Matt Thompson
Matt Thompson
**Description** The supported thermodynamic outputs in the LAMMPS driver are [hardcoded](https://github.com/openforcefield/openff-interchange/blob/35ac30006146b9b5eb3e344660b84ffa9933f2fb/openff/interchange/components/mdconfig.py#L265). I think these cover the important bits, but it might be useful to expose these options to users.
For historical reasons, all virtual sites are created using `openmm.LocalCoordinatesSite`, which does a lot of fancy math to handle all virtual sites with one object. Unfortunately, this comes at the...
See the tail end of this comment for some good suggestions: https://github.com/openforcefield/openff-interchange/pull/756#issuecomment-1612478743 These changes would need to be pushed to the `central-examples` branch to go live until a release is...
Blockers: * No AmberTools packages * ~~No OpenEye Toolkits packages~~
OpenMM conventionally groups proper and improper torsions into the same force, from which I don't know how to split contributions from each. ```python In [23]: torsion_force Out[23]: In [24]: [val...
**Description** Most of the time exporting large systems is spent dealing with units. This looks promising: https://pint.readthedocs.io/en/stable/advanced/performance.html
Evaluator needs this to evaluate gradients with non-LJ potentials.
While nearing a 0.1.0 release, I am taking some time to look at import times. The first pass of checks is, quite frankly, awful. The quickest way I know how...
There's some usage in [this](https://github.com/openforcefield/openff-interchange/blob/v0.3.1/examples/protein_ligand/protein_ligand.ipynb) and other examples but it might be useful to have something standalone that's focused on the features of just this module.
The current design of `GROMACSSystem` nearly allows for different physics to be swapped in just by re-assigning one of its `.molecule_types`: ```python3 complex_system: GROMACSSystem = from_files("complex.top", "complex.gro") complex_system["LIG"] = reparametrized_ligand...