Add option --concat-playlist

Closes #1855, related: #382

yt_dlp/YoutubeDL.py

@@ -1596,6 +1596,19 @@ class YoutubeDL(object):
     def _ensure_dir_exists(self, path):
         return make_dir(path, self.report_error)
 
+    @staticmethod
+    def _playlist_infodict(ie_result, **kwargs):
+        return {
+            **ie_result,
+            'playlist': ie_result.get('title') or ie_result.get('id'),
+            'playlist_id': ie_result.get('id'),
+            'playlist_title': ie_result.get('title'),
+            'playlist_uploader': ie_result.get('uploader'),
+            'playlist_uploader_id': ie_result.get('uploader_id'),
+            'playlist_index': 0,
+            **kwargs,
+        }
+
     def __process_playlist(self, ie_result, download):
         # We process each entry in the playlist
         playlist = ie_result.get('title') or ie_result.get('id')
@@ -1695,17 +1708,7 @@ class YoutubeDL(object):
 
         _infojson_written = False
         if not self.params.get('simulate') and self.params.get('allow_playlist_files', True):
-            ie_copy = {
-                'playlist': playlist,
-                'playlist_id': ie_result.get('id'),
-                'playlist_title': ie_result.get('title'),
-                'playlist_uploader': ie_result.get('uploader'),
-                'playlist_uploader_id': ie_result.get('uploader_id'),
-                'playlist_index': 0,
-                'n_entries': n_entries,
-            }
-            ie_copy.update(dict(ie_result))
-
+            ie_copy = self._playlist_infodict(ie_result, n_entries=n_entries)
             _infojson_written = self._write_info_json(
                 'playlist', ie_result, self.prepare_filename(ie_copy, 'pl_infojson'))
             if _infojson_written is None:

yt_dlp/__init__.py

@@ -591,6 +591,12 @@ def _real_main(argv=None):
     # XAttrMetadataPP should be run after post-processors that may change file contents
     if opts.xattrs:
         postprocessors.append({'key': 'XAttrMetadata'})
+    if opts.concat_playlist != 'never':
+        postprocessors.append({
+            'key': 'FFmpegConcat',
+            'only_multi_video': opts.concat_playlist != 'always',
+            'when': 'playlist',
+        })
     # Exec must be the last PP of each category
     if opts.exec_before_dl_cmd:
         opts.exec_cmd.setdefault('before_dl', opts.exec_before_dl_cmd)

yt_dlp/options.py

@@ -1397,6 +1397,16 @@ def create_parser():
         '--xattrs',
         action='store_true', dest='xattrs', default=False,
         help='Write metadata to the video file\'s xattrs (using dublin core and xdg standards)')
+    postproc.add_option(
+        '--concat-playlist',
+        metavar='POLICY', dest='concat_playlist', default='multi_video',
+        choices=('never', 'always', 'multi_video'),
+        help=(
+            'Concatenate videos in a playlist. One of "never" (default), "always", or '
+            '"multi_video" (only when the videos form a single show). '
+            'All the video files must have same codecs and number of streams to be concatable. '
+            'The "pl_video:" prefix can be used with "--paths" and "--output" to '
+            'set the output filename for the split files. See "OUTPUT TEMPLATE" for details'))
     postproc.add_option(
         '--fixup',
         metavar='POLICY', dest='fixup', default=None,

yt_dlp/postprocessor/__init__.py

@@ -7,6 +7,7 @@ from .embedthumbnail import EmbedThumbnailPP
 from .exec import ExecPP, ExecAfterDownloadPP
 from .ffmpeg import (
     FFmpegPostProcessor,
+    FFmpegConcatPP,
     FFmpegEmbedSubtitlePP,
     FFmpegExtractAudioPP,
     FFmpegFixupDuplicateMoovPP,

yt_dlp/postprocessor/ffmpeg.py

@@ -1123,3 +1123,48 @@ class FFmpegThumbnailsConvertorPP(FFmpegPostProcessor):
         if not has_thumbnail:
             self.to_screen('There aren\'t any thumbnails to convert')
         return files_to_delete, info
+
+
+class FFmpegConcatPP(FFmpegPostProcessor):
+    def __init__(self, downloader, only_multi_video=False):
+        self._only_multi_video = only_multi_video
+        super().__init__(downloader)
+
+    def concat_files(self, in_files, out_file):
+        if len(in_files) == 1:
+            os.replace(in_files[0], out_file)
+            return
+
+        codecs = [traverse_obj(self.get_metadata_object(file), ('streams', ..., 'codec_name')) for file in in_files]
+        if len(set(map(tuple, codecs))) > 1:
+            raise PostProcessingError(
+                'The files have different streams/codecs and cannot be concatenated. '
+                'Either select different formats or --recode-video them to a common format')
+        super().concat_files(in_files, out_file)
+
+    @PostProcessor._restrict_to(images=False)
+    def run(self, info):
+        if not info.get('entries') or self._only_multi_video and info['_type'] != 'multi_video':
+            return [], info
+        elif None in info['entries']:
+            raise PostProcessingError('Aborting concatenation because some downloads failed')
+        elif any(len(entry) > 1 for entry in traverse_obj(info, ('entries', ..., 'requested_downloads')) or []):
+            raise PostProcessingError('Concatenation is not supported when downloading multiple separate formats')
+
+        in_files = traverse_obj(info, ('entries', ..., 'requested_downloads', 0, 'filepath'))
+        if not in_files:
+            self.to_screen('There are no files to concatenate')
+            return [], info
+
+        ie_copy = self._downloader._playlist_infodict(info)
+        exts = [traverse_obj(entry, ('requested_downloads', 0, 'ext'), 'ext') for entry in info['entries']]
+        ie_copy['ext'] = exts[0] if len(set(exts)) == 1 else 'mkv'
+        out_file = self._downloader.prepare_filename(ie_copy, 'pl_video')
+
+        self.concat_files(in_files, out_file)
+
+        info['requested_downloads'] = [{
+            'filepath': out_file,
+            'ext': ie_copy['ext'],
+        }]
+        return in_files, info

yt_dlp/utils.py

@@ -4695,6 +4695,7 @@ OUTTMPL_TYPES = {
     'annotation': 'annotations.xml',
     'infojson': 'info.json',
     'link': None,
+    'pl_video': None,
     'pl_thumbnail': None,
     'pl_description': 'description',
     'pl_infojson': 'info.json',