Merge 6c8ede8188 into 96da952504

Authored by: seproDev, pukkandan

test/test_YoutubeDL.py

@@ -874,32 +874,33 @@ class TestYoutubeDL(unittest.TestCase):
         }), r'^30fps$')
 
     def test_postprocessors(self):
-        filename = 'post-processor-testfile.mp4'
-        audiofile = filename + '.mp3'
+        filename = 'post-processor-testfile'
+        video_file = filename + '.mp4'
+        audio_file = filename + '.mp3'
 
         class SimplePP(PostProcessor):
             def run(self, info):
-                with open(audiofile, 'w') as f:
+                with open(audio_file, 'w') as f:
                     f.write('EXAMPLE')
                 return [info['filepath']], info
 
         def run_pp(params, PP):
-            with open(filename, 'w') as f:
+            with open(video_file, 'w') as f:
                 f.write('EXAMPLE')
             ydl = YoutubeDL(params)
             ydl.add_post_processor(PP())
-            ydl.post_process(filename, {'filepath': filename})
+            ydl.post_process(video_file, {'filepath': video_file})
 
-        run_pp({'keepvideo': True}, SimplePP)
-        self.assertTrue(os.path.exists(filename), '%s doesn\'t exist' % filename)
-        self.assertTrue(os.path.exists(audiofile), '%s doesn\'t exist' % audiofile)
-        os.unlink(filename)
-        os.unlink(audiofile)
+        run_pp({'keepvideo': True, 'outtmpl': filename}, SimplePP)
+        self.assertTrue(os.path.exists(video_file), '%s doesn\'t exist' % video_file)
+        self.assertTrue(os.path.exists(audio_file), '%s doesn\'t exist' % audio_file)
+        os.unlink(video_file)
+        os.unlink(audio_file)
 
-        run_pp({'keepvideo': False}, SimplePP)
-        self.assertFalse(os.path.exists(filename), '%s exists' % filename)
-        self.assertTrue(os.path.exists(audiofile), '%s doesn\'t exist' % audiofile)
-        os.unlink(audiofile)
+        run_pp({'keepvideo': False, 'outtmpl': filename}, SimplePP)
+        self.assertFalse(os.path.exists(video_file), '%s exists' % video_file)
+        self.assertTrue(os.path.exists(audio_file), '%s doesn\'t exist' % audio_file)
+        os.unlink(audio_file)
 
         class ModifierPP(PostProcessor):
             def run(self, info):
@@ -907,9 +908,9 @@ class TestYoutubeDL(unittest.TestCase):
                     f.write('MODIFIED')
                 return [], info
 
-        run_pp({'keepvideo': False}, ModifierPP)
-        self.assertTrue(os.path.exists(filename), '%s doesn\'t exist' % filename)
-        os.unlink(filename)
+        run_pp({'keepvideo': False, 'outtmpl': filename}, ModifierPP)
+        self.assertTrue(os.path.exists(video_file), '%s doesn\'t exist' % video_file)
+        os.unlink(video_file)
 
     def test_match_filter(self):
         first = {

test/test_networking.py

@@ -785,6 +785,25 @@ class TestHTTPImpersonateRequestHandler(TestRequestHandlerBase):
                 assert res.status == 200
                 assert std_headers['user-agent'].lower() not in res.read().decode().lower()
 
+    def test_response_extensions(self, handler):
+        with handler() as rh:
+            for target in rh.supported_targets:
+                request = Request(
+                    f'http://127.0.0.1:{self.http_port}/gen_200', extensions={'impersonate': target})
+                res = validate_and_send(rh, request)
+                assert res.extensions['impersonate'] == rh._get_request_target(request)
+
+    def test_http_error_response_extensions(self, handler):
+        with handler() as rh:
+            for target in rh.supported_targets:
+                request = Request(
+                    f'http://127.0.0.1:{self.http_port}/gen_404', extensions={'impersonate': target})
+                try:
+                    validate_and_send(rh, request)
+                except HTTPError as e:
+                    res = e.response
+                assert res.extensions['impersonate'] == rh._get_request_target(request)
+
 
 class TestRequestHandlerMisc:
     """Misc generic tests for request handlers, not related to request or validation testing"""

