feat: automatic audio language metadata for embedded audio tracks

- Add intelligent embedded audio language detection at mux stage
- Automatically set audio language metadata when no separate audio tracks exist
- Respect user flags (-V, --no-audio) to avoid unnecessary processing
- Smart video track selection based on title language with fallbacks
- Improved default track selection to prioritize title language matches
- Enhanced FFmpeg repackaging with audio stream metadata injection
- Works automatically for all services without service-specific code

unshackle/commands/dl.py

@@ -1147,8 +1147,9 @@ class dl:
                     with Live(Padding(progress, (0, 5, 1, 5)), console=console):
                         for task_id, task_tracks in multiplex_tasks:
                             progress.start_task(task_id)  # TODO: Needed?
+                            audio_expected = not video_only and not no_audio
                             muxed_path, return_code, errors = task_tracks.mux(
-                                str(title), progress=partial(progress.update, task_id=task_id), delete=False
+                                str(title), progress=partial(progress.update, task_id=task_id), delete=False, audio_expected=audio_expected, title_language=title.language
                             )
                             muxed_paths.append(muxed_path)
                             if return_code >= 2:

unshackle/core/tracks/track.py

@@ -420,7 +420,7 @@ class Track:
             for drm in self.drm:
                 if isinstance(drm, PlayReady):
                     return drm
-        elif hasattr(cdm, 'is_playready'):
+        elif hasattr(cdm, "is_playready"):
             if cdm.is_playready:
                 for drm in self.drm:
                     if isinstance(drm, PlayReady):
@@ -567,8 +567,7 @@ class Track:
         output_path = original_path.with_stem(f"{original_path.stem}_repack")
 
         def _ffmpeg(extra_args: list[str] = None):
-            subprocess.run(
+            args = [
-                [
                 binaries.FFMPEG,
                 "-hide_banner",
                 "-loglevel",
@@ -576,6 +575,24 @@ class Track:
                 "-i",
                 original_path,
                 *(extra_args or []),
+            ]
+
+            if hasattr(self, "data") and self.data.get("audio_language"):
+                audio_lang = self.data["audio_language"]
+                audio_name = self.data.get("audio_language_name", audio_lang)
+                args.extend(
+                    [
+                        "-metadata:s:a:0",
+                        f"language={audio_lang}",
+                        "-metadata:s:a:0",
+                        f"title={audio_name}",
+                        "-metadata:s:a:0",
+                        f"handler_name={audio_name}",
+                    ]
+                )
+
+            args.extend(
+                [
                     # Following are very important!
                     "-map_metadata",
                     "-1",  # don't transfer metadata to output file
@@ -584,7 +601,11 @@ class Track:
                     "-codec",
                     "copy",
                     str(output_path),
-                ],
+                ]
+            )
+
+            subprocess.run(
+                args,
                 check=True,
                 stdout=subprocess.PIPE,
                 stderr=subprocess.PIPE,

unshackle/core/tracks/tracks.py

@@ -305,7 +305,14 @@ class Tracks:
             )
         return selected
 
-    def mux(self, title: str, delete: bool = True, progress: Optional[partial] = None) -> tuple[Path, int, list[str]]:
+    def mux(
+        self,
+        title: str,
+        delete: bool = True,
+        progress: Optional[partial] = None,
+        audio_expected: bool = True,
+        title_language: Optional[Language] = None,
+    ) -> tuple[Path, int, list[str]]:
         """
         Multiplex all the Tracks into a Matroska Container file.
 
@@ -315,7 +322,28 @@ class Tracks:
             delete: Delete all track files after multiplexing.
             progress: Update a rich progress bar via `completed=...`. This must be the
                 progress object's update() func, pre-set with task id via functools.partial.
+            audio_expected: Whether audio is expected in the output. Used to determine
+                if embedded audio metadata should be added.
+            title_language: The title's intended language. Used to select the best video track
+                for audio metadata when multiple video tracks exist.
         """
+        if self.videos and not self.audio and audio_expected:
+            video_track = None
+            if title_language:
+                video_track = next((v for v in self.videos if v.language == title_language), None)
+                if not video_track:
+                    video_track = next((v for v in self.videos if v.is_original_lang), None)
+
+            video_track = video_track or self.videos[0]
+            if video_track.language.is_valid():
+                lang_code = str(video_track.language)
+                lang_name = video_track.language.display_name()
+
+                for video in self.videos:
+                    video.needs_repack = True
+                    video.data["audio_language"] = lang_code
+                    video.data["audio_language_name"] = lang_name
+
         if not binaries.MKVToolNix:
             raise RuntimeError("MKVToolNix (mkvmerge) is required for muxing but was not found")
 
@@ -332,12 +360,20 @@ class Tracks:
                 raise ValueError("Video Track must be downloaded before muxing...")
             events.emit(events.Types.TRACK_MULTIPLEX, track=vt)
 
+            is_default = False
+            if title_language:
+                is_default = vt.language == title_language
+                if not any(v.language == title_language for v in self.videos):
+                    is_default = vt.is_original_lang or i == 0
+            else:
+                is_default = i == 0
+
             # Prepare base arguments
             video_args = [
                 "--language",
                 f"0:{vt.language}",
                 "--default-track",
-                f"0:{i == 0}",
+                f"0:{is_default}",
                 "--original-flag",
                 f"0:{vt.is_original_lang}",
                 "--compression",
@@ -363,6 +399,18 @@ class Tracks:
                     ]
                 )
 
+            if hasattr(vt, "data") and vt.data.get("audio_language"):
+                audio_lang = vt.data["audio_language"]
+                audio_name = vt.data.get("audio_language_name", audio_lang)
+                video_args.extend(
+                    [
+                        "--language",
+                        f"1:{audio_lang}",
+                        "--track-name",
+                        f"1:{audio_name}",
+                    ]
+                )
+
             cl.extend(video_args + ["(", str(vt.path), ")"])
 
         for i, at in enumerate(self.audio):