-
Notifications
You must be signed in to change notification settings - Fork 7
/
Copy pathCheckers.py
776 lines (734 loc) · 39 KB
/
Checkers.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
###############################################################################
# Checkers.py
#
# Revision: 1.00
# Date: 11/11/2020
# Author: Alex
#
# Purpose: Contains all functions necessary to implement the Checkers
# game environment.
#
# Notes:
# 1. Run this module to see a demonstration game of Checkers.
# 2. Rules are according to the World Checkers Draughts Federation (WCDF).
# 3. https://www.wcdf.net/rules.htm
#
###############################################################################
"""
import numpy as np
from copy import deepcopy
from tabulate import tabulate
import time
class Checkers:
"""Class to represent a game of Checkers."""
def __init__(self, neural_net=None):
"""Initialize the class with the pieces in their starting positions.
Get a list of valid (legal) next moves. Moves are not explicitly
represented; instead possible next states of the game board are
generated. The player's move is implied by selecting one of these
possible next states.
The game state is a 3 dimensional NumPy array of 15 8x8 arrays.
These 8x8 arrays represent the 8x8 game board, where:
0. Array 0 represents the locations of player 1's uncrowned men.
1. Array 1 represents the locations of player 1's kings.
2. Array 2 represents the locations of player 2's uncrowned men.
3. Array 3 represents the locations of player 2's kings.
4. Array 4 indicates the current player (all 0s for P1, all 1s for P2)
5. Array 5 is the draw timer; counts in increments of 1/80
6. Arrays 6, 7, 8, 9 are normal moves (UL, UR, BL, BR)
7. Arrays 10, 11, 12, 13 are jumps (UL, UR, BL, BR)
8. Array 14 contains the indices of the parent state's action
"""
self.state = np.zeros((15,8,8), dtype=float)
self.init_board()
self.history = [self.state]
self.legal_next_states = self.get_legal_next_states(self.history)
self.move_count = 0
self.done = False
self.outcome = None
self.player1_man = 'x'
self.player1_king = u'\u0416'
self.player2_man = 'o'
self.player2_king = u'\u01D1'
self.neural_net = neural_net
def step(self, next_state):
"""Execute the player's (legal) move. Check to see if the game
has ended, and update the list of legal next moves.
"""
if any((next_state[:5] == x[:5]).all() for x in self.legal_next_states):
self.state = next_state
self.history.append(self.state)
self.legal_next_states = self._check_moves(self.history)
self.done, self.outcome = self.determine_outcome(self.history,
legal_moves=self.legal_next_states)
self.move_count += 1
return self.state, self.outcome, self.done
else:
raise ValueError('Illegal next state (invalid move)!')
def get_legal_next_states(self, history):
"""If the game is not done, return a list of legal next moves given
a history of moves as input. The next moves are actually board states;
the move to achieve those states is implied.
This function calls determine_outcome() which also must check
the legal next states in order to determine the outcome of the game.
Redundant computation can be avoided by checking the legal next states
here first and then passing them as an optional argument to
determine_outcome().
"""
legal_next_states = self._check_moves(history)
done, outcome = self.determine_outcome(history,
legal_moves=legal_next_states)
if done == True: return [] # Game over
return legal_next_states
def _check_moves(self, history):
"""Method intended for internal use. Creates a list of the locations
of all of the pieces on the board divided up into four categories (P1's
men, P1's kings, P2's men, and P2's kings).
Checks for all possible ordinary moves of men and kings for the current
player only. Calls two other internal methods, _check_jumps()
and _check_king_jumps() to determine if there are jumps possible for
the player's men and kings, respectively. Per the rules jumps are
mandatory moves, and so if jump moves exist they are returned by the
function instead of the ordinary moves.
The function also determines if an ordinary move results in a man
reaching King's Row, and kings the man if so.
