notebooks/usage-LinearTree.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pprint\n",
    "import numpy as np\n",
    "\n",
    "from sklearn.linear_model import *\n",
    "from lineartree import LinearTreeClassifier, LinearTreeRegressor\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "from sklearn.datasets import make_classification, make_regression\n",
    "\n",
    "import warnings\n",
    "warnings.simplefilter('ignore')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((5000, 10), (5000,))"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X, y = make_regression(n_samples=5000, n_features=10,\n",
    "                       n_informative=3, n_targets=1,\n",
    "                       random_state=0, shuffle=False)\n",
    "\n",
    "t = X[:,0] > np.quantile(X[:,0], 0.5)\n",
    "y[t] += X[t][:,3]*X[t][:,4]\n",
    "y[~t] += X[~t][:,1]*X[~t][:,2]\n",
    "\n",
    "t = X[:,5] > np.quantile(X[:,5], 0.7)\n",
    "y[t] += X[t][:,6]*X[t][:,7]\n",
    "y[~t] += X[~t][:,8]*X[~t][:,9]\n",
    "\n",
    "X.shape, y.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### default configuration "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearTreeRegressor(base_estimator=LinearRegression())"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr = LinearTreeRegressor(LinearRegression(), criterion='mse')\n",
    "regr.fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.        , 0.15034349, 0.17039916, 0.16322357, 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.1249401 , 0.39109368])"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.feature_importances_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABHIAAAIYCAYAAAAB9MvqAAAABmJLR0QA/wD/AP+gvaeTAAAgAElEQVR4nOzdf3RU9Z3/8de1cuqP1YxnawLiCX7bLhSLBNFyoFayJuxRaGfs7iGU2JPjnpbQyVqO9DCeUk9mESeiXZMtGthAJl0bs5LReGrNCLauiYR2SURaZ6zUSi01c4ruTHXNtdYfq4f7/YOda4ZMksnPm5l5Ps7JIfncz/3M+04+c+C++Xze17AsyxIAAAAAAABmvLOcDgAAAAAAAACZIZEDAAAAAACQJUjkAAAAAAAAZImznQ4AAIB89cEHH+jxxx93Ogxkueuuu04XX3yx02EAAIBpYlDsGAAAZ8Tjcc2ePdvpMJDlDh48qNLSUqfDAAAA04QVOQAAOIwbcYwHiUAAAPITNXIAAAAAAACyBIkcAAAAAACALEEiBwAAAAAAIEuQyAEAAAAAAMgSJHIAAAAAAACyBIkcAAAAAACALEEiBwAAyO/3y+/3j9gnkUgoFArJ4/FMU1SZm8mxAQAATKaznQ4AAABkh23btmnPnj1Oh5HWVMYWDocVDAYlSdXV1XK73VPyOgAAAJkwLMuynA4CAIB8FI/HNXv2bB08eFClpaVOh5MRwzAkSTPxnw9TEVsoFNK+ffvU1tYmSdq6dauWLl2q6urqSXuN8crG+QMAACaOFTkAAABpxGIxVVZWqre3VwUFBZIkr9erJUuWaNmyZSopKXE4QgAAkI+okQMAQJ4brr6MaZoKhUIyDEMej0fHjx+flPHD4bA9ZiwWG/Y1DcNQMBhUIpEYMmamsSUSCTU0NNj9uru77WOj1QU6fPiwJOmSSy6x2+bMmSNJOnLkSIZXDwAAMLlYkQMAQJ7bsGGDwuHwkPaqqirNnTtXAwMDKigoUCgUmvD4fX19crvd6u/v17x58zR37lw1NTWlvKbb7ZZlWUokEva5bW1t9qqYTGNLnn/TTTfJsix1d3ervLxckUgko9U0PT09kqTi4mK7rbCwUNLpZNRM2F4FAADyDzVyAABwyEyqcXJmfZlwOCyPx6OXX35Z8+fPl3R6FYzL5UrpN97x07UlEy3xeNxOmPT19WnFihVqb2/X+vXrxxRbKBRSZWXlkNesra1VIBAYV8wjtU+3mTR/AADA9GFrFQAAGOLAgQOSZCdKJKWsiJkKHR0dkj5e9SJJCxculCTt27dvzLElz0lu00omYOrq6iY5cgAAgOlDIgcAAAzhxGPG071mMkEzeOtXprElz7Esa8hXJkZ6zLjX681oDAAAgMlGIgcAAMwIycRJuuLGE0mcjLdIc7p4ksWZly5dOu54AAAAJoJEDgAAGKK5uVmSFI1Gp+01b7rpJknSiRMn7DbTNCVJFRUVY44t2a+trc0eJ/kUq0xcf/31Q+J57bXXUo4BAABMNxI5AADkucErTpLfJxMVfr/fXoUy+NHdNTU14xo/mVBJ/jn4+OrVq+V2u7Vjxw677cknn5TX61VZWZndP9PYbrzxRkmna+K4XC4ZhqGioiI7KTTa48eLi4vV3Nys1tZWmaYp0zTV2tqq5ubmlCdZAQAATCcSOQAA5LmioqIh3xcXF6u/v19z587VvHnzVFNTo0WLFsntdqu9vV3bt28f1/jJJ0sl/xx8vKCgQC0tLXK73SoqKrKLE99zzz0p42UaW2Fhofr7+1VbWyvp9Pas/v7+MSVhqqurtWbNGrlcLlVVVamiooLHjgMAAEfx+HEAABzC46MxEcwfAADyEytyAAAAAAAAsgSJHAAAAAAAgCxxttMBAACA7JSsYTMadnEDAABMHhI5AABgXEjQAAAATD+2VgEAAAAAAGQJEjkAAAAAAABZgq1VAADkuDNr2YxlS9REzgUAAMDkY0UOAABZIBaLyTCMIV8NDQ12n+7u7iHHB7MsKyURk0gkFAwG7b6hUGjI6555TjYIh8PyeDzyeDwKh8MZnWOapvr6+hQMBuXxeEYd2zAMeTyetO/ZYMn3d/DrpPs9Dvf+AwAAnIlEDgAAWaC4uFgDAwPq6uqS2+2WJLW3t2vLli12n7KyMkUiEbndbvX29o6YgDFNUxs2bJB0OlkTj8e1b98++f3+KYk/Go1O2diDhUIhBYNBtbW1qa2tTQcOHFAwGBz1vPr6eu3fv18bN24cNvnT0NAgj8ejQCAgy7IUCARUWVmZkkwbLBqNauPGjSltL7300rAxlJWVjRonAACAYWXbf7MBAJAj4vG4Zs+erYMHD6q0tDTj8xKJhJ2EaWtrU0FBgaTTq3a+/e1vq6WlRYWFhXb/5IqQwX/lh0IhVVZWamBgwD4/Go1qyZIl6urqGpJUSDdGJnF2d3erp6dHpaWlKisrS4lrssViMc2bN0+9vb1avny5pI+vKRKJqKSkZNQxRrrOdMcMw5Db7VZnZ2dKX9M0VV9fr7q6upRzQqGQvvjFL6q4uNjum0gk1NjYqEAgMJbLHff8AQAA2Y0VOQAAZJnCwkIFAgGFw2G1tLTY7XffffeQJM5w9u3bJ0l2EkeSLrvsMklSR0fHhOJLrr5pbGzUwoUL1dTUpPXr109pEkeSDh8+LEm65JJL7LY5c+ZIko4cOTLh8evr6yVJfX19kk4njiSlTcC0tLRo06ZNQ9rLyspSkjjS6S1xa9eunXB8AAAgP1DsGACALFRSUqLOzk55PB5deeWV+v3vf6/t27dnnCxJt30omdTZs2ePmpqaxhTPmatvNm3aNGwsZ9buGc5YFw339PRIUkqiJBlDOBxWdXX1mMY705YtWzQwMKAVK1aot7dXr776quLx+JDr7O7u1jXXXJP2+tO19fT0aP369ROKDQAA5A9W5AAAkKXcbrfq6+tVXl6uz3zmM2Na8eL1eiVJx48fn5RYioqKdPLkSd1zzz2jrr5JFlAe7Wus9uzZM+yxTIsejyYQCMjr9WrFihU6duyYPvnJT6YcTyQS+v3vf29v7RpNNBplWxQAABgTEjkAAGSxiooKud1u7dy5U6ZpZnzezTffLEn6wQ9+YJ8XjUYlfbyFaCzi8bjmzp2rrVu3KhQKKZFIjHmMsRjp6VxTqaGhQaWlpRoYGJAkVVVVpbzvjz/++JhW/jz66KMUOQYAAGNCIgcAgCyVSCT0yiuvaNeuXUPq5Yxm+fLl6urq0smTJ+VyuRQMBvXmm29KklatWjXmWAoLC7V+/Xo1NTVp4cKFamxslN/vt5NDgw33+O2JJmeST/NKJ7kCaSJCoZB8Pp9Wr16tgoICVVVVKRwO65FHHpF0etXP9ddfn/F4yWTXVNcOAgAAuYVEDgAAWerxxx+3i+c2NzfL5/PZhXgzUVZWps7OTlmWperqaj3//POqra3N6OlOIykpKVEgENCmTZv00ksvqaamJmWVzmRsrUrXN5nIGbwaKFmQeOnSpRO6JkmqrKyU9HEtoaKiIkmyHzHu8Xg0b968tMmodIkpihwDAIDxIJEDAECWMU1TDQ0NKVt4qqur5Xa7tWPHjnHVvQmFQurp6ZHP55u0ONOt0plKydUwJ06csNtee+21lGMTceaKn2RCJ9k+UjIqXWKqp6dnwkkzAACQf0jkAACQRaLRqKqqqtJuf2pra1M4HNaCBQvU3d096limaSoajaqmpkYnT55UZ2dnyuPIJ1Nylc5USq5Mam1tlWmaMk1Tra2tam5uTnmSld/vl9/vH3L+4Fo36eoNbd68WdLppJf08WPIk+1jQZFjAAAwXiRyAADIEoZhaMmSJQqHw1qyZElKsiYWi8nlctk/l5eXj1hnxjAMuVwuHTlyRF6vV1u2bJnS2KdLdXW11qxZI5fLpaqqKlVUVGRUfDj5fiS5XK4h719ZWZm6urrU09MjwzDU2tqqrq6ucRUrpsgxAAAYL8Maz/M9AQDAhMXjcc2ePVsHDx6c0tUZyYTERP7Kn4wxMLmma/4AAICZhRU5AAAAAAAAWYJEDgAAAAAAQJY42+kAAADA9BjP9qiR6uwAAABg+pHIAQAgx02krg01cQAAAGYWtlYBAAAAAABkCRI5AAAAAAAAWYJEDgAAeSSRSCgUCsnj8TgdCgAAAMaBRA4AAHlk27ZtqqysVDgcdjqUtEzTVF9fn4LB4LiTTdFo1D7/zGLN4XDYbvd4PAqFQinHY7GYampqZBiGampq1N3dPexrGIZhf9XU1IwrVgAAgLEikQMAQB5pampyOoQR1dfXa//+/dq4ceO4kk0NDQ3y+/2aPXu2du3alVKsuaGhQR6PR4FAQJZlKRAIqLKyUg0NDZJOJ5Gi0aiampo0MDCg0tJSlZeXp43jyJEjKT+vWbNmzLECAACMh2HxOAoAABwRj8c1e/ZsHTx4UKWlpdP2uuN5DPl0G0+MNTU1+tSnPiWfz6eCgoKMxjQMQ263W52dnQqHw3K73RnFka7vdHNq/gAAAGexIgcAAMg0TYVCIXurUDAYVCKRGNKvoaEh5fiZW5dGOu73++X3+6ck/uS4gUAgbRJHOr3aR5L6+voknd5GlTxH0rCJGa/Xm/JzLBaTx+OR3++3xwIAAJguJHIAAICqqqr05z//WZZlKR6PKxwOa8OGDTJN0+7T0NCgiooKWZaldevWqbGxMWWM0Y5PlWg0qrq6Oq1Zs0bBYNCuf3NmfZstW7aotrZWK1asUF9fnw4fPqx4PK6SkpK04yav/cxtU9FoVJJUV1enFStWyOPxpE16AQAATAUSOQAA5Lnu7m6Fw2HdeOONkqTCwkLdfvvtCofDevLJJ+1+Pp9P55xzjiSpoKBAmzZtShlntOOBQMBe/TKZnn76aUlScXGxqqurNTAwoLlz56q8vHzIiplAICCv16sVK1bo2LFj+uQnPznsuL/85S/ldru1cuXKlHa3262BgQFFIhHV1tYqHA7r8ccfn/TrAgAASIdEDgAAea6jo0PS6QRO0sKFCyVJ+/bts9u8Xq+KiooUCoVkmqYKCwtTaseMdnyq+Hw+SbJX1hQUFNjboVpbW1P6NjQ0qLS0VAMDA5JOr0QavOposJ07d+r2229Pu1WroKBAJSUlCgQCam5unrFPAQMAALmHYscAADhkphQ7Hq6g75ntx48fl8/ns5MW9fX12rJli91/tOMTiXE8fc9sD4VCqqys1MDAgAoKCnT8+HEtWLBAzc3Nqq6uTjk3FArpz3/+85D2dEzTlMvlmvbi0RQ7BgAgP7EiBwCAPJcs8puuzsvgQr/z589XZ2enIpGIvF6vfD6f/ejuTI5PlWSM6VbWDC5gXFlZKUn2CpuioiJJ0saNG1POiUajOnbsWEZJnOR4ZxZEBgAAmCokcgAAyHM33XSTJOnEiRN2WzIpUlFRYbcZhiHTNFVSUqKmpiZFIhF7W1Mmx6dKMsZXX311SPzJa5OGPpUqmdAZ3J5IJPT000+n1PKJRqOqqakZ9vVN00x5nwAAAKYSiRwAAPLI4FU3ye9Xr14tt9utHTt22G1PPvmkvF6vysrKUs6vr6+3H9t90UUX2Y/0zuR4po8fH7yyJt0qmzPHKSsrU21trfx+vx3/I488IrfbrfXr19v9Nm/eLOn0tinp48eQJ9sTiYQ2bNggn89nP4bdMAwtWbLEfnJVKBRKeRpWLBbToUOHhrxPAAAAU4VEDgAAeSS5nWjw9wUFBWppaZHb7VZRUZFdW+aee+4Zcv6mTZvU0dEhwzDU0dExpAbOaMdHYxiGXC6X/bPL5bLjGUkgEBgSf1tbW0qfsrIydXV1qaenR4ZhqLW1VV1dXXYSZtu2bcMWLV6wYIEk6fzzz1d5ebkMw5Df79dbb701ZKUPAADAVKLYMQAADqFYLSaC+QMAQH5iRQ4AAAAAAECWIJEDAAAAAACQJUjkAAAAAAAAZAkSOQAAAAAAAFmCRA4AAAAAAECWIJEDAAAAAACQJUjkAACAaZNIJBQKheTxeJwOBQAAICuRyAEAANNm27ZtqqysVDgcnvbXjkajMgzD/qqpqRnSJxwOy+PxyOPxDBvjZPUBAAAYDxI5AABg2jQ1NTn22keOHEn5ec2aNSk/h0IhBYNBtbW1qa2tTQcOHFAwGJySPgAAAON1ttMBAAAATIfZs2fLsqy0x2KxmCorK9Xb26uCggJJktfr1ZIlS7Rs2TKVlJRMWh8AAICJYEUOAAA5qKGhQYZhKBgMKpFIyDAM+5hpmgoGg/YWI7/fr0QiIWloDZtwOGxvQ4rFYpJOrzg5sy2RSNjbiSTZ49fU1Oj48eOjxptIJOyYPR6Puru7M74ev98vv98/4vixWEwej0d+v199fX1Djh8+fFiSdMkll9htc+bMkfTxSp7J6gMAADARJHIAAMgxDQ0NqqiokGVZWrdunRobG1OOb926VRs3blQ8Hld/f7/q6uq0bds2SdKGDRvsGjbRaFRut1u9vb3as2eP7r77bvX19Wn9+vXq7++32ySpqKjIrgfT19en6upqDQwMSJIWLFgwYjInkUhow4YNmjt3rizL0ubNm1VeXq5oNJrR9WQiOVZdXZ1WrFghj8djJ68kqaenR5JUXFxstxUWFkqSXeNmsvoAAABMhGENt8YYAABMqXg8rtmzZ+vgwYMqLS2dtHENw1A8HrcTCIlEQkVFRfa2Ir/frzfeeMOuV5Nc3ZI8fubPmbal6xONRrVkyRLV19dry5YtafuFQiFVVlYOGbu2tlaBQGDU68mUaZp69dVX9eijj6qurk7Nzc2qrq4eNvYz2yerz2SZqvkDAABmNlbkAACQY7xer4qKihQKhWSapgoLC1MSCIFAQE1NTYrFYmpoaJjSWJI1YXw+37B99u3bJ0kpT5SSTq+ekUa/nkwVFBSopKREgUBAzc3NrJABAABZiUQOAAA55jvf+Y7cbrcqKyvlcrnSJmuCwaC+/e1vy+12OxBhqmRCxbKsIV9SZtczVuvWrUtJ5Iz0Pni93kntAwAAMBEkcgAAyDHz589XZ2enIpGIvF6vfD5fSvIjFApp48aN2rVrl+bPnz8tMWWSxBiujs5o1zMeBQUFKTElEzCD6+YkCzkvXbp0UvsAAABMBIkcAAByjGEYMk1TJSUlampqUiQSSdnaVFlZKSm1IO9USSZn1qxZM2yf5uZmSVJbW5tM05T08VOspNGvZzxM01RFRYX98/XXXy9JOnHihN322muvpRybrD4AAAATQSIHAIAcVF9fb68Eueiii1RfX28fS64aicViKatgEolEykqSwUmVwX2Ga0sKhUL2+W1tbXK73WlXqiS/v/HGGyWdronjcrlkGIaKiopSEi0jXc9ojx8PhUIpjzOPxWI6dOiQysrK7Lbi4mI1NzertbVVpmnKNE21traqubnZTnhNVh8AAICJIJEDAEAO2rRpkzo6OmQYhjo6OuwnRkmnix1Lp+vkuFwu1dbWyuv16v3331dRUZHdz+VySVJKW/L7dG1JCxculMfjkcvlUnFxsdra2tL2TX5fWFio/v5+1dbWSjq9Dau/vz8l8THS9Yzm/PPPV3l5uQzDkN/v11tvvZW2lk11dbXWrFkjl8ulqqoqVVRU2E+1muw+AAAA48XjxwEAcEiuPT56Kh6xjeHl2vwBAACZYUUOAAAAAABAliCRAwAAJmykmjkAAACYPCRyAADAhI1UMwcAAACT52ynAwAAANmPujgAAADTgxU5AAAAAAAAWYJEDgAAAAAAQJYgkQMAAAAAAJAlSOQAAAAAAABkCYodAwDgsDvuuEOf+tSnnA5jVB999JHOPju//unw4YcfatasWU6HkdYHH3zgdAgAAMABrMgBAMAh55xzjtauXZsVSZwPPvhAP/3pT/X66687Hcq0OXbsmH7+85/P2CdyffKTn9TatWt18cUXOx0KAACYRoY1U/91AgAAZoyvfe1r6uvr04svvqgLLrjA6XCmxW9/+1tdddVV8vl82r59u9PhAAAASGJFDgAAGMX+/fv1yCOPaO/evXmTxJGkz33uc7r33ntVV1ennp4ep8MBAACQxIocAAAwgrfffluf//znVV5erh/96EdOhzPtLMvSjTfeqBdffFGRSEQXXnih0yEBAIA8x4ocAAAwLJ/Pp/fff1/33nuv06E4wjAM/fCHP9R7772nW2+91elwAAAASOQAAID0enp61NLSot27d+d1Qd2LL75YDzzwgFpbW/Xwww87HQ4AAMhzbK0CAABDvPfee1q8eLEWLFigJ554wulwZoSamhqFQiFFo1EVFxc7HQ4AAMhTJHIAAMAQt912m5qbm3Xs2DFdeumlToczI7z77ru66qqrVFRUpO7ubp11FgubAQDA9ONfIAAAIEUkEtF9992n+vp6kjiDnHfeeXrooYfU29urnTt3Oh0OAADIU6zIAQAAto8++kjLli3ThRdeqGeeeUaGYTgd0oxz9913a/v27Xr22WdVUlLidDgAACDPkMgBAAC2uro67dixQy+88II++9nPOh3OjHTq1CmtWrVK8XhcR48e1bnnnut0SAAAII+wtQoAAEiSXn75Zd1111268847SeKM4KyzztKDDz6o119/Xd/73vecDgcAAOQZVuQAAACdOnVKpaWleuedd/Tcc8/p7LPPdjqkGe+hhx5SVVWVnnjiCa1Zs8bpcAAAQJ7gX2kAAEC7d+9WX1+fjhw5QhInQ1//+td14MABfeMb39ALL7ygwsJCp0MCAAB5gBU5AADkuVgspkWLFmnz5s268847nQ4nq5imqZKSEl155ZV67LHHnA4HAADkARI5AADkudWrV+sPf/iDIpGIzjnnHKfDyTo///nPdd1112nv3r365je/6XQ4AAAgx1HsGACAPPbggw/qqaeeUktLC0mccbr22mvl8/l066236vjx406HAwAAchwrcgAAyFNvvPGGLr/8clVWVuq+++5zOpys9uGHH+pLX/qSTp06pcOHD2vWrFlOhwQAAHIUK3IAAMhTt9xyi84991zV1dU5HUrWmzVrllpbW/Wb3/yG9xMAAEwpEjkAAOSh/fv365FHHtHevXt1wQUXOB1OTvjc5z6ne++9V3V1dTp06JDT4QAAgBzF1ioAAPLM22+/rc9//vMqKytTa2ur0+HkFMuydOONN+rFF19UJBLRhRde6HRIAAAgx7AiBwCAPOPz+fT++++rvr7e6VByjmEY+uEPf