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/cdm/decrypt_labs_remote_cdm.py

@@ -10,8 +10,8 @@ from pywidevine.cdm import Cdm as WidevineCdm
 from pywidevine.device import DeviceTypes
 from requests import Session
 
-from unshackle.core.vaults import Vaults
 from unshackle.core import __version__
+from unshackle.core.vaults import Vaults
 
 
 class MockCertificateChain:
@@ -353,17 +353,19 @@ class DecryptLabsRemoteCDM:
         Generate a license challenge using Decrypt Labs API with intelligent caching.
 
         This method implements smart caching logic that:
-        1. First attempts to retrieve cached keys from the API
-        2. If required KIDs are set, compares cached keys against requirements
-        3. Only makes a license request if keys are missing
-        4. Returns empty challenge if all required keys are cached
+        1. First checks local vaults for required keys
+        2. Attempts to retrieve cached keys from the API
+        3. If required KIDs are set, compares available keys (vault + cached) against requirements
+        4. Only makes a license request if keys are missing
+        5. Returns empty challenge if all required keys are available
 
         The intelligent caching works as follows:
+        - Local vaults: Always checked first if available
         - For L1/L2 devices: Always prioritizes cached keys (API automatically optimizes)
         - For other devices: Uses cache retry logic based on session state
         - With required KIDs set: Only requests license for missing keys
         - Without required KIDs: Returns any available cached keys
-        - For PlayReady: Combines cached keys with license keys seamlessly
+        - For PlayReady: Combines vault, cached, and license keys seamlessly
 
         Args:
             session_id: Session identifier
@@ -372,7 +374,7 @@ class DecryptLabsRemoteCDM:
             privacy_mode: Whether to use privacy mode - for compatibility only
 
         Returns:
-            License challenge as bytes, or empty bytes if cached keys satisfy requirements
+            License challenge as bytes, or empty bytes if available keys satisfy requirements
 
         Raises:
             InvalidSession: If session ID is invalid
@@ -381,6 +383,7 @@ class DecryptLabsRemoteCDM:
         Note:
             Call set_required_kids() before this method for optimal caching behavior.
             L1/L2 devices automatically use cached keys when available per API design.
+            Local vault keys are always checked first when vaults are available.
         """
         _ = license_type, privacy_mode
 
@@ -393,6 +396,31 @@ class DecryptLabsRemoteCDM:
         init_data = self._get_init_data_from_pssh(pssh_or_wrm)
         already_tried_cache = session.get("tried_cache", False)
 
+        if self.vaults and self._required_kids:
+            vault_keys = []
+            for kid_str in self._required_kids:
+                try:
+                    clean_kid = kid_str.replace("-", "")
+                    if len(clean_kid) == 32:
+                        kid_uuid = UUID(hex=clean_kid)
+                    else:
+                        kid_uuid = UUID(hex=clean_kid.ljust(32, "0"))
+                    key, _ = self.vaults.get_key(kid_uuid)
+                    if key and key.count("0") != len(key):
+                        vault_keys.append({"kid": kid_str, "key": key, "type": "CONTENT"})
+                except (ValueError, TypeError):
+                    continue
+
+            if vault_keys:
+                vault_kids = set(k["kid"] for k in vault_keys)
+                required_kids = set(self._required_kids)
+
+                if required_kids.issubset(vault_kids):
+                    session["keys"] = vault_keys
+                    return b""
+                else:
+                    session["vault_keys"] = vault_keys
+
         if self.device_name in ["L1", "L2"]:
             get_cached_keys = True
         else:
@@ -404,7 +432,7 @@ class DecryptLabsRemoteCDM:
             "get_cached_keys_if_exists": get_cached_keys,
         }
 
-        if self.device_name in ["L1", "L2", "SL2", "SL3"] and self.service_name:
+        if self.service_name:
             request_data["service"] = self.service_name
 
         if session["service_certificate"]:
@@ -441,17 +469,22 @@ class DecryptLabsRemoteCDM:
             """
             cached_keys = data.get("cached_keys", [])
             parsed_keys = self._parse_cached_keys(cached_keys)
-            session["keys"] = parsed_keys
+
+            all_available_keys = list(parsed_keys)
+            if "vault_keys" in session:
+                all_available_keys.extend(session["vault_keys"])
+
+            session["keys"] = all_available_keys
             session["tried_cache"] = True
 
             if self._required_kids:
-                cached_kids = set()
-                for key in parsed_keys:
+                available_kids = set()
+                for key in all_available_keys:
                     if isinstance(key, dict) and "kid" in key:
-                        cached_kids.add(key["kid"].replace("-", "").lower())
+                        available_kids.add(key["kid"].replace("-", "").lower())
 
                 required_kids = set(self._required_kids)
-                missing_kids = required_kids - cached_kids
+                missing_kids = required_kids - available_kids
 
                 if missing_kids:
                     session["cached_keys"] = parsed_keys
@@ -585,15 +618,15 @@ class DecryptLabsRemoteCDM:
 
         license_keys = self._parse_keys_response(data)
 
-        if self.is_playready and "cached_keys" in session:
-            """
-            Combine cached keys with license keys for PlayReady content.
+        all_keys = []
 
