def test_find_even_three():
hand_card = [
3, 3, 3, 4, 4, 4, 5, 6, 6, 6, 7, 7, 7, 8, 8, 8, 10, 10, 10, 11, 12, 12,
12, 13, 13, 13, 14, 14, 14
]
k = 3
hand_card_status = HandCardUtils.obtain_hand_card_status(hand_card)
res = HandCardUtils.find_even_pair(hand_card_status, k)
print(res)
def test_find_continues():
hand_card = [
3, 3, 3, 4, 4, 4, 5, 6, 6, 6, 7, 7, 7, 8, 8, 8, 10, 10, 10, 11, 12, 12,
12, 13, 13, 13, 14, 14, 14
]
k = 5
hand_card_status = HandCardUtils.obtain_hand_card_status(hand_card)
res = HandCardUtils.find_continues(hand_card_status, k)
print(res)
def _calc_put_card_seq_score(self, hand_card):
ctj = CardTypeJudge()
score = 0
for item in hand_card:
hand_card_status = HandCardUtils.obtain_hand_card_status(item)
cts = ctj.judge_card_type(hand_card_status)
one_hand_score = HandCardUtils.value_map(cts.primary_item,
cts.card_type,
cts.card_count)
score += one_hand_score
return score
def _gen_agent(self):
#random.shuffle(all_card.ALL_CARD_NO_COLOR)
#main_agent = all_card.ALL_CARD_NO_COLOR[:20]
#low_agent = all_card.ALL_CARD_NO_COLOR[20:38]
#up_agent = all_card.ALL_CARD_NO_COLOR[38:]
main_agent = [
3, 3, 4, 4, 6, 7, 7, 7, 8, 10, 10, 11, 11, 12, 12, 12, 12, 15, 15,
16
]
low_agent = [3, 3, 4, 5, 8, 8, 8, 9, 9, 10, 11, 13, 13, 13, 14, 14, 15]
up_agent = [4, 5, 5, 5, 6, 6, 6, 7, 9, 9, 10, 11, 13, 14, 14, 15, 17]
main_agent_status = HandCardUtils.obtain_hand_card_status(main_agent)
low_agent_status = HandCardUtils.obtain_hand_card_status(low_agent)
up_agent_status = HandCardUtils.obtain_hand_card_status(up_agent)
return main_agent_status, low_agent_status, up_agent_status
def test_find_even_pair():
#hand_card = [3,3,4,4,5,6,6,7,7,8,8,10,10,11,12,12,13,13,14,14,15,15,16]
hand_card_status = [0, 0, 0, 2, 2, 2, 1, 1, 1, 1, 1, 1, 2, 3, 0, 0, 0, 0]
k = 3
#hand_card_status = HandCardUtils.obtain_hand_card_status(hand_card)
res = HandCardUtils.find_even_pair(hand_card_status, k)
print(res)
def hand_card_color_seq(self, color_seq):
self._hand_card_color_seq = sorted(color_seq,
key=lambda x: x[1],
reverse=False)
# 将有花色序列转换为无花色序列
self._hand_card_seq = list(
map(lambda x: x[1], self._hand_card_color_seq))
self._color_seq = list(map(lambda x: x[0], self._hand_card_color_seq))
# 记录每张牌的状态
self._hand_card_status = HandCardUtils.obtain_hand_card_status(
self._hand_card_seq)
# 记录每张牌的个数
self._hand_card_count = sum(self._hand_card_status)
# 得到手牌的初始价值
self._hand_card_value = HandCardUtils.hand_card_init_value(
self._hand_card_seq)
def obtain_reward(hand_card_status, surround, *args):
info = dict()
put_card_status = surround.has_put_card_status
last_player_role, last_card_type_struct, last_action = args[0], args[
1], args[2]
if len(args) != 3:
last_player_role, last_card_type_struct, last_action = None, None, None
primary_item = None
if last_card_type_struct is not None:
primary_item = last_card_type_struct.primary_item
exist_card = list(
map(lambda x: x[0],
filter(lambda x: x[1] >= 1, enumerate(hand_card_status))))
if isinstance(exist_card, int):
exist_card = [exist_card]
if primary_item is not None:
exist_card = list(filter(lambda x: x > primary_item, exist_card))
if isinstance(exist_card, int):
exist_card = [exist_card]
if len(exist_card) == 0:
info['error'] = True
else:
rnd = random.randint(0, len(exist_card) - 1)
put_card = exist_card[rnd]
score = HandCardUtils.value_map(put_card, CardTypeEnum.CT_ONE, 1)
hand_card_status[put_card] -= 1
put_card_status[put_card] += 1
info['put_card'] = [put_card]
info['primary_item'] = put_card
return score, info
def step(self, action):
score = 0
ERR_CARD_SCORE = -1
done = False
if ENV_DEBUG:
print('Action: {}'.format(action))
# 单牌
if action == ActionTypeEnum.ACTION_PUT_ONE.value:
exist_card = list(
map(
lambda x: x[0],
filter(lambda x: x[1] >= 1,
enumerate(self.hand_card_status))))
if isinstance(exist_card, int):
exist_card = [exist_card]
rnd = random.randint(0, len(exist_card) - 1)
put_card = exist_card[rnd]
score = HandCardUtils.value_map(put_card, CardTypeEnum.CT_ONE, 1)
self.observation[put_card] -= 1
self.observation[18 + put_card] += 1
self.hand_card_status[put_card] -= 1
if ENV_DEBUG:
print('Put card %s' % put_card)
# 对子
elif action == ActionTypeEnum.ACTION_PUT_DOU.value:
exist_card = list(
map(
lambda x: x[0],
