Disallow scalar conversion for non-0D arrays

[vlad-perevezentsev]

This PR disallows converting non-0D dpnp.ndarray to Python scalars (int, float and complex) and raises TypeError` to align with NumPy 2.4.0

• 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?

[vlad-perevezentsev]

NumPy allows converting non-0D ndarrays to bool and also for __index__ method which contradicts the Python Array API. Should we disallow this as well to comply with the Python Array API?

[github-actions]

View rendered docs @ https://intelpython.github.io/dpnp/pull/2694/index.html

[github-actions]

Array API standard conformance tests for dpnp=0.20.0dev1=py313h509198e_27 ran successfully.
Passed: 1229
Failed: 0
Skipped: 7

[vlad-perevezentsev]

NumPy allows converting non-0D ndarrays to bool and also for __index__ method which contradicts the Python Array API. Should we disallow this as well to comply with the Python Array API?

bool() is truth-value testing, not a numeric conversion.
Since NumPy only deprecated/disallows numeric scalar conversions and the Python Array API does not explicitly forbid ndim>0 scalar conversions so the current behavior can be kept

__index__ is not part of numeric scalar conversion semantics

[antonwolfy]

Do you consider implementing that check in dpctl? Since current dpctl implementation seems breaking Python array API spec:

array.int() → int
Converts a zero-dimensional array to a Python int object.

[coveralls]

coverage: 81.455% (+0.005%) from 81.45%
when pulling 7e07ffe on disallow_conv_to_scalar_ndim
into 21718ad on master.