Mirrors
/
yt-dlp
mirror of https://github.com/yt-dlp/yt-dlp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

[AbemaTV] Add extractors (#1688) 

Authored by: Lesmiscore
This commit is contained in:
Lesmiscore (Naoya Ozaki) 2022-02-25 11:14:04 +09:00 committed by GitHub
parent a539f06570
commit 3e9b66d761
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 515 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

488
yt_dlp/extractor/abematv.py Normal file
View File

@ -0,0 +1,488 @@
import io
import json
import time
import hashlib
import hmac
import re
import struct
from base64 import urlsafe_b64encode
from binascii import unhexlify
import typing
if typing.TYPE_CHECKING:
from ..YoutubeDL import YoutubeDL
from .common import InfoExtractor
from ..aes import aes_ecb_decrypt
from ..compat import (
compat_urllib_response,
compat_urllib_parse_urlparse,
compat_urllib_request,
)
from ..utils import (
ExtractorError,
decode_base,
int_or_none,
random_uuidv4,
request_to_url,
time_seconds,
update_url_query,
traverse_obj,
intlist_to_bytes,
bytes_to_intlist,
urljoin,
)
# NOTE: network handler related code is temporary thing until network stack overhaul PRs are merged (#2861/#2862)
def add_opener(self: 'YoutubeDL', 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(self._opener, compat_urllib_request.OpenerDirector)
self._opener.add_handler(handler)
def remove_opener(self: 'YoutubeDL', 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 = self._opener
assert isinstance(self._opener, compat_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(compat_urllib_request.BaseHandler):
handler_order = 499
STRTABLE = '123456789ABCDEFGHJKLMNPQRSTUVWXYZabcdefghijkmnopqrstuvwxyz'
HKEY = b'3AF0298C219469522A313570E8583005A642E73EDD58E3EA2FB7339D3DF1597E'
def __init__(self, ie: 'AbemaTVIE'):
# the protcol 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._downloader.params.get('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(license_response['k'], self.STRTABLE)
encvideokey = bytes_to_intlist(struct.pack('>QQ', res >> 64, res & 0xffffffffffffffff))
h = hmac.new(
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 = compat_urllib_parse_urlparse(url).netloc
response_data = self._get_videokey_from_ticket(ticket)
return compat_urllib_response.addinfourl(io.BytesIO(response_data), headers={
'Content-Length': len(response_data),
}, url=url, code=200)
class AbemaTVBaseIE(InfoExtractor):
def _extract_breadcrumb_list(self, webpage, video_id):
for jld in re.finditer(
r'(?is)</span></li></ul><script[^>]+type=(["\']?)application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
webpage):
jsonld = self._parse_json(jld.group('json_ld'), video_id, fatal=False)
if jsonld:
if jsonld.get('@type') != 'BreadcrumbList':
continue
trav = traverse_obj(jsonld, ('itemListElement', ..., 'name'))
if trav:
return trav
return []
class AbemaTVIE(AbemaTVBaseIE):
_VALID_URL = r'https?://abema\.tv/(?P<type>now-on-air|video/episode|channels/.+?/slots)/(?P<id>[^?/]+)'
_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',
}]
_USERTOKEN = None
_DEVICE_ID = None
_TIMETABLE = None
_MEDIATOKEN = None
_SECRETKEY = b'v+Gjs=25Aw5erR!J8ZuvRrCx*rGswhB&qdHd_SYerEWdU&a?3DzN9BRbp5KwY4hEmcj5#fykMjJ=AuWz5GSMY-d@H7DMEh3M@9n2G552Us$$k9cD=3TxwWe86!x#Zyhe'
def _generate_aks(self, 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(self._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(urlsafe_b64encode(tmp).rstrip(b'=') + nonce)
mix_once(self._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 urlsafe_b64encode(tmp).rstrip(b'=').decode('utf-8')
def _get_device_token(self):
if self._USERTOKEN:
return self._USERTOKEN
self._DEVICE_ID = random_uuidv4()
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',
})
self._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
self._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': 'bearer ' + self._get_device_token()
})['token']
return self._MEDIATOKEN
def _real_initialize(self):
self._login()
def _login(self):
username, password = self._get_login_info()
# No authentication to be performed
if not username:
return True
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': 'bearer ' + self._get_device_token(),
'Origin': 'https://abema.tv',
'Referer': 'https://abema.tv/',
'Content-Type': 'application/json',
})
self._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 specifing 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'<link\s+rel="canonical"\s*href="(.+?)"', webpage, 'canonical URL',
default=url)
info = self._search_json_ld(webpage, video_id, default={})
title = self._search_regex(
r'<span\s*class=".+?EpisodeTitleBlock__title">(.+?)</span>', webpage, 'title', default=None)
if not title:
jsonld = None
for jld in re.finditer(
r'(?is)<span\s*class="com-m-Thumbnail__image">(?:</span>)?<script[^>]+type=(["\']?)application/ld\+json\1[^>]*>(?P<json_ld>.+?)</script>',
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: (example is 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'<p\s+class="com-video-EpisodeDetailsBlock__content"><span\s+class=".+?">(.+?)</span></p><div',
r'<span\s+class=".+?SlotSummary.+?">(.+?)</span></div><div',),
webpage, 'description', default=None, group=1)
if not description:
og_desc = self._html_search_meta(
('description', 'og:description', 'twitter:description'), webpage)
if og_desc:
description = re.sub(r'''(?sx)
^(.+?)(?:
アニメの動画を無料で見るならABEMA！| # anime
等、.+ # applies for most of categories
)?
''', r'\1', og_desc)
# canonical URL may contain series and episode number
mobj = re.search(r's(\d+)_p(\d+)$', canonical_url)
if mobj:
seri = int_or_none(mobj.group(1), default=float('inf'))
epis = int_or_none(mobj.group(2), default=float('inf'))
info['series_number'] = seri if seri < 100 else None
# some anime like Detective Conan (though not available in AbemaTV)
# has more than 1000 episodes (1026 as of 2021/11/15)
info['episode_number'] = epis if epis < 2000 else None
is_live, m3u8_url = False, None
if video_type == 'now-on-air':
is_live = True
channel_url = 'https://api.abema.io/v1/channels'
if video_id == 'news-global':
channel_url = update_url_query(channel_url, {'division': '1'})
onair_channels = self._download_json(channel_url, video_id)
for ch in onair_channels['channels']:
if video_id == ch['id']:
m3u8_url = ch['playback']['hls']
break
else:
raise ExtractorError(f'Cannot find on-air {video_id} channel.', expected=True)
elif video_type == 'episode':
api_response = self._download_json(
f'https://api.abema.io/v1/video/programs/{video_id}', video_id,
note='Checking playability',
headers=headers)
ondemand_types = traverse_obj(api_response, ('terms', ..., 'onDemandType'), default=[])
if 3 not in ondemand_types:
# cannot acquire decryption key for these streams
self.report_warning('This is a premium-only stream')
m3u8_url = f'https://vod-abematv.akamaized.net/program/{video_id}/playlist.m3u8'
elif video_type == 'slots':
api_response = self._download_json(
f'https://api.abema.io/v1/media/slots/{video_id}', video_id,
note='Checking playability',
headers=headers)
if not traverse_obj(api_response, ('slot', 'flags', 'timeshiftFree'), default=False):
self.report_warning('This is a premium-only stream')
m3u8_url = f'https://vod-abematv.akamaized.net/slot/{video_id}/playlist.m3u8'
else:
raise ExtractorError('Unreachable')
if is_live:
self.report_warning("This is a livestream; yt-dlp doesn't support downloading natively, but FFmpeg cannot handle m3u8 manifests from AbemaTV")
self.report_warning('Please consider using Streamlink to download these streams (https://github.com/streamlink/streamlink)')
formats = self._extract_m3u8_formats(
m3u8_url, video_id, ext='mp4', live=is_live)
info.update({
'id': video_id,
'title': title,
'description': description,
'formats': formats,
'is_live': is_live,
})
return info
class AbemaTVTitleIE(AbemaTVBaseIE):
_VALID_URL = r'https?://abema\.tv/video/title/(?P<id>[^?/]+)'
_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,
}]
def _real_extract(self, url):
video_id = self._match_id(url)
webpage = self._download_webpage(url, video_id)
playlist_title, breadcrumb = None, self._extract_breadcrumb_list(webpage, video_id)
if breadcrumb:
playlist_title = breadcrumb[-1]
playlist = [
self.url_result(urljoin('https://abema.tv/', mobj.group(1)))
for mobj in re.finditer(r'<li\s*class=".+?EpisodeList.+?"><a\s*href="(/[^"]+?)"', webpage)]
return self.playlist_result(playlist, playlist_title=playlist_title, playlist_id=video_id)

