scnc
/
phono3py
mirror of https://github.com/phonopy/phono3py.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
phono3py/phono3py/api_phono3py.py

2762 lines
106 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

"""Phono3py main class."""
# Copyright (C) 2016 Atsushi Togo
# All rights reserved.
#
# This file is part of phono3py.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
#
# * Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
#
# * Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in
# the documentation and/or other materials provided with the
# distribution.
#
# * Neither the name of the phonopy project nor the names of its
# contributors may be used to endorse or promote products derived
# from this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
# FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
# COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
# BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
# ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
# POSSIBILITY OF SUCH DAMAGE.
from __future__ import annotations
import copy
import warnings
from collections.abc import Sequence
from typing import Literal, Optional, Union, cast
import numpy as np
from numpy.typing import NDArray
from phonopy.harmonic.displacement import (
directions_to_displacement_dataset,
get_least_displacements,
get_random_displacements_dataset,
)
from phonopy.harmonic.dynamical_matrix import DynamicalMatrix
from phonopy.harmonic.force_constants import (
set_permutation_symmetry,
set_translational_invariance,
symmetrize_compact_force_constants,
symmetrize_force_constants,
)
from phonopy.interface.fc_calculator import get_fc2
from phonopy.interface.mlp import PhonopyMLP
from phonopy.interface.pypolymlp import (
PypolymlpParams,
)
from phonopy.interface.symfc import SymfcFCSolver
from phonopy.physical_units import get_physical_units
from phonopy.structure.atoms import PhonopyAtoms
from phonopy.structure.cells import (
Primitive,
Supercell,
get_primitive,
get_primitive_matrix,
get_supercell,
guess_primitive_matrix,
shape_supercell_matrix,
)
from phonopy.structure.symmetry import Symmetry
from phono3py.conductivity.init_direct_solution import get_thermal_conductivity_LBTE
from phono3py.conductivity.init_rta import get_thermal_conductivity_RTA
from phono3py.interface.fc_calculator import FC3Solver
from phono3py.interface.phono3py_yaml import Phono3pyYaml
from phono3py.phonon.grid import BZGrid
from phono3py.phonon3.dataset import forces_in_dataset
from phono3py.phonon3.displacement_fc3 import (
direction_to_displacement,
get_third_order_displacements,
)
from phono3py.phonon3.fc3 import (
cutoff_fc3_by_zero,
set_permutation_symmetry_compact_fc3,
set_permutation_symmetry_fc3,
set_translational_invariance_compact_fc3,
set_translational_invariance_fc3,
)
from phono3py.phonon3.imag_self_energy import (
get_imag_self_energy,
write_imag_self_energy,
)
from phono3py.phonon3.interaction import Interaction
from phono3py.phonon3.real_self_energy import (
get_real_self_energy,
write_real_self_energy,
)
from phono3py.phonon3.spectral_function import run_spectral_function
from phono3py.version import __version__
class Phono3py:
"""Phono3py main class.
Attributes
----------
version
calculator
fc3 : getter and setter
fc2 : getter and setter
force_constants
sigma : getter and setter
sigma_cutoff : getter and setter
nac_params : getter and setter
dynamical_matrix
primitive
unitcell
supercell
phonon_supercell
phonon_primitive
symmetry
primitive_symmetry
phonon_supercell_symmetry
supercell_matrix
phonon_supercell_matrix
primitive_matrix
unit_conversion_factor
dataset : getter and setter
phonon_dataset : getter and setter
band_indices : getter and setter
phonon_supercells_with_displacements
supercells_with_displacements
mesh_numbers : getter and setter
thermal_conductivity
displacements : getter and setter
forces : getter and setter
phonon_displacements : getter and setter
phonon_forces : getter and setter
phph_interaction
"""
def __init__(
self,
unitcell: PhonopyAtoms,
supercell_matrix=None,
primitive_matrix=None,
phonon_supercell_matrix=None,
cutoff_frequency=1e-4,
frequency_factor_to_THz=None,
is_symmetry=True,
is_mesh_symmetry=True,
use_grg=False,
SNF_coordinates="reciprocal",
make_r0_average: bool = True,
symprec=1e-5,
calculator: Optional[str] = None,
log_level=0,
):
"""Init method.
Parameters
----------
unitcell : PhonopyAtoms, optional
Input unit cell.
supercell_matrix : array_like, optional
Supercell matrix multiplied to input cell basis vectors. shape=(3, )
or (3, 3), where the former is considered a diagonal matrix. The
elements have to be given by integers. Although the default is None,
which results in identity matrix, it is recommended to give
`supercell_matrix` explicitly.
primitive_matrix : array_like or str, optional
Primitive matrix multiplied to input cell basis vectors. Default is
the identity matrix. When given as array_like, shape=(3, 3),
dtype=float. When 'F', 'I', 'A', 'C', or 'R' is given instead of a
3x3 matrix, the primitive matrix defined at
https://spglib.github.io/spglib/definition.html is used. When 'auto'
is given, the centring type ('F', 'I', 'A', 'C', 'R', or primitive
'P') is automatically chosen.
phonon_supercell_matrix : array_like, optional
Supercell matrix used for fc2. In phono3py, supercell matrix for fc3
and fc2 can be different to support longer range interaction of fc2
than that of fc3. Unless setting this, supercell_matrix is used.
This is only valid when unitcell or unitcell_filename is given.
Default is None.
cutoff_frequency : float, optional
Phonon frequency below this value is ignored when the cutoff is
needed for the computation. Default is 1e-4.
frequency_factor_to_THz : float, optional
Phonon frequency unit conversion factor. Unless specified, default
unit conversion factor for each calculator is used.
is_symmetry : bool, optional
Use crystal symmetry in most calculations when True. Default is
True.
is_mesh_symmetry : bool, optional
Use crystal symmetry in reciprocal space grid handling when True.
Default is True.
use_grg : bool, optional
Use generalized regular grid when True. Default is False.
SNF_coordinates : str, optional
`reciprocal` or `direct`. Space of coordinates to generate grid
generating matrix either in direct or reciprocal space. The default
is `reciprocal`.
make_r0_average : bool, optional
fc3 transformation from real to reciprocal space is done
around three atoms and averaged when True. Default is False, i.e.,
only around the first atom. Setting False is for rough compatibility
with v2.x. Default is True.
symprec : float, optional
Tolerance used to find crystal symmetry. Default is 1e-5.
calculator : str, optional.
Calculator used for computing forces. This is used to switch the set
of physical units. Default is None, which is equivalent to "vasp".
log_level : int, optional
Verbosity control. Default is 0. This can be 0, 1, or 2.
"""
self._symprec = symprec
if frequency_factor_to_THz is None:
self._frequency_factor_to_THz = get_physical_units().DefaultToTHz
else:
self._frequency_factor_to_THz = frequency_factor_to_THz
self._is_symmetry = is_symmetry
self._is_mesh_symmetry = is_mesh_symmetry
self._use_grg = use_grg
self._SNF_coordinates = SNF_coordinates
self._make_r0_average = make_r0_average
self._cutoff_frequency = cutoff_frequency
self._calculator: Optional[str] = calculator
self._log_level = log_level
# Create supercell and primitive cell
self._unitcell = unitcell
self._supercell_matrix = np.array(
shape_supercell_matrix(supercell_matrix), dtype="int64", order="C"
)
pmat = self._determine_primitive_matrix(primitive_matrix)
self._primitive_matrix = pmat
self._nac_params = None
if phonon_supercell_matrix is not None:
self._phonon_supercell_matrix = np.array(
shape_supercell_matrix(phonon_supercell_matrix),
dtype="int64",
order="C",
)
else:
self._phonon_supercell_matrix = None
self._supercell: Supercell
self._primitive: Primitive
self._phonon_supercell: Supercell
self._phonon_primitive: Primitive
self._build_supercell()
self._build_primitive_cell()
self._build_phonon_supercell()
self._build_phonon_primitive_cell()
self._sigmas = [
None,
]
self._sigma_cutoff = None
# Grid
self._bz_grid = None
# Set supercell, primitive, and phonon supercell symmetries
self._symmetry: Symmetry
self._primitive_symmetry: Symmetry
self._phonon_supercell_symmetry: Symmetry
self._search_symmetry()
self._search_primitive_symmetry()
self._search_phonon_supercell_symmetry()
# Displacements and supercells
self._supercells_with_displacements = None
self._dataset = None
self._phonon_dataset = None
self._phonon_supercells_with_displacements = None
# Thermal conductivity
# conductivity_RTA or conductivity_LBTE class instance
self._thermal_conductivity = None
# Imaginary part of self energy at frequency points
self._gammas = None
self._scattering_event_class = None
# Frequency shift (real part of bubble diagram)
self._real_self_energy = None
self._grid_points = None
self._frequency_points = None
self._temperatures = None
# Force constants
self._fc2 = None
self._fc3 = None
self._fc3_nonzero_indices = None # available only symfc
# MLP
self._mlp = None
self._mlp_dataset = None
self._phonon_mlp = None
self._phonon_mlp_dataset = None
# Setup interaction
self._interaction = None
self._band_indices = None
self._band_indices_flatten = None
self._set_band_indices()
@property
def version(self) -> str:
"""Return phono3py release version number.
str
Phono3py release version number
"""
return __version__
@property
def calculator(self) -> Optional[str]:
"""Return calculator interface name.
str
Calculator name such as 'vasp', 'qe', etc.
"""
return self._calculator
@property
def fc3(self) -> NDArray | None:
"""Setter and getter of third order force constants (fc3).
ndarray, optional
fc3 shape is either (supercell, supercell, supercell, 3, 3, 3) or
(primitive, supercell, supercell, 3, 3, 3),
where 'supercell' and 'primitive' indicate number of atoms in
these cells.
"""
return self._fc3
@fc3.setter
def fc3(self, fc3):
self._fc3 = fc3
@property
def fc3_nonzero_indices(self) -> NDArray | None:
"""Setter and getter of non-zero indices of fc3.
ndarray, optional
Non-zero indices of fc3.
"""
return self._fc3_nonzero_indices
@fc3_nonzero_indices.setter
def fc3_nonzero_indices(self, fc3_nonzero_indices):
self._fc3_nonzero_indices = fc3_nonzero_indices
@property
def fc2(self) -> NDArray | None:
"""Setter and getter of second order force constants (fc2).
ndarray
fc2 shape is either (supercell, supercell, 3, 3) or
(primitive, supercell, 3, 3),
where 'supercell' and 'primitive' indicate number of atoms in
these cells.
