RAFFLE

RAFFLE (pseudoRandom Approach For Finding Local Energetic minima) is a Python and Fortran package for structure prediction applied to interfaces. RAFFLE can be utilised as a Python package, a Fortran library, or a standalone Fortran executable. The Python package provides a high-level interface to the Fortran library, which contains the core functionality.

The Python package interfaces seemlessly with ASE (Atomic Simulation Environment), allowing for easy reading, writing, and manipulation of atomic structures. Although the package comes with a built-in atomic structure reader and writer, it is recommended to use ASE due to its greater functionality and wide-reaching support.

The code is provided freely available under the GNU General Public License v3.0.

An example

# A simple example of how to use RAFFLE to generate 10 structures of diamond and write them to a single file
from ase import Atoms
from ase.io import write
from ase.calculators.singlepoint import SinglePointCalculator
from raffle.generator import raffle_generator
from mace.calculators import mace_mp

generator = raffle_generator()
generator.distributions.set_history_len(10)
calc = mace_mp(model="medium", dispersion=False, default_dtype="float32", device='cpu')

host = Atoms('C', positions=[[0, 0, 0]], cell=[10, 10, 10])
host.calc = calc
generator.set_host(host)

generator.distributions.set_element_energies( { 'C': 0.0 } )
generator.distributions.create(host)

num_structures_old = 0
for i in range(10):
    structures = generator.generate(
        num_structures = 2,
        stoichiometry = { 'C': 7 }
    )
    for structure in structures:
        optimiser = FIRE(structure)
        optimiser.run(fmax=0.05)

    generator.distributions.update(structures)
    num_structures_old += len(structures)
    if generator.distributions.is_converged():
        break

structures = generator.get_structures(calc)
for structure in structures:
    structure.calc = SinglePointCalculator(
        structure,
        energy=structure.get_potential_energy(),
        forces=structure.get_forces()
    )

write('structures.traj', structures)

Contents: