train_final_step.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From DQN.py:109: arg_max (from tensorflow.python.ops.gen_math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use `argmax` instead\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "DQN.py:168: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison\n",
      "  if action_index not in availble_actions:\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('loss', 100, 209.24817)\n",
      "('loss', 200, 202.13641)\n",
      "('loss', 300, 213.10597)\n",
      "('loss', 400, 204.6724)\n",
      "('loss', 500, 200.87831)\n",
      "('loss', 600, 168.094)\n",
      "('loss', 700, 182.11964)\n",
      "('loss', 800, 171.72006)\n",
      "('loss', 900, 189.35129)\n",
      "('loss', 1000, 190.7134)\n",
      "('loss', 1100, 209.22356)\n",
      "('loss', 1200, 175.96536)\n",
      "('loss', 1300, 184.6994)\n",
      "('loss', 1400, 192.12706)\n",
      "('loss', 1500, 198.02753)\n",
      "('loss', 1600, 172.36725)\n",
      "('loss', 1700, 184.72969)\n",
      "('loss', 1800, 159.8089)\n",
      "('loss', 1900, 171.00862)\n",
      "('loss', 2000, 210.03072)\n",
      "('loss', 2100, 211.84155)\n",
      "('loss', 2200, 199.64958)\n",
      "('loss', 2300, 198.50264)\n",
      "('loss', 2400, 155.4646)\n",
      "('loss', 2500, 177.7889)\n",
      "('loss', 2600, 179.58261)\n",
      "('loss', 2700, 181.51672)\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0mTraceback (most recent call last)",
      "\u001b[0;32m<ipython-input-1-371dd7f67a77>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     48\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mepisode\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m100000\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     49\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mepisode\u001b[0m \u001b[0;34m%\u001b[0m \u001b[0;36m10000\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m0\u001b[0m \u001b[0;34m:\u001b[0m\u001b[0;32mprint\u001b[0m \u001b[0;34m\"episode\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mepisode\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 50\u001b[0;31m             \u001b[0mRLModel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mRLModel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexerience_replay_final_step\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     51\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     52\u001b[0m         \u001b[0;31m# 将下一个 state_ 变为 下次循环的 state\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/data/opt/jupyter/fanyu.zhang/GoldenFlower/DQN.pyc\u001b[0m in \u001b[0;36mexerience_replay_final_step\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    233\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    234\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mexerience_replay_final_step\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 235\u001b[0;31m         \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmemory\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    236\u001b[0m         \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    237\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrandom\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msample\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbatch_size\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "#-*- coding:utf-8\n",
    "import sys\n",
    "import os\n",
    "os.environ['CUDA_VISIBLE_DEVICES'] = \"\"\n",
    "stdi, stdo, stde = sys.stdin, sys.stdout, sys.stderr\n",
    "reload(sys)\n",
    "sys.setdefaultencoding('utf-8')\n",
    "sys.stdin, sys.stdout, sys.stderr = stdi, stdo, stde\n",
    "\n",
    "import GameEnv\n",
    "import random\n",
    "import numpy as np\n",
    "import DQN\n",
    "\n",
    "gameEnv = GameEnv.GlodenFlower([20000,20000])\n",
    "RLModel = DQN.DQN(embedding_size=50,sequence_length=20,learning_rate=0.01,batch_size=1000)\n",
    "gameEnv.debug = False\n",
    "memory = []\n",
    "for episode in range(500000):\n",
    "    # 初始化环境\n",
    "    gameEnv.reset()\n",
    "\n",
    "    playerI = gameEnv.getStartTurn()\n",
    "    observation_this = [[],gameEnv.playerCards[\"A\"],gameEnv.personStatus[\"A\"]]\n",
    "    if playerI == \"B\":\n",
    "        availble_actions = gameEnv.chooseAvailbleAction(playerI)\n",
    "        action = RLModel.choose_action(np.array([observation_this]),availble_actions)\n",
    "        #print (playerI, action, gameEnv.deskMoney, gameEnv.nowPrice)\n",
    "        observation_next, reward, done = gameEnv.step(action, \"B\")\n",
    "        playerI = \"A\"\n",
    "        if done:\n",
    "            continue\n",
    "\n",
    "    while True:\n",
    "        # DQN 根据观测值选择行为\n",
    "        availble_actions = gameEnv.chooseAvailbleAction(playerI)\n",
    "        action = RLModel.choose_action(np.array([observation_this]),availble_actions)\n",
    "        # 环境根据行为给出下一个 state, reward, 是否终止\n",
    "        #print (playerI, action, gameEnv.deskMoney, gameEnv.nowPrice,gameEnv.personStatus[playerI])\n",
    "        observation_next, reward, done = gameEnv.stepA(action,RLModel)\n",
    "\n",
    "\n",
    "        # DQN 存储记忆\n",
    "        # RL.store_transition(observation, action, reward)\n",
    "        RLModel.store_transition(observation_this, action, reward,done,observation_next,gameEnv.playerCards[\"B\"])\n",
    "\n",
    "        # 控制学习起始时间和频率 (先累积一些记忆再开始学习)\n",
    "        if (episode > 100000):\n",
    "            if episode % 10000 == 0 :print \"episode\",episode\n",
    "            RLModel.train(RLModel.exerience_replay_final_step())\n",
    "\n",
    "        # 将下一个 state_ 变为 下次循环的 state\n",
    "        observation_this = observation_next\n",
    "\n",
    "        # 如果终止, 就跳出循环\n",
    "        if done:\n",
    "            break\n",
    "\n",
    "        # end of game\n",
    "print('game over')\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('loss', 200, 218.0011)\n",
      "('loss', 300, 219.64523)\n",
      "('loss', 400, 233.80363)\n",
      "('loss', 500, 217.5003)\n",
      "('loss', 600, 201.68799)\n",
      "('loss', 700, 204.67659)\n",
      "('loss', 800, 218.53188)\n",
      "('loss', 900, 216.44981)\n",
      "('loss', 1000, 202.11006)\n",
      "('loss', 1100, 202.05333)\n",
      "('loss', 1200, 239.83394)\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0mTraceback (most recent call last)",
      "\u001b[0;32m<ipython-input-2-0bd9009090b5>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mwhile\u001b[0m \u001b[0mTrue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m     \u001b[0mRLModel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrain\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mRLModel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexerience_replay_final_step\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/data/opt/jupyter/fanyu.zhang/GoldenFlower/DQN.pyc\u001b[0m in \u001b[0;36mexerience_replay_final_step\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    226\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mexerience_replay_final_step\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    227\u001b[0m         \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmemory\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 228\u001b[0;31m         \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    229\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrandom\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msample\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbatch_size\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "while True:\n",
    "    RLModel.train(RLModel.exerience_replay_final_step())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.7560779\n",
      "0 闷_2 32.60102\n",
      "1 闷_4 -0.5648137\n",
      "2 闷_8 16.131987\n",
      "3 闷开_0 0.89452285\n",
      "4 看_2 35.17862\n",
      "5 看_5 -1.4472319\n",
      "6 看_10 -2.6187477\n",
      "7 看_20 -2.5867748\n",
      "8 开_0 20.885717\n",
      "9 丢_0 -0.42460057\n"
     ]
    }
   ],
   "source": [
    "for i in range(100):\n",
    "    _, global_step,loss = RLModel.sess.run([RLModel.train_op, RLModel.global_steps, RLModel.loss], feed_dict=feed_dict)\n",
    "print loss\n",
    "res = RLModel.sess.run(RLModel.predictions,feed_dict=RLModel._feed_dict(np.array([[[], ['heart_J', 'club_Q', 'diamond_Q'],\"看\"]])))\n",
    "for i,v in  zip(RLModel.action_reverse_index_dicts.keys(),res[0]):\n",
    "    print i, RLModel.action_reverse_index_dicts[i],v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 闷_2 0.0\n",
      "1 闷_4 0.0\n",
      "2 闷_8 0.0\n",
      "3 闷开_0 0.0\n",
      "4 看_2 6.737506\n",
      "5 看_5 11.0771885\n",
      "6 看_10 15.052254\n",
      "7 看_20 24.87912\n",
      "8 开_0 11.454239\n",
      "9 丢_0 -10.331483\n"
     ]
    }
   ],
   "source": [
    "# RLModel.train(RLModel.exerience_replay_final_step())\n",
    "res = RLModel.sess.run(RLModel.predictions,feed_dict=RLModel._feed_dict(np.array([[[],  ['club_10', 'heart_10', 'spade_10'],\"看\"]])))\n",
    "for i,v in  zip(RLModel.action_reverse_index_dicts.keys(),res[0]):\n",
    "    print i, RLModel.action_reverse_index_dicts[i],v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[array([16.51234], dtype=float32), 551.67017, None]\n",
      "[array([35.27028], dtype=float32), 22.370262, None]\n",
      "[array([39.938026], dtype=float32), 0.0038407235, None]\n",
      "[array([52.92187], dtype=float32), 166.97476, None]\n",
      "[array([-51.1008], dtype=float32), 8299.355, None]\n",
      "[array([60.855316], dtype=float32), 434.9442, None]\n",
      "[array([133.87975], dtype=float32), 8813.406, None]\n",
      "[array([73.1089], dtype=float32), 1096.1993, None]\n",
      "[array([25.075468], dtype=float32), 222.74165, None]\n",
      "[array([-6.5888815], dtype=float32), 2170.524, None]\n"
     ]
    }
   ],
   "source": [
    "# RLModel.train(RLModel.exerience_replay_final_step())\n",
    "res = RLModel.sess.run(RLModel.predictions,feed_dict=RLModel._feed_dict(np.array([[[],  ['club_10', 'heart_10', 'spade_10'],\"看\"]])))\n",
    "for i,v in  zip(RLModel.c.keys(),res[0]):\n",
    "    print i, RLModel.action_reverse_index_dicts[i],v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "None\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "_dict = RLModel._feed_dict(np.array([[[], ['heart_7', 'heart_4', 'heart_10'],\"看\"]]))\n",
    "_dict[RLModel.actionInput] = RLModel._one_hot([8])\n",
    "_dict[RLModel.yInput] = np.array([30])\n",
    "res = RLModel.sess.run(RLModel.train_op,feed_dict=_dict)\n",
    "print res\n",
    "\n",
    "_dict = RLModel._feed_dict(np.array([[[], ['heart_7', 'heart_4', 'heart_10'],\"看\"]]))\n",
    "_dict[RLModel.actionInput] = RLModel._one_hot([9])\n",
    "_dict[RLModel.yInput] = np.array([-30])\n",
    "res = RLModel.sess.run(RLModel.train_op,feed_dict=_dict)\n",
    "print res"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_data = RLModel.exerience_replay_final_step()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_observation_this = train_data[:,0]\n",
    "train_action =  train_data[:,1]\n",
    "train_reward = train_data[:,2]\n",
    "train_done = train_data[:,3]\n",
    "train_observation_next = train_data[:,4]\n",
    "statusMap = {\"闷\":1,\"看\":0,\"开\":0}\n",
    "\n",
    "playSequenceStr = [i[0] for i in train_observation_this]\n",
    "playCardStr =  [i[1] for i in train_observation_this]\n",
    "personStatus = [i[2] for i in train_observation_this]\n",
