Acrobot-v1 (SARSA)-Tiling.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import gym\n",
    "import itertools\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "import numpy as np\n",
    "import sys\n",
    "import sklearn.pipeline\n",
    "import sklearn.preprocessing\n",
    "from lib import plotting\n",
    "import time\n",
    "\n",
    "from sklearn.linear_model import SGDRegressor\n",
    "from sklearn.kernel_approximation import RBFSampler\n",
    "\n",
    "matplotlib.style.use('ggplot')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "env = gym.make('Acrobot-v1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 6.6720024e-02,  6.9492832e-03,  2.7780622e-01, -8.2222879e-01,\n",
       "        8.5213413e+00,  1.4480220e+01], dtype=float32)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "env.observation_space.sample()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "class TileCoder:\n",
    "    \n",
    "    def __init__(self, tiles_per_dim, value_limits, tilings, offset=lambda n: 2 * np.arange(n) + 1):\n",
    "        tiling_dims = np.array(np.ceil(tiles_per_dim), dtype=np.int) + 1\n",
    "        self._offsets = offset(len(tiles_per_dim)) * \\\n",
    "          np.repeat([np.arange(tilings)], len(tiles_per_dim), 0).T / float(tilings) % 1\n",
    "        self._limits = np.array(value_limits)\n",
    "        self._norm_dims = np.array(tiles_per_dim) / (self._limits[:, 1] - self._limits[:, 0])\n",
    "        self._tile_base_ind = np.prod(tiling_dims) * np.arange(tilings)\n",
    "        self._hash_vec = np.array([np.prod(tiling_dims[0:i]) for i in range(len(tiles_per_dim))])\n",
    "        self._n_tiles = tilings * np.prod(tiling_dims)\n",
    "  \n",
    "    def __getitem__(self, x):\n",
    "        off_coords = ((x - self._limits[:, 0]) * self._norm_dims + self._offsets).astype(int)\n",
    "        return self._tile_base_ind + np.dot(off_coords, self._hash_vec)\n",
    "\n",
    "    @property\n",
    "    def n_tiles(self):\n",
    "        return self._n_tiles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([  80,  201,  322,  443,  564,  685,  817,  927, 1049, 1170, 1291,\n",
       "       1412, 1544, 1665, 1775, 1896, 2017, 2138, 2259, 2391])"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tiles_per_dim = [10, 10]\n",
    "# value limits of each dimension\n",
    "lims = [(0.0, 10.0), (0.0, 10.0)]\n",
    "# number of tilings\n",
    "tilings = 20\n",
    "T = TileCoder(tiles_per_dim, lims, tilings)\n",
    "T[3.6, 7.21]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "test = np.array([3.6, 7.21])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([  80,  201,  322,  443,  564,  685,  817,  927, 1049, 1170, 1291,\n",
       "       1412, 1544, 1665, 1775, 1896, 2017, 2138, 2259, 2391])"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "T[test]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "samples: 100000 batch_mse: 0.0021264814737211517\n",
      "elapsed time: 2.4881842136383057\n"
     ]
    }
   ],
   "source": [
    "# target function with gaussian noise\n",
    "def target_ftn(x, y):\n",
    "    return np.sin(x) + np.cos(y) + 0.05 * np.random.randn()\n",
    "\n",
    "# linear function weight vector, step size for SGD\n",
    "w = np.zeros(T.n_tiles)\n",
    "alpha = 0.1 / tilings\n",
    "\n",
    "# take 10,000 samples of target function, output mse of batches of 100 points\n",
    "timer = time.time()\n",
    "batch_size = 100\n",
    "for batches in range(1000):\n",
    "    mse = 0.0\n",
    "    for b in range(batch_size):\n",
    "        x = lims[0][0] + np.random.rand() * (lims[0][1] - lims[0][0])\n",
    "        y = lims[1][0] + np.random.rand() * (lims[1][1] - lims[1][0])\n",
    "        target = target_ftn(x, y)\n",
    "        tiles = T[x, y]\n",
    "        w[tiles] += alpha * (target - w[tiles].sum())\n",
    "        mse += (target - w[tiles].sum()) ** 2\n",
    "    mse /= batch_size\n",
    "print('samples:', (batches + 1) * batch_size, 'batch_mse:', mse)\n",
    "print('elapsed time:', time.time() - timer)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-0.442202147349454"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tiles = T[2.5, 3.1]\n",
    "w[tiles].sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-0.40066300616932293"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sin(2.5)+np.cos(3.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Function_Approximator():\n",
    "    \n",
    "    def __init__(self):\n",
    "        tiles_per_dim = [2,2,2,2,8,18]\n",
    "        lims = [(-1.0,1.0),(-1.0,1.0),(-1.0,1.0),(-1.0,1.0),(-4*np.pi,4*np.pi),(-9*np.pi,9*np.pi)]\n",
    "        tilings = 8\n",
    "        self.T = TileCoder(tiles_per_dim, lims, tilings)\n",
    "        self.q0 = np.zeros(self.T.n_tiles)\n",
    "        self.q1 = np.zeros(self.T.n_tiles)\n",
    "        self.q2 = np.zeros(self.T.n_tiles)\n",
    "        self.alpha_sgd = 0.1/tilings\n",
    "        \n",
    "    def predict(self,state):\n",
    "        \n",
    "        tiles = self.T[state]\n",
    "        val_q0 = self.q0[tiles].sum()\n",
    "        val_q1 = self.q1[tiles].sum()\n",
    "        val_q2 = self.q2[tiles].sum()\n",
    "        \n",
    "        return np.array([val_q0,val_q1,val_q2])\n",
    "    \n",
    "    def update(self,state,action,target):\n",
    "        \n",
    "        tiles = self.T[state]\n",
    "        if action == 0:\n",
    "            self.q0[tiles] += self.alpha_sgd*(target-self.predict(state)[0])\n",
    "        elif action == 1:\n",
    "            self.q1[tiles] += self.alpha_sgd*(target-self.predict(state)[1])                                             \n",
    "        elif action == 2:\n",
    "            self.q2[tiles] += self.alpha_sgd*(target-self.predict(state)[2])        \n",
    "        \n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_epsilon_greedy_policy(estimator, epsilon, nA):\n",
    "    \n",
    "    def policy_fn(observation):\n",
    "        A = np.ones(nA, dtype=float) * epsilon / nA\n",
    "        q_values = estimator.predict(observation)\n",
    "        #print(q_values)\n",
    "        best_action = np.argmax(q_values)\n",
    "        A[best_action] += (1.0 - epsilon)\n",
    "        return A\n",
    "    return policy_fn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def sarsa(env, estimator, num_episodes, alpha=1, discount_factor=0.99, epsilon=0.1, epsilon_decay=1.0):\n",
    "    \n",
    "#     alpha = 1\n",
    "    \n",
    "    stats = plotting.EpisodeStats(\n",
    "        episode_lengths=np.zeros(num_episodes),\n",
    "        episode_rewards=np.zeros(num_episodes))\n",
    "    \n",
    "    for i_episode in range(num_episodes):\n",
    "        \n",
    "        policy = make_epsilon_greedy_policy(estimator, epsilon * epsilon_decay**i_episode, env.action_space.n)\n",
    "        \n",
    "        # Initialize 's'\n",
    "        state = env.reset()\n",
    "        \n",
    "        # Choose 'a' and 's' using policy derived from Q\n",
    "        action_probs = policy(state)\n",
    "        action = np.random.choice(np.arange(len(action_probs)), p=action_probs)\n",
    "            \n",
    "        for t in itertools.count():\n",
    "            \n",
    "            # Take action 'a' and observe r, s'\n",
    "            next_state, reward, end, _ = env.step(action)\n",
    "            env.render()\n",
    "            \n",
    "            # Choose a' from s' using policy derived from Q\n",
    "            next_action_probs = policy(next_state)\n",
    "            next_action = np.random.choice(np.arange(len(next_action_probs)), p=next_action_probs)\n",
    "            \n",
    "            stats.episode_rewards[i_episode] += reward\n",
    "            stats.episode_lengths[i_episode] = t\n",
    "            \n",
    "            # Get Q(s,~)\n",
    "            q_values_this = estimator.predict(state)\n",
    "            \n",
    "            # Get Q(s',~)\n",
    "            q_values_next = estimator.predict(next_state)\n",
    "            \n",
    "            # [r + gamma*Q(s',a')]\n",
    "            td_target = alpha*(reward + discount_factor * q_values_next[next_action]) + (1-alpha)*q_values_this[action]\n",
    "            \n",
    "            # Update Q value\n",
    "            estimator.update(state, action, td_target)\n",
    "            \n",
    "            # Policy update\n",
    "            policy = make_epsilon_greedy_policy(estimator, epsilon * epsilon_decay**i_episode, env.action_space.n)\n",
    "                \n",
    "            if end:\n",
    "                break\n",
    "                \n",
    "            state = next_state\n",
    "            action = next_action\n",
    "        if (i_episode+1) % 10 == 0:\n",
    "            avg_reward = np.sum(stats.episode_rewards[i_episode-10:i_episode])/10\n",
    "            print(\"\\rEpisode {}/{} Avg. reward = {}\".format(i_episode + 1, num_episodes, avg_reward))\n",
    "    \n",
    "    return stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Episode 10/1000 Avg. reward = 0.0\n",
      "Episode 20/1000 Avg. reward = -424.6\n",
      "Episode 30/1000 Avg. reward = -395.8\n",
      "Episode 40/1000 Avg. reward = -281.2\n",
      "Episode 50/1000 Avg. reward = -325.6\n",
      "Episode 60/1000 Avg. reward = -309.0\n",
      "Episode 70/1000 Avg. reward = -240.6\n",
      "Episode 80/1000 Avg. reward = -201.9\n",
      "Episode 90/1000 Avg. reward = -222.6\n",
      "Episode 100/1000 Avg. reward = -202.3\n",
      "Episode 110/1000 Avg. reward = -244.4\n",
      "Episode 120/1000 Avg. reward = -250.5\n",
      "Episode 130/1000 Avg. reward = -166.8\n",
      "Episode 140/1000 Avg. reward = -212.5\n",
      "Episode 150/1000 Avg. reward = -196.8\n",
      "Episode 160/1000 Avg. reward = -191.2\n",
      "Episode 170/1000 Avg. reward = -176.3\n",
      "Episode 180/1000 Avg. reward = -185.8\n",
      "Episode 190/1000 Avg. reward = -201.4\n",
      "Episode 200/1000 Avg. reward = -167.0\n",
      "Episode 210/1000 Avg. reward = -173.7\n",
      "Episode 220/1000 Avg. reward = -176.0\n",
      "Episode 230/1000 Avg. reward = -183.8\n",
      "Episode 240/1000 Avg. reward = -182.5\n",
      "Episode 250/1000 Avg. reward = -192.5\n",
      "Episode 260/1000 Avg. reward = -154.7\n",
      "Episode 270/1000 Avg. reward = -145.5\n",
      "Episode 280/1000 Avg. reward = -147.4\n",
      "Episode 290/1000 Avg. reward = -160.4\n",
      "Episode 300/1000 Avg. reward = -168.0\n",
      "Episode 310/1000 Avg. reward = -155.9\n",
      "Episode 320/1000 Avg. reward = -222.4\n",
      "Episode 330/1000 Avg. reward = -149.4\n",
      "Episode 340/1000 Avg. reward = -152.2\n",
      "Episode 350/1000 Avg. reward = -147.0\n",
      "Episode 360/1000 Avg. reward = -155.5\n",
