Bringing in MACE-MP workflows to CARMM

[ikowalec]

Now removed the [WIP] tag.

TODOs:

• minimise code repeats and merge react.py and react_mace.py
• vibrations
• MD (missing for FHI-aims also)

Done:

• geometry optimisation with restarts
• NEBs with restart
• Added MACE to the CI test
• choice of MACE_MP, MACE_OFF and MACE_ANICC models added

[ikowalec]

The CI tests will likely fail due to missing imports. Rebuilding torch as part of CI is not an option due to size - @GabrielBram ideas?

Edit: for now using a dry_run flag and EMT() calculator

[codecov[bot]]

Codecov Report

Attention: 19 lines in your changes are missing coverage. Please review.

Comparison is base (68fddf9) 86.58% compared to head (672132b) 87.50%.

Files	Patch %	Lines
carmm/run/workflows/react_mace.py	87.39%	15 Missing :warning:
carmm/run/workflows/helper.py	71.42%	4 Missing :warning:

Additional details and impacted files

```diff @@ Coverage Diff @@ ## master #138 +/- ## ========================================== + Coverage 86.58% 87.50% +0.92% ========================================== Files 72 74 +2 Lines 2847 3033 +186 ========================================== + Hits 2465 2654 +189 + Misses 382 379 -3 ``` | [Flag](https://app.codecov.io/gh/logsdail/carmm/pull/138/flags?src=pr&el=flags&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Andrew+Logsdail) | Coverage Δ | | |---|---|---| | [unittests](https://app.codecov.io/gh/logsdail/carmm/pull/138/flags?src=pr&el=flag&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Andrew+Logsdail) | `87.50% <90.10%> (+0.92%)` | :arrow_up: | Flags with carried forward coverage won't be shown. [Click here](https://docs.codecov.io/docs/carryforward-flags?utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Andrew+Logsdail#carryforward-flags-in-the-pull-request-comment) to find out more.

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[GabrielBram]

Look through the MACE ASE calculator, they unfortunately import torch at the head so I don't believe there's a way around this. Would this require Andy to update the CI test runner?

[ikowalec]

Look through the MACE ASE calculator, they unfortunately import torch at the head so I don't believe there's a way around this. Would this require Andy to update the CI test runner?

We would have to cache the pytorch installation somehow - we can't just download and rebuild 2.5gb of packages on each pull request. Something to think about.

Edit: Apparently there is a cached version of pytorch, so dependancies are building only ~60 sec longer with torch, torchvision and mace in the CI test

[ikowalec]

Look through the MACE ASE calculator, they unfortunately import torch at the head so I don't believe there's a way around this. Would this require Andy to update the CI test runner?

now resolved by amending the main.yml