自適應增強分類器(Adaptive Boost Classifier).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 自適應增強分類(Adaptive Boost Classifier)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### 先引入我們需要的packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os \n",
    "import numpy as np\n",
    "import random\n",
    "import matplotlib.pyplot as plt\n",
    "from random import choices"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x1296 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x=np.zeros((100,2))\n",
    "y=np.zeros(100)\n",
    "for i in range(0,100): \n",
    "    f=-0.1\n",
    "    if(i<50):\n",
    "        y[i]=1\n",
    "        x[i,0]=random.random()+0.3\n",
    "        x[i,1]=random.random()+0.3\n",
    "    else:\n",
    "        y[i]=-1\n",
    "        x[i,0]=random.random()\n",
    "        x[i,1]=random.random()\n",
    "plt.rcParams[\"figure.figsize\"] = (18,18)\n",
    "plt.plot(x[:50,0],x[:50,1],\"bo\",markersize=20)\n",
    "plt.plot(x[50:,0],x[50:,1],\"rX\",markersize=20)\n",
    "plt.xticks(fontsize=30)\n",
    "plt.yticks(fontsize=30)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Testing Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "N_test=20\n",
    "x_test=np.zeros((N_test,2))\n",
    "y_test=np.zeros(N_test)\n",
    "for i in range(0,N_test): \n",
    "    f=-0.1\n",
    "    if(i<N_test/2):\n",
    "        y_test[i]=1\n",
    "        x_test[i,0]=random.random()+0.3\n",
    "        x_test[i,1]=random.random()+0.3\n",
    "    else:\n",
    "        y_test[i]=-1\n",
    "        x_test[i,0]=random.random()\n",
    "        x_test[i,1]=random.random()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Decision Stump"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def Decision_Stump(x,y,D):\n",
    "    # comp_x0: 用來放橫軸加權錯誤率候選值\n",
    "    # comp_x1: 用來放縱軸加權錯誤率候選值\n",
    "    comp_x0=np.zeros(len(y)-1)\n",
    "    comp_x1=np.zeros(len(y)-1)\n",
    "    # x0_sort: 所有資料點橫軸座標排序\n",
    "    # x1_sort: 所有資料點縱軸座標排序\n",
    "    x0_sort=np.sort(x[:,0])\n",
    "    x1_sort=np.sort(x[:,1])\n",
    "    # x0_sort_index:所有資料點橫軸座標排序後所對應的index\n",
    "    # x1_sort_index:所有資料點縱軸座標排序後所對應的index\n",
    "    x0_sort_index=np.argsort(x[:,0])\n",
    "    x1_sort_index=np.argsort(x[:,1])\n",
    "    # 這裡的演算法核心就是我們以排列好的橫軸和縱標\n",
    "    # 從第一個資料開始分成兩半，然後並計算impurity值\n",
    "    for i in range(1,len(y)):\n",
    "        # 分成左右subtree\n",
    "        s_left,s_right=x0_sort_index[:i],x0_sort_index[i:]\n",
    "        # 檢查分到的兩邊為1或為-1\n",
    "        if(np.count_nonzero(y[s_left]==1)/len(y[s_left])>=0.5):\n",
    "            class_left,class_right=1,-1\n",
    "        elif(np.count_nonzero(y[s_right]==1)/len(y[s_right])>=0.5):\n",
    "            class_left,class_right=-1,1\n",
    "        comp_x0[i-1]=sum(D[s_left]*(y[s_left]!=class_left))+sum(D[s_right]*(y[s_right]!=class_right))\n",
    "        # 分成上下subtree\n",
    "        s_left,s_right=x1_sort_index[:i],x1_sort_index[i:]\n",
    "        # 檢查分到的兩邊為1或為-1\n",
    "        if(np.count_nonzero(y[s_left]==1)/len(y[s_left])>=0.5):\n",
    "            class_left,class_right=1,-1\n",
    "        elif(np.count_nonzero(y[s_right]==1)/len(y[s_right])>=0.5):\n",
    "            class_left,class_right=-1,1\n",
    "        comp_x1[i-1]=sum(D[s_left]*(y[s_left]!=class_left))+sum(D[s_right]*(y[s_right]!=class_right))\n",
    "    # 回傳分類線和錯誤率\n",
    "    if(min(comp_x0)<min(comp_x1)):\n",
    "        x_split=(x0_sort[np.argmin(comp_x0)]+x0_sort[np.argmin(comp_x0)+1])/2\n",
    "        return x_split,0,min(comp_x0)\n",
    "    else:\n",
    "        x_split=(x1_sort[np.argmin(comp_x1)]+x1_sort[np.argmin(comp_x1)+1])/2\n",
    "        return x_split,1,min(comp_x1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Adapative Boost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The training error rate:  0.2\n",
      "The testing error rate:  0.15\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABDEAAAQECAYAAABjvYCTAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOzdf3Rb6X3f+c/jKwIGXW+TtcfNSWxLAu06yJoMlcrL8SreDpxuHB9N2z3dDZK2Ygq2XZPmnuw6yU5P43R3ZnaTODvJpulJlLBst6RKNU2RZNvskdZxNjGU08OkSMYmI3JV2TIhyU7SU//Ib4NDaKC7fzy8EUQRxAXug4t7gffrHB1c4D73wTMSBgQ+/D7PY3zfFwAAAAAAQNK9ZtgDAAAAAAAACIMQAwAAAAAApAIhBgAAAAAASAVCDAAAAAAAkAqEGAAAAAAAIBVODXsAw/LGN77RP3PmzLCHAQAAAAAA2nziE5/4ou/7Tx13bmxDjDNnzujll18e9jAAAAAAAEAbY8z9TueYTgIAAAAAAFKBEAMAAAAAAKQCIQYAAAAAAEgFQgwAAAAAAJAKhBgAAAAAACAVCDEAAAAAAEAqEGIAAAAAAIBUIMQAAAAAAACpQIgBAAAAAABSgRADAAAAAACkAiEGAAAAAABIBUIMAAAAAACQCoQYAAAAAAAgFQgxAAAAAABAKhBiAAAAAACAVCDEAAAAAAAAqUCIAQAAAAAAUoEQAwAAAAAApAIhBgAAAAAASAVCDAAAAAAAkAqEGAAAAAAAIBUIMQAAAAAAQCoQYgAAAAAAgFQgxAAAAAAAAKlAiAEAAAAAAFKBEAMAAAAAAKQCIQYAAAAAAEgFQgwAAAAAAJAKhBgAAAAAACAVCDEAAAAAAEAqEGIAAAAAAIBUIMQAAAAAAACpQIgBAAAAAABSgRADAJBKOzvS8rJUKEjZrGSMvS0U7OM7O8MeIQAAAFw7NewBAADQi3pdmp+XtrelgwOp1Xp0rtmUbt+W7tyRrlyRZmeljQ0pnx/eeAEAAOAOlRgAgNSoVKTpaalWkxqNxwOMdq2WPV+r2faVSrzjBAAAwGBQiQEASIVKRSqXpf398NcEYUa5bO+XSoMYGQAAAOJCJQYAIPHqdWlhobcAo93+vr3+7l234wIAAEC8CDEAAIk3P2/Xv4ji4EC6dMnNeAAAADAchBgAgES7edMu4tlp/YuwWi1pa4tdSwAAANKMEAMAkGgrK9GrMALNpu0PAAAA6USIAQBItGo1ehVGoNWy/QEAACCdCDEApFoul1Mulxv2MBLrQx/6kD70oQ8NexiR1Otu+9vbc9sfgP5EeX/a2ZGWl6VCQcpmJWPsbaFgH2faGPrF5wog+dhiFUCqvfLKK8MeQqJtb28PewiRNZtu+3vwwG1/APrTz/tTvW4X+t3ettPM2qu0mk3p9m3pzh3pyhVpdlba2JDyeYeDxsjjcwWQfFRiAAASLZNx29/EhNv+AMSjUpGmp6VaTWo0Ok8za7Xs+VrNtq9U4h0nAGCwCDEAAInm+reoU1Nu+wMweJWKVC6fHF4cFYQZ5TJBBgCMEkIMAECiFYuS57npy/NsfwDSo16XFhak/f3+rt/ft9ffvet2XACA4SDEAIAxksbF8JaW7BhdyGRsfwDSY34++jbLBwfSpUtuxgMAGC5CDAAYA/W6dOGC9PTT0uqqXfwuWDAzWAxvddWev3DB/Y4gUczM2AX6olZjeJ507pydIw8gHW7etIt4Rt1mudWStraSGdQCAHpDiAEAI+7zn/+LqV8Mb2MjejVGNitdvepmPADisbISvQoj0Gza/gAA6UaIAQAj7POf/4v61Ke+N/WL4eXz0tqalMv1d30uZ68/e9btuAAMVrUavQoj0GrZ/gAA6UaIAQAjql6XPvWp79XDh6/t6/qkLYZXKknr69LkZPipJZ5n26+v2+sBpIvrqW17e277AwDEjxADAEbU/Lz08OGpSH0kbTG8UsnOaZ+bs9UVncIMz7Pn5+ak3V0CDCCtgrV7XHnwwG1/AID4EWIAwAgKFsOTooUYSVwML5+XNjft2h2Li3ZnlUzG7rSSydj7i4v2/OYmU0iANMtk3PY3MeG2PwBA/KJ9ugUAJNIgFsO7fNlNf65MTydvTADcyuft7kmuTE256wsAMBxUYgDACGIxPACjoFiMvr1ywPNsfwCAdCPEAIARxGJ4AEbB0lL07ZUDmYztDwCQboQYADCCWAwPwCiYmZFmZ6NXY3iedO6cnYYGAEg3QgwAGEEshgdgVGxsRK/GyGalq1fdjAcAMFyEGAAwgvJ5t/2xGB6AYcnnpbU1u21yP3I5ez07FQHAaCDEAIARxGJ4AEZJqSStr0uTk+Hf2zzPtl9ft9cDAEYDIQYAjCAWwwMwakolaWdHmpuz1RWdwgzPs+fn5qTdXQIMABg1hBgAMIKCxfCkVyP1w2J4AJIkn5c2N6VaTVpclAoFG7QaY28LBft4rWbbMYUEAEbPqWEPAAAwGBsb0tvf/qoePuz/rX5igsXwACTP9LR0+fKwRwEAGAYqMQBgROXz0jve8RG95jWv9N3Hw4fSpUtSve5wYAAAAECfCDEAYIS96U2/qne84yM9LYbXrtm0ZdnT01Kl4n58AAAAQC8IMQBgxL3pTb/6p4vhZTK9X99qSY2GVC4TZAAAAGC4CDEAYAzk83aNjNdEeNff35cWFqS7d92NCwAAAOgFIQYAjIn5eenBg2h9HBzYNTIAAACAYSDEAIAxcPOmtL1tp4ZE0WpJW1vSzo6bcQEAAAC9IMQAgDGwsmKrKFxoNm1/AAAAQNwIMQBgDFSr0aswAq2W7Q8AAACIGyEGAIyBet1tf3t7bvsDAAAAwiDEAIAx0Gy67S/qAqEAAABAPwgxAGAMZDJu+5uYcNsfAAAAEAYhBgCMgXzebX9TU277AwAAAMIgxACAMVAsSp7npi/Ps/0BAAAAcSPEAIAxsLQkZbNu+spkbH8AAABA3AgxAGAMzMxIs7PRqzE8Tzp3TpqedjMuAAAAoBeEGAAwJjY2oldjZLPS1atuxgMAAAD0ihADAMZEPi+trUm5XH/X53L2+rNn3Y4LAAAACOvUsAcAAIhPqWRvFxakgwOp1ep+jefZCoy1tUfXAwAAAMNAJQYAjJlSSdrZkebmbHVFp3UyPM+en5uTdncJMAAAADB8hBgAMIbyeWlzU6rVpMVFqVCwu44YY28LBft4rWbbMYUEAAAAScB0EgAYY9PT0uXLwx4FAAAAEE5PIYYxxpNUkHRe0l84vP16ScEycS/6vv+Cq8EZY94q6f2S/svD53mrpNdK+kNJn5L0cUn/1Pf9z7p6TgAAAAAAkEy9VmJUJP21QQzkKGPMv5H0VySZY06/8fDPBUl/zxjzD3zf/5E4xgUAAAAAAIaj1xDj6PJvvyfpS5Le7mY4j3mnHgUYvympKunTkv5I0psl/TeyIUZW0g8bY7K+7//AAMYBAAAAAAASoNcQ4zck/XtJn5D0Cd/37xpjypLWXA9M0iuSLkv6cd/3P3XM+X9ojPmQpH94eP8FY8zP+r7/6QGMBQAAAAAADFlPIYbv+z84qIEc4z2+7//+SQ183/8xY8w3ylZlnJL0NyU9H8fgAAAAAABAvBK7xWq3AKPNz7YdTw9iLADgws6OtLxsty/NZu12ptmsvb+8bM8DAAAA6GwUtlj947bjXMdWADAk9bo0Py9tb0sHB1Kr9ehcsyndvi3duSNduSLNzkobG1I+P7zxAgAAAEmV2EqMHryz7fj+0EYBAMeoVKTpaalWkxqNxwOMdq2WPV+r2faVSrzjBAAAANIg1SGGMeaUpIW2h64PaywAcFSlIpXLJ4cXRwVhRrlMkJFWTBsCAAAYnFSHGJL+J0lfe3i8oy4hhjHmA8aYl40xL3/hC18Y+OAAjK96XVpYkPb3+7t+f99ef/eu23FhcOp16cIF6emnpdVVO02o2bTngmlDq6v2/IULtj0AAAB6k9oQwxhTlPS/Hd59VdKi7/sPT7rG9/1V3/fP+75//qmnnhr4GAGMr/l5u/5FFAcH0qVLbsaDwWLaEAAAQDxSGWIYY75WdleSYGHSD/u+/+tDHBIA/KmbN+0inmGnkHTSaklbW0w/SDqmDQEAAMQndSGGMeaspP9X0hsOH/ox3/d/eIhDAoDHrKxEr8IINJu2PyQT04YAAADilaoQwxjzFkkfl/Tmw4dWfN//riEOCQCeUK1Gr8IItFq2PyQT04YAAADilZoQwxjz1bIBxpnDh/6ZpOWhDQgAOnC9YOPentv+4AbThgAAAOKXihDDGPNVsgHG2w4f+heS/jvf9/3hjQoAjhfsSOHKgwdu+4MbTBsCAACIX+JDDGPMU5J+RdI7Dh/6V5L+VredSABgWDIZt/1NTLjtD24wbQgAACB+iQ4xjDH/qaRflvR1hw/9X5Iu+b7v6GMjALiXz7vtb2rKbX9wg2lDAAAA8RtKiGGMWTfG+Id/XujQ5s9K+iVJM4cP/d+Svt33/VdjGiYA9KVYlDzPTV+eZ/tD8jBtCAAAIH6neml8uL3p3zny8Ezb8XuNMUf7/Hnf97f6GNtHJf2Fw+PflV0H46Ix5qRrGr7v/1IfzwUAziwtSVeuSI1G9L4yGdsfkieTcRtkMG0IAACgu55CDEmnJX3fCeffc/in3Wck9RNivLvt+Ktl18Lo5r4e7V4CAEMxMyPNzkq1WrQ1EzxPOndOmp52Nza4k89Lt2+7649pQwAAAN0lek0MAEirjQ0pm43WRzYrXb3qZjxwj2lDAAAA8espxPB9/4bv+6bHP+vH9FNuO/9Ch+fq9XmM7/tn+vpbAADH8nlpbU3K5fq7Ppez158963ZccGdpKXpQFWDaEAAgqp0daXlZKhTszydj7G2hYB/f2Rn2CAE3ep1OAgAIqVSytwsL0sFBuKklnmc/cKytPboeycS0IQBAEtTr0vy8tL395OeNZtNOfbxzx67XNTtrq0Vd76QGxInpJAAwQKWS/c3H3Jytrug0/cDz7Pm5OWl3lwAjLZg2BAAYpkrFhuC1ml1QvFOo3mrZ87WabV+pxDtOwCVCDAAYsHxe2ty0HxwWF21ZZyZjyzwzGXt/cdGe39xkCkmaMG0IADAslYpULp8cXhwVhBnlMkEG0ovpJAAQk+lp6fLlYY8CrjFtCAAQt3rd/tzZ3+/v+v19e/273kWQjvShEgMAgIiYNgQAiNP8vA3Oozg4kC5dcjMeIE6EGAAAOMC0IQBAHG7etIt4RllUWrLXb22xawnSh+kkAAA4xLQhAMAgraxEr8IINJu2P35uIU2oxAAAAACAlKhWo1dhBFot2x+QJoQYAAAAAJAS9brb/vb23PYHDBohBgAAAACkRLPptr8HD9z2BwwaIQYAAAAApEQm47a/iQm3/QGDRogBAAAAACmRz7vt72u+xm1/wKARYgAAAABAShSLkue56+9zn5MuXHC/1gYwKIQYAAAAAJASS0tSNuuuv1dflWo1u0V4peKuX2BQCDEAAAAAICVmZqTZWbfVGK2W1GhI5TJBBpKPEAMAAAAAUmRjw201RmB/X1pYkO7edd834AohBgAAAACkSD4vra1JuZz7vg8OpEuX3PcLuEKIAQAAMGQ7O9LyslQo2N+uGmNvCwX7+M7OsEcIIGlKJWl9XZqcdD+1ZGuL9x0kFyEGAADAkNTrdleAp5+WVlel27elZtOeazbt/dVVe57dAwAcVSrZsGFuzm2Q0WxKKyvu+gNcIsQAAAAYgkrF7gZQq9kF9Vqt49sFC+6xewCA4+Tz0uam9Na3uuuz1ZKqVXf9AS6dGvYAAAAAxk2lYncB2N8Pf0377gGS/Q0sAAR+53fc9re357Y/wBUqMQAAAGJUr9vV/3sJMNqxewCA4wRT0Vx58MBtf4ArhBgAAAAxmp+3q/9Hwe4BAI7KZNz2NzHhtj/AFUIMAACAmNy8KW1vd17/Iix2DwBwVD7vtr+pKbf9Aa4QYgAAAMRkZSV6FUaA3QMAtCsW3e1Q4nm2PyCJCDEAAABiUq1Gr8IIsHsAgHZLS1I266avTMb2ByQRIQYAAEBM6nW3/bF7AIDAzIw0Oxu9GsPzpHPn7JbOQBIRYgAAAMSE3QMADNLGRvRqjGxWunrVzXiAQSDEAAAAiAm7BwAYpHxeWluTcrn+rs/l7PVnz7odF+ASIQYAAEBM2D0AwKCVStL6ujQ5GX5qiefZ9uvr9nogyQgxAAAAYsLuAQDiUCrZLZjn5mx1Raf3Hc+z5+fmpN1dAgykAyEGAABATNg9AEBc8nlpc1Oq1aTFRalQsO8bxtjbQsE+XqvZdkwhCWdnR1petn9/2az9+8xm7f3lZXseg3Vq2AMAAAAYF8HuAbVatK1W2T0AQFjT09Lly8MeRfrV69L8vLS9LR0cPP4e3mxKt29Ld+5IV67Y9/mNDfdTCGFRiQEAABAjdg8AgHSpVGwYVKtJjUbnELrVsudrNdu+Uol3nOOCEAMAACBG7B4AAOlRqUjl8snhxVFBmFEuE2QMAiEGAABAzNg9AACSr16XFhak/f3+rt/ft9ffvet2XOOOEAMAAGAI2D0AAJJtft6ufxHFwYF06ZKb8cAixAAAABgSdg8AgGS6edMu4hllEWbJXr+1xa4lLrE7CQAAwJCxewAAJMvKSvQqjECzafvjfd4NKjEAAAAAAGhTrUavwgi0WrY/uEGIAQAAAABAm3rdbX97e277G2eEGAAAAAAAtGk23fb34IHb/sYZIQYAAAAAAG0yGbf9TUy47W+cEWIAAAAAANAmn3fb39SU2/7GGSEGAAAAAABtikXJ89z05Xm2P7hBiAEAAAAAQJulJSmbddNXJmP7gxuEGAAAAAAAtJmZkWZno1djeJ507pw0Pe1mXCDEAAAAAADgCRsb0asxslnp6lU344FFiAEAAAAAwBH5vLS2JuVy/V2fy9nrz551O65xd2rYAwAAAAAAIIlKJXu7sCAdHEitVvdrPM9WYKytPboe7lCJAQAAAABAB6WStLMjzc3Z6opO62R4nj0/Nyft7hJgDAohBgAAAAAAJ8jnpc1NqVaTFhelQsHuOmKMvS0U7OO1mm3HFJLBYToJAAAAAAAhTE9Lly8PexTjjUoMAAAAAMBY2NmRlpdt5UQ2ayspsll7f3nZnkeyUYkBAAAAABhp9bo0Py9tbz+5QGezKd2+Ld25I125Is3O2u1V8/nhjRedUYkBAAAAABhZlYqdBlKrSY1G5x1GWi17vlaz7SuVeMeJcKjEAAAAAACMpEpFKpel/f3w1wRhRrls77PLSLJQiQEAAAAAGDn1urSw0FuA0W5/315/967bcSEaQgwAAAAAwMiZn7frX0RxcCBduuRmPHCDEAMAAAAAMFJu3rSLeHZa/yKsVkva2mLXkiQhxAAAAAAAjJSVlehVGIFm0/aHZCDEAAAAAACMlGo1ehVGoNWy/SEZCDEAAAAAACOlXnfb396e2/7QP0IMAAAAAMBIaTbd9vfggdv+0D9CDAAAAADASMlk3PY3MeG2P/SPEAMAAAAAMFLyebf9TU257Q/9I8QAAAAAAIyUYlHyPDd9eZ7tD8lAiAEAAAAAGClLS1I266avTMb2h2QgxAAAAAAAjJSZGWl2Nno1hudJ585J09NuxoXoCDEAAAAAACNnYyN6NUY2K1296mY8cIMQAwAAAAAwcvJ5aW1NyuX6uz6Xs9efPet2XIjm1LAHAAAAAADAIJRK9nZhQTo4kFqt7td4nq3AWFt7dD2Sg0oMAAAAAMDIKpWknR1pbs5WV3RaJ8Pz7Pm5OWl3lwAjqQgxAAAAAAAjLZ+XNjelWk1aXJQKBbvriDH2tlCwj9dqth1TSJKL6SQAAAAAgLEwPS1dvjzsUSAKKjEAAAAAAEAqEGIAAJByOzvS8rIthc1mbWlsNmvvLy/b8wAAAKOA6SQAAKRUvS7Nz0vb20+uuN5sSrdvS3fuSFeuSLOz0saGnRMMAACQVlRiAACQQpWKnddbq0mNRuct41ote75Ws+0rlXjHCQAA4BKVGAAApEylIpXL0v5++GuCMKNctvfZNg4AAKQRlRgAAKRIvS4tLPQWYLTb37fX373rdlwAAABxIMQAACBF5uft+hdRHBxIly65GQ8AAECcCDEAAEiJmzftIp6d1r8Iq9WStrbYtQQAAKQPIQYAACmxshK9CiPQbNr+AAAA0oQQAwCAlKhWo1dhBFot2x8AAECaEGIAAJAS9brb/vb23PYHAAAwaIQYAACkRLPptr8HD9z2BwAAMGiEGAAApEQm47a/iQm3/QEAAAwaIQYAACmRz7vtb2rKbX8AAACDRogBAEBKFIuS57npy/NsfwAAAGlCiAEAQEosLUnZrJu+MhnbHwAAQJoQYgAAkBIzM9LsbPRqDM+Tzp2TpqfdjAsAACAuhBgAAKTIxkb0aoxsVrp61c14AAAA4kSIAQBAiuTz0tqalMv1d30uZ68/e9btuAAgTjs70vKyVCjYYNYYe1so2Md3doY9QgCDcmrYAwAAAL0pleztwoJ0cCC1Wt2v8Tz7AX9t7dH1AJA29bo0Py9tbz/5/tdsSrdvS3fuSFeu2Ol3Gxvud3YCMFxUYgAAkEKlkv1N49ycra7otE6G59nzc3PS7i4BBoD0qlTsWj61mtRodA5wWy17vlaz7SuVeMcJYLAIMQAASKl8XtrctB/UFxdtGXUmY8uqMxl7f3HRnt/cZAoJgPSqVKRy+eTw4qggzCiXCTKAUcJ0EgAAUm56Wrp8edijAIDBqNft9Ln9/f6u39+317/rXYS5wCigEgMAAABAYs3P2/Uvojg4kC5dcjMeAMNFiAEAAAAgkW7etIt4hp1C0kmrJW1tsWsJMAoIMQAAAAAk0spK9CqMQLNp+wOQboQYAICxtLMjLS/bxS+zWbsYZjZr7y8v89s6AEiCajV6FUag1bL9AUg3FvYEAIyVet3Or97etr/da/9w3GxKt29Ld+5IV65Is7PSxobdBQQAEL963W1/e3tu+wMQPyoxAGAEdKoq+M3fvKJPf/pDVBUcqlTsTh612snb9AXb8tVqtj1b8wHAcDSbbvt78MBtfwDiRyUGAKRYt6qCZvO0Go2v0dNPU1VQqUjlcm9b9AVhRrls75dKgxgZAKCTTMZtkDEx4a4vAMNBJQYApFTYqgLp1NhXFdTr0sJCbwFGu/19e/3du27HBQA4mevgfWrKbX8A4keIAQApFFQVnBxePK69qmDcgoz5+eir2x8cSJcuuRkPACCcYlHyPDd9eZ7tD0C6EWIAQMpQVdCbmzftdJuoq9u3WtLWFruWAECclpbsGk8uZDK2PwDpRogBAClDVUFvVlai/30Fmk3bHwAgHjMzdk2nqNUYniedO2enVQJIN0IMAEgRqgp6V61G//sKtFq2PwBAfDY2oldjZLPS1atuxgNguAgxACBFqCroXb3utr+9Pbf9AQBOls9La2tSLtff9bmcvf7sWbfjAjAcbLEKAClCVUHvXG7NJ0kPHrjtDwDQXbDF9cLCk1uKd+J5tgJjbY0tsoFRQiUGAKQIVQW9y2Tc9jcx4bY/AEA4pZKdBjk3Z6srOq2T4Xn2/NyctLtLgAGMGkIMAEgRqgp6l8+77W9qym1/AIDw8nlpc1Oq1aTFRalQsGG1Mfa2ULCP12q2HVNIgNHDdBIASJFMxm2QMQ5VBcWidOeOm2k4nmf7AwAM1/S0dPnysEcBYBioxACAFKGqoHdLS9FXtQ9kMrY/AAAADAchBgCkSLHYeQ5wr8alqmBmRpqdjf735nnSuXP2t38AAAAYjp5CDGOMZ4x5pzGmbIz5cWPMrxtjGsYY//DPC4MYpDHmdcaY7zbGbBpjPm+MecUYc98Y86+MMe8bxHMCQBJRVdCfjY3of2/ZrHT1qpvxAAAAoD+9rolRkfTXBjGQTowx5yT9nKSjRdRvPfxTMsb8tKQF3/cdL3kHAMkSVBXUatHWeBi3qoJ83m6xVy5L+/u9X5/L2etZIA4AAGC4ep1OcrQY9/ck3XE0licYY05L+qgeBRi/Iel/kPQ3JH1E0pcOH/8bkv7PQY0DAJKEqoL+lErS+ro0ORl+aonn2fbr62zRBwAAkAS9hhi/IemHJH2rpLzv+2+Q9IPOR/XIj0n6c4fH/0zSu33f/3Hf9/+l7/sflvQNkj57eP6SMebiAMcCAIkQVBXkcv1dP85VBaWStLMjzc3Zv4dOYYbn2fNzc9LuLgEGAABAUvQ0ncT3/UEGFo8xxny9pP/68O5nJf33vu8/PDKezxpjPijp+uFDL7QdA8DICr5ULyxIBwfhppZ4nq3AWFsb7y/l+by0uWnDjJUVqVqV9vakBw/slrNTU3bB06Wl8ZluAwAAkBa9rokRp29rO171ff+VDu0+Kukzkt4m6bwxZsr3/b2Bjw4AhqxUks6fl+bnpa0tqdnsFGa8qlzulM6ds1NIxrEC4zjT09Lly8MeBQAAAHqR5BDjm9uOf7FTI9/3fWPMx2RDDEl6n6SfHOTAhuXUqVNqRVnJDxhhxphhD2HI3ilpUdJ7JU1JmpD0QNKepI9rf/8f69d+bVf5o0skA8CQ8f6NJOJ1iVHieZ5effXVYQ/DmUSGGMaY10gqHN59VdJvdbnk5bbjdw5kUACQaLuSvnPYgwAAAAAGKpEhhqQ3S5o8PP4d3/e7xUb3247//GCGNHyjlJ4BrgS/KfF9f8gjSaZnnnlGknTjxo2hjgMYadWqXaBmbc0uqBK13Zjg/QlJxOcKIPl63Z0kLl/RdvzFEO2/1Hb8FZ0aGWM+YIx52Rjz8he+8IW+BwcAACDJBhPPPivdv29vO30hP9quWo11mAAAjIqkhhh/pu2404Ke7fbbjl/fqZHv+6u+75/3ff/8U0891ffgAAAA/jSYaDTs/UZDunjxySDjuHYEGQAA9CWpIUY7arkAAECyHA0mAkeDjJPaEWQAANCzpIYYf9J2nAvRvr3NHzseCwAAwCOdgolAEGS89FL3dgQZAAD0JKkhxh+0Hb8hRPv2Nn/QsRUAAEBUCwudg4lAoyG9+GK4dgsL7sYGAMCIS2qI8duSgp/6bzbGdNtF5XTb8acHMyQAAADZ3UUmJ4aPaeIAACAASURBVLu36xZgSLaf9fXIQwIAYFwkMsTwff+hpH9/ePeUpK/vcsn5tuPdgQwKAABAstujXrsWLsg4yeSkdP26dLjVKAAA6C6RIcahj7Udv69TI2M3c24//7FObQEAAJyIGmQQYAAA0JckhxiVtuNFY8xrO7R7v6S3HR6/7Pv+3mCHBQAAoP6DDAIMAAD6NpQQwxizbozxD/+8cFwb3/d/S9K/Obz7Vkk/YYx5bLzGmLdK+qm2h47tCwAAYCCKRen558MHGZOT0gsvEGAAANCnbgtmPsYYc1bS3zny8Ezb8XuPWYTz533f3+pncJI+JOndkv7c4fO+0xizIelLkqYlLerRziT/wvf9630+DwAAQO+q1XC7kAQaDRtivOtdBBkAAPShpxBDdheQ7zvh/HsO/7T7jKS+Qgzf9+8bY94v6eck5SXNHf456qcl/e1+ngMAAKAv1ar07LPhA4xAoyFdvCh9//dL/+gf2d1OisXBjBEAgBGT5DUxJEmHVRwzkr5H0q9L+qKkA0mfk/Szkt7v+/7f9H2/ObxRAgCAsdJvgBFoNKTv/m7p/n3bT7XqdnwAAIyonioxfN+/IclEfVLf98uSyj20/7KkHz38AwAAMDxRA4yjGg3b37VrVGQAANBF4isxAAAAEmVhwV2AEQiCDCoyAAA4ESEGAABAL9bWet9WNQyCDAAAuiLEAAAA6EWxaKd+DCrIWFhw3y8AACOCEAMAAAxHtSqdOdO98iBsuzgNKsiYnJTW1932CQDACCHEAAAA8QsWxwx257hxI1y7UQ4yJiel69elZ55x0x8AACOIEAMAAMTr6O4ejYZ08eKTQcZx7ZIaZJw+Lf3oj/YfaBBgAAAQCiEGAACIT6ftSY8GGSe1S2KQce+e9F3f1V9lBgEGAAChEWIAAIB4dAomAkGQ8dJL3dslLcgIFIvS88+HDzImJ6UXXiDAAAAgJEIMAAAQj4WFzsFEoNGQXnwxXLsk7uJRrYYbf6DRsCFGpzVBAADAYwgxAABAPNbWwlUohAkAkriLR7dKk046rQkCAACeQIgBAADi4Wo3jySuIdFvgBEgyAAAIBRCDAAAEJ+oQcYoBhgBggwAALoixAAAAPHqN8hIYoAhhVvrQwo/laZcjjwkABimnR1peVkqFKRsVjLG3hYK9vGdnWGPEGlGiAEAAOI3Srt4hFnrIxh/mHZra86GBgBxqtelCxekp5+WVlel27elZtOeazbt/dVVe/7CBdse6BUhBgAg3apV6cyZ7ttthm2HeIzSLh7dKkuCCpLnnuve7to12x8ApEylIk1PS7WafctutY5v12rZ87WabV+pxDtOpB8hBgAgvYK1CO7ft7edvuAebUeQMVyjuItHpyDj6BSYk9oRYABIqUrFzoQ7Kbw4KggzymWCDPSGEAMAkE5Hvwh3+oJ7XDuCjOEZ5V08jgYUndbwOK4dAQaAlKrX7dJA+/v9Xb+/b6+/e9ftuDC6CDEAAOnT6Yvw0S+4J7UjyIjfOOziEQQUp0+fvAhpezsCDAApNj8vHRxE6+PgQLp0yc14MPoIMQAA6dLti3DwBfell7q3I8iI17js4lEsSvfudV+ENGhHgAEgpW7elLa3w08h6aTVkra22LUE4RBiAADSJcwX4UYj3KKRjYbtD/FgFw8AGCkrK9GrMALNpu0P6IYQAwCQLmG+CEvhf+O/vh55SAiJXTwAYKRUq9GrMAKtFsWRCIcQAwCQLt2+CIfVadFFDBa7eADAyKjX3fa3t+e2P4wmQgwAQPpEDTIIMIaLXTwAYCQ0m277e/DAbX8YTYQYAIB06jfIIMBIBnbxAIDUy2Tc9jcx4bY/jCZCDABAehWL0vPPhw8ygkUjCTCSgV08ACDV8nm3/U1Nue0Po4kQAwCQXtVquF1IAo2GDTFu3BjkqAAAGAvFouR5bvryPLJqhEOIAQBIp2pVevbZ8AFGoNGQLl4kyAAAIKKlJSmbddNXJmP7A7ohxAAApE+/AUaAIAMAgMhmZqTZ2ejVGJ4nnTsnTU+7GRdGGyEGACBdogYYAYIMAAAi29iIXo2RzUpXr7oZD0YfIQYAIF0WFsIFGGEW+2w0pHI58pAAABhX+by0tiblcv1dn8vZ68+edTsujC5CDABAuqytdQ8ogl1IwrRbW3M2NAAAxlGpJK2v2x+rYaeWeJ5tv75urwfCIsQAAKRLsShdu9Y5oJiclK5fl557rnu7a9dYCh0AAAdKJWlnR5qbs9UVncIMz7Pn5+ak3V0CDPSOEAMAkD6dgowgwHjmme7tCDAAAHAqn5c2N6VaTVpclAoFu+uIMfa2ULCP12q2HVNI0I9Twx4AAAB9CQKKYJHPowHGSe0IMAAAGJjpaeny5WGPAqOKSgwAQHoFAcXp08cHGMe1I8AAAABILSoxAADpVixK9+65awcAAIDEohIDAAAAAACkAiEGAAAAAABIBUIMAAAAAACQCoQYAAAAAAAgFQgxAAAAAABAKhBiAAAAAACAVCDEAAAAAAAAqUCIAQAAAAAAUoEQAwAAAAAApAIhBgAAAAAASAVCDAAAAAAAkAqEGAAAAAAAIBUIMQAAAAAAQCoQYgAAAAAAgFQgxAAAAAAAoM3OjrS8LBUKUjYrGWNvCwX7+M7OsEc4vk4NewAAAAAAACRBvS7Nz0vb29LBgdRqPTrXbEq3b0t37khXrkizs9LGhpTPD2+844hKDAAAAADA2KtUpOlpqVaTGo3HA4x2rZY9X6vZ9pVKvOMcd1RiAAAAAADGWqUilcvS/n74a4Iwo1y290ulQYwMR1GJAQAAAAAYW/W6tLDQW4DRbn/fXn/3rttx4XiEGAAAAACAsTU/b9e/iOLgQLp0yc14cDJCDAAAAADAWLp50y7i2Wn9i7BaLWlri11L4kCIAQAAAAAYSysr0aswAs2m7Q+DRYgBAAAAABhL1Wr0KoxAq2X7w2ARYgAAAAAAxlK97ra/vT23/eFJhBgAAAAAgLHUbLrt78EDt/3hSYQYAAAAAICxlMm47W9iwm1/eBIhBgAAAABgLOXzbvubmnLbH55EiAEAAAAAGEvFouR5bvryPNsfBosQAwAAAAAwlpaWpGzWTV+ZjO0Pg0WIAQAAMO6qVenMme57A4ZtBwApMTMjzc5Gr8bwPOncOWl62s240BkhBgAAUfDlD2lXrUrPPivdv29vb9wI147XMoARsbERvRojm5WuXnUzHpyMEAMAgH7x5Q9pF7w2Gw17v9GQLl588rV8XDteywBGRD4vra1JuVx/1+dy9vqzZ92OC8cjxAAAoB98+UPaHX1tBo6+lk9qx2sZwIgolaT1dWlyMvzUEs+z7dfX7fWIByEGAAC94ssf0q7TazMQvJZfeql7O17LAEZEqSTt7Ehzc7a6olOY4Xn2/NyctLtLgBE3QgwAAHrBlz+MgoWFzq/NQKMhvfhiuHYLC+7GBgBDlM9Lm5tSrSYtLkqFgt11xBh7WyjYx2s1244pJPE7NewBAACQKoP48nfvnrPhAaGsrZ0csgW6nZce1VIDwAiZnpYuXx72KHAcKjEAAOjF2pr90tYNX/6QZMWidO1auNfySSYnpevXpWeecTIsAAC6IcQAAKAXfPnDqIj6WuY1DAAYAkIMAAB6xZc/jIp+X8u8hgEAQ0KIAQBAP/jyh1FRLErPPx/+tTw5Kb3wAq9hAMBQEGIAANAvvvxhFFSr4RaiDTQa9nUcbCUMAECMCDEAAOgXX/6Qdt22DO4k2EqY1zIAIGaEGAAA9IMvf0i7fl/DAV7LwFDt7EjLy1KhIGWzkjH2tlCwj+/sDHuEwGAQYgAA0Cu+/CHtor6GA7yWgdjV69KFC9LTT0urq9Lt21Kzac81m/b+6qo9f+GCbQ+MEkIMAAB6wZc/jIKFhXCv4TDrvTQaUrkceUgAuqtUpOlpqVaz/+u1Wse3a7Xs+VrNtq9U4h0nMEiEGAAA9IIvfxgFa2vdX6PBQrRh2q2tORsagONVKvZHxknhxVFBmFEuE2RgdBBiAADQC778YRR02yI42Ar4uee6t7t2zfYHYGDqdZuh7+/3d/3+vr3+7l234wKGgRADAIBe8OUPo6LTazl4DQdbAZ/UjtcwEIv5eengIFofBwfSpUtuxgMMEyEGAGCgRnL1dL78YVQcfY0efQ2f1I7XMBCLmzel7e3wU0g6abWkra2U/twF2hBiAAAGYuRXT+fLH0ZF8Bo9ffr41/Bx7XgNA7FZWYlehRFoNm1/QJoRYgAAnBub1dP58odRUSxK9+51fg0fbcdrGIhNtRq9CiPQatn+gDQ7NewBAABGS7B6ei+Lj7Wvni5JpdIgRjYgwZc6V+0AAGjjulJxb89tf0DcqMQAYFWr0pkz3eP5sO0wllg9HQAAt4KpmK48eOC2PyBuhBgAbCDx7LPS/fv29saNcO0IMnAEq6cDAOBWJuO2v4kJt/0BcSPEAMZdEEw0GvZ+oyFdvPhkkHFcO4IMtGH1dAAA3Mvn3fY3NeW2PyBuhBjAODsaTASOBhkntSPIwCFWTwcAwL1iUfI8N315HuvyIv0IMYBx1SmYCARBxksvdW9HkAGxejoAAIOwtCRls276ymRsf0CaEWIA42phoXMwEWg0pBdfDNduYcHd2JBKrJ4OIIl2dqTlZalQsF8EjbG3hYJ9nKlrSLqZGWl2Nno1hudJ587ZLc2BNCPEAMbV2po0Odm9XbcAQ7L9rK9HHhLSjdXTASRJvS5duCA9/bS0uirdvv3ofarZtPdXV+35CxfcB7GASxsb0asxslnp6lU34wGGiRADGFfFonTtWrgg4ySTk9L169IzzzgZFtKL1dMBJEWlYn/bXKvZLL7TVLdWy56v1Wz7SiXecQJh5fP290+5XH/X53L2+rNn3Y4LGAZCDGCcRQ0yCDDQhtXTASRBpSKVyyeHF0cFYUa5TJCB5CqVbOHr5GT4qSWe96hgtlQa5OiA+BBiAOOu3yCDAANHsHo6gGGr1+0STfv7/V2/v2+vv3vX7bgAV0olu47L3Jytruj0c9fz7Pm5OWl3lwADo4UQA0ibalU6c6b71g1h20n22+Lzz4cPMiYnpRdeIMDAY1g9HcCwzc9H3+r54EC6dMnNeIBByOelzU07DWpx0S5Sm8nYRWszGXt/cdGe39xkCglGDyEGkCbBtqj379vbGzfCtQsTeITZhSTQaNgQo9PzYyyxejqAYbp5U9rejr7Vc6slbW2xawmSb3paunxZunXLhm8PH9rbW7fs4/wcxagixADSIggmgqCh0ZAuXnwySDiu3UlBxtH2YXV6fow1Vk8HMCwrK9GrMALNpu0PAJA8hBhAGnQKGo4GCSe1Oy7I6DfA6PT8GHusng5gWKrV6FUYgVYr3GxMAED8CDGApOsWNARBwksvdW/XHmREDTCOPj9BBg6xejqAYajX3fa3t+e2PwCAG4QYQNItLHQPGhqNcGtaNBq2v7D9SuEW+wz2pQMOsXo6gLg1m277e/DAbX8AADcIMYCkW1sLHyR0E/yqO2y/wS4kYdqtrXV/fowVVk8HEKdMxm1/ExNu+wMAuHFq2AMA0EWxKF27Fn3qx+SkdP36o21Ru/Xb3v78+ZPbXbtm+wOOEayeDgCDlM9Lt2+7629qyl1fAAB3qMQA0iAIHMJUZBznaIDRrd9Ogcdx7QgwAAAJUCxG3+I54Hn8aAOApCLEANKi3yCjU4DRqd+wgQcBBgAgQZaWom/xHMhkbH8AgOQhxADSpFiUnn8+fJARrGnRKcBo7/faNen06XCBx+nTBBgAgESZmZFmZ6NXY3iedO6cnQoHAEge1sQA0qRaDbcLSaDRsCHGu94VLsi4d697n2HbAQAQs40NGz5EWUIqm5WuXnU3JgCAW1RiAGlRrfa3uGejIV28KN24MZBhAQCQFPm83Swrl+vv+lzOXs9uSQCQXIQYQBr0G2AECDIAAGOiVLK7iU9Ohp9a4nmPdiEvlQY5uphVq9KZM/bWRTsASABCDCDpogYYAYIMAMCYKJWknR1pbs5WV3QKMzzPnp+bk3Z3RzDAePZZ6f59e9vp5//RdgQZABKOEANIuoWFcAFGmMU+Gw2pXI48JAAAki6flzY3pVpNWlyUCgW764gx9rZQsI/XarbdSE0hOfoLkE6/yDiuHUEGgIQjxACSbm2te0AR7EISpt3amrOhAQCQdNPT0uXL0q1b0sGB9PChvb11yz4+cruQdKrgPBpknNSOIANAghFiAEkXbGvaKaCYnLTboj73XPd2bIsKAMDo6jYFNQgyXnqpezuCDAAJRYgBpEGnICMIMILtU09qR4ABAMBoCzMFtdEIt117o2H7A4CE6TnEMNa3GWOuGWN+2xhzYIz5D8aYXzHG/F1jzCnXgzTGXDTG/LQx5o4x5k+MMU1jzBeNMb9mjPlBY8yU6+cEEudoQHE0wDipHQEGAACjL8wUVCn8Wlvr65GHBACu9RRiGGO+UtIvS/oZSRclfY2kjKSvkvReSf9EUs0Y81YXgzPGvNEYU5V0TdJfl/Q2Sa+TNCHpDZLeLel7Jd0yxvw9F8+ZGmyZNZ6CgOL06eMDjOPaEWAAADAeuk1BDavTL0oAIAFCV00YYzKSfkHSew4f+pykVUmfkfRmSX9bUkHSN0j6qDHm3b7v/1G/Azus6PiopPOHD70iaUPStqTfl/QWSX9Z0jfKBin/uzHmT3zf/8l+nzM12uc7Pvts5x8yR9vxZXY0FIvSvXvu2gEAgNERBBn9bs9OgAEg4XqpxPigHgUYn5T09b7vf7/v+z/j+/6PyIYXHzs8/3WS/ueIY/vrehRgfE7S1/q+/wHf93/S9/1/6fv+S77vv0fSB9queXEQ01kShS2zAKTUzo60vGy3Ncxm7TaH2ay9v7xszwMAHOi3IoMAA0AKhAoxDoOB7zu860v6Dt/3f7+9je/7r0j6DklfPnzoO40xb4gwtve1Hf+Q7/v3j2vk+/4/kfSJw7tvlK0GGU1smQUghep16cIF6emnpdVV6fZtqdm055pNe3911Z6/cMG2BzBETFkdDcWi9Pzz4YOMYLt2AgwACRe2EuO9kp46PP4V3/f/v+Ma+b7/edn1MiQpK+mvRhjbm9qO73Rp++m249dFeM7kYsssAClUqUjT01KtZt9+Wq3j27Va9nytZttXKvGOE8Ch4PPG/fv29milZ6d2fK5Inmo13C4kgUbDhhid/s0BICHChhjf3Hb8i13atp//lt6G85j/2Hb89i5tg/MtPR5ojA62zAKQMpWKVC6fHF4cFYQZ5TJBBhA7pqyOjm6//Oqk0785ACRI2BDjnW3Hn+jYynq5w3W9+oW2479vjDl9XCNjzN/Vo7UzNnzf/70Iz5lcbJkFIEXqdZuV7u/3d/3+vr3+7l234wLQAVNWR0e/AUaAIANAwoUNMf582/G9Lm1/W7YiQpLebowxvQ7q0M9L+teHx2+RdNsY84+NMR80xny7MeY5Y8y/ld3WVYdtv7PP50o+tswCkCLz89LBQbQ+Dg6kS5fcjAfACZiymkrHLZb8zRNV7X9ThAAjQJABIMHC7uTxFW3HXzypoe/7rxpj/kjSVx72/zpJf9LrwHzf940x3yrpf5UNJ16vx3ciCXxS0v8i6f/xfd/v9XlShS2zAKTAzZvS9nb4KSSdtFrS1pb9oD497WZsAI4xiCmrbPE9MPW6DYq3t23Y2/5eu/rqgnIKWZUb5t+yXObfEkDihK3E+DNtx6+EaN9eQPz68MN5nO/7LUk/LOlFSZ1+p/cNkv6+pHd3688Y8wFjzMvGmJe/8IUv9Dus4WLLLAAJt7ISvQoj0Gza/gAMEFNWU6PbYskLWtOX1eXfMtiFpNu/+eSkfW0AQMKEDTGGwhjzLbLTV35E0q/Lbrv6FbI7n7xd0odlt3T9Rkm/Yow5cTcU3/dXfd8/7/v++aeeeuqkpsnGllkAEqxajV6FEWi1qEyHY2wf+iSmrKZCmMWSb6ioZ3Wtc5AR/Bs999zJ/+aTk/Z8sehk7ADgUtgQo306yGtDtM+1Hf9x+OE8chhgXJf0ZyX9nKRv8n3/l3zf/0Pf95u+73/G9/2PyG7/enA4rg1jzFf183ypwpZZABKsXnfb396e2/4wxtg+tLOoQQYBxkD1slhypyDjy5rUf/inbf9Gnf7NCTAAJFzYEOMP2o7fcFJDY8wpSf/J4d1XZSsl+vF/yI7voaT/0ff9h8c18n3/NyStH959vaRyn8+XDmyZBSDhmk23/T144LY/jCm2D+2OKauJ1etiyUeDjC9rUn/lNdf13/7EM483PPpvToABIAXChhifbjs+06XtmyV5h8d3+lls0xhzVtLXHd695fv+73a55Jfbjv/zXp8vNdgyC0AKZDJu+5uYcNsfxhDbh4bHlNXE6Xex5CDIuKfTuqjr+vjDZ/50seTHBEHG6dMEGABSIWyIsdt2fL5L2/bzux1bneyr247/KET7P2w7fl2fz5lsUQOMAEEGgAHL5932NzXltj+MGbYP7Q1TVhMnymLJN1TUWd3Tr+oZSScsllws2l1ICDAApEDYEONjbcfv69L2W9qOf7G34fyp9uDiLSHan247/lKfz5lsYbY/k8KvLl4uRx4SABynWJQ8r3u7MDyPz9SIaBDbh44qpqwmEoslA8DjwoYYVUnBnqR/yRjznx3XyBjzJknffnj3FUm/0Oe4PqNHW7m+xRjzX3Rp/+1txy/3+ZzJFmb7M7bMApAAS0tSNuumr0zG9gf0je1Dw2HKamKxWDIAPC5UiOH7/quSfuDwrpH0z40xX9nexhjzWklX9Gg6x0/4vn9sVYQxZt0Y4x/+eeGY59vX4wHIFWPMWzv09WFJ33R490BSJcx/U+p0W2yLLbMAJMTMjDQ7G70aw/Okc+ek6Wk348KYYvvQ7piymmgslgwAjwtbiSFJPyXp3x4ef4Ok3zLGfJ8x5tuMMd8j6ZN6NJXklqTvjzi2D0v6vcPjt0naNcb8pDHmO4wx32qM+R5jzL/To3BFkl70ff+3Iz5vcp20FdZ1tswCkBwbG9GrMbJZ6epVN+PBmGP70JMxZTXRWCwZAB4XOsTwfb8p6a9K+vjhQ2+RDSp+RtKPSCocPv5JSe/3ff8Pn+ikB77v1yX9V5LuHD70ekkflK32qBw+59zhuVcl