test075_siemens_cios_alpha and spekpy issue

[tbaudier]

With numpy version > 2.4.0, spekpy leads to error:

  File "<stdin>", line 1, in <module>
  File "/home/hbaldau/Software/spekpy_release/spekpy/SpekPy.py", line 159, in __init__
    self.set_state_parameters(kvp=kvp, th=th, dk=dk, physics=physics,
  File "/home/hbaldau/Software/spekpy_release/spekpy/SpekPy.py", line 246, in set_state_parameters
    self.spectrum_from_model()
  File "/home/hbaldau/Software/spekpy_release/spekpy/SpekPy.py", line 326, in spectrum_from_model
    self.model = SpekModel().get_spectrum_parameters(self)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/hbaldau/Software/spekpy_release/spekpy/SpekModel.py", line 148, in get_spectrum_parameters
    s = aniso.SpekAniso(kvp=E0,x=x,y=y,z=z,target=anode_material,
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/hbaldau/Software/spekpy_release/spekpy/SpekAniso.py", line 104, in __init__
    self.anode_self_filtration_char, self.t_char = self.__Char()
                                                   ^^^^^^^^^^^^^
  File "/home/hbaldau/Software/spekpy_release/spekpy/SpekAniso.py", line 813, in __Char
    x = arange(0,dmax+0.5*dx,dx)
        ^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: only 0-dimensional arrays can be converted to Python scalars

After investigation, we cannot do anything on our side, the problem has to be resolved by spekpy:

diff --git a/spekpy/SpekAniso.py b/spekpy/SpekAniso.py
index 649fa18..7b5cf61 100644
--- a/spekpy/SpekAniso.py
+++ b/spekpy/SpekAniso.py
@@ -804,6 +804,7 @@ class SpekAniso:
             dmax = 20.*csda_range

         # Define the escape thickness depending on transmission or reflection
+        dmax = dmax[0]
         x = arange(0,dmax+0.5*dx,dx)
         if self.trans:
             x_escape = self.thick - arange(0,dmax+0.5*dx,dx)
@@ -922,7 +923,7 @@ class SpekAniso:
                                sin(abs(self.varphi))**-1 * cos(self.vartheta)**-1 )
             # Characteristic freq. dist. per solid angle per keV, with the
             # ... self-filtration removed (applied again outside this class)
-            char_kx[i,:] = divide(char_kx[i,:],attn_factor,
+            char_kx[i,:] = divide(char_kx[i,:],attn_factor, out=None,
                         where=attn_factor>finfo(dtype='float64').resolution)

         if self.shape == 'kqp' or self.shape == 'sim':