파이썬으로 AB 테스트하기 (1).ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 기본 세팅\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "pd.set_option('display.max_rows', 100)\n",
    "pd.set_option('display.max_columns', 100)\n",
    "\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt  # 파이플롯 사용\n",
    "from IPython.display import set_matplotlib_formats\n",
    "\n",
    "set_matplotlib_formats('retina')  # 한글코드를 더 선명하게 해주는 조치, 레티나 설정\n",
    "matplotlib.rc('font', family='AppleGothic') # 폰트 설정\n",
    "matplotlib.rc('axes', unicode_minus=False) # 마이너스 폰트가 깨지는 경우가 있으므로 조치\n",
    "\n",
    "import seaborn as sns\n",
    "sns.set_style('whitegrid')\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "warnings.filterwarnings('ignore', 'This pattern has match groups')\n",
    "warnings.filterwarnings('ignore', 'The iterable function was deprecated in Matplotlib')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 통계 세팅\n",
    "\n",
    "import scipy as sp\n",
    "from scipy import stats\n",
    "from scipy.stats import chisquare\n",
    "from scipy.stats import f_oneway\n",
    "from scipy.stats import chi2_contingency\n",
    "\n",
    "\n",
    "import statsmodels.api as sm\n",
    "from statsmodels.graphics.factorplots import interaction_plot\n",
    "from statsmodels.formula.api import ols"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 분포가 있는 두 샘플의 비교 t-test : stats.ttest_ind(x1, x2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Ttest_indResult(statistic=0.26833823296238857, pvalue=0.788494433695651)\n",
      "Ttest_indResult(statistic=0.26833823296238857, pvalue=0.7884945274950106)\n"
     ]
    }
   ],
   "source": [
    "np.random.seed(12345678)\n",
    "\n",
    "rvs1 = stats.norm.rvs(loc=5,scale=10,size=500)\n",
    "rvs2 = stats.norm.rvs(loc=5,scale=10,size=500)\n",
    "\n",
    "print(stats.ttest_ind(rvs1,rvs2))\n",
    "\n",
    "print(stats.ttest_ind(rvs1,rvs2, equal_var = False))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 바이너리 분포 t-test 비교하기"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "P1 = 41/195 = 0.21 (that’s 21%)\n",
    "\n",
    "P2 = 351/605 = 0.58 (that’s 58%)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "p_1 = 41/195\n",
    "p_2 = 351/605"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "statistic, p_value = stats.ttest_ind_from_stats(mean1=p_1, \n",
