notebooks/AudioNetworkNetworkPredictions.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import os \n",
    "import glob\n",
    "import sys\n",
    "\n",
    "import math\n",
    "import json\n",
    "import scipy\n",
    "import scipy.stats\n",
    "%matplotlib inline\n",
    "import matplotlib.pylab as plt\n",
    "import matplotlib\n",
    "import pickle as pckl\n",
    "from PIL import Image\n",
    "import seaborn as sns\n",
    "\n",
    "from model_analysis_folders import all_model_info\n",
    "import urllib.request, json\n",
    "from scipy.io import loadmat\n",
    "\n",
    "import matplotlib\n",
    "matplotlib.rcParams['pdf.fonttype'] = 42\n",
    "matplotlib.rcParams['ps.fonttype'] = 42\n",
    "\n",
    "matplotlib.rcParams.update({'font.size': 6})\n",
    "\n",
    "from notebook_helpers import * \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "models_to_evaluate = [\"kell2018\",\"cochresnet50\",\n",
    "                      \"cochresnet50_l2_1_robust_waveform\", \"cochresnet50_l2_p5_robust_waveform\", \"cochresnet50_linf_p002_robust_waveform\",\n",
    "                      \"cochresnet50_l2_1_random_step_waveform\", \"cochresnet50_linf_p002_random_step_waveform\",\n",
    "                      \"kell2018_l2_1_robust_waveform\", \"kell2018_linf_p002_robust_waveform\",\n",
    "                      \"kell2018_l2_1_random_step_waveform\", \"kell2018_linf_p002_random_step_waveform\",\n",
    "                      \"cochresnet50_l2_1_robust_cochleagram\", \"cochresnet50_l2_p5_robust_cochleagram\",\n",
    "                      \"cochresnet50_l2_1_random_step_cochleagram\",\n",
    "                      \"kell2018_l2_1_robust_cochleagram\", \"kell2018_l2_p5_robust_cochleagram\",\n",
    "                      \"kell2018_l2_1_random_step_cochleagram\",\n",
    "                     ]\n",
    "\n",
    "# Set the colors for the models we are measuring. \n",
    "model_cmap = [\"#000000\",\"#004949\",\"#009292\",\"#ff6db6\",\"#ffb6db\",\n",
    " \"#490092\",\"#006ddb\",\"#b66dff\",\"#6db6ff\",\"#b6dbff\",\n",
    " \"#920000\",\"#924900\",\"#db6d00\",\"#24ff24\",\"#ffff6d\"]\n",
    "\n",
    "plot_color_idx = [0,3,\n",
    "                10,11,12,\n",
    "                10,12,\n",
    "                1,2,\n",
    "                1,2,\n",
    "                10,11,\n",
    "                10,\n",
    "                1,2,\n",
    "                1,\n",
    "                ]\n",
    "\n",
    "\n",
    "assert (len(np.unique(plot_color_idx)) <= len(model_cmap)), 'Not enough colors to plot all of the networks!'\n",
    "model_cmap_dict = {m:model_cmap[plot_color_idx[m_idx]] for m_idx, m in enumerate(models_to_evaluate)}\n",
    "\n",
    "# Set the line styles for the models we are measuring\n",
    "\n",
    "plot_style = ['-','-',\n",
    "              '-','-','-',\n",
    "              ':',':',\n",
    "              '-','-',\n",
    "              ':',':',\n",
    "              '--', '--',\n",
    "              '-.', \n",
    "              '--', '--',\n",
    "              '-.',]\n",
    "\n",
    "model_style_dict = {m:plot_style[m_idx] for m_idx, m in enumerate(models_to_evaluate)}\n",
    "\n",
    "\n",
    "plot_num = [0,0,\n",
    "            1,1,1,\n",
    "            1,1,\n",
    "            2,2,\n",
    "            2,2,\n",
    "            3,3,\n",
    "            3,\n",
    "            4,4,\n",
    "            4] \n",
    "\n",
    "network_network_folder = 'network_network_evaluations'\n",
    "\n",
    "models_by_type = {'kell2018':'standard',\n",
    "              'cochresnet50':'standard',\n",
    "              'cochresnet50_l2_1_robust_waveform':'adv_trained_wav',\n",
    "              'cochresnet50_l2_p5_robust_waveform':'adv_trained_wav',\n",
    "              'cochresnet50_linf_p002_robust_waveform':'adv_trained_wav',                 \n",
    "              \"cochresnet50_l2_1_random_step_waveform\":'rand_trained_wav',\n",
    "              \"cochresnet50_linf_p002_random_step_waveform\":'rand_trained_wav',                  \n",
    "              'kell2018_l2_1_robust_waveform':'adv_trained_wav',\n",
    "              'kell2018_linf_p002_robust_waveform':'adv_trained_wav',                  \n",
    "              \"kell2018_l2_1_random_step_waveform\":'rand_trained_wav',\n",
    "              \"kell2018_linf_p002_random_step_waveform\":'rand_trained_wav',                  \n",
    "              'cochresnet50_l2_1_robust_cochleagram':'adv_trained_coch',\n",
    "              'cochresnet50_l2_p5_robust_cochleagram':'adv_trained_coch',                  \n",
    "              'cochresnet50_l2_1_random_step_cochleagram':'rand_trained_coch',\n",
    "              'kell2018_l2_1_robust_cochleagram':'adv_trained_coch',\n",
    "              'kell2018_l2_p5_robust_cochleagram':'adv_trained_coch',                  \n",
    "              'kell2018_l2_1_random_step_cochleagram':'rand_trained_coch',\n",
    "             }\n",
    "\n",
    "model_types = ['standard','adv_trained_wav','adv_trained_coch','rand_trained_wav','rand_trained_coch']\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAf0AAAJXCAYAAACZo42dAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOydeZxUxbX4v6eX2Xs2ZmXYmQFkG2TYJSyKyy/RmE1jXJJnfDEmJhrMe8YkL5GQxPgSozExiyZG8xK3GLe4AMqisqjIILIKCDMwwMDAMMMMs3f3+f1xb7c9TU/PgLNT38+nP327TlXdunX63lNVt+qUqCoGg8FgMBj6P46eLoDBYDAYDIbuwRh9g8FgMBjOEozRNxgMBoPhLMEYfYPBYDAYzhKM0TcYDAaD4SzBGH2DwWAwGM4SeqXRF5EYEfmdiDwgIg+KyNwOpFkkIuPDwr4oIn8VkT+IyA9FZJ6IlIlIoojEicijdjyviBTZx0/a37eKyMMi8hcRyQnJU+wyBT4ZIvKoiCSdwXVGTScijpDjYSJyz+meowNl6Ky6/lBE/igiz4hIYgfy+A8RKRYRh4iMEZFFbcT7V8jxHhH5k4jcaP+eLyJ/E5HHRGRgSLx+pSODwWDoLFw9XYA2+BrwiqougaBhuhy4DIgDFgEtwM+ACuA5O92NIuIB3gT+DXxSVb8SyAOYBawEFgKhD+flwHeBq0PC5qnqZ0VkGvCf9rkA0oAUVb3KzjcJmAT8QEQeB+YBIwC/qv6X3bDYDeQBz9tl+ytQAkyw87jWziMZuBn4IZAOvCci9cAC4CTgO/2qbJePXdeq+giwSVW/ISLfB/JFpBT4CSBALXA/8HugFHjJzmMLcC2w3j63C/g5EAs4gb8AY+0GwT12HcQDZXb6m4AvAWOBG4Cf2uH9TUcGg8HQKfRWoz8OeCrwQ1WbReTLqvp5ERkKfBvwA4tVdTeAiCwAnlLVtSLyFLANy6iE5gHwLtbDOzvkfCeB1SLyyZCwh0TkD8BxICMkn+Mi8m8R+TPQDHwf2ATcpaonReQCoAGYJyJZdrK/APVYhisWeE1VHxGRySHn82IZnXPt34Frec5ufEwHrji9auwQnVHXjwCFIvJ/WIbwf4GvYxnocmAUkIBlEJ8H3sIyuv8Cvgxstk9/ITAM2AEMBaqA7aq6yD7vuViNiJeBJYCoql9E9gGDQq6hv+nIYDAYOoUzHt4PG07N65ziBNkGFIXkHxMiC7gQFCxjFMqJkDh7gOAQdFgevwB+EJb2z1i9RSeAqi5R1W8Cq4APQiOq6uOq+jVgNfDZQDlEZABQqKo/BHZhGTqAOqzecqxd7mY7vMn+vlJV78Dq8QbSBK6lJSxuZ9MZdQ3wvqoGDPgorP/Wy6q6SFWvVtV9wC1YPecfh+TzOzscO81aO831qro/JH9U1a+qPqDRHlb3299DgAOhhetnOjIYDIZO4eP09P8TWGwf3wrc/vGLE+TPwL0ichmWEX4S+IeIPIj1wP0p1oN2kYiUYw3lt0JVK0XkVRF5GGjEMgpv2bI99pBsbEh8r4g8BvwKQESuA2bacQJGCRFJt+OcxBot+G8gBvgl8CCQICL/hWX4IvEqcL+IZGL1agHKReR2YBrwRlj8Z0RkcWhZO5mPXddh/Aa4F2sI/AER+QRW/fwFq1EVh/U6JQlAVVeLyG3AfmAZ8CcR+RWQijXKcERE7gaewPqfAbxu9/AfsvN1A98LFKAf6shgMBg6BTkT3/si8jTWe9RtWD2xbaq6OHoqg8FgMBgMPckZGX0AERliD78iIhmqeqxTS2YwGAwGg6FTifpOX0TGi8gdIvJjEflxmPiWkOPbOr9oBoPBYDAYOpP23unfhvV+tiWCLC3kOCOC3GAwGAwGQy+iPaO/VVW3tiF7SkSewZoV/ZdomYjIKKzZ8s+r6vN22FewGguJqrrYXotdC1So6t9P4xoMHcTowWAwGM5u2jP680VkHtbsd1T1yhDZVuBRVX1RRCZFy0RVd9kOUFJDgiep6kL71UEalqOUX4vIfcApxsb2wnYjwNixY4tee+01nn/+eVwuFw6HA5fLhcvlwul0Br/dbjdutzs0D+bPn09MTEx49r0J6aqMP64ewnWwbds2tmzZwo4dO4iLiwOgoaGBlJQUampqAEhMTKS2tpaEBGuVW3NzM+np6VRXV+NyuUhOTqampoaUlBSam5upq6sjKyuLAwcO4PF4yMnJ4fjx4+Tl5dHc3Ex9fT15eXmcPHmSxMREkpKSqKioIDs7mxMnTtDY2MiwYcMoLS0lMTGR+Ph4jh07Rm5uLpWVlTQ3NwflHo8Ht9sdzN/v95OSktLlejAYDIaeIqrRV9XLRGScfbwtTPxzLMc1LwJfpfU7/tNB2/gOL8tDwEMAU6ZM0YaGhqDhUNVWH5/Ph8PhCE0b/F6/fj233norycnJZ1jcfku7egjXAcCmTZsoKSnB5XIF4iAiwTr3+XxBWQCHw4Hf7w/KQieTiggffGC5RWhubsbtdrfSXyC9z+fjiiuuYMCAAaSkpBATE0NSUhKxsbE4nU5SUlKIjY3F7XYH5R6PB6/XG5THxcUFj91uN35/uCsCg8Fg6F9ENfoicj9wzD7+uqqGGvYqPnJOEhWxfNd/AYgXkRSs9dibROS7AKpaJSJO+/fGjuR54sQJduzYQWZmJg6HAxEJfh85coTs7GxUNfggV1Wam5tJSkrimWeeYcaMGZxzzjkdOVW/oSv0UFBQwIcffsjw4cMD5wjqQUTYvXs3I0aMOMWwOxwO9u3bR25uLs3Nzfh8Pnw+X7AxENBXRsZH00X8fj9er5eamhp8Ph8vvfQSQ4cO5XOf+xwOh4PExI9c/ufm5gaPPR4PAFVVVaxatYo5c+aQm5sbbIyYBqDBYDhbaG9436uqPwUQkV+GyXYBXxZr45X10TJR1cPAt8KC/xYWZ1G7pQ2hpKSE999/n5iYmFa9fL/fT2NjI7Gxp/pJqa+vp7q6mquvvhpVZcuWLcyfP5+MjAxsF739mq7Qw2233cb7778fNPShiAg+nw8Rwe/3R/wEGgfhH1WlpaUFl8tFbGwsMTExxMXFER8fj8PhYMGCBSxcuJDi4mJ+85vfcNlll1FQUHBK+ZqamiguLmbXrl3ExsZSV1fH+vXrqa2tDTYIk5OTyczMxOVycd55553O5RsMBkOfoj2j7xKRO+3j+ECgWE/33Vi+0kVVT3ZR+dpk/vz5rF+/nkGDBpGcnMyJEyeCDYCysjKGDx9OXV0dIkJCQgJ1dXW43W5efvll/vrXv/KZz3yGyZMn88gjj5CQkMDJkyeDowMtLS3k5ORQWVlJQ0MDw4YNo7m5mcLCQmpqaqiurqaoqIgVK1YwdOhQMjIy2LNnD6NGjaK8vJyysjIuuOACiouLSU9PJzk5mdLSUpKTk2lpaaGhoYGioiKKi4vJysoiLS2NpqYmhg0b1t3V+LH56U9/ytKlSxk+fHirYfjA8cGDBxk4cGCrhlnAsB8+fJisrKxWYYFRAKfTSUlJCRkZGTQ1NdHS0hIcDVi+fDk7duxg3759fPrTn6ayspLnnnuOlJQUrrrqKmJjYzlw4ADFxcVUVlaSnp5OQUEBTU1NeL1e0tPTSU9Px+FwEBMTQ11dHYcPH+bAgQPG6BsMhn5Ne+/0b430Tl9VVUSmAqtUtUdehMbGxpKWlsbYsWPx+Xzk5OTgdDoREbxeL/n5+af0PFU1aMT//ve/U1FRwTXXXIPL5aK6upphw4YxcKC1pUBDQwNZWVls3bqV2NhYmpubWb16dXC+QG1tLVVVVUyaNInExESKiopwu91kZ2fT3NyMx+Nh3rx5wXMHJo+FGvZQeWlpaRfWVteRl5dHVlZWmw2WxsZGRo4c2Wq+RUtLCy0tLdTU1DB48GBiYmKCowKBIXyv10tKSgp5eXnBkYKmpiaam5uZOHEiy5cvZ8iQIaxevZqLL76Yb3/727z88sv84Q9/oLm5mcTERDIzM8nPzyc3N5ecnBw8Hg/79u1rs6yvvPJK11WUwWAw9ALaNPr2cP6wj36Khs3ePw/4t4g0YLUDrgzPoytxOBwMGDAAh8NxytByWlpaxOF6VSUxMZGf/exnZGRk8PDDD/PrX/+ar3zlK7hcLg4ePMj27duJjY3F4/EQFxeHw+Fg4MCBNDU1BT8tLS2cOHGC8vJy3nzzzeD5/X4/sbGxHD58mHPOOYfU1NRW54/Wk++LvXyAuLg4UlNTg7Pzw0lNTW31rj2U48ePM2DAAACcTifJycmkpKSQkpJCYmIiL7zwAjNnzqS2tpaTJ09SX1+P3++nvr6eJ554gjfffJNLLrmEjRs3UlRUxGWXXcbRo0d54okn+Pa3vx3xnNHq+ZOf/GSbMoPBYOgPtGn0VfV2EclT1YMAIhLcutQe3n9EVZ/thjJGJD4+ngkTJjBjxoxW4arKW2+9xaRJk2hubqa5uTnYQ2xpaUFE2LZtGxdccAFOp5MlS5Zw77338tnPfpYBAwaQmZmJqlJVVYXD4aCiooL169eTl5dHYWEhGRkZwZUBL730EpMnT6a8vJyamhpUlaamJvbv388bb7zB6NGjGTNmTE9UT7eRnZ1Nfn4+U6dObbVsMkBcXBzTpk2jqamJhoYGGhoaaGxspKGhgfj4eDIyMoJLK10uF263G5/PR319PdnZ2WRmZpKbmxscxQFrlGf48OG88847XHPNNZSVlVFRUUFWVhaZmZlMmzYtYln37dvHxo0bKSwsZOjQoa3KaTAYDGcD7b3Tv4GPdtK7BXsnPXt4f7yIvIU9g19V67uslBFoaGjggw8+oLGx8RTZvn37cDqdxMTEEBMTE5wIFh8fz9tvv82YMWMoKCigvLwch8PB9u3beeSRR4iLi6OoqIj8/HwKCwuZP38+zz77LAUFBVRVVbF582ZUlUGDBjFq1CguvfRSgOArAbCWqB08eJD6+npKSko4fPgws2fPPmXZWn+hpqaGw4cPs337dlpaWvB6va1WTOzbt4+mpiZiY2OJj48PfpKTk/F4POTm5gbTtbS00NjYGDw+ePAgmzZtCg73Bzh06BAFBQVs2bKFTZs2MWPGDJYuXcpFF12Ey+U6pSFYW1vL22+/TW5uLikpKYgIb775Jn6/P9hoCfgZMBgMhv5MtOH9p4Gx9ux8sHbUC2U41lr9AF/t5LJFJT4+nnnz5kUcrg1/dx7KvHnzSElJYcOGDRQUFFBdXU1hYSF5eXkMGTKEjRs3snnzZl555RUWLlxIZmYmRUVFwaWBJ0+eJCYmhtdff52WlhYuuuiiVj4BnE4nCxYsICkpiWXLluHxeHj11VeZMWMG6enpXVQbPUd2djYzZ85ss76j6WLatGmtluSFk5qaGjHtzp07KS4uZsWKFRw6dAiAWbNmsWbNmlbzJHw+Hxs2bKCpqYk5c+YQGxsbLE9g4mFFRUUwjsPhYP78+R29dIPBYOhzRN1lL3R4P4LMDVxk/1ymqt5I8bqCKVOm6IYNGz52PkeOHOHdd99lx44d1NfXB9eTx8fHc+LECQ4ePMjy5cspLy/nqaeeoqqqimnTprFhwwZSUlKorKzE5/MxZsyY4Dr1UHw+H0uXLg16fHM4HK2MUhv0ibWDnaWDM6G2tpbf//73/OY3v2HhwoV85jOfYfTo0WzdupW4uDjy8/MpKSlh586dTJkyJWrDIpQVK1ZwwQUXBH72CT0YDAbD6dDmLnv2RL77ROSfIvK0iPwzLMqfgHT782AXlrHLyM7O5tJLLw2u766srKS+vp7GxkZSUlKYM2cOCxYsYOjQoVx11VUMHTqUNWvWMG3aNPLy8igrKyM7O5u6ujqWLVtGdXV1q/ydTief+tSnKCoqYteuXWzdupVVq1bh8/l64nL7DR6PB4/Hw6hRo9i0aRP79u0DYPz48ZSUlPD000/T0NDAxRdf3GGDDzBy5MiuKrLBYDD0Cto0+qp6u6peaX+uiDA7v0ZV/25vylIV7SQiMllEfiYi94pIoh32DRFZKCIv27+fEpHviMjMj3tRp8vll1+Ox+PhvPPOIykpiWPHjrF//3527NgBwMUXX8z48eNZsGABo0aNYu3atfj9fsaNG0deXh7l5eWkpqayefNm1qxZw4cfftgq/7Fjx3LNNddQU1NDeXk5L7/8MidOdMiZYafR23VwuiQmJlJQUMC6devweDxUVlYClv+G9PR0xo4de9oOl/rqCgqDwWDoKG0afQAR+bWIPCYij4vI42HiKhF5WEQeBmraOc+XgEXA81gOfVDVPwLbgUftOIextuuN+L5BRG4UkQ0isuHo0aPtnO70EBFmzZrF1KlTufbaa1m4cCFXXHFFcPOV+Ph4Zs2axYIFC7jgggtITk5mz549wY1eLrzwQkaOHBl0J7tkyRJaWlrvRuzxeDj//POZOnUqFRUVPP7442zatKlTr6MderUOTpekpCQGDx7MsWPHqKmpYcuWLQC4XC7TYzcYDIY2iGr0gcOqeo2qXq2qVwcCRSQTuAv4DrAQ+GMHzhVpI5fPAM+B5QhIVe8Ero2YWPUhVZ2iqlMyMzM7cLrTI3zG95AhQ7jmmmvweDzU19ejqkyfPp3LLruMa6+9lvLyclasWEFdXR0AGRkZLFiwgAkTJuBwOFi+fPkpw/gzZsygoKCAG264gZkzZ7Ju3Tr+8pe/cODAge5yztOrdXA6DB48mLi4OEaOHMmyZcuCvvrhzHvsJSUlnVhCg8Fg6H20t45stj1EWgegqn+ww3+uqjdi7buOiPwJuClKPk9i9TITgBIRcQBDgLLABEAR+SEQC7x/RlfSRcyaNYv9+/dTUlJCY2MjhYWFxMTE8KMf/YihQ4eycuVKPvWpTwVn8Kenp5Obm8vUqVNZsWIFF154YUSf9JMmTWLcuHGsWLGC5557jqSkJK6//vquvJQ+q4NIjBw5kjVr1pCfn88777zDHXfcwdatW5k8eXLUdB9++CHJyclUVVVx/PjxVks+jxw5EnFCpsFgMPQX2jP692L1CoXWvcPwxfEtREFVi4HisOBSrNGCQJyf0wuZMWMG06dP57777qO+vp7Bgwdzzjnn4HA4uPfee7nmmmtYuXIlCxYsCKaZNm0aFRUVFBYW8vrrrzNv3ryI75fdbjeXXHIJ06ZN48knn+zS6+jLOojEgAEDiI2NJT8/n2XLllFbW0tlZWXQj384lZWVbNy4kUOHDlFUVERaWhpDhgwhPj64pUSfdYVsMBgMHaW94f1L7c+ngHkikm+HJ4lIOoD9ndR1Rex5RITrr78er9dLSUkJ+fn5nHPOObhcLh5++GEOHDjAu+++G4yfl5fHwYMHg45f1q1bFzX/9PT0dnuohtaICLGxsaSkpJCZmcm///1vBg8eTFlZWat4VVVVLF++nD179jB//nzmzp3L+PHjycvLa2XwwUzkMxgM/Z/2jL4D+CfwFDAA+L4dvgj4tYg8Cdxj/+7XpKWlMX78eFpaWti5cycjRozgm9/8JgMHDuS5555j06ZNwdn+IoLH4wluKJOdnU17a9rD5xQY2ic2NpbExERGjhzJ2rVrGTVqFLt27QLgxIkTrFixgp07dzJ37lymTZuGy+Uyht1gMJzVtGf0k1R1gz00HAscAlDV/ap6vapepapfVdV9XV7SXsBll12G0+nE6/Xy4Ycfkpuby5e//GWamppYt24da9asYefOnZSWljJhwgS2bt0KQH5+PvHx8axcubKHr6B/kZ2djdPppKCggPXr11NfX09CQgLPPPMM27ZtY/bs2cyYMQO3293TRTUYDIZeQXtGf4XtmOdpYCXQYxvs9AacTmdwI5+0tDR8Ph+VlZXcfffd7Nixg+3bt/POO++wePFiEhMTqaurC+4rP27cOCoqKoJLywwfn5EjR+L3+xk4cCB1dXWsXLmSadOmkZKSwqxZs4iNje3pIhoMBkOvoj2j/yzWGu5HgWdV9b2uLlBvZ/78+SQkJFBSUoLP56OgoID333+fJ598kmXLlrFz587gjPChQ4eyf//+YNoZM2agqmzevLmnit+vCPjPj4+PJz8/n1dffRWXy0V+fn77iQ0Gg+EspD2j/yBtuNoVkR/YTnueiOC4p9/idDoZOnRo0OC89957jBo1ivXr17NkyRL+8Y9/sH//fmpqasjPz2/lnW/YsGFMnDgREeH993v9qrheT2AnxdTUVEaMGMH69evx+/3mvb3BYDC0QXtGP5qrXbGd9nwp1HHP2cDs2bOpq6sjNjYWEaG8vBy3201paSkvvvgi69ev580338ThcOB2u2lqamqVfsKECTidzu72yNcvcbvdxMbGMnz4cLZt28bOnTt7ukgGg8HQa2nP6EdztTtSRK4QkU+KyCejZdKG3/c/237eL7Z/LxKR74rIdWd8Nd2Ex+MhMzOTuLg4XC4Xe/fuZcyYMZSVlZGUlISI8Oc//xlVZdy4ccEJfaGMHz8et9vNe+91zxuT/qaDACkpKTQ3N5OQkEBBQQHPP/981PiLFi1qU2bW6RsMhv5Oe0b/51iudr8D/CxM9gYQD2TYn2ic4vcdy8+7B3CKSBrgV9VfAxEXrPcmv+8AY8aMCRr4oUOHsnTpUiZNmsRbb72F0+lk5cqVbNy4kQEDBnD8+PGIeYwbN47Y2FiKi4u7w+D0Ox2A9crE6/WSlJTEkCFDWLduXUTD7vf7qaqq4tFHH2XHjh1s3LiRdevW8frrrwc/99xzT/dfgMFgMHQj7Xnk+5Wq3gYgIncDd4TI1mC53lXgoQ6cq5Xfd1X9kZ3v74G3wuWnJFZ9KHCeKVOmRIzTnQwfPpyNGzcyYcIEVq1aRWpqKsXFxeTk5PC1r32No0ePcuedd/LSSy+RmZlJRUUFWVlZp+QzduxYdu/ezZ49e7rjXXS/0gFAQUEBb7/9Njk5OQwbNox//vOf7N+/n4suuqjVaxWHw0FKSgoOh4O0tDQGDhxIfHw8brc76MFv/vz5PPDAAz11KQaDwdDltGf0E0KOw73u3Qn80D7+GfCVKPlE8vt+M9YEwf2qWiUiThH5LrCxg2XvUQI9/KysLLKzs6mpqaG+vp6mpiZycnK46aabmDZtGjt27GDs2LGsW7cuotEHy3AdPHiwq4vc73QAkJqait/vJyUlJeia9+DBg0yePJm4uDhaWlqora0NfkpKSvD7/SQmJuJytff3NxgMhv5Fe0+9tSLyLOAHXgyTNalqGYCINETLpA2/778Li7Oo3dL2MiZOnMjbb7/N5ZdfzsMPP4zT6aS5uRm3282hQ4e4/PLLue2221iyZAl+vx+fz4fT6YyYV1f38vurDkQEt9tNS0sLLpeLiRMn8tJLLzFkyBBqa2tpaWnB4/EEPwCTJ0+msrKS1NRUsrOzyc7OJicnB7BeAwQ2TzIYDIb+RntG/1ngA1V9V0TywmQvisi/sIaC/9ElpevlxMbGBo353LlzWbNmDYcOHSIhIYHdu3fz1a9+lSuuuILi4mJGjRrFmjVrmDt3bsS8zDKzMycuLo7m5mY8Hg/nnXceL730Elu2bMHj8RAfHx8cvvd6vYwbN47HHnuMmpoaqqurOX78OFVVVVRVVTFw4EBj8A0GQ7+mPaN/P9AEvAv8AGs4GBGJBZYDK+x4veL9bk8wfvx4tmzZwuTJk9m9ezdNTU0cPXqUnJwcYmJi+NznPsftt9/O8uXLWbp0KZ/4xCeMYelkBg4cyLFjxxg6dGjQ42F2dnZQrqps376dgwcP0tDQwOjRo0lMTDxFD7/4xS+6tdwGg8HQ3bRnfWqBI/Zx6Ha6PwAewBoefsD+nJVkZmZy7NgxVJWpU6cyduxYVJX09HS2b9/Opz/9aXbt2sWqVavIzc3ljTfeCLrmNXQO+fn51NXVkZSUhKoyefJkqqur8fl8HDhwgKVLl5KUlMRFF11EWVkZHo/HNLwMBsNZSXs9/aPAHBG5B+u9PgCqeqeIfEpVXwYQkcu7sIy9nsGDB3PgwAEGDx7M1q1bcbvdHD9+PPie+eqrr+aOO+5g5syZLFy4kPXr1zN9+vSeLna/IS8vj6amJkQEl8vFzJkzeeGFF6iqqiItLY1Bgwaxd+9eSkpKuOWWW6isrCQ9PT047G8wGAxnC1GNvqreZU/kcwBBf7Ii8ingSyIS6LJ+CXihy0rZyxk9ejQrV65k8ODBTJ8+nZ07d3LixAmysrIoKytj1qxZvPLKK/z2t7/l/vvvp66ujq1btzJ+/PieLnq/ICEhITh6kpKSQktLC3V1dQwcOJCWlhb27dsX3O544MCBvP3227hcLtxuN4mJiQwZMoTs7GxuvfXWHr4Sg8Fg6FraNPoiMgGYh7XU61NYjno+a4szgGYgE2sE4K4uLWUvx+FwkJCQQG1tLR6PJ7jlq9/vp7GxEYfDwXXXXcf3vvc9duzYwbhx41i/fj0lJSUMHz68p4vfL3C