Configure graph-app-kit and add integrations

Most settings can be configured by creating a custom Docker environment file src/docker/.env (see src/envs/*.env for options). You can also edit docker-compose.yml and /etc/docker/daemon.json, but we recommend sticking to .env.

Integration settings that deal with external systems such as TLS, accounts, and notebooks require having launched them. If you are not integrating into existing ones, see the initial setup section for how to quicklaunch a Graphistry server.

Monitoring

By default, graph-app-kit logging uses the Docker json file driver:

• Inspect recent activities: cd src/docker and then sudo docker-compose logs -f -t --tail=100

• Setup alternative logging drivers

Core

• Streamlit: URL base path: BASE_PATH=dashboard/ and BASE_URL=http://localhost/dashboard/
• Graphistry: None - set GRAPHISTRY_USERNAME=usr + GRAPHISTRY_PASSWORD=pwd (see src/envs/graphistry.env for more, like GRAPHISTRY_SERVER if using a private Graphistry server)
• Log level: LOG_LEVEL=ERROR (for Python's logging)

Databases

• Amazon Neptune guide for TinkerPop/Gremlin integration

TLS with Cadddy

• Auth: See Caddy sample reverse proxy example for an authentication check against an account system, including the one shipping with your Graphistry server (requires sudo docker-compose restart caddy in your Graphistry server upon editing /var/graphistry/data/config/Caddyfile)

Public+Private views

• To simulatenously run 1 public and 1 private instance, create two graph-app-kit clones public_dash and private_dash, and for src/docker/.env, set:
  • COMPOSE_PROJECT_NAME=streamlit-pub and COMPOSE_PROJECT_NAME=streamlit-priv
  • Override default ST_PUBLIC_PORT=8501 with two distinct values
• See Caddy sample for configuring URI routes, including covering the private instance with your Graphistry account system (JWT auth URL)