refactor: improve Docker speech container audio configuration and user permissions

- Update Dockerfile to enhance audio setup and user management
- Modify setup-audio.sh to add robust PulseAudio socket and device checks
- Add proper user and directory permissions for audio and model directories
- Simplify container startup process and improve audio device detection

.gitignore

@@ -87,4 +87,6 @@ site/
 # Python
 __pycache__/
 *.py[cod]
-*$py.class
+*$py.class
+
+models/