"""
return self._fc2
@fc2.setter
def fc2(self, fc2):
self._fc2 = fc2
@property
def force_constants(self) -> NDArray | None:
"""Return fc2. This is same as the getter attribute `fc2`."""
return self.fc2
@property
def sigmas(self) -> list:
"""Setter and getter of smearing widths.
list
The float values are given as the standard deviations of Gaussian
function. If None is given as an element of this list, linear
tetrahedron method is used instead of smearing method.
"""
return self._sigmas
@sigmas.setter
def sigmas(self, sigmas):
if sigmas is None:
self._sigmas = [
None,
]
elif isinstance(sigmas, float) or isinstance(sigmas, int):
self._sigmas = [
float(sigmas),
]
else:
self._sigmas = []
for s in sigmas:
if isinstance(s, float) or isinstance(s, int):
self._sigmas.append(float(s))
elif s is None:
self._sigmas.append(None)
@property
def sigma_cutoff(self) -> Optional[float]:
"""Setter and getter of Smearing cutoff width.
This is given as a multiple of the standard deviation.
float
For example, if this value is 5, the tail of the Gaussian function
is cut at 5 sigma.
"""
return self._sigma_cutoff
@sigma_cutoff.setter
def sigma_cutoff(self, sigma_cutoff):
self._sigma_cutoff = sigma_cutoff
@property
def nac_params(self) -> Optional[dict]:
"""Setter and getter of parameters for non-analytical term correction.
dict
Parameters used for non-analytical term correction
'born': ndarray
Born effective charges
shape=(primitive cell atoms, 3, 3), dtype='double', order='C'
'factor': float
Unit conversion factor
'dielectric': ndarray
Dielectric constant tensor
shape=(3, 3), dtype='double', order='C'
"""
return self._nac_params
@nac_params.setter
def nac_params(self, nac_params):
self._nac_params = nac_params
if self._interaction is not None:
self._init_dynamical_matrix()
@property
def dynamical_matrix(self) -> Optional[DynamicalMatrix]:
"""Return DynamicalMatrix instance.
This is not dynamical matrices but the instance of DynamicalMatrix
class.
"""
if self._interaction is None:
return None
else:
return self._interaction.dynamical_matrix
@property
def primitive(self) -> Primitive:
"""Return primitive cell.
Primitive
Primitive cell.
"""
return self._primitive
@property
def unitcell(self) -> PhonopyAtoms:
"""Return Unit cell.
PhonopyAtoms
Unit cell.
"""
return self._unitcell
@property
def supercell(self) -> Supercell:
"""Return supercell.
Supercell
Supercell.
"""
return self._supercell
@property
def phonon_supercell(self) -> Supercell:
"""Return supercell for fc2.
Supercell
Supercell for fc2.
"""
return self._phonon_supercell
@property
def phonon_primitive(self) -> Primitive:
"""Return primitive cell for fc2.
Primitive
Primitive cell for fc2. This should be the same as the primitive
cell for fc3, but this is created from supercell for fc2 and
can be not numerically perfectly identical.
"""
return self._phonon_primitive
@property
def symmetry(self) -> Symmetry:
"""Return symmetry of supercell.
Symmetry
Symmetry of supercell
"""
return self._symmetry
@property
def primitive_symmetry(self) -> Symmetry:
"""Return symmetry of primitive cell.
Symmetry
Symmetry of primitive cell.
"""
return self._primitive_symmetry
@property
def phonon_supercell_symmetry(self) -> Symmetry:
"""Return symmetry of supercell for fc2.
Symmetry
Symmetry of supercell for fc2 (phonon_supercell).
"""
return self._phonon_supercell_symmetry
@property
def supercell_matrix(self) -> NDArray:
"""Return transformation matrix to supercell cell from unit cell.
ndarray
Supercell matrix with respect to unit cell.
shape=(3, 3), dtype='int64', order='C'
"""
return self._supercell_matrix
@property
def phonon_supercell_matrix(self) -> NDArray | None:
"""Return transformation matrix to phonon supercell from unit cell.
ndarray
Supercell matrix with respect to unit cell.
shape=(3, 3), dtype='int64', order='C'
"""
return self._phonon_supercell_matrix
@property
def primitive_matrix(self) -> NDArray:
"""Return transformation matrix to primitive cell from unit cell.
ndarray
Primitive matrix with respect to unit cell.
shape=(3, 3), dtype='double', order='C'
"""
return self._primitive_matrix
@property
def unit_conversion_factor(self) -> float:
"""Return phonon frequency unit conversion factor.
float
Phonon frequency unit conversion factor. This factor
converts sqrt(<force>/<distance>/<AMU>)/2pi/1e12 to THz
(ordinary frequency).
"""
return self._frequency_factor_to_THz
@property
def dataset(self) -> Optional[dict]:
"""Setter and getter of displacement-force dataset.
dict
Displacements in supercells. There are two types of formats.
Type 1. Two atomic displacement in each supercell:
{'natom': number of atoms in supercell,
'first_atoms': [
{'number': atom index of first displaced atom,
'displacement': displacement in Cartesian coordinates,
'forces': forces on atoms in supercell,
'id': displacement id (1, 2,...,n_first_atoms)
'second_atoms': [
{'number': atom index of second displaced atom,
'displacement': displacement in Cartesian coordinates},
'forces': forces on atoms in supercell,
'supercell_energy': energy of supercell,
'pair_distance': distance between paired atoms,
'included': with cutoff pair distance in displacement
pair generation, this indicates if this
pair displacements is included to compute
fc3 or not,
'id': displacement id. (n_first_atoms + 1, ...)
... ] }, ... ] }
Type 2. All atomic displacements in each supercell:
{'displacements': ndarray, dtype='double', order='C',
shape=(supercells, atoms in supercell, 3),
'forces': ndarray, dtype='double',, order='C',
shape=(supercells, atoms in supercell, 3),
'supercell_energies': ndarray, dtype='double',
shape=(supercells,)}
In type 2, displacements and forces can be given by numpy array
with different shape but that can be reshaped to
(supercells, natom, 3).
In addition, 'duplicates' and 'cutoff_distance' can exist in this
dataset in displacement pair generation. 'duplicates' gives
duplicated supercell ids as pairs.
"""
return self._dataset
@dataset.setter
def dataset(self, dataset):
if dataset is None:
self._dataset = None
elif "first_atoms" in dataset:
self._dataset = copy.deepcopy(dataset)
elif "displacements" in dataset:
self._dataset = {}
self.displacements = dataset["displacements"]
if "forces" in dataset:
self.forces = dataset["forces"]
if "supercell_energies" in dataset:
self.supercell_energies = dataset["supercell_energies"]
else:
raise RuntimeError("Data format of dataset is wrong.")
self._supercells_with_displacements = None
self._phonon_supercells_with_displacements = None
@property
def phonon_dataset(self) -> Optional[dict]:
"""Setter and getter of displacement-force dataset for fc2.
dict
Displacements in supercells. There are two types of formats.
Type 1. Two atomic displacement in each supercell:
{'natom': number of atoms in supercell,
'first_atoms': [
{'number': atom index of first displaced atom,
'displacement': displacement in Cartesian coordinates,
'forces': forces on atoms in supercell,
'supercell_energy': energy of supercell}, ... ]}
Type 2. All atomic displacements in each supercell:
{'displacements': ndarray, dtype='double', order='C',
shape=(supercells, atoms in supercell, 3),
'forces': ndarray, dtype='double',, order='C',
shape=(supercells, atoms in supercell, 3),
'supercell_energies': ndarray, dtype='double',
shape=(supercells,)}
In type 2, displacements and forces can be given by numpy array
with different shape but that can be reshaped to
(supercells, natom, 3).
"""
return self._phonon_dataset
@phonon_dataset.setter
def phonon_dataset(self, dataset):
if dataset is None:
self._phonon_dataset = None
elif "first_atoms" in dataset:
self._phonon_dataset = copy.deepcopy(dataset)
elif "displacements" in dataset:
self._phonon_dataset = {}
self.phonon_displacements = dataset["displacements"]
if "forces" in dataset:
self.phonon_forces = dataset["forces"]
if "supercell_energies" in dataset:
self.phonon_supercell_energies = dataset["supercell_energies"]
else:
raise RuntimeError("Data format of dataset is wrong.")
self._phonon_supercells_with_displacements = None
@property
def mlp_dataset(self) -> Optional[dict]:
"""Return displacement-force dataset.
The supercell matrix is equal to that of usual displacement-force
dataset. Only type 2 format is supported. "displacements",
"forces", and "supercell_energies" should be contained.
"""
return self._mlp_dataset
@mlp_dataset.setter
def mlp_dataset(self, mlp_dataset: dict):
self._check_mlp_dataset(mlp_dataset)
self._mlp_dataset = mlp_dataset
@property
def phonon_mlp_dataset(self) -> Optional[dict]:
"""Return phonon displacement-force dataset.
The phonon supercell matrix is equal to that of usual displacement-force
dataset. Only type 2 format is supported. "displacements", "forces", and
"supercell_energies" should be contained.
"""
return self._phonon_mlp_dataset
@phonon_mlp_dataset.setter
def phonon_mlp_dataset(self, mlp_dataset: dict):
self._check_mlp_dataset(mlp_dataset)
self._phonon_mlp_dataset = mlp_dataset
@property
def mlp(self) -> Optional[PhonopyMLP]:
"""Setter and getter of PhonopyMLP dataclass."""
return self._mlp
@mlp.setter
def mlp(self, mlp) -> Optional[PhonopyMLP]:
self._mlp = mlp
@property
def band_indices(self) -> list[NDArray]:
"""Setter and getter of band indices.
list[NDArray]
List of band indices specified to select specific bands
to computer ph-ph interaction related properties.
