1
0
mirror of https://codeberg.org/polarisfm/youtube-dl synced 2024-11-27 02:44:33 +01:00
youtube-dl/youtube_dl/extractor/imggaming.py

134 lines
5.0 KiB
Python
Raw Normal View History

# coding: utf-8
from __future__ import unicode_literals
import json
import re
from .common import InfoExtractor
from ..compat import compat_HTTPError
from ..utils import (
ExtractorError,
int_or_none,
str_or_none,
try_get,
)
class ImgGamingBaseIE(InfoExtractor):
_API_BASE = 'https://dce-frontoffice.imggaming.com/api/v2/'
_API_KEY = '857a1e5d-e35e-4fdf-805b-a87b6f8364bf'
_HEADERS = None
_MANIFEST_HEADERS = {'Accept-Encoding': 'identity'}
_REALM = None
_VALID_URL_TEMPL = r'https?://(?P<domain>%s)/(?P<type>live|playlist|video)/(?P<id>\d+)(?:\?.*?\bplaylistId=(?P<playlist_id>\d+))?'
def _real_initialize(self):
self._HEADERS = {
'Realm': 'dce.' + self._REALM,
'x-api-key': self._API_KEY,
}
email, password = self._get_login_info()
if email is None:
self.raise_login_required()
p_headers = self._HEADERS.copy()
p_headers['Content-Type'] = 'application/json'
self._HEADERS['Authorization'] = 'Bearer ' + self._download_json(
self._API_BASE + 'login',
None, 'Logging in', data=json.dumps({
'id': email,
'secret': password,
}).encode(), headers=p_headers)['authorisationToken']
def _call_api(self, path, media_id):
return self._download_json(
self._API_BASE + path + media_id, media_id, headers=self._HEADERS)
def _extract_dve_api_url(self, media_id, media_type):
stream_path = 'stream'
if media_type == 'video':
stream_path += '/vod/'
else:
stream_path += '?eventId='
try:
return self._call_api(
stream_path, media_id)['playerUrlCallback']
except ExtractorError as e:
if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
raise ExtractorError(
self._parse_json(e.cause.read().decode(), media_id)['messages'][0],
expected=True)
raise
def _real_extract(self, url):
domain, media_type, media_id, playlist_id = re.match(self._VALID_URL, url).groups()
if playlist_id:
if self._downloader.params.get('noplaylist'):
self.to_screen('Downloading just video %s because of --no-playlist' % media_id)
else:
self.to_screen('Downloading playlist %s - add --no-playlist to just download video' % playlist_id)
media_type, media_id = 'playlist', playlist_id
if media_type == 'playlist':
playlist = self._call_api('vod/playlist/', media_id)
entries = []
for video in try_get(playlist, lambda x: x['videos']['vods']) or []:
video_id = str_or_none(video.get('id'))
if not video_id:
continue
entries.append(self.url_result(
'https://%s/video/%s' % (domain, video_id),
self.ie_key(), video_id))
return self.playlist_result(
entries, media_id, playlist.get('title'),
playlist.get('description'))
dve_api_url = self._extract_dve_api_url(media_id, media_type)
video_data = self._download_json(dve_api_url, media_id)
is_live = media_type == 'live'
if is_live:
title = self._live_title(self._call_api('event/', media_id)['title'])
else:
title = video_data['name']
formats = []
for proto in ('hls', 'dash'):
media_url = video_data.get(proto + 'Url') or try_get(video_data, lambda x: x[proto]['url'])
if not media_url:
continue
if proto == 'hls':
m3u8_formats = self._extract_m3u8_formats(
media_url, media_id, 'mp4', 'm3u8' if is_live else 'm3u8_native',
m3u8_id='hls', fatal=False, headers=self._MANIFEST_HEADERS)
for f in m3u8_formats:
f.setdefault('http_headers', {}).update(self._MANIFEST_HEADERS)
formats.append(f)
else:
formats.extend(self._extract_mpd_formats(
media_url, media_id, mpd_id='dash', fatal=False,
headers=self._MANIFEST_HEADERS))
self._sort_formats(formats)
subtitles = {}
for subtitle in video_data.get('subtitles', []):
subtitle_url = subtitle.get('url')
if not subtitle_url:
continue
subtitles.setdefault(subtitle.get('lang', 'en_US'), []).append({
'url': subtitle_url,
})
return {
'id': media_id,
'title': title,
'formats': formats,
'thumbnail': video_data.get('thumbnailUrl'),
'description': video_data.get('description'),
'duration': int_or_none(video_data.get('duration')),
'tags': video_data.get('tags'),
'is_live': is_live,
'subtitles': subtitles,
}