filter(lambda x: x[1] == 2,
enumerate(self.hand_card_status))))
if isinstance(exist_card, int):
exist_card = [exist_card]
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
rnd = random.randint(0, len(exist_card) - 1)
put_card = exist_card[rnd]
score = HandCardUtils.value_map(put_card, CardTypeEnum.CT_DOU,
2)
self.observation[put_card] -= 2
self.observation[18 + put_card] += 2
self.hand_card_status[put_card] -= 2
if ENV_DEBUG:
print('Put card %s,%s' % (put_card, put_card))
# 三不带
elif action == ActionTypeEnum.ACTION_PUT_THREE.value:
exist_card = list(
map(
lambda x: x[0],
filter(lambda x: x[1] == 3,
enumerate(self.hand_card_status))))
if isinstance(exist_card, int):
exist_card = [exist_card]
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
rnd = random.randint(0, len(exist_card) - 1)
put_card = exist_card[rnd]
score = HandCardUtils.value_map(put_card,
CardTypeEnum.CT_THREE, 3)
self.observation[put_card] -= 3
self.observation[18 + put_card] += 3
self.hand_card_status[put_card] -= 3
if ENV_DEBUG:
print('Put card %s,%s,%s' % (put_card, put_card, put_card))
# 三带一
elif action == ActionTypeEnum.ACTION_PUT_THREE_ONE.value:
exist_card = list(
map(
lambda x: x[0],
filter(lambda x: x[1] == 3,
enumerate(self.hand_card_status))))
if isinstance(exist_card, int):
exist_card = [exist_card]
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
rnd = random.randint(0, len(exist_card) - 1)
put_card = exist_card[rnd]
other_exist_card = list(
map(
lambda x: x[0],
filter(lambda x: x[1] >= 1 and x[0] != put_card,
enumerate(self.hand_card_status))))
if isinstance(other_exist_card, int):
other_exist_card = [other_exist_card]
if len(other_exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
rnd = random.randint(0, len(other_exist_card) - 1)
other_put_card = other_exist_card[rnd]
score = HandCardUtils.value_map(put_card,
CardTypeEnum.CT_THREE_ONE,
4)
self.observation[put_card] -= 3
self.observation[18 + put_card] += 3
self.hand_card_status[put_card] -= 3
self.observation[other_put_card] -= 1
self.observation[18 + other_put_card] += 1
self.hand_card_status[other_put_card] -= 1
if ENV_DEBUG:
print('Put card %s,%s,%s,%s' %
(put_card, put_card, put_card, other_put_card))
# 三带一对
elif action == ActionTypeEnum.ACTION_PUT_THREE_DOU.value:
exist_card = list(
map(
lambda x: x[0],
filter(lambda x: x[1] == 3,
enumerate(self.hand_card_status))))
if isinstance(exist_card, int):
exist_card = [exist_card]
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
rnd = random.randint(0, len(exist_card) - 1)
put_card = exist_card[rnd]
other_exist_card = list(
map(
lambda x: x[0],
filter(lambda x: x[1] >= 2 and x[0] != put_card,
enumerate(self.hand_card_status))))
if isinstance(other_exist_card, int):
other_exist_card = [other_exist_card]
if len(other_exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
rnd = random.randint(0, len(other_exist_card) - 1)
other_put_card = other_exist_card[rnd]
score = HandCardUtils.value_map(put_card,
CardTypeEnum.CT_THREE_DOU,
5)
self.observation[put_card] -= 3
self.observation[18 + put_card] += 3
self.hand_card_status[put_card] -= 3
self.observation[other_put_card] -= 2
self.observation[18 + other_put_card] += 2
self.hand_card_status[other_put_card] -= 2
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s' %
(put_card, put_card, put_card, other_put_card,
other_put_card))
# 两连对
elif action == ActionTypeEnum.ACTION_PUT_2_DOU.value:
exist_card = HandCardUtils.find_even_pair(self.hand_card_status,
k=2)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_DOU, 4)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 2
self.observation[18 + put_card] += 2
self.hand_card_status[put_card] -= 2
log_put.extend([put_card, put_card])
if ENV_DEBUG:
print('Put card %s,%s,%s,%s' % tuple(log_put))
# 三连对
elif action == ActionTypeEnum.ACTION_PUT_3_DOU.value:
exist_card = HandCardUtils.find_even_pair(self.hand_card_status,
k=3)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_DOU, 6)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 2
self.observation[18 + put_card] += 2
self.hand_card_status[put_card] -= 2
log_put.extend([put_card, put_card])
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s' % tuple(log_put))
# 四连对
elif action == ActionTypeEnum.ACTION_PUT_4_DOU.value:
