[nova:embed] Fix extraction (closes #24700)

youtube_dl/extractor/nova.py

@@ -6,6 +6,7 @@
 from .common import InfoExtractor
 from ..utils import (
     clean_html,
+    determine_ext,
     int_or_none,
     js_to_json,
     qualities,
@@ -33,42 +34,76 @@ def _real_extract(self, url):
 
         webpage = self._download_webpage(url, video_id)
 
-        bitrates = self._parse_json(
+        duration = None
-            self._search_regex(
-                r'(?s)(?:src|bitrates)\s*=\s*({.+?})\s*;', webpage, 'formats'),
-            video_id, transform_source=js_to_json)
-
-        QUALITIES = ('lq', 'mq', 'hq', 'hd')
-        quality_key = qualities(QUALITIES)
-
         formats = []
-        for format_id, format_list in bitrates.items():
+
-            if not isinstance(format_list, list):
+        player = self._parse_json(
-                format_list = [format_list]
+            self._search_regex(
-            for format_url in format_list:
+                r'Player\.init\s*\([^,]+,\s*({.+?})\s*,\s*{.+?}\s*\)\s*;',
-                format_url = url_or_none(format_url)
+                webpage, 'player', default='{}'), video_id, fatal=False)
-                if not format_url:
+        if player:
-                    continue
+            for format_id, format_list in player['tracks'].items():
-                if format_id == 'hls':
+                if not isinstance(format_list, list):
-                    formats.extend(self._extract_m3u8_formats(
+                    format_list = [format_list]
-                        format_url, video_id, ext='mp4',
+                for format_dict in format_list:
-                        entry_protocol='m3u8_native', m3u8_id='hls',
+                    if not isinstance(format_dict, dict):
-                        fatal=False))
+                        continue
-                    continue
+                    format_url = url_or_none(format_dict.get('src'))
-                f = {
+                    format_type = format_dict.get('type')
-                    'url': format_url,
+                    ext = determine_ext(format_url)
-                }
+                    if (format_type == 'application/x-mpegURL'
-                f_id = format_id
+                            or format_id == 'HLS' or ext == 'm3u8'):
-                for quality in QUALITIES:
+                        formats.extend(self._extract_m3u8_formats(
-                    if '%s.mp4' % quality in format_url:
+                            format_url, video_id, 'mp4',
-                        f_id += '-%s' % quality
+                            entry_protocol='m3u8_native', m3u8_id='hls',
-                        f.update({
+                            fatal=False))
-                            'quality': quality_key(quality),
+                    elif (format_type == 'application/dash+xml'
-                            'format_note': quality.upper(),
+                          or format_id == 'DASH' or ext == 'mpd'):
+                        formats.extend(self._extract_mpd_formats(
+                            format_url, video_id, mpd_id='dash', fatal=False))
+                    else:
+                        formats.append({
+                            'url': format_url,
                         })
-                        break
+            duration = int_or_none(player.get('duration'))
-                f['format_id'] = f_id
+        else:
-                formats.append(f)
+            # Old path, not actual as of 08.04.2020
+            bitrates = self._parse_json(
+                self._search_regex(
+                    r'(?s)(?:src|bitrates)\s*=\s*({.+?})\s*;', webpage, 'formats'),
+                video_id, transform_source=js_to_json)
+
+            QUALITIES = ('lq', 'mq', 'hq', 'hd')
+            quality_key = qualities(QUALITIES)
+
+            for format_id, format_list in bitrates.items():
+                if not isinstance(format_list, list):
+                    format_list = [format_list]
+                for format_url in format_list:
+                    format_url = url_or_none(format_url)
+                    if not format_url:
+                        continue
+                    if format_id == 'hls':
+                        formats.extend(self._extract_m3u8_formats(
+                            format_url, video_id, ext='mp4',
+                            entry_protocol='m3u8_native', m3u8_id='hls',
+                            fatal=False))
+                        continue
+                    f = {
+                        'url': format_url,
+                    }
+                    f_id = format_id
+                    for quality in QUALITIES:
+                        if '%s.mp4' % quality in format_url:
+                            f_id += '-%s' % quality
+                            f.update({
+                                'quality': quality_key(quality),
+                                'format_note': quality.upper(),
+                            })
+                            break
+                    f['format_id'] = f_id
+                    formats.append(f)
+
         self._sort_formats(formats)
 
         title = self._og_search_title(
@@ -81,7 +116,8 @@ def _real_extract(self, url):
             r'poster\s*:\s*(["\'])(?P<value>(?:(?!\1).)+)\1', webpage,
             'thumbnail', fatal=False, group='value')
         duration = int_or_none(self._search_regex(
-            r'videoDuration\s*:\s*(\d+)', webpage, 'duration', fatal=False))
+            r'videoDuration\s*:\s*(\d+)', webpage, 'duration',
+            default=duration))
 
         return {
             'id': video_id,