Return complex arrays from eig / eigvals

[antonwolfy]

NumPY 2.5 changed numpy.linalg.eig and numpy.linalg.eigvals so that for general (non-symmetric) matrices they always return complex-typed eigenvalues, instead of value-dependently casting the result back to a real type when the imaginary parts happen to be zero. For real input, float64 now yields complex128 and float32 yields complex64.

dpnp.linalg.eig / eigvals are thin NumPy fallbacks (the OneMKL geev routine is not yet wired up), so they inherit this behavior directly from the installed NumPy.

This PR brings dpnp's documentation in line with the new always-complex output.

• 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/2953/index.html

[coveralls]

coverage: 78.116% (+0.003%) from 78.113% — adopt-to-complex-dtype-in-eig-eigvals into master

[github-actions]

Array API standard conformance tests for dpnp=0.21.0dev1=py313h509198e_8 ran successfully.
Passed: 1355
Failed: 5
Skipped: 16

[vlad-perevezentsev]

LGTM
Thank you @antonwolfy