diff options
Diffstat (limited to 'youtube_dl')
-rw-r--r-- | youtube_dl/extractor/teamcoco.py | 25 |
1 files changed, 20 insertions, 5 deletions
diff --git a/youtube_dl/extractor/teamcoco.py b/youtube_dl/extractor/teamcoco.py index 64235b0f6..63fd4fe1c 100644 --- a/youtube_dl/extractor/teamcoco.py +++ b/youtube_dl/extractor/teamcoco.py @@ -16,7 +16,7 @@ from ..utils import ( class TeamcocoIE(InfoExtractor): - _VALID_URL = r'https?://teamcoco\.com/video/(?P<id>([^/]+/)*[^/?#]+)' + _VALID_URL = r'https?://teamcoco\.com/(?P<id>([^/]+/)*[^/?#]+)' _TESTS = [ { 'url': 'http://teamcoco.com/video/mary-kay-remote', @@ -70,6 +70,15 @@ class TeamcocoIE(InfoExtractor): }, { 'url': 'http://teamcoco.com/video/the-conan-audiencey-awards-for-04/25/18', 'only_matching': True, + }, { + 'url': 'http://teamcoco.com/italy/conan-jordan-schlansky-hit-the-streets-of-florence', + 'only_matching': True, + }, { + 'url': 'http://teamcoco.com/haiti/conan-s-haitian-history-lesson', + 'only_matching': True, + }, { + 'url': 'http://teamcoco.com/israel/conan-hits-the-streets-beaches-of-tel-aviv', + 'only_matching': True, } ] @@ -84,7 +93,7 @@ class TeamcocoIE(InfoExtractor): display_id = self._match_id(url) response = self._graphql_call('''{ - %s(slug: "video/%s") { + %s(slug: "%s") { ... on RecordSlug { record { id @@ -94,6 +103,9 @@ class TeamcocoIE(InfoExtractor): thumb { preview } + file { + url + } tags { name } @@ -111,15 +123,15 @@ class TeamcocoIE(InfoExtractor): record = response['record'] video_id = record['id'] - srcs = self._graphql_call('''{ + video_sources = self._graphql_call('''{ %s(id: "%s") { src } -}''', 'RecordVideoSource', video_id)['src'] +}''', 'RecordVideoSource', video_id) or {} formats = [] get_quality = qualities(['low', 'sd', 'hd', 'uhd']) - for format_id, src in srcs.items(): + for format_id, src in video_sources.get('src', {}).items(): if not isinstance(src, dict): continue src_url = src.get('src') @@ -146,6 +158,9 @@ class TeamcocoIE(InfoExtractor): 'format_id': format_id, 'quality': get_quality(format_id), }) + if not formats: + formats = self._extract_m3u8_formats( + record['file']['url'], video_id, 'mp4', fatal=False) self._sort_formats(formats) return { |