demo.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Copula-based conformal bounding box prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/lab/bcyusamu/miniconda3/envs/TBD/lib/python3.8/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "from src.reporting import UQEvalResults, SummaryReporter, GifReporter"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Uncertainty quantification results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "folds, version = 10, \"___release\", #10\n",
    "approaches = ['Bonferroni',\n",
    "              'Max additive',\n",
    "              \"Independent DE-CCP\",\n",
    "              \"Gumbel DE-CCP\",\n",
    "              'Empirical DE-CCP']\n",
    "epsilons = [0.01, 0.05, 0.1, 0.2] #, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9]\n",
    "detectors = ['Correlated D.', 'Uncorrelated D.', 'YOLOv8 D.']\n",
    "filename = f\"results/UQ_experiments{version}.pkl\"\n",
    "res = UQEvalResults()\n",
    "res.load(filename)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Validity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">> Reporting validity ...\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABNwAAAH0CAYAAADmCzDsAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/OQEPoAAAACXBIWXMAAA9hAAAPYQGoP6dpAACssklEQVR4nOzdd1hTVwMG8DeEkUDYKENRVNwbJ2oddeO2dddVa2tdtbhqrQO1rtZRZ7W2rjr7qVSto2pxz4qoVdwiDiwisldI7vcHJXIJUxIC5P09Tx7Nvefee3Igecm5554rEQRBABEREREREREREemEiaErQEREREREREREVJKww42IiIiIiIiIiEiH2OFGRERERERERESkQ+xwIyIiIiIiIiIi0iF2uBEREREREREREekQO9yIiIiIiIiIiIh0iB1uREREREREREREOsQONyIiIiIiIiIiIh1ihxsREREREREREZEOscONiIiIdO7kyZOQSCSiR2aZ1588eVK0vnXr1qL1s2fPLpzKF6Lc2oCIiIiIiidTQ1eAiIiI9Ov06dPYt28fzpw5g2fPniEyMhJmZmZwdXVF3bp10bFjR/Tt2xd2dnaGrmqxFxUVheXLl4uWTZgwgW1LREREZGTY4UZERFRC3blzBx9//DEuXLigtU6pVOLhw4d4+PAh9u7dC19fXzx48AAuLi4GqGnJERUVBT8/P9GyYcOGscONiIiIyMiww42IiKgE+vPPP/HBBx8gLi4uT+Xj4+ORlJSk51qJPX78WPTcGDv72AZEREREJRM73IiIiEqY4OBg9OnTR6uzrU2bNhg5ciRq1aoFMzMzPHv2DKdPn8bmzZsRGhpa6PX08PAo9GMWNWwDIiIiopKJN00gIiIqYXx9fRETEyNa5ufnh7/++gsDBgxA7dq1Ua1aNbRr1w5z5szBo0ePsHLlSshkMk356Oho/PTTT/jss8/QtGlTVKhQATY2NjA3N4eTkxOaNWuG6dOnF6ij7l1uGKBSqbBs2TLUq1cPVlZWcHR0RLdu3XDx4sUsy8+ePVt0jNatWwMANm3ahGbNmsHOzk507IcPH2Lp0qUYOHAg6tWrh7Jly0Iul0Mul6NMmTLo0KEDVq5cqdWZGRISAolEggoVKmjVoUKFCtne/CGvbfD06VNMmzYNjRo1gqOjI8zNzVGqVCm0aNECc+fORURERJbbbdq0SbT/9A6+M2fOoEePHihVqhRkMhmqV6+O+fPnQ6lUZt/4efDixQvMnj0bLVq0QOnSpTW/L3Xr1sWYMWNw9uxZUXkPDw9R/TZt2qS1z9zKDBs2TLR+2LBhSE1NxbJly+Dl5QVra2tIJBJcuXJFVM7GxibbUZ3Vq1cXld25c6dWmbNnz+Ljjz9G1apVYW1tDZlMBg8PDwwaNAinTp3Kto3UajV+/fVXdO3aFe7u7pDJZJDJZHB3d0fjxo3x2WefYePGjXj16lXuDU5ERERFl0BEREQlxq1btwQAokfnzp3zvZ8LFy5o7Serh0KhEPbv36+1fUBAgFbZzDKvDwgIEK1v1aqVaP1XX30lvP/++1nWw8TERNiyZYvWMWbNmiUq17JlS2Hw4MHZHnvBggV5et0VK1YUHjx4oDnO48eP87QdAGHWrFl5bgNBEIQff/xRMDc3z3GfNjY2gr+/v9a2GzduFJUrX768MG/ePEEikWS5nw8++CCb34jcrV69WrCwsMixnj169BBtU758edH6jRs3au03tzJDhw4VrR8wYIDQtm1brWM/fvxYqF69umjZ3r17tY53/fp1rbZNSEjQrI+LixMGDBiQ6895+PDhQnJysmjfKpVK6NKlS55+T7JqCyIiIio+OMKNiIioBDl8+LDWsokTJ+rteHFxcRgwYADCwsL0dox0P/zwA/76668s16nVaowYMQI3btzIcR/nzp3D1q1bC1yXR48eoV+/fgXeT262bNmCUaNGISUlJcdyMTEx+PDDDxEQEJBjudDQUHzzzTcQBCHL9Xv27Mnydyg3q1atwpgxY5CcnJzvbXVt9+7dOHHiRJbrBg8erFU2q+0z6tOnD+RyOQBAEAQMGDAAO3bsyLUeGzduxOjRo0XL9uzZgz/++CPXbYmIiKj4Y4cbERFRCZK5w8nU1BQtWrTI935MTEzQsGFDfPvttzh48CAuX76Me/fu4cqVK1i5ciUcHBw0ZePj47Fhw4YC1z03iYmJqFOnDv744w8EBgbCz88PJiZv/5RRKpWYMWNGjvtQqVQwMTHBtGnTcO3aNdy8eRMbNmzQ3KzAwsICnTt3xtq1a3H06FEEBQXh7t27OHPmDKZMmSI63tWrV3HmzBkAQNmyZfH48WPN84zOnDmDx48fax4TJkzI0+uNiYnRKlumTBns2rULN27cwM8//yy6+2lqaio+//xzqFSqbPcpCALMzMywdOlS3Lp1C+vXr4eFhYWoTFadUDl59uwZJk+eLFqmUCiwcOFCBAYG4s6dOzh48CCGDBkCMzOzfO37XahUKsjlcixatAg3b95EYGAgfvjhBygUCgwePFj0Mzxw4AASExNF2//222+i50OHDtX8f/fu3Thw4IDmuVwux+LFixEYGIgbN25g2bJlokuzf/75Z9FltKdPnxbte/Dgwbh06RLu37+Pq1evYseOHRgzZgzn9iMiIioJDDzCjoiIiHSoc+fOosvSnJ2d9XKcJUuWiI7ToUMH0Xp9XFIql8uF8PBwUZmJEyeKykilUiE6OlqzPvMlpQCEmTNnvvPr7tatm2hf8+fPF63P6tLSx48fZ7u/nNrg559/1lofFBQk2v7333/XKvPXX39p1me+pBSZLmkVBEEYP368aH2DBg3y1SZ+fn5ax/jzzz+zLBsVFSV6ro9LSgEIv/zyS7b1zXy56W+//aZZd+3aNdG6ChUqCGq1WrO+devWovU//vij1v5nzJghKvPRRx9p1o0aNUq07uLFi1nWUa1WC3Fxcdm+BiIiIir6OMKNiIioBJNIJO+87cuXLzFnzhy0atUKrq6ukMlkmknkM1+m+uLFi4JWNVedO3dGqVKlRMuGDBkieq5SqRAYGJjtPszNzXMdYRYUFITx48ejYcOGmhsUpL/ujKObAP2+7sw3GKhfvz7q1q0rWtatWzetNsm8XWajRo0SPa9cubLoeVRUVL7qmXnUVsuWLdG+ffssy9ra2uZr3+/C1dVV69LRjDL/zuzatUvz/8yj+wYPHqx5D6lUKpw/f160ftSoUVo3vpg7d66oTMafR/369UXrunXrhuHDh2PJkiU4fPgwnj17BiDtfWtlZZXbSyUiIqIizNTQFSAiIiLdydz5EhERgeTkZK3LBnMTEBCAHj16IDY2Nk/lM9+1Ux+yuswuq2X//vtvtvtwd3eHvb19tuu/++47TJ06Nds5zjLT5+t++fKl6HlWd0BNv/Noxjta5jSfnrW1teby2XTp85OlS01NzVc9M3c61qlTJ1/b61qNGjVgapr9n7gffPABRo8ejfj4eADAoUOHEB8fDysrK63LSTN2zr1+/TrXufSykvHn+NFHH2HVqlW4efMmAODVq1dad12tXLkyPv/8c4wbNy7H10FERERFG0e4ERERlSCZOztSU1NzHfGUWUpKCgYNGpTnzjYAee6g0rX8juBzc3PLdt3169fx1Vdf5eu16PN153Xfmcvl1CYZ595LJ5VK81exXBRkVGVW889FRkbmax85/YwBwMrKCh988IHmeUJCAg4cOIDAwEA8ePBAs7x58+aoVKmS5vm7/qyTkpI0/7e0tMSFCxcwb9481K5dO8u2un//Pnx9fTFmzJh3Oh4REREVDexwIyIiKkE6d+6stWzJkiX52sf58+e1RkmNHTsWFy9exMOHD/H48WPMmzevQPV8FyEhIVrLHj9+rLXM2dk5231knDA/s3379kGtVmuey+VyLF++HEFBQZobHnTt2jV/lS4AV1dX0fOsXqsgCHjy5IloWeYRbPqWuYMrtzvFZpS5sy8hIUH0/OXLl/nq+AVy/hmny3xZ6e7du7UuJ81cxsnJSeumD7/++qvohhjZPTKysrLC9OnTcePGDcTExODKlSv49ddf8cknn4g64DZs2IDXr1/n6TUTERFR0cMONyIiohKkRo0a6NSpk2jZ4cOHteaVykilUmH16tWaS98yXyJoZ2eHlStXokmTJqhYsSI8PDxynCdNXw4fPiy6dBIAtmzZInoulUrh5eX1TvvP/Lo7duyIL774AnXr1oWHhwdKlSqVa2dSVnfhzHwXzLzKfHfZa9eu4fr166JlBw4c0GqTd7krbUG0bNlS9PzUqVM4ceJElmWjo6NFzzNf3nv37l3R8/Xr1+ughtratGkDd3d3zfPDhw9jx44dmucWFhbo27evaBupVApvb2/Rsv3798PDwyPbR0REhGhOvFevXolG8SkUCjRs2BCDBg3CTz/9JJqjT61W4/79+7p6yURERFTI2OFGRERUwixduhQ2NjaiZTNnzsT777+PnTt34p9//sGdO3dw/PhxzJw5E5UqVcLYsWM1l745OjqKto2KisLcuXNx+/ZtXLhwAcOHD8fevXsL7fWkS0xMRLt27XD48GEEBQVh7ty5WLZsmaiMj4+P1mvPq8yv+6+//sLOnTtx9+5dHDlyBB06dEBoaGiO+3BwcNAaYbVixQrcvn0bISEhCAkJyfMcaR9++KFWh1SXLl2we/du3Lx5Exs3bsTQoUNF66tUqaLVAaZvw4cP15ojsHfv3li8eDGCgoJw7949HD16FJ9++ilGjhwpKlerVi3R8w0bNmDDhg0IDg7GihUr8O233+qlziYmJvjoo480z5OSkkQ/2x49esDOzk5ru8w3nNi9ezd69OiBQ4cOITg4GLdu3cLhw4fh5+eH+vXro1GjRggKCtKU37VrF8qVK4cxY8Zg586duHr1Kh48eICbN29iyZIl+Oeff0T7540TiIiIijHD3SCViIiI9OXo0aOClZWVACDPj8ePHwuCIAhxcXGCvb19jmWdnZ1Fz8uXLy86fkBAgNY2mWVeHxAQIFrfqlUr0XpTU9Mc62RqaioEBQWJ9jFr1ixRmVatWmXbZufPn8+1jTK/7qFDh2rtp379+jnu4/79+3lug82bN+f552dqaiqcOHFCtP3GjRtz/DnltUxuli9fnqc69ujRQ7TdH3/8kes2EolE9Hzjxo2ifQwdOjTXn0lWgoODsz3mwYMHs9xGpVIJXbp0ydf7KmN9V65cmeft3N3dBZVKlY+fAhERERUlHOFGRERUAnXo0AFXrlxB06ZN81TeysoKMplM8/81a9ZkOxdWs2bNMGvWLJ3VNa/GjRuHRo0aZbnOxMQE69evF12Sl1/e3t4YPXp0tuu//PJLrct1szJ16tR3rkNmQ4YMwdq1a2Fubp5jORsbG/z22294//33dXbs/Pjiiy/www8/5FrPzHx8fNC7d+9s18+aNQvlypUraPWyVK1atSx/n0qXLo2OHTtmuY2JiQl27dqFQYMG5ekY5ubmWncOzgtra2ts3bo1T/PRERERUdHEFCciIiqhqlevjgsXLuDkyZMYP348vLy8UKpUKZiZmUEul6NixYro1asX1q1bh2fPnokm2+/fvz8CAgLQoUMH2NjYQC6Xo0aNGpg3bx4CAgIgl8sL/fXY2NjgzJkzmD9/PmrVqgW5XA47Ozt06dIFZ8+exfDhwwt8jNWrV2PDhg1o0KABZDIZbGxs0KxZM2zfvh1Lly7N0z769euHffv2oXXr1rCzsyvQXTuBtMsY79+/j6lTp6JBgwaws7ODqakpHBwc4O3tDT8/Pzx48AA9e/Ys0HEKavz48Xj48CFmzJiBZs2awdHRUVPP2rVr4/PPP8fEiRO1ttu5cyfmz5+PatWqwdzcHPb29vDx8UFAQABmz56t1zpnviQXAAYOHAhTU9Nst7GyssKvv/6Kixcv4tNPP0XNmjVhY2MDqVQKW1tb1K5dG4MHD8amTZsQFhaGLl26aLYdMmQIfv/9d0yaNAnvvfceKlWqBIVCoWmnJk2aYPr06bhz5w5atWqll9dMREREhUMiCHq8nz0REREREREREZGR4Qg3IiIiIiIiIiIiHWKHGxERERERERERkQ6xw42IiIiIiIiIiEiH2OFGRERERERERESkQ+xwIyIiIiIiIiIi0iF2uBEREREREREREekQO9yIiIiIiIiIiIh0iB1uREREREREREREOsQONyIiIiIiIiIiIh1ihxsREREREREREZEOscONiIiIiIiIiIhIh9jhRkREREREREREpEPscCMiIiIiIiIiItIhdrgRERERERERERHpEDvciIiIiIiIiIiIdIgdbkRERERERERERDrEDjciIiIiIiIiIiIdYocbERERERERERGRDrHDjYiIiIiIiIiISIfY4UZERERERERERKRD7HAjIiIiIiIiIiLSIXa4ERERERERERER6RA73IiIiIiIiIiIiHSIHW5EREREREREREQ6xA43IiIiIiIiIiIiHWKHGxERERERERERkQ4ZtMPt9OnT6NatG9zc3CCRSODv75/rNqdOnUKDBg0gk8lQsWJF/Pjjj/qvKBERFUvMGSIi0jdmDRERZcWgHW7x8fGoW7cuVq1alafyjx8/ho+PD9577z1cu3YNX3/9NcaPH489e/bouaZERFQcMWeIiEjfmDVERJQViSAIgqErAQASiQT79u1Dz549sy0zdepU7N+/H8HBwZplo0aNwvXr13HhwoVCqCURERVXzBkiItI3Zg0REaUzNXQF8uPChQvo0KGDaFnHjh3x888/Q6lUwszMTGub5ORkJCcna56r1WpERkbC0dEREolE73UmIirpBEFAbGws3NzcYGJSvKcGfZecAZg1RET6ZuxZw5whItIvfeRMsepwe/nyJZydnUXLnJ2dkZqaioiICLi6umpts2DBAvj5+RVWFYmIjNbTp09RtmxZQ1ejQN4lZwBmDRFRYTHWrGHOEBEVDl3mTLHqcAOgdQYn/YrY7M7sTJs2Db6+vprn0dHRKFeuHO7duwcHBwf9VbQYUSqVCAgIQJs2bbIdvWFM2B7a2CZixtge8dHJ2L/sApKTzRBv+hqtLm2ARXIMpg23h/f9AViwbjSsra0NXU2dyG/OAMya3BjjeyYnbA9tbBMxY2yPhPgU7Fh4AjbqJOzx3Iy5m6IAAOM/k8JUsEWHW/3gt36M0WYNcyZnxvieyQ3bRIztoc3Y2iRVpUbfdRfQIDYALeR74LbHFkBazthJVCj3pDF+/v4XneZMsepwc3FxwcuXL0XLwsPDYWpqCkdHxyy3sbCwgIWFhdZyBweHbLcxNkqlEpaWlnB0dDSKN1pu2B7a2CZixtgejk+2YrDCDzuEuZCllseDWuNRM3gz+jwegkS1JYCcO6SKi3fJGYBZkxtjfM/khO2hjW0iZoztYXFrG1wlJohBRbwfOg4W0mWQJUdhRJwJXr8cAwtBBsB4s4Y5kzNjfM/khm0ixvbQZoxtsrR2CKoEbUQHtzL4JTUVAKCWCQgzl6Cc2zkAus2ZYjUBgre3N44dOyZa9ueff6Jhw4ZG8wtCRFToArcC+8fhxT4zNDmzArLEV0iSl8JVr0kQUktDlvja0DXUGeYMEZEBBG6F4q8x6G4zA4JpOGxSnBBYbwKibSrA9OlXsE12QpxppKFrqTPMGiIiAwjciuqXF+GYtCxMUt92qn0eEQOLFAEP1KV1fkiDdrjFxcUhKCgIQUFBANJukR0UFITQ0FAAaUOnhwwZoik/atQoPHnyBL6+vggODsYvv/yCn3/+GZMmTTJE9YmISj6VEvjDF0DapS6nIkIRdeE7UZEq938zQMXyhjlDRFTEZciZF/vM4XFiEY5d+hEJMkdc9ZqEJLUzZImv0PrST4auabaYNURERVzgViT7T8Wex354dW8aqsx+gyiVCgDQbLMVflpljT5Bw3V+WIN2uP3999+oX78+6tevDwDw9fVF/fr1MXPmTABAWFiYJqgAoEKFCjh06BBOnjyJevXqYe7cuVixYgU++OADg9SfiKjEk5oBXZYCkOBlo6eY/G8Ypjy4hkcvb2mKnPfqa7j65YI5Q0RUxGXImTepqRj38DZ+D/oNh/7erClSM3gzLJJjDFfHXDBriIiKsP9O7CgFGZLNFNh4eQs2vXqBEc9fQC0ISLKwQ2C9CUiS6/7yfIPO4da6dWvNBKFZ2bRpk9ayVq1aITAwUI+1Spu0NDU1Far/ejxLOqVSCVNTUyQlJRnNa86JsbeHVCqFqalpiZgjhXTEazBuPXyG3ou+QapKQF2P5qhm7YTagd/jauMhUKQU3cmai2rOAIBKpYJSqdT7cYoCY/9czczY24M5Q1q8BiMhMRmTNoxCiFIFW4UjWtToplkd0HQY/iy/CiiiA8CKatYY03caY/9czYqxtwmzhjT+O7Gj2D8O9x5+jkv3omAiMUF77zGIjb2JWw0HI0ldCnEmobnvK5+K1U0TCkNKSgrCwsKQkJBg6KoUGkEQ4OLigqdPn/IDCWwPALC0tISrqyvMzc0NXRUqAp49e4aOE1YjKkmNii618EHnT9Hw3ArIkqOwu/oqtLn9iaGrWOzExcXh2bNnOX5BK0n4uSrG9mDOkJhKpcKg7//AhecqWFpYYqzPYriYCKgZ+D0Cmg6DTYoTOjwcjpM4Y+iqFhvG9p2Gn6va2CbMGsrAazDW7fkL353fBADo0eYTVKncEVfREVAD0RYROFppo84Pyw63DNRqNR4/fgypVAo3NzeYm5sbxYeTWq1GXFwcFAoFTEyK1X009MKY20MQBKSkpODVq1d4/PgxKleubHRtQGLR0dHw8fHB8xfPUaaUByZ+MAvl3WZhiK8J0iIkGtcqFd15dYoilUqFZ8+ewdLSEqVKlWLOGCFjbg/mDGUmCALGjx8Pf39/mEnN8WnHb1HOSQ6vs8shS47Ckaor0OXeOJgn8k6ceWWM32mM+XM1O8bcJlllDRm3/fv3Y/TCLQCAaa0scLPLc