from typing import List, Union
import numpy as np
from _icet import _ClusterExpansionCalculator
from icet.input_output.logging_tools import logger
from ase import Atoms
from icet import ClusterExpansion
from icet.core.structure import Structure
from icet.core.sublattices import Sublattices
from mchammer.calculators.base_calculator import BaseCalculator
[docs]class ClusterExpansionCalculator(BaseCalculator):
"""A ClusterExpansionCalculator object enables the efficient
calculation of properties described by a cluster expansion. It is
specific for a particular (supercell) structure and commonly
employed when setting up a Monte Carlo simulation, see
:ref:`ensembles`.
Cluster expansions, e.g., of the energy, typically yield property
values *per site*. When running a Monte Carlo simulation one,
however, considers changes in the *total* energy of the
system. The default behavior is therefore to multiply the output
of the cluster expansion by the number of sites. This behavior can
be changed via the ``scaling`` keyword parameter.
Parameters
----------
structure : ase.Atoms
structure for which to set up the calculator
cluster_expansion : ClusterExpansion
cluster expansion from which to build calculator
name
human-readable identifier for this calculator
scaling
scaling factor applied to the property value predicted by the
cluster expansion
use_local_energy_calculator
evaluate energy changes using only the local environment; this method
is generally *much* faster; unless you know what you are doing do *not*
set this option to `False`
"""
def __init__(self,
structure: Atoms, cluster_expansion: ClusterExpansion,
name: str = 'Cluster Expansion Calculator',
scaling: Union[float, int] = None,
use_local_energy_calculator: bool = True) -> None:
super().__init__(name=name)
structure_cpy = structure.copy()
cluster_expansion.prune()
if cluster_expansion._cluster_space.is_supercell_self_interacting(structure):
logger.warning('The ClusterExpansionCalculator self-interacts, '
'which may lead to erroneous results. To avoid '
'self-interaction, use a larger supercell or a '
'cluster space with shorter cutoffs.')
self.use_local_energy_calculator = use_local_energy_calculator
self.cpp_calc = _ClusterExpansionCalculator(
cluster_space=cluster_expansion.get_cluster_space_copy(),
structure=Structure.from_atoms(structure_cpy),
fractional_position_tolerance=cluster_expansion.fractional_position_tolerance)
self._cluster_expansion = cluster_expansion
if scaling is None:
self._property_scaling = len(structure)
else:
self._property_scaling = scaling
self._sublattices = self.cluster_expansion._cluster_space.get_sublattices(structure)
@property
def cluster_expansion(self) -> ClusterExpansion:
""" cluster expansion from which calculator was constructed """
return self._cluster_expansion
[docs] def calculate_total(self, *, occupations: List[int]) -> float:
"""
Calculates and returns the total property value of the current
configuration.
Parameters
----------
occupations
the entire occupation vector (i.e. list of atomic species)
"""
cv = self.cpp_calc.get_cluster_vector(occupations)
return np.dot(cv, self.cluster_expansion.parameters) * self._property_scaling
[docs] def calculate_change(self, *, sites: List[int],
current_occupations: List[int],
new_site_occupations: List[int]) -> float:
"""
Calculates and returns the sum of the contributions to the property
due to the sites specified in `local_indices`
Parameters
----------
sites
index of sites at which occupations will be changed
current_occupations
entire occupation vector (atomic numbers) before change
new_site_occupations
atomic numbers after change at the sites defined by `sites`
"""
occupations = np.array(current_occupations)
if not self.use_local_energy_calculator:
e_before = self.calculate_total(occupations=occupations)
occupations[sites] = np.array(new_site_occupations)
e_after = self.calculate_total(occupations=occupations)
return e_after - e_before
change = 0.0
try:
for index, new_occupation in zip(sites, new_site_occupations):
change += self._calculate_partial_change(occupations=occupations,
flip_index=index,
new_occupation=new_occupation)
occupations[index] = new_occupation # Safe because we work with a copy
except Exception as e:
msg = 'Caught exception {}. Try setting parameter '.format(e)
msg += 'use_local_energy_calculator to False in init'
raise RuntimeError(msg)
return change * self._property_scaling
def _calculate_partial_change(self, occupations: List[int], flip_index: int,
new_occupation: int):
"""
Internal method to calculate the local contribution for one
index.
Parameters
----------
occupations
entire occupation vector
flip_index
lattice index for site where a flipped has occurred
new_occupation
atomic number of new occupation at site `flip_index`
"""
cv_change = self.cpp_calc.get_cluster_vector_change(occupations=occupations,
flip_index=flip_index,
new_occupation=new_occupation)
return np.dot(cv_change, self.cluster_expansion.parameters)
@property
def sublattices(self) -> Sublattices:
"""Sublattices of the calculators structure."""
return self._sublattices