Hugging Face's logo

Cainiao-AI
/
TAAS

Feature Extraction
Transformers
PyTorch
Chinese
TAAS
custom_code
Model card Files Community
TAAS
File size: 4,838 Bytes 
c08e521
 
54a9a32
c08e521
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
 
#!/usr/bin/env python
# -*- coding: utf-8 -*-

import operator
import pickle
import numpy as np
import pandas as pd

s2_label_dict = {
    '0': 0,
    '1': 1,
    '2': 2,
    '3': 3,
    '4': 4,
    '5': 5,
    '6': 6,
    '7': 7,
    '8': 8,
    '9': 9,
    'a': 10,
    'b': 11,
    'c': 12,
    'd': 13,
    'e': 14,
    'f': 15
}
s2_label_decode_dict = {v: k for k, v in s2_label_dict.items()}

s2_weights = [0.025, 0.025, 0.025,
              0.025, 0.025, 0.025,
              0.025, 0.025, 0.025,
              0.0325, 0.0325, 0.0325,
              0.035, 0.035, 0.035,
              0.0375, 0.0375, 0.0375,
              0.04, 0.04, 0.04,
              0.0425, 0.0425, 0.0425,
              0.045, 0.045, 0.0475,
              0.025, 0.025, 0.025,
              0.0, 0.0, 0.0]

def generate_s2_index(s2_label):
    result = [0 for _ in range(33)]
    for i, char_ in enumerate(s2_label):
        result[i] = s2_label_dict[char_]
    return result


def decode_s2(x):
    result = []
    for i in x:
        result.append(s2_label_decode_dict[i])
    return ''.join(result)


def sample_csv2pkl(csv_path, pkl_path):
    # df = pd.read_csv('/Users/liujianlin/odps_clt_release_64/bin/addr6node_small1.csv', sep='^', encoding="utf_8_sig")
    df = pd.read_csv(csv_path, sep='^', encoding="utf_8_sig")
    # print(df)
    data = []
    for index, row in df.iterrows():
        node_s = []
        label = []
        node1 = [row['node_t1'], row['poi_address_mask1'], row['node1'], generate_s2_index(row['node1'])]
        node2 = [row['node_t2'], row['poi_address_mask2'], row['node2'], generate_s2_index(row['node2'])]
        node3 = [row['node_t3'], row['poi_address_mask3'], row['node3'], generate_s2_index(row['node3'])]
        node4 = [row['node_t4'], row['poi_address_mask4'], row['node4'], generate_s2_index(row['node4'])]
        node5 = [row['node_t5'], row['poi_address_mask5'], row['node5'], generate_s2_index(row['node5'])]
        node6 = [row['node_t6'], row['poi_address_mask6'], row['node6'], generate_s2_index(row['node6'])]
        label.extend(node1[3])
        label.extend(node2[3])
        label.extend(node3[3])
        label.extend(node4[3])
        label.extend(node5[3])
        label.extend(node6[3])
        node1.append(label)
        node2.append(label)
        node3.append(label)
        node4.append(label)
        node5.append(label)
        node6.append(label)
        node_s.append(node1)
        node_s.append(node2)
        node_s.append(node3)
        node_s.append(node4)
        node_s.append(node5)
        node_s.append(node6)
        data.append(node_s)
    # print(data)

    with open(pkl_path,'wb') as f:
        pickle.dump(data,f)


def calculate_multi_s2_acc(predicted_s2, y):
    acc_cnt = np.array([0, 0, 0, 0, 0, 0, 0])
    y = y.view(-1, 33).tolist()
    predicted = predicted_s2.view(-1, 33).tolist()
    # print(y.shape, predicted.shape)
    for index, s2 in enumerate(y):
        for c, i in enumerate(range(12, 33, 3)):
            y_l10 = y[index][12:i+3]
            p_l10 = predicted[index][12:i+3]
            # print(y_l10, p_l10, operator.eq(y_l10, p_l10))
            if operator.eq(y_l10, p_l10):
                acc_cnt[c] += 1
        # print('==='*20)
    # print(acc_cnt)
    return acc_cnt

def calculate_multi_s2_acc_batch(predicted_s2, y, sequence_len = 6):
    acc_cnt = np.array([0, 0, 0, 0, 0, 0, 0])
    y = y.view(-1, sequence_len, 33).tolist()
    predicted = predicted_s2.view(-1, sequence_len, 33).tolist()
    # print(y.shape, predicted.shape)
    batch_size = len(y)
    for batch_i in range(batch_size):
        for index, s2 in enumerate(y[batch_i]):
            for c, i in enumerate(range(12, 33, 3)):
                y_l10 = y[batch_i][index][12:i+3]
                p_l10 = predicted[batch_i][index][12:i+3]
                # print(y_l10, p_l10, operator.eq(y_l10, p_l10))
                if operator.eq(y_l10, p_l10):
                    acc_cnt[c] += 1
        # print('==='*20)
    # print(acc_cnt)
    return acc_cnt



def calculate_alias_acc(predicted, y):
    tp, fp, fn, tn = 0, 0, 0, 0
    acc = 0
    for index, label in enumerate(y):
        if int(label) == int(predicted[index]):
            acc += 1
        if int(label) == 1:
            fn += 1
            if int(predicted[index]) == 1:
                tp += 1
    if fn == 0:
        precision = 0
    else:
        precision = tp / fn * 100
    return tp, fn, acc


def calculate_aoi_acc(predicted, y):
    tp, fp, fn, tn = 0, 0, 0, 0
    acc = 0
    for index, label in enumerate(y):
        if int(label) == int(predicted[index]):
            acc += 1
        if int(label) == 0:
            fn += 1
            if int(predicted[index]) == 0:
                tp += 1
    if fn == 0:
        precision = 0
    else:
        precision = tp / fn * 100
    return tp, fn, acc