[openrec] Add extractors (#624)

Authored by: nao20010128nao
This commit is contained in:
The Hatsune Daishi 2021-08-04 18:14:37 +09:00 committed by GitHub
parent 3e376d183e
commit 9997eee4af
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
3 changed files with 131 additions and 1 deletions

View file

@ -88,7 +88,7 @@ # NEW FEATURES
* **Aria2c with HLS/DASH**: You can use `aria2c` as the external downloader for DASH(mpd) and HLS(m3u8) formats
* **New extractors**: AnimeLab, Philo MSO, Spectrum MSO, SlingTV MSO, Rcs, Gedi, bitwave.tv, mildom, audius, zee5, mtv.it, wimtv, pluto.tv, niconico users, discoveryplus.in, mediathek, NFHSNetwork, nebula, ukcolumn, whowatch, MxplayerShow, parlview (au), YoutubeWebArchive, fancode, Saitosan, ShemarooMe, telemundo, VootSeries, SonyLIVSeries, HotstarSeries, VidioPremier, VidioLive, RCTIPlus, TBS Live, douyin, pornflip, ParamountPlusSeries, ScienceChannel, Utreon
* **New extractors**: AnimeLab, Philo MSO, Spectrum MSO, SlingTV MSO, Rcs, Gedi, bitwave.tv, mildom, audius, zee5, mtv.it, wimtv, pluto.tv, niconico users, discoveryplus.in, mediathek, NFHSNetwork, nebula, ukcolumn, whowatch, MxplayerShow, parlview (au), YoutubeWebArchive, fancode, Saitosan, ShemarooMe, telemundo, VootSeries, SonyLIVSeries, HotstarSeries, VidioPremier, VidioLive, RCTIPlus, TBS Live, douyin, pornflip, ParamountPlusSeries, ScienceChannel, Utreon, OpenRec
* **Fixed/improved extractors**: archive.org, roosterteeth.com, skyit, instagram, itv, SouthparkDe, spreaker, Vlive, akamai, ina, rumble, tennistv, amcnetworks, la7 podcasts, linuxacadamy, nitter, twitcasting, viu, crackle, curiositystream, mediasite, rmcdecouverte, sonyliv, tubi, tenplay, patreon, videa, yahoo, BravoTV, crunchyroll playlist, RTP, viki, Hotstar, vidio, vimeo, mediaset, Mxplayer

View file

@ -936,6 +936,10 @@
OoyalaIE,
OoyalaExternalIE,
)
from .openrec import (
OpenRecIE,
OpenRecCaptureIE,
)
from .ora import OraTVIE
from .orf import (
ORFTVthekIE,

126
yt_dlp/extractor/openrec.py Normal file
View file

@ -0,0 +1,126 @@
# coding: utf-8
from __future__ import unicode_literals
from .common import InfoExtractor
from ..utils import (
ExtractorError,
traverse_obj,
try_get,
unified_strdate
)
from ..compat import compat_str
class OpenRecIE(InfoExtractor):
IE_NAME = 'openrec'
_VALID_URL = r'https?://(?:www\.)?openrec\.tv/live/(?P<id>[^/]+)'
_TESTS = [{
'url': 'https://www.openrec.tv/live/2p8v31qe4zy',
'only_matching': True,
}, {
'url': 'https://www.openrec.tv/live/wez93eqvjzl',
'only_matching': True,
}]
def _real_extract(self, url):
video_id = self._match_id(url)
webpage = self._download_webpage('https://www.openrec.tv/live/%s' % video_id, video_id)
window_stores = self._parse_json(
self._search_regex(r'(?m)window\.pageStore\s*=\s*(\{.+?\});$', webpage, 'window.pageStore'), video_id)
movie_store = traverse_obj(
window_stores,
('v8', 'state', 'movie'),
('v8', 'movie'),
expected_type=dict)
if not movie_store:
raise ExtractorError('Failed to extract live info')
title = movie_store.get('title')
description = movie_store.get('introduction')
thumbnail = movie_store.get('thumbnailUrl')
channel_user = movie_store.get('channel', {}).get('user')
uploader = try_get(channel_user, lambda x: x['name'], compat_str)
uploader_id = try_get(channel_user, lambda x: x['id'], compat_str)
timestamp = traverse_obj(movie_store, ('startedAt', 'time'), expected_type=int)
m3u8_playlists = movie_store.get('media')
formats = []
for (name, m3u8_url) in m3u8_playlists.items():
if not m3u8_url:
continue
formats.extend(self._extract_m3u8_formats(
m3u8_url, video_id, ext='mp4', entry_protocol='m3u8',
m3u8_id='hls-%s' % name, live=True))
self._sort_formats(formats)
return {
'id': video_id,
'title': title,
'description': description,
'thumbnail': thumbnail,
'formats': formats,
'uploader': uploader,
'uploader_id': uploader_id,
'timestamp': timestamp,
'is_live': True,
}
class OpenRecCaptureIE(InfoExtractor):
IE_NAME = 'openrec:capture'
_VALID_URL = r'https?://(?:www\.)?openrec\.tv/capture/(?P<id>[^/]+)'
_TESTS = [{
'url': 'https://www.openrec.tv/capture/l9nk2x4gn14',
'only_matching': True,
}, {
'url': 'https://www.openrec.tv/capture/mldjr82p7qk',
'info_dict': {
'id': 'mldjr82p7qk',
'title': 'たいじの恥ずかしい英語力',
'uploader': 'たいちゃんねる',
'uploader_id': 'Yaritaiji',
'upload_date': '20210803',
},
}]
def _real_extract(self, url):
video_id = self._match_id(url)
webpage = self._download_webpage('https://www.openrec.tv/capture/%s' % video_id, video_id)
window_stores = self._parse_json(
self._search_regex(r'(?m)window\.pageStore\s*=\s*(\{.+?\});$', webpage, 'window.pageStore'), video_id)
movie_store = window_stores.get('movie')
capture_data = window_stores.get('capture')
if not capture_data:
raise ExtractorError('Cannot extract title')
title = capture_data.get('title')
thumbnail = capture_data.get('thumbnailUrl')
upload_date = unified_strdate(capture_data.get('createdAt'))
channel_info = movie_store.get('channel') or {}
uploader = channel_info.get('name')
uploader_id = channel_info.get('id')
m3u8_url = capture_data.get('source')
if not m3u8_url:
raise ExtractorError('Cannot extract m3u8 url')
formats = self._extract_m3u8_formats(
m3u8_url, video_id, ext='mp4', entry_protocol='m3u8_native',
m3u8_id='hls')
self._sort_formats(formats)
return {
'id': video_id,
'title': title,
'thumbnail': thumbnail,
'formats': formats,
'uploader': uploader,
'uploader_id': uploader_id,
'upload_date': upload_date,
}