73TidToqKili1ahWf//znqa2tpaqqipMnT6Kqwdn8Bw4coLGxEZfLRWxsLLt27cLr9TJy5MievgyDwWDoUqL19BcD/wssA36kqgGDj6r+TURyVfVvXV3AvkJhYSHvvfces2fPZvbs2ezZs4e1a9fi8XjYv38/EyZMwOFwMGfOHL7whS+waNEiduzYQUJCAtnZ2SxatCiqi1hDdFJSUmhoaCAtLY2srCy2b9+Ow+GgsLCQ1NTUYLza2lquvfZafD4fhw8fZs+ePZSXlwdHBwwGg6E/E202U5mqvg2sC7y7D2O2iHxfRL4pIt+MdpI2/L5/U0R+JCI/sX8/ZfuBn3nGV9ODJCYm0tjYiM/nIysri9raWhITE4PhIsLYsWP54IMPiI+PZ9y4caxevZq3336bEydO8JOf/KRLy9ffdTBkyBAqKiqorKzk/PPPZ/78+UyfPp1du3bx2muvUVJSgqoye/ZswNqhLy8vjzlz5vDFL36R66+/nri4uB6+CoPBYOhaohn9fBH5JzBCRJ62j0P5FbAW2Apsa+c8p/h9V9U/YI0k5NhxDgNptLH8rzf6fQ9nzJgxfPDBB4C17n7atGkcPXoUt9tNSUkJgwcPJiUlhXvvvZd33nmHrVu3cvPNN3PnnXd2h3e4fq2D4cOHB4fxAz4P4uLimDZtGgsWLMDr9fLaa6QKJH8AACAASURBVK+xY8eOiHshiAgzZszo5lIbDAZD99Km0VfVT6rqlfb3Fap6ZViURuA/gRuw1vK3Ryu/7iISB/wCez6Aqt6qqncC17ZRnodUdYqqTsnMzOzA6bqfvLy8oDvdgoICjh8/TkJCApmZmVRXVzNo0CCWL1+Oz+dj5MiRPPXUUzzzzDNs2LABr9fbHUXstzrIysqipaUFp9OJz+drJRMRCgoKuOiii3A4HOzbt48VK1awatUq9uzZE6x7Y/QNBkN/5+N0L7+BZfQF+APwdpS4kfy+P4U1QnAR8GcR+SEQC7z/McrUo4gIWVlZvPfee5x77rmkpqZSWFjIu+++i9PpZMyYMUyfPp3ly5dz4YUX4nA4mD59OqtXr+6OdeP9WgexsbGoKllZWbz22mssWLAg4ujJ6NGjgyMBPp+PsrIy1q5di9frxel0Mm/evO4tuMFgMHQjH3dMOdBzjLrguQ2/75eHxfn5xyxLr2D8+PE8+uijTJo0iYkTJ7Ju3Tri4uLIzMxk586dlJaWMmXKFJYvX86CBQtwOBzdsl78bNBBfHw8LpeLlpYW3nzzTRwOB+PGjSN0VCLU3bHT6WTYsGHBsOXLl3dziQ0Gg6F7ibZk72kgEYgBfICq6iUhUf4I/BXL4J+1HvnCcblcFBYWsnHjRoqKihAR8vPzg+/6X375ZQYMGMD555/PypUrueCCCxAR7rzzzh4ued8nNTWVvXv3MmHCBIYNG0ZzczPbtm1j06ZNZGRkMG7cOGJiYtpMn5+f342lNRgMhu6nTaOvqleIyGLgJ1hr8b8bFmU/1tCwAvu6rIR9kKlTp7JmzRoqKyspLCxk27ZtJCQkkJGRwbXXXssjjzzCs88+S2JiIiUlJcyePZvrrruup4vd5xk2bBjbt2/n4osvBiAmJoZzzz0XgKNHj7Ju3TqOHz9OVlZW8JOSkhIcaTGbHhkMhv5Oey+SRwF5WLO7w73ILAKWACuB+zq9ZH2cmTNn8s477wTXj+fm5lJXV8eLL77IwIEDSU5ORlU5fvw4r7zyCsuWLevpIvd5Bg8eTH19fURZZmYm8+bNY/LkyUydOpWEhAT27t3L66+/zsqVK1m5ciWrVq3q5hIbDAZD99LeO/3FwC328e/CZNtU9X0AEdnd2QXr6zidTqZMmcL69espKCigvLw8OIO8srKSAwcOsG3bNtLT0zl06BDHjh3r6SL3eTIzM2lpaUFV25wnEejNDxo0iEGDBrWSrVy5squLaDAYDD1Kez39T6vqf9mfD8JkV4rIqyKyEmtTnvB1/Gc9WVlZiAhxcXHU1tbi9Xr58MMPycnJYdasWXzxi18kIyODm2++mbS0tJ4ubp/H7Xbjcrl47733zij9iBEjOrlEBoPB0Ltor6c/W0S+D5yAoDMX7OPZIuKxj2u7roh9m6lTp7J06VKys7PJzMwkNzeXt956C6/Xy4gRI5gxYwZr1qyhqKiop4vaL/B4PGzevJmqqiqSk5MZM2YMHo+nQ2nNO32DwdDfac/o/4o2luWJyP8AQ+3jMlVd3PnF6/s4HA5mzJjB5s2bOXz4MCNHjmTu3Ln4/X727t1LcXExDoeDQ4cO9XRR+wWpqalUVlaSmZlJfHw8W7Zsob6+nqSkJMaMGRP0w6+qeL1empqaaG5upqmpCafTSVZWVg9fgcFgMHQd7Rn9RuBmLMP/xzBZuqp+DUBEftUFZes3pKenk5ycjIiwc+dOamtrcTgcDBo0iLlz5+Lz+c54SNrQmpEjR5KSkoLH4+HkyZNBpzsnTpxgyZIl1NXVBbfiFRGcTicOhwOn04nT6eTyyy9v/yQGg8HQR2nP6EfzupcqIl/BahAMiJaJiEwGPoflDe5HqlonIrdhLQVUrEmCvwSqgA2q2u+msp977rm89NJL+P1+RASfz8fWrVtZu3YtPp+PK68M93LcuZwtOhg1ahQxMTFUV1dTV1cXrO+YmBiysrKorKwMLtML//j9/p4uvsFgMHQpHfHIp/YnfDr0L7CW9AlwUzt5fAn4PjALa7OX54HBqrpQRO4DCoHNqvp/9u9TDI6I3AjcaP88KSI7gQwg0rT3tsKjyXo0rxtuuCEgWxrmBKmz6CodwJnVUXuyrsq3o7Ku0oPBYDD0GO0Z/T8Cj9jHvw+TfVZVf3ka59LT/G6dWPUh4KHQMBHZoKpTwuO2FX4maXo6r06m03UAZ35d7V1zV+TbC3RgMBgMPUY0N7y/xnK88402Zud/WUQuAGqwXPRGG5+OtNlLmYh8ByjF2uDlanulwKtnciGGdjE6MBgMhrOcaD393wLnA/eLSAqwU1V/ECKfo6qnbkwegTY2ewn34vffHcnLcGYYHRgMBoMhmtEvA3YA2fYnPiAQkUeBRhEpUdX/7dISRueUoeZ2ws8kTU/n1Rc40+tq75q7It/+qgODwWBoF1GN+PoWESkHngbuVtVDYbI/qepNIvI7Vf12N5TTYDAYDAbDxySa0U8G5gIzsXr5lar6M1u2EViONfy/EkBVb++OAhsMBoPBYDgzog3vO+xPPDAMawJYgM/a3+Ez+g0Gg8FgMPRSovX0HwVeB1ap6r5uLJPBYDAYDIYuoE2j35sRkWRVrRGRFFU9ERoOoKo1UdKKRrjoSOFtxbVlA1S18syvou/Tlh7C4gxW1bJOOl8B0AycA+xX1e0R4nxCVVd3xvkMBoOhv9ERj3y9CtuwXy0ijwNfx3Idi4j8yI6itq3+eViaAN8A/tcOvx9IwlqeONf+DuwlEFjPPlhVbw3J61778FwReU9Vb7PDb+Ujz4Wqqr8NSfMlVX1CRK5U1X+2F78v0JYebNmnA4fApcDX2shjnKpuCwsbCeSq6poI8m8BecBiLC+Qt4SkWwu8A0wWkY0BvdiyaLrp03owGAyG06HPGX1gIZAPZAE7Q8LrVPVeANunfCgvAKuwHuzTsY0+cFhVfyEii2i9f8ARIEFV7xGR8LXrJViTGC9S1ftDwp+3vyO5LB5ofw9pI35fpC09gDXn4xGsemgMFYhIwFALMAf4fFjaO4A19l4BAKFGfy+We9zNInIgQnkKgX1heoHouunrejAYDIYO0+eMvqr+RETuxHp4jw4RHbO9CAJsCkt2l6q+BiAioRvXB+Yq/AS7l2/zPpZhB6gOO//vRORLQEFY+D4R+RYwHtiMtUFRgBJ7hGB1WPxvAkWAE/iPNi+6FxJFD2DV924AEfkgTDYcyymQAOkRsv5QVf8mIhcCP6a1Xp4BKuzjN8LKs95eZnpphLK2qZu+rgeDwWA4Hfqc0bf5DdYD+rpAgL1RzNzAz9DIqvqaiMwJz0RVHw8Jfzo8vojAqb1YgEPA0yIyR1XfbJ2l3mQbGCC4SU156HBzCAlYDYGmKNfamzlFDwCquttuEAT0sDhEfBdwQlWbReTuCHm+LCIxtg5OhuV7wG5wPaGq70RIO1tVw7eADkneWjch9HU9GAwGQ4dw9HQBzpC5WEPD4T3MVKxh+msjpEnD2kUtXHa64dHOM0xEFtN6GD8WuFlE7g2ZDxCgDHADk+mbtKUHsBoEDwCtJvip6lHsIX1VbQhPpKpbQ+RvRcj3KyLy+wh12Z4skm4C9HU9GAwGQ4foqz39E1j7wN8ZFr4KqyEzNkKatmSnG96mTFX/W0SyVLUiJOx3IlIC1HLqe+N/Yb3v75N719O2HsBqEDiI3CD4iojMBpraGAGJJr8K6xXB/gjp2pRF0k0IfV0PBoPB0CH6qtGfDozAerjfFRK+EMsIReohtiU73fA2ZfZwdYuIuFX1jhCRE2tk4BIg9HXAg8AWrMZAX5w13pYeIHqDIJrhbk++CDiI1Yj6U0dlUXQDfV8PBoPB0CH66vD+flW9CQju8mcvvarG2uq3VQ+8LdnphrcnAw6q6o+wDE8oq7DcFYcbsTdV9f4+vEzsFD0AiMh9wGeAzwHfj5BuEbAAuKKNfKPJj9ifSJMAo8na0g30fT0YDAZDh+irPf1vicjFwCgRGWUPAT/PR8PnkZZlRZKdbnh7MhGRPwFbw8LbGjWYaa8m8LUxzN3biaQHVHWhiLiwdmeMNJweMM6D2sg3mvwDrMbEc6cpa0s30Pf1YDAYDB2iTxp9VZ0dIWyfiNwOJGMty9rXnux0w9uTAS8CDXy03C90ZACskYHldvgUrBGAJvroGvFIegjhV1jXXQGEz6iPZpzbk48FvtvGu/loslN0A/1DDwaDwdBR+qTRF5EHsNyxEtYz8wK7gJwIydqSnW54NNlXsQzVrcB37LC2RgZuA7ZjvSbok0TRA1jD6BVYznvCiWac25M/CNwgInnA/aq6p4OySLqBfqAHg8Fg6Ch90ugDa1X1iQjhs4AYIg/htiU73fBosqNAJXAo4Jc+ysjAb7EmrHWKX/oeoi09gDWR77PAkgiyaMa5PfkCrIbEMuBG4HsdlJ2iGzu8P+jBYDAYOkSf23BHRAYBD2MNx3pV9XMhsluBx7GWetWEpYsoO93wdtIEHNIE/LgvtsNvA44Bqf1lslg0Pdjyear6ehtprwKmYk1unKOq3+uoXESmqOoG+zg5rP6jySLqxmAwGM4m+mJPfzCWwY1EKtaGOtDaC1w02emGR5MF3t23YBmsANFGDfoq0fQA1mY8mVgNo3+HyT5U1ScBRCTSjnjR5Bfb+yE4VfULpyFrSzcGg8Fw1tDnevq9GRH5MdaEscsBl6r+jx3e5qhBfyXEJTKq+oaIDFLVA7bsh8BEIhvnqHIR+Q7Wq4PG8FcL7cgi6sZgMBjOJvpiT783kwRUYfly3xUSHm3UoF+iqm+EBZ0P/J99XAcsJWwHvhAiykWkEIgH4oAZwBMdkdm0pRuDwWA4azBGv3P5O5ZDmX8AuwOBqvqTHitR70GgfePcjnwS1mZHcOpyvmgyaEM3BoPBcDZhjH7n8jWgGNipqmbHttYEvPa1Z5zblKvq39rKPJrMxujGYDCc9Zh3+p2MiMwE/qiqk3q6LL0BEblYVXvFRjZGNwaD4WzH9PQ7ERH5OZbDnqt7uiw9iYj8FWu2vGD13Hvc6BvdGAwGQ9/dcKe3kgnEAj/o6YL0MDtU9TZVXQj8tacLY2N0YzAYznpMT79z2Yy1DvxETxekJ1HVX4lImqpWqerfe7o8NkY3BoPhrMcY/c7lTaylYYU9XZBewC0iEg88q6rre7owGN0YDAaDmchn6DpEZD5wt6pO7+myGAwGg8G80zd0ESJyH5AOnNfTZTEYDAaDhenpG7oEERkLFGH9x/6vvfgGg8Fg6HpMT9/QVdyANXnu/Z4uiMFgMBgsjNE3dBW1gBNrO1uDwWAw9ALM7H1DV1ECjLePN/dkQQwGg8FgYXr6hq6iGZgFGD/3BoPB0EswRt/QVeSo6k3AwJ4uiMFgMBgsjNE3dBVDReSrwJCeLojBYDAYLMw7fUNX8QGwBYjv6YIYDAaDwcL09A1dRYaqvovloMdgMBgMvQBj9A1dxesicg+wsqcLYjAYDAYL45HPYDAYDIazBNPTNxgMBoPhLMEYfYPBYDAYzhL6rNEXkRgR+Z2IPCAiD4rI3A6kWSQi48PCPhSRP4rIMyKS2IE8/kNEikXEISJjRGRRG/H+FXK8R0T+JCI32r/ni8jfROQxEemz69iNDnoHRg+9A6OHnsfooH368pK9rwGvqOoSCCr7cuAyIA5YBLQAPwMqgOfsdDeKiAd4U1UfATap6jdE5PtAvoiUAj8BBMt//P3A74FS4CU7jy3AtcB6+9wu4OdALJa/+b8AY23F3wOcxFq6Vmanvwn4EjAWa2Oan3ZetXQrRge9A6OH3oHRQ89jdNAOfdnojwOeCvxQ1WYR+bKqfl5EhgLfBvzAYlXdDSAiC4CnVHWtiDwFPAIUisj/YS0t+1/g61iKKAdGAQmAD3geeAsYAfwL+DIf+ZS/EBgG7ACGAlXAdlVdZJ/3XKw/y8vAEqwJlH4R2QcM6vyq6TaMDnoHRg+9A6OHnsfooB36stHfhrVf+zKwWnQhssCSBMFScCgnwuK8r6pfFpG7sJTpAF5W1X8HEojILcDnsJS4zw7+HXALsN9Os1ZVfxuSJrgsQlX9dlijiDgAv/09BDhw+pfeazA66B0YPfQOjB56HqODdujLRv/PwL0ichnW0MmTwD9E5EGsVthPsYZxFolIOfDvNnOy+A1wL3Az8ICIfAKIwRqSuQFraGg5kASgqqtF5DYs5S4D/iQivwJSsVqTR0TkbuAJ4Fb7HK/bLbmH7HzdwPc+dk30HEYHvQOjh96B0UPPY3TQDmadvsFgMBgMZwl9dva+wWAwGAyG08MY/Q4iIv8t1lKQEVHiSHeW6WzE6KHnMTr4iNO9TglZstVTiMgnROQpewJbW3H6lG0weug4ffmdfruIyBVAPZAFTMFarnEH8FfgGiAb6x3KKOC4qr4oIo9gvXu5FcgEPHa6a4BXgCYR+TtQA2xW1QdFZBvwd+BZ+73Mej6a6TkduENVd3THNfdGjB56HqODzkNEhgGPAS8CT4i1znoAsFRVnxeRLcCjWBPKbgCmAjcCgdniMcCDWJPHjqvqYhF5A3gTGA28bqdZrqqP2Wkygf8C/og1oWyciPzDzvcHWO+M3wdeBb6lqv8tIt/EmjmeBcwEkoHfYumzGTgkIr8FvIBbVb8tIi8B64B3ReR/opWppzF6ODM99KnW3BnwOjAHa91jIzAPq/KasCZgHAGuA54BPivWOk0vMNhOV42llFxgE3AX8Fngr6p6MxBooR1S1bvtuLtU9XasJRcPYC33uKSLr7O38zpGDz3N6xgddCbb7etswFqHfQSrMQRwQFV/jbWUaxLwDeA/gT/Z8ouw1oN/BxgqIslAk6r+CFiLNXP7q8DlgZOp6lGshtn5wFoRmWSfU7E6b8eBK1V1HzDE7iF+AngD+BaW/o4A07DWlT+O9fyvUtXbgGNiOahxAL9U1dfaK1MvwejhNOnXRj9EQWBV1pewHnS3Ys3KfBBIUNWTWEr7CvAsVr1sU9VFqvoNVV0fkq3w0bKOACdCjmvs7yZVrcF6+MV23lX1PYweeh6jg04ncJ3XYc0AvwtrJASgzv5uwbpeP1YDqskOj1RvwboCatSaYR3+fD4BzMAyWouw9PdJLMP3Yz4auV2JZRhK7WVhDbb+vq+qD4XkF1qOwHe9qnpPo0w9jdHDadKvh/dtaoE9wNvAdapaJSKrsJZEHAmJ9wxWb2SUqnpFxC8i92INTd4VEu8J4D4R+RywqluuoH9g9NDzGB10PuuwPKmdh9WoicSDwC+Bw/bvV7GWck0AylS1poOvpFdjrQvfZJ/veiANuFtEcrGWqAH8E/gQazQHrCVrD2H1hl8OZKaqW0Tk6/aSsnhV3dqHp2IYPXQQs2TPYDAYDIazhN42VGMwGAwGg6GLMEbfYDAYDIazBGP0DQaDwWA4S+jTRl9EXuym80SdVSEi94tIwhnkO0JEHpbWeywXicjXz6Sc3Uk/qPtEsfau/rOIXBMmmyciq8Xa63peb9BJd9W3fa6eqPNWskhxReQz9u8XROSi3qCXcPrBfdEnnknmfjjz+6HPGn2xHDOUfoz0g0RklYgsFMsrUryI3CsivxeRn4lIjoisE5E7gBwReUNEbrcr/usi8i8RGS8i6YCqar2IpNmG4l4ROb+9MqjqXlW9ISysGGtdZ6+lP9Q91uzbf6nq14BPh8kUa6/rOKy1vj2qk49b33Yevb3Ow2WnxFXV5+3f/wF8saf1Ek5/uC/6wjPJ3A8f737oy0v2JgMbAz9E5FNYnsb2qOo/7LAJwC/C0n1VVSuAQuB5Vb1fRB7D8jz2uKpuEKuVey7wpKr+VkT+H/CsHfd5rJ2cqrH2SM4Cttt5J2N5NLtfVVfaZcjHcsoQymJVPR7l2hpFJFtVj0SJ05P0h7ofBGyxw3xhcVar6hsiko21w9Y19KxOWtV3JNqpb+j9dR4uixb3f4Df28e96V7pD/dFW/Taeo6EuR/a1tMZG30RGaiqh+zjPFU9eKZ5nSFFwFOBH6r6slguFL8ZErYFuLSN9IVYzkfA6tmNA+4XyzVjPZYHp+dt+STgORFxA5VqbYM4Hkv507D+AGA5PLlUVetDziOcfj1XYf2BesMNFon+UPcHsG6kTYSNeNmONMDSQ2zIcU/ppFV9i8hQrD28fcAqVS1tp76hl9d5BNkpcUVEgLuBJaoaeOj3pnulP9wXbdFr69ncD6d3P3ycnv5/Aovt41uB2z9GXmfCZGCAiHiBp7EcJvw3cF8H0+cDu0QkA8tZwyrgISyF/8LOa2doXGAilg9lgGGqul9EUvnItagDay/nGuD7qupT1d2c2tIDQEQGAD8HzhWR76tqoGWah7Ufc2+lz9c91g3/gN0bexFARGZitfAPAxdj+dF+wI7fkzoJr+9YLFe6bmAMHRvq7O11/rcw2fPhcbF6YwuAFBHJV9U/0bvulT5/X/SRZ5K5HyzO6H44I+c8IvI0lg/vbVitpG2qujh6qq5FRH6EtSnIUlV9pRvPK8AfVfWmTsovCbins/LrDvpL3Uc5T6/SiYgMwbrZHcAKVS3pxnN3S513sCy9Si/h9Jf7og/Us7kf6Lieztgjn4gMUdX99nGGqh47o4z6ASLyCaA4bFjnTPMaBfhUdc/HL1n/pzPrPso5jE5C6I4672A5jF7awDyTuo++dj9ENfr2e4tLgRiA0N68iNyjqv9lH9+lqj/ojIIbDAaDwWDoGtp7p38b1uzllgiytJDjjE4rkcFgMBgMhi6hPaO/VVW3tiF7SkSewdqu8C/RMrGHHX6AtUTieTvsK1iNhURVXSwii7B2AatQ1b+fxjUYOojRQ89jdNA7MHroHRg9dD/tGf35IjIPa2YkqnpliGwr8Kiqvigik6Jloqq7RORRrNnQASap6kIR+bGIpAF+Vf21iNwHnKJYEbkRuBFg7NixRU8//TRPP/00Dsep/oVUFVVFojhSUtWIaQH8fn8rmYgE8/L7/SxYsIC8vDyysrKIi4uLdulnQpftqfhx9RCug23btvHggw9y+PDhDtdlR2SBV06hOgrUf+A7NjaWW265hZiYmA5e/WnTJXrointh27ZtLF5svXlzOBzB+hOR4H3g8/lwOBytwuzy4Pf7cTqdp6QB8Hq9uFyu4D3lcDjw+/2ISPBeGD16NDU1NTQ1NTFs2DAOHTpEYmIisbGxVFdXk5GRQXV1Nc3NzeTl5XHgwAGSkpJwuVxUV1eTk5PDsWPH8Pl85Obm4nK5cLvdwcvtZBUErrvT9fDYY4/xwgsv4HA4WtVTaD0H/s/hcr/fH6xfEcHlcrXSS1NTE6mpqTQ1NeF2u3G73TQ3N5OUlERjYyPjx49nypQpFBcXk56eTnJyMqWlpSQnJ+P1eqmvr6eoqIji4mKysrKIjY2lrKyMtLQ06urq8Hq9TJw4kU2bNpGbm0tycnJQn12lA7teuuyZ5HQ6g/9bp9OJ3+8P1nXgmR7+vw7oJvQeCOijubmZmJgYXC4XXq+XmJgY/H4/fr8ft9tNS0sLl1xySbDuzznnHPbt20d9fT1ZWVlUVFSQmZlJXFwcZWVljB8/nt27d3P06FHmzJkTrHsR4dChQxQWFrJ3716KioqCl9sZdR7V6KvqZSIyzj7eFib+OXAca/nAV4FbzrAM2sZ3eFkewlpSwZQpU3TAgAGICNnZ2acYBIDy8nIGDhwY+YSqHDp0iOzsbOCjGy7wOX78OAMGDAjG9fv9+Hw+fD4fJ0+e5LXXXiM9PT3453G73SQmJpKbm8uCBQvOsBp6nHb1EK4DgGnTpvHiiy+Sn58PwMmTJ8nIyKCyshIRoaKiggEDBpCYmIjf76ehoYGMjAyOHTvGkSNHOOecc6iqqsLj8eD1emlsbCQ7O5uKigoOHTrEOeecw/Hjx0lISKC+vp6mpibi4+M5ePAgd999Nx6Ph+TkZKZMmUJzczN1dXXBGyz04TdmzBj279/PoUOHuOCCCyguLj7lBty/fz+TJ0/u+ppumzO6F8B6UOXk5BAbG0tTUxNA8EHlcDiC//eYmBiam63txgMPqiNHjpCdnY3f7w/KRQSn00l5eTk5OTmtjFZLSws+n4/KykpWrFhBU1MTEydODJ4r9GEb2gBxOp3B+8XpdOJwOHC73cGHasDg9QLOSA+ZmZk4HA5yc3MD8lbfR44cIScnBxEJXmtoo3bfvn0MGjQo+KwJ/TQ2NuL1eoPGK6DjEydOUF9fT2NjI0lJScybNy9YvmHDhlFaWhow3ACt5CNHjqS0tJTCwsKI8tLS0tOps67gjJ5JI0aMoKKigoyMyG+djx49Sk5OTvA/GPh2OBzs3buXQYMGBf/vXq8Xr9eLz+cLGn2fz4eIBO8jEaGhoYGWlhY2bNjA1VdfHazzgI0pLS09pe4BMjIyKC0tJTU1tZV81KhRAEE71JlENfoicj9wzD7+uqqGGvYq4ERHTiIiOcAXgHgRSQGWAZtE5LsAqlolIk77d1RPSwE2bNjAww8/THZ2NnFxcTQ0NOByWZfj9XopLy8PVrjb7aahoSF4w3i9Xo4dO0ag4eByuWhsbAy24Kqrqxk6dCjNzc3ExcWRlJREc3MzaWlp1NbWcuGFF1JdXc348eMZOHAgTqeTXbt28dZbb+HxeBARxo0bx5YtW0hNTSUpKYn9+/eTkJCAz+ejvr6eSZMmBQ1Penp6aKu6y+gKPTzxxBO8+OKL5OXlBRtPgdavqnLkyJFgHYV/du/ezejRo1sZh1B2797N2LFjgzdm4OYUEWJiYrj++utZvXo1x48fZ8OGDbjdbgYMGMDevXu59NKP/HIE6jUnJ4fS0lI8Hk/EG/DkyZNnUq2nRVfooKamhscff5yRI0eeooPA70AjONDDCTWuZWVlDB8+vNXDL/DZuXMnY8eODRqowCeQ7qqr/j97bx4fVX3v/z8/mclkX8mekJ2ELIYtBmUJhFVw31esy622tt9q0Axq4wAAIABJREFUe21/t/XeSu3VetHqVXur7a3WtteqVVArKCA7CYSwQwIBQoAsZA9JyEYymc/vjzNznOwsmcxkOM/HYx6ZnPc5Zz7nvM75rO/P+3MfeXl5nDlzhiVLlmAwGIiJiVHPHRCguP54e3ur26Kjo/vZPT0vOXT5FWELHfbt28enn35KTEyMeu8thbbJZKKmpobg4OABK0QAjY2NRERE4OrqisFgwGAwqBWklpYW1ebq6qr2hlgqYCdOnKC7u5ubb765l7bD5SlD2W2dH4FtdFi5ciWbN28mLCwM8znUj8lkoq6uDn9/f1UfS6FuadT5+vqi0+nQ6/W98quGhgYSExPx8/PD398fX19ffHx88PLyoqKigpCQEIKDg3n77bdZvHgxqampapoGu5c9PT00Njbi5uamlk0tLS3qp7GxccR1GK573yil/DWAEGJFH9tx4GGzh3/BUCeRUlbTP0DBX/rss3zY1FoRFBREVlYW/v7+vbZbMrTGxkYCAwN71aot36WU1NfXqxmO5TiLrbW1tZcIloeiu7ubgoICnnjiCTw9Pdm/fz86nY4lS5aQkpLCxx9/zO7du3n00Ufx8PBg0qRJakYaHBxMZWUl0dHR6m/NmTNHfUFHo1ZtCx08PT1JSEhQKzt9C5Rx48bh7e3drzdFSklcXJxaUbM+zpIZxsXFIaWku7ubzs5OtfutoqKCpqYm7rzzTmbOnMmZM2eoqqrCx8eH4OBgysrK+rVwLNg7k7OFBnq9nvDwcNzd3Xv1elk/9waDAX9//34aWHSwPO9SSoxGo/o9NDSUlpaWfr8phKCgoIBnn32WyMhItcBLSUlh1qxZthx2GRFsoYPJZCIoKEjtvbDOcywaWFpuA/VoeHh4EBAQoOY7HR0dtLS0YDQaaWxspL6+HqPRSHd3t/qxDJ/8z//8D4cPH+bTTz/lrrvuuugeE4veFy5coKurq9dfvV5PfHz8pdyCS8YWOoSEhJCUlNSvbLAQGhqKn59fr1a+paepubkZHx+ffj0tJpNJrTS3trbS2NhIV1eX+jl58iQ5OTnMnz+f1NRU1qxZw+HDh7nttttwc3Pr9fudnZ0cP36cw4cPU19fT1tbG35+fmpvmL+/P8HBwcTGxtLcfFHt6ktiuEJfL4R43vzdw7JRKE/UCZTQh0JKafsmUh/i4+PJyMggMjIS+LaL3vK9traWkJCQfg+/5SWsqKggLi6uV8Zo+Zw8eZIJEyb0a5lKKQkKCuL73/8+K1eu5KmnnuIPf/gDJ06c4Ec/+hGPPvoov/vd7ygpKWHSpEl4eKi3bNgXaDQKHFuwYMECdDod48eP79dtKYTg1KlTauFtXduWUnLmzBm1EmTd0rccf/r0aWJiYnod29PTQ01NDf/93//Njh07+OlPf4qrqyve3t6cOnWK4uJiuru7OXbsmJouZ8fT05Ps7Gw1U7Jg/f3s2bOMHz9ebanrdDpcXV1xcXGhqKiIjIyMXhUGiyb79u0jPT1dLWSsdero6OCll15i+/bthIWFsWHDBvbv3099fT0ZGRm4u7uP2ef6cpg6dSpz5sxRx2Whd2W2oqKCqKioXjbr+1lWVkZwcLDayLB+7puamvD29u73ntTW1vKnP/2Juro67rrrLr744gs++OADHnzwwUEL/oaGBtatW6cWOJ6enr16cizf3d3dbV7o24JZs2bh6upKaGhovwafi4sLZWVl/fIVy32trq5W/UoMBgNubm7qZ9euXcyaNYuuri46Ozvp6OhQh1ZWrVpFaWkpnZ2deHh48Mgjj/DFF1/w3nvvkZ2dzfjx4zl06BDl5eU0Njbi4uJCXFwcN9xwA/n5+dx4440AdHd3U1dXR21tLdXV1TapPA83pv/0QGP6UkophLgWJc6xadAT2BDLWO706dP7PawuLi589dVXLFq0aNDjP/vsM2bMmKHW4qzH7hsbG5kwYUKvblJLN1BYWBj33HMP99xzDxs3buQXv/gFb7/9Ni+//DIxMTEsXbqUNWvWMHHixH41PGckMTGRvXv3Wjub9MIyDDKYbbDjTCYTDQ0NJCUl0d3drbZwjEYjHh4e3HHHHbz11ls8+OCDXHfddYAyZvrVV19RVlZGd3c3JSUl6HS6QX07nAkfHx/1PgzEpk2bBrzXUkq1e9/yvzXHjh0jPDxcbQFaF0bTpk3jrbfeYt++fUydOhWDwcD27dupqKigtbWVpqYmvv/97/eq/Doz48aNw8/PjxkzZvQaCrHkS2vWrGHGjBm9WurWn9raWpKSkvo1Qiy9KnPmzOl33qamJt599111XPjee+9l1apVvP/++zz00EOqQ2RPTw8HDx5k//79dHZ2EhERwQMPPMDXX3/NQw89NGAF4auvRi2I4IiSlpZGcXEx2dnZ6PX6fp9PPvmEG264Qc3XrYdh8vPzycjI4MKFC+qnq6uL1tZWDAYDDQ0N6v338PBQK0wZGRnk5ubS2trKwYMHWbBgAffeey/r169n3759bNq0iXHjxuHp6cltt92mNlYBtcAHZSg6IiJCzbNs0QM8aKFv7s6P/fZfIWVv7/2ZwD+FEB0o9YB7+p7Dlri4uBAQEGDt5QugFtK+vr50dw8UXkAhKChI7dbpO17p7+9PYGBgv8Kmq6uLoKAghBA88MADLF26lLy8PJ566ik2btxIbm4uy5Ytw9PTk02bNrFkyRJb3wa74+rqSkBAwKAVHIsX90AMdRxAYGAgISEh/ZxtSkpKiIqKYvr06Tz88MPs2rVLdep89NFH+fWvf43RaMTNzY2jR4+i0+lU/w5nxeKRPRhD3euAgAC8vLwGHNOPiooiPT0dNzc3tWfAwokTJ0hJSeGNN97gvffeIz09HSEE+/bto729nfPnz/P+++8TFxdHTk7OsJXgnp4eqquraWxs5Jprrrm8G2FHLHmSZcgKevdA+vr64urqioeHR69CyDJuX1VVxcyZM3v1alnys5MnT6q+J5YeAikl/v7+xMfHs3v3brKzs2lvb+e2225j9erV/PWvfyUtLY2GhgaqqqowGAxERERw7bXXqk5us2fPHrRHYOnSpTa+Y7bBx8eHcePG4e7uTnd3Nx0dHb3ycVdXV8rLy9X7bz1+39PTg6enJ+7u7r0qA0ajkfLyctLS0vD19VUdXi2fs2fPYjAYKC4u5v777yc3N5ecnBxuvPFGduzYQVFRETfccAOBgYGXdC226CkbtNCXUv5MWK2eJ4SIstjM3ft/llKuGux4W9PT04Ner6emZuDFhIayAeh0OsrKBl6XwGQycepU7/DNlkzQzc2N7Oxstm/fzm233cbixYvJy8tj/vz55ObmUlZWxtKlS1m1ahWVlZW9anTOSFtbGzqdjrNnzw5ot3iOW2OZIqPX66mvr+9X0Fg+lmlD1tt0Oh1BQUHExsbS3d3NRx99xOuvv85PfvIT9fzz589n8+bNREREkJKSwqFDh5gyZcqg3rzOgJub27DP+1AalZeX99pm0ainp4e9e/cC/cehfXx8yMrKYtWqVezcuZNZs2aRlpaGEEKdiqTX6zlx4gTHjx8nLi6O1NRUdVjt/PnzVFRUUFNTo3aZ6vX6MVvoW7zrq6qqBrRb5zl976WUko6ODvLz89V9rZ1XTSYTJ06c6NXzaDkmLi6OI0eOEBERwf79+7nhhhtYsmQJGzZsIC8vj/j4eKKiopgyZUq/yq8zDr/U1dUhpRywlSylxN3dvddYubUWHR0dnDt3Tu3a9/T0VL8fP34cvV5PWVkZ58+fV4/R6XSEh4eTlJTEoUOHyMzMJCQkRK2ozZw5k8jIyEsu8AF1eHQkGW5M/3G+XUnvR5hX0jN376cLIXZi9uCXoxx32