    "playSequenceIndex = [[RLModel.seq_action_index_dicts[i] for i in j] + [len(RLModel.seq_action_index_dicts)] * (len(RLModel.seq_action_index_dicts) - len(j)) for j in playSequenceStr]\n",
    "playSequenceLength = [len(i)+1 for i in playSequenceStr]\n",
    "personIndex = [statusMap[i] for i in personStatus]\n",
    "playCardIndex = [sorted([RLModel.card_index_dicts[i] for i in j]) for j in playCardStr]\n",
    "actionIndex = [RLModel.actions_index_dicts[i] for i in  train_action]\n",
    "\n",
    "next_status = np.array([[i[0],i[1],i[2]] for i in train_observation_next])\n",
    "maxQNext = RLModel.get_max_action(next_status)\n",
    "y = []\n",
    "for i in range(RLModel.batch_size):\n",
    "    if train_done[i] == True:\n",
    "        y.append(train_reward[i])\n",
    "    else:\n",
    "        print 1\n",
    "        y.append(train_reward[i] + RLModel.sigema * maxQNext[i])\n",
    "feed_dict = {RLModel.playSequenceInput:np.array(playSequenceIndex),RLModel.playCardsInput:np.array(playCardIndex),\n",
    "             RLModel.actionInput:RLModel._one_hot(actionIndex),RLModel.playSequenceLengthInput:np.array(playSequenceLength),\n",
    "             RLModel.yInput:np.array(y),RLModel.personStatusInput:np.array(personIndex)}\n",
    "_, global_step,loss = RLModel.sess.run([RLModel.train_op, RLModel.global_steps, RLModel.loss], feed_dict=feed_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4 (['diamond_2', 'diamond_3', 'heart_Q'], 3)\n",
      "7 (['diamond_6', 'diamond_J', 'diamond_5'], 61)\n",
      "10 (['diamond_5', 'spade_1', 'club_3'], 9)\n",
      "11 (['spade_10', 'diamond_J', 'heart_8'], 1)\n",
      "18 (['diamond_1', 'club_10', 'diamond_3'], 3)\n",
      "19 (['club_K', 'diamond_K', 'club_4'], 13)\n",
      "20 (['club_9', 'diamond_K', 'spade_2'], 21)\n",
      "21 (['diamond_2', 'spade_1', 'heart_Q'], 9)\n",
      "27 (['diamond_7', 'diamond_9', 'diamond_2'], 3)\n",
      "30 (['heart_K', 'heart_7', 'heart_8'], 9)\n",
      "31 (['club_J', 'club_10', 'heart_6'], 21)\n",
      "33 (['club_4', 'club_1', 'diamond_4'], 11)\n",
      "35 (['heart_9', 'heart_10', 'spade_10'], 5)\n",
      "44 (['heart_K', 'club_2', 'diamond_2'], 1)\n",
      "47 (['heart_8', 'spade_4', 'club_4'], 1)\n",
      "50 (['spade_1', 'club_5', 'heart_K'], 9)\n",
      "52 (['heart_2', 'spade_1', 'heart_5'], 1)\n",
      "58 (['diamond_J', 'club_10', 'club_K'], 49)\n",
      "61 (['spade_4', 'spade_5', 'spade_6'], 1)\n",
      "65 (['heart_6', 'spade_5', 'club_K'], 21)\n",
      "67 (['spade_9', 'club_8', 'club_K'], 7)\n",
      "72 (['heart_6', 'spade_Q', 'club_6'], 11)\n",
      "74 (['club_Q', 'club_6', 'spade_Q'], 6)\n",
      "78 (['spade_7', 'diamond_J', 'spade_K'], 1)\n",
      "87 (['club_8', 'spade_4', 'heart_8'], 13)\n",
      "98 (['spade_10', 'heart_7', 'club_3'], 45)\n",
      "99 (['club_4', 'club_1', 'diamond_5'], 9)\n",
      "101 (['club_4', 'club_K', 'club_9'], 3)\n",
      "104 (['diamond_K', 'spade_J', 'diamond_6'], 13)\n",
      "106 (['heart_1', 'spade_3', 'club_1'], 11)\n",
      "108 (['heart_1', 'spade_4', 'diamond_3'], 5)\n",
      "112 (['club_J', 'spade_J', 'spade_4'], 1)\n",
      "116 (['club_7', 'diamond_10', 'club_K'], 3)\n",
      "120 (['spade_9', 'heart_Q', 'heart_10'], 1)\n",
      "125 (['spade_Q', 'spade_1', 'club_7'], 1)\n",
      "129 (['diamond_8', 'diamond_K', 'spade_4'], 1)\n",
      "132 (['heart_8', 'heart_Q', 'spade_9'], 6)\n",
      "135 (['club_10', 'spade_1', 'diamond_5'], 9)\n",
      "146 (['heart_J', 'heart_2', 'spade_1'], 21)\n",
      "148 (['heart_1', 'heart_7', 'diamond_6'], 16)\n",
      "151 (['club_3', 'club_5', 'spade_4'], 6)\n",
      "156 (['diamond_1', 'spade_1', 'club_4'], 1)\n",
      "161 (['heart_1', 'club_1', 'heart_6'], 21)\n",
      "164 (['heart_Q', 'club_6', 'club_9'], 1)\n",
      "165 (['club_K', 'diamond_2', 'club_8'], 23)\n",
      "166 (['spade_K', 'spade_8', 'spade_10'], 1)\n",
      "167 (['diamond_8', 'spade_10', 'spade_J'], 41)\n",
      "176 (['spade_1', 'heart_5', 'club_5'], 11)\n",
      "179 (['diamond_9', 'spade_K', 'diamond_7'], 7)\n",
      "185 (['heart_J', 'heart_1', 'spade_10'], 45)\n",
      "188 (['heart_9', 'club_Q', 'club_9'], 1)\n",
      "190 (['club_J', 'spade_J', 'diamond_5'], 9)\n",
      "197 (['heart_7', 'diamond_10', 'diamond_K'], 9)\n",
      "200 (['club_9', 'spade_9', 'club_4'], 5)\n",
      "203 (['heart_Q', 'spade_J', 'diamond_3'], 5)\n",
      "208 (['diamond_1', 'spade_2', 'spade_8'], 21)\n",
      "212 (['heart_Q', 'club_J', 'club_1'], 3)\n",
      "219 (['diamond_10', 'spade_4', 'heart_9'], 6)\n",
      "236 (['club_K', 'heart_7', 'spade_5'], 1)\n",
      "241 (['spade_4', 'heart_4', 'spade_1'], 9)\n",
      "246 (['club_Q', 'heart_Q', 'spade_5'], 9)\n",
      "249 (['club_K', 'diamond_K', 'club_10'], 17)\n",
      "250 (['heart_7', 'club_K', 'diamond_Q'], 9)\n",
      "257 (['spade_6', 'heart_Q', 'spade_4'], 1)\n",
      "261 (['heart_J', 'spade_3', 'diamond_6'], 11)\n",
      "265 (['diamond_1', 'club_6', 'diamond_9'], 11)\n",
      "291 (['spade_J', 'diamond_Q', 'heart_4'], 9)\n",
      "294 (['spade_2', 'heart_K', 'spade_9'], 9)\n",
      "297 (['diamond_7', 'diamond_J', 'diamond_8'], 9)\n",
      "298 (['diamond_7', 'club_7', 'club_3'], 1)\n",
      "307 (['spade_Q', 'spade_7', 'heart_Q'], 1)\n",
      "312 (['club_4', 'diamond_Q', 'diamond_3'], 3)\n",
      "313 (['heart_6', 'heart_8', 'spade_7'], 9)\n",
      "316 (['heart_J', 'spade_J', 'spade_9'], 1)\n",
      "318 (['spade_9', 'club_8', 'club_7'], 29)\n",
      "319 (['diamond_1', 'heart_K', 'club_6'], 21)\n",
      "325 (['heart_8', 'diamond_10', 'club_K'], 5)\n",
      "337 (['heart_5', 'diamond_8', 'diamond_5'], 9)\n",
      "339 (['spade_8', 'club_6', 'heart_1'], 1)\n",
      "342 (['club_Q', 'heart_4', 'diamond_3'], 6)\n",