"""
return self._band_indices
@band_indices.setter
def band_indices(self, band_indices):
self._set_band_indices(band_indices=band_indices)
def _set_band_indices(self, band_indices=None):
if band_indices is None:
num_band = len(self._primitive) * 3
self._band_indices = [np.arange(num_band, dtype="int64")]
else:
self._band_indices = [np.array(bi, dtype="int64") for bi in band_indices]
self._band_indices_flatten = np.hstack(self._band_indices).astype("int64")
@property
def masses(self) -> NDArray:
"""Setter and getter of atomic masses of primitive cell."""
return self._primitive.masses
@masses.setter
def masses(self, masses):
if masses is None:
return
p_masses = np.array(masses)
self._primitive.masses = p_masses
p2p_map = self._primitive.p2p_map
s_masses = p_masses[[p2p_map[x] for x in self._primitive.s2p_map]]
self._supercell.masses = s_masses
u2s_map = self._supercell.u2s_map
u_masses = s_masses[u2s_map]
self._unitcell.masses = u_masses
self._phonon_primitive.masses = p_masses
p2p_map = self._phonon_primitive.p2p_map
s_masses = p_masses[[p2p_map[x] for x in self._phonon_primitive.s2p_map]]
self._phonon_supercell.masses = s_masses
@property
def supercells_with_displacements(self) -> list[PhonopyAtoms]:
"""Return supercells with displacements.
list of PhonopyAtoms
Supercells with displacements generated by
Phono3py.generate_displacements.
"""
if self._supercells_with_displacements is None:
self._build_supercells_with_displacements()
return self._supercells_with_displacements
@property
def phonon_supercells_with_displacements(self):
"""Return supercells with displacements for fc2.
list of PhonopyAtoms
Supercells with displacements generated by
Phono3py.generate_displacements.
"""
if self._phonon_supercells_with_displacements is None:
if self._phonon_dataset is not None:
self._phonon_supercells_with_displacements = (
self._build_phonon_supercells_with_displacements(
self._phonon_supercell, self._phonon_dataset
)
)
return self._phonon_supercells_with_displacements
@property
def mesh_numbers(self):
"""Setter and getter of sampling mesh numbers in reciprocal space."""
if self._bz_grid is None:
return None
else:
return self._bz_grid.D_diag
@mesh_numbers.setter
def mesh_numbers(self, mesh_numbers: Union[int, float, Sequence, NDArray]):
self._set_mesh_numbers(mesh_numbers)
@property
def thermal_conductivity(self):
"""Return thermal conductivity class instance."""
return self._thermal_conductivity
@property
def displacements(self):
"""Setter and getter displacements in supercells.
There are two types of displacement dataset. See the docstring
of dataset about types 1 and 2 for the displacement dataset formats.
Displacements set returned depends on either type-1 or type-2 as
follows:
Type-1, List of list
The internal list has 4 elements such as [32, 0.01, 0.0, 0.0]].
The first element is the supercell atom index starting with 0.
The remaining three elements give the displacement in Cartesian
coordinates.
Type-2, array_like
Displacements of all atoms of all supercells in Cartesian
coordinates.
shape=(supercells, natom, 3)
dtype='double'
For setter, only type-2 dataset format is allowed.
displacements : array_like
Atomic displacements of all atoms of all supercells.
Only all displacements in each supercell case (type-2) is
supported.
shape=(supercells, natom, 3), dtype='double', order='C'
"""
dataset = self._dataset
if self._dataset is None:
raise RuntimeError("displacement dataset is not set.")
if "first_atoms" in dataset:
num_scells = len(dataset["first_atoms"])
for disp1 in dataset["first_atoms"]:
num_scells += len(disp1["second_atoms"])
displacements = np.zeros(
(num_scells, len(self._supercell), 3),
dtype="double",
order="C",
)
i = 0
for disp1 in dataset["first_atoms"]:
displacements[i, disp1["number"]] = disp1["displacement"]
i += 1
for disp1 in dataset["first_atoms"]:
for disp2 in disp1["second_atoms"]:
displacements[i, disp2["number"]] = disp2["displacement"]
i += 1
elif "displacements" in dataset:
displacements = dataset["displacements"]
else:
raise RuntimeError("displacement dataset has wrong format.")
return displacements
@displacements.setter
def displacements(self, displacements):
disps = np.array(displacements, dtype="double", order="C")
natom = len(self._supercell)
if disps.ndim != 3 or disps.shape[1:] != (natom, 3):
raise RuntimeError("Array shape of displacements is incorrect.")
if self._dataset is None:
self._dataset = {}
elif "first_atoms" in self._dataset:
raise RuntimeError("Displacements are incompatible with dataset.")
self._dataset["displacements"] = disps
self._supercells_with_displacements = None
@property
def forces(self):
"""Setter and getter of forces in displacement dataset.
A set of atomic forces in displaced supercells. The order of
displaced supercells has to match with that in displacement dataset.
shape=(displaced supercells, atoms in supercell, 3)
getter : ndarray
setter : array_like
The order of supercells used for calculating forces has to
be the same order of supercells_with_displacements.
"""
return self._get_forces_energies(target="forces")
@forces.setter
def forces(self, values):
self._set_forces_energies(values, target="forces")
@property
def supercell_energies(self):
"""Setter and getter of supercell energies in displacement dataset.
A set of supercell energies of displaced supercells. The order of
displaced supercells has to match with that in displacement dataset.
shape=(displaced supercells,)
getter : ndarray
setter : array_like
The order of supercells used for calculating supercell energies has
to be the same order of supercells_with_displacements.
"""
return self._get_forces_energies(target="supercell_energies")
@supercell_energies.setter
def supercell_energies(self, values):
self._set_forces_energies(values, target="supercell_energies")
@property
def phonon_displacements(self):
"""Setter and getter of displacements in supercells for fc2.
There are two types of displacement dataset. See the docstring
of dataset about types 1 and 2 for the displacement dataset formats.
Displacements set returned depends on either type-1 or type-2 as
follows:
Type-1, List of list
The internal list has 4 elements such as [32, 0.01, 0.0, 0.0]].
The first element is the supercell atom index starting with 0.
The remaining three elements give the displacement in Cartesian
coordinates.
Type-2, array_like
Displacements of all atoms of all supercells in Cartesian
coordinates.
shape=(supercells, natom, 3)
dtype='double'
For setter, only type-2 dataset format is allowed.
displacements : array_like
Atomic displacements of all atoms of all supercells.
Only all displacements in each supercell case (type-2) is
supported.
shape=(supercells, natom, 3), dtype='double', order='C'
"""
if self._phonon_dataset is None:
raise RuntimeError("phonon_dataset is not set.")
if "first_atoms" in self._phonon_dataset:
num_scells = len(self._phonon_dataset["first_atoms"])
natom = len(self._phonon_supercell)
displacements = np.zeros((num_scells, natom, 3), dtype="double", order="C")
for i, disp1 in enumerate(self._phonon_dataset["first_atoms"]):
displacements[i, disp1["number"]] = disp1["displacement"]
elif (
"forces" in self._phonon_dataset or "displacements" in self._phonon_dataset
):
displacements = self._phonon_dataset["displacements"]
else:
raise RuntimeError("displacement dataset has wrong format.")
return displacements
@phonon_displacements.setter
def phonon_displacements(self, displacements):
if self._phonon_supercell_matrix is None:
raise RuntimeError("phonon_supercell_matrix is not set.")
disps = np.array(displacements, dtype="double", order="C")
natom = len(self._phonon_supercell)
if disps.ndim != 3 or disps.shape[1:] != (natom, 3):
raise RuntimeError("Array shape of displacements is incorrect.")
if self._phonon_dataset is None:
self._phonon_dataset = {}
elif "first_atoms" in self._phonon_dataset:
raise RuntimeError("Displacements are incompatible with dataset.")
self._phonon_dataset["displacements"] = disps
self._phonon_supercells_with_displacements = None
@property
def phonon_forces(self):
"""Setter and getter of forces in fc2 displacement dataset.
A set of atomic forces in displaced supercells. The order of
displaced supercells has to match with that in phonon displacement
dataset.
shape=(displaced supercells, atoms in supercell, 3)
getter : ndarray
setter : array_like
The order of supercells used for calculating forces has to
be the same order of phonon_supercells_with_displacements.
"""
return self._get_phonon_forces_energies(target="forces")
@phonon_forces.setter
def phonon_forces(self, values):
self._set_phonon_forces_energies(values, target="forces")
@property
def phonon_supercell_energies(self):
"""Setter and getter of supercell energies in fc2 displacement dataset.
shape=(displaced supercells,)
getter : ndarray
setter : array_like
The order of supercells used for calculating supercell energies has
to be the same order of phonon_supercells_with_displacements.
"""
return self._get_phonon_forces_energies(target="supercell_energies")
@phonon_supercell_energies.setter
def phonon_supercell_energies(self, values):
self._set_phonon_forces_energies(values, target="supercell_energies")
@property
def phph_interaction(self):
"""Return Interaction instance."""
return self._interaction
@property
def detailed_gammas(self):
"""Return detailed gamma."""
return self._detailed_gammas
@property
def grid(self):
"""Return Brillouin zone grid information.
BZGrid
An instance of BZGrid used for entire phono3py calculation.
"""
return self._bz_grid
@property
def mlp(self):
"""Return MLP instance."""
return self._mlp
@property
def phonon_mlp(self):
"""Return MLP instance for fc2."""
return self._phonon_mlp
def init_phph_interaction(
self,
nac_q_direction=None,
constant_averaged_interaction=None,
frequency_scale_factor=None,
symmetrize_fc3q: bool = False,
lapack_zheev_uplo="L",
openmp_per_triplets=None,
):
"""Initialize ph-ph interaction calculation.
This method creates an instance of Interaction class, which
is necessary to run ph-ph interaction calculation.
The input data such as grids, force constants, etc, are
stored to be ready for the calculation.
Note
----
fc3 and fc2, and optionally nac_params have to be set before calling
this method. fc3 and fc2 can be made either from sets of forces
and displacements of supercells or be set simply via attributes.
Parameters
----------
nac_q_direction : array_like, optional
Direction of q-vector watching from Gamma point used for
non-analytical term correction. This is effective only at q=0
(physically q->0). The direction is given in crystallographic
(fractional) coordinates.
shape=(3,), dtype='double'.
Default value is None, which means this feature is not used.
constant_averaged_interaction : float, optional
Ph-ph interaction strength array is replaced by a scalar value.
