feat: Bump version to 0.2.5, enhance logging, normalize Cyrillic characters, and improve database info formatting

renamer/extractors/fileinfo_extractor.py

@@ -1,4 +1,13 @@
 from pathlib import Path
+import logging
+import os
+
+# Set up logging conditionally
+if os.getenv('FORMATTER_LOG', '0') == '1':
+    logging.basicConfig(filename='formatter.log', level=logging.INFO, 
+                        format='%(asctime)s - %(levelname)s - %(message)s')
+else:
+    logging.basicConfig(level=logging.CRITICAL)  # Disable logging
 
 
 class FileInfoExtractor:
@@ -10,6 +19,7 @@ class FileInfoExtractor:
         self._modification_time = file_path.stat().st_mtime
         self._file_name = file_path.name
         self._file_path = str(file_path)
+        logging.info(f"FileInfoExtractor: file_name={self._file_name!r}, file_path={self._file_path!r}")
 
     def extract_size(self) -> int:
         """Extract file size in bytes"""

renamer/extractors/filename_extractor.py

@@ -11,6 +11,18 @@ class FilenameExtractor:
     def __init__(self, file_path: Path):
         self.file_path = file_path
         self.file_name = file_path.name
+        self.file_name = self._normalize_cyrillic(self.file_name)
+
+    def _normalize_cyrillic(self, text: str) -> str:
+        """Normalize Cyrillic characters to English equivalents for parsing"""
+        replacements = {
+            'р': 'p',
+            'і': 'i',
+            # Add more as needed
+        }
+        for cyr, eng in replacements.items():
+            text = text.replace(cyr, eng)
+        return text
 
     def _get_frame_class_from_height(self, height: int) -> str | None:
         """Get frame class from video height using FRAME_CLASSES constant"""

renamer/extractors/mediainfo_extractor.py

@@ -73,10 +73,10 @@ class MediaInfoExtractor:
             return 'HDR'
         return None
 
-    def extract_audio_langs(self) -> str:
+    def extract_audio_langs(self) -> str | None:
         """Extract audio languages from media info"""
         if not self.audio_tracks:
-            return ''
+            return None
         langs = []
         for a in self.audio_tracks:
             lang_code = getattr(a, 'language', 'und').lower()