allow elastically unstable structures

[matthewkuner]

Summary

Introduce a task_document_kwarg to the ElasticDocument that allows the elastic workflow to still complete in the event that the structure is elastically unstable. This specifically affects the DerivedProperties field in the ElasticDocument, which will be set to none in the event that the structure is elastically unstable.

@utf I made it so that elastically unstable structures are, by default, allowed to complete. Let me know if you think this should be changed.

[codecov]

Codecov Report

Merging #355 (0861f26) into main (2f54128) will decrease coverage by 0.15%.
The diff coverage is 100.00%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #355      +/-   ##
==========================================
- Coverage   64.74%   64.60%   -0.15%     
==========================================
  Files          74       74              
  Lines        7157     7159       +2     
  Branches      938      938              
==========================================
- Hits         4634     4625       -9     
- Misses       2224     2236      +12     
+ Partials      299      298       -1     

Impacted Files	Coverage Δ
src/atomate2/vasp/jobs/elastic.py	78.31% <ø> (ø)
src/atomate2/common/schemas/elastic.py	88.65% <100.00%> (+0.11%)	⬆️
src/atomate2/vasp/flows/elastic.py	82.50% <100.00%> (+0.44%)	⬆️

... and 1 file with indirect coverage changes

[mkhorton]

Can you think of a way of adding logic here that might distinguish between truly dynamically unstable materials and materials for which the fit is just poor (e.g., has gone into a non-linear regime)? It seems important to distinguish these possibilities.

Separately, how is "elastic instability" determined -- negative eigenvalues of Voigt tensor, or some other method?

[matthewkuner]

@mkhorton I was just looking to allow workflows to complete when the bulk/shear modulus are negative (which indicates elastic instability). Honestly, I'm not familiar enough with the fine details of elastic constant fitting methods to write logic for the fitting method not being applicable / identifying dynamic instability

[mkhorton]

That makes sense, perfectly valid!

[matthewkuner]

@mkhorton There is someone in Mark Asta's group, Pedro, who is more familiar with the math underlying elastic constant fitting/calculations. I will see if he has any ideas about automatically determining if the fitting method is not applicable. That probably belongs in another PR though

[matthewkuner]

@utf let me know if this is good to go or if you'd like any changes made

[utf]

Thanks @matthewkuner. Thanks for this.