tangram-core 0.5.0

A framework for real-time air traffic management research

tangram_core

tangram_core is the foundation of the tangram platform. It provides the essential scaffolding for custom geospatial visualisation tools.

While often used for aviation data, tangram_core itself is domain-agnostic. It handles the infrastructure (displaying maps, managing state, handling connections) so plugins can focus on the domain logic (decoding ADS-B, processing maritime signals, simulating weather).

• Documentation: https://mode-s.org/tangram/
• Repository: https://github.com/open-aviation/tangram

Components

• Backend: A Python application (FastAPI) that manages the lifecycle of plugins and background services.
• Channel: A high-performance Rust service that bridges Redis pub/sub with WebSockets for real-time frontend updates.
• Frontend: A Vue.js + Deck.gl shell that dynamically loads widgets and layers from installed plugins.

Usage

CLI

This package is rarely used alone. It is typically installed alongside plugins:

# with uv
uv tool install --with tangram-jet1090 --with tangram-system tangram-core
# with pip
pip install tangram-core tangram-jet1090 tangram-system
# launch!
tangram serve --config /path/to/your/tangram.toml

Programmatic

You can also launch tangram from a Python script or a Jupyter notebook using [tangram_core.Runtime][].

This provides direct access to the backend state (e.g. Redis client, HTTP client), enabling you to spawn data or interact with services programmatically.

For scripts, use the async context manager to safely configure the server within the async with block and then explicitly yield control to the server loop until a signal (Ctrl+C) is received.

import tangram_core

async def main() -> None:
    async with tangram_core.Runtime() as runtime:
        await runtime.state.redis_client.set("my_key", "value")
        await runtime.wait()  # required to run the server indefinitely

if __name__ == "__main__":
    import asyncio

    asyncio.run(main())

For interactive environments (Jupyter) where execution spans multiple cells, you must start and stop the server manually:

# %%
import tangram_core

runtime = tangram_core.Runtime()
await runtime.start()
# %%
await runtime.state.redis_client.set("my_key", "value")
# %%
await runtime.stop()
# %%