Merge branch 'main' into feature/add-rest-api

unshackle/core/__main__.py

@@ -1,9 +1,3 @@
-import warnings
-
-# Suppress SyntaxWarning from unmaintained tinycss package (dependency of subby)
-# Must be set before any imports that might trigger tinycss loading
-warnings.filterwarnings("ignore", category=SyntaxWarning, module="tinycss")
-
 import atexit
 import logging
 from pathlib import Path

unshackle/core/constants.py

@@ -6,6 +6,7 @@ DOWNLOAD_LICENCE_ONLY = Event()
 
 DRM_SORT_MAP = ["ClearKey", "Widevine"]
 LANGUAGE_MAX_DISTANCE = 5  # this is max to be considered "same", e.g., en, en-US, en-AU
+LANGUAGE_EXACT_DISTANCE = 0  # exact match only, no variants
 VIDEO_CODEC_MAP = {"AVC": "H.264", "HEVC": "H.265"}
 DYNAMIC_RANGE_MAP = {"HDR10": "HDR", "HDR10+": "HDR10P", "Dolby Vision": "DV", "HDR10 / HDR10+": "HDR10P", "HDR10 / HDR10": "HDR"}
 AUDIO_CODEC_MAP = {"E-AC-3": "DDP", "AC-3": "DD"}

unshackle/core/tracks/tracks.py

@@ -14,7 +14,7 @@ from rich.tree import Tree
 from unshackle.core import binaries
 from unshackle.core.config import config
 from unshackle.core.console import console
-from unshackle.core.constants import LANGUAGE_MAX_DISTANCE, AnyTrack, TrackT
+from unshackle.core.constants import LANGUAGE_EXACT_DISTANCE, LANGUAGE_MAX_DISTANCE, AnyTrack, TrackT
 from unshackle.core.events import events
 from unshackle.core.tracks.attachment import Attachment
 from unshackle.core.tracks.audio import Audio
@@ -294,11 +294,14 @@ class Tracks:
         self.videos = selected
 
     @staticmethod
-    def by_language(tracks: list[TrackT], languages: list[str], per_language: int = 0) -> list[TrackT]:
+    def by_language(
+        tracks: list[TrackT], languages: list[str], per_language: int = 0, exact_match: bool = False
+    ) -> list[TrackT]:
+        distance = LANGUAGE_EXACT_DISTANCE if exact_match else LANGUAGE_MAX_DISTANCE
         selected = []
         for language in languages:
             selected.extend(
-                [x for x in tracks if closest_supported_match(x.language, [language], LANGUAGE_MAX_DISTANCE)][
+                [x for x in tracks if closest_supported_match(str(x.language), [language], distance)][
                     : per_language or None
                 ]
             )

unshackle/core/utilities.py

@@ -24,7 +24,7 @@ from unidecode import unidecode
 
 from unshackle.core.cacher import Cacher
 from unshackle.core.config import config
-from unshackle.core.constants import LANGUAGE_MAX_DISTANCE
+from unshackle.core.constants import LANGUAGE_EXACT_DISTANCE, LANGUAGE_MAX_DISTANCE
 
 
 def rotate_log_file(log_path: Path, keep: int = 20) -> Path:
@@ -114,6 +114,14 @@ def is_close_match(language: Union[str, Language], languages: Sequence[Union[str
     return closest_match(language, list(map(str, languages)))[1] <= LANGUAGE_MAX_DISTANCE
 
 
+def is_exact_match(language: Union[str, Language], languages: Sequence[Union[str, Language, None]]) -> bool:
+    """Check if a language is an exact match to any of the provided languages."""
+    languages = [x for x in languages if x]
+    if not languages:
+        return False
+    return closest_match(language, list(map(str, languages)))[1] <= LANGUAGE_EXACT_DISTANCE
+
+
 def get_boxes(data: bytes, box_type: bytes, as_bytes: bool = False) -> Box:
     """
     Scan a byte array for a wanted MP4/ISOBMFF box, then parse and yield each find.

unshackle/core/utils/tags.py

@@ -294,7 +294,7 @@ def _apply_tags(path: Path, tags: dict[str, str]) -> None:
     for name, value in tags.items():
         xml_lines.append(f"    <Simple><Name>{escape(name)}</Name><String>{escape(value)}</String></Simple>")
     xml_lines.extend(["  </Tag>", "</Tags>"])
-    with tempfile.NamedTemporaryFile("w", suffix=".xml", delete=False) as f:
+    with tempfile.NamedTemporaryFile("w", suffix=".xml", delete=False, encoding="utf-8") as f:
         f.write("\n".join(xml_lines))
         tmp_path = Path(f.name)
     try: