[foxnews:insider] Remove extractor (#15810)
Now covered by foxnews:article
This commit is contained in:
parent
f51f526b0a
commit
91aa502d91
@ -373,7 +373,6 @@ from .foxgay import FoxgayIE
|
|||||||
from .foxnews import (
|
from .foxnews import (
|
||||||
FoxNewsIE,
|
FoxNewsIE,
|
||||||
FoxNewsArticleIE,
|
FoxNewsArticleIE,
|
||||||
FoxNewsInsiderIE,
|
|
||||||
)
|
)
|
||||||
from .foxsports import FoxSportsIE
|
from .foxsports import FoxSportsIE
|
||||||
from .franceculture import FranceCultureIE
|
from .franceculture import FranceCultureIE
|
||||||
|
@ -76,7 +76,7 @@ class FoxNewsIE(AMPIE):
|
|||||||
|
|
||||||
|
|
||||||
class FoxNewsArticleIE(InfoExtractor):
|
class FoxNewsArticleIE(InfoExtractor):
|
||||||
_VALID_URL = r'https?://(?:www\.)?foxnews\.com/(?!v)([^/]+/)+(?P<id>[a-z-]+)'
|
_VALID_URL = r'https?://(?:www\.)?(?:insider\.)?foxnews\.com/(?!v)([^/]+/)+(?P<id>[a-z-]+)'
|
||||||
IE_NAME = 'foxnews:article'
|
IE_NAME = 'foxnews:article'
|
||||||
|
|
||||||
_TESTS = [{
|
_TESTS = [{
|
||||||
@ -107,6 +107,9 @@ class FoxNewsArticleIE(InfoExtractor):
|
|||||||
'params': {
|
'params': {
|
||||||
'skip_download': True,
|
'skip_download': True,
|
||||||
},
|
},
|
||||||
|
}, {
|
||||||
|
'url': 'http://insider.foxnews.com/2016/08/25/univ-wisconsin-student-group-pushing-silence-certain-words',
|
||||||
|
'only_matching': True,
|
||||||
}]
|
}]
|
||||||
|
|
||||||
def _real_extract(self, url):
|
def _real_extract(self, url):
|
||||||
@ -122,47 +125,3 @@ class FoxNewsArticleIE(InfoExtractor):
|
|||||||
|
|
||||||
return self.url_result(
|
return self.url_result(
|
||||||
FoxNewsIE._extract_urls(webpage)[0], FoxNewsIE.ie_key())
|
FoxNewsIE._extract_urls(webpage)[0], FoxNewsIE.ie_key())
|
||||||
|
|
||||||
|
|
||||||
class FoxNewsInsiderIE(InfoExtractor):
|
|
||||||
_VALID_URL = r'https?://insider\.foxnews\.com/([^/]+/)+(?P<id>[a-z-]+)'
|
|
||||||
IE_NAME = 'foxnews:insider'
|
|
||||||
|
|
||||||
_TEST = {
|
|
||||||
'url': 'http://insider.foxnews.com/2016/08/25/univ-wisconsin-student-group-pushing-silence-certain-words',
|
|
||||||
'md5': 'a10c755e582d28120c62749b4feb4c0c',
|
|
||||||
'info_dict': {
|
|
||||||
'id': '5099377331001',
|
|
||||||
'display_id': 'univ-wisconsin-student-group-pushing-silence-certain-words',
|
|
||||||
'ext': 'mp4',
|
|
||||||
'title': 'Student Group: Saying \'Politically Correct,\' \'Trash\' and \'Lame\' Is Offensive',
|
|
||||||
'description': 'Is campus censorship getting out of control?',
|
|
||||||
'timestamp': 1472168725,
|
|
||||||
'upload_date': '20160825',
|
|
||||||
'thumbnail': r're:^https?://.*\.jpg$',
|
|
||||||
},
|
|
||||||
'params': {
|
|
||||||
# m3u8 download
|
|
||||||
'skip_download': True,
|
|
||||||
},
|
|
||||||
'add_ie': [FoxNewsIE.ie_key()],
|
|
||||||
}
|
|
||||||
|
|
||||||
def _real_extract(self, url):
|
|
||||||
display_id = self._match_id(url)
|
|
||||||
|
|
||||||
webpage = self._download_webpage(url, display_id)
|
|
||||||
|
|
||||||
embed_url = self._html_search_meta('embedUrl', webpage, 'embed URL')
|
|
||||||
|
|
||||||
title = self._og_search_title(webpage)
|
|
||||||
description = self._og_search_description(webpage)
|
|
||||||
|
|
||||||
return {
|
|
||||||
'_type': 'url_transparent',
|
|
||||||
'ie_key': FoxNewsIE.ie_key(),
|
|
||||||
'url': embed_url,
|
|
||||||
'display_id': display_id,
|
|
||||||
'title': title,
|
|
||||||
'description': description,
|
|
||||||
}
|
|
||||||
|
Loading…
Reference in New Issue
Block a user