DPFunc/DPFunc_main.py at main · CSUBioGroup/DPFunc · 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
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
from ruamel.yaml import YAML
from logzero import logger
from pathlib import Path
import warnings

import torch
import numpy as np
from dgl.dataloading import GraphDataLoader

from DPFunc.data_utils import get_pdb_data, get_mlb, get_inter_whole_data
from DPFunc.models import combine_inter_model
from DPFunc.objective import AverageMeter
from DPFunc.model_utils import test_performance_gnn_inter, merge_result, FocalLoss
from DPFunc.evaluation import new_compute_performance_deepgoplus

import os
import pickle as pkl
import click
from tqdm.auto import tqdm

@click.command()
@click.option('-d', '--data-cnf', type=click.Choice(['bp', 'mf', 'cc']))
@click.option('-n', '--gpu-number', type=click.INT, default=0)
@click.option('-e', '--epoch-number', type=click.INT, default=15)
@click.option('-p', '--pre-name', type=click.STRING, default='temp_model')


def main(data_cnf, gpu_number, epoch_number, pre_name):
    yaml = YAML(typ='safe')
    ont = data_cnf
    data_cnf, model_cnf = yaml.load(Path('./configure/{}.yaml'.format(data_cnf))), yaml.load(Path('./configure/dgg.yaml'))
    device = torch.device('cuda:{}'.format(gpu_number))

    data_name, model_name = data_cnf['name'], model_cnf['name']
    run_name = F'{model_name}-{data_name}'
    logger.info('run_name: {}'.format(run_name))

    data_cnf['mlb'] = Path(data_cnf['mlb'])
    data_cnf['results'] = Path(data_cnf['results'])
    logger.info(F'Model: {model_name}, Dataset: {data_name}')

    train_pid_list, train_graph, train_go = get_pdb_data(pid_list_file = data_cnf['train']['pid_list_file'],
                                                         pdb_graph_file = data_cnf['train']['pid_pdb_file'],
                                                         pid_go_file = data_cnf['train']['pid_go_file'],
                                                         train = data_cnf['train']['train_file_count'])
    logger.info('train data done')
    valid_pid_list, valid_graph, valid_go = get_pdb_data(pid_list_file = data_cnf['valid']['pid_list_file'],
                                                         pdb_graph_file = data_cnf['valid']['pid_pdb_file'],
                                                         pid_go_file = data_cnf['valid']['pid_go_file'])
    logger.info('valid data done')
    test_pid_list, test_graph, test_go = get_pdb_data(pid_list_file = data_cnf['test']['pid_list_file'],
                                                      pdb_graph_file = data_cnf['test']['pid_pdb_file'],
                                                      pid_go_file = data_cnf['test']['pid_go_file'])
    logger.info('test data done')

    train_interpro = get_inter_whole_data(train_pid_list, data_cnf['base']['interpro_whole'], data_cnf['train']['interpro_file'])
    valid_interpro = get_inter_whole_data(valid_pid_list, data_cnf['base']['interpro_whole'], data_cnf['valid']['interpro_file'])
    test_interpro = get_inter_whole_data(test_pid_list, data_cnf['base']['interpro_whole'], data_cnf['test']['interpro_file'])

    assert len(train_pid_list)==len(train_graph)
    assert len(train_pid_list)==train_interpro.shape[0]
    assert len(train_pid_list)==len(train_go)

    assert len(valid_pid_list)==len(valid_graph)
    assert len(valid_pid_list)==valid_interpro.shape[0]
    assert len(valid_pid_list)==len(valid_go)

    assert len(test_pid_list)==len(test_graph)
    assert len(test_pid_list)==test_interpro.shape[0]
    assert len(test_pid_list)==len(test_go)

    mlb = get_mlb(Path(data_cnf['mlb']), train_go)
    labels_num = len(mlb.classes_)
    with warnings.catch_warnings():
        warnings.simplefilter('ignore')
        train_y = mlb.transform(train_go).astype(np.float32)
        valid_y = mlb.transform(valid_go).astype(np.float32)
        test_y  = mlb.transform(test_go).astype(np.float32)

    idx_goid = {}
    goid_idx = {}
    for idx, goid in enumerate(mlb.classes_):
        idx_goid[idx] = goid
        goid_idx[goid] = idx

    train_data = [(train_graph[i], i, train_y[i]) for i in range(len(train_y))]
    train_dataloader = GraphDataLoader(
        train_data,
        batch_size=64,
        drop_last=False,
        shuffle=True)

