import copy

import keras

import numpy as np

import pandas as pd

#from model import get_model

from keras.models import load_model

x_train , x_test , y_test , y_train = pd.read_csv('sudoku.csv')

adam = keras.optimizers.Adam(lr = 0.001)

model.compile(loss='sparse_categorical_crossentropy' , optimizer = adam)

model.fit(x_train , y_train , batch_size = 32, epochs = 2)

model =load_model('sudokumodel.py')

def norm(a):

return (a/9)-.5;

def denorm(a):

return (a+ .5)*9;

def inference_sudoku(sample):

feat = copy.copy(sample)

while(1):

out = model.predict(feat.reshape((1,9,9,1)))

out = out.squeeze()

pred = np.argmax(out, axis=1).reshape((9,9))+1

prob = np.around(np.max(out, axis=1).reshape((9,9)), 2)

feat = denorm(feat).reshape((9,9))

mask = (feat==0)

if(mask.sum()==0):

break

prob_new = prob*mask

ind = np.argmax(prob_new)

x, y = (ind//9), (ind%9)

val = pred[x][y]

feat[x][y] = val

feat = norm(feat)

return pred

def solve_sudoku(game):

game = game.replace('\n', '')

game = game.replace(' ', '')

game = np.array([int(j) for j in game]).reshape((9,9,1))

game = norm(game)

game = inference_sudoku(game)

return game

game = '''

0 8 0 0 3 2 0 0 1

7 0 3 0 8 0 0 0 2

5 0 0 0 0 7 0 3 0

0 5 0 0 0 1 9 7 0

6 0 0 7 0 9 0 0 8

0 4 7 2 0 0 0 5 0

0 2 0 6 0 0 0 0 9

8 0 0 0 9 0 3 0 5

3 0 0 8 2 0 0 1 0

'''

game = solve_sudoku(game)

print('solved puzzle:\n')

print(game)