    "                                                std1=np.sqrt(p_1*(1-p_1)), \n",
    "                                                nobs1=195, \n",
    "                                                mean2=p_2, \n",
    "                                                std2=np.sqrt(p_2*(1-p_2)), \n",
    "                                                nobs2=605)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0000000000000000000292903403051350772759983424636849837225482136640275851327888667618992712959880009\n"
     ]
    }
   ],
   "source": [
    "print(format(p_value, '.100f'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-9.4758349718035077557942713610827922821044921875000000000000000000000000000000000000000000000000000000\n"
     ]
    }
   ],
   "source": [
    "print(format(statistic, '.100f'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### AB 테스트에 대해서도 시도"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-1.9443675727573038 0.052\n"
     ]
    }
   ],
   "source": [
    "p_1 = 375/1597 # experiment ctr\n",
    "p_2 = 439 / 1661 # control ctr\n",
    "\n",
    "# 귀무가설 이 두개의 ctr이 같다\n",
    "# 대립가설 : 이 두 개의 ctr은 다르다\n",
    "statistic, p_value = stats.ttest_ind_from_stats(mean1=p_1, \n",
    "                                                std1=np.sqrt(p_1*(1-p_1)), \n",
    "                                                nobs1=1597, \n",
    "                                                mean2=p_2, \n",
    "                                                std2=np.sqrt(p_2*(1-p_2)), \n",
    "                                                nobs2=1661)\n",
    "print(statistic, format(p_value, '.3f'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### ABC 테스트의 검정 - 일원적 카이제곱"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0026177122087117222874741440108437018352560698986053466796875000000000000000000000000000000000000000\n",
      "-1.9443675727573037637085917594959028065204620361328125000000000000000000000000000000000000000000000000\n"
     ]
    }
   ],
   "source": [
    "\n",
    "observed, expected = [56, 42, 33], [55, 55, 55]\n",
    "statistics, p_value = chisquare(observed, f_exp=expected)\n",
    "\n",
    "print(format(p_value, '.100f'))\n",
    "print(format(statistic, '.100f'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA9gAAAL5CAYAAABLrYxWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAAWJQAAFiUBSVIk8AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAgAElEQVR4nOzdeZyWdb3/8fewDC44pmAeAhQFwVyIJUE9pqFyUlPSSjwKIupRIZcWzTVPlh21U+YaLqShKBSm4nZckILAo6aJW4poMoIcLUMJcEEG5veHj5mfyGAJXxgGns+/9Lqu+7o+M/hweM33vq67ora2tjYAAADAKmnW2AMAAADAukBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYA64RXX3013bp1S7du3XLJJZf8w+OPPPLIdOvWLTfddNMamO6TueKKK9KtW7eccsopjT3KavHcc89l6NCh6d27d3r27JmDDz74H75m7733Trdu3XLbbbctt2/p0qWZOXPm6hgVAD4RgQ3AOue6667L9OnTG3sMGjB//vwcffTRefjhh9OsWbNsu+222WqrrVb6fM8880y+/vWvZ9y4cQWnBICV06KxBwCA0hYvXpxzzz03v/71r9Osmd8lr00efvjhzJs3L61bt84DDzyQzTbbbJXOd/PNN+dPf/pT+vbtW2hCAFh5/tYBwDqnoqIiTz/9dG688cbGHoWPePPNN5MkXbp0WeW4BoC1jcAGYJ1z2GGHJUkuu+yyvPrqq408DR+2ZMmSJEllZWUjTwIA5QlsANY5p5xySjp27Jh33nkn3//+9//p1334QWlvv/32cvtnzJhRv//DzjzzzHTr1i0PPvhgnnnmmQwbNix9+vRJz549c/jhh2fq1KlJknfeeSc/+clPsvfee2ennXbK3nvvnUsuuSSLFy9e4Uwvv/xyTjrppOyyyy7157vzzjtXePzChQtz5ZVX5qCDDsrnPve59OrVK//+7/+ecePG1cfth9U9PGzWrFn59re/nR49emSXXXbJ6aef/k99z/7617/moosuype+9KXsvPPO6d27dw4//PCMGzcuNTU19cc9+uij6datW84///wkyR/+8If67+XK/BKk7s/q9ttvT5Jcf/316datW84888xljps7d25+/OMf50tf+lK6d++eXXbZJUcddVTuu+++Bs9b9/2YP39+/ud//icDBw5Mjx490rdv35x88sl55ZVXkiSzZ8/Oqaeeml133TU777xzBgwYkDvuuKPBc06YMCHHHHNM+vTpk5122il77LFHvvGNb9T/dwHAusM92ACsczbYYIP88Ic/zNFHH52pU6dm/Pjx/9STqlfVpEmTMn78+LRs2TKdOnXK7Nmz88QTT+T444/PVVddlYsvvjgvvvhittpqq7Rr1y6zZs3K1Vdfnb/97W/5r//6r+XON3PmzAwcODBvv/12tttuu7zzzjt54okn8sQTT+SRRx7JBRdcsMzxr776ao455pi88soradGiRTp16pSlS5dm2rRpmTZtWh544IGMGDGiwdXj7373u3nmmWfStWvXvP766/nMZz7zD7/eadOmZdiwYZk3b14qKyuz3Xbb5e23366f8Z577smIESOy8cYbZ5NNNkmvXr3yxhtvZPbs2WndunW6du2aJGnVqtUn/l63atUqvXr1yiuvvJK5c+dmyy23TPv27dOpU6f6Y/70pz/luOOOy9y5c1NZWZltttkm77zzTh555JE88sgj+epXv5oLLrggFRUVy53/Zz/7WcaOHZu2bdtm6623zp///Oc88MADmTZtWi655JIMHz48ixYtyjbbbJM33ngjL7zwQv0vJb7yla/Un+faa6/NxRdfnCTp0KFDOnbsmNdeey0TJ07MxIkTc+6552bw4MGf+OsHYO1kBRuAddLuu++eQw45JEly0UUX1d/7uzrdcsst2XvvvTNlypTcfvvtmTRpUnbYYYcsWbIkw4YNy8KFC3Prrbfm/vvvz4QJE3LGGWckSW6//fbMmzdvufPNmDEjm2++ee6+++7ceeedefDBB3P55ZensrIyt956a+655576Y5csWVK/wrr33ntn8uTJueeee3LvvffmnnvuSefOnTNlypT85Cc/aXD25557LjfffHPGjx+fKVOm5D/+4z8+9mudN29efVzvt99+mTJlSm677bbcf//9ueWWW9KuXbs88sgjOe+885IkO+ywQ8aOHZuhQ4cu8+9jx47NFlts8Ym/11tssUXGjh2bPffcM0ny5S9/OWPHjs2wYcOSJAsWLMiJJ56YuXPn5tBDD83DDz9c/z0cM2ZMPv3pT+e2227LDTfc0OD5x44dm9NOOy2///3vc8cdd+SOO+7IRhttlDfeeCNDhgxJ9+7dM3ny5Nx55535/e9/n