N3dnbCwMK6//voB7d3d3YPaQHlBrR3prNmyZUsv727rmvfevXtJTU1lz549TJgwgc7OTm666SY2btxIdHQ0q1at4plnnsHf35+8vDzuuuuuQeepOwNRUVFERERc9L2Eb+/nli1byMzM7NXNZv0pLy8nMDCwn62mpob29nYCAwO59dZb+dWvfsWdd96peo1fd911bN++ncrKSurr61mwYAGbN2/m2muv7ffitbW1UVFRQVVVFeXl5Sxbtsxm98qWjBs3bsSed/hWo61bt5KTkzPgOU+fPs348ePx9fVl06ZNpKen4+/vT2pqKm5ubpw6dYqUlBTq6+tpbGykoqKCxsZGcnNzaWlpITAwUPV+9vT0xMfHh7a2Nhobh1rW3XGxTNu9lPtsTVxc3KCFcEhIyIC2np4eNmzYwNq1azlw4ACpqakcOXKEtLQ0cnJy+OCDD5g+ffolD3FZfG7GYqUgISGB8vLyQe/1UDrExsYOes1ZWVkD2o1GI+vWrWPChAls27aNJ554gpiYGHbu3ElERAQeHh4XfR9bWlqoqKigtrZWnWkwaoW+MK+kZ/bOB2VFPWviUObqW3hsRFM2DEII4uPje3WlWTOUDZQXbLBurb4CWTuCTJ48mby8PJ588kneeOMNpk+fTmtrK/feey9TpkxRhxuWLFnCJ598QkFBwZBjrWMdi0PKxd5L+PZ+xsfHD+mokpaWNmC3fHh4OMXFxTz55JOsWLGCZcuW8cADD5Cbm6vqdN1115Gbm8uTTz7Jzp07WbBgAd988w1+fn7odDp1ep6Xlxd+fn60t7ergVDGIiP5vMO3Gg2V4SQnJ9PT00NmZibr1q1j+vTpLF68GFAy3ltuuYWWlhba29uJj4+nsLCQ5uZmoqKi1EAllmlJQgg8PT3VeAxjEcszfSn3+WLtg9ksPiteXl7s37+fJUuW8M0335CUlERISAjXX3/9sAW+0Wikrq6O6upqmpqa1O1SyjFZ6MPlz9K5HJteryc5OZm4uDj+8Y9/qA6ws2fPZvv27SxcuHDAZ8JkMlFdXc3+/ftVvxdfX1+ioqKYOHGi6sw80gzaBJVS3g0sklLeY/78qs8uTwArzZ8nRjxlF8Hlinclx86ZM4e6ujo8PT1JTk6mra2N7Oxsurq6+PWvf42LiwvFxcUEBgYSHBysTi9zZmzxgg1lt3S7GY1GQkJCiIyMpKGhgd///vfqPtOnT8fT05PVq1cTEBBAeXk5CxYsoLKykpSUFObNm6eOM9fW1rJo0SJ1Xu9YxFbP+1C25ORk3N3dSUpK4vTp07S1tXH8+HHVHh8fz+TJk5k3bx4dHR2EhYWxaNEiamtr6ejooLa2VnWIMplMtLa20t7e7igBei6LK3neL5eAgADi4+M5cuQIFRUVZGVlUVCgzKIerMFRW1vLpk2b+Oyzz9ixYwfNzc3ExcUxd+5ccnJyyMnJGZOe+xZGW4eIiAh8fHyIi4tj7dq16vbExEQKC5VI9lJKGhoa2L17Nxs3bmTr1q00NzdjMBiYN28e8+bNIzMzk7CwMLV32BZpHbTQNzvyvS6E+IcQ4hMhxD/67PIOEGj+/GGoHxFCTBVC/KcQ4jUhhJd52/eFED8WQqwx//+xEOIZIcTgfZQOQmpqKtu3b+euu+6is7OTpKQkbr31VkDpZrV4vS5atIimpiZ27tw5YPCZ0cTZNEhMTOTgwYM8/vjjdHd389JLL/Hcc89RWlrK8uXLcXd3JyUlhbNnzxIVFcWxY8cwGo1kZmbi4+NDa2sr69atIyAggNmzZ1NQUDAq3crOpENAQAAXLlzA3d2dzMxM/vnPf3Lq1Kl+DrR6vZ7p06czY8YMzp49S1JSEu7u7oSHhxMaGsqcOXN45JFHeOSRR4iPj7dJFLK+OJMOQUFBxMXFUVhYSElJiRqPYaA53mfPnmX9+vVUVFSQnZ1NQEAA2dnZTJw4EX9//0sK7HOlOJMGnp6eeHh4MGHCBLZu3cqUKVPYt28f8fHxNDQ0sGrVKjZv3kx5eTnJycnMnz+fnJwcUlJSmDBhwqimdaiW/s+sWvl3D+Cd3yKl/JtUFkA4N8zv3A8sBz5HmduPlPJt4AjwvnmfapSV+wYsHYUQTwgh9ggh9tTV1Q3zc7YlKyuL0tJSjEYjS5cuZc+ePaSnpzN+/Hhqa2tpb2/HZDLh7e1NREQERqORI0eO2DuspVNpMG3aNEpKSjAYDISFhXH8+HHuu+8+vvOd7/CrXymdUtdeey3+/v58+OGHzJ49m9zcXGJjYykqKqKgoIB58+YRFRWl+mNkZmaORtKdRgfLUMqUKVNISUnhiy++ICEhQXVG64u7uzuzZ89m6tSphIaGMn/+fGbOnElYWBiNjY2sXbuWsLAwNaSzjXEaHSzjzMeOHaOlpQWTydSrtQ9KDADL3PwFCxYwdepU9Hq9vbvvnUYDUIYKk5OTOXDgAK2trep0vzlz5hAQEMC8efOYPHkyvr6+vY4bbQ2G9DATQvxWCPGBEOLvQoi/9zGfE0K8K4R4F+gfsqs/A8VPvg34DJSYAFLK54GHBjxYyj9KKTOllJnBwcEX8XO2w2AwEBISwu7du8nIyFDDMM6YMUNdBGjHjh2AEsva4si0detWzp0brn5kU5xGA+sIco888ghdXV3cd999vVo348aNIzw8nPb2djUQ0N/+9jfc3NyYN28eJpOJdevWMW3aNHWceZRwGh0sMyk8PDxYunQpH374Ia6urtTX1w96jJeXF3PmzAGUcc38/HyOHTvGokWLiIyMHM1M0Cl0iI2Nxdvbm6CgIEpKSjh9+rRaGV6zZg3r1q2jtbWVRYsWkZGR0cux2AHG7J1CA/h2qqC/vz+ff/45U6ZMYf/+/cP6xow2w7mVV0spH5RSPiClfMCyUQgRDLwEPAP8GHh7mPN8hFKjuxUYL4RwEULEAuVSSqP5nM8JIV4ADl7OhYw2iYmJ1NXV0d7ezv333095eTlZWVlqob5r1y5A6faJjo6mvb0dg8HAwYMH+eyzz8jNzaW0tJSSkhKKi4s5efKkrZPsdBoEBgZSWlqqzj/Oz8/nmWeeQQihFv6TJ08mODiYtWvXkpiYSFRUFImJiTQ3N7Nx40ZycnIICAjg1KlT6tQ0G+NUOqSmpnLixAn0ej1xcXG8/fbbTJo0ifz8fPLy8mhraxvwuNjYWKqrq1m3bh0JCQnMmDFDdaQcpYLIaXTw8fHB3d2duLg4Dhw4oPYopqenYzQaWbRokbp+goPhNBoAanTRCRMmsGXLFnWtFqPR6AiVK5XhpuzNMj8obQBSSoun1ItSyidQ1jhGCPEO8L3BTiKl3Av0zVFPo1QcLPu8yBgiMTGRsrIyCgoKmDt3LlFRUZSWljJ9+nSOHTtGcnIyra2teHt7M2vWLL744gtqa2uJjIwkODiY1tZWduzYgcFgIDo6muPHj6vBN2yBM2qQkZHB/v37SUhIYNmyZaxYsYJx48YRERHByy+/zG9+8xtiY2PZvXs3bW1tbNy4kQkTJlBVVcXhw4dZvHgxer2eiooKvv7661HphXE2HWJjY9myZYu6gFRKSgqff/45d955J1999RVFRUVqwS+EICgoiNDQUPLz8wkLC2Px4sX9prQOtm7CSOJMOggh8PDwIC4ujj179vDYY4/R09ODTqdj0qRJjljYA86lASgzVrZu3UpKSgqrV6+mo6ODjIwMDh8+zJQpU+ydPJXhWvqvAbtQpusVWm3v7LPf4KHvnBRLBCdXV1eampq46667MJlMTJ06la+//hohBJs2bQKUrmjLuE5YWBghISHEx8eTlZVFamoq9fX1o7LCm7MRHR2tjmG6ubkRHh7OsWPHyM7O5ve//z0VFRUIIUhMTCQ4OJjCwkJOnDhBSUmJun71F198wccff4yHh8dVEy52JLEU2Dk5OUgpWbZsGW+++SYGg4HJkyeTlZWleoNnZ2cTEhJCVVUVer2ea6+9dsAC/9ChQ/a4lDGNt7c30dHRlJaW4uvrS0lJCXBlvSYOsLTumMIS9TI2NpampiZyc3MJCQmhrq6O0tJSeydPZbiW/k18O9bSKoQ4K6UsAbyFEIFSykYhRCDgPfgpnJfY2FhMJhO7d+9m4cKFuLq64uXlRUxMDPX19Rw/fhwppRoiNjs7e8CXMD09fVDnJ43B0ev1+Pj4UFZWRmxsLA8++CCvvvoqaWlpuLu78+yzz/LRRx+RkZGhTqM8ePAg48eP58MPP8TX15eGhgauueYaFi1axOeff27vSxqTuLm50dXVhU6no6uri5qaGgoKCvjqq69Yvny5up+LiwshISGEhIT0c2YCKCwspL29fUCbxtCEh4dTUVFBZGQku3btIikpieTk5Is+XkrJ+fPnqampoaamRl1cyZG6pR0dIYQa+jcxMZHVq1ezcOFCpkyZwpo1azhz5kyvfb29vfH19aWlpUVd6dLNzU0NDQ+26fUartB3AT5EKfgfBn6OEqVvOfBbIYQH0G7+/6ojLi6OLVu2MG7cOKqqqpg7dy7btm1j6tSp7NmzhyVLllBRUaGu9DaUeM4cwMeWjBs3jpKSEtWZLCwsjLNnz/Lcc8+RmZmpLgYUGRlJa2srNTU1zJw5k7q6OlpaWnBzc2PhwoWcP3/eYbtBHZ2YmBiKioqYOHEiR48e5Qc/+AFvvvkmH3zwQa9C3xrrd0FKSUFBAd7e3mRlZfHMM88MGblOoz+W8fzY2Fg2bdpEWloa3d3d/dYmsSClpLi4mGPHjqmVLF9fX8LCwsjKysJgMLBly5ZRvALnwN3dnfPnz5OcnExubi5SSoKDg5k7d26vZ95kMtHW1kZLSwuHDh3CYDBw4cIFOjs7MRqN6voKdXV1I17oD9e97y2l3GMee3EDzgJIKcuklI9KKe+TUj4mpTwz9GmcE0vX5DXXXMPBgweZMWMGUkomTpzIgQMH6OnpYfv27XZOpXMTFRXF+fPn6enpAeCBBxR/05MnT/Lzn/+cp556ClCWPW1qalIjwwUGBuLi4kJsbCwbN26ksLAQT09Pu13HWCYpKYny8nJ1RsSSJUvUWBUXLlwY8liTycSWLVsIDQ0lLS0NgDfeeMPmaXY2LEvFxsXFsW/fPpKSknoFSrKmrKyMtWvXqtEorQPDREVFqVEytVb+pRMSEgIoZcKRI0cGHWZxcXHBx8eHyMhIMjMzycjI4Nprr2X27Nnk5OSoQZIs5xtJhiv0N5oD83wCbALstsCOo5KQkKB2wezYsYPo6GiMRiNz586lpKSEyspKurq67J1Mp2X8+PHo9Xp1gSQvLy/0ej1Hjx7l8ccfp6qqinfffRcfHx8CAwM5d+4cs2bN4vTp05SWlqLT6Zg3bx7XX389a9assfPVjE2CgoLo6OggMDAQnU5Hfn6+ei8TEhJ49dVXey1QArB8+XKMRiPffPMNaWlphIWF8dlnn3HPPaO6WKfTYDAYMBgMREZGUlNTQ2dnZ7+Ff5qamli/fj0tLS3ccMMNREdHDzmVTCv0L52EhASEEAQHB6vP9HCMdvTA4Qr9VShBEt4HVkkp9494CsY40dHRlJWVkZycTHFxMbfccgtdXV1kZGSoc/UPHz5s51Q6L+7u7nh4ePRaMfEHP/gBLS0tFBYW8sorr/DKK6/Q0NDA5MmTaW9vZ9euXdTX17Ns2TJSUlJwcXGhvb2dN954w+6RE8ciLi4uuLq60traSkxMDJWVlepw1ZdffsnevXuJi4vj3//939VYCCtWrGDNmjW0tbXxs5/9jIiICN566y0WLlxoz0sZ07i7u+Pu7k5sbCwrV67Ezc2Nzs5Ourq62L59O0eOHCEnJ0ddAhm0gn2ksdxPnU5HcnLyFff02qPQ/wODhNoVQvzCHLTnwwEC91w1CCHUdZiDg4NVr/6QkBCam5s5f/78sIX+YMtAalwclmEWo9EIgL+/P25ubhw7doxbbrkFHx8fVqxYQWhoKOPHj8fLy4vs7Gz1uIKCAqZOnQpAZ2ffiSkaF4Ofnx+nT58mOzsbk8lERUUFAFOmTOHDDz9k165dNDQ0MHHiRB5//HGCg4P57ne/y4oVK5gyZQqFhYVs2rSJ7373u3a+krGLr68vnp6exMTEsGXLFiZOnMgnn3zCtm3bmDRpEjNmzBh0jF9jZNDr9RgMBrq7u0lJSSE/P5+WlouJXTd6DFfoDxVqV5iD9txvHbhnIAaJsfy/5pjKi83/LxdC/KsQYsytbTphwgROnDjB9ddfz7Fjx5g8eTKtra3ceuutHDp0iMbGxkEjlEkpWb9+vc2D8zizBqGhoQQGBva6h/fddx81NTUcOnSIFStW8PHHH7N161YSEhIwmUwEBwfT1dXFL3/5S26++WZeeEFZQdrW0/acVYfIyEjKysqIjo5Gp9Oxfv16nn76adWekJDAm2++yZ///GeklJSXl5Ofn09+fj5PP/10r5Xgnn/+eZun1xl1iIqK6jWuHxwcTFBQEAsWLLjkGRE9PT02b4g4owYArq6uCCHUVSi/+eYbeyepF8MV+kOF2k0QQtwthFgqhFg6zHn6xVhGiansA+iEEAGASUr5W2DqQCdwpBjLfYmIiKCyspKgoCDq6+tZuHAhPT09pKens337dnp6ejh48OCAL1FeXh6TJk0ajWVdnVaD6OhoLly4oLYuLdt0Op06bz89PZ2vvvqKAwcOMH36dAoLC7nuuuvYu3cvBw4c4J577hmVwgYn1SE8PJzW1lZ1eurZs2eZNGkSoFRsjxw5wubNm5kzZw7vvfceUspBV3EbzON/hHE6HRITE+nu7iYkJISOjg6OHDlCSkrKRR1rMpmorKwkLy+PjRs3sn379tHQwek0AKXHxbL2SlJSEuvWrRvyXj7zzDOD2myhwXBT9l4ELC7NfeNpbgU8zJ+LoVeMZSnlfwAIIf4H2NnX3u9gKf8I/BEgMzPToQZehRDqlAs3Nze6u7vx8PBASklYWBjV1dVUV1dTXl7ea4ymtrYWV1dXPDw8RqvbzSk18PHxoa2tDZ1O12ua0qJFi1i7di2FhYW8/PLLzJ8/nwULFlBbW8uKFSv4zW9+w+OPP66Ob45SYQNOqENISAhdXV20t7eTmZnJ6tWrKS8v59ChQ1SSqaX3AAAgAElEQVRWVpKcnMyiRYvsncy+OJUOERERSCnx8PAgISGBlStX8h//8R+D7l9XV8err77K4sWLEUIQERHBtGnTcHd3B5SAS++//76tk+1UGoAyk6KxsRG9Xs/EiRPJy8vjxIkTPPPMMzQ1NdHc3Exzc7M64+jgwYO9pkfqdDo8PT3x9PTknXfeGfF8abhC/xUp5U8AhBAvA/9mZctFCb0rMd/0IbDEWPYETgkhXIAfoPgKlEkpzwkhdEKIfwX2XfJVOADJycnk5eWRnp7OsWPHmDNnDuvWreOWW25h586dxMbGqtGaQHnhLKsydXV1jcaYvtNrYBlmSU1NBSAzM5O1a9dSVFTE/fffz0033cR7771HRUUFBQUF/TyX8/PzRyNeglPqoNPp8PX15cyZM6Snp7Nt2zaamppYt24dc+fO7deqH6VelaFwOh2EEGqAsPHjx7Nly5YBC/329nZ27tyJv78/v/vd7/iv//ovO6QWcEINQMmHCgoKMJlMpKSk8PHHH5OVlcWpU6fw8/Nj/PjxpKWlodcrxe+8efPYvHmzerzRaKSjo0NdrXWkGa7Qt5643Dfq3vPAc+bv/wl8Z7CTDBJj+a0++ywfJi0OTUhICGvWrGHevHkcOnSIBQsWsHbtWlJSUnjrrbd46KGHqKur48CBA3R1dVFfX090dDSVlZX4+Pjw97//nTvvvNNm6XN2DQICAnB3d+fo0aNqoQ9KfP6ioiJOnjzJb37zG9577z02b97cL/xrbW0tubm5Ni/0nVkHLy8vqqurSUlJISIigtbWVgwGA9u3b6eiooLbbrvtonpV7Bh7f8zr4ObmhoeHB7GxsfzlL3/h+eefV5eallJy4MABzp07x4wZM/Dw8KC9vV21nT17lgMHDrB//352795t87Q6qwbjxo3DZDLR1dVFaGgosbGx5OXl8fe//x2dTqd+XFxc1GHd1157TZ126ebmpn63xZDFcGP6eUKIVUKIT4G+T8EFKWW5lLIc6BjxlI0xhBCEh4fT1NSEm5sbFy5cwM/Pj7q6OnXt97KyMvz8/Ojq6uLuu+9m/vz5JCUlYTQa+eKLL+x9CWOa6OhoysvLMRgMnDhxQt1+00030dPTw969e9VAF9YFvslkYufOnXz55Zf4+fmNerqdCV9fXzUYzw033IC7uzsPPvggvr6+nDp1ij/+8Y+9lj7uS21tLRs2bGDVKi0cyOUSEBBAV1eXGgfe0oqvrq5m7dq1hISEMG/ePBobG/noo48AZRgsNDSUKVOm8OKLL3L48GGmT5+uRai8TCw9Lm5ubnh5eXHjjTfi7e1NcHAwfn5+eHh4oNfrkVKq70tFRQXFxcXs3r2bTZs28c9//pMPPvjAJukbrqW/CiiWUu4WQkT2sX1prgxI4P9skroxxty5c9m7d68aDevGG2/kgw8+YPHixaxatYq9e/cSEBDAhQsX+Mc//sHp06epqKhQa9sal09gYCAHDhxg8uTJrF27lgkTJgBKAR8REcGZM2c4e/Zsr27lhoYGduzYQWdnJ3fccQeffvqpvZLvFFhCIHd0dODj40NWVhYFBQXcddddbNy4kTNnzvC3v/2NmTNn0traqq5UWVRURH19PT09PQghho3ipzE4cXFx5Ofn4+rqSlJSEiUlJWzevJm2tjbq6+v55S9/ydatW2lqamL27NmAMtPlscceIywsjLS0NCzr0j/33HND/ZTGELi5uQHKFL6YmBhuu+02dYXJ1NRUNagYwIsvvshrr7024HlsUfEartB/A7iA0sr/Bcp4C0IIN2ADsNG8n0M4UNgbDw8POjs7CQoK4vDhw8yfPx8XFxeio6M5d+4cnp6eHDhwgPT0dG6++WYmTJhAYmIi4eHh6gOgcXlYXo5x48bh5eVFe3u7Glb33nvv5fXXX2fHjh0sX74ck8nEnj17aG9vp6urCw8PD1auXDkqS+s6M6Ghoej1ejVY1ezZs+nq6mLjxo3MmzePwsJC9u7dy+7du6mtrWX37t2q41l8fDwJCQkEBATw4Ycf2vtSxiyWmO+BgYFERkZSUFDAo48+SkdHB3PmzGHOnDn8+Mc/ZuLEiezdu5fW1lZmz55NQkJCvyEvjcsnNDSUhoYGPD090ev1VFZWkpOTQ0NDA3v27KGgoIDAwEC1B8CyMBsooaubmppoamoiMTFxxNM2XElznm/n51tHLfkFMN78XaAU+o+NbNLGJomJiZw8eVIN0BAeHk5lZSUvvvgiv/rVr9SY5H15/PHHRzmlzoe3tzetra0sXryYnTt3Mm/ePEAZa/b29qaiooLt27fT0tKCv78/xcXF+Pn5qQGVtHDJV4bBYMDd3Z2qqip1hTeDwcCiRYvYuHEjkydPJjg4mC1btmAymYiJiSEyMpK0tDTa2to4ePAgOp2OqKgoO1/J2CUgIABQfIwSEhLo7Oxk7dq1JCcn9ypUNmzYwNSpU9mwYYPaK2bBaDRSUFAwGtOInZbExEQqKyvp6OggKioKV1dX9u5V3Bd8fX0xGo3U19dTVVXFzTffzDvvvKM6w3p6emIymejp6VH1HEmGq9rVATOEEK8CqhuhlPJ5YKV5sZ1HAW1A2kxMTAxnzpxRPckXL15MV1cXU6ZMobi4WN3PEvxi69atbNy4kfvvv9+OqXYOLCGRPTw88PX1VUO+Atxxxx20tbWxa9cuTCYTxcXFeHp60tLSQltbG5MmTcLb+6pcIXpEEUL08zjW6XQsWLCAwsJCvL29ufvuu3F3d+fOO+8kPj6enTt3UlVVxcKFC8nIyFC7nTUuD51Oh6urKwaDAR8fHyZOnKgW+OfOnWPjxo1kZ2f3W8zFZDKRn5/Pn//8Z86dO+eIUyzHDDExMVy4cAEXFxeEEAQFBTF//nyCg4Opqamhurqa9vZ2dDod3t7euLi40NXVRVVVFaWlpdTU1ODq6jpoLIsrYciWvpTyJSHEKpTKQYlluxDiRuB+IYSlW/9+tIIf+HadZC8vL4qKikhNTUWn07F7926eeOIJ9uzZQ3NzMzqdjvDwcBISEjh37lwv5zONyyM0NJTi4mJSU1OZOnUq69evV+cgR0dH4+rqSmdnJydOnMDNzQ0/Pz9uv/12TCYTe/fu1ZY3HgEsQ1ydnZ3qfG9QfCtycnLIy8sjPDycsLAwNmzYQEBAAPPnz+/VqtTiwV8ZBoOBnp4eXFxcmDlzprr9zJkzlJSUsGjRIjo6OigtLeXZZ59l3bp1VFZWcu7cOdzd3fHy8qK2tlYr9K8ALy8vpJT4+Pioi3qVlpaSnp5Odna2unSupQLw8MMP4+HhQVdXF+fOnaOhoYHS0lImT5484mkbtNAXQlwDzEWZS3kjSqCe283mIKALCEbpAXhpxFM2hsnIyODgwYO4urrS3d1NXFwcp0+fJi0tjfr6etzc3JBSUltbS1BQEGFhYezcudPeyR7zuLi4qK1MFxcXJkyYwPHjx9Wu5jlz5rB27Vri4uKYPXs2QUFB9PT0sHbtWhYvXkx5ebk9k+8UhIWFUVNTQ3l5eb9uYyEEs2bNYt++feh0OubOnav5stiAoKAgGhoa8PX1JTMzk48++oimpiaMRiPe3t789a9/RafTodfrGTduHGfOnMHFxYVp06YRFxdHREQErq6uoxGYx6nR6XTq2iwJCQlMmzatl10IQVhYGDNnzlTDfxsMBkJDQ1WHP8t6IiPJUN37L6A48K0D6qSUlgIfKeVfgONSyr+YY/MfGvGUjWEsjmSJiYmUlJSQk5ODyWSiubmZU6dOcezYMY4dO8bRo0fJzc3ls88+05zIRgg3Nzd1HfGEhAROnTqlvjhZWVl4eHhw++23ExQUBMCOHTuYPn069fX15Obm2i3dzkJ4eDg9PT1UVlYOus/UqVOJjIwcsMCXUtp8HQpnZ8KECZw/f56wsDD0ej1VVVUYjUY8PDzUcePx48eTkJCAlJIFCxbwL//yL8ydO5eYmBg1oqUt1nK/mvD29qa7u5uOjo4he6+G6mEc1ZY+UC6lzBdC7JBSDrTQ+CwhxM+BZgAp5e8HO5EQYipwB0qwn/+QUrYJIZ4CxgF6KeXzQoiPUcIu7pJSjvlmb1xcnLqmdUpKCqmpqRQWFhIbG8uFCxfo6enBZDJhNBrR6/WDLsgzUlwtGiQmJvL111+TkJCATqcjMzOTvXv3qvOO09PT1X1LSkrw9/entbWVM2fOMGvWLJunz9l1cHd3V5/v0tLSQcckB8oEa2tr2bNnD42NjSQkJNg0nc6sQ3JyMps3byYwMBCdTsdDDz2kTsMDZSXJ6upqmpubiYyM7KdRdXU1R48e7TU8YwucWQNQFkCqra3Fy8uLqqoqAgIC8Pb2viQHSVsMdQ1V6CcKIf4BeAshPgGklPIeK/srfDtVb7jJhPcDPwdmoCys8LmU8vdCCAPfRmCqBgIYZPqfEOIJ4AlQHLYcnfj4eDZu3Iher8doNDJjxgzOnj1LTk7OgC+TJVCGDbkqNAgLC2PatGmsW7eO6dOnExQURGFhIW1tbXh5ebF0qbI21Pnz5zl9+jSRkZHU1NQwZ86c0UriVaFDRkYGq1evprGxkUmTJvVbW8I6M+vo6GDbtm2cO3eOoKCg0VqHwml18PLyUruV/f39exX4oFTMLPffEpDKaDRSVFRETU0NYWFhzJ49ezR6vpxWA4CkpCROnDihLkDl5uZGa2urGnPfQnNzM1FRUQQEBBAQEICfn59NZ04MWuhLKYdbOa8TZd6+BN6+iN/qtXCCEMIdxU/gJfPvPW3e/jsgf4D0ONzCCkMhhMDDw4OQkBBOnjzJxIkTufHGG/sV+FJKDh06RGZm5mgk66rQICMjg7S0NHbs2IGvry9ZWVnk5+erU/hMJhPbt28nKCiI7u5upk+fPtpJdGodDAaD6qDn7e3Ntm3b8PDwYOrUqb2ef5PJxLZt2ygtLSUuLo6bb74ZLy+v0ViHwoLT6mCZReHu7q7GddfpdISEhBAWFoafnx9CCHx8fNi6dSsmk4nU1FR1ZUQYNYdKp9UgLCyM7u5udDod06ZNG/R+bt68meTkZM6dO0d5eTmFhYVqxUBKqeZbI8WVeNF8H/gXlFb+7xlAECsGWljhY6AIWAT8rxDiOcANOHgFaXIorrnmGoqKimhvb2fixIn9RK+trWXv3r1cc801lJSU2CQQgxVXlQY6nY7Zs2dz6tQpcnNz8fX1pbq6WnWadHFxISAgoJ+z2Sjg9DqEhoZSXV2tPu/z58+npaWFXbt2ATBlyhRWr17N+fPnSU5O5uGHH+41vj9KhY1T62AwGGhvb2fBggXq/TQajdTW1lJaWkpzc7M6F3zOnDlqBDlrRkEHp9bA3d0dKZV6SExMzKD7xcXF4ePjg4+PT7/eCuvV90aKK3Wdvaju/UEWVri1zz4vXmFaHA5fX19aW1vVLn5LxtbV1UV+fj7u7u4sXryYxsZGDAaDTdNytWoQFxdHaGgo27ZtY9u2bWRmZnLmzBmys7PtEgTmatAhPDycoqKiXhmYr68vkyZN4sCBA3z22Wc0Njby9NNP2y0KnLPrEBwcTH19fa9nXK/XExERQUREhLpty5YtAxb4UkrOnDlj04Lf2TUA1FUPv/rqK2bMmDFgsJ2h7vGojumbx/G9AAPQgzKmf4PVLm8D76EU+L8b8ZQ5CTExMTQ3N3Py5EmSk5M5fvw4p06d4rrrrsPLy4udO3ei0+m0eOM2xNPTk8WLF7N+/XpWrlzJI4880m+cU2Pk8Pb2pq2tDVCivx09epT6+nr8/PzIyspi7ty55Ofna2FfbcjEiRPZvn07tbW1vQr5vlgKFSkl9fX1nDp1itbW1n52jcvD19cXk8mETqfj9OnT7N+/H1AqZXFxccMGBBvVQl9KebcQ4gXgVyhz8f+1zy5lKN0wEjgz4ilzEhITE9m0aRM1NTUcOnSISZMmqXPC8/LyuO666wgMDBzNccyrEiEEixcvpq2tTSvwR4ELFy6watUqgoKCSElJ6Tf1SAuEZFuSkpLYuHEj1dXVaqFvmS1k+XR2dnL8+HFOnjypRo1LSUnBx8cHQMuTRoDo6GhKS0uZOXNmrwpWXV0dRUVFtLa20tLSonr2e3t7q139lkh9I81w3ftJQCTQDcT1sS1HGcs3AK8D94104pwBFxcX3NzciImJoaysjJiYGDZv3kxAQABLlixRw2NqNerRYerUqfZOwlVBdnY2BQUFZGdn2zspVyWW8eSysjKam5sRQiCEQK/Xo9fr1TC9JpNp0Mh7Wp505SQkJHDkyJFe91IIQUhIiBoH4fTp00RHR9PW1kZrayvnz5+nurpa9fQfbUe+F4Afmb+/1cdWJKU8aL4ILYbsEFxzzTUcP34cNzc3tmzZwowZM9TatMboomVko4OXlxdxcX3bCRqjiRCCSZMmDamDtt6EbQkNDR12IS9LnmRp4YeHh6s2W/S2DFfo3yKlfHYQ2z1CiJvM59AJIf7RZx6/hhl/f3+am5txc3Nj8eLF9k6OhsaooFWw7IvBYFDDuw6GppFtMRgMCCG4cOHCgA6TwzHawXlgiKh7UspZQggf8/fzI54yJ2PhwoXaGJmGhsaoERoaSlFREWFhYYAyltzS0kJ1dTU1NTX09PTQ3NyMv7//gMdbQvVqXBkeHh6sXr2awMDAfjadTkdHRwfx8fH4+/vj5+dn85lcwxX6g0bdE0L8OxBj/l4upXxh5JPnXGi1ag0NjdHCEo7XxcUFKSVCCPz8/AgLCyM+Ph5XV1dOnz5NbGwsUkpMJlOvT3V1tb0vwSnw9fXF39+ftLQ0fHx8cHd3V325enp6KC4uxt3dnZqaGo4fP95rOMAewXmGiroXKKX8LoAQ4pWhTjJIjOWfoMwKkCj+AiuAc8AeKeW6S70QjaHRNHAMNB0cg6tBhwkTJrBu3ToaGhrQ6XS4uLjQ2tpKVVWVugJcS0sLxcXFahAZi7e4EMIma7lbczVoAJCSksLp06fJz8+nu7sbo9GIi4sLLi4u6HQ62tvb8fHxQQiBi4uL6nQJ9Jo+OVIMV+gPFXXPXwjxHRRxxg1znn4xloHxUsofCyFeByYBh6SUfzX/309c6xjLQKsQ4hjKEr+DrVQzlG04uy1sF3vs2j7xEEYKW2lgnfaBGEsaWdvGmg6Odi9H6rzOpMNY1WisaWCd9oFw5Od9KNuI6HAxEfmk+dM36t5vUKb0CeB7F3meS/nb+2CrGMsWhBB7pJQDBq0fynYlx9rjN0eQEdcAHO9+2So9I8iovgvD2ceSfiOMlifZX4cxkyc5on6Xw3CF/tvAn83f/6eP7XYp5YqL/J2BYiyXCyGeAU6jxFV+wOw0uP4iz6lxaWgaOAaaDo6BpoP90TSwA0OF4f0tsAn4/iDe+Q8LIeYDLfRfdrcXg8RYfr3P/z+9uCRrXA6aBo6BpoNjoOlgfzQN7MNQLf03gXnAG0IIP+CYlPIXVvZsKWWjTVM3PP26dS7SdiXH2uM3HR1Hu1+2So8j42j30pbndWQc7X5pedLo2Oz1m5eMsHht9jMoXS1ZwFwgFEBK+WOz7X0Uz/5TUsr/GulEaWhoaGhoaIw8Q7X0K4FPgJellGf72DqllN8TQvQNzauhoaGhoaHhoAzV0vcF5gDXAx5Ag5TyP822fcAGlO7/TQBSyp+NRoI1NDQ0NDQ0Lo+hWvou5o8HEIviYWnhdvPfvh79GhoaGhoaGg7KUC3994EtwGYp5ZlRTJOGhoaGhoaGDRi00HdkhBC+UsoWIYSflLJ5iP3GSynLB7GlSSmLrP5PAMKllLkD2CYAXUAKUCalPDLA+WZLKbf32fY03wY2klLKNy/GNla4GB0uRQPzNk2HS2SkdRhKA7Nd06EPWp7kGGh50vBcTEQ+h8Lsa/CAEOLvwJMosZmt7bdYvgI3Ad+1sv3IypYN3Gl16L8BuUKJBw1gLfwPgUjgBZTog5bzIITIA3YBU4UQ+6SUP7E67nPz34EiGlrbxhxD6XAFGoCmwyVhIx2G0gA0HXqh5UmOgZYnXRxjrtAHfgwkAiHAsQHst6NEERQo0wqtiUMJ/iCAvusclkgp/yKEWAj8EiVOgYVSlLjHh4QQFQOkZxJwRkr5hrVBSnlGCPFDIB04hLJ+gbXtKWAaoAMeGeqiHZChdLhcDUDT4VKxhQ5DaQCaDn3R8iTHQMuTLgYp5Zj7AM+bP78cwDbB6ntoH1swYDB/9+hjS7eyXd/HFmVlmz7Ab45HiVw4UFp/YP77wwFsz5oFvd/e93QkdbhcDTQdHEOHoTTQdLg0DWylg6bB6LwLzqiDso7i2OO/UZZd7DdmI6U8IYR4XgjxS5QuHmtbHeauGyllRx9boZVtZx9bhZVt1wDpmSWl7Lv0sIVYIcQLQPQAtnLAFZg6gG0sMKAOl6uBeZumw6UzojoMpYF5m6ZDf7Q8yTHQ8qRhGKuOfLegTCdcJKV8agC7H0q3yDLZp1tFCLEWOAlckL3HWGxiM9tDpJS1A2zXAREoOpQNf+WOxVA6XK4Gw9k1HfpjCx1spZHZ7nQ6aHmSY6DlScMzVgv9OYAJKDbX0vrahxLeH2UMp0xK2TAKtpeBbsBVSvlvfWx/Ag4zdj1lB9XhcjUYzq7p0B9b6GBDjZxSBy1Pcgy0PGl4xqIjH8B0IB4oA14awN6MIvzzA9iWo4QYPg+8Mwq2SinlW0KI/zdAWrZJKf86wPaxwlA6XK4Gw9kv16bp0J/lXN69vJJjnVUHLU9yDLQ8aRjGaqFfJqVcIYT4Xl+DEMJ6aUYJ9O1SqTF/ogY4ry1sQgjxDlA4gO16IcQ0oGegrp8xwIA6XKEGw9k1HfpjCx1spZGz6qDlSY6BlicNw1gt9H8ohFgMJAkhkqxvipTyx0IIPcrKgP3GSoBi4A7gs1GyfQl0oKxVoCKEyAQ2AxcYo/NiGUSHK9RgOLumQ39soYOtNHJWHbQ8yTHQ8qRhGJOFvpRy1jC7vAI0oYjb13MyFfjXgZwnbGR7DEXwp4FnrLb/BDgCtAx1IY7MMDpcrgbD2TUd+mAjHWylkVPqoOVJjoGWJw3PmJyyJ4T4nRDiNSHEa4PsUgmcArwGsP0BeFQI8aZQwiva2lYHNABnhRIxysKbQBDKONOgYTsdmWF0uFwNhrNrOvTBRjrYSiOn1EHLkxwDLU8anjFZ6AN5UsqfDDHW0YwSgWmgmtIClIhN64AnRsEWAHwHZbVCtTYnpcyXUj4jpfyLlPIvg1yHozOUDperwXB2TYf+2EIHW2nkrDpoeZJjoOVJwzDmpuwJIaKAd1HGO4xSyjsG2GeulHLLIMdnSin3mL/7SilbbGx72vy1G9gkpSy+3Gt3JIbT4XI1GM6u6dAbW+lgQ42cTgctT3IMtDzp4hiLLf3xwN+BlcAXg+zzgBDibvHtIgvWLBZCfCyE+LSvsDay+QHbUGp7Dw1/eWOG4XS4XA2Gs2s69MZWOthKI2fUQcuTHAMtT7oIxpwjn1TCIPYLDdqHD6z/EUJESSVcIkAbsJb+iy7YyuYNnAM8gePDpHvMcBE6XK4Gw9k1HaywoQ620sjpdNDyJMdAy5MujrHY0h8WKeVWy8e8aR6AEGISyvhJKHCv9TG2sJn5G3A38H/0eeicmcvRYDi7psOlM9LvwpUei6aDlifZCS1PUhhzLf3LxLJe8WTgrPl733mTtrCBsm7zXuCYlPLCJaTZ2bgYDYazazpcOVf6LlzpsZoOClqeZH+uyjxpzDnyXQ5CiJullF/a8fevB96WUk62Vxrsjb01MKdB00HTwSGwtw6aBvbXwJyGUdfBqQt9IcRiKeU6O6fhRSAM+K2U8og902IPHEEDczo0HTQd7I4j6KBpYH8NzOmwiw5OV+gLId5DibokgMlSyhw7p+ePKI4aSCmdyVN2UBxNA9B0QNPBbjiaDpoG9tcA7KeDM47pH5VSvgIghFhm78QAh1DmYY7JCFeXiaNpAJoOmg72w9F00DRwDOyig9O19AGEEAFSynP2TgeAECIDZWrGJCll33jPTosjaQCaDvZOhwVNB/ujaeAY2EsHZy30n0eZKrFKSllg7/RcjWgaOAaaDo6BpoP90TRQcMpCH0AIkQO8LKWcbu+0XK1oGjgGmg6OgaaD/dE0cNLgPEKI14FAYKa903K1omngGGg6OAaaDvZH00DBKVv6QohUYBrK9f3V3um5GtE0cAw0HRwDTQf7o2mg4JQtfeBxFM/Ig/ZOyFWMpoFjoOngGGg62B9NA5y30D8P6ADn68YYO2gaOAaaDo6BpoP90TTAOefpA5wC0s3fD9kzIVcxmgaOgaaDY6DpYH80DXDeln4XMAO4mheTsDeaBo6BpoNjoOlgfzQNcN5CP0xK+T0gwt4JuYrRNHAMNB0cA00H+6NpgPMW+jFCiMeAaHsn5CpG08Ax0HRwDDQd7I+mAc47pl8MHEaJvqRhHzQNHANNB8dA08H+aBrgvC39ICnlbpRADBr2QdPAMdB0cAw0HeyPpgHOW+hvEUK8Cmyyd0KuYjQNHANNB8dA08H+aBrgpBH5NDQ0NDQ0NPrjrC19DQ0NDQ0NjT44ZaEvhDAIId4SQvxOCPEHIcScizhmuRAivc+2e4UQ7wkhfi+EeE4IMVcIUS6E8BJCuAsh3jfvZxRCTDN//8j892khxLtCiD8JIcJscJkOjaaBY6Dp4BhoOmg4Cs7qvf9d4Csp5degvnC3AjcD7sByoBv4T6AW+Mx83BNCCB9gG/BPYKmU8juWc6AEdtgE/Bh41er3NgD/CjxgtW2ulPJ2IUQW8C/m37qa0DRwDDQdHANNBw2HwFkL/TTgY8s/UsouIcTDUso7hRAxwP8DTMALUsoTAEKIBcDHUso8IcTHQBHK9A7rcwDsBiYDoVa/1wpsF0Istdr2RyHE74FGIMgWF+ngaBo4BpoOjoGmg4ZD4JTd+ygvxzTLP+YasQWL56JAecmsabba5yTfxmnue47fAL/oc+z/otwxOmAAACAASURBVKzipAOQUn4tpXwK2IwyP/RqQ9PAMdB0cAw0HTQcAmdt6f8v8JoQ4maUB/4j4P+EEH8APIFfo3SlLRdCVKF0m/VCStkghFgvhHgX6AQqgJ1m20khRDvgZrW/UQjxAfAKgBBiGXC9eZ8f2exKHRdNA8dA08Ex0HTQcAi0KXsaGhoaGhpXCc7ava+hoaGhoaHRB63Qv0iEED81T7mJH2IfMZppuhrRdPiWS71OIcSntkrLJaRhthDiY7OT2mD7jJl8SdNAY6zhrGP6AAgh7gbagRAgE2VazL8B7wEPoni7/glIAhqllF8KIf6M4kn7NBAM+JiPexD4CrgghPgb0AIcklL+QQhRBPwNWCWE+CNQgLKoQxUwHfg3KeXR0bhmR0TTYeQQQsQCHwBfAh8KIZ4AxgFrpZSfCyEOA++jOI09DlwLPAFYPMINwB9QHMQapZQvCCG2okwJSwa2mI/ZIKX8wHxMMPAs8DawRkqZJoT4P/N5fwH4AweB9cAPpZQ/FUI8BRxF0fx6wBd4E0XPLuCsEOJNwAi4Sin/nxBiNbAD2C2E+Peh0mRPNA3sr4HG5ePstbktQDaQiuL4MhflIb6AMje2BlgGrARuN8+HNQLjzcc1obwc4cAB4CXgduA9KeUPAEtN+ayU8mXzvsellD8DooDfAf8F3GDj63R0tqDpMJIcMV9nB4pTVg1KZQigQkr5WxQHr8nA91HmZL9jti8Ctkkpn0FZatQXuCCl/A8gD8U57DHgVsuPSSnrUCpm84A8IcRk829KlIZDI3CPlPIMEG1uJc4GtgI/RNGvBsgCVgN/R8l7zkkpfwLUCyUIjQuwQkr5zXBpcgA0DTTGJE7d0pdS1gkhQoE6lIf2fpSa9wsoHq0C+JWUslUIIYHvAKtQHvwiKeVyy7nMtWbMx/T1fmy2+t5i/ntBStkihOjCyqP2akTTYcSxXOcyFC/vXcAX5m1t5r/dKNdrQqlAXTBvH+i+qfcKaJFSygG6d5uB61AKruXAn4GlKIXfX4UQm837bUIpHE5LKU1CiI4++j0yQDosf9ullMZLSJM90TTQGJM4daFv5jzK/NZ8YJmU8pz55fj/UGq+FlaitAiTzFNdTEKI11C6h1+y2u9D4HUhxB0o8101Lg5Nh5FnB/A9YCZK78ZA/AFYAVSb/18PvCOEuAYoN1eILua3tgN3oPS0zAQeBQKAl4UQ4ZjnggP/AEpQenNAmZb2R5QW8RrLyaSUh4UQTwohXgE8pJSFY9QVQ9NAY0yhTdnT0NDQ0NC4StC6ajQ0NDQ0NK4StEJfQ0NDQ0PjKkEr9DU0NDQ0NK4StEJfQ0NDQ0PjKmFMF/pCiC9H6XeGdGkVQrwhhPC8jPPGCyHetY7SJYSYJoR48nLSaStG6z6bf8tW99pLCPEXIcT/CiEeHMo20L5CiNvM/3/x/7N33uFRHecefr8t6qsuUEVCSBQJAQJELwYuJYC7cbdjYpu4xCXN98aObeLETpzYvnFwTa4dYjuJHTsOTlzApveOQIgugeioIiEhrbTS3D+2WF0I1Faa93n22d2Zc+bMmd+Z852ZM/ONiMzoajrputD+6HrQ9euBpmXc1uiL3SvWsSvYP1pEVonID8XuktJbRF4RkddF5FciEi4iG0Xkf4BwEVkjIk84KsH3ReQTERksIsGAUkpdFJEgEXnLkc7UlvKglMpWSt1bL2wHdqcaXYIrLWdHGp1e1tinOn2ilLofuKaFuAbbKqWWOP7fA9zSlXTSdaH9aesydoTpeqDpcC57nr6IRCqlTjt+RymlTrVdti6J4cDOWvmZg93Fa5ZS6gNHWAr2daZr8z2lVC4wFFiilHpV7MtPPgL8TSm1XeytjVTgQ6XUH0TkO8Cnjm2XYF8m8zwQi93F5T5H2v7YXcm+qpRa6chDAnaPWLV5TilV2My5VYhIb6XUuWa26SjqlHNjtFDO0DXKOhrIcIRV19umflxz2/4ceN3xu6vo1GxdcBN9mqJLlnFjtPJ+A7oeaDqBK2np31fr92NXmpHLYASww/lHKfUF8DIQWSssQyk1t96n9o1umXNTIBnIELtf7IvY3Wd+44gfBiwTETNQoJSqAQZjrxDB2Csj2D3NzVVKOT1zgd3rlanepyWKsFfmrkCdchaRWBG5T0TmO1o/LZUzdI2yPon9JgYNr/v6cQ22FTsvAl8ppZw3/66iU7N1wU30aYouWcaXUQ/qlzHoeqDpBC6rpS8iHwNJYvfVrIDMNs3VpTEcCBERG/Axdm9VPwX+9xL3TwAOiUgodk9Zq4A/Yq98v3akdbD2tsAQ7AtYAMQppY6LSCDf+nQ3AK+ISAnwM6VUtVLqMA2fugEQkRDgeSBVRH6mlHK2EqKA45d4Hu1N/XL2xO4/3wwM5NK6PDu9rLG79X3N0Qr+D4CIjMXeuvpLvbgl9bfF3ir7LyBARBKUUm/RdXTSdaH9udJ6UL+Mwe45r1PrgbMOKKXeqB+H+9UDzSVw2R75RKSPUuq443eoUiq/TXPW+vw8jX01tqVKqS878LgCvKmUeqCN0vMDXmqr9NoaEemDvdIbgBVKqaMdeOw2LesrzEuX1UnXhfZH1wNXXrqsRprGadboO1rycwEPAKXUc7XiXlJK/cTx+wWl1JPtnNcui4hMBHYopS62QVr9gWqlVNaV56z70ZZlfYX50Do1gq4LHYOuB5rLpSWj/y7wCvbVolBKHawV945ztK2I/FEptaCd86rRaDQajeYKaOmd/l6l1N4m4j4SkX9iXzby/5pLxPE0+CT20atLHGHfBUIBX6XUcyKyEPtKbLlKqfdbcQ6aS0Tr0PloDboGWgdNT6Uloz9FRK7CPmAFpdTNteL2AouVUv8RkWHNJaKUOiQii4HAWsHDlFI/FJFnRCQIqFFKvSwi/ws0qGAisgBYAJCUlDRi8+bNbN68GaPRiMFgQERc30ajEZPJhKenJx4eHq6P0WjE398fHx8fpOsuIdluGbtSHeprkJmZSUZGBjk5OZjNZgwGAyaTCZPJ5NLAZDK5wp3fXl5exMTEdGUNoJ10aI+6kJmZyf79+ykrK8NqtVJZWYnVaqWqqoqqqirMZjPe3t54enpiNBpRSmGz2RgwYADh4eHtcZptidvosGzZMpYvX+66rmtf3waDAR8fHywWC76+vvj6+uLn54fFYqFXr17tcYptSZeuqJrW0azRV0pdLSLJjt/1R+g/DxRiH9X5PeDRy8yDauK7fl7+iH2kKyNHjlQnTpxg48aNGI1GZ7xr25qaGgwGQ+19G/0N9oopIiilCA0NJSAgwGW0DAYDFRUVJCYmEhkZSa9evTCbzZd5ml2eFnWorwHArl27OHbsWKPl7dShfpkrpRARTCYTgYGB9O3bl9jYWMLDwwkMDKyTVg/jsuoCwD/+8Q/XtVzf6FRXV7uu8VppsHr1ap544gm8vb3b41zcmcvSITc3l+PHj2MyNbyt2mw21wNX/XuVj48PP/3pT9v0BDSapmjW6IvIq0C+4/f3lVK1DXsRUHwpBxGRcOAmwFtEArDPV00XkR8DKKWKRMTo+N+sAwwnhYWFbNmyhaCgoAZxRUVFBAcHYzAYGvQCFBcXExkZiYeHh8u4K6WoqKigurqavLw8VzpGo5Hq6mry8/MxGo14e3tjsVhcv81mM//1X/91KdntErSHDjU1NWRlZdVprThvaufOnSMmJsbV4jebzZhMJs6ePUt8fDylpaXk5+eTn5/P9u3b8fDwoLKykoiICIxGI2azmdDQUCIiIjCZTCQnJ7dxiXQ87aGB1WrlxIkTREZGNhqfn59PWFgYIoLZbHb1emVnZ/Pyyy9zzz33EB0d3ei+3ZX20OHChQvs2bOH0NBQV6PB+cnNzSUmJgaj0VgnzmazUVpaytq1a5k0aVK7na9G46Sl7n2bUuqXACLy23pxh4C7HSP8tzaXiFLqLA3njf6l3jYLW8xtLUpKSigrK6OysrJBXGVlJZWVlS7j43y6rq6upqCggOzs7DrdoJWVlVy4cAGr1YrBYCAgIAA/Pz88PT0xmUzExsYyadIkgoODKSgoQCmFn58fJ06cICcnh7S0NFJSUrp6d3W76LB06VK2b9+Ot7e368HK+amoqGDPnj1UV1djs9lcnwsXLlBaWorJZCIsLIzAwEB8fHzw9/cnKCiI4uJiQkNDMZlMnDt3jn379lFaWsry5csJDg7G398fs9mM2WymoqKCOXPmuE3vQHtocPHiRVatWoWnp6dzf9e3Uoqqqiq8vLwaPAAfOnSIWbNmYTKZGDRoEFOnTsVisVz2ubkT7aFDXl4e2dnZnDlzBpPJRGVlpauH5cKFCxw8eNDVuKipqaGyspLS0lICAgKorKzkzJkz9OvXD39/f0JCQkhPTyclJcVVXxITEzl8+DClpaX079+f06dPExUVRWFhIeXl5SQmJrJhwwb69u2Lt7c3J06coF+/fuTk5HD69GmmTp1KRkYGYWFheHl5kZubi4eHByJCTU0Nffv2JSsri+DgYHx9fSkpKSEuLu7KClrT5WjJ6JtE5FnHb1cfoGOe6GFgOvYZAKXtlL8mGT58ONOnTycwMLBBXFFREQEBAa6bnlKKmpoawN5DEBISAlCnGzQ3N5egoCCsVitlZWWUl5dz8eJF8vLyOH36NC+++CIVFRUMHDiQ1NRUEhISXA8ca9as4euvvyYwMJC4uDjKy8uZOHFio3nrbnzve98jLi6OqKioOt2WIsLJkydd74xra1FUVERQUBBlZWUUFhZSVFTk+mRlZfHFF19gtVrp1asX4eHhro/FYqGoqIjCwkLCwsKIjo4mMzOTlStXMm3atC7/0NVeBAUFMXfuXAYPHozRaMRmswH2nqrKykqys7NJSEiguroak8lEVVUVBoOB7du3s2TJEk6dOkVFRQV79+7F09OTyspKoqOjCQoKorq6Gm9vb5d25eXljBkzhqCgIMxmM9XV1ZjNZo4dO0ZcXFyd1whOYyIi2Gw2qqur8fDwoKysDKPRyIkTJ+jduze+vr6Ul5ejlMLX15cTJ04QHx/fmUV6WQwbNoyZM2cSFhbWIC4vL4/w8PA6D11gf/Xy3nvv8X//93/cdtttHD9+nPDwcMLCwjhy5Ah9+/bFx8eHoKAgPDw86NOnD1u3bsVisdCnTx98fHzw8PBwlW11dTWRkZGYzWZCQkIwm80EBwcTFBSEj48PqampLo0sFgvbt29nxIgRKKXw9PQkJibG1djZs2ePNvrdkJbe6T/W2Dt9pZQSkTRglbK7h+xwLBYLnp6eJCQkNIjLyMhw3eRqampc3zU1NVitViIiIqipqaljiMrKyggLC3P1CDj3y83NZcKECdhsNs6dO0dOTg7p6el8+umnmM1mpkyZwqBBg/D09CQ/P5+zZ89SU1NDeHg458+fJzIykgEDBjT6nq87MHDgQNLT00lLS8PDw4O8vDyGDh3KwYMHMRgMXHvttWRkZBASEkJ5eTl5eXkopYiKisJkMhEUFMSZM2cICAigvLycgwcPMnLkSNLT0zl69Chnz57l3LlzbN68mb/97W8kJCTQv39/4uPjOXz4MACbNm1i3759xMfH4+PjQ1paGtu3b0cpRXx8PMeOHWPQoEEcO3aM8vJyevXqRW5uLr169cLDw4MTJ06QkpJCTk4OI0aM6OQSvTxiYmJcPSUlJSUYDAb8/PwoLS2ltLSUqKgoKisrsVgsrviEhARef/11nn32WRYvXswdd9zB8OHD2bZtG0ajkaKiIi5evOgy8FarldOnT5OXl4fJZMLHx4eamhqGDBnCli1b6N+/P6mpqezatYvg4GAsFgvp6emMHTuWU6dOUVlZyeDBgzl8+DAhISEcPnyYiooK+vbty8mTJ6muriY+Pp7s7Gy3NPqRkZGEhIQwc+bMOoOKRYQlS5YwdepUrFarq3fR2dN4zTXXsGvXLt59913mzZuHwWDAarVSXFzMxo0bXT1lzt6swsJCTp8+jaenp2tQoMlkQilFXl4eW7ZswdPTk9DQUEJDQwkMDKRfv34AeHl5ufLr5eVFv3796jROgoODXb+1we+eNGmJHN35cd/+FVVv9P544N8iUo79OeDm+mm0JwaDgcDAwDoXsRNnd3FjFBcXu564a2O1Whu9yCsqKkhNTQXs769zc3M5cuQISikyMzPZsmULy5cvZ9KkSQwfPhwvLy/KysrYuHEjQ4cOJTc3l4MHD+Ll5YXRaGTmzJlXfvJdiJqaGoKCgqioqKCiogJPT08OHDgAQEBAAEeOHMHb25uLF+0+REJDQykqKiIiIsKVRkxMDAD+/v6UlJTg7e3N2LFjGTt2LGVlZWRnZ5OXl0dISAiHDh1i8+bNfPnll0RFRREdHc3EiRMpKysjPT0dgM2bN7taocOHD+eqq64CoHfv3gAcO3bMFQa4HhxLSzu8w6rNCAkJcZ1faGioKzwoKIjAwED8/PxcLXDnjd3Ly4s+ffrw3HPP8dZbb/H6669TXFxMREQEnp6e9O7dm+HDh+McoFZdXc2FCxcYPHiwa7ZGaWkp+/fvp6CggKysLLKy7D5ajh8/jre3NyUlJfTq1avOmI/aD1bOOhcQEOAKa+xB3l0IDAyktLS0TkOjpqYGPz8/zp8/j4eHB97e3q5XVEopsrKyePPNN1m4cCF//etfKSwsZMqUKVRX29e4MRgMWCwWPDw8MJvNXLhwgbi4OKxWKxUVFRQXF7tmxVRVVZGUlIS/vz8FBQUcP36cPXv2kJ+fz8yZMxu8vmnOsGuj3z1p0ugrpZ6QWqvniYhrpI+je//PSqlPOyCPjeLt7U1KSgpjxoxpEOfl5dVoeHNx58+fr2MImtteKcWHH35ITU0NL7zwApmZmTz//POsXLmS8ePHk5iYCMCZM2eYMWMGJSUl5Ofns3z5cnJyckhOTiYtLY1z5865phWeP3/eLStZXFxcAx2cr1O2bNnCmDFj6owoB3uZjh49GpvN5mr5OFs/VVVVjB07Fk9Pzzr7vPLKK9hsNlJTU7n66qs5evQoBQUF/Pa3v6W4uJg777yTiRMnYjabsdlsHDt2jL179/L222/j6elJ37598ff3x2QyNfne2h3L30lcXBypqakYDAbXQDHnt9lsZsiQIXXK2Wq1kpycjMViISgoiPnz5xMdHc17773HxIkT6d27N4WFhRw4cIDExEQmTJiAyWSib9++hIWFuV7HiAhhYWEUFxfXKT+bzUZ5eTmnT5/mww8/5JZbbmnwoN1UeburDiJCSEgIFoulziA+o9HIxYsXiYuLa6CNyWQiLS2NgwcPsmDBAry9vfnwww8pLi5m0qRJXLhwAV9fXy5evEhRURFVVVUUFxejlGLmzJnYbDaXFoWFhRw5coTVq1dTUVGBr68vcXFxjBo1ip07d7Jjxw4CAgIYNmxYj30Vpmn5nf69gNP17qPAE+Dq3h8sIptwjODvDHeQhw8fprKykiFDhpCenk54eDgGg4EDBw4wYMAAMjMzEREGDhxIRkYG0dHRnDhxgoqKCoYPH87OnTvx8vIiPj6es2fPcvToUcrKysjPz2fEiBHs2LGD06dPExsby8GDB4mPj6ewsJDz588ze/Zsjhw5wqeffsqsWbP4zW9+Q1VVFS+99BLvv/8+jz76KIcOHaKgoICrr77a9Y46NTWVvXv3kpmZSUhICOPGjcPDw8P1TtTdOHbsGAcOHKCiosIV5nxvmZ2dTVVVVYMpe8eOHaOiosI1ktzT07POw8+uXbvqpAe43u/v3r2b/fv3M2jQICwWC/fddx9r167lzTffJCoqioiICPr168eIESNcBu/kyZPk5uaSl5dHamoqq1evZsCAAd1qqpqzpV37dZbzOycnxzVP31nOQUFBrp6RpKQkvLy8SElJISoqisWLF7Njxw7mzZvHXXfdRUZGBhkZGYSGhroevCIiIoiOjnYZj759+za4fqurqzl37hyVlZW8++673Hzzzd16oKDzeg4JCanTyne+VrRarQ20qays5MCBA0yfPp2amhpGjRqFj48PS5cu5eWXXyY5OZmYmBiSkpJITU1lypQpfPbZZxw/fpzFixdz1VVX0bdvX1dPSkREBHFxcVRVVXHmzBkOHz7MP//5T86ePct1112Ht7c3S5cuZfTo0XW68jU9hybd8IpjJT2+XUEvUyn1i1rxf6bW/FWl1PfaMZ91GDlypNq+fXuThrI5A9rafZpL68CBAy5nHPfccw++vr4opUhMTCQxMZFx48YB4Ofnxw033IBSypVWdXU1mzZt4sCBA/Tq1QsR4eqrrwY3cYTh1ADatkybilu9ejVXXXUVSinX4LNZs2axdOlSZs2axfz588nJyeHXv/41gYGBjB071nVTKysrY+nSpRQWFhIQEMChQ4dITU1l9uzZzbV43EqHy73mY2JiyMjIIDc3F6vVSmZmpuv1y1tvvUVBQQH/8z//w3XXXcfOnTs5fPgwd9xxB+fPn6ewsND1QOfh4cH48eMblOexY8cIDQ3ls88+w2q1MmHCBPr379+aU+wWOjSnz9GjR7l48SInT56kf//+fPrpp1RUVLBjxw5Gjx7N7t272bdvHzk5OVRWVhIeHs4HH3yAiHDs2DH8/f0ZPHgwffr0aTJ/69atIzc3l6KiIiZPnkxubi5Go5FevXpdyvgJt9BAc2m05Hvf1b3fSJwZmOH4u0wpZWuH/DVKbYPT2Rw+fJg1a9ZQVVXFfffdh9lspqysjOTkZH7wgx9QVVVFdXU1FosFPz8/5s+f38CRTUZGBtu2bePee+1LGXTaybSCjtag/k1z+/btbNiwAR8fH/r06cP06dO5+uqr2bVrF8888wxgH1h1zTXXuPYpLS3lP//5D0ePHiUsLIzk5GTXg1kj9Cgdampq2L17N+vXr6ewsJDg4GACAgI4e/Ys77zzDkVFRTz77LNs3ryZqVOnEh0dTUJCgkuTJUuW4OvrS3h4OMnJyS6nWU6qq6tZsmQJFy9edI3DuMTBrT1Gh5qaGvbs2cPJkyfJyMigsrKSsLAw1ysDs9lMdnY2ixYtIjIykp/+9KckJyeza9cuwsPDyc/Px2w2M3bs2CZ7scrKyvj8888pLS0lLS2NrVu3MnXq1JYMv1tooLk0mmvp1xnIR73BeiLyDrDa8fcq5Vh8pyPoSkYf4KuvvqK8vJzCwkLmz5+P0Whk06ZNzJ49m+eff971DrWiooIRI0Ywd+7cBmls3rzZ+V7cLSpYV9Dg9OnTvPPOO0yYMIHY2FhiYmK45pprSE9P55e//CV5eXmMGTOG8ePH4+Hh4dpv0aJFWCwWrFYr06dPb+qG1yN12LdvHx9//DHBwcFUVFRgNBqxWCxkZmby+eefc/ToUU6cOMGePXvo27cvx48fx8fHB5vNxuTJkzlz5gx79+4lMDCQYcOGNfBguWzZMvLz8wkKCiIyMpJhw5r14A09UAebzcbHH3/MoUOH8Pf3x2q1IiJ4e3vj6+vL0aNHef/993nggQcYOnQoU6dOZfXq1QwbNoxVq1bRp08fKioq8Pb2JjExkdDQUHJycuo8NJeUlPD5559z+vRpBg0aBEBaWlpTLoHdQgPNpdGkRxOl1BNKqZsdn3mNjM4vUUq971iIoqi5g4jIcBH5lYi8IiK+jrAHReSHIvKF4/9HIvK4iIy90pPqaGbNmkVxcTHx8fH89a9/JSsri7Fjx/Lwww/z0ksvYTKZXCOrMzMz+frrrxuk0dTAw7aiO2rgnCKVmZnJvn37MBqNLF68mBEjRvDMM8+45jmvXLmSo0e/Xe587ty5eHh44OXlxVdffUVJSUmH5bmr6+AcKxEfH8/gwYNds2Ouuuoq7r33XuLj41mwYAGpqalkZ2czZcoUhg0bRlZWFocPHyYiIoLp06fTr18/1q1b1+BanzlzJv369SM3N5e1a9eybdu2BmM+OoKurIPJZOK2227Dx8eHIUOGMGDAAIYNG4bFYqG4uBiz2cztt9/Oiy++iNVqZdu2bcycOZMjR45QUVHB+PHjmTZtGkOHDuX06dOsXLmS+g8k/v7+3H777QQFBWEymcjNzeU///kPf/7zn9m7dy9VVVUcO3asI09b00E068ZMRF4Wkb+KyN9E5G/1ootE5B1Hi7+lu+ZtwEJgCXaHPiil3gT2AYsd25wFgmjCz7WILBCR7SKyvbar3K6AiHDTTTdx5MgRoqOj+fLLLwH48Y9/zMiRI/n8889dA9cGDRrE1q1b+eKLLzr6ZtctNZg9ezbe3t6UlpaSnp5O7969+clPfkJqaioff/wxO3fuZMaMGZSWlrJq1SoqKyvp27cv119/PSKCr68vH3zwgcuhTQfQpXUQEW644QZGjBjhmqPfu3dvTp48yfDhw7n11ls5deoUN954I8nJyXz99deYzWamTp1KZWUlq1evxmazERwczNSpUykqKmpgPMaMGcPo0aNdXjWXLl3KhQsX2vQ8LoEurQPAxIkTmTZtGldffTURERFERUUxbdo0wO6n5NZbb2XBggWUl5eTlZXFhAkTGDBgAN98841r9P7QoUOZNm2ay9FYfZKTk5k5cyZ33303o0aNIjQ0lF27dvHhhx+ydu3aNj8nTefTku/Ss0qpO5RStyulbncGikgY8ALwOPBD4M1LOFZji1dcB/wL7I6AlFLPAnc2urNSf1RKjVRKjWzM41VnY7FYmDBhAmfOnKG0tBSbzUZQUBD33nsvhYWFrpHSJpOJa6+9ll27dvH3v/+9I40NdEMN4uLimDdvHrm5uWRlZWG1Wpk8eTJz586ld+/evPPOOxw6dIiUlBTS0tJYtWoVO3fuxNvbm+nTp7tWYFy8eLHLa2MH0KV1iIuLIzw8nBtvvJGkpCROnTrFyJEjOXToEB4eHnz3u9/FbDZzxx13YDAYWLp0KSdOnCA5OZlhw4bx9ddfuwzMqFGjOHjwYAN32U63vwUFBeTl5bF69Wr28Ap44gAAIABJREFU79/f5ufSAl1aB2fvn8lkYsiQIcyYMYPw8HC8vb2prq5m9OjRjBkzhp/85CccPnyYkpISxo0bx/jx41mzZg05OTmutCZMmMC2bdsa6OA8htFoJCUlhblz5zJhwgSCg4Pd2m+FpmlaMvoTROSnIvKQiDxUK/x5pZRNKXVBKVUC/LKFdD7E/lR9LRAjIgYRiQNOOAcAishTIvIcsPtyTqQrkJSUREBAAAaDgc2bNwMwbtw4/vu//5slS5awe/duzpw5Q2hoKHfeeSdHjx7lT3/6E2VlZR3RldZtNfD392fgwIEUFRWxfv16RIQbb7yRefPmsWPHDhYvXkxVVRV+fn7MmDGDXbt2AfbpTU7NbDYb77zzTkdk1610GDp0KPfccw+HDh0iMjKSmpoa/P39ue666/D392fRokWuMSxgd04za9YsDh06xM6d9nVqxo0bx8aNGxukPW7cOG688UbGjRvH+fPnSU9P5/PPP2f9+vUdcWpupYOT3r17M2PGDGJjYzl58iTz58/Hz8+Pt99+m3Xr1pGdnY2Pjw8zZsygqKiIdevWuTwdTpw4scXWu4jQt29f5syZ0+QCThr3pqXR+5OxPwU7B/KtdYT/QdVacU9EFimlHmnvzDrpCoPImqK6upoXX3yRsLAw7r//fsDum7+oqIh58+bxwgsv0KdPH26++Ways7NZsmQJIoKPjw8PPPAAuMmgma6mQXV1Ne+88w4iwrx58wgMDOSNN97g1KlTvP7667z11lvceuutAHzzzTcMGzbM5SN9+fLlFBcXc+DAAZ566ilnklqHWlRXV/Pll1+ye/duTCYT/fr149y5cyxfvpyamhri4uK49dZb68yGOH78OCtXruSee+5h9+7dBAYGEhsb22j6ZWVlrFq1CoAjR47w+OOPO6O0Do2wbt061q5dS2hoKBUVFSxatIi5c+dSWVnJG2+84dquoKCALVu20KdPHwYPHkx2djYXLlxg6NChLR6j1owZt9BAc2m01NKf6/jMAa4SEad/TD8RCQZwfPu1XxbdC6PRSO/evRERsrOzAfuo2KCgIGw2m8vVaWZmJv369WPSpEn4+/s3+r5Nc+kYjUYmTZqEUoqvvvoKsA+wDAsLY/bs2Tz11FOcPXsWsA9Ky8jIcO07efJkvLy8Gow013yL0Wh0jZ/w9vYmPz+f6Ohorr/+ekwmE4sWLWLfvn11eqz69OmDh4cHBQUFDBkyhP3791NVVdVo+r6+vsyZM8e1LK2mecaMGUNERAR5eXkMHz6cRx99lA8++IDs7Gw2b97sGi8UEhLCjBkz2LJlCwDx8fGUlZVx7ty5Fo/hjs7CNC3TktE3AP8APgJCgJ85whcCL4vIh8BLjv8aB7fddhs1NTWubkofHx/X1KfRo0fz9ttvs2XLFmw2GyNHjnQtham5MgYOHEhISAhFRUUcOnTI5TJ21qxZhIeHM3/+fADXsrzl5eWu/+PGjdMatIDRaOQ73/kOHh4elJSUUFJSgsVicZXr7t272bhxI7m5ua59Ro0axe7duxGRJrv5nYgIY8aMca3FoGka53x8s9nM1q1bCQoKYtGiRSxbtowVK1bw9ddfs2fPHmpqajCZTPTp08dl6MeMGcPOnTuxWq2dfBaazqAlo++nlNqulNoBeAKnAZRSx5VS85VStyqlvqeUymk+mZ6Fj48Pfn5+lJeXu0Ylp6Wl8dBDD/Hee++5HGSsW7cOEWHq1KluvchIV2LSpElYLBa++eYbNm/eTGxsLBUVFdx1113s3buXl19+mYULFzJs2DDX4EqwL0wTFRXViTl3D5KSkpg+fTomk4mioiKsVislJSXceeedru7/bdu2cf78ecC+eI6HhwelpaX4+/sTEBDAiRMnmj3GJczd12AfDBkYGEh5eTnR0dFcvHgRg8HA73//e3JzcwkLC2PFihVs376d2NhYdu+2D00wGAxMmjSJNWvWdMp0SU3n