2021-10-27 Thompson/Wagner Check in

General updates

Biopolymer functionality walkthrough

(JW and MT worked on loading+running a PLBenchmarks protein-ligand complex, but runtime was too slow for the large protein use. MT showed an example of loading MainChain_ALA and _ARG into a single topology, assigning parameters and getting equivalent energies out in AMBER, GROMACS, OpenMM, and LAMMPS)

Reducing number of toolkit tests

• MT – Too many tests

  • QC tests are failing

  • Many tests just loop over large datasets

  • GBSA energy tests are of questionable use

  • Status repo shows lots of stuff broken

  • Pint had a release with breaking changes for evaluator

    • Opened

De-ParmEd-ing examples

• MT – I’m working on a PR for this.

• JW – Do we want the examples to be a mix of ParmEd and Interchange, or should we do the whole switchover at once?

• MT – I’d like to minimize the amount of ParmEd overall.

• JW – Would like to be careful here about what we’re advertising and its suitability for use. Interchange is suitable for small molecules, but I expect that there’s some bugs/edge cases waiting to be found.

• MT – There can be microrepos or monorepos. OpenFF is, in practice, more of a monorepo. People googling openforcefield will be directed to the toolkit. So if we don’t have interchange examples there then many users won’t find them at all.

• (MT and JW discussed, here are steps forward)