data_exploration.py

import scipy.io
from itertools import islice
import pprint
import matplotlib.image as mpimg
import pickle
import numpy as np


class DataGAN:

    def __init__(self):

        self.actor_age = None
        self.actor_image_url = None
        self.actor_id = None
        self.image_array = None

    def load_data(self):

        data_dict = scipy.io.loadmat('data/celebrity2000_meta.mat')
        celebrity_data = data_dict['celebrityImageData']

        actor_age = []
        actor_image_url = []
        actor_id = []

        for i in celebrity_data[0][0][0]:
            actor_age.append(i.tolist()[0])

        for i in celebrity_data[0][0][7]:
            actor_image_url.append(i.tolist()[0][0])

        for i in celebrity_data[0][0][1]:
            actor_id.append(i.tolist()[0])

        return actor_age, actor_image_url, actor_id

    def read_image(self, image):

        # Reading the image from the directory
        img1 = mpimg.imread('CACD2000/' + image)
        return img1

    def create_image_array(self, actor_image_url):

        image_array = []

        # image_array = np.empty([163446,250,250,3])

        for i, url in enumerate(actor_image_url):

            # image_array[i]=read_image(url)

            image_array.append(self.read_image(url))

            if i == 10:
                break

        return image_array

    def data_preprocess(self):

        self.actor_age, self.actor_image_url, self.actor_id = self.load_data()
        # self.image_array= self.create_image_array(self.actor_image_url)

        # list of ages, list of actor ids, list of 3d image arrays
        return self.actor_age, self.actor_id, self.actor_image_url


x = DataGAN()
actor_age, actor_id, image_array = x.data_preprocess()

print(len(image_array))

with open('data/image_urls.pkl', 'wb') as f:
    pickle.dump(image_array, f)

with open('data/image_ages.pkl', 'wb') as f:
    pickle.dump(actor_age, f)

# with open('GAN_Image_Dump.pkl','rb') as f:
#     x = pickle.load(f)
# print(x[0].shape)
# print(len(x))


# with open('GAN_Age_Dump.pkl','rb') as f:
#     x = pickle.load(f)
#     print(x.shape)