search

compSPI / simSPI

Generative Models
9 stars 10 forks source link

Replace mdtraj with gemmi #22

Open fredericpoitevin opened 2 years ago

fredericpoitevin commented 2 years ago

I was very much into MDtraj when I wrote the material you used to make this function, but gemmi might be a better choice, since MDtraj is no longer maintained I think.

Originally posted by @fredericpoitevin in https://github.com/compSPI/simSPI/pull/19#r755671560

geoffwoollard commented 2 years ago

When you tackle this, I think it would be good to also come up with some solution for "atom selection". Maybe tem_simulator should just use the atoms it gets in the pdb file, and the atom selection should happen upstream in some other tool (gemmi).

We want to know exactly what pdb file the tem_simulator used. It would be confusing if there's an all atom pdb file associated with the particle stack (in the bundle of particle stack, meta data, and input structure), but the tem_simulator only used only some of the atoms.

geoffwoollard commented 2 years ago

https://github.com/compSPI/simSPI/pull/19/files#r764198998

fredericpoitevin commented 2 years ago

Both MDtraj and gemmi have rather well document atom selection API: