NodeKit: a Python library for behavioral tasks

NodeKit is a Python library for writing behavioral tasks. Any task written using NodeKit can be run in the browser (for people) or in Python (for models).

Why use NodeKit?

• No HTML/JavaScript/CSS. Instead, write tasks with Python.
• Humans and models do the same task. One NodeKit task can feed model simulations and web browser experiments.
• A task is one JSON document. NodeKit tasks are fully described by a single .json file, with no dependencies on external libraries or plugins.

Why not use NodeKit?

• No HTML/JavaScript/CSS. Arbitrary logic and visual customizability are not accessible in NodeKit.
• Fixed layout philosophy. NodeKit tasks are intended to be pixel-for-pixel reproducible, and do not react to window resizing or device type.
• No audio support. NodeKit focuses on purely visual tasks.

Project roadmap

NodeKit is a research tool under active development. Its releases on PyPi are versioned, but do not yet follow semantic versioning rules. Semantic versioning will be introduced with NodeKit's v1 release.

Current goals include:

• Refining and expanding the documentation site
• Improving determinism in the web browser
• Enriching the simulator with an end-to-end headless renderer