import warnings
warnings.filterwarnings("ignore")
import os
os.environ['TF_CPP_MIN_LOG_LEVEL']='3'
import numpy as np
import cv2
import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.applications import InceptionV3
from tensorflow.keras.layers import Conv2D, Dense, Dropout, MaxPool2D, Flatten, BatchNormalization
from tensorflow.keras.optimizers import Adam
from tensorflow.keras.callbacks import TensorBoard, EarlyStopping, ReduceLROnPlateau, Callback
from tqdm import tqdm as tqdm
path = 'your_path/'

data = []

def load_data(path):
    global data
    print('Loading in Data ....')
    for f in tqdm(os.listdir(path)):
        p = path + f
        file = np.load(p, allow_pickle=True)
        try:
            data = data + file
        except:
            data = file
load_data(path)

print('Separating images and labels...')
x = np.array([np.array(img) for img, label in data]).reshape(-1, 256, 480, 3) / 255
y = np.array([label for img, label in data])
print(x.shape)
print(y[0])


class model():
    def __init__(self):
        self.model = Sequential()
        self.name = 'version0.1'
        self.epochs = 10
        # Callbacks
        self.tensorboard = TensorBoard()
        self.reducelr = ReduceLROnPlateau(patience=2)
        self.WIDTH = 256
        self.HEIGHT = 480
        self.optimizer = Adam(lr=0.001)
        self.make_model()

    def make_model(self):
        self.model.add(InceptionV3(include_top=False,
                           input_shape=(self.WIDTH, self.HEIGHT, 3),
                       weights="inception_v3_weights.h5"))
        self.model.add((Dense(32, activation='relu')))
        self.model.add(Dense(7, activation='softmax'))
        self.model.layers[0].trainable = False

    def train(self):
        self.model.compile(optimizer=self.optimizer, loss='categorical_crossentropy', metrics=['accuracy'])
        self.model.fit(x=x, y=y, batch_size=64,
                   validation_split=0.15, shuffle=True, epochs=self.epochs,
                   callbacks=[self.tensorboard, self.saver, self.reducelr],)

conv_model1 = model()
conv_model1.train()