      "Episode 370/1000 Avg. reward = -145.7\n",
      "Episode 380/1000 Avg. reward = -146.3\n",
      "Episode 390/1000 Avg. reward = -156.3\n",
      "Episode 400/1000 Avg. reward = -139.9\n",
      "Episode 410/1000 Avg. reward = -142.6\n",
      "Episode 420/1000 Avg. reward = -161.9\n",
      "Episode 430/1000 Avg. reward = -176.3\n",
      "Episode 440/1000 Avg. reward = -166.5\n",
      "Episode 450/1000 Avg. reward = -139.0\n",
      "Episode 460/1000 Avg. reward = -145.5\n",
      "Episode 470/1000 Avg. reward = -135.7\n",
      "Episode 480/1000 Avg. reward = -168.9\n",
      "Episode 490/1000 Avg. reward = -179.9\n",
      "Episode 500/1000 Avg. reward = -144.9\n",
      "Episode 510/1000 Avg. reward = -132.9\n",
      "Episode 520/1000 Avg. reward = -146.6\n",
      "Episode 530/1000 Avg. reward = -135.7\n",
      "Episode 540/1000 Avg. reward = -130.1\n",
      "Episode 550/1000 Avg. reward = -141.5\n",
      "Episode 560/1000 Avg. reward = -148.9\n",
      "Episode 570/1000 Avg. reward = -158.1\n",
      "Episode 580/1000 Avg. reward = -146.6\n",
      "Episode 590/1000 Avg. reward = -154.0\n",
      "Episode 600/1000 Avg. reward = -126.8\n",
      "Episode 610/1000 Avg. reward = -138.1\n",
      "Episode 620/1000 Avg. reward = -154.5\n",
      "Episode 630/1000 Avg. reward = -160.2\n",
      "Episode 640/1000 Avg. reward = -146.4\n",
      "Episode 650/1000 Avg. reward = -126.3\n",
      "Episode 660/1000 Avg. reward = -133.9\n",
      "Episode 670/1000 Avg. reward = -135.8\n",
      "Episode 680/1000 Avg. reward = -157.1\n",
      "Episode 690/1000 Avg. reward = -135.5\n",
      "Episode 700/1000 Avg. reward = -131.7\n",
      "Episode 710/1000 Avg. reward = -131.5\n",
      "Episode 720/1000 Avg. reward = -128.9\n",
      "Episode 730/1000 Avg. reward = -116.9\n",
      "Episode 740/1000 Avg. reward = -143.1\n",
      "Episode 750/1000 Avg. reward = -140.7\n",
      "Episode 760/1000 Avg. reward = -149.3\n",
      "Episode 770/1000 Avg. reward = -138.9\n",
      "Episode 780/1000 Avg. reward = -137.2\n",
      "Episode 790/1000 Avg. reward = -124.3\n",
      "Episode 800/1000 Avg. reward = -138.8\n",
      "Episode 810/1000 Avg. reward = -156.0\n",
      "Episode 820/1000 Avg. reward = -118.1\n",
      "Episode 830/1000 Avg. reward = -120.0\n",
      "Episode 840/1000 Avg. reward = -134.7\n",
      "Episode 850/1000 Avg. reward = -134.9\n",
      "Episode 860/1000 Avg. reward = -141.9\n",
      "Episode 870/1000 Avg. reward = -122.6\n",
      "Episode 880/1000 Avg. reward = -151.8\n",
      "Episode 890/1000 Avg. reward = -123.3\n",
      "Episode 900/1000 Avg. reward = -151.3\n",
      "Episode 910/1000 Avg. reward = -107.9\n",
      "Episode 920/1000 Avg. reward = -186.3\n",
      "Episode 930/1000 Avg. reward = -123.4\n",
      "Episode 940/1000 Avg. reward = -175.2\n",
      "Episode 950/1000 Avg. reward = -136.5\n",
      "Episode 960/1000 Avg. reward = -166.7\n",
      "Episode 970/1000 Avg. reward = -112.9\n",
      "Episode 980/1000 Avg. reward = -140.3\n",
      "Episode 990/1000 Avg. reward = -122.8\n",
      "Episode 1000/1000 Avg. reward = -125.9\n"
     ]
    }
   ],
   "source": [
    "path=\"/home/shiladitya/Desktop/CartPole3D_RL-master/Exp_data_SARSA_Tiling/\" # Just change the path here\n",
    "\n",
    "estimator = Function_Approximator()\n",
    "\n",
    "# Initialize Parameters \n",
    "discount_factor=0.99\n",
    "epsilon=0.1\n",
    "epsilon_decay=1.0\n",
    "alpha=1\n",
    "N_episod=1000\n",
    "f_name=\"/N_ep= \"+str(N_episod)+\" Dis_fact= \"+str(discount_factor)+\" Epsi: \"+str(epsilon)+\" dcay: \"+str(epsilon_decay)+\" alpha: \"+str(alpha)\n",
    "\n",
    "stats = sarsa(env, estimator, N_episod, alpha, discount_factor, epsilon, epsilon_decay)\n",
    "np.save(path+f_name,stats)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 499.,  499.,  497., ...,  149.,  131.,  124.],\n",
       "       [-500., -500., -497., ..., -149., -131., -124.]])"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Fisrt row -> Episode Length\n",
    "# Second row -> Episode Reward\n",
    "np.load('Exp_data_SARSA_Tiling/N_ep= 1000 Dis_fact= 0.99 Epsi: 0.1 dcay: 1.0 alpha: 1.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmcAAAFRCAYAAAAxeQo8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOydeZjUVNb/v6mq3hd6KWgUFcWFUcRXBBHHrcVWX4Wfzjg6MIO4jMs7orjPiDoOoogooo6AOOMCrqOOKO4OIiAIos2iLC4tCCiy9U7v3VW5vz+qkk5SWauSSqr6fJ6nn04lN/eeJPfenJxz77kcY4yBIAiCIAiC8AQ+twUgCIIgCIIguiHljCAIgiAIwkOQckYQBEEQBOEhSDkjCIIgCILwEKScEQRBEARBeAhSzgiCIAiCIDwEKWcE0UO54oorUFFR4Xg5HMfhpZdecrwcN1m2bBk4jsPOnTvdFsVx5s+fj0Ag4LYYBJHWkHJGECnGFVdcAY7jYv7y8/Mt5fOPf/wD//nPfxyS0l7uvfdeHHHEEW6LAQA44ogjcO+997othu1s375dtV5J/8rLyzFmzBj88ssvbotLEGkNff4QRApy2mmn4fXXX5ft8/msfWv16tXLTpGIFIIxhlAohIyMDHHfwQcfjN27d4u/X3vtNdx2220ya2BmZiZycnKQk5OTVHkJoqdBljOCSEEyMzPRt29f2V+fPn3E4+Xl5fjTn/6ESZMmIRgMorCwENdeey3a29vFNEq35ubNm3HuueeiqKgIeXl5OProo/Hiiy+Kx3fv3o2xY8eiqKgIOTk5KC8vx5o1a2RyLV26FMcddxyys7Nx3HHHYenSpTGy7927F1dccQV69+6NgoICnHLKKVi+fHlC96O5uRk33XQT+vXrh9zcXAwZMgRvvvmmeFywCr3++usYPXo0cnNzMWDAAMyfP1+Wz7Zt23DOOecgOzsbBx98MObMmYPy8nJcffXV4n3dunUrpkyZIlqTtm/fLp7/7bff4vTTT0dubi6OOeYYfPjhh4ayP//88zjmmGOQmZmJgw46CH/7298QCoUAAE8//TR69eole24A8NBDD+GQQw4Bz/MAgC1btuB3v/sdioqKUFxcjHPOOQcbN24U0wuuyKVLl2LIkCHIysrC4sWLZXn6/X5ZfRKUd+m+kpKSGLemNO/BgweLdWPXrl1Yvnw5hgwZgry8PFRUVMRY3D7++GOccsopyMnJQb9+/XDllVeitrbW8J4RRLpDyhlBpClvvPEGamtrsWLFCrz88stYuHAh7rzzTs30f/jDH1BaWopVq1Zh48aNePTRR1FcXAwgYmn5zW9+g++++w7vvfcevvzyS5SVleHss89GTU0NAGDXrl0YPXo0hg4dinXr1mHmzJm46aabZGW0tbXhzDPPRFNTEz788EOsX78e559/Ps4++2x8++23cV0nYwz/7//9P3z99dd47bXXsGnTJlx33XUYO3YsPvnkE1naSZMm4bLLLsOGDRswduxYXH311aiqqhLz+e1vf4vGxkYsX74c7777Lt5//32sX79ePP/NN9/EoYceittuuw27d+/G7t27cfDBB4vHb7/9dtx11134+uuvcdJJJ2HMmDGor6/XlP3999/Hn/70J4wfPx6bNm3CzJkzMWfOHEyZMgUA8Pvf/x6dnZ14++23Zee98MILuPTSS+Hz+bB3716ceuqp6NOnD1asWIHVq1dj4MCBKC8vR3V1tXgOz/O444478Oijj+K7777DsGHD4rrfavA8jylTpuCZZ57BypUr8csvv2DMmDH4+9//jrlz52LlypXYuXMnbr31VvGcJUuW4MILL8TYsWOxYcMGLFy4ENu3b8dFF10EWlWQ6PEwgiBSissvv5z5/X6Wl5cn+xs9erSY5owzzmD9+/dnoVBI3PfPf/6TZWVlsebmZjGfs846SzxeWFjI5s2bp1rm4sWLGQC2efNmcV97ezvr27cvmzJlCmOMsbvvvpsdcsghrKurS0zz7rvvMgDsxRdfZIwxNm/ePNavXz9ZGsYYO/PMM9lNN92kec2TJ09mhx9+uOqxpUuXsqysLNbQ0CDbf+WVV7ILL7yQMcbYtm3bGAA2c+ZM8XgoFGL5+fnsqaeeYowxtmjRIgaA/fDDD2Ka2tpalpOTw6666ipx3+GHH84mT54cIwMAtmDBAnHfnj17GAD20UcfaV7Xqaeeyi655BLZvscff5xlZ2ezjo4OxhhjY8aMYeeff754vLKykgFg3333nXhvTjrpJFkePM+zAQMGsMcee4wxFrnvANjy5cs1ZVEyb9485vf7DfcLea9fv17c9/DDDzMAbM2aNeK+Rx99lJWWloq/zzjjDHbHHXfI8t6xY0dMXgTRE6ExZwSRgpx00kl4/vnnZftyc3Nlv4cPHw6/3y/+PuWUU9DR0YGtW7fiuOOOi8nz9ttvx9VXX4358+ejvLwcF1xwAU444QQAEZdnaWkpjjnmGDF9VlYWTjrpJGzevBkA8M0332D48OEyl9epp54qK6OyshJ79uxBUVGRbH9HR0fc45gqKyvR2dmJfv36yfZ3dnbiyCOPlO07/vjjxW2/348+ffpg7969ovzBYFA28aCkpAQDBw40LYs0/7KyMvj9fjF/NTZv3owxY8bI9p1xxhlob2/H1q1bcfTRR+Pyyy/HBRdcgH379qFPnz544YUXMHz4cFGuyspKrF27NmZCSFtbG3744QfZvhNPPNH0tViB4zgMHjxY/N23b18AkNWzvn37ora2FuFwGH6/H5WVlVi9ejVmz54dk98PP/wgu5cE0dMg5YwgUpCcnBzbZy/ec889GDduHD766CMsWbIE06ZNw1//+ldMnTrVtjJ4nsfRRx+Nt956K+aYUrm0kmevXr1QWVkZcywzM1P3N8dx4rgt4XciKPMX5EuEc845B8FgEK+88gquv/56vPrqq7LZojzP46yzzlJVcqSTPvx+P7KzsxOSRQufzyf7EBDuo3TCgbCPRV2Wgpt1/PjxMfkJyh1B9FRIOSOINKWyslK0UgDAqlWrkJWVhcMPP1zznAEDBmDChAmYMGECpk+fjhkzZmDq1KkYNGgQamtr8c0334jWs46ODnzxxReYMGECAOCYY47Biy++KCtz5cqVsvyHDRuGF154AYWFhbIJDIkwbNgwNDQ0oL29Hccee2zc+RxzzDGorq7G1q1bxXtUX1+PqqoqDB06VEyXmZmJcDicsNwAMGjQICxfvhw33HCDuO/TTz9FTk6OKIPf78e4cePw4osvYsCAAWhsbMTYsWPF9MOGDcP8+fNx0EEHOaZ8OcGwYcOwefNmz4RIIQgvQRMCCCIF6ezsxJ49e2L+mGQgdW1tLa6//np8++23eP/993HPPffg//7v/5CXlxeTX3NzM66//nosWbIE27Ztw/r16/HRRx+JitjIkSMxfPhw/PGPf8TKlSuxadMmXHbZZWhvb8d1110HALjuuutQXV2Na6+9Ft9++y0++eQT3H333bJyxo0bh8MOOwyjRo3CokWLsH37dnzxxRd48MEHsXDhQsNr/uqrr2R/GzZswMiRI1FRUYGLLroICxcuxI8//oi1a9di1qxZePrpp03f04qKCvzP//wPxo8fj8rKSnz99dcYP348AoGAzKJ22GGHYeXKlfjpp59QU1OTkGXszjvvxIIFCzB9+nRUVVXh9ddfx7333ovbbrtNZoW77LLLsG7dOkyePBmjR49GSUmJeOyGG25AOBzGhRdeiBUrVmD79u347LPPcPfdd2PVqlVxy+Y09913H95++23ceuut+Oqrr7B161Z89NFHuOqqq9DW1ua2eAThKqScEUQKsmLFChxwwAExf9IwBBdffDEKCgpw6qmnYuzYsRg9ejSmT5+uml8gEEB9fT2uuuoqHH300Tj33HNRVlaGV155BUDEJbVw4UL86le/wqhRo3DiiSdiz549+PjjjxEMBgEA/fr1w7vvvosvv/wSxx9/PG666SY8+uijsnKys7Px6aefYtiwYbjyyitx1FFH4aKLLsKXX36J/v37617zzz//jCFDhsj+hg8fDo7j8M477+Ciiy7CLbfcIsr4/vvv61oJlXAch7feegt5eXk47bTTMHr0aJx33nkYOHCgzCI1ZcoUNDQ0YODAgejduzd++ukn02UoOf/88/Hcc8/h+eefx7HHHotbbrkFEyZMwOTJk2XpjjvuOBx//PH46quvcNlll8mOlZWV4fPPP0cwGMRFF12EgQMHYty4cdixYwcOOOCAuGVzmjPPPBNLlizBhg0bcNppp+G4447DLbfcgoKCApk7lCB6IhxjNGeZINKN8vJyHHHEEXjmmWfcFiWlaWpqwkEHHYSpU6di4sSJbotDEEQPgcacEQRBRHnnnXcQCARw9NFHY9++fWKw2d///vdui0YQRA+ClDOCIIgora2tuO+++7B9+3bk5eVh6NCh+Oyzz1BWVua2aARB9CDIrUkQBEEQBOEhaEIAQRAEQRCEhyDljCAIgiAIwkOQckYQBEEQBOEh0mpCwK5duxzNPxgMoqamxtEyCOvQc/Em9Fy8Bz0Tb0LPxXsk45kceOCBmsfIckYQBEEQBOEhSDkjCIIgCILwEKScEQRBEARBeAhSzgiCIAiCIDwEKWcEQRAEQRAegpQzgiAIgiAID5G0UBrXX389srOz4fP54Pf7MX36dDQ3N+Oxxx5DdXU1evfujVtuuQX5+flgjGHevHlYv349srKyMGHCBAwYMCBZohIEQRAEQbhGUuOcTZ48GYWFheLvhQsXYvDgwfjNb36DhQsXYuHChbj00kuxfv167NmzB0888QR++OEHPPPMM5g2bVoyRSUIgiAIgnAFV92alZWVOOOMMwAAZ5xxBiorKwEAa9aswemnnw6O43DUUUehpaUF9fX1bopKEARBEASRFJJqOXvggQcAAGeffTYqKirQ2NiI4uJiAEBRUREaGxsBAHV1dQgGg+J5paWlqKurE9O6AavZi9Z1K8F4BtbaAmTngMvJBWtrBdfvUKCxDmAMOPo4gPMB334NHHkMuIxMsKpNwKFHgsvM0s5/xxagpA+4gkLNNMmE/bwNKCwC10v9nrM9vwB+P7jefZMsGUEQBJFKsJq9QKgLXN+DEs+rowPYsQXcUYNskMy7JE05u//++1FSUoLGxkZMnTo1ZtkCjuPAcZylPBcvXozFixcDAKZPny5T6OymfetmNM59SLaPKf4DQP64/4MvWIb9/7gPBdfchswhJ6F2xl3ILj8PvW66RzP/vddcAH/ZgQg+9Yb9wsfB3msuAJeTi96vLFY9XvePe+HLzUPRnQ+pHk8mgUDA0WdPxAc9F+9Bz8SbpPtz2XvNBQCAsrdWJZxX42P3on35IgT/uQD+PgcknJ8Wbj+TpClnJSUlAIBevXrhxBNPxJYtW9CrVy/U19ejuLgY9fX14ni0kpIS2ZpWtbW14vlSKioqUFFRIf52ch0sdvARyPvjtWh55V/gzvkN2KKFqulatm8BW/s5AKC5tRXcLz8DANp/rEKXgXzhvbs8tb4aa2vVlCfc0gzwvCfkpXXpvAk9F+9Bz8Sb9JTnYsc1hrd8BwCo2/ULOF9Gwvlp0SPW1mxvb0dbW5u4vWHDBhxyyCEYNmwYPv30UwDAp59+ihNPPBEAMGzYMCxfvhyMMVRVVSE3N9dVlyYAcJlZyLv4cvimPgVu2Gn6icOh6AaLuDoBwJeGUUsYM05DEARBEIQlkmI5a2xsxCOPPAIACIfDOPXUU3H88cfj8MMPx2OPPYYlS5aIoTQAYMiQIVi3bh1uvPFGZGZmYsKECckQ0xCO48CVHQi2c7t+Qp8/8j8UAnheONlR2ZIOY6ScEQRBEMkl3d6lGiRFOSsrK8OMGTNi9hcUFODvf/97zH6O43D11VcnQ7T4COiZUrluK1k4BLCockaWM4IgCIIgTJCGGkMSyDDwc3PR2xoKATyT70sXSDEjCIIgCEdIM40hSegpZxwH+KJm11BIMuYsDU2xpKARBEEQyaSHvHdIOYsHPbcmx3Vby8Jd3W7NdLOcAd3XRhAEQRCEbaShxpAEdMecoVtpSfsJAW4LQRAEQRDpByln8WCknPES5SxtQ2kwkHZGEARBuEK6GTwUpJvGkBQ4PUWL4wA+HNmWztZMx4rUQ3z/BEEQhMdI8/cPKWdOoGY5S7cxZxTnjCAIgiAcIc00Bo/Axz/mLHzfTWChLocEsxlSzgiCIAg3SEdvlARSzmyHExUyJp2taXbM2c/bgLpqh2SzEdLLCIIgCMIRSDmzGw7dY85kljPtW81S0gJFbk2CIAiCcAJSzpxAUMjCJmdrpqqSk6pyEwRBEISHIeXMbjhONuaMRQPScrr+cab705PQhACCIAiCcARSzmyHkwehNTPmLGV1nJQVnCAIgiA8CylnccL97++0D6q5Na3MLEkVi1SqyEkQBEEQKQQpZ3HCnXq29kHRrdllLpRGjJKTAkoPLd9EEARBEI5Aylm8aOlaHBRBaM2E0khRLYcWPicIgiCSSQ/x2JByFi+aoTE4ueXMzAoBqVjXXGggjDGw5v1JL5cgCIIgkgkpZ04gWJQYM7lCQArO1gSSrqCxlYvB33Ip2M/bklouQRAEQSQTUs7iRctNKQ2lIQ03QXHOEuebryLF7vopueUSRA8kfM0F4F9/1m0xCEIDWr6JUEWjYkiVM8DUCgGxlrIUUNYozhlBpD3s47fdFoEg5IheqPR+/wTcFiBl0XNTyixn+hMC2I/fgy39QLEzFSpdKshIEARBpBUp8X5MHFLO4sWMcgYAvH6cM37m34DODvnOVKl7qSInQRAEQaQQ5NaMF93x/WFhwziUhqrSlgJaD0OP+YIhCIIgiGRCylm8SMeQKceTiW5NxLlCQArED5O6bAmCIAgiQVgohPCt48FXrjCRmiYEEGpI60VmlmS/YramqVAaCnh3LVKMLGIEQRBEsmlpApoawf79L7clcR1SzuJFai3LzJQekIw5M+HWVMNt5chU+TRbkyAIgiCcgJSzuJFYwqSWM+V4McNQGioWtVRxF5JyRhAEQdhFensqLUHKWbz4JLUoQ2I5kyosDPG5NV1XekyU77qMBEEQRFpBrxURUs7iRmo5kyhnfFiShpkLQqvEbcXHbPGpYuEjCIIgUgcrxow0hZSzeJFUHm7k6O79YYnCIp3RqFXXVCNpuK2cmSyfvnIIgiAIwnZIOYsXqXJ2/Iju/UyhnEnDapjFdYuUSbem20okQRCOQDO2CcJdSDmLF05jW9mphSUBac3icigNU6IyZjIhQRAEQZiAPgpESDmLF60gtLzC6sUkMc9U81Gbrel2BTXr1nRbToIgHIHaNuEmpsacpXcdJeUsbjjVze4xZpzcrWmlIrnu1jQJdeAEkaZQ2yY8Tpq/f0g5ixdpKA2p5UyoMD4/ZLM1reB2pTM9ISC9GwdB9FioaROuQBVPgJSzuNEYdCYoY36fPM6ZpQkBbitnZtK4KCNNsyYIgujhpLciR8pZvEgVBJ+Kcsb55b9VlBm2aS3Q1hqbt8OKD+vsAKuv1UthJhf33K9uK68Eke5QGyNchT7ASTmLF07dcsaY1HImVWBiOzv+ifvV83ZY6eGfuA/8X6/UTkBxzgiih0ONm3ATM+GcnJfCTUg5ixepcsapWM6EMWeCoqNakTRql9Nfrd9vNEhgsmE4JCdragS/YpF2AnJrEoSzpPmLjyC8TsBtAVIWLeUsFIr89/sjMc50LGea9PA4Z/w/Hwa+3wg2cDC4PgdolE0QBEGkJyY+wNP8PUCWszjhOI1QGmFBOQsgYjmLJ/cUqXROidlYH/kv3EuCIJJMivRBRHphqdqldx0l5cwOpKE0pJYzwDgIrRpuxzkzJauLEwLIrUkQzpLmVgnCo1C9EyHlzAZkVjTB2hPIiE4IiO63UufcdmuaXiHAWSkIgnAJatuEK1gxYjgnhRcg5cxuQhK3JkN8Y87c/nowaThL+9ZBEARBJA/hlULOEVLObCfUFfkfEMacCbM1VRQZLSXMbbem6ThnpJwRRHpCbZtwA0vR2h2TwguQcmY3guUsEIi6Na1XIOa20kPLNxFEz4baNuEGVO9ESDmzG8Fy5o9GKYmnsrldQc2G0nBbToIgHILaNuECzIJfM83fP6Sc2U1MKA0dt6YWqVDpHIxzRhAEQRC6pPnrJ6lBaHmex6RJk1BSUoJJkyZh3759ePzxx9HU1IQBAwZg4sSJCAQC6OrqwuzZs/Hjjz+ioKAAN998M/r06ZNMUeNHOuaMSVyUCoWLdbRr55ESY87MJyMIIsWgtk24QSoYJpJEUi1nH3zwAfr16yf+fumllzBq1CjMmjULeXl5WLJkCQBgyZIlyMvLw6xZszBq1Ci8/PLLyRQzMZRjzjRiafCzp2rn4XYoDdNjztxWIgmCcAR6SRKuQBMCBJKmnNXW1mLdunU466yzAEQsSps3b8aIESMAAOXl5aisrAQArFmzBuXl5QCAESNGYNOmTe4PkjdLVDnjBLemlqL13QbtPJJ0rdr31MzamvGufgDwSz8Aq6uxfmKq1AGCSHmorREepYe8B5KmnM2fPx+XXnqpGLC1qakJubm58Ecj6ZeUlKCurg4AUFdXh9LSUgCA3+9Hbm4umpqakiVqYsjGnAFiJ2fpgyBJFinNUB6mM7BeZGM92CtPgX9iiuVzCYIgiDTG0nsyvZW0pIw5W7t2LXr16oUBAwZg8+bNtuW7ePFiLF68GAAwffp0BINB2/JWIxAIyMrYG/0fDAbFbaHCZOXnox1AZkYGOgFkZ2ehUOVcNQry85Hj4LVI5eZ8sfp52AfUSNKosY/jwBizfM/DHEMNAF9bi+a5NX4/wgCKi4oRkKRpyM5CB4DCggJkS/YrnwvhDei5eA+zz4RvzkR1dJueofOke1uRvnP0CHW0ohaAz+8zfD8UFRUhw8F75vYzSYpy9v3332PNmjVYv349Ojs70dbWhvnz56O1tRXhcBh+vx91dXUoKSkBELGi1dbWorS0FOFwGK2trSgoKIjJt6KiAhUVFeLvmpo4XGUWCAaDqmWo7esIhQDG0NnRAQBob2tDp0n5mvbvR4tKWtbSBC4v9j7ES01NNTifP7achjpJGnWZWdS6V11dLV++ygBWVwsgMjlEK+9wOAwAqG+oB5eTL+7n2yP3cn9TE5ol52o9F8Jd6Ll4D7PPhLV0eyroGTpPT2krRtfI6iPvHp5nhu+Hhvp6cA7es2Q8kwMPPFDzWFLcmn/84x/x1FNPYc6cObj55ptx7LHH4sYbb8SgQYOwevVqAMCyZcswbNgwAMDQoUOxbNkyAMDq1asxaNAgSwqAJ/ArgtAmaK7lV30C/uZxYD9vs0c+XZnMjDkT/sdrWvbm82S1+8AvWui2GAThLmnuMiI8iplql2q6QJy4Guds3LhxeO+99zBx4kQ0Nzdj5MiRAICRI0eiubkZEydOxHvvvYdx48a5KWZ8BJRra1pA7ZzNX0UO/bI9IbHk5SQw5kw8N4mdeBIaJT/1FrD/PAfW1up4WQRBEIQUk5PRegBJjXMGAIMGDcKgQYMAAGVlZXjwwQdj0mRmZuLWW29Ntmj2ErNCQAoFobUUMNc5MWLLSkJhzU3JK4sgvApVf8IN0i1YewLQCgFOwPmiVp44VwjQjXNmp/UogVAaiOO6Uop0vS6CMAPVf8IFLKzelO6QcuYEfn9EOYt7/Um9c2zsNG3QzdJWOUvX6yIIM1D9J1yBgtAKkHLmBP4ARNU/rrU1kxV5PxHtLJ608aRXIRkDQunlRBAEkVyo3xUh5cwJ/P5us2w8lS1pbs1ESNCtmYiClYwGTH0E0aOhBkC4gQW/ZppXUVLOnMDvR7flLGoFs22go51uTa3Zmh6dEJBU0vbCCMIYqv6EG9AKASKknDmBXzIJVrSCedCtqenVtDCd2RFZDaxy5NYkCGeh+k+4CtU/Us6cQJgQAEgsZxbOT1rHmEA5LGYjeSTl/lDnQBAEkVQs9e3p3UeTcuYE/kC3y5y3ya3phLHIs27N6MW6GQk6vds9QRhADYAwR/iRu8H/5zmbcqNYGgKknNmE765Hun/IxpzFF4SWNdaDdXZI9iUqoRU8HufMIaWNSa+F3DpET4aqP2GW7zeC2bXkHRnOREg5swnusKOAAw+J/PBJbms8L3nGg793IvjrLwHjw/YIqF6Qpd3yNHEu32SHe9cpxUmWb5q3fILQgz5OCFewEmQzvSHlzE4Ei464QgAAQbmyUp94BjTvj2zX10bztEVCOXbUccuduIcblsxy5p4YBEEQPRKhDzblHUnvTpqUMzsRKpSq5SzOMWdOfsEmZJmK061px3g2p8aipYBbM/zwJPCvPOW2GETa4836T6Q5pqqdcshQekLKma1IlDPRchbPhABJeApXKqCDEwLELyMnMk8Ub7o1WVMjWEd75McP34At/cBdgYj0xzvVn+hRkFtTgJQzOxEUDo6ThNIQLEwW8pHpCE5WxETGnFlJLD3PrmC8DuBRtyZ/63jwU29xWwyiR+GhBkD0HLz8fkgypJzZiapbM44grbJznHRrWj4gSRJvEFoPT5X28oSAPb+4LQFBEISzpLnCZQVSzuyEi95Onw+i8hHXCgHJsuC4EOfMUno3LWfUSRA9GKr/BOEqpJw5gU+68HmiY84UlinPBMe3NiGA/+JT8C/O8UCAW73ySDkjCABU/wl3oHonEjBOEmHXrl3Yvn072tvbZftHjhxpu1ApC6czIcAKybKcJdIQLI45Y8/MBABwZ19ooQwXLWcEQRCEd0nz/tqUcvbmm29iwYIF6N+/P7KysmTHSDmTII1zppzua6Ui8VLlTKHc2TlUK5ExZwK8xQbi5fZEljOCIAgXoSUCBEwpZx988AGmTZuG/v37Oy1PaiNaziQaVFwveR3Lma310YY4Z44OOnPTcpbeDZ8gdKGPE8INqNqJmBpzlpmZiX79+jktS+ojWyEguk8S54xV7wELhYzzYTqWs6SQhDhn8eTt+AvDm6E0CHdg329CeMZdYGEnl1DzKtQACBfw8pjkJKOpnPE8L/6NGTMGzz33HOrr62X7+XjGU6UzslAa8uWbWEsT+LuuBfv3v1DdIQcAACAASURBVIzz4ZMVSiOROGdxWs68PFtT5k5O85ZPyGA8Dxbqku3jn3sMqNoENNS6JJWLUPUnXEEIUu7BUEtJRtOt+Yc//CFm3yeffBKz77XXXrNXopRGZUKA8JJvbor8/PYr42zUJgQ4UlkTcGuKupmjsTTkON5gya3ZU2EvPQm2YhH8T7/jtigE0XOxNOolvftoTeVs9uzZyZQjPRB0B5/ErakM1mpGwVALpeFERbQjS0fX1mT6v+3G9WWzCLdgKxa5LYK3oPpPuIIFw0Cao+nW7N27t/j3+eefy34Lf1988UUyZfU+0SC0HKeyQoDgquRMDPNTWM5YQy1Ya4tNQkqyXv85wo/crXbEzNnRf3FazkwpqRazThSZ4ayH9ACEMT2yLvTEayZch1l4P6R5HTU1IWDBggWW9vdYVMecCZYvXnLMAMWEAP4vVwIb19gmppj1y08B328Ei7FQmTnZhTFnTrs1Pbq2JuESPXncC9V/wg3MfAiJXilHJXEd3VAamzZtAhCZHCBsC+zduxc5OTnOSZbKqI05E2Z8cVysMqREOjssKSsCMMWLyCOzNe0813L+ad7yCYIgCM+iq5zNnTsXANDZ2SluAwDHcSgqKsKf/vQnZ6VLNdSC0AqWs66u7jRGs1x56dT9JCgJiYztshrqI5ExZ05DQWh7PIwxcD3ZYiZC9Z9wAUv9bnrXUV3lbM6cOQAikwNuuOGGpAiUUvTrL7dyyZZviu4TlBdhmj7nA8IGsc6keSYjXAnjAfglvw2SW3T/sXgtUq4u35TeDZ8gdKGPE8IVaEKAgKkVAkgxU8d/7yz5DtmYsyiiWzOqkPk4wCAQLeOTHPQypiO2+etFqoza0enHkQf79muwn3+E75zf6qVS3SR6EDEu/h4K1X/C66T5B4Qp5ey6665T3Z+RkYHS0lIMHz4c55xzDvx+v2q6noN0hQDBrRmtQIJCxvnkljE1eJVQGk5ieX1Mi+6/zk5JemtFGZZvEv7ReyIbesoZT6E0CHruBOEa1PxETM3WPO+885Cfn49LLrkEf/7zn3HxxRejoKAA5eXlOPnkk/Hhhx/i3//+t9Oyeh+fymxNpVvT5wPCXTGnypApCbZKqI5SAbR7XFiXRDnzsltTXriLZUclaG8F21Dpthg9C/cfu0egG0HYD6vdh/A1F4BtWqeRgJZvEjBlOVu2bBn+9re/oaSkRNw3ZMgQTJ06FY8++iiOPfZY3H///bj00ksdEzQ1UBlzJiCN35KI5cyRYLRWJwRYtJzFbZFK9pizJCvFBvDzngDWrXJbjJ4FWUwj0H0gHIBt/S7yf+VicMeeoJbCSm62yORVTFnO6uvrkZ2dLduXlZWF+vp6AMABBxyAlhb7g6SmHGpxzpT4fIZjznRDaTjRaSomHfAfvqGfnmn+0Egf5yxIraROvTisXpfT7P3FbQl6IB547p6A7gPhAEZ9N1U7EVOWs6FDh2LGjBn47W9/i9LSUtTW1mLhwoUYOnQoAKCqqgq9e/d2VNCUQG1CgFoao9mauqE0nLCcKaxzX39p4dw4E3lx0LXXlm/yggw9DbrlBOEiqTLsxXlMKWfXXnst/vOf/+Dpp59GXV0diouLcfLJJ+Piiy8GAJSVleHOO+90VNCUgFOZEBCTxoTlTGrJUobScCK0RiLrY1pZJN1qWZppk2A5S+92T2hCDx4A3QbCWbTej5aWb0pvTClnmZmZGDduHMaNG6d6vKioyFahUh4fp+nVtG45SwKWZ4RadVPGq/W4OObMC2+nNP8ytBP+hdlAIADfH/+cWEZ0zyPQfSCcwFa3ZnrXUVPKGQDs2rUL27dvR3t7u2z/yJEjbRcqZREWNef80B1zZqicJXlCgNVQGjKsjjmzIWun2mS8Fj7CddiKRZGNhJUztX09sS70xGsmHMfKZLMejinl7M0338SCBQvQv39/ZGVlyY6RcibBH1XOfAm6NU1MCGBdnWAfLgB33sXgMjLik1eRp/n0Fs+1e1kkxyYE0Jgzgu45QbiG2OeZcGumeVM1pZx98MEHmDZtGvr37++0PCkN5w9E6ovehACfmVAaUuVM3XLGFi0Ee/ffQFY2uHP1ot6bIJFAt5YbiIfdmp5r7V6TpwegphD3RCU5CZfM1q4CjjkeXE6u84UR3sBOy1mat0tToTQyMzPRr18/p2VJffxRXVdPOYPx8k26QWiFCikEte3ssCKhOpYrudQSxoN/8Unwzz5mLn8rExpcnRDggYbvARF6Hqp+zaRL4T7OXjPbvRP8U9PBz/+Ho+UQKUZPbGoamFLOxowZg+eeew719fXgeV72R0gIRJUzjtOZjcJbG3OmGUpDsQJBIlh9jgo3JVv+EdjqpXonxFeWloLk2JgzmhDgJiwUAmve77YYsfSsxxDB6brX0Rb5X1vtbDmEx7CzXqV3wzTl1nzyyScBAJ988knMsddee81eiVIZYW1RvTFnPA9mpJxJ3Z7KwfriVGPht2UpY3G6I+blyhwAb06V9lwoDU8IkTTYC7PAPl8K3z/fAudzaZ1emhCQHOiW9kwMvZoWxzCnMaaUs9mzZzstR3ogc2tqKB/hsLFbU8+CIyo3PvXj8WDZcqYij9kTbFnInSYEpCPsy+WRDZ6ZtOk7IQS5NQH0yEsmkoHRx7mVMWcJC+NpTClnQvR/nufR2NiI4uJiR4VKWQTLmd5LlTfh1pRZzjRCaQiVO6EwGIo8zZ9g7VybgtAyp1uj18ac9VjcvPdqEwKSL4X79MiLJpxGfJ9pDftJmiSex5Ry1tLSgmeeeQarV69GIBDAiy++iDVr1mDLli0YO3as0zKmDoJyFgppzwTmTVjO9FYIUCpndliiEpqtadFyZkmZ1BpzlgTLmRd6iZ6qIHpMN/NEXUg2PbXuES5jZdhLetdRU86Dp59+Grm5uXjyyScRiA56P+qoo7Bq1SpHhUs5BLdmOAxN7YznLYXSYCrKGetoB1v4kvg7YRJZvslyEFoziiBTzZoTJ0E4pZzZHI+NiBOPaWdUFwgiOZhpa5zD7wGPYEo527hxI6688kqZO7OwsBCNjY2OCZaSCJazcFhnQkAYCHfp5yONcxazlBMDe18yCUOlgjLGwK/6BKyjPeaYKk7HOTNwF/ILX0L4mgvADJRW592aHlPOvCCDG7h53apxzpIvhus4/gx64k0ljFBWO9ZQB7Z+tX6iNMWUWzM3NxdNTU0y5aympsb02LPOzk5MnjwZoVAI4XAYI0aMwO9//3vs27cPjz/+OJqamjBgwABMnDgRgUAAXV1dmD17Nn788UcUFBTg5ptvRp8+feK7wmQiWs503JrhsAnLmYFbs6ND/ltJ1Sawef8AtnwL7rIbDMWWuhqZI2tl6lvO2McLIxuhUFTBdevLqGc0es/jqnJmeidhB16ctU04h8UgtPwjdwN7f4Fv7pvgAnJ1hTFT6wikLKYsZ2eddRZmzpyJTZs2gTGGqqoqzJkzBxUVFaYKycjIwOTJkzFjxgw8/PDD+Oqrr1BVVYWXXnoJo0aNwqxZs5CXl4clS5YAAJYsWYK8vDzMmjULo0aNwssvvxz/FSYTqeVMC563OOZMkRdj+isIAEB7JIYQa6yP/O/oAGvSiR9l1e0obT9mZnoaWaSEmacxZcvTOu7WVAv54SZekMFBGM+Df+ohsC3fRncw+X9XILcmAOevuSfeU8IYZaiomj3CATekcRVTytmFF16IX//613j22WcRDocxd+5cDBs2DKNHjzZVCMdxyM7OBgCEw2GEw2FwHIfNmzdjxIgRAIDy8nJUVlYCANasWYPy8nIAwIgRI0Sl0PNILWdqOr3PZ1I5kyhfSkWPMcXam8b3hX/or+BvvVQ7gSyEhGF2cSCxzKlOCNBQumJ+JtPV4oX65gUZHKS5EWztSvBPTlMc8JhbM92fg5uQ5axnYfS+ijmuVz/Su12acmtyHIfzzz8f559/vriP53m89tprGDNmjKmCeJ7HHXfcgT179uDcc89FWVkZcnNz4Y9am0pKSlBXVwcAqKurQ2lpKQDA7/eLbtXCwkJLF5d0pBMC1DqdQCCieMWMI1MQtmI5M/Gl//M2/fIsW84SCaWhkr/P6szTZEwIcKYISyiV01T4QLFCTD9sY3iYuPHGmDPWWA9WtQm+E09LfuERCRzOPs3qMmEOoY83rZOrTw7rCZhSztQIh8N48803TStnPp8PM2bMQEtLCx555BHs2rUr3qJFFi9ejMWLFwMApk+fjmAwmHCeegQCAd0y2oqKsB9AVsCPzMICKB2JXCATHIDsrEy06hUkUb7yc3PRJDmUl5uLUEYmhKH+2dlZKFTI1F5YiEZE1kQtDgaxN7pfkH0v5PQqLERm9Bjr6sQ+yTG16+VbsiAsulJYWIgGnbQA0NVUh7rodkFeHvYjonQL6ff5/GAASouK4SvshRq/H2EAhYUFyJLk2ZCZhQ4A+fn5yJXsN3ou0mvWS9e5uxD14nXJy3aDah8HqboaLC0Vn43Tdd0OjJ5L2M+hBpG+IRgMYm+0Ay4tLYEvr8BSWWaer5nzS4pL4C+OfBhW+3zgARQV9UJGku937bTbENr2A0pOr7B8L/Qw01YAoHNfL7EtOFHXOmuLUI/IkJeSFKjLTmP2uXgRM21PSJOfn48mAFnZ2eilkr4tPz/6fojcDzHv0hJwmVkA0P1+yM9HtoP3zO1nErdyFi95eXkYNGgQqqqq0NrainA4DL/fj7q6OpSUlACIWNFqa2tRWlqKcDiM1tZWFBTEdlAVFRWycW81NTWOyh4MBnXL4FsjY706WlvQsb8p5jjz+8G6utDW0qJfkOSrsnm/XMVraW4GWrtVu/bWVnQqZGLRczo7O2Xyasle/+Ad8M2YD47jwLo6Zceqq6vBKayArLVZ3N7f0CBua+XP6uvF7aaobOFwWEwvXG1tTTW4zi6Eo27b/Y2N4CR5hqOLvDc3NaNVst/ouUjRS8ck16Is2w14hUu7pqZasu2ubGYwei7CmEie56PpIjWhtqYWXFuH5nl6JHpf6mprwYWZKBcANNTXgytI7v0O74uMtandVw2uIL57oYbZtsJMtOtEEPLv6upKibrsNFb6MK9iRv7mpsi7o6OjQzU93xR5b4bDoejxyLunproGXFZW9Fj0/bB/P5odvGfJeCYHHnig5rGkLJKyf/9+tEQVks7OTmzYsAH9+vXDoEGDsHp1ZJrssmXLMGzYMADA0KFDsWzZMgDA6tWrMWjQoBgFwYtwBx4c2RjwK3W3pj8QMesazdaUohJKw9CtaSmQH4DG+u6B/Vpreapkr/JDHSO3pujOUs5M1cqvZ7o1PSGTk9gZWDluvOHWTOtpaFJSoF8ntLF/qIVmp29zOd5H13K2adMmzWMho0HtEurr6zFnzhzwPA/GGE4++WQMHToUBx10EB5//HG8+uqrOOywwzBy5EgAwMiRIzF79mxMnDgR+fn5uPnmm02X5SbcoUfCN+1fQLAMbM1nsQkCgchMSqMxZ1JUFBYmi4NmU6UVG5kZjcCqEiNJJFyPtFPWUs6SjscnBKT7OB1xtqa7YsTiOYGSAI05I0yQyNJ/Vg73wPqiq5zNnTtX92Sz/tj+/fvj4YcfjtlfVlaGBx98MGZ/ZmYmbr31VlN5ew2ud18AAFP79PULEwIsKCExszUVljeLVgbNLx3xxWjSemU+gfkQFQahNCyVqXUmY9pWWM+H0vCATEnBY3HOvFAXkk2yLpksZymOzRVFbGvKetHz2qCucjZnzpxkyZF2cJxKdQoEIoqZFcuZmrKUiIVJS5kT9ptpA9KXlSnLndpsUEnj80W961rriOpkZxnGdF4IHnNrKkl3JcELbk3PrRDgVuFpXtcIe7BsODM6QXFc6KrV3jNp3h8mZcwZEcXvjy7fZOHlE6OwKJQ7q4qaljIlejVNKkgxJ+olMVDmNF/KTjQ+FhWJgX/nFbBdP4GFQhGLotU1Q50mJu6bB2SyEzHgpCLOneeUoTS77wThFpqWMeG45olmEqUVSZ+t2WPQmhAQ5vUVKs4nV1KUipzScmaXz1/TcmZgSTAV58zAXchpWM60MzSZTufU5iawd18F+2wxUF8D7sxR4I49QV9Ot/GgSIlh4GJ3A8+5NV1y+9HamoQZrNYT5QoAsQmixznV3SZ2pg1kOXMMldqXkWHs1hSWgBJQna0pVc5UFBq9OmvLmLMEvmJUZ2tG/ydjtqZwbme7TB627APvW868IJOdKK2oSguaK/Q894kqPfGaiTiwe0KA8rgHhjq4BClnTqH2ZRANpcH0lDPl4q67f5Yfj5kQYF6kWNed7KBGfnHOrlHLW7ktoGU5c+QFIShn0XhugYyoDJxCTgeKtkq6uzXF6N8a/93Ac/fYa/LYRIpcFqvdB9ZhX5y5tMPu52hlnHGK1KF4Ma2cNTU1Yfny5Xj77bcBRJZYqq2tdUyw1EfDrQkAXV3d20oUyhm+/lL+m8F44XMtTClnJhQkq8s9qU4IkKAZSsOJCQHR/9GAttrKmQdafrpbzkyNL0kyZpZDSwrWLAYsFAKzEj/RK3h8tiY/6Wrws+