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
@@ -3208,7 +3219,6 @@ class YoutubeDL:
         # info_dict['_filename'] needs to be set for backward compatibility
         info_dict['_filename'] = full_filename = self.prepare_filename(info_dict, warn=True)
         temp_filename = self.prepare_filename(info_dict, 'temp')
-        files_to_move = {}
 
         # Forced printings
         self.__forced_printings(info_dict, full_filename, incomplete=('format' not in info_dict))
@@ -3236,13 +3246,11 @@ class YoutubeDL:
         sub_files = self._write_subtitles(info_dict, temp_filename)
         if sub_files is None:
             return
-        files_to_move.update(dict(sub_files))
 
         thumb_files = self._write_thumbnails(
             'video', info_dict, temp_filename, self.prepare_filename(info_dict, 'thumbnail'))
         if thumb_files is None:
             return
-        files_to_move.update(dict(thumb_files))
 
         infofn = self.prepare_filename(info_dict, 'infojson')
         _infojson_written = self._write_info_json('video', info_dict, infofn)
@@ -3316,13 +3324,12 @@ class YoutubeDL:
                for link_type, should_write in write_links.items()):
             return
 
-        new_info, files_to_move = self.pre_process(info_dict, 'before_dl', files_to_move)
+        new_info, _ = self.pre_process(info_dict, 'before_dl')
         replace_info_dict(new_info)
 
         if self.params.get('skip_download'):
             info_dict['filepath'] = temp_filename
             info_dict['__finaldir'] = os.path.dirname(os.path.abspath(encodeFilename(full_filename)))
-            info_dict['__files_to_move'] = files_to_move
             replace_info_dict(self.run_pp(MoveFilesAfterDownloadPP(self, False), info_dict))
             info_dict['__write_download_archive'] = self.params.get('force_write_download_archive')
         else:
@@ -3436,9 +3443,6 @@ class YoutubeDL:
                         info_dict['__files_to_merge'] = downloaded
                         # Even if there were no downloads, it is being merged only now
                         info_dict['__real_download'] = True
-                    else:
-                        for file in downloaded:
-                            files_to_move[file] = None
                 else:
                     # Just a single file
                     dl_filename = existing_video_file(full_filename, temp_filename)
@@ -3452,7 +3456,6 @@ class YoutubeDL:
 
                 dl_filename = dl_filename or temp_filename
                 info_dict['__finaldir'] = os.path.dirname(os.path.abspath(encodeFilename(full_filename)))
-
             except network_exceptions as err:
                 self.report_error('unable to download video data: %s' % error_to_compat_str(err))
                 return
@@ -3523,7 +3526,7 @@ class YoutubeDL:
 
                 fixup()
                 try:
-                    replace_info_dict(self.post_process(dl_filename, info_dict, files_to_move))
+                    replace_info_dict(self.post_process(dl_filename, info_dict))
                 except PostProcessingError as err:
                     self.report_error('Postprocessing: %s' % str(err))
                     return
@@ -3644,8 +3647,6 @@ class YoutubeDL:
                 os.remove(filename)
             except OSError:
                 self.report_warning(f'Unable to delete file {filename}')
-            if filename in info.get('__files_to_move', []):  # NB: Delete even if None
-                del info['__files_to_move'][filename]
 
