Install

Install the Elodin Python SDK using pip:

pip install -U elodin

The SDK is only supported on Linux and macOS. Windows users can still use Elodin by running the simulation server in WSL. See the Windows section below for instructions.

Download the Elodin Client:

FilePlatformChecksum
elodin-aarch64-apple-darwin.tar.gzApple Silicon macOSsha256
elodin-x86_64-unknown-linux-gnu.tar.gzx64 Linuxsha256
elodin-x86_64-pc-windows-msvc.zipx64 Windowssha256

Start Simulating

  1. Download the three-body.py example code from our Github repo: elodin-sys/elodin.
  2. Launch the simulation using the elodin CLI.

    This command is only supported on Linux and macOS. Windows users can still use Elodin by running the simulation server in WSL. See the Windows section below for instructions.

    elodin editor three-body.py
    

Windows

To use Elodin on Windows, the simulation server must be run in Windows Subsystem for Linux (WSL). The Elodin Client itself can run natively on Windows. Follow these steps to get started:

  1. Install WSL by following the instructions here.
  2. Install the Elodin Python SDK in WSL.
    pip install -U elodin
    
  3. Run the simulation server in WSL.
    The --watch argument enables live-reload. The simulation will restart to pickup the changes whenever the file is updated.
    python three-body.py run --watch
    
  4. Run elodin.exe in Windows to launch the Elodin Client.

Perform Analysis

To analyze simulation data, use the Exec API to run the simulation for some number of ticks and collect the historical component data as a Polars DataFrame. The DataFrame can then be used to generate plots or perform other methods of data analysis.

Run the ball/plot.py example code to see this in action.

The ball/plot.py example depends on matplotlib. Install it using pip:

pip install -U matplotlib
python examples/ball/plot.py

Monte Carlo

Run Monte Carlo simulations to explore the state space.

  1. Create an account at https://app.elodin.systems to receive 60 free minutes of hosted simulation time (per month).
  2. Authorize elodin to access the Monte Carlo platform.
    elodin login
    
  3. Clone the Elodin Github repository locally to access the bouncing ball example code, which includes random sampling and asserts.
  4. Start a 300 sample Monte Carlo run with a maximum sim duration of 15s.
    Add --open to automatically open the dashboard url in the browser.
    elodin monte-carlo run --name ball examples/ball/main.py --max-duration 15 --samples 300
    

Next Steps

Try out the following tutorials to learn how to build simulations using Elodin: