Getting markdown content on collab server #1394

VincentK-ZA · 2024-06-26T16:14:46Z

VincentK-ZA
Jun 26, 2024

When using the collab plugin, it relies internally on the y-prosemirror project, meaning the content synced into the yjs doc is an xml fragment.

i.e. if my markdown content is:

# hello world

then yDoc.getXmlFragment("prosemirror").toJSON() will give:

<heading id="hello-world" level="1">hello world</heading>

Is there an easy way to get the markdown content on the server running the yjs connection provider?

I've currently got something working in a very hacky way, whereby I create an instance of the Milkdown editor on the nodejs + express server, relying on jsdom

export function setup_jsdom() {
  const dom = new JSDOM(`<!DOCTYPE html><p></p>`);
  global.window = dom.window;
  global.document = dom.window.document;
  global.addEventListener = dom.window.addEventListener;
  global.removeEventListener = dom.window.removeEventListener;
  global.dispatchEvent = dom.window.dispatchEvent;
  global.CustomEvent = dom.window.CustomEvent;
}

export async function create_milkdown_editor() {
  const editor = await Editor.make().use(commonmark).use(gfm).use(emoji).create();

  let serializer = null;
  let schema = null;
  editor.action((ctx) => {
    serializer = ctx.get(serializerCtx);
    schema = ctx.get(schemaCtx);
  });

  return { serializer, schema };
}

export async function get_markdown(doc, serializer, schema) {
  const fragment = doc.getXmlFragment("prosemirror");

  const node = yXmlFragmentToProseMirrorRootNode(fragment, schema);

  const markdown = serializer(node);

  console.log(markdown);
}

So I setup jsdom once and create the editor once.

I can then use the serializer and schema to create the markdown from the yXmlFragment

setup_jsdom();
const { serializer, schema } = await create_milkdown_editor();
//...
ysocketio.on("document-update", function (doc, update) {
  get_markdown(doc, serializer, schema);
});

So this works, but I'm sure there must be a way to setup the necessary plugins and create the serializer and schema only, without needing the view.

The reason for needing this functionality is since we'd like to only work with the markdown as a source of truth, and have the server periodically persist the markdown to a nosql database.
I do realize that we would need to ensure that the plugins used on the frontend match those on the backend, else content may get lost.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Milkdown

Getting markdown content on collab server #1394

{{title}}

Replies: 0 comments

Select a reply

Milkdown

Getting markdown content on collab server #1394

VincentK-ZA Jun 26, 2024

Replies: 0 comments

VincentK-ZA
Jun 26, 2024