Migrating to aiida v2.5 and optimade-python-tools 1.0

[eimrek]

Hi @CasperWA, I would really need some fixes that are in aiida v2.5.0 (namely, mutating nodes with iterall()) to enable the optimade api for some large DBs on materials cloud (would be great to have before submisison of the optimade paper).

However, aiida v2.5.0 migrated to pydantic 2 and therefore it looks like the optimade-python-tools dependency should also be upgraded to 1.0.

Additionally, aiida v2.5 removed the "legacy testing" that was used in this repo. This i managed migrate to the new format.

However, i am not able to adapt this repo to optimade-python-tools 1.0. No specific error is raised, but just the functionality is not correct (e.g. it re-calculates the optimade extras every time; there seem to be spurious optimade attributes in random places, etc).

Would be great if you could take a look! Feel free to contribute directly to this PR.

[CasperWA]

Indeed, the upgrade to pydantic v2 (and more importantly to OPTIMADE v1) is a large task.
Mainly due to the somewhat "hacky" way the OPTIMADE attributes are calculated by this package, stored, and handled --- it could be better. But alas.

I don't have much (if any) time to spend on this. Especially not at the moment. However, I may take a peek after hours in a couple of weeks.
If this is too late, feel free to take complete control and do whatever you wish here :) I would love to contribute and help where I can, but I am afraid this will take more time than I can afford currently.

[eimrek]

thanks for the info @CasperWA. As it's a lot of work, I think for now, we might adapt a different approach (https://github.com/materialscloud-org/optimade-maker) for the larger DBs.

And also pragmatically, as maintaining this repo is a constant source of work (due to aiida-core getting updated), we might need to consider moving all the mcloud optimade APIs to a simpler approach. We'll discuss in the team.