pepepper/youtube-dl
1
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl synced 2025-01-22 03:10:11 +09:00
Code Issues Projects Releases Wiki Activity

Finished rewrite to parse JSON as JSON. This looks much better.

Browse Source
This commit is contained in:
Tim Mann 2021-02-13 17:34:34 -08:00
parent fd61f317bf
commit fb35bd50b0
1 changed files with 15 additions and 30 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

45
youtube_dl/extractor/pac12.py
View File

@ -1,13 +1,8 @@
# coding: utf-8
from __future__ import unicode_literals
from __future__ import print_function #XXX
import pprint #XXX
import re
from .common import InfoExtractor
class Pac12IE(InfoExtractor):
_VALID_URL = r'https?://(?:[a-z]+\.)?pac-12.com/(?:embed/)?(?P<id>.*)'
@ -17,8 +12,8 @@ class Pac12IE(InfoExtractor):
'info_dict': {
'id': 'vod-VGQNKGlo9Go',
'ext': 'mp4',
'title': '2020 Pac-12 Women\'s Basketball Media Day - Arizona, Cal & Stanford | Pac-12',
'description': 'During the 2020 Pac-12 Women\'s Basketball Media Day, Ros Gold-Onwude moderates a discussion with Arizona\'s Adia Barnes & Aari McDonald, Cal\'s Charmin Smith & Evelien Lutje Schipholt & Stanford\'s Tara VanDerveer & Kiana Williams.',
'title': '2020 Pac-12 Women\'s Basketball Media Day - Arizona, Cal & Stanford',
'description': 'During the 2020 Pac-12 Women\'s Basketball Media Day, Ros Gold-Onwude moderates a discussion with Arizona\'s Adia Barnes & Aari McDonald, Cal\'s Charmin Smith & Evelien Lutje Schipholt & Stanford\'s Tara VanDerveer & Kiana Williams. ',
}
}, {
'url': 'https://pac-12.com/article/2020/11/24/sonoran-dog-dish-presented-tums',
@ -26,7 +21,7 @@ class Pac12IE(InfoExtractor):
'info_dict': {
'id': 'vod-YLMKpNLZvR0',
'ext': 'mp4',
'title': 'Sonoran Dog | The Dish, presented by TUMS | Pac-12',
'title': 'Sonoran Dog | The Dish, presented by TUMS',
'description': 'Pac-12 Networks introduces "The Dish," presented by Tums. Jaymee Sire is bringing fans a closeup to game day treats from around the Conference with each treat connecting to a Pac-12 school, bringing the flavor and recipes fans know and love right to the dish! As Arizona and USC basketball seasons tip off, the first feature item from "The Dish" is the Sonoran Dog, a beloved treat by Trojans & Wildcat fans.',
}
}]
@ -39,33 +34,23 @@ class Pac12IE(InfoExtractor):
self._search_regex(
r'<script[^>]+type="application/json"[^>]*data-drupal-selector="drupal-settings-json">([^<]+)</script>',
webpage, 'drupal settings'), video_id)
pprint.pprint(drupal_settings.get('currentVideo'))
video_url = drupal_settings.get('currentVideo', {}).get('manifest_url')
cv = drupal_settings.get('currentVideo', {})
manifest_url = cv.get('manifest_url')
vod_url = None
if (video_url is None) or ('vod-' not in url):
vod_url = self._search_regex(r'(https?://(?:embed\.)?pac-12\.com/(?:embed/)?vod-[0-9a-zA-Z]+)',
webpage, 'url', default=None)
if video_url is None:
if manifest_url is None:
# Video may be embedded one level deeper
vod_url = self._search_regex(
r'(https?://(?:embed\.)?pac-12\.com/(?:embed/)?vod-\w+)',
webpage, 'url', default=None)
if vod_url is None:
return None
return self.url_result(vod_url)
video_url = re.sub(r'\\', '', video_url)
if 'vod-' not in url and vod_url is not None:
video_id = self._match_id(vod_url)
title = self._html_search_regex(r'<title>(.+?)</title>',
webpage, 'title')
description = self._og_search_description(webpage, default=None)
if description is None:
d = self._search_regex(r'"description":"(?P<description>[^"]+)"',
webpage, 'description', default=None)
if d is not None:
description = d.encode('utf-8').decode('unicode_escape')
return {
'id': video_id,
'title': title,
'description': description,
'url': video_url,
'id': cv.get('id'),
'title': cv.get('title'),
'description': cv.get('description'),
'url': manifest_url,
'ext': 'mp4',
}