Example_13.1_Short corridor with switched actions(REINFORCE).ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "#######################################################################\n",
    "# Copyright (C)                                                       #\n",
    "# 2018 Sergii Bondariev (sergeybondarev@gmail.com)                    #\n",
    "# 2018 Shangtong Zhang(zhangshangtong.cpp@gmail.com)                  #\n",
    "# Permission given to modify the code as long as you keep this        #\n",
    "# declaration at the top                                              #\n",
    "#######################################################################"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "這次的範例主要參考[ShangtongZhang](https://github.com/ShangtongZhang/reinforcement-learning-an-introduction/blob/master/chapter08/maze.py)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Chapter 13進入policy gradient，我們不再需要表格式條列的去記錄每一個state action pair所得到的action value，而是利用學習參數的方式來計算value，同時給出action。這種同時計算action與value方法稱為actor-critic，其中actor為policy，critic為value function，不過要注意的是，通常為state-value function。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "下面給出範例說明"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Example 13.1 Short corridor with switched actions\n",
    "\n",
    "考慮下圖小插畫中的小走廊網格世界(small corridor gridworld)。一如往常的，每一個step的rewrad都是$-1$。在這三個nonterminal states都只有兩個actions可以選擇，右跟左。在第一、第三個states，這兩個actions都會有很常見的結果，就是左、右(第一個state的左是停在原地)，但是在第二個state是反過來的，選擇左會往右，選擇右會往左。這個問題是很難的，因為所有的states在function approximation下看起來都是一樣的。特別是，我們定義對於所有的$s$，其$\\mathbf{x}(s, \\text{right})=[1, 0]^T$、$\\mathbf{x}(s, \\text{left})=[0, 1]^T$。使用$\\epsilon$-greedy的action-value methods會迫使我們在兩個策略之間做選擇：\n",
    "1. 有很高的機率，高到有$1 - \\epsilon/2$的機率在所有steps中選擇右\n",
    "2. 相同高的機率在所有的time steps選擇左\n",
    "\n",
    "如果$\\epsilon=0.1$，那這兩個policies在start state的value就會小於$-44$與$-82$，如下圖所示：\n",
    "![](https://hackmd.io/_uploads/ByRpCogvF.png)\n",
    "\n",
    "如果有一個方法可以學習用一個特定的機率來選擇右，那這方法會明顯得到較好的結果。這最佳機率大約是$0.59$，得到的value大概是$-11.6$。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "首先載入需求套件"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "from tqdm import tqdm"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "然後針對範例說明來設置一個環境"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "class ShortCorridor:\n",
    "    \"\"\"這個範例很簡單，就4個states\n",
    "    \n",
    "    從最左邊的Start state開始，一路到最右邊的Goal state\n",
    "    \"\"\"\n",
    "    def __init__(self):\n",
    "        \"\"\"初始化的時候先重置狀態\"\"\"\n",
    "        self.reset()\n",
    "        \n",
    "    def reset(self):\n",
    "        self.state = 0\n",
    "    \n",
    "    def step(self, action):\n",
    "        \"\"\"執行action之後的結果\n",
    "        \n",
    "        唯一要注意的就是在第二個state的時候，執行的action跟結果會是相反的\n",
    "        其它兩個state就是很正常，\n",
    "        過程中的reward都是-1，只有在最後進到目標狀態的時候才會是0\n",
    "        \n",
    "        action: 0: left, 1: right\n",
    "        \n",
    "        return reward, next_state, is_finish_flag\n",
    "        \"\"\"\n",
    "        if self.state == 0 or self.state == 2:\n",
    "            if action == 0:\n",
    "                self.state = max(0, self.state - 1)\n",
    "            else:\n",
    "                self.state += 1\n",
    "        else:\n",
    "            if action == 0:\n",
    "                self.state += 1\n",
    "            else:\n",
    "                self.state -= 1\n",
    "        \n",
    "        if self.state == 3:            \n",
    "            reward = 0\n",
    "            is_finish_flag = True\n",
    "        else:\n",
    "            reward = -1\n",
    "            is_finish_flag = False\n",
    "            \n",
    "        return reward, self.state, is_finish_flag                "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "下面我們做個測試"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(-1, 1, False)\n",
      "(-1, 2, False)\n",
      "(0, 3, True)\n"
     ]
    }
   ],
   "source": [
    "sc = ShortCorridor()\n",
    "print(sc.step(1))\n",
    "print(sc.step(0))\n",
    "print(sc.step(1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "然後我們會需要一個實作REINFORCE演算法的類別，下面給出Monte-Carlo Policy Gradient的pseudocode"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](https://hackmd.io/_uploads/BJ_GRJIPY.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "softmax在機器學習中對於多類別輸出的問題非常實用，最重要的是負值也能轉正，你總是可以輸出一個正值的機率就是"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def softmax(x):\n",
    "    \"\"\"softmax function\"\"\"\n",
    "    prob = np.exp(x - np.max(x))\n",
    "    return prob / np.sum(prob)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.05021127, 0.94978873])"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "softmax(np.array([-1.47, 1.47]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "class REINFORCE:\n",
    "    \"\"\"Monte Carlo Policy Gradient Control\"\"\"\n",
    "    def __init__(self, alpha, gamma):\n",
    "        \"\"\"\n",
    "        alpha: step size\n",
    "        gamma: discount factor\n",
    "        \"\"\"\n",
    "        # 參數初始值\n",
    "        self.theta = np.array([-1.47, 1.47])\n",
    "        self.alpha = alpha\n",
    "        self.gamma = gamma\n",
    "        self.epsilon = 0.05\n",
    "        \n",
    "        # action，left: (0, 1), right: (1, 0)\n",
    "        self.actions = [0, 1]\n",
    "        self.x = np.array([[0, 1], [1, 0]])\n",
    "        \n",
    "        # 跟蹤記錄每個time step的reward與action\n",
    "        self.track_rewards = []\n",
    "        self.track_actions = []\n",
    "    \n",
    "    def _softmax(self, x):\n",
    "        return softmax(x)\n",
    "    \n",
    "    def policy(self):\n",
    "        \"\"\"回傳當前的policy\n",
    "        \n",
    "        我們採用的是linear function，然後會經過softmax做轉換處理\n",
    "        \"\"\"\n",
    "        # (2, )x(2, 2)=(2, )\n",
    "        # 把參數跟action做內積\n",
    "        h = np.dot(self.theta, self.x)\n",
    "        # 內積之後的結果經過softmax來output probility distribution\n",
    "        prob = self._softmax(h)\n",
    "        \n",
    "        # 為了避免policy的機率過傾，做個小判斷來調整\n",
    "        min_idx = np.argmin(prob)\n",
    "        \n",
    "        if prob[min_idx] < self.epsilon:\n",
    "            prob[:] = 1 - self.epsilon\n",
    "            prob[min_idx] = self.epsilon\n",
    "        \n",
    "        return prob\n",
    "    \n",
    "    def action_select(self):\n",
    "        \"\"\"選擇action\n",
    "        \n",
    "        有了policy，我們就可以做action的選擇\n",
    "        這印象中是可以有兩種做法：\n",
    "         1. 一種是確定性的作法，也就是取機率最大的那一個\n",
    "         2. 利用softmax回傳的機率分佈做sample\n",
    "         \n",
    "         範例中會利用回傳的機率分佈做sample\n",
    "        \"\"\"\n",
    "        prob = self.policy()\n",
    "        # 測試之後發現，這樣子做在初始的時候會一直選擇機率大的那個action\n",
    "        # 造成學習上的困擾，所以調整成跟作者一樣的作法\n",
    "        # action = np.random.choice(self.actions, p=prob)\n",
    "        \n",
    "        # 如果機率比選右的機率還要小，那就選右\n",
    "        action = np.random.uniform() <= prob[1]\n",
    "        return action\n",
    "    \n",
    "    def terminal(self):\n",
    "        \"\"\"當episode結束的時候要做的事情\"\"\"\n",
    "        # 首先記錄經過幾個time steps\n",
    "        T = len(self.track_rewards)\n",
    "        return T"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "模擬一下互動結果"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "time step:  0 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  1 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  2 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  3 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  4 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  5 action:  right reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  6 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  7 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  8 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  9 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  10 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  11 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  12 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  13 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  14 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  15 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  16 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  17 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  18 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  19 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  20 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  21 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  22 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  23 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  24 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  25 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  26 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  27 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  28 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  29 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  30 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  31 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  32 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  33 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  34 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  35 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  36 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  37 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  38 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  39 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  40 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  41 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  42 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  43 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  44 action:  right reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  45 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  46 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  47 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  48 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  49 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  50 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  51 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  52 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  53 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  54 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  55 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  56 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  57 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  58 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  59 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  60 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  61 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  62 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  63 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  64 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  65 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  66 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  67 action:  right reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  68 action:  right reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  69 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  70 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  71 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  72 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  73 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  74 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  75 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  76 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  77 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  78 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  79 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  80 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  81 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  82 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  83 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  84 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  85 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  86 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  87 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  88 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  89 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  90 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  91 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  92 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  93 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  94 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  95 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  96 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  97 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  98 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  99 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  100 action:  right reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  101 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  102 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  103 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  104 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  105 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  106 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  107 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  108 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  109 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  110 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  111 action:  right reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  112 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  113 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  114 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  115 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  116 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  117 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  118 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  119 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  120 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  121 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  122 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  123 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  124 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  125 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  126 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  127 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  128 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  129 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  130 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  131 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  132 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  133 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  134 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  135 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  136 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  137 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  138 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  139 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  140 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  141 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  142 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  143 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  144 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  145 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  146 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  147 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  148 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  149 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  150 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  151 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  152 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  153 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  154 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  155 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  156 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  157 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  158 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  159 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  160 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  161 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  162 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  163 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  164 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  165 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  166 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  167 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  168 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  169 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  170 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  171 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  172 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  173 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  174 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  175 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  176 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  177 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  178 action:  left reward:  -1 next_state:  0\n",
      "is_finish_flag:  False\n",
      "time step:  179 action:  right reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  180 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  181 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  182 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  183 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  184 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  185 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  186 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  187 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  188 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  189 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  190 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  191 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "time step:  192 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  193 action:  left reward:  -1 next_state:  1\n",
      "is_finish_flag:  False\n",
      "time step:  194 action:  left reward:  -1 next_state:  2\n",
      "is_finish_flag:  False\n",
      "time step:  195 action:  right reward:  0 next_state:  3\n",
      "is_finish_flag:  True\n",
      "episode finish.\n",
      "gradient:  [-0.05021127  0.05021127] probs:  [0.94978873 0.05021127]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [ 0.95 -0.95] probs:  [0.95 0.05]\n",
      "gradient:  [-0.17647386  0.17647386] probs:  [0.82352614 0.17647386]\n",
      "gradient:  [-0.13086067  0.13086067] probs:  [0.86913933 0.13086067]\n",
      "gradient:  [-0.10385639  0.10385639] probs:  [0.89614361 0.10385639]\n",
      "gradient:  [-0.08605338  0.08605338] probs:  [0.91394662 0.08605338]\n",
      "gradient:  [-0.07344673  0.07344673] probs:  [0.92655327 0.07344673]\n",
      "gradient:  [-0.0640556  0.0640556] probs:  [0.9359444 0.0640556]\n",
      "gradient:  [-0.05679069  0.05679069] probs:  [0.94320931 0.05679069]\n",
      "gradient:  [-0.05100416  0.05100416] probs:  [0.94899584 0.05100416]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [ 0.95 -0.95] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [ 0.95 -0.95] probs:  [0.95 0.05]\n",
      "gradient:  [ 0.95 -0.95] probs:  [0.95 0.05]\n",
      "gradient:  [-0.07408938  0.07408938] probs:  [0.92591062 0.07408938]\n",
      "gradient:  [-0.06454416  0.06454416] probs:  [0.93545584 0.06454416]\n",
      "gradient:  [-0.05717453  0.05717453] probs:  [0.94282547 0.05717453]\n",
      "gradient:  [-0.05131364  0.05131364] probs:  [0.94868636 0.05131364]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [ 0.95 -0.95] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [ 0.95 -0.95] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05118589  0.05118589] probs:  [0.94881411 0.05118589]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [ 0.95 -0.95] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [-0.05  0.05] probs:  [0.95 0.05]\n",
      "gradient:  [ 0.95 -0.95] probs:  [0.95 0.05]\n"
     ]
    }
   ],
   "source": [
    "episodes = 1\n",
    "for i in range(episodes):\n",
    "    sr = ShortCorridor()\n",
    "    rl = REINFORCE(1, 1)\n",
    "    idx = 0\n",
    "    while True:        \n",
    "        # 選擇一個action\n",
    "        _action = rl.action_select()\n",
    "        # 跟環境互動一下，得到reward、next state、is finish flag\n",
    "        _reward, _next_state, _is_finish_flag = sr.step(_action)\n",
    "        print('time step: ', idx, 'action: ', 'left' if _action == 0 else 'right', 'reward: ', _reward, 'next_state: ', _next_state)        \n",
    "        print('is_finish_flag: ', _is_finish_flag)\n",
    "        # 寫入action跟reward\n",
    "        rl.track_actions.append(_action)\n",
    "        rl.track_rewards.append(_reward)\n",
    "        idx += 1\n",
    "        # 判斷是否terminal\n",
    "        if _is_finish_flag:\n",
    "            print('episode finish.')            \n",
    "            # 取得terminal的time step\n",
    "            T = len(rl.track_rewards)\n",
    "            for t in range(T):              \n",
    "                # 計算expected return\n",
    "                G = np.sum([np.power(rl.gamma, j + 1) * k for j, k in enumerate(rl.track_rewards[t:])])\n",
    "                # 判斷選擇的action是左還是右，1就是右，0就是左\n",
    "                j = 1 if rl.track_actions[t] else 0    \n",
    "                # 計算梯度，我們用的是softmax\n",
    "                # 然後我們希望我們的output跟實際的output愈接近愈好\n",
    "                # 相關推論這邊就不寫\n",
    "                pmf = rl.policy()\n",
    "                gradi_ln_pi = rl.x[:, j] - np.dot(rl.x, pmf)\n",
    "                print('gradient: ', gradi_ln_pi, 'probs: ', pmf)    \n",
    "                # 參數更新\n",
    "                rl.theta = rl.theta + rl.alpha * np.power(rl.gamma, t) * G * gradi_ln_pi\n",
    "            break\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "執行結果來看應該是沒有什麼問題"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "把上面的學習過程弄成一個function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def run_rl(episodes, rl_agent):\n",
    "    \"\"\"學習\n",
    "    episodes: episodes\n",
    "    rl_agent: REINFORCE\n",
    "    \"\"\"\n",
    "    sr = ShortCorridor()    \n",
    "    rl = rl_agent()\n",
    "    \n",
    "    # 記錄每個episode的的total reward\n",
    "    rewards = np.zeros(episodes)\n",
    "    \n",
    "    for i in range(episodes):\n",
    "        sr.reset()        \n",
    "        rl.track_rewards = []\n",
    "        rl.track_actions = []\n",
    "        idx = 0\n",
    "\n",
    "        reward_sum = 0        \n",
    "        while True:        \n",
    "            # 選擇一個action\n",
    "            _action = rl.action_select()\n",
    "            # 跟環境互動一下，得到reward、next state、is finish flag\n",
    "            _reward, _next_state, _is_finish_flag = sr.step(_action)\n",
