Jupyter Hooks System (Python ↔ Frontend)

This chapter documents the Jupyter hooks system: how your notebook Python code can drive the Cellucid viewer (commands) and react to what a user does in the viewer (events).

Context (so the ecosystem is clear):

• Cellucid is the web app (the interactive viewer UI).

• cellucid-python is the helper package you use from Python/CLI to prepare data, run servers, embed the viewer, and integrate with notebooks.

• cellucid-annotation is a helper repo for community annotation workflows.

• cellucid-r prepares R data for the same web export contract; Python owns notebook embedding, servers, and hooks.

Where to start (pick one)

• “I want a minimal working loop (select → highlight)” → Quickstart: minimal round-trip (select → highlight)

• “Does this work in Colab / VSCode / JupyterHub?” → Supported environments matrix

• “I need the mental model and vocabulary” → Glossary and mental model

• “I need message + event schemas” → Reference (hooks, commands, schemas, endpoints)

• “Something is broken” → Troubleshooting (hooks)