examples/od_if_kddcup.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Isolation Forest outlier detection on KDD Cup '99 dataset\n",
    "\n",
    "## Method\n",
    "\n",
    "[Isolation forests](https://cs.nju.edu.cn/zhouzh/zhouzh.files/publication/icdm08b.pdf) (IF) are tree based models specifically used for outlier detection. The IF isolates observations by randomly selecting a feature and then randomly selecting a split value between the maximum and minimum values of the selected feature. The number of splittings required to isolate a sample is equivalent to the path length from the root node to the terminating node. This path length, averaged over a forest of random trees, is a measure of normality and is used to define an anomaly score. Outliers can typically be isolated quicker, leading to shorter paths.\n",
    "\n",
    "## Dataset\n",
    "\n",
    "The outlier detector needs to detect computer network intrusions using TCP dump data for a local-area network (LAN) simulating a typical U.S. Air Force LAN. A connection is a sequence of TCP packets starting and ending at some well defined times, between which data flows to and from a source IP address to a target IP address under some well defined protocol. Each connection is labeled as either normal, or as an attack.\n",
    "\n",
    "There are 4 types of attacks in the dataset:\n",
    "\n",
    "- DOS: denial-of-service, e.g. syn flood;\n",
    "- R2L: unauthorized access from a remote machine, e.g. guessing password;\n",
    "- U2R: unauthorized access to local superuser (root) privileges;\n",
    "- probing: surveillance and other probing, e.g., port scanning.\n",
    "\n",
    "The dataset contains about 5 million connection records.\n",
    "\n",
    "There are 3 types of features:\n",
    "\n",
    "- basic features of individual connections, e.g. duration of connection\n",
    "- content features within a connection, e.g. number of failed log in attempts\n",
    "- traffic features within a 2 second window, e.g. number of connections to the same host as the current connection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import os\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from sklearn.metrics import confusion_matrix, f1_score\n",
    "\n",
    "from alibi_detect.od import IForest\n",
    "from alibi_detect.datasets import fetch_kdd\n",
    "from alibi_detect.utils.data import create_outlier_batch\n",
    "from alibi_detect.utils.fetching import fetch_detector\n",
    "from alibi_detect.utils.saving import save_detector, load_detector\n",
    "from alibi_detect.utils.visualize import plot_instance_score, plot_roc"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load dataset\n",
    "\n",
    "We only keep a number of continuous (18 out of 41) features."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(494021, 18) (494021,)\n"
     ]
    }
   ],
   "source": [
    "kddcup = fetch_kdd(percent10=True)  # only load 10% of the dataset\n",
    "print(kddcup.data.shape, kddcup.target.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Assume that a model is trained on *normal* instances of the dataset (not outliers) and standardization is applied:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(400000, 18) (400000,)\n",
      "0.0% outliers\n"
     ]
    }
   ],
   "source": [
    "np.random.seed(0)\n",
    "normal_batch = create_outlier_batch(kddcup.data, kddcup.target, n_samples=400000, perc_outlier=0)\n",
    "X_train, y_train = normal_batch.data.astype('float'), normal_batch.target\n",
    "print(X_train.shape, y_train.shape)\n",
    "print('{}% outliers'.format(100 * y_train.mean()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "mean, stdev = X_train.mean(axis=0), X_train.std(axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Apply standardization:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = (X_train - mean) / stdev"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load or define outlier detector\n",