+De2/V/Vd6K6qnhuKDj47LDLYOUlBSo1Wq4u7vD0tLS0NUpNGq1GikpKZDJZEb3YZwVY28PuVwOMzMzPHnyRNMOZJySk5PRq1cv3Lx5Ey4uLjgRcAxlnMtC8TgRLQ6OxxWZBeQC4Nl4DPbjvKGrW2wolUoIgoBSpUpBLpcbujqFwtg/VzMz9vZgzlBGCxcuxJo1ayCRSDCu/xzU82iMnh1CoHa+g8syC3wb9wa1eiux+5CdoatabBjjdxpj/1zNirG3SeaskUqlhq4SGcjFixfRv39/qNVqjBgxAl8NboOdv5kg48QuHe5/hFZtnuMXTNTpsdnhlgVj/EAiyojvAQLS5r6oXLkyrly5gsOHD6NqNc+0FfaDYQug3R++QJeleF3eB8DnhqxqsVTSRxsQ5YQ5Q+k8PDxgZmaG77//Hp+NHA1lUioU9s0AhQrt/8sZeH2ETu4vgMWGrm3xwvcZGTu+BwgA7O3t4ezsjBo1auC7+cvx+4obUKYmQSF9CQfntXgZPhpIdsb1C3Y6PzY73IiIKEumpqb48ccfMWXKFFSqVEm80mswULd/2iSkr18bpoJERFTsDRgwAI0bN9bkjIX8v68nGXMGgJWthaGqSERERUxyqgrHb4cjJYcbgphLpWhXozSqVq2KCxcuAEpTHFxxAzERSbBxkqFnB8D61B3E9hXg/6cM4eFxOq8nO9yIiIydSqn5QgMAx48fR+vWrTV3dtLqbEuXYRsiIqIcZciamzdvwsHBAWXKlAEA5gwREeVL4JMojNme4U7PJgkwtXoIQW0BVXwlqFNSkBL+CPv8RsC7kiNcXFyQnJgKuXXaDTR6+nrB2qEZ0LwPrKVm6FkrCTsWndJ5PTnGkvIlISEBH3zwAWxsbCCRSBAVFWXoKmkJCQmBRCJBUFCQoatCVPQFbgXmu6X9C+B///sfOnTogB49eiAlJcXAlSNjxJwhKoEyZE1ISAg6dOgAb29v3Lt3L/dtifSAWUNUvDXysIe7vRwSAKbWN2FddQ7kZbfBstwvsCy/CBH75yB8x9d4dOmoZhsLuSm6ja+HXhO9YO3w3/yx/53YsXaQodNntXVeT3a4lRDDhg2DRCLRPBwdHdGpUyfcuHFDp8fZvHkzzpw5g/PnzyMsLAy2trY63b8uuLu7IywsDLVq1TJ0VYiKnORUFf64EYZ9154h0H8Fog6Ox58WUpw9Ngnrvx2HgYM+giAIcC9XHmZmHFlAbzFn3mLOEOUsc9YI+8dCUKUgYudotG7RBC9fvoRUpoCtA+88SmLMmreYNUTZM5Wa4Mv2VSCYJEPmtlOzXBAEhO24jcSHN2FmbgbPTCOoLeSmUNhnfbMmfUxdwEtKdSg/1xFbmOr+LimdOnXCxo0bAQAvX77EN998g65duyI0NFRnx3j48CGqV69eoA9+lUoFiUSiNYllSkoKzM3NC1pFSKVSuLi4FHg/RCVR+vDrPtKTsLTdiqHlywIAkp4nIWT+GqSmqCGv3BQffTmbk/oXQcyZvGHOEBlWxqzpYboeAJCUKqDHjkQ8eR4HaxtrKNtPxaNowNnJsHUlMUPnDMCsScesIcpZ97puWHTmHlSpqdi6JO0zq1UZa0SdiQIkwPrN69GkSROD1pEj3HQo/Y+LL3ddz/YxZnsgAp9E6eX4FhYWcHFxgYuLC+rVq4epU6fi6dOnePXqFYC0+TLef/99yOVyODo64tNPP0Vc3NuJAYcPH46ePXvi+++/h6urKxwdHTFmzBgolWk3zG3dujWWLFmC06dPQyKRoHXr1gDSQmXKlCkoU6YMrKys0KRJE5w8eVKz302bNsHOzg4HDx5EjRo1YGFhgSdPnsDDwwPz5s3DsGHDYGtri5EjRwIA9uzZg5o1a8LCwgIeHh5YsmSJ6HV6eHhg/vz5+Pjjj2FtbY1y5cph/fr1mvUcfk2UvUYe9vjM6iImmuzFtNJp33KUb5R4suQJUhPUcKlghWkDWqNpJX4DKoqYM8wZouKgkYc9PrM+h0Wm6yEB8EYiQfvfU3D+mQp2MuDcQDXGlrmHRh72hq4qZWLonAEKnjXDhg1j1hAZAVOpCRpWfntyYHdUFF79dQUA4DbUDbVa1TBU1TTY4aZDGa8jzopEApRzsCyUPy7i4uKwbds2eHp6wtHREQkJCejUqRPs7e1x5coV/Pbbbzh+/DjGjh0r2i4gIAAPHz5EQEAANm/ejE2bNmHTpk0AgL1792LkyJHw9vZGWFgY9u7dCyDtC9S5c+ewc+dO3LhxA3369EGnTp1w//59zX4TEhKwYMECbNiwAbdu3ULp0qUBAN999x1q1aqFq1evYsaMGbh69Sr69u2L/v374+bNm5g9ezZmzJihqUO6JUuWoGHDhrh27RpGjx6Nzz//HHfu3NFfgxKVEKor21E53A3+b+bBOt4OP89LBr4KgTJSCZmzHCPa/QDPyHJQXdlu6KpSFpgzzBmi4sD0+jZ8pVwNCYBjZjKM+0GCc7eSITGVwHNcOVi7WWBqyiqYXt9m6KpSJkUpZ4C8Z824ceNE2zFriEo+dUICXOXJsFACJ+PiMOfflwAA166l4NrMHlHJUYatIHhJqU6lX0fsu/t6lusFAfiyfWWYSvXTz3nw4EEoFAoAQHx8PFxdXXHw4EGYmJhg27ZtSExMxJYtW2BlZQUAWLVqFbp164YFCxZALpcDAOzt7bFq1SpIpVJUq1YNXbp0wYkTJzBy5Eg4ODjA0tIS5ubmmuHNDx8+xI4dO/Ds2TO4ubkBACZNmoQjR45g48aNmD9/PgBAqVRizZo1qFu3rqjO77//PiZNmqR5PmjQILRt2xYzZswAAFSpUgW3b9/Gd999h2HDhmnK+fj4YPTo0QCAqVOnYtmyZTh58iSqVaum62YlKjkCt0L5x2wkqudCpXJGp7vjcVuYg9CUFDhYSTG802K4mlRAvPAvlH98AwtzAfAabOhaUwbMGeYMUZEXuBXYn9b5kWAiwXRre0QlPIYEQIWPyyCppg3mJ5hh3b+vNOWYNUWHoXMGePesmT59OmxsbAAwa4iMwV2vBvgAwAcA5sTHQQ2gt60t5t5zgGSJCgnd7QxbQXCEm851r+uW5Vmh9LNB3eq46e3Ybdq0QVBQEIKCgnDp0iV06NABnTt3xpMnTxAcHIy6detqggkAmjdvDrVajbt372qW1axZE1Lp2/kYXF1dER4enu0xAwMDIQgCqlSpAoVCoXmcOnUKDx8+1JQzNzdHnTp1tLZv2LCh6HlwcDCaN28uWta8eXPcv38fqgxzSWTcl0QigYuLS471JDJ6KiXwhy8U0gj0sJkBlVk4bFKcoGo+A2uqe2NsZz9UkNVAjHkEyrrNhUIaAfzhm7YdFSnMGeYMUZH1X9YAApJSrPCnSVmYm5ni13LlsdTNDV97SGCRIuCO2gUvJNYABGZNEWTInAHePWsyjkRj1hAZlxmlnTHfxRWznF0081C7W7sbuFYc4aZz2Z0VKoyzQVZWVvD09NQ8b9CgAWxtbfHTTz9BEIRsJ0DPuDzzXQklEgnUanW2x1Sr1ZBKpbh69aoo1ABozkwBgFwuz/L4GcMSQJb1FARBa7v81pPI6EnNgC5Lgf3j8HyfOWpLv8ejhpORJC8FvDcXpQHIEl+h2YXlUPePACBJKy/lnUqLGubMW8wZoiLmv6xJ9p+KjXemwMLaFav+txwyaRQ6WtsAmwEvC2sE1huH47Zv0N9uDiy6L2LWFDGGzBmAWcOsIcqb0ieOQyo3Rc/dnbB2eRJ62trik/FSJJsBjVwaY61laUNXkSPc9CHzWaHCOhuUWfpdcxITE1GjRg0EBQUhPj5es/7cuXMwMTFBlSpV3vkY9evXh0qlQnh4ODw9PUWPd7mrTo0aNXD27FnRsvPnz6NKlSpa4UdE+eQ1GOi+Er++icQHd4OAKz+IVtcM3gxZchTcU1OB7it5iU8RxpxhzhAVWV6DcbX8WEw6OB2H759GYL0JSLKwAwAkWdilPZeXQrLaFsp23zFriqiikjNpx85b1mTspMsvZg1R8RIXF4dOPXti+IhRmNR0umZ5shngaF8GX7WaZcDavcUONz1IPyuUfg6jsM4GJScn4+XLl3j58iWCg4Mxbtw4xMXFoVu3bhg0aBBkMhmGDh2Kf/75BwEBARg3bhwGDx4MZ2fndz5mlSpVMGjQIAwZMgR79+7F48ePceXKFSxatAiHDh3K9/4mTpyIEydOYO7cubh37x42b96MVatWieZEIKJ3t+exJRa8eoU3KhX+Z+8hWhfQdBjWfmqN0l1+4BegIo45w5whKqqePHmCD2esQqIyATdCTyLewgGB9SYg2qYCApuPR5K8FGLMI1Cr63MoWg4ydHUpG4bKGeDdsuajjz7S3MDgXTBriIqP1NRU9OvXD3///TfOnDmDqqZv5zz8oXxv/NH7D5S3KW/AGr7FDjc9ST8rBBTe2aAjR47A1dUVrq6uaNKkiebOPa1bt4alpSWOHj2KyMhINGrUCB9++CHatm2LVatWFfi4GzduxJAhQzBx4kRUrVoV3bt3x6VLl+Dunv9rpr28vLB7927s3LkTtWrVwsyZMzFnzhzR5KJE9G7Onj2LQYMGQRAEvF+3C7rUGgFZ4is0CPweMeYRsElxQv0Xfoj16GPoqlIeMGeYM0RFTWRkJDp16oSwsDDUqlUL33zXFfGyN0iSl8JVr0lIUjsj2iICv9daCc+WQwxdXcqFIXIGeLesWblyZYGPy6whKvoEQcCoUaNw6NAhyOVyHDx4EJVTrqF6/zBU7/8Czc+vgGnQDkNXU0MiZHUxeQkWExMDW1tbREREwNHRUbQuKSkJjx8/RoUKFSCTyQp8rL2Bz+C7+zqW9auLXvXLFnh/+qJWqxETEwMbGxuYmLAPlu2h/V5QKpU4dOgQfHx8tOaaMEbFsT3SJ+998+YN6ldugeGtZ8LORo1ax2ZDlhyFsP4SPI9fjphYU9g4ydBrohcU9nn7HHz9+jWcnJwQHR2tuTuYsSusrGHOFE9sD+ZMbopjeyQlJaF9+/Y4e/YsypYtiwsXLsDe2R79Vw9EpztfaMrtq7UMFaq44KcOP+Vr/8waMeaMGD9XtbFNxO8FqVRa7D5X9a04Zo2fnx9mz54NExMT7Nu3D93LRv93x+uM3VqSd5oeRx85w5sm6FGv+mVQqZQCdcraGroqRGTEXrx4gU6dOuHNmzdo0rgJxnVdAHWyCXr6esF64GjgD19U77IUsR4t4b80EHJrc5jJGA/FAXOGiIoClUqFjz76CGfPnoWtrS0OHz6MsmXLIjYyCd1ejkfGe5B2eDgM3bvXM1RVKZ+YM0RUVGzYsAGzZ88GAKxZsyabzjakPd8/Lu2/Bp4mh9+o9EgikaCuu52hq0FERm7WrFkIDQ1F5cqVcfCPg7C2soMyKTVtBJvDYKBuf0BqBmsAvSZ6wUxmCgs546E4YM4QUVFw4MAB7NmzB+bm5vD390etWrUQG5kE/6WBUEZJoHA0h0OneLw8ZAa8sceV9f+ijG8ZWDsU/IoS0i/mDBHpWnKqCsdvhyNFpcq2jLlUinY1SsPCNO0mI69fv4avry8AYPr06fiskSWS/adCKThAIX2ttX2cygFm/lNhARi0043fqIiISrgffvgBgiDg66+/hpOTEwCIO9Skb4eQ5/UyUiIionQ9e/bE0qVL4ebmhtatWyPuTVpnW0xEEmycZGkjqh1kiK35drn/0sB8TV9AREQlQ+CTKIzZHphruR0jm8K7Utol846Ojvjzzz+xa9cuzO1VBcn+U3HgzQwkqm3R0+EbWGfodItVOcI/ch7kJtHoZuBON3a4ERGVQIIgQCKRAAAsLS2xYcMGA9eIiIhKmoxZ8+WXX2qWm8lMIbc2BwBNZxsAWDukdb5x+gIiIuPVyMMe7vZyPHuTqHUxKABIJIC7vSUaediLcqZp06Zo2qgBMN8NSsEaiWpbxKhc4B85T9Pplt7ZFqNyAQAoBRks/vDVXNFT2JhyRETFRH6GX/vNnAFTU1P4+flpQoqIiEhX/vrrLyxevBg7d+6EnZ2daJ2F3BTdxtd7O31BBtYOMk5fQERkxEylJviyfRX47r6etsAkAaZWDyGoLaCKrwRBkOLL9pUR+ToCH3zwAX744Qd4eXmllZWaAV2WQrF/HHo6fKPpXPOPnId2tstxPHoCYlQusJG+RE+Hb6CQRgJdVhqksw1ghxsRUbGR1+HXH8pvY8mCBQCA9u3b47333tN31YiIqKRSKbW+qNy4cQO9evVCTEwMFi5ciIULF2ptZiHPvkONl5ESERm37nXdsOzYPYSlXoa87DbNcnWKHexixuL9SrZo364trly5gmHDhiEoKOjt3Xb/uzzUOlOn297ItCxK72yzlka+091Kdck47w9MRFQMpQ+/zjhezRSpmv9LJIDli6tY6jcVADB37lx2thER0bsL3ArMd0v79z+hoaHo3LkzYmJi0LJlS80d44iIiPLKVGqC0e+Xhcxtp2i5iXkUHN0PYtDAAbhy5QocHBzw22+/ve1sS+c1GOi+EtbSSLSzXS5a1c52eZHobAM4wo2IqNjIPPy6h9mf6GT9P+xStsfJ2N5IfHoPL/43H4Ig4NNPP8X06dMNXGMiIipOMk5dUP7JXtQPmpG2Yv9YXAt9g1v272P6x73x4sUL1KxZE/7+/pDJOFqNiIjyT2YXDIlJ2lQ58Y9Hw1RxH+ZOf+Lsmj/w5tQbyGQyHDx4EFWrVs16B16DERsnxfFd4sXHoyegZz/A2mugnl9B7jjCjQzOw8MDy5cvz7GMRCKBv78/ACAkJAQSiQRBQUE5btO6dWtMmDBBJ3UkKiq613WDu70c79nuxF+ef2GKswOulr0KhdXXeL3PDynJyejWrRtWr17NuduI/sOcIcqb9KkLzv/vB9S7NgMQkDaqWgCqXfkGXw/ujKeP7qOUsysOHz4Me3t7Q1eZqEhgzhDl39PYJ7BIEbB7QSr+2LkC0ujyeLX/Fd6cegMTExPs2LED3t7e2W4fG5kE/z89NHO29Xb4CjbSl2lzuv3pgdjIpEJ8NVljh5u+qZSFcphhw4ZBIpFg1KhRWutGjx4NiUSCYcOGFUpd9M3d3R1hYWGoVasWAODkyZOQSCSIiooSldu7dy/mzp1rgBoS6Y+p1AQLqgTilus1zTJ1shr3VtxDakIcmjRpgp07d8LUlAOYjQZzRueYM2SsGnnY4zPrc1hkuh4SpE1VgP/+HXsoCc+ePoOFhQWOHjkMd3d3g9aVChFzRueYM0RAdHK06Lk05AHC94UDAMbOGYuePXtmu23cmyT4Lw1ETEQSbJxk6NkPcJU/Rs9+gI2TDDERaevj3hi2040dbvqUxbwX+uTu7o6dO3ciMTFRsywpKQk7duxAuXLlCqUOhUEqlcLFxSXXDgUHBwdYW1sXUq2ICkngVrx+uBQmSmD3glTsXpCKL6JjULpXaVi4WeCX2b1gaWlp6FpSYWHO6AVzhoyV6fVt+Eq5GhIAgkqC4J1uCN7phnOmMnzhbY5K9hIc6idFffUNQ1eVCgtzRi+YM2Ts1AkJsBcsYZGhP39Ez6awqaWAS2cnDPlkSI7bm8lMIbc2T+ts8/WCdcuBwNcvYN1yIHr6esHGSQa5tTnMZIYdhMAON30J3ArsHweoUtL+LYSQ8vLyQrly5bB3717Nsr1798Ld3R3169cXlT1y5AhatGgBOzs7lCpVCv369cPDhw8167ds2QKFQoH79+9rlo0bNw5VqlRBfHx8lsd/+PAhevToAWdnZygUCjRq1AjHjx8XlQkPD0e3bt0gl8tRoUIFbNu2TWs/9+/fR8uWLSGTyVCjRg0cO3ZMtD7jEOyQkBC0adMGAGBvby8685VxCPa0adPQtGlTrWPVqVMHs2bN0jzfuHEjatasCRcXF9SoUQNr1qzJ8rUSGUTgViT7T8UTlBItbpyYBPvm9qjzjRdir64ptD+KycCYM8wZIl367zMlWW2JeLUjTsjfzs32hXMpTPFyx74xFdGsvE2hfeaQgTFnim3OVK9eHZaWlmjcuDHWrl2b5WslMqS7Xg3Q9uMN2LBCpVnWafo8nE0pgxMPHeFunfMoagu5KbqNr4deE71g7fBfXv13R21rBxl6TfRCt/H1sr1bdmFhh5s+pIcThP8WCIUWUsOHD8fGjRs1z3/55Rd8/PHHWuXi4+Ph6+uLK1eu4NixYzAxMcEHH3wAtVoNABgyZAh8fHwwaNAgpKam4siRI1i3bh22bdsGKyurLI8dFxcHHx8fHD9+HNeuXUPHjh3RrVs3hIaGasoMGzYMISEh+Ouvv/C///0Pa9asQXh4uGa9Wq1G7969IZVKcfHiRfz444+YOnVqtq/X3d0de/bsAQDcvXsXYWFh+OGHH7TKDRo0CJcuXRKF8K1bt3Dz5k0MGjQIAPDTTz9h+vTpmDt3Li5duoR58+ZhxowZ2Lx5c7bHJyo0/3W2HYicATyZBrs4W6yNiMCr1FQoUgQ4xNmiV/AE/BP+NZL9p/KLUEnHnGHOEOmSSgn84YtktRwHI2diz+t5WGpZCUdiY3A5IR4WSiAp2RHHo77FwTczkayWA3/4FtqlhmQAzJlinTPffvstbt26hRkzZmDmzJnMGSqynqakYFNkJAQh7bPGVCKBRCJBacvSuW5rITeFwj7rG/co7GUG72wDeJdS3dMKp3T/hRSg11vTDh48GNOmTdOcNTl37hx27tyJkydPisp98MEHmv+r1WqsXLkSlStXxu3btzVzCaxbtw516tTB+PHjsXfvXsyaNQuNGjXK9th169ZF3bp1Nc/nzZuHffv2Yf/+/Rg7dizu3buHw4cP4+LFi2jSpAkA4Oeff0b16tU12xw/fhzBwcEICQlB2bJlAQDz589H586dszymVCqFg4MDAKB06dKws7PLslytWrVQp04dbN++HTNmpN1xa9u2bWjUqBGqVKkCAJg7dy6WLFmC3r17IyYmBrVr18adO3ewbt06DB06NNvXTaR3/30RUgrWiI51gFReChYbS2Pl6/PYHxONXb/XxxCvcUiSO0EwVUEpyGDxhy9Qt7/mTA+VIMwZzXPmDJGOSM2ALkuh3Dcb0bEOSJKXQq0jbTElbB4kADYuq4mkZuOQJC+FOKkKSsESFl1mM2NKKuaM5nlxzhm1Wg1HR0eEhIQwZ6jIqRp4Fa9evUL7lt54/iocqYKAi36l4WDnipVtVxm6ejrDEW66lG04pdP/mSEnJyd06dIFmzdvxsaNG9GlSxc4OTlplXv48CEGDhyIihUrws7ODvXq1QMA0dkbe3t7/Pzzz1i7di0qVaqEr776Ksdjx8fHY8qUKahRowbs7OygUChw584dzT6Dg4NhamqKhg0barapVq2aKFSCg4NRrlw5TTgByPHOJPkxaNAgzZBvQRCwY8cOzdmgV69e4enTpxgxYgRsbGxQtmxZ2NjYYN68eaKzSEQG8d8XIYU0El5By3EhaBsO3fQHAAwsUxX/ePkiSV4KssRXaOc4EwppJNBlKb8IlUTMGeYMkb54DYai12x4BS1HxIurWHVyKVIFAS3snJHg/bUmZ6q4zIWi12y9driQATFnSkTOKBQKTdZ8++23zBkqcpIA9OjXD89Dw+BmaooetrZY3GEF9vU/DA/nqoauns5whJuu5BpO6fR/Zujjjz/G2LFjAQCrV6/Osky3bt3g7u6On376CS4uLoiJiUGzZs2QkpIiKnf69GlIpVK8ePEC8fHxsLGxyfa4kydPxtGjR/H999/D09MTcrkcH374oWaf6cNEJem3vMpCepmMciqfHwMHDsRXX32FwMBAJCYm4unTp+jfvz8AaIae//TTT2jUqBHi4uKgUChgYmICqVSqk+MTFch/nxchtz7Hjj0XAQBdGg5FmQZDkAQgxjwCd6t8jxExr4DuK/lFqCRizjBniPTNazAcZr/E0mHTkZSiRCWX2ujeZTFSTM0RYx6B3bVXYmaNPsyYkoo5U2JypkmTJlCr1ZqsMTPjSVgqOlJTU9G/f39cunQJDg4OWG9ji1Kmpqjq1hQmJiWri4oj3HThv8u9cg+ndIJe573o1KkTUlJSkJKSgo4dO2qtf/36NYKDg/HNN9+gbdu2qF69utYtqAHg/PnzWLx4MQ4cOAAbGxuMGzcux+OeOXMGw4YNQ69evVC7dm24uLggJCREs7569epITU3F33//rVl29+5d0bFr1KiB0NBQvHjxQrPswoULOR7X3NwcAKBSqXIsV7ZsWbRs2RLbtm3Dtm3b0K5dOzg7OwMAnJ2dUaZMGTx69Aienp6oWLEiPD094enpiQoVKuS4X6LCciHZEwMPpEAtAB3qO6NThj9yr1fchIkJj9nZVlIxZwAwZ4j0LSoqCl1mbEZMjBIKZwU+6zQXZqZpv/8nqmxFpCIadjV6GLiWpBfMGQAlK2cyZg1zhooKQRAwduxYHDhwADKZDPv370eXx49Q/U4wTCwtDV09nStZ3YeG8t/lXnk7IwQAEr1e7iWVShEcHKz5f2b29vZwdHTE+vXr4erqipCQEHzzzTeiMrGxsRg8eDDGjRuHzp07o1y5cmjYsCG6du2KPn36ZHlcT09P7N27F926dYNEIsGMGTM0Z1oAoGrVqujUqRNGjhyJ9evXw9TUFBMmTIBcLteUadeuHapWrYohQ4ZgyZIliImJwfTp03N8veXLl4dEIsHBgwfh4+MDuVwOhUKRZdlBgwZh9uzZSElJwbJly0TrZs+ejfHjx8Pa2hotWrSAmZkZAgMD8ebNG/j6+uZYByJ9u3v3Lrp164akZCU6etdD/zq+iMfbs6Vd7n8Eh34fAV4DDVhL0hvmDADmDJE+JSUloWfPnrh16xYcSjngk85zYWlhrVn//v3B2F9zZa53jqNiijkDoOTkjI2NDTp27IjXr1/jzp07iI6OZs5QkTB//nysW7cOEokE27dvR/PmzQ1dJb3iCDdd8RqcNrIEuQ0XlhTKCBQbG5tsh0ubmJhg586duHr1KmrVqoWJEydizpw5ojJffPEFrKysMH/+fABAzZo1sWjRIowaNQrPnz/Pcr/Lli2Dvb09mjVrhm7duqFjx47w8vISldm4cSPc3d3RqlUr9O7dG59++ilKl357BxITExPs27cPycnJaNy4MT755BN8++23Ob7WMmXKwM/PD1999RWcnZ01w8+z0qdPH7x+/RoJCQno2bOnaN0nn3yCDRs2YPPmzWjevDnatGmDTZs28YwQFQkWFhZwcnKCV/0G+LDF94hHGdhIX6K3w1ewkb5EjMoF/n96IDYyydBVJX1hzjBniPRIIpHAzc0NNtY2+LLHUpSzqIZoiwjsq7UM0RYRsE12Qt+7kyBPyv5yPCrmmDMlJmc2bdqEunXromvXrtiyZQtzhooMZ2dnSKVSrFixAr169Xq7ooTe9VoiZHWReQkWExMDW1tbREREwNHRUbQuKSkJjx8/RoUKFSCTZX172VzlOPdB4YRTfqnVasTExMDGxgYmJuyDZXtovxeUSiUOHToEHx8fzgEBGKw9njx8joOrgoBEOWycZOjZIQTWp75AbKsf4P+nB2IikmDjJEOviV7Z3iJbH16/fg0nJydER0fnOC+KMdFr1jBnij22B3MmN4Zqj5jXCfjxm/2wEkrDzE7ALxVnI94iClbJdugdPAFWifYGyRmAWZMZc0aMn6va2Cbi94JUKmXOZGKorLlz5w6qVav2dkHg1rRL1LssNehniz5yhpeU6lr6L4hWSBXNcCKiokupVOLSpUto0aIFAMDFzRnOLs5IjE1BT18vWDs0A5r3gbXUDD1rJcF/aSDk1uYwk/GjvURjzhCRDp06dQrvvfceTExMYGFpjooenpqc6Wd1EFdeXoHcVI6aveviwPIbzBljwJwhIh26efMmXFxcUKpUKQDQ7mxL/6zR881YDIFpqQ9aIcVwIqL8EQQBn376KbZs2YIff/wRI0eOhIXcFN3G14MyKfXtyIL/5k6xdkgbcWAmM4WFnB/tJR5zhoh0YNeuXejfvz/69OmD7du3Z5EzMrQr305TnjljRJgzRKQDjx49Qrt27WBra4tjx46hfPnyb1dqjaYteZ1uTEt9Sf8FKQJDI4mo+Jk5cyY2bdoEqVQKNzc3zXILefZfdAr78h4yMOYMERXAyZMnMWTIEACAi4uLZmJ65gxpMGeIqAAiIiLQqVMnhIeHw9XVFfb29m9XZnvpesnqdGOHmz55DQbq9tfb3XuIqGRat24d5s2bp/l/ly5dDFwjKrKYM0T0Dv755x/07NkTKSkp6N27N5YtWwaJJLeJ8skoMWeI6B0kJCSgW7duuH//PsqVK4dDhw69nRctx3kigZLU6WacsycWJoYTEeXD/v37MXr0aADArFmzMGLECAPXiIo85gwR5cOzZ8/QqVMnREdHo0WLFvj11181o9uIssScIaJ8SE1NxYABA3Dx4kXY29vjyJEjb6/Y+a+zLVktR5zKMcvt41SOSFbL0zrdArcWYs11jx1uRERFxMWLF9G/f3+o1WqMGDECs2bNMnSViIioBImKikLnzp3x/PlzVK9eHb///jvkcrmhq0VERCWEIAgYN24c9u/fDwsLC+zfvx/Vq1dPW6lSAn/4Ilktx4E3M7Evch5iM3W6xaocsS9yHg68mZnW6faHb9p2xRQ73IiIioiDBw8iMTERPj4++PHHH3l5DxER6dS1a9dw//59uLq64vDhw3BwcDB0lYiIqAR58+YNjh8/DolEgu3bt6NFixZvV0rNgC5LoRQskai2RYzKBf4ZOt1iVY7wj5yHGJULEtW2UAqWafNHFuNRtpzDjYioiJg7dy48PT3x4YcfwtSUH89ERKRbbdq0wfHjx6FQKMR3iiMiItIBBwcHnD9/HgEBAejdu7d2Aa/BUADouW8G/CPnajrd2tkux/HoCYhRucBG+hI9HWZA0Wt2sZ/Djd/oiIgMKD4+HmZmZjA3N4dEIsGwYcMMXSUiIiphoqKiYGdnBwDi0QZEREQ6kDFnSpUqhb59+2Zf2GswrCHudNsbuRAANJ1t1iWgsw3gJaWUiUQigb+/v6GrkSezZ89GvXr1DF0NonemVCrRp08fdO3aFbGxsYauDlGhYM4QFa7ly5ejZs2auHHjhqGrQlQomDNEhSswMBAVKlTAL7/8kveNvAbDutdstLP9QbS4ne0PJaazDWCHW4kxbNgw9OzZ09DVKPY2bdqUp/lMNm3aBIlEAolEAqlUCnt7ezRp0gRz5sxBdHS0qOywYcM0ZTM+OnXqlOtxAgIC4OPjA0dHR1haWqJGjRqYOHEinj9/rikjCALWr1+PJk2aQKFQwM7ODg0bNsTy5cuRkJAAIC3MM9bX3d0dn3zyCV69epXPFiJdEQQBn332GQ4fPoxz587h/v37hq4SUY6YM7rBnKHC9Ntvv8HX1xcvXrzAsWPHDF0dohwxZ3Rj06ZNmpFGuZVjzlBBPX78GD4+PoiKisL27duhVqvzvG2sRx8cV84VLTuunItYjz66rqbBsMNNj6KTo3HsyTGce34OqepUQ1eHdMzGxgZhYWF49uwZzp8/j08//RRbtmxBvXr18OLFC1HZTp06ISwsTPTYsWNHjvtft24d2rVrBxcXF+zZswe3b9/Gjz/+iOjoaCxZskRTbvDgwZgwYQJ69OiBgIAABAUFYcaMGfj999/x559/asrVrFkTYWFhCA0Nxdq1a3HgwAEMGTJEt41CeTZ79mxs3LgRJiYm2LVrF7y8vAxdJSqGmDMlG3OGCuL06dP46KOPIAgCxowZA19fX0NXiYoh5kzJxpwxLsmpKvxxIwz7rj3DvmvPcPBG2s/44I0XmmV/3AhDcqoqT/t7/fo1OnfujH///Rd16tTBnj17YGKSty6m2Mgk+C8NREysKWysU9HbaQZsrFMRE2sK/6WBiI1MeufXWaQIRiY6OloAIERERGitS0xMFG7fvi0kJiYW+DhHHx8Vam2qpXl0+K2DEBIdUuD9Zmfo0KFCjx49NM9btWoljBs3Tpg8ebJgb28vODs7C7NmzRJtc+/ePeG9994TLCwshKpVqwpHjhwRAAj79u3TlHn27JnQt29fwc7OTnBwcBC6d+8uPH78WOu4s2fPFkqVKiVYW1sLn376qZCcnKwpo1arhUWLFgkVKlQQZDKZUKdOHeG3337TrA8ICBAACMePHxcaNGggyOVywdvbW7hz546ovgsWLBBKly4tKBQK4eOPPxamTp0q1K1bV1Tml19+EapVq6Z5TatXr9ase/z4sQBA2LNnj9C6dWtBLpcLderUEc6fPy+qR8ZH5jZLt3HjRsHW1lZr+b///is4OTkJgwYNyvZnkxdPnz4VzM3NhQkTJmS5/s2bN4IgCMKuXbsEAIK/v79WGbVaLURFRQmCIAizZs3Saqt58+YJJiYmQkJCgta2md8LKSkpgr+/v5CSkpKv11FSFbQ91q1bp/kdW79+vY5rV/giIiIEAEJ0dLShq1JkFEbWMGeYM9n9bPKCOVO0FbQ9/vnnH8HOzk4AIPTq1UtITU3VcQ0LH7NGjDmThjmTe87MnDkzy/Yu6TkjCOL3AnNGEM4/iBDKTz2oeVT5+oDg7+8vVPn6gGj5+QfanyuZJSQkCN7e3gIAwd3dXXj+/Hme6xEbmShsmX5OWPXZCWHL9HNCzOtEQUhNEWJei5fHRha8XyY/9JEzHOGmB/HKeHx15ivRshfxL/DtxW8LtR6bN2+GlZUVLl26hMWLF2POnDmaywnUajV69+4NqVSK8+fPY8mSJZg2bZpo+4SEBLRp0wYKhQKnT5/G2bNnoVAo0KlTJ6SkpGjKnThxAsHBwQgICMCOHTuwb98++Pn5adZ/88032LhxI9auXYtbt27hyy+/xEcffYRTp06Jjjd9+nQsWbIEf//9N0xNTfHxxx9r1u3evRuzZs3Ct99+i7///huurq5Ys2aNaPuffvoJ06dPx7fffovg4GDMnz8fM2bMwObNm7WOM2nSJAQFBaFKlSoYMGAAUlNT0axZMyxfvhw2Nja4c+cOnj9/jkmTJuWrzUuXLo1BgwZh//79UKnydmYgK7/99htSUlIwZcqULNenDxPftm0bqlatih49emiVkUgksLW1zfYYcrkcarUaqak8W1mYDh48iM8//xwAMGPGDIwcOdLANaLiiDnDnGHOUHaeP3+Ozp07IyoqCs2bN8e2bdsglUoNXS0qZpgzxT9nnj9/jjt37mDixIn5anPmTMnVyMMe7vZySLJZL5EA5Rws0cjDPsf9qFQqDBw4EBcuXICdnR2OHDkCNze3PNfDTGYKubU5bJxk6OnrBWsHGSA1g7VD2nMbJxnk1uYwk5WAe3zqrOuumCiMs0G/P/hdcyYoKDxIWBu0VvP83/h/C7Tv7GR1RqhFixaiMo0aNRKmTp0qCIIgHD16VJBKpcLTp08FlUolvHnzRvjjjz9EZ4R+/vlnoWrVqoJardbsIzk5WZDL5cLRo0c1x3VwcBDi4+M1ZdauXSsoFApBpVIJcXFxgkwm05x1STdixAhhwIABgiCIzwilS69L+s/C29tbGDVqlGgfTZo0EZ3lcHd3F7Zv3y4qM3fuXMHb21sQhLdnhDZs2KBZf+vWLQGAEBwcLAjC2zM9b968EVQqVVZNLSqXlbVr1woAhH///VfTRlKpVLCyshI95syZk+3+P//8c8HGxibb9emqV68udO/ePddymc8IBQcHC56enkLjxo2zLM+RBzl71/ZISEgQXFxcBADC8OHDRe+t4oyjDrTpO2uYM8wZ5kzJVpD2GDRokABAqFq1apafQcUVs0aMOcOcSZddzqS3SXZZU9JzRhA4wi0re64+zXGE297Ap7nuY/v27QIAwdzcXDh16tQ71SMpQZntCLbYyEQhKUH5TvstCH3kTAnoMix6QqJDAAAWUgvUcaoDQRCwOmg1AOBp7FOUtixdKPWoU6eO6LmrqyvCw8MBAMHBwShXrhzKli2rmdjQ29tbVP7q1at48OABrK2tRcuTkpLw8OFDzfO6devC0tJS89zb2xtxcXF4+vQpwsPDkZSUhPbt24v2kZKSgvr162dbX1dXVwBAeHg4ypUrh+DgYIwaNUpU3tvbGwEBAQCAV69e4enTpxgxYoRoxFBqaqrWWZHsjlOtWjXogiAIANLOyKRr06YN1q5dKyqXPmn2qFGj8Ouvv2qWx8XFQRAE0fY5HSsv5QDg5s2bUCgUUKlUSE5ORuvWrbF+/fo8bWvsklNVOH47HCnpZ/nUKpggbb4DmKSNGjCXStGuRmlYmGY/ikAul+Pw4cP44YcfsG7dujz/7IgyY84wZwDmDGVtzZo1EAQB3377LRwdHQ1dHSqmmDPMGYA5UxJ1r+uGZcfu4dmbRNFyiQRwt7dEtzq5j1Tr378/7t+/j+rVq6Nly5bvVA8LuSks5Fl3RynsZe+0z6KIHW56EJ2cdlcXS1NLSCQSWJu//YCPSo4qtHqYmZmJnkskEk0YpX+IZl6fkVqtRoMGDbBt2zatsqVKlcr1+BmP98cff6BMmTKi9RYWFtnWN70ueb3LSXq5n376CU2aNBGty3wZRUGOkxfBwcGwsbER/ZFrZWUFT0/PLMvPmTNH65KiKlWqIDo6GmFhYZoQzUqVKlUQHBycp3pVrVoV+/fvh1QqhZubm1b7U/YCn0RhzPZAzXMLqYDFjYGv9t5Esurt+2bHyKbwrpTzl5t69eph48aNeqsrGQfmjPbxmDPMGUpjY2OT5XuKKD+YM9rHY84wZ0oCU6kJvmxfBb67r4uWCwLwZfvKMJXmPuuYRCLBzJkz9VXFEoVzuOmBrUXaGYiE1AQIgoDYlFjNOjsLOwPVSqxGjRoIDQ0V3X3mwoULojJeXl64f/8+SpcuDU9PT9Ej41mW69evIzHxbQ/5xYsXoVAoULZsWdSoUQMWFhYIDQ3V2oe7u3ue61u9enVcvHhRtCzjc2dnZ5QpUwaPHj3SOk6FChXyfBxzc/MCzVUQHh6O7du3o2fPnnm+Q0vm9gWADz/8EObm5li8eHGW20RFRQEABg4ciHv37uH333/XKiMIguiW3ubm5pr2YDjlT0HmO4iPj0fnzp1x7tw5/VaSjApzhjnDnKGMvvrqK3z33XdZdkAQvQvmDHOGOVNyda/rJvpuk/5dJqfRbceOHUPv3r0RHx9fOJUsITjCTQ/K25QHACSrknEj4gYuhr39IHW3zvuHsj61a9cOVatWxZAhQ/Ddd98hLCxMq5d60KBB+O6779CjRw/MmTMHZcuWRWhoKPbu3YvJkyejbNmyANKGU48YMQLffPMNnjx5glmzZmHs2LEwMTGBtbU1Jk2ahC+//BJqtRotWrRATEwMzp8/D4VCgaFDh+apvl988QWGDh2Khg0bokWLFti2bRtu3bqFihUrasrMnj0b48ePh42NDTp37ozk5GT8/fffePPmDXx9ffN0HA8PD8TFxeHUqVPw9vaGQqEQDS/PSBAEvHz5EoIgICoqChcuXMD8+fNha2uLhQsXisomJyfj5cuXomWmpqZwcnLKct/u7u5YtmwZxo4di5iYGAwZMgQeHh549uwZtmzZAoVCgSVLlqBv377Yt28fBgwYgBkzZqB9+/YoVaoUbt68iWXLlmHcuHHo2bNnnl47ZS+7M0HpsjsjlJqair59++LIkSP4559/8ODBA/5xQDrBnGHOMGdKloJMXbBixQosWrQIANCqVSs0bty4MKtOJRRzpvjnzIkTJ1CxYkWYmppCoVBkWZY5Y5zSv9tM23cBgDlM5A8wvl2vbEe3Xbt2Db1790ZcXBy+++47zJ49u1DrW5yxw00P2pVvB78LflCqlfjo0Eea5U1dmxbafAe5MTExwb59+zBixAg0bdoU5cqVw4oVK+Dj46MpY2lpidOnT2Pq1Kno3bs3YmNjUaZMGbRt2xY2Njaacm3btkXlypXRsmVLJCcno3///qI34dy5c1G6dGksWLAAjx49gp2dHby8vPD111/nub79+vXDw4cPMXXqVCQlJeGDDz7A559/jqNHj2rKfPLJJ7C0tMR3332HKVOmwMrKCrVr18aECRPyfJxmzZrhs88+w8cff4zIyEjMmjUr2w+UmJgYuLq6QiKRwMbGBlWrVsXQoUPxxRdfiNoHAI4cOaI1lLpq1aq4c+dOtnUZPXo0qlSpgu+//x69evVCYmIiPDw80LVrV03gSiQSbN++HevXr8cvv/yCefPmwdTUFJUrV8aQIUPQsWPHPL92yll+5zsQBAGjRo3CoUOHIJfL8dtvv7GzjXSGOcOcYc6ULO86dcGePXs0v3/z589nZxvpDHOmeOfMqFGjMGDAALx+/RozZ84U3XE1I+aM8ZLb3YK15yIAM2BZ9lesf3QIjTx/0nS2pwsJCYGPjw/i4uLQpk0brTsBU84kgpGNPY+JiYGtrS0iIiK0JpJNSkrC48ePUaFCBchkBZuo78+QPzHx1NtbMLtZuWF9h/Vav8BFgVqtRkxMDGxsbPI8bDjdsGHDEBUVBX9/f/1UzgAK0h4lReb3glKpxKFDh+Dj46M1l4Yx2Rv4DL67r//3RUiFKZelSFZJsKxfXfSqX1ZU1s/PD7Nnz4aJiQn8/f3RrVs3A9Va/16/fg0nJydER0dr/XFmrAoja5gzxRdzhjmTWapKjTbfn8SzN4kQAK2cST+589fEVpoRCGfOnEH79u2RnJyMzz//HKtXry7RN+Nh1ogxZ8SYM9qYNeL3glQqNeqcySheGY/3dr4HE7UJZtjNwNyouUhGMrxdvbG+w9ubUERGRqJ58+a4c+cOateujTNnzmjdwKMk0UfOcISbnnTw6ICzrmdx5eUVyE3laOLaBKYmbG6i4ix9lNurmAQA2c93sGHDBs1Z0TVr1pTozjYyHOYMUcmR36kLbt++jR49eiA5ORk9e/bEypUrS3RnGxkGc4aoZDoRegJKtRIWSLv6ZmSdkVh1YxUuhF1AeEI4SluWRmJiIrp37447d+6gbNmyOHToUInubNMX4+zqLiS2FrZoV74dmpdpznAiKgHSvxAJJmmXlabNd1BRNN/B6dOnNbd8/+abb/DZZ58ZpK5kHJgzRCVH5kms02U+uRMbG4vOnTvjzZs38Pb2xvbt27XuYEikK8wZopInJDoEAGAuNQcANHJupFn3NPYpAOCzzz7DuXPnYGdnhyNHjmjmO6T84acmFcimTZsMXQWiQpXbfAeNGzdGr169oFAoMGfOHMNWlqgEYM6QschulFvm0W3W1taYOnUqVq9ejQMHDkAulxuiukQlBnOGjE10ctqdXy1N027apDB/e1ONqOQoAMD48eNx6tQpbNmyBTVr1iz0OpYUHOFGRJRH8cp4TD8nnig0LD4M3178VvNcJpNh586dWL9+PS/vISKifMk8yi27qQtGjx6NwMBArbm7iIiIcmNrkXZpaEJq2jQ5cSlxmnV2FnYAgIYNG+LevXto1apVodevJGGHGxFRHqXPd5BuZKmmAIAzd89g5ryZUKvVAACpVGr0k7ESEVH+pY9y62lyFgDQU3IWX7avDBMJsGDBAkRGRmrK8s7XRET0LtKvzElRpQAArvx7BQAQfTkar+6+0pRjzhQcO9yIiPIoJDoEFikCNi1NBQA0CtwPdbIaT5Y/wdwZc/HVV18ZuIZERFTc9RD+whyzjQCAOWYb0V39F77++mt8/fXXaN26NZRKZS57ICIiyl678u1gZvJ2cMBPN35C3K04PF//HB90/gD//POPAWtXsrDDjYgoj6KfXxE9t1Cm4umap0h8lAgbexuMGDHCQDUjIqISIXArpAfGixatnfEZFi1aBACYOHEiR1ATEVGBWJlZYVEjP80gAuXDeDxd9RTqVDW6deuGGjVqGLiGJYfBO9zWrFmDChUqQCaToUGDBjhz5kyO5bdt24a6devC0tISrq6uGD58OF6/fl1ItSUiY6U+/zPs756CxX8DCwRBwIyDMYi9HguJmQTLvx2AqlWrGraSlC1mDREVeYFbgf3jAAiaOdz8bydh/OEkAMC80R9i6NChBqse5Yw5Q0TFSdtybQEA4eHheP7jK6gSVWjdujU2b94MExODdxOVGAZtyV27dmHChAmYPn06rl27hvfeew+dO3dGaGholuXPnj2LIUOGYMSIEbh16xZ+++03XLlyBZ988kkh15yIjErgVtz9+Hu03WCFDStUAIDdu3djX2A8TACsKOWGLhG7074sUZHDrCGiIi9DZ1u64OBgDN0TAwHAqAZm+NrpT+ZMEcWcIaLiQp2QkPZITER0airmzJmD169eo1aNGtizbRvnbdMxg3a4LV26FCNGjMAnn3yC6tWrY/ny5XB3d8fatWuzLH/x4kV4eHhg/PjxqFChAlq0aIHPPvsMf//9dyHXnLLSunVrTJgwoUD72LRpE+zs7HRSHyKd0HwJemtf5Gvs2LEDAPBNaWe0tbZGaZUqrRy/DBU5zJqSgzlDJVKmzjZ1qgSHt9nj22+/RVIq0K2qKVb5yCCRgDlTRDFnSg7mDJV0d70a4K5XA/zj3Qzjn4Tg2bNnKG1mhh+SkvHv+20NXb0Sx9RQB05JScHVq1e1Jhnv0KEDzp8/n+U2zZo1w/Tp03Ho0CF07twZ4eHh+N///ocuXbpke5zk5GQkJydrnsfExAAAlEql1qSzSqUSgiBArVZr7jZYnLx8+RILFy7EoUOH8OzZM9ja2qJy5coYOHAghgwZAktLyyy3EwRB829BX3dB95G+bU77kEqlmv9bWlrCzc0NzZo1w9ixY9GgQQPNupMnT6Jt26w/NJ4/fw4XF5cs1wmCgH///RfffPMNDh06hOfPn6N06dKoW7cuvvjiC9E+r127hgULFuDMmTOIjo5GuXLl0LJlS0yaNAlVqlRBSEgIKlWqpClvZ2eH2rVrw8/Pr0jfYlmtVkMQBCiVSkilUs17xegmalYpgcNfAyYWqNj3DQDgtJkc5pssYGpqCve2pXGzRykMefUKSsjStjn8NVDjA0BqPHPsFOXfC2aNbjFnmDO6wpz5T4acSRdgJQfMzSGXyyEpDSjHV8CTqNdwT02ba8cYcwYour8bzBndYs7oJmcAICwsDH5+fjh+/LhR5gwgzpr0n0dR/SwpLKr/RrAp1WrYmJnB0tISaypURGmpFCoYd/vo47UbrMMtIiICKpUKzs7OouXOzs54+fJllts0a9YM27ZtQ79+/ZCUlITU1FR0794dK1euzPY4CxYsgJ+fn9bygIAArQ9sU1NTuLi4IC4uDikpKe/wqgwnJCQEnTp1gq2tLaZPn44aNWogNTUVDx8+xK+//go7Ozv4+PjkuI/Y2NgC1SE1NRUpKSmaPwDeRVJSEgRByHUfq1evRtu2bZGcnIwHDx5g8+bN8Pb2xqpVq9C/f38AQEJCAgDgypUrsLa2Fm0vk8myPUZoaKimLWfNmoWaNWtCqVTir7/+wpgxY3D58mUAwJEjRzB06FC8//77WLduHSpUqIBXr17h999/x9dff41ffvkFcXFxAAB/f39Uq1YNr169wty5c9G1a1ecP38e5cuXf+e20qeUlBQkJibi9OnTSE3/Ax/AsWPHDFgrA6m1SmtRtaXAkidPUK5cOUgkEvzjBoju5XPUuNop/b1WFDFrdIc5w5zRJeZMBlnkDOYBiyIjYWpqChsbG9wsA9zMuN7IcgYoulnDnNEd5gxzRteyyhqjzJmM5rz9HBmnUuHFixcwdXfHw/+WPTx0yDD1KgL0kjOCgTx//lwAIJw/f160fN68eULVqlWz3ObWrVuCq6ursHjxYuH69evCkSNHhNq1awsff/xxtsdJSkoSoqOjNY+nT58KAISwsDAhJSVF9IiJiRFu3bolxMfHCyqVqkAPZWyscLtqNeF21WqCMja2wPvL7dGhQwehbNmyQkxMTJbrU1NTBZVKJTx8+FAAIFy9elWzLiIiQgAgHD9+XFCpVMKJEycEAMKhQ4eEevXqCTKZTGjTpo0QFhYmHDx4UKhWrZpgbW0t9OvXT4jN8NpatWoljB49Whg9erRga2srODg4CF9//bXm2CqVSkhMTBQmTZokuLm5CZaWlkLjxo2FEydOaNb//PPPgq2tbY6vFYCwZ88ereWDBw8WrK2thYiICNHreP36db7aslOnToKbm5sQHR2ttS59X7GxsYKTk5PQo0ePLPeRXi6r9g4NDRUACGvWrNH778W7PuLj44Vbt24JMTExQkpKihAfHy/4+/sL8fHxWu8bo3hc2SKk+DkLKX6lhaQZzsL1Ro0Ff39/IW5WeSHFr/R/D+e0coauqwEeYWFhAgAhOjpapzmhCyU5a5gzzBnmTAl6XNki3KxTV/NIz5nrjRprlj2f52q0OVOUs4Y5w5wpqjlTpkwZ4dmzZ6K6G1POZM4ao8+ZTI+k6Oi332levzZ4fYrCQx85Y7ARbk5OTpBKpVpnfsLDw7XOEKVbsGABmjdvjsmTJwMA6tSpAysrK7z33nuYN28eXF1dtbaxsLDIcuI/MzMzrduqq1QqSCQSmJiYFPzOHBm218n+cvD69WscO3YM8+fP1zrzoV0tE606SSQSzb8Zl8+ZMwerVq2CpaUl+vbti/79+8PCwgLbt29HXFwcevXqhdWrV2Pq1Kma/W/ZsgUjRozApUuX8Pfff+PTTz+Fh4cHRo4cCQAYMWIEQkJCsHPnTri5uWHfvn3w8fHBzZs3UblyZVH9cnsdmcv4+vpi69atOHHiBPr27Zvla81NZGQkjh49im+++QYKhUJrOwcHBwBpZ0YiIiIwderULPedXi6rOigUCgBpv29F9Q4wJiYmkEgkWu+TrN43RqHh4LQZL/ePA0wEVO8agscAzE0SYaZOAiABuq8EvAYbuKKGUZR/J0p01jBnmDNgzpQYDQdDmjxfa7FJcjKk/11GGNZyLNwaGmfOAEU3a5gzusGcSaPLnJk3bx6srKw0bZLOWHIGEGdN+mW8RpszmaiVSpj8ly+mpqZsE+gnZwz27jA3N0eDBg20hnQeO3YMzZo1y3KbhIQErTd0+htHEISsNil0Ge/6oVmWmKhZrg8PHjyAIAioWrWqaLmTkxMUCgUUCoUoRPJq3rx5aN68OerXr48RI0bg1KlTWLt2LerXr4/33nsPH374IQICAkTbuLu7Y9myZahatSoGDRqEcePGYdmyZQCAhw8fYseOHfjtt9/w3nvvoVKlSpg0aRJatGiBjRs3vnsD/KdatWoA0oajZ1S2bFlNOygUCq12yii9LatUqZLjse7fvy86Zl7Fx8dj2rRpkEqlRX7OA8rEa3BapxokmVYYd2dbUVcSs4Y5w5zJCXOm+KoaeBX7vq2PT8a/ndvJrEcEBk+UYvBEKaLKNzFg7Sg7zBndYM6k0WXO5FQGYM4YOxNLS1T9+0ra/+VyA9em5DLYCDcgrQd/8ODBaNiwIby9vbF+/XqEhoZi1KhRAIBp06bh+fPn2LJlCwCgW7duGDlyJNauXYuOHTsiLCwMEyZMQOPGjeHm5mbIl6Jx16uB1rL7zVto/l/9TrDejp1+Zifd5cuXoVarMWjQINEkq3lVp04dzf+dnZ1haWmJihUripalX/+frmnTpqJ6eHt7Y8mSJVCpVAgMDMzyS0ZycjIcHR3zXb/M0v9AydwOZ86cEZ0pMzU11Szv3LmzZvm6devg6emZ5T6yO1ZeNWvWDCYmJkhISICrqys2bdqE2rVr52sfVASkd6odnPzfAna2FQclLWuYM8yZrDBnij8TS0soKjRFcuQNzbJ4cxMkm6f9rthZ2BmoZpQb5ozuMGeYM0QliUE73Pr164fXr19jzpw5CAsLQ61atXDo0CHNxIthYWEIDQ3VlB82bBhiY2OxatUqTJw4EXZ2dnj//fexaNEiQ72EIsHT0xMSiQR37twRLU8PE3mGHuv0s2kZP2CzuxtHxiGV6UNxM5JIJPm6g49arYZUKsXVq1dFd+YB3g5LLojg4LTwr1Chgmh5hQoVsrw1d8OGDREUFKR57uzsDKVSCYlEgrt37+Z4rPSQvXPnDry9vXOt265du1CjRg3Y2dnpJIzJgLwGA2oAzwH4fMfOtmKAWVNwzJk0zBnSt/I24snHr9bqCkSfAQC4W7sbokqUB8yZgmPOpNFlzty5cwfvv/9+tsdizhDpn0E73ABg9OjRGD16dJbrNm3apLVs3LhxGDdunJ5r9e6qBl4FkDbsOv1MUOVzZ/U6TNPR0RHt27fHqlWrMG7cOFhZWWVbtlSpUgDSgr9+/foAIPqQLqiLFy9qPa9cuTKkUinq168PlUqF8PBwvPfeezo7Zrrly5fDxsYG7dq1y1N5uVyuOQOUUYcOHfDzzz9j8uTJWnNIREVFwc7ODh06dICTkxMWL16Mffv2ae0jvVw6d3d30e20qZir2x94fijtXyoWSlLWMGeYMxnLpWPOlAztyreDn8wPw3xNMAPAuui0kTdNXZuitGVpw1aOcsScKRjmTBpd5UzHjh2xZs0aDB06FDY2NqJ1zBmiwlN0ZzgspkwsLdMeGc/CyOWa5fqyZs0apKamomHDhti1axeCg4Nx9+5d/Prrr7hz547mDIxcLkfTpk2xcOFC3L59G6dPn8bMmTN1Vo+nT5/C19cXd+/exY4dO7By5Up88cUXANLOogwaNAhDhgzB3r178fjxY1y5cgWLFi3CoXzefjgqKgovX77EkydPcOzYMXz44YfYvn071q5dq3X2Jzw8HC9fvhQ9sjsLBqTdolulUqFp06bYs2cP7t+/j+DgYKxYsUJz9sfKygobNmzAH3/8ge7du+P48eMICQnB33//jSlTpmguISAi0jXmDHOGOVNyWZlZYeF7C0XL3Kzc8E3TbwxUIzJGzJninzNr1qyBSqVCu3btmDNEBmTwEW6kG5UqVcK1a9cwf/58TJs2Dc+ePYOFhQVq1KiBSZMmic64/fLLL/j444/RsGFDVK1aFQsXLkSnTp10Uo8hQ4YgMTERjRs3hlQqxbhx4/Dpp59q1m/cuBHz5s3DxIkT8fz5czg6OsLb2xs+Pj75Os7w4cMBADKZDGXKlEGLFi1w+fJleHl5aZXNasLQCxcuoGnTplnuu0KFCjh58iRWrlyJiRMnIiwsDKVKlUKDBg2wdu1aTbkePXrg/PnzWLBgAQYOHIiYmBi4u7vj/fffx7x58/L1eoiIijrmDHOGCkcHjw5oUKoBzh4/i2VtlsG7rDdMTfgnO5V8zBnd5szff/+N2bNnY/LkycwZIgORCEXhVjiFKCYmBra2toiIiNC69jwpKQmPHz9GhQoVIJPJDFTDwqdWqxETEwMbG5sifVvnwsL20H4vKJVKHDp0CD4+PkZ7y+jkxFQok1KhsNduj7g3STCTmcJCbpxfiF6/fg0nJydER0drXbZgrJg1YvxcFWN7MGdyw/bQxqwRY86I8XNVG9tE/F6QSqX8XM1E+fdWHHpuA58yMTBryHmp9ZEzxvnOIyLKh+TEVBxYEYR9SwIRG5kkWhcbmYR9SwJxYEUQkhNTDVRDIiIiIiKiPArcChyanPb/Q5PTnpPOscONiCgXyqRUJMamICYiCf5LAxH3Ju229HFvkuG/NBAxEUlIjE2BMokdbkREREREVIQFbgX2jwOQfrGjkPacnW46xw43IqJcKOxl6OnrBRsnGWIiknBwVRAA4OCqIMREJMHGKW29wt44LtsgIiIiIqJiSKuzLR073fSBHW5ERHlg7fC20y02Mm2EW2xksqazzdqBnW1ERERERFREBW5Fsv9UxKkcslwdp3JAsv9UdrrpEDvciIjyyNpBhnbDa4qWtRtek51tRERERERUdP3X2XbgzQzsi5yHWJX4ZiuxKkfsi5yHA29msNNNh9jhRkSUR7GRSTi+8ZZo2fGNt7RupEBERERERFQkqJTAH75QCjIkqm0Ro3KBf+Q8zUi3OJUD/CPnIUblgkS1LZSCDPjDN207KhB2uBER5UFsZJLmBgnWDhYAAGsHC82NFNjpRkRERERERY7UDOiyFAppJHo6fAMb6UvEqFxw8M0MAMDBNzMQo3KBjfQlejp8A4U0EuiyNG07KhB2uBER5SLuzdvONhsnGbqOrQcA6Dq2nuZGCml3L2WnGxERERERFTFeg4HuK2GdodMtVu0MAIhVO2s626ylkUD3lWnlqcDY4UZElAszmSnk1uYZ7kaaNsJNYW+huZGC3NocZjJTA9eUiIiIiIgoCxk63drZLhetame7nJ1tesAONyqwYcOGoWfPnrmWk0gk8Pf319lxPTw8sHz5cp3tjyg7FnJTdBtfD70mat+N1NpBhl4TvdBtfD1YyNnhRqQPzBkiItIn5gwZDa/BiH1/DY5HTxAtPh49AbHvr2Fnm46xw62EGDZsGCQSidajU6dOej/2Dz/8gE2bNuVaLiwsDJ07d9Z7fdLNnj1b0w6mpqZwcnJCy5YtsXz5ciQnJ4vKtm7dWlNWKpXC3t4eUqkUo0aNyvU4e/bsQevWrWFrawuFQoE6depgzpw5iIyM1JRJSUnB4sWLUbduXVhaWsLJyQnNmzfHxo0boVSmTUaZ8WdoZmaGihUrYtKkSYiPj9dtw9A7sZCbQmGf9d1IFfYydrZRicec0cacISLSHeaMtnfNmYxZ8/nnn+d6HOaM8YiNTIL/nx6IUbnA2uRfAIC1yb9pN1L404PzUusYO9x0LDkxNdt5nOLeJCE5MVVvx+7UqRPCwsJEjx07dujteOlsbW1hZ2eX7fqUlBQAgIuLCywsLPRen4xq1qyJsLAwhIaGIiAgAH369MGCBQvQrFkzxMbGisqOHDkSYWFheP78Oe7cuYPnz59j8eLFOe5/+vTp6NevHxo1aoTDhw/jn3/+wZIlS3D9+nVs3Zp2K+WUlBR07NgRCxcuxKefforz58/j8uXLGDNmDFauXIlbt97e9TL9Z/jo0SPMmzcPa9aswaRJk3TfMERUbDFntDFnmDNEpDvMGW3FLWcyZs2iRYty3D9zxnhozUv9Qdryrh+A81LrCTvcdCg5MRUHVgRh3xLtOxbGRiZh35JAHFgRpLeQsrCwgIuLi+hhb2+vWS+RSLBu3Tp07doVlpaWqF69Oi5cuIAHDx6ga9eusLa2hre3Nx4+fKjZZvbs2ahXrx7WrVsHd3d3WFpaok+fPoiKitKUyTwEu3Xr1hg7dix8fX3h5OSE9u3ba46fcQj2s2fP0L9/fzg4OMDKygoNGzbEpUuXAAAPHz5Ejx494OzsDIVCgUaNGuH48eP5bhNTU1O4uLjAzc0NtWvXxrhx43Dq1Cn8888/WuFjaWmpaTdnZ2e4uLjAxsYm231fvnwZ8+fPx5IlS/Ddd9+hWbNm8PDwQPv27bFnzx4MHToUALB8+XKcPn0aJ06cwJgxY1CvXj1UrFgRAwcOxKVLl1C5cmXNPtN/hu7u7hg4cCAGDRqk02HrRFS8MWfSMGeYM0SkH8yZNCUhZ9KzhjlD6bTmpW7WBwCgaNaH81LrCTvcdEiZlIrE2BRNz3B6SMVGvu1JToxNgTJJf2eFcjN37lwMGTIEQUFBqFatGgYOHIjPP/8cX375JS5fvgwAGDt2rGibBw8eYPfu3Thw4ACOHDmCoKAgjBkzJsfjbN68Gaampjh37hzWrVuntT4uLg6tWrXCixcvsH//fly/fh1TpkyBWq3WrPfx8cHx48dx7do1dOzYEd26dUNoaGiB26BatWro3Lkz9u7dW6D9bNu2DQqFAqNHj85yffpZsm3btqFdu3aoX7++VhkzMzNYWVllewy5XK4Zok1ExJx5iznDnCEi3WPOvMWcYc6UNJyXuvCxw02HFPYyTc9wekiFPYwWDdtMu8Nh1vNAFdTBgwehUChEj7lz54rKDB8+HH379kWVKlUwdepUhISEYMCAAWjbti2qV6+OL774AidPnhRtk5SUhM2bN6NevXpo2bIlVq5ciZ07d+Lly5fZ1sXT0xOLFy9G1apVUa1aNa3127dvx6tXr+Dv748WLVrA09MTffv2hbe3NwCgbt26+Oyzz1C7dm1UrlwZ8+bNQ8WKFbF///6CNxTSQiokJES0bM2aNVAoFLCxsUHZsmVhY2ODzZs3Z7uP+/fvo2LFijAzM8vxWPfv38+yDXJz+fJlbN++HW3bts33tkRUMjFn3mLOiMsxZ4hIF5gzbxXnnMmYNcwZyojzUhcutqaOWTukhVB6KO397ioAaMIpc0+yLrVp0wZr164VLXNwcBA9r1Onjub/zs7OAIDatWuLliUlJSEmJkYz/LhcuXIoW7aspoy3tzfUajXu3r0LFxeXLOvSsGHDHOsaFBSE+vXra9UvXXx8PPz8/HDw4EG8ePECqampSExM1MkZIQAQBAESiUS0bNCgQZg+fTrUajXi4uKgUCg0r69z5844c+YMAKB8+fK4detWlvvI67Gyk/5HRmpqKpRKJXr06IGVK1fm89URUUnGnEnDnMn5WNlhzhBRbpgzaYpzzgDQZE3FihUBMGeIDIEdbnpg7SBDu+E1NeEEAO2G19RrOAGAlZUVPD09cyyT8exF+odmVsvSh0JnJb1MTh+6OQ0rBtKGFudk8uTJOHr0KL7//nt4enpCLpfjww8/1ExYWlDBwcGoUKGCaJmtrS08PT2hVqs1AW1ikjYIdMOGDUhMTATwtr2qVKmCs2fPQqlU5nhWqEqVKggODs5TvdL/yDAzM4Obm1uuZ5uIyDgxZ5gzGTFniEjXmDPFO2cAiLIGYM4QGQIvKdWD2MgkHN94S7Ts+MZbxfYWu6GhoXjx4oXm+YULF2BiYoIqVaq88z7r1KmDoKAg0a2mMzpz5gyGDRuGXr16oXbt2nBxcdEaMv2u7ty5gyNHjuCDDz7I8zZlypSBp6cnPD09Ub58eQDAwIEDERcXhzVr1mS5TfpErAMHDtTM3ZBZamqq6DbZ6X9klC9fnuFERNlizuSOOZOGOUNE74I5kzvmTBrmDFH22OGmYxknFLVxkqH35AaiORD0GVLJycl4+fKl6BEREVHg/cpkMgwdOhTXr1/HmTNnMH78ePTt2zfb4dd5MWDAALi4uKBnz544d+4cHj16hD179uDChQsA0uZM2Lt3L4KCgnD9+nUMHDgwx7NU2UlNTcXLly/x4sUL3Lx5EytXrkSrVq1Qr149TJ48WVQ2ISFB027//vsvXr58iTdv3mS77yZNmmDKlCmYOHEipkyZggsXLuDJkyc4ceIE+vTpo5kvYcKECWjevDnatm2L1atX4/r163j06BF2796NJk2a4P79+/l+XURkvJgzecOcYc4Q0bthzuRNUc+Z9KxhzhAZDjvcdCjuTZLWhKKulWy1Jh6Ne6OfkDpy5AhcXV1FjxYtWhR4v56enujduzd8fHzQoUMH1KpVK9uzIHllbm6OP//8E6VLl4aPjw9q166NhQsXQiqVAgCWLVsGe3t7NGvWDN26dUPHjh3h5eWV7+PcunULrq6uKFeuHFq3bo3du3dj2rRpOHPmDBQKhajsTz/9BFdXV5QpUwbVqlVDmTJlMGDAgBz3v2jRImzfvh2XLl1Cx44dUbNmTfj6+qJOnTqa22hbWFjg2LFjmDJlCtatW4emTZuiUaNGWLFiBcaPH49atWrl+3URkXFizuQdc4Y5Q0T5x5zJu6KcMxmzZuDAgTnunzlDpD8SQRAEQ1eiMMXExMDW1hYRERFwdHQUrUtKSsLjx49RoUIFyGT5n58gOTEVB1YEITE2RWtC0fQzRXJr8yJ3q92s5pJJN3v2bPj7+yMoKMgwlTOAnNrDWGR+LyiVShw6dAg+Pj4cGg6wPTJ5/fo1nJycEB0drZknxNjpK2uYMyUDc4Y5kxu2hzZmjRhzRow5o41ZI34vSKVSfq5mwqwR00fOFJ1PyRLAQm6KbuPrQZmUqnWrXWsHGXpN9IKZzLRIhRMRERUfzBkiItIn5gwRke7wk1LHLOTZB1Dm0CIiIsov5gwREekTc4aISDeMc2wp5dns2bONbvg1EREVHuYMERHpE3OGiAyFHW5EREREREREREQ6xA43IiIiIiIiIiIiHWKHGxERERERERERkQ6xw42IiIiIiIiIiEiH2OFGRERERERERESkQ+xwIyIiIiIiIiIi0iF2uJHIpk2bYGdnp3k+e/Zs1KtXT/N82LBh6NmzZ6HWycPDA8uXL9frMTK/7oULF8LLy0uvxyQiMkbMmTTMGSIi/WDOpPHz88N7772n12MSUc7Y4VaCvHz5EuPGjUPFihVhYWEBd3d3dOvWDSdOnHjnfU6aNKlA2+dH5pBId+XKFXz66aeFUod0Y8eOxbFjxwr1mERERR1zRneYM0RE2pgzujNx4kT8/vvvhXpMIhIzNXQFSDdCQkLQvHlz2NnZYfHixahTpw6USiWOHj2KMWPG4M6dO++0X4VCAYVCUaC6paSkwNzc/J23L1WqVIGO/y4UCgVsbGwK/bhEREUVc0a3mDNERGLMGd1SKBRQq9WFflwieosj3EqI0aNHQyKR4