Madeorsk/youtube-dl
1
0
Fork 0
mirror of https://codeberg.org/polarisfm/youtube-dl synced 2024-11-26 02:14:32 +01:00
Code Issues Releases Wiki Activity

[ITVBTCC] Improve reliability

Browse Source
This commit is contained in:
Fran Hermoso 2020-05-14 17:04:54 +02:00
parent e4587a2aac
commit 0381cbe576
1 changed files with 14 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
youtube_dl/extractor/itv.py
View File

@ -20,6 +20,7 @@ from ..utils import (
merge_dicts, merge_dicts,
parse_duration, parse_duration,
smuggle_url, smuggle_url,
try_get,
url_or_none, url_or_none,
xpath_with_ns, xpath_with_ns,
xpath_element, xpath_element,
@ -285,7 +286,7 @@ class ITVBTCCIE(InfoExtractor):
'id': 'btcc-2019-brands-hatch-gp-race-action', 'id': 'btcc-2019-brands-hatch-gp-race-action',
'title': 'BTCC 2019: Brands Hatch GP race action', 'title': 'BTCC 2019: Brands Hatch GP race action',
}, },
'playlist_mincount': 12, 'playlist_count': 12,
} }
BRIGHTCOVE_URL_TEMPLATE = 'http://players.brightcove.net/1582188683001/HkiHLnNRx_default/index.html?videoId=%s' BRIGHTCOVE_URL_TEMPLATE = 'http://players.brightcove.net/1582188683001/HkiHLnNRx_default/index.html?videoId=%s'
@ -295,28 +296,29 @@ class ITVBTCCIE(InfoExtractor):
webpage = self._download_webpage(url, playlist_id) webpage = self._download_webpage(url, playlist_id)
json_map = self._html_search_regex( json_map = try_get(self._parse_json(self._html_search_regex(
'<script id="__NEXT_DATA__" type="application/json">(.*?)</script>', '(?s)<script[^>]+id=[\'"]__NEXT_DATA__[^>]*>([^<]+)</script>', webpage, 'json_map'), playlist_id),
webpage, lambda x: x['props']['pageProps']['article']['body']['content']) or []
'json_map'
) # Discard empty objects
video_ids = []
for video in json_map:
if video['data'].get('id'):
video_ids.append(video['data']['id'])
entries = [ entries = [
self.url_result( self.url_result(
smuggle_url( smuggle_url(
self.BRIGHTCOVE_URL_TEMPLATE % video_id['data'].get('id'), { self.BRIGHTCOVE_URL_TEMPLATE % video_id, {
# ITV does not like some GB IP ranges, so here are some # ITV does not like some GB IP ranges, so here are some
# IP blocks it accepts # IP blocks it accepts
'geo_ip_blocks': [ 'geo_ip_blocks': [
'193.113.0.0/16', '54.36.162.0/23', '159.65.16.0/21' '193.113.0.0/16', '54.36.162.0/23', '159.65.16.0/21'
], ],
'referrer': url, 'referrer': url,
} }),
),
ie=BrightcoveNewIE.ie_key(), video_id=video_id) ie=BrightcoveNewIE.ie_key(), video_id=video_id)
for video_id in self._parse_json( for video_id in video_ids]
json_map, playlist_id
)['props']['pageProps']['article']['body']['content']]
title = self._og_search_title(webpage, fatal=False) title = self._og_search_title(webpage, fatal=False)