Get Started#

  • Programming Language: Python 3

Prerequisites#

Clone the Repository#

Note

Ensure you have the Git LFS extension (for handling scene bundles) installed before cloning the repository.

Clone the repository:

git clone --filter=blob:none --sparse --branch main https://github.com/open-edge-platform/metro-ai-suite.git
cd metro-ai-suite
git sparse-checkout set smart-building-digital-twin
cd smart-building-digital-twin

Run setup#

From the project root, run:

./setup.sh

The setup script:

  • prompts for an admin password (SUPASS) and a database password (DATABASE_PASSWORD)

  • generates TLS certificates

  • starts all services

  • waits for the API

  • imports the included Showcase scene automatically

  • performs a best-effort telemetry check

If xpu-smi is already installed on the host, ./setup.sh also grants the needed host access for xpu-smi, starts the host GPU telemetry bridge, and verifies that the analytics service can read telemetry. If you install xpu-smi after the initial deployment, rerun ./setup.sh.

The analytics service learns the SideDoorEntry door-state baseline only after the first complete replay loop, preventing partial data from a mid-loop startup from affecting the baseline.

After Setup#

After setup completes, you can access these runtime endpoints:

  • Scenescape web UI: SCENESCAPE_UI_URL from .env (accept the self-signed certificate) This is the Scenescape management interface for the simulated building scene. You can use the interface to edit object classes, camera transforms, and regions. Its default is https://PUBLIC_HOSTNAME; the browser must accept the generated self-signed certificate. The setup output provides the admin credentials.

  • Analytics dashboard: DASHBOARD_URL from .env This is a separate custom FastAPI web UI served by scripts/dashboard.py. It displays live scene state, people, bag, and door counts, region occupancy, system telemetry, narrated events, alerts, and camera snapshots. Its default is http://PUBLIC_HOSTNAME:7000.

Next Steps#