     @staticmethod
     def post_extract(info_dict):
@@ -3662,8 +3663,7 @@ class YoutubeDL:
 
     def run_pp(self, pp, infodict):
         files_to_delete = []
-        if '__files_to_move' not in infodict:
-            infodict['__files_to_move'] = {}
+
         try:
             files_to_delete, infodict = pp.run(infodict)
         except PostProcessingError as e:
@@ -3675,10 +3675,7 @@ class YoutubeDL:
 
         if not files_to_delete:
             return infodict
-        if self.params.get('keepvideo', False):
-            for f in files_to_delete:
-                infodict['__files_to_move'].setdefault(f, '')
-        else:
+        if not self.params.get('keepvideo', False):
             self._delete_downloaded_files(
                 *files_to_delete, info=infodict, msg='Deleting original file %s (pass -k to keep)')
         return infodict
@@ -3691,23 +3688,27 @@ class YoutubeDL:
         return info
 
     def pre_process(self, ie_info, key='pre_process', files_to_move=None):
+        if files_to_move is not None:
+            self.report_warning('[pre_process] "files_to_move" is deprecated and may be removed in a future version')
+
         info = dict(ie_info)
-        info['__files_to_move'] = files_to_move or {}
         try:
             info = self.run_all_pps(key, info)
         except PostProcessingError as err:
             msg = f'Preprocessing: {err}'
             info.setdefault('__pending_error', msg)
             self.report_error(msg, is_error=False)
-        return info, info.pop('__files_to_move', None)
+        return info, files_to_move
 
     def post_process(self, filename, info, files_to_move=None):
         """Run all the postprocessors on the given file."""
+        if files_to_move is not None:
+            self.report_warning('[post_process] "files_to_move" is deprecated and may be removed in a future version')
+
         info['filepath'] = filename
-        info['__files_to_move'] = files_to_move or {}
         info = self.run_all_pps('post_process', info, additional_pps=info.get('__postprocessors'))
         info = self.run_pp(MoveFilesAfterDownloadPP(self), info)
-        del info['__files_to_move']
+        info.pop('__multiple_thumbnails', None)
         return self.run_all_pps('after_move', info)
 
     def _make_archive_id(self, info_dict):
@@ -4294,10 +4295,11 @@ class YoutubeDL:
             sub_filename = subtitles_filename(filename, sub_lang, sub_format, info_dict.get('ext'))
             sub_filename_final = subtitles_filename(sub_filename_base, sub_lang, sub_format, info_dict.get('ext'))
             existing_sub = self.existing_file((sub_filename_final, sub_filename))
+
             if existing_sub:
                 self.to_screen(f'[info] Video subtitle {sub_lang}.{sub_format} is already present')
                 sub_info['filepath'] = existing_sub
-                ret.append((existing_sub, sub_filename_final))
+                ret.append(existing_sub)
                 continue
 
             self.to_screen(f'[info] Writing video subtitles to: {sub_filename}')
@@ -4308,7 +4310,7 @@ class YoutubeDL:
                     with open(sub_filename, 'w', encoding='utf-8', newline='') as subfile:
                         subfile.write(sub_info['data'])
                     sub_info['filepath'] = sub_filename
-                    ret.append((sub_filename, sub_filename_final))
+                    ret.append(sub_filename)
                     continue
                 except OSError:
                     self.report_error(f'Cannot write video subtitles file {sub_filename}')
@@ -4319,7 +4321,7 @@ class YoutubeDL:
                 sub_copy.setdefault('http_headers', info_dict.get('http_headers'))
                 self.dl(sub_filename, sub_copy, subtitle=True)
                 sub_info['filepath'] = sub_filename
-                ret.append((sub_filename, sub_filename_final))
+                ret.append(sub_filename)
             except (DownloadError, ExtractorError, IOError, OSError, ValueError) + network_exceptions as err:
                 msg = f'Unable to download video subtitles for {sub_lang!r}: {err}'
                 if self.params.get('ignoreerrors') is not True:  # False or 'only_download'
@@ -4339,6 +4341,7 @@ class YoutubeDL:
                 self.to_screen(f'[info] There are no {label} thumbnails to download')
                 return ret
         multiple = write_all and len(thumbnails) > 1
+        info_dict['__multiple_thumbnails'] = multiple
 
