automixin class generated looks good, but then when I go to use the methods I get errors with the Graphic classes not being defined even though I am importing them in the file

seem to be having some importing issues, not sure how to fix but the generated class should be looking better!

not sure what is happening, need to keep investigating, but it seems like the generated methods are not registering *args and **kwargs

if I do something like:

import fastplotlib as fpl
import numpy as np

data = np.random.rand(512, 512)

plot = fpl.Plot()

plot.add_image(data)

when I print the *args and **kwargs, I get an empty tuple and dictionary, respectively (even when I try with data=data as kwarg)

The only way I could get the generated add_<graphic> methods to work was to include the non-default arguments (mainly data=data) in the instantiation of the given graphic for a method

all of the graphics have a non-default data arg except for text, which for now I just manually changed, will need to think of a long-term solution as more graphics are added

I think it has to do with the nature of a mixin class, both the mixin class and the graphic class expect the kwarg arg data and if it is not specified in the graphic instantiation in the mixin method then it is not getting the data arg at all regardless of how you pass it to the mixin graphic method

I think maybe my screenshots will be different because of the things that you changed in recent PRs?

Actually let's just set the type annotation as the Graphic that is returned and put in the docstring that it always returns a proxy. I think that makes inspection tools and tab completion etc. to work properly, rather than putting them as proxy[Graphic]

last question, docstrings and tab completion work in jupyterlab?

yeah, works for me!