6j33ntPt956q9PhAACAHEQiBwCAPNLT06OWlhbt3r1bF198sdPh5KSLL75YDzzwgFpbW/Xwww87HQ4AAMgxbK0CACBPvPfee1q8eLEWLFigJ554wulwcl5NTY1CoZCi0aiKi4udDgcAAOQIEjkAAOSJ2267Tc3NzTp27JguvfRSp8PJee+++66uuuoqFRUVqbu7W2edxUJoAAAwcfyLAgCAPBCJRHTfffepvr6eJM40Oe+88/TQQw+pt7dXO3fudDocAACQI1iRAwBAjvvoo4+0bNkyXXjhhXrmmWdkGIbTIeWVu+++W9u3b9ezzz6rkpISp8MBAABZjkQOAAA5rq6uTjt27NALL7ygz372s06Hk3dOnTql8vJyJRIJHT16VOeee67TIQEAgCzG1ioAAHLYyy+/rLvuukt33nknSRyHnHXWWWpra9Prr7+u733ve06HAwAAshwrcgAAyFGnTp1SaWmp3nnnHT333HM6++yznQ4prz300EOqqqrSE088oTVr1jgdDgAAyFL8iw4AgBy1e/du9fX16ciRIyRxZoCvf/3rOnDggL7xjW/ohRdeUGFhodMhAQCALMSKHAAAclAsFtOiRYu0efNm3XnnnU6Hg/9jmqZKSkp05ZVX6rHHHnM6HAAAkIVI5AAAkIM8Ho+OHz+uSCSic845x+lwMMjPf/5zXXfdddq7d6+++c1vOh0OAADIMhQ7BgAgxzz44IPav3+/WlpaSOLMQNdee618Pp9uvfVWHT9+3OlwAABAlmFFDgAAOeSNN97Q5ZdfrvXr1+v+++93OhwM48MPP9Q111wjy7J0+PBhzZo1y+mQAABAlmBFDgAAOeSWW27Rueeeq7vuusvpUDCCWbNmqbW1Vb/5zW/4XQEAgDEhkQMAQI7Yv3+/Ojo6tHfvXl1wwQVOh4NRLFy4UPfee6/uuusu9fb2Oh0OAADIEmytAgAgB7z99tv6/Oc/r7KyMrW2tjodDjJkWZY8Ho+OHTumSCSiCy+80OmQAADADMeKHAAAcoDP59P777+v+vp6p0PBGBiGoX//93/Xe++9p82bNzsdDgAAyAIkcgAAyHI9PT1qaWnR7t27dfHFFzsdDsbo4osv1gMPPKAf/ehHevjhh50OBwAAzHBsrQIAIIu99957Wrx4sRYsWKAnnnjC6XAwATU1NQqFQopGoyouLnY6HAAAMEORyAEAIIvddtttam5u1rFjx3TppZc6HQ4m4N1339VVV12lSy+9VD/72c901lksnAYAAEPxLwQAALJUJBLRfffdp/r6epI4OeC8887TQw89pEOHDmnnzp1OhwMAAGYoVuQAAJCFPvroIy1btkwXXnihnnnmGRmG4XRImCR33323tm/frmeffVYlJSVOhwMAAGYYEjkAAGShuro67dixQy+88II++9nPOh0OJtGpU6dUXl6uRCKho0eP6txzz3U6JAAAMIOwtQoAgCzz8ssv66677tKdd95JEicHnXXWWWpra9Prr7+u22+/3elwAADADMOKHAAAssipU6dUWlqqd955R88995zOPvtsp0PCFHnooYdUVVWlJ554QmvWrHE6HAAAMEOQyAEAZOSnP/2pfvjDHzodRt778MMPdeTIES1atEgFBQVOhzOs1atX6xvf+IbTYUyY0/P+l7/8pT71qU9p3rx5jsWQz3JlHgMAcgtbqwAAGfnd736ncDjsdBh5b9asWbrmmmtmdBLn4MGD+tWvfuV0GJPC6Xl/1VVXkcRxSC7NYwBAbmE9NgAgYwUFBero6HA6DMxwK1eudDqEScW8z0+5No8BALmDFTkAAAAAAABZgkQOAAAAAABAliCRAwAAAAAAkCVI5AAAAAAAAGQJEjkAAAAAAABZgkQOACAr+P1++f3+EWL+A7cAACAASURBVPskEgmFQiF5PJ5piipzMzk2OI/5DQAAMsXjxwEAOWPbtm3as2eP02GkNVWxmaapl156Sb/+9a8VDofV2dk56a+BmSEf53csFtPdd9+tPXv2yOv1qqKiQmVlZZP+OgAAZBNW5AAAskIgEFAgEBixT1NT0zRFM3ZTFVt9fb3279+vjRs3KhwOT8lrYOoxv4cyTVPRaFRNTU0aGBhQaWmpysvLmecAgLxHIgcAgCyWSQIAyEaHDh2S2+2WJBUUFGj9+vWSxPYtAEDeI5EDAJjxhqu/YZqmQqGQDMOQx+PR8ePHJ2X8cDhsjxmLxYZ9TcMwFAwGlUgkhoyZaWyJREINDQ12v+7ubvtYJnVTkP2Y3+klkzhn8nq9o14zAAC5jBo5AIAZb8OGDWm3U1RVVWnu3LkaGBhQQUGBQqHQhMfv6+uT2+1Wf3+/5s2bp7lz56ZsG6mqqpLb7ZZlWUokEva5bW1tKigoGFNsyfNvuukmWZal7u5ulZeXKxKJqKSkZFzXguzD/M6MaZqSpDVr1oz5XAAAcooFAEAG7r//fquwsNCx15dkDf5rq7Oz05Jkvfzyy3bbwMDAkH7jHT9dW1dXlyXJisfjdltvb68lyWpvbx9zbO3t7Wlfs7a2dlLid8q1115r3XLLLU6HMSmma94zv0fX1dVlud1ua2BgYFznj1UuzWMAQG5haxUAICsdOHBAkjR//ny7bfCKganQ0dEhSSosLLTbFi5cKEnat2/fmGNLnpPcxmIYhiSprq5ukiNHtmF+D7Vz507dfvvtU/4+AAAw07G1CgCQlZx4DHO610zeVA7eGpNpbMlzLMuahOiQS5jfqUKhkNxut5YvXz7hsQAAyHasyAEAIEPJ4qvpir9OpADreIvYApNpps7vaDSqY8eOqbq6ekLjAACQK0jkAACyUnNzs6TTN3nT5aabbpIknThxwm5LFmCtqKgYc2zJfm1tbfY4yaf8IL8xv2X3f/rppxUIBOy2aDSqmpqajMcAACDXkMgBAMx4g1cIJL+//vrrJZ1+hHHyEcqDH208lhu9weMnbziTfw4+vnr1arndbu3YscNue/LJJ+X1elVWVmb3zzS2G2+8UdLpmiEul0uGYaioqMi+ac708eODYx38PbID83v4uDds2CCfz5dSZ2fJkiU8uQoAkNdI5AAAZryioqIh3xcXF6u/v19z587VvHnzVFNTo0WLFsntdqu9vV3bt28f1/gulyvlz8HHCwoK1NLSIrfbraKiIrt46z333JMyXqaxFRYWqr+/X7W1tZJOb1/p7+9XcXFxxrEbhpESa/KGGdmD+Z3etm3b0j6WXZIWLFiQ0RgAAOQiw6LCIgAgA42Njaqrq1M8Hnc6FMxwK1eu1OLFi7Vr1y6nQ5kw5n3+yqV5DADILazIAQAAAAAAyBIkcgAAAAAAALLE2U4HAADAVMm0Vgy7jJGNmN8AAOQnEjkAgJzFDSxyGfMbAID8xNYqAAAAAACALEEiBwAAAAAAIEuwtQoAMKOcWfdjLNtHJnIuMFPwGQAAACNhRQ4AYNLFYjEZhjHkq6Ghwe7T3d095PhglmWl3ITGYjHV1NTIMAzV1NSou7t7yOueec5MEg6H5fF45PF4FA6HJ+28RCKhYDBov4ehUChtH7/fP2qf0cZB5qbiM5DJ75HPAHMXAJD7SOQAACZdcXGxBgYG1NXVJbfbLUlqb2/Xli1b7D5lZWWKRCJyu93q7e0d8ebTNE1Fo1E1NTVpYGBApaWlKi8vH9PN4FhEo1H5/f5JGy8UCikYDKqtrU1tbW06cOCAgsHghM8zTVMbNmyQdPoGPh6Pa9++fSmxJxIJnThxQoFAQJZlqb29XZWVlSkJhUzGwdhM9mcgk9/jZMqlzwAAADnHAgAgA/fff79VWFg45vPi8bjldrstt9ttDQwM2O39/f2W2+224vF4Sn9J1pl/PXV2dg4ZN12/TI6NFGd7e7vl9Xqt9vb2IXGNV39/vyXJ6u3ttdsikYglyYpEIhM6r7293ZKU8r4m+3R1dVmWZaWcn3Tm+5PJOGNx7bXXWrfccsuYz5uJxjvvB5uMz0Amv8dMj40UZy5+BsYrl+YxACC3sCIHADClCgsLFQgEFA6H1dLSYrfffffdamlpUWFh4ahjJFc0nMnr9U44vuTKg8bGRi1cuFBNTU1av359RnFl4vDhw5KkSy65xG6bM2eOJOnIkSMTOm/fvn2SpIKCArvPZZddJknq6OiQJC1fvjxlXNM0JUm1tbV2WybjYPwm4zOQye9xvHL9MwAAQK4hkQMAmHIlJSXq7OyUz+dTd3e3gsGgtm/fPu4bxeRN7Jo1a8Z1fiKRUCgUUk1NjV566SVt2rRJgUBAJSUlQ/qmq3OS7ms4PT09kk5vtUlKXvdIW8MyOS/d+ckb2j179gw5FovFVF9fL0mqqqqy28c6DsZuMj8Dw/0exyJfPwMAAOQCEjkAgGnhdrtVX1+v8vJyfeYzn5nQ//b/8pe/lNvt1sqVK8d1flFRkU6ePKl77rln1JUH1v8Vjx3tazgj3UyOdBObyXnJFUnHjx8ftm9SLBbTvHnzVFdXN+S1xzIOxm8yPgMj/R7HIh8/AwAA5AoSOQCAaVNRUSG3262dO3faq2rGY+fOnbr99ttTtlOMRTwe19y5c7V161aFQiElEolxx+Kkm2++WZL0gx/8wH4/o9GoJNkrNpKKi4tlWZYikYhqa2vl8/nsorFjGQcTM9HPwEi/x7HIx88AAAC5gkQOAGBaJBIJvfLKK9q1a9eQWiFjEQqF5Ha7h9QMGYvCwkKtX79eTU1NWrhwoRobG+X3++0bwMEmuq1kuPo+0sg1fjI5b/ny5erq6tLJkyflcrkUDAb15ptvSpJWrVqV9tySkhJ7O87GjRvHPQ7GbrI+A1L63+NY5PNnAACAbEciBwAwLR5//HGVlZWpuLhYzc3N8vl86uvrG9MY0WhUx44dU3V19aTFVVJSokAgoE2bNumll15STU1NygqFiW4rSd6MDl7xEIvFJElLly6d8HllZWXq7OyUZVmqrq7W888/r9ra2rS1TpLmz58/pG0842BsJuMzMFi63+N45ONnAACAbEYiBwAwpUzTVENDQ0rypbq6Wm63Wzt27Mi4tkUikdDTTz+tQCBgt0WjUdXU1ExKnOlWKEyG66+/XpJ04sQJu+21115LOTZZ54VCIfX09Mjn840YU3ILSnt7+4TGQWYm6zOQblxp+N/jWOXTZwAAgGxGIgcAMGWi0aiqqqrSbnFoa2tTOBzWggUL1N3dPeI4iURCGzZskM/nS9nKsWTJknE/uWokyRUKkyG5+qK1tVWmaco0TbW2tqq5uTnlaTx+v19+v3/M55mmaSe0Tp48qc7OzpTaQR6PRw0NDfZKBtM0VV9fr9raWq1fvz7jcTA+k/UZyPT3OFly6TMAAECuOdvpAAAAuWlwzYxwOKyuri6VlZVJ+vjJO0nl5eWSNOz2jG3btg37dJsFCxZMVshTprq6WuFwWC6XS263W5s3b7bfi4mcl3yPm5ub5fV6024lqa6ulsfjsVco1NfX68tf/nJKjaFMxsHYTeZnIJPf40zm5GcAAIBcY1gjbWoGAOD/NDY2qq6uTvF4fEpfJ3ljNpG/niZjDIzfypUrtXjxYu3atcvpUCZsuub9YHwGZoZcmscAgNzC1ioAAAAAAIAsQSIHAAAAAAAgS1AjBwAwI41na8jgmiRAtuMzAAAA0iGRAwCYUSZS04N6IMgFfAYAAMBI2FoFAAAAAACQJUjkAAAAAAAAZAkSOQCAGSWRSCgUCsnj8TgdCjDpmN8AAGCiSOQAAGaUbdu2qbKyUuFw2OlQ0jJNU319fQoGgxO6GQ8Gg2kL04bDYXk8Hnk8nmHfg7H0MQxDHo9HoVBo3LFi8uTy/I5GozIMw/6qqakZ09imaaacP/jrzPmbyWcAAIBcRbFjAMCM0tTUpD179jgdxrDq6+slSXV1deMeIxqNauPGjUPaQ6GQ9u3bp7a2NknS1q1b9d///d+qrq4eU5+Ghgb5fD5FIhF1dnYqGo1qyZIlOnnypLZs2TLuuDFxuTy/jxw5kvLzmjVrxjT2Sy+9NOzYZWVl9veZfAYAAMhlJHIAABiDQCAgafyJHNM09eijjw5pj8ViqqysVG9vrwoKCiRJXq9XS5Ys0bJly1RSUpJRH0ny+XySZP+c/LOnp4dEDkY0kfk9e/bsEZ+aNdrYr776qvr7+1VcXGy3JRIJNTY2qrCwUFJmnxMAAHIdW6sAAFnBNE2FQiF7q0UwGFQikRjSr6GhIeX4mduXRjru9/vl9/un9DpaWlq0adOmIe2HDx+WJF1yySV225w5cyR9vNIhkz7Sxysf+vr6JJ2++ZU+vpHGzJPt8zsWi8nj8cjv99vzbqzKyspSkjiS1N3drbVr19o/Z/oZAAAgl5HIAQBkhaqqKv35z3+WZVmKx+MKh8PasGGDTNO0+zQ0NKiiokKWZWndunVqbGxMGWO041Otu7tb11xzjb26YLCenh5JSrmRTfZL1gDJpI8kbdmyRbW1tVqxYoX6+vp0+PBhxeNxVivMYNk+v6PRqKTTq21WrFghj8eTNhE1kuE+F4PnbaafAQAAcpoFAEAG7r//fquwsHBaXkuSNfivqK6uLkuSFY/H7bbe3l5LktXe3p5y3uA+8Xg8ZZzRjk8kxtHE43Grubl52POHG29weyZ9BvN6vZYkq7a21hoYGMg41om69tprrVtuuWXaXm8qTcW8z8X5bVmWNTAwYEUiEau2ttaSlDLfxzN2JBJJuf6Rzh1PvKPJpXkMAMgtrMgBAMx4HR0dklL/x37hwoWSpH379tltXq9XRUVFCoVCMk1ThYWFKTU7Rjs+lR5//PFpLcba0NCg0tJSDQwMSDq94mPw6g7MHLkwvyWpoKBAJSUlCgQCam5unvAKmUcffTSlyDEAADiNRA4AYMZL95SfZKHTwTeL3/nOd+R2u1VZWSmXy6WGhoaUc0Y7PlXC4bCuv/76Efu43e5hj3m93oz7SKef6uPz+bR69WoVFBSoqqpK4XBYjzzyyBgjx3TI9vmdzrp16yaUyEluyzpzu1WmnwEAAHIZiRwAwIyXvHlLV3Nj8M3b/Pnz1dnZqUgkIq/XK5/Pl3IzO9rxqeLxeDRv3jy7kO3gArTJ79NdY7JI8dKlSzPuI0mVlZWSPk4GFBUVSVLaR57Dedk+v9MpKCiYUGLlzCLHSZl+BgAAyGUkcgAAM95NN90kSTpx4oTdltwmVFFRYbcZhiHTNFVSUqKmpiZFIhH7UdyZHJ8qlmUN+Rp8TJK9YmfwNb722mspxzLpIw1dtZBM6Iy0mgHOyfb5nY5pmimxj9WZRY6TMv0MAACQy0jkAABmlMH/0578fvXq1XK73dqxY4fd9uSTT8rr9Q6poVFfX2//D/1FF11kP4o7k+OZPp55cK2ZdHVnxvOY5+LiYjU3N6u1tVWmaco0TbW2tqq5udl+Qk8mfSRp8+bNkk5vsZI+fgx5sh3OycX5HQqF1N3dbf8ci8V06NChtPVtRhtbOv0ErNLS0rTHMv0MAACQy0jkAABmlOQ2oMHfFxQUqKWlRW63W0VFRfZ2pHvuuWfI+Zs2bVJHR4cMw1BHR4e2bNkypuOjMQxDLpfL/tnlcqVslZqI6upqrVmzRi6XS1VVVaqoqBhSIDmTPmVlZerq6lJPT48Mw1Bra6u6urooHDsD5OL8Pv/881VeXi7DMOT3+/XWW2+lXf2V6dijFTnO5DMAAEAuM6zpfJwBACBrNTY2qq6uTvF43OlQMMOtXLlSixcv1q5du5wOZcKY9/krl+YxACC3sCIHAAAAAAAgS5DIAQAAAAAAyBIkcgAAAAAAALIEiRwAAAAAAIAsQSIHAAAAAAAgS5DIAQAAAAAAyBIkcgAAGCSRSCgUCsnj8TgdCjDpmN8AAGQ/EjkAAAyybds2VVZWKhwOT+vrxmIx1dTUyDAM1dTUqLu7O22/cDgsj8cjj8czbIzJPoZhyOPxKBQKpRxPJBLy+/0yDEOGYQw5jtw1k+d3IpFQMBgccV6apqm+vj4Fg8G0ySjTNO3zz/xingMAcgWJHAAABmlqapr21zRNU9FoVE1NTRoYGFBpaanKy8uH3GyHQiEFg0G1tbWpra1NBw4cUDAYTOnT0NAgj8ejQCAgy7IUCARUWVmphoYGSadvlk+cOGEfb29vTzmO3DZT57dpmtqwYYMkybIsxeNx7du3T36/P2Ws+vp67d+/Xxs3bkybjHrppZeGjaOsrGySrggAAGcZlmVZTgcBAJj5GhsbVVdXp3g87nQoU84wDEmnbyinQzgcltvtHjGGWCymefPmqbe3V8uXL5ckRaNRLVmyRJFIRCUlJcPGbhiG3G63Ojs71dfXZ58/3GtN1MqVK7V48WLt2rVrUsZzUi7O+5k4v0OhkCorKzUwMKCCggJJH8/vrq6uIUmY4a4hFArpi1/8ooqLi+22RCKhxsZGBQKBMcWdS/MYAJBbWJEDAHBEQ0ODDMNQMBhUIpGwb8yk0/87P3iLhd/vVyKRkDS0xkc4HLa3a8RiMUmnb+bObEskEvaWI0n2+DU1NTp+/Pio8SYSCTtmj8czZGvISNfj9/uHrCwY7Myb3CSv12t/f/jwYUnSJZdcYrfNmTNHknTkyBG7rb6+XpLU19cnSfb1J29iz0zimKYpSaqtrR02Powd8/tjmczvffv2SZKdxJGkyy67TJLU0dExavxJZWVlKUkcSeru7tbatWszHgMAgBnPAgAgA/fff79VWFg4KWPV19db/f39lmVZ1sDAgFVbW2sN/ivJ6/Vakqx4PG719/dbkiyv12tZlmW53W5LkiXJikQilmVZVm9vr92nt7fXsixryHnJcyTZfQYGBuzXevnll+3XT/ZLisfjltvtttrb2y3Lsqyurq6U1x/tempra63a2tqM35+BgQFLktXZ2TnkPTmTJMvtdqe0JV+/t7fXam9vt+LxeNrX6e/vt/sOvv6Juvbaa61bbrll0sZz0njmPfN7ZOnm95kxjbc9neR7NFa5NI8BALmFRA4AICOTmchJ3sQmxePxITeGg2++zrxpS3cTl0lbuj6RSMSSZNXX1w/br729Pe3YyZvX0a5nrLq6uiy3220NDAyMGPtI7ckb+Nra2pRxkpKJgOTX4OufqFy6AR7PvGd+jyzd/E6XcBrumkZqP1MkErETVGOVS/MYAJBb2FoFAJh2Xq9XRUVFCoVCMk1ThYWFKbUuAoGAmpqaFIvFprwIb7K2jM/nG7ZPctvH4CfgSFJdXZ2k0a9nrHbu3Knbb789ZZvJWDQ0NKi0tFQDAwOSpKqqKnsLVVJxcbEsy1IkElFtba18Pt+QwskYH+b3yNLN75tvvlmS9IMf/MCeq9FoVNLH2wXH49FHH6XIMQAg55DIAQBMu+985ztyu92qrKyUy+VKezMbDAb17W9/e9j6GtMp+XQc6/RK1pQvKbPryVQoFJLb7R5Sy2ak92FwrZFQKCSfz6fVq1eroKBAVVVVCofDeuSRR9KeW1JSoqqqKknSxo0bxx03Psb8Ht5w83v58uXq6urSyZMn5XK5FAwG9eabb0qSVq1aNa7XStYdKiwsHHe8AADMRCRyAADTbv78+ers7FQkEpHX65XP50u5OQyFQtq4caN27dql+fPnT0tMg5MhwxmuaOxo15OpaDSqY8eOqbq6esix5A1/8uZU+riQ8dKlS+22yspKSR8XjS0qKpI0cpJmut7jfMH8Tm+k+S2dLlTc2dkpy7JUXV2t559/XrW1tfaqorGiyDEAIFeRyAEATDvDMGSapkpKStTU1KRIJJKy9SOZjDjz6TNTIXnzumbNmmH7NDc3S5La2trsbR/Jp/xIo19PJhKJhJ5++umURyRHo1HV1NRIkq6//npJ0okTJ+zjr732WsoxaejKnWRCZ6SVH8lram9vH1PMSI/5PdRo8/tMoVBIPT09Y36dwXp6esadBAIAYCYjkQMAcER9fb29ouSiiy5KqYORTDrEYrGUVQKJRCJlRcrgm87BfYZrSwqFQvb5bW1tcrvdaVe8JL+/8cYbJZ2uGeJyuWQYhoqKilRRUZHR9Yz2eOZEIqENGzbI5/Ol1ClZsmSJfQNeXFys5uZmtba2yjRNmaap1tZWNTc3pyQENm/enHKNyceQJ9s9Ho8aGhrsWE3TVH19vWpra7V+/fphY8TYML8/lsn8TsabTO6cPHlSnZ2daetEDa73dGbtp6RoNKrS0tJhYwIAIJuRyAEAOGLTpk3q6OiQYRjq6OjQli1b7GPJ/7UPBoNyuVyqra2V1+vV+++/b28VkiSXyyVJKW3J79O1JS1cuFAej0cul0vFxcVqa2tL2zf5fWFhofr7+1VbWyvp9DaV/v7+lATKSNczmm3bttl1Ss60YMEC+/vq6mqtWbNGLpdLVVVVqqioGLJNpaysTF1dXerp6ZFhGGptbVVXV5dd8LW6ulo+n0/z5s2TYRhqaWnRl7/85ZSVEpg45vfHMpnfhmHI5XLpyJEj8nq9w46f7JeUTDydiSLHAIBcZlgTeewAACBvNDY2qq6uTvF43OlQxi15w8dffVNr5cqVWrx4sXbt2uV0KBOWTfOe+T25cmkeAwByCytyAAAAAAAAsgSJHABAXhippgiQ7ZjfAADkDxI5AIC8MFJNESDbMb8BAMgfZzsdAAAA04G6IchlzG8AAPIHK3IAAAAAAACyBIkcAAAAAACALEEiBwAAAAAAIEuQyAEAAAAAAMgSJHIAAACANE6ePKm33nrL6TAAAEjBU6sAABlLJBIyDMPpMJAFFi9e7HQIk4Z5n78uuugiXXHFFWppadENN9zgdDgAAEiSDIvnVQIAMvDKK6/oV7/61Yh9PvzwQz344IN66qmnVFFRobVr105TdM4wTVMbN27Utm3bdPnllzsdzozyN3/zN7ryyiudDmPCMpn3uYr5Lc2ePVsPPfSQmpubVVVVpd27d+uCCy5wOiwAQJ4jkQMAmBSxWExf+9rXdOzYMQWDQX3ta19zOqQpF4/HNXv2bB08eFClpaVOhwNMKub3xx599FHV1NTo/PPP1wMPPKDrrrvO6ZAAAHmMGjkAgAkLh8NasmSJ3n77bfX19eVFEgdA/li7dq2OHTumpUuXqry8XN/61rf0l7/8xemwAAB5ikQOAGDcPvroI91xxx366le/qq985St67rnn8nYLBoDcVlhYqB//+Md6+OGH9eijj6qkpES/+MUvnA4LAJCHSOQAAMblj3/8o/72b/9W//Iv/6K9e/fqwQcf1Hnnned0WAAwpSoqKvTiiy9q4cKFuu6667R161Z98MEHTocFAMgjJHIAAGPW3d2tq6++Wm+88YaeffZZbdiwwemQAGDazJkzR52dndq9e7d2796tq6++Om+LYgMAph+JHABAxizL0ve//3393d/9nVatWqWjR4/qiiuucDosAJh2hmFo48aNeuGFF/TXf/3XWr58ubZu3aoPP/zQ6dAAADmORA4AICN/+tOfdMMNN2jbtm3613/9V/3Hf/yH/uqv/srpsADAUf/v//0/PfPMM9q1a5caGxv1pS99Sb/97W+dDgsAkMNI5AAARnXo0CEtWbJEL7/8sg4dOqRbb73V6ZAAYMZIrs557rnndOrUKV155ZX6/ve/r1OnTjkdGgAgB5HIAQAMy7Is3XfffVq1apW+8IUv6Pnnn9eyZcucDgsAZqTLL79cvb29uuOOO/TP//zPWrlypV555RWnwwIA5BgSOQCAtN5880195Stfkc/nUyAQ0GOPPaaLLrrI6bAAYEY7++yz9d3vfldHjx7VX/7yF5WUlOi+++6TZVlOhwYAyBEkcgAAQxw9elRf+MIX9Otf/1o9PT367ne/K8MwnA4LALLGFVdcoSNHjui2227Tli1btHr1av3xj390OiwAQA4gkQMASNHc3KxrrrlGn/70p3X06FF98YtfdDokAMhKs2bN0h133KFf/OIX+sMf/qBFixapubnZ6bAAAFmORA4AQJL09ttva926dfqnf/onfe9739NTTz2lwsJCp8MCgKy3fPlyRSIReb1e1dTUqKKiQn/605+cDgsAkKVI5AAA9Pzzz+uqq65ST0+PfvrTn+qOO+7QWWfxVwQATJZzzz1X99xzj372s5/pyJEjWrRokR577DGnwwIAZCH+lQ4Aee7BBx/Ul770JV166aWKRCJatWqV0yEBQM5atWqVfv3rX+urX/2q/uEf/kHr1q3T//zP/zgdFgAgi5DIAYA89c477+jrX/+6/vEf/1GbNm3S008/rTlz5jgdFgDkvAsvvFB79+7Vk08+qcOHD2vRokV64oknnA4LAJAlSOQAQB767W9/qxUrVuipp57SgQMHdM899+gTn/iE02EBQF654YYb9OKLL2rVqlXyeDz61re+pXfeecfpsAAAMxyJHADIM21tbbr66qt17rnn6rnnntMNN9zgdEgAkLdcLpcefPBBPfzww/rxj3+sK664QgcPHnQ6LADADEYiBwDyxPvvv69bb71VN998s775zW/qv/7rv3TZZZc5HRYAQFJFRYVefPFFlZSUqKysTN/61rf07rvvOh0WAGAGIpEDAHngd7/7nZYvX64f/ehHeuSRR3Tfffdp1qxZTocFABikqKhIP/nJT/Twww+ro6NDX/jCF3T06FGnwwIAzDAkcgAgx/3kJz/RsmXL9IlPfELPP/+81q5d63RIAIARVFRUKBKJaM6cOVqxYoW2bt2q//3f/3U6LADADEEiBwBy1EcffaStW7fq7//+77Vu3TodPnxYn/70p50OCwCQgeLiYv3nf/6ndu/erd27d+vqq69WJBJxOiwAwAxAIgcAclAsFtO1116rf/u3f1N7e7v27t2rT37yk06HBQAYg//P3p3HN1Xl/+N/hU0WtTAjYAyIKwAAIABJREFULYtQYLCAA7TsBYRKAaFoAqOCLQqMn4GSwiDysR9BTb/gtAqMqajFAVNmkEWSARylYdW2toi0IEgqgiyCtAiSoJgIOBSB8/uD370mbdombdqbpK/n45GH5S7nvm/MOTk59ywqlQqJiYkoKipCq1atMHDgQCxatAg3b95UOjQiIlIQG3KIiILM1q1b0adPHzgcDhQUFCA+Pl7pkIiIqAa6dOmCTz75BK+99hqWLFmCBx54AMePH1c6LCIiUggbcoiIgsSNGzewaNEijB8/Hg8//DAOHDiAP/7xj0qHRUREPtCgQQPMnTsXBw8exK+//oo+ffpg6dKluHXrltKhERFRHWNDDhFREDh37hxGjBiBpUuXYuXKlVi7di2aN2+udFhERORjf/zjH1FYWIiFCxfi//2//4eYmBicOnVK6bCIiKgOsSGHiCjAffLJJ+jfvz8uXryIffv2YcaMGUqHREREtahRo0aYP38+Pv/8c1y+fBl9+/aFwWBQOiwiIqojbMghIgpQQggsXboUo0ePxpAhQ7Bv3z707t1b6bCIiKiO9O7dG4WFhUhKSsKsWbMQFxeHc+fOKR0WERHVMjbkEBEFoIsXLyIuLg4LFy5Eeno63n//fYSEhCgdFhER1bGmTZtiyZIl+PTTT3Hq1Cn07NkT69atUzosIiKqRWzIISIKMLt370ZUVBSOHTuG/Px8zJ07V+mQiIhIYYMHD0ZRURFmzpyJP//5z5g0aRJ++OEHpcMiIqJawIYcIqIAIYTAm2++iVGjRqF///44dOgQBg0apHRYRETkJ5o1a4YlS5Zg586dKCwsRM+ePbFlyxalwyIiIh9jQw4RUQBwOBx4/PHHkZycjNTUVHz44Ydo1aqV0mEREZEfGj16NA4fPozx48djwoQJmDRpEn766SelwyIiIh9hQw4RkZ87cOAA+vTpg/379yMvLw/z58+HSqVSOiwiIvJjISEheOedd7Bt2zZ89tlniIqKQk5OjtJhERGRD7Ahh4jIjxkMBgwdOhRdunTBgQMHMHToUKVDIiKiADJu3DhYLBYMGDAAo0ePxsyZM3HlyhWlwyIiohpgQw4RkR+6fPkynnjiCcyaNQsvvPACPvroI4SFhSkdFhERBaDWrVtj8+bN+Pe//433338fvXv3xu7du5UOi4iIqokNOUREfsZisaBv377Iy8vDjh07sGjRIjRowOKaiIhqZuLEifjqq6/Qq1cvjBgxAnPnzkVpaanSYRERkZf4y4CIyI+sXbsWQ4cORfv27WGxWDB69GilQyIioiDSpk0bbNmyBatXr8bq1avRr18/HDx4UOmwiIjIC2zIISLyA//9738xffp0/PnPf8b06dORnZ2Ntm3bKh0WEREFqalTp+Lw4cMIDQ1FdHQ0FixYgOvXrysdFhEReYANOURECjt27BgGDhyILVu2YNu2bXjzzTfRqFEjpcMiIqIgFx4ejpycHLz99ttYvnw5BgwYgKKiIqXDIiKiKrAhh4hIQevXr0f//v3RrFkzfP7554iLi1M6JCIiqkdUKhUSExPx5ZdfIiQkBNHR0Vi6dClu3rypdGhERFQBNuQQESng2rVrmDt3LqZOnYq//OUv2LNnDzp16qR0WEREVE916dIFubm5WLRoERYuXIhhw4bh5MmTSodFRERusCGHiKiOnTx5EoMHD8a7776LjRs34s0330STJk2UDouIiOq5Ro0aYf78+Thw4ABKS0sRGRmJpUuX4tatW0qHRkRETtiQQ0RUhz788EMMHDgQDRo0wBdffIHHH39c6ZCIiIhc9OzZE4WFhVi4cCFSUlIwduxYnD17VumwiIjo/8eGHCKiOnDjxg0sWLAAjz76KNRqNfbs2YM//OEPSodFRETkVuPGjTF//nzs2bMHZ8+eRa9evWAwGJQOi4iIwIYcIqJad/bsWQwfPhxvv/023nvvPaxduxbNmjVTOiwiIqIqDRw4EIcOHYJWq0VSUhIefvhhnD9/XumwiIjqNTbkEBHVoq1btyIqKgp2ux2FhYVISEhQOiQiIiKvNG3aFEuWLMGnn36KEydOoGfPnnjvvfeUDouIqN5qpHQARETB6MaNG0hLS0NqaiqefPJJrFixAi1atFA6LHLj2rVrsNlsOH/+PC5evAir1Yrvv/8eNpsNFy5cgN1ux82bN/Hzzz8DAOx2O4QQuHLlCq5fv46mTZti/PjxaNSoEUJCQgAAISEhaNCgAVq0aIEmTZrgzjvvRNu2bREaGoqwsDCXv9u0acPPBhEFhCFDhqCoqAgvvPACpkyZgi1btmDFihX4/e9/r3RoRET1ikoIIZQOgogomNhsNkyePBmfffYZ3nzzTSQmJiodUr1XWlqKkydP4vjx4zhx4gSOHz+OY8eO4eTJk7h06ZLLsSEhIWjbti1at26NNm3aoFWrVmjUqBHuuusuAEDLli2hUqlw5513onHjxvJ5zo09DocDt27dwi+//ILS0lJcuXIF58+fh81mg81mww8//OByzbvuugv33XcfIiIi0L17d3Tr1g0RERHo1q0bG3lIMVarFW3atEFeXh5iYmKUDof8zK5duzB9+nTcuHED77zzDjQajdIhERHVG+yRQ0TkQ5988gkmT56Mu+++G/v27UPv3r2VDqnesdvt2LdvH/bt24f9+/fj66+/RnFxMW7evIkGDRogPDwc3bp1Q3R0NKZNm4Z7770XrVu3Rrt27RAaGoqmTZvWeozXr1/HxYsXceHCBVy4cAHff/+93MC0YcMGfPvtt/j1118BAB06dEC3bt0wYMAAREdHY9CgQQgLC6v1GImIKjNmzBh89dVXeP755zF+/HhMmTIFb7/9ttzoTUREtYc9coiIfEAIgb///e946aWXoNFosHr1anmYDdWemzdv4quvvkJhYSEKCwuxb98+HDt2DEIIdO7cGYMGDUKvXr3k3i0RERG44447lA67Sr/++itOnz6NY8eO4cSJEzhy5Aj2798v31unTp0QHR0tN+z06dMnIO6LAgt75JCnNm/ejFmzZqF58+ZYvXo1RowYoXRIRERBjQ05RD6ya9cuWK1WpcMghRQUFGDVqlWYPHkyRo4cWeXxY8aMYa+Karp48SLy8vJgNpuxdetW/PTTT7jzzjsRGRmJfv364YEHHsDw4cOD8v29fPkyioqK8Nlnn2HPnj3Yt28fLl68iGbNmmHo0KEYNWoUNBoNevTooXSoAefYsWPYv3+/0mH4lcuXL+Pll1/GrFmz0KVLF6XD8Svdu3fHwIEDlQ7DL0j1H4fDgXfffReHDh3C1KlTERsbq3Ro5IdY/yHyDTbkEPnI8OHDsXfvXjRqxBGL9ZUQAiqVqsrjSktL+YTbCzdv3kRhYSF27NiBnTt34osvvsAdd9yBmJgYxMXFITY2Fvfffz8aNmyodKiKOHHiBPLy8rBz505kZ2fj8uXL6NatG8aNG4e4uDgMHz6cvXU8kJGRgWeeeYbvFVXp+vXrmDVrFpYvX650KH6hbP1HGsbqyfch1S+s/xD5Dn9xEvmQVqtlxY4qJQ1VoMoJIVBQUID169dj48aN+PHHH9GlSxfExcXh5ZdfxogRI9C8eXOlw/QLERERiIiIQGJiIq5fv449e/Zg586d2LFjB5YtW4Y777wTEyZMwFNPPYVRo0bV2wYvT4SGhrJnJVVp+PDhSofgd1j/oaqw/kPkW2zIISIiv/HNN99g/fr1WL9+PU6dOoWePXvi+eefx4QJExAREaF0eH6vSZMmiI2NRWxsLP7+97/j7NmzyMrKwnvvvYexY8eibdu2SEhIwJQpUxAVFaV0uERERERUDQ2UDoCIiOq3q1evYuXKlRgyZAjuu+8+GAwGjB8/HocOHcLhw4fx/PPPsxGnmjp06IDZs2dj7969OHnyJBITE7Flyxb06dMHvXr1wmuvvYYff/xR6TCJiIiIyAtsyCEiIkWcO3cOL7zwAjp27IjnnnsOXbt2xa5du3D27Fmkp6ezx4iPde3aFYsWLcLJkyfx2WefYdiwYVi8eDE6duyIpKQknDhxQukQiYiIiMgDbMghIqI6ZbFYMHPmTHTt2hWrV6/GnDlzUFxcjLVr1+Khhx7iHC61TKVSYciQIfjHP/6B8+fPY8WKFcjPz0f37t0xevRomM1mcB0EIiIiIv/FhhwiIqoTubm5GDFiBPr06YN9+/Zh5cqVKCkpwaJFi3DPPfcoHV691LRpU0ydOhVfffUVPvjgA1y/fh0ajQb9+vXDpk2b2KBDRERE5IfYkENERLXqyy+/RFxcHEaOHIk77rgD2dnZsFgsmDZtGpo0aaJ0eASgQYMGGD9+PPLz83HgwAF069YN8fHxiI6ORl5entLhEREREZETNuQQBYGUlBSkpKRUeozNZoPJZIJGo6mjqDznz7FR9X333XeYOXMm+vbtix9++AE5OTnYuXMnRo4cqXRoVIl+/frBaDTi8OHDCA8Px4gRIzB69GhYLBalQwtqLMeJvMd8Q0T1FRtyiOqJhQsXIiEhAWazWelQyqmt2Gw2G1JSUqBSqaBSqWAymXyaPrnncDjw/PPP47777kNOTg7ee+897N+/H7GxsUqHRl64//77sXHjRmRnZ+Onn35Cv3798PTTT+PcuXNKh1Zv1bdy3OFwyOV32RfLc/JUfcs3ErPZDI1GA5VKBY1GwzxDFGwEEfnEsGHDxOzZs5UOo1IAhL9me1/HZrVaRUFBgfxvo9EoAAi9Xu+za1THhQsXBACRl5enaBy1xWw2i3vvvVe0bt1avPXWW6K0tFTpkMgHbt26JTZs2CA6d+4sQkJCRGZmprh165bSYfnUW2+9JUJDQ5UOo0r1qRwvKCiQ0yz7slqtPruOtwLh+74uBcL7UZ/yjRBC6PV6AUBYLBYhhBAWi0XxOlCw13+I6hp75BBRUDp9+jSio6Plf8fHxwMAkpOTlQopqP3000+YOXMm1Go1Bg8ejKNHj2LOnDmcAydIqFQqJCQk4MiRI9BqtdBqtYiJicHJkyeVDo2C2JkzZ1BcXAwhhPyyWq3Q6XQIDQ1VOjwivyXVdSIjI13+m5+fr1hMRORbbMghCnAVja92OBwwmUxyl9oTJ074JH2z2SynWVJSUuE1VSoVMjMzYbPZyqXpaWw2mw3p6enycbm5ufK+qsbFOzfiSNcEAJ1O59mNk8fMZjN69uwJs9mM//znP9i4cSNXoQpSzZo1w5IlS3DgwAFcuXIFUVFRWLp0KW7evKl0aAGN5bh7sbGx6Nixo8u23NxcPP744x7fOwUv5puK6fV6AEBhYSEAyPGmpqZ6ePdE5O8aKR0AEdXM9OnT3Y6tnjJlCtq3bw+73Y6QkJBqj412Tr+wsBBqtRrFxcUIDw9H+/btsWLFCpdrqtVqCCFgs9nkc9etW4eQkBCvYpPOnzx5MoQQyM3NxciRI2GxWOQnS54qKSlBZmamfG3yDbvdjunTp+M///kPpk+fjtdee83l/zMFr6ioKBQWFiItLQ0pKSnYvn073nvvPdx7771KhxaQWI67567XTX5+vtzDkuo35puKPffcc7Db7Rg8eDAKCgpw5swZWK1W9mQjCiaKDeoiCjJKjhFHmfHVWVlZAoA4fvy4vM1ut1d7HLa788puy8nJKTdvgTS/gdFo9Do2aU6bstfU6XRexV5cXOwyrwLnyPGNgwcPii5duoi2bduK7OxspcMhBVksFtGjRw/RunVrsXPnTqXDqTal58hhOV41i8XiEodSAmFOmLrE+o//5hutViufa7fbvT7fl4Kl/kPkLzi0iigIbd++HQAQEREhb6vtnhKbNm0C4PoEtUePHgCADRs2eB2bdI7zKiUAkJaW5lVcHTt2hBACFosFOp0OycnJcu8cqp41a9ZgyJAh6Ny5Mw4dOsTlxOu5yMhI7N+/Hw899BDGjRuHtLQ0CCGUDivgsRwvb/PmzVz9jirFfPOb9PR0xMTEwG63A7jdG0gaZk5EgU8lWNsi8onhw4ejd+/eWL58eZ1fW/qSl7Jz2X9XdFx106/JNWt6XE2cOHEC3bp183m63rBarWjTpg3y8vIQExOjSAzVdevWLbz00ktYunQpnn/+ebzyyito2LCh0mGRH3n77bfx7LPPYtKkSfjnP/+Jpk2bKh2SxzIyMpCWlgar1arI9VmOV85msyEjI8Mv5vhQ8vveH7H+43/5xmQyISEhQR7CJdV/DAYDZsyYUe10ayKQ6z9E/og9cojIJ9RqNQC4ndxPq9VWO93qTlLojvNTMPLO9evXkZCQgGXLlmH16tVYsmQJG3GonNmzZ2P79u3Yvn07Ro0ahZ9++knpkMgL/lyOc5Jj8lf+mG8SEhIA/NbjJywsDACQmJhY7TSJyL+wIYcoCBkMBgBAUVFRnV1z8uTJAG4v+y2RuvBOnDjR69ik49atWyenI63iUF1SOkajsdpp1EdXr16FWq3Gzp07sXPnTkybNk2RODxZqaOiVUz8gT/H5kujR4/G3r17UVxcjAcffBAXLlxQOqSAxHLcVX5+vtcT3VP9w3xzm9S4JJEadMpuJ6IAVstz8BDVG0pN9me1WuWJ8qSJ9qQJftVqtSguLhZC/DYZHwCh1Wqrlb40UZ7z5HzSNe12u1Cr1UKtVsvbjEZjuWt5GpvzdZ1f0jk6na7Sif/UarXQ6/Xy8Xa7vcpz6kKgTfZ3+fJlMWTIEBEaGioOHjyoaCye/P+TJnb0x6+3uorNYDD4xf2fOXNGREREiK5du4pz584pHU6VlJzsmOV45fxlkmMJJzt2xfqP/+UbKV0p30iTL+fk5Hh8/74WaPUfIn/HHjlEAU7qLuv8d8eOHVFcXIz27dsjPDwcSUlJ6NmzJ9RqNYxGI15++eVqpd+yZUuX/zrvDwkJwapVq6BWqxEWFiaP8V6yZIlLep7GFhoaiuLiYuh0OgC3uycXFxejY8eOHsU9Y8YMJCcnIzw8HCqVCqtWrcLDDz/sF/MrBIr//ve/UKvVOHXqFPLz89G3b19F40lNTa3y/5/zcrD+pi5iKyoq8puu8+Hh4di9ezcaN26MUaNGuR12QLexHK8cJzkmd5hvKhYbG4ucnBzk5+dDpVJhzZo1yMnJYT4iCiKc7JjIRzj5IXkiUCb7u3HjBjQaDfbv349PPvkEvXr1Ujokj9XGRNm+UpuxORwO6PV6eWUTf7n/8+fPIyYmBi1atMDu3btx9913Kx2SW0pPdkyBg9/3rvh+kCcCpf5DFCjYI4eIiMqZPXs2du/ejR07dvhFI05F88s4HA6YTCaoVCpoNJpqTw5ZNn2z2SynWVJSUuE1VSoVMjMz3fY28TQ2ae4D6bjc3Fx5nyfzAklWrVqFOXPmeHrLdaZdu3bIzs6GzWbDpEmTcOPGDaVDIiIiIgpojZQOgIiI/Et6ejpWrVqF999/HwMGDFA6HADA9OnTYTaby22fMmUK2rdvLy+xajKZapx+YWEh1Go1iouLER4ejvbt27sMi5oyZQrUajWEELDZbPK569atkyeU9DQ26fzJkydDCIHc3FyMHDkSFovFq4ldc3NzMXToUISGhlbr/mtbeHg4srKy8OCDD+KZZ57BP/7xD6VDIiIiIgpY7JFDVE9JvQmqelH9smfPHixYsACLFy/GhAkTlA5HlpWVVW6b2WyG2WzGvHnz5AaUuLi4GqcfHR0NAPJ8BCtXrpT35ebmwmw2Y/z48QBuz2Xw4osvwmw2Y8eOHV7HJqUXHx8PAPL8BZs3bwbg2bxANpsNp06dkuP2V/3798fatWuxcuVKrF+/XulwggLLcSLvMd8QUTBgQw5RPSWE8OhF9ccPP/yA+Ph4jBs3Dv/3f/+ndDhV2r59OwAgIiJC3ubcI6Y2bNq0CQBcer706NEDALBhwwavY5POKfvjQZrnxhNbtmzBjBkzPD5eSY8++ijmzZsHrVaLo0ePKh1OwGM5TuQ95hsiCgZsyCEiIgDArFmz0LBhQ7z77rsB8TTSuaeMkteUGmich355Gpt0TnV/RJjNZowZM8ajY/3FkiVL0LNnT0ybNo3z5RARERFVAxtyiIgIGzduxObNm2EwGNCqVSulw/FbarUaANxObqzVaqudbnUnadZoNAgPD3c7HMBfG+MaN26MNWvW4MiRI9Dr9UqHQ0RERBRw2JBDFGBqMoab47/JncuXL2Pu3LmYPn16QPXuMBgMAICioqI6u+bkyZMBAKdPn5a3ORwOAMDEiRO9jk06bt26dXI60ipWnqisJ48/Dw3o1q0bFi5ciJdffhnffvut0uHUKZbhRNXDvENE9Bs25BApoKSkxO3Ees4/3nJzcyuteLgbfmE2m6HRaORljMuukuPP476l2DUajdvViap7nsPhQGFhITIzM8stXe1pOg6Ho8LJEKu7SpI/Wbx4MUpLS7F48WKlQ6mQcw8Y6W+p0SklJUVeItx56e6kpKRqpS81qEj/dd4fFxcHtVqNV199Vd62Y8cOaLVaeaJib2KTJk1OS0tDy5YtoVKpEBYWJjcKebP8eKD53//9X4SHh2PBggVKh+K12ijDPSmrgq0Mt9lsSElJ8ao8zczMdPtDvKioyOW9ds7/wV6GB5LayDs2m03+XFT0/zTY8o6zoqIiudwo+15Vli88Tac6+ZSI6oAgIp8YNmyYmD17tsfH2+12kZOTI9RqtQAgjEZjuWMsFotQq9WioKBA3gZAuMu