OWR-ImageClassification/utils.py at main · rubenIng93/OWR-ImageClassification · 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
import torch
import csv
import statistics
import plotly.graph_objects as go
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
try:
    import cPickle as pickle
except:
    import pickle

def plotConfusionMatrix(method, confusionMatrixData):
    fig,ax=plt.subplots(figsize=(10,10))
    sns.heatmap(confusionMatrixData,cmap='Greens',ax=ax)
    plt.ylabel('True label')
    plt.xlabel('Predicted label')
    plt.title("Confusion Matrix {}".format(method))

    filename = "cm_{}.jpg".format(method) # ex. cm_lwf_30
    plt.savefig(filename, format='png', dpi=300)
    plt.show()

def unpickle(file_path):
    with open(file_path, 'rb') as f:
        data = pickle.load(f)
    return data

def save_data(filename, data):
    with open(filename, 'wb') as f:
        pickle.dump(data, f, 2)

def map_label_2(map_f, labels):

    mapped_labels = list(map(lambda lb: map_f[lb], labels.cpu().numpy()))
    # back in tensor and cuda
    return torch.LongTensor(mapped_labels).cuda()


def map_label(labels, actual_classes, split=None):
    '''
    Function that maps the label in
    the range [0, actual_classes-1]:
    '''
    if split != None:
        map = {k: (v + 10 * split) for v, k in enumerate(actual_classes)}
        labels = labels.cpu().numpy()
        for i in range(len(labels)):
            labels[i] = map[labels[i]]
        return torch.LongTensor(labels).cuda()  # remove cuda if GPU busy
    else:
        map = {k: v for v, k in enumerate(actual_classes)}
        labels = labels.cpu().numpy()
        for i in range(len(labels)):
            labels[i] = map[labels[i]]
        return torch.LongTensor(labels).cuda()  # remove cuda if GPU busy


class FileWriter():
    '''
    Class able to save in a file the result of the training/ testing
    process
    '''

    def __init__(self, filename, open_world=False):

        self.datafile = open(filename, 'w')

        if not open_world:
            print('seed\tsplit_0\tsplit_1\tsplit_2\tsplit_3\tsplit_4\tsplit_5\tsplit_6\tsplit_7\tsplit_8\tsplit_9',
              file=self.datafile)
        else:
            print('seed\tsplit_0\tsplit_1\tsplit_2\tsplit_3\tsplit_4', file=self.datafile)

    def register_seed(self, accuracy_list):
        print(*accuracy_list, sep='\t', file=self.datafile)

    def close_file(self):
        self.datafile.close()


def trend_chart(filename, title):

    split_0, split_1, split_2, split_3, split_4 = [], [], [], [], []
    split_5, split_6, split_7, split_8, split_9 = [], [], [], [], []

    with open(filename, 'r') as _file:
        values = csv.reader(_file, delimiter='\t')
        header = True
        for row in values:
            if header:
                header = False
            else:
                split_0.append(float(row[1])*100)
                split_1.append(float(row[2])*100)
                split_2.append(float(row[3])*100)
                split_3.append(float(row[4])*100)
                split_4.append(float(row[5])*100)
                split_5.append(float(row[6])*100)
                split_6.append(float(row[7])*100)
                split_7.append(float(row[8])*100)
                split_8.append(float(row[9])*100)
                split_9.append(float(row[10])*100)

    mean_values = []
    list_of_lists = [split_0, split_1, split_2, split_3,
                     split_4, split_5, split_6, split_7, split_8, split_9]

    for _list in list_of_lists:
        mean_values.append(statistics.mean(_list))

    x = np.arange(10)

    fig = go.Figure(data=go.Scatter(x=x, y=mean_values))
    # Edit the layout
    fig.update_layout(title=title,
                      xaxis_title='Split',
                      yaxis_title='Accuracy [%]',
                      xaxis=dict(
                          tickmode='array',
                          tickvals=x
                      ))
    fig.show()


def compute_h_mean(closed_dict, open_dict):

    '''
    Compute the harmonic mean
    the input dictionaries have as key the threshold
    (retrieve it as str()!) and as value the accuracy

    Returns: a dictionary k=threshold, value=harmonic mean
    '''

    result = {}

    for t in closed_dict.keys():
        result[str(t)] = 2 / (1/closed_dict[str(t)] + 1/open_dict[str(t)])
    return result


def compute_a_mean(closed_dict, open_dict):
    '''
    Compute the aritmetic mean
    the input dictionaries have as key the threshold
    (retrieve it as str()!) and as value the accuracy

    Returns: a dictionary k=threshold, value=harmonic mean
    '''

    result = {}

    for t in closed_dict.keys():
        result[str(t)] = (closed_dict[str(t)] + open_dict[str(t)]) /2
    return result