[daystar] Add extractor (#2924)
Closes #2887 Authored by; hatienl0i261299
This commit is contained in:
parent
5bcccbfec3
commit
761fba6d22
2 changed files with 49 additions and 0 deletions
48
yt_dlp/extractor/daystar.py
Normal file
48
yt_dlp/extractor/daystar.py
Normal file
|
@ -0,0 +1,48 @@
|
|||
from .common import InfoExtractor
|
||||
from ..utils import js_to_json, urljoin
|
||||
|
||||
|
||||
class DaystarClipIE(InfoExtractor):
|
||||
IE_NAME = 'daystar:clip'
|
||||
_VALID_URL = r'https?://player\.daystar\.tv/(?P<id>\w+)'
|
||||
_TESTS = [{
|
||||
'url': 'https://player.daystar.tv/0MTO2ITM',
|
||||
'info_dict': {
|
||||
'id': '0MTO2ITM',
|
||||
'ext': 'mp4',
|
||||
'title': 'The Dark World of COVID Pt. 1 | Aaron Siri',
|
||||
'description': 'a420d320dda734e5f29458df3606c5f4',
|
||||
'thumbnail': r're:^https?://.+\.jpg',
|
||||
},
|
||||
}]
|
||||
|
||||
def _real_extract(self, url):
|
||||
video_id = self._match_id(url)
|
||||
webpage = self._download_webpage(url, video_id)
|
||||
|
||||
src_iframe = self._search_regex(r'\<iframe[^>]+src="([^"]+)"', webpage, 'src iframe')
|
||||
webpage_iframe = self._download_webpage(
|
||||
src_iframe.replace('player.php', 'config2.php'), video_id, headers={'Referer': src_iframe})
|
||||
|
||||
sources = self._parse_json(self._search_regex(
|
||||
r'sources\:\s*(\[.*?\])', webpage_iframe, 'm3u8 source'), video_id, transform_source=js_to_json)
|
||||
|
||||
formats, subtitles = [], {}
|
||||
for source in sources:
|
||||
file = source.get('file')
|
||||
if file and source.get('type') == 'm3u8':
|
||||
fmts, subs = self._extract_m3u8_formats_and_subtitles(
|
||||
urljoin('https://www.lightcast.com/embed/', file),
|
||||
video_id, 'mp4', fatal=False, headers={'Referer': src_iframe})
|
||||
formats.extend(fmts)
|
||||
subtitles = self._merge_subtitles(subtitles, subs)
|
||||
self._sort_formats(formats)
|
||||
|
||||
return {
|
||||
'id': video_id,
|
||||
'title': self._html_search_meta(['og:title', 'twitter:title'], webpage),
|
||||
'description': self._html_search_meta(['og:description', 'twitter:description'], webpage),
|
||||
'thumbnail': self._search_regex(r'image:\s*"([^"]+)', webpage_iframe, 'thumbnail'),
|
||||
'formats': formats,
|
||||
'subtitles': subtitles,
|
||||
}
|
|
@ -348,6 +348,7 @@ from .daum import (
|
|||
DaumPlaylistIE,
|
||||
DaumUserIE,
|
||||
)
|
||||
from .daystar import DaystarClipIE
|
||||
from .dbtv import DBTVIE
|
||||
from .dctp import DctpTvIE
|
||||
from .deezer import (
|
||||
|
|
Loading…
Reference in a new issue