main.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from math import floor, ceil\n",
    "import keras.models\n",
    "from keras.layers.convolutional import Conv1D, ZeroPadding1D\n",
    "from keras.layers.recurrent import Recurrent, LSTM, GRU\n",
    "from keras.layers.pooling import MaxPooling1D, AveragePooling1D\n",
    "from keras.utils import plot_model, np_utils\n",
    "import keras.backend as K\n",
    "from sklearn.preprocessing import LabelBinarizer\n",
    "from sklearn.metrics import accuracy_score\n",
    "% matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# preprocess"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train, test sizes (1, 884682, 2) (1, 9698, 2)\n"
     ]
    }
   ],
   "source": [
    "PAD_WITH_DERIV = True\n",
    "def pad_with_deriv(x_train):\n",
    "    x_deriv = np.zeros(x_train.shape)\n",
    "    x_deriv[:,:-1,:] = np.diff(x_train,axis=1)\n",
    "    x_train_2 = np.zeros((x_train.shape[0],x_train.shape[1],2))\n",
    "    x_train_2[:,:,0] = x_train[:,:,0]\n",
    "    x_train_2[:,:,1] = x_deriv[:,:,0]\n",
    "    return x_train_2\n",
    "\n",
    "def load_and_format_data():\n",
    "    # load data\n",
    "    x_train = np.array([])\n",
    "    y_train = np.array([])\n",
    "    for i in ['7']:\n",
    "        dataset = 'train/'+str(i)\n",
    "        x_data = pd.read_csv(dataset + '.train.calcium.csv')\n",
    "        y_data = pd.read_csv(dataset + '.train.spikes.csv')    \n",
    "        for key in x_data:\n",
    "            x_train = np.concatenate((x_train,x_data[key].dropna()))\n",
    "            y_train = np.concatenate((y_train,y_data[key].dropna()))\n",
    "    x_test = pd.read_csv('train/10' + '.train.calcium.csv')['9'].dropna()\n",
    "    y_test = pd.read_csv('train/10' + '.train.spikes.csv')['9'].dropna()\n",
    "    num_y = int(np.max(y_train+1))\n",
    "    # binarize labels\n",
    "    lb = LabelBinarizer()\n",
    "    lb.fit(y_train) # this lb takes 0 -> [1,0,0,0,0], 1-> [0,1,0,0,0], etc.\n",
    "    y_train_orig = y_train\n",
    "    y_train = lb.transform(y_train)\n",
    "#     y_test = lb.transform(y_test)\n",
    "    # reshape data\n",
    "    x_train = x_train.reshape((1, len(x_train), 1))\n",
    "    y_train = y_train.reshape((1, len(y_train), y_train.shape[1]))\n",
    "    x_test = x_test.values.reshape((1, len(x_test), 1))\n",
    "#     y_test = y_test.reshape((1, len(y_test), y_test.shape[1]))\n",
    "    if PAD_WITH_DERIV:\n",
    "        x_train = pad_with_deriv(x_train)\n",
    "        x_test = pad_with_deriv(x_test)\n",
    "    return x_train, y_train, x_test, y_test, y_train_orig, num_y\n",
    "x_train, y_train, x_test, y_test, y_train_orig, num_y = load_and_format_data()\n",
    "print(\"train, test sizes\",x_train.shape,x_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "code_folding": [
     4,
     28,
     43
    ],
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from tensorflow.contrib.metrics import streaming_covariance\n",
    "# gives Nonetype error which means gradient is 0\n",
    "def custom_loss(y_true, y_pred):\n",
    "    y_true_k = K.clip(y_true,0,1)\n",
    "    print(\"calling...\",y_true,y_true_k,y_pred)\n",
    "    return streaming_covariance(y_true,y_pred)\n",
    "#     loss = K.switch(K.equal(y_true,-1),0,K.square(y_true-y_pred))\n",
    "    return -1*K.sum(y_pred * y_true)\n",
    "\n",
    "def matthews_correlation(y_true, y_pred):\n",
    "    y_pred_pos = K.round(K.clip(y_pred, 0, 1))\n",
    "    y_pred_neg = 1 - y_pred_pos\n",
    "\n",
    "    y_pos = K.round(K.clip(y_true, 0, 1))\n",
    "    y_neg = 1 - y_pos\n",
    "\n",
    "    tp = K.sum(y_pos * y_pred_pos)\n",
    "    tn = K.sum(y_neg * y_pred_neg)\n",
    "\n",
    "    fp = K.sum(y_neg * y_pred_pos)\n",
    "    fn = K.sum(y_pos * y_pred_neg)\n",
    "\n",
    "    numerator = (tp * tn - fp * fn)\n",
    "    denominator = K.sqrt((tp + fp) * (tp + fn) * (tn + fp) * (tn + fn))\n",
    "\n",
    "    return numerator / (denominator + K.epsilon())\n",
    "\n",
    "def pos_acc(y_true, y_pred):\n",
    "    y_pred_pos = K.round(K.clip(y_pred, 0, 1))\n",
    "    y_pred_neg = 1 - y_pred_pos\n",
    "\n",
    "    y_pos = K.round(K.clip(y_true, 0, 1))\n",
    "    y_neg = 1 - y_pos\n",
    "\n",
    "    tp = K.sum(y_pos * y_pred_pos)\n",
    "    tn = K.sum(y_neg * y_pred_neg)\n",
    "\n",
    "    fp = K.sum(y_neg * y_pred_pos)\n",
    "    fn = K.sum(y_pos * y_pred_neg)\n",
    "\n",
    "    return tp/(tp+fn)\n",
    "\n",
    "def neg_acc(y_true, y_pred):\n",
    "    y_pred_pos = K.round(K.clip(y_pred, 0, 1))\n",
    "    y_pred_neg = 1 - y_pred_pos\n",
    "\n",
    "    y_pos = K.round(K.clip(y_true, 0, 1))\n",
    "    y_neg = 1 - y_pos\n",
    "\n",
    "    tp = K.sum(y_pos * y_pred_pos)\n",
    "    tn = K.sum(y_neg * y_pred_neg)\n",
    "\n",
    "    fp = K.sum(y_neg * y_pred_pos)\n",
    "    fn = K.sum(y_pos * y_pred_neg)\n",
    "\n",
    "    return tn/(tn+fp)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# create model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "sample_weight = np.ones((1,len(y_train_orig)))\n",
    "sample_weight[0,y_train_orig==0]=1/10000\n",
    "kernel_sizes = [5,13,31] # should be odd\n",
    "pad_sizes = [floor(s/2) for s in kernel_sizes]\n",
    "model = keras.models.Sequential()\n",
    "model.add(ZeroPadding1D(padding=pad_sizes[0], input_shape=(None, x_train.shape[-1])))\n",
    "model.add(Conv1D(filters=num_y, kernel_size=kernel_sizes[0], activation=\"linear\"))\n",
    "# model.add(MaxPooling1D(pool_size=2,strides=1,padding=\"same\"))\n",
    "model.add(ZeroPadding1D(padding=pad_sizes[1], input_shape=(None, x_train.shape[-1])))\n",
    "model.add(Conv1D(filters=num_y, kernel_size=kernel_sizes[1], activation=\"linear\"))\n",
    "model.add(ZeroPadding1D(padding=pad_sizes[2], input_shape=(None, x_train.shape[-1])))\n",
    "model.add(Conv1D(filters=num_y, kernel_size=kernel_sizes[2], activation=\"softmax\"))\n",
    "model.compile(loss='mse', optimizer='adam', metrics=['accuracy',pos_acc,neg_acc],\n",
    "              sample_weight=sample_weight, sample_weight_mode=\"temporal\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# fit the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "1/1 [==============================] - 1s - loss: 0.1980 - acc: 0.0051 - pos_acc: 2.0346e-05 - neg_acc: 0.9883\n",
      "Epoch 2/10\n",
      "1/1 [==============================] - 1s - loss: 0.1942 - acc: 0.0083 - pos_acc: 1.6955e-05 - neg_acc: 0.9902\n",
      "Epoch 3/10\n",
      "1/1 [==============================] - 1s - loss: 0.1894 - acc: 0.0188 - pos_acc: 1.4695e-05 - neg_acc: 0.9923\n",
      "Epoch 4/10\n",
      "1/1 [==============================] - 1s - loss: 0.1853 - acc: 0.0461 - pos_acc: 1.5825e-05 - neg_acc: 0.9939\n",
      "Epoch 5/10\n",
      "1/1 [==============================] - 1s - loss: 0.1812 - acc: 0.1245 - pos_acc: 1.1190e-04 - neg_acc: 0.9954\n",
      "Epoch 6/10\n",
      "1/1 [==============================] - 1s - loss: 0.1768 - acc: 0.3181 - pos_acc: 8.8732e-04 - neg_acc: 0.9968\n",
      "Epoch 7/10\n",
      "1/1 [==============================] - 1s - loss: 0.1724 - acc: 0.6312 - pos_acc: 0.0043 - neg_acc: 0.9986\n",
      "Epoch 8/10\n",
      "1/1 [==============================] - 1s - loss: 0.1681 - acc: 0.8586 - pos_acc: 0.0104 - neg_acc: 0.9996\n",
      "Epoch 9/10\n",
      "1/1 [==============================] - 1s - loss: 0.1638 - acc: 0.9445 - pos_acc: 0.0181 - neg_acc: 0.9996\n",
      "Epoch 10/10\n",
      "1/1 [==============================] - 1s - loss: 0.1592 - acc: 0.9712 - pos_acc: 0.0266 - neg_acc: 0.9995\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x11cdb9668>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "epochs = 10\n",
    "model.fit(x_train, y_train, epochs=epochs, verbose=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# evaluate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "code_folding": [
     17
    ],
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------training----------\n",
      "pred nonzeros\t 0.0299316590594\n",
      "acc\t\t 0.980767100495\n",
      "corr\t\t -0.00795557921929\n",
      "----------testing----------\n",
      "pred nonzeros\t 0.0182511858115\n",
      "acc\t\t 0.963910084554\n",
      "corr\t\t -0.0139195785673\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Library/Frameworks/Python.framework/Versions/3.4/lib/python3.4/site-packages/ipykernel/__main__.py:5: FutureWarning: comparison to `None` will result in an elementwise object comparison in the future.\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABQkAAAGqCAYAAACoOcKmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3XecJHWZP/BPdc6T487mhV1gl2VZclwVFBUQRFQ85dAz\nBzzPM9yd97vFcJyHkUNQQVBBWBEBQUBANsCyLLA554k7OXes/Pujump6eqrjVHdXbz/v14sXy0zT\nXTvTXfWt5/sERpZlEEIIIYQQQgghhBBCKpel1AdACCGEEEIIIYQQQggpLQoSEkIIIYQQQgghhBBS\n4ShISAghhBBCCCGEEEJIhaMgISGEEEIIIYQQQgghFY6ChIQQQgghhBBCCCGEVDgKEhJCCCGEEEII\nIYQQUuFspT4AFcMwHQAmAEgAeFmWLyjtERFCCCGEEEIIIYQQUhlMEySEEhxcI8vyWKkPhBBCCCGE\nEEIIIYSQSmKmcmMG5joeQgghhBBCCCGEEEIqgpmCcjKAlxmGeZthmM+U+mAIIYQQQgghhBBCCKkU\nZio3vlSW5T6GYRqgBAsPyrK8udQHRQghhBBCCCGEEELIqc40QUJZlvvi/x5iGOYpABcA0IKEDMPI\npTo2QgghhBBCCCGEEELMQJZlphDPa4ogIcMwHgAWWZZDDMN4AbwbwB3Jj5NlihMScqpYu3Yt1q5d\nW+rDIIQYhD7ThJxa6DNNyKmFPtOEnDoYpiDxQQAmCRICaALwVDxb0AbgD7Isv1TiYyKEEEIIIYQQ\nQgghpCKYIkgoy3I7gHNKfRyEEEIIIYQQQgghhFQiM003JoRUkDVr1pT6EAghBqLPNCGnFvpME3Jq\noc80ISQbTLn0+WMYRi6XYyWEEEIIIYQQQgghxGgMwxRscAllEhJCCCGEEEIIIYQQUuEoSEgIIYQQ\nQgghhBBCSIWjICEhhBBCCCGEEEIIIRWOgoSEEEIIIYQQQgghhFQ4ChISQgghhBBCCCGEEFLhKEhI\nCCGEEEIIIYQQQkiFoyAhIYQQQgghhBBCCCEVjoKEhBBCCCGEEEIIIYRUONMECRmGsTAMs4NhmGdK\nfSyEEEIIIYQQQgghhFQS0wQJAXwVwIFSHwQhhBBCCCGEEEIIIZXGFEFChmHaALwPwAOlPhZCCCGE\nEEIIIYQQQiqNKYKEAH4K4BsA5FIfCCGVZu3zB7Cje6zUh0EIIYQQQgghhJASKnmQkGGY9wMYkGV5\nFwAm/g8hpEie29+PJ3f3lvowCCGEEEIIIYQQUkK2Uh8AgEsBXM8wzPsAuAH4GYb5vSzLtyY/cO3a\ntdqf16xZgzVr1hTrGAk5pb14cADfe/+ZYBiK0RNCCCGEEEIIIWaxceNGbNy4sSivxciyeSp8GYa5\nEsDXZVm+Xud7spmOlZBTxfl3rQcAvPjFy1DrdZT4aAghhBBCCCGEEJIKwzCQZbkgGT4lLzcmhJSW\n06acBobDLMKcgP19kyU+IkIIIYQQQgghhBSbqYKEsixv0ssiJIQUjttuxdJGHw4PhvDgGx247ZFt\npT4kQgghAPonY/joQ2+W+jAIIYQQQkiFMFWQkBBSfFFeREyQ8N0XDkKQqKSfEELMYnvXGI4Ph8EJ\nUqkPhRBCCCGEVAAKEhJSwURJBidIWNVWDQAQRCVIKFH/T0IIKbnRCA8AGItyJT4SQgghhBBSCShI\nSEgFiwkiXHYr/uM9y7C00YfeiSgAIBgTSnxkhBBCQpxyLh6PBwsJIYQQQggpJAoSElLBopwIt90K\