Skip to content

Latest commit

 

History

History
97 lines (72 loc) · 2.84 KB

README.md

File metadata and controls

97 lines (72 loc) · 2.84 KB
Raw

Buildless JSX

Create an interactive Colab cell with JSX

Now you can use a framework like React or Preact in your Colab notebook. Just add Buildless JSX to the bottom of the cell!

How to use

Paste this cell into your notebook:

%%html
<script type = "importmap">
  {
    "imports": {
      "preact": "https://esm.sh/*preact@10.8.2"
    }
  }
</script>
<script>window.onerror = () => true // silence initial JSX parse error</script>
<script type = "module">
  window.onerror = undefined;

  import {
    createElement,
    render,
  } from 'preact';

  const root = document.getElementById('output-body');

  render(
    <div>
      This &lt;div /&gt; was rendered with JSX!
    </div>,
    root
  );
</script>

<script type = "module" src = "https://esm.sh/buildless-jsx"></script>

How it works

Buildless JSX scans all the <script> tags in the cell for JSX markers </ and />. If it finds one, it converts all the JSX in that cell to use the factory createElement.

It's the same philosophy as @babel/standalone, but it works in a notebook with zero configuration. We use it in Colab, but it should work in any cell that renders HTML, e.g Jupyter/IPython. Buildless JSX currently uses tsc for transpilation.

Buildless JSX is focused on making JSX usable in a notebook. It doesn't watch the DOM for new <script> tags. It doesn't expose hooks to configure or swap out the transpilation engine. It doesn't let you change the name of the JSX factory.

Because it must live in every cell, the import statement is whittled down to a single line.

Syntax highlighting

Colab only provides JavaScript syntax highlighting when type is set to a recognized value like module. However, this means that the browser will also try to excute these tags, before Buildless JSX has had a chance to process them.

Luckily, the JSX markers </and /> are not valid JavaScript. The browser cannot parse them, so we don't need to worry about the code being run twice. However, the browser will throw a parse error when it encounters a <script> tag with unprocessed JSX.

We can suppress this error with a bit of cleverness. Before your JSX <script> tag, add this line:

<script>window.onerror = () => true // silence initial JSX parse error</script>

Then at the beginning of your JSX <script> tag, add this one:

window.onerror = undefined;

This will prevent errors from being written to the console while the invalid JavaScript is in the DOM. As soon a Buildless JSX converts the tag to valid JavaScript, the second line will run, and the global error handler will be re-enabled.

Addendum

This is not an officially supported Google product. While we do use it internally, open-source support is provided on a best-effort basis.

License

Apache 2.0