    "            reward_sum += _reward\n",
    "#             print('time step: ', idx, 'action: ', 'left' if _action == 0 else 'right', 'reward: ', _reward, 'next_state: ', _next_state)        \n",
    "#             print('is_finish_flag: ', _is_finish_flag)\n",
    "            # 寫入action跟reward\n",
    "            rl.track_actions.append(_action)\n",
    "            rl.track_rewards.append(_reward)\n",
    "            idx += 1\n",
    "            # 判斷是否terminal\n",
    "            if _is_finish_flag:\n",
    "#                 print('episode finish.')            \n",
    "                # 取得terminal的time step\n",
    "                T = len(rl.track_rewards)\n",
    "                for t in range(T):              \n",
    "                    # 計算expected return\n",
    "                    G = np.sum([np.power(rl.gamma, j + 1) * k for j, k in enumerate(rl.track_rewards[t:])])\n",
    "                    # 判斷選擇的action是左還是右，1就是右，0就是左\n",
    "                    j = 1 if rl.track_actions[t] else 0    \n",
    "                    # 計算梯度，我們用的是softmax\n",
    "                    # 然後我們希望我們的output跟實際的output愈接近愈好\n",
    "                    # 相關推論這邊就不寫\n",
    "                    pmf = rl.policy()\n",
    "                    gradi_ln_pi = rl.x[:, j] - np.dot(rl.x, pmf)\n",
    "#                     print('gradient: ', gradi_ln_pi, 'probs: ', pmf)    \n",
    "                    # 參數更新\n",
    "                    rl.theta += rl.alpha * np.power(rl.gamma, t) * G * gradi_ln_pi                \n",
    "                break\n",
    "        \n",
    "        rewards[i] = reward_sum     \n",
    "    return rewards"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ -14.,  -95., -148.,  -45.,  -45.,  -36.,  -58.,  -86.,  -21.,\n",
       "        -19.])"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "run_rl(10, lambda: REINFORCE(2e-4, 1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "設置一個函數來做執行"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 100/100 [02:08<00:00,  1.28s/it]\n",
      "100%|██████████| 100/100 [06:46<00:00,  4.06s/it]\n",
      "100%|██████████| 100/100 [04:09<00:00,  2.49s/it]\n"
     ]
    }
   ],
   "source": [
    "num_trials = 100\n",
    "num_episodes = 1000\n",
    "\n",
    "agent_generators = [lambda : REINFORCE(alpha=2e-4, gamma=1),\n",
    "                    lambda : REINFORCE(alpha=2e-5, gamma=1),\n",
    "                    lambda : REINFORCE(alpha=2e-3, gamma=1)]\n",
    "\n",
    "labels = ['alpha = 2e-4',\n",
    "          'alpha = 2e-5',\n",
    "          'alpha = 2e-3']\n",
    "\n",
    "rewards = np.zeros((len(agent_generators), num_trials, num_episodes))\n",
    "\n",
    "for agent_index, agent in enumerate(agent_generators):\n",
    "    for i in tqdm(range(num_trials)):\n",
    "        reward = run_rl(num_episodes, agent)\n",
    "        rewards[agent_index, i, :] = reward"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA7gAAAIaCAYAAAAKmuEKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzsnXd4FNX6x7/bW7Jpm14ISSCE3nsXULBgQfT+7OXKtVz1WrHg9arXhthRsWBBQRQrIqAIoQZCgNBCSO91Uzbb+++PmTM7syUJCpbc83keHrI7s7NnZ87Onu/5vud9RV6vFxQKhUKhUCgUCoVCofzVEf/RDaBQKBQKhUKhUCgUCuVsQAUuhUKhUCgUCoVCoVD6BFTgUigUCoVCoVAoFAqlT0AFLoVCoVAoFAqFQqFQ+gRU4FIoFAqFQqFQKBQKpU9ABS6FQqFQKBQKhUKhUPoEVOBSKBQKhUKhUCgUCqVPQAUuhUKhUCgUCoVCoVD6BFTgUigUCoVCoVAoFAqlT0AFLoVCoVAoFAqFQqFQ+gTSP7oBZwOdTudNT0//o5tBoVAoFAqFQqFQKJRzwKFDh/Rerze2p/36hMBNT09HQUHBH90MCoVCoVAoFAqFQqGcA0QiUXVv9qMhyhQKhUKhUCgUCoVC6RNQgUuhUCgUCoVCoVAolD4BFbgUCoVCoVAoFAqFQukTUIFLoVAoFAqFQqFQKJQ+ARW4FAqFQqFQKBQKhULpE1CBS6FQKBQKhUKhUCiUPgEVuBQKhUKhUCgUCoVC6RNQgUuhUCgUCoVCoVAolD4BFbgUCoVCoVAoFAqFQukTUIFLoVAoFAqFQqFQKJQ+ARW4FAqFQqFQKBQKhULpE1CBS6FQKBQKhUKhUCiUPgEVuBQKhUKhUCgUCoVC6RNQgUuhUCgUCoVCoVAolD4BFbgUCoVCoVAoFAqFQukTUIFLoVAoFAqFQqFQKJQ+ARW4FAqFQqFQKBQKhULpE1CBS6FQKBQKhUKhUCiUPgEVuBQKhUKhUCgUCoVC6RNQgUuhUCgUCoVCoVD6NB6PF4tX5eHnouZf9frvCutx3/rCkNs7zA5s+5XHppxdqMClUCgUCoVC+YMpaujCqcauP7oZlD85J+oN+Olk0x/djL8knVYn8ivbkV/Z9qtef8/nhfj6SH3I7bd/dgi3flKADrPj1zaRcpagApdCoVAoFArlD2bB67sx/7Xdf3Qz/vR4vd5zenyrw41jdZ3n9D1+Cxe9sQe3rTnU434n6g1oMth+hxb9eeiyOfHwhmOobbcE3d5msgMA2s3O3/Q+Trcn6PPlrWYAwO4yPcpbTT0ex+HyYO2BGrg957ZP/y9CBS6FQqFQKBTKn5Tadgumv7gDdR3BB+1nE6vDfc7f47dw68cF6P/Ij+f0PR7YcBSXvLkX7b/Bhfs9zqMrhMgiLFlzCK/8XHLO29FbGg1WPPL1MdicPZ+bmz7Mx+2f9izi/bn380KsL6jF1hAOt97EXNN2sz3kMRwuD5asKcB/NxVxba3vtCL78c3cPkabK+hrJSIRAODudUewfMvpHtu7em8lHv3mOL46XNfjvj2x+Xgjnvz+5G8+Tl+BClwKhUKhUCh/Sr44WItWY+jB6P8CpS1G1LRbUNrSsyP0a/iusB4f7a1EfmU7cp7YgrzyXxe+6fF4kfHIJrydW35W2lXbbsF3hcJw0G2nzs36xt2lrfiioBYAsOt0KwCgvsPa4+vcHi8e3nAMRQ2+0PIT9QbkPLEFv5yjthJaTaG/F16vF60mO1qMv4+Du62oGUZb967oPesKsS6/FodrOno83o7Trdh84szCsL1eL7YXtwAAnO7gjmibmTi4oScvylpM2HqyGe/trkRhLePk17RZYHf5JhRCfVaJWMT9rTfZ4fZ4UdJsDPleZNIqlGA+E27/7DA+2lfV6wiHkmbjOY+G+COhApdCoVAoFMqfjjaTHQ99dQw3fph/xq89XmfAP9cd6dbl2ny8EQte2w3Pnzw80GRnXCTTWRgEB+Oezwvx5MYiHKxqBwDsON0Scl+DxRlyUKw32+HxAi9sKT4r7brsrX245/PCoNewNy7gmbD0q+N4aMMxrD1QAzv7ft055m0mO/aW6dFitGF9QS0WrtzDbdtXrgcA7C37dRMFALDs2xPdXgcAaO4KLXBtTg8cLg/aLb8tFLcnDtd0YO2BGtz6SQHu++IoACZ81+3xwu3xor7TN0mQz/avLivTj+0u91kVWFZen+i0OmC0OWHw+/xtrIPb1p3A5YUWG6zM6y0O4XcvlCAV81RVu9mBx789gXmv7Ao50WB1MH3NbP9t323+eSTntzv2lOox75Vd+PLQb3eO/6xQgUuhUCiUX8XbueVYf7Dmj24GpY/iYIVGcVNoByQU939ZiI1HGwSDVX8e+PIoihq7UN5qwqyXcrG7tPVXt5WPx+PtMXz0TCDC9rcOgntCLZcw7xfifdpMdox7dhtyTwvPU3WbGW9uL0Wz4ew67XrWoTQHCfcNJlD0Jjue23wKdteZi182shSbTzTC4SICN7SDu+idPFzz/gFOTPAdQ3K9wpTSM24HwPSfNfurcdOHBwXPWxwuweRBsPW1ZS0m/O3d/ahuZ9aCnutkR5e/tQ+PfnMcAHCgghH0N36Yjye/P4k1eVWY8vx2FDV0CdraarKj0+JA9uNb8MGeSgDA6j2V2HSsEUDv11ifqDegus3MPSZiFGAmYqa9uAMjnvpJ8BrfGtzuHVxCF3tM/+9EVygHV+RzcNstDqzLrwn5fh6PF/WdzCTKyh1leGFLMR795jjyK9tDti0UlXrfefB39r89Uo+jtcI15cVNTMQBP/Kgr0EFLoVCoVB+FS9sKcbDXx3/o5vxl+B0kxHpSzfhRL3hd3/vmjYLHt7Qu7VvfyZsTkZo9CYBS32nFRuPNnCPI1VyAIDeGHogG69VAgC+PlKPSr0Zz28+O87jv74oRNZjm3vekQc/ac3DG46hps3nHhJhG0p4ni2IAxZKSLcY7XC4PAHO5k0fHcRLP5WgkJeY6WwmNwrWnrYg4bmPfXMcq3ZWYG+Znnuups2ClTvKuhVNHo8XzV1Mew9U+MRFMAfXZHfh0pV7OUFBBAoALrFRFytwedGqZ0Qwd7C5y4ZVOysE4d+kzXw2Hm1AXkUbJxY7LL9fNl+zww2X24OTbDbwQzVMf/jqcJ3AwWw12lHDnquvDzMh6E/9UIQ71x4G4Dt/QPdi9971hfjvplPcY77A7bQ40RnEvSYTIxaHGztLWvHxvqqAfcpbTIhQyQRtsfhNsoRySUU8gct//2D7L1y5F/vZ/mZ3efB2bjnWHqjB4lV5Z+xsF1T7wr71vO+Gx+PFvesLsXDlXsH+5J4q5XVSgzV0dMZfESpwKRQKhXLGnE2H6vfE6/UKZrt/L8jaML4I+714c0cp1hfU4gd20PtHYXe58XZueUiH7f3dFThU3S7Yv7fc+nEB/rnuCDfIjdYwApcvQPzRhSkAANtPMdcmOVLV6/cz211IX7oJn+RVBWz7rvDMrzFfxK0vqMU/Pz/CPSbC9mys0+sOMiAnYZwAUN5q4ga9JEyThEwTqtjvU1GDb/LmZEPvJnI+3FuJH451f76e/qEoYB99EIF7sIoZ5FfqmWvucHmwvqAGy7ee7taNbTXZ4XR7kRSh5KIGgOAO7tHaTm5dJgDUtvv2IbVVibjnC64zodMaKEonPPsLXvulVPAcX+B+f7QBm4834gBb/mYb26eNNlfIjL9nG7fHi8LaTnRanGgx2rlETpuONQrEXqvRzgnNMEWgy813na0hJuW8Xi/qOiyCa8R/jzZeEin+xB6/b9+wOh//DpKUqazFhJGpkQB8Dq7/JEuoNbihJoeC9YXj3Ux2rsuvDbktGN/z7jn8z9jE6yMNvHBx0ickEkbg1rZbMO6ZbdhV6psc+qtDBS6FQqFQzpjmv2jin88P1mLWS7kCIfVbeXFLcUAImD9yKfNzy09U8nuRwDqVPYkOl9tzTrO/vr+7Ei9sKcbnQQZvLrcHz20WbiMObij2V7Rx60bJoJiU5tCqmIFzd8LGxgro02wSmMQIZch9zXaXYK0uERfv7a4AwAy4l317gmsP+Uy9xd+d5Q9GyaD5XIQo89tIziFxLo/UdOC8FTvx6f5qto3M+fJfj0hOy7E6X//q6MXaT6/Xi/9sLMJda490u9/mE024a+0RgXupNwlF4Fu5ZVwY6OGaDtS2WzD2mZ+xcgfzmvJWE9pMdixelReQYZf0kRnZcdxz0wboUNZqClif7V9+hv94IyvCybrTngSu0ebEvrJAQeF/7kKJKb54uXvdEdz+2WEcYV1Tfj3lUC6u1+vFgtd248uCwO+j1eHGoerQyaCq9Ga43B4utJvc3/awn6e5y4aSZua72Gqyc22QSURoNdrR2Mm0/XSzEde+f0Bw7HZee0M5pV1WF2xOj+AckPOdHKnCqUbfsgb+d6ktSPZkj8crcC7rO63or9MgTCHlJpXMfpM65Pm6DgvnlvOf98e/L5D7rEgEKGU+KdZfp8H0gbFY9t2JXifXa+6yYW+5HtdN7AdAOPnDL1VEJlkBXx+zs/fYvWV6ONweVPaitNFfBSpwKRQK5RzhdHsw9pltv5tr12F2BGQdJVgcLsx7ZWe3g5YzgT9oOJc0GWxnNayXZIitbjs7JVfMdhfeyi3H1e/u73Y/OTtTHsqV1JvsSF+6qUc3C2ASKBHR0Rtc7CC9JxF+59rDyHliS9AwtX1legz791bB+rQzhQz8gx2/xchkHG0w+PoV33nxF94dZgeufnc/rnwnDwAQG864saR9vVlH6b8ujh9eSPB6mcHv2Ge24fK393HPk9BFKZtVxmh3Yc3+amzhZX61nEFIuP8Amj+4NTvOXYgyPxyUDIwbOm3weLyoYGt6EuffJ7R9beW7gycbuqBghU5v3MvGbsKYg4XT89ef8l0qm9ONl38qwdzB8RiVFolNxxox7cUdgs9W0WrGrtJW5Fe2Y8maQwL3k9zLZmbHAgBSolS4bFQyqtss+PKQUPz5R3+QUNtrJ6bhSE0nVu0s5/r514frsfDNPQjFp/trcM0HB6A32fHYN8dRwE6OdPoJUnId/CE1V/kh/HaXB+kxasF+HWYm9HTxO3mY/VIuStkJneImI4oau/DghmMBx37tl1IsXpUXVFwfrzNg5ku5eP2XUni9wG3TM7Bv6WyIRcA+9v5qd3nQarRDF6ZgvtesoM2KC0eryY4mg28SYI+fyOc7uKH6ERG27WYHrA43atos3L5p0WrBd+X5zcVo6bLB6/Wius2CJL+JrIxHf8TDXx1DWYsR7+2qgMnuQmy4AlqllFtraw6SZKqsxYQFr+3GnWsPw2hz4q61h0M6zl1+n4OI+OcuG8YJ09umZ+Ce8wbgn7Oz4PZ4caLeAIPViQtf3y1YK1vTZsHClXtR0mzES1tPc9fhgqEJEIv8BC7vfk367sodZdzaZ9Iusu73XCcl+z2hApdCoVDOEV1WJzd4+T24d30h7vm8ECt+Oo0tJ4ThqMfrDChpNuH5zacCXvf+7gr8/ZOCM3ovfgmN37K20+n2dBtCN2P5Dlz0RuhB4plC2ir5tQvk/CCDiZ6OR5zbUA7uaTaR0pq84MKVCC0AuPjNPXj82xOCgS352+X24PP8GoEzRwZpPSVr2nqSCbHkh10S3theBqPdhc8O9F5Y+0MGoJogYYlEFBBnBxD2q+3FLQLnjS/wHS4P59gSgUsSE3WXCbfD7MCQJC332O5i1hDyJyHe2VmBC17dDavTjcLaTq7ECRmEe71eNHRa0cJmtOUL09NNxl6L0u72I66Qye5Cu9nRYzmW7nC4PILMsnwx1cR+BofbgwaDFbXsuSOTDkTg8h1c/4mirLgwAMy9z+Px4roPDuDWjwuCJuXhD9pJ377900NY9m3P7hV/De7JBgNcHi8Wj01FWrQ66P7lrSYcrvZN8BxgB/Rer5cTZZMzY/DwBYOw7u8TcdmoZGTGagJK1VTozciI1eDZy4YBAGrZ++DdswfggiEJeHHracHEydE6Q8jrVdFqgtfLiMLPDtRg9d5KrNxRxrmd5JZSoQ+cVJqapUNxYxdMdhc+90v0NyVLJ3jcbnZgd6ke+VXtqNCb8eG+Knx1qI5LFjY4USvY3+Px4vvCerg93gCnHAB2lzGv28mGsw5PiYAuTIHUaHVAgqQJ/aMBADVs0qusuDDojfaQkxsej1dw/spaTHjk6+MBvzGNvImwBzccxfTlOzjXup+fwP+pqBkrfipBdZsFLUY7lszIxGMLcvCvOQO5fb4oqMOid/Lw3x+Z38fYMAW0KlnIEOVXtpVgzss7uUmUdfk13S4BIetbvyiohcnu4u4fURo5HrpgEPYtnY1HF+Tg0lHJGJQQDoDp17tLW3GyoQs3fJiPF7cUo9PCTGIfre3EvFd24c0dZfjsAHP9dWEKRGsUXHmiQ9UdqNCbEa6QYlBCOKrbzPB6vXiHFwlB7snk+3Cuk5L9nlCBS6FQKOcIsp7LFkLUfLyvCu/uOjs1IwHfj/4b28vwj08PC7aRLJ9SceBtP6+8DTtPt/ZYLmXxO3l4amMRAAjKP/yWRCZTnt+OSc/9EnL72Q7pJccrbTYJEvn8WshAnGSgDQVxvewhwm6JqJJJgv8sz1iei/9774AgzJi894ZDdch89Efc+dlhrC+oxdKvj+NjnlAmYX4Wh1sgig9Vd+Annmgka1BJmZPqNjO2FTWj0WBFXkUblDIxvjlS36uyOvmV7QGJhshgMdj3gbhoDQYrJ3b4Icp3rj2MJWsO4fujDajvtAr638DHN3MlWYjAJY5vQ2fwgbTN6YbZ4cb8oQm4bFQy937XvH8Ag5Ztwaqd5TjdZMS+cr1AYBCXjawhrGqzYPLz27lQQH4ynSvfycPFvZycCRZ+TLLS8pNMjX76Z8x6aWevjgkwQpDvut+wOl+QWZZ/HhsNVm5dcmmLiVtbW9vOnO9gya7yKoSlcFKiVAhXMM7XoZoO7C7VY9upZsEaQUIRL4y23ezA/oo2bD7RhDX7q3us33qszsBd48Ja5jsxIiUCjy7IwRMXDQ7Yv7zVhMM1HZjQPxpquYS7jrklrViXX4NLRiQhXCnD7TMzkRqthkgkQlq0GnqTXSCuKvVmDIgL4/pMHevgRqhleOzCHHi9XoQrpJiUEcO9JlS0SDX7WvId/PF4E5ZvPc1FmZB7QTAHd3JWDOwuD5Z+dQyPfXNCsG0i770B5v789eE6xGjkiAtXYO2BGtz/5VG8s5P57fH/SThY1Y4GA3FIAycaSPuIC0v6TH+dJmDfGawrXtNugUYuQVKkEq1Gu6Df8TE5XILfk2c2FWFdfk1AFBTfgSfC8mR9F8Qi332MP+cYrpRiP9tXp2Tp8PfpGYjTKgTH5K/hjQ1XIJx1cF/dVsKJyFC8v7sy5DbyfTjZ0IWHNhzD14fruM8YrZFDJhEjibf+P1wpQ3qMGkWNXZwAbTXa8VZuOUY+9TNW/FwCABiaLJyYiNbIoQuTo9Voxzs7y3HF2/uw4VAd0nUa9ItRo7rNgvpOK4y872+XzSm4n7b/jknJzjVU4FIolD5PRaspoB7e7wEZgDlCiLR/f38Sz/54djK3AsHFK4EMSmXSwH2aumxwuD3QBxnMAMwAu7ipC/lV7Vi9l/kh54coB8tW2VtajPagLoE/nRbHWUlsRQarb+4o41zrDrPjV4eRE5EZzJXkQ0LcNh1vxJvbSwO2k5BL4gS73B4sXpWH3aWt6LQ4UNNuQV5Fm8AhIYOSY2z22tzTLdy1aOENAPkOEl+cLF6Vh9vWHEIFK87IGlQiWmYsz8WtnxRwbv2E/jFMdlKrEzanG2/8UopOiyNoCYxbPj7IDaAJXE1JXhsaDVa8v7uCm2ywOT3osDixs6QVGw4Frg28e90RLH4nTzBI41PFlg0hLqN/aCHBN8BU4JWrRmJAXBhsTjcOVLbD6wWe21yMDYdqcbrJKCgBQ0JS/Z0OIiL9Q6Ir9eaQmUk9Hi/Sl27COzvLgwrc61fnw+X2cJMjxMn1T7Dk8Xjxdm45mgw2wXUHgDHPbMOclxlB7PV6uWvr9nixJq8K133gqzHcaXFifP8o5vM0m1CpN6O/TgOxiIksIK44P6PsxsIGDIgL4yZ4JmbEQKuSwWB14tsj9ZCzIo2069FvjnOTemTtM8CEK/O/gyRhFGHVdWO4v4cma5Ff1Y5XtpVgw6E6FNZ2IilCiTitEvFaJW6e2h/zBsfjyjEpkIhFGJkaibIWM4qbjBjTLwqj06K445ex60SfXjg04PzrwhQ4Ud+FYU9uxeGaDng8XtS0WZAeo4FKLoFSJobR7oJcIoZCKkFqtBpL5w/CfxYOQWq0T7AcqGwPGuVCStz8xCanIpBQUolYhE6LAxV+YdHJkSpMH8AIx8NBlpyM6RcleNxmdqCqzYKcRC1SeQ43Xzzx+Y53HZq77Ljw9d346lAdvF4vVu4ow+5SPbeNOU9MQrcMXVhAW1JY4VbdZkGkWo4MnQYOt0eQ9ZePweJEu9l3vyLfC/97TDAH+GSDAREqGVRsX5w+MJbb5mIdzRiNHJmxjBDvblKSCVGWocvqCrkcZPmi4dzfLbxz+OKi4di3dDYA4LxBcdCqZPhwbxUe+ZqJ5DrVaOQ+T5RaHvTYg5O0ONnQxd1v/BnfPxo//HMarhyTwj0XqZYhJUqNqjYLF61jcbiRGKFEvxgNSltMeHHLacFxDlZ14FG2XeFKaZ9ycH9dkS7KmTNzZuBzixcDd9wBWCzAggWB22+8kfmn1wOLFgVuv/124KqrgNpa4LrrArfffz9w8cXA6dPAkiWB2x9/HJgzBygsBO69N3D7s88CkycD+/YBjz4auP3VV4GRI4Ft24BnngncvmoVkJ0NbNwIrFgRuH3NGiA1FVi/Hnj77cDtGzYAOh3w0UfMP39+/BFQq4G33gK++CJwe24u8/9LLwE//CDcplIBm9kyDk8/Dfzi5yDFxABffcX8/cgjQF6ecHtKCvDpp8zf997LnEM+AwcC777L/H3bbUBJiXD7yJHM+QOAa68F6vyKbU+aBDz3HPP3FVcAbX4F4887D1i2jPl7/nzA6jcbetFFwAMPMH/TvodrRi3BxW3FeLSGdT5+p75n++BDABnMc+Q68PsegX+NfkPfkw29FghL5B57brsNFTV6ZNna0Rk7DMi8APJTRQDGMzuwfa959B2AXIOGq29E3PD+AX3vo6QJeCl1Glcs0uPxCoRSx5I7gS5WjJxp35v4oHBf/77Hbh/51M+4vPUEXp6XzvU92/U34Yu4Ybi2+Qg3W/vNdffjX6VinLw2A5q7bg94+7bzHuL+Lm/shHPWbDyedSE2xQzCwKV3I3vZfb3qewe+/AnvbCvGVEM1kD4b6uoKYOaTIe97pv5zgfiRAICXfirBXa2HYXh3NSLczMBInzQeSJsBmZsddK7+FPmVMQIBAgDHecmxKu9/HCP1RWgeeAkQPRBmh5sbrNrzC4AXmc/fNeT/gHDGcTLbXYhQyeB+6mm4PaMBACuXrsSK8s0wj7oZUMTgWJ0BjUv/DdJPji99Bkg/D4MLdmKndjBajDY0vvgaVshHc25C1f7l3H3P5nTDaHOhZeNW4OU7ufZ2jr8TEKsZoXTttXDWN+CSUUvQKg+DzmEG5MzAs6HTihtWCz83n/pOK8IbazHU7cKJsATBNhIqbDlVCmhiYTFZgZkz4QHQcvnfkHD3EsBiQfvfbgSG34Do558GOkqhHHodKgw68IdFxRu3oyWyv+D4te1WoLYWbe99BCRP4J4v+/RrIHZI0DW/Jy66GgMtesgeXYoXXCkYaOvAt1uP4O+NB4GcK/H85mIsL/8RyPTdix+q2YkX02agaf6lMGVfCmjiUdLgC7H9ZtEdeCN5Ij45tQHN8jC8MPQabo1q6SQPZK+sQIM8HBj9D+YFM2ei/LX3uNc3zF+IZaMC783pMRroJB4Uf/wlKqOysFB/CoNlKnyww41ILSNWzJU1wMwH0SWRI3/cPbinbi9eS5kCAJgxMBbrfzyMrm3HkaeJw3mmRhwKT0bLl98B8x7CWtYJK/voC5wKT0GUVIkOmRqNj/wbzeddw7Vj7Tf7AWUk9zhixQtA3BwAwGOV2/GEKx3v7vK1+zJ7LYDzmAe33YZ32d/c5QCWp07FyuRJAJj1md7vN+K9sGy4Zs5Cfb9ZCIsdCu1TTwDPC+97sanTgOSJcLq9OPX+50h58CY43B6krF0NvHgQUaOWoFGhhcZmYsYZDzyA26ZnAjNnYk/mfCCWEc1P/1CEkh0H8MKyq7n7nlUsRfP4fwEInAAtLmsEpEpYHG6MfOpnaF02QMpMPmVFKbBt81NwbRFBPOE+zmnlE3vZhcCE+wEAMo8bDSvfR236RMwblQZDY4tgX6XbCb3BA8/evRBPmQLnnr34cU8jxlv1yNemIveVj3Aybjju//IoumxOLN96Gpe2FkEEL76JHQIA0P3tSuCt1zAzOxZf7q/A4I5aHNCmIcneBe3ddwDDb0BdhxVDkrTIOn4AgDrkpG/nF1+j9WA9FDHZsItlMFkdgEiMej07GcL+5jb3n4fwmGwYpb71tGaHG7EuKyJeXwFkLsDV37+HV7tqMXfkzeiyOlHXaUWmtQ2iWbMAAKqoLCD7sqDt0N14DbRp03E0Ih16eaAzDQBXPnAdLocImROZ37usuDCUtZgQ+d//IKmjDNuU0UjO78LlY29FvTicy5pc/Mt+ZK8vAvrPQdT/XQmMGxnwm5uVMgVbkiei9NufgajMgPe+ePt6YN0j0LF9NEIlg0wixqCfv8WO5AkYVHQQ0OUAABJqSpGcxfT/74NM5O4sYULOxyeFob4PCVzq4FIolD5Pm0wNvSz4j9RvZZ82DekTH0S7NLDEiM3bu3WeRknwWdwzReIVDhpeU2VjzshbUKGMQoeMaZ8Mwn2cIjH3A96oEIY8tYsVSJ/4ID5IHAsvLwFPXYdVkPTHEOSz90RFmwW3Dgw+uAjF17FD4fH6wnnfSp6AJ/rPwXe6wahRRKBdqsIbtUw7q7qCu8rNdp+b5hJLUK2IRAfb/gJWBAZjS9QAXDn4b/Cwbtw1h53YEZWJE5rWvo2NAAAgAElEQVR4AIBZLMfLKVNC1mw1+13j01YxRoy7G1/rmHBK0j8LG0145ecSWDzB+85Pxa1ckpQHshZgeeo0NMvCoHYLswg74Xt9l8QXikcmJiq9voHhzoj+8ACwiJjaj5V6M7bAF+Z4OCwJADDEyYjrli47WkXC8D6HyDecIO6EXuZzi7wAmtjXEAd3W1QWWuWM66OXa7i+GSp8kU+xSget245thR8gzsF8ZhmYiZe3cstQoWJcLIdYCodIjPcSx2FiQ4ov7FYRAQCIdjEOidLjwmm7cM4/X5sCf8i6VPJ9IpSqYgL2Jdw+4BJkT7gPR7u8WLWzAvcf6MDuyHRcn3Mlt49ZLOwfQ82MEKlXaGFir5/R6etbD2ZcgApVDKaOXoIrhl4jeG09ayb9FJXFPecQibGr1heqvDNCKNwJUWo5Bihc+Cp2KLqkSkzqqsHTlduQaW1HKxtlYGHva+1S5vqm2Qy4RM8sXeiv0yBC5MbRsAQ0KrSYYqhGvMOEZplG4GR/ETccFapoTGInxhqlYWjusmNyZgzk8KBGGYlhJl/4fITX930O97qQYxEKtSvsocuqJDh8nzs+Qon+LiNcYgkaFVrUK7RIdnQhSH4x6Jw+5/Qzlw63s8s+ktjjRbmYfhrmDhQF/O8cAOyzMd+3t/dW4+GM81GjYMS73BPo3PNFGwB08R6Hsa6jFF5EO5n3j3Ga8dIA5rsTIRdD5vXgjvr9uL7pMFLsBpxSx6HNCaRGqxCnErqWU7qq4RaJ0elgrk1hhxudMhWuaT4KANgaPYDb99vCBiQrgVfKNyHDxtwLpB43ItyMyJ4+MBbHJ3qxvmg9Xiv9AV+d/IybwAMYhzFL4fv9ybAGZrLvdIvQIA/HIAsjujzsfaWMDdMuc0qxMmkCapSR6GfrxFOV26Dhnf8IsRuXtxZh4/FPcEFHKSLdNsS6LOiyOdHQaUWyyLev2h068ijGZYHWbQ8qbu+qy8PuI4yBIYGvT3900zisMh7A3I4yAECWrR0qjwt2P6l1Wq1Dm0wNkdeLCFfwUPwUuwEekRh7I9K453IStfiqdAOOHnwd1zYzE9s6J3M/ImXRsq2tcIvEqFL6XPx4iRvDUyK4x5FOK25uLMAkg8+ZvnRkEnQa2e9aN/lcQx3c3wviJgZDre5+u07X/fbU1O63Z2d3v33kyO63T57c/fY5c5h/obj4YuZfKK66ivkXCuLohOKOO5h/oXjgAZ+jFIxly3xuaDDIzFooiBMbCuLkhoK4caEgbl4o/N1Af/7H+55rwYVwPLYZ5pmzgQ8eEm4/C33vHeVYoKQVR1d+jFm8MhMAYFt8NfAeWwLB73PwE9k0r/sK4XHhgcc/w74nXZUH8EJYt6SPA5qNaHvnA3QWNQO7KuAe5gurwqefoqXTCjy/HQBQ/6+lwLQMbvOJ/74GrM5Hu0yYtONUUxesTib0qdFgQ8fDjwMT0gT7nGwwIO/p1bhpSn8u7Db3dAvyKtpwcb0Br26vwrZo3wA84DqTvrd0k+Dp57XD8e7jW1DyzHxYrr4G2FOJln/ej+mbixGjkTPZdJuMaIiMxxC/Y1ocLhif2Cp4rnzFW5DsrwZK9Thw2Y24ZvIoZoNf39v77Qkc3F+N9gGDoQPgYsc1p4dPAhq7UKWKwuspkzE3PBHDgID7nvnjg8Ap36C8cuxUoPgwNl54Iy6/aTz0nx8BChugt7jw2i+lGHfLfOADYfkMADA53BifEcM5N+sHTIVELMa0fpHYerKZq9G5Q5GIVf/5AEtmZML47C9IFDFhfSTM9fglfwPWH8Vt0zPw7q4KFK/9HubV+UiSiNBgsGHf4MkAGzp5OGs0RAYrcp59DFixE81dNrRedDmw1Rfu1vDlRqTrNHhqYxEXxq7PGIRDD38NQIQYjRyOl5jzaXa4gU8/xVcfFyCurhNLZmRiTV4V7piVhYc2HMM2v5BNuUQMh9sDuUSMf80dyDmVYZPGI+vtO5C4ci9aajsxIDECRY1dTBie2De8sWzagj3rjgClelS1mRGljsKDIxZBJ5Uge/1HgFoG5QcHAL8akHaxTPA4QatEXbsVnuQUtM2YC5zytbMsKgkIsb66jnUiP3bGAggeCm/+x52C85n8/pvAip2o/8/zMG86Bfg5Ky5x6PDKqhETkJ6bi5qNRQB7LQzfb8Z+XrK73RdfB5xsRoJWKSi1EqGW4fIFY3Hw6+MYmx6F+c+tgkgkwvANR1FcwEQambXRQG4uOms7gZV7Efn0E3g1Ow4vebwQiUTQDsxAK3sNp7z1LHI3FaGh08atqb5v7kC8zDr/o6+/FJt/PIXWq69DS0EtchJjETU8GZuONSJz6mgcZ9fuRr66HMqXcmFzeqB9/BHkHG/E92w/GJqsxaRneZE8fr+5cSebgDWHADDXULX0HuDd/djz9Bso31WBdJ1GeJ9nf3N1RxuAdUwJoyKPGmDDapNeewFI1CLq/f1AWRvC+qcB9/AmGXJzoVx7GDjWiEi1DJ0WJ7TxzATIC79UAHHDkXLNIuDnEvx38WjkVbQhLVqNV7cFLl0gSMQiuD1ehGt9v9m6V3dB32REXFoiMucOBUr3ISZCBeTmgvzSXb86n8lS7PEiLVoNEVKBEt+ymEnXXoxfNp1Ca84IRAPYp0qESGTEjDWvQfP8dhiggljElII6WtuJsf2iIMrNRVphPfB5IR68cAhEL+b6Gsre9xayDw1WJ/AfZs13pEqOiGuu4u7p088bjYp9VYLP2Tl1FhospzE8JRJVJa1cGHVZmxUGqxNzGpOBNGYick5OPK6/4RZc5/Viwet7cKqxC6mDMyB5ZgdzD2aJeHc/OixONBlsSJoxHvjv9QAAVXUHwMuMzke2Yzu0P50GtpcFbMt+8A6kjkjyPcF+npQoNVJWPhWwf+1jPwK8ZQ4WiRyFsy5BRL0B0twdwp3ZvpdW0Qa8u19wD5qVHYsx93wo2F3HXgcicAe99RLwyi6c0iZy9bsS5s3EqLQoHFk2FwqZGFaHGzFhi/DQhqPIK6jDsosG4+Yp6Xhhy2l0FLXB6/UGzSj/V4M6uBQKpU9DSnX4l+HoLS63R5Cx0x+2+kvQdXahEgoBwnWr3ZXLOBNkEuGPEkmE02lxcjOz/msS+YmA/BPyhCqxUqU3w8YKXECYZGpNXhW+K6zHle/k4ZlNp/CPTw+hpNmIHcUtuPHDg1i1swIrfjoNhUz489ObxEUA8O4upu7orpJWqNl1r2SdVpvZwdVj9K9XCQgzPxMqWs3c8yRDbjBIAq8mg02wTom/jhAQZjwtbzXhQ1Zg+GfJJYP9rhDrKv2PC4DLrpkS5XMPOyxMpu7M2DAkRig5AdtgsOG5zcUoqGpHl83JXSuT3YXqNjNKm00Qi4AbJqcDYGrKWhwuTGAT1OznJQ+q77QiWi1HUgTzvi1Ge8C6PbJWjIhb5lyYccXbebji7X2CGowWB1NTdmdJCxYMS8QtU/sj98FZWDw2FYMSwvHlIeGSjcRIpu33zxuIm6akc8+HsddfxfalgfHCNYDhSrZ/ONzcWlCb04PTzUYYbS68uGgYItTMAFIh7T5JGAAMiA+Dw+1Bi9Ee4HT0VLMXYBywUPjXzySJcho6rYK+E6aQ4uObx2N8/2jcPTsLwSDXosEvEVx+VTsy2ERAu0v16K/TYOu/puPxC3O4/aLUclw5NhWHHp+Lj24azw1048L5oaBMe4j4iFDJIBaLuO+eVsmc03itAukxasRplWgx2ri14GQwDjCOr0YuRZfViVajHfFaJVc2ZXgKL0RZJUO/aKbtKjZZEcCU5/nhn9O6zWJO6kADzDrz9BjmOI98fRwVejN3rv0ha0v9IQmByJpTpSxwKP2fS4bg0QWDcGTZXFw7MY27H5Fmrvi5BHKpGJeOSsbLi0fi3jkDUfX8hVwWan9IZuAw3lp/ktwpSi3j1nLGaIRt7het5sqEpUWrEc9LrJSgVWJoMuPske/z3nI9hiRpEamWI4Y9/hWjfZEMJDHThcMS8dXtk3HbdN+EaDDCFVLOHSffNSl7EkhbLx+djDW3MMshOi0ONHbakBSpRHa8b9K3xWgPyNCcEMG0RSQScdnAb5oSGJmgVUlR2myEy+NFcqRvsranxIBx4Yqgz6tkwtftfmgWtt03I+RxyBr+Mf2icMdMJtx4d6ke0SHW3wIQrJUm/TNeG1inO5brA8yx0nUayCViQSRRAnvvj9LIoZZLuetK8kaMSouESCRCtEYGh9sjWGP/V4YKXAqF0qexsMI2VLKZnvgkrxozX8oNWb9UzK1LDdzGTyziH7rKT5qxv6LtjBIoudwePP7tcS5JCcE/yRQpat9hdnCF3U02F7aebOISZ5BslFKxCPWdQlFYEaLou9nugtXpRpRaDqVMDIPViXX5NUhfugnLvjuJez4vhMXhxrUT0/BzUTNe/6WUW380Oi0SBqsTEr8Z4jO9Pt8dbeAGGnwHqrv6p/7JWgCgtMXI7dvSZecmKj7Jq8K8V3zZautZ8d9osOEYry6v/3XlZzy9/K19+M/GIiZjr92NMf2iuOyqpPQKySysNwoFUylP4CZolXj3ujEYz5bc0CplWHbRYDx4fja3T7xWyQ14+Xywp5JJNMIOkg5VtWPG8ly8lVsOrUqGpAglwpVSVOrNsDjcSI1WM8lzbC6oZBIuWY4uTAGVXIJwpRRv7SjDpuPCkhihkqEQtpxoQr8YNXIStTDb3TDaXXC6vQKxDoD7jHzIQDNaI4dSJuEERZiSCFymH2TGCsUBqY1rsbu4jLQdbMIuAOjPS4oTTKT4Q8L8QtWTJuVWotSygG098UVBLSfCmfZIoAuTo7rNIsgkPiM7FjMGxuKLJZNw37xsbL13esCxqvS+8j7ka7ajmElANm8Is2bZ4nBjWHIEIlQyQdZd0vYItQxK3kCeL4z0JgfKW02cqIj0+7wRKubx8BRm4BwfroTe5OAytIYrpdCw4iItWo0whRTV7RZ4vECcVomJGTH46V/TuckXgLnGH940Dv++eDBiwxU4f0gCbpuegfvnZqMnyABfIRUjQiULEC5SSXBxHErgaNl+NymTOW/+tXEBICZMgdumZ3LZmLtsLtz6cQH4t4uchPCArOnkd2HBMOHaciLKhQKXETVRajkncKL9BC6/bBIjcJlz8fSlQ7Hxn1M58f/+ngosfHMP8ivbMXNgnOBYcwbHc5NFRFBJJWKM6RfVo9MnFotA5n5To5i27Hl4NnY/NAtO9mSkRKowvn80JGIRPjtQA4fbg+RIFQYlMgKXfL9/LhKWbEqM8N07ViwegXvnDAhIsAUw90sykUgmRgChwCWf7+7ZWfjsVmZtPfmu+KP0E7ip0eqQExMA8PHN43HXrCx8dftk3D8vm+v7/uWM+PAnZeYOZpbBBBO4OraPkskCmUSMTL+2BHsdADwwLxvvXz8Wo9OYc0b6ULCkgX9FqMClUCh9Gi6baoisq7XtFq5kRDAX9ni9AUabC82G4C4u+YF3815L1qfaeKI4oNA770dk5Y7ybmvo+VPWasKn+2tw+6eHmdIarNjw+LWfPGwzOzjX0WR34fP8Gry3uwKPfH2cc9ZGpUWitEUoaIMJQrEIMNndsDrcUMoliFTJ0WF2YPUeYZkErVKKpxcOxYT+0WjpYhwvjVyCuHDGZWz2y/Z6prPGh6s7uAkEMqAHfE5EbZD6p/7lNpIilDhQ0Q6H24O0aDUcbg+au+y4e90RPPHdSZQ0m7j34BzcLpug7Io//MEucbi6rE6Y7S4kRCgxnx24Enetk62P2OxXFqWEJ3CnD9Rh3pAEXMAOuAbEh+GWqf1x56wsbgAbF64IEBqAr0QJcQG28cKkI1QybgBO6vCGKSScQAlTSjGQDZ2P0jDPuT1eQTIrwpvby7otu1RQ3Y4x/aKgkUuYUHHWzSMDS8KcHGYwlxHrW/tGBtrkf6597GBfLZeybRQO7slg3Oxwc+5im8mOmjazoJwI4Bu0xvqJGv4gdHImU1+0y+ZCXLgCc3KESxImZcbg8LK5+OX+mQGfX+4nZPxdw3azAy8tHiF4LilSxTn5F49IwpyceK7+KoFkhBWLmAgOlUyCmnYz9pbpUdZi4kT3c5uLEamW4apxqdxrs9mIABVvoB8ZwlWK4w2S3R4vzluxk5sYilAJX0Ncyhz2vYk4Jt8/rVKG164ehZxELfrFaKBRSLht8ez5HxgfLnBlRSIRkiJVnEOnlEnw6IKcgGsejBiNHGIRI3RFIhHEvOPGaOS4cFhi0NcFmzAibQF85Xh6yiKfwgq7bWxI+7QBTD+SBikJZmfvNwv82sQ5uMpABzdCLUO4UgqxiMkKLnxvFfeekWo5hiZHYPagOJw3KA6x4Qr0i1FjZGokck+34mgdM3F3KVsCKUYjh0jE1LIl0Rv+348zYQab1TghQonUaDXc7KywlM1C/cRFg7nsv0kRKm7Cinzvt51qQX+dhnOA+SJwcqYO9/Lq2vLRqnz3Rf6EGr/fL180HOkxatw8tT9XRzheqwxwa5nXnZl0mjEwFg+wk5ESsQhW9hpfPjpwjT+B3/dJre54beC551x83vcgJ0G43IlM8PijUUgxhxXPANNPx/aLCplH4q8GFbgUCqXP0WK0cY4rEU6hQpTf3F6Gu9YeQUuXDdmPb8GWE8JZYiJYGgzBw3XJ7xBxWY7UdGDYk1tR224RhC12diNwgd4l1iEQl7K23YKVO8rwzCamOH2wUhQA41pxIcp2N7psLlS3WbAuvwYbDtVBIhZhUkYMqvRmWHniJZiDqwtTwOJwweb0QCWTIFItQ4fFKXAKAGbwLBKJEKdVotlog8HiRKRajnClNKjA5Ydh/ni8EQerAhOQEC4cnoj6TisnOoubfPU0SVh2sBDlilYTYsMVmMC6hJOzdNx5H5XGhEOu3FEmyDTZZLDB4nBxg9hl357A0z8UCcQkPzQ8mJvTZXPCZHchTC7lhBQJB2812nG4pgOdFqdgQF3a7Dv3ZFA5OUuHnQ/OxCW89V/nsSIrQiULWnKCnGcSosyvPUqEYlq0GqfYc6iWS7nnwxVSXDmWEUQn6pntoSYimrpsmL58R9BtALMcLEoth0Yhhdnh5kJyw5VCUT59YCyOPjFP4MyRQTsZxJFBJwmx+z92/TcpnULgO7jkfqA3OVDdbkFSpIoTvYDPwU3QKqHgPT+CFyarkktwFXs+vr9rKjcQJqREqRCtkQucITIYH5UWiRsnp3MDVyL+CPfPHYh5vMEmwFwzMtExKzsW798wlrs23PElYux8cCaOPXk+Cp+Yh2kDdDhS04lr3j8Ai8PNCVwAuGZCmsDBIuHKfEcqlPsczM0k9Vv920SEb38dc0/ox7qPpK5vuJIZWG++ZxrkUjHCFFLOVQ/lNv0WpBIxYsMVAkH05v+Nwht/G4VDy+ZibHpg1ADAfK6pvGucFKEUhMhHa+S4ZESSoFxMMJL8JjOundgP8wbH4zFeaDjhk1sm4PELcwImQIiDG85zcKNZB1chFUMsFuG26Zm4aLhQGM8aFIeHLxiEt68dw32m1TeO49okEonwyPxByIzVICdRi3HpUZwbOW2ADgtHJCFSLeeWCfBD1c+UnESh8JrJ5qwg5Xz4jn1ipJL7TSH/t5sdGJYcwTnuoYSbP1rePYZ/LcjEGABMytQh98FZARM8eY/Mxs4HZ7LHYfbvzXKG7njiosEIU0g5ZzYUS+cPwuMX5mD2oDjcNj0DQ5IiAvaJUstw2ahkzMr23fvIxNXQZC0WDEsQuP7dMSkzBhtun8ysSe8D0CRTFAqlT+H1enH+K7vwjxmZWDIj0ydwQ4TAtpkZ8Xe4phMOtwf/+PQQKp9bwM3Sk7qaDSEEKAlRtrLHP15vgMvjRaPBJhCcTDifBq1GOxa8vpubzd63dDamvLA9pDglbD7eiNs/O4xdD87i2mK0u1DfwQg9p9vDzQwHfEYTL0TZ7gpwk2M0cgxO0sLjZeqX7q9ow/5HzxOE/hLCFFKY2BBllUyCKLUcBqsDTrcX0wbocPW4NNy59jAGsuun4sIVnIMbpZEhXCmD0eZEh8WL8enREImYOpEklNzr9eKOz5hspS/7OVoHHj0PPxU1IzVKhU3HGrkw0S6/9YsJWiVONxtR2mzEq9tK8Y8ZmRiWEoEKvRkZOg1WXTcGpS0mnKg3YAOTewazB8Xhu8IGrqYsodFg49aw8cmMDePef1CClgvBJnVP+aF7BtbB1Sj4AtfXn654mykFNTkzhhPX/DqvsTzhS8QCYen8QRiZGolJmTHYelI4OaOUiTnBnxhkMEiESWq0mhOcGj8Hd05OHJIjVfjHDGat3YorR+DtneUCF/vhCwZBJRPjyY1FAe/BJ1wphUYhQUOnlXs/rTJQUBFHikDC78j/RCSSfaZk6VD1/IUAGLfjZAMjxsmEwbu7KziHWm+yo77TGjAho2QHreFKKXRhCsSEyZGToMXcwfHcNVFKJfjPwiG4a3YWEiKUuGZCPzjdHq6WNRElfIEcr1WivtOKMIUUT14yBHKpGO/uqkBmnAbbmHkplDwzH3KpOCCCJFoj5ybJgrnzBH6fSNdpBHVVs3luzui0KMHgnAxk+S6Vv1jlfw5/jtYZoJZLBBMFADPhsOVkE6axEw6j+0VCLZdwUSr+kxr8GtL+DuHb14zm7l2/hYUjkwXfgYuGJ3WzN4NIJMKnt05AXnkb7vn8CDbfOz3g/Lz+t1E9HmdwohZXj0vF5weZTM/9dRq8e/3YoPuO6ReFMf2icIo3EQX4rhXfwSV9ljhuS+cPCjieUibB7TMDy8zwmZARg1/unwm3xyvogzfy1rMm/gYH98VFwyECAsKZx6VHC35rAeDbO6fgtW0lyIwNQ06CFssuGowFwxIw6TkmEeLwlAjsON0Cm9PTe4GrYs5ZXLhCIGr5/V4TYj1upFqOSLUcVc9fiPd3V+CZTacEzu+v4cYp/QXnNhT/mOG7bo8uCJwMAZhz+spVIwXPke/84rGpuH5S+q9v6F8c6uBSKJQ+Ad+x7bA4OReBhChbQji4BqsDdpdHsOaRJBvqMDs4544vSOo7rbjxw3ys3FEGsuyVCGnyvlanW+Dg7itvg9nuwol6A1qNduwoboFIxPzohiukAtFpc7rx5vZSQSke4tJOX74D/2DLVQBMKK7Hy7QvWKIbuUSMhk4r2s0OiEWM0Df4CVxdmAKDErRcOz1eJrTV5vRwYZhPLxyC5YuGQ62QwOJww+JwQSX3ObjtZgeiNXIMTWaOM5gNq4oLV8DqdKOuw4oo1sE1O9ywuzyYNyQe98xhylCY7C7sKmnFRW/s4dp13xdHBe0kCWjIsctbA91SALhzFjMweGDDMWw63ojrVh9AcVMXqvRmZMRqEKmWY1x6tGDNJnEtSJgeodFgRV2QcGcnb800X0SY7C60Gu2C7UdqOmF2uBGmkHCDqvpOK8b2i8LisUyY2vj0aCy7aDAWj00JcPd03QwqFVIJFo5MhkgkCgjXzOCtMdUGES5ansAl8B3cMIUUUokYe5fOxnXsQOmKMSlYfcM4wXHOy4nDjVP648ObxmFkaqRgWxJvEBqulEEtl8JsDx2iTOAP5LPiwhCukHKDa7LWPJgzsenuabiaDcOVctm7W7lEbnqTHdVtloD1b2TiIUwhxcSMGEwfEIsXFg1HGm8/pUwMpUzCnS+5VMzUPWUhoaj8ATsJKySfJ1gyICIQ/QUA303yDwMOhb9w5/elgfFCB424gsS9DldKg4bNAqGFTWSQfjV9YCyqnr+Qm2BQSCWYnBnDTbb4X3P+dfRfQzp/WCLn0P8WHl2QEzQBUW+YlBmD/MfmhBT/PSGXivH8FT6XN9hkkz85iVp8dNM4XMxGawxKCMeNk9