"""
state = history[-1]
player = int(state[4,0,0])
xman1, yman1 = np.where(state[0] == 1) # Locations of P1's men
xking1, yking1 = np.where(state[1] == 1) # Locations of P1's kings
xman2, yman2 = np.where(state[2] == 1) # Locations of P2's men
xking2, yking2 = np.where(state[3] == 1) # Locations of P2's kings
piece_locs = [np.column_stack((xman1, yman1)), np.column_stack((xking1, yking1)),
np.column_stack((xman2, yman2)), np.column_stack((xking2, yking2))]
board = np.sum(state[0:4], axis=0) # All pieces on one 8x8 grid
idx = player * 2 # State index of player's pieces
opp_idx = 0 if idx else 2 # State index of opponent's pieces
fwd = 1 if player == 0 else -1 # Sets forward direction of player's men
legal_moves = []
jump_moves = []
# Get legal moves including jumps for men
for x, y in piece_locs[idx]: # Men
if y+1 < 8 and -1 < x+fwd < 8:
if board[x+fwd,y+1] == 0: # Diagonal-right space open
temp_state = deepcopy(state)
temp_state[5:] = 0 # Erase NN layers from previous state
temp_state[4] = 1 - player # Toggle player
temp_state[idx,x,y] = 0 # Piece no longer in prev location
if (fwd == 1 and x+fwd == 7) or \
(fwd == -1 and x+fwd == 0): # On King's row, king the man
temp_state[idx+1,x+fwd,y+1] = 1
else: # Not on King's row, man does not become king
temp_state[idx,x+fwd,y+1] = 1
if fwd == 1:
state[9,x,y] = 1 # NN layer representing BR move
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
9, x, y
else:
state[7,x,y] = 1 # NN layer representing UR move
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
7, x, y
legal_moves.append(temp_state)
if y-1 > -1 and -1 < x+fwd < 8:
if board[x+fwd,y-1] == 0: # Diagonal-left space open
temp_state = deepcopy(state)
temp_state[5:] = 0 # Erase NN layers from previous state
temp_state[4] = 1 - player # Toggle player
temp_state[idx,x,y] = 0 # Piece no longer in prev location
if (fwd == 1 and x+fwd == 7) or \
(fwd == -1 and x+fwd == 0): # On King's row, king the man
temp_state[idx+1,x+fwd,y-1] = 1
else: # Not on King's row, man does not become king
temp_state[idx,x+fwd,y-1] = 1
if fwd == 1:
state[8,x,y] = 1 # NN layer representing BL move
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
8, x, y
else:
state[6,x,y] = 1 # NN layer representing UL move
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
6, x, y
legal_moves.append(temp_state)
# Check to see if man can jump any of opponent's pieces
jump_moves.extend(self._check_jumps(x,y,fwd,state,idx,opp_idx,board,player))
# Get legal moves including jumps for kings
for x, y in piece_locs[idx+1]: # Kings
for xmove in range(-1,2,2):
for ymove in range(-1,2,2):
if -1 < x+xmove < 8 and -1 < y+ymove < 8:
if board[x+xmove,y+ymove] == 0: # Diag space open
temp_state = deepcopy(state)
temp_state[5:] = 0 # Erase NN layers from previous state
temp_state[4] = 1 - player # Toggle player
temp_state[idx+1,x,y] = 0 # Piece no longer in prev location
temp_state[idx+1,x+xmove,y+ymove] = 1
if xmove == 1 and ymove == 1:
state[9,x,y] = 1 # NN layer representing BR move
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
9, x, y
elif xmove == 1 and ymove == -1:
state[8,x,y] = 1 # NN layer representing BL move
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
8, x, y
elif xmove == -1 and ymove == 1:
state[7,x,y] = 1 # NN layer representing UR move
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
7, x, y
elif xmove == -1 and ymove == -1:
state[6,x,y] = 1 # NN layer representing UL move
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
6, x, y
legal_moves.append(temp_state)
# Check to see if king can jump any of opponent's pieces
jump_moves.extend(self._check_king_jumps(x,y,state,idx,opp_idx,board,player))
if jump_moves:
state[6:10] = 0 # Clear all possible non-jump moves
return jump_moves # Jumps are mandatory
return legal_moves
def _check_jumps(self,x,y,fwd,state,idx,opp_idx,board,player):
"""Method intended for internal use. Checks to see if a jump is
possible for a man given its position and the game state. Function
recursively calls itself in case multiple jumps are possible in the
same turn. All jumps are mandatory moves, so a double jump takes
precedence over a single jump, a triple jump over a double jump, etc.