/79+ydJRo0atcz36PLLL0+rVq1y0003ZeLEibn11lszZcqUfPvb306SXHzxxXn33Xc/8dcPwNpJYAOwzjrzzDPTpk2bvPXWWw2uEJe26aab5oILLkjr1q2TJK1bt87hhx+e5IPPaj7vvPOyww471B8/ZMiQVFZWZsmSJXnxxReXO19FRUWuuOKKdO7cuX7bl770pQwfPjzJBx9HVmfChAl57rnnss022+TSSy9N27Zt6/d16dIll156aZo1a5axY8dm7ty5y13r3/7t39KzZ88kScuWLeu/hhW5+eabM2/evHTt2jUXX3xxPvWpT9Xv6969e0aMGJGKiorcddddeemllz72XKvDuHHj8tprr6VPnz45//zzl/l6evfunR/96EdJPlhhbugt+rvvvnuOO+64NG/ePEnSuXPn7L333kk+uH/8kksuyeabb57kg+9X3S8OXnjhhdTW1iZJqqurs3jx4myzzTbZZZdd6s/dvHnzDBs2LP37989+++3X4C9XAGiaBDYA66xPfepTOfvss5Mkd999d37/+9+v1uv16tVruTCte6t1ixYtsuuuuy6zr0WLFvVh2tA93717917ufu8k+drXvpbkg7dA18XyxIkTkyT77rtvg2+57tq1a7p27ZrFixfnkUceWW5/jx49/uHX92GTJ09OkgwcODAtWix/x9kOO+yQ3r17p7a2NpMmTfpE5y6h7vtxwAEHNPgW8D333DObbrpp5s6dmz/96U8N7v+o9u3bJ0l22mmnbLrppsvsq/uFxpIlS+pXpNu3b5/mzZtn+vTp+elPf5rZs2cv85orr7wyF154Ydq1a7cSXyEAayP3YAOwTjvwwANz1113ZdKkSTnvvPNy1113ZeONN14t19pyyy2X29ayZcskH6xmN3Tvc93+ulXPD/vsZz+7wutssskmWbBgQWbOnJk2bdrkz3/+c5Lkvvvuyx//+McGX/f6668n+eDe7o/6pG/Trq6u/tgZkw8i+/HHH68/dk2q+36MHj16hQ+Fq1u5njlz5nK/YPi4P8u6lesP+/AvGer+LLfYYosceeSRGTVqVEaOHJmRI0emU6dO2WOPPbLXXntlt912qz8nAOsGgQ3AOu+8887LAQcckDlz5uSSSy7J9773vdVynY022miNnW+jjTbKggUL8t577yX54OnhyQdPt/7oSulHLViwYLltn/RBY3Ur7h/3VvK6+RtanV/d6r4fdaH9cRr6fmy44YYrPL6hFfEVOeuss7LjjjtmzJgxefLJJ1NdXZ3q6urcdNNN2WyzzfKd73wnAwcO/KfPB8DaTWADsM5r165dvvOd7+RHP/pRbr755hx00EH/8DUNrSjXxeya8s4776xwX120VlVVJfn/QXjZZZdlv/32W+2zbbTRRpk/f359yDakbl/pXzz8MzbccMMsWLAgv/nNb7Lzzjuv8et/2IABAzJgwIDMnTs3jzzySB566KH87ne/y5tvvplzzz03m2++efbdd99GnRGAMtyDDcB6YdCgQenRo0eWLl2a733vew0+2OrDb/N9//33l9v/17/+dbXO+FEremv17Nmzs3DhwjRr1izbbrttkmTrrbdO8sHnZq/ItGnTMmPGjCK/KNhmm22SfPD08RWpu7e5brY16Z/5fjz66KP585//3OCfdQnvvvtunnvuufpV9DZt2uTLX/5yLrjggkyePDm77757knzs55oD0LQIbADWC82aNcv555+fli1bZsaMGXnyySeXO6ZuNThp+D7l3/72t6t1xo969NFH89prry23fezYsUmSz3/+8/Vv0f7iF7+YJBk/fnwWLVq03Gtmz56dwYMH56CDDsq0adNWeba99toryQcfTVZTU7Pc/meffbb+e/yv//qvq3y9Fal7u/ZH33FQ9/0YN25cg+9GeOyxxzJkyJB8+ctfzv/93/+tltlGjx6dQw45ZLnPK08+eBJ57969k3zwhHkA1g0CG4D1RteuXXPcccclafgt4BtttFH9U7t/9rOf1d+bW1NTkxtvvDG33377mhs2H6yin3TSSfnLX/5Sv+2WW27JqFGjUlFRkRNPPLF++4EHHphOnTrllVdeycknn5w33nijfl91dXW+8Y1vpKamJp/97Gez2267rfJshx9+eDbffPPMmDEjp5122jIfNfX000/n5JNPTm1tbfbbb7/suOOOq3y9Fal7YN1HI/mII47IZpttlscffzxnn332MvdZP/PMM/nOd76TJNlnn33SqVOn1TLb/vvvn+bNm2fq1Kn5xS9+scwvImbMmJFf//rXSRp+YjkATZN7sAFYrwwfPjz333//Ch9+dcopp+Tkk0/O448/nr322iudOnXKa6+9ljfffDNHHXVUbrvttgYfirU6fOELX8hjjz2WffbZJ9ttt13eeuut+hXt008/fZmP/aqsrMzPf/7zHHvssZk8eXK++MUvpkuXLlm8eHGqq6uzZMmS/Mu//EtGjBhRZLbNN988V155ZYYPH5577703EydOzHbbbZe33367/q3tffv2zfnnn1/keitS9wuR+++/PwcddFD69OmTc889N23atMkVV1yRb3zjG7nttttyzz33pEuXLlm4cGFeeeWV+tdeeOGFq222jh075qyzzsqPfvSj/OQnP8m1116bDh065O23384rr7yS2trafPGLX8xXv/rV1TYDAGuWFWwA1iuVlZU5//zzV/gk6H333Tc33HBDvvCFL6RZs2Z5+eWX06FDh/z3f/93/Wdqryk77rhjxo4dm759+6a6ujrz58/P7rvvnl/+8pc59thjlzu+S5cuueOOOzJ8+PBsu+22qa6uzqxZs7LVVlvlmGOOye23317/udwl9O7dO3fddVeGDBmSdu3a5cUXX8y8efOyyy675IILLsioUaOWedv96nDIIYdk6NChadOmTaqrqzN9+vT6fbvssssy87300kt5/fXX07Vr15xyyikZO3bsap/vyCOPzIgRI7LnnnumefPmeeGFF/LWW2/l85//fH70ox/lqquuavBzxAFomipqG3qPHAAAAPCJWMEGAACAAgQ2AAAAFCCwAQAAoACBDQAAAAUIbAAAAChAYAMAAEABAhsAAAAKENgAAABQgMAGAACAAgQ2AAAAFNCisQdoip555pksWbIkrVq1auxRAAAAKGjRokVp3rx5dt5550/8WoG9EpYsWZKlS5c29hiw3pk/f36SpKqqqpEnAYDVz889aByr0noCeyXUrVx/9rOfbeRJYP0yYcKEJEnfvn0beRIAWP383IPG8fzzz6/0a92DDQAAAAUIbAAAAChAYAMAAEABAhsAAAAKENgAAABQgMAGAACAAgQ2AAAAFCCwAQAAoACBDQAAAAUIbAAAAChAYAMAAEABAhsAAAAKENgAAABQgMAGAACAAgQ2AAAAFCCwAQAAoACBDQAAAAUIbAAAAChAYAMAAEABAhsAAAAKENgAAABQgMAGAACAAgQ2AAAAFCCwAQAAoACBDQAAAAUIbAAAAChAYAMAAEABAhsAAAAKENgAAABQgMAGAACAAlo09gDQGDqdeU9jj7DWq77oy409AgAANClWsAEAAKAAgQ0AAAAFCGwAAAAoQGADAABAAQIbAAAAChDYAAAAUIDABgAAgAIENgAAABTQorEHAAAAlte/f//GHgH4hKxgAwAAQAFWsAEA1gOdzrynsUdYq1Vf9OUP/uG8TRt3kLXdeX9Pkux8w86NPMja75mjnmnsEWgEVrABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACWjT2AJ/UpZdemquuuqrBfQcccEAuueSS+n8fP358Ro0alerq6lRVVWX//ffPKaecko033nhNjQsAAMB6oskF9vTp01NZWZnjjz9+uX3bbbdd/T9fc801+dnPfpZu3bpl8ODBmTFjRkaNGpWnnnoqN954YyorK9fk2AAAAKzjmlxgz5gxI126dMnJJ5+8wmPmzJmTyy+/PD179szo0aPTsmXLJMlll12WESNGZNy4cRk8ePCaGhkAAID1QJO6B3vhwoWZM2dOunXr9rHHjRs3LjU1NTnhhBPq4zpJhg0bltatW+eWW25Z3aMCAACwnmlSgT19+vQk+YeB/dhjjyVJ+vTps8z2Vq1apUePHpk+fXoWLFiweoYEAABgvdSk3iL+wgsvJEnefPPNHH300Xn22WeTJLvttlu+9a1vZdttt02SzJo1K23btm3wYWbt27dPksycOTPdu3df6Vnmz5+fCRMmrPTraRz9+/dv7BGanLXpv/P58+cnWbtmAljb+dkHjcvfW5qeqqqqVFVVrdRrm9QKdl1gX3/99WndunUOPfTQdO/ePffff38GDhyY559/Pkkyb968bLLJJg2eo277woUL18zQAAAArBea1Ap28+bN0759+1x44YXp27dv/fY777wz3/3ud3P22Wfn9ttvT01NzQqfEl63fdGiRas0S1VV1TIzwLpqbVr5qPsN8No0EwDAx/H3lqanbuF2ZTSpwP7+97/f4PYBAwZk3Lhxeeyxx/Lyyy9ngw02yOLFixs89v3330+SbLjhhqttTgAAANY/Teot4h9nhx12SJK8+uqrqaqqWuFDzOq2r+gt5AAAALAymkxg19TU5Omnn85TTz3V4P733nsvyQdPCu/UqVPmzp1bv+3D5syZk2bNmmXrrbderfMCAACwfmkygb106dIcccQROe6447JkyZJl9tXW1mbatGlp0aJFPvvZz6Z3795ZunRpHn/88WWOW7RoUZ588sl06dIlrVu3XpPjAwAAsI5rMoFdWVmZfv365e9//3uuvfbaZfZdf/31mTFjRg488MBUVVXlwAMPTPPmzXPllVfW33OdJFdffXUWLlyYww47bE2PDwAAwDquST3k7Iwzzsi0adNy6aWX5g9/+EO23377PPvss/nDH/6QLl265Mwzz0ySdO7cOcccc0xGjhyZgw8+OP369ctLL72USZMmpVevXhk4cGAjfyUAAACsa5rMCnaSdOjQIbfeemu+9rWv5cUXX8zo0aMzZ86cHHPMMfnVr36VzTbbrP7YU089Nf/5n/+ZioqK3HjjjXnxxRczdOjQXHvttSv8CC8AAABYWU1qBTtJttxyy1xwwQX/8LiKiooMGjQogwYNWgNTAQAAsL5rUivYAAAAsLYS2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABbRo7AEA/ln9+/dv7BEAAGCFrGADAABAAVawgY933qaNPcHa7by/N/YEAACsJaxgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAU0OQD+8c//nG6deuWRx99dLl948ePz8EHH5wePXpkzz33zIUXXpi33367EaYEAABgXdekA/vpp5/ODTfc0OC+a665JmeccUaWLl2awYMHZ/vtt8+oUaNy7LHH5v3331/DkwIAALCua9HYA6ys999/P2effXaWLFmy3L45c+bk8ssvT8+ePTN69Oi0bNkySXLZZZdlxIgRGTduXAYPHrymRwYAAGAd1mRXsK+++upUV1dn9913X27fuHHjUlNTkxNOOKE+rpNk2LBhad26dW655ZY1OSoAAADrgSYZ2NOnT8+1116bE044IV26dFlu/2OPPZYk6dOnzzLbW7VqlR49emT69OlZsGDBGpkVAACA9UOTC+wlS5bknHPOydZbb50TTjihwWNmzZqVtm3bZuONN15uX/v27ZMkM2fOXK1zAgAAsH5pcvdgX3fddXnuuecyZsyYVFZWNnjMvHnz0qFDhwb3bbLJJkmShQsXrtIc8+fPz4QJE1bpHKx5/fv3b+wRWEf5/wGwtvKzDxqXvyM0PVVVVamqqlqp1zapFeyZM2fmyiuvzBFHHJGePXuu8LiampoVxnfd9kWLFq2WGQEAAFg/NZkV7Nra2pxzzjlp06ZNvvOd73zssRtssEEWL17c4L66j+jacMMNV2meqqqq9O3bd5XOAaw7rBABAA3xd4Sm5/nnn1/p1zaZwL755pvzxz/+Mddee22D91Z/WFVV1QofYla3ve6t4gAAAFBCkwns+++/P0ly/PHHN7h/yJAhSZKJEyemU6dOeeyxx/Lee+9lgw02WOa4OXPmpFmzZtl6661X78AAAACsV5pMYB9yyCHLfexWkkyZMiVPPfVUDjnkkLRv3z5VVVXp3bt3Hn300Tz++OPZY4896o9dtGhRnnzyyXTp0iWtW7dek+MDAACwjmsygf3Vr361we3z58+vD+y6e6IPPPDAXHPNNbnyyivTp0+f+gebXX311Vm4cGEOO+ywNTY3AAAA64cmE9ifROfOnXPMMcdk5MiROfjgg9OvX7+89NJLmTRpUnr16pWBAwc29ogAAACsY9bJwE6SU089Ne3atcuYMWNy4403ZosttsjQoUNz0kknrfAjvAAAAGBlNfnAPuecc3LOOecst72ioiKDBg3KoEGDGmEqAAAA1jfNGnsAAAAAWBcIbAAAAChAYAMAAEABAhsAAAAKENgAAABQgMAGAACAAgQ2AAAAFCCwAQAAoACBDQAAAAUIbAAAAChAYAMAAEABAhsAAAAKENgAAABQgMAGAACAAgQ2AAAAFCCwAQAAoACBDQAAAAUIbAAAAChAYAMAAEABAhsAAAAKENgAAABQgMAGAACAAgQ2AAAAFCCwAQAAoACBDQAAAAUIbAAAAChAYAMAAEABAhsAAAAKENgAAABQgMAGAACAAgQ2AAAAFCCwAQAAoACBDQAAAAUIbAAAAChAYAMAAEABAhsAAAAKENgAAABQgMAGAACAAgQ2AAAAFCCwAQAAoACBDQAAAAUIbAAAAChAYAMAAEABAhsAAAAKENgAAABQgMAGAACAAgQ2AAAAFCCwAQAAoACBDQAAAAUIbAAAAChAYAMAAEABAhsAAAAKENgAAABQgMAGAACAAgQ2AAAAFCCwAQAAoACBDQAAAAUIbAAAAChAYAMAAEABAhsAAAAKENgAAABQgMAGAACAAgQ2AAAAFCCwAQAAoACBDQAAAAUIbAAAAChAYAMAAEABAhsAAAAKENgAAABQgMAGAACAAlo09gAA64Kdb9i5sUdYqz1z1DONPQIAwGpnBRsAAAAKENgAAABQgMAGAACAAgQ2AAAAFCCwAQAAoACBDQAAAAUIbAAAAChAYAMAAEABAhsAAAAKENgAAABQgMAGAACAAgQ2AAAAFCCwAQAAoACBDQAAAAUIbAAAAChAYAMAAEABqzWw33nnndV5egAAAFhrFA3s2trajB07Noceemh23nnnfP7zn0+S3HTTTTnrrLPyt7/9reTlAAAAYK3RotSJampq8o1vfCNTpkxJixYtsvHGG+fvf/97kuTVV1/N7bffnj/+8Y/51a9+lc0337zUZQEAAGCtUGwF+/rrr8/vf//7DB06NH/4wx8yaNCg+n2nnXZaTj755MyaNSvXXHNNqUsCAADAWqNYYI8fPz69evXKGWeckQ033DAVFRX1+1q0aJETTzwxu+66ayZNmlTqkgAAALDWKBbYs2fPrr/nekV22mmnvP7666UuCQAAAGuNYoFdVVWVOXPmfOwxs2bNyiabbFLqkgAAALDWKJUxAK0AACAASURBVBbYu+22WyZMmJDnn3++wf1PPvlkfvvb32bXXXctdUkAAABYaxR7ivgpp5ySSZMm5fDDD8/Xv/71vPLKK0mS22+/Pc8880x+85vfpLKyMsOHD1+l67z11lv5+c9/nkmTJuWvf/1rOnTokEMOOSRHH310WrRY9ssZP358Ro0alerq6lRVVWX//ffPKaecko033niVZgAAAICPKraCvdVWW+WGG25Ix44dc9NNN2XKlCmpra3N2WefnTFjxuTTn/50Ro4cmc6dO6/0NRYuXJgjjjgio0ePTpcuXTJo0KBssskm+elPf5qTTjoptbW19cdec801OeOMM7J06dIMHjw422+/fUaNGpVjjz0277//fokvGQAAAOoVW8FOkh133DF33XVXnnrqqTz77LNZsGBBNtpoo3Tr1i277LJLmjVbtZ6/9tpr8/LLL+ecc87JkCFD6refeuqpufvuuzN58uR88YtfzJw5c3L55ZenZ8+eGT16dFq2bJkkueyyyzJixIiMGzcugwcPXqVZAAAA4MOKrWB/2Oc+97kMGjQow4YNy5AhQ9K3b99VjuskmTNnTtq1a5cjjjhime0HHHBAkmTatGlJknHjxqWmpiYnnHBCfVwnybBhw9K6devccsstqzwLAAAAfFixFeyJEyf+08fus88+K3WNiy++uMHtL7/8cpKkbdu2SZLHHnssSdKnT59ljmvVqlV69OiRqVOnZsGCBZ5oDgAAQDHFAvvEE09MRUXFP3Xsip40/knU1tbmzTffzH333Zcrrrgin/nMZzJgwIAkH3wcWNu2bRt8mFn79u2TJDNnzkz37t1XeQ4AAABI1kBgv/vuu5k1a1YmT56cz33ucznqqKOKXO+yyy7LVVddleSDlevrrrsum266aZJk3rx56dChQ4Ovq1u1Xrhw4Spdf/78+ZkwYcIqnYM1r3///o09AqzX/H8T1jw/+6Bx+dnX9FRVVaWqqmqlXlsssE8++eSP3f/cc8/liCOOyIIFC4pcr2PHjjnuuONSXV2diRMnZtCgQfnFL36RHXfcMTU1NamsrGzwdXXbFy1aVGQOAAAASAo/Rfzj7LDDDtlvv/1y/fXX55BDDlnl833ta1+r/+ff/e53GT58eM4444zcdddd2WCDDbJ48eIGX1f3EV0bbrjhKl2/qqoqffv2XaVzAKxvrKQBsL7xs6/pWZVbmlfLU8RXZLPNNssrr7xS/Lz9+vXLbrvtlhdffDGzZs1KVVXVClfK67Z7wBkAAAAlrbHAfvPNN3P//fdniy22WKnX19TU5H//93/z0EMPNbj/M5/5TJLkrbfeSqdOnTJ37ty89957yx03Z86cNGvWLFtvvfVKzQEAAAANKfYW8ZNOOqnB7UuXLs27776bp59+Ou+8805OPPHElb7GsGHDsvHGG2fq1Klp3rz5MvumT5+eioqKdOjQIb17986jjz6axx9/PHvssUf9MYsWLcqTTz6ZLl26pHXr1is9BwAAAHxUscB+8MEHP3b/pptumqFDh2b48OErdf4WLVqkf//+ufvuu3Pdddfl+OOPr983ZsyYPPvss+nXr1/atm2bAw88MNdcc02uvPLK9OnTp/7BZldffXUWLlyYww47bKVmAAAAgBUpFtgTJ05scHtFRUVatmyZNm3apFmzVXtH+umnn57HH388F198cR599NF07do1zz//fB5++OF06NAhP/jBD5IknTt3zjHHHJORI0fm4IMPTr9+/fLSSy9l0qRJ6dWrVwYOHLhKcwAAAMBHFQvs9u3blzrVCm255Zb