nodeBook: CNL knowledge graphs in posts (concept maps, inference, Petri-net simulation)

Introducing a new component for drawing Knowledge Graphs

nodeBook renders Controlled Natural Language (CNL) written in ```nodeBook code fences as interactive knowledge graphs โ€” in posts, chat messages, and the composer preview. It works like diagram extensions such as graphviz, mermaid.

This component allows users to include: mindmaps; concept maps with features supporting inference and query; named edges and draggable nodes; a click-to-open node inspector with polymorphic states (morphs); inferred relations (transitive closure + Prolog inverse/symmetric rules) drawn dashed with proof tooltips; a containment view nesting class hierarchies; and Petri-net process notation with click-to-fire token simulation โ€” chemical reactions, biological pathways, and system models come alive in a post. ```nodeBook-schema fences let a topic define its own types. A composer toolbar entry inserts a starter graph. The graphs are rendered using cytoscapjs.

Install: Admin โ†’ Customize โ†’ Themes โ†’ Install โ†’ From a git repository โ†’ the repo URL above, as a component. No server plugin โ€” works on hosted Discourse. Setting: graph_height.

Built on the open-source (AGPL, like Discourse) npm packages @nodebook/dom / @nodebook/core โ€” the same engine behind nodeBook for Logseq and Obsidian. Live example on a production forum: Testing nodeBook component - LivingAcademy - ChatShaala ยท Guided tutorial: nodeBook which is a https://hedgedoc.org/ fork, where it is developed and tested.

1 Like