2. creating_querying_labeled_kd_trees.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "383fbfaa",
   "metadata": {},
   "source": [
    "# Creating & querying labeled k-d trees\n",
    "This notebook contains the core algorithms I developed to produce the results in this research paper.\n",
    "With the goal to quickly identify if the labeled objects are physically close, I created an efficient spatial search algorithm. \n",
    "In the first part, I create labeled point clouds, while in the second part, I store and query them in a custom k-d tree utilized to spatially search labeled data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c217770e",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from scipy.spatial import KDTree\n",
    "import matplotlib.pyplot as plt\n",
    "import pickle\n",
    "import timeit\n",
    "import itertools"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4e634488",
   "metadata": {},
   "source": [
    "# Creating labeled point clouds"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "63db9391",
   "metadata": {},
   "source": [
    "Below is the function takes a single row with 3D coordinates local radius/diameter and bodyid. It returns 14 points around it, without radius."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "3ae9cb22",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Example generated point - x: 0.0 y: 0.0 z: 1.0 bodyId: 1000.0\n"
     ]
    }
   ],
   "source": [
    "def spherical_sample_dataframe_row(array):\n",
    "        results_array = np.empty((0,4),np.float64)\n",
    "        point = np.array([array[0],array[1],array[2],array[4]])\n",
    "        #if diameter_or_radius == \"diameter\":\n",
    "        # If input is diameter, divide by 2 to get radius\n",
    "        r = array[3]\n",
    "        points_array = [point + np.array([0,0,r,0]),point + np.array([0,0,-r,0]),point + np.array([0,r,0,0]),point + np.array([0,-r,0,0]),\n",
    "                        point + np.array([r,0,0,0]),point + np.array([-r,0,0,0]),point + np.array([r * 1/2,r * 1/2,r* np.sqrt(2)/2,0]),point + np.array([r * 1/2,-r* 1/2,r* np.sqrt(2)/2,0]),\n",
    "                        point + np.array([-r* 1/2,r* 1/2,r* np.sqrt(2)/2,0]),point + np.array([-r* 1/2,-r* 1/2,r* np.sqrt(2)/2,0]),point + np.array([r* 1/2,r *1/2,-r* np.sqrt(2)/2,0]),\n",
    "                        point + np.array([r* 1/2,-r* 1/2,-r* np.sqrt(2)/2,0]), point + np.array([-r* 1/2,r* 1/2,-r* np.sqrt(2)/2,0]),point + np.array([-r* 1/2,-r* 1/2,-r* np.sqrt(2)/2,0])\n",
    "                                    ]\n",
    "        results_array =   np.append(results_array,np.array(points_array),axis=0)\n",
    "        return np.array(results_array).astype(np.float64)\n",
    "    \n",
    "x = 0\n",
    "y = 0\n",
    "z = 0\n",
    "radius = 1\n",
    "bodyId = 1000\n",
    "spherical_points = spherical_sample_dataframe_row([x,y,z,radius,bodyId])\n",
    "print('Example generated point -','x:',spherical_points[0][0],'y:',spherical_points[0][1],'z:',spherical_points[0][2],'bodyId:',spherical_points[0][3])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "7502368f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAcsAAAHBCAYAAAAGmZAhAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8rg+JYAAAACXBIWXMAAAsTAAALEwEAmpwYAAEAAElEQVR4nOz9eZQj930dit8qFIAG0OgFve89PT37vvSQsmVFDmVSYpKhlxOJkk8kP5qRKUs2Yyf6vXFOVr+jY+olsWOHjqTYcqJYkvkkxxIlZSiJUkx5ocUhOezpfd/3biyNfav6/v7o+RYLhQJQG9DoYd1zdDTsRi2orvre+mz3MoQQWLBgwYIFCxYKgz3sE7BgwYIFCxaqHRZZWrBgwYIFCyVgkaUFCxYsWLBQAhZZWrBgwYIFCyVgkaUFCxYsWLBQAhZZWrBgwYIFCyXAlfi9NVdiwYIFCxaOOhijO7AiSwsWLFiwYKEELLK0YMGCBQsWSsAiSwsWLFiwYKEELLK0YMGCBQsWSsAiSwsWLFiwYKEELLK0YMGCBQsWSsAiSwsWLFiwYKEELLK0YMGCBQsWSsAiSwsWLFiwYKEELLK0YMGCBQsWSsAiSwsWLFiwYKEELLK0YMGCBQsWSsAiSwsWLFiwYKEELLK0YMGCBQsWSsAiSwsWLFiwYKEELLK0YMGCBQsWSsAiSwsWLFiwYKEELLK0YMGCBQsWSsAiSwsWLFiwYKEELLK0YMGCBQsWSsAiSwsWLFiwYKEELLK0YMGCBQsWSsAiSwsWLFiwYKEELLK0YMGCBQsWSsAiSwsWLFiwYKEELLK0YMGCBQsWSsAiSwsWLFiwYKEELLK0YMGCBQsWSsAiSwsWLFiwYKEELLK0YMGCBQsWSsAiSwsWLFiwYKEELLK0YMGCBQsWSoA77BOwYMFMCIKARCIBQgjsdjs4jgPLsmAY5rBPzYIFC0cYDCGk2O+L/tKChWoBIQQ8zyOTySCbzYLnefF3DMOA4zjxfxZ5WrDwjoPhB94iSwtHHoQQpNNpCIIAhmGQzWbFf9PfC4Igft4iTwsW3nGwyNLCOxuCICCdToMQAoZhwDAMMplMDlnKoUSeNGVrs9ks8rRg4cGDRZYW3pkghCCbzSKbzYokSVGKLJX2JQgCkskkdnZ20NfXJ5Inx3F5+7dgwcKRg+EH2GrwsXDkIE+7GiUyhmHEiDIajYJhGKTTaaRSKQAAy7Kw2+2w2+2w2WwWeVqw8A6ERZYWjhSy2SwymQwAlI20KHkCB8QMAOl0Gul0GsABeXIcJ5Iny1oTWBYsPOiwyNLCkQBNu46Pj6O1tRVNTU2mH4NhGMjLEpSMLfK0YOGdDYssLVQ9BEHQXIcsFyzytGDhnQmLLC1ULaSzk8ABCbEsmxf9mQWlyFLNNkAuedKaqpQ85d22FixYOFqwyNJCVYIQgkwmA57nc2qTagiNZdmc0ZBKQl5HpeSZSqXEhqFwOIzW1lbYbDax29aCBQvVDYssLVQdlGYnKdSQZSQSgSAI8Hg8mohIT2SpZp9y8pydnUVdXZ34M5vNlhN5WuRpwUL1wSJLC1UD+eykUrqyGKERQrCwsIDd3V3Y7XbE43HU1taisbERjY2NqKmpOXQiouQpTdvSGU8KizwtWKg+WGRpoSqgdnayEFmmUimMjIygoaEBV69eFT8Ti8UQDAYxOzuLZDKZR55q9l1OKEWeFnlasFB9sMjSwqGDNvEopV3lUCK03d1dzMzM4PTp02hqahKjU5ZlUVtbi9raWvT09IAQgkgkgmAwiKmpKaTTadTV1YnkWc7mIbUoRJ6JRCKnmcgiTwsWKguLLC0cGtSkXeWQkqUgCJidnUUkEsH169fhdDpLbltXV4e6ujr09fVBEASEw2EEg0FsbGwgk8kgnU5jd3cXDQ0NsNvtpnxPI6DkSa+NRZ4WLBwOLLK0cCiQz05qWeAJIYjH4xgdHUVLSwuuXbumiyBYlkVDQwMaGhoAHGjK3r17F5FIBCsrKyCEoL6+Ho2NjWhoaADHHf7jooY8pY4qFnlasGAODv/pt/COgnx2UitRsiyLUCiEpaUlnDt3TiQ6M0DFBAYGBgAcSOvt7+8jGAxiaWkJDMOgoaEBjY2NqK+vF5t0DhNK5MnzPLa2tpDNZtHe3i4KJFh2ZBYs6IdFlhYqBjo7+dZbb+HSpUuaF22e57G5uQme53Hjxg3T06TyeijHcWhqahKl9bLZLEKhEPx+PxYWFsTIlJJnNYgNUPIUBAE8z4NlWfA8j2w2K/7e8vK0YEE7LLK0UBFIZyfj8bjmBToajWJ0dBRut/vQ6okcx6G5uRnNzc0ADiTuQqEQdnZ2MDc3B47jxGYhr9d76ORJiVPeMCQXo7fI04KF0rDI0kJZIU27qm3ikW+/vr6OlZUVXLhwAfv7++B5vug2ehd7raMjDocDra2taG1tBXAwvhIMBrG5uYnp6Wk4nU6x3un1equChJTIM5PJWORpwUIJWGRpoWww6juZzWYxMTEBhmFw48YNcByHSCQiphSrDU6nE+3t7WhvbwcAJJNJBINBrK2tIRqNoqamBo2NjeB5XhyTOWxIBRKAwuQpFYWvhvO2YKHSsMjSQlmgZXZSCeFwGGNjY+jv70dnZ2fO78oppG4mampq0NHRgY6ODhBCkEgkEAwGkU6ncefOHXg8HjHydLvdph5f7zUqRJ7pdFr8O9JmIapra5GnhXcCLLK0YCrUzk4WiqwIIVhZWcHGxgYuXboEj8eT8/tqEA7QA4Zh4Ha74Xa7sbGxgevXryMejyMYDGJhYQGJREIkz8bGRrhcLlOOacY+5OSZTqdFUfhEIgG32w2PxyOOqVjkaeFBhEWWFkyD2tlJWhuU/z6dTmNsbAwulwsPPfRQUaI96mAYBh6PBx6PB93d3SCEIBqNIhgMYmZmBqlUCl6vVyTPUoILlYJc13ZzcxNNTU3i39Py8rTwoMIiSwuGoXV2klpoSRfSYDCIiYkJDA4Ooq2trei2DwJZysEwDLxeL7xeL3p7eyEIgkiek5OTedJ8DofjsE85R0GI4zjLCNvCAw2LLC0YgjztqiYFJ+06pU4he3t7uHr1qqr044NIlnKwLFtQmm99fR08z+eoC8lHaQ7jGikZYQMWeVp4MGCRpQXd0CtZR8mSOoXU19djaGhIszbsOwlSab5jx46B53lRXYhK89HfU1WjStQOi3X1KpEnrXlKyVOqa2uRp4VqhUWWFjTD6Owky7KiCg51ClGLdypZymGz2eDz+eDz+QDkS/NlMhm4XC64XK6ySvNpGYFRmvGkL020YUhKnrTb1oKFaoBFlhY0gRCC3d1d2Gw2eDwezYuZIAiIxWJYWVlR5RQih0WWypBL862srCCRSGBvbw/z8/Ow2WxiytZMaT4jfws15Gk5qlioFlhkaUE1qGTd7u6u6BOpBYlEAiMjI2BZFufPn9fV4WmRpTpQL8+uri4A5ZXmM4vAihlh7+3tiS8EFnlaOAxYZGmhJJRmJ7US1tbWFubn53H27FksLy/rPpdSZEkIweLiIvb39+Hz+cSZxXfioir9zoWk+TY2NhCJRERpvsbGRtTW1qq+XuVUIpKSZyKRgMvlEsmTwoo8LVQKFllaKAolyTotZMnzPKanp5FKpUSnENqQogfFyFLaMNTf349QKIS5uTkkEglxZtHn81XNzOJhQi7Nl0gkEAqFsLq6miPN19jYqCvdbjYIIWBZNicCtoywLVQSFllaKAi5OwVdeKgFVClQp5Curi6cOXNG3N7IrGQhsvT7/ZiamsLp06fR2NiIdDoNr9eLnp4eEEIQiUTEWc5MJiOOXTQ2Nh6Kg0m1gTYDyaX5lpaWEIvF8tSF6N+yUhq3SsdRY4RtkacFs2CRpYU8lJqdpKICxbC+vo7l5WWcP38edXV1Ob9TS7ZKkJMlIQRzc3MIhUJiw5B83wzD5M0s0s7R1dVVceyisbHxgamHGm28odJ8XV1dIIQgFoshGAxifn4e8XgcXq8XDQ0N4j1SbqghZTXkKXVUscjTghZYZGkhB2pmJ4uRnZJTiNL2ZkSWNO3a2NiI69evq174WJYVoyR6zvv7+wgEAojH43jzzTfF31eLqbMemNl4Qxu6aKRO1YUikYiY+i6nNJ+eCFaJPKVG2ABEgQTLjsxCKVhkaQFAvmRdMYJgWVb8nBTFnEKkMIMs9/b2MD09rXlOUwnSsYtQKISLFy8iGAxiZ2cHs7OzcDgcOZ2j7/QFVSrNt7+/j+PHjyOTySimuRsaGkyR5jMj3avUbSslT0qsdrsddrvdIk8LObDI0oJow8TzvColHqVUaDGnEDnUpHGLIRwOI51O65rTVAO73Z7TOSr1pYxEInC5XGKzkNnWWkcNtPGmvr5ebKySSvOtra2VlOZTexyzr7MSeSaTSbz11lu4ePGi6OVpGWFbACyyfMeDzk5q8Z2Ukl06ncb4+DicTmdRpxAp9EaWyWQSY2NjAKAp7WoUhXwpFxYWEI/HUVtba6q1lhk4zNqrWmk+muZWStXLUYlGImna1mazibV7uRG2RZ7vTFhk+Q6FWt9JJVCyU+sUUmh7LaBp12PHjmFra+vQFiml5he5tVa1uINUS+NNMWm+xcVFMAyTQ55K0nyV6rqVuuEoRZ6ZTCaPPKWi8BZ5PriwyPIdCKXZSS1gGAaBQACBQEC1U4h8e7VkKe92pR6Kao5RCShZa0UiEQQCgRx3EJ/Ph4aGBlVR1FGCHhKTS/NlMhmEQqE8ab7GxkbU1dWJo0aHTf4Mk2+ELSdPuSi8RZ4PDh6sJ9dCSfA8j9XVVTAMg7a2Ns0PcyqVwuzsLBiGUZ12lUNtzTKZTGJkZAQ+n09Mu6ZSqaoe75DW7+QpyKWlJfE7BAKBsgqcHyXY7Xa0tLSgpaUFwNvSfNvb25iZmYHD4UA6nRbnPcvZnSz3WS0GJfJMp9NFReEt8jy6sMjyHQJp2jWbzep6U6ep0K6uLsTjcd2LlprIslC361HThpWnIDOZDF5//XUxijJTo1WKSl2jckR8StJ8o6Oj2N3dxcrKim5pPjUw8n2KkSctdUjTthZ5Hi1YZPkOgHx2stDoR7HtZ2dnEYlEcP36dSQSCcRiMd3nU4zwBEHA/Px8jsiA2m2PAuhYwsmTJwEU1mj1+XyGZeYOO21pFpxOJ2pqajAwMAC32y02WK2srCAajcLtdovkabQ7WUtkWQpS8lQyws5kMuA4DrW1tZaX5xGARZYPMOSzk9JuP7WjG9QppKWlBdeuXRPTiEZGPwoRnlLaVe22clSqxmUUShqtwWAQy8vLiEajBWXm3mmQ/j2pNF9nZ2dOd/Li4mJRaT6txzETUvk9epy9vT0AQEdHBwDkRZ4WeVYXLLJ8QFFsdlItWUqdQqjaDWA8ulPShlUrMqD22EeVVOREQGXm5ubmkEwmUVtbK7qpFJsxPcppWCUUiviKSfNJrxklz5qaGl3HMRtUBcvhcIhjKjRtSyNPac3TIs/Dh0WWDyBKzU6WEjJXcgqRb280sqTbC4KAubk5hMNhVSIDasjS6OJdLVGpXGZOEAREo1EEAgFMTEwgm83mjKnI/04PShpWy3GKSfPRe5o60Ci9cFTyb19qTEVuhG2R5+HCIssHCNK0a7HZyWLaroWcQuTbGxXqpr6EIyMjaGpqElO8arY9yjVLI2BZVhSE7+/vB8/zolKOXBDeyMtMNUIviRUa7aEzwtIXjoaGhopFlkDxKFYNeVJHFZvNJnbbWigfLLJ8QKBldrJQZFjMKUTN9mrBMAwikQjW1tZw5swZsVNU7bblBCXjo7DwSOcRgYNh/1AohEAggJ2dHdhsNsRisZx5RbNRbZFlKShJ89HRnrW1NaRSKdjtdtFVpZxzsfRZVQMl8rSMsCsLiywfAGiVrJOTnRqnECmMRHeCIGB7exvRaBQPPfSQZcRsIjiOQ3NzM5qbm8FxHJxOJziOw9bWljivSDttzR65KDfKRcpyB5qNjQ2Ew2Hs7+9jeXk5J1pvaGgwdS6W6urqgUWelYdFlkcYeiXrpGSp1imk0PZaQNOuDocDnZ2dZSPKdDqNnZ0dNDQ0aD7Gg5TmpcP+ckH41dVVRCIRU0YujlpkWQoMw8Dj8aCnpwdAbrROpfmk6kJGyNPsMZVi5BmLxcBxnKjFa5GndlhkeUShxneyECjZLS8vq3YKkUIPoezu7mJmZgZnzpxBJpNBJBLRtL1ahEIhjI+Po7GxERsbG8hmszmRwYMmN6cFckH4eDyeJwhPO21LdY1WGofVdSuN1oG3pfl2d3cxNzcHjuPE+0trqruc9VH5mhAMBuFwOOBwOHLGWKzIUz3euSvHEUWh2UktyGazCAQCsNvtuHHjhua3Yy2RpVK3687OjunRGyEEy8vL2NrawpUrV8SHn8rNBQIBUW6OpiKVFrdKRZaHXRelEZTH40F3d3dO1+jU1BTS6bQqQfhKfYdKRrDFCExJmi8YDOalutV4n1a6mYjK7gFvR56JREI8R6mjikWe+bDI8giBpl1HRkbQ19cHr9ereR/BYBDj4+OoqanB2bNndZ2HWkIp1O1arBtXDzKZDMbGxuB0OnHjxg0wDCPOqinJzQWDQVF3VK6Y86BAK+ErdY3STlsqCH+YEXolxzm0vDw6HA60tbWJrjs01b2+vo5IJIKamhrxmsnrxIfZeSsVKAHyO+kBiAIJlh3ZASyyPCKQpl3pW6EWEEKwsLCAvb09XLhwAbOzs7rPRc1DI027yrtdS815agGtuR47dkxUQim2b7mxs1wxJ51OY3NzE83NzVXjTakXRhY3JU/KUCiUIwjf0NCAbDYLnucfGEF4I003QH6qm5KnkjQfz/NVMaYCFCZPqiNNPT7fyeRpkWWVQ2l20mazaSLLVCqFkZER1NfXY2hoCDzPl20OT6ojOzQ0pJi+MyPVSQjB2toa1tbWNNdcpZAr5rz55pvgeT7Hm5LW8eRD/+8k2Gw2RVutjY0N3L17t2yC8JWG2U038vtLWife39/HzMwMmpqa0NDQUFY5Q60Rs1LDECVP+vt3mhG2RZZVjEKzk1pqhlRG7tSpU2KTAh1wNhtqRQaMkiUhBKOjo+Koi1lRDV0AOjs7xRm8cDiMQCCAtbU1CIJQtjGCowZau1taWsLQ0FCeIDxNPzY2NhoWhK8kylkbldeJ33jjDfT29iIcDuuS5tMCo1GsEnlms9k8I+wHmTwtsqxSFJudVEOWcqcQ6QiF2TVDoHjaVQ4jadhoNIpYLIa+vj50d3fr2odaSFORQO4YwcLCglgPrcZoqpLasECuILxU3HxpaSlH3Nzn81V1eluLUIBREELg9XpRX18vSvNRdSEtTVZqYHZ9VIk8qRF2JpPB/v4+Ojo6ckThjzp5WmRZZVAzO1mKLJWcQqQws+NTTdpVDr1kvbGxgaWlJbhcrrIRZbFrIx8jSKfTCAQCedGUz+crOrdYKZWgw1qcGKawuPns7CySyWRRfdbDhNGapVbIm26onGFfX1+ONB8dgyqmBVwM5a4rM8zbdmTxeBx+vx/Nzc05kadc1/aokadFllUEtZJ1xciykFOIFGbdpKVIuRC0krUgCJicnEQmk8GNGzdw584dvaes6tzUwuFwKEZTdG6REoLP56sqQjATavV85YLwcn3W+vp6Mb19mLXhSpJlqWdALs0n1QJeW1vT1KFc6c5bm81W1AhbSp5U17baydMiyyqBPP+vVdu1lFOI2chkMrh7925RUi4ELWQZj8cxMjKCjo4O9Pb2VmzWTiuUoikpIWQyGXFhe1AUgvRCiQSoPuvKyoooMefz+VBfX1/R2nAl07BaIdcClnYoLy8vA4BYNpDX1CtJlkpRrDTyBAqTJ03bViN5WmR5yJCnXdXcIHKyVOMUYhZo2jWTyeBd73qXrjqK2prlzs4OZmdnce7cObFuWG6YlaKWp9RoVBAIBBCPx3H37l0x6qyvrzd9ITtKhCyfhaW1Yb/fj/n5edhsNqRSKYRCobIJwlNUkiyNHkfeoSyX5qM190qPqahJ+UrJk96r1MszGAziueeewxe+8IWyn6sWWGR5iNArWceyLHieB6DeKcQMSNOuLpdLd8NBKUKihByNRgvWQQ9bAUcrpFFBMBjEhQsXEA6HsbOzg7m5OdjtdtNFzo/S9ZFCqTb8xhtviCo5VEiisbHRdEH4StcszYSSNF8wGMTOzg7i8TiGh4crMt6jlZil8nsAEIlEsL29XZZzMwKLLA8Bcsk6rTcty7Li7CQAVU4hRkGjPJp23draMuQxWKjmSsdPmpubcfXqVcX9l5MEKiV3J5dNUxI5p5225Zy/OwpwOByw2+04ffo0gLeFJJQG/fUKwlMctZewYpAKcITDYZw9exahUAibm5uYnp7WJM2nBVpnOuWIxWJwu92mnIuZsMiywqAt1jzP687LJ5NJrKys4OTJk+jq6irDWb6NQlEeTQXreSgKERKdCVUzfvIgLWqAssh5IBAQ5++8Xq9InkZGCI4i5PdKsUF/aWOVnlnFStX2yjXrXAxOp1NRmm9tbQ3RaNS02Vie5w29vMdiMdTW1urevlywyLKCEARBjB5Onz6t+WYkhGBlZQXr6+toaWkpO1FK067yKM9IBCbflhCC+fl5BIPBvJlQJZgpl1fq3A7jGNLhdWn3aCAQUK3TepRqlqVQ7MVISRBePqso7bQt9aJRLYLtlYD8BS2RSCAUConSjx6PR7zPtETsPM8b6v62yPIdDGkTDwCxA0wLpGLhp0+fht/vN+W8Cp2HPO0qh15PS/m26XRalOK7fv266uvyIJFBKUi7R5V0WqUGxtIGmGrtHNZzDC1jSfJZRakKU6kXjUpFlpXsTlUDaTd3oYhdqi5UTFjCaDMRFbGoNlhkWWbIZye16roCEMcPBgcH0dbWhmAwaFiBp1AaVU1zjXR7PaCRFf1eJ0+eFGt3as+91CKtlygqVbM0AnkXZDqdRigUymmA4XlejAyOerraSLSnpMJEx1Sklm2NjY2or68/NN/Mch9L63dSitiphZtUWIK+eEjT3UZrltFo1Ios32mgTTxSyTqbzSZ2spaC1Cnk6tWr4tucEaKiUCJLmnZtbW0t2FxDYZRUkskkZmZmcr6XFlQ7oVUSDocjz0llamoKW1tbWF1dFSMCn89nuqlzpaJXs47DcVyeIDztGJ2bm0M8Hsfq6iqamppMbXqR4zDtufSAYfIt3Ar5n6bTacMNPhZZvkNQTLJOLdGlUimMjo6irq4OQ0NDuvZRDPJ9lEq7Km2vh7AymQxGR0chCELe91ILNceOx+OoqanRvP+jEFmWgsvlgsfjQUtLCxoaGsRFjYpW0BqeUSeVSurPlou05JZtr732GmpqakxvepGjkvOc5SBmlmUV093BYBDBYBDRaFS8blpVmeLxuPgyU02wyNJklJqdVEN0Sk4hWvdRCpRwBEHAzMwMYrGYam1XQJ++6/7+PsbHxzEwMIBkMmnoAS60UFNpvHA4DJ7nUVNTIw68v9NGMOj9J48IaBpydXXVkFrOg0CWcrAsm9f0IhWEV1u3K4XDVtQxG9J0dyQSwfHjx0WBAarKJG20KtYta42OPOCQz04WGgspFhUJgoC5uTmEw+GiXaFmkCXDMIjH47h37x5aW1tx6tQpTQuSlgiMEILV1VWsr6+L3pMLCwt6T73gsROJBO7du4f29nYcP34cDMMgkUhoGsF4ECLLYpA2AwFvq77s7e1hfn5e9KX0+Xyq0pCVqu8dxkuOkoQhjdKp36lU/1fLSE8lu2Er3UzE8zzsdrvoNAMo14ppvVP+khaNRuH1ekse56mnnsJ3vvMdtLa2YmxsLO/3hBA8++yzuH37Nubn50cA/BIh5C4AMAzzfgC/D8AG4I8JIc+VOp5FlibAjNlJLaLkZpBlOp3G+Pg4zp8/r1nbVcs5ZLNZjI+Pw2azmeY9qURo1CLs7NmzaGhoQDqdzlnsuru780YwqD9lpTVIqym6lau+UF/K9fV10UmFvlzIxwfeSWMWgHLdjt5PY2Nj4Hk+J8VdTcLmlbx+Sg0+SrViuaTh5OQkfD6f6prlL/3SL+FTn/oUPvrRjyr+/qWXXsLs7CxmZ2fBsuzHAXwOwEMMw9gA/CGAnwGwBuB1hmG+RQiZKHY8iywNopjvpFpsb29jbm5OU71QL1nStGsikcCFCxd0EaXac4hGoxgZGUFfX5+pM6FSspTOaNI0cqHIUD6CoRRV8TyPWCxW1uYOet7lhF4iU/KlpP6ddOCfkmelXi6qVYBCaaRHLggv7bQ9LGFzo92peo5X6rvJFazS6TRWV1fx9a9/HT/+8Y8xPj6Omzdv4qd/+qdx/fp1xZrne97zHiwtLRU8xosvvoiPfvSjdL34McMwDQzDdADoBzBHCFkAAIZhXgDwBACLLMsBadpV3sSjFnqdQvSSpbTbtbm52ZDKRql0JfWevHDhgqqUip5j0xnNuro6xRnNUousUlQ1MTGB7e1trKyswOPx5EjOvdMgj8zpwH8gEBCdVNLpNHZ3d0tGUkZQrWQph5IgfDAYFF/GqD6wz+eruLB5pSNzrX8vh8OBJ554Ak888QR+8Rd/Eb/1W7+FpaUl/Mmf/Al+9Vd/FV/+8pdx7tw5TftcX19HT0+P9EdrALru/29V9vOHSu3PIksdUOs7WQxGnEL0kKW823ViYsJQKrfQOfA8j8nJSfA8XzbNWoZhEA6Hsbi4iBMnToidjNLf66k9Op1OeDwetLa2or6+HrFYDIFAQKxP1dfXw+fz3W9QsIHPhACSBRgONns9DrI7Dy6kA//9/f2ijm84HMby8nLezKJZC/RRIUs5OI7Li56okXMwGARz35aqHILwUlSbAEIpxGIxDAwM4F3vehc+/OEPA9CXiSmwDQGgdKFLHsAiS40wI+1KIyK9TiFajknTrvF4XFHbVS+UyIg2DHV1daGnp6csDz9VFllYWMDVq1dN75qj34th3jYslnaRBvzb2Fr9OziZVdQ4AYfDCYfDAdZWA6f3LJyeQbA2c2cZqxU2mw12ux3Hjx8HkDuzODs7C4fDIUbmRsjgqJKlHA6HQ9Rm3d7eRiQSAcdxoiA8bYgxWzy/0pGl0RIDLYNIoedadHd3Y3V1NedHADYAOAD0KPy8KCyyVAlCCFKpFJaXl9HX16fr5stms5iYmADP87h+/XrZBbGlaVd5t6tRspR39W5tbWF+fh7nz59HfX29ofMuBJ7nMT4+jnQ6jfPnz1e0vZxlWdTX1cCWmkXWEQDYVqRSAhLJJILhBDhbAjXhv0KNexyNnR8AZy+vXVopVIJg5MeQzyxSoW45GdAxHr3HeRAgCAKcTic6Oztz5OWkndtSMQkjWquVrlkaRTKZNPR9KW7evInnn38eTz75JFiWfRjAPiFkk2GYXQAnGIY5BmAdwJMAPlJqfxZZqgCdneR5Hpubm+jv79e8j3A4jLGxMfT19SGRSJT9TY82DRUyTjYjshQEAYIgYHp6GolEQlPdVStisRhGRkbQ09MDm81WtutXKH1LCI/o3l+Cz+yDcxzUODkOooZlJptFMpFAeH8Pe3tfBql5Nxp9LYYXumpGKRKTC3XHYrEcuTSq+FLKSaWSXbeVgjw1KhfPlwrC0/qwXjGJo9h5q2YfH/7wh/HKK69gb28P3d3d+Pf//t+Lo3vPPPMMHn/8cdy+fRuDg4MA8EcA/i8AIIRkGYb5FIDv4WB05E8IIeOljmeRZREo+U5qfaCoU8jGxgYuXryI2tpabG5uGraxKQQpeZXSdjWyOLAsi2Qyiddffx2tra26XFTUYnt7W4xa6+rqEA6HKz4LmUluIJvaBedQVhaxcxzs98cKsmk/4CQIJw4ahrLZbI549zsR0rR2MScVWhOWRkKVJMtKquoUi/aUBOGVxCToPVVsX3TusRIwmvLV8lz/2Z/9WdHfMwyDP/zDP6T/eUF2nNsAbms5N4ssC0AuWafnIZI6hUhnDM2Yk1RCPB7H6Ogo2traSpKXHgUeKWKxGPb29nDp0iXd4yelQEXdqboQfeDLKRxQaN/JyARYVl2EyNrcYIRF9PU9gf7+/jyXkHg8juXlZTQ3N5fNsb7a3EDkKOSkQsdUpJ2jlRIlqGQEpnV2VElMYn9/H4FAAIuLi6KCjlJzVSXTsGYdqxrT7hZZKqCUZJ0ayJ1CpCgHWZZKu8qh9xwIIZibm0MwGERfX58hoiy22KZSKdy7dw9NTU24cuVKzucqrbJDCI9schs2u7rvyrA14LNBECEJxubKcwl58803UVNTg42NDUQiEbhcrpxanlkLRaVrlkag5KQi7xx1Op2ma7RKUcnI0mgEJh/yp84z0uYqSq6VHlMxQpaHpdakBhZZSmDG7GQhpxAptDiPlDoWIURV2lUOPWSZSqUwMjKCxsZG9PX16TllEdKuUzkCgQAmJydx+vRpRUHlikeWhAfDqCcfhmHAgAEhyn9jlmXR2toqSqhpleSrFpTzhUXaObq3twe/3w+WZXM0Wuk1MstJpZojy1KQO8/Q5qq1tTX4/X6EQiEkk0nNRs5aYZQsE4lEVerCAhZZijBjdrKYU4gUZgmhx2IxjI2Nob29XXPNkGVZsRarBpTAqLj7+vq6pu2Vji9fnAghWFpaws7ODq5du1ZwEay4fivDAQwLQnhVs5SECAAIWLY00VWrJJ9aVKqW6HA40NXVlafRSu2hzHBSqXTNspzELG2umpqaQmNjIzKZjGYjZ60wSpbVavwMWGQJQPvspNJDVcopRAqWZQ1HltlsFsPDwzh//ryuphG1DT6EECwuLmJ3dzeHwIwSlnz7bDaL0dFR1NTUlLTuqnRkyTAsHJ4TSEfnYLOXHosR+Cjsrl4wKshSDjWSfFQlptjsYrXXLI0cR0mjVd78UkhmrhiOYteoGhBC4Ha74fV684ycqeCG2s7kUjCa8rXIskpRzHeyEOSmyWqdQqSw2WyGtF2np6eRTqdx48YN3Sapahp80uk0RkdH4fF4TPfUlJJSJBLB6Ogojh07ho6ODk3bVgpOzwmkIlMlo0tCBBAhA2ftaVOOqyTJFwgEcmYXKXnKI/Fqrf1oRak6llLzi1RmTq2TyoMUWUohj/aUXjaoF6W0M1mNnZYcRht86D1djXjHkqXetCutN9psNk1OIVLoJZp4PI6RkRG0t7cbTsmVOof9/X2MjY0pNigBxgmLRrZUQ5aO1aiBmmPrPb9C29nsjaipu4hk+B5s9kZFwiREAJ8JwFl7Cpwz/5qZAafTmTe7GAgEctKRtIO03KhW1xG5zBx1UllbW0MkEoHb7VZsqDrqhsx6jyX1opQKwgcCgZJ2WnKYkYbVGwCUG+9IssxmsyV9JwuBRoVanULk+9CahpV3u4ZCobJou9K50M3NTVy5cqVgsd2Muuv09DQEQdCsIXsYkSXDMHDVXwLDMEiER8AQgOE8YBgOIDx4Pg6AR433LFwNxV+czDr/gpJ8gQCCwSDi8Tiam5vh8/lQV1dn+uJ8VOYf5U4q8XgcwWAwp6GK1jqrhcAO81hyQXhqpyWP1BsbG/NGn6ya5QMCM2YnGYYRF3kt3adSaCGaQgo5ZsjVybfPZrMYGxuD3W7H0NBQyaFpvQt+IpFAKBRCb28vBgcHNf8d1Bw7lUqBZVnNwg/F9s0wLFz1l+HwHEc6toB0bB5ESACMHTXeM3B6BlXVNMsFaToymUyiq6sL6XQa29vbmJmZgdPpFBfBcnZEmg0zSVmqlEMbqqLRKAKBAFZXV5FMJsGyrEQwvzxL5GGmYbVCbqdFI3U6+kTvKzqmYkQAwYosqwBmCKDTlFd3dzdOnjxpaCBbDdFJ067yblezhdBp3bC/vx+dnZ0lt9d7fNoI5fV60dnZqfuFpRhZrq6uYnl5GQBEMW+fz2fafJ6N88JVfwmu+kuG91VO2O12NDQ0iOMEdERlcXFRFKumi5weSb5qTcNqAcuyolJOfX09dnZ20NzcnJOCpClbM6Pzao4sS0EaqQMH9xUV3AiFQnC5XBAEQZcgfDQatcjysEBnJ7e3t7G7u4szZ87o2s/6+jqWl5fR1NSEtrY2Q4uEzWZDKpUq+plSwuRmRpZra2tYWVkxvW4oBZ0/DQQCuH79Oqanp3VHpoWOTe3BBEEQ/S1pQwydz6ME4fP5FLMCh5HiVQThAbCAiWTkcrlyxi+k3pRUkk9Jbq7gKR6RNKxaCIKQ020MvO2kIo3OafRuxEml0rZZ5bx+LpcLLpcLnZ2dmJubQ01Njdj4KE1zq3kpsyLLQwIhRBRA19uBSp1CAODGjRuYm5szPPZRjOho2jWZTBYVJjeDLHmex+joKAghmuuGWo5Pu2q9Xi+uXbsGlmUNkZLStolEAvfu3UNnZyd6enqQzWYhCAJqampynB0oQYyNjSnOMB5aapIQsGQOXPb74Pg3AWQAcMiyV5DlHoPAnjSVOKXao3JJPiqfRqPOurq6Q03ZHiYpy51UaBQl70bWOq9YzUo1RiAIAmpra9HQ0CBqANM0N30pq6urE1/K5OtbPB7P86etFjywZClPu3Icp5nkpE4hXV1dAMxR3ylENMXSrnIYPY9kMom9vT2cPHkS3d3dZakbAm931cpNmo1o08qPTVO7pZqt5AQhnWGcm5s78KVkWXi93sraQpE0HJk/Ase/AYCAwA0wHoAIsAl3YUu/BYG9hJTjEwCjLl2q9UVESW4uEAhgY2MD09PTcLlcOaRQTIHJbFTTcaRRlNRJRTqvSImglJPKUTJkVgv5nKU0zU1fyuiYysrKiigIn06n0d3drbrB57vf/S6effZZ8DyPp59+Grdu3cr5/X/4D/8BX/nKVwAcBDzj4+M8gBZCSIBhmCUAEQA8gCwh5Lqa7/bAkWWh2Ukt5KLkFEJhZEZSug/5uWj1gzQSWW5tbWFubg5erxc9PT2lN9BxfEII1tbWsLa2pthVazSyFARBFEzY29tTPeMqhXyGMZlMYn5+Hnt7e9jZ2SmZstWLnAWZCHBkPg8u+yYI482NHhkWgBcgBDb+LTjTf4iU45/d/7nG42iEw+Eo2kFaV1cHh8PxQI2oaE2NSruRaRRFiWBtbU3MXCg5gzyIHp1A6TlLKpAvnYkNhUL4yle+gi9/+csQBAEbGxtoaWnBQw89pPjc8TyPT37yk3j55ZfR3d2NoaEh3Lx5E2fPnhU/8+lPfxqf/vSnAQDf/va3cfPmzR8RQgKS3fw0IWRPy3d7oMiy2OykWrIs5BRCYYb6jpRoBEHA1NQUUqmUJj9IPWQpPdbly5cxPT2t+dwpipEdz/M5qWulh8coWVIFI5fLhevXr5vyll5TUyPKpnV0dOSkbHmeFxs9zJSdY4VJcPzdfKKUgmFA4IVNGIFNGAFvu2zKsdVCqYM0EolgfX0dwWAQb7zxhi7FHLWoVrKUQz6vSIlAyUnlQYXWzlv6wvrss8/i2WefxSc+8QmcPHkSX/nKV/Drv/7raG9vx+/+7u/m9JrcuXMHg4ODGBgYAAA8+eSTePHFF3PIUor7Vl7F/bzUnKvRHVQLqAB6oW5XNWRZzClEy35KgRIdTbt2dHTgzJkzmrVdtZAlFVBoa2vDmTNnRFcVvSh0fGrS3N3dXTS9ayQyTqVSWFlZwenTp1Up/miBNL1YKGU7Pz8Pu91uqMuWvijYsy+BgCldj2QYEMKCy75UcbKUg0ry8TwPh8OB/v7+vGtDScFIEwxFNaVhtUCeuZCmtumzT69TuUZ5Kt2sZnRMJZPJ4ObNm7h48SIAYGVlJa+0sr6+npMR6+7uxmuvvaa4v3g8ju9+97sA8L8kPyYAvs8wDAHwBULIf1NzbkeeLNVK1hUjOan+aSGnEOl+0um0oXO22WyIRqMYHh7GuXPnVKVd5dBCNtS2R1rTM3v0BHhbOEFNKllvZLm1tYXV1VW0t7ebTpTFoJSyVeqy1TSGQdKwCaMgUNv954ZNmAZIHGCKOzNUUhtW6dpIm2CoaLeSJJ+W45Qb5T6ONLUdiUQwODiIYDAoipvTrlGfz6drlEcJle66NUMb1uv1iv/d29ub9xmle7vQ3+3b3/42fvInfxLf/va3pSnYnySEbDAM0wrgZYZhpgghf1Xq3I40WWrxnSz0O7VOIRRG07CCIGBhYQGJRALvfve7dQ/wqnENkerWygUUzBw9oSbN0WhUtVCDVrIUBAEzMzOIx+MYHBxEMplUfX5aoLbxqFCXLe34U5eyTeJgPETl4nI/umSQOGgCUvFdDgNSx4tiknyNjY2qOrArPTpSKVC3GamTCnX3SafTOfVOvetEJb0sAeONS2pGR7q7u7G6uir+99raWsHZ8BdeeAEf/vCH5ee4cf//dxiG+QaAGwAeTLKU+k4C2iXrKLQ4hVAYScPS1EtzczPS6bQhpYtSZJBMJjE6Ogqfz6eoW2uWtiv1uPT5fLh69arqv4OW40uPceXKFezs7JTcppLpJ6UxjGAwCL/fL8qDNTU1iSnbt+EAIACEqBsLIQQMeBBUh8elGhIrJsm3vLysauj/QYksi4Fh3hY37+vrE7tGqWg+AHHMqa6uTpOTSiVt3YxeQzVkOTQ0hNnZWSwuLqKrqwsvvPACvvrVr+Z9bn9/Hz/60Y/w5S9/WfwZwzAeACwhJHL/348C+G0153bkyFI6O6mXJGnEtb+/r7mLUi9Zbm1tYWFhAefOnYPb7UYgECi9UREUI0u/34+pqamC5smA8aiDYRik02m88cYbml42KNRGfqFQCOPj4zh58qQot2Vk7KQUzBAlsNlsiinb5eVlRKNRZDIZ7OzsoLW1FTVsPxhhDVARKQJJCEwnoDptW17oWRjlDiF06H9rawszMzOoqanJq+NVasyiUilLNddN3jVK9Vl3d3cxNzcniidQfdZC+6t0GtYo0ul0ycwUx3F4/vnn8dhjj4HneTz11FM4d+4cPv/5zwMAnnnmGQDAN77xDTz66KPyUZQ2AN+4f704AF8lhHxXzbkdKbI0Q7KONro0NzeLKi9aoDW9x/M8pqenkUqlMDQ0BLvdDp7nTTF/lu+DquT4/f6i5slGQUdraCpZj3lsKVIihGB1dRUbGxt5oydHreVenrJ96623kE6nMTExgbqaEzjVOQfGZofNxhX+boSAQRZp7h+qikKrQoVIBZSG/mn3KK3jpVIpZLPZsp9LNXfdKumzBgKBHCcV+RwsUPk0rBnXT835Pv7443j88cdzfkZJkuKXfumX8Eu/9Es5PyOELADQpVN5JMhSmnYt1sRTCplMBnfv3tXlFEKhJbKMxWIYHR3N63Y1w7FDvg+qklNbW2vaKIUSpGLrbrdbt8u62tETJUF3tZGlnsWv3HJ3tCGmu7sbTqcTfPYMhMQYWH4F8aQDDMPCxnHgOA429r7cHSFgEIXAdIK3DWk6VjlRDnJRkuSbnp7GwsICFhYWNEvyaUGlojAz1HvkVm3xeByBQCBPYo7juIqlYY0+N9U+e1r1ZEkIQSAQEFMyei4mje6y2Sxu3Lihe4EH1IsSbG5uYnFxUbHb1ay3L3oeNFUpV8kxG9FoFCMjI6LY+quvvqp7X4VIidZ1qWyd0rWq5gdKLeh3t3E1EGr/JRypz8LhWINAGGSzBy8/PJ+FnRPA2QTA1oVMzacBpjrqlUD5o1daC/Z4POjt7YXL5cqbW6TjO8VSkWpxVIXhpXOwVBwhEokgGAxiZ2dHdODRY+asBWZcv2omzKomS5p2HRkZwcMPP6zrItK5v87OTlNcA0p1wyqlXcsBeh5LS0vY2toq6j1pBij5X7hwIae1Wy9og5AUtOGKenYWQrXXLLUftA5J578Gl/0xOP5/w2Hfg91uA8AgK/iwE3kXlrf6kc7MlTWy0opKN94UkuRbW1tDNBrNkeTT8yxUKrIsd2qUzsHW19ejtrYWwWAQPp9PdAaR1o3NdFIxOmNZ6ZSxVlQlWcpnJ2k0p/VCUqeQ8+fPo66uDoFAwLCgQLE0rJSYtYoMaIUgCPD7/bDZbLhx40bZbjK5sLtZb6VSwpM7kpRquKrWN09DYGqQtb8XWe7vgSG7OBgrqQFhmtHgYdHQDlHsnEZWUocM+fB/pYjsMLtUlST5pKnIOq8T9e4QnLYtMMiCtXlQ4z0Lh+c4GCZ/UT+qkWUxCIIAu92e95IRCoXEpirqpGLUxs4oWcbj8ao1fgaqkCyVZiepCLrahVruFEK3M0N9p1Aatlja1WxQgXen01lQ4skMJJNJ3Lt3D21tbSWF3bWCkmUmk8Ho6CjcbrfoSKJm23JFf4cSWeaeAAijnEqXR1a0yUM6/C+1lyo3qmmkQy7JFwu8isjuXyIbSSORPZBssXM2xPYnYXd44G19HDW1gzn7qGTN8jC9LB0Oh6KTChXYoCISWp1UzBAkKGd2zCiqhiyLzU7abDZks1lVIx5KTiEUZpClPAXI8zympqaQyWTKmnYFcsXJz507h8XFRVP2qbQQ0fGTM2fOlGXxZVkWsVgMr7/+OgYGBkQjWTUoN1keFcibPKRWSOFwGIuLi2hpaamKlK0RaCFlQghie68gHroDu9MLRw0n/jydSSOTziCRCCMW+TLY2kfQ2HpZjMqruRtWL9REe0pOKoFAQHRSoXrJjY2NRdc3ozOd1exlCVQJWcrTrnp0Xemowfr6ekETY7PIkoKmXbu6ugo2o5Q6Z7Xb0GiZYRjcuHHDtPET+TlIx0/0OHmoRSgUwvb2NoaGhjQ/IGrIMh6PI5lMor6+Xrdu61GCfKh9eHgYDQ0NqlK2elFNkSVFJrGMeOh1sFw9GIkyEsMwcDqccDqcAGrBZ5PIpP4Wq8tuRGJZ1NbWIh6Pl1TFMgOVjiy1lE6URCSoOMLq6qpoqaWkTmU0DUuzI9WKQydLNZJ1pUiulFOI2v1ogdG0K+1mVXNzRaNRjI6OoqenB93d3QAOFhCjZEmjZPrgZjIZjIyMlHX8hNZA9/f30dXVpevhKEWWtFbtcrkwPT2dk55UUw+tRJdnJQjZ5/OJEXuxlK3eF6JqJMt46HUwDJdDlEqwcTVgkEJ/kwBXww1Eo1FMTExgYWEBgiDkRFNmd48edhpWC6ROKsDbllpSdSpa78xms4bJ0qpZKkDL7CTHcQWHkqlTyPHjx0um8swgS57nkUgksL29bajhhZ5LqZtrY2MDS0tLeV2oZs9qhsNhjI6OFnVcMYpUKoV79+6hubkZx44dQzQa1bWfQmRDiTiVSuH69eviC5jcqb0cVlvVjmIp22w2W1VdtnKoJUshG0M6vgzWpq5bm2FdSOy/CXfjELxeL2pqanD69GlwHIf9/X0Eg0HVknxaUG1pWC2Qi+anUikEg0Gsr68jEAjAZrOJDUNanVSsBp8iKJR2lUOJ5KROIWrHJoyKoNO0q81mw6VLlwy9VZciO0EQMDk5iUwmo0jKZkQn9BzW1tawurqKy5cva75Z1S5kdBaUSuPt7OwYOn/5tpSIW1pacPr0aQiCIJKlXHMzGAzm2ElR3VapvNqDDKVrorbLVopKdXWqJks+BgYaZrEZO/hMWNy/dERF2iilRpJPC8wQJdByrHL+jZxOp9iRvLm5iVgsBgC6nFTkjiPVhkMjSxpNqlmY5GRJnUK8Xq8qpxAK2lWrBzTtev78eYyNjenahxTFyFLqc9nb21u2wXyGYTA1NQWWZYumr4ttX2ohk8rWSe3PjETG8vtmf38fY2NjqjRq5bqtUnm1RCIBh8MhOrqUs1mr3NBC+Fq6bKULXqVeKtSnYbU+E7mymYWIRS7JF4/HRWutRCKRc33UOO5UUty80ilfqsDU3d0tKjDR7F82my2a3o7FYlZkWQhq3+KlJKfHKYRCja2VHNJuVxrhaak3FkKhlDD1niw1mG8U9IHv7e3F4OCgLvIt9bLD8zzGx8fBsmyebJ3RCI5uS6NivaIMUnk1QRCwvr6O7e1tjIyMAAAaGxvR1NQEr9db1QPTStD7QqWUspUueDRlWylLK7VkabPX37cxy4JhVFh/CUlwzrdLN2qPI7fWovZsY2Nj4Hm+ZEq7kpGl2WnYUseSvixQBaa6ujoxg6GU3mZZFm1tbaq7Yb/73e/i2WefBc/zePrpp3Hr1q2c37/yyit44okncOzYMQDAz//8z+Pf/Jt/Q8/p/QB+H4ANwB8TQp5T+/0OvcFHDWw2G1KpFGZmZnQ5hVBwHCemCdSgULer2npjMcgjKz2ekHpBCbm+vh7t7e26H1zaIKR0HeLxOO7du4fu7u4cV3PptnrJkl67iYkJ8SXGjAWBZVl4PB40NDRgcHBQTL9tbGwgEomICjE+n8+QZOJRgjRl29vbK6Zsg8Egtre3wXEcUqkUmpqaTOuylUMtiTGsAzV1l5DcvwuGK910R0gG7sYb4n/ricKkhEDt2eQpbZqGpJJ8R6nBRwtKrYlK6e1QKIQ///M/xx/90R/B5XLh/PnzeOihh3Dp0iXF8+Z5Hp/85Cfx8ssvo7u7G0NDQ7h582bevPlP/dRP4Tvf+U7Oz5gDJYo/BPAzANYAvM4wzLcIIRNqvt+RIEue57G6uoqenh5dTiEUanVdgbcba6j6j979FIKULJPJpOiEosUTUisIIZidnUUkEsHQ0BBmZmYMG0ArEd7u7i5mZmaKRsdGyDKVSiEUCqGpqQl9fX2miyXQ85Km36QKMTMzM3kmxtXWFFMuSFO2drsdLMuC4zisrq4iEomIA+0+n89U1xu1f2NX/RUk9+9BEFJg2cIv1AIfhY2rh8M9oOs4haCU0g4GgznuIPQFpBKoZmKmTiqf+MQn8IlPfAK/9Vu/BZZl8bu/+7u4d+8eTp8+jX/9r/81Lly4IG5z584dDA4OYmDg4O/25JNP4sUXX1QrznIDwNx95xEwDPMCgCcAVD9Zqrkxt7e3RbeB48ePGzqemgYfnucxOTkpiq4rpZmMNgoBb0enNK1cruF/inQ6jXv37qGxsVEkZKMdtXKBBqlsXanoWG29Wo5QKISxsTG4XC709/frOW1dkItV05RSIBDA4uKipjnGSjQRVaqe6HA40NbWJqZs6UA7LV00NDSINapKvFBwDh/qOn4W4c1vQBAyYGy58m2E8CB8FIzNg4bOfwyGLe8SKG2AoS9cc3Nz2N7exvb2Nurq6sQXrnLUyCudhjVyLEIInnjiCTzyyCMghGBiYkK0JKNYX1/PyVR1d3fjtddey9vX3/3d3+HSpUvo7OzEf/yP/xHnzp0DgC4Aq5KPrQF4SO35VW1kSQXJk8kkzp49i62tLcP7LNXgQ+cZS4kMmCVusL6+jnQ6Xdbhf0DZQBkwPn4i3Z7K1nk8HlWydXrE0KWiE9PT07rPu9R5qW06k6aU5E0xXq9XU9NHOVDuupg8PSofaJe+UCwtLYmGxuVM2QKA0zOAhu6PIOb/a6QTy/RswYAFwMDpPQdP00/BxlW285K+cNXV1aG2thZNTU15A//SsSYzIsJqSsOWgrTBh2EYSnA5UHo25ffR1atXsby8jNraWty+fRs/+7M/i9nZWUC5A0z1W2VVkqVckDwWi5kiJlAsIiyWdpXDaBo2nU5ja2sLXq/XUFoZKF7PoSbNm5ubOZ2oFGZElrTBYXR0VJNsnZboitYnBUHA0NAQBEGouvEOeVOMtOlDEIScRfBBQalaotILRTAYFFO2Ho9H/L3ZRuX2mg40dH0QfDqITGoDRMiCYZ1wuPvA2g633kwJTGngn9pqzc3NwW63i9dHr8D5USPLUg0+3d3dWF19OzhcW1tDZ2dnzmek6/fjjz+OX/3VXwXDMM04iCSlDRTdADbUnl/VpWHlTiGAeco7SvuhaVee51WLDBhJw9KuQpp6MWNWU+kGlZo0F3IlMZoOZFkW29vb2NraKigxWAhqj03F3Nvb28UxGkJIyW2NNC0ZJWJ504d0EZydnRVfltrb23Nc7R90yFOS8pRtOWrANkcjbA59Ru/lQiEC4zgOLS0tYvYnmUwiEAhgeXlZjLq0qi5V2uHEqJB6qVru0NAQZmdnsbi4iK6uLrzwwgv46le/mvOZra0ttLW1gWEY3LlzhwYEfgAhACcYhjkGYB3AkwA+ovb8qiayLOQUAhibj5RCTpY07drd3Y3u7m7Vi5Ye8iaEYGlpCTs7O7h69Sp2d3cNNwnRCFe+sFCTZiUxeSmMRJaCIGB/f1+3dZeamiV9sZDXc4+acIB8ERweHgYA0UpKShJmjWJU4voYkbsrlbJlWVYkhqP0t1YDtaRSU1ODzs5OUeCcqi5RsRJ635TT0FkLKhFZchyH559/Ho899hh4nsdTTz2Fc+fO4fOf/zwA4JlnnsGf//mf43Of+xw4joPL5cILL7yAn/iJnyAAsgzDfArA93AwOvInhJBxted3+FcYxZ1CgLddR4xCSnJa0q5yaCUZWs9zuVyiiALLsoa/E41wpY0BWkya9ZIlVcvhOA4nT57U9aAWIzypkMG1a9fyUnTlJMtKEDHHcWhvb4fb7RZfOmgEQY156WynkaizEjVLs/YjT9lKjZ3j8TjGxsbKlrKtNPSOqMhVl+jMotTQmY6oHMZMsBl+lmqyU48//jgef/zxnJ8988wz4r8/9alP4VOf+pTitoSQ2wBu6zm/QyfLlZWVok4hgP7OSaX9CIIgDg/r1XbVElnu7+9jfHw8r55ns9mQTqc1H1sKKdlJTZrVWoXpIUsa7Z0+fRrb29u6/y6FGnx4nhczDHIhA+m2D4pFl9S1HsglCWldr6mpqaxNYHqh53od1HTD2N7aQigUvO81aUdbe/v971mTY+wcjUbR39+vmLKtlqhKC8yoIyq9XIRCIWxubmJ6eho1NTWicMRRMbXOZDKH1gynBod6lwUCAUSjUdOGykshGo0iFouht7dXU9pVDq2WYZcuXcqTcTJTCJ3W9VpbWzWZNGvpSFVqFjKSSlYivEQigXv37qGzs7NoN/JRiZj0HENKEtK6nlQ9p6mpqSpE4PUswplMBnNzMwiHw3DYHaitPYies9ks1tfXsLa2iv6+Y2iViPlrSdnqjcYrmeoth4KP1NCZECIaOqdSKbz++us5Gq3VTEjVjEMly6amJs0pUL2gaVeXy6WoKKMFpaLCbDaL8fFx2Gy2gi8CZsxqsiyLYDCIlZUVXXOaas+hkGydURUe6ba0FnP27FkxyjoMVMqiS+3n5CQhtUei3ZJNTU15gt6HSfiFwPM8ZmamkYjHUV+X2xXMcRy8tV7wPI+FxXmAAVpbld1viqVs9XbZVir6oscq54sOwzBwu91wuVzY2NjA9evX87qz6fxrtbjMHIW69NHKX+iAvNv1zp07hvdZjGTUNtcYHT+hBf9IJKJY11MDNVq5VLZO6qUp3d5IZElTRCsrK9ja2tL9Pd4pkKvD0G7JxcVF0eFBSiKVWPy1HGNvbxfRaCSPKKWw2Wzw1nqxvLyExkafqnJCoWhcS8q2kmTJ83xFjiX1CNYqyacHZhBeNXeGV93oSDFovaELdbsafTAKpWELeU8qwQjR0IYhQghOnjypm2BKnQOVrTt//rzifOAB4WUhCGkwjF3TNaVkOTo6Kkas1SBUfpQ6beXdknTAfW1tDbFYDIuLi2L2phzXVstzRAjB5uYGPCrE7m02GwghCAYCaLnv9KEWelO2lRQ3r9Q4R6GxslKSfB6PRyRPtRrIRp8Zo8bRlcCRiSwpQakt5pcyTTZTBF3PrKbe2VHaOTwwMID9/X3N20tRiCwJIZifn0cwGFSUrROENOKhUbDRl5AIB7G2YwPD2FHrG0Jt07tgryltHp1MJhGNRgsKrRvFzs4Odnd3TR/JqFYwDIP6+nrU19fj2LFjeP3111FbWyt6MJZDBF4LWcbjMaTTadR41ZVdapw12N3b1UyWcqhN2dbW1h6ZeUSzj6MkyUc1kFOpVI6tVqFI34wZy2q25wKOEFlyHIdsNlty0StFXGY4hkiJjqYpS0nkyaEnslxfX8fKyorYMBSJREzVdgUOotaRkZGC6kKZ1B52F/4b+EwEjJAFmBrYbA4QwiPq/ztE/a+hvv1ReFveW/Ba+P1+TE1Nwel0mk6UhBDMzc1hf38fXV1dCIVCWF5ezlk0i0mtHaXIshRaWlrQ1tamuABKbaQq8SLB84Imt0mWZZHNpE1PjyqlbIPBIObn5xEOhzEzM1P261KplC/P87pGVKQayIIg5EjyARDvHakkn9EAxCLLEtByw6iJxGjataenB11dXYr7N0MNiO5je3sb8/PzOHfunGYZMy1kKX0BGBoaEh9iM7RdpcRAZeuOHz+Otrb86JDPhLEz/3kI2eiBZJiQBpVbZBgbGJsbhPDY3/oeGIaDt+WncrYnhGB5eRnb29u4fv063nzzTd3nroRsNouRkRF4PB5cuXIF2Wy2oKGx1+tFU1MTfD51dbGjCHr/Ky2AtGZFNVvVisBLoWXR17poHzTBcGUlFmnKtrm5GbOzs2hubja1y7bYscsNM0ymi0nyzc7Owul0orGxER6Px3BkqUUB7DBw6JGlFuHqYiRHZfJK1QvNEkEPhULIZrOqZxr1nkc8Hhd1cuWRq5lC6DRtXWzeNbzzf8BnI7DZaN0p/4FnGBvAOhDaug134zXYuIPP8jwvyu+Voz4Zi8Vw7949HDt2DB0dHXnXRUm71e/3Y21tDcDbJs92u/2BiSwLQUoCQP6LBBX5LjVmoOU6uVwusBqyOslUEl1d3RWLwmgaUSllu76+Ltpr0bT+UfAzLUe6t5Ak39ramjhTrlWSD4B431UzDp0s1aKQ5B0dYieEqKoXGiXLRCKB0dFRMAyDK1eu6H6Q1RBdKV9IM8iSRq2pVKro9RP4FKKB1xU8ApVcAGwQhDRiwTdQ1/Keoh21ZoBepwsXLqgaRZJ2Bx47dizH5Hl/fx88z2NjY6NsQgDVluqVv0hEo1H4/f6cMQM62ylffLVIRLa1tWNzY6Nk8xvV/m1qajrUgXp5ytasVHal/vaVsOeiTWa1tbVYW1tDT0+PLkm+WCwGt4rmr8PEkSFLJck7NWlXpf3oJUu6KJ88eRJLS0umiKArQVp3K+YLaZQss9ksdnd3cezYsZJiBqnYIgByEDneBwOg0HPPMCziwbtIM2cxPT1dsKPWCKje7t7enuJ1UrvQSk2eE4mE6GdKhQBo1GmWbVKloOf+lMqqURH4UCiUk3ajUafWDtLW1jbs7e4gmUygpkY5MiOEYD8cRmdnJ5zOGqRSqYpc81L3ilIqW95la5ZMoVmotOMIx3GKknzya6Qkyac2Dfvd734Xzz77LHiex9NPP41bt27l/P4rX/kKPvvZzwIAamtr8bnPfQ6XLl0CADAMswQgAoAHkCWEXNfyHY8UWUpJTm3atdR+1ICSVygUwvXr12G32zE/P69pH3IUkvBLp9MYGRlBfX09rl27VvShUzMnWQhUtq62thbHjh0r+XlBSCIvimSKkCVsSMSD2A0tlMWvkxCCkZER2O32gv6ZesmC4zj09vaKowfBYBC7u7uYnZ0VZcSampqORCrOKDiOQ3NzM5qbmwEcZFYCgYD4PAiCgPb2dlUdxw6HA6dOn8H09BTCkX24XW5w3EEJgxCCZDKBdDqNjvYO9PT0ij+vZBpWLZRkCoPBoJiyLdZ9XCkirSRZKtVHlTqRaQYnEomgpqZGJFk1aVie5/HJT34SL7/8Mrq7uzE0NISbN2/i7Nmz4meOHTuGH/3oR2hsbMRLL72Ej3/843Jz6J8mhOzp+Y6HTpZaa5Za066F9qMWqVQKIyMjaGhoELtDCSGGpeqUUMikuRBsNhuSyaSmY0ibbM6fP4+lpSVV27GsA/k1SgZKaVgCIJlKAGwdrl+/bvoDm0gkEI/H0dfXVzCta5ZFl81myyGLQqk4M22lqhkulwtdXV3o6urC2NgYmpqaEIlEsLKyAgDii0Sh6MrlcuPcuQsIBPzY3NxAPJEAAwYEBI0NjWhrb4fXW2faTLRaGJ2zdDgcaGtrK9h9LHWWqRQqkYbVciz5NUokEvjrv/5r/Jf/8l+wvLyM9vZ2nDhxAo888oji+nfnzh0MDg5iYGAAAPDkk0/ixRdfzCHLn/iJnxD//fDDD4s9CWbg0MlSLTiOQywWw507dzSlXeXQQpY0937q1ClxsQTMfzOUumxcuXJFde5eaxqWyvBxHIehoSGk02nV2zvcfQAICBHAMAfkxxycfM7nBCIgkUjAzgG+zneZTpT0b+J0OstS/ywFt9sNt9uN7u7unK7SxcVFcBwnpij1mvUeNdTX16OjowPA25FDKdk5u92OtrZ2tLa2IZ1OgxABNhun2Ch3VETApSiWsl1eXkY8Hsfi4iJ8Ph/q6urK9v2q2fiZSvI99thjeOyxx/D8888jHA5jbm4OX/jCFxCPx/E7v/M7+Pt//++L26yvr+eMmnV3d8ujxhx88YtfxAc+8AHpjwiA7zMMQwB8gRDy39R/wyNElvv7+9jZ2cH169c1pV3lUCuCvri4iN3d3bJLsEl1ZAu5bBSCFiH0WCyGkZGRnCYbLWRr4zxw119CLPSWpBs2F9lsFqlUCs4aB1jw8PhuqPsiKiCXxbt7965p+5ZCS/ONUlep3+/H0tISYrEY6urq0NTUVHSY+yhDfp3kkYNcdk7aEGOz2cAwTMn0fCXJslzHkaZseZ7H3bt34Xa7sbGxgenp6bIIRgCVJ0sjx0omk7h48SI+/OEP41/9q3+FaDSa16Oi9FwW+pv95V/+Jb74xS/ib/7mb6Q//klCyAbDMK0AXmYYZooQ8ldqz/HQybLUDZrNZjE5OYlEIoHOzk5DRAkckGUqlSr4+3Q6jdHRUXg8nrJLsPE8j9dff113l6hafVnanCFvstEamda1vQ+J8DgEPgXW5gQYRkzCptJp8NksXK4aEJJCbfNPgbOb09AjCIKYepf+TcqxwBnZn9PpFOXnBEEQx1PoMDdNUVZTJ6xRFBN3UBKBp3qkVAS+VBRerTVLI8fhOE51ytaIMIIWxTOjEATBkJuJvMFHqX7Z3d0tPksAsLa2hs7OzrzPjYyM4Omnn8ZLL70kzlkDACFk4/7/7zAM8w0ANwAcHbIsBipK3tvbi66uLmxvbxveZ7HIcn9/H2NjYxgcHFQcyjcTW1tbSCQSeOihh3R3iZYiu1JdtVp9Qu3OZrQM/FPsLn4RPB8HITaAEMQTCbAMA6eTAUgatb4hNHQ8XnqHKkDtx9rb29Hb25szaF+uhdQs71QqPwccKCPReTS/349MJoO2tjY0NTUdWcskLde/kAj80tJSngi8NAp/ECJLKeSkXCpla8TU2QxRArUwWh9Vo+AzNDSE2dlZLC4uoqurCy+88AK++tWv5nxmZWUFP//zP48//dM/xcmTJ8WfMwzjAcASQiL3//0ogN/Wco5VS5Zra2tYWVkRu13p/JtRKJGl1KtRS81QzwMmCAJmZmZEV3AjkXIxsqQRstfrLdhVqyWNS+F096Lj5D9HNPAagpuvgM/GYXc4wNlYOGsHUdfyHjhrT5iy8NCGp9OnT+e8IdJzL0eEVq4F86BOdxBNTExMoKmpCclkUpxlpAviO2E8BSguAk8Iqbh7SrXotcq7bOlLFu0g1ZKyreaapRx0PSwGjuPw/PPP47HHHgPP83jqqadw7tw5fP7znwcAPPPMM/jt3/5t+P1+/Oqv/qq4zRtvvAEAbQC+cf9e4gB8lRDyXS3neOhkKX8Q6HwbgJxuVzOUd5T2k81mc5Rl1P7B9QiyJ5NJjIyMoLm5GadOncKdO3cM3dCFyJKKrReSraPQuwjZ7HVIMZewGnGg1s3g/NkzYG0eUa1HCwq9cKytrWF1dVU0mlY693KlMyvhZ+l2u9HW1oa+vr48CbFy1bDMhlnXSS4CT0Uitra2EAwGQQjB+vp6Wa9HtaZ7pS9ZWlO2RuuIWmCULNUq+Dz++ON4/PHcrNUzzzwj/vuP//iP8cd//Md52xFCFgBc0n2CqAKylIJqk/b29ubV8Aop+GiFtM5Hj9ff36+Y+y61Hy03CO3ilEZJWp1U5FAiS7nYutmgjiShUAjnzl3A2toa7M7SYy5KoGlg6SIlCAKmp6dFRaFC17eckWWla4pSCTHaUu/3+zEzM4N0Op3XGFMtKBfBSEUi6FwezciUSwS+WiLLYiiVsmUYJkfLtpJpWDNcR4z2o5QbVUOW8rSrHGZFlizLIpvNiscrpoVaaj9qUphUZWZnZyevs9ZMbVdBEDA1NYV0Op0jtm4mpELlV69eRTweN0QsNA1MH7J0Oo179+6hqamppKJQtUnGmQUadbrdbvT09IgqKH6/X2yMoeMpbrf7UMdTKnX9HQ4Henp6cgiCdh3TjuSmpiZNIvByVNJj0qzjKKVspUP/mUwGDocDdru97BmKSqRhDxuHTpY8z2NkZAQMwxQVGVCSu9MDKoLOsqwuUQPp+ZQi70wmg7GxMdTU1Ch21lJtVr2gZEmbYNra2nDmzJmCCwbP89jb28XS4gL29/dBCIE/4Edvbzfa2jqKjjfQZisqVC49vl5ICY+mjk+cOKFKkOFBiiyLQa6CkkwmReJMJBI54ymH4dlZbrKWR69yglASgafkqaVxinaplhvljGClETkhBMPDwwCA2dlZJJNJ07pslVCJBp/DxqGTZTQahc/nKzk6obVzUwnUmcJms+HChQuGHvRSYxs0xSslF637KAWWZZFKpfDmm2/izJkz4oKqhGg0ijfffB3xWAxOp0O8MXd2dzE6OoKpyUlcvXYdPl9T3rbUikzJSNtoZEkIwebmJhYXFzWljquN1CqFmpoaUUGHeg36/X6xc7KS4ymVqPOVOkYhEfjx8XHwPJ+Tsi1GUkdFKUgtGIYBy7Lo6uqC0+kUI/JgMJijtqSny1YJRsmS5/mqn0U+dLJsaGioyBvF1tYWFhYWcPbsWUxOThq+YYtFhVS3tlSK10hkRlV/EokEfuqnfqqocEI8Hsed1/4OhAh57iUcx6GhvgGpVAp37ryGhx96Fxruv7XT0ZNwOKxoRaanm1a+/dzcHJLJpOYovxRZEkKQTqc1a9JWioTNGk+Reg1SS6mVlRXE43FMTEyIC2I5xlOqjZALicDv7e1hbm4OTqdTvB7yFPZRqFkaOVahlO3m5qYpwghGvtdReek9dLIs91sWreWlUimxlmfGH0cpDSsIgmhNo2bx15uGpao/drsdbre7pMLQ7Mw0Mtks6ooU0Onb5+joPbz7p/6eWJ/0er24evVqwdETvdcyk8kgHA6LRs1a74Nix+Z5HuPj44hEIiCEiBZT1dIgU657XmopdefOHXR3dyMQCIjjKXQxrKurM1XarZwwsggXEoGXprBpWvIoyuqVQrFuWHnKll4becq2oaFBVcRnxveqdnnIQyfLciKRSGBkZKRkLU8P5CnURCKBe/fuoaOjI2d4Xss+1IDK1lGhhldffbXo55PJJDY3i3sIEhzovLpcLoRCIbGjdmBgAO3t7QW30xsZ0/qny+VCX1+fbncQJbKUihicOnUqpyFE2iBDXUPkx672B1YtpJ6dNMoKBALY2trCzMwMXC6X2CikV86xGtKwWiAVgacp7EAggNXVVSSTSSSTSdjt9rLqtVZynEMtgUmbyqjmMb02NGUrtR87SnPAZuLIkaXah4fOrJ09e7YsSv/SqJD6XGo9llayod9JrckxAAQC/qIPDcPc99m6f02TySTeeOMO3ve+R0t2p+khy52dHczNzeHChQuYn5/XHZkqkSVVYDp9+jR8Ph/S6XReg4zUYoq+QdMGmWqIOssFjuNyIol4PA6/35+j20o9O6vpOpSLkOUp7KmpKTgcDlGv1e12K4rAG0UlI0tA38uf/NrIU7bUpo6ms40inU5Xfb0SOGJkSWcti6U3BUHIqbOVS0qMdufOzc0hGAzq8mxUOw6j1gxaCZlMGsWeF5F0GAZ7e3uIJxI4feqUqjZuLWlYQggWFhbEa+VwOAw1CMmPvbGxgeXlZVGBqdB+5dGFkmsIz/MVS8sdBqTzelLd1r29PczPz8PhcOR4dh6mbmul/g4Mw4gvC2pE4PWikjVLs2qBhVK29IWzrq4OmUwGmUxGF+nFYjFTSLfcOHSy1PIgUIIqRJZUIaepqamkcbJR0PnJ1tZW0edSK9REZlrMoJVgs3EFDZqBg+ufzWaxs7uLGqcTrS0tcDrVvUmrPZdsNovR0VG4XC5cvXpVXCyM1DzpsQkhmJ2dRTQa1TxfKncNoWMZ6XQar7/+uhh1mjn8Xo2Q67YqRd9KIwcPEllKj1NKBN6IFVulRlTKBaWU7f7+Pvb29jAyMgIAonSj2tq4XES9WlEVfzW1i2axSIymlJR0RJWOZ+QNb39/H8vLy2hubs4R69WKUmQZDocxOjqKEydOoLW1VdcxqJB3oUWHEIL1jQ00NzfDW1uL0H4ILTqPpYR4PI579+6hr68vTyXJSDctwzBIp9N466234PV6dTUJyUHHMjY2NnDt2jWx1imNOpuamgyLAVR7xCqPvqUqMTSt/SCNpwDFIz75y4Tcik0qAl8q61PJyLISYFkWdXV1cLlcuHbtWo5U4czMTE7KtlCWwiLLMkBJ8o6m9/x+v2rvSUq6Wm9aOq6xvr6OY8eOGRZJKGYXRptsLl++XHK0ptiC4vXWobHRh3g8CpcrN9URjkSQSqXQ2dkJj9uNbDYLO2dHc7M++To56AuM3BqMwkhkyfM8JiYmMDg4WHCO1QiUht9pk1A8Hs+pdR7lSKEUCokALC8vI5FIiKLwcrcQs1CpGp8WUpZasRFCRCs2qSh+U1OTYmRVKbKsZBlBupaqSdnSLAW9X9QKEnz3u9/Fs88+C57n8fTTT+PWrVs5vyeE4Nlnn8Xt27fhdrvxP/7H/8DVq1cBAAzDvB/A7wOwAfhjQshzWr/nkXrK5So+1FmjtrYW169fV30TUrLU8nDTcQSq/BMIBBAKhbR+hRwoRZZ0/CSbzapKK6qxqjp16jR+/ONXwXF22O12EEKwt7eHdDoNt8sF+/2XkHAkjIsXLxtu8CCEYHl5WTTrLlTL1Vuz9Pv92Nvbw6lTp8pClEqQe1UWirbUpuWOymyZHFIRgNdeew3d3d3w+/1YW1sDgByiMGOxrobIshikXcf0BToYDGJ7ezsnsqL13weVLJXWjGJdtqurq1heXsarr76KgYGBkrOdPM/jk5/8JF5++WV0d3djaGgIN2/exNmzZ8XPvPTSS5idncXs7Cxee+01fOITn8Brr71GA6w/BPAzANYAvM4wzLcIIRNavmdVkKWeNCy1b9KTotSqM0vHNaQmzUal6ug+pGQpla1TO1JB91HsAWz0+XD12jUMv/UWotEYIpEw3B4POru6sLmxiWg0CpZlcOb0WfT09Br6TvSlwmazlXyB0ZqGpVZqW1tbaGtrOzR5rEJRJ03LUQk6n8/3wEedUqKQapNOTU3B4/GI10Fr8xvFUfOzVBLFl84vEkLAcRzq6urKem9UWvxAzQu2vMv22LFjiMVi+OY3v4mRkRGsr6/jZ37mZ/Doo49icHAw5+9x584dDA4OYmBgAADw5JNP4sUXX8whyxdffBEf/ehHwTAMHn74YYRCIWxubmJpaQkA5u47j4BhmBcAPAHg6JGlWtDIcmlpCVtbW5q8J+X7UUt029vb4qiDdFzDDGF36T6oK0kp2To51I5vtLV14OIl4Mc/fhXOGifsdg6RcBipVAKdXZ04ffoMGhvVH1cJyWQSw8PD6OzsRG9vadLVkoYVBAETExMghGBoaAjT09Mlv3el3qzlUSeVoFtZWTFN6PsoQJ6Cox2lExMTyGazYtSpxbOz2iPLYlCKrEZHRxGPxzE8PJzTYOb1ek39nkbl5ypxLJ/Ph3/yT/4Jampq8J73vAdPPvkkvv/97+PTn/40bt68iaeeekr87Pr6Onp6esT/7u7uxmuvvZazP6XPrK+vY319HQBWJR9dA/CQ1vM9UmQJAIuLi2hoaMCNGzd039xqiE4QBLHL8saNG3kpWzPIkkani4uLiq4kavehhizpaMWjj74fDocDyWQSAMHi4iJ6enoNz6IGg0FMTExomjVVS5apVAr37t1Da2urGHFXa5OE/O1ZWuOjUWcqlTLFFKCaIe8opdJzdFZYnp4shKMWWRYDy7Kw2+3o6emB1+sV5QnX1tYQiUREEXgjkTjFUTJ+pg0+g4ODGBwcFI2bpVBaJ+R/r0KfKbDGaK6DHBmyDIfDWF1dRVNTU07orQeliI4uzk1NTQWl3ow6blDs7e2hpaVF0ZVEDUqdB/UBTCQSOTVQSv52u8Nw/Yw2PWklezVkSTuCT506JcqWUZhx/csNudB3OBzGzMwM5ubmsLy8XPGoU+Dj4DMhACxsDh9YVv8cspb7Ri49JzcxpnOMcnGIoxxZKkHasCSVJ6Qi8PJIXI0IvBIqTZZGvSxLOQ11d3djdfXt4HBtbS2vu77QZ9LpNAD0SD8KYEPreVYFWRZ7GKhL+urqKvr6+sqm60pBIySlxVntPtQgFothfHwcDofDEPkXI0vqD+nz+XDq1CnTSV8QBCQSCQSDQQwNDWl+uyx1bCp+r9QRXK2RZTEwDIP6+nrU1dWhvb0dLpcrx17K6/WWrbM0FZtDZOf7SEWnwDAcAAJCBLgbrqO25X2w11SmUYpCnp5UEofw+XwVc+k4bFKWisD39fWB53kEg0FVIvBKqKSsnlGT6VgshmPHjhX9zNDQEGZnZ7G4uIiuri688MIL+OpXv5rzmZs3b+L555/Hk08+iddeew319fXo6OigRHyCYZhjANYBPAngI1rPsyrIshCy2SwmJibA3Pe69Pv92N/fN7xfJaKjHZzb29uqIiQj9lrU8urkyZM0n64bhQiHSr+dPHmy6FubXrKk0TfLsjh//rzuTkKll59SbifSzxVDtarw0HOSRxa01kk7S2nUabSetb/1LcR2/w8ISQMACHk7BRwPvoZE6A009HwU7oZrBr6VfsjFIaQNU8FgEJFIRBzJKJcs2mGTpRw2m62oCDx9sSp0TYwSmBaYlYYtBo7j8Pzzz+Oxxx4Dz/N46qmncO7cOXz+858HADzzzDN4/PHHcfv2bQwODsLtduO///f/Lm4L4FMAvoeD0ZE/IYSMaz3PqiXLaDSK0dHRnA5UjuNMqfXIyZIqzDidTtXpUD3dsFRtJhKJYGhoCIIgiELFeqFEdnRGU00DlB6ypER86tQpzM3NaT5nCiWypH8Lt9tdMAVOz/uojl8A+URPo046j0rrWaurq2LUSclTC2FEd/9PDlHmQwAhAoKr/xM2rhbO2lN6v5JpkDZMzczMwO12IxqNiik2s14i5KgmspRDLhRBZzul14Sq5hgVXdGKShk/P/7443j88cdzfvbMM8+I/2YYBn/4h3+ouC0h5DaA27pPElVClvKblJoBy82GzWiqofvJZDIA1Jk0K0EryUhl6ygJZDIZw3U3KWkLgoDp6ekcOzI122s5B7kG6/z8vKEZNSlpFFP7UcJRJstSkEeddHGkkmJqCEMQ0ghvf6sIUUpAMgitfw1tp/61mV/DFNTW1uZ4dgaDwbymmKamprLpQJsJM0iMZdmcFyvpyA71pnQ4HBWdszQyBhOPxy0FH60o5QdpJlkmk0lsbGxgaWmppEmzErTciDQSk8+EGknlUlCyo2nRlpYWnD59WvX5qe1IJYRgeno6r1HISM1TGh3S0Zlz587lGVQbPe9qTMVqgXzwPZPJ5HVR0hqflDASoTdxYL6mDnzaj3RiFQ5XT+kPVwjyv5/D4UBbWxva2tpymmKk6jk+n0/TeEolUY6ITz6yE4/Hsby8jP39/TyN43KkZiuRhq0GVA1ZxuNxjIyMFPWDVJK70wOGYbC9vQ23263KpNkI1tbWsLq6qpgSNaKNSsGyLKLRKGZmZko2JRXaXq2Ye2NjY16jkFExdJqK3tzc1NRNW+q4DMMYOrdqht1uzyMMv9+P0dFREELg8/nA8zwS+3dBBGU5RSUQkkUqMqGaLA+78UbeFEPVc+h4isvlytElrRaU87pRR5nGxkZ4vV50dnaKalNUBF6r2lQpGK2PRqNRiyzVYn9/H8PDwyWjCrncnR4kEgnMzc2B4zhcunSpbDcuHaLneR43btwoKAdlFNFoFJFIBNevX9cl0FCq9krT1IODg4pKSUZHaDY3N+FyuXD9+nVND1wpIoxGoxgfHxdVZMr1Vn3YkBJGf3+/GHWm02mEQ7twavrKAgQ+qfrT1SakrqSe4/f7MTMzg3Q6bZrNlhFU6uWNdsPK/Vzp3K8eEfhixzJyPePxeFFz+mpBVZBlbW2tKp9Go2nYvb09TE9Po7e3F/v7+2UjykQigZGREbS3txeMko1CEARMTU0hkUjg2LFjuv3gWJYV67dy0K7dYmlqvdFbOp3G8vIyXC4XLly4oPkaFTvu7u4uZmdncerUKbFRZmFhAQ6HQ3SPqKZIw0zQqHNlZQV1De1IRXbUb8zYwHKHIyFYCHrT6FL1nJ6eHvA8L7rILCwswG63i6nro+ClqBWCICiup/K530gkYjiNbQZZHoW/QVWQJcdxqhZcvR2Q1JkkEAjg+vXr4tu3GZA/zNRpQ4uSjVakUikMDw+boo+qdE3Vjm7Q7bVGlpFIBCMjI2hpadHdiKBElnIBd/ozqU+jPNKgUWel3esrEWF4fA8hHZvWkIpl4aq7WNZz0gqzas7yCIt6l9JRjFQqhd3d3bK7yFSqfq6mNiqthff39+elsaX2WsXIzOhM51Hx+KyKMyznDSR1Jrl27Zq4uJtR+6T7stlsohn07u6uLtk6taAC8tS3c21tzdB3kddNs9ksRkZG4PF4io5uSLfXsvDTaPXSpUuIRqOIxWK6z1t6XJr2BiAKuN9X7hDhcrnQ3d2N7u5u0dCXDn3T+lYpL9SjhJq6iwCjVoMVIGwLogkH6uzV47lYrgYt6l1KRzFee+01hMNhLC8vl1XPt9JpWC2QprGBt1WWSpmAG6lZHqWegqogy3KBdqEODg6ira1N/LmZXbWCIIAQgrGxMTidTk1WYVpBZeWuXr0qphGLpVHVQBoZUneV/v5+1WM0aiNLGt1TtR+73Y5YLGa4OQg4eCEaHh7O0Y4tBamhr7S+NTk5iVgshvn5+YKehEcFDGNDQ9cvIrj6JYAUv0dY1gGb7wlsbW1henoabrdbvD5GdUqNoBLdzAzDgOM4HD9+HMDbM65SZSWjdb1KwwxRArnKktSWTiokkc1mDR/rKHSsP5BkSQjB2toa1tbWFKXSzCJLlmURiUQwNTWF/v5+VbOBhc632M1CoyZBEPJk5Yw22NDtd3d3MTMzk+euUgpqIkue50XRh6tXr4rkY0RYgB6XNiCdOHGipL5ksX1J61t37txBfX296ElIicMMgetKw91wFURII7T+ZwAEgOTe9wzjABgWTcc+CafnONo6oegYQhfGSo9kVIIslcZT5DOu0rqeVABAy7Wo5BiT2SMqcls6uQj81NSUrmfEiiw1QssNVEqdgud50cqpUBeqWSLomUwG4+PjuHz5su5uLmkqVwnU47JQs5DR78IwDEKhkKgqpPXNudTxE4kEhoeHc5SYpMfWe+4MwyASiWBjY0PXnGypfVOpMUocfr8f4+Pj4mJppsFxueHxPQxn7QlE/X+FuP9vQISDjleWq0NtyyPwNL4rp7FHyTGEGhpLRzKqrRtWL4qtJ0p1vUAggK2tLczMzMDlcokkUar0UmlD5nK+1EhfKCKRCPr7++H3+/NE4Ovr64tGnel0+shE61VBllpAo0KlG4EqwNC6VLH5LCMghIhOHkaIEnj7+yjdUFTUvZjHpRGy5Hke8/PzyGazui3PikWWpWy79Da6EEKwu7uL/f19PPzww2V52OjCJiUOOssXCAREg+NCggDVBs7RhIaOn0NDx8+BCBmAYcEw6lJn8pGMeDyOvb09JJNJvP7667p8KtXiMCLLYuA4Lk8AgDb1ZTIZsWlMiSSq0ZDZDMifEWk/wPz8fF7nsfRaq5W6qwZUDVmqXTgpucg7NGkH1/nz50UZqHKAOnk0NjaipaXF8IOsRHaEEKyurmJjY6Nks5BeskwkEqINWTKZ1P0QFzo+TYNL66ty6CFLQRAwPj6OdDqNzs7OihOUfLGkggBjY2OiIIApuqXhMJhIBMTrBTSkxdWAYfWLkdOhd5fLhd3dXVy+fDnPp5LWOs1ocpNaWpULRuQaPR4PPB4Pent780jC4XDkeHZWmiwrdSz5fS7tBwDyO4+9Xq8465pIJDRnhQKBAD70oQ9haWkJ/f39+NrXvpb3Mr66uoqPfvSj2NraAsuymJiYeJYQ8vv3z/ffAfinAHbvf/xf3teOLYqqIUu1kNcbBUHA3Nyc7jSiFsidPKampkwxgJaSDU0jA1Ble6WHLKXSchzHYX5+XvuJ34dSV6pUn7bY+WslS9rI09bWhpaWFkSjUd3nXQz0vNR0AisJAtA6TjHLLcXvncmAe/FFOH7v98BOTgJ2O5DJQDhzBunf+A1kn3ji4GdVAHp9pD6VNNIKBAJ5kZbe8ZxKWHSZOZ4iJQnqFEK7SWtra5HJZJDNZss+KmF09tFMyDuPI5EIJicn8Zu/+ZvgeR4ulwuvvvqqajW15557Do888ghu3bqF5557Ds899xw++9nP5nyG4zj8p//0n3D16lVEIhHU1dV9kmGYlwkhE/c/8nuEkP+o5XscObKUOo+kUimMjIzA5/OpGnMwAiXZOjO0XaXkT6O9zs5O9PT0qPo+WsiSEIKVlRVsbW2JEWs8HjelQQg4qOHSqFuNPq2WBh86m0kl/XZ2dlRZdFUyHSWXoaPi52tra2AYpvhoSjAI9xNPgJ2dBUPHae7f57bRUdT82q9B+IM/QPzFF4Eyze9qQSFXehppUSEAqScjjTq1yM8dds3SCOROITs7OwiHwxgeHhbnPs2UnZOikmbWWkBF4B9++GG8+uqr+MEPfoDnn38ef/Inf4JnnnkGJ06cwMc//nE89thjBffx4osv4pVXXgEAfOxjH8N73/vePLKkwgsAaJlsEkAXgAnoRNWQpdY0rFqT5lLHU2OiOjk5CUEQ8hqG9Nh0yUHJhkZ7WsUM1J4D7aglhOTYkJnRIETTkSMjIzh+/HjOmE6pbdUce2dnB3Nzc7h06ZKYsil2v1CSpP9P/22z2cAwjKphbaPNK3Lxc+k4QiAQQCwWE7tMOUIOiHJiAkxa2SGEicXATkzA/cQTiP/wh1URYZZ6duSejLS+p0UUotpqlnrBsixqa2tRV1eHs2fPirJzy8vLZTH+rhRZGj0Ox3E4f/48fv/3fx+EEExNTeXNR8uxvb0tEmFHRwd2doqrVC0tLQHAFQCvSX78KYZhPgrgDQD/nBASLHmupT5QbWBZFpubm4hGo0XrYWr2U8pahkZ6hcTdzRhBYRgGGxsbCIfDusQM1ERnxTpqjfpCsiyLUCikaKlWCqVIiRCCxcVF+P1+XL9+PSfFXmhbKUFyHCcKRvA8LxJzNpsFy7Li/yoBafcgre1Fo1GsrKyg/ZVXcHJ6uiBRUjDpNNjZWXDf+hayv/ALFTnvQtBDMHL5OakohNPpPDQpwkoSC71mSrJzcuNvqT+lHlSi89Zoulcqos4wDM6cOQMAeN/73oetra28z3/mM5/RvP9fOHhW/hkhJHz/x58D8P8AIPf//z8BeKrUvo4UWWazWezu7qKmpka1SXMhUAeTQmSpRrbOaFRGI2Qj36dUKliu+COHkfENQogYJT300EOa68XFyFIQBIyNjYHjOFF5qdS2lBSlnax0waAPNFVvov9PX3ZYlhWjznLL0dF0pc/nw8DAAFy//uuwJRLqto3F4Pi936sKsjQCeX2P1jpnZmaQSqXEqPNBiSzpcZSecXkWQupPOTU1JZoBVOOcb7nsuX7wgx8U3KatrQ2bm5vo6OjA5uamosEDcFAW+oVf+AX84i/+It58882/oD8nhGzTfzMM80cAvqPmXKuGLEvdrDTNV1dXB5/PZ4qBqlJUSKOZvb09XL9+vejNKTWR1goatbpcLvT09BjqRi0U3a6vr2NlZUXRHky6vR6y5HkeY2NjyGaz6O3t1dVYVSiqpdq3NKJXgpzQlIiy0DHptaYRqDRVS/drxhyuKoTDsE1NadqEnZgAwmHTu2S1wkyCkarF0KgzEAiI42C0vlcOwe1KRZZqZx/l/pR0zlc6w1iuUR2tMDrPGY1GNY+O3Lx5E1/60pdw69YtfOlLX8ITTzyR9xlCCH75l38ZZ86cwW/+5m/m/I5hmA5CyOb9//w5AGNqjls1ZFkMm5ubYppvf3/fNKk6+X6y2SxGR0dRU1OjSraOmkhrBY1az507B7/fb1qDDYW0I7WQMEOx7UshmUxieHgYXV1dYBhG999DKaoNh8MYHR0tGAlLt6VkSQhBNptVVY+UQk6cdH6Tkm46nQbDMLDZbGVblJhI5KD+qMF6jnAchFAI7CGSZTmjMWnUGQqFcPLkSfj9frGrlFptNTY2mtK8ddiRZTEozfnKR3Wk4ymVhtEGung8rlpak+LWrVv44Ac/iC9+8Yvo7e3F17/+dQDAxsYGnn76ady+fRt/+7d/iz/90z/FhQsXcPnyZdy7d28Yb4+I/L8Mw1zGQRp2CcCvqDluVZMlXfSTyaTYVhyNRksWgNVATpY0cj127JjqP57Wblgqtk5dMZxOJ0KhkCHyl5MdnQNtampS1ZGqdZGgaV0qlLCxsaE7upZHh1RkXUmisNC2NKVaLJpUA5ZlsbKygp2dHVy7dg0cx+WkapXStUYgEr3XC2i8fkw2i+GFBdiCwUOr81VSjUYqgC8IAkKhEPx+PxYXF8WBd3oN9JzTUWmGAZAzqgMop68rpa4EmGPPpXXOsqmpCT/84Q/zft7Z2Ynbtw/GJd/97nfLr8Fl+g9CyD/Rc65VS5a0KaW1tTVn0ae1RqOQkuXW1hYWFhY0N6ho6YalaUu73W56NyoFjcroHKjZUErrGmkQotsqiayrAU2fGiVKQRAwMzODbDabp11LSZOmZunfm+d5kTT1RAsi6uognDkD2+io+vM9exZX3/veopZjDyqkAt5A/ixjfX09mpqaNEWdlSL+cpCyXOycvkjE43G89dZbYq2zHOMpgDkNPpaCj0ZI/5A0Takk82az2cQ5SyOg+5menkYsFtO0SEv3oYYsad1FSR+VZVlTvg9NVauJyrSCyvvF43EMDQ3lNEUZ1XcVBAGjo6PgOC6HqEqdj9PpRCwWwxtvvCG+aetRzaGWZI2NjTh16pTi9vScbDYb7Ha7SJo0sqX/VjuaIkf6N34DNb/2a2/PVxYB8XiQ/o3fAFDcciyZTGJ9fd00JZ2886hgZFkM8llGavC8uLgIjuPEqFMusybFUYosi4G+SNTV1SEcDuPs2bPw+/1YWlpCPB4Xx1MaGxtNGU8BjNcs9USWh4WqIUvgbRsnOiqg1FxjlmMIIQSzs7Po6OjAlStXdD34atKwe3t7mJ6eLijDZ7PZDKWVCSFIJpPY3NxUrYChBZlMBiMjI6ivr8fly5fzrpORztF0Oo1wOCyOtKgBJSiO40Qjb7/fL86r1dfXo7m5+WB+scS1SCQSoiWZ2tlQQLnWKR1N0Rp1Zp94AsIf/EHROUsAIA4HhBMnkL15M+93csuxO3fuQBAEUUnH7KaQanSLkDtjJJNJBAIBUWatkB/jUY4sCx3HZrPB6XSis7MTnZ2donKO3+/H6uoqAJgizViubthqRNWQJZUy83g8RZtrzCDLUCiE9fV1tLe3ix52elAsDau2q9ZIGpYSGQDdhF8MsVgM9+7dw8DAANrb2xU/o/f8acrY6XRqIkoahdP7w+Fw5Myr7e/vY3d3V4wsaNQpj7ZDoZAoAmFES1gadQLqRlPyYLcj/uKL+Qo+0u/u8UA4ceJAwUdFVMCyLHp6enKUdGhTCHXKMOpVWQ2RZTHU1NTkkAWNOpeWlsSo0+fzld2hg4LO/lbiOPLvQ5Vz6L2eTqcRDAZFacba2lqRPLV0thtt8InFYoaMKCqJqiHLeDyOrq6ukm/4Urk7PVhdXcXa2hr6+voMd9IVIu5sNqvaDFov+dOGpIGBAaRSKdMXLhoRl/K31BNZ0hrxpUuXMKqiVicVGihWn2QYBg0NDWLNLplMYm9vT2x+aGxsRHNzM1KpFFZXV3H58mXTG2NKRZ3ZbFZ5NKWxEfEf/hDct751oA07MfG2NuzZswfasDdv6lLukSrpSJ0yJiYmwPO8LsuxaknDqoU86kylUmK6NhwOo6amBjU1NXlRp5mophEVh8ORI80oNQQQBCHHYqvYvpRMLbTAqlnqQGNjo6pwXC+5UNk6ct/ncmdnBwmVg+CFoBRV0fpkb28vurq6dO2jFKj8G21IWlhY0LR9MRBCsLy8nNOxWwxatWmljTwcx6nWd9XTyFNTU5NT0wsGg5ifn0c0GkVjYyMCgQCam5t1R1fxRArReAqEENQ47Kivy50BVIo6qalyc3Oz2EVMf8/a7cj+wi8cCA4YcB0pdk3lThnUq1Kr5Vg1pmG1QJqiXF1dRSqVQjgcxvLyMliWFSNvMxtjKp2GVQu5IQC9J6SZCNpUJX+5NKMb1oosywQ9ZKkkUG5GOle+j93dXczMzGiyCdNDNoFAIE/+zeibPu1KHR8fB8MwquZM6fmrWThpN7DD4dDUyGNWxyvDMNja2kJ9fT2uX7+ORCKB3d1djI2NidFVS0uLqujKH4xgan4T61sBMGBA7p9rY70HZ050orvdp7iPSCQi+nt6vV7xu0nFFMR0bV0diIE5SrXXSu5VqcVy7ChFlqXg8XjEkTGq27q0tIRYLIa6ujrxBcJI1FkJBxV6HKPqZnL/Uul4Co06GxoaDKewE4nEocyH6kHVkKXam0jrqAJNJ547dy6npd4MsqREJyUxrTZhas+DCia4XK48+Te1ovCFwDAMkskkRkZGFPVjS21biuzpGBB9WVEDtYo8apBOpzEyMoK2tjbx+DS6kltrhcNheL1eNDc3o6mpKS/FtLi6izvDc6hxOtDsqwPLSoxs4yn8zZ0ZnBxox5VzfTl/o52dHbFbmS4O8nStVE3IyGiKXsgjjGKWY5WSoasE5CQm1W2VNsasrKyIHadNTU2ora3VdA2qKQ2rFkpOMlRdaWFhAel0WuxOL9ZxXAiVdAUyiqohS7NRqsHGLBF0QgiGh4cVSUwN1ESWpVK7dB96HxBBEPDmm2/izJkzRVVzlFDq5YV6gCqNARWCUiOPXkSjUYyNjWFwcLCgO43cWiscDmNvb09cHGm9L5bk8eO7c2j2eWG35z/gHrcTrhoHZha34Kqx4+yJgzGhlZUV7O7u4urVq4r1HVOahMqAYpZjlMxpc0g5iLMalHWkjTEDAwNIp9MicUajUTHqVDOOoUfBRw/KSUByTV869kU7jqXXo1QUftRS+Q8kWUqjsELpRDPIMhaLIRaL4dixY+js7NS1j1LnQWdOi6V2jXTUbm5uIh6P46GHHirayFMIxSJL2shTTJtWDrMUeYCDazc7O4vz58+rbk9nGEZcHI8fPy42gszPz+Nv35gH53DAXXMQgSndVyzLoLWpDpOzGzje24alpQVkMhlcuXJF9UJpdDSlHIuQXOybjuvIScNoqlKKahzpkHdfh8Nh1VHngzLPKUdHRwdcLhcEQRCvB639lorCzXjOK4WqIUuzLpha2TqjZEmbbFwul26iBAoTnZZGGz3fhc6Z0tlEvU0uhdw/5ufnsb+/r3r208z6JHBg1r25uYmrV6/qEnmnoI0gntoGTCzH4HVxCIfD2Nragp2z3ycQLxyS62ezschks/jhj/4Op4534eTJk7q/T6GokxIojcDlggjlXoA4joPH48GpU6cUScOMBplKRpZ6jiN9saJRJ/UsVfKoPIppWDXHEhvUWDanG13petBGIYfDoeulLhAI4EMf+hCWlpbQ39+Pr33ta4quUP39/fB6vbDZbOA4Dm+88QYAgGEYH4D/D0A/DnRhP0hUeFkCVUSWWkCjGfkNoUW2Ti9ZUiIIhUK4fv26+EfQCyWy5Hke4+PjsNlsqhpttEaWVLWmtrYWV65cwVtvvaU7MpUfm+d5UYz+6tWrqhYhM4mSqg2l02lcvXrVtHRUOJoAyzCo9XpRe//eoh2Uq2tr4LNZ1Hq9qKurg9PhwObmOuqO9xqa41WC2tGUSqa45KQhbZCJx+OaUnNSVDKyNOM4Us9SJY/KbDaLeDxeNuk5ikrWAYsRs9L1CAQCuHv3Lv7Fv/gXYn9HNptVfV8899xzeOSRR3Dr1i0899xzeO655/DZz35W8bN/+Zd/qVR6uQXgh4SQ5xiGuXX/v/9vNceuKrJUO7NHiU66aEjl2NTM/eghS0oyHo8np6PTyEMtFzagjh6dnZ2qh/W1kCWtf/b394uRt5E0rrRmSc+djmuohVmNPHS+1ev1GormFM9READk7s/pdIpdg4IgIBIOY3d3F+FwGDbWDl4QkE6nDUW2xVAo6tzc3ITD4RBHU8pR6yx2z8sbZPb390XypEIRaoTPq6FmqRfytHUmk8Hw8DC2t7exuLiYF2WZiUqJH9BjqSFm6fXo7+/HN7/5Tfyv//W/8MMf/hDXrl3DyZMn8dhjj+H9739/0bXjxRdfxCuvvAIA+NjHPob3vve9BcmyAJ4A8N77//4SgFdwFMlSLSjR2e12pFIpjIyMwOfzaVKx0UqWhdK7lCyMptkAIBgMiqMFhQynC+1DDdkVqn8aEUOnUb7eRh6GYTA1NYWWlhZDlku0m7enp0ez5Y8aOB3FX8BYlhWlC0+ePIlAKAaOZXDv3j0AB04JevVr1YJhGKytrSEUCuH8+fPii5i01mkWcaq956ViAMePH0cymcyz26Li7/K//VEmSznsdjvsdjtOnToFu90uRp1UlMMM6TkKnucrahKt53ybm5vx6KOP4q//+q/xzW9+E5OTk/je976Hv/iLv8Cv//qvF9xue3tbfL47Ojqws7NT8JweffRRMAyDX/mVX8HHP/5x+qs26mVJCNlkGEbZOVoBR5osqV3UqVOnCnY6FoKWPzC1jlJK79IFyejDtrq6ivX1dVy9elXz3FEpsiSEYGVlBVtbW4r1TyNi6CzLIpVKYWJiQlcjz7Vr13IEwGtqatDS0qJJLCAcDou2YeVy3GhurAXHschmeXBcPqEHAgH4/X4MDAzAbrfDFk3joevn4WuoFTso9ejXqgUhBFNTUyCE4NKlS0U7bKXuKdJ6k9bj6UFNTU2O8Dl1yVhYWIDD4cix2zpqaVg1x6EvK/KoUzqio1YYotRxqh2xWExMSZ89exZnz54FALzvfe/D1tZW3uc/85nPqN733/7t36KzsxM7Ozv4mZ/5GfzKr/zKewghf2XkfKuKLLWkYdfW1hAMBnWRi1oQQjA3N4dwOFwwvSuNcvVAEAQkEglR1UZPZFWMLAVBwMTEBAghOdZgarcvBlq/TaVSeNe73qWrkUfaDAIcPEB7e3uiWACNyAqJBWxvb2Npaaks0nVS2O0cThxrx9TcBlqb347KCSHY2tpCMpnE8ePHwbIswpEEmho8aKw/kPGSdlDSjkGqX2u328XRFLUvGnLQOjFdfAs5p0jLFnLS1BN1GiUY2i0ptduilmOpVAperxeZTKbsi/9hiwXIR3Tk0nNa5QiPGlnK8YMf/KDgNm1tbdjc3ERHRwc2NzfR2qocGNKmy9bWVvzcz/0cRkZGbgD4KwDbDMN03I8qOwAoh6YKqCqyVAMaUbpcLt3kogZUpNzr9RZtVDHSVZtKpXDv3j2wLIsLFy4YSuUqkR0Vp29tbUVfX1/B/eshS2kjj8vl0kSU0oVZDjoA3dfXJzqKrK6uIhKJoK6uDi0tLfD5fLDZbFhaWhJfmMyyHCqG0wOd2N4Lwx+IoMnnBREErKyugrNx6O/vP5g7jCWR5Xk8dPlMQdJS0q+dnp7O0a9tbGxUteBRs++uri7VXdlKTUJavTrLEfXJLce2t7cRCoXwxhtvoKamRnypMttyrFLzj2qumZIwhFSO0OPxiNehUNRpVIKuUtDjOHLz5k186Utfwq1bt/ClL30JTzzxhOJ+BUGA1+tFLBbD97//fQAYu//rbwH4GIDn7v//i2qPfaTIMh6PY2RkBC6XC729vWW7IaQi5YXcNij0RmW0xnfq1CnMzMwYWniUzoG6eqhJUWv9DvJGnldffbXkNlJFnkJEKYfdbs/pptvf38fe3h4WFxeRSqXgcrlw9uzZihAlADgcHN5z4xTuDM9jZWMPO9ubaGk+iHzD0QQSiTS8tS488hPnUOdVF+Uq6ddS2US32y1GnUopafo8nDhxQrOYBIU0XVvIq5N+Thp1lrvb1mazob6+HnV1dTh37pwo/l4Oy7FqHlGx2+1obW1Fa2trQcFzGnVKX4AqRf5GoIcsb926hQ9+8IP44he/iN7eXnz9618HAGxsbODpp5/G7du3sb29jZ/7uZ8DcND095GPfASvvvrqd+/v4jkAX2MY5pcBrAD4x2qPXVVkWexGksrWBQIB3TU2OeQ3MB0/uXjxYtmE3Tc2NrC8vCzW+GZmZgw3CUmvB/0Oao2gtTT40DqxliYkM6TrGObAUcTj8WB/f1+0l6IpO5quNcuvsRCcDjuune9FKrqDjuYTyAo2CERAS1Mdjve2osWnLFagBnJ3EHlKWqpfGw6HMTExgfPnz5sqRK1WEKESqUvpM+F2u+F2u8tiOVYpcjF6veRRZzabRSAQwObmJqanp8WoM5PJVGye00jAosdxpKmpCT/84Q/zft7Z2Ynbt28DAAYGBsTGOjkIIX4Aj2g+WVQZWSpBLh7udDqxv79vyKaLghIddb+YnZ1FJBJRPX4i3YcaEEIwPT2NRCIhum4AxrVdpRq1VAxAy3dQ2+BDSV5LnViqyGP0AY7FYhgdHcXx48fR0tICAOLiSRcN6pxBScfsqHN/fx8TExO4cf2yLsUjtWAYBrW1taitrc3TaQ0EAuB5HgMDA6anJKUoJogQiUTAMAzS6XSeIIJZKPRMlLIcU4q2iuGoRGJycByXE3XGYjH4/X7s7+9jfHxcfIFQex20wgwvy6Ni/AxUOVlmMhmMjo7C4/Hk6K6aIVVH9yMIAjKZDO7du4f6+nrVg/QUalOY9BgNDQ24fPlyzjGMaruyLCvOcbndbtO/A210oi8Sh6HIQ10Pzp07lxdJ2Wy2POeM3d1dDA8Pg2GYHANoI+dRqWYiJdAmkEwmg3g8joGBAYRCIbz11ls5+rXlHHinqdiFhQXE43Gxe1EqiEAjUzMWZzW1RKnQt9RyTB5tFavxVSoNW05IX66CwSBOnjyJWCyGra0tTE9Pw+12m2L4LYXRKYB4PK5pHvuwUVVkKb1hI5EIRkdHFeuGHMchnU4bPp7NZsP+/j5mZ2dx/PjxksbThfZRirjpdxkcHFTs3pJGuHqQzWaxsrKCkydP6pLeK0aWVGfX7XarnmM1myg3Njawvr6OK1eulHzQpakqKkG2t7eH+fl5xOPxnAYatW/FdPTG7/dXrJlI6Rzod6DKRE1NTaJ+rfQ7NjQ0iKMpZtb16XgKAFy6dCnn71oO8Xc9JCa3l5KmsmlnqXze9UEgSykEQYDD4YDb7c6x2aLRdzabNaXmazQNG4vFdHeAHwaqiiwpNjc3sbi4WLBuaLPZTEnDptNpTE9P4/Lly7rTATQ6LQQ6o1msBmpEQYdqL7a2turWqC10/EQigeHhYdVG1oC51lo0opUShFY4HA7R5FcQBASDQU0znYIgYHp6GoIg4PLly4fSkk/Hf+x2u2LXtNPpzJtd3N3dxfz8PJxOpxh1GomGqR9pbW0tBgYG8s6h0GiKEUEEo3XRYqlsqeVYJbVUKwH59ykUfcvNnbV2GptRszwqxs9AlZGlIAiYnJxEMpksKsBtNA0rCAJmZ2eRSCRw4cIFQ3lzuVwdhZoZzVL7KAUqZDAwMCBKm+mBElkeViMPBV2c3W43Ll68aMqbv9aZThpV19fXF5xfLDeoxGJTUxP6+vpKfl4+uxiPx7G3t4fJyUlkMhkxstISUdASQnt7u6q0mZw4gbdr13JBhGLnYHbEV8hyLJFI4O7du+K9odZyLJlMYmZmCpMT40gkErBxHLq7unH+wkU0N7fk7KPS0Wux66pk7izvNKbmzsX2Y0Ya1qpZ6sTe3h6cTidOnz5d9MYyQpbUCLihoQGtra2Gb2CbzZZHVHIN2VLHKBWdyiEIgnhjDw0NIRAIIJlM6jp/IL/BR08jj3QRNPqWTudPabRULhSb6fR4PIhEIujr6zu0ugqV8Ovr69NVIgAOukh7e3vR29sLnufh9/vFRiiPx4OWlpai9Txq3H3s2LGCA+DFQO8FedRJU7XF0rXlJBipis7e3h4uXLiQ45BRynJsemoSf/VXr0Aggih1yKdTmJmZxuzsDLq6uvEzj75fvK6VFArQ0kgkjzql40s0+1JovtVq8DlEtLW1qdIV1ZuGpbOHJ06cQGtrK2ZnZw03CtlsthyiisVi4uKiVqNUSxqWDqE3NTXhzJkz4gJjZJRG2k07OzuLWCymqZGHEIJEIgGn02l4cYtEIuL8qVqNWTMgnekMh8MYGRlBfX091tfXsbOzg+bmZrS0tFSssYeaVp86dUqTTnAx2Gy2vJm9vb29gvq1tPvYzHOgUSfHcSUFESoZjckdMopZjs1MT+GVV/4PODsHuy03Y2Sz2UAIwdraKv73//4W/tE/+lnxux6FVK+805iqKtGXc6mWrxk1S4ssywyO43TPNkpnD/WmP6WQEhUdJr9w4YKmsQK151GoUcgMssxmsxgeHobH48nr1i0EutD19vaKgtCUVPR0ZdI628WLFzXPX5kFahh95coV8RwSiQT29vYwNTVVkZnOYDCI6elpTabVWiFthDp27Fiefq3L5UIkEsHFixcLmo4bRSlBhFQqJf6sHKMphcAw+T6Vfr8fS0tLCIVCuDf8Jux2DjZWmSgYhoHD6cDO9hYmJsZw8eLlI0OWUjAMkzffKtVxBg4yF4lEQtdLpEWWFYCWNCxt0Egmk3nRkhkjKDTKXVxcxO7urujRpnUfpciOmk0rNQoZJctMJoP19XWcOnVKdZOQtOO1o6MDnZ2dip2n1E2kVG1qZWUFe3t7uHbt2qF0mwLA+vo6NjY28gyjXS4Xenp6KjLTubW1hZWVFVWdv2ZCql+7u7uL6elpNDc3Y2pqyhT9WjWQ1jr9fj/W19dx9uzZPEEEs0ZT1EJ6be4NvwXWxgJgkEgkABy8vIv11/vvhwzDwMZxuHdvGBcuXDqSZCkH7cCmNf/5+fn7ddsZUaqR1jrVRJyxWMxq8NELs+21aMrS5/Mp1kHNmtfc3d1Fc3OzKqNmJRQjOyrKQIXWlRZkI2QZDAYxPz8Pn8+niSiVGnmUOk+l0m2081RKRNJu0ytXrhzKgkLHMmKxWMmu23LOdC4vL4vjKZXyI5SDjuncuHFD/DsZ1a/Vit3dXSwsLODq1aviC4NaGb5yY3JyHHaOA3f/OSQCAc9nkU6nxRoeJU+bzYZkIoHd3d2yGz5TVDJ1TVO2bW1tYtTp9/vFLmxa8y30gpVMJssqqmE2qoos1UKNPBvVXj158qSo9iKHzWZDKpXSfR6JRAJTU1NwOBzigLYeFCJtKlbudDpzzKbl0EuW6+vr4nxmKBRStQ0hRKwXF1ugpHUeOu+2u7sr1seam5vR0NCAhYUF+Hw+UYi80uB5HuPj46ipqdHcdWvWTCchBDMzM8hms4c2nkIIwfLysihMLz1fI/q1WkHJWj7PWkqGL5vNlk1JSIpEIgFWcm0YlgHH2g/Ik0Ak8lQ6BZY5WKdCoSDcbndF/q6VjGClDT7yqJPWOmdnZ5FKpQr6lmo510AggA996ENYWlpCf38/vva1r+XV0qenp/GhD31I/O+FhQX89m//Nv7ZP/tnYBjm3wH4pwB27//6XxJCbqs9/pEky1KgJFDKX9FIZBkIBDA5OYmBgQFsb2/rPVUAymSXSCRw7949cZHSun0x0MU5Ho9jaGgI0WgUgUCg5DZ6hQak8260PraxsYHh4WFwHAePx4NAIFC2SKUQaGd0W1sbenp6DO+v1EwnredKSYWOyHg8Hpw8efJQXhikZC31wlRCIf3a0dFRceif6tfqiawDgUDJ6L6QDJ98NKUcUafNZkMmW2BMiwFsnA02zgYHHCACQTKZxObGBgKBIBiGQSAQKDmSYQSVJMtiDT5yB5n9/X3Rt/T3fu/3cOHCBc3lqueeew6PPPIIbt26heeeew7PPfccPvvZz+Z85tSpUxgeHhbPr6urSxRVv4/fI4T8R00Hvo+qIkujCwUdqUin06q6OfWQJSEEq6ur2NzcxLVr1wAciCgYgZzsgsEgJiYmVM84aiFLOtbi9XrFRh41cndmKvLE43Fsbm7i6tWr8Hq9qtK1ZoM6dkh1Zs2EmplOGll3dHQc2niKIAgYGxsTHVz0vgRJh/7p+I3X60VzczOampqK1nNpGjyRSJQkayUUE0SgaclCxKlVr7Wzswvz87OqanIEBDbOhocefhjxeAJra2s5L080TWlmh3UlBRbUdsPabLac2d9/+2//Lb797W9jbW0NV69exbvf/W584AMfwHvf+96iwc2LL76IV155BQDwsY99DO9973vzyFKKH/7whzh+/Liq+WQ1qCqyNAI6m9fS0iKOVJSCVrKkSiqEEFy/fh02mw3pdNqUJiGaDl5bW8Pq6qqmGUe1ZBmPx3Hv3j309/fnjLUUS2ubTZSbm5tYXV3FlStXxHpFqXSt3u7aQgiFQpicnMS5c+fKKoYuhXymc3NzEyMjI+A4Dvv7+3A4HAVn+soF+uLU3NyM3t5ew/uTD/2Hw2Hs7e2J4xdK9VxCCCYnJ8GyLM6fP2/4b6yUrpWSp7xJSGuN78LFy1hYmFO1XSadweCJk6ipcSEeT8DtdmNwcBAARCGA6elpUy3HjM4+aoFeYj516hQGBwfx4osv4sc//jH+5m/+Bi+99BK+/e1v43Of+1zB7ba3t8V1q6OjAzs7xX2bX3jhBXz4wx+W//hTDMN8FMAbAP45ISSo9ryrjiypA4eaz9GUg9QbspR3oxRayJIOZ7e3t6O3t1d8UMxoEqKjI5OTk0ilUrhx44amG14NWdJo9dy5c6LxMEUh1xGzpesWFhYQiUQKNrAopWv1dNcWw/b2tmiPdljNBYlEQqzL1dXV5fh0VqrzlL5c9vb2lvRs1QPp+IWSfi3tnNzc3CwooWcUSulauSACfXbVpi9bW1vR0dGFjY01OIrMFWezWbA2FleuXFXcf7ksxw6rZqkVicTBy4PT6cQjjzyCRx45cM163/veh62trbzPf+Yzn9G0/3Q6jW9961v4nd/5HemPPwfg/wFA7v//fwLwlNp9Vh1ZqgUlqY2NDTFS0bq4qCU6Kv12+vTpPJNdo2MbwAGRrK+vo7u7u6R6kRJKncPa2hrW1tZw7do1RYJQ2l5tI48a8DyPiYkJOByOPAHuYtDTXVsItIGF1sQOq9uUpuEuXbok3q8NDQ1oaGjA4OCgONNJO0/LMdNJU9AnT56smPCDXL+WDrrTyG9tbc2wfm0pyAURBEHA0tIS6urqROIsNZrCMAwee+wD+M53XsTu7g5sHJdzLwmCgHQqDY6z4QPv/wfw+ZrEnxfap5mWY9WYhlVCLBZTnKX+wQ9+UHCbtrY2bG5uoqOjA5ubm0UVpV566SVcvXo1R/mKECI2lzAM80cAvqPlnI8sWbIsi8nJSRBCNEdiFGrIslSzkNG34Wg0itnZWXi9Xhw/flzXPgqdA/XPpDOmha6RlCzNTrvS8R2jdTk13bWF0rW0lg3g0LpNgcJznFKUe6YzHA5jfHy8oiloObLZLJaWljA4OIiOjg5T9GvlyGR5xOJJCAKB2+VEjTP3ejEMg9nZWTAMg9OnT+dkUXieF18UlTpsHU4nbj7x8xgfH8W94beQSqVwcMsxEIiAEydO4srVayJRAuojvkKi59Rqq5TlWKXTsEbIUqsgwc2bN/GlL30Jt27dwpe+9CU88cQTBT/7Z3/2Z3kpWIZhOgghtMHk5wCMaTl+1ZGlmjRsKpVCOBxGV1eXoQ7CYmRJ5/9SqZRq6TetoFHS8ePHsb+/b+q+s9ms6NFZKpqTyt2ZSZRUsu3EiRN5EbkRFErXLiwsIBaL5aRrBUHA6OgoGhoaDm08RZ6CVrvAmD3TSVv5pVFtpUHLGQMDA2JjlVS/NpvN5rwgqNGvlWI/HMe9yWW8NjyPdIbHfQ7DhVO9uH7xGDpaGyAIAsbHx+FyuXD8+HHxXi82miKPOjmOw6VLV3Dx4mXs7GwjmUzAZuPQ3NyMmpr86FhvelTJcszv9+dYjtGoU1qaqgSMHCsajWq+B2/duoUPfvCD+OIXv4je3l58/etfB3AwbvT000/j9u2DKZB4PI6XX34ZX/jCF+S7+H8ZhrmMgzTsEoBf0XL8qiPLUqAp0bq6OrS3txta/AqRpVR/VU9atBQIIVhaWsLe3h6GhoaQTCZLjm5oAW3kOXbsmKp6FCVLM4mSphvLKdlGUShdS1922tvb0dXVdShESaNahmE0paDlMDrTSZWBikW15QbVmj19+nRe3ZyC47iS+rUtLS2KziDL63t44Vt/h2xWQFNjLRz2g+WN5wWMz6xheGIJj/zEWbi5hPjypIRioylKgghtbaWfMTNITPqSSBvFgsEgNjY2xMyD3W6vqGC73vtZT2TZ1NSEH/7wh3k/7+zsFIkSOHj58vv9eZ8jhPwT7Wf6No4UWVJLqqtXr2JpacmUxhp5FEvF1ouJGRgBHYLnOA7Xrl0Dy7Ki+ocZoPOf58+f16TpmUwm4ff7TZkBW11dxfb29qEszDRd63A4EAgEcPr0aTGaAcrTXVsI1OKrHFGtlplOKiV4mLVamv49f/68aomzQvq1S0tLiEajqK+vF02u94IxfOWbr8LrroHHndsUY7OxaG2qQzqdxp9980f4R49cweXL/arPvZQgghqvznKkR+12e96LxdLSEiKRCPb398V0tlrLsUriqOnCAlVIlkp/VDqyIQiCWHvTI6ZeCltbW1hYWMgRW1cLNW9ZdNHu6OjIadU3o0kIKN3IowSaej1z5gx2dnYwMzOD2tpaMfWlpTZGh9szmUxRxaFyQ9pEQ/+OpdK1Zp8r7Tbt6elR7T6jF4VmOkdHRxGPx8FxHM6ePVuxWpYcgUAAMzMzhtO/Uo1WQRByuoj/6o1lJOM8fPXKDUK8wGN7Zxs9XS0Ynt7Ce38yC4dD+/KnVhCB/l5KsuV8HuiLRVNTExobG9Ha2qpoOdbY2Gia9rIRAi7U4FPNqDqylKPYyIYemy4lUGuqaDRa1HS6EGidtdjNQ8dbzpw5k9eBaHT8hBAi6ncWa+SRgz7k0voXNcXd3d3FyspKTt2sWKei1Cj5sJRogIMXhq2tLcWo1szu2mKg6cZKdptK4fF4RNcQj8eDxsZGrK2tYXJyEnV1dWhpaanYTOfOzg4WFxdNF4ZnWRaNjY1obGxEcD+G23+7gro6F/Z2d5HN8nC5XPB43HC53OAFHpsbm2j0NaLWU4v17QDmV3ZwZlCdFnKp81ASRJBajtHRsEq8PNKmGy2WY4fxrEajUSuyNBN0NrAcBENBCMHdu3dRV1eHK1eu6Lpx6LkUehioPVihjlojkWUmk8HIyAgYhsHFixdVPZDFGnmkprjHjx9HMpnE7u6u2KlIa0ZSObNEIoHR0dGyzeypASEEc3NzSCQSuHLlSskXBiPdtcVABQ+0pBvNBs/zGBkZQWNjo1iXo4tmJWc6NzY2xO7fcjrJbGwHwXF2+BoagYZGCERAIpFANBbD7u4esnwWdXV1YrbF5XRgen7DFLKUopAgAs/zYnSXzWbL6poiCELey14xy7FYLIb6+nox6lT7AmU0ExaPxy2yNAoapa2urmJjY6NgStEMsoxGo4jFYhgYGDCUKqPnIl8Q5BqshW5Evd8lHo9jeHgYAwMDyGQyqsQcpA8wrbMUQ01NjTjKkM1m4ff7sba2hnA4jPr6erjdbmxsbODs2bMFmzbKDVoHdrlcuHDhguYXHi3dtcUWOWkUdViCB7Q5raurK89FhmGYis100plWNS8uRpHJ8DhocDwAy7DwuD2wc3YkEgm0tLSAz/LY2toCEQgE2BAMucrq0EGvH8uymJ2dhcfjgc/nK7tripp0rzydTaPO5eXlHEF0t9td8PoYrcHGYrGy9ISUE1VHllRYGkDRlCLHcYYcQ6hahsfjKTrcqgZKfpQ04qurqytppqxWtUgKOtRNjaZXVlZK3sDSWTI1RCkHx3E5cmaLi4tYXl6G3W7H4uKimK6tpA+jWXOcUuhJ166urmJnZ6fsUVQxUPH9wcFBVUpW5ZjpNKrzqgdKtcdUKoWt7S20t7WL92NjYyN4gcf65h6SiSh+/OMfq9av1QPaa1FTUyOOqNCflxpN0Qut6V6WZcUXKODgulHB80QikRN1StcWIzOWwAFZHtb4kl5UHVnu7Oygvr4ePT09RRdzvdEYnXkLBAIYGhrCvXv3DP/haU2CIhaLibNkatKSWklLKepWI4ZupnQddY3/yZ/8SXAcl+c+QdOY5ezEo7VBteSgB6XStU1NTUgkEuB5/tD8OAEgEolgbGwMZ8+e1dQFTWHGTKfZOq9q0dvZBJZhwAsCbCyLZDKJ7Z1tdLR35KUkbawNHOfAoz/9EAZ6W1Xp1+oBFainUn5SFGoSkkadhBBdlmNGIz6n05nzokjdQmjaXtpIZpQsj5LxM1CFZNnZ2YlMpoAFjgR6yDKbzWJsbAxOp1Mc2zAjnSvdB01r0YjPTFChhHQ6LQq5UxQjS2kjj9HFnL4tcxyXo4YjFwqndbFydZ0Gg0FMTU1VtDYoT9cmk0mMjIwgk8mAZVnMzMyUrbu2GGi36cWLF03pMNQz00nJwePxlEXntRhqPTW4cLoH4zPr8Hrs2NvdQ2dnJ+xcfqQYS6RQ66lBf3eLKv1aNX6kcgiCINaM1ThelBpNkdY5Kyl3J22iAg4yF4FAAHNzc4jH4wAgjptpJU5rdKSC0EpydFC/t7cXXV1duvejBBpZLi0tYWdnB9evXzc9FZnJZHDv3j00NjYqCiUU0nc1W7puZGQEra2tRV0q7HZ7Tk1EmsakiizF0ns8zyOZTAI4SBXKH34l55JKI5PJYHx8HB0dHejp6Slrd20xSIXhy5X+LjXT6fP5sLOzU/K+KCfe89BpjEwuYWl5BydPHANny1/akqk0QvsxfORnfwI2Wz6hyPVrQ6EQdnd3MTc3B6fTKUadxbrCaXNVc3OzLo9UrYII8mejnCMqLpdLvD7BYBArKysIBAKYn5+H0+kUo041+r7W6EgFwXGc6tERWt9TGtQ3iywXFhbgcrlw/fp1029WmtY9fvx4jjCw/BykZGk2UdKUp1b/R3kak6b33nrrLbAsK6b+3G43gsEA7r75BkZH70EQDmq4NhuLixcv48rV66ivrxfTv4c5YE/HmY4dOybWu9V015o9IE7rpFeuXKlYnVQ+0xkKhTA6OgqbzYatrS1kMhk0NzfrMn82gmQ8gofPt2J8yY3tvQjqPDWo9dSAYRgkU2kE9mOw2Wz44D98CMf7lJ8hKViWzfFgVKNfSyUmqWKUGdAqiFApbVhCCDwej27LMSsNawLUPmBqSI46TWxvbxeM9pSac7QglUpha2sLPp8P586dM32BkDfyFIJcDN2s+iTwdprv3Llzhm5weXovlUqJsnQry0uYm5vOO+dsNou7d9/A8PBdnL9wGd3dPRVrHFECrQ2eOXOmYPdvoe5as9LStO4ejUZx+fLlQxMbSCaTmJ6expkzZ9Dc3IxMJgO/3y+aP1dqpnNtbQ3b29t473vehfe8G5hb2sKrd+ewuRMCAVBX68L73n0e5052w+vRl4kopV/b2NiIzc3NsopQSKNOKVlKZzvNmj0vBXmfh1bLMa1p2K9//ev4d//u32FychJ37tzB9evXFT/33e9+F88++yx4nsfTTz+NW7duAQAYhvEB+P8A9ONAF/aDWrwsgSokS7UoRZZ0nIBlWQwNDRW1x9F7g1GhAep4bpSU5K3sKysr2NzcVJXWlYqhm0mU1CmjHGk+p9OJ7u5upFJJzM/PQiAEIAAhAgAGDMuAOZDCPuguvncXg4ODh0aUemuDZoohUK1ZlmVx8eLFQxN/UNJ5tdvtOYPwlZjppCMq0peGsye7cfZk9/3nAYopVyOQ69cGg0FxrVlbWxPHVcrZ3CYdTQEO7oudnR0x0qR9H2aPplAUa4osZjn2+7//+6ivr0c6ndZkx3b+/Hn8xV/8BX7lVwprn/M8j09+8pN4+eWX0d3djaGhIdy8eRNnz54FgFsAfkgIeY5hmFv3//v/1vKdH0iyLCQrV2g/eiLLzc1NcaZub2/PlFQuTaHQBTGbzeY18hTbPpPJIJvNmvJwUFWjZDKpySlDz3G+/73bEAQeNsk5E0IOUsnkYMFjmIPr8/L3v4t/+vFPVJwkpHVSIy8NRtK1PM9jdHQUdXV1OHbs2KERpRqd13LPdNLoOhaLFcw0VOKlKpPJYG5uDqdPn0ZLS0tR/dpyRte0o5c+q9J5asC80RQKvZZj//k//2fcvn0bP/rRj/Dud78bp0+fxuOPP473v//9BUtMAHDmzJmSx7pz5w4GBwfF7uMnn3wSL774IiXLJwC89/5HvwTgFbxTyFJJBB14W/Xn7NmzYhdXMWitWVKlmHA4LErj2Ww2VR28as5DEAQMDw+jqalJ9YJI5yb39vbgdrsNG+jSWVePx6NryF8LlpeX7vsB5h6D/jf9bgCBIBCEw/u4c+c1nDlzFl6vt+yEIR2TMbtOqiVdy/O8WA8za55UD/TqvJo500nFPnieL/v9WQypVArDw8MYHBwUa7jF9Gup3ZbZ0fXOzo7Y5CW9fjabDXa73dTRFAqe53U9Cw0NDfjIRz6C//pf/yveeustTExM4Pbt2/j0pz+N//k//6fm/Umxvr6e01TV3d2N1157jf5nG/WyJIRsMgyjebi+6sjSyI0vdSVRSxgsy6pOw2azWYyMjKC2thZXr17N0ak1IpBAzyMajWJqagqDg4OqhRLoG2R3dze2trZEd/Wmpia0trZqTgXRcYju7u48BZhyYHbmoCFAHrnS7/W2eAIjdh3v7mzD7fYgGo2ioaFBrIuZHUkQQjA1NQVCSEXqpIW6Tmk01tHRYVhAwwjM0nk1MtNJZzk5jsOZM2cOjSiTySSGh4eL6v8qjV5I/57S0RS999b29jZWVlZw+fLlgi8aapqEtEadPM/rvgekJaLz58/j/PnzAID3ve992Nrayvv8Zz7zmaJGz9L9ymHm/VF1ZAloV7SRpi21CIkDB/WHdDpd8nN09KSvry+PROSiBHpARxEuX76suolGWp90OBzo6+szNOdI02unT59WFZWbgUQikfcz+kCzrA1K97rNZsOFCxfy2vtdLpdp4xqHnfKk6Vqn0wm/369oNVZJ+6Vy6bxqmelkGEYc9D/MNDRVSirmy6kEeXQtr13Tv6laEtra2hJLA2qjPLMEEcyY55T//X7wgx8Y2l93dzdWV1fF/15bW5Ou1dsMw3Tcjyo7AOxo3X9VkqUWUCuklpYWXZ6BatKwxUZP1O6jGJaXlxGPx3Hp0iVdRClv5Ck250jtt5qbm3MeMBo1GLVR0gp5BoA2ZbA2Gwr9JWvubyNt71eq/9HoRes8VzF91UqCirJfvHhR7BwsR3dtKVRS57VQdD07O4t0Oo3GxsZDM/IGDhqbRkZGcO7cOUOiI/ImGLkCFq3pFhrB2drawtramiaiVIJeQQQjIyo0kjUbQ0NDmJ2dxeLiIrq6uvDCCy/gq1/9Kv31twB8DMBz9///Ra37P9JkKQgC3njjDZw6dUq33FmpqJB2pBbziNTrGiIIAiYnJ8HzPFpaWlTf9IQQ1Y088oYSqf0Wx3Fobm5GKpVCOBw+FF3TEydOYnx8FNksD0IOZkILdS8SQsBxdpw4cTLvd0r1v93dXbFJyefzoaWlpWRDSTwex8jISFkl9NRgd3cXCwsLisILhQjFbDEEWp9PJpOHMq5D7936+npEo1G0tbXBZrNhbGxMLDVUcqYzGo1idHTUdNUo6b3b39+PTCaDQCAgjuB4vV6x1GC327G5uYn19XVcvnzZ1Bq6FkGEbDarmyzj8bjmF9hvfOMb+LVf+zXs7u7iH/yDf4DLly/je9/7HjY2NvD000/j9u3b4DgOzz//PB577DHwPI+nnnoK586do7t4DsDXGIb5ZQArAP6x1vNmSqQ7tal7m4RMJlOSfDY2NjA+Po4bN27o0sKkCAaD2Nrayuu2khLZuXPnit4Y4XAYy8vLuHDhgurj0uilubkZ/f39mJycREdHR9H0p9lCA/F4HGNjY0gmk6JCiZ46pxEQQvBf/+sfIBaNlqyZCIKAuro6/NOP/6qm86MNJbu7u9jf3xcXn6amppzFZn9/HxMTE4ajBqNYX1/H5uYmLl26pFnEnEbXe3t7APSna6U6r6dOnTq0SC6TyWB4eBg9PT05Ost0pnNvb68iM520RGGWpKBaUC/Kvb09+P1+ZLNZCIKAixcvVqTBjULu1Uk1mel9peVFanNzE7/2a7+G733ve2U84zwYvlBHLrIkhGB6ehqJRAINDQ2GZ/+UIst0Oo3h4WHVqV2tadhoNCpGL7Rho9Q+zCbKTCaDyclJtLa2oq+vD9lstqKpPYp4PI7+/gFMTU6g2LsZTfs8+ugHdKXapQ0ldPFZWloS5/9sNhtWV1dx+fJlw93EekE7b/f393WlPM0SQxAEAaOjo/B6vYdaG6TdpgMDA3mqUZWc6dzf38fk5GTFSxRArhdlTU0NNjY20N7ejsXFRUP6tVpBX2RptqG+vh61tbXiukQzXTabreR6cRSl7oAjRpZUH7WhoQGXL1/GyMiIYcUKOUlFIhGMjIzg5MmTqmXdtMxq0m44+mZIUSyVW46IcmRkBAMDAyJZq6lzFrIxIoRgxx9GKpWBw86hpalO1SA4HUP4e3/vp9Hf34/vf+8l8LwAhskdHQEOrvEHPvAP0X9soNguS0Iunp1IJDA7Owu/3w+Xy4X19fU8c+tKgL4E0qjBjBcUPela2vHd0tKiS9vULNAmmmLdphTlnOkMBoOYnp7GpUuXDu0lCjjINmxvb4tzlFI9YqrTS7NDLS0tZdFMpvcox3E4ceJEjuWYNFVbSr82Go0eORF14AiRZSQSEbVJ6fCq2Y4h29vbmJ+fx+XLlzW9+ajphqXSezs7OxgaGsq3DirwXcxW5KFuHcVSjaXqnDRKY1kOb4wu4OW/HsNuIHJwboTA66nBIz91Hu+6Mgi3Sznylw/5nzt3AZ0dXXjz7hsYk2jDcpztvjbsNTQ0mNuhSwjB+vo6CCF4z3veA0JIjlxbfX29mNor51s7dexwu905vodmQo0YQn19Pebm5soq2aYGtIlGr92YWTOdfr8fs7OzuHz58qEJ9gMHXZ07Ozu4dOlSntOQVKeX6tdOTEwU1K/VCzrbyjBMDlHS82BZFhzHiU1CxQQRotHokYwsq7Jmmc1mc4iDkpi0KxAAJicn0dbWVvLNsxhoytXn82F/fx8XL17U3OTC8zxef/11PPzww4q/p7ZWhBCcO3dO8calKUGpADNt5AHMUSPZ2NjA2toaLl68qPvhTyQSB8S5toFv/mAcu8Ekmn11qK/zgGEOzjGRTCMUjqPF58Wv/1+Poqnx7QiaKq9EIhFcuHBBkYR4nkcikQDDADU1rrIQFf2b2O12nDx5Mo+g6ED57u4uAoEAampqxEjMTNk/Kr7d2tp6aJFcOp3G5uYmFhYWcgbnyzG7WgpUe/fChQumRx/SmU6/3190plPaYFVO15hSWF1dxd7eHi5evKjpOaD6tXt7e9jf3xdfEpqamjR/H0qUhBDN9WvpaArNnH3nO9/Bq6++is9//vOazsMgHuyapVQtZ2hoKI/EtDiPFANtEJAKDWhBsRSqvJGn0P7lQuhmpl2pe30sFsO1a9cMkc+BTU83/tf3x5FIs+hsbUAymcTOTgxOpxOuGhdcTidcrQ3wByP4L//jZfz/PvEP4a5x5HhhXrp0qeD3stlsZU3T0FRjU1NTQb9B+UA5jcRGR0dBCBHTXUZMgunYU19fX1Gpr3Ijk8mI+r91dXU54xqVshoDDkZlpqamylYbVDvTmc1msbKyUjVEqacTWa5fS18SpKblavRrqeylHqIE8kdTotEo/uAP/gCPPPKIpv1UA6qWLOmC5vF4CpKY0TRsIpHA8PAwbDYbTp8+rXs/hW4g2shz4sSJkvVPKplnNlFSQXmXy2Wa8Pb0wibmlrbR1tIAhmHg9ngAQpBKpZBIJhDaD8Fut8Nd48L2Xgh3Rxfx0OUBsRZ2WJ6HwNsE1dvbm9NdWQpU37K/vz9vkaVjKQ0NDaoXNSpErqYmV04o6bzK07V7e3tlF0M4jJSnUk13eXkZoVAIjY2N2NnZQUtLS9l8QouBekWaMbKj9JLg9/vFxq9C+rU0WOF5XtFDVytSqRR+8Rd/EZ/4xCfwy7/8y4b2dRioSrKMx+N488030d/f//9v77zDoyrz9n/PpPc+k0YKoYVAGrAiqwgiIC0JCIgu4iXr4q4NXt3XxfVVUfen7oqru2LZXV1EV0RI6AlFUVSKQTCVhJDeMyW9TT3n9wf7HGcmM5kzM2dKwvlcl9clyZyTc5KZ832eb7nvUYfCbQmWREM2JSUFFRUV1l6qSciMn2Hq2BSk7slloFQqlSgtLUVUVBSneqLfXKiAp6e7/vUJBPDy9oaXtzcAGmqVGsOKYWhUw9h78BtohyVImjjRqbqmAwMDKC8vx9SpU21SKDL2kJVKpaiqqmJVEyMjKlzP61mKOZ1Xw/k/Y921XKRrpVIpGhoakJmZ6bSdnFAohFKpZOrXSqUScrncKTOdJGBz1ehlyGj6taSTOCwsDO3t7VCr1ZzICioUCvzqV7/CunXrsHnzZo7uxLG4ZLAcHh7GjBkzzM66WWuvRTRkRxMasBbdRp7Zs2ez/vALhUIMDw8zw762vjlJ7Wfq1Kmc7lxomkZFTStEYaM1Xgjg4ekJD09PeHv7oLahFb5+wZBKpWhra2NSQI6cEyNdjTNmzOA0xTuauTUZWdEdYdDdQTmzu5IEKEt0XkdT17E2XUsG7B1pYG0M0kRDrL7c3d3h5+fHyEc6yqeTjA7NnDnTIfViU/q1xcXFUKvViIqKQnd3t0VZE0OUSiU2bdqEZcuW4ZFHHnHaKJKtuGSwJHUDc7i7u1skYE40ZNVqtcUasmzPT3aps2fPZv3momkaAQEBkEql+PHHHxEQEACRSISwsDCrrlEmkzENUVx3nd1Q2gGEQvNv+OHhYfT39yE4OBhx8fEQhwcxD57GxkYMDAw4ZJ5TIpGgsbHR7ik+w3SXQqHQG2Hw8vJiLM+ckdojENEDWwKUse5aS9O1JEA5QkZvNBobG9Hd3T2i25TgqJnO+vp6pvHNWZ6tPj4+UKlUCA4OxtSpU9HT08NkTazRr1Wr1di8eTPuuOMOPPnkk2M2UAIu2g1LURQryyuZTIbu7m5MmTJS/swQ0mhjzPrqwoULmDdvnk3XfO7cOXh6ejJD/mzfFCT1StKuZGheKpWis7OT6cKMiIhgtWJvamqCVCpFamqqXVJaNE3jf176D4ICfeHubvoBNzDQD8WwAiGhoZB19eO1P9yLQP+ROrBknrO7uxt+fn7MIoGrXUZjYyM6OzuRmppqVz9Bc9TX10MikcDP74ZbCtmdWLsgspaGhgZ0d3db3F1pCSRdK5fLTaZrie2ZqY5oR0EC1IwZM6wKUGQnJpfLbZrprKurw8DAgNXXwRV1dXUYGhpCSkqK3jNMd0Ekl8tZ6ddqNBr8+te/Rnp6Ov74xz86O1Da/MPHdLDs6uqCRCIxawxKZjRNWV9dvHgRt9xyi9Vv0v7+fly8eBHp6ekWW2uZq0+SLkyZTAaBQMAETsNVLEVRuH79OjQaDaZPn27XD9y+oxdx/ko1RGHG0uQ0enp6QFE0QkNC0NM/hAlRodj262WjnpOkMMkiQVd1x5p0JWl3V6vVdv99mLsOIgJOroPsTsgIg5eXF5PCtNfOlzRrKJVKh/4+dNO1XV1d8PHxYTxKnbmDIiNMw8PDnP0+DGUV2c50kkYxU2NljqK+vp4J2OYCm6HcIJGQDAgIgK+vL7RaLX77299i0qRJ2LFjh7MDJXCzB8u+vj40NTUxfmjGkEqlqKmpGbXR5tKlS1anpMj5BQIBZs+ezdq81ppGHqVSyQROpVLJaLn6+PigrKwMISEhVjmvWEpLRxdee/coIkL1lXpomkZXVyc8PDwQGBgImgbapT149IG7MHOaZTOECoWCuVeNRmNRnZN0ANtzyJ8NJC3v6ek5YpBbl6GhIUbPVavVMmMpXHWcEp1XNzc3ozOljoKiKFy9evW/M7Q3rsHRVmPAzwsYjUZjN09MNjOdugHbcCfnaBoaGtDf329VwCbZMJlMhieeeIIR9EhKSsI///lPpy4AdBifwZKmaVYek4ODg0yzhLFz1NfXo7OzE2lpaaOmJK9cuYKUlBSLVvVEx5PMQZWVlbE6B1eKPBqNBp2dnWhvb0dnZydCQkIQHx9vdy1X4MY9HDp1Gae/L0NkeDDc3ISgtFrIOzvh7+8HX18/0DSNdmkPMlLi8et7F7CSvzMFWcXKZDKzdU4iiRgZGenUzls2s5zGGBoaRtm1WkgkUqjVKkxJnIDo6Eir/66uovNKjAk8PT0xadIkCAQCVulariGSbQAcKhBP7lUmk2FoaAjBwcFQq9UQCoVOD5SNjY3o7e3lJAVMURSeeOIJSKVSeHl5oa6uDrfddhtWr17t7NnK8S1KYA5ToyNarRbl5eXw9PTErFmzzL4BLB1BIStkoVDInJ+NTReXijzu7u7w8vLC8PAwMjMzQVGUnpYrqf3Zo04nEAiQs+TGfX/5XRkomgKtUSA0NATu7h6QdfZCS9GYkzYRG3N+aVOgBPQbLAx1a/38/JgUJglQxoS3HQmpj8fGxrKWjRsYUuDc5UqcuVCKIYUSAtzYeXx9uQFpU6IwMToIovBQps7Jph7tCupAwM8BmxhpE+zRXTsaZIdtqG3qCHTvlWQ+BgYGIBAIGNESZ8x0kjEVLlLiFEVh+/bt8PPzw7FjxyAUCqFSqfD999+joaGBmwt2ImN6Z6nRaPDTTz/hF7/4BfM1hUKB4uJixMbGst5ZlJWVIT4+npUtE5HHE4vFiIuLYz5wpaWlSExMNDk3Z9jIYysdHR1obGxEamqqXk2PaLmS2p+npydT+7PHB/Ha9QYc+7IQ7V0qqNQUvDzdkZmSgNt/MRUTosPs+kDSrXNKpVIMDw8jNjYWEyZMcNpYBhEAnzx5MqPZaY7Onn78fU8+ZF29EIcFwdvr5+CgUKoh6exFaJA/Hl63ELTmRoctSeuZMrcmAdvQ2srRaLVaxpydbcDWbSaRyWQAbE/XkpS4j48PJk6c6LSdHKkd684v6jbOOHKms6mpCd3d3ZwFyhdeeAEDAwP44IMPXCX1qsv4TMMCYDUSQtM0Ll68yHSy9vT04OrVq5g+fbpFQ+cVFRVmvSSBnx1JjJlNX716FbGxsSOEn+0hXVdXV4e+vj7MnDnT7M6R1MNkMhlommYCJxcjJW1tbWhtbUVqaqpTRyHI7OLUqVOZOpFarWaCiaPmOclsqyUC4EqVGn/55yF09w5AHB5s8nUSeS8C/X2w/bdr4O3lyQzNy2QyPXPr4OBgxtrK2QbWJCVOdlTWQhRnZDKZVelaIlRPUtHOgtRKR1PEcZRPZ3NzM9MlzoXI+iuvvIKOjg589NFHTu1uHoWbO1gCP499tLa2oqmpyaph76qqKmY1ZwpzjULGRN111feJXY0taLVapmHEmkYNlUrFBE6FQoGwsDCIRCKLV7C6WrMzZsxw6oeDBGzDurQldU4uIGo4M2fOtGghcqW8Fh/u/woJMebTxg2tMjy4ZgHmpk/V+7puF2Z3dzdUKhXi4+MRFxfntHEZkoHhWvfWsLvWXLqWoiiUlpYyNX1nYY0Yue5MZ2dnJ2cznS0tLZDJZJxI6dE0jT//+c+ora3Fnj17nDqeZYbxGyxVKhXMXBsA4Pz58wgLC4NCocCMGTOs+mPV1NQgMDDQ6NgH20ah69evMw9kchyX1loqlQqlpaUQi8Wc1J+0Wi06OzshlUqZ7jWRSGR2tW7OrcNRkL8L2WGPFrCNzXOSBywX85xE9CAtLc3iHfZrH+Shr38IwYHmA2zfwDC8PD3w/OPrjP7e+/r6UF5ejsTERAwODjIPWJJNcJTmqkKhQElJCZKSkuy6szWXrqUoyuIUsL2uk4umIi5mOrkOlG+//TaKi4uxd+9epyowseDmbvBRq9UYHh6Gh4eHTW9CU7J5JH3j7u5utlFIt0mI60BJNE25TKu5ubkxrgQURaGnp4fRszUVTNRqNUpLS12iYeTatWsQCASs0kjGJOmkUqmeJJ2185zNzc2QSqXIzMy0eKGmVKnR3CZHPItdJQAE+vugsU2OoWEl/Hz1Ax/Z2WZkZDD3QYyQZTIZrl69ytTD7JmaJsbiturvssGYdi0RCB8YGIBGo0FkZKSe7Z2joWka165dg1AotHlxaatPZ2trK+OLyUWgfO+99/Djjz9i//79rh4oOWHMBkvi6OHh4YGkpCSbzmWsG5bUfaKioli5ZJBuWN1GHi7SfaQex7WmqS5CoRChoaEIDQ0dEUyIv6G/vz+qqqqQlJTk1E5TrVaL0tJSBAcHWzVTqitJl5SUxMxzVlZWWlTnJKnooaEhZGRkWPW31mi0ELCQDdRFKBBArdF/r46m8+rj44O4uDjExcWNkBoMDg5man9cpNIHBgaYESo2zXJcQwTCw8PDUVRUBJFIBK1Wi8LCQkaqja0SFhfYs/tWd5GnO9NZXFxsdKazra0NEonEpKSfJdA0jQ8//BBnz57FwYMHnWpj5khcNliSoV1jkLGB1NRUlJeXg6IomwITscci9PX1oaysDNOmTWPd0SgUCqFWqznteG1paUF7e7tD3RgMg8nw8DCamppQXV0NHx8fDAwMwMfHxyYfR2uxZiTDHN7e3sxqna1uLZkZdHNzw8yZM63+PXh7ecJNKIRGox1VOpCg1VKAAPD1+TkgWqLzajiCQ7IJtbW18PHxsWlUg1h92cO02RLUajWKioqQkJDAlFUcaTVGIIHSw8ODmSu1F+Z8Oj09PaFWq5GZmclJoNyzZw8KCgpw5MgRpzb2ORqXrVmSwKMLEQKQyWRIT0+Hp6enTeo7hI6ODgwODiIpKQkSiQS1tbVIS0tj3ahB0zRzXFRUFMRisU31IdIMQCTSnNlAI5FI0NDQwGirkgah4eFhhIaGQiQSISgoyO6Bk/g/WjKSYQum6pwhISG4du0agoKCOFFL2nf8HC4WVSFaZD5l2S7tRuaMJGxavQAAdzqvNE3rdU0DsMjcmji6pKWlOdVJhTQVJSYmjpr9sLW71hw0TaOiogJeXl5OVZACbiymmpubERwcjJ6eHnh7e9s00/mf//wHX3zxBY4dO2YXg247Mn4bfAyDJRnkdXNzQ3JyMvOm/umnn5CcnGzTh5Q8EN3c3Bj3AbbBV7c+Sdr5pVIptFqt3pgG2w+MRqNh2tydPQ/W2NiIrq4uzJw5c8Tvg9RMpFIp+vr6mAYprlJ6uvT09KCystJp/o8kzdXR0YHm5mZ4e3sjNjbW6jqnLs3tcrz6fh5iI0PhPsrvTaul0NzRiT9sWY34mAi76rzqqs0MDw8zO2xjNk1kB5OWluawBiJjkKaiSZMmWbSY0u2u7e7uho+Pj03pWpqmGbN1Z35+gRv2Z21tbYztGACbZjr379+P3bt3Iz8/36nZAysZv8FSo9EwdUTyQTBWPyRdd7b88eRyOSorKxEWFoZp06ZZZK1lqpFHrVbr7cLYjGkoFAqUlpZiwoQJnKUZrYGiKFRVVYGiKL2FiSlIi7tUKmXEsslCwdbCv1QqRX19vcs8jCdOnIiAgADmb8vFPGf+N5dx9MyPmBAVDg8j6ViNVoumNjmWzc9E1l1zmDqYI7qRtVots8Pu6emBv78/oyLU1dXFWJ85s25FhCBsbSqyVQxBN1Da2kdhKx0dHWhpaUF6errJxjNLZjoPHTqEDz74AMePH2c9Q+xijP9g2dvbi/LycpP1w/LyckyYMMHqP6BSqcTly5fh7u6OW265hfVxpJEHMC9dZzimERwcDJFIpFcLIzWf5ORkBAcHW3UvXKDRaFBWVoagoCCrtETJA4cEE1u6TUmnaWpqqlO77UjjirG/jUajYR6upGnG8G9rDpqmcfpcMY6e+REAEBroDw8PN6jVWnT1DQAAlt8xC0tvT8PVq1edpvNK1KFkMhna29uhVquRkJCAyMhIp6VfSfdtcnIy5w9xS9K1RALTz88PEydO5PQ6LEUikaC5uXnUQGmI4UynQCDAmTNnkJ2djfr6erz11lvIz8+3e4ezHRnfwbK5uZmZXzOVH6+srGR0UC2FNPLEx8ejq6sLqampZo+xVZGHNFZIpVJ0d3cjICAAnp6ezM93Zh3AHjtbXfcQsgsTiUSjrtSJJNjw8LDT/f16enpw7do1Vt3Ihn9bS+c5e/oGcbmsBoXF1zGoUMHH2xO3pE3G7JmTEOjn7RI6r8CNRYxMJsPUqVMZMQS1Ws2Mpdhbpo0wODiI0tJSh3TfjpaudXd3dwmFIOBGJqaxsREZGRk2CQR0dXXhww8/xPHjx1FbW4uNGzdi3bp1mDdvnkXn3bx5M44fPw6RSITy8vIR36dpGlu3bkVBQQF8fX3x8ccfIzMzEwBw8uRJbN26FVqtFg8//DC2b99u9f1gPAfLuro6SCQSs6a91dXVzEC9JXR0dKCurg5paWkQCASoqqpCRkbGqMdwLV1H0p3Ev9HHxwcikYhT8Wi2EKm2adOm2W31qFarmV3Y4OCgnkQbCYhkhe7l5eVwsWtDbEkB67bzy+Vym3bYpHElLi7OqTqvAPSEIHQXMcQFRyaTMSIXJKVnjwY1stt3RvetYbp2cHAQgYGBmDx5skOtxgwhgTI9PZ2TTMzXX3+Nl156CQcOHEB5eTmOHz+OCxcuYPPmzdi2bRurc3z33Xfw9/fHpk2bjAbLgoICvPPOOygoKEBhYSG2bt2KwsJCaLVaTJkyBV9++SViY2MxZ84cfP7555g+fbq1tzN+RQmInqS5N56ljiFEW7W7uxtz5syBh4cHVCqV2XPYI1BWVlZCKBRi3rx5EAqFGBwchFQqRUlJCYRCoU3D8pZAZjlTU1M50Yw1hYeHB6KiohAVFQWKotDV1YWOjg5UVVUhMDAQISEhaGtrg0gkYjXbak9aWlrQ0dGBzMxMqx48hu381s5zEmMAZ+u8kt2+SqUyKrzt7u4OsVgMsVisZ25dW1vLmFtzJeZPShb2fr+agogh+Pr6ore3F6GhofD19UV9fb3DrMYMkclknAbK7777Di+++CLy8/MRGRmJhIQErFy5kvGuZMv8+fNHdRw5cuQINm3aBIFAgLlz56Knpwft7e1oaGjApEmTmJT2hg0bcOTIEVuCpc24bLD08PAwqqpjiCXBUte6KzMzk3kjmzsH14o8RAknPDxcz7nEz88PiYmJSExMZB6uFRUVjCGwSCTifL7RGbOcwI06L2meoGmaCSRubm7o6upidmKO3mETGb3+/n5kZGRwtivSneckdU5dcQBjdU6SZnR2HZuo0AgEAkyfPt3s+08gECA4OBjBwcGYPHkyswsrKysDRVE2mVuTtPhopRlHYExzliwCSbq2pqbG5u5aNshkMtTX19s8Qke4ePEinn32WRw/fnxEJkMgEHBaG25tbdUrK8TGxqK1tdXo1wsLCzn7udbgssGSLe7u7nqCAqYg3YzR0dEjaj5CodBksOTSgxL4eV7QnBKO4bA8adEnnbURERE2zTfq1gW5GFa2hcHBQWb8IDg4WG+HLRAImF2JvR+ORMOToiikpqbaLZ3m7u4+QhxAKpXq+XN6enoywuzObNMnWsDe3t5Wzwz6+fnBz88P8fHxepJ0ZBfGVuCezHOmp6c7tTOaBMrQ0NARGRBDaUWyUCgtLQVN05yLIcjlck4D5Y8//oinn34aR48edYhMoLEyoClBGmeWZAAXDpZsfzFubm5QKBSjvoZ01CYnJ+u5goz2s7hOuwI3iuZVVVUWzwvqpi9JZ21raysqKyut6r4kM6ve3t42KdBwAdE01U2p6e6wlUolZDIZrl27xjSRiEQizrVNSdbB39/fofNxxqQGm5qa0NHRgYCAAGaX7YxuUxIUiLQgFxBJOt1dmDEjb8MHf2dnJ2pqaoxK+jkSIrcYHh5uttFqNO1aLtK1nZ2dqKur4yz1WlRUhCeeeAKHDx92WBkkNjYWzc3NzL9bWloQHR0NlUpl9OvOxGWDJVvMpVBJI09GRgbrnYk9AiVJLWRmZtr0YTcmgE52Jf7+/kyDkKmdItfuJbbQ3t6O5ubmUR+AXl5ejJE3aSIxJ0dnKa4iEC8QCDA0NITBwUHcdttteulpey4UjEFMm0UiEWsTdUsxJnAvk8lQVFSkV7MfHBxkdk/OnOckgTIiIsKq34nhQoHIDVqTrtVdPHDxOykrK8Pvfvc75ObmOnT0JSsrC7t27cKGDRtQWFiIoKAgREVFISIiAtXV1aivr0dMTAz27duHvXv3Ouy6jOGy3bAURbFKr3Z3d6OjowPJycl6XydC1729vUhLSzPb7kx8Me1h1lxTU4OhoSG7ej+SGTipVAq5XA4vLy+IRCK9Dx+ZSXOFZpHGxkbGpd2aFndDOTqyUAgLC7PofGRchmvfRWsgOq/GFKTIQkEqlTJ1TtJtynUTiVqtRnFxMacavJZCMgotLS0YGhpCdHQ0IiMjHSKtaAx7Lh5007VyuRw0TTOlFmPp2q6uLlRXV3MWKCsqKrB582bs27eP8waa++67D2fPnoVcLodYLMZLL73EPNd/+9vfgqZpPP744zh58iR8fX2xe/duzJ49G8CNTtlt27ZBq9Vi8+bNeO6552y5lPE7OkLTNFQqldnX9fX1obGxETNnzmS+ptVqUVZWBm9vb9bWXRcuXMCtt97KaSMPSe35+vraXUzZEF1hAJIOIs7ozpCMI5C6oFarZaUOxPacfX19kMlk6OzshKenJ6vuS1I/doSdlDkaGhrQ09Nj1psTgN6upKuri1N/TjKmoitE7iza29vR2tqKGTNmMH9fIq1oSmnGHpBAKRaLHVLHG00MoaenB9XV1UhPT+ckHX39+nVs2rQJn332md4zdBzCB8vBwUHmzQP83GpPUndsuXDhAmbPns2ZtZZSqURJSQliYmKc6qcH3Bgib2hogLe3N9NkYI/OWnM4qi44NDTE7LBpmtbT6CX09vaioqLCaXqzBJqmUV1dzYjmW/reYzvPOaxQ4cKVapRXNUOp0kAUHog7bklG4oSfm8zIZ8dRYvWj0draio6OjhFZId2xlK6uLnh6ejLpS3s0/Wi1Wsaqzxk1M92FERF/mDhxIqKiomzeVdbV1eH+++/Hnj17zM6YjwP4YKlQKHD16lXMmjXLbCOPqZ9DURRqa2shlUoZJ43g4GCrH+ZkwH/q1Kmsr8Me6IqhE3EH0lkrlUod6hxC7LWio6MdunhQqVSQyWSQSqWMu7ynpycjMO1Mlwwya8ulzquhYlJISCguFLfg+Del0GgoADeyJlrtjTJDUrwYTz28DKFB3igtLcW0adOcOqYC/KwQxMZ7cWhoiBEH4NrcWqPRMJrUzm4uIWYC06ZNY2TpzKVrR6OxsREbNmzAhx9+iDlz5tjxyl2G8RssgRu7M3NoNBpcuXIFcXFxaGhosGj+yrA+SdM046TR29uLoKAgiMViixpIyCD2zJkznTIwTaAoCteuXQMAk+LwpLOWpLfsVQdzlVqpVqtFdXU1Ojo64OnpyTQIOXJ4XPdaysvLERgYyInVlzHUajV2/uMYzv5wDZ4ebvD28oCHhwfc3T0gEPy3zj2ogLubEL9aMR2L7pjr1F02AKaWnZqaavHfhAiD6+r0Wvv31Wg0KC4uRkxMjFNNDYCfZ0sNR2astRprbW3FunXr8P777+PWW291xC24AnywpCgKZ8+eRVBQEKtGHoK5Rh6appkGkq6uLgQEBDANJMZWuzRNo6mpCXK53OnC3xqNhpkDi4+PZ/UgNtQ1tbZhxhCS7nSEfqc5dP0fBQLBiLofuV97/+3IjkUsFtut0xQAzv5QiTf/VYDgAF8IBAJotBqo1Wpo1Bq4uQnh4eEBgUCIrp4+RIpC8dFftsDNzXk6vHV1dRgYGOBED9iwruvr68va3JoEytjYWKfLC/b29qKystLsbKnu/Y5mNdbe3o61a9fi7bffxh133OGIW3AVxnewVKlURodTCaSRp6urCwsXLmS9OrdUkYc0kEilUnR2duppuHp4eFhsaWVPiPhCfHy81R903c5a0jBj2FnLBrLLdrYpMDHTVqvVRuuCpO5H6pweHh5M3Y/rOpijdF5pmsZjz38MWWcffH28DL4HUJQWSqUSKpUKbm5uUKppPPd4FubNnma3axrtWmtra6FQKOziz6nrhCOXywHAaB0b+LkTOC4uzund0WwDpSGG3bUnT57EwMAAFi5ciJdffhk7d+7EokWLrLomc+Lmb7zxBj777DMANxYdlZWVkMlkCA0NRUJCAgICAuDm5gZ3d3dcvnzZqmuwkps3WOo28jQ1NWHevHmszmmrIg95I5IHq5ubG5RKJSIiIhze8WoIqZVyLY9m2FkbEREBkUg0agAk2qqpqalOnY0jwuze3t6s/z7Dw8NMnZOiKKtMvE2dt6SkxCENNLWNEjz1ymcICvAxes1qtRoKhYIJFt29A0iIDsYDWWkOdQ8hDU4ajQbJyckO+fyQOrZMJoNCoWAE/X19fVFaWuoSgbKvrw8VFRWcLDQlEgk+//xzfPLJJ1CpVFi2bBlWrVqFhQsXWtRRa6m4+bFjx/DWW2/h66+/BgAkJCTg8uXLzirFjF8h9dHo6enB1atXMX36dISEhKCpqYnVccSD0paxEF1VjsjISJSUlCAwMBA9PT24cuUKq0BiD4gWpT00M4lcWUJCApRKJaRSKSorK6HRaEZ01hKh+oGBAU61Va2BpKOJBi9bfHx8EBcXh7i4OKhUKj2pQWsbokbzxLQHss5+CIVCo9eoUqmgVKrg5+8P4X+/7+/ni2G1ALNmzUJnZyeam5sZ71V71XXJGBEAhwVK4IY4AOlS12q16OrqQmtrKyQSCaN7qtFoHDKWYgwuAyVwQwHs8OHDeOONN3D33Xfj+++/x7Fjx/Dss8/i4MGDrG3FLl26ZJG4+eeff4777rvP5ut3FVw6WBrTCGxra0NjYyMyMzNZv5HsochDutOmT5/OfMDIIDUJJGRHYm9tz+bmZkgkEoeIoXt5eZnUrA0JCcHQ0BC8vb3tqq3KBq7SnZ6enowDju6DtbKykrUNFanbOlTnVQAIjCSGlEoV1GoV/P1H7pIFAoGee4huHay6uprTeU6aplFRUQFPT0+nZmTc3NwQHByM+vp6zJw5E15eXpDJZGhoaICHhwdT93PU4re/vx8VFRVITU3l5Gf29vZi3bp12L59O1atWgUAuPPOO3HnnXeOWuIyhiXi5kNDQzh58iR27drFfE0gEGDJkiUQCAR45JFHsGXLFivuyHm4dLDUhSjh9Pf3Y86cOXqrPoFAAIqijK58SaDUarUmV9qWoivTpltL0JVmMwwkZAfGpVQZSWEpFAqn7OJ0NWuVSiWKioogEAigUCgYU25ndJqS7luu0526M4w0TevJlfn6+urVsQnE/szRYyrRomBQNM3U5QFAoVBCo9EYTScrlCrMnDrSYMBQt5bI0dniz0lS476+vg7V4TUGWVQlJiYyxgZBQUGYNGkShoeHIZfL9eQG7Zme7u/vZ6zHuMgO9ff3Y/369di2bRvWrFkz4vuW3oMl4ubHjh3DL3/5S73RufPnzyM6OhpSqRSLFy/GtGnTMH/+fIuuwZmMiWCp0WhQVlYGX19fZGRkjPgDubu7Q6PRjNhV6TbycBEoSYqxv78fmZmZo6ZpDMXPdS2ZuJjlJGLoPj4+ThdDJwIMCQkJek4aZEfCVWctG4jXob27bwUCAUJCQhASEqJXx9YNJAKBAO3t7U4R/46LCUfiBBGa2uTw8/GGQqkApdX+N1Dqv/bGQ1CAVYtND6ab8+dkG0goikJZWRkCAwNZp//sBQmUEydONFpH8/Hx0bNV001Pc21uPTAwgPLycs4C5eDgIO69915s2bIF9957r83nA0yLnhtj3759I1Kw5LUikQirV6/GpUuXxlSwdOkGH41Gg4GBASadZmqY/aeffkJycrLeCpdrD0qtVsukjWwZICemx7qznJbuwMiAf1RUlF1HD9hAPuRTpkwxKsDAVWctG3QdTJzpdahQKHD9+nW9zmlr/Rtt4YeiGry66wg83AGhQAAfH98RgRIA+gaGESUKwa6XN1mVBSCBRCaTjVrnJELkYWFhTjf3VqlUKCoqQlJSksUNJxRF6akIeXt7M+lpaxZFpJ7NlZn18PAw1q9fj1/96lfYvHmzzecjaDQaTJkyBWfOnEFMTAzmzJmDvXv3IiUlRe91vb29SExMRHNzM3M/g4ODoCgKAQEBGBwcxOLFi/HCCy/g7rvv5uz6zDC+G3y6u7tRVlbGNPKYwtB5RLeRh4sUIJfBydD0mMw26u7ARnMNIXqmriBJ1t3djWvXro1aixMIBAgMDERgYCAmTZrESNHpelVy0RDV0dGBpqYmp1s4ATfS9BRFYf78+aAoivEcHBwc5CSrwJZb0pNw26x4nLlYjeBA/xGBkqJo9A0MIzDABy9sXW31Z2W0OidJT4eEhKCiosKuLiZsUSqVKC4uxqRJk6z6DAmFQiarAPzcLV5WVsbISbLtniafZ64CpUKhwP33349169bhoYcesvl8uri7u2PXrl1YunQpI26ekpKCDz74AMANYXQAOHToEJYsWaJ3PxKJBKtXrwZwI+jef//9jgyUnODSO8uWlhb4+/ubfZCWl5djwoQJCAwM5LyRh+yc7B2cDF1DvL29mR0JqYGR4ORsPVPgxpu/sbERqampVs8ikoYoqVQKtVrN1HUt3YHpikE4q4MRMK/zqtVq0d3dzWQViCC4KaELW6AoihHxr2kdxGeHLqB/UAGNVguh4Ebrj5tQiFkzE/G7jYsQHsr9+4nUOSUSCZqamuDt7Y2YmBindIsTSKCcPHmyXaQoSfe0TCbD0NAQM5YSHBw84v0wODiI0tJSzhq/lEolHnjgASxZsgRPPPGE082SXYzxPWep1WqZmcjRIM0kwcHBnAZKMo4xY8YMh7vVk2YKmUwGd3d3eHl5ob+/3+ku8YB9ghNpiNJ9yJgb0SBNXwqFAikpKU4Vg7BU51VXELyzs5NZHLFRmDEHSXcSBacbX6NQVNGI63XtUCrVCA8NwK2Zk+0SJHUhQ/4TJkxAcHAwo0vsiIYZQ0gTmqmSAdeQkotMJkNPTw/8/f2ZxRHxleUqUKrVajz44IO47bbb8PTTT/OBciR8sARu2MwEBAQgPDycs0BJxjGcPVRP1Gc6OzuZHSYxf3b06tycEg5XkBENqVSKvr4+o3Vde4iQ23K9tuq86gpd2JKeJlJ6kZGRTne7Gc3ui22dkyuIiImz7Nh0VaLIgpB0zttaX9doNNi8eTMyMzPx7LPP8oHSOHywpGlar+YnFottSmuRQWmNRmPXgMAGEhCEQiGmTp0KoVCol7o0JgpgL0j3ra+vL5KSkhz2gdSt6xIN1/DwcLS3tyMsLIy19q29sIfOq0Kh0NuBkRqYubEj3V2cszVNSbqTTQONMR1XY2M41kICpSs4quiONRGlKOKGExERYbHYhVarxSOPPILJkydjx44dfKA0zc0dLHUbeQCM0G8Vi8UIDw9nnSokIypBQUFITEx06htPrVajrKxsVDF0Q7utsLAwiEQiztNaarWaCQi6Q8mOhrjCXL16FUKhkBmSF4lETtn9k51TfHy83eTRNBoNk54eGBhASEgIU3LQXciR4DRx4kRmXtBZkOBkTbrTUE5SKBTa1ATmSoGSyB1Onz5db6xJo9Ew6Vpdc2tzi36tVosnnngCUVFRePXVV/lAOTrjO1hSFAW1Wj3i62wcQ3SFsT09PSEWi/WaZQwZHh5GaWmpTQLkXEGuJSEhgfVDmNhtSaVS9Pf3Mw/VkJAQm/VMS0tLkZiYOCKV5mjIw4ZYfZHOWks0a7m+Fkd2JVMUxTQI9fT0ICAggOm6HG18x5GQ3wtXwclwl21JnZPra7EFU4HSEMNatpeXFzOWoturQFEU/ud//gcBAQHYuXOnUzNgY4SbL1haI11HVqqkWYbU/MhuhMiROUq3czTIUL0t10IeqhKJxOpZTuBnYXZdST9nQWbRTF0LV521XFyLIyBOOG1tbWhra0NgYCCio6PtMr/KFpJiTE5OtsvvxbDOOdr7mgQne12LJdhyLUNDQ0yjX21tLYqKipCdnY28vDwIhUL8/e9/typQmnMPOXv2LLKzsxnhiDVr1uCFF15gdayLcnMFSy6EBoaHhyGVSiGVSiEQCODj44O+vj6Hy5EZg3wguByqN6z5sVXTITJtzh7wB26MzFRVVbE21CYmwFKplPPZRqfovJqALGaIghN5qAKmLajsBRmDcNRYkym/yoiICKjVapSWlprdxTkCkgbmImh3dnZi//792LdvH+rr63HfffchOzsb8+fPt2iBxMY95OzZs9i5cyeOHz9u8bEuyvgWJdB9sHGlyOPj44P4+HjExcWhpqYGMpkMHh4eKC8vZ9J4zggOpPt21qxZnJoPG8qykVnO+vp6o7OcwA2x+tbWVocIs5uDXKslIzMeHh6IjIxEZGQk01nb1taGa9euMTJlYWFhFq/InaXzaoyenh5cu3ZNz2VG1xlGLpfj+vXrUCqVTIOQPTVNSdB21ALCULeWCANcuXIFQ0NDiImJcerMLaBfL+VidxsaGgqZTIaUlBR89913OH/+PI4ePYrf//73eOaZZ3D//fezOo+l7iFcHTvWcelgSSAelFwp8lAUhYqKCri7u2Pu3LkQCoVQqVSQSqW4du0a1Go1Ezjt/eEn4xgqlQqZmZl2rT0YqukY6pmKRCIoFAoMDg4iMzPTqfZawA1RCuKmYu0Cwpj4uVQqRU1NDfz8/JiuS3MPViLC4AoLCCLrZ2oB4eXlxVhQGWqaBgcHM7VsLt5rxE6KKwUaayC2eQKBAB0dHUhLS8Pw8DCqqqqgUqkcPs8J/NxwxVW9lKZp/PnPf0ZzczM++eQTuLm54a677sJdd90FmqahVCpZn4ute8jFixeRlpaG6Oho7Ny5EykpKRY5j4w3XDpYkt0kl0IDZBhYJBLp6VN6enrqOYYQNwmFQoHw8HCIxWLO619kPs/Pzw8zZsxweDebn58fEhMTkZiYiKGhIZSXl2N4eBi+vr5obm522i5b1xMzPT2ds6BtuMsmTWCNjY3w8PBgdtmGcnm6QdvZuxWZTIb6+nrWQduYFJ1UKsX169eZIXlLOsZ10d3dOnunbSwNHBsba1QAnSwW7LUYJOIHU6dO5SxQvvXWW6isrMTevXtHXLdAILBIqISNe0hmZiYaGxvh7++PgoIC5OTkoLq62iLnkfGGSwfL/Px8vPfee8jOzsaKFSts7jokOoxJSUmjttd7eHgwHoakdZ9oe4aFhUEsFtu8SlUqlSgtLUVMTIxJ5X5HodVqcf36dURERCAhIQFqtdroLtves5zAjQ9yZWUlBAKBXT0xdV00kpKSmM7a0tJSAD/X/Ig0HZdB21ra29vR0tKCjIwMq3bahqnL/v5+yGQys4sFY5jb3ToS0nBlLA1sTreW/J25Kn2QHeWUKVM4ET+gaRrvvvsuLl++jP3793NynWzcQ3RrvcuXL8ejjz4KuVxukfPIeMOlG3xomsa1a9eQm5uL48ePIzAwEFlZWVi1ahVjgcSWrq4uVFVV2dSAYGw8QywWW9w44ii9WTYQkXhTQdtRs5zAzzvtgIAAp865ks7a+vp6aLVaxMbG2iWzYAktLS2QSqV2078ljW8ymQwURektkAzp7OxETU0N0tPTnS5aP1qgHA3dOqdMJmPmOSMiIqzOphAnE650Z2maxr/+9S989dVXyMvL4+x3zcY9pKOjA2KxGAKBAJcuXcLatWvR2NjINPiYcx5xQcZ3N6wuNE2jtrYWeXl5OHLkCDw9PZGVlYXs7GxERkaO+hBrbW1FW1sbUlNTOXvDGbPaEovFZmtBZEXuDL1ZQyw1STY1y2lMJNpSiPBBZGSk010piHKSh4cHEhMT7dZZy5aGhgb09PRg5syZDtndEjFwqVQKhUKhpy5Dsizp6elOr91yaW1F5jllMplVdU4SKK11MjGEpml8/PHHOHr0KI4cOcL57r2goADbtm1j3EOee+45PfeQXbt24f3334e7uzt8fHzw17/+FfPmzTN57Bjg5gmWutA0jaamJuTl5eHQoUOgaRqrVq1CTk4OYmNjmTc32akIBAKkpKTY7UGjWwvq7u5GYGAgM56hG0Ta2trQ0tKCtLQ0p6/IyQiEtTttwwF5U/fMBoVCgZKSEpcQPtBqtYyKk6HOK0VRzJwfcQ0hc372eG+RBeLw8LDThOLJAomMaFAUhalTpyIiIsKpaWnSgWuPxiJj85yjGT0TJaekpCTOMkX/+c9/8MUXX+DYsWNOH90aJ9ycwVIXmqbR3t7OBM6hoSGsXLkSd911F/7whz9g6dKl2Lp1q0O1THt7exnZPdJE0d/fj8HBQYftDkaDjGOkpqZy0phheM9+fn6MRq+5lCGpIztL4FoXS3RejWnWsu2sZQPpkqYoCtOmTXN6EwWpl5KdNrln0iDE5biTOewZKA3RrXN2d3fDx8eHuWdPT0+o1WoUFRVh4sSJFptIm2L//v3YvXs38vPznZ59GkfwwVIXIqq+e/duvPHGG0hJScHChQuRk5PjFGcKEkQqKiqgUqkQHBzMyO45q6uyubmZqX3Z4wGn62FI5LpMWU/19PSgsrLSJQb8bdF5NZRXtLRZxtj5Kioq4OnpiUmTJjk9ULa2tqKjo0OvyYncs0wmg1wuZ0aPIiIi7Nrwozuq4ugdl2GdUyAQQKFQIDExkbPSwaFDh/DBBx8gPz/f6YIK4ww+WBry008/YfPmzXjnnXcwffp0HD58GAcPHkRHRweWLl2K1atXIzk52SEpLaIkEhERgQkTJmBwcBASiYTRqyUPF0fUfoj34/DwMGbMmOGwlJ6u1KDuA7W/vx91dXWc7W5tgWudV12JMpqmLRK7IKbNpMnJ2TQ3NzPepaNlRIiDhkwmg1artYsbTm9vLyorK11CVUqtVuPKlSsICgqCQqHgRPzh+PHjePvtt5Gfn+/0LMs4hA+WuvT19WH58uXYs2cPkpKS9L7X09ODo0eP4uDBg2hoaMDixYuxevVqpKam2iVwmBMgHxoagkQi0QsiIpHILrVMIsLg6emJyZMnO22nolAoIJVK0dLSAoVCgfj4eERFRTn1wWdvnVcidkEaR0bTrCWmzWFhYXozwM6ioaEBvb29mDlzpkWfEd0Oal0jb1uaokigdIWZTmKFFh8fz3y2La1zGnLq1Cm8/vrrKCgocHqH/DiFD5aGUBRl9oPd39+P/Px85OXloaqqCosWLUJ2djZmz57NSeAkqVe2D2CyKpdKpQDAqXsGsfoKDw93+gOYpmmms3PatGlMN7G5IGIvHK3zSmZ2jXXWarValJSUICoqyiXm1ogohK1ZCCI3qNsUxcZ+ShdXEj/QaDQoKirSC5SGUBTF1PB1dWuNlSIA4Ouvv8ZLL72E/Px8pze4jWP4YGkrQ0NDOHHiBPLy8lBWVoY77rgD2dnZmDt3rlWNOLY2zyiVSkboXavVIiIiAmKx2Krdl0KhYGzH7OW3yBbSsKLRaEakwXWDyNDQkNVGuJZAdF6d9QDWHT3q6emBSqVCTEwMJk6c6NQGMNKBq1AokJKSwunv39B+yrBZxhgkULqC+IFGo2HMtdl+nnTrnHK5HAKBAF5eXhAKhYzG63PPPYf8/HyrrQHNuYB89tln+POf/wwA8Pf3x/vvv4+0tDQAQEJCAgICAuDm5gZ3d3dcvnzZqmsYA/DBkksUCgW+/PJL5Obm4sqVK5g3bx5Wr16NX/7yl6wacpqamiCTyThrnlGpVMyOk+y+xGIxqzoQSS+6gu0YqcP5+voiKSlp1Gu35ywngei8usKsIJFGi4yMhEqlYrqJndFl6sgOXGOiAKSeTRYvxG3GlQJlbGysTX63SqUSP/zwA/70pz8xc6zvvfceVqxYYdX7m40LyIULF5CcnIyQkBCcOHECO3bsYPRcExIScPnyZc46eV0YPljaC5VKha+//hp5eXm4cOECbrnlFuTk5Bi1w6FpGlVVVdBoNJg+fbpdaqCGSjokbRkQEDDioUaED9haWtkTjUaD0tJSq9LApmY5bZlrJDqvaWlpTtd5JY1FumMzup21nZ2djP+qtZ21bCFqWUKh0Cmd4wqFggmcarUafn5+6O3tRWZmptNTr1qtFsXFxYiJieHMGP7HH3/EU089hS1btuDChQsoKirCrbfeit/85jeYPXs26/NcvHgRO3bswKlTpwAAr732GgDg2WefNfr67u5uzJgxA62trQD4YGnRCfhgaR6NRoNvv/0Wubm5+P7775GRkYGcnBwsXLgQKpUKW7ZswR//+EfGV9DeaLVaJnAODAwwEnRBQUGQSCRoampyCeEDpVKJkpISxMXF2fyQ0Z3lJHUgS+YaSb2UNKw4e9aVzJea81zUlaGztLOWLa42qiKTyVBVVQV/f38MDw/bJbvAFhIoo6OjERUVxck5i4qK8Oijj+Lw4cNMx7NGo8GFCxfg7e2NX/ziF6zPlZubi5MnT+LDDz8EAHz66acoLCzErl27jL5+586duHbtGvP6xMREhISEQCAQ4JFHHsGWLVtsvDuXZXz7WboK7u7uWLRoERYtWgStVovz588jNzcXf/zjH6FUKrFq1SqHPmTc3NwYcWjSQNHS0oKSkhIIhUJMmzbNoek7YxApvSlTpnCikykQCBAcHIzg4OARjiHmxnB066X26n62hL6+Ply9epVVYxHxX42Pj2fS8rrWU6ayC2yhKApXr15lHGicHSg7OztRV1eHOXPmwMvLi6ntSiQSVFVVISAggElR23vBQwJlVFQUZ4GyrKwMv/vd75Cbm6s3GuTu7o758+dbfD5LXEC++eYbfPTRRzh37hzztfPnzyM6OhpSqRSLFy/GtGnTrLqOmwE+WFqIm5sb5s+fj7CwMHz//ffYunUramtrceedd2Ly5MnIycnB0qVLHTZk7+bmhvDwcHR2djKOKMRejNgRhYaGOjRAkGBgi2j9aBg6hpDaF1kskN2Xt7e33tjM9OnTnR4MjJk2s8XT01PPp1Iul6OhoYHprI2IiGB2CWygKEpP2s/ZEIH2jIwMZtEjFAoRHh6O8PBw0DSNvr4+pomOCF7YY1bZHt3JFRUV+M1vfoMvvvgCU6ZM4eScbF1ASktL8fDDD+PEiRN6oynktSKRCKtXr8alS5f4YGkCPg1rBd999x22bduGzz77DMnJyQBuPHiKiopw4MABnDp1CnFxccjKysLy5cvtMr9HIPq3/v7+mDhxIvOgJHJsEokE3d3dCAgIYLRb7bkid3aXKal9kW5ijUaDsLAwTJ061eHXYggJBmlpaZw2rBiK+rOp7braTKdcLkddXZ1FTVe6DUIAOEtRk0ApFosRExNj07kIVVVVePDBB/HZZ59h5syZnJwTYOcg0tTUhDvvvBOffPIJI4YO3Pj9URSFgIAADA4OYvHixXjhhRdw9913c3Z9LgRfs3QGdXV1jP6pMUj3Z25uLjM7lZ2djZUrV3KSkiQQe63o6OhRP9S6K/LOzk6L631sIfqhaWlpTu8yJYPjPj4+UKvVTpvlJEilUjQ0NNi9A9dQp1f3b01S8yQYiEQipzu8ANYFSkOIrZpMJrNJTYeiKJSUlCAiIoKz301dXR3uv/9+7NmzBxkZGZycUxdzDiIPP/ww8vLyEB8fDwDMiEhdXR1Wr14N4EbQvf/++8eKg4g18MHS1dH15Dx27BiCgoKQnZ2NVatWITw83OqHNumknDRpkkWdbLrarXK5HN7e3kwqy5Y6Z2NjIzo7O+3mt2gJxnReDQUBdJui7B0429ra0NbWhrS0NIePghhq1oaFhUEikWDChAmc1eFsgfiGWmtobQxDNZ3g4GCIRCKz9nkkUIaHh2PChAmcXEtjYyM2bNiADz/8EHPmzOHknDxWwQfLsQTRZ83Ly8PRo0fh5eXFeHISo1U2kJogFxJtunq1ZExBJBKxXuGTe1IqlXYbm7EENjqvpClKKpWir6+P9cPUGpqbmyGTyZCWlub0Dty+vj6UlJTAzc0NHh4ejNmxs8aLyG6by0BpCBk/Iq4hxAXIMKtCURSTluYqULa0tGD9+vV4//33ceutt3JyTh6r4YPlWIWmaTQ2NiIvLw+HDx8GAMaTMyYmxmTglMvlqKmpsYuY9NDQEDOmoDskbqq+5iqaswRrdF4NvUi5rO3W19ejr6/PYm1Ve0B22wkJCRCJRHqCFyRtaWtnrSWQTub09HSH7bZpmkZ/fz+ToibuMGFhYaiqqkJoaChn9dv29nasXbsWf/vb3/iGGdeAD5bjAZqm0dbWxnhyKhQKrFy5EtnZ2Xrt/F9++SWCg4MdUhMkoudSqRQ0TTM7TtK0QxpEQkJCXKKTkgudV0M5Nmtru66221YqlYw5sbGUPUlbkrlde881kllgRwZKY5DFYUNDA9zc3BAbG8s4pdiCRCLBPffcgzfeeAOLFi3i6Gp5bGT8B8sDBw5gx44dqKysxKVLl0yqW5jSR+zq6sK9996LhoYGJCQkYP/+/S5tf0M8OQ8ePIiDBw+it7cXy5YtQ2NjI+rr63HkyBGHN88Q5wypVAqNRoOQkBB0dnYyriHOxh4duIb1PraWaqRGLRAIMHXqVKfvthUKBYqLi1nPuxrrrLVU+Hw0SKDMyMhwem2bNOIFBgYiOjqaqWkrFAqr9YnlcjnWrFmDP/3pT+O1q3SsMv6DZWVlJYRCIR555BHs3LnTaLAcTR/xmWeeQWhoKLZv347XX38d3d3djKjwWKC9vR3r169HV1cXvL29sWzZMuTk5CA5OdkpD+L+/n4UFxfD09NTT1HGGR2mANDR0cHsUuy5iNBNUQsEAmanrZuiJmlpb29vsxq4joDUb6dNm2aVPrChapKPj8+IzlpL6OjoQEtLC9LT010qUBpmRog+sUwmQ19fH+t55a6uLqxZswYvvPACVq5caec74LGQ8R8sCQsWLDAZLEfTR5w6dSrOnj2LqKgotLe3Y8GCBaiqqnLotVvLwMAA7r33XixcuBBPP/00ent7cfToUeTl5aGpqYnx5HRUTay/vx/l5eVMTdCYW4hIJLLa/NZSnKXzajjLSRplamtrXWbAf3BwEKWlpZz5dBLhc7LTdnd3Z+6bzcxoe3s7WltbXSJQ0jTNzCabM9gmNW2ZTIauri6TIvc9PT2455578MwzzzDjGDwuBS93BwCtra16HWyxsbGMqr5EImFShVFRUYxn5Fjg3//+NzZu3Ij77rsPABAcHIxNmzZh06ZN6OvrQ35+Pt58801UV1cznpyzZs2yS+Ak4uypqalMTcfd3R2RkZGIjIxkVuPNzc3o7+/nxPDXFETnta+vD+np6Q7vMvX29saECRMwYcIEJkV9+fJluLm5wc/PD/39/U7baQM/NzpxqaAkEAjg7+/PiF8QD9by8nLQNM00CBmr97W1taG9vR0ZGRlO7wimaVpP3s8cQqEQoaGhCA0N1UvNFxUV4euvvwYALF++HL///e+xbds2mwKlOastmqaxdetWFBQUwNfXFx9//DEyMzNZHctjOy4RLO+66y50dHSM+Pr/+3//D9nZ2WaPt0QfcSzx5JNPmvxeYGAg7rvvPtx3330YGhpCQUEBPvjgA5SXl2PBggXIzs7GLbfcwlmdqbGxERkZGSbF2d3c3JjUJKl7tbW14dq1a5yOZujqvLpCl6lAIEBHRwemTp2KiIgIdHZ2or6+3uGznASy+7e3obWPjw/i4uIQFxfHdNZev359RGdte3s72tvbnbKoMYQESh8fH0ycONHi4w1lFsPCwrB3715s3LgRAoEANTU1zEiXpX9vrVaLxx57TK+UlJWVpWe1deLECVRXV6O6uhqFhYX43e9+h8LCQlbH8tiOSwTLr776yqbjR9NHFIvFaG9vZ9Kw49GJ3NfXF2vXrsXatWuhUChw+vRpfPLJJ9i2bRt++ctfYvXq1Zg3b55V6a/m5mZIpVKLZuF09TxJGksikeD69euMFFtYWJjFgc7VdF4NxzEAjBC4b21tRWVlpV1nOQm9vb2orKy0SnfWFgw1azs7O9HY2Iju7m4IBAKn1dd1Ic4qpJ7MBVFRUSgsLMQf//hH5OTkoKCgAC+++CJqamqwZ88exmCZDZcuXcKkSZOYIL5hwwYcOXJEL+AdOXIEmzZtgkAgwNy5c9HT04P29nY0NDSYPZbHdlwiWNrKnDlzUF1djfr6esTExGDfvn3Yu3cvACArKwt79uzB9u3bsWfPHlY71bGMt7c3srKykJWVBZVKhTNnziA3NxdPP/005s6di5ycHNx+++1mm2FomkZtbS2GhoaQkZFh9QPeMI3V29sLiUSCmpoa+Pv7QywWs+q01Gq1KCsrQ3BwsEvUBBUKBaOgZEz8wM3Njanp6c5yXr9+3S46vboC7c70f3R3d4dYLGYkBidMmACZTIbq6mrOO2vZQtM0Kisr4eXlxVmgVCgU+NWvfoV7770XDz30EAQCATZu3IiNGzdCoVBYvDgYrZQ02mtaW1tZHctjOy4fLA8dOoQnnngCMpkMK1asQHp6Ok6dOoW2tjY8/PDDKCgogLu7O3bt2oWlS5cy+ohESHj79u1Yv349PvroI8TFxeHAgQNOviPH4enpiWXLlmHZsmVQq9X47rvvcODAAWzfvh2zZs1CdnY27rzzzhGpVYqiGCNgLj06DW22iF5tXV0dfHx8IBaLjc40qtVqxgGCK2FrWyD2Y2y7TA0XDIb3bUuHKfBzPTk9PZ1TgXZraW5uhlwuZ1KvIpFIr7OWq/tmAwmUHh4enHUoK5VKPPDAA1i5ciW2bNky4pzW/A3YlJJMvWa8lqFcDZcPlqtXrzZaNI+OjkZBQQHz7+XLl2P58uUjXhcWFoYzZ87Y9RrHAh4eHnqenOfOnUNubi5efPFFzJw5Ezk5Objrrrug1Wrx61//Gs8++yzS0tLs9qETCAQICgpCUFAQJk2aNMKfUiwWM7uykpISPZ1XZzIwMIDy8nKkpKRY1TxjeN9EbvCnn35iFGUskRuUy+Wora0dtZ7sSEigTEtL08tGGC6USGdtUVERE1DZdtayhcy8uru7c+Y3q1ar8dBDD+HOO+/E448/ztnng43VlqnXqFQqVjZdPLYxZkZH7A0b8YKqqirce++9zL/r6urw8ssvY9u2bdixYwf+9a9/ISIiAgDw6quvGg3ergZFUfjhhx+Qm5uLU6dOYXh4GIsXL8Yrr7ziME9OQ8iDVCKRYGhoCDExMUhMTHS6kwmpCc6cOdMueqrGZjkjIiJMplUd5WTClqamJnR1dVlssE06a2UyGTOKY6uSDk3TqKqqgkAgwJQpUzgJahqNBps3b0ZmZiaeffZZTheSbKy28vPzsWvXLhQUFKCwsBBPPvkkLl26xOpYnptoztLeWCpeoNVqERMTg8LCQsTHx2PHjh3w9/fH73//ewdeNXe0tLRgzZo12LBhAyQSCU6dOoWEhARkZWVh2bJldvXkNAYZf0hKSsLw8DCkUqlJMQBH0N3djaqqKofVBI3NcuoGEFeRjCM0Njaip6fH5g5lY5q1llptkY5pAJwFSq1Wi0ceeQRTpkzBiy++aJeMizmrLZqm8fjjj+PkyZPw9fXF7t27mblzY8fy6MEHS66wVLzg9OnTeOmll3D+/HkAGNPB8tq1a9iwYQPee+89xhyWoiiUlZUhNzcXBQUFEIvFyM7OxooVKzj15DRGT08PKisr9WY6gRu1IiK7R1EUZ2a/5iCpzvT0dKekOtVqNRNAFAoFvLy8oFQqMWvWLJcIlA0NDejt7eV8lMcazVoSKGma5kxuUKvV4oknnkBUVBReffVVvh44NuGDJVcEBwejp6eH+XdISAi6u7tNvp6kYx5//HEAN4Llxx9/jMDAQMyePRtvvvmmS2vQ6tLS0oK+vj6TreakSSI3NxfHjx9HcHAwY2ZN0s5cQVxVzO3gyA5EIpFArVbrye5xiavt4JqamtDa2go/Pz8MDg7aVfyBDfX19ejv78eMGTPsOvNKrLakUil6enqMdhTTNI3q6mpotVpMmzaNk98HRVHYtm0bgoKC8MYbbzh9rpfHavhgaQmjiR88+OCDrIOlSqVCdHQ0rl69yjSdSCQSxsz5+eefR3t7O/7973/b5T6cCXHUyM3NxdGjR+Hj44NVq1ZZ7MlpDGt1XtVqNeRyOSQSCRQKBcLDwyEWi21W0WltbUVHR4fD5fRM0dTUxBhsu7m5jRA9DwoKglgstusspy6OCpSGGLrD+Pj4ICIiAn19faAoirO5Toqi8Mwzz0AoFOLvf/87HyjHNnyw5ApL0rBHjhzBu+++i9OnTxv9fkNDA1auXIny8nJ7XrLTIbJzxJNTKBRi5cqVZj05jUHED2wNTCR1R5qDQkNDIRaLLdarNQxMzsZcqtOevpzGqKurw+DgIFJSUpwaRIgEXVVVFQYGBpj7trWzlqIoPP/88xgaGsL777/PB8qxDx8sueJ///d/ERYWxjT4dHV14S9/+YvR127YsAFLly7FQw89xHyNqAQBwFtvvYXCwkLs27fPIdfuCtA0jdbWVsaTU6VSMZ6cCQkJJgMVTdN6OxQuH+xEr1YqlaK/vx8hISEQi8WjpizJ9QwMDDh8x2Tqeurq6jA0NMQ6MOnOcpKdF5czjbW1tRgeHkZKSopL1O9qa2uhUCgwffp0pjHKls5amqbx8ssvQyqV4sMPP3SJxRKPzfDBkis6Ozuxfv16NDU1MeIFoaGheuIHwI32/gkTJqCurk6vQ/SBBx5AcXExBAIBEhIS8I9//MMlvB6dAU3TkEgkjCdnX18fVqxYgezsbEyePJl5wGq1WhQVFcHHxwfJycl2r3mZS1mSmpdarXYJOT2S8lapVFZfjzG3ELLzsrRZiag6EVNrZ/9+gNEDt0qlGuFRac4Vh6ZpvP7666ivr8eePXv4QDl+4IPlWIOtGXVCQgICAgLg5uYGd3d3XL582aLjXQm5XI7Dhw8jLy8PMpkMy5Ytw/Lly/HKK68gNTUVzz//vEMfvDRNM80i3d3djAybTCaDu7s7Z+MGtl7j9evXQVEUZ80qAJgxHDKKQ3Ze5sZhXDFQ6u64zV2PVquFXC6HTCZjsgwREREjFktvvfUWSktLsXfvXpeoU/NwBh8sxxps5zkTEhJw+fJlhIeHW3W8q9Ld3c0oB02YMAELFixATk6O0xxEaJpGT08PKioqoNFomFRteHi403YVpPvYzc3NroFbqVQyIykajYbRsjXsKCY7XLVa7RKi6AD0UuWWXo9uZ219fT12796NrKwsyOVyXL58GV988QUnIg9sFrbNzc3YtGkTOjo6IBQKsWXLFmzduhUAxqzQiYvCB8uxBttGIlPBciybWQM3VHBWr16NjRs3Yu3atTh+/Djy8vJQU1ODRYsWIScnB5mZmQ4LnESgPSQkBHFxcejv72dSlo7SL9WF2EgRdwxHBSbDWU5is+Xv74+amhpOxzFshXiZclFT1mq1+Pbbb/H+++/jhx9+wO2334577rmHk3liNgtbYmGWmZmJ/v5+zJo1C4cPH8b06dPH9Oy2C8IHy7EG23nOxMREhISEQCAQ4JFHHsGWLVssOt4VoWkaK1euxMMPPzxC73dgYAAnTpxAbm4uKioqmB3nL37xC7vt8DQaDUpKSiAWixEbGzvi+0SvViaTwdPTk6n12UtajqIolJeXMwbLzoKkLEmDkJeXF6ZOncq8H51JY2Mjent7OWu+omkaH3/8MY4dO4bDhw+jvr4ehw4dwvHjxzFr1iy88847Vp/bmoVtdnY2Hn/8cSxevJgPltzCB0tXhIt5zra2NkRHR0MqlWLx4sV45513MH/+/DEdLIEbDVLmFHcUCgVOnTqF3NxcFBUV4bbbbkNOTo7VnpzGUKvVKC4uxoQJExAZGcnqukmtT9fomitFH4qiUFpa6jIWZERblaZphIeHQyaTMY1RIpEIoaGhDk+bcyWpp8unn36K/fv349ixYyPel8PDwzZJG1r6WW1oaMD8+fNRXl6OwMDAMS104oLwwXKsYc1qU3eFOdbTsJaiVCoZT87CwkLceuutjCentalRpVKJkpISJCYmWqVApFAomMAJgHWTjCm0Wi1KS0sRHh6u50voLEyJkJP6rlQqRVdXF/z9/Zk0tb3ru9aKtI/GF198gT179iA/P99q0XauhE4GBgZwxx134LnnnsOaNWsA3DxCJw6CD5ZjDTbznIODg6AoCgEBARgcHMTixYvxwgsv4O6777ZoHnS8oVar8e233yI3Nxfff/89Zs+ejezsbCxcuJD1Dm94eBglJSWYMmUKJxq3unq1ZK5PLBaz1qvVarUoLi5GZGSkS3h1ElsrNzc3vTEfY68zrO+SBiGu67vNzc2MQARXgfLgwYP4xz/+gfz8fAQGBnJyTkPYLmzVajVWrlyJpUuX4qmnnjJ6rptF6MSO8MFyrMFmnrOuro6p6Wk0Gtx///2Mi4Cp4282NBoN48l59uxZpKamIjs7G3fddZfJHd7g4CDKysqQnJxsFxcVXccMlUrFyO75+fkZDToajQbFxcWIiYlxiZlcXaNkS/0fSX3X1llOQ1paWiCTyUb4Y9rC8ePH8fbbbyM/P9+uaU02C1uapvHggw8iNDQUb7/9tt73bnahE47hgyUPj1arZTw5z5w5gylTpmD16tVYsmQJk167dOkSJBIJFixYYJVps6VoNBomcA4PDyMsLAxisRgBAQEQCARMzTQuLs4lTK1pmkZFRQW8vLxs7sIls5wymQw0TTP1XUvT1CRQcik5eOrUKbz++usoKChAWFgYJ+c0BZuF8blz53D77bfr1WHJiAgvdMIpfLDkMc7NOuNFURSuXLmCAwcO4PTp00hISEBqair27t2Lzz77DDNnznT4Nel2lw4MDCA4OBjd3d2YNGkSRCKRw6/HEC4DpSGGs5xkJMXUbpvQ2toKiUSCtLQ0zgLl119/jZdeegkFBQWcu+XwuDx8sOQxDj/jdSNwfvjhh9ixYwfi4uIQFhbGeHI6q6tweHgYV65cga+vL5RKJSuPRntC5jp9fHyQlJRk159F3GF0d9vG5Ofa2toYtxeuAuV3332H5557Dvn5+ay6n3nGHTYHS17PaZxy5MgRnD17FgDw4IMPYsGCBSOCZVRUFJPWCQgIQHJyMlpbW036Wo418vPz8e9//xtFRUUQiUSoqKhAbm4ucnJyEBoaynhyGgo/2AuFQoGSkhJMnz4doaGhjJKMRCJBVVWVw8cyKIrC1atX4efn55C5Tg8PD+Y9R0Tum5ubGfk5kUiE4eFhdHR0ID09nbNAeeHCBTz77LM4fvw4Hyh5rIbfWY5TbvYZr6GhIWzcuBEfffTRiOsmgunEk9PX1xdZWVnIysqy2ZPTFKQLd9q0aQgODh7xfcOxDHtbbBEBhICAACQmJnJ+fkuvpauri1HmEYlEiIyM5GTRcOnSJWzbtg1Hjx5FXFwcR1fMMwbh07A3M/yMl+0QSy7iyenm5oZVq1YhJycH0dHRnARO0oU7ffp0VmMKhhZbvr6+zDwjF6IMJFAGBga6hAACcMP4u6WlBenp6cxIiu4sZ1hYmMX3/tNPP+Gxxx7D4cOHnb4g4HE6fLDkMQ4/42U5NE2jpaWF8eRUq9VYtWoVsrOzER8fb1XgHBgYQFlZGWbMmGFVFy4xN5ZIJJDL5fD29mbGMqyZZ6QoCmVlZQgKCnKZQCmRSNDc3Iz09HS9gGg4y2nJvZeWlmLLli3Iy8vD5MmT7X0LPK4PHyx5jMPPeNkGTdPo6OhgPDkHBgYYT062M4j9/f0oLy9Hamqq1QoxhhBvSmInRsYy2OjVkkAZHByM+Ph4Tq7HViQSCZqampCRkWF256h776NJDlZUVOChhx7C/v37kZycbM/L5xk78MGSxzj8jBe3yGQyxpOzs7MTy5YtQ1ZWlknLqt7eXlRWViI1NZW1mo+lDA8PQyKRQCaTQSgUMrsub2/vEa8l2rOhoaEuU7uTSqVobGxkFSgNGR4eZkZSKIrC119/jaysLLi5uWHTpk3Yu3cvZ2NCN6MH7TiED5Y8PI6mq6sLR48eRV5eHlpaWrB06VLk5OQwThinTp1Cc3MzfvWrX9kkxG0JRK9WJpOBoig9IQCKolBSUuIy2rPAz4EyPT3dZnm8oaEh/OMf/8CRI0dQX1+PdevW4bHHHuPMpPpm96AdJ/DBksc5nDx5Elu3boVWq8XDDz+M7du3632fpmls3boVBQUF8PX1xccff4zMzExWx44lent7cfz4cRw8eBA1NTVITU3FDz/8gEOHDjnNZkulUjF6tWq1GhqNBpGRkXafo2SLTCZDfX09MjIyONORbWxsxIYNG/C3v/0NbW1tOHjwIGpra7F8+XL86U9/silo3uwetOMEPljyOB6tVospU6bgyy+/RGxsLObMmYPPP/9cbz6zoKAA77zzDgoKClBYWIitW7eisLCQ1bFjlYMHD+IPf/gD0tPTce3aNSxcuBA5OTmYM2eO3V05jKHValFUVARvb29oNBoolUo9U2dneFPaI1C2tLRg/fr1+OCDDzB37lzm60NDQ7h8+TLmz59v0/lvZg/acQQvSsDjeC5duoRJkyYxO6cNGzbgyJEjegHvyJEj2LRpEwQCAebOnYuenh60t7ejoaHB7LFjkcOHD2Pnzp344YcfEBYWhuHhYZw+fRq7d+/Gk08+idtvvx05OTm49dZbOfPkHA2tVouSkhJERUUxbiYajQZyuRz19fUYGhoyqaBjL8jP5iL1Smhvb8e9996Lv//973qBEgB8fX1ZB8rRxrDYcv78eT0P2mnTptkcqHlcBz5Y8lhMa2urXu0rNjYWhYWFZl/T2trK6tixBkVROHPmDAoKChjBAR8fH2RnZyM7OxtKpRJfffUV9u3bh6eeegrz5s1DTk4ObrvtNs7trICfbb+ioqIQHR3NfN3d3R2RkZGIjIwcoaATGhrKyO7ZI3B2dnaitrYWGRkZrDp32SCRSLBu3Tq8+eabNgelr776yuT3xGIx0x3e3t5uUs+X/K5FIhFWr16NS5cuYf78+ayP53FtHC9GyTPmMZa6N3zAmnoNm2PHGkKhEO+8845RZR4A8PLywooVK7B7924UFxdj/fr1OHr0KObNm4dHH30Up06dglKp5ORaTAVKQ8joxYwZM3DLLbcgLCwMbW1t+OGHH1BZWYnOzk5QFMXJNXV2dqKmpobTQCmXy7Fu3Tq89tpruPPOOzk5pymysrKwZ88eAMCePXuQnZ094jWDg4Po7+9n/v/06dOYd4/csQAADnJJREFUMWMG6+N5XB9+Z8ljMbGxsWhubmb+3dLSMuLBbOo1KpXK7LHjGQ8PDyxevBiLFy9mPDkPHDiA559/HmlpacjOzsaiRYus6qK11h9TKBQiPDwc4eHhoCiKkd27fv06AgMDGQUda6Tnurq6OA+UXV1dWLt2LXbs2IGlS5dycs7R2L59O9avX4+PPvqIGcMCoDeGJZFIRnjQ3n333aMezzO24Bt8eCxGo9FgypQpOHPmDGJiYjBnzhzs3bsXKSkpzGvy8/Oxa9cupsHnySefxKVLl1gdezOi1Wpx8eJF5OXl4auvvsK0adOQk5Oj58k5GiRQxsbGciYWTtM0ent7IZFIGOk5sVjMWq+2q6sL169fR0ZGhs0m0ISenh7cc889eOaZZ5jgxMPDAr4blsc5FBQUYNu2bdBqtdi8eTOee+45fPDBBwCA3/72t6BpGo8//jhOnjwJX19f7N69G7NnzzZ5LM/PUBSFy5cvIzc3F6dOncLEiRORnZ2Nu+++26i2rEajQVFRkV2NpIn0nEQiQWdnJ3x8fCAWi03q1XZ3d6OqqorTQNnX14e1a9fiySefxPr16zk5J89NAx8seXjGM0R558CBAygoKEBMTAzjyRkcHAy5XI4//elPePbZZ+0WKA2haRqDg4OMXq2npyfEYjGj2drT04Nr164hPT3dqJqQNQwMDGDdunX4zW9+g40bN3JyTp6bCj5Y8ox/zIkYfPbZZ4wiir+/P95//32kpaUBMC1BNhYhRs25ubnIz89HQEAA2tra8Mgjj+CRRx5x2nXparbSNA2lUon09HRWDitsGBoawvr16/HAAw/goYce4uScPDcdfLDkGd+wETG4cOECkpOTERISghMnTmDHjh3MOIopVZWxTmdnJxYvXowZM2bg+vXr8PPzYzw5RSKRUzqMe3t7cfXqVYjFYnR1dUEgEDCye9buMBUKBTZs2IB77rkHW7ZsGfOd0zxOgxcl4BnfsBFAmDdvHvP/c+fORUtLi8Ov05F0dnZi5cqVeOWVV7BixQrQNI26ujrk5eVh48aNcHd3Zzw5o6KiHBJgiHB8RkYGfHx8kJSUBKVSCalUiqtXr4KiKEREREAkErEWllcqlXjggQewatUqPlDyOB1+zpLHpTElbmCKjz76CMuWLWP+LRAIsGTJEsyaNQv//Oc/7XqtjuKbb77BCy+8gBUrVgC4cY9JSUl45plncO7cOfznP/+Bu7s7Nm/ejCVLluBvf/sbmpqajM64ckFfXx8qKyuRlpamN/Li5eWFCRMmYNasWUhLS4OHhweqqqpw6dIl1NXVYWBgwOQ5VSoVHnroIdx55514/PHH+UDJ43T4YMnj0lgiYvDNN9/go48+0nN0OH/+PH766SecOHEC7777Lr777ju7XaujWLt2rd6CQBeBQIAJEyZg27Zt+Pbbb5Gbm4vAwEA89thjWLRoEd58803U1NRwFjj7+vpw9erVEYHSEE9PT8TExCAjI4PZfdbW1uKHH35ATU0N+vr6GBEEjUaDhx9+GLfccgueeuopTgJlV1cXFi9ejMmTJ2Px4sVGtVmrqqqQnp7O/BcYGMj4vO7YsQMxMTHM9woKCmy+Jp6xBV+z5HFpLl68iB07duDUqVMAgNdeew0A8Oyzz+q9rrS0FKtXr8aJEycwZcoUo+fasWMH/P398fvf/96+F+2iyGQyHDp0CHl5eejq6sLy5cuRlZWFadOmWRWQiLl1Wlqa1Z6dWq0WcrkcpaWlePrppzFv3jx0d3cjMzMTO3bs4GxHaalNllarRUxMDAoLCxEfH3/Tv3fGATa/kfidJY9LM2fOHFRXV6O+vh4qlQr79u1DVlaW3muampqwZs0afPrpp3qBcjQJspuRiIgIbNmyBadOncKpU6eQkJCAF198EbfffjteeeUVlJWVsZa44yJQAjdk98RiMRYvXoxz585hYGAAzc3NOHToELZu3Ypvv/0WWq3W6vMTjhw5ggcffBAA8OCDD+Lw4cOjvv7MmTNISkpCfHy8zT+bZ3zA7yx5XB5zAggPP/ww8vLymAcbGRGpq6sbIUHGCyCMpLe3F8eOHcPBgwdRV1eHu+66Czk5OUhPTzcqcTcwMICysjKkpqayUhdiA0VR2LZtG4KCgvDGG29Ao9Hgm2++wcGDB9HX14fPP//cpvNbapO1efNmZGZm4vHHHwdwIyvx8ccfIzAwELNnz8abb76JkJAQm66Jx6HwoyM8PDzc0d/fj4KCAuTl5Y3w5BQKhbhy5Qq++eYb/O53v+M0UD7zzDNwc3PD3/72N6s0aIHRbbYefPBB1sFSpVIhOjqaGYMBbjichIeHQyAQ4Pnnn0d7ezv+/e9/W3WdPE6BD5Y8PFxiTgDh7NmzyM7ORmJiIgBgzZo1eOGFF1gdO9YYHh7GqVOnkJubi+LiYmRmZuLcuXP4+OOPGelCW6EoCv/3f/8HhUKB9957z+pAaY6pU6fi7NmzjE3WggULUFVVZfS1R44cwbvvvovTp08b/X5DQwNWrlyJ8vJyu1wrj13ga5Y8PFyh1Wrx2GOP4cSJE6ioqMDnn3+OioqKEa+7/fbbUVxcjOLiYiZQsj12LOHj44OcnBz85z//wb59+3Du3DnMmjULjz76KLZu3YpvvvkGarXa6vPTNI2XX34Zvb29ePfdd+0WKAHLbLI+//xz3HfffXpfa29vZ/7/0KFDN3Xt+2aFD5Y8PP9FVwDB09OTEUCw97GuTnV1NTZu3IijR4/iwIEDKCoqwrp163DkyBHMmzcPjz32GE6fPg2VSsX6nDRN47XXXkNbWxs+/PBDVi4mtrB9+3Z8+eWXmDx5Mr788ktm19/W1obly5czrxsaGsKXX36JNWvW6B3/zDPPYObMmUhNTcU333yDt956y67Xy+N68Ao+PDz/xZgAApHN0+XixYtIS0tDdHQ0du7ciZSUFNbHjkV2796NTz75hNlNeXh4YMmSJViyZAk0Gg2+//57HDhwAP/3f/+H9PR0xpPTlMQdTdP461//iurqanz22Wd2D5QAEBYWhjNnzoz4enR0tN7MpK+vLzo7O0e87tNPP7Xr9fG4Pnyw5OH5L2wEEDIzM9HY2Ah/f38UFBQgJycH1dXVFoknjDVeffVVk99zd3fHwoULsXDhQmi1Wly4cAF5eXl4+eWXkZyczHhykvESmqaxa9cu/PTTT9i/f79Rey8eHleEf6fy8PyX2NhYNDc3M/9uaWlBdHS03mt0nTSWL1+ORx99FHK5nNWx4x03NzfcfvvtuP3220FRFH788Ufk5ubiz3/+M5KSkpCdnY329nacP38eubm58PDwcPYl8/Cwhu+G5eH5LxqNBlOmTMGZM2cQExODOXPmYO/evUhJSWFe09HRAbFYDIFAgEuXLmHt2rVobGxk3FFGO/ZmhaIolJSU4NNPP8WJEydQVFTEmc8lDw9LeNcRHh6ucHd3x65du7B06VJGACElJUVPACE3Nxfvv/8+3N3d4ePjg3379kEgEJg8lgcQCoWMJuxf//pXZ18OD49V8DtLHh4eHp7xDj9nycMzVjl58iSmTp2KSZMm4fXXXx/x/TfeeINxuZgxYwbc3NzQ1dUF4Iap9cyZM5Gens6ZQAAPD49p+J0lD48TIDXOL7/8ErGxsZgzZw4+//xzPVNrXY4dO4a33noLX3/9NYAbwfLy5csIDw935GXz8IxV+J0lD89YxFIRA2OqMjcbBw4cQEpKCoRCIS5fvmzydaZ27Gw8LXl4TMEHSx4eJ2BMxKC1tdXoa4eGhnDy5Encc889zNcEAgGWLFmCWbNm4Z///Kfdr9cVmDFjBg4ePIj58+ebfM1osoOvv/46Fi1ahOrqaixatMho6puHxxR8NywPjxOwRMTg2LFj+OUvf4nQ0FDma+fPn0d0dDSkUikWL16MadOmjRpExgPJyclmX6O7YwfA7NinT5+OI0eO4OzZswBueFouWLBgVANoHh5d+J0lD48TsETEYN++fSNSsOS1IpEIq1evxqVLl+x3sWOI0XbsEokEUVFRAICoqChIpVKnXCPP2IQPljw8TmDOnDmorq5GfX09VCoV9u3bh6ysrBGv6+3txbfffqvnkjE4OIj+/n7m/0+fPj1uXDDuuusuzJgxY8R/bEXpx7PsII9z4dOwPDxOgI0AAnDDDmrJkiV6RssSiQSrV68GcEN16P7778fdd9/t+JuwA1999ZVNx4+2YxeLxWhvb2c8LUUikU0/i+fmgh8d4eHhGVMsWLAAO3fuNDpfOppk4f/+7/8iLCwM27dvx+uvv46uri785S9/ccId8DgBfnSEh4eHHZs3b4ZIJDKZsqVpGk8++SQmTZqE1NRU/PTTT8z3zAkoOIJDhw4hNjYWFy9exIoVK7B06VIA+p6Uujv25ORkrF+/npEdNOVpycPDBn5nycNzk/Ddd9/B398fmzZtQnl5+YjvFxQU4J133kFBQQEKCwuxdetWFBYWWiygwMPjgvA7Sx4eHnbMnz9fb/zEkCNHjmDTpk0QCASYO3cuenp60N7ebrGAAg/PeIQPljw8PABMj11YIqDAwzNe4YMlDw8PANNjF/w4Bg8PPzrCw8PzX0yNXahUKtYCCjw84xV+Z8nDwwMAyMrKwieffAKapvHDDz8gKCgIUVFRrAUUeHjGM/zOkofnJuG+++7D2bNnIZfLERsbi5deeglqtRrADRGE5cuXo6CgAJMmTYKvry92794NwLSAAg/PzQQ/OsLDw8PDM97hR0d4eHh4eHjsDR8seXh4eHh4zMAHSx4eHh4eHjPwwZKHh4eHh8cMfLDk4eHh4eExAx8seXh4eHh4zMAHSx4eHh4eHjOYEyXgBSB5eHh4eG56+J0lDw8PDw+PGfhgycPDw8PDYwY+WPLw8PDw8JiBD5Y8PDw8PDxm4IMlDw8PDw+PGfhgycPDw8PDY4b/D96waUEHx0AcAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# All generated points plotted in 3D\n",
    "fig = plt.figure(figsize=(12,8))\n",
    "ax = plt.axes(projection=\"3d\")\n",
    "ax.scatter3D(x, y, z, c ='red',s=150)\n",
    "ax.scatter3D(spherical_points[:,:3][:,0], spherical_points[:,:3][:,1], spherical_points[:,:3][:,2], c=spherical_points[:,:3][:,2], cmap='cividis',s=150);"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "989102ed",
   "metadata": {},
   "source": [
    "# K-d trees for spatial neighborhood search\n",
    "With the point cloud I can create from the original data, I store it into a large k-d tree to speed up my spatial search"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "6745a044",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Loading datasets from all regions\n",
