[extractor/cbc] Fix live extractor, playlist _VALID_URL (#6625)

Authored by: makew0rld

yt_dlp/extractor/cbc.py

@@ -8,14 +8,16 @@
     compat_str,
 )
 from ..utils import (
+    ExtractorError,
     int_or_none,
     join_nonempty,
     js_to_json,
     orderedSet,
+    parse_iso8601,
     smuggle_url,
     strip_or_none,
+    traverse_obj,
     try_get,
-    ExtractorError,
 )
 
 
@@ -404,7 +406,7 @@ def _real_extract(self, url):
 
 class CBCGemPlaylistIE(InfoExtractor):
     IE_NAME = 'gem.cbc.ca:playlist'
-    _VALID_URL = r'https?://gem\.cbc\.ca/media/(?P<id>(?P<show>[0-9a-z-]+)/s(?P<season>[0-9]+))/?(?:[?#]|$)'
+    _VALID_URL = r'https?://gem\.cbc\.ca/(?:media/)?(?P<id>(?P<show>[0-9a-z-]+)/s(?P<season>[0-9]+))/?(?:[?#]|$)'
     _TESTS = [{
         # TV show playlist, all public videos
         'url': 'https://gem.cbc.ca/media/schitts-creek/s06',
@@ -414,6 +416,9 @@ class CBCGemPlaylistIE(InfoExtractor):
             'title': 'Season 6',
             'description': 'md5:6a92104a56cbeb5818cc47884d4326a2',
         },
+    }, {
+        'url': 'https://gem.cbc.ca/schitts-creek/s06',
+        'only_matching': True,
     }]
     _API_BASE = 'https://services.radio-canada.ca/ott/cbc-api/v2/shows/'
 
@@ -473,8 +478,9 @@ def _real_extract(self, url):
 
 class CBCGemLiveIE(InfoExtractor):
     IE_NAME = 'gem.cbc.ca:live'
-    _VALID_URL = r'https?://gem\.cbc\.ca/live/(?P<id>\d+)'
-    _TEST = {
+    _VALID_URL = r'https?://gem\.cbc\.ca/live(?:-event)?/(?P<id>\d+)'
+    _TESTS = [
+        {
             'url': 'https://gem.cbc.ca/live/920604739687',
             'info_dict': {
                 'title': 'Ottawa',
@@ -488,34 +494,74 @@ class CBCGemLiveIE(InfoExtractor):
                 'uploader': 'CBCC-NEW',
             },
             'skip': 'Live might have ended',
+        },
+        {
+            'url': 'https://gem.cbc.ca/live/44',
+            'info_dict': {
+                'id': '44',
+                'ext': 'mp4',
+                'is_live': True,
+                'title': r're:^Ottawa [0-9\-: ]+',
+                'description': 'The live TV channel and local programming from Ottawa',
+                'live_status': 'is_live',
+                'thumbnail': r're:https://images.gem.cbc.ca/v1/cbc-gem/live/.*'
+            },
+            'params': {'skip_download': True},
+            'skip': 'Live might have ended',
+        },
+        {
+            'url': 'https://gem.cbc.ca/live-event/10835',
+            'info_dict': {
+                'id': '10835',
+                'ext': 'mp4',
+                'is_live': True,
+                'title': r're:^The National \| Biden’s trip wraps up, Paltrow testifies, Bird flu [0-9\-: ]+',
+                'description': 'March 24, 2023 | President Biden’s Ottawa visit ends with big pledges from both countries. Plus, Gwyneth Paltrow testifies in her ski collision trial.',
+                'live_status': 'is_live',
+                'thumbnail': r're:https://images.gem.cbc.ca/v1/cbc-gem/live/.*',
+                'timestamp': 1679706000,
+                'upload_date': '20230325',
+            },
+            'params': {'skip_download': True},
+            'skip': 'Live might have ended',
         }
-
-    # It's unclear where the chars at the end come from, but they appear to be
-    # constant. Might need updating in the future.
-    # There are two URLs, some livestreams are in one, and some
-    # in the other. The JSON schema is the same for both.
-    _API_URLS = ['https://tpfeed.cbc.ca/f/ExhSPC/t_t3UKJR6MAT', 'https://tpfeed.cbc.ca/f/ExhSPC/FNiv9xQx_BnT']
+    ]
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
+        webpage = self._download_webpage(url, video_id)
+        video_info = self._search_nextjs_data(webpage, video_id)['props']['pageProps']['data']
 
-        for api_url in self._API_URLS:
-            video_info = next((
-                stream for stream in self._download_json(api_url, video_id)['entries']
-                if stream.get('guid') == video_id), None)
-            if video_info:
-                break
-        else:
+        # Two types of metadata JSON
+        if not video_info.get('formattedIdMedia'):
+            video_info = traverse_obj(
+                video_info, (('freeTv', ('streams', ...)), 'items', lambda _, v: v['key'] == video_id, {dict}),
+                get_all=False, default={})
+
+        video_stream_id = video_info.get('formattedIdMedia')
+        if not video_stream_id:
             raise ExtractorError('Couldn\'t find video metadata, maybe this livestream is now offline', expected=True)
 
+        stream_data = self._download_json(
+            'https://services.radio-canada.ca/media/validation/v2/', video_id, query={
+                'appCode': 'mpx',
+                'connectionType': 'hd',
+                'deviceType': 'ipad',
+                'idMedia': video_stream_id,
+                'multibitrate': 'true',
+                'output': 'json',
+                'tech': 'hls',
+                'manifestType': 'desktop',
+            })
+
         return {
-            '_type': 'url_transparent',
-            'ie_key': 'ThePlatform',
-            'url': video_info['content'][0]['url'],
             'id': video_id,
-            'title': video_info.get('title'),
-            'description': video_info.get('description'),
-            'tags': try_get(video_info, lambda x: x['keywords'].split(', ')),
-            'thumbnail': video_info.get('cbc$staticImage'),
+            'formats': self._extract_m3u8_formats(stream_data['url'], video_id, 'mp4', live=True),
             'is_live': True,
+            **traverse_obj(video_info, {
+                'title': 'title',
+                'description': 'description',
+                'thumbnail': ('images', 'card', 'url'),
+                'timestamp': ('airDate', {parse_iso8601}),
+            })
         }