Update action version numbers

[zm711]

I'm dropping test of 3.10 because it is EOL and we will need to start testing 3.15 shortly. No rush in merging this.

[zm711]

@alejoe91 what do you think about jumping our python release to 3.14? I don't remember why we chose 3.10 to do the release in the action in the past. We could also jump to 3.12 or 3.13 if we don't want to go to advanced?

[h-mayorquin]

I think 3.14 is fine for releases, but question, can we just drop python 3.10 as its own thing, that can be done now and I feel would make things easier?

And I think we should do it now, quoting one of the masters

“With Python 3.10, it is not a question of whether to drop support, but of when. If you ask why not drop it in the next release, I say why not today? If you say today at 5 o’clock, I say why not one o’clock?”

— John von Neumann, on Python version support, allegedly

[zm711]

Nice (for 3.14!)

You're saying you'd prefer to drop 3.10 away from dropping NumPy? The reason why I did it this way is:

updating actions is basically free and doesn't affect our code so get this PR merged first then in quick succession merge #1922.

The reason why I pooled NumPy and Python is that we usually keep the lowest Python that NumPy supports for Neo. This is different from SpikeInterface and I know causes a little friction between keeping the packages playing nicely together, but as this is our long standing policy I didn't want to change it. Thus I viewed this more as update actions then drop NumPy and because we are dropping NumPy then we have to drop 3.10. Technically 3.10 had one final release that we could support for a little longer, but if we are updating NumPy there is no point. I didn't see anyone else respond to Andrew's suggestion so I also didn't want to jump to make the change of NumPy + Python without everyone agreeing.

Does that all make sense?

[h-mayorquin]

Yes. I should have answered yes to Andrew suggestion though, I agree with it.