Create repo for nanite models in AFM-Analysis

[PinkShnack]

As discussed in a previous PR, it would be good to have the external Hertz KVM model in a repository for nanite models.

• The KVM model exists here: https://data.mendeley.com/datasets/c2gccnfkgd/2, but a repo here would be practical for when we need to update the code due to e.g. numpy version changes.

Todo

• Create a repo, similar to https://github.com/AFM-analysis/nanite_model_sneddon_spher I guess, perhaps we could call it "nanite_model_extensions" or "nanite_external_models"?
• add the up to date models. For example, the KVM model should have two versions, one (original from paper) for numpy v1 and one for numpy v2. Reason:
  • fit\_stop = int(np.argwhere(force < (max\_force * fraction_force))[0])\ fails for numpy v2

  • # numpy 1 and 2 compatible code
    _mask = force < (max_force * fraction_force)
    if not np.any(_mask):
        raise ValueError("No force values below threshold")
    fit_stop = int(np.argmax(_mask))

• Have a section in the nanite docs website referring to this repo
• Show how to load this extension in nanite.
• Should nanite have this as a dependency?

[paulmueller]

I created https://github.com/AFM-analysis/nanite_community_models

I think this is a good place to put all models that people come up with. I gave you write access.

I don't have any preferences on how to incorporate this into the nanite workflow. A separate nanite-community package with an extra_requires in the nanite pyproject.toml is probably most user convenient and easy to maintain.

For a start however, I think it is fine to just have a list of files that people can selectively download and load into their PyJibe.

[PinkShnack]

Savage, good name too