django-models/event/management/commands/artist_popularity_model.py at main · garciadiazjaime/django-models · 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
from django.core.management.base import BaseCommand

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns

import tensorflow as tf
from tensorflow.keras import layers


def plot_loss(history):
    plt.plot(history.history["loss"], label="loss")
    plt.plot(history.history["val_loss"], label="val_loss")
    plt.ylabel("Loss")
    plt.xlabel("Epoch")
    plt.legend(["train", "test"])
    plt.grid(True)
    plt.show()


def plot_values(train_dataset):
    sns.pairplot(
        train_dataset[
            [
                "twitter_followers",
                "popularity",
            ]
        ],
        diag_kind="kde",
    )
    plt.show()


def plot_xy(x, y, feature, labels):
    plt.scatter(feature, labels, label="Data")
    plt.plot(x, y, color="k", label="Predictions")
    plt.xlabel("twitter_followers")
    plt.ylabel("popularity")
    plt.legend()
    plt.show()


def get_cleaned_data(raw_dataset):
    dataset = raw_dataset.copy()

    dataset = dataset[["twitter_followers", "popularity"]]
    dataset["popularity"] = dataset["popularity"] / 100
    # dataset["twitter_followers"] = np.log2(dataset["twitter_followers"])
    dataset["twitter_followers"] = dataset["twitter_followers"]

    return dataset


def get_split_data(dataset):
    train_dataset = dataset.sample(frac=0.8, random_state=0)
    test_dataset = dataset.drop(train_dataset.index)

    return train_dataset, test_dataset


def get_training_data(train_dataset, test_dataset):
    train_features = train_dataset.copy()
    test_features = test_dataset.copy()

    train_labels = train_features.pop("popularity")
    test_labels = test_features.pop("popularity")

    return train_features, test_features, train_labels, test_labels


def get_model(x, y):
    x_normalized = layers.Normalization(
        axis=None,
    )
    x_normalized.adapt(np.array(x))

    model = tf.keras.Sequential([x_normalized, layers.Dense(units=1)])

    model.compile(
        optimizer=tf.keras.optimizers.Adam(learning_rate=0.1),
        loss="mean_squared_error",
    )

    history = model.fit(
        x,
        y,
        epochs=2,
        # Suppress logging.
        verbose=0,
        # Calculate validation results on 20% of the training data.
        validation_split=0.2,
    )

    hist = pd.DataFrame(history.history)
    hist["epoch"] = history.epoch

    return model, history, hist


class Command(BaseCommand):
    help = "Artists Popularity Model generator"

    def add_arguments(self, parser):
        parser.add_argument(
            "--plot",
            action="store_true",
            help="Displays available plots",
        )

    def handle(self, **options):
        uri = "./data/artists.csv"

        raw_dataset = pd.read_csv(
            uri,
            na_values="?",
            comment="\t",
            sep=",",
            skipinitialspace=True,
        )

        dataset = get_cleaned_data(raw_dataset)
        print(dataset.tail())
        print(dataset.isna().sum())

        train_dataset, test_dataset = get_split_data(dataset)
        print(train_dataset.describe().transpose())
        if options["plot"]:
            plot_values(train_dataset)

        train_features, test_features, train_labels, test_labels = get_training_data(
            train_dataset, test_dataset
        )

        model, history, hist = get_model(
            train_features["twitter_followers"],
            train_labels,
        )
        model.summary()
        print(hist.tail())
        if options["plot"]:
            plot_loss(history)

        test_loss = model.evaluate(
            test_features["twitter_followers"], test_labels, verbose=2
        )
        print("\nTest loss:", test_loss)

        if options["plot"]:
            # x = tf.linspace(0.0, np.log2(max(raw_dataset["twitter_followers"])), 1000)
            x = tf.linspace(0.0, max(raw_dataset["twitter_followers"]), 1000)
            y = model.predict(x)
            plot_xy(x, y, train_features["twitter_followers"], train_labels)

        model.save("./data/model/artist_popularity_model.keras")
        model.export("./data/model/saved_model")