[cleanup] Refactor ffmpeg convertors

yt_dlp/__init__.py

@@ -31,7 +31,6 @@
     preferredencoding,
     read_batch_urls,
     RejectedVideoReached,
-    REMUX_EXTENSIONS,
     render_table,
     SameFileError,
     setproctitle,
@@ -45,6 +44,13 @@
 from .extractor import gen_extractors, list_extractors
 from .extractor.common import InfoExtractor
 from .extractor.adobepass import MSO_INFO
+from .postprocessor.ffmpeg import (
+    FFmpegExtractAudioPP,
+    FFmpegSubtitlesConvertorPP,
+    FFmpegThumbnailsConvertorPP,
+    FFmpegVideoConvertorPP,
+    FFmpegVideoRemuxerPP,
+)
 from .postprocessor.metadatafromfield import MetadataFromFieldPP
 from .YoutubeDL import YoutubeDL
 
@@ -209,25 +215,25 @@ def parse_retries(retries, name=''):
     if opts.playlistend not in (-1, None) and opts.playlistend < opts.playliststart:
         raise ValueError('Playlist end must be greater than playlist start')
     if opts.extractaudio:
-        if opts.audioformat not in ['best', 'aac', 'flac', 'mp3', 'm4a', 'opus', 'vorbis', 'wav']:
+        if opts.audioformat not in ['best'] + list(FFmpegExtractAudioPP.SUPPORTED_EXTS):
             parser.error('invalid audio format specified')
     if opts.audioquality:
         opts.audioquality = opts.audioquality.strip('k').strip('K')
         if not opts.audioquality.isdigit():
             parser.error('invalid audio quality specified')
     if opts.recodevideo is not None:
-        if opts.recodevideo not in REMUX_EXTENSIONS:
+        opts.recodevideo = opts.recodevideo.replace(' ', '')
-            parser.error('invalid video recode format specified')
+        if not re.match(FFmpegVideoConvertorPP.FORMAT_RE, opts.recodevideo):
+            parser.error('invalid video remux format specified')
     if opts.remuxvideo is not None:
         opts.remuxvideo = opts.remuxvideo.replace(' ', '')
-        remux_regex = r'{0}(?:/{0})*$'.format(r'(?:\w+>)?(?:%s)' % '|'.join(REMUX_EXTENSIONS))
+        if not re.match(FFmpegVideoRemuxerPP.FORMAT_RE, opts.remuxvideo):
-        if not re.match(remux_regex, opts.remuxvideo):
             parser.error('invalid video remux format specified')
     if opts.convertsubtitles is not None:
-        if opts.convertsubtitles not in ('srt', 'vtt', 'ass', 'lrc'):
+        if opts.convertsubtitles not in FFmpegSubtitlesConvertorPP.SUPPORTED_EXTS:
             parser.error('invalid subtitle format specified')
     if opts.convertthumbnails is not None:
-        if opts.convertthumbnails not in ('jpg', 'png'):
+        if opts.convertthumbnails not in FFmpegThumbnailsConvertorPP.SUPPORTED_EXTS:
             parser.error('invalid thumbnail format specified')
 
     if opts.date is not None:
@@ -480,10 +486,10 @@ def report_args_compat(arg, name):
         opts.postprocessor_args['default'] = opts.postprocessor_args['default-compat']
 
     final_ext = (
-        opts.recodevideo
+        opts.recodevideo if opts.recodevideo in FFmpegVideoConvertorPP.SUPPORTED_EXTS
-        or (opts.remuxvideo in REMUX_EXTENSIONS) and opts.remuxvideo
+        else opts.remuxvideo if opts.remuxvideo in FFmpegVideoRemuxerPP.SUPPORTED_EXTS
-        or (opts.extractaudio and opts.audioformat != 'best') and opts.audioformat
+        else opts.audioformat if (opts.extractaudio and opts.audioformat != 'best')
-        or None)
+        else None)
 
     match_filter = (
         None if opts.match_filter is None

yt_dlp/options.py

@@ -5,7 +5,6 @@
 import re
 import sys
 
-from .downloader.external import list_external_downloaders
 from .compat import (
     compat_expanduser,
     compat_get_terminal_size,
@@ -18,11 +17,18 @@
     get_executable_path,
     OUTTMPL_TYPES,
     preferredencoding,
-    REMUX_EXTENSIONS,
     write_string,
 )
 from .version import __version__
 
