feat: add storage and training modules for snore detection

- Implemented `storage.py` for managing metadata storage, including sample addition, retrieval, and review state management.
- Created `training.py` for training a local model using Random Forest, including functions for training and predicting samples.
- Developed a web interface in `app.js` for capturing audio samples, managing labels, and training the model.
- Added HTML structure in `index.html` for the SnoreStopper control room with sections for sample capture, overnight gathering, training, and status display.
- Styled the application with `styles.css` to enhance user experience and interface aesthetics.

requirements.txt

@@ -0,0 +1,10 @@
+fastapi>=0.115,<1.0
+uvicorn[standard]>=0.30,<1.0
+numpy>=2.0,<3.0
+sounddevice>=0.5,<1.0
+soundfile>=0.12,<1.0
+scipy>=1.13,<2.0
+matplotlib>=3.9,<4.0
+scikit-learn>=1.5,<2.0
+joblib>=1.4,<2.0
+python-multipart>=0.0.9,<1.0