-
Notifications
You must be signed in to change notification settings - Fork 17
Open
Labels
enhancementNew/improved feature or requestNew/improved feature or request
Description
Summary
This may be possible some other way but it was not obvious to me (write_kwargs that goes to ase.io?) but for example MACE has the node_energy result which is the list of predicted per-atom energies. This would be very useful for thermal-conductivity and more analysis.
This is stored in atoms.calc.results['node_energy']) after a single_point.
from janus_core.calculations.single_point import SinglePoint
from ase.io import read
atoms = read("LiF.extxyz")
sp = SinglePoint(
struct=atoms,
arch="mace",
write_results=True
)
sp.run()
print(atoms.calc.results['node_energy'])
# array([-3.04302047, -5.38119082, -2.85913875, -5.24258456, -2.88435103, ...
# The predicted "Per-atom" energies, useful for thermal-conductivity.At the moment though in the written results here we just have these properties,
Properties=species:S:1:pos:R:3:momenta:R:3:masses:R:1:mace_mp_forces:R:3:mace_forces:R:3:forces:R:3
But If you do,
atoms.set_array('atom_energy', atoms.calc.results['node_energy'])
write("LiF-sp.extxyz", atoms)Then atom_energy is in the properties
Properties=species:S:1:pos:R:3:momenta:R:3:masses:R:1:mace_mp_forces:R:3:mace_forces:R:3:atom_energy:R:1:forces:R:3
Example
User controlled properties like so (atom_energy),
Properties=species:S:1:pos:R:3:momenta:R:3:masses:R:1:mace_mp_forces:R:3:mace_forces:R:3:atom_energy:R:1:forces:R:3
Alternatives
No response
Additional context
No response
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
enhancementNew/improved feature or requestNew/improved feature or request