PLly/jwww9RpUoV1KxZE76+vrh48aKm3NKlS1G7dm1YWVnB3d0do0ePRlxcXLb7zTwEO52fnx9Kly4NGxsbfPbZZ0hJSdGsa926NcaOHQtfX184OTmhffv2uR775MmTGD58OKKjoyGRSCCRSDB79mwA2kOwQ0ND0aNHD82Xlb59++Lff//VqvPWrVvh4eEBW1tb9O/fH7GxsXluz8yX+qQPPf/+++/h6uoKR0dHjBkzBkqlUlMmJSUFU6ZMQZkyZWBlZYUmTZrg5MmTeT4mEVFRxpxhzhAR6RNzRrc5k/mSUuYMUeFjh1sexcfHZ/tISkrKc9nExMQ8lc2PyMhIHDlyBGPGjIGVlZXW+ozDmk1MTLBixQr8888/2Lx5M/766y9MnTo1X8c7ceIEgoODERAQgB07dmDfvn3w8/MTldm8eTNMTU1x7tw5rFu3LttjT5kyBQDQrFkzLF++HDY2NggLC0NYWBgmTZqkdWxBENCzZ09ERkbi1KlTOHbsGB4+fIh+/fqJyj18+BD+/v44ePAgDh48iFOnTmHhwoX5ep2ZBQQE4OHDhwgICMDmzZuxadMmbNq0SbN++PDhOHfuHHbu3IkbN26gT58+6NSpE+7fv1+g4xKRcWDOvMWcYc4Qke4xZ95izjBniAqFYGSio6MFAEJERITWusTEROH27dtCYmKi1joA2T58fHxEZS0tLbMt26pVK1FZJyenLMvlx6VLlwQAwt69e/O1nSAIwu7duwVHR0fhzZs3gkqlEjZu3CjY2tpq1s+aNUuoW7eu5vnQoUMFBwcHIT4+XrNs7dq1gkKhEFQqlSAIgtCqVSuhXr16eT52uszHTle+fHlh2bJlgiAIwp9//ilIpVIhNDRUs/7WrVsCAOHy5cuaOltaWgoxMTGaMpMnTxaaNGmSbV0yHlulUglTp07Vet3ly5cXUlNTNcv69Okj9OvXTxAEQXjw4IEgkUiE58+fi/bbtm1bYdq0aTk3RBGU+b2QkpIi+Pv7CykpKQauWdHA9hCLiIgQAAjR0dGGrkqR8S5Zw5xJw5xhzhDbIyvMGjHmzFvMmayPPXPmTKFWrVqa12RsOSMI4vcCP1e1sU3E9JEznMOtBBAEAQAgkUhyLRsQEID58+fj9u3biImJQWpqKpKSkhAfH5/nuWTq1q0LS0tLzXNvb2/ExcXh6dOnKF++PACgYcOG+Tp2VmeyshIcHAx3d3e4u7trltWoUQN2dnYIDg5Go0aNAKQN27a2ttaUcXV1RXh4eJ6OkZ2aNWtCKpWK9nnz5k0AQGBgIARBQJUqVUTbJCcnw9HRsUDHJSIyNOYMc4aISJ+YM8wZopKIHW55lNO8ABk/tADk+EFoYiK+ijckJKRA9QKAypUrQyKRIDg4OMdbXD958gQ+Pj4YNWoU5s6dCwcHB5w9exYjRoxAampqgeuRMSAzB05Ox844b0BuBEHIMogzLzczM9OqW0EnDc1pn2q1GlKpFFevXtX6fSjoJK1EZByYM7ljzjBniOjdMWdyx5xhzhDpEjvc8iivZyz0WTY7Dg4O6NixI1avXo3x48dr7TMqKgp2dnb4+++/kZqaiiVLlmiCcvfu3fk+3vXr15GYmAi5XA4AuHjxIhQKBcqWLZvtNnk5trm5OVQqVY7HrlGjBkJDQ/H06VPNWaHbt28jOjoa1atXz/dr0ZX69etDpVIhPDxcNDkpEVFeMWfeYs5oY84QUUExZ95izmhjzhDpXoFumpB5ck0ynDVr1kClUqFx48bYs2cP7t+/j+DgYKxYsQLe3t4AgEqVKiE1NRUrV67Eo0ePsHXrVvz444/5PlZKSgpGjBiB27dv4/Dhw5g1axbGjh2rdbYro7wc28PDA3FxcThx4gQiIiKQkJCgtZ927dqhTp06GDRoEAIDA3H58mUMGTIErVq1ynLYd2GpUqUKBg0ahCFDhmDv3r14/Pgxrly5gkWLFuHQoUMGqxdRccecKTqYM8wZopKKWVM0MGeYM0QlTb473NRqNebOnYsyZcpAoVDg0aNHAIAZM2bg559/1nkFKW8qVKiAwMBAtGnTBhMnTkStWrXQvn17nDhxAmvXrgUA1KtXD0uXLsWiRYtQq1YtbNu2DQsWLMj3sdq2bYvKlSujZcuW6Nu3L7p166a55XV28nLsZs2aYdSoUejXrx9KlSqFxYsXa+1HIpHA398f9vb2aNmyJdq1a4eKFSti165d+X4durZx40YMGTIEEydORNWqVdG9e3dcunRJND8DEeWOOVM0MWeYM0QlCbOm6GHOMGeISpz83mXBz89PqFixovDrr78KcrlcePjwoSAIgrBr1y6hadOmOrmTgz69611KSzKVSqW5qw+xPQSBd4/LDdtDTNd39CnuOSMIzJrM+LkqxvZgzuSG7aGNWSPGnBHj56o2tgnvUpobtomYPu5Smu8Rblu2bMH69esxaNAg0WSKderUwZ07d3TUDUhERMaKOUNERPrGrCEiIn3Ld4fb8+fP4enpqbVcrVbn6+4sREREWWHOEBGRvjFriIhI3/Ld4VazZk2cOXNGa/lvv/2G+vXr66RSRERkvJgzRESkb8waIiLSN9P8bjBr1iwMHjwYz58/h1qtxt69e3H37l1s2bIFBw8e1EcdiYjIiDBniIhI35g1RESkb/ke4datWzfs2rULhw4dgkQiwcyZMxEcHIwDBw6gffv2+qgjEREZEeYMERHpG7OGiIj0Ld8j3ACgY8eO6Nixo67rUmQIgmDoKhAZFN8DZGglPWcAvs/IuPH3n4qCkp41fJ+RseN7gAwt3yPcSjIzMzMAQEJCgoFrQmRY6e+B9PcEEelO+t3wUlJSDFwTIsNhzmhLTkxF3JukLNfFvUlCcmJqIdeIiit+pyFKw6whQ8v3CDd7e3tIJBKt5RKJBDKZDJ6enhg2bBiGDx+ukwoWJqlUCjs7O4SHhwMALC0ts3ytJY1arUZKSgqSkpJgYsI+WGNuD0EQkJCQgPDwcNjZ2Wk6BogKU0nOGQAwNTWFpaUlXr16BTMzM6P4nDHmz9WsGHN7MGeylpyYigMrgpAYm4Kevl6QWb9tl9jIJPgvDYTc2hzdxteDhfydLlChTEpy1hjjdxpj/lzNjjG3SVZZo1arDV0tMkL5TuyZM2fi22+/RefOndG4cWMIgoArV67gyJEjGDNmDB4/fozPP/8cqampGDlypD7qrFcuLi4AoAkoYyAIAhITEyGXy0t8GOcF2wOws7PTvBeICltJzxmJRAJXV1c8fvwYT548MXR1CgU/V8XYHsyZzJRJqUiMTUFMRFrnWtfxdQAAcW+ScXDFDcREJGnKscNNN0p61hjbdxp+rmpjmzBryPDyndhnz57FvHnzMGrUKNHydevW4c8//8SePXtQp04drFixoliGU/oXodKlS0OpVBq6OoVCqVTi9OnTaNmyJYfbgu1hZmbGEQdkUCU9ZwDA3NwclStXNprLSo39czUzY28P5ow2hb0MPX294L80EDERSTi4Kgg2DYCDq4IQE5ECG6e09Qp7maGrWmKU9Kwxtu80xv65mhVjbxNmDRUF+e5wO3r0KBYtWqS1vG3btpg4cSIAwMfHB1999VXBa2dAUqnUaN6gUqkUqampkMlkRvlhnBnbg8iwjCVnTExMIJMZx5dnfq6KsT0oK9YObzvdYiMTYQMgNjIZNk5y9PT1grWDcXxeFBZjyRpj+U7Dz1VtbBMiw8v3xdwODg44cOCA1vIDBw7AwcEBABAfHw9ra+uC146IiIwOc4aIjJW1gwzthtcULWs3vCY72/SAWUNERPqW7xFuM2bMwOeff46AgAA0btwYEokEly9fxqFDh/Djjz8CAI4dO4ZWrVrpvLJERFTyMWeIyFjFRibh+MZbomXHN97iCDc9YNYQEZG+5bvDbeTIkahRowZWrVqFvXv3QhAEVKtWDadOnUKzZs0AQDMMm4iIKL+YM0RkjNLvRhoTkQSb0hYA4mDtYIGY8LTl7HTTLWYNERHp2zvd5qh58+Zo3ry5rutCREQEgDlDRMYl7k2GzjYnGbqOrYPTF06g69h6mruU+i8NRK+JvHGCLjFriIhInwp0X/HExEStu97Y2NgUqEJERETpmDNEZAzMZKaQW5sDAHr6ekFmnTbJvcLeQnMjBbm1OcxkBfrTnbLBrCEiIn3Id2onJCRgypQp2L17N16/fq21XqVS6aRiRERknJgzRGRsLOSm6Da+HpRJqVDYy0SdP9YOMvSa6AUzmSks5Oxw0xVmDRER6Vu+71I6efJk/PXXX1izZg0sLCywYcMG+Pn5wc3NDVu2bNFHHYmIyIgwZ4jIGFnITbO9XFRhL2Nnm44xa4iISN/yndwHDhzAli1b0Lp1a3z88cd477334OnpifLly2Pbtm0YNGiQPupJRERGgjlDRET6xqwhIiJ9y/cIt8jISFSoUAFA2twGkZGRAIAWLVrg9OnTuq0dEREZHeYMERHpG7OGiIj0Ld8dbhUrVkRISAgAoEaNGti9ezeAtLNEdnZ2uqwbEREZIeYMERHpG7OGiIj0Ld8dbsOHD8f169cBANOmTdPMe/Dll19i8uTJ+a7AmjVrUKFCBchkMjRo0ABnzpzJsXxycjKmT5+O8uXLw8LCApUqVcIvv/yS7+MSEVHRpOucAZg1REQkxu80RESkb/mew+3LL7/U/L9Nmza4c+cO/v77b1SqVAl169bN17527dqFCRMmYM2aNWjevDnWrVuHzp074/bt2yhXrlyW2/Tt2xf//vsvfv75Z3h6eiI8PBypqan5fRlERFRE6TJnAGYNERFp43caIiLSt3x1uCmVSnTo0AHr1q1DlSpVAADlypXLNkhys3TpUowYMQKffPIJAGD58uU4evQo1q5diwULFmiVP3LkCE6dOoVHjx7BwcEBAODh4fFOxyYioqJH1zkDMGuIiEiM32mIiKgw5KvDzczMDP/88w8kEkmBD5ySkoKrV6/iq6++Ei3v0KEDzp8/n+U2+/fvR8OGDbF48WJs3boVVlZW6N69O+bOnQu5XJ7lNsnJyUhOTtY8j4mJAZAWtEqlssCvoyRIbwe2Rxq2hza2iRjbQ0yX7aDLnAGYNUUF3zNibA9tbBMxtoe2opo1zJmige8ZbWwTMbaHNraJmD7aId+XlA4ZMgQ///wzFi5cWKADR0REQKVSwdnZWbTc2dkZL1++zHKbR48e4ezZs5DJZNi3bx8iIiIwevRoREZGZjvnwYIFC+Dn56e1PCAgAJaWlgV6DSXNsWPHDF2FIoXtoY1tIsb2SJOQkKDT/ekqZwBmTVHD94wY20Mb20SM7fFWUc0a5kzRwveMNraJGNtDG9skja5zBniHDreUlBRs2LABx44dQ8OGDWFlZSVav3Tp0nztL/OZJUEQsj3bpFarIZFIsG3bNtja2mqO9+GHH2L16tVZnhGaNm0afH19Nc9jYmLg7u6ONm3awNHRMV91LamUSiWOHTuG9u3bw8zMzNDVMTi2hza2iRjbQ+z169c63Z+ucwZg1hga3zNibA9tbBMxtoe2op41zBnD4ntGG9tEjO2hjW0ipuucAd6hw+2ff/6Bl5cXAODevXuidfkZlu3k5ASpVKp15ic8PFzrDFE6V1dXlClTRhNMAFC9enUIgoBnz56hcuXKWttYWFjAwsJCa7mZmRl/qTJhm4ixPbSxTcTYHml03Qa6yhmAWVPUsD3E2B7a2CZibI+3imrWMGeKFraHNraJGNtDG9skjT7aIN8dbgEBATo5sLm5ORo0aIBjx46hV69emuXHjh1Djx49stymefPm+O233xAXFweFQgEgLSBNTExQtmxZndSLiIgMS1c5AzBriIgoa/xOQ0RE+mbyrhs+ePAAR48eRWJiIoC0YdP55evriw0bNuCXX35BcHAwvvzyS4SGhmLUqFEA0oZODxkyRFN+4MCBcHR0xPDhw3H79m2cPn0akydPxscff5ztBKNERFQ86SJnAGYNERFlj99piIhIX/I9wu3169fo27cvAgICIJFIcP/+fVSsWBGffPIJ7OzssGTJkjzvq1+/fnj9+jXmzJmDsLAw1KpVC4cOHUL58uUBAGFhYQgNDdWUVygUOHbsGMaNG4eGDRvC0dERffv2xbx58/L7MoiIqIjSZc4AzBoiItLG7zRERKRv+e5w+/LLL2FmZobQ0FBUr15ds7xfv3748ssv8/1FaPTo0Rg9enSW6zZt2qS1rFq1aryLBhFRCabrnAGYNUREJMbvNEREpG/57nD7888/cfToUa35BSpXrownT57orGJERGScmDNERKRvzBoiItK3fM/hFh8fD0tLS63lERERWd45h4iIKD+YM0REpG/MGiIi0rd8d7i1bNkSW7Zs0TyXSCRQq9X47rvv0KZNG51WjoiIjA9zhoiI9I1ZQ0RE+pbvS0q/++47tG7dGn///TdSUlIwZcoU3Lp1C5GRkTh37pw+6khEREaEOUNERPrGrCEiIn3L9wi3GjVq4MaNG2jcuDHat2+P+Ph49O7dG9euXUOlSpX0UUciIjIizBkiItI3Zg0REelbvke4AYCLiwv8/Px0XRciIiIAzBkiItI/Zg0REelTvke4VahQATNmzMDdu3f1UR8iIjJyzBkiItI3Zg0REelbvjvcxo0bhyNHjqB69epo0KABli9fjrCwMH3UjYiIjBBzhoiI9I1ZQ0RE+pbvDjdfX19cuXIFd+7cQdeuXbF27VqUK1cOHTp0EN3ph4iI6F0wZ4iISN+YNUREpG/57nBLV6VKFfj5+eHu3bs4c+YMXr16heHDh+uybkREZMSYM0REpG/MGiIi0pd3umlCusuXL2P79u3YtWsXoqOj8eGHH+qqXkRERMwZIiLSO2YNERHpQ7473O7du4dt27Zh+/btCAkJQZs2bbBw4UL07t0b1tbW+qgjEREZEeYMERHpG7OGiIj0Ld8dbtWqVUPDhg0xZswY9O/fHy4uLvqoFxERGSnmDBER6RuzhoiI9C3fHW537txBlSpV9FEXIiIi5gwREekds4aIiPQt3x1u6cF09epVBAcHQyKRoHr16vDy8tJ55YiIyPgwZ4iISN+YNUREpG/57nALDw9H//79cfLkSdjZ2UEQBERHR6NNmzbYuXMnSpUqpY96EhGRkWDOEBGRvjFriIhI30zyu8G4ceMQExODW7duITIyEm/evME///yDmJgYjB8/Xh91JCIiI8KcISIifWPWEBGRvuV7hNuRI0dw/PhxVK9eXbOsRo0aWL16NTp06KDTyhERkfFhzhARkb4xa4iISN/yPcJNrVbDzMxMa7mZmRnUarVOKkVERMaLOUNERPrGrCEiIn3Ld4fb+++/jy+++AIv/t/e3QdZVZh3AH4XdlkQBwyoxA9EiF8Q1CiogLWOKFikOo5tQ2oUVEizUccPkliIaQVq4phWQoyCX4iTVA1GNKUtie6kfqBkbEGIGmy0iK4aEFlHRE1ggdM/GNYczmL2wjn3LrvPM8MM9+y5d999Z+/9zf727r2/+13zsbfffjuuvfbaOPPMM3MdDoCOR84AUDRZA0DRSi7cbrvttti4cWMcfvjh8bnPfS6OOOKI6N+/f2zcuDF++MMfFjEjAB2InAGgaLIGgKKV/Bpuffv2jeeffz7q6+vjf//3fyNJkhg0aFCcddZZRcwHQAcjZwAomqwBoGglF247jBo1KkaNGpXnLADQTM4AUDRZA0BRSv6T0quuuipuvfXWzPHbbrstrrnmmjxmAqADkzMAFE3WAFC0kgu3BQsWxKmnnpo5PmLEiHj44YdzGQqAjkvOAFA0WQNA0Uou3BobG6Nnz56Z4z169Ij169fnMhQAHZecAaBosgaAopVcuB1xxBHxi1/8InP85z//eQwYMCCXoQDouOQMAEWTNQAUreQ3TZg8eXJceeWV8e6778bIkSMjIuKXv/xl3HLLLTFr1qy85wOgg5EzABRN1gBQtJILt8suuyw2bdoU3/nOd+Kf/umfIiLi8MMPjzlz5sT48eNzHxCAjkXOAFA0WQNA0Uou3CIivva1r8XXvva1ePfdd6Nbt26x77775j0XAB2YnAGgaLIGgCLtVuG2wwEHHJDXHACQIWcAKJqsAaAIJb9pAgAAAACwawo3AAAAAMiRwg0AAAAAcqRwAwAAAIActepNE2699dZW3+BVV12128MA0DHJGQCKJmsAKKdWFW7f//73W3VjVVVVwgmAkskZAIomawAop1YVbqtXry56DgA6MDkDQNFkDQDl5DXcAAAAACBHrXqG287eeuutWLhwYTQ0NMTmzZtTH5s5c2YugwHQcckZAIomawAoUsmF2y9/+cs477zzon///vHb3/42Bg8eHK+//nokSRInnnhiETMC0IHIGQCKJmsAKFrJf1I6derU+PrXvx4vvfRSdO3aNRYsWBBvvvlmnH766fE3f/M3RcwIQAciZwAomqwBoGglF24vv/xyTJgwISIiqqur4/e//33su+++MWPGjLj55ptzHxCAjkXOAFA0WQNA0Uou3Lp37x6bNm2KiIiDDz44Vq1a1fyx9evX5zcZAB2SnAGgaLIGgKKV/Bpuw4YNi2effTYGDRoUY8eOja9//evx4osvxiOPPBLDhg0rYkYAOhA5A0DRZA0ARSu5cJs5c2Z8+OGHERExbdq0+PDDD2P+/PlxxBFHxPe///3cBwSgY5EzABRN1gBQtJILtwEDBjT/f5999onZs2fnOhAAHZucAaBosgaAopX8Gm7/8z//E88991zm+HPPPRdLly7NZSgAOi45A0DRZA0ARSu5cLviiivizTffzBx/++2344orrshlKAA6LjkDQNFkDQBFK7lwW7lyZZx44omZ4yeccEKsXLkyl6EA6LjkDABFkzUAFK3kwq22tjbeeeedzPE1a9ZEdXXJLwkHAClyBoCiyRoAilZy4TZq1KiYOnVqbNiwofnY+++/H9/61rdi1KhRuQ4HQMcjZwAomqwBoGgl//rmlltuiT//8z+Pfv36xQknnBAREStWrIg+ffrEj3/849wHBKBjkTMAFE3WAFC0kgu3Qw45JF544YW4//7749e//nV069YtLr300vjbv/3bqKmpKWJGADoQOQNA0WQNAEXbrRco6N69e/zd3/1d3rMAQETIGQCKJ2sAKFKrCreFCxfGmDFjoqamJhYuXPip55533nm5DAZAxyFnACiarAGgnFpVuJ1//vmxdu3aOPDAA+P888/f5XlVVVWxdevWvGYDoIOQMwAUTdYAUE6tKty2bdvW4v8BIA9yBoCiyRoAyqlTKSc3NTXFGWecEa+88kpR8wDQgckZAIomawAoh5IKt5qamnjppZeiqqqqqHkA6MDkDABFkzUAlENJhVtExPjx42Pu3LlFzAIAcgaAwskaAIrWqtdw+2ObN2+Oe+65J+rr62Po0KHRvXv31MdnzpyZ23AAdDxyBoCiyRoAilZy4fbSSy/FiSeeGBGRed0DT8sGYE/JGQCKJmsAKFrJhdsTTzxRxBwAEBFyBoDiyRoAilbya7gBAAAAALvWqme4XXDBBXHfffdFjx494oILLvjUcx955JFcBgOg45AzABRN1gBQTq16hlvPnj2bX8ugZ8+en/qvVLNnz47+/ftH165dY8iQIbF48eJWXe/ZZ5+N6urq+MIXvlDy5wSgbSkyZyJkDQB+pgGgvFr1DLd58+a1+P89NX/+/Ljmmmti9uzZceqpp8add94ZY8aMiZUrV8Zhhx22y+tt2LAhxo8fH2eeeWa88847uc0DQGUUlTMRsgaA7fxMA0A57fZruK1bty4WL14czzzzTKxbt263bmPmzJkxceLEmDRpUgwcODBmzZoVffv2jTlz5nzq9b761a/GhRdeGMOHD9+tzwtA25dHzkTIGgB2zc80ABSl5Hcp/eCDD+KKK66In/zkJ7F169aIiOjcuXOMGzcubr/99lY/BXvz5s2xbNmymDJlSur46NGjY8mSJbu83rx582LVqlXxr//6r3HjjTf+yc+zadOm2LRpU2r+iIimpqZoampq1azt3Y492Md29pFlJ2n2kZb3HvLKmQhZ01a4z6TZR5adpNlHVlvNGjnTNrjPZNlJmn1k2UlaEXsouXCbNGlSrFixIv7jP/4jhg8fHlVVVbFkyZK4+uqr4ytf+Uo