AngleAxis and getunit error with Numpy 2.4

Reproduce the SO3.AngVec(theta=0, v=[0, 0, 1]) failure with NumPy 2.4, then inspect spatialmath/pose3d.py and spatialmath/base/transforms3d.py, including getunit and angvec2r. Done means the call no longer raises a TypeError under NumPy 2.4 and retains the expected identity rotation behavior.

Hi,

When calling SO3.AxisAngle in Python 3.11 with numpy 2.4, I get the error:

Python 3.11.11 | packaged by conda-forge | (main, Dec  5 2024, 08:47:03) [Clang 18.1.8 ] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy as np
>>> np.__version__
'2.4.0'
>>> from spatialmath import SO3
>>> SO3.AngVec(theta=0, v=[0, 0, 1])
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/opt/homebrew/Caskroom/miniconda/base/envs/vp_311/lib/python3.11/site-packages/spatialmath/pose3d.py", line 825, in AngVec
    return cls(smb.angvec2r(theta, v, unit=unit), check=False)
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/homebrew/Caskroom/miniconda/base/envs/vp_311/lib/python3.11/site-packages/spatialmath/base/transforms3d.py", line 760, in angvec2r
    R = np.eye(3) + math.sin(θ) * sk + (1.0 - math.cos(θ)) * sk @ sk
                    ^^^^^^^^^^^
TypeError: only 0-dimensional arrays can be converted to Python scalars
>>> 

However, with numpy 2.2, I get the expected behavior:

Python 3.11.11 | packaged by conda-forge | (main, Dec  5 2024, 08:47:03) [Clang 18.1.8 ] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy as np
>>> np.__version__
'2.2.6'
>>> from spatialmath import SO3
>>> SO3.AngVec(theta=0, v=[0, 0, 1])
SO3(array([[1., 0., 0.],
           [0., 1., 0.],
           [0., 0., 1.]]))

The error seems to stem from getunit, which when given a scalar now returns a 1-dimensional array instead of a 0-dimensional array. That then gets fed to math.cos, not np.cos, in angvec2r and the error occurs.

Thank you for all your work on this library,
JF