+from .downloader.external import list_external_downloaders
+from .postprocessor.ffmpeg import (
+    FFmpegExtractAudioPP,
+    FFmpegSubtitlesConvertorPP,
+    FFmpegThumbnailsConvertorPP,
+    FFmpegVideoRemuxerPP,
+)
+
 
 def _hide_login_info(opts):
     PRIVATE_OPTS = set(['-p', '--password', '-u', '--username', '--video-password', '--ap-password', '--ap-username'])
@@ -1123,7 +1129,9 @@ def _dict_from_options_callback(
         help='Convert video files to audio-only files (requires ffmpeg and ffprobe)')
     postproc.add_option(
         '--audio-format', metavar='FORMAT', dest='audioformat', default='best',
-        help='Specify audio format: "best", "aac", "flac", "mp3", "m4a", "opus", "vorbis", or "wav"; "%default" by default; No effect without -x')
+        help=(
+            'Specify audio format to convert the audio to when -x is used. Currently supported formats are: '
+            'best (default) or one of %s' % '|'.join(FFmpegExtractAudioPP.SUPPORTED_EXTS)))
     postproc.add_option(
         '--audio-quality', metavar='QUALITY',
         dest='audioquality', default='5',
@@ -1134,15 +1142,14 @@ def _dict_from_options_callback(
         help=(
             'Remux the video into another container if necessary (currently supported: %s). '
             'If target container does not support the video/audio codec, remuxing will fail. '
-            'You can specify multiple rules; eg. "aac>m4a/mov>mp4/mkv" will remux aac to m4a, mov to mp4 '
+            'You can specify multiple rules; Eg. "aac>m4a/mov>mp4/mkv" will remux aac to m4a, mov to mp4 '
-            'and anything else to mkv.' % '|'.join(REMUX_EXTENSIONS)))
+            'and anything else to mkv.' % '|'.join(FFmpegVideoRemuxerPP.SUPPORTED_EXTS)))
     postproc.add_option(
         '--recode-video',
         metavar='FORMAT', dest='recodevideo', default=None,
         help=(
             'Re-encode the video into another format if re-encoding is necessary. '
-            'You can specify multiple rules similar to --remux-video. '
+            'The syntax and supported formats are the same as --remux-video'))
-            'The supported formats are also the same as --remux-video'))
     postproc.add_option(
         '--postprocessor-args', '--ppa',
         metavar='NAME:ARGS', dest='postprocessor_args', default={}, type='str',
@@ -1250,11 +1257,15 @@ def _dict_from_options_callback(
     postproc.add_option(
         '--convert-subs', '--convert-sub', '--convert-subtitles',
         metavar='FORMAT', dest='convertsubtitles', default=None,
-        help='Convert the subtitles to another format (currently supported: srt|ass|vtt|lrc) (Alias: --convert-subtitles)')
+        help=(
+            'Convert the subtitles to another format (currently supported: %s) '
+            '(Alias: --convert-subtitles)' % '|'.join(FFmpegSubtitlesConvertorPP.SUPPORTED_EXTS)))
     postproc.add_option(
         '--convert-thumbnails',
         metavar='FORMAT', dest='convertthumbnails', default=None,
-        help='Convert the thumbnails to another format (currently supported: jpg, png)')
+        help=(
+            'Convert the thumbnails to another format '
+            '(currently supported: %s) ' % '|'.join(FFmpegThumbnailsConvertorPP.SUPPORTED_EXTS)))
     postproc.add_option(
         '--split-chapters', '--split-tracks',
         dest='split_chapters', action='store_true', default=False,

yt_dlp/postprocessor/ffmpeg.py

@@ -290,13 +290,12 @@ def _ffmpeg_filename_argument(self, fn):
 
 
 class FFmpegExtractAudioPP(FFmpegPostProcessor):
-    COMMON_AUDIO_EXTENSIONS = ('wav', 'flac', 'm4a', 'aiff', 'mp3', 'ogg', 'mka', 'opus', 'wma')
+    COMMON_AUDIO_EXTS = ('wav', 'flac', 'm4a', 'aiff', 'mp3', 'ogg', 'mka', 'opus', 'wma')
+    SUPPORTED_EXTS = ('best', 'aac', 'flac', 'mp3', 'm4a', 'opus', 'vorbis', 'wav')
 
     def __init__(self, downloader=None, preferredcodec=None, preferredquality=None, nopostoverwrites=False):
         FFmpegPostProcessor.__init__(self, downloader)
-        if preferredcodec is None:
+        self._preferredcodec = preferredcodec or 'best'
-            preferredcodec = 'best'
-        self._preferredcodec = preferredcodec
         self._preferredquality = preferredquality
         self._nopostoverwrites = nopostoverwrites
 
@@ -315,7 +314,7 @@ def run(self, information):
         path = information['filepath']
         orig_ext = information['ext']
 
-        if self._preferredcodec == 'best' and orig_ext in self.COMMON_AUDIO_EXTENSIONS:
+        if self._preferredcodec == 'best' and orig_ext in self.COMMON_AUDIO_EXTS:
             self.to_screen('Skipping audio extraction since the file is already in a common audio format')
             return [], information
 