/QPf9z8S5TlT4bitsI77YMWWWQCGKJ+3Vfy5XH/X53L2+rNn3Y4LY4ztQztjymqisVgyADyul0oM+b7/+5L+kuwaFNcl/a6kpqT/KBtufEDSnO/7n3UxON/3PylpRtLfkvSvJd2T1JANLb4k6d9J+iFJ7/B9/wc6dDN62rfCOumDFVtmARiiUskuHzA5GX5qiec9WnagVBrk6DCW2D70eExZTTQWSwaAxxnf94c9hqE4f/68//LLo7kGKDBOjDGSpHF9L+vmmcMvXzeGOEe9Xpfm56WtLTu3+7hV9j3PlkyfO2enkFCBgYHoZ+2HmCoxdnakn/opO8R63f6/ksnY38IXi9IHPxjD+jDH/f0c99/fqV2PAUYS3p/S4OZN6d3vjr5kiWSr3Go11ho6CZ8rgGQwxnzC9/3zx53rqRIDAIBe5fPS5qb94Ly4KBUK9suZMfa2ULCP12q2HQEGBiKh24fW69KFC9LTT0urq9Lt248Wcmw27f3VVXv+wgX3O1U8himricRiyQDwOEIMAEAspqely5elW7ekgwPp4UN7e+uWfZwP1hiYhG4fWqnY132tZp/iuColyT7eaDz6DXplkPuwMWU1kVgsGTjezo60vGx/IZLN2l+QZLP2/vKyPY/Rc2rYAwAAABgY19uHOppaUqnYTTz298NfE4QZweYfA1s3pliU7t1z1w6RBYsl9/qaCbBYMkZNMFV1e9v+QqQ9BA6q2O7cka5csZVMGxvuF8nF8FCJAQAARlcCtw+t1+2w+vkyKtnrFhaku3cjD+Vk1ap05oy9ddEOkbBYMmAlsooNsSLEAAAAoyuB24fOz9vfHEZxcCBduhR5KJ0FFSz379vbTlNpjraLMcgYxzLyUsn+d83N2eqKTmGG59nzc3PS7i4BBkZHUMV2UnhxVHsVG0HGaCDEAAAAoyth24fevGnLn8N++O6k1bI7/gzki/rRKTid1gQ5rl0MQUaiFkMdAhZLxrhKTRUbBo4QAwAAjLZOQcbR3TdOaudo8cqVlehVGIFm0/bnVKc1RI4GGSe1G2CQQRn5IyyWjHGTiio2xIIQAwAAjL6EbB9arUavwgi0Wo6zgm6LoAZBxksvdW83gCCDMnJgfKWiig2xIcQAAADjIQHbh7qe2rC357CzMIugNhrSiy+Ga7ew4GxolJED4y3xVWyIFSEGAAAYH8G2oG0BxrELRH5LUYXcPS3/bNHpb+yCtRtcefDAYWdhFkGVwu/2sr4eeUgBysiB8ZboKjbEjhADAACMpWEsEJnJRO+j3cSEw866LYIaVqepOn2ijHzA2EoXKZDoKjbEjhADAKLgwx+QSsNaIDKfj3b9UVNTbvuLHGQ4DjAkysgHKgVb6QJSwqvYEDtCDADoFx/+gFQa5gKRxaLkef1f387znC7X8Ui/QcYAAgyJMvKBSfhWukC7RFexIXaEGACGJ81VDHz4A1Jp2AtELi3ZdTdcyGRsfwNRLErPPx8+yJiclF54wXmAIVFGPhAJ30oXOCrxVWyIFSEGgOFIcxUDH/6A1Br2ApEzM9LsbPRqDM+Tzp2zU1wGoloNtwtJoNGwIUan9/IIKCN3LOFb6QLHSUUVG2JDiAEgfmmuYuDDH5BaSVkgcmMjejVGNitdvRqtj466vc910um9PCLKyB0bxla6aa68RCKkpooNsSDEABCvtFcxDOPDHwAnkrJAZD5vdzPN5fq7Ppez158929/1J+o3wAgMIMigjNyxuLfSTXPlJRIjNVVsiAUhBoD4jEIVQ9wf/gA4k6QFIksl+7//5GT4D+We9+hto1Tq/7k7q5SnSAAAIABJREFUihpgBBwHGZSROxbnVrpprrxE4iS+ig2xIcQAEJ9RqGKI88MfAKeStkBkqWSnpMzN2eqKTl/UPc+en5uTdncHFGBI4d6jpfBBbrkceUgSZeQDEcdWummvvETiJLqKDbEixAAQn1GpYojjwx8A55K4QGQ+L21uSrWatLgoFQr2i7Yx9rZQsI/XarbdQD98h3mPDnYhCdNubc3JsCgjH5BBbqU7CpWXSKREVrEhdoQYAOIzSlUMg/zwB2AgkrxA5PS0dPmydOuWXbfj4UN7e+uWfTyWL97d3teC96/nnuve7to1p/M2KCMfkEFtpTsKlZdIrMRVsSF2hBgA4jVKVQyD+vAHYCBYIDKETu/RR997T2rnOMCQKCMfmEFtpTsqlZdIrERVsSF2hBgA4jcqVQyD+vAHYCBYIPL/Z+/+g9xa7/u+fx6B3NXiyrKv5etobNkkQcs2bC1LqlSXyY4q4ka1ZHETjcaTtTsl5d3YFSnaTtNpmUnj2ORV7NbDuI0zEUfUJjH2hvSMBNup1SxH0rQS1m05DSZXIs2l6a2uCC5t1Yp9nciWdLF3wQue/vHgeHFBHOAszgPg/Hi/Zjj49eDBw728WJwPvuf5htT9Hh303ttr3AgCDB9l5I6NspVumiovMRYbG9L58zZ8mJ62YcT0tL19/nxwS+tYVLFh7AgxAExG0qsYRvnhD8BIsEHkHvgHoQcO9D+I7Bw3wgDDRxm5I+NopZumykuMTL0uzc9LJ05IKyvS5ubu/kXNpr29smIfn593v0EzkokQA8BkJLmKYRwf/gA4xwaRe1QqSVtbgw8i/XFjKk2hjDyicbbSTUvlJUaiUrHvo7Wa/ecU1AK71bKP12p2fKUy3nUifggxAIxfkqsYxvnhD4BzbBCZHpSRD2ncrXSTXnmJkahU7D+dfuFFNz/MWFoiyMg6QgwA45X0KoZxf/gD4BQbRCLzxt1KN8mVlxiJet1+nNreHu7529v2+Q8euF0XkoMQA8D4pKGKYdwf/gA4xwaRyLRxttJNcuUlRubMGVs1FcXOjnT6tJv1IHkIMQCMTxqqGMb54Q/AyLBBJDJtHK10k155iZG4c0e6fTv8KSRBWi3p1q3griVIN0IMAOOTliqGcXz4AzBybBCJTBtlK900VF5iJK5ejV6F4Ws27XzIHkIMIMmqVengQXvpYtyopamKYZQf/gCMFRtEIrNG1Uo3DZWXGIlqNXoVhq/VmvxHW0wGIQaQVP63HA8f2sugbym6x0363T5NVQyj+vAHAMC4jKKVbloqL+Fcve52vvv33c6HZCDEAJKou0wzqNyy17g4BhlJrmIYxYc/AABiYmNDOn/enl41PW1Pt5qetrfPnw/YkyBNlZdwqtl0O9+jR27nQzIQYgBJE3SeaXeQ0W9cnIIMqhgAAIidel2an5dOnJBWVqTNzd0D0GbT3l5ZsY/Pz/f4hj1NlZdwZmrK7Xz797udD8lAiAEkyaCNsvwg4/LlwePiEmRQxQAAQKxUKnYvmFrNfmQI2sOg1bKP12p2fKXSNSBNlZdwolBwO9/hw27nQzIQYgBJEmajrEZDeu65cOOWl92tDQAAJF6lYvfR7BdedPPDjKWlPkEGlZeQ/c8a1NJ6r3I5/plkFSEGkCRhNsqSwu8IvroaeUkAACAd6nX7/cb29nDP3962z3/woOsBKi/Rdu6c3VPFhakpOx+yhxADSJJBG2WFFVTOCQAAMuvMGdteOIqdHen0aTfrQfocOSIdPRq9GiOXk44dowV2VhFiAEkTNcggwAAAAF3u3JFu3w5/CkmQVku6dSugawkg6dq16NUY09PS9etu1oPkIcQAkmjYIIMAAwAA9HD1avQqDF+zaecDeikU7BnSMzPDPX9mxj7/0CG360JyEGIASVUqSRcvhg8y8nnp0iUCDAAA8IRqNXoVhq/VmnwDNMTb4qLdmi2fD39qSS63u6Xb4uIoV4e4I8QAkqpaDdeFxNdo2BBjfX2UqwIAAAlUr7ud7/59t/MhfRYX7WlHc3O2uiIozMjl7ONzc9LduwQYIMRA0lSr0sGDg+P9sOOSqlqVFhbCBxi+RkM6dYogAwAAvEaz6Xa+R4/czod0KhSkmzelWk06e1YqFm3XEWPsZbFo76/V7DhOIYEk7Zv0AoDQOg/cFxaC93boHpe2XuPDBhg+P8hgbwwAANA2NeU2yNi/391cSL/ZWenKlUmvAklBJQaSofvAPaiioNe4hYX0VGREDTB8VGQAAIAOhYLb+Q4fdjsfAPgIMRB/QQfu3Qfi/calJchYXg4XYITZ7LPRkJaWIi8JAAAkX6kUfoPFQXK5dBXBAogXQgzE26DKAz/IuHx58Lg0BBnl8uCAwu9CEmZcuexsaQAAILnOnZOmp93MNTVl5wOAUSDEQLyFqTxoNMJ16Wg07HxJVirZPT6CAop83u51ceHC4HFp2ysEAAAM7cgR6ejR6NUYuZx07Jjd4wAARoEQA/EWpvJACn+Kxepq5CVNXFCQ4QcY/mad/cYRYAAAgC7XrkWvxpielq5fd7MeAOiFEAPxNqjyIKzuA/yk6/65BP39eo0jwAAAAD0UCvb7o5mZ4Z4/M2OfTxtMAKNEiIH4ixpkpC3A8Pk/lwMH+v/9OscRYAAAgD4WF23haj4f/tSSXG634HVxcZSrAwBCDCTFsEFGWgMMX6kkbW0N/vv54wgwAADAAIuL0saGNDdnqyuCwoxczj4+NyfdvUuAAWA8CDGQHKWSdPFi+CDD79KR1gADADAa1ap08ODgjlZhxwEJVChIN29KtZp09qxULNquI8bYy2LR3l+r2XGcQgJgXPZNegFAaNVquC4kvkbDhhjveAdBBgAgnM7W3gsLwdV83eM4XQ8pNTsrXbky6VUAwC4qMZAMnR8W96LRkE6dktbXR7IsAECKdP+uCfod0mvcwgIVGQAAjAEhBuJv2ADDR5ABABgk6HdN9++QfuMIMgAAGDlCDMRb1ADDR5ABAAgy6HeN/zvk8uXB4wgyAAAYKUIMxNvycrgAI8xmn42GtLQUeUkAgJQJ87um0Qi3L1OjYecDAAAjQYiBeCuXBwcUfheSMOPKZWdLAwCkRJjfNVL4UH11NfKSAABAb4QYiLdSye74HvThMp+3O8dfuDB4HDvHAwB6GfS7Jiz/dxIdsQAAGBlCDMRf0IfL7g+L/cYRYAAA+okaZBBgAAAwFoQYSIbuD5dBHxZ7jSPAAACEMWyQQYABAMDYEGIgOfwPlwcO9P+w2DmOAAMAsBelknTxYvggw9+XiQADAICx2DfpBQB7UipJW1vuxgEA0KlaDdeFxNdo2BDjHe8gyAAAYAyoxAAAAJBsgLGwED7A8DUa0qlT0vr6SJYFAAB2EWIAAAAMG2D4CDIAABgLQgwAAJBtUQMMH0EGAAAjR4gBAACybXk5XIARZrPPRkNaWoq8JAAA0BshBgBgTzY2pPPnpWJRmp6WjLGXxaK9f2Nj0isE9qhcHhxQ+F1Iwowrl50tDQAAvBYhBgAglHpdmp+XTpyQVlakzU2p2bSPNZv29sqKfXx+3o4HEsFvzR0UUOTztrX3hQuDx9HaGwCAkSLEAAAMVKlIs7NSrWar5Vut3uNaLft4rWbHVyrjXScwtKAgww8w/Pap/cYRYAAAMHKEGACAvioVe4p/v/Cimx9mLC0RZCBBugOK7gCj3zgCDAAAxoIQA9iLalU6eNBeuhgHxFy9bvc83N4e7vnb2/b5Dx64XRcwMn5AceBA7wCj1zgCDAAAxoYQAwjLb8H38KG9DGqh1z2OIAMJduaMtLMTbY6dHen0aTfrAcaiVJK2toIDjO5xBBgAAIwNIQYQhh9M+C34Gg3p1Kkng4xe4wgykFB37ki3b4c/hSRIqyXdukXXEgAAAERHiAEM0h1M+LqDjH7jCDKQQFevRq/C8DWbdj4AAAAgCkIMoJ+gYMLnBxmXLw8eR5CBhKlWo1dh+Fot/vljcjY2pPPnpWJRmp6WjLGXxaK9nyohAACSgxAD6Gd5OTiY8DUa0nPPhRu3vOxubcCI1etu57t/3+18wCD1ujQ/L504Ia2sSJubtipIspebm/b+EyfsONf/5vshWAEAYDiEGEA/5fJuC71+BgUYkp1ndTXykoBx8Q/2XHn0yO18QD+VijQ7K9Vq/dsD++2AazU7ftQtgeMcrAAAkASEGEA/fgu9MEFGP/l8/1Z9QAxNTbmdb/9+t/MBQSoVaWmpf3jRzQ8zlpZGF2TENVgBACBJCDGAQaIGGQQYSKhCwe18hw+7nQ/opV63Z+5tbw/3/O1t+/wHD9yuK67BCgAASUOIAYQxbJBBgIEEK5WkXM7NXLmcnQ8YtTNnonfV2dmRTp92sx4pvsEKAABJRIgBhFUqSRcvhg8y8nnp0iUCDCTWuXN2o0EXpqbsfMAo3bkj3b4dvatOqyXduuVuc804BisAACQVIQYQVrUarguJr9GwIcb6+ihXBYzMkSPS0aPRqzFyOenYMXtuPzBKV69GDwt8zaadL6q4BisAACQVIQYQRrUqLSyEDzB8jYZ06hRBBhLr2rXo1RjT09L1627WA/RTrUYPC3ytlp0vqjgGKwAAJBkhBjDIsAGGjyADCVYo2E7DMzPDPX9mxj7/0CG360I2bGxI589LxaINw4yxl8Wivb+7KsF1O9L796PPEcdgBQCAJCPEAPqJGmD4CDKQYIuL0uqq3eYl7KkluZwdv7pqnw/sRb0uzc9LJ05IKyvS5qatQpDs5eamvf/ECTvODy/8Ma48ehR9jjgGKwAAJBkhBtDP8nK4ACPMZp9+nzwggRYX7bfec3O2uiIozMjl7ONzc9LduwQY2LtKxe6fUqv1b0fqtx+t1ez4SsVuIOvS/v3R54hjsAIAQJIRYgD9lMuDAwq/C0mYceWys6UB41YoSDdv2oPGs2dtSf/UlC3xn5qyt8+etY/fvMkpJNi7SsVmvf3Ci25+mLG0JH3Hd7hdz+HD0eeIY7ACAECS7Zv0AoBYK5WktbXgU0ryeenGDdtG9fjx/uPW1ux8QMLNzkpXrkx6FUibet0Wv21vD/f87W1bpfC610mPH0dfTy7n5i27ULCnv7jiIlgBACDJqMQABvGDjO5Ki84AY9A4AgwA6OvMmehdPFyEF76pKencuejzlErR2xT7XAUrAAAkGSEGEEZ3QNEdYPQbR4ABAH3duSPdvh29i4cfYrwu4qebXE46dsxWHUV17lz0NsU+V8EKAABJRogBhOUHFAcO9A4weo0jwACAga5ejV6F4TPG/olielq6ft3Neo4ckY4ejV6N4TJYAQAgydgTA9iLUkna2nI3DgCgajV6FYav1ZK++7ul//gfh9tfY2bG7sHscmPaa9ds+BClW7fLYAUAgCSjEgMAAExUve52vpdeklZX7Rl9YSsgcjk7fnXVfWvgQsEGIzMzwz1/FMEKAABJRYgBAAAmqtl0O9+jRzaI2NiQ5uZsCBAUZuRy9vG5OenuXfcBhm9xMV7BCgAASUWIAQAAJmpqyu18+/fby0JBunlTqtWks2elYtG+ljH2sli099dqdtyoKx3iFqwAAJBE7IkBAAAmqlCQNjfdzXf48Gtvz85KV664mz8KP1jZ2LAbmlar0v37tnpk/3679lLJdiFhE08AAJ5EiAEAACaqVJJefNHN5p65XDKaQsUpWAEAIEk4nQQAAEzUuXO2+4YLU1N2PgAAkE6EGAAAYKKOHJGOHg2/4WWQXE46dozTMAAASDNCDAAAMHHXrkWvxpielq5fd7MeAAAQT4QYAABg4goFqVy2XTmGMTNjnz/qDiMAAGCy2NgTAADEgt9KdHlZ2tkJt9FnLmcrMMplWpECAJAFVGIAAIDYWFy07Ufn5mx1RdA+GbmcfXxuTrp7lwADAICsIMQAAACxUihIN29KtZp09qxULNquI8bYy2LR3l+r2XGcQgIAQHZwOgkAAIil2VnpypVJrwIAAMQJlRgAAAAAACARCDEAAAAAAEAiEGIAAAAAAIBEIMQAAAAAAACJQIgBAAAAAAASgRADAAAAAAAkAiEGAAAAAABIBEIMAAAAAACQCIQYAAAAAAAgEQgxAAAAAABAIuw5xDDWjxtj1owxXzHG7BhjvmqM+Zwx5qeNMftGsVBjTM4Y87eMMZ8wxnzZGPNNY8zXjTEvGmM+bYz5e8aYt47itQEAAAAAwOTtKXAwxjwt6bckPdv10Jvbf56V9GFjzAc8z