Add dpnp.tensor.broadcast_shapes to align with the 2025.12 Python array API

[antonwolfy]

The 2025.12 revision of the Python array API standard adds a standalone broadcast_shapes function that broadcasts one or more shapes (as opposed to arrays) against one another.

The function was already present in the top-level dpnp namespace but missing from dpnp.tensor. This PR closes that gap.

• 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/3009/index.html

[github-actions]

Array API standard conformance tests for dpnp=0.21.0dev3=py314h509198e_22 ran successfully.
Passed: 1375
Failed: 2
Skipped: 5

[coveralls]

coverage: 78.414% (+0.008%) from 78.406% — add-broadcast_shapes-to-dpnp.tensor into master

[ndgrigorian]

Suggested change

	normalized = [tuple(operator.index(dim) for dim in sh) for sh in shapes]
	normalized = [tuple(map(operator.index, sh)) for sh in shapes]

map is often a bit more efficient than a generator expression from my understanding

[ndgrigorian]

other than efficiency point LGTM