MLP_shap.ipynb 189 KB
Newer Older
thecml's avatar
thecml committed
1
2
3
4
{
 "cells": [
  {
   "cell_type": "code",
5
   "execution_count": 10,
6
   "metadata": {},
thecml's avatar
thecml committed
7
   "outputs": [
thecml's avatar
thecml committed
8
9
    {
     "name": "stderr",
10
     "output_type": "stream",
thecml's avatar
thecml committed
11
12
13
14
15
16
     "text": [
      "`Model.state_updates` will be removed in a future version. This property should not be used in TensorFlow 2.0, as `updates` are applied automatically.\n"
     ]
    },
    {
     "name": "stdout",
17
     "output_type": "stream",
thecml's avatar
thecml committed
18
     "text": [
19
20
21
22
23
24
      "Test loss: 0.6338703265403236 \n",
      "Test accuracy: 0.575419008731842\n",
      "Test precision: 0.7761194109916687 \n",
      "Test recall: 0.4601770043373108\n",
      "Test roc_auc: 0.6585733294487 \n",
      "Test pr_auc: 0.7767161726951599\n"
thecml's avatar
thecml committed
25
     ]
thecml's avatar
thecml committed
26
27
    }
   ],
28
29
30
31
32
33
34
35
36
37
   "source": [
    "import tensorflow as tf\n",
    "import shap\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from tools import data_loader, preprocessor\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import train_test_split\n",
    "from pathlib import Path\n",
    "import paths as pt\n",
38
    "from utility.settings import load_settings\n",
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
    "tf.compat.v1.disable_v2_behavior()\n",
    "\n",
    "def make_model(input_dim):\n",
    "    model = tf.keras.models.Sequential()\n",
    "    model.add(tf.keras.layers.Dense(80,\n",
    "                                    input_dim=input_dim,\n",
    "                                    activation='relu'))\n",
    "    model.add(tf.keras.layers.Dropout(0.35))\n",
    "    model.add(tf.keras.layers.Dense(20, activation='relu'))\n",
    "    model.add(tf.keras.layers.Dropout(0.15))\n",
    "    model.add(tf.keras.layers.Dense(10, activation='relu'))\n",
    "    model.add(tf.keras.layers.Dropout(0.15))\n",
    "    model.add(tf.keras.layers.Dense(1, activation='sigmoid'))\n",
    "    metrics = [\n",
    "      tf.keras.metrics.BinaryAccuracy(name='accuracy'),\n",
    "      tf.keras.metrics.Precision(name='precision'),\n",
    "      tf.keras.metrics.Recall(name='recall'),\n",
    "      tf.keras.metrics.AUC(name='roc_auc'),\n",
    "      tf.keras.metrics.AUC(name='pr_auc', curve='PR')\n",
    "    ]\n",
    "    model.compile(loss='binary_crossentropy',\n",
    "                  optimizer=\"Adam\",\n",
    "                  metrics=metrics)\n",
    "    return model\n",
    "\n",
64
65
66
    "# Load data\n",
    "data_settings = load_settings(pt.CONFIGS_DIR, 'data.yaml')\n",
    "target_settings = load_settings(pt.CONFIGS_DIR, 'complete.yaml')\n",
67
    "\n",
68
69
70
71
72
73
    "ats_resolution = data_settings['ats_resolution']\n",
    "converters = {str(i)+'Ats':str for i in range(1, ats_resolution+1)}\n",
    "dl = data_loader.CompleteDataLoader(file_path=pt.PROCESSED_DATA_DIR,\n",
    "                                    file_name=\"complete_emb.csv\",\n",
    "                                    settings=target_settings,\n",
    "                                    converters=converters).load_data()\n",
74
75
76
77
78
79
80
81
82
83
84
85
86
87
    "features = dl.get_features()\n",
    "X, y = dl.prepare_data()\n",
    "\n",
    "# Calculate class weight\n",
    "neg, pos = np.bincount(y)\n",
    "class_weight = preprocessor.get_class_weight(neg, pos)\n",
    "\n",
    "# Make a train, validation and test set\n",
    "X_train, X_rem, y_train, y_rem = train_test_split(X, y, train_size=0.7,\n",
    "                                                  stratify=y, random_state=0)\n",
    "X_valid, X_test, y_valid, y_test = train_test_split(X_rem, y_rem, test_size=0.5,\n",
    "                                                    stratify=y_rem, random_state=0)\n",
    "\n",
    "# Upsample\n",
88
89
90
    "#neg = (pd.Series(y_train == 0))\n",
    "#X_train = np.concatenate((X_train, X_train[neg]), axis=0)\n",
    "#y_train = np.concatenate((y_train, y_train[neg]), axis=0)\n",
91
92
93
94
95
96
97
98
99
100
101
    "\n",
    "# Shuffle data\n",
    "#idx = np.arange(len(X_train))\n",
    "#np.random.shuffle(idx)\n",
    "#X_train = X_train[idx]\n",
    "#y_train = y_train[idx]\n",
    "\n",
    "# Make model\n",
    "model = make_model(input_dim=X_train.shape[1])\n",
    "history = model.fit(X_train, y_train, epochs=50,\n",
    "                    class_weight=class_weight,\n",
102
    "                    validation_data=(X_valid, y_valid),\n",
103
104
105
106
107
108
    "                    batch_size=32, verbose=False)\n",
    "results = model.evaluate(X_test, y_test)\n",
    "print(f'Test loss: {results[0]} \\nTest accuracy: {results[1]}' +\n",
    "      f'\\nTest precision: {results[2]} \\nTest recall: {results[3]}' +\n",
    "      f'\\nTest roc_auc: {results[4]} \\nTest pr_auc: {results[5]}')"
   ]
thecml's avatar
thecml committed
109
110
111
  },
  {
   "cell_type": "code",
112
   "execution_count": 11,
113
   "metadata": {},
thecml's avatar
thecml committed
114
115
116
   "outputs": [
    {
     "data": {
117
      "image/png": "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",
thecml's avatar
thecml committed
118
      "text/plain": [
119
       "<Figure size 432x288 with 1 Axes>"
120
      ]
121
     },
122
123
124
     "metadata": {
      "needs_background": "light"
     },
125
     "output_type": "display_data"
thecml's avatar
thecml committed
126
127
128
    },
    {
     "data": {
129
      "image/png": "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",
thecml's avatar
thecml committed
130
      "text/plain": [
131
       "<Figure size 432x288 with 1 Axes>"
132
      ]
thecml's avatar
thecml committed
133
     },
134
135
136
     "metadata": {
      "needs_background": "light"
     },
137
     "output_type": "display_data"
thecml's avatar
thecml committed
138
139
140
    },
    {
     "data": {
141
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEWCAYAAABxMXBSAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAA+SElEQVR4nO3deXxU1dnA8d9DkE1AZLGyh1UEhVACVqwWrAsqglVEkVrQKuqrr0uLC6WKouDrUqVWisUVAcWtIiiWIu6KSDBhC6CICYKoEFllDXneP86dZDLMTGYmM5kk83w/n/nMzLnn3nvuZHKfOcs9V1QVY4wxJlI1kl0AY4wxVYsFDmOMMVGxwGGMMSYqFjiMMcZExQKHMcaYqFjgMMYYExULHKbcRORtERkR77zJJCJ5InJGArarItLRe/2EiNwZSd4Y9jNcRP4bazmNCccCR4oSkd1+jyIR2ev3fng021LVc1R1WrzzVneqeq2q3lve7YhIuhdkavpte6aqnlXebQfZVz/v+7JbRHaJyFoRuSIgj4jIrSLylfe92iAi94tI7YB8fURknohsF5GfROTzwG2F2L+KyO1B0jcGyf++iFzl976ziLwiIltFZIeILBeRP4lIWmyfSGqywJGiVLW+7wFsAM73S5vpy+d/MjLG8533vWkI3AI8KSLH+S1/DBgF/AFoAJwD/BZ42ZdBRE4G3gU+ADoCTYDrvLzhjAB+8rYdFRHpACwGvgVOVNWjgIuBTK+cJlKqao8UfwB5wBne637ARuB24HtgOnA08CawBdjmvW7lt/77wFXe65HAx8DDXt5vgHNizNsO+BDYBbwDTAZmhDiGSMp4L/CJt73/Ak39ll8O5AMFwFj/zyRgPyd5n0uaX9rvgOXe6z7AImA7sBl4HKjll1eBjt7r54D7/Jbd6q3zHXBlQN7zgGxgJ+7Ed7ffehu8vLu9x8m+z9YvT19gCbDDe+4b6WcTcPz9gI0BaT8CF3uvOwGHgD4BeVoD+4HTvfcfA5Oj/J4e6ZXvUuAAkBmuXEG+bzOAt5L9/1YdHlbjMMEcCzQG2uJ+OdYAnvXetwH24k6IoZwErAWaAg8CT4uIxJD3BeBz3K/Ru3En91AiKeNlwBXAMUAtYDSAiHQFpnjbb+Htr1WwnajqYuBn4PSA7b7gvT6E+xXeFHcC/y3wP2HKjVeGAV55zsSdfAP7V37G/cpuhAsi14nIBd6y07znRupqjIsCtt0YeAtXE2gCPAK8JSJNAo7hsM+mjDLXEJFB3rGu85J/izuBf+6fV1W/BT4DzhSRerjP5tWy9hHgQlxgfAWYj6t9ROOMGPZpgrDAYYIpAsap6n5V3auqBar6mqruUdVdwATgN2HWz1fVJ1X1EDANaA78Ipq8ItIG6A3cpaoHVPVjYE6oHUZYxmdV9UtV3YtrNsnw0ocAb6rqh6q6