Scaled Sortation

Scaled Sortation combines a browser-based fluid simulation prototype with a Python control stack for Gymnasium experiments.

Repository Layout

• 3dfluid/ contains the WebGL prototype for browser-based simulation and speaker control.
• scaled_sortation/ contains the Python package, desktop demo, shader assets, and Gymnasium environment.
• tests/ contains smoke tests for the environment and a legacy-name audit.

Install

python -m pip install -e .[dev]

Run The Desktop Demo

python -m scaled_sortation.simulation_demo

Use The Gymnasium Environment

import gymnasium as gym
import scaled_sortation

env = gym.make("ScaledSortation-v0", launch_demo=False)
observation, info = env.reset()
observation, reward, terminated, truncated, info = env.step([1.0, 0.5])
env.close()

The action and observation vectors are [amplitude, frequency].

Set launch_demo=True if you want the environment to mirror its commands into the desktop demo window.

Browser Prototype

Serve the repository over HTTP before opening the browser prototype. Modern browsers block the shader XMLHttpRequest calls when the page is loaded from file://.

python -m http.server 8000

Then open http://localhost:8000/3dfluid/ in a WebGL-capable browser.

License

This repository is publicly visible but not open source. Noncommercial use is allowed under LICENSE.md. Commercial use is forbidden without a separate written license. AI or machine-learning training use, including model training, fine-tuning, benchmarking, and dataset creation from repository materials, is forbidden without a separate written license.