@@ -400,6 +399,8 @@ def run(self, information):
 
 
 class FFmpegVideoConvertorPP(FFmpegPostProcessor):
+    SUPPORTED_EXTS = ('mp4', 'mkv', 'flv', 'webm', 'mov', 'avi', 'mp3', 'mka', 'm4a', 'ogg', 'opus')
+    FORMAT_RE = re.compile(r'{0}(?:/{0})*$'.format(r'(?:\w+>)?(?:%s)' % '|'.join(SUPPORTED_EXTS)))
     _action = 'converting'
 
     def __init__(self, downloader=None, preferedformat=None):
@@ -419,14 +420,14 @@ def _options(target_ext):
         return []
 
     def run(self, information):
-        path = information['filepath']
+        path, source_ext = information['filepath'], information['ext'].lower()
-        target_ext = self._target_ext(information['ext'].lower())
+        target_ext = self._target_ext(source_ext)
         _skip_msg = (
             'could not find a mapping for %s' if not target_ext
             else 'already is in target format %s' if source_ext == target_ext
             else None)
         if _skip_msg:
-            self.to_screen('Not %s media file %s; %s' % (self._action, path, _skip_msg % source_ext))
+            self.to_screen('Not %s media file "%s"; %s' % (self._action, path, _skip_msg % source_ext))
             return [], information
 
         prefix, sep, oldext = path.rpartition('.')
@@ -708,6 +709,8 @@ def run(self, info):
 
 
 class FFmpegSubtitlesConvertorPP(FFmpegPostProcessor):
+    SUPPORTED_EXTS = ('srt', 'vtt', 'ass', 'lrc')
+
     def __init__(self, downloader=None, format=None):
         super(FFmpegSubtitlesConvertorPP, self).__init__(downloader)
         self.format = format
@@ -816,6 +819,8 @@ def run(self, info):
 
 
 class FFmpegThumbnailsConvertorPP(FFmpegPostProcessor):
+    SUPPORTED_EXTS = ('jpg', 'png')
+
     def __init__(self, downloader=None, format=None):
         super(FFmpegThumbnailsConvertorPP, self).__init__(downloader)
         self.format = format
@@ -841,31 +846,29 @@ def fixup_webp(self, info, idx=-1):
                 info['__files_to_move'][webp_filename] = replace_extension(
                     info['__files_to_move'].pop(thumbnail_filename), 'webp')
 
-    def convert_thumbnail(self, thumbnail_filename, ext):
+    @staticmethod
-        if ext == 'jpg':
+    def _options(target_ext):
-            format_name = 'JPEG'
+        if target_ext == 'jpg':
-            opts = ['-bsf:v', 'mjpeg2jpeg']
+            return ['-bsf:v', 'mjpeg2jpeg']
-        elif ext == 'png':
+        return []
-            format_name = 'PNG'
+
-            opts = []
+    def convert_thumbnail(self, thumbnail_filename, target_ext):
-        else:
-            raise FFmpegPostProcessorError('Only conversion to either jpg or png is currently supported')
         # NB: % is supposed to be escaped with %% but this does not work
         # for input files so working around with standard substitution
         escaped_thumbnail_filename = thumbnail_filename.replace('%', '#')
         os.rename(encodeFilename(thumbnail_filename), encodeFilename(escaped_thumbnail_filename))
-        escaped_thumbnail_conv_filename = replace_extension(escaped_thumbnail_filename, ext)
+        escaped_thumbnail_conv_filename = replace_extension(escaped_thumbnail_filename, target_ext)
-        self.to_screen('Converting thumbnail "%s" to %s' % (escaped_thumbnail_filename, format_name))
+
-        self.run_ffmpeg(escaped_thumbnail_filename, escaped_thumbnail_conv_filename, opts)
+        self.to_screen('Converting thumbnail "%s" to %s' % (escaped_thumbnail_filename, target_ext))
-        thumbnail_conv_filename = replace_extension(thumbnail_filename, ext)
+        self.run_ffmpeg(escaped_thumbnail_filename, escaped_thumbnail_conv_filename, self._options(target_ext))
+
         # Rename back to unescaped
+        thumbnail_conv_filename = replace_extension(thumbnail_filename, target_ext)
         os.rename(encodeFilename(escaped_thumbnail_filename), encodeFilename(thumbnail_filename))
         os.rename(encodeFilename(escaped_thumbnail_conv_filename), encodeFilename(thumbnail_conv_filename))
         return thumbnail_conv_filename
 
     def run(self, info):
-        if self.format not in ('jpg', 'png'):
-            raise FFmpegPostProcessorError('Only conversion to either jpg or png is currently supported')
         files_to_delete = []
         has_thumbnail = False
 

yt_dlp/utils.py

@@ -1716,8 +1716,6 @@ def random_user_agent():
     'wav',
     'f4f', 'f4m', 'm3u8', 'smil')
 
-REMUX_EXTENSIONS = ('mp4', 'mkv', 'flv', 'webm', 'mov', 'avi', 'mp3', 'mka', 'm4a', 'ogg', 'opus')
-
 # needed for sanitizing filenames in restricted mode
 ACCENT_CHARS = dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
                         itertools.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],