paradizelost/youtube-dl
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[appleconnect] Fix extraction (#29208)

Browse Source
This commit is contained in:
kikuyan 2021-06-17 06:12:13 +09:00 committed by GitHub
parent d156bc8d59
commit 8fe5d54eb7
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 8 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
youtube_dl/extractor/appleconnect.py
View File

@ -9,10 +9,10 @@ from ..utils import (
class AppleConnectIE(InfoExtractor): class AppleConnectIE(InfoExtractor):
_VALID_URL = r'https?://itunes\.apple\.com/\w{0,2}/?post/idsa\.(?P<id>[\w-]+)' _VALID_URL = r'https?://itunes\.apple\.com/\w{0,2}/?post/(?:id)?sa\.(?P<id>[\w-]+)'
_TEST = { _TESTS = [{
'url': 'https://itunes.apple.com/us/post/idsa.4ab17a39-2720-11e5-96c5-a5b38f6c42d3', 'url': 'https://itunes.apple.com/us/post/idsa.4ab17a39-2720-11e5-96c5-a5b38f6c42d3',
'md5': 'e7c38568a01ea45402570e6029206723', 'md5': 'c1d41f72c8bcaf222e089434619316e4',
'info_dict': { 'info_dict': {
'id': '4ab17a39-2720-11e5-96c5-a5b38f6c42d3', 'id': '4ab17a39-2720-11e5-96c5-a5b38f6c42d3',
'ext': 'm4v', 'ext': 'm4v',
@ -22,7 +22,10 @@ class AppleConnectIE(InfoExtractor):
'upload_date': '20150710', 'upload_date': '20150710',
'timestamp': 1436545535, 'timestamp': 1436545535,
}, },
} }, {
'url': 'https://itunes.apple.com/us/post/sa.0fe0229f-2457-11e5-9f40-1bb645f2d5d9',
'only_matching': True,
}]
def _real_extract(self, url): def _real_extract(self, url):
video_id = self._match_id(url) video_id = self._match_id(url)
@ -36,7 +39,7 @@ class AppleConnectIE(InfoExtractor):
video_data = self._parse_json(video_json, video_id) video_data = self._parse_json(video_json, video_id)
timestamp = str_to_int(self._html_search_regex(r'data-timestamp="(\d+)"', webpage, 'timestamp')) timestamp = str_to_int(self._html_search_regex(r'data-timestamp="(\d+)"', webpage, 'timestamp'))
like_count = str_to_int(self._html_search_regex(r'(\d+) Loves', webpage, 'like count')) like_count = str_to_int(self._html_search_regex(r'(\d+) Loves', webpage, 'like count', default=None))
return { return {
'id': video_id, 'id': video_id,