nAGgOuNA1FgEAjEUoa4UQQkotzCpBwhBLGzeEEEIIIaTwKEhISAWL8iLcduU00OR3oWtMySQci1LW\nCiGElJoaHAxzYomPhBBCCCGEVAIKEhJSwSIJmYRNfqf2dcokJISQ0guzAuxWBhGOMgkJIYQQQkjh\nUZCQkAoW40W4HfEgYUAJEjKg0jZCCDGDECeg0edEhDIJCSGEEEJIEVCQkJAKppQbK0HCeq8SJKz3\nOai0jRBCTCDEimj0uxCmTEJCCCGEEFIEFCQkpIJFOBGeeJDQ45gKFlImISGElNZnH9uBQwNBNPqd\ntHFDCCGEEEKKgoKEhFSw8SiPao8DALSy4zqvQ5uoSQghpDR29owDAJUbE0IIIYSQojFFkJBhGCfD\nMG8yDLOTYZj9DMP8d6mPiZBK0DUWQWuVCwDgsdsAAHVUbkwIISUlybL250a/kwaXEEIIIYSQojBF\nkFCWZRbAO2RZXgXgbADvZBjm0hIfFiGnvN6JGNqq3QCmlxtT/ytCCCmd0fDUhPlaD23cEEIIIYSQ\n4jBFkBAAZFmOxP/ohHJcYyU8HEIqwmCIRaNfySScGmDiwMuHBiFIUikPjRBCKtaj27q1P3scVio3\nJoSY1rrt3fiPZ/eV+jAIIYQYxDRBQoZhLAzD7ATQD2CjLMsHSn1MhJSjH68/gu3d2cXYJ6M8qt12\nAIDVwuAvn70Y82o9AIDDA6GCHSMhhJDUjg+HtT8rQULK7iaEmNPz+/vx0qHBUh8GIYQQg9hKfQAq\nWZYlAKsYhgkAeIlhmCtlWd6U+Ji1a9dqf16zZg3WrFlT1GMkpBys296DwSCL1XNrMj42yArwO6dO\nA61VboRZJWOFEymTkBBCSsFhs+DmVXOwrWscXoeNyo0JIaZlszKlPgRCCDnlbdy4ERs3bizKa5km\nSKiSZXmSYZjnAJwHIGWQkBCSmtOWOUlYlmWEWAE+5/TTwGmNPly0oBYxnm5KCSGkFGK8iEsX1eOb\nVy1F91iE+sQSQkzLbs285uRFCe0jYZze6C/CERFCyKknOUnujjvuKNhrmaLcmGGYeoZhquJ/dgO4\nGsCu0h4VIeXLkUWQkBUkWBhG97FOmwWsUFmZhIIoIRjjS30YhBCCGC/CbVfOzV6HjXoSEkJMy27J\nnEn47N4+/MPv3i7C0RBCCJktUwQJAbQA2BDvSbgVwDOyLL9S4mMipOxw8cCeBZkXbHpZhCqX3Vpx\nmYS/f6sL7/y/10p9GIQQgigvasOkZjO45F+f2oN127szP5AQQnLwvb8dxHhU2VjNJpPw8GCw0IdE\nCCHEIKYIEsqyvFeW5XNlWV4ly/JKWZZ/VOpjIqQc3fzgVgBAJIsAX7ogodNmQazCMglHI1ypD4EQ\nQgAAUV7SgoROmwWiJEPIo0/spmPD+Ou+PoxFONz0wBuQZdnoQyWEVBhRkvHM3j7s75sAAFjimYSi\npH9+6RmP4sndvUU7PkIIIbNjiiAhIZXmC3/cgTfaRwx9TlmW0TsRA4Cssk5CrAC/K0Umoa3yMgmz\n6eNICMmdLMv4378frrhzymzEeBGueJCQYRh4HNa8h5ccHgzh3b/YjK6xKA1AIYTMWohVeqSeHFfW\nnGz83K5+fYaEzQnJZBsVRwaDeGJnD44NhTBOm8WEEAKAgoSElMS2rnG8cKDf0OdUewh+dHUbIlk0\nuQ+yAnwOq+73KrEnocum/Cy2dY2V+EgIObUEWQF/2nkSfZOxUh9K2YgJU0FCQC05nv3wkqEQO+vn\nIIRUNrV/80BQOadHeWW9OJmirzObkAVttv6qj23vxg//fgS3/PYt3PXKkVIfDiGEmAIFCUnZm4jy\nWi++cpKiKiNvIVZArceB957ZnFW2SLpyY4fNUpY/01mJt3H8wh93lvY4CDnFjISV7IyUWSYmMhRi\n8b9/P1zqw4j3JJxaonkcNkOyAPMJEr56bBi7esZn/drEHB7a2oEfvHio1IdBythk/FzePxlDz3gU\n41EeTpsFo2H9TLzETWezBQn9Trv2Z6PX5YQQUq4oSEjK3s0PbsVP1h8t9WGUXDAe9PM6bAhncTMe\nZAX4UpQbKz0JzbWQK7SKC4oSUiRqkDCb81KpvX5iBH/aebKkxyBIEkRJhiNhGEBbtRsdo5FZP/dw\nHkHCrz+1B19/as+sX5uYw1/29OLpPdQfjuSnfSSMf3x4G5w2C/omY7jx/jfQNRbB0kY/Osf0z1HT\ng4Tmug54EypqXHa6LSaEEICChKRMXX3Pa1q57liEx9548+RyYnRfFk6Q4LRZ4M2yd1WYFeBzpAoS\nWsHl0SS/nFGQkJDCUANT5dAPz5HFlM5Ci8WHljDM1JT6VW1V2J5HKwSPfXpLiaFQfj23Ku16cCqb\nV+sBABpiQ/KiZiMvqvdiYHJq02FxvRf9KVpKsLwIu1U5nz2yravwB5kDISF9MMSa/xpFCCHFUPrV\nMCF5GI/y2HJiavBHjC+fGxg1OGj0+lyQZNisDHxOG8JZ7NSmG1zitFnAltHP1AisKOEb7zodLrul\nLMoiCSkX5VRuDCgnZraEmdRRXoTTNj24N6/Wk/IGPB2PIzlImF9PQl6ggNKpwhoPPpdD0J6Yj3oe\nr/U4MJIw6GNOtRuj4RQ9CQUJFy2oAwD8ZU9f4Q8yB2rVzM2r5iDE6h8/IYRUmooPEp5/13ocHQyV\n+jBIHiZjUzec5VQaq/ZjySaQlwtBlGCzMHDaLBBEGXyGzI9gmp6ETpsFT+3pLZObemOwggiX3YJm\nv0trxk0ImT2t3LgMghIxQW3AX7pzX3I/QgAIuOwI5nE+TpzaXuOx51VuDABihWadPbClHZ95dHup\nD8NQalYoTXIl+QjGz42SLMNmmcp2nlPtxmiK91RMEOG0WfCjG1fgssV1RTnObLG8hK+/8zRcv6JV\n+7sRQkilq+ggoVpqsb9/ssRHQvIRZHmI8TIBszVCTkedChcyeDEiSDJsFgsYhkFLlQs949G0jw+l\nKTdWyy86DeiBVS7Ucu16nxPDeZbkEUJmGotyqPU4yqInISukn9JZDDFehDupTNjvtOUVuJRk4IGP\nrQYA1HudeZUbq2WClWjLiRHsOjmBfb3l19IkFXUDcSxKWVMkd+p5aEVr1bQgYZ3XkTJIyMbXV267\nFTHeXOt1VpDgc9rgc9ry2oghhJBTUUUHCdUsKZryVp6CMQFhToDXYQUnSKZbeKQyGRNgZRjDFyNK\nkFBZsC1p8OHYkH6G7MnxKO559TgmYwICbrvuY6pcytdve2QbBKkyyo5ZQYLDaoXfaauoDEpCCi3K\niWjwOfDrLe149dhwqQ8nLfU6UtpMQmlGkDDgsmsbTLkQJAnNAScAoN7nwHA490xCM/RpLBV1uNft\nT+wu8ZEYhxNlBFw2jFOQkORhMsbjc5cuxKcvXgBb/Nzwg+vOQo3HnnK6cYyX4LJZ4bZbETVZK5uY\nIMJlV9Z+lElICCEKU6z8GIZpYxhmPcMw+xmG2cswzO3FeN2JKK819RZp7n3ZmYwJWl89n9OKvjz6\nNZVC11gETQGn4YEoQZJgi2d8LK734vhwWPdxr58Ywe/e7ET7SBg1KYKEa06rR6NPubEspyzN2WAF\nCU67BX4XBQlJboIx/pTKNDJahBdRHz+fdIzon5fMQs0kHI+UNpPQpZNJmM/GkpphDgCNPieGQmzO\nAyvUh5fLRpyRAvENs1StOcoRL0po9DkpSEjyEowJqHLbwTAMTmvwotptx7uXNaHOkzmT0GWzImqy\n84h6bGo/b6OHChJCSDkyRZAQgADgX2RZPgvAxQC+xDDMskK/aEyQ0FzlQsBFKeblYn/fpFYqE2R5\nrWR2fq0XO7rHS3x02fn3Z/ejdyKG4TCHnjHjynkFcepmsMnvTNmgXi0d652IoSpFkJBhGC2DolKC\nhJwowWm1UMkJydkftnXjk3/YTtNCU4hyImo8yrnGbjPLskOfklFswb6+0gV9o7wId9LAEZfdAlGS\ncx6oIkoyrBYG33jX6fjYefPgtlsxkUNwSJJlxAQRdV4HJkpYgl0qasCgtcpV4iMxDi9KaPA7qSch\nyctEjEcgvj784QdW4Il/ugiAEkjnRVl3M4EVRDjtFrjtFtNtNrDxfolWCwOvw4ZJCp6TEjHbZ4NU\nNlOs1mVZ7pdleVf8zyEABwHMKfTrcqJyMxBw2Uvaf4hk77ZHtuHZfX2wMgx4UcZkTBm+sWpudcod\nTLOxMMAPP7AcAPD4zpOGPW9iuXG6vnoWZqqHTEuaGx814FEpQUJ1N9nvtOG+107k3eCfVA5OkPDw\nW13a9aMcBnOUQoQXtfOO2RfBMV7ERQtrsb+vdL2K1ZvWRAzDKBuaOZbDKZtHDD58bhsW1XtR78ut\nL2GEE+GyWVHrcZQ0u7JUWEHCdctb4EnRv9csOEFCJMthaJwgwWphcPem4wU+KnIqCsZ4LcM24LJr\nm80MwyglxzprcS2T0G5FzGzlxvFSaABoDrjQH6S1n1Ge3tOL/X2T4ATJ0KSIU9XlP9uUslVUoTy1\n+ySuuue1or4mKQ+mCBImYhhmAYBzALxZ6Nfi4hkDflfl7RyVc3n1YJCFJMvwOW3om4zB57QpfeRM\n3EtEkmUc7J/EW52jsFoYXLaoHt97/5kps/3ykVhuXO9N3XtKLae7fkWLlnmo56yWAACYrjSkENpH\nwtjfNwlHvOSEEyW8dGig1IdFTG5b9xju3nQMe04qWWdUpq4vyolY2ugHANP3fGIFCXOr3RhO0Vur\nGDhR1u0DyApSzj0dEzePgHjJcQ59CdWsxiq3vSLLU2O8iGq33fTB7Wf29uKeV7ML+vGijI+eOxcA\nKPuZ5GySFbQgYbJar0O3L6ESJLTCabPknA1daBFehCeeuT2nyoXeDEP/SPZ+8OIhrNvejd9s7cCN\nD2wt9eHM8GbHKH68/kipDwNAwjDVIm9Q7ugez6m6gFQOUwUJGYbxAXgCwFfjGYUFxYmSEhRw2Cou\nA+SiH29A70R5XQjVhUXPeBQOmwUeuxXffeEgXHYrfE4bQlnuopfCoYEgbn14G770+C7wogyHzYIG\nvxMjeTSRT0XNGAHSZxKqNzuJN456/vOaM7CqrRphE/9cjfLc/n4AgMduhTfee6oQzfrf7BitmEEw\nlWAsfjN0bEjps2f2AFipRHgR7zi9Af/27qWYZM29GI0JIubWeEqaScwJEuw6558wJ+LOlw/n9Fxi\nUpCw3ufI6e8WjU9arvHYMR4tj2x9I8V4SQkSmiywkWwixmMwywwoTpRwWoMPDqtF2zQkJFuTUR5+\nl35mbZ3HgRHdTEIRLpsFTpvVdO+5xB6wLVVu9JZJf/Ny4ognMJgtQeXxnT1Yt72n1IcBAFqG7ebj\nxR3uprfWIAQATFM/wTCMDUqA8GFZlv+i95i1a9dqf16zZg3WrFkzq9dUMwkdNgvCFZQBovbY6Z+M\nobXKXeKjyV6IVRbpRwZDcNksEOIXmygvKkFCE/8Oh3QW7/Xe3Mq+MklsUF/jUUroBVHSps+polqQ\nMP2FwcIw8LtsCLPmvjkyQp3XAQCo9jig3k5zorGLmRgv4st/2oVffnQVVs+tMfS5SWmom0uiLMNp\nsyBo8gBYqUQ5JdBk9oxvQMl4qfM6EOZE9IxH0VZd/GskJ0ozyo3zIcsyRFnpSahq8KXuV6snxis3\n97UeB0ZKmF1ZKqwgotpjN12JZLIIJ6acLJuMFyXYrQx8TitCrDBjSA4h6YQ4Eb4U5fc1KdoSKIPh\nrLBbGYiSrPVKNYMIJ2pDLFuqXOgZK68ECrNjmKng4EiYQ6PfWeIjmuI30UAqNSHjaJHLjc3yOSTZ\n2bhxIzZu3FiU1zLPpwN4EMABWZZ/nuoBiUFCI3CCkknodVjxrWf24fFPXYiFdV5DX8OM1Eyyiai5\nb9aShVgBNR67MpXXY9cGmEQ4Qbn5NHGQcDTC4QMrWlDnc+LIQBAAUO22G5rinVhWZrNYUOt1YCjE\noiUpEByL7+IubfJlfE6/yTM0jcIKIj64shU+pw2XLapDW7UbIYMDPgPxQPGekxMUJDxFJJ5zWqtc\nFZeRng1ZlhHlRbjsSv9fsw8FUoNiAHDbw2/j71+5oujHwKfIJNzyL2twxc82ZX2DLcQfxyT0oa3x\nOHK6CY7yEtx2Kxp9zqwz1U4lMUFClQHlxuNRHjYLU7ApyVFe1M3g0sOLMuzxIV0hTkA9zHPTTsyP\nEyQ47fqbGJl6EjIMAxnARx56Uxt4UmpqtjQANPtd2NY1VuIjOjUktjJQkw0GgjFTBQnVMvP+yRia\nA6UdThXhRdR6HBgOc5Bledp1u5DU+8ZivibJX3KS3B133FGw1zJFjinDMJcC+AcA72QYZifDMDsY\nhrmm0K+r7tZ74ztixY7el4o6iKLcSodCrIAmv3ISj/KilklY73PqTqT9n5cO4/Ed5kgjV0uMv3DZ\nIvz0ppUAlElwEU40LP0+sSchALSkaMAc5UX885oluH5Fa8bnDLMCvvvCQUOOz8winKgtXKo9Dty8\nao7hwYz+eAnL7pOlm5pKjBXmBC3jq7XKjWgRg4Q/XX/U9H3SAOU6a7MysFnKY3J4jJfgsltx9bJG\nnFuiYL4yVG3mYt1utaDKbc86EzC51BhQrju5tJCI8SKcdisa/U4MVsgwJ0GStIoLVhBR7XagezyK\n9pFw3s/58d+9hXfc/Sqe3GXcsLJEEU7ManibLMvg4612nDYr7nzpMPUlJDlRhz7qqfM6ZrwPDw8E\nlfNIQnZ056g5hlgIogRBlLWgZ67tGEhqalk5L8oI88o1Z8BkG01c/Bi/9PjOEh+Jcg6v9zlgYZii\nbjir1zpONHe2PCk+UwQJZVl+XZZlqyzL58iyvEqW5XNlWf5boV+Xi5dcqD3ChAr5gKjlpuXWhDzE\nCvA5lV0fTphaxP+/a86A12mdUcb2590n8cKB/qIfpx5elGaU91otDLxOq2E3zUpPwqnXaA640Dcx\ns7dKjBdT9pNJZrYLeqFEOHHa9EqfM/cpopn0TkRx6aI67Dk5AVYQi953hBgvzIpo8CnB5Xqvo2j9\nOyOcgEe3d5d0Am+2EidH1nkduq0XzCTMCfA6bLjmjGYtW73YlHJQ/eXZqrZqbO0Yzep5koeWAIDX\nYc2pvUosnmXT6K+cTMJrf7kFF/5oA06OR7WehADw4Qfzn6en9gi+b/MJQ44xWYQTEeMzTzgWJBkW\nCwMLw2AixmNH93jZrQVJ6UiyDEnnvKIKuOxYt71HuxYGYzw+/vu3sfvkhKlKO1UnJ2JoDjhhYaaG\n/lViW4VCUCstoryICCei1mPXNsvNQg3GndEcKPGRqGsPa9Hfg2owtxJaS5HcmCJIWCqsIMFptWI0\n3j+jkB9KM+3UapmEOn1DzGwiyqPa7YDVwkCSAfVH6rJb4XPY0D0exZO7p+/SZxrOUSy8pASkkwVc\ndgRjxvwekm8IG3xO3R1RVpC00opM/vv65QBQFhlL+WAFEefftR5jEQ5ex9TPxO+0G16+vqtnAivn\nVMFmZfDnXSfxtSf3GPr8pPhCnKB9ljwOm3ZuLfjrxhdzhweDRXm92YgJSqkxADT5nQhzoqlbQ4Q5\nAV6nFQGXDRMGnZtzlWq6MQCc3uhD91h2WTiCKOlmEoZyeJ+qv79GvyunXoblTF0L/nnXScQEUQsS\n5qJzNIK9vVNZ43ab8nsoVP8nNTiYqS9hYhaYenNIQRGSLXWoUqqyxMsW1QGYyhRU76+CrIDaeO9n\nM+mbiE7rzV4bz4Q00z1buVJbFUU4JUi4qN6LgaDZgoQCbji7VUueKaUIJ8Jtt6HO68CRwWDBss6T\nqfd3lTCkkuSmooOEasnFbRfOx7waN8YKFDQTJRkX/GgDJkt0w5FMPRmOlVm58WiEQ63HAWd8gStj\n6iJe41EWH11JJQxm6a/Ai/KMASIAEHDZMGlQxhovTi83bvA5MJQ0PXkgGMPfDw9m3ai8rdqNRr8T\nY1n2Oio3alDn6FBoWiah32VsWeRElMfzB/pxxZJ6tFW5cWIWZWvEPMKsgGuXt+CmlXPgcViLFiRU\nJ61uPGr+bNTETEKGYbCg1jOrss1CC7MivA6bcg6Y5bl5LMLh/LvW57z45kUJ9hSDS5oDLvRlmY0h\n6PQu9DpsWZWlqtSehA0+pcetVEE3z6wggeWlvPoI3vPqcXzqD9u1/1bXLUZvXHKChBPDYUR4ERYG\nGMmwjuWFqVJ29eYwl/cDqWxqBVYq1R4HLl5Yq60Z1SzVr1yxGAtqlZ7vW/5lDawWxhSBuPEojxrP\n1CaA02aF227FhMkHbJUDNTNtZ884tpwYwYJaL/onzbHRFGIF7OgeQ4gV4vf/pT8HRjkRHocVXqcN\n97x6HHe+XJxWEGo25WtU3USSVHSQkBWUi91ZLQH844XzCxY0+8WrxwHANGnWkTItNz4+HEad16Fl\npfzyI+fi17ecC0DZnf/nNUuQ3N7PJDFCCKIEu87NQcBlNyxbJTmTsN7r1EqcVPt7lfJESw4/mFqP\nA4MGTmE2EzWTon0kkpRJaOwgnPaRMFa0BrC43oe2GjdODClBkkppcXCqCnMCljX58O13L1XKOIu0\nE8vyEuq9DhwbCpniRisdVhCnbUosqDNvkFCWZUQ4EV6HFVUu+6w39t7qVMqCD/XnlvHJCal7fikZ\n4tmdjxMn3qu8TitODIfxwJb2rJ5DGeRihdNmhddhK9hmqhkdHQrCZmXyyv5T4ygnx5UhMWr5eKoy\n8nz9efdJfOShNxHlRDQHXBkzCXlpasPynDlVAGCKG2RSHjgh8+T1xEno4xEOly+ux60Xzte+b7da\n4LJZTDHoazxeoZSozkt9CY3AiRKq4q2NZMBUmYSvHR/G59btxN7eSbRVe0xxXVPLjVfOqUJvvFXU\niwcHCv66UV7E9Sta8LcivBYpL2UVJIzxoqEnmMSLXbXbUbCTxMNvdwEAHnyjw7AhFbOhLCadZVdu\nvKV9BJctroMznpVy9pwqrGqr1r4fcE/d1Km9pMzw8wampgkm87tseLvTmElqyTeEc6rdOD4cnhZE\nGI//fFbGbw6ycWggiE8/uj3zA8uQGiQEMKMnoZFBwiAroMql7Faf1uDH3ngvOdqtLm+heNYZUNxy\n45ggoinggiTLpi7dBdRMwqnz0sI6LzpGzNG0PtlETIDHYYXNaolv4AizCsIeGVSGoR3INUiYJlun\n3ufAzp7xrAYgCZI8LbscUFopAEDPeHYTjtXJ1ADifQnNcZNXSD6nDYvrvdjZM6EFax+77QLt55AN\nNTD+uXU7AAAOW2GChOrGYIQX0VbtzpgVmBiAvvvmc3DTyjkYCZfXWpCUTrp+qapaz9T91FhSpp6q\n2m3HuAnRHMy4AAAgAElEQVSC00qQcPrx1XmdVIJvAC4+GV61qM6LAZNkEibu+7RVmyOTUO2Nfsvq\nudrX/uv5AwV9zfEIh4P9Qdxwdis6RiJUckymKasg4aPbunHtL7dgw5EhQ56Pi5cbA0Ctx16wk0Sj\nz4nF9V68cmQIfRPZLcwL6dvP7IPDasGJkXBZ7ZZFORFNfheWNvnRWjVzVH21265lR6olvGa5gU7V\nk3B/3yQeebvLkGwgIenG8oxmP0Isj66xqffcSJjDpy5ekFP51HvOaAIA05TLG4kVpoI60zIJDSwD\nB5Tm3erP/PIlddrXzfL+JPlR+tcpv1e33aplaReakt1lUTI2TLC4TScmiNrkSABYUOtF+6g5MwmH\ngizq44NoHDYl22U2GfdHBkO4amkjDg3kNmAm3fTQudUeAMrE0G1dY2nLmfWmG9d5Hbh+RUvWxxIT\nJC3gpU44Ho/y+ObTe7N+jnLT4HPgv69T+vGqEx8X13shiLI2DTMTXpRw08o52nnfY7fiM5csMLy/\nr9oTdTjEKUHCTJmECUEeC8Og0e+kcmOStXT9UlW1CYMXusci+ut1jwNjJqhmGo/y0wJZgHKOHAmX\nz72RWXGiNK3f45xqNyZjfNbn0EKK8soxrGqrwsI6D2KCVPLjisTLjV12Ky5eWIsHPrYadqtFG65a\nCPdtPgFOlFDjtmNZkx/7es0/DI8UT1kFCdWF6suHlZTYD97/Bp7fn//0WmUhrjxntadwmYQhVtAy\nFkudOaQGRWK8BFaQ8I8Pbyvp8eRCbaB+53Vn4fFPXTjj+1UuOybii45gjIfTZsHx4bAphm4kTx5W\nXXuWcrNmRNmF0n9q6jUsDIMzmgM4NhTSvjYS5lDnya159PevPQvnzq2eMRTmVJCYSahOqQWUTBJO\nkAx77wRZQZsoPa/Go32dgoTlTZm4rvxevQ4rXj8xUpS2EmrgptbryBgUKLXEnoQAMLfGjZ6x0m+W\n6RkKs2j0TZ0fV7VVY1tX/pne7SNhXLaoLuc+TLwwtYGZzGphcNuF8xFkeXzhjzsBIGV2hiBJuqWy\n717WlPWk4hgnar+/Zr8LfRMxdI9FsOHo0Cl7/lIy/xntz4DSTzOX9iCcKOO0Rp8WLOFEGRcvrDN8\ns03tMc2JEtqq3Rk3DdikctGaAm6Qk+J6q3O04NUz6fqlquriwz8AoGM0goXxXoSJGn1ODJhgWvpE\nlEd1UqZjvdeRc/Y3mYmPr1PuufkcAMq62izZ6BFOxEdXt+HXt6yGzWqBKMnY0j5S0mNSy40B4O4P\nnYOVc6rQ4HMWdL2kVjv4XHbMr/VkXWFAKkN5BQnjFya/04Z33v0qusejWffV0cMlNHBWR44b3ZRb\nkCTEBFELEJV6Ua2Ww6lTxgq5Q2EkUZIhxHcwbVaLVnKcqNozPZOwrVrZwfr9W51FPVY9qTIJP3Pp\nQsyrcWPYgF3L5J6EANBa5ZpWoh9KCFbl4l2nN6JvovQXdqOxgoRz51bjsdsumLabbGEYNPgcGDQo\n0zYYE+CPB5MsDIMrl9Sj0e80bLI1KT5BkhDmRATinyefSylRv38W16RUZFmelm3MCkoJr3IzZu73\n0JYTI9POS23VbnSMRvC7N0t/Xk4kSBI6RsJo8E9tFixu8OW9aP73Z/dhIMhieWtVzud3XtJvT6Gq\niV/rFtYpN9+pnj/V5pSaEZgNVpC0Mts51W6cnIhq64iTp+gNBRfPtku+ZrdUuWYMR0v5HIKEBp8T\nIVaAIErg49kaUV40NJATTdhgnFPlxp93ndQ2S/XwSRUH5bDRQLLzpcd3zWpTIxvpspxVdV6HtgnR\nMRLGgrqZQcKmgNMUfdrHItyMcuOLFtTijzt6Sp5ZVu5YUdmQUNfWbrsV9V4nhkxwvonyAjwJvZKt\nFgaHBkobGNbLal1U78WJ4cJUXmzvHsO+vkn86MYVqHbb0ZLDUDRSGcoqSKikLrvw5O5ebfLobIY+\nsAm79S67FT6n1fDFkjopcUmDcpEsdZBQzZz62U0r8dMPno25CVlNZsbGS9bSTSuudidkErI8Gv1O\nnDu3OqchHYUipOhJCCj9T7JtRJ/2NSRpRv+peq8TQwnPHWaVnlu5ag640DUWOeUmW7KCCKfNgiUN\nvhnfa/S7ss62ySTICvA5py7+P7rxbKxoCSBkgsbdJD/jER4Bl007v6ycU4VbVs8tyECoFw7048qf\nv6r9d4wX4bRZ4w3iS5+Nkc6fd5/EkYRsZrUi4J74QC+zeHp3L3664RhaA1PlUTVue96lmC8fGgSg\n9BDM9TnSDS4BlGvdWISHzcJgXo0n5bpFlGduHAFKkLBzNLvzeSx+jgSA5S0B7OgeRzi+jvnL3t5s\n/jplR4gHCZOzMM9qDuDwYCjF/zUdJyrB1RqPHSMRDpwowWlXhr8YuQ6MxnsRAsDlS+rhdVix6Vjq\nljzJQZ7WgBsnRsKmH4BE0lOrHmJCYdcUmc5NAHBmcwBHh0IYDXPom4xhbo17xmOq3bMfDGWEkTCH\nOu/06pqLFtZhXo2yIULyp25IzK1x44azW2G1MKbJXI5wItwJ90KfvnhByc+ByqTt6e/F0xt8OFCA\n4KUoyfj8up1Y1VaFK5c0AFA2wczQEo2YR1kFCXlRxtJG/4yv5bvg4sXpZRfNAZc2UcgoaubWV9cs\nwZol9SUPEsZ4EfNqPKjzOtBa5U6742wmySVrenxOGyK8CEGUMBkTEHDZcdmiOi2gXErJu/eJ6n3G\nTFJTskamv0aDz4GhhOcOcyJ8jtwzCVsCLmzrGsevXzc+S6qUWF6CM8WCt8nvxIABu2r3bT6BN9pH\nZmRw+pw2yiQsY8lTEW0WC645s8mQ90yy3okYoryISLz3XCw+TMLsWUDq9O47r18+7etfXbMkp76o\nxTAc/zk2J/TPqplFG5JF9V48dtsF8Nit4EU5p0nmehs+iao9DkxEeYRYAfNq3Cl7ewmirFtu7HXY\nUOWyYX9f5v5DStaqcu09vdGHztEIQvH34Z92nsT27sJmLunpHivs4BsuXm5sTdpgzCXgy8eDcT6n\nDa+fGAEnKIFHv8tm6Jokwou4dnkz/uf65bBbLZhX68H3/nYoZQA4eYjakgYvnDYrdvSMG3ZMpPjU\n81euvZQFScppSIfSyz39xrvPacOSei/e6hyF32nT3SA3ejhcvobDHOq9zhlfn1frQWc8a/jwQBDf\nfeFgsQ+t7KkBZa/Dhv94zzIA0yu+SinCifDYp9YgVW47JqKlfT+OR3ltwKHqooV1eP34sOGvpfYx\n/tbVS7WvtQZcp2TFGMlfmQUJJcyv9eCc+GTW2y6cjygv4o48T97KCWwq8FTjceCHfz9syLGqQqwA\nr8MGj8OGlip3yS+KiaVDAZfNFDt52WATshlSsTAMmv1O9IxHEYwpvcL8LjuCJpggy4sy7DplXwDA\nALjzpdm/75KnGwNAvc85LQAZ5vLLJGytVm6cT7ULCBvP7tDjsFnwwBsdeT3vRJTHRJRHjBfx4Bsd\naB+JzAiKJJbkkPIzFuFmTG2s9zq1mzUjxeIZ4I9u61beV/HATZ3JB5eEORE+pw1nNgemff2DK1u1\nCfRmoQaE1KwsYHb92pRArhUMw+R8Q8xnGA5Q41aOK8QKmFvjSXmMgiTpZhICwOp5NVmdfxJ72Hkc\nNrjsVpwcj+KqpY0AgM+v25nxOYz2wQe24qcbjhbs+XlRgt2ilMlNK83NIXNXDQq2j0Rw50uHlUxC\nqwUNPqeh2RoxXkS124F3xX8f37pKuem797UTWLe9e+ZxidP7XTIMgxvObsEze/oMO6ZTVYgVcP5d\n64uacSTJMu7f0p6xRF1d5+W68f/4jpO45t7NWT8+m+nGALCgzotdPROocuv3wPY5bCVfm8d4EZwg\naS1DEi2o9aIjPmBr/ZFBPLuPPh+5Sj7XAECN24FxE7RIifLTMwmr3Nn3my2UYIxHwD39vbi8JYCe\niajhU4eDMQHNAScW109VUc2r9eDwYKisBpqSwjJFkJBhmN8wDDPAMMyedI9Td2bVcqUvXbEYALDx\naH7TjtmkHbHhEIsjgyEtW0M1m+EFiT3gfE5rybPalDI1NUhox0RMKHmKdTYSJyyms6TBh2PDYYQ5\nJUgYMEm2Fi9JsKVYWH3uskUZG0FnQ5TkGdknDb7p5cYRTtSmsebC67Bh7XvPKHgpS7ElN3FPdHZr\nFXrGo3l9Pm5+cCuuuuc1XP6zTdrXkvverGitwpb2Eep7U6bGojxqkn6ntV5ll9zo5vHjUeUz/KvX\n2/HAG+2IxIP9dV4HRsOlP7+lEkmxKaFOZI2aqNyeEyV86qL5WNVWrX2txuPAW51jefUqUjIVlL+n\n35lb9limG3G13DjCKaWmqbId9a4JqhpPdllxnCBOu9Fr9CmlyonB1GJSP1uPbusu2NpFkJRMwh9c\nexZ+dtNK7eu1OXze1LLeD65sBaAOfGCwtNGPu145YtixRzhR+zwBwFktAfidNvzuzU7cvenYzOMS\nZr63rljSUPBedqcCtQdnMft2TcYE/Pr19mkVIXo+89gOALkHCdVrS7Y4IfN0YwBYWOfBjp7xGese\nlc9l0zKSS2U4Xmqs18ZoYZ0H7cPTM5bNMASxnOj1r6x22zGW43uuEBKvz4Ay+LLUSTOTMQEB5/TP\ni9XCYEGtF+0G9yUMscKMqrIajwOr51Vjfz9NOCYKUwQJATwE4D2ZHsSJSglO4s74w7eej8X1M5vi\nZiO5t8Z3rjkDAKalQo9HeVz+s015Zz0EE6ZfmiG9PpZQOuSwWeCwWhApgwtfjBe1wTXpLG7w4dhQ\nSBsU4XfZMGmCkgYhYVpistYqFyKcMOvMGr0LcoPPOS3zIcxNb9abi7k1Hqw/MjRtWnK5S5eh+oGz\nW+Fz2nIu3wGge9M+s4ygFgf6g7jjhQM5Pz8pvfEIj+qk/jE2iwVVLpvhPXcmIjyaA0pJ1LrtPTg+\nHIbHoUw3NnNPwjAnatP6EjEMg9osg1TFomSUzJxyCQCvn8i93CfKTwVvcg4S6gyhSlTjUYYqibIc\nb5MS1Q24CpI8o2RWew63HW92jGbsS5i8kTIR4/HKkSF4HFb87hPnae/LYkm8ThZi7SLLMkRJKdNe\n3lqFC+bXat+r9TiwpX0EPVmUO/PxLJpvXb0UDABJVs4PH1ndhvaRSE4lnukkvs9U6jozOYMXUNYJ\nyde8OVUuhDmhbNrPlIoaJDqSZV9KI6ib3JkyiRr9Tiyu9+b8vlLfO9kOMcwlk7B9JDwj217lc9q0\n3qalMhJiZ/QjVC2KHz8AbaOdehTmhhdm9mJXsvNLf55JziT0u2x4q3MMOw1ou/DwW514TCeLOx1Z\nlhFMMVhyUZ0XJ0YMDhJy+q91WoMPR4t4fiPmZoogoSzLmwFk3MZUS3CuXtaId5ymNNps9OVf3sUJ\nEhwJfe6WNfmxrMk/LUioLrby7VWYGK03R5BQGQCi8jiMH9ZSCMqNShaZhPVeHB4IYjDEKpmELjt2\ndI/juRKXCqRbWNksFtR7nbMuPeWEman9XocVoixrC8xIvPwvH63xXl0PbGnHmx2jszrWfIxHOJx/\n13pDn5Pl07+vGv2z/72oqpMWyzaLBfd++JxpmZ6kfIxFuBmZhAAgA9h0zNgeMmNRflpZyPojQ/A4\nbKYLtCWLcCI8KXqgmqU3kYrVKY2q9znx4VVtOWeGipKs9DyOX2uV/qPZX/uFDDfiLrtVu7lf0RrA\nW51juO0P22Y+jySnzGCv8Tiw6dhwxnN58rX3zCalL7TXYcOSeh9GwlxRB1qxCZnXhVhPifHAql52\nUWN88nU2bShYQYLDysDCMNOyaefVeLCiJZD31OxkUV6cka37q1vOxR3vOxNRnSCqXn9khmFQ53Xg\na0/uNuSYTlXqz/O148P4wh93YHsRsi/VzYUwmz4g7rZbccWSejy7r0/rpZcNQVQ+u9lmyLJZTDcG\ngIW1ylDEmhTlxv4czonrtnfjl5tPZPXYXCj9CPWPb0GdF+2jYUiyrAUHH3yjw/BjOJWxOhsS1R4H\nBiZjJa9gS84kVDcI/5hjcE/P3ZuO4xc5DmaL8iLsVkb3ur+43ovjQwYHCWOCblXZaQ0+HOgv7ZRn\nYh6mCBJmSw20vO+sFvzvDSsAKDcaav+vXOntqFa7p9+4qDuHPeP5NcoOJWYSOkq/cxZL2nUeCXP4\n8INvlvCIsqM26s/ktAYfNp8YwYsHB+BzWrWdkrdLXEqj9CRMnRkyFuXw+I4e3e+xgpjVLq9eJiHD\nMGiIT08WRAmCKGfs7ZhKvc+JL16+CK8cGcKX/7Qrr+eYjfH4jrqRJR+sTmA1UZPfiYHg7EqLWgIu\nvP+sZtR6Zi5GG/zOjGVExHxkWUbHaAQNvplZVDee3Zp3C4xUxiIcGpNeK7HcuNQL7lQiaTKXjdgY\nMVKqqZ3zaz05l3QrE4Gt2uTrXIdVcGkGXanUiaE1Hgfuv+Vc9E/EZgTrxDQZiepjMwVqY0mZhOra\nS5BkOGwWeOzWomagJQYJMwVO8iGkKdFu8Dlx86o5WT1P4sZg8s3YnGq3Vro6W1FuZibhqrZqnDu3\nGkcGQ3hoa8e076V6n49GeOztnTRdSaWZ2nFEeQkrWgPYcGQI27rGsbUIm6VqIC1TJmGIFXB6fLBj\nLkFCtYXMcJYZ6XyGNZOqKaBsKrtT9MD2OW04NhzGJT/ZkDHY/+P1R/GbNzqyOr5c3L3pWMr2O0q7\nIjv6JmI4MhjCVUsb8dKhQcNbiRTboYFg0RJDeJ3WBg1eJ3adnMAfts0+GDcbyZsrrVUunNHsz6sd\nk57kyqFM9EqNVYvqC5BJqFNuDCiDUvb0Thh2fSLlrQyDhNMXb+oi/LPrduT8fHpBlYDLhkmdIGH3\nWH4fGCVIqJyIPA4rNh4bLmlT0CgvzVhQ5hs0KqZsexLOr/Xg+9eeCUC5yKu7mMl/52JL15MQAN69\nrAmvt4/ofu/Lj+/C7X/KvMPPp1j81/ucODYUQohTLop6GRLZuuHsVu3PmRZLo2HO0OBFJF5O127g\nxZIVpbRl7EZkErrsFqx935m6U0brvc6yyOQl0204OoS/Hx7UboQSXbSwTnuvGkGSZQwEWSxrVm4A\nn/v8pQCU64nLbkWt146uPK9PhRZJUW4MAPNq3OgcM3bhOxt6TdYBZcBQrsNhkgM3J8ej+Ldn9qX9\nfzYcGdIyVZIn0OpxJWT3ndNWjYDbht6khb0g6U83BpQm5QDwZsdo2gzt5JYM6vVDLSOciAn4z7/u\nT3usRuLiQY1Gn7MgmYRChlLvq5c1YWvHaMZp1VzChlxy9v6cardhmYQRnXJjYKoH7r2vTc/A4lK8\ntz4cD36aqWn9rp5xXPrTjaU+DE1MENFa5caquUrf0kyBfCOo7/FwhmtKmBNw6cI63Hh2a04bjzFe\neR8PZ1nRoHffpEe9N6ty6wdd/PHPBC/KOJGm35pa2p9vBUwqMV5E70QsZbkxAJzZ7Mdz+/vgsVtx\n5/XLsaTei7c7i19FY6RP/P7tggRc9ei9V5Y0ePG1dyzBCwf6i3IMqSRvYFoYBl++YvGsz8tq249c\n76uDMQE+nfJfQKlyPNA/aWiAOsKL8OoECavddlxzRhPW7ShtEJeYg/mjQwle/v09+NOvfoK1a9di\n48aN2tfPm1ejpbbnQi+LKOCyT+tBpk6V3JznCPLETMLlrcpU5t0nJ/J6LiMk96+57yOrsDRePmRm\n2Uw3BpQbmKuXNQEAqt0OuB1W/ODas/DErpMlbcydrichANy8qi3lwmsoxGaVCalXLgcADT4Hvv3M\nPvztQH9ek40T1SRkw2Xqu/aeezdjXYrsyHyogZdbH95m2I1MpvdVnceBZ2ZZqp48cTqR12EFL0lg\nT7GBMKc6te/TGTrnTqMbYI+EOXgdVnxgRSv+9sXL0Oh34rHbLsBli+oBACvnVGOXAX10CiHMpy43\nXljnxRM7Txo+tS8Xgihpmw6coEyfTZbPhGPlOjv1XJ+7bBEApH1f/GTDEdwXL6nLVG4MAC1V0wPU\nS+qVoV2J0rW5uGRhHW6/cglei69tUmWQcTqtPjbcfgXee2az9jxvdo5lDJoZhRUlLKzzYlG9tyCD\nD9IFVgFofd+ez3CTmzggJDljY16NB8cNakSf3FtLlbipmjhNmRNF3bXGFy5fjOaAEzc+sNWQ4zJC\nRzwjziyT0KPx3thqgKsYvdXULN3O0dTvF0GSwAsyXHYLGvxODOZQ/cDmmkmYRZaz6sUvXoaPnz9P\n93t+lx3/fd1ZePeyRnSm6fF54wNbcfnieqT5SOZlNKKUGn85PgBTz1XLGnH/lg6c3qi0+njfWS2G\ntxIphWJdc/Wq9RiGwUdXz8V4hDd0wz9XEV6EW+e83JVDFq6ezrEIGv3OnNvATLI8AikC4fU+JxxW\nS9af0WxEuJltKlT/cN5cPLu3L+s+pbO14ciQqTanzG7jxo1Yu3at9k8hmSlIyMT/SemSj34Bt33l\nm1i7di3WrFmjff2jq9vA5rGI0Cu7YBjgrleOaKU4UV7EhfNrcHw4jI40F+lUQtxU3b/PacNnL1mI\nH7x4qKg9fBIpZbsJzVpz7JWUiizLBe37F+WlaZkT6VgYBk9/5mKcGc+8UTMeitE/JpVMzZ7r0gwg\nWFCnDOb55eYTaTPzUi3e/PEU9p7xqCGp9J+IL/qyuQgambKeWKr/3vtex5cf3znr58zUk/C9ZzVj\nf98k7t44c0pkttLdcDIMo00qJeUjxov4h/PmolYnCyHgsmEiTTDopUMDePnQQNavNTAZQ1PABauF\n0bIeljT4tPfV0kY/vv/ioaw3stYfHsT27uKcC1NNNwaApU1+DARZvHJ4sCjHouf5A/1auw1OlHSn\nzNd6HRgMsjllRSdvxl2+uB6nNfjS9jbun5w6//MZNpUA4NtXL8VfP3+J9t/zajx4aGsHNh8f1m4C\nM93Q1/sc2qZov05g4YEt7RgOczNu9BKzen7+oZVYVO/FkSINtFIHqXgd1oK0bxFEKW0mYcBlx9fe\nsQSHB1L/fWVZ1gaXAJjxGbhkYS3e6hw1JPgV5cWUJf0fOLsFZ7dWTdtk5AVZ930OTAW9zFDiK8sy\nXj+hVFcYlXWZaCzC4X33bc7pcx3jlYqWT128AJctqivIje36I4O4KSFQq05W/+Xm9pSbiWFWhNep\nVIk0+pwYzCWTUJAwv9aTdd/1VBnXemq9jrSbpFcva8LCOm/awIzXYcV/vfcMsIJkaObwSJhDU8CV\ntrJm5RwlY/T0+GbgsiYfjg+X/1CHWRQT5SRVOx8Lw+DK0+rxagkDrsk9CQGlcogTpVm1ABqPcJhf\n44EgyTm1bgjFBPjTlCjX+5xZZ/tmI5JicwkAWqrcmFPtxoG+4vQm/OZf9mbV55co1qxZU1lBQoZh\nHgWwBcDpDMN0MQzzSb3HpQq0VLsdGM/jJlvvYrcmPhBFfb4YL8LvsuPsOVX47dbOnF8jzIrTdpGv\nW9GCICugK4vpeIWg7oSqlF5Jsw9QjEV4rH3hoOFTPVVhVoDXmX0W3Jxqt3bxX95ShTqvA6XsJMKL\nUtqehDUeO0YjPJ7Z2zvje5H4Dd9v3uhIu0hiU2TCqGWKPeNReA0ou759zRKcP68mbZmsenH8444e\ndBi0W5hcbvNm5+wDHcmTO5MtqFUCtE/v6U1bEpMocfetwefAjStb0zxamXqZOAltIsrrTio91cR4\nEU/sNC7TtJj0JoqqqtxKNnqqm8//eHY//t9z2U+0DrGClrmip8GnBA6znbz4rWf24VtP78369Wcj\nMZM+2bImPz64shXf+9uhkk3TU0vtxiNcynNBW7UbVW47Xj6UfTAzys1cgDelaV2gnjNq4xtavCTB\nnubmGlACdU3+qWzCBr8TB/qD+NqTe/CbLR0AUvefUy2p92nvrZ+sPzojYPtWvLRObwpiouUtgZze\n07Oh/p2CrIB/f9b4Mmel3Dj9z35+rReP7+zB51O0uVGzEdWSy+SNomqPA41+JzpGZr8OjOj0JFR9\n5z1n4JJFtdPaEaQLKqrvFSMzVvK1p3cSG48OwW5lDPk5Jesai2IoxGUs402k9vSeV+PB5y9bpGU6\nGumtzjF0jUW0NcCxoZDWsiHVZmKIFbTSwUa/E0M5tEhheRFLG31Z9zFMVa6erxWtVfjtm514fn8f\nNh4dwpO7T2rfk2QZUV4Ztnfxglo8v9+4EtXhEJe21BhQztkAtISDJQ0+7OyZwL7e0lWDzYZ6nSlW\njkqq7HwAuHBBbclKt5MHi6ksDIOz51Rh7yyq/aK8pPWLzmUTYTLDOq/e6zA4kzB1v2gAuGB+rXb9\nL6SfbjgKAFm1MCDFZ4rfiizLH5NluVWWZacsy/NkWX5I73GpdtdrPHbs65vEr3KcfqUXVLlgfi1O\na/BpO3FRXilHPH9eDZ7b3z+tbCMbIXb6BKHmgAsXLajNqbGwkSaivDbFCVB2xY3IJOybVHYhsw2k\n5CrM6TdZzYbbYcVnLlmYMqj1+I4ebDhi7KCBZHyaKZPAVEnq9/52aMbuUyihOftommA4nyIT5oaz\nW/HVNUvQMRI2rClvkBXwlSdS90m8/GebtD8fMSgAoGbHNAdmDovIVzZl7D+6cQWCrIDvZNl3K7Ef\n2Z3XLccHV6ZvdD8S5vBovIlzjBdx1T2v4TvP7TdNiVWhdI5G8MO/H8HX/rxb6ztULvR6u6rsVgtE\nSU57QyPk0FsmkuamHlDabQBKVmw2GCh95Irx/pqMCdOuN8nULOl05WaFpJbyXf2LzSlvaGwWCz5x\nwTz8dX/2mfJ6QeQGnxN/3denm4kS4URYLQxGIzy+8dQe8GLq4RmpJE7pVM+VmW7oT2v0Yf3tV+CL\nly/CG+2j+HZS38TWKjc+cm6bFuxK5QuXLULnaCSnDNl8qcHcL16+uCD9lMU0g0tU6tCY7d36Zf5s\nUnD2kxfOx6cvXjDtMUsafDg6y+zLVDe70491egldusD9gx9fDQD4lyf3zOq4jDAYjOGShXX4p4sX\nYJ8pH+UAACAASURBVPMJ4zOO1A3tkRx6AseEqQF6Sxp8GI1whmcTBuNZ6OrAtCAraJ/hLSn6Vif+\nThvyyCRc2uhH52gE3/vbQRzL8J7