    "roi_list = ['ME(R)']\n",
    "path = 'neuron_regions_points/'\n",
    "points_df_dict = {}\n",
    "for roi in roi_list:\n",
    "    skeleton_df = pd.read_csv(path+ roi + '.csv')\n",
    "    points_df_dict[roi] = points_df[['x','y','z','radius','bodyId']].astype('float')\n",
    "    points_df_dict[roi] = points_df_dict[roi][pd.notnull(points_df.bodyId)]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "34795071",
   "metadata": {},
   "source": [
    "Now I introduce a function that creates a point cloud and stores into a k-d tree, that I will query later"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "3a880042",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading: ME(R)\n",
      "Point cloud construction time: 96.94\n",
      "K-d tree construction time: 8.01\n"
     ]
    }
   ],
   "source": [
    "def dataframe_to_bodyid_kd_tree(points_df):\n",
    "    start_pc = timeit.default_timer()\n",
    "    # Convert pandas dataframe to numpy (for speed) and apply spherical sample function (introduced above)\n",
    "    points_bodyid_numpy = points_df.to_numpy()\n",
    "    points_bodyid_numpy = np.apply_along_axis(spherical_sample_dataframe_row,1,points_bodyid_numpy )\n",
    "    points_bodyid_numpy = np.concatenate(points_bodyid_numpy)\n",
    "    end_pc = timeit.default_timer()\n",
    "    print('Point cloud construction time:',np.round(end_pc - start_pc,2))\n",
    "    start_tree =  timeit.default_timer()\n",
    "    # Since my data is sorted by neuron bodyId use np.unique function to find the first time each unique bodyId occurs\n",
    "    # I obtain two arrays that store this information, but I will only need the second one\n",
    "    dummy_body_id_array, occurence_start_array = np.unique(points_bodyid_numpy[:,3],return_index = True)\n",
    "    occurence_start_array = np.sort(occurence_start_array)\n",
    "    # I load the created point cloud into the k-d tree\n",
    "    big_kd_tree = KDTree(points_bodyid_numpy[:,:3],compact_nodes=True,balanced_tree=True,leafsize = 400)\n",
    "    end_tree = timeit.default_timer()\n",
    "    print('K-d tree construction time:',np.round(end_tree-start_tree,2))\n",
    "    \n",
    "    return big_kd_tree,occurence_start_array,points_bodyid_numpy\n",
    "\n",
    "for region_key in roi_list:\n",
    "    print(\"Loading:\",region_key)\n",
    "    spherical = True\n",
    "    points_df = points_df_dict[region_key]\n",
    "    # Removing null values\n",
    "    points_df = points_df[pd.notnull(points_df.bodyId)]\n",
    "    big_kd_tree,occurence_start_array,points_bodyid_numpy = dataframe_to_bodyid_kd_tree(points_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "a527643d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[9.19200000e+03, 1.75520000e+04, 9.11200000e+03, 5.43702186e+08],\n",
       "       [9.19200000e+03, 1.75520000e+04, 9.08800000e+03, 5.43702186e+08]])"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Points and bodyids are loaded in the efficient numpy array \n",
    "# The first three columns are coordinates x,y,z, while the last one is bodyd)\n",
    "points_bodyid_numpy[0:2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "efd6478b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([     0,  43008, 223986, 352996, 586992], dtype=int64)"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Occurences of different bodyIds (sorted in ascending order, but the order doesn't correspond to the previous array)\n",
    "occurence_start_array[0:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "dd8cf25f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([list([0, 6, 7, 8, 9]), list([1, 10, 11, 12, 13])], dtype=object)"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# K-d tree in which the entire dataset in stored (as a point cloud) which doesn't have labels included\n",
    "# It can be queried with any points and radius. For example, I choose two points two query\n",
    "query_points = points_bodyid_numpy[:2,:-1]\n",
    "radius = 100\n",
    "within_radius_indices = big_kd_tree.query_ball_point(query_points,r=10,return_sorted=True)\n",
    "#It returns sorted indices of points that are within radius \"r\", where each list corresponds a queried point\n",
    "within_radius_indices"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "b0c9ff80",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 9192.        , 17552.        ,  9112.        ],\n",
       "       [ 9198.        , 17558.        ,  9108.48528137],\n",
       "       [ 9198.        , 17546.        ,  9108.48528137],\n",
       "       [ 9186.        , 17558.        ,  9108.48528137],\n",
       "       [ 9186.        , 17546.        ,  9108.48528137]])"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# For example, our first querying point, points_bodyid_numpy[0,:-1] is within radius with the 5 following points\n",
    "points_bodyid_numpy[within_radius_indices [0],:-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "5869f988",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0,  6,  7,  8,  9,  1, 10, 11, 12, 13], dtype=int64)"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# I also use numpy.fromiter function to create a single list from the results (this is used later on)\n",
    "within_radius_indices_single = np.fromiter(itertools.chain.from_iterable(within_radius_indices), np.int64)\n",
    "within_radius_indices_single"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "395b1f52",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[9.19200000e+03, 1.75520000e+04, 9.11200000e+03, 5.43702186e+08],\n",
       "       [9.19800000e+03, 1.75580000e+04, 9.10848528e+03, 5.43702186e+08],\n",
       "       [9.19800000e+03, 1.75460000e+04, 9.10848528e+03, 5.43702186e+08],\n",
       "       [9.18600000e+03, 1.75580000e+04, 9.10848528e+03, 5.43702186e+08],\n",
       "       [9.18600000e+03, 1.75460000e+04, 9.10848528e+03, 5.43702186e+08],\n",
       "       [9.19200000e+03, 1.75520000e+04, 9.08800000e+03, 5.43702186e+08],\n",
       "       [9.19800000e+03, 1.75580000e+04, 9.09151472e+03, 5.43702186e+08],\n",
       "       [9.19800000e+03, 1.75460000e+04, 9.09151472e+03, 5.43702186e+08],\n",
       "       [9.18600000e+03, 1.75580000e+04, 9.09151472e+03, 5.43702186e+08],\n",
       "       [9.18600000e+03, 1.75460000e+04, 9.09151472e+03, 5.43702186e+08]])"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# This is useful to filter through all points at once (In this example, I queried two points at once)\n",
    "points_bodyid_numpy[within_radius_indices_single,:]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "90a28f55",
   "metadata": {},
   "source": [
    "## Querying labeled k-d trees\n",
    "Now that I have the k-d tree and the supporting arrays, I want to query them for each neuron (a complex 3D object), and find which other neurons (also complex 3D objects) are spatially close (within certain radius) of each other."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "a0f6ce4e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# I create function that takes the k-d tree, occurence start array and points with bodyids, along with the required query radius\n",
    "def query_tree_with_neuron_bodyid(big_kd_tree,occurence_start_array,points_numpy,radius):\n",
    "    query_results_dictionary = {}\n",
    "    # Going through each unique occurence\n",
    "    for i in range(0,occurence_start_array.shape[0]):\n",
    "            # Recording the bodyId (label) of a specific occurence\n",
    "            body_id = int(points_numpy[:,3][occurence_start_array[i]])\n",
    "            # If the index is not the last index in the array\n",
    "            if i <occurence_start_array.shape[0]-1:\n",
    "                # Identify starting and ending indices of a bodyId (label) in the points_bodyid numpy array\n",
    "                interval_start_index = occurence_start_array[i]\n",
    "                interval_end_index = occurence_start_array[i+1]\n",
    "                results = big_kd_tree.query_ball_point(points_numpy[interval_start_index :interval_end_index, :3], r = radius,return_sorted=True)\n",
    "                # I create a single list of results, where indices of points are within one single list\n",
    "                results_array = np.fromiter(itertools.chain.from_iterable(results), np.int64)\n",
    "            # Do the same procedure, but for the last index in the occurence array\n",
    "            else:\n",
    "                interval_start_index = occurence_start_array[i]\n",
    "                results = big_kd_tree.query_ball_point(points_numpy[interval_start_index :, :3], r = radius,return_sorted=True)\n",
    "                results_array = np.fromiter(itertools.chain.from_iterable(results), np.int64)\n",
    "            unique_results = np.unique(results_array)\n",
    "            query_results_dictionary[body_id] = []\n",
    "            # Going through occurences again to check for each neuron if it's in the neighborhood\n",
    "            for j in range(0,occurence_start_array.shape[0]):\n",
    "                    # Identify the bodyid (label) of the neighbor\n",
    "                    body_id_neighbor =  int(points_numpy[:,3][occurence_start_array[j]])\n",
    "                    # Exclude the neuron bodyid (label) from the outer loop \n",
    "                    if body_id != body_id_neighbor:\n",
    "                        # If the labeled object is not the last one in the array\n",
    "                        if j <occurence_start_array.shape[0]-1:\n",
    "                            interval_start_n_index = occurence_start_array[j]\n",
    "                            interval_end_n_index = occurence_start_array[j+1]\n",
    "                            # Create a mask boolean that check indices if the indices of query results\n",
    "                            # are between starting and ending intervals of a bodyId (label) in the occurence array\n",
    "                            mask_inverted = np.logical_and((interval_start_n_index  <= unique_results),(unique_results < interval_end_n_index))\n",
    "                        else:\n",
    "                            interval_start_index = occurence_start_array[-1]\n",
    "                            mask_inverted = np.logical_and((interval_start_n_index  <= unique_results),True)\n",
    "                        #If the neighboring neuron (object) with this bodyid (label) is present within the spatial search radius\n",
    "                        if mask_inverted.any():\n",
    "                            # Add the label to the dictionary associated with the original body_id (index i) from the outer loop\n",
    "                            query_results_dictionary[body_id].append(body_id_neighbor)\n",
    "                        # Create a mask to remove all points in the query results associated with the neighbors bodyId (label)\n",
    "                        # I only need to detect one occurence of two unique objects being close (at a distance less than radius)\n",
    "                        mask = np.logical_not(mask_inverted) \n",
    "                        unique_results = unique_results[mask]\n",
    "                        # If the array is emptied, stop the process\n",
    "                        if unique_results.size == 0:\n",
    "                            break\n",
    "    return query_results_dictionary"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "466c2aac",
   "metadata": {},
   "source": [
    "The labeled k-d tree querying algorithm is used to detect all unique pairs objects that are below the threshold (querying radius) distance. These pairs are represented as an edge in graph (in the research paper its named \"metagraph\"):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "c06a4e81",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Region: ME(R)\n",
      "Point cloud construction time: 57.9\n",
      "K-d tree construction time: 4.72\n",
      "Querying radius 0.0\n",
      "Total time taken for querying radius: 251.29\n"
     ]
    }
   ],
   "source": [
    "save_path = 'querying_results/'\n",
    "for region in roi_list:\n",
    "    print('Region:',region)\n",
    "    points_df = points_df_dict[region]\n",
    "    big_kd_tree,occurence_start_array,points_numpy = dataframe_to_bodyid_kd_tree(points_df)\n",
    "    # Querying radii were chosen based on the datasets and the research needs. \n",
    "    # The idea is to capture the distances starting from the resolution limit to the several orders of magnitudes higher\n",
    "    step = 0.01\n",
    "    radius_list = list(np.arange(0,1 + step ,0.01)) + [1, 2, 3, 4, 5, 10, 15, 20, 25, 30, 40, 45, 50, 60, 70, 80, 90, 100, 150,200]\n",
    "    for radius in radius_list:\n",
    "        start_total = timeit.default_timer()\n",
    "        print('Querying radius',radius)\n",
    "        query_results_dictionary = query_tree_with_neuron_bodyid(big_kd_tree,occurence_start_array,points_numpy,radius)\n",
    "        end_total = timeit.default_timer()\n",
    "        print('Total time taken for querying radius:',np.round(end_total-start_total,2))\n",
    "        with open(path + region+\"_querying_results_radius_\" + str(radius)+\".pkl\", \"wb\") as h:\n",
    "                    pickle.dump( query_results_dictionary, h)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "75f7cce7",
   "metadata": {},
   "source": [
    "All querying results (for each queried radius) are saved in the pickle format, are going to be used to produce specific results.\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}