-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathget_moleculenet_embeddings.py
168 lines (118 loc) · 6.64 KB
/
get_moleculenet_embeddings.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
import os
from time import time
from fnmatch import fnmatch
import pandas as pd
from pandarallel import pandarallel
import to_selfies
import torch
from transformers import RobertaTokenizer, RobertaModel, RobertaConfig, AutoTokenizer, AutoModel
from dmgi_model import load_heterodata, load_dmgi_model
import argparse
import numpy as np
from unimol_tools import UniMolRepr
parser = argparse.ArgumentParser()
parser.add_argument("--dataset_path", required=True, metavar="/path/to/dataset/", help="Path of the input MoleculeNet datasets.")
parser.add_argument("--model_file", required=True, metavar="<str>", type=str, help="Name of the pretrained model.")
parser.add_argument("--heterodata_path", required=True, metavar="/path/to/heterodata/", help="Path of the input heterodata.")
parser.add_argument("--dmgi_model", required=True, metavar="/path/to/dmgi_model/", help="Path of the input dmgi_model.")
args = parser.parse_args()
os.environ["TOKENIZERS_PARALLELISM"] = "false"
os.environ["WANDB_DISABLED"] = "true"
os.environ["CUDA_VISIBLE_DEVICES"] = "0"
model_file = args.model_file # path of the pre-trained model
config = RobertaConfig.from_pretrained(model_file)
config.output_hidden_states = True
tokenizer = RobertaTokenizer.from_pretrained("./data/RobertaFastTokenizer")
model = RobertaModel.from_pretrained(model_file, config=config)
scibert_tokenizer = AutoTokenizer.from_pretrained("allenai/scibert_scivocab_uncased")
scibert_model = AutoModel.from_pretrained("allenai/scibert_scivocab_uncased")
unimol_model = UniMolRepr(data_type='molecule', remove_hs=False)
heterodata = load_heterodata(args.heterodata_path)
dmgi_model = load_dmgi_model(args.dmgi_model, heterodata)
def generate_moleculenet_selfies(dataset_file):
"""
Generates SELFIES for a given dataset and saves it to a file.
:param dataset_file: path to the dataset file
"""
dataset_name = dataset_file.split("/")[-1].split(".")[0]
print(f'Generating SELFIES for {dataset_name}')
if dataset_name == 'bace':
smiles_column = 'mol'
else:
smiles_column = 'smiles'
# read dataset
dataset_df = pd.read_csv(os.path.join(dataset_file))
dataset_df["selfies"] = dataset_df[smiles_column] # creating a new column "selfies" that is a copy of smiles_column
# generate selfies
pandarallel.initialize()
dataset_df.selfies = dataset_df.selfies.parallel_apply(to_selfies.to_selfies)
dataset_df.drop(dataset_df[dataset_df[smiles_column] == dataset_df.selfies].index, inplace=True)
# dataset_df.drop(columns=[smiles_column], inplace=True)
out_name = dataset_name + "_selfies.csv"
# save selfies to file
path = os.path.dirname(dataset_file)
dataset_df.to_csv(os.path.join(path, out_name), index=False)
print(f'Saved to {os.path.join(path, out_name)}')
def get_sequence_embeddings(selfies, tokenizer, model):
torch.set_num_threads(1)
token = torch.tensor([tokenizer.encode(selfies, add_special_tokens=True, max_length=512, padding=True, truncation=True)])
output = model(token)
sequence_out = output[0]
return torch.mean(sequence_out[0], dim=0).tolist()
def get_text_embeddings(text, tokenizer, model):
torch.set_num_threads(1)
if type(text) == str:
token = torch.tensor([tokenizer.encode(text, add_special_tokens=True, max_length=512, padding=True, truncation=True)])
output = model(token)
text_out = output[0][0].mean(dim=0)
else:
text_out = torch.zeros(768)
return text_out.tolist()
def get_unimol_embeddings(smiles, model):
unimol_repr = model.get_repr(smiles, return_atomic_reprs=True) # UniMolRepr model
# CLS token repr
print(np.array(unimol_repr['cls_repr']).shape)
# atomic level repr, align with rdkit mol.GetAtoms()
print(np.array(unimol_repr['atomic_reprs']).shape)
def get_kg_embeddings(chembl_id, heterodata, dmgi_model):
node_idx = heterodata['Compound'].id_mapping[chembl_id] if chembl_id in heterodata['Compound'].id_mapping else None
if node_idx:
output = dmgi_model.Z[node_idx]
else:
output = torch.zeros(64)
return output.tolist()
def generate_embeddings(model_file, heterodata, dmgi_model, args):
root = args.dataset_path
model_name = model_file.split("/")[-1]
prepare_data_pattern = "*.csv"
for path, subdirs, files in os.walk(root):
for name in files:
if fnmatch(name, prepare_data_pattern) and not any(substring in name for substring in ['selfies', 'embeddings', 'results']):
dataset_file = os.path.join(path, name)
dataset_name = dataset_file.split("/")[-1].split(".")[0]
print(f'-------Processing {dataset_name}-------')
generate_moleculenet_selfies(dataset_file)
selfies_file = os.path.join(path, name.split(".")[0] + "_selfies.csv")
print(f'Generating embeddings for {dataset_name}')
t0 = time()
dataset_df = pd.read_csv(selfies_file)
pandarallel.initialize(nb_workers=10, progress_bar=True) # number of threads
# selformer embeddings
print(f'\n\nGenerating SELFormer embeddings using pre-trained model {model_name}')
dataset_df["sequence_embeddings"] = dataset_df.selfies.parallel_apply(get_sequence_embeddings, args=(tokenizer, model))
# scibert embeddings
print(f'\n\nGenerating SciBERT embeddings')
dataset_df["text_embeddings"] = dataset_df.description.parallel_apply(get_text_embeddings, args=(scibert_tokenizer, scibert_model))
print(f'\n\nGenerating UniMol embeddings')
dataset_df["unimol_embeddings"] = dataset_df.selfies.parallel_apply(get_unimol_embeddings, args=(unimol_model,))
print(f'\n\nGenerating KG embeddings')
dataset_df["kg_embeddings"] = dataset_df.chembl_id.parallel_apply(get_kg_embeddings, args=(heterodata, dmgi_model))
dataset_df.drop(columns=["description", "chembl_id"], inplace=True) # not interested in selfies data anymore, only class and the embedding
file_name = f"{dataset_name}_{model_name}_embeddings.pkl"
# save embeddings to file
path = os.path.dirname(selfies_file)
dataset_df.to_pickle(os.path.join(path, file_name))
t1 = time()
print(f'Finished in {round((t1-t0) / 60, 2)} mins')
print(f'Saved to {os.path.join(path, file_name)}\n')
generate_embeddings(model_file, heterodata, dmgi_model, args)