/tDN0uUjDHHJP1LScd6PPwtkr5P0nslfZekv+vqdQEAAIBhbGxIH/uYVK1K9brUbEpTU1KhIJVK0oc/LM3OTnqVAJA8oUMMY8yUpE9Jemf7rj+StCLpy5LeIulvSypKerukTxtj/qrneV+PukBjzF+T9GlJb2zf9e8k/RtJD9q3v6f9uu+L+loAAABAFPW6dOaMdPu2tLMjtVq7jzWb0uam9OKL0vPPS0ePSteu2WADABDOXioxPqzdAOOLkt7ted7X/AeNMR+V9DuS3iPphyT9gqQLURZnjPkrkv432QDjFUnLnud9ImBsTtJ3Rnk9AEibb37zkP74j9+vYpFvAgFg1CoVaXn5yfCiW6slNRpSrWbfg8tlaXFxfOsEgCQLtSdGe5+Ln2/f9CR9sDPAkCTP816R9EFJL7fv+jljzJsiru+fSvLnWAoKMNqv3/I876sRXw8AUqFel+bnpVu3PqavfvVvaHPTBhjS7jeBKyvSiRN2XL0+2fUCQNJVKtLSkg0n+gUYnfwwY2nJPh8AMFjYjT2flfRM+/rnPM/7/V6DPM/7U0l+0DAt6f3DLswYc0DS32rf/D89z/vksHMBQJZUKvabvVpNevz49ZJyPcd1fxPIB2gAGE69biswtreHe/72tn3+gweDxwJA1oUNMX6k4/pnBoztfPy9e1vOa/ykdtf3LyLMAwCZwTeBADB+Z87YU0ii2NmRTp92sx4ASLOwIcbbOq5/YcDYFwKet1f/ecf1zxtj3mSMec4Ys9HRXvX3jTH/zBjz/RFeBwBSgW8CAWD87tyxm3iGDY6DtFrSrVu2qwkAIFjYEKMzJNgaMPYrkvy38bcaY8xeF9V2vH35F5IOSLor6Rdlg5GnZFur/pCkn5V0zxgTaRNRAEg6vgkEgPG7ejX6e6+v2bTzAQCChQ0xvq3j+p/1G+h53quS/Naq+2QDhz0xxkxL+tb2zZakNUlvllSX9A8l/YRstxT/1JWcpMvGmP9mr68FAGnAN4EAMBnVavT3Xl+rZecDAAQLG2K8oeP6KyHGdxYzf0v45fylpzuuf3v79mck/bDneb/sed4nPc+76nnej0o61zH2sjHmLUGTGmM+ZIx5wRjzwksvvTTEsgAgnvgmEAAmw3V3p/v33c4HAGkTNsQYt+51fUPS6XYb19fwPO/jkn67fXNKtkKjJ8/zVjzPO+553vFnnnkmaBgAJA7fBALAZPjtq1159MjtfACQNmFDjG92XH99iPEzHde/EX45gc/5157n/Yc+41c6rr97iNcDgETjm0AAmIypKbfz7d/vdj4ASJuwIcafd1x/U7+Bxph9kt7YvvmqpJeHWNc328/1fXHA+M6OKYeHeD0ASDS+CQSAySgU3M53mE+yANBX2BDjSx3XDw4Y+xbZjTYl6UXP87y9Lqr9nBc77vqLAU/pfPxbA0cBQErxTSAATEapJOVyg8eFkcvZ+QAAwcKGGHc7rh8PHPXk43cDRw12p+P6GwNHWZ3BxaDAAwBSh28CAWAyzp2TpqfdzDU1ZecDAAQLG2J8tuP6ewaMfW/H9c8Ejhrs0x3X/9MBYzsf/1LgKABIKb4JBIDJOHJEOno0+ntwLicdOybNzrpZFwCkVdgQoyrJ70n6bmPMD/caZIz5Tkk/0b75iqRPRVjbpyQ12tc/YIzptxfHhzquRwlOACCR+CYQACbn2rXo78HT09L1627WAwBpFirE8DzvVUm/3L5pJP0rY8zTnWOMMa+X9Lykp9p3fTSoo4gxZtUY47X/XAp4zT+X9D+3b75R0rX2a3TP9SFJP9a++bKkj4X5OwFAmvBNIABMTqEglcvSzMzgsb3MzNjnHzrkdl0AkEb79jD2Y7JhwTslvV3S7xljPi7py7Kbef6UpGJ77D1Jv+Rgfb8i6Udl99n4UUl3jTG/3n7NpyW9v32/76zneS89MQsAZMC1azZ8aDQGjw3CN4EAMJzFRXu5vCzt7Eit1uDn5HL2fbdc3n0+AKC/sKeTyPO8pmxo8Pn2Xd8jG1R8QtKvajfA+KKkH/U8L/IGm57nNSS9T9J6+67DshUhn5R0VbsBxrakJc/zfiPqawJAUvFNIABM1uKitLEhzc3Z99Sg6rhczj4+NyfdvUuAAQB7ETrEkCTP874m6d2y+17ckPTHkpqS/kQ23PiQpDnP8/7Q1QLblRXPSlqU3Sfjj9qv+ReSbkn6nyR9n+d5z7t6TQBIqsVFaXVVyufDn1qSy9nxq6t8kAaAqAoF6eZNqVaTzp6VikW715Ax9rJYtPfXanYcwTEA7I3xPG/Sa5iI48ePey+88MKklwEgImOMJCmr72VB6nXpzBnp3/7bV/T48T71Onswl7MfqI8ds6eQ8EEawDidPHlSkrS+vj7RdQCd+FwBxIMx5gue5x3v9dieKjEAAMngfxN47NiH9V3f9W/4JhAAAACpsJeNPQEACfOGNzzQW9/6T7W+/oFJLwUAAACIjEoMAAAAAACQCIQYAAAEqValgwftpYtxAAAAiIQQAwCAXqpVaWFBevjQXgZtPtg9jiADAABgZAgxACDpqBZwzw8mGg17u9GQTp16MsjoNY4gAwAATNDGhnT+vN3IfXrabuw+PW1vnz9vH08yQgxkAwd5SCuqBdzrDiZ83UFGv3H8jAEAwJjV69L8vHTihLSyIm1uSs2mfazZtLdXVuzj8/N2fBIRYiD9OMhDWoWsFjj6ta9RLRBWUDDh83/Gly8PHsfPGAAAjEmlIs3OSrWa/RjSavUe12rZx2s1O75SGe86XSDEQLpREo60ClktcPRrX9Ov3L1LtUBYy8vBwYSv0ZCeey7cuOVld2sDAADooVKRlpb6hxfd/DBjaSl5QQYhBtKLknCk1R6qBX7l7l29/vHj4HH8G3+tclnK5wePGxRgSHae1dXISwIAAAhSr9vvTLa3h3v+9rZ9/oMHbtc1SoQYSCdKwpFme6gWCAwwOsdRLbCrVJLW1sIFGf3k89KNG9LJk06WBQAA0MuZM9LOTrQ5dnak06fdrGccCDGQTpSEI82oFhitqEEGAQYAABiDO3ek27fDn0ISpNWSbt1KTtcSQgykEwd5SDOqBUZv2J8xP1MAADAmV69Gr8LwNZt2viQgxEA6cZCHtKNaYE+G6pdeKkkXL4b/Gefz0qVLmfmZAgCAyapWo1dh+Fqt5JxBT4iB9OIgD2lHtcBAkfqlV6vhTjnzNRo2xAhq4wwAAODQaz63OHD/vtv5RoUQA+nGQR7SjmqBQJH6pQ/aHDhIUBtnAKk3VMUXAETgfzHjyqNHbucbFUIMpB8HeUgzqgV6itIv/dfPVPXqe4cIMHwEGXvGwR+SLFLFFwBEMDXldr79+93ONyqEGEg/DvKQVlQL9BSlX/pJVfXbzQXtaw4ZYPhS/jN2hYM/JF2kii8AiKhQcDvf4cNu5xsVQgykGwd5SKth/237UvxvPEq/9LKW9ZRCdi0apNGw5SDoiYM/JF2Uiq+lJf4tA4iuVJJyOTdz5XJ2viQgxEB6cZCHtIr6b9uXwn/jUfulL6uslzUgoPBPORsUZOTztt0znsDBH5IuSsWXZJ+3vCw9eOB2XQCy5dw5e/qlC1NTdr4kIMRAOnGQhzRbXg73bzuD1QJR+6Wvq6QFrQUHGf6mvxcu9N80OJ+3jyflK40x4uAPaRCl4su3syOdPu1mPQCy6cgR6ejR6NUYuZx07JiteEwCQgykEwd5SLNyOVwVwKVLeuV1A97mU1Yt4KJfemCQ0d21KKj7EQFGXxz8IemiVnz5Wi3p1i02rgUQzbVr0asxpqel69fdrGccCDGQTns4yKMkHIkzqHVwR7XA33/b24KDjBQebLva+PGJICOo7XL3f4sU/kxd4uAPaRC14qtTs2nnA4BhFQr2UGVmZrjnz8zY5x865HZdo0SIgXTaw0EeJeFIpH5VAB0H27efflp//21vy0y1gMt+6esq6W9oTTpwoHeA4fP/Wxw4kMqfqUsc/CENXFR8+VotOx8ARLG4KK2u2o93YU8tyeXs+NVV+/wkIcRAeoU8yKMkHInVqwqgx8H27aefzky1gOt+6TenStLWVnCA4Su1x6XwZ+oSB39IA9etfu/fdzsfgGxaXLQVinNztroiKMzI5ezjc3PS3bvJCzAkQgykXciDPErCkVidVQBUC2S2X3pScPCHNHBZ8SVJjx65nQ9AdhUK0s2bti352bNSsWi/4DHGXhaL9v5azY5L0ikknQgxkH4c5MVbtSodPDj4K9Ww47KoRLWAL6v90pMiUwd/vLelluuKr/373c4HALOz0pUr0r179jTOx4/t5b179v6kdCEJQoiBbOAgL578VrgPH9rLoFa23eP4sI8AWe2XnhSZOfjjvS3VqPgCgMkixAAwGf6Hd78VbqMhnTr15If9XuP4sI8AWe2XnhSZOPjjvS31qPgCgMkixAAwft0f3n3dH/b7jePDPgJksV96UqT+4I/3tkyg4gsAJosQA8B4BX149/kf9i9fHjxuYUEnR7ZQJFUW+6UnRaoP/hy/txFkxBcVXwAwWYQYiD82R0uX5eXgD+++RkN67rlQ48ruVoYUyVq/9KRI9cGf4/c2LS+7Wxuco+ILACaHEAPxFrfN0QhUoiuXd1vZ9jPoQ74k5fNairwgpFWW+qUnSWoP/hy/t2l1NfKSMDpUfAHA5BBiIL7itjla3AKVpPJb2Yb5sN9PPi/duKHfdbMqpFT5KNqjAAAgAElEQVRW+qUnSdSDv9e/Xpqfl973PhtmGGMvi0Xp/HkbXE2E4/e2gd20MHFUfAHAZBjP8ya9hok4fvy498ILL0x6GQjS79zizg94g8atrbnZ+a3X6/T6oBk0ztU60mTQ+eP9dPzsjTGSpKy+lw1ysv3vcz0odAMmpFKxZ0zs7Eit1uDxuZzkedK+fXZ8r+fkcjbQOHrUVny47oYSiqP3tixIy/tTvS6dOSPduiU1m8H/Nqem7GlQ168TmMYZnyuAeDDGfMHzvOO9HqMSA/ETt83R2G1+NIb91jJjH/KBtNrL6T5TUzbAkIIPEiV7f6NhK2tmZ21QMna8t2UOFV8AMF6EGIifOG2OFrdAJW1KJenixfAf9vN56dIlPuQDKRHm4O+v/3V7+/Fj+ycMP8xYWppgkMF7W+bMzkpXrkj37tkKo8eP7eW9e/b+WG1ECwAJxukkiJ8opbjdon6zdfCg3dsizOuEWe+BA9LW1nBrSaNh/lt3/Tel7LO/tJRrI5vqdXvgF+XXQT5vN2wd67ffDt7bsoD3J8QRnyuAeOB0EiRLnDZHY7f50Rk2rAra4BVA6pw5Y7/JjmJnRzp92s16QuG9DQCAkSLEQDxFDTJcfaMVp0AlTaJW2/BhH0i9O3ek27fDbfrZT6tlN1wcS9cS3tsAABg5QgzEV1w2R4tLoJIWrk4Xan/Yf5ebVQGImatXo1dh+JpNO99IOX5vI8gAgPTa2LBtwYvFmLULTwhCDMRbXDZHi0ugkgZhNm6VQp/Gsxp5QQDiqFqNXoXha7XGsK+y4/c2LS1FXhIAIF7qdWl+XjpxQlpZkTY3bdAu2cvNTXv/iRN2XL0+2fXGFSEG4q1aDdeFxNdo2BBjFN9gxSVQSbow+4z4P7sQ4yL0ngEQY64/uN2/73a+Jzh+b1O57GxpAIDJq1TsZtW1mj1kiXW78JgjxEB8xW1ztDgFKkk2qKrFr165cGHwuLU1rY9soW3Vqu1SM+hr3LDjAITifzPlyqNHbud7guP3NpVKo1srAGCsKhVbYNcvvOg28XbhMUaIgXiK2+ZocQtUki7ow3736Tf9xo3jQ77/3/3hQ3sZ9N+xexxBBhDZ1JTb+fbvdztfT0l5bwMAjE29bs843N4e7vnb2/b5Dx64XVeSEWIgfuK2OVrcApW06P4QH7R/SK9x4www/P/uQf8de40jyAAiKxTcznf4sNv5AsX9vQ0AMFaJbBcec4QYiJ84bY4Wt0AlbfwP8QcO9N8AtXPcJAIMX/d/x37jCDKASEolKZdzM1cuN+ZsIK7vbQCAsUpku/AEIMRA/MRpc7Q4BSppVSpJW1uDN0D1x00qwPD5Qcbly4PHEWQAQzt3zrabc2Fqys43VnF7bwMAjF3i2oUnBCEG4idOm6PFKVDBeIQJrhqNcJu8Nhp2PgB7duSIdPRo9GqMXE46dszu8A4AwDglrl14QhBiIJ7isjlanAIVjEeY4EoKX6Gzuhp5SUBWXbsWvRpjelq6ft3NegAA2IvEtQtPCEIMxFdcNkeLS6CC8RgUXIUV9O8VQGiFgs0VZ2aGe/7MjH3+oUNu1wUAQBiJaxeeEIQYiLe4bI4Wl0AF4xE1yEhTgFGtSgcPDq5fDDsO2KPFRVvQlM+HP7Ukl9sthFpcHOXqAAAIlsh24QlAiIH4i8vmaHEJVDAewwYZaQswFhakhw/tZVB3ne5xBBlwbHHR7sg+N2erK4LCjFzOPj43J929S4ABAJisxLYLjzlCDGAv4hKoYDxKJenixfBBhr/Ja5oCDH/vj6A2wb3GEWRgBAoF6eZNqVaTzp6VikX7DZcx9rJYtPfXanYcp5AAACYt0e3CY4wQAwCCVKvhupD4Gg0bYgRVLCRFUJvZ7iCj3ziCDIzI7Kx05Yp0755tW/f4sb28d8/eTxcSAEBcJL5deEwRYgBAL0EH6IMEVSwkxaC/t//3u3x58DiCDAAAkGG0Cx8NQgwA6DZsgOFLcpCxvDz4791ohKtQaTTsfAAAABlFu3D3CDEAoFPUAMOX1CCjXA63B0iYn4/fHgJActCRCACcol24e4QYANApTCWCFP5Af2kp8pLGKmp7WV+aurQAWUFHIgAYCdqFu0WIAQCdwlQi+F1Iwowrl50tbWyiBhkEGEDy0JEIAEaKduHuEGIAQKdBB/D+AfqFC4PHra0ltxfWsEEGAQaQPHQkAoCxoF24G4QYANAt6AC++wC937gkBxi+Ukm6eDF8kOFXqBBgAMlBRyIAGDvahUdDiAEAvXQHFEEVBr3GpSHAkOzBSJguJL5Gw4YYSdvMFMgyOhIBABKGEAMAgvgBxYED/U+R6ByXpgBjmC4tSe3KAmQVHYkAAAlDiAEA/ZRK0tbW4FMk/HFZDjB8BBlActCRCACQMIQYAIBdUQMMH0EGkBx0JAIAJAghBgBgV5jz46Xw5edLS5GXBGAM6EgEAEgIQgwAwK4w58f7XUjCjCuXnS0Nw9vYkM6ft63bpqdtK7fpaXv7/Hn7OEBHIgBAEhBiAAB2Dfo21v/W9cKFwePSsslpgtXr0vy8dOKEtLIibW5KzaZ9rNm0t1dW7OPz83Y8MoyORACABCDEAAC8VlCQ0V023m8cAcbEVSq2z3ytZo81W63e41ot+3itZsdXKuNdJ2KCjkQAgIQgxAAAPKk7oAg6773XOAKMiatU7HYk/cKLbn6YsbREkJE5dCQCACQIIQYAoDc/oDhwoP/GfZ3jCDAmrl63+7Nubw/3/O1t+/wHD9yuCzFFRyIAQMLsm/QCAAAxVipJW1vuxmHkzpyRdnaizbGzI50+Ld286WZNiLG9dCQaNM4v5eG9AAAwQlRiAACQEnfuSLdvhz+FJEirJd26RdeSTKAjEQAgYQgxAABIiatXo1dh+JpNOx9Sjo5EAICEIcQAACAlqtXoVRi+VsvOhwygIxEAIEEIMQAASIl63e189++7nQ8xRkciAEBCEGIAAJASzabb+R49cjsfYo6ORACABKA7CQAAKTE15TbI2L/f3VxICDoSAQBijkoMAABSolBwO9/hw27nAwAAiIoQAwCAlCiVpFzOzVy5HGcJAACA+CHEAAAgJc6dk6an3cw1NWXnAwAAiBNCDAAAUuLIEeno0ejVGLmcdOyYNDvrZl0AAACuEGIAAJAi165Fr8aYnpauX3ezHgAAAJcIMbKuWpUOHrSXLsYBACaqUJDKZWlmZrjnz8zY5x865HZdAAAALhBiZFm1Ki0sSA8f2sv19XDjCDIAYOI2NqTz56Vi0VZOGGMvi0X7dv6Rj0j5fPhTS3I5O351VVpcHOXKAQAAhrdv0gvAhPjBRKNhbzca0qlT0o0b0smT/cctLEhra2xbDwATUK9LZ85It29LOztSq7X7WLMpbW5KL75oA40f/EF7/x/8gX2sc6wvl7ObeB47Zk8hoQIDAADEGZUYWdQdTPj8IMOvyOg3jooMABi7SsVutlmr2bfiXqGEZO9vNKTf+z0banzkI9LZs7ZKY2rKVm1MTdnbZ8/a+W7eJMAAAADxR4iRNUHBhM8PMi5fHjyOIAMAxqZSkZaW+ocX3fww4xd/UXrXu6R792z1xuPH9vLePenKFbqQAACA5CDEyJrl5eBgwtdoSM89F27c8rK7tQEAeqrX7dvt9vZwz9/ets9/8MDtugAAAMaNECNrymW7c9sggwIMaXcHOADASJ05YysnotjZkU6fdrMeAACASSHEyJpSyW7KGSbI6Ceff3ITUACAc3fu2E08w55CEqTVkm7dsl1NAAAAkooQI4uiBhkEGAAwNlevRq/C8DWbdj4AAICkIsTIqmGDDAIMABirajV6FYav1WI/ZgAAkGyEGFlWKkkXL4YPMvJ56dIlAgwAGKN63e189++7nQ8AAGCcCDGyrFoN14XE12jYEGN9fZSrAgB0aDaHf+5JVfVAB3VSu+UXjx71GFitSgcPUqYBAABijxAjq6pVaWEhfIDhazSkU6cIMgBgTKamhnveSVW1pgUd1EOtaUHv0rokaf/+roH+74OHD+0lQQYAxMrGhnT+vFQsStPTkjH2sli097NhM7KGECOLhg0wfAQZADA2hcLen+MHGE/Jvs8/pYZu6JTepXUdPtwxsPv3QaNBkAEAMVGvS/Pz0okT0sqKtLm5W53XbNrbKyv28fl596cfAnFFiJE1UQMMH0EGAIxFqSTlcuHHdwcYPj/I+ND3r9s7gn4fEGQAwMRVKtLsrFSr2bfloA2eWy37eK1mx1cq410nMAmEGFmzvBwuwAiz2WejIS0tRV4SACDYuXO2bDiMoADD95Qa+jufPSVdvtw/0CbIAICJqVTsR+x+4UU3P8xYWiLIQPoRYmRNuTw4oPC7kIQZVy47WxoA4ElHjkhHj4arxihrOTDA8L3ulUa4TZ0bDRt8AwDGpl63b73b28M9f3vbPv/BA7frAuKEECNrSiVpbS04oMjnpRs3pAsXBo9bW7PzAQBG6tq1cNUYyyrrZYWspBskn5dWVwePAwA4c+aMtLMTbY6dHen0aTfrAeKIECOLgoIMP8A4eXLwOAIMABibQsEWvs3M9B+3rpIWtBYuyOin+/cBAGDk7tyRbt8OfwpJkFZLunWLriVIL0KMrOoOKII+sPYaR4ABAGO3uGgLI/L5/qeWRA4yCDAAYCKuXo1eheFrNu18QBoRYmSZH1AcOND/A2vnOAIMJBD91ZEWi4v23+vcnK3KCAoz/q9cST82tabt1+0xyCDAAICJqVajV2H4Wi32ZkZ67Zv0AjBhpZK0teVuHBAj9bo9t/T2bfvNRucHA7+/+osvSs8/bzdOvHbNlu0DcVYoSDdv2jDj6lX7IfX+fenRI2n/funwYfuWfe5cSTOfvhhuE09pd1NnAgwAmIh63e189++7nQ+IC0IMAKlUqdjdubvDi27d/dXLZfttNxB3s7PSlSt9BlSr4QMMyY67dEl6xzsIMgBgAppNt/M9euR2PiAuOJ0EQOrQXx2ZV61KCwvhAwxfoyGdOiWtr49kWQCAYFNTbufbv9/tfEBcEGIASBX6qyPzhg0wfAQZADARrk9pPXzY7XxAXBBiAEgV+qsj06IGGD6CDAAYu1Kpf/epvcjl2Isf6UWIASA16K+OzFteDr+J5yD++VUAgLE4d852T3NhasrOB6QRIQaA1KC/OjKvXB4cUPhdSMKMK5edLQ0A0N+RI7ZbWtRqjFxOOnbMbgANpBEhBpKrWpUOHhzcBDvsuLTLwM+L/urIvFJJWlsLDijyeenGDenChcHj1taoRQaAMbt2LXo1xvS0dP26m/UAcUSIgWTyz/t++NBeBp233T0uq0elGfl50V8dUHCQ4QcYfvvUfuMIMABgIgoFWwQ3MzPc82dm7PMPHXK7LiBOCDGQPN0b1wVtQNdrXAIPzCPL0M+L/upAW3dA0R1g9BtHgAEAE7W4KK2u2rfksKeW5HJ2/OqqfT6QZoQYSJagnfe7D8z7jUvYgXkkGft50V8d6OAHFAcO9A4weo0jwACAWFhctBuMz83Z6oqgMCOXs4/PzUl37xJgIBsIMZAcg1oH+gfmly8PHpegA/OhZfDnRX91oEupJG1tBQcY3eMIMAAgNgoF6eZNqVaTzp6VikX7hY0x9rJYtPfXanYcp5AgK/ZNegFAaGFaBzYa0nPPhRu3vGw/tKdVBn9epZL04otuNvekvzoAAIiD2VnpypVJrwKIjz1XYhjrx40xa8aYrxhjdowxXzXGfM4Y89PGmJEHI8aYbzfG/HtjjNfx5+CoXxcTFqZ1oDT4gFzaPWkwzTL486K/OgAAAJBuewoxjDFPS/o/JH1C0ilJ3y1pStKbJT0r6Z9LqhljvtfxOrv9L5L+yohfA3EzqHVgWEEb3KVNBn9e9FcHAAAA0i10iGGMmZL0KdmwQpL+SNIvSPovJV2Q9Aft+98u6dPGmDc6XGfnOv4LST8p6bGkV0bxGoixqAfmCTogdyKDPy/6qwMAAADptZdKjA9Lemf7+hcl/See5/2S53mf8DzvV2XDi8+2H/8h2YDDKWPMU5I+3r55RdKfuH4NJMCwB+YJPCB3ImM/L/qrAwAAAOkVKsRo73Px8+2bnqQPep73tc4xnue9IumDkl5u3/Vzxpg3uVpo2y9JOiTp/+tYD7KoVJIuXgx/YJ7PS5cuJe6A3JmM/bzorw4AAACkU9hKjGclPdO+/jnP836/1yDP8/5Udr8MSZqW9P5oy9tljPnPJP2d9s2f8zzvG67mRgJVq+G6avgaDXtQvr4+ylXFVwZ/XvRXBwAAANInbIjxIx3XPzNgbOfj793bcnozxuyX9C9l1/spz/P+VxfzIqGqVWlhIfwBua/RkE6dSvSB+VAy/POivzoAAACQLmHbob6t4/oXBox9IeB5UfwP7bm+KelnHc2JJBr2gNznH5gncK+HofDzkkR/dQAAACAtwlZifH/H9a0BY78iqdW+/lZjjNnrojoZY4qS/kH75s97nveVKPMhwaIekPtSUGEQCj8vAAAAACkTNsT4to7rf9ZvoOd5r0r6evvmPklPDbEuSZIx5nWS/oXs/hovSProsHO15/uQMeYFY8wLL730UpSpMAnLy+EOyMNsXtloSEtLkZcUa/y8AAAAAKRM2BDjDR3XXwkxfrvj+reEX84TfkbSX5Ot7PiQ53mPI8wlz/NWPM877nne8WeeeWbwExAv5fLgA26/q0aYceWys6XFEj8vAAAAACkTNsQYO2PM90r6H9s3f83zvFuTXA9ioFSS1taCD7jzebt3w4ULg8etrdn50oyfFwAAAICUCRtifLPj+utDjJ/puD5sK9SPyVaAPJR0ccg5kDZBB+b+Abm/+WS/cVk6IOfnBQAAACBFwoYYf95x/U39Bhpj9kl6Y/vmq5Je3uuijDGnJb2vffNnPM/b8xxIse4D7u4D8n7jsnhAzs8LAAAAQEqEbbH6JUmH2tcPylZHBHmLpFz7+oue53lDrOun2pf/XtIxY8yxgHHf2nH9Z40xftjyzz3P+5MhXhdJ4R9wLy9Lq6vB7T87x5XL2T0g5+cFAAAAIAXChhh3Jb2nff24pN/tM/Z41/OG4bdlfbOkfxTyOf9dx/U1SYQYaVcqSVtb7salHT8vAAAAAAkX9nSSz3Zcf0/gKOu9Hdc/s7flAAAAAAAA9BY2xKhKeql9/d3GmB/uNcgY852SfqJ98xVJnxpmUZ7nnfQ8zwz6o9ee1nKo47Hbw7wuAAAAAACIr1Ahhud5r0r65fZNI+lfGWOe7hxjjHm9pOclPdW+66Oe5/2HXvMZY1aNMV77z6WhVg4AAAAAADIl7J4Ykm15+mOS3inp7ZJ+zxjzcUlflt3M86ckFdtj70n6JYfrBAAAAAAAGRc6xPA8r2mMeb+k35L0rKTvUe+g4ouSPuB53l+4WSIAAAAAAED4PTEkSZ7nfU3Su2X3vbgh6Y8lNWU7gXxe0ockzXme94eO1wkAAAAAADJuL6eTSJI8z/MkfbL9Zyie5y1JWhr2+R3zHIw6BwAAAAAASIY9VWIAAAAAAABMCiEGAAAAAABIBEIMAAAAAACQCIQYAAAAAAAgEQgxAAAAAABAIhBiAAAAAACARCDEAAAAAAAAiUCIAQAAAAAAEoEQAwAAAAAAJAIhBgAAAAAASARCDAAAgA4bG9L581KxKE1PS8bYy2LR3r+xMekVAgCQXYQYAAAAkup1aX5eOnFCWlmRNjelZtM+1mza2ysr9vH5eTseADoRggKjR4gBAAAyr1KRZmelWk1qNKRWq/e4Vss+XqvZ8ZXKeNcJIJ4IQYHxIcQAAACZVqlIS0v9w4tufpixtESQAWQdISgwXoQYAAAgs+p1aXlZ2t4e7vnb2/b5Dx64XReAZCAEBcaPEAMAAGTWmTPSzk60OXZ2pNOn3awHQHIQggKTQYgBAAAy6c4d6fbt8N+eBmm1pFu32LAPyBpCUGAyCDEAAEAmXb0a/QDE12za+QBkAyEoMDmEGAAAIJOq1egHIL5Wy84HIL0626ceO2b3tXCBEBTYm32TXgAAAMAkuG5xeP++2/kAxEO9bk8duX3bVm+5Cj99hKDA3lCJAQAAMqnZdDvfo0e97+/89nZ6WjLGXhaL9n7KyIH4Cts+NSpCUCA8KjEAAEAmTU25DTL273/t7X7f3jab0uam9OKL0vPPS0ePSteuSYWCu/UAiMZvnzps95G9CApBATyJSgwAAJBJrgODw4d3r4f99rbVso/XanZ8peJ2TQCGE7V96l51h6CuUAmGNKISAwAAZFKpZCshXJSH53J2Pmm4b2/9MGNpyd5eXIy+JgDDc9E+dS86Q1AXqARDmlGJAQAAMuncOfuNpAtTU3a+qN/ebm/b5z944GZdAIYx66R9alidIagLVIIh7QgxAABAJh05Yr+BzOWizZPL2XaLs7Nuvr3d2ZFOn442B4Aozo61CsMPQV3wK8H2sglpZyUYQQaSgBADAABk1rVr0asxpqel69elO3fk5NvbVku6dYtz1YHJeXasVRh+CBoVlWDICkIMwFetSgcPDm7UHXYcACD2CgWpXJZmZoZ7/syMff6hQ9LVq+7OoW827XwAJmF8m0P4IagLVIIhKwgxAMkGEgsL0sOH9nJ9Pdw4ggwASLzFRWl1Vcrnw59aksvZ8auru5twVqvuzqFvtfgVA0yOo81yBugMQaOiEgxZQogB+MFEo2FvNxrSqVNPBhm9xhFkAEAqLC7aD+1zc/bAIijMyOXs43Nz0t27r+0iUq+7XdP9+27nAxDWaDfE6BWCRkUlGLKEEAPZ1h1M+LqDjH7jCDIAIBUKBenmTbtT/9mzUrFoN9wzxl4Wi/b+Ws2O6/72tNl0u55Hj9zOByAsx4lkW78QNCoqwZAl+ya9AGBigoIJnx9kXLwoPfdc/3ELC9Lamtv+WACAiZidla5c2fvzpqbcBhn797ubC8BefF65XNFZKPC610k/8AP2Y+K5c2428exGJRiyhBAD2bW8HBxM+BqN/gFG57jlZWlry9nyAADJUihIm5vu5jt82N1cAPbi45qe/pmBH//CmJmx1VujCC46UQmGLOF0EgRLe7eOctmekDhImN9g/omNAIDMKpXCbww6SC5HcR8wORs6ejT6/88u26cOMjXldj4qwRBnhBjoLQvdOkolewpImCCjn3xeunFDOnnSybIAAMl07pxtl+jC1JSdD8BkXLsW/f9nl+1TByk47gpLJRjijBADT8pSt46oQQYBBgCg7cgRJe7bWwC9FQq2aHdmZrjnu2yfGgaVYMgSQgy8Vha7dQwbZBBgAAC6JO3bWwDBFhft2cL5fPiAYBTtU8OgEgxZQoiBXWG7dVy+PHhcEoOMixfDBxn5vHTpEgEGAOA1kvbtLYD+FheljQ3bFnVmJjjMGGX71DCoBEOWEGJg1yi6dSRFtRru7+VrNGyIEbRXCAAgs5L07S2AwQoF6eZN22Xk7FmpWLTVCsbYy2LR3l+r2XGTCiGpBENWEGJgV1a7dQyqQAkStFcIACDzkvLtLYDwZmelK1eke/eknR3p8WN7ee+evX/S1QtUgiErCDGwK4vdOoYNMHwEGQCAAEn59hZAelAJhizYN+kFIGb8IGPYA/ssBRg+P8hIyt8bADBW/re3ADAOi4vS8ePSmTPSrVtSsym1Wk+Oy+VsoHrsmD2FhCAVSUElBp6UlW4dYfYAkcKfYrO0FHlJAAAAQFRUgiHNCDHQWxa6dYTZA8T/e4UZVy47WxomZ2NDOn/e/nKfnra/7Ken7e3z5+3jAAAASRD3fTyAYRBioLcsdOsYVHHiV5ZcuDB43NqanQ+JVa9L8/PSiRPSyoq0uWnLLyV7ublp7z9xwo6r1ye7XgAAACCLCDHwpCx16wgKMrpPjek3jgAj8SoV+01ErWb/Gfc6b1Sy9zcadtzsrH0eAAAAgPEhxMBrZbFbR3dAEbS3R69xBBiJV6nY7Uz6hRfd/DBjaYkgAwAAABgnQgzsct2tI4lBxoED/Tcn7RxHgJF49brd33V7e7jnb2/b5z944HZdAAAAAHqjxSp27aVbx6Bx/tfUW1suVjYepVK49YYdh9g7c8ZubhXFzo50+rTd2RsAAADAaFGJgV1060CG3Lkj3b4d/hSSIK2W7cFO1xIAAABg9AgxsItuHciQq1ejV2H4mk07HwAAAIDRIsTAa9GtAxlRrUavwvC1WnY+AAAAAKNFiIEn0a0DGVCvu53v/n238wEAAAB4EiEGeqNbB1Ku2XQ736NHbucDAAAA8CS6kyAY3TqQYlNTboOM/fvdzQUAAACgNyoxAGRSoeB2vsOH3c4HAAAA4EmEGAAyqVSScjk3c+VynEkFAAAAjAMhBoBMOndOmp52M9fUlJ0PQETVqnTw4OB2P2HHAQCA1CHEAJBJR45IR49Gr8bI5aRjx6TZWTfrAjKrWpUWFqSHD+3l+nq4cQQZAABkivE8b9JrmIjjx497L7zwwqSXsSf79u1Tq9Wa9DKAFDkkaUPSUxHmeFnS2yRtuVgQkEknJa3ptf8nvizplKTfDTFuQdL6KBcIAECC5XI5vfrqq5Nexp4YY77ged7xXo9RiQEgwx5IWpLUGPL5jfbzt9wsB8igk3oymFD79g1J7woxbq39OAAASD9arCZI0tIzYByMMZKkKFVllYq0vCzt7Ehhip1yObufRrmc1+Libw79uuNw8uRJSdJ6UGk+MEn+qSGN3kHiU5LW83np4kXpuef6jqvm89LaGrvsJgjvT4gjF58rAIwWlRgAMm9xUdrYkObmpJmZ4H0ycjn7+NycdPeufR6ACJaXA4OJv9Ro9A0wXjNuednd2gAAQCwRYgCApEJBunlTqtWks2elYtF2HTHGXhaL9v5azY47dGjSKwZSoFyW8vnB4wYFGJKdZ3U18pKA1KL7D4CU4HQSAOgwOytduTLpVQAZUSrZU0D6nFISSj4v3bghtU9PANCl81j3VwoAACAASURBVNSthYXg/1+6x3GKFoAYohIDAABMjh9khKnI6IUAA+ive++ZRkM6derJNsa9xtHGGEAMEWIAAIDJGjbIIMAA+gvaPLc7yOg3jiADQMwQYgAAgMkrlWwXkrBBRj4vXbpEgAEEGdD95y+DjMuXB48jyAAQI4QYAABg8qrVcF1IfI2GDTFozwn0RvcfAClFiAEAACZr0DfGQYLO7Qcw3u4/dD4BMEaEGAAAYHKGDTB8BBlAb1E3zfUN2nvG/3/44UN7GfT/Yvc4ggwAQyLEAAAAkxE1wPARZAC9jbr7D51PAEwAIQYAAJiMMOfsS+FL4peWIi8JSJ1Rdf+h8wmACSHEAAAAkxHmnH2/C0mYceWys6UBqeK6+w+dTwBMECEGAACYjEHfEPvfBF+4MHjc2pqdD8CTXHf/ofMJgAkixAAAAJMTFGR0l7L3G0eAAQQbRfefcXY+AYAuhBgAAGCyugOKoHPxe40jwACCjar7z7g6nwBAD4QYAABg8vyDogMH+h/UdI4jwACCjbr7z6g7nwBAgH2TXgAAAIAke1C0teVuHJBle+n+E2bfiqWlJ/+/84OMvYYlBBgAIqASAwAAAEibcXX/cd35BAAGIMQAAAAA0mZc3X9cdz4BgAEIMQAAAIA0GnX3n1F0PgGAAQgxAAAAgLQaVfefUXU+AYABCDEAAACANHPd/WfUnU8AoA+6kwAAAABp57L7zzg6nwBAACoxAAAAAIQ3rs4nANADIQYAAACA8MbV+QQAeiDEAAAAALA3o+58AgABCDEAAAAA7N2oOp8AQB+EGAAAAACG47rzCQAMQHcSAAAAAMNz2fkEAAagEgMAAAAAACQCIQYAAEDaVKvSwYP20sU4AABighADAAAgTapVaWFBevjQXq6vhxtHkAEASABCDAAAgLTwg4lGw95uNKRTp54MMnqNI8gAACQAIQYAAEAadAcTvu4go984ggwAQMwRYgAAACRdUDDh84OMy5cHjyPIAADEGCEGAABA0i0vBwcTvkZDeu65cOOWl92tDQAAhwgxAAAAkq5clvL5weMGBRiSnWd1NfKSAAAYBUIMAACApCuVpLW1cEFGP/m8dOOGdPKkk2UBAOAaIQYAAEAaRA0yCDAAAAlAiAEAAJAWwwYZBBgAgIQgxAAAAEiTUkm6eDF8kJHPS5cuEWAAABKBEAMAACBNqtVwXUh8jYYNMdbXR7kqAACcIMQAAABIi2pVWlgIH2D4Gg3p1CmCDABA7BFiAAAApMGwAYaPIAMAkACEGAAAAEkXNcDwEWQAAGKOEAMAACDplpfDBRhhNvtsNKSlpchLAgBgFAgxAAAAkq5cHhxQ+F1Iwowrl50tDQAAlwgxAAAAkq5UktbWggOKfF66cUO6cGHwuLU1Ox8AADFEiAEAAJAGQUGGH2CcPDl4HAEGACDmCDEAAADSojug6A4w+o0jwAAAJAAhBgAAQJr4AcWBA70DjF7jCDAAAAmxb9ILAAAAgGOlkrS15W4cAAAxQSUGAAAAAABIBEIMAAAAAACQCIQYAAAAAAAgEfYcYhjrx40xa8aYrxhjdowxXzXGfM4Y89PGGGf7bBhjvtcYc9YY8xvGmLvGmK8bY5rGmJeMMf+3MeYjxpjvdfV6AAAAAAAgvvYUOBhjnpb0W5Ke7Xroze0/z0r6sDHmA57n/WGUhRljfkfS35Rkejz8He0/85L+njHmH3qe96tRXg8AAAAAAMRb6BDDGDMl6VOS3tm+648krUj6sqS3SPrbkoqS3i7p08aYv+p53tcjrO1t2g0w/p2kqqQvSfp6+/V+TDbEmJb0j40x057n/XKE1wMAAAAAADG2l0qMD2s3wPiipHd7nvc1/0FjzEcl/Y6k90j6IUm/IOlChLW9IumKpH/med7/2+Pxf2KM+buS/kn79iVjzG96nvelCK8JAAAAAABiKtSeGO19Ln6+fdOT9MHOAEOSPM97RdIHJb3cvuvnjDFvirC2d3qe97MBAYb/mr8m6bfbN/dJ+q8ivB4AAAAAAIixsBt7Pivpmfb1z3me9/u9Bnme96eSPtG+OS3p/cMurDsk6eM3O67PDvt6AAAAAAAg3sKGGD/Scf0zA8Z2Pv7evS1nKN/ouD4zhtcDAAAAAAATEDbEeFvH9S8MGPtCwPNGpfM1Ho7h9QAAAAAAwASEDTG+v+P61oCxX5HUal9/qzGmV4tUJ9p7dSx33HVjVK8FAAAAAAAmK2yI8W0d1/+s30DP816VbYMq2c02nxpiXWH995J+sH19QwNCDGPMh4wxLxhjXnjppZdGuCwAAAAAAOBa2BDjDR3XXwkxfrvj+reEX054xpiSpH/UvvmqpLOe5z3u9xzP81Y8zzvued7xZ555pt9QAAAAAAAQM2FDjFgxxvygbFeSfe27/oHnef/PBJcEAAAAAABGLGyI8c2O668PMb6zS8g3AkcNwRhzSNL/LulN7bt+zfO8f+zyNQAAAAAAQPyEDTH+vOP6mwJH6S8323xj++arkl4eYl1Bc3+PpM9Lekv7rque5/23ruYHAAAAAADxFTbE+FLH9YMDxr5FUq59/UXP87y9LqoXY8x3yQYY/uv/uqTzLuYGAAAAkALVqnTwoL10MQ5A7IQNMe52XD8+YGzn43cDR+2BMebNsgHG97Xv+g1J/7WrgAQAAABAwlWr0sKC9PChvVxfDzeOIANIlLAhxmc7rr9nwNj3dlz/zN6W8yRjzDOSPifpB9p3fVLSTw7qRAIAAAAgI/xgotGwtxsN6dSpJ4OMXuMIMoBECRtiVCW91L7+bmPMD/caZIz5Tkk/0b75iqRPRVmcMebb///27j7asrOuD/j3kRTCjbwEDFgaMgMKCPLOKAtb2tyI8pJUSilCFWVGFIr4QrXpUqzNREHaYFe1sopQZQYQCgVU7EwDVJjYVYtLhyAlQBekmAkRhRB5v0AaePrH3oc5OTnnnjPnfd/z+ax11t37nN/dzzNzn3XuPt+79/Mk+cMkD2qf+t0kz6y1fnWW4wIAAHvEYDDRMxhk7FYnyIDOmCjEqLXekuTF7W5J8ppSyrn9NaWUs5O8Osk57VMvq7XeNOx4pZSjpZTaPg6PqLlLknckeWj71B8keUbbFwAAYNONCiZ6ekHGFVeMr7vkkly4sI4C83LWGdS+PMlTkzw2ySOTvK+U8ook16aZzPPZSR7Y1n4wyYtm7NuVSR7Vbn88zTwYF5dSdvuenVrrO2ZsFwAA6IJDh0YHEz07O8nll09UdyTJfebWOWARJg4xaq03l1KenOTNSS5Kcu8MDyquTvKUWutnZ+zbY/q275VmLoxxTmX86ikAAMBecOTI7ldY9Ix7PUm2tnJwkjpgpSadEyNJUmv9dJLHpZn34niaKyRuTvKJNKuHPCfJo2ut18+5nwAAALe2vZ0cO5Zsbc12nK2t5Pjx/NF8egUs0JncTpIkaZc1fWMmuzJi1DEOJjk4pmbX+0YAAAC+HmRMckXGMG2AkQsvnHvXgPk7oysxAAAA1s60V2QIMKBzhBgAAED3bW8nl102eZCxtZUcPizAgI4RYgAAAN134sRkq5D07Ow0IcZVVy2yV8CcCTEAAIBuO3FiujkxdnaSiy8WZECHCDEAAIDumjbA6BFkQKcIMQAAgG6aNcDoaYOMfzCfXgELJMQAAAC66dChyQKMSSb73NnJ0Zk7BCyaEAMAAOimI0fGBxS9VUgmqDs0t44BiyLEAAAAuml7Ozl2bHRAsbWVHD+eXHrp+Lpjx3LVwjoKzIsQAwAA6K5RQUYvwLjwwvF1x441rwNrT4gBAAB022BAMRhg7FYnwIBOEWIAAADd1wso9u0bHmAMqxNgQOecteoOAAAAzMX2dnLddfOrA9aOKzEAAACAThBiAAAAAJ0gxAAAAAA6QYgBAAAAdIIQAwAAAOgEIQYAAADQCUIMAAAAoBOEGAAAAEAnCDEAAACAThBiAAAAAJ0gxAAAAAA6QYgBAAAAdIIQAwAAAOgEIQYAAADQCUIMAAAAoBOEGAAAAEAnCDEAAACAThBiAAAAAJ0gxIDdnDiR7N/ffJ1HHQAAAFMTYsAoJ04kl1ySnDrVfL3qqsnqBBkAAAALIcSAYXrBxM5Os7+zk1x88W2DjGF1ggwAAICFEGLAoMFgomcwyNitTpABAAAwd0IM6DcqmOjpBRlXXDG+TpABAAAwV0IM6Hfo0OhgomdnJ7n88snqDh2aX98AAAA2nBAD+h05kmxtja8bF2AkzXGOHp25SwAAADSEGNBvezs5dmyyIGM3W1vJ8ePJhRfOpVsAAAAIMeC2Zg0yBBgAAAALIcSAYaYNMvZCgHHiRLJ///hJSSetAwAAmBMhBoyyvZ1cdtnkQcbWVnL4cPcDjEsuSU6dar72lpMdVyfIAAAAlkCIAaOcODHZKiQ9OztNiDHqg/+6G1xetrec7OC/Z1idIAPY61ylBgBrQYgBwwx+UJ/UqA/+627Uv3fw37NbnSAD2KtcpQYAa0OIAYOmDTB6uhZkjPv39v49V1wxvs5JO7DXuEoNANaKEAP6zRpg9HQpyDh0aPy/d2dnsltrdnaa4wHsBa5SA4C1I8SAfpN8oE8mm+xzZyc5eHDmLi3ckSOT/3vG2dpKjh6duUsAK+cqNQBYS0IM6DfJB/reKiST1B05MreuLcy0y8kO2gvLywL0uEoNANaSEAP6jftA3/ugfuml4+uOHWuO1wWzBhkCDGCvcZUaAKwlIQYMGvWBfvCD+m51XQoweqYNMgQYwF7kKjUAWEtCjE1n3fvhBk9eR52EDqvrYoDRs72dXHbZ5CftvVtrnJwDe5Gr1ABg7QgxNpl173fXO3ndt2/3k9D+ui4HGEnzs53k/u6enZ0mxOjCKiwA03CVGgCsFSHGprLu/WS2t5Prrht/Etqr63qAMc3ysl1aThZgGq5SA4C1IcTYRNa9Z9C0AUaPIAPYy1ylBgBrQ4ixaax7z6BZA4weQQawF7lKDQDWihBj01j3nkGTjIlk8qUGDx6cuUsAa8FVagCwdoQYm8a69wyaZEz07u+epO7Ikbl1DWBlXKUGAGtJiLFprHvPoHFjovezvvTS8XVdX50FoMdVagCwloQYm8i69wwaNSYGf9a71QkwgL3EVWoAsJaEGJvKuvcMGhwTo37Ww+oEGMBe4yo1AFhLQoxNZt17BvVO2vft2z2s6q9zcg7sVa5SA4C1I8TYZNa9Z5jt7eS668aHVb06J+fAXuYqNQBYK0KMTWXdewCYjKvUAGBtnLXqDrAC81r33twYAGyK3tVn86oDAKbiSoxNY917AAAAOkqIsWmsew8AAEBHCTE2jXXvAQAA6Cghxqax7j0AAAAdJcTYRNa9BwBYnBMnkv37m6/zqAPg64QYm8q69wAA89ebRP3UqebrqEnQB+sEGQATEWJsMuveAwDMz+AqcKNWcxtWJ8gAmIgQY9P11rMfFWAM1gkwAABua9Qy9oNBxm51ggyAsYQYAAAwi1HBRE8vyLjiivF1ggyAXQkxAABgFocOjQ4menZ2kssvn6zu0KH59Q1gjxFiAADALI4cGb0sfb9xAUbSHOfo0Zm7BLBXCTEAAGAWo5alP1OjVosD4OuEGAAAMKtZgwwBBsBEhBgAADAP0wYZAgyAiQkxAABgXra3k8sumzzI2NpKDh8WYABMSIgBAADzcuLEZKuQ9OzsNCHGVVctslcAe4YQAwAA5uHEieSSSyYPMHp2dpKLLxZkAExAiAEAALOaNsDoEWQATESIAQAAs5g1wOgRZACMJcQAAIBZHDo0WYAxyWSfOzvJwYMzdwlgrxJiAADALI4cGR9Q9FYhmaTuyJG5dQ1grxFiAADALLa3k2PHRgcUW1vJ8ePJpZeOrzt2rDkeAEMJMQAAYFajgoxegHHhhePrBBgAYwkxAABgHgYDisEAY7c6AQbARIQYAAAwL72AYt++4QHGsDoBBsDEzlp1BwAAYE/Z3k6uu25+dQB8nSsxAAAAgE4QYsBuTpxI9u9vvs6jDgAAgKkJMWCUEyeSSy5JTp1qvl511WR1ggwAAICFEGLAML1gYmen2d/ZSS6++LZBxrA6QQYAAMBCCDFg0GAw0TMYZOxWJ8gAAACYOyEG9BsVTPT0gowrrhhfJ8gAWA7zFwHAxhBiQL9Dh0YHEz07O8nll09Wd+jQ/PoGwG2ZvwgANooQA/odOZJsbY2vGxdgJM1xjh6duUsAjGD+IgDYOEIM6Le9nRw7NlmQsZutreT48eTCC+fSLQAGmL8IADaSEAMGzRpkCDAAFsv8RQCwsYQYMMy0QYYAA2DxzF8EABtLiAGjbG8nl102eZCxtZUcPizAAFg08xcBwMYSYsAoJ05M9le8np2dJsQYNTM+APNh/iIA2FhCDBhm3P3Wo4yaGR+A+TJ/EQBsJCEGDJo2wOgRZAAsh/mLAGDjCDGg36wBRo8gA2A5zF8EABtFiAH9JpnxPpl8QrmDB2fuEgC7MH8RAGwUIQb0m2TG+95f8SapO3Jkbl0DYID5iwBg4wgxoN+4+6t791Ffeun4umPHmuMBMH/mLwKAjSTEgEGjgozBieB2qxNgACyO+YsAYGMJMWCYwYBi1Ez2w+oEGACLZf4iANhYQgwYpRdQ7Nu3+1J8/XUCDIDFM38RAGyss1bdAVhr29vJddfNrw6A2fXC41G3lPRfPXfgwO51wmcA6BRXYqyrEyeS/fubr/OoA4C9xPxFALCRhBjrqDdh2alTzddRE44N1gkyANgk5i8CgI0jxFg3gzOuj5o5fVidIAOATWP+IgDYKEKMdTJqybjBIGO3OkEGAJumNy/RqABjsE6AAQCddcYhRmk8vZRyrJRyQynlK6WUvyqlvLOU8qOllLlPFlpKOaeU8jOllD8upXyylPLlUsqpUsobSymPn3d7KzFuzftekHHFFePrBBkAAADsQWcUOJRSzk3y5iQXDbz0ze3joiTPK6U8pdZ6/Tw6WEp5RNvmfQdeuqB9fH8p5fVJDtVab55HmysxyZr3OzvJ5ZdPVnfokNUyAAAA2FMmvhKjlHL7JG/N6QDjY0l+Mck/TXJpkg+1zz8yyZWllDvP2rlSyr4kV+Z0gPGnSX4qyQ8keUmSm9rnfyDJb8/a3kpNsuZ9Mj7ASJrjHD06c5cAAABgnZzJ7STPS/LYdvvqJA+rtb6o1vqGWuuvpgkv3t6+/qA0Acesfi3JPdvtVyV5TK31N2qt/7nW+sK2zd4VH88spVw8hzZXY9QScGdq1MzsAAAA0HEThRjtPBe/0O7WJD9ca/10f02t9ctJfjjJF9unfrKUcvdpO1ZKeViSf9TuXp/k+bXWrw20eX2acKXn8LTtrYVZgwwBBgAAAHvYpFdiXJTkvHb7nbXWDwwrqrV+Mskb2t07JHnyDH17et/2K9uQZJgrk1zbbh8opXzLDG2u3rRBhgADAACAPW7SEON7+7bfNqa2//UnnFl3zrzNWmvN6dtYkqT7q5VsbyeXXTZ5kLG1lRw+LMBgI5199tk5++yzV92NtfXwhz88D3/4w1fdDYDb8P7EOnJeAeuvNBnAmKJS3pbT4cB2rfWqXWr3J/mLdvdDtdYHnXGnSvmGJJ9PspXkliR3rLXeskv9wSRH2t2X11p/fFwbBw4cqCdPnjzTri3HuOVWh3ElBgAAAHtAKeU9tdYDw16b9EqM+/dtXzem9oYkX22371dKKRO20e/8NAFGkvzlbgFG61Tf9v1HVnXBNAFG0tRffHFy1VUL6RYAAACs2qQhxl37tj+1W2EbOHyu3T0ryTlT9Gvi9lo39W3fdWTVups2wOgRZAAAALCHTRpifGPf9qgJNvt9qW/7TpN3Z7HtlVKeU0o5WUo5eeONN07RrQWaNcDoEWQAAACwR00aYqzS+Ek7Jj1Qra+stR6otR4477zzxn/DMh06NFmAMclknzs7ycGDM3cJAAAA1smkIcYX+rYnma73jn3bn5+8O0Pbu+PIqvm1t3pHjowPKHqrkExSd+TI7jUAAADQMZOGGJ/p2777boWllLOS3LndvSXJF6fo18TtDan5zMiqdba9nRw7Njqg6K0+cuml4+uOHWuOBwAAAHvIpCHGh/u294+pPT/J7drtj9RJ1nC9rRuS9O6tOL8NRnazr2/7wyOr1t2oIGNw+dTd6gQYAAAA7FGThhjX9G0PXat1xOvXjKzaRa31a0k+1O6eleRhi25zbQwGFIMBxm51AgwAAAD2sElDjLf3bT9+TO0T+rbfdmbdOfM2Syll4PW3j6rtjF5AsW/f8ABjWJ0AAwAAgD2uTHK3R3s7x8eTnJdmtZCH1Fo/MKTuHkk+muScNEujnl9rvWmqjpXysCR/3u5en+QBtdbbLLdaSnlSkuPt7sla63dMcvwDBw7UkydPTtM1AAAAYEFKKe+ptQ69C2SiKzFqrbckeXHveEleU0o5d6CRs5O8Ok2AkSQvGxVglFKOllJq