If a jump lands a man on King's row, the man is kinged and the player's
turn is over.
"""
jump_moves = []
more_jumps = []
for ydir in range(-1,2,2):
if -1 < y+ydir < 8 and -1 < x+fwd < 8:
if state[opp_idx,x+fwd,y+ydir] == 1 or \
state[opp_idx+1,x+fwd,y+ydir] == 1: # Opponent's piece on diag space
if -1 < y+2*ydir < 8 and -1 < x+2*fwd < 8:
if board[x+fwd*2,y+ydir*2] == 0: # Piece is jumpable
temp_state = deepcopy(state)
temp_state[5:] = 0 # Erase NN layers from previous state
temp_state[idx,x,y] = 0 # Piece no longer in prev location
temp_state[opp_idx,x+fwd,y+ydir] = 0 # Opponent's piece jumped (if man)
temp_state[opp_idx+1,x+fwd,y+ydir] = 0 # Opponent's piece jumped (if king)
if (fwd == 1 and x+2*fwd == 7) or \
(fwd == -1 and x+2*fwd == 0): # On King's row, jump is over
temp_state[idx+1,x+2*fwd,y+2*ydir] = 1 # Man is kinged
else: # Check for multiple jumps
temp_state[idx,x+2*fwd,y+2*ydir] = 1
more_jumps = self._check_jumps(x+2*fwd,y+2*ydir,
fwd,temp_state,idx,
opp_idx,board,player)
if more_jumps:
#jump_moves.extend(more_jumps)
more_jumps = [] # Don't toggle player
else:
temp_state[4] = 1 - player # Toggle player
if fwd == 1 and ydir == 1:
state[13,x,y] = 1 # NN layer representing BR jump
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
13, x, y
elif fwd == 1 and ydir == -1:
state[12,x,y] = 1 # NN layer representing BL jump
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
12, x, y
elif fwd == -1 and ydir == 1:
state[11,x,y] = 1 # NN layer representing UR jump
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
11, x, y
elif fwd == -1 and ydir == -1:
state[10,x,y] = 1 # NN layer representing UL jump
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
10, x, y
jump_moves.append(temp_state)
return jump_moves
def _check_king_jumps(self,x,y,state,idx,opp_idx,board,player):
"""Method intended for internal use. Checks to see if a jump is
possible for a king given its position and the game state. Function
recursively calls itself in case multiple jumps are possible in the
same turn. All jumps are mandatory moves, so a double jump takes
precedence over a single jump, a triple jump over a double jump, etc.
