DL_Notes/Dataset_Support.py at master · daddyjab/DL_Notes · 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
# File: Dataset_Support.py
# Owner: Jeff Brown

# Dependencies: Standard libraries
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import math

# Machine Learning - Keras (Tensorflow) -  Dataset Generation
from tensorflow.keras.datasets import mnist      # Images: Handwritten digits 0-9 (28x28 grayscale, 60K train, 10K test)

# Function to import the Keras MNIST handwritten digits sample dataset
def mnist_load_ds():
    (X_train, y_train), (X_test, y_test) = mnist.load_data()
    return (X_train, y_train), (X_test, y_test)


# Function to plot a list of up to 10 digits on a single subplot
def mnist_plot_digit_list( a_X_list = None, a_y_list = None, a_find_all_digits = False):
    # The first 10 digits from the specified list

    # If no list is specified then return None
    if (a_X_list is None):
        return None

    else:
        X_list = list(a_X_list)

    if (a_y_list is None):
        return None

    else:
        y_list = list(a_y_list)

    # Find All Digits flag
    #   If True => Find and plot all digits 0-9 within the list, starting at index 0
    #   If False => Plot up to the first 10 digits in the list
    if a_find_all_digits:

        # Flag is True: Get indices of samples for each of the digits 0-9 within the 1000 sample subset
        # If the digit is not present in the input list then move on to the next digit
        d_i_list = []
        for d in range(10):

            try:
                # Add the index at which this digit can be found to the list
                d_i_list.append( y_list.index(d) )

            except ValueError:
                # Digit is not present in the input list -- move on to the next digit
                pass

    else:
        # Flag is False: Get the indices for up to the first 10 values in the list
        d_i_list = range( min(10, len(y_list) ))

    # The iterpolation method to use for ploting the digit images
    i_type_selected = 'lanczos'

    print("Indices:", d_i_list)

    # Plot Classification Performance results: Best Score vs. Mean Fit Time (ms)
    fig = plt.figure(figsize=(20,9))

    # Create subplots for each of the sampled digits
    for i in range(len(d_i_list)):
        # Create a subplot for this iteration
        ax = fig.add_subplot( math.ceil(len(d_i_list)/min(5, len(d_i_list))), min(5, len(d_i_list)), i+1 )

        # Display a note for each subplot
        point_text = f"Label: {y_list[d_i_list[i]]}"
        point_text += f"\nSample Index: {d_i_list[i]}"
    #     ax.text(1, 2+1.4*point_text.count("\n"), point_text )
        ax.set_title(point_text)

        # Display the image
        ax.imshow(X_list[d_i_list[i]], cmap=plt.cm.Greys, interpolation=i_type_selected)

    # Return the number of digits plotted
    return i+1

# Function to plot example images
def plot_examples( a_X_list = None, a_y_list = None, a_find_labels=False, a_label = None):
    """
    Plot examples from the specified list
    a_X_list:   List of example inputs
    a_y_list:   List of example labels
    a_find_labels:
                False:  Plot all examples
                True:
                        If a_label is None, plot one example of each unique label
                        If a_label not None, plot all examples with the single specified label
    """

    # Parse args
    X_list = list(a_X_list.squeeze() )
    y_list = list(a_y_list.squeeze() )

    # Parse flags to determine which examples to plot
    if a_find_labels == False:
        # Plot all examples
        d_i_list = range( len(y_list) )

    if a_find_labels == True:
        # Plot examples based upon the label

        if a_label is None:
            # Plot examples for each unique label in the input
            label_list = sorted(set(y_list))

            d_i_list = []
            for label in label_list:
                try:
                    # Add the index at which this label can be found to the list
                    d_i_list.append( y_list.index(label) )

                except ValueError:
                    # Digit is not present in the input list
                    # => Shouldn't happen since the list of labels is drawn from the input list
                    pass

        else:
            # Plot all examples with the single specified label
            d_i_list = []
            for i in range(len(y_list)):
                if y_list[i] == a_label:
                    d_i_list.append(i)

    # print("Image Indices to display:", d_i_list)

    # Plot images
    fig = plt.figure(figsize=(20,10))

    # Create subplots for each of the sampled digits
    for i in range(len(d_i_list)):
        # Create a subplot for this iteration
        ax = fig.add_subplot( math.ceil(len(d_i_list)/min(5, len(d_i_list))), min(5, len(d_i_list)), i+1 )

        # Display a note for each subplot
        point_text = f"Label: {y_list[d_i_list[i]]}"
        point_text += f"\nSample Index: {d_i_list[i]}"
    #     ax.text(1, 2+1.4*point_text.count("\n"), point_text )
        ax.set_title(point_text)

        # Display the image
        ax.imshow(X_list[d_i_list[i]], cmap=plt.cm.Greys, interpolation='lanczos')

    # Return the number of digits plotted
    return i+1