download-quick-draw.py

import tensorflow as tf
import tensorflow_datasets as tfds
from tensorflow import keras 
import numpy as np
import os
import urllib.request

#Names of 10 Classes:
class_names = ['cloud', 'sun', 'pants', 'umbrella', 'table', 'ladder',
               'eyeglasses', 'clock', 'scissors', 'cup']

def download():
    """
    Function to download dataset from google api
    """
    root = 'data'
    os.mkdir('data')
    print('Downloading data with selected classes...')
    base = 'https://storage.googleapis.com/quickdraw_dataset/full/numpy_bitmap/'
    for c in class_names:
        path = base+c+'.npy'
        print(path)
        urllib.request.urlretrieve(path, f'{root}/{c}.npy')
    print("Data Downloaded !")

def load(test_split = 0.2, max_items_per_class = 10000):
    """
    Function to load the dataset as train and test
    """
    print("Loading data....")
    root = 'data'
    #initialize variables 
    x = np.empty([0, 784])
    y = np.empty([0])

    #load each data file 
    for idx, file in enumerate(class_names):
        data = np.load(f'{root}/{file}.npy')
        data = data[0: max_items_per_class, :]
        labels = np.full(data.shape[0], idx)

        x = np.concatenate((x, data), axis=0)
        y = np.append(y, labels)

    data = None
    labels = None

    #randomize the dataset 
    permutation = np.random.permutation(y.shape[0])
    x = x[permutation, :]
    y = y[permutation]

    #reshape and inverse the colors 
    x = 255 - np.reshape(x, (x.shape[0], 28, 28))

    #separate into training and testing 
    test_size  = int(x.shape[0]/100*(test_split*100))

    x_test = x[0:test_size, :]
    y_test = y[0:test_size]

    x_train = x[test_size:x.shape[0], :]
    y_train = y[test_size:y.shape[0]]

    print('Training Data : ', x_train.shape[0])
    print('Testing  Data : ', x_test.shape[0])
    return x_train, y_train, x_test, y_test, class_names


if __name__ == "__main__":
    download()
#    x_train, y_train, x_test, y_test, class_names = load()