This PR implements a Node for classical sampling of configurations using https://github.com/alanwilter/acpype and the GAFF as well as PACKMOL and RDKit for generating boxes from smiles.
I've added some trajectory conversion utilities. The energy reading code works in principle but I am fairly certain that the energies I obtain are wrong in some way. There is also an issue that the gmx energy command works on the CLI but not from within the node so I have opted to set the energies to 0 for now.
TODO
[x] remove maxwarn, currently identifiers don't seem to be unique
[x] box size
[x] packmol pbc issue
[ ] make energy / force saving to ASE atoms optional
[ ] configure gromacs mpi/omp parameters for better performance
[ ] clean up files? When converting to ASE we have a large duplication w.r.t to the gromacs output files. Do we want to keep them?
PythonFZ
commented
1 month ago
This PR implements a Node for classical sampling of configurations using https://github.com/alanwilter/acpype and the GAFF as well as PACKMOL and RDKit for generating boxes from smiles.
I've added some trajectory conversion utilities. The energy reading code works in principle but I am fairly certain that the energies I obtain are wrong in some way. There is also an issue that the gmx energy command works on the CLI but not from within the node so I have opted to set the energies to 0 for now.
TODO
maxwarn, currently identifiers don't seem to be unique