5zWwwPY9VzYmM5XuijHcSUcvbNN9/g5ptvxooVK7BgwQIAwJ49e/D00087KlzaISheoa5Y96WAltImEGM506mgMX57pt3Zi/uV1hrVxAbHtZNbmxDgBILlzEA584QiZOZZpDDiOnuc/L/nOl2PjYFTgb/hEvCTrraxXIev2XPPWIfvN7otgXex+zkaGQ96EKaUs/nz5+Pmm2/G3XffLS5UfsQRR2Dr1q2OCpfSqHwRcjLLmZZyprFfIMZyZqHS8jqTEYRsYpQno/ytTghQs5xpuTUTKFMD/sY/RDY6omPOMgTlzOdBt6bytxeEshEtt63X3JquDoEzWXg4DDTY6clI1oQAb1vO7IY11CF8zQVgG9e6LYo9WB5yZvF9ovvBlmb9oQJTyll1dTUGDx4s2xcIBMQ1rggV1PocOyxnMRMCrAw601POtCxnBi8rU+VL0gjKn6kVAhz4igp1Rf4LEwKEZ8Jx8vI8MQA13d2aXvxK9pr7xKWyk6WbeRjmxIolO7YAAPhlH9ifd0pg1Q2q/8HGdmxBePINYO1tCcrlPUwpZwcddBC++uor2b6NGzfikEMOcUSo9CCqcBQWde8SLWed2kqY0RgM3oRypvfS01TOWHf+8gO64phaW02mzOnM1rQcYy1+WLugnHW7NZnk2j1hpIqxLLkjhnNY+UpOEkZu/GQhfrCk3UOPILq03RVDFyfqoSc6FjuJU9lKKM5Z9y7+jfnArp+AbVXW5EgBTMU5Gz9+PB566CEMGTIEnZ2d+Ne//oW1a9fiL3/5i9PypS7KcTSAxHIWMnZfaiJXsCwtPKtrORPyMaEQWB44b5DeIM4Za20GW/mJJBsbOrhOhXIGheXMk5pQ4jKxnduBPTvBDTs1cXESRal4aK7vmkS85tZ0zXJGY84cXevX4xMhTGPiOTK194XW9SvzS/ePFB1MKWdHHXUUZsyYgRUrViA7OxvBYBDTpk1DaWmp0/KlAZJK6Je6NbUsZwbGTMYszdZkX1d2/1Ba3ZT5Sv/r5xp7HqC9bqXhmDN1tyZjADo7wP/1KqCjDcgviCk+bjoUbk2fF2drKn8nLhM/5UYAgN8LypkTs3ETRlK4J5aTStOCjV7SXsBJ5SxtsDjmOO78DMaceaG/thnTKwSUlJTgwgsvdFKWNIOT/QMgV86yc3VP04QxzdmarKkRaKwHPze6mPwvO8DPvl+SltdxeQorBJgYc6YpGw9wKhZBs8s3qYw5Y6/8M6KYAZGJFHYRVc64QEakiXsxlEbaz9bUcmt67KVIY85SsYDEYU6MqU6B67aC5QkBcR438uCkIZrK2axZs9StIApuuOEGWwVKG/QmBHR1Aflat97gnsdYzrorKH/reHnalmb5b15xrizf2PzkBzR2SV+kPFMfxRin5QyMyYPwaimQ8RANpdEdEFgZSsMDpH2cM40PgR6oC2mSrtYb8SONLGcpjVVPi9m0MbqHQR5etsDGiaYPrW/fvigrK0NZWRlyc3NRWVkJnudRUlICnudRWVmJ3FwN6w8hqSzSMWfR8U2hTu0xZ0aVTMdyZgjjta0SmpYzDRnUjmvKYtJypiZbKKQvS7wIsgr30otuTTMzZ1MZIyuuK6h9jKTZfTeF09ecAveUlDMHMHjuMdb06H+1Z+G5/tpeNC1nl1xyibj9wAMPYNKkSTj66KPFfd99950YkJZQQ2VCgKCQdYUAnz0TAqwpZybGnFkN32BmtQA1S5tUB9VbvikcQiw2NEReoZwp45x54eURI4IHZLITTbdm8kUR8Vo8JbdeOjQhgJQzU5gZcybdNnqfJCRMWmEqlEZVVRWOPPJI2b4jjjgCVVXpN33VdqRKiODWZHz8szVjQmnodSCKmq4bhJbvzl+23+hlJbWKGSl+GvlFX8psz06wtlb5MZnlzE63l1I5S4W1NR0sqqMdbPdO5wpQLVSrrrmpDJnclzTSdMxZKihnqSCj25jyaqr0/5oOIi1rOrk1VTnssMPw73//G53RxaI7Ozvx6quv4tBDD3VSttRGa+FztW1LKJQzK1OMdUNpxGyo/tQ+D+bcmjpjztgb88HP/Js8v5B0EoD1zpJ1dakHKBTjukmUM8/N/lE+C+e+5Pk5D4D/+wRroVkShdya2qR7CIFUmq3pZRldx+b6GVMvVGIfeqJvdh5TGsKECRPwxBNP4PLLL0d+fj6am5tx+OGH48Ybb3RavhRGbcyZVDlLYMyZ7OVll1tTa7C9hQkBptyaKgOBpde8YwvQ96Du32G1MWfmr5mfdjuwc5vKAV6RP6e4N8npAPilH4A79gRwvfvGHkzmfIBvv46WwZL3MlJ2xOTW9FDZ5NZ0RDlLgcu2hB1xLrXSGpWT5gqbKeWsT58+mDp1KmpqalBfX4/i4mIEg0GnZUtt1Bq0GcuZGeXM7Jgz5SFeb0IAk//XzV+jUZiaEKAzW1PtvJDKmDMrDVFFMYtYhwTLmUYHnIS2zjrawV55CqwkCP9Dz6mlUPxMhlDJ7OQ0ymreD/Z1Jbj/ORGstQX4fiO4ISMSLo3/1wywr7+Af84b1mTqkbqZ435Nh/O3AZlV3WbSxRpn+TGqfJybyk9yQLx3KVCHEsCUWxMAmpubsXnzZmzatAmbN29Gc3Oz8Uk9GdFwZtFy5jMRhFYWvsKCC4gxbTeJZhBaCw3AKG/V/KETLRqKCQE2NUbGuuUQlD8fF79FMm45ouW1tGgcT6bpTKtMB9GoL/ys+8HPvh+spRn8c4+Bf3Ia2L7dCRfHKlcA0aEZ2olM70wSaWo5SwV3LSO3pi3Y8aiNhs+k4TMyPSFg4sSJ+Pjjj7Fjxw4sXrwYEydOpAkBVjGwnHG/uVS+FqcqCgVL92WqOMbzOnHOtJQznbQx2wYx1AQZlGitisCYMxMCpMpZWBrnTJHGayRDJC9YzrqiChTPA4JS1mWgVNkmkkfGnHmhbEdJgesS+yo73ZopcN2WsHg9hl5NRQK9MEte768TxJRbc/78+bj66qtxyimniPtWrVqFefPm4cEHH3RMuNQmNpSGGI0eAOf3x9RT7oz/BavapJ8tr7Cc6Q2eVnNR2h7nTOqyNOHWVGtEPp3OL444Z4wPg733GriKC7QSdMshdV0wk+5iu7A8/CIp2lkSyohiZPXlNH84iEfcmm4vAu+4V1Mx3tCLCPVTr3+KGw9ftxWcHnNm9EHu5fqTIKYsZ7t378bJJ58s2zdixAjs2bPHEaHSAtUxZ371bek5liuvXlJlWhOzNU2NOZNgZvybYSw0nTFnqukNZNqwBuzdV8FefUZbHqXljOOS/xI2DB3hwpgzL7mbpO3BRd3MVSsPxTlzD7XJS4QCi54WIb3eUBazB6T9eBpiSjnr27cvVq1aJdv3+eefo6yszBGh0gqrY86MrAlqCpcWYYWb0a7ZmlJMxVxLYMyZ2g7DQIaR46xNYywXL2nUrobSMHcdmr8dIYmdnbK+6E0+SdoXssfcmmk65kwM2eJly4cjoTTSTJmwejmG6S24NdMcU27NK664AtOnT8eHH36IYDCI6upq7N69G5MmTXJavtQlrtma3ZYCbugpYGtXxibReGGrxqdSVmjdILQaVhwjSwozsZSU0dgAzUkQcSonguKrNtMTkLs1hQkHHKe41tiy+PdeA3r3he+kM2KzbG0BP+8f8I2fAE5n3CDbUAlWtQm+i6/05pd5MjtBM+NP+CS/xL0WSiPN3uUiqWDxcDLOmYeafEKYGqMs3bZofFDLQ7rTy8p9gphSzgYOHIhZs2Zh3bp1qK+vx9ChQ3HCCScgPz/faflSH5nlLKN7W9WtCbHycmeO0lDOlL8VSoaUGMuZCeVMqYyp5SttQGETM0dl49JsePkbdQiC4qu69BMUbk2hA/ZBbjlTOe3tlyMbasrZZ4uAr1aDBcvAjblKUzR+1v2RjYuvtO5C9MI4OFvLUtSFmHAmDN0CudgJu6JHuGwxSNaYMy/jpHKWApfvCJbH2Qr7PTIWNImYDlOfn5+P008/HQCwd+9etLa2knKmh1qDDpi3nGm7+hSd9bYqsK4uINylnl52rtYYLmhaztiqJeAu/KMysbo8hnlDvZFJZ2tKt+PtwAXFV6mcivlqWM4SmRBgpBBqyWEJ+3ojxhg4tTrmpuVMzWWfbPeXquHMjbcAk/1zrXzHsk+BN6uTcc7SBv3nyHZsAT/1VskOI4VXq72nQH2xGVNjzh5//HF8//33AIClS5fi1ltvxW233YYlS5Y4Klxqo7JCgJUJAUbxzqRs/VbbhSeFZ9rWGuErUWHZYu+9CvbTVp08zSwlpWI5kzY+aTuUzoyKtwPXsyYKxw0nBCRDObNqObOWPK6yHeoDWWsLwtdfDPbNV5KdRi4OyXbS3o8ec2u6VTZNCJAoEhb64p6G0aiXb762eILyEn/PhwAAIABJREFUuN6s5RSoQwlgqtZt2rQJhx9+OADgvffewz333INp06Zh4cKFjgqX0qiurSl1a6pYzjgYm9LV3II8L65Byf3x/7RlYjw0Y5EB4F+eC7Z5XeyBLoVVTtom7JitKbOcmXgLm51BqqWw8hJ3mfB17PMlaDkTrHXxWM5MlpXgC43V10p/aSRyyHL28zagsxP8e69KytIY/CuVxcQ127oeqNfcJ6mgxMRDKlyX2Bc7kHe6GOOsPkerE97iLScNMOXWDIVCCAQCqKurQ3NzM371q18BABobGx0VLrWJjXNmya2pYTljWrPJBEVEOq5Nic6EALZnJ9iyD9XPU1r5tMaQabzY5TOp1dyaknvkk4z9irdBCgqXpuWM77byqa7dGUfZouVMW/mNwXLYivg7KFazF/yd18jLVqtiTvWBaiuuKCdEKO+5dFatnlxx1BPGGNj7r4MbUQ4uWCY9oJbacv6J426cM8eLTYXZmk5M2Ek7HcPiILK4vSFq+9LuZsowZTk79NBD8dZbb+GNN97ACSecAACoq6tDTk6Oo8KlBdJ2LVXOMjLV06t0Wlz5+ZLjKudILGea+Qp5a325bF4f+Z9XEHtMax1QwKS1yWBCgE9jzJl2ofqHBQVJc8yZxHIWllgqZddiQgwpwrM1416Ot5BE+iKZ1UwnM8fGnKlqZ/qnGI1VNJuPGjV7wd5+GfycB4zTuvkScK1sx7Uzh/O3gSSE0gg/NR3hazSCZacClr8vzbZ5TvavJ4bSMKWc/fnPf8ZPP/2Ezs5OjB07FkBkSadTTz3VUeFSGrFBa4w5y1CxcEkH7MtcfZI0asoN40WlgFPLV5KOaVnOGusjG0UlsQdjxsdZs5xJGyRTdWtyim2DhW2NOgRBKQtpTJKQusukA1QTGHPGRRVYFteYM7OdfwIvtJjB9g6UoZZbfS3Y7p3qEe+VlkO1QcDic9KRKx6Rhfw62tX36+9KHukahNaJpZHsxgHljClDw6xdpZ04HTFdr0xY3FJAv08EU27Nvn374qabbpLtGzFiBEaMGOGIUGmBWnuWWogyVSxcgQyNDkGyrRXaQtiva+XSsZy1NHWnURIzFkiyLZsQEOdsTVlZ0tma+km1YLyB5YxXuQ+cT77uaNxjzqy4NS1+DSYUvV/Z2WmUbfMKAfxfrwQA+O6YbkImpVtTOuZMTzmLQ2ahTsc8A6+4NZ2H1dUARcXgfBoBse0qp3YfUNJbfXawl3HCcpZ2FiCr1n+LblA7y04xNN/ky5cvF0Nn6M3KHDlypP1SpQUqY86knaBibJjvn29FOkm1MWfSPDauiS2K8d1KgdbKA4B+EFohmr6goAQyui1Peg3KzIQAWXqNMXMCPhOhNIzKMeHWjBm7F7N0lsWGLyiVVtyalpUKGzujJFnOulGxnJlxcZiynFmTWXcCgVfGtjjszmENdeDv+BO4c38L7uIrVRLYc81s63fgp/8V3GU3gDvtHMkBBwfb24UjylmaKRSWZ5wbtXlhQy3mYc9CUzlbuXKlqJytWLFCMwNSzjRQa9BS5UwxNkz4euUOOhTslx1ATp5+XlKkliC9MVt6s99aospZVKHhLr4S7NV/Rc9TniN1TUkUIC2ri9FsTSmmQogYNFTRcmbCrSnAcQpF04QYsjwNJiGonmP3V6feucrfybGcxcqho5ypxjYyoZzF8/WutbC4qvvEzTFnDuXbFJnMxTatiwREdgi266fIxo/fAzLlzLEi7UNo03aG0mDedufyq5eBO+Z/wBUWmzsh3j5MM46nRp9g2C5ToUJZQ1M5u/POO8XtyZMnJ0WYtERaCSUuRy4jQ7U6ceNvAHfGeeBKe5svg7FuhUTXcsbkypSUjrZommjnkdEtK3//zeBO/1/41MJ0SBoI/+yjQEkQ/lvvVyZSTa+6T9Zo42xwZiYEqHUCVqw6SsTZnxbcmlZDaSTUAeko2LaVoVe8CaVHdbamzvl6eRvJommR81gn75RiaKQb2FWupvXbwej7duFEKA2nP34SgO1vAHv2UbBDj4T/7pkOFWJkCdboE0xb3NIH058ELS0tWLFiBd555x2sWLECLS0ai0oT2kgH6wfUZ1VyWVngjjxGsdOgd5Auy6RneZJOONBCyEfqdg2HwZa+L89HTC9RRvb+AnyrDDoIecMxVM5MjDkT2mtDLcLT/wq+oU5+XLgGvReDkeUs3rEUJi1njDHrHUpCupmK4mMmnV2ofRSYmrllxq1pVRZeu9P3iluzu3CH8nU3VEdKvEwdCaWhfuHMykedLDsGVr0nEYm6EdpofY0VASyWYXBPLQ23SOBjOgUwHYT2+uuvx4cffogtW7bgo48+wvXXX4+NGzc6LV/qwsWOOeOkipNeyIvYzHSPMulAf73BvXprawqIFji9iQXS9CbGxFhxa1qwnLEl7wFbv0Pb4nflB7Ssg+JxLctZIm7N6LlmO1nGm/iKVJ6TiFvTpIXIqU5OuC9W3ZpGijYQx33kdfI1eAkkDWPliVkK26LM3meQv83XrBZgOHLA3nJshCVzQoCV4RDS7Ba9Bf6ua8F2bo9fJgGhTliaqGT3FyZT/Bd+mvmQSy9MzdZ89tlnce211+LXv/61uO/zzz/Hs88+i8cff9wx4VIaowYdCMB3/5Pg75lgnJfPaMwZL490r5vOoOEJY8403K6qeRoitbSpjDmQdlh2jDmTrpeperqGcpaA5YxZtJzJFcQkhNIwm5VjljMVZciMW7P7oE7mibg1+dhjCWZvKxrPg//iU7BnZsJ3/1xwfftZz1eccKClpOuItHEtUFwK7qBDTRRkPX/PkMQJAXFbzqo2RzZq9wGmnocOeuO74sXMmE6148qxaYbtMhUqlDVMWc7q6+tjwmYMHz4cDQ0NjgiVHhg06IwMoJdKTLF48rLi1jRrOdNbaUBrQoBwtGav/EtO9o6Nlr/rJ/DL/9stl4BMaTOy7nDRn4p0YaOvci3LWQJmcuG+7vkFbPN6hK+5oLvj1JTBqsVHKp5cPsYY2MY1mnHsYvMyEfbETuJya+q4H2XpLMqiazlTS+9ix6/1Ml//eWRj57b48hXbWXf+rHk/2NbvdMsFAP6JKeCn3GiuHLGpcuBfmI3wvROj+1NpzFkSZmvGaTlzBgv13XLbM2rzFs5LpL9OAUwpZ6effjo++ugj2b5FixaJszkJHbQadiDDvMHEKJ10oH+ibk1BsdFTzmRjzmIbBX/nNfLOW2MVAfbmC9ENycm+eOKcKS0uwiwrjeSqbk1f5N4I5cfb6YRD4B+fHNm14r/66RNZvkmpnK1eBv6J+8CWa5RpZmyV/oHE4FXcmmrhTGJkMaOcxTHuhanJo5WX9zp+LuqCUirp4ftv7g4orZ9D5J+0Kc+4C/z0v9okYRRpe1+xCPhlh/DL3nLihIW6EH54EtiWb2IPOqmcKbNMxEVtG0z2z9wpVq3WFicEOCVHCmDKrblt2zZ8/PHHeOedd1BSUoK6ujo0NjbiyCOPlM3knDJlimOCphycZOOwo4BtVfLjgQBs085kY84SnRBgxnImzdPqmDM1K5l0n1R+A5eIVqcpugi03Jp8bN7CmDNfVEmzJbiiznOTKYiR/6yrC/D75WMTled0Fyg/Vlct/mddXWD/eQ7cBX8Al1+onj7ZoTQEpV9SLlOWpTtbUy/zOF4QQtmmLGfWsrcF0Z2j8Zy0guj+9CPYqiXgzvudQQEqFyWEvdA6bidembW4bzfwwzfgn58N//1Pyo/xTobSUO4Ow/Xxd2KbcHDMmVHyGEu5XjtIb8uZKeXsrLPOwllnneW0LOmFRHHw/WVa7DIxgQzzX2SWZmtqW84Yr2KtUbrzhA5Jbxko2ee21Sj3EveW2ngCS2PONO6L0YQANauVMObM5wcQcuCLUC29vAx+wu/AnXo2uMsnGpehFE8yRoNVrojMru3q1M4rjlAarK0V7PMl4M4cZTnau7q71cjFIVFg9e6viWclK5+XfKR43XKmVbTegH6/iTYkUU75zz4GN3hYd5E8b+Mla7VVj7xM9T70tKxcCZWncd2hEOAz+UFsJr94ENualXOkpzPjfsEwPInGR5OhTB6pTzZiSjkrLy93WIx0pHvhVi4jM3Z2ZiADplu9iZhE4oBS3QkBYRXFhYOqsqUXL02KmYGsTCV/reNSZdFsx6NMpmKlkafXspwxiVvTBsuZnuFs4hj47orGEupoFxc/Zp99DGgpVLqrMUhegmqTE5TJeQYWCoG/7iJw4/6sfx3CoVeeAlu9DNyBhwC/Ok4znSqCLNLslR21k25N2dJcEsuZmfhvDn2Vs4Y6oK4a3ICBOok0yvapfNiIx0woZ0LbaKgFe34W2KFHdh+Lx3KsWU70f8yj9cjLVG/smxPrf0brHafIk4XD8SlnIjbIaOZDKPYk+fkxs3I1LPZWXcVGEwK8Yom1Ed1W/Nxzz8l+K5dxeuSRR+yXKG3otmSoEggYtiff7Q/Ad/9cmJoQIFR6va9mM25NgxmE/Cv/BHv/P5LkZhqFgVtT03Km5dZUfNHGTAiIKgJWg9CKlrM4sGpBBOKwtum4NaWPTcVbHFsWAzo7IlsLntcoQ3FG8/7IRlenOXmliM/CgpJlOgitmfIViq3qGDiNvPTuSWN9932xCH/vRPAP/sUglUbZmmuDAuBM1GGlAr9fMrnLyPIcD0YvbdfQ6aedmLRgcygNW4nHcibPIPEkmsfjPjFl0VXOPv30U9nvF198Ufab4pyZQW9CgH6j5wYOjkyTN2MqFl4+uss3sdjOXCtrjfAdbOn7YGs+k5dthExJkCpnKrLKZmsa5Kt1X4SXi5ZsahMCIBlzppTTDPGsgWi1jK4Oybka5XM+qGpnKpYz8RlLlVhdmRKYaq+2Tqvhh4I0FpxemXFYzrSssxavjb/9cvC3XGrpHJGWJu1jRqEuxHqqcg/NuDWF89QsDrIF5xWnWX72Rh9YHNiWb8F/uEA7hz2/gH9jXhxlxy8egJQIpWEroqvfQl/GdPoYvTKM8jP6AFcWmH66mb5y5khj6CkY3Tsrbk1TEwLinK2p2fGYlM3qhABp+UYvIO0M9Q/zBp2M6goB0fNEd65FmVTN6iYsnkqiwX/V2h7/2GTpyYq8JEWqWSTVLGeCzNKZYmbGb3W0g1/0lvmwHYC6tS1GP1axrgh99ZL3tfsjq25Nabw/TfewxfydQqtsyZizmPtixvqrmIwigw9rlxuPhRhATFuQ5M8/dAfYm89DC372VLD/vgXYFQlfTQ61D0Ung9Aq8/TCbE29OqF5jtr5CQmhsdugXcbzcexxdMecWR30q0VNTQ3mzJmDhoYGcByHiooKnH/++WhubsZjjz2G6upq9O7dG7fccgvy8/PBGMO8efOwfv16ZGVlYcKECRgwYIAtsiQdjXvI+XzRGTpm8jA4Lh3grPfVbCYIrVimyRlKpr74NNyaasczMrs7Ks2XscGOeNbWZNH9nPVQGuGHJwE/qEzFN7NgvRJBOTScVaulVHDqbk3lcxeuV1mWGeVswfNA7T5wwb7ACScbpgdgznKmasWK7GOVK8CddzFw8GEqApl4WErroKblTO1kN5Uzjf3S2Zwxypm1CQEAFG5Nnbpn1f2WiEItprUQk84qeoP+JcoZa2sFfvwe3KAh9pSnxEuWM0vjt5jGtgZG7x8Ws2Eqa8Zcn+tqO7qtOBwOY9OmTeIfz/Mxv83g9/sxfvx4PPbYY3jggQfw3//+Fzt37sTChQsxePBgPPHEExg8eDAWLlwIAFi/fj327NmDJ554Atdeey2eeeaZxK802Wh98FZcaD0vM7M1BbemTsfMnp+l0jg08jZb0y27NaXphXEzDBg4GDjmeIVb09CvqZ7OcLYmD3VLksRyZuVFoKaYmUF1pl20fMMOUsMdx2kMOlMbmKuiAPLTbpNZxMJPTQe/7ANJ3gDaIuvqMuUMZD1UlTODc5RKtJZ1wcyz+nlb97Z0RQ0Vy1l4xl3gv/hUuith2C87rFkajQoXl9pR+dAwYzlTthHh+QAGyplFJUJq0ZXtl9ZXAzgdF27C6MghUc74px8B//hksIbaxIrTaNcs0TFnWlGDdm4z7wETDWdWFGdrrkVmOMGHyf7phtJgmj/SAl3LWa9evTB37lzxd35+vux3YWGh2mkxFBcXo7i4GACQk5ODfv36oa6uDpWVlbj33nsBAGeccQbuvfdeXHrppVizZg1OP/10cByHo446Ci0tLaivrxfzSA3UG71vzFXAmKuiv5yIc2bQMau588Rtn8Ts7oNv0sPGQSl1OkzW2QEuMwuaSoI0zhnHqYT10Mi7eT/452cBefndSb9cDnR1wXfKWcYvELUXmjBZwuSYM37FInC5eeCGnqKdyPDxqpThM2k507L4aLk1YyxnKvcAiChAXV1AVlbk99pVYGtXge/qAjati5YRx8uyqyt2n/J81UHjZjrd7jRq0/nZjq1iYGAxX71QGlWbwKo2AcEy9TQWYTu2gp96C7jfXApu1O8tnqxRtnS