exist_card = HandCardUtils.find_even_pair(self.hand_card_status,
k=4)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_DOU, 8)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 2
self.observation[18 + put_card] += 2
self.hand_card_status[put_card] -= 2
log_put.extend([put_card, put_card])
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s,%s,%s' % tuple(log_put))
# 五连对
elif action == ActionTypeEnum.ACTION_PUT_5_DOU.value:
exist_card = HandCardUtils.find_even_pair(self.hand_card_status,
k=5)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_DOU, 10)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 2
self.observation[18 + put_card] += 2
self.hand_card_status[put_card] -= 2
log_put.extend([put_card, put_card])
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s,%s,%s,%s,%s' %
tuple(log_put))
# 两连三不带
elif action == ActionTypeEnum.ACTION_PUT_2_THREE.value:
exist_card = HandCardUtils.find_even_three(self.hand_card_status,
k=2)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_THREE, 6)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 3
self.observation[18 + put_card] += 3
self.hand_card_status[put_card] -= 3
log_put.extend([put_card, put_card, put_card])
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s' % tuple(log_put))
# 三连三不带
elif action == ActionTypeEnum.ACTION_PUT_3_THREE.value:
exist_card = HandCardUtils.find_even_three(self.hand_card_status,
k=3)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_THREE, 9)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 3
self.observation[18 + put_card] += 3
self.hand_card_status[put_card] -= 3
log_put.extend([put_card, put_card, put_card])
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s,%s,%s,%s' %
tuple(log_put))
# 两连三带一
elif action == ActionTypeEnum.ACTION_PUT_2_THREE_ONE.value:
K = 2
exist_card = HandCardUtils.find_even_three(self.hand_card_status,
k=K)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
other_card = list(
map(
lambda x: x[0],
filter(lambda x: x[0] not in exist_card,
enumerate(self.hand_card_status))))
if len(other_card) < K:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
tmp = list()
for _ in range(K):
rnd = random.randint(0, len(other_card) - 1)
tmp_card = other_card[rnd]
tmp.append(tmp_card)
self.observation[tmp_card] -= 1
self.observation[18 + tmp_card] += 1
self.hand_card_status[tmp_card] -= 1
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_THREE_ONE,
3 * K + K)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 3
self.observation[18 + put_card] += 3
self.hand_card_status[put_card] -= 3
log_put.extend([put_card, put_card, put_card])
log_put.extend(tmp)
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s,%s,%s' %
tuple(log_put))
# 三连三带一
elif action == ActionTypeEnum.ACTION_PUT_3_THREE_ONE.value:
K = 3
exist_card = HandCardUtils.find_even_three(self.hand_card_status,
k=K)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
other_card = list(
map(
lambda x: x[0],
filter(lambda x: x[0] not in exist_card,
enumerate(self.hand_card_status))))
if len(other_card) < K:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
tmp = list()
for _ in range(K):
rnd = random.randint(0, len(other_card) - 1)
tmp_card = other_card[rnd]
tmp.append(tmp_card)
self.observation[tmp_card] -= 1
self.observation[18 + tmp_card] += 1
self.hand_card_status[tmp_card] -= 1
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_THREE_ONE,
3 * K + K)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 3
self.observation[18 + put_card] += 3
self.hand_card_status[put_card] -= 3
log_put.extend([put_card, put_card, put_card])
log_put.extend(tmp)
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s' %
tuple(log_put))
# 两连三带一对
elif action == ActionTypeEnum.ACTION_PUT_2_THREE_DOU.value:
K = 2
exist_card = HandCardUtils.find_even_three(self.hand_card_status,
k=K)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
other_card = list(
map(
lambda x: x[0],
filter(lambda x: x[0] not in exist_card and x[1] == 2,
enumerate(self.hand_card_status))))
if len(other_card) < K:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
tmp = list()
for _ in range(K):
rnd = random.randint(0, len(other_card) - 1)
tmp_card = other_card[rnd]
tmp.extend([tmp_card, tmp_card])