      "355 (['club_K', 'heart_10', 'heart_4'], 1)\n",
      "356 (['club_1', 'club_8', 'spade_6'], 17)\n",
      "358 (['diamond_7', 'diamond_Q', 'spade_K'], 3)\n",
      "360 (['heart_7', 'club_5', 'club_6'], 9)\n",
      "362 (['heart_2', 'heart_Q', 'club_6'], 9)\n",
      "367 (['spade_9', 'heart_K', 'heart_3'], 17)\n",
      "371 (['spade_9', 'diamond_9', 'spade_3'], 1)\n",
      "374 (['diamond_7', 'spade_9', 'heart_8'], 41)\n",
      "375 (['diamond_6', 'spade_6', 'spade_10'], 5)\n",
      "377 (['heart_7', 'spade_J', 'spade_9'], 13)\n",
      "382 (['heart_9', 'spade_8', 'spade_7'], 1)\n",
      "386 (['heart_7', 'spade_8', 'club_1'], 7)\n",
      "387 (['heart_Q', 'heart_9', 'heart_2'], 5)\n",
      "388 (['diamond_3', 'spade_3', 'diamond_8'], 21)\n",
      "394 (['club_3', 'spade_8', 'club_1'], 61)\n",
      "404 (['diamond_10', 'club_5', 'diamond_9'], 5)\n",
      "409 (['spade_Q', 'diamond_Q', 'club_7'], 1)\n",
      "412 (['heart_7', 'heart_1', 'spade_8'], 9)\n",
      "416 (['diamond_6', 'spade_10', 'club_10'], 1)\n",
      "421 (['diamond_8', 'spade_8', 'spade_4'], 9)\n",
      "428 (['diamond_10', 'spade_K', 'spade_1'], 5)\n",
      "430 (['club_7', 'heart_3', 'diamond_1'], 1)\n",
      "435 (['diamond_Q', 'club_Q', 'club_9'], 21)\n",
      "439 (['spade_6', 'heart_8', 'spade_1'], 1)\n",
      "446 (['club_2', 'heart_Q', 'spade_6'], 5)\n",
      "455 (['spade_9', 'club_9', 'club_10'], 9)\n",
      "459 (['diamond_7', 'heart_5', 'diamond_1'], 25)\n",
      "460 (['heart_6', 'spade_K', 'heart_7'], 11)\n",
      "463 (['heart_7', 'club_3', 'club_7'], 9)\n",
      "465 (['club_10', 'diamond_4', 'heart_10'], 21)\n",
      "473 (['spade_2', 'spade_6', 'spade_4'], 6)\n",
      "474 (['diamond_10', 'club_9', 'heart_J'], 9)\n",
      "477 (['club_5', 'heart_9', 'diamond_5'], 11)\n",
      "480 (['diamond_5', 'spade_K', 'club_6'], 25)\n",
      "486 (['heart_9', 'diamond_1', 'diamond_8'], 5)\n",
      "487 (['diamond_10', 'club_4', 'heart_10'], 23)\n",
      "488 (['heart_9', 'diamond_4', 'heart_4'], 43)\n",
      "494 (['diamond_2', 'spade_3', 'club_K'], 1)\n",
      "497 (['club_1', 'diamond_4', 'diamond_6'], 21)\n",
      "498 (['spade_5', 'heart_7', 'club_K'], 1)\n",
      "501 (['heart_10', 'spade_9', 'club_J'], 5)\n",
      "504 (['spade_6', 'spade_7', 'diamond_1'], 6)\n",
      "509 (['club_3', 'club_1', 'heart_8'], 26)\n",
      "510 (['club_J', 'spade_3', 'club_1'], 21)\n",
      "512 (['diamond_1', 'diamond_2', 'club_J'], 41)\n",
      "513 (['diamond_4', 'diamond_J', 'diamond_7'], 3)\n",
      "516 (['spade_Q', 'spade_5', 'club_5'], 31)\n",
      "517 (['spade_9', 'spade_10', 'diamond_3'], 21)\n",
      "530 (['diamond_9', 'diamond_2', 'diamond_10'], 1)\n",
      "533 (['spade_Q', 'club_10', 'heart_J'], 6)\n",
      "535 (['spade_2', 'club_5', 'diamond_Q'], 69)\n",
      "536 (['diamond_8', 'diamond_J', 'diamond_3'], 5)\n",
      "538 (['diamond_5', 'diamond_J', 'spade_J'], 5)\n",
      "544 (['diamond_9', 'diamond_5', 'diamond_8'], 5)\n",
      "545 (['diamond_8', 'diamond_1', 'diamond_J'], 25)\n",
      "546 (['spade_J', 'heart_1', 'diamond_J'], 1)\n",
      "554 (['heart_9', 'club_9', 'diamond_5'], 1)\n",
      "557 (['spade_1', 'heart_8', 'club_Q'], 1)\n",
      "559 (['heart_3', 'club_K', 'diamond_10'], 23)\n",
      "560 (['spade_10', 'spade_5', 'spade_1'], 9)\n",
      "561 (['spade_4', 'club_K', 'diamond_K'], 6)\n",
      "567 (['club_Q', 'spade_10', 'heart_4'], 1)\n",
      "573 (['spade_Q', 'club_6', 'club_5'], 1)\n",
      "580 (['club_4', 'diamond_2', 'spade_3'], 26)\n",
      "582 (['heart_8', 'diamond_3', 'club_8'], 21)\n",
      "584 (['spade_8', 'club_J', 'heart_J'], 11)\n",
      "585 (['spade_3', 'spade_K', 'club_J'], 25)\n",
      "589 (['club_3', 'heart_Q', 'spade_1'], 3)\n",
      "591 (['diamond_6', 'spade_J', 'club_2'], 5)\n",
      "597 (['spade_Q', 'club_5', 'heart_5'], 1)\n",
      "598 (['diamond_J', 'diamond_8', 'diamond_3'], 21)\n",
      "601 (['diamond_4', 'spade_8', 'club_7'], 21)\n",
      "608 (['club_5', 'spade_6', 'heart_1'], 5)\n",
      "611 (['heart_3', 'spade_J', 'diamond_J'], 51)\n",
      "612 (['heart_2', 'club_Q', 'diamond_J'], 37)\n",
      "623 (['club_6', 'spade_5', 'spade_7'], 65)\n",
      "626 (['heart_10', 'club_1', 'spade_4'], 1)\n",
      "632 (['club_1', 'club_4', 'spade_6'], 1)\n",
      "639 (['spade_J', 'spade_Q', 'spade_K'], 1)\n",
      "642 (['club_3', 'club_K', 'club_1'], 11)\n",
      "644 (['spade_1', 'heart_4', 'spade_Q'], 6)\n",
      "645 (['heart_2', 'heart_3', 'heart_8'], 1)\n",
      "650 (['club_3', 'club_1', 'spade_10'], 31)\n",
      "654 (['heart_Q', 'heart_J', 'diamond_3'], 21)\n",
      "655 (['heart_3', 'diamond_10', 'diamond_2'], 11)\n",
      "657 (['club_1', 'spade_4', 'club_8'], 9)\n",
      "663 (['club_4', 'club_J', 'heart_1'], 1)\n",
      "671 (['club_6', 'club_J', 'heart_J'], 1)\n",
      "672 (['spade_7', 'spade_6', 'diamond_5'], 1)\n",
      "673 (['club_1', 'club_10', 'spade_4'], 29)\n",
      "674 (['diamond_K', 'spade_2', 'diamond_9'], 3)\n",
      "675 (['diamond_10', 'spade_4', 'heart_7'], 41)\n",
      "677 (['diamond_Q', 'spade_1', 'spade_7'], 21)\n",
      "688 (['spade_4', 'spade_3', 'diamond_J'], 3)\n",
      "691 (['club_6', 'club_2', 'diamond_K'], 33)\n",
      "692 (['diamond_2', 'diamond_4', 'heart_9'], 21)\n",
      "696 (['spade_7', 'heart_5', 'spade_8'], 1)\n",
      "697 (['heart_5', 'heart_1', 'club_4'], 11)\n",
      "699 (['heart_1', 'club_K', 'club_8'], 21)\n",
      "700 (['diamond_4', 'diamond_1', 'club_5'], 1)\n",
      "702 (['diamond_4', 'heart_10', 'heart_K'], 1)\n",
      "703 (['heart_9', 'diamond_9', 'heart_K'], 1)\n",
      "706 (['club_2', 'diamond_8', 'heart_8'], 21)\n",
      "708 (['diamond_J', 'diamond_8', 'club_8'], 11)\n",