"""
jump_moves = []
more_jumps = []
for ydir in range(-1,2,2):
for fwd in range(-1,2,2):
if -1 < x+fwd < 8 and -1 < y+ydir < 8:
if state[opp_idx,x+fwd,y+ydir] == 1 or \
state[opp_idx+1,x+fwd,y+ydir] == 1: # Opponent's piece on diag space
if -1 < x+2*fwd < 8 and -1 < y+2*ydir < 8:
if board[x+fwd*2,y+ydir*2] == 0: # Piece is jumpable
temp_state = deepcopy(state)
temp_state[5:] = 0 # Erase NN layers from previous state
temp_state[idx+1,x,y] = 0 # Piece no longer in prev location
temp_state[opp_idx,x+fwd,y+ydir] = 0 # Opponent's piece jumped (if man)
temp_state[opp_idx+1,x+fwd,y+ydir] = 0 # Opponent's piece jumped (if king)
temp_state[idx+1,x+2*fwd,y+2*ydir] = 1 # Move piece to new location
more_jumps = self._check_king_jumps(x+2*fwd,y+2*ydir,
temp_state,idx,
opp_idx,board,player)
if more_jumps:
#jump_moves.extend(more_jumps)
more_jumps = [] # Don't toggle player
else:
temp_state[4] = 1 - player # Toggle player
if fwd == 1 and ydir == 1:
state[13,x,y] = 1 # NN layer representing BR jump
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
13, x, y
elif fwd == 1 and ydir == -1:
state[12,x,y] = 1 # NN layer representing BL jump
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
12, x, y
elif fwd == -1 and ydir == 1:
state[11,x,y] = 1 # NN layer representing UR jump
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
11, x, y
elif fwd == -1 and ydir == -1:
state[10,x,y] = 1 # NN layer representing UL jump
temp_state[14,0,0], temp_state[14,0,1], temp_state[14,0,2] = \
10, x, y
jump_moves.append(temp_state)
return jump_moves
def determine_outcome(self, history, legal_moves=[]):
"""Given game history as an input, determine if the game is over.
If the game is over, determine the winner (or if the game is a draw).
The win condition is to either capture all of your opponent's pieces or
to make the last move - meaning that the opponent is unable to move
their pieces due to being blocked.
According to the WCDF, a game shall be declared a draw when both of
the following conditions are true:
1. Neither player has advanced an uncrowned man towards the king-row
during their own previous 40 moves.
2. No pieces have been removed from the board during their own previous
40 moves.
I interpret this to mean that the conditions should be checked over the
last 80 moves of the game (last 40 moves for each player).
"""
state = history[-1]
current_num_pieces = np.sum(state[0:4]) # Number of pieces on board
current_player = int(state[4,0,0])
last_player_to_move = 1 - current_player
if not legal_moves:
legal_moves = self._check_moves(history) # Can player make a move?
man_moved, piece_jumped = True, True # Default if <80 moves played
if len(history) >= 80:
man_moved = False # Draw condition #1
piece_jumped = False # Draw condition #2
for cnt, move in enumerate(reversed(history[-80:])): # Check for draw conditions
if np.sum(move[0:4]) != current_num_pieces:
piece_jumped = True
state[5] = cnt / 80 # NN layer representing draw counter
break
if not ((move[0] == state[0]).all() and (move[2] == state[2]).all()):
man_moved = True
state[5] = cnt / 80 # NN layer representing draw counter
break
if np.sum(state[2:4]) == 0: # All player 2 pieces jumped
done = True
outcome = 'player1_wins'
elif np.sum(state[0:2]) == 0: # All player 1 pieces jumped
done = True
outcome = 'player2_wins'
elif not legal_moves: # No legal moves for player
if last_player_to_move == 0:
done = True
outcome = 'player1_wins'
else:
done = True
outcome = 'player2_wins'
elif (not man_moved) and (not piece_jumped):
done = True
outcome = 'draw'
state[5] = 1 # NN layer representing draw counter
else:
done = False # Game is still in progress
outcome = None
return done, outcome
def print_board(self):
"""Print a visual representation of the current game state to the
console."""
player = int(self.state[4,0,0])
player_mark = self.player1_man if player == 0 else self.player2_man
board_pieces = (self.state[0] - self.state[2]) + \
2*(self.state[1] - self.state[3])
board_table = []
for row_idx, row in enumerate(board_pieces):
row_list = []
for col_idx, square in enumerate(row):
if square == 1: row_list.append(self.player1_man)
if square == -1: row_list.append(self.player2_man)
if square == 2: row_list.append(self.player1_king)
if square == -2: row_list.append(self.player2_king)
if square == 0:
if (not (row_idx % 2) and not (col_idx % 2)):
row_list.append('.') # Unused dark squares
elif ((row_idx % 2) and (col_idx % 2)):
row_list.append('.') # Unused dark squares
else:
row_list.append('') # Empty squares
board_table.append(row_list)
print(tabulate(board_table, tablefmt='fancy_grid'))
if not self.done:
print('Move #{}: It\'s now Player {}\'s turn ({})'
.format(self.move_count+1, player+1, player_mark))
else:
print('Game over after {} moves! The outcome is: {}'
.format(self.move_count+1, self.outcome))
def current_player(self, state):
"""Return which player's turn it is for a given input state."""
