1

src/environment/chengdu_majiang_env.py

@@ -1,7 +1,7 @@
 import gym
 import numpy as np
 from gym import spaces
-
+from src.engine.actions import draw_tile, discard_tile, peng, gang, check_blood_battle
 from src.engine.calculate_fan import calculate_fan, is_seven_pairs, is_cleared, is_big_pairs
 from src.engine.chengdu_mahjong_engine import ChengduMahjongEngine
 from src.engine.scoring import calculate_score
@@ -25,19 +25,40 @@ class MahjongEnv(gym.Env):
         return self.engine.state.hands[self.engine.state.current_player]
 
     def step(self, action):
+        """
+        每一步游戏，基于玩家的动作更新游戏状态。
+        action：表示玩家的动作，可以是摸牌、打牌、碰牌等。
+        """
+        done = False
+        reward = 0
+
+        # 根据action类型选择执行的动作
         try:
-            # 执行玩家动作
-            self.engine.discard_tile(action)
+            if action == 0:  # 0代表摸牌
+                reward, done = draw_tile(self.engine)
+            elif action == 1:  # 1代表打牌
+                tile = self.engine.state.hands[self.engine.state.current_player][0]  # 假设选择第一张牌
+                discard_tile(self.engine, tile)
+                reward, done = -1, False
+            elif action == 2:  # 2代表碰牌
+                tile = self.engine.state.hands[self.engine.state.current_player][0]  # 假设选择第一张牌
+                peng(self.engine, tile)
+                reward, done = 0, False
+            elif action == 3:  # 3代表杠牌
+                tile = self.engine.state.hands[self.engine.state.current_player][0]  # 假设选择第一张牌
+                gang(self.engine, tile, mode="ming")  # 暂时假设为明杠
+                reward, done = 0, False
+
             # 检查是否胡牌
             if self.engine.state.can_win(self.engine.state.hands[self.engine.state.current_player]):
                 reward, done = self.handle_win()
-            else:
-                reward, done = -1, False  # 默认小惩罚
 
-            # 检查是否有玩家分数 <= 0
-            if any(score <= 0 for score in self.scores):
+            # 检查游戏结束条件
+            check_blood_battle(self.engine)
+
+            if self.engine.game_over:
                 done = True
-                reward = -100  # 游戏结束的惩罚（可根据需求调整）
+
         except ValueError:
             reward, done = -10, False  # 非法操作扣分
 
@@ -62,7 +83,7 @@ class MahjongEnv(gym.Env):
         is_self_draw = True  # 假设自摸（后续可动态判断）
 
         conditions = {
-            "is_cleared": is_cleared(hand),
+            "is_cleared": is_cleared(hand, melds),
             "is_seven_pairs": is_seven_pairs(hand),
             "is_big_pairs": is_big_pairs(hand),
             # 添加其他条件...
@@ -79,10 +100,11 @@ class MahjongEnv(gym.Env):
 
         # 奖励设置为赢家得分
         reward = scores["winner"]
-        done = True  # 胡牌结束当前局
-        return reward, done
+        self.engine.state.winners.append(winner)  # 添加赢家到列表
+
+        # 如果有玩家分数 <= 0，可进行其他处理，如记录惩罚或结束游戏
+        if any(score <= 0 for score in self.scores):
+            self.engine.game_over = True  # 设置游戏结束标志
+
+        return reward, True  # 胡牌结束当前局
 
-    def render(self, mode="human"):
-        print(f"当前轮数: {self.current_round}")
-        print("玩家分数:", self.scores)
-        print("当前玩家状态:", self.engine.state.hands[self.engine.state.current_player])