         if thumb_filename_base is None:
             thumb_filename_base = filename
@@ -4360,7 +4363,7 @@ class YoutubeDL:
                 self.to_screen('[info] %s is already present' % (
                     thumb_display_id if multiple else f'{label} thumbnail').capitalize())
                 t['filepath'] = existing_thumb
-                ret.append((existing_thumb, thumb_filename_final))
+                ret.append(existing_thumb)
             else:
                 self.to_screen(f'[info] Downloading {thumb_display_id} ...')
                 try:
@@ -4368,7 +4371,7 @@ class YoutubeDL:
                     self.to_screen(f'[info] Writing {thumb_display_id} to: {thumb_filename}')
                     with open(encodeFilename(thumb_filename), 'wb') as thumbf:
                         shutil.copyfileobj(uf, thumbf)
-                    ret.append((thumb_filename, thumb_filename_final))
+                    ret.append(thumb_filename)
                     t['filepath'] = thumb_filename
                 except network_exceptions as err:
                     if isinstance(err, HTTPError) and err.status == 404:
@@ -4378,4 +4381,5 @@ class YoutubeDL:
                     thumbnails.pop(idx)
             if ret and not write_all:
                 break
+
         return ret

yt_dlp/extractor/patreon.py

@@ -219,7 +219,29 @@ class PatreonIE(PatreonBaseIE):
             'thumbnail': r're:^https?://.+',
         },
         'params': {'skip_download': 'm3u8'},
+    }, {
+        # multiple attachments/embeds
+        'url': 'https://www.patreon.com/posts/holy-wars-solos-100601977',
+        'playlist_count': 3,
+        'info_dict': {
+            'id': '100601977',
+            'title': '"Holy Wars" (Megadeth) Solos Transcription & Lesson/Analysis',
+            'description': 'md5:d099ab976edfce6de2a65c2b169a88d3',
+            'uploader': 'Bradley Hall',
+            'uploader_id': '24401883',
+            'uploader_url': 'https://www.patreon.com/bradleyhallguitar',
+            'channel_id': '3193932',
+            'channel_url': 'https://www.patreon.com/bradleyhallguitar',
+            'channel_follower_count': int,
+            'timestamp': 1710777855,
+            'upload_date': '20240318',
+            'like_count': int,
+            'comment_count': int,
+            'thumbnail': r're:^https?://.+',
+        },
+        'skip': 'Patron-only content',
     }]
+    _RETURN_TYPE = 'video'
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
@@ -234,58 +256,54 @@ class PatreonIE(PatreonBaseIE):
                 'include': 'audio,user,user_defined_tags,campaign,attachments_media',
             })
         attributes = post['data']['attributes']
-        title = attributes['title'].strip()
-        image = attributes.get('image') or {}
-        info = {
-            'id': video_id,
-            'title': title,
-            'description': clean_html(attributes.get('content')),
-            'thumbnail': image.get('large_url') or image.get('url'),
-            'timestamp': parse_iso8601(attributes.get('published_at')),
-            'like_count': int_or_none(attributes.get('like_count')),
-            'comment_count': int_or_none(attributes.get('comment_count')),
-        }
-        can_view_post = traverse_obj(attributes, 'current_user_can_view')
-        if can_view_post and info['comment_count']:
-            info['__post_extractor'] = self.extract_comments(video_id)
+        info = traverse_obj(attributes, {
+            'title': ('title', {str.strip}),
+            'description': ('content', {clean_html}),
+            'thumbnail': ('image', ('large_url', 'url'), {url_or_none}, any),
+            'timestamp': ('published_at', {parse_iso8601}),
+            'like_count': ('like_count', {int_or_none}),
+            'comment_count': ('comment_count', {int_or_none}),
+        })
 
-        for i in post.get('included', []):
-            i_type = i.get('type')
-            if i_type == 'media':
-                media_attributes = i.get('attributes') or {}
-                download_url = media_attributes.get('download_url')
+        entries = []
+        idx = 0
+        for include in traverse_obj(post, ('included', lambda _, v: v['type'])):
+            include_type = include['type']
+            if include_type == 'media':
+                media_attributes = traverse_obj(include, ('attributes', {dict})) or {}
+                download_url = url_or_none(media_attributes.get('download_url'))
                 ext = mimetype2ext(media_attributes.get('mimetype'))
 