player = int(state[4,0,0])
if player == 0:
return 'player1'
else:
return 'player2'
def reset(self):
"""Reset the board so it's ready for the next game."""
self.state = np.zeros((15,8,8), dtype=float)
self.init_board()
self.history = [self.state]
self.legal_next_states = self.get_legal_next_states(self.history)
self.move_count = 0
self.done = False
self.outcome = None
def init_board(self):
"""Place Checkers pieces on board in their starting positions."""
for row in range(0,8):
for col in range(0,8):
if row % 2 != col % 2: # Row and column not both odd or even
if row < 3:
self.state[0,row,col] = 1
elif row > 4:
self.state[2,row,col] = 1
def predict(self, state):
"""Use the supplied neural network to predict the Q-value of a given
state, as well as the prior probabilities of its child states.
Masks any non-valid probabilities, and re-normalizes the remaining
probabilities to sum to 1.
"""
nn_inp = np.moveaxis(state[:14],0,-1)
nn_inp = nn_inp.reshape(1,8,8,14)
prob_planes, q_value = self.neural_net.predict(nn_inp)
prob_planes, q_value = prob_planes[0].reshape((8,8,8)), q_value[0][0]
action_mask = state[6:14]
prob_planes *= action_mask
prob_planes = prob_planes / np.sum(prob_planes)
return prob_planes, q_value
def set_prior_probs(self, child_nodes, prob_planes):
"""Takes as input a list of the parent node's child nodes, and the
probability vector generated by running the parent's state through the
neural network. Assigns each child node its corresponding prior
probability as predicted by the neural network.
"""
for child in child_nodes:
layer = int(child.state[14,0,0]) - 6
x = int(child.state[14,0,1])
y = int(child.state[14,0,2])
if x % 2 == y % 2: raise ValueError('Invalid (x,y) locations for probabilities!')
if not (0 <= layer <= 7): raise ValueError('Invalid layer for probabilities!')
child._prior_prob = prob_planes[layer, x, y]
class Checkers_GUI:
"""A class to display a Pygame representation of the Checkers board."""
def __init__(self, game_env):
"""Initializes a Pygame GUI to visualize the game of Checkers."""
# Reference to game environment
self.game_env = game_env
# Initialize Pygame display
import pygame
self.pygame = pygame
pygame.init()
pygame.mixer.quit() # Fixes bug with high PyGame CPU usage
self.game_width = 600
self.game_height = 600
self.gameDisplay = pygame.display.set_mode((self.game_width,
self.game_height))
pygame.display.set_caption('Checkers')
# Define GUI variables
self.sq_dim = 50 # 50x50 pixel squares
self.board_width = 8*self.sq_dim # Pixels
self.board_height = 8*self.sq_dim # Pixels
self.board_offset = 100 # Pixels
self.move_delay = 1 # Seconds between animations
self.p1_pwin = 50
self.p2_pwin = 50
self.GREEN = (1,50,32)
self.RED = (135,14,14)
self.BLACK = (0,0,0)
self.BROWN = (101,67,33)
self.WHITE = (255,255,255)
self.BLUE = (3,53,252)
self.RED2 = (255,0,0)
self.BLACK2 = (128,120,120)
self.gridfont = pygame.font.SysFont('Segoe UI', 32)
self.statusfont = pygame.font.SysFont('Segoe UI', 28)
self.probfont = pygame.font.SysFont('Segoe UI', 20)
# Load images
self.board = pygame.image.load('img/board.png').convert_alpha()
self.red_checker = pygame.image.load('img/red_checker.png').convert_alpha()
self.black_checker = pygame.image.load('img/black_checker.png').convert_alpha()
self.red_king = pygame.image.load('img/red_king.png').convert_alpha()
self.black_king = pygame.image.load('img/black_king.png').convert_alpha()
self.ghost_checker = pygame.image.load('img/ghost_checker.png').convert_alpha()
self.ghost_king = pygame.image.load('img/ghost_king.png').convert_alpha()
self.move_checker = pygame.image.load('img/move_checker.png').convert_alpha()
self.move_king = pygame.image.load('img/move_king.png').convert_alpha()
self.select_sq = pygame.image.load('img/select_sq.png').convert_alpha()
self.move_sq = pygame.image.load('img/move_sq.png').convert_alpha()
# Draw display
self._set_board(game_env.state)
# Define if a human player is playing
self.human_player = False
def _set_board(self, state):
"""Blit Checkers pieces on GUI board in their starting positions."""
