# coding: utf-8
# Copyright (c) Pymatgen Development Team.
# Distributed under the terms of the MIT License.
from __future__ import division, unicode_literals
"""
This module provides so-called "strategies" to determine the coordination environments of an atom in a structure.
Some strategies can favour larger or smaller environments. Some strategies uniquely identifies the environments while
some others can identify the environment as a "mix" of several environments, each of which is assigned with a given
fraction. The choice of the strategy depends on the purpose of the user.
"""
__author__ = "David Waroquiers"
__copyright__ = "Copyright 2012, The Materials Project"
__credits__ = "Geoffroy Hautier"
__version__ = "2.0"
__maintainer__ = "David Waroquiers"
__email__ = "david.waroquiers@gmail.com"
__date__ = "Feb 20, 2016"
import abc
import os
import json
from monty.json import MSONable
from pymatgen.symmetry.analyzer import SpacegroupAnalyzer
from pymatgen.core.operations import SymmOp
from pymatgen.core.sites import PeriodicSite
import numpy as np
from scipy.stats import gmean
from pymatgen.analysis.chemenv.coordination_environments.coordination_geometries import UNCLEAR_ENVIRONMENT_SYMBOL
from pymatgen.analysis.chemenv.utils.coordination_geometry_utils import get_lower_and_upper_f
from pymatgen.analysis.chemenv.utils.func_utils import CSMFiniteRatioFunction
from pymatgen.analysis.chemenv.utils.func_utils import CSMInfiniteRatioFunction
from pymatgen.analysis.chemenv.utils.func_utils import DeltaCSMRatioFunction
from pymatgen.analysis.chemenv.utils.chemenv_errors import EquivalentSiteSearchError
from pymatgen.analysis.chemenv.coordination_environments.coordination_geometries import AllCoordinationGeometries
from pymatgen.analysis.chemenv.utils.defs_utils import AdditionalConditions
from six import with_metaclass
from pymatgen.analysis.chemenv.coordination_environments.voronoi import DetailedVoronoiContainer
from collections import OrderedDict
module_dir = os.path.dirname(os.path.abspath(__file__))
MPSYMBOL_TO_CN = AllCoordinationGeometries().get_symbol_cn_mapping()
ALLCG = AllCoordinationGeometries()
[docs]class StrategyOption(with_metaclass(abc.ABCMeta, MSONable)):
allowed_values = None
[docs] @abc.abstractmethod
def as_dict(self):
"""
A JSON serializable dict representation of this strategy option.
"""
pass
[docs]class DistanceCutoffFloat(float, StrategyOption):
allowed_values = 'Real number between 1.0 and +infinity'
def __new__(cls, myfloat):
flt = float.__new__(cls, myfloat)
if flt < 1.0:
raise ValueError("Distance cutoff should be between 1.0 and +infinity")
return flt
[docs] def as_dict(self):
return {'@module': self.__class__.__module__,
'@class': self.__class__.__name__,
'value': self}
[docs] @classmethod
def from_dict(cls, d):
return cls(d['value'])
[docs]class AngleCutoffFloat(float, StrategyOption):
allowed_values = 'Real number between 0.0 and 1.0'
def __new__(cls, myfloat):
flt = float.__new__(cls, myfloat)
if flt < 0.0 or flt > 1.0:
raise ValueError("Angle cutoff should be between 0.0 and 1.0")
return flt
[docs] def as_dict(self):
return {'@module': self.__class__.__module__,
'@class': self.__class__.__name__,
'value': self}
[docs] @classmethod
def from_dict(cls, d):
return cls(d['value'])
[docs]class CSMFloat(float, StrategyOption):
allowed_values = 'Real number between 0.0 and 100.0'
def __new__(cls, myfloat):
flt = float.__new__(cls, myfloat)
if flt < 0.0 or flt > 100.0:
raise ValueError("Continuous symmetry measure limits should be between 0.0 and 100.0")
return flt
[docs] def as_dict(self):
return {'@module': self.__class__.__module__,
'@class': self.__class__.__name__,
'value': self}
[docs] @classmethod
def from_dict(cls, d):
return cls(d['value'])
[docs]class AdditionalConditionInt(int, StrategyOption):
allowed_values = 'Integer amongst :\n'
for integer, description in AdditionalConditions.CONDITION_DESCRIPTION.items():
allowed_values += ' - {:d} for "{}"\n'.format(integer, description)
def __new__(cls, integer):
intger = int.__new__(cls, integer)
if intger not in AdditionalConditions.ALL:
raise ValueError("Additional condition {:d} is not allowed".format(integer))
return intger
[docs] def as_dict(self):
return {'@module': self.__class__.__module__,
'@class': self.__class__.__name__,
'value': self}
[docs] @classmethod
def from_dict(cls, d):
return cls(d['value'])
[docs]class AbstractChemenvStrategy(with_metaclass(abc.ABCMeta, MSONable)):
"""
Class used to define a Chemenv strategy for the neighbors and coordination environment to be applied to a
StructureEnvironments object
"""
AC = AdditionalConditions()
STRATEGY_OPTIONS = OrderedDict()
STRATEGY_DESCRIPTION = None
STRATEGY_INFO_FIELDS = []
DEFAULT_SYMMETRY_MEASURE_TYPE = 'csm_wcs_ctwcc'
def __init__(self, structure_environments=None, symmetry_measure_type=DEFAULT_SYMMETRY_MEASURE_TYPE):
"""
Abstract constructor for the all chemenv strategies.
:param structure_environments: StructureEnvironments object containing all the information on the
coordination of the sites in a structure
"""
self.structure_environments = None
if structure_environments is not None:
self.set_structure_environments(structure_environments)
self._symmetry_measure_type = symmetry_measure_type
@property
def symmetry_measure_type(self):
return self._symmetry_measure_type
[docs] def set_structure_environments(self, structure_environments):
self.structure_environments = structure_environments
if not isinstance(self.structure_environments.voronoi, DetailedVoronoiContainer):
raise ValueError('Voronoi Container not of type "DetailedVoronoiContainer"')
self.prepare_symmetries()
[docs] def prepare_symmetries(self):
try:
self.spg_analyzer = SpacegroupAnalyzer(self.structure_environments.structure)
self.symops = self.spg_analyzer.get_symmetry_operations()
except:
self.symops = []
[docs] def equivalent_site_index(self, psite):
return self.equivalent_site_index_and_transform(psite)[0]
#return self.structure_environments.structure.index(psite.to_unit_cell)
[docs] def apply_strategy(self):
"""
Applies the strategy to the structure_environments object in order to define the coordination environments of
the sites in the structure_environments object
:param structure_environments: StructureEnvironments object containing all the information needed to define the
coordination environment of the sites in the structure
:return: The coordination environment(s) of the sites
"""
raise NotImplementedError()
[docs] @abc.abstractmethod
def get_site_neighbors(self, site):
"""
Applies the strategy to the structure_environments object in order to get the neighbors of a given site.
:param site: Site for which the neighbors are looked for
:param structure_environments: StructureEnvironments object containing all the information needed to get the
neighbors of the site
:return: The list of neighbors of the site. For complex strategies, where one allows multiple solutions, this
can return a list of list of neighbors
"""
raise NotImplementedError()
@property
def uniquely_determines_coordination_environments(self):
"""
Returns True if the strategy leads to a unique coordination environment, False otherwise.
:return: True if the strategy leads to a unique coordination environment, False otherwise.
"""
raise NotImplementedError()
[docs] @abc.abstractmethod
def get_site_coordination_environment(self, site):
"""
Applies the strategy to the structure_environments object in order to define the coordination environment of
a given site.
:param site: Site for which the coordination environment is looked for
:return: The coordination environment of the site. For complex strategies, where one allows multiple
solutions, this can return a list of coordination environments for the site
"""
raise NotImplementedError()
[docs] @abc.abstractmethod
def get_site_coordination_environments(self, site):
"""
Applies the strategy to the structure_environments object in order to define the coordination environment of
a given site.
:param site: Site for which the coordination environment is looked for
:return: The coordination environment of the site. For complex strategies, where one allows multiple
solutions, this can return a list of coordination environments for the site
"""
raise NotImplementedError()
[docs] @abc.abstractmethod
def get_site_coordination_environments_fractions(self, site, isite=None, dequivsite=None, dthissite=None,
mysym=None, ordered=True, min_fraction=0.0, return_maps=True,
return_strategy_dict_info=False):
"""
Applies the strategy to the structure_environments object in order to define the coordination environment of
a given site.
