Add simple_efficiency tests for numpy array inputs

[aman-coder03]

this PR adds test coverage for vectorized use of pvlib.transformer.simple_efficiency with numpy.ndarray inputs.

the tests verify:

• support for NumPy array inputs for input_power, load_loss,
  and no_load_loss
• consistency between vectorized and scalar evaluations

No functional changes are introduced.

fixes #2649

[aman-coder03]

the CI failures are due to deprecation-enforcement tests in tests/test_irradiance.py, unrelated to this PR.
this PR only modifies tests/test_transformer.py.
please let me know if you would like me to rebase

[cwhanse]

@aman-coder03 this is the scope of #2658

Could you complete #2646 first? It needs a whatsnew note.

[aman-coder03]

@cwhanse shall I create pvlib-python\docs\sphinx\source\whatsnew\v0.13.2.rst ?
and then add

* Fix a division-by-zero condition in
  :py:func:`~pvlib.transformer.simple_efficiency` when ``load_loss = 0``.
  (:issue:`2645`)

in Bug Fixes

[cwhanse]

It should be v0.14.1.rst. Thanks!

[aman-coder03]

@cwhanse i have completed that, you can have a look!!

[aman-coder03]

@cwhanse as suggested by you in #2646 , i have implemented the additional tests in this separate PR to avoid scope creep in #2646. Let me know if you would like any adjustments.

[cwhanse]

@aman-coder03 I think it is premature to test with all inputs as np.array. In other discussions we don't have consensus to allow the parameters no_load_loss and load_loss to be anything other than float, and intend to edit the docstring accordingly.

If we agree in the future to allow vectors for these parameters, then this test would be needed.

I don't want to include it now so to prevent someone's AI from deciding we have missed a combination of types in other tests.

[aman-coder03]

Would you prefer that I remove the test that uses array inputs for no_load_loss and load_loss, and keep only the vectorized input_power case? Or should I drop the vectorized tests entirely for now?

[cwhanse]

I think drop test_simple_efficiency_numpy_array_inputs and keep test_simple_efficiency_vector_equals_scalar

[aman-coder03]

implemented the suggested changes @cwhanse , thanks for your guidance!!

[cwhanse]

Looks OK now. Please add a brief note in the Testing section of this file: https://github.com/pvlib/pvlib-python/blob/main/docs/sphinx/source/whatsnew/v0.15.1.rst

[aman-coder03]

added whatsnew entry in the testing section, please have a look.
thanks again!! @cwhanse

[echedey-ls]

I suggest to merge the tests into one, most functions are type-tested in the same test for any numeric implementation, and it can be shortened with the @pytest.mark.parametrize.

[echedey-ls]

I suggest to merge the tests into one, most functions are type-tested in the same test for any numeric implementation, and it can be shortened with the @pytest.mark.parametrize.

@aman-coder03 , for my previous comment, I suggest you give it a try doing it without AI, it's not difficult to understand, you have an example in the same file, documentation is public and easily accessible, and we prefer contributors to know what they are doing rather going back and forth to an LLM model.

[echedey-ls]

Suggested change

	(:pull:`2661`)
	(:issue:`2649`, :pull:`2661`)

[aman-coder03]

I refactored the numeric tests into a single parametrized test following the pattern used elsewhere in the file. I reviewed the pytest documentation to make sure I understood the approach
Please let me know if this structure looks good @echedey-ls thanks!!