0pLRXyEiHztc8q4EOm2BHXdjypQpAKxYsQIAT09Phg0bxtmzZ/nqq6/YsGED77//PmVlZYiIvtG1EWFhYURFRWE2m1m3bh0pKSn06tULT09PHn/8cZ577jmee+45LBYLZWVlrpXMAIYPH96JOXcfEhISiImJwWg0UlBQQFhYGElJSXz66ae8++67HD58mI0bN1JWVgbgWm4X7AMDnfPAm0J3K186kydPxs/Pj82bN9OnTx8eeughFi1axI9+9COOHj3K9OnTiY2NZdOmTYSGhrpa+76+vq7pfZqeRUtG/1Psc1YXA58qpXa1d4a6CxEREZjNZs6dO+eaCjZ69Gj27dtHVFQUTz75JJ9++ikvv/wyoG90bcnYsWPp3bs3VqvV5cXM29ub8PBwbr/9dry8vKiuriYpKanONDGtwaUzb948l9e3gwcP4uHhwZEjR/jggw/41a9+hb+/P6tXr+bQoUP4+vpis9morKx0dfNv2rRJO39pAwYMGEBQUBAA+fn5xMbGEh0dza9//Wuuu+46srOzCQsLw8/Pj8GDB7ta+wCxsbEUFBRw5MiRzsq+phNoyei/DQQ7Pm/XjhCRJx1Oe/7eiOMeDfYufaWUaxBN3759GT9+PBs2bODBBx9k/vz5vPzyy6xfv56FCxd2bma7Ed7e3kRHR+Pt7c1nn31GeHg44eHh2Gw27rvvPvz9/XnjjTeIjIzk9OnTnZ1dt0REGDVqFKWlpa5XWeXl5fTv35+FCxdy++23k5qayptv2l14DB061GVwAgIC8Pf3r/N6RXN5iAjx8fH06tWLo0ePYjAYyMrKYuLEiTz++OMu18gZGRlkZWUREhJSZ8zF2LFjOXv2LCdPnuzEs9B0JC0Z/eZc7YrDac9ttR33aL5l8ODBGAwG9u7d6wrz9/cnJCSEM2fOMG3aNO69917mzZvHL37xi2ZS0rSW4cOH4+fnx4kTJ4iKiqK6uprQ0FDS09NJSkrimWeeITc3l+joaH3Du0zS0tKIjIx0vSaZMGECy5Yt44EHHmDKlCnMmzePDRs2UFxcTHBwMOfPn3f1eg0dOpTCwsLO8MTX7UhJSSEyMhKj0ciFCxe4/vrrWbduHQsWLGDOnDlMmTKFX/ziF5w4cYJhw4bVae3HxcUxfvx4Dh48SH5+fieehaajaMnoN+dqt5+IzBOR2SIyu7lEmvBz/SeHX+uZjv8LReTHInLXZZ9NF0NESEhIwGaz1fGOlZKSwoEDB5gyZQqDBg1i8uTJdRaEaae89CgNRITx48cTEBDAZ599hsFgICgoiICAAAYPHsz48eN54oknGDhwIAcOHOjIfHUbHby9vQkKCiI8PByAf//734wYMYJ///vfvPPOO1itVnbv3u2aHjlo0CDX6xQRYeLEiaxbt64jPSG66E46BAYGulx8A+zcuZM5c+bwr3/9i1deeYXBgwcD9h6WkpISgoOD67T2RYQpU6awfft2/RDWA2jJ6D+P3dXu48Cv6sWtAbyBUMenORr4ucbu19oCGEUkCKhRSr0MNDqaqiv6fb8UJk6ciIjU8YTlNEhbt24lPj6e73//+w3cY7YDPU6DlJQUUlJSKCgoICgoCF9fX5eRuuWWW/jqq6/YtGkTfn5+HbnoTrfSoV+/fiQnJ2Mymbh48SK+vr5YrVb27dvncjy1ceNGampqXL0qzhHjiYmJjBo1ig0bNnRG1ruVDiEhIYwcORKj0cjGjRvx9fVl/PjxfPLJJ/zhD3/Ax8fH9QBWv7UP9hH906ZNY+3atVRUVHTSWWg6gpaM/u8crnYvYPe1X5v1QAowGGh68u231PFzrZR6Win1S+yOfxrEN9i5C/p9vxTMZjOhoaFcuHCBixcvusItFguhoaHExMSQk5PDDTfc0BHZ6XEapKWlERISwvr168nNzcVsNmOxWPDx8eGuu+7i4YcfJjk5uaPfL3cbHeLi4jh16hRjxozBbDbz+eefM3HiRPbt24eXlxeJiYm8/fbbHDx4ELAPHqvd6xUaGkpwcDCHDh3qjOx3Gx2SkpLIysripptuQinFX/7yF5KSkoiNjWXr1q0MGTKEn//8567R+/Vb+4Br4aQVK1Y0O7tC4960ZPR9av2u73XvWeAPwCLg6RbSaczP9SMi8ixwXClVhP3p+sfAzkvNvLswY8YMAP7617/WCXe6xRw4cCBz5sxpbNe2pEdqICJcd911XLx4kdLSUhITE/H29qasrIxx48bh6enJX/7yF2w2W0fd6LqVDuJY5z0yMpI5c+ZgtVrZuXMnoaGhbNy4kfPnzxMdHe2apTJgwIAGBj45OZlTp05RVNTsCt1tTbfSwdPTE6vViqenJ3fccQdFRUWsWLGCyZMnU1NTw4gRI7j++ut57733OHjwYJ2BlbXx9vZm4sSJrFixwuVRVNO9MLUQv0FEPgVqgP/Ui7MqpU4AiEh5c4k4nPvsqBe8qN42C1vMrZvi7++Pr68vJ0+epKioyDXFRkSYMGECW7dubfd3+j1Zg6CgIIYOHcqePXvo1asXRqORwMBACgsLmT9/Pk8//TRr1qxh1apVrge09qI76jBkyBCWLFnCXXfdRWJiIhkZGVxzzTWcPn2aG2+80eUm9ic/+QkDBw4kNDSUvLw8areOJ06cyLJly5g5cyYmU0u3pSunO+qQmJjItm3buOmmmxgwYACbNm0iKSmJ6667jv379zNkyBB+9KMf8fHHH/Pss88SHBxMenp6Ax8h/v7+jBs3jh07dhAfH99JZ6NpLy5lnv6vlVI3Acvrxf1HRD5xOO75ql1y142YPHkyAB999FEdhzB+fn706dMHHx+fpnbVtAEzZsxARNi/fz8jRoygsrISi8VCTEwM1113Hb/97W/517/+1dnZdEssFgtVVVVUV1dzyy23ICKsXLkSPz8/BgwYQExMDPfdd59zQSmGDBnCnj176szTN5lMjB8/nrVr1+r5+5dJnz59XOMlbrrpJjw9PXnvvfeoqqoiLi6O0tJSfve73/HGG2+Qk5PD0KFD2b59e6Ne+fz9/enbt29Hn4KmA2jJ6L8K3OP4/aQzUEQ8sT8EfNcRr906tUBcXBxKKZKTkxt08yckJBAcHNxJOesZiAg333wz586dIycnBz8/PwICAjh9+jSzZ89m+fLlvPXWW52dTbdl2LBhZGRkICLMnTuXixcvcvjwYaqrq7FarcyePZvDhw/zxRdfYDKZ8PX1Zd++fXXSCAwMJCYmRs/fv0xEhKSkJE6ePInRaOSaa66hpqaGd999l9jYWNdaCXPmzOHxxx/HZDLRu3fvJpcy1s6quictGf0LgHM5ptpDOp8EXsPeHfaa46NpgYEDB7J3715CQ0P54osv6sTpCtb+9OnTh5CQEHbu3Elqairnzp0jLCwMHx8ffvYz+1pSNputk3PpngwfPpzc3FyUUgwZMgSLxUJeXh5lZWVUVVVRXl7Os88+y0MPPUR5eTmjRo3i1KlTDcZRJCYmEhgY2Eln4f7UnhY5YMAAlx+F9evXM2jQIHJycnj++eddS/OuWLGC8PBwl5tkTfenJaOfB4wTkZewv9cHQCn1LPBPx2I784HP2jGP3YaRI0dyyy23kJOTQ3FxMbXX39ZGv2Nwrgi3Y8cOPDw8SExM5Pjx4/Tr1w+gU+aMdxf69etHVlYWYH9H7+vrS0VFBTabjfLycqZNm0ZsbCwvvvgiBoOBffv2sW7dugbp6Lpw+RgMBtd4FYAbbrgBg8GA1WqlV69eVFVVkZ2dzQsvvMD8+fN59dVXSUxMBNDueHsIzRp9pdQL2Ofovws85QwXkTnAbbUc89zWrrnsJsTFxREcHMzMmTMpLy8nPT29zvQlTfvj6enpatmEhYWxZ88eQkJCsNlsREREtPuAyu5MfHy8a8T3kCFD8PT0REQ4fvw4NpuNQ4cO8cgjj/Daa6+RlZXF73//e6Kiohp082ujf2XUHplvsVhITU0F4Ouvv6Z///5kZmZy7733upwqgX1RpDNnznD27NlOybOm42jS6ItIiog8AhQAo4CPakWHApVAGBBCwzn8mmaIj48nJSUFT09Pli1b1tFTlXo83//+910jl4uLixk5ciSFhYUMHDiws7Pm1ogI4eHhnDlzBhEhOTkZT09PbDYbVquVs2fPMn78eO68804eeeQRwN4FnZubq+tAG+Lh4eFylgT25aZNJhPV1dWcPn2a6upqMjMzeeqppxARfvCDH3DhwgUmTJjAnj17KC4u7uQz0LQnzbX0nwO2AcuAPKXU9c4IpdRfgENKqb84fPPvaed8djtGjRpFaGgofn5+LFmypLO8kvVYbrzxRmpqarh48SLbt28nICCAAQMGdHa23J7k5GQyMzMB++ssZ9jZs2ex2WwcP36cmTNncvz4cQBKSkqYOHEimzZtqjOrRXNlDBs2zDUg0mg0MmjQIGJjYzl79iyhoaFs2rSJqVOnMnbsWMrKykhJSeGbb75h6tSprF+/noqKCr0IWDelOaN/Qim1GdiolPqikfgJIvIzEXlIRB5q7iBN+Ll+SESeFpFfOP5/5PB7Pfayz8bNmDlzJmazGQ8Pj0YdZbQlWoO69OrVi/j4eLy8vDhw4ACxsbGkpaW1+3G7uw5GoxE/Pz+Ki4sxmUwEBQVRVVWFn58fpaWlZGRkEBERwT/+8Q/APvDso48+YsyYMa5R5B1hbLq7DhaLhfLyctfA1GuvvZb8/Hyio6M5deoUlZWVrlb/888/z9tvv82CBQu4//77SU1NZeXKlfzyl7/s5LPQtAfNGf0EEfkHEC8iHzt+1+Z3wAZgL5DZwnEa+LlWSr0BvAg4XyydBYJowuVlV/Jz3VYYDAbmzJmDh4dHp/je7+kaJCUlER4ejo+PD+vXr8doNHbEYbu9Dqmpqa5W5tSpUykuLmbatGlUVlZy/vx5AgMDXS5gP/nkE373u99xww03UFxczMGDBztqxclur0NKSgp79tg7YY1GI1FRUfTv3x+A6upqli9fztatW9mzZw+TJ08mIyMDX19f0tLSKCkp0YNauylNGn2l1Gyl1M2O73lKqZvrbVIB3AfcC1gv4Vh1/FiLiBfwaxzjAZRSjzlmBdzZRH66jJ/rtsTHx4epU6d21Dx9rUEtEhMT8fPzIyQkhLKyMjrwxt2tdfD09ASgoqKChIQE+vfvz/79+xk2bBjV1dV8/fXX1NTU4OHhwdixY9m+fTvz5s3j3nvv5emnn8ZisXRUVru1Dr169SI/P9/lfGf69Ons2LGD2267DavVSnFxMc888wyTJk1izZo1+Pn58dprr/Hhhx/yzDPPdHLuNe1FS1P2muNB7EZ/AXB/C9s28HONfWCgFZgBICJPichzQPv2c3dBQkJCmDRpUnsfRmtQDxEhKCiI0aNHYzQaG/VM1g70CB1qv1MePnw4lZWVJCQk4OnpSW5uLkOHDuXhhx+mpqYGo9HIww8/TGZmJt7e3litl9KGuGJ6hA6JiYkcPnwYsA/wCw4OpqCggOHDh1NVVYWPjw9ms5nk5GR27LB7JZ44caKrh0DT/ZDLvdGJyGLsxl6At5RS32vDfDXLyJEjVe057t0M6ewMXArdRQOr1cqmTZsoLS1lz549PPnkt44nOzNfl0pX1mHFihVcddVVGI1Gtm/fzvHjxxk/fjx/+tOfCAkJYdy4cRQWFjJlypQ6+4lI7QcwrcMVoJTim2++ca0pUVRUxGeffcZ3v/tdXnvtNQoLC4mNjeX2228nPT2dqKgooqKigDo6uIUGmkujuSl7H4vIlyKyXESWicjSepu8iX3+/rvAH9szkxpNe+Hsip4yZQoGw5V0fGnqk5yczOrVqwFci7qcO3eOgQMHkpuby5YtWwgLC6MrGsvugogQGRnJtm3bAPviUxaLhX379nHfffcBdk1+//vfEx8fT0ZGBuXlza6fpnFzmnunPw/YDswEZtFwwZ3j2AfwZQDaw4zGbRk6dCj79u0jOjq6s7PSrQgPDycnJ4cLFy5gMpmIi4sjKyuLa6+9Fg8PD86cOcOaNWtIT0/nk08+YcOGDRw9epTHHnuss7PerUhKSmL37t2UlpYC9oGWGRkZeHp6kpCQQFBQEEop3nrrLcrKyvjyyy9RSvHss892cs417UFLTZv+QBT20az1l1xaiH11vZXA/7Z5zjSaDiIoKIiioiLGjx/f2VnpdkyaNIm1a9ditVoZOnQoRqORnTt38vjjjwNw/vx58vLyOHbsGCUlJdhstnZf3rinYTAYmDx5MmvWrKGqqoq+ffvi6elJRkYGd9xxBwsWLGDGjBmYTCYOHDhARUUF77//PnfffXdnZ13TDrRk9J8DHgV+TL21poFMpdRupdQ24HB7ZE6j6SgSExMbLP6iuXISEhKYMmUKK1asAOyLHjmn66WlpfHggw8SEREBwK5du1iyZIl2Td0OJCYmMnHiRFauXAlATEwMx44dczlESk1N5bHHHiMwMJDs7GwuXrzI3r17OzPLmnaiJaN/jVLqJ47PgXpxN4vI1yKyEpjUyDx+jcZtiIuLc41y1rQtPj4+jBkzhpUrV5KSkoLBYGDr1q3Mnj2b4OBg7r77bkaNGkVkZCSenp6UlZV1dpa7Jf7+/gwZMoQNGzYwZMgQTCYT33zz7aro3t7ePPzww4wcOZL8/Hxt9LspphbiJ4jIz4BicDmvwPF7gohYHL8vtF8WNZr2x+krXtM+BAcHk5SUxKZNm4iJiSE/P58lS5aQlpZGZGQkkyZNIjY2lrVr11JQUNDZ2e22REREUFJSwoEDB/D19eXs2bOsWLECDw8PEhISCA8P5zvf+Q4jRozgz3/+c2dnV9MOtGT0f8e3DizqTNsQkZ8DsY7fJ5RSz7V99jSajkOv7ta+REVFUVZWRmlpKVVVVQwePNhlgIxGIwkJCdx4440sXVp/opCmLRkwYABbt24lJCQEq9XKtGnTsFqtZGVlsW/fPgwGA7GxsR3hO0TTCbTUvd+c171gpdT9Sqn7gQ5zoaXRaNyX/v37IyL4+fmRkZHBmTNn8Pb2JiYmhvPnz7Np0yZ8fX07O5vdnrS0NPLy8qioqGDlypVs3LgREWH06NEuY9+7d+9OzqWmPWippe/0uifAG8DmWnGBIvJd7D0BIc0lIiLDgRsAH+BppVSZiPwIqHHsvwj4LVAEbFdKLbuMc9E0g9aga6B1sA8a27ZtG2VlZQQGBlJaWsqZM2coLS2lpqaGgICAds9DT9dBRJg8eTIffvghAQEBKKU4cuQIO3fudK0DMmHChE7OpaY9aMnog70CKBp6Zfo19il9AjzQQhq3AT8DxmFf3GIJEKOU+qGI/C8wFNijlHrP8b9BBRORBdhd/gKUishBIBTIb+R4TYU3F9dV0lqqlJrVxDZXQntpUDvv9ekqZXo54e6mgzuU6eXs01106Epl2tq02ksDTSfQktF/E3CO5ni9Xtz1SqnftuJYqpXfdXdW6o/U8/wnItuVUiPrb9tU+OXs09lptTFtrgG4X5leTlptTIfVhebiOrtMe5IOXbVMu8A9SdPBNGn0ReRl7I53HmxidP7dIjINKAFUI6vw1ca5uIUPcNSxuMUJEXkcOIZ9QYvbHTMFvr6cE9G0iNaga6B16BpoHTQ9kuZa+n8ApgKvikgAcFAp9WSt+ElKqcJLOYhSagewo15wfS9+P72UtDSXh9aga6B16BpoHTQ9leaM/glgP9Db8fF2RjhW2KsQkaNKqRfbNYfN09RCP80tANTafTo7LXfA3cr0ctLq6rhjmfYkHbpqmXbXe5KmCZpcWldEzgAfA79RSp2uF/eWUuoBEVmklHqkA/Kp0Wg0Go3mCmnO6PsDk4Gx2Fv5BUqpXznidmJfdW8q9vf+KKWe6IgMazQajUajuTya6943OD7eQBz2AS9Ornd81x/Rr9FoNBqNpovSXEt/MbAaWKWU0steaTQajUbj5jTphlcpdY9SanFXNPiOVw84ZhXUCXfGNbNvfSdDTYY3ta0jrlkvhD2By9WhhXK9ZB20BnbaWofLCO/xOuh7ksZdaLKl31VxVKDbgb8BDzgdBInI045NFHa/Ac/X28fJg84ZByLyKuCHfXriZKXUHxzhv8Mxfxe7l67HaqX1iuNnKrBLKfUjR/hjfOu5UDnTcsTdppT6u4jcrJT6R0vbuwOt1aEpDRxxrdKhKQ0ccVoHWq9DUxo44lqlQ0tl2t100PckjTtxKW54uxo/BBKAXsDBWuFlSqlXAMTuQ7s2nwGrsF/MowGnwTmrlPq1iCyk7voB5wAfpdRLIlJ/ru5R7IMYZyilXq0VvsTx3ZjL4kjHd58mtndHWqtDUxpA63VoSgPQOjhprQ5NaQCt16E5DaD76aDvSRq3we2MvlLqFyLyLPaLeECtqHyHF0GA9Hq7vaCU+gZAREbUCne+uvgF9idrJ7uxVyKA8/WOv0hEbgMS64XniMgPgMHAHuwLFDk56ngaX1dv+4eAEYARuKfJk+6CXIYOTWkArdShKQ0ccVoHO63VoSkNoJU6tKABdDMd9D1J4064ndF38HvsF+VdzgBlXxhjsvNv7Y2VUt+ISIPFoZVSf6sV/nH97R2vzw7W3w84DXwsIpOUUmvrJqkecFQ0AOeiHGdqd0HXwgd7pau/bLG7cMk6NKWBI+5ydGhKA8euWofW6NCUBrX3aaUODTSAbq2Dvidp3IImB/J1cSYDk6j7VA0QiL1L7M5G9gnCvmpU/bjWhjd3nDgReY66XWaewMMi8kqtd29OTgBmYHgjx3AHWqtDc2XaWh2a01rrYKe1OlyOPq2pC9B9ddD3JI1b4K4t/WLs614/Wy98FfYHmaRG9mkqrrXhTcYppX4qIr2UUrm1whaJyFHgAg3flX2C/d2au67V3VodOkQfrYOLdr/mmwpvTANHeHfVQd+TNG6Buxr90UA8cBx4oVb4D7FXvE2N7NNUXGvDm4wTkd8AVSJiVkr9T60oI/an8FlA7a63t4EM7BXPHUfKtlar0R7AAAAgAElEQVSHDtFH6+Ci3a/5psKb0QC6pw76nqRxC9y1e/+4UuoBwLXKn2O6yXnsS/3WedptKq614S3FAaeUUk8Dp+qFr8Lurvh4vfC1SqlX3XhqzCXr0FH6ONA6dMA1f5l1AbqnDvqepHEL3LWl/wMRmQn0F5H+jgEpS/i2q6r+9JSm4lob3lKciMhbwN564U09oY91jNytbmJQTVenNTp0lD6gdbgcHdpSH2haA+ieOuh7ksYtcEujr5Sa0EhYjog8Afhjn56S01Jca8NbigP+A5Tz7dSa2k/hYH8KX+4IH4n9aduKm86LbY0OHaWPA61DB1zzra0L0H110PckjbvglkZfRF4DKgHqPY3agENAeCO7NRXX2vDm4r4H/At4DHjcEdbUU/iPgH3Yu+TcksvQoaP00TrY6YhrvjV1AbqpDvqepHEX3NLoAxuUUn9vJHwc4EHjXYpNxbU2vLm4PKAAOC0i/kqpkmaewv8A3Ip9ioy70lodOkofrYOdjrjmL7kuQLOtUnfXQd+TNG6B2xl9EYkG7hG7ByqbUuqGWtHrsPu/bsyxRFNxrQ1vLi4I+C72p+fHgecc4Q2ewpVSm4HNjaTtFlymDh2lj9bBTkdc862tC9DNdND3JI074XZGH4jBfnE3RiDwoOP3c5cY19rw5uKc78mqsI+MddLcE7q7cjk6dJQ+Wgc7HXHNt7YuQPfTQd+TNG6D262y15URkWewD5y5FjAppX7uCH8Mx1O4s5tT035oHTqfpjRwxGkdOghdFzT1cdd5+l0VP6AIu//qY7XCnU/hjzeyj6bt0Tp0Pk1pAFqHjkTXBU0ddEu/DRGRFOwerpYBh5VS5Z2cpR6J1qHz0Rp0DbQOmvroln7bcj+QCxzUlatT0Tp0PlqDroHWQVMHbfTbEKXUo9hHxG7p7Lz0ZLQOnY/WoGugddDUxx1H73dZROR57FNgbu/svPRktA6dj9aga6B10NRHt/TbljDsa1U/2dkZ6eFoHTofrUHXQOugqYNu6bcte7DPhy3u7Iz0cLQOnY/WoGugddDUQRv9tmUt9ikyQzs7Iz0crUPnozXoGmgdNHXQU/Y0Go1Go+kh6Hf6Go1Go9H0ELTR12g0Go2mh6CNvkaj0Wg0PQRt9DUajUaj6SFoo6/RaDQaTQ9BG32NRqPRaHoI2uhrNBqNRtND0EZfo9FoNJoegjb6Go1Go9H0ELTR12g0Go2mh6CNvkaj0Wg0PQRt9DUajUaj6SFoo6/RaDQaTQ9BG32NRqPRaHoIbmv0RcRDRBaJyGsi8raITL6EfRaKyOB6YUdE5E0R+aeI+F5CGveIyA4RMYjIQBFZ2MR2n9T6nSUib4nIAsf/KSLyFxH5q4hEtniyXRStQddA66DRaC4VU2dn4Aq4H/hSKfUVuG581wJXA17AQqAK+BWQC/zLsd8CEbEAa5VSfwbSlVIPisjPgAQROQb8AhDgAvAq8DpwDPjckUYGcCew1XFsE/A84AkYgf8Dkhw3wZeAUsAbOOHY/wHgNiAJuBf4ZdsVS4eiNegaaB00Gs0l4c5GPxn4yPlHKVUpIncrpW4UkVjgEaAGeE4pdRhARP4L+EgptUFEPgL+DAwVkfeAYOBF4PvYb0pngP6AD1ANLAE2AfHAJ8DdwB7H4acDccB+IBYoAvYppRY6jpuK/cb5BfAVIEqpGhHJAaLbvmg6DK1B10DroNFoLgl3NvqZwAhgGdhbN7XilONbsN/salNcb5vdSqm7ReQF7Dc2A/CFUurfzh1E5FHgBuw3tBxH8CLgUeC4Y58NSqk/1NrHmT5KqRpHWIWIGIAax3cf4GTrT73LoDXoGmgdNBrNJeHORv9PwCsicjX2bsQPgQ9E5G3sLZJfYu/SXCgiZ4B/N5mSnd8DrwAPA6+JyETAA3v35L3Yu0mXA34ASql1IvIj7De6ZcBbIvI7IBB7y+qciPwG+DvwmOMYqx2tmj860jUD/33FJdF5aA26BloHjUZzSYhSquWtNBqNRvP/7J15fFTV3f/f39mykJ1A2CJLQkgwCEhkEUSq4m5b1z6trWurtn2qtdrW2sfHpe51ebTW1mqtrVr3Fmvlp2gBQSsgVATCKkICRLZsZCGZSeb7++PeGYZJMoGQIZnJeb9e85o793vvOeeez9x77tm+x2CIeWJ29L7BYDAYDIbDo88W+iIyQkQeCtv3nIikHGG4fTZPDxejwaEhInKYx7/e+VHdQ3hc9jS+c48wzLjSz2DoTcRyn/5B2FOFnsSaWrQaeAurb3IfsEpVn7L7GV1YfY9/B04QkQeBXOBbIWHNBs7BGrn8BrADuBTIwep//BT4I7AFOAs4Gasf9d/AJyIyBmtks19VbxaR5+w4i4BFWIOWXKp6Y7TyoycwGnQfIjICeBErD1+y57X3B95R1Tkishp4DmsA39XACcA1QGB0vgd4CmuwXpWq3iUiH2Bd+xhgoX3O+6r6Yki8P8XKmxpVvU1EHgdaALeq/khELgOmA/uBm4FsEbkbmGDHHwhnOHAT1gDCzcAfgFux+vk/U9WnReQ+2z4GS8sSrJkDn4rIbixNBwI/Ab4KfMWO90usMQDFwCWq6j2SvDYY+hLx9EZ9OvCJqt6kqs9hzf19VlV/CJwmImMBr6reqKqP2udsVNWfAbuAQSFhXQ/UYD1cJgPNWIOXdgHfwRq5PE9V/wfYaZ/jAB5U1fewBlPtB6aJyEDb/nusedIjVfWnWIVcvGE06F7Wqur9WNeRgHXtl9q27ar6MNbUuQnA94HvYl0jWFosUtUfA8NFJA1oVtXbgI+wRspfBXwtEJndwjJOVX9kF/jFQLWq/gTYa/8+X1WvVdUfq2oL4LM1eAarkA7wAzvdlcA4rBkCLqAKuERE0oEcVb0F60UtwCu2z4BWLD3dwGm27V1VvQ6Yace5DGu6osFgOETipqZP2ylJwoGpSO3Z4cCUpSash2oAB3C3/VBDRH4D/NoOI+CsJBB24LtRVVtEpD8wXlWvEpE/YY2eBqu2m2Z/xytGg+4lkDffwRpxvxR4097XYH/7sPLNj1Ujb7b3h+c9HLjuZmCfqmpYU3p7eoXncXiYofplhOx3AM+r6ioAEbkQ6yXmLyKyIEKYgfC+r6pfE5HLOVg/gD32t5eD/zMGg6ET4qnQnwf8VkQKsByFvAQ8KiIXAAtUtVREkuzm5TIOeBRrj8eBZ0SkClgOLMCaTrQrJK5nRGQ0MATrgRegBkgWkZux5jr3JYwG0eHfWJ7rpmMVdO3xFPAgB1o95mFNnRsHbFPVfZ0NDVDVOhFZKyKPYnUJ/EpErrX1SlLVNSLyloj8Fuul49YIwT0B3GtPEawDngfuF5HBgFNVa0Vktz2V71gsjUNZKyK/xOqOeT9iwg0GwyFjpux1ERG5Cas5ep+qGtehPYDRILaxXwZPAAYAPzB98wZD9DGFvsFgMBgMfYR4GshnMBgMBoMhAqbQNxgMBoOhjxDThb6IvHWU4ok4AkpEHhOR5EjHdHDeKBH5oxy83vgkEbm2K+k8msRB3vcTax33p0Xk0jDbLBFZLNa677OMJm3iMfeDwRCjxGyhbzsv2XoE5w8TkQUicqOIvGKPKn9ERH4rIneLyCAR+beI3AIMEpEPRORndkFxrYi8LiLFIpIFqKo2ikimXVA8IiKndJYGVf1CVa8O27cCOKmr13U0iIe8x1op7nVV/R6W45dQFGvd90Ss+fBGE3M/GAxxQSxP2Tse+E/gh4icg+XRa7OqvmDvGwfcF3beVaq6GxgPzFHVx0TkRazVwP6qqsvtmsZE4GVVfVxEzgL+Zh87B2tVsxqs9cIHAmvtsNOwpog9pqrz7TTkA/8dloa7VLUqwrU1iUiOqu6KcExPEg95PwzLayBYjmBCWayqH4hIDtZqc5diNDH3g8EQB3S50BeRIapaYW8PVdUd3ZesQ2IS8Ergh6q+LZab0R+E7FsNdOQHfDzwt8ChWHOFHxPLfWkjlpezObZ9AvB3EXEDlfaSoMVYD7vJWA88sDzQnauqjSHxCIefz9VYD8ze+pCLh7zfjlXwrySsxUvtNd+xdEgI2TaaWJj7wWCIUY6kpv9d4C57+wbgZ0eenMPieKC/iLQArwGLgZ8Cj0Y86wD5wEYRycZyaLIAyz94I1Zt6KfAhtBjgeOAdfa+EapaLiIZwJn2PgfWuub7gF+oaquqbqJtzQYAsTzH3QNMFJFfqGqgFjYUy098byXm8x6rgHvCrhG/BSAi07BqtDuBM7A8zD1hH280MfeDwRDzdGmevoi8BowFSrFqBaWqelfks6KLiNyG5eTjHVWdexTjFeB3tk/w7ggvBXiou8I7GsRL3keIx2hy6PH2+fvBYOjNdNk5j4gco6rl9na2qu7t1pTFECJyErAirBmzq2EVAK2quvnIUxb/dGfeR4jDaHIYmPvBYOi9RCz07X66cwEPQGhtXkQeUtWb7e17VTWSH26DwWAwGAw9TGd9+j/BGr3sa8eWGbKd3W0pMhgMBoPBEBU6K/TXqOqaDmyviMgbWEt6PhMpELuJ7lasKUFz7H2XY70s9FPVu0TkDqzVuHar6vOHcQ2GQ8To0PMYDQwGQ0/SWaH/FRGZhb1sqapeEmJbAzynqm+JyIRIgajqRhF5joPX256gqjeKyP+KSCbgV9WHxVrWs82DTkSuAa4BGDt27KTXX3+dV199FafT2W6cra2t7drEdibW0tKCy+VCRIL7Anafz4fH4wn+Dv00NzeTmZlJcnIyaWlpZGRkkJ6eTmZmJhkZGWRkZBwUXhc4opMjcaQ6hGtQWlrKCy+8wObNm3G73QD4/X6cTid+vzXrrbW1FYfDgcPhaGNvaWnB4/EEj1FVVBWn00lrayutra0kJibS2tqK2+0O2vv160ddXR15eXnk5eVRW1vLmDFj2LlzJ3V1dQwcOJDdu3eTlZVFWloaW7dupbCwkPLycioqKjj11FNZsWIFAwYMIDExkW3btlFcXExZWRmTJk2Kqg7RuBdKS0v51a9+hcPhQESCeS8iqCoictD/PbAvkOctLS1B/VwuF36/P2jfv38/SUlJwbx3uVxBvZqbm5k8eTITJ06kqqqK+vp6Ro8ezbp160hNTaW+vh4RYcSIEezZswefz0deXh5lZWWkpaXhdrupqqpi6NCh7Nq1C7/fz9ChQ3G5XMH7L1o6GAx9lYiFvqqeJyLH2tulYeZ7gCqs6U5XAdd3MQ3awXd4Wv6ANYWIkpISHThwIE6nk+zsbBISEmhqago+1Hw+H1VVVfTv3x8Aj8dDc3MzTqcz+JCrrq4mNTU1WMh4vd7gA7O5uRmPx4PPd6BXo6WlBYfDgc/nY+/evYE0BR+ggbERo0ePxuVyMXLkSJxOJwMHDiQxMZHKykpaWlpITk7G6/UyfPhwysrKSE1NJSEhgX379jFy5MguZmG30KkO4RoADBo0iK1bt5KZmRl+OACVlZVBHdqzZWRkBMJu811ZWUlycjKqit/vx+/3o6rs378fr9fLxo0b2bBhA6rKhx9+GHwxczqd3HLLLcGXrxEjRhCa1tTUVGbNmhVMR15eHgD19fWHkk/RpEv3AoDD4SAjIwO3201rq+VryOVyBf+3e/bsISMjA4fDgdfrPehlwOv14vF42vyXVTVYwAfyNhAmgM/nY9WqVaxatYqSkhLGjh1LYmIiY8aMwe12s23bNnJyckhMTMThcARfPNLT00lMTMTpdJKamorL5SIlJSX4Qhh4QTQYDN1PxEJfRB4D9trb16pqaMFeDdQeSiQiMgi4CEgSkXTgXWClWOuho6rVIuK0f/8nQlBBVq9ezWuvvUZ+fj5utxuv1xusQba2trJ582by8/NxOBy43W5aWlqCNX9VZd26dUyYMAG3201SUhKtra0kJCSgqnz00UeccsopNDU1BR94gZeCHTt2MGvWLPLz80lJSSE1NZWmpib27t3LypUr2bx5MzNmzCA3N5fS0lK+/PJLnE4nSUlJVFRUkJ6eTmtrK9u3b6exsRGXy0VCQgIJCQlRL/SjocPKlSt56623KC4ubtMqArB3714GDBiAHf5B5+7du5ecnBycTudB5wXyfMeOHQwbNuyg8EJr+zNmzMDn8+Hz+Whubqa+vp76+noqKyt5+eWXufDCC0NrjMCBF4D2iGTrLqKhQX19PX//+9+ZOHFim5enQK1/7dq1HHfccTgcDpxOZ7BwdTqdVFRUMHr0aBISEoL7A/m9c+dO0tPTAavVJlBwu91umpubmTBhAsXFxcybN4+VK1dSVFTEzJkz8Xg8jB49OpjGwEt0TU0NtbW1fPbZZ6SkpOD1eoMa+nw+vF4vCQkJfP3rX++G3DYYDOF0Nnr/YVW9yd5+UFV/FmK7DrgMy7PZMlV9KNqJDVBSUqL33XcfDzzwACkpKeFpBqChoSFYSwwQeAACNDY2kpCQEKxBBmqRqkpTUxOJiYnB8EILo927dwcfXlVVVbhcLoYNG8awYcOora3lvPPOw+12U1BQwLnnnhtslq6qqmLFihWcdtpp7dZktm7dGih0YqI5s6SkRJcvX86NN97IRx99REZGRjD/Qj+NjY0kJ7e/9kqg6TiU0MK9ubk5WHMN1PZVFZ/Px/r168nPz+cb3/gGZ599drC52Ol08txzz9HS0kJ+fj5nnnkmWVlZXbnEmNFh4cKFnHzyyWRndzyetrq6moyMjIP+54E8ra+vJyEhIdilEnpPNDQ0BDVoaWkJdsu0tLTQ0NDA66+/jqpy+umnU1ZWxsKFC2lqaiI7O5vq6mr69esXLNgDOrrdbqqrqxkyZAgulwun04nb7Q5u19XV8Y1vfCOQ9JjQwWCIFTrr03eJyO32dvDpbDvg2ATMxnpxOOrtouPHj+ekk05i6NChB9UEA9vl5eUMHz78oHMCzZd+vz9oD6+dighLly5l0qRJwYdg6GfHjh2MGDGC1tZWGhsbg7VLr9fLH//4R6644gqqqqrYunUrzz//PKeccgojRoygf//+FBQUdNh0eTRqmdHgzDPPxOVyMWLEiOC1hdbW161bx9ixY4PHh/Ypl5aWcuyxx3YYdkf2lpYWtmzZQm1tLW+++SYPPvggF198MVdddRVTp07l5ptv5v/+7//44osvmDdvHiUlJeTn53f/xfcSUlJSOPnkk8nLy2vTWhLYXrt2LQUFBQd1lQTYtm0bubm5wd8BHVWV8vJyBg0aFOznDyAiPPLII/zzn//krrvuYv78+Zx22mmMHDmS0tJS1q1bR319PSNGjCApKYns7Gyys7PJysrC5XIxd+5czj777HavZ+vWrdHJKIPB0Gmf/g3t9emrqorICcACPeCn/KiSnp5Oeno6kydPbtdeU1PDpEmTgoPIQj8iwj//+U9mzJjRpqbf2trKnj17OO6449r0JQdqRYMHDw62JPTv35+Wlhb279/Pt7/9bW6//XZefvllli1bhsfj4e2332bkyJHBwj/emDx5MqWlpZx55pnBrpXQvKyurubEE0/E5XIFm4WdTicul4u6ujrOOeccWlpagjXJ0O2qqiqOPfbYYNNvS0tLcHvPnj1kZmYyYsQIBg4cyIcffsi3v/1tPB4PJ5xwAo8//jhPPvkkW7Zswel0smfPHqZOndpmkKXP52Pbtm1s3LiRiooKrrrqqh7KySMjNzeXs84666D/c+inurqamTNn4nK52rx4zp07l1mzZtHc3ExjYyNNTU00NTXR3NzMvn37yM3NDY6VCL2PioqK+PTTT9m1axfFxcUsWrSIk08+meLiYoqLi1myZAlTp05tN70dFfgGgyG6dFjoi8iDwIgDP0XDRu9PB/4hIvux3gMuCQ8j2qSmpgYHFbVn83q9HZ6bnJzMvn372rWlpaW1ecAFfmdmZpKenk5aWhoOh4MBAwaQmppKY2MjtbW1TJkyhQcffJDrr7+eNWvWcPbZZ7N06VJefPFFGhsbSU1NbTfOESNGHDS4LFZoamoiPT2d2tr2h3ekpKTQ0NAQLHxCm48TExP5/PPPgy8DoS8GiYmJDBkyhOHDh7exNzU1sXPnTi6//HJee+01tm3bxrHHHsu3vvUtFi5cyD333MPTTz/NWWedxTvvvEN5eTmqSk1NDbm5ubhcLr788kv27dvHvn37SE5OZujQoZSVlR3l3Os+MjMzqamp6dCekZFBdXV1sIk+gKqSmZnJ3r17cbvdJCcnk5WVRVJSEomJibS0tDB9+vTgS1fop7CwkD/96U8sWbKEK664gqamJpYuXcqUKVMA2hT4ra2trF+/noqKChoaGoJjBcJfxOrq6uLyBdlg6A10WOir6s8kZPU8ERkWsNnN+39S1b91dP7RID09vcPpcenp6e0WNIFPQkICDQ0NbVoBnE5nsKAKPNxC40hJSaFfv35kZ2fTv39/Ghsb+fLLL6mpqSE7O5vjjz+eN998k7KyMgYOHMh7773HpZdeyieffMLmzZspLi4mLS2N1NRUUlNT6devHyLCCy+8cLSyrVsRkYg6DBw4EJer/b9ZdnZ2cEqe1+ulqampTcvL+vXrDxpRHpidkZaWRllZGZdddhlVVVW89NJLLF26lGHDhpGbm8vvfvc7fvzjH7Nx40bKysrYtWsXDoeDhQsXcswxx5CUlMSwYcPw+XzU19cHa7ixSlJSEnV1de2+qDocjuALcmiBH06gFaWurg4g2LpVXl6Ox+MhISEBj8dDSkoKHo+H4cOHBweoBrpxmpqa+Oyzzxg/fnww3KqqKlavXk1rayuFhYWMHTuWsrKyDgt207xvMESPzvr0r+bASnrXY6+kZzfvF4vIx9gj+KPp+7w9AtN8Ro0a1a59165djBkzpk2NPfBZsGABJ554YpuXgtbWVvbu3cvEiRODA4tCC7QFCxYwduxY9u7dy6ZNm4IPSBEJ1lDPO+88brvtNhYtWsSCBQuYM2cO3/ve95gzZw4FBQXU1dVRW1vL9u3baWhoQFXbDGiLFTIyMjrVIZKtsLCw3RcvEWHhwoXttn6oKu+88w4rV65kyJAhZGVl8cMf/pBPPvmEhQsXcs4553Dfffdx5ZVXcv755/Pss8/S2NjInj17EBGys7NpaGigvLyc/fv309zcjMPhCGoZi+Tk5HDSSSe1ebkN/LcrKyuZPn16u74rQgaRHoSqsmDBAsaNG4fX66W5uRmv14vX66WhoQGAoqIiVq5cybhx4ygsLAw2+a9bt46dO3fS0tJCVlYWU6dOJSEhIRh2T8+iMBj6KpGa918Dxorlfx+sFfVCGYk1Vz/AUe0MdTqdFBYWBpsIwyksLOywKR0IzqMPjBwOpaCgIDh6v73zUlJSSElJafNw+vzzzxk7diyrVq3i0ksv5Xvf+x6vvvoqr7/+Oq+//jotLS2kpaWRlpbWJtxYrd0kJiZ2qkMkW/jsi1A6eviLCGeddRZr167l1Vdf5corr8TpdHLCCSeQl5fHb37zG2bMmMEDDzzAvffey+mnn878+fPZv38/tbW1lJaWBueI9+vXL+ibIZbnhwcGUnZ0DaNHj+7QkVWkfB41ahTJycntzsBYtmwZhYWF/P3vf+e73/0uK1eu5Pjjj2fixImsWbOG+vp6zjvvvC5fk8Fg6H46fMqp6sXA6ap6if25M+yQa4A37M81UUxjhxxJbaGr50ayTZkyhX379pGQkEBeXh4Oh4MXXniBGTNmsHv3bjZu3MiLL77IunXr2L9//2GltzcTjbw8FPvYsWOZPn06f/nLX4LjN7KysnA6nRx//PE89dRT7Nixg1GjRpGTk8PgwYNxOp2MGjWK3NxcPB4Pfr8/OCAw4E8gFjnSvOzKeXl5ecEukvLycqqqqoIOrYqLixk3blyX4jQYDNGjw0LfHsj3qIi8KiKvicirYYf8HsiyP09FikREjheRu0XkERHpZ+/7vojcKCJv279fEZEfi8i0I7qiHuS4444jKSmJsWPHUl9fz0UXXcRDDz2Ey+WiuLg4OONg8eLFvPLKK/zzn/9k/vz5fPzxx3z++edRTVu8alBUVBSsbQYGE55zzjn4/X6uvPJKbr/dmnF69tlnU1NTg9vtprGxkd27d+PxeCgsLOSMM87g1FNPZfbs2VFPbzzpMGrUKJKSkhgzZgyffvopw4cPZ8WKFUF7LL/IGgzxSqSa/s9CavkXtzM6f5+qPm8vCFLdSTzfBO4A5mDN7UdVfwesBZ6zj9mJtXJfu96CROQaEVkuIsv37NnTSXQ9x5QpU6itrcXlcpGRkcG1117L5ZdfzrRp0xg/fjypqalcfPHFDB8+PDiALCUlhVWrVkU7aXGrwdSpU0lLS+Pf//4327ZtIzMzExFh4sSJ/OMf/2Dt2rU4nU7OPfdcfD4fQ4YM4cILL+Sss86iqKgoONDwKBVScaNDwMV0UVER77//Pl9++SVNTU1tWrEOlYaGBtavX8+nn37azSk1GAwBInZiisjDIvKiiPxVRP4aZq4WkT+KyB+B9ue+HUx7/sS/DvwdLJ8Aqno78O12T1b9g6qWqGpJb26Gzc3NpbW1lYKCAiorKxk+fDiDBw/mzjvv5IMPPuC4447j448/Zty4cXz1q18lIyODzZs3dzh9sJuJSw1EhJNPPhmHw0FZWRmrV69m5MiRlJWVccstt3DLLbcAlu/9oqIiJk2a1OGYjaNEXOjgcrlISkpixIgRVFRUsH37dkpKSvjkk086PfeLL76goqKCZcuWMX/+fObPn8+aNWtITU2lurqzOoTBYOgqnY3e3xlwwxuKiAwA7sXy0idAQvgxYbyMVbtJBraIiAM4Btimqi12mL+0w/nscC6gN1JSUsK6detoaWlh5MiRnHvuudx3333s2LEDv9/Paaedxpo1a0hPT2f69Ol4vV7mzp0b7WTFtQbJyckUFBRQXV3Npk2buPTSS7n//vs5/fTTefzxx1m0aBEzZ85kwoSOF4S84447uOOOO6Kd1LjSwePxkJyczJgxY97Zn/UAACAASURBVCgtLWXGjBk4nU727dt30IDVpqYmtm3bRkVFBa2trZSXlzNlyhSKioraDLgNXejKYDB0L5353n8T+BBoAFDVJ+39f1DVa0KO+72qXhfltAYJ+H3vragqzz77LElJSdTW1pKTk0NOTg4zZszg5z//OfPmzeOLL75gypQpjBkzhuuuu45HHnmEZ555BmLE13hv1WDx4sVs2rSJK664gkceeQSPx0NWVhZPPPEEH3/8cYf+BGpqasjMzAxdq8HocAjMnTuXDRs2sHz5crZs2cI999zDiSeeyMKFCw+aFZOYmEhubi6DBw/G5XJ1OB2zHWJCB4MhVuhsjtIjwFKs6XprQvaHezExr+YhiAhjxowhMzOThoYGhg0bxoYNG7jhhhu45pprWL58ORs2bOCyyy6jsrKSGTNm8Nxzz/V0suOCE044AZ/Px5YtW7jooouoqanh61//Os3Nzbzxxhttjl+/fj0//OEPg74EYtlBT08QmKVSUFDAypUrqampweVyMX78eJqamvjKV77CKaecwoknnhj0hghmkJ/B0FN0Vuifa3/OAWaJSGDVkhQRyQKwvzuebN1HmTp1Ktu3byctLY3169fzX//1X+Tn57NixQpeeuklVq1axUknncQLL7xAZWVlcA10w5GRmJhI//79gw5nHA4H7777Lg888AC33nort912G36/n7lz53LmmWcya9Ys+vfvz5o1a4LnGw6dkSNH4nA4SEtLIy8vj82bN7N582YGDRrE2LFjO2xZMYW+wdAzdNan7wBewhpwdBnwCywvfXcAD4tIEtBo/zaE4HK5yMnJITU1lc8++4xPP/2UUaNGcfbZZ1NZWcnq1atZunQp//nPf8jIyODSSy/t6STHDXl5eVRUVNDS0sL48eNZtWoVt956K8OHD+fuu+/mlVdeISUlhRtuuIE5c+YEC/rA9D7DoePxePB4PACMGTOGDz74gIkTJ1JQUGAKdoOhF9JZTT9FVZer6gqsgUUVAKparqpXqup/qepVqhq7K5VEkdmzZwdXeWtoaKCqqgq/30///v2ZNWsWF1xwAWPGjAk2gxq6h8LCQkSEzz//nHPPPZfW1lZKS0t59NFHAXj22WdZsWIFl19++UE1+yuuuKKHUhzbBBZCKioqYv78+fh8voiLXXWEqrJ9+3ZWrlwZhVQaDAbovND/l+2Y5zVgPtCjC+zEGklJSSQkJHDsscfyxRdf4Ha7eeedd4KroTmdToqKijjzzDPp379/D6c2fkhKSsLtdrN9+/bggkDvvfcexcWWR+kZM2a0aXaura1l/vz5PZHcmCclJQWn00n//v3Jzc1lx44drF+//pDO9fl8bNy4kfnz57NgwQIaGhqorKyMcooNhr5LZ837f8MeuQ+8G5hSZDh0Zs+ezbx58wDYtGkTI0aMYN68eeTn5zNhwoSgr/RBgwb1ZDLjjqysLPbu3YvX6+XSSy/lySefbHf+t9frZdmyZezdu5fdu3f3QEpjn+HDh7Ny5UqysrIoKChgwYIFDB06lJaWFoYMGXLQsYEVE9esWYOI4HK5GDVqFLNmzQreC6EL8xgMhu6ls5r+U3TgaldEbrWd9rzUjuMeg83AgQNpbm7mggsuIDU1lUGDBlFTU8O//vUvHn74Yf7xj3+wcOFCNm7c2NNJjSsKCwsB2LhxI9nZ2TidTt59991gv72qsmrVKubNm0d5eTl79uzpUpO0werL93q9iAiFhYW8/fbbFBcX88UXX1BTU0N1dXXwU1VVRVVVFV6vl1NPPZWTTz6Z3NzcgxYKMmMBDIbo0VlNf5/tZhcRGR9mE1X9VnSSFV+ceOKJLFmyhJycHE477TTAKnRKS0v517/+hcvlCi5Vauge+vXrh9PpZOfOnRQXFzNjxgwWL17ML3/5S7Zt28bq1avx+/1s3boVVaVfv34djjQ3RCYzMxOHw4HL5SI7O5sBAwawdu1aPvzwQy666KJ2z2lv1b4AR8lJksHQJ+msph/J1W6eiFwsImeLyNmRAulgkZGn7UVFzrB/3yEiN4nId7p8Nb2UoqIidu/eTXl5OQsWLKC8vByv10txcTHXX389M2fOjHqB0xc1SE9PR1XZv38/M2fOxO/38+c//5nPP/+cTZs2BX3yB5qXj0YNM151cLvdiAjZ2dnk5+czZ84cHnvssQ6Pj+SX4s47wxf0NBgM3UVnNf17sNyFwoG+/QAfYLnhTTqEeL6JNd3vRKxFRuZgLSqSCjhFJBPwq+rDIvIo8Hx4ACJyDfYSvsccc8whRNm7uPrqq3nmmWdITU1l+fLl1NXV4ff7SUxMJC0t7Wj0Y/Y5DYqKili6dCnr169n4sSJZGdns2PHDnbt2kVWVhZTpkxh/PjxJCVZf+GO1qLvZuJSh8TERHw+H9nZ2RQUFPDqq+GLch7MnXfe2aY27/P52LlzJ5mZmVFMqcHQt+ms0P+1qv4EQETuB24JsX0IXIc1h/8PhxDXQYuMqOptdri/BT4Ot7c5WfUPgXhKSko69h3cSxERxo0bR0lJCSUlJYDVxF9dXc327dspKzsqsx77lAYpKSmoanA0+OWXX85DDz3EFVdcwdChQ9scfxT7kuNOh5ycHHbv3k1rays5OTkUFBTw5ZdfMnv2bE466SRmzpzJlClTgi9YTqeTiooKKioqqK2tpbq6ml27dlFXV0d9fX0PX43BEL90VuiHdryFe927HfilvX03cHmEcNpbZOSHWAMEy1W1WkScInIT8J9DTHvMMXXq1IN+iwhZWVlkZWXR2NgY7ej7pAZpaWk0NDTQ0NBAv379SElJabfAB4Je/KJMXOqQl5fH1q1bycjIICUlhYsuuogPPviAG264gUWLFvHzn/+cNWvWMHLkSMaNG8eQIUP47ne/S0VFRXA66+jRoxk9erRZcMdgiCKdFfoficjfAD/wVpitWVW3AYhIxAW0bec+K8J2/ybsmDs6TW0cE/5C0N30VQ0KCwv59NNPWbt2LSeccALTp09v97jGxkY+/vjjqBf68arDsGHDaG5uJjk5mZycHD7//HPGjh1Lv379OPPMM/n6179O//79KSsr4+OPP+bll1/mvvvuY/To0eTn55OVlRUM64UXXujBKzEY4ptDmae/XlU/EZHw6tFbIvI6VhOkuUsNvZK0tDR8Ph+1tbVA+y9Xu3fvZsWKFfTk2vSxTlJSEg6HA6/XS2ZmJi6XiwkTJjB9+vSgm16wpvedfvrp3HHHHcb1tMHQA3Q2cukx4Ap7+9bAThFJAN7HatK/AngvCmkzGLqFlJQUEhISggV/KBs2bGD9+vWcccYZ5Ofnt3O24VBJSEjA4XDg9/sZOHAg2dnZvPrqq+3mu8Fg6Bk6K/TrgF32duiao7cCT2A1Sz5hfwyGXknAF/+6deuC+1SVJUuW0NzczMyZM3E4HMYpzBGSkZFBQ0MDbrebqVOnMnToUKqrq3n11VfZsmXLIYdzww03RDGVBkPfprNCfw9woog8hNWvD4Cq3g68YS+2cyXwZhTTaDAcERkZGTQ1NQVHhbe0tPDee+8xbNgwjjvuuOBxW7du7aEUxge5ubnU1taiquzcuZOZM2cGnfO8+eabzJ07F7/feoyEr2gYOGfx4sVMnjz5qKfdYOgrROzTV9V77YF8DuDzwH4ROQf4pogEpgt9E1PwG3oxycnJiAhLliyhsrKSGTNmkJ6eHrTv27ePJUuWmNr+EZCbm8unn36KiODz+TjmmGMYPHgw3/72t3n77bfZtm0bTz/9NDNnzuQ73/kOra2tlJWVUVZWht/vx+l08uWXX7Jr167OIzMYDF2iw0JfRMYBs7CmGJ2D5ajnfNucDXiBAVgtAPdGNZUGwxFSWFjIxo0bWb16NVdccQVutxuwav3Lli3D7/ebgXxHyIABA/D5fLjdbo499lj+8Y9/MGTIEJKSkrjwwguZP38+ZWVlrFixgp07dwZXlty3bx+tra0kJiaSm5uLaq9wPWAwxCWRavp3AQ8A7wK3qWqgwEdV/ywig1X1z9FOoMHQHWRlZVFfX8/o0aNxu92oKhs2bKCsrIzJkyeTmZlpmvePkISEBJxOJykpKTQ1NTFz5kzmz5/P1KlTycrK4tRTT6W0tJSPPvoIn8+H3+9n8ODBnH/++WRkZATDMVP2DIboEanQ36aqS0Tk36r6djv2GSLyC6AWQFWf7CggETkeuADLIcltqtogIj8A+gMuVb1dRF7B8ka2VFU/7uoFGdrHaGBNK8vOzmbPnj2sWLGC0aNHc8YZZwTtR8v3PnGsQ8AZ0v79+znxxBPJz89n4cKF5OfnM3z4cI499lhyc3P5y1/+wtVXX91uGDNmzDjKqTYY+g6RBvLli8irwCgRec3eDuXXwEfAGqC0k3i+ieWFbA6Wv/HAS8IDQGAh+Z1AJh24HhWRa0RkuYgs37NnTyfRGdqhz2tQWFjIRx99xJYtW5g9ezZ5eXk9kYy41iErK4sdO3bQ3NwMWO52TznlFHbt2sWqVasA68Ug4Iq6PUzzvsEQPTqs6atqxJXzsKbw/RDrwfS7Q4jrIH/iIpKINU7gXju+G+z9TwBL2klPr/M3HoP0aQ369+/P8OHDe8Po8LjVYdiwYSxbtoxBgwaxYsUKJk2ahIgwefJkNmzYwIcffsj06dODTpIC60+Ul5dTWVmJiNDQ0MDIkSN7+EoMhvikM498kfg+8F1AgCdp5+EUQnv+xl/BaiE4HXhaRH4JJACfHUGaDB1jNICeqt2HEtc6DBw4EJ/Px8yZM3n55Zepq6tj2rRpJCQkMGbMGNLT05k3bx6qGvTUl5mZyTHHHMP48eMRETO2wmCIIkdS6MOBGkvExeA78Df+tbBj7jnCtBgiYDSw6OkpefGuQ3Z2Nj6fj9bWVqZNm0Z2djaLFi1i6NChFBUVMWjQINLS0li8eDGnnHJKu2H0tEYGQzwTacrea0A/wAO0AqqqZ4Yc8jvgWawC33jkMxgMeDwekpKSKCsrC7aqzJ49m82bN/Puu+8ydepUMjIyGDNmTIdhHKXVDg2GPkmkPv2LReQu4E6sufg3hR1SjtUkqcBRWQzeYDD0fjIyMti8efNBXSl5eXkMHz6cJUuW4HK5EBHq6+upq6ujqemAh28Roa6uzhT6BkOU6Kx5vwAYCviA8JE1d2D15XuAR4H/6u7EGQyG2CMzM5OdO3e22e9yuZgxYwaVlZUsWrSIGTNmkJ+fT0JCAiIHeghNn77BED06K/TvAq63t38TZitV1c8ARGRTdyfMYDDEJoMGDWLr1q0dNtP379+fiRMndugB0dTyDYbo0Vmh/1VVvbkD2yUicq4dhlNEXlXVS7o3eQaDIdYYOHAgbreb+fPnU1hYSFFREZmZmQcdYwp2g6Fn6KzQ79DrnqrOEJFUe7suekk0GAyxRHZ2NsnJyXi9XlJSUti4cSN1dXWkpqZSVFREWlpa8FhVpb6+nurqaqqqqqitrSUlJYVJkyb14BUYDPFLZ4X+r+lgWp6I/A8w3N7epqp3dX/yDAZDrOFyuRgwYAAjRozA5XJRVVUVtH322Wc0NzdTX18f9Lffr18/kpKSSEpKIjExEa/X21NJNxjins4K/Uhe97JU9XsAIvLrSIF04G/8J1izAhRrvMCDQDWwXFXfPdwLMUTGaNA76Cs6tLS00NTUREJCAnl5eXg8HhoaGti2bVuwdh9Y2rihoQGwXhZSU1ODq+8ZDIbup7NCP5LXvQwRuRzrQdXZXfpN4BfAiVj+xucAuap6o4g8CowHVqnqX+zfbR50InINcI39s15ENmAt8bu3gzgj2TqzR8N2qOe+E+YPobuIlgahaW+PWNIo1BZrOvS2vOyucKOlg8HQJzkUj3xqf8K97t2HNaVPgOsOMZzD+T745BB/4wFEZLmqtrtyRyTbkZzbE3F2I92uAfS+/IpWerqRo3ovdGaPJf0MBsOR0Vmh/zvgT/b2b8Ns56vqg4cYT3v+xreJyI+BrVg+xr9lDxqcd4hhGg4Po0HvwOhgMBh6jEhueB8G5gPf72B0/mUiciqwD8tFb4fT9TrwN/5o2O+fHlqSDV3BaNA7MDoYDIaeJFJN/3HgFOAxEUkHNqjqrSH2mapa1f6pR402Tc2HaDuSc3sizt5Ob8uvaKWnN9Pb8jKa4RoMhi4iqu0vx203O04GZgE5AKp6o217Dmtk/xZVfeBoJNRgMBgMBsOREammvwN4DbhfVSvCbE2qep2IhLvmNRgMBoPB0EuJVNNPA04GpgFJQKWq3m3b/gO8j9X8Px9AVX92NBJsMBgMBoOha0Sq6TvsTxIwAmu0cYDz7e/wEf0Gg8FgMBh6KZFq+s8BC4EFqlp2FNNkMBgMBoMhCnRY6PdmRCRNVfeJSLqq1kY4LldVt3VgO1ZVS0N+5wGDVfXDdmyjAS9QBJSr6tp2wjtJVReH7buBA46NVFUfPxRbrHAoOhyOBvY+o8Nh0t06RNLAthsdDIYY5VA88vUq7LEG3xKRvwLXYvkpD7V/NbAJnAt8L8R2fYhtJnBhyKm3AB/avtEBQh90/w0MBe7C8j4YCAcR+QhYChwvIv9R1Z+EnDfH/m7Po2GoLeaIpMMRaABGh8MiSjpE0gCMDgZDzBJzhT5wI5APDAQ2tGM/H8uLoGBNKwxlJJYjFAGywmyfq+qfRWQ28L9YfgoCfIHlA3yViGxvJz3jgTJVfSzUoKplIvLfQDGwCmv9glDbD4BJgBO4ItJF90Ii6dBVDcDocLhEQ4dIGoDRwWCIWWKu0FfVO0XkdqyH1Zh2DrlXVTcBiMj6cBtQq6peEbk/zPa2iHhU9T0RqQ+zvQHstrc/CEvPMhH5Eqsm1UGS9Tr7YRdOMrAYaO7g3F5LJzp0VQMwOhwWUdIhkgZgdDAYYhZHTyegi/wf1hKkbfovVXWTiNwuIv+L1dwZatuD3YypqvvDbGtCbB+H2baH2Ja2k54Zqhq+9HCAESJyF3BMO7ZtgBs4vh1bLNCuDl3VwN5ndDh8ulWHSBrY+4wOBkOMEqsD+b6K9cJyuqr+oB17OlYT4XfCmxhF5B1gM9Ac1t8YFZttH6iqu9vZ7wSGYOlQ3vmV9y4i6dBVDTqzGx3aEg0doqWRbY9LHQyGWCBWC/2TAT+w3q6xhNsjPQQzsPozy1W18ijY7gd8gFtVbwmzPQOsJkZHK0fSoasadGY3OrQlGjpEUaO41cFgiAVirk/fZgowCijH6psMpxbrIXh7O7Y7sFwM1wG/Pwq2Har6GxH5UTtpWaSqf2lnf6wQSYeuatCZvas2o0Nb7qBreXkk58azDgZDrydWC/1yVX1QRK4LN4hI6DKlCoQ3L+6yP8PaCTcaNhGR3wNr2rFNE5FJQGt7zaAxQLs6HKEGndmNDm2Jhg7R0iiedTAYej2xWuj/t4icARSISEHoA0JVbxQRF9bKgG36DYH1wAXA34+S7S1gP9ZaBUFEpARYgDVSOfb6WCza1eEINejMbnRoSzR0iJZG8ayDwdDriclCX1VndHLIr4EarAdd+CjiscBN7Q0kipLtKqyH3w3Aj0P2/wRYC+yLdCG9mU506KoGndmNDmFESYdoaRS3OhgMsUBMTtkTkSdE5BEReaSDQ3YAW4B+7dieAq4UkcfFcjcabdseoBKoEMt7WoDHgWysPtcOXQn3ZjrRoasadGY3OoQRJR2ipVHc6mAwxAIxWegDH6nqTyL0+9VieSNrr9ZwGpb3sneBa46CLRO4HGu1wmDNRlWXqOqPVfXPqvrnDq6jtxNJh65q0Jnd6NCWaOgQLY3iWQeDodcTc1P2RGQY8Eesvr8WVb2gnWNmqerCDs4vUdXl9naaqu6Lsu0Ge9MHzFfVcM9oMUlnOnRVg87sRoeDiZYOUdQoLnUwGGKFWKzp5wJ/xXIF+mYHx3xLRC6WAwuOhHKGiLwiIq+HP8iiZEsHFmHVfL7d+eXFDJ3p0FUNOrMbHQ4mWjpES6N41cFgiAlibiCfWm5B27gGDePF0B8iMkwt16EADcA7tF2AJFq2FKAay6/4xk7SHTMcgg5d1aAzu9EhhCjqEC2N4lIHgyFWiMWafqeo6geBj73rFAARGY/Vl5gDfCP0nGjYbJ4HLgZeIOwBHM90RYPO7EaHw6e774UjPZc+qoPB0FuIuZp+Fwms3T0BqLC3w+cQR8MG1hrmK4ANqtqXVw87FA06sxsdjpwjvReO9Fyjg8HQg8TcQL6uICLnqepbPRj/NOB3qjqhp9LQ0/S0BnYajA5GB4OhTxPXhb6InKGq7/ZwGu4BBgEPq+rankxLT9AbNLDTYXQwOhgMfZ64K/RF5FksD2QCTFDVr/Rwev6ANWgJVe0To5V7mwZgdMDoYDAYiM8+/XWq+msAEflOTycGWIU1J7kveRnrbRqA0cHoYDAY4q+mDyAimapa3dPpABCR47CmKY1X1XDf53FLb9IAjA49nY4AfVUHg6G3EK+F/u1Y04b+pqrLejo9fRGjQe/A6GAwGEKJy0IfQES+AtyvqlN6Oi19FaNB78DoYDAYAsSlcx4ReRTIAqb3dFr6KkaD3oHRwWAwhBKXNX0RGQtMwrq+v/R0evoiRoPegdHBYDCEEpc1feBqrFHCn/V0QvowRoPegdHBYDAEiddCvw5wAvHXjBE7GA16B0YHg8EQJB7n6QNsAYrt7VU9mZA+jNGgd2B0MBgMQeK1pu8FTgTMgh49h9Ggd2B0MBgMQeK10B+kqtcBQ3o6IX0Yo0HvwOhgMBiCxGuhP1xErgKO6emE9GGMBr0Do4PBYAgSr33664HVWJ7IDD2D0aB3YHQwGAxB4rWmn62qn2A5JTH0DEaD3oHRwWAwBInXQn+hiDwEzO/phPRhjAa9A6ODwWAIEpce+QwGg8FgMLQlXmv6BoPBYDAYwjCFvsFgMBgMfYS4LPRFxCMivxGRJ0TkKRE5+RDOuUNEisP2fUNEnhWRJ0XklyIyS0S2iUg/EUkUkefs41pEZJK9/bL9fYOI/FFEnhGRQVG4zF6N0aB3YHQwGAyhxOuUve8Bc1X1/0Hwwfc14DwgEbgD8AF3A7uBv9vnXSMiqcAi4B/A2ap6eSAMLM9m84EbgYdC4nsfuAn4Vsi+Wap6vohMBr5rx9WXMBr0DowOBoMhSLwW+scCrwR+qKpXRC5T1QtFZDjwI8AP3KWqmwBE5DTgFVX9SEReAUqx5jeHhgHwCTAByAmJrx5YLCJnh+z7g4g8CVQB2dG4yF6O0aB3YHQwGAxB4rJ5H+shNSnww66ZBAhMVxCsh10otSHHbObAQiXhYdwH3Bp27tNYy5g6AVT1/6nqD4AFWA5S+hpGg96B0cFgMASJ15r+08AjInIe1oPnZeAFEXkKSAZ+hdWkeYeIfInVfHkQqlopIvNE5I9AE7Ad+Ni2bRaRRiAh5PgWEXkR+DWAiHwHmGYfc33UrrT3YjToHRgdDAZDEDNP32AwGAyGPkK8Nu8bDAaDwWAIo88W+iIyQiz3pKH7nhORlCMMt8/m6eFiNDg0xB41dxjHvx6ttHQWl4hcISLnHmGYcaWfwdCbiJs+fRFxAU8CdVgjjd8C/g/YB6xS1adE5NdY11yONTXpBBF5EMglZIqRiMwGzsFamewNYAdwKdYo5WeAT4E/AluAs4CTsfpK/w18IiJjgFGAX1VvtucwlwNFWFOgjgFcqnpjtPKjJzAadB8iMgJ4ESsPXxKRa4D+wDuqOkdEVgPPYQ3Suxo4AbgGCIzA9wBPYQ3Iq1LVu0TkA6xrHwMstM95X1VfDIn3p1h5U6Oqt4nI40AL4FbVH4nIZcB0YD9wM5AtIndjjeK/JiSc4VhT9wRrIOAfsAb8ZQCfqerTInKfbR+DpWUJ1sJAn4rIbixNBwI/Ab4KfMWO90vAjTW48BJV9R5JXhsMfYl4eqM+HfhEVW9S1eeAbwLPquoPgdNEZCzgVdUbVfVR+5yNqvozYBcQ6jTkeqAG6+EyGWjGmtO8C/gOMBuYp6r/A+y0z3EAD6rqe1gDpvYD00RkoG3/Pdb85JGq+lOsQi7eMBp0L2tV9X6s60jAuvZLbdt2VX0Ya0DdBOD7WHPgf2/bTwcWqeqPgeEikgY0q+ptwEdYg/GuAr4WiMxuYRmnqj+yC/xioFpVfwLstX+fr6rXquqPVbUF8NkaPINVSAf4gZ3uSmAc1iwAF9a0vUtEJB3IUdVbsF7UAryiqn8CWrH0dAOn2bZ3VfU6YKYd5zKsKYkGg+EQiZuaPm2nHQkHpiS1Z4cD05KaCBl9jPWwudt+qCEiv8EaiSzAnWFhB74b7VHL/YHxqnqViPwJa4Q0WLXdNPs7XjEadC+BvPkO1qj6pcCb9r4G+9uHlW9+rBp5s70/PO/hwHU3A/tUVcOa0tvTKzyPw8MM1S8jZL8DeF5VVwGIyIVYLzF/EZEFEcIMhPd9Vf2aiFzOwfoB7LG/vRz8nzEYDJ0QT4X+POC3IlIArAJeAh4VkQuABapaKiJJdvNyGfDPCGE9DjwjIlXAcqz5xT/HqmkF4npGREYDQ7AeeAFqgGQRuRko6L7LiwmMBtHh38B1WM3qHTVlPwU8yIFWj3nA70VkHLBNVfd1NjRAVetEZK2IPIrVJfArEbnW1itJVdeIyFsi8lusl47w+fmhPAHca08DrAOeB+4XkcGAU1VrRWS3iNyPVVtfHnb+WhH5JVZ3zPsRE24wGA4ZM2Wvi4jITVjN0ftU9Vc9nZ6+iNEgtrFfBk8ABgA/MH3zBkP0MYW+wWAwGAx9hHgayGcwGAwGgyECptA3GAwGg6GPENOFvoi8dZTiiTgCSkQeE5HkSMd0cN4osdYZfz1k3yQRubYr6TyaxEHe9xORP4vI0yJyaZhtlogsFpHf29tGk4PjMfeDwRCjxGyhbzsv2XoE5w8TkQUicqOIvGKPKn9ERH4rIneLyCAR+beI3AIMEpEPRORndkFxrYi8LiLFIpIFqKo2ikimXVA8IiKnuYp8kAAAIABJREFUdJYGVf1CVa8O27cCOKmr13U0iIe8By4AXlfV72E5fglFsZaITcSaD280MfeDwRAXdHnKnogMUdUKe3uoqu7ovmQdEscD/wlJzzlYHr02q+oL9r5xWEt/hnKVqu4GxgNzVPUxsVYE+xHwV1Vdbtc0JgIvq+rjInIW8Df72DlYK5fVAMOxPIattcNOw5oi9piqzrfTkA/8d1ga7lLVqgjX1iQiOaq6K8IxPUk85P0wDqwR3xp2zGJV/UBEcoBHsBziGE3M/WAwxDxHUtP/bsj2DUeakC4wCVgR+KGqbwMPY83ZDuxbrarnhn122+bxwLuBQ7HmCq8Wy31pI5aXs/ds+wTgXRFxA5Wq6sdyAboay21ojX3cN4FzVTXgQAUsByeusE9nVGM9MHsr8ZD327EKfgi7D+w4wNIhIWTbaGJh7geDIUbpUk1fRF4DxorlllOB0m5N1aFxPNBfRFqA14DFwE+BRyOedYB8YKOIZGM5NFmA5R+8Eas29FNgQ+ixwHHAOnvfCFUtF5EM4Ex7nwNr7fJ9wC9UtVVVN9G2ZgOAWJ7j7gEmisgvVDVQCxuK5Se+txLzeQ/8DXjCrhG/BSAi07BqtDuBM7A8zD1hH280MfeDwRDzdHmevogco6rl9na2qu7t1pQdfnpuw3Ly8Y6qzj2K8QrwO9sneHeElwI81F3hHQ3iJe8jxGM0OfR4+/z9YDD0ZiIW+nZN/lzAA6Cqd4XYHlLVm+3te1U1kkvOuEZETgJWqGpjN4RVALSq6uYjT1n80515HyEOo8lhYO4Hg6H30lnz/k+wBjL52rFlhmxnd1uKYhBVXdyNYW3srrD6At2Z9xHiMJocBuZ+MBh6L50V+mtUdU0HtldE5A2s1b2eiRSI/bZ+K9bo4Dn2vsuxXhb6qbXW9x1YC3PsVtXnD+MaDIeI0aHnMRoYDIaepLNC/ysiMgt7BTNVvSTEtgZ4TlXfEpEJkQJR1Y0i8hwHL705QVVvFJH/FZFMwK+qD4u1wlebB52IXANcAzB27NhJK1asYOfOnXg8HlwuF263G7fbjcvlwuPx4HC0nZigqkgnK431AqKWwCPVIVyD0tJStm7dyq5du3C5XDidTtxud/A7ISGBfv36BfcBtLS04Pf7SUtL6+1aRCVx0bgXSktL2bDBGmOXkJCAx+MhISEBh8OBquJwOHA6nTgcjjbb7d0nvYxe/ScxGGKNiIW+qp4nIsfa2+Ej9O8BqrBGPl8FXN/FNHS2XncgLX/AGk1MSUmJbtq0ib/97W/tFhyh4xTC7aEPQbfbTVJSEqmpqTQ1NTFixAgSExNxuVyoKq2trVRXV5ORkREeBapKY2MjkyZNIicnp7cXYIdCpzqEawDw/vvvs2PHjnYLD7/ff1C+BLb9fj8OhwMRQURwOBzBFza/38/gwYPp168fycnJuFzWX7SmpoaMjAxEhISEBBITE0lISCA1NZWhQ4fGQ/5DF+8FgFdeeQUROei/H+k+COy77bbbuiPdBoMhRohY6IvIY8Bee/taVQ0t2KuB2kOJREQGARcBSSKSjjUfeKVYS6OiqtUi4rR//ydCUEHcbjcVFRUMGDAAv9+aVq2qqCq7du1i8ODBB9VoAgVMTU0NgwcPBqxaZ3V1NVVVVagqO3fubDcuESExMZH09HRSUlJISkrC4/Gwd+9eFi1aRHNzMy6Xi+zsbEaOHElubi4JCQnthtWTREOHYcOGsW7dOoYOHRosZFQVv9/Pzp07GTIkOE08aK+urmbgwIGA9QLQ0tKCz+ejqamJ1tZW6urqDorD4XDg9/vJzMwkKyuLAQMGICL4/X4aGxtZu3Yt2dnZjBs3DrfbfWSZFGWioUEgDwcPHhx8+QoU8iLCli1byM/Pxw4fv98f1OeNN97gwgsv7O7LNBgMvZTORu8/rKo32dsPqurPQmzXAZdhOTlZpqoPRTuxAUpKSvSGG27g5z//OUlJScECPfBpbGwkKSkJv98frLH7/f5ggeLz+fB6vSQnJ5OSkkJqaip1dXXk5eUFfycnJ+PxeNiwYQOFhYU0NTXh8XiCdrfbjaoydOhQpk2bRnZ2Nrt372br1q18/vnnnHTSSUyaNKkrNdCYqLKWlJTo8uXLOf3001m+fHmwdSQUn8+H0+kMvowFPl6vF7/fH7S73W48Hg9utxsRYeTIkQwcOJCsrCwSEhIQEbxeL/379yclJYWEhAScTiciQlNTE+PHj8fj8eD1eklNTWXChAk4nU4qKioYNWoUra2teDyeYAtDoKuhE2JGh7lz55KbmxvsYgGrsA/kvddrLVPv8/kOyvMhQ4ZwySWXMHnyZPLz89m7dy+TJk1i/vz5DBs2jCFDhrBx40ZGjRpFVVUV1dXV5OTksHv3btLT08nOzmbz5s0UFBSwZs0aPB4PkyZNYsWKFWRlZZGWlsbKlSuZNm0aW7dupbGxkZKSElasWMGAAQOoqanB5/NRXFzMxo0baWlp4bjjjmPXrl2MGTMmcIkxoYPBECt01qfvEpHb7e2kwE57Lu4mYDbWi0P9/2/v3OOjqs69/32STO73C7kQSAjkRoBwCQoYMYYS6amC8pG2+GK9VGv1HKn2VE8v1tup2mq9V6v08Gp9+2pLRbGWqiAEQQEJKBAIIpAEkph7QiYJk0wu6/wxF0OSmZCY2yTr+/nMZ2bWmr33s/dvZj+z1nrWs4bIPocsXbqUI0eOEBcXd14LUylFaWkp0dHR9hZN127ms2fPEhISQmdnJy0tLZw7dw6TyUR1dTU+Pj7ntTqbm5vx9/fnxIkTNDQ00NjYiNFopK6uDi8vL4KDg4mOjmbTpk1ER0cTGxtLSEgIbW1t5OXl8cknn5CcnIy3tzcGg4Hm5mYWLlxIQEDAcF+uIeP+++/n3XffJTExERHBZDIRFBREY2MjxcXFpKamYjQa8fT0pL29HbPZTGVlJVFRUXYH3NTUhFIKk8lESUkJnp6e1NbWUlVVRUtLC0ajkdLSUjo7Ozl79iwBAQEEBgYSHBxMSEgINTU1XHTRRfj7+1NYWMjp06dpb2+nvb0df39/ysrKmDp1KsXFxZhMJkJDQ6murmbChAl4enpSWlrKzJkzKSoqIiMjY6Qvab+ZMGECd9xxh70137WVD1BcXMzEiRPp6Oigo6MDs9mM2Wzmn//8J6+88gqtra1UV1czZ84cqqurCQgIYO7cubi5udl7aiZPnmzfV1ZWlv3YkyZNAsBsNhMfHw9wXj1AZGQkkZGR9ve2+uLiYvs24eFfTwA6e/YsGo1maOhrTP8nvY3pK6WUiMwHcrukLB1WPD09CQ4Oxtf368W8bC2dpqYmYmNjzxszBstNsKCggEmTJtHe3m7fTilFYWEhUVFR9tZQe3s7HR0dtLS04OXlhYeHBz4+Pvj5+eHm5kZLSwvvv/8+YWFhHDlyhIMHD1JVVUVAQABJSUnceeedzJ07l48//pikpCQyMjLYvn07BQUFNDY2EhISQmpqqt3+rjdAV8LPz4+goCB7YF5goCVbqre3Nw0NDfj7++Pn53feNq2trYSGhtr/rAUGBhIQEEBwcDB79+4lPT2djo4OTCYTra2ttLS0cOrUKaKjo2lra6OhocE+LFNcXMz69et59913ueiii3jwwQeZOnUqn332Gbm5uUycOJHYWEu2XZtjKS4uJi0tzW6PzVl2H1ZwJWy9ILbvvA2lFEFBQURGRuLh4WH/jdied+/ezfr167nuuutobGxk4sSJlJeXU1JSYtfHNrxlMBioq6tj8eLFBAYG2ntr2traqKurIzo6usewlrPvtKM6V/wdaDSugkOnLyKPA/FfvxXVLXr/EuAfImLC8j/gu933MZSEhoYSGRlJQkICs2bN4uDBg/bWY1NTE0uWLOHo0aOICCkpKeTn5xMbG0tjYyMRERHMmDGDzz//HIPBQGRkJF999RXz58/HbDbT0NBARkYGhw4dIj8/nxUrVpCXl0drayuVlZWYzWbc3d1JT0/HYDCQkJBgH1etqanhgw8+4L777uOGG25g0qRJlJSU8NlnnxEeHk5oaCju7u5ERETw5ptv4unpSUxMDDU1NS55swsJCSEiIsI+Rt+Vs2fPEhER0aO8oaHhvHO19brU1dVhMBhoamqyOy/bMIqfn599VkZERATh4eEopUhJSSEnJ4djx46Rl5fHvHnzSE9PJzMzEz8/P958801Wrlx5Xpf+WHQ2YWFhxMXF9ShXStHa2npeS9tWHhERweOPP86TTz7Ja6+9xqpVq+xOPCIiwh4s6e3tDUBLS4s9jsVsNtv/DERFRXHq1Cnq6+vp6OjAzc0Nb29voqOjqays5LLLLhuWa6DRaPrGodNXSt0rXVbPExHb4iS27v1XlFJvDYONDklISGDx4sXA+V2KBQUFhISEkJmZaS+z1efn53PxxRdz7tw5UlNTMZlMmEwmIiIiMJlM9vHfw4cPIyKEhoZSVFREREQEHh4eJCYm2oPF3n33XXJycjCbzRiNRoxGI+fOnSM0NJRt27bx3HPPcdtttxEaGkpoaCinTp3iqquuIigoCIAf/OAHKKWoqqqitrZ2eC7aIBMfH8/SpUt7dZheXl7Mnj3bfo1NJhPnzp0jKCiIgIAA3N3dezw6OjpYvHixfdqZbaz/r3/9K1dddRVNTU00NzfT1NREU1MT+/btIyMjg8zMTDIyMqiqqiIvL48XX3yRzMxMfH19WbduHfHx8SxatMh+7ccaqampPbrVbfj4+LBgwYIe5V5eXhiNRq655hq8vLx44403WL58OcnJyZjNZpqbm2lvb7fPeLHNnLjlllvsgZS1tbWcPHmSc+fOYTQa7YGC7e3tnDx5koqKCoqLi1m+fDkhISG9WKfRaIaTvsb0fwjYUu+uBe4Fe/f+DBHZgzWCfyjToPaGyWSitLSUHTt29KirqKjotdzmYAsKCvDx8bFP15swYQLV1dXMmzfP7nxsXaQ7duwgKyvrvChzW/d/eHg4CQkJ5815dnd358knn2TFihVMmDCB559/nptvvtm+v3Xr1vH973/fPhYqIkRGRjJnzpwhu1ZDSXFxMQcPHqS4uLjXOhGxX2sfHx9CQkKoqKggMTHxvADLjo4O2tvb8fT0pKGhAbPZTGtrK2azmba2NlpbWzl27Jh9uGDSpEn4+/tjNpu5+OKLaW5uZtq0aeTm5hIbG8u8efN4+eWXmTdvHv7+/gQGBrJp0yb8/Pzw8vLiyiuvHCvT/ACorKzs9TsPFh1aWlp6lBcWFrJ06VKSk5MpLS3Fx8eH119/nS+++IKVK1eSnZ1NRkYGnZ2d1NTUUFlZyebNm/nTn/7ExIkT8fX1xdvbm5CQEFJSUlixYkWPY7z22mtERETwyiuvEB0dTU5ODmFhYYN9+hqN5gJxGL0v1pX0+HoFvaNKqYe61L9Cl3nESqmbh9DO87BFjjsaB3c2Pt7fbQa6r/fffx8fHx8OHjzI+vXrufHGGwkNDcXX1xcfHx/Cw8OZO3cuSUlJ3Z2PS3gimwYwfNf0QsqVUuzcuZPjx4+zYcMGSkpKWLNmDX5+fqxatYrGxkY2b97MrFmzCAgIID09vUfMgRWX0mEg17SwsJCWlhZKSkqYPHkymzZtwmQy8ec//xk/Pz/Ky8tpamqyD6XFxsZiMBhYu3Ytvr6+KKVIT0+noaGBbdu2ER0dTWpqKhMnTuxx7JaWFv7xj39QV1eHr68vl19+OWVlZb32QHTDJXTQaFyFvqbs2bv3e6kzADnWtx8opdp7+9xQ0NXhjFaOHDnCZ59Zpll7eXnx05/+lLCwMG666Sa8vb2ZP38+hYWFtLW1kZKSQmtrK4sWLQIXucmNdg22b9/Orl27WL9+PdnZ2SQmJhIUFMTUqVOpqqoiMjKSBQsWcPToUUwmE4mJiXR2djJlyhTbLsaNDkopzpw5w+eff87hw4epq6tj+fLl9l6vkpISDhw4wL59+zh8+DCbN28mPDyc8PBw9u3bx6xZsygpKWH+/Pl88cUXlJWVERMTQ2pqao/pkQUFBRQWFtLQ0EBJSQk33ngjUVFRzsxzCR00GlfBYQ5OayDf0yKyQUT+LiIbun3kJSDU+nh5CG10SWbMmIHZbCY0NBSj0ciHH35IaWkpn332Gc3NzZw8eZKrr76a7OxsSktLOXbsWN871VwwtvH8733ve2zatAmDwYDRaKSyspLm5mbi4+PZs2cPF198MdnZ2bS2trJ3796RNntEEBHi4uK4+uqrSUhIICQkhLy8PAoKCuzj9dOnT2fNmjWkpKRw++23U1RUhMFgYNmyZVRWVnLixAnc3NxIS0sjJyeH0NBQduzYwdatW8871vTp07nssssIDQ1l8uTJbN26lb/85S/U1dWN0NlrNOMLh05fKXWvUuq71seqXqLzjUqp/2ddEKTe2UFEZK6I/EZEnhIRP2vZ7SJyt4hstr7/m4jcJSILv+lJjRaysrJoaGggLi6O3bt3s2DBAv71r3/R2tpKSUkJO3bsIDo6mhUrVgz5OOd408DT05OIiAji4+PJyspi8+bNREREUF5eTk1NDV988QVpaWls3bqVjo4OkpKSWLhw6E97tOuwatUqRISwsDBMJpP9+3v99dezevVqli1bhojw9NNPs2vXLpRSzJ07l6CgILZt20ZHRwcAMTExLFmyhPr6egoLC887RkBAAMuWLaOjo4P09HSys7N555132LhxI/X19b3Gh2g0msHB6WobIvKkiPx/EXldRF7vVl0vIutFZD1g7OM4q4EHgU1YEvqglPojUAC8av1MBZblensdbxCRH4nIfhHZX11d3cfhRgfTpk2jra0Ng8HAtGnTyMnJYf78+XzyySd4eHhw4sQJewt/9mynaxYNBuNOgwULFtDR0cGll15KRUWFPeuiUoqTJ0+yc+dOMjIy2Lp1K21tbcM1ZW9U6+Dl5UVCQgJr1qzhe9/7HjNnzqSgoIDf//73vPTSS3h6enLbbbdx4sQJXn31VXvvyKxZs5gzZw5btmyxZwAEmD9/PidPnuwRSCgiXHrppfj6+vLFF1/YexHeeecdPvroo0E9J41G8zV9LbFVoZT6P0qp65RS19kKRSQCeBS4C7gb+OMFHKu3RUSuBt4GSyIgpdQDwJpeN1ZqnVIqQymV0dvc79FKdnY2ZWVlnDt3jpaWFu655x727t1LQ0MDTU1N7N27F6PROFwOZ1xpkJycjIgQFBTED3/4Q37zm9/g6emJp6cnkyZNorKykjfeeIOFCxeydetWTpw4MVymjWodMjMz8fb2ZuLEiWRnZ7N27Vp+/vOfM3v2bHsK60cffZTdu3fzwgsvkJ+fD1hyBSxevJitW7fS1GRJ0jllyhQyMzPZtWtXj+PEx8czbdo05s6dy5YtW4iNjWXNmjXnJc7SaDSDS19OP1NE7hGRO0Tkji7ljyil2pVSjUopI/Dffeznr1haNyuASSLiJiLxQIktAFBEfiUiDwOHBnIio5XJkyfj5uZGWFgYAQEBmEwmrrvuOnvq3srKSt577z2KioqG2pRxp4GI2IPE/P39ufrqq3n22WcpKyvDx8eHlJQU+zz+hIQE9u3bNxxmjXodevsDKiIsXbrUntXPaDRy9913s23bNl5++WUee+wxwJKhcenSpezatYuamhoAfH19iYuLcxi3EhwczBVXXMH+/fs5c+aMfUEsjUYz+PQVvX8ZltaIYJmev9Na/lzXFfdE5Hml1J1DbayN0R453p2vvvqKDz/8kNbWVry9vZk1axarV69m5cqVREZGcu7cOQIDA7n99tvBRaKVXUWDyspKNmzYQGBgIBEREfzsZz+jpaWFjRs34unpSVFREc3NzZSVlSEi3H333bZNtQ690NzczNNPP01UVBQpKSn861//Yt26daSnp/PUU0+Rnp4OWLIs5ubmEhAQwEUXXQTAtm3bWLBggaMpkgAcPHiQM2fOsHz5cluRS+ig0bgKfbX0r7Q+vgNkicg0a7m/iIQCWJ/9h85E1ycmJsa+ap/BYODcuXPceuutvPzyy+Tk5ODj42NvFWkGl8jISAwGA+7u7pSWlvKrX/2K4uJivvzyS0pKSvjWt75FYGAg0dHRNDc3j7S5ox4/Pz8WLlxIRUUFR44c4dprr+W2225j586d5Ofn27v63dzcyM7OJj8/3x7cZ+vmd9bQmD17tn39Bo1GM/j05fTdgA3A34Aw4BfW8geBJ0Xkr8Dvre81TsjMzKSmpsaeFXDp0qVkZWVx//33c9FFF+Hj49P3TjQDIj4+ntbWVvz8/EhKSiIhIYHf/va3zJo1i0OHDnHFFVcwYcIE+8JMGucsWbIET09PRISioiJmzJhBWFgY9957L1VVVRw9asnnJSKkp6dTUFAAWIIEU1JSOHz4sNP9u/IaCBrNaKevu5y/Umq/UuoA4AV8BaCUOqOUukkp9X2l1M1KqdNDbqmLExUVRUdHB15eXnR2dmIwGFi1ahVbtmzBYDDgCl3lrsrixYsxmUz2DIkrV64E4A9/+IN9eeXs7OxeFw3S9M4NN9xAVVUV9fX1zJ8/nxtuuIFrr72Wu+66i4aGBvv4fUZGBuXl5fbWfXx8PGfPnnW6fK52+hrN0NGX099mTczzd2A7MKIL7Lg6iYmJ9qViT58+TWJiInfddRe33norf/vb30bavDGLLUd8c3MzAQEBLFiwgP/6r//ihRdeIDIykk8//RQR4Vvf+tZIm+oyREZGEhkZSU1NDXv37mX69OlkZmYSGRnJ7373OxobGzl+/DhgCWYtKSmxb3vJJZewe/fuHvP3NRrN0NOX038Ly9zhV4G3lFKfD7VBY5mrrrqKlpYWlFK4u7sTGxvLggULaGxs7JGuVDO4zJw5E7PZTFhYGMePHycsLIxbb72Vm2++meDgYL766ivdwuwnOTk5uLm5UVdXR0tLC2azmWeeeYaioiI2bNiA0Wjk448/Jikpyf4HAMDDw4OLL76Y7du397oQkEajGTr6cvov4yDVroj80pq0541eEvdoesHDw4OIiAjc3Nwwm832JYDfeuste7CTZmjIyMigtbWV2tpa+9K+N910Ew0NDezZs4dDhw45DTDT9CQuLo7k5GSam5uprKzkO9/5Drt37+all17i7bffZt++fWzcuBGwTMvrmmo3LCyMzMxMtm3bRn2904SeGo1mEOnL6TtLtSvWpD2ruybu0Thnzpw5tLW1UV1dTUdHB2lpabhCdjtXx93dHX9/f5RSJCYmcvToUcrKyrjvvvt44IEH8Pf3Jzc3d6TNdClEhIiICBISElBKkZeXx/Tp06mvr+f+++/nkUce4S9/+QsnTpwgPT2dQ4fOTzuQkpLCsmXLOHDgAKdP67AgjWY46MvpO0u1O1VEVonIv4nIvznbiYN843+y5he/wvr+QRH5TxG5fsBn4wLExMTYl9KdMGECX375Jf7+/tx4441DelytASxatIjm5maOHTuGr68vIkJmZiarV6/mvvvuG5a4irGmw4wZM4iLiwPgwIEDzJ07l9raWjIyMnjooYeoqakhLy8PT09P3NzcenTnu7u7s2TJEiorK3v8KdBoNINPX07/ESypdu8CftOt7iPABwi3PpzRI984lvziAYC7iIQAnUqpJ4G5ve3A1fK+O0JEmDx5Mu7u7hQXF1NdXW2Pfh5ixr0G06ZNo7OzE6UUaWlplJSUcOrUKZYvX05bWxvr1q0bDjPGlA6BgYE0NjYyZ84cOjs7efXVV1myZAkHDhwgNTUVHx8fXnrpJc6ePcvs2bM5ePBgj32IiH3a6s6dO3WAn0YzhPTl9J+wptptxJJrvysfAzOBGcDuCzjWefnGlVK/Vkr9N5bEPz3qe2zsYnnfnTF79mza29upqqoiNjaW0tLSEcm9Px41CAkJwcPDg+LiYry9vTEajSQnJ/OLX1hSUJw7d244zBhTOsTGxhIdHc1NN91EdXU1xcXFTJgwgaamJhYtWkRtbS0vvPACQUFBNDU1OYxfSUpKIjk5mSNHjgzzGWg044e+nL5vl9fds+49ADwHPA/8uo/99JZv/E4ReQA4o5Sqx9LK+U/gsws13lUJDAzE399yOb29vc+LbB5CtAbA5ZdfTmNjI21tbaSkpFBVVUV1dTVeXl6AZXrfEDPmdEhKSiI/P5+YmBhmzpxJbm4uCQkJ1NfXM336dJ544gmeeeYZzp49y/Tp0x3m4AfLVECdkU+jGTo8+qj/RETeAjqBd7vVtSqlSgBExORsJ9bkPge6FT/f7TMP9mntGCIxMZHDhw/z+eefM3nyZE6dOjWkrX2tgYXw8HBEhICAAEpKSvDx8eH06dMkJSURExMz5Mcfizq4u7sTGRmJ0Whk5cqVnDhxgs2bN7Nw4UJKS0vx8/Ozr9b32muvkZ+fT1pamj22pTt66qRGM3RcyDz9x5RS1wIfdqt7V0TetCbueW9IrBvDpKam0tnZSV1dHfPmzdPz9IeRjIwMSktLMZvNJCcnU1NTg7e3N9dfP+rj5kYt2dnZ9vH6a665BpPJRH5+PgaDgfLycp555hn++c9/sn///h7Jerqjnb5GM3T05fSfBW60vv6lrVBEvLD8CbjBWr91CGwb0xgMBsLDLfGPJ0+e1De6YWTJkiX4+vraW/v+/v4cPnyYnJyckTbNZfHx8aGjowOz2WzvNampqcFkMtHS0kJ5eTl33nknt9xyC1OnTuX48eMUFxePtNkazbijL6ffCFRaX3eda/NL4A9YuiX/YH1o+klaWhpeXl7s379fO/1h5qqrrqK0tJS2tjYSEhKoqKjQufe/IV2j8xctWkRUVBRmsxkPDw/y8/O59957cXd358UXXyQqKorHH398hC3WaMYffTn9amCRiPwey7g+AEqpB4CN1sV2bgLeGUIbxyyTJ0/G09OThoYGdu++kAkQmsHCz8+PkJAQ/Pz8KC8vJygoiD179oy0WS5NSEgIDQ0NdHZ2kpiYSFBQEADCoK4BAAAOfklEQVR1dXUYjUZqa2u58847eeSRRwgODmbnzp18+umnOhOiRjOMOHX6SqlHsczR/7/Ar2zlIvIdYHWXxDyrh9TKMYqIEBMTg6+vr9OIZs3QcOWVV9pb+3FxcdTW1o60SS5PSkoKx48fR0RITk7G09OTtrY23Nzc2LZtG1lZWaxevZq1a9dy9OhRIiMj2bFjB52dnX3vXKPRfGMcOn0RmSkidwK1wEVA13Rl4YAZiADC6DmHX3OBzJw5Ew8PDwwGw0ibMu7w8vIiNjYWb29vKioqtAaDQGxsrD1ILzU1FX9/f5KSkjCZTNTU1BAYGMgVV1xh/5Pr7+/PjBkz2LJlC21tbSNpukYzLnDW0n8YyAM+AKqVUtfYKpRSfwa+VEr92Zqb//AQ2zlmCQkJsU8j0ww/OTk5VFZW0t7eTkBAwEib4/KICBMnTqS0tBQ3Nzd7nERAQADu7u68/fbbhIWFsXnzZsDSM7B+/Xrmzp3Lli1baG5u5sEHHxzBM9BoxjbOnH6JUmovsFsptbmX+kwR+YWI3CEidzg7iIN843eIyK9F5CHr+79Z848vHPDZuCgJCQn4+fkN6TG0Br1jMBiYOnUqXl5ew5KNbzzokJqaam/Jz58/Hzc3N/uiPBUVFUyZMoXy8nIAPvnkE/Ly8pg3bx7l5eXk5uby/PPPO9u9RqP5BjhLzjNNRDYA/ta5+Eop9d0u9U/wdZrQvpqpq4FfAIuw5BvfpJR6UUQ8+ToxSQUQgoPUoyLyI+BHYAmAG0skJiYOR9ey1sABWVlZrF+/Hje3vuJaB4Uxr4ObmxuBgYGcPXuWhIQEmpubKSwsZMqUKZw8eZKNGzeSlJSEu7s7ycnJbNy4kd27d3PPPffQ2NhIc3PzSJ+CRjNmcXiXU0r9m1Lqu9bnVd0cPlim8N0C/BBovYBjnZdPXES8gcewxgMopX5inRWwxoE9oyrf+GAzErn3tQYW3N3dmT59+nBGkY95HbpO30tLS8PNzY2QkBCCgoKoq6sjJiaGtWvX2gP4Fi1axMcff8zDDz9Ma+uF3E40Gs1A+CZNm9uxOP0fAbf28dke+caxBAa2AjkAIvIrEXkY0OtrDg1aAycsWrRouJz+uNDBy8sLEaGlpQU3Nzfi4+MxGo1MnToVg8HAe++9x6JFi9i+fbv9uosIV1999QhbrtGMbfrKvd8XF9S97yDf+Ipun3nkG9qicYLWwDk2xzTUjCcdZs+eTW5uLt/+9rdJS0vj9OnTeHh4MGXKFE6cOMGxY8f49re/zY4dO8jKytLBrBrNMODQ6VvH8f0AT6ADy5j+si4f+SOW+fuCzsinGQPMnj17pE0YUwQFBVFZWUl9fT0hISHExcVRVFREZGQk1dXV1NTU8NFHHxEcHExNTQ3XXHMNHh4ePPDAAyNtukYzZnE2pr8K2A9cASyj54I7Z4CjQD5weqgM1GiGC50KefC59NJL2b17N+3t7faxfYCpU6fi6elJWFgYNTU1FBYW8j//8z/k5uayatWqEbZaoxm79DWmnwRMBKKAKd3qHsSyut524OlBt0yj0bg8U6dOJTMzk48++sg+hNLQ0EB0dDRhYWEEBwczZcoUkpOTqaur4/Dhwxw40H30Q6PRDBZ9jek/DKy1vu4+efaoUuoQgIicGGzDNBrN2CAoKIgpU6Zw8OBBZs2aRXFxMUajkZycHGJiYigqKmLv3r1MnjyZsrIyzpw5M9ImazRjlr6c/nKl1M8c1H1XRK607sNdRDb0Mq1Po9FoSEhIYM+ePVRWVhIfH09JSQlHjx7F19eXpKQkkpOTKSkpIT8/n4KCgpE2V6MZs/Tl9DNF5BdAA4BS6kVbhVIqU0QCrK8bh85EjUYzFliwYAEffPABl156KWVlZUyZMoWKigry8/Nxc3MjNjaWyy+/XCfn0WiGkL6cvsOseyJyHxBnfV2ilHp48M3TaDRjBREhKyuL3Nxc0tLSyMvLIzY2lgULFuDj40NpaSn79u0jMDBwpE3VaMYsfTn9FuDfsTj+P3arC1VK3QogIk8MgW0ajWaM4e3tzZw5cygsLGTevHmEhIRw6NAhmpubmThxIpdccgmlpaUjbaZGM2bpy+nbsu4J8CKwt0tdsIjcgOUPQZiznYjIXGAl4Av8WinVLCI/BTqt2z8PPA7UA/uVUh8M4Fw0TtAajA60DhAVFYXRaOTgwYMEBwcD4OHhwcmTJ/n000+Jj4/X0yc1miHiQjLyKeuje7qsx7BM6RPgx33so8ciI8AkpdTdIvI0kA4cVkq9Zn3f40bXdZERoElEjgPhQE0vx3NU7qxutOzr/W5JkAaLodKgq+3dGS3XdCDlrqaDK1zTgWwzVDpoNOOSvpz+H4FXrK9f6FZ3jVLq8X4cS/Xz+fyNlVoHrOtaJiL7lVIZ3T/rqHwg24z0vgaZQdcAXO+aDmRfg8yw/Rac1Y30NR0FOmg04w5naXifxJJ453YH0fk/EJElgJGey+52x7bIiC9QZF1kpERE7gKKsSwscp11psCWgZyIpk+0BqMDrYNGoxkxnLX0nwOygWdFJAg4rpT6ZZf6xUqpugs5iINFRrpn8bvnQvalGRhag9GB1kGj0Ywkzpx+CXAMiLQ+fGwVIvIq0CIiRUqp3w2phc7p0dXcR/lAthnpfbkCrnZNB7Kv0Y4rXtOxqINGM6oRR2uIi0g58Hfgt0qpr7rVvaSU+rGIPK+UunMY7NRoNBqNRvMNceb0A4HLgIVYWvm1SqnfWOs+w7LqXjaWcX+UUvcOh8EajUaj0WgGhrPufTfrwweIxxJ4ZOMa63P3iH6NRqPRaDSjFGct/VeBHUCuUur0MNqk0Wg0Go1mCHDo9EczIhKolDKKSJBSqqFrOYBSyuhkW1G9nHRv5Y4+a60LU0rVDvwsXJ+B6tDHdb1gHbQGFgZbhwGUax00GhfB5Zy+9UZ2HfA68GNbgiAR+bX1IwpL3oBHum1j43bbjAMReRbwxzI98TKl1HPW8iewzqPGki3tJ1329ZT15Rzgc6XUT63lP+HrzIXKti9r3Wql1Bsi8l2l1Ia+Pu8K9FcHRxpY6/qlgyMNrHVaB/qvgyMNrHX90qGvazoWddBoXIULScM72rgbmAZMAI53KW9WSj0FYM1l3pV3gFwsN5SLAZvDqVBKPSYiD3L++gGVgK9S6vci0n3OdBGWIMYcpdSzXco3WZ97S1kcY32e7ODzrkh/dXCkAfRfB0cagNbBRn91cKQB9F8HZxrA2NRBo3EJXM7pK6UeEpEHsNxMkrtU1VizCAIc7LbZo0qprQAiMq9LuS1W4SEsLRwbh7DczADOdjv+8yKyGkjsVn5aRP4DmAEcxrJAkY0ia6toV7fP3wHMA9yBGx2e9ChkADo40gD6qYMjDax1WgcL/dXBkQbQTx360ADGoA4ajavgck7fyjNYbgzX2wqUZYGSy2xvu35YKbVVRBZ334lS6vUu5X/v/nkRgfNbTza+Av4uIouVUjvP36X6sfWGB9gXRynv2gXdBV8sN75WJ+c6mrlgHRxpYK0biA6ONLBuqnXojw6ONOi6TT916KEBjHkdNJpRj9tIGzBALgMWc37LBiAYS9fkml62CcGyelf3uv6WOztOvIg8zPndll7Av4vIU13GQG2UAAZgbi/HcAX6q4Oza9pfHZxprXWw0F8dBqJPf34LMLZ10GhGPa7a0m/Asv74A93Kc7H8kZneyzaO6vpb7rBOKXWPiExQSlV1KXteRIqARnqOV76JZXzTVddM768Ow6KP1sHOkH/nHZX3poG1fCzroNGMelzV6V8MJABngEe7lN+N5ea3p5dtHNX1t9xhnYj8FmgTEYNS6uddqtyxtIaWAV27QF8G8rHc/FwxWrm/OgyLPloHO0P+nXdU7kQDGLs6aDSjHlft3j+jlPoxYF/lzzrl5yyWpX7Pa3U4qutveV91QJlS6tdAWbfyXCzpis90K9+plHrWhacnXbAOw6WPFa3DMHznB/hbgLGrg0Yz6nHVlv5/iMgVQJKIJFmDgjbxdXdh92lCjur6W95XnYjIS8CRbuWOWkkLrRHUHQ4Cm0Y7/dFhuPQBrcNAdBhMfcCxBjB2ddBoRj0u6fSVUpm9lJ0WkXuBQCzThE73Vdff8r7qgHcBE19Pb+raGgJLa+hDa3kGlhZPKy46N7k/OgyXPla0DsPwne/vbwHGtg4ajSvgkk5fRP4AmAG6tQjagS+BqF42c1TX33JndTcDbwM/Ae6yljlqDf0UKMDSNeqSDECH4dJH62BhOL7z/fktwBjWQaNxBVzS6QOfKKXe6KV8EeBJ712Kjur6W+6srhqoBb4Saz50J62h54DvY5mm5Kr0V4fh0kfrYGE4vvMX/FsAp70DY0EHjWbU43JOX0RigRvFkgmsXSm1skv1Liw5yHtL7uGorr/lzupCgBuwtGDuAh62lvdoDSml9gJ7e9m3SzBAHYZLH62DheH4zvf3twBjUAeNxlVwOacPTMJyk+mNYOB26+uHL7Cuv+XO6mxjlW1YopNtOGtBuSoD0WG49NE6WBiO73x/fwswNnXQaFwCl1tlbzQjIvdjCWBaAXgope6zlv8Ea2tIOVn2VzM4aB1GHkcaWOu0DhrNCOGq8/RHK/5APZYc4sVdym2tobt62UYz+GgdRh5HGoDWQaMZMXRLfxARkZlYsox9AJxQSplG2KRxidZh5NEaaDSjE93SH1xuBaqA4/omN6JoHUYerYFGMwrRTn8QUUqtxRKV/OlI2zKe0TqMPFoDjWZ04orR+6MWEXkEyzSk60balvGM1mHk0RpoNKMT3dIfXCKwrBf+y5E2ZJyjdRh5tAYazShEt/QHl8NY5iU3jLQh4xytw8ijNdBoRiHa6Q8uO7FMVUofaUPGOVqHkUdroNGMQvSUPY1Go9Foxgl6TF+j0Wg0mnGCdvoajUaj0YwTtNPXaDQajWacoJ2+RqPRaDTjhP8FkPNvbR15MiIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 576x720 with 25 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Raw data -- all models. \n",
