BUG: avoid soon-to-be-deprecated direct mutations of ndarray.shape attributes (1/2)

[neutrinoceros]

PR Summary

xref: numpy/numpy#29536

This is incomplete at the time of opening, I just want to take what I already have for a spin on CI

PR Checklist

• New features are documented, with docstrings and narrative docs
• Adds a test for any bugs fixed. Adds tests for new features.

[matthewturk]

Your work on making this backward- and forward-compatible is admirable, and I wish it didn't have to be so.

[neutrinoceros]

yeah it's a bit fiddly. Hopefully it should be straightforward to find-and-replace when the time comes.

[matthewturk]

What I was trying to say without saying it is that I wish we didn't have to make this change, and I agree it looks a bit fiddly to use the COPY object etc, but I guess we have to because of upstream changes. So, thank you.

[neutrinoceros]

Oh yeah, this has been on my radar for a while but I was constantly putting it off, but it seems to be moving forward upstream...

[neutrinoceros]

Clearly I messed up somewhere. Help is welcome, but in the meantime, I'll be off duty for today.

[neutrinoceros]

ok this is now reasonably stable, so, opening for review.
Note that I avoided dealing with yt/volume_rendering to save whatever little time, but this needs to be done too (preferably by someone else, that I can then review).

[neutrinoceros]

... I couldn't resist. Volume Rendering is now handled in its own PR to hopefully make reviewing swifter. See #5309

[cphyc]

Thanks so much for this! This is handy work :)

Upon reviewing the code, I found several instances where we manually set the shape (primarily in the .pyx files).

• https://github.com/neutrinoceros/yt/blob/f1f1a923e47ba39beea9ce2ef962e13cd540fc7b/yt/geometry/grid_geometry_handler.py#L347-L348
• https://github.com/neutrinoceros/yt/blob/f1f1a923e47ba39beea9ce2ef962e13cd540fc7b/yt/geometry/particle_deposit.pyx#L191-L192
• https://github.com/neutrinoceros/yt/blob/f1f1a923e47ba39beea9ce2ef962e13cd540fc7b/yt/geometry/particle_deposit.pyx#L289-L290
• https://github.com/neutrinoceros/yt/blob/f1f1a923e47ba39beea9ce2ef962e13cd540fc7b/yt/geometry/particle_deposit.pyx#L347-L348
• https://github.com/neutrinoceros/yt/blob/f1f1a923e47ba39beea9ce2ef962e13cd540fc7b/yt/geometry/particle_deposit.pyx#L399-L400
• https://github.com/neutrinoceros/yt/blob/f1f1a923e47ba39beea9ce2ef962e13cd540fc7b/yt/geometry/particle_deposit.pyx#L438-L439
• https://github.com/neutrinoceros/yt/blob/f1f1a923e47ba39beea9ce2ef962e13cd540fc7b/yt/geometry/particle_deposit.pyx#L548-L549

[neutrinoceros]

I have no idea how/if the C side of things will change, so I haven't updated it yet.

[chrishavlin]

changes LGTM. I'm a little worried the changes in the frontends might not get picked up by tests at the moment... but all the changes look functionally identical on a careful read. So unless anyone has bandwidth to run some nosetests locally, I say we merge