6er1eABAWi0U+F4DQ6/Xljq0oDU9ZLBah0+mqfX5ZRqNRqNVqYbfbhd1uF1qtVhgMBp+cp9PphE6nq/Seq0qnoKBAPr/sy2q1enWvFy5cEABEXl6eV+fVlrNnz4qmTZuK119/XelQKlX2fZcUFxcLrVYrAAitViusVqtQq9XCaDR69f/GXfoVXdNqtQqDwSBvNxqNwm63l0vT09iKi4vlz6hWqxXFxcXyPunzW533KhCYzWahUqnE3r17lQ5FvPXWWyI0NNTj431dhntSVlWVhqf8oQy3Wq0u74vRaKzwO8s57oru3TlPAhBZWVnyPl+W4UJ4/30f7JSs/9jtdqFWq+XPm1TOVvT5Doa840yv1wu1Wi2ysrJcvjskleULT9KpTj6tiL/Vf4gCXWDU9IgCQHUrdlKlQ/oilxQXFwu1Wl2ukllRJcTddgBCrVZ7dKwncRqNRqHVar3+kVyZ4uJiAcCloiBV1qVGKV+cV9E9e5KO0WgsV0GyWq3Vqsz5W0Vm7ty54t577xWlpaVKh0L11NChQ8XDDz+sdBheN+RIfFWGe7rf02PcxelPZbjz8ZLK7stut1fa0FXRD1QhfFuGC8GGnLKUrP9IDQvO50ufv5ycnHLXDIa8I9FqtUKn07l9kCCpLF94ko63+bQy/lb/IQp0HFpFpLDQ0FCkpqbCbDZj1apV8vbFixdj1apVLsscV0aaNLSwsBAA5OEcqampNYqvqKgIKSkpyMjIQI8ePbBixQrEx8d7HFdV9u7dCwBo166dvK1t27YAgP379/v8vOqkExsbi44dO7qcl5ubi8cff9zj6/ijixcvIjMzE8nJyWjSpInS4VA99cILL2D79u11OteRL/mqDK8t/lqGR0dHu/xbGsqo0+ncHr9q1SrMmTPH7b6SkhJoNBqkpKTI34HOgrUMD3S+yDsbNmwA8NvqgQDQqVMnAMCmTZtqFJ+/5h0A8pDb1NRUl3t3VlW+8CQdb/MpEdUhpVuSiIJFTZ/QZWVlyU+QDAZDhU98UMmTEOlpZUFBQaVPjSpLQwjvnj5JaVX1qog09MRduu56E1X3vIriqMn1q8OfnkgtXbpUtGzZUly5ckXpUKgeu3XrlvjjH/8oZsyYoWgc1e2RI/FFGe7Jfk+OCYQy3Jnz8MLjx4+X25+TkyP3DHAXj/TeSy93vTncxV1d7JHjSsn6T0WfT2+3SwIh70i9drKysuShU2q1ulwPpKryhafpSKrKp1Xxp/oPUTBgQw6Rj/iiYifNc1PRl6gQVVdCpIpBZd1tPfkhodfrK+2u6yvVrWz5qvJWnetbLBa3Y/o94U8VmR49egT9j5GaVrSpbrz++uvizjvvFD///LNiMdS0IUcI35ThvmjICYQyXCINL5FeZefekOakqipdu90uz18CoNJ5RmpShgvBhpyylKz/SHWesg0LNalb+HveKTsnojS3DoByQ6EqyxfepFNVPvWEP9V/iIIBh1YR+ZGJEydCrVbjjTfecFkxx1Pp6emIiYmB3W4HAEyZMqVa6VitVrRv3x4LFiyAyWRyWdGHgM2bN7usUBSI9u3bh6+//hpPP/200qHUKuFmeW53L1LWlClTcP36dXz44YdKh1IjNS3DfSWQyvCOHTtCCAGLxQKdTofk5GRkZmbK+7ds2YIZM2ZUmU5ISAgiIyORmpoKg8FQ6eo/wVCGB5vq5p1p06YBAJYtWyafJw3TlIaceyMQ8k5ycjIAIDIyEsDtz75WqwUArFmzxuXYyvKFN+lUlU+JqO6xIYfIT9hsNnzzzTdYvnx5ufHinjCZTEhOTkZcXBxCQkIwZcoUmM1mbNy40etYQkNDER8fjxUrVqBHjx7IyMhASkqK2zksKlrStbKlQ52p1eoK90kVCl+eV9N0pEqd0vNe1NSWLVvQqVMn9OvXT+lQiHDPPffgwQcfrNbSu/6ipmW4LwVCGV5WZGQkpkyZAgBITEwEcHtZ5jFjxnichmTSpEkVfpaCpQwPJjXJO9HR0cjJycG5c+fQsmVLZGZm4scffwQAjBo1yutYAjHvAL81xqxcubLCYyrLF56m4y6fEpEy2JBD5Ce2bNkiT8hoMBiQnJxc4eR07iQkJAD4bcK/sLAwADX/opWe5MyZMwdff/01kpKSXJ5S1bTHg1SRcX7qJU3U3LdvX5+fV9N0gmWCzK1bt0Kj0SgdBpFMrVbjo48+wq+//qp0KNVS0zK8tvhrGe5ORESEy781Gg3Cw8Pd/iiu7Aeyc8+CsoKlDA8mNc07sbFeT2gKAAAgAElEQVSxyMrKghACM2bMwKFDh6DT6eRGiery17wjfbbd9VyqrHGobL6objpl8ykRKYMNOUQKczgcSE9Pd+k6PmPGDKjVarz66qs4ceKER+mU/dKVGnQq+zL2hrunVL4gPW09ffq0vO38+fMu+3x5Xk3Tyc/Pr3HlUGmXLl3CV199hdGjRysdils2mw0mk4kNTfXM6NGj4XA4YLFYlA7FK74qw2ubv5Xh7kg/KI1GIwD3P5Qllf1AdjgcmDhxott9wVCGB4vayDsmkwn5+fnysCFf8Le8I322z5w5I2+T8s7kyZMrPK9svqhJOsBv+ZSIFOLzWXeI6qnqTPZnsViEWq2WJ5pzZrfb5UnlnCf/QwWT4OXk5AgA8gSOBQUFFU4cWFEaSjEYDEKr1Qq73S5Ptld2okqdTid0Op3X5wnh+l66m8DQ03RqOkGmEP4x2d/OnTsFgCpXdVGKNNmiP31GndntdlFQUCAMBoPHq/K4I60S4u0xxcXF8nuk1WrL5XHnz3vZV00/v7Xp1q1bolWrViIjI0OR61dnsmNfluFlz6lsslV/yx/VKcPVarXQ6/WiuLhYCHH73t2V82WVvXej0ejy/hYXF4usrCy35/qiDBeCkx2XpXT9R5rQV6vVVjkJbzDkHWmb8ypUZb+PPM0XVaVT3Xzqjj/Uf4iCif+UZEQBztuKTNkfWGW/cN39CHM+z52cnJxKf+CVvbY/kZbJrGjpy4oqDlWdV9EPWm/TkWKoaeOHP1RkFi9eLMLDwxW7vif88TMqkT6LNYlRWva1svPdHWO32+XKuN1uF0ajUQBwqaBLjbjuXv7aeCcZPXq0ePrppxW5trcNOb4uwz0tqypLQ0neluFll0bW6/XlVspxp+y9O6ej0+ncNgw4x+CLPMCGHFdK1n+kbQaDodL/92WP9yfVrf9IDf3S/Ts3/nqTLzxNx5t86o4/1H+IgolKCC7XQeQLw4cPR+/evbF8+fJavY40L0BNsq4v0qDqsVqtaNOmDfLy8hATE6NIDDNmzMCZM2fw8ccfK3J9TwTCZ7S6MTocDuj1eqSlpVV4fkXHmM3mcsMly8ZhMpkwZMgQdOzYUT7GZrMhIyMDqampXsVa12bNmoUjR44gPz+/zq+dkZGBtLQ0WK3WWr0Oy/DAV1ff94GC9R/yhD/Uf4iCCefIISKqZ7799lt07txZ6TC85nA4YDKZ5ElPMzMz3S4Nm56e7rK/7KSole1PSUlBSkpKrd7HqlWrMGfOnGodU9GcV84TWEqThjoLlAleO3fu7DJfAxERERGVx4YcIqJ65uLFi/KqZoFkypQpuHz5MoQQsFqtMJvNmD59usuKG+np6Zg4cSKEEJg0aVK5SSmr2l/bcnNzMXTo0EqXPvbkGIl07+PGjZO3uTsvUCZ4bdOmTa33iCEiIiIKdGzIIQpQZZdira1zKPhcuXIFLVq0UDoMr+Tm5sJsNmP8+PEAbjdWvPjiizCbzdixY4d8XHJyMpo2bQrg9sptZXu1VLU/NTW11oYf2Ww2nDp1CtHR0TU6xtnBgwehVqsxfPjwCo8pKioKmG7sLVq0QGlpKW7cuKF0KLWOZThR9TDvEBGxIYco4IgKlmKt7XMpeFy7dg3NmjVTOgyvbNq0CYBrb5MePXoAADZs2CBv02q1CAsLg8lkgsPhQGhoqMtnvar9tWnLli0uy+xW9xhnb7zxBl588UWEhIRUeMzmzZsRGxvrcZpKat68OQDgl19+UTiS2sMynKh6mHeIiH7DhhwionqmcePGuH79utJheGXlypXltkmNF2azWd42b948qNVqJCQkoGXLlkhPT3c5p6r9tcVsNmPMmDE1PsaZyWSCWq2usocP4H64lT8qLS0FALnXFBERERGVx4YcIqJ6pkWLFrh69arSYXhFmuTX3eTGzhP9RkREICsrCxaLBVqtFsnJyS6NNVXtry0ajQbh4eFy937nLv7S354cIykqKsKRI0eq7L0TKJMcS65evYrGjRujSZMmSodCRERE5LfYkEMUYGw2G0wmEzQajdKhUID63e9+h0uXLikdhlcmT54MADh9+rS8TZrod+LEifI2lUoFh8OByMhIrFixAhaLBcnJyR7vry1lu/U7d+2X/vbkGOB2GZCdne0yl09RURGSkpLKXTdQJjmW/Pjjj2jVqpXSYdQqluFE1cO8Q0T0GzbkEAWYhQsXIiEhwWU4iT9xOBwoLCxEZmamV5WtkpISJCUlQaVSISkpCbm5uW6PM5vN0Gg00Gg0bt8Dm82GlJQUuUeDyWSq1jHBLDw83K+XeHbudSP9HRcXB7VajVdffVXetmPHDmi12nLzv+j1epSUlAAAWrVqBb1e7/F+T5cfd14py/lvb9Pxls1mw/Tp05GcnOzScycqKspl5SogsCY5lpw5cwadOnVSOoxaFaxleHXK1czMTLcT0BYVFbl8vt01UnqSDgWX+p53qqr/OB+jUqmg0WjKpWWz2eT8Uh/rP0TBhA05RAFmxYoVSodQKb1ej23btiExMdHjypbD4UBRURFWrFgBu92OmJgYjBw5stz5JpMJmZmZWLduHdatW4ft27cjMzNT3m+z2XD69GmkpqZCCAGj0YiEhASXoTOeHBPsOnfujFOnTikdRoWcl0aX/g4JCcGqVaugVqsRFhYm/2hbsmRJufPnzJmDTZs2QaVSYdOmTXjuuee82l8VlUqFli1byv9u2bJlnf2IXLhwYYX5qlu3bi7/DqRJjiWnT59G586dlQ6jVgVjGV6dcrWoqAiJiYlu9+3fv9/l32UbKT1Nh4JLfc47VdV/ACA9PR0ajUZOKzU11SUth8OB6dOnA7jdy9NqtWLDhg218tCBiOqAICKfGDZsmJg9e3adXAuA8Pfs602MWVlZVZ5fXFwsAIiCggJ5m8ViEQCExWIRQgiXfRWl48kxtenChQsCgMjLy6uT67mzceNG0bBhQ3HlyhXFYiByp0OHDuLVV19V5NpvvfWWCA0NrZNrBVsZ7m25arfbhU6nq/AYd98J1UmnttTl930gYP3Hla/zjif1n4quC0Co1WohhBBGo1EAEHa7vVw6OTk5nt1cDfhD/YcomLBHDlGQcDgcMJlMcnfZzMxMtxPDpqenu+wv25Ogsv21NVxEmsi2LOdJbPfu3QsAaNeunbytbdu2AH57elt29R5pyItOp5O3eXJMsBs4cCBu3ryJL774QulQiGTff/89zp49i0GDBikdiiICuQz3tlxdtWoV5syZ43ZfSUkJNBoNUlJSUFhYWOl1K0uH6o9gzzue1H8AyMOEpXwjDSGW5lPbsGEDgN9WfAQgD2XdtGlTDe+EiOoaG3KIgsSUKVNw+fJlubus2WzG9OnTXebvSE9Px8SJEyGEwKRJk5CRkeGSRlX764oUs3N3+vz8fABAx44d5W3SksruujCXlJTIlZopU6a4vY4nxwSj8PBw3HvvvcjLy1M6FCJZXl4eGjdujAEDBigdiiKCpQyvqlzNzc3F0KFD5fK7rKKiIgBAWloaBg8eDI1G4/ZHeVXpUP0R7HnH0/rPc889B51Oh8GDB6OwsBB79+6F1WqVJ7x3V1eSGnVWrlzp47sholqnWF8goiCjZNfinJwcAUBYrVZ5W0FBgQAgjEajy3nOx1itVpd0qtpfkxi9kZOTI9RqtUv334rSc7dd6oYsvfR6fbnzPDmmNvhL1+KZM2eKAQMGKBoDkbOEhAQxcuRIxa6v5NCqYCnDqypXrVarMBgMVV7DbrcLi8UiD5tyPsebdGoLh1a5Yv2n8hg9UVne8ab+I4QQWq1WABA6nc6lHiVtP378eI3jrQ5/qf8QBQv2yCEKAlKXWOcnkz169ADwW1da4PZQpbCwMJhMJjgcDoSGhrosa1zV/rryxhtv4MUXX3Tp/uuNjh07QggBi8UCnU6H5OTkcpMCenJMMHvkkUdw4MABfPfdd0qHQoTS0lLs2rULDz/8sNKhKCJYyvCqytUtW7ZgxowZVaYTEhKCyMhIpKamwmAwlOtJ4Gk6FPzqS97xVHp6OmJiYmC32wHc7tkj9UyaNm0aAGDZsmXyNqkHXNnVHYkoACjVgkQUbJR8IlX23xVtP378uFCr1RU+La1qf01i9JTRaCz39FUIIcfl7jparbbC9I4fP15lLJ4c4yv+8kSqtLRUtG7dWqSlpSkaB5EQQphMJtGwYUNRUlKiWAxK9sgJpjLcORbnNLKyskRxcbHX17Db7S7HVDcdX2KPHFes/1R+bW+VzTue1n/KTmYspeNcp5J6PEvbpR5NzpMm1xZ/qf8QBQv2yCEKAtJkwe7mEXCeMDgiIgJZWVmwWCzQarVITk52WeKyqv21raioCEeOHHH7pNXdPUoT+fXt27fCNCMiIqq8rifHBJsmTZrgySefxL/+9S/cunVL6XConvvnP/+JuLg4dOjQQelQFBEsZbizsuWqRqNBeHi4PCGt80SyZSeddRYSEuLyHlQ3HQpO9SHveFr/SUhIAPDbvDdhYWEAgMTERPmY2NhYZGVlQQiBGTNm4NChQ9DpdPI8OkQUONiQQxQEJk+eDAA4ffq0vE3qNjtx4kR5m0qlgsPhQGRkJFasWAGLxYLk5GSP99cmm82G7OxseXUF4HbDTlJSEgBgzJgxAFzv8fz58y773JHeB6PRWKNjglFiYiK+/fZbbNmyRelQ/JrNZoPJZIJGo1E6lKB0+PBhZGdnY+bMmUqHophgKMPLKluuCiHKvSSikiEsDofD5T2objoUnOpD3vG0/lN2BVCpQaeilUFNJhPy8/MVu08iqhk25BAFGOcnMtLfcXFxUKvVePXVV+VtO3bsgFarRWxsrMv5er1efpLTqlWrcuOiK9vv6fKbzitFOP9dUTo2mw3Tp09HcnKyy1PWqKgoeeWqjh07wmAwYM2aNXA4HHA4HFizZg0MBoO8koNGo0F6erocv8PhgF6vh06nQ3x8vMfH1Bc9evTA+PHjsXjxYqVD8WsLFy5EQkKC2xU/apPNZkNmZqacH0wmk8t+h8Phkl+cX87H2mw2pKSkVJiO0hYvXowePXq4rFIXzIKxDPdVuWoymZCbmyv/u6SkBLt37y73HlD9VF/zjif1HwB49tlnAUAu46VlyKXtUvrSQ7Jz584hKyur2vMREpHCFBnQRRSE6mqMOJxWNXDOwtIqHtJ2o9HoslqBdK7VahV6vd7tGPCq9ut0OqHT6byKr2yc7tKRVlJw9yq7ukJWVpYAINRqtcjJyXG7T3rp9XpRUFDg9TG1yd/GiO/fv1+oVCphNpuVDsWvufsc1ya73S7UarU8t4HVahVqtdol30grs7h7SauvWK1Wl8+3NIdCXa3SVpWvvvpKNGzYUGzYsEHpUOpsjpxgLMOrW66WTds5HZ1O5/G8HXWdPzlHjivWf9zH5+u8U1n9R5KTkyPXqbRarctx0jUMBkOdzIlTlr/Vf4gCnUoI9kMl8oXhw4ejd+/eWL58udKhkB+zWq1o06YN8vLyEBMTo3Q4AIBJkybhyy+/xOHDh9G4cWOlw/FL0twbdfWVaTKZkJCQALvdLj8tLSoqQlRUFHJychAbGwuTyYQhQ4a4PJG12WzIyMiQhygWFhYiOjpa0XupzNixY2Gz2XDgwAE0aKBsJ+GMjAykpaXBarUqGgf5P37fu+L7QZ7wx/oPUSDj0Coionpu6dKlOHPmDDIyMhSNIz09HSqVCpmZmbDZbC4TlzocDpdhRikpKXI3+rJz2JjNZqhUKiQlJcnd1U0mU7ltNpsNZrNZPk9KPykpCSdOnKgyXpvNJses0WhchoRUdT9VddOXls117vLeqVMnAL8ttxsbG+vSiAMAubm5ePzxx+V/l23Ekbr663S6Ku+vtmVlZWHXrl144403FG/EISIiIgokrDkREdVznTt3xoIFC5CSkoJvvvlGkRjS09MxceJECCEwadKkco1KCxYsQGJiIqxWK4qLi5GWloaFCxcCAKZPny7PYVNUVAS1Wo2CggKsXLkSixcvRmFhIeLj41FcXCxvA26v6KHRaGA2m1FYWIgZM2bAbrcDALp161ZpY440r1P79u0hhMCzzz6LkSNHoqioyKP7qYq7+XikRp2VK1cCAEJDQ8sdk5+fX+HqIyUlJfKcD1OmTPEqHl+z2+2YNWsWnnrqKQwfPlzRWIiIiIgCDRtyiIgIL774Iu677z785S9/UWQ58uTkZDRt2hTA7QaLOXPmuOy/5557oNVqERoaKvdCkRo0srKy5OOkRgypJ8rKlSvlv8ue5zy0SDrGeanjyiY3zs3NhdlsliejlCbV3Lx5s0f3k5qa6rJCW1lSDJ70DJIUFRVV2F29pKQE4eHhSEtLA1D5vdWF//3f/8WNGzfwxhtvKBoHERERUSBiQw4REaFJkyZ49913sW/fPvztb3+r8+trtVqEhYXBZDLB4XAgNDTUpaElNTUVK1asQElJCdLT02s1FqkxqLIlWaWhT84rRQGQG0qqup+qTJs2DQCwbNkyeTiU1Nun7Eorks2bN1e4uk/Hjh0hhIDFYoFOp0NycjIyMzM9jseX3nvvPbz77rt455138Pvf/16RGIiIiIgCGRtyiIgIABAVFYVly5YhNTUVH330UZ1ee968eVCr1UhISEDLli3dNtZkZmbir3/9K9RqdZ3G5o7Uo0UIUe4FeHY/lYmOjkZOTg7OnTuHli1bIjMzEz/++CMAYNSoUeWOl+YLcjfcyllkZKQ8rCoxMdGrmHzh6NGjmDlzJubNm4fx48fX+fWJiIiIgkEjpQMgIiL/kZSUhE8//RTx8fHYu3cvunfvXifXjYiIQFZWFoqKirBy5Uq5N8xzzz0H4PZkxYmJiSguLi43wW9tkYY3VebEiROIiIgot72q+/FEbGysSw+b9PR06HQ6t3PglJ3kuDLu4q0LFy9ehEajQVRUFJYsWaJIDERERETBgD1yiMgrZVcIouDzr3/9C927d8e4cePknh61TaVSweFwIDIyEitWrIDFYnEZ2pSQkAAAddKII81LM27cuAqPMRgMAIB169bJQ5+kVayAqu/HWyaTCfn5+RWmUdkkx2VJ8RqNxmrH461r165hwoQJuHnzJt5//30uc68gluFE1cO8Q0T+hA05ROSVhQsXyisE1SWHw4HCwkJkZmZWWIny1TEA5GWppVWN6pOmTZviww8/hEqlwsMPPyz/8K9ter1eXhq8VatWLnPBSMOpSkpKXCYAttlsLo1Nzo0qzsdUtE1iMpnk89etWwe1Wi1f09150rCgtLQ0tGzZEiqVCmFhYZg4caJH91PV8uNSLEVFRUhKSsK5c+eQlZXlshy5pLJJjjUaDdLT0+U4HA4H9Ho9dDqdPFFzbfv1118xadIkfP3119i+fTvCwsLq5Lrknj+X4cDtz7Pz3FNJSUkuaTjvc35JeRi4nU9TUlLc7iOqLn/OO9X5zGdmZsrzu9XkGCJSiCAinxg2bJiYPXu20mHUCQCirosPnU4ndDpdpdf21TFGo1Go1Wpht9uF3W4XWq1WGAwGn9zHhQsXBACRl5fnk/Rq05kzZ0R4eLiIjo4WP//8c61eC4CwWq1Cr9cLAEKv17vst1gsAoDQ6XTCarUKnU4ntFqtKC4ulv9fOv8/9XabxWIRarVaABAGg0HY7XaX2MqeJ4QQxcXF8mdJisXT+5E+h5W9H1IsFoul0vdOek/cycrKcoldr9eLgoKCStPzpRs3boj4+HjRokULsXv37jq7bnW89dZbIjQ0VOkw6oS/luFCCGEwGFw+s1lZWfK+goKCcvlRekl5wGq1unzGjUaj2zxYE/Xp+94T9en98Me8U53PvPSdWtm9eHKMNwKp/kMUCNiQQ+QjrMj4z7VrcozUMOBcKZIqM1X9oPZEoFVkvv76axEWFiaGDBkiLl26pHQ4PqfkZznYXbt2TTz22GOiefPmIj8/X+lwqsSGHP+4tnPDTVlGo9GlwVQIITfsStw1VPr6fuvT970n6tP74Y95x9vPvN1ur7JR1ZNjvBVo9R8if8ehVUR+Kj09HSqVCpmZmbDZbC5dWx0Oh9zdVaVSISUlxWX4iPMYbrPZLHdPl4ZXmEymcttsNps8nAj4rTttUlKSy1CWikjzg6hUKmg0GuTm5np8P54MM6kre/fuBQC0a9dO3ta2bVsAwP79+xWJSUndu3dHXl4ezp49iwceeADnzp1TOiQKAFevXsX48eOxa9cubNmyBcOHD1c6pDrHMtx7JSUl0Gg0SElJQWFhYbn9sbGx5ebJKjvRd3R0tMt+abilTqercXxUN5h3vOPtZ37VqlWYM2dOpWl6cgwRKUzpliSiYOHLJ1J6vV5+6uj8VESi1WrlruRSDxKtViuEEPLwEDj1IJG6o2u1WvnJTdnzpHPg1BtFGlYEQBw/fly+Pso8obFarUKtVguj0SiEECInJ8fl+lXdT1XDTJyVvbavj5Hu193xarXaoxgrE6hPpIqLi0VERITo3LmzOHz4sNLh+ITVai03LINq7syZM6Jv374iNDRUHDx4UOlwPObLHjkswytWWflcdiigWq2uMm9K9++O8/BH5/uvqfrUA8UTrP8on3ckVX3mc3Jy5HusKD1PjqmOQK3/EPkrNuQQ+YgvKzJlf1hKPzgl0vwgzsc773f3xevJNnfHSMOKnMdalz1OGo9dNm2pclLV/XijthtyvN3urUCuyFy8eFEMGzZM3H333WLbtm1Kh1NjzpV3PtfwjezsbHHPPfeI3r17i1OnTikdjld82ZDDMrxiVeU3u90uLBaL/GO0svnJLBaL/AO6rLLzZ3GOnNrD+o9/5J2qPvNWq9UlP7lLz5NjqiuQ6z9E/og1VyIf8WVFRnoKZDQaXSZdLau4uFieTLW2KjKeHOf8FMzdj2NP78cTbMhRVmlpqfjzn/8sVCqVmD9/vrhx44bSIZGfeOedd0Tjxo3FE088Ia5cuaJ0OF7zZUMOy/CKeVOWGgyGSntCVjbRt8TTRiFvsCHHFes//pV3KvrMl/38u0vPk2OqK9DrP0T+hg05RD7iy4rM8ePHXSoH7p4kShXc48ePK16RqeqL3pP78VRtN+RIcbo7vrIu/J4KlopMRkaGuOOOO8RDDz3EYUn13JUrV0R8fLxo1KiReO2115QOp9p82ZDDMrxi3vwwtNvtFR5bdpLjyrh7j2uCDTmuWP/xv7xT9r3JysoqN1F42fQ8OaYmgqX+Q+Qv2JBD5CO1UbGzWCzy0xznL3+pK6/0hVsXFRlPujJXNQdBRffjjdpuyJGWvnVunJC6K/viaW4wVWQ+//xz0blzZ9G6desKhzdQcMvNzRVdunQR99xzj8jOzlY6nBqpjVWrWIaX5+0Pw4oa0I1Go1crCbIhp/aw/uOfecf5eOnvil6eHlMTwVT/IfIHXLWKyA+pVCo4HA5ERkZixYoVsFgsSE5OlvcnJCQAQLnVO2qDtGLDuHHjKjzGYDAAANatWyevliCt4gBUfT/+ZMyYMQCA06dPy9vOnz/vso9u69+/P7766iv8z//8D5588kk88sgj+O6775QOi+rAzz//jLlz52LUqFHo0aMHLBYLRo4cqXRYfoNluG84HA5MnDjR7b78/HxERkZ6nA4AGI1Gn8VGtYN5xzfKfubF7Yf3Li+J9LcnxxCR/2BDDpGf0uv18tKYrVq1gl6vl/ep1WoAt5dqdV4a02azyctwAnCpVDgfU9E2iclkks9ft24d1Gq1fE13540fPx4AkJaWhpYtW0KlUiEsLMylAl7Z/Xi6/KZ0P2X/9uUxHTt2hMFgwJo1a+BwOOBwOLBmzRoYDIY6qTgGmubNm2PJkiXIz8/HyZMn0bNnTxgMBlb6gtiuXbvQq1cvrFu3DitWrMDWrVvRvn17pcPyOyzDy6us7DWZTC7LNpeUlGD37t2IjY0tl05RURFiYmLcXkOj0SA9PV2O1eFwQK/XQ6fTIT4+vsoYSXnMO+VVlnf4mSeqpxTpB0QUhGpj1QZpIr+y3XCllRSkiR6lVRzKrliACrrLVrXNYrHIY7oNBoPLBH3uzhPCdclLKRZP78eT5TfLXrfs9X15jBC/LYOrVqtFTk5OpbF5I5i7Fl+9elU8++yzokGDBmL06NFBs0w53VZcXCyeeuopAUDEx8cLm82mdEg+VRurVrEMd31PKit7nZce1+l0lQ6bqmyS47JLmOv1enkpZV/h0CpXrP/4T97x5jNfUf3H22M8Fcz1HyIlqITgY1MiXxg+fDh69+6N5cuXKx1KtalUKgDsQlubrFYr2rRpg7y8vAqfKAe6goIC/PWvf4XFYsHUqVPxt7/9DR06dFA6LKqmS5cuYfHixVi+fDnatWuH119/XX4KHUwyMjKQlpYGq9WqdCjVxjK8bgTD970vBcP7wbxT++pD/YeoLnFoFRER+dTgwYNx4MABmEwm7NmzB127dsXMmTNx8eJFpUMjL1y/fh1vvvkmunbtitWrV2PRokU4evRoUDbiEBEREQUSNuQQEYDKx4wTeUulUmHixIk4cuQIXnvtNXzwwQfo1q0bli5dWuHcReQfrl27BoPBgD/84Q946aWX8Ne//hXffvst5s+fjzvuuEPp8KgCLMOJqod5h4gCERtyiAgAEBYW5vZvoppo0qQJnnnmGXzzzTeYPXs2Xn31VXTo0AHz5s3Dt99+q3R45MRms2HRokUIDw/HM888A7VajW+++QZ/+9vfcNdddykdHlWBZThR9TDvEFEgYkMOEQEov+wkkS/dfffdSE1NxXfffYfU1FR88MEH6Nq1K9RqNbKzs5UOr147efIk5s6di86dO+Ptt9/G008/jVOnTuEf//gH2rRpo3R45CGW4UTVw7xDRIGIDTlERFRn7rrrLsydOxfffPMNNmzYAJvNhtGjR2Pw4MFYvXo1fv75Z6VDrBeuXbuGzZs3Y+zYsejWrRs++ugjLFu2DCUlJViyZAmXEyciIiLyY0W2sbgAACAASURBVGzIISKiOteoUSM88cQT2LdvH/bs2YOOHTsiKSkJbdq0QUJCArZt24YbN24oHWZQEUJg9+7dmDFjBtq2bYv4+HioVCqYzWYcPXoUiYmJaNasmdJhEhEREVEVGikdABER1W9Dhw7F0KFDYbfbsWnTJqxbtw5qtRqtW7dGfHw8nnrqKQwYMEDpMAPW119/jffeew/r169HcXExoqKikJKSgoSEBLRt21bp8IiIiIjIS2zIISIiv9CyZUvMmDEDM2bMwNmzZ/Gf//wHq1evxltvvYXQ0FCMGTMGarUao0ePRsuWLZUO129du3YNe/bsQXZ2NrKzs3Hw4EG0b98ejz32GKZNm4a+ffsqHSIRERER1QAbcoh8aNu2bSguLq7yuJs3b6Jhw4Z1EBH5m+vXrysdQkDo0KED5s6di7lz5+KLL77A9u3bsW3bNiQkJKBBgwZ44IEHEBcXh7i4OPTs2VPpcBV3+vRp7NixA9u3b0deXh5++eUXREZGIi4uDsuWLcPQoUPRoAFHU1fFbrdDrVZ7fZ40QapKpfJ1SOSHjhw5gt69eysdhl/xtP5DwcmTei3rP0S+pRKcnp3IJ1555RUcO3as0mN+/fVXFBYW4ubNm3jwwQfrJjAFnT9/HiUlJYiOjlY6FL/z0ksvoXv37kqHEXAuXbqEnJwcZGdnY+vWrTh//jxCQkIwYMAADB06FP369cPQoUPxu9/9TulQa82vv/6KL7/8Env27MHBgwdx8OBBHD16FM2bN8eQIUPwyCOP4NFHH0WHDh2UDjWgfPzxx1i7dq3X5127dg27d+9GWFgYIiMjayEy//Ltt9/C4XAgKipK6VAUNXr0aEydOlXpMPyCJ/Wf+qY+1X8uXryI3bt3Y/DgwWjXrl2Vx7P+Q+QbbMghqiNHjx7FxIkTcfHiRaxfvx4PPfSQ0iHVuoyMDLzyyiu4cOGC0qFQELp16xYOHDiAPXv2oLCwEIWFhTh79iwaNGiA+++/H4MGDcKgQYPQq1cvREREBGTjzuXLl3HixAkcOXIE+/fvx759+1BUVIRff/0VrVu3lu/xgQcewJAhQ9CkSROlQ65XDh8+jPHjx+PGjRv48MMP68WwtZdffhlGo5E/3IkqUZ/qP//973/xzDPP4J///CfmzJkDvV6Pxo0bKx0WUdDj0CqiOrB27VokJSWhf//++Pjjjz16YkFElWvQoAEGDhyIgQMHytvOnz+Pffv2yQ07JpMJV69eBQC0bt0a3bt3R7du3dCtWzf573vvvVfR1ZpKS0vx/fff48SJEzh27BiOHTuGEydO4Pjx4/juu+8AAE2aNEGfPn0wZMgQzJs3D9HR0ejSpYtiMROwdetWPPnkk+jduzfef/99hIaGKh1SnWjXrh3Onz+vdBhE5CeaNWuGzMxMDBs2DElJSbBYLDAajazrEtUyNuQQ1SI+pSCqW+3atcOf/vQn/OlPfwJwu9dOSUkJjh8/juPHj8uNJDt37pQbSQDgzjvvRLt27RAaGorQ0FC0a9cOrVu3Rps2bdCqVSs0atQId911F1QqlTzR8t13342GDRvijjvuQGlpKQDgp59+AgD8/PPPuHnzJn755ReUlpbiypUrOH/+PGw2G6xWK77//ntcvHgRFy5ckM8BXBubxo4dKzc4de7cGY0a8SvbHwgh8Pe//x0vvvgipk+fjuXLl9ercr1du3a4fPkyrly5gjvvvFPpcIjIT0ydOhX9+/fHxIkTERUVVW96nxMphbVColpy7NgxTJo0CefPn8e2bdsQFxendEhE9U6DBg3QqVMndOrUCWPGjHHZd/nyZZw8eVJuYDl//jwuXrwIq9WKL7/80m1DS3WVbSjq1asXQkND0aZNG/l13333oVWrVjW+FtWeq1evYtq0adiyZQuWLVuGZ555RumQ6pz0lP38+fOIiIhQOBoi8if3338/CgsLkZiYiLFjx+L555/HK6+8wgU+iGoBG3KIasH69euh1Wpx//3348CBA+jUqZPSIRFRGXfddRf69u3r0bwmN2/exM8//wwhBOx2OwDA4XDg1q1b+OWXX9C8eXMAkBtipN46LVq04Lw1QeK7777DhAkTcObMGezatQuxsbFKh6QINuQQUWXuuusuGI1GxMXFQavVYv/+/diwYQPatGmjdGhEQYUNOUQ+dO3aNcyfPx8ZGRmYM2cOXnvtNf6IIwoCDRs2lBtpAnHSZKqZzz77DI899hhCQ0Px+eefo3PnzkqHpJjWrVujcePGnCeHiCo1depU9OzZE5MmTUJkZCQ2bNiAkSNHKh0WUdBooHQARMHixIkTiI6OxrvvvouNGzfizTffZCMOEVGAy8zMRGxsLAYMGIA9e/bU60Yc4PZwxbCwMDbkEFGV+vbtiy+++AIxMTEYO3YsFi1ahFu3bikdFlFQYEMOkQ988MEHGDRoEBo1aoRDhw7h8ccfVzokIiKqgRs3bmDBggWYOXMm5s2bhy1btuDuu+9WOiy/0K5dO3z//fdKh0FEAeDuu+/Gv//9b+j1eixevBgajQY//vij0mERBTw25BDVQGlpKebOnYtHH30UkyZNwt69e7kkMBFRgLt06RLGjh2L5cuXY9OmTViyZAkaNGCVScKGHCLyhkqlwty5c7Fnzx4cPXoUffr0wd69e5UOiyigsVZCVE0lJSV48MEHsXr1aphMJrzzzjscSkVEFOAOHz6MAQMG4Pjx48jPz8djjz2mdEh+p127dhxaRUReGzBgAD7//HP06tULMTExWLp0KYQQSodFFJDYkENUDVlZWYiKisK1a9fwxRdf4IknnlA6JCIiqqHt27dj2LBhaNu2LQ4cOIB+/fopHZJfatu2LRtyiKhafv/732Pr1q3Q6/VISUnBhAkT8NNPPykdFlHAYUMOkRekORMmTJiARx55BJ999hm6du2qdFhERFQDQggsXboUarUakyZNQm5uLsLCwpQOy2+1a9cO586dUzoMIgpQ0lCr7OxsHDhwAFFRUdi3b5/SYREFFDbkEHno7NmziImJwVtvvYXMzEysXbsWzZs3VzosIiKqgWvXrmHq1KnQ6XR49dVXYTAYOEy2Cm3btsUvv/wCh8OhdChEFMCGDx8Oi8WCHj16ICYmBm+++abSIREFDDbkEHlg27Zt6NOnDy5duoT9+/fjL3/5i9IhERFRDZ07dw7Dhg3D9u3bsXPnTsyfP1/pkAJCu3btAIATHhNRjbVu3Ro7duzAyy+/jOeeew6PPvoo7Ha70mER+T025BBV4saNG1i0aBE0Gg3GjRuHAwcOoGfPnkqHRURENbR37170798fpaWl+PzzzzFy5EilQwoYUkMO58khIl9QqVSYP38+Pv74YxQWFmLQoEEoKipSOiwiv8aGHKIK2Gw2jB07FkuXLsXrr7+OtWvXokWLFkqHRURENbRq1SqMGDECffv2xaeffoouXbooHVJAueeee3DHHXewIYeIfGrEiBGwWCzo2LEjoqOjOdSKqBL/X3t3H99kfe9//J0WCpSblhtLW+/rDTA9IEgLAhsgMAEJyhw6UdBNYXFnjm36E7eVUQ87Do/VY9nNmZUpZdNwYAoWJjood4ItUCE4ZRQFqcpNCkoA6R0tn98fPVdI09xcSZN8r+R6Px8PHg9t0+ST9pWrV7+5coULOUQ+bNq0CYMGDcLnn3+OHTt2YO7cuapHIiKidmpubsaTTz6JOXPm4Gc/+xnWrFmDtLQ01WPFHYvFgszMTC7kEFHEZWRkuF/q+vOf/xwzZ87EuXPnVI9FZDhcyCHyoL1zyYQJEzBixAjs3LkTAwcOVD0WERG101dffYWJEyeiqKgIy5Ytw6JFi5CUxN2gcGVlZfEcOUQUFcnJySgoKMA//vEPrF+/HkOHDsWHH36oeiwiQ+EeDNH/OXHiBCZOnIgFCxbgueeew+uvv85naomIEsCBAwcwYsQI/Otf/8K7776L+++/X/VIcS87O5tH5BBRVI0bNw6VlZXo3bs38vLysGTJEtUjERkGF3KIAGzduhU33XQTqqqqsGXLFr6UiogoQaxbtw55eXno3bs3KisrMXToUNUjJQQu5BBRLFx22WXYvHkznnjiCcyZMwezZs1CbW2t6rGIlONCDpmaiKCoqAjjx49Hbm4u9uzZg2HDhqkei4iIIqCoqAhTpkzB9OnTsWnTJmRmZqoeKWFkZWVxIYeIYqJDhw4oKCjA6tWrsXbtWuTm5mLfvn2qxyJSigs5ZFpffvklbr/9djz++ONYuHAhVq1ahZ49e6oei4iI2qm+vh4PPPAAHnvsMTz99NN46aWXkJKSonqshJKdnc1z5BBRTE2dOhUOhwM9evTA8OHDsXz5ctUjESnTQfUARCrs2rULd999N5qbm7FlyxaMGDFC9UhERBQBR44cwbRp03Dw4EG8/fbbGD9+vOqRElJ2djbq6upw6tQpPglCRDFzxRVXuF9qde+992LTpk1YvHgxOnXqpHo0opjiETlkKtpLqUaNGoUbb7wRDoeDizhERAmivLwcQ4cOxenTp/Hee+9xESeKsrOzAYAvryKimOvUqROKiorwxhtvYMWKFRgxYgQOHjyoeiyimOJCDpnGmTNncPfdd+Pxxx/HL37xC7z55pvo1auX6rGIiCgCXnvtNYwbNw6DBw/Gzp070a9fP9UjJbSsrCwAXMghInWmTZuGnTt3orm5GUOGDMHKlStVj0QUM1zIIVPYvXs3Bg8ejPLycmzatAkFBQVISmL+RETxrrm5GU8++STuu+8+zJ49G2vXrkVaWprqsRJer1690LlzZy7kEJFS1113HSoqKvDggw/innvuwdy5c9HY2Kh6LKKo41+ylPCKi4sxYsQIXHnllaisrMSoUaNUj0RERBFw5swZ3HnnnSgqKkJJSQmKioq4SB8jFosFWVlZPOExESnXuXNnFBUVYdmyZXj55ZcxatQofPrpp6rHIooq7u1Qwjp79iy+973v4ZFHHsFPf/pTbNiwgW89S0SUID7++GMMGzYMe/bswdatWzFr1izVI5kO37mKiIzk/vvvR2VlJerr65Gbm4u33npL9UhEUcOFHEpIDocDQ4YMwcaNG7Fu3TosWrSIz9ISESWIt99+G3l5eejZsycqKyuRm5ureiRTys7O5kuriMhQ+vXrhx07dmDatGmYMmUK5s6di/Pnz6seiyji+JctJZxly5Zh5MiRuPTSS+FwOPDtb39b9UhERBQhRUVFmDJlCiZPnoyysjIeaakQF3KIyIi6dOmCl156CUuXLsWSJUswfvx4bqso4XAhhxJGXV0dHn74YTz44IN4+OGHsX79evfboxIRUXxraGjAgw8+iMceewz/+Z//iVdffRVdunRRPZapZWVl8Y8jIjKsWbNmobKyEidPnsRNN92Ed955R/VIRBHDhRxKCPv370deXh7efPNN/P3vf0dRURE6duyoeiwiIoqAo0eP4lvf+hZWrVqF1atXY968eapHIlxcyBER1aMQEfk0YMAAVFRUYNy4cZg0aRKefPJJNDc3qx6LqN24kENx7y9/+QuGDh2KLl26YNeuXZg0aZLqkYiIKEJ2796N4cOH49SpU9ixYwemTJmieiT6P9nZ2WhsbMRXX32lehQiIr+6d+8Ou92OpUuXYvHixZgwYQKOHz+ueiyiduFCDsWt+vp6zJ07Fw888AAeeughbN++HVdddZXqsYiIKELsdjtGjRqFG264ATt37kT//v1Vj0QetJcv8+VVRBQPZs2ahW3btuGzzz7DoEGDsGHDBtUjEYWNCzkUlw4cOIBhw4ahpKQEK1eu5EupiIgSSHNzM5588knMmDEDs2fPxtq1a5Genq56LPLChRwiijdDhgzB7t27MXr0aEycOBEFBQW4cOGC6rGIQsaFHIo7b7zxBvLy8tCxY0fs3r0bd911l+qRiIgoQs6ePYtp06bhhRdewNKlS1FUVITk5GTVY5EP6enp6Nq1KxdyiCiu9OjRAytWrMAf//hH/Pa3v8W3v/1tOJ1O1WMRhYQLORQ3GhoaMHfuXNx1112455578N577yEnJ0f1WEREFCGffPIJhg0bhvfffx9bt27FAw88oHokCiIzM5MLOUQUl+bMmYPt27fj0KFDGDp0KLZv3656JCLduJBDcaG6uhqjR4/GK6+8guXLl+PFF19ESkqK6rGIiChC3nnnHeTm5qJLly6oqKhAXl6e6pFIh+zsbBw7dkz1GEREYRk6dCh27dqFgQMHYsyYMXjmmWf4TnwUF7iQQ4b35ptvYvDgwWhsbMTu3btxzz33qB6JiIgiqLi4GFOmTMGkSZOwbds2XH755apHIp2ys7N5RA4RxbXevXtj7dq1KCwsxPz583HnnXfi1KlTqsciCogLOWRYTU1NePLJJzFt2jRMmTIF27Ztw7XXXqt6LCIiipCGhgZ8//vfx49+9CP85je/wWuvvYYuXbqoHotCwIUcIkoEFosFc+fORVlZGSorK3HTTTehoqJC9VhEfnEhhwzp888/x+jRo/GHP/wBf/3rX7Fs2TKkpqaqHouIiCLk2LFjGDNmDF5//XWsWrUK8+bNUz0ShSErK4svrSKihPHNb34Te/fuxYABAzBmzBgUFRWpHonIJy7kkOGsXbsWN910E7766iuUl5djxowZqkciIqII2rNnD4YPH46TJ09ix44dsFqtqkeiMGnnyOHb9xJRoujTpw/WrVuHp556Co899hi+853vwOVyqR6LqBUu5JBhNDU1oaCgAHfccQduv/12VFZW4sYbb1Q9FhERRdDy5csxatQoDBgwALt27cKAAQNUj0TtkJ2djfPnz+PkyZOqRyEiihiLxYJ58+Zhw4YN7hPw7927V/VYRG5cyCFDOHLkCMaOHYtnnnkG//M//4Nly5aha9euqsciIqIIEREUFBRgxowZuP/++7F27Vqkp6erHovaKSsrCwB4nhwiSkhjxoyBw+HAlVdeieHDh/OlVmQYXMgh5TZt2oShQ4eipqYGO3bswJw5c1SPREREEXT27FlMmzYNixYtwssvv4wXX3wRHTp0UD0WRUB2djYALuQQUeLKyMjA22+/jXnz5uHnP/85Zs6ciXPnzqkei0yOCzmkjIjgmWeewfjx4zFixAjs3LkTAwcOVD0WERFF0CeffILhw4dj586d2LJlCx588EHVI1EE9ejRA926deMJj4kooSUnJ6OgoAD/+Mc/sH79egwdOhT//Oc/VY9FJsaFHFLixIkTmDhxIhYsWIDnn38er7/+OtLS0lSPRUREEfSPf/wDeXl56NSpEyoqKjBs2DDVI1EU8C3Iicgsxo0bh8rKSvTu3RvDhg3DkiVLVI9EJsWFHIq5LVu2YNCgQaiqqsLWrVsxd+5c1SMREVGEFRcXY8qUKbjtttuwbds2XHHFFapHoijR3rmKiMgMLrvsMmzevBlPPPEEfvjDH2LWrFmora1VPRaZDBdyKGZEBEVFRZgwYQLy8vLgcDiQl5eneiwiIoqghoYGPPTQQ7DZbPjlL3+J1157DampqarHoijiETlEZDYdOnRAQUEBVq9ejb///e/Izc3FRx99pHosMhEu5FBMnDx5Erfffjsef/xxLFy4EKtWreK7lRARJZiTJ0/itttuw8qVK7Fq1SoUFBTAYrGoHouiLCsriws5RGRKVqsVe/bsQY8ePXDLLbdg+fLlqkcik+BCDkXdrl27kJubiw8//BBbt27FvHnzuGNPRJRgHA4Hhg4diiNHjqCiogJ33HGH6pEoRriQQ0RmdsUVV2Dr1q340Y9+hHvvvRezZs1CXV2d6rEowVlERFQPYWR1dXWoqanBsWPHUFNTA6fTiePHj7s/5nQ6UVdXh8bGRpw7dw4XLlzA6dOnAQAulwue397U1FR06tQJKSkp6Nq1K5KTk9GjRw9YLBZkZGTgkksuQVZWFvr27YuMjAxkZWUhIyPD/bl4IyJYvHgx/t//+3+47bbbUFJSgl69eqkei0LA/snM2L9+K1aswPe//32MGjUKy5cvR8+ePVWPRO0USv9ffvklvv76ayQnJwMwX/+UeLj9p3CtXr0a3//+95GTk4MVK1bgmmuuUT1SyNh/fOBCDoDm5mZUV1fjwIED2L9/P6qqqnDgwAFUVVXhyJEjrS6blpbmDiwzMxN9+/Z1B5qamuqOU7tsUtLFg57OnTuHxsbGNtFfuHABJ06cQE1NDY4fPw6n04mamhqcP3/e/bVdunRBv379cP311+P6669H//793f+v3Z6RnD59Gg8//DBWr16NX/3qV/j1r3/d6ntBxsH+yczYf/uICJ566in8x3/8B2bPno3f//736Nixo9KZSD/2T2bG/ilaPv74Y9x99904dOgQlixZgunTp6seqQ32H/9Mt5DT1NSEf/7znygvL8eOHTuwe/dufPzxx2hoaAAA9O3bF/3798f111+Pfv364brrrkPfvn2RnZ2NSy65BJ07d47ZrFrcNTU1OHz4sPvBtX//fhw8eBCNjY0AWg5pvvHGGzFs2DD3P5UrmO+//z7uvvtuNDY2wm63Y9SoUcpmodbYP5kZ+4+sr7/+GjNnzsS6devwxz/+ET/4wQ9icrsUHvZPZsb+Kdbq6+sxb948/O53v8Ojjz6KZ599FikpKUpmYf+JKeEXco4dO4aKigp3uJWVlaitrUWPHj2Ql5eH3Nxc9O/f3x1vvJyAt6mpCYcPH3aH/cEHH2DHjh04cOAARATXXHMNhg8f7g57yJAh6NChQ9TnKi4uxqOPPopRo0bhtddeQ9++faN+m+Qf+49t/2Qs7D96/R88eBB33HEHnE4n/va3v2H06NERvX5qP/bP7b+ZsX/2bxSvvvoqbDYb+vfvjxUrVuDqq6+O+m2yf3P0n3ALOc3NzXA4HFizZg3Wrl2L3bt3IykpCf369cPNN9+Mm2++GaNGjcLgwYMT8qU+Z86cwQcffIDt27dj27ZtqKiowMmTJ9G1a1eMHTsWVqsVkyZNwuWXXx7R2z179ixmz56Nv/3tb8jPz+dLqRRh/2r6J2Ng/7Hpf+vWrfjud7+LSy+9FKtXr8aVV14ZoXtA7cH+uf03M/bP/o2sqqoK06dPx9GjR7Fs2TJMnjw5otfP/s3Zf0Is5Bw5cgRvv/021q1bhw0bNuD06dO4/vrrMWnSJEyaNAkjRoxA9+7dVY+phIhg//79WL9+PdatW4ctW7agvr4egwYNcn9/brnllnatVu7Zswd33303zpw5g7/+9a+YMGFCBO8BBcP+/YtF/6QW+/cvGv0XFxfjxz/+MaZNm4ZXXnkFqampUbwHFAz794/b/8TH/v1j/8ZTV1eHuXPnYsmSJXj00UdRWFjYrnPKsX//TNO/xKlTp05JSUmJjB8/XiwWi3Tp0kXGjx8vixYtkn379qkez7Bqa2tl/fr1Mm/ePPnGN74hAKRnz54yZ84ceffdd+XChQshXV9JSYmkpqbK6NGj5ejRo1Gamryx//BEun9Sg/2Hpz39NzQ0yEMPPSQWi0UWLFjAx4pC7D883P4nBvYfHvZvHNrfTt/61rfkyJEjIX0t+w9PovYfVws59fX18vrrr8udd94pKSkp0rVrV7n//vvlrbfekrq6OtXjxaUDBw7I008/7Y76uuuuk6eeekoOHjwY8OvOnj0rM2bMEIvFIvPmzZOmpqYYTWxe7D/ywu2fYo/9R57e/k+cOCFjxoyRbt26yapVqxRNa27sP/K4/Y8f7D/y2L9a+/btkxtuuEH69Okjb7/9dsDLsv/IS5T+42IhZ8+ePTJnzhzp2bOnJCcny4QJE2TZsmVy9uxZ1aMllPfff19++tOfSmZmplgsFhk5cqQsWbKkzUbiX//6l9x4443Sp08fWbdunaJpzYP9x4be/im22H9s+Ot/x44dctVVV8k111wjH374oeoxTYf9xwa3/8bE/mOD/atx5swZuffee/0+Kc7+YyOe+zfsQk5zc7OsWbNGbr31VgEgN9xwgxQWFoZ8CBqF7vz58/LWW2/JjBkzpFOnTpKRkSELFiyQ48ePS0lJiXTt2lXy8vLk8OHDqkdNWOxfnUD9U2ywf3W8++/evbvk5OTI/v37VY9mGuxfHW7/1WP/6rB/NUpKSqRLly4yZswY+eKLL9i/IvHYv+EWcurr66WkpMR9qNPIkSOltLQ0bl+7Fu+cTqcsWrRIsrOzJSUlRXJycuS+++6TxsZG1aMlJPZvLN79z5w5Uz744APVYyUs9m8s7D+22L+xsP/YYv/Gwv5jq6KiQi677DLJyclh/wYQL/0bZiHn7NmzUlBQIH369JHOnTvL7Nmz5aOPPlI9Fv2f2tpaefHFF2XAgAFisVjktttuk4qKCtVjJQz2b2zsP7rYv7Gx/+hi/8bG/qOL/Rsb+48u9m9sRu9f+UJOY2Oj/OEPf5C+fftKenq6FBQUSE1NjeqxyI8LFy7IW2+9Jd/85jfFYrHIXXfdJVVVVarHilvsP76w/8hi//GF/UcW+48v7D+y2H98Yf+Rxf7ji1H7V7qQU1paKtddd52kpKTInDlzxOl0qhyHQrR+/XoZNGiQdOjQQebMmcO3Hw8R+49v7L992H98Y//tw/7jG/tvH/Yf39h/+7D/+Gak/pUs5Gzbtk2GDRsmSUlJcv/99/OkuXGsqalJ/vznP8tll10m3bp1kwULFkhtba3qsQyN/ScO9h869p842H/o2H/iYP+hY/+Jg/2Hjv0nDqP0H9OFnLNnz8qPf/xjSUpKkgkTJsiePXtiefMURbW1tfLb3/5W0tLS5LrrrpOtW7eqHslw2H/iYv/Bsf/Exf6DY/+Ji/0Hx/4TF/sPjv0nLtX9x2whZ/PmzXLttddKenq6vPjii7G6WYqxo0ePyrRp08RiscicOXPkzJkzqkcyBPZvDuzfN/ZvDuzfN/ZvDuzfN/ZvDuzfN/ZvDqr6j/pCjsvlkjlz5ojFYpEpU6bIkSNHon2TZAAr2LegewAAIABJREFUVqyQPn36yJVXXinvvPOO6nGUYf/mxP5bsH9zYv8t2L85sf8W7N+c2H8L9m9Ose4/qgs5GzdulKysLMnMzJTXX389mjdFBuR0OmX69OlisVjkkUcekbq6OtUjxRT7Nzf2z/7NjP2zfzNj/+zfzNg/+zezWPYflYWc5uZmWbhwoSQnJ8tdd90lX375ZTRuhuLEihUrJC0tTQYPHiyffPKJ6nGijv2TJ/bP/s2M/bN/M2P/7N/M2D/7N7NY9B/xhZwzZ87IlClTJCUlRYqKiiJ99RSnPvnkExkyZIikpaXJunXrVI8TNeyffGH/ZGbsn8yM/ZOZsX8ys2j3H9GFnMOHD8u//du/Sd++feW9996L5FVTAqirq5OZM2dKcnKyLF68WPU4Ecf+KRD2T2bG/snM2D+ZGfsnM4tm/8kFBQUFiACHw4GxY8eiW7du2LRpE2688cZIXC0lkA4dOmDatGno2LEjnnjiCbhcLtx2222wWCyqR2s39k/BsH8yM/ZPZsb+yczYP5lZVPuPxGrQu+++K+np6TJu3DhlbzeXn58v+fn5AS/jdDrFbreL1WqN0VT6GXm2aFi+fLmkpKTIrFmz5Pz586rHaRf2335Gni0a2H9ksf/4wv7bh73HN/bfPuw/vrH/9mH/8S3S/bd7IWfjxo2Smpoq3/nOd6S+vr7dA4VLT9g2m00ASITWryIqmrM5HA73dQMQm80W8dsIx7p16yQ1NVW++93vxu3GnP1HRixmczgcUlxcLFar1RDfA/YfOezfN6fTKcXFxe7rttvtEb3+9mD/4WPvvrlcLikvL3dv5/0pLS0Vq9UqVqtVSktLIzqDXuw/fOzfNz39632MRBv7Dx/7901P29XV1e7bt9lsUlZWFtEZ9Ipk/+36Lm7fvl26desm3/ve96Spqaldg8SKUcMWid5snjvyAJTtuPiyefNmSU1Nlfvuu0+am5tVjxMS9h9Z0ZytsLDQvdNeXV0dldsIB/uPLTP173K5xGq1SnFxsYi0LOpYrdagO4CxxP6jy0y9i1z8AyfQdWvPBLtcLnG5XGKz2dyPkVhj/9HF/sO7TKyw/+gyws/YHxX9u1wu99+/LpdL7Ha70r+JI9V/2N/FvXv3Slpamtx5553S2NgY9gCxZrawRcRQCze+rF+/Xjp37iyPPPKI6lF0Y/+RF63ZbDab5Ofni8vlivh1RwL7jx0z9a/tpHh2rx2dqepZKF/Yf/SYqXc9111dXS0ApLy83P0x7THhcDiiMksw7D962L+a2w8F+48eo/yMfVHRv6+/hVV/jyLRf1jTHzlyRC6//HIZM2aMNDQ0hH3jkeLv9XaeK25Wq1WqqqrC+qF5X39paan7Or2f3fe8TQBSXFwsTqezzXXqnc3pdEphYaH7cp474HoOr9N2XPLz81vtvBjNqlWrJCkpSZ599lnVowTF/uOn//z8fMO8lDAQ9h8+9u+br5cQulwu9yHFRsL+9WPvwfm739oMnvfD6XS6Z1eF/evH/oOLp4UcEfYfCvYfXCj32wj7Q+3tP+RHcW1trQwZMkT69esnX331VVg3GmnaDqv3D85qtYrNZnM/I+kZXLjXry2GaAsk3gH4OpRdO4w31Nm0r9fOa1BWVtbqmSM9YWsPQu2f1Wr1+UAzgueff16SkpJk9erVqkfxi/3HT//aM62lpaXulxd6/3IwEvYfHvbvm7/7aqQdeE/sXx/2Hpy/+62dG8HX5VWfeJP968P+g4u3hRwR9q8X+w9O7/3WntgywqtW2tN/yI/i2bNnS3p6unz88cch31g0ef/gtAWMqqoq98e0H1o4Gy9fX+f9MS08z4WS8vJyAVqfZFLvbFrs3rcZ6jkOXC6XOBwO92sHVT7zFMzDDz8s6enpcvDgQdWj+MT+/X/MaP1rK/vaLwLtfAiev6CMhv2Hh/23pbXueTv+7otRsH992Hvo84fz8Vhj//qw/9DnD+cyscb+9WH/oc/vS1lZmc+FJ1XC7T+kn/Crr74qFotF3njjjZBuJBa8f3CBnnmJVti+blML1vPZHr2zea6Mev8Ll+oz1QdTV1cngwcPliFDhih9Fxxf2H989e/rstpROqoPpfSH/YeH/bel7VR5PhOm9V9YWKjrOmKN/evD3kOfP5yPxxr714f9hz5/OJeJNfavD/sPfX5frFaroZ7UDbd/3d+hL774QtLT0+UnP/lJWANGm/cPLtK/sPWErfc223u59tAeaEb28ccfS7du3eRXv/qV6lHc2H/89W/0nXZ/2H/o2L9v2jNOQMuRmN6HKxsR+w+OvQfm77p8nTdKu7xRFvfZf3DsPzA912XU/SD2Hxz7D0zPddntdkO+OiWc/nV/1+644w7JycmRr7/+Oqzhos0IYWs7Cd7noPHeSQg1bO9D49vLKDssgfz+97+XDh06yK5du1SPIiLsPx77154J8D5sElB/PoRg2H9o2L8+hYWFhnr7cX/Yf2DsPfT5RcR9rjTPmbXzPxhpp579B8b+Q58/nMuowv4DY/+hz+9JO9WIUYXav66f8MqVK8Visci7777bruGiyfsHp/3C9n7mMZpha6/x8zxUSzsCxvMEq3pn0y7n+dbJ2lm9w+VyuQx7sldPzc3N8q1vfUtyc3PlwoULSmdh//HZv6+jD7R5PF/Da0TsPzTsPzjtnSiM8nrwQNh/YOw99PlFfL/9uPYSRO93ZFGJ/QfG/kOfP5zLqML+A2P/oc+v8XWdDofDUAc4hNp/0J9wQ0ODXHvttTJr1qyIDBgN2ttHeq4Oar+wPd8yTfvDznvFMJTr1wLzPFmTdpsul8t91m7tY3a7vc1t6Z3N83Y9/2lfE+ws3na7vdUDqrq62hBn59bro48+kg4dOsiyZcuUzcD+47d/7TKe8xj9HFGe2L8+7N8/7UT3NpvNsOfF8Yf9+8beA/Oc1deiZXFxsfu8UdoJ8I10NI6G/fvG/gML1r/ey6jG/n1j/4EFalt7Zyxft2O0v41D6T/oQs7zzz8vqamp8vnnn0dkuGjw/oFoqqur3S+vsNlsrd7eLJS34PZ1/f5u0+l0ulcWgZZn/n1tKPXOVl1d7X63KZvN1upZo2Bhe771eH5+vqHPi+DPnDlz5LLLLlN24jP2H7/9azznKS4uNuyOiy/sPzj2H3ju4uLiuNz2i7B/X9i7/tm959Vo+0ZWq9XQRyiz/7bYv/7ZffWv9zFiBOy/Lfavf3bvebUZfP2L9ClMIkFv/xYREfjR0NCAnJwczJgxA88++6y/ixFFzfHjx3H11VejqKgIc+bMielts39Sjf2TmbF/MjP2T2bG/snM9PafFOhKli5dii+//BI/+9nPIj4gkR6ZmZmYNWsW/uu//gvNzc0xvW32T6qxfzIz9k9mxv7JzNg/mZne/gMekTNw4EDk5ubiz3/+c1SGJNLjwIED6N+/P9asWYPbb789ZrfL/skI2D+ZGfsnM2P/ZGbsn8xMT/9+F3J27dqFvLw8lJeXY/jw4VEdVBWLxaLrcgHWuihGxowZg969e+P111+Pye2x/4vYv3rsP/LYf/xg/+3H3uMX+28/9h+/2H/7sf/4Fax/vy+t+stf/oIBAwYkTMS+SMvJnoP+I/UeeughrFmzBqdOnYrJ7bF/9m8k7D/y2H/8YP/tx97jF/tvP/Yfv9h/+7H/+BWsf78LOWvWrMFdd90VtcGIQjF16lRYLBa8/fbbMbk99k9Gwv7JzNg/mRn7JzNj/2Rmwfr3uZDzz3/+E4cPH8aUKVOiOhyRXmlpaRg1ahT+/ve/R/222D8ZDfsnM2P/ZGbsn8yM/ZOZBevf50LOu+++i7S0NOTm5kZ1uHDV1NRg+fLlmDp1qupRKIbGjx+PrVu3Rv122D8ZEftvwf7Nif23YP/mxP5bsH9zYv8t2L85Berf50LOzp07kZeXh6SkgO9OrsyCBQtw7733Ys2aNapH8en06dOoqKjASy+9FNKDLZyve+mll3yexGrv3r2wWCzuf4888kirz3/22Wd45JFH3J/buHGj7jlVycvLw+eff46jR49G9XbYf/uE27/eJgO1ffr06Vaf8/y3fPnyVtezZs0aTJ06FVOnTjXs99IT+29h9v41e/fudd+G9+8APW1rl7FYLJg6dWqbx4fRsP8Widq/N3/7NqF0q+cxwv5bY//tE639/1D2bTwFexxx/6c19t8+0dz+19TUuD8eqPtgf/96X9bf7wgjCdi/+DBw4EB58sknfX3KMACIn/GVy8/Pl/z8/JBnDPXrHA6H38sWFxe7PwdASktL3Z9zuVzu/3e5XGK329tcxohcLpdYLBZZu3ZtVG+H/bdPOP2H0mSgtsvLy1t9zvOf0+l0X85ut4vVahWXyyUul0tsNpsUFxe3855HF/u/yMz9i4gUFhaK1WqV0tJSqa6ubvU5PW0XFhYKAHE4HCJy8XdJYWFhqHc3Ztj/RYnWvzd/+zahdBvoMcL+/WP/7ROt/X+9+zae/D2OuP/jH/tvn2ht/10ul1itVnenTqdTrFar5Ofnt/n6QH8jeAr0O8JoAvXv87vcvXt3WbJkSdQHaw8jh6wJd0Y9X+dyuQI+WAItyvj6XDx8P0VELrnkElm8eHFUb4P9R0YoM4bSZKC27XZ7mw2y0+lstbGvrq4WAFJeXu7+mPaLQ9uxNyr238LM/dtsNsnPzxeXy9Xmc3rb9nXdAMRqteqaVxX23yLR+vcUaN9Gb7eBHiOhXI/RsP8Widy/v6/Ts2/jyd/jiPs/gbH/yIj09l97cstzm651W1ZW1uo69ByUEOx3hBH567/NsWMnT57E2bNncdVVV3l/yvBOnz6N5cuXuw+neumll1BTU9Pmcs8991yrz3sfThXo8/Pnz8f8+fOjfl+CWbJkCR599FGfn/vss88wdepUzJ8/HxUVFW0+b7VafX6dzWaL6IzRkJOTg08//TRq18/+1fSvt8lgbd9666244oorWn1s48aN+O53v+v+//feew8AkJ2d7f5YVlYWgJbDao2M/ftnhv612164cCHS0tLaXF5v24WFhQDgfgx99tln7us1MvbvXzz37ynQvo2eboM9RvRejxGxf/8SpX9/9OzbePL3OOL+j3/sX33//rp97bXXAKDVNl37Oa1cudL9sWB/IwD6fkcYkb/+2yzknDhxAgDQt2/f6E8VYTNnzsTZs2chInA6nVizZg0efvhhnD592n2Z5557DtOnT4eI4O6778bvfve7VtcR7PNGsHHjRowcORIZGRk+P793714AwG9+8xvccsstmDp1qs8HtEb7/kyePDnyw0ZYRkaGu9FoYP/G6N9fk8Ha9vWY2LJlCwYNGtTq/wG02inSvs6orzvWsH//Er3/vXv34je/+Q0mT57sfp341KlTW51LR2/bjz32GPLz83HLLbegoqIC7733HpxOZ6vHiRGxf/8Sof9g+zbButXzGNFzPUbF/v1LhP4D0bNvown0OOL+j3/s37jbf19taoswf/rTn9wfC/Y3gt7fEUbkt3/vQ3R27dolAOTQoUNRODAocuB12FVZWVmb14pqrym12+2tvs7zMk6ns9X1BPt8e2aMxNc5nc5Wr2X1d1mXyyUOh8N9iFqg17+WlZW5Xy9rdN/73vdk2rRpUbt+9q++f5HATYbStsPhaHX/A83Vnnljhf23MGP/3uf10M5tAI/D5ENtW/v6eDnEmP23SMT+9e7biPjvVs9jRM/1GBX7b5GI/Yf6db72bUSCP464/+Mf+zfu9l/bVldVVQW9nUB/I4T6O8JI/PXf5oic+vp6AECnTp28P2Vo2qFVnit5AwYMAHDxkCyg5VD1vn37Yvny5Th9+jQyMjIgIro/r9qbb76J2bNnB71cWloaBg0ahIULF6K4uDjgSvsLL7yAX/7yl3FxiFlqairq6uqidv3s3xj9B2oylLb/9re/4dZbb43mqDHF/n0zQ/+PP/44ALifgU1LS3O/9KqkpCTk23juuecwevRouFwuAC3P6Hk+e2dE7N+3ROhf775NoG5DeYyw/7bYvzG2/3r427fR+ziKR+zft0ToP1i3DzzwAADgv//7v93bae3oG+2lsppAfyNEej8qlvz2772ys3PnTgEgn376abQWlSICYa4yV1VVidVqdX/c+10Kgn2+PTO29+t8nVlbz224XC6/l7Hb7YY/W72nGTNmyJ133hm162f/6vsPpclAbfs7EaB2/7wBEJvNFtqwMcb+W5ixfz33UW/b3icOrKqqavPMlRGx/xaJ1r/efZtg3er9PrB/39i/+u2/nq/zt2+j53HE/R//2L+xt//akcratlo7EinQSbq9/0bQ+70yIn/9t5n6ww8/FADy0UcfxWSwcPnbOHm/DZ+/jZPD4XAfTuUr1mCfD2fG9n6d9nF//wLx9z3wd8Z7o7rjjjvkvvvui9r1s399nw9nRj3CadLfzofdbve5gdfemtDze6W9k4PRd+TZfwsz9q/N4/0SEODiu+3obdt7Nm1nx+g7Muy/RaL1r3ffJli3eh4jeq7HqNh/i0TrP9Sv87dvo+dxxP0f/9i/vs+HM6Oey4b6t21hYaGuvxc8vwd6f0cYkb/+23yHjh07JgBky5YtMRksXN4/YO0ZFs/XuGm/nD3fmsz7B6i9fZnez7dnxmh8nZ7LulyuNm/P5nQ62zxAtQevkY0cOVIeffTRqF0/+1fXfzhN+mpb4+/rfL39pvZ6Yu9nBYyG/bcwY/++nn3S7qP2Oni9bft6VjYedmTYf4tE7F/P1wfrVs9jRM/1GBX7b5HI/ev5ulD2072vj/s//rF/9f3r/Xq73a7r3K7efyPo/R1hRP76b/NdunDhgqSmpsrSpUtjMlg4tBMwea5AulwusVqtYrVa3R+z2+1tNnhAy4nttA1WdXV1q53nYJ/Pz8/XtQLo+QyPr9D8XU+wr/Pm6wHtGW11dbWUlpa2+hqn09nq8DnPf96XNZrs7Gx5/vnno3b97F9N/3qa1NO2xt+JADXFxcVis9nE5XK5T3Zm9GejRNi/iHn7177O8z4WFxe3+eNTT9vazoz2GNF25P0tihoF+0/M/n3xtSOvp1s9jxH27xv7N/7+f7B9G2++Hkfc//GN/Rt7+6+dxNhms/k8Ukjv3wh6fkcYkb/+25zs2GKx4Morr8TBgwe9P2UYnm8Np/13WloalixZAqvVir59+8JisQAAFi1a1ObrH330UaxcuRIWiwUrV67EY489FtLng7FYLEhPT3f/f3p6unueaHydp65du2LcuHGwWCyYP38+Tp06BavV2uoyCxYs8HuC2H79+oV0e7FUW1uL48eP46qrrorabbB/Nf3raVJP25pgJzmePXs2Jk+ejPT0dMycORPTp083/AkC2X8Ls/YPAAsXLmxzH//yl7+0uryetm+99VaUlZVhy5YtsFgsKCkpQVlZmaFPDM7+WyRi/3rp6VbPY4T9+8b+jb//H4k3cOD+j2/s37jbf+16d+7cCZvN5nMuvX8j6PkdYTSB+reItD0l9YwZM3DmzBmsXbs2FvMR6bJt2zZ885vfxOHDh3HllVdG7XbYPxkR+yczY/9kZuyfzIz9k5kF6r/NETkAkJubi507d8ZkOCK9duzYgYyMjKhuxAH2T8bE/snM2D+ZGfsnM2P/ZGaB+ve5kDNy5EicOHECH374YdSHI9Jr8+bNGDlyZNRvh/2TEbF/MjP2T2bG/snM2D+ZWaD+fS7kDB06FJmZmX5fs08Ua3V1ddi4cSOmTJkS9dti/2Q07J/MjP2TmbF/MjP2T2YWrH+fCzlJSUm4/fbbsWrVqqgOR6TXO++8g/r6ekyePDnqt8X+yWjYP5kZ+yczY/9kZuyfzCxY/z4XcoCWEz7t2rULH3zwQdSGI9Lr5Zdfxrhx45CZmRmT22P/ZCTsn8yM/ZOZsX8yM/ZPZhasf78LOWPHjkVOTg5eeeWVqA1HpMeRI0ewbt06/OAHP4jZbbJ/Mgr2T2bG/snM2D+ZGfsnM9PTv9+FHIvFgh/+8Id4+eWX4XK5ojJgoqipqcHy5csxdepU1aMkpKKiIvTp0wfTpk2L2W2yf/3Yf3Sxf2Nj/9HF/o2N/UcX+zc29h9d7N/Y2H906epfAjh9+rT07NlTFi5cGOhipmez2QSABPl2RoXD4XDfNgCx2WytPu90OiU/P9/9ebvdHvQ6i4uLldwXX1wul6SlpckzzzwT89tm//oYtX+Xy9Xqc57//D0OHA6HFBcXi9VqNcRjgP0bn6r+nU6ne1sdqOnS0lJ3z1arNejvAG7/W7B/fYy6/dcE61/v40gF9m98KvoPZd9Gz/Zfu4zVapXS0tJY3Y2g2L/xGX3773lZPfv28bj/E3Ta+fPnS69eveTLL7+M2HCJSFXInjsgAFpthJ1Op5SXl7v/3263CwApLCz0e32eDwwj+MUvfiG9evWSM2fOKLl99q+PEfsvLy/3u7PjdDrbXFdhYaF7R6a6ujqWd8Mv9h8fVOzIW61WKS4uFpGWbb3VapX8/PxWlyssLBQA4nA4ROTi9t3f7wBu/1tj//oYcfsvErx/vY8jVdh/fIh1/3r3bfRs/+12u1itVnG5XOJyucRms7kfD6qx//hg1O2/Ru++fbzu/wSd9syZM5KZmSlz586N2HCJSNUPP9DquecijibQnC6Xq9XRO6odOnRIOnfuLC+88IKyGdi/Pkbs3263t9loa0eoebPZbJKfny8ulyviM4aL/cePWPevLcp79qrthJSVlQWcS3tm1hu3/22xf32MuP0XCd6/3seRCuw/fqjY/uvZtwnWf3V1tQBo9beC1r+2+KMK+48fRt3+i+jft4/n/R9d0/7pT3+Sjh07yocfftju4dpDW10uLi4Wp9PZ6pvtcrlarc7l5+e7V6adTqd71Vmk5YevHYalbQy1X+ieH3M6ne5DDkUurv7ZbDapqqpqNZuvH77T6XTPbLVa2+wYBLo/+fn5QZ8V0jbC+fn5PhdtvGmHY/q73sLCQvccRgh52rRp0q9fP2lsbFQ6B/uPz/59HXVjt9vb7KDk5+cHPCRTFfbfGvu/yNfhwdr23bNl7Ta0x4f2mPG1k87tv2/s33j9i+jb/wnWv97HkQrsvzX23/q6vfnatwnWv3a/PReFtFlUH5XD/ltj/63p2f6Hsm8fz/s/uqZtamqSYcOGSW5urpw/f77dA4ajsLDQHZjnyplGe52e0+l0/4C1H6D2y9pzA6Ydmmiz2dps5LSv077Gc0OoHXoIoFXM3j987RBd7fWoZWVlrW4/2P3RE7L2gNT+Wa1Wnxt47b5pt+H9INTm0+6jEUJesWKFWCwWWb9+vdI5RNh/IvTv+X3ypD37VFpa6v5F5euXTqyx/9bYf2v+ttG+Pq5dd3l5udjtdp+PEW7//WP/xutfRP/2P1D/oTyOYon9t8b+g/P3B2ug/rX74k17PKnC/ltj/20F2/6Hsm8f7/s/uqfdt2+fdO7cWZ5++umwh2sPLVKNrxU8zw2Z9w/D1w9Hz8d8XcbX60y9L6etcHpftxZnsPujl8vlEofD4X4g+FpF1x6g2j/v8yNoJ/sLdJ9jyel0SkZGhsyePVvZDN7Y/0Xx1r/n3N4n+vN+DbnnLyo9R7lFA/tvi/235mtnyt+8npf3dYgxt//Bsf+LjNC/Ru/231//oT6OYoH9t8X+A/O1b+PJX//+Omf/rbH/i4zUf6Dtv959+0TY/wlp2sLCQunYsaNs27Yt5OHaS/sB2O32gK91q66udv8AoxWynst5roJ6/wvl/oRCOyO3P/6C9975URlyc3OzjB8/Xq6++mpDna9EhP2Hcjkj9u95uKm/uUUu/qJScWg9+/eN/bfm+Yya9vX+TmRcWFjovp38/Hz3iS013P7rw/71X85I2/9A/YfyOIoF9u8b+w/M176NJlD/eu9frLB/39h/cN7bf7379omw/xPStBcuXBCr1SqXXnpp0JcwRFpVVVWrOHz9ktV+kFVVVcpDDhaDnvsTKu213YF4f298ncVbZci//vWvpVOnTlJZWank9gNh//ovZ7T+/Z3k2Gg7MuzfN/bfVllZmfs6iouL2xy+LNL2ZK7a90bbeeH2Xz/2r/9yRtn+B+tfRN/jKFbYv2/s3z9/+zYiwfv391bM3n/sxgr79439B+e9/dcze6Ls/4Q87VdffSU5OTkyYsQIqaurC/XL283hcLhX87zfQg+4eNKuWISs51A2X+ej0XN/wqVn4+s5q/bf/v7F0v/+7/+KxWKRP/3pTzG93VCw/4sfj6f+fZ0IULu8546OBoj9a8TZf3Ds37/CwsKg71qi7exw+x8e9n/x40brX6Tt9j9Y/774ehzFAvsPjv235W/fxtdc3v1r5w7xXJjQTsUQ65Mds//g2H9gnjPp2bdPlP2fsCbdt2+fpKenyz333CMXLlwI5ypC5v0D0Q6R8vx8KP+v92O+LqOteHq+9Zn35bQNpOdrUrWzeOu5P+FwuVxBT9KqbcgDvZ5WRcTbt2+Xzp07x8Xb/LH/+Ovf3wKPr2df9TxGIo39B8b+A9PelcJ7p8XXM66eOzK+cPsfGPs3Xv8ivrf/ofbv73EUbew/MPbvX6Anb4P17+vtx7WXG3ofqRBN7D8w9h+c9/Y/3H37eNz/CXvasrIySUlJkX//93+PScxaFNrGRXstoEbbYFVXV7c6tMzpdLZ6SzHPqDwv4+9j2v9rP3jP15lqgl2X5z/PFdNA9yfYWbvtdnuraKurq1s9sLTvia+zgwd7tinWITscDunVq5dMnTpVmpqaYna77cH+jd+/JtiJALX7o80e7Fw7kcb+g2P/bWkn+rPZbH6fzdJ2ZrT5tZ30QAv+3P4Hx/7jY/uvp389j6NoYv/BsX/fgu3b6Om/uLjYfY4o7YSwsTwah/0Hx/5b07v9D2ffPh73f9o17apVq6RDhw7y05/+tD1Xo4sWiXYiJ+9fuNqKnnbSL+0s3t7v2KQLMUXDAAAKWUlEQVT9gEL9mMPhaPU6al9nfvcOwPMtv7VZ9N6fYCF7vvVafn6+z0Mrvd+erbCwUNc78cQy5P3790tmZqbceuutUltbG5PbjBT2L36/TkR9/57XFew1zdozCL7uXzSxf33Yf9vvhzZLsHN5lJWVuQ8zttlsQY/a5PZfH/Yvfr9OxDjb/0D9h/I4igb2rw/7903Pvo2e7b/2ePL39szRwv71Yf+thbL9D3XfPh73f9o97auvvirJyckyd+7cmB1mFkux/KGaze7duyUjI0NGjRolX3/9tepxwsL+KVzs3/jYf/Swf+Nj/9HD/o2P/UcP+zc+9h89key/A9ppxowZSE1Nxb333ouTJ0/ilVdeQceOHdt7tZTgduzYgcmTJ2PgwIEoLS1F165dVY8UFvZP4WD/ZGbsn8yM/ZOZsX8ys4j3H6HFJSkrK5Pu3bvLhAkT5NSpU5G6WqV8vfaP2u+vf/2rpKamytSpU5Wc+T0a2D/pxf7jA/uPDvYfH9h/dLD/+MD+o4P9xwf2Hx3R6D+ix0zt2rVLsrOzZcCAAfLJJ59E8qqV0CIGDy+LiMbGRvnxj38sFotF5s2bFzcnNtOL/VMg7D++sP/IYv/xhf1HFvuPL+w/sth/fGH/kRXN/iP+0zly5IgMHTpUevToIStXroz01VOcqqmpkbFjx0q3bt1kxYoVqseJGvZPvrB/MjP2T2bG/snM2D+ZWbT7Ty4oKCho34uzWuvevTtmzpyJw4cP41e/+hXOnTuHMWPGIDk5OZI3Q3GkvLwc48aNQ11dHdavX4+xY8eqHilq2D95Y//s38zYP/s3M/bP/s2M/bN/M4tJ/xFfGvKwdOlS6dq1qwwcOFAqKyujeVNkQHV1dfLEE09Ihw4d5Pbbb0+Y147qxf7Njf2zfzNj/+zfzNg/+zcz9s/+zSyW/Uf9hW+HDh2ScePGSYcOHeQnP/mJnDt3Lto3SQbw3nvvSf/+/aVHjx7ywgsvSHNzs+qRlGD/5sT+W7B/c2L/Ldi/ObH/FuzfnNh/C/ZvTrHuPyZnMLpw4YKUlJRIz5495ZprrpFNmzbF4mZJgdraWpk3b54kJyfLxIkTpbq6WvVIyrF/82D/bbF/82D/bbF/82D/bbF/82D/bbF/81DVf0xPRf3ZZ5/JxIkTJSkpSWw2mxw7diyWN09RdOHCBVm5cqXk5ORIz549ZenSpapHMhz2n7jYf3DsP3Gx/+DYf+Ji/8Gx/8TF/oNj/4lLdf9K3lNs2bJlctlll0m3bt3k17/+tZw5c0bFGBQhmzZtkry8PElKSpL77rtPjh49qnokQ2P/iYX9h4b9Jxb2Hxr2n1jYf2jYf2Jh/6Fh/4nFCP0re3P4hoYGeeGFFyQ9PV169+4tixYtkvr6elXjUBj27dsn06dPFwAyfvx4ef/991WPFDfYf/xj/+Fj//GP/YeP/cc/9h8+9h//2H/42H/8M1L/yhZyNCdPnpSf//zn0qlTJ7n22mulpKREGhoaVI9FAXz00Ucya9YsSUpKkptvvlnWr1+veqS4xf7jD/uPHPYff9h/5LD/+MP+I4f9xx/2HznsP/4YsX/lCzmaw4cPywMPPCAdO3aU7Oxsefrpp+XLL79UPRZ5WL9+vUyaNEksFov069dP7Ha7XLhwQfVYCYH9Gx/7jx72b3zsP3rYv/Gx/+hh/8bH/qOH/Rufkfs3zEKO5tixY7JgwQLp1auXdOrUSWbOnCn79u1TPZZpNTQ0yIoVKyQ3N1cAyMiRI2XFihXS1NSkerSExP6Nhf3HFvs3FvYfW+zfWNh/bLF/Y2H/scX+jSVe+jfcQo7m7NmzsnjxYrnmmmskKSlJrFarvPHGG3wdYYzs379f8vPzJTMzU1JSUmTmzJmye/du1WOZBvtXi/2rxf7VYv9qsX+12L9a7F8t9q8W+1cr3vo37EKOprm5Wd544w0ZP368JCUlSc+ePeWHP/yhvPvuu4Y5rClR1NTUyOLFiyUvL08AyKWXXiq//OUv5ciRI6pHMy32Hzvs33jYf+ywf+Nh/7HD/o2H/ccO+zce9h878dy/4RdyPH3xxRfy7LPPysCBAwWA5OTkyPz58+Vf//qX6tHi1tdffy3Lly+XKVOmSMeOHaV79+7ywAMPyIYNG6S5uVn1eOSB/Uce+48f7D/y2H/8YP+Rx/7jB/uPPPYfP9h/5CVK/xYREcShjz76CCtXrkRJSQkOHz6Mq6++GhMmTMD48eMxceJEdO/eXfWIhnXo0CFs2LABa9aswYYNG3D+/HmMHTsWM2fOxHe+8x1069ZN9YgUBPsPH/uPf+w/fOw//rH/8LH/+Mf+w8f+4x/7D18i9h+3Czma5uZmlJeX46233sK6deuwd+9edO7cGWPGjMHkyZMxadIkXHPNNarHVKq2thabN292f48OHTqEnj174tvf/jYmTZqEyZMn45JLLlE9JoWB/QfH/hMX+w+O/Scu9h8c+09c7D849p+42H9wZug/7hdyvJ04cQKbN2/GmjVrsGbNGrhcLmRmZmLo0KG4+eabcfPNN2PUqFHo2bOn6lGj5ujRo9i+fTu2bduG999/H5WVlWhoaMA3vvENWK1WjB8/HqNHj0bHjh1Vj0oRxv7Zv5mxf/ZvZuyf/ZsZ+2f/Zsb+zdl/wi3keGpqakJFRQXKy8tRUVGBHTt24MiRI0hOTsYNN9yA4cOHIzc3F/3790f//v3Rp08f1SOHpLGxEQcPHsT+/fvxwQcfYMeOHdixYwe++uordOrUCYMHD8awYcMwbNgwjB49GtnZ2apHphhi/+zfzNg/+zcz9s/+zYz9s38zY//m6T+hF3J8+eKLL1BRUeEOe8+ePTh37hwAoHfv3ujXrx/69++P66+/Hv369cP111+PzMxM9OrVS8m8DQ0NqKmpweHDh1FVVeX+t3//fnz66adoamqCxWJBTk6OO9phw4Zh8ODBSElJUTIzGRf7JzNj/2Rm7J/MjP2TmbH/xGS6hRxvIoLPP/8cBw4ccAdSVVWFAwcO4LPPPoP27UlJSUFGRgYyMzPRt29fZGRkIDs7GxkZGejcuTO6deuGjh07okuXLujcuTM6duyIbt26ISkpCZ07d0ZtbS2amppw9uxZAMCpU6cAAGfOnEFzczNqampw4sQJHD9+HMePH0dNTQ2OHz8Ol8vlnrVbt27uB1f//v3Rr18/9/+npqbG/ptHcY/9k5mxfzIz9k9mxv7JzNh/YjD9Qk4gtbW1OHToEJxOJ44dO4YTJ07g2LFjcDqdOHHiBI4ePYqamhrU19fj3LlzaGxsDPk20tLSkJSUhEsuuQQZGRno27cvsrKycMkllyAzMxOZmZnIyMjA5ZdfjksvvTQK95LIN/ZPZsb+yczYP5kZ+yczY//xgws5EVZXV4f6+no0Njbi3LlzaG5uRn19Pbp27Yrk5GT06NEDABL6ZFNkXuyfzIz9k5mxfzIz9k9mxv7V4EIOEREREREREVGcSFI9ABERERERERER6cOFHCIiIiIiIiKiONEBwCHVQxARERERERERUXD/HzvV0+OU+3a3AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.plot_model()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{0: {'children': (1, 2),\n",
      "     'col': 9,\n",
      "     'loss': 1.6825,\n",
      "     'models': (LinearRegression(), LinearRegression()),\n",
      "     'samples': 5000,\n",
      "     'th': -0.1847},\n",
      " 1: {'children': (3, 4),\n",
      "     'col': 2,\n",
      "     'loss': 1.4008,\n",
      "     'models': (LinearRegression(), LinearRegression()),\n",
      "     'samples': 2083,\n",
      "     'th': 0.0939},\n",
      " 2: {'children': (7, 8),\n",
      "     'col': 3,\n",
      "     'loss': 1.517,\n",
      "     'models': (LinearRegression(), LinearRegression()),\n",
      "     'samples': 2917,\n",
      "     'th': 0.09},\n",
      " 3: {'children': (5, 6),\n",
      "     'col': 8,\n",
      "     'loss': 1.2033,\n",
      "     'models': (LinearRegression(), LinearRegression()),\n",
      "     'samples': 1110,\n",
      "     'th': 0.0107},\n",
      " 4: {'loss': 1.4427, 'models': LinearRegression(), 'samples': 973},\n",
      " 5: {'loss': 1.1452, 'models': LinearRegression(), 'samples': 534},\n",
      " 6: {'loss': 1.257, 'models': LinearRegression(), 'samples': 576},\n",
      " 7: {'children': (9, 10),\n",
      "     'col': 1,\n",
      "     'loss': 1.3454,\n",
      "     'models': (LinearRegression(), LinearRegression()),\n",
      "     'samples': 1573,\n",
      "     'th': 0.3457},\n",
      " 8: {'children': (11, 12),\n",
      "     'col': 9,\n",
      "     'loss': 1.3089,\n",
      "     'models': (LinearRegression(), LinearRegression()),\n",
      "     'samples': 1344,\n",
      "     'th': 0.6563},\n",
      " 9: {'loss': 1.2608, 'models': LinearRegression(), 'samples': 984},\n",
      " 10: {'loss': 1.4868, 'models': LinearRegression(), 'samples': 589},\n",
      " 11: {'loss': 1.174, 'models': LinearRegression(), 'samples': 780},\n",
      " 12: {'loss': 1.4956, 'models': LinearRegression(), 'samples': 564}}\n"
     ]
    }
   ],
   "source": [
    "pprint.pprint(regr.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((5000,), (5000,), (5000, 13), 0.9998658336679382)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.predict(X).shape, regr.apply(X).shape, regr.decision_path(X).toarray().shape, regr.score(X,y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### using linear_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearTreeRegressor(base_estimator=LinearRegression(),\n",
       "                    linear_features=[0, 1, 2])"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr = LinearTreeRegressor(\n",
    "    LinearRegression(), criterion='mse',\n",
    "    linear_features=[0,1,2]\n",
    ")\n",
    "regr.fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.32123838, 0.2288477 , 0.40982133, 0.01760775, 0.        ,\n",
       "       0.        , 0.        , 0.02248484, 0.        , 0.        ])"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.feature_importances_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((5000,), (5000,), (5000, 15), 0.9998467954961635)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.predict(X).shape, regr.apply(X).shape, regr.decision_path(X).toarray().shape, regr.score(X,y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### using split_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearTreeRegressor(base_estimator=LinearRegression(),\n",
       "                    split_features=[0, 1, 2, 3, 4])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr = LinearTreeRegressor(\n",
    "    LinearRegression(), criterion='mse',\n",
    "    split_features=[0,1,2,3,4]\n",
    ")\n",
    "regr.fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.67775477, 0.        , 0.10825879, 0.10132923, 0.11265721,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ])"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.feature_importances_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((5000,), (5000,), (5000, 15), 0.999865840718042)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.predict(X).shape, regr.apply(X).shape, regr.decision_path(X).toarray().shape, regr.score(X,y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### using linear_features + split_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearTreeRegressor(base_estimator=LinearRegression(),\n",