Default is None, which means this feature is not used.
frequency_scale_factor : float, optional
All phonon frequencies are scaled by this value. Default is None,
which means phonon frequencies are not scaled.
symmetrize_fc3q : bool, optional
fc3 in phonon space is symmetrized by permutation symmetry.
Default is False.
lapack_zheev_uplo : str, optional
'L' or 'U'. Default is 'L'. This is passed to LAPACK zheev
used for phonon solver.
openmp_per_triplets : bool or None, optional, default is None
Normally this parameter should not be touched.
When `True`, ph-ph interaction strength calculation runs with
OpenMP distribution over triplets, and over bands when `False`.
`None` will choose one of them automatically.
"""
if self._bz_grid is None:
msg = "Phono3py.mesh_numbers of instance has to be set."
raise RuntimeError(msg)
if self._fc2 is None:
msg = "Phono3py.fc2 of instance is not found."
raise RuntimeError(msg)
self._interaction = Interaction(
self._primitive,
self._bz_grid,
self._primitive_symmetry,
fc3=self._fc3,
fc3_nonzero_indices=self._fc3_nonzero_indices,
band_indices=self._band_indices_flatten,
constant_averaged_interaction=constant_averaged_interaction,
frequency_factor_to_THz=self._frequency_factor_to_THz,
frequency_scale_factor=frequency_scale_factor,
cutoff_frequency=self._cutoff_frequency,
is_mesh_symmetry=self._is_mesh_symmetry,
symmetrize_fc3q=symmetrize_fc3q,
make_r0_average=self._make_r0_average,
lapack_zheev_uplo=lapack_zheev_uplo,
openmp_per_triplets=openmp_per_triplets,
)
self._interaction.nac_q_direction = nac_q_direction
self._init_dynamical_matrix()
def set_phonon_data(self, frequencies, eigenvectors, grid_address):
"""Set phonon frequencies and eigenvectors in Interaction instance.
Harmonic phonon information is stored in Interaction instance. For
example, this information store in a file is read and passed to
Phono3py instance by using this method. The grid_address is used
for the consistency check.
Parameters
----------
frequencies : array_like
Phonon frequencies.
shape=(num_grid_points, num_band), dtype='double', order='C'
eigenvectors : array_like
Phonon eigenvectors
shape=(num_grid_points, num_band, num_band)
dtype='complex128', order='C'
grid_address : array_like
Grid point addresses by integers. The first dimension may not be
prod(mesh) because it includes Brillouin zone boundary. The detail
is found in the docstring of
phono3py.phonon3.triplets.get_triplets_at_q.
shape=(num_grid_points, 3), dtype=int
"""
if self._interaction is not None:
self._interaction.set_phonon_data(frequencies, eigenvectors, grid_address)
def get_phonon_data(self):
"""Get phonon frequencies and eigenvectors in Interaction instance.
Harmonic phonon information is stored in Interaction instance. This
information can be obtained. The grid_address returned give the
q-points locations with respect to reciprocal basis vectors by
integers in the way that
q_points = grid_address / np.array(mesh, dtype='double').
Returns
-------
tuple
(frequencies, eigenvectors, grid_address)
See more details at the docstring of set_phonon_data.
"""
if self._interaction is not None:
freqs, eigvecs, _ = self._interaction.get_phonons()
return freqs, eigvecs, self._interaction.bz_grid.addresses
else:
msg = (
"Phono3py.init_phph_interaction has to be called "
"before running this method."
)
raise RuntimeError(msg)
def run_phonon_solver(self, grid_points=None):
"""Run harmonic phonon calculation on grid points.
Parameters
----------
grid_points : array_like or None, optional
A list of grid point indices of Phono3py.grid.addresses.
Specifying None runs all phonons on the grid points unless
those phonons were already calculated. Normally phonons at
[0, 0, 0] point is already calculated before calling this method.
Phonon calculations are performed automatically when needed
internally for ph-ph calculation. Therefore calling this method
is not necessary in most cases.
The phonon results are obtained by Phono3py.get_phonon_data().
"""
if self._interaction is not None:
self._interaction.run_phonon_solver(grid_points=grid_points)
else:
msg = (
"Phono3py.init_phph_interaction has to be called "
"before running this method."
)
raise RuntimeError(msg)
def generate_displacements(
self,
distance: Optional[float] = None,
cutoff_pair_distance: Optional[float] = None,
is_plusminus: Union[bool, str] = "auto",
is_diagonal: bool = True,
number_of_snapshots: Optional[Union[int, Literal["auto"]]] = None,
random_seed: Optional[int] = None,
max_distance: Optional[float] = None,
):
"""Generate displacement dataset in supercell for fc3.
Systematic displacements
------------------------
Unless number_of_snapshots is specified, this method systematically
generates single and pair atomic displacements in supercells to
calculate fc3 considering crystal symmetry.
For fc3, two atoms are displaced for each configuration considering
crystal symmetry. The first displacement is chosen in the perfect
supercell, and the second displacement in the displaced supercell. The
first displacements are taken along the basis vectors of the supercell.
This is because the symmetry is expected to be less broken by the
introduced first displacement, and as the result, the number of second
displacements may become smaller than the case that the first atom is
displaced not along the basis vectors.
Random displacements
--------------------
Unless number_of_snapshots is specified, displacements are generated
randomly. There are several options how the random displacements are
generated.
Note
----
When phonon_supercell_matrix is not given, fc2 is also computed from the
same set of the displacements for fc3 and respective supercell forces.
When phonon_supercell_matrix is set, the displacements in
phonon_supercell are generated unless those already exist. If a specific
set of displacements for fc2 is expected, generate_fc2_displacements
should be called.
Parameters
----------
distance : float, optional
Constant displacement Euclidean distance. Default is None, which
gives 0.03. For random direction and random distance displacements
generation, this value is also used as `min_distance`, is used to
replace generated random distances smaller than this value by this
value.
cutoff_pair_distance : float, optional
This is used as a cutoff Euclidean distance to determine if each
pair of displacements is considered to calculate fc3 or not. Default
is None, which means cutoff is not used.
is_plusminus : True, False, or 'auto', optional
With True, atomis are displaced in both positive and negative
directions. With False, only one direction. With 'auto', mostly
equivalent to is_plusminus=True, but only one direction is chosen
when the displacements in both directions are symmetrically
equivalent. Default is 'auto'.
is_diagonal : Bool, optional
With False, the second displacements are made along the basis
vectors of the supercell. With True, direction not along the basis
vectors can be chosen when the number of the displacements may be
reduced.
number_of_snapshots : int, "auto", or None, optional
Number of snapshots of supercells with random displacements. Random
displacements are generated by shifting all atoms in random
directions by a fixed distance specified by the `distance`
parameter. In other words, all atoms in the supercell are displaced
by the same distance in direct space. When “auto”, the minimum
required number of snapshots is estimated using symfc and then
doubled. The default is None.
random_seed : int or None, optional
Random seed for random displacements generation. Default is None.
max_distance : float or None, optional
In random direction and distance displacements generation, this
value is specified. In random direction and random distance
displacements generation, this value is used as `max_distance`.
"""
if distance is None:
_distance = 0.03
else:
_distance = distance
if number_of_snapshots is not None and (
number_of_snapshots == "auto" or number_of_snapshots > 0
):
if number_of_snapshots == "auto":
if cutoff_pair_distance is None:
options = None
else:
options = {"cutoff": {3: cutoff_pair_distance}}
_number_of_snapshots = (
SymfcFCSolver(
self._supercell,
symmetry=self._symmetry,
options=options,
).estimate_numbers_of_supercells(orders=[3])[3]
* 2
)
else:
_number_of_snapshots = number_of_snapshots
self._dataset = self._generate_random_displacements(
_number_of_snapshots,
len(self._supercell),
distance=_distance,
is_plusminus=is_plusminus is True,
random_seed=random_seed,
max_distance=max_distance,
)
if cutoff_pair_distance is not None:
self._dataset["cutoff_distance"] = cutoff_pair_distance
else:
direction_dataset = get_third_order_displacements(
self._supercell,
self._symmetry,
is_plusminus=is_plusminus,
is_diagonal=is_diagonal,
)
self._dataset = direction_to_displacement(
direction_dataset,
_distance,
self._supercell,
cutoff_distance=cutoff_pair_distance,
)
self._supercells_with_displacements = None
if self._phonon_supercell_matrix is not None and self._phonon_dataset is None:
self.generate_fc2_displacements(
distance=_distance, is_plusminus=is_plusminus, is_diagonal=False
)
def generate_fc2_displacements(
self,
distance: Optional[float] = None,
is_plusminus: str = "auto",
is_diagonal: bool = False,
number_of_snapshots: Optional[Union[int, Literal["auto"]]] = None,
random_seed: Optional[int] = None,
max_distance: Optional[float] = None,
):
"""Generate displacement dataset in phonon supercell for fc2.
This systematically generates single atomic displacements in supercells
to calculate phonon_fc2 considering crystal symmetry. When this method
is called, existing cache of supercells with displacements for fc2 are
removed.
Note
----
is_diagonal=False is chosen as the default setting intentionally to be
consistent to the first displacements of the fc3 pair displacements in
supercell.
Parameters
----------
distance : float, optional
Constant displacement Euclidean distance. Default is None, which
gives 0.03. For random direction and random distance displacements
generation, this value is also used as `min_distance`, is used to
replace generated random distances smaller than this value by this
value.
is_plusminus : True, False, or 'auto', optional
With True, atoms are displaced in both positive and negative
directions. With False, only one direction. With 'auto', mostly
equivalent to is_plusminus=True, but only one direction is chosen
when the displacements in both directions are symmetrically
equivalent. Default is 'auto'.
is_diagonal : Bool, optional
With False, the displacements are made along the basis vectors of
the supercell. With True, direction not along the basis vectors can
be chosen when the number of the displacements may be reduced.
Default is False.
number_of_snapshots : int, "auto", or None, optional
Number of snapshots of supercells with random displacements. Random
displacements are generated by shifting all atoms in random
directions by a fixed distance specified by the `distance`
parameter. In other words, all atoms in the supercell are displaced
by the same distance in direct space. When “auto”, the minimum
required number of snapshots is estimated using symfc and then
doubled. The default is None.
random_seed : int or None, optional
Random seed for random displacements generation. Default is None.
max_distance : float or None, optional
In random direction and distance displacements generation, this
value is specified. In random direction and random distance
displacements generation, this value is used as `max_distance`.
