Adds copy parameter to __array__ for numpy 2.0

[andrew-s28]

xarray has been raising some deprecation warnings whenever np.array is called on an xr.DataArray. This is the same message as documented in #9312, but the issue is not specifically related to the interpolate_na method. Instead, the issue can be produced when using NumPy>=2.0.0 as:

>>> import xarray as xr
>>> import numpy as np
>>> da = xr.DataArray([0])
>>> np.array(da)
DeprecationWarning: __array__ implementation doesn't accept a copy keyword, so passing copy=False failed. __array__ must implement 'dtype' and 'copy' keyword arguments.

It is caused by the AbstractArray.__array__ method not containing a copy kwarg that NumPy 2.0 now expects.

A fix was proposed in #9312 by @kmuehlbauer, but upon testing it I found that this fix is not backwards compatible when using NumPy versions<2.0:

>>> import xarray as xr 
>>> import numpy as np
>>> da = xr.DataArray([0])
>>> np.array(da)
ValueError: NoneType copy mode not allowed.

The only way I found that would be able to address the deprecation warning and maintain backwards compatibility was to use a conditional in the AbstractArray.__array__ method based on the installed NumPy version, which is how this PR is structured (based on the suggested implementation from the migration guide, from this SciPy PR).

Using a conditional based on a version doesn't feel great, but the only other option I see is just ignoring the copy the kwarg argument that is passed into the __array__ method, which is ignoring user input and also isn't great. I'd be interested to see if there are any more elegant ways to deal with this deprecation warning, though.

That said, if xarray ever drops support for NumPy<2.0, then this can be reverted to the more straightforward fix proposed in #9312.

• Closes storing results of interpolate_na into numpy.array #9312
• User visible changes (including notable bug fixes) are documented in whats-new.rst

[welcome]

Thank you for opening this pull request! It may take us a few days to respond here, so thank you for being patient.
If you have questions, some answers may be found in our contributing guidelines.

[max-sixty]

Thanks @andrew-s28, this looks good.

Could we add a test that there's no warning?

[andrew-s28]

@max-sixty test is now included (and seems to be passing).

I accidentally synced my feature branch (instead of main) with the upstream and it goofed up the commit history a bit, so my apologies. I seem to be making it worse trying to rebase, so let me know if there's anything I should do to fix it up.

Think I have it cleaned up now. Sorry about that, this was the first time where merges to the upstream came in after my PR that I had to manage. Some lessons learned. 👍

[welcome]

Congratulations on completing your first pull request! Welcome to Xarray! We are proud of you, and hope to see you again!

[max-sixty]

Thanks a lot!

(No prob re commits — it all gets squashed on merge anyway!)