# Get current game state information
state = self.game_env.state
move_count = self.game_env.move_count+1
player = player = int(state[4,0,0])
player1_man = 'red'
player2_man = 'black'
player_mark = player1_man if player == 0 else player2_man
done = self.game_env.done
outcome = self.game_env.outcome
# Blit green background
self.pygame.draw.rect(self.gameDisplay, self.GREEN, (0,0,
self.game_width,
self.game_height))
# Blit game board
self.gameDisplay.blit(self.board,
(self.board_offset,self.board_offset),
(0, 0, self.board_width, self.board_height))
# Blit grid numbers for each row and column
for num in range(8):
grid_text = self.gridfont.render(str(num+1), True, self.WHITE)
self.gameDisplay.blit(grid_text,
(self.board_offset+self.sq_dim//2-5+num*self.sq_dim,
self.board_offset-30))
self.gameDisplay.blit(grid_text,
(self.board_offset-30,
self.board_offset+self.sq_dim//2-10+num*self.sq_dim))
# Blit status message
if not done:
status_str = 'Move #{}: It\'s now Player {}\'s turn ({})'.format(move_count, player+1, player_mark)
else:
status_str = 'Game over after {} moves! The outcome is: {}'.format(move_count, outcome)
status_text = self.statusfont.render(status_str, True, self.WHITE)
self.gameDisplay.blit(status_text,
(self.board_offset,
self.board_offset+self.board_height+30))
# Blit Checkers pieces
for row in range(0,8):
for col in range(0,8):
if row % 2 != col % 2: # Row and column not both odd or even
if row < 3:
self.gameDisplay.blit(self.red_checker,
(self.board_offset+self.sq_dim*col,
self.board_offset+self.sq_dim*row))
elif row > 4:
self.gameDisplay.blit(self.black_checker,
(self.board_offset+self.sq_dim*col,
self.board_offset+self.sq_dim*row))
self._blit_possible_moves(state)
self.update_screen()
def _blit_possible_moves(self, state):
"""Blits a visual indicator of which pieces can make a move, and the
locations that the pieces can move to.
"""
player = int(state[4,0,0])
player_idx = player * 2
possible_moves = np.sum(state[6:14], axis=0)
poss_x, poss_y = np.where(possible_moves > 0)
for x, y in zip(poss_x, poss_y): # Highlight pieces that can move
self.gameDisplay.blit(self.select_sq,
(self.board_offset+self.sq_dim*y,
self.board_offset+self.sq_dim*x))
# Place ghost pieces to indicate which squares pieces can move to
shift = [(-1,-1), (-1,1), (1,-1), (1,1), (-2,-2), (-2,2), (2,-2), (2,2)]
for idx in range(6,14):
move_x, move_y = np.where(state[idx] > 0)
for x, y in zip(move_x, move_y):
if state[player_idx, x, y] == 1:
ghost = self.ghost_checker
else:
ghost = self.ghost_king
self.gameDisplay.blit(ghost,
(self.board_offset+self.sq_dim*(y + shift[idx-6][1]),
self.board_offset+self.sq_dim*(x + shift[idx-6][0])))
def update_screen(self):
"""Update Pygame display after blitting operations."""
self.pygame.display.update()
def render(self, root_node=None, best_child=None):
"""Renders the new board state every time the step() method is called.