4
yt_dlp/extractor/extractors.py
View File

@ -14,6 +14,10 @@ from .abcotvs import (
ABCOTVSIE,
ABCOTVSClipsIE,
)
from .abematv import (
AbemaTVIE,
AbemaTVTitleIE,
)
from .academicearth import AcademicEarthCourseIE
from .acast import (
ACastIE,

23
yt_dlp/utils.py
View File

@ -2554,6 +2554,13 @@ def url_or_none(url):
return url if re.match(r'^(?:(?:https?|rt(?:m(?:pt?[es]?|fp)|sp[su]?)|mms|ftps?):)?//', url) else None
def request_to_url(req):
if isinstance(req, compat_urllib_request.Request):
return req.get_full_url()
else:
return req
def strftime_or_none(timestamp, date_format, default=None):
datetime_object = None
try:
@ -5172,6 +5179,22 @@ def variadic(x, allowed_types=(str, bytes, dict)):
return x if isinstance(x, collections.abc.Iterable) and not isinstance(x, allowed_types) else (x,)
def decode_base(value, digits):
# This will convert given base-x string to scalar (long or int)
table = {char: index for index, char in enumerate(digits)}
result = 0
base = len(digits)
for chr in value:
result *= base
result += table[chr]
return result
def time_seconds(**kwargs):
t = datetime.datetime.now(datetime.timezone(datetime.timedelta(**kwargs)))
return t.timestamp()
# create a JSON Web Signature (jws) with HS256 algorithm
# the resulting format is in JWS Compact Serialization
# implemented following JWT https://www.rfc-editor.org/rfc/rfc7519.html