from .common import InfoExtractor
from .gigya import GigyaBaseIE
-
-
from ..compat import compat_HTTPError
from ..utils import (
ExtractorError,
strip_or_none,
float_or_none,
int_or_none,
+ merge_dicts,
parse_iso8601,
)
class CanvasIE(InfoExtractor):
- _VALID_URL = r'https?://mediazone\.vrt\.be/api/v1/(?P<site_id>canvas|een|ketnet|vrtvideo)/assets/(?P<id>[^/?#&]+)'
+ _VALID_URL = r'https?://mediazone\.vrt\.be/api/v1/(?P<site_id>canvas|een|ketnet|vrt(?:video|nieuws)|sporza)/assets/(?P<id>[^/?#&]+)'
_TESTS = [{
'url': 'https://mediazone.vrt.be/api/v1/ketnet/assets/md-ast-4ac54990-ce66-4d00-a8ca-9eac86f4c475',
'md5': '90139b746a0a9bd7bb631283f6e2a64e',
'url': 'https://mediazone.vrt.be/api/v1/canvas/assets/mz-ast-5e5f90b6-2d72-4c40-82c2-e134f884e93e',
'only_matching': True,
}]
+ _HLS_ENTRY_PROTOCOLS_MAP = {
+ 'HLS': 'm3u8_native',
+ 'HLS_AES': 'm3u8',
+ }
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url)
format_url, format_type = target.get('url'), target.get('type')
if not format_url or not format_type:
continue
- if format_type == 'HLS':
+ if format_type in self._HLS_ENTRY_PROTOCOLS_MAP:
formats.extend(self._extract_m3u8_formats(
- format_url, video_id, 'mp4', entry_protocol='m3u8_native',
+ format_url, video_id, 'mp4', self._HLS_ENTRY_PROTOCOLS_MAP[format_type],
m3u8_id=format_type, fatal=False))
elif format_type == 'HDS':
formats.extend(self._extract_f4m_formats(
'season_number': 1,
'episode_number': 1,
},
- # 'skip': 'This video is only available for registered users'
+ 'skip': 'This video is only available for registered users'
}]
_NETRC_MACHINE = 'vrtnu'
_APIKEY = '3_0Z2HujMtiWq_pkAjgnS2Md2E11a1AwZjYiBETtwNE-EoEHDINgtnvcAOpNgmrVGy'
def _login(self):
username, password = self._get_login_info()
if username is None:
- self.raise_login_required()
+ return
auth_data = {
'APIKey': self._APIKEY,
def _real_extract(self, url):
display_id = self._match_id(url)
- webpage = self._download_webpage(url, display_id)
+ webpage, urlh = self._download_webpage_handle(url, display_id)
+
+ info = self._search_json_ld(webpage, display_id, default={})
- title = self._html_search_regex(
+ # title is optional here since it may be extracted by extractor
+ # that is delegated from here
+ title = strip_or_none(self._html_search_regex(
r'(?ms)<h1 class="content__heading">(.+?)</h1>',
- webpage, 'title').strip()
+ webpage, 'title', default=None))
description = self._html_search_regex(
r'(?ms)<div class="content__description">(.+?)</div>',
webpage, 'release_date', default=None))
# If there's a ? or a # in the URL, remove them and everything after
- clean_url = url.split('?')[0].split('#')[0].strip('/')
+ clean_url = urlh.geturl().split('?')[0].split('#')[0].strip('/')
securevideo_url = clean_url + '.mssecurevideo.json'
- json = self._download_json(securevideo_url, display_id)
+ try:
+ video = self._download_json(securevideo_url, display_id)
+ except ExtractorError as e:
+ if isinstance(e.cause, compat_HTTPError) and e.cause.code == 401:
+ self.raise_login_required()
+ raise
+
# We are dealing with a '../<show>.relevant' URL
- redirect_url = json.get('url')
+ redirect_url = video.get('url')
if redirect_url:
- return self.url_result('https:' + redirect_url)
- else:
- # There is only one entry, but with an unknown key, so just get
- # the first one
- video_id = list(json.values())[0].get('videoid')
+ return self.url_result(self._proto_relative_url(redirect_url, 'https:'))
- return {
+ # There is only one entry, but with an unknown key, so just get
+ # the first one
+ video_id = list(video.values())[0].get('videoid')
+
+ return merge_dicts(info, {
'_type': 'url_transparent',
'url': 'https://mediazone.vrt.be/api/v1/vrtvideo/assets/%s' % video_id,
'ie_key': CanvasIE.ie_key(),
'season_number': season_number,
'episode_number': episode_number,
'release_date': release_date,
- }
+ })