@linkdlab/funcnodes_nodebuilder

Interactive “node builder” UI for FuncNodes: write Python code in the browser, execute it in a Pyodide worker, and preview the resulting Node inside a FuncNodes React Flow canvas.

This package is meant for development/prototyping: it executes user-provided Python via exec(...) in Pyodide and extracts the last defined funcnodes_core.Node subclass.

Features

• Split view: FuncNodes canvas + Monaco Python editor
• Runs Python in Pyodide (dedicated or shared worker)
• Supports lightweight dependency installation via # requires ... directives
• Persists editor content in localStorage (optional)

Installation

This is a React/Vite library with peer dependencies. Install it together with its peers:

yarn add @linkdlab/funcnodes_nodebuilder \
  @linkdlab/funcnodes_pyodide_react_flow \
  @linkdlab/funcnodes_react_flow \
  pyodide react react-dom

Usage (React)

import NodeBuilder from "@linkdlab/funcnodes_nodebuilder";
import "@linkdlab/funcnodes_react_flow/dist/funcnodes_react_flow.css";

export function App() {
  return (
    <div style={{ height: "100vh" }}>
      <NodeBuilder id="nodebuilder" show_python_editor={true} />
    </div>
  );
}

# requires ...

At the top of your Python code you can add one or more lines like:

# requires numpy
# requires pandas, scipy

Each module spec is passed to micropip.install(...) before evaluation.

Usage (Browser entry)

The browser entry mounts into a DOM element and also exposes globals:

import "@linkdlab/funcnodes_nodebuilder/src/browser_index";

window.NodeBuilder("app", { show_python_editor: true });

Development

yarn install
yarn watch

Tests / typecheck

node --test tests/*.test.js
./node_modules/.bin/tsc --noEmit