EMGen/opt.py at main · CodingWZL/EMGen · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
from xgboost.sklearn import XGBRegressor
import numpy as np
import os
from bayes_opt import BayesianOptimization
from pymatgen import Lattice, Structure, Composition
from matminer.featurizers.composition import BandCenter
from StructureFeature import structure_symmetry, average_atomic_volume
from ComponentFeature import meredig, band_center

#enreg = XGBRegressor()

def calculate_features(id):
    features = []
    structure = Structure.from_file("../ML/cif/"+str(id) + ".cif")
    ###### Structure Feature ######
    crystal_system, centrosymmetric = structure_symmetry(structure)
    average_atom_volume, average_volume_atom = average_atomic_volume(structure) # dimension:2

    ###### Component Feature ######
    a = structure.composition
    elemental_property = meredig(Composition(a))
    bandcenter = band_center(Composition(a))
    feature = crystal_system
    feature.append(centrosymmetric)
    feature.append(average_atom_volume)
    feature.append(average_volume_atom)
    feature = feature + elemental_property + bandcenter

    features = features + feature
    features = np.array(features)
    features = features.reshape(-1,28)
    np.savetxt("features.txt", features)


def main():
    print("****************************************")
    print("*** Welcome to BayesOptimization !!! ***")
    print("* Please input your cif id:")
    cif_id = input()
    calculate_features(id=cif_id)
    print("* Features have been calculated successfully !!!")
    print("* Please input your elements:")
    global elements
    elements = input()
    elements = elements.split(" ")
    global element_number
    element_number = len(elements)
    if element_number == 2:
        print("* Two elements are obtained successfully !!!")
        print("* Please input the fraction range of each element:")
        fraction = input()
        fraction = fraction.split(" ")
        band_opt(fraction)
    elif element_number == 3:
        print("Three elements are obtained successfully !!!")
        print("* Please input the fraction range of each element:")
        fraction = input()
        fraction = fraction.split(" ")
        band_opt(fraction)
    elif element_number == 4:
        print("Four elements are obtained successfully !!!")
    elif element_number == 5:
        print("Five elements are obtained successfully !!!")
    #print("Please input ")


def model(feature):
    enreg = XGBRegressor()
    enreg.load_model('xgb-1.model')
    p_1 = enreg.predict(feature)
    enreg.load_model('xgb-2.model')
    p_2 = enreg.predict(feature)
    enreg.load_model('xgb-3.model')
    p_3 = enreg.predict(feature)
    enreg.load_model('xgb-4.model')
    p_4 = enreg.predict(feature)
    enreg.load_model('xgb-5.model')
    p_5 = enreg.predict(feature)
    enreg.load_model('xgb-6.model')
    p_6 = enreg.predict(feature)
    enreg.load_model('xgb-7.model')
    p_7 = enreg.predict(feature)
    enreg.load_model('xgb-8.model')
    p_8 = enreg.predict(feature)
    enreg.load_model('xgb-9.model')
    p_9 = enreg.predict(feature)
    enreg.load_model('xgb-10.model')
    p_10 = enreg.predict(feature)
    return (p_1+p_2+p_3+p_4+p_5+p_6+p_7+p_8+p_9+p_10)/10


def calculate_band(a,b,c):
    if element_number == 2:
        formula = elements[0]+str(a)+elements[1]+str(b)
    if element_number == 3:
        formula = elements[0]+str(a)+elements[1]+str(b)+elements[2]+str(c)
    composition = Composition(formula)
    features = np.loadtxt("features.txt").reshape(1,28)
    element_property = np.array(meredig(Composition(composition)))
    BC = BandCenter()
    bandcenter = BC.featurize(composition)
    features[0][10:27] = element_property[0:]
    features[0][27] = bandcenter[0]
    result = model(features)
    return result[0]*(-1)

def band_opt(fraction):
    #fraction = [0.9, 1.0, 0.9, 1.0]
    #element_number = 2
    if element_number == 2:
        opt = BayesianOptimization(calculate_band,\
        {'a':(fraction[0],fraction[1]),\
        'b':(fraction[2],fraction[3])})
        opt.maximize(init_points=200, n_iter=20)
    if element_number == 3:
        opt = BayesianOptimization(calculate_band,\
        {'a':(fraction[0],fraction[1]),\
        'b':(fraction[2],fraction[3]),\
        'c':(fraction[4],fraction[5])})
        opt.maximize(init_points=200, n_iter=20)


if __name__ == '__main__':
    main()
    #band_opt()