self.observation[tmp_card] -= 2
self.observation[18 + tmp_card] += 2
self.hand_card_status[tmp_card] -= 2
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_THREE_DOU,
3 * K + 2 * K)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 3
self.observation[18 + put_card] += 3
self.hand_card_status[put_card] -= 3
log_put.extend([put_card, put_card, put_card])
log_put.extend(tmp)
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s,%s,%s,%s,%s' %
tuple(log_put))
# 三连三带一对
elif action == ActionTypeEnum.ACTION_PUT_3_THREE_DOU.value:
K = 3
exist_card = HandCardUtils.find_even_three(self.hand_card_status,
k=K)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
other_card = list(
map(
lambda x: x[0],
filter(lambda x: x[0] not in exist_card and x[1] == 2,
enumerate(self.hand_card_status))))
if len(other_card) < K:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
tmp = list()
for _ in range(K):
rnd = random.randint(0, len(other_card) - 1)
tmp_card = other_card[rnd]
tmp.extend([tmp_card, tmp_card])
self.observation[tmp_card] -= 2
self.observation[18 + tmp_card] += 2
self.hand_card_status[tmp_card] -= 2
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_THREE_DOU,
3 * K + 2 * K)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 3
self.observation[18 + put_card] += 3
self.hand_card_status[put_card] -= 3
log_put.extend([put_card, put_card, put_card])
log_put.extend(tmp)
if ENV_DEBUG:
print(
'Put card %s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s'
% tuple(log_put))
# 四带二单
elif action == ActionTypeEnum.ACTION_PUT_FOUR_ONE.value:
exist_card = list(
map(
lambda x: x[0],
filter(lambda x: x[1] == 4,
enumerate(self.hand_card_status))))
if isinstance(exist_card, int):
exist_card = [exist_card]
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
rnd = random.randint(0, len(exist_card) - 1)
put_card = exist_card[rnd]
other_exist_card = list(
map(
lambda x: x[0],
filter(lambda x: x[1] >= 1 and x[0] != put_card,
enumerate(self.hand_card_status))))
if isinstance(other_exist_card, int):
other_exist_card = [other_exist_card]
if len(other_exist_card) <= 1:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(put_card,
CardTypeEnum.CT_FOUR_ONE,
6)
self.observation[put_card] -= 4
self.observation[18 + put_card] += 4
self.hand_card_status[put_card] -= 4
pre = -1
t = 0
tmp = list()
while t < 2:
rnd = random.randint(0, len(other_exist_card) - 1)
while True:
if rnd != pre:
pre = rnd
break
rnd = random.randint(0, len(other_exist_card) - 1)
other_put_card = other_exist_card[rnd]
tmp.append(other_put_card)
self.observation[other_put_card] -= 1
self.observation[18 + other_put_card] += 1
self.hand_card_status[other_put_card] -= 1
t += 1
log_put = [put_card] * 4
log_put.extend(tmp)
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s' % tuple(log_put))
# 四带二对
elif action == ActionTypeEnum.ACTION_PUT_FOUR_DOU.value:
exist_card = list(
map(
lambda x: x[0],
filter(lambda x: x[1] == 4,
enumerate(self.hand_card_status))))
if isinstance(exist_card, int):
exist_card = [exist_card]
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
rnd = random.randint(0, len(exist_card) - 1)
put_card = exist_card[rnd]
other_exist_card = list(
map(
lambda x: x[0],
filter(lambda x: x[1] >= 2 and x[0] != put_card,
enumerate(self.hand_card_status))))
if isinstance(other_exist_card, int):
other_exist_card = [other_exist_card]
if len(other_exist_card) <= 1:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(put_card,
CardTypeEnum.CT_FOUR_DOU,
8)
self.observation[put_card] -= 4
self.observation[18 + put_card] += 4
self.hand_card_status[put_card] -= 4
pre = -1
t = 0
tmp = list()
while t < 2:
rnd = random.randint(0, len(other_exist_card) - 1)
while True:
if rnd != pre:
pre = rnd
break
rnd = random.randint(0, len(other_exist_card) - 1)
other_put_card = other_exist_card[rnd]
tmp.extend([other_put_card, other_put_card])
self.observation[other_put_card] -= 2
self.observation[18 + other_put_card] += 2
self.hand_card_status[other_put_card] -= 2
t += 1
log_put = [put_card] * 4
log_put.extend(tmp)
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s,%s,%s' %
tuple(log_put))
# 连子(5)
elif action == ActionTypeEnum.ACTION_PUT_5_CONTINUE.value:
K = 5
exist_card = HandCardUtils.find_continues(self.hand_card_status,
