Enable Asynchronous Execution of Python Functions with Timer Component about streamsync

[ramedina86]

If you have something that you want to store in state but avoid serialising, you can prefix them with an underscore. More info here.

Numpy arrays are serialisable, however you may have something inside them that's preventing serialisation. For example, complex numbers.

Here's an example of a non-serialisable Numpy array that can be kept in state, but isn't serialised (nor send to the frontend).

initial_state = ss.init_state({
    "_k": np.array([1+2j, 2, 3])
})

def print_k(state):
    print(state["_k"])

Hope this helps!

[ramedina86]

Hello, I'm glad to hear you're enjoying this library and that it's working well for you.

I have good news, I've just released a new version which runs all event handlers asynchronously to better support long-running tasks. It seems to be perfectly in line with what you were requesting. There's also other new features targeting background/long-running tasks.

It seems like you're building something interesting by the way.

[sarim-zafar]

Yes I can see that however I can't seem to put some of the variables in the state dict as they are non serialisable (numpy arrays). What's the proposed way of tackling this issue as I want to avoid using global variables. Is there anyway I can send in other variables with events?