-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathtest.py
201 lines (176 loc) · 8.27 KB
/
test.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
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
import argparse
from transformers import AutoTokenizer, AutoModelForCausalLM
from distutils.ccompiler import new_compiler
import pandas as pd
from model import Text2Mol
from dataset import MoleculeDataset, MoleculeGraphDataset
from torch.utils.data import DataLoader
from torch import nn
import numpy as np
import torch
from transformers import BertTokenizer, AutoTokenizer
from utils import calculate_similarity, is_valid_smiles, set_seed
from NoamOpt import NoamOpt
import json
import selfies as sfs
from rdkit import Chem, rdBase
import pickle
import random
import wandb
# Disable error messages from rdkit
import warnings
warnings.filterwarnings('ignore')
from rdkit import RDLogger
lg = RDLogger.logger()
lg.setLevel(RDLogger.CRITICAL)
rdBase.DisableLog('rdApp.error')
def main(args):
set_seed(42)
if torch.backends.mps.is_available():
mac_run = True
else:
mac_run = False
run_name = "text2mol_" + args.text_encoder + "_" + args.molecule_decoder+ "_" + args.dataset_name + "_test"
if args.use_wandb:
print("Using wandb for logging.")
try:
wandb.login(key=args.wandb_key)
except:
pass
run = wandb.init(project=run_name, config=args)
# run = wandb.init(project=run_name, config={"epochs": 100, "learning_rate": 0.005, "batch_size": 8}, name="Trained on CheBI-20 ChemT5 MolGen7B " + dataset_name)
else:
print("Not using wandb for logging.")
corrected_smiles = []
corrected_selfies = []
corrected_corpus = []
if args.dataset_name == "ChEBI-20":
import csv
with open("ChEBI-20/test.txt") as f:
reader = csv.DictReader(f, delimiter="\t", quoting=csv.QUOTE_NONE, fieldnames=['cid', 'SMILES', 'description'])
for n, line in enumerate(reader):
try:
smiles = Chem.MolToSmiles(Chem.MolFromSmiles(line['SMILES']))
selfies = sfs.encoder(smiles)
except:
continue
corrected_smiles.append(smiles)
corrected_selfies.append(selfies)
corrected_corpus.append(line['description'])
test_data = list(zip(corrected_corpus, corrected_selfies, corrected_smiles))
elif args.dataset_name == "PubChem_filtered":
with open("PubChem/PubChem_filtered.json", "r") as f:
dataset = json.load(f)
for i in range(len(dataset)):
try:
smiles = Chem.MolToSmiles(Chem.MolFromSmiles(dataset[i]['smiles']))
selfies = sfs.encoder(smiles)
except Exception as error:
print(error)
continue
corrected_smiles.append(smiles)
corrected_selfies.append(selfies)
corrected_corpus.append(dataset[i]['input'])
data = list(zip(corrected_corpus, corrected_selfies, corrected_smiles))
test_data = random.sample(data, 3000)
elif args.dataset_name == "PubChem_unfiltered":
with open("PubChem/PubChem_unfiltered.pkl", "rb") as f:
dataset = pickle.load(f)
corrected_corpus = [dataset[i]['description'] for i in range(len(dataset))]
corrected_selfies = [dataset[i]['SELFIES'] for i in range(len(dataset))]
corrected_smiles = [dataset[i]['SMILES'] for i in range(len(dataset))]
data = list(zip(corrected_corpus, corrected_selfies, corrected_smiles))
random.shuffle(data)
test_data = data[int(len(data) * 0.99):]
molecule_tokenizer = AutoTokenizer.from_pretrained("zjunlp/MolGen-Large")
cls_idx = molecule_tokenizer.cls_token_id
eos_idx = molecule_tokenizer.eos_token_id
mask_idx = molecule_tokenizer.mask_token_id
pad_idx = molecule_tokenizer.pad_token_id
model = Text2Mol(args.text_encoder, args.molecule_decoder)
# Total parameters
total_params = sum(p.numel() for p in model.parameters())