"""
if distance is None:
_distance = 0.03
else:
_distance = distance
if number_of_snapshots is not None and (
number_of_snapshots == "auto" or number_of_snapshots > 0
):
if number_of_snapshots == "auto":
_number_of_snapshots = (
SymfcFCSolver(
self._supercell, symmetry=self._symmetry
).estimate_numbers_of_supercells(orders=[2])[2]
* 2
)
else:
_number_of_snapshots = number_of_snapshots
self._phonon_dataset = self._generate_random_displacements(
_number_of_snapshots,
len(self._phonon_supercell),
distance=_distance,
is_plusminus=is_plusminus is True,
random_seed=random_seed,
max_distance=max_distance,
)
else:
phonon_displacement_directions = get_least_displacements(
self._phonon_supercell_symmetry,
is_plusminus=is_plusminus,
is_diagonal=is_diagonal,
)
self._phonon_dataset = directions_to_displacement_dataset(
phonon_displacement_directions, _distance, self._phonon_supercell
)
self._phonon_supercells_with_displacements = None
def produce_fc3(
self,
symmetrize_fc3r: bool = False,
is_compact_fc: bool = False,
fc_calculator: Optional[Union[str, dict]] = None,
fc_calculator_options: Optional[Union[str, dict]] = None,
):
"""Calculate fc3 from displacements and forces.
Parameters
----------
symmetrize_fc3r : bool, optional
Only for type 1 displacement_dataset, translational and permutation
symmetries are applied after creating fc3. This symmetrization is
not very sophisticated and can break space group symmetry, but often
useful. If better symmetrization is expected, it is recommended to
use external force constants calculator such as ALM. Default is
False.
is_compact_fc : bool, optional
fc3 shape is
False: (supercell, supercell, supercell, 3, 3, 3) True:
(primitive, supercell, supercell, 3, 3, 3)
where 'supercell' and 'primitive' indicate number of atoms in these
cells. Default is False.
fc_calculator : str, optional
Force constants calculator given by str.
fc_calculator_options : dict or str, optional
Options for external force constants calculator.
"""
fc_solver_name = fc_calculator if fc_calculator is not None else "traditional"
fc_solver = FC3Solver(
fc_solver_name,
self._supercell,
symmetry=self._symmetry,
dataset=self._dataset,
is_compact_fc=is_compact_fc,
primitive=self._primitive,
orders=[2, 3],
options=fc_calculator_options,
log_level=self._log_level,
)
fc2 = fc_solver.force_constants[2]
fc3 = fc_solver.force_constants[3]
if fc_calculator is None or fc_calculator == "traditional":
if symmetrize_fc3r:
if is_compact_fc:
set_translational_invariance_compact_fc3(fc3, self._primitive)
set_permutation_symmetry_compact_fc3(fc3, self._primitive)
if self._fc2 is None:
symmetrize_compact_force_constants(fc2, self._primitive)
else:
set_translational_invariance_fc3(fc3)
set_permutation_symmetry_fc3(fc3)
if self._fc2 is None:
symmetrize_force_constants(fc2)
self._fc3 = fc3
self._fc3_nonzero_indices = None
if fc_calculator == "symfc":
symfc_solver = cast(SymfcFCSolver, fc_solver.fc_solver)
fc3_nonzero_elems = symfc_solver.get_nonzero_atomic_indices_fc3()
if fc3_nonzero_elems is not None:
if is_compact_fc:
self._fc3_nonzero_indices = np.array(
fc3_nonzero_elems[self._primitive.p2s_map],
dtype="byte",
order="C",
)
else:
self._fc3_nonzero_indices = np.array(
fc3_nonzero_elems, dtype="byte", order="C"
)
# fc2 as obtained above will not be set when "|" in fc-calculator setting.
if fc_calculator is not None and "|" in fc_calculator:
fc2 = None
if fc_calculator_options is not None and "|" in fc_calculator_options:
fc2 = None
# fc2 should not be set if phonon_supercell_matrix is available.
if self._phonon_supercell_matrix is not None:
fc2 = None
# Normally self._fc2 is overwritten in produce_fc2
if self._fc2 is None:
self._fc2 = fc2
def produce_fc2(
self,
symmetrize_fc2=False,
is_compact_fc=False,
fc_calculator=None,
fc_calculator_options=None,
):
"""Calculate fc2 from displacements and forces.
Parameters
----------
symmetrize_fc2 : bool
Only for type 1 displacement_dataset, translational and
permutation symmetries are applied after creating fc3. This
symmetrization is not very sophisticated and can break space
group symmetry, but often useful. If better symmetrization is
expected, it is recommended to use external force constants
calculator such as ALM. Default is False.
is_compact_fc : bool
fc2 shape is
False: (supercell, supercell, 3, 3)
True: (primitive, supercell, 3, 3)
where 'supercell' and 'primitive' indicate number of atoms in these
cells. Default is False.
fc_calculator : str or None
Force constants calculator given by str.
fc_calculator_options : dict
Options for external force constants calculator.
"""
if self._phonon_dataset is None:
disp_dataset = self._dataset
else:
disp_dataset = self._phonon_dataset
if not forces_in_dataset(disp_dataset):
raise RuntimeError("Forces are not set in the dataset.")
self._fc2 = get_fc2(
self._phonon_supercell,
disp_dataset,
primitive=self._phonon_primitive,
fc_calculator=fc_calculator,
fc_calculator_options=fc_calculator_options,
is_compact_fc=is_compact_fc,
symmetry=self._phonon_supercell_symmetry,
log_level=self._log_level,
)
if fc_calculator is None or fc_calculator == "traditional":
if symmetrize_fc2:
if is_compact_fc:
symmetrize_compact_force_constants(
self._fc2, self._phonon_primitive
)
else:
symmetrize_force_constants(self._fc2)
def cutoff_fc3_by_zero(self, cutoff_distance, fc3=None):
"""Set zero to fc3 elements out of cutoff distance.
Note
----
fc3 is overwritten.
Parameters
----------
cutoff_distance : float
After creating force constants, fc elements where any pair
distance in atom triplets larger than cutoff_distance are set zero.
"""
if fc3 is None:
_fc3 = self._fc3
else:
_fc3 = fc3
cutoff_fc3_by_zero(
_fc3,
self._supercell,
cutoff_distance,
p2s_map=self._primitive.p2s_map,
symprec=self._symprec,
)
def set_permutation_symmetry(self):
"""Enforce permutation symmetry to fc2 and fc3."""
if self._fc2 is not None:
set_permutation_symmetry(self._fc2)
if self._fc3 is not None:
set_permutation_symmetry_fc3(self._fc3)
def set_translational_invariance(self):
"""Enforce translation invariance.
This subtracts drift divided by number of elements in each row and
column.
"""
if self._fc2 is not None:
set_translational_invariance(self._fc2)
if self._fc3 is not None:
set_translational_invariance_fc3(self._fc3)
def run_imag_self_energy(
self,
grid_points,
temperatures,
frequency_points=None,
frequency_step=None,
num_frequency_points=None,
num_points_in_batch=None,
frequency_points_at_bands=False,
scattering_event_class=None,
write_txt=False,
write_gamma_detail=False,
keep_gamma_detail=False,
output_filename=None,
):
"""Calculate imaginary part of self-energy of bubble diagram (Gamma).
Pi = Delta - i Gamma.
Parameters
----------
grid_points : array_like
Grid-point indices where imaginary part of self-energies are
caclculated.
dtype=int, shape=(grid_points,)
temperatures : array_like
Temperatures where imaginary part of self-energies are calculated.
dtype=float, shape=(temperatures,)
frequency_points : array_like, optional
Frequency sampling points. Default is None. With
frequency_points_at_bands=False and frequency_points is None,
num_frequency_points or frequency_step is used to generate uniform
frequency sampling points.
dtype=float, shape=(frequency_points,)
frequency_step : float, optional
Uniform pitch of frequency sampling points. Default is None. This
results in using num_frequency_points.
num_frequency_points: Int, optional
Number of sampling sampling points to be used instead of
frequency_step. This number includes end points. Default is None,
which gives 201.
num_points_in_batch: int, optional
Number of sampling points in one batch. This is for the frequency
sampling mode and the sampling points are divided into batches.
Lager number provides efficient use of multi-cores but more
memory demanding. Default is None, which give the number of 10.
frequency_points_at_bands : bool, optional
Phonon band frequencies are used as frequency points when True.
Default is False.
scattering_event_class : int, optional
Specific choice of scattering event class, 1 or 2 that is specified
1 or 2, respectively. The result is stored in gammas. Therefore
usual gammas are not stored in the variable. Default is None, which
doesn't specify scattering_event_class.
write_txt : bool, optional
Frequency points and imaginary part of self-energies are written
into text files.
write_gamma_detail : bool, optional
Detailed gammas are written into a file in hdf5. Default is False.
keep_gamma_detail : bool, optional
With True, detailed gammas are stored. Default is False.
output_filename : str
This string is inserted in the output file names.
"""
if self._interaction is None:
msg = (
"Phono3py.init_phph_interaction has to be called "
"before running this method."
)
raise RuntimeError(msg)
if temperatures is None:
self._temperatures = [
300.0,
]
else:
self._temperatures = temperatures
self._grid_points = grid_points
self._scattering_event_class = scattering_event_class
vals = get_imag_self_energy(
self._interaction,
grid_points,
temperatures,
sigmas=self._sigmas,
frequency_points=frequency_points,
frequency_step=frequency_step,
frequency_points_at_bands=frequency_points_at_bands,
num_frequency_points=num_frequency_points,
num_points_in_batch=num_points_in_batch,
scattering_event_class=scattering_event_class,
write_gamma_detail=write_gamma_detail,
return_gamma_detail=keep_gamma_detail,
output_filename=output_filename,
log_level=self._log_level,
)
if keep_gamma_detail:
(self._frequency_points, self._gammas, self._detailed_gammas) = vals
else:
self._frequency_points, self._gammas = vals
if write_txt:
self._write_imag_self_energy(output_filename=output_filename)
return vals
def _write_imag_self_energy(self, output_filename=None):
write_imag_self_energy(
self._gammas,
self.mesh_numbers,
self._grid_points,
self._band_indices,
self._frequency_points,
self._temperatures,
self._sigmas,
scattering_event_class=self._scattering_event_class,
output_filename=output_filename,
is_mesh_symmetry=self._is_mesh_symmetry,
log_level=self._log_level,
)
def run_real_self_energy(
self,
grid_points,
temperatures,
frequency_points_at_bands=False,
frequency_points=None,
frequency_step=None,
num_frequency_points=None,
epsilons=None,
write_txt=False,
write_hdf5=False,
output_filename=None,
):
"""Calculate real-part of self-energy of bubble diagram (Delta).