      "709 (['club_8', 'diamond_4', 'spade_J'], 1)\n",
      "711 (['diamond_8', 'heart_9', 'spade_K'], 29)\n",
      "713 (['club_4', 'heart_4', 'heart_9'], 45)\n",
      "719 (['club_Q', 'spade_1', 'diamond_6'], 1)\n",
      "720 (['diamond_6', 'diamond_8', 'diamond_10'], 11)\n",
      "723 (['club_9', 'diamond_3', 'diamond_1'], 9)\n",
      "726 (['heart_4', 'diamond_7', 'spade_Q'], 3)\n",
      "730 (['spade_7', 'spade_10', 'spade_2'], 21)\n",
      "734 (['diamond_Q', 'club_2', 'club_K'], 1)\n",
      "737 (['spade_Q', 'club_J', 'heart_6'], 3)\n",
      "738 (['club_4', 'diamond_K', 'spade_Q'], 21)\n",
      "739 (['diamond_J', 'spade_K', 'heart_5'], 3)\n",
      "741 (['spade_K', 'diamond_K', 'spade_3'], 9)\n",
      "742 (['spade_8', 'spade_6', 'diamond_6'], 9)\n",
      "743 (['spade_4', 'spade_2', 'spade_8'], 1)\n",
      "744 (['heart_6', 'spade_4', 'heart_10'], 11)\n",
      "746 (['spade_10', 'diamond_2', 'spade_5'], 1)\n",
      "748 (['diamond_3', 'diamond_Q', 'heart_2'], 1)\n",
      "751 (['heart_9', 'spade_J', 'diamond_2'], 19)\n",
      "757 (['heart_3', 'diamond_6', 'heart_K'], 9)\n",
      "758 (['heart_J', 'diamond_8', 'spade_8'], 31)\n",
      "759 (['spade_4', 'diamond_Q', 'club_1'], 3)\n",
      "770 (['diamond_10', 'club_9', 'heart_2'], 9)\n",
      "772 (['club_2', 'club_7', 'diamond_Q'], 1)\n",
      "779 (['spade_1', 'diamond_1', 'spade_J'], 1)\n",
      "780 (['diamond_2', 'heart_Q', 'diamond_5'], 11)\n",
      "783 (['spade_K', 'spade_9', 'spade_4'], 5)\n",
      "788 (['heart_7', 'diamond_Q', 'heart_9'], 11)\n",
      "792 (['club_9', 'diamond_7', 'heart_7'], 29)\n",
      "794 (['heart_4', 'diamond_K', 'heart_K'], 1)\n",
      "798 (['club_5', 'heart_K', 'spade_10'], 1)\n",
      "801 (['diamond_1', 'heart_K', 'heart_4'], 21)\n",
      "807 (['spade_10', 'diamond_J', 'heart_1'], 21)\n",
      "810 (['diamond_K', 'spade_10', 'club_1'], 3)\n",
      "811 (['diamond_9', 'heart_K', 'diamond_8'], 5)\n",
      "814 (['diamond_1', 'spade_7', 'club_J'], 21)\n",
      "817 (['club_Q', 'diamond_7', 'diamond_2'], 9)\n",
      "821 (['diamond_6', 'heart_7', 'club_7'], 5)\n",
      "824 (['club_K', 'heart_Q', 'spade_6'], 16)\n",
      "834 (['diamond_J', 'heart_8', 'club_3'], 3)\n",
      "836 (['diamond_8', 'spade_1', 'spade_2'], 1)\n",
      "837 (['club_2', 'heart_2', 'spade_1'], 26)\n",
      "839 (['diamond_4', 'club_5', 'heart_1'], 1)\n",
      "847 (['spade_9', 'heart_7', 'club_K'], 9)\n",
      "850 (['spade_Q', 'club_5', 'club_9'], 1)\n",
      "851 (['heart_1', 'spade_2', 'spade_1'], 21)\n",
      "858 (['club_8', 'diamond_Q', 'heart_Q'], 5)\n",
      "861 (['club_J', 'spade_7', 'diamond_K'], 11)\n",
      "864 (['diamond_10', 'heart_3', 'heart_10'], 1)\n",
      "872 (['heart_K', 'club_J', 'diamond_8'], 23)\n",
      "876 (['club_K', 'heart_5', 'diamond_1'], 1)\n",
      "883 (['heart_10', 'diamond_6', 'spade_6'], 41)\n",
      "885 (['spade_4', 'spade_8', 'spade_10'], 17)\n",
      "886 (['heart_10', 'diamond_1', 'club_5'], 33)\n",
      "888 (['diamond_K', 'heart_5', 'heart_9'], 9)\n",
      "892 (['diamond_10', 'diamond_7', 'club_1'], 1)\n",
      "902 (['diamond_8', 'heart_J', 'club_Q'], 6)\n",
      "917 (['diamond_10', 'diamond_2', 'club_1'], 6)\n",
      "927 (['spade_4', 'diamond_5', 'spade_3'], 1)\n",
      "928 (['diamond_10', 'club_8', 'diamond_6'], 11)\n",
      "930 (['club_J', 'spade_1', 'heart_9'], 1)\n",
      "935 (['diamond_8', 'club_8', 'club_2'], 9)\n",
      "937 (['club_6', 'diamond_1', 'diamond_10'], 9)\n",
      "938 (['club_7', 'spade_K', 'diamond_7'], 29)\n",
      "941 (['spade_K', 'spade_1', 'heart_Q'], 5)\n",
      "943 (['spade_4', 'club_Q', 'spade_6'], 3)\n",
      "946 (['spade_Q', 'club_2', 'club_7'], 5)\n",
      "950 (['heart_7', 'heart_10', 'spade_5'], 9)\n",
      "954 (['diamond_3', 'diamond_8', 'spade_3'], 3)\n",
      "958 (['club_9', 'club_3', 'spade_1'], 11)\n",
      "960 (['heart_8', 'club_10', 'spade_1'], 1)\n",
      "968 (['heart_5', 'spade_Q', 'spade_K'], 6)\n",
      "977 (['club_1', 'diamond_4', 'club_3'], 1)\n",
      "978 (['spade_5', 'club_6', 'club_K'], 16)\n",
      "981 (['diamond_Q', 'spade_K', 'diamond_4'], 1)\n",
      "984 (['diamond_K', 'diamond_4', 'club_K'], 26)\n",
      "989 (['heart_Q', 'heart_J', 'spade_5'], 23)\n",
      "995 (['club_K', 'heart_6', 'heart_3'], 1)\n"
     ]
    }
   ],
   "source": [
    "for i,j in enumerate(zip(playCardStr,y)):\n",
    "    if j[1] > 0:\n",
    "        print i,j"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "l = RLModel.card_index_dicts.keys()\n",
    "input_data = []\n",
    "for i in range(10000):\n",
    "    input_data.append([[], random.sample(l,3),\"看\"])\n",
    "res = RLModel.sess.run(RLModel.predictions,feed_dict=RLModel._feed_dict(np.array(input_data)))\n",
    "# for i,v in  zip(RLModel.action_reverse_index_dicts.keys(),res[0]):\n",
    "#     print i, RLModel.action_reverse_index_dicts[i],v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "l = RLModel.card_index_dicts.keys()\n",
    "l.remove(\" \")\n",
    "card_list = []\n",
    "input_data = []\n",
    "for i in range(200000):\n",
    "    cards = random.sample(l,3)\n",
    "    card_list.append(cards)\n",
    "    input_data.append([[],cards ,\"看\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "res = RLModel.sess.run(RLModel.predictions,feed_dict=RLModel._feed_dict(np.array(input_data)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "df = pd.DataFrame(data={\"card\":card_list,\"score\":res[:,8]})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "import GameEnv\n",
    "gameEnv = GameEnv.GlodenFlower([2000,2000])\n",
    "d = {\"豹子\":10,\"同花顺\":9,\"金花\":8,\"顺子\":7,\"对子\":6,\"单\":5}\n",
    "reverse_dict = {v: k for k, v in d.iteritems()}\n",
    "\n",
    "def f1(x):\n",
