yt_dlp/extractor/viously.py

@@ -1,9 +1,9 @@
 import base64
-import re
 
 from .common import InfoExtractor
 from ..utils import (
     extract_attributes,
+    get_elements_html_by_class,
     int_or_none,
     parse_iso8601,
 )
@@ -18,18 +18,16 @@ class ViouslyIE(InfoExtractor):
         'info_dict': {
             'id': 'F_xQzS2jwb3',
             'ext': 'mp4',
-            'title': 'Turbo du 07/09/2014\xa0: Renault Twingo 3, Bentley Continental GT Speed, CES, Guide Achat Dacia...',
+            'title': 'Turbo du 07/09/2014 : Renault Twingo 3, Bentley Continental GT Speed, CES, Guide Achat Dacia...',
-            'description': 'Turbo du 07/09/2014\xa0: Renault Twingo 3, Bentley Continental GT Speed, CES, Guide Achat Dacia...',
+            'description': 'Turbo du 07/09/2014 : Renault Twingo 3, Bentley Continental GT Speed, CES, Guide Achat Dacia...',
             'age_limit': 0,
-            'upload_date': '20230328',
+            'upload_date': str,
-            'timestamp': 1680037507,
+            'timestamp': float,
-            'duration': 3716,
-            'categories': ['motors'],
         }
     }]
 
     def _extract_from_webpage(self, url, webpage):
-        viously_players = re.findall(r'<div[^>]*class="(?:[^"]*\s)?v(?:iou)?sly-player(?:\s[^"]*)?"[^>]*>', webpage)
+        viously_players = get_elements_html_by_class('viously-player', webpage) + get_elements_html_by_class('vsly-player', webpage)
         if not viously_players:
             return
 
@@ -42,19 +40,19 @@ class ViouslyIE(InfoExtractor):
         for video_id in traverse_obj(viously_players, (..., {extract_attributes}, 'id')):
             formats = self._extract_m3u8_formats(
                 f'https://www.viously.com/video/hls/{video_id}/index.m3u8', video_id, fatal=False)
-            if not formats:
-                continue
             data = self._download_json(
                 f'https://www.viously.com/export/json/{video_id}', video_id,
                 transform_source=custom_decode, fatal=False)
+            if not formats or not data:
+                continue
             yield {
                 'id': video_id,
                 'formats': formats,
                 **traverse_obj(data, ('video', {
-                    'title': ('title', {str}),
+                    'title': 'title',
-                    'description': ('description', {str}),
+                    'description': 'description',
                     'duration': ('duration', {int_or_none}),
                     'timestamp': ('iso_date', {parse_iso8601}),
-                    'categories': ('category', 'name', {str}, {lambda x: [x] if x else None}),
+                    'categories': ('category', {lambda x: [x['name']]}),
                 })),
             }