Merge eecd76dd92 into 96da952504

Authored by: seproDev, pukkandan

yt_dlp/YoutubeDL.py

@@ -2136,6 +2136,11 @@ class YoutubeDL:
 
     def _check_formats(self, formats):
         for f in formats:
+            working = f.get('__working')
+            if working is not None:
+                if working:
+                    yield f
+                continue
             self.to_screen('[info] Testing format %s' % f['format_id'])
             path = self.get_output_path('temp')
             if not self._ensure_dir_exists(f'{path}/'):
@@ -2152,33 +2157,44 @@ class YoutubeDL:
                         os.remove(temp_file.name)
                     except OSError:
                         self.report_warning('Unable to delete temporary file "%s"' % temp_file.name)
+            f['__working'] = success
             if success:
                 yield f
             else:
                 self.to_screen('[info] Unable to download format %s. Skipping...' % f['format_id'])
 
+    def _select_formats(self, formats, selector):
+        return list(selector({
+            'formats': formats,
+            'has_merged_format': any('none' not in (f.get('acodec'), f.get('vcodec')) for f in formats),
+            'incomplete_formats': (all(f.get('vcodec') == 'none' for f in formats)  # No formats with video
+                                   or all(f.get('acodec') == 'none' for f in formats)),  # OR, No formats with audio
+        }))
+
     def _default_format_spec(self, info_dict, download=True):
+        download = download and not self.params.get('simulate')
+        prefer_best = download and (
+            self.params['outtmpl']['default'] == '-'
+            or info_dict.get('is_live') and not self.params.get('live_from_start'))
 
         def can_merge():
             merger = FFmpegMergerPP(self)
             return merger.available and merger.can_merge()
 
-        prefer_best = (
-            not self.params.get('simulate')
-            and download
-            and (
-                not can_merge()
-                or info_dict.get('is_live') and not self.params.get('live_from_start')
-                or self.params['outtmpl']['default'] == '-'))
-        compat = (
-            prefer_best
-            or self.params.get('allow_multiple_audio_streams', False)
-            or 'format-spec' in self.params['compat_opts'])
+        if not prefer_best and download and not can_merge():
+            prefer_best = True
+            formats = self._get_formats(info_dict)
+            evaluate_formats = lambda spec: self._select_formats(formats, self.build_format_selector(spec))
+            if evaluate_formats('b/bv+ba') != evaluate_formats('bv*+ba/b'):
+                self.report_warning('ffmpeg not found. The downloaded format may not be the best available. '
+                                    'Installing ffmpeg is strongly recommended: https://github.com/yt-dlp/yt-dlp#dependencies')
 
-        return (
-            'best/bestvideo+bestaudio' if prefer_best
-            else 'bestvideo*+bestaudio/best' if not compat
-            else 'bestvideo+bestaudio/best')
+        compat = (self.params.get('allow_multiple_audio_streams')
+                  or 'format-spec' in self.params['compat_opts'])
+
+        return ('best/bestvideo+bestaudio' if prefer_best
+                else 'bestvideo+bestaudio/best' if compat
+                else 'bestvideo*+bestaudio/best')
 
     def build_format_selector(self, format_spec):
         def syntax_error(note, start):
@@ -2928,12 +2944,7 @@ class YoutubeDL:
                 self.write_debug(f'Default format spec: {req_format}')
                 format_selector = self.build_format_selector(req_format)
 
-            formats_to_download = list(format_selector({
-                'formats': formats,
-                'has_merged_format': any('none' not in (f.get('acodec'), f.get('vcodec')) for f in formats),
-                'incomplete_formats': (all(f.get('vcodec') == 'none' for f in formats)  # No formats with video
-                                       or all(f.get('acodec') == 'none' for f in formats)),  # OR, No formats with audio
-            }))
+            formats_to_download = self._select_formats(formats, format_selector)
             if interactive_format_selection and not formats_to_download:
                 self.report_error('Requested format is not available', tb=False, is_error=False)
                 continue