Pi = Delta - i Gamma.
Parameters
----------
grid_points : array_like
Grid-point indices where real part of self-energies are
caclculated.
dtype=int, shape=(grid_points,)
temperatures : array_like
Temperatures where real part of self-energies are calculated.
dtype=float, shape=(temperatures,)
frequency_points_at_bands : bool, optional
With False, frequency shifts are calculated at frequency sampling
points. When True, they are done at the phonon frequencies.
Default is False.
frequency_points : array_like, optional
Frequency sampling points. Default is None. In this case,
num_frequency_points or frequency_step is used to generate uniform
frequency sampling points.
dtype=float, shape=(frequency_points,)
frequency_step : float, optional
Uniform pitch of frequency sampling points. Default is None. This
results in using num_frequency_points.
num_frequency_points: Int, optional
Number of sampling sampling points to be used instead of
frequency_step. This number includes end points. Default is None,
which gives 201.
epsilons : array_like
Smearing widths to computer principal part. When multiple values
are given frequency shifts for those values are returned.
dtype=float, shape=(epsilons,)
write_txt : bool, optional
Frequency points and real part of self-energies are written
into text files.
write_hdf5 : bool
Results are stored in hdf5 files independently at grid points,
epsilons, and temperatures.
output_filename : str
This string is inserted in the output file names.
"""
if self._interaction is None:
msg = (
"Phono3py.init_phph_interaction has to be called "
"before running this method."
)
raise RuntimeError(msg)
if epsilons is not None:
_epsilons = epsilons
else:
if len(self._sigmas) == 1 and self._sigmas[0] is None:
_epsilons = None
elif self._sigmas[0] is None:
_epsilons = self._sigmas[1:]
else:
_epsilons = self._sigmas
# (epsilon, grid_point, temperature, band)
frequency_points, deltas = get_real_self_energy(
self._interaction,
grid_points,
temperatures,
epsilons=_epsilons,
frequency_points=frequency_points,
frequency_step=frequency_step,
num_frequency_points=num_frequency_points,
frequency_points_at_bands=frequency_points_at_bands,
write_hdf5=write_hdf5,
output_filename=output_filename,
log_level=self._log_level,
)
if write_txt:
write_real_self_energy(
deltas,
self.mesh_numbers,
grid_points,
self._band_indices,
frequency_points,
temperatures,
_epsilons,
output_filename=output_filename,
is_mesh_symmetry=self._is_mesh_symmetry,
log_level=self._log_level,
)
return frequency_points, deltas
def run_spectral_function(
self,
grid_points,
temperatures,
frequency_points=None,
frequency_step=None,
num_frequency_points=None,
num_points_in_batch=None,
write_txt=False,
write_hdf5=False,
output_filename=None,
):
"""Frequency shift from lowest order diagram is calculated.
Parameters
----------
grid_points : array_like
Grid-point indices where imag-self-energeis are caclculated.
dtype=int, shape=(grid_points,)
temperatures : array_like
Temperatures where imag-self-energies are calculated.
dtype=float, shape=(temperatures,)
frequency_points : array_like, optional
Frequency sampling points. Default is None. In this case,
num_frequency_points or frequency_step is used to generate uniform
frequency sampling points.
dtype=float, shape=(frequency_points,)
frequency_step : float, optional
Uniform pitch of frequency sampling points. Default is None. This
results in using num_frequency_points.
num_frequency_points: Int, optional
Number of sampling sampling points to be used instead of
frequency_step. This number includes end points. Default is None,
which gives 201.
num_points_in_batch: int, optional
Number of sampling points in one batch. This is for the frequency
sampling mode and the sampling points are divided into batches.
Lager number provides efficient use of multi-cores but more
memory demanding. Default is None, which give the number of 10.
write_txt : bool, optional
Frequency points and spectral functions are written
into text files. Default is False.
write_hdf5 : bool
Results are stored in hdf5 files independently at grid points,
epsilons. Default is False.
output_filename : str
This string is inserted in the output file names.
"""
if self._interaction is None:
msg = (
"Phono3py.init_phph_interaction has to be called "
"before running this method."
)
raise RuntimeError(msg)
self._spectral_function = run_spectral_function(
self._interaction,
grid_points,
temperatures=temperatures,
sigmas=self._sigmas,
frequency_points=frequency_points,
frequency_step=frequency_step,
num_frequency_points=num_frequency_points,
num_points_in_batch=num_points_in_batch,
band_indices=self._band_indices,
write_txt=write_txt,
write_hdf5=write_hdf5,
output_filename=output_filename,
log_level=self._log_level,
)
def run_thermal_conductivity(
self,
is_LBTE: bool = False,
temperatures: Optional[Sequence] = None,
is_isotope: bool = False,
mass_variances: Optional[Sequence] = None,
grid_points: Optional[Sequence[int]] = None,
boundary_mfp: Optional[float] = None, # in micrometer
solve_collective_phonon: bool = False,
use_ave_pp: bool = False,
is_reducible_collision_matrix: bool = False,
is_kappa_star: bool = True,
gv_delta_q: Optional[float] = None, # for group velocity
is_full_pp: bool = False,
pinv_cutoff: float = 1.0e-8, # for pseudo-inversion of collision matrix
pinv_method: int = 0, # for pseudo-inversion of collision matrix
pinv_solver: int = 0, # solver of pseudo-inversion of collision matrix
write_gamma: bool = False,
read_gamma: bool = False,
is_N_U: bool = False,
conductivity_type: Optional[str] = None,
write_kappa: bool = False,
write_gamma_detail: bool = False,
write_collision: bool = False,
read_collision: bool = False,
write_pp: bool = False,
read_pp: bool = False,
write_LBTE_solution: bool = False,
compression: str = "gzip",
input_filename: Optional[str] = None,
output_filename: Optional[str] = None,
log_level: Optional[int] = None,
):
"""Run thermal conductivity calculation.
Parameters
----------
is_LBTE : bool, optional, default is False
RTA (False) or direct solution (True).
temperatures : array_like, optional, default is None
Temperatures at which thermal conductivity is calculated.
shape=(temperature_points, ), dtype='double'.
With None,
`is_LBTE=False` gives temperatures=[0, 10, ..., 1000].
`is_LBTE=True` gives temperatures=[300, ].
is_isotope : bool, optional, default is False
With or without isotope scattering.
mass_variances : array_like, optional, default is None
Mass variances for isotope scattering calculation. When None,
the values stored in phono3py are used with `is_isotope=True`.
shape(atoms_in_primitive, ), dtype='double'.
grid_points : array_like, optional, default is None
List of grid point indices where mode thermal conductivities are
calculated. With None, all the grid points that are necessary
for thermal conductivity are set internally.
shape(num_grid_points, ), dtype='int64'.
boundary_mfp : float, optional, default is None
Mean free path in micrometer to calculate simple boundary
scattering contribution to thermal conductivity.
None ignores this contribution.
solve_collective_phonon : bool, optional, default is False
This is an option for the feature under development.
use_ave_pp : bool, optional, default is False
RTA only (`is_LBTE=False`). Averaged phonon-phonon interaction
strength is used to calculate phonon lifetime. This does not
reduce computational demand, but may be used to model thermal
conductivity for analyze the calculation results.
is_reducible_collision_matrix : bool, optional, default is False
Direct solution only (`is_LBTE=True`). This is an experimental
option. With True, full collision matrix is created and solved.
is_kappa_star : bool, optional, default is True
With true, symmetry is considered when sampling grid points
at which mode thermal conductivities are calculated.
gv_delta_q : float, optional, default is None, # for group velocity
With non-analytical correction, group velocity is calculated
by central finite difference method. This value gives the distance
in both directions in 1/Angstrom. The default value will be 1e-5.
is_full_pp : bool, optional, default is False
With True, full elements of phonon-phonon interaction strength
are computed. However with tetrahedron method, part of them are
known to be zero and unnecessary to calculation. With False,
those elements are not calculated, by which considerable
improve of efficiency is expected.
With smearing method, even if this is set False, full elements
are computed unless `sigma_cutoff` is specified.
pinv_cutoff : float, optional, default is 1.0e-8
Direct solution only (`is_LBTE=True`). This is used as a criterion
to judge the eigenvalues are considered as zero or not in
pseudo-inversion of collision matrix. See also `pinv_method`.
pinv_method : int, optional, default is 0.
Direct solution only (`is_LBTE=True`).
0. abs(eigenvalue) < `pinv_cutoff`
1. eigenvalue < `pinv_cutoff`
pinv_solver : int, optional, default is 0
Direct solution only (`is_LBTE=True`). Choice of solver of
pseudo-inversion of collision matrix. 0 means the default choice.
1. Lapacke dsyev: Smaller memory consumption than dsyevd, but
slower. This is the default solver when MKL LAPACKE is
integrated or scipy is not installed.
2. Lapacke dsyevd: Larger memory consumption than dsyev, but
faster. This is not recommended because sometimes a wrong
result is obtained.
3. Numpy’s dsyevd (linalg.eigh). This is not recommended
because sometimes a wrong result is obtained.
4. Scipy’s dsyev: This is the default solver when scipy is
installed and MKL LAPACKE is not integrated.
5. Scipy’s dsyevd. This is not recommended because sometimes
a wrong result is obtained.