# Trainable parameters
trainable_params = sum(p.numel() for p in model.parameters() if p.requires_grad)
# Untrainable parameters
untrainable_params = total_params - trainable_params
print(f"Total Parameters: {total_params}")
print(f"Trainable Parameters: {trainable_params}")
print(f"Untrainable Parameters: {untrainable_params}")
device = torch.device("mps") if not torch.cuda.is_available() else torch.device("cuda:0")
model_weight = torch.load('model_parameters_' + args.text_encoder + "_" + args.molecule_decoder + "_" + args.dataset_name + '.pth', map_location="cpu")
if args.freeze_encoder != True:
model.encoder.load_state_dict(model_weight['encoder'], strict=False)
model.structural_adapter_attn.load_state_dict(model_weight['attn'], strict=False)
model.structural_adapter_ffn.load_state_dict(model_weight['ffn'], strict=False)
model.to(device)
batch_size = args.batch_size
test_dataset = MoleculeDataset(test_data)
test_dataloader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False, drop_last=False)
descriptions = []
ground_truth = []
output_smiles = []
model.eval()
with torch.no_grad():
test_invalid_selfies = 0
test_similarities = []
for batch_idx, batch in enumerate(test_dataloader):
input = batch
selfies = input['selfies']
input['description'] = ["Write in SMILES the described molecule: " + i for i in input['description']]
input['description'] = [i.replace(" with data available.", ".") for i in input['description']]
output, sequence = model.sample_ar(input, temp=1, cls_idx=cls_idx, greedy=False)
eos_indices = []
for g in sequence:
eos_position = torch.nonzero(g == eos_idx, as_tuple=True)[0]
if len(eos_position) > 0:
first_eos_index = eos_position[0]
else:
first_eos_index = g.shape[0] - 1
eos_indices.append(first_eos_index)
sequence_np = sequence.detach().cpu().numpy()
selfies_output = []
for i in range(sequence_np.shape[0]):
line = sequence_np[i][:eos_indices[i]]
selfies_output.append(molecule_tokenizer.decode(line))
descriptions.extend(input['description'])
for i in range(output.shape[0]):
ground_truth.append(input['smiles'][i])
try:
smiles_output = sfs.decoder(selfies_output[i])
except:
output_smiles.append("None")
test_invalid_selfies += 1
continue
if is_valid_smiles(smiles_output) and smiles_output != "":
output_smiles.append(smiles_output)
similarity = calculate_similarity(smiles_output, input['smiles'][i])
if similarity is None:
test_invalid_selfies += 1
continue
test_similarities.append(similarity)
else:
output_smiles.append("None")
test_invalid_selfies += 1
if args.use_wandb:
wandb.log({"test_similarity": np.mean(test_similarities),
"training invalid selfies number": test_invalid_selfies,
"Validity": len(test_similarities)/(len(test_similarities)+test_invalid_selfies)})
with open(args.text_encoder + "_" + args.molecule_decoder +"_" + args.dataset_name + ".txt", 'w') as f:
f.write('description' + '\t' + 'ground truth' + '\t' + 'output' + '\n')
for desc, rt, ot in zip(descriptions, ground_truth, output_smiles):
desc = desc.replace("\n", "")
f.write(desc + '\t' + rt + '\t' + ot + '\n')
if args.use_wandb:
wandb.finish()
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument("--dataset_name", type=str, default="CheBI-20")
parser.add_argument("--batch_size", type=int, default=16)
parser.add_argument("--text_encoder", type=str, default="ChemT5")
parser.add_argument("--molecule_decoder", type=str, default="MolGen")
parser.add_argument("--freeze_encoder", type=bool, default=True)
parser.add_argument("--use_wandb", type=bool, default=False)
parser.add_argument("--wandb_key", type=str)
args = parser.parse_args()
main(args)