3.screen_with_solubility_and_band_gap_criteria.py

import pandas as pd
import timeout_decorator
from pymatgen.analysis.phase_diagram import PhaseDiagram
from pymatgen.core import Composition, Element
from pymatgen.ext.matproj import MPRester
from retrying import retry
from tqdm import tqdm


@retry(stop_max_attempt_number=20)
@timeout_decorator.timeout(300)
def recheck_e_above_hull(material_id, key_element):
    with MPRester(api_key='') as mpr:
        entry = mpr.get_entry_by_material_id(material_id)
        pretty_formula = entry.name
        chemsys = Composition(pretty_formula).chemical_system + '-' + key_element
        # using GGA and GGA+U mixed scheme as default, namely compatible_only=True
        entries = mpr.get_entries_in_chemsys(chemsys)
    
    phase_diagram = PhaseDiagram(entries)
    e_above_hull = phase_diagram.get_e_above_hull(entry)
    # to avoid kind of values -1.77635683940025E-15, and -0.000 after rounded
    if 0 > e_above_hull >= -phase_diagram.numerical_tol:
        e_above_hull = 0.0
        
    return(e_above_hull)


key_element = 'Li'
tieline_dataframe = pd.read_csv('Tables/{element}/tieline_distinct.csv'.format(element=key_element))

# remove noble gas
boolean_gas = tieline_dataframe.pretty_formula.apply(lambda x : True not in [e.is_noble_gas for e in Composition(x).elements])
no_nobel_gas_dataframe = tieline_dataframe[boolean_gas]

# screen phases without the key element
boolean_element = no_nobel_gas_dataframe.pretty_formula.apply(lambda x : Element(key_element) not in Composition(x).elements)
vanishing_solubility_phases_dataframe = no_nobel_gas_dataframe[boolean_element]
vanishing_solubility_phases_dataframe.reset_index(drop=True, inplace=True)
vanishing_solubility_phases_dataframe.to_csv('Tables/{element}/tieline_without_solubility_and_gas.csv'.format(element=key_element), index=False)

# pick up those phases with band gap >= 3eV
material_id_list = vanishing_solubility_phases_dataframe['material_id'].to_list()
with MPRester(api_key='') as mpr:
    # entry id is an alias of task id
    candidates = mpr.query(criteria={'task_id':{'$in': material_id_list}, 'band_gap':{'$gte':3.0}}, properties=['material_id', 'pretty_formula', 'theoretical', 'band_gap', 'density', 'e_above_hull', 'icsd_ids'])

candidates_dataframe = pd.DataFrame(candidates)

# Due to bug in Materials Project database, e_above_hull needs to be checked
# https://discuss.matsci.org/t/materials-project-database-release-log/1609
# utilize progress_apply instead of apply to show process bar
# https://github.com/tqdm/tqdm

tqdm.pandas(desc="pandas_apply_process_1st")
candidates_dataframe['e_above_hull'] = candidates_dataframe.material_id.progress_apply(recheck_e_above_hull, key_element=key_element)
candidates_dataframe = candidates_dataframe.round({'band_gap':3,'density':3,'e_above_hull':3})
# sort with band gap order, in the meanwhile, make experimental structures shown on top
candidates_dataframe.sort_values(by=['theoretical', 'band_gap'], ascending=[True, False], inplace=True)
candidates_dataframe.to_csv('Tables/{element}/candidates.csv'.format(element=key_element), float_format='%.3f', index=False)