The solver choices other than --pinv-solver=1 and
--pinv-solver=4 are dangerous and not recommend.
write_gamma : bool, optional, default is False
RTA only (`is_LBTE=False`). With True, Write mode thermal
conductivity properties into files at each grid point. With
`band_indices` or multiple `sigmas` is specified, the files
are made for each of them, too.
read_gamma : bool, optional, default is False
RTA only (`is_LBTE=False`). With True, dead files created by
`write_gamma=True` instead of calculating phonon-phonon
interaction strength and imaginary parts of self-energy.
is_N_U : bool, optional, default is False
RTA only (`is_LBTE=False`). With True, categorization of normal
and Umklapp scattering is made and imaginary parts of self energy
for them are separated.
conductivity_type : str, optional
"wigner", "kubo", or None. Default is None.
write_kappa : bool, optional, default is False
With True, thermal conductivity and related properties are
written into a file. With multiple `sigmas`, respective files
are created.
write_gamma_detail : bool, optional, default is False
RTA only (`is_LBTE=False`). With True, detailed information of
imaginary parts of self energy is stored into files such as
those made by `write_gamma`.
write_collision : bool, optional, default is False
Direct solution only (`is_LBTE=True`). With True, collision matrix
is written into a file. With multiple `sigmas` specified,
respective files are created. Be careful that this file can be
huge.
read_collision : bool, optional, default is False
Direct solution only (`is_LBTE=True`). With True, collision matrix
is read from a file.
write_pp : bool, optional, default is False
With True, phonon-phonon interaction strength is written into
files at each grid point. This option assumes single value is in
`sigmas`.
read_pp : bool, optional, default is False
With True, phonon-phonon interaction strength is read from files.
write_LBTE_solution : bool, optional, default is False
Direct solution only (`is_LBTE=True`). With True, eigenvectors of
collision matrix is written in a file as the row vectors except
unless `pinv_solver=3` (for this, column vectors). With multiple
`sigmas` specified, respective files are created. Be careful that
this file can be huge.
compression: str, optional, default is "gzip"
When writing results into files in hdf5, large data are compressed
by this options. See the detail at h5py documentation.
input_filename : str, optional, default is None
Deprecated. When specified, the string is inserted before filename
extension in reading files.
output_filename : str, optional, default is None
Deprecated. When specified, the string is inserted before filename
extension in writing files.
"""
if input_filename is not None:
warnings.warn(
"input_filename parameter is deprecated.",
DeprecationWarning,
stacklevel=2,
)
if output_filename is not None:
warnings.warn(
"output_filename parameter is deprecated.",
DeprecationWarning,
stacklevel=2,
)
if self._interaction is None:
msg = (
"Phono3py.init_phph_interaction has to be called "
"before running this method."
)
raise RuntimeError(msg)
if log_level is None:
_log_level = self._log_level
else:
_log_level = log_level
if is_LBTE:
if temperatures is None:
_temperatures = [
300,
]
else:
_temperatures = temperatures
self._thermal_conductivity = get_thermal_conductivity_LBTE(
self._interaction,
temperatures=_temperatures,
sigmas=self._sigmas,
sigma_cutoff=self._sigma_cutoff,
is_isotope=is_isotope,
mass_variances=mass_variances,
grid_points=grid_points,
boundary_mfp=boundary_mfp,
solve_collective_phonon=solve_collective_phonon,
is_reducible_collision_matrix=is_reducible_collision_matrix,
is_kappa_star=is_kappa_star,
gv_delta_q=gv_delta_q,
is_full_pp=is_full_pp,
conductivity_type=conductivity_type,
pinv_cutoff=pinv_cutoff,
pinv_solver=pinv_solver,
pinv_method=pinv_method,
write_collision=write_collision,
read_collision=read_collision,
write_kappa=write_kappa,
write_pp=write_pp,
read_pp=read_pp,
write_LBTE_solution=write_LBTE_solution,
compression=compression,
input_filename=input_filename,
output_filename=output_filename,
log_level=_log_level,
)
else:
if temperatures is None:
_temperatures = np.arange(0, 1001, 10, dtype="double")
else:
_temperatures = temperatures
self._thermal_conductivity = get_thermal_conductivity_RTA(
self._interaction,
temperatures=_temperatures,
sigmas=self._sigmas,
sigma_cutoff=self._sigma_cutoff,
is_isotope=is_isotope,
mass_variances=mass_variances,
grid_points=grid_points,
boundary_mfp=boundary_mfp,
use_ave_pp=use_ave_pp,
is_kappa_star=is_kappa_star,
gv_delta_q=gv_delta_q,
is_full_pp=is_full_pp,
is_N_U=is_N_U,
conductivity_type=conductivity_type,
write_gamma=write_gamma,
read_gamma=read_gamma,
write_kappa=write_kappa,
write_pp=write_pp,
read_pp=read_pp,
write_gamma_detail=write_gamma_detail,
compression=compression,
input_filename=input_filename,
output_filename=output_filename,
log_level=_log_level,
)
def save(
self, filename: str = "phono3py_params.yaml", settings: Optional[dict] = None
):
"""Save parameters in Phono3py instants into file.
Parameters
----------
filename: str, optional
File name. Default is "phono3py_params.yaml"
settings: dict, optional
It is described which parameters are written out. Only
the settings expected to be updated from the following
default settings are needed to be set in the dictionary.
The possible parameters and their default settings are:
{'force_sets': True,
'displacements': True,
'force_constants': False,
'born_effective_charge': True,
'dielectric_constant': True}
"""
ph3py_yaml = Phono3pyYaml(settings=settings)
ph3py_yaml.set_phonon_info(self)
with open(filename, "w") as w:
w.write(str(ph3py_yaml))
def develop_mlp(
self,
params: Optional[Union[PypolymlpParams, dict, str]] = None,
test_size: float = 0.1,
):
"""Develop machine learning potential.
Parameters
----------
params : PypolymlpParams or dict, optional
Parameters for developing MLP. Default is None. When dict is given,
PypolymlpParams instance is created from the dict.
test_size : float, optional
Training and test data are split by this ratio. test_size=0.1
means the first 90% of the data is used for training and the rest
is used for test. Default is 0.1.
"""
if self._mlp_dataset is None:
raise RuntimeError("MLP dataset is not set.")
self._mlp = PhonopyMLP(log_level=self._log_level)
self._mlp.develop(
self._mlp_dataset,
self._supercell,
params=params,
test_size=test_size,
)
def save_mlp(self, filename: Optional[str] = None):
"""Save machine learning potential."""
if self._mlp is None:
raise RuntimeError("MLP is not developed yet.")
self._mlp.save(filename=filename)
def load_mlp(self, filename: Optional[str] = None):
"""Load machine learning potential."""
self._mlp = PhonopyMLP(log_level=self._log_level)
self._mlp.load(filename=filename)
def evaluate_mlp(self):
"""Evaluate machine learning potential.
This method calculates the supercell energies and forces from the MLP
for the displacements in self._dataset of type 2. The results are stored
in self._dataset.
The displacements may be generated by the produce_force_constants method
with number_of_snapshots > 0. With MLP, a small distance parameter, such
as 0.01, can be numerically stable for the computation of force
constants.
"""
if self._mlp is None:
raise RuntimeError("MLP is not developed yet.")
if self.supercells_with_displacements is None:
raise RuntimeError("Displacements are not set. Run generate_displacements.")
energies, forces, _ = self._mlp.evaluate(self.supercells_with_displacements)
self.supercell_energies = energies
self.forces = forces
def develop_phonon_mlp(
self,
params: Optional[Union[PypolymlpParams, dict, str]] = None,
test_size: float = 0.1,
):
"""Develop MLP for fc2.
Parameters
----------
params : PypolymlpParams or dict, optional
Parameters for developing MLP. Default is None. When dict is given,
PypolymlpParams instance is created from the dict.
test_size : float, optional
Training and test data are split by this ratio. test_size=0.1
means the first 90% of the data is used for training and the rest
is used for test. Default is 0.1.
"""
if self._phonon_mlp_dataset is None:
raise RuntimeError("MLP dataset is not set.")
self._phonon_mlp = PhonopyMLP(log_level=self._log_level)
self._phonon_mlp.develop(
self._phonon_mlp_dataset,
self._phonon_supercell,
params=params,
test_size=test_size,
)
def save_phonon_mlp(self, filename: Optional[str] = None):
"""Save machine learning potential."""
if self._mlp is None:
raise RuntimeError("MLP is not developed yet.")
self._phonon_mlp.save(filename=filename)
def load_phonon_mlp(self, filename: Optional[str] = None):
"""Load machine learning potential."""
self._phonon_mlp = PhonopyMLP(log_level=self._log_level)
self._phonon_mlp.load(filename=filename)
def evaluate_phonon_mlp(self):
"""Evaluate the machine learning potential.
This method calculates the supercell energies and forces from the MLP
for the displacements in self._dataset of type 2. The results are stored
in self._dataset.
The displacements may be generated by the produce_force_constants method
with number_of_snapshots > 0. With MLP, a small distance parameter, such
as 0.01, can be numerically stable for the computation of force
constants.
"""
if self._mlp is None and self._phonon_mlp is None:
raise RuntimeError("MLP is not developed yet.")
if self.phonon_supercells_with_displacements is None:
raise RuntimeError(
"Displacements are not set. Run generate_fc2_displacements."
)
if self._phonon_mlp is None:
mlp = self._mlp
else:
mlp = self._phonon_mlp
energies, forces, _ = mlp.evaluate(self.phonon_supercells_with_displacements)
self.phonon_supercell_energies = energies
self.phonon_forces = forces
###################
# private methods #
###################
def _search_symmetry(self):
self._symmetry = Symmetry(
self._supercell, symprec=self._symprec, is_symmetry=self._is_symmetry
)
def _search_primitive_symmetry(self):
self._primitive_symmetry = Symmetry(
self._primitive, self._symprec, self._is_symmetry
)
if len(self._symmetry.pointgroup_operations) != len(
self._primitive_symmetry.pointgroup_operations
):
print(
"Warning: point group symmetries of supercell and primitive"
"cell are different."
)
def _search_phonon_supercell_symmetry(self):
if self._phonon_supercell_matrix is None:
self._phonon_supercell_symmetry = self._symmetry
else:
self._phonon_supercell_symmetry = Symmetry(
self._phonon_supercell,
symprec=self._symprec,
is_symmetry=self._is_symmetry,
)
def _build_supercell(self):
self._supercell = get_supercell(
self._unitcell, self._supercell_matrix, symprec=self._symprec
)
def _build_primitive_cell(self):
"""Create primitive cell.
primitive_matrix:
Relative axes of primitive cell to the input unit cell.