2pjKNmZnWetek59aMW4kwDqXBeF49xp/e5IdE0ZsAJXVBCjHg1OqxpfIU1yDEVbRpQgBrawWXkxvZXrcK/Nzp4K65Hdzw0y3I5qblTKm8qbhajVycaYKucjZnzhzbC9y3bx+2bduGI444Ao2NjaLCVVRUhMbGRgBAXV0dgsGgeE5paSnq6upSSzkzU1lMxzkzKosHhDpv5NLQm2mXmQV0tHWXaUY+ncbGX38J/E+/o/i6UlHOmKCc+RRp1fNmn30c2SjtE90BsKejM4dPOctcnDO1lzLPS9y5THFY8fuF2WBA5PriRe3eCc/P8EWkYgkDoFVZYqw2UgXFKG8A7PVnVZJZ6BBVLWca1j+ZjCrFfrUaKCgCd/ivYsUV6pJ01+6fIN/BtENpSH+qBUmOh2iAYKYMRK2LweSLRN2ausqZTigNy+43jXxUJyKE1WU33SbiQNTNdOKcAd1txVQcRr3yVJ4Vzyc+IYABbM1n4P/5MHx/exRc/yPAdu6IHNv9szXZ4lV0zJwnlqF13Mp+A4UtxTEV58wu2tvbMXPmTFxxxRXIzc2VHeM4zvIYt8WLF2Px4sUAgOnTp8sUOicIBAKmy+iq24s6AIGMDJRqnMPCYeyLbuvl25yXjxadsvJycgGOQzOA0t69Ua2TNivUCZkzSjK2zJedDT6qnBUXl4BlZaJOkjQYDGKvIr8Mvw9635LBYBCteXkQ5qX5fT4I3ZDP70cwGERdIAAuMxMIZIBv84P5/QgDyM3J0b3uQF4+QrX74JN0mMFgEI2BAPQcbr0KC9Ds98vkzggE4MvMRCgjA2GOQ25ODvIlz0T5rPaqbCvJyspGL8VzlaYtLCiAcnVaX2YmgsEg+LYW3edYWlIKX26e+LspNxetAHLz8hDo1QuNADIzM1EULb89Pw+NkvOLi4vBZWWjRi3v4mIxb7Vr4/x+MAD5ubnI1am3gUB395Lt96MNgN/vE+u6UC98Pl/0Psrbf6/CAtRLOt2iXr2QEQxi75xpAICyt1YBADq2fy/ex2BpKTiF5ajV74N0XmRxr0KE25sj5zAme4aFhd3PRKir+Xl5mtcprQdatEueR7EknXBuaWlpTN9X7feBB1BQUIBslbz35+aiDZFnkFNaItZNACjs1Uv1HKC7D+vcWwCtRZ5KiorQkZcv3jPptYURFuuMmb6wNTcXTQCyc3PQJjmvOTfStrMyMyHMQQ4WFYHLzonJozYjAyF0P3876arbp9lPN0f7n0AgAzwXeR4lpaXw5eehY8MaZA8/zXJ5wnVnZmWhKBjE3mjf5Y/WQ6vUZ2aiExEvVvuXy9EOIK9uH3KHjkBzdjZaAOTmF8j6Mi1CrftRC4htwgxdTfXiOyJYWgouK1t2vCUvD82S31kZGegAkJ2dhUK1ep2VIesn9nE+MMjbQU30/ZCXm4vO6PUX5Ocjx+a6YeV97wRJU85CoRBmzpyJ0047DSeddBKAiNtUcFfW19eLbtKSkhLU1HS/Nmpra1FSErtIeEVFBSoqKsTf0nOcIBgMmi6D1Ue6vlAopHmO1Jqhly/f1irf0fcgYM9O8WdLc3f1r63XWIw+Owdob0N7rfYrn5cs2VTf2AAoylWTsaujI2af8hxeIl9Y4hbgeR41NTWRff7o7NWuTvHrvLW5WZmdjFA0sC8v+eqsqakB39qqdQoAoLGhIeKmk15HVxfQ1hb9OufQ2tKKdsn1Msl4J+l90HtuHR0dusf318e+HnlwqKmpAWvVv/bamhpwuW3ib741osa2trYCTZFzO9vbxfL5xkbZ+fW1tRFLqVre1dXg8tpUjwHdVsTmpv1o1bk+acfW3hx51YdD4W6ZovuEeqA0+jU2NMi+iBsaG8Cp3PvwA3+R7VMqZ7xCxvq6OqChUbgY2TPa39jdfsLR9tnc3Kx7nVJZ1GBNkevs1KgPNdX7wCmGIwh1uml/I5pVzuHbI+2uuakJLYrjTc0tqucA3X0Ya9DoJwDU1VSDSdqeVGZWXa26Xwu+JVIv29u6P5f23n87uAMPBgB0dHZb8mv27QMn+eAQEJ5DQ30duAJ7+3hWH1EtQuFwzPUI/VYoHBbdjnX1DWDPPgG2eimaJj8B7qBDLZXHN0fuR2dnZ7TOR5SzUGcHGuJ4f4Wj92///v1gbZE2K9RXoX21trfL+jIthHcWGMO+9xeALf8v/H+ZZu4cRNueQjkTnr9AR/R90d7Whk61eh1NH46+N4W+pqmxux2Eo22jpaUFLHr9TU37Y9pBolh538fLgQceqHnMxkXDtGGM4amnnkK/fv0wevRocf+wYcPEWGqffvopTjzxRHH/8uXLwRhDVVUVcnNzU8ulaRbTlkJ5Om5EufywNPK/ltk9Nzo+q1kntpLsZS0dWK6DqQkBZtya0d9Sd4dR3sKqCwr3nOEsWDVXkHTMGQfE2tHjMJsbztZUkVOcEGDVrRlbJoPevdQYc2ambJ9kMLpZulTcmkZraxqMOVMd6Ky2r701No3Wwudq15Swy8TARakok/H/v73vjtOiuP9/zz7Pcwd3x3GNIigWBA1KU7ALWGKskWiisUWsQVTEFpEklpAQokGxoSa2RJOfmqJRv19NvsSAGjVBUGOJHbuUu+OOK8DdPTu/P3Znn9nZmZ3Z555rMO/XS7lny8zs7pT3fGrW2ygA6sfn45xJnBq0iPvGMXHOkttGSWxvV70oV6GpVHukC9WakjIppXD/+ADo6nfZgdB1dO0X3h9t8RtTKURbR/YdO6vWlM1ZJuGVZG0DQO9Z5Nleavt+QtWi7hrle1CsGwxbYYaAbpGcvfvuu3juuecwYsQIXHmlt8s95ZRTMH36dNx888149tlng1AaADBx4kSsWrUKs2fPRlFREWbNmtUdzSwsDLyRjNW44nUiAeNtqFQDsaTUs30Ro5nzZfPkjBno66AZbFS07+KJVMTmjITP6wzO/YUi4umkm8SlDgFuzuZMdExgbQz+LNBEIJuI/Dhn5i7n7Dfrbw6IQ7zToi2WeL8y+blmocjHFovZOobq1NyvclpgaGoAysVNm+R6UfIcl/hcSvjim6lFXKol1h7+5+/vDjxi1YnPOccVk2eI1NldNmea+kPkTEH8ZItxoRA4UjmhY/Svf+YuouH2mqTKU0F8bvYdC+EQwF6PaKNnEpQYUGxMXMRmnKDKH4rrNXZtqrknzlbVtO4+hm4hZ7vvvjseffRR6blrrrkmcowQgnPPPberm9U9KESsOJ3BL6VBp5Z6OwEeOQM8cpZK5SYYvn3FAjlTZAkIQTepZIUk4q6CnMF3CPjik9x53WTsqxppi6AC1C0g0gXNP+4wiSGF++/nQH/9Szg3PRR+N8ax4gjoui+Bpo0543UesonIOEOBYkEmwf/C14h1SSUu7FrN8+UhyaDMISArC3jrtzcmCK0UTRuj5Ex2vSA5c6+/BM7Mq+TtzCdBuxYx4QCAqOTspWdzf1OAvvQP0PtuhrP49yClZaBtW0BffYm7V5S8Ub3MO1ZyllU/cuI4Z6rjkhOqcduVDgGyMSg+Iz9f8FK0fMiZ/62DDZ5fBu1sEFo2fwHIhShK2E7ZN9GmtKPSP7XX+3XR9WuAhnqQUWO841lBsujPCZRq+vTWx826R625baKAvUUcNEKAWPrE70G/+ix+EPb3yVnblpx0RkRIcuaYBaLdoIn709YGteibG26EIDL6dJOxv+C7TWF7KqM4ZzK1JaX+c/s/mVfoJ++rCaYG7g9nKsORSD20jCVnClWWKk5cJOCr5BiDlpzlIclg5CwuebtK1ayCIZFitjihY7yairPflMdT6uRY1gkaW5uFNoQqB33+r96fn6/2jvzhvly0/FASd2kBcmglZwokjnOm0CAEZIerSzVuC6X6k0GWBUCsJ0TOuPiAeUnOhHAVAfHsrLdmbk6jv7kV7sO/5iR8ydWaAbIJCLHJOHHDz+/OOx/uDXO5+hQByHVl164BFdeBPg5LznoByH6HxF8gi40jYuWLsYMwZGib5skZNylxDgHGoTR0SZ/b28JrhUpyRgiISDqfUefbAxBMPu7G3KCkWzabSc5k9k6BK7+n1iQDPakMbWwIf4N8En/LIFsETW3OVGpNcOro0HuXqXETqhYYiEYSJAOzOZPm8aRwH7lHEotNp56WtFM2iW+RODdwdbk/niU9zhUa3w4t4tmZe+PV4TaEqqYgFdXenw2e8TrlnXpkdnlGdqA60psr0/3dXTkyW6g4Z4nUmgUiMDIE74onZzLJGSfRCQidIekRywJyY48USnLGSZkpBf37k3lIzmTHDObSuAJUJEup4TdQa0qO0acegXvlDGUz+yIsOesqBGtlPMFxbn8U5KzZZmUFNyk+W9wgZA4BQFhyxjePv99UcgYAAwaqz7VtQegB+KjywZrlmtu48fAnNLchF/DDvegkA3KmsKmh1HcI8EVnPjlDY3348rYEkfHjIFuM4hJXZ/NPAAAgAElEQVRahyBRy7L7ZORJlhEhX7Vm0usALs6ZRHKWzYIulcSL0+2WZfVLFxjZt04gNVQuJIbPr7M5E3NGivY1fD+MnFc4t+gQd002HISWLvtfgI2xQsU54xdp1l+Vak2mpsvCfXkZ6MfvJ2xDDAKixR2LlZzRzknwxOTigVSws+RMMqcFRMdwXs1HcqbhZtHrNRclkdRF7u0FKbAKCEvOugqi/l8BUtwv4kYfLUvosCrSFBcZvKgod56XkIXK5dWMiG52VYsRCwYrwxcfg773Vu53KESEIOExJYMM/oSW5e3UAAO1JqKTBMs7ytSrFDlCu1GQnGnChwTQkE266qXoQVP7GplDA7uPFwLc/lNkZ383Wh6fj1WEqYSiE2pNumUz6CP3xJeTDzkzlXIpU0ElUGsak1MzNbB0fFGa2/w01EfPyzwrjchZTNupG32NaX/uyJdERIZCTmqqVVuS3Hl6701wf3Z5fm2QgUmGiEZyxjuQiAQrUX2iiYr/b2ft6WR2tG7CdqpszuJvUvytqUPnEBCyoYXiGWhhbLp7KSw56zJ0VhUSU5RKQhZLbgjgp/UIkTiuc4fUijLJmWrRlsQmCm65YwHwnxWKJgnpm0wcEHioFljNDop+8oHcDZ5JzgKjM//FtzbDXXxd7ro8XOilaos3XpFc6P+bNJQGgyssrK//2/NW7ArJ2eefIDv7u6D1BrGA2gWbs0YuxpspOVMtPnHXqI6pCHYSwmdKYk1t9IJ+Kyx47H5ZGByZ04SR7U+c5EzyDoiGQKmgqoYnO1rJWXfYnHFznawenpC53EYocX1MrRmuN7+8q3zzJP2A9U/T9yYjQCbZVmR/a69XXKsi/1KJuGGdfRSWnHU1CkHsdTkIGeLUmsTJOQWobM4ikjOhHpXImcUbyxeBt2bCl5VV5CbQTCj0fx6Vq5LcbDiURpZTC/Cpd4wTfnPPY2qnpsr5GLlO/M1N+rIykqRvYtItnaTn5X8Am1pBX31Zfp6/X3QI4L+RamGKODEYkDMTuxdArZrWJljm6889g/uP/4X7wv+BbpTE3ddJCxiy7aBNGwW7PESMqMNtcPXvSYZYh4BstK5ABV2AOGf8cSBWtef+4X7gvTe5dhUYsj4UkZxx0kn+e+RDDFTju4AOAQFYP3Kz8n4pQkbYTex3g7815wHu+VVtUPT1PN413dwKulkdTLu3w5KzrkJBCb0oCleQmNiEx9TLEgCovTVDdTjRehQTCFGpSbV1cOU7JLlaUyU5c93knlRMLRA4BCAqYmfIJ/hku+E9wYSap1qTV0npyJnOIcBUHaJasPk6RckZn6FBVY+EgIYkDFKHgNiW5i77y+/kJ6TSOEX7uM0K/f1doL+5De7lZ4K+sVK4XyMtYGjvgHvZ6cKC58b3BVlMMpPvplVrqshZQhIR9LEYhwBmWO9m4S79C+jaL3OX/e2xoE4qbA5pSxPcJ36vDzodg8Bj2omRnFHIJWedUmsKpLUL1Zr0vbfgXn4m3H8/pytEcihJuwzup5E/wmDvngLu0idy0mKVRDxmQ+9e/F24l50R2+LeDEvOugyqHWMeMI1+HEdusllOciZ4ZfqgojRDJDiqSTBfyVloYsrfISCCbDZ5m5gXFrM5y3ZwE4UwweTjEGBqpxZM/LoFR9wlM4mUQiIWWWwlx8Q2qPqdeB9Hzujnq5E975ug774RXnA6fMkh60O89DGJ5CxEzkztwxJIBKSLgPxSVbuDyPLidbrd/yZJyq6Q5EzWBklMMgMpQ6waLasmfIF63nTzo/O+ozRnztDaDPrIvXB/NBN0/Zqo5JbrZ9nzp8Odcxrokw8Db6wya4u0HUklZ1S9aROLue5iZOedL5TFJNwFJmeyTAesnR9/4P37/luRa6Rt45HNgm5sUErHtfcrN5GajWFHe84m1bvBvE4ehfKs7wFYctZVMHQIMCzM7LI4yRl15TZnfPv4Ad6vBJG2qyQkmQJIzkxDd4Tao5j4sx1qpwclOMlZ9WDQV17ISciEiY8aOwRwf5tOEjpixNrwyL3CfWzByOYWtVjJmUQdJl5rapPGk7P//sf797V/CZKzjvC9vPRRUU9kcXZdA3WoTPWXQIxdCJuzCFnSqHIYhDyEwU1xNk7ZrJzEKtD88D3I3nh1/KImqycIpcEM6IWwN+vXyG0PRcPu4DhXB7O74saVO+/86HzDv2/+/o5OLMCmcc4CG7mozRn94hNkLzsDtF7IW/zFJ2pPXBquVxrzMAlk0k7RhCAfBxvXhXvrT+AuWQAqZtqIlGkw9lzNWMgq2mq66dLA/X+/Qva2+Ynv6wlYctZVGLwdAIBMOqjzZZn2wbjdbDYLwshZWqHW9AcA+e75IMXFEbJEn/h/8vvytjnLGUqTfCRnceQiKWEMJGcOyAGHAk2NQEtTrjwe+ag1Te8xVCnSV17w/t1QB7qpNZconffeCwX4lNmc6VQLKnWesGhmZYsmCUtiO9q5S9yw9FEpOZPsunXkTPpInSRnuvekq0/2PWRokRj8N9TnpKiuG8Q6CxAjOaP16yMEt+WR+4D33opvC42RxrFvL4xVd975cK86O1qWLI4YX15gSoCoLWe7YFOqeN+0M3kVTeKceZX4/3L3MGniy8uApkbQF5bq64u891yYkLzAv0exn4rha3SvSeqtmQXWfxXTRir9M3dM1c80Y0o2X8W10xD02afUDmq9DN2SvmlbBKkaBGfJH/OQ4EhgqvePU2u6ri8Ng6DWlEjOmGRN8J6ky5+Rl5235Cwo2VcnFmivkJdaE957dhwg42VKYCmHIhIcU4cA/t22JZWc6b+5+6/lXoJi8X6pekGjIpRdqyQlIjmTpGRySLh8/hrXDUsflTZnEpLDlylV/Ybvof99Hfjgv/LyZZC9kw21cJ98GOTYk8P5cE0DZhqqrGhrVK1JH38IZMqR3t8rngdd8Tyw3Q65C7ISCSh1g77hXDYf+Nr4aGU6b03VMyTNK6mVvnJ2QxFyJmxoTNXfSWAS5yzU53jJmf9s223vnfpstb4+3nHHOxCtIx+4LqhIeEQ7T33jpOXGan9CgjMDtaauLWxMRdLNKcwNbCgNi3xAMkXmyc3jYLpRiI36neXUmoogtCI5MyVLmWKQb50RG1JDCj6SfT6hNFRoqEtOioPE5yS38HQIqjh26bKnDQvlCGdSyZnJhPrf1yX3c4s1FY6HECM50xk8twrqN34x4yOfxy2m/EKsDKUhUcXyRuES72H36vPCv2/6sbxsFSTPTJc9DfrE74HPPxYqS6rW1AxkmeRMdryBS5kmlZxxUm5VmTGSGip1MmCSs6TkTPHsfJoeFudRzOQgbmiUYRYKIDnj5+m4aP0ymzPW/s8+0tcnmhwwKWdBHAKEY1lx/tKpNRWSs7jzOrWm6caLQekQpVBrqkwiuPvpJx+AfvSu9LreDEvO+gJMd4a+fQM58PDcMUZSXDfnEKDy1mQdmt1jSiwzGThHfwfkxBlm1wfgItnnkyFA06bEYCoWtvAwOzHx/a/9IkE70kFZutAUANTESBLol/5TokbhJ+k4taZroNY0VRfJ1JoOUROAbDas1lSSM+G3S0NlShc0mV1MEsR9IyZJfe9NZGeeADQ2qAqRl5knOaOyoKgMWVnYC1eesopHnN1kXFqsbDKHAPrOf3Jt4hEkuObK2ixKzgRyJqo5g0o6QWx4NXxLE+inH0VJYN263OWffxydF1g/rFunH+MiUfKvb/3Tb+GueD6vR8iVI6o1BXKma5vUIcDNzcubW0E/fEe8SfqnskxtSjaF5kAnoRPBbYbdn14G9+dX6u/pZbDkbCuEM2M2MH4f7wcjKW4W6O+F0iC8zRkvHWMDJ7HkzFchJiVXoeCcnVRr8knbgTwkZ8zmjJecsXyQnYiK7rfDfeqR0CSvRHs7aENd/moOWdwrQDKhGag1jaPtc4smF1wzu+4rdRtNnCq03prdFPeKVV+71rvkqUeAbIdxGiEqBh5VQeoQgPjwFW42Ktn9v78EQX5ph4LQbI4hsa6rlv655pIz+voKgC3mkc0BF3BXZXMmSs5UTjWuG843agDa0Y7sHQsApookBO4v5sKdPyf2fdMHbuV+SGwJRaeASAGsL0QJO/3VjabNl5QrUUUHsQU7GYSWmcXdvxjuwh+ANipipklt1sw2KpHcrTKpuaw+1ZpjHIuy98KSs76Azoi8GUnJuoogtNF6CJOsmXItRgB1aahEdNZbkweL4caQDzmjrpdKi5FEtlNX7di1ZSIn8fp8dUTlJkVjPdwrz4pKgEzfTcghgIaPi23L11tThCSZOf34A9T/4Fz59aJaUwWdQ0BnJCbKOmPI2a9u9DzymFq3uFhxoWhnY7g4qlSQcZ6+srnBJ5EA1Cq6TQoiGJSpsTkz2EjRDZz3pkB4KK/WDCQzgrRPtDlTEE36/N/gzj0nmerqs9XAay+DPv8377fjAF995rfVcDMmk/JsVElTWWNd+b/stGgyYAqZWpN9+6zheJaSK659LCQHPzeFbjFQa6rmloC0KrIaSOqh//Oo+pl0kuM+AEvO+gTMFknnql/kfrDJk1Nrkn4Sm7NQNTGSM5Z8WYZCSM466xAgDtK8vTUlkjOV9EFfaN5xjKhIXkzfrRiTiUEaSqNQak2JQ8C6L+XXsraY2ODJJnB+0tYkSabvv62vQ9a2ODRvzOWHVfVX1cKjWRxlDgEAouSM1yS1bdGoYlXkLGbxcrPqZ0ii1uRtSEXC08HbnLFQGiI5E8adahz6Dh90zefhJrtZuMuelksPYwJxG4e1ENWagH4jFyHuggRp5T/N6hbhytSaYW9N+sL/gb7+b/O2AeFnC/KA8u8nqVrTJ1ZK8q8af4o+rnLCEFXkfRCWnPUFmEoweLskNpCYlEzlECBbwAPJGTe5xkmi2LmkUfn5NrDgr/lCnIATh/eguYVCdAiIMxCOg0vNCU6kOcJ9huSMvrVKLqkRjtFPP1JLiZImdpZIzmL7guuaSSNlhNJArckWY/eGufo6dHWKSKVzUicVWVCSM51aUyE5E+18+K6wZbOGnAltZBsvnVpTPHTdxaBNG4M4dkpD7LdfzdldxeWrDIgDp9YUJWcRtaaO+Ag///l30N/dCfrXx6LXiv0z5BBguBmTfVddzDVRxS2+R53kTdkWmVozOm/RV2LInyqURiTepWS8ez/0ZQbvTLxOUJvHlaOT1gFayTyl1AtD9MWnsdf1JCw56wswXeB5NQubbHziRF03R85Cak2vbDL1SM4hwD/P73xjJFGE2Xv1pM2ZMIGQfEJp+JIzIjoExKiVYr2s4gK96hBZIAzfbX0t6N+fkpQXnqjpH+5T9iuqCgSpABXCZACIV3G7rllQ3kgQUhpRa9LmjdH72CKf1HuYtS0OHR05taapulsXeJPBNOQHX07t2qj3LA/hHbKxKg0oyiCzOQPg/vwK4ANfGqnoG+7N14ZVhYp2hPojkducUUG6SrU2m0Kbmhq9f2X5FSOqSz7sjalntcSWsF2zkYt4ryqkXXw1997kBXWOLVdmzC/pn3zA6Nq14flLanPmRuf1DomkXPibsuC7KnKWNHerMs5ZnjZn2SzcH5wF97qL4q/rQVhy1idgSM6K+nG3+PeEHALkQWidux8HOe2C+FAacZKzTJ6Ss2DQM8lZHqKzMRNRctzJIIccLbQpj/RNLM5ZErVmHIGJiyWmg7jrjc2bKuDLT3P1M8gm6nxDaYiQ5cmMa6+bNZNOiIskFciZ68K99PTofYx47DRaX4cI5SLgo6M9936UzyCxlZMdzxtcObVr4S64XH2p0I9IsT9HxErOJB6gQDjafVzfYB6aSSVnujhncSo5GVidsrlLHF/81GMckzDrkRs3geRMVHFHAlxH76cvL4N7x8805RpKzvwxRdd8Affq80D/9w/RtvHIdoScfACo8wQzleXq9+DOOx/0y08l/UQz56jUmqq5SrVm6FTT+Tp5dSMsOesLMJRghLww2T18KI3SAd5EyGzPAI8XOY4Xjy1ftaZPhEjeak14k2Mecc5IaRkGnH1J+JkAtdODDIOG5lRmxMlJfdoNyFkc+RIn7iSIqGnzCQ3CLxqSyUhG2ADO0NmQTIRUMRKVlqR8aiJ1Ej06I2pNxbtl5CwfqWVIkiB5fl7ipwztIP7OLciUUrjPPQOqUmGaIMlzcYGU3eXP5CQlsTZnBuU3NymDrlK2OeDmg0hyct7mjCjUmkkdcVTkRDYXxJkqqMiHAHfJAriLfhR6X6xf0w9y9o509fug777p/S2qM8U2G9YdgWojIeL1f4OuegnwnTWCUCeKMtw/3M+p2/35WUdemcRyY0O0TImXqvdbo9YMx9KIrx/Qky+ZhPKfS0GbJJL4HoIlZ30BpjZnPMSYZa4L0r8Ezg8Wgux/iLzsiEMAR5biJCFJ46IxEALa0uxPFHmqNQMVrHBvEknT8J28fyM2