[ie/crunchyroll] Fix stream extraction (#10005)

[yt-dlp.git] / yt_dlp / extractor / periscope.py

diff --git a/yt_dlp/extractor/periscope.py b/yt_dlp/extractor/periscope.py
index fc8591a2ce61c55a102cc841aae47b6e7f50f6dc..d2351df1a2778bd6a5770e674b8511ecdce83682 100644 (file)
--- a/yt_dlp/extractor/periscope.py
+++ b/yt_dlp/extractor/periscope.py
@@ -1,11 +1,10 @@
-import re
-
 from .common import InfoExtractor
 from ..utils import (
     int_or_none,
     parse_iso8601,
     unescapeHTML,
 )
+from ..utils.traversal import traverse_obj
 
 
 class PeriscopeBaseIE(InfoExtractor):
@@ -22,22 +21,26 @@ def _parse_broadcast_data(self, broadcast, video_id):
         title = broadcast.get('status') or 'Periscope Broadcast'
         uploader = broadcast.get('user_display_name') or broadcast.get('username')
         title = '%s - %s' % (uploader, title) if uploader else title
-        is_live = broadcast.get('state').lower() == 'running'
-
         thumbnails = [{
             'url': broadcast[image],
-        } for image in ('image_url', 'image_url_small') if broadcast.get(image)]
+        } for image in ('image_url', 'image_url_medium', 'image_url_small') if broadcast.get(image)]
 
         return {
             'id': broadcast.get('id') or video_id,
             'title': title,
-            'timestamp': parse_iso8601(broadcast.get('created_at')),
+            'timestamp': parse_iso8601(broadcast.get('created_at')) or int_or_none(
+                broadcast.get('created_at_ms'), scale=1000),
+            'release_timestamp': int_or_none(broadcast.get('scheduled_start_ms'), scale=1000),
             'uploader': uploader,
             'uploader_id': broadcast.get('user_id') or broadcast.get('username'),
             'thumbnails': thumbnails,
             'view_count': int_or_none(broadcast.get('total_watched')),
+            'concurrent_view_count': int_or_none(broadcast.get('total_watching')),
             'tags': broadcast.get('tags'),
-            'is_live': is_live,
+            'live_status': {
+                'running': 'is_live',
+                'not_started': 'is_upcoming',
+            }.get(traverse_obj(broadcast, ('state', {str.lower}))) or 'was_live'
         }
 
     @staticmethod
@@ -67,6 +70,7 @@ class PeriscopeIE(PeriscopeBaseIE):
     IE_DESC = 'Periscope'
     IE_NAME = 'periscope'
     _VALID_URL = r'https?://(?:www\.)?(?:periscope|pscp)\.tv/[^/]+/(?P<id>[^/?#]+)'
+    _EMBED_REGEX = [r'<iframe[^>]+src=([\'"])(?P<url>(?:https?:)?//(?:www\.)?(?:periscope|pscp)\.tv/(?:(?!\1).)+)\1']
     # Alive example URLs can be found here https://www.periscope.tv/
     _TESTS = [{
         'url': 'https://www.periscope.tv/w/aJUQnjY3MjA3ODF8NTYxMDIyMDl2zCg2pECBgwTqRpQuQD352EMPTKQjT4uqlM3cgWFA-g==',
@@ -92,13 +96,6 @@ class PeriscopeIE(PeriscopeBaseIE):
         'only_matching': True,
     }]
 
-    @staticmethod
-    def _extract_url(webpage):
-        mobj = re.search(
-            r'<iframe[^>]+src=([\'"])(?P<url>(?:https?:)?//(?:www\.)?(?:periscope|pscp)\.tv/(?:(?!\1).)+)\1', webpage)
-        if mobj:
-            return mobj.group('url')
-
     def _real_extract(self, url):
         token = self._match_id(url)
 
@@ -135,7 +132,6 @@ def add_width_and_height(f):
             }
             self._add_width_and_height(rtmp_format)
             formats.append(rtmp_format)
-        self._sort_formats(formats)
 
         info['formats'] = formats
         return info