Fix .groups() arguments

- https://docs.python.org/3/library/re.html#re.Match.groups
> Match.groups(default=None)

yt_dlp/extractor/duoplay.py

@@ -1,19 +1,24 @@
 from .common import InfoExtractor
 from ..utils import (
-    traverse_obj,
+    extract_attributes,
+    get_element_text_and_html_by_tag,
+    int_or_none,
+    join_nonempty,
+    str_or_none,
+    try_call,
     unified_timestamp,
-    strip_or_none,
 )
+from ..utils.traversal import traverse_obj
 
 
 class DuoplayIE(InfoExtractor):
-    _VALID_URL = r'https://duoplay\.ee/(?P<id>\d+)/'
+    _VALID_URL = r'https://duoplay\.ee/(?P<id>\d+)/[\w-]+/?(?:\?(?:[^#]+&)?ep=(?P<ep>\d+))?'
     _TESTS = [{
         'note': 'Siberi võmm S02E12',
         'url': 'https://duoplay.ee/4312/siberi-vomm?ep=24',
         'md5': '1ff59d535310ac9c5cf5f287d8f91b2d',
         'info_dict': {
-            'id': '4312',
+            'id': '4312_24',
             'ext': 'mp4',
             'title': 'Operatsioon "Öö"',
             'thumbnail': r're:https://.+\.jpg(?:\?c=\d+)?$',
@@ -21,19 +26,18 @@ class DuoplayIE(InfoExtractor):
             'upload_date': '20170523',
             'timestamp': 1495567800,
             'series': 'Siberi võmm',
-            'series_id': 4312,
+            'series_id': '4312',
             'season': 'Season 2',
             'season_number': 2,
             'episode': 'Operatsioon "Öö"',
             'episode_number': 12,
-            'episode_id': 24,
         },
     }, {
         'note': 'Empty title',
         'url': 'https://duoplay.ee/17/uhikarotid?ep=14',
         'md5': '6aca68be71112314738dd17cced7f8bf',
         'info_dict': {
-            'id': '17',
+            'id': '17_14',
             'ext': 'mp4',
             'title': 'Episode 14',
             'thumbnail': r're:https://.+\.jpg(?:\?c=\d+)?$',
@@ -41,36 +45,37 @@ class DuoplayIE(InfoExtractor):
             'upload_date': '20100916',
             'timestamp': 1284661800,
             'series': 'Ühikarotid',
-            'series_id': 17,
+            'series_id': '17',
             'season': 'Season 2',
             'season_number': 2,
-            'episode_id': 14,
+            'episode': 'Episode 14',
+            'episode_number': 14,
         },
     }]
 
     def _real_extract(self, url):
-        def decode_quot(s: str):
+        telecast_id, episode = self._match_valid_url(url).groups()
-            return s.replace("&quot;", '"')
+        video_id = join_nonempty(telecast_id, episode, delim='_')
-
-        video_id = self._match_id(url)
         webpage = self._download_webpage(url, video_id)
-        manifest_url = self._search_regex(r'<video-player[^>]+manifest-url="([^"]+)"', webpage, 'video-player')
+        video_player = try_call(lambda: extract_attributes(
-        episode_attr = self._search_regex(r'<video-player[^>]+:episode="([^"]+)"', webpage, 'episode data')
+            get_element_text_and_html_by_tag('video-player', webpage)[1]))
-        ep = self._parse_json(episode_attr, video_id, decode_quot)
+        if not video_player or not video_player.get('manifest-url'):
+            self.raise_no_formats('No video found', expected=True)
+
+        episode_attr = self._parse_json(video_player.get(':episode') or '', video_id, fatal=False) or {}
 
         return {
             'id': video_id,
-            # fallback to absolute "episode_id" value
+            'formats': self._extract_m3u8_formats(video_player['manifest-url'], video_id, 'mp4'),
-            'title': traverse_obj(ep, 'subtitle') or f"Episode {traverse_obj(ep, 'episode_id')}",
+            **traverse_obj(episode_attr, {
-            'description': strip_or_none(traverse_obj(ep, 'synopsis')),
+                'title': (None, ('subtitle', ('episode_id', {lambda x: f'Episode {x}'}))),
-            'thumbnail': traverse_obj(ep, ('images', 'original')),
+                'description': 'synopsis',
-            'formats': self._extract_m3u8_formats(manifest_url, video_id, 'mp4'),
+                'thumbnail': ('images', 'original'),
-            'timestamp': unified_timestamp(traverse_obj(ep, 'airtime') + ' +0200'),
+                'timestamp': ('airtime', {lambda x: unified_timestamp(x + ' +0200')}),
-            'series': traverse_obj(ep, 'title'),
+                'series': 'title',
-            'series_id': traverse_obj(ep, 'telecast_id'),
+                'series_id': ('telecast_id', {str_or_none}),
-            'season_number': traverse_obj(ep, 'season_id'),
+                'season_number': ('season_id', {int_or_none}),
-            'episode': traverse_obj(ep, 'subtitle'),
+                'episode': 'subtitle',
-            # fallback to absolute "episode_id" value
+                'episode_number': (None, ('episode_nr', 'episode_id'), {int_or_none}),
-            'episode_number': traverse_obj(ep, 'episode_nr') or traverse_obj(ep, 'episode_id'),
+            }, get_all=False),
-            'episode_id': traverse_obj(ep, 'episode_id'),
         }