                 # if size_bytes is None, this media file is likely unavailable
                 # See: https://github.com/yt-dlp/yt-dlp/issues/4608
                 size_bytes = int_or_none(media_attributes.get('size_bytes'))
                 if download_url and ext in KNOWN_EXTENSIONS and size_bytes is not None:
-                    # XXX: what happens if there are multiple attachments?
-                    return {
-                        **info,
+                    idx += 1
+                    entries.append({
+                        'id': f'{video_id}-{idx}',
                         'ext': ext,
                         'filesize': size_bytes,
                         'url': download_url,
-                    }
-            elif i_type == 'user':
-                user_attributes = i.get('attributes')
-                if user_attributes:
-                    info.update({
-                        'uploader': user_attributes.get('full_name'),
-                        'uploader_id': str_or_none(i.get('id')),
-                        'uploader_url': user_attributes.get('url'),
                     })
 
-            elif i_type == 'post_tag':
-                info.setdefault('tags', []).append(traverse_obj(i, ('attributes', 'value')))
+            elif include_type == 'user':
+                info.update(traverse_obj(include, {
+                    'uploader': ('attributes', 'full_name', {str}),
+                    'uploader_id': ('id', {str_or_none}),
+                    'uploader_url': ('attributes', 'url', {url_or_none}),
+                }))
 
-            elif i_type == 'campaign':
-                info.update({
-                    'channel': traverse_obj(i, ('attributes', 'title')),
-                    'channel_id': str_or_none(i.get('id')),
-                    'channel_url': traverse_obj(i, ('attributes', 'url')),
-                    'channel_follower_count': int_or_none(traverse_obj(i, ('attributes', 'patron_count'))),
-                })
+            elif include_type == 'post_tag':
+                if post_tag := traverse_obj(include, ('attributes', 'value', {str})):
+                    info.setdefault('tags', []).append(post_tag)
+
+            elif include_type == 'campaign':
+                info.update(traverse_obj(include, {
+                    'channel': ('attributes', 'title', {str}),
+                    'channel_id': ('id', {str_or_none}),
+                    'channel_url': ('attributes', 'url', {url_or_none}),
+                    'channel_follower_count': ('attributes', 'patron_count', {int_or_none}),
+                }))
 
         # handle Vimeo embeds
         if traverse_obj(attributes, ('embed', 'provider')) == 'Vimeo':
@@ -296,36 +314,50 @@ class PatreonIE(PatreonBaseIE):
                     v_url, video_id, 'Checking Vimeo embed URL',
                     headers={'Referer': 'https://patreon.com/'},
                     fatal=False, errnote=False):
-                return self.url_result(
+                entries.append(self.url_result(
                     VimeoIE._smuggle_referrer(v_url, 'https://patreon.com/'),
-                    VimeoIE, url_transparent=True, **info)
+                    VimeoIE, url_transparent=True))
 
         embed_url = traverse_obj(attributes, ('embed', 'url', {url_or_none}))
         if embed_url and self._request_webpage(embed_url, video_id, 'Checking embed URL', fatal=False, errnote=False):
-            return self.url_result(embed_url, **info)
+            entries.append(self.url_result(embed_url))
 
-        post_file = traverse_obj(attributes, 'post_file')
+        post_file = traverse_obj(attributes, ('post_file', {dict}))
         if post_file:
             name = post_file.get('name')
             ext = determine_ext(name)
             if ext in KNOWN_EXTENSIONS:
-                return {
-                    **info,
+                entries.append({
+                    'id': video_id,
                     'ext': ext,
                     'url': post_file['url'],
-                }
+                })
             elif name == 'video' or determine_ext(post_file.get('url')) == 'm3u8':
                 formats, subtitles = self._extract_m3u8_formats_and_subtitles(post_file['url'], video_id)
-                return {
-                    **info,
+                entries.append({
+                    'id': video_id,
                     'formats': formats,
                     'subtitles': subtitles,
-                }
+                })
 
