Remove optional ase dependency

[agoscinski]

The new ase release break some example notebooks. The ase dependency comes from one roy dataset (see

scikit-matter/src/skmatter/datasets/_base.py

Line 121 in bc76938

def load_roy_dataset():

)
and can easily removed as dependency by just storing the arrays inside the ase frames. We could even try to just use numpys loadtxt, then we don't need to change the binary data file.

[PicoCentauri]

I would extract the arrays and keep them either as plain ascii files or in the numpy format. We can keep the xyz file if we like.

[PicoCentauri]

What basically is to be done is saving the energies as an .npz to /src/skmatter/datasets/data and change the code loading the energies via ase to pure numpy:

scikit-matter/src/skmatter/datasets/_base.py

Lines 121 to 150 in bc76938

	def load_roy_dataset():
	"""Load and returns the ROY dataset, which contains structures,
	energies and SOAP-derived descriptors for 264 polymorphs of ROY, from [Beran et Al,
	Chemical Science (2022)](https://doi.org/10.1039/D1SC06074K)
	Returns
	-------
	roy_dataset : sklearn.utils.Bunch
	Dictionary-like object, with the following attributes:
	structures : `ase.Atoms` -- the roy structures as ASE objects
	features: `np.array` -- SOAP-derived descriptors for the structures
	energies: `np.array` -- energies of the structures
	"""
	module_path = dirname(__file__)
	target_structures = join(module_path, "data", "beran_roy_structures.xyz.bz2")
	try:
	from ase.io import read
	except ImportError:
	raise ImportError("load_roy_dataset requires the ASE package.")
	import bz2
	structures = read(bz2.open(target_structures, "rt"), ":", format="extxyz")
	energies = np.array([f.info["energy"] for f in structures])
	target_features = join(module_path, "data", "beran_roy_features.npz")
	features = np.load(target_features)["feats"]
	return Bunch(structures=structures, features=features, energies=energies)