Mxe5AvLJ44mcEm4H4NErEo5PUnbf41Anfx2FQuEsQf/z4/MjUSH940HkqZBGKxCLdM7Y/ECBXnng5PicSiMcw9M6GXbj+ZUPFfJysRi6CQiqGUiUN+bj5aXmTLn5lRqVEYlBDOra39X+XPfZUoFAqlF+wp1ePaDw7g2zuncD96RMQRkehwe7jQXr7jQPYrrO2ESiaBF158fbgeY/pFo5KXxKmBl234ze1lyD3diuJGI8amMz8iROCS0Firw825slOzdFi1sxzLeWVA2swOxGjkkErErKvpc+x2l+rx0k8lSNdpOKchVAgzydBY224N6gJnxGo4d3FQghZFjV1o8Vs/qQtXII1NlJGgVWJPmZ5buzkjOw7v8wTNhkN1MNmYEGWlTIJItRydFgdsTg+iNXL0i9HgiyWTMCKVCacizk9piwnZCeGCwW1ihAoauS8T8r+/P8l9zucvH4alX/vKmvCJC1dy5TeCMbpfFMalR3NZN+USMW75qABtZge3loycG4DJvhtqvV2jwYb9FW1QyyWYmqXDvvI2mOwuzBwYi2N15LwKhcNtaw7h0pE+h4hMTkSwSbkAJmtulEaOF64YjiUzMqELUyBCJcOLi0bgns+PcA5OuEIaIHhDQcSTUibGgmGJUEjFXEhyMDHID1EmaORSbiAXKrQtLUaNT2+ZgMLaDrz8cwkXvjorOw6zsuOQ8cgmSMVirFg8AokRSix6J4/7LBq5BM1GO0rYEOxQApc/iLxkRBLmDI7n3Bfi4GpCtI8cM9i4/2RDF4w2lyBDK+ATeUqZBCt4kQNKnuPpn1zGn+SowGy85FyStsZw64kVWH1jYMjx4xfmcIly+FlWu3Nw+RDRKhYBP94zjZu4AhCQ0Is4UeQzBgtxJ8gkYkzMiIbN6UFhbSdGp0XicE2nYD1rd0zMiOHWf4eFELgqmaTHc/xX5qvbJ2HryeYABzsUM7PjcKCyHRKxCNEaOZ68ZIhgO5lMmj5AF+zlZ5VRaZGI1sgFEz6/J4kRKhjtRgxJ0mJ0WiTumJkV8vvvD/mND5YISiWXBCQ8DMXFw5OQoFUK1qP/GYlQy7AlSPK5/zWowKVQ/iXew7YAACAASURBVMdZl18DtZxxgf4o7C43WrrsAifpTHhvN1M6plJv4kKNyI+ameeCLt9ajPd2V6L46Qu4gRRf4I5IjUBcuBKbjjfiqYVDUck6hOEKqcDBJWtU2y0OuFiBSYQ0SW5kc7phczHJbe6bNxCXvxVYb498Xq1KJgizPc2uhyQiIFSCLD61HZagAjczNoxL3DEoMVywBpOgC5NDLBbh539NR1mLCXNf2cWtifN3aDQKKZe9mQlRlkFvYpwZMnjnZ8Llr92LUssFIakJEQpukGJ2uJAcqeIELslQGgpdmCKkwI3XKjElU4f8ynaEK6W4aUp/rnYqP9QyOVKFJy4ajPOHJgQ4R4TyVhM2H2/CwpFJeP6K4fB6vWgx2hGjkeN1tkYiCd/LiNWgotWMwtpOrh4tIT1GjUWjU1Da4osUYJK4iAKy/xLnNTFCibxHzuv2PPAhImhgfDheXjwSz/54itvGF6vDkiNwvN7AiSt+UhK1QsJdI38hwmfqAB2TFTpTF7Bu7ei/5wFgnBN+pu9wpRRqhRRuj5e7HqEG+/z3FotFgvYT5yqUk7J4bCre212Jy0YlC8oWAb5yPEOShevZyP3AP6stP7tyT+IrmMNH+jdpfzQv3HrWoLiA/W/l1aHmi9pgTmkwiDOdHKUSiFvA55bFaOTcZB8ArsRPT9mfP79tEjweL040GNBhceKG1fkhS4n5w49yCJMLr5vG79z0Vcb0i8aYfsHX+4bixsnpGJ8eHdAvAcbNPPbkvKBh/mebaQNicXjZ3HP+PqFIi1FDLBZxfeVMnOSFI5Owv6IN95w3IGCbWibplXsLMGKYrBem/PmhApdC+R/nEdYlOxcC93STEf1i1D0ODP+zsQhrD9T0+sfa6/Xivd0VqGm34OmFQ5HHJi+RScScUOyyEQfXJw7XHqgBwISMEgFFBG6b2YELYsMwNj0K3x9twI7iFpxs6IJYBIzrHy0YqJNwYofLg3I2EZN/iHJJsxFHazuhlIoxOi0KH988Hn//pEAwsCTOWbhSii6bEy9uKUal3sz94JawwtQ/u3EwCms6g4Yo88PhchK0AOoD9iFrPEUiEVersIb9vP6Ddo1CyglalUwMCa8QvX8GT8BXNxFgEmLwHbt4rZJz45jkVz7B6p+YxR8yGJ6apcOeMj33vFQsQrRajkmZMXhlGxMyzE+sw2+PSCTCzVO7D1s8XN0Bq9PN9ReRSMSJ5I13TUV1u5kTXP2i1YJMzVqllOuPt07LCCi9EioDbGYc48K1mc6sXANx4Egf468x4wvEwYnMmmHyXeM7EmEKX4hyT3UiVXJJ0LIcfNHKd1nClTJBDWHmuVAubOj7AOkz/ms/CQPiw1H1/IUB2UAHxodxk0b+rjhZNyuXCt0cpZzv4AZ/vzW3jEd5iO8omXAj539cejQuG5Xcq/DBKEGIcu9ETGKEEjKJiKsZCwAvXjGc+14DwN6lswNep2KjMXpCLBZheEokVyKot/AFrr/rS/oIydRN8RGvVXabeOv3ELd/Bv576VA4zqCUHh9dmALvhVjzrJJLIP+NSaMof07oGlwK5S9IWYsRb24v/d3er67DgtFP/xyQxOZgVTuG/ntr0LppnRYHzn91Fx7+6liPxz/MJus56ud6hSKvog3P/liMT/fXoM3s4Ab0VoebW9sXzMEljikJXbW7hOtkU6LUGN+fETK3flKA1XsrkRKlxtAkLSr1ZhgsTrz8cwkau6xcSCsRn1Yns66QuIpv5ZZjf0U7J2hmDIzFsosGCz4HqTGqZTML/3CsEVtPNuEQm0GYlAhpCpHBmdRBTIpQYn1BLUx2F0alRXLrOgFgcpbPCSXCKfA4PtFHBtJVbIhygMCVS7hzy6zB9Q2Kgzkw/MEZk0WZOR6z/ljJDW4tDhcaOq2YPzQBOx+cKXAsHr8wB/mPCp1M8tnVcgmOLJuL/142FADjGIvFIi5E+rbpGYKMpL3J0sovx0Cc7GBrWIelROCi4Uncelu+eN18zzTkPzaHe0wmGsg+6vQ3offuD/r+xME90wEdOf/kdRoFT+AqpciI1SBGI0ckKySCiSa13Jdk6mws7QsTCFypQOgAoV1RdTeTYkR499Q8iZ+Q4mct9Q+/JmI4wMHlhyiHGAhPGxAbMmkM6d/k/wi1DK9cNVIgOENBBF+4IvTaWH+kEjHm5MRjJi+r6uJxqTifV9NTGSQUOEwhFawL7olItRyx4Qrc0sMEESE2xBIAwDfJoZFTz4USnDitklvjfjZRy6UhE0xR/trQuwmF8hfkqlX70WZ24KYp/Xu9DoXg9nixZE0B/j4tAxMyfOLHP/Mrn8e+OYF2swP7yvSCgdLbueUw2V04VN0RkPKehMVuPNqA164ehboOC+QSccD6lboOC9JjNChuMqKgqoPLvNkdep6zVc8LkbO5PD4HlxVh/LImJMtwflUbgAEByZbitYqA8gz9dRouw/DHeVV4/RdmYuH8IfHYetKXrVRvcghqlxL4A0n/BDDksVYpw4H2di5zboPBBrlUjKo2pmxPoyEwmzHAZJzcV96GWdlxGPffbQCA2dlxuGhEEr4oqAMAQSmTtOgQAjfcN7BVSJlESGQtcTAHl//Z+FkpgwnccKUMujAF9CY7InkOboxGAblUzImQJoMNXTYXRqZGBmQKzooLC+g35L00CimiNHJksWG+sex+CqkEp566AAqpWJANujuB+/HN46GWS/Do18fR3MWEoROR151TQgRXcqQKr1w1Ak63N8AhTOAErhiABxJVHba0rMBy3BhwvP6/skwDEW1EcKvYayMWMQ7dT/dOh8cLfMDWLOZf22kDdNhdqkeYQso5o56zoHAV7DV2e7wIU0pxw6R0xIQpcPe6I92+rru1nS8uGoH3dlVglF9Cq56wOT1442+juMznfMhadn+BK5OIIBYxia16u94UABaPTYHX67smYYozH0QTB7c3YpgPKQlzJixfNBzxvcxKSzjIm8DpiVC/LYDP5Q7lyFMo54qFI5Nov+ujUIFLofwFIcLMaHNhweu78eD52b0qe8C8xoltp1owIiVSIHANVmfIEDWyblPpN9NJ6n86/Zwmr9fLleXweJnHt3xUgNPNRmy8ayqGpTDuWkOnFVNf8NXOJGVXesLCW5PKXxtrc7i5urRdVhe8Xq8gRLmJFYqkBql/uRxS4++lK0dg5Y4yVOrNUEjFGJzItHd7cQu3b3qMBmq5hBPQPxc142deeQ4Cye4MBApcMoANV0oFZWEA4IF5A/Hsj8XYXtyCilYz5FIxRPDV2wUYx3nxWDXbdgVajHao5BJBaKlCKsEj8wfBG+T9Cf6CMlIt40S1f3IbvsAldXAJodZFzR+agDX7qyES+dyahAhmX4lYBKVMzLnVwRL1BHP5SDg0CS0lznA8rw0kiU5cuAISMSNUultnSEo3qYMmZAr9czkrOw4vXjEcl4xMCulIknW6KpkEEDF9QoTgg37S7rFBwn+7g4j3JWxmXyIcNAopRCIRF0pNzgFf4K66bgwKazsRqZZzkw7+Ib6/BpFIBI1cgi6bC+FKKcRiEYYm9S5pFuCrM80nOVIVkHCnN2gUUi4rrT/E9fYXuCKRCCo2o+uZ8OIiJlHVOzvLufc+U8j9obcJpn4Lk7POfaKiH/45NWgpM3Ju5L10qSmUswV/zTulb0EFLoXyF8TD1qprMFhR3WbBA18e7bXAJYLMf6DR1GULKXA5N9TuRkuXDS1GO4YmR3BhcyS5kcvtwb++OIojNR2YkukbMFXqzZyA+dt7+6GSS3Df3IGCWppkvzP5DIAwu7HV6YabPTcOtwc2p0cQokzEISkd5O/gkvWZi8akYEBcGBau3IvkKBVSolQIV0gFiYO0Khmi1HJYHIHhwyIRQMoJ8ms4+rvDpNwNCX+VSUR48PxspEVrMGNgLJ79sRh3rmVKAvWLUcPicHMJrvyJ1sjRYrRDIZME1Olbwqunx6dfjBqf3Dw+QOBGqHwCN6BmJe/YarkUmbFh6Bejxtyc+JDO49L5gxCpluH/2XvveCmq+338me27t1LupQgCIja8YO8Fe43GFlui0ZimRmOKwWj82j8mmqKJRpOo0Z9GwY6AoggIogiI9N65vd+7faf9/ph9nz3Tdmcv99Kc5/XyJXd3dubMzJkz5znP+/28zxs7GPFsuabB5blrcWB1KeZt0PJoh1uEoYUtSCOFU9I9JeJspdD6vB7mrp1PSTKeo0dwpuAKgoDvHWtdBoNApDLEEVyPYD+hX/7AuUVP+MMBL6sHC+SUZWM4LuU30yKDtq0PJ2WfWXIVldXeKT9SGtRykekaFsqvJky745ReMR4K+jx48JKx+O6R9j4DtFhnlRMc6gHBJRhNpopBedgPQQAqHZYI2tNxuMHYi2Bn8OXChQsXPYVLcF242MORyEj4eFUTLj1iKJuc06S7ORtGaWUuZAcitkaC29SdxiGDzdunRJmRiHhGwnGPfQoA2Pr4RfBnJ32dCREfr2rEtrYEPlhWDwD4YnMrBpeH0NidwvQVDWx/sbSEWFrC/e+vZGGBhJZYOm+oNCHBqbINnIKbFGWdYtqVFJHMyDrCqf1eZt/zGFSWI0fjh1fivzcdi+NHDYDHI+CgwWU6hbky4kdF2G9Zvqd/JMBqPvKmUgGfB1N/cQqG94+gNOhjShlNqscOrdDV1awqCzJCO6AkAJ9HtCW4A7I5qSGfJ2/+Io/ykN9EbuncqF3GPEadghvwoH9JAJ/99oy8xykJ+vDrcw8GABYeypOrG04YibvfXo7B5SHUcJPgkN+DlKiYCDuQM2iixRUiT4NtwixHDIg4ViQp7HpIRc7V2SoHtxhQnw75vRAErf96BPv71BvmMXQeRnJ16piBmHbHKTiwuszqZyxv+5pjd77+KKDdf79XYGq701IwY4daE6Ji8PV9Z8Pn9RQ0abrhxJGIpSTLnFKN4Pbs+BSa3BOC6/UIqAz7iw5R3tsgKebybS5cuHCxM3AJrgsXezj+9NE6/PeLraguDzKFhdAStc7NzAcqZ5PMGAmufl+bW2IY3j+iC+Nt5NRSVVUZWelMZPDQ1NW63+9oT+La44ZjXWMUz8zWwvR4x1sjuQU0MhhN55Se7z33JcrDfvznRr0DIq/gktLo9wpIiTIr2wNoTsqJjIRBZSFdHmYiI0NVVRPBLTeEoU44OFfGY0x1qY7gVoT9tq6f4YAXsBGjrVQMUkmNjrT/vuEY/HPORsxY1YSGrlTe/FEK2fUIAnxeD84fOxhXHj3MdnuvRzCdL4HIgHXZE+d1Qa0Q8ntxzmGDMOGg3LW99MihWFnfhRtOHKlTykJ+r1Zv18Lch1RWIrgVET/+dvUROMWmJuQfrxgHp4IknePQyhDqOpMIeD2MnBWLt39+ErqSuZxxr0eA36c1xJtHwe0NRLgQZR6CIOQlj8P6RXRK8M6iJOhDaTZMmlCzX4VlXcrexoA85kY8Qn4vfpVdhDF/57HM23WCMdVlKA/5bNMDCuGGE0fi0CHWCxH7GnoSxu3ChQsXVnBHExcu9nC0xDTFzkq5a7ZR8/KBD1GWuNzZJo68Lt7ajiuf+xIPXjIWJ3H1SL/Y1KbbTzyrpLbblIwYUhHGoUPKsWT7KgDAD04cARUq5m9sM21bEfajK6kplERwF2bdhO3OAdDCtEN+D8pDfqREWUfcu5IiEhkZgytCaI9nWJ6drKjIyIqpjmo+5dg4Ga8MB2xDB4sNLaXzNZYOOWJ4JZ68ajxmrPoY44dVsuttBQrZ7cyS9ud+YG00895tJ2PqsnpMWVZvqxISsbXK/eOdTq1Ch53AWLIh6PPioUsPN20X9nvRCRFWt4XINR+JkC8EtZgay6R8Dq4IA+hAedjnKLTZClZldIJ+rR96PX3r3knnYVeKZ1ehNOgzhbp/8ItTdlNrikfI7+0xwT18vwosf+C8Hh/7rnMO6vFv9xZcWDMEy3Z0WdYpdeHChYuewI0HceFiDweFm76zpA5TsuG/BF51VR3KUxTem8zIujIkbVypn+fnbgYAtMczJpWT0JkUEcsaI5HjLKCVvCEMrgjhuFG5wvYHDCzBP649ynJ/o6u0UNnWaBopUcY323NqKe/mev/7K/HuN7larvWdSQwoCSIc8GbLBEmMEHUlRLTGMhhYGmAGSKTEJTMy2uMZSxMbKxgJboVN6ODlR+5XdO7icaO02pinHWRWH8tCfnx812l48nvjGSF96NKx+OiXp+q2O3J/zVG22sboiXDE8Ercd/FhuPa4/XFhzRDLbQ4ZrOUGn3eYOWa9VBei3LcE7V8/OAaXjB9qmbN55P6VOHL/Stx74WEWv9w5kDo8NBvunK8ma08Q9Gv9w9PHr2Cm4O7m8ivjh1fgqP2LczzekxAJeHu8mOOiMII+Lx64ZKxtXWgXLly4KBauguvCxS7CyInTcOuE0bj7/EOK+h1NUj9b34LP1rfgEs4FlFdw87kg80hxObhpLne3g1NhP88a/gR8HkZw+XxQQAtLjmVL8vCleo4e0Q+z17UA0Gp/juHy/AZVhFAS8JlyYgFgdFUplmzvREssjbeX1OIP761k3x3w++m4/Mj98NjlNXh94XZdeHNTdxo1+1UgI2mmUtGUhP37R7CtLYHG7hRaoikcMbwC9Z0p1HUmUV0exI72JBIZGW3xDPpFAjj94CrmWGqH0VV6glsa8pkU0LUPn4+A14PTntCcoS84fDB+cWZhVaKqLIi/Xn2E7fcHDdKuIYUUX3/8CFNu7CXjh2J4/4jj0in5lKGbTxmFH5400tJY56gR/TC0IoSmaLpgXuPOomZYBZ6+9kjL7yIBH9699eQ+OS65KFPt2vJeVkCDfhVp7AIFl/I/d7OC+9vzihvz9jTccdYYW8drFy5cuHCx58EluC5c7AKQuc2zczYVTXDz+eLwyunWtgSOcEBwE1wOLl9yhtx+42mJhX0mMhIjuEMqQjqC25UQWWkb3mjp4MHlWLilHfGMjCEVIR0RK8vm4VWE/abw4AOyBLI1mkZ7QjSd9zvf1KE7JTFyG/B6mALdvySAzqSIpCgjmpZwYFUpGrpS2NwSR1s8g6qyECIBzbG5uizECG57LIMBpQH85Xv25JIwrF8Y9198GERZwRMz1mFweYipOtcetz+uOXY4C5ul+33xuKHMKbk3cNCgMoypLjWRW0ALrzaGOO8M7FxjQ34vPv/dmYhlJF0N3H0JpOCuSryLwMAdKA9f06v7Zwpun+fg9tzB10UOTmpzu3DhwoWLPQduiLILF7sAKYvaf3aYuboJh/+/GYimRPz2zWX4cGWj7nveJZg3mbr+3wvQnRJZTdod7Ql8sroJN/93EV5fuJ1tx+fg8g6/nVkFl1eFkxmFEdzBBoOjzqSIaMpcbueQwWUsHHhwtvbnyzcfh3suOITlMfa3IOL9In54PQJaYxl0WeT0/uCEEZi5Jldnljd46l8SQNjvQVKUkcxIKA35MLxfGF9v74CqamG7fE1U7TpIaIunHZchEQQBN58yCj89fTQ2PnYhwgEvwgFtCI0EvBjPKadSluD2tnJ208mj8MmvTu/VffYEHo/QKy6/eypIwf2o/iUEq2b2+rlec5yWK+zN46LcGwj7vfB5hD5X2l24cOHChYs9CS7BdbHPY0Vtl47I7Q5YFbe3w19nrkcsLWHq8ga8+XUtWmN6IykilQDQGtOI4C2njEI8I2PcAx/j/L/NBQDc8cY3+PErizFrbTPueWcF1jR0a23RKbjavwNeDwtRbu7m68rqFVwAzA20MyGyEGXC/Rcfhu+MH4qqsiDKss6pAHD6QVW6WqxW5kWyqmJASQAt0TQzSiLcedYYkxLKhxSXhXxZx10Z8YyMSMCHkQNKsCxbt7a6LMjUVsrzohDlASXOXFatQIqtcQGDFFxXOds7MbqqRFfv185tuqc4coQWct7XCq7XI+Clm47F908Y0afHceHChQsXLvYkuATXxT6NbW1xfOcfn+Ox6Wt2azuKUXBJbXn5i62W30cNpDLs9+L0g3MhdJta4lBVFX5D4cYLnpqHX09ehsc/WgsA2Nwaxzl/1chwdXkQHXERqqpiB5dPm8zI6EqKOhfUEQM0gtsWSyMpygj5c8c5ZHAZvB4BBwwsxZhB9iVAeNVUELQyHGccXI2BpUG0xtIsXBoAXrvleNx1zkGMYBOMBDecJbiJtIRIwIv9B+Rcc6vLQ0zB9WZVZDKZovqxPUGO4OoXUMidene717roGU4aPRArH8w5314y3t6duScQZa1/2ym4L6x4AQ9/+bDtbxc1LnJ8rFPHVLGIChcuXLhw4eLbAJfgutinQTmqK+u6dms7iiG43Slt8ru2MWr5fTSlVzdLgj4MNNR6rO1IYlg/s/Ps20tqmbrIY3B5CLG0hIemrsZv3lwGQHOQTWQJbkXYzwxrysN+BH0efLZeM5Laj3O4JRL5wCVj8d+bj7M9R94M67iR/bH24QswtDKsGVnF0roQZSL8RiddPuyyNOhH2O9FIiMjIcooCXgxckAJ+76qLIgTD9DKHR00WFPPulMiOhOi4xBlKxyfdYi+eLzejdhVcPd+8GWBTuRKZRXC3Nq5mL19dt5tMorWv+0U3L8t+Rsmr59s+d2Ti5/EzTNuxrr2dY7b5MKFCxcuXHyb4BJcF/s0iFgG/X3X1d/7pg6TFm3Pu41R4bODqqrY1powfT5+WAXGD6sAAFPIcmnQa1Ihl+7oRFKUIQjAE1eOK3jcQVl19KX5W9ln1eUhJEUZXQkR5WE/KzUS8HqQlhQs3qaV8RnMKatkahMOePPmLY4cEEG/bJgyX4JlYGkQrYYQZSKyRgWXF6jLQj4E/V50JUSoKhAO+HDGwdXs+6rSIK4+djg+/fXpmHCQpnbXd2qh2AN2guCOGFCCrY9fpDsWkMvBLXEJ7rcOt316G+6YfUfebURF698+T/H9Y227FoHRnekuvnEuXLhwsRuREBPY3LW5z/bfnenGGZPPwNLmpX12DBd7B1yC62KfBtVpDXj7rqv/ctJS/O7tFfh6W4fOAIqHVQ6uqqp4ZvZGlvOalmSc+9e5iKYl/PS0A/CP647EuYcNAgAcOqQc3z1SC5PckiXArMZl0If+kQA4wQkr6rqQFGWM268CVxw1DL87/xCcfKC9CkUGUpRfS/tPZmS0JzIYUBKAP3sN/V69u+4RnLlSxGFd1B+fdgBm/up0+DyCrgRLVVkQrbEMOriavOVhMxEGgIyUU6IpRJlcnUuCWojyE1eOw3fGD0XA54EgCBhdVcpU5h0d2nUcUNr74Zu3TjhQa0cf14l1sXeCQpR7koMrKVof93tc4ygXLlwAixoX4T8r/rO7m+EIP5/5c1z63qV9tv8VLSvQmmzFs0uf7bV9KqqCC9+5EG+vf7vX9umi7+ESXBf7NMg4Kejre6JxxT+/wP3vrWJ/x9MSnp2zEYmMxIydeDR0pfDEjHWYsVpzBl5V340NzVopmxMOGICLxw3FIdlw2nDAy8jj8tpO+L0CDh2imS6VBH3weT06Z+JNzTEkMzJCfi88HgE/nzAaPz1tNOxABJeI5ZjqUo3gijI64hn0LwlAUjQV2u/14KlrjsBz3z8Kax46H+eNHcz245TgBn1eDCgN4rCh5SxkGAAGlgaQkRV0c3nGZTYqKJUIArRQYHI0BsAMpa46Zjj+bqijSkr0pmYqG9T7BPfOs8dg6+MXwdeHCysu9l6Qgtsjgqtqz0ZP1F8X3w4c/f8djd/N/d3uboaLXYQZW2fgxRUv7u5mOMKS5iUAAFlxnrZVDGhcpHGyN5CRM9gR3YEHvnyg1/bpou/hzr5c7NOgWqt9GaLM4+vtHezfv568DH/6aB0+Wd2ky8FVVU15JFU3nf3/kmzI72OX1eDUMQMB5HJVZUVlhkYrarswamAJC90lUklhymOHlmNTSwwpUWZqJQCcdlAV3vrZiZbtri7XSF40LeHoEf0w/c5TEfJ7mRFT/5IAc6L2ez249Ij9cP7hQxAOeHW5sMXWRZ1y+yn4GeeubGWGw9dj5VXoNHdNy0J+hLhFjHyhwSG/B4KQy8seXWVvhuXCRV+ACG6hMkGqqqIj1YGPtnzEPiMFt69LDLnYe5FRMpi+ZfruboaLXQRJkXqV0O0K0BjY26BxsTcJdF+1FQA2d25mET0uehcuwXWxT4Nqu1oZK/UFiMgmMzI+WqXVr63tSOpClKkObTqbl0tK5DfbO7FfZRjXHb8/U/6I1KZFhZHHza1xjKkuwwEDNROlQWWa+kpGUyceMADb2xPoSopMySRUl+nzWAl8jdt+ES0cORLwojsloTsloV8kgOrsNgdUleh+yxPc0E4uJFRx4cJnHFyFH50ySvf9a7ecgE/uOs3UDk3BzZ1rPiVZEASE/V7EMzL6lwRY2SAXLnoT+SZFThVcURExfct0/Hbub9GV1hZkaOKmYPeWPnPhYldBURU8+OWD2NCxYXc3ZY+EpEh9SsIA7R40xBp6bX+0UNfbYApuL+6fv7YJ0eyR0lO0JFpw6fuX4vGFj/faPl3k4BJcF/s0qNyMVYhwXyCdVTkbuVqyW1vjjMwCubzglKQnuivruzB+eIVuf0GfJ7tfWUdWD6wuxe8vPBTT7jgF91x4CACN4JYGfagZVgFFBba2JcwEt9w6HJc3iqLSNmG/lxla9S/x4zvjhuClHx6L7x+vr6nJ58byzrM9Aa/g3nDSSPzh4sNM24wZVIbXbjke/+87Y3Vt5k2tCinJRIBHG8i6Cxe9hVgmZvtdRs7vokxIySkkJa1sV0rSxhRSahTFJbgudi9EWcTVU6/GgoYFfXqculgd3lr/Fu6Yld+87dsKSZUgKRKLDusLPL/8eZz79rmojdb2yv5Scgq/mPULrGhZ0Sv7I/RFiDKvsK7r6D33+s50JwDg66ave22fLnJwCa6LfRoUopwoQHDf/aYWFzw1D13J4lZBFYMyTApuYxdHcNvi6qR9CgAAIABJREFUOgV3RW0XpiyrZ9umJQWKoqK+M4nhnMkTABw0SMtPPWJ4pS7HdOTACDweAWOHVrAw5uuP3x+/O/9gHDAwF3IbNiiZIb+1slkZDjAyTaVtQtxv+5UEIAgCzjikWhcyDABeT3GkNi2nMWfHHMvveIJLJXiscPKBA3XnUhbyoYoj74VygUl13q/SXErJhYveQD6Cy0KUPfn7aUpKMTJM/ydlQlZ3zaKdi75BfaweK1tX5t1mVesqfNXwVVH77avcRitExShWt63OW7Lq1pm34uEvH0bNyzVojDc63nddrA63f3o7EmKCLQSp2DWRWHsbaExQ1L5b9FpQry1i1Mfqe7wPvn1bu7Zizo45uHvu3TvdNh5E8nkFd137OtS8XIOtXVt7tE9ewe1IdVhus75jPVqTrUXtl67HzgoDLqzhElwX+zQoRDmRpw5tc3cKd01ahjUN3djUYj8ptYLRHZkU3OaoRnCP2r8SW1oTuhzcW15ZjDte/wZpUYGvbAViYhSt8TREWTURrpphFZjzmwm48aSRCPtzquSgcnOo8fEHDMAPThypI4lGBRcAC23mEfB50C9LlEuzCm6EOx5vYLWzeHLRk/jFrF9gecty03eVkQD+dOU4fP67M4rK5y0N+lioNlCY4E684FAAwDEj7Um0CxeEnigjUdG6jjXgPAc3LaWRlrUoCqqdS7/ty8msi77HeW+fh2unXZt3m2umXYNbPr6lqP1Sf9kVIDKdLxx0YeNCVtP5i/ovHO/7z4v/jM9qP8Pc2rmsr3+b+vxds+/C9M3O8qjZfejDPFy/V1sU3plQaEqzAHKLFb0dWk0Lf/xCz/ub3gcAfFb7WY/2ybfRbmHxjll34LllzxW1X7oGPTEbdFEY7lV1sU+D6qkmM/YDf3M0NyFo5kKLjXhz8Q6MnDhNV8Imntbvl4yYmrL7Of6AAWiNpU21awFgbdsGhIe9hoXR51hN1iEVZkVx5MASLW+UI21WBJdQGcmF6hoVXACY9ZsJWPngebjiqGHss4DPw3JRScHlFePezFPd0r0FABATrRcTvnfMcAzrF7H8zg4+rweDOAW3UP3Zcw4bhKX3n4Prjtu/qON8WzG3di7OfvNsPLHoid3dlN2Cnkwc8yq42ZC3QsQ5KSeZckvEhSZuroK7Z6Et2aabwDvFj2b8CJ9u/7TX2pGS7d9hvQ3qg3YEV1VV1n+B4sg3heSHfeFdolDuaZi5fSZ+N8+ZEzZdn77KawVyZckSkj4H9cWVLzp+L/AKpzEipS3Z1iu1vamP8OMjjbfFOM93pbtw7+f3ojvTrSO4dte4O92N7nRx7aex3CW4fQP3qrrYp0Ehx1Z1aAm8usqHFhvx5tda7smq+twgFrcJfW7qTiPs92LUAE0tbew2v9gbY9pkKC63oqFTy7MbUmFPXCMOCW7I7wVFDduFJJcGffjz98azv70eAf2yxDhHcDkFtxcJLr2AFjQswGc7eraiaoV+nMpsReyNqIwETOHWLqyxonUFmhJNPV4B39vRk4ljPgXX6YQ9LaXNIcqqG6K8J2LC5AmYMHlC0b9b2LgQq1pX5d3mg00foD3V7mh/aWnXKbg08bdbAJJUSRdWzJPdQqDcc4/gYb/ryxzTPQnFnidd/74kuAGP9n41kri/fv1XvLL6FUf74Aku3V/qQxMmT8B5b5230+2kcZEnpfTvYmqHv7PhHUzZNAUvrHjBEcFNyamiF5dowUeAOw/pC7gE18U+jTTnamyHFGcA1RRNozmawsiJ0zBtud4xkMhnQ5c2MC/b0WmZs1sXrcOmrg0YVB5kSmJbLI2SgBd3nHkgbjtDK4vTnlWCVQD1WWKdLyeUJ6ulBRRKqvtrFaJsBxaiHMyZTBm/s8OSP5yDxfed7eg4tGr50sqXcPus2x23rxB4shop4rxdFAatgNOk5NuGnkwcneTgFiKpKTmVC1E2KB6uydSeh54SjEJhmr///PeYOHeio33tUgW3QIiysfxJMeGozFxNTuX6/G5yDhdlEctalu2y4xXbj2h7URHx3sb3EBfjvd4mClHuyhQfpUDg81fp/vLnahfVVQxYiDKv4Gb7XcDrfKG+f0hLX6qN1ur6sdW9kRUZoiKyqAOnoLHdVXD7Bu5VdbFPQ8qaQOUzmeIV3KbuFOasawEAvLpgG2aubmLfUSmdNQ1RfLW5DZc+Mx8/+I/ZAOT8d87H19J9qC4PIRLUiFZbLIOSoA+/OvdgnHygVuO2Pa4NbqoK1HcmEfJ7dOHFRhTKK+VB5XqcKJkEOjbl4ErZ8kWHDSlHwJd/qOhfEmBligqhN5Wnj+86DW9ma/smxAS8kU0AwMosuegd0ASh2Bf4vgKnE3Neeck3WSOyWkjB1ZlMKa7J1L4KJ8pmW6rN0b52aQ6uRb4jD+N59SREOSWl2PO3uxTcJxY/ge9P/z42d23eJcejZ90paEz4sv5L/GH+H/DnxX/u9TbRte9JGD6BHxOtCC4PURF7lJ9LC398nySCWoyCSwaADfGGgjm41K/5xaX2VHveRU7AuZu+i57Bvaou9mmIWZKWFGXblyOV6ykN+tDcncbqbAjyl5vbcMsri/HFRi2sxufVFMIX52/B1f/SHAWjafuV1v0qwyjJhvm2xdOMbJKDb0eCwq6AdY1RDOsXyeum5y+CtDlVcM8fO5j9m1TasqyCO354JQDg4e8e7vi4TtCbeVQHDSrDsVmjqPu/uB+REf+G4Ovstf270ECTkN1BcFe2rsS0zdN2+XF5OFVU+FDNaKawyZTVZIl/PlJyik12aTLE6uDuwnxEVVUxae2knZrc7ouYWzu3V/IGnUzkwz5nju99+YzWRmt16iCvHFrBSNScEPmudBe6M92MACWlZI7g7iYX5bXtawHYO+j2NooJ5QZy40hdrA5A7yihRtD92Jn+zo+JxhBlI05+/WSc/aazqDAe+RTcYgguhfrXxeoKhigzgss9e6dPOh2nTjo1/zEoRNl1Ue4TuATXxV6PaErErLVNps9VVYUoqwj6PFBVfSgyD/p8xIAIGrtTWFWvn8T9fdZGADkDKaeYcHAVSrIKbmssg1CWdFJZn7as8VQyo+Dzja24qGZIUfvPB1Jw7XJwCc/94GhsffwiADkjKVJwDx1Sji3/dyGOHtGv19oF9J3ytKFjAwBg/Aj7/GQXPQO94DNKZpeWIQGAa6ddi4nzJu7W/DunBJcPZXMSomxFUvljpSSLEGXKt+tDx1QjamO1eOSrRzBr+6xddsw9HU3xJtz26W24Z949O70vJ4TGjuDe8OEN+P7077O/+1LBveCdC3DTRzexvwvlfvZEwb1n3j24d969TA1LSkn2XNkt6mzo2ICNHRt1n3WkOpAQExBlEdu6txU8bj5QjiSNQYsbF6Pm5Ro0xc3zjt5AsQSXrj/laZcHynu9TWQuZbfI5eS9wBPvhKjtT4VqObYnpaTjvHMe1EescnCLAfW/9lS77n44VXAB7b4YF0U+r/uclcyiBaBCbvouegaX4LrY6/GH91bi5v8uNpX4ofDk8qxiamc0RSHKIwZE0NSdwqYWff5KfTbntliCe/ahg5iCC+TqyjIFN6kNiqKstfP64ws7+t5+xoF48YfHFNyOKbhFhCjvVxmCIAADuFDjvlhZ7CuCRJOQJ64c1yf7/zZDR7p2YY4fj8707lPmHRNcbiJldBu12s5qv/w+0nLaFKJMcKrgNsYbUfNyDZY2L3W0vRVoMtrbJT12F9qSbY7I1oaODfjZJz+z3JbUqB3RHbrPe3KNnISkhnzWC3ffNH+jyw/t6xDlNe1r2L8LlacxnpeT3NBt3duwLbrNUsG16/OXT7kcl025TPfZLR/fgmeWPoMHv3wQF7978U4pj/QeJAX55dUvA0DBOsY9RU9DlIlMOSW4KSnlWPGnMcDOKdiJP4OVggsUf775YBU279TzgAf/HPHvHqsxm66hlcHb7B2zdX+/uvpVAFrdXBrbXZOpvoFLcF3s9djRoQ2UzQanYilLHMuzimTCplQQEdzqshBiacm0XTqr8Gay4c7P/+BoHLl/Jf569XjddpURP1NOAa1UDV+uJpz9riTghdcjQM4OlCqAoM+D6jzOyITfnHcwzjxkUMHtWA5uEWZL5xw2GFN/cUpeo6veQF8puMZJiIveAz9p39VGUz5Be4a2R7fv0uPyqIvV4Y5ZdxScJPOTn3zXKZ8ixavAfA6ukbg4fY6+rP8SAPDm+jctv29JtBTMFSOyvq8Q3AmTJ+AnH/+k4HaLGhdhfv18S6WOFnrIXZZQbKkQwNkCitNFFpps03PTW7A6fqHyNEaTKZ7gKqpi2e/aUm1oiDXkTKakXJi+UenLR85ak61oTbYy53ciaD2BUcGl8ygJmGvK9waM160QaIGBCG6J31m7Tnz9RJzz1jm6z5JSEosaF6Hm5Rqdu3ehEOV8C3oEnYLLbd+bYfU0pvJ90mjQ5wQ8WeX7Tt4QZW7xlxYZjGMHkXy/x++aTPUxdstVFQThCUEQ1gqCsFwQhHcFQajkvrtHEISNgiCsEwRh5z3DXezzINffJkMNWyKkTMG1MZpKZ5XZyojfMpSZcnQzkor9KsM4b+xgvHvrybjsyGG67UoCPhOhpBBlIEc2BUFAZdgPCHQcAWWh3p2MkIKrFBHW6fUIGDu0olfbYcTqttV9rsR9W0pJ7AwWNS7CjK0zHG/vlLj1BQaVaAs6RqVsZ7GjewfLrSuEF1a+gNk7ZuPdDe/m3Y4ngPkmbfkUBd0+LFyUCU5dlOkYdjUgz3zzTHz3/e/m3QdN8PqyDEmxaE224pvmb4r+HY0PS5qXFNyWHGPzKbgBb0A35uTLvbaDXUhqxJerB+6UoFFbySSnt2B1DfKZTImKaApn5QntU0uewomvn6gjvUkpibgY1xGFlJzKLQhxLsr1sXoc+9qx+Nfyf7HP+H2RSRG12wkBs4Nx8ZTOI+Ttm3QYp4qmqqqYvX02I2NkROZ0kVdSJNP7+MYPb8TNM24GoIXTEvgQ5SNfORL/XPpP3e+cvBdimRhbeOG37813Sr4cXFYjV5Hx4soX8x6X74N837FUcOWcIRqBnmnjoiAtECSkhGsy1cfYXVf1EwCHq6o6DsB6APcAgCAIhwG4BsBYAOcDeFYQ3OB0F/lBBLeuUz9YkQswhQTHbAyhUqIMj5DbzghewTW6CfPOweGAFzXDKnXfh/1eUJQvH/pbEfYDQm4ALgs5Nz9wgupyZ47GfYXpm6fjhRUv6D6TFRlXT71aVwsP6D1CSpOQPWkSvqfi5hk34zef/cbx9k6Jmx3akm1sEvzm+jd15Przus8xt3au7W+rI9UAzAS3PdWOmpdr8MGmD4puDwBc+O6FuOqDqxxtS6vxzYnmvNs5VbppAmup4HL72NS5iU2ujCTIqYJLx8g3iWpK5FSGWCZmIlPUhj3p2Tr/7fNxw4c3FP27YkLsiaBZEQ6e4PIhuj0JhTXe2x3dO6Coii43r5BxEI2jdH52Cxo9hdVzn0/BvXPWnfjRxz/Sfcafw1vr39I+40hvW9LsFK0zmeLeFdRn//7N39lntdFaXdsycoaR495QcNe2r8W69nXs3vdVNJJVfyCPCR7TtkzDHbPvYNEtlLO6M5EWZFQF6BcUaDxrTbZCUiU8u+xZ3f1wcn2jmSgqQ5W6/QF9o+BaEVz67KOtH+GvX/9V13eM4Bd0+HOzzMGVcgquqmo5xfR7W4Ir5giuazLVN9gtBFdV1Y9Vlb0RFgAgKexSAG+oqppWVXULgI0AjtsdbXSx94CUWiPBpdxWKu/TFrNeFU2JMkJ+r20ZnrSkOTBnJBkBg5Px9DtPwV++p4Uqh/we/O3qI3TfC4LAarLyZLg87IdABFftfQX30e/WYOIFh+C4rMPwrsaMrTPYBIZgVz/PKn9rR3fxah1NQuxe7mQMYjQjsUNXugs1L9dgyqYpRbdlX8POKLgJMYEJkyfg8YWPAwAe+vIhHbn++cyf47ZPb7P9Pd1XfvIK5AjvG2vfKKo9+fCPb/6BZ5c+C0A/mSaC25Joyfv7QtdpbfvarPmd1kcb4g14ZMEjtoYoUzdPZQY5PSW4tJ1TI5MTXz8RF7xzge6zQiU9AGDmtplojDc6OoYTbO7arAuRNMJJrumixkWYvG6y7jNe5SsUmk0E12getqZtTY7gegK673tCcPl73pZsw3fe+w4+3f6p7vNC+at0j2iyXYxjrBNYXW9GcC3G8Hl180yf8QSXSA3/nBgXP30en20OrhUp4hfBRFlTcKltcTGORxY80iPVnwjIk4ufxJUfXMnuRV/5SfDXWlIkXPjuhbh8yuWm7Yx59Yzg5glxbku24bL3L7MkzIBeraQxUFVVJMUk/B4/W+zxe/xFp67ExTj6hfqZtu8LBZeHqYZ4th/lM7Hi+5eVc7hu2+yikqIqkBQJkiIxFd04H6FxIy7Gdfe5NdmKLV1b8pyZi2KxJ+jiNwP4MPvv/QDwM9va7GcuXNiiO6kNIPUmgqsNYkOzOaUtMesJUUpUEPJ7bR2HFVUzrMpIZgW3uiyE0w6qAqCptf1LzIXE5exLYmBp7rsBJQEANBALTIXuLVRE/PjZ6aPh8eyelUFZlU0hYe1J65eJceL+ed3nuPDdC/HR1o+KOmYhgktmD3zYVT7QZP0/K/5TVDv2ReyMgkuTgzfW5SeicTGOfy79J5sg3P7p7VjavJT1D2MuExG2YlSU5kSzzpnYiOeXP49/Lvsn5uyYo5tMUxtakmaC+/b6t/Hptk8B5FdwV7WtwlUfXIUVrSt0k6RJ6yZhUcMi9rfd5DSjZHSk26nJFE3Ci3HqNE78WIiyjZmQpEi4a85duPHDGx0foxAufe9SXDPtGsvveCKd7zrcPONmPLzgYd1nvBpjnFAaFUSm4Gbvf3uqHT+f+XN8b+r32IJdwBvQ3Xe7HNypm6filVWvWH7HK8Rd6S7Iqoz6WD0ySgYnDDkBp+53akGC253pxvqO9WyBxk7BlRW5R0ZUvPJN/ZCePaeKYTyTOwc6Z/45MV7/URWjekRwFVWBpEq6d0t3phuT1k3qkepvNAGKihpJsSI7Gzs2YmPHRtz7+b2OF1ON4MeAfORve7e1L0E+h/UNnRuwsXMjy03WHVcRdefEOxJLqoQR5SPYd5XBStswXjvExBj6B7VFd/45dGI+RigU8ZXPmZ6+C3o1scHKFIrAPyM6gmtxbfltk3JS9zd/LyVFYn05LsZ1xPvMyWfikvcusWxLR6oDf1n8F9296Up3YV6teRHJRQ59RnAFQZgpCMJKi/8u5ba5F4AE4LUe7P8ngiAsFgRhcUtL/lV1F3sX5q5vwXOfbXK8fVeW4NZ1WBPcwRWagms0oSKkRBkhnweRgD3JTImyZYgykCvFY0eQU6KC8P7P49O2XPH1wRUhLgcXva7g7m5IimR6admtlhonW+s71gMAVrYU51BJq+x2+WxlgTIAuclJIdCLqCc5dfkgKRL+8vVf9oiaohMmTXCUiysqIpsU0KTmvs/vw3GvFQ6wcbo6//SSp/HssmcxY+sM1Mfq8VntZ5g4byKbCFN+GYFCbmVVxrsb3sVzy54reIyz3jwL10+/3nZiOLhEqwu9vGW5bqLEh+cZ8cCXD+CXc34JINdnAp6A6bxp8t6WbDMRApo0zdo+y7Z/NsYbsbQlp9jIqoyklMR1067Lq3QSCbEKUc43WayN1rLvC4Uo07nWx+tt99eb4K+DE3LFj0X8v7dFc+Vj5tbOxYTJE/BF/RdsTCISm1EyqI/V4/RJp7Nj04KdV/Dqxhy78WLqpql4a8Nblt/RJPiJRU/gicVPaPtPtUNRFRw16CiM6TfGkgTwfTSaieLWmbey/uMVvGhNtpoiUH792a9xzKuFnfiN4MmAUcE09gu7RTCrvs0TI+MzXuIr0Rmt8SGzVqHm8+vn69rDLxzsjPeDkeDmU64vm3IZLptyGaZsmoLLplzWo9Bovt38OKKqKtZ3rMdra17D7+f93tZ4L1+kBfVbes/Sfo3HAnL9iz7fvyxX6aEiWGFrxGSHaCZqqeA6vTdd6S6Me2Uc3lj7Bn4959e4+N2LTdvY5YMDwJRNU/DHhX9kjuT50hV0IcrZPipAyOuiTP/WEVxufOJdt3kFV1TEvHnTf1z0R7y06iVdKs9dc+7CrZ/eqs0jGlcA6d6vfby3o88IrqqqZ6uqerjFf+8DgCAIPwRwMYDr1dxbtg7AcG43w7KfWe3/X6qqHqOq6jFVVVV9dRoudgNueHEhHv+wsPmLJCs45y+fYW2j9tI0mkxRmaCSgA/9In40R60Hs5Sk5A1RBjQjqoykmEKUASDks645yw+0vpItWNmVG5yGVoYBgQZKAaXB3g0n292QFAlpOa0b3O0IrpGQerLDEj+ZscN/VvwH98+/H0BhBZcRXIeElbbriStqPszcNhMvrXwJf17858Ib9zHaUm2OwqIkRWLXj17272963xF5dUpwO9KaA6iiKmwiHPFHdOoZD15Fmr5lOt7f+H7e/dNkbX3Hemzqsl5A4yeufP+jc7ZScAk1L9fgk62fAADKg+WmyRNf9sSqXmJjvBF3zr6T5RSftf9Zum2mbJqiU58URUFDvAErWldgResK23bReVspuPmUngveuQCvrNYUx0ImU1bkKyP3Xc1kPlzdiRrJK758W/lJOUV2/PSTn+KYV4+Bqqo6Bdd471tT2mJHWtGPc3YhynExrlMwedDvX1n9CmsHLaYEPAGU+kshKmLeurLRTFSn2iqqgqs/uBr3fn4vdkR3QFIkPL/seXy6/VPLNhQCfywi/tR/jPfZru5sUkqaFDb+HhgXkEK+EFJSyjIH10iiw74wvmr4Cmva1rDteQJmtaC4qXOTo5Byu/x1J/27J+Z4/H3mx8+Pt32MK6ZcgccXPo4PNn+gy5flkS9EmcZRPkSZqemifqymxbHEds2JfUSFvYJbaJxXVdU2RNlpvdvNXZsBAO9seAcfb/vY1M9ak62mRRIgdz1WtK7Aq2teZYuN+aKRrEKUw76w5T3nn420lLYluPfMuwdDSoYA0N4pjOBy90uURdMzQuMGv9CyrUs792S6G/jP2cDX/7U9l28rdpeL8vkA7gZwiaqq/LLPFADXCIIQFARhFIAxABbujja62D1I2dSqtUJ7IoMNzblVq+6UhPe+qcOO9qwpS9Yd2ecVUF0WQnM0N+jM29CCk/7vUyQzMlKijGCeEGUgS3Bl1VLB9Xk98HkEM8HNEzo5pCIEgXNRDgf2hGwBDU3xJjTEGnZqHzSo85MXqxcPYCa4zLHSgfnU8pblLKeKfmdHcMmRtFDeHYFear1Zow/IrRr3hkHJjugOU46hHe6afZclCXRCQEVZZAS32BBlXb1DG3WdjgFo4WNU7iLii7D72ZHqsK5tqMjoTHfa5ngT+Pw/OzJME2JJkXTHonOIi/G8k9p3N2ouy+WBctN11RFcwzWUVZlNtGmSf80h1+DQ/ofaHktSJTYhzRceyHJwLVx1C5Ujob5VKAfX6vhXTLkCL658EW+sfQPHvXZcr7qb8+Hq+foUlfBpiOfGM76t+ZTX1W2rdSZTxpBfuk9pKa3bj124ZVyKIy5Zf2d1DpTv7ff4EfFrY9fbG97G88ueZ9vwBK47041BkVwJuZZkC5qTmina8pblmL1jNv6x9B/s+2LvB09mPtzyIVRVtVUy7QguYB6f+fuxrXsbqiPVqBlYg4nHTUTYF9aFKMuqjL99/Tdk5Izp+bpktBbeubpttWVKgdVC5Xff/y5+MP0H9idNsMn0cTKGf7jlw6Lzfvl3Tr6QejvkW7giMrmxMxc+fc+8e3DTRzfhrjl36bZNSklAlpCYfD0AYERZjuCWBcp0/bZQiHJSSkJWZfQLmgkujfcEu7SD+pgWIRLw6lPBtnZtxf3z78cZk8/AU0ueYp9THzf2uQ82a4uI+RbHrEKUQ74Q6/OxTIz926j28gs1/LGbEk24cNSFGBQZhLgYz9VD5+7XOW+dY3o/iaq2HT8G0b9TqQ5ASgE2KWDfZuyuWfU/AJQB+EQQhKWCIDwHAKqqrgIwGcBqAB8BuE1V+8imzsUeiSXbcgMduSDz2Ngcw/GPzURjVwqdidzAQQZOv5y0FFc9p602koIb8HpQVRZEC0dwH522BvVdKWxujWVNpjx5FdyUKFvm4BJCfi+rc8vanydMaEhFOOeirAIBb9+bhYuKaBliacTZb52Nc98+19E+7cgOb+xBsHLIBMwvGVotd5JjyOe0kPJbaNJuDJNrTbbqwrUIPCHqzck59Qu7nMhVratsV+aNoBxDJyrWzO0zcd/8+0yfOwktk1TJluAWuk/8xMdoFMWDziHgDbBJWMSXU3BVqLpQNj4vrz3Vjlgmlpd88hNcKg9krBWalHNEzi7fL99EjhZPygJlpuvEm+oY75ekSOw+EKnye/wI++xrUiuqwp6vfDlsdJ2s+hvfjt/P+71JzaIQyEJ1cI3KT0pKYWv3VqxpX4PHvnrMUrV2Cqtj8q7P+QguqUU8weWvFX/+xnOfsXUG+0yURdO4QuNZRs7o2mjXPxJiAnExbuucbTwPUoj9Xj9K/aUAgMe+ekxHUvlrGs1EURWxjmZb3rKcjY+EYhfueDL91JKnsLp9NSN4xnddvpDTtJzWjTmr2lahIdYAVVXxddPXOLL6SPzvov/h+kOvR8gXwrqOdZi6aSrb/oWVL+DRrx419acBoQEA9K7LvDLIt4nvA6QK5oMxRJmQj0jy7S0279cuB9dI7OyQN0TZQi39ZNsnWNO+xhQJkhATgCIh7tH6zsDwQLZQrEK1VXATYsI0xtGzZKXgTt8y3VH7KbXE+H58+pun2eIiD2qfcQxZ3rJc970VUnKKPXcJKQG/xw+f4GN9/sTXT8TEeRO1bbmx/v8W/h+un349+5u/l7Iiwzv/aZTCqw9R5rZpS7WZUj0k2TxnIIKbSGWKmk09AAAgAElEQVTHLbEH42smDmz4pPjf7SXYXS7KB6qqOlxV1SOy//2M++5RVVVHq6p6sKqqH+bbj4s9H10JEdvanBsIbGzhCoGLMlpjaTw5Yx3S2Vq0m1piaOpOY3t7Au3x3At61MBcvcCmbCgyEWRNwdUTXG/WfCklKtkcXK+uhq2R7KZFxdJFmXDUiH44fD99Ddncy89MjoZWhrgyQQL8vr43g3rgiwdwxuQzHBuCFCI+27q34eTXT8YXdV+YvqMXFD+RcBqiTBNAYw7kh1vMwwFPcAspuPS5UcG96J2LcMWUK0zbO2l7Iby74V0sblys+4xImJ0JzDXTrsH5b5/vaP+dKW3SVojU55vwODEH4RVco3JSSNHlt+eJiRE04fYIHpYnVuIvQUbJoH9IMybh7wOds6RI6Eh1QIWat5QKr9BRmyU1R2RlRdaVPeHVGf4c+H5hJNR0DmWBMqTltKUKnJSSpmsoKRK7DxSq7ff4Wa6YFf635n+sBmi+Z5WIST53UUBTNazy3lNSqmCIsrEPUTmlpngTm5T1tEyL1e/4kON8izvlQc39mo9I4fenMyEyKHxfN33N+kZtrBa1Mf3iDC0WpuSUreL29JKnsbBBC0KjfmMVMWEZAp3IElyPHyX+Esvz45+97ky3icQC2iLR8pbl8Hv1aTDFGk0ZyUBjvNE2BzffokNGzrA+DgAvrXwJ5759LupidWhKNOHoQUez72iBx5hrOrd2rmncofIzCSlh+Q7gQ5S3dm/VjSWLGheZtudhR3D7KgTfLkTZKd7b+B5qXq6xXJi1i6Q6bb/TTJ8lJI3gdmcJbkWwgi2iJMSEri746rbVOPKVI7GlawvOfutsXDvtWt2+/rvqvwCAQ/ofAgGC7ryMiwxW/WdZyzK8uuZVAOZQdvJOMIIiR+xC+/OaTEnpXFqOqBFcr8cLURFZ3yP/Cv5Z+rrpa91++IVYBQp8ioiSdMyUg8vD2F76nt+OInISmezCTU/G12m/Bl67Emh2Vg9+b8OeExfpYp/E+U/NxelPzHG8PYUVA0AiLeOVL7biH7M34tUF2gsunf1+3oYW/H8LcmFQfAmekqxZFJUP8nk8qCoPoq4ziVX12kvOlyWqHfFM1kVZr+CWZ+vS+r1C9rj2JlMA8MrNx+GGE0fqPqOX35hB5knq4IpQrkwQgKANce5N0GDstJYlb4hghU+2fYKMksHM7TNN3xVFcBXrlw9vuvDYV4/h7rl3Y0WLfoVZVnOkpFAOLgstEmNoTbYyomtH8HginI+YzdkxBzUv11g6sN7/xf24acZN+nZkFz6sCG6xSjEz1iqgxvD3wXgMJxMoSZVQ5rdWcAsRZH77vCFhXHgw9RWP4IEoiyz0UkdwyQRMjLJ/5zPu4hU6frJOkwljaQ5+gcXO8dPu2jG120LhsFJw03KaqaC8gkvGXoRT9zuV/bs2VouvGr/S2iclMK92HjMU4ssZ2U2iAHO/sQqhbYw3FjSZMpJQIriN8UYWkeFkIQUAZm+frXv2rK5xU6KJLXrM2DqDKTt/XPhH/PSTn7Lt+HJMBDsF1xiizBPaZ5c+i3vm3aP7nhTBtJy2VNya4k3494p/45ezf6k7rlWKREbJmEpQ8QsdPMHlow6M7bdSpEZVjEJLssV0/fNN7q1Ax6Lr3pxozoVYGk2mCpj3GENSAY3AAMBR1Uexz+zeGd3pblO/KPWXwu/xIyEmLBf8+BSGRxc8iqum5Gpg3zzj5rwk165OqaRIuH/+/Xh51cu2v+0J7EymiiW7Vu8Fu2s6qmKU6TNScInglgfKUTOwBgCwsHEhJq2bxLadsXUGJFXC62tfRzQTxYaODbox950N7+DiAy7GEdVHwOfxsb5z0tCTdMectnma5XneOetOy3xpUcktvlohKSVt34/5FmfTchqlAU3BjYtx+L1++DyagstHA9w//37L/sy3D+AWtlUgEijDxs6NLA/aODbb/c0/7zkFN9uv85yLpEiYOG+iOVKtLetFkbRv/94Ml+C66FM0dBUXNpHhwpITGQmRbPmcGSu1FXsiwH+ftRHTlucmLVbqq5StgxvwCTh9TBXKQz7c8bqWC+PLKrhPzFiHFXVdWpkgnuCGteOWZYluSrQ3mbIDvfRfvHmc6bugzwu+TJAdce5N0IS9EMHdr1SrzEUTDiO60l2ojdYyR78v6780bUODuROCe//8+3V1b+nlxq9iburUBmKjCiUpUm4yk52D2BFc+m00E8Vtn96my9Xh20zg2/7s0mfx449/bElAX1r5EgBzftSs7bMA5ELn+DYD1iGjxU5gaB/5FBMgfx0/pwpuwBtA0BssqOC2JFpQ83INPt76MQB79dMIUl8lRWIqQ1pOI6Nk2Ap9W7INS5qW4Pllz1uSWqcEl78GtB9+Up5PweVVYrtrR3Vz+d/RderOdENRFZ0qlJSSOQU3xSm4Xv3i2BVjrsAzZz1jOl5cjOPWT2/FvZ/fi8/rPseZb57JSkjQeVk998Z+Y5WnmJJTjHzbhWTaKbgtyRa2UOVEwW2MN+KO2XfgzDfP1J2bsc3tqXbm6PrM0mdw0bsXAQBeXfMqvqj/QrctoM/xIxIf9oXZhFFRFdPk2WnUhtFMj67FgoYFADTiQDVZ+ePzkBSJXTMj/B4/jhl0DB486UH8ZNxPIKm5MY/vX9FMlJ3PsNJh7PNSf6mWs2oIIzcusmTkDC57/zLbMmq0/aSLJ8En+NAUb8qFKKsSYpkYU9atxiPq73YElz7j1Ti7VI2MkkFnulMXwh/0BhH2hR0ruJSfTODv99TNU3Hsq8eysk62Cq4q492N7+LJxU/2ahqLnYJrvG4hb4jlmVvBylCRX4g9fsjx7N9WBDcuxgFFQhcR3GA5/u/U/8Mxg+xduPmFgtnbtdJ8iqogKSUxvEzzkKUazQIE3HeCPm1m4ryJ+NuSv5n2WxGsMH0GaNcqH1HNZwyYVvLn4NKiLlNwBa9uARbQfBfsSuAFPAGT+uqDihLBi+ZEM+vfxrHZuEDDu4JLioRHFzzK0n0SdI9F+7nDho4NmLZ5mmmRDhTV4TCqb2+DS3Bd7FEQpdxLIpGRWY3bJds7cOOLC/HGQrMt/g9PGok7zhrD/i7JkmKRU3BPOnAgfnjyKGxpjSMtyYzgrmvSBgcthzZHOCrC2oNP5XvSUv4cXCvIqozFjYtx1TTr2ma3nZl7oewKgmtnuGAEGZrYOc1+f/r3ccE7F2B5y3L0D/VHbayWmT8QrHJw7cjN1u6t+MMXf2B/08uKnzQTqTDmIOlClEnBzb4clrUswxmTz2ATG3pJRDNRtKfadWFyxuPxxwSAz2o/w4KGBbrJs3E7YxghEf/RlaNNbQasFdxiQ6FJHStEcPlzMbn75nkxEiRFYiGzduZJhMVNWkg2uQE7JbhEMPgJRFJKQlIk5jzZnmrHjR/diH8s/Yf1JDaP0ZSdezZdO36SZFRwdeeQca7g6vLSDASW7y98GDARQr/Xb+rvlaFKk+IB6K8rLUz9Z8V/8MqqV5hSRzn4FN7XlmzDnB1zdPuxUkhSUoqdhyiLqI3W6hxYAXPYL5E1FSq7T04Wb+j4xnw+HvPqNOJOk2UrsHua7et83n1CTCDkDSHiizDSdtNHN+WdCOdDWsopuD6Pj7WXCO7IipH6cdDCSdkqRJkQ8Abg9/px+ZjLmUEP9SWjuU1KSuHYwcfiglEXsM9L/CVahICUn+C2JluxsXOjKUqGP09AWxioilShKdHEFgWb4k048fUTcd206wBYK7ikhomyaNnPaLLPPxcPn/ywaTtCc6JZR3qC3iAi/oim4BrGhjJ/mY7gWqnovFPy1E1TkZJTeGHlCwDsFVx+3F3VZl+qq1jw++Xvk5HgDi4ZbGkeR7Aab/l3zPkjc6kwVmG+MTEGRc6gO7uwXxHQrrddeg2gN68iszE6B4pIoXB5n8eHoSVDTfvgy56xOt4eLyYMm4Afjv2hblt+fLKC3cIRkOvTk9dNxmtrXtO7IctptlCZUTJaDq7HB0mRWGpQIZQGSk1GUl4V8BgWQ4zPA99//7Loz1jTvkZrh5zBF/Vf4I11b7BzTmSi2O7zoSa5hOUWG0HbmjwdiOAWSHHaW+ESXBd7BNY0dKOuM8lIKQDE0xI6kxQCpeKz9S1YvM288vvAJWMxrF/uwSWVVZRpoqj9fcDAEigqsL0tAVnRDzApUYbf62EhyRSiXBokgqsUTXAlRcKfFv3JdlKvkoIrKBhcYW8m01ugiXMhBZe+N5JWwtburQA0An/KfqcAgGnCaxWinC88lcj3uvZ1TAm1IsemlU6uML0xRHlL1xatbEB2xZoPUSbTmHzup1b5nO9vMrvv0u+MpSRoEmrlmAtoL/cbP7wRt8y4hX1XbK1GK4Lbnek23TueJBjb48hFWRHh8/iYq2m+39NElSZMduonoA+XJgIqqTmCS58NDA9EwBPQhYoXUmmMsCv3ZKngqnqCyyuXvAJnp0rSxIi/1nQd6B6TWQsALG9drqvtCljn4FYGKy2Vf74d1M+XNC/BE4ufyBmtyCLum38f7v38XgDATz75iUkpsVoESEgJXYjyBe9cgMunXG7aht+HVUi/EwXXSonh9y0rMibOnYj9SvfDhOETdNvxYwO5LNOYwxOauBhHxB9B0BtEWk5DURUsaV5SsG22beZycCsCFew+03Mgq7Kuz1w3/Tp8tuMz3T5kVbatzUxqF5CboNIxdARXTCAjZxD0BnUEpDRQauk6bCShNMle1roM/1z6T5MiSdsHvUEMigxCc6KZXXP6LZF0qwU3UsPSctqS4K7rWIewL6xr+9gBY/GbY36j247G+aZ4E0LeELsmQV8QEV/EUsEdEB6gGxusao6ylBUxgYWNWt50Z7oTsiLbKrj8IulXDV/Zqrh26qMd+JBa/h4b3w9DSobYtg0wj7e00PHz8T/H/y78H04YcgL7jvLVeWzs3IhTp16BLo8HYUVhxJTvk0AubJ3QL9gPZf4y9m4kIknjGf3e5/FZEnR+/KD3ZTwTR3mwHJVBLdeaf/elpBQGRQbh3BFmc0xj6D8P6tMPL3gYjy98nEVdAdpYRIsy1GafxwdZkdGeNi9EWynpJf4StvglS9lFMFXFhox+Hms1ryG8tPq/7N8ZOYOlzfr3RCITw7yIdl2nbs6asa18Oxd+DBuCu/DfQFc2QsIluC5c9B0ueGoeTn58lo7gJkQZXQlnD56PCx2OZ7TBgvZFpPWAKm1leHNrHNGUfkChWrphvxd+r4BwNlyZFNyUKEO0KRNkB0mR8rqg0qA2ZlAE3xk3xPF+ewqnIco0IDtx8iU1iV+1BXIDtI7g5sn5oknNlR9cifn18wFY5z0a2y6pEiMjxnxUPjToi/ovGDGncClJlXRmFXExDlEWdaYwxpcWX56EwEJrVeuXlFEZ5kOUlzQvYXmUQM8VXD7U6tL3LsV5b59n2UbAPLF1FKKsiPB7/KgMVpqUb3p5bujYgPc2vmfK88mn4FqFAZNpFJCbOAe8AYysGMlC1QHrSXS+mpbdmW5LcrihYwMyckbXP0XFXIuQUCg/FLBWcIm8MYdof47gztkxJzc5ycLvsVBwg5UQBME0qTWSQB58Dm59rJ4911bO4XYKLjOZsgtRFvUEtznRbHKodtLPrK5nQ7yBnVNMjCElp/D9Q7+PYWXDdNvxpWko59aO4Jb4SxDwBli4MwCcNsxssuMESSmJX8z6BQCNyBjJpyiLJnL/p0V/Mu1n8nrrcl9WBJeuJR3LJ/g0BVdOIeQN6fp5ib8EkipZhnrzIAI4v24+nl32rEn54lW4QSWD0JRoMvUHNh7JaVN4PZGFtJy2XEhZ37GeudbyMEbG7F+uhaY3J5oR9oXZcUiVT0pJU4hn/1D/gj4F1Pf//s3fISoizht5HhRVQUe6w5IQA3pFddK6SSa/hdcveh1l/jLHvheEjJxhzzh/n/ix97pDrsOPx/3YtkYvAKxsWalrI133/qH+qKmq0V1bWpQzoluMod3rRbmSGw+NCm51pFr392nDTkNJoIT1OXrnUP+lPm0kygT+fTWvbh5EWURciiPii7B7QdEMaTmNlJQyLY5c3a2dq52plhVWta7CW+vfYvvlc3vJRVlURcsQe6v+VeovRVuyDZPXTYaUvRY+ACM99uaBgL1pZFpOmxbjElIcUrav+Dw+QFGAt24G/p1L86D7zuajYgqY/hugLStMFIgA21vhElwXuwS0stkcTWHBZvsBR5eDm5bRmcxg3LAKFlLsBEReJYUIrtbNRw3MEtyWOLpT+gFkULkWOhMOaPVwSQWmHNy0qGgmU0Xk4MqqjLBfT3D5FV4iQBVhj20IFOHpJU+j5uUax8e2QiEFN5aJ4fW1r7N2tSRaLAda3vjmwMoDUR2u1hEP/hg6Qx7ZXin0eXymyS0/Iab9GFfmeddbIiS8uy79/5559zADHgBsEsSHBcbEGP625G/40cc/wqrWVYhlYugf1q9MW4U7UduMxIKOb+WYC5idIAH9hMlJmSSa3PD3yWq/OoJrVHCLCFEeXDJYZ9jD7+/yKZfjD/P/gPXtGnGicOF84aZWKqykSGyyzRPc0ZWjdf3MigwVysEtC5SZJme3fnqrqdSSpEgmtYQmO3yftiNtVgTXpOByBNcKVjm4pAQZVQ+dgmsgHfzCQXuqnTlUW8GK4PL5wXZjB69QRjNRtCZbMbxcH0LshOBaKbh/mP8HPL/8ebZvQCNLxsWnDR0bGLEjUkztbU42o+blGsyvm4+WZAtK/CVMwaVFq8vHXM6UKJ6cG+8BD1IlCRXBCnaeRExERTSRSyt3bEmRWC1XHrz7sZ2C2z/UH0lRMy8LeAO6/kGk0difjdfaeO+NucJpKY2gNwhBEFAdqUZTvMnUHxRVYYtF5GpsbEdGzqA73W26rl3pLku3aCPpPaT/IQC0sPOQL8SuZcAbsA1RHhDW+yBY4X9r/odHFzyKV9e8iqsPvhrnjdQWCduSbbZuyfyCZEO8Qeeg6xE8OHzg4bj6kKsLOlYrqoKrPrgK0zZPA6BdI1oQ4Bck+dDYKw66AscOPtY0d+Cv1yNfPYLHvnqM/U3PD41P/BhE4cdWWB0IoEJ2TnAvHn0xSnw5gkv9le45/Z7+P+2yaTg/Zh3p9svZv8SjXz3KFqZoDkVpK6RKh3whXbt+2KX1Z7vyhAT+3r68+mU8+OWDrGYvfy39Xs1FWVZkE8HtH+pvmTpSGihFc7IZDy94GCuzqSM+VcWjodF49cJXbdtkNKbiPzfOtRJiAlK2C/gEX84wiusr9F5kBNdIaB3MAfZGuATXxS4BhQtf9dyXuOZfC3RET+HChdOS3mSqMyFiYKkPI6sK14m94UStCHksLUFVVZbP68squGUhP6rKgtjSGmO5vQDw1DVH4JnrNOfGSMCXVXGzBDcbohxNay/yYhVcPgwR0E8+aXLgpFj8v1f8G0Dv1GO1m6TeN/8+PPbVY2hKNMHv8UOFipnbZ+KPC/+oOy7VsQOAQZFBGvHoKkxw01LattyFV/BiR3SH7rO17WuxoGEBU1QBe4LLhyob/8/b+hMoN5BKcQBaCBSVK3hkwSNY0rzEFHrVnGg23QO7+2iXe0jbrW5bDSN0Zjh58lUJVBaEJtN2iiKf92ecbDlVcH0eH4aUDNGVaLE6Ji0C8PmUtChiPCcrgssTKpqQBTwBHFh5oK4+IE/qAp4AIr5IQYJbHii3VA2+avhKN5H8vO5zXD31at02FBrnxEXZKkSZ9m+Vg2sFv9ePs/Y/S/cZTeCMSnS+SAk6XlJKoivdhagYtZ1wWylrfI6bExdlynEnEyirbazwyIJHzCYoWXxW+5mufWWBMpPDdHemmy0A1Mfr2TnyY/DPZv4MixoXoSJQgaA3iIycYf15cMlgNgEfGBmoHcdfZqsyAebQ04pgBVsw4gmu8dztyj8d2v9Q02c6BTe7YErHoP7VL9QPCUmrPRryhXQEnfpZZ7oT5YFy/O7Y3+naRzA+O8aQ/pScYte8MliJlJyyXJCg8ifGdx+v4NKzaISVGy7/nPzq6F/h9iNuZ3+HvDmCq1NwDeNKoWcN0PoMmQVdc/A1GBjW+kBrstX2HU3kiTf1ItC1CngDmmldnpJC0UwUa9vXsrqqoiIi5NUIG/888wouEVLjWGA81/l183PpOVydbkAfVmssI8VjS8CvU3CNz0RVOFd/+UeH/wjHDT4OJf4SxGMNgKrm1H9fUPd7Gs/2L98fR6fsFwFmbJ0BSZFQ4i/BtdUn4PqDr8VNh2tqeVpOIykndQquoAKVWUJeSMG1yhenRVyTgpvNwe1Y9Tb7vNRfillXzcIzZz1jGpP4exHLLiz6AJRLGYyvGo8JwyZY5iAbo9BYW2NNiBoXosQEpOwih9fjBeLZRXtu8aIru7gVpoVw4zjuYK6xN8IluC52CShceFub9qLvTuYeMD5cuK4jyUhkIiOjMyFim/dfaOqvz8Mh9IvkBtqHLj0cEy84BLKiIiUqEA0KLgAMrQyjrjOJeCb3srl43FAMyJYZCvm9iAS8rC4thShHs4pvUS7KqmRapeYnFUR2iwlfclrDtif74HM7RpRriwV3z70br655VTeRJ9IX8ARQEazAqIpR2Na9Da+ufhXfNGsu1SxvJjtwKqqCjJJBic96orGydSV+PefXps9//PGP8aOPf8T+NoUoc2SWJ7T8/3kjKuNv+VyfmBhjiujKNq1EEk+AAe3FY0ei7JwQ+Ung1q6tjMiTcQS/DZ/bY2WEYgSt3tMEgs+F5tuTj5QlpETehRNFVSCrMvweP4aUDEE0E80bpkvXmibISTHJJrN8O0jtMcIqTJsUXGO7CQPDA1ERrCgYolwWKDOFzgKa2kL3wC6tIOQLIegN6hXcAjm4OgU3S0ro2TASACP8Hj/GDhyLFTeuwE/H/RSHDThM114eVoZsBLqefLSCXVkLq+sXE2Osf1kR4FgmxsL/afu2VBuGlg7VtbPQQsqkdZN0ZlA8BoQGQFVzdY7LA+Wm8O2klGT9uCnexNpMRIXH70/4PQLeANJyGo2JLMGNDGYT8IEh7TflwfK8C5C06EGoCFSw54mfpBrVUKs+CFgTvHw5uNRnK0OVrPxU0BvUXXdSobrSXSgLlDHnXKNruLFPGPsCH3ZM7bDqD0RwjZN9XsGNZqK6nE9SAK2IKH9Nbjr8Jp0aG/Lpc3DD/qyLsiHyyNiWQhhdOZq537el2mzf0fRs8c8mgfonHTufimtcTMjIGeZab+WAP65qHIsuMo4Fxj4UFaNY0aoZhxkVXEEQcMNhN+D5s583tak6rFdlKxyEKI8bOA6/PPqX8AgelEhpbG9ahic+uAEvrngRABD22oco+2zCwIHcmBZRFJQ+dwomtrWxhXYKUeZD871QUaKq8MJjGdHEw+p9XhfV0jgivgh7VilEWVZldERzJcRCvhC8Hi98Hp+p//J/p7P30auqTDH9+1l/xzWHXGM6PotCMyxWdi7+N2RDZFdCTkLOhih7BS8Qz0aZBbP9YPrd6FzyXwCA0pUVEVwF14WL3gOV7CG0xXMPbhenpm5sjjEHY43gZtCsaLbzEw6u0u3jzEOqMfNXp+s+K2WKqwhRMhPcqtIgNrdoA80Rwytx9THD4eXCnyMsRJlycLW2EAkvykVZkfMW8GbmAw4UXAI/KaFQwGJh97LmVzqJ4BL4iTpNcqoj1RAEARXBCsTFOP7+zd/x/sb3dcegiR293EsC1gS3I92hmyTbwYmCa6zNKCqiJcEVFVFXYD4uxk2ThSMHHcn+TeGIxvISxrYY28pPvL/z3nfwybZPTL+lyaQuX8pmsg8AO6I7MHHeREbQaTLNE1yeWPOTJOPElRYf7EDn5ff6mXEUr+IaiQv1UV7BZQRX0k/WrPoiTRr5sDe/x29SSfg+OTAykBlgibKIde3rTPuNZqIoD5RbOoB6BS9rt53aQ/VIeQJZTA6ucdt8IcoCBF07bz/ydky6OFdz0qja5Mtzpj7Fh9fbhe1ZERZ+wYEnyXTvfjbzZ/iq4SvWprgYRzQTRf9Qfx2JNy4GqKqKh758CEuaChs89Qv1w91z78bNM24GoJElI8FNSSld36MxxxieOqx0GA6oOECn4AY8AfQP9TcpuBXBCtP4PPt7s9m/K0JmBVdWtXGfKbiyOUTZzkyOCDa/MOo0RDkhJpCWNBLKhyiT6ktldYhw8e+imz66CU9/87SuLcaJf0pKsWtO99VqjLIjuPRMZJSMScEdHBms24aH8Xks9Zeyvhbyhdi1CnqzJlMWIcrFEFx6r/EKrt07k54tKxOpoCeoO3Y+p3tjfyDXXiPBBYAjqo7Aaxe+xsih8Z1llcdMpZ/ofvHb/PbY3+Kk/czhtZcfdDlW3LgCldltyxWFmRHxxPShkx5i58+bMpXIEur8PrzSsRQfbv0QgNlkSkdwHQSnlZD3wKZZrC9ahSh7oFUNLPOFCoYoWz2L26Pb2fnwxlqk4LZ4ueeLWxA1LlryC/qx7HF8qgpkEsC6D4FlkywjRNgivWE85o9LSEqpXIiyx5dTcGkBqXk1uhIG8zdjX8w4q1G+t8EluC56Fc3dKWxpNYc7iIp+1ak9nnvAeIJb15lEWcgHQdA+55XWoFefazKmupQpr4Sc4ipBUvQhygBQVRZktXl/cMII/PFKfY3aqtIgqsqCTMGNBL0QBLCQZjuCWxutNU0OJUXKW5rBGErrBDxpOeetc3DG5DMc/5bgRAXev2x/ZuIA6EkM/Z4ICG98wsxoiOBmQ2Op3XYKrlNIioRoJopbZ96Kte1rcyRWFpkiblRw7WrkSYqEjZ0bWf2/uBjXEYejqo/CQyc9xP4mUxueKPDKp12IsgrV1liFQJNJfpt8Cu51065j+VpAbrGEJ6927sVWJQ7yhY+yskaCj+U98aHCdiSPKbhSMpe/yoVKi4pomeNNExJeQfB7/aYQOr7N/UP9WU7lfZM1hXIAACAASURBVPPvw5UfXGkZclketCa4HsHDQgGtJoiANsEp9Zc6qoNLE7585l75FNzRlaPzmsfkKw1i7GeUe88TVbuwPSunaf53/IIajWVUloj6Py269A/115F4qzJcb65/Ez/55Cem9AQjMnIGH239iP1tFaJMCiagPTszts4AYK5DTYSPFNymeBMGlwyGIAhsnxRyWRmsNIWW8vfNpOBm7zs5GgPa82MkKVZRChFfhJGi8kA5Cx/lJ8B0bKbgyil4BA/KA+WIS3GWg8srxHROnelO7RjZv1NyClu6tuBPi/5kcvAGbBRcnzMFl9ycedBzRSHKZYEy3HT4TTi0/6FMjXSSgysIAspVgbWDKbhUJsjCRdkpwb364Kvx+kWvA9AWoMK+sJaDa7EI7RW8LGT42kOuNX1P45UTBZcnWeTub6XgAubwdo9hGm9cRA77wphfNx8LGhawmu1WCwlG0MLBif3HAgASggBk30k0hpYHynHZmMtyBNdfCjSvATbNRomFqzALUfbqQ5ShqhrxK4BS6h+KzNqXlnImU0zBze6r3BMsGKJs9Z4lw7qyQBkj0j6v5vgcy8SwJpg7Nx3BNSxa8otw9N71CQIgJoDXrwHe/Ynl+4gWnCVDSbFWC4KbkNN6k6lYluCGsgRXSqFb0K5HRhaxqnUVZhlrXbshyi5cFMZxj32KM56cY/rcrODmCG5nUr+aFPB6UBLwoaFLP3EOBvT7CPrNDzspuGf9+TPUd6bY/ghVZbkXHZFhQl2sDhed0IG/fO8I9puA14Ogz4MY5eDahChf8M4FuG7adTpjIFmVTZP/aZun5SY+qnXu5rr2dbYho88ufRZfNWiuu/yLr5jcXCtCbVzl9Hl8OKjfQexv/jyo/bfUaOVtdCpNNjyPhSgbFNyhpeZ8Ex78Ma0gKiJeXf0q5tXNw/sb37cOUTaYTNkRMDJsGF81HoA+RBnQcvIi/lyIEu/eSeD3bWUIQUhJKcucWwJNcPhJkFWZIuP2BPodT6LsVD2jCzIAvLDiBdw5607LY9F58AouT0jsFhC6M91ojDdiVdsqhH1h+D1+k4JrtdhCk/9BJYPYZwFPwLTSzZ9fmb8MIV8IaSmN6Vums3atbV/LJjBGkyl+0usRPOwa5iO4/z973x2uR1Wt/07/+ukn9ZCQQkgCJAQSCKEXBaSDFGkqckEuFhQRLIjYAa+NYuHnFaJXQbAgipemiEE6gVBDS68np3/nqzPz+2Nm7dl7z57vnAS4os9Zz5Mn3/nK1D17r3e971ora2UFYJ0UGCBHnQcA8jhslBeoysXkTVUNmqxRHjJZUrVuGbBYupX4XTqfuW1zhfeJ3W9LtQnXUr5W1L6j4lZw9G+Obni81FuZjHc8yYZqQ2z+/cfGf+C6J68LjkNicEkmSQGRbeVt7DvkgBPAbbKbYkW7+P3K94HyIkv1UsMcXNU9ytk5tn/HdBhoGKmKsmM4yFpZ9Ff64cNnkkkyGuf9lX6Bwe2v9OO0u0/D0heXxo4FAL75+Ddx1SNXsb/5HFxy5JMAbtktx+4PsXvdw0Ev5oJdwKf2+hRuP/Z2FoRQPXs8K0hWCOdaVQ4uH1wAgme7EcD93fG/Y/f74K6DBeVIW6otkcGle58xM5jZMhM3HHaD8Dntk/4//I7D8cSmJ5THwM/n64fWo+pVYeujA7iNikwBwOE7HY4Xtr2A8+89n/XqTQK4dB0yZoat1ZfNOA2O5+Hg4RIQAi4ak3SPadzn7Txw477A0hOUAJeePZqD2dh2a0jurBtZhrY5uAn2G38FIFZRZqx2+P0mwxmRwVWpEAjg8jUbLB8w1zyGNwfeRE3TMLsSjDEe4MpzOv/sUtFF084LkmAlg+smMLimAuB6FbjhEAgkylIObq2EvhAYV70qTv/j6fjEk9+QdjgmUR6zMdth49v/AAGDu7G/hDN+/Ci+9Jeb4XRGbJRt6kjbBtb1lgBEoMG2xG2krPjwJYALACvWhxEzXWRwySZIvWfP/OOZ+Nwjl6A9ZzNZs23qSFnGqCTKqwZWCWC15tViTu13n/4url9+ffC5BMQA4ME1D+KUP5wSaxlCduerd+Ij935EeO/FbS9ij1v3EBbPV3tfZc5e1a3iwvsvFI5LNjnKaRmWADB4B63m1XDElCNwwOQDAECoFF2sFYXz6Sn14LonrmNO8sFdB+M7B39HWWDkmGnH4FN7fUp53mzfbg1/eOMPAIIqijxrKzO3dH2TWLZNxU3oq/Rh19ZdkTJSGK4NCwCXr84JBAW1ANE55VmOpDZBQOCIPt/9fOJ5EXNWc2uMOVc5jzWvpnyfIr58QZJEgKvIv7zlxVvw4NoHY+/z52FqJjrSHTA0Q2jJogogtKXa4Pouzr7nbAzVhtBX6YNt2LFx1Ajg8o6mbcQBLn9fc3YOjuHEAP7ZfzobS19cCt/3YxJl3jHRNZ2dR5KMfnskyqZuImNm2L3yfX+7AO6SSUsSP6PjTbLRtJqSK2GTyVL1vJ1nDqJ8/SmwQQ71rJZZAMByWlvTrcI5ys8hL3fmTe6DCsQrg2etbIz5SAL2MYBrRgxu1a2ir9LHwCSxQvQbVW9Qfr+y7JS+X6wVhRzc1QOr0eK04OmznsYZu56hbDuTs3JRYaKwvgHQOAf3he4XMLUwVQgyOoYjAG9eypk20wxEv9b32og9sO98NSqmQ1WU+ePg5yKat4r1gMFNmSkGpHRNZ/Pnjc/eGABc7tpSXQeVbF+ldNg71GQOVAeQMlNM0p820/DhC8+ooRkCwJXBrm3YbH4khQpZW7oN28oJDG4YRKD7JG+XLzJF9v+e/3/s9UB1AFc9chWGqkPC2H2973XG4NqGHQt0Ekjkz483+Xod1nUoZGPfKfUCpQhc/+mkP+GxDzyGe0+5l1WRbreyeGL1OhxTHGYAl54BUhnQveSBc1YB3FR9cAEAteFRMbhsm24FqXuCgnS8RJnuiR5uqqBZI6aANWJwcz2rYZeDtTk9uBlGWKFY933sVyoJ5wTExy8/V/SHa7zhFAIGNzRLkY/PfJjwexPqdeT9SKLc6kbnVHSrLAfX9+pRDu6Wl4D/fh9Q3Ip+ndoJRj5KP9+ZpDrG4I7ZmCltqDqU2LeLjOTClO+6baiC5Wv68I83tmFj9WmYhRXsu5ahI2sbeHXzIDRzkHtfnABTimgW9a8FgDe7i9A0CDm2HZykeeY4MdJJIK9UL0UAN2Rwqa2QCuAKMlWOxXO9iMHlK+WRw1jz4zm4L/e8DCDKAUkyfp9PbgrYjQfWPAAAWLF1BU666yQWmX+p5yUsW7+MfV8VjZbZR0u38PE9P84iurwjVPNqwqItM7g8aHm9/3Xc8uIt+PFzPwYQOEaHTzmcLQrt6XbMbJkJIJD7UWuJqYWpuHLxlWw79Lrslhl7WPWqAqil85IZ8iQnjhzuKYUpyFiZgMHlpkRavMlRIWedv1985D3G4Lo15vyX6iVljjGxwxTdrXpV5uypFt5bXrgFJ/7+xNj7dM48A0zn7fmekDOblP9H35WNz8E1dAPjMuOwZiAanzxbRUYAgfbbmekMGNzayAwukyhLAFcGNDxYztt5pIyUcP5bhreg7JaxobgBZbeMmlcTAC4/bkfD4Jq6GUiUq8kS5SsXX4kbD7sx2I4dyZk/cu9HUKqXlNVtZbvxsBtx9M6NGU2+D67cLkfF0LNz0AIQ8NK2lxK/Q2ZoBtJmmgFmOShFALfklrCgcwFuOeoWAFGv6BanRZgnhmvDKNVL+PaT30apXlK23ALE5+jSvS/FoQoHXR4LLU5L4riWi0zxktayW0ZfpY9JjelZb0+3wzGcWPsT2eTWIDSm+MBXzavh2a3PYl7nPFiGJaR+8EZBGiAY7wzgctJ8mjdL9RL6yn1YvnU5Dpx8oBCscQxHyeDSudPfG4eCIMele1+KD+z6AfadJRPVwRVeokznyT8LdLy3vXwb3ux/E7Zh46HTHsLTZz+NO4+9E0dOPVLYHr8O0XaVTKlCjn9+KZinJmQnMBZX0zQGMHjAaGiGMA6zUr6hrdvsc1KokLWn27GtpC4yRc8yq0gsMdb0t5BPzYG+nz3/M9z56p349cpfo39TUJwxbabxt/V/Y1Jzx3BiAHckBld+NiYrgsnsN7+9ELiLq0ptppCxMmhymqIgmlePZpvwfjNWMxybxODyc6cK4MqsNrse9fLocnD5wE3oA6mqKBthACkl9QtXmUopRb3C8+uXw64Ea1arZjEQ3uG6aAtJG4HBlVKw+Ge3P9yP6TQJANdU1L9gyrRwvfz0kIvxrseqJY+rc/3jvTrLwa3Xy0AxDAYOdwOr/w4MbcZgCHCr3Hq72uLuzxiDO2ZjprbFv1yMj97/0dj7Qs9X10PN9eCGQHdbsRpVT9Yr0PQy0qHk2DZ0pG0TA+U6dCtyWmwp0JVSSJTnTmzCxYfMAAB0D1Vh6WKP2c6C0/D3QOCcWGHermXocEyOwVVIlHmnWiVRPmbaMfji4i+y92lyV+XgEgM1Us4Q70DRJEpBBgLHJEeSQYsS4EpRTlMzMS47jsmuZImyKi8MCHuy+fHty+dFzsGXFn8Jh3QFecRZKxu1V3HLwsJBTsSGoSj3s+pWR1VkaiSWImNmWG6lriczuHSe/PXjmSVVkSlWTdctKaW8BGbJIau6VZaPRtKpk+86GRfdfxGAADDylZ/JXu97HT3lHqVE+dYXbsXyrcux17i9ADRm+FSLPY0rch7GZ8fHGFyZVeaBwcXzL8ZXl3w1BnD5Qjy8EfMlAFzdjlWelSXKjumw6pdAxCR2l7oZQ85LlPlou6EZKNfLMDQjsYWLpVtoSbWgp9yDvnIfLnvoMrzQ/YKwjffv8n6mbMhbeQYAHt8Utrri/C1Vtea0mcYBkw8YsS82PdMXzrsQ1x92fcPv8lZwCjho8kH4y9q/jPhdYsQYwJXYTBprlXoFTU4Tc+QpqNGabhXu2UPrHsKSXy7Bz174Gf7npf9JZHB5dtMxHKVEVbaWVEu812NoVBGZjO4vSZT7K/0MdDIAZ2Xw86N/LgA/lS2asAjPnfMc+5vGFA+2twxvwaqBVdizMyhYp2KFgXAMcwCXQDcfwNA1PSimVivh0Y2PwvM9HDT5IGEsywwuv5akzBR0TYelW2we2X/S/rhinyvYMfzwiB8qj48vHMUYXE7eSQD3yc1PwvVdOIYDS7dg6RZmtMyI5dDz7eZGk6fKj6VJdQ/3ehNx4bwLMbNlJuuNqwowGLr4TGeleiCO4eDm99yMTyz4RCzo1J5uT5Qoy0oQOdBE6wYPfHlJbiVcD3z46HvhThRcFwd3HYwH1zyI4dowUmYqYHCrjQGuzODKSotCrUEP3oENwGB8PRGMP3eJwaV9taeDIn+T8pPYV3NaHODStaK5PWJwSw2rKJNlEZ1rKvQvSTXGV1Em4+9IUuXyRrUx8vUqk/+2aSYD4VnPhxPun59v6N4smbQE1x50rXAv+gjgpiSJcjXuF7D0hjAYY6aa4HCsbSf32oXLcnDrtRKwOa4UGw5Jnmp1CEb43dUWdz1GaOH2r2pjAHfM3hZ7bNNjsfeqnCy55npCMameYpWxoppRAfQKmjLhpGnqmNgUSllSUSTWMkeWKBu6hkvfOwvT2oOFypQKU3XkkoEjn5tETK1l6iikTWzoC+WLTnyS5IGL3Oe2VC8JUXN+P6yKMsdYkMRUbi8kG89+0KROwIDAKk3ostOgAqDyAk7OiJzzBYQMLreY8xLl4dowOx9VbgktABSVNzSDMXFZK8sY43PnnCtIsQpW4BTyzKFw3Ru0CRqp/yblsBVrRYHBlaPN5ETyAQMe4KqKTJEzW6qpAS4VPWIA16siZ+Vg6iZzbFb2rsTD6x+OnTNvv175axzzm2MEiTJ99+8b/o5ZLbOY/LtRjqZSFu2LlTPHZ8cLRabK9XLsdwdOPpC9Pnb6sWhONcPSLYGxVUmUqVI1BR3ILMOKMRM8wJ3bPheO4QgONwGtbaVt7PgKToE9F7xjomkac+BlR4jYUsuwAkanvA1/XfdX3LPqHqHVk3x8OTvHjoeel8sWXsY+5/f/wbkfxA2H3YA7j70TozECgVkzq5T7y0bXtWAX8MG5HxzVPgjg0vyRxOCW3TKr3GtqJoZqQzB1E3krH2Pf6H5rmsZycBuZbdgMdExvmp74vZyVY+Na7l3dSKJMFX9lmWnKSGHX1l1HBNdpMy0EIwhg8QCXxuke7UFBw6T7lbWywpyjYnBpn6V6Ca/1vQZd07Fr666iRNkUGVweYNG5p4wUez4ISP/qmF/hN8f/JnZctE7xbYJoLuSfQbmScKMg7emzTmc1HPjjSuoVe9sxt+Gek++J3vDqmOD5sA0bZ84+E7cedSuACDTfsfIO9lVd04VrkPNEIGUbNma1zhKOh6wt1Ya+Sl9sDT18p8PZNVZJkfm/+etA7/38eztj2UtBz13DB/p1Hc2eh33G74O+Sh9WDaxC2kiPrsiUlK7Az0N7lCvIVxsEeOtlgO5hZTCQtcrmcr5BRSwyRWM/Z+dwz0n34KjJkdoiQ+BXj/s+xJQzv6NWGpVEOYM4g0vPPV9FmVZoh9skSfNlSyrmaGgG0rUyhsKgd5uvMxCe8zykQz+Af/aMeuCD7du6O46ceqSYgxs+b6bTLADKelh8qr0uprcBQC2891aqGTb5db6PJg7g1n0PlRDAbtj0NJaXNgEc0PcBlFg7wTKa68E1f1NgcMcA7piNWcySpMk118N37+d6crp+DOASK6rpFWiaD8cmllTD7pODxTKTidimOMBNLrSyU1sw6VgS40oM7scPmxn7DS0cg9VBQaLckrFRCVsOtWbjhRN41szjosN1PwK4qjwq1gfXjzO4SSwSGQ9waQGkqB85CbTwyRP4aBlc/lgFibJbE5wuWaJM2xccnnChIQeJVTvUDRy181EAgEO6DkHKTGHFuStw1pyzlAwuL92WQfeOMri2YaNgFzBUFYtMyY4LMSPbw+Dy7WJUDeVN3USr08rk8TU3kH/nrbySTW3EcAzWBllVVdqn53t4ofsFzOuYx65nI4nylx75Ep7Z8gy+//T3sWDpAnZMQARwJ2QnCCC/5JZiFVezVhZ3HHsHPr7nx1lem+yoqyTKdIwFpyA8A7Zux6so14cxpTAFd594NxaOXxgLCpED313qZsfHFw2Rc9XK9bLgJJHRPbR0C52ZTni+h7+t+xv7nFpqqX5Hz54HD+ftdh6On348+5yegd3adsOn9/40Dpx8ILoKXRiN0fU3dVMIMCUZFUjLWTnMbZ/L8iEbGQFcMhmYbShuwEfu/QjWD66PcuvCe5S38tA0jZ3j7u27C791DCdRoix/jwIdPDskW9pMs3EtV03O23l8eLcPC9+lbZPJEuWR5l8yeczRNSLwzgcCWV6vdB1nNM+Aruk4d+65SomyHHAhgLtmYA0mZifCMiyRwdUdZRVl/txtw4bru9CgsXOf2zY3JtEFotSdcr3MQIIqLzYpB1Vle43fS/j82OnH4oQZJ+DCeRcqvz+nbQ7G96yJ2pl4NZFZDG2/ifvh0K5DhfXM1ExhXzKDKwNTbHgGuKoJ2Pwiu2d8MObX6zfi2oOuZWNbxdQC8TZBQDCH+L6P77c04XWEVf6Ht6LP0NHseiygUvfqcEwHtmGzdYMVYJRwoKz20LYEaU4X9/bhFxs3I1seSM7Zr5WAWrguPfpD4CeHAXKQgb/O4VoaFXOKttuWboNR4vy1cI7q4AAuBQVo/mG1P+olJHtzkWW5c7UAGNDZc58202zc+2FQ0uFAs6paMZDcji9v56HVhjHAAK7GQHjW9+CEgRJ+rtDD83eK3YDvw3rqFvYZ5cGa6WaAWz+3hHLo6TUx+AuEjCwAK93KAH3B8wTgVvd9VMLr8vu+F3DJuHHA7u9nn5c1DX74eQ0aquHrdSZ3PcbaBI3ZmMUtaXL47TPrcdNfX2d/1z2Rwe0dDgBu1jZgmgEwcyyq1qpjjxDgVrQod9A0xIUp3QjgthLAFSd/xzTw6teOwiWHxwEuLUSBRJmKTGkCqG1TAFweQPFgtebWWHU/QaYUOoEEiPioNW3LMRw83/08VvauVJ4f3++U2CmWExdugxYTmV1T5T3GcnDDYyTHSS4OpALsgJiDS3k5ACdRDh0kWmxMzcQeHXtgxbkrMLVpqnAM/MJBCz8xuBkzI8pduTZBz2x5Bit7VyqLTKmqzzqGg7ydx0B1QAlw+f8NzRCcpyQG1/d91L06Y874Fia8kRSc8harblA9M2fnlGyqCiTzVnErzGHtrfTitLtPw1BtCLt37M7uE8/galKO0mMbH8M595yDn6z4CWpeLTiP8LrSPZMLsdTceOErS7cwq3UWzt/jfOaAyYy+qk0Q7SNv54U2QXLrEyBgxS3dYgBTjtCTBLOn3MOcIKHIFAcMK/UKVg+uRmuqtSHAJYUBL/Gd0zZHOHb2OyuPweogUxdQfhsZyeF5WfxojQCupVujar1FAJfO5VfH/Aof2u1DDX9j6ZYIcCVp7dObn8ZjGx9D3a/HgkGyUuOcOecIhbNSRipRoswbqSuAxkW5UmaKPee89JX2dclel2BBZxCwGQ3AHW1bGTpPKoZFNQQIvPPHzAdveJuUm4Rnz3kWe3TsIRSZOmH6Cbh80eUxFjxtplGsFbFqYBUb+zEGN0GiTPMoHXeT06TMcb3+0Osxv2M+gCDvf/H/LEZPuSeSKCuCKvLzHQOOnMntldJmGl9Z8pUY286s3A/89EjguYD1hFdnPVl5M3VTUI8AwRgUQD4HegzNiAOfFSH7+9r97Hj4db3geTB1M1aNXZYoq8aSoRnYMrwFJe6Z7x9cjz7dQMHzRLmrkRICKBQkeq3vNfH8pDXNff3+4HjC09Tuvwr5pDmiNhywuAAwsD5oF1OScvh5gBuCYZnBZTYUBa3IH2njr0t4zyiQwioc18qwGjC4Pzr8Rzht1mkwJT/F5tRPaTPNxnLYRQoOByRVijIgIgByXrz+BGpFuOE5tnlgIDzn+UwinTbTwOYXgfVPwaDje/pW4L9mw9r6MttUObznhjQ/nZCdhuOmvBcX9UbrctWt4ovLvohzXgjST8x0K7ufBdcTggE1eChz96Fop6IWQYjY2+AYNBRDtpf/zVgO7piNmcKKSdXXpLmqxjG4k5rTKFZcDJZryKcsWFYAcC0rmBAtQ8duk0JwZEVOkAxwHYVEGQhAh51ZD7vtrzEGl7avynGjhejuN+7GQD2QYNqGgZZMMEHrGlBIxydJXjbKA8XrnrwOPvyYRPnN/jdx6wu34s3+NwGIzB9JTA3NwBl/PAMn33Wy8hxf6X2FvSamj5x4AhuMwa2NzODKsmU5v0guMiXk4HJOu+d7bH88g0sMmorBTTKhOqGZCdhTvx60H7FzAuiuuBXm9Hu+h5PvOllZZEqV92gbNgO4PDOpBLi6Idzj7lI3A7H8+7RvcmbLbhnlejkm4zN1E+Oz41m+aNULqmdSTrDc/omXIKusUq+wfMIXt73Iipbt2bknu558O6SRJJg8COUlyrxV3WqssJHKmZDfkxncw3c6XBh305unC781dEMA5MP1YWGbSQyu67usMFlSFeV1Q+vwxKYncNhOhyUCXFM30Z5pZ8dORuBGPj8qMkWBDTo+KnrEnoEGLX+SjMaFqZvKyrOy0Zig+92ebh+xYrmpm8IzSMwjjWG+8joDjbrIgFJQQtM09pwAwTOnquYtm23Y7PxSRgpLj1qK3xwXl9HyxylLlOX7zefgkjGJsinmmI5k9BwtPWopHj/z8SC9QDMbA1yOwd1/0v44YsoR7G/WJshwMK15Gs6cfWZsn62pVvSUe7BmcA0LXPBjgOTiQBDA4oEmXRt6TwaaZAd1HYRLF14KAPj5Sz/HUG0INa/GxrCt27FxKz83jdJs+ODnqKwyBPhuAHSBgGVM6B/fkekQ/pbbBKW4OVUJwsnZt9KxAmUAGBAbicGV++ACwTrF1y8AgP7iFvQaOlpdN5bPyW9z3wn7AgDL8SeTfRmPgscELPvXIh9KZ0/d5VTcd8p94rmS/0Lsa1GsWK5icBnAlYs4FSN/baYWnPf7TO5+hGv2+EywhjDfpFZqWGRq8cTF+MK+XwCkmg2OFjG4vPqGNuVwoJUF1aU1deiNoDgnAdxJuUAp0lfpE6oLt61fzjG4mghwb1oM/ORQBnBdtwIMboSlOCdTqtLf1LcGX9v7MrR4okT5d6/9jv1tZTsYG93qudDC145uw+UYXACoeDX43LgphYDW8n0M6Dq88LtUsOp/8jmcb8X7n/872BjAHbO3ZDyD+8CaB5DZ+fvQrB642hCM9Jugqabu+ugfDgFuSxqD5ToGy3XkUhpzAInJtQwdnfkUzlsyBalMD6YWpgIADGNkifLrfa/jkNsPwW0bLoXT+Wfo5uilF+T43Lf6Pjy05ZfhsUQMbnPGFioykxGrZumWAJCoOEvKTAnR3XvevAfXPnktA14880fXQpYMyxIjWiT5XE3aHwFckiyPSqLsqSXKpm7C1u3oWD0Xru8mSpSBqOUNz1T0l6M8GSACtkmyISBevIGXyNq6LTCzKmZTlYOrkh4SwB2sDgpAn5xNJlE2A+mfzOBSOyWZvQciZ7ZUCyTKxACSUVXiTcVN8H2fFfAieSsPAF3PbShRputAzjoBiM/t8zlMKUxhjpbne8zxHsnRLNVLbLzwRabI0mYaVa+Ku18X21qp7qssMa65UQ7ub4/7La476DoBjGiahssXXQ5bt9lY4rcrj0OZdeMrR7/R/waAZIBLAZKjdz46xhSTRNbSLYFV/s7B38HDpz3MQIMyB7c6yJ4dGnt3nXAX7jvlPvZMN2r5k2QeIomySi4qGz07o8nXJaOKy3SMNGZIAswXK5MlyiyQFT7nrucKwRQffiy3UGUpM8XmppSZwvzO+azquvA9Qw1w9x63N2N06Vx4mS4Z+44RZ3dHY5ZhsTHbohu9nAAAIABJREFU5DQx9QAPcOka8ffg/N3Px/EzItl6Uj4nbx2ZDrza9yqKtSJjcGNVlLngCX8urJhWeL3kYABvqs9oTtQ0LRYEsHSLqRlGOgc50DeiEQgjOa1bUzK4QLxqtixRtnmAq+jVygNcWe4e/D74f6QcXNX7pXoJqwclgFvuwTbDQJvrIcUdT8pMCcc9PjseT531FE6fdbrw+xiDG7LrBrfO58NnaPeO3aP52/cDwEnXlJjboqSs4H2DmihRbsTgzvVN/H31OhxjcEGUUArbmZWqk9cbF5li+5HuuaMZrDtBa6qVXQvywmwO4NJnBEzpHIZC/4AA7ozmGdzxRnNUa98aBsJzhoOUlHYFIAK44d8qVtqQA5KP/AB44mbhu7LKzsqNY59PqdUZcDM1HXX4Ahvr+R7qnJqpFK4vza6LYU45UAt/8o32VjyqQuL/BjYGcMfsLRkPnlZsXQEjtQG5Gdfg6y+ciszUH8FsehqAWGRqcnMaQ5UaBis1ZFPRJGyYgfNum8GTd/6h7ah6ZebQxACuok0QX2UXADRn3ajPhV/stlXWwUivQsoy0MIArnoSIAfWNmxlAaeMmWm42Lu+yxgZAmpyTqMONcB1PZcFGRjADf+m45Jl5EqJsq+WKAOBHG24Nozh2jC+8XjQINyWFmLeqOUND57oPbmKclJlQ0B02jJmhi1I1E6DB7g/WfGT2O/pPEdicB3DQcEuBHmyHBuvYnB1XRcCBNtK2xjAdaUILAAhB7dSr8QALjG4pXoJg7VB1oIpZwUFinhwPlQbGlGiXK6XUbAL0DWdAdzZrbOF8wAixztjZRKlW3TcdC4swJCLJMpZK4tV/auwbMMyLJ6wmL0/Ggb35udvZrltaSuQl8ng88zZZ+Kps5+Kxo0EIgUGVxqH/DP0Rl8AcHN2LlYBlEyDhimFKUqgCgSgjHeeF01YxIpnAfGxnLfyqHm1SL0QHl/ezmN8dvxbYnB5ibKhGw3ZsmlN0wTp92jNMqyoqJCZYdsg2SaTFyLuzNP9o3NzfVdgcGtuTZljLhufg6g6R0u3YGoi00zH1+w047+P/G923PQdlUSZANdhUw7DxfMvFpjN90x5z4jHyVtLqkXJ4Mo9Q4G4goUqHDcEuOkOFhwggDs+Ox6Tc5OxR8cemFKYEm1XUwP5kRhcIJ7LDIj3QA6smLqJ2465DWfNPitxm2SJ++15E3j1vvj7BDppfvbqyQyuNMd2ZjrFIo8ckHCUDG445xqOUjJtceqJYBtqWbsqn7tcL2N1vwhw11f7UdM0tLmuIJ9OGSLApXERy7kNWdRrJh2FP63dAJ+CyF4dmPleYPf3o1AOxgv/DLJrSevjcAhw77sSeO2B6HsNJMqyXyJU8K0MosnzxB6r4bWl82KBjtro2gRBUjA50NBT7oGhGZjdOhvWfVdFHxo2HO7YaZ6m3FlWCTwEfVR8jFcOgfNvHT9qP5TRDDghUOXl+kZ4fFR5WVU4y1Ipbpb/suH5m/kJqIX3fUqtxnhz3fdRB1B2RFa4akRrEVVQbpIk2HWZfU8o8PavbGMAd8zekhF4MjWTgRjeNCOY3HiAO6kljXLNQ2+xhjQHcHUjBLihrJgkcNRDNs7gquTHotPqWWvZ61teuAVf+cdXEs+FB36riy8jM/WH+Niy4/DjN8+Dnl6Nre2fxiMbHon9jkCRzOCSUcP2RkYAk7YlA3XZyGn24TPGdKA6gJoX5UMSKJWrlT6+8XHc9vJtQi5mEoMLREVNntz8JG575TZ2rmTUuoIdW3g8fJSeyTTlKsoNJMq8k8znPFFOMw/++Jxkts9wsRmtRBkQWalYXmHYhoAPBmwtbWVyK4HBlQBu2S0HDG4mDnAJIG8ubkbVrbLWKEPVIQHED1QHRpYoh70q02aayYYZu6Zo7WTrdkM5Jl80jIEkK88c95yVY8e4W/tu7HejAbjL1i/Dt5/6tvCZil1ttA3+bxXrRg7dmsE1yFrZYByFY1sGlpSPqMqlBUJpPje/EBPHpG8J0ubu4UDyJwM0Ym5HagmkMl9ysmWZMhWkufaga3HHcVFFWbnH7x3H3oErFl2h3Iepm0wCW6wV2XWjwJWKwZWd+mlN0wAEwIpncMtuuWHxtx8d/iO8d+p7YWomk4AvHL8w9r1lZyzDsjOWCeOFQJzMjMsAl5+Taa7qynfhgnkXCPfkmgOvwVNnPSWcXyNrdprZ2kXPia5FaTGO4SQGbICAgVRJY8l4ADclP4Xt556T78Evjv4FcnaOjW0dujCXEyt70OSDAADFejKLnrEyuHTvS3HV4qvYe0JFZikPl86FzrkRQ58459y4L/CLU+LvMzBWDphHP1mizDPPVy2+Ct855DuJDK4yuEdg2qsFa43c/oeevfC60jVJykHmr1l3qRsPr39YAD6ve8H82ea6SNUjHyRlpmBzjGiS2onG1eRqGV31Olwz2J8GAF2LgN1ORiFc34WUFHae9aBSMkmUNzwN/Pyk6HsNJMrCM+bWgee5KvCVMLBe4YL1XK7nr4/9dZRuUBseVRVlWaLcqQXnOq15GjJWBkZYkdgHgFQTbK7asMzg2oYN0/cxFAJAKj5G42dCdgJQLeLAYa5FYnits5qBDisPBxq68lFhwIPC27dPKVinfcXcbvCB2H0+Cux7EdC/pmEOslWYiE1GcPxTa3UYlF/te/A0oCz5URVurFAObrMr+qY1+dgGNwJbXgb+8vWRW0f9i1gyfTJmYzYKo0XMMiz0leMA13eDxa7uBTm4GTvKad3YX0JnexWgdrhGlIMLRNJSKtxhaCIIU0mU5ZYwdSuquvvw+ofxZt+brCftw+sexovbXsQF8y4AAGVPTgIJZjYo7PDkpiex38T9hO/wAFfV5qC71B0DuCSJZcfp1WHqJtsWNRoHQsmJghkm45mq/ko/+7tYK+KKh6/Ag2sfFL7/4NoH8eDaBzFUG8J5u58X7F/avixBHq4PszxG+XMgAsEAcOerwSInR+k1aMxZGA17xTukmqYxB4IAbqN2NwBYERseJKokyqZmNgS4LO8sLHREALfiVlCsFVlPPyEHN3QK0mYahmag6lZRcSso2AVMzE5kbXYszWIAecPQBia7JYky7yQOVAdGZnDdMhzDEarK0vFrmsZ6fzqGE8jPwxxHWTFAdtlDl0Vy5hAEaJqGCdkJeK3vNYER5J0npUS5AVNM42IkgNuIwVWBj65CF17c9iJ6yj2sOFYSIKXxmsTg0vxw6d6XYqf8TrFjSvodjUN57BHzEmNBRmE0BvmK0D2Ixu6M5hlY2bsSWTMLS7dYgEm+rrNaZ8WOm3I8qQAcEATSeACja7oycET3kcbcR3b/COa2z8X+k/YXWnwlPbuHdB2Cc+acg73H7439JgXz7OKJi/HXU/+qZNPkvFo6fiA+t9Ax0W8m5ybD0Ax8aq9PNRybhm7AgIEbDrsBOzftDAC4fNHliVWg+SJX9OzIx5K386iUKkoFC4HUJKMgmamZgpqCNx6A8GCdju3Du30YfZU+BnST7Ny55wrBw0YMLlPYhEGURgx9YlCHMbQewBdf4xlcAlwJEmU+aHryLkENC0GZwzO4SoAbrhdhMDFn54R1wQQA32f3lOYd+ZxUxade6X0FGjR8tbsHn+8IxjMdTZvrIs2dU8pIwVn3FPtbXnPZ+ZIsdzBIyfA8F9ABw/cBMwVM2hv5ELwpAS4QAFe5uBQQsHp8azpJoiyQgOueCOTNe5wGPHdbBHDLAwANG84/o97Fwf7LjB0FgGu2dOP2XfbDk31SkU3pnn/Ynognaz0sdYRAsg8ATgGOW4vReHT/Dc2A4/usDRBdo9Tzv8Pvu4fRdMovgJuW4IaerezIKNc1o5loTeXxsNGB1Ph92LYX9m3Bik3RPKfiRE1+DrYzwORFwKM3NpRoW9lx2BhWPZ5Sq+PJUB9thOcyJP22YvAAN5QocwxuGhqTKJN5PW9AH9oCPPStoApzfuRK++92G2Nwx+wtGYG0JMChhaCUGNymtIVcKnj4eodrSNkcsCKAG/agJeeeHGtfF6cLVRVlHswEv4laonQPd6O30svYjz+v+jOWvrQUt718G1b2rlQC3OhEqIVRcpug3kovzrv3POGzjnQHjpl2TOx3XfkutKZaGUBgDK4bZ3Dl45KZib5KH3tvoDrA7slQbQhPbHoi8ZQoTwxAjHnmnV4CrzzolqPavOP87NZnAcTzrFJmijkBo2Fwbd2GBi0qWCMzuPV4fvUX9/0irjvoOgBRDqbgiBtx4KRpGmPjRmRwuSJTJAEmtkWVg2vpFhzDYUWmUmYK/3vK/7KKmIZusJwoyiWy9ajIFA9wB6uDQg6uKveyUg8Y3JSRYr8V2u1wrIKlW0GLEWk7/Ph6pfcVPL3laTQ5TQKoI9aZl2AKfWtHweAKnxmizHdHAK4qeEEFQwCxWJRqW7TPWLGccLt0T8+dey4O2emQ2DHJ50djip4z+fiol+2OVFGWGVy5wrDMqMp5wLzxxz2lMAWfXfhZ9j7J2/l9ybmB/H5YDi6n1Nh/0v4ARHl0UoGpi/e8GHuP3zv2fmJ13dB44EV5pvI8Kefg7t6xO54++2mcM/echtsmO3DygYytOXP2mbhkr0uU3+MDe3Rf5DEls/+8dWQ6GqoqiMGdnJ+cyOolqQPoGTV0A59Z+BksmrAocT9kQn4vl9snHyMdC7H+jVjoEU3uTcoY3EoEcBMYXJUJc6CQgxtev1oJ6F0dvQYYW6jMcXdr7NqqfILzdz8f75v2PgDxNa411Yrjhoq4Z+16nDvnXPZ+m+shXYvmd8d0YHMg1NLU8+en9voUJuUmYeZQsHa5oXpIBwDTAXIdrIqyIFHmAW6pT+yFmgmft5+9D/jTpbHfKItMFcOAT0cIXIm5rQzg9+s24A9rNyT3W62VhIJMRxWH8dWZ8QJrMoO7v2vic/t8Dp/f5/NB0IEBXA1IhQA3NJpvHW7udHyfVbTOhhJlZ+1jmDbYjbaV9zJ5NZ0l5bo6mgGkmpEuD4jPWM8bwvHVFYEck5+rdQvIBc9zozRYy0qjI2Sju+q8RDl4ryhB6Qq3pjCJMsfgNvtaTKJcf/QmYPMLwR8jFJ/8V7ExgDtmb8koB9fWbfSpovJaMMEUq8PoGy6jKW0h70SLMlVQBgBfExlcigCTw+JJLONoGNy2fPQQd5e7UfNqDBgVa0UMVYfw1ce+ipPvOlkJmNhphABXxRKR81j36kLbGAD44RE/RFu6LRapTxkpPHTaQ7ho3kXstwBiDG7GzKDqiZO6nBvVV+lj7NRgdZAB3FUDqxLbOAGNe7iqJMrrByOAK0eTVZVcZTkkf+1o+0mLNhAVMiFHivZJsjH5XgNRr1IgkkXzADfJIaS8OJ5Vl0GCbdiCRJnAcFu6DRo0ZQ6upVtBC5PaMFzfZY44OaCmbqIz04m0mWZMCUmmffjCPRqsDgoSZfn6AhyDy8kHeeefPxdLt2IS5YXjFyrz46Y1TRMWchpvPGjhQZaKbZDzM3kjkMWz9CqLSZQTikzR6/Z0O3s9kqSYxrC8DwqKyM8hWdL26DmlYJUsUaag0g5VUYYa4F628DL85rjfsAAUXfNG/bVpGx3pDtx94t0seEEMv2y2EZe103Zpv6r7x49XUhfI564KQI3G+P0RGJbbxcgSZWDHCnyNZCqAK58njcUduffE4BKQVFmSDH9H5PBy6xr2foJE+ZCuQ3D9odfjnDnxwMHSo5bipsNvGnmnMsAlMFYrcQBXzeACHLEYAhR+HPMMrk3r3G1nAd/bIyy8RGxxMNcqq5S7VbYOOLoN/O06oC9i7s6Zcw7Lj5atLWTRJ9ddIWWlzXWRqonB2FSVA7gJDO7e4/fGn0/+MzJDwVrhhesQA7gA8vlgvs7xYJ1fPwc3ihttCmW3a/4RvWfnWaCB9cHln59yCGizHeLflUFMq9UxtV5PbkdTL8cYTEOVEyoVWdSqQzhj1zOC/uHVISZJ9TSEEuVkgKtrOlJetM88FY0in+SNv4j5w4gAblozgXRzUNWbP05p3KoYXIN/bgwLCBWKjSTKpm7i+s1b8cNNW+D4QGsIVptrwZpU8WWAGz33TKLMHWez58UkyrWV92DTP76HezNpDO7APPFutDGAO2ajsuVbluNDf/5QrG8lsUR8Lihvmh58/xsvnYSX3B+jwDG4AGCaXJVYULGm4OEiAENMoBsubK1ZG18/cS5sBVbhQY8GDZ4WTIhVt8oYZgImxVpRyKfsq/ThkK5D8McT/4hP7/VpYbvtYV0QJYNbT5aNkoOjaRo+t8/nYiwCRXcJYBIoI8lo3avHGFy5EmrdqwsAl4IOqmrJvPG5uY2KTGWsDIZrw1g3FBXsUjG4MuBqS7cJEWM+Ej4aBhcInFK55QgxuCrAYeom2lMic8CPiaT9CZHt0Fj1ZC6v0NRNBmT7wnyf1lRrwOz6aoDrGA4be4zZ0qIq0rqmY0phCut5TAwuILLsAxVRopyV2g0AAZh3DEfojcqDGl42R0VLyIH75gHfxE/f+1NlEIeCBmQ03pIYXJX0kpwiFbOTxKaO9D1VmyBTM4WqwfS80P98vu+z5zyLKxdfKewzJm0lBjfBoeYrjvNGUlkGcCVwSeNlR0CWLFEmsHFw18GY2TIz1o+WAiOqQk0y2ycz0g+8/wHcf8r9bG4iCTxvcvVhFZDm5ZGU+iEHU+RexqM1/nja0+2456R7cMU+Vyi/ozq2t9NUEuUYg6uoCj5aIwaXV1TIxtoEvQ2OKg/w+LmBjmN262zMapmFU2aewvZ5UNdBynOb3zmfMfoNrSIFZnkGl3wQV7G+9a8DfnshHu6u4K+r1wX5hNJxU5GhJtdFlt5/LegdG0hyRQaXxsv7d3k/bt64mX1G64BTHgQe/Arw66ivtAxGV5y7AqfNOg0A0MqN+f06FrDXzZ4HvTrEWts41WHYHMAacayE7X1obmASZQCHdB2KM/sH0fzbi6Lv82BzIApeBz9WgGknx0CxksGtSACXlyiTVYvA4Cbg1fvFbdeGY0WWTN4n6V0V/C/L0vlxMtwj5vE6BaTqHIESBhSZRLnUJwQ7cuFz6xCIHtgIud8l9bJ1dAtINQHlvhgI5o2KhmW5dUooMmXYwXbQOF/U0i10ui6WlILn4NyBIXyhYz+cPKhOA6hyawox1HyRqaa6i7o0NdQ0DctTDj49rgNb6iMXAPxXsDGA+y63f2z4B771+Lf+2YeBKx6+Ak9ufhIbihuw3zcewNJHAzkPMYQVt4KBmorBjSaYAePxQKLMMbiGGTywhmag5gcPlR5KKojBJSeoKRO8/5n3zsIa7Vc4/77zY7vjWdiufBdj7/iKnwRMVIztzk07Y6fCTjh7ztnC+wunJxeJapQXyQONM3Y9I1bRlq8yqtpW1avGALSqEqrA4HKsbaPqqpQbCCiKTHGLac7KYaA6IEiUZZZrUm4SFo5fKFTQTJkpPPKBRzCvYx4AUbI82gqyaTPNnGeKwGas5KrUpm7G5Iw8MNGgKQGF6pryubf0P/XiBYCeShAoaXaag9xcFYNrBACXGCu5uAyBo6mFqRHANWwGBm56NmI7+ip9wrnwY0s4bjMlgF+hwAovUTYsWLoVc/zlazu7dTYuX3S58N4x047BxfMvxrhMlKfDAxgV20DvyVVOgQhk0fntUJGpEBzpWlTpumAX2L2VJcqWbgVRfK5SMP85GV2XpBSGJAa3NS0B3LeRwZX3TWNBruhK97LkJkuUaf+UCy6fT2emE+Oy49j527odC1LQtZfbBPHGP2M0B/NgMOn4RmN0bLu17QZd0zE5Pzk2VlQM7jthfJGjkRjcHQG4GSuDKxdfiTN2PSPxO6zI1I4w1ENibrGgAOECEFRALGfncMdxdzDm/20xGeAyVrUUMWaqgNOfLgOe/SWaBjejzfMYMOIDsiRR/nJ3Dy7f5QPi771aBFjCMUXrT1duMvYph+DHq7M126b0Hm6dVt1XGnc8wJ2ZjXKoDQAo9yMduuWpvrVweIkyGswTtRIQqo8+UAyOY2G5EgAoALvsfxkud3aCvum56Dd8kbeBkMFddEG0PdnsXFRFmetvzYwxuOHcQGwmz7rWSsBPjwR+cbLUfqgcKzJl8QGM780LJOTyHMyD51JPxOACQKoAm5N8wxcZXKMyJPREzoVsOgVAYqAfUQ5uAHCbA2l3A4D73uIwPtfdg0smHMreizG4YbBLg7rqMqBY98wUTpt6tADQhePk2lmqiky11qtxBleLimjZOxhofLfZGMB9l9t/3Pcf+PlLP/9nHwYDTa6rY0N/GV/8XVAO/vFVwSRQqpdQ4gDjvvbX4XsWNL2OqCsZAokyx+CmnWASWzR+EdZWHoVmb6V5CMVaERo05hhZlo9V33wfzli0E5ZvWY7nu59nuWhkfM5iVyECuLzUk9gDlcSVHEJDNwR2inLGVH1IR8PgytuXK9v+8Y0/4uYVNytZV142C0RAgnc82jPt7DjrXp05OLPbZuPWo27F3uPEvLa0mUZ3qZtdvxiDy02o47LjsGFog3BtZQDzlSVfwTUHXoOP7fkx9h7L2QuPk2+PQZ+NxODyEmUe/Cj7FyIADRkrk1jt1IevzAVVAVy6V/M752PJpCWJObjE4KqqKFMPRrlNEs/gAsDUpqkRK2dYjFHm89qpMitZEhhwDIf9ngp08Z/RuR0w6QAsHL+QATty5OTr84kFn4gBmom5ibhg3gXCdR5tDm6j3Dwa/0nnNhoG19AN9pzm7Ty7tzJrRk6aDKrlfVDV4KQesrK8mswxHOSsHCso9nYyuPK+iWWTpcl0f46ddiwAYFbLrNg2OjId+OqSr+J7h34PAJc+IJ0PA7iGLQQ2AMRUFkoGVyFRjgHcBgG5RrZrW5D3d9miyxK/M79jPvabuB8LBr5TNjk/mb1mAFd/+wAuELCJjSTKfBVlALh66zZ8fzNXTX/TCmDT8/EfrrwXuG6mwLDx8wd/fwjgrhlYA6y4IygM9XaZrAYbbQ5uz+vi32HhJP4csuGat1ulimmS2gduNQJNoeqBQL0hVAzmJMoE1DhQoFKw0PzSwquFnr4FPxnScbUZjpnKAFJUa6LnTQHAWN+ZGz9fMupdm2rG/OEiVhx+C8a5LmNwoRvATotZH1oAIojtD9VZC84BdjtZDXCdvAiKIc1dlYEABBOAUwUEa8NA75vBa74oab3E4LtOAFSWKA9uirUJwuAmBlwx3MPl4AJwmuBwvpkvA1xNj177PtK5YE5joFcGuB95IKpab9iBRNmtAMNiWhpvOoAzBoeQz09k75m88sqwADPyZZJkyrF5wrCBlqmJgJi/SsO6BtP3WZXoFtdF2nVjObi1SXuhEr61vX3A3602VkV5zEZlJD/eOhQ2xQ5Z2Dd6emDm4nLYei0F37MCBlcTAW7O4eSvqWASu3rJ1TjijvfAKiyH6y0BEADQrJVlzlrdq+OZLc/g0r9eii2lIMrcU+4RGLvh+jDSZhqfXPBJ9FX6sGz9MrieK7CV5Fyp2hjIvULpOwSKVQC3UbsLGQjS9mWJ8n899V+J25Ar3PIOe3k4mMBbnYA1IBDUkmpBd6kbc9vmYs/OPTEpNwlPbn6SbaMr34WVvSsxWBtEwS7Ec3C5CXVcZhxjT9l5KZx5AELAgRZ5chD4vrgM4DXogwsAc9rmsBxsvjJxEoNL1yprZZX3yvM9oaosWdpMw9RMAaSSM7f/pP2ZrI7Pwe0t98LQDOTtfMDscteQFZkyLDimg81FsdAQFRZiALcwlf3W1m3Map2Fgl2IVcjmjbZ18fyL0ZZuw5f/8WV23BQEkRcqxuDqNpPmvrDtBWF7qorfScZ/lw/mNHLc+byzP5zwB7zS+wr7eyQGVx4vQm4d12OZxmvBiRhcBiokSTEFuggkysc+r3MevrjvF/Heqe9VH1MCgwsEwY81g0FuXozB9d5GBje89jKwpf/fN+19rOiNyo6fcTx7LbdsIuMlylQYjYyCI14YzFQBVVI61P06m4NliXKjQmSNbF7HPCw/e3nDgNnMlpn40RE/2qHtb4/xbUPovsQKkDk7noM7GpMlyicOSWvdPZcDmgZ88G7x/fXhGrHuCWDm4bHtqhjczcObgTvPC0DRgrNjv9khS2RwyxFz63titeVqEdimBri8vac4jCbXCwCgvEZUhuISZUp9gARwQwBmh6wmuLm20TjM8HPyA1djXwCYtATQHgVqJaRhAKjD6X4VM6oRS201YAoZwG3eCdj0XBQQ4MATrAzA+zx8gH/jswA0oHUaYKYD6e1D14r7cHLsPtBzLkiUywMBG9lIhcHvs9QLZEK1Q6mPgbU9KsF1N1VyZP69TBswvC34l20HSr0M0PiaBhQmwOFYYC9ct1PNU4DaVhj1KgO4lu/Dzo0D+gHH98JrJREgTV34chH49XA/9hhXYLmzLDhg5+K54wDwkQfguP3Ay8GfuilJlDlL6oVr6ibQORfYEhaBMlNAy9RETr/i11EH8Fg6hZKmI63pKKaCdXDPcgUW/FgV5ZqTYwzuvwvAHWNw/0VM1V/1/9LIYewuBhNcU5rKxKtle7V6GvAtnLCgA3yqfVPaCvJu9Qr01Fp42hBzmDJ6C3SrD24oERmqDSFjZZhzUHNruHLZlQzcAhDyQoHAWe1Id+ADsz/AmKmyWxYZ3JB5UwFcnhnk8zKpl6VKptiIwY1tX2JwR+PgEIN7/u7n46fv/WkspxAI8jEt3WJSbHIcqTep7KxS/hadV4zB5Yo/yc4skOyI8pFy1lhdbuqOyAEYicH42v5fw6ULgyqOAoObAHAJBKoKMAEBAFcdu6ZpsWIiqn3wEuXeSi+anWbomi7k5gIRGLd0CykjxYAqnycKiDJQfr9t6Tbcd8p97D25rRQQjdWMlRGYMMd02PnLTJosXQXiuYkyO57EXMrbGYnBpUAQz+BObZoqAMcRAW5CNVqAY1o457KRRJmvCgxE/bZTmvdzAAAgAElEQVTlfRiagVNnnRqrCs5/DqiDNbxc9R1lcMO5jgJqOxV2QovT0jA4kWRJgJ0ArorBpWtP61RSDu4v3vcLpM002xZ/fYC3ljM6khrk/8ooIAckS5R3LuyMvJ2PFzB66hbg6Vt3bMelXuC+KwG3NnIKSKk3kFfKRmNRWg/IUhwnMjE3UfywrNje9hqNuVgObghEa2WRueVlyn1rxb8nzFMC3LTv4+ASAWZpPadesNw+I3UC9z23FkmUCeCqZJ2bngeW/zLYdDj/ZVTgwXQC0FIvs446qY0rMKMWnY8JiJJcsrVPAD8JJbDNIatP149/Du1McO3onHmWduNyoHXn4DtWOgCNf/mquB8ryyTKFMgWc3D7gVQhBtoE4/fJj5dNK2Bl2rB0wybcsDnw70yZwa0MiKxw28zgf6pcXOoVGc2WnYWK2YzBDVtrGX6dA7jAzPxUTKn7mFx3A6AvW6oJ4zQTF/f1QzdDBheIcq8756jPeeKewnqmWdw9kdZJvtAW7xPomh4Eo066OfxikLtrWuo0pWq9gu+3NOPC8Z14NO0gnW7HLjODAOfZA4MwfT9W4blqZZkEO8m/+lezMYD7L2Kq/qr/DOsJAW4hbaFUdVkbIN58z0SlqgOehZovMrie7+OTD30U+VlfQnbnG/CrV37JHv6c0Q7N6sf68gosfXEpirUiclaOOXFv9L/BWqmQrRuMA1xyKmi7pXqJAVxTM9FbDloFqSTKAoPLSUn4XGPZRupNyhudCy+nHMkIHC2asAgLxy9kQIJ37m09qLxLAHda0zTYuo0FnUERC5m9IXnb1tJWXPbQZbj+mevF4+SYZ8rN4200EyA5yPRdZQ7udjilPIObFGE8pCto3yJLw8n4fp6AuEDLv1Htw9ANxrz1lnsZsOT74wLxIlPk3ERyNxEU8deGB65kaSMtAFzHcIQgAV99NmWk2BhJYjz5+0fgiBw5Gcg1ZHA5MMxXVVUBXHre5IJVvPH3WGVJxXqAuPwbUBeZkgHciTNOxJcWfwlnzTlLuY+RAGgjBpdXl8jj6eCugzG/Yz7+c/5/Ntz+aPZ99M5H45K9LmHX/bCdDsPfTv/bDkXik87nhBknAAD2m7hfYr4lrVNJEvM5bXOiVjWh+gEAdmnZBdcedK3yN/9qppL0yvPcUTsfhQff/2D8/vzh48BdH8MO2X1fApZ9D3jx9+wZSAwYVAZZzqZ48OFxJvgbDvcsGLqBj877KL6XC3olN2TuRms0b8QALs/gcsfGg92yVANk4oIIxCdVqJXXc17C64qF2Vw+J9StsrnKod+oAO4PlwC/uxBABHDTquCwmQp+X68gTdV8Q7B999oNuHrrtmCl6lsd/+2jN0avm8PKzQTs+fFF61utCGx5Cfg9N+9UhyKAZiXcRyvN7gMDuHIOrpNXXwcguLc8C033ptQbyJa79sH8ShWFkOAw5OBDZTACuM1TgEPCInLbXgfWPQk8/F8iA9oyVZB4+0xdEq4TfiRXtuBjVvM03F1vDYoxqQCulYrGJ+XgAtHYHKcAuIYN6IZYPI+/J5LCj5coU29fZplWYMri4DUFXbLxehYAUPFqeDYVrM19uoG0mcFcs4Dlb67B3uUKLB+oyRJlhwO4CSlg/2o2BnDfZdZf6cfqgfgkJjNs/1e2YOkCXHDfBezva57/D6S7bkZT2sSG/hKgizISr56B76YxXHXh+xaGa0PQjChq55gG65NKRg52zmyHbvbhf3uuwjVPXMPAKp+nKhtf+AiIJMpA5HCXagHAbU21oiXVgr5KH6peVZCjkvGOv6qy7ltlcMlpZEVZGrTJIWO9hsNJh5xCHohYuoWCXWBA/sDJB+Jvp/8NE8Jo5cLxC7Hi3BXs+8TgbixuxD2r7okV3BIkyqEz2+K0JOYbNjpXcnp5ibKcBzka49k91f5XnLsCs1qDHMNcQh83kiiT8WBIltiqwLcsUSaAa+iSRJlycHVTcGJlh5f+56WaSS12eLkyvwCZuikAipQZSZTlecPRndjv2fMSguQvLf4SPjj3g+zz0TK4/HVVgT257ZfKtrfIFH9s9Nm05sg54XNw6XmRAZyhGzhll1MSx+RbAbjEUKaMVAxs5O08lh69tGEu5UhG5zyjZQY+vNuHd3g7vCX19Z3fOR8rzl2ByfnJjMHVoOHEGSeyIFgjBlc+5qyVZeNwZstMHDn1yLfl+N9NRs+yql3P217NmdgxTu2la3q88iwQsGEVhZyS5L4JijH5iC+afxEO1cN10lI8s1teAra8PMKB8/sPn6EHrgbWPBa9TyxpXWJwH70pOm8CuCfdDBx/A9DcFYCPWimeu0kmr928xLQuVlGu8KlIXA6uTYHyRiSE57J5pOkv34x/bjqBNLheRpqeoRDoTanXI4l5L+cbunXgmunAhqej94jBJfDIAys7DJhWi8ADX4kfQ2fY81rVFgkI7m94H8i3EoKVFZIoKwCuboY5vNz1JhC+YXnwf5fYk1mTuyS8+DvgudsD8PnJ54Cd9gsUBw9cDdx8GDC4AQafStUyVWRwXWr1GNZZgc+useX7wXnnwmB+23T1NaC12bAjgEvWvkvw/06LgQ/fG7y2FJXbDW6dUEiUKVf2oK6DFPt3hP/Npq74dwBUChPRF+6nZKeDYPmMI5ik2YQfr6JsZVDVNJjQ3jVqmLdqYwD3XWan/uFUHPPbY2LvqwCu67n41cu/irXueTut5tXwyIZHhPfM3GtoSltY31sCNBEkusMz4JUnBeyub2PZhmXIzbiGfX7qPvECM+TM5q0OaFYkWyGJsqEb0DUdZbeMPdr3wAV7BIC72WlmDO53n/ouDvjVARiuDTPmizG4bglbS1vRlm5Dzs5hqDYUY2+JVUjKKSRTMbiydLSRkQNIYGI0EwnJbgkoyTmFQFiYyM6juxwAXMd0EllMIGJwH1r7kPJz3rltTbXC0i1Mzk9WSlyTjBxlYrjfTgZ3pP0nVRj2fV8AkDzooEBLe7odR0w5Qvl7UzfZcQxUB1gQhAe+L/e8jEv+egmAKAeXjF7LEmWBweXO7fZjbse3D/p2HOAadtRLV1MA3HA8y7nVqvtHfZoJFHdkOvDpvaM2WUn9F+Xt6JrO7msjBjdv56FBwycWfCL2ne2VKPPXrSXVgh8c+gN85+DvsPdGI1EeyUZKI2i0PZJjyznsb5ftaJGi0Wyz0bYJuO/WvhuuXnI1e44pN48CKY22n7WybGztaO4ts+d/owZs/yQ7b7fz0JZqSyzY9Y4YA6VaJH+HHq/y6vsBGyazpAAnUVYDXEflMhLAVJ3jjfsCN+4zioMPjSTGtWFg6QnR+zyDy/s7D34FeP3B4DVJXifuCex5FpAOA2mlPqBexo2btuDnGzaJ+5NZQv5auaJEWai14dbYWu6Ui9Exk/HgHACqQ7h4z4txYW8/jizGlWMCg+uJxZAEW3lPwFYWu4FX/hQUOKIWOkCQlwpE14IHVuQPVIeBnEJFQwA3KfBiptg5zu+Yj6v3uxqf2+dzwWfbXg8k4qmCyFCy36YDsFfkCjLRMVJl58khwKV5R/ZrX38wGB+0fTMotITBDcDkhcCH/gxz+mHR950cUukoBcIPx5ZD6WF+mG+LEODaOSAfAtxWDuAuOCdidGmMGxZQkIrVtc0I/rdzQMcu0XmjQfE8mcGFj4WlMu4++EZ1lXTKqSY/Yq8PKjdbMQz05YN7POxVg/V05wOAK3uBCfNg+T48TRP69NbsNCqapn7G/0VtrMjUu8yo4qZsquq6z297Hl977GvoyndhyaQl7/ShCWbqOjb0laBJALe88QTAy6BYqAN+fMHbMBQUXPn+Id/HFX+/AsVakQHSJrMjrLocWLFWZA6iqZmo+lXM65yH/5z/n7hgjwvwgT99AGsH1+Lzf/887nr9LgCBIzEpPwlANKmU6iVsK21De6od28rbUHNrQv6tqZko2AUM1YbEqrAKFlAFcDcVN8XeSzKZ4RiN40PAhpxBJrnkqxKH5/ByTxAtH0li0pnphGM4uH/N/crPeeeW+rROb54eMOa17WNwqQenDHCTWvYk2WgkymSqHrFAvIoyv38ah0dOPRKfXfRZ5e8NzcDjWx7Hufeci4HKALKtUZ9LYqppLAJRDi4Zex3iasbo8618uHs3u202ZrfNxs0rbhbGmWM4TF4tM7gzm2eiGDpq8rxBgJQHrcdMOwa7te+2Q/ma8jgzdRO+5ysDF/TsZK0snjv3udjn/PHuCMAFAtkvb1krGwsIJRVRIpPrHYyWwVU9E3uN2wuAet54O+ydALhyhW+VdeW78MkFn4wV3mI9OBsErmi7VNgNeIsAcOOzwB0fAvY4HTjpnS8iNRr75F6fxCf3+iRTKwlBEs8DVv4Z2OXIiDF9O4zGraaxANd+k/aLA9xaKcixdd0ARPBO9kgA11OALmLlEvpEj9p8X2T4slwwnBhcOQcXiGTFxOCGfUUjgNsLaBoOKCmUVvJzyYNUYnDDMVreyDGlgkQ5DBRUhyKK+6fvEbdbGUIh3YL/7FO0UgSiHNxaCWnPhQlA+UQ88/Pgn8rO+k10/IzB5eZnYnBrRXUFaiZRbsDghvdH0zScaHUC654GHvoWsOrh4DtOQg6unQnOr39t9F6pF1jzaJCnnO2IwGWmPQDu9Qo+OljGPpMPAuzeSFLPj9cTfxzkD889Cci2wdjwOLDyxWi3zVMABDm9PlXoD8kFHUCHFhyrBw2ws8CkBcDqZUDTpGgfx/0gek3zmmEBhYnAZ1cD3wpl4V2LgKOuBeYcF11DFYPLmwRw21wX7a6LKa0zUVH9hoH7EOByPqCu6WztqrgVoSglW091HbAysOiR4QL8NSuFqqbBeYcK3/0zbAzgvkvN9VzB0VIxuAQc3innqZHVXA8b+gIG13cdaEYFvq8BXvBQDldd6H58onu9P6hyOL15OrJmUKWYHr6j5szG3/4efZfPpyVmoMVpgaZpsAwLTXYTHtv0GLA5+s3W0tao7QmXg7u1tBU7N+2Mweogql5VALg5O6fMTVQVKpIlyq7nBlUkFaaS31LuCk14jWRqtm6j6lWxqn8VgMhB36VlF3RmOjG9KYoyEoNLDNhIADBjZtCebhck3vwEKTu3PzriR0iZKTy28TH23ZGMnDpicHkAZermdjvn9AzIDO7x04/HPhNEliAJIMkSZf48aKw1Oi5y3J/eEjg7BIr54lO8FJtycMnoftM4UDn2KnbaNmxBRm4btsA+8zm4pFQAkhlcuTfuLi27JJxxY5OP1dKtWOsusm8d+C388uVfsuqrKttuBtdWF34i0zQNB046EGfNPgtTCoEjYoRVL5PSAzyIzv1IDG4jwEwA952yd4IZHE0agqZpOG/382LvU356o/mBtps2028Pg0sArm/Njm/jnbCHroVbDtYGAfCvewL41RnAh+4BpuwXvS8/N7Vy4Og3TcaojAFcHc2pZvzppD8FBQJ7Vonf45nbymBUybZWihjNBLmtso0JMbgJfaJHbbJcuJNriyPk4ErgjD4jRjAVOv08wFXJp4G4dJmXKFMV5WJQ16KymlOy8UWmSqGypjqMi/a9CM898u34fl67v3GOsuEwBjfleXDINd/9VGDF7dH39voQ8NR/x3/fOReYcViwHyCS/woMLkmUh0UmFQiYSWIg+RzcQ78APBgWm7LSYXDED0DgzxRV2Z18AKIMWxwPqabgvT4O4Pasira90+II7GXaWL7tRdu6gTlTg+0SwOXnlq6FwT86Rek+2wvPB5Z/DQDgUUCC8ld9H9PS4wD0Y4NlBgB3rw8G/9aHwQx5LeYlykBUaAoImNt9/iN47fvBcYbHk+iP0XbCAmM/2Lw1yCO2s2qSwpAYXG7NIRIICPxUHjMI66mZYs/xEBdgq5kpVDQN9jsQNP1n2b8PF/1vZqV6SZAkqopMkTOoYndHsoHqQKxA0/ZYzfXQM1yFptfhh6A2ALfBkCrXXOiIP6CrB1bD1ExMyk1i4IAA6eS8KJsZqg2xz+gcecaGZzB5Y0WmwslluDaM7lI32tJtsAwLNa8mgIWcpQa4KkZLDiZsLW2F67uxCeyGw27AA6c+EPs9WxR1sZpyo/N4bNNjmNM2h1UzntY8DQ+8/wGhn6OlWwLjPJKEN2Nl0JEWCxTwjrwMzjsznSjYBdbiJQnE8EZOK+Uo80xm2kzHK4iO0mSA+/EFH8ex048VvpMkCRoNwG3kbMvXhQFcLgdXriQtSJSpnVIoWVUFQVQLGz++CDTzx+5IeU8M4PpqBnc0RSQ0JBSoUWyPzNTNREnz9Obp+MK+X2jI7m1vH9ykysZXLr6SFVYblx2Hzy76bPDbra/A/MeNym2RyWP7reTgmrqJE2acgKN2PqrhNnbU/lkS5SSjAFmja8ZXr6Zn7S1V7Ryh8u9bthfvAn7zH9v/u1fvhb7uKQBSv18CUcM94vdlsHXnecB35m5Hf9lw3IbzT1e+K7i+cusSGeCSfW18VDk34VrG8iL545blvsKhjUKiL/dezXAVtonBhR9v30KflfsDSSjNhTzATaqTEQO4fJGp4HymWYGfsRNX0RhujflkDjHHtWF8dI8LcRPfb5jsDx8HfiuNoX25Ik9cFeVjhiv4aGYacFU/cDRXeG3eGcD+l6jPg1hrWlMJ7PPPFamaasNBW6GOXYEP/il4r32XCLzx6/KkvbljTAHwg2u2aln0fmFSxHKSXFoGdKmm4PfFqAMG+rmAVNv06DeZ1oB5rgwEQRszJQYoVNL60Axp3TDnnR79QX2LzYjBnTbruOhzXvVF11POs+UlymTHXR9IpAUlhBZcx/C4kwK2bHvjgmBOwfORCfOBlQXidD34jaKgIr+u3vqiWIld8LWsDKvWXNR5BtcJJMpjAHfM3mkr1opCbqfruyjVS7hp+U0M2Mr/b4996M8fwlG/2XGnq+756C/VQwY3ZKXc6CGquT40BcAdqg4ha2dh6AZjSFmRG2kSGKwOxkCQAHATit9Myk0Strd5eDPqXh0d6Q7Yuh2TKOftfJSbyDn+J8w4AV/f/+vs752bdo4BXJKN0j7JphSmxFpgABGwoT6oKiBG14U/9+OmHxf7Hj+hWbolXL+RGNyUkYr1oOQd2iQQ8t2Dv4vP7P0ZTM4nswptqTZhG9QXkt/f2XPO3uGelGkzHSuyJFuj8xfk19wUSNevkWMvO+6U68vn4FLPaCBg1vl7LDOOquucVGSKrGAXAoAbHrvneyjYBezauiuuO+g6ANiuHNwkW3bGMiw7Y1nD76gY3LfCxs1uDfLAeEZa3j5vSbLq9+/yfnzzAEUxl+FtbHF/uwFu0nl/ZclXcM2B1yg/e6v2TjC4I51Pog1sQEsYdMyYycErgcHdjqJ1iTaCrPYt2+1nA8/dth1AMzS3gnmuh0v2ugRX73d19D4Fq2XgKUuJXw571Dbosy4Yx+DCc4HfXxzIP6sSIOTmJ3YMMfZY3OfVZRsHF4fVxZrqCgbX94EhDuhVFO1teOt+Ld7Sh8/B5K+BnGtd54pMpbj5lcBJI4Abq6JM90Bj57okPRG/2LAJpw+K7C4FD21q3VMtxkF6I5vFFVXjcnAXlEo4Nx8USxRAk2FFoB0Ico13CqvqEpNIQLCkysHlikwNdwPjdwemLglAMOXfxn6Tjr+ul6J+rEBwnacF3QvQFSqpTGl9STWJ2zLTwMDG6O/W6SKDaziR5NzKiGNBrpbNmdkoaB4+v3a4zhvTDsbUxZ+MPucBbi6sEn/YleI2qEAUP18tOBv4iCLVy0qzaz4ig/uB24FTl0bvN6pNYjpqBpd7LbfAFHxrOwsrfNyHNZ7BdVDVtH+bCsrAmET5XWvD9WEB4Na9Op7Y9ARufPZGLJ64GPM7578lgLuydyWAgMkt2AXc/sRa/OKx1fj9xfuP6vdvaDehq3QqoNeZLNl3RadUl5LYgYCZpodNZlrlScCHH3uPB0kqBvfM2Wfi9F2DqB05WWsGgkhhe7odpmFiuD7MJgBDM5Czcyway8vCx2fH49jpx+Jzfw8KKYzPjEexXsRL216CYziY1jwNG4vBJD0pNwlv9L/BnbvaKab9UORNZqr4qtF8kai9x+0N2fjonaVbglM5EoAxdAPFepEdq+d7o+rJ25HpwDlzz2n4naVHLcWyDcvYvfvyfl/G8TOOR1chqvjXkmoRWY3tsJSZEs5vewCu53twzCB/1YfPAg38bxo52zIgpSCEpVvs3vLKC1Mz2bHykkwCUapjV0qUuUXnpJknodlpxiu9rwTbClsf/frYX7PvjARwRwMoRpOTq8zBfQsFlX5w2A/wRt8biSywfNzbzTJqOmslkRTE2V6Jslww7P/S3olql7qm48rFV2LR+EUjf5nMrQP/NRtfnHs89trn85jXMS/xq0IO7o6Cad7o/rzTveLLfSKrOJLVK9A0I17dmp5JmYmSixVqRsA6VYdF5zvJGEjVgJ43gWeWBv8KUjBSxeDKxyIBtRMrHk7s6VbLkIlB5cHi364F/vK16O/hHhF88ua5wPV7RRLZE34IPPxtMae3xgHUWE5x+FmpT5SMEhgs94m/500G7AQMnHx0PuU+7FGRztutRkWmCHBXiw3BV8z4+YIY3HJ/AMaJheafC8MOjovGxcQFAHxgzT+ia0tARsngUg5uKFGmFjP7XgRMOzj6HgFRwxYrItP7tRKw9vHo/XRzULH6sjejoIK8/joFANwYa5oEDHD1ZuxcdKzZ9gAgE0jn2hMF55Z8jY0kphSA57sAdFgEDu28uNbyr51cwKDLxhjcUYBAK82CBYlrA61z2fYgd3c0xt0Xft6U59CP7fkx/OCZgFkXAG7rNJhvxhncz6y4EchmsNsOKuvejTbG4L5LTQa4ru+yaskEaCkfdEckyuTorh0MciJWbh7Ea9bVOPT2Q9l35GIrvA0YT+FN75fQNJ9JlHkGFwA0Pb4YDtWGGKMlS5RVoER29gWAKzG4R0w5ApcvulxwoADgly8Hjdbb0+2wdRs95R5c+tClAICzZp+F46cfz/atagNE5hgOqm4Vp959Ko7//fEAkhncpPY35DwTAJbPOW/n2Wc8wFXlLcrRO57xHQ1DRznctG1DN7D0qKW4eP7FI/62kXUVuliQAQjksgdOPvAtbRMAy7PVNV0EuIprLUt2yTwEEmUKLPAMriwfVpm8L7pHhh7k4G4qbsKWUiTD4hlcHjD6DVjEkSTKH97twzhn7jlMQqySi5NEWT4Xum470h9VZfJ23iqDW7ALmN85P/Hztw4iNZh+sjwciD/LIzG4hm5gYnYiJudGmSv5/9n77nhJqjLtp0LHm2fu3Mk5MDMMwwBDBslJyTmK4UNgTcjqiphAxVX8dnH9QFwVVFSMK2tYgiAsCEqWHAeYAYYJTLj5dt8O9f1x6q1z6tSp6upw08x5fr/7677dlbu7qp7zPO/zTgCcseQMr2Y5FvrZebD1xf/B2UvPDu+/CiDhfm99NbgRSd2xMdIEd3BrddMXc2pCGEZwZWWSvney0lsJTslPNnuFUqRyWSK47rIHJFutrESSgqZScMkyLFqUH/mef5ohyY4tgrZn6xr2mEizG/8wBVceCAhTcIkMDm0PV8EDFmWX4Cab+P7kFOqzmKJcGGKko1xgymhciOdJUnALg2w5NPAtnu/MBLO+EonPdACUEkz7TSS2fxMju2KIGd1PDG5l3ykK8TrqamDhYcJ0IsEVBuApofndl4A+QX0lUpudxNcnn6fTbf7a3pbpfDChbTaw+1ls/w0LaOpyFVwiuFn/AEWEoGMraq1nNM3A+YUkz72ga7/7+/rqu1vxzc1bPGt/JLwa3Bjnq5bpQIu6X3hgedUgO9n7LMQBTnkg9oAZB3jXOB/BnbILEp5FOXhtS2aDnU4mKjTBHacYLEgEt1wK1NzWo+BS/eZbvYzgFkplmOkNeHeIX+gqEeeCq1hxi3IGKVv4SikI7vbc9qCC6/6vIiXyzbJoUZbr72S7Ly235JRw1NyjsEfXHkiYCV+w0sf3/DhOWnQSztrlLACsTlDGH07+A+49414kraS/VQCAzYObkbWzATU57KaYTrJErGRbdnOi2TtRiYRVdQMoK7jVWJQBVqd44sITsXLKSm95q7pW4eLdL64w59jg+sOvx1/OYHXNu3Ts4r0eR8Fd1L4Ie03dC1fueyUSZsKbx1cv6x7jqEGOgIJr85CpRzY8gqN+exQeWv8QFrUvwo+O+ZGvBlcckHEUJIs+v7CQKXk6+o6pCHnKSmFa0zR8ef8vB14PW0ctkL+XtmmPaEuURqiktISwZe3RtQd+e8Jvvf/jBKrddfpdOG3JaXVvW1w0aoBCCccBfnw88FKw73goSI2J0dvVdq2rGZiNsShTvWhUH9J6QL8VOZinEorDQQssEKHgSsoknWvk18NABL9UDO/7Whr2r5fuMfqloETJ4ugtT3VuJAIovicPBoj1xuUS8MfLgA3PsO+NTODtDCNI4v2HSHDkgYBCCME1DEYCh7ZHHw8RIsGlz05lrx7u91wKCYD3TxVttyqIrfvEa4mn4Pby/wFGGOn8Q+daUqYz7dxRQPtN9yFD23nYFoHIL/XSbfJncPBtocRd26/gtsxgj6/d558+I9Wpiq/RPUq6jS8X8LcpOu0mN4Qqway6e3+YrdezKKfDLebypisU3LtOvwuftafjPUPse9LiHiO6lp/cP4D3qlo3qUDXjDjE9OxbgWP+NXqaWq7D5/8XcAjr8iDeP8jXszmtc7yBfF/JyJSlnotpIBV0aY3otWWUoS3K4wzUZ3OwMOgjU0WnGEpwa1FwpzZNxdretZ6CW1DE/1dabrHELqiigtuStpHvdy8aRvCi0p3v9ogpEVwicqp61CiCKyu4st1XnPcL+30hEIDz0+N+6v2Yj5h7BJ698Fnlfs5vm+8tf13vOt97vcO9aE+1B9SgMOvg/jP2x/1v3++pI/LJhNoVAdF1bAoUgaYAACAASURBVPI6Epbfohx2klo1ZRWeepc1Vt9l0i645qBr8I1HWa1imKI1XpC2095nPKtlFn5zwm/w5KYnlWRN3v+mRBN+fOyPAQDLJy9H2k7j5W0v47I9eQ0OKafDqiAVF/IoqRgyJWJW8yysnsZs5fS9VtWMixeln7/357j3zXsj9ydtpb11ETkPc1rcffrdgddUYWr1QFabR7rfJ31H9+zaE3tP27vC1AoYBpLu4ELUMdhlEh9AiVIjxwq3nXQbXut+bWQWXsyzth/r/gZ8OUJ5E9HrDhqGOCdEJKhVmqD21/W9oetUnCCjWpDIMiJUjToHsJtyRScBb3sDNbgC8SyXODmQQ5XCQOeBP38hfJrvHQiIvUKJ7PZv9k/Xt5ElyU5dweyiRARlolgucwIYlaIsEtzuN1kSMKUBf1iqX0ykgwS3mANSbUC+x9/3FfAT3M7F/veI4IbVxhbz/tpqj+A28wENlSX29k/jxhO+jfXLPwTjjVOYUtfzJv8dhGHSAmCTe59hJfh+Ul0lrUu8lzETjGwHCG6Hn0ACTLUmyOSFBsy7KxDcMAWX+r5S3+Epy4B3XwwGMQHAqT9gn9NrfwEe/T4j26Ti2xlOxMX1AcDiI911J/w1uDFFHCvMXpudjC+u3YZLunvQeygbFIhTkqVYgfsY4/rZFEMJrUXB7eDOmrAaXIDdb6TtNPoL/X4hZfJCL0V58JivAY98zTefJrgaI4akmUSxXMRAYcBHcEvlUlC5dS081z52La597NpQgqYC3VC82fcmW3YpeJNcSRkulMuw4Vdw0wkLCctAoeQARpAgd+e7MTXLbBtEDuIqV+K0gELBlQiueGNKgU/iDXm1P2TV9vXme9Gaag0QnDCyeO7Sc3H03KO9NGL5pNScbIY1yPtQfnjFh0PtvYEaXFHxDblhvOW4WwKKH007EvV8I4mlk5Zi6aSlyvfkz3a3zt2855fsfolyHvp8oxRc+fOiQQX58xYVfVpuJYvy4o7FWNwh3aBJyxC///SdrjQQIqKaGtxqlkcY6TpUGqC6eOXFrMdn1TCw31AOlxebfC6AiYbZLbO98LYAcr3AHZ8Fjv1XtbpSCUR6YqixHkjBjXFOpbNMRiS49ViUiQj1bQBe+TOw5Ojo6atFsonZJatVcEvDgKrcwVNw5XRj4f9iXlBwK1iUN7Beux7B7Y/oy751DbcCi+uULcrvvgT84DBgtzOA037IyWsglKkP3j7SNHJYFOC3KMuEcZs0UGNnghZlstTme4Jtcoo5NrgxtN1PnABBwY2owRXP9zSYkGrhZFVlUQbQ8uo9WDrzCvYPhRKJdaUqTJrHCa5puwS1yJOC6Z5PHCiyiOC69x9pwaJM51siuKbFyPlwf5DgmhZbD7XTClVwU3y94jmgeSoAA9j0HBtsmLqcEVzVOaZrKft79S72f6qVK7jJJn9NuYqUWil/De5RXwHudgOfZoa3XrMTIbXq2UlIAJhRLGHAJdQ1EdxqFNxYy4tezlcO+ErkudEXDKrYHxpcz/gCvlJcwS0G3SGNGvweD9AW5XEG+nINFv0KbskJtyjXApr3pW0vYd9b98W7hZe8975550t4e/tg5eWb7vaRglvOIJOwkLDY12ra8AcC9j6fRdlNoPVIgmkHiELUjXhAwVUowGcuOdOXYCkur1qCK05PZKVnuAdtybbAycVU1DYAjHQTuVVhfut8b1m2YeOyvS7DnlP3VE4rntziWpQNwwh8Jh7B3YEafL9n1nuwdNJS7zicuvjUivPEqcOOq+CKx59+xyLBDes3XGnbxEGMS3a/BJ/f9/M4dv6xYbOFLqdRo7SNTlGuhHpa2AAADANZx8EHS5kJN6ATG4/9EHj6VuBv/6+2+UmRS9RAcGO06im516+MYTdWwR3cAtx6BrPoNhL0m1MpuLke4L8/ym/GRRRzfpJG8AiuRJzkHqxeDW6IglsYYn1P7/wccOeV1SnY9PuhdcoKLuHlO5nV2rMoS/sjk3KAWY9liAqufBy3ve7/P5FmN/6iMp/vY4m/KhSGgC2vskGIqcv971UiuKW8/z1RwfVqcHv81mJCuo3vf4trUe6LILj7XgLsKlyHTJsPYgjJuOz/jH86IKjgptu51VesmyRiq2qjmMhyizLV1IbBSgSJNhH5rqX+7QgDKeeJLD+fJLPs+HrbpAiGspPwBk4SGeDAT7LQp395A/hAeOlEqIKb4eFwlrtPNZ3/6TzVqGtcBTJ5yuJTcPyC40PfVym4X9j3C/j9yb8HwEUfuRQucdoPAQTTlgGt4GqMIEhhHCwMIlfiJ95iOdyiXAtI/V3bsxb5Uh594LUjN/7vGkxvS+Oo3aK/6KbFTu7l4UkY3r4Piv1LkekigltCkzEDVx9wNb740Be9eRw43o+OgnB8xMxOoVjgNyhRo0kywVX1Vf3i/l/0/S8ur9qRKnH6fDGP3X7CVMGj5h4VID612H2/csBXcOLCE3H/2/cDqHwC9vWtNe2qUpRF0AjheLRi1oq2VBt+c8Jv8MLWF/DYxsdClVERdJMtt4ISIX8mZLOnz7sz04ktQ1vQk+cqBdXS+2pwqV1UjPpOgH+e4iBOykr5wrzigH57YSFc1UIeXU5ayRH9HtFxrpmcenWaI2RnHQ+gm56YdWsBeApueCJpAKR2DW5jhCTiO1B0SXBaKBmpj+BKpLowAFghib21gJTRAUXI1PongKd+Biw/ya8cl0uMoKlqP0MtysLNZu96fgyHB5iNVh40/fMXgcd+wJ53LQfs6YiNbCcjbqTcDigIbnYyq6Vd9yAfuAiEMgn78NStbB4VyRBVW7n/r0xw7Qxrx0LEvphnx6xjHnDqD4Hf/R//9MUc8Mqd7PniY/zvZTqYyqhKUTYstmxxn+gzSDVzZTffCzRPAbZLZCCR5b+VNjdgbmOEi+64bzLbN8FTcOGGTAkDSjKxBDipEi3KXcuAs34GLDqST59qZW4GRX0lks1cTQ9TcNvnAstPZqRSJjukvM49kJPzKJcIKeKJDD+f2Bl2fAkqgmuFkP0KKeZG2KAczWclYbvOP+/+6fzfBUOxwkBtgmpQOU9ceGKwH66szn7yGfXvJwSyyAEASyYt8YJDaX2yy8t2P/tBRfnDjqTgaoI7zkAX/IHiAHLCDYqYokz91wrSaGrZKce+YSaSTCR6uCxcAIwScoVSxRpcw2Ynd8dJIr+RjUymBQXXtgyl0iL/6ERimrJSvlElIvyf2+dzXm0qQe7jqlJwZTRKwRXrNFuTQYty3M9BRDaRhWVa3om3kqIa1Qe3mvXTMVGl8U50LJ+8HMsnL688IYAF7eyiEGZ7BhQKrvsdpt/k/Lb5AYJ73PzjcNua23DB8guq2nYRdNGp1wK83/T9cMU+V3j9ZusFEc4zlpwBALho5UUoV9svtJr11avgVlGvecy8Y3DX2rtqW89Ygm6Oo2oio0A37bUouOUCm18OuBFQdAlpxrC8c09dN1XydWp4ILwlTVyQemcnuQqlSlEmciSrkvR6uRAkp0TIAxZloQb3RsF+v/YB4LaLgbN+Ciw7gb/eLeRBDHVzFTEO2ucwy+/TvwQOuUJtv97zQuDv1/vDxuTvlLgPTgl4+hf+nqoEMZxRXtdWyaKcaXcV3ALw5C3AfW4v+lSrul1SYQh4435WE9o+O7isoW51inK61SW4IQqu1yaol5FBufY318OJ+6Ijgcdu4nbxMIjEVVRw5ZY8cg0uTUP7RI+G4f9OANEKLhFLOxPeesqygTN/on6P7OSLj+JkvVoF17QrW5RV7YniIGzglrZxxp6wXJLqXcsXHaGeRwXvs6h+QO6ag64JvigvR6ivjQNRSKH9Ee/9QhVcdz/ke2pgx1JwtUV5nIGsi0OFoUANbiUFN0p5Itxw3xq8/m5/YN5CWZjXGEauUPZP4wRH5A3LncfhP7J0wkLSYtPalqm8EaUfnZyiDARJKhH+c5edi4+s/IjvveZkM/790H/HovZFvuVGQVScqlWxSiEpna2p1qCCW8MNOJ1YyN5caRniyS1hJZQKdhx4BHdHVrViYPcpu+O2E2/DecvOC51G/pzpO0yDMofPPhxHzDkC/7LPv3jTdGY68bsTf4dZLUIbGfdQU6ufSqDvRr222rSdxnnLzqtpAEYFwzDwxPlP4Av7sWCb/abvV2NtbDx4BLfWQDQnvoJ77XuuxZPnP1lxunEHuhEOS46thFoUXNGiWyGMiRTcjGFhccdifHDFB70WYDVBRXDrxb8t4SSTrsOq/aIbeJm0FSPamtDAdCBFOaTW9h8/A+D41T/ATwxyPdGDNhlJ+Wqfw5JY873Ay7er1921HJh3EPDiH/lrAQVXbnXUB7z1GNC1q/91MeRJHigQFdwFhzLbLYUv/eHjvCVNsllNyoo5pgq3zQq+l+lg+yh/JxYeDnQuAba+6q8ZFhVcp8zs7rke1rpGxrO/5mpydjJw2JXBaQi7vI89WpIy67Mohym4kkV51t7AtJXhCiwNLqkGmdrnsMemKfHa4shYchzfBjHNOQwewRUUXNP0W5RV92HigFc19zVWClds3YZfr5fSrGlfFx3JBYRarqXVtAmKtbz61FJxH6Y3MweH6n46jOCqLMpj0c99pKAJ7jgDkddAyJRQg0vKrVwrmK9wQ9M9OIxv3