:param site: Site for which the coordination environment is looked for
:return: The coordination environment of the site. For complex strategies, where one allows multiple
solutions, this can return a list of coordination environments for the site
"""
raise NotImplementedError()
[docs] def get_site_ce_fractions_and_neighbors(self, site, full_ce_info=False, strategy_info=False):
"""
Applies the strategy to the structure_environments object in order to get coordination environments, their
fraction, csm, geometry_info, and neighbors
:param site: Site for which the above information is seeked
:return: The list of neighbors of the site. For complex strategies, where one allows multiple solutions, this
can return a list of list of neighbors
"""
[isite, dequivsite, dthissite, mysym] = self.equivalent_site_index_and_transform(site)
geoms_and_maps_list = self.get_site_coordination_environments_fractions(site=site, isite=isite,
dequivsite=dequivsite,
dthissite=dthissite, mysym=mysym,
return_maps=True,
return_strategy_dict_info=True)
if geoms_and_maps_list is None:
return None
site_nbs_sets = self.structure_environments.neighbors_sets[isite]
ce_and_neighbors = []
for fractions_dict in geoms_and_maps_list:
ce_map = fractions_dict['ce_map']
ce_nb_set = site_nbs_sets[ce_map[0]][ce_map[1]]
neighbors = [{'site': nb_site_and_index['site'],
'index': nb_site_and_index['index']}
for nb_site_and_index in ce_nb_set.neighb_sites_and_indices]
fractions_dict['neighbors'] = neighbors
ce_and_neighbors.append(fractions_dict)
return ce_and_neighbors
[docs] def set_option(self, option_name, option_value):
self.__setattr__(option_name, option_value)
[docs] def setup_options(self, all_options_dict):
for option_name, option_value in all_options_dict.items():
self.set_option(self.STRATEGY_OPTIONS[option_name]['internal'], option_value)
@abc.abstractmethod
def __eq__(self, other):
"""
Equality method that should be implemented for any strategy
:param other: strategy to be compared with the current one
:return:
"""
raise NotImplementedError()
def __str__(self):
out = ' Chemenv Strategy "{}"\n'.format(self.__class__.__name__)
out += ' {}\n\n'.format('='*(19+len(self.__class__.__name__)))
out += ' Description :\n {}\n'.format('-'*13)
out += self.STRATEGY_DESCRIPTION
out += '\n\n'
out += ' Options :\n {}\n'.format('-'*9)
for option_name, option_dict in self.STRATEGY_OPTIONS.items():
out += ' - {} : {}\n'.format(option_name, str(getattr(self, option_name)))
return out
[docs] @abc.abstractmethod
def as_dict(self):
"""
Bson-serializable dict representation of the SimplestChemenvStrategy object.
:return: Bson-serializable dict representation of the SimplestChemenvStrategy object.
"""
raise NotImplementedError()
[docs] @classmethod
def from_dict(cls, d):
"""
Reconstructs the SimpleAbundanceChemenvStrategy object from a dict representation of the
SimpleAbundanceChemenvStrategy object created using the as_dict method.
:param d: dict representation of the SimpleAbundanceChemenvStrategy object
:return: StructureEnvironments object
"""
raise NotImplementedError()
[docs]class SimplestChemenvStrategy(AbstractChemenvStrategy):
"""
Simplest ChemenvStrategy using fixed angle and distance parameters for the definition of neighbors in the
Voronoi approach. The coordination environment is then given as the one with the lowest continuous symmetry measure
"""
# Default values for the distance and angle cutoffs
DEFAULT_DISTANCE_CUTOFF = 1.4
DEFAULT_ANGLE_CUTOFF = 0.3
DEFAULT_CONTINUOUS_SYMMETRY_MEASURE_CUTOFF = 10.0
DEFAULT_ADDITIONAL_CONDITION = AbstractChemenvStrategy.AC.ONLY_ACB
STRATEGY_OPTIONS = OrderedDict({'distance_cutoff': {'type': DistanceCutoffFloat, 'internal': '_distance_cutoff',
'default': DEFAULT_DISTANCE_CUTOFF},
'angle_cutoff': {'type': AngleCutoffFloat, 'internal': '_angle_cutoff',
'default': DEFAULT_ANGLE_CUTOFF},
'additional_condition': {'type': AdditionalConditionInt,
'internal': '_additional_condition',
'default': DEFAULT_ADDITIONAL_CONDITION},
'continuous_symmetry_measure_'
'cutoff': {'type': CSMFloat,
'internal': '_continuous_symmetry_measure_cutoff',
'default': DEFAULT_CONTINUOUS_SYMMETRY_MEASURE_CUTOFF}})
STRATEGY_DESCRIPTION = ' Simplest ChemenvStrategy using fixed angle and distance parameters \n' \
' for the definition of neighbors in the Voronoi approach. \n' \
' The coordination environment is then given as the one with the \n' \
' lowest continuous symmetry measure.'
def __init__(self, structure_environments=None, distance_cutoff=DEFAULT_DISTANCE_CUTOFF,
angle_cutoff=DEFAULT_ANGLE_CUTOFF, additional_condition=DEFAULT_ADDITIONAL_CONDITION,
continuous_symmetry_measure_cutoff=DEFAULT_CONTINUOUS_SYMMETRY_MEASURE_CUTOFF,
symmetry_measure_type=AbstractChemenvStrategy.DEFAULT_SYMMETRY_MEASURE_TYPE):
"""
Constructor for this SimplestChemenvStrategy.
:param distance_cutoff: Distance cutoff used
:param angle_cutoff: Angle cutoff used
"""
AbstractChemenvStrategy.__init__(self, structure_environments, symmetry_measure_type=symmetry_measure_type)
self._distance_cutoff = DistanceCutoffFloat(distance_cutoff)
self._angle_cutoff = AngleCutoffFloat(angle_cutoff)
self._additional_condition = AdditionalConditionInt(additional_condition)
self._continuous_symmetry_measure_cutoff = CSMFloat(continuous_symmetry_measure_cutoff)
@property
def uniquely_determines_coordination_environments(self):
return True
@property
def distance_cutoff(self):
return self._distance_cutoff
@property
def angle_cutoff(self):
return self._angle_cutoff
@property
def additional_condition(self):
return self._additional_condition
@property
def continuous_symmetry_measure_cutoff(self):
return self._continuous_symmetry_measure_cutoff
[docs] def apply_strategy(self):
return None
[docs] def get_site_neighbors(self, site, isite=None, dequivsite=None, dthissite=None, mysym=None):#, neighbors_map=None):
#if neighbors_map is not None:
# return self.structure_environments.voronoi.get_neighbors(isite=isite, neighbors_map=neighbors_map)
if isite is None:
[isite, dequivsite, dthissite, mysym] = self.equivalent_site_index_and_transform(site)
ce, cn_map = self.get_site_coordination_environment(site=site, isite=isite,
dequivsite=dequivsite, dthissite=dthissite, mysym=mysym,
return_map=True)
nb_set = self.structure_environments.neighbors_sets[isite][cn_map[0]][cn_map[1]]
eqsite_ps = nb_set.neighb_sites
coordinated_neighbors = []
for ips, ps in enumerate(eqsite_ps):
coords = mysym.operate(ps.frac_coords + dequivsite) + dthissite
ps_site = PeriodicSite(ps._species, coords, ps._lattice)
coordinated_neighbors.append(ps_site)
return coordinated_neighbors
[docs] def get_site_coordination_environment(self, site, isite=None, dequivsite=None, dthissite=None, mysym=None,
return_map=False):
if isite is None:
[isite, dequivsite, dthissite, mysym] = self.equivalent_site_index_and_transform(site)
neighbors_normalized_distances = self.structure_environments.voronoi.neighbors_normalized_distances[isite]
neighbors_normalized_angles = self.structure_environments.voronoi.neighbors_normalized_angles[isite]
idist = None
for iwd, wd in enumerate(neighbors_normalized_distances):
if self.distance_cutoff >= wd['min']:
idist = iwd
else:
break
iang = None
for iwa, wa in enumerate(neighbors_normalized_angles):
if self.angle_cutoff <= wa['max']:
iang = iwa
else:
break
if idist is None or iang is None:
raise ValueError('Distance or angle parameter not found ...')