ZwbemrFpcAqo1swn0wTfbSQ7Saoymv3oHdCmRvN+t/bLnJ2RKH2VNevxh4DXDBIpi2otSsM7a9VEzshZHsSY/k1in8SD/y5KmzNFGABKgdXvgT64BPR3dyVuWwDFc5FzLo0eZCTknf+APrQElBHpOAmPKsuEeNlPLpGfYF6acfkq+XfHNmWi2jJfohLU6dchI2fixoR/XFPJGQC896awCfLtHX+Rs3d0F1wO95fz/B9CajTxPQt162MjcuPOQHIGwHMKCJmUxNQlOFkACDYo9LPVoBu4gMiix2VHu6QfqWzOhHtF8BvFkFZTseZoPKTdOaeFq1/zBegDt8K955ex93UnbPqmvoB8yJlMcgaAjNxd2B0YOgTEESemQkwqOaMU7i3X5dqRj1pTRQxNwnqM3wfOAYeCvvqyNzGJ3ppMSqLaZcdFwQc6p9YUCSF7zmEjcmpLHbhFg8qeQWGXQZc/A/r2a17qHxNsavG8GAcMzPW7mL5A//2cWbmiWtMVQmmo3i1beDudTDpeGkGf+ZNZOby0xJcqKD0zU2n9rl+l+pH1efbdxXcZJzmSeWsmgUo9JvtNqbquJCRJVi97Rs4Byn3uGeDLz0DGTBDu5e3GEtabxFszsOeU1Cu733RzZ0ioAQBFRaCyGISGcxVtbwOBjJwLG7T29miZXBBengzSl5f5AYAN89UyKNaMxInk2fjI1yGjC2AlZ30CfsccNQYAQIaNML+Vtzlj4Psz3+cDY26/W/CELI54BaE0kpMzrH7P+zvr5ic5S6kkZ/p9B6kZArLXAd4Al9mc6WyLCOInRDdKzsh0iY2UDCovuySeo/y1MhujuFhA69ckW6DZZO9KyBmTXBh8kxAiNmeGas0grlP0PDno6wkaEH1+I3Ws+O34RajZVxGJqaAYTJwYFAu21KyAtSVi1K1R1XcmtiLrAzwhi0jO2Lk4cpaf5Mxd8QLoKy/knpGTnNEHl4D+/cnYb5SYFPJqzX8uBX01RiqskZxRUVqoCReTu9F8rNIvPgF98A72y7wMJuFUkVcmFPPbTNvbJOpL/3dHe9g54eFfg77495hGK37owgEZg5ljENAtmw0kll0PS876AvzJjhx4OJxf3AcycnftLWTAQO+PFBdKIzipYGdUJGfcdXHkrCjPOGchg/WO/EJpqNSaJnYJQXtJTspFnChJzNfmjBnpFnHOCaa2Y6o6dQbHPPjFsVli4yTzYuORZIJqaQL94G3Q997wfvPvsKTM+zepajYiOYs6BEihCmQJgHzLkBwD8i5kQM7oP/4X9O3XvL9dF3Tpk/4JN6fyU0VZLy0zb58I2RhlfSAJOevoSETOoh7LjJzxMelibM5U/SzP4M/0VzfAvfsG0OXPeAckm4KIJDlEzhKSQv75P/kA7pIF+msDNZ5ocyaMb+Po/tRcysYH69U5wIQQb3MWZMxgbW5vl6j4/TreXAX6giT9nBIqRwJVijiJM4ROywEAn62Ge9FJoC/9I0HbugaWnPUF0ByrJ1U18mv2OiD8mwWNZQtwqGNyE7UszhlJSM4SemuSaUdF2kSz2c5JzsR7TXZO7PkIQbCDd5yoeijOqzHW5swnfHxqKVPpkUpClmTB4sv4XGK8rXM3T0DO3Nvme3Y2n37k38u9F5aZIiM8+25jQQ4/Xl29mGuQUiHJukpylvXUNl98Ej2XpI9J1ZrCwlQzRH7rG694f/xnRS4gLwXA2+jIUDrAvH0iZM/W0Q5aXwv31p8Ix+MM4tvMVVzr10SvZd8+RM7EUBrMWxPqujprc8YgK1/MJsBLNxPWS5/4vfnFkTA18TZn2u/Ax1ozJdRCnDL3T79B9ooZ+vHOlgPF+6H3LPL+CNSaMZIzAHSFoXkDEJDY7KWnhzc2yg2a5LhiTXAfuQfuTy8LH/xSMnd0Myw56xPQe8ClLhACbg6s8v5t9j1nsgZqTcFeiJiSs3QytSY54lvAuMnhHWq2I0+bM4XkzGTHGZAzAPW1niE579pvgph6KEs3lBc5k5MwMukg87bpbJd0UbRNduJiTlMGnkSy1Fqi5CybTSYtNQxCCzcL94ffl6tf+D42dpKuwughMY6VSg3pk59QpgdK40NYADkpYz6QvcuODtD335Icj5OctQPUTGLjXntR9DsE8b/iJGd+/eu+VLclqXpx/Rpkb5wnaaTEGebvTwoHuH6VtF4Gk7HtCuSMr7df/1jJGRWybrgP3Aq89ar3o6EeyMfTkFLPdrKx3nwzpns/vDmIKrcmoN4gMK1P+EbvHzGuoXKDJvNOVzhILH1CXkYPw5KzvgDWwRMsZKTCl5z5CyLZc2/FlTLJmaSiQjoEOCmvDpbLDvAX6k6QM/FWE1sNXq3pg370bjJy1hGXlJqpNXlypnBU2G4HkKO+zZWrIGfj94Hz818btk1Hzgqg1lSRCX4iVJEzGrYzdC7/Kcgp58OZp/CYksQ5k14WJzXlvi0Zv4/6OkCuSRG/i+r5Zd+PuqCahY0kVGuSfaZyP6L9lnZIVEtAwSRn3rViHkSJWlNMEs87KvDx08SyE4A+/UfPe1JEXH9gsR87o9ZkMDFZECVnPHkRyBl13ZB02/3hTNCNOakR/WdOLUif/5tcOp4Eug2t3yfo038MQoPElhOn1gTkY6SkDGSfKdHjlMrTcKm+rex4ku+qmxu7AdZbs09ALzmLYOwkkH2mghx/KlDcT9iNyNWaZOfRXhRuMWaYiFQq3PnTEu9O3f0RNWRHfumfUsKEyIz7jWw1/Pby9Q6sTNaOOOmUTK2psLtyvnkKkCnK8QFRQhOSYhqGCdEEeo3kMhRhEudMaZDLtV8R/DgSfdxx4Bx6LKgY3yyoy9RbM+bb8/1Ou5jKJGcCYVBIzui7b4D+93XhINXbDCaVnIXIvmT8JbQfA+A9Y3E//XUAsMtu0Y2QNmYVjIKA0qRekyrEtMP56Z1wb5gX7sf51pvOANCNKcFZhQrk7POPQevXg1QNAn3kHtBnhWwfa78CymPyHHcGOvLC5pO2LbnQIDxG7g7atiWXLkum1uTHlMxEI5WSryOURuMeArHS88ihq8+VXytDLyBnVnLWF5CH5wgpKoZz3uUgg7cDGVgZ9uRSkCgy4xI4P7oJZEB5fOFC9P1A/WlKHlPpaMDZzqo12Ssq89suFY0LkAS9dc66JKFaU2NzRt1w5gZVcNzSAaGFWbpLZGWmDckZCwapglZyZrCoqxZZTjJDVJIz1w2TC9Z/VMGIXSrEOVNIzu5frG5vyItUQ87EcUdIhPASFZlavwbuTT+OlqdTCSUlZ3x/kr03PqOBKWQhEFRwUhLJGatbs3HRId+ctiLinqW4BCgf6KkFGfJVa5o4vIjOKvwY2+iNV/fem7xTy5+O3E5VUsZ8wX+HuJReQPz3yBQB6QzoU48Aa7/wr2/zTDt48N9CRgadFKSbDErl3uWqjZisrQm+K9XZ43YDLDnrCzCIHZUIob7PSc6Ki0F23FV/vyo1Eq8yOucy+TWAfHeUr0NAQM58j9aJ+4GcfSnIcacY3Mw7BMATqZeUJWoHfexB9UnX9QgFJzkjKrVmWXlYChMn9TKVnOlQCLWmyeTIJLEyu0DeTkfmJSy0JxRp3tSTjQdXNklKztKZqFSlv0bKzD9L80a9E4bKhk+FkJ2T5L1lO9RhOxSgidWaKslZJ8JxAOYJyHVobADlyRePdNob77yJhUxyZvJdTNSavLE8EO5jLPZda2v0HMPaL/1Thht2Xbv5cjojLUr78fk2cXle29ujwmeerMny4jqOwnyH6scOuPfS2b5nJWcWJmC7D5KPZElbuOF1fN2qSYi7xtlvmrqsVApEHIEdHSAT9gWZfLBhg1hZoiSKwNn/EBCTiTIgYSxJvCoshxp05T9jTlLPsDpkc6ZoV+mAnFcjEL8wJY0XpoJu8TOUnJFTzpceD8AkZyJcNyz5kdgAhpqz9AnQR+4J388wek99W4Hwt02afzWTiS4QmQRkastm4JMP5OcGD0Pq108kSzsGCJJHyfn33ooav6swYT9P4iyzI1Ohoz3G7kfI63nMSWZlMuQZSkMEXfoXuFfOiJ5gAafFeVUm0TEZcyaSM/auZOSMhUgqLo6eY2AEzjQ3ZAJyRv/6Z7MyZWCmLvy7lNk78o8kI1uquZdC78AEcJ7C3PvJJ7uKJWcWRuBCaRQGilAapncfd4qcRJmSmlQKNBIxPAsyZBjImbOTNSZYzPIIGsheA1MHpZOFBNHC9y4kIZszxSRfNiAc40pQMZAdR3p/DKxUS87KK0CO/o78nGyC0k3wJq+0ox1k7wMlx/MgZ45GrbnuS8+NfuwkT/3HkzPVNxMllbxUVEeEuLHhXHKd9w5Fu5ckacLiKzNrk4gQue1cv3WOPMHLQNFZhwAG8XjSTUWhJGcqsJzC4rwqCxVh8l1M5g32TtraItIvZ8Ylnhf7h+/A/effpXMz/dcy0M8/NieuJuSsKOEmRYZUxiee3LuU2Zzp+pXsewDeXKqzkeXL5/tePs9nyZn6unUAACAASURBVJmFCZwjv+0tRqPHRs6RaUeBHHhYsgL5zj9kWOL2kLJyOOdfKTmRwOZMnHgZUUgl7JLihJ+Ev+YjOUtiE8S8NYu5CVJW9vAdPQLHT6TNG0PPRo4/Hc4PF4HssLOSEJADDgM55Bh5Wyqqosd0Nj0mkjPXlRM/Xr3BjMvFibpmSPj7BfH14vsAKR/ovUe+fSrCGiFnvARYN2lzKv899/JsatqE3XsCcka+ro7pFrybpAQm9P6S3RpBOuNJB5PYnLW3xwQCFY4nJZ5dTc4YIpIziVrTRPpi8u3YO8l2RMdfUTHI9jsBAOgDt8jv39QK9/rZcC8xMduAmWNHWjIOkm5QU6mo3XC7JLemjmApbc5gpNYM+i3f95JKyGFYVxfDkrM+ALLzKKRu+b3UUN857QJvx5UEvETANH+i4v4Q2I5TEz+KEBKdmBg5SzgpBHZDsibttX/8giBq0QwkZ44YTy4OMm9NgUQ4l16P1HW3ec0gBOTU73snNjaGJGkklQLZaZT3Q0UI+LygkYbLApTqyJmhNFK36LLn54srr4Bz7mVSmzOt+t5Jee/ASHIWbhtJotYUHz+VikosdOSMf4f9Y1Iz0TwlZzqbsyRIZ7yF2lRyRgiwoRb0qUcjpygLGMzb5CWVMvYUOZPZnJl8F5NrQgbxEqKvs2FkKJTNGSCXLCW1a2XjkXuXUttFTb8i35iuOENBTaRZMnKWj3RbF4+wG2DJ2TYJf2AX9wMprzC7hZ/AVBKVkV8DOe67cOKcARgikjN/MCVVzYgTItfO1AVXw7ntEfW9KslZXBsM2kemHQ2M3oOLc8ZNfgKJIGMmhn+zeHRbNimldEryQhz1RCSbbHVqTVPpiU6qIJEokDETPecLqc2ZBinfiJtvv4qcxX0vU7Xm18b79aaiCzd7NpXUhF8k4hbevCVnvM1ZJ8lZJgNSVJSLT1UzBI6/cZDCl8rQl56Nnst2eM/Oty/psxXI5oyBqiR84qIvIz4mC7wROeOCyooZMJyUPoxRUuhUepTKxy8hyWy10pmo5KyjI5HZDDnsODhTjpSHp6FmDgHS1G35xK3b1GpGBrsQlpxtiwjGS56TuWK8kVQKzjdPNQukqZCcJXZ6CCZ8RaPiFoTA5swfBoHkLGYiNmmf4+fnZBNFRqPW5MEvAknzLDpETUZki4vW5kyIwaSsV/NMkvOBelDmralDKuVdy/ch1XeJW1QNFlyy/6FwZl/jt09GzlJwbn8Uzo0PyAsIxXvTJzXXepCK4K/vLDljas0OX+LhOPHfJE5lls16/Z8fS0klGKZG76Zg6srBvikHe18mITsK4BDg/s+jQEtzcB393V1CHQkkZ6bQqTWpq5auJbHVCmJfijZnCbwmfXJLph0dPee6ZnZggeSM+6YtzeZt4JBd91Ve9xUKlpxti2AT7nbb53e/P+CcG+5XL0oy8GqkCDnLIywCEAmlIRLOWLInEhkTmzPTBZCQ3DM5jhnxE88njXkVp9aUkTbNO3cfvde79bQL4Cy8R3mdXg0pqZvtZvMxaPfJWSgWnErKF7eoaskgBfqX5AiTQnJGivupySu/249TMeWt1pTEicsXmYw3Rts4ciZ+29IBObOFODISSM548t0Dak2+ja2+HST7VsxzWkYCxX5jQiw1BI4+/pD3TKq+kkqBFFxypnMIgBfGR/zOBMk8kVPpaDo2aRDaGLBvJZOcZTtynqpxkDkETNzPvA1AECMzu9aSM4tuBinuB+eSa+HMvjav+wPn0cpqEJmhuQLOT++Ec+UC74fKISApOuMQENyTwFvTtHxCwnZ0psnh+UXg84+BTBHIgYebVTl+3xi1Zh42Z6vf88odMgxElpB7yHBAlwIJkC/KPmkhqTzUmo5Ecqbaoas8P1Xt4kGFNjmpqLE4e98qRxaOnJFMnNq8lzgEpD21JnVdP82a0NYhw0CYdCdO7Z3Nev/x/a4r1Jo6QspvCNniztrPvplskyKG4jEhlibhewA1OXM6KTkbMjxyiOjImetLOFkYj9ydySRnLM6ZziEgtgw/1aBsHmhvS0bOuG9KKqrDzji6Oct3ksuu71lyZtM3baNQ59o0QBJRNV9n1SCgapD3QxJKIy90xg08WDj8ySBlIDkzmWso9W2iZJIznQowtwiQCfvAOXWmQYWAc9VCkB1HgqoIVz5qzaBweZtTP73T6HbiOP5r414eezd5qzVTYYKfT9DJMgnhDEEMQpsGmhU2ZyqSwEvOmKTGIYDY3VlVnQmlUQiHgEzGc+xxs3LJmcM5YsRJtrId0SDDXeEQIHuXPDKZIKOS+9TD3h+B5Mxvm8wWTUzFZKICT6XNAvooJWdOp8gZ2WcK6JP/L3ywSKPWZOm9ZG1K4uUYSM4EtaZuXJaV54LRxpHb9nYz2zGZQwAhoXFChm4P+ulHwIZaaRGksgbY/1CkhuapWSoQrOTMwgwhh4BkMcWcC+aCnHhm+KC4K843VQtzA88jzFnwTEy6IpNuiYulKTF1OMkZIbmyeTsyGRnhJDDk5PPM6gJyk2sSb01jSBb9JAstq5sjKs5Zc7w/8nEIYN6a7Z0jZ6RfCZxbH1ZfQKkgOXOiak2dhylvJ1Ptb0x2HaOuszNqzTgpoQnSaa8fsYTbUnLmeFI1IH7MZrNeLEOufSTps5ls2HR9hldrrnzRuyWQnKXV9YjEpFAOAYCanBEn3rZTB1kbdZKzts2+hFNybyJy5qfx4jcpHe36PKV8mr04NXl7G6iJ5Myfn2mEnHFl68ZJKg3n7Dko3iuhOrTAsOTMIjmSBq4dOxnOkSeGj0UkZ9xvwYMxFpHdVoIFSohIT2RSEJHUGGdUECVn/kRXUQXn0p/IywbCkrMkBMi3VSFdQc6EBdD5xb1wbvyN+f3smfxyyOSDQYb6Kph0HjZJqZQ32fI7aZUnXll8nlgSJ6no6BAkP+nIpkJLOHi15rARcK69Bc7xp+XODxvh/Rt4ayZ0CChgKA1CSE6a2NTo9eFIv6FGkjP6wv8Bq9/vnLemCQSJJdnvEKCqhqtT0qcG+onD2beTkUzxVRYqQwCgJGCEkM6pNWXjR0fOGuq9PiqLB5iEnKUzEpuzdn0+S056TSqqlZfR9jbPsH+3sXCuu11dnswhgBBB/a+ZZwoWWLpzsOTMIjGMUiPxkNnj+BM7OeNC7ze3iKYuvd5LZVMzRF92oNbMI4tCEPTU/y1TPUbImCE7cxyAJSkmqRyJJA4wfEf/GskkkG/eTN2OuzP5OP2dpjPvl3AW/AqkapA05h7gOYmEfv/o5iAIL9lxVzgX/QjkzItzFyhszsjJ56jbk0pFUykppCxk9Fg4F1ytLisOHe3hxdJJRaPHayZyKgTdJNvvHCakwUYnP4cAEid51BEKWYiTUv+7bmzIeR3zoDQgwuTgI5RF06ceAerXh/tdoXLC8hBJ+dDhcOZz6nbZ92F2soHkTELORMl+a0v0GgDkrFyMSXLAobrWeohTNRp49Cohm2czBv1pzedytbzMZERF9lgQWp6ddbR5/8WBl5yxkDUyfPQeULcOpLwCZPgI9XWBQwAnSSdE6IckXsDQFZuIPGDJmYUeO4/O/b3LbsDE/RPdTmSTsr9bJeMmgcy4BM7VN0puNCBabHeXT4qrYKIR4pzxE7o4iE2lhqI6LJ3xvLH49CSyhSPfXZuWnHVGcuar7nYeDTJoaPy1A8Nx88iOI4HdxoJ852yQ0y4AGb+P593IoLA5I7uPU9eRSkskZ4JaM+izFGSvZP01BDGIqihlkRGc47jo7dL8gZJvnLfNWUwoDV1Z3HdI/foJ7w+2SWpq9ENpCGVSBO+a7DERzhU/i68jFOOvUOnnOIikPJXSegAHEho21mRSV5GwffqRtHqyz5Tc37uOgbPgV9omB7Eld/1a9FwqBeeOP2jLkEIjiQeQk9SKkBE7GUlRSLdIKu31i6bG3EEDyRkpy5EzIvPSZGio8zYMMsckHkGcM/77EaHv6dSaVnJm0QfgLPkjnB8szP3+xglq1VkSsEUuUwznwMNABm8XvUaYWJ0lf4xe45Mzwow3/fQnRmDkTMytGecBRimcJX8E+cYJ8WXzJI7ZnAnqPdkkkG9ye218rFQKqB6cV9lJCK+MiBNC4BwxXR7/LqTWFDwjVXBS3nfnpVIiOSuKSR6dBP00Ee5lx/jnkCVr3mFngKl1/XaSr/lkVFgQnbk3xLePD7eSVHLWzydn/DMwNROlcskZM/QHvO9QWYM4hFRVnf0WJkilw30wVnLmn5NJXUUSvtOu8vrEfmoidakaBOfWh+EcrUgEnykyU2+Kki0n5UmleWcv4fmdHy/25lH2nX3tRHTcEvnGQhlP0Htu+sL/5Y61t+ltzjRmB5Hqx2oc2ZiEM2RzFvzPh6YfWsmZRV8AyRT5qpMC73qZYX2cilTYaROZDQRzv95zLzjX3WYcdgIASLFCchbrrUnl7RDL5t3aO9q9CZeVy4hEZ+NSmWL4jnC+dQaca26Bc9VC/fUiurKdaYUtSNw3CNSanORMXGDzyacnARHVmpG2yCZynpxFA2cSx4Fzip+mq7gfnPl3gnzPV/UKakoSCXEggCdn4jNr02r55IwnoLxkwklFF+OWZjgnnOmp5nfdHahU2wkBCAdBHbWHPIZVHMRn0Nmjit9I/N2vf64/x6k1uWPOgl/JcwlDYuNpYvKRcrx+xeXcLd7n4FyZhOQC5cZBVDE6KTgHfR1k0kGhY6H2ptMgmSI4P70bzi/u83LGeg8SLV+WwigJwWaOJXEYoPOYzoEceizIuMnxFzVu8P4NkTMnvHy5VKPWtJIzC4t4I1qTVEm8Kmz4jskkT2xyE+KcxUsG/UEt223v5iWmJ/tOBTnxe8B2O3jHW/2o4IHkzL8+QYw4Ec61t8D56V36CwGkrrsNZMRIkJJST/Uy5/qcUbRRZfp36lx/O5zrYwx1lY1TeBvGSc5SKRCWA5JBUE0FKhLJJEwOO84gjIYPjpxJHTRkx/g+qIqVFNxHQYYOz9lxSu3RYsBLI8VUbLrxw9rAS6359yKTnDU3guw40utT/Ur0GxXuPOlfAmfBr/XtUdwPQE84xTLE32XlubHLyspKPH35OFmDhpp7UZpIXVjfHr6T9/P7P0DF1b8IXUIkMcsiEMkZe1YxNp8EpLQMpKomN//K1JpJ0hepvNh14S84taYWO4/SN+PNlcjOOS0cJiOyJlCgZaO6ECs5s9iWQQ44zPu3ENH4Tes8TYgZxia3Tb4o3ETE7qoXTLLdDkj9+gmQkbuDOCmQKb7BdGuLt6P2J0BSXgnyvYtyaYHEcs67As78JfHPsv3OIEMMdteye/eYGJ99QFR/GnwHMmwEiMqeJQ68dyJPBHRpl0QJhajWLOXUcwKc756H1M2/M2ufLnG3bPHj35dKcsCeVexPEm9NZ+ZVIFOOlJfDkzORcOrCvgzazsv+cOEPc81KZ8KOMSIxlxnGx6jgyHhB0qGQLJFvnorUXY9FT3SWnInzS3E/oMbr34G9GP+eVNHkTaXHsvbtspu0jaS0zJsveEkXw44jDeqKqjW9f/lxlPvbufR6SRmK+IsEwCaJ5Ew1F+SpsiYJ1Jom2RPo838DWpqAunXcjUJaO0ojauuQxsVKziz6FIIxWRi7EXLmxXIbMh6FsG3jixNztvnkjG6o99pUFW8/48F7fnLsydFT4qDu7y+crS3eosct5M7BRyhdx519puRs6DoB5+If57xhIyfV7zZijN9das2kNmc8VOSss/1Va3MmWYzZAhangmHPKhIoiSSZ7H1gQCgi53jJlPjONAsmyWTgTDsKRJSiBraYErWmBM6iB+HMnJsrd8o3gnZHgl2rJG2qeoT3QUylgUHjxBRtGZCqQXBu/0Mkh6Oz4FdwVLEFTTeKAjkjR58EMlkgXyabna8fD2feoviLxGdjv1n5vJcicUBkKuG4tHIycqYAjdsIjIqJ62cqwQY6EWaECFY5knGx40hg9J7e39sSOVuyZAnOPfdcXH755cGx5uZmzJ8/H7Nnz8b8+fPR3OwFmKOU4r777sPFF1+MK664Ah99JPeSsehmsAFfIJte4jh6lUiBJWcAQM6+NPeDLUINdd6/GuNmALmIHaVlIN+7KHxOGNSBZ2NFpTex6GIO5QFy9HdAvnmq/Ny4yXD8hTICGTkbviOcC64GOe2C8ETYBU52AVSedap0SIBCciaomZlEiX2vU84HOejrydvHq7NkC5hUkuPHc4sLBcOeVSRQKsnQJx96/+q8ZXmwBXOCQhqkqovZickSn0skHSSTCRunM8Iq8azjzQ7IiWfmJFWqsS7OEdr0Z8Izid+MJdcuLo6aQJSUAv1VqZXMBkGEIBcXIxJXS2eHBWZ3JnGS4qEiooEELRVSn0vB+oAkiLNz4