckey)) def abematv_license_open(self, url): url = request_to_url(url) ticket = urllib.parse.urlparse(url).netloc response_data = self._get_videokey_from_ticket(ticket) return urllib.response.addinfourl(io.BytesIO(response_data), headers={ 'Content-Length': len(response_data), }, url=url, code=200) class AbemaTVBaseIE(InfoExtractor): _USERTOKEN = None _DEVICE_ID = None _MEDIATOKEN = None _SECRETKEY = b'v+Gjs=25Aw5erR!J8ZuvRrCx*rGswhB&qdHd_SYerEWdU&a?3DzN9BRbp5KwY4hEmcj5#fykMjJ=AuWz5GSMY-d@H7DMEh3M@9n2G552Us$$k9cD=3TxwWe86!x#Zyhe' @classmethod def _generate_aks(cls, deviceid): deviceid = deviceid.encode('utf-8') # add 1 hour and then drop minute and secs ts_1hour = int((time_seconds(hours=9) // 3600 + 1) * 3600) time_struct = time.gmtime(ts_1hour) ts_1hour_str = str(ts_1hour).encode('utf-8') tmp = None def mix_once(nonce): nonlocal tmp h = hmac.new(cls._SECRETKEY, digestmod=hashlib.sha256) h.update(nonce) tmp = h.digest() def mix_tmp(count): nonlocal tmp for i in range(count): mix_once(tmp) def mix_twist(nonce): nonlocal tmp mix_once(base64.urlsafe_b64encode(tmp).rstrip(b'=') + nonce) mix_once(cls._SECRETKEY) mix_tmp(time_struct.tm_mon) mix_twist(deviceid) mix_tmp(time_struct.tm_mday % 5) mix_twist(ts_1hour_str) mix_tmp(time_struct.tm_hour % 5) return base64.urlsafe_b64encode(tmp).rstrip(b'=').decode('utf-8') def _get_device_token(self): if self._USERTOKEN: return self._USERTOKEN AbemaTVBaseIE._DEVICE_ID = str(uuid.uuid4()) aks = self._generate_aks(self._DEVICE_ID) user_data = self._download_json( 'https://api.abema.io/v1/users', None, note='Authorizing', data=json.dumps({ 'deviceId': self._DEVICE_ID, 'applicationKeySecret': aks, }).encode('utf-8'), headers={ 'Content-Type': 'application/json', }) AbemaTVBaseIE._USERTOKEN = user_data['token'] # don't allow adding it 2 times or more, though it's guarded remove_opener(self._downloader, AbemaLicenseHandler) add_opener(self._downloader, AbemaLicenseHandler(self)) return self._USERTOKEN def _get_media_token(self, invalidate=False, to_show=True): if not invalidate and self._MEDIATOKEN: return self._MEDIATOKEN AbemaTVBaseIE._MEDIATOKEN = self._download_json( 'https://api.abema.io/v1/media/token', None, note='Fetching media token' if to_show else False, query={ 'osName': 'android', 'osVersion': '6.0.1', 'osLang': 'ja_JP', 'osTimezone': 'Asia/Tokyo', 'appId': 'tv.abema', 'appVersion': '3.27.1' }, headers={ 'Authorization': f'bearer {self._get_device_token()}', })['token'] return self._MEDIATOKEN def _call_api(self, endpoint, video_id, query=None, note='Downloading JSON metadata'): return self._download_json( f'https://api.abema.io/{endpoint}', video_id, query=query or {}, note=note, headers={ 'Authorization': f'bearer {self._get_device_token()}', }) def _extract_breadcrumb_list(self, webpage, video_id): for jld in re.finditer( r'(?is)]+type=(["\']?)application/ld\+json\1[^>]*>(?P.+?)', webpage): jsonld = self._parse_json(jld.group('json_ld'), video_id, fatal=False) if traverse_obj(jsonld, '@type') != 'BreadcrumbList': continue items = traverse_obj(jsonld, ('itemListElement', ..., 'name')) if items: return items return [] class AbemaTVIE(AbemaTVBaseIE): _VALID_URL = r'https?://abema\.tv/(?Pnow-on-air|video/episode|channels/.+?/slots)/(?P[^?/]+)' _NETRC_MACHINE = 'abematv' _TESTS = [{ 'url': 'https://abema.tv/video/episode/194-25_s2_p1', 'info_dict': { 'id': '194-25_s2_p1', 'title': '第1話 「チーズケーキ」 「モーニング再び」', 'series': '異世界食堂2', 'series_number': 2, 'episode': '第1話 「チーズケーキ」 「モーニング再び」', 'episode_number': 1, }, 'skip': 'expired', }, { 'url': 'https://abema.tv/channels/anime-live2/slots/E8tvAnMJ7a9a5d', 'info_dict': { 'id': 'E8tvAnMJ7a9a5d', 'title': 'ゆるキャン△ SEASON2 全話一挙【無料ビデオ72時間】', 'series': 'ゆるキャン△ SEASON2', 'episode': 'ゆるキャン△ SEASON2 全話一挙【無料ビデオ72時間】', 'series_number': 2, 'episode_number': 1, 'description': 'md5:9c5a3172ae763278f9303922f0ea5b17', }, 'skip': 'expired', }, { 'url': 'https://abema.tv/video/episode/87-877_s1282_p31047', 'info_dict': { 'id': 'E8tvAnMJ7a9a5d', 'title': '第5話『光射す』', 'description': 'md5:56d4fc1b4f7769ded5f923c55bb4695d', 'thumbnail': r're:https://hayabusa\.io/.