import glob

#wget http://vis-www.cs.umass.edu/lfw/lfw.tgz

#tar -xvzf lfw.tgz

face_images = glob.glob('lfw/**/*.jpg')


# In[2]:


import cv2
import numpy as np
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split

from tensorflow.keras import Model, Input, regularizers
from tensorflow.keras.layers import Dense, Conv2D, MaxPool2D, UpSampling2D
from tensorflow.keras.callbacks import EarlyStopping
from keras.preprocessing import image

import PIL
from tqdm import tqdm
import warnings;
warnings.filterwarnings('ignore')


# In[3]:


all_images = []

for i in tqdm(face_images):
  img = image.load_img(i, target_size=(80,80,3))
  img = image.img_to_array(img)
  img = img/255.
  all_images.append(img)


# In[ ]:


all_images = np.array(all_images)

# split data into train and validation data
train_x, val_x = train_test_split(all_images, random_state=32, test_size=0.1)


# In[ ]:


# function to reduce image resolution while keeping the image size constant

def pixalate_image(image, scale_percent = 40):
  width = int(image.shape[1] * scale_percent / 100)
  height = int(image.shape[0] * scale_percent / 100)
  dim = (width, height)

  small_image = cv2.resize(image, dim, interpolation = cv2.INTER_AREA)
  
  # scale back to original size
  width = int(small_image.shape[1] * 100 / scale_percent)
  height = int(small_image.shape[0] * 100 / scale_percent)
  dim = (width, height)

  low_res_image = cv2.resize(small_image, dim, interpolation = cv2.INTER_AREA)

  return low_res_image


# In[ ]:


# get low resolution images for the training set
train_x_px = []

for i in range(train_x.shape[0]):
  temp = pixalate_image(train_x[i,:,:,:])
  train_x_px.append(temp)

train_x_px = np.array(train_x_px)


# get low resolution images for the validation set
val_x_px = []

for i in range(val_x.shape[0]):
  temp = pixalate_image(val_x[i,:,:,:])
  val_x_px.append(temp)

val_x_px = np.array(val_x_px)


# In[ ]:


# get low resolution images for the training set
train_x_px = []

for i in range(train_x.shape[0]):
  temp = pixalate_image(train_x[i,:,:,:])
  train_x_px.append(temp)

train_x_px = np.array(train_x_px)


# get low resolution images for the validation set
val_x_px = []

for i in range(val_x.shape[0]):
  temp = pixalate_image(val_x[i,:,:,:])
  val_x_px.append(temp)

val_x_px = np.array(val_x_px)


# In[ ]:


Input_img = Input(shape=(80, 80, 3))  
    
#encoding architecture
x1 = Conv2D(256, (3, 3), activation='relu', padding='same')(Input_img)
x2 = Conv2D(128, (3, 3), activation='relu', padding='same')(x1)
x2 = MaxPool2D( (2, 2))(x2)
encoded = Conv2D(64, (3, 3), activation='relu', padding='same')(x2)

# decoding architecture
x3 = Conv2D(64, (3, 3), activation='relu', padding='same')(encoded)
x3 = UpSampling2D((2, 2))(x3)
x2 = Conv2D(128, (3, 3), activation='relu', padding='same')(x3)
x1 = Conv2D(256, (3, 3), activation='relu', padding='same')(x2)
decoded = Conv2D(3, (3, 3), padding='same')(x1)

autoencoder = Model(Input_img, decoded)
autoencoder.compile(optimizer='adam', loss='mse')