def data(args):
train_data = data_process.read_files(args.data_path, "train")
test_data = data_process.read_files(args.data_path, "test")
train_feature, train_mask, train_label, train_seq_len = data_process.get_data_label(
args, train_data)
test_feature, test_mask, test_label, test_seq_len = data_process.get_data_label(
args, test_data)
vail_feature, vail_mask, vail_seq_len = data_process.read_vail(
args, "test.txt")
#print(train_feature[0:2],train_mask[0:2],train_seq_len[0:2],train_label[0:2])
#print(test_feature[0:2],test_mask[0:2],test_seq_len[0:2], test_label[0:2])
#转化为tensor
train_feature, train_mask, train_label, train_seq_len = data_process.to_tensor(
train_feature, train_mask, train_label, train_seq_len)
test_feature, test_mask, test_label, test_seq_len = data_process.to_tensor(
test_feature, test_mask, test_label, test_seq_len)
train_set = torch.utils.data.TensorDataset(train_feature, train_mask,
train_label, train_seq_len)
test_set = torch.utils.data.TensorDataset(test_feature, test_mask,
test_label, test_seq_len)
train_iter = torch.utils.data.DataLoader(train_set,
batch_size=args.batch_size,
shuffle=True)
test_iter = torch.utils.data.DataLoader(test_set,
batch_size=args.batch_size,
shuffle=True)
vail_feature, vail_mask, _, vail_seq_len = data_process.to_tensor(
vail_feature, vail_mask, vail_mask, vail_seq_len)
vail_set = torch.utils.data.TensorDataset(vail_feature, vail_mask,
vail_seq_len)
vail_iter = torch.utils.data.DataLoader(vail_set,
batch_size=args.batch_size,
shuffle=False)
return train_iter, test_iter, vail_iter
#%% Main Function
if __name__ == '__main__':
#%% ########## Read Data ##########
print('read data')
ECG_data, ECG_label = read_data.extract_data(
read_data.read_data(Args)) # read data
#%% ########## Data Processing ##########
ECG_data = data_process.cut_out(ECG_data,
Args.len) # cut out the ECG signals
ECG_data = data_process.axis_change(ECG_data) # change the axis
ECG_label = data_process.label_from_0(ECG_label) # label from 0
# split data
train_x, test_x, train_y, test_y = data_process.train_test_split(
ECG_data, ECG_label, trainratio=Args.trainratio, random_state=0)
# change to Tensor
train_x, train_y = data_process.to_tensor(train_x, train_y)
test_x, test_y = data_process.to_tensor(test_x, test_y)
#%% ########## Load Data ##########
# change to dataset
train_dataset = Data.TensorDataset(train_x, train_y)
test_dataset = Data.TensorDataset(test_x, test_y)
# load data
print('load data')
loader_train = Data.DataLoader(
dataset=train_dataset, # torch TensorDataset format
batch_size=Args.batch_size, # mini batch size
shuffle=True, # shuffle or not
num_workers=0, # multi-worker for load data
)
loader_test = Data.DataLoader(test_dataset, Args.batch_size)
del ECG_data, ECG_label, train_x, test_x, train_y, test_y