-            This ensures we have both the cached keys (obtained earlier) and
-            any additional keys from the license response, without duplicates.
-            """
+        if "vault_keys" in session:
+            all_keys.extend(session["vault_keys"])
+
+        if "cached_keys" in session:
             cached_keys = session.get("cached_keys", [])
-            all_keys = list(cached_keys)
+            for cached_key in cached_keys:
+                all_keys.append(cached_key)
 
         for license_key in license_keys:
             already_exists = False
@@ -606,16 +639,16 @@ class DecryptLabsRemoteCDM:
                 license_kid = str(license_key.key_id).replace("-", "").lower()
 
             if license_kid:
-                    for cached_key in cached_keys:
-                        cached_kid = None
-                        if isinstance(cached_key, dict) and "kid" in cached_key:
-                            cached_kid = cached_key["kid"].replace("-", "").lower()
-                        elif hasattr(cached_key, "kid"):
-                            cached_kid = str(cached_key.kid).replace("-", "").lower()
-                        elif hasattr(cached_key, "key_id"):
-                            cached_kid = str(cached_key.key_id).replace("-", "").lower()
+                for existing_key in all_keys:
+                    existing_kid = None
+                    if isinstance(existing_key, dict) and "kid" in existing_key:
+                        existing_kid = existing_key["kid"].replace("-", "").lower()
+                    elif hasattr(existing_key, "kid"):
+                        existing_kid = str(existing_key.kid).replace("-", "").lower()
+                    elif hasattr(existing_key, "key_id"):
+                        existing_kid = str(existing_key.key_id).replace("-", "").lower()
 
-                        if cached_kid == license_kid:
+                    if existing_kid == license_kid:
                         already_exists = True
                         break
 
@@ -623,12 +656,23 @@ class DecryptLabsRemoteCDM:
                 all_keys.append(license_key)
 
         session["keys"] = all_keys
-            session["cached_keys"] = None
-        else:
-            session["keys"] = license_keys
+        session.pop("cached_keys", None)
+        session.pop("vault_keys", None)
 
         if self.vaults and session["keys"]:
-            key_dict = {UUID(hex=key["kid"]): key["key"] for key in session["keys"] if key["type"] == "CONTENT"}
+            key_dict = {}
+            for key in session["keys"]:
+                if key["type"] == "CONTENT":
+                    try:
+                        clean_kid = key["kid"].replace("-", "")
+                        if len(clean_kid) == 32:
+                            kid_uuid = UUID(hex=clean_kid)
+                        else:
+                            kid_uuid = UUID(hex=clean_kid.ljust(32, "0"))
+                        key_dict[kid_uuid] = key["key"]
+                    except (ValueError, TypeError):
+                        continue
+            if key_dict:
                 self.vaults.add_keys(key_dict)
 
     def get_keys(self, session_id: bytes, type_: Optional[str] = None) -> List[Key]:

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):

unshackle/vaults/MySQL.py

@@ -28,26 +28,33 @@ class MySQL(Vault):
             raise PermissionError(f"MySQL vault {self.slug} has no SELECT permission.")
 
     def get_key(self, kid: Union[UUID, str], service: str) -> Optional[str]:
-        if not self.has_table(service):
-            # no table, no key, simple
-            return None
-
         if isinstance(kid, UUID):
             kid = kid.hex
 
+        service_variants = [service]
+        if service != service.lower():
+            service_variants.append(service.lower())
+        if service != service.upper():
+            service_variants.append(service.upper())
+
         conn = self.conn_factory.get()
         cursor = conn.cursor()
 
         try:
+            for service_name in service_variants:
+                if not self.has_table(service_name):
+                    continue
+
                 cursor.execute(
                     # TODO: SQL injection risk
-                f"SELECT `id`, `key_` FROM `{service}` WHERE `kid`=%s AND `key_`!=%s",
+                    f"SELECT `id`, `key_` FROM `{service_name}` WHERE `kid`=%s AND `key_`!=%s",
                     (kid, "0" * 32),
                 )
                 cek = cursor.fetchone()
-            if not cek:
-                return None
+                if cek:
                     return cek["key_"]
+
+            return None
         finally:
             cursor.close()
 

unshackle/vaults/SQLite.py

@@ -19,22 +19,30 @@ class SQLite(Vault):
         self.conn_factory = ConnectionFactory(self.path)
 
     def get_key(self, kid: Union[UUID, str], service: str) -> Optional[str]:
-        if not self.has_table(service):
-            # no table, no key, simple
-            return None
-
         if isinstance(kid, UUID):
             kid = kid.hex
 
         conn = self.conn_factory.get()
         cursor = conn.cursor()
 
+        # Try both the original service name and lowercase version to handle case sensitivity issues
+        service_variants = [service]
+        if service != service.lower():
+            service_variants.append(service.lower())
+        if service != service.upper():
+            service_variants.append(service.upper())
+
         try:
-            cursor.execute(f"SELECT `id`, `key_` FROM `{service}` WHERE `kid`=? AND `key_`!=?", (kid, "0" * 32))
+            for service_name in service_variants:
+                if not self.has_table(service_name):
+                    continue
+
+                cursor.execute(f"SELECT `id`, `key_` FROM `{service_name}` WHERE `kid`=? AND `key_`!=?", (kid, "0" * 32))
                 cek = cursor.fetchone()
-            if not cek:
-                return None
+                if cek:
                     return cek[1]
+
+            return None
         finally:
             cursor.close()