    "\n",
    "The pretrained outlier and adversarial detectors used in the example notebooks can be found [here](https://console.cloud.google.com/storage/browser/seldon-models/alibi-detect). You can use the built-in ```fetch_detector``` function which saves the pre-trained models in a local directory ```filepath``` and loads the detector. Alternatively, you can train a detector from scratch:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "load_outlier_detector = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "filepath = 'my_path'  # change to directory where model is downloaded\n",
    "if load_outlier_detector:  # load pretrained outlier detector\n",
    "    detector_type = 'outlier'\n",
    "    dataset = 'kddcup'\n",
    "    detector_name = 'IForest'\n",
    "    od = fetch_detector(filepath, detector_type, dataset, detector_name)\n",
    "    filepath = os.path.join(filepath, detector_name)\n",
    "else:  # define model, initialize, train and save outlier detector\n",
    "    \n",
    "    # initialize outlier detector\n",
    "    od = IForest(threshold=None,  # threshold for outlier score\n",
    "                 n_estimators=100)\n",
    "    \n",
    "    # train\n",
    "    od.fit(X_train)\n",
    "    \n",
    "    # save the trained outlier detector\n",
    "    save_detector(od, filepath)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If `load_outlier_detector` equals False, the warning tells us we still need to set the outlier threshold. This can be done with the `infer_threshold` method. We need to pass a batch of instances and specify what percentage of those we consider to be normal via `threshold_perc`. Let's assume we have some data which we know contains around 5% outliers. The percentage of outliers can be set with `perc_outlier` in the `create_outlier_batch` function."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5.0% outliers\n"
     ]
    }
   ],
   "source": [
    "np.random.seed(0)\n",
    "perc_outlier = 5\n",
    "threshold_batch = create_outlier_batch(kddcup.data, kddcup.target, n_samples=1000, perc_outlier=perc_outlier)\n",
    "X_threshold, y_threshold = threshold_batch.data.astype('float'), threshold_batch.target\n",
    "X_threshold = (X_threshold - mean) / stdev\n",
    "print('{}% outliers'.format(100 * y_threshold.mean()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "New threshold: 0.08008174509752322\n"
     ]
    }
   ],
   "source": [
    "od.infer_threshold(X_threshold, threshold_perc=100-perc_outlier)\n",
    "print('New threshold: {}'.format(od.threshold))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's save the outlier detector with updated threshold:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "save_detector(od, filepath)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Detect outliers\n",
    "\n",
    "We now generate a batch of data with 10% outliers and detect the outliers in the batch. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1000, 18) (1000,)\n",
      "10.0% outliers\n"
     ]
    }
   ],
   "source": [
    "np.random.seed(1)\n",
    "outlier_batch = create_outlier_batch(kddcup.data, kddcup.target, n_samples=1000, perc_outlier=10)\n",
    "X_outlier, y_outlier = outlier_batch.data.astype('float'), outlier_batch.target\n",
    "X_outlier = (X_outlier - mean) / stdev\n",
    "print(X_outlier.shape, y_outlier.shape)\n",
    "print('{}% outliers'.format(100 * y_outlier.mean()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Predict outliers:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "od_preds = od.predict(X_outlier, return_instance_score=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Display results\n",
