How to script or edit UI?
I'm working on a realtime scripting system: https://docs.openbrush.app/alternate-and-experimental-builds/runtime-scripting
14 Replies
Currently UI for the new Plugin Scripts is limited to declaring that a given parameter is to be controlled from a slider but I am planning other widget types. (The existing HTTP API allows you to use any language/UI but of course the snag is that it appears outside the app)
If you want to play around in the C# code, the developer docs have some info on how the UI system works: https://docs.openbrush.app/developer-notes
@enigmatron_ :
Thanks, I'll play around with the code a bit. I just wanted make a few API calls to a web server first and take it from there.What kind of calls? I'm working on outgoing calls myself for something else. The bit I'm uncertain about is whether and how to go async (or not to bother) Scripts currently run on the main thread.
I wanted integrate some AI stuff. Stable Diffusion, Point-E, just make the call and bam.
Ah OK. And would the HTTP API not be an easier way to do this?
I can run a script on the side but you know switching apps isn't that flowy.
PC or standalone?
Standalone.
Yeah. It's a couple of clicks (assuming Meta don't stop breaking the ability to switch to the browser without quitting the app)
I presume you'll need a keyboard for entering the prompts?
Yes. It would work for just the prompts but if we want to render the depthmaps, canny maps a UI would be really helpful. I also find posing characters easy on VR.
How do I run the experimental builds side by side with the stable one?
On the Quest? They install to "Unknown Sources" - you can have many builds installed.
I meant "how are you implementing a keyboard?" We have a pull request for one so best to use that.
Oh thanks.
GitHub
Feature/xr keyboard by andybak · Pull Request #406 · icosa-gallery/...
Including a basic "rename sketch" feature.
I guess I'm trying to think how what you're planning might be part of a broader feature. Much more likely to get accepted for a pull request if it can be generalized.
I've been holding off on text-to-3d AI because they are fairly rough but there's been a lot of activity recently.
text-to-image is great but without some 3d element, there's not a huge amount of value to adding it to the main app. If we got automatic depth maps and rendered them with parallax - now that would start to sound interesting!
I'm missing a good parallax shader at the moment.