my_cn = None
my_inb_set = None
found = False
for cn, nb_sets in self.structure_environments.neighbors_sets[isite].items():
for inb_set, nb_set in enumerate(nb_sets):
sources = [src for src in nb_set.sources
if src['origin'] == 'dist_ang_ac_voronoi' and src['ac'] == self.additional_condition]
for src in sources:
if src['idp'] == idist and src['iap'] == iang:
my_cn = cn
my_inb_set = inb_set
found = True
break
if found:
break
if found:
break
if not found:
return None
cn_map = (my_cn, my_inb_set)
ce = self.structure_environments.ce_list[self.structure_environments.sites_map[isite]][cn_map[0]][cn_map[1]]
if ce is None:
return None
coord_geoms = ce.coord_geoms
if return_map:
if coord_geoms is None:
return cn_map[0], cn_map
return (ce.minimum_geometry(symmetry_measure_type=self._symmetry_measure_type), cn_map)
else:
if coord_geoms is None:
return cn_map[0]
return ce.minimum_geometry(symmetry_measure_type=self._symmetry_measure_type)
[docs] def get_site_coordination_environments_fractions(self, site, isite=None, dequivsite=None, dthissite=None,
mysym=None, ordered=True, min_fraction=0.0, return_maps=True,
return_strategy_dict_info=False):
if isite is None or dequivsite is None or dthissite is None or mysym is None:
[isite, dequivsite, dthissite, mysym] = self.equivalent_site_index_and_transform(site)
site_nb_sets = self.structure_environments.neighbors_sets[isite]
if site_nb_sets is None:
return None
ce_and_map = self.get_site_coordination_environment(site=site, isite=isite, dequivsite=dequivsite,
dthissite=dthissite, mysym=mysym,
return_map=True)
if ce_and_map is None:
return None
ce, ce_map = ce_and_map
if ce is None:
ce_dict = {'ce_symbol': 'UNKNOWN:{:d}'.format(ce_map[0]), 'ce_dict': None, 'ce_fraction': 1.0}
else:
ce_dict = {'ce_symbol': ce[0], 'ce_dict': ce[1], 'ce_fraction': 1.0}
if return_maps:
ce_dict['ce_map'] = ce_map
if return_strategy_dict_info:
ce_dict['strategy_info'] = {}
fractions_info_list = [ce_dict]
return fractions_info_list
[docs] def get_site_coordination_environments(self, site, isite=None, dequivsite=None, dthissite=None, mysym=None,
return_maps=False):
return [self.get_site_coordination_environment(site=site, isite=isite, dequivsite=dequivsite,
dthissite=dthissite, mysym=mysym, return_map=return_maps)]
[docs] def add_strategy_visualization_to_subplot(self, subplot, visualization_options=None, plot_type=None):
subplot.plot(self._distance_cutoff, self._angle_cutoff, 'o', mec=None, mfc='w', markersize=12)
subplot.plot(self._distance_cutoff, self._angle_cutoff, 'x', linewidth=2, markersize=12)
def __eq__(self, other):
return (self.__class__.__name__ == other.__class__.__name__ and
self._distance_cutoff == other._distance_cutoff and self._angle_cutoff == other._angle_cutoff and
self._additional_condition == other._additional_condition and
self._continuous_symmetry_measure_cutoff == other._continuous_symmetry_measure_cutoff and
self.symmetry_measure_type == other.symmetry_measure_type)
[docs] def as_dict(self):
"""
Bson-serializable dict representation of the SimplestChemenvStrategy object.
:return: Bson-serializable dict representation of the SimplestChemenvStrategy object.
"""
return {"@module": self.__class__.__module__,
"@class": self.__class__.__name__,
"distance_cutoff": float(self._distance_cutoff),
"angle_cutoff": float(self._angle_cutoff),
"additional_condition": int(self._additional_condition),
"continuous_symmetry_measure_cutoff": float(self._continuous_symmetry_measure_cutoff),
"symmetry_measure_type": self._symmetry_measure_type}
[docs] @classmethod
def from_dict(cls, d):
"""
Reconstructs the SimplestChemenvStrategy object from a dict representation of the SimplestChemenvStrategy object
created using the as_dict method.
:param d: dict representation of the SimplestChemenvStrategy object
:return: StructureEnvironments object
"""
return cls(distance_cutoff=d["distance_cutoff"], angle_cutoff=d["angle_cutoff"],
additional_condition=d["additional_condition"],
continuous_symmetry_measure_cutoff=d["continuous_symmetry_measure_cutoff"],
symmetry_measure_type=d["symmetry_measure_type"])
[docs]class SimpleAbundanceChemenvStrategy(AbstractChemenvStrategy):
"""
Simple ChemenvStrategy using the neighbors that are the most "abundant" in the grid of angle and distance
parameters for the definition of neighbors in the Voronoi approach.
The coordination environment is then given as the one with the lowest continuous symmetry measure
"""
DEFAULT_MAX_DIST = 2.0
DEFAULT_ADDITIONAL_CONDITION = AbstractChemenvStrategy.AC.ONLY_ACB
STRATEGY_OPTIONS = OrderedDict({'additional_condition': {'type': AdditionalConditionInt,
'internal': '_additional_condition',
'default': DEFAULT_ADDITIONAL_CONDITION},
'surface_calculation_type': {}})
STRATEGY_DESCRIPTION = ' Simple Abundance ChemenvStrategy using the most "abundant" neighbors map \n' \
' for the definition of neighbors in the Voronoi approach. \n' \
' The coordination environment is then given as the one with the \n' \
' lowest continuous symmetry measure.'
def __init__(self, structure_environments=None,
additional_condition=AbstractChemenvStrategy.AC.ONLY_ACB,
symmetry_measure_type=AbstractChemenvStrategy.DEFAULT_SYMMETRY_MEASURE_TYPE):
"""
Constructor for the SimpleAbundanceChemenvStrategy.
:param structure_environments: StructureEnvironments object containing all the information on the
coordination of the sites in a structure
"""
AbstractChemenvStrategy.__init__(self, structure_environments, symmetry_measure_type=symmetry_measure_type)
self._additional_condition = additional_condition
@property
def uniquely_determines_coordination_environments(self):
return True
[docs] def get_site_neighbors(self, site):
[isite, dequivsite, dthissite, mysym] = self.equivalent_site_index_and_transform(site)
cn_map = self._get_map(isite)
eqsite_ps = (self.structure_environments.unique_coordinated_neighbors(isite, cn_map=cn_map))
coordinated_neighbors = []
for ips, ps in enumerate(eqsite_ps):
coords = mysym.operate(ps.frac_coords + dequivsite) + dthissite
ps_site = PeriodicSite(ps._species, coords, ps._lattice)
coordinated_neighbors.append(ps_site)
return coordinated_neighbors
[docs] def get_site_coordination_environment(self, site, isite=None, dequivsite=None, dthissite=None, mysym=None,
return_map=False):
if isite is None:
[isite, dequivsite, dthissite, mysym] = self.equivalent_site_index_and_transform(site)
cn_map = self._get_map(isite)
if cn_map is None:
return None
coord_geoms = (self.structure_environments.
ce_list[self.structure_environments.sites_map[isite]][cn_map[0]][cn_map[1]])
if return_map:
if coord_geoms is None:
return cn_map[0], cn_map
return coord_geoms.minimum_geometry(symmetry_measure_type=self._symmetry_measure_type), cn_map
else:
if coord_geoms is None:
return cn_map[0]
return coord_geoms.minimum_geometry(symmetry_measure_type=self._symmetry_measure_type)
[docs] def get_site_coordination_environments(self, site, isite=None, dequivsite=None, dthissite=None, mysym=None,
return_maps=False):
return [self.get_site_coordination_environment(site=site, isite=isite, dequivsite=dequivsite,
dthissite=dthissite, mysym=mysym, return_map=return_maps)]
def _get_map(self, isite):
maps_and_surfaces = self._get_maps_surfaces(isite)
if maps_and_surfaces is None:
return None
surface_max = 0.0
imax = -1
for ii, map_and_surface in enumerate(maps_and_surfaces):
all_additional_conditions = [ac[2] for ac in map_and_surface['parameters_indices']]
if self._additional_condition in all_additional_conditions and map_and_surface['surface'] > surface_max:
surface_max = map_and_surface['surface']
imax = ii
return maps_and_surfaces[imax]['map']
def _get_maps_surfaces(self, isite, surface_calculation_type=None):
if surface_calculation_type is None:
surface_calculation_type = {'distance_parameter': ('initial_normalized', None),
'angle_parameter': ('initial_normalized', None)}
return self.structure_environments.voronoi.maps_and_surfaces(isite=isite,
surface_calculation_type=surface_calculation_type,
max_dist=self.DEFAULT_MAX_DIST)
def __eq__(self, other):
return (self.__class__.__name__ == other.__class__.__name__ and
self._additional_condition == other.additional_condition)
[docs] def as_dict(self):
"""
Bson-serializable dict representation of the SimpleAbundanceChemenvStrategy object.
:return: Bson-serializable dict representation of the SimpleAbundanceChemenvStrategy object.
"""
return {"@module": self.__class__.__module__,
"@class": self.__class__.__name__,
"additional_condition": self._additional_condition}
[docs] @classmethod
def from_dict(cls, d):
"""
Reconstructs the SimpleAbundanceChemenvStrategy object from a dict representation of the
SimpleAbundanceChemenvStrategy object created using the as_dict method.
:param d: dict representation of the SimpleAbundanceChemenvStrategy object
:return: StructureEnvironments object
"""
return cls(additional_condition=d["additional_condition"])
[docs]class TargettedPenaltiedAbundanceChemenvStrategy(SimpleAbundanceChemenvStrategy):
"""
Simple ChemenvStrategy using the neighbors that are the most "abundant" in the grid of angle and distance
parameters for the definition of neighbors in the Voronoi approach, with a bias for a given list of target
environments. This can be useful in the case of, e.g. connectivity search of some given environment.