bywp37M5jVuLieTDgqFHMmd4GKpmWzENeRMKoEE5B7HIlKp3HhJ6jHdRegWcjZt2jTMmzcvdOzxxx/H2LFjceutt2Ls2LF4/PHHAQCvvvoq1qxZg1tvvRXnn38+7rlHnXDZoneDTDlSmu/NGF0QF8nZ/5DcBM7I0gZGzmSSLHEkc78jnloCORu9B5zZ14IcdyrI0SfBmXlV3u1WwfnWGXCO+24eN0bfrTPnepC99vcWW5ZmC+hayVnI5owP/qv+9iQVlZyRMROEctkE630v59Bj4fDx1SRwLvpx9CCvNpRJyWS2e2LcPBmCzY7QvxSLFDnIU7uQYyQSWw2ISjqhal9gixm2OXN+uAjONbfI7+EN/9niplvkSkpBmHe1yqsvadw3nc1ZXF9OZwIyTr7xLbHkRO0IwLxeebQaSqR0zy4G4GXqUkZGUk5uHKmkqGyTI25uCAHZc284l1wXHHJ+eFP4GpYd49iT4x0YMkUg20lMHgYMjNj7xkKj1iTnXKY6I6g13YjUFOkMFwh6G5KcjRkzBmVlYRuXFStWYOrUqQCAqVOnYsWKFQCAV155BVOmTAEhBKNHj0ZLSws2bNjQHc3sU3CuXJCf5103wjljlnH+xW4F88RiYQz2m+b9LpcstGxO2/sA719e3cgCkvpBZcnXBIIAgIzdGySTAamqAdlR4Y7fE5BJKvgdphjEsavA282IseFUcFKhydyZfS1QI0iUWFFJbGF22yNcxHlXhEOUyNoki7sUPEdM3YHNWVhiQQgBOfX70cv33NuzT2J9lYNz6U/kITeYPVskxIa/+KgW/yJOmsFLunYaBaLyzuQXTpGcKWNrpQEW9LZFkZpHDGmh2yiovDUDqVJMX05nQJyU954POTZZvXEQ3/8mgzREgJ4kRNSagkOAk9JLpFj/jkjOWBmcw5XgcEK+dzHIaTM9yb0iSC8AbyMlm27OmJUr34Sc6SRnqv5MCAgfU44CzmkzvQDZ7JJ+/XMx2Yo7kUKrgOgxm7PGxkZUVnqLYUVFBRobveB19fX1qKnJiU+rq6tRX1/fI23szSCj9wSJy9G3NaDANmcRMHJ24plw7vxTbCoe5+BvwFnyp3AQVj+UAxkzwTsXF+G6t4F5YvE7Xn4BCqU/6brvEHrnhpIzpNJhtabjRG23ArVhksYIkk8xjpRE9SP1DmbHXBoQFEc0do8hj2TfqTFNTEVsNcmYCfKQG9v5GwleNQVwzhKKSnwpGJF5a6oQcpxgHoDe+3SuvSX6/IAXToJJ9Vqa5OWKEjVFZopc3cIYZv0kw0kDFQhJ3SI5RTuxQRHnMVNbLvFZdFLBlEBE02l9CjglOfMhjkPuNZCycjjTjvbGQAzhJEVF8vc+bnLuuEnoG87EQOxPLCUemX66pAEEZMdduXN+x+e/S3H/QN0cm9KtG9ErlKuEkLyCby5duhRLly4FACxcuDBE6roC6XS6y+vorWgoKsYWAAMGlKFfN72D9gvnoum+W9D+9usAgJqaGqwVrqmpqdF+l7XctaHf2w1TErKgHr9bDqyqQtF2YRuQ5rSDFgAlFVUo205jH9KNEJ9XhlpQZAEMOPF0NN3pSV2qq2vglHuu7Q0jd8OWdV4YiMqqKqQTfHP+G5mMl6C9gwYH8wBt24J1iuvLKyuQbalEU/C7Eu3r+qMFANJplJ7wPTgDytAEoF+/YpRr2hDUP3hQqM7ymsEo5u7dWJSBGFyAfz5WTmlpGZoB9O/fD6UP/A/Q3g5HCJvS3liHegDplINqoX1ua3+sl5TPY+PR30Z6+I4oEc6zNqR3/Roqrvo5aNsWpLbbHut+mwtzkiorR3ZjA4odgoGS8jeUlqENQHH//igflHsncd/SLcoEbS4pH+CNi5JSlEnuYW0sHzoM6OhAA4BM22ZUSsa3k+0Aow3Vd/4BW158FnFyp/KqShRz5RSXlmILPBszumUTMpkMqhTvjH++rENRyx2n7W3S/sjPSeL8AgAlJf2RKi/HRgCZ0XsA6TTKz7ggMp5Uc1jofTipUIiPTFER+CRT5RUVKK6pwZaBFWgAkK4ejNKKSrCY/bLyN1VWYSOAonQa/S7/CVqf/jPa334NhBD/udtD3782lQJrQXllZTA+Ok77Pho+fAfZNV9E6iivrkF6hxGoFY7XDBqMbLYddQCc4mLpWAIAUjoA1Tc9gJRgg7fWcQDXRb9pR2LgFE/t7377e1j/+EOh60pKvX7YUlHhjcviYgyoqUFH60b4Bi0YOGQoGtu3wIXXL4sN1pWuRo+Rs4EDB2LDhg2orKzEhg0bUF7u7YiqqqpQW5v7jHV1daiqkseDOvzww3H44TkXWP6+rkBNTU2X19FbkfVFvk1NTWjurncwsAa4dD5w3jcBCN93t7EgZeWora3Vfhdy+PGgb74SuaauoUHfhvJKoH49Gre0gQj3ux3eNNVaVIzNvbBfxL0T9j2bt+Sm97oN9SBt3m966gXAS8sAABsaG0CK89tNJhkvdXV1wd9UkSsTADZu3AjKpYXZ2NQM6qvFyBEnYPPXp8N97q8AgM2bNqHNsA21gvnExk2bQ9/cZaq3vQ+Ac+x3gf4loedz5t8JOARMLrJ5/H7YsslXfQttoL6moKO9PfKOKBdfSvn+vvU9tAFoVZynV/0CGwCgqD/AvVcAyPpSki1NG6XlZ31pxpb2DtTW5+6N+5aUS4Le2uq1v3XTpthxsbG9I1C9tm2ol5bvcsbzDeliuJuisbec626He+fPgbVfYGNjo/fNxu8DuC7a/DFK/U1YezarfA7+OG1sCB2nsuTo7J6dRoEM31FabmtLK1Dk9Yj2on5IXXItGrwbQ9cZrS2CFKtdiM+2sbkFpLYW1O/HHeWVaGrJqRul79fPYLFl0yZ07D4BdOgOwJVngYJ6z83ZotXW1iLLjcuNzc258ZHpB/qDhcBlZ0Tq2LhpM9AWfX91dXWgtR6ldx1H+fzUdbHByUTeGZNubtnSFtwb+k5jJwFvvIJNX5uAzbW1cFu877Bp0yZsqa0FbczFOmvc0gbXl2hubO8AMVhXCoFhw9S2ej1GziZNmoTly5dj+vTpWL58OSZPnhwcf+aZZ3DggQfi/fffR0lJSaD+tLBw5lwP9Ouvj5zO33PyOQCXSJvsfwjoS/8wu3fGbNDatYDEXowcdpxn7HqwwiOyN8OfZEkmk9Nu8fZF/fr7yYyzeas1K39yG2JCPcZDayvD52x0cio6X/1EdtkNFAAZv495neJzit61/mJIxk3OGbLztw/NOb8EuSpVYM8ns4nr6swRPjmjigChpKif9zoTqDW16cNk6F8S2JzxOSpDENWaEg0LGT4CGLYDsPaLwG4oddGPAADu/b4DA+svkvud+XcCaz4TDor2XGrNTuqHcSEvaK7OPLRD5PDjQZf+xftRUeU9I4OofvPbTP0ArKR6kNapiqT98c/MAiLxzuK8pjVqV4ZMUdhhhAdLMB/Xf1TOIsQBkA2F4yDptPcMrguy48hwLMkgfA0z/OfVmv1yXrTbklpz8eLFePvtt9HU1ISZM2fipJNOwvTp03HzzTfj2WefxaBBg3DppV6Ig4kTJ2LVqlWYPXs2ioqKMGvWrO5oooUpuiM/XgzIHpr0LSZlnDUHZMZss4sHlMNR2JKRdAbk0GOl53o92I44wxEQcSIOyFl+9jZFY/eO7nYlcH58M+jbr4WOac0cMiI5YymxfHK2/U5w7npMb3MTV6cYroPZ4RUigrjCIQBA5+ybTMDelWrRY2mFHCe/XK9sjtDNFf1LQUrLPPshFRmPkaDyIEXFoABo2+aw7Tn7/kVqckaGDs/lOg0OCtfJSEpcqBQGimj8sQQgJ53tZRtxHKBuPeiyp0H/5ttbiXZarO9XD/Y2JnvsFR+SBojmFxXtLuPaLJYtCaDM2qksxydWZOwkdT2qeIFMWiuGQEqnPZIeGadCG1gcSsCzZ2NjsX/CFGNdhG4hZ3PmzJEev+aaaIR0QgjOPffcrm6SRUKQ7bb3dlhiipjuqPuw45IlNNeVR0hkElKiC8J59AowosETEGni7LbEiwo5+Vy1953s+hEjQUYYRETn76mqyUn8nFSOCHCLaBJiJoWw+DEVT6fLBYLduUwC16XesUBuEWpTkLOiHDnrUrB3IHufI3cHPnwnetxX+ZJ9p4L+a3nu+PY7Af9aDlIqOAywsjNqciaFGKZCuM+ZfY2XK1QH2knJGSE5qdKgocC3Z+TImUjs/d9kwr5wfnY3yODtQN/5T3wFkXy/Qc25NuwzBWSfqZHjqhhyAOD8eDHc+f66LzrWcCBVg+AsvEeZo9W54X5t5pZI8HA2t4v2xGL4Gj7EDB/PcFuSnFn0fZDjTgHZbSwIi6LcjXC+q4ja3ZXI+Lkbe0metYKD7ZR5AiKTnAGJFxXn8G92omGG4PPsVVaDTD4Y9KlHQCYdmLgo8p2zQZ/5U/SEqGph76wAhJ1U1cD5wUJARkq7Wq1ZMwRoqINz9Lfl57uJnMWpQp0Lfwj3sjO8NGOtHNFnIRtGjATWfQXip0YiR3zLI/livDtHWKhN362uz1cN9lT/MozfbtnBYQAAEvxJREFUB3j93/4PGg1x0QmESKLoocznA2XG84yIqAJgB+RMLaF0zrsi96OyGljzuV+JIkMBEN7EazwxiZgujj+nCt3CQ8y0wnmrhgsLZ+UIBUTm+6KJ52g3wKZvsjACSaX6VqiIziJJ5Oq+iCDWG0/OhGsCYtrFkhwNnEuujRwjTgrOFT+Dc/WNIBVVIMNGePGpWLL5JOUfMR2pmx6MnoioNf0FrECEnYwaA1IsWTS7WHJGDvo6Unc9plYlMdLR3i4/r4Bz1UIv5VAnJEW5wlJwLrgazrxfho+zMBT9S5Ca98tg40YcJ0rM/HJCMG1TjGqZnHWJZ+emQOqiH4GwwNAUuaCwhf6ugbQrhmBVe8GkiSoAM3s/7F4/xpfKXMM5/8rcD1Hzyz8fT3D8v50f3QznigXydgiIBImNg0juVLHTNFluWAibvFT5XYCtVCxgYdFJFBV7O/atlpz5k3HaQHKmS6DdRSCnfh9k59EgO43yJuC6cDADstvYrqv7iOnRpMyBzVkX94kuXBwG//mfIa9YKXxJBN1o4M3MgcVdpG+/Gnudc+UC0LVfagojIHvtDyqo7ugmT3JGSspkd0URsakyVWuqpaPOAYeZFCA5VJh+Q06fBTJ8BNynHvEOZHwbK4l9Hhk8DM6df1bHcAzImS9NymSQ+vUToAp7QVJW7kktP/0wvpH8xsbfABKTZO6snlO/D+e0mWYXi1k62DdX2pxxzzZ4GLDO64vOuZeDKrMMdD8sObOwkIFJlBQu9H0egbpDEaEfyE1uhkbZhYZzyDG5H0yaU1ENdEPwZXL0d6IHCyw5U9bdheTMpOzAnq+xa4J/k9F76s0jmNetKoCrqV1QvuSsYE4ZFDTI1FCYEp2pR3r/fvc8uI/eBzJsBOhf/wz4MQpFxAXXjticsXtM3lPcNbzUKp1cTZhkDESuZSrciOTM/9fNkTPnmsUA57XcW6RmgCVnFhZSON85C+69N0ftGbY2hCLtCxNTzRBg/ZreIT0cMhz44hM4l1wD0gUJ5COQhVw4+jtwP/yv3E6sl8BZ8KvOF8K83xrzTJtn6q0ZB4WUyTnqRLgfvgPsZJgKjUmGmGefseSsk4t0yPjcew+kwOYBZOj2SM2+BjSbBRk7Kb+MMamw5MwM+u9KHMez+XLdeHLYGXxtPMjOu0WPs7AdkXqjkjNS3E8d5qOHYcmZhYUEZMJ+SN32SE83o+sRE0rD+f4PQN9cBWISMqCL4cyYDTp+H5Dtd+6eCiXkgIyZgNSdf+6e+vNEKL1YvmBG2NyCTfY/1LwNe0wE/cvv4sMjaAuRExkyZiJSQvqqWASSM18CbuwQ0MkNCdv0pNM5ktpFmxySSgG75emo5eekJQcerrlQWnP86UxRQtKXDKnL5stPsPRjooRbjHPWy2HJmYXFVgbn0uvN8/dxkjNRpE9KB8TmeexOkP4lIAeYE4TOV9gL1BudITedACkqBjn5HJDdPQcgbUBd8f6dRye+J4JCERlGzphzg+ln7eT3J4ceC7Q0g3z9W6CvvdypsroSZEB58m9lSm7SmZ4hQgrJGRm+kycz6yM5qS05s7DYykDGJAjU20vcxnsdepicOTc9CPTruXhLzuHH91jdAEKSK3LosSB77pVfOU5YckZMMx50khySomKQE88EgJxxfVeHSOkmOOdcBvrUI8DwHeMvFL2duwvM7EGwFyYjd4dzw/1m4Tl6AbaO3mJhYZEXCCEgZ8zyAlxaFCYMRCGaMWAgSE8tbr0B3Pt3Tjk/fxWp71lLRu0BFBWBHDE92f1igNN8QAvrENDTIMN3hPP9H+htydJFeTkDdBbEDwdC2zZHz/URYgZYyZmFxTYJMvVI0OXPAACcKUcCU47s4Rb1FhD4wal6uiHbNgpFjv2FGpXVSN2RwFYNgHPxj4Eddul8G7YyyVkcnLk35NJ/ZTJyx5rrb0+UQSQxmNdqF9q7dQcsObOw2AbhnD4LON3mrY3AIUAWW42Uo6+iUCENyP6HAnXrQI48Ifm9spyOo/KwV9rKJGdxICN3z/3IFEnJGRmmDuBbkDYcfypQ1A9knyldWk9Xw5IzCwsLC4bRewL/fR3bxEraC+FcuQD0lX8WrDySToNMP70gZTmLf69OgxSLbUdyFkJpWY88M+lXAvKtwnzznoQlZxYWFhY+nFnzgNo1XRebqYdAjjoR9PNPeroZWhgFqO0hkFLDrAQiIknFtw0437uop5vQp7F1zUAWFhYWnQDp1x/orlhq3QjnhDN7ugkW25jkjFQN6ukm9GlsW73FwsLCwsKiGxEETt59XM82xKJPwUrOLCwsLCwsughk51FwbnoIZEB5TzfFog/BSs4sLCwsLCy6EJaYWSSFJWcWFhYWFhYWFr0IlpxZWFhYWFhYWPQiWHJmYWFhYWFhYdGLYMmZhYWFhYWFhUUvgiVnFhYWFhYWFha9CJacWVhYWFhYWFj0IlhyZmFhYWFhYWHRi2DJmYWFhYWFhYVFL4IlZxYWFhYWFhYWvQiWnFlYWFhYWFhY9CIQSint6UZYWFhYWFhYWFh4sJKzBJg7d25PN8FCAvtdeifsd+l9sN+kd8J+l96Hnv4mlpxZWFhYWFhYWPQiWHJmYWFhYWFhYdGLkLruuuuu6+lG9CXssssuPd0ECwnsd+mdsN+l98F+k94J+116H3rym1iHAAsLCwsLCwuLXgSr1rSwsLCwsLCw6EVI93QD+gpee+013H///XBdF4cddhimT5/e003aJlBbW4s77rgDDQ0NIITg8MMPx9FHH43m5mbcfPPNWL9+PQYNGoRLL70UZWVloJTi/vvvx6uvvori4mLMmjXLqgu6EK7rYu7cuaiqqsLcuXOxbt06LF68GE1NTdhll11w8cUXI51Oo729Hbfffjs++ugjDBgwAHPmzMHgwYN7uvlbHVpaWnDXXXfhs88+AyEEF1xwAYYNG2bHSg/jqaeewrPPPgtCCHbYYQfMmjULDQ0Ndqx0M5YsWYJVq1Zh4MCBWLRoEQDktZYsW7YMf/7znwEAJ5xwAqZNm1b4xlILLbLZLL3ooovomjVraHt7O73iiivoZ5991tPN2iZQX19PP/zwQ0oppa2trXT27Nn0s88+ow8++CB97LHHKKWUPvbYY/TBBx+klFK6cuVK+rOf/Yy6rkvfffddevXVV/dY27cFPPnkk3Tx4sX05z//OaWU0kWLFtEXXniBUkrp3XffTf/6179SSil95pln6N13300ppfSFF16gN910U880eCvHbbfdRpcuXUoppbS9vZ02NzfbsdLDqKuro7NmzaJbtmyhlHpj5B//+IcdKz2At956i3744Yf0sssuC44lHR9NTU30wgsvpE1NTaG/Cw2r1jTABx98gKFDh2LIkCFIp9M44IADsGLFip5u1jaBysrKYLfSv39/DB8+HPX19VixYgWmTp0KAJg6dWrwPV555RVMmTIFhBCMHj0aLS0t2LBhQ4+1f2tGXV0dVq1ahcMOOwwAQCnFW2+9hf322w8AMG3atNB3YbvL/fbbD2+++SaoNXctKFpbW/Hf//4Xhx56KAAgnU6jtLTUjpVeANd10dbWhmw2i7a2NlRUVNix0gMYM2YMysrKQseSjo/XXnsN48aNQ1lZGcrKyjBu3Di89tprBW+rVWsaoL6+HtXV1cHv6upqvP/++z3Yom0T69atw+rVq7HrrruisbERlZWVAICKigo0NjYC8L5VTU1NcE91dTXq6+uDay0KhwceeACnn346Nm3aBABoampCSUkJUqkUAKCqqgr19fUAwmMolUqhpKQETU1NKC8v75nGb4VYt24dysvLsWTJEnzyySfYZZddMGPGDDtWehhVVVU47rjjcMEFF6CoqAjjx4/HLrvsYsdKL0HS8SHyAf7bFRJWcmbRJ7B582YsWrQIM2bMQElJSegcIQSEkB5q2baJlStXYuDAgdZGqRchm81i9erVOOKII3DDDTeguLgYjz/+eOgaO1a6H83NzVixYgXuuOMO3H333di8eXOXSFosOo/eND6s5MwAVVVVqKurC37X1dWhqqqqB1u0baGjowOLFi3CwQcfjH333RcAMHDgQGzYsAGVlZXYsGFDsKusqqpCbW1tcK/9Vl2Dd999F6+88gpeffVVtLW1YdOmTXjggQfQ2tqKbDaLVCqF+vr64N2zMVRdXY1sNovW1lYMGDCgh59i60J1dTWqq6sxatQoAJ5K7PHHH7djpYfxxhtvYPDgwcF733ffffHuu+/asdJLkHR8VFVV4e233w6O19fXY8yYMQVvl5WcGWDkyJH46quvsG7dOnR0dODFF1/EpEmTerpZ2wQopbjrrrswfPhwHHvsscHxSZMmYfny5QCA5cuXY/LkycHx5557DpRSvPfeeygpKbFqmi7Aqaeeirvuugt33HEH5syZgz333BOzZ8/GHnvsgZdffhmA59HExsnee++NZcuWAQBefvll7LHHHr1mh7q1oKKiAtXV1fjyyy8BeKRg++23t2Olh1FTU4P3338fW7ZsAaU0+C52rPQOJB0fEyZMwOuvv47m5mY0Nzfj9ddfx4QJEwreLhuE1hCrVq3Cb37zG7iui0MOOQQnnHBCTzdpm8A777yDa665BiNGjAgmqFNOOQWjRo3CzTffjNra2oj787333ovXX38dRUVFmDVrFkaOHNnDT7F146233sKTTz6JuXPnYu3atVi8eDGam5ux88474+KLL0Ymk0FbWxtuv/12rF69GmVlZZgzZw6GDBnS003f6vDxxx/jrrvuQkdHBwYPHoxZs2aBUmrHSg/j0UcfxYsvvohUKoWddtoJM2fORH19vR0r3YzFixfj7bffRlNTEwYOHIiTTjoJkydPTjw+nn32WTz22GMAvFAahxxySMHbasmZhYWFhYWFhUUvglVrWlhYWFhYWFj0IlhyZmFhYWFhYWHRi2DJmYWFhYWFhYVFL4IlZxYWFhYWFhYWvQiWnFlYWFhYWFhY9CJYcmZhYbFNY8GCBUFcqULh0Ucfxa233lrQMi0sLLYd2AwBFhYWWwUuvPBCNDQ0wHFye85p06bhnHPOib1v3rx5Xd00CwsLi0Sw5MzCwmKrwVVXXYVx48b1dDMsLCwsOgVLziwsLLZqLFu2DH//+9+x00474bnnnkNlZSXOOeccjB07FgBw3XXX4eCDD8Zhhx2GNWvW4M4778THH3+MdDqNPffcE5deeikAL5/oAw88gC+//BLDhg3DjBkzsNtuuwEA1q1bhzvuuAOrV6/GqFGjMGzYsFAb3nvvPfz2t7/F559/jkGDBmHGjBnYY489uvdFWFhY9BlYmzMLC4utHu+//z6GDBmCe++9FyeddBJ++ctform5OXLdww8/jPHjx+P+++/HnXfeiaOOOgoA0NzcjIULF+Koo47Cfffdh2OOOQYLFy5EU1MTAOCWW27BLrvsgnvvvRcnnnhikKsP8BIjL1y4ECeccALuu+8+nHHGGVi0aBE2btzYPQ9vYWHR52DJmYWFxVaDG2+8ETNmzAj+W7p0KQBg4MCBOOaYY5BOp3HAAQdg2LBhWLVqVeT+dDqN9evXY8OGDSgqKsLuu+8OwMutO3ToUEyZMgWpVAoHHXQQhg0bhpUrV6K2thYffvghTj75ZGQyGYwZMwZ77713UOZzzz2HiRMnYq+99oLjOBg3bhxGjhwprd/CwsICsGpNCwuLrQhXXnllxOZs2bJlqKqqAiEkODZo0CDU19dH7j/99NPx8MMPY968eSgtLcWxxx6LQw89FPX19Rg0aFDoWlZGfX09SktL0a9fv9C52tpaAEBtbS1efvllrFy5MjifzWatWtPCwkIJS84sLCy2etTX14NSGhC02tpaTJo0KXJdRUUFZs6cCQB45513MH/+fIwZMwZVVVX417/+Fbq2trYWEyZMQGVlJVpaWrB58+aAoDFiBgDV1dU4+OCDg3ItLCwsdLBqTQsLi60ejY2NePrpp9HR0YGXXnoJX3zxBSZOnBi57qWXXkJdXR0AoLS0FABACMHEiRPx1Vdf4YUXXkA2m8WLL76Izz//HHvttRcGDRqEkSNH4tFHH0VHRwfeeeedkJTs4IMPxsqVK/Haa6/BdV20tbXhrbfeCuqxsLCwEEEopbSnG2FhYWHRWcjinI0bNw6TJ08OeWtWVFTg7LPPxvjx4wGEvTUfeughPP/882htbUVFRQWOP/54HH744QA8Sdr999+PNWvWYOjQoTjrrLMCm7S1a9cG3pqjR4/GsGHD0NLSgtmzZwPwHBIeeughfPrpp3AcB7vuuivOO+881NTUdPNbsrCw6Auw5MzCwmKrBgulMX/+/J5uioWFhYURrFrTwsLCwsLCwqIXwZIzCwsLCwsLC4teBKvWtLCwsLCwsLDoRbCSMwsLCwsLCwuLXgRLziwsLCwsLCwsehEsObOwsLCwsLCw6EWw5MzCwsLCwsLCohfBkjMLCwsLCwsLi14ES84sLCwsLCwsLHoR/j9AaxgSYuA/oAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "(<Figure size 720x360 with 1 Axes>,\n",