H7jT+wxCeREYBiAiDYBzvTRUdamqfqaqhaqaB/wrSDmCGeqVb6Wq/owLlP7H976qrlDVIlVd7u0vku2CCzRfqep0r1wvAmuA8/3yhPpsgmkhIttxwfl14E+qmu0ta4qrNQWz2Vt+NO7cEypfKCOAl7zvygvApSJyRBTrN4lhnyYICxwmmC2qus/3RkTqici/RCRfRHbimo8ahelQ/N73QlX3eC/rR5m3BfCTXxq4JpqgIizj936v9/iVqYX/tr0Td0GofeFOWhd6nb0XAl+oar5Xjs4i8qaIfO+VYyLuZFmWUmXANZv5H99JIvKeiGwRkR3AtRFu17ft/IC0fKCl3/tQn00w36lqI1wfx2OUrn1txQX/YJp7y7fhAnOofIcRkdZAf8DX//YGUAcXFAEKgWBB5AjgoPe6IJp9mtAscJhgAqdM/jNwHHCSqjakpGkkVPNTPGwGGnvNGj6tw+QvTxk3+2/b22eTUJlVNRd34j2H0s1U4Jq81gCdvHL8JZYy4Jrb/L2Aq3G1Vtep+4Tfdsua4vo7XBOevzbApgjKFZJXO7sdONGv2exdoLWI9PHP6534fwUs9H4MLAIuimJ3l+POV3NF5HtgPS5w+JqrNgBNRaQ44HlNnm0pCZrvRLlPE4IFDhOJBrhmie1ee/m4RO/Q+wWfBdwtIrW8UTjnh1mlPGV8FRgoIr8WkVrAeMr+33gBuAkXoF4JKMdOYLeIdMGNFIrEy8BIEenqBa7A8jfA1cD2eSfly/yWbcH9gm8fYtvzgM4icpmI1BSRS4CuuAEE5aKqB4C/AXd577/EBbWZIvIrEUkTkW7Aa8A7qvqOt+pt3vHe6utrEZEeIjIrxK5GAPfgmtB8j4uAc0WkiapuwI2YekBE6nu1wVtxtY3PvG2MA/qKyEMicqy3z44iMkNEGpX3s0glFjhMJCYBdXHNDJ8B/6mg/Q7HdaIWAPcBL+FG5gQziRjLqKqrgOtxwWAzrinlsGsCAvj6GN5V1a1+6aNxJ/VdwJNemSMpw9veMbyL62h+NyDL/wDjRWQX7iT9st+6e3B9Op9410T8KmDbBcBAXK2sAHfSHhhQ7vJ4BmgjIr7AfgPwFG4U027c3+J9/H7tq+qnuCau04H1IvITMBUX5ErxjqctbhTW936PObjPapiX9RJc5/46XG3qt8B5vmZXVf0a931KB1Z5TX6v4X6g7IrLJ5EiRNVu5GSqBhF5CVijqgmv8RhjQrMah6m0RKS3iHTwhn0OAAYDs5NcLGNSnl0VbCqzY4F/4zqqNwLX+Q37NMYkiTVVGWOMiYo1VRljjIlKSjRVNW3aVNPT05NdDGOMqVKWLl26VVWbBaanROBIT08nKysr2cUwxpgqRUQCZxwArKnKGGNMlCxwGGOMiYoFDmOMMVGxwGGMMSYqFjiMMcZExQKHMSYuZs6E9HSoUcM9z5xZ1hqmqkqJ4bjGmMSaORNGjYI93m238vPde4Dhw5NXLpMYVuMwxpTb2LElQcNnzx6XbqofCxzGmHLbsCG6dFO1WeAwxpRbm8Ab3ZaRngpi6fOJdp1k9StZ4DDGlNuECVCvXum0evVcejJVxIk12D58fT75+aBa0ucTbv/RrhPLPuJGVav9o1evXmqMSawZM1TbtlUVcc8zZiR3HzNmqNarp+pOq+5Rr17Z5YrHPpo0KZ3me7RtG3pbbdtGt060+WMBZGmQc2pK3I8jMzNTbZJDY6q2wJFb4Go1U6cGH7mVnu5+hQdq2xby8hK7j1BEoKgo+LIaNdypP9J1os0fCxFZqqqZh+07Pps3xpj4Cdb8E27kVrD8sXTYRzs6LNrOf1+fT7DyRttPFC494U10waoh1e1hTVWmMquIJp6qJFTzT7BmmVDLY20uEgm+jkjw/KGai5o0Cd1MFur4rrsu/DqB35FYthMtQjRVJf2kXhEPCxymIlVEO3x1FupknJYWXXq4k7dq8L9TtP0G4f5+ob4H4fYRTYAItY949n1Y4DCmAkQbCCqigzPZoq1RhfrVH6pmESqvSOh9x/PXeryOL9paTajvSLTbD8cChzEVIJn/5BUl0TWqaH+RxxJ8o91HPCX6O2I1DgscpopJ5j95RYhnjSra2kCofcQSnJIZsBNdK41n86cFDmMqQDL/ycOJ5Vd0PH7dR9vsFK7/IZ7Hl+yAnehaW7xqTRY4TEpJ1kilZP6Th9pWrGWKtj8hmGg7uivyxF2VBiUk6/uclMABDADWAuuAO0LkGQrkAquAF7y0/kCO32MfcIG37DngG79lGWWVwwJH1RbunyZeJ8pY9h3PdeKhIq5ijvaEH68AlAiJrtVUBxUeOIA04GugPVALWAZ0DcjTCcgGjvbeHxNkO42Bn4B6WhI4hkRTFgscVVdZQxHjdaKMdt/JFk0zUiwn6VibmMpb1urSt1NdJCNwnAzM93s/BhgTkOdB4KoytjMKmOn33gJHCgl3gonnibIqndxiuUAu1HHE61qDeB1DZTkRb96seuhQyfvK+l1ItGQEjiHAU37vLwceD8gz2wsenwCfAQOCbOddYKDf++e85q/lwKNA7RD7HwVkAVlt2rRJ0MdqEi3c6Jdwv4qrc3NKtM1IoS6EK+tK5USf2Ctj009Rkeojj6jWqKF6000l6VVx2HQ8VNbA8SbwOnAE0A74Fmjkt7w5sAU4IiBNgNrANOCusspiNY6qK5YaR1lXDEe6j7La8xPdRh5LzSKeVxjHcmLPyVFdurTsfJXRnj2qv/+9+wxatHDHvWSJWxZJjePgwWSUOrEqa1PVE8AVfu8XAr393t8ETA2zj37Am2WVxQJH1RVLH0eoE2UosbTnV8S1BrEEtHgcd6y/olesUK1fX7VZM9Wff45tG8mSn6/6y1+6Y7/3XtVt21SbN3dpBw+G//sdOqQ6ZIhqRkbp5q3qIBmBoyaw3qtJ+DrHuwXkGQBM81439WocTfyWfwb0D1inufcswCTg/8oqiwWOqi3aUVXRiqU9P9o271jayKNpioulGSme7fYFBart26s2bOi28fjj0W8jWT74wAW7hg1V58wpSX/pJXcsjz7q3of6LtxzT8ln9+GHiSvnRx+pzpvnmtMqSoUHDrdPzgW+xI2uGuuljQcGacnJ/xHccNwVwKV+66YDm4AaAdt818u7EpgB1C+rHBY4TDgVceVxLL/uQ53Y69ZVrVnT/SL2bWPSpIo57mAOHlQ94wzVWrVUP/1U9eSTXdkPHIi+TBWpqMgFuJo1VY87TnX16sOXDxjgalHffht8G2+95T7/oUPdZ3fNNYkp6xdfqNap4/5GJ52k+s47idlPoKQEjsrysMBhypLoK49j+XUf7MReq5Z7/tvfXJ5vvlE98kjVs8+O7ZdoPGpst9ziyvTMM+79G2+495WhszuU/ftVr77alXPgQNXt24PnW7/eBerf/e7wZevWqTZq5Jqofv5Z9bLLVI8+2m07ngoKVNu1U23ZUvUf/1Bt3dqVu39/1U8+ie++AlngMJVKXp77x6uqKqKPw7ee78TeooVq7dqqp59eui39scfc9qZNi9vhRey559y+b7yxJO3QIdWuXVVPOKFim1Ui9eOPqqee6so9ZkzZ/RL33+/y+jdj/fyzavfuLlCsX+/S3nrL5Zs9O35lLSx0tZ4jjlBdtMil7dun+ve/qx5zjNvfueeqLlsWv336s8CRAvLzK+c/aqBt29xJsFYt137s+8cN9+v3889V//Qnd4yVRUVeeXzwoGsCatRIdcOG0ssOHVI95RR3Etu8ObpjKI/PPisJZIHNUtOmubPLm29WXHkikZPjPvs6dVRfeCGydQ4cUO3WTbVNG9Vdu9z/2GWXub/j22+XztesmerFF8evvHfe6T7HKVMOX7Z7t+rEie47Ua+e6tdfx2+/PhY4qjlf88ADDyS7JGW74go3Muj0012ZBwxQnTw5+C/yKVNcu7Gvj6BhQ9Vnn60aATKe7r3XHX+ok92aNe4kPmRIxZRn0ybXx9KunerWrYcvP3DANan8+tcVU55I/PvfrlmvRQv3QyQaH3/sPv8//9n92gfV++47PN8NN7i/Q6imr2jMmeP2M3Jk+O97fr5qgwaqv/1t/P8vLHBUY0VFbtigb6hmots9y+Ptt7W4iaCoSPWf/3S//mrUKB00fI8aNdwx3XKL+7V42mkufdCg2H5dx/uisx9/dM0JifT55+4zGDYsfL6JE91n89pr0W1/6VL3d9iyJfL8PXq4k/Dy5aHz+U6wH38cXXni7dAh1fHjXVn69FH97rvYtnPVVe7vULOm6uDBwZu4PvtMS/X3hBPuJP/VV6pHHeX+r/fsKXtbU6a4/T71VNl5o2GBoxqbN8/9JR95xA2JbNPGdahVNtu3q7Zq5dq/9+0rSV+5MnjQ8D38228PHXLHWbu2u9DvlVci3388RhEVFalmZamOHeuaL8B1jn7wQeTbiMbPP6t27uw+t59+Cp/3wAHVnj1Vf/GLsv/+Bw6ovviiat++JZ9Fgwaq48ap7tgRfJ3Nm1WvvNIF3aZNXZt+OLt3u7/R+eeHzxdPBw+6YDZtmrvy+9RT3XGBu7hv797Yt11Q4PoVOnUKXaMoKlLt0MHVpsNZvNh997p2VR01SnX6dDfQoajIfW4nnKDauLFLi8ShQ6q/+Y0LNps2RX5MZbHAUU0VFbm27zZt3GiOJUtcR9qgQWVXWz/++PAhiIl09dWuBrF48eHL2rQJHjRCzRaTm6uamenyDBsW+mTnrzzXLbz/vmuGaNVKi2tC/fqp3nVXySiXIUMi/0eP1J/+5LYd6fDLL75wv4o7dHDt8BMmqL7+uuqXX7qa0fffu1/fLVq47Xbo4PqZFi9Wvegil9a4seqDD5ZcxLd3r6vN1K/vvlt//rPrp4qE7xqHFStiOPgoFBWp3nZbyZBV34+Cvn3d3+2VV+LTjPPdd2U3Q40b54Lrxo3Bl+/dq3r88e5vcM45Jde+gPt+de/u1v/Pf6Ir21dfueOP5H8/UhY4qqmFC91fcfLkkrRJk7S4BhLMvn1uFAy48esVcbXrggVuf7feGr+p0A8ccCfBtDTVyy8vuwyxXinta3KpW1f1ggvcSCL/Zp2ff3YnyLp1XU1o7FjXiVpeixa5sl17bXTrvfCC6zcKDMa1a7sTP7jhu2++efjfPivLrQuqxx6r+te/qqana3Hz4JdfRleWrVtdk1Ykf59Aubmla6bh+EaWXXyx+87k5ia+CTGUL790ZXn44eDL77jDLfcFhsJC1wz7j3+oXnKJ+7wfeii2fT/0kNv2rFmxrR/IAkc11b+/66T0r4IXFbk22COOOPzX/Zo1rmkFXNUWXFNXWcrTN7Bzp1unc2fX9huPaUL8+UaelDWCJ5Yax+efu89x4MCyp9H49lvV4cPdNlu0cP0y997rgvoLL7gTxeLFkXWc7tvnfpW2bh1ZbSqUnTvdPp95RnX0aPdYs6bs9T78sGTI6gknuMAfq1tuccE9Ly/ydZYtc9+D004r+/jnz3c1wFD9DsnQu7f7Pwu0eLEr61VXJWa/Bw+6fTdrFnmfVTgWOKoh30iPYDWLggL3izM93TUrFBWVnLSbNHEjNvbvdye4M88Mv5/yXoPgW+euuxIzPfX+/e7k1qJF+CaUGTNcrSDS49i2zY0airbP6NNP3WiiI48MfqxNm7pfmOGMHRt5UE+UoiJ3jUJ5J+/79lsXfP/3fyNf56qrXA2pZk3XJBls5Jaqa2o96ijXWR+PWl68+GqpK1eWpO3d6/o0WrWKz6irUFascJ/3ZZeVf1sWOKqhc85xJ6Hdu4MvX7SoZATIpZe6v3b//qXbXn0jccK1QcfrqudETlW+ZIn7VXvFFaHzFBWVDAEG988VbHy8L++FF7rP79NPYy/X/v2uXyE31412e+01d+Jo3Nj1RwSTne2O5Q9/iH2/lc0VV7ig/f33ZectKHB5r75ade5cF0C6dTt8NNTWra6P5phjKtf1ParuONPS3OhBn8AmqkQaN87ta+7c8m3HAkcSFRWpvvxyfEc7ZGW5v97EieHz+do809JcR2lgu+/Wre6fNFzVOZ7zLCXyXtNjxrht+V+U5e+++9zyO+5weerUUT3xRNUffjg8r6/NPNa25nC+/trVYho1Kpm22yeakVFVydq1rolm9Oiy8/q+s77RdO++62pvHTqUDD44cMANTvDNj1UZDRjgvteHDrkmzxo1VP/4x4rZt68W3rJl+Wo3FjiSaMUK90l37hyfdkdV10nbqFFJ+2+o/gHf8FXfdAU+/vl9o2V+/DH4vkIFgWOOCb7vZ54Jnj9UzSNeNwjat6+kKSCwXfzZZ92+Lr+8ZMTJO++4oHn88aV/zfpGpg0cmLg282++cc2IRx1Vuh9qwgRXzmivxagKhg93f+tggdqnsNB9LqedVjp90SL3fW/VyvXRjBrlPqfp0xNb5vKYPt2VccGCimmiCrR4sftu5+bGvg0LHEn0t79p8aiWPn1CNy1Favlyt71x49z7eMybBG44abAA9Oyzh9c6fO8HDz683yDcdODxuvVoKL7Ox1GjStLmzXM1nTPPPHwCuvffd79mO3VybfG+fo3WrUO3q8dLXl7JVOSffur+wWvViu+UFZXJ6tXub3777aHz+K6WDnZ9zrJl7seKr+/IvxmoMtq1y/2fNW6sYWvCiVTekWUWOJLo7LNd5J89253UzjmnfFNOX3qpqyX4mjLiNVOrSPDO44sv1uIahn+t4o9/DB0gGjRI3j2lb71Vi3/pLVniTjQ9e7oRRsF88ok7ebdv7/42FXn1/YYNqh07ur9nt27uJBNJP0BVNWyY+3uEqnmfeab7ZR6qQ37tWvd3Gjq08oygCueyy9x38cork12S2FjgSJK9e93J2Dd76NSp7lMfMSK2i3TWrDn8V1u87g0RruM61BxI4dZJ1j2l9+xx16e0aeOCXXp62dOTLF7smkLAXfxWkTZudM2YULmnIo+HVavc9yFYbWH1avcZBJsDyl9hYdWZqywryzUrR3rBZGVjgSNJ3nnHfcr+1xj4rqYNV2UP5tAh1yFYv37pduJ41TjCPUJdBZuI4bXx8Mkn7gTVuHFk1y2ouibAv/89Ob9kf/zRXeFdVU6I5TF0qPsOBzYF3nCDa6oL1wdiKpYFjiS57TbX0eo/xryoyF0NDNHduc032iRwArV43Rsi1HUHRx8dukzxuotcIrz1lvuFayoX32CRsWNL0nbscMEklivMTeJY4EiSnj3dFdqBCgvddQLBmieCNfHk5LgAdOGFwX+VxuPeEMFGQ4mUfYOgZDVJmapryBDXD+abuPEf/3Dft2inOzeJZYEjCX74wX3CEyYEX753r2t6qlHDzX+kGvwXfN26rsOwefPoh/NGe1IfPLj0vu+6K7r9GROJZctKvl+HDrk+qT59kl0qEygpgQMYAKwF1gF3hMgzFMgFVgEv+KUfAnK8xxy/9HbAYm+bLwG1yipHsgLHzJll/4r6+WfVM85w+f71r/D9D9FecRpLM9KmTSUT4VXUxUomNf3ud+46lpdfdt+3ynxNRqqq8MABpAFfA+2BWsAyoGtAnk5ANnC09/4Yv2W7Q2z3ZeBS7/UTwHVllSVZgWPkSNc5W9ZY6r17Vc87L3TA8D2iFWvH9RVXuEnSEn0dg0lt2dklNepjjol8JlxTcUIFjhokTh9gnaquV9UDwCxgcECeq4HJqroNQFV/DLdBERHgdOBVL2kacEE8Cx0vqvDf/8IZZ0BaWvi8derAv/8NF14YOk/r1tGXYcOG6NJ9pkyBNWugSZPo92lMpDIyYPBg2LsXRo2C2rWTXSITqUQGjpbAt37vN3pp/joDnUXkExH5TEQG+C2rIyJZXvoFXloTYLuqFobZZqWQmwvffQdnnhlZ/lq14KWX4OSTD19Wpw7cf3/0ZWjTJrp0n9q1oXHj6PdnTLQmToT+/eH665NdEhONRAaOSNTENVf1A4YBT4pII29ZW1XNBC4DJolIh2g2LCKjvMCTtWXLljgWOTILFrjnSAMHQM2a8NFHcNppJWmNGsFTT8Hw4dGXYcIEqFevdFq9ei7dmMqga1d491049thkl8REI5GBYxPg38DSykvztxHX8X1QVb8BvsQFElR1k/e8Hngf6AkUAI1EpGaYbeKtN1VVM1U1s1mzZvE5oij8979w3HHQtm1066WlwXvvwW23weWXQ0FBbEED3HpTp7oyiLjnqVNj354xxkBiA8cSoJOItBORWsClwJyAPLNxtQ1EpCmu6Wq9iBwtIrX90k8Bcr3OmveAId76I4A3EngMMdm/Hz74ILrahr8aNeCBB+D5591rn5kzIT3dpaWnu/dlGT4c8vKgqMg9W9AwxpRXwgKH1w9xAzAfWA28rKqrRGS8iAzyss0HCkQkFxcQblXVAuB4IEtElnnp/6equd46twN/EpF1uD6PpxN1DLH69FPYswfOOit+25w503Ug5ue7jvf8fPc+kuBhjDHxJO5HfPWWmZmpWVlZFba/MWPg4Yfhp5+gQYP4bDM93QWLQG3bupqEMcbEm4gs9fqaS0l253i1tGCBGx0Vr6ABsQ+tNcaYeLPAEWdbtsAXX8S3mQpiH1prjDHxZoEjzhYudH0QsXaMh2JDa40xlYUFjjhbsMBde5F5WKtg+djQWmNMZVGz7CwmUtFMMxKL4cMtUBhjks9qHHG0Zg1s3Bj/ZipjjKlMLHDE0bJl7rlv3+SWwxhjEskCRxz5rqdIT09mKYwxJrEscMRRfr6birx+/WSXxBhjEscCRxzl50c/qaExxlQ1FjjiKC/PmqmMMdWfBY448U08aDUOY0x1Z4EjTrZudTPiWuAwxlR3FjjixDdzrTVVGWOqOwscceILHFbjMMZUdxY44sSu4TDGpAoLHHGSnw8NG7oJDo0xpjqzwBEneXnWTGWMSQ0WOOIkP9+aqYwxqcECR5yEu4Zj5kwXVGrUcM8zZ1ZkyYwxJr4SGjhEZICIrBWRdSJyR4g8Q0UkV0RWicgLXlqGiCzy0paLyCV++Z8TkW9EJMd7ZCTyGCKxfTvs2BG8xjFzJowa5QKL7yLBUaMseBhjqq6EBQ4RSQMmA+cAXYFhItI1IE8nYAxwiqp2A272Fu0B/uClDQAmiUgjv1VvVdUM75GTqGOIVLihuGPHugsD/e3Z49KtJmKMqYoSeQfAPsA6VV0PICKzgMFArl+eq4HJqroNQFV/9J6/9GVQ1e9E5EegGbA9geWNmW8obrDAsWFD8HV8NQ9fUPG9B7vLnzGmcktkU1VL4Fu/9xu9NH+dgc4i8omIfCYiAwI3IiJ9gFrA137JE7wmrEdFpHawnYvIKBHJEpGsLVu2lO9IyhDuqvE2bYKvk5YWuiZijDGVWbI7x2sCnYB+wDDgSf8mKRFpDkwHrlDVIi95DNAF6A00Bm4PtmFVnaqqmaqa2axZs4QdALjAUbcuNG16+LIJE6BevdJp9erBoUPBtxWqhmKMMZVFIgPHJqC13/tWXpq/jcAcVT2oqt8AX+ICCSLSEHgLGKuqn/lWUNXN6uwHnsU1iSWV7xoOkcOXDR8OU6eWLG/btuR9MKFqKMYYU1kkMnAsATqJSDsRqQVcCswJyDMbV9tARJrimq7We/lfB55X1Vf9V/BqIYiIABcAKxN3CJEp6xqO4cNdcCkqcs/Dh4euiUyYkMCCGmNMHCQscKhqIXADMB9YDbysqqtEZLyIDPKyzQcKRCQXeA83WqoAGAqcBowMMux2poisAFYATYH7EnUMkYrlqvFQNRHrGDfGVHaiqsEXiDQDmqlqbkB6V2CLqia2xzmOMjMzNSsrKyHb3r0bGjSA+++HO4JeqWKMMVWTiCxV1czA9HA1jn/gftEHagL8PV4Fq+psOnVjTKoJFzg6quqHgYmq+hHQPXFFqloscBhjUk24wNEgzLIj4l2Qqsru/GeMSTXhAsc6ETk3MFFEzgHWJ65IVUteHtSqBccem+ySGGNMxQg35cjNwFsiMhRY6qVlAicDAxNcriojP99de1Ej2ZdSGmNMBQl5ulPVr4ATgQ+AdO/xAdDdfy6pVGc3cDLGpJqwkxz6XZ1tQsjPh3MPa9AzxpjqK2TgEJFdgP9FHgpsxV2od7t3oV5K27cPvv/eOsaNMaklXFNVA1Vt6Pc4CtfHsQp4osJKWIn5JiS0pipjTCqJqktXVbep6qNAhwSVp0qxobjGmFQU9VggETmCxN4AqsoIdwMnY4yprsL1cVwYJPlo4BLg1SDLUk5+vrshU8vA21MZY0w1Fq7GcX7AYyDuBkp/V9XxFVC2Si8/H1q1gpo17f7hxpjUEbLGoapXhFomIr1VdUliilR1+K7hmDnT7h9ujEkdEfdxiEhXEblXRNYBUxJYpirDdwOnsWPt/uHGmNQRtpNbRNJx9wIfBhwE2gKZqpqX8JJVcgcPwqZNrsYR6j7hdv9wY0x1FLLGISKLcPf8rglcpKq9gF0WNJyNG92tYNu2DX2fcLt/uDGmOgrXVPUDbmr1XwDNvLTgtwtMQb6huOnpdv9wY0xqCXfl+AW4SQ6XAneLyDfA0SLSp4LKVqn538DJ7h9ujEklYTvHVXWHqj6rqmcBJwF3Ao+KyLeRbFxEBojIWhFZJyJB78gtIkNFJFdEVonIC37pI0TkK+8xwi+9l4is8Lb5mIhIREcaZ/n5Lki0bu3eDx/uaiFFRe7ZgoYxprqKeFSVqv6oqo+r6inAr8vKLyJpwGTgHKArMExEugbk6QSMAU5R1W64e4AgIo2Bcbhg1QcYJyJHe6tNAa4GOnmPAZEeQzzl5UHz5lC7djL2bowxyRPT7YdUNT+CbH2Adaq6XlUPALOAwQF5rgYmq+o2b7s/eulnAwtU9Sdv2QJggIg0Bxqq6meqqsDzwAWxHEN55efbVCPGmNSUyPvWtQT8m7Q2emn+OgOdReQTEflMRAaUsW5L73W4bVaIvDyb3NAYk5qSfcPTmrjmpn64a0WeFJFG8diwiIwSkSwRydqyZUs8Nlns0CH49lurcRhjUlO46zgeEpFrgqRfIyL/F8G2NwGt/d638tL8bQTmqOpBVf0G+BIXSEKtu8l7HW6bAKjqVFXNVNXMZs2aBcsSs82bobDQAocxJjWFq3GcDkwNkv4kbsLDsiwBOolIOxGpBVwKzAnIMxtX20BEmuKartYD84GzRORor1P8LGC+qm4GdorIr7zRVH8A3oigLHHlfw2HMcakmnBTjtT2OqBLUdWiSIbAqmqhiNyACwJpwDOqukpExgNZqjqHkgCRCxwCbvXdklZE7sUFH4DxqvqT9/p/gOeAusDb3qNC+V/DYYwxqSZc4NgrIp1U9Sv/RG8I7d5INq6q84B5AWl3+b1W4E/eI3DdZ4BngqRnASdEsv9E+eYb92xTihhjUlG4wHEX8LaI3Ie7ehzcPcfH4F1vkaqWL4d27eDII5NdEmOMqXjh7sfxtohcANwK/K+XvAo34eGKCihbpZWdDRkZyS6FMcYkR9hp1VV1JTBCROp773dXSKkqsV27YN06+MMfkl0SY4xJjrDXcYjI/4jIBiAfyBeRfBH5n4opWuW0bJl77tkzueUwxphkCXcdx19xw277qWoTVW0C9AfO8ZalpOxs92yBwxiTqsLVOC4HLlTV9b4E7/VQ3PUTKSk7G5o2hRYtkl0SY4xJjnCBQ1V1X5DEvUBR4opUuWVnu9pGciZzN8aY5AsXODaJyG8DE0XkdGBz4opUeR04AKtWWTOVMSa1hRtVdSPwhoh8TOnrOE7h8OnRU0JuLhw8aIHDGJPawt06dhXuCu0PgXTv8SFwgrcs5VjHuDHGlH0dxz4Cpv0QkRoiMlxVZya0ZJVQdjbUqwcdOya7JMYYkzzhhuM2FJExIvK4iJwpzg242WuHVlwRK4/sbOjRA9LSkl0SY4xJnnCd49OB44AVuFu8vgdcDFygqinXx1FU5C7+s2YqY0yqC9dU1V5VTwQQkadwI6naBBuimwrWr3fTjVjgMMakunA1joO+F6p6CNiYqkEDrGPcGGN8wtU4eojITu+1AHW994K7OLBhwktXiWRnu76Nbt2SXRJjjEmucNOqWxewn+xs6NoV6tRJdkmMMSa5ws6Oa0rk5FgzlTHGgAWOiHz/vXv07AkzZ0J6OtSo4Z5nptzVLMaYVBf2AkDj+DrGt26FsWNhzx73Pj8fRo1yr4cPT07ZjDGmoiW0xiEiA0RkrYisE5E7giwfKSJbRCTHe1zlpff3S8sRkX3ebWwRkedE5Bu/ZRmJPAYoCRzPP18SNHz27HHBxBhjUkXCahwikgZMBs4ENgJLRGSOquYGZH1JVW/wT1DV94AMbzuNgXXAf/2y3Kqqryaq7IGys6FdO8jLC758w4aKKokxxiRfImscfYB1qrpeVQ8As4htVt0hwNuquqfMnAni6xhv0yb48lDpxhhTHSUycLQEvvV7v9FLC3SRiCwXkVdFpHWQ5ZcCLwakTfDWeVREagfbuYiMEpEsEcnasmVLTAcAsHMnrFvnAseECW6SQ3/16rl0Y4xJFckeVTUXSFfV7sACYJr/QhFpDpwIzPdLHgN0AXoDjYHbg21YVaeqaqaqZjZr1izmAi5b5p579nQd4FOnQtu27g6Abdu699YxboxJJYkcVbUJ8K9BtPLSiqlqgd/bp4AHA7YxFHhdVf2nP/HdfXC/iDwLjI5biYPwdYxnZLjn4cMtUBhjUlsiaxxLgE4i0k5EauGanOb4Z/BqFD6DgNUB2xhGQDOVbx0REeACYGV8i11adjY0awYtWiRyL8YYU3UkrMahqoXe/TvmA2nAM6q6SkTGA1mqOge4UUQGAYXAT8BI3/oiko6rsXwQsOmZItIMN2dWDnBtoo4BSjrGRRK5F2OMqToSegGgqs4D5gWk3eX3egyuzyLYunkE6UxX1dPjW8rQDhyAVavg7LMrao/GGFP5JbtzvFJbtQoOHrQ5qowxxp8FjjACO8aNMcZY4AgrOxuOPBI6dUp2SYwxpvKwwBHG6tXQo4ebCdcYY4xjs+OGMX8+bNuW7FIYY0zlYr+lw0hLg6ZNk10KY4ypXCxwGGOMiYoFDmOMMVGxwGGMMSYqFjiMMcZExQKHMcaYqFjgMMYYExULHMYYY6JigcMYY0xULHAYY4yJigUOY4wxUbHAYYwxJioWOIwxxkTFAocxxpioJDRwiMgAEVkrIutE5I4gy0eKyBYRyfEeV/ktO+SXPscvvZ2ILPa2+ZKI1ErkMRhjjCktYYFDRNKAycA5QFdgmIh0DZL1JVXN8B5P+aXv9Usf5Jf+APCoqnYEtgF/TNQxGGOMOVwiaxx9gHWqul5VDwCzgMHl2aCICHA68KqXNA24oDzbNMYYE51EBo6WwLd+7zd6aYEuEpHlIvKqiLT2S68jIlki8pmIXOClNQG2q2phGds0xhiTIMnuHJ8LpKtqd2ABrgbh01ZVM4HLgEki0iGaDYvIKC/wZG3ZsiV+JTbGmBSXyMCxCfCvQbTy0oqpaoGq7vfePgX08lu2yXteD7wP9AQKgEYi4rtX+mHb9Ft/qqpmqmpms2bNyn80xhhjgMQGjiVAJ28UVC3gUmCOfwYRae73dhCw2ks/WkRqe6+bAqcAuaqqwHvAEG+dEcAbCTwGY4wxAWqWnSU2qlooIjcA84E04BlVXSUi44EsVZ0D3Cgig4BC4CdgpLf68cC/RKQIF9z+T1VzvWW3A7NE5D4gG3g6UcdgjDHmcOJ+xFdvmZmZmpWVlexiGGNMlSIiS72+5lKS3TlujDGmirHAYYwxJioWOIwxxkTFAocxxpioWOAwxhgTFQscxhhjopKw6ziMMZXPwYMH2bhxI/v27Ut2UUwlUqdOHVq1asURRxwRUX4LHMakkI0bN9KgQQPS09Nxk02bVKeqFBQUsHHjRtq1axfROtZUZUwK2bdvH02aNLGgYYqJCE2aNImqFmqBw5gUY0HDBIr2O2GBwxhjTFQscBhjQpo5E9LToUYN9zxzZvm2V1BQQEZGBhkZGRx77LG0bNmy+P2BAwfCrpuVlcWNN95Y5j769u1bvkJ63n//fY466igyMjLo0qULo0ePLrV89uzZdO/eneOPP54TTzyR2bNnl1r+8MMP06VLFzIyMujduzfPP/980P0UFhbSrFkz7rjjjlLp6enpbN26tVR5Bg4cWPz+7bffJjMzk65du9KzZ0/+/Oc/l/OIo6Cq1f7Rq1cvNcao5ubmRpx3xgzVevVUoeRRr55Lj4dx48bpQw89VCrt4MGD8dl4HLz33nt63nnnqarqnj179LjjjtOPP/5YVVVzcnK0Q4cOun79elVVXb9+vXbo0EGXLVumqqpTpkzRs846S3fs2KGqqjt27NDnnnsu6H7mzZunffv21fbt22tRUVFxetu2bXXLli1By7NixQpt3769rl69WlVVCwsL9Z///Ge5jjfYdwM3k/lh51SrcRhjgho7FvbsKZ22Z49Lj6eRI0dy7bXXctJJJ3Hbbbfx+eefc/LJJ9OzZ0/69u3L2rVrgdK/uO+++26uvPJK+vXrR/v27XnssceKt1e/fv3i/P369WPIkCF06dKF4cOHo95s4PPmzaNLly706tWLG2+8sdQv+WDq1q1LRkYGmza5+8Y9/PDD/OUvfykehdSuXTvGjBnDQw89BMDEiROZMmUKDRs2BKBhw4aMGDEi6LZffPFFbrrpJtq0acOiRYsi+swefPBBxo4dS5cuXQBIS0vjuuuui2jdeLDAYYwJasOG6NLLY+PGjXz66ac88sgjdOnShY8++ojs7GzGjx/PX/7yl6DrrFmzhvnz5/P5559zzz33cPDgwcPyZGdnM2nSJHJzc1m/fj2ffPIJ+/bt45prruHtt99m6dKlRHJr6W3btvHVV19x2mmnAbBq1Sp69epVKk9mZiarVq1i586d7Nq1i/bt25e53X379vHOO+9w/vnnM2zYMF588cUy1wFYuXLlYfuvSBY4jDFBtWkTXXp5XHzxxaSlpQGwY8cOLr74Yk444QRuueUWVq1aFXSd8847j9q1a9O0aVOOOeYYfvjhh8Py9OnTh1atWlGjRg0yMjLIy8tjzZo1tG/fvri2MGzYsJDl+uijj+jRowctW7bk7LPP5thjj43D0ZZ488036d+/P3Xr1uWiiy5i9uzZHDp0CAg+0qmyjIizwGGMCWrCBKhXr3RavXouPd6OPPLI4td33nkn/fv3Z+XKlcydOzfk9QW1a9cufp2WlkZhYWFMecI59dRTWbZsGatWreLpp58mJycHgK5du7J06dJSeZcuXUq3bt1o2LAh9evXZ/369WVu/8UXX+Sdd94hPT2dXr16UVBQwLvvvgtAkyZN2LZtW3Hen376iaZNmwLQrVu3w/ZfkSxwGGOCGj4cpk6Ftm1BxD1PnerSE2nHjh20bNkSgOeeey7u2z/uuONYv349eXl5ALz00ktlrtOuXTvuuOMOHnjgAQBGjx7N/fffX7yNvLw8Jk6cWDyyacyYMVx//fXs3LkTgN27dx82qmrnzp189NFHbNiwgby8PPLy8pg8eXJxc1W/fv2YPn06AIcOHWLGjBn0798fgFtvvZWJEyfy5ZdfAlBUVMQTTzxRjk8lOhY4jDEhDR8OeXlQVOSeEx00AG677TbGjBlDz549o64hRKJu3br885//ZMCAAfTq1YsGDRpw1FFHlbnetddey4cffkheXh4ZGRk88MADnH/++XTp0oXzzz+fBx98kIyMDACuu+46+vfvT+/evTnhhBM49dRTqVGj9On29ddf5/TTTy9VKxo8eDBz585l//793Hnnnaxbt44ePXrQs2dPOnbsyO9//3sAunfvzqRJkxg2bBjHH388J5xwQkQ1nHixe44bk0JWr17N8ccfn+xiJN3u3bupX78+qsr1119Pp06duOWWW5JdrKQK9t1Iyj3HRWSAiKwVkXUickeQ5SNFZIuI5HiPq7z0DBFZJCKrRGS5iFzit85zIvKN3zoZiTwGY0z18+STT5KRkUG3bt3YsWMH11xzTbKLVKUkbHZcEUkDJgNnAhuBJSIyR1VzA7K+pKo3BKTtAf6gql+JSAtgqYjMV9Xt3vJbVfXVRJXdGFO93XLLLSlfwyiPRNY4+gDrVHW9qh4AZgGDI1lRVb9U1a+8198BPwLNElZSY4wxEUtk4GgJfOv3fqOXFugirznqVRFpHbhQRPoAtYCv/ZIneOs8KiK1A9fx1hslIlkikhXJBT7GGGMik+xRVXOBdFXtDiwApvkvFJHmwHTgClUt8pLHAF2A3kBj4PZgG1bVqaqaqaqZzZpZZcUYY+IlkYFjE+Bfg2jlpRVT1QJV3e+9fQoovoZeRBoCbwFjVfUzv3U2e/Nv7QeexTWJGWOMqSCJDBxLgE4i0k5EagGXAnP8M3g1Cp9BwGovvRbwOvB8YCe4bx1x195fAKxM1AEYY+Krf//+zJ8/v1TapEmTwk7Q169fP3zD6c8991y2b99+WJ67776bhx9+OOy+Z8+eTW5uydicu+66i3feeSeK0geXitOvJyxwqGohcAMwHxcQXlbVVSIyXkQGedlu9IbcLgNuBEZ66UOB04CRQYbdzhSRFcAKoClwX6KOwRgTX8OGDWPWrFml0mbNmhV2vih/8+bNo1GjRjHtOzBwjB8/njPOOCOmbQU69dRTycnJITs7mzfffJNPPvkEgGXLljF69GjeeOMNVq9ezZw5cxg9ejTLly8H4IknnmDBggV8/vnn5OTksHDhQkJdW7dgwQI6d+7MK6+8EjJPoJUrV3LDDTcwY8YMcnNzycrKomPHjuU+3oQNxwVQ1XnAvIC0u/xej8H1WQSuNwOYEWKbp8e5mMakpJtvBm/qpbjJyIBJk0IvHzJkCH/96185cOAAtWrVIi8vj++++45TTz2V6667jiVLlrB3716GDBnCPffcc9j66enpZGVl0bRpUyZMmMC0adM45phjaN26dfFssU8++SRTp07lwIEDdOzYkenTp5OTk8OcOXP44IMPuO+++3jttde49957GThwIEOGDGHhwoWMHj2awsJCevfuzZQpU6hduzbp6emMGDGCuXPncvDgQV555ZXiqcyDiWb69enTpzNx4kTef//9qKZfnzJlCosWLYrohlWJmn492Z3jxpgU0rhxY/r06cPbb78NuNrG0KFDEREmTJhAVlYWy5cv54MPPij+VR7M0qVLmTVrFjk5OcybN48lS5YUL7vwwgtZsmQJy5Yt4/jjj+fpp5+mb9++DBo0iIceeoicnBw6dOhQnH/fvn2MHDmSl156iRUrVlBYWMiUKVOKlzdt2pQvvviC6667rszmsFSZfj2hNQ5jTOUVrmaQSL7mqsGDBzNr1iyefvppAF5++WWmTp1KYWEhmzdvJjc3l+7duwfdxkcffcTvfvc76nnT9w4aNKh42cqVK/nrX//K9u3b2b17N2effXbY8qxdu5Z27drRuXNnAEaMGMHkyZO5+eabAReIAHr16sW///3vkOXp0aMHX331FTfffHPCp1+/9957mTRpEmlpaUmZft1qHCHE+17Lxhhn8ODBLFy4kC+++II9e/bQq1cvvvnmGx5++GEWLlzI8uXLOe+880JOp16WkSNH8vjjj7NixQrGjRsX83Z8fJMQhpuWPdWmX7fAEcTMmTBqFOTnuzst5+e79xY8jCm/+vXr079/f6688sriTvGdO3dy5JFHctRRR/HDDz8UN2WFctpppzF79mz27t3Lrl27mDt3bvGyXbt20bx5cw4ePMhMv3/aBg0asGvXrsO2ddxxx5GXl8e6desAmD59Or/5zW9iOrZUmX7dAkcQFXWvZWNS1bBhw1i2bFlx4PBNHd6lSxcuu+wyTjnllLDr//KXv+SSSy6hR48enHPOOfTu3bt42b333stJJ53EKaecUqoj+9JLL+Whhx6iZ8+efP11yUQUderU4dlnn+Xiiy/mxBNPpEaNGlx77bUxH1sqTL9u06oHUaOGq2kEEnH3JTCmqrJp1U0olWZa9aqqIu+1bIwxVY0FjiAq8l7LxhhT1VjgCCJZ91o2piKkQvO0iU603wm7jiOE4cMtUJjqp06dOhQUFNCkSZOEj/U3VYOqUlBQQJ06dSJexwKHMSmkVatWbNy4EbtHjfFXp04dWrVqFXF+CxzGpJAjjjiieN4kY2JlfRzGGGOiYoHDGGNMVCxwGGOMiUpKXDkuIluA/DKyNQW2lpGnOrLjTi123KmlvMfdVlWbBSamROCIhIhkBbu0vrqz404tdtypJVHHbU1VxhhjomKBwxhjTFQscJSYmuwCJIkdd2qx404tCTlu6+MwxhgTFatxGGOMiYoFDmOMMVFJ+cAhIgNEZK2IrBORO5JdnkQSkWdE5EcRWemX1lhEFojIV97z0cksY7yJSGsReU9EckVklYjc5KVX6+MGEJE6IvK5iCzzjv0eL72diCz2vvMviUitZJc13kQkTUSyReRN7321P2YAEckTkRUikiMiWV5a3L/rKR04RCQNmAycA3QFholI1+SWKqGeAwYEpN0BLFTVTsBC7311Ugj8WVW7Ar8Crvf+xtX9uAH2A6erag8gAxggIr8CHgAeVdWOwDbgj8krYsLcBKz2e58Kx+zTX1Uz/K7fiPt3PaUDB9AHWKeq61X1ADALGJzkMiWMqn4I/BSQPBiY5r2eBlxQkWVKNFXdrKpfeK934U4mLanmxw2gzm7v7RHeQ4HTgVe99Gp37CLSCjgPeMp7L1TzYy5D3L/rqR44WgLf+r3f6KWlkl+o6mbv9ffAL5JZmEQSkXSgJ7CYFDlur8kmB/gRWAB8DWxX1UIvS3X8zk8CbgOKvPdNqP7H7KPAf0VkqYiM8tLi/l23+3GYYqqqIlItx2eLSH3gNeBmVd3pf/e76nzcqnoIyBCRRsDrQJfkliixRGQg8KOqLhWRfkkuTjL8WlU3icgxwAIRWeO/MF7f9VSvcWwCWvu9b+WlpZIfRKQ5gPf8Y5LLE3cicgQuaMxU1X97ydX+uP2p6nbgPeBkoJGI+H40Vrfv/CnAIBHJwzU9nw78nep9zMVUdZP3/CPuh0IfEvBdT/XAsQTo5I24qAVcCsxJcpkq2hxghPd6BPBGEssSd1779tPAalV9xG9RtT5uABFp5tU0EJG6wJm4Pp73gCFetmp17Ko6RlVbqWo67v/5XVUdTjU+Zh8ROVJEGvheA2cBK0nAdz3lrxwXkXNxbaJpwDOqOiG5JUocEXkR6IebavkHYBwwG3gZaIOben6oqgZ2oFdZIvJr4CNgBSVt3n/B9XNU2+MGEJHuuM7QNNyPxJdVdbyItMf9Gm8MZAO/V9X9yStpYnhNVaNVdWAqHLN3jK97b2sCL6jqBBFpQpy/6ykfOIwxxkQn1ZuqjDHGRMkChzHGmKhY4DDGGBMVCxzGGGOiYoHDGGNMVCxwGBMjETnkzULqe8RtokQRSfefxdiYysSmHDEmdntVNSPZhTCmolmNw5g48+6J8KB3X4TPRaSjl54uIu+KyHIRWSgibbz0X4jI6959M5aJSF9vU2ki8qR3L43/eld/IyI3evcXWS4is5J0mCaFWeAwJnZ1A5qqLvFbtkNVTwQex81MAPAPYJqqdgdmAo956Y8BH3j3zfglsMpL7wRMVtVuwHbgIi/9DqCnt51rE3NoxoRmV44bEyMR2a2q9YOk5+FuoLTem2Dxe1VtIiJbgeaqetBL36yqTUVkC9DKfwoMbwr4Bd7NdxCR24EjVPU+EfkPsBs3Xcxsv3tuGFMhrMZhTGJoiNfR8J9L6RAlfZLn4e5c+Utgid+sr8ZUCAscxiTGJX7Pi7zXn+JmbAUYjpt8EdztPK+D4hsvHRVqoyJSA2itqu8BtwNHAYfVeoxJJPulYkzs6np31/P5j6r6huQeLSLLcbWGYV7a/wLPisitwBbgCi/9JmCqiPwRV7O4DthMcGnADC+4CPCYd68NYyqM9XEYE2deH0emqm5NdlmMSQRrqjLGGBMVq3EYY4yJitU4jDHGRMUChzHGmKhY4DDGGBMVCxzGGGOiYoHDGGNMVP4fhZU2O8ojFe4AAAAASUVORK5CYII=",
thecml's avatar
thecml committed
142
      "text/plain": [
143
       "<Figure size 432x288 with 1 Axes>"
144
      ]
145
     },
146
147
148
     "metadata": {
      "needs_background": "light"
     },
149
     "output_type": "display_data"
thecml's avatar
thecml committed
150
151
152
    },
    {
     "data": {
153
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAA5eElEQVR4nO3deXxU1d348c83YZNFUcClQAgqiFAllAh1QdFWRUUpFhWkPlBbUdRafGotiltR+vxaW2ut1D64VyPoY9WCorgr7gQDCEGQ3SggRoJg2PP9/XHuJJPJncnMMHcmyXzfr9e8Zu6527mTyf3ec86954iqYowxxkTKyXQGjDHGNEwWIIwxxviyAGGMMcaXBQhjjDG+LEAYY4zxZQHCGGOMLwsQZp+IyIsiMibVy2aSiKwRkR8HsF0VkSO9z/8UkZvjWTaJ/YwWkZeTzacxIRYgspCIbAt7VYnI9rDp0YlsS1XPUtVHU71sU6eqV6jq7fu6HRHJ94JJs7BtF6nqGfu6bZ99DfZ+L9tEZKuILBORn0fkI/Q7WiMiE+PYpojIKhEp9ZlXJ1CLyFgReSdsuoWI3CYin4nId946D4lIfgoOOes1q38R09SoatvQZxFZA/xSVV+NXE5EmqnqnnTmzTR4X6pqFxERYBjwtIh8CFR689ur6h4RKQTeEpH5qvpKjO2dDBwMNBOR41R1XoL5eRroAlwMlABtgJ8BPwIeTHBbJoKVIEw17wqxTER+JyIbgIdF5EAReV5ENonIZu9zl7B13hSRX3qfx4rIOyLyZ2/Z1SJyVpLLdheRt70r1VdFZKqIPB4l3/Hk8XYRedfb3ssi0jFs/iUislZEykVkUozvZ6CIbBCR3LC04SKyyPs8QETeF5EKEVkvIveKSIso23pERO4Im/6tt86XInJpxLLniEiJiHwrIp+LyG1hs9/23iu8K/fjfa6yTxCReSKyxXs/Id7vJhp1ngM2A7195hcDS4CCejY1BvgPMNv7HDevdHE6MExV56nqHlXdoqpTVdWCQwpYgDCRDgUOAroB43C/kYe96TxgO3BvjPUHAsuAjsCfgAe9q81El30C+AjoANwGXBJjn/Hk8WLg57ir1RbAdQAi0hu4z9v+97z9dcGHqn4IfAecFrHdJ7zPe4FrveM5HncVe2WMfOPlYYiXn9OBHkBk+8d3wH8B7YFzgPEi8hNv3snee3tVbauq70ds+yDgBeAe79juAl4QkQ4Rx1Dnu6knzzkiMtzL0yc+838IfB9YEWMbrYERQJH3GhktoEbxY+AjVf08gXVMAixAmEhVwK2qulNVt6tquar+W1UrVXUrMAU4Jcb6a1X1flXdCzwKHAYcksiyIpIHHAfcoqq7VPUdYGa0HcaZx4dVdbmqbgeeoubKdgTwvKq+rao7gZu97yCa6cAoABFpB5ztpaGq81X1A+9Kdg3wvz758HOhl7/FqvodLiCGH9+bqvqJqlap6iJvf/FsF1xA+UxVH/PyNR34FDg3bJlo342f74lIBfA1cCtwiaouC5v/tYhsB94H/gE8F2Nb5wM7gZdxQay5l994dQDWJ7C8SZAFCBNpk6ruCE2ISGsR+V+vCuZbXJVG+/BqlggbQh9UNVQv3TbBZb8HfBOWBhD1KjHOPG4I+1wZlqfvhW/bO0GXR9sXrrRwvoi0xJ3gPlbVtV4+enrVWxu8fPwBV5qoT608AGsjjm+giLzhVaFtAa6Ic7uhba+NSFsLdA6bjvbd+PlSVdur6kGqWqCqMyLmd/TW/w0wGHfSj2YM8JQXuHYA/6Z2NdMen/WbA7u9z+W4iwoTEAsQJlJk976/AY4CBqrq/tRUaUSrNkqF9cBBXhVESNcYy+9LHteHb9vbZ4doC6tqKe4Eexa1q5fAVVV9CvTw8nFjMnnAVZOFewJXguqqqgcA/wzbbn3dMX+Jq3oLlwd8EUe+kqKqe1X1LmAHUarYvDai04CfeQF1A640d3ZYG8g6ID9i1e7UBLxXgQHh7U0mtSxAmPq0w9XpV3j12bcGvUPvirwYuE3cbYzHU7tKJJV5fBoYKiInefXfk6n//+IJ4Ne4QPR/Efn4FtgmIr2A8XHm4SlgrIj09gJUZP7b4UpUO0RkAC4whWzCVYkdHmXbs4GeInKxiDQTkYtwjcrPx5m3ffH/gOtFpJXPvEuA5bjAXuC9egJleFV4wJPABBHpJU4hcCkwA8C78+4V4FkR6e8dXzsRuSKyod8kxwKEqc/dwH64OucPgJfStN/RuIbecuAO3MliZ5Rl7ybJPKrqEuAq3El/Pe6unLJ6Vgu1Abyuql+HpV+HO3lvBe738hxPHl70juF1XKPu6xGLXAlMFpGtwC24gBJatxLX5vKud/fUDyO2XQ4MxZWyyoHrgaER+Q7KC7jv8zKfeWOAf6jqhvAXrnQUqma6H3fzwSxgC/AvYJKqhv99R+CC4JPeMouBQlzpwuwjsQGDTGMgIk8Cn6pq4CUYY4xjJQjTIInIcSJyhHc75RDcQ1nPZThbxmQVe5LaNFSHAs/gGozLgPGqWpLZLBmTXayKyRhjjC+rYjLGGOOryVQxdezYUfPz8zOdDWOMaVTmz5//tap28pvXZAJEfn4+xcXFmc6GMcY0KiIS+aR9NatiMsYY48sChDHGGF8WIIwxxviyAGGMMcaXBQhjjDG+Ag0QIjJE3MDmK8RnAHMRyfP6uS8RkUUicnbYvBu89ZaJyJlB5tMYY5JVVAT5+ZCT496LijKdo9QJ7DZXb7CWqbhhFMuAeSIy0+tPP+Qm3IAh93lDP84G8r3PI4E+uAFPXhWRnt7IY8YY0yAUFcG4cVDpDW21dq2bBhg9OnP5SpUgSxADgBWqukpVd+H6cB8WsYwC+3ufD8ANboK33Axv2MvVuC6QBwSYV2OMSdikSTXBIaSy0qU3BUEGiM7UHkaxjNrDHIIbe/dnIlKGKz38KoF1EZFxIlIsIsWbNm1KVb6NMY1Mpqp51q1LLD3Vgj7uTDdSjwIeUdUuuMHfHxORuPOkqtNUtVBVCzt18n1S3BiTYg2tzj1UzbN2LajWVPOkI195kYPD1pOeSuk47iADxBfUHme3C3XHwf0F3uhYqvo+0Ao36Hk86xpj0iyTJ+NokqnmSVWQmzIFWreunda6tUtPJb/8pqV6S1UDeeEawFfhBhlvASwE+kQs8yIw1vt8NK4NQnCN0wuBlt76q4DcWPvr37+/GmOC1a2bqgsNtV/dumUuTyL+eRJRffxxlzcR9/744+7VunXtZVu3dunJ8NtHshLJr98xh447EUCxRjuPR5uRiheu2mg5sBI3liy4QeHP8z73Bt71gsEC4IywdSd56y0DzqpvXxYgjAlerJNxqiR6wo0WtDp08D+xduiQniAX7ThipSeS39zc1BxHxgJEOl8WIIwJXtAliGSu7hM9sSZ75Z1I4IqWp/Hjox9ftO821isVJSELEMaYlEh19UykZAOQ38k7Wmkn2ivWPhI97mjHEeuqP5n8pqJ6ywKEMSYhsU48qapzT+SkHrq6T2TfiVY9JbOtaEEl0ZN96HhSld9EWIAwxsQt2VJCKqpgYrUPJJqvWMsnGgATbXtJpgSRbH73lQUIYwIQ5D9tJvedTDVPoie3ZK6Wk81XIt9TMoErke3EaoNIJr+pYAHCmBQLui4+2X2n4gSTzJ1KiZ7wY1W1RDuGdNxBlcpqnkTvYsoUCxCmUWto/1CqmX0eIOi66mSOLdE692Ru0UzHd57oMxVNgQUI02hl8ko9lnRczSa672ivRKtgkvnO03GLZjp+Cw3xQcCgWYAwjVZD/YdNR314ovtOtNomlVVVyTY6J/p9BH0V31AvSIJkAcI0Wpm8Uo8llXfUpGrf0U7G6XqiOB3dWqRDU61KisYChGm0GmoJQjU19+Sn8ko6XU8UJyrbTriNjQUI02g1xitQP7HaDVJ5R1LQTxSbpscChGnU0nEFmsp9JHLff7S7eVL59GymntA1jYMFCNMoJNMwmqouH4JuH4j2gFQiV/bJXt1n6gld0zhYgDANXroafRO5uk+mfSDRbSVzR1IyLBCYaCxAmAYl0ZO0n1R2BxHrhJxoEEr0rqtUde1gTLJiBQhx8xu/wsJCLS4uznQ2TD1CQ1aGD5XYunXdoRNDRKCqqm56To47Zca7PLihGteurZuemwt798af3q0brFmT2D5irRMaPnLdOjeWcWi4Sr/vado0GD3afzvGJENE5qtqod+8IMekRkSGiMgyEVkhIhN95v9VRBZ4r+UiUhE2708iskRElorIPSIiQebVpEe0cXRzc/2XT3RQ+FiDxa9b55++d6//uMJ+wSHWdiC5MYpHj3bBo6rKvY8e7V7TprnAIuLeLTiYtItWtNjXF5CLGzL0cGrGpO4dY/lfAQ95n0/ADUWa673eBwbH2p9VMTUOydzu6SeV3UEk2j5QXzWP1febxoRMtEEAxwNzwqZvAG6Isfx7wOlh684H9gNaA8XA0bH2ZwGicUhlg3CquoNIx9PPxjRUmQoQI4AHwqYvAe6Nsmw3YD2QG5b2Z6AC2AJMibLeOC94FOfl5QX1/ZkUStdJN1VdLVtpwDR1jSFA/A74e9j0kcALQFvv9T4wKNb+rAQRrKAfJEslu/I3Jn6xAkSQjdRfAF3Dprt4aX5GAtPDpocDH6jqNlXdBryIq3YyGRC682jtWne6XbvWTRcVJbc9v0bZVIrWED5pUmr3Y0xTF2SAmAf0EJHuItICFwRmRi4kIr2AA3GlhJB1wCki0kxEmgOnAEsDzKuJobGdcKPdZRTr7iNjTF2BBQhV3QNcDczBndyfUtUlIjJZRM4LW3QkMMMr6oQ8jbsD6hPc3U8LVXVWUHk1sTW2E24yt8AaY+pqFuTGVXU2MDsi7ZaI6dt81tsLXB5k3kz88vL8H/5qqCfcKVP8HzKL9SyCMaauQB+UM5lXVOSe7s3Jce/JtBvEevgrFdtPNXvIzJjUsADRhKWqcTnaCRdib7+oyJUyRNIfPIJuCE+X9eth7lzYsyfTOTFZKdrtTY3tlU23ue7ZE99yyTyUlqpR0h57TLV589rprVpF394336j+61+q55+vesYZqrffrvrWW6rbt8d3rE3Fnj2q772nOmmSar9+Nd/d4YerTp2qWlmZubzt3Kn66quqv/616vHHqz75ZObykm5ff616xx2q+fmqf/pTsPvatk31u+/i/z/fV1hnfU3Dp5/Cz38OCxfCuefCqFEwZAi0auW/fLQO7aBuB3mtW8OYMfDoo/F3EBdr+23awHff1U1v2RLuuQcuusjN/89/4Jln4M033VVy587QoQN88onbdsuWMHAgnHIKnH02/PCH/vvbV7t3u2N/911X6gi99u51723buu/8rLOif9/74sMP3ffy0kvwzTeub6oTTnDH3KUL/P3v8NFH0KkT/OpXcNVVcNBB8W9/7164805YsQLOOANOPx0OPLD+9crL4cUXYdYsl7dvv3V/k0MPhc8/h8ceg4svTv64g7ZlC0ydWvN3Va393rq1+00NHgzHHQctWtRef9kyuPtu99vYvt2ViMvK4NVX4dRT69//t9/CO++4/YnUfu3a5Urda9bA6tU1r4qKmvWbN3ffd6tW7vWjH7m7B3v0SN13FKuzPgsQjUBVlTtB/Pa3NSesnBz3vv/+cP75MHKk+/E0C7vtIOjeS6Ntv0UL9+OPJXyZnj3dMQwfDoWF7ti++cb9Y739Nrz1Fnz8sTveP/wBJk50/2CpsHcvTJ8Ot90GK1e6E1+rVi4POTnuO8nJgY0bXZ7atYNhw1yAO/1098+7L6qq4E9/gptugvbt4ZxzXFA444zaJ3BVV9X0xz/C7NkuAP/yl3DjjXDwwbH3sXmzO4m/9JLL/9at7riOP94FvLPOgqOPhuXLYfFiWLLEvRYvhlWr3L4POQSGDnVB8sc/dtsdOtT9fR59FH72s337HuJVVgb33w9du8J550U/9ooK+Nvf3Mm9ogKOOabm7ypS8755M5SWunX22899J6ecAkcd5YLfCy+4v/HPfgYTJrjf/HHHufUWLHC/l2i++cZta/Hi2MfUqpXbbvfu7tW1q8vfjh3utXOne6+ogGefddOjR7vfTM+eiX1/fmIFiIxXDaXq1VSrmFavVh082FUz5OTUrrJp2VL15JNVDzjATbdpo9q5s2qPHqoFBao9e9ZdJ5lRzBIZy0DE5atTJ/9t5eWpfvSR6n//t6tGWrJEtaqq/u9hyxbVUaPcNi67THX37tjL793rqkPmzlXdtKnu/Koq1X//W7VPH7fNvn1VZ82Knpfdu1Vffln1F79QPfBAt84BB6iOGaM6fbrqhg31H0OkDRtcdRqoXnihakVFfOstWqT6X/+l2qyZavv2qvfeG706orTU/R6aN1edNs0dx7vvqt50k+oPfuD/N8rNVT36aNURI1QnT1b98EP3fUb67jvV005zf/NHH038+BOxfr2r2mrZsiafOTmqgwap3nWX+z9RVS0vV735ZtX993fLDBumOn9+7G1v2qT6zDOq11zjfgehDiU7dVK97TbVjRtrL//JJ6r77eeOPdr3/u23qgMGqLZo4X4fH36o+sEHqu+/76oQ333Xpa1fH9/vP2TDBtXf/MbtPydHdfRo1aVL41/fDzZgUONTVaV6//2qbduqtmsXewCZHTtUn3vO/cAvvVR15EjVc89V/dGPVI88siZIHHZYcuMjx+rm6vHH3fxQcOjYUXXBgmC6u9i7V/XGG922hgxx/4R+3nxTtX//2vvu0EH1hBPc9zNlimphoUs/6ihXl+53Aoxm507V2bNdcGjfvmYfffqoXn21O9mUl8fexmuvqR56qGuX+ec/EztJhCxdqvrjH7t99+vnTjzh/vMf99s55BDVd97x38b69aoPP6x6663uRLZokfs9xeu779zvTET1kUcSP4Y331S95BLVP/zBBfTIIPn116rXX+9+O7m5LkCvXq1aUqJ6yy2qxxxT8/0fe6w7XnBtWSUliedH1bWHvf127Pavhx5y+7n11rrzKivdRV1urvsbBGHDBtXrrnPfi4j7LSbzG1K1AJFRu3a5f85771WdN6/+P+LWre5q7KST3F/n1FPdP0SiI5WFW7rUBYdOndzVT7ST92WX+QeJ1q3dieiGG1Tvu89dof385y7tqKNcyQXc1Vz41VZQfS5Nm+byWVCgWlZWk/7pp+6KEVS7dnUnrBdfVP3rX1Uvv1z1lFPcyTIUWB96qP6SSH327HF/1z/+UfXMM2u+VxEXnIcNcw3OTzyhunChO6HefLOb36uXOyHvi6oq1aeeciVHcAFw40Z35Q8uEH7++b7toz6Vlaqnn+6O6aGH4l/vnXfclXDbtrV/b716uRLShAnuhC/irpSXL/ffzmefqd55pytNjxy5799pPKqqXB5FVF95pSZ9507Vc85x6UVFwedj40YXQCdMSH4bFiDSqKrKnQjuusv9UCJ//J07q155peqcOe7HpOquXl9/3V0FhE62hx/ugkroyjbZsQlCli1T/d733BX+okV1T9733ONOJiI1V8WHHeaCxuWXu+qIZs1qTn7f+57qwIGqP/2p+3H+4x81x5MOL77ovtsuXVTfeEP1qqtc0GjXzl2Nxrrbp6Ji3wNDNDt3uhPfHXeoXnCBq6rxC7pjx7q7VVJl61Z3omjWrObusUsuSd9dT5WVrrpMRPXuu+u/EJo/31UD9ezproa/+cZV4d1xh+p559UE8p/+VHXx4vQcQ6K2bVPt3Vv14INVv/zSXSxceKHL9//+b6ZzFz8LEGny4IM1P2xwP/4rr3R13StXuiva4cNrrjL3399Nh6po2rVzRei33677DxaryibeK/Xly12A6tDBVQOFfPihCwZt28YuEldWuqvRXbv29ZtKjZISF6hC1WPjx9etL24IduxwQfmJJ1zd/9NPB7ev0lJ3Uv3b35KvckjW9u01Jbjhw91J309pqbtQyctTXbfOf5mqqtQG0KAsWeL+D085xZXewJVmGhMLEAGrqnLVCKFqlocfjv7DV3Un2pkzXTDIy3NVE0884aofYvELBInW9X/2mbvqPugg1Y8/dsXgli3d/d3pKJqn2rp1qhMnupOOyby9e1X/8hdXisnLc42x4VatckH9kEOiVxk1No88UvO/d/PNmc5N4ixABGjHDtWLL3bf5C9+kdjVdSrq6JOpelq50v3zhgLLySerfvVV4vs2JpqPPnLVpLm5rspv717VL75waQce2DgvRmK57TZXPZbuUlsqWIAIyDffuKIlJP7jSEWVkWryjderVrk7QK68Mr1tByZ7VFSoXnSR+z2efrqrr2/b1t3uaRqOWAHCHpRL0urV7oGmlSvh4YcT7+sn2kNmHTq4JzbjfZo52naiPdxmTDqpwgMPwDXXuM8vveSeWjYNR6wH5ayzviQsWOAez9+wAV5+ObmO4KKNpVBentjgPLF6WjUm00Tgsstc1ykffWTBobGxAJGEW291V0PvvZf8Dz7RsRSiBRTr2to0BkceCccem+lcmEQFGiBEZIiILBORFSIy0Wf+X0VkgfdaLiIVYfPyRORlEVkqIqUikh9kXhOxeLHrqOvoo5PfRrQr/w4d/JePFVCaStfWxpiGJbAAISK5wFTgLKA3MEpEeocvo6rXqmqBqhYAfweeCZv9L+BOVT0aGAB8FVReE7F9u2t/6N27/mVjiXbl/7e/WZWRMaZhCHLI0QHAClVdBSAiM4BhQGmU5UcBt3rL9gaaqeorAKq6LcB8JmTZMle9tK8BAlyQiHa1P2mSq1bKy3PBwUoFxph0CzJAdAY+D5suAwb6LSgi3YDuwOteUk+gQkSe8dJfBSaqG6s6fL1xwDiAvDQNkBzqGnhfqpfqEytwGGNMujSURuqRwNNhAaAZMAi4DjgOOBwYG7mSqk5T1UJVLezUqVNaMrp0qetLP5UDdhhjTEMUZID4AugaNt3FS/MzEpgeNl0GLFDVVaq6B3gO+EEQmUxUaam7I2NfB4oxxpiGLsgAMQ/oISLdRaQFLgjMjFxIRHoBBwLvR6zbXkRCxYLTiN52kValpcFWLxljTEMRWIDwrvyvBuYAS4GnVHWJiEwWkfPCFh0JzNCwR7q9qqbrgNdE5BNAgPuDymu8du1yY/qmooHaGGMauiAbqVHV2cDsiLRbIqZvi7LuK0CDerRmxQrYs8cChDEmOzSURupGIR13MBljTENhASIBS5e6h9p69cp0TowxJngWIBJQWup6T4180tkYY5oiCxAJsDuYjDHZxAJEnPbudd1sWAO1MSZbWICI0+rVsHOnBQhjTPawABEnu4PJGJNtLEDEyQKEMSbbWICI09Kl0LkzHHBApnNijDHpYQEiTqWl1v5gjMkuFiDiUFXlShBWvWSMySYWIOJQVgbffZd8CaKoyD1gl5Pj3ouKUpk7Y4wJRqCd9TUVoQbqZAJEURGMGweVlW567Vo3DTZqnDGmYbMSRBz25Q6mSZNqgkNIZaVLN8aYhswCRByWLoVOnaBjx+jLRKtGWrfOf/lo6cYY01BYFVMc6ruDKVY1Ul6em46Ul5f6fBpjTCoFWoIQkSEiskxEVojIRJ/5fxWRBd5ruYhURMzfX0TKROTeIPMZi2r9nfTFqkaaMqVu76+tW7t0Y4xpyAIrQYhILjAVOB0oA+aJyExVrR5bWlWvDVv+V0C/iM3cDrwdVB7jsXEjVFTELkHEqkYKNURPmuSm8/JccLAGamNMQxdkCWIAsEJVV6nqLmAGMCzG8qOA6aEJEekPHAK8HGAe6xXPHUzRqotC6aNHw5o17nmKNWssOBhjGocgA0Rn4POw