5zW9+k8svvzy/+93v8sgjj+TTn/50jjrqqAwfPjybbbZZ/bGnnnpq2rVrlzFjxuTGG2/MFltskaFDh+akk05a4Ud4AQAAwMpa6cCePn36Sl90++23X+nXbrHFFjn//PP/4XEVFRUZNGhQBg0atNLXAgAAgH/WSgf2wQcfnIqKipV67ap8rhgAAACsjRolsAEAAGBds9KBfdFFF5WcAwAAAJq0VXus90qYPXv2mr4kAAAArHbFniKeJJMnT85dd92VN998M0uWLEltbW2SpLa2NjU1NZk3b16qq6vdgw0AAMA6p1hgP/DAA/nmN79ZH9UN2XDDDbPPPvuUuiQAAACsNYq9RfyXv/xlmjdvnksvvTQPPfRQdthhhwwcODAPPfRQbrjhhuy4446pqKjIaaedVuqSAAAAsNYoFtgzZszIvvvum/322y9t2rRJr1698sc//jFt2rRJ3759c91116WysjJXX311qUsCAADAWqNYYC9atChbb711/b9vu+22qa6uzvvvv58k+dSnPpV99903Tz75ZKlLAgAAwFqjWGC3bds2b775Zv2/b7XVVlm6dGlefPHF+m2bbbZZ/vKXv5S6JAAAAKw1igX2LrvskgceeCAzZ85Mkmy//fZJkokTJ9Yf88QTT2TTTTctdUkAAABYaxQL7OOOOy7vvfdeDjrooNx3331p27Zt+vXrl2uuuSbf+ta3cuSRR+aJJ57I7rvvXuqSAAAAsNYo9jFdXbt2zejRo3P55Zdnk002SZKce+65mT17du67774kSffu3XPqqaeWuiQAAACsNYoF9g9/+MN85StfyS9+8Yv6be3atctdd92V6dOnp1WrVunUqVMqKipKXRIAAADWGsUCe8yYMRk7dmw6duyYAw88MAcddFC22WabJP//fmwAAABYVxW7B/vXv/51Bg8enPfeey8jRozIAQcckK997Wu54YYb8re//a3UZQAAAGCtVCywP/e5z+Wcc87J5MmTc+ONN2bgwIF57bXXcuGFF2avvfbKsccem/Hjx+ftt98udUkAAABYaxQL7DoVFRXp06dPfvCDH2TKlCm57rrr8tWvfjUvvPBCzjrrrOyxxx6lLwkAAACNrnhgf9jSpUuzaNGi1NbWJklqa2vTvHnz1XlJAAAAaBTFHnJWZ8mSJZk6dWruvffeTJw4MQsXLkzz5s3zhS98IQMGDMg+++xT+pIAAADQ6IoF9kMPPZR77703EyZMyPz581NbW5sePXpkwIABOeCAA/KpT32q1KUAAABgrVMssI899tgkSadOnTJkyJAMGDAgHTt2LHV6AAAAWKsVC+zBgwdnwIAB6d69e6lTAgAAQJNRLLC/973vlToVAAAANDmr9SniAAAAsL4Q2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKEBgAwAAQAECGwAAAAoQ2AAAAFCAwAYAAIACBDYAAAAUILABAACgAIENAAAABQhsAAAAKKBFYw/wSb3xxhu54oorMnny5MydOzebbrppdtttt3zzm99Mx44dlzl2/PjxGTVqVKqrq1NVVZX9998/p5xySjbeeONGmh4AAIB1VZNawX7jjTdy6KGH5te//nU6d+6cI488MjvvvHPuvvvufP3rX091dXX9sddcc03OOOOMLF26NIMHD87222+fUaNG5dhjj83777/feF8EAAAA66QmtYJ9xRVX5LXXXsuZZ56Zo48+un77HXfckdNPPz0XXXRRrr766syZMyeXX355evbsmdGjR6dly5ZJkssuuywjRozIuHHjMnjw4Mb6MgAAAFgHNakV7AcffDCbb755jjrqqGW2f+UrX8lWW22VqVOnZunSpRk3blxqampywgkn1Md1kgwbNiytW7fOLbfcsqZHBwAAYB3XZFawlyxZkhNOOCEtWrRIs2bL/16gsrIyixcvTk1NTR577LEkSZ8+fZY5plWrVunRo0emTp2aBQsWZJNNNlkjswMAALDuazKB3bx58+VWruv8+c9/zssvv5ytttoqlZWVmTVrVtq2bdvgw8zat2+fJJk5c2a6d+++WmcGAABg/dFkAntFli5dmvPPPz9Lly7NwIEDkyTz5s1Lhw4dGjy+btV64cKFq3Td+fPnZ8KECat0Dta8/v37N/YIsF7z/01Y8/zsg8blZ1/TU1VVlaqqqpV6bZO6B/ujamtr85//+Z95+OGHs9NOO9WvcNfU1KSysrLB19RtX7Ro0RqbEwAAgHVfk13BrqmpybnnnpvbbrstHTt2zIgRI+rjeYMNNsjixYsbfF3dR3RtuOGGq3T9qqqq9O3bd5XOAbC+sZIGwPrGz76m5/nnn1/p1zbJwH733