-
Notifications
You must be signed in to change notification settings - Fork 1
/
Game_setting.py
executable file
·201 lines (180 loc) · 8.23 KB
/
Game_setting.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
import logging
from my_moduler import get_module_logger
import util_ability
from my_enum import *
import numpy as np
#import matplotlib.pyplot as plt
mylogger = get_module_logger(__name__)
class Game:
def mulligan(self, Player1, Player2, virtual=False):
assert Player1.player_num != Player2.player_num, "same error"
Player1.mulligan(Player1.deck, virtual=virtual)
if not virtual:
print("")
Player2.mulligan(Player2.deck, virtual=virtual)
assert len(Player1.hand) == 3 and len(Player2.hand) == 3,"{},{}".format(len(Player1.hand),len(Player2.hand))
def start(self, f, virtual_flg=False,history_flg = False):
turn = 1
win, lose, lib_num = 0, 0, 0
f.secret = bool(virtual_flg)
self.mulligan(f.players[0], f.players[1], virtual=virtual_flg)
while True:
end_flg = False
# (win,lose,lib_num,turn,end_flg)=self.play_turn(f,0,win,lose,lib_num,turn,virtual_flg)
(win, lose, lib_num, turn, end_flg) = f.play_turn(0, win, lose, lib_num, turn, virtual_flg)
if end_flg:
break
(win, lose, lib_num, turn, end_flg) = f.play_turn(1, win, lose, lib_num, turn, virtual_flg)
# (win,lose,lib_num,turn,end_flg)=self.play_turn(f,1,win,lose,lib_num,turn,virtual_flg)
if end_flg:
break
for player in f.players:
policy = player.policy
if policy.policy_type in [3,4]:
policy.current_node = None
policy.prev_node = None
if history_flg:
#mylogger.info("state_value_histroy")
#for i,value in enumerate(f.state_value_history):
# mylogger.info("Turn {}:{:.3f}".format(i+1,value))
turns = np.array(range(len(f.state_value_history)))
state_values = f.state_value_history
#if f.players[0].name == "Bob":
# state_values = [1 - value for value in state_values]
state_values = np.array(state_values)
return win, lose, lib_num, turn, [turns,state_values]
return win, lose, lib_num, turn
def start_for_train_data(self, f, virtual_flg=False,target_player_num=0):
turn = 1
win, lose, lib_num = 0, 0, 0
f.secret = bool(virtual_flg)
self.mulligan(f.players[0], f.players[1], virtual=virtual_flg)
train_datas = []
reward = 0.0
accumulate_turn = 0
while True:
(win, lose, end_flg, train_data) = f.play_turn_for_train(0)
if target_player_num == 0:
train_datas.extend(train_data)
if end_flg:
break
(win, lose, end_flg, train_data) = f.play_turn_for_train(1)
if target_player_num == 1:
train_datas.extend(train_data)
if end_flg:
break
accumulate_turn += 1
assert accumulate_turn < 100,"infinite loop"
for player in f.players:
policy = player.policy
if policy.policy_type in [3,4]:
policy.current_node = None
policy.prev_node = None
#reward = int(target_player_num == 0)*(2*win-1) + 1 - win
reward = win if target_player_num == 0 else lose
return train_datas, reward
def start_for_dual(self, f, virtual_flg=False,target_player_num=0):
turn = 1
win, lose, lib_num = 0, 0, 0
f.secret = bool(virtual_flg)
f.players[0].draw(f.players[0].deck, 3)
f.players[1].draw(f.players[1].deck, 3)
self.mulligan(f.players[0], f.players[1], virtual=virtual_flg)
train_datas = [[],[]]
#reward = 0.0
accumulate_turn = 0
while True:
(win, lose, end_flg, train_data) = f.play_turn_for_dual(0)
train_datas[0].extend(train_data)
#if target_player_num == 0:
# train_datas.extend(train_data)
if end_flg:
break
(win, lose, end_flg, train_data) = f.play_turn_for_dual(1)
train_datas[1].extend(train_data)
#if target_player_num == 1:
# train_datas.extend(train_data)
if end_flg:
break
accumulate_turn += 1
assert accumulate_turn < 100,"infinite loop\n{}".format(f.get_observable_data(player_num=target_player_num))
#print("accumulate_turn:{}".format(accumulate_turn))
for player in f.players:
policy = player.policy
if policy.policy_type in [3,4]:
policy.current_node = None
policy.prev_node = None
#reward = win if target_player_num == 0 else lose
#reward = 2*reward - 1.0
reward = [2*win-1,2*lose-1]
return train_datas, reward
from collections import namedtuple
Detail_State_data = namedtuple('Value', ('hand_card_categories_and_ids', 'hand_card_costs', 'follower_card_ids',
'amulet_card_ids', 'follower_stats', 'follower_abilities', 'follower_is_evolved',
'life_data'))
import Embedd_Network_model
def get_data(field, player_num=0):
return Embedd_Network_model.get_data(field,player_num=player_num)
"""
hand_card_categories_and_ids = []
hand_card_costs = []
for hand_card in f.players[0].hand:
hand_card_categories_and_ids.append(Card_Category[hand_card.card_category].value * (hand_card.card_id + 500))
hand_card_costs.append(hand_card.cost)
for j in range(len(f.players[0].hand), 9):
hand_card_categories_and_ids.append(0)
hand_card_costs.append(0)
follower_card_ids = []
amulet_card_ids = []
follower_stats = []
follower_abilities = []
follower_is_evolved = f.get_able_to_evo(0)
for i in range(2):
for card in f.card_location[i]:
if card.card_category == "Creature":
follower_card_ids.append(card.card_id + 500)
follower_stats.extend([card.power, card.get_current_toughness()])
follower_abilities.append(card.ability[:])
amulet_card_ids.append(0)
else:
follower_card_ids.append(0)
follower_stats.extend([0, 0])
follower_abilities.append([])
amulet_card_ids.append(card.card_id + 500)
for k in range(len(f.card_location[i]), 5):
follower_card_ids.append(0)
follower_stats.extend([0, 0])
follower_abilities.append([])
amulet_card_ids.append(0)
life_data = [f.players[0].life, f.players[1].life, f.current_turn[0]]
datas = Detail_State_data(hand_card_categories_and_ids, hand_card_costs, follower_card_ids, amulet_card_ids,
follower_stats, follower_abilities, follower_is_evolved, life_data)
return datas
input_field_data = []
for hand_card in f.players[player_num].hand:
input_field_data.append(Card_Category[hand_card.card_category].value)
input_field_data.append(hand_card.cost)
input_field_data.append(hand_card.card_id)
input_field_data.extend([0,0,0]*(9-len(f.players[player_num].hand)))
for side_num in range(2):
i = (side_num + player_num) % 2
for card in f.card_location[i]:
if card.card_category == "Creature":
input_field_data.append(card.card_id)
input_field_data.extend([card.power, card.get_current_toughness()])
#embed_ability = [int(ability_id in card.ability) for ability_id in range(1, 16)]
input_field_data.append(card.ability[:])
# input_field_data.extend([card.card_id, card.power, card.get_current_toughness(),
# int(KeywordAbility.WARD.value in card.ability)])
else:
input_field_data.append(0)
input_field_data.extend([0, 0])
input_field_data.append([])
for k in range(len(f.card_location[i]), 5):
input_field_data.append(0)
input_field_data.extend([0, 0])
input_field_data.append([])
input_field_data.extend([f.players[player_num].life, f.players[1-player_num].life, f.current_turn[player_num]])
#print(input_field_data)
return input_field_data
"""