MtPGQqwsX1OI/r1mG/3r+IL7x9F7c+dJh7XthVoDbboXVwmDV\n3BpDkyuGw2zKycYqhmGw8atX4MolSvKI2m7nf185YthxFJP6Xtebel4Iqe5JAGW40t7eSe39Xkxq\nUF3vunZGkx+HBvPPoIvyynu2zuvAcA7nlmCMT7sZV+dzYMTATMJomnJjALhoQS22nNA/3xhJvfcp\nUnIrRsJcyeIv5cgUQcJscSkzCZWbkJdy2OlXy0Hstplvzcb4xFFJlrVd2g+tasP582py3jnUW5A1\nB1z416f2orsEN0YDQRbNCc32PQ4rYrw0635CvfHg6YH+yVk9TyohVpxV42IlU4/Dvt6JGRk+d71y\nBN/8y96CTvXjMmSsAcCHzpkDv9OGx7Z3T1t4hjkB58cDAalKfGVZxmQs9U5Ua8CFkxOxjJMss/X1\nd56mvW6yxFT9T100P68yfT0RTsQNZ7fiT5+6yJDnAzJPNwaU4SIA0J3lhPMwJ8LvtOGDK1uxuMGX\n8fFPfVr5+zy0tQO3P6FMjX712DB+/MrRrF7PDNYfHsR4jlnEkfjnbfOJkYJOifz16ycMH3CQ3LZB\nz9oXDup+PdfFkNI7Js0Os8+JL12xKG2Jc6I51UqAoxiTOSdjPAJpFr5L472efrXZ+N9RNhJ/Zqwg\npjwXXLSgFrt6JrJuHK4X2G2pcmHD0SH8SOcGM8wKWlbLxmPD07LQsrW4fupco04A5rMcMnDTOVMb\nGYnXwRArYPXcmoz/f3188va/P7sf3//bwYJO21aDhIviAWajCZIMa4affWvCGkqvBDT55nhlW7XW\nl1J1WoMPR2ZxMwrEg1YJU7T1zK/xYEdCe4EwJ6QcJrSg1oszm/2G9UucjfEoj+aAE1cva8KbBThX\nqZvGqdq86FHuB5TzmdXCYH6tx/BSaHWNNhDPBgyxAv7n+uUAgDtfOqzbTy7ETQ1IbPI70T4SyXrj\nO8aLWNLoQ/tIGM/s7cOOFIFvFZ9DuXG21IqNZEF2apOpye/UfiZGGA5nziQEMGO4w39esyyr/8+M\n1ODIcIgrypTm5M2SRAGXHaIs453/91rRJ6rf9cqRlIHSlXOq8eSuk3n3CVc3kBt8uQ09DMaEtGul\nJr9LtyVIvsIZgoSr51ajayyC/knjXjNZ4pwAdS22r3cCH/vtW7juV68b+lohVsBPNxzF2ucP4EO/\n2YrhEEu94LNQVkFCIcXFSQ2MeB1WrNvendUJR5RkMAyjW1ZS73NgKMTiBy8ewo/XH9VuBufVenLe\nyUqcbqxSbx5KUWI1HGJR75u6wFkYBqIs4xev5ZaFmexgfxDLmvz4q4HlAIkmY/ysgoR1Hge6xqL4\n5B+2477NJ3SDomracyFkE4z61tVLcfWyRtz72gnckRBgCHMifnDdWbhiST2e39+vewPWPR6F32lL\nGUyoipexDRh0kTmnTenV8sAbHdjaPgJJlrHx6BBkWcb77lNO7l97x2lYUOdFu0GlQiFWQEvABZfd\niic/fZEhkwVZIXOvy8Z4yUmMl7I6t6gNgf/t3cuyes+21ShDl+597QR29kyVObzdNWaK3lB6tneN\nTSuR/tYz+3IeUqNOvv3E+fPQNRYtSIl1iBVw/5YOw8uDovxURomeBl/qm4hcM8TSNZhWVbsdWW8A\nWBgGly6qw788uafgvXGDGTIJz51bg89eshAdo5GCBAIyCXMCbr1gHubVuONN1vV/zgGXHX6nLevA\nql4QWW1l4nPO/HmEuembYPncwJ3W6MN/vGcZgKkJtan6LCZL3FxKzKzSW7+k8oEVLQCAv+ztK+hO\nvdomRi3nNvq8IUhSxs+ozWrB2994J5w2Cx7XOe9lamMBKFUrrx4bnlVANcpl3qyYX+vBREzAieEw\nZFlG/ySb9rp094fOgctuKfkN1ESUR5XbjrZqN8ajvOHDFtTSvc+t25mxTY4sy4hwwoxp7c0Bl+E3\n0eNRHovrvdpaLcQJOKPZj6p4MkSfTu/AMDs1qVT93X7zL9m1lIgJEqrddiyMB93lDI15jM4kBDDt\nniRRMDaVZKEmbxileyyC1ip3zv/fGU0BdIxEcHggWLRhTUb5+cZj+MCKFuzpncAvXj1e0Nfa1jWG\nt7vG0p4H1c2sQvQcTefFgwMpqzkuXFCL5oAL+/vyS3hRE4uWNvqxP4ekmSAr6K4NVHOqXHhoa6dh\nQ+qiaYbKAco17rLF9fjHh7dpLa+MIMsyfvjyYciyjF+8ehzXr2jBDz+wXEvM+uu+fhwdCqF/kjV0\n4/iVw4N4dFu3toZ7732v4+ez6DVfKcoqSMil6OvGMAweufV8HBwI4sfrj2b1IWLF1L0Sqlx23PnS\nYTy3Twl4qVMKlclHuZcbJy/IPn3JArzjtAatxGoyVrwpT5MxAVU6N22PvN01q4mcD7/dhZtXzcFQ\nkC1IM+djQyFtEZOPWq9D+xk/tLUTu05OKBPhEi7ygyEWD23tMPSECADP7+9TytazWFipWZ7qAl2W\nZaUfo8MGSZLx9J5e3Ruw7rEIlqTJWju3rRq//8R5+M3HVuf5t5jpnDlV+PXr7fjKE7uxbns3vvH0\n3mmBoo+dNxcLaj2G3TCGuanPUlu1G5I0+6mH2dzM1SZMdL78Z5vwPy8fTnsTn25YQyp6mR1dYxE8\n8nZXTs9TLD9efxQ/WT89qP6bNzpyajcQ5UVcuKAWZ7YEsLd3Alf8fBPOv2u9oQtv9b1n9CI0lqYn\nIQA8cusFqHLbsa93YsZrqxtT2QYHwlm8n6pcduzoHserx4Yzlu+EWAEfO28uAP0bTiNlyiQEpm5K\nf/NGR9FLjyKciDlVboxF+IwbOSNhDv/8591ZXSf1+sQtrPPi/lvOxYjO9TGcFHzI11VLG/HlKxZj\nIMhiw5Eh/Pr1djiyCEonNu9P3AhNV5qa7DvXnKH9nW9+8E0cLFBVAZfQIqLGY0f7aBhvpCjDzIcg\nZu5JqLpueQvu3nR8RrBdyfBL/xxnNiuDe/64oyfvm6HEDLJUXHYr3nl6A44Nh7ClfRR7eie0gJOe\ngMuGGC/hxvvfSPu8u09OQJJldI1FDNt8TDQR41HlssPCMJhX40H3mHHncEGUMJRQupcpIPDyoUFc\n+fNX8ey+vmnn4maDs3sApR/6siY/Xjw4gL8fHkQsnpmkXhf1ftb9kzGtFDbxs5zNNUa9lt1/y7lo\n9Du1FgypcDlMN86WXmbemp9vwuHBoFaC2eBzoms0YliW8tHBEJbFB5LkYk61Cz3jUXz892/nVL1W\nalFORPtIGLdeOB8ACn7PqfaWTHdN/eLlSnZ1MYOE2Wy8r55Xg9159iWMxQeCnNUSwKPburP+OWda\nK7VVK8kEv3/LmHuCiWjqacqqW1bPxWiEw2EDE5pivIQndp3Ezp5x9E3G8O2rl6LR58RbnWP4446e\naRt092wyJpC9/vAgvv/ioRlff+34cEGuXaeSsgoS8mn666glVACyyqjghNQ77GozcDF+MVIXvvNr\n3egcDedUIhLiZi6y/S47LltcpzVj/sn6o9qExUJLdyLaOYsdCq/Dinee3ohz2qoNnZx5YjiM/X2T\n6BiN4PTGzKWbqbRUuZAYX+6diOK+107gkp9sBADcef1ybD4+gntfO4Erf/7qLI96StdYBP/1vJIV\nmK4noaoxXra1s2cCm44N4bdvdkKQZDhsFozEd7v1Si76JmJoSSiBSsYwDM5oDmBuPGvNCN99/5na\nn3+6QdmRUQNHW7/+DgBKFkPXWGRWpQ0xXsQnH9kW7yOifBYZhoHPOfvplukGUKisFgZv/es7tOyw\nP+86qZXX6wlnKA/V8/RnL8FvP37ejK/nM7W9GMajUzdYiYNatuUwQTzCKb1bWqtc2Jdwk5ZLH5d0\nntvXp5VvG3mDCaTvSQgow0tCrIBP/mH7tP5ekixri9THtnenfR+plOyV9O+nao8dPeNRfP2pPfjV\n6+1pHxvmBKxoqcI5c6oMv8FNNpEhkxCYKuc/OhRK2XOrUMKsgKaAC0FWQO9ELO2GwZ3XL0eNx45/\n/vPujMOYorykOzmw3uvQzuMqWZYxEeXhddiw8atX4F2nN+T3l4GSRXTTOXPQORrRMolyPfXe/sRu\n7cYmyKafuJjsiX+6CN+/VrkubD6u/C5DrIDz71qPa+7NbZJsKrGEYK7PYcMPXjyE29P0x82VKMtp\npxsn+tbVS+G2W/HVpNdX2oukv64wDINPX7IAP15/FB95KL/132Q08+cLUDa4O0cjGAqxcNosaYMj\napBpKMTh1t+/ja06n8mhEItPP7odu09O4KYHtuLaX27R7SvbMx7Nu5RwIsprrYTm1bjxdwOnoL//\nl6/j6T29uP3KJWj0O6cNdtGTeE1KDBI2BZz4xasnDCuXlGUZ41EeZzYH8FbnGP4t3iOUYRj8383n\nYH6tR3djZ8PRQVyyqE7778duu0D59/bM2f3qMMMqtx2fOH8e7t/Sgc88uj3lZzXbFga5UD8rd7zv\nDFgtDD6/bgfCnIgXDw5ogYwGnxMOmwWbDJqIOx7ltcBqLjwOG3778fNw3fIWHBwozEZINvacnMgp\nq2wsqpRXt1W7cdXSxpwG9uRD/eym66f8yYsW4MOr2vDsvuz7/c6W2rLp17ecm/Ix58ypxr2vncD1\nv9qS8/NH4vcUK+dUAQD+muXfLRgT0vYkXN4awC8+fA4O9k8aUirePxlDc1Xqe0YAWNrkx1VLGw1t\njaZmhH9u3U4sqvPCbrVon8Mjg0EMh1i0Vbvx3fefib8dnF1l4v6+SQwEY/hW/Dz6whcunfb9/kml\nYpSkVmZBQv2ehICyOL7rhhW4ZGEd/vO5Axmzl5ReCfoLwQ+tmoNXv3olblmtZFuoF7B5NR682TmG\nf/jd21kdryBK4ARJ9yR5VksA648O4kevHMFzBSrRnXE8koQYL80YXrHukxfgUxfNx46u/IKEgiTF\nU5etuHZ5c15ToFP5zl/347ZHtiHgsmUsq0nHZrHgla9cgQ23X4GPrm7DSwcHtN3jR249Hw0F6jHy\n4BsdOT2+LSHY/acdPbg3oQz8ny5aAEB/J7lvMoaWDCd8o7VUufHIredrZZcLapUA5F03rNCmOXoc\nNvidNrzdOYrRMAdJlqcFlbIxGGKxr28SLxwYmPYemIgJ+NQfts/q75Bt1h/DMLhuxdSU4g8+sBV3\nvqR/ccmmPDRZtduuTdL70Q0rtK+nm4id7OhgKO3EaaNIsoyhEKf11kucuLj+SA4TYOPnjGVNfvzw\nA8u1r2fKZMjWb97owGRMwHnzarC1c/T/s3ff8U2V+x/AP0/3orulu5RRKENm2RsVcYCKExUVxb2R\nq97h9vpTXOhVFHGBiuICRBBEKBvKpkAZZbWle++kSZ7fHyc5Tdq0TToopZ/363Vfl6ZJe1p7Tp7z\nfb6jRUtrGwsuOzoINdBgfrNVoa0pU35/YwqmLdyBvY0EVm35ezK/fm09Vf9NVFW1HnqDhJuzA0K8\n3ZCc1byeaI0xlQw25Pr+YZhlHOrw68GMZm1W2auiVplvQze/E2KD8Mt9w5GUUYLHfz7QYMDL1Ly8\nNn9PF+SXay1eu+10Pub8dgieLo7wdHHC4KjGewA2xMvVyaJg0NZrwgOjYtQekbd8uQu5ZRrlxsWO\nNh/BnZQectP6haKgQrnemzYN88u1Fi0KmkoZMKP8blPyyltsMJaJTi/rTCNuyP0jumDn2QKLQJEt\nGeqAcpP82Nhu6rU0r0xjV0llqabhZvcmMQEeWLjtDE7llWHm0KhGp9POmdgDgZ4uSM4uRcJJy+uJ\nVmdQW4o8sLRmuvMzv1kOO/lyx1nc8PkOLGliNnxxpU69dkzqGYxvdp1rsWCcaQDcjCGRuGNIVKOb\nNWWamvck89+dqXfl8eyWuY5WVuvh4KBsrtbWL8wH1/YNwU8HzlusoT7YeBJ7Uosw2ixIaKoqsaWF\nTpWuZsPLFHA+cL643s26iuq6k9tbwk0DwjEiJgAezo7qQKDEc4Vq72JHB4Fp/UJbJEPZICVKq3Tw\nseHcsaZPqDeu6ROCIxltFyR8bkUSZi+1Pl3dmsKKavi5u8BBCDwzsUeL9Qqvj7MxG9vUr7Y+1/QN\nweaUPGSVVLVKy5naCiq0iA32wkBjyyRrTAG+zJIqu6sbTGtDN2dHzLu+n83VNQ31lAeUNjHxUX7o\nFuiJ4e9ubHTAUEO0OgOKKqvVDdqGhHi74bU/j2HuctOmY/PW0ebBaVObgXBfdzwxrjtWHc7C3ydy\n8eo1vXFVXGeUa/X4ab99bYxMdAYD7vl2D679VAn0vnJ1bwR6udZ5X7bnvb4jaldBQq2+/sAeoPT8\n6RvqDQA4m1/eYAqzkmJv/YR0EALuLo64z3jjYnpDNA/C2DLluMyYASKsNJXuGuCJqmoDfrSzj1dz\nlFUppSm1m1x3C/TCyJgA/LgvHeuP5+CVNUft6v9i6ofiIATG9whCSl454udtaJFjDjP+zht7o7GF\nl6sTvFyd0DO4E3adK8TetCL8/uBI9OzcCUHGAM0dxjK8VYcz8exvzZ/wJyXs6qXYP8IXu56dgEfG\ndMWuc5bBg3E9gnD3sGir0/gyS6osBtJcKD07d8I3d8bjs9sG4oUre2Lm0Ci195ZJqI8bHv/5IN5e\nfxxvrz/RaBmTucziSotdrL5h3hafT2tmmYItmYQmXrUWxr8ezLAa8CypqrbrxtrE39MF9wyLRn+z\nxctn287YXK4945tEzF66r0UzLqxJN/a/dHAQyCyuRG6ZBt0CPfH1nUOwP70IhRVamwITppJMByEw\nMTYYUcaJoS3Rh6SoshoFFVrcPjgS/7yyJ0oqq7F0TxrOt1BZiy2DSzQ6A3oEeSG/XKv+TKbp8t/O\njFeft+ZoVoPB84rqxgPZEX4e+PX+4fjH5bE4X1yFBVut95gtqVJuvIUQKNPqMD8hBSm5Za3SD9CU\nIefj3vC54CAE7h0WjekDwrEvrQgP/7Bf/ZxpeiugTP8r0+iwbF+6xSCG5iirVebbUJ9JQKkCSHx2\nAqr1EpkN9CKrb6qzh4sTNDoDzhYoDcG1OoN6fpuyj6cPCEfCk2Ob8uOo7h0ejacnKMOl/G3cAJs9\nMgbf3j1U/fi9DSeNfZLsu5Y5CIFBkX74+cB5PP7TATz7W01vtPc3pjS7z6r5ELCYgJpgSku1CdEZ\nDDZnEgLAzGHRCPV2wwu/H8aTPx9EamGFTT2ITW4dFIHsUg3eXn8cD/6w366qkvxyrdX2MbVd3rMz\n3Jwd8MPedJvWCdP6haHUeM1ychRIL6zAksRzSC+swN1LGt4kzyyuxP3f71WvQU1tY1NcVbPBcEWv\nzgj3cbNram9Dovzc8emtA+HoINTy0YaUVOnw9IQemDOxB3qH1KxBrusXij6h3vh2d2qLtAYpMmZP\nDo32w439w+p8fsbgKOSUajDi3QSsOpyJ4spqHMooxlPju9d5PzJVltRX+fTx5lPIKK60eC8znev+\nHi54f+NJq+9Jldq6Q5lawnNX9ISfh4v6d2divt7tEuCJL3eeww/N3GwwbdbZUt1Tn7iQTjidX273\nwLaWUFWtV+/jbJ3OXVihha+xN3mgpwsMEi3a47G28modXriyZ6PvH71DvNEv1BvXfbYdY+dvalbb\nK1vkl2sbDY6Z3jNdnRzw33XHUVWttzkL3nxqcLdATxw8X4wpn2xtdF1bVKm1aG1kjRACb07tizAf\nN+w40/Q1W05pFYK8XG0KkJnO9YSTuXh+RRKGvbOxyd8XgEVswXzDZXgXf/XfUX4eatzk7fX2TRI/\nnKFk2G6rNZn5qt6dAQBbnx6P3XMnYvFdSuXW+aLKC9burT1qV0FCnV42elE3/QFuOJGLGz7fUW9Z\nkC09gEw7tKbApHmfmqkLdzQ6uaihfj5CCHx220CLx1q7iXxxVf2lQ71DvaGXEi+sPIxVh7Nw2I4d\nMqUkSfk5zQOQLXGzX6bR4YPpl2HhbfWnhttrfI8gTOndGS6ODuqC2TScoo8xyPzKmuQWKWvILdNg\niJ2ZIQ5CqD3DagvycsHe1EL8evC8xTCG03nliGrBUmJ7dA30xKBIPwyK9MPj47rX+XyA8Q25oEKL\ng+lFyCqxfWFy73d78dQvSrDW38NFbQUAAA+NjgFQk63528HzdvcorKhuvJSzNn+PmnMop1SDT7ac\nwudmJZ7ZpVXo3ISArYMQeHRsN/i6O2P33Il4eUocANi1k5ZaWIEXVh7Gsn3pmPXtHqw+kml35mZj\niiqqEeXvAb1BYurCHdickodofw/0CfVG/3BfPL/yMCZ/srXRr2PKJDSZP70/uvh7oKSq+deNlNwy\nxAZ74ZmJPRDp54Eofw98kJCCT+sJntmrStfw4BIA+PX+4fhixmCUa3WY8OFmVGh1aruH2GAvPDy6\nK+4f0QUrkjLx1C8HMeLdBKtfx5ZyYwCI9PPAzQMjACjnxLrk7DrPmbv8kHpTcf1lyk3o7V8n4rGf\nDjT69e1VrtXD2dGh0bJLQOmbNjJGyYTRS4lX1yQjp1SDqxdsUzcVJn60BQ/9sA/z/j6BB80Cic1R\nYZz0+sdDShmKLccqhECvzp3w2dYz2HEmH48uqzkWjU6Pz7efMZZKWv9vFuHrjqd/OYjrPtuO6V/s\nwAfG5tmmc8FBCLuvSbU9MqYbZgyJxLpHR+Pe4dF2vfY/V/VCjyAvdbOhKTvtpvfTRONG15+PjMbf\nj49BlJ8HUvKal/lnHoAz77HbUhN5dQZp93ChvHINtp7Kx/Yz+Zi+aCf+veqIzdOpTQGan/afR2ph\nhV3rph/2pqtDxBri4uSAn2YNBwCL99D6uLs4qkMFHIXA70ey8OGmU7hh0U6k1Po9D4lSvv+Z/Aoc\nzijG74cz1X5e1/QJQXJWaZPKzIsqtBZZyMGd3JDdQkNCyrV6NVsv3Me90c0j03X7tsGRFmt6VydH\nXN07BAkpeZifkAIpZbPW8aZsLyGEeg6Zc3FywOBI5fedkluGqZ9tR1JGCS7vFVznuSseGIEQb1fc\n8c1uHKyVnS2lxNe7zmHawh0WWa+h3m4I7uQKPw9n/HUsBy/+cbTO161oQjsVe8wYEonHx3ZDD2M2\n5Ohugern4kKUMvm1ydl4+Mf9TW5fYNqsaw4PFyeM7xFk0Yc7v1yLWd/usfg+M5fsbvFJwnOXJyGn\nTIPhXfyx/Uy+2tKhoXvSwspqNQglhEDXAE88tzzJooe8QcoWy+arPeSnIe9N76/+O9VsU3xfWiGm\nLbS/5LchuWWaBgfLmWx5ahz+enQMTuWVYcwHmzD0nY34etdZaHUGZJdWWW2vAFj2kA73dUepRoe8\nci12n2t4YzO/XGvThl4nN2c8OCoG3+1JxRtrjzUpuzqrVGNzUsn4HkGYZVxD/G3jxPSGVGj1GGS8\nhmnMNgzDfJXj+eqOwep1/7u7lc30Ob8dwmYb78fv/W4vZi/dh2d/S8JtgyPUx2u/H5veB8/kV2Bm\nIxtfHVm7ChJqbeiFMWOIMqlw9VGlhPfmL3fh+z11d/hsaQZv+qMyv8B/dttAvDglDv4eLo02yy4y\n27mxZlCkH96cWlNmV1JZjRlfJ7bK4A9AycSorx+hs6MDfpo1DAAwMiYA523IlDSpHQw1lQ4WVWib\nHaAoqdIh0NO1WaXGtXm5OuGVq3tjwxNj1MecHBwwfUA4+oR64+aB4erjzQ3c5pZpcNvgCDx3eaxd\nr3N1csRU47TID8zeQAM9XbHrXCHeXHcc9363FwXlWuSVKcNimtKE+UIwXZoziqtg71+DeUZH7Xuu\n+0YoQcIFW08jKaMY/1133K4hKXqD0h+usWCPiSnwZ34D/+nW0/hq5zks3K4ECc/kl+Pz7WfRNaD5\nAVtT9mzCyVysOZqF0qpqm7Nl5v19AkmZJXhpdTJ+sKEvkT3KtJZlEV/sOItJPZWblH5h3tiXZlu5\nqJJJWPN1IvyUQOOy/WlIPFeAn5tYZgAoA0vMg+YhnUwDgVomYGpLBmqknwfcXRzx+4NKAGrc/M04\nX1ylZvLNGtEF8dHKBoJpV9haBneF1r7yrmmXKdeNf606AkDJOjAFfI5mlaoLsHHdgzB3Us11qaSq\nGidzylosC7zIrKeYLboHearXsN8PZ+Kxnw6goEKL7FINVh9R+vqYN9A2bQA2p89duVYPT1cnBHdy\nxe65E21+XbiPssZ46peDSDxXiPh5G/D+hpM4kVOGhdvO4M/k7HpvQt+/8TKcN/YVM98waY09Qj8P\nF5