Highlight groups/nodes from description text using references

Core Development
1

Hi There,

So after duplicating the preview feature, I have another feature request or rather I'm wondering whether this is being worked on. The central itch being scratched is the abiliity to highlight nodes or groups in flows from description text.

I write a text about a flow and have to write things like "the node in the top left corner" instead of having a highlight mechanism "the <a ref-to-node-in-flow>node</a> does this that and ...."

What is shown in the animation is how I implemented this feature. I have a pencil in the info panel and when I highlight a node (in the info panel), the pencil generates a ref link that I can include in my markdown description.

When the description text is displayed in the bottom half of the info panel, I can click on the link and the corresponding node is highlighted in the flow. The reference links are highlighted in pink to indicate that there highlight links.

node-highlight

Can this be done already by some other trick? Or is being worked on and how it is being implemented? I would like to keep my solution in sync with the "official" node-red solution if there is one.

2

It seems feasible to me, even if I don't know where to add it yet. I wonder if it already exists?

3

I should also point out that I do this not only for in-node-red gratification but rather also for website hosting of flows --> for example, the flow show in the animation is over here as a web page.

So it might well make little sense in node red itself to do this highlighting but for external documentation it makes much sense - IMHO.

4

To me, that seems a bit niche to go into core?

Wouldn't it be better to make sure that core has a node selection capability (by id, position, text/numbers in any field, text in comments, etc).

Then people could create different tools to use that. I could imagine, for example, an LLM AI tool that would do the selection for you. But locking Node-RED core into an LLM would, in my opinion, be a mistake and would likely increase the size of core very significantly. So, better in a custom node. Doubtless there are other ways of achieving this as well.

5

To niche for it to be worked on for core. I was asking whether there is any work going on in that direction. True in my own self-interest so that I don't duplicate features again.

It does and that is what I'm using:

RED.view.reveal(ndeIds[0], true)
RED.view.redraw();

and

RED.view.selectNodes({
    selected: ndeIds,
    onselect: function(selection) { RED.tray.show(); },
    oncancel: function() { RED.tray.show(); }
});

Won't the logical conclusion of this argument end with: why do I both to code at all? And why using Node-RED since AI is far better for textual code? After all an LLM/AI can also code the solution to my problem.

Hm. The solution I currently have is approximately 80 lines of JS code.

Compared to the 55k lines of the un-minified red.js that is 0.15% - if that is significant then so be it.

But you are perfectly correct, if the core was better documented and exposed to usage by third-party nodes, then this could be pushed into a plugin for the sidebar.