Refactor Python wrapper method la.Vector.norm into function

[nate-sime]

Describe new/missing feature

I would like to move the Python wrapper method la.Vector.norm

https://github.com/FEniCS/dolfinx/blob/main/python/dolfinx/wrappers/la.cpp#L51-L55

to a function defined in the declare_functions function

https://github.com/FEniCS/dolfinx/blob/main/python/dolfinx/wrappers/la.cpp#L179

To better match the C++ structure from the Python interface the norm method

https://github.com/FEniCS/dolfinx/blob/main/python/dolfinx/la.py#L261-L270

would be refactored into a standalone function of the form

def norm(x: Vector, type: _cpp.la.Norm = _cpp.la.Norm.l2) -> np.floating:
    ...

This would further permit declaring la.Vectors containing int32_t, int64_t etc, which is very useful for, e.g., mesh associated data. E.g:

from mpi4py import MPI
import numpy as np
import dolfinx

def pprint(*msg):
    print(f"[{MPI.COMM_WORLD.rank}]: {', '.join(map(str, msg))}", flush=True)

Nx, Ny = 2, 2
mesh = dolfinx.mesh.create_rectangle(
    MPI.COMM_WORLD, [[0.0, 0.0], [1.0, 1.0]], [Nx, Ny],
    cell_type=dolfinx.mesh.CellType.quadrilateral)

im_v = mesh.topology.index_map(0)
v = dolfinx.cpp.la.Vector_int32(im_v, 1)
v.array[:im_v.size_local] = np.arange(*im_v.local_range)
pprint("before:", v.array)
v.scatter_forward()
pprint("after:", v.array)

yields

[0]: before:, [0 1 2 3 0 0 0 0 0]
[1]: before:, [4 5 6 7 8 0 0 0 0]
[1]: after:, [4 5 6 7 8 2 0 1 3]
[0]: after:, [0 1 2 3 5 7 4 6 8]

Suggested user interface

No response