The coordination environment is then given as the one with the lowest continuous symmetry measure
"""
DEFAULT_TARGET_ENVIRONMENTS = ['O:6']
def __init__(self, structure_environments=None, truncate_dist_ang=True,
additional_condition=AbstractChemenvStrategy.AC.ONLY_ACB,
max_nabundant=5, target_environments=DEFAULT_TARGET_ENVIRONMENTS, target_penalty_type='max_csm',
max_csm=5.0, symmetry_measure_type=AbstractChemenvStrategy.DEFAULT_SYMMETRY_MEASURE_TYPE):
SimpleAbundanceChemenvStrategy.__init__(self, structure_environments,
additional_condition=additional_condition,
symmetry_measure_type=symmetry_measure_type)
self.max_nabundant = max_nabundant
self.target_environments = target_environments
self.target_penalty_type = target_penalty_type
self.max_csm = max_csm
[docs] def get_site_coordination_environment(self, site, isite=None, dequivsite=None, dthissite=None, mysym=None,
return_map=False):
if isite is None:
[isite, dequivsite, dthissite, mysym] = self.equivalent_site_index_and_transform(site)
cn_map = self._get_map(isite)
if cn_map is None:
return None
chemical_environments = (self.structure_environments.ce_list
[self.structure_environments.sites_map[isite]][cn_map[0]][cn_map[1]])
if return_map:
if chemical_environments.coord_geoms is None or len(chemical_environments) == 0:
return cn_map[0], cn_map
return chemical_environments.minimum_geometry(symmetry_measure_type=self._symmetry_measure_type), cn_map
else:
if chemical_environments.coord_geoms is None:
return cn_map[0]
return chemical_environments.minimum_geometry(symmetry_measure_type=self._symmetry_measure_type)
def _get_map(self, isite):
maps_and_surfaces = SimpleAbundanceChemenvStrategy._get_maps_surfaces(self, isite)
if maps_and_surfaces is None:
return SimpleAbundanceChemenvStrategy._get_map(self, isite)
current_map = None
current_target_env_csm = 100.0
surfaces = [map_and_surface['surface'] for map_and_surface in maps_and_surfaces]
order = np.argsort(surfaces)[::-1]
target_cgs = [AllCoordinationGeometries().get_geometry_from_mp_symbol(mp_symbol)
for mp_symbol in self.target_environments]
target_cns = [cg.coordination_number for cg in target_cgs]
for ii in range(min([len(maps_and_surfaces), self.max_nabundant])):
my_map_and_surface = maps_and_surfaces[order[ii]]
mymap = my_map_and_surface['map']
cn = mymap[0]
if cn not in target_cns or cn > 12 or cn == 0:
continue
all_conditions = [params[2] for params in my_map_and_surface['parameters_indices']]
if self._additional_condition not in all_conditions:
continue
cg, cgdict = (self.structure_environments.ce_list
[self.structure_environments.sites_map[isite]]
[mymap[0]][mymap[1]].minimum_geometry(symmetry_measure_type=self._symmetry_measure_type))
if (cg in self.target_environments and cgdict['symmetry_measure'] <= self.max_csm and
cgdict['symmetry_measure'] < current_target_env_csm):
current_map = mymap
current_target_env_csm = cgdict['symmetry_measure']
if current_map is not None:
return current_map
else:
return SimpleAbundanceChemenvStrategy._get_map(self, isite)
@property
def uniquely_determines_coordination_environments(self):
return True
[docs] def as_dict(self):
"""
Bson-serializable dict representation of the TargettedPenaltiedAbundanceChemenvStrategy object.
:return: Bson-serializable dict representation of the TargettedPenaltiedAbundanceChemenvStrategy object.
"""
return {"@module": self.__class__.__module__,
"@class": self.__class__.__name__,
"additional_condition": self._additional_condition,
"max_nabundant": self.max_nabundant,
"target_environments": self.target_environments,
"target_penalty_type": self.target_penalty_type,
"max_csm": self.max_csm}
def __eq__(self, other):
return (self.__class__.__name__ == other.__class__.__name__ and
self._additional_condition == other.additional_condition and
self.max_nabundant == other.max_nabundant and
self.target_environments == other.target_environments and
self.target_penalty_type == other.target_penalty_type and
self.max_csm == other.max_csm)
[docs] @classmethod
def from_dict(cls, d):
"""
Reconstructs the TargettedPenaltiedAbundanceChemenvStrategy object from a dict representation of the
TargettedPenaltiedAbundanceChemenvStrategy object created using the as_dict method.
:param d: dict representation of the TargettedPenaltiedAbundanceChemenvStrategy object
:return: TargettedPenaltiedAbundanceChemenvStrategy object
"""
return cls(additional_condition=d["additional_condition"],
max_nabundant=d["max_nabundant"],
target_environments=d["target_environments"],
target_penalty_type=d["target_penalty_type"],
max_csm=d["max_csm"])
[docs]class NbSetWeight(with_metaclass(abc.ABCMeta, MSONable)):
[docs] @abc.abstractmethod
def as_dict(self):
"""
A JSON serializable dict representation of this neighbors set weight.
"""
pass
[docs] @abc.abstractmethod
def weight(self, nb_set, structure_environments, cn_map=None, additional_info=None):
pass
[docs]class AngleNbSetWeight(NbSetWeight):
def __init__(self, aa=1.0):
self.aa = aa
if self.aa == 1.0:
self.aw = self.angle_sum
else:
self.aw = self.angle_sumn
[docs] def weight(self, nb_set, structure_environments, cn_map=None, additional_info=None):
return self.aw(nb_set=nb_set)
[docs] def angle_sum(self, nb_set):
return np.sum(nb_set.angles) / (4.0 * np.pi)
[docs] def angle_sumn(self, nb_set):
return np.power(self.angle_sum(nb_set=nb_set), self.aa)
def __eq__(self, other):
return self.aa == other.aa
def __ne__(self, other):
return not self == other
[docs] def as_dict(self):
return {"@module": self.__class__.__module__,
"@class": self.__class__.__name__,
"aa": self.aa
}
[docs] @classmethod
def from_dict(cls, dd):
return cls(aa=dd['aa'])
[docs]class NormalizedAngleDistanceNbSetWeight(NbSetWeight):
def __init__(self, average_type, aa, bb):
self.average_type = average_type
if self.average_type == 'geometric':
self.eval = self.gweight
elif self.average_type == 'arithmetic':
self.eval = self.aweight
else:
raise ValueError('Average type is "{}" while it should be '
'"geometric" or "arithmetic"'.format(average_type))
self.aa = aa
self.bb = bb
if self.aa == 0:
if self.bb == 1:
self.fda = self.invdist
elif self.bb == 0:
raise ValueError('Both exponents are 0.')