-        if can_view_post is False:
+        can_view_post = traverse_obj(attributes, 'current_user_can_view')
+        comments = None
+        if can_view_post and info.get('comment_count'):
+            comments = self.extract_comments(video_id)
+
+        if not entries and can_view_post is False:
             self.raise_no_formats('You do not have access to this post', video_id=video_id, expected=True)
-        else:
+        elif not entries:
             self.raise_no_formats('No supported media found in this post', video_id=video_id, expected=True)
+        elif len(entries) == 1:
+            info.update(entries[0])
+        else:
+            for entry in entries:
+                entry.update(info)
+            return self.playlist_result(entries, video_id, **info, __post_extractor=comments)
+
+        info['id'] = video_id
+        info['__post_extractor'] = comments
         return info
 
     def _get_comments(self, post_id):

yt_dlp/networking/_curlcffi.py

@@ -132,6 +132,16 @@ class CurlCFFIRH(ImpersonateRequestHandler, InstanceStoreMixin):
         extensions.pop('cookiejar', None)
         extensions.pop('timeout', None)
 
+    def send(self, request: Request) -> Response:
+        target = self._get_request_target(request)
+        try:
+            response = super().send(request)
+        except HTTPError as e:
+            e.response.extensions['impersonate'] = target
+            raise
+        response.extensions['impersonate'] = target
+        return response
+
     def _send(self, request: Request):
         max_redirects_exceeded = False
         session: curl_cffi.requests.Session = self._get_instance(

yt_dlp/networking/common.py

@@ -497,6 +497,7 @@ class Response(io.IOBase):
     @param headers: response headers.
     @param status: Response HTTP status code. Default is 200 OK.
     @param reason: HTTP status reason. Will use built-in reasons based on status code if not provided.
+    @param extensions: Dictionary of handler-specific response extensions.
     """
 
     def __init__(
@@ -505,7 +506,9 @@ class Response(io.IOBase):
             url: str,
             headers: Mapping[str, str],
             status: int = 200,
-            reason: str = None):
+            reason: str = None,
+            extensions: dict = None
+    ):
 
         self.fp = fp
         self.headers = Message()
@@ -517,6 +520,7 @@ class Response(io.IOBase):
             self.reason = reason or HTTPStatus(status).phrase
         except ValueError:
             self.reason = None
+        self.extensions = extensions or {}
 
     def readable(self):
         return self.fp.readable()

yt_dlp/postprocessor/embedthumbnail.py

@@ -224,4 +224,8 @@ class EmbedThumbnailPP(FFmpegPostProcessor):
             thumbnail_filename if converted or not self._already_have_thumbnail else None,
             original_thumbnail if converted and not self._already_have_thumbnail else None,
             info=info)
+
+        if not self._already_have_thumbnail:
+            info['thumbnails'][idx].pop('filepath', None)
+
         return [], info

yt_dlp/postprocessor/ffmpeg.py

@@ -662,6 +662,10 @@ class FFmpegEmbedSubtitlePP(FFmpegPostProcessor):
         self.run_ffmpeg_multiple_files(input_files, temp_filename, opts)
         os.replace(temp_filename, filename)
 
+        if not self._already_have_subtitle:
+            for _, subtitle in subtitles.items():
+                subtitle.pop('filepath', None)
+
         files_to_delete = [] if self._already_have_subtitle else sub_filenames
         return files_to_delete, info
 
@@ -698,6 +702,7 @@ class FFmpegMetadataPP(FFmpegPostProcessor):
                 infojson_filename = info.get('infojson_filename')
                 options.extend(self._get_infojson_opts(info, infojson_filename))
                 if not infojson_filename:
+                    info.pop('infojson_filename', None)
                     files_to_delete.append(info.get('infojson_filename'))
             elif self._add_infojson is True:
                 self.to_screen('The info-json can only be attached to mkv/mka files')
@@ -1016,9 +1021,6 @@ class FFmpegSubtitlesConvertorPP(FFmpegPostProcessor):
                     'filepath': new_file,
                 }
 