+zg8os33Jfn9dveCJC8rpdyqv6WUC5K8vO+poccCAAAAum/chJn9Xp7kqUkem+SRSd5XSnlFkmvTTOb57CQPbGs/mORFc+jfC5I8Jsk92+M/uJTy2iQ3JXlIkufm9Mokr6u1Hh96FAAAAKDzJg4xaq03l1KenOTNSS5Kcu8MDyquTvKUWutnZ+1crfVUKeWJbZv3TfLo9jHo9Ul+ZNb2AAAAgPU16cSeSZJa66eTPC7JM9LMQ/HxJDcn+USSdyV5TpJH11qvn1cHa63vTfLQJD+b5N1JPpXkK0k+luRNSZ5Ya/3BWuvN82oTAAAAWD8TTey5F5nYEwAAANbPzBN7AgAAAKyaEAMAAADoBCEGAAAA0AlCDAAAAKAThBgAAABAJwgxAAAAgE4QYgAAAACdIMQAAAAAOkGIAQAAAHSCEAMAAADoBCEGAAAA0AlCDAAAAKAThBgAAABAJwgxAAAAgE4QYgAAAACdIMQAAAAAOkGIAQAAAHSCEAMAAADoBCEGAAAA0AlCDAAAAKAThBgAAABAJwgxAAAAgE4QYgAAAACdIMQAAAAAOkGIAQAAAHSCEAMAAADoBCEGAAAA0AlCDAAAAKATSq111X1YiVLKjUlOLaGpb0ryqSW0w2YxrlgE44pFMK5YBOOKeTOmWATjanr7aq3nDXthY0OMZSmlnKy1Hlh1P9hbjCsWwbhiEYwrFsG4Yt6MKRbBuFoMt5MAAAAAnSDEAAAAADpBiLF4r1x1B9iTjCsWwbhiEYwrFsG4Yt6MKRbBuFoAc2IAAAAAneBKDAAAAKAThBgAAABAJwgxAAAAgE4QYgxRGk8vpRwrpdxQSvlKKeWvSinvLKX8aCnlrAW0eU4p5WdKKX9cSvlkKeXLpZRTpZQ3llIeP+/2WL5ljqtSygWllOeWUl5XSrmmlPK5UsrNpZQbSyn/s5TyS6WUC+bVHquziverIX24Wynlr0spte+xf9HtsjirGlellNuVUp5WSnlDKeXaUsoX2vevj5RSriyl/MtSyv0W0TaLt6Lzq4tLKa9vx9AX2t+Fnyql/K9Syq+UUr5l3m2yHO37xYNLKQdLKb9RSnl3KWWn7/fQ4QW165x9D1v2uHLOPh0Tew4opZyb5M1JLtql7OokT6m1Xj+nNh/RtnnfXcpen+RQrfXmebTJci1zXJVSfj/J9yUpY0q/kuRf1Vp/dZb2WJ1VvF+N6MfRJM8aePo+tdbrFtUmi7OqcdX+LvztJI8YU/rrtdYXzKtdlmPZ46qU8k1J3pTkwjGlNyf5xVrrFbO2yXKVUt6S5B/vUnJ5rfXwnNt0zr7HLXNcOWef3sL/QtclpZTbJ3lrkse2T30szbI41yY5P8mPJHlgkkcmubKU8pha6+dmbHNfkiuT3LN96k+T/E6STyV5SJLnJLl7kh9I8rUkPzRLeyzfCsbVg3P6zfDPkpxI8uEkn2vbe2qSv5vkDkleWkq5Q631xTO0xwqs4v1qRD++J02A8bU0HwbOnncbLM+qxlUp5bvS/C68c/vUnyX5r0n+ot2/d9vuk2Zti+Vb9rhqr+i4MsmB9qkvJ3ltkj9P8uk04+kfJvl7SW6f5N+WUr5Qa/2P07bJStxuYP9vktyUZCFXazln3xjLHFfO2adVa/VoH0l+OkltH+9Jcu7A62cneVtfzUvn0Obv9R3vt5N8w8DrFyQ51Vdz8ar/nzzWe1wluSbJy5I8YJeaF/S19/+S3H/V/08e6z2uRvThnCQfbY//H5Jc19fe/lX/H3lM9TNdxe/Be6b5EFCTfCnJM3apvV2Sv73q/yePM/4ZL/v34A/1Hev6JPtG1P1YX92NSc5a9f+Vxxn9nF+Y5CVJ/kmaq/+S5GDfz/TwnNtzzr4Bj2WOK+fs0z/cTtJqU/uPJzkvzSB5SK31A0Pq7pHmpP2cNJf2/J1a601TtvmwNH8VSJpfsg+otX55SN2Tkhxvd0/WWr9jmvZYvhWNq3NrrZ+eoO7NaRLeJPmlWutl07TH8q1iXI3ox79P88v1L9P8FfX9Sfa1L7udpGNWNa5KKW9I8vR29xm11jdOeyzWz4p+D/5Okh9sd59fd7nCopRyMsmj2t2H1lrfP02brIdSysEkR9rdeV7275x9gy1wXDlnn5KJPU+7KM0v2CR557BfsElSa/1kkje0u3dI8uQZ2nx63/Yrh70Ztq5Mc8llkhwwCVWnLH1cTfJm2HpT3/ZDpm2PlVjF+9WtlFK+M8lPtbs/WWv9/LyOzcosfVy1l2c/rd39HwKMPWkV71f36Nv+yJjaD/dtnzNDm+xtztmZO+fs0xNinPa9fdtvG1Pb//oTFt1mbS6XeXvfU2Y+7o5VjKtJ9X/ovOMS2mN+VjquSil/K+2ltEneWmv9vXkcl5Vbxbh6Vk6fi/zWDMdhfa1iXH2ib3vcfey917+aWwca0M85O6vknH2AEOO0B/dtv2dM7ckR3zexUso3pLn8OkluSfK+RbfJSix1XJ2h/jZOAZdgbwAACMBJREFULaE95mfV4+rn22N9IclPzOmYrN4qxtXf79t+Vynl7qWUy0sp7y+nl1f9QLvM3f1naIfVWcW4emvf9s+1V/zcRinlR3N68s/X1lr/ZoY22aOcs7MGnLMPsDrJaf0nR9eNqb0hTWJ/uyT3K6WUeuaTi5yfZKvd/sta6y1j6vsHrBO57lj2uJpIe4/yob6njo+qZS2tbFyVUh6YZtKrJPmFWusN0x6LtbOKcdX7APnZNPOpvCXJNw/UPKh9PK+U8vO11pdO0Q6rs4px9ZY0kzA+Jc1KJP+nlPKa3Hp1ku9LszpJ2tqfnKIdNoNzdlbGOftwQozT7tq3/andCmutt5RSPpfk3DT/h+ek+YvkQtpr9U9uddeRVaybZY+rSf2LJN/Wbr8/3hC7ZiXjqv1r1G+luV/9ZJoZtdk7ljquSil3SHKXdverSY61x/toklelua/83DRzIzwhzQfbK0opN9daf/1M2mKllv5+VWutpZSnJfmlNOHEndIsfzno6iT/Osl/W9QfDdgTnLOzSs7Zh3A7yWnf2Lc9arKefl/q275TB9pjNdbu51xK2U7yy+3uLUmeW2v92iLaYmFWNa6en+S70nzgfI5xs+cse1yd27d9t3b/bUm+vdb64lrrG2utv1lrfWKSf9ZXe0Up5fwp2mM1VvJ+VWv9apKXJrk8zWonwzwyyc8lecy07bAR1u5cjs3gnH00IcZ6kP6zFKWUb0szw3HvKqwX1lrfvcIu0RGllAuS/Eq7+2u11veusj/sCYPnIJ9P8sxhs/7XWl+R5haBJLl9kuctuG90XCnlCWluX/nVJO9OM8HiXdNcSXa/NLfFfTHNLSXvLKXMbfUm9jTn7CyFc/bdCTFO679c8ewJ6vtnhp1macH+9iaZZXbW9liNZY+rkUop90ny35PcvX3q19xb3lmrGFcvT/PXqFNJNn598j1q2eNq8Ht+t9Z609DKxiv7th83RXusxtLfr9oA43ia25XenOS7a63vqLV+ttZ6c6312lrrS9Is//qVtl+vLaUMzscCiXN2lsw5+3hCjNM+07d995FV+foEK3dud29Jk+QvrL0hNZ8ZWcW6Wfa4GnXseyd5V5rJqZLkN2ut/3xex2fpljquSinPTPKkdvf5tda5jU3WyrLfr77Qfm/P1WPq+1e2+JYp2mM1VvF78N+lOcf9WpKfHnX5da31T5McbXfvlOTglO2xtzlnZ2mcs0/GxJ6nfTjJfdrt/dl9+Zrz00wwliQfmXIyqBuS7KSZ7fj8UspZY2Y77l8ezDrm3bHscXUbpZR7pXkz3N8+9aokPz6PY7Myyx5Xz26//nWSR5RSHjGi7i592z9RSumdvP2nWusnpmiX5VrquGonX/xITi9d+Nkx39L/+l1GVrFuljqu2r9gPqjd/WCt9eNjvuUPkzy33f7OM22PjeCcnaVwzj45IcZp16S5XzJplnz7o11qD/RtXzNNY7XWr5VSPpTkUWl+Dg/L7uunz9wmK7HUcTWovTT2XUm+tX3qdUl+zCzsnbfscVXar9+c0xNMjfOzfdvHkggx1t8q3q/+d06HGHferTC3Di7GBR6sj2WPq3v1bX9ugvr+sXTOlG2yhzlnZxmcs58Zt5Oc9va+7cePrGo8oW/7bYtus5RSBl5/+6ha1s4qxlWSpJRyXpJ3JnlA+9QbkzzLrMZ7wsrGFXvaKsbVlX3bjxpT2/+6v252x7LHVX9wce8J6vv/ar7bnCxsNufsLIxz9jMnxDjtRJIb2+3HlVK+fVhRKeUeSZ7R7n45yVtnaPO/9G0/t5QyasKrJ+Z0Kney1vp/Z2iT5VrFuEop5W5pLpHtXVL7u2lm/f/qLMdlbSx1XNVaL6y1lnGP3Poy8fv0vfbn07TL0q3i/eqtaS7TTpKnlFJ2u9/8OX3bArnuWPa4ujanl8G8dynlu8bUP6Nv++SUbbL3OWdnIZyzT0eI0WrvbXtxu1uSvKaU0r+Gfdo3rFfn9OWGLxs1k3op5WgppbaPwyPafF+S3293L0jyslLKrX4m7bKGL+97auixWE+rGFellLskeUeSh7ZP/UGSZ4y5f5MOWcW4Yu9b0e/Bz6SZhDFpbid57bAPB6WU5yR5arv7xdz69yJrbNnjqtb6pdw6AHl1ey417FgvTPLd7e5XcusPqmwI5+wsgnP2xTInxq29PM1J0mOTPDLJ+0opr0iT6p+fZnK73r27H0zyojm0+YIkj0lyz/b4Dy6lvDbNJY0PSTPZVO8vU6+rtR6fQ5ss17LH1ZU5fdn1x9PcU3dxc4XjSDu11nfM2C7LtYr3K/a+VYyrf5Pmr5cH2q/XlFJe1bZ5bpInt8/3PLfWeuNtjsI6W/a4emGS70lytzR/Fb+mlPI7Sf4kyZfSfAh9WpJH933P5bXWG2ZslyVqJ3F99sDTD+3bvqhd8abfW2qt752ySefsG2DJ48o5+5SEGH1qrTeXUp6cZk3xi9LcSznsF+nVSZ5Sa515YrFa66lSyhPbNu+b5hfqo4eUvj7Jj8zaHsu3gnH1mL7te6W5r26cUzk9EzIdsIr3K/a+Ff0e3CmlPCnNX8EvTLN86ouHlH4pyfNqra+btU2Wa9njqtb60VLK9yR5Q5L7pVk+9XntY9AtSQ7XWl8yS5usxL4kv7DL649tH/2uTTJViOGcfWMsc1w5Z5+S20kG1Fo/neRxae6RPJ4mFbs5zcz670pzT+6ja63Xz7HN96ZJ+H42ybuTfCrNZY0fS/KmJE+stf5grfXmebXJcq1iXLH3GVcswop+D96Y5sPt96e5FeBjbZufTXNi+JIk31prffW82mS5lj2uaq1Xpzm3elaS30tyXZr5V25J85fzP0lzFdADaq3DQjO4DefssB6KVVsAAACALnAlBgAAANAJQgwAAACgE4QYAAAAQCcIMQAAAIBOEGIAAAAAnSDEAAAAADpBiAEAAAB0ghADAAAA6AQhBgAAANAJQgwAAACgE4QYAAAAQCf8f/gDplKAkyR/AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1296x1296 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "N=100\n",
    "# T: sub-models數量\n",
    "T=20\n",
    "# D:存放權重的矩陣\n",
    "# a:存放每次回圈所得到的alpha值\n",
    "D=np.ones(N)/N\n",
    "new_D=np.ones(N)/N\n",
    "a=np.zeros(T)\n",
    "# result用來存放資料最後AdaBoost分類的結果\n",
    "result=np.zeros(N)\n",
    "result_test=np.zeros(N_test)\n",
    "plt.rcParams[\"figure.figsize\"] = (18,18)\n",
    "# 執行迴圈\n",
    "for i in range(0,T):\n",
    "    # 執行弱分類演算法\n",
    "    x_split,axis,error=Decision_Stump(x,y,D)\n",
    "    # 劃出分割線\n",
    "    if(axis==0):\n",
    "        plt.plot([x_split,x_split],[min(x[:,1]),max(x[:,1])],\"k-\")\n",
    "    else:\n",
    "        plt.plot([min(x[:,0]),max(x[:,0])],[x_split,x_split],\"k-\")\n",
    "    # 計算alpha值\n",
    "    a[i]=0.5*np.log((1-error)/error)\n",
    "    # 確認分類線劃分的兩邊為1或-1\n",
    "    temp=y[x[:,axis]<x_split]==1\n",
    "    if(sum(temp/len(temp))>=0.5):\n",
    "        cl,cr=1,-1\n",
    "    else:\n",
    "        cl,cr=-1,1\n",
    "    \n",
    "    left_index=np.array(np.where(x[:,axis]<x_split))\n",
    "    right_index=np.array(np.where(x[:,axis]>x_split))\n",
    "    left_index_test=np.array(np.where(x_test[:,axis]<x_split))\n",
    "    right_index_test=np.array(np.where(x_test[:,axis]>x_split))\n",
    "    # 四種情況\n",
    "    # 1.分到其中一邊，且正確 | 2.分到其中一邊，但錯誤 | 3.分到另外一邊，且正確 | 4.分到另外一邊，但錯誤\n",
    "    new_D[left_index[y[left_index]==cl]]=D[left_index[y[left_index]==cl]]*np.exp(-a[i])/sum(D)\n",
    "    new_D[left_index[y[left_index]!=cl]]=D[left_index[y[left_index]!=cl]]*np.exp(a[i])/sum(D)\n",
    "    new_D[right_index[y[right_index]==cr]]=D[right_index[y[right_index]==cr]]*np.exp(-a[i])/sum(D)\n",
    "    new_D[right_index[y[right_index]!=cr]]=D[right_index[y[right_index]!=cr]]*np.exp(a[i])/sum(D)\n",
    "    # 每次計算後加乘結果\n",
    "    result[left_index]+=cl*a[i]\n",
    "    result[right_index]+=cr*a[i]\n",
    "    result_test[left_index_test]+=cl*a[i]\n",
    "    result_test[right_index_test]+=cr*a[i]\n",
    "    # 更新權重\n",
    "    D=new_D\n",
    "# 輸出結果\n",
    "print(\"The training error rate: \",round(1-sum(np.sign(result)==y)/N,2))\n",
    "print(\"The testing error rate: \",round(1-sum(np.sign(result_test)==y_test)/N_test,2))\n",
    "plt.plot(x[:50,0],x[:50,1],\"bo\",markersize=20)\n",
    "plt.plot(x[50:,0],x[50:,1],\"rX\",markersize=20)\n",
    "plt.xticks(fontsize=30)\n",
    "plt.yticks(fontsize=30)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}