How biomedical-signal-processing/sleepyland works
Project Type and Purpose
• This is a full-stack, Docker-orchestrated Python web application (“SLEEPYLAND”) for FAIR, local sleep staging analysis.
• It leverages a Flask-based GUI, a lightweight Python API (“manager-api”), data-download scripts, model-inference services, and Jupyter notebooks for tutorials and advanced analysis.
Key Directories & Services
• Top-level docker-compose.yml defines multiple containers:
• gui (Flask web UI)
• manager-api (internal REST API)
• nsrr-download (data ingestion service)
• usleepyland (model-inference CLI)
• notebook (Jupyter environment)
• wild-to-fancy (data harmonization CLI)
• Shared volumes:
• nsrr_file_path (input EDF/XML)
• output (predictions, metrics)
• gui/app/ contains Flask routes, service modules (download, prediction, metrics), utils (file handling, validation).
• manager-api/src/api.py exposes REST endpoints; utils.py holds shared logic.
• nsrr-download/ and usleepyland/ hold standalone scripts and environment configs.
• notebook/ hosts tutorial notebooks and resources.