-            info['__files_to_move'][new_file] = replace_extension(
-                info['__files_to_move'][sub['filepath']], new_ext)
-
         return sub_filenames, info
 
 
@@ -1083,16 +1085,15 @@ class FFmpegThumbnailsConvertorPP(FFmpegPostProcessor):
         return imghdr.what(path) == 'webp'
 
     def fixup_webp(self, info, idx=-1):
-        thumbnail_filename = info['thumbnails'][idx]['filepath']
+        thumbnail = info['thumbnails'][idx]
+        thumbnail_filename = thumbnail['filepath']
         _, thumbnail_ext = os.path.splitext(thumbnail_filename)
         if thumbnail_ext:
             if thumbnail_ext.lower() != '.webp' and imghdr.what(thumbnail_filename) == 'webp':
                 self.to_screen('Correcting thumbnail "%s" extension to webp' % thumbnail_filename)
                 webp_filename = replace_extension(thumbnail_filename, 'webp')
                 os.replace(thumbnail_filename, webp_filename)
-                info['thumbnails'][idx]['filepath'] = webp_filename
-                info['__files_to_move'][webp_filename] = replace_extension(
-                    info['__files_to_move'].pop(thumbnail_filename), 'webp')
+                thumbnail['filepath'] = webp_filename
 
     @staticmethod
     def _options(target_ext):
@@ -1130,8 +1131,6 @@ class FFmpegThumbnailsConvertorPP(FFmpegPostProcessor):
                 continue
             thumbnail_dict['filepath'] = self.convert_thumbnail(original_thumbnail, target_ext)
             files_to_delete.append(original_thumbnail)
-            info['__files_to_move'][thumbnail_dict['filepath']] = replace_extension(
-                info['__files_to_move'][original_thumbnail], target_ext)
 
         if not has_thumbnail:
             self.to_screen('There aren\'t any thumbnails to convert')

yt_dlp/postprocessor/movefilesafterdownload.py

@@ -1,16 +1,22 @@
 import os
+from pathlib import Path
 
 from .common import PostProcessor
 from ..compat import shutil
 from ..utils import (
     PostProcessingError,
-    decodeFilename,
-    encodeFilename,
     make_dir,
+    replace_extension
 )
 
 
 class MoveFilesAfterDownloadPP(PostProcessor):
+    # Map of the keys that contain moveable files and the 'type' of the file
+    # for generating the output filename
+    CHILD_KEYS = {
+        'thumbnails': 'thumbnail',
+        'requested_subtitles': 'subtitle'
+    }
 
     def __init__(self, downloader=None, downloaded=True):
         PostProcessor.__init__(self, downloader)
@@ -20,34 +26,79 @@ class MoveFilesAfterDownloadPP(PostProcessor):
     def pp_key(cls):
         return 'MoveFiles'
 