else:
self.fda = self.invndist
elif self.bb == 0:
if self.aa == 1:
self.fda = self.ang
else:
self.fda = self.angn
else:
if self.aa == 1:
if self.bb == 1:
self.fda = self.anginvdist
else:
self.fda = self.anginvndist
else:
if self.bb == 1:
self.fda = self.angninvdist
else:
self.fda = self.angninvndist
def __eq__(self, other):
return self.average_type == other.average_type and self.aa == other.aa and self.bb == other.bb
def __ne__(self, other):
return not self == other
[docs] def as_dict(self):
return {"@module": self.__class__.__module__,
"@class": self.__class__.__name__,
"average_type": self.average_type,
"aa": self.aa,
"bb": self.bb
}
[docs] @classmethod
def from_dict(cls, dd):
return cls(average_type=dd['average_type'], aa=dd['aa'], bb=dd['bb'])
[docs] def invdist(self, nb_set):
return [1.0 / dist for dist in nb_set.normalized_distances]
[docs] def invndist(self, nb_set):
return [1.0 / dist**self.bb for dist in nb_set.normalized_distances]
[docs] def ang(self, nb_set):
return nb_set.normalized_angles
[docs] def angn(self, nb_set):
return [ang**self.aa for ang in nb_set.normalized_angles]
[docs] def anginvdist(self, nb_set):
nangles = nb_set.normalized_angles
return [nangles[ii] / dist for ii, dist in enumerate(nb_set.normalized_distances)]
[docs] def anginvndist(self, nb_set):
nangles = nb_set.normalized_angles
return [nangles[ii] / dist**self.bb for ii, dist in enumerate(nb_set.normalized_distances)]
[docs] def angninvdist(self, nb_set):
nangles = nb_set.normalized_angles
return [nangles[ii]**self.aa / dist for ii, dist in enumerate(nb_set.normalized_distances)]
[docs] def angninvndist(self, nb_set):
nangles = nb_set.normalized_angles
return [nangles[ii]**self.aa / dist**self.bb for ii, dist in enumerate(nb_set.normalized_distances)]
[docs] def weight(self, nb_set, structure_environments, cn_map=None, additional_info=None):
fda_list = self.fda(nb_set=nb_set)
return self.eval(fda_list=fda_list)
[docs] def gweight(self, fda_list):
return gmean(fda_list)
[docs] def aweight(self, fda_list):
return np.mean(fda_list)
[docs]def get_effective_csm(nb_set, cn_map, structure_environments, additional_info,
symmetry_measure_type, max_effective_csm, effective_csm_estimator_ratio_function):
try:
effective_csm = additional_info['effective_csms'][nb_set.isite][cn_map]
except KeyError:
site_ce_list = structure_environments.ce_list[nb_set.isite]
site_chemenv = site_ce_list[cn_map[0]][cn_map[1]]
if site_chemenv is None:
effective_csm = 100.0
else:
mingeoms = site_chemenv.minimum_geometries(symmetry_measure_type=symmetry_measure_type,
max_csm=max_effective_csm)
if len(mingeoms) == 0:
effective_csm = 100.0
else:
csms = [ce_dict['other_symmetry_measures'][symmetry_measure_type] for mp_symbol, ce_dict in mingeoms
if ce_dict['other_symmetry_measures'][symmetry_measure_type] <= max_effective_csm]
effective_csm = effective_csm_estimator_ratio_function.mean_estimator(csms)
set_info(additional_info=additional_info, field='effective_csms',
isite=nb_set.isite, cn_map=cn_map, value=effective_csm)
return effective_csm
[docs]def set_info(additional_info, field, isite, cn_map, value):
try:
additional_info[field][isite][cn_map] = value
except KeyError:
try:
additional_info[field][isite] = {cn_map: value}
except KeyError:
additional_info[field] = {isite: {cn_map: value}}
[docs]class SelfCSMNbSetWeight(NbSetWeight):
DEFAULT_EFFECTIVE_CSM_ESTIMATOR = {'function': 'power2_inverse_decreasing',
'options': {'max_csm': 8.0}}
DEFAULT_WEIGHT_ESTIMATOR = {'function': 'power2_decreasing_exp',
'options': {'max_csm': 8.0,
'alpha': 1.0}}
DEFAULT_SYMMETRY_MEASURE_TYPE = 'csm_wcs_ctwcc'
def __init__(self, effective_csm_estimator=DEFAULT_EFFECTIVE_CSM_ESTIMATOR,
weight_estimator=DEFAULT_WEIGHT_ESTIMATOR,
symmetry_measure_type=DEFAULT_SYMMETRY_MEASURE_TYPE):
self.effective_csm_estimator = effective_csm_estimator
self.effective_csm_estimator_rf = CSMInfiniteRatioFunction.from_dict(effective_csm_estimator)
self.weight_estimator = weight_estimator
self.weight_estimator_rf = CSMFiniteRatioFunction.from_dict(weight_estimator)
self.symmetry_measure_type = symmetry_measure_type
self.max_effective_csm = self.effective_csm_estimator['options']['max_csm']
[docs] def weight(self, nb_set, structure_environments, cn_map=None, additional_info=None):
effective_csm = get_effective_csm(nb_set=nb_set, cn_map=cn_map,
structure_environments=structure_environments,
additional_info=additional_info,
symmetry_measure_type=self.symmetry_measure_type,
max_effective_csm=self.max_effective_csm,
effective_csm_estimator_ratio_function=self.effective_csm_estimator_rf)
weight = self.weight_estimator_rf.evaluate(effective_csm)
set_info(additional_info=additional_info, field='self_csms_weights', isite=nb_set.isite,
cn_map=cn_map, value=weight)
return weight
def __eq__(self, other):
return (self.effective_csm_estimator == other.effective_csm_estimator and
self.weight_estimator == other.weight_estimator and
self.symmetry_measure_type == other.symmetry_measure_type)
def __ne__(self, other):
return not self == other
[docs] def as_dict(self):
return {"@module": self.__class__.__module__,
"@class": self.__class__.__name__,
"effective_csm_estimator": self.effective_csm_estimator,
"weight_estimator": self.weight_estimator,
"symmetry_measure_type": self.symmetry_measure_type
}
[docs] @classmethod
def from_dict(cls, dd):
return cls(effective_csm_estimator=dd['effective_csm_estimator'],
weight_estimator=dd['weight_estimator'],
symmetry_measure_type=dd['symmetry_measure_type'])
[docs]class DeltaCSMNbSetWeight(NbSetWeight):
DEFAULT_EFFECTIVE_CSM_ESTIMATOR = {'function': 'power2_inverse_decreasing',
'options': {'max_csm': 8.0}}
DEFAULT_SYMMETRY_MEASURE_TYPE = 'csm_wcs_ctwcc'
DEFAULT_WEIGHT_ESTIMATOR = {'function': 'smootherstep',
'options': {'delta_csm_min': 0.5,
'delta_csm_max': 3.0}}
def __init__(self, effective_csm_estimator=DEFAULT_EFFECTIVE_CSM_ESTIMATOR,
weight_estimator=DEFAULT_WEIGHT_ESTIMATOR,
delta_cn_weight_estimators=None,
symmetry_measure_type=DEFAULT_SYMMETRY_MEASURE_TYPE):
self.effective_csm_estimator = effective_csm_estimator
self.effective_csm_estimator_rf = CSMInfiniteRatioFunction.from_dict(effective_csm_estimator)
self.weight_estimator = weight_estimator
if self.weight_estimator is not None:
self.weight_estimator_rf = DeltaCSMRatioFunction.from_dict(weight_estimator)
self.delta_cn_weight_estimators = delta_cn_weight_estimators
self.delta_cn_weight_estimators_rfs = {}
if delta_cn_weight_estimators is not None:
for delta_cn, dcn_w_estimator in delta_cn_weight_estimators.items():
self.delta_cn_weight_estimators_rfs[delta_cn] = DeltaCSMRatioFunction.from_dict(dcn_w_estimator)
self.symmetry_measure_type = symmetry_measure_type
self.max_effective_csm = self.effective_csm_estimator['options']['max_csm']
[docs] def weight(self, nb_set, structure_environments, cn_map=None, additional_info=None):
effcsm = get_effective_csm(nb_set=nb_set, cn_map=cn_map,
structure_environments=structure_environments,
additional_info=additional_info,
symmetry_measure_type=self.symmetry_measure_type,
max_effective_csm=self.max_effective_csm,
effective_csm_estimator_ratio_function=self.effective_csm_estimator_rf)
cn = cn_map[0]
inb_set = cn_map[1]
isite = nb_set.isite
delta_csm = None
delta_csm_cn_map2 = None
nb_set_weight = 1.0
for cn2, nb_sets in structure_environments.neighbors_sets[isite].items():
if cn2 < cn:
continue
for inb_set2, nb_set2 in enumerate(nb_sets):
if cn == cn2 and inb_set == inb_set:
continue
effcsm2 = get_effective_csm(nb_set=nb_set2, cn_map=(cn2, inb_set2),
structure_environments=structure_environments,
additional_info=additional_info,
symmetry_measure_type=self.symmetry_measure_type,
max_effective_csm=self.max_effective_csm,
effective_csm_estimator_ratio_function=self.effective_csm_estimator_rf)
this_delta_csm = effcsm2 - effcsm
if cn2 == cn:
if this_delta_csm < 0.0:
set_info(additional_info=additional_info, field='delta_csms', isite=isite,
cn_map=cn_map, value=this_delta_csm)
set_info(additional_info=additional_info, field='delta_csms_weights', isite=isite,
cn_map=cn_map, value=0.0)
set_info(additional_info=additional_info, field='delta_csms_cn_map2', isite=isite,
cn_map=cn_map, value=(cn2, inb_set2))
return 0.0
else:
dcn = cn2 - cn
if dcn in self.delta_cn_weight_estimators_rfs:
this_delta_csm_weight = self.delta_cn_weight_estimators_rfs[dcn].evaluate(this_delta_csm)
else:
this_delta_csm_weight = self.weight_estimator_rf.evaluate(this_delta_csm)
if this_delta_csm_weight < nb_set_weight:
delta_csm = this_delta_csm
delta_csm_cn_map2 = (cn2, inb_set2)
nb_set_weight = this_delta_csm_weight
set_info(additional_info=additional_info, field='delta_csms', isite=isite,
cn_map=cn_map, value=delta_csm)
set_info(additional_info=additional_info, field='delta_csms_weights', isite=isite,
cn_map=cn_map, value=nb_set_weight)
set_info(additional_info=additional_info, field='delta_csms_cn_map2', isite=isite,
cn_map=cn_map, value=delta_csm_cn_map2)
return nb_set_weight
def __eq__(self, other):
return (self.effective_csm_estimator == other.effective_csm_estimator and
self.weight_estimator == other.weight_estimator and
self.delta_cn_weight_estimators == other.delta_cn_weight_estimators and