uHcJhM7ReG/1zVC54ujvjr0dFN+r6BxpsdU4AxwNMF3m7O6BfmjeUHM/DiH0ew+1wB/jqmDGfY\ndjq/wT6PlVq9ug7T6GuCGubrDlPriYPpRVb76NlKb7B9cInJhzcNsPg4v1zbaN9Kc4vvGqKetybV\negOSs0rqvTnSG5ShIdf0CbHpe4R4u+G7u+PVa05jTKW03+9Ns2ibMndSrHqeLps1DHMn9VQHWfzn\nj6NIMttgfmpCD+SXa/HG2mPYYEeGu5QSBRXVFgMtAr1csC+tSL0p3nAip8mtKcq1OvV9PNzHHZkl\nVQ2u9Rqa0n5D/zB8MP0yeLo44pMtp/FQMzYw8so0CDAGMKb3D8ek2CCM7R5o8RzT33xBhVYtxQ22\nUmkjhICLsSz//qX7sPtcTeZR7Y04U9ZOj2Av/PHQKDXb969jOXUCXOVNaKdij6cn9MDMYdH4YsZg\nLJ89Ao+N7aZ+rou/J9Y/NgaHM0uwJ7UQQ9/Z2KQ1unmWanPcMywa3+9JU/8mT+WVISmzBEWV1ajU\n6vHvVUeQnFWK4e9ubNFAoakX9xW9gvHqmmRsO61cIxLP1Z9dVlShtXg/vrJXMJIySzDF2DoAAP44\nkoWx8ze10DHavgHv6+6MF67sCR93Z9z73V58v0fJFE3JLUdGcRXyyjQ4mlXS5InD5raeyrdpEdXj\neQAAIABJREFUo8TN2RHuLo4WAyE/3nwa+9OL8OxvSbhh0U6rrys329Q1D0wlnKw/wLbmaBbKtXqb\nK5BGdg3EFb2CsfxQBsZ8sAlnC8rV86BCq4NGp8epBjbjskqqEOJtW3Vej2AvPDymm8Vj1uY+SCmR\nV6bBcyuSGlyTmQZxArDoo+/p4oRPbx2I3qE1mdqxwcr7zOaUPMz57VCjAdHa3/eZCT0s1rjm/D1d\n8PuDIwEo8y7IunYVJCyqrG60hCO4k6vaS9AUZFl9JAtSSovdwgrjzYFtak70QZF+uK5vKB4cHYOf\n9qc3WDqTX6G1OrXL3OU9g7HpybEI83HDNZ9ux8ncMrz6Z3KrBApLNA03ue4S4IkNj49BvzBvZBZX\nYfmhDJt2Kco0lkGDibHBCPJywQ2LdtabHWPzMdswHbMphBB1skaev6Inwnzc8ajZomTV4UwYpGzy\njWh+uRbdA71w08CIxp9cyx1DovDo2K4YZdZrpneIZSBw8idbcbagAhF+Hhdtb4VnJvbAZ7cNRFFl\ntbq40RkMjZadGYxNvE2snUuvX9sbbs4O6g3i7V8n2two2BTcttYOwJoregZjxzPj8eFN/RFp7B25\n5qhlttYtX+5CXEgnXNmrs01fsyFDo/2w5alxyC7V4MU/jmLiR1vw6DKlF9rx7FJc8+k26AwG/PP3\nw/i/v46rr3N1crC44WzucBdzJ3JK8fyKw/BydYKDAOKNWbKmXX/zv9W5vx2qd2d7X5oyyaz2zUZe\nmQZ7Uovw6LIDeMvOMgMTKSXWH89GtH/NNHRP4wTQliplKdPobS7DDO7kqt5Ev7/xpMU1eGCELxbe\nPghPje+OroGeFrvoJuVaHTzt2CT59+Q49d8VWp3aAP8/q47A2VFYNG827egCwKSPtuC1P5UhS/Hz\nNuDXg+eb9fuypR+huTAfdyyZGY//3TwAw7v4W5SAmIY/AVBvJp5dnoRfD57H0Hc2Iq9MgyWJqTZf\np0uqqqHR6Zs0iRyA2j/U/L7v+71pFj1SuwZ61n4ZgJp+Y89O6mHxeEtn/DZHXIg3Ep4cB98mNPYH\nlJ/xs9sGIqxWtkJssBd+O5SBNUez8ciyA/jn70fwwsrDeOqXg2pzcWuuW7gdz69IQlW13qLcuPa1\n++31x3H/0n14vBmDTHQG+3oSArCoFpgUG4TYYC8MjfZv4BWW4kK8cXlsTTZY/LwNGPleAuatP4E5\nvx3CksRz2HQy1+ImOaukCr7uznZtoMYGd7I5w/Ef9Wxserk6mZV7e6JroCe+uVPJ+EgvqsTu1EK1\nTYyvuzPmTOyBFUmZeG7lYVRodTado6a+VeZBBn8PFyzacRYLjVn7z604jF8OnLfpZzGn0xtQrZPq\n5qC7iyMCPV0w7J2NeGPtMaulZ8qk5fo32Ed1DYSTg8D3e9KwP73IpnYbtXuTAkrbGNM6x9fDBf83\nrR/eveEyi+fc2D8cI2L81bXHrmcn1LuG+fiWAbgsTOmvZprmLKXE9EU1LV9MGeXmzL/emqNZWH4o\nA3qDxOfbzyC1sKLZmc62cHdxtBhIaeLj7mzRV+x4dinKNDoUV1bjSGaJTSW0JZUtc1/RJcATvTor\n17QrP96iZovtOJOPdceysdOsjYetZcG1PbB0L/6uFWCv0hmwZGY8ru0biomxQfh8+1kAwImcMpwr\nqMDiXeewrFYrq8IKy0EtNw2MwCJjJrbeoNwfn2pGn7vayrU6O+6xlb/r6/oq69b3N57EofPFmPe3\nsgb8+0QO7l6yB/+tpw+4PQ5lFOOavrZtrAA1gSzTNe1YdimO1dODtKpajz2phRYZlH89NgbLZ4/A\nppRcLElMtfravamFuHlguM33Isp1NVZtiXbzF7uwLjkb5woq8PCP+zHm/U247atEiwD6AbPrUnPb\nU13xvy11Hvtp/3lMWbANG07kWvQpr81UxZnw5Fg8aKwGMxkc5dfg+9OGEzn1nt/v/H3ColIozMcN\nQgjcNDAcKx8YYfU1Id5u+O91faA3SKw4lFHv9+3I2lWQEIBNzaAHR/lhxuBIxAQoi/T0okosSUzF\n9Z/vgM5YjlhuNim1IR9MvwyTetadNnh17xCUanSYunA74udtwN7Uwjop/dklGqs7fLV5uDjh9Wv7\nQGs8th1nCnD3t3vw9vrjuPObRKw2G2ySXljR5OEEtrwxdnJzVqeNvrH2mPoml1lcqQZ1smqVfFhr\nbl6mqQkuVmqVC6e9DWDVr93MsgB7ebo4YdVDIxHuozRsve7T7XUmhs76bg+W7UuH3lB/AFGrM6BK\np2/yYqRroCfuGdbF4rFQH3dseWqcxWMP/7i/1QavtIQQbzcMivTDwAhfdTLzrO/24rZaEx2llBZB\n6bwyLTxdnNAz2AsPjIzBR7WyNQBgclwI+oRY9ik0z/DV6Q319g9VAtC2/20JIeDk6KD0f5s9Ajvn\nTEC4WZ/Se41lJp/dOsjmN/vGvl/tG8DDmSXYejofdy7ejZxSDc4XVeGvYzkWN0vLZ4/AGLMSHQPQ\nYrvYfx7NRmW1EiDbMWcCnr+yJ4CagTv9jDckAJCQkoe9aYUo1+pQWlVtcZ6Ypg3XvhkfFGmZ5fLG\n2mOYn5BidVhPfZQsgyKLaeimJtrHskuxtpkT00wlOfbcKJl+zpxSjcX1QAiBgRG+uCM+Cl0DPOsM\nsSqpqkZSRkmTy7vGzd+MOcbeqn8mZyPIy9WiD0xns5K2cB83JJstYN9cdxwLtjS9PLuosv4b64YM\n6+KPl6bE1fv5KD8PvHBFT5wrqMCb65Tg+KaUPHy4KaXBxam5SR9twdzlSXB2dLA7awwAZsRH4sOb\n+mNMt0B0C/RUg+QAcGP/MOyeOxFhPnVvcAHlv/nuuRMxzNiD543r+gBouXP0YiCMfQlfmhKHj2+p\nuW6b/55qa6jMrbiyGgkpefj3qiPYeTbf4rq48oER2PHMeADKjYpJbplG/Z02FoBNLaxQM2h1etun\nG5vrG+qNB0bF4P+m9cPiu+Lx0tX1/w1bfX2YD0Z3C8Dr1/ZW31eSjEHBDzedwrPLk3DPt3twIL0I\n645l448jmVYHXLSUG/qHWQ0Uhvu44dZBERYbUX5mFTN6g1Q3jwBl4+iHe4diYIQvxs3fjA82pjT6\nvc8XVdYZxGbacEg8V9BoD8GGKNnDjhbv0aa/y+WHMnDLl7sw9bPtFu9XtqxD3Zwd1TX85E+2In7e\nBvxjeVK9z7/qk60WU10NUuLLnWcxtlvD080vC/fB29P6IcjLBeO6BzZ4Ux3i7aZmDOWUVSGtsALZ\npRoUGzMJ46P88K/Jveq8bs6kHuqU8lfWJOONtccw/N2NWLjtDAI9XWwq12xNfz4yGo8bN/NnLtmD\nCR9uxuX/24J7vt2Dx386gKySKquVXou2n8G20/koaoFyY5NRXQPxxY6zKKyoxte7lKGNL/5xFK+v\nPWaxSXbNp9uwbF86jmeX4u/jOTieXWpTae/+9OI62cSmII+DELi8ZzDOFVRgcKQvlu5Nw01f7MRH\nm0+pATaTwkqtxXkK1AzoGP7uRlz/+Q58Z8zgyynV4NU1yXh5dd1yc1vZeo9tzrSZ6ubsgJWHlaBN\n90BPvPP3SfQI8sKJnDLc+PkO/HE4E3llGjWBw9q95eM/HcDNX+zESbPqg6SMYugN0iKDrTFBxnv4\nu+KjAAD/23xK/Vzt4SEbjOW45sFRX3dnhPu6I9rfAx9uSsGjy/ZbXL+OZJbgbEEFRsQEwF4Pje6q\ntqbal1aEm77YiaNZpergsueWJ+GFlYdxNKsEs5fuw3sbT0JKiYziSruDhLUnA/91LBtrk7OwYMsp\n5JRqcNpsc+VMfrlFb3QppZrtaMri9nRxsnvt9dLqZLy6JhmLd53DjjP5SC+swH/XHoNWp8x3+HLn\nOYR6u2HLU+Pwy31Kew0HIRBaz1oMUBK1AOD1tcfwypqjDVYwVVXrsamBjNBL0UURJBRCXCWEOCaE\nOCGEeK6h5zY2lQ1QdjefnthD3cUu1+rxkfHE/ut4Ds4XVWJPWiGCbUg5HtU10GqfIjdnR/QO8VZ3\nhx76cT/uX7oPlVo9pi3cjrwyDZKzS9Az2LYS0H5hPkh4ciymGJtr5pRq8NP+8zieU4aXVh9Vd9Fv\nWLQTD/6wr0mZbXnlWoudpPqEeLthj3Hi5tzlSTiTX46pC3fgm8Rz+PNoFq77bLtFMKdUUw2vWjeD\nlWafn/FNIh7+cT8mfrTFrsWdTm+ARm+w+42mJXTu5Ib/Tu2LfqHeyCnT4K9jORZvREkZJVi04wyG\nv7sR3+1Js1omWFCh/L5bImBkzlrmgLONTdLb0hXGvjmdO7kiOasUaUWV2HIqT+0juOZoNsZ8sAkH\n0osgpcSx7FKE+7rh27uHYvaomHr7dZjekE3T/XanFiKrpAqZxZUY8V4Cpi7cYfUGXBku0PTdZEcH\nYbHAPmy8oWvpyX+3D4606CuWWVwTMKs9CXZSbBD8PV0wtlug+povd5zFv36vP0vHHtXGm21vN2VQ\nUZSfB1Y9NFK9Ljs7OuDFKXH4adYw3D0sGmmFFZj13V5M/GgLnvj5IM7ml2NfWiG+3HkWAOoMopo1\nogu6mwVdlh/KwLe7U7H1lO3lg1klVRjdNUANwgDA9IHheHlKHMq1evx71VG89ddxfLPL9insu84W\nqKU8ZRodnJ2EXdlGC28bhM9vV/qq1nfpvnlgOH49eB6VWj1O5pRBSontxubLkX71L3KssZYdAqBO\nZp95NuS/r4pTf++m/9foDHjtz2Q8/tMBi3NoZVIGskqqED9vQ73lXuuPZ9cbKGtMoJcrQr3drLZp\niPb3QA+zALCrk4O6ID1TYLlQNb9BqKrWY9vpfMz4OhGAshnX1PJzJwcHjIgJwHs3Xoal9wzF9/fU\nDPyw1pfVmig/D7w5tS+u6BmM8d0D1Z6Ml5L+Eb4WGXVxId7wdXfGOmMZ82VhPgj0dIGrkwM0OoP6\n36u0qlrdDDVf62xKycOZ/AqLIF6oj7vaq9p00wsAVy/Yhhf/OAKNTo8R7yYgs7gShzOKLcpev9+T\nBr1BqpsWlVq93YNLTL66cwhmj1QyIxwdhM0ZeyZerk54/8b+mBwXguUPKGVQPsYhFkDNTdrspfvw\nr9+P4PPtZy2avbc0IQSu7hOCF6fE4WfjDdfMoVHoH+GLK3p1tshYDvRyxe8PjsTkOGX9+vi47lhm\nbF8jhEC3QC/0Mw4dS0ixvMmytpY9nlOK2FoBZYPx+nM8p0ztU2qt7K0xpWblbia31+oBnVlShYzi\nmsmmtrROmDMxFq9e09visY0nc7FsXzq2nMpDmUaHlNwyFJoNusgu0eBYdikMUvkbdHNyrFNebI2b\nsyNWPzwa79TKMrTmGmN21rZT+bhx0U58tfMsfN2dsfD2Qfjk1oFWXzMyJgCT40LwxrV96nzurWn9\nWnxNay8vVyfMHBaNh0d3rfO5U3nluO6z7Xji54P482iW2qtv0fYz+GzbGTz1y0G8tPqoTQN/bDGq\nawCKK6sxupty/XY2W9M8MDLG4loy7+8TuHPxbjy/8jDuXLwbY+dvQvy8DYiftwGTPtpcpzem6T5r\n/fEcbDmlBArzyjSQUqobcKYpvTf0D1PPv9o0Oj1WHc6Cr3vdNfTdw5R14vQBNW2Wrvl0G34/nIk/\njmQhrbCiSVnu5Vrbqy1MpvQOwaNju6Kq2oAVh5QA+j+uUDai74xXztG0okq8vCYZUxZsw/j5m7Hm\naDYmfmSZ2fbIj/ux82wBzhZUYMY3ifhs62lcvWArZn23F2UanV1/v6br+E0DI/CWWXswQKlc2pNa\niKV705BVUqW2RbG2XvvPVXF4cUocSqp0+GhTCso0Osz4OhH3fLsHB88XW2zY2uPO+CjcFR+F34yZ\ncOZ/bwkpeVh/XMnCBJRM06HvbMSqw1l1svwbE+jlio1PjMUi4zr2n78fwb9XHcWXO8/hmk+3qSXv\ngBIPeWHlYWSVVEFvkLj960Tc/MUubDiRa8wwtf0eybydRu+QTth5tgAfbT6FJ34+iBsW7cRvhzIw\n6v2EmuP0dIGbs6PNQ4mEEOhnLHFedTgLixNTkVOqwdK9aeraNrO4EpVaPXLKNDZNjL+UtH7OeCOE\nEA4A/gdgEoAMALuFECuklHXyiu0NFk2IDcKmJ8fip/3nsfpoFrJLqiwa8f7nqro7aPa4Mz4KvUO8\n4erkoO7amPo5zF2ehMySKtw6yPoACms8XZzq3d36+0Qubh6oBCHOFlRg6d40zBgSVed589afQJ/Q\nTri6T2idzx3OKLZpAWLq5fX0hO54f2MK9hoDhtmlVWqpx6rDmUjKKMEr1/Q2lhtbHvdT47urDdnN\nA4NLEs9h6+l8rHxgJD7efApphRWYZ2WhY5ASB88Xw9vN9nLQltY7xBt3DY3GP1YkIbWwArd+tQuD\nIv3UxuKmrJX5CSmYn5CCr+4YjN6h3jiaWYLdqYXwdXdutNy8qWICPC3KYkpbYNhDaxsc6QcvVyd8\nML0/bjfeqD/z6yG4ODrgyzsG4yXjruXspftw7/Bo/Lg3XV3ANOSBkTHoH+aj7mQVVlTjus8smx0X\nVSr9jdTMT6GUEDV3N9m0I7nu0dG48uPGh3U0xTMTlbLE+0fEYPyHmyx2iN9YW3OZ/OHeoegWWHNj\nNWt4F2QUV2Ftcjb+PpGLn/enY+H2M1gxeyTcXRxRWlWNV/88hren9bXpHNMZDKgw7nybZ3TX7u9i\nKhmJ8nPHa3/WHN/OswW417hIA5R2EFdYKcv+fMZgOAglC84ko55s0EJjnx3z488v19bJQony80CU\nnwd83J3x9K+H8LMx8/KrnWex8Ymxjf78//njCAorqvHc5bFYsjsVVdX2LZrjzLJdy+rpKzko0g/R\nfh7qe8jDo7vir2PZmDWii909Wf81uRfCfNzwSa1MwNp9FE0/96IZg9E/3Aff3zMUB9OLEOrjjuTs\nEovJtBtP5qJvqLIx9tqfx9Ry2XXJ2egb5oNwY3mHydGsUrx4lX3ZVOZWPjgSGp0eb60/gc6dXNUs\n5EBPZePluctj8db6E+gX5q1uar26Jhn3j4zBdX1DcftXiXB1csCGJ8YiragCT/x0sMGJxE1l+plH\nxgQgyt/d5hsjUxYIAKvvgZciL1cn/PVYTS/gcd0DMdN4jf9+Txoe+nEfnBwc4OrkgAqtDs9MjEVR\npRJUmRQbhHMFFThbUFFvO4dIP3d8fvsgvLImGX8cycK6YzlYd0wJCk5dWFNiueKBETAYJN7feBKV\n1TqsTFJu7h5eth9HMkswwCzY2FbemtoXfp4u6OTqhEMZxWqPWnN9w1r3OD1dnNTrOdDw0JMQbzcM\njPDF8Wyl72ntDQlTxr/eIJGUUYwz+eXQ6Ax4e/0JxIV0gpODwOyRMRga7Y/Uwso6g9hmDIlEuK87\nXlmjtB6IDfbC0r1pGBDug96hSvD5jm92w9/DGc9OikWojxt2nyvEpJ7B0Oj0cHQQcHJwQKmVfnSD\nIv0wOa4z1poNe7r+8x2I9HXHY+O6KVnRjQQJJ8QqGYALtpy2uM78b/Mpi03zUG833GQMyCzcfgYL\nt5/BezdehtKqanVwXksaGROAedf3w9zlSYj298CvBzMwc2iUGlxqyMSeQcAqy8cC2ziL0Nzdw6Ix\npXdni3Pb3H+M93vrj+dYVAMZpPX2NU1hqliLj/LHqJgAGCTUNVqXAA/cPSwaX+w4qz7/mj4huOGy\nMNy/dJ/F1ymp0uGJnw9i3vX9sPFkDu4bEYNdxkourd6AZ349hGcm9oCniyNGdQ1Q33cCvVwxvIs/\n4qP8MTkuBA+P7orrjUH0rafyMCTKD5OMQbTa7YoA4LGx3VBaVY1bB0VgQLiP+jszeWTZfmSVaHBN\nnxAcSC/Cdf1C0T/cF4MifVFVrcfa5Gzc0D/c4jVKT1Gt3evrmABPxAR44se96cgr1+Kvx8bA190Z\nr13TGxNigyzajgBKMorpdxQ/bwN2z52I/HItdqdaDglZfTRLLbdfMjPermMa2z1QfQ+KMmZuv3vD\nZWqFxsM/Kj1ICyu02HWuEPcMi7aaIdc9yAvdAj3h5+6M1/5MxoQPN1t83pakJWuCvFzx+LhuGNc9\nEAFergj0dMFNX+xEfrkWAyN8EeHrrgYQze8Zezahh72XqxP6R/hi7qTYOpmq1gZSXvfZdvQN9VaH\nipVWVePjzafVVnC2ePnq3vjjSBau6BWMN67tg7fXn8CN/cMx45tEq8/XNyGB6os7BmPoOxsRH+Wn\nlKN/qvTp9PdwRv9wX0xduAPX9Q3FDf3DWqSXaXsimtP0u0UOQIjhAF6SUk4xfvw8ACmlfKvW8+TG\nEzkY36PhVPyGPPXLQXUs9gtX9sQ1fUJsmmbYmLwyjUXzV3PbnxlvU/ajSVGFFoczS1BcqZSZbT2d\nh+xSDfqFeaN3iDd+NPaZmNAjCK9d2xv/XXsczk4C6YWVcHVyxPYz+YgJ8MCyWcMhpcTu1EI8uuwA\nds+diFu+3Ik3ru1rkYXREL1B4s11x5B4rhCZJVUYFu2HXbUmNO2cMwHD392IO+Oj8OR4ywyKBVtP\nQwAWb5AmKx4YgZdXH8X+9GJsN5YK5ZRq1B4kRzJLcM+3exDq7YaVxuaibaVMo6tzQbfm3uHR2J9W\nhAPna6YOD4r0xWctOJnZRKc3QAiBvHIN5iek4Kq4EJsCwBcDvUFi+Lsb4ebsoAZbeod0wtGsur06\nEp4ca3NZ56HzxdibVohgL1e8vCa5zucHR/pir9lQjYdGx+BMfgVet7Jbbg9T2fKuswU4kVOKu4ba\nN03UHqfzynGrWZl2mI8bMoqrcPvgSDw1oXud7JXtZ/LxpJX+XG9O7Ys1R7OwOSUPvz84EtmlGhzL\nLsHNAyPgIASklNDqDdiUkqfekD/960E1o2/m0KhGM6b2pRWqU2j7hnqrmZaAUkLy8tW9G1yo1B6g\nMTDCB8O6+GNavzBodAaE+7orJV2Xx+LmgRFINU6b+z0pE27ODupgm9pyy5SpuSbT+4cjws8d8xNS\nsPSeocqkcH8PPPLjfmx4YiwAYPT7CRZZZ9MHhON54w63Pf6z6gim9AmpN2usoFxbZzL0I2O64t7h\nXez+XoDSnsK8wfb0/uFqibiJvp4ebGcLynHzF8rf2otT4vCqlXPK3KLbByGrtAq/HDiPfmE+WJyY\natf5W5/4eRuwbNYwdXfetABfmZSB1/48hjUPj6rz/vvgqJg6LSLqY8/AEmo56YUVCPVxV//2KrQ6\njJu/GT2CvHCyVhnXgHAffD5jMEqqquEghNVA7PHsUoT71gRpq/UGTF+0s05guJOrE0o1Org4Oqjl\nobUpPZPGWf1cW/p48ylE+Xuo5+K6R0fbVB3SEnJKNQj0cmk0Q9IgpdXnGKREaZUO13y6Df4eLvUG\n7Id38cfOswV45eo4q5vdVdV6bE7Jg5+HMx5ZVv9U9kmxQfj7RC6GRvsh8VwhhkT54ZNbBiDxXCEW\nJ57Dx7dYZtHllWmw4URunZtfE1uvEwYpMeydjbgszAen8srU/ooNuTM+CuUaHaL9PXBHfN3N/+Y6\ndL4Y932/F5ueHItD54vRPcjLatDZmkxjaeKKpEy8sfYYtj09vtHe8BeaqRXSp1tPY8luZVClo4Oo\nU0Hi6eKI+Cg/JKTk4V+Te9WbcW+vV9ckY8aQSHQP8jJOuAZ+OXAeU/uFws3ZEdMWbkdGcRViAjzw\n1IQeGBkTgJmLd1u096htclxnRPt7IKO4EuE+7ur72Y39wxAT4InbBtefgPLljrNYsNVyg3ByXGeb\n1rrJWSWYuWRPo88L83FDTqkGOoPEPcOiEe3vgdwyDcZ0C8SSxFScyS/HYjsDciYpuWVwd67bk3L7\nmXx4uTjhvu+V3r+Rvu5IM0tCiY/yswgQrnl4FF5Zk2zRG7I57/canR5P/nwQ79xwWb33hHMm9mjw\nv43p+lBbYgO9RZuiWm+AgLKJuSe1EPnlWry0+iimDwjH0xO6Nzv2YZASeWVaNaBmLsLXHXqDrPca\n/9FN/THcjsqJvDINvFydLDbLR7y7ETorFWIvT4lTs6ftET9vA56e0B2pBZX45eB5+Lg7o7iyGuN7\nBCGntArpRZV47oqe+ONwFubf1L/xL3gBCeWerVWyqS6GIOF0AJOllA8YP74TwFAp5RO1nif3pRXa\ntPtVn5TcMiRlFGPzqTy8d8NlLXpCSimxdG8a3t+Ygst7BqvlTy1xA1JUoUV6USXu/W6v2oT4UEZx\ng68ZEuWLEG83rDqs9LhafNcQzF66D38/Psaui8OKQxl4fW3jzWJ/mz1C7UtW27pj2dDqDFiZlIH9\n6TXH7eQgoDNI/O/mAXjsJ2Wx98S47gj1ccPqI1lqev3FcBOXklumZr+Ze2CUEohYuO2M1UBX10BP\n/HjvsAtyjO1RQ5NUbQlEWZNXpsG7G05i/fEcRPt7oKpar2Yh1XbzwHD843L7gz1tKa9Mg9u+2oWJ\nscG4cUA47lq8u94FRmphBe76ZjfenNYXT/58EA7CctiCNasfHmURRJsc1xnrj1tOOXz+ip4W5SnW\nSCnx1voT+OXAeax7dDRySjW4c/Fum2/ACyu0eOuv4/j7RN0eIKHeboj0c0fiuUL0DPbCW9P64Y5v\nEtUbsv9c1QtT+9V/A/Dz/nRUG5TeY2+bDUiZMTgS3+9Nw7R+oVhhzC767f7hdSbZfTszvkk7sbaY\n+tl2i8XVhsfHNKsva/y8DRhu7PPn7eZs8w2elBJHs0pRVa1HtL9HvRthDWmJa3daYQUi/er2Xksv\nrMAb645hwa2DED9vg3pTOLprALaezsdLU+LUrCPza3Ptc+BieH8hRblWh2q9xNI9qUgrqsRfxizA\nmwaE47kmBOV3nyvAnN+S8OCoGHQP8oKTg8BDPzY+ffbjWwbYNXjkQjNtFrX0jeWFcMuXu9SMlrmT\nYrH1dB52nCnA6odH4bkVSep05C9mDMZlDWR0llRVY976E/gzObve51gT5OWC/uG+eLPRlyL7AAAg\nAElEQVRW6aBJ7Q0hQKlWmD3K+qaTNfHzNuCNa/tgaBd/iyb/82/qX2fTLtjLFTnGZvyLbh+E/s24\nx6mPTm/AzrMFGN2t6RvJa5Oz8O9VR9vF9VJKid8PZ1pUMgDKJtmM+EhMX7QTn902sE4P5Nby1C8H\ncSKnFKsfrpkWX6nV47mVSdhxpv6JxEBN0OnzbUrWKQB8feeQRrNO5/52CAlmvQx3zBlvcw+4g+eL\ncf/3e/H42G5qq6761o5BXi5qlp6Ji6MD1j02utUG3Kw5moUX/ziKCT2CsLGeHnGmc6l2kkdL/f3G\nz9uAT24ZgHBfd5zOK8fTvyqZhf+8smedzMradp0tQHJ2CY5klKCyWo/Xr+3T5CFhtpJSYug7G3Hv\n8Gg8UmtacXO8/mcyViRlIsrPHamFlXhgZAzuG9kFW07lIadUg7fXn8D3dw9FQkquWonYEv8NZn23\nB0kZJRgc6YvT+eUorKhu1tctM/bQNGUiOgqBocZg7qzh0fhxXzrKtXpMjA3CW9P6Nfv4W1JrBgnb\nvNzYHks+mocVxlTP8ePHY/z48Xa9vnuQF7oHeTV6AjeFEAIzhkTh1kGROJxZghBvN9wysGW+j6+H\ni3oB8XJ1wjV9QxoNEu5JtexVZtoZsnf3wNQnYfqAcHU4Qp9Qb3XKngDw+rV96g0QAlAzkU7mlGF/\nejGi/T1wrqACOoNE75BOaoAQAD7cVNPU2tPFUZ1U3da6Bnri/hFdsGjHWcyZ2APvbjiJK3oFY/bI\nGEgpcfvgSExZUJMB1KtzJ+gN0mpPLaqx/Znx2HGmAC6OAuG+7rhx0U706twJx7JLrfabsUWglyve\nnNoXd2QUw9fDRR0o8vuDIxHcyRXrj+XgX6uOAECLNa++kAK9XLHqoVFwclTKpxKerL9cNsrPA5uM\ng262PzMeSxJT1R3mEG9XxHX2VhdZphL2q2sFg8xLsAZG+OLdG/rZVFIphEAXY3mGr7sz/DxcEO7j\nVm/2Tm1+Hi544cpeuH9kDA5lFKtDKgCgSqdHojGr+XhOGZ7+9SBcnRzVIGFj2TWmaeMGKbEyKRPH\nsktxZa9grD2m/KymACEA3LBoJ/w9XPCvyb3wza5z9Zb+tZQ3p/bF2fxyeLk6ITrAo9mDm4ZG++Ha\nviF2H7MQwuJG5PbBkVi6N63R102MDcKGE7mYO6llrn3WAoQAEOHngQW3KlnaEb7uuCquM+6Ij4Kz\no0BOqQYRvu4oqqzGlN4hCPB0wcH0Irg4OaCwshpP/nwQj4/thuWcaHdRMd1UPjymGzaeyEVWcRVe\nv64P/Jt4AxUf7Y/NZoO+TL21ruwVrJYh+7g7Y/1jYzDnt0PqgICLOUAIKOuRHXPGt7sAIQBE+7kj\ns6QSI2MCMLZ7IEbE+CO9qBJBXq549ere6oZMl4CGh7J4uznjtWv74LVr+0BKiTKNDlq9xIItpyyu\n34Ayqdm0GZRbprUYsFXb13cOwZn8cnQP8oKPmzOmLNiK+0Z2setnXDZrGKL8PODoILDxibF4f+NJ\nbE7Jw8iYAMydFIsoP3e8tPooCiqq8eT47up6pJeVctCW4OTo0KwAIQCM7RaEDy+yDJr6CCEwtV8Y\negR5YeaSPRAAdj47AQ5CQGcwYMbgSPQNvXAtBV68Kg6142vuLo548ao4rDqSiY83n8bb0/rhHyuU\n9h5X9grG8JgAvLomWW2dMntUDBZuP4POnVxtKkt/7do+GPPBJvVje4ZE9Avzxv9N7YtBkb7YcDIX\n/7qyF5ydBLr4e2JvaiEe+nE/JsYGoV+YD/qGemN2rdJprd7QqhOwTV97zqQeeHB0DG77SkngGN89\nEAkpefjqjsFqKwZ7+yLaauecCWoWfJiPO67pE4KBEb6Y0rvxycnDuvhjWBd/SCkv2DXc9H1qDxlt\nrn9fFYereocg3McNGSVV6NW5ExyEwLjuStXn8C7+iPTzsDpIqDn+d7MyEM3DxQkpuWVqy5mmMv2d\nOJn993hodAw+3XoGwZ2UYV1f7jxn83C81pSQkICEhIQL8r0uhkzC4QBellJeZfy43nLjtj7WtrYp\nJRfRfh4orKzGA0v3qZl4JgGeLuoglbjOnSxS2a/rG4q7hkap/TNslVemwf/9dRz/N60vTuSUwWCQ\n8HV3houTI4K87BvKUVWtR2GFFi+tTlYHLkzvH45fDirBx8vCfCyCn4+P7ab2K7qY6AwG5JZq6kxM\nGvVeArR6AxKfnQAA7XIB39bWJmchLsS7Tj+i5vjn74dRrtFbpIibMhibU8rZXu0+V4BFO87is9sG\nQUqJ4zlluGvxbnwxYzBe+zMZZwsq1OeOiPHHwfRifHHHYBRXVmNwlH0770cyS/CfVUfw6+wRAJTs\nQJ1BqhPj7GH6bzYg3AcHzhcj1NgDa/VRJVt6clxnzB4Zg9KqavQK6WTzovh8USWu/3yHmnn29ITu\n2J9ejFsHRaB7oCeu+HgrFt81BHEh3moGzy7jDUdHU6034FReOTTVerWn0uBIX9w6KBKrj2RCJyVe\nvbo3dAZ5wcog7VWu1eHDhBS8cGXzehJT+xQ/bwOeGt8dk+M6w0EIuDs7wt3FUR2WlVOqwbhmtLWh\nhlVq9XBwqH/DWkqJrJKqBidSNsQgJXafK8R3e1Ixa3gX6KXEZWE+GPleAmYOjcLixNQLWqZdH1O2\n1u65E1Gp1eNsQblF71pqGUsSU+Hq5IBbBkW09aFYlV5UiWd/O4SPbxmIXw+ex8JtZzBjcCSujOuM\ne77dg+/ujkescQBm/LwN6BnshW/vHtrIV1WYSvMhUG8fV3uZ2muZWmmZKg76hHpDo9Pj7+O5MEiJ\na5tQ8mmrtMIKPPbTAawwDniauXg3bhkUgS7+Hrj3u711Mqzj522AoxD4xxWxuLEVkoTai8MZxYgN\n7tQm7QIqtXo8+tN+FFfq8Mv9wy/492+qkzlliAnwgJOjA4a/sxETewbhv9dZz0JvK5d6ubEjgONQ\nBpdkAkgEcLuUMrnW8zp8kNAktbAC0xftxOYnxyGnTIPs0ip0C/SCj5sTViZl4s2/jmPBrQORlFGM\nT7acxs0Dw/HEuO52N79vLQXlWhRXVUOrMyA22Au/HcxAcCdXdafT1JfL1GusvRj+zkbopWwX5Rgd\nTe0dO1O/tZZOvW+PTOUYfz02Bnd+k4jsUg2+v3so7lyciLuGRuOxsRfH7+euxbsR4u2Gf1wei6sX\nbMMDo2Jw84BwvPD7YexJLbKpBKc+OoMBx7JKce93e9WAoEntnn1ZJVUIsXMy3KVo5uLdmBAbpAbZ\nTe/P3Byhi9mZ/HJE+Lrb1Sua2r/tZ/LRP9wHSxJT8VATqxSIWtuSxFQM7+KPoE6uuOJ/W7DxibFq\nllP8vA3oE+qNr+8c0qbHWF/f0YtBpVYP91pDThPPFSDKz4PrtjZm6qVvrQd2e1BUWQ03J4eLJpZi\nckkHCQFACHEVgPkAHAB8IaX8PyvPYZDQTHpRpdUSX4OUmPPrIbx3o9JzMX7eBrx+bW9Mjms8Bfpi\nkpJbhi7+HjaPMb8YrEvOhl5Km9LNqe19suUUrusbWm85Y0dSVa2Hm7Mjtp3OQ2FFNa7tG4qqaj2c\nHR0uyjf0s/nlCPF2U9+sWyJwpzMYsPFELi7vGcxAFxEREbUZ07rM5PW1yRjeJQCX9wxuw6MioovJ\nJR8ktAWDhE2TWliBCF/3i3bXh4iIiIiIiIiIbMMgIRgkJCIiIiIiIiKijq01g4Ttp5aTiIiIiIiI\niIiIWgWDhERERERERERERB0cg4REREREREREREQdHIOEREREREREREREHRyDhERERERERERERB0c\ng4REREREREREREQdHIOEREREREREREREHRyDhERERERERERERB1cmwcJhRA3CSEOCyH0QohBbX08\nREREREREREREHU2bBwkBJAG4AcCmtj4QIrpwEhIS2voQiKgF8ZwmurTwnCa6tPCcJiJbtHmQUEp5\nXEp5EoBo62MhoguHCxWiSwvPaaJLC89poksLz2kiskWbBwmJiIiIiIiIiIiobTldiG8ihPgLQGfz\nhwBIAP+SUv5+IY6BiIiIiIiIiIiIrBNSyrY+BgCAEGIjgDlSyn31fP7iOFAiIiIiIiIiIqI2IqVs\nlZZ9FyST0A71/pCt9QsgIiIiIiIiIiLq6Nq8J6EQ4nohRBqA4QBWCSHWtPUxERERERERERERdSQX\nTbkxERERERERERERtY02zyS0hRDiKiHEMSHECSHEc219PERUlxAiQgixQQhxRAiRJIR4wvi4nxBi\nnRDiuBBirRDCx+w1LwghTgohkoUQV5o9PkgIcch4zn/QFj8PESmEEA5CiH1CiJXGj3lOE7VTQggf\nIcRPxnP0iBBiGM9povbLeI4eMZ6P3wkhXHhOE7UfQogvhBDZQohDZo+12DlsvCb8YHzNDiFEVGPH\ndNEHCYUQDgD+B2AygD4AbhdC9GrboyIiK3QAnpFS9gEwAsCjxnP1eQDrpZQ9AWwA8AIACCF6A7gF\nQByAKQA+EUKYeo8uAHCflDIWQKwQYvKF/VGIyMyTAI6afcxzmqj9mg9gtZQyDkB/AMfAc5qoXRJC\nRAOYDWCglPIyKPMGbgfPaaL25CsosS5zLXkO3wegQErZA8AHAN5u7IAu+iAhgKEATkopz0kpqwH8\nAGBaGx8TEdUipcySUh4w/rsMQDKACCjn6zfGp30D4Hrjv6cC+EFKqZNSngVwEsBQIUQIgE5Syt3G\n5y02ew0RXUBCiAgAVwNYZPYwz2midkgI4Q1gjJTyKwAwnqvF4DlN1F6VANAC8BRCOAFwB3AePKeJ\n2g0p5VYAhbUebslz2Pxr/QxgUmPH1B6ChOEA0sw+Tjc+RkQXKSFEFwADAOwE0FlKmQ0ogUQAwcan\n1T63zxsfC4dynpvwnCdqO+8DmAvAvIExz2mi9ikGQJ4Q4itjC4GFQggP8JwmapeklIUA3gWQCuX8\nLJZSrgfPaaL2LrgFz2H1NVJKPYAiIYR/Q9+8PQQJiagdEUJ4QdmleNKYUVh7OhKnJRG1A0KIawBk\nGzOERQNP5TlN1D44ARgE4GMp5SAA5VBKmvg+TdQOCSG6AngaQDSAMCgZhXeA5zTRpaYlz+GG1vQA\n2keQ8DwA8+aKEcbHiOgiYyx1+BnAEinlCuPD2UKIzsbPhwDIMT5+HkCk2ctN53Z9jxPRhTUKwFQh\nxGkASwFMFEIsAZDFc5qoXUoHkCal3GP8+BcoQUO+TxO1T0MAbJNSFhgzhH4DMBI8p4nau5Y8h9XP\nCSEcAXhLKQsa+ubtIUi4G0B3IUS0EMIFwG0AVrbxMRGRdV8COCqlnG/22EoA9xj/fTeAFWaP32ac\nuBQDoDuARGNKdbEQYqixEetMs9cQ0QUipfynlDJKStkVynvvBinlXQB+B89ponbHWLqUJoSINT40\nCcAR8H2aqL06DmC4EMLNeC5OgjJojOc0UfsiYJnh15Ln8Erj1wCAm6EMQmmQUzN+kAtCSqkXQjwG\nYB2UoOYXUsrkNj4sIqpFCDEKwB0AkoQQ+6GkRf8TwFsAlgkhZgE4B2UiE6SUR4UQy6AsZqoBPCKl\nNKVSPwrgawBuUKYw/nkhfxYiatD/gec0UXv1BIDvhBDOAE4DuBeAI3hOE7U7UsqDQojFAPYC0APY\nD2AhgE7gOU3ULgghvgcwHkCAECIVwEtQ1to/tdA5/AWAJUKIkwDyoWz8N3xMNV+TiIiIiIiIiIiI\nOqL2UG5MRERERERERERErYhBQiIiIiIiIiIiog6OQUIiIiIiIiIiIqIOjkFCIiIiIiIiIiKiDo5B\nQiIiIiIiIiIiog6OQUIiIiIiIiIiIqIOjkFCIiIiIiIiIiKiDo5BQiIiIiIiIiIiog6OQUIiIiIi\nIiIiIqIOjkFCIiIiIiIiIiKiDo5BQiIiIiIiIiIiog6uVYOEQogIIcQGIcQRIUSSEOKJep73oRDi\npBDigBBiQGseExEREREREREREVlyauWvrwPwjJTygBDCC8BeIcQ6KeUx0xOEEFMAdJNS9hBCDAPw\nKYDhrXxcREREREREREREZNSqmYRSyiwp5QHjv8sAJAMIr/W0aQAWG5+zC4CPEKJzax4XERERERER\nERER1bhgPQmFEF0ADACwq9anwgGkmX18HnUDiURERERERERERNRKWrvcGABgLDX+GcCTxozCpnwN\n2bJHRURERERERERE1L5IKUVrfN1WDxIKIZygBAiXSClXWHnKeQCRZh9HGB+rQ0rGCal9EkJg69at\nGDVqVFsfSoOEELj66qvxxx9/tPr3evnll/Hyyy+3+vchogujvZ3TQgi89957ePrpp9v6UCwIIfDQ\nQw9hwYIFbX0o1E4IIXDo0CH069evRb9uezunqXXl5uYiODjY7vux0NBQZGVl8T7uItAW5/SwYcOQ\nmJjI//5ELUyIVokPArgw5cZfAjgqpZxfz+dXApgJAEKI4QCKpJTZF+C4iIiIiIiIqJUwOERE1L60\naiahEGIUgDsAJAkh9gOQAP4JIBqAlFIulFKuFkJcLYRIAVAO4N7WPCYialhr7koQERERUfvD9SER\nUcfQqkFCKeU2AI42PO+x1jwOIrr4jB8/vq0PgYhaEM9poksLz2miSwvPaSKyxQWbbkxEZI4LFaJL\nC89poksLz2lqCSw3vnjwnCYiWzBISEQWWE5CREREROa4PiQi6hgYJCQiIiIiIiIiIurgGCQkIiIi\nIiIiIiLq4BgkJCIiIiIiohbHnoRERO0Lg4REF0h7WSSx5wwRERERmWvq+rC9rH+JiEjBICERERER\nEREREVEHxyAh0QXCDD0iIiIiIiIiulgxSEhEFhjMJCIiIiJzXB8SEXUMDBISERERERERERF1cAwS\nEhERERERERERdXAMEhIREREREVGL43RjIqL2hUFCIrLAnjNEREREZI7rQyKijoFBQiIiIiIiIiIi\nog6OQUKiC4TlFkRERETUkXD9S0TUvjBISERERERERI1i0I+I6NLGICHRBdJeerm0l+MkIiIiogvD\nFBxkkJCI6NLGICEREREREREREVEHxyAhERERERERNcreTEJmHhIRtS+tGiQUQnwhhMgWQhyq5/Pj\nhBBFQoh9xv/9uzWPh4gax3JjIiIiIiIioo7HqZW//lcAPgKwuIHnbJZSTm3l4yAiIiIiIqImYE9C\nIqKOoVUzCaWUWwEUNvI0pi1Rh8BFFRERERF1JFz/EhG1LxdDT8IRQogDQog/hBC92/pgiIiIiIiI\nqC4G/YiILm2tXW7cmL0AoqSUFUKIKQCWA4it78kvv/yy+u/x48dj/PjxrX18RB0OexISERERUUtg\nUJGIqPkSEhKQkJBwQb5XmwYJpZRlZv9eI4T4RAjhL6UssPZ88yAhERERERERtT72JCQiaju1k+Re\neeWVVvteF6LcWKCevoNCiM5m/x4KQNQXICRq75ihR0RERETtGYOERESXtlbNJBRCfA9gPIAAIUQq\ngJcAuACQUsqFAG4SQjwMoBpAJYBbW/N4iKhxDGYSERERERERdTytGiSUUs5o5PMfA/i4NY+BiIiI\niIiImo7lxkREHcPFMN2YqEPgooqIiIiIOhKuf4mI2hcGCYmIiIiIiKhRDPoREV3aGCQkIgvsSUhE\nRERELYFBRSKi9oVBQiIiIiIiIqoXexISEXUMDBISERERERERERF1cAwSEhERERERUaOYSUhEdGlj\nkJDoAmkvvf7ay3ESERER0YXBcmMioo6BQUIiIiIiIiIiIqIOjkFCIiIiIiIiapS9mYTMPCQial8Y\nJCS6QNrLIonlxkRERETUEtrL+peIiBQMEhIREREREVG92JOQiKhjYJCQiIiIiIiIiIiog2OQkIiI\niIiIiBrFTEIioksbg4REZIE9CYmoo+D1ji4V/FumixWDikRE7QuDhERERNQh8eaVLhX8W6bWxp6E\nREQdA4OERBcId/mJiIiIqD1jkJCI6NLGICHRBdJeFlUMZhIRERFRS/j/9u4+Wo76vu/45yMEJoSH\nGCeWgwSKn7CBNibYwRCn5VLq8uAEtYkdg2mJsePiBIpTt44fTs5BitPm4Zz6gRrs0AIFjgl27DYo\nAaccim9dE5fwJANCAhEKXF2kK5B0L7oSkq6kb//Y3au9q71792Fmd34779c5OrozOzvz3Zn5/Wbm\nu7/fb1O5/wUAVJAkBAAAAADMi+7GAFAOJAkBAAAAAACAkss1SWj7JtsTth9vscx1tjfYXmP79Dzj\nAQAAAAB0h5aEADDc2koS2n5dO/OauEXS+S3We6Gkt0bE2yVdKekb7cQDID+MSQigLKjvMCw4l1FU\nJBUBIC3ttiT8UZvz5oiIH0ra3mKRFZJuqy77oKTjbC9pMyYgCdwcAUAxUT9jWHAuI2+MSQgA5dAy\nSWj7TbbfLeknbJ9R929E0lEZbH+ppLG66fHqvKbGxsZ04MABRYR27typ6enprje8f/9+PfHEE9q2\nbVvL5WZmZrRnz56uthERGh8f1wsvvNDV+7O2b98+bdy4UVNTU3r11VczW+/evXs1PT2tvXv3ZrbO\nmtoxnpyc1Msvv6zt27frwIED2r59u/bv36+JiQm9/PLLh7xv3759mpyczDyeTtWfoxGh6elpjY2N\ndbT/2znPp6enNTExobGxMW3durXpMrXzcevWrZqamtLWrVu1du1abd68WS+88IL27dvXdkyDsHv3\n7sxj3LRpk3bu3Klt27Zp48aNmp6e7