-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathgen_pre.py
116 lines (93 loc) · 3.74 KB
/
gen_pre.py
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
# -*- coding: UTF-8 -*-
import torch
import numpy as np
import matplotlib
from utils import process
matplotlib.use('Agg')
from matplotlib import pyplot as plt
from rdkit import Chem
from rdkit.Chem import AllChem
from rdkit import DataStructs
import pandas as pd
from pandas import DataFrame
def tok(ms, word2idx):
all_ids = []
all_smiles = process(ms)
i = 0
for smiles in all_smiles:
i += 1
id = []
words = smiles
if len(words) < 141:
for word in words:
id += [word2idx[word]]
while len(id) < 141:
id += [0]
all_ids.append(id)
else:
print(i, words)
return torch.LongTensor(all_ids)
def predict(ms, pt_path):
word2idx, idx2word = torch.load("data/opv_dic.pt")
data = tok(ms, word2idx)
model = torch.load(pt_path, map_location=torch.device('cpu'))
model.eval()
out = []
i = 0
for input in data.unsqueeze(1):
out += [model(input[:, :-1]).detach().item()]
i += 1
return out
def draw(x, y, z, x_range, y_range, label_x, label_y, label_z,save_name):
plt.figure()
plt.scatter(x, y, c=z, cmap='rainbow')
plt.xlim(x_range)
plt.ylim(y_range)
plt.axis("square")
plt.tick_params(direction='in', labelsize=16)
plt.xlabel(label_x, {'weight': 'normal', 'size': 20})
plt.ylabel(label_y, {'weight': 'normal', 'size': 20})
plt.colorbar().set_label(label_z)
figure_fig = plt.gcf()
plt.tight_layout()
figure_fig.savefig(save_name, format='eps', dpi=300)
def getfp(smiles):
ms=[Chem.MolFromSmiles(smile) for smile in smiles]
fps=[]
for i in range(len(ms)):
fps+=[AllChem.GetMorganFingerprint(ms[i],2)]
return fps
if __name__ == '__main__':
with open('results/5sample.txt', 'r') as f:
ms = f.readlines()[:10000]
fps = getfp(ms)
h_pred = predict(ms, "results/saved_models/50pre.pt")
l_pred = predict(ms, "results/saved_models/51pre.pt")
p_pred = predict(ms, "results/saved_models/54pre.pt")
draw(h_pred, l_pred, p_pred, x_range=(-7.4, -5.5), y_range=(-4.9, -1.8),
label_x='$HOMO(eV)$', label_y='$LUMO(eV)$', label_z='PCE(%)', save_name='results/saved_models/figure7-1.eps')
ori_df = pd.read_csv('data/opv.csv')
with open('data/smi_c.txt', 'r') as f:
ms0 = [m[:-1] for m in f.readlines()]
fps0 = getfp(ms0)
simi = []
nn_list = []
for i in range(len(ms)):
simi_list = [DataStructs.DiceSimilarity(fps[i], fp0) for fp0 in fps0]
max_simi = max(simi_list)
simi += [max_simi * 100]
nn_list += [ms0[simi_list.index(max_simi)]]
nn_data = []
for nn in nn_list:
i = ori_df[ori_df['smiles'].isin([nn])]['index']
nn_data.append([ori_df.loc[int(i), s] for s in ['homo','lumo','PCE']])
draw(np.array(h_pred), np.array(nn_data)[:,0], np.array(simi),x_range=(-8, -5), y_range=(-8, -5),
label_x='$HOMO_{gen}(eV)$', label_y='$HOMO_{nn}(eV)$', label_z='Similarity(%)', save_name='results/figure8-1.eps')
draw(np.array(l_pred), np.array(nn_data)[:,1], np.array(simi),x_range=(-6, -2), y_range=(-6, -2),
label_x='$LUMO_{gen}(eV)$', label_y='$LUMO_{nn}(eV)$', label_z='Similarity(%)', save_name='results/figure8-2.eps')
draw(np.array(p_pred), np.array(nn_data)[:, 2], np.array(simi), x_range = (0, 11), y_range = (0, 11),
label_x = '$PCE_{gen}(eV)$', label_y = '$PCE_{nn}(eV)$', label_z = 'Similarity(%)', save_name = 'results/figure8-3.eps')
dic={"SMILES":ms,"HOMO_pred":h_pred,"LUMO_pred":l_pred,"PCE_pred":p_pred,"Similarity":simi,
"HOMO_nn":np.array(nn_data)[:,0],"LUMO_nn":np.array(nn_data)[:,1],
"PCE_nn":np.array(nn_data)[:,2],"SMILES_nn":nn_list}
DataFrame(dic).to_csv('results/Gen.csv')