k=K)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_CONTINUE, K)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 1
self.observation[18 + put_card] += 1
self.hand_card_status[put_card] -= 1
log_put.extend([put_card])
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s' % tuple(log_put))
# 连子(6)
elif action == ActionTypeEnum.ACTION_PUT_6_CONTINUE.value:
K = 6
exist_card = HandCardUtils.find_continues(self.hand_card_status,
k=K)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_CONTINUE, K)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 1
self.observation[18 + put_card] += 1
self.hand_card_status[put_card] -= 1
log_put.extend([put_card])
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s' % tuple(log_put))
# 连子(7)
elif action == ActionTypeEnum.ACTION_PUT_7_CONTINUE.value:
K = 7
exist_card = HandCardUtils.find_continues(self.hand_card_status,
k=K)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_CONTINUE, K)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 1
self.observation[18 + put_card] += 1
self.hand_card_status[put_card] -= 1
log_put.extend([put_card])
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s,%s' % tuple(log_put))
# 连子(8)
elif action == ActionTypeEnum.ACTION_PUT_8_CONTINUE.value:
K = 8
exist_card = HandCardUtils.find_continues(self.hand_card_status,
k=K)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_CONTINUE, K)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 1
self.observation[18 + put_card] += 1
self.hand_card_status[put_card] -= 1
log_put.extend([put_card])
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s,%s,%s' % tuple(log_put))
# 连子(9)
elif action == ActionTypeEnum.ACTION_PUT_9_CONTINUE.value:
K = 9
exist_card = HandCardUtils.find_continues(self.hand_card_status,
k=K)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_CONTINUE, K)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 1
self.observation[18 + put_card] += 1
self.hand_card_status[put_card] -= 1
log_put.extend([put_card])
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s,%s,%s,%s' %
tuple(log_put))
# 连子(10)
elif action == ActionTypeEnum.ACTION_PUT_10_CONTINUE.value:
K = 10
exist_card = HandCardUtils.find_continues(self.hand_card_status,
k=K)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_CONTINUE, K)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 1
self.observation[18 + put_card] += 1
self.hand_card_status[put_card] -= 1
log_put.extend([put_card])
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s,%s,%s,%s,%s' %
tuple(log_put))
# 连子(11)
elif action == ActionTypeEnum.ACTION_PUT_11_CONTINUE.value:
K = 11
exist_card = HandCardUtils.find_continues(self.hand_card_status,
k=K)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_CONTINUE, K)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 1
self.observation[18 + put_card] += 1
self.hand_card_status[put_card] -= 1
log_put.extend([put_card])
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s' %
tuple(log_put))
# 连子(12)
elif action == ActionTypeEnum.ACTION_PUT_12_CONTINUE.value:
K = 12
exist_card = HandCardUtils.find_continues(self.hand_card_status,
k=K)
if len(exist_card) == 0:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
score = HandCardUtils.value_map(exist_card[-1],
CardTypeEnum.CT_CONTINUE, K)
log_put = list()
for put_card in exist_card:
self.observation[put_card] -= 1
self.observation[18 + put_card] += 1
self.hand_card_status[put_card] -= 1
log_put.extend([put_card])
if ENV_DEBUG:
print('Put card %s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s' %
tuple(log_put))
# 炸弹
elif action == ActionTypeEnum.ACTION_PUT_BOMB.value:
qu_ja = [CardEnum.QU.value, CardEnum.JA.value]
comm_bomb = list(
map(
lambda x: x[0],
filter(lambda x: x[1] == 4,
enumerate(self.hand_card_status))))
master_bomb = list(
map(
lambda x: x[0],
filter(lambda x: x[0] in qu_ja,
enumerate(self.hand_card_status))))
if isinstance(comm_bomb, int):
comm_bomb = [comm_bomb]
if isinstance(master_bomb, int):
master_bomb = [master_bomb]
if len(comm_bomb) == 0 and len(master_bomb) < 2:
score = ERR_CARD_SCORE
done = True
if ENV_DEBUG:
print('Can not accept the card')
else:
rnd = len(comm_bomb)
put_card = qu_ja
if len(comm_bomb) >= 1 and len(master_bomb) == 2:
