# -*- coding: utf-8 -*-
"""
Created on Sat Nov 28 20:43:31 2020

@author: dzyzh
"""

import numpy as np
import random as rd
import time

class player(object):
    
    def __init__(self,number):
        self.number = number
        self.average_reward = 0.0
        self.count = 0
        self.value = np.random.rand(10,19,20)
        self.decision = np.random.rand(10,19,20,2)
        self.frequency1 = np.random.randint(1,10,[10,19,20])
        self.frequency2 = np.random.randint(1,10,[10,19,20,2])    
        self.memory = np.empty(100)
             


            
    def choose_action(self,concealed_card,open_card):

        t = np.mean(self.memory)
        if(abs((t-self.average_reward))>abs(self.average_reward)):
            explor = 100
        elif(abs((t-self.average_reward))>0.1*abs(self.average_reward)):
            explor = 10
        else:
            explor = 1
        explor = 100
        UCT_value=np.empty(20)
        current_value = self.value[concealed_card,open_card]
        current_frequency = self.frequency1[concealed_card,open_card]
        i = 0
        while(i<20):
            total_frequency = sum(sum(sum(self.frequency1)))
            UCT_value[i] = (current_value[i] + np.sqrt(explor)*np.sqrt((np.log(total_frequency)/current_frequency[i])))
            i = i+1
        choice = UCT_value.argmax()
        return choice + 1
    
    def decision_choose(self,concealed_card,open_card,chip):

        t = np.mean(self.memory)
        if(abs((t-self.average_reward))>abs(self.average_reward)):
            explor = 100
        elif(abs((t-self.average_reward))>0.1*abs(self.average_reward)):
            explor = 10
        else:
            explor = 1

        explor = 100
        UCT_value=np.empty(2)
        current_value = self.decision[concealed_card,open_card,chip-1]
        current_frequency = self.frequency2[concealed_card,open_card,chip-1]
        i = 0
        while(i<2):
            total_frequency = sum(sum(sum(sum(self.frequency2))))
            UCT_value[i] = (current_value[i] + np.sqrt(explor)*np.sqrt((np.log(total_frequency)/current_frequency[i])))
            i = i+1
        choice = UCT_value.argmax()
        return choice    
    
    
    def update_value(self,banker,concealed_card,open_card,chip,decision,reward):
        #p = int(given_value/2-1)
        #q = int(choice-1)
        if(banker==1):
            origin_frequency = self.frequency1[concealed_card,open_card,chip-1]
            origin_value = self.value[concealed_card,open_card,chip-1]
            self.frequency1[concealed_card,open_card,chip-1] = origin_frequency + 1
            self.value[concealed_card,open_card,chip-1] = (origin_value*origin_frequency + reward) / self.frequency1[concealed_card,open_card,chip-1]
        else:
            origin_frequency2 = self.frequency2[concealed_card,open_card,chip-1,decision] 
            origin_decision = self.decision[concealed_card,open_card,chip-1,decision]
            self.frequency2[concealed_card,open_card,chip-1,decision] = origin_frequency2 + 1
            self.decision[concealed_card,open_card,chip-1,decision] = (origin_decision*origin_frequency2 + reward) / self.frequency2[concealed_card,open_card,chip-1,decision]
        self.average_reward = ( self.average_reward * self.count + reward ) / (self.count + 1)
        self.count = self.count + 1
        self.memory[0:99] = self.memory[1:100]
        self.memory[99] = reward

 


def game(player1,player2,i):
    win_1 = 0
    win_2 = 0
    chip_1 = 0
    chip_2 = 0
    banker = rd.choice([0,1])
    card_library = [0,1,2,3,4,5,6,7,8,9]
    open_card = rd.choice([-1,1])*rd.choice(card_library)
    concealed_card_1 = rd.choice(card_library)
    concealed_card_2 = rd.choice(card_library) 
    decision_1 = 0
    decision_2 = 0
    if((concealed_card_1 - concealed_card_2) > -open_card):
        win_1 = 1
    elif((concealed_card_1 - concealed_card_2) < -open_card):
        win_2 = 1
    else:
        win_1 = 1
        win_2 = 1
    if(banker == 0):
        chip_1 = player1.choose_action(concealed_card_1,open_card+9) 
        decision_2 = player2.decision_choose(concealed_card_2,-open_card+9,chip_1)
        if(decision_2 ==1):
            if(win_1 ==1 and win_2==0):
                reward_1 = chip_1
                reward_2 = -chip_1
            elif(win_1 ==0 and win_2==1):
                reward_1 = -chip_1
                reward_2 = chip_1            
            else:
                reward_1 = 0
                reward_2 = 0
        else:
            reward_1 = 1
            reward_2 = -1
    else:
        chip_2 = player2.choose_action(concealed_card_2,-open_card+9) 
        decision_1 = player1.decision_choose(concealed_card_1,open_card+9,chip_2)
        if(decision_1 ==1):
            if(win_1 ==1 and win_2==0):
                reward_1 = chip_2
                reward_2 = -chip_2
            elif(win_1 ==0 and win_2==1):
                reward_1 = -chip_2
                reward_2 = chip_2            
            else:
                reward_1 = 0
                reward_2 = 0
        else:
            reward_1 = -1
            reward_2 = 1       
            
    if(banker == 0):
        banker_1 = 1
        banker_2 = 0
    else:
        banker_2 = 1
        banker_1 = 0
    
    print("现在是第{0}局游戏，当前正在游戏的玩家是：player{1},player{2}".format(i,player1.number, player2.number))
    print("当前的庄家是：{0},{1}; player{2}的底牌是：{4}，player{3}的底牌是：{5}，player{2}的公共牌是：{6}，player{3}的公共牌是：{7}".format(banker_1,banker_2,player1.number,player2.number,concealed_card_1,concealed_card_2,open_card,-open_card))
    print("player{0}下注{2}; player{1}下注{3}; player{0}跟注为{4}，player{1}跟注为{5}，player{0}收益为{6},player{1}收益为{7}".format(player1.number, player2.number,chip_1+1,chip_2+1,decision_1,decision_2,reward_1,reward_2))
    return(banker_1,banker_2,concealed_card_1,concealed_card_2,open_card+9,-open_card+9,chip_1,chip_2,decision_1,decision_2,reward_1,reward_2)







d = {}
for i in range(10):
    d['player'+str(i)] = player(i)
    
k=0
while(k<5000000):
    current_waiting = [0,1,2,3,4,5,6,7,8,9]
    while(current_waiting != []):       
        a=rd.choice(current_waiting)
        current_waiting.remove(a)
        b=rd.choice(current_waiting)
        current_waiting.remove(b)
        
        result = game( d['player'+str(a)] , d['player'+str(b)] , k )
        d['player'+str(a)].update_value(result[0],result[2],result[4],result[6],result[8],result[10])
        d['player'+str(b)].update_value(result[1],result[3],result[5],result[7],result[9],result[11])
    k = k+1

doc = open("value2.txt",'w')
for i in range(0,10):
    for j in range(0,19):
            for k in range(0,20):
                print([i,j,k],file=doc)
                print(d['player'+str(5)].decision[i,j,k],file = doc)    
doc.close()