Fix dpnp.mgrid/dpnp.ogrid inconsistency for complex steps with non-integer magnitude

[antonwolfy]

For a complex step, the integer part of its magnitude is the number of points to generate (start and stop inclusive). The single-slice and tuple-of-slices code paths disagreed when the step magnitude was not an integer (e.g. 2.5j):

• the tuple path used int(abs(step)) - 1 as the divisor (correct), while
• the single-slice path used the float magnitude abs(step) - 1 as the divisor.

As a result dpnp.mgrid[0:10:2.5j] and dpnp.mgrid[0:10:2.5j,] produced different arrays, and the single-slice form disagreed with NumPy.

>>> dpnp.mgrid[0:10:2.5j]      # before: [0. , 6.667]   after: [ 0., 10.]
>>> dpnp.mgrid[0:10:2.5j,][0]  # always: [ 0., 10.]     (matches NumPy)

• Have you provided a meaningful PR description?
• Have you added a test, reproducer or referred to an issue with a reproducer?
• Have you tested your changes locally for CPU and GPU devices?
• Have you made sure that new changes do not introduce compiler warnings?
• Have you checked performance impact of proposed changes?
• Have you added documentation for your changes, if necessary?
• Have you added your changes to the changelog?

[github-actions]

View rendered docs @ https://intelpython.github.io/dpnp/pull/2971/index.html

[coveralls]

coverage: 78.066% (-0.003%) from 78.069% — fix-issue-in-nd_grid into master

[github-actions]

Array API standard conformance tests for dpnp=0.21.0dev1=py313h509198e_27 ran successfully.
Passed: 1357
Failed: 3
Skipped: 16