    "\n",
    "F1 score and confusion matrix:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "F1 score: 0.3315\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "labels = outlier_batch.target_names\n",
    "y_pred = od_preds['data']['is_outlier']\n",
    "f1 = f1_score(y_outlier, y_pred)\n",
    "print('F1 score: {:.4f}'.format(f1))\n",
    "cm = confusion_matrix(y_outlier, y_pred)\n",
    "df_cm = pd.DataFrame(cm, index=labels, columns=labels)\n",
    "sns.heatmap(df_cm, annot=True, cbar=True, linewidths=.5)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plot instance level outlier scores vs. the outlier threshold:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_instance_score(od_preds, y_outlier, labels, od.threshold)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see that the isolation forest does not do a good job at detecting 1 type of outliers with an outlier score around 0. This makes inferring a good threshold without explicit knowledge about the outliers hard. Setting the threshold just below 0 would lead to significantly better detector performance for the outliers in the dataset. This is also reflected by the ROC curve:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmkAAAFNCAYAAABbpPhvAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOzdeXhU5fnG8e/DEpBFUMSl4pIQiCwKioIsyiYURGixlkUFsRRFRbG4gNa21mo1AaXgiqBSlEVkF1EUbBQp/FgNhD1sgqAgyBaWhPD8/shII7IMIZOZSe7Pdc3lnGXOuSfvlfjwvuec19wdEREREYksRcIdQERERER+SUWaiIiISARSkSYiIiISgVSkiYiIiEQgFWkiIiIiEUhFmoiIiEgEUpEmIiIiEoFUpIlIxDCzDWZ2wMz2mdl3ZjbczMocs08DM/vczPaa2W4z+9DMqh+zz9lm9i8z+yZwrLWB5fPy9xuJiOSeijQRiTRt3b0MUBu4Gnjipw1mVh/4FJgM/AqIBVKA2WYWF9gnBpgJ1ABaAWcD9YEdQN1QhTazYqE6togUTirSRCQiuft3wHSyi7WfJAEj3H2Qu+91953u/hQwF3g6sE9X4FKgvbsvd/cj7r7N3f/h7tOOdy4zq2Fmn5nZTjP73syeDKwfbmbP5tiviZltzrG8wcz6mtkSID3wftwxxx5kZoMD78uZ2VtmttXMvjWzZ82s6Bn+qESkgFKRJiIRycwqAa2BtMByKaAB8MFxdh8LtAi8vwn4xN33BXmessAM4BOye+fiye6JC1ZnoA1QHhgD3Bw4JoECrAMwKrDvcOBw4BxXAy2BP57GuUSkEFGRJiKRZpKZ7QU2AduAvwXWn0v236ytx/nMVuCn680qnGCfE7kF+M7dX3T3g4Eeuv87jc8PdvdN7n7A3TcCi4D2gW3NgP3uPtfMLgBuBh5293R33wYMBDqdxrlEpBBRkSYikea37l4WaAJcwf+Krx+BI8BFx/nMRcAPgfc7TrDPiVwCrM1V0mybjlkeRXbvGsDt/K8X7TKgOLDVzHaZ2S5gCHD+GZxbRAowFWkiEpHc/QuyhwcHBJbTgTnA74+zewf+N0Q5A/i1mZUO8lSbgLgTbEsHSuVYvvB4UY9Z/gBoEhiubc//irRNwCHgPHcvH3id7e41gswpIoWMijQRiWT/AlqYWa3Acj/gLjN7yMzKmtk5gQv76wN/D+zzLtkF0Xgzu8LMiphZBTN70sxuPs45pgIXmdnDZlYicNx6gW1fk32N2blmdiHw8KkCu/t2IBl4B1jv7isC67eSfWfqi4FHhBQxs8pm1jgXPxcRKQRUpIlIxAoUPCOAvwaWvwJ+DdxK9nVnG8m+AL+Ru68J7HOI7JsHVgKfAXuAeWQPm/7iWjN330v2TQdtge+ANUDTwOZ3yX7ExwayC6z3g4w+KpBh1DHruwIxwHKyh2/HcXpDsyJSiJj7sT31IiIiIhJu6kkTERERiUAq0kREREQikIo0ERERkQikIk1EREQkAqlIExEREYlAxcId4HSVL1/e4+Pjwx1Dcik9PZ3SpYN9xqhEErVddFP7RS+1XXRbuHDhD+5eMTefjboi7YILLmDBggXhjiG5lJycTJMmTcIdQ3JBbRfd1H7RS20X3cxsY24/q+FOERERkQikIk1EREQkAqlIExEREYlAKtJEREREIpCKNBEREZEIpCJNREREJAKpSBMRERGJQCEr0szsbTPbZmapJ9huZjbYzNLMbImZXROqLCIiIiLRJpQ9acOBVifZ3hqoEnjdA7wewiwiIiIiUSVkRZq7fwnsPMkuvwFGeLa5QHkzuyhUeURERESiSTivSbsY2JRjeXNgnYiIiEjUW7169Rl9Pirm7jSze8geEqVixYokJyeHN1CYPDAznfTMcKfIA598FO4Ekltqu+im9otearuocSTzIOlLZxDzqysoVu6CMzpWOIu0b4FLcixXCqz7BXd/E3gTICEhwQvrRLPpn3zEhhfahDvGGdFEwdFLbRfd1H7RS20XHdLT03nppZd45ZVXuKF+fZ599AFq1qyJDe6c62OGs0ibAvQyszFAPWC3u28NYx4RERGR0/LNN9+wa9cuqlatyvbt20lOTqZatWp5cuxQPoJjNDAHSDCzzWbW3cx6mlnPwC7TgHVAGjAUuD9UWURERETyUmpqKl27duXqq69mzpw5lCxZksGDB+dZgQYh7Elz95P277m7Aw+E6vwiIiIiecnd2bFjBxUqVOChhx6iRYsWDB48mPLly4fkfFFx44CIiIhIuBw5coQpU6aQmJjIueeey0cffcTnn38e8vOqSBMRERE5DnfHzOjWrRsrVqygb9++tG/fPt/OryJNREREJIfdu3czZMgQ3n//febOncugQYMoX748ZpavOTTBuoiIiEjAyJEjiYuLIyUlhbfffpvixYtzzjnn5HuBBirSREREpJBbvXo1999/Pzt27KBu3bosWLCAkSNHUqtWrbDmUpEmIiIihdLq1av53e9+R6NGjTj//PMpVqwYVapUITY2NtzRAF2TJiIiIoWIuzN9+nSqV69OkSJFaNy4MSNGjKB06dLhjvYL6kkTERGRAi8zM5ORI0dSu3ZtHn/8cb777jvi4+N56KGHIrJAA/WkiYiISAGWnp7Onj17KFasGCNGjOCFF16gVatWYbkR4HSpJ01EREQKnB9++IGnn36a2NhYRo8eTcWKFZk+fTqtW7eOigIN1JMmIiIiBUhWVhYADRo0oHHjxsyaNYuEhIQwp8odFWkiIiIS9VJSUkhKSmL//v1MnDiRpUuXUqJEiXDHOiMa7hQREZGo1rNnT1q3bk2tWrUYPnw4QNQXaKAiTURERKJMVlYW48ePp0uXLrg7vXv3Zv369Tz++OOUK1cu3PHyjIo0ERERiRpTpkyhWrVq9O/fn1tvvRV3p1q1agWi5+xYKtJEREQkou3atYuBAweSkZHBueeey9ChQ5kzZw7t27enSJGCW8oU3G8mIiIiUW3r1q089thjVK5cmUWLFrFnzx4aNWpE48aNo+YxGmdCd3eKiIhIRFm5ciUXXXQR69atIyMjg0WLFnHZZZeFO1a+U5EmIiIiEWHOnDkkJiYyZ84cJk6cSMOGDWnYsGG4Y4WNhjtFREQkbNydAwcOsH79eu68805atGjB+vXradCgQbijhZ2KNBEREcl3mZmZjBgxgquuuoqXX36Z2NhY1qxZwwMPPECpUqXCHS8iaLhTRERE8lVmZiY1atTgkksu4cUXX6RFixYABfpOzdxQkSYiIiIht23bNl5++WX27NnDoEGDmDFjBpdeemm4Y0U0lawiIiISUk8++SRXXHEF27dv58EHHwRQgRYEFWkiIiKS5xYvXswLL7wAQNOmTVm+fDlvvPEG8fHxYU4WPVSkiYiISJ756quvaNmyJW3btqV48eK4Oy1atODCCy8Md7SooyJNREREzkhWVhbTpk3D3dm0aRO3334769at45FHHikUMwOEim4cEBERkVw5cOAAw4cPZ8CAAVx44YVcf/31dO7cOdyxCgwVaSIiInJafvzxR8qUKcOHH37Ixx9/zIgRIwr1zAChouFOERERCcqmTZt45JFHqFy5MgsWLKBDhw5MmTJFBVqIqEgTERGRk3J3li9fTq1atQBISUmhfv36YU5V8Gm4U0RERI7rq6++IikpiV//+tfcf//9rFu3jvLly4c7VqGhIk1ERER+5sCBA7Ro0YLvvvuORx99lLvuugszU4GWz1SkiYiICBkZGYwaNYr09HQeeOABnn32WW644QaKFi0a7miFlq5JExERKeQGDx5M5cqVGTVqFDVq1ACgSZMmKtDCTD1pIiIihdD333/P7NmzufXWW4mJiWHy5Mlcc8014Y4lOagnTUREpBBJS0ujZ8+eVKtWjVmzZuHu9OzZUwVaBFJPmoiISCGQlpZGfHw8Y8aMoWLFiqxcuZLzzz8/3LHkJFSkiYiIFFDuzqeffkpiYiJr164lNTWVp556KtyxJEga7hQRESlgDh8+jLvz1ltv8cgjj9CtWzfS0tIoW7ZsuKPJaVCRJiIiUkDs37+fV155hSpVqrBo0SK6du3K0qVL6dq1K8WLFw93PDlNKtJEREQKgCVLlhAbG8vMmTMZNWoUderUISYmBjMLdzTJJV2TJiIiEqU2btzIwIEDadasGa1ateKLL77giiuuCHcsySMh7Ukzs1ZmtsrM0sys33G2X2pm/zGzxWa2xMxuDmUeERGRgmDv3r106dKFa665huLFi3PttdcSExOjAq2ACVlPmpkVBV4FWgCbgflmNsXdl+fY7SlgrLu/bmbVgWnA5aHKJCIiEq3cnS+//JL09HRat25Nw4YNefnllzWfZgEWyp60ukCau69z9wxgDPCbY/Zx4OzA+3LAlhDmERERiTruzsSJE6lfvz5//OMfSU9Px8zo2bOnCrQCLpTXpF0MbMqxvBmod8w+TwOfmtmDQGngphDmERERiRqHDh1i+fLlmBnJyck89thj/Pa3v9V8moVIuG8c6AwMd/cXzaw+8K6Z1XT3Izl3MrN7gHsAKlasSHJycv4njRDR/t337dsX9d+hsFLbRTe1X/TYt28fH374IRMmTKBWrVo8/PDDtG/fHoBZs2aFOZ3kp1AWad8Cl+RYrhRYl1N3oBWAu88xs5LAecC2nDu5+5vAmwAJCQnepEmTEEWOcJ98RLR/9+Tk5Kj/DoWV2i66qf0iX3p6OqVLl6ZHjx7s37+fGTNmUKtWLbVdIRbKIm0+UMXMYskuzjoBtx+zzzdAc2C4mVUDSgLbQ5hJREQkoqxatYoBAwYwbdo00tLSGDJkCEWK6DGmEsIbB9z9MNALmA6sIPsuzmVm9oyZtQvs9gjQw8xSgNFAN3f3UGUSERGJJG+99RY33HADF198MSkpKZx11lkq0OSokF6T5u7TyH6sRs51f83xfjnQMJQZREREIoW788knn9C/f3+GDBlC+/bt6dSpE6VLlw53NIlA4b5xQEREpFBYunQpd9xxBwB9+/bl8ssv13yaclIq0kREREIkPT2dt956iwYNGhAXF0diYiKtWrXSfJoSFA18i4iI5LF9+/bxt7/9jdjYWL744gtKlSrFueeeS+vWrVWgSdDUkyYiIpJHNmzYwO7du0lISGD37t3MmjWLhISEcMeSKKWeNBERkTOUkpLCHXfcQZ06dZg3bx4lS5bkX//6lwo0OSMq0kRERHLB3dm2bRvuziOPPELt2rVZt24dPXr0CHc0KSA03CkiInIasrKymDhxIklJSVxwwQV8+OGHzJgxI9yxpABSkSYiIhIEd8fM6Nq1K+vWrePJJ5+kXbt2p/6gSC6pSBMRETmJXbt28frrrzN27FjmzZvHa6+9xtlnn627NCXkdE2aiIjICbz33ntUrlyZFStWMGLECIoXL065cuVUoEm+UJEmIiKSw4oVK7j33nvZuXMn119/PYsXL2bEiBFceeWV4Y4mhYyKNBEREWDVqlX89re/pUmTJlSqVIlixYoRHx/PpZdeGu5oUkjpmjQRESm0jhw5wrRp07jyyispWrQoLVq0YNSoUZQqVSrc0URUpImISOGTmZnJ6NGjSUpKonjx4gwdOpRrr72W+Pj4cEcTOUpFmoiIFBr79u1jz549FC9enDFjxvDSSy/RokUL3QggEUnXpImISIG3bds2/vKXvxAbG8sHH3xAxYoVmTZtGi1btlSBJhFLPWkiIlJgHT58GDPjhhtuoGnTpsyZM0dDmhI1VKSJiEiBs2jRIpKSkjh06BATJ05k6dKlxMTEhDuWyGnRcKeIiBQoPXr0oF27dlx33XWMGDECQAWaRCUVaSIiEtUOHz7M+++/z+2334678+ijj7Ju3ToeeeQRypYtG+54IrmmIk1ERKLWpEmTSEhIYPDgwXTq1AmAhIQE9ZxJgRBUkWZmMWamKy1FRCTsdu7cyYABA8jIyKBixYqMGDGC2bNn065dO92pKQXKKYs0M2sDLAU+CyzXNrOJoQ4mIiKS05YtW+jTpw/x8fEsW7aMvXv30rBhQxo2bBjuaCIhEczdnc8A9YD/ALj71+pVExGR/LJs2TIqVarE+vXrMTOWLFlCpUqVwh1LJOSCKdIy3X3XMV3IHqI8IiIiAHz11VckJiYyf/58Jk6cqF4zKXSCuSZthZl1AIqYWayZDQTmhjiXiIgUQkeOHGH//v2sW7eOu+++mzZt2rB+/Xrq168f7mgi+S6YIq0XUAc4AkwADgG9QxlKREQKl4yMDN555x1q1qzJa6+9RlxcHKtWraJnz56cddZZ4Y4nEhbBDHf+2t37An1/WmFmt5JdsImIiJyRzMxMqlevTlxcHC+//DLNmjUDoEgRPSVKCrdgirSn+GVB9ufjrBMREQnK999/z6BBg9i3bx+DBw8mOTlZNwOIHOOE/0wxs18Hrj+72MxeyvEaRvbQp4iIyGnr168f1apVY/fu3Tz88MMAKtBEjuNkPWnbgFTgILAsx/q9QL9QhhIRkYJlwYIFfPrppzz55