Blits an indicator of the move selected in the previous state, updates
the screen, pauses for a moment, and then redraws the game board with
the current state's piece positions and possible moves.
"""
# Get current game state information
prev_state = self.game_env.history[-2]
state = self.game_env.state
move_count = self.game_env.move_count
if move_count == 1 and not self.human_player:
self._blit_probs(root_node)
# Blit selection animation
old_xy, new_xy = self._states_to_piece_positions(prev_state, state)
self._blit_selected_move(prev_state, state, old_xy, new_xy)
self.update_screen()
time.sleep(self.move_delay)
# Blit board and pieces after move
self.gameDisplay.blit(self.board,
(self.board_offset,self.board_offset),
(0, 0, self.board_width, self.board_height))
self._blit_pieces(state)
self.update_screen()
time.sleep(self.move_delay)
# Blit next player's possible moves, prior probs, update status text
self._blit_possible_moves(state)
if not self.human_player: self._blit_probs(best_child)
self._blit_status(root_node, best_child)
self.update_screen()
def _states_to_piece_positions(self, prev_state, state):
"""Given a previous state and the current state, return two (x,y)
coordinates. The first coordinate will be the location of the piece
that was moved, and the second coordinate will be the location that
the piece moved to.
"""
board = prev_state[0] + 2*prev_state[1] + 3*prev_state[2] + 4*prev_state[3]
nboard = state[0] + 2*state[1] + 3*state[2] + 4*state[3]
board_diff = board - nboard
xnew, ynew = np.where(board_diff < 0)
xnew, ynew = xnew[0], ynew[0]
new_val = abs(nboard[xnew,ynew])
xold, yold = np.where(board_diff == new_val)
try:
xold, yold = xold[0], yold[0]
except IndexError: # Man promoted to king
new_val -= 1 # Value of man is 1 less than king
xold, yold = np.where(board_diff == new_val)
xold, yold = xold[0], yold[0]
return (xold,yold), (xnew,ynew)
def _blit_selected_move(self, prev_state, state, old_xy, new_xy):
"""Once player has selected a move, this function blits an indication
of which piece will be moved and the square it will be moved to.
"""
move_sq, sq_dim = self.move_sq, self.sq_dim
player = int(prev_state[4,0,0])
player_idx = player * 2
if state[player_idx, new_xy[0], new_xy[1]] == 1:
new_piece = self.move_checker
else:
new_piece = self.move_king
self.gameDisplay.blit(move_sq, (self.board_offset+sq_dim*old_xy[1],
self.board_offset+sq_dim*old_xy[0]))
self.gameDisplay.blit(new_piece, (self.board_offset+sq_dim*new_xy[1],
self.board_offset+sq_dim*new_xy[0]))
def _blit_pieces(self, state):
"""Iterates through the first four layers of the state array and blits
the corresponding game pieces to the board.
"""
piece = [self.red_checker, self.red_king,
self.black_checker, self.black_king]
for idx in range(4):
xloc, yloc = np.where(state[idx] == 1)
for x, y in zip(xloc, yloc):
self.gameDisplay.blit(piece[idx],
(self.board_offset+self.sq_dim*y,
self.board_offset+self.sq_dim*x))
def _blit_probs(self, node):
"""Blit the prior probability associated with each possible move."""
