Improve emulator re-initialisation

[radka-j]

Closes #748
Closes #757
Closes #874

Overview:

• emulators save all their input args so that all input values can be retrieved
• replace any **kwargs with optional keyword arguments to match use (they are mostly used to handle for scheduler kwargs)
• update HMW so that user can pass emulator as well as result
• re-initialize emulators when refitting in AL

TODOs:

• add tests
• should the reinitialisation code be a method to avoid duplication of code?
• check attribute access error issue in HMW (lines 383 and 388)

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[sgreenbury]

Just adding a note here as ran into this when working with a GP subclass for the error quantification. This call:

autoemulate/autoemulate/core/compare.py

Line 578 in 76689ae

model_class = get_emulator_class(result.model_name)

fails since:

autoemulate/autoemulate/emulators/__init__.py

Lines 68 to 70 in 76689ae

	emulator_cls = EMULATOR_REGISTRY.get(
	name.lower()
	) or EMULATOR_REGISTRY_SHORT_NAME.get(name.lower())

doesn't also look at:

autoemulate/autoemulate/emulators/gaussian_process/exact.py

Lines 460 to 463 in 76689ae

	GP_REGISTRY = {
	"GaussianProcess": GaussianProcess,
	"GaussianProcessCorrelated": GaussianProcessCorrelated,
	}

@radka-j - adding here as it might be addressed by the upcoming changes to this API? But if not happy to open a new issue to look at this. An option could also be to revisit having a central registry class to handle this uniformly.

[radka-j]

@sgreenbury I don't think we should ever use the GaussianProcess or GaussianProcessCorrelated classes so this to me feels like correct behaviour. If we want a GP class for an RBF + constant kernel we should add that specifically to the registry.

[sgreenbury]

It was the GP context (passing a create_gp_subclass instance to AutoEmulate) I ran into this issue and a workaround might have been to also look at GP_REGISTRY since this maintains a registry of all GPs including the created subclasses.

But thinking more about it, it affects any subclass used by AutoEmulate currently if reinitialize is called, e.g. in the advanced tutorial:

class SimpleFNN(PyTorchBackend):
    ...
ae = AutoEmulate(x, y, models=[SimpleFNN])
ae.fit_from_reinitialized(x, y)

since SimpleFNN is constructed at runtime the class is not found in the lists of emulators.

I think if the emulator becomes the entity that does the refitting in this PR then a global emulator registry including all custom subclasses would not be needed for this but might still be useful?