Source code for abipy.tools.tensors

# coding: utf-8
"""
This modules provides subclasses of pymatgen tensor objects.
"""
from __future__ import annotations
import numpy as np
import pandas as pd

from pymatgen.core.tensors import Tensor, SquareTensor
from pymatgen.analysis.elasticity.elastic import ElasticTensor  # noqa: F401
from pymatgen.analysis.elasticity.stress import Stress as pmg_Stress
from pymatgen.analysis.piezo import PiezoTensor # noqa: F401
from abipy.iotools import ETSF_Reader


class _Tensor33:

    def _repr_html_(self):
        """Integration with jupyter notebooks."""
        return self.get_dataframe()._repr_html_()

    def get_dataframe(self, tol=1e-3, cmode=None) -> pd.DataFrame:
        """
        Return |pandas-Dataframe| with tensor elements set to zero below `tol`.

        Args:
            cmode: "real" or "imag" to include only the real/imaginary part.
        """
        tensor = self.zeroed(tol=tol)
        if cmode == "real": tensor = tensor.real
        if cmode == "imag": tensor = tensor.imag

        return pd.DataFrame({"x": tensor[:,0], "y": tensor[:,1], "z": tensor[:,2]}, index=["x", "y", "z"])

    def get_voigt_dataframe(self, tol=1e-3) -> pd.DataFrame:
        """
        Return |pandas-DataFrame| with Voigt indices as colums.
        Elements below tol are set to zero.

        Useful to analyze the converge of individual elements.
        """
        tensor = self.zeroed(tol=tol)
        columns = ["xx", "yy", "zz", "yz", "xz", "xy"]
        d = {k: v for k, v in zip(columns, tensor.voigt)}
        return pd.DataFrame(d, index=[0], columns=columns)


[docs] class Stress(pmg_Stress, _Tensor33): """ Stress tensor. rank2 symmetric tensor with shape [3, 3]. .. rubric:: Inheritance Diagram .. inheritance-diagram:: Stress """
[docs] class DielectricTensor(SquareTensor, _Tensor33): """ Subclass of |pmg-Tensor| describing a dielectric tensor. rank2 symmetric tensor with shape [3, 3]. .. rubric:: Inheritance Diagram .. inheritance-diagram:: DielectricTensor """
[docs] def reflectivity(self, n1=1, tol=1e-6) -> pd.DataFrame: """ If the tensor is diagonal (with off diagonal elements smaller than tol) returns the three components of the reflectivity :math:`|n1 - n2| / | n1 + n2 |` """ d = np.diag(self) if np.max(np.abs(self - np.diag(d))) > tol: raise ValueError("The tensor is not diagonal.") n2 = np.sqrt(d) return np.abs((n1 - n2) / (n1 + n2)) ** 2
[docs] class DielectricDataList(list): """ A list of tuples with (DielectricTensor, structure, params) Useful for convergence studies. Example: diel_data = DielectricDataList() diel_data.append((eps0, structure0, params0)) diel_data.append((eps1, structure1, params1)) df = diel_data.get_dataframe() """
[docs] def append(self, obj) -> None: """Extend append method with validation logic.""" if not isinstance(obj, (list, tuple)): raise TypeError(f"Expecting list or tuple but got {type(obj)=}") if len(obj) != 3: raise TypeError(f"Expecting two items but got {len(obj)=}") if not isinstance(obj[0], DielectricTensor): if isinstance(obj[0], np.ndarray): obj = list(obj) obj[0] = DielectricTensor(obj[0]) else: raise TypeError(f"Expecting DielectricTensor instance but got {type(obj[0])=}") from abipy.core.structure import Structure if not isinstance(obj[1], Structure): raise TypeError(f"Expecting Structure instance but got {type(obj[1])=}") if not isinstance(obj[2], dict): raise TypeError(f"Expecting dict instance but got {type(obj[2])=}") return super().append(obj)
@property def eps_list(self) -> list: return [obj[0] for obj in self] @property def structures(self) -> list: return [obj[1] for obj in self] @property def params_list(self) -> list[dict]: return [obj[2] for obj in self]
[docs] def has_same_structure(self) -> bool: """True if all structures are equal.""" if len(self) in (0, 1): return True structures = self.structures structure0 = structures[0] return all(structure0 == s for s in structures[1:])
def __str__(self): return self.get_dataframe().to_string()
[docs] def get_dataframe(self, with_params=True, with_geo=False, with_spglib=True, **kwargs) -> pd.DataFrame: """ Dataframe with the components of eps_infinity. Args: with_params: True to add parameters. with_geo: True to add info on structure. with_params: True to add calculations parameters. kwargs: Optional kwargs passed to add_geo_and_params. """ structures, eps_list, params_list = self.structures, self.eps_list, self.params_list comps2inds = {"xx": (0,0), "yy": (1,1), "zz": (2,2), "xy": (0, 1), "xz": (0, 2), "yx": (1, 0), "yz": (1, 2), "zx": (2, 0), "zy": (2, 1)} rows = [] for structure, eps, params in zip(structures, eps_list, params_list, strict=True): d = {} for k, ind in comps2inds.items(): d[k] = eps[ind] if with_params: d.update(params) if with_geo: geo_dict = structure.get_dict4pandas(with_spglib=with_spglib, **kwargs) d.update(geo_dict) rows.append(d) return pd.DataFrame(rows)
[docs] class ZstarTensor(SquareTensor, _Tensor33): """ Born effective charge tensor (for a single atom). .. rubric:: Inheritance Diagram .. inheritance-diagram:: ZstarTensor """
[docs] class NLOpticalSusceptibilityTensor(Tensor): """ Subclass of |pmg-Tensor| containing the non-linear optical susceptibility tensor. .. rubric:: Inheritance Diagram .. inheritance-diagram:: NLOpticalSusceptibilityTensor """
[docs] @classmethod def from_file(cls, filepath: str) -> NLOpticalSusceptibilityTensor: """ Creates the tensor from an anaddb.nc netcdf file containing ``dchide``. This requires to run anaddb with ``tnlflag`` > 0 """ with ETSF_Reader(filepath) as reader: try: return cls(reader.read_value("dchide")) except Exception as exc: import traceback msg = traceback.format_exc() msg += ("Error while trying to read from file.\n" "Verify that nlflag > 0 in anaddb\n") raise ValueError(msg)