if self.game_env.neural_net is None: return
player = int(node.state[4,0,0])
color = self.RED if player == 0 else self.BLACK
shift = [(-1,-1), (-1,1), (1,-1), (1,1), (-2,-2), (-2,2), (2,-2), (2,2)]
offsetx = 20
offsety = 20
corner = [(offsetx*-np.sign(x[0]),offsety*-np.sign(x[1])) for x in shift]
for child in node.children:
layer = int(child.state[14,0,0])
x = child.state[14,0,1]
y = child.state[14,0,2]
prob = str(int(np.round(child.p*100, 0)))
prob_text = self.probfont.render(prob + '%', True, color)
self.gameDisplay.blit(prob_text,
(self.board_offset+self.sq_dim//2-10+corner[layer-6][1]+
(y + shift[layer-6][1])*self.sq_dim,
self.board_offset+self.sq_dim//2+-5+corner[layer-6][0]+
(x + shift[layer-6][0])*self.sq_dim))
def _blit_status(self, root_node, best_child):
"""Blit status text containing move count, current player, and player's
confidence in winning. If game is over, display outcome.
"""
move_count = self.game_env.move_count+1
prev_state = self.game_env.history[-2]
state = self.game_env.state
player = int(state[4,0,0])
prev_player = int(prev_state[4,0,0])
player1_man = 'red'
player2_man = 'black'
player_mark = player1_man if player == 0 else player2_man
done = self.game_env.done
outcome = self.game_env.outcome
self.pygame.draw.rect(self.gameDisplay, self.GREEN, (0,
self.board_height+self.board_offset,
self.game_width,
self.game_height-self.board_height-self.board_offset))
if not done:
status_str = 'Move #{}: It\'s now Player {}\'s turn ({})'.format(
move_count, player+1, player_mark)
status_text = self.statusfont.render(status_str, True, self.WHITE)
self.gameDisplay.blit(status_text, (self.board_offset,
self.board_offset+self.board_height+30))
if self.game_env.neural_net and not self.human_player:
if prev_player == 0:
self.p1_pwin = best_child.pwin
else:
self.p2_pwin = best_child.pwin
status_str = 'P1/P2\'s confidence in winning: {:4.1f}%'\
' / {:4.1f}%'.format(self.p1_pwin, self.p2_pwin)
status_text = self.statusfont.render(status_str, True, self.WHITE)
self.gameDisplay.blit(status_text, (self.board_offset,
self.board_offset+self.board_height+62))
else:
status_str = 'Game over after {} moves!'.format(move_count)
status_text = self.statusfont.render(status_str, True, self.WHITE)
self.gameDisplay.blit(status_text, (self.board_offset,
self.board_offset+self.board_height+30))
status_str = 'The outcome is: {}'.format(outcome)
status_text = self.statusfont.render(status_str, True, self.WHITE)
self.gameDisplay.blit(status_text, (self.board_offset,
self.board_offset+self.board_height+62))
def _create_board_image(self):
"""Assemble Checkers board into a single image and save to disk.
Only needs to be run when a change to the Checkers board design is
desired.
"""
board = self.pygame.Surface([self.board_width, self.board_height])
wood = self.pygame.image.load('img/wood5.png').convert()
black_sq = self.pygame.Surface((self.sq_dim, self.sq_dim))
black_sq.set_alpha(160)
black_sq.fill(self.BROWN)
board.blit(wood,(0,0), (0, 0, self.board_width, self.board_height))
for row in range(8):
for col in range(8):
if row % 2 != col % 2:
board.blit(black_sq, (self.sq_dim*col, self.sq_dim*row))
self.pygame.image.save(board, 'img/board.png')
def close_gui(self):
"""Close Pygame GUI."""
self.pygame.quit()
def test_game():
"""Test function to validate functionality of Checkers class.
Plays a game of Checkers by randomly selecting moves for both players.
"""
GUI = True
checker = Checkers()
checker.print_board()
if GUI: checker_gui = Checkers_GUI(checker)
input('Press enter to continue...')
while not checker.done:
legal_next_states = checker.get_legal_next_states(checker.history)
move_idx = np.random.randint(0,len(legal_next_states))
checker.step(legal_next_states[move_idx])
checker.print_board()
if GUI: checker_gui.render()
input('Press enter to continue...')
if checker.done and GUI: checker_gui.close_gui()
if __name__ == '__main__':
test_game()