1
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl.git synced 2024-11-25 03:32:05 +00:00

[TVer] Support redirect from landing page

This commit is contained in:
vvto33 2021-07-15 16:36:21 +09:00
parent a803582717
commit 2a4c11ddb4

View file

@ -6,6 +6,7 @@ import re
from .common import InfoExtractor
from ..compat import compat_str
from ..utils import (
ExtractorError,
int_or_none,
remove_start,
smuggle_url,
@ -14,7 +15,7 @@ from ..utils import (
class TVerIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?tver\.jp/(?P<path>(?:corner|episode|feature)/(?P<id>f?\d+))'
_VALID_URL = r'https?://(?:www\.)?tver\.jp/(?P<path>(?:corner|episode|feature|lp)/(?P<id>[fc]?\d+))'
# videos are only available for 7 days
_TESTS = [{
'url': 'https://tver.jp/corner/f0062178',
@ -29,16 +30,33 @@ class TVerIE(InfoExtractor):
# subtitle = ' '
'url': 'https://tver.jp/corner/f0068870',
'only_matching': True,
}, {
# redirect "f"
'url': 'https://tver.jp/lp/f0009694',
'only_matching': True,
}, {
# redirect "c"
'url': 'https://tver.jp/lp/c0000239',
'only_matching': True,
}]
_TOKEN = None
BRIGHTCOVE_URL_TEMPLATE = 'http://players.brightcove.net/%s/default_default/index.html?videoId=%s'
def _extract_redirect(self, url):
webpage = self._download_webpage(url, "")
newpath = re.search(r'to_href="(.+?)";', webpage)
if newpath is None:
raise ExtractorError('Not found redirect path', expected=True)
return re.match(self._VALID_URL, "https://tver.jp" + newpath.group(1)).groups()
def _real_initialize(self):
self._TOKEN = self._download_json(
'https://tver.jp/api/access_token.php', None)['token']
def _real_extract(self, url):
path, video_id = re.match(self._VALID_URL, url).groups()
if path[0:2] == "lp":
path, video_id = self._extract_redirect(url)
main = self._download_json(
'https://api.tver.jp/v4/' + path, video_id,
query={'token': self._TOKEN})['main']