Any implement plan for more constraints?

[alphalm4]

Dear torchsim devs,

is there any plan for implementation of more constraints (like here)?
For example, dihedral angle constraint during atomic position relaxation is extremely useful when benchmarking molecular torsion barrier.

[CompRhys]

We're very open to receiving additional constraints PRs. I personally am unlikely to find scope to implement these and I am not sure about other maintainers. For me the big near term goals are 1) Finish the push to externalize models, 2) Settle the extensible attributes API, 3) fire-fight any issues in the externalized model interfaces caused by 2. I think once those land we would be open to a discussion about what the next steps on the technical wishlist might be and whether any maintainers have scope to implement those features.