github/yt-dlp
1
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-11-21 20:46:36 -05:00
Code Issues Projects Releases Packages Wiki Activity

[rcs] fixed embeds detection, fixed tests

Browse source
This commit is contained in:
nixxo 2020-11-09 10:59:25 +01:00
parent 508649e6f5
commit 5867a16789
2 changed files with 29 additions and 33 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

3
youtube_dlc/extractor/extractors.py
View file

@ -937,8 +937,7 @@
) )
from .rbmaradio import RBMARadioIE from .rbmaradio import RBMARadioIE
from .rcs import ( from .rcs import (
CorriereIE, RCSIE,
GazzettaIE,
RCSEmbedsIE, RCSEmbedsIE,
RCSVariousIE, RCSVariousIE,
) )

59
youtube_dlc/extractor/rcs.py
View file

@ -14,7 +14,7 @@
) )
class RCSIE(InfoExtractor): class RCSBaseIE(InfoExtractor):
_ALL_REPLACE = { _ALL_REPLACE = {
'media2vam.corriere.it.edgesuite.net': 'media2vam.corriere.it.edgesuite.net':
'media2vam-corriere-it.akamaized.net', 'media2vam-corriere-it.akamaized.net',
@ -237,7 +237,11 @@ def _real_extract(self, url):
# if no video data found try search for iframes # if no video data found try search for iframes
emb = RCSEmbedsIE._extract_url(page) emb = RCSEmbedsIE._extract_url(page)
if emb: if emb:
return self._real_extract(emb) return {
'_type': 'url_transparent',
'url': emb,
'ie_key': RCSEmbedsIE.ie_key()
}
if not video_data: if not video_data:
raise ExtractorError('Video data not found in the page') raise ExtractorError('Video data not found in the page')
@ -247,7 +251,7 @@ def _real_extract(self, url):
description = (video_data.get('description') description = (video_data.get('description')
or clean_html(video_data.get('htmlDescription'))) or clean_html(video_data.get('htmlDescription')))
uploader = video_data.get('provider') or mobj.gruop('cdn') uploader = video_data.get('provider') or mobj.group('cdn')
return { return {
'id': video_id, 'id': video_id,
@ -258,8 +262,7 @@ def _real_extract(self, url):
} }
class RCSEmbedsIE(RCSIE): class RCSEmbedsIE(RCSBaseIE):
IE_NAME = 'rcs:rcs'
_VALID_URL = r'''(?x) _VALID_URL = r'''(?x)
https?://(?P<vid>video)\. https?://(?P<vid>video)\.
(?P<cdn> (?P<cdn>
@ -279,6 +282,16 @@ class RCSEmbedsIE(RCSIE):
'description': 'md5:65b09633df9ffee57f48b39e34c9e067', 'description': 'md5:65b09633df9ffee57f48b39e34c9e067',
'uploader': 'rcs.it', 'uploader': 'rcs.it',
} }
}, {
'url': 'https://video.gazzanet.gazzetta.it/video-embed/gazzanet-mo05-0000260789',
'md5': 'a043e3fecbe4d9ed7fc5d888652a5440',
'info_dict': {
'id': 'gazzanet-mo05-0000260789',
'ext': 'mp4',
'title': 'Valentino Rossi e papà Graziano si divertono col drifting',
'description': 'md5:a8bf90d6adafd9815f70fc74c0fc370a',
'uploader': 'rcd',
}
}, { }, {
'url': 'https://video.corriere.it/video-embed/b727632a-f9d0-11ea-91b0-38d50a849abb?player', 'url': 'https://video.corriere.it/video-embed/b727632a-f9d0-11ea-91b0-38d50a849abb?player',
