feat: improve language detection using trakit 0.2.1

knowit/rules/general.py

@@ -1,6 +1,5 @@
 from logging import NullHandler, getLogger
 
-import babelfish
 from trakit.api import trakit
 
 from knowit.core import Rule
@@ -15,7 +14,9 @@ class GuessTitleRule(Rule):
     def execute(self, props, pv_props, context):
         """Language detection using name."""
         if 'name' in props:
-            guessed = trakit(props['name'])
+            language = props.get('language')
+            options = {'expected_language': language} if language else {}
+            guessed = trakit(props['name'], options)
             if guessed:
                 return guessed
 
@@ -30,13 +31,4 @@ def execute(self, props, pv_props, context):
 
         guess = pv_props['guessed']
         if 'language' in guess:
-            guessed: babelfish.Language = guess['language']
-            if 'language' not in props:
-                return guessed
-
-            lang: babelfish.Language = props['language']
-            if guessed.alpha3 != lang.alpha3 or str(lang).count('-') >= str(guessed).count('-'):
-                logger.debug('Discarding %s: Language %r and guessed %r', self.description, lang, guessed)
-                return lang
-
-            return guessed
+            return guess['language']

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "knowit"
-version = "0.5.0"
+version = "0.5.1"
 description = "Know better your media files"
 authors = [
     "Rato",
@@ -46,7 +46,7 @@ enzyme = "^0.4.1"
 pint = "^0.20.1"
 pymediainfo = "^6.0.1"
 pyyaml = "^6.0"
-trakit = "^0.1.2"
+trakit = "^0.2.1"
 
 [tool.poetry.group.dev.dependencies]
 pytest = "^7.2.0"