rnd = random.randint(0, len(comm_bomb))
elif len(comm_bomb) >= 1:
rnd = random.randint(0, len(comm_bomb) - 1)
if rnd < len(comm_bomb):
put_card = comm_bomb[rnd]
card_count = 4 if isinstance(put_card, int) else 2
if card_count == 2:
score = HandCardUtils.value_map(put_card[-1],
CardTypeEnum.CT_BOMB,
card_count)
for item in put_card:
self.observation[item] -= 1
self.observation[18 + item] += 1
self.hand_card_status[item] -= 1
if ENV_DEBUG:
print('Put card %s,%s' % tuple(put_card))
else:
score = HandCardUtils.value_map(put_card,
CardTypeEnum.CT_BOMB,
card_count)
self.observation[put_card] -= 4
self.observation[18 + put_card] += 4
self.hand_card_status[put_card] -= 4
if ENV_DEBUG:
print('Put card %s,%s,%s,%s' %
(put_card, put_card, put_card, put_card))
# 不出
elif action == ActionTypeEnum.ACTION_NO_PUT.value:
pass
if self._is_done():
score += 1
if ENV_DEBUG:
print('Current card: %s' % self.hand_card_status)
print('Current env: %s' % self.observation)
print('is game over: %s' % (self._is_done() or done))
return self.observation[3:], score, self._is_done() or done
def put_card_process(self,
sess,
env,
curr_player_id,
net_out,
last_card_type_struct=None,
last_action=None):
curr_player = self._players[curr_player_id]
print('玩家[ID=%s ROLE=%s]当前手牌: %s' %
(curr_player_id, self.get_player_role_title(curr_player_id),
curr_player.hand_card_struct.hand_card_seq))
if curr_player_id == None:
#if curr_player_id == human_player_id:
while True:
input_card = input('请玩家[ID=%s]出牌(多张牌以空格分隔)' % (curr_player_id))
if len(input_card) == 0 and last_action is not None:
return None, [], None
input_card = input_card.split()
input_card_temp = []
new_dict = {v: k for k, v in CARD_MAP.items()}
for card in input_card:
card_temp = card.capitalize()
if card_temp in new_dict:
input_card_temp.append(new_dict[card_temp])
elif card.isdigit() and int(card) >= 0 and int(card) <= 10:
input_card_temp.append(int(card))
input_card = list(map(lambda x: int(x), input_card_temp))
print(input_card)
is_contain = HandCardUtils.is_contain_card(
curr_player.hand_card_struct.hand_card_status, input_card)
is_find, cts = HandCardUtils.is_one_hand(input_card)
is_meet = True
print(is_contain)
print(is_find)
print(is_meet)
if is_contain and is_find and last_action:
is_meet = False
last_primary_item = last_card_type_struct.primary_item
card_type = cts.card_type
primary_item = cts.primary_item
card_count = cts.card_count
if last_action == ActionTypeEnum.ACTION_PUT_BOMB.value:
if card_type == last_action and card_count == 2:
is_meet = True
elif card_type == last_action and card_count == 4 and primary_item > last_primary_item:
is_meet = True
else:
if card_type == ActionTypeEnum.ACTION_PUT_BOMB.value:
is_meet = True
elif card_type == last_action and primary_item > last_primary_item:
is_meet = True
if is_contain and is_find and is_meet:
self._update_curr_player(curr_player_id, input_card)
if len(curr_player.hand_card_struct.hand_card_seq) == 0:
self._game_is_over = True
return cts.card_type, input_card, cts.primary_item
else:
print('所出牌型不合理')
action_reward = list()
n_action = config.N_ACTION
n_input = config.N_INPUT
hcs = curr_player.hand_card_struct
hand_card_status = hcs.hand_card_status
put_card_status = self._put_card_status
obser = env.specify_env(hand_card_status, put_card_status,
curr_player.player_role)
#print('curr obser:{}'.format(obser))
for action in range(n_action):
act1 = [0] * n_action
act1[action] = 1
act2 = [0] * n_action
act3 = [0] * n_action
candidate = np.random.choice([action, 35, 36], 1)[0]
act2[candidate] = 1
if candidate != 36:
candidate = np.random.choice([candidate, 35, 36], 1)[0]
else:
candidate = np.random.choice([action, 35, 36], 1)[0]
act3[candidate] = 1
act = act1 + act2 + act3
x = copy.deepcopy(obser)
x.extend(act)
x = np.reshape(x, [1, n_input + n_action * 3])
out = sess.run(net_out, feed_dict={'input_x:0': x})[0][0]
action_reward.append((action, out))
order_action = list(
map(lambda x: x[0],
sorted(action_reward, key=lambda x: x[1], reverse=True)))
last_primary_item = None
if last_card_type_struct is not None:
last_action = last_card_type_struct.card_type
last_primary_item = last_card_type_struct.primary_item
order_action = list(
filter(
lambda x: x == last_action or x == ActionTypeEnum.