    valid_data = [(valid_graph[i], i, valid_y[i]) for i in range(len(valid_y))]
    valid_dataloader = GraphDataLoader(
        valid_data,
        batch_size=64,
        drop_last=False,
        shuffle=False)

    test_data = [(test_graph[i], i, test_y[i]) for i in range(len(test_y))]
    test_dataloader = GraphDataLoader(
        test_data,
        batch_size=64,
        drop_last=False,
        shuffle=False)

    del train_graph
    del test_graph
    del valid_graph

    logger.info('Loading Data & Model')

    model = combine_inter_model(inter_size=train_interpro.shape[1],
                                inter_hid=1280,
                                graph_size=1280,
                                graph_hid=1280,
                                label_num=labels_num, head=4).to(device)
    logger.info(model)
    optimizer = torch.optim.AdamW(model.parameters(), lr = 1e-4)
    loss_fn = FocalLoss()

    used_model_performance = np.array([-1.0]*3)

    for e in range(epoch_number):
        model.train()
        train_loss_vals = AverageMeter()
        for batched_graph, sample_idx, labels in tqdm(train_dataloader, leave=False):
            batched_graph = batched_graph.to(device)
            labels = labels.to(device)
            inter_features = (torch.from_numpy(train_interpro[sample_idx].indices).to(device).long(),
                            torch.from_numpy(train_interpro[sample_idx].indptr).to(device).long(),
                            torch.from_numpy(train_interpro[sample_idx].data).to(device).float())
            feats = batched_graph.ndata['x']

            logits = model(inter_features, batched_graph, feats)

            loss = loss_fn(logits, labels)
            train_loss_vals.update(loss.item(), len(labels))

            optimizer.zero_grad()
            loss.backward()
            optimizer.step()

        plus_fmax, plus_aupr, plus_t, df, valid_loss_avg = test_performance_gnn_inter(model, valid_dataloader, valid_pid_list, valid_interpro, valid_y, idx_goid, goid_idx, ont, device)
        logger.info('Epoch: {}, Train Loss: {:.6f}\tValid Loss: {:.6f}, plus_Fmax on valid: {:.4f}, AUPR on valid: {:.4f}, cut-off: {:.2f}, df_shape: {}'.format(e,
                                                                                                                                train_loss_vals.avg,
                                                                                                                                valid_loss_avg,
                                                                                                                                plus_fmax,
                                                                                                                                plus_aupr,
                                                                                                                                plus_t,
                                                                                                                                df.shape))

        if e > min(used_model_performance):
                replace_ind = np.where(used_model_performance==min(used_model_performance))[0][0]
                used_model_performance[replace_ind] = e
                torch.save({'epoch': e,'model_state_dict': model.state_dict(),
                            'optimizer_state_dict': optimizer.state_dict()},
                        './save_models/{0}_{1}_{2}of{3}model.pt'.format(pre_name, ont, replace_ind, 3))
                logger.info("\t\t\t\t\tSave")

    cob_pred_df = []
    for i_t_min in range(3):
        if os.path.exists('./save_models/{0}_{1}_{2}of{3}model.pt'.format(pre_name, ont, i_t_min, 3)):
            checkpoint = torch.load('./save_models/{0}_{1}_{2}of{3}model.pt'.format(pre_name, ont, i_t_min, 3))
            model.load_state_dict(checkpoint['model_state_dict'])
            pred_df = test_performance_gnn_inter(model, test_dataloader, test_pid_list, test_interpro, test_y, idx_goid, goid_idx, ont, device,
                                        save=True, save_file='./results/{0}_{1}_{2}of{3}model.pkl'.format(pre_name, ont, i_t_min, 3), evaluate=False)
            cob_pred_df.append(pred_df)
            print(i_t_min, 'epoch:', checkpoint['epoch'], pred_df.shape)

    final_result = merge_result(cob_pred_df)
    with open('./results/{}_{}_final.pkl'.format(pre_name, ont), 'wb') as fw:
        pkl.dump(final_result, fw)
    logger.info("Done")
    go_file = './data/go.obo'
    new_fmax, new_aupr, new_t = new_compute_performance_deepgoplus(final_result, go_file, ont)
    logger.info('Final Result: plus_Fmax on test: {:.4f}, AUPR on valid: {:.4f}, cut-off: {:.2f}'.format(new_fmax, new_aupr, new_t))

if __name__ == '__main__':
    main()