6TIvrQ4R6QZ0B173pnOAvwDXxdqBiIwTkWIRKd60aVNKMh0pnjuYrBrJGNMUNZS7mEYCT6vqXm/6SmC2qpbFWklVp6lqoaoWdurUKZCMLV3q+l867LDoy4weDdOmQbdubkjSbt3ctJUUjDGNWZB3MX0BdA2b7uKl+RkJXBU2fTwwSESuBNoCLURkm6rWaegOWugOJpHYy40ebQHBGNO0BBkg5gE9RKQ7LjCMBC6OXEhEegEHAu+H0lR1dNj8sUBhJoIDuAAxdGgm9myMMZkVWBWTqu4BrgbmAEuBp1R1iYhMFpHzwhYdCcxQVQ0qL8kqL4evvrJeXI0x2SnQB+VUdTYwOyLtlojp2+rZxiPAIynOWlyWLnXvFiCMMdmooTRSN0g2ipwxJptFDRAicqaIjPBJHyEipwebrYZh6VJ3u6p1i2GMyUaxShC3AG/5pL8JTA4kNw1MqIuNHCtnGWOyUKxTX0tVrfP0map+DbQJLksNx5IlVr1kjMlesQLE/iJSpxFbRJoD+wWXpYZh0yb44gsoKMh0TowxJjNiBYhngPtFpLq0ICJtgX9685q0khL33i+y+0BjjMkSsQLETcBGYK2IzBeRj4HVwCZvXpMWChBWgjDGZKuoz0F4D7pNFJHfA0d6yStUdXtacpZhJSWuT6WDDsp0TowxJjOiBggROT8iSYH2IrJAVbcGm63MKymx6iVjTHaL9ST1uT5pBwHHisgvVPX1gPKUcdu2wWefWed7xpjsFquK6ed+6d7YDU8BA4PKVKYtXOiGGrUShDEmmyX8CJiqrgWaB5CXBiPWHUxFRZCf7x6ey89308YY0xQl3FmfiBwF7AwgLw1GSQl07AidI8a/KyqCceOgstJNr13rpsGqo4wxTU+sRupZuIbpcAcBhwGXBJmpTAs1UEcOEjRpUk1wCKmsdOkWIIwxTU2sEsSfI6YVKAc+U9VdwWUps3btgsWL4dpr685bt85/nWjpxhjTmMVqpPbrqA8ROUlERqnqVX7zG7vSUti927/9IS/PVSv5pRtjTFMTVyO1iPQTkTtFZA1wO/BpnOsNEZFlIrJCROoMGSoifxWRBd5ruYhUeOkFIvK+iCwRkUUiclH8h7RvYjVQT5niuv8O17q1SzfGmKYmVhtET2CU9/oaeBIQVT01ng2LSC4wFTgdKAPmichMVS0NLaOq14Yt/ysgdFquBP5LVT8Tke8B80VkjqpWJHJwySgpgTZtoEePuvNC7QyTJrlqpbw8Fxys/cEY0xTFaoP4FJgLDFXVFQAi4lMzH9UAXNccq7x1ZwDDgNIoy48CbgVQ1eWhRFX9UkS+AjoBFQnsPyklJdC3b/QxIEaPtoBgjMkOsaqYzgfWA2+IyP0i8iNAYiwfqTPwedh0mZdWh/fwXXegztPZIjIAaAGs9Jk3TkSKRaR406Y6Q1ckrKoKFiywDvqMMQZiBAhVfU5VRwK9gDeACcDBInKfiJyR4nyMBJ5W1b3hiSJyGPAY8HNVrfLJ4zRVLVTVwk6dOu1zJlaudN1s2BPUxhgTRyO1qn6nqk+o6rlAF6AE+F0c2/4C6Bo23cVL8zMSmB6eICL7Ay8Ak1T1gzj2t89sDAhjjKmRUFcbqrrZu2r/URyLzwN6iEh3EWmBCwIzIxcSkV7AgcD7YWktgGeBf6nq04nkcV+UlECzZvD976drj8YY03Al3BdTvLzxJK4G5gBLgadUdYmITBaR88IWHQnMUNXwp7YvBE4GxobdBlsQVF5DSkqgd29o2TLoPRljTMMntc/LjVdhYaEWFxcnvb4qHHoonHUWPPJI6vJljDENmYjMV9VCv3kJlyBEJEdEmtyNnuvXw1dfWfuDMcaERA0QIrK/iNwgIveKyBni/ApYhasCalKsgdoYY2qL9aDcY8BmXOPxL4Ebcc9B/ERVFwSftfQKBQh7BsIYY5xYAeJwVT0GQEQewD00l6eqO9KSszRbsACOOAL23z/TOTHGmIYhVhvE7tAH7wG2sqYaHKBmDAhjjDFOrADRV0S+FZGtIrIVODZs+tt0ZTAdtmyBVassQBhjTLhY40HkpjMjmbRggXu3AGGMMTVidffdCrgCOBJYBDzkPfzW5NgdTMYYU1esKqZHgULgE+Bs4C9pyVEGlJS4h+QOPTTTOTHGmIYj1l1MvcPuYnoQ+Cg9WUo/a6A2xpi64r2LqUlWLQHs2OHGobYAYYwxtcUqQfQNu1tJgP28aQFUVZvEEwMVFXDOOTBoUKZzYowxDUvW38V06KHwn/9kOhfGGNPwBNbdtzHGmMbNAoQxxhhfFiCMMcb4CjRAiMgQEVkmIitEZKLP/L+GjRi3XEQqwuaNEZHPvNeYIPNpjDGmrsAChIjkAlOBs4DewCgR6R2+jKpeq6oFqloA/B14xlv3IOBWYCAwALhVRA4MKq9+ioogPx9yctx7UVE6926MMZkXZAliALBCVVep6i5gBjAsxvKjgOne5zOBV1T1G1XdDLwCDAkwr7UUFcG4cbB2rRuKdO1aN21BwhiTTYIMEJ2Bz8Omy7y0OkSkG9AdeD2RdUVknIgUi0jxpk2bUpJpgEmToLKydlplpUs3xphs0VAaqUcCT3vjTsRNVaepaqGqFnbq1CllmVm3LrF0Y4xpioIMEF8AXcOmu3hpfkZSU72U6Lopl5eXWLoxxjRFQQaIeUAPEekuIi1wQWBm5EIi0gs4EDf2dcgc4AwROdBrnD7DS0uLKVOgdevaaa1bu3RjjMkWgQUIr4O/q3En9qXAU6q6REQmi8h5YYuOBGaoqoat+w1wOy7IzAMme2lpMXo0TJsG3bqBiHufNs2lG2NMtpCw83KjVlhYqMXFxZnOhjHGNCoiMl9VC/3mNZRGamOMMQ2MBQhjjDG+LEAYY4zxZQHCGGOMLwsQxhhjfFmAMMYY48sChDHGGF8WIIwxxviyAGGMMcaXBQhjjDG+LEAYY4zxZQHCGGOMLwsQxhhjfFmAMMYY48sChDHGGF8WIIwxxvgKNECIyBARWSYiK0RkYpRlLhSRUhFZIiJPhKX/yUtbKiL3iIgEmVdjjDG1NQtqwyKSC0wFTgfKgHkiMlNVS8OW6QHcAJyoqptF5GAv/QTgROBYb9F3gFOAN4PKrzHGmNqCLEEMAFao6ipV3QXMAIZFLHMZMFVVNwOo6ldeugKtgBZAS6A5sDHAvBpjjIkQZIDoDHweNl3mpYXrCfQUkXdF5AMRGQKgqu8DbwDrvdccVV0aYF6NMcZECKyKKYH99wAGA12At0XkGKAjcLSXBvCKiAxS1bnhK4vIOGAcQF5eXrrybIwxWSHIEsQXQNew6S5eWrgyYKaq7lbV1cByXMAYDnygqttUdRvwInB85A5UdZqqFqpqYadOnQI5CGOMyVZBBoh5QA8R6S4iLYCRwMyIZZ7DlR4QkY64KqdVwDrgFBFpJiLNcQ3UVsVkjDFpFFiAUNU9wNXAHNzJ/SlVXSIik0XkPG+xOUC5iJTi2hx+q6rlwNPASuATYCGwUFVnBZVXY4wxdYmqZjoPKVFYWKjFxcWZzoYxxjQqIjJfVQv95tmT1MYYY3xZgDDGGOPLAoQxxhhfFiCMMcb4sgBhjDHGlwUIY4wxvixAGGOM8WUBwhhjjC8LEMYYY3xZgDDGGOPLAoQxxhhfFiCMMcb4sgBhjDHGlwUIY4wxvjI95KgxJgN2795NWVkZO3bsyHRWTJq0atWKLl260Lx587jXsQBhTBYqKyujXbt25OfnIyKZzo4JmKpSXl5OWVkZ3bt3j3s9q2IyJgvt2LGDDh06WHDIEiJChw4dEi4xBhogRGSIiCwTkRUiMjHKMheKSKmILBGRJ8LS80TkZRFZ6s3PDzKvxmQbCw7ZJZm/d2BVTCKSC0wFTgfKgHkiMlNVS8OW6QHcAJyoqptF5OCwTfwLmKKqr4hIW6AqqLwaY4ypK8gSxABghaquUtVdwAxgWMQylwFTVXUzgKp+BSAivYFmqvqKl75NVSsDzKsxJoaiIsjPh5wc915UtG/bKy8vp6CggIKCAg499FA6d+5cPb1r166Y6xYXF3PNNdfUu48TTjhh3zLpefPNNznggAMoKCjg6KOP5ve//32d9F69enHdddfF3M6ECRPo3LkzVVU117q33XYbf/7zn2stl5+fz9dffw3Ahg0bGDlyJEcccQT9+/fn7LPPZvny5Sk5rngE2UjdGfg8bLoMGBixTE8AEXkXyAVuU9WXvPQKEXkG6A68CkxU1b3hK4vIOGAcQF5eXhDHYEzWKyqCceOg0rtEW7vWTQOMHp3cNjt06MCCBQsAd5Js27ZtrRPsnj17aNbM//RUWFhIYWFhvft47733ksucj0GDBvH888/z3XffUVBQwLnnnlsrffv27fTr14/hw4dz4okn1lm/qqqKZ599lq5du/LWW29x6qmn1rtPVWX48OGMGTOGGTNmALBw4UI2btxIz549U3ZssWS6kboZ0AMYDIwC7heR9l76IOA64DjgcGBs5MqqOk1VC1W1sFOnTmnKsjHZZdKkmuAQUlnp0lNp7NixXHHFFQwcOJDrr7+ejz76iOOPP55+/fpxwgknsGzZMsBduQ8dOhRwweXSSy9l8ODBHH744dxzzz3V22vbtm318oMHD2bEiBH06tWL0aNHo6oAzJ49m169etG/f3+uueaa6u1G06ZNG/r378+KFStqpe+3334UFBTwxRdf+K735ptv0qdPH8aPH8/06dPj+j7eeOMNmjdvzhVXXFGd1rdvXwYNGhTX+qkQZID4AugaNt3FSwtXBsxU1d2quhpYjgsYZcACr3pqD/Ac8IMA82qMiWLdusTS90VZWRnvvfced911F7169WLu3LmUlJQwefJkbrzxRt91Pv30U+bMmcNHH33E73//e3bv3l1nmZKSEu6++25KS0tZtWoV7777Ljt27ODyyy/nxRdfZP78+WzatKne/JWXl/PBBx/Qp0+fWumbN2/ms88+4+STT/Zdb/r06YwaNYrhw4fzwgsv+OYx0uLFi+nfv3+9ywUpyAAxD+ghIt1FpAUwEpgZscxzuNIDItIRV7W0ylu3vYiEigWnAaUYY9IuWu1tELW6F1xwAbm5uQBs2bKFCy64gO9///tce+21LFmyxHedc845h5YtW9KxY0cOPvhgNm7cWGeZAQMG0KVLF3JycigoKGDNmjV8+umnHH744dXPBYwaNSpqvubOnUu/fv0444wzmDhxYnWAmDt3Ln379qVz586ceeaZHHrooXXW3bVrF7Nnz+YnP/kJ+++/PwMHDmTOnDlA9DuLGsodZoEFCO/K/2pgDrAUeEpVl4jIZBE5z1tsDlAuIqXAG8BvVbXca2u4DnhNRD4BBLg/qLwaY6KbMgVat66d1rq1S0+1Nm3aVH+++eabOfXUU1m8eDGzZs2Keg9/y5Ytqz/n5uayZ8+epJaJZdCgQZSUlDB//vxaVT6DBg1i4cKFLFmyhAcffLC6XSXcnDlzqKio4JhjjiE/P5933nmnupqpQ4cObN68udbyW7dupX379vTp04f58+cnlM9UC7QNQlVnq2pPVT1CVad4abeo6kzvs6rqf6tqb1U9RlVnhK37iqoe66WP9e6EMsak2ejRMG0adOsGIu592rTkG6jjtWXLFjp37gzAI488kvLtH3XUUaxatYo1a9YA8OSTTya9re7duzNx4kT++Mc/1pk3ffp0HnjgAdasWcOaNWtYvXo1r7zyCpWVlZx88snMnDmTrVu3AvDMM8/Qt29fcnNzOe2009i5cyfTpk2r3taiRYuYO3du0vlMVKYbqY0xjcDo0bBmDVRVufeggwPA9ddfzw033EC/fv0SvuKPx3777cc//vEPhgwZQv/+/WnXrh0HHHBA0tu74oorePvtt6sDDkBlZSUvvfQS55xzTnVamzZtOOmkk5g1axbHHnssV199NSeddBIFBQX885//5IEHHgBcNdOzzz7Lq6++yhFHHEGfPn244YYbfKuxgiKh1vzGrrCwUIuLizOdDWMahaVLl3L00UdnOhsZt23bNtq2bYuqctVVV9GjRw+uvfbaTGcrMH5/dxGZr6q+9w1bCcIYk7Xuv/9+CgoK6NOnD1u2bOHyyy/PdJYaFOvN1RiTta699tomXWLYV1aCMMYY48sChDHGGF8WIIwxxviyAGGMMcaXBQhjTNqdeuqp1d1NhNx9992MHz8+6jqDBw8mdCv72WefTUVFRZ1l/LrPjvTcc89RWlrTc88tt9zCq6++mkDu/TXFbsEtQBhj0m7UqFHVXViHzJgxI2Z/SOFmz55N+/btk9p3ZICYPHkyP/7xj5PaVqRBgwaxYMECiouLefzxx/n4449rpZeUlPD888/z7rvv+q4f2S14PELdgg8ePJiVK1cyf/58/ud//se3T6pE2W2uxmS5CRPApwuhfVJQAHffHX3+iBEjuOmmm9i1axctWrRgzZo1fPnllwwaNIjx48czb948tm/fzogRI6qvxMPl5+dTXFxMx44dmTJlCo8++igHH3wwXbt2re4B9f7772fatGns2rWLI488kscee4wFCxYwc+ZM3nrrLe644w7+/e9/c/vttzN06FBGjBjBa6+9xnXXXceePXs47rjjuO+++2jZsiX5+fmMGTOGWbNmsXv3bv7v//6PXr16RT2+8G7BDz64ZqDMeLsFv+iii5g+fXpc40ZE6xY8FawEYYxJu4MOOogBAwbw4osvAq70cOGFFyIiTJkyheLiYhYtWsRbb73FokWLom5n/vz5zJgxgwULFjB79mzmzZtXPe/8889n3rx5LFy4kKOPPpoHH3yQE044gfPOO48777yTBQsWcMQRR1Qvv2PHDsaOHcuTTz7JJ598wp49e7jvvvuq53fs2JGPP/6Y8ePH11uN1VS6BbcShDFZLtaVfpBC1UzDhg1jxowZPPjggwA89dRTTJs2jT179rB+/XpKS0s59thjfbcxd+5chg8fTmuvu9nzzjuvet7ixYu56aabqKioYNu2bZx55pkx87Ns2TK6d+9ePVrbmDFjmDp1KhMmTABcwAHo378/zzzzTNT89OvXj5ycnOpuwd98883qbsE/++wzJkyYELNb8Lvuuot27dpVdws+dOjQjHULnvUliFSPtWuMic+wYcN47bXX+Pjjj6msrKR///6sXr2aP//5z7z22mssWrSIc845J2o33/UZO3Ys9957L5988gm33npr0tsJCXUZHqu78KbWLXhWB4jQWLtr14JqzVi7FiSMCV7btm059dRTufTSS6sbp7/99lvatGnDAQccwMaNG6uroKI5+eSTee6559i+fTtbt25l1qxZ1fO2bt3KYYcdxu7duykK+6du165ddffa4Y466ijWrFlTPZzoY489ximnnJKKQ63W2LoFz+oAka6xdo0x/kaNGsXChQurA0Tfvn3p168fvXr14uKLL+bEE0+Muf4PfvADLrroIvr27ctZZ53FcccdVz3v9ttvZ+DAgZx44om1GpRHjhzJnXfeSb9+/Vi5cmV1eqtWrXj44