    "\n",
    "    return reverse_dict[gameEnv.score(x)]\n",
    "    \n",
    "\n",
    "\n",
    "df['card_level'] = df['card'].apply(lambda x:f1(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>card</th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>card_level</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>单</th>\n",
       "      <td>149333</td>\n",
       "      <td>149333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>同花顺</th>\n",
       "      <td>410</td>\n",
       "      <td>410</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>对子</th>\n",
       "      <td>33988</td>\n",
       "      <td>33988</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>豹子</th>\n",
       "      <td>446</td>\n",
       "      <td>446</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>金花</th>\n",
       "      <td>9764</td>\n",
       "      <td>9764</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>顺子</th>\n",
       "      <td>6059</td>\n",
       "      <td>6059</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              card   score\n",
       "card_level                \n",
       "单           149333  149333\n",
       "同花顺            410     410\n",
       "对子           33988   33988\n",
       "豹子             446     446\n",
       "金花            9764    9764\n",
       "顺子            6059    6059"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby(\"card_level\").count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>score</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>card_level</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>单</th>\n",
       "      <td>-12.270060</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>同花顺</th>\n",
       "      <td>14.447891</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>对子</th>\n",
       "      <td>5.290378</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>豹子</th>\n",
       "      <td>14.180468</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>金花</th>\n",
       "      <td>10.442841</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>顺子</th>\n",
       "      <td>11.867952</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                score\n",
       "card_level           \n",
       "单          -12.270060\n",
       "同花顺         14.447891\n",
       "对子           5.290378\n",
       "豹子          14.180468\n",
       "金花          10.442841\n",
       "顺子          11.867952"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby(\"card_level\").mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>card</th>\n",
       "      <th>score</th>\n",
       "      <th>card_level</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>123378</th>\n",
       "      <td>[spade_1, diamond_1, club_1]</td>\n",
       "      <td>21.382235</td>\n",
       "      <td>豹子</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>157308</th>\n",
       "      <td>[club_1, spade_1, diamond_1]</td>\n",
       "      <td>21.382235</td>\n",
       "      <td>豹子</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63375</th>\n",
       "      <td>[spade_1, diamond_1, club_1]</td>\n",
       "      <td>21.382235</td>\n",
       "      <td>豹子</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>168887</th>\n",
       "      <td>[spade_1, diamond_1, club_1]</td>\n",
       "      <td>21.382235</td>\n",
       "      <td>豹子</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>111806</th>\n",
       "      <td>[club_1, spade_1, diamond_1]</td>\n",
       "      <td>21.382235</td>\n",
       "      <td>豹子</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52438</th>\n",
       "      <td>[heart_Q, heart_7, heart_1]</td>\n",
       "      <td>21.167383</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90385</th>\n",
       "      <td>[heart_7, heart_Q, heart_1]</td>\n",
       "      <td>21.167383</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59546</th>\n",
       "      <td>[heart_1, heart_7, heart_Q]</td>\n",
       "      <td>21.167383</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17468</th>\n",
       "      <td>[heart_1, heart_Q, heart_7]</td>\n",
       "      <td>21.167383</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>180910</th>\n",
       "      <td>[heart_K, diamond_Q, heart_1]</td>\n",
       "      <td>21.159143</td>\n",
       "      <td>顺子</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45727</th>\n",
       "      <td>[heart_1, heart_K, diamond_Q]</td>\n",
       "      <td>21.159143</td>\n",
       "      <td>顺子</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>155988</th>\n",
       "      <td>[diamond_Q, heart_K, heart_1]</td>\n",
       "      <td>21.159143</td>\n",
       "      <td>顺子</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>138189</th>\n",
       "      <td>[heart_1, heart_K, diamond_Q]</td>\n",
       "      <td>21.159143</td>\n",
       "      <td>顺子</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>174498</th>\n",
       "      <td>[heart_K, diamond_Q, heart_1]</td>\n",
       "      <td>21.159143</td>\n",
       "      <td>顺子</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90492</th>\n",
       "      <td>[diamond_Q, heart_K, heart_1]</td>\n",
       "      <td>21.159143</td>\n",
       "      <td>顺子</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>192915</th>\n",
       "      <td>[heart_1, heart_K, diamond_Q]</td>\n",
       "      <td>21.159143</td>\n",
       "      <td>顺子</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>94489</th>\n",
       "      <td>[heart_1, diamond_Q, heart_K]</td>\n",
       "      <td>21.159143</td>\n",
       "      <td>顺子</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146612</th>\n",
       "      <td>[heart_7, heart_1, heart_8]</td>\n",
