debug branch

core/robot.py

@@ -5,12 +5,7 @@
 
 from core import rule
 from core.player import Player
-from core.predictor import Predictor
-from tensorpack import *
-from core.DQNModel import Model
 from handlers.protocol import Protocol as Pt
-import numpy as np
-from core.extra.card import Card
 
 logger = logging.getLogger('ddz')
 
@@ -21,12 +16,6 @@ def __init__(self, uid: int, username: str, player: Player):
         from handlers.loopback import LoopBackSocketHandler
         super().__init__(uid, username, LoopBackSocketHandler(self))
         self.room = player.room
-        self.predictor = Predictor(OfflinePredictor(PredictConfig(
-            model=Model(),
-            session_init=SaverRestore('C:/Users/44762/PycharmProjects/doudizhu-tornado/core/res/model-92500'),
-            input_names=['state', 'comb_mask', 'fine_mask'],
-            output_names=['Qvalue']
-        )))
 
     def to_server(self, message):
         packet = json.dumps(message)
@@ -74,58 +63,11 @@ def auto_call_score(self, score=0):
 
     def auto_shot_poker(self):
         pokers = []
+        if not self.table.last_shot_poker or self.table.last_shot_seat == self.seat:
+            pokers.append(self.hand_pokers[0])
+        else:
+            pokers = rule.cards_above(self.hand_pokers, self.table.last_shot_poker)
 
-        def to_char(pokers):
-            cards = rule._to_cards(pokers)
-            for i, card in enumerate(cards):
-                if card == 'w':
-                    cards[i] = '*'
-                elif card == 'W':
-                    cards[i] = '$'
-                elif card == '0':
-                    cards[i] = '10'
-            return cards
-        handcards_char = to_char(self.hand_pokers)
-        last_cards_char = to_char(self.table.last_shot_poker)
-        print(handcards_char)
-        print(last_cards_char)
-        if self.table.last_shot_seat == self.seat:
-            last_cards_char = []
-
-        total_cards = np.ones([60])
-        total_cards[53:56] = 0
-        total_cards[57:60] = 0
-        remain_cards = total_cards - Card.char2onehot60(handcards_char +
-                                                        to_char(self.table.history[self.seat] +
-                                                                self.table.history[(self.seat + 1) % 3] +
-                                                                self.table.history[(self.seat + 2) % 3]))
-        next_cnt = len(self.table.players[(self.seat + 1) % 3].hand_pokers)
-        next_next_cnt = len(self.table.players[(self.seat + 2) % 3].hand_pokers)
-        next_state = remain_cards * (next_cnt / (next_cnt + next_next_cnt))
-        next_next_state = remain_cards * (next_next_cnt / (next_cnt + next_next_cnt))
-        prob_state = np.concatenate([next_state, next_next_state])
-        assert np.all(prob_state < 1.) and np.all(prob_state >= 0.)
-        # print(self.table.last_shot_poker)
-        # print(self.hand_pokers)
-        # print(self.table.players[self.seat].hand_pokers)
-        intention = self.predictor.predict(handcards_char, last_cards_char, prob_state)
-        print(intention)
-
-        def to_pokers(cards):
-            for i, card in enumerate(cards):
-                if card == '*':
-                    cards[i] = 'w'
-                elif card == '$':
-                    cards[i] = 'W'
-                elif card == '10':
-                    cards[i] = '0'
-            return rule._to_pokers(self.hand_pokers, cards)
-        # if not self.table.last_shot_poker or self.table.last_shot_seat == self.seat:
-        #     pokers.append(self.hand_pokers[0])
-        # else:
-        #     pokers = rule.cards_above(self.hand_pokers, self.table.last_shot_poker)
-        pokers = to_pokers(intention)
         packet = [Pt.REQ_SHOT_POKER, pokers]
         # IOLoop.current().add_callback(self.to_server, packet)
         IOLoop.current().call_later(2, self.to_server, packet)
-