Ye54IILOOaYY8jJyalVCkiVxtQteKDdfYvIEOBvQC7wgKr+P59lLgRuAxRYqKoXh83bHzfU6HOqenWsfSXT3XdOjis51M2T6/femKbKuvvOTol29x1YI7WI5AJTgdOBMmCeiMxU1dKwZXoANwAnqupmETk4YjO3A28Hlce8PFet5JdujDHZLsgqpgHAClVd5Q0XOgMYFrHMZcBUVd0MoKpfhWaISH/gEODloDKYzrF2jTGmsQkyQHQGPg+bLvPSwvUEeorIuyLygVclhYjkAH8BYj6TLiLjRKRYRIo3bdqUcAYzNdauMQ1BUxlN0sQnmb93pp+DaAb0AAYDXYC3ReQY4GfAbFUti3Wfr6pOA6aBa4NIJgOjR1tAMNmnVatWlJeX06FDh8DvpTeZp6qUl5fTqlWrhNYLMkB8AXQNm+7ipYUrAz5U1d3AahFZjgsYxwODRORKoC3QQkS2qerEAPNrTNbo0qULZWVlJFPyNo1Tq1at6NKlS0LrBBkg5gE9RKQ7LjCMBC6OWOY5YBTwsIh0xFU5rVLV6mt6ERkLFFpwMCZ1mjdvTvfu3TOdDdPABdYGoap7gKuBOcBS4ClVXSIik0Uk9Dz8HKBcREqBN4Dfqmp5UHkyxhgTv0Cfg0inZJ6DMMaYbBfrOYisfpLaGGNMdE2mBCEimwCfx95q6Qh8nYbsNETZeux23NnFjjtx3VS1k9+MJhMg4iEixdGKUk1dth67HXd2seNOLatiMsYY48sChDHGGF/ZFiCm1b9Ik5Wtx27HnV3suFMoq9ogjDHGxC/bShDGGGPiZAHCGGOMr6wJECIyRESWicgKEWmy/TqJyEMi8pWILA5LO0hEXhGRz7z3AzOZxyCISFcReUNESkVkiYj82ktv0scuIq1E5CMRWegd9++99O4i8qH3e39SRFpkOq9BEJFcESkRkee96Ww57jUi8omILBCRYi8t5b/1rAgQYaPbnQX0BkaJSO/M5iowjwBDItImAq+pag/gNW+6qdkD/EZVewM/BK7y/sZN/dh3Aqepal+gABgiIj8E/gj8VVWPBDYDv8hcFgP1a1xfbyHZctwAp6pqQdjzDyn/rWdFgCC+0e2aBFV9G/gmInkY8Kj3+VHgJ+nMUzqo6npV/dj7vBV30uhMEz92dbZ5k829lwKnAU976U3uuAFEpAtwDvCANy1kwXHHkPLferYEiHhGt2vKDlHV9d7nDbihXJssEckH+gEfkgXH7lWzLAC+Al4BVgIVXo/K0HR/73cD1wNV3nQHsuO4wV0EvCwi80VknJeW8t96pkeUM2mmqioiTfbeZhFpC/wbmKCq34aPltZUj11V9wIFItIeeBboldkcBU9EhgJfqep8ERmc4exkwkmq+oWIHAy8IiKfhs9M1W89W0oQ8Yxu15RtFJHDALz3rzKcn0CISHNccChS1We85Kw4dgBVrcCNq3I80F5EQheATfH3fiJwnoiswVUZnwb8jaZ/3ACo6hfe+1e4i4IBBPBbz5YAUT26nXdXw0hgZobzlE4zgTHe5zHAfzKYl0B49c8PAktV9a6wWU362EWkk1dyQET2A07Htb+8AYzwFmtyx62qN6hqF1XNx/0/v+6NRNmkjxtARNqISLvQZ+AMYDEB/Naz5klqETkbV2eZCzykqlMym6NgiMh0YDCu+9+NwK24oV2fAvJwXaJfqKqRDdmNmoicBMwFPqGmTvpGXDtEkz12ETkW1yCZi7vge0pVJ4vI4bgr64OAEuBnqrozczkNjlfFdJ2qDs2G4/aO8VlvshnwhKpOEZEOpPi3njUBwhhjTGKypYrJGGNMgixAGGOM8WUBwhhjjC8LEMYYY3xZgDDGGOPLAoQx9RCRvV6vmaFXyjr8E5H88J53jWlIrKsNY+q3XVULMp0JY9LNShDGJMnrk/9PXr/8H4nIkV56voi8LiKLROQ1Ecnz0g8RkWe9sRsWisgJ3qZyReR+bzyHl70nohGRa7zxLRaJyIwMHabJYhYgjKnffhFVTBeFzduiqscA9+Ke1Af4O/Coqh4LFAH3eOn3AG95Yzf8AFjipfcApqpqH6AC+KmXPhHo523nimAOzZjo7ElqY+ohIttUta1P+hrcYD2rvI4CN6hqBxH5GjhMVXd76etVtaOIbAK6hHf94HVN/oo3yAsi8juguareISIvAdtwXaU8FzbugzFpYSUIY/aNRvmciPC+gvZS0zZ4Dm4kxB8A88J6KTUmLSxAGLNvLgp7f9/7/B6uh1GA0bhOBMENAzkeqgf5OSDaRkUkB+iqqm8AvwMOAOqUYowJkl2RGFO//bwR20JeUtXQra4HisgiXClglJf2K+BhEfktsAn4uZf+a2CaiPwCV1IYD6zHXy7wuBdEBLjHG+/BmLSxNghjkuS1QRSq6teZzosxQbAqJmOMMb6sBGGMMcaXlSCMMcb4sgBhjDHGlwUIY4wxvixAGGOM8WUBwhhjjK//Dy1+MmXpZZLuAAAAAElFTkSuQmCC",
thecml's avatar
thecml committed
154
      "text/plain": [
155
       "<Figure size 432x288 with 1 Axes>"
156
      ]
157
     },
158
159
160
     "metadata": {
      "needs_background": "light"
     },
161
     "output_type": "display_data"
thecml's avatar
thecml committed
162
163
164
    },
    {
     "data": {
165
      "image/png": "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",
thecml's avatar
thecml committed
166
      "text/plain": [
167
       "<Figure size 432x288 with 1 Axes>"
168
      ]
169
     },
170
171
172
     "metadata": {
      "needs_background": "light"
     },
173
     "output_type": "display_data"
174
175
176
    },
    {
     "data": {
177
      "image/png": "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",
178
      "text/plain": [
179
       "<Figure size 432x288 with 1 Axes>"
180
      ]
181
     },
182
183
184
     "metadata": {
      "needs_background": "light"
     },
185
     "output_type": "display_data"
thecml's avatar
thecml committed
186
187
    }
   ],
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
   "source": [
    "# Plot the loss curves for training and validation.\n",
    "history_dict = history.history\n",
    "loss_values = history_dict['loss']\n",
    "val_loss_values = history_dict['val_loss']\n",
    "epochs = range(1, len(loss_values)+1)\n",
    "\n",
    "plt.figure(1)\n",
    "plt.plot(epochs, loss_values, 'bo', label='Training loss')\n",
    "plt.plot(epochs, val_loss_values, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "# Plot the accuracy curves for training and validation.\n",
    "acc_values = history_dict['accuracy']\n",
    "val_acc_values = history_dict['val_accuracy']\n",
    "epochs = range(1, len(acc_values)+1)\n",
    "\n",
    "plt.figure(1)\n",
    "plt.plot(epochs, acc_values, 'bo', label='Training accuracy')\n",
    "plt.plot(epochs, val_acc_values, 'b', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "# Plot the ROCAUC curves for training and validation.\n",
    "acc_values = history_dict['roc_auc']\n",
    "val_acc_values = history_dict['val_roc_auc']\n",
    "epochs = range(1, len(acc_values)+1)\n",
    "\n",
    "plt.figure(1)\n",
    "plt.plot(epochs, acc_values, 'bo', label='Training ROC AUC')\n",
    "plt.plot(epochs, val_acc_values, 'b', label='Validation ROC AUC')\n",
    "plt.title('Training and validation ROC AUC')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('ROC AUC')\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "# Plot the PRAUC curves for training and validation.\n",
    "acc_values = history_dict['pr_auc']\n",
    "val_acc_values = history_dict['val_pr_auc']\n",
    "epochs = range(1, len(acc_values)+1)\n",
    "\n",
    "plt.figure(1)\n",
    "plt.plot(epochs, acc_values, 'bo', label='Training PR AUC')\n",
    "plt.plot(epochs, val_acc_values, 'b', label='Validation PR AUC')\n",
    "plt.title('Training and validation PR AUC')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('PR AUC')\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "# Plot the precision curves for training and validation.\n",
    "acc_values = history_dict['precision']\n",
    "val_acc_values = history_dict['val_precision']\n",
    "epochs = range(1, len(acc_values)+1)\n",
    "\n",
    "plt.figure(1)\n",
    "plt.plot(epochs, acc_values, 'bo', label='Training precision')\n",
    "plt.plot(epochs, val_acc_values, 'b', label='Validation precision')\n",
    "plt.title('Training and validation precision')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Precision')\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "# Plot the recall curves for training and validation.\n",
    "acc_values = history_dict['recall']\n",
    "val_acc_values = history_dict['val_recall']\n",
    "epochs = range(1, len(acc_values)+1)\n",
    "\n",
    "plt.figure(1)\n",
    "plt.plot(epochs, acc_values, 'bo', label='Training recall')\n",
    "plt.plot(epochs, val_acc_values, 'b', label='Validation recall')\n",
    "plt.title('Training and validation recall')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Recall')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
thecml's avatar
thecml committed
274
275
276
  },
  {
   "cell_type": "code",
277
   "execution_count": 12,
278
   "metadata": {},
thecml's avatar
thecml committed
279
   "outputs": [
thecml's avatar
thecml committed
280
    {
281
     "data": {
282
      "image/png": "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",
283
      "text/plain": [
284
       "<Figure size 432x288 with 1 Axes>"
285
      ]
286
     },
287
288
289
     "metadata": {
      "needs_background": "light"
     },
290
     "output_type": "display_data"
thecml's avatar
thecml committed
291
    },
thecml's avatar
thecml committed
292
293
    {
     "data": {
294
      "image/png": "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",
thecml's avatar
thecml committed
295
      "text/plain": [
296
       "<Figure size 432x288 with 1 Axes>"
297
      ]
thecml's avatar
thecml committed
298