       "      <td>20.898483</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14343</th>\n",
       "      <td>[heart_7, heart_8, heart_1]</td>\n",
       "      <td>20.898483</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42857</th>\n",
       "      <td>[heart_1, heart_7, heart_8]</td>\n",
       "      <td>20.898483</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13806</th>\n",
       "      <td>[heart_7, heart_8, heart_1]</td>\n",
       "      <td>20.898483</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126747</th>\n",
       "      <td>[heart_1, heart_8, heart_7]</td>\n",
       "      <td>20.898483</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42023</th>\n",
       "      <td>[heart_8, heart_7, heart_1]</td>\n",
       "      <td>20.898483</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>191231</th>\n",
       "      <td>[heart_1, heart_8, heart_7]</td>\n",
       "      <td>20.898483</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>66534</th>\n",
       "      <td>[heart_7, heart_1, heart_8]</td>\n",
       "      <td>20.898483</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>177997</th>\n",
       "      <td>[heart_8, heart_1, heart_7]</td>\n",
       "      <td>20.898483</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63946</th>\n",
       "      <td>[heart_8, heart_1, heart_7]</td>\n",
       "      <td>20.898483</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199665</th>\n",
       "      <td>[heart_1, heart_7, heart_8]</td>\n",
       "      <td>20.898483</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>133558</th>\n",
       "      <td>[heart_8, heart_7, heart_1]</td>\n",
       "      <td>20.898483</td>\n",
       "      <td>金花</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34073</th>\n",
       "      <td>[heart_1, diamond_K, heart_Q]</td>\n",
       "      <td>20.879963</td>\n",
       "      <td>顺子</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>161220</th>\n",
       "      <td>[heart_3, diamond_2, spade_10]</td>\n",
       "      <td>-45.713802</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82226</th>\n",
       "      <td>[diamond_2, spade_10, heart_3]</td>\n",
       "      <td>-45.713802</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>128785</th>\n",
       "      <td>[diamond_2, spade_10, heart_3]</td>\n",
       "      <td>-45.713802</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>181954</th>\n",
       "      <td>[spade_10, diamond_2, heart_3]</td>\n",
       "      <td>-45.713802</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>102329</th>\n",
       "      <td>[spade_10, heart_3, diamond_2]</td>\n",
       "      <td>-45.713802</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42578</th>\n",
       "      <td>[diamond_2, spade_10, heart_3]</td>\n",
       "      <td>-45.713802</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97830</th>\n",
       "      <td>[diamond_2, heart_3, spade_10]</td>\n",
       "      <td>-45.713802</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>116784</th>\n",
       "      <td>[diamond_2, heart_Q, spade_9]</td>\n",
       "      <td>-47.310329</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23465</th>\n",
       "      <td>[heart_Q, diamond_2, spade_9]</td>\n",
       "      <td>-47.310329</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>94463</th>\n",
       "      <td>[heart_Q, diamond_2, spade_9]</td>\n",
       "      <td>-47.310329</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39092</th>\n",
       "      <td>[spade_9, diamond_2, heart_Q]</td>\n",
       "      <td>-47.310329</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49040</th>\n",
       "      <td>[spade_9, diamond_2, heart_Q]</td>\n",
       "      <td>-47.310329</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12159</th>\n",
       "      <td>[spade_9, diamond_2, heart_Q]</td>\n",
       "      <td>-47.310329</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5562</th>\n",
       "      <td>[diamond_2, heart_Q, spade_9]</td>\n",
       "      <td>-47.310329</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34108</th>\n",
       "      <td>[diamond_2, heart_Q, spade_9]</td>\n",
       "      <td>-47.310329</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97503</th>\n",
       "      <td>[club_7, spade_9, diamond_2]</td>\n",
       "      <td>-52.121597</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132369</th>\n",
       "      <td>[spade_9, club_7, diamond_2]</td>\n",
       "      <td>-52.121597</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57945</th>\n",
       "      <td>[club_7, diamond_2, spade_9]</td>\n",
       "      <td>-52.121597</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114115</th>\n",
       "      <td>[spade_9, diamond_2, club_7]</td>\n",
       "      <td>-52.121597</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>176469</th>\n",
       "      <td>[spade_9, club_7, diamond_2]</td>\n",
       "      <td>-52.121597</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>139200</th>\n",
       "      <td>[spade_9, club_7, diamond_2]</td>\n",
       "      <td>-52.121597</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>109435</th>\n",
       "      <td>[diamond_2, club_7, spade_10]</td>\n",
       "      <td>-52.421940</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>128529</th>\n",
       "      <td>[diamond_2, spade_10, club_7]</td>\n",
       "      <td>-52.421940</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>133913</th>\n",
       "      <td>[club_7, spade_10, diamond_2]</td>\n",
       "      <td>-52.421940</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>65005</th>\n",
       "      <td>[spade_10, diamond_2, club_7]</td>\n",
