numpy>=2 support

[vincefn]

It seems there is an issue related to numpy>=2:

In [1]: import skcuda.fft as cu_fft
/home/esrf/favre/.conda/envs/pynx-devel-numpy2-x86/lib/python3.12/site-packages/skcuda/cublas.py:284: UserWarning: creating CUBLAS context to get version number
  warnings.warn('creating CUBLAS context to get version number')
---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
Cell In[1], line 1
----> 1 import skcuda.fft as cu_fft

File ~/.conda/envs/pynx-devel-numpy2-x86/lib/python3.12/site-packages/skcuda/fft.py:20
     15 from .cufft import CUFFT_COMPATIBILITY_NATIVE, \
     16      CUFFT_COMPATIBILITY_FFTW_PADDING, \
     17      CUFFT_COMPATIBILITY_FFTW_ASYMMETRIC, \
     18      CUFFT_COMPATIBILITY_FFTW_ALL
     19 from . import cudart
---> 20 from . import misc
     22 class Plan:
     23     """
     24     CUFFT plan class.
     25 
   (...)     60         indicates whether the caller intends to allocate and manage the work area
     61     """

File ~/.conda/envs/pynx-devel-numpy2-x86/lib/python3.12/site-packages/skcuda/misc.py:637
    633 diff.cache = {}
    636 # List of available numerical types provided by numpy:
--> 637 num_types = [np.sctypeDict[t] for t in \
    638              np.typecodes['AllInteger']+np.typecodes['AllFloat']]
    640 # Numbers of bytes occupied by each numerical type:
    641 num_nbytes = dict((np.dtype(t),t(1).nbytes) for t in num_types)

KeyError: 'b'

Also:

  File "/home/esrf/favre/.conda/envs/pynx-devel-numpy2-x86/lib/python3.12/site-packages/numpy/__init__.py", line 400, in __getattr__
    raise AttributeError(
AttributeError: `np.sctypes` was removed in the NumPy 2.0 release. Access dtypes explicitly instead.

PR incoming