MCCFN/data_loader.py at main · Miraitwo/MCCFN · 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
import pickle
import torch
import numpy as np
import torch.utils.data as Data
import h5py


def Load_Dataset(dataset,
                 logger
                 ):
    if dataset == '2016.10a':
        classes = {b'QAM16': 0, b'QAM64': 1, b'8PSK': 2, b'WBFM': 3, b'BPSK': 4,
                   b'CPFSK': 5, b'AM-DSB': 6, b'GFSK': 7, b'PAM4': 8, b'QPSK': 9, b'AM-SSB': 10}
    elif dataset == '2016.10b':
        classes = {b'QAM16': 0, b'QAM64': 1, b'8PSK': 2, b'WBFM': 3, b'BPSK': 4,
                   b'CPFSK': 5, b'AM-DSB': 6, b'GFSK': 7, b'PAM4': 8, b'QPSK': 9}
    elif dataset == '2016.04c':
        classes = {b'8PSK': 0, b'AM-DSB': 1, b'AM-SSB': 2, b'BPSK': 3, b'CPFSK': 4,
                        b'GFSK': 5, b'PAM4': 6, b'QAM16': 7, b'QAM64': 8, b'QPSK': 9, b'WBFM': 10}
    elif dataset == 'rml22':
        classes = {'QAM16': 0, 'QAM64': 1, '8PSK': 2, 'WBFM': 3, 'BPSK': 4, 'CPFSK': 5, 'AM-DSB': 6, 'GFSK': 7,
              'PAM4': 8, 'QPSK': 9, 'AM-SSB': 10}
        raise NotImplementedError(f'Not Implemented dataset:{dataset}')

    dataset_file = {'2016.10a': 'RML2016.10a_dict.pkl',
                    '2016.10b': 'RML2016.10b.dat',
                    '2016.04c': '2016.04C.multisnr.pkl',
                    'rml22': 'RML22.01A'}

    # file_pointer = './dataset/%s' % dataset_file.get(dataset)
    file_pointer = r'H:\desktop\my_ws\work_station\dataset\%s' % dataset_file.get(dataset)

    Signals = []
    Labels = []
    SNRs = []

    if dataset == '2016.10a' or dataset == '2016.10b' or dataset == '2016.04c' or dataset == 'rml22':
        Set = pickle.load(open(file_pointer, 'rb'), encoding='bytes')
        snrs, mods = map(lambda j: sorted(list(set(map(lambda x: x[j], Set.keys())))), [1, 0])

        for mod in mods:
            for snr in snrs:
                Signals.append(Set[(mod, snr)])
                for i in range(Set[(mod, snr)].shape[0]):
                    Labels.append(mod)
                    SNRs.append(snr)

        Signals = np.vstack(Signals)
        Signals = torch.from_numpy(Signals.astype(np.float32))
        Labels = [classes[i] for i in Labels]
        Labels = np.array(Labels, dtype=np.int64)
        Labels = torch.from_numpy(Labels)
    logger.info('*' * 20)
    logger.info(f'Signals.shape: {list(Signals.shape)}')
    logger.info(f'Labels.shape: {list(Labels.shape)}')
    logger.info('*' * 20)

    return Signals, Labels, SNRs, snrs, mods


def Dataset_Split(Signals,
                  Labels,
                  SNRs,
                  snrs,
                  mods,
                  logger,
                  val_size=0.2,
                  test_size=0.1
                  ):
    global test_idx

    n_examples = Signals.shape[0]
    n_train = int(n_examples * (1 - val_size - test_size))

    train_idx = []
    test_idx = []
    val_idx = []

    Slices_list = np.linspace(0, n_examples, num=len(mods) * len(snrs) + 1)

    for k in range(0, Slices_list.shape[0] - 1):
        train_idx_subset = np.random.choice(
            range(int(Slices_list[k]), int(Slices_list[k + 1])), size=int(n_train / (len(mods) * len(snrs))),
            replace=False)
        Test_Val_idx_subset = list(set(range(int(Slices_list[k]), int(Slices_list[k + 1]))) - set(train_idx_subset))
        test_idx_subset = np.random.choice(Test_Val_idx_subset,
                                           size=int(
                                               (n_examples - n_train) * test_size / (
                                                       (len(mods) * len(snrs)) * (test_size + val_size))),
                                           replace=False)
        val_idx_subset = list(set(Test_Val_idx_subset) - set(test_idx_subset))

        train_idx = np.hstack([train_idx, train_idx_subset])
        val_idx = np.hstack([val_idx, val_idx_subset])
        test_idx = np.hstack([test_idx, test_idx_subset])

    train_idx = np.array(train_idx, dtype='int64')

    Signals_train = Signals[train_idx]
    Labels_train = Labels[train_idx]

    val_idx = np.array(val_idx, dtype='int64')
    test_idx = np.array(test_idx, dtype='int64')

    Signals_test = Signals[test_idx]
    Labels_test = Labels[test_idx]

    Signals_val = Signals[val_idx]
    Labels_val = Labels[val_idx]


    logger.info(f"Signal_train.shape: {list(Signals_train.shape)}", )
    logger.info(f"Signal_val.shape: {list(Signals_val.shape)}")
    logger.info(f"Signal_test.shape: {list(Signals_test.shape)}")
    logger.info('*' * 20)

    return (Signals_train, Labels_train), \
           (Signals_test, Labels_test), \
           (Signals_val, Labels_val), \
            test_idx

def Create_Data_Loader(train_set, val_set, cfg, logger):

    train_data = Data.TensorDataset(*train_set)
    val_data = Data.TensorDataset(*val_set)

    train_loader = Data.DataLoader(
        dataset=train_data,
        batch_size=cfg.batch_size,
        shuffle=True,
        num_workers=cfg.num_workers,
    )

    val_loader = Data.DataLoader(
        dataset=val_data,
        batch_size=cfg.batch_size,
        shuffle=True,
        num_workers=cfg.num_workers,
    )

    logger.info(f"train_loader batch: {len(train_loader)}")
    logger.info(f"val_loader batch: {len(val_loader)}")

    return train_loader, val_loader