'match_only': True 'match_only': True
@ -324,17 +337,17 @@ def _extract_url(webpage):
return urls[0] if urls else None return urls[0] if urls else None
class CorriereIE(RCSIE): class RCSIE(RCSBaseIE):
IE_NAME = 'rcs:corriere'
_VALID_URL = r'''(?x)https?://(?P<vid>video|viaggi)\. _VALID_URL = r'''(?x)https?://(?P<vid>video|viaggi)\.
(?P<cdn> (?P<cdn>
(?: (?:
corrieredelmezzogiorno\.| corrieredelmezzogiorno\.
corrieredelveneto\.| |corrieredelveneto\.
corrieredibologna\.| |corrieredibologna\.
corrierefiorentino\. |corrierefiorentino\.
)? )?corriere\.it
corriere\.it)/.+?/(?P<id>[^/]+)(?=\?|/$|$)''' |(?:gazzanet\.)?gazzetta\.it)
/(?!video-embed/).+?/(?P<id>[^/\?]+)(?=\?|/$|$)'''
_TESTS = [{ _TESTS = [{
'url': 'https://video.corriere.it/sport/formula-1/vettel-guida-ferrari-sf90-mugello-suo-fianco-c-elecrerc-bendato-video-esilarante/b727632a-f9d0-11ea-91b0-38d50a849abb', 'url': 'https://video.corriere.it/sport/formula-1/vettel-guida-ferrari-sf90-mugello-suo-fianco-c-elecrerc-bendato-video-esilarante/b727632a-f9d0-11ea-91b0-38d50a849abb',
'md5': '0f4ededc202b0f00b6e509d831e2dcda', 'md5': '0f4ededc202b0f00b6e509d831e2dcda',
@ -356,18 +369,6 @@ class CorriereIE(RCSIE):
'uploader': 'DOVE Viaggi', 'uploader': 'DOVE Viaggi',
} }
}, { }, {
'url': 'https://video.corriere.it/video-embed/b727632a-f9d0-11ea-91b0-38d50a849abb?player',
'match_only': True
}, {
'url': 'https://video.corriere.it/video-360/metro-copenaghen-tutta-italiana/a248a7f0-e2db-11e9-9830-af2de6b1f945',
'match_only': True
}]
class GazzettaIE(RCSIE):
IE_NAME = 'rcs:gazzetta'
_VALID_URL = r'https?://(?P<vid>video)\.(?P<cdn>(?:gazzanet\.)?gazzetta\.it)/.+?/(?P<id>[^/]+?)(?:$|\?)'
_TESTS = [{
'url': 'https://video.gazzetta.it/video-motogp-catalogna-cadute-dovizioso-vale-rossi/49612410-00ca-11eb-bcd8-30d4253e0140?vclk=Videobar', 'url': 'https://video.gazzetta.it/video-motogp-catalogna-cadute-dovizioso-vale-rossi/49612410-00ca-11eb-bcd8-30d4253e0140?vclk=Videobar',
'md5': 'eedc1b5defd18e67383afef51ff7bdf9', 'md5': 'eedc1b5defd18e67383afef51ff7bdf9',
'info_dict': { 'info_dict': {
@ -378,16 +379,12 @@ class GazzettaIE(RCSIE):
'uploader': 'AMorici', 'uploader': 'AMorici',
} }
}, { }, {
'url': 'https://video.gazzetta.it/video-embed/49612410-00ca-11eb-bcd8-30d4253e0140', 'url': 'https://video.corriere.it/video-360/metro-copenaghen-tutta-italiana/a248a7f0-e2db-11e9-9830-af2de6b1f945',
'match_only': True
}, {
'url': 'https://video.gazzanet.gazzetta.it/video-embed/gazzanet-mo05-0000260789',
'match_only': True 'match_only': True
}] }]
class RCSVariousIE(RCSIE): class RCSVariousIE(RCSBaseIE):
IE_NAME = 'rcs:various'
_VALID_URL = r'''(?x)https?://www\. _VALID_URL = r'''(?x)https?://www\.
(?P<cdn> (?P<cdn>
leitv\.it| leitv\.it|