gradient

PlasmaCalcs.quantities.patterns.vector_derivatives.gradient(array, components=[Component('x', 0), Component('y', 1), Component('z', 2)], *, _post_slices=None)

return gradient of array as a single vector array.

The result will have ‘components’ dimension.

components: iterable of str or Component objects

which components to include in the result.

result[‘component’][i] will be components[i],

and result.isel(component=j) will be the derivative along str(components[j]).

E.g., if components=[Component(‘x’, 0), Component(‘y’, 1)],

then result[‘component’][1] == Component(‘y’, 1),

and result.isel(component=1) = d(array)/dy.

The derivative is taken along the corresponding dimension,

e.g. d(array)/dy is xarray_differentiate(array, dim=’y’).

Note that any components provided here but missing from array will become 0 in result.

e.g., components=[x,y,z] but ‘z’ not in array –>

np.all(result.isel(component=2) == 0).

_post_slices: None or dict of indexers

[EFF] if provided, apply these indexers to results after taking each derivative,

but before merging into the final array.

Equivalently, could just result.isel(**_post_slices).

But, indexing before merging can help avoid ever creating a large array.