react-jupyter-notebook

A fork of react-jupyter-notebook that allows for the rendering and execution of Jupyter notebook files (.ipynb).

Demo: desousak.github.io/react-jupyter-notebook

Install

npm install --save https://github.com/Desousak/react-jupyter-notebook.git

Features

• Nearly identical looking to original JupyterLab interface.
• Can render codes, images, outputs, and markdown (with table & equation support).
• Enable resizing the height of the scrolled output.
• Execution of code with user-defined messaging.
  • Features support for Pyolite - allowing for execution of Python code through JS (see notes).
• (SOON!) Can change the alignment of the media outputs.
• (SOON!) Customisable code block styling.

Usage

// If you want code execution capabilities - you must provide your own connection code
// For example:
import { KernelMessenger } from 'react-jupyter-notebook';

class ExampleMessenger extends KernelMessenger {
  get ready() {
    // Returns a promise which resolves when the kernel loads in initially
    return Promise.resolve();
  }

  get kernelInfo() {
    // Return a promise containing info about the kernel
    return Promise.resolve({
      status: 'ok',
      protocol_version: 'X.Y.Z',
      implementation: '',
      implementation_version: 'X.Y.Z',
      language_info: {
        name: '',
        version: 'X.Y.Z',
        mimetype: '',
        file_extension: '',
        pygments_lexer: '',
        codemirror_mode: '',
        nbconvert_exporter: '',
      },
      banner: '',
      debugger: false,
      help_links: [{ text: '', url: '' }],
    });
  }

  get connected() {
    // Returns true if connected, false otherwise
    return false;
  }

  set connected(bool) {
    // Allows manually updating connected status if needed
    this.#connected = bool;
  }
  
  deconstructor() {
    // Will be run when the class is about to be removed
    // Can be used to disconnect from servers gracefully
    // Returns a promise, which will resolve when ready to be removed
    return Promise.resolve();
  }

  connectToKernel() {
    // this.updConnState(bool) is a callback that should be run on a connection change
    // (connect = true, disconnect = false, connecting = null)
    // Can be re-run multiple times for when states change
    this.updConnState(null);
  }

  runCode(code, callbackFunc) {
    // Run the code
    // Returns a promise which resolves once execution is completed
    return new Promise();
  }

  signalKernel(signal) {
    // Returns a promise which resolves if signalling was successful
    return new Promise();
  }
}

import JupyterViewer, {
  useKernelMessenger,
  useKernelReady,
  getIpynb
} from 'react-jupyter-notebook';

// If you want to update the kernel dynamically a provided hook can be used
const [messenger, changeMessenger] = useKernelMessenger(ExampleMessenger);
// A hook also exists for the kernel's state (whether it's ready, loading, or disconnected)
const kernelReady = useKernelReady(kernelMessenger);

function component() {
  // If you need to obtain the jupyter data (for example: saving purposes)
  // getIpynb() can be used
  useEffect(() => {
    const interval = setInterval(() => console.log(getIpynb()), 1000);
    return () => clearInterval(interval);
  }, []);
  return <JupyterViewer rawIpynb={rawIpynb} messenger={messenger} />;
}

import React from 'react';
import ReactDOM from 'react-dom';
import JupyterViewer from 'react-jupyter-notebook';
import nb_test from './nb_test.json'; // You need to read the .ipynb file into a JSON Object.
import ExampleMessenger from './ExampleMessenger'; // Replace with your own messenger class

ReactDOM.render(
  <React.StrictMode>
    <JupyterViewer rawIpynb={nb_test} messenger={new ExampleMessenger()} />
  </React.StrictMode>,
  document.getElementById('root')
);

Props

Prop name	Type	Description	(default) Values
rawIpynb	Object	The JSON object converted from the .ipynb file.	{ cells: [] }
KernelMessenger	Function	The class responsible for handling interactions with a Jupyter kernel	KernelMessenger

Todo

• [   ] Customizable code block styling
• [   ] Reimplement customizable styles
• [   ] Customizable output cutoff thresholds
• [   ] Support for more ascii values in output (other than \b and \r)
• [   ] Make rapid execution output updates faster (through memoization?)
• [   ] Fix vertical scaling of code cells & output (add mobile support?)
• [✓] Refactor such that prop-drilling isn't needed to pass along the KernelMessenger
• [✓] Add a default kernel messenger that interacts with Python via WASM (with JupyterLite)

Notes

codeStatus Reference:

• Within Source.js an integer variable titled codeStatus is used to keep track of the state of the code's execution within the kernel.
• See below for an explanation of what the variable can be:

  Value	Description
  -2	An error occurred during last execution
  -1	Code has never been sent (first load of cell)
  0	Code hasn't been sent to the kernel
  1	Code is sent but not running yet
  2	Code is sent and currently running

Pyolite:

• The current implementation of the Pyolite kernel has four parts:
  1. pubic/pyodide.js: Pyodide is a Python distribution for the browser and Node.js based on WebAssembly. It is used to run the Jupyter Kernel.
  2. src/Pyolite-Kernel/: A copy of the Kernel class made by @joyceerhl.
  3. src/ExampleMessenger.js: A class that interfaces with the Pyolite Kernel and forwards code & results to the library.
  4. public/Wheels/: The pure-python wheels that are loaded into Pyodide. Provided from the JupyterLite Project.