Fixed embedding filepath issue for subs and infojson

yt_dlp/YoutubeDL.py

@@ -1808,7 +1808,7 @@ class YoutubeDL:
                     info_copy['id'] = ie.get_temp_id(ie_result['url'])
                 self.add_default_extra_info(info_copy, ie, ie_result['url'])
                 self.add_extra_info(info_copy, extra_info)
-                info_copy = self.pre_process(info_copy)
+                info_copy, _ = self.pre_process(info_copy)
                 self._fill_common_fields(info_copy, False)
                 self.__forced_printings(info_copy)
                 self._raise_pending_errors(info_copy)
@@ -1982,9 +1982,8 @@ class YoutubeDL:
                 'playlist', ie_result, self.prepare_filename(ie_copy, 'pl_infojson'))
             if _infojson_written is None:
                 return
-
-            description_file = self._write_description('playlist', ie_result, self.prepare_filename(ie_copy, 'pl_description'))
-            if description_file is None:
+            if self._write_description('playlist', ie_result,
+                                       self.prepare_filename(ie_copy, 'pl_description')) is None:
                 return
             # TODO: This should be passed to ThumbnailsConvertor if necessary
             self._write_thumbnails('playlist', ie_result, self.prepare_filename(ie_copy, 'pl_thumbnail'))
@@ -2885,13 +2884,13 @@ class YoutubeDL:
             # which can't be exported to json
             info_dict['formats'] = formats
 
-        info_dict = self.pre_process(info_dict)
+        info_dict, _ = self.pre_process(info_dict)
 
         if self._match_entry(info_dict, incomplete=self._format_fields) is not None:
             return info_dict
 
         self.post_extract(info_dict)
-        info_dict = self.pre_process(info_dict, 'after_filter')
+        info_dict, _ = self.pre_process(info_dict, 'after_filter')
 
         # The pre-processors may have modified the formats
         formats = self._get_formats(info_dict)
@@ -3202,7 +3201,7 @@ class YoutubeDL:
             info_dict.clear()
             info_dict.update(new_info)
 
-        new_info = self.pre_process(info_dict, 'video')
+        new_info, _ = self.pre_process(info_dict, 'video')
         replace_info_dict(new_info)
         self._num_downloads += 1
 
@@ -3229,8 +3228,8 @@ class YoutubeDL:
         if not self._ensure_dir_exists(encodeFilename(temp_filename)):
             return
 
-        description_file = self._write_description('video', info_dict, self.prepare_filename(info_dict, 'description'))
-        if description_file is None:
+        if self._write_description('video', info_dict,
+                                   self.prepare_filename(info_dict, 'description')) is None:
             return
 
         sub_files = self._write_subtitles(info_dict, temp_filename)
@@ -3314,7 +3313,7 @@ class YoutubeDL:
                for link_type, should_write in write_links.items()):
             return
 
-        new_info = self.pre_process(info_dict, 'before_dl')
+        new_info, _ = self.pre_process(info_dict, 'before_dl')
         replace_info_dict(new_info)
 
         if self.params.get('skip_download'):
@@ -3677,7 +3676,10 @@ class YoutubeDL:
             info = self.run_pp(pp, info)
         return info
 
-    def pre_process(self, ie_info, key='pre_process'):
+    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)
         try:
             info = self.run_all_pps(key, info)
@@ -3685,13 +3687,17 @@ class YoutubeDL:
             msg = f'Preprocessing: {err}'
             info.setdefault('__pending_error', msg)
             self.report_error(msg, is_error=False)
-        return info
+        return info, files_to_move
 
-    def post_process(self, filename, info):
+    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 = self.run_all_pps('post_process', info, additional_pps=info.get('__postprocessors'))
         info = self.run_pp(MoveFilesAfterDownloadPP(self), info)
+        info.pop('__multiple_thumbnails', None)
         return self.run_all_pps('after_move', info)
 
     def _make_archive_id(self, info_dict):
@@ -4233,27 +4239,27 @@ class YoutubeDL:
             self.report_error(f'Cannot write {label} metadata to JSON file {infofn}')
             return None
 
-    def _write_description(self, label, info_dict, filename):
+    def _write_description(self, label, ie_result, descfn):
         ''' Write description and returns True = written, False = skip, None = error '''
         if not self.params.get('writedescription'):
             return False
-        elif not filename:
+        elif not descfn:
             self.write_debug(f'Skipping writing {label} description')
             return False
-        elif not self._ensure_dir_exists(filename):
+        elif not self._ensure_dir_exists(descfn):
             return None
-        elif not self.params.get('overwrites', True) and os.path.exists(filename):
+        elif not self.params.get('overwrites', True) and os.path.exists(descfn):
             self.to_screen(f'[info] {label.title()} description is already present')
-        elif info_dict.get('description') is None:
+        elif ie_result.get('description') is None:
             self.to_screen(f'[info] There\'s no {label} description to write')
             return False
         else:
             try:
-                self.to_screen(f'[info] Writing {label} description to: {filename}')
-                with open(filename, 'w', encoding='utf-8') as descfile:
-                    descfile.write(info_dict['description'])
+                self.to_screen(f'[info] Writing {label} description to: {descfn}')
+                with open(encodeFilename(descfn), 'w', encoding='utf-8') as descfile:
+                    descfile.write(ie_result['description'])
             except OSError:
-                self.report_error(f'Cannot write {label} description file {filename}')
+                self.report_error(f'Cannot write {label} description file {descfn}')
                 return None
         return True
 
@@ -4282,7 +4288,7 @@ class YoutubeDL:
             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)
                 continue
 
             self.to_screen(f'[info] Writing video subtitles to: {sub_filename}')
@@ -4293,7 +4299,7 @@ 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)
                     continue
                 except OSError:
                     self.report_error(f'Cannot write video subtitles file {sub_filename}')
@@ -4304,7 +4310,7 @@ 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)
             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'
@@ -4324,6 +4330,7 @@ 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
@@ -4345,7 +4352,7 @@ 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)
             else:
                 self.to_screen(f'[info] Downloading {thumb_display_id} ...')
                 try:
@@ -4353,7 +4360,7 @@ 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)
                     t['filepath'] = thumb_filename
                 except network_exceptions as err:
                     if isinstance(err, HTTPError) and err.status == 404:

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')

yt_dlp/postprocessor/movefilesafterdownload.py

@@ -11,7 +11,6 @@ from ..utils import (
 
 
 class MoveFilesAfterDownloadPP(PostProcessor):
-    TOP_LEVEL_KEYS = ['filepath']
     # Map of the keys that contain moveable files and the 'type' of the file
     # for generating the output filename
     CHILD_KEYS = {
@@ -33,12 +32,7 @@ class MoveFilesAfterDownloadPP(PostProcessor):
             return
 
         output_file_type = output_file_type or ''
-        current_filepath = relevant_dict['filepath']
-        # This approach is needed to preserved indexed thumbnail paths from `--write-all-thumbnails`
-        # and also to support user-defined extensions (eg: `%(title)s.temp.%(ext)s`)
-        extension = ''.join(Path(current_filepath).suffixes)
-        name_format = self._downloader.prepare_filename(info_dict, output_file_type)
-        final_filepath = replace_extension(name_format, extension)
+        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:
@@ -46,6 +40,17 @@ class MoveFilesAfterDownloadPP(PostProcessor):
         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