Source code for matador.utils.pmg_utils

# coding: utf-8
# Distributed under the terms of the MIT License.

""" This file implements some light wrappers to
the pymatgen, via ASE.


import pickle
import os
import getpass
import copy
from typing import Union

from import AseAtomsAdaptor
from pymatgen import MPRester, Structure

from matador.utils.ase_utils import ase2dict, doc2ase
from matador.utils.cell_utils import calc_mp_spacing
from matador.crystal import Crystal


[docs]def get_chemsys(elements, dumpfile=None): """ Scrape the Materials Project for the chemical system specified by elements, e.g. for elements `['A', 'B', 'C']` the query performed is `chemsys='A-B-C' & nelements=3`. Requires interactive user input of their MP API key (unless set by environment variable $PMG_MAPI_KEY). Parameters: elements (list): list of chemical symbols. Keyword arguments: dumpfile (str): optional filename to dump the pickled response after conversion to matador documents. """ cursor = [] count = {} api_key = os.environ.get('PMG_MAPI_KEY', None) if api_key is None: api_key = getpass.getpass('MP API key:').strip() with MPRester(api_key=api_key) as m: data = { 'criteria': { 'chemsys': '-'.join(elements), 'nelements': len(elements), }, 'properties': REQ_PROPERTIES } response = m.query(**data) for result in response: cursor.append(mp2dict(result)) if dumpfile: with open(dumpfile, 'wb') as f: pickle.dump(cursor, f) print('ICSD structures: {}'.format(sum(1 for doc in cursor if 'icsd' in doc))) print('PF structures: {}'.format(sum(1 for doc in cursor if '_mp_pf_ids' in doc))) return cursor, count
[docs]def doc2pmg(doc: Union[dict, Crystal]): """ Converts matador document/Crystal to a pymatgen structure, via ASE. """ ase_atoms = doc2ase(doc) pmg_structure = AseAtomsAdaptor.get_structure(ase_atoms) = {} if isinstance(doc, Crystal):["matador"] = copy.deepcopy(doc._data) else:["matador"] = copy.deepcopy(doc._data) if "_id" in["matador"]["_id"] = str(["_id"]) return pmg_structure
[docs]def pmg2dict(pmg: Structure, as_model=False) -> Union[dict, Crystal]: """ Converts a pymatgen.Structure to a matador document/Crystal. Parameters: pmg (pymatgen.Structure): the structure to convert. Keyword arguments: as_model (bool): if True, return a Crystal instead of a dict. Returns: Union[dict, Crystal]: the converted structure. """ from matador.utils.ase_utils import ase2dict return ase2dict(AseAtomsAdaptor.get_atoms(pmg), as_model=as_model)
[docs]def mp2dict(response): """ Convert a response from pymatgen.MPRester into a matador document, via an ASE atoms object. Expects certain properties to be requested in order to construct a full matador document, e.g. `structure` & `input`. Parameters: response (dict): containing one item of the MPRester response. """ if 'structure' not in response: raise RuntimeError('`structure` key missing, nothing to scrape...') ase_struct = AseAtomsAdaptor.get_atoms(response['structure']) doc = ase2dict(ase_struct) doc['source'] = [] if 'material_id' in response: doc['_mp_id'] = response['material_id'] doc['source'].append('{}'.format(doc['_mp_id'])) if 'pf_ids' in response: if response['pf_ids']: doc['_mp_pf_ids'] = response['pf_ids'] for _id in doc['_mp_pf_ids']: doc['source'].append('pf-{}'.format(_id)) if 'icsd_ids' in response: if response['icsd_ids']: doc['icsd'] = response['icsd_ids'][0] doc['_mp_icsd_ids'] = response['icsd_ids'] if 'formation_energy_per_atom' in response: doc['formation_energy_per_atom'] = response['formation_energy_per_atom'] doc['enthalpy_per_atom'] = response['formation_energy_per_atom'] doc['enthalpy'] = doc['enthalpy_per_atom'] * doc['num_atoms'] doc['formation_energy'] = doc['formation_energy_per_atom'] * doc['num_atoms'] if 'e_above_hull' in response: doc['hull_distance'] = response['e_above_hull'] if 'doi' in response: doc['_mp_doi'] = response['doi'] if 'snl' in response: doc['_mp_references'] = response['snl'].references if 'input' in response: incar = response['input']['incar'] if not isinstance(incar, dict): incar = incar.as_dict() doc['cut_off_energy'] = incar['ENCUT'] doc['xc_functional'] = 'PBE' kpts = response['input']['kpoints'] doc['kpoints_mp_grid'] = kpts.kpts[0] doc['kpoints_mp_offset'] = kpts.kpts_shift doc['pressure'] = 0.0 doc['stress'] = 0.0 doc['kpoints_mp_spacing'] = calc_mp_spacing(doc['lattice_cart'], doc['kpoints_mp_grid']) doc['spin_polarized'] = bool(incar['ISPIN'] - 1) return doc