feature request: Spinner/Loader on "use_thread"

[Lundez]

Hi!

I suggest that we add a mode where a spinner is automatically applied whenever one uses use_thread. This could potentially be

1. An input variable
2. A new function (e.g. use_thread_and_load)

I write a lot of code where I wrap this, i.e.

    df = solara_run_inference.use_thread(
        Path(model),
        Path(file),
        aggregate_duration=10,
    )

    if df.state == solara.ResultState.RUNNING:
        Progress("Running...")
    elif df.state == solara.ResultState.FINISHED:
        DfSelector(df.value, file)

and it's starting to be quite ugly as it's in multiple places.

[maartenbreddels]

Hi Hampus,

we totally recognize this boilerplate issue, but I didn't yet come up with a solution that was satisfying and something I would feel comfortable in solara.
What about a custom component that takes the function to run in the thread as an argument? Would that work?

Regards,

Maarten

[Lundez]

Hi Hampus,

we totally recognize this boilerplate issue, but I didn't yet come up with a solution that was satisfying and something I would feel comfortable in solara.
What about a custom component that takes the function to run in the thread as an argument? Would that work?

Regards,

Maarten

Hi!

That's a sound approach. The Streamlit approach is also a smooth one (using with context), alas it isn't compatible with the Solara way of execution.
I'll see if I can make a mini-component to share 👍

BR,
Hampus

[Lundez]

Still not fully solved, but at least task is an improvement. Closing for now. 👍

[maartenbreddels]

Sorry for keeping this open. We've been working on this internally a bit, but it was difficult to make something general. Keeping it open for a reminder, I think it would be good to have a good pattern for this.