       "      <td>-52.421940</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24155</th>\n",
       "      <td>[spade_10, diamond_2, club_7]</td>\n",
       "      <td>-52.421940</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121851</th>\n",
       "      <td>[spade_10, club_7, diamond_2]</td>\n",
       "      <td>-52.421940</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>199828</th>\n",
       "      <td>[spade_10, diamond_2, club_7]</td>\n",
       "      <td>-52.421940</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34963</th>\n",
       "      <td>[diamond_2, spade_10, club_7]</td>\n",
       "      <td>-52.421940</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>118331</th>\n",
       "      <td>[spade_10, club_7, diamond_2]</td>\n",
       "      <td>-52.421940</td>\n",
       "      <td>单</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>199900 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                  card      score card_level\n",
       "123378    [spade_1, diamond_1, club_1]  21.382235         豹子\n",
       "157308    [club_1, spade_1, diamond_1]  21.382235         豹子\n",
       "63375     [spade_1, diamond_1, club_1]  21.382235         豹子\n",
       "168887    [spade_1, diamond_1, club_1]  21.382235         豹子\n",
       "111806    [club_1, spade_1, diamond_1]  21.382235         豹子\n",
       "52438      [heart_Q, heart_7, heart_1]  21.167383         金花\n",
       "90385      [heart_7, heart_Q, heart_1]  21.167383         金花\n",
       "59546      [heart_1, heart_7, heart_Q]  21.167383         金花\n",
       "17468      [heart_1, heart_Q, heart_7]  21.167383         金花\n",
       "180910   [heart_K, diamond_Q, heart_1]  21.159143         顺子\n",
       "45727    [heart_1, heart_K, diamond_Q]  21.159143         顺子\n",
       "155988   [diamond_Q, heart_K, heart_1]  21.159143         顺子\n",
       "138189   [heart_1, heart_K, diamond_Q]  21.159143         顺子\n",
       "174498   [heart_K, diamond_Q, heart_1]  21.159143         顺子\n",
       "90492    [diamond_Q, heart_K, heart_1]  21.159143         顺子\n",
       "192915   [heart_1, heart_K, diamond_Q]  21.159143         顺子\n",
       "94489    [heart_1, diamond_Q, heart_K]  21.159143         顺子\n",
       "146612     [heart_7, heart_1, heart_8]  20.898483         金花\n",
       "14343      [heart_7, heart_8, heart_1]  20.898483         金花\n",
       "42857      [heart_1, heart_7, heart_8]  20.898483         金花\n",
       "13806      [heart_7, heart_8, heart_1]  20.898483         金花\n",
       "126747     [heart_1, heart_8, heart_7]  20.898483         金花\n",
       "42023      [heart_8, heart_7, heart_1]  20.898483         金花\n",
       "191231     [heart_1, heart_8, heart_7]  20.898483         金花\n",
       "66534      [heart_7, heart_1, heart_8]  20.898483         金花\n",
       "177997     [heart_8, heart_1, heart_7]  20.898483         金花\n",
       "63946      [heart_8, heart_1, heart_7]  20.898483         金花\n",
       "199665     [heart_1, heart_7, heart_8]  20.898483         金花\n",
       "133558     [heart_8, heart_7, heart_1]  20.898483         金花\n",
       "34073    [heart_1, diamond_K, heart_Q]  20.879963         顺子\n",
       "...                                ...        ...        ...\n",
       "161220  [heart_3, diamond_2, spade_10] -45.713802          单\n",
       "82226   [diamond_2, spade_10, heart_3] -45.713802          单\n",
       "128785  [diamond_2, spade_10, heart_3] -45.713802          单\n",
       "181954  [spade_10, diamond_2, heart_3] -45.713802          单\n",
       "102329  [spade_10, heart_3, diamond_2] -45.713802          单\n",
       "42578   [diamond_2, spade_10, heart_3] -45.713802          单\n",
       "97830   [diamond_2, heart_3, spade_10] -45.713802          单\n",
       "116784   [diamond_2, heart_Q, spade_9] -47.310329          单\n",
       "23465    [heart_Q, diamond_2, spade_9] -47.310329          单\n",
       "94463    [heart_Q, diamond_2, spade_9] -47.310329          单\n",
       "39092    [spade_9, diamond_2, heart_Q] -47.310329          单\n",
       "49040    [spade_9, diamond_2, heart_Q] -47.310329          单\n",
       "12159    [spade_9, diamond_2, heart_Q] -47.310329          单\n",
       "5562     [diamond_2, heart_Q, spade_9] -47.310329          单\n",
       "34108    [diamond_2, heart_Q, spade_9] -47.310329          单\n",
       "97503     [club_7, spade_9, diamond_2] -52.121597          单\n",
       "132369    [spade_9, club_7, diamond_2] -52.121597          单\n",
       "57945     [club_7, diamond_2, spade_9] -52.121597          单\n",
       "114115    [spade_9, diamond_2, club_7] -52.121597          单\n",
       "176469    [spade_9, club_7, diamond_2] -52.121597          单\n",
       "139200    [spade_9, club_7, diamond_2] -52.121597          单\n",
       "109435   [diamond_2, club_7, spade_10] -52.421940          单\n",
       "128529   [diamond_2, spade_10, club_7] -52.421940          单\n",
       "133913   [club_7, spade_10, diamond_2] -52.421940          单\n",
       "65005    [spade_10, diamond_2, club_7] -52.421940          单\n",
       "24155    [spade_10, diamond_2, club_7] -52.421940          单\n",
       "121851   [spade_10, club_7, diamond_2] -52.421940          单\n",
       "199828   [spade_10, diamond_2, club_7] -52.421940          单\n",
       "34963    [diamond_2, spade_10, club_7] -52.421940          单\n",
       "118331   [spade_10, club_7, diamond_2] -52.421940          单\n",
       "\n",
       "[199900 rows x 3 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.sort_values(\"score\",ascending=False)[100:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gameEnv.compare(['spade_K', 'spade_1', 'spade_6'],['heart_2', 'diamond_1', 'diamond_5'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}