self.symmetry_measure_type == other.symmetry_measure_type)
def __ne__(self, other):
return not self == other
[docs] @classmethod
def delta_cn_specifics(cls, delta_csm_mins=None, delta_csm_maxs=None, function='smootherstep',
symmetry_measure_type='csm_wcs_ctwcc',
effective_csm_estimator=DEFAULT_EFFECTIVE_CSM_ESTIMATOR):
if delta_csm_mins is None or delta_csm_maxs is None:
delta_cn_weight_estimators = {dcn: {'function': function,
'options': {'delta_csm_min': 0.25+dcn*0.25,
'delta_csm_max': 5.0+dcn*0.25}} for dcn in range(1, 13)}
else:
delta_cn_weight_estimators = {dcn: {'function': function,
'options': {'delta_csm_min': delta_csm_mins[dcn-1],
'delta_csm_max': delta_csm_maxs[dcn-1]}}
for dcn in range(1, 13)}
return cls(effective_csm_estimator=effective_csm_estimator,
weight_estimator={'function': function,
'options': {'delta_csm_min': delta_cn_weight_estimators[12]
['options']['delta_csm_min'],
'delta_csm_max': delta_cn_weight_estimators[12]
['options']['delta_csm_max']}},
delta_cn_weight_estimators=delta_cn_weight_estimators,
symmetry_measure_type=symmetry_measure_type)
[docs] def as_dict(self):
return {"@module": self.__class__.__module__,
"@class": self.__class__.__name__,
"effective_csm_estimator": self.effective_csm_estimator,
"weight_estimator": self.weight_estimator,
"delta_cn_weight_estimators": self.delta_cn_weight_estimators,
"symmetry_measure_type": self.symmetry_measure_type
}
[docs] @classmethod
def from_dict(cls, dd):
return cls(effective_csm_estimator=dd['effective_csm_estimator'],
weight_estimator=dd['weight_estimator'],
delta_cn_weight_estimators={int(dcn): dcn_estimator
for dcn, dcn_estimator in dd['delta_cn_weight_estimators'].items()}
if ('delta_cn_weight_estimators' in dd and dd['delta_cn_weight_estimators'] is not None) else None,
symmetry_measure_type=dd['symmetry_measure_type'])
[docs]class CNBiasNbSetWeight(NbSetWeight):
def __init__(self, cn_weights, initialization_options):
self.cn_weights = cn_weights
self.initialization_options = initialization_options
[docs] def weight(self, nb_set, structure_environments, cn_map=None, additional_info=None):
return self.cn_weights[len(nb_set)]
def __eq__(self, other):
return (self.cn_weights == other.cn_weights and
self.initialization_options == other.initialization_options)
def __ne__(self, other):
return not self == other
[docs] def as_dict(self):
return {"@module": self.__class__.__module__,
"@class": self.__class__.__name__,
"cn_weights": {str(cn): cnw for cn, cnw in self.cn_weights.items()},
"initialization_options": self.initialization_options,
}
[docs] @classmethod
def from_dict(cls, dd):
return cls(cn_weights={int(cn): cnw for cn, cnw in dd['cn_weights'].items()},
initialization_options=dd['initialization_options'])
[docs] @classmethod
def linearly_equidistant(cls, weight_cn1, weight_cn13):
initialization_options = {'type': 'linearly_equidistant',
'weight_cn1': weight_cn1,
'weight_cn13': weight_cn13
}
dw = (weight_cn13 - weight_cn1) / 12.0
cn_weights = {cn: weight_cn1 + (cn - 1) * dw for cn in range(1, 14)}
return cls(cn_weights=cn_weights, initialization_options=initialization_options)
[docs] @classmethod
def geometrically_equidistant(cls, weight_cn1, weight_cn13):
initialization_options = {'type': 'geometrically_equidistant',
'weight_cn1': weight_cn1,
'weight_cn13': weight_cn13
}
factor = np.power(float(weight_cn13) / weight_cn1, 1.0 / 12.0)
cn_weights = {cn: weight_cn1 * np.power(factor, cn - 1) for cn in range(1, 14)}
return cls(cn_weights=cn_weights, initialization_options=initialization_options)
[docs] @classmethod
def explicit(cls, cn_weights):
initialization_options = {'type': 'explicit'}
if set(cn_weights.keys()) != set(range(1, 14)):
raise ValueError('Weights should be provided for CN 1 to 13')
return cls(cn_weights=cn_weights, initialization_options=initialization_options)
[docs] @classmethod
def from_description(cls, dd):
if dd['type'] == 'linearly_equidistant':
return cls.linearly_equidistant(weight_cn1=dd['weight_cn1'], weight_cn13=dd['weight_cn13'])
elif dd['type'] == 'geometrically_equidistant':
return cls.geometrically_equidistant(weight_cn1=dd['weight_cn1'], weight_cn13=dd['weight_cn13'])
elif dd['type'] == 'explicit':
return cls.explicit(cn_weights=dd['cn_weights'])
[docs]class DistanceAngleAreaNbSetWeight(NbSetWeight):
AC = AdditionalConditions()
DEFAULT_SURFACE_DEFINITION = {'type': 'standard_elliptic',
'distance_bounds': {'lower': 1.2, 'upper': 1.8},
'angle_bounds': {'lower': 0.1, 'upper': 0.8}}
def __init__(self, weight_type='has_intersection', surface_definition=DEFAULT_SURFACE_DEFINITION,
nb_sets_from_hints='fallback_to_source', other_nb_sets='0_weight',
additional_condition=AC.ONLY_ACB, smoothstep_distance=None, smoothstep_angle=None):
self.weight_type = weight_type
if weight_type == 'has_intersection':
self.area_weight = self.w_area_has_intersection
elif weight_type == 'has_intersection_smoothstep':
raise NotImplementedError()
# self.area_weight = self.w_area_has_intersection_smoothstep
else:
raise ValueError('Weight type is "{}" while it should be "has_intersection"'.format(weight_type))
self.surface_definition = surface_definition
self.nb_sets_from_hints = nb_sets_from_hints
self.other_nb_sets = other_nb_sets
self.additional_condition = additional_condition
self.smoothstep_distance = smoothstep_distance
self.smoothstep_angle = smoothstep_angle
if self.nb_sets_from_hints == 'fallback_to_source':
if self.other_nb_sets == '0_weight':
self.w_area_intersection_specific = self.w_area_intersection_nbsfh_fbs_onb0
else:
raise ValueError('Other nb_sets should be "0_weight"')
else:
raise ValueError('Nb_sets from hints should fallback to source')
lower_and_upper_functions = get_lower_and_upper_f(surface_calculation_options=surface_definition)
self.dmin = surface_definition['distance_bounds']['lower']
self.dmax = surface_definition['distance_bounds']['upper']
self.amin = surface_definition['angle_bounds']['lower']
self.amax = surface_definition['angle_bounds']['upper']
self.f_lower = lower_and_upper_functions['lower']
self.f_upper = lower_and_upper_functions['upper']
[docs] def weight(self, nb_set, structure_environments, cn_map=None, additional_info=None):
return self.area_weight(nb_set=nb_set, structure_environments=structure_environments,
cn_map=cn_map, additional_info=additional_info)
[docs] def w_area_has_intersection_smoothstep(self, nb_set, structure_environments,
cn_map, additional_info):
w_area = self.w_area_intersection_specific(nb_set=nb_set, structure_environments=structure_environments,
cn_map=cn_map, additional_info=additional_info)
if w_area > 0.0:
if self.smoothstep_distance is not None:
w_area = w_area
if self.smoothstep_angle is not None:
w_area = w_area
return w_area
[docs] def w_area_has_intersection(self, nb_set, structure_environments,
cn_map, additional_info):
return self.w_area_intersection_specific(nb_set=nb_set, structure_environments=structure_environments,
cn_map=cn_map, additional_info=additional_info)
[docs] def w_area_intersection_nbsfh_fbs_onb0(self, nb_set, structure_environments,
cn_map, additional_info):
dist_ang_sources = [src for src in nb_set.sources
if src['origin'] == 'dist_ang_ac_voronoi' and src['ac'] == self.additional_condition]
if len(dist_ang_sources) > 0:
for src in dist_ang_sources:
d1 = src['dp_dict']['min']
d2 = src['dp_dict']['next']
a1 = src['ap_dict']['next']
a2 = src['ap_dict']['max']
if self.rectangle_crosses_area(d1=d1, d2=d2, a1=a1, a2=a2):
return 1.0
return 0.0
else:
from_hints_sources = [src for src in nb_set.sources if src['origin'] == 'nb_set_hints']
if len(from_hints_sources) == 0:
return 0.0
elif len(from_hints_sources) != 1:
raise ValueError('Found multiple hints sources for nb_set')
else:
cn_map_src = from_hints_sources[0]['cn_map_source']
nb_set_src = structure_environments.neighbors_sets[nb_set.isite][cn_map_src[0]][cn_map_src[1]]
dist_ang_sources = [src for src in nb_set_src.sources
if src['origin'] == 'dist_ang_ac_voronoi' and
src['ac'] == self.additional_condition]
if len(dist_ang_sources) == 0:
return 0.0
for src in dist_ang_sources:
d1 = src['dp_dict']['min']
d2 = src['dp_dict']['next']
a1 = src['ap_dict']['next']
a2 = src['ap_dict']['max']
if self.rectangle_crosses_area(d1=d1, d2=d2, a1=a1, a2=a2):
return 1.0
return 0.0
[docs] def rectangle_crosses_area(self, d1, d2, a1, a2):
# Case 1
if d1 <= self.dmin and d2 <= self.dmin:
return False
# Case 6
if d1 >= self.dmax and d2 >= self.dmax:
return False
# Case 2
if d1 <= self.dmin and d2 <= self.dmax:
ld2 = self.f_lower(d2)
if a2 <= ld2 or a1 >= self.amax:
return False
return True
# Case 3
if d1 <= self.dmin and d2 >= self.dmax:
if a2 <= self.amin or a1 >= self.amax:
return False
return True
# Case 4
if self.dmin <= d1 <= self.dmax and self.dmin <= d2 <= self.dmax:
ld1 = self.f_lower(d1)
ld2 = self.f_lower(d2)
if a2 <= ld1 and a2 <= ld2:
return False
ud1 = self.f_upper(d1)
ud2 = self.f_upper(d2)
if a1 >= ud1 and a1 >= ud2:
return False
return True
# Case 5
if self.dmin <= d1 <= self.dmax and d2 >= self.dmax:
ud1 = self.f_upper(d1)
if a1 >= ud1 or a2 <= self.amin:
return False
return True
raise ValueError('Should not reach this point!')
