import numpy as np
from PIL import Image
import matplotlib.pyplot as plt
from sklearn import datasets
from sklearn.metrics import accuracy_score
from sklearn.neural_network import MLPClassifier
import pandas as pd
import cv2
import keras

# 28x28
train_data = np.genfromtxt('dataset/train.csv', delimiter=',', skip_header=1, max_rows=20000, encoding='utf-8')
test_data = np.genfromtxt('dataset/test.csv', delimiter=',', skip_header=1, max_rows=20000, encoding='utf-8')

# train_data = pd.read_csv('dataset/train.csv')
# test_data = pd.read_csv('dataset/test.csv')

# training
# recznie napisane cyfry

digits = datasets.load_digits()
y = digits.target
x = digits.images.reshape((len(digits.images), -1))

# print(type(y[0]), type(x[0]))
# ogarnac zbior, zwiekszyc warstwy

# x_train = train_data.iloc[:, 1:].values.astype('float32')
# y_train = train_data.iloc[:, 0].values.astype('int32')
# x_test = test_data.values.astype('float32')

x_train = train_data[0:10000, 1:]
y_train = train_data[0:10000, 0]
x_test = train_data[10001:20000, 1:]
y_test = train_data[10001:20000, 0].astype('int')

print(type(y_test[0]), type(x_test[0]))

# x_train = x[:900]
# y_train = y[:900]
# x_test = x[900:]
# y_test = y[900:]

# 500, 500, 500, 500, 500
mlp = MLPClassifier(hidden_layer_sizes=(150, 100, 100, 100), activation='logistic', alpha=1e-4,
                    solver='sgd', tol=0.000000000001, random_state=1,
                    learning_rate_init=.1, verbose=True, max_iter=10000)

mlp.fit(x_train, y_train)
predictions = mlp.predict(x_test)

print("Accuracy: ", accuracy_score(y_test, predictions))

# image

img = cv2.cvtColor(cv2.imread('test5.jpg'), cv2.COLOR_BGR2GRAY)
img = cv2.blur(img, (9, 9))  # poprawia jakosc
img = cv2.resize(img, (28, 28), interpolation=cv2.INTER_AREA)
img = img.reshape((len(img), -1))

# print(type(img))
# print(img.shape)
# plt.imshow(img ,cmap='binary')
# plt.show()

data = []

rows, cols = img.shape
for i in range(rows):
    for j in range(cols):
        k = img[i, j]
        if k > 225:
            k = 0  # brak czarnego
        else:
            k = 255

        data.append(k)

data = np.asarray(data, dtype=np.float64)
# print(data)
print(type(data))

predictions = mlp.predict([data])

print("Liczba to:", predictions[0].astype('int'))