89NBDrbqd9evXx9atW6NPnz6p43369Im1a9e2eJ1XX301pkyZEosXL47q6taNftNNN8X06dMzx5944onYZ599WnUbHUV9fX2lR2hT7CPLTtLsY7uPP/4419vLK2ciZE1b4z6TZh9ZdpJmH59oq1kjZ9oW95ksO0mzjyw72S7vnInYjcLtP//zP+Oxxx6LP/uzP2s+dvbZZ8fdd98df/EXf1HyADteuHSHJEkyxyIitm7dGhdeeGFMnz49jjrqqFbf/tSpU2Py5MnNlz/44IPo27dvnHHGGdG7d++S522Pmpqaor6+PkaNGhU1NTWVHqfi7CPLTtLsI62xsTHX28s7ZyJkTaW5z6TZR5adpNlHVlvPGjlTWe4zWXaSZh9ZdpKWd85E7Ebh1rt37xafYt2zZ8/4zGc+0+rb2X///aNz586Z3/ysW7cu8xuiiIiNGzfG0qVLY/ny5XHllVdGRMS2bdsiSZKorq6Oxx9/PEaOHJm5Xm1tbdTW1maO19TU+KbaiZ2k2UeWnaTZx3Z57yCvnImQNW2NfaTZR5adpNnHJ9pq1siZtsU+suwkzT6y7GS7InZQ8psmfPvb347JkyfHmjVrmo+tXbs2vvnNb8Y//MM/tPp2unTpEkOGDMk8fbG+vj5GjBiROb9Hjx7x4osvxooVK5r/1dXVxdFHHx0rVqyIU045pdQvBYA2KK+ciZA1ALTMzzQAFK3kZ7jNmTMn/u///i/69evX/DbXDQ0NUVtbG++++27ceeedzec+//zzn3pbkydPjosvvjiGDh0aw4cPj7vuuisaGhqirq4uIrY/dfrtt9+OH/3oR9GpU6cYPHhw6voHHnhgdO3aNXMcgL1XnjkTIWsAyPIzDQBFK7lwO//883P75OPGjYvGxsaYMWNGrFmzJgYPHhyLFi2Kfv36RUTEmjVroqGhIbfPB0Dbl2fORMgaALL8TANA0Uou3G644YZcB7j88svj8ssvb/Fj991336ded9q0aTFt2rRc5wGgsvLOmQhZA0Can2kAKFrJhdsf+/DDD2Pbtm2pYz169NijgQBgBzkDQNFkDQBFKPlNE1avXh1jx46N7t27N7+Lz2c+85nYb7/9Sn73OADYmZwBoGiyBoCilfwMty9/+csREXHvvfdGnz59oqqqKvehAOi45AwARZM1ABSt5MLthRdeiGXLlsXRRx9dxDwAdHByBoCiyRoAilbyn5SedNJJ8eabbxYxCwDIGQAKJ2sAKFrJz3C75557oq6uLt5+++0YPHhw1NTUpD5+3HHH5TYcAB2PnAGgaLIGgKKVXLi9++67sWrVqrj00kubj1VVVUWSJFFVVRVbt27NdUAAOhY5A0DRZA0ARSu5cLvsssvihBNOiAcffNALjAKQOzkDQNFkDQBFK7lwe+ONN2LhwoVxxBFHFDEPAB2cnAGgaLIGgKKV/KYJI0eOjF//+tdFzAIAcgaAwskaAIpW8jPczj333Lj22mvjxRdfjGOPPTbzAqPnnXdebsMB0PHIGQCKJmsAKFrJhVtdXV1ERMyYMSPzMS8wCsCekjMAFE3WAFC0kgu3bdu2FTEHAESEnAGgeLIGgKKV/BpuAAAAAMCutbpwO+ecc2LDhg3Nl7/zne/E+++/33y5sbExBg0alOtwAHQccgaAoskaAMql1YXbY489Fps2bWq+fPPNN8d7773XfHnLli3x29/+Nt/pAOgw5AwARZM1AJRLqwu3JEk+9TIA7Ak5A0DRZA0A5eI13AAAAAAgR60u3KqqqqKqqipzDADyIGcAKJqsAaBcqlt7YpIkcckll0RtbW1ERPzhD3+Iurq66N69e0RE6rUQAKBUcgaAoskaAMql1YXbhAkTUpcvuuiizDnjx4/f84kA6JDkDABFkzUAlEurC7d58+YVOQcAHZycAaBosgaAcvGmCQAAAACQI4UbAAAAAORI4QYAAAAAOVK4AQAAAECOFG4AAAAAkCOFGwAAAADkSOEGAAAAADlSuAEAAABAjhRuAAAAAJAjhRsAAAAA5EjhBgAAAAA5UrgBAAAAQI4UbgAAAACQI4UbAAAAAORI4QYAAAAAOVK4AQAAAECOFG4AAAAAkCOFGwAAAADkSOEGAAAAADlSuAEAAABAjhRuAAAAAJAjhRsAAAAA5EjhBgAAAAA5UrgBAAAAQI4UbgAAAACQI4UbAAAAAORI4QYAAAAAOVK4AQAAAECOFG4AAAAAkCOFGwAAAADkSOEGAAAAADlSuAEAAABAjhRuAAAAAJAjhRsAAAAA5EjhBgAAAAA5UrgBAAAAQI4UbgAAAACQI4UbAAAAAORI4QYAAAAAOVK4AQAAAECOFG4AAAAAkCOFGwAAAADkSOEGAAAAADlSuAEAAABAjipeuM2ePTv69+8fXbt2jSFDhsTixYt3ee4jjzwSo0aNigMOOCB69OgRw4cPj8cee6yM0wKwN5I1ABRJzgCws4oWbvPnz49rrrkmrr/++li+fHmcdtppMWbMmGhoaGjx/KeffjpGjRoVixYtimXLlsUZZ5wR5557bixfvrzMkwOwt5A1ABRJzgDQkooWbjNnzoyJEyfGpEmTYuDAgTFr1qzo27dvzJkzp8XzZ82aFdddd12cdNJJceSRR8Z3v/vdOPLII+Pf//3fyzw5AHsLWQNAkeQMAC2prtQn3rx5cyxbtiymTJmSOj569OhYsmRJq25j27ZtsXHjxujVq9cuz9m0aVNs2rSp+fIHH3wQERFNTU3R1NS0G5O3Pzv2YB/b2UeWnaTZR1pb3oOsaRvcZ9LsI8tO0uwjq63uQs60De4zWXaSZh9ZdpJWxB4qVritX78+tm7dGn369Ekd79OnT6xdu7ZVt3HLLbfERx99FF/84hd3ec5NN90U06dPzxx/4oknYp999ilt6Hauvr6+0iO0KfaRZSdp9rHdxx9/XOkRdknWtC3uM2n2kWUnafbxibaaNXKmbXGfybKTNPvIspPtisiZihVuO1RVVaUuJ0mSOdaSBx98MKZNmxb/9m//FgceeOAuz5s6dWpMnjy5+fIHH3wQffv2jTPOOCN69+69+4O3I01NTVFfXx+jRo2KmpqaSo9TcfaRZSdp9pHW2NhY6RH+JFlTWe4zafaRZSdp9pHV1rNGzlSW+0yWnaTZR5adpBWRMxUr3Pbff//o3Llz5jc/69aty/yGaGfz58+PiRMnxk9/+tM466yzPvXc2traqK2tzRyvqanxTbUTO0mzjyw7SbOP7dryDmRN22IfafaRZSdp9vGJtroHOdO22EeWnaTZR5adbFfEDir2pgldunSJIUOGZJ6+WF9fHyNGjNjl9R588MG45JJL4oEHHoixY8cWPSYAezFZA0CR5AwAu1LRPymdPHlyXHzxxTF06NAYPnx43HXXXdHQ0BB1dXURsf2p02+//Xb86Ec/iojtwTR+/Pj4wQ9+EMOGDWv+TVK3bt2iZ8+eFfs6AGi7ZA0ARZIzALSkooXbuHHjorGxMWbMmBFr1qyJwYMHx6JFi6Jfv34REbFmzZpoaGhoPv/OO++MLVu2xBVXXBFXXHFF8/EJEybEfffdV+7xAdgLyBoAiiRnAGhJxd804fLLL4/LL7+8xY/tHDhPPvlk8QMB0O7IGgCKJGcA2FnFXsMNAAAAANojhRsAAAAA5EjhBgAAAAA5UrgBAAAAQI4UbgAAAACQI4UbAAAAAORI4QYAAAAAOVK4AQAAAECOFG4AAAAAkCOFGwAAAADkSOEGAAAAADlSuAEAAABAjhRuAAAAAJAjhRsAAAAA5EjhBgAAAAA5UrgBAAAAQI4UbgAAAACQI4UbAAAAAORI4QYAAAAAOVK4AQAAAECOFG4AAAAAkCOFGwAAAADkSOEGAAAAADlSuAEAAABAjhRuAAAAAJAjhRsAAAAA5EjhBgAAAAA5UrgBAAAAQI4UbgAAAACQI4UbAAAAAORI4QYAAAAAOVK4AQAAAECOFG4AAAAAkCOFGwAAAADkSOEGAAAAADlSuAEAAABAjhRuAAAAAJAjhRsAAAAA5EjhBgAAAAA5UrgBAAAAQI4UbgAAAACQI4UbAAAAAORI4QYAAAAAOVK4AQAAAECOFG4AAAAAkCOFGwAAAADkSOEGAAAAADlSuAEAAABAjhRuAAAAAJAjhRsAAAAA5EjhBgAAAAA5UrgBAAAAQI4UbgAAAACQI4UbAAAAAORI4QYAAAAAOVK4AQAAAECOFG4AAAAAkCOFGwAAAADkSOEGAAAAADlSuAEAAABAjhRuAAAAAJAjhRsAAAAA5EjhBgAAAAA5UrgBAAAAQI4UbgAAAACQI4UbAAAAAOSo4oXb7Nmzo3///tG1a9cYMmRILF68+FPPf+qpp2LIkCHRtWvXGDBgQNxxxx1lmhSAvZWsAaBIcgaAnVW0cJs/f35cc801cf3118fy5cvjtNNOizFjxkRDQ0OL569evTrOOeecOO2002L58uXxrW99K6666qpYsGBBmScHYG8hawAokpwBoCUVLdxmzpwZEydOjEmTJsXAgQNj1qxZ0bdv35gzZ06L599xxx1x2GGHxaxZs2LgwIExadKkuOyyy+Jf/uVfyjw5AHsLWQNAkeQMAC2prtQn3rx5cyxbtiymTJmSOj569OhYsmRJi9f51a9+FaNHj04dO/vss2Pu3LnR1NQUNTU1mets2rQpNm3a1Hx5w4YNERHx3nvv7emX0G40NTXFxx9/HI2NjS3usKOxjyw7SbOPtB2Pp0mSVHiSLFnTNrjPpNlHlp2k2UdWW80aOdM2uM9k2UmafWTZSVoROVOxwm39+vWxdevW6NOnT+p4nz59Yu3atS1eZ+3atS2ev2XLlli/fn0cdNBBmevcdNNNMX369Mzxo446ag+mB2BnjY2N0bNnz0qPkSJrANqXtpY1cgagfckzZypWuO1QVVWVupwkSebYnzq/peM7TJ06NSZPntx8+f33349+/fpFQ0NDmwrrSvrggw+ib9++8eabb0aPHj0qPU7F2UeWnaTZR9qGDRvisMMOi169elV6lF2SNZXlPpNmH1l2kmYfWW09a+RMZbnPZNlJmn1k2UlaETlTscJt//33j86dO2d+87Nu3brMb3x2+OxnP9vi+dXV1dG7d+8Wr1NbWxu1tbWZ4z179vRNtZMePXrYyR+xjyw7SbOPtE6dKv7G1xmypm1xn0mzjyw7SbOPrLaWNXKmbXGfybKTNPvIspO0PHOmYonVpUuXGDJkSNTX16eO19fXx4gRI1q8zvDhwzPnP/744zF06FB/cwxAhqwBoEhyBoBdqeiviCZPnhz33HNP3HvvvfHyyy/HtddeGw0NDVFXVxcR2586PX78+Obz6+rq4o033ojJkyfHyy+/HPfee2/MnTs3vvGNb1TqSwCgjZM1ABRJzgDQkoq+htu4ceOisbExZsyYEWvWrInBgwfHokWLol+/fhERsWbNmmhoaGg+v3///rFo0aK49tpr4/bbb4+DDz44br311virv/qrVn/O2trauOGGG1p8SnZHZSdp9pFlJ2n2kdbW9yFrKs8+0uwjy07S7COrLe9EzlSefWTZSZp9ZNlJWhH7qEra2ntrAwAAAMBerG296igAAAAA7OUUbgAAAACQI4UbAAAAAORI4QYAAAAAOWqXhdvs2bOjf//+0bVr1xgyZEgsXrz4U89/6qmnYsiQIdG1a9cYMGBA3HHHHWWatHxK2ckjjzwSo0aNigMOOCB69OgRw4cPj8cee6yM0xav1O+RHZ599tmorq6OL3zhC8UOWGal7mPTpk1x/fXXR79+/aK2tjY+97nPxb333lumacuj1J3cf//9cfzxx8c+++wTBx10UFx66aXR2NhYpmmL9fTTT8e5554bBx98cFRVVcXPfvazP3kdj6tZ7X0nciZNzmTJmjQ58wk50zI5kyZnsmRNmpxJkzOfqFjOJO3MT37yk6Smpia5++67k5UrVyZXX3110r179+SNN95o8fzXXnst2WeffZKrr746WblyZXL33XcnNTU1ycMPP1zmyYtT6k6uvvrq5Oabb07++7//O3nllVeSqVOnJjU1Ncnzzz9f5smLUeo+dnj//feTAQMGJKNHj06OP/748gxbBruzj/POOy855ZRTkvr6+mT16tXJc889lzz77LNlnLpYpe5k8eLFSadOnZIf/OAHyWuvvZYsXrw4+fznP5+cf/75ZZ68GIsWLUquv/76ZMGCBUlEJI8++uinnu9xNau970TOpMmZLFmTJmfS5EyWnEmTM1myJk3OpMmZtErlTLsr3E4++eSkrq4udeyYY45JpkyZ0uL51113XXLMMcekjn31q19Nhg0bVtiM5VbqTloyaNCgZPr06XmPVhG7u49x48Yl3/72t5MbbrihXYVTqfv4+c9/nvTs2TNpbGwsx3gVUepO/vmf/zkZMGBA6titt96aHHrooYXNWCmtCSiPq1ntfSdyJk3OZMmaNDmza3JmOzmTJmeyZE2anEmTM7tWzpxpV39Sunnz5li2bFmMHj06dXz06NGxZMmSFq/zq1/9KnP+2WefHUuXLo2mpqbCZi2X3dnJzrZt2xYbN26MXr16FTFiWe3uPubNmxerVq2KG264oegRy2p39rFw4cIYOnRofO9734tDDjkkjjrqqPjGN74Rv//978sxcuF2ZycjRoyIt956KxYtWhRJksQ777wTDz/8cIwdO7YcI7c5Hlez2vNO5EyanMmSNWlyZs+158fUCDmzMzmTJWvS5EyanNlzeT2mVuc9WCWtX78+tm7dGn369Ekd79OnT6xdu7bF66xdu7bF87ds2RLr16+Pgw46qLB5y2F3drKzW265JT766KP44he/WMSIZbU7+3j11VdjypQpsXjx4qiubld3md3ax2uvvRbPPPNMdO3aNR599NFYv359XH755fHee++1i9c82J2djBgxIu6///4YN25c/OEPf4gtW7bEeeedFz/84Q/LMXKb43E1qz3vRM6kyZksWZMmZ/Zce35MjZAzO5MzWbImTc6kyZk9l9djart6htsOVVVVqctJkmSO/anzWzq+Nyt1Jzs8+OCDMW3atJg/f34ceOCBRY1Xdq3dx9atW+PCCy+M6dOnx1FHHVWu8cqulO+Pbdu2RVVVVdx///1x8sknxznnnBMzZ86M++67r138RmiHUnaycuXKuOqqq+If//EfY9myZfGLX/wiVq9eHXV1deUYtU3yuNq681s6vreSM2lyJkvWpMmZPdPeH1Mj5MzO5EyWrEmTM2lyZs/k8Zjarqrt/fffPzp37pxpbdetW5dpJ3f47Gc/2+L51dXV0bt378JmLZfd2ckO8+fPj4kTJ8ZPf/rTOOuss4ocs2xK3cfGjRtj6dKlsXz58rjyyisjYvuDc5IkUV1dHY8//niMHDmyLLMXYXe+Pw466KA45JBDomfPns3HBg4cGEmSxFtvvRVHHnlkoTMXbXd2ctNNN8Wpp54a3/zmNyMi4rjjjovu3bvHaaedFjfeeONe/Vvl3eFxNas970TOpMmZLFmTJmf2XHt+TI2QMzuTM1myJk3OpMmZPZfXY2q7eoZbly5dYsiQIVFfX586Xl9fHyNGjGjxOsOHD8+c//jjj8fQoUOjpqamsFnLZXd2ErH9N0GXXHJJPPDAA+3q77ZL3UePHj3ixRdfjBUrVjT/q6uri6OPPjpWrFgRp5xySrlGL8TufH+ceuqp8bvf/S4+/PDD5mOvvPJKdOrUKQ499NBC5y2H3dnJxx9/HJ06pR9OO3fuHBGf/CakI/G4mtWedyJn0uRMlqxJkzN7rj0/pkbImZ3JmSxZkyZn0uTMnsvtMbWkt1jYC+x4+9u5c+cmK1euTK655pqke/fuyeuvv54kSZJMmTIlufjii5vP3/F2r9dee22ycuXKZO7cue3qLbSTpPSdPPDAA0l1dXVy++23J2vWrGn+9/7771fqS8hVqfvYWXt7R59S97Fx48bk0EMPTf76r/86+c1vfpM89dRTyZFHHplMmjSpUl9C7krdybx585Lq6upk9uzZyapVq5JnnnkmGTp0aHLyySdX6kvI1caNG5Ply5cny5cvTyIimTlzZrJ8+fLmtxX3uCpr5EyanMmSNWlyJk3OZMmZNDmTJWvS5EyanEmrVM60u8ItSZLk9ttvT/r165d06dIlOfHEE5Onnnqq+WMTJkxITj/99NT5Tz75ZHLCCSckXbp0SQ4//PBkzpw5ZZ64eKXs5PTTT08iIvNvwoQJ5R+8IKV+j/yx9hZOSVL6Pl5++eXkrLPOSrp165YceuihyeTJk5OPP/64zFMXq9Sd3HrrrcmgQYOSbt26JQcddFDy5S9/OXnrrbfKPHUxnnjiiU99TPC4KmuSRM7sTM5kyZo0OfMJOdMyOZMmZ7JkTZqcSZMzn6hUzlQlSQd8fiAAAAAAFKRdvYYbAAAAAFSawg0AAAAAcqRwAwAAAIAcKdwAAAAAIEcKNwAAAADIkcINAAAAAHKkcAMAAACAHCncAAAAACBHCjcAAAAAyJHCDdqpF154IYYPHx7HHXdcfPe7342//Mu/rPRIALQzsgaAIskZ9mYKN8jZ008/Heeee24cfPDBUVVVFT/72c/KPkNTU1NMmDAh5s2bFy+88EI888wzceyxx5Z9DgCKIWsAKJKcgT2ncIOcffTRR3H88cfHbbfdVrEZHn744Tj11FPjmGOOiYiIgQMHxnHHHVexeQDIl6wBoEhyBvacwg1yNmbMmLjxxhvjggsuKPTz3HHHHXHsscdGt27domfPnjFy5Mjmj7300kupMPrNb34jnADaEVkDQJHkDOy56koPAJRuwYIFMWXKlLjrrrti2LBhsXHjxnj99debP967d+949dVXIyLiv/7rv+Lpp5+Oo48+ukLTArA3kjUAFEnO0N4p3GAv9Morr8Rhhx0Wo0ePjv322y8iIj7/+c83f/yiiy6Kc845J4YNGxYjRoyIU045Jaqr3d0BaD1ZA0CR5AztnT8phQqaNm1aVFVVfeq/pUuXZq73la98JTp37hy9evWKfffdN1atWpX6ePfu3WPp0qWxZMmS6NSpU1x00UXl+pIAaGNkDQBFkjPQMvUwVNCVV14ZX/rSlz71nMMPPzx1uampKb70pS/FSSedFHfffXfst99+MWDAgNQ53/ve92LBggVRXV0dY8eOjcsuuyzv0QHYS8gaAIokZ6BlVUmSJJUeAtqrqqqqePTRR+P888/P7TYfeuihuO6661KvbwBAxyVrACiSnIHd409KIWcffvhhrFixIlasWBEREatXr44VK1ZEQ0NDLre/efPmWLNmTfz4xz+O119/PV566aW48847o6mpKZfbB6DtkzUAFEnOwJ7zDDfI2ZNPPhlnnHFG5viECRPivvvu2+Pb37JlS/z93/99PPTQQ/HOO+9Er1694swzz4z7779/j28bgL2DrAGgSHIG9pzCDQAAAABy5E9KAQAAACBHCjcAAAAAyJHCDQAAAABypHADAAAAgBwp3AAAAAAgRwo3AAAAAMiRwg0AAAAAcqRwAwAAAIAcKdwAAAAAIEcKNwAAAADIkcINAAAAAHKkcAMAAACAHP0/gjUyByqVfEcAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 1500x500 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "validity_tbls = SummaryReporter(epsilons, approaches, detectors, k_folds=folds).report(res)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>$1 - \\epsilon^g$</th>\n",
       "      <th>Bonferroni</th>\n",
       "      <th>Max additive</th>\n",
       "      <th>Independent DE-CCP</th>\n",
       "      <th>Gumbel DE-CCP</th>\n",
       "      <th>Empirical DE-CCP</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.99</td>\n",
