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
| import numpy as np
import pandas as pd
from sklearn.utils import shuffle
def read_fa(file, mode='train'):
assert mode in {'train','test'}
labels = []
seqs_info = []
cates_id = []
seq = ''
with open(file,mode='r') as f:
line = f.readline().strip()
while line:
if line[0]=='>':
info = line[1:].split(' ')
cates_id.append(info[0])
if mode == 'train':
label = ''.join(info[1].split('.')[:2]
label = label[0]+'.'+label[1:]
labels.append(label)
if seq:
seqs_info.append(seq)
seq = ''
else:
seq += line
line = f.readline().strip()
seqs_info.append(seq)
return cates_id,seqs_info,labels
def load_data():
train_file = '/kaggle/input/textfiles/astral_train.fa'
test_file = '/kaggle/input/textfiles/astral_test.fa'
train_sample_id, train_seqs_info, train_labels = read_fa(train_file, mode='train')
test_sample_id, test_seqs_info, _ = read_fa(test_file, mode='test')
train_data = {
'sample_id': train_sample_id,
'seq_info': train_seqs_info,
'label': train_labels
}
test_data = {
'sample_id': test_sample_id,
'seq_info': test_seqs_info,
}
train = pd.DataFrame(data=train_data)
train = shuffle(train,random_state=2021).reset_index(drop=True)
test = pd.DataFrame(data=test_data)
return train,test
def split_windows(sentence,w = 3):
new_sentence = []
for i in range(len(sentence)-w+1):
new_sentence.append(sentence[i:i+w])
return new_sentence
data = pd.concat(load_data(),ignore_index=True)
label2idx = { l:idx for idx, l in enumerate(data[~data['label'].isna()]['label'].unique().tolist())}
idx2label = { idx:l for l,idx in label2idx.items()}
data['label'] = data['label'].map(label2idx)
data['new_seq_info'] = data['seq_info'].apply(lambda x:split_windows(x,w = 1))
train,test = data[~data['label'].isna()].reset_index(drop=True),data[data['label'].isna()].reset_index(drop=True)
max_features= 1000
max_len= 256
embed_size=128
batch_size = 24
epochs = 50
from keras.preprocessing.text import Tokenizer
from keras.preprocessing import sequence
tokens = Tokenizer(num_words = max_features)
tokens.fit_on_texts(list(data['new_seq_info']))
x_data = tokens.texts_to_sequences(data['new_seq_info'])
x_data = sequence.pad_sequences(x_data, maxlen=max_len)
x_train = x_data[:9472]
y_train = data['label'][:9472]
x_test = x_data[9472:]
from keras.layers import Dense,Input,LSTM,Bidirectional,Activation,Conv1D,GRU
from keras.layers import Dropout,Embedding,GlobalMaxPooling1D, MaxPooling1D, Add, Flatten
from keras.layers import GlobalAveragePooling1D, GlobalMaxPooling1D, concatenate, SpatialDropout1D
from keras.callbacks import Callback
from keras.callbacks import EarlyStopping,ModelCheckpoint
from keras import initializers, regularizers, constraints, optimizers, layers, callbacks
from keras.models import Model
from keras.optimizers import Adam
import keras
sequence_input = Input(shape=(max_len, ))
x = Embedding(max_features, embed_size, trainable=True)(sequence_input)
x = SpatialDropout1D(0.2)(x)
x = Bidirectional(GRU(128, return_sequences=True,dropout=0.1,recurrent_dropout=0.1))(x)
x = Conv1D(64, kernel_size = 3, padding = "valid", kernel_initializer = "glorot_uniform")(x)
avg_pool = GlobalAveragePooling1D()(x)
max_pool = GlobalMaxPooling1D()(x)
x = concatenate([avg_pool, max_pool])
preds = Dense(245)(x)
model = Model(sequence_input, preds)
model.compile(loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True),
optimizer=keras.optimizers.Adam(1e-3),
metrics=['accuracy'])
model.fit(x_train, y_train,
batch_size=batch_size,
validation_split=0.2,
epochs=epochs)
|