fUyLrjp0SDBdfgFuXnJ1/B6/5O+aQxEkMGy0OokYcJ2FdyEaSBhBE8E9KNb0rEEM5tn+votqmr6onDU3KO8vpWxCG4dCu7L218GwPqqiRBrcKc1TcOzFz5b00mC1OpaFFzbsKsKGxJBxySq7/HOgkUdiyLJn3xRpO8cXSimZKfg24d9G0s6lkSuR+6HXAn0XR2PF5+klWwYYa4E+q7WruC6BDeGgmsaZmP6s4426Byqqv+MA4/gSufHNx4IJ82FIU6iZAuqBI/ggim4l+91uTJhPDYCBLfKvrEqDG1n5MdxBAVXsV+VFFz5ORBuUabwnzDI+ymShMJA+OfduYSl2oromAtMcUPW+jep63zbZjHLr9hCSB5EJwWXfv+FQbYtk6W+7T6CKx0r6nd6xk9YmxiAkTq5ljrV7N9nb9k5tnyV0kckrE/YBzsDXHAbcMi/MFX2wX/n74kKLgD86ZOsrraSIyDVAqw6FzjhO8C8g/3vHXctcM6t7rqF+xvTFqyxhkRw0/7pAE6qFh8FXPLXcJJFyq3KojyZiQGVrL6hOP1m4PKX2Lrn7MesyvQ9UoHU5c7FXME1TP9AhYpo16rgWjbO6+3HsmHpt7DidOD4bwMHXeZdO+oLmarTxkttpeocsBavg1cfcDX+9eB/9d17hCm4NF+/4lxZT+njeIMmuOMMRHAHi4MBizK951mUpS+iWLOrwrCblJwvlgPzFgUF1zDKeLT3ZuSFxu2Gw0845aF5KA7wC5jjiATXQjbphiRZhvIGkX5s89rm4c7T7kRXlo+OyiQ1jnWNlpexKp8IfTW4ZnUnKbJhHjL7EN/rYopyXEUuatvijjCKJ2jLtIL1HTFBn5EmuJUhHvOsnfWIHV0o5GTvMJAdPO73xbMoj/NWTiONum5OgJ2jBpeIUAxHjxJeyJRwPtn0PPCTE4C7Pq+epzDAFbQKacNcwW3Q7YfsrJGtv/Vg2+tcjc2pgqRC7MvisZdtvaoU5VKBBUZFnT9ExXfziwh8h1XpxYBrf5Wude1z2OebbGGfl6oVUftsYK7kxihK+0I3yHJgkawYywpux3yWjNvp3oybNqtj9vq/JoIhXKmWEAV3iG2/SunzCO47nDSSi2PRkYxobHlV2E6X4La49cz/+Bl7FK2jKhJLIVR7XQgcJv1GfGqsRHDP+w2wxwUsvCmuRbkSPItyBMGttXwhmeXtgiYvBD54ezT53+cjwJXvAK0zuIJrWOH2aIJPwa3hvma/j/r/t5PA6g8CViK2gKDergYpuBf9hfXJrRMiwW1JtgQCqbxyQOm3QQPFPcPBFlhRIZsTDZrgjjMQ8ZRDpuJYlCt9MfMFSm41AvW7RScv/T+MwYLwWjkBx3GV2YGDURoSTvgO/7Fnkhbas67KYpnKG/IopVVWVeOEj9CPuFoFt9ownENnH4pnL3wWUzJ+a5Co4NZDcD0bqpCiHAVZxaqZ4O7ANbiNhnjMmwU1gRTctmR1tX9xv4ONsihPdNSt4NLN7Y78VQ/rWRoXqjZBFEb07kvB6QFGYFqZk0bZN1SclPrgNur2YyQsyoTX7+PPlUnJ7jGWA6giFVz3Ozjcx50Ebz3CjtvyE8O3hQhfuQx8dz/gqZ/73w+zhlvJYFgQEbimTvbZDg9wUjF5MfD+P7ABi2bJmhtQcN3vikxw5f9FgjuwlX1XDvwk/84km/1KnpUI9pRNhhDcQo4NQqhqxong9m7gROzIq4T32/0KNcAURrGOeNV5wP4f4//P2Z8TRYJYY52UiLb4O5IJ7vTdgZOuZ/NXDJmKec7zeuIqXBG03arv8kjAMPhn5lNwFeRbBM1jWNWTyat6gGO/Hvo2XUPHNEW5bRaw9H31LQOV7xHDLMp0/ezJ9wTuWbWCqzFiIPI6VGQ1uFTjV3R4ivJTm5/C/771v1UruPkiu7BaphGYtwT/vEUnh/5hfjFzyjbgsBNCykrAKfEfjOPwE0U6YaEjy07iCTM6ZEoFuQa3KgU3BsFrRAsTeZtaki3eftaTIEvLJVWwGgU3zvRhoGNSitHiY2cHHfOFbQvxjYO/4b3uEdy4Cm6VDEu2r++sWNi+ELOaZ6Ez01l5YhUqBOftEPB6ltYZMiWeu+m4qW6yHYcpaBRyVIHgFl2Cm2rUgFqjLcqiSrnpefaYamP7JQeoFULqc4sxFFynzJVTaq2z4NDw7aLPJSwdW+5lS5ADjABOLJumsATlwhCv6UxmgQWuS0lWYmUbNB1rebom6fdJSvcD3wLe/BuQdYknfWcCPVsV1+lUSA1uYdC1KCsUXNrPrWsY8buqB9hfUPfSbUH13Ur6Ceyup/hTf00rMkQtsB0+ZVZBXJXTqSzKDVRwqzkPZiZFfy/jwqvBjaHg0uc2AvckRAprKquxqlTTRxiV7jdDQ6bc/ejN9wbU3ThZPhMFmuCOMxDxHC4PI1fMeQS3VOYpyve8eQ8+fu/HgyFTFUbsc6TgWiqCm5f+H8aAQHDLZRtwFdyUnfR637I3+Yk6nfAruCqLcpTSKpPHODbiahTcRkSgyyqzmF5di4JL8wYIbhU1uPVAh0zFBx3zOa1zsPe0vb3XKVU7bi1htRZlz5q7kyu4K6esxB2n3eFTz6vCzmBRputAvQqueI0gIqciuPQeqYKVFFy3/twOCe2rGo1WcEWCTHW3rdMBOMC6h4CvTQO632KvewpuNQRXOK5kU+55i9lc6RjKmLSApfkC4QQ3rMTElgjuB24Hpq1gz5u7uEWZCK74ucjWZvk7le9nipxsUw2zKD/+I/ZIoUvTWKu9wGemOs+FWpRzrkVZMcA9aT57LBeYMi1DZa+1kqxGlXrLTl4obZsdtGqLCBBcUcGVUpTDphNV4GpJlVeDq7gWtUxj9t1zfhlvWQDw2TeA9/8+/vRhoO9gHIsyhWGNAOgaWlO5j5d6PTGyGZZOWooVk1cEhCba975CH5ps/2ehLcoaI4KyU/ZUtEKpgKHiEJoT7EZODJkiyF9EUcF99t1n8fjGxwGwm2nHcTBU4ApusVz0ETWZ4JYxjMECX365bHtKbdpOolwUbjCFGtyEaaAt4yq4luE7idDNfFQ7H5mkjqRFuVaIx23ZpGXYo2sPXoNbg4JLx4WWS8uoRGYaVY+pFdz4UEXxi6ABqUqgwYS43xf6bHb2Gty64cQPmZqwIJVNvD70vwvc/62gAqkCEVxRrSPSqjonEXnJtDOVRlWrKqDoHnu70mfgOMH6WhUaTXDFWldS94h4PvxdpkY+91/sfyKbcgCVSELDQqbEdfW8xWpew5L9Jy/mFuVq+xtbST85mncgf97UCfRvZseMVNeoY/7YD7jaDLDBgGRz8HsRsCgPskGB3vXAsd8AVp3DXp+5F3scko6fikAkm9UkrzDIjqmK4CYyQJvbOmiKIvhPSXDddXcuYUpyu9S6xbR5evS0lX7LM6CwKIukVjhO8jVEvH8RBzq8GtyY535SblUqs2Ew++6sveItq6Gg1gFmDII7O/r9OmAbNtpT7ZiSDUmhjkIdbYLGAicsPAG/OP4XgddF4UlWcFdPXT3i2zVa0AR3HEHsO1souwTXVSqK5aKitY///9tevQ1v9rIkxnNvPxcfvOuDcBwHNzx1A1beshL9eXZhTJgmCuUCWgS1qYzgaP/WQR5a4ZQTANjJOZtIopzjScJiyJRpGuhwFdx8sewbOWp1RxSjrMSBGtwYJxLPhhEjZKoRqaiiCvyr43+FbCJbF/EgkiOry5WW2TAFV4dMxUYYwf318b/G5/b5XGzCesy8YwCwvrlx4BHccZiiPKGwMyi4ZDETlcjbPw3c9zVg7V8rz0/ziUSMCK7KVUI220SGkdwKCu61pVYcPjCImVaF1PdbTgK+PrPy9sqEbEQUXPd6RySE+v4S2cz3+FU9X8iUZOsVt5eIUvdbrC4v7PqUnVzZohwGKxVOnJumMHu1U+L1qpUGOm8+lj3meoDt6xjxlAcZ5JTeQg548+/s+Zz9+evTd1evQzUQLdfpygi7r5jkBmJ2KtJ+RYJLvVJp3bscCyx9b5C8mzaw/GT2/LSbgIM+5X8/SsEVt1/eF1pP+xypHrlKi3L7bABGdX2RRwN0fyG3CVKhbWQV3NtPvR2nLDqlhpnHl0W5Voj3ESLBvfv0u3HB8gvGYpNGBPpuaRxBJLjDJWZRjlJw5f//+Pofcf/b9+Ohcx7yXnu953Xc8sItAIC+PBtppxrc5kQztgyxup0yghfNp7Y8yv9xuEU5YyfhFIX6DqFNkGUYnkV5IF+EbfKRutZkK3ryPVWFTMWxFO8+ZXfs0bUHJmcmV5y20QouEZq6LMruOBMtl+yrlQisSkVcPXV11fWJOmQqPryACulGf9nkZVg2eZlqFiVOWXQK3rfgfbFbVVEP5p29BrduVNEmaMKCFEM5pReoSD7ZfL3+ecT5nDJL6yUyBAi9LrOMMFRYx25FB/+xeUvlOsA37meP5VJ0O42Agtunni4uxONGyiIpakR++ySCCzC1lxJmfRZl4fmbj7B6UHldPW8BM/YIhkER0q3colyokuDaipApgtjflSyt8vE0TL/9mZwBN+zHjkPnkqBKLRJcO8O+I6/ezUgk2ZIBNiiSavWTXkCtVpoV9JiwtN3Ji9h3qTNCwU02u4Mz3Zy8HPhJ9fJMm7Xq+XJ3eIub9/078NajwDO/DD/2Mkj1lglztSnK8w8BLntmRG2+NYFs613Lg9Z3GbW2MYqJlkohV2GQWzZNUIj3wVk7i68f9HUkzASmNY2zQZE6oQnuOIJIWIfLw8iV/DW4RenCU5T7xAHoHWYXwdZkK3qHe/Hg+ge9m2MKjbIsB2WnjCab/8gdY8i3HLvUhUfevYe/7/CQqUwyBfiIHL/5MAWL8kC+5PshUX1iVHF/LSnKq7pW4Zbjbqk4HcBreuvp26ki3TQiVstyF3csxvNbnw/say1k5kfH/qjqeXQf3PioK6BCgGEYVfVhPnjWwdh72t74xJ6fqGu9Oz2cnUHBda8johJJlkBVOxgZqhpcIq2v3AF8c57/5p5aqyQysQiu0kIdhW2vsz6aMsplRsYaZVEe2OJXSgHBouze+PW4yb6k4Ipkc3CLmuDS83IZuPlo/zrzfawH7eBWprzJN84nXg/sejLw0HfYwEO5HE/BPezzjEg/8yt1myCCGAbVMY89rjzbP02yhSnUBFKzieSbCf9giGH52x2lmlkA1su3AytODQ5WfHZdkLzW4lRRhUwBwKzVwLO/VfdrFQkuXdcrkRfa/ig1ee8Ph/eTDkPXMuDTa4BmyTpLxyLu4LxhjD9yCwAzVgEX/DfrnVsJdYR1jiiaprDPQVXfPIGQNJMwYMCBg6ZEE05YeMJYb9KIQBPccQQfwXUVXBppUtbglsNvENpT7egd7sVzW57ziMuAezG2THaT15TgNhHH9BPc9NAh6G/+DX+hnPBqcLMJdiEw8vPhpN6ASHYtwaLcly/6rBAzmmfg+a3PB4i6iJRdvYJbDciOa9bhzlcRk3qUtRuPvBEvbH3BU7apPnO07Kh0jHXIVGWQglsvwa0WTYkm3HzMzaO6zh0SXoLtDvxdJ8VQJGpEcOMkDA+7JLiksCgTyiWusskKbv9m/7T3XAU8eB0nxUT4wgjuUDfw0Lf5/5ueVxPcv32HtcrZXSJk1RLcQo7N862FrKZy+kr2ukjWidT1vs0et73OvkMi2RSDplQhU1teDq57uB/oXseetylqcDMdrKYy1QLAYYMJMsE1TEbMxRRlKyEQtigFVyBTLdOBKzcEldBkk5/gyupaz1v+utN0q1+BTbW4rYj6ubVXhEqZJTInq8cizv8d8O7LwF2fY/+r2gQBwO7nAMtO9Cche9vq2pLtlP94RSHudZkId4xsEA8yuQUa13t1PGDhYfx5+xweNqbCSd8df1bgZScAH39ixBXmkYZlWuhId2BbbhuydoVSkQkMTXDHEYj4WYaFwcIgSk6JtwkqFwO9a+X/ReTci+C63nVe/d7gMBFcdsHICgouTClFueA/KTuCRbkpmQIwjOzWS/Gh93Th6y9u5IsxgDaX4A4KBNcyLHxpvy9hcfviyCJ2OYCqEZZiEaTg1tPOR0lw6wiZ6kh34MCZfFTTsyiPkh1VK7jxUSlkSmOcY2eowaVa0GKOKWtWQiC4McgfESiVgktwSvBuH7waXJfgbnnVP+2D17nLK7JtqaTg/uVq4HFhMGfzC0zFlLF1DfuTnUzVtgm64zPMTgoAG59hfwDQPJXvt9wCaWg7I7nFHCMwxZy/3YxIQmk/33qEv2ba7Hjke4FX7mKvzd5XcUPvISSzGgAAIABJREFUfk8pMCjXGyS4bbM5iSSIwVJim6BDrvDPK/a5TWSCAUkAI4ai65us0oR8r/+zTLX6SaBYb9kVs4zD6/2aYHWwlFotYtER/oGqMAXXMNTkFuAKrp2Obz+NS3Cbu9i0mfZ401da345AcEVc9mz0+3ucNzrbUQ1MC+iYW3m6CYDOTCcjuGG/mx0A+i5tHIEU2mwii74Cu6JUU4ML8BrQoRIbVV/bu9ZT5gZdBdd0CW7GCk+y29ovfTUcHjLFCC6QsFJoT/kbwZuGgeYUOyEPSAS3Pd2OS1ddGllbKpJH27AbTiSIzNVSK0tQqcpEfOpZLoE+r9FqCaMJbnzQZ6MJ7gTFTqHgCmRD7mk7FJ1wDIATKFUNLkEMSvIU3AoWZa8/Lym4IQO0eYmgiqm9Iob7mbqXlwhXtQru9rVMCZQh1qdmO4Pk84372bGinp1EcHM9wJNCyQypuWsf5K+RcpjvZ4nMM1ezG2eZxNAxomTcfJ9fHW6eCnzkf/1KLOAnuHaKkbyreoDDPiftozCfitwCfHDkff/GtpNSssmGbGf8n2W61W+nFe2c2Zj5EJ4t1wLO+DFw0V/U02WFWvCwGtwoeAR3BBTcZScAH3ss2BO4WuwgwUYaI4daVNgpGfbb1wRXY1TgEVzhy+oR3HI8gktEa6g4hKZEE4aK3HpMbX9Mi92cZKxgoX2xfzEGXv8EUPJfLJwyr8FtJoJrmrAkxXJWRxadzez901fP9shTXLutSHAbkXgsw7Mo10FQoizKjSC4hNFqCaNDpuKD3BA67GmCwhvE2YG/62Ko0ZCbhE8EZGh7cHoZSgVXIpFimQlZmpNNnOCqziW0DZUsynKK7/on1MsjIiyT9rgEt1QEXr7TXY67/OOv4++LdtF0Gw/WmrKUkdrX72f70jIdgMEtyr84F3jnSWE9w0zVfu53wvJa2Tzd65hivNS1asrXFo/gukQs38sHFAB2zLOT/EoswEiRHaOmNN3Ok7HDWpyd8p/A8pOAPd4PzD+YHe9ymYV5zdkfuPQhScFt85PAlHCfUSlcSNx+oDKZzEhhVtXCGwRIV0FwY577TYsnONeDCdaaRmN0cd+Z9+HuM+6uer5JafbbidvacCJCE9xxgA39G9Cd6/YsyuKIimdRdoJtglQNmS3TQqFcQLFcxNJJS33vkYJbdthyUgoFt9C7ChieAacs1Y04NhyyKKdIwTVhmZzQ3XThahyz61SkExZeveY4fOrIxR55iqtGHjjzQJyx5AzYht3w+lugMRZlpYJbh0VZxqgruNQmCFrBrYSy20e0EZ+zxijAcfzkiJTHrWuAR74/Nts0uI2phiMFsV0NkT8ivXK/URUoOCmqBlckoWKboHQbe294AHj4e8Dbj/PpiBQTaXviJ8BL/xNcv9z2Z2Az658qg6zIMmkfHmDb+6fLgUd/ADx8Y3BeAHjoOuAXZ/nJqNjz1FMeDfac1MdkMzBrb2Djs4xsJrOMZA5uYfXH6wSlFmDk77EfMoLStSt7zUww4vfGA+z/WXuzR/HaMnU3ltYLcJvr0Ha/gksEMCPVBPosyhEhR6bJVdwwBXfKLsCZtzBymm5nn+/AZjZYtMtxwOSF/DuXnQy0TGXLpUHkVA2JtbRfqoHoD97JSDfgT/OuRcGl5OK5+wvteCqQ6tE+99P2NLhcS2PHQGem0wtwrQbELaLadk50aII7DnD0fx2NI35zhFLBbUqGpyjnS1I0P5iylHcvgLt0+FMDc+7rw+6NS8pt4UPElf1jYVJTCk6JE1ynbMMpNcN2R3pbkuw92zJgCgT3oMWd3o1/wjJhGIan3MatpZ3fNh9f2v9LSNmphtffitsxUgpuQ+CMwDIjQJ+RtihXhlZwJxiubgfuupL/L55D7/hMsH5zNHDDvsB/hPT/bARKeXjBf0T+iBRVUnAdh5NhcUA1L1uUhXOFbFGm9dz5WeCHRwjbJVmUh7YBvzw3uA3iZzRzL/YoEmVvm0IIbjHHVN/Hb2L9f++8IjgvwJOQRSSFQV8aaE63MsJGwTLJLFNw+zbyGtxsJ7Mov/lwcJnFYeCVO1n7ljbXzmzajChvXQPA4P1gRXJ16YN8nUTEBrcCgiuLE1ypzlOuwY0CEdw4VkVaD9XE0iDAcd9gx+RDdwHH/Ku7be71mwhuNTfSUQR37v48WEzsY1sLwe1aClx0L3D4l+Ifr9FGtW2CNDRigIS0qNDXiQ5NcMcJhsvD3hdNtAyQRbnslAMKbk7RLqBYLnq25JnNM6Xp2Q0GLcc22AXBKQoXdcfElJYUnDK/WAy8fhkK3as9ItScYifahOW3KMt2ZQC+GtxqYJu2p7Y2EiOdojwhFVw9MhwbugZ3lFEYqjxNGIi8Pvxd/ppsfx2Li/uAmzI8UiUBpQIPRfIsyi65HIwguJueV6f/AsE+p0oFN8t6jgLAuy8Fl//TU4EHvhU85t1vSssWyDP1TKWkYRHU7zZAcPN+m7LYskaEKt02kQX2vZTVmhJhIhJFJDPZzI5vYYAFO9lpVmc5sJXZjeVzw6bnmGK/5GiBRNmc+HUu5iFSYaB1D2wJUXA7/NOLKcqVbMFUIxqH4FLtMH0etN1L3wdc/gLbl5ap/m2j/QwLelLBsyhXuAaK19taCC7ABlEsm7fWGaXuBbGxI6Uoa4wbkJA2GKd13ASFvksbR6BUZFHBzdpZmIbJLMpSKAcFSYkYLg+jz73wt6fbfWSMFFwi0oaTQKFndxT7hWRDx0JncxIo84uiU+gEnARslwg1exZlA5bwDRLtygTTMGEZVtUtbxJmYkQsyh6Zq4OHqraLiG1DQqbcG9+4NbjzWufVtT5NcOODVG6t4I4C1v0duGYa8Np9tc3vDQAKv0nZ/tpogvvOU8wWGwaxb+pI3VgU8yx8CAgquINb1MT6hT8ANx4APH0r+98w+QCB4wQJrnjcCoOsjtNKukqkAax/EgG8+yI7PjLWuAFCax8E3vir/zMi4lYIDuaGKriFIV4XvODQcGKg6k+abGJq5EV/4YSPUoA9gtvE2wYNbmXLyU5mx3bDM0DnEuCorwKz9mHTvPQn9rj4GL5O0+ZqManUIhYe7v8/1cKUvE3P+dVsuq6m61BwqX43zKIswlNw3UGJsMEDgKvRRGyTVdT6edfpKs6ztRJcwow92eN2xWBKI1Htfc2OmqKsMaaYkmXODZVgs6NgnA1V7dwoOuymISOcqNN2GpZhKUOmwqwFW3MszTFjZ9CWbMPmIaYY5D2Cy5bjOBZy75wDK7sGyQ7WJsGBhdZMAioGmLBs5ADYBgVHmTCJ2Bnh6qVt2lUTgoSZGJE+sI1o86JSlomUNjJkKs4x+/s5f687jGskwrx2VBDB1TW4o4A3/8Ye37jf3z8xLoiUieeRkSa43z+EPe5zkfp9sRfq4NbqbvoroZBj6meJgo8QVHD7NgD3fBk46ivq7SLylGrh6nm5iEAol5yinMi67VhaGMF7+zH1Noq9eQlEln7sBi0tOpK/l2xin19RoeR7NbhCXXEi6yq47nvpNv9n/OIfgTX3ACf8R4iCmwk+J0JCSqed5go5TWdarLY61wPMOxg48BPA/h8DvtLBPueu5UD7bF4Pa9qcSM3lLeIAAJ9dF1RTDYOt/5lf+V+nc7dSwY1pba3GopyWCG6U8ky/O7ImJ6uoxfVSlKu4B6g3DXamS3DffVH9/gdu97d6qhWfWRM8D0XBI7jaoqzROJy48ET0DffhrF3OGutNGTFogjuOoFJw03Yatmkr2wSFYcsQS3NMW2m0plo5wS3lkZr6B3SX2YitU7YAOHAEtRaOGZqmmzQTcBwTJffthM1DplT2ZIJt2lWTVdscmZApqjs4Z+k5NS+DrMNHzuE3Yo20rlZjUW5OVmH7CsFopTXvCNAK7mjCPafUWhtOCq74O1JZlB2Hkej5h4x8gMym5/nzwa3cFtkIXL830PMmS/lNZJi65im4OaYUJpuAV+8JElw6PxMBTbXy5GQaKJi6G7DJ7V0pW5RFYjhjD+CVO9TbqOpRK9vQ+3hfddhpRpBkBbdUFNoZCVbqRJaRTFLH021+MvGr89njmr/4+8aK83vPXQJMqispuOUiH0Cg9xNZ9nk6JWDKEva6KVwLKCyK7MKmDfS7+zlPIrhhfVOzk9kAhQj63BYeBqz+EO8fLPa+rXQdXXGqq/jGGOikY7B1DXtMRRFcaXnioEAleMFKVVxP671f6HQ/tzBFa96Bwc+qFqQjVG8VWqaxGu9RKlnS2DlgmzYu3PXCsd6MEYW+sx1HUNXgZu0sLMNCsRxMURbRnmpHoVzAQGGAE1w7jRZh1HSw3IPkpL+BlvLKxkEAGcARCa6FUjlIcFO2Cdu0YRQtFIrshjMphEyp7MmEWtTYhJkYkRrcpJXE0+9/um6l9a9n/dX3OTkNbDtCyxot4mkYBi7a7SIcOvvQUVnfRAYNYGhb9yiABovqJbjioJOs2JaLwD9+BvzhY8CpPwBWnhlcTqnArJLV3GyHgXqlys8bgR5XVSvm2U16pp2rm8VhRhYnL2bpvzLo+0wENNkMwGHkkAjkHuczdfjuL/lJ4/CA397a3BXeC1fucQswdVZ8fdNz/LmdZkRTVnCHFUowwLZjcIufqKtU+p63QuZXhEwRcfLs0kMSwc2w+Yj0EwkUsccF7rJIwbWY0rv2r0DHfPW2yMhOCr4m2lePvw54+peM3Mt9cKMwcy+1TVqF1hnsMyGreZSCS4R59r7AoZ8DVn843jqA2izK9Q5OmRZw3m+Bttn1LafRWP1hYDfFeUlDQyMSmuCOI+RK7IZMbBOUttKwTAslJ5iiLOKGI25Af6EfF999sUdws3bWp4IOl/03BS+8MwAgIym4Fj522GJs7M3jdWHaTNKCZdoALCyYwlTDI5dN9ZTbKIJbq0V5pKyzjVBZ2+WapwbmxZCCPlohUwDwiT0/MWrrmsg4c5czsb5/PT68WxU3axq1wSO4Nf644lqU3/kHey73U831sFCf/7cnsOJ04PSbatsOEaLaOBijZQ+hMMT+bv8MsOocv41XtQ4ryQiZZ1HOM9LaOp29VshxhfKZX/NjQESTgoFyPcD2N9hzO8ltuuUSe+/Oz7E0YtF+GkWowhTcMMJpp9QKroooA0DbHGafHdzK5rOS8W3oZsKvYsoWZfq/mGdE2s4w4m0n/aqr2LLnhO8AU1ew8CVAUHATwNk/Z/sem5gpppMHjmkwyGdRbuB11LSASQuBza4TIVLBtfj6Dw1Jsg4DbXOca/Up3wfWVN8HVAlS2scT7CRgKwZNNDQ0IqFDpsYYYmuW7jy7wZItymE1uCJMw/QSl7cObfXmFVXQYcdPcN/c6t5siYFSjoUl05rx+4/6rTiZhIUFk1uRSaSwqKsZT3/5aJy9z5wRsyg3JZq8/ZkImJxhF6CDZx3csGVqG+z4Q8bO4Mp9r9yhm6OPG1Sr4JYKwI0HAq/82f0/718OELQolwpM8QOCCtnPTmPkFgCe+2387Qb8bXTk9RGqUXB/egpw7Xy2Hc//d/S0pWGXdHX4Q6bsFFcer5kKrHNrnH93Ed8/IqAUDPSj9wI/cAOPrJTwmZSAR78PPPVzYN1D/oTcqNASVQ1uYTCYpEwgBXe4H+gXLMVkkSVMWgCcfSuw/ET2/+BWRkJNm7U7ijNIIocsybWPVLNLzoBdT+aviwOe4vdorwuBWYI6KtbgJjJqVTYM+d7gawGCSz3mknxdjQ6R6XSTspMt0TXkpMLWkqXhzRvjGrj7WcBpP6x+HRoaGjs0tII7xhBV2e6cn+AaMJCyUrANG0WnGKngGoaBJptdbLbkuEVZVHALjjTq7bCLh6jgJkwbKZu9ftPRN+HBN17Dd15kBPekXQ7DrHY2YtuWoX6yjNiaUQquYVetRl59wNUTKvxoWtM0/OWMv2Byuv6R1tFuE6QxgZDvY6ripJi2xokMI6IG9/Efsdf3FpT0oe3M3rr5edaSxVNwhd9RQMEtsfYuQDCFVQ5KGtgKNMX8fTsleOPHA1uYUtc+m5FPIomVCO6dVwJTlzNr8Jt/569vedW//fdd49/W4rBrUe4ANr8A9G1ySW/Kb6195HvA3AP868yLFmX4A3fsFD+W5RJ8iqKYBRDVlkYmuMlmdmyI4E7fHdjwNH/fSjIC+dKf2N+/vAH0bwJ+erJ/OXaatap54ifs/8FtPKAKYN+VSgOGckgRqe20P9NXsUf6zp30XWDJscD897D2QIRMBGmtZBeOgsr2HbhGigQ3ZpugatE+lz0uPjJafa4n/Ze+Z+OtZY+GhsaEgVZwxxiiKrs9z0bbyaKcttMwDAOWaSl73oowYHjEeJtbd5WxM36CCz/BdVyCC4dfRLJJPv0+0/fBgdOOZtuSsHDs/GPx+f0+71uGFacG10pUXU86r21eoI/veEdXtquhpFQruBoB/OQE4DurxnorRgdRCu6fLgP+53L/a0SeaCDQq8GNIrhFruBKbdi89FcCpTrHgTgYefMxwLdXsHUX82y5mY7KBPfhG4DffzT4+paXmVJXLgGv3wf89d+ANx7g75fyjFSsPIvZru/+Iq/LFQlux3zeCogwLNSuyrCSnHCUi37iElfBlWs5Wqa7BHcdm0+uBbXTfuLZtyEYtARwQkQq6+A2INHEw4ri2JQDBNf9PtD+NE8BruoBlp3grtNkKm52kj/FWFWDS6DrsewkiINjvxkMbpKvN6JFmci0Ki26HlBbocVHR0/nBUXVQFKtGmpwNTQ0NARogjvGEFXZDQPswt3m9pbLuDdYllGZ4Nqm7RHjbTlGcNOW36JcNtQKLmDCKbMLSlPSP9qbsNlXJJNUX2ioD67ZYIvyzoyxqMHVmCCgWsmdAXRzG7cGd3iAPRJpUyq4cshUgRNN+T2ZqPS6xOrOK1nd6qYXwrdFJNJkp33jAUaarITbN7UKi3JTF38+tJ0lAP/8dGajlkFq7S7HskTjvg0usU6yGlyCUw7aXvOSRVmEnRI+k5KfuIhW1WoUw5ZpjOBuWQNMXhhMlbZTvFYYYMdMrL8la7DXkibFpxMV3DgEV7YoLz6aJese/M+V5w2zKMug7aslOG3J0cAVkpVbJryiRXnuAcAxX2chT43EPh8BTr8ZWHl29HTV2IxD59W3qBoaGrVBnz3GGKKC++SmJ9GZ6URnhgV5EMG1TdsLoJLx6dWfxuV7XY7F7Yu5gksEV7Iowxr0z+zwi6NTZhfepoR/9D3hXmAyCfVFyvRCpsL30TY0wa0Fpv55akRheAB44P8GVbgdBWEW5bD9JYJL51SVgqtqEzTgKriyupuVeouSuvnwDaxu9cb9gad+od4WkVBNWcoeX/g9D4BKNgXb44ggck6gXqWE+64BXruXPZ+2m/89p8xVx2QTOy4lV8EVidjjNwPX7Sptt3vsUoqepVZSsCiX/bXNvpApSTE84kvBZREy7S7BfYUFMWU7/e9TmyBC/2a/VbdtFnv0CC4puEINLsDShcPqfAlyXX12EvCxx4CupdHz0X7QMqJsyHQ9rqYPqoiwOmGCp+C6bX/2/2hjQ6YAtn8rTqtMPmvpZevN637PtIKroaFRI/Qd9BijINjieod7sWzSMq8FSdpiF2vLsLzgKBkzm2figys+CMMwPCtwoVxA2krDNEyctctZHrk0bUZwB9d9GENvnw+nxEbp0wnTC5o6fz9/bV/CNtxpwhTcyiFTJy06CcfNPy7iKGiIoBpcY6R7cmpMbNz3deDer1YfgFQPSoXwAKVGg9QomeCq+pcCgoJLBNetoRRvxFU1uGSblUP8iAwaJvvL9wXnJ5IpQ5yOyOrgVlfBdesjxURlGTlJWaW+rouOYtv1xI9ZTejFDwDn/y44P6moRHCLrqprGMD7f8+I4HA/X64MVX9tWwqZEl1FPouyMKh6wX8Dy04K389EFsh1A9vXMrVUVs1lBbdvI5seAM79NTDVJehEomhaSo2m1//ncuCHYQm5Bg99qhXJFnZsKoVG1aPgqmDJ5JEU3HGQX2HVETJFvx/tYtLQ0KgRmuCOMeTgqGWTl3mqa9odjbZMCy9vf1k5v9yPM+NepJvdG5RdJu2Cf1zwD0xJc2taKTcbxb4V3v8d2aRnUT5uxSz/8q1oi3KckKnzlp2HkxedHPq+hh8ewa2zV6/GDgzH4Tf6sto3kvhqJ/BzhS12JECEUyYDqhpMgCuscg1uVJsgUQ2Ubay03qYuRirzfTyVmNDztnpbRKWYbMCFQUa8bCK44an4Puvw8AD7W3EacMaPgRO+zay8p/wnC2Vq7grOTyQz2SwouO5rCw4FupaFrxsIUXBTfsuvqED7QqYEBdNKKkgYmHX2qK8yUtm3gR2vzl2ABYew0CYKMgoouJtYXbFhMrJP2ykruAAjzyJB6t8Y3I65BwFXdTPlNSoRuBJME0i3+WtxVSBlPW7rIhX2/xivkQ4jj1YVNvGRQj1BUXR8tIKroaFRIzTBHWPIrX92nbyrVzfrEVz3JL/PtH3QkfJfQGWCSzZlucWOKQZml/0Xv45sEnDYa3IYVJIIbkL9VYkTMqWhodFgOGWupMZVORwHePhGnhpc9TpddShMtawGpWLl2tpSGMFVEBVAsChLBLd3A/DHT7LlyRZlmkecj0Dzt0xlRIoSrEWEEVxxWRR+NTzILcqmXUHBFXry9m9m29kynSmlu54CfPKZaOusJSi4+T63Llcgf1H9SwE1wbVFi3LJr/6GhUzZKTXZWnYicOAn/MFOnYvY9l74R2Dmnnx+UcHt38SOTbqNkUqP4Fp8eoJYgxsGW7By10NwAUZuKyq4dVqUAeCYa5j1GBjnBLeOGlynynObhoaGhgRNcMcQuWIOr2x/BQBw1f5X4frDr8ehsw/12uNQDW7e7ee47/R9UYb/Zk9upUPztAg1UeWyA8dNSraQAmBiUhO/AHa1prxWQXKtrKfg1mFR1qgOTtxQHY2dAwNbmWololziZC2uyrHxWeDOK4DbLq5tO3Ldlaf5+Zmsd2oUymXgq5NZum/kdC5JlH8Pfe+opw9YlF1luzjELL3rHgoSC1H9lt8r5NixPeSzjBDm+4LBUD1vRW97IceJbGGAh0xVY1Hu38zmFVXSSudbkbjR5yaGP6UrEFyVRdmSQqZ8Cq5YgyvmPiSCQUgAJ7Yi6SbVFuCqbZiCS/bxKAU3FsF1pz/uG8B+l0ZPWwl7vh/Y7czoaawGWZS9+lbp2FIK9XjIvKjHotwyjT3OPbBx26OhobFTQRPcMcSVD16JzzzwGQBAZ6YTh8w+BKZheqoskdU3+1g4xoK2BShL9W8BBTcRVHCvvetlbOguutOzC7pIcK89bSVmtbHk5iDBdWtw67Aoa9QGXYOrAQD41gLg2gX+15xS7XVqYRbfSuiNMd+rdzEiGQVqy/PoD6Kn8whumIIr/T6G3XRdOWTKg6FQaQWSJluGiznWBmbp+7iCOygpuHDUNclD3cD1+wBv3M9fKwxxBddK8PVteRV47r/884sW5e1r2WM1CqNoUfZeE9RNN6k/dF5VErJPwS37Ca6o4IokM8yiTPWuooIrBmCRamunJQXXDZmiUCfZqhsguBVqUak+e/lJLHG6Hhz0KWCP86KnoeNaS5sgEVaIOnr+74CL7q08ADIaqCdkqnMx8LHHgUOvaOw2aWho7DTQBHcM8fA7D3vPRaLq1eC6IVNUp7ugbQEuX+3v/SgruJ5FWbix+d79r3k9b1Mmu7GY7FNw09hj1lQACoJrawV3tDE5zYJWxBZPGjs5Aum/ooIb8zRON/9R6b1RqJUYy+hdzx4r1SsSAZTDn7w2MRK5pNflNkEEwwgex4JAggPkN8+PWaqZkU7ZogwA214PvrbmHtav9s+uSm3azKJM/WjFGtzrVwO//ZB/flHB3f4Ge6yJ4IrtewSCG6Xg2mm1xdVKCQS36LcoJ0NCpihQS7UOgBNdO+MPA/MU3JQiRdm1KAPV1eCqMNpuGa8Gt06CS/cLcphUpj3YS3isEKYyx0XnYm1R1tDQqBnjwMey80JU6ESiSsQmY/tTHWe3zsaC9gVYPnk5zvrTWWy+EAWXLMo9Q3STyKZLWWyZTSn/R0/hVJZkd8wkLHS1pDC/U31zZWkFt+G49pBrce+b92Je27yx3hSN8QqnxJXNuASXpq81lIoIbpT6Fwc9MQkuEVvZyiuS1FIecAftAm2CSvJ+GgqLchTBzQkEtwXY+pq6d+3bj7L6URGkuhKRbJ7GLcpUlxppURbCr7pdG7TKNhwG0aJMEIlm1HfGTqtJidwHN1bIVIhFmcgsEVy5766n4Kb86cZD24GhdqB1pjtfRA1upqOyetioNOO4oO9Tveutx/47WvAIriapGhoaow+t4I4hTOEmQ1ROLdOCZVheyBSByKxIQgMpyi4pJovy42tZT1y4Cm7ayrrr8BPSrJ2FbdoBW2zCMvHo54/ESatmqvfBpG0O2UmNqjEpPQmnLzl9rDdDYzyjLFiU494sy+FL1YIsyhkFwX3m18BVwuuFiHX0ujW0FQmuu39FiQiKRFQk64GQKYWCWw/BVYVMAcBbjwZfI4JLBKxlqj9kSrQoq5DvBWtfk+CKt0wCRRz4Sf//RIDCFNyodada1G1mxD64uV5/XXhYyBTtaxiIvMrq9LITgYM/zUiSOL9TYoQ/E6MGt2NuDII7ygpuI0KmgPrV0dGAlWADItrdpaGhMQYYx8N/Oz5EgisT1Y+s/AgOmHEAAODHx/7YC5oC/GRYni/l3lyQRfmljSzBc0FnO97Mc3JsSwR372l7h/bajYK2KGtojAHKgoIbRVZ880jE77oVwNwDgFO/H29+CndScYInfuL/P9cNJKapl0OErVwCHr8Z2OuD6pvgUogSG0pwZYuyRLKdctCiXJHgumQtrAZ39n7A248Ft717nbtNLulumQ6sf4Kpnh7BlZVphx+HXC9bp53iynmURfnIq4Gu5TxAjEhmmLI6qUpdAAAgAElEQVQapR53LePq7NQVwKbn+Pw0uPrfl/jn8a1HsihHXRs8giuR9xmr2B/AP5d0G1O2S/l4IVPtc1l9cxRGW8Gl7WyYgjuO1VHTHt8Ks4aGxg4NrbuNIcQ+pzJR/adV/4RVXewCv9fUvTyyC/gV3KRU31R2L5wtCXbhX7O5H9Pb0rj8gLMBALkyu+GyLRNHLuvC2XvPBgAcPudwXHvItVXvAxFb3SZIQ2MUIYZMyTWqYfAUUZe89LwFPPOr+OsceNc/vwiZMMn9YkWQgvvWw8CfPqUmiIBgNZb2T6y7LSkIbllKURa3USaxYTW4pSL731NwW9l+kzJLmLYbsO2NoBJI05GluZllHCDXHd4H19daqJetM93GLd1RFmXD8FuQVRbltKCwU0pt1/LgsqauAGbvC+x7CXDmLfx1JWExmF04O5m/JCu4UaCQqah9o+NCxxAAmqawRzlkSqzjbZsVQ8GtU0mtFqLFux6E1eCOJ2iCq6GhMYbQBHcMEaXgRiFKwS25F05ScF/Z1IfFU1twxJwjcN2h1+GEGZ90l2HghxfujW+ctrLm7Qd47a0muBoaDcRD3wmSKRFiyFS1Cq5M8uKC1qOyH8skO5LgrldvV2CZIVZjcfr1T/DtCrQJkrazVIi2KIvHkYhzQrAoA8CGZ4SZDaB9NqutFVOPRQy6JSIt09njUDdXcOVjJq4/18Pqd9NtwLDbR7dSyJRIdlQW5Y75/PlBlwOn/gC49G/MCiyiaykj4cd9E2ibLeyuEVQMV5wGXP6CP+nYroLgEqL2jT7vpi7+Wru7XR7BVSiZdkqd4Cxi1BVcoY9wPbDqSCgeLdjp8U3ANTQ0dmhogjuGEOtd5fTiKEQRXFJw01YaZ37v73j+nV4smsLI7pFzj8TU1CJ3GY0hpF7IlLYoa2g0Br0bWI/YW88Kn8an4MYkrLUSW3l+pYJbBcHNSWRQPIdtWQM8fKO7THd9sjos7sdvPgDc+1X23FNwQ4hxaThIaFR9cB0H+MfP2XNbIrgiKbWSPOworIXSEBFcV30sDoX3wRWXPbSd1SiLqmslgiseR5VFWSSrdhJYeWZQ+QWAzl2E5UgERQ6nSvhzItiypZApGWKQIQ1KRNUXk2LfLBDctlnufJJFWcZ4DZkS1ehaQJ/1eCa4+1wEnPzdsd4KDQ2NnRSa4I4halVwfSFT7g3EaTf+DR+79UmU3Js00zDxqBswddBibh8ruDcLdoNSobSCqzFh8eQt3C47nkDtV6LCoMJqcPP9wBt/Vc9Try3Sa9tT5ORTfo8gBhDJkJVLcbtuPRO48wpG8OKkKAPA+ifZY6WQqVKBvTfnAODor7nTCGSd5nvtXuAO1p/cV4NLIFJrJbky2xfxPTJMP1Gl1jnlot9uLap6g9sYwU0J7XwqpSiLaqVnUc6q3/fN5157WqYDJ90ATBVsy/LApayUqhK5iVybCXX9rbgMammz+kPB6Qj7XASsPAs44kv8NSLrdoqtRyZ6pPaON4LbNhM4+UbgrJ/Wt5yJkKI8eSHrIa2hoaExBtAEdwxhCodf7mcbBcsMpig/sW47/vTMBs+ibBkW2jIJvH//uTh8KR8tLpVZrVjDFFxNcDUmIga3AX/4OPCzcZhWTaRBrGWU4ZT5zblIGH//T8BPjud1myJE5VMkVnHbBokETFZxoyzKW9aw/qWq7ZDXT4So9x0+XZSCC3AF0uuDG1aD61qUExlgj/OD09A+iO1vqFXbrH34a1NchdNKAK0uwQ1TcGn7EmK7nhQnKOJxK0kKbnYSJ8Z2pnLqtFnBohy1fQAj7nRcQtchESqyYPumMdl0dsj3V1xGx1zgqh5gwaHh68y0syC09rn8tWwnezQMNu/03fl7n3oB+Nhj6u2VIarVo4VV5/rV6FowERRcDQ0NjTGEJrhjCJ9F2Yh/ofKSkA0bpmGiWOI3q9ReZtfOXVEuOwHiabshHC3pxlwYtUVZY0IjSnkbSZRLwOaX1O8RebSTfiIqz0+EU1RTN7/IHsmuK89DEMOZVL1dlesUCJhchxtlUb5+L+DbK8OnFQksKaI9bwtEVSa4koJrJdj2yH1wAzW4bsiUaXGCoEpRFm24RNJap3MyQaRIVHBfucPfu9a3fSlJSRV6w4r7Tut3HGZtzggEt3125XYrotWYBkcSVRDcOIOsUp/0UCu6SOIrLSMuxAAp8fn5v/UT87aZvI2QqBYfeBlwwn/w/0+/GTjmmtq2ZaxBaryucdXQ0NBQQg//jSF8KcpVXKioBpfm2djLb9LeM+s9ePbCZwEAJeepQPueM1bPwua+HC5+z8Kat1sE74OrCa7GBIKnfo5yiirh5TuAX18AXP4iT7QlDLsWZTsdnpDslITQKHEa93eosl6GtdcZ2AK0zqi8zb75KxBcmTT7rMBy0JOwLXQset4SFFw5ZEqa/5U7gWuEmkYi/AVJZS4Ns+NmWJys+lKUS8AD3wLW/IW/Jrad+dhjzP494KrRdpK3uXnxj0DrLChhJXhaMMAtyoD/uNHnODzAtjXTwV+LU7MpXkOoVY+dZCFSS98bMV8VgUWyRTlMIbaTCA2YMkdxXF3cp5VnMfL7R7dn8MIjwlXm8Q6t4GpoaGhEQp8dxxCO0FCylhRlenynW12rV1IouAnLxGVHLql2U0OhFVyNCQlP/YyZQNxoDLzLSGiuN0hwqQbXSoVvX7nEiZ84DamPlQhuvo8/j6vgiutRqaMiqKWQCoH0YIHAUruZnvVCH1yFRdlOh9co0/LlZOMtrzK7dPtcTgxE4l0qAPd+zT+PSIAmLWB/D7kqIBE46s+aC6k7tpJ+IkgpyvI20neSVNHsJE7+xRreMIhkRySXR3wxej5PwY2ReCwS3EkLWBKzCna69uCnKJz503jHwluXaNuWSHfchOfxiIlQg6uhoaExhtAW5THEsHDjVk2KMlmUE2YCz7zdjfXd7IY4k/CPrpcdxwuBGinwGtwRXY2GRmPhVJlA3GgQcVGtn+zFdipI7ghOSVhGXIIrhhhtVT+Pgji/rI7K+0E1tyqLdVQNLm13z9tCWJSC4IqKaNjyZYL76H8yQmtanKj5anAVnwUptCJkQnjxX6MJt5XwL0dUN8UwLiL0lLycmcQJTBxS52sTVIUyWatFee+LeA2yapmNtigDwPITgQWHxJ/eR/rtIOGdqDCrUN01NDQ0dkLos+MYIi8oF6bcfiECRHANJHDi9Q/hkCWs6X1Xq/+mplR2AhblRsMwDNZpQluUNSYS6u0JWy9IOVRZkD2LciqiR2yJEyqxBteIaVEWSe3AlnjbXC6wVN98b2WLMtl4SY32TVt01Wn3/KcimT1v830hkl8YAp74CVNKk02cCIowLLZ8x2EqdbI5WI9sWG5PV9tP1FXHWmVhJeJGjx1zgcmL1YFLQIiC65IrUfWl9dNyMh3AttfZczFNOQy0TMOs3P/VN18FNfCQz7LjCfgV3ChCbKeCLYUIqp61IwXf9ibd/w0AzuhuR6PRtQxYdT4wZ/+x3hINDQ2NcQlNcMcQw2HqTAUYhgHLsGCCXaBf3cTshmIysuM4KDsYcQUXYDZlbVHWmFDw1EgncrIRA9V+KhVcNyzJjrAoO2WubPoUXCKFiuWKCqwYDqQioSqUi4ww5nvV9a0iBrawbaB9kZeTagEG88F56Xj0b+J25XKBEd5fXQC847YECku/zbSz6Yf72THKTFIQXPcYmTYn13Y6uE9hUFl6023hSridAtLt/vmJHIrBVPQ5ihblxUexnsi7n115u4igVqtMVlJwD7uSPzdiElwrFR6KRa2BRgMiaae2RbTdE/malcgAJ98w1luhoaGhMW6hCe4YwXEc5EoRfS4rwDIsWAa7UPfn2U0htQACAHo60gouwEi0VnA1JhTGKlyKQAqoajsKLim0IizKPgVXnMb9HaqUYZFMiwQ37kBbqQikmoE+BBVcsTdtup0pk1+dDJz7G8V2FIBUFzDoKsc+BdddTq7Hb8u9blf/MpIhFuV0O5uXaoyzHUDPm/5pSCE1E3w/7LRaEVZZoZUEtxXY9pp6mywiVkl2rK1ECMGlGlzBotwylbXRiQNPWa4yOKmqGtyQOl8ZdhKA4prwoT8D01ZUtXl1QbYoAxPbmqyhoaGhEQua4I4RCmHpqDFhmzYst7VQrsjsiIUSJ7hEdkejNtYyjFEh0hoaDYMzXghuhIJrWuEWZbEG12dRdn/wKtIqLktUbWP3wXUVXICrnQ98C9j4rH+61pncervhKfVyRJJRyjML7EPfBvrcfrK5bqAcUt8JhLe/ybQz4pxz629JBRZB9cGmxT+HRMZvMU63AeffxnveilApnum2aIsyADRNAXrXuy10ompw3cGHSn1vZVB9qV2tgksW5Rg1uKKtN2r6ZDOU7og5+1a1aXVDRch13aqGhobGDg99ph8j5OXWF1XCMrmCO+wSXL+Cy56PikXZNEZlPRoaDcN4UXBVFmSvn2spOkXZI8kqi3IFBXdYILgqMrzxOWD9E8BeFwrzF5i1WNz+J37CWvqIaJ0ObH6ePZdJX7nMrMPigFhxGFj/JHDPVf5tzfWoa2iBaAW3XOIBU5lJwWn6N7FH0/YHeg1IFuNZIVZaleKZag22M5Knb+p0Ca6o4AoEt3c9cPOxjGybidqJas0W5Ri3A+LnFmVRPvYb1W3DSEFFyLWCq6GhobHDQxPcMUCumMMRvzmirmXYBldwCUUhsdRTcEfDomyMzno0NBqG0QqXeuT7wKT5rJZSRJwa3HIx3D5cynMVWtUmSEVwxeApkTSqFNzvHcge93w/JzVUOwtwgqsiC2Lbo27JHuztr0hwc2pyOLAVWHIMsNvpwC/P9b8XRlIy7WzfieBmFQSXrNFWgm+PnQaGhdZJqhpmed1yDW7o9C6xamJhgDBtIWRKsB//6VN8u+OESoWtp1aCW0WrOt/6VJi63P//5S+Fp0yPJHzJ0mI4mL5eaWhoaOzI0M1dxgBv972NoWLMQJMQWKYFUxqf8FmUHbIoj/yF3LZMreBqTCyMlkX5gW8Bj/8o+Lpcg7v2IeCqNtan1SO4BT8Bfv1+/jws/ZcIbqUaXNGiHOUm8U0nENyCqv7XRcc8/lxWd2m7lhzN+tHSMlSpz4UBRgblPsFAeAJu2g2ZirIoe8sQzp92WtrOKIKrIJLpCEJKNbHZTvaY7xMsykIttNjWSNWeqBI8i3K1Nbh1EuM4aJ3OBnpGG+JnTAM1VqK6lGkNDQ0NjQkHTXDHAP0Frp5ctNtFuOnom6pehmVYMCUF12dRdp+PRrrxRQcvwAkrI+rlNDTGG0bDolwuM7WQ6kpFyPbiZ37FHtc+wEmlbFG+5UT+XLQYv/Qn4FY3ZTesBvfvNwB//TdhfiHdWO4zmxOIlkjARAWXtlGlyrXMAK54i6maAYLrEsemLuCyZ5iqWczzNjQyrATQoji3GJbafpxpZ2SZrL/yNK0zgQ/fw56LJLkqghtTwbUzfB8AZlEG2HeCXgtNXk6rX4+C3L4o9nxVWJRFTIRaVtU2mgltU9bQ0NDYwTGmBNcwjH82DMMxDKPT/d8wDOM7hmGsMQzjGcMw9hzL7Rsp9OS5LW3Xzl2xz/R9ql6Gbdow4b+RKZQUFuVRUFYvPXQhDljUOeLr0dBoGEaD4A5tZ2Srb2PwPTlkihRMw/RblMPC6ApS+x2qeUVIm6CXbgcG3uX/e62IMkEFd+Mzwj6IfVoLvJ8rKcgyOQYYwUq3MvVUJMhXtbHaXpoG4EnRooIrtqIxbUaGDdMfumTawGffAFZ/2L9uSj32esm2+98/6FPA7L3dZQjnT1n1jFL4VSFTKksx1QnT9Pt/FJi1N7D7uXzdct2vtx+1KLhuj9daU5SrtihPAJJoKJR+K1n9vmpoaGhoTCiMGcE1DGM2gKMBiEVaxwFY7P59BMCNY7BpI47teX7Tl6r2ZsQFI7j+i7eo4JZGMWRKQ2PCYTQsygNuWm//JqbmigjU4NJv14hXgxumuoYpuPL/VIObauHLWvsQ8M4/gO1r+XREUB2Hpx8nskwh3fQCI+pdUgsfUkZV9uBn3bZBpKzZSbcGWFBwm4TBMtNmymJTF9A2y/86ECSmRDqHtgMwgkRR/F9U9+Tp9r0kuO3yOiopuJT0TNO1zgD+zz2s9Q+9Jg46hG1nNbCSdViUG1iDO15gKm5xLHtibLuGhoaGRs0YSwX3OgD/An8vgZMA3OIwPAyg3TCMHc77Kiq4fWKwSRW4dPdLsar9fb7XimUHjkts6X5ahz9pTEi8+Efgmul+K241uKoNuOfq8PejLKhhdtlqQeTFKfFgI4Jcg+vxW5nghmynrLrm+9h2GyF9cAME1z2uaSH998fvBb5/KJAXAqjI6kvbadqM4D76feDG/dm8y44H3v+HoBKoshB7LWmoJ2mKLUP8PERiTNNP2w2Yvoq/riIu4roHt6lV1TCCS6Qw2Qx8cUt0CnBYH1wZnoKrIFOeRXlL8D2A25urhZjQHHsehSIdBxPBoqyClZwY6rOGhoaGRs0YE4JrGMZJANY7jvO09NZMAGLR1tvuazsUuvPc9je/rbbgjfcteB9mpHYNvE4qLg+ZqmnxGhpji7u/xOo8e9+pfl4iqA/+e/g0YRbltQ8BX58Z3tO0GlC/VSBoUw60CRIUXAqgK5fCLcoyYXVKzDbsKbgSMZYJPdXQplqCKcpiwjIpuDQ/EVwRVhJYcAivGyXio1JwPWIr9GzN9/sHMkRiTNOf8wvg+G9zyym9LodT0euDWxnplD9nkThaipCpVCvbtqiBQRUhzCpKNIhMqxRVWka5GDyeAJCooQYXcBVvbVGORKoFSDWP9VZoaGhoaIwgKg7BGoaxBMwqPNVxnBWGYawEcKLjOF+rMN89ABTxl/g8gCvB7Mk1wzCMj4DZmDFnzpx6FjXq6M53Y1J6Eu467S6kawkTcVEsB5WmYtmBbY1uyJSGRsNBBK2WtNM4LYDCCO72tay+tW+jusVMNRgQ1Lm+jcD0lfx/qmFV1eDSvke2CVIQ33yv0Ae3GotyFMElBdddn2kHe9DKxJaIn+r40TSexTgNvH4f+/OWlwRSbUC+J7hMKwEUS0GCu8/FwKpzmMUaAIa2sX2TCbBKwTVtQXWWanZVUFmUVUnPskXZtwzhtexkoEdyKtSj4FbbPzfdBiw5Dpi9b/Xrmoh47/8d6y3Q0NDQ0BhhxNH3fgDgcwAKAOA4zjMAzq40k+M4RzqOs0L+A/A6gPkAnjYMYy2AWQCeNAxjGoD1AGYLi5nlvqZa/vcdx1ntOM7qKVOmxNiN8YOefA/aU+11kVtATXApaGo0Q6Y0NBoOUTGsdd4oiDW4oiWZ5i1UYY3ufxf4y1eCpFmsr+yXFdy8f320DYbByWiURVnVuzbfF94mSGVRNhOMYAYI7gBrt2PaQQXXSgTrQ0mhlO3HkxcGt5EIuEdcFWTMtIFsh3+Z3nvu/6TkEoHtXAzM2EOwKG9lamxsgusuTwyyCgNts6jMqlTahJSi7FuGqP4qlO56anCrVVZNCzj3l8CcnYTgTl6o/m5qaGhoaOwwiENws47jPCq9FuMOUg3HcZ51HKfLcZx5juPMA7Mh7+k4zkYAfwDwfjdNeT8APY7jKHpsTGxsz21HeyqGUlABxVKwd2TQoqwJrsYEBBG0WuphVepmYPkCGRWnj0Nwi3lgg1BdccdnWAue1+71TzewGUi6bXUG5BpcV8HteQvoeRs+izLte1SKsmofr18NvP6/6vfl/4cHGLm1km6KsXCc8/1M/Uy38xrcUoRFmcidKRHclWcFt5GU67CQKHpv2kr/MgmWpACLyjcg1La6Cm5SsqKGElz3eTqOghuzZtUjuCoFtxLBrXHws2Me0DFK/WZ1ErGGhoaGxjhFHIK7xTCMhXDvwAzDOB3ASJHO28EU3jVgyvE/jdB6xhTd+e6GENxCSaXgUsiUtihrTGAQIasl7TiWRVmYRuzlSsSXiJgKT/8C+P5hQsKwS7Lyvf7pBrYA7bOD6wC4avrgdcB1u/oJJpHJUjGcrMshU4H3KxDcwgCzstopti2iwjvcz4hhpj1mDS4puJKdONkEnPdfrC2Ot95B/zQqC7Zpcbtsr2Tg8Ui0pOASwSWimutmNbhLjgFO+A8+v7jtHsG1+PbEsSgnm5itt3VW9HRR6cSyRVlGrRblD/wPcPgXapu3WkzUGlwNDQ0NjR0ecQjuRwH8J4ClhmGsB3AZgEsbtQGukrvFfe44jvNRx3EWOo6zm+M4jzdqPeMJvfletKUUbSWqRFFuPQKt4GrsIJBrU6tBHAVXJM4iyYqj4Pasd5OR3SAqqrWUE5/7NwPNXUwpFQmz4wQJL+2nqNqWowhuSG2u+P7bTwD/+BlrSUTTG8IpX1Rwxf0dHmAkLtOhrsENsyjLCi4ALD6S1cYSZAVX1QfWtIE5+7PnfdJYqiUR3Bluq/TOJf73AWZRNgxgrw/491lcDz3S8zgW5UQa+NQLwIrT/K/LajEdD1Xok7hOkeCSglyrRdkwogOyGomJalHW0NDQ0NjhUbHAzXGc1wEcaRhGEwDTcZza+tpoeBguDyPZgNFvsQbXNICyE6zB1fxWY0KCiJ1iEKciRFvvHVcAj9wIXNUjTSMQXJ+CSwQ3QsEddElZzl0mhS7J8wy8C0xawMiK+J6qfpYsyuWCvwY3zKIs98GVUSoAPzycPTdtfjytJN9fqtcs5v3kfLifEVwryeuIfTW4IRZluQbXe18glTLBHQwhuDP3BA7/IrDi1OB74uOe7wfmHsBqcOV1p1qCyxa3XdxeGmCIY1EG1Cm82cn+gC5Z0RZhGEDnLsDm5/0EN+PawmsluKMJTXA1NDQ0NMYpQgmuYRiXh7wOAHAcJ6IHh0YUSk4JFoWk1IGiYFFOJywMDpc80kuOR21R1piQ8GpwayG4gv34kRtDphEJrkA44yi4Q65yS5ZkIiOFAf90A1uApinMbuojuAryTD9YeVtqVXBzvBUZNj0vEGXhfGCn2F8pH1Rwm6YwMrj1VXd9gkU5kKKc4u+Jj/L7tGxxmjCCaxjAez6teE9qE2QYnNwC/rpQVW9asbbVa49jc2Iax6IchqYpQPe64LYYIUapqbsygpts4q9lOliSd50BhKMCXYOroaGhoTFOEWVRbnH/VoNZkme6f5cA2HPkN23HRdkpwwy76akCYshUOsFu/EplnaKssQOAiG0tNbiq5GH5NSeM4MaowSVrco4IrsKiPDzACG9TJyPAxZgKrkg0y6XaCa7Yx3fzC2qLsqfgDnPiCbA05mRTDTW4ITWnPgVXqsFVKdRmxOAfbX/YAKHYVkpVBiLWtpIKa1psn4F4FuUwnH4TsOq84LYi5Bw8Yw/2KA5G1GtRHk1oBVdDQ0NDY5wilGU5jnO14zhXg7Xq2dNxnH92HOefAewFYGI1nh1nKDvlxii4gkU54xJcCpmiGlxTE1yNiYywfrWR8yhIk1zzGhoyFceiLCu4LoETySlZe5u7GCEUl6daNhF68b3IFOVKBNdVRg0L2PAMf110dFjJcAWXanBzvewziKzBpT64YRZlQcGVLcpHXh3c9sjWUNRmKOT8KaqKKouyKVzyKOHatDnBl+toq0HHPODk7wovCK2fVFj9IeCATwD7XsJfI4I9EQhu1EDEeMKeFwKHfm6st0JDQ0NDYxQRp8nkVADi3dSw+5pGjSiVSzDN2hXc25/dgHVbB30W5VTCdJftT1G2tEVZYyKjUSFTxZy/brJcIWRqWLIbi/Asyq7q5yjU136X4DZNYQS4Ug2utwyJ4Naq4NI2zt4XePNv0nr+P3tnHmdJVZ7/59yll+mejRmGYdjBAVR2BRkFdwSNiqLGBVGjxqigqJEsRhPU4BI1iUYN+nOPGuMGoomiJooLOiq77ODIOgPMMEvPdPddqs7vj1Nv1alTp+rW7e57e+b28/18+nPr1q3l9MJwn/u87/MqADpycIfNz7hhRStIivLIMnPc9PYOPbjO2JzCHtzJ9DGnvMWcd/k7kmOKBK47R9dl3Ppfk12i/OIvAzd/L32s/D2oqhmNZO+bC3QHgVsfAZ7x3vQ+EbgzTVEmWZ77sfleASGEkD5TRuB+CcBvlFKXwLwzOhPAF3q5qEFntg7uG79yNQDg+cfvF+8bqYmDyxJlMkDMaEyQLQojMee6pnouQqYiB1fEctPj4I7t7QmZ8l3bEcmq2l3I1GNeBVz1hewa93mUI3ADIw7DlhGWIk6lFBkwP4+h8URsTW+zenCr+SXIlbxQJevfINfBtc8XCgWujAPK+fdzr0OT7WFL4D7yOebLRnpfw1Yi8H2u74yxZhuXJXZw94AeXEIIIWQ3paONqLW+CMCfAdgKYAuAP9Nav7/XCxtk2ro9JyXKt25KXJcRx8GNS5Tp4JI9mRmFTFniVYKErvkysOn3/mPscuVOIVPNyUQQS4myiGU7ZMoWuLXR9PVaTrm0vR4RgPVR4/RuvSt7LJDMwV19jHl81oeBFVbYkgjWsVXp83RoCdGhpH/WFriAcTIlcGlqq9WDW0/P7JXrAPkO7vgq4ORzzXbT6cGV79WmVIlyzjGpHtwOYlXKkdsN4OBTzPbiNcXn9Br5mbsuOSGEEEJKU7ZONgAQWl9khoTRG/bZCNx9l5pP92/auCPeNzqU7sGV6Sp0cAk+ehzwk/fN9ypmxkzGBNllvSKkrvgAcPETrOvaJco+gZvj4NqpvzImSM5p7DT9uRcuBX77/8y+OGTKErVuPzCQlByLEK6PGsF882X+dYiD+9x/Ay6403yfvtFj43unn4dBWuDGDu629HHSgyuvyfddqWVdddfBdcWnUsAZ7wOW7JctUbbPFwpDpjr04Nr4UpRtRAC3pkyp8PnXZX9es6FTibIPCZnaE1KUCSGEkN2UjgJXKXU+gK8AWAlgFYAvK6Xe1OuFDSpB9OZwNinKByzPfrovJcptSVHWUhKzOeIAACAASURBVKI849uQQWHrBuCKD87d9Ro7gYlNc3e9Imbk4FoCN8/p87m29naewJ2y0onFwY0F7gSw+TazvekGAMoIFTdkqlDgWg5uEXJ8bcSIaMCfaptxcINEHFbriYM77Qrc8URsXf0l4OvnJOe4v5OaNW7HfnSp1tNhVfH5XZQoi4Nb9AHhEX9iHn0pyjaxgztt1rb84OLju2YGJcoiuveEkClCCCFkN6VMD+5rADxOa70LAJRSHwTwKwD/1suFDSqxgzuLBMq2x9WSMUFtJ2SKJcpkzvnp+4E7fgycu77395rtmCCfq+let5sS5djBVdke3MaOdPludci4d/WR9PW8AjcSfiJw7ZChtacDt1/uHB8JXPvfEa+Duyq7L+Xg5pQo10cTB/emS61zq9kSZRGoeWOCfOuzRazbbzqbkCkAeMFngI3XAWMr8o8Bkh5c3+9jLpAPArr5MPOgxwPHvhRY9ajerIkQQghZAJT5P6+CKVEWAnT1kTSxCcLZO7itQGf2xQJXxgQxZIr0isktwK7NwO8+Bzx0a2/vNWsHN+eDpI49uHklypGDu2Q/y8EVgTuRXq8IuvqidN+trwdXkpXjEuWR5NxHnZk9vozArY2kg5aEVA+uJ2RKXpN+0NS5dTPrde8jrWOHk9fs67vY67NFsJsYPJseXAAYWgQctK7gGhFzGijl4ZiXmEfp7y3D+Crg+Reb74EQQgghM6KMyvo8gPVKqQuVUu8G8GsAn+3tsgaXuejBbbZ9Dq6ETKVLlOngkjknbJuv/347cN3XZnetr/wp8LvPF9xrJinKlmDNE8gpgevZznVwI4G7/ODEwdU5AldKd+tRyJQ4nz7HUO4XlyhHAqc65HcAReDa/45UHdE3vNifxmunHec5uOLuumFHlRqwZN+0ey8jz6olSpTdNQDpObnA3PXgdmI2M2/LcMipwIXbgRWH9fY+/Waoxx8MEEIIIbOkTIryP8OkKD+MJEX5X3u9sEFFenBnI3BlFJCNOLhJyBQdXNIjgpYRWDroPI+1E7dfDnzvLfmv55Uob7zehDltudO/vvh8578VmW+bKlG2ji/bg7vsQDMv1j5HhL8gjmVtBIBOflY+gSvpwiJ0pey3NuwXcxIyZQtF18EdXuyfpxr34A4lonPHxvQxIjpHHBc3r/wYMA6uquaHKlVyBG43KcpzKXDncubtQuJtNwF/tWG+V0EIIYTkUiZk6jAAN2qtPwbgBgCnKqU8tWukDHMRMtUocHCnmkF0Hwpc4uEX/wp857zZXSNsJwIwyJnTWga3l9O3P8+Bvek75vHar/rXl3e+jO/pGDJV0IM7stSII/kZxALXEfx2ibJ9TRG49odcMmLIDZnq5OAWlSgPL/aHFaUc3Oj1Hfemj5FrSR9ufG6BsKzWiwVwXg/uTEKmCo8pSa8d3EFlZAmwaK/5XgUhhBCSSxmV9S0AgVLqEQAuBnAAAM+7SlKGuShR9jm4o5GD+1ffuh6XXXd/3IPLEuUFjisif/wPwDX/Mbtrhm3ECbHhLARunktqi+a8MUHj+5jH7fcUn58RuJuj6840ZOphYHQv43BK36xcK2yn3dmqVaIMJN+v9ODaYrDplih3ErjRvVMlyo64HF7SQeAOAauPSvbbicvi4Lp9uLYLK78DYen+pjc5D3t9qR7cbkKmop/FHMwR75nAff0vgdf/ojfXJoQQQkhHygjcUGvdBnAWgI9rrS8AsG9vlzW4xCFTldmETOWXKAPAj296ACEdXALMLKSpE7YgnI2DmycibZGYt35xPLfdnX2tjIOrO/XgFpQoL9rLOJ/tKfMBQny+Tp8nItEVuO1pIy7tDx/aIn6lRHk0uYbPNZXvq8jBXbq/31G1S5RHliJ2RQ84KXutjINric9z15vZscLJ5xYLu9IpymV6cOfAwXV7lueK1UcBq4/uzbUJIYQQ0pEyKqullHopgFcA+F60r6AOjRTRq5CpYUvgjtarEA1cpYO7sOkmpGnDz0xf664txccFnp7VmSD9sJnrW2W+eT24EvDk68ENCxzgG77hiFL4Hdz2tN89ntwCLFqRiNcNVwA7H0het78nEZc+gVsb9Yt3ub+IvupQsVtpi0FxV2sjwPhq4LT3+s8RYS3re0Z03P4nWmuPxGhRD+7o8vTs2GqtOP1XznX7dOc6RXlP5EVfBF5YELZGCCGEkNKUEbh/BmAdgIu01huUUocAmGWN48Klrc0b2NmVKGd7F0dtgTtUTebgztwoJoNAN3Nkf/kx83jfVcXHpRzcpinb3fT77tdmi8HGzmS7jIMrI3p2PZgELsVrstfXSLaPeQnw+28BD9xoBK4IwjzB3va4uJNbTYmyiNYvnQnc+b/J67YrHYdMjaa/r9aUEchF7nrNFrgF/xHbQk/E7ilvA/7yFmB8b/85cl9Z37rzgL/bBIyttO6fV6I8i9Lgas4YoUyKcpmQqTn8h23NCXN3rZny6OcBR50136sghBBCBoIyKco3aa3frLX+z+j5Bq31B3u/tMFEHNyZhkxprdEMQiweSb8JlJApAKhVFEOmiKEbB1f+JjuJYrdE+Qt/Alz8hO7XZovB9+8H3He12W5bojRv/dPb0+vRGrjth8Z1tR1c2w0+4ozovlPmexQRmRc45StTdh1cl6ZH4IoYFkHfbhiHtujnbI8JKizZtR3cavJYVLnhClylzBrt/tq4RLmgB7db5Jpu2bS71sIP/+bYwf3b+4BXXz431yKEEELIbkGuylJKfT16vEEpdb31dYNS6vr+LXGwmMmYoF/esRn/d4spgxT3dtmi9JvEkVpyvV3NdhwyxRLlBU5eCbEvwTgWuB36dl3H88Gbou0u+33dEuWN10bXt0uUc64pJcqAEYo3fhv46ouA334mvy94eEmy5rCdiNS8mbhuj3Bz0vT+LlruH78DJKODgETQDY2lr9eeygYruYgornUqUbaEnvz+Ook/Edau0PT18xb14HZL7OB2+Lev6N8seW0uQqYAk4ZdG+p8HCGEEEL2GIpsxPOjx2cDeI71Jc/JDAgjEdBNyNTZn1mPV3/hdwCAZtRcu3TUEbhWifLEdDsOmarQwV3Y5AlE3/zaIoH7tbOBKz5ktt0SZcFX0itc8U/AZ5+R3ucKSBFPZUqUXQdX0pE335bv+saCtmWOqQ4BUOUd3Lt+aR73PS7fwbW/JzlG0nqlDLvdKC9wq8PFgs8Wi6UFruPgxteyR/fMYA5uJ+R+Phf47bcDj4z+t5I3PgqwUpT57xohhBBC/OS+E9Jab4we71JKrQZwEsxskN9qrTf1aX0Dx0wcXJtWFDC1bDT95tQuUd7ZoINLIvLEXruRFWnyt+ITGJtusARiTopyczJxK11+clF2n+vgyn8Tdk9t3vobloMbholgbE/ljy4SAXfpucD2u4GlB5h9YU4Prru+2y43pcMHn5ruu019T3aJciTkhiOBK+5uq4SDK2utFZQoq0pa6MXir8OHZ3kC1xavcn8pUV53HvCYP5tlD25OiTIAjK9KyrJRIHClRLkX6eCEEEIIGQg62ohKqdcC+A3MmKAXAvi1UurVvV7YoDLbHtxWjoNbrybX27R9Grc9MGHuQwd3YZPX52n3uQpFDq4OE/GXF8qUN/YnD1dAiniyneAyJcphOxFHren8EmURbduj0UKqEglcx8GtOanHwh9/Dhz0BNM/m9uDa5coi4M7ln5NenDLUB3OL8d198f/phQIxJVHZFOUhVS5c/TvhpQoj68CVj6i1JJzketLqXiGgg9Y3HUVHUMIIYSQBU2ZhqoLAByvtd4CAEqpFQCuBPC5Xi5sUBEHt6Zm1svWiBzcpU4PbrWicNCKRbhryyRu2TSBWzZNxPvJAibPAQ26FLhhkFwrz8HtVuC6x4toadklygUhU/Ux0xMbthOx1ipycB1BWKlFAjcAfvQPZr5tGADDi43IdgVuY6cRekB+D64vRVlKlOOQqSlgJE/kRcjPolaQoux+P3Jc3u/8Db8Clu4HfPLx6fXF1/P8myQCd3hx8XrLIPdbkSOUYze6RIlyoctLCCGEkIVMGRtxC4AJ6/lEtI/MABG4s3Vwl3kc3CsueArOOn6/1H6WKC9wunJwo78VnwOqg0T45vXgztbBFdHdycENA6A5YQSprC2eXzuVHhNk4wq4StXMbg3bwG0/AG7/kdkW8dmaBO76lZkNvOkGI5zlGnkOrP09SXhRpWoEcSP6Z7RMD65dRpxXFux+P51CwvY+AhhZWq4HV1h+CHD6+4FHnlm83jLI9fc6JOeALv6tYokyIYQQQnIoYyPeAWC9Uuo7MB+bnwngeqXU2wBAa/3PPVzfwCElymV7cLVTipcXMlWrmjeH4874IM7BXeAU9eC6iECyQ57s64iIzIQyKQA63X9aBlfgyppsB9dd/2dOM7NvAeMubr8nSUUGIgfXEbirHg087xN+QSglyo0JIzrDduJWtqaADVeY7d9/24j5eLZtCYFrC8jh8eS11lQSImVTqUfusxV8NZMS5TzxJ0I5FrhOibKvN1YpYN0b/dfrFvm9LT/Y/3qZ8mOWKBNCCCGkA2Xkz50ALkVSE/YdABsALI6+SBcE0Rv2sinK0630m9VW2/waFo84AjcqRR4fTr+Jp4O7wMkTOz4RK4LJl7CsrRJl2+ENLFfTNzc2cx1LmLiOr6ypKEX53t8AW/9othetMI9hmKxZSpRtMbvvscCa47MCN2iafUHLlB83JtIC97qvJqJ06wbjDIsIzBO4LU/IFGDKlO0eXF8Pr5QDV+uJi7xk3+5LlDu5m904uHPJtqj3eekBOQeUKFEekxLxkj3MhBBCCFlwdHxHo7V+NwAopRZprbu0aIhLtw7uzkbiRv3VN6/DYw8yZZnuHNxaFDIVOM4Ge3AXOHlzcL1jgqK/FZ+7G4ZJubNboixJxK1d2fN86xHh5zq4rQ4C152zu2S/5JoiutvTRojWFyVJyyIEXUHYmk4c3OaEEZ214SQE6Q8/NV+AEdVBs7PATTm4logdGrfGBE35e3gX7WVczkodOO5sI1iPeQnwwA3+e+UKXEcgPv7NwJUfS57Pl8BdssY8rlzrfz3WtwUC99n/DBy0Djjw5DldGiGEEEIGhzIpyuuUUjcBuCV6fqxS6pM9X9mA0m0P7i5L4H79d/fip7eZMr9li+r45uvX4R3POhIAMFwz19u2K90/qejgLmxyS5R9Dm70N+kLoNI5IVNhuzsHtyh1WdbUyunBndiYPl4Ek12i/NAtwLVfTpcAxwLXKcFtT5vXGhPmPq6Da/PwhkjMR9fwlRgDzpggt0S5rINbM+s6/uVmO+/DsLI9uM94L3ChNTc4r0S51wL3jA8Cr/ofYMVh/tfL3H9kKXDiazkHlxBCCCG5lHlH868ATgdwGQBora9TSj2xp6saYLqdg7urmXbgJpvm/OFaBY85aC+sXbUY+ywZwcpx84Z5bLjHb1LJ3KI18NCtwKoje3T9PIHrcXDFQvO9FrYLSpQjYeU6sj6CViIO83pw2zk9uNvuSh8vAlcHWUfadljlvzVXQLUbRrBObYueTwGtYb87Ox0dE/fgujOEK0Y4tjwhU4AZFTS5xYjo9rSZL3vGB4AfvitJfY4FruOs5n0YltuDm/M7F8rMwe0Fw+PAwU/If/1p/2Aej3lxb9dBCCGEkIGmlI2otb7H2dXhHRTJo9sS5V2N9I9aBK7MvV26qI4zj0uSk99++uF45lGr52KppB9c8x/AJx8H3PmT3ly/Gwc3TiLOCZnylSiHXfbgphxc53ivg2utX3pvASPuxvdJrukmP9uCt5IncKfMPhGvgNkuEnrV6BquCJZyZPv7swXkUBQytfk283zFWuDkNwDnX5ccIwLXdZpzU5Qr/uM69uDKHNw+lyh3YtFewHM+Wn5GMCGEEEKIhzIC9x6l1OMBaKVUXSn1dgA393hdA0u3IVN2iTIATDkC12XRUA3nrDtoFiskfeX+a8zjljt6c/08N2/jtdnUY3ESi0KmtE5fM2glTmKZMUG2EHX7g20HV0SeLdZsgTu2dyLQwjArcCcfTgRb/OgIRR2a16a2pvcXCb2qNfrHpjaUfyxg3MvGTmDz7eb53kdEx1hi1i5RtskNmXKPk4ThsiFTfS5RJoQQQgjpA2VU1usBnAtgPwD3ATguek5mQNcObqZE2TwfquX/6kbq5a5NdgPETetVT6EbzCT8/CPA996S3hcHNTk9uFobUWT3utrnxCN6cgSuHRqUCqhqpUWgzL9tT5uSXiDtQD90S7I9vipxMMN2Is7jazcSkSwC0SfgKlWPwC3478d1V4Wqp6fWdXCntwH3X2s+EFh+SHZNUrqdKVF21jM07t8v11zW4QOu+QqZIoQQQgjpA2VSlDcDOLsPa1kQzCZkCkhKlIdyHFwgCZwiewIi/nokcIv6MTdel34uAtcNmRJBpIOsUxq2gCASWnlzcFM9ta309tA4MPVwdJw1B3dozAhC24184Cbg0KcAf/gJsHh1Ish8JcqAcSjbU/klyrLPFeaFDm6OwPWFRrkCtzUJrP93I5LF8bWvJwJXBGy8Hue/56ExE1jlrvOoFxhn+5AOEQnz1YNLCCGEENIH+JF9n+l+TJC/B5cO7oDQcwfX+vtx3VyZIxu/Lg6uVaK88Xrg+v9Kzncd3LDt7521sfcH7fR2fRSQl+M5uFOJgytirDUFPHwn8OjnA/f+1vTfisAT4V0fA85dD/zrUWZ/pkTZ88+dT9SVKVEus98uW57cnGw/5pXWvaz7n/BKYHx1NojJ/TBsaBzAA54xQQo49Em5S49hiTIhhBBCBhi+o+kzbW3e4JcVuJMZB9c8z+vBBShw9yx67OC6I31sFu2Vfm7PkhU+d3ricOrAP1c3LlHOSVG2S55DpwfXFm+2g1tfFN0zEmMP3Wq293kU8Lx/Nz2skn4sJcrVOrB0/+R6IuDyUpS72ede08V2cOtj5mdhi95DngRc9QXg/OuBZQf6rze2EjjeUyyTKVGOxH9RKXURFLiEEEIIGWBYy9pnuu/BTTu4rcAIokIHlyXKew6xg9uj35ldouyK09EcgWuHTLmhUL5SYCHPwW1b+90EZglWAiwH1xK44kBLCNfeRwKPeq4RuHGJcjQmqDqUdsLdcKlKBZkPEnwisVIFXvfT7M8HyBe4VWckEJDuyz3qLOBdW4DlBzlrLPHvgHuM/GxK/huS4fFvitbHHlxCCCGEDB4d31UrpYaVUi9TSr1DKfX38tWPxQ0a90zcg+9v+D6A8inK7cAfElSv5jt+dHD3IPoZMuUGMY0uc451QqbaDUcg5zi4gm8ObhgAv/1s8tydobvqkcArLjO9tXapc23YCDhxGyUIyi6rFuEXBqbcWcTnEy8AnvVhxGLWFoiuiMtzcNccDxz0eM9rlsB99eVm3UDawY0FriOG3XTksrgffsx2jM7T/wG4cHtWOLMHlxBCCCEDQBmV9R0AZwJoA9hlfZEuefF3X4xf3PcLAMUO7ss/sx7nfuVqAEA71N5j6gUCmQJ3ntAaeOi2bk+KHudY4E5tBX7xr45j6gZOOfe0HdyJB4B/XJUOedJBViTb+Obn3ngJ8KuPW2twSqYrddM3OroMePBmMzap3TC9uaqSCOzGhHkcXpycHwtcq0QZAJ76TuCkP08+NFAlBO6Qfd3oOj7BZ7ueB54M7PeY7H4RuL7gqZng/ltRi8Koin4XM4EOLiGEEEIGgDLvaPbXWp/R85UsACZaE/F2UYryL+4wgTSfAND2jHmpVxUqlXxBVC14jcwhf/gp8KUzgQvuNP2T96w3Pavn/hbY+/By13BLlLf+EVi87+zF0XffAtx0KbDuvGiHyrqvbsKyvN5uANvuyl4zDD0i2cIubX54QyJKU8c4Jc8iUmsj5vmnnwyseIR5Xqkm92vuNEKvZrmXqZCpZnaEj+rg4I4sTZ4Pjxsx39pllTb7BK7zT2bVI4ZjBzcnkKpbMjN3o7+Ntmde8VzcZ80Jc3tdQgghhJA+UsbBvVIpdXTPV7LAKDsmKPA4uEUBU6SHaJ1OAf7lx8zj/deYx8lo3M30ti6uGX2AoZQJV/rkOuCaL6ePueV/gAuXArs2Z8/PY8f96efKI3DdD09EoLYb6dm18fEdenBtwfWx44BPneopg3bn4EbC0Bb0renIwbVKlBsTxr21S7mVXaLc8jiuPgc32j72ZcCbr7UE7mJgfO/0MT6Bmjdax/7v+eBTgeNeDqx6VPZ8H8/9N9Pzm4dbvi4iP5hjgQsAr7sCeMWlc39dQgghhJA+UUYpnQLgKqXUrUqp65VSNyilru/1wgadmipXDtgOskKjKGCK9JArPwa8dwUwvd08j8Vp9PsQ8VbUp5rBKlFuTZovV8j++ELzuO3u8peVfthY9KmsOHUd3NQcXI/ALSxRVn7B5YpoN0VZXFJbHLenjIhTFUfgLklfyy5R9glc+b3Yrq0cs9ehJkU6LlEeB8ZWpY/3jhByk4etn6+weDXwvE8AQ4uy5/s44RWm5zePTIly9GFALwTumuOMs00IIYQQsodSRmU9s+erWICUDpmig7v7II5tY6cRASIQRUjJ86IyXhc7ZEp6WING+pjNUV9vN2NhZGSPCEdVSYT3ae8Bfvzu7DrtEmXtCTcrCpkaXux3d93jA9fBjf4J2rrBWnvk4FZcgTuevpadohy2PCXKInA9JcpyX7tEWQR0fEzk1tZGkt9NXvKwL715rsikKEc9uL0QuIQQQgghezgdlZLW+i4AywA8J/paFu0js6DsmCCfwB2iwJ0fJh1nVdxJO80XmJmDqyrJmJ12w39MUXmwizi4MqJHqWR9S/c343lyHdxmjlgN0gLVZmg8K8yBrOObcXAjMTixKdlvO7ihFTJlB0wBHgfXEZ8icO3y4UrN/zi8BBiPHFz5PYiDKz21QLYH13ePueq9de8h9LJEmRBCCCFkD6fMmKDzAXwFwKro68tKqTf1emGDiC1qy/fgZp00lijPMyIMYwfXEbiucCy8VvT71To9nkewR+8UCZp2M53g3JxMX8t2cCs1Iw6ntgG3XZ69/sRG4NI3eNbaycH1lSg7x3//b4C7fmW+X9t1fUE0SmhkmfmZ1Eb8Pbg2qZCplkd8qvRxgDUTV8KhrBLl8X3M9q4Ho9cioZoSuDki2i5RnutxO+6HYeLgznXIFCGEEELIAFBGKb0GwOO01n+vtf57ACcD+PPeLmswsUVtWQe35enBLZqBS/pALGRdB7edfr0bdJCUwdrjdrbcmWwXCdzL3wF84sQkXCouUZZrqcQ9rdSMaLrx28BX/xR4/wHAl56XFqM77sveIyzowR0e9wsu+VksWhld917g1v9J9osY3P8xwImvTdZbG86OCXIFrv3BQrclyrHQtUqUx6KQqZ0PRGsTgesZIRTfQ4KsPC7xXOGWKO99pHmkg0sIIYQQkqGMwFUA7HfsAeZ8aOfCwBa1ZQWuL0WZDu48I8I2nMMeXB0mbqsIl613JQnNQHGJ8l1Xmkc3oKrIwRUaO4A//MRc/+gXJXNWM2stcHCHxv2CS9b8ws8l+1pTabEtqKrV61rPjgkacntw7RLlZkGJsk/gOo9D48BRLzAJyKe8LVkD4Di4OSJa9dLBtf57v+BOYJ+jzHY3lQKEEEIIIQuEMlbD5wGsV0pdEj1/HoDP9m5Jg0u1Uo0/KihbotwONSoKsHVumZCpb7x+HZptT1AQmRl22bBbiqxcB7eLHlxbLEu/bHvaCN+PHpO/BhcRe67oaXl6cCu1bF8nYETi8oOBo84Crv1K9vWwnd+DO7zYiNZffQI4+k+tcyIhW7dEc2sq+RnZYjBVShytUT4A8KYoWyFTQTtbogxfibKTkGyPCRpdBrzqe8mxIpjtcKs8gWsz5z24lngeW1n8d0AIIYQQssDpKHC11v+slPopzLggAPgzrfU1BaeQHGxRq9zZljm0gxDDtSqmWolwKRMydeLBe3W/QJKPHYLkOrXyu5xJD64ESOkg3YN796+zhxaVpNpiz55hG4shle/g2mupDqXFaOrlML9EWdzVy9+RPl/uL8FIgCmfDjwObqqUuBr14Abme2ruLA6ZCgtCpuzEcrc0WZ671wasFOXR7D73Hr0sUQaAU98OHBEF2uf9fgghhBBCSCkHF1rrqwFc3eO1DDxlZ9/atEON0SFH4LJEuf/seijZdntwodPPZ1KiHNo9uA3g5suyxxaVKNtiT1xbILmmqqRFZV6JfKUG1Avmt+a5hyvXJtstq4e4FYVd5Tm4uQK3lowJau40+/LGBOnAiP/c/tiiEmUpQ3auDSRubZ7LDPSnRBkAnvauZLvo90MIIYQQssChUuojZcuSbYJQY8QRtJyDOw/YwU/aEbixSJ1JyJSIYydFefLh7KFlHNzWFDC9LbtuhRIOLow4KytwJTjqSX+TdkDtn5WI7ZSDO5mI7U4lymFg5g4DxSFTvhLlOEW5RA9ukYNrr7FMirIrtOea2nBvr08IIYQQsgdDpdRHygZL2bSDECP19Hmcg9sn/nAF8NUXm3m3trgMnRJl17ntakyQXaIcCcOgkfTj2hQJXPnbak8D0zuS/bGba/XgVosc3HpxCawtXpcfDFzwB+Apf5sWgfZoI9m2RXMqZKpI4EZjghoTZl9eyNSd/wdM3F9QolymB9fn4A6lH+3z3Hv0cg6uS8n2BkIIIYSQhUgPmsVIHtU816yAdqgx7AjcOkuU+8PXzgaaE+bLDleKnVsRtK6DWyJk6tJzjYi1S5RbVolyu2Ec0kkrFbnQwY3+RlqT6ePKpCjbSHpxHnLt6pA5dmxF9NxyFae3J9sisOsj6X124JWg3B7caExQXObsOMtKmXPu/lV0Tp74dK5r31eeD/kcXEcEu9upe9glyn36Z/Wkv+jPfQghhBBC9iBy34kppSYQ109m0VovyXuN+JlpifLoEB3ceUHET7tZzsHtZkzQTn7/kAAAIABJREFUtV82j0c8K7mG3YPbngb2OsQI3COfDdzyvXIC95uvTu+3e3DLCtyiv9O45Hg0LfZs19IncHNLlIt6cKvpn4vPWa5UgcCZqRsjJcqeAKhSJcrSg2t9b657KsLdHiXU6xJlALhwe+djCCGEEEIWILkCV2u9GACUUu8FsBHAf8C8YzwbwL59Wd2AMZMS5VYQYnktXfI4VGOJYl8Q8dOeSqcHa6cU2S1RnsmYIB0YIQ0YgduaNoLurzYYcfi+fcv14LrEDq5Kz54tKlGuFZTYyvXqI47AtURdSuBOAlBp8Vq6RLliysNjF9gncGvJz8UV7co3JshxZWXdhSXKBYL1yGcDT34HcPLrgau/1Pl4QgghhBDSU8pYgc/VWn9Saz2htd6htf53AGf2emGDyExKlOngziN2cJOdYBw6pcnQzn5r/vCuLcA1X86/hzdFedp81UaARXslwk7WsPG6RAy7a3WJe2adObgzDpmKxObifc1MVqFWUKLsCr7ckCnHwZUSZfkebBfYPk7Y+WD6NW+JsiNs93sscOhTgLFV2Wv7UpQz968CT/5rYGRp9jxCCCGEENJ3yiilXUqps5VSVaVURSl1NoBdHc8iGbTOrfj2EoYa7VBjpM4U5Xkh1dfqcXAzIVOeHtxvvQb4zrnA5jtybmKNGEqFTE0ngk4p4yYGTeDhDcCnngj88J3+tbrEzrNOlyjnObirjykWuOKmvvjLwLM+lOwvcnBd8Z3Xg+uODFLRmKAiB9dm293p594ZtU4P7gEnAq+41O9ai4PbbclxP0qUCSGEEEKIlzJpKC8D8NHoSwP4ZbSPdEnQTbouTMBUO9AYqTFkal4QEdia8vfguiFTvh5ccRV9qcipcz09uLZjWR0yrq3M472/y7HUWhf34D7zQ8Da00zf73RBf6ekIi9Zk75GKmTKGlPkE7hBM/l5FM7BraadbZ+Da/9eMh8gSSm/td8NlyrCl6JcBjq4hBBCCCHzRqHAVUpVATxfa82S5Dkg6Go+qilPNinKaUHLEuUc7v61KZVdc/zcXM92cENPinKZMUEScJT3u7fPtefgtqbTycPVuhFzso6pbcDtPzKiFEinPOfdJ7B7cJ2/oUc8zYhboNgpbU35BXJRyJSvfFpGGdli0C0lljFBRQ6ufE97Hwmc9en0a/I9ap/ALSFC4xLlLlOR88rFCSGEEEJIzylUSlrrAMBL+7SWgafdTfgQgHYYIgjDTEnyEB1cPz98J/CT983d9VI9uEUOridk6huvAn724US02aLXFrvyN2H3mkIDzZ1ZB9cWuFtuB77yQuCh24DJh4sDqGSNoZU27ApUOwV4qEOJsu3WxuuzBKOM9ZFtn+BrRAK3U8iUDvxJzIL8XNedByw7IP2ahEzZPdFuenIRM3ZwezwHlxBCCCGE5FLGavilUurjAP4LVu+t1rrLGknS1p0Frt2nG0QlytVKOjWZDm4O7UZnodcNeSFTcSmyCCe3RDkEbrzEfK05wTkWSZkvkB45JHNwgUjgWkKyOmzWIC6v8IkTgeGlwJpjO3wzdolyNduDa/fdFvbg7vL3q9Y8ohfwh0wBQGMiWkunMUG6eExQvGbPa16BG92jTBmxLXAPXAfcf23nc8pemxBCCCGE9IQyAve46PE91j4N4Klzv5zBpkwPbmhVU7YCU6JccwRuvcoxQV5sl3IuiAXuZE6KstOD6wuZEkFlC++W1Y8bO7hWD65Qs0SblCg3Pfluje3p9fno1INrO7idSpS9Dm6Oa9maAkaXZffHJcodQqbCyMFV1WLh6HN34xJln4NbpgfXGin06h90Pj6+BwUuIYQQQsh80VHgaq2f0o+FLATK9OCGjoMbhBq1TIly9+OGFgRhOy1mZkvFCpnyzsGVGbbumCDr9yxOqS1eW7aDawtcx51NObhD+QIX6CxwGzuAn3/EbLspyrWRtOArdHAnixOHXZq70uOE4vVEfbqVvDFB9ahEuWV+dp0SlOsFAtcbMlVChI7tDaw8HFj1qM7H2lRYYUEIIYQQMl+USkNRSv0JgEcDiN9Faq3fk38G8VGmBzewLNx2GKIdhnRwyxIGPRK4k/4e3LyQKfv3LCLLFq9Nq0c1dn2DrINb9zi4dn+rja80uzaSvub0NlNqWx1KizBX0BY5pd06uDrwi8lSIVNWD67PobXxifJlB5nHocXp69qPRQyNAef9tvNxhBBCCCFkt6Gj1aCUuhjAiwG8CWbuxosAHNTjdQ0kZXpwbQe3FWiEGqhWFG557xm46PlHAWDIVC46mNsSZRkzk+nBFefWdXI9Y4LiEuVI4N78PeDiJySv2yK5k4N72w8SF9ZFxgfZ+FzPMz5gelNtMXnsS/zX9NGc9PfbFgUreUOmpAe36j/OHRNU5CoDfgF8xvuBF34eOPBx1nW76MElhBBCCCF7HGWU0uO11q8AsFVr/W4A6wAc3ttlDSZSonz0yqNzj7F7cBttc3y9WsFIvRrPw2XIVA5hO10ePOvrRaLWFbiugwu3B9cuUXYc3Bsvce5hpyhPpftubdEmonJio3+tvv0+USgCTx73e6wRgi5PfSew9vTs/tYuv5j1lS2793zrjcA5l5rtjinKUQ+ujAnylSDb+MR8fRQ46ixnLV304BJCCCGEkD2OMkpJEnEmlVJrALQA7Nu7JQ0mQRhAQ+ONx70RX/2TrxYclyjcRss4g5KiXItKk92xQSQinOOQqcASuL4e3Dg9uWAObixwo1Jhtx81VaLcAEaWJq/VnDm43eJNFq6mH/OE3hMvSGbs2uSVCxc5uLL2pfsDyw402+LgVvN6cGUObuTgdipR7vS6fV2AQVCEEEIIIQNKGaX0PaXUMgAfAnA1gD8CyFdoxIskKNc7vLG2xwTd/uBOAIh7cEXYskQ5Bx3MrYMrfa15c3Dj+4rALUhRbkfnL1qRPtedg5srcGcwW9UncF0Hs6gX1XfPoFkuZMotN3bX1NgZvVYgcGVMUGuqRMhUhxLmeJ1WMjIhhBBCCBk4OiolrfV7tdbbtNbfgum9PVJr/fe9X9pgIQFTVXf+qIPt4L79G9cBSASuK3SJQ9hOz5udLSJKZUyQCE5XRGvHybXLmV0H1xWCbg/u4tXJa3ZZbqeUZB9lSpSLSnXzRLUvZEo5wWeL11j3tESs/AybJXpwlUp6cDuGTJV0cMf3AUb3Yg8uIYQQQsiAUiZk6hdKqYuUUmcAGNJab+/DugYOCZjqJHDtHlyhGgnakXo1eqTA9RL20sFtJX2wuQ5ukD4PsMYENbKvAZbrG/WarrTa221RJz2rRZz0F6Z3Nj7fJ0Qr6XUVOrg5IrCo31ZYvE+ybYtYWZM4uLkpylXz/KGbgXvWd3Zwax1eF44/B3jzNezBJYQQQggZUMoopXMA3ArgBQCuVEr9Tin1L71d1uAhAVO1DqWRdoqyIM7tusNW4ANnHY1j91829wscBDqlKF/1BeBHXRQfBLaD20xEVGYUkRMy5Y77AZIUZTcpOU5ebpkvu0fXFrjTBQJXRGu1nrillbq/zzTj4Bb8PfoEMuB3cF3GVvlFdOzg+kqUXQe3kj0vj7KObLUGjPK/H0IIIYSQQaVMifIGAD8C8L8AfgZgEYBH9nhdA4f04M5G4NarFbzkpANRqXAOrpcwLHZwv3s+8MuPlr9eKmSqXeDgSolytN8WsRJOJaI3cASufQ8gLSpTAregcELKgav1ROjZ2zZuyFRRRcHyg3MSk3ME7qv+BzjlrWZ7bEVyrhskVaknHx7k9epWakkQFZAWuz7cEmlCCCGEELIgKVOifCeASwHsA+CzAI7SWp/R64UNGjPpwRUkPZl0IGzPcYqy6+Dm9eA6IVO2gysCVkRv2ylRlns0d5nH1GggW+Bu869xaHESTFUdSkRlJUfgdtODu/po4O8eyO7P6809+AmJy71opbUW50MdKTdWFaBi/ROU2q4BD/8heb793vx1EkIIIYQQElGmRPljAO4G8FIAbwbwSqXUYT1d1QASC9wOvX8eAxfVCntuS6EDT/nwTK+lE/HZnjZCVYKM8hxcCbhKObjt9L6gAYztDTz1Xen9rUnzWBsBRpdH25aQdHt3hZEliWCs1BMxWa35S5QzY4I6pAn7/vbyHFwgEczLD7YSi52/eTnfXZ/r4NoCd+uG4nXuDhzyJGD/k+Z7FYQQQgghC5oyJcof1Vq/CMDTAVwF4EIAt/V4XQNH2RJln4NbZ0lyOcIOPbjdXkt6a9uNDinKBQ6uCNdY4LZMD6uIPilhbloC9/FvNtvDS5LrrDvPv87lhyQC1y5LrtQ6OLiV5LhuKRpZtO484FkfBo5/ubUWZx3i8rp9tXF1g4rWZ/23cOC67tfZb155GfDaH833KgghhBBCFjRlSpQ/opRaD2A9gGMA/D2Atb1e2KAhDm5NdRC4Hgu3SoHbGa3ndg6u7Zi2p40QFUHmjiISgevrwRXhaodM1Yayva8tKVEeNn2s73wwHYZ0+kXAhduBZ/8LcPCpZt+T/ho4++uJI9pNyFRZB9dHUeBTfQQ46c/NffJmzsp6h8edtdXSj2d8EDjoFOD864Dnf8p/v9ddAZxzaXfrJ4QQQgghA0uZd7e/AvBPWmtPMx4pS/kSZfbgds2uzUlZ71w5uCJwVQVoTRf34BalKEtvrV2iXB3Oij7bwVUqvwz4sa8GJh4A/vhz4/AOjSXrqg6Z0mTAPE5uNttL9gd2RD2sqose3DzKjAkCTH+wrCV1frTeobH0fjfZ+eTXm68i1hxXbi2EEEIIIWRBUKa589sATlNKvQsAlFIHKqXYaNYlcYlyJwfX00JaYw9uPnf9CvjQYcBNkYs3Vz24Eg41vCTqwW0nwi4M0i6uOwfXdnDFmRXR226a67jCUkqZ6x3G4QDJOuTYuKe1lnZwN99utvd5dHJuxsGdgcAtMyYIMEnKsi4bWfeQ6+BGayk78ocQQgghhBCHMsrpEwDWAXhZ9Hwi2ke6QObgdnJwi8YELQgaE8DtXfQxbrrBPN7+Y/M41yXKI0vMNVuTxiFV1Wjebjs5Ng6ZEoHrc3CbwIM3G1e1Opwde+NLUc5DjokfJWTKKQt++E6zvfqo5FxX2HYav+O9f0mBu2hlspbU+dG6c0uUZyC6CSGEEEIIQTmB+zit9bkApgFAa70VQMkaRSK0dckeXE/I1ILqwf3OecBXXghsubPc8RKw1NxpHt3+2JkiAleCnpo7o77WauTgtpJji3pw7VFDnzwZuO8qIxBdESfnlhGPckzNcXDbjcTBrdaAs78JHP2iRGgCWQcXM/jbKgqZshkTgeuGTImDuzi93+3BJYQQQgghpEvKCNyWUqqKqNFQKbU3gDlSEQuHOGSqw5t335igWnUBlSjLaJjGjnLHi1gSgVvGwfX9kF1igRuJsMaEcUd9Du7tPwJu+W+/gytMbEq2q56QKaEbB1fEvTy2p5N+10odWHsa8ILPmJ5eQRxbKXtXMxC4ZR3csb3No1s2nhcyNZvgK0IIIYQQQlB+Du4lAFYppS4C8AsA7+vpqgYQ6cHtVKLsS1FeSAZuXGIbtIuPE0RoNsTBLSFwwxLXdgVu0IxSiqvGJbbXd+t/A197WXJd3/Un7k+2fQ5u/FoJgSs9sDXnsTVtObiWa2qXIYugzRPYZSjr4C6KenDdDytqHXpwKXAJIYQQQsgM6fhOUmv9FaXUVQCeBlPP+Dyt9c09X9mAEacodxAWvh7cVlDCcRwURKDZJcBFtKfMo/SwlnFww3bnICO3RBnI78EVyvb/ztrBFWHrzJNtT/tnz/r6bCuzKFEu7eBGJcpT29L76+zBJYQQQgghvaFQ4EalyTdqrY8EcEt/ljSYSMhU3Teb1CL09OBOt+YoOGlPIHZwSwrcVlQO3Jwwj3Pm4EqKstUnWqmZ0t4wR+CWHVFU84wJil8rIXCX7mdE65J9zfO1pwE/uQh4xNPLi0QR2DMpUS6boiwO7rQjcOXemR5cOriEEEIIIWR2FJYoa60DALcqpQ7s03oGFgmZ6lii7BG4yxctoEwvETfi4GoNbP1j/vHi4EqJMnTnHtsy4tktUQYcB9dzjTyB65b0VoeSHliXMu7ofo8BLrgT2OtQ83zN8cCF24GD1iUfEOSVKAszcUnXnGDuufcR5Y4fWWYep7en98vPKTMHlyFThBBCCCFkdpR5J7kcwI1Kqd8A2CU7tdbP7dmqBpAyJcr3b5vCR350W2rfJW98PI7ef2lP17Zb4fbgrr8Y+MHfAK//ZXrcjdASgWv1eYZBErbko4zT6i1RtlOUPdfIc4b3eTRw/zXJ89rw7EqUAWDRXv79lZIlynFpchcO7ut+Uv5YIHFwXSErHw7klihT4BJCCCGEkJlR5p3ku3q+igVAmTm4b/rPa3DVXVtT+44/cHlP17Xb4fbg3v1r8/jQLcUCNzWXNkDhn3aZEuXJh83j4tXJPjtF2ecC5/XgugK3mhMyVakVC/MyVK0xQYK3DLkPfd1L9gWe+3HgEU9L75efvyvm44Rn9uASQgghhJCZUSZk6op+LGTQaYbGERyq5JcbL6he2zxEmImArC8yj81d/uN9I3k6ObRlAqy232MeVzzCWttwkqLsW0/efZcdnH5ec0OmFABd3r0tomzIlJRxz6QHtxtOOCe7T363rlOrlPm50MElhBBCCCEzZAENWJ1fmlHJ61DBiJWKIzaOO2BZT9e0WxI7uJHLNxQJ3Nak//iWR+B2SjMu4+Buv9eU2I5aDvrociMWdZANTgIKenCdYDFVTbuUMse2bDpxEXGJsu3g+v4zFwd3HmZQyc/JJ2QrFLiEEEIIIWTmUOD2iVbkGhY5uPa823958bH48msf1+tl7X64Kcri4MYhUg4SMmXT0cEt4ZRvuwdYun9adC5akfTgurNdgXxhXa0Df3M3cOrbo/u30g5uPPZnDh3cTiFT8WvzIXCj361vVFOlRoFLCCGEEEJmDAVun2hFgq1eMH9VWWLjsL3HMT484G/0tQY2XpfeJ85mnKIcmseph/3X8Dq4YfF9y6Qob78XWHpA4q4CJthJenDdZGAg3xmu1IGRpYkb3G6mHVwRtnMhcL1BTR4R2ylp2ubc3wCv+dGslpXixNeaxzXHZ1+r1NiDSwghhBBCZkyuglJK3QB/Eo0CoLXWx/RsVQNImR5c28F1y5UHkqu/CHz3fODsb5pZrkBSYtuOUowlRGrXZv81fA5uJ4HbqURZa9ODe+iT8h3caY+Dm3dd6SuWawUNv8C1xfRMKe3gdlGiXHYsUFkOP92MNfKhKnRwCSGEEELIjCl6J/nsvq1iASA9uLWCN+9VS+EuCIH7wE3mccsdicCNS5Qb5lEE7mSOwPU5uB1LlDsI3OntQHNnVKJsuaqj4uCG/hLlPES0S/91u5kuURZhO+SMzZkJZccE9StkqlsqtfTaCSGEEEII6YJctaW1vqufCxl0mmETQ5WhVBmyiy1qbbE7sIjwsh1XETdXfcGUCLdn4uDOgcAFTElxSuAuAyqVyMHdDgwtBpoTxdcCEtGe6+BG+925sDMhTlG2rq98M2/nMWSqCPbgEkIIIYSQWdCxB1cpdbJS6rdKqZ1KqaZSKlBKdWFfEcD04Bb13wKuwO31inYD5Pu1+0FFmG25A/jGKy0Hd4v/Gq2prPM5Wwe3HbnHtZG0w1mpWj24O4zgLWLxmui8SLDJyKHVRzshU3Po4CoF1MfS15Lvwffhym7n4FbZg0sIIYQQQmZMGavk4wBeAuAbAB4L4BUADu/logaRVtgq7L8FjDkoFDm9g4N8j3art9P2LeOBJvNCpqaA4SWmpDi+xGwFbiSqfaFPcQ/udmBkWTIvNyaaaQsAyw8CJu5PXNX9Hwu8/pfAqkcBG69JTqlH95kLBxcAXnEpsNdh1pLkD8v6m+omZKqfcEwQIYQQQgiZBaV8Qq31HQCqWutAa/15AGf0dlmDRzNo5jq4Wmu857s34ZaNSblrdSEIXPkef/hO4P89zWy7wksc3PaUX5S1p4GRJel9nRzcTinK4uDWPQJXHNzGDpOM7GKHUi070Dw2rDLm1UeZTzKUJ2Rq2Pk+ZsoBJwFjK+xFRw/2f+67qcBdsh+wZM18r4IQQgghhOyhlBG4k0qpIQDXKqX+SSn11pLnFaKUepNS6hal1I1KqX+y9v+tUuoOpdStSqnTZ3uf3QXpwfUx2QzwuV9uwJZdzXjfHhkytWszcOsPyh9vf4/3/c48ugnILavH1idMW9NZodkxRbmDAG6VcXB3GGHtBjhVI4GrqkasAcDERs91LJdSwqfmokTZh6xReRzc3e3v7JxLgadfON+rIIQQQggheyhlagHPgRG05wF4K4ADAJw1m5sqpZ4C4EwAx2qtG0qpVdH+R8GUQz8awBoAP1ZKHa51p5rT3Z+iHtx2kHXTKntiD+5/PB/YdD3wjo3A0KISJ/jms7oCdzLZbk8DNedDgvZUVuB2ErCNHcD6TwEnvMI/mifuwY1ee/I7TLkxkKQoT28Hhh+d/R5qQ0ADRgivOw946FbghFdl75HqM41+/3NVouziK1HeXUOmfK45IYQQQgghJSkjcJ+ntf4ogGkA7wYApdT5AD46i/u+AcAHtNYNANBaPxjtPxPA16L9G5RSdwA4CcCvZnGv3YIiB7cVZh3HPdLB3Xx7tFGy/NU7vqbAwRXhKYQBEDSzpb3yecjOB80xk5uBvR+ZvP6t15jHJfsBj/RMw5IeXBFbT/7r5LVKxYz5aWw3wlpV0j2/VatEeWwF8NKvZq8PpEuUg6gnuJ8OLiGEEEIIIQNIGZ/wlZ59r5rlfQ8HcKpSar1S6gql1InR/v0A2Kk990b79nhaQQtDVb/A9Tm4e+SYIBF6ZQOMfILL5+DWx8x225l5K8+HF6f3i4P74bXAPx8JXHwKcMnrsvfKTWaOrusrUZYe3OYuYGgsK9Kl7zZoZs+1sR1cOdb9PuYK5enBPfTJ5vFwttMTQgghhJDBIdfBVUq9FMDLAByilLrMemkJgJxI29T5Pwaw2vPS30X33QvAyQBOBPB1pdShXawbSqnXAXgdABx44IHdnDovNMMm6hV/iXIryDq4PTHb2g3gH1cBT/5b4Ml/M/fXF3FatqK8rIM7tgpo7Uo7uJf/HXDXlWbbFYa++//+W9l9ux6yztHAzgeAxasT4VzUgxu2TR+t+4va51HA3Vdmz3Oxv/cg+r567uBa91xzPHDh9t7cjxBCCCGEkHmiqET5SgAbAawE8BFr/wSA6ztdWGv99LzXlFJvAPBtrbUG8BulVBjd5z6YHl9h/2if7/qfBvBpAHjsYx+7m0bCJrSCFobthF2LduhxcHuhcBvRKJ31n+qxwO0Q8hRTwsENmsDocmDHvWkH91cfT7YzDm7J++/aHN1TA19/BXDzZcBbfl8scFUVCKOwq0otK9JXParcvW0HV8qw+9qDSwghhBBCyOCRW6Kstb5La/1TrfU6ALcAWBx93au17jBItCOXAngKACilDgcwBGAzgMsAvEQpNayUOgTAWgC/meW9dguKenDbHge3JyXKMv+1V3NGRZy2poALlwI/+1Dx8WUcXAAYXWYeg0b2NSDrfJZ1kHdFrd9bNxhxC5hQKBG4vsCjStX04AJREpjze1pZckS03YPbjIK0euXgescEEUIIIYQQMnh0fMerlHoRjMh8EYA/BbBeKfXCWd73cwAOVUr9HsDXALxSG24E8HUANwH4AYBzByFBGTBzcPN6cFueHlzVCwc3dh6rxcfNlqmt5vHXFxcfV6YHF0gE7o7706FTgut8dkpRFqREWZxtwIjdwh7cStIz63Nw3Zm8edgfMkhSdM97cHtzeUIIIYQQQnYXylh57wRwoiQdK6X2BvBjAN+c6U211k0AL8957SIAF8302rsrrbCV7+B6Smp74uBKD2uvHNyu7+MTuJ5q89Hl5vG/Xg7sfxLw2h+lX884uCVKlEf3SkqUQ2u+7tY/mrm0lbr/g4BKNZnHq6pZgSuBWJ3wlSgPlTy3W1iiTAghhBBCFghlahYr1hgfANhS8jxiUTQH1+fg9iREWZzH2ZaqTjxgSpDvypne1NxlHjsJXJ8Q9e2z59ze66lYd2fZ6iApI85j+cFmjBCQjOkBjMBtT/vdW8CI2sAS8O7vaWgR8LJvAM/9t+L7+0KmfDN554LYwaXAJYQQQgghg00ZK+8HSqnLAfxn9PzFAL7fuyUNJkUpyr4e3J7MwZ0rB/ePPzePv/k0cNC67OuxwO1QCh16Wrl95cVDHUp33dLvMACaO/3HCssPBu6/JklEBoDaqBG446v8/beA04Prc3AXAYc/o/jecq5wzqXA9V8HRpZ1Pm8m+FKUCSGEEEIIGUA6Kh2t9QVKqbMAnBLt+rTW+pLeLmvwKOrB9aUo79YCN04Z9qdCx+Kyo4PrEbM+B7dTurC7Dl1C4K44DIA2glZKlFc8Ath2F7D6GCN2faQcXI/ALVtmbIdMrTnOfPUKligTQgghhJAFQpmQqQ9qrb+ttX5b9HWJUuqD/VjcIFHUg+ubg9uTHlxbmM0GEcq5ArdkibLPrfUJ3E7pwhkHN0wHR/l45HPN442XJCXKYyuAxgTQnsr/3ipVq9S7ioxozClD916nX8QOLgUuIYQQQggZbMrULJ7m2ffMuV7IoNMKWvkObr96cO3S2tkgAq86W4HrKVHWIbDyCODxb0r2uQ6uG0Tl/lx1kKwhj1WPBA442QhccXCHlwDQwPSO4hJlEeGV2sxFo+qjwOWYIEIIIYQQskDIfcerlHqDUuoGAEcopa63vjYAuL5/S9zzCXWItm7n9+B6UpTnZEzQrs3psTpBv0qUy/bg5ji4qpJeo+vgBk6AlLuOMACaE8X3rtZN//BDtyaO9HA04mfq4fyQKXtdlao/9bkMvU6ytmEPLiGEEEIIWSAUveP9KoDnALgsepSvx2itvSN+iJ9mJMi6SVGeEz7/LOC85fGsAAAgAElEQVRnH0qez1kPrpQo54jAoh7chzcAN33HbLsObhh2FriqksyNFXwObqcSZQBYur9xbyc2mucyh3Zqa0mBWwMwU4HbR7HJHlxCCCGEELJAyFU6WuvtALYDeGn/ljOYNEMjcLuZgzsn7HrIfAnifPbNwa2ZmbFBy4zPAYBPPwmY3g78w7aswNWhcURdgWuPz6kOp11p3zraTeC+33X+PpYeYB4f3mAeR8TB3QqsWOs/x17XnuKIKpYoE0IIIYSQhQHf8fYBcXDzenB75uCGgXFFBXFei/o/N/wc2HpX8XWllzfn+0kJ3C89D3jfvslr09uTtbglyjqIHFyVLm+2ne/aUFbgur3Av/scsP7i4u8BMA4uAGyNBK6UKE9vL+7BjbdrMy9R7iecg0sIIYQQQhYIfWwEXLi0I6cyfw5uj0SSDtIuaVAiZOqLzzaPF27PP0Z6efMEU1yiXE1m5vqOcccEhUFSomyLcPvn5nVwHaE9XbB2GxG44uAOW/N288YE2WupVDHjEuV+whJlQgghhBCyQKCD2wc6Obi5JcoP3Qp85jQzumYmhO20iGx3MSbowqXAljv9r0mJsi8kCuicYAyY7ylTohz4e3BTDq5H4LoObtmU6JGlxrXd+kfzPCVw88YEOT24e4SDKyFT87sMQgghhBBCeg0Fbh8QB7eW0/uaW6L8v+8B7v0N8IefzuzGYZAWod2mKOfdV4Sy68AKInB9Y4DsY7w9uB6B64rdtitwnQ8OxEGWkuMiluyXjAkasY6vL/Ifn+rB3UMcXI4JIoQQQgghCwS+4+0DYTQ3tZIjMNpBjoMrxxcJxcIbuw5uVKJcdgZrXpKwCNw851kEbtDKv3ZzZ9YBDks4uGGQdnArtWwicWPCfI9/8pH8+wu2qLUF8egy//GpHlznvmuf0fl+8wHHBBFCCCGEkAUCe3D7QAgjBFVOjWg7zHEBRUzllQIX3jQEoP0Ori6Z2pxXptvRwY0cVHdeLWBElg7NGJ9MyJTt4Ob04Ibt9JggtzwZMAK7vsjvVC/eF3j+p6zzLffXFrgjS7PnAmmxXaklBu7p7wMe93r/OXmc9h5gzQndnTMT2INLCCGEEEIWCBS4fUBHfZp5Dm7LcXAfuW8ktNQsBK6Iz9Dj4LYmgY3XA/sekz6n7H1EYOY5y7GD6xO4VSNim54e3JSDa6coW3+mQRNoTSfP3YApWd/YKn/K83EvAw59knW+JZDtHtw8gZtXolwbKd/7Kzzh/O6OnykcE0QIIYQQQhYIfMfbB3QkglRO6rCdovyaUw7B988/1TwRMZXnlBYh4lF7HNw//hz41KnAtnvS57glxe1peJES4U4hU1Nbk30SxiTfU2Nn+Tm4toMbtDo7uICZnVv1pFa7Is8+f3g82S4jcFPbXYrbvsIxQYQQQgghZGFAgdsH4h7cnB+37eDWKpYImVWJsji4loh0Bas7TscVnG5acbw/Eph5wrsVCdydD2bXI6LQNyYoNQfX6cF96rvMdtBKr0sc3NPeC6y2HOmhMX+JsitwxcGt1NLBUrkCt5reFuG+W7uj8gEKBS4hhBBCCBlsdud35QODlCjnObh2inLVJ3Bn4+D6SpQFV5SFJR3cOCW507qs3mL5HiSYqZuQqUodeOLbgVPeatZoC1xxYJ/wZuBV/53sz3NwXZEXC9x6+vhSDq5Vorw7C9w9QoQTQgghhBAye/iOtw9IiXJuirKVRpwSuHEP7gxSlCVIyg6UkhLl+PrOegLXwc0rURYHt2RYFWAJ7uic3BLlwBMyFa2zUjc9uBJiBaT7bO1z6ov8PbiZEuXoGFcM5wpc6zhlO7i7cYmy/J5YokwIIYQQQgYcCtw+ICXKeSnKtoObOkLcwrxxPIU3LePgOusp6+DGPbhdCG85VmbYNvMErsfBFUSMNnYkAtQOmbJFZn00EaMpUVpQomxTugd3T3BHWaJMCCGEEEIWBrvzu/KBYcZzcCuzcHB9Pbiug5s5x7lPNyFTnUR4GJj+WbmHb0yQlChXqn5HVNKUp7YCQ4tNebIdEpVxcKPj7aRkV+OJaHYFrj0yyMbtwfVt727ELjMFLiGEEEIIGWwocPtAnKJcYg5uaiKumoMeXF3g4LoC001R9oVMhWFyTfvanUR4GKSv5+vB1Tk9uIKI0altJkSqNlzg4C5KnFu7VDnPwXVLlPMEq+vg7gnikSXKhBBCCCFkgcA5uH0gLlHODZlK3E9tK1zpPZ2rObiug+sK5zIOrl3GnHJwW9lj3Wu3rW+u4ZmDmypR9ghMEaxTW43ArQ45Dq4lXuujibCtjST7MwJ3xL8/j9Qc3Ar2iBLlWODuxmskhBBCCCFkDuA73j4gKcr5Jcq2g2uJwFnNwfW4rG1H4LoCUxzcsVVGDPkcXNvlTQVYdRC42nFwW5PZ+4eBfw6uIC7rtO3gOnNwxcWtj1olynbYlPMhg4jg1CcLBdhOb0rs7sYlyvJhwdDi+V0HIYQQQgghPYYCtw90k6Kc0llqNnNwPSFTgVui7PTNyjnP+Siw6tF+B9e+RthNiXI7LXCDZvk5uIKIy6ltwNA4MLzYPNqM7W0eh8YSx7fQwR1O7g0Ax74UOOJZ+d/HnjgHd/UxwFPfCbzws/O9EkIIIYQQQnoKS5T7gC9F+f5tU3jzf16DT53zmFSKcpgqUZYU5Vk4uEUCN69EuVID6iM5Pbjt7HZrCvjocZ3XIwnKqmJGEmUc3A4lyuK2Tm8DhhYBz/8UMLosfcxehwI7NzklynbIVM6YIHGjn39x8feRl6K8O4dMKQU88YL5XgUhhBBCCCE9Zze2nQYHXw/uZ3+xAb+7ayu+edW9aQcXnlLZGaUoe0Km3DLivBLlas24nm5JM5AWyXLtrXcBzYn0ce4M2jBI5uoOLzE9u2EArD4aWLE2ul6HMUH2vqExYM1xwPKD08fI8+pwUqJs9+m6fdAifst+iOCWJe8JDi4hhBBCCCELBL4r7wO+EuWRutlutMOUg5vSt3FQVIf+Vu9NfWOCPCnKuzYD/7gauHt9cp9KPRK4HXpwRRS6CcSAcVBT92on1xteEo0MCoA1JwDP/GCy5jIOLpAtTRZE4O7cZJUoFzm4UqJcctZwysGtYo8ImSKEEEIIIWSBwHflfUBCpuwS5aGqEXCNdpCag5vyb0V0dQpw8hGXKNtBUJ4S5YmNRnhu3ZCI4Wo9KlH2pShbgjleXzN7XM0jcGMHd3EyE7dSs+b9dhoTZAnpobHs6wCw32OieyxJji8SuLVuBa4dMmWJcApcQgghhBBC5h324PYBKVG2HdwgEp6TzSA9B9dOmRKROhuBW1iiHCRlyO1p0xcLGHFZG81xcJ2QqWv/E7hnffa4+kj6+fR24FuvMdsjS4HJzZHArVrzfjuUKJcRuGufDrz8W8DBT0yuYYdMuSnKXQtcO2SqxhJlQgghhBBCdiMocPuAlCjbDu6OaSMmt+xspkuUUycWOKSdKJWiHCRJye2GVaJcy3dwUyXKbeCaLwN3/SJ7nOvgPnynGQ108KnAyrXAQ7eY+1dqiTjUHRxc2z3NK1EGgEc8PdlW1bQwdntw3ZCpTmRGA1HgEkIIIYQQsrvAd+V9wJ6De/XdW/HrP2zBjikjFDfvbKRLlFM9uLMoUXZ7cLX2lyjHDm7DCplyenAfuAloTmbXogOgtct/f5k9K+5pMzru8W82+4KWOb9SdUqUozm4PsGY6sHNcXAz59TTwnjWJcpOirLeA1KUCSGEEEIIWSBQ4PYBO0X5rE9eiZd8+texg3v3w5O4b1tSCuzvwZ2Fg+sLm4qPcQRuPCaonqQot6aBTz8ZuPYr0TlOyFQzR+BK7684rSKQa8NGGIZWD26mRDlvDq61r74o91tPUal3cHBnkaLMkClCCCGEEEJ2K/iuvA+EyM7B3TFthOK9W6cw2Qzw5qeZUTnHH2jNdY17cGcicJ2QKbmGLcTCdlKiHNgCt2pSkNvTQHOnea2xI7sWHSbCNXP/SAiL0ypjhGojRnBKyJSqAhUpUe40JqhkibJNte6UFbsObpclyrbItsWyooNLCCGEEELIfMMe3D4gJcrbJhP3U0qUAWD1khGc/7S1eOlJB2DfpVbv6lykKGtHJNdGk7JiHSTXbk87JcqRszm1LbqerCUSwbWRyMHd6b+/XCt2cKN71oaNUJV12T24HVOU7RLlkg7uEy8w83Kv/qJ5nhG4UQm1noGDa0MHlxBCCCGEkHmHArcPSMjUvVuT0KaJ6Tae/sh98Nzj1uDY/ZeiWlFpcQtkxWk3xCFT0aMIzvpIInDDMCdkqp6Iycb2aC2OE1wbjnpwcxzclWtNsNSBjwMevNEqUR5JlwynSpS7GBO08oji719Y90ZnxxyGTKUuS4FLCCGEEELIfMN35X1AenBtgbtjuoX9l4/iuceuwUErcgKTRHSFswmZ8ji48cLaQDva324k7qwdzDRtCdyJTcD9VyfXaU2lxfdp70m21z4DeMOVwPHnmOfi9NaGHYHrhEyFQSRwPSW/9nljK4q//zzyQqbKkidwK/xPiRBCCCGEkPmG78r7gJQoP7jDiMFqRWFno40lo/Wi02Y5B9cJmRIhas+n1e6YIKsHV3pNY4EbAD/7MPCzDyXXccuTj3x2sl2pAfs8OhGEcYnySLqXtlK1xgRJD27VL3CHF5vHk88t/t6LyBsTVJa8tGQ6uIQQQgghhMw7LFHuA+LgBqFOPS4Z6fDjlxE0MypRjtxfHUYjgiLxWrcdXCtFOXBKlGMHd0dyneltybm1UWDXQ+l71izxLMJWHltWinJuiXKYjAnyOaWjy4G/vA0YX1X8vRcxawc350MJhkwRQgghhBAy71Dg9gFJUQ51ev/ijgJ3DnpwASNkfSXKOjDCFsjOwRVn0y5RbiXjjFAbBhoT6Xva4lmcztjBlRJlTw9uag5uQQ8uACzex7+/LK7Arc5RiTIdXEIIIYQQQuYdvivvA1KiHIbp8tjx4Q4lykUpyloDl70ZuPeqZN/kw8CHDzf77FRgbQncjIMrJcrT6Tm4IkJlPFDoBErVRxNxLHgd3Ei8uinKgqp4SpRVDx1Rt0Q5WmfZsUMUuIQQQgghhOy28F15H5AUZdfBHRvuIOKK5uC2Js3omz/8X7LvnvXAzgeAKz7gcXAlRTmnRLndjM5RJjBJhJxdomzPvPWV9tr7YgfXEriVetTfawncat0SuLaD26M/TbcHFwCedzHwFz8rd35eD27efkIIIYQQQkjfoMDtA9KD6zq4nUuUIwf34T8AN3wz/ZoI1sASslJWHDQTcQwY4RqXKLshUyJwozm4Ij69Jcq2wHVGGrnJx8otUZ5MxLXtgtYX+UuUhbWnY07xOa3HvRRYcVjJ8z0COe+6hBBCCCGEkL7CHtw+kJQop/ePDZfswQWAb70GOPqFyXMRsPYIIXFQ247ATZUoL7Ku0XZCptpJ+XDVMybI7sG105iBbC+riFgRus2dwNjK6Fgrubg24p+DCwBvvx0YWYo5JU+g7q7XJYQQQgghhJSGArcP5JUoj3cUuGH+ayJs7f5cEZV2Py1glHVcomwJ09AZExS0kp5UuVbDKlFOhUy5AtcZt+OmKEMn59glyraDG/fgRgJ3NmnJefTKaWWKMiGEEEIIIfMO6yr7QFyi3K3AdS3f1Gvt9KO9HTTzQ6YyKcrRfhHFIkjdEmU3ZMoNW6q5AtfpwQUShzlVojyaiM4wzJYozzU9E7j8T4kQQgghhJD5hu/K+4AI3LbTg9u5RNkSuG7Kb+BxcEXgthtZ4StC1nZPUw5u07jCmRLlHAfXDVXKOLg+getzcEfSIVPQPRaLPSolZsgUIYQQQggh8w4Fbh+QHlztWLj1aocfvw6AfY4Cjj8nOyrI14MrgVNBwwmZslKUbfc0DIywBaKQqXYiPuMSZXFw20DbErhuSW7VGXmUKVFG4uDaYtguUZY10sElhBBCCCGEzAC+K+8DcQ9uqDqXJadODI0oXHqAR7R6UpRlnxsy1ZoCttxhtm3BqZ0e3LDlKVGOHFy7PNm9DpAfMpUSuJGDa8/BrVkOrrjOFLiEEEIIIYSQGcCQqT4gJcqBNrNvdzbaHc6QEwMjnIai5OPmLmBkSfSa9OB6SpSDRroH97LzzIxcIAmRkuPtFOXUmCARoTq5t02nEmVxeJWnB7fqjAmSY2KB28NE4p6lKLNEmRBCCCGEkPmGtlMfsFOUx4ZqWIKd+OPIy4CrvtjhxNAIJxntY7uovh5c2XZ7cEXcAmlHdXo7MPVwcq92w++8AlmBGzuWkWAsFTLlcXDrIwNSoswxQYQQQgghhMw3FLh9IHZwQxMstZ/aYl5Y/6niEyVRuG45uPFFPT24qZApy8G1sYXrVV8AdtyXPL/t+4nYdB3ZxoRznei40WX+4+M5uCpxN/N6cDMObi8FLkOmCCGEEEIIGVQocPuAhEyFITBSryAU17Nozq28XqkmJcq2g+vrwRUHVAdpB9eeWes6sy6bbjCPbmhUxsEVgbuX/3j7PiL+4hRl67Vq3erB7YOD26sUZfbgEkIIIYQQMu/wXXkfkBLldqhQUQoB7LE4BYSBcRzrY+Z50xa4BT24cq5Qt2bfduoVffa/mMeKK3B3pp+LaF0kAtcNmbLuI2I3noPriuHo5xEwZIoQQgghhBAycxgy1QekRFmHQLWmEIrAtUXophuAvQ5L3Fo5QdkOruWieufgWtu2eK4vAqa2ygv5Cz3tPcBjX222Ozm4IlpHl0fH55Qo29u+ObiCqgKbb422d3OBe/w5ibCPr8sSZUIIIYQQQuYbCtw+IAK3HQLVisI3Xn8K8HkkIrTdAC4+BTjsqcA5lyQn6iDdg9uy5tDGPbh2ibK1bR9rO7i+suizvwnc/F3gxD9P9mUEruPgiqAbWeY/3nZwRVTGPbgegasD4M7/y54718xFD+6ZH/dclw4uIYQQQggh8w0Fbh+wU5SrFYUViyOhJ2JTZtGKwBPCIOrB9ZUo+xxcS+A2diTb9jG22AWAx70BWHua+bJxy4hdYSxlxcPjpjy55pQo245m0DSP4va613bZ3R3cfl6XEEIIIYQQUhoK3D4gIVNBANQqCoieI4xEY5AzF9dNUbZLlL09uNb29PZk23Zzj3u5KVf+9cXAzk3Z8T5CJxdVBGx9kRG3RSFTEo41trd59Dm4qWv3cg5uj4QoU5QJIYQQQgiZd2g79QEpUQ41UFGWwNWeUT/a6pF1e3AbO4HbLjfHBL4UZWs77rlF4hADJsH4lLcm17QTlm2UyvbVAmbfmuMTQTc0DizdH1iyX/o4X1qzCNz5dHB7lqLMObiEEEIIIYTMN3Rw+0CIpAe3VlVJua/00UoJLwDseggYX2W2dWhKgSVF+Rf/AkxuBl74+eI5uACw86Fk23ZwBVmDW1psU6mn1wYAf3mrCVj69b+b50OLgNf8KHsdn6M5tjJ6rYOA3RNLlAkhhBBCCCHzDt/t9wF7Dq5xcCNxKQ6u3SO75Y5kO4xCpqp14+RObjb7Jzbm9OBa2zs3Zfef+vZkn7i9eQ4ukJ5XK4hwlRLlobGoD7cgZEoQB7cTFLiEEEIIIYSQGcB3+33ADpkyPbiOg2s7r3bvrJQoK5UETcl5vh5cW+y6oVAHnQI87V3WNaJjfWXIgu81KT0WF7Y+lj3GPs5m0cr8e9n0VOCylJgQQgghhJBBhQK312iN8MGbUYFCO9CoVBTiWbQiQm1hKv2yEw+Y0uJ4xI7ltOrQ34MbBvmOrOuwhiUcXF+vrPI4uN5zPQLXnvFbxJ4kcNedN7fXI4QQQgghhMwY9uD2GqWgb/0fVJaMI9Q67eDGAtfqc203TC3zRw43z6XUNyUYdU4PbssI1toIML3NiFApg3bd2DICV0qU7evE6xGBmyNaVYdU4ae/24RVec/dg0qUT7/IfBFCCCGEEELmHQrcPhBG4rIdalQ7lSi3G8CuB5PnIshsgWuXKAdOiXK1DowsNQK3NpyM6HEd3LgHtyBkSkTx8HhSOp1xcMf953Yam3PKW/Jf2xNTlAkhhBBCCCHzDkuU+4CuDqECIHQFri9kqt0Att2TPFeOYwqYMUHi3NriOGwbISy9rmNWz+tMHFwpUR5aLItJem9FcNfzHFxLSB7wOOART8+/T+bcPcjBJYQQQgghhOw20MHtA7o2hAqaaIYaVeVzcJ0e3O13J89FKNoOrg79Dm7YNqJUXjv0KcDVXzTbGYEbnVcrCpmK7il9trbIXrkWWLK/mYHbidf8sPMxNr0MgqLAJYQQQgghZGDhu/0+EFaHobQ4uJU4Ywo6AH74rmIH19eDq8OkxNhNUa7WgAd+b57brqlbohzPwS3qwbVKlN017HcC8LYbzUzcueCJf5Vs70khU4QQQgghhJDdBgrcPhBW61DQUQ8u0iN8rvwYsPOB5Hl7GthmO7ieHlzbwdWhCaUCjNit1IFnfhBYsh+w5rjknLxxQEU9uHGJcuTgdgqOmg1P/Tvg0WeZbXG2ewEdXEIIIYQQQgYWvtvvA9KDG+hoTJA7o3Z6R7IdNIDtnXpww2x6MmCEYaUGPOZVwNtuSoviXIFb5OBGAlf6bDsFR82W+qh5bE318CZ0cAkhhBBCCBlUKHD7gK7WobRGEDpjgoTJLcl2u5F2dH0Orj0mCLBm4raSvlkgLV7dEuV4f1EPbt2IW7l3rwWurFdmAfcCOriEEEIIIYQMLHy33wfCah0VrRGEoQmZiptwI6YeTrbb00BjZ/Lc14MbBk64VCt5rFhCdnhxsj0TB7cSCVzpW+1liTLQYwdXvgf+yRNCCCGEEDKo8N1+HwirdVSgMYxWFDLlOriRwFX/v737j46qvPc9/vlmJiQGAQlo1YCG9ogJvxp+2cSgjbT8aGurIqtocZWWcipWez31Evx1va13NctTy7JtKCAWLPbcE29PQW690kstWKi9FypIU6QKpKc3R9JqKwoxCoH8eO4fe08ySSbBJDN7JpP3ay3WzDx7z97PkO2YD9/neXbIq+Cefb99W1sFt/MQ5ajbA7UtONXcaVhyphTy59j2dYhyZk7sPiRCZD5w85nEnYNFpgAAAIC0xW2CAuAyMmWSztfprotMSe1DlLOGeRXcs1EV3Jjh0sWeg9vS3HUocma2N6+3uyHKPS0yNTxPOn0iah5wgi+XtiHKCajgmnn3DybgAgAAAGmLCm4AXCgsc9L5drrnCm7WMKmpu4DbaYhyrDm4rU1dQ2jYH/bbl1WU535bWry5vQ+JHqJ8ean3OPZjCTg4Q5QBAACAdEcFNwCtGd4Q5fPaKrjdzMEdcn7H+bhS7Dm4LU2d5uB2M0RZag+w3QXcnqqy4SGShkT1IcHhML9UurdWOm9kAk9CBRcAAABIVwTcALhQWCZp2LkquEOGSu8f77gtZgW3qdMc3J6GKEcquN0MUf4gQ3ZjruTcjRU1HecQ91aiwq2Zt7YXFVwAAAAgbRFwA9CaEWqfg2vqWsFt8RdVyhomHT/qPR9yvjdUOdZ9cFuaurkPbqwhyv681p5uB3QuvRmifP5FfT9PEAi4AAAAQNrit/0AOAsrQ1KOGhUKxajgSpLMq+Ceedd7GalkxqqetpztZg5urCHK3QTcCTd88A8Q1CrKCRWZg8sQZQAAACBdUcENgDMpw0khdXMfXMkbQhy94NN5F0j1x9rnvX6QObgtTbFXUY4cP9rnf9K1ktyddAi4xiJTAAAAQLrjt/0AtEoyOYWtReEMi13BzYi6Z60kZV/gPcas4HYzB7e12TtOtMgqyrHC6QetZga1inIQCLgAAABA2uK3/QB4Ader4GZ0F3A7V3CzR3iPkX07zME92zHgtkYH3E4hNHLMlrN9/wAZMeYBDzjW6REAAABAuiHgBsDJ+4sOqdW/TVB3ATe7/fV5fgU3Mtc2unoaWUU5c6j3utlfpCrmEGW/gtvU2PcP0JtVlFMVQ5QBAACAtMdv+wFwksxJIbXEvk2Q5A0t7jAH119kKlKpjTUHd+go73Vjffu+XYYo+6G5+XTfP0BaDVGmggsAAACkKwJuAFrl/UWHI4tMxVrcqXMFN8sfohwz4PpDlIf6t+SJ3Ee3panrMOLpS7zHj8zu+weIdauigYqACwAAAKStATzmdOBw8haZ8iq43czBda3tFVzLkML+bX0iC0jFWmRq2MXe69MnpLPvS82NUtbwjse9dKr0rfr+fYBIKBzQAZchygAAAEC647f9ADjz4lVYrV0D7pDzvcez77W3ZY9oD7SRObgdFpnyA244y9v39AnpzVckOeniyfH/AJFzp8UQZS55AAAAIF3x234AWhW5D27kNkFRQ5TPy/Uez74vDR3tPf/0qva5tN0NUW4567WdN1I6/Y7012pv26VT4/8B0mmRKVZRBgAAANIWATcArc55c3CtxbtNkKICbo4fcFvOSh+9Vfr6AWnywvaqaWusIcpnpeaz3pzd83K9Cu4b1dL5H5KGXxL/D9AWcAdyBZdgCwAAAKS7AVySGzicXNt9cMOdhyhHqraSt9DUqI+0P5diV3Bbm72QG87yKrin3pFOn5QuLEjMB4gMTWZ4LwAAAIAURmIJQKtrlckUUqsyrFPAzRkV+02RQNsSCbjRc3DPeve+DfkB9/QJqeWMlJmTmA+QDhXcK+Z4jwN5mDUAAACAHhFwA9BewW1ROPRBA27nCm7ngNvoVXBzcr05uC3N7VXfeGu7D+4AvlxuWi/d/QcpM/vc+wIAAAAYkAZwYhk4nHPKkCncVsGNXmRqZOw39TgHt6njEOXGei/wJirgZqTBZZKZLY3MT3YvAAAAACQQ4zUD4A1Rjl5FOaqC210ozR3nPV5S5D1GB9ymU95jOEvKHOo9P/V2e9U33tKhggsAAAAg7RFwA9CqVslMYbV0vQ9ud6Hx0qnSHXvaF46KNXc0lNU+5PZMgxRK0I+zrY+sRAwAAAAgdRFwg+CkDH+Rqa4BNyR94d+k8/ieY44AABzJSURBVC/q+r4PTWh/HmuBp3CWFBoSdZJEVXAH8OJSAAAAAAYNAm4AWuUNUW6r4EbfB/cj10kfmnjug8Sq4IazowKuWGQKAAAAwKCWlMRiZkVmttfMqs1sv5ld5bebmVWa2Z/M7KCZTUtG/+Kt7TZB1mmRqfJ//2DhVuom4GZ1DLUJn4PLEGUAAAAAqStZJblHJT3snCuS9F/915L0KUlX+H++KmldcroXX845yR+inBl9m6DeVES7C7jRoTZRc3AH8v1vAQAAAAwayQq4TtJw//kISX/1n98g6SfOs1fSBWZ2STI6GE/efXDNvw9uRsc5uB9UJGRm5rS3hbI6DlFmFWUAAAAAg1iy5uD+k6RfmtkqeSH7ar89T9KxqP3q/LY3gu1efLW6VoUkhdWqIWff9W7pI/UuMEYWesoa3vE2QdHHSPQcXFZRBgAAAJDCEhZwzWyHpItjbHpQ0ickfcM5t8XMPi9po6RP9vL4X5U3jFmXXXZZP3ubWM61V3AvXDu+fUNfhihnDZPee9N7Hs5Sh9CZ8IALAAAAAKkrYQHXOddtYDWzn0i623/5M0kb/Od/kTQ2atcxflus4z8h6QlJmjFjhou1T6poda2SvPvgdtCXgJs9vL0tnNW+YJXEIlMAAAAABrVkleb+Kunj/vPZkmr8589K+qK/mnKxpHrn3IAenixFz8HtNPe2TxXcqIAb6rSKcsIquJFgS8AFAAAAkLqSNQf3HyX9wMzCkhrlDzWW9AtJn5b0J0mnJH05Od2Lr8htgrpWcHsRGCOLTGUNa28LZ0utTVH7JOjHGakSk28BAAAApLCkBFzn3G8lTY/R7iTdGXyPEqutgmtxqOB2HqLcEsAcXEWGQZNwAQAAAKSuZFVwB5X2ObjxGKI8or0tnNXxlkOJmoPbVsEl4AIAAABIXQTcALQFXIvDIlPhrPa2cJbUGnVMKrgAAAAABjECbmBMmZ0ruL0JjJE5uNHzbENZUqg5ap9Ez8El4AIAAABIXdzgNACtrlVypnCXObj9DLjhgFZRpoILAAAAYAAg4AYg9n1wrZcB1w+2GVE/MrOOoTZRc3CjzwcAAAAAKYqAGwAnJ3VeRbk382+lqIDbaRhyaEjU8wQvMkUFFwAAAEAKI+AGwLv7UacKbm+rod0F3IxQe1hOdMClggsAAAAghRFwAxDzNkG9ruD6c3At1HVbpIqb6CHKVHABAAAApDACbgCcnJzLUKhDBTdOQ5Sl9mCb6EWmqOACAAAASGHcJigAJ0+fUe7ZVmX0p4KbmSPJpCFDpWU7pff+1r4tEmwTdpugSL8JuAAAAABSFwE3AO81Nss5dRyi3NuwmJMrffkX0qVTpczzOm6LDFFOWAXXRwUXAAAAQAoj4AbAzEnq5xBlSbr86tjtbRVcVlEGAAAAMHgxBzcAZk5OppCaoxrj+FcfCbihBA9RpoILAAAAIIURcAPhBdxM19TeFM+smPBVlKngAgAAAEh9BNwAmEVWUW4998590VbBHZKY43/4Ou+x8LOJOT4AAAAAxAFzcAPTqfrpYu/VJ4m+TdDFk6Rv1Sfm2AAAAAAQJ1RwA+BVcDsH3DhWc9uGKPPvFQAAAAAGLxJRILxVlLu2xUkowRVcAAAAIE01NTWprq5OjY2Nye4KJGVnZ2vMmDHKzOxbtiHgBsGcWrsMUY5nwE30IlMAAABAeqqrq9OwYcOUn58v464hSeWc09tvv626ujqNGzeuT8dgiHIgYlRw4zpEOXIf3FD8jgkAAAAMAo2NjRo1ahThNgWYmUaNGtWvajoBNxDeKsodm+IccDMyuU8tAAAA0AeE29TR358FATcQ3n1wO7fFTWgI828BAACANHD48GGVlJQoKytLq1at6rBt6dKluuiiizRp0qQej7Fr1y4VFRVp4sSJ+vjHP57I7qYcAm4AnJxauwxRjnPAZf4tAAAAMODl5uaqsrJSK1as6LLtS1/6krZv397j+0+ePKmvfe1revbZZ/XHP/5RP/vZzxLV1ZREwA3AR867RsMaR3RqjWPAzQhLIdYLAwAAAAa6iy66SDNnzoy5ivC1116r3NzcHt9fVVWlBQsW6LLLLms73mBCKgrAJy78uv69+kfSkKjGeM7BnXSzlPvh+B0PAAAAGKzKyrq2ff7z0te+Jp06JX360123f+lL3p/jx6WFCztu27Ur/n3swdGjR9XU1KSysjI1NDTo7rvv1he/+MVA+5BMBNwAZIUz1KxOKxzHc4jyR67z/gAAAAAY1Jqbm/Xyyy9r586dOn36tEpKSlRcXKzx48cnu2uBIOAGIJxhOqPOQwziGHABAAAAxEdPFdecnJ63jx7dp4rtmjVr9KMf/UiS9Itf/EKXXnppr48RMWbMGI0aNUpDhw7V0KFDde211+oPf/jDoAm4zMENQCiUodPKSnY3AAAAAKSgO++8U9XV1aquru5XuJWkG264Qb/97W/V3NysU6dO6Xe/+50KCwvj1NPURwU3AOEM02lHwAUAAADQszfffFMzZszQu+++q4yMDH3/+9/Xq6++quHDh+vWW2/Vrl27dPz4cY0ZM0YPP/ywvvKVr+jxxx+XJC1fvlyFhYWaP3++pkyZooyMDC1btuyctxVKJwTcAIQyTKeo4AIAAAA4h4svvlh1dXUxtz399NMx25cvX97hdXl5ucrLy+Pet4GAIcoBCGcYQ5QBAAAAIMEIuAEIMUQZAAAAABKOgBuAcEYGQ5QBAAAAIMEIuAG4fFSOGjUk2d0AAAAAgLRGwA3ApLwR2vaNTya7GwAAAACQ1gi4AfmHDw1LdhcAAAAAIK0RcAEAAAAgBeXn5+v48eP93ideysvLVVBQoClTpuimm27SyZMn+3ScLVu2yMy0f//+OPeQgAsAAAAA+ADmzJmjQ4cO6eDBgxo/frweeeSRXh+joaFBP/jBD/Sxj30sAT0k4AIAAABAUt14442aPn26Jk6cqCeeeKLL9traWhUUFGjx4sUqLCzUwoULderUqbbtq1ev1rRp0zR58mQdPnxYkvTSSy+ppKREU6dO1dVXX60jR470u59z585VOByWJBUXF6uurk6S1NLSovLycs2cOVNTpkzR+vXruz3GQw89pHvvvVfZ2dn97k8s4YQcFQAAAAAGmIf/1x/16l/fjesxJ1w6XN/87MQe93nyySeVm5ur06dPa+bMmbr55ps1atSoDvscOXJEGzduVGlpqZYuXaq1a9dqxYoVkqTRo0frwIEDWrt2rVatWqUNGzaooKBAL774osLhsHbs2KEHHnhAW7Zs6XDMhoYGXXPNNTH7VFVVpQkTJvTY50WLFkmSNm7cqBEjRmjfvn06c+aMSktLNXfuXI0bN67Dew4cOKBjx47pM5/5jL773e/2+HfSVwRcAAAAAEiiyspKbd26VZJ07Ngx1dTUdAm4Y8eOVWlpqSTptttuU2VlZVvAXbBggSRp+vTpeuaZZyRJ9fX1WrJkiWpqamRmampq6nLeYcOGqbq6utf9raioUDgc1uLFiyVJzz//vA4ePKjNmze3nbumpqZDwG1tbdU999yjTZs29fp8vUHABQAAAADpnJXWRNi1a5d27NihPXv2KCcnR2VlZWpsbOyyn5l1+zorK0uSFAqF1NzcLMkbCnzddddp69atqq2tVVlZWZdj9qWCu2nTJj333HPauXNnWx+cc1q9erXmzZvXYd8HH3xQ27ZtkyTt3r1bhw4dauvHm2++qc997nN69tlnNWPGjJh96AsCLgAAAAAkSX19vUaOHKmcnBwdPnxYe/fujbnf66+/rj179qikpERVVVWaNWvWOY+bl5cnSd1WTXtbwd2+fbseffRR7d69Wzk5OW3t8+bN07p16zR79mxlZmbq6NGjysvLU0VFhSoqKtr2i17tuaysTKtWrYpruJVYZAoAAAAAkmb+/Plqbm5WYWGh7rvvPhUXF8fc78orr9SaNWtUWFioEydO6I477ujxuCtXrtT999+vqVOntlV1++uuu+5SQ0OD5syZo6KiIi1fvlyStGzZMk2YMEHTpk3TpEmTdPvtt8ftnL1lzrmknDieZsyY4RJxD6W4+9aITq/rk9MPAAAAAJKk1157TYWFhcnuRo9qa2t1/fXX69ChQ8nuSiBi/UzM7GXn3DnLvVRwg/ThsmT3AAAAAADSFgE3SLf+VPrHXye7FwAAAAAGkPz8/EFTve0vAm6QMrOl4XnJ7gUAAAAApCUCbtAyQsnuAQAAAACkJQJu0Ai4AAAAAJAQBNygGQEXAAAAABKBgBs0KrgAAAAAPoD8/HwdP3683/vES3l5uQoKCjRlyhTddNNNOnnyZK/ev2nTJl144YUqKipSUVGRNmzYEPc+EnCDRgUXAAAAwAA0Z84cHTp0SAcPHtT48eP1yCOP9PoYixYtUnV1taqrq7Vs2bK495GAGzQquAAAAACi3HjjjZo+fbomTpyoJ554osv22tpaFRQUaPHixSosLNTChQt16tSptu2rV6/WtGnTNHnyZB0+fFiS9NJLL6mkpERTp07V1VdfrSNHjvS7n3PnzlU4HJYkFRcXq66uTpLU0tKi8vJyzZw5U1OmTNH69ev7fa6+CiftzIMVFVwAAAAgNf3v+6Q3X4nvMS+eLH3qn3vc5cknn1Rubq5Onz6tmTNn6uabb9aoUaM67HPkyBFt3LhRpaWlWrp0qdauXasVK1ZIkkaPHq0DBw5o7dq1WrVqlTZs2KCCggK9+OKLCofD2rFjhx544AFt2bKlwzEbGhp0zTXXxOxTVVWVJkyY0GOfFy1aJEnauHGjRowYoX379unMmTMqLS3V3LlzNW7cuC7v27Jli37zm99o/Pjx+t73vqexY8f2+HfTWwTcoGVQNAcAAADQrrKyUlu3bpUkHTt2TDU1NV0C7tixY1VaWipJuu2221RZWdkWcBcsWCBJmj59up555hlJUn19vZYsWaKamhqZmZqamrqcd9iwYaquru51fysqKhQOh7V48WJJ0vPPP6+DBw9q8+bNbeeuqanpEnA/+9nP6tZbb1VWVpbWr1+vJUuW6IUXXuj1+XtCwAUAAAAA6ZyV1kTYtWuXduzYoT179ignJ0dlZWVqbGzssp+Zdfs6KytLkhQKhdTc3CxJeuihh3Tddddp69atqq2tVVlZWZdj9qWCu2nTJj333HPauXNnWx+cc1q9erXmzZvXYd8HH3xQ27ZtkyRVV1d3CO3Lli3TypUrY567Pwi4AAAAAJAk9fX1GjlypHJycnT48GHt3bs35n6vv/669uzZo5KSElVVVWnWrFnnPG5eXp4kL5TG0tsK7vbt2/Xoo49q9+7dysnJaWufN2+e1q1bp9mzZyszM1NHjx5VXl6eKioqVFFR0bbfG2+8oUsuuUSS9Oyzz6qwsPADn/uDYrwsAAAAACTJ/Pnz1dzcrMLCQt13330qLi6Oud+VV16pNWvWqLCwUCdOnNAdd9zR43FXrlyp+++/X1OnTm2r6vbXXXfdpYaGBs2ZM0dFRUVavny5JK8aO2HCBE2bNk2TJk3S7bffHvOclZWVmjhxoj760Y+qsrKy2+DdH+aci/tBgzZjxgy3f//+ZHfjg/vWCP+xPrn9AAAAAAa51157LSGVxHiqra3V9ddfr0OHDiW7K4GI9TMxs5edczPO9V4quAAAAACAtEDABQAAAIAUlp+fP2iqt/1FwE2WCy5Pdg8AAAAAIK2winIy3P0HKfuCZPcCAAAAANIKATcZRuYnuwcAAAAAkHYYogwAAAAASAsEXAAAAABIQfn5+Tp+/Hi/94mX8vJyFRQUaMqUKbrpppt08uTJXr3/8ccf1+TJk1VUVKRZs2bp1VdfjXsfCbgAAAAAgHOaM2eODh06pIMHD2r8+PF65JFHevX+L3zhC3rllVdUXV2tlStX6p577ol7Hwm4AAAAAJBEN954o6ZPn66JEyfqiSee6LK9trZWBQUFWrx4sQoLC7Vw4UKdOnWqbfvq1as1bdo0TZ48WYcPH5YkvfTSSyopKdHUqVN19dVX68iRI/3u59y5cxUOe8s4FRcXq66uTpLU0tKi8vJyzZw5U1OmTNH69etjvn/48OFtz99//32ZWb/71BmLTAEAAACApO+89B0dfudwXI9ZkFuge6+6t8d9nnzySeXm5ur06dOaOXOmbr75Zo0aNarDPkeOHNHGjRtVWlqqpUuXau3atVqxYoUkafTo0Tpw4IDWrl2rVatWacOGDSooKNCLL76ocDisHTt26IEHHtCWLVs6HLOhoUHXXHNNzD5VVVVpwoQJPfZ50aJFkqSNGzdqxIgR2rdvn86cOaPS0lLNnTtX48aN6/K+NWvW6LHHHtPZs2f1wgsv9Pj30hcEXAAAAABIosrKSm3dulWSdOzYMdXU1HQJuGPHjlVpaakk6bbbblNlZWVbwF2wYIEkafr06XrmmWckSfX19VqyZIlqampkZmpqaupy3mHDhqm6urrX/a2oqFA4HNbixYslSc8//7wOHjyozZs3t527pqYmZsC98847deedd6qqqkrf/va39dRTT/X6/D0h4AIAAACAdM5KayLs2rVLO3bs0J49e5STk6OysjI1NjZ22a/zcN7o11lZWZKkUCik5uZmSdJDDz2k6667Tlu3blVtba3Kysq6HLMvFdxNmzbpueee086dO9v64JzT6tWrNW/evA77Pvjgg9q2bZskdQnSt9xyi+64446Y5+4PAi4AAAAAJEl9fb1GjhypnJwcHT58WHv37o253+uvv649e/aopKREVVVVmjVr1jmPm5eXJ8kLpbH0toK7fft2Pfroo9q9e7dycnLa2ufNm6d169Zp9uzZyszM1NGjR5WXl6eKigpVVFS07VdTU6MrrrhCkrRt27a25/HEIlMAAAAAkCTz589Xc3OzCgsLdd9996m4uDjmfldeeaXWrFmjwsJCnThx4pzVz5UrV+r+++/X1KlT26q6/XXXXXepoaFBc+bMUVFRkZYvXy5JWrZsmSZMmKBp06Zp0qRJuv3222Oe84c//KEmTpyooqIiPfbYY3EfnixJ5pyL+0GDNmPGDLd///5kdwMAAADAAPPaa6+psLAw2d3oUW1tra6//nodOnQo2V0JRKyfiZm97Jybca73UsEFAAAAAKQFAi4AAAAApLD8/PxBU73tLwIuAAAAACAtEHABAAAADGrpsC5Ruujvz4KACwAAAGDQys7O1ttvv03ITQHOOb399tvKzs7u8zG4Dy4AAACAQWvMmDGqq6vTW2+9leyuQN4/OIwZM6bP7yfgAgAAABi0MjMzNW7cuGR3A3HCEGUAAAAAQFog4AIAAAAA0gIBFwAAAACQFiwdVgszs7ck/Uey+9GD0ZKOJ7sTGPS4DpEquBaRCrgOkSq4FpEKBsJ1eLlz7sJz7ZQWATfVmdl+59yMZPcDgxvXIVIF1yJSAdchUgXXIlJBOl2HDFEGAAAAAKQFAi4AAAAAIC0QcIPxRLI7AIjrEKmDaxGpgOsQqYJrEakgba5D5uACAAAAANICFVwAAAAAQFog4CaQmc03syNm9iczuy/Z/UF6M7OxZvZrM3vVzP5oZnf77blm9iszq/EfR/rtZmaV/vV50MymJfcTIJ2YWcjMfm9mz/mvx5nZ7/zr7admNsRvz/Jf/8nfnp/MfiO9mNkFZrbZzA6b2WtmVsJ3IoJmZt/w/798yMyeNrNsvhMRBDN70sz+bmaHotp6/R1oZkv8/WvMbEkyPktvEHATxMxCktZI+pSkCZJuNbMJye0V0lyzpP/snJsgqVjSnf41d5+knc65KyTt9F9L3rV5hf/nq5LWBd9lpLG7Jb0W9fo7kr7nnPsHSSckfcVv/4qkE3779/z9gHj5gaTtzrkCSR+Vd03ynYjAmFmepP8kaYZzbpKkkKRbxHcigrFJ0vxObb36DjSzXEnflPQxSVdJ+mYkFKcqAm7iXCXpT865Pzvnzkr6H5JuSHKfkMacc2845w74zxvk/SKXJ++6e8rf7SlJN/rPb5D0E+fZK+kCM7sk4G4jDZnZGEmfkbTBf22SZkva7O/S+TqMXJ+bJX3C3x/oFzMbIelaSRslyTl31jl3UnwnInhhSeeZWVhSjqQ3xHciAuCc+42kdzo19/Y7cJ6kXznn3nHOnZD0K3UNzSmFgJs4eZKORb2u89uAhPOHNE2V9DtJH3LOveFvelPSh/znXKNIlO9LWimp1X89StJJ51yz/zr6Wmu7Dv3t9f7+QH+Nk/SWpB/7w+U3mNlQ8Z2IADnn/iJplaTX5QXbekkvi+9EJE9vvwMH3HcjARdIM2Z2vqQtkv7JOfdu9DbnLZvO0ulIGDO7XtLfnXMvJ7svGPTCkqZJWuecmyrpfbUPxZPEdyISzx/KeYO8f3C5VNJQpXj1C4NHun4HEnAT5y+Sxka9HuO3AQljZpnywu2/Ouee8Zv/Fhlm5z/+3W/nGkUilEr6nJnVypuaMVvePMgL/OF5Usdrre069LePkPR2kB1G2qqTVOec+53/erO8wMt3IoL0SUn/zzn3lnOuSdIz8r4n+U5EsvT2O3DAfTcScBNnn6Qr/FXyhshbUODZJPcJacyfo7NR0mvOuceiNj0rKbLi3RJJP49q/6K/al6xpPqoIStAnzjn7nfOjXHO5cv73nvBObdY0q8lLfR363wdRq7Phf7+afevyQiec+5NScfM7Eq/6ROSXhXfiQjW65KKzSzH//905DrkOxHJ0tvvwF9KmmtmI/0RCXP9tpRl/DeTOGb2aXlz0UKSnnTOVSS5S0hjZjZL0ouSXlH73McH5M3D/TdJl0n6D0mfd8694/+P9ofyhkqdkvRl59z+wDuOtGVmZZJWOOeuN7MPy6vo5kr6vaTbnHNnzCxb0r/ImzP+jqRbnHN/TlafkV7MrEjeYmdDJP1Z0pfl/eM+34kIjJk9LGmRvLsd/F7SMnlzGPlOREKZ2dOSyiSNlvQ3eash/0/18jvQzJbK+51Skiqccz8O8nP0FgEXAAAAAJAWGKIMAAAAAEgLBFwAAAAAQFog4AIAAAAA0gIBFwAAAACQFgi4AAAAAIC0QMAFACBFmdl/M7NPxuE478WjPwAApDpuEwQAQJozs/ecc+cnux8AACQaFVwAAAJkZreZ2UtmVm1m680sZGbvmdn3zOyPZrbTzC70991kZgv95/9sZq+a2UEzW+W35ZvZC37bTjO7zG8fZ2Z7zOwVM/t2p/OXm9k+/z0PB/35AQBIJAIuAAABMbNCSYsklTrniiS1SFosaaik/c65iZJ2S/pmp/eNknSTpInOuSmSIqF1taSn/LZ/lVTpt/9A0jrn3GRJb0QdZ66kKyRdJalI0nQzuzYRnxUAgGQg4AIAEJxPSJouaZ+ZVfuvPyypVdJP/X3+u6RZnd5XL6lR0kYzWyDplN9eIqnKf/4vUe8rlfR0VHvEXP/P7yUdkFQgL/ACAJAWwsnuAAAAg4jJq7je36HR7KFO+3VYIMM512xmV8kLxAsl3SVp9jnOFWuRDZP0iHNufa96DQDAAEEFFwCA4OyUtNDMLpIkM8s1s8vl/f94ob/PFyT9NvpNZna+pBHOuV9I+oakj/qb/q+kW/zniyW96D//P53aI34paal/PJlZXqQvAACkAyq4AAAExDn3qpn9F0nPm1mGpCZJd0p6X9JV/ra/y5unG22YpJ+bWba8Kuw9fvvXJf3YzMolvSXpy3773ZKqzOxeST+POv/z/jzgPWYmSe9Jus0/JwAAAx63CQIAIMm4jQ8AAPHBEGUAAAAAQFqgggsAAAAASAtUcAEAAAAAaYGACwAAAABICwRcAAAAAEBaIOACAAAAANICARcAAAAAkBYIuAAAAACAtPD/AZZ6cnXEYRNpAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1152x648 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(16, 9))\n",
    "plt.plot(np.arange(num_episodes) + 1, -11.6 * np.ones(num_episodes), ls='dashed', color='red', label='-11.6')\n",
    "for i, label in enumerate(labels):\n",
    "    plt.plot(np.arange(num_episodes) + 1, rewards[i].mean(axis=0), label=label)\n",
    "plt.ylabel('total reward on episode')\n",
    "plt.xlabel('episode')\n",
    "plt.legend(loc='lower right')\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}