       "                    linear_features=[0, 1, 2, 5, 6],\n",
       "                    split_features=[0, 1, 2, 3, 4])"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr = LinearTreeRegressor(\n",
    "    LinearRegression(), criterion='mse',\n",
    "    linear_features=[0,1,2,5,6],\n",
    "    split_features=[0,1,2,3,4]\n",
    ")\n",
    "regr.fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.31934141, 0.24955976, 0.40197897, 0.02911987, 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ])"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.feature_importances_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((5000,), (5000,), (5000, 15), 0.9998477419282291)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.predict(X).shape, regr.apply(X).shape, regr.decision_path(X).toarray().shape, regr.score(X,y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### using linear_features + split_features + categorical_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearTreeRegressor(base_estimator=LinearRegression(),\n",
       "                    categorical_features=[3, 7],\n",
       "                    linear_features=[0, 1, 2, 5, 6], split_features=[0, 1, 3])"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr = LinearTreeRegressor(\n",
    "    LinearRegression(), criterion='mse',\n",
    "    linear_features=[0,1,2,5,6],\n",
    "    split_features=[0,1,3],\n",
    "    categorical_features=[3,7]\n",
    ")\n",
    "regr.fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.29751995, 0.66368927, 0.        , 0.03879078, 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ])"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.feature_importances_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((5000,), (5000,), (5000, 15), 0.9998453110604035)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.predict(X).shape, regr.apply(X).shape, regr.decision_path(X).toarray().shape, regr.score(X,y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### mae criterion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearTreeRegressor(base_estimator=LinearRegression(), criterion='mae')"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr = LinearTreeRegressor(LinearRegression(), criterion='mae')\n",
    "regr.fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.12267756, 0.        , 0.13440127, 0.26154094, 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.14018907, 0.34119116])"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.feature_importances_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((5000,), (5000,), (5000, 15), 0.999870061052249)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.predict(X).shape, regr.apply(X).shape, regr.decision_path(X).toarray().shape, regr.score(X,y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### poisson criterion "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearTreeRegressor(base_estimator=LinearRegression(), criterion='poisson')"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr = LinearTreeRegressor(LinearRegression(), criterion='poisson')\n",
    "regr.fit(X, np.abs(y))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.84497365, 0.        , 0.        , 0.        , 0.        ,\n",
       "       0.02035242, 0.06404556, 0.03695365, 0.        , 0.03367473])"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.feature_importances_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((5000,), (5000,), (5000, 15), -0.8826707898650505)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.predict(X).shape, regr.apply(X).shape, regr.decision_path(X).toarray().shape, regr.score(X,y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### multi-target regression with weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((5000, 10), (5000, 2))"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X, y = make_regression(n_samples=5000, n_features=10,\n",
    "                       n_informative=3, n_targets=2,\n",
    "                       random_state=0, shuffle=False)\n",
    "\n",
    "t = X[:,0] > np.quantile(X[:,0], 0.5)\n",
    "y[np.ix_(t,[0])] += (X[t][:,3]*X[t][:,4]).reshape(-1,1)\n",
    "y[np.ix_(~t,[0])] += (X[~t][:,1]*X[~t][:,2]).reshape(-1,1)\n",
    "\n",
    "t = X[:,5] > np.quantile(X[:,5], 0.7)\n",
    "y[np.ix_(t,[1])] += (X[t][:,6]*X[t][:,7]).reshape(-1,1)\n",
    "y[np.ix_(~t,[1])] += (X[~t][:,8]*X[~t][:,9]).reshape(-1,1)\n",
    "\n",
    "W = np.random.uniform(1,3, (X.shape[0],))\n",
    "\n",
    "X.shape, y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearTreeRegressor(base_estimator=LinearRegression())"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr = LinearTreeRegressor(LinearRegression(), criterion='mse')\n",
    "regr.fit(X, y, W)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.15719169, 0.16621532, 0.13664283, 0.        , 0.15800745,\n",
       "       0.        , 0.        , 0.        , 0.23880011, 0.14314261])"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.feature_importances_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((5000, 2), (5000,), (5000, 13), 0.9999051032996715)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regr.predict(X).shape, regr.apply(X).shape, regr.decision_path(X).toarray().shape, regr.score(X,y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Hyperparameter Tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(estimator=LinearTreeRegressor(base_estimator=Ridge()), n_jobs=-1,\n",
       "             param_grid={'base_estimator__alpha': [1, 5, 10],\n",
       "                         'max_depth': [3, 5, 10]})"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "parameters = {'base_estimator__alpha':[1, 5, 10], 'max_depth':[3, 5, 10]}\n",
    "search = GridSearchCV(LinearTreeRegressor(Ridge()), parameters, n_jobs=-1)\n",
    "search.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'base_estimator__alpha': 1, 'max_depth': 3}"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "search.best_params_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((8000, 15), (8000,))"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X, y = make_classification(n_samples=8000, n_features=15, n_classes=3, \n",
    "                           n_redundant=4, n_informative=5,\n",
    "                           n_clusters_per_class=1,\n",
    "                           shuffle=True, random_state=33)\n",
    "\n",
    "X.shape, y.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### default configuration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearTreeClassifier(base_estimator=RidgeClassifier())"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf = LinearTreeClassifier(RidgeClassifier(), criterion='hamming')\n",
    "clf.fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.09379374, 0.24276051, 0.        , 0.42029795, 0.        ,\n",
       "       0.        , 0.        , 0.11931671, 0.12383108, 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ])"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.feature_importances_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.plot_model()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{0: {'children': (1, 2),\n",
      "     'col': 3,\n",
      "     'loss': 0.1401,\n",
      "     'models': (RidgeClassifier(), RidgeClassifier()),\n",
      "     'samples': 8000,\n",
      "     'th': 1.4162},\n",
      " 1: {'children': (3, 4),\n",
      "     'col': 8,\n",
      "     'loss': 0.1125,\n",
      "     'models': (RidgeClassifier(), RidgeClassifier()),\n",
      "     'samples': 5333,\n",
      "     'th': -0.9453},\n",
      " 2: {'children': (9, 10),\n",
      "     'col': 1,\n",
      "     'loss': 0.0817,\n",
      "     'models': (RidgeClassifier(), RidgeClassifier()),\n",
      "     'samples': 2667,\n",
      "     'th': -0.5967},\n",
      " 3: {'children': (5, 6),\n",
      "     'col': 0,\n",
      "     'loss': 0.0562,\n",
      "     'models': (RidgeClassifier(), RidgeClassifier()),\n",
      "     'samples': 2439,\n",
      "     'th': 0.3997},\n",
      " 4: {'children': (7, 8),\n",
      "     'col': 7,\n",
      "     'loss': 0.114,\n",
      "     'models': (RidgeClassifier(), RidgeClassifier()),\n",
      "     'samples': 2894,\n",
      "     'th': -0.1735},\n",
      " 5: {'classes': array([0, 1, 2]),\n",
      "     'loss': 0.0508,\n",
      "     'models': RidgeClassifier(),\n",
      "     'samples': 1064},\n",
      " 6: {'classes': array([0, 1, 2]),\n",
      "     'loss': 0.0604,\n",
      "     'models': RidgeClassifier(),\n",
      "     'samples': 1375},\n",
      " 7: {'classes': array([0, 1, 2]),\n",
      "     'loss': 0.1588,\n",
      "     'models': RidgeClassifier(),\n",
      "     'samples': 1354},\n",
      " 8: {'classes': array([0, 1, 2]),\n",
      "     'loss': 0.0747,\n",
      "     'models': RidgeClassifier(),\n",
      "     'samples': 1540},\n",
      " 9: {'classes': array([0, 1, 2]),\n",
      "     'loss': 0.0467,\n",
      "     'models': RidgeClassifier(),\n",
      "     'samples': 1370},\n",
      " 10: {'classes': array([0, 1, 2]),\n",
      "      'loss': 0.1187,\n",
      "      'models': RidgeClassifier(),\n",
      "      'samples': 1297}}\n"
     ]
    }
   ],
   "source": [
    "pprint.pprint(clf.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((8000,), (8000, 3), (8000,), (8000, 11), 0.914375)"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.predict(X).shape, clf.predict_proba(X).shape, clf.apply(X).shape, clf.decision_path(X).toarray().shape, clf.score(X,y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### crossentropy criterion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearTreeClassifier(base_estimator=LogisticRegression(),\n",
       "                     criterion='crossentropy')"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf = LinearTreeClassifier(LogisticRegression(), criterion='crossentropy')\n",
    "clf.fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.16432087, 0.14521427, 0.        , 0.17889246, 0.        ,\n",
       "       0.        , 0.03135475, 0.        , 0.48021766, 0.        ,\n",
       "       0.        , 0.        , 0.        , 0.        , 0.        ])"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.feature_importances_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((8000,), (8000, 3), (8000,), (8000, 13), 0.933375)"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.predict(X).shape, clf.predict_proba(X).shape, clf.apply(X).shape, clf.decision_path(X).toarray().shape, clf.score(X,y)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:prova]",
   "language": "python",
   "name": "conda-env-prova-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}