ACTION_PUT_BOMB.value, order_action))
else:
order_action.remove(ActionTypeEnum.ACTION_NO_PUT.value)
#print('可出牌型: {}'.format(order_action))
can_accept = False
accpet_action = None
for action in order_action:
obser, _, done, info = env.step(action, last_primary_item,
last_action)
error = info['error']
if not error:
if done:
self._game_is_over = True
can_accept = True
accpet_action = action
break
env.restore()
put_card = []
primary_item = 0
if can_accept:
put_card = info['put_card']
primary_item = info['primary_item']
self._update_curr_player(curr_player_id, put_card)
return accpet_action, put_card, primary_item
def judge_card_type(self, hand_card):
assert len(hand_card) == 18, 'the size of hand card status must be 18.'
cardTypeStruct = CardTypeStruct()
card_count = sum(hand_card)
cardTypeStruct.card_count = card_count
find_one = list(
map(lambda x: x[0],
filter(lambda x: x[1] == 1, enumerate(hand_card))))
find_two = list(
map(lambda x: x[0],
filter(lambda x: x[1] == 2, enumerate(hand_card))))
find_three = list(
map(lambda x: x[0],
filter(lambda x: x[1] == 3, enumerate(hand_card))))
find_four = list(
map(lambda x: x[0],
filter(lambda x: x[1] == 4, enumerate(hand_card))))
# 单牌
if card_count == 1:
assert len(find_one) == 1, 'hand card status error.'
cardTypeStruct.card_type = CardTypeEnum.CT_ONE
cardTypeStruct.primary_item = find_one[0]
# 分值[-7,7]
cardTypeStruct.card_score = HandCardUtils.value_map(
find_one[0], CardTypeEnum.CT_ONE, card_count)
# 对子(包含连对)
if card_count % 2 == 0 and card_count > 0 and len(find_two) >= 1:
# 列表中只含对子且对子中不含2
if len(find_two
) * 2 == card_count and CardEnum.TW.value not in find_two:
if HandCardUtils.is_find(find_two):
cardTypeStruct.card_type = CardTypeEnum.CT_DOU
cardTypeStruct.primary_item = find_two[-1]
# 分值[-7,7]
cardTypeStruct.card_score = HandCardUtils.value_map(
find_two[-1], CardTypeEnum.CT_DOU, card_count)
#三带一单(连三带一单)
if card_count % 4 == 0 and card_count > 0 and len(find_three) >= 1:
# 三带一单
if len(find_three) == 1 and len(find_one) == 1:
cardTypeStruct.card_type = CardTypeEnum.CT_THREE_ONE
cardTypeStruct.primary_item = find_three[0]
# 分值[-7,7]
cardTypeStruct.card_score = HandCardUtils.value_map(
find_three[0], CardTypeEnum.CT_THREE_ONE, card_count)
elif len(find_three) * 3 + len(
find_one
) == card_count and CardEnum.TW.value not in find_three:
if HandCardUtils.is_find(find_three):
cardTypeStruct.card_type = CardTypeEnum.CT_THREE_ONE
cardTypeStruct.primary_item = find_three[-1]
# 分值[0.5,7.5]
cardTypeStruct.card_score = HandCardUtils.value_map(
find_three[-1], CardTypeEnum.CT_THREE_ONE, card_count)
#三带一对(连三带一对)
if card_count % 5 == 0 and card_count > 0 and len(find_three) >= 1:
# 三带一对
if len(find_two) == 1 and len(find_three) == 1:
cardTypeStruct.card_type = CardTypeEnum.CT_THREE_DOU
cardTypeStruct.primary_item = find_three[0]
# 分值 [-7,7]
cardTypeStruct.card_score = HandCardUtils.value_map(
find_three[0], CardTypeEnum.CT_THREE_DOU, card_count)
elif len(find_three) * 3 + len(
find_two
) * 2 == card_count and CardEnum.TW.value not in find_three:
if HandCardUtils.is_find(find_three):
cardTypeStruct.card_type = CardTypeEnum.CT_THREE_DOU
cardTypeStruct.primary_item = find_three[-1]
# 分值[0.5,7.5]
cardTypeStruct.card_score = HandCardUtils.value_map(
