import cv2

import numpy as np

from nilearn import plotting

import nibabel as nib

import matplotlib.pyplot as plt

class BayesPreprocessor:

    def __init__(self, image):

        # store target image width, height and interpolation

        # inter is an optional argument

        self.image = image

        # self.height = height

        # self.kernel = kernel

        # self.inter = inter

    @staticmethod

    def bayes_noise_removal(self,o):

        # implement image processing

        S1=self.image

        S2 = 255 - S1

        S3 = S1 + S2

        dv = self.calculate_params(S1, S2, S3, o)

        dv[dv < 0] = 0

        return dv

    def calculate_params(self, S1, S2, S3, o):

        try:

            w = S2 / S3

        except ZeroDivisionError:

            w = 0

        w = np.divide(S2, S3)

        p = np.divide(S3, w)

        try:

            p = S3 / w

        except ZeroDivisionError:

            p = 0

        dv = np.log(np.i0((p * np.sinc(w) * S1) / o ** 2)) + np.log(np.i0((p * np.sinc(1 - w) * S2) / o ** 2)) - p * (

                ((np.sinc(w)) ** 2 + (np.sinc(1 - w)) ** 2) / (2 * o ** 2))

        return dv

# plt.imsave('slice.png',slice_2l)

# cv2.imshow('dw.jpg',slice_2l)

# cv2.waitKey()

img = nib.load('../../mr_train/mr_train_1020_image.nii.gz')

label=nib.load('../../mr_train/mr_train_1001_label.nii.gz')

# display = plotting.plot_anat(img)

# plotting.show()

# plotting.plot_img(img)

# plotting.show()

#

img_data=img.get_fdata()

shape_0=int((img_data.shape[0]-1)/2)

shape_1=int((img_data.shape[1]-1)/2)

shape_2=int((img_data.shape[2]-1)/2)

slice_1=img_data[shape_0,:,:]

image=cv2.imread('../tmp/pred_0.png')

image = cv2.resize(image, (128,512), interpolation=cv2.INTER_AREA)

cv2.imshow('daw',image)

cv2.waitKey(0)

# bayes=BayesPreprocessor(image)

# # image = image / 255

# cv2.imshow('dfdw',bayes.bayes_noise_removal(90))

# # plt.imshow(slice_1)

# # plt.show()

# cv2.waitKey(0)

##img=cv2.imread("mri.png",0);

#

# # plt.figure(1)

# S1=slice_1

# # plt.figure(2)

# S2=255-S1

# S3=S1+S2

#

#

# # plt.figure(3)

# try:

#     w = S2/S3

# except ZeroDivisionError:

#     w = 0

# w=np.divide(S2,S3)

# p=np.divide(S3,w)

# try:

#     p = S3/w

# except ZeroDivisionError:

#     p = 0

#

#

#

# # cv2.imshow("pikseli",np.uint8(p))

# # cv2.imwrite("pikseli.png",np.uint8(p))

#

# o=40

#

# dv=np.log(np.i0((p*np.sinc(w)*S1)/o**2))+np.log(np.i0((p*np.sinc(1-w)*S2)/o**2))-p*(((np.sinc(w))**2+(np.sinc(1-w))**2)/(2*o**2))

# # plt.figure(4)

# dv[dv<0]=0

# cv2.imshow("popravljena.png",np.uint8(dv))

# cv2.waitKey(0)