     },
299
300
301
     "metadata": {
      "needs_background": "light"
     },
302
     "output_type": "display_data"
thecml's avatar
thecml committed
303
304
    }
   ],
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
   "source": [
    "# Plot ROC and PR curves\n",
    "from sklearn.metrics import roc_curve, precision_recall_curve\n",
    "from sklearn.metrics import auc\n",
    "y_pred = model.predict(X_test).ravel()\n",
    "\n",
    "fpr, tpr, _ = roc_curve(y_test, y_pred)\n",
    "roc_auc = auc(fpr, tpr)\n",
    "plt.figure(1)\n",
    "plt.plot([0, 1], [0, 1], 'k--')\n",
    "plt.plot(fpr, tpr, label='MLP (area = {:.3f})'.format(roc_auc))\n",
    "plt.xlabel('False positive rate')\n",
    "plt.ylabel('True positive rate')\n",
    "plt.title('ROC curve')\n",
    "plt.legend(loc='best')\n",
    "plt.show()\n",
    "\n",
    "precision, recall, _ = precision_recall_curve(y_test, y_pred)\n",
    "pr_auc = auc(recall, precision)\n",
    "plt.figure(1)\n",
    "len_samples = len(y_test[y_test==1]) / len(y_test)\n",
    "plt.plot([0, 1], [len_samples, len_samples], 'k--')\n",
    "plt.plot(recall, precision, label='MLP (area = {:.3f})'.format(pr_auc))\n",
    "plt.xlabel('Recall')\n",
    "plt.ylabel('Precision')\n",
    "plt.title('PR curve')\n",
    "plt.legend(loc='best')\n",
    "plt.show()"
   ]
334
335
336
  },
  {
   "cell_type": "code",
337
   "execution_count": 13,
338
   "metadata": {},
339
   "outputs": [
thecml's avatar
thecml committed
340
341
    {
     "name": "stderr",
342
     "output_type": "stream",
thecml's avatar
thecml committed
343
     "text": [
344
345
      "keras is no longer supported, please use tf.keras instead.\n",
      "Your TensorFlow version is newer than 2.4.0 and so graph support has been removed in eager mode. See PR #1483 for discussion.\n"
thecml's avatar
thecml committed
346
347
348
349
     ]
    },
    {
     "name": "stdout",
350
     "output_type": "stream",
thecml's avatar
thecml committed
351
     "text": [
352
      "WARNING:tensorflow:From C:\\Users\\cml\\miniconda3\\envs\\py38-air\\lib\\site-packages\\shap\\explainers\\tf_utils.py:28: The name tf.keras.backend.get_session is deprecated. Please use tf.compat.v1.keras.backend.get_session instead.\n",
thecml's avatar
thecml committed
353
354
355
      "\n"
     ]
    },
356
    {
357
358
359
360
361
362
363
364
365
366
367
     "ename": "TypeError",
     "evalue": "unsupported operand type(s) for -: 'str' and 'str'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-13-bccb225d6bff>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[1;31m# Print shap features\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      2\u001b[0m \u001b[0mexplainer\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mshap\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mDeepExplainer\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mdata\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mX_train\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 3\u001b[1;33m \u001b[0mshap_values\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mexplainer\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshap_values\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX_test\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      4\u001b[0m \u001b[0mshap\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msummary_plot\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mshap_values\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mX_test\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mfeature_names\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mfeatures\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\miniconda3\\envs\\py38-air\\lib\\site-packages\\shap\\explainers\\_deep\\__init__.py\u001b[0m in \u001b[0;36mshap_values\u001b[1;34m(self, X, ranked_outputs, output_rank_order, check_additivity)\u001b[0m\n\u001b[0;32m    122\u001b[0m             \u001b[0mwere\u001b[0m \u001b[0mchosen\u001b[0m \u001b[1;32mas\u001b[0m \u001b[1;34m\"top\"\u001b[0m\u001b[1;33m.\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    123\u001b[0m         \"\"\"\n\u001b[1;32m--> 124\u001b[1;33m         \u001b[1;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mexplainer\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshap_values\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mranked_outputs\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0moutput_rank_order\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mcheck_additivity\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mcheck_additivity\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;32m~\\miniconda3\\envs\\py38-air\\lib\\site-packages\\shap\\explainers\\_deep\\deep_tf.py\u001b[0m in \u001b[0;36mshap_values\u001b[1;34m(self, X, ranked_outputs, output_rank_order, check_additivity)\u001b[0m\n\u001b[0;32m    310\u001b[0m                 \u001b[1;31m# assign the attributions to the right part of the output arrays\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    311\u001b[0m                 \u001b[1;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mlen\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mX\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 312\u001b[1;33m                     \u001b[0mphis\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0ml\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mj\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;33m(\u001b[0m\u001b[0msample_phis\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0ml\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mbg_data\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0ml\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m*\u001b[0m \u001b[1;33m(\u001b[0m\u001b[0mX\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0ml\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mj\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m-\u001b[0m \u001b[0mbg_data\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0ml\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmean\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    313\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    314\u001b[0m             \u001b[0moutput_phis\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mphis\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;32mif\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mmulti_input\u001b[0m \u001b[1;32melse\u001b[0m \u001b[0mphis\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mTypeError\u001b[0m: unsupported operand type(s) for -: 'str' and 'str'"
     ]
368
369
    }
   ],
370
371
372
373
374
375
   "source": [
    "# Print shap features\n",
    "explainer = shap.DeepExplainer(model, data=X_train)\n",
    "shap_values = explainer.shap_values(X_test)\n",
    "shap.summary_plot(shap_values[0], X_test, feature_names=features)"
   ]
thecml's avatar
thecml committed
376
377
378
  }
 ],
 "metadata": {
379
  "interpreter": {
380
   "hash": "1257d43d6e3967ffdae7723e8889b746915ea50e5b681a3d1d09455fe4a03787"
381
  },
thecml's avatar
thecml committed
382
  "kernelspec": {
383
384
   "display_name": "Python 3.8.11 64-bit ('py38-air': conda)",
   "name": "python3"
thecml's avatar
thecml committed
385
386
387
388
389
390
391
392
393
394
395
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
396
   "version": "3.8.8"
thecml's avatar
thecml committed
397
  }
thecml's avatar
thecml committed
398
399
400
 },
 "nbformat": 4,
 "nbformat_minor": 2
401
}