def __eq__(self, other):
return (self.weight_type == other.weight_type and
self.surface_definition == other.surface_definition and
self.nb_sets_from_hints == other.nb_sets_from_hints and
self.other_nb_sets == other.other_nb_sets and
self.additional_condition == other.additional_condition
)
def __ne__(self, other):
return not self == other
[docs] def as_dict(self):
return {"@module": self.__class__.__module__,
"@class": self.__class__.__name__,
"weight_type": self.weight_type,
"surface_definition": self.surface_definition,
"nb_sets_from_hints": self.nb_sets_from_hints,
"other_nb_sets": self.other_nb_sets,
"additional_condition": self.additional_condition}
[docs] @classmethod
def from_dict(cls, dd):
return cls(weight_type=dd['weight_type'], surface_definition=dd['surface_definition'],
nb_sets_from_hints=dd['nb_sets_from_hints'], other_nb_sets=dd['other_nb_sets'],
additional_condition=dd['additional_condition'])
[docs]class MultiWeightsChemenvStrategy(AbstractChemenvStrategy):
"""
MultiWeightsChemenvStrategy
"""
STRATEGY_DESCRIPTION = ' Multi Weights ChemenvStrategy'
# STRATEGY_INFO_FIELDS = ['cn_map_surface_fraction', 'cn_map_surface_weight',
# 'cn_map_mean_csm', 'cn_map_csm_weight',
# 'cn_map_delta_csm', 'cn_map_delta_csms_cn_map2', 'cn_map_delta_csm_weight',
# 'cn_map_cn_weight',
# 'cn_map_fraction', 'cn_map_ce_fraction', 'ce_fraction']
DEFAULT_CE_ESTIMATOR = {'function': 'power2_inverse_power2_decreasing',
'options': {'max_csm': 8.0}}
DEFAULT_DIST_ANG_AREA_WEIGHT = {}
def __init__(self, structure_environments=None,
additional_condition=AbstractChemenvStrategy.AC.ONLY_ACB,
symmetry_measure_type=AbstractChemenvStrategy.DEFAULT_SYMMETRY_MEASURE_TYPE,
dist_ang_area_weight=None,
self_csm_weight=None,
delta_csm_weight=None,
cn_bias_weight=None,
angle_weight=None,
normalized_angle_distance_weight=None,
ce_estimator=DEFAULT_CE_ESTIMATOR
):
"""
Constructor for the MultiWeightsChemenvStrategy.
:param structure_environments: StructureEnvironments object containing all the information on the
coordination of the sites in a structure
"""
AbstractChemenvStrategy.__init__(self, structure_environments, symmetry_measure_type=symmetry_measure_type)
self._additional_condition = additional_condition
self.dist_ang_area_weight = dist_ang_area_weight
self.angle_weight = angle_weight
self.normalized_angle_distance_weight = normalized_angle_distance_weight
self.self_csm_weight = self_csm_weight
self.delta_csm_weight = delta_csm_weight
self.cn_bias_weight = cn_bias_weight
self.ordered_weights = []
if dist_ang_area_weight is not None:
self.ordered_weights.append({'weight': dist_ang_area_weight, 'name': 'DistAngArea'})
if self_csm_weight is not None:
self.ordered_weights.append({'weight': self_csm_weight, 'name': 'SelfCSM'})
if delta_csm_weight is not None:
self.ordered_weights.append({'weight': delta_csm_weight, 'name': 'DeltaCSM'})
if cn_bias_weight is not None:
self.ordered_weights.append({'weight': cn_bias_weight, 'name': 'CNBias'})
if angle_weight is not None:
self.ordered_weights.append({'weight': angle_weight, 'name': 'Angle'})
if normalized_angle_distance_weight is not None:
self.ordered_weights.append({'weight': normalized_angle_distance_weight, 'name': 'NormalizedAngDist'})
self.ce_estimator = ce_estimator
self.ce_estimator_ratio_function = CSMInfiniteRatioFunction.from_dict(self.ce_estimator)
self.ce_estimator_fractions = self.ce_estimator_ratio_function.fractions
[docs] @classmethod
def stats_article_weights_parameters(cls):
self_csm_weight = SelfCSMNbSetWeight(weight_estimator={'function': 'power2_decreasing_exp',
'options': {'max_csm': 8.0,
'alpha': 1.0}})
surface_definition = {'type': 'standard_elliptic',
'distance_bounds': {'lower': 1.15, 'upper': 2.0},
'angle_bounds': {'lower': 0.05, 'upper': 0.75}}
da_area_weight = DistanceAngleAreaNbSetWeight(weight_type='has_intersection',
surface_definition=surface_definition,
nb_sets_from_hints='fallback_to_source',
other_nb_sets='0_weight',
additional_condition=DistanceAngleAreaNbSetWeight.AC.ONLY_ACB)
symmetry_measure_type = 'csm_wcs_ctwcc'
delta_weight = DeltaCSMNbSetWeight.delta_cn_specifics()
bias_weight = None
angle_weight = None
nad_weight = None
return cls(dist_ang_area_weight=da_area_weight,
self_csm_weight=self_csm_weight,
delta_csm_weight=delta_weight,
cn_bias_weight=bias_weight,
angle_weight=angle_weight,
normalized_angle_distance_weight=nad_weight,
symmetry_measure_type=symmetry_measure_type)
@property
def uniquely_determines_coordination_environments(self):
return False
[docs] def get_site_coordination_environments_fractions(self, site, isite=None, dequivsite=None, dthissite=None,
mysym=None, ordered=True, min_fraction=0.0, return_maps=True,
return_strategy_dict_info=False, return_all=False):
if isite is None or dequivsite is None or dthissite is None or mysym is None:
[isite, dequivsite, dthissite, mysym] = self.equivalent_site_index_and_transform(site)
site_nb_sets = self.structure_environments.neighbors_sets[isite]
if site_nb_sets is None:
return None
cn_maps = []
for cn, nb_sets in site_nb_sets.items():
for inb_set, nb_set in enumerate(nb_sets):
#CHECK THE ADDITIONAL CONDITION HERE ?
