get_radial_basis_covariance gives error

[tavanfran]

Dear librascal developers,

I am trying to use the jupyter notebook called "optimized_radial_basis_function" in your example directory. I get the following error:

cov = get_radial_basis_covariance(spex, feats)
Traceback (most recent call last):
File "", line 1, in
File "/usr/local/lib/python3.10/site-packages/rascal/utils/radial_basis.py", line 258, in get_radial_basis_covariance
].reshape((n_environments, max_radial, (max_angular + 1) ** 2))
ValueError: cannot reshape array of size 0 into shape (1000,30,81)

The variable feats gives me:

feats
{(1,): array([[ 1.05488364e-06, 4.92687408e-09, -3.85978917e-10, ...,
6.56854997e-10, -1.71375982e-10, -1.80129393e-09],
[ 5.16306113e-07, 1.15980425e-08, -2.97132210e-09, ...,
-8.48864522e-08, 1.09817793e-08, 1.19521158e-07],
[ 3.02115704e-07, 1.14544918e-08, 4.79069661e-10, ...,
3.21842435e-11, 8.18395871e-12, -4.99493908e-11],
...,
[ 2.12966817e-04, -1.22929392e-19, -1.30212210e-19, ...,
3.61752414e-09, 3.84588223e-09, 5.44978080e-09],
[ 2.12966817e-04, -5.04722147e-15, 1.64968448e-15, ...,
-6.05091733e-09, -1.25810418e-09, 2.85810710e-09],
[ 2.12966817e-04, 5.04728299e-15, -1.64968696e-15, ...,
3.75128458e-13, -4.00480187e-13, 7.70920793e-14]]), (6,): array([], shape=(0, 0), dtype=float64), (8,): array([], shape=(0, 0), dtype=float64)}

which seems that it finds only H-atoms and not C and O atoms.
I tried with a different input structure, but I get the same error.

I am able to run librascal on my systems and get the output model in json format and when I load all packages I get no errors.

Could you help me?

Thank you

[Luthaf]

So this notebook runs fine for me with the latest master. Did you make any changes to it by chance? In particular to the spherical_expansion_hypers?

If not, what Os are you running on, and what's your Python version?

Running [print(k, v.shape) for k, v in feats.items()] gives me

(1,) (1000, 1620)
(6,) (1000, 1620)
(8,) (1000, 1620)