       " <Figure size 720x360 with 1 Axes>,\n",
       " <Figure size 720x360 with 1 Axes>)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plotting.plot_episode_stats(stats, smoothing_window=25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Stats(episode_lengths=array([499., 499., 499., 499., 499., 452., 499., 499., 499., 340., 475.,\n",
       "       468., 498., 499., 256., 499., 499., 372., 332., 340., 445., 324.,\n",
       "       182., 269., 249., 352., 175., 396., 499., 124., 499., 307., 211.,\n",
       "       354., 472., 360., 217., 327., 312., 328., 345., 499., 268., 186.,\n",
       "       163., 213., 228., 246., 240., 208., 297., 147., 277., 336., 158.,\n",
       "       343., 121., 236., 345., 278., 426., 242., 252., 307., 201., 187.,\n",
       "       141., 130., 232., 165., 171., 161., 243., 242., 229., 499., 181.,\n",
       "       190., 147., 203., 140., 329., 192., 170., 212., 278., 107., 338.,\n",
       "       347., 170., 280., 418., 179., 157., 288., 294., 396., 205., 152.,\n",
       "       267., 121., 112., 174., 266., 209., 263., 135., 233., 217., 211.,\n",
       "       208., 274., 215., 129., 229., 224., 178., 303., 282., 253., 189.,\n",
       "       130., 269., 204., 154., 274., 145., 178., 253., 114., 261., 158.,\n",
       "       171., 197., 159., 258., 177., 395., 168., 184., 293., 252., 124.,\n",
       "       176., 156., 254., 158., 262., 228., 147., 247., 206., 436., 236.,\n",
       "       189., 215., 183., 109., 162., 230., 178., 204., 191., 156., 308.,\n",
       "       236., 215., 121., 223., 148., 155., 195., 130., 200., 171., 157.,\n",
       "       129., 198., 127., 124., 220., 134., 148., 155., 155., 147., 182.,\n",
       "       176., 129., 214., 138., 228., 120., 247., 168., 124., 143., 132.,\n",
       "       198., 154.,  99., 104.,  87., 131., 180., 244., 236., 181., 121.,\n",
       "       179., 142., 226., 140., 168., 101., 135., 166., 192., 145., 139.,\n",
       "       204., 255., 148., 165., 140., 186., 148., 121., 136., 128., 117.,\n",
       "       129., 192., 108., 133., 143., 198., 124., 198., 118., 211., 184.,\n",
       "       171., 212., 155., 148., 219., 175., 112., 289., 206., 283., 275.,\n",
       "       167., 158., 240., 163., 193., 156., 139., 209., 161., 136., 149.,\n",
       "       137., 190., 125.,  91., 115., 112., 207., 158., 122., 166., 201.,\n",
       "       134., 201., 107., 133., 155., 154., 107., 162., 141., 125., 249.,\n",
       "       141., 109., 189., 170., 143.,  96., 198., 168., 116., 181., 263.,\n",
       "        95., 191., 105., 212., 128., 119., 128., 139., 113., 147., 163.,\n",
       "       195., 118., 308., 210., 163., 143., 175., 107., 195., 191., 126.,\n",
       "       113., 137., 178., 143., 130., 127., 101., 104., 206., 131., 155.,\n",
       "        96., 203., 200., 235., 184., 144., 147., 120., 130., 121., 268.,\n",
       "       133., 129.,  99.,  92., 121., 115., 110., 114., 174., 193., 129.,\n",
       "       217.,  95., 291., 142., 143., 101., 214.,  88., 291., 110., 157.,\n",
       "       250., 111., 123., 160., 158., 171., 177., 128., 167., 150., 177.,\n",
       "       100.,  79., 127., 163.,  96., 171., 157.,  92., 204., 135., 151.,\n",
       "       110.,  75., 155., 107., 105., 151., 113., 160., 390., 149., 139.,\n",
       "       107., 128., 130., 201., 110., 163., 100., 121., 232., 134.,  95.,\n",
       "       121., 130., 112.,  97., 132., 162., 130., 207., 151., 139., 131.,\n",
       "       154., 259., 126., 103., 110., 116., 115., 176., 128., 145., 117.,\n",
       "       123., 159., 155., 152., 237., 117., 261., 165., 136., 142., 124.,\n",
       "       123., 107., 221., 110., 107.,  99., 135., 154.,  86., 172.,  85.,\n",
       "       122., 168., 163.,  98., 103., 151., 119., 107., 141., 117., 117.,\n",
       "       173., 111., 208., 126., 134., 127.,  85.,  91., 120., 122., 165.,\n",
       "       122., 217., 112., 101., 137., 158., 133., 155., 225., 113., 130.,\n",
       "       152., 160., 159., 125., 133., 234., 118., 252., 324.,  88., 114.,\n",
       "       164., 144., 181., 105.,  90.]), episode_rewards=array([-500., -500., -500., -500., -500., -452., -500., -500., -500.,\n",
       "       -340., -475., -468., -498., -500., -256., -500., -500., -372.,\n",
       "       -332., -340., -445., -324., -182., -269., -249., -352., -175.,\n",
       "       -396., -500., -124., -500., -307., -211., -354., -472., -360.,\n",
       "       -217., -327., -312., -328., -345., -500., -268., -186., -163.,\n",
       "       -213., -228., -246., -240., -208., -297., -147., -277., -336.,\n",
       "       -158., -343., -121., -236., -345., -278., -426., -242., -252.,\n",
       "       -307., -201., -187., -141., -130., -232., -165., -171., -161.,\n",
       "       -243., -242., -229., -500., -181., -190., -147., -203., -140.,\n",
       "       -329., -192., -170., -212., -278., -107., -338., -347., -170.,\n",
       "       -280., -418., -179., -157., -288., -294., -396., -205., -152.,\n",
       "       -267., -121., -112., -174., -266., -209., -263., -135., -233.,\n",
       "       -217., -211., -208., -274., -215., -129., -229., -224., -178.,\n",
       "       -303., -282., -253., -189., -130., -269., -204., -154., -274.,\n",
       "       -145., -178., -253., -114., -261., -158., -171., -197., -159.,\n",
       "       -258., -177., -395., -168., -184., -293., -252., -124., -176.,\n",
       "       -156., -254., -158., -262., -228., -147., -247., -206., -436.,\n",
       "       -236., -189., -215., -183., -109., -162., -230., -178., -204.,\n",
       "       -191., -156., -308., -236., -215., -121., -223., -148., -155.,\n",
       "       -195., -130., -200., -171., -157., -129., -198., -127., -124.,\n",
       "       -220., -134., -148., -155., -155., -147., -182., -176., -129.,\n",
       "       -214., -138., -228., -120., -247., -168., -124., -143., -132.,\n",
       "       -198., -154.,  -99., -104.,  -87., -131., -180., -244., -236.,\n",
       "       -181., -121., -179., -142., -226., -140., -168., -101., -135.,\n",
       "       -166., -192., -145., -139., -204., -255., -148., -165., -140.,\n",
       "       -186., -148., -121., -136., -128., -117., -129., -192., -108.,\n",
       "       -133., -143., -198., -124., -198., -118., -211., -184., -171.,\n",
       "       -212., -155., -148., -219., -175., -112., -289., -206., -283.,\n",
       "       -275., -167., -158., -240., -163., -193., -156., -139., -209.,\n",
       "       -161., -136., -149., -137., -190., -125.,  -91., -115., -112.,\n",
       "       -207., -158., -122., -166., -201., -134., -201., -107., -133.,\n",
       "       -155., -154., -107., -162., -141., -125., -249., -141., -109.,\n",
       "       -189., -170., -143.,  -96., -198., -168., -116., -181., -263.,\n",
       "        -95., -191., -105., -212., -128., -119., -128., -139., -113.,\n",
       "       -147., -163., -195., -118., -308., -210., -163., -143., -175.,\n",
       "       -107., -195., -191., -126., -113., -137., -178., -143., -130.,\n",
       "       -127., -101., -104., -206., -131., -155.,  -96., -203., -200.,\n",
       "       -235., -184., -144., -147., -120., -130., -121., -268., -133.,\n",
       "       -129.,  -99.,  -92., -121., -115., -110., -114., -174., -193.,\n",
       "       -129., -217.,  -95., -291., -142., -143., -101., -214.,  -88.,\n",
       "       -291., -110., -157., -250., -111., -123., -160., -158., -171.,\n",
       "       -177., -128., -167., -150., -177., -100.,  -79., -127., -163.,\n",
       "        -96., -171., -157.,  -92., -204., -135., -151., -110.,  -75.,\n",
       "       -155., -107., -105., -151., -113., -160., -390., -149., -139.,\n",
       "       -107., -128., -130., -201., -110., -163., -100., -121., -232.,\n",
       "       -134.,  -95., -121., -130., -112.,  -97., -132., -162., -130.,\n",
       "       -207., -151., -139., -131., -154., -259., -126., -103., -110.,\n",
       "       -116., -115., -176., -128., -145., -117., -123., -159., -155.,\n",
       "       -152., -237., -117., -261., -165., -136., -142., -124., -123.,\n",
       "       -107., -221., -110., -107.,  -99., -135., -154.,  -86., -172.,\n",
       "        -85., -122., -168., -163.,  -98., -103., -151., -119., -107.,\n",
       "       -141., -117., -117., -173., -111., -208., -126., -134., -127.,\n",
       "        -85.,  -91., -120., -122., -165., -122., -217., -112., -101.,\n",
       "       -137., -158., -133., -155., -225., -113., -130., -152., -160.,\n",
       "       -159., -125., -133., -234., -118., -252., -324.,  -88., -114.,\n",
       "       -164., -144., -181., -105.,  -90.]))"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "# state = env.observation_space.sample()\n",
    "state = env.reset()\n",
    "# plt.figure()\n",
    "# plt.imshow(env.render(mode='rgb_array'))\n",
    "while True:\n",
    "    q_values = estimator.predict(state)\n",
    "    best_action = np.argmax(q_values)\n",
    "#     plt.figure()\n",
    "#     plt.imshow(env.render(mode='rgb_array'))\n",
    "\n",
    "#     action_pge = get_pge_action(best_action)\n",
    "    \n",
    "    next_state, reward, end, _ = env.step(best_action)\n",
    "    if end:\n",
    "        break\n",
    "        \n",
    "    state = next_state\n",
    "    env.render()\n",
    "# env.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-0.11945403,  0.29413375, -1.4568335 , -0.17493604,  0.40899193,\n",
       "        0.73602676,  1.3300323 ,  0.1892223 ], dtype=float32)"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "env.observation_space.sample()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "env = gym.make('CartPole-v1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([ 0.00353907,  0.19214232, -0.0343606 , -0.34522076]), 1.0, False, {})"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "env.reset()\n",
    "env.step(env.action_space.sample())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([-0.74117499, -1.35688824,  0.28442935,  1.79704102]), 0.0, True, {})"
      ]
     },
     "execution_count": 123,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "env.step(env.action_space.sample())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([ 0.99191838,  0.12687758,  0.93816387, -0.34619151, -0.26319339,\n",
       "         0.28025021]),\n",
       " -1.0,\n",
       " False,\n",
       " {})"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "env.step(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "env.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}