nkbteP82muvaf/+/ZqentbU1JR27do1W5fs2rVL+/fvz+Ij\n9N2BAwe0c+fOOfO6qYv37NmTS50hVcr/a6+9lsu6Jenll1/W9PS0XnrpJUWEXnnllbbPm0731dTU\nlLZv3z5n/TMzM3rppZc6Wk+32+/F9u3btWXLFj399NNdX0f7Zdu2bZqZmdHY2FjPZXP//v0aHx/v\nqJw3Hpfx8fHZf/Ody9PT0y0f1vfu3asNGzbMXg8by+1Cdu3apS1btmjjxo1z5s/MzGj37t0drSsr\ne/fuXfC+rRcRMXstkCrHodP9JklbtmzRgQMHZqe3bt2aS2JlfHy8o/N17969euaZZwp//dmzZ4/2\n7NnT1b5fyM6dOzs6h1977TXt27dPe/bs0fbtrdo99G7r1q2amZmZnZ6enm6rzt66dWtXdfuOHTs6\nug7X6qKpqanZZ7LafXizc6ox/k6vy+1+/k5t3bpVk5OTmpiY0PPPPz97D9jM7t275xyTQevmubR2\nX9qv6//MzIy2b9+uZ599VgcOHOjrfUc7duzYoZ07d2pqairTe8W9e/dqbGys47p+3759Ghsb0/j4\n+GyeI8999tprr2l8fFxTU1PasmWLdu3aldu26tXyN0hERMz7T9JvSvq+pB2S7q/7d5ekf9HqvXXr\nWC7p8Xle+ytJv1Q3fZ+kM+ZZNiTFueeeG7/yK78StelufehDH2prHRdeeGEsX768q218+ctfnt3G\n6tWru1pHlq644orZeJYtW5bZen/5l385JMUv/uIvZrbOiIixsbHZ41OLW1LccMMNISmuueaaeY/h\ntddeG0cddVSm8XRDUvzgBz8ISfEHf/AHs/Eee+yxbb3/nnvuaes8r98/8y0/Ojp6yHL1/2pl4tJL\nL+3oM/bLscceG5dddllm69u9e3dIivPOO2/OfvjEJz7R03rvuOOO2XVdfvnlISmWLFkSkuKUU06J\niMrx+tznPpfFx+i76667bs459uSTT3ZVF59yyikxMjKSZWizfuu3fisOP/zwXNYdMbe8ffGLX5w9\n1gvZtGlTx/uqtp2PfvSjs/M++MEPdn39kxRr1qzp6r2d2LJlS1v10iBJii996Uuzfx9//PEhKb7+\n9a/3tN7aOdFuOX/00Ufn7J9m9fN88X/3u9+dd72/+7u/O6eOO/vsszv6HPXb37Vr1+z8X//1X483\nvvGNHa0rK5/5zGdyPZcee+yx2c+8Z8+ekBSnnnpqx+uRFHfdddec6dHR0SxDnV3vqlWrQlI8/vjj\nCy7/Mz/zM23XV4P0rne9K7d6o7beI488sq3lbccnP/nJ3Ouxl19+OSTFihUrZue1u81uY3vnO98Z\nH/vYx9pe/j3vec+ceuH000+f/fumm26as2ztHktSvO51r+sqxrzPgcZ/d9999yHLLl26NC6++OLM\nY+jWBz7wgTjppJM6es+3vvWtvl6Hf+3Xfm12e1/60pcKdw9w+umnxznnnJN5XG94wxualoWFfP7z\nn5+N5Q//8A/jtttuy3V/tXuPkbWbb765UOdBir7//e/HtddeO/uvuj8XzMd182+xWoiIWyXdavvf\nVU+i2oAnIel426dHxJpW61jAuKQT66aXVefN68QTT9QJJ5ygv/7rv+5hs9Jjjz3W9nKbN2/uahvP\nPPPM7N/dtvrI0hNPPDH793zfmHXj4YcfliQ99NBDma1TqnzT08yLL74oSXrqqafmfe/TTz/dt29G\nFjIxMSFJc86jdlsSZnneLNSycs2aSlEu6rhGr7766pxzuFe1llmPPPLInPlr167tab2181M6WM/U\nzoF169bNvlZfP6TkueeemzPdbauedevWaXy8ZXXftSeffLJv3/zX9seTTz654LK9fINav/4f/vCH\nXa9H6v6YdaIo9W8navul1/1TX7bbKeevvPJK19tqVYaef/55SQfruB/9aMFRYuZV30roxz/+sbZs\n2dL1unqRd8+M+hY6tZaAre41WmksA1NTU90H1sKGDRvaXrbW8yLLa2kefvzjH+e+jXZbEkZE1+dA\nJ2otmvrx2WvWr1+vww8/vO3la/f7NbX7RunQe8yi905pZtOmTYfMGx8f12GHHTaAaJp77LHHOn42\nGBsbW3ihDNWfF3//93/f1223Y82aNTrqqCw6RM5V68nV6XWqvn559tlnM42pSGr3JOjeyMiIRkZG\nZqdXrVqV27baHZPw3ZI+KelnJZ2gyo+MXCDpv9j+vQXeax1MLjZaLelySbJ9lrQP3D4AACAASURB\nVKTJiJhotbLIqLtGVusB2sU5BwwfyjXKoCznef3nHObPPMyfDQAA9KZlS8I6y1TpBjwtSbavlXS3\npH8s6RFJf9rsTbbvkDQi6Q22X5R0raQjVGkaeWNE3GP7ItvPStop6YqFAiFJiNRwrgHDi/KNMijL\neZ5SkrAWXzdxFv2zARgO1DVAmtpNEr5RUv0oqTOSlkTEa7bnHT01Ij6y0Ioj4uo2YwCSlsqFsqjd\njQEga9R3GBacyyiqVO5/AQAV7SYJvynpQdt3Vad/VdIdtn9SUv4DddShJSFSw7kGDC/Kd9o4fu0p\ny35KuSVhJ/EW/bMBGA7UNUCa2koSRsQXbX9P0vuqsz4ZEbXRay/LJTJgyHChBAAAAAAARdVuS0JV\nk4IPL7hgzmhJiNRwrgHDi/KNMijLeZ5yS8Ju3gsUFefocOA4Amlq99eNAfQolQsl4xoBKAvqOwwL\nzmUUVSr3vwCAiuSShLQkRDNFPp69fNsPoNgo12nj+LWnLPsp5ZaEjEmIYcI5Ohw4jkCakksSAqnh\nAgkML8o3io5zFI04JwCUWZHrwCLHhvJILklIS0I0k8LxTCFGqTxdlmjhCaAs9V2vylJPptiSsIZz\nGUXSa1kqevkDgGFGkjBHXOD6p8j7usixlRlJQmSB8ydtw3z8svxsw7yf6qWYJDxw4MCc6U7ei/QV\n9Vj2GldRPxc6U6ufiqbIX6oUOTaUR3JJQiBV3PAAAAAAAICiSi5JSEtCNFPkfV3k2OqlEmdWaEmI\nLHD+oKiyrOPKcp6n2JKQrpzlVtRjSUtCSBxHIFXJJQmBVKVyoaSZO9C+VMo1mitDfcc5Wg6dnMuc\nE+gnzjcUTZHPySLHhvJILklIS0I0U+R9nUqLtaLHl7VUjguA/Axz+WdMws6l3JKw6PEiH0U97rQk\nhMRxBFKVXJIQAICi4AYYRcc5ikacEwAAYD7JJQkbb2y6vdGhJWG6mu3XIu/rVL7lL3p8WUvluADI\nTxm6G2ehLPVkii0JaziXy6mo52mvZamonwsAyiC5JCGQmtRudHjQANqXWvnGXMN8/PgipBy6Oc6c\nEwD6gboGSFNySUJaEiK1loQ1RY+x6PFljQdoAGhPWerJlFsSopyKeh4wJiEApIskYY5SutlMSWpJ\nwiLHVi+VOLNCkhBZ4PxBUR04cEBSNudoWc7zlO7bavHVjnM370X6inos6W4Mqbv6qR+K3GuqyLGh\nPJJLEgKpSuWGh4sT0L5UyjWaK0N9xzlaDp2cy5wT6CfONxRNkc/JIseG8kguSUhLQtCSMB+pxJmV\nfrQkLNs+BVJDGW1PWfZTFvdt/WqlntX9MNJW1ONOd2MASFfuSULbF9heb/sZ259t8vo5tidtP1r9\n9/t5xwQMAjc8wPChXKOoGFKhHOjKCWSDcpE99imQpsV5rtz2Iklfk3SepJckPWT7rohY37DoDyLi\n4nbWSUtCpNqSsMgxSsWPL2u0JARQhu7GWShLXZZyS0KUU1HPg6K2JCzq/gKAIsm7JeGZkjZExAsR\nMSPpTkkrmizHXTpQEDw0A+3jgQNFlcoXVOgNLQmHC8dmcNj32WOfAmnKO0m4VNJY3fTG6rxGZ9te\nY/tu26e2WiEtCUFLwnwUPb6s0ZIQAGW0PWXZTym2JEzlHgOtFfl5phu0JASAdOXa3bhNj0g6KSJ2\n2b5Q0l9KOnm+hdevX69XXnmlb8EBveKGBBhelG8UFcmjcqAl4XDh2AwO+z577FMgO6OjoxodHe3L\ntvJOEo5LOqluell13qyImK77+3u2b7B9fERsa7bCd7zjHTrttNP0wAMP1N7TVWC0JMxe1t1U51tf\nai0Ja7qJsZ9df2vxlaW7cV4P0PX7L4XzEr3hGDeXyn7Jo77rZx3azrayOBapHM9epdiSsN/vRT66\nPW+KeixTbEk432tluS8GUGwjIyMaGRmZnV61alVu28q7u/FDkt5me7ntIyRdIml1/QK2l9T9faYk\nz5cglCoVeJY3cXkqW5Iw68843/pSSxKmciNf5H2Yh7we5MpW7oFmUjn384izKPV2lnGkcjx7lWKS\n8MCBA33ZHvJV5OeZbqSYJMzyPahg3wFpyrUlYUTst321pHtVSUjeFBHrbF9ZeTlulPRB278taUbS\na5I+nGdMwKBwoQSGD+UaRUV343Ig6TFcODaDw77PHvsUSFPuYxJGxN9IekfDvD+r+/t6Sdd3sL6W\n092uJw+0KMpHqi0JixyjVPz4ssYPl6Df6s+5Ye++lMq5P+zHISupHM9epdiSEMOhyM8z3RimloQA\nUDZ5dzcGkBgemoH28cCBokrlCyr0hpaEw4VjMzjs++yxT4E0JZckpCUhaEmYj6LHlzVaEqLfsjrn\nUqhTihxbvVTiHLSy7KcUWxKmUB9gYUV+nukGLQkBIF3JJQkBACgKHjhQVCSPyoGWhMOFYzM47Pvs\nsU+BNCWXJGwcy6nI37yVrSVhv7qplqklYT+7/hZ5H+YhrwfoLOonpKOTY0xLwuLJo47tZ73dzray\nOBapHM9epdiSsN/vRT66PW+KeixTbEk432sMwwOgbJJMEmZ5E5fVcgu9twgX8bwvuFmvf771Npuf\nVTf0PPSyf7L8HAutq/Z6kW+G8tgfWZ877Zb7Ip2jnchyfw1Dd6JOyndWdUGvdW6/vyQrsjy+dOr0\n2p9XGcry2lyU+5l+Jt6yThLmfY/UzRcWaE8/68x+Jwmz+mx53ktlqZv7siKVl0E/S3S6vSLtu37p\npQwP8/4a5s82jJJLEgKpVTKpxYvscOwBAHniOgOgqKifgDQllyTM6puqdt/XS6uqon0zkFcLsbzW\nu9C3qq26dRaxNdygz4GF9smg42tHlsd10N1KiniOtqMx7l4+R2p1Uq/bzGrogV5biPVj/6RQn0jN\n90WW+6eddeV9Dc16XYOsu/Ledp4tCbOOPYVhB4ZFkevMXo9/Vp+tcT0ptiRMQTfHq991dtmH3el0\nf9cvn+qzQTuG+bMNo+SShADyRSUOtK+MN8BIA+dmOaTQ/RDt49gMDvs+e+xTIE3JJQn73ZKwF0Vr\nSZiaTsb1KfL+TeXb/qLHlzXGZUO/ZVUXpFCnFDm2eqnEOWhl2U95tiTMWi9jEqJ4ivw80w1aEkJi\nfwOpSi5JCABAUXADjKLi3CwHWhIOF47N4LDvs8c+BdKUXJIwIjIZ66CMLQnzjqFf4+2UqSVhP8bg\na/f1YZPX5y37WCxl080vitKSsDjyGF6hn0M29GtbqRzPXuXRkjCv8trL+spyPFPS7XlS1GOZYkvC\nQY9VDQBFkVySEEC+uBkCAAAA0IuiJrEBtJZckpAxCbtXhBg6QUvCfOLo9vVhw5iEyAItCZsrcmxl\nkdevGw+zFFsSdrP+shzPlNCSMNv3d7Peou7LlLFPgTQlmSTM8iYuq+UWem8RKsm8L7h53wC3mp9V\n8jgPveyffj7kFWmfzacf+yPLG9thvBnNsqwNw8DknZTvrOqCXutcEuSt9asOyGJ7/apjinI/k/e2\nU0wSHjhwoOv3oj39rDP7nSTM6rNlfR9epCRhkcrLoJ8lOt1eN/VT6nopw0U617I2zJ9tGCWXJARS\nq2RSi5fuxtlJ7dgDAAAAWeA+GEhTcknCfnc37iVhUrRvBvKKIa+k0kLfqrb6gYgiJroGfQ4MQ0vC\nfvyQS7/OnSKeo+1ojLuXz5HXPujnvq2dR+1ss5cWRc3qu27LbD/2Twr1idQ8ziz3TzvryvsamvW6\nBll35b3tlFsSplLmUlTkOrPX457VZxuGloQp6OZ49bvOrt9eGVsSdrq/65dP9dmgHcP82YZRcklC\nAAAAtJb6wzDak0L3Q7SPYzM47PvssU+BNCWXJIyIli3IOllP3srSkrAm628IFhqTcKF5RZFV66Gs\n4uj29WHTj5a1ZdunZdTJMc6qRVFeLZOyVOTY6uXxzXY/vy3v17ZSOZ69SrElYb/fi3wMakzCrAxD\nS8JB9zABgKJILkkIIF/cDAEAAADoRVGS2AA6k3uS0PYFttfbfsb2Z+dZ5jrbG2yvsX16q/X1e0zC\nXpStJWHWaEmYTxzdvp610dHRvm6vUb/rAAwnWhIeVF+mixZbPxXls+c1JuEwS7ElYTfrb3fZQV+n\ny4SWhK3Xl5VuWhIOk36X6TLsU2AY5ZoktL1I0tcknS/pNEmX2n5nwzIXSnprRLxd0pWSvpFnTACK\ngYcPYLhQpoH+y/MhnDINDBeShADakXdLwjMlbYiIFyJiRtKdklY0LLNC0m2SFBEPSjrO9pL5VkhL\nwu4VIYZO0JIwnzi6fX3Y0JIQ3ejlGjTsLQnrFTm2hRS1BUynaEnYuZRbEnbzXhQHLQlbry8rZW9J\n2G9l3Kdl/MwYPotzXv9SSWN10xtVSRy2Wma8Om+i2Qq3b9+uycnJ2elNmzZpZmam48D27t07+/dL\nL70073I7d+5ccJn51Me5devWrtaRpR07dsyZziqe3bt3S6pUill+xpdfflnSoXHWPkf959m8eXPT\nZQa9z6WD58GuXbvmzG8ntm3btrW97ELrrq1rPq+++urs//3Ybzt27Oh4O928Zz4TE02rmJ4///bt\n22f/bjzm0sFjMzU1VYjzs1O187kW+yuvvCJJGh8f73g8y+np6Vz2QZ7lv/Hmr3aM2zk3t2zZMhvX\nMccc09F268/L+uvS0Ucf3dF6pErdmtd+r6239lnrFfF837Zt2yFx9VoH1F+b2innW7duldR6/8z3\nWrP4a2rnSTvrWcjExIR+6qd+SlKl3Payrl70ck/WjvrrZP01opPt1a4Bk5OTeumll7Rnz57ZdWcZ\nd+2+p3bt3rJlS9vrb/f6k+U1t1t5br/dddf2cSfv6VStzmy2z1tts76uHeR1uPGcmu8eq91t1d/X\nZ7nP65/NGs33rJbXvUo3uqkD6+9L+1Gma9cIqXJe1BRlH0oHn12l7OPqtK6v31+Tk5Ozx6uf+6sf\n22p8fkjNokWL9KY3vWnQYfSN88x22/51SedHxL+uTv9LSWdGxDV1y/yVpD+KiL+tTt8n6fci4tGG\ndZGWBwAAAAAAQKlFRC6/OJp3S8JxSSfVTS+rzmtc5sQFlsltBwAAAAAAAABll/eYhA9Jepvt5baP\nkHSJpNUNy6yWdLkk2T5L0mRENG+jDgAAAAAAACBzubYkjIj9tq+WdK8qCcmbImKd7SsrL8eNEXGP\n7YtsPytpp6Qr8owJAAAAAAAAwFy5jkkIAAAAAAAAoPjy7m6cCdsX2F5v+xnbnx10PAAOZXuZ7ftt\nr7X9hO1rqvNfb/te20/b/p+2j6t7z+dtb7C9zvY/q5t/hu3Hq2X+K4P4PAAqbC+y/ajt1dVpyjSQ\nKNvH2f6Lahlda/u9lGkgXdUyurZaHr9p+wjKNJAO2zfZnrD9eN28zMpwtU64s/qeH9mu/82Qpgqf\nJLS9SNLXJJ0v6TRJl9p+52CjAtDEPkmfjojTJJ0t6apqWf2cpPsi4h2S7pf0eUmyfaqk35B0iqQL\nJd1gu/YDRV+X9PGIOFnSybbP7+9HAVDnU5KeqpumTAPp+qqkeyLiFEnvkrRelGkgSbaXS/qEpF+I\niJ9XZSixS0WZBlJyiyq5rnpZluGPS9oWEW+X9BVJf7pQQIVPEko6U9KGiHghImYk3SlpxYBjAtAg\nIjZHxJrq39OS1qnya+UrJN1aXexWSf+8+vfFku6MiH0R8bykDZLOtP0mScdExEPV5W6rew+APrK9\nTNJFkv5r3WzKNJAg28dK+kcRcYskVcvqlCjTQKpelbRX0k/aXizpJySNizINJCMifihpe8PsLMtw\n/bq+I+m8hWJKIUm4VNJY3fTG6jwABWX75ySdLun/SlpS+8XyiNgs6Y3VxRrL9nh13lJVynkNZR4Y\nnC9L+oyk+gGMKdNAmt4s6RXbt1SHELjR9lGiTANJiojtkv6TpBdVKZ9TEXGfKNNA6t6YYRmefU9E\n7Jc0afv4VhtPIUkIICG2j1blW4pPVVsUNv46Er+WBCTA9gckTVRbCLvFopRpIA2LJZ0h6fqIOEPS\nTlW6NHGdBhJk+y2S/q2k5ZJOUKVF4WWiTAPDJssy3OqeXlIaScJxSfWDKy6rzgNQMNWuDt+RdHtE\n3FWdPWF7SfX1N0naUp0/LunEurfXyvZ88wH01/skXWz7OUl/Lumf2L5d0mbKNJCkjZLGIuLh6vR3\nVUkacp0G0vQeSQ9ExLZqC6H/IemXRJkGUpdlGZ59zfZhko6NiG2tNp5CkvAhSW+zvdz2EZIukbR6\nwDEBaO5mSU9FxFfr5q2W9NHq378p6a66+ZdUf3HpzZLeJunvqk2qp2yfWR2I9fK69wDok4j4QkSc\nFBFvUeXae39E/CtJfyXKNJCcatelMdsnV2edJ2mtuE4DqXpa0lm2j6yWxfNU+aExyjSQFmtuC78s\ny/Dq6jok6UOq/BBKS4t7+CB9ERH7bV8t6V5Vkpo3RcS6AYcFoIHt90m6TNITth9TpVn0FyT9iaRv\n2/6YpBdU+UUmRcRTtr+tys3MjKTfiYhaU+qrJP03SUeq8iuMf9PPzwKgpT8WZRpI1TWSvmn7cEnP\nSbpC0mGiTAPJiYgf275N0iOS9kt6TNKNko4RZRpIgu07JI1IeoPtFyVdq8q99l9kVIZvknS77Q2S\ntqryxX/rmA6uEwAAAAAAAEAZpdDdGAAAAAAAAECOSBICAAAAAAAAJUeSEAAAAAAAACg5koQAAAAA\nAABAyZEkBAAAAAAAAEqOJCEAAAAAAABQciQJAQAAAAAAgJIjSQgAAAAAAACUHElCAAAAAAAAoORI\nEgIAAAAAAAAll2uS0PbrbD9o+zHba23/x3mWu872BttrbJ+eZ0wAAAAAAAAA5lqc58ojYo/tcyNi\nl+3DJD1g+30R8UBtGdsXSnprRLzd9nslfUPSWXnGBQAAAAAAAOCg3LsbR8Su6p+vq25ve8MiKyTd\nVl32QUnH2V6Sd1wAAAAAAAAAKnJPEtpeZPsxSZsljUbEUw2LLJU0Vjc9Xp0HAAAAAAAAoA/60ZLw\nQET8gqRlkv6x7XPy3iYAAAAAAACA9uU6JmG9iHjV9t2S3iPpf9e9NC7pxLrpZdV5c9iOfCMEAAAA\nAAAAii0inMd6c00S2v5pSTMRMWX7JyS9X9KqhsVWS7pK0rdsnyVpMiImmq0vgjxhO7zKOuGYE/TS\njpckSUcdfpRef+Tr9fFf+LhWndu4+4vnqZef0mk3nKarfvEqfe2irw06HORk5cqVWrly5UC2XSsj\n458+5PuI0ntx6kUt/8py3X/5/Tr3zecOOpymvMr6N2f+G/3ReX+ko//oaEnS2cvO1t9+/G8HHFm5\nDbJMA73YsWeHjv3jY/XwJx7Wu09496DDKQzKNDBcKNPA8LBzyQ9Kyr8l4c9KutWVT7BI0u0R8b9s\nXykpIuLGiLjH9kW2n5W0U9IVOcdUCvUJVZKrAAAAAAAAaCXXJGFEPCHpjCbz/6xh+uo84yij0NzE\noO1D5hVVLalJchPov9nyl0h9AQC9qtV31HsAAKDscv/hEgzGnJaECln5NUcFujEyMjLoEABkiDIN\nDBfKNDBcKNMA2kGScEjxbTiKjhsV9Kp+LI48x+VAeyjTwHChTAPDhTINoB0kCQEAAAAAAICSI0k4\npBrH87OdzBh/jA0EDM5s+UukvgCAXjEWMgAAQAVJwiFVn2CLYExCAAAAAAAAzI8k4ZBq/OESAAAA\nAAAAYD4kCYdUY2LQdjLJQrr9AIMzW/4SqS8AoFcMcwIAAFBBknBIHTImId2NAQAAAAAAMA+ShCVA\nizwAw44vQgAAAACgNyQJhxRdZgAMOxKDAAAAAJCdXJOEtpfZvt/2WttP2L6myTLn2J60/Wj13+/n\nGVNZHNLd2E6mRSFjAwGDM1v+EqgvqCMAZIGxkAEAACoW57z+fZI+HRFrbB8t6RHb90bE+oblfhAR\nF+ccS6nUPzyHghY3AAAAAAAAmFeuLQkjYnNErKn+PS1pnaSlTRYlg5Wx+m/D+WYcAAAAAAAArfRt\nTELbPyfpdEkPNnn5bNtrbN9t+9R+xTTMmnXDS61rHslN5IFuZa2lsl8iIplYARQbw5wAAABU5N3d\nWJJU7Wr8HUmfqrYorPeIpJMiYpftCyX9paSTm61n5cqVs3+PjIxoZGQkl3iHQbMxCVPBgz8weDws\nAwAAAMDgjY6OanR0tC/byj1JaHuxKgnC2yPirsbX65OGEfE92zfYPj4itjUuW58kRPsYkxDAsGn8\n4iOlL0IAAAAAoF2NjeRWrVqV27b60d34ZklPRcRXm71oe0nd32dKcrMEITpDKyAAAAAAAAC0K9eW\nhLbfJ+kySU/YfkxSSPqCpOWSIiJulPRB278taUbSa5I+nGdMZdGsu3Eq3XgZGwh54vxqbXb/JFBf\ncAwBZIGxagEAACpyTRJGxAOSDltgmeslXZ9nHGVU//AcQXdjAAAAAAAAzK9vv26M/uLbcAAAAAAA\nALSLJCEAIHm0lgYAAACA3pAkHFJzuhsrKmMSJjJ+F2MDIU+cX63N7p8E6guOIYAsMFYtAABABUnC\nIXXID5fQygYAAAAAAADzIEk4pPg2HAAAAAAAAO0iSTikDmlJaCfTNY9uP8gT51drs/sngfqCYwgg\nCwxDAQAAUEGScEg1PjzT3RgAAAAAAADzIUkIAAAAAAAAlBxJwiFFlxkAZWLTWhoAAAAAepFrktD2\nMtv3215r+wnb18yz3HW2N9heY/v0PGMqi0O6G9vJjN/F2EDIE+dXa7P7J4H6gmMIIAuMVQsAAFCx\nOOf175P06YhYY/toSY/Yvjci1tcWsH2hpLdGxNttv1fSNySdlXNcpcOYhAAAAAAAAJhPri0JI2Jz\nRKyp/j0taZ2kpQ2LrZB0W3WZByUdZ3tJnnEBAAAAAAAAOKhvYxLa/jlJp0t6sOGlpZLG6qbHdWgi\nERlIrWse3X6QB7qVtZbKfomIZGIFUGwMQwEAAFDRlyRhtavxdyR9qtqiED0afX5Um3Zsanv5lAb1\n58EfGDwelgEAw2Djqxu1/pX1Cy8IAAByH5NQtherkiC8PSLuarLIuKQT66aXVecdYuXKlbN/j4yM\naGRkJLM4U3Purefqk+/+pL7+K19va3nGJAQwTFL64gMAMDjvv/39Wv/KesW1fPkFAEjT6OioRkdH\n+7Kt3JOEkm6W9FREfHWe11dLukrSt2yfJWkyIiaaLVifJIQ0PUOjTACQ+CIEANDc7n27Bx0CAAA9\naWwkt2rVqty2lWuS0Pb7JF0m6Qnbj0kKSV+QtFxSRMSNEXGP7YtsPytpp6Qr8oyprGwn0413dmyg\nROJFWhh7qrWUyh/HEEAWGKsWAACgItckYUQ8IOmwNpa7Os84QCsbAAAAAAAAzK9vv24MAAAAAAAA\noJhIEgIAkscPmQAAAABAb0gSloTtZMbvmh0bKJF4kRbGnmotpfJXfwxTiBdAMTFWLQAAQAVJwpJg\nTEIAAAAAAADMhyQhACB5dDcGAAAAgN6QJCwJ28l0r5zt9pNIvEgL3cpaS6n8cQwBZIFhKAAAACoW\nt3rR9g5p/jumiDg284iQC7obAwAAAAAAYD4tk4QRcYwk2f6ipE2SbpdkSZdJ+tncowMAAAAAAACQ\nu3a7G18cETdExI6IeDUivi5pRZ6BAQDQLlpLAwAAAEBv2k0S7rR9me3DbC+yfZmknXkGhmzZTmb8\nrtmxgRKJF2lh7KnWUip/9ceQ4wmgW4xVCwAAUNFukvAjkn5D0kT134eq81qyfZPtCduPz/P6ObYn\nbT9a/ff77QaOztDKBgAAAAAAAPNpOSZhTUQ8r+66F98i6T9Luq3FMj+IiIu7WDcAAJL4IgQAAAAA\netVWS0LbJ9v+X7afrE7/fDut/iLih5K2L7T6dmJA71LrjpdavEgD3cpaS2W/REQysQIoNoahAAAA\nqGi3u/F/kfR5STOSFBGPS7okoxjOtr3G9t22T81onWhgp5OL5cEfGDwelgEAAACgXNrqbizpqIj4\nu4ZE074Mtv+IpJMiYpftCyX9paSTM1gvGtAVD8AwSemLDwAAAABIQbtJwldsv1WqNC2x/UFJm3rd\neERM1/39Pds32D4+IrY1W37lypWzf4+MjGhkZKTXEAAAQ4CkIQAAAIBhNDo6qtHR0b5sq90k4VWS\nbpT0Ttvjkv6fpMvafK81z7iDtpdExET17zMleb4EoTQ3SYjO2E6mG+/s2ECJxIu0MPZUaymVv/pj\nmEK8AIqJsWoBAECRNTaSW7VqVW7bWjBJaHuRpPdExD+1/ZOSFkXEjnZWbvsOSSOS3mD7RUnXSjpC\nUkTEjZI+aPu3VRnr8DVJH+7uY2AhdDcGAAAAAADAfBZMEkbEAdu/J+nbEbGzk5VHxEcWeP16Sdd3\nsk4AABrR3RgAAAAAetPurxvfZ/vf2z7R9vG1f7lGBgAAAAAAAKAv2h2T8MOq/GjJ7zTMf0u24SAv\ntpMZg212bKBE4kVaGHuqtZTKH8cQQBYYqxYAAKCi3SThqaokCH9ZlWTh/5H0jbyCQvYYkxAAAAAA\nAADzaTdJeKukVyVdV53+SHXeb+QRFAAAneCLEAAAAADoTbtJwn8QEafWTX/f9lN5BIT8pNI1b7bb\nTyLxIi10K2stpfJXfww5ngC6xTAUAAAAFe3+cMmjts+qTdh+r6SH8wkJeeCXPwEAAAAAADCfdlsS\nvlvS39p+sTp9kqSnbT8hKSLi53OJDpmhKx6AYUYdBwAAAAC9aTdJeEGuUQAAAAAAAAAYmLaShBHx\nQt6BIF+2kxmza3ZsoETiRVoiQpYZe2oeKZU/jiGALDBWLQAAQEW7YxIicXTFAw5ijE4AAAAAAObK\nNUlo+ybbE7Yfb7HMdbY32F5j+/Q84wEADCcSvwAAAADQm7xbEt4i6fz5XrR9oaS3RsTbJV0p6Rs5\nx1NqqXXNSy1epIFuZa2lsl8iYk6s1BcAujU7zAL1CAAAKLlck4QR8UNJ9B2/aQAAC6VJREFU21ss\nskLSbdVlH5R0nO0lecY0TDq5mU2plU0qSQqki+73C+NhGQAwDLieAQDQvkGPSbhU0ljd9Hh1HjKW\nUlJk7/69gw4BQy6lpHk/RYT27NsjSZo5MDPgaFprPIYcUwAAAADoTVu/blwUK1eunP17ZGREIyMj\nA4ulCIb1ofj9t79/0CEApXTnk3fqI//9I5Kky/77ZfrIP/zIgCNq3+uPfP2gQwAAFNCw3i8DAMpj\ndHRUo6OjfdnWoJOE45JOrJteVp3XVH2SEJ2xnVw33tTiRRoiQpbpftTExM6JQYfQkfpj+JbXv2WA\nkQBIGWPVAgCAImtsJLdq1arcttWP7sau/mtmtaTLJcn2WZImIyKtp9REpNTdGMgbrQoAAAAAAJgr\n15aEtu+QNCLpDbZflHStpCMkRUTcGBH32L7I9rOSdkq6Is94hg0toQAAAID5cb8MAED7ck0SRsSC\nA1pFxNV5xoCDUrtJSi1epIFuZcMhIuYcQ+oLAN2q1R/UIwAAoOwG/evG6EEnXSbpXgkcRPd7AADK\ngXtgAADaR5KwJEiKAAfxwJA+jiEAAAAAZIskIQAAAAAAAFByJAlLJLUx2FKLF2lg7KnhUX8MqS8A\ndIuxagEAACpIEpYEXfOAg+h+3xyJUwAAAAAoL5KECevkgZ6kCHAQSXMAAMqBL8AAAGgfSUIAgCSS\npwAAAABQZiQJE9bJA73t5L5JTS1epCEUsszYU0Og/hhSXwDoFmPVDje+AAMAoH0kCUuC7sbAQTww\nAAAAAAAwV+5JQtsX2F5v+xnbn23y+jm2J20/Wv33+3nHBAAAAAAAAOCgxXmu3PYiSV+TdJ6klyQ9\nZPuuiFjfsOgPIuLiPGMZRp12i0mte2Vq8SINdCubX0r7JCLmxEt9AaBbtfqDemQ4pXRtAwBg0PJu\nSXimpA0R8UJEzEi6U9KKJsvR9y9ndK8EDqL7PQAAAAAAc+WdJFwqaaxuemN1XqOzba+xfbftU3OO\naWh09MMlJEWAWSTNm0tpv6QUKwBgcLheAADQvly7G7fpEUknRcQu2xdK+ktJJw84JgAAAAAAAKA0\n8k4Sjks6qW56WXXerIiYrvv7e7ZvsH18RGxrXNnKlStn/x4ZGdHIyEjW8Q4t28mNyZJavEhDKGSZ\nsaeGQP0xpL4A0C3GqgUAAEU2Ojqq0dHRvmwr7yThQ5LeZnu5pE2SLpF0af0CtpdExET17zMluVmC\nUJqbJERn6G4MHETXIwAAAABAChobya1atSq3beWaJIyI/bavlnSvKuMf3hQR62xfWXk5bpT0Qdu/\nLWlG0muSPpxnTACA5mhFAwAAAADllfuYhBHxN5Le0TDvz+r+vl7S9XnHMYw6faBPrXtlavEiDXQr\nGw4RMecYUl8A6Fat/qAeGU5c7wEAaF/ev26MgqB7JXAQ3e+bo54AAAAAgPIiSZiwTh7oSYoAB5EM\nSx/HEADQDq4XAAC0jyQhAAAAAAAAUHIkCUsktTFZUosXaWDsqeFRfwypLwB0i7FqAQAAKkgSJqyT\nm1m6WgAH0f2+OR6QAQDDhmsbAADtI0lYEiRFgINImgMAAAAAMBdJwoSR6AAAAADmx/0yAADtI0lY\nEraTG4MttXiRhoiQZbofDYH6Y0h9AaBbjFULAABQQZJwiNV3Maa7MXAQrQqa4wEZAAAAAMqLJOEQ\nIxECoBO0rgQAAACA8so9SWj7AtvrbT9j+7PzLHOd7Q2219g+Pe+YyqKx9WBqrYRIWCAPdCubX0r7\nJCLmxEt9AaBbtfqDegQAAJRdrklC24skfU3S+ZJOk3Sp7Xc2LHOhpLdGxNslXSnpG3nGNEw6uZml\nVSGKZnR0dGDbpvt9czwgoxeDLNMAsjcsZZprG1AxLGUaQL7ybkl4pqQNEfFCRMxIulPSioZlVki6\nTZIi4kFJx9leknNcpVCfGCQpgqIZaJKQpHnyOIbFw8MHMFwo08BwoUwDaEfeScKlksbqpjdW57Va\nZrzJMmhioYdkEoMAOpFSd2MAANrBl0oAALRv8aAD6MSv/vmvDjqEQrnzyTs1uXty3tePOvwoTe2Z\nkiQtXrRY9z13X1L78MHxB5OKF515+omn9cifP9L37e7et1uLvEiWOb8aPLvt2TnTRd4/33nqO3py\ny5Oz06ufWa3nJp8bYEQYVJkGerVjzw5J0n/4P/9BN6+5ecDRFMewlOnnJ5+XVOxrGtAPw1KmgX77\n6aN+WresuGXQYfSN8xynw/ZZklZGxAXV6c9Jioj4k7plviHp+xHxrer0eknnRMREw7po4gIAAAAA\nAIBSi4hcmsrn3ZLwIUlvs71c0iZJl0i6tGGZ1ZKukvStalJxsjFBKOW3AwAAAAAAAICyyzVJGBH7\nbV8t6V5Vxj+8KSLW2b6y8nLcGBH32L7I9rOSdkq6Is+YAAAAAAAAAMyVa3djAAAAAAAAAMWX968b\nZ8L2BbbX237G9mcHHQ+AQ9leZvt+22ttP2H7mur819u+1/bTtv+n7ePq3vN52xtsr7P9z+rmn2H7\n8WqZ/8ogPg+ACtuLbD9qe3V1mjINJMr2cbb/olpG19p+L2UaSFe1jK6tlsdv2j6CMg2kw/ZNtids\nP143L7MyXK0T7qy+50e2T1oopsInCW0vkvQ1SedLOk3SpbbfOdioADSxT9KnI+I0SWdLuqpaVj8n\n6b6IeIek+yV9XpJsnyrpNySdIulCSTfYro09+nVJH4+IkyWdbPv8/n4UAHU+JempumnKNJCur0q6\nJyJOkfQuSetFmQaSVB33/xOSfiEifl6VocQuFWUaSMktquS66mVZhj8uaVtEvF3SVyT96UIBFT5J\nKOlMSRsi4oWImJF0p6QVA44JQIOI2BwRa6p/T0taJ2mZKuX11upit0r659W/L5Z0Z0Tsi4jnJW2Q\ndKbtN0k6JiIeqi53W917APSR7WWSLpL0X+tmU6aBBNk+VtI/iohbJKlaVqdEmQZS9aqkvZJ+0vZi\nST8haVyUaSAZEfFDSdsbZmdZhuvX9R1J5y0UUwpJwqWSxuqmN1bnASgo2z8n6XRJ/1fSktovlkfE\nZklvrC7WWLbHq/OWqlLOayjzwOB8WdJnJNUPYEyZBtL0Zkmv2L6lOoTAjbaPEmUaSFJEbJf0nyS9\nqEr5nIqI+0SZBlL3xgzL8Ox7ImK/pEnbx7faeApJQgAJsX20Kt9SfKraorDx15H4tSQgAbY/IGmi\n2kLYLRalTANpWCzpDEnXR8QZknaq0qWJ6zSQINtvkfRvJS2XdIIqLQovE2UaGDZZluFW9/SS0kgS\njkuqH1xxWXUegIKpdnX4jqTbI+Ku6uwJ20uqr79J0pbq/HFJJ9a9vVa255sPoL/eJ+li289J+nNJ\n/8T27ZI2U6aBJG2UNBYRD1env6tK0pDrNJCm90h6ICK2VVsI/Q9JvyTKNJC6LMvw7Gu2D5N0bERs\na7XxFJKED0l6m+3lto+QdImk1QOOCUBzN0t6KiK+WjdvtaSPVv/+TUl31c2/pPqLS2+W9DZJf1dt\nUj1l+8zqQKyX170HQJ9ExBci4qSIeIsq1977I+JfSforUaaB5FS7Lo3ZPrk66zxJa8V1GkjV05LO\nsn1ktSyep8oPjVGmgbRYc1v4ZVmGV1fXIUkfUuWHUFpa3MMH6YuI2G/7akn3qpLUvCki1g04LAAN\nbL9P0mWSnrD9mCrNor8g6U8kfdv2xyS9oMovMikinrL9bVVuZmYk/U5E1JpSXyXpv0k6UpVfYfyb\nfn4WAC39sSjTQKqukfRN24dLek7SFZIOE2UaSE5E/Nj2bZIekbRf0mOSbpR0jCjTQBJs3yFpRNIb\nbL8o6VpV7rX/IqMyfJOk221vkLRVlS/+W8d0cJ0AAAAAAAAAyiiF7sYAAAAAAAAAckSSEAAAAAAA\nACg5koQAAAAAAABAyZEkBAAAAAAAAEqOJCEAAAAAAABQciQJAQAAAAAAgJIjSQgAAAAAAACUHElC\nAAAAAAAAoOT+P1TraRlYPEE5AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x118e54f60>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def visualize(calcium, spikes, spikes2=None):\n",
    "    plt.figure(figsize=(18, 6))\n",
    "    t = np.arange(len(calcium))\n",
    "    plt.hold(True)\n",
    "    if spikes2==None:\n",
    "        plt.subplot(211)\n",
    "        plt.plot(t, calcium, color='#348ABD')\n",
    "        plt.subplot(212)\n",
    "        plt.plot(t, spikes, color='black',label='gt')\n",
    "    else:\n",
    "        plt.subplot(311)\n",
    "        plt.plot(t, calcium, color='#348ABD')\n",
    "        plt.subplot(312)\n",
    "        plt.plot(t, spikes, color='black',label='gt')\n",
    "        plt.ylabel('gt')\n",
    "        plt.subplot(313)\n",
    "        plt.plot(t, spikes2, color='g',label='pred')\n",
    "        plt.ylabel('pred')\n",
    "#     plt.ylim([-6, 3])\n",
    "    plt.xlim([0,10000])\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "    \n",
    "def corr(gt, pred):\n",
    "    return np.corrcoef(gt,pred)[0,1]\n",
    "\n",
    "print('----------training----------')\n",
    "y_pred_train = np.argmax(model.predict(x_train)[0,:,:],axis=1)\n",
    "print('pred nonzeros\\t',str(np.sum(y_pred_train) / y_pred_train.size))\n",
    "print(\"acc\\t\\t\",accuracy_score(y_pred_train,y_train_orig))\n",
    "print(\"corr\\t\\t\",corr(y_pred_train,np.argmax(y_train[0,:,:],axis=1)))\n",
    "\n",
    "print('----------testing----------')\n",
    "y_pred = np.argmax(model.predict(x_test)[0,:,:],axis=1)\n",
    "print('pred nonzeros\\t',str(np.sum(y_pred) / y_pred.size))\n",
    "print(\"acc\\t\\t\",accuracy_score(y_pred,y_test))\n",
    "print(\"corr\\t\\t\",corr(y_pred,y_test))\n",
    "# visualize(x_train[0,:,0],y_train_orig,y_pred_train)\n",
    "visualize(x_test[0,:,0],y_test,y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "# visualize model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from IPython.display import SVG\n",
    "from keras.utils.vis_utils import model_to_dot\n",
    "# SVG(model_to_dot(model).create(prog='dot', format='svg',))\n",
    "# plt.savefig(\"network.pdf\")\n",
    "# from keras.utils import plot_model\n",
    "plot_model(model, to_file='network.png')"
   ]
  }
 ],
 "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": "3.4.1"
  },
  "latex_envs": {
   "LaTeX_envs_menu_present": true,
   "autocomplete": true,
   "bibliofile": "biblio.bib",
   "cite_by": "apalike",
   "current_citInitial": 1,
   "eqLabelWithNumbers": true,
   "eqNumInitial": 1,
   "hotkeys": {
    "equation": "Ctrl-E",
    "itemize": "Ctrl-I"
   },
   "labels_anchors": false,
   "latex_user_defs": false,
   "report_style_numbering": false,
   "user_envs_cfg": false
  },
  "toc": {
   "colors": {
    "hover_highlight": "#DAA520",
    "running_highlight": "#FF0000",
    "selected_highlight": "#FFD700"
   },
   "moveMenuLeft": true,
   "nav_menu": {
    "height": "84px",
    "width": "252px"
   },
   "navigate_menu": true,
   "number_sections": true,
   "sideBar": true,
   "threshold": 4,
   "toc_cell": false,
   "toc_section_display": "block",
   "toc_window_display": false,
   "widenNotebook": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}