       "      <td>0.992±0.002</td>\n",
       "      <td>0.989±0.002</td>\n",
       "      <td>0.990±0.002</td>\n",
       "      <td>0.989±0.002</td>\n",
       "      <td>0.989±0.002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.95</td>\n",
       "      <td>0.951±0.007</td>\n",
       "      <td>0.950±0.005</td>\n",
       "      <td>0.949±0.006</td>\n",
       "      <td>0.946±0.007</td>\n",
       "      <td>0.945±0.007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.90</td>\n",
       "      <td>0.911±0.009</td>\n",
       "      <td>0.901±0.006</td>\n",
       "      <td>0.904±0.010</td>\n",
       "      <td>0.899±0.010</td>\n",
       "      <td>0.894±0.012</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.80</td>\n",
       "      <td>0.831±0.012</td>\n",
       "      <td>0.802±0.012</td>\n",
       "      <td>0.813±0.013</td>\n",
       "      <td>0.804±0.013</td>\n",
       "      <td>0.762±0.094</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   $1 - \\epsilon^g$   Bonferroni Max additive Independent DE-CCP  \\\n",
       "0              0.99  0.992±0.002  0.989±0.002        0.990±0.002   \n",
       "1              0.95  0.951±0.007  0.950±0.005        0.949±0.006   \n",
       "2              0.90  0.911±0.009  0.901±0.006        0.904±0.010   \n",
       "3              0.80  0.831±0.012  0.802±0.012        0.813±0.013   \n",
       "\n",
       "  Gumbel DE-CCP Empirical DE-CCP  \n",
       "0   0.989±0.002      0.989±0.002  \n",
       "1   0.946±0.007      0.945±0.007  \n",
       "2   0.899±0.010      0.894±0.012  \n",
       "3   0.804±0.013      0.762±0.094  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "validity_tbls[0] # Experiment N°1: correlated dissimilarities"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'0.992±0.002'"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "method, vldts, j = \"Bonferroni\", [], 0\n",
    "for k in range(folds):\n",
    "    all = res.logs[method, detectors[0], k, epsilons[j]][\"test_res\"]\n",
    "    gt_boxes = all[\"Ground truths\"].xs(\"Exact value\", level=\"Bounds\", axis=1)\n",
    "    outer_boxes = all[method].xs(\"Upper bound\", level=\"Bounds\", axis=1)\n",
    "    inner_boxes = all[method].xs(\"Lower bound\", level=\"Bounds\", axis=1)\n",
    "    vldt = (  \n",
    "        (gt_boxes[\"x_min\"] >= outer_boxes[\"x_min\"]) \n",
    "        & (gt_boxes[\"y_min\"] >= outer_boxes[\"y_min\"])\n",
    "        & (gt_boxes[\"x_max\"] <= outer_boxes[\"x_max\"]) \n",
    "        & (gt_boxes[\"y_max\"] <= outer_boxes[\"y_max\"])\n",
    "        & (gt_boxes[\"x_min\"] <= inner_boxes[\"x_min\"])\n",
    "        & (gt_boxes[\"y_min\"] <= inner_boxes[\"y_min\"])\n",
    "        & (gt_boxes[\"x_max\"] >= inner_boxes[\"x_max\"]) \n",
    "        & (gt_boxes[\"y_max\"] >= inner_boxes[\"y_max\"])\n",
    "    )\n",
    "    vldt = vldt.value_counts().get(True, 0) / len(vldt)\n",
    "    vldts.append(vldt)\n",
    "f\"{np.mean(vldts):.3f}\\u00B1{np.std(vldts):.3f}\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>$1 - \\epsilon^g$</th>\n",
       "      <th>Bonferroni</th>\n",
       "      <th>Max additive</th>\n",
       "      <th>Independent DE-CCP</th>\n",
       "      <th>Gumbel DE-CCP</th>\n",
       "      <th>Empirical DE-CCP</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.99</td>\n",
       "      <td>0.990±0.001</td>\n",
       "      <td>0.990±0.003</td>\n",
       "      <td>0.988±0.001</td>\n",
       "      <td>0.986±0.003</td>\n",
       "      <td>0.987±0.002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.95</td>\n",
       "      <td>0.949±0.004</td>\n",
       "      <td>0.951±0.006</td>\n",
       "      <td>0.946±0.004</td>\n",
       "      <td>0.946±0.003</td>\n",
       "      <td>0.937±0.030</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.90</td>\n",
       "      <td>0.904±0.006</td>\n",
       "      <td>0.902±0.009</td>\n",
       "      <td>0.896±0.006</td>\n",
       "      <td>0.897±0.005</td>\n",
       "      <td>0.894±0.005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.80</td>\n",
       "      <td>0.822±0.009</td>\n",
       "      <td>0.803±0.010</td>\n",
       "      <td>0.791±0.012</td>\n",
       "      <td>0.793±0.011</td>\n",
       "      <td>0.635±0.190</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   $1 - \\epsilon^g$   Bonferroni Max additive Independent DE-CCP  \\\n",
       "0              0.99  0.990±0.001  0.990±0.003        0.988±0.001   \n",
       "1              0.95  0.949±0.004  0.951±0.006        0.946±0.004   \n",
       "2              0.90  0.904±0.006  0.902±0.009        0.896±0.006   \n",
       "3              0.80  0.822±0.009  0.803±0.010        0.791±0.012   \n",
       "\n",
       "  Gumbel DE-CCP Empirical DE-CCP  \n",
       "0   0.986±0.003      0.987±0.002  \n",
       "1   0.946±0.003      0.937±0.030  \n",
       "2   0.897±0.005      0.894±0.005  \n",
       "3   0.793±0.011      0.635±0.190  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "validity_tbls[1] # Experiment N°2: independent dissimilarities"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>$1 - \\epsilon^g$</th>\n",
       "      <th>Bonferroni</th>\n",
       "      <th>Max additive</th>\n",
       "      <th>Independent DE-CCP</th>\n",
       "      <th>Gumbel DE-CCP</th>\n",
       "      <th>Empirical DE-CCP</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.99</td>\n",
       "      <td>0.990±0.004</td>\n",
       "      <td>0.985±0.006</td>\n",
       "      <td>0.985±0.005</td>\n",
       "      <td>0.985±0.006</td>\n",
       "      <td>0.979±0.007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.95</td>\n",
       "      <td>0.958±0.010</td>\n",
       "      <td>0.927±0.011</td>\n",
       "      <td>0.955±0.011</td>\n",
       "      <td>0.951±0.009</td>\n",
       "      <td>0.920±0.010</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.90</td>\n",
       "      <td>0.920±0.011</td>\n",
       "      <td>0.869±0.012</td>\n",
       "      <td>0.911±0.011</td>\n",
       "      <td>0.902±0.013</td>\n",
       "      <td>0.859±0.014</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.80</td>\n",
       "      <td>0.842±0.018</td>\n",
       "      <td>0.760±0.021</td>\n",
       "      <td>0.823±0.017</td>\n",
       "      <td>0.803±0.016</td>\n",
       "      <td>0.753±0.020</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   $1 - \\epsilon^g$   Bonferroni Max additive Independent DE-CCP  \\\n",
       "0              0.99  0.990±0.004  0.985±0.006        0.985±0.005   \n",
       "1              0.95  0.958±0.010  0.927±0.011        0.955±0.011   \n",
       "2              0.90  0.920±0.011  0.869±0.012        0.911±0.011   \n",
       "3              0.80  0.842±0.018  0.760±0.021        0.823±0.017   \n",
       "\n",
       "  Gumbel DE-CCP Empirical DE-CCP  \n",
       "0   0.985±0.006      0.979±0.007  \n",
       "1   0.951±0.009      0.920±0.010  \n",
       "2   0.902±0.013      0.859±0.014  \n",
       "3   0.803±0.016      0.753±0.020  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "validity_tbls[2] # Experiment N°3: YOLOv8's dissimilarities"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Efficiency"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">> Reporting efficiency ...\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x500 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "efficiency_tbls = SummaryReporter(epsilons, approaches, detectors, k_folds=folds, summarize_validity=False).report(res)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>$1 - \\epsilon^g$</th>\n",
       "      <th>Bonferroni</th>\n",
       "      <th>Max additive</th>\n",
       "      <th>Independent DE-CCP</th>\n",
       "      <th>Gumbel DE-CCP</th>\n",
       "      <th>Empirical DE-CCP</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.99</td>\n",
       "      <td>1.16e+04±5.86e+02</td>\n",
       "      <td>1.52e+05±1.04e+04</td>\n",
       "      <td>1.09e+04±5.17e+02</td>\n",
       "      <td>1.06e+04±4.09e+02</td>\n",
       "      <td>1.04e+04±3.61e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.95</td>\n",
       "      <td>5.09e+03±2.73e+02</td>\n",
       "      <td>5.20e+04±2.66e+03</td>\n",
       "      <td>4.94e+03±2.34e+02</td>\n",
       "      <td>4.79e+03±2.55e+02</td>\n",
       "      <td>4.73e+03±2.44e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.90</td>\n",
       "      <td>3.31e+03±1.61e+02</td>\n",
       "      <td>2.58e+04±1.26e+03</td>\n",
       "      <td>3.17e+03±1.69e+02</td>\n",
       "      <td>3.04e+03±1.54e+02</td>\n",
       "      <td>2.94e+03±1.58e+02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.80</td>\n",
       "      <td>1.93e+03±7.48e+01</td>\n",
       "      <td>9.77e+03±6.77e+02</td>\n",
       "      <td>1.77e+03±6.73e+01</td>\n",
       "      <td>1.68e+03±5.94e+01</td>\n",
       "      <td>1.46e+03±3.67e+02</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   $1 - \\epsilon^g$         Bonferroni       Max additive Independent DE-CCP  \\\n",
       "0              0.99  1.16e+04±5.86e+02  1.52e+05±1.04e+04  1.09e+04±5.17e+02   \n",
       "1              0.95  5.09e+03±2.73e+02  5.20e+04±2.66e+03  4.94e+03±2.34e+02   \n",
       "2              0.90  3.31e+03±1.61e+02  2.58e+04±1.26e+03  3.17e+03±1.69e+02   \n",
       "3              0.80  1.93e+03±7.48e+01  9.77e+03±6.77e+02  1.77e+03±6.73e+01   \n",
       "\n",
       "       Gumbel DE-CCP   Empirical DE-CCP  \n",
       "0  1.06e+04±4.09e+02  1.04e+04±3.61e+02  \n",
       "1  4.79e+03±2.55e+02  4.73e+03±2.44e+02  \n",
       "2  3.04e+03±1.54e+02  2.94e+03±1.58e+02  \n",
       "3  1.68e+03±5.94e+01  1.46e+03±3.67e+02  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "efficiency_tbls[0] # Experiment N°1: correlated dissimilarities"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'3.17e+03±1.69e+02'"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "method, effs, j = \"Independent DE-CCP\", [], 2\n",
    "for k in range(folds):\n",
    "    all = res.logs[method, detectors[0], k, epsilons[j]][\"test_res\"]\n",