JPcdNNN9OnTh/PPPz/csUQi2gmLNHdfDCw2s5HufjAfM4mISAHx5Zdf8ve//501a9bwyCOP4O7cdNNN4Y4lEhWCuSrzYjMbY2ZLzGz1T6+QJxMRkah0+PBhPvzwQ9ydLVu2cNddd7F27Vp69+6taZtETkMwNw4MB54FBgCtgbvRw2xFROQY+/fv5+233+bFF1/kkksuoWHDhkcnPReR0xdMT1opd58O4O5r3f0psos1ERERduzYQWZmJlOnTmXGjBmMGjWKL7/8knPPPTfc0USiWjBF2iEzKwKsNbOeZtYWKBviXCIiEuE2btxI7969qVKlCgsXLqRDhw5MmjRJswOI5JFgirQ/AaWBh4CGQA/gD6EMJSIikcvdWbZsGddccw0lSpQgNTWV66+/PtyxRAqcU16T5u7/F3i7F+gCYGYXhzKUiIhEFnfnyy+/JDExkTZt2nD//fezbt06ypUrF+5oIgXWSYs0M7sOuBj4yt1/MLMaZE8P1QzQkwdFRAqB/fv307x5c3bs2MFjjz1Gly5dMDMVaCIhdsIizcyeB34HpABPmdlU4H4gEeiZP/FERCQcDh06xLvvvsvBgwfp1asXiYmJNGzYkKJFi4Y7mkihcbJr0n4D1HL33wMtgceA6939RXffny/pREQk3w0cOJDY2FjGjx9PrVq1ALjxxhtVoInks5MNdx509wMA7r7TzFa7+7p8yiUiIvlo69atzJ49m9tuu41SpUrx8ccfHy3QRCQ8TtaTFmdmEwKviUBsjuUJ+RVQRERCZ9WqVfTo0YMaNWowd+5cAO69914VaCIR4GQ9ab87ZvmV0z24mbUCBgFFgWHu/sJx9ukAPE32LAYp7n776Z5HREROz+rVq6latSrjxo3j4osvZvXq1Zx33nnhjiUiOZxsgvWZZ3JgMysKvAq0ADYD881sirsvz7FPFeAJoKG7/2hm55/JOUVE5MTcnY8//pikpCQ2btzIkiVL+POf/xzuWCJyAsE8zDa36gJp7r7O3TOAMWTfjJBTD+BVd/8RwN23hTCPiEihdPjwYdydYcOG0a9fP3r06MHq1aspW1aTx4hEsmAmWM+ti4FNOZY3A/WO2acqgJnNJntI9Gl3/+TYA5nZPcA9ABUrViQ5OTkUeaNCtH/3ffv2Rf13KKzUdtHnwIEDfPTRR4wbN44nnniCGjVqMGjQIMyM2bNnhzueBEm/e4VX0EWamZVw90MhOH8VoAnZD8f90syudPddOXdy9zeBNwESEhK8SZMmeRwjSnzyEdH+3ZOTk6P+OxRWarvokpKSwu9//3tuvPFGpkyZwv79+9V+UUq/e4XXKYc7zayumS0F1gSWa5nZy0Ec+1vgkhzLlQLrctoMTHH3THdfD6wmu2gTEZHTtH79eh588EEmT55MtWrVmD17NuPHj6du3brhjiYiuRDMNWmDgVuAHQDungI0DeJz84EqZhZrZjFAJ2DKMftMIrsXDTM7j+zhTz2LTUTkNOzZs4fbb7+da6+9ltKlS1O3bl1iYmKoWrVquKOJyBkIZriziLtvNLOc67JO9SF3P2xmvYDpZF9v9ra7LzOzZ4AF7j4lsK2lmS0PHPMxd99x2t9CRKSQcXeSk5PZv38/N998M02aNOH111/XfJoiBUgwRdomM6sLeOCxGg+SPSx5Su4+DZh2zLq/5njvQJ/AS0RETsHdGT9+PImJiezdu5fnnnsOM+Oee+4JdzQRyWPBFGn3kT3keSnwPTAjsE5ERPLJwYMHWbZsGXXq1GH27Nn8+c9/pl27dhQpEsonKYlIOAVTpB12904hTyIiIr+wa9cuXn/9dQYPHkyLFi0YMWIEAwcODHcsEckHwRRp881sFfA+MMHd94Y4k4hIobdv3z7KlCnDo48+SkZGBp9++ilXXnlluGOJSD46ZZHm7pXNrAHZd2f+3cy+Bsa4+5iQpxMRKWRWrFhB//79mT59OmvXrmXo0KEcc+OWiBQSQV3M4O7/dfeHgGuAPcDIkKYSESmEhg4dSpMmTYiNjWXJkiWULFlSBZpIIRbMw2zLmNkdZvYhMA/YDjQIeTIRkQLuyJEjTJ06lcaNG7NmzRpuu+021q9fz1/+8hcqVKgQ7ngiEmbBXJOWCnwIJLn7rBDnEREpFFJSUrjjjjsoXrw4ffv2JTY2lmLFQjmdsohEm2D+IsS5+5GQJxERKeD27dvH0KFDueGGG4iLi+Oll16iRYsWGtIUkeM64XCnmb0YeDvezCYc+8qnfCIiUW/v3r089dRTxMbGMnfuXMqUKcO5555Ly5YtVaCJyAmdrCft/cB/X8mPICIiBc26devYs2cPV1xxBQcPHmTOnDnEx8eHO5aIRIkT9qS5+7zA22ruPjPnC6iWP/FERKLPokWL6NSpE3Xr1mXhwoWULFmSAQMGqEATkdMSzCM4/nCcdd3zOoiISDRzd77//nvcnX79+lG3bl3Wr19P9+76cykiuXPC4U4z60j2A2xjj7kGrSywK9TBRESiweHDhxk/fjxJSUlUqlSJyZMn8+mnn4Y7logUACe7Jm0esAOoBLyaY/1eYHEoQ4mIRDp3x8zo0qULmzZt4umnn6ZNmzbhjiUiBcgJizR3Xw+sB2bkXxwRkci2c+dOXnvtNcaNG8f8+fMZMmQIZ599drhjiUgBdLJHcHwR+O+PZrYzx+tHM9uZfxFFRCLDiBEjiI+PZ+3atYwaNYrixYurQBORkDnZcGfTwH/Py48gIiKRaNmyZQwcOJCkpCQaNmzIkiVLqFSpUrhjiUghcLJHcPw0y8AlQFF3zwLqA/cCpfMhm4hI2KxcuZK2bdvSvHlz4uLiKFasGJUrV1aBJiL5JphpoSYB15lZZeAdYCowCrgllMFERPLbTxOe16pVi+LFi3PLLbcwduxYzjrrrHBHE5FCKJgi7Yi7Z5rZrcDL7j7YzHR3p4gUGBkZGYwcOZL+/ftTqlQphgwZQp06dahcuXK4o4lIIRZMkXbYzH4PdAF+G1hXPHSRRETyx549e9i7dy8xMTFMmDCBl19+mWbNmmk+TRGJCMHOONAUSHL3dWYWC4wObSwRkdD57rvvePLJJ4mLi2PChAlUrFiRDz/8kObNm6tAE5GIccqeNHdPNbOHgHgzuwJIc/fnQh9NRCRvZWZmUqRIERo3bkyLFi2YP38+sbGx4Y4lInJcpyzSzOwG4F3gW8CAC82si7vPDnU4EZG8MH/+fBITEzly5AgTJkwgNTWV4sV11YaIRLZghjsHAje7e0N3bwC0AQaFNpaISN7o3r07t912GzfccAMjRowAUIEmIlEhmCItxt2X/7Tg7iuAmNBFEhHJvcOHDzNq1Cg6deqEu9OvXz/S0tLo3bs3ZcqUCXc8EZGgBVOkLTKzN8ysUeD1OppgXUQi0IQJE6hSpQpDhgyhS5cuAFSpUkU9ZyISlYIp0noC64DHA691ZM86ICISdjt27CAxMZGMjAwuvPBCRo8ezRdffEGbNm10p6aIRLWT3jhgZlcClYGJ7p6UP5FERE5t8+bN9O/fn3fffZdbb72Vffv20aBBg3DHEhHJMycs0szsSaA7sIjsaaGecfe38y2ZiMhxLFmyhMsuu4xvvvmGEiVKkJqayq9+9atwxxIRyXMn60m7A7jK3dPNrCIwDVCRJiL5zt358ssvSUxMZPHixUycOJEGDRqo50xECrSTXZN2yN3TAdx9+yn2FRHJc1lZWaSnp7Nu3Truvfde2rdvz/r167n++uvDHU1EJORO1pMWZ2YTAu8NqJxjGXe/NaTJRKTQOnToECNGjGDAgAHce++99OnThxUrVuhGABEpVE5WpP3umOVXQhlERAQgIyODatWqkZCQwJAhQ2jcuDGACjQRKXROWKS5+8z8DCIihdeWLVsYNGgQ6enpvPLKK3z11Ve6GUBECj1dZyYiYfXYY49Rs2ZNDhw4wKOPPgqgAk1EBBVpIhIGc+fO5R//+AcArVu3ZvXq1QwePJjLL788vMFERCJI0EWamZUIZRARKfiSk5Np3LgxnTp1onz58rg7zZo147zzzgt3NBGRiHPKIs3M6prZUmBNYLmWmb0c8mQiUiBkZmYyadIk3J3vv/+ee++9lzVr1vDggw/qZgARkZM46bRQAYOBW4BJAO6eYmZNQ5pKRKJeeno6w4YN46WXXiIuLo7GjRvTsWPHcMcSEYkawRRpRdx94zH/4s0KUR4RiXLbt2/nnHPO4aOPPmLWrFl88MEH1K1bN9yxRESiTjDXpG0ys7qAm1lRM3sYWB3iXCISZdavX0+vXr2oWrUqCxcupEOHDowbN04FmohILgVTpN0H9AEuBb4Hrg+sOyUza2Vmq8wszcz6nWS/35mZm9m1wRxXRCKHu5Oamsp1111H2bJlWb58OfXq1Qt3LBGRqHfK4U533wZ0Ot0Dm1lR4FWgBbAZmG9mU9x9+TH7lQV6A/93uucQkfBwd/7zn/+QmJjIb37zG+677z7Wr19P2bJlwx1NRKTAOGWRZmZDAT92vbvfc4qP1gXS3H1d4DhjgN8Ay4/Z7x9AIvBYMIFFJLz2799PkyZN2Lt3L4899hh33EXZ4Q4AAB6HSURBVHEHZqYCTUQkjwVz48CMHO9LAu2BTUF87uJj9tsM/GwMxMyuAS5x94/MTEWaSIQ6ePAgU6ZMYcmSJTz00EMMHDiQ+vXrU6SInoctIhIqwQx3vp9z2czeBb460xObWRHgJaBbEPveA9wDULFiRZKTk8/09FEr2r/7vn37ov47FDbvv/8+Y8eOJS4ujri4uKPt9+WXX4Y3mJwW/e5FL7Vd4RVMT9qxYoELgtjvW+CSHMuVAut+UhaoCSQHHu9xITDFzNq5+4KcB3L3N4E3ARISErxJkya5iF0AfPIR0f7dk5OTo/47FAbffvsts2fPpkOHDqxZs4YHHniAH374QW0XxfS7F73UdoVXMDMO/GhmOwOvXcBnwBNBHHs+UMXMYs0shuybD6b8tNHdd7v7ee5+ubtfDswFflGgiUj+WbFiBX/4wx+48sorWbRoEQA9evSgZs2aYU4mIlL4nLQnzbK7uGrxvx6wI+7+i5sIjsfdD5tZL2A6UBR4292XmdkzwAJ3n3LyI4hIflmxYgXVqlVj0qRJxMbGkpaWxrnnnhvuWCIihdpJizR3dzOb5u65+me0u08Dph2z7q8n2LdJbs4hIrlz5MgRPvroIxITE/nuu+/4+uuveeKJYDrJRUQkPwRza9bXZnZ1yJOISL7IzMzE3Rk2bBh/+9vfePDBB1m5ciVlypQJdzQREcnhhEWamf3Uy3Y12Q+iXWVmi8xssZktyp94IpJX9u7de3Sy85SUFO6++24WLlxIx44dKVYsN/cQiYhIKJ3sL/M84BqgXT5lEZEQWbx4MS1btqRZs2ZMmjSJ2rVrhzuSiIicwsmKNANw97X5lEVE8tDatWt58cUXadmyJTfffDNz5swhPj4+3LFERCRIJyvSKppZnxNtdPeXQpBHRM7Qnj176NGjBzNnzqRnz540aNCAmJgYFWgiIlHmZEVaUaAMgR41EYlc7s6MGTM4ePAgt9xyCy1btmTYsGGaT1NEJIqdrEjb6u7P5FsSETlt7s7YsWNJSkri4MGDPPfcc5gZ3bt3D3c0ERE5Q6e8Jk1EIs+BAwdITU3luuuuY968eTz99NO0adNGE56LiBQgJ/uL3jzfUohIUHbu3Mmzzz5LbGwsr7/+OgAvvvgibdu2VYEmIlLAnLAnzd135mcQETmxvXv3UrZsWfr27UtWVhaff/451atXD3csEREJIT3BUiSCLVu2jKSkJGbOnElaWhpvvvkm2VPqiohIQafxEZEINWTIEJo3b05CQgJLly6lZMmSKtBERAoRFWkiEeLIkSNMnjyZRo0akZaWRocOHVi/fj1PPvkk55xzTrjjiYhIPtNwp0gE+Prrr+ncuTOlS5emb9++xMbGUrRo0XDHEhGRMFKRJhIme/bs4c0336Rx48bEx8fz6quv0rRpUw1piogIoOFOkXy3Z88ennjiCeLi4li0aBFnn30255xzDs2aNVOBJiIiR6knTSSfrFmzhn379lGtWjWysrKYP38+sbGx4Y4lIiIRSj1pIiE2f/58fv/739OwYUNSUlIoWbIkSUlJKtBEROSkVKSJhIC7s3XrVtydp556ikaNGrFu3Tq6desW7mgiIhIlNNwZgWr9/VN2H8j8xfpyZxUPQxo5HYcPHz464XlcXBwTJkxg+vTp4Y4lIiJRSEVaBNp9IJMNL7QJdww5De6OmXHnnXeydetW/vnPf9K6detwxxIRkSgWlUXaiXqaCgr1mEWPHTt28MorrzB+/HgWLVrE0KFDKVu2bLhjiYhIARCVRZp6miQSDB8+nD59+nDrrbfywQcfUKxYMRVoIiKSZ3TjgMhpWLJkCXfffTc7d+7kxhtvJDU1lWHDhpGQkBDuaCIiUsCoSBMJwooVK7j55ptp1aoVV1xxBTExMcTFxfGrX/0q3NFERKSAisrhTpH8kJWVxeTJk6lTpw4xMTG0b9+eCRMmULJkyXBHExGRQkBFmsgxDh06xIgRIxgwYADly5fnjTfe4Oqrr6Zy5crhjiYiIoWIijSRgN27d7Nv3z5iYmKYOnUqQ4YMoXHjxppPU0REwkLXpEmht2XLFh5//HHi4uKYNGkSFStWZPLkyTRp0kQFmoiIhI160qTQyszMpEiRIjRp0oTWrVuzaNEiLrvssnDHEhERAVSkSSE0d+5ckpKSAJgwYQLLli2jeHE9QFhERCJL1A13frP3iJ7IL7nWrVs3OnfuTLNmzXj33XcBVKCJiEhEiroi7YhDyt9ahjuGRInMzEzeffddOnTowJEjR3jqqadYs2YNvXr1onTp0uGOJyIickJRV6SJBOuDDz6gcuXKvPPOO/zhD3/AzIiPj6dYMY3yi4hI5FORJgXK9u3bef7558nIyKBSpUqMGzeOzz//nFatWulOTRERiSoq0qRA+Oabb+jVqxcJCQls2LCB9PR06tevT926dcMdTUREJFc07iNRbfHixcTFxbF582bKli3L8uXLufDCC8MdS0RE5IypSJOo4+58/vnnJCUlkZqaysSJE2nQoAENGjQIdzQREZE8o+FOiRpZWVmkp6ezdu1aHnzwQTp27Mi6des0pCkiIgWSijSJeAcOHOCNN94gISGBoUOHEh8fz7Jly/jDH/5AiRIlwh1PREQkJDTcKREtIyODatWqUbNmTd555x0aNWoEoDs1RUSkwFORJhFn8+bN/Otf/+LAgQO8+uqrzJ07VzcDiIhIoaPhTokoffr04aqrriIrK4u+ffsCqEATEZFCKaRFmpm1MrNVZpZmZv2Os72PmS03syVmNtPMLgtlHolM//3vf3nmmWcAaNu2LWlpaQwcOJBLL700zMlERETCJ2RFmpkVBV4FWgPVgc5mVv2Y3RYD17r7VcA4IClUeSTyzJw5k0aNGnHnnXdSsWJF3J2mTZty7rnnhjuaiIhI2IWyJ60ukObu69w9AxgD/CbnDu7+H3ffH1icC1QKYR6JAJmZmUyYMAF3Z8eOHfTq1YvVq1dz33336WYAERGRHEJ548DFwKYcy5uBeifZvzvwcQjzSBjt3buXoUOH8sILL1CrVi2aNm1Khw4dwh1LREQkYkXE3Z1mdidwLdD4BNvvAe4BiLkwnuTk5PwLJ2fkxx9/pGzZsnz11Vd88cUXPPnkk9SuXZuUlJRwR5PTtG/fPv3uRTG1X/RS2xVe5u6hObBZfeBpd/91YPkJAHd//pj9bgJeBhq7+7ZTHbfERVX80NY1IUgseWnt2rUMGDCA999/n08++eTorADJyck0adIkvOEkV9R20U3tF73UdtHNzBa6+7W5+Wwor0mbD1Qxs1gziwE6AVNy7mBmVwNDgHbBFGgS+dydpUuXUq9ePSpUqMDKlSs1bZOIiEguhGy4090Pm1kvYDpQFHjb3ZeZ2TPAAnefAvQHygAfBC4a/8bd24Uqk4SGuzNjxgySkpJo37499913Hxs2bKBMmTLhjiYiIhK1QnpNmrtPA6Yds+6vOd7fFMrzS+jt37+fG264gYMHD/L444/TuXNnzEwFmoiIyBmKiBsHJLocOHCAd955h8zMTHr37s0rr7xCvXr1KFJEE1iIiIjkFf1fVU5LYmIil19+OdOnT6devewnqtSvX18FmoiISB5TT5qc0qZNm/jvf/9Lx44dOe+880hOTqZatWrhjiUiIlKgqftDTig1NZW77rrrZ8816969uwo0ERGRfKAiTX7G3Vm2bBkAU6dO5YorriAtLY1//vOfYU4mIiJSuGi4UwA4cuQIU6ZMISkpiR9++IFFixbRr1+/cMcSEREptNSTVshlZGTg7rz55ps899xz9OnThxUrVugRGiIiImGmIq2Q2rNnD/379yc2NpYlS5bwxz/+kXnz5nHbbbdRtGjRcMcTEREp9FSkFUKLFi0iLi6OxYsX89FHH1GrVi2KFStGYNYHERERiQC6Jq2QWL16NQMGDKB169a0adOG+fPnExsbG+5YIiIicgLqSSvgdu/ezW233UajRo246KKLaNSoETExMSrQREREIpx60gogd2f69OlkZGTQtm1bbrnlFv79739TunTpcEcTERGRIKknrQBxd0aOHEnt2rV5/PHHATAzunXrpgJNREQkyqgnrQBIT08nNTWVevXq8fXXX/P888/TunVr3QggIiISxdSTFsV++OEHnn76aWJjYxk6dCgA/fv35+abb1aBJiIiEuXUkxaFdu/eTbly5XjiiScAmDVrFgkJCWFOJSIiInlJRVoUSUlJISkpiS+++IK0tDTefPNN9ZiJiIgUUBrujBKvv/46rVu35qqrrmLZsmWULFlSBZqIiEgBpiItQmVlZTF+/Hjq169PWloanTt3Zv369fTt25dy5cqFO56IiIiEmIY7I9DixYvp2LEj55xzDn379iU2NlbzaYqIiBQyKtIixK5du3jjjTdo3rw58fHxDB06lBtvvFFDmiIiIoWUhjvDbM+ePTz++ONUrlyZ1NRUypcvzznnnEPjxo1VoImIiBRi6kkLk5UrV7J//36qV69O0aJFWbRoEZdddlm4Y4mIiEiEUE9aPpszZw7t27fnxhtvJDU1lZIlS/L888+rQBMREZGfUU9aPnB3tmzZwkUXXcQzzzxDmzZtGDlyJKVKlQp3NBERCbPMzEw2b97MwYMHj7u9XLlyrFixIp9TyekqWbIklSpVonjx4nl2TBVpIZSZmcmYMWNISkoiISGBcePG8fHHH4c7loiIRJDNmzdTtmxZLr/88uNei7x3717Kli0bhmQSLHdnx44dbN68mdjY2Dw7roq0EDhy5AhFihThjjvu4IcffmDAgAG0bNky3LFERCQCHTx48IQFmkQHM6NChQps3749T4+rIi0Pbd++ncGDBzNx4kS+/vpr3nnnHUqXLh3uWCIiEuFUoEW/ULShbhzII2+//TZVq1Zl27ZtTJw4kWLFiqlAExGRqFCmTBkANmzYwFlnnUXt2rWPvjIyMk75+UmTJmFmrFy58ui65ORkbrnllp/t161bN8aNGwdkXxLUr18/qlSpwjXXXEP9+vWDviTo0KFDdOzYkfj4eOrVq8eGDRuOu9+gQYOoWbMmNWrU4F//+tfR9Y899hhXXHEFV111Fe3bt2fXrl1Htz3//PPEx8eTkJDA9OnTAVi1atXPfiZnn332z44XKirSzsDixYvp2rUrO3bsoGnTpqxYsYIhQ4ZQpUqVcEcTERHJlcqVK/P1118ffcXExJzyM6NHj6ZRo0aMHj066PP85S9/YevWraSmprJo0SImTZrE3r17g/rsW2+9xTnnnENaWhp/+tOf6Nu37y/2SU1NZejQocybN4+UlBSmTp1KWloaAC1atCA1NZUlS5ZQtWpVnn/+eQCWL1/OmDFjWLZsGZ988gn3338/WVlZJCQkHP15LFy4kFKlStG+ffugv2tuqUjLhWXLltGyZUvatm3LVVddRYkSJYiNjeXCCy8MdzQREZF8tW/fPr766iveeustxowZE9Rn9u/fz9ChQ3n55ZcpUaIEABdccAEdOnQI6vOTJ0/mrrvuAuC2225j5syZuPvP9lmxYgX16tWjVKlSFCtWjMaNGzNhwgQAWrZsSbFi2Vd8XX/99WzevPnocTt16nT0/+vx8fHMmzfvZ8edOXMmlStXzpdHZ6lIC1JWVhZjx45l48aNlCxZkk6dOrF27VoeffTRo93EIiIi0W7t2rVHh/UeeOABABYsWMAf//jH4+4/efJkWrVqRdWqValQoQILFy485TnS0tK49NJLOfvss4+7vWPHjj8bXvzpNWLECAC+/fZbLrnkEgCKFStGuXLl2LFjx8+OUbNmTWbNmsWOHTvYv38/06ZNY9OmTb8419tvv03r1q1/cVyASpUq8e233/5s/zFjxtC5c+dTfse8oBsHTuHgwYMMHz6cAQMGcP755/Paa69Ru3ZtKleuHO5oIiJSAF3e76M8P+aGF9oEve9Pw505XXvttQwbNuy4+48ePZrevXsD0KlTJ0aPHk2dOnVOeCF9MBfYv//++0HnPZFq1arRt29fWrZsSenSpalduzZFixb92T7PPfccxYoV44477gjqmBkZGUyZMuXo8GioqUg7gR9//JH09HRKlCjBp59+yvDhw2nUqFG4Y4mISAF3bEEVyc9J27lzJ59//jlLly7FzMjKysLM6N+/PxUqVODHH3/8xf7nnXce8fHxfPPNN+zZs+e4vWkdO3Zk1apVv1jfp08funbtysUXX8ymTZuoVKkShw8fZvfu3VSoUOEX+3fv3p3u3bsD8OSTT1KpUqWj24YPH87UqVOZOXPm0cLxp+P+ZPPmzVx88cVHlz/++GOuueYaLrjggtP8SeWOhjuPsXnzZh555BEqV67Mhx9+SMWKFZkwYYIKNBERkWOMGzeOLl26sHHjRjZs2MCmTZuIjY1l1qxZVKlShS1bthydLWHjxo2kpKRQu3ZtSpUqRffu3endu/fRu0e3b9/OBx98AGT3pOW8eeGnV9euXQFo164d//73v49maNas2XF76LZt2wbAN998w4QJE7j99tsB+OSTT0hKSmLKlCk/m/2nXbt2jBkzhkOHDrF+/XrWrFlD3bp1j24fPXp0vg11goq0ozIyMjh8+DDNmzfH3UlJSeG+++4LdywREZGwO9E1aaNHj/7FXY6/+93vGD16NCVKlOC9997j7rvvpnbt2tx2220MGzaMcuXKAfDss89SsWJFqlevTs2aNbnllltOeI3asbp3786OHTuIj4/npZde4oUXXgBgy5Yt3HzzzT/LUr16ddq2bcurr75K+fLlAejVqxd79+6lRYsW1K5dm549ewJQo0YNOnToQPXq1WnVqhWvvvrq0SHS9PR0PvvsM2699dbT/Onlnh17N0SkK3FRFT+0dU2eHW/27NkkJiZSvHhxxo8fz+HDh4/e8SF5Lzk5mSZNmoQ7huSC2i66qf0i14oVK6hWrdoJt0fycKf83PHa0swWuvu1uTleoe5J69KlC127dqV169a89957ACrQREREJCIUqookIyODUaNGMXXqVMaOHcvf//53Lr30UhVmIiIiEnEKTU/a2LFjqVy5MiNHjqRnz56YGXFxcSrQREREJCIV6CLt+++/59lnn+XQoUNceumlTJo0ic8++4ybbrpJk9mKiEjEiLbrw+WXQtGGBbJI27hxI/fddx/VqlVjy5Yt7N+/n+uvv546deqEO5qIiMjPlCxZkh07dqhQi2Luzo4dOyhZsmSeHrdAjfUtXLiQ+Ph4tmzZQoUKFVi5ciXnn39+uGOJiIicUKVKldi8eTPbt28/7vaDBw/m+f/8Je+VLFnyZw/LzQshLdLMrBUwCCgKDHP3F47ZXgIYAdQBdgAd3X3D6ZzD3fnss89ITExk9erVTJgwgfr161O/fv28+RIiIiIhVLx4cWJjY0+4PTk5mauvvjofE0mkCNlwp5kVBV4FWgPVgc5mVv2Y3boDP7p7PDAQSAz2+IcPH2bfvn1HJzm/6667WLt2Ldddd11efQURERGRsAnlNWl1gTR3X+fuGcAY4DfH7PMb4N+B9+OA5naKK/rdnddee42qVavy9ttvEx8fT0pKCl27diUmJibPv4SIiIhIOIRyuPNiYFOO5c1AvRPt4+6HzWw3UAH44UQHzdy+kenTp/Pee+/RoEEDAN2pKSIiIgVOVNw4YGb3APcEFg9NmTIldcqUKeGMJLl3HicpwiWiqe2im9oveqntoltCbj8YyiLtW+CSHMuVAuuOt89mMysGlCP7BoKfcfc3gTcBzGxBbufAkvBT+0UvtV10U/tFL7VddDOzBbn9bCivSZsPVDGzWDOLAToBx3Z/TQHuCry/Dfjc9aAYERERkdD1pAWuMesFTCf7ERxvu/syM3sGWODuU4C3gHfNLA3YSXYhJyIiIlLohfSaNHefBkw7Zt1fc7w/CPz+NA/7Zh5Ek/BR+0UvtV10U/tFL7VddMt1+5lGF0VEREQiT4Gcu1NEREQk2kVskWZmrcxslZmlmVm/42wvYWbvB7b/n5ldnv8p5XiCaLs+ZrbczJaY2UwzuywcOeX4TtV+Ofb7nZm5memuswgSTPuZWYfA7+AyMxuV3xnl+IL423mpmf3HzBYH/n7eHI6c8ktm9raZbTOz1BNsNzMbHGjbJWZ2TTDHjcgiLdRTSknoBNl2i4Fr3f0qsmeaSMrflHIiQbYfZlYW6A38X/4mlJMJpv3MrArwBNDQ3WsAD+d7UPmFIH/3ngLGuvvVZN9o91r+ppSTGA60Osn21kCVwOse4PVgDhqRRRohmlJK8sUp287d/+Pu+wOLc8l+hp5EhmB+9wD+QfY/jA7mZzg5pWDarwfwqrv/CODu2/I5oxxfMG3nwNmB9+WALfmYT07C3b8k+ykVJ/IbYIRnmwuUN7OLTnXcSC3Sjjel1MUn2sfdDwM/TSkl4RVM2+XUHfg4pInkdJyy/QLd9Je4+0f5GUyCEszvX1WgqpnNNrO5Znayf/1L/gmm7Z4G7jSzzWQ/OeHB/IkmeeB0/98IRMm0UFIwmdmdwLVA43BnkeCYWRHgJaBbmKNI7hUje8ilCdm92F+a2ZXuviusqSQYnYHh7v6imdUn+zmjNd39SLiDSWhEak/a6UwpxcmmlJJ8F0zbYWY3AX8G2rn7oXzKJqd2qvYrC9QEks1sA3A9MEU3D0SMYH7/NgNT3D3T3dcDq8ku2iS8gmm77sBYAHefA5Qke15PiXxB/b/xWJFapGlKqeh1yrYzs6uBIWQXaLoeJrKctP3cfbe7n+ful7v75WRfU9jO3XM9N53kqWD+dk4iuxcNMzuP7OHPdfkZUo4rmLb7BmgOYGbVyC7StudrSsmtKUDXwF2e1wO73X3rqT4UkcOdmlIqegXZdv2BMsAHgXs9vnH3dmELLUcF2X4SoYJsv+lASzNbDmQBj7m7RiHCLMi2ewQYamZ/Ivsmgm7qnIgMZjaa7H/8nBe4ZvBvQHEAd3+D7GsIbwbSgP3A3UEdV+0rIiIiEnkidbhTREREpFBTkSYiIiISgVSkiYiIiEQgFWkiIiIiEUhFmoiIiEgEUpEmInnKzLLM7Oscr8tPsu/lZpaaB+dMNrNVZpYSmO4oIRfH6GlmXQPvu5nZr3JsG3a8iebPMOd8M6sdxGceNrNSZ3puEYk+KtJEJK8dcPfaOV4b8um8d7h7LeDfZD+L77S4+xvuPiKw2A34VY5tf3T35XmS8n85XyO4nA8DKtJECiEVaSIScoEes1lmtijwanCcfWqY2bxA79sSM6sSWH9njvVDzKzoKU73JRAf+GxzM1tsZkvN7G0zKxFY/4KZLQ+cZ0Bg3dNm9qiZ3Ub2nLIjA+c8K9ADdm2gt+1oYRXocXsllznnkGOCZTN73cwWmNkyM/t7YN1DZBeL/zGz/wTWtTSzOYGf4wdmVuYU5xGRKKUiTUTy2lk5hjonBtZtA1q4+zVAR2DwcT7XExjk7rXJLpI2B6a+6Qg0DKzPAu44xfnbAkvNrCQwHOjo7leSPcPKfWZWAWgP1HD3q4Bnc37Y3ccBC8ju8art7gdybB4f+OxPOgJjcpmzFdlTNP3kz+5+LXAV0NjMrnL3wcAWoKm7Nw1M4/QUcFPgZ7kA6HOK84hIlIrIaaFEJKodCBQqORUHXglcg5VF9nyRx5oD/NnMKgET3H2NmTUH6gDzA1OInUV2wXc8I83sALABeBBIANa7++rA9n8DDwCvAAeBt8xsKjA12C/m7tvNbF1g7r01wBXA7MBxTydnDNlTo+X8OXUws3vI/rt8EVAdWHLMZ68PrJ8dOE8M2T83ESmAVKSJSH74E/A9UIvsHvyDx+7g7qP+v707Zo0qCKMw/J5aNJBCy6hYWAhWimBnJxaChECQIFY2lloI+hOsBAlW0UIFiyBIEIOIIEQF0aigNrYWFkEkYJXPYu6Chg1mm3AN79PtMjsz91Zn55thkrwCTgELSS4AAW5X1ZVNjHH2z4vek4wPa9TdkXiUdlH1JHARODHCs9wHpoDPwHxVVVpi2vQ8gTe0/Wg3gDNJ9gGXgCNVtZJkjnZ59noBFqtqeoT5SvpPWe6UtBXGgG9VtQbM0C6Q/kuS/cDXrsT3kFb2ewpMJtndtRlPMrHJMb8Ae5Mc6D7PAM+7PVxjVbVAC4+Hh/z2J7Bzg37ngdPANC2wMeo8u0uxrwHHkhwEdgGrwI8ke4CTG8zlJXB88ExJdiQZtiopaRswpEnaCjeBc0mWaSXC1SFtpoCPSd4Bh4A73YnKq8CTJO+BRVop8J+q6hdwHniQ5AOwBszSAs+jrr8XDN/TNQfMDg4OrOt3BfgETFTV6+67kefZ7XW7DlyuqmXgLW117i6thDpwC3ic5FlVfaedPL3XjbNEe5+StqG0P3SSJEnqE1fSJEmSesiQJkmS1EOGNEmSpB4ypEmSJPWQIU2SJKmHDGmSJEk9ZEiTJEnqIUOaJElSD/0G8uyq4mQ1LyAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "roc_data = {'IF': {'scores': od_preds['data']['instance_score'], 'labels': y_outlier}}\n",
    "plot_roc(roc_data)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:cdod]",
   "language": "python",
   "name": "conda-env-cdod-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}