    "\n",
    "current_row = 0\n",
    "subplot_idx=0\n",
    "num_rows = 5\n",
    "num_columns = 5\n",
    "fig, axs = plt.subplots(num_rows, num_columns, figsize=(2*4,2*num_rows),\n",
    "                        gridspec_kw={'hspace': 1.5, 'wspace': 0.75})\n",
    "\n",
    "all_other_nets_evals = {}\n",
    "\n",
    "# These will group together the MEASUREMENTS from a given model type. \n",
    "all_other_nets_evals_by_type = {}\n",
    "\n",
    "for met_mod_idx, metamer_model in enumerate(models_to_evaluate): \n",
    "    all_other_nets_evals[metamer_model] = {}\n",
    "    all_other_nets_evals_by_type[metamer_model] = {}\n",
    "    metamer_model_layers = all_model_info.ALL_NETWORKS_AND_LAYERS_AUDIO[metamer_model]['layers']\n",
    "    model_name = all_model_info.ALL_NETWORKS_AND_LAYERS_AUDIO[metamer_model]['paper_name']\n",
    "    for layer in metamer_model_layers:\n",
    "        all_other_nets_evals[metamer_model][layer] = []\n",
    "        all_other_nets_evals_by_type[metamer_model][layer] = {m_type:[] for m_type in model_types}\n",
    "    for measured_model in models_to_evaluate:\n",
    "        measured_model_info = all_model_info.ALL_NETWORKS_AND_LAYERS_AUDIO[measured_model]\n",
    "        pckl_path = os.path.join(measured_model_info['location'], network_network_folder, 'accuracy_on_metamers_from_%s.pckl'%metamer_model)\n",
    "        eval_pckl = pckl.load(open(pckl_path, 'rb'))\n",
    "        # Store the data here. \n",
    "        for layer in metamer_model_layers:\n",
    "            if measured_model!=metamer_model:\n",
    "                all_other_nets_evals[metamer_model][layer].append(eval_pckl['average_correct'][layer])\n",
    "                all_other_nets_evals_by_type[metamer_model][layer][models_by_type[measured_model]].append(eval_pckl['average_correct'][layer])\n",
    "        if measured_model==metamer_model:\n",
    "            this_plot_linestyle = '--'\n",
    "        else:\n",
    "            this_plot_linestyle = '-'\n",
    "        axs[current_row, subplot_idx].errorbar(range(len(metamer_model_layers)), \n",
    "                 [eval_pckl['average_correct'][layer] for layer in metamer_model_layers],\n",
    "                     yerr=[eval_pckl['bootstrap_sem'][layer] for layer in metamer_model_layers],\n",
    "                 color='grey', \n",
    "                 linewidth=0.5,\n",
    "                 ls=this_plot_linestyle) \n",
    "    # plot the average\n",
    "    axs[current_row, subplot_idx].errorbar(range(len(metamer_model_layers)), \n",
    "             [np.mean(np.array(all_other_nets_evals[metamer_model][layer])) for layer in metamer_model_layers],\n",
    "                 yerr=[np.std(np.array(all_other_nets_evals[metamer_model][layer]))/np.sqrt(len(all_other_nets_evals[metamer_model][layer])) for layer in metamer_model_layers],\n",
    "             color='k', \n",
    "             linewidth=1,\n",
    "             ls='-') \n",
    "    axs[current_row, subplot_idx].set_title(model_name, fontsize=7)\n",
    "    axs[current_row, subplot_idx].set_ylim([0,1])\n",
    "    if subplot_idx==0:\n",
    "        axs[current_row, subplot_idx].set_ylabel('Word Recognition \\n Proportion Correct', fontsize=7)\n",
    "    axs[current_row, subplot_idx].set_xticks(range(len(metamer_model_layers)))\n",
    "    axs[current_row, subplot_idx].set_xticklabels(labels=[layer.split('_fake_relu')[0].split('_fakerelu')[0].split('_after_preproc')[0] for layer in metamer_model_layers], rotation=90, fontsize=6)\n",
    "    # Hide the right and top spines\n",
    "    axs[current_row, subplot_idx].spines['right'].set_visible(False)\n",
    "    axs[current_row, subplot_idx].spines['top'].set_visible(False)\n",
    "\n",
    "\n",
    "    try:\n",
    "        if plot_num[met_mod_idx+1]!=current_row:\n",
    "            for k in range(subplot_idx+1,num_columns):\n",
    "                axs[current_row,k].set_axis_off()\n",
    "            current_row+=1\n",
    "            subplot_idx=0\n",
    "        else:\n",
    "            subplot_idx+=1\n",
    "    except IndexError:\n",
    "        for k in range(subplot_idx+1,num_columns):\n",
    "            axs[current_row,k].set_axis_off()\n",
    "        continue\n",
    "        \n",
    "# plt.savefig('audio_model_model_comparisons_all_models_no_color.pdf')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Visual Model-Human Recognizability Scatter Plot\n",
    "# Figure 8F"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "experiment_list = [1,3,4,7,8] # do not compare Spectemp experiment\n",
    "\n",
    "all_experiment_dictionaries = []\n",
    "\n",
    "for exp_num in experiment_list:\n",
    "    experiment_name = all_model_info.TURK_AUDIO_EXPERIMENTS_GROUPINGS['experiment_%d'%exp_num]['paper_experiment_name']\n",
    "    jsin_configs = all_model_info.TURK_AUDIO_EXPERIMENTS_GROUPINGS['experiment_%d'%exp_num]['experiment_params_web']\n",
    "\n",
    "    with urllib.request.urlopen(jsin_configs) as url:\n",
    "        experiment_params = json.loads(url.read().decode())\n",
    "\n",
    "    plot_model_order = experiment_params['experiment_info']['networks']\n",
    "    \n",
    "    exp_response_dict = {}\n",
    "    for n_idx, network in enumerate(experiment_params['experiment_info']['networks']):\n",
    "        participant_data_matrix, model_layers, conditions, participants = unpack_experiment_mat(\n",
    "            '../AuditoryBehavioralExperiments/EXPERIMENT_%d/AudioExperiment%d_network_vs_humans_datamatrix_%s_public.mat'%(exp_num, exp_num, network))\n",
    "        \n",
    "        exp_response_dict[network] = responses_network_by_layer_mat(participant_data_matrix, model_layers, conditions, experiment_params)\n",
    "\n",
    "    all_experiment_dictionaries.append(exp_response_dict)\n",
    "\n",
    "combined_experiment_dict = combined_experiment_response_dictionaries(all_experiment_dictionaries)\n",
    "   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "R^2=0.949, p=0.000\n",
      "rho=0.863, p=0.000\n",
      "tau=0.713, p=0.000\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 126x126 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "matplotlib.rcParams.update({'font.size': 7})\n",
    "\n",
    "combined_human_means = []\n",
    "combined_human_sem = []\n",
    "combined_model_means = []\n",
    "combined_model_sem = []\n",
    "\n",
    "for model in combined_experiment_dict.keys():\n",
    "    for layer in all_model_info.ALL_NETWORKS_AND_LAYERS_AUDIO[model]['layers']:\n",
    "        human_average = np.mean(combined_experiment_dict[model][layer])\n",
    "        human_sem = np.std(combined_experiment_dict[model][layer])/np.sqrt(len(combined_experiment_dict[model][layer]))\n",
    "\n",
    "        other_models_average = np.mean(all_other_nets_evals[model][layer])\n",
    "        other_models_sem = np.std(all_other_nets_evals[model][layer])/np.sqrt(len(all_other_nets_evals[model][layer]))\n",
    "\n",
    "        combined_human_means.append(human_average)\n",
    "        combined_human_sem.append(human_sem)\n",
    "\n",
    "        combined_model_means.append(other_models_average)\n",
    "        combined_model_sem.append(other_models_sem)\n",
    "    \n",
    "plt.figure(figsize=(1.75,1.75))\n",
    "plt.errorbar(combined_human_means, \n",
    "             combined_model_means,\n",
    "             fmt='.',\n",
    "             xerr=combined_human_sem,\n",
    "             yerr=combined_model_sem,\n",
    "             color='k',\n",
    "             linewidth=0.75,\n",
    "             markersize=2)\n",
    "plt.plot([1/794,1/794],[0,1],':',color='grey', linewidth=0.75)\n",
    "plt.plot([0,1],[1/793,1/793],':',color='grey', linewidth=0.75)\n",
    "\n",
    "plt.xlim([0,1])\n",
    "plt.ylim([0,1])\n",
    "plt.xlabel('Human Recognition', fontsize=8)\n",
    "plt.ylabel('Model Recognition (all Models)', fontsize=8)\n",
    "plt.title('Model and Human Recognition \\n for Audio Models in Figures 2,6', fontsize=8)\n",
    "\n",
    "import scipy.stats\n",
    "r_value, p_value = scipy.stats.pearsonr(combined_human_means, combined_model_means)\n",
    "print('R^2=%0.3f, p=%0.3f'%(r_value**2, p_value))\n",
    "plt.text(0.01,0.9,'R^2=%0.2f'%r_value**2, fontsize=7)\n",
    "\n",
    "# Note: in our paper we just report the spearman correlation, but other stats are \n",
    "# included here for completeness (all are significant)\n",
    "rho_value, p_value = scipy.stats.spearmanr(combined_human_means, combined_model_means)\n",
    "print('rho=%0.3f, p=%0.3f'%(rho_value, p_value))\n",
    "plt.text(0.01,0.8,'rho=%0.2f'%rho_value, fontsize=7)\n",
    "\n",
    "tau_value, p_value = scipy.stats.kendalltau(combined_human_means, combined_model_means)\n",
    "print('tau=%0.3f, p=%0.3f'%(tau_value, p_value))\n",
    "plt.text(0.01,0.7,'tau=%0.2f'%tau_value, fontsize=7)\n",
    "\n",
    "plt.xticks([0, 0.2, 0.4, 0.6, 0.8, 1])\n",
    "plt.yticks([0, 0.2, 0.4, 0.6, 0.8, 1])\n",
    "\n",
    "plt.savefig('ScatterModelHumanAudioModels_update_size.pdf')\n",
    "\n",
    "# Change back for the other plots in this notebook\n",
    "matplotlib.rcParams.update({'font.size': 6})\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Average CochResNet50 Plots\n",
    "## Figure 8D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'Metamer Generation Layer')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 144x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "model_groupings = {'Generated: Standard':['cochresnet50'],\n",
    "                   'Generated: waveform_adversaries':[\"cochresnet50_l2_1_robust_waveform\", \"cochresnet50_l2_p5_robust_waveform\", \"cochresnet50_linf_p002_robust_waveform\",],\n",
    "                   'Generated: cochleagram_adversaries':[\"cochresnet50_l2_1_robust_cochleagram\", \"cochresnet50_l2_p5_robust_cochleagram\"],\n",
    "                  }\n",
    "cp = sns.color_palette(\"colorblind\")\n",
    "color_values = ['k', cp[4], cp[6]]\n",
    "                   \n",
    "    \n",
    "plt.figure(figsize=(2,2))\n",
    "all_group_data = {}\n",
    "all_group_data_human = {}\n",
    "all_group_data_recognition_averages = {}\n",
    "\n",
    "for group_idx, metamer_group in enumerate(model_groupings.keys()):\n",
    "    all_group_data[metamer_group] = {}\n",
    "    all_group_data_human[metamer_group] = {}\n",
    "    for met_mod_idx, metamer_model in enumerate(model_groupings[metamer_group]):\n",
    "        all_group_data[metamer_group][metamer_model] = {}\n",
    "        if (group_idx==0) and (met_mod_idx==0):\n",
    "            metamer_model_layers = all_model_info.ALL_NETWORKS_AND_LAYERS_AUDIO[metamer_model]['layers']\n",
    "        else:\n",
    "            # For this type of plot, the layers need to be *exactly* the same between the models. \n",
    "            assert metamer_model_layers == all_model_info.ALL_NETWORKS_AND_LAYERS_AUDIO[metamer_model]['layers']\n",
    "            \n",
    "        all_group_data_human[metamer_group][metamer_model] = [np.mean(combined_experiment_dict[metamer_model][layer]) for layer in metamer_model_layers]\n",
    "        \n",
    "        for measured_model in models_to_evaluate:\n",
    "            if measured_model != metamer_model:\n",
    "                measured_model_info = all_model_info.ALL_NETWORKS_AND_LAYERS_AUDIO[measured_model]\n",
    "                pckl_path = os.path.join(measured_model_info['location'], network_network_folder, 'accuracy_on_metamers_from_%s.pckl'%metamer_model)\n",
    "                eval_pckl = pckl.load(open(pckl_path, 'rb'))\n",
    "                average_correct = [eval_pckl['average_correct'][layer] for layer in metamer_model_layers]\n",
    "                # Save the average group data. We will do the plotting afterwards. \n",
    "                all_group_data[metamer_group][metamer_model][measured_model] = average_correct\n",
    "\n",
    "    # First average across the metamer generation models\n",
    "    average_data_from_group = {}\n",
    "    for measured_model in models_to_evaluate:\n",
    "        collapse_data = []\n",
    "        collapse_data_human = []\n",
    "        for metamer_model in model_groupings[metamer_group]:\n",
    "            if measured_model != metamer_model:\n",
    "                collapse_data.append(all_group_data[metamer_group][metamer_model][measured_model])\n",
    "        if len(model_groupings[metamer_group])>1:\n",
    "            average_data_from_group[measured_model] = np.mean(np.array(collapse_data),0)\n",
    "        else:\n",
    "            if measured_model == model_groupings[metamer_group][0]:\n",
    "                average_data_from_group[measured_model] = np.ones((len(metamer_model_layers)))*np.nan\n",
    "            else:\n",
    "                average_data_from_group[measured_model] = np.squeeze(np.array(collapse_data))\n",
    "                \n",
    "    collapse_group_human_data = []\n",
    "    for metamer_model in model_groupings[metamer_group]:\n",
    "        collapse_data.append(all_group_data_human[metamer_group][metamer_model])\n",
    "    if len(model_groupings[metamer_group])>1:\n",
    "        average_data_from_group_human_data = np.mean(collapse_data)\n",
    "    else: \n",
    "        average_data_from_group_human_data = np.squeeze(collapse_data)\n",
    "    \n",
    "    all_group_data_recognition_averages[metamer_group] = average_data_from_group\n",
    "    \n",
    "    average_over_all_measured_models = np.squeeze(np.nanmean(np.array([average_data_from_group[measured_model] for measured_model in average_data_from_group.keys()]),0))\n",
    "    sem_over_all_measured_models = np.squeeze(np.nanstd(np.array([average_data_from_group[measured_model] for measured_model in average_data_from_group.keys()]),0)/np.sqrt(len(models_to_evaluate)-1)) # N-1 because we don't use the same model for predictions\n",
    "    plt.errorbar(range(len(metamer_model_layers)),\n",
    "                 average_over_all_measured_models, \n",
    "                 yerr=sem_over_all_measured_models,\n",
    "                 linewidth=2,\n",
    "                 label=metamer_group,\n",
    "                 color = color_values[group_idx]\n",
    "                 )\n",
    "    \n",
    "plt.legend(loc='center left', bbox_to_anchor=(1, 0.5), fontsize=10)\n",
    "plt.title('Auditory: CochResNet50')\n",
    "plt.ylim([0,1])\n",
    "plt.ylabel('Proportion Correct\\n by Recognition Models')\n",
    "\n",
    "layer_labels = [layer.split('_fake_relu')[0].split('_fakerelu')[0].split('_after_preproc')[0] for layer in metamer_model_layers]\n",
    "plt.xticks(range(len(layer_labels)), layer_labels, rotation=90)\n",
    "plt.xlabel('Metamer Generation Layer', fontsize=8)             \n",
    "\n",
    "# plt.savefig('2023_07_17_CochResNet50HumanRecModelGroupings.pdf')\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}