find_three[-1], CardTypeEnum.CT_THREE_DOU, card_count)
# 连子
if card_count >= 5:
# 必须都是单牌
if len(find_one) == card_count:
arr = set(find_one)
exclude_arr = {
CardEnum.TW.value, CardEnum.JA.value, CardEnum.QU.value
}
# 连子中不能包含2和大小王
if len(list(arr.intersection(exclude_arr))) == 0:
if HandCardUtils.is_find(find_one):
cardTypeStruct.card_type = CardTypeEnum.CT_CONTINUE
cardTypeStruct.primary_item = find_one[-1]
# 分值[-6,8]
cardTypeStruct.card_score = HandCardUtils.value_map(
find_one[-1], CardTypeEnum.CT_CONTINUE, card_count)
# 四带两单
if card_count == 6 and len(find_four) >= 1:
if len(find_four) == 1 and len(find_one) == 2:
cardTypeStruct.card_type = CardTypeEnum.CT_FOUR_ONE
cardTypeStruct.primary_item = find_four[0]
# 分值[0,7]
cardTypeStruct.card_score = HandCardUtils.value_map(
find_four[0], CardTypeEnum.CT_FOUR_ONE, card_count)
# 四带两对
if card_count == 8 and len(find_four) >= 1:
if len(find_four) == 1 and len(find_two) == 2:
cardTypeStruct.card_type = CardTypeEnum.CT_FOUR_DOU
cardTypeStruct.primary_item = find_four[0]
# 分值[0,7]
cardTypeStruct.card_score = HandCardUtils.value_map(
find_four[0], CardTypeEnum.CT_FOUR_DOU, card_count)
# 三不带(连三不带)
if card_count % 3 == 0 and card_count > 0 and len(find_three) >= 1:
if len(find_three) == 1:
cardTypeStruct.card_type = CardTypeEnum.CT_THREE
cardTypeStruct.primary_item = find_three[0]
# 分值[-7,7]
cardTypeStruct.card_score = HandCardUtils.value_map(
find_three[0], CardTypeEnum.CT_THREE, card_count)
# 连三不带里面不能包含2
elif len(
find_three
) * 3 == card_count and CardEnum.TW.value not in find_three:
if HandCardUtils.is_find(find_three):
cardTypeStruct.card_type = CardTypeEnum.CT_THREE
cardTypeStruct.primary_item = find_three[-1]
# 分值[0.5,7.5]
cardTypeStruct.card_score = HandCardUtils.value_map(
find_three[-1], CardTypeEnum.CT_THREE, card_count)
# 炸弹(王炸)
if card_count == 2:
if len(find_one
) == 2 and find_one[0] == CardEnum.QU.value and find_one[
1] == CardEnum.JA.value:
cardTypeStruct.card_type = CardTypeEnum.CT_BOMB
cardTypeStruct.primary_item = find_one[-1]
# 分值 25
cardTypeStruct.card_score = HandCardUtils.value_map(
find_one[-1], CardTypeEnum.CT_BOMB, card_count)
# 炸弹(普通炸弹)
if card_count == 4 and len(find_four) >= 1:
if len(find_four) == 1:
cardTypeStruct.card_type = CardTypeEnum.CT_BOMB
cardTypeStruct.primary_item = find_four[0]
# 分值[7,21]
cardTypeStruct.card_score = HandCardUtils.value_map(
find_four[0], CardTypeEnum.CT_BOMB, card_count)
return cardTypeStruct
def _gen_agent(cards1, cards2, cards3):
main_agent_status = HandCardUtils.obtain_hand_card_status(cards1)
low_agent_status = HandCardUtils.obtain_hand_card_status(cards2)
up_agent_status = HandCardUtils.obtain_hand_card_status(cards3)
return main_agent_status, low_agent_status, up_agent_status
def gen_policy_net_input():
card_process = HandCardUtils.read_from_file()[0]
samples = PreprocUtils._gen_state_action_pair(card_process, target=0)
train_inputs, train_output = PreprocUtils._gen_policy_net_input_core(
samples)
return train_inputs, train_output
def _decode_put_card(put_card):
p_action, k_action, card_dict = HandCardUtils.decode_put_card(None)
if put_card != "PASS":
p_action, k_action, card_dict = HandCardUtils.decode_put_card(
put_card)
return p_action, k_action, card_dict