Relative axes to the supercell is calculated by:
supercell_matrix^-1 * primitive_matrix
Therefore primitive cell lattice is finally calculated by:
(supercell_lattice * (supercell_matrix)^-1 * primitive_matrix)^T
"""
self._primitive = self._get_primitive_cell(
self._supercell, self._supercell_matrix, self._primitive_matrix
)
def _build_phonon_supercell(self):
"""Create phonon supercell for fc2.
phonon_supercell:
This supercell is used for harmonic phonons (frequencies,
eigenvectors, group velocities, ...)
phonon_supercell_matrix:
Different supercell size can be specified.
"""
if self._phonon_supercell_matrix is None:
self._phonon_supercell = self._supercell
else:
self._phonon_supercell = get_supercell(
self._unitcell, self._phonon_supercell_matrix, symprec=self._symprec
)
def _build_phonon_primitive_cell(self):
if self._phonon_supercell_matrix is None:
self._phonon_primitive = self._primitive
else:
self._phonon_primitive = self._get_primitive_cell(
self._phonon_supercell,
self._phonon_supercell_matrix,
self._primitive_matrix,
)
if (
self._primitive is not None
and (self._primitive.numbers != self._phonon_primitive.numbers).any()
):
print(" Primitive cells for fc2 and fc3 can be different.")
raise RuntimeError
def _build_phonon_supercells_with_displacements(
self, supercell: PhonopyAtoms, dataset
):
supercells = []
positions = supercell.positions
magmoms = supercell.magnetic_moments
masses = supercell.masses
numbers = supercell.numbers
lattice = supercell.cell
if "displacements" in dataset:
for disp in dataset["displacements"]:
supercells.append(
PhonopyAtoms(
numbers=numbers,
masses=masses,
magnetic_moments=magmoms,
positions=positions + disp,
cell=lattice,
)
)
else:
for disp1 in dataset["first_atoms"]:
disp_cart1 = disp1["displacement"]
positions = supercell.positions
positions[disp1["number"]] += disp_cart1
supercells.append(
PhonopyAtoms(
numbers=numbers,
masses=masses,
magnetic_moments=magmoms,
positions=positions,
cell=lattice,
)
)
return supercells
def _build_supercells_with_displacements(self):
magmoms = self._supercell.magnetic_moments
masses = self._supercell.masses
numbers = self._supercell.numbers
lattice = self._supercell.cell
supercells = self._build_phonon_supercells_with_displacements(
self._supercell, self._dataset
)
if "first_atoms" in self._dataset:
for disp1 in self._dataset["first_atoms"]:
disp_cart1 = disp1["displacement"]
for disp2 in disp1["second_atoms"]:
if "included" in disp2:
included = disp2["included"]
else:
included = True
if included:
positions = self._supercell.positions
positions[disp1["number"]] += disp_cart1
positions[disp2["number"]] += disp2["displacement"]
supercells.append(
PhonopyAtoms(
numbers=numbers,
masses=masses,
magnetic_moments=magmoms,
positions=positions,
cell=lattice,
)
)
else:
supercells.append(None)
self._supercells_with_displacements = supercells
def _get_primitive_cell(
self, supercell, supercell_matrix, primitive_matrix
) -> Primitive:
inv_supercell_matrix = np.linalg.inv(supercell_matrix)
if primitive_matrix is None:
t_mat = inv_supercell_matrix
else:
t_mat = np.dot(inv_supercell_matrix, primitive_matrix)
return get_primitive(supercell, t_mat, self._symprec, store_dense_svecs=True)
def _determine_primitive_matrix(self, primitive_matrix):
pmat = get_primitive_matrix(primitive_matrix, symprec=self._symprec)
if isinstance(pmat, str) and pmat == "auto":
return guess_primitive_matrix(self._unitcell, symprec=self._symprec)
else:
return pmat
def _set_mesh_numbers(
self,
mesh: Union[int, float, Sequence, NDArray],
):
# initialization related to mesh
self._interaction = None
self._bz_grid = BZGrid(
mesh,
lattice=self._primitive.cell,
symmetry_dataset=self._primitive_symmetry.dataset,
is_time_reversal=self._is_symmetry,
use_grg=self._use_grg,
force_SNF=False,
SNF_coordinates=self._SNF_coordinates,
store_dense_gp_map=True,
)
def _init_dynamical_matrix(self):
if self._interaction is None:
msg = (
"Phono3py.init_phph_interaction has to be called "
"before running this method."
)
raise RuntimeError(msg)
self._interaction.init_dynamical_matrix(
self._fc2,
self._phonon_supercell,
self._phonon_primitive,
nac_params=self._nac_params,
)
freqs, _, _ = self.get_phonon_data()
gp_Gamma = self._bz_grid.gp_Gamma
if np.sum(freqs[gp_Gamma] < self._cutoff_frequency) < 3:
for i, f in enumerate(freqs[gp_Gamma, :3]):
if not (f < self._cutoff_frequency):
freqs[gp_Gamma, i] = 0
print("=" * 26 + " Warning " + "=" * 26)
print(
" Phonon frequency of band index %d at Gamma "
"is calculated to be %f." % (i + 1, f)
)
print(" But this frequency is forced to be zero.")
print("=" * 61)
def _get_forces_energies(
self, target: Literal["forces", "supercell_energies"]
) -> NDArray | None:
"""Return fc3 forces and supercell energies.
Return None if tagert data is not found rather than raising exception.
"""
if self._dataset is None:
return None
if not forces_in_dataset(self._dataset):
return None
if target in self._dataset: # type-2
return self._dataset[target]
elif "first_atoms" in self._dataset: # type-1
num_scells = len(self._dataset["first_atoms"])
for disp1 in self._dataset["first_atoms"]:
num_scells += len(disp1["second_atoms"])
if target == "forces":
values = np.zeros(
(num_scells, len(self._supercell), 3),
dtype="double",
order="C",
)
type1_target = "forces"
elif target == "supercell_energies":
values = np.zeros(num_scells, dtype="double")
type1_target = "supercell_energy"
count = 0
for disp1 in self._dataset["first_atoms"]:
values[count] = disp1[type1_target]
count += 1
for disp1 in self._dataset["first_atoms"]:
for disp2 in disp1["second_atoms"]:
values[count] = disp2[type1_target]
count += 1
return values
return None
def _set_forces_energies(
self, values, target: Literal["forces", "supercell_energies"]
):
if "first_atoms" in self._dataset: # type-1
count = 0
for disp1 in self._dataset["first_atoms"]:
if target == "forces":
disp1[target] = np.array(values[count], dtype="double", order="C")
elif target == "supercell_energies":
disp1["supercell_energy"] = float(values[count])
count += 1
for disp1 in self._dataset["first_atoms"]:
for disp2 in disp1["second_atoms"]:
if target == "forces":
disp2[target] = np.array(
values[count], dtype="double", order="C"
)
elif target == "supercell_energies":
disp2["supercell_energy"] = float(values[count])
count += 1
elif "displacements" in self._dataset or "forces" in self._dataset: # type-2
self._dataset[target] = np.array(values, dtype="double", order="C")
else:
raise RuntimeError("Set of FC3 displacements is not available.")
def _get_phonon_forces_energies(
self, target: Literal["forces", "supercell_energies"]
) -> NDArray | None:
"""Return fc2 forces and supercell energies.
Return None if tagert data is not found rather than raising exception.
"""
if self._phonon_dataset is None:
raise RuntimeError("Dataset for fc2does not exist.")
if target in self._phonon_dataset: # type-2
return self._phonon_dataset[target]
elif "first_atoms" in self._phonon_dataset: # type-1
values = []
for disp in self._phonon_dataset["first_atoms"]:
if target == "forces":
if target in disp:
values.append(disp[target])
elif target == "supercell_energies":
if "supercell_energy" in disp:
values.append(disp["supercell_energy"])
if values:
return np.array(values, dtype="double", order="C")
return None
def _set_phonon_forces_energies(
self, values, target: Literal["forces", "supercell_energies"]
):
if self._phonon_dataset is None:
raise RuntimeError("Dataset for fc2 does not exist.")
if "first_atoms" in self._phonon_dataset:
for disp, v in zip(self._phonon_dataset["first_atoms"], values):
if target == "forces":
disp[target] = np.array(v, dtype="double", order="C")
elif target == "supercell_energies":
disp["supercell_energy"] = float(v)
elif "displacements" in self._phonon_dataset:
_values = np.array(values, dtype="double", order="C")
natom = len(self._phonon_supercell)
ndisps = len(self._phonon_dataset["displacements"])
if target == "forces" and (
_values.ndim != 3 or _values.shape != (ndisps, natom, 3)
):
raise RuntimeError(f"Array shape of input {target} is incorrect.")
elif target == "supercell_energies":
if _values.ndim != 1 or _values.shape != (ndisps,):
raise RuntimeError(f"Array shape of input {target} is incorrect.")
self._phonon_dataset[target] = _values
else:
raise RuntimeError("Set of FC2 displacements is not available.")
def _generate_random_displacements(
self,
number_of_snapshots: int,
number_of_atoms: int,
distance: float = 0.03,
is_plusminus: bool = False,
random_seed: Optional[int] = None,
max_distance: Optional[float] = None,
):
if random_seed is not None and random_seed >= 0 and random_seed < 2**32:
_random_seed = random_seed
dataset = {"random_seed": _random_seed}
else:
_random_seed = None
dataset = {}
d = get_random_displacements_dataset(
number_of_snapshots,
number_of_atoms,
distance,
random_seed=_random_seed,
is_plusminus=is_plusminus,
max_distance=max_distance,
)
dataset["displacements"] = d
return dataset
def _check_mlp_dataset(self, mlp_dataset: dict):
if not isinstance(mlp_dataset, dict):
raise TypeError("mlp_dataset has to be a dictionary.")
if "displacements" not in mlp_dataset:
raise RuntimeError("Displacements have to be given.")
if "forces" not in mlp_dataset:
raise RuntimeError("Forces have to be given.")
if "supercell_energy" in mlp_dataset:
raise RuntimeError("Supercell energies have to be given.")
if len(mlp_dataset["displacements"]) != len(mlp_dataset["forces"]):
raise RuntimeError("Length of displacements and forces are different.")
if len(mlp_dataset["displacements"]) != len(mlp_dataset["supercell_energies"]):
raise RuntimeError(
"Length of displacements and supercell_energies are different."
)
Reference in New Issue View Git Blame Copy Permalink