    "    outers = all[method].xs(\"Upper bound\", level=\"Bounds\", axis=1)\n",
    "    inners = all[method].xs(\"Lower bound\", level=\"Bounds\", axis=1)\n",
    "    eff = (outers - inners).abs().product(axis=1).mean()\n",
    "    effs.append(eff.item())\n",
    "f\"{np.mean(effs):.2e}\\u00B1{np.std(effs):.2e}\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>$1 - \\epsilon^g$</th>\n",
       "      <th>Bonferroni</th>\n",
       "      <th>Max additive</th>\n",
       "      <th>Independent DE-CCP</th>\n",
       "      <th>Gumbel DE-CCP</th>\n",
       "      <th>Empirical DE-CCP</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.99</td>\n",
       "      <td>2.22e+03±3.72e+00</td>\n",
       "      <td>6.31e+04±3.82e+02</td>\n",
       "      <td>2.21e+03±3.62e+00</td>\n",
       "      <td>2.21e+03±4.61e+00</td>\n",
       "      <td>2.21e+03±3.97e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.95</td>\n",
       "      <td>2.13e+03±5.57e+00</td>\n",
       "      <td>5.38e+04±4.50e+02</td>\n",
       "      <td>2.12e+03±6.69e+00</td>\n",
       "      <td>2.12e+03±6.48e+00</td>\n",
       "      <td>2.09e+03±6.70e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.90</td>\n",
       "      <td>2.03e+03±8.13e+00</td>\n",
       "      <td>4.36e+04±6.53e+02</td>\n",
       "      <td>2.00e+03±9.84e+00</td>\n",
       "      <td>2.00e+03±7.38e+00</td>\n",
       "      <td>2.00e+03±8.36e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.80</td>\n",
       "      <td>1.83e+03±9.59e+00</td>\n",
       "      <td>2.72e+04±6.80e+02</td>\n",
       "      <td>1.77e+03±1.23e+01</td>\n",
       "      <td>1.77e+03±1.36e+01</td>\n",
       "      <td>1.43e+03±4.14e+02</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   $1 - \\epsilon^g$         Bonferroni       Max additive Independent DE-CCP  \\\n",
       "0              0.99  2.22e+03±3.72e+00  6.31e+04±3.82e+02  2.21e+03±3.62e+00   \n",
       "1              0.95  2.13e+03±5.57e+00  5.38e+04±4.50e+02  2.12e+03±6.69e+00   \n",
       "2              0.90  2.03e+03±8.13e+00  4.36e+04±6.53e+02  2.00e+03±9.84e+00   \n",
       "3              0.80  1.83e+03±9.59e+00  2.72e+04±6.80e+02  1.77e+03±1.23e+01   \n",
       "\n",
       "       Gumbel DE-CCP   Empirical DE-CCP  \n",
       "0  2.21e+03±4.61e+00  2.21e+03±3.97e+00  \n",
       "1  2.12e+03±6.48e+00  2.09e+03±6.70e+01  \n",
       "2  2.00e+03±7.38e+00  2.00e+03±8.36e+00  \n",
       "3  1.77e+03±1.36e+01  1.43e+03±4.14e+02  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "efficiency_tbls[1] # Experiment N°2: independent dissimilarities"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>$1 - \\epsilon^g$</th>\n",
       "      <th>Bonferroni</th>\n",
       "      <th>Max additive</th>\n",
       "      <th>Independent DE-CCP</th>\n",
       "      <th>Gumbel DE-CCP</th>\n",
       "      <th>Empirical DE-CCP</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.99</td>\n",
       "      <td>1.13e+10±7.34e+09</td>\n",
       "      <td>3.42e+10±3.52e+10</td>\n",
       "      <td>6.14e+09±3.67e+09</td>\n",
       "      <td>5.02e+09±3.15e+09</td>\n",
       "      <td>8.63e+08±6.48e+08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.95</td>\n",
       "      <td>5.72e+07±5.47e+07</td>\n",
       "      <td>5.62e+05±3.99e+05</td>\n",
       "      <td>1.13e+07±1.25e+07</td>\n",
       "      <td>6.89e+06±9.47e+06</td>\n",
       "      <td>3.20e+05±2.05e+05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.90</td>\n",
       "      <td>5.11e+05±4.32e+05</td>\n",
       "      <td>5.60e+04±2.33e+04</td>\n",
       "      <td>2.45e+05±1.83e+05</td>\n",
       "      <td>1.52e+05±9.69e+04</td>\n",
       "      <td>3.70e+04±1.52e+04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.80</td>\n",
       "      <td>2.83e+04±1.44e+04</td>\n",
       "      <td>7.70e+03±2.31e+03</td>\n",
       "      <td>1.86e+04±8.45e+03</td>\n",
       "      <td>1.31e+04±5.25e+03</td>\n",
       "      <td>6.04e+03±1.81e+03</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   $1 - \\epsilon^g$         Bonferroni       Max additive Independent DE-CCP  \\\n",
       "0              0.99  1.13e+10±7.34e+09  3.42e+10±3.52e+10  6.14e+09±3.67e+09   \n",
       "1              0.95  5.72e+07±5.47e+07  5.62e+05±3.99e+05  1.13e+07±1.25e+07   \n",
       "2              0.90  5.11e+05±4.32e+05  5.60e+04±2.33e+04  2.45e+05±1.83e+05   \n",
       "3              0.80  2.83e+04±1.44e+04  7.70e+03±2.31e+03  1.86e+04±8.45e+03   \n",
       "\n",
       "       Gumbel DE-CCP   Empirical DE-CCP  \n",
       "0  5.02e+09±3.15e+09  8.63e+08±6.48e+08  \n",
       "1  6.89e+06±9.47e+06  3.20e+05±2.05e+05  \n",
       "2  1.52e+05±9.69e+04  3.70e+04±1.52e+04  \n",
       "3  1.31e+04±5.25e+03  6.04e+03±1.81e+03  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "efficiency_tbls[2] # Experiment N°3: YOLOv8's dissimilarities"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Object detection results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Examples"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "ename": "TypeError",
     "evalue": "__init__() missing 1 required positional argument: 'num_frames'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[11], line 1\u001b[0m\n\u001b[0;32m----> 1\u001b[0m giffer \u001b[38;5;241m=\u001b[39m \u001b[43mGifReporter\u001b[49m\u001b[43m(\u001b[49m\u001b[43mdest\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mMax_additive_vs_Empiricial_DE-CPP.gif\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m)\u001b[49m\n\u001b[1;32m      2\u001b[0m giffer\u001b[38;5;241m.\u001b[39mreport(res, save\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mTrue\u001b[39;00m)\n",
      "\u001b[0;31mTypeError\u001b[0m: __init__() missing 1 required positional argument: 'num_frames'"
     ]
    }
   ],
   "source": [
    "giffer = GifReporter(dest=\"Max_additive_vs_Empiricial_DE-CPP.gif\")\n",
    "giffer.report(res, save=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## DELETE ME"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(3007, 16), (2756, 16), (744, 16)]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[res.logs[approaches[0], det, 0, epsilons[0]][\"cal_res\"].shape for det in detectors]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(9781, 16), (9674, 16), (2740, 16)]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[res.logs[approaches[0], det, 0, epsilons[0]][\"test_res\"].shape for det in detectors]"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {
    "vscode": {
     "languageId": "raw"
    }
   },
   "source": [
    "from pandas import DataFrame\n",
    "from PIL import Image, ImageDraw\n",
    "from typing import List, Any\n",
    "\n",
    "\n",
    "log_idx, mode = -1, \"cal_res\"\n",
    "method = res.logs[log_idx][\"cfmz_name\"]\n",
    "fname, _ = res.logs[log_idx][mode].iloc[0].name\n",
    "gt_bboxes = res.logs[log_idx][mode][\"Ground truths\"].loc[fname]\n",
    "conf_intevals = res.logs[log_idx][mode][method].loc[fname]\n",
    "inner_bboxes = conf_intevals.xs(\"Lower bound\", level=\"Bounds\", axis=1)\n",
    "outer_bboxes = conf_intevals.xs(\"Upper bound\", level=\"Bounds\", axis=1)\n",
    "bboxes = zip(gt_bboxes.iterrows(), inner_bboxes.iterrows(), outer_bboxes.iterrows())\n",
    "\n",
    "\"\"\"\n",
    "frames = [{\"name\": fname, \n",
    "          \"Ground truths\": gt_bboxes,\n",
    "          method1: conf_intevals1,\n",
    "          method2: conf_intevals2\n",
    "          }]\n",
    "\"\"\"\n",
    "\n",
    "src = \"datasets/FilteredKitti/raw/training/image_2\"\n",
    "img = Image.open(f'{src}/{fname}')\n",
    "canvas = ImageDraw.Draw(img)\n",
    "\n",
    "for item in zip(bboxes):\n",
    "    gt, inner, outer = item[0]\n",
    "    canvas.rectangle(gt[1], fill=None, outline=\"green\")\n",
    "    canvas.rectangle(inner[1], fill=None, outline=\"red\")\n",
    "    canvas.rectangle(outer[1], fill=None, outline=\"blue\")\n",
    "img.show()\n",
    "\n",
    "\n",
    "class GifReporter():\n",
    "    def __init__(self, dest: str, frames: List = None) -> None:\n",
    "        self.gif = None\n",
    "        self.dest = dest\n",
    "        self.frames = frames\n",
    "\n",
    "    def report(self, res: UQEvalResults, filter: dict,\n",
    "               save = False, split_num: int = 0, \n",
    "               method1: str = \"Empirical DE-CCP\",\n",
    "               method2: str = \"Max additive\") -> Any:  # Gif\n",
    "        filter.update({\"split_num\": split_num})\n",
    "        self.current_results = res\n",
    "        if self.frames is None:\n",
    "            self.random_frames(filter, [method1, method2])\n",
    "        else:\n",
    "            self.find_frames(filter, [method1, method2])\n",
    "        self.draw_bounding_boxes()\n",
    "        self.save_gif() if save else None\n",
    "        return self.gif\n",
    "    \n",
    "    def random_frames(self, filter: dict, methods: List, \n",
    "                      num_frames: int = 10) -> None:\n",
    "        i, self.frames, assigned = 0, [], [] \n",
    "        while len(self.frames) < 10:\n",
    "            fname = None\n",
    "            filter.items() + \n",
    "            if (all(self.current_results[k] == v for k, v in kvs)\n",
    "                and fname not in assigned):\n",
    "                gt_bboxes, conf_intevals1, conf_intevals2 = [None] * 3\n",
    "                self.frames.append({\n",
    "                    \"name\": fname, \"Ground truths\": gt_bboxes,\n",
    "                    methods[0]: conf_intevals1,methods[1]: conf_intevals2})\n",
    "            i += 1 \n",
    "\n",
    "    \n",
    "    def find_frames(self, filter: dict, methods: List) -> None:\n",
    "        pass\n",
    "\n",
    "\n",
    "giffer = GifReporter(dest=\"KITTI_results.gif\")\n",
    "giffer.report(res)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "TBD",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.18"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}