+    def move_file_and_write_to_info(self, info_dict, relevant_dict=None, output_file_type=None):
+        relevant_dict = relevant_dict or info_dict
+        if 'filepath' not in relevant_dict:
+            return
+
+        output_file_type = output_file_type or ''
+        current_filepath, final_filepath = self.determine_filepath(info_dict, relevant_dict, output_file_type)
+        move_result = self.move_file(info_dict, current_filepath, final_filepath)
+
+        if move_result:
+            relevant_dict['filepath'] = move_result
+        else:
+            del relevant_dict['filepath']
+
+    def determine_filepath(self, info_dict, relevant_dict, output_file_type):
+        current_filepath = relevant_dict['filepath']
+        prepared_filepath = self._downloader.prepare_filename(info_dict, output_file_type)
+
+        if (output_file_type == 'thumbnail' and info_dict['__multiple_thumbnails']) or output_file_type == 'subtitle':
+            desired_extension = ''.join(Path(current_filepath).suffixes[-2:])
+        else:
+            desired_extension = Path(current_filepath).suffix
+
+        return current_filepath, replace_extension(prepared_filepath, desired_extension)
+
+    def move_file(self, info_dict, current_filepath, final_filepath):
+        if not current_filepath or not final_filepath:
+            return
+
+        dl_parent_folder = os.path.split(info_dict['filepath'])[0]
+        finaldir = info_dict.get('__finaldir', os.path.abspath(dl_parent_folder))
+
+        if not os.path.isabs(current_filepath):
+            current_filepath = os.path.join(finaldir, current_filepath)
+
+        if not os.path.isabs(final_filepath):
+            final_filepath = os.path.join(finaldir, final_filepath)
+
+        if current_filepath == final_filepath:
+            return final_filepath
+
+        if not os.path.exists(current_filepath):
+            self.report_warning('File "%s" cannot be found' % current_filepath)
+            return
+
+        if os.path.exists(final_filepath):
+            if self.get_param('overwrites', True):
+                self.report_warning('Replacing existing file "%s"' % final_filepath)
+                os.remove(final_filepath)
+            else:
+                self.report_warning(
+                    'Cannot move file "%s" out of temporary directory since "%s" already exists. '
+                    % (current_filepath, final_filepath))
+                return
+
+        make_dir(final_filepath, PostProcessingError)
+        self.to_screen(f'Moving file "{current_filepath}" to "{final_filepath}"')
+        shutil.move(current_filepath, final_filepath)  # os.rename cannot move between volumes
+
+        return final_filepath
+
     def run(self, info):
-        dl_path, dl_name = os.path.split(encodeFilename(info['filepath']))
-        finaldir = info.get('__finaldir', dl_path)
-        finalpath = os.path.join(finaldir, dl_name)
-        if self._downloaded:
-            info['__files_to_move'][info['filepath']] = decodeFilename(finalpath)
+        # This represents the main media file (using the 'filepath' key)
+        self.move_file_and_write_to_info(info)
 
-        make_newfilename = lambda old: decodeFilename(os.path.join(finaldir, os.path.basename(encodeFilename(old))))
-        for oldfile, newfile in info['__files_to_move'].items():
-            if not newfile:
-                newfile = make_newfilename(oldfile)
-            if os.path.abspath(encodeFilename(oldfile)) == os.path.abspath(encodeFilename(newfile)):
+        for key, output_file_type in self.CHILD_KEYS.items():
+            if key not in info:
                 continue
-            if not os.path.exists(encodeFilename(oldfile)):
-                self.report_warning('File "%s" cannot be found' % oldfile)
-                continue
-            if os.path.exists(encodeFilename(newfile)):
-                if self.get_param('overwrites', True):
-                    self.report_warning('Replacing existing file "%s"' % newfile)
-                    os.remove(encodeFilename(newfile))
-                else:
-                    self.report_warning(
-                        'Cannot move file "%s" out of temporary directory since "%s" already exists. '
-                        % (oldfile, newfile))
-                    continue
-            make_dir(newfile, PostProcessingError)
-            self.to_screen(f'Moving file "{oldfile}" to "{newfile}"')
-            shutil.move(oldfile, newfile)  # os.rename cannot move between volumes
 
-        info['filepath'] = finalpath
+            if isinstance(info[key], list) or isinstance(info[key], dict):
+                iterable = info[key].values() if isinstance(info[key], dict) else info[key]
+
+                for file_dict in iterable:
+                    self.move_file_and_write_to_info(info, file_dict, output_file_type)
+
         return [], info

yt_dlp/utils/_utils.py

@@ -2092,7 +2092,9 @@ def prepend_extension(filename, ext, expected_real_ext=None):
 
 def replace_extension(filename, ext, expected_real_ext=None):
     name, real_ext = os.path.splitext(filename)
-    return '{}.{}'.format(
+    ext = ext if ext.startswith('.') else '.' + ext
+
+    return '{}{}'.format(
         name if not expected_real_ext or real_ext[1:] == expected_real_ext else filename,
         ext)