Gamma surface & stacking fault tools

[thomas-rocke]

Classes to generate stacking fault & gamma surface (generalised stacking fault) images, plus utils for managing relaxation, evaluation, & plotting.

[pgrigorev]

Hi @thomas-rocke, this looks like a very good idea and nice work. Gamma surface and stacking faults are strongly related to dislocation properties and this is why I had an implementation of a gamma_line function. The idea was that you can pass the unit_cell from the dislocation class to this function and use it to provide related gamma line and stacking fault energy. I see that you added a deprecation to the function in dislocation module. I guess your implementation is more general and I will be happy to use it in the future. It would be great if we could discuss how I can use your new implementation instead of one in the dislocation module. Also we should think about a tutorial example ideally closely related to the dislocation examples. For example how to use this module to get stacking faults for 1/2<111>(110) dislocation and/or other dislocations in BCC. Let me know what you think?

[thomas-rocke]

Hi @pgrigorev , I see your point about dislocations and gamma surfaces being very closely related - there's definitely an argument in trying to modify my solution to more directly coexist with/around the dislocation tools, and I have a few ideas on how we could achieve this. I can also make a reasonable tutorial on utilising the code, once we've settled on the exact final solution. I'm happy to meet to discuss this at some point, maybe also including @jameskermode ?

[pgrigorev]

Sure I will be happy to discuss together. Also to understand better what you did. For the tutoral I was thinking to add a section to plasticity documentation about calculation of stacking fault energies relevant for different dislocations. You would just to need to create a Jupiter notebook with and it will be automatically rendered for the documentation.

[thomas-rocke]

Successfully added the ability to generate stacking faults from dissociated dislocations (e.g. StackingFault(alat, DiamondGlideScrew, symbol="Si")) and get correct structures from CNA. Docs are a WIP, and I think I need more testing of the results before this is ready for release.

[thomas-rocke]

@jameskermode @pgrigorev New docs are done, and I think the features are also finished (I did relax the dislocation constraint so any CubicCrystalDislocation is fine). If I'm being critical, I should maybe try to have a few more unit tests, but I'm unsure what the best practice should be - I don't think it's useful to have a bunch of individual tests for very basic things, if they are just going to get picked up by the energy fits anyway.

Assuming there's nothing massively wrong with the docs, we should be good to try to merge this on Weds.

[jameskermode]

Approved subject to remaining small changes and @pgrigorev being happy with the docs

[pgrigorev]

There were still a few complaints left. I does not have to be perfect, but unused variables can be very confusing. Once it is fixed and everything works, I am happy with this to be merged.

Great work. Thanks a lot!

[thomas-rocke]

@jameskermode this should be finished and ready to merge. Most of the new changes are just small QOL fixes.

One new feature of note has been added. @pgrigorev had spotted that gamma_surface.ipynb was silently failing in the CI Docs build due to timeout. He extended the timeout or that notebook, and I've added the nb_execution_raise_on_error=True arg to the sphinx conf.py, which should make the CI fail if there are any future issues with notebooks. It definitely works locally with a notebook I knew would fail.

[pgrigorev]

@jameskermode this should be finished and ready to merge. Most of the new changes are just small QOL fixes.

One new feature of note has been added. @pgrigorev had spotted that gamma_surface.ipynb was silently failing in the CI Docs build due to timeout. He extended the timeout or that notebook, and I've added the nb_execution_raise_on_error=True arg to the sphinx conf.py, which should make the CI fail if there are any future issues with notebooks. It definitely works locally with a notebook I knew would fail.

I did not know about this flag, this is great that you found it. I think it is in shape to be merged. Fantastic work!

[jameskermode]

thanks @thomas-rocke and also to @pgrigorev for reviewing!