+', 'series': '相棒', 'episode': '第5話『光射す』', }, 'skip': 'expired', }, { 'url': 'https://abema.tv/now-on-air/abema-anime', 'info_dict': { 'id': 'abema-anime', # this varies # 'title': '女子高生の無駄づかい 全話一挙【無料ビデオ72時間】', 'description': 'md5:55f2e61f46a17e9230802d7bcc913d5f', 'is_live': True, }, 'skip': 'Not supported until yt-dlp implements native live downloader OR AbemaTV can start a local HTTP server', }] _TIMETABLE = None def _perform_login(self, username, password): if '@' in username: # don't strictly check if it's email address or not ep, method = 'user/email', 'email' else: ep, method = 'oneTimePassword', 'userId' login_response = self._download_json( f'https://api.abema.io/v1/auth/{ep}', None, note='Logging in', data=json.dumps({ method: username, 'password': password }).encode('utf-8'), headers={ 'Authorization': f'bearer {self._get_device_token()}', 'Origin': 'https://abema.tv', 'Referer': 'https://abema.tv/', 'Content-Type': 'application/json', }) AbemaTVBaseIE._USERTOKEN = login_response['token'] self._get_media_token(True) def _real_extract(self, url): # starting download using infojson from this extractor is undefined behavior, # and never be fixed in the future; you must trigger downloads by directly specifying URL. # (unless there's a way to hook before downloading by extractor) video_id, video_type = self._match_valid_url(url).group('id', 'type') headers = { 'Authorization': 'Bearer ' + self._get_device_token(), } video_type = video_type.split('/')[-1] webpage = self._download_webpage(url, video_id) canonical_url = self._search_regex( r'(.+?)', webpage, 'title', default=None) if not title: jsonld = None for jld in re.finditer( r'(?is)(?:)?]+type=(["\']?)application/ld\+json\1[^>]*>(?P.+?)', webpage): jsonld = self._parse_json(jld.group('json_ld'), video_id, fatal=False) if jsonld: break if jsonld: title = jsonld.get('caption') if not title and video_type == 'now-on-air': if not self._TIMETABLE: # cache the timetable because it goes to 5MiB in size (!!) self._TIMETABLE = self._download_json( 'https://api.abema.io/v1/timetable/dataSet?debug=false', video_id, headers=headers) now = time_seconds(hours=9) for slot in self._TIMETABLE.get('slots', []): if slot.get('channelId') != video_id: continue if slot['startAt'] <= now and now < slot['endAt']: title = slot['title'] break # read breadcrumb on top of page breadcrumb = self._extract_breadcrumb_list(webpage, video_id) if breadcrumb: # breadcrumb list translates to: (e.g. 1st test for this IE) # Home > Anime (genre) > Isekai Shokudo 2 (series name) > Episode 1 "Cheese cakes" "Morning again" (episode title) # hence this works info['series'] = breadcrumb[-2] info['episode'] = breadcrumb[-1] if not title: title = info['episode'] description = self._html_search_regex( (r'(.+?)

(.+?)[^?/]+)' _PAGE_SIZE = 25 _TESTS = [{ 'url': 'https://abema.tv/video/title/90-1597', 'info_dict': { 'id': '90-1597', 'title': 'シャッフルアイランド', }, 'playlist_mincount': 2, }, { 'url': 'https://abema.tv/video/title/193-132', 'info_dict': { 'id': '193-132', 'title': '真心が届く~僕とスターのオフィス・ラブ!?~', }, 'playlist_mincount': 16, }, { 'url': 'https://abema.tv/video/title/25-102', 'info_dict': { 'id': '25-102', 'title': 'ソードアート・オンライン アリシゼーション', }, 'playlist_mincount': 24, }] def _fetch_page(self, playlist_id, series_version, page): programs = self._call_api( f'v1/video/series/{playlist_id}/programs', playlist_id, note=f'Downloading page {page + 1}', query={ 'seriesVersion': series_version, 'offset': str(page * self._PAGE_SIZE), 'order': 'seq', 'limit': str(self._PAGE_SIZE), }) yield from ( self.url_result(f'https://abema.tv/video/episode/{x}') for x in traverse_obj(programs, ('programs', ..., 'id'))) def _entries(self, playlist_id, series_version): return OnDemandPagedList( functools.partial(self._fetch_page, playlist_id, series_version), self._PAGE_SIZE) def _real_extract(self, url): playlist_id = self._match_id(url) series_info = self._call_api(f'v1/video/series/{playlist_id}', playlist_id) return self.playlist_result( self._entries(playlist_id, series_info['version']), playlist_id=playlist_id, playlist_title=series_info.get('title'), playlist_description=series_info.get('content'))