Editing the `@ct.lattice` executor in a Jupyter Notebook is not always reflected if importing a `@ct.electron`

[Andrew-S-Rosen]

Environment

• Covalent version: 0.234.0rc0
• Python version: 3.10.13
• Operating system: Ubuntu

What is happening?

I have built a Python package containing a bunch of pre-made @ct.electron-decorated functions. When using a Jupyter Notebook, however, I get some unfortunate mutability issues, described in greater detail below. This isn't necessarily a bug per se, but I'd love to get your input on how to better address this conundrum.

How can we reproduce the issue?

First, I installed the necessary packages:

pip install --force-reinstall --no-deps https://gitlab.com/ase/ase/-/archive/master/ase-master.zip
pip install quacc[covalent]
covalent start

Then, in a Jupyter Notebook I ran the following in a cell:

import covalent as ct
from ase.build import bulk
from quacc.recipes.emt.core import relax_job

atoms = bulk("Cu")

@ct.lattice(executor="dask")
def workflow(atoms):
    return relax_job(atoms)

ct.dispatch(workflow)(atoms)

This will properly use the "dask" executor. However, when I edit executor="dask" to be executor="local" and re-run the same cell, the executor in the GUI still shows as "dask" when it should be "local".

I believe the reason for this comes down to mutability. It seems that doing

relax_job.electron_object.executor

when I run it the first time yields "dask". However, when I rerun with the updated executor, it still shows "dask" even though we are running with "local".

What should happen?

The imported @ct.electron should have its executor dynamically updated.

Any suggestions?

None, but I'm open to suggestions! This seems potentially related to #1889, but I think it may be slightly different. Feel free to close if it's a duplicate.