yt_dlp/extractor/_extractors.py

@@ -2379,6 +2379,7 @@ from .xhamster import (
     XHamsterEmbedIE,
     XHamsterUserIE,
 )
+from .xiaohongshu import XiaoHongShuIE
 from .ximalaya import (
     XimalayaIE,
     XimalayaAlbumIE

yt_dlp/extractor/xiaohongshu.py

@@ -0,0 +1,85 @@
+from .common import InfoExtractor
+from ..utils import (
+    float_or_none,
+    js_to_json,
+    url_or_none,
+)
+from ..utils.traversal import traverse_obj
+
+
+class XiaoHongShuIE(InfoExtractor):
+    _VALID_URL = r'https?://www\.xiaohongshu.com/explore/(?P<id>[a-f0-9]+)'
+    IE_DESC = '小红书'
+    _TESTS = [{
+        'url': 'https://www.xiaohongshu.com/explore/6411cf99000000001300b6d9',
+        'md5': '2a87a77ddbedcaeeda8d7eae61b61228',
+        'info_dict': {
+            'id': '6411cf99000000001300b6d9',
+            'ext': 'mp4',
+            'uploader_id': '5c31698d0000000007018a31',
+            'description': '#今日快乐今日发[话题]# #吃货薯看这里[话题]# #香妃蛋糕[话题]# #小五卷蛋糕[话题]# #新手蛋糕卷[话题]#',
+            'title': '香妃蛋糕也太香了吧🔥不需要卷❗️绝对的友好',
+            'tags': ['今日快乐今日发', '吃货薯看这里', '香妃蛋糕', '小五卷蛋糕', '新手蛋糕卷'],
+            'duration': 101.726,
+            'thumbnail': r're:https?://sns-webpic-qc\.xhscdn\.com/\d+/[a-z0-9]+/[\w]+',
+        }
+    }]
+
+    def _real_extract(self, url):
+        display_id = self._match_id(url)
+        webpage = self._download_webpage(url, display_id)
+        initial_state = self._search_json(
+            r'window\.__INITIAL_STATE__\s*=', webpage, 'initial state', display_id, transform_source=js_to_json)
+
+        note_info = traverse_obj(initial_state, ('note', 'noteDetailMap', display_id, 'note'))
+        video_info = traverse_obj(note_info, ('video', 'media', 'stream', ('h264', 'av1', 'h265'), ...))
+
+        formats = []
+        for info in video_info:
+            format_info = traverse_obj(info, {
+                'fps': 'fps',
+                'width': 'width',
+                'height': 'height',
+                'vcodec': 'videoCodec',
+                'acodec': 'audioCodec',
+                'abr': 'audioBitrate',
+                'vbr': 'videoBitrate',
+                'audio_channels': 'audioChannels',
+                'tbr': 'avgBitrate',
+                'format': 'qualityType',
+                'filesize': 'size',
+                'duration': ('duration', {lambda x: float_or_none(x, scale=1000)})
+            })
+
+            formats.extend(traverse_obj(info, (('mediaUrl', ('backupUrls', ...)), {
+                lambda url: url_or_none(url) and {'url': url, 'ext': 'mp4', **format_info}})))
+
+        thumbnails = []
+        for image_info in traverse_obj(note_info, ('imageList', ...)):
+            thumbnail_info = traverse_obj(image_info, {
+                'height': 'height',
+                'width': 'width'
+            })
+            for url in traverse_obj(image_info, (('urlDefault', 'urlPre'), {url_or_none})):
+                thumbnails.append({
+                    'url': url,
+                    **thumbnail_info
+                })
+
+        return {
+            'id': display_id,
+            'formats': formats or [{
+                'url': self._html_search_meta(['og:video'], webpage, fatal=True),
+                'ext': 'mp4'
+            }],
+            'thumbnails': thumbnails or [{
+                'url': self._html_search_meta(['og:image'], webpage, default=None)
+            }],
+            'title': self._html_search_meta(['og:title'], webpage, default=None),
+            **traverse_obj(note_info, {
+                'title': 'title',
+                'description': 'desc',
+                'tags': ('tagList', ..., 'name', {str}),
+                'uploader_id': ('user', 'userId'),
+            }),
+        }