cn_maps.append((cn, inb_set))
weights_additional_info = {'weights': {isite: {}}}
for wdict in self.ordered_weights:
cn_maps_new = []
weight = wdict['weight']
weight_name = wdict['name']
for cn_map in cn_maps:
nb_set = site_nb_sets[cn_map[0]][cn_map[1]]
w_nb_set = weight.weight(nb_set=nb_set, structure_environments=self.structure_environments,
cn_map=cn_map, additional_info=weights_additional_info)
if cn_map not in weights_additional_info['weights'][isite]:
weights_additional_info['weights'][isite][cn_map] = {}
weights_additional_info['weights'][isite][cn_map][weight_name] = w_nb_set
if w_nb_set > 0.0:
cn_maps_new.append(cn_map)
cn_maps = cn_maps_new
for cn_map, weights in weights_additional_info['weights'][isite].items():
weights_additional_info['weights'][isite][cn_map]['Product'] = np.product(weights.values())
w_nb_sets = {cn_map: weights['Product']
for cn_map, weights in weights_additional_info['weights'][isite].items()}
w_nb_sets_total = np.sum(w_nb_sets.values())
nb_sets_fractions = {cn_map: w_nb_set / w_nb_sets_total for cn_map, w_nb_set in w_nb_sets.items()}
for cn_map in weights_additional_info['weights'][isite]:
weights_additional_info['weights'][isite][cn_map]['NbSetFraction'] = nb_sets_fractions[cn_map]
ce_symbols = []
ce_dicts = []
ce_fractions = []
ce_dict_fractions = []
ce_maps = []
site_ce_list = self.structure_environments.ce_list[isite]
if return_all:
for cn_map, nb_set_fraction in nb_sets_fractions.items():
cn = cn_map[0]
inb_set = cn_map[1]
site_ce_nb_set = site_ce_list[cn][inb_set]
if site_ce_nb_set is None:
continue
mingeoms = site_ce_nb_set.minimum_geometries(symmetry_measure_type=self.symmetry_measure_type)
if len(mingeoms) > 0:
csms = [ce_dict['other_symmetry_measures'][self.symmetry_measure_type]
for ce_symbol, ce_dict in mingeoms]
fractions = self.ce_estimator_fractions(csms)
if fractions is None:
ce_symbols.append('UNCLEAR:{:d}'.format(cn))
ce_dicts.append(None)
ce_fractions.append(nb_set_fraction)
all_weights = weights_additional_info['weights'][isite][cn_map]
dict_fractions = {wname: wvalue for wname, wvalue in all_weights.items()}
dict_fractions['CEFraction'] = None
dict_fractions['Fraction'] = nb_set_fraction
ce_dict_fractions.append(dict_fractions)
ce_maps.append(cn_map)
else:
for ifraction, fraction in enumerate(fractions):
ce_symbols.append(mingeoms[ifraction][0])
ce_dicts.append(mingeoms[ifraction][1])
ce_fractions.append(nb_set_fraction * fraction)
all_weights = weights_additional_info['weights'][isite][cn_map]
dict_fractions = {wname: wvalue for wname, wvalue in all_weights.items()}
dict_fractions['CEFraction'] = fraction
dict_fractions['Fraction'] = nb_set_fraction * fraction
ce_dict_fractions.append(dict_fractions)
ce_maps.append(cn_map)
else:
ce_symbols.append('UNCLEAR:{:d}'.format(cn))
ce_dicts.append(None)
ce_fractions.append(nb_set_fraction)
all_weights = weights_additional_info['weights'][isite][cn_map]
dict_fractions = {wname: wvalue for wname, wvalue in all_weights.items()}
dict_fractions['CEFraction'] = None
dict_fractions['Fraction'] = nb_set_fraction
ce_dict_fractions.append(dict_fractions)
ce_maps.append(cn_map)
else:
for cn_map, nb_set_fraction in nb_sets_fractions.items():
if nb_set_fraction > 0.0:
cn = cn_map[0]
inb_set = cn_map[1]
site_ce_nb_set = site_ce_list[cn][inb_set]
mingeoms = site_ce_nb_set.minimum_geometries(symmetry_measure_type=self._symmetry_measure_type)
csms = [ce_dict['other_symmetry_measures'][self._symmetry_measure_type]
for ce_symbol, ce_dict in mingeoms]
fractions = self.ce_estimator_fractions(csms)
for ifraction, fraction in enumerate(fractions):
if fraction > 0.0:
ce_symbols.append(mingeoms[ifraction][0])
ce_dicts.append(mingeoms[ifraction][1])
ce_fractions.append(nb_set_fraction * fraction)
all_weights = weights_additional_info['weights'][isite][cn_map]
dict_fractions = {wname: wvalue for wname, wvalue in all_weights.items()}
dict_fractions['CEFraction'] = fraction
dict_fractions['Fraction'] = nb_set_fraction * fraction
ce_dict_fractions.append(dict_fractions)
ce_maps.append(cn_map)
if ordered:
indices = np.argsort(ce_fractions)[::-1]
else:
indices = list(range(len(ce_fractions)))
fractions_info_list = [
{'ce_symbol': ce_symbols[ii], 'ce_dict': ce_dicts[ii], 'ce_fraction': ce_fractions[ii]}
for ii in indices if ce_fractions[ii] >= min_fraction]
if return_maps:
for ifinfo, ii in enumerate(indices):
if ce_fractions[ii] >= min_fraction:
fractions_info_list[ifinfo]['ce_map'] = ce_maps[ii]
if return_strategy_dict_info:
for ifinfo, ii in enumerate(indices):
if ce_fractions[ii] >= min_fraction:
fractions_info_list[ifinfo]['strategy_info'] = ce_dict_fractions[ii]
return fractions_info_list
[docs] def get_site_coordination_environment(self, site):
pass
[docs] def get_site_neighbors(self, site):
pass
[docs] def get_site_coordination_environments(self, site, isite=None, dequivsite=None, dthissite=None, mysym=None,
return_maps=False):
if isite is None or dequivsite is None or dthissite is None or mysym is None:
[isite, dequivsite, dthissite, mysym] = self.equivalent_site_index_and_transform(site)
return [self.get_site_coordination_environment(site=site, isite=isite, dequivsite=dequivsite,
dthissite=dthissite, mysym=mysym, return_map=return_maps)]
def __eq__(self, other):
return (self.__class__.__name__ == other.__class__.__name__ and
self._additional_condition == other._additional_condition and
self.symmetry_measure_type == other.symmetry_measure_type and
self.dist_ang_area_weight == other.dist_ang_area_weight and
self.self_csm_weight == other.self_csm_weight and
self.delta_csm_weight == other.delta_csm_weight and
self.cn_bias_weight == other.cn_bias_weight and
self.angle_weight == other.angle_weight and
self.normalized_angle_distance_weight == other.normalized_angle_distance_weight and
self.ce_estimator == other.ce_estimator)
def __ne__(self, other):
return not self == other
[docs] def as_dict(self):
"""
Bson-serializable dict representation of the MultiWeightsChemenvStrategy object.
:return: Bson-serializable dict representation of the MultiWeightsChemenvStrategy object.
"""
return {"@module": self.__class__.__module__,
"@class": self.__class__.__name__,
"additional_condition": self._additional_condition,
"symmetry_measure_type": self.symmetry_measure_type,
"dist_ang_area_weight": self.dist_ang_area_weight.as_dict()
if self.dist_ang_area_weight is not None else None,
"self_csm_weight": self.self_csm_weight.as_dict()
if self.self_csm_weight is not None else None,
"delta_csm_weight": self.delta_csm_weight.as_dict()
if self.delta_csm_weight is not None else None,
"cn_bias_weight": self.cn_bias_weight.as_dict()
if self.cn_bias_weight is not None else None,
"angle_weight": self.angle_weight.as_dict()
if self.angle_weight is not None else None,
"normalized_angle_distance_weight": self.normalized_angle_distance_weight.as_dict()
if self.normalized_angle_distance_weight is not None else None,
"ce_estimator": self.ce_estimator,
}
[docs] @classmethod
def from_dict(cls, d):
"""
Reconstructs the MultiWeightsChemenvStrategy object from a dict representation of the
MultipleAbundanceChemenvStrategy object created using the as_dict method.
:param d: dict representation of the MultiWeightsChemenvStrategy object
:return: MultiWeightsChemenvStrategy object
"""
if d["normalized_angle_distance_weight"] is not None:
nad_w = NormalizedAngleDistanceNbSetWeight.from_dict(d["normalized_angle_distance_weight"])
else:
nad_w = None
return cls(additional_condition=d["additional_condition"],
symmetry_measure_type=d["symmetry_measure_type"],
dist_ang_area_weight=DistanceAngleAreaNbSetWeight.from_dict(d["dist_ang_area_weight"])
if d["dist_ang_area_weight"] is not None else None,
self_csm_weight=SelfCSMNbSetWeight.from_dict(d["self_csm_weight"])
if d["self_csm_weight"] is not None else None,
delta_csm_weight=DeltaCSMNbSetWeight.from_dict(d["delta_csm_weight"])
if d["delta_csm_weight"] is not None else None,
cn_bias_weight=CNBiasNbSetWeight.from_dict(d["cn_bias_weight"])
if d["cn_bias_weight"] is not None else None,
angle_weight=AngleNbSetWeight.from_dict(d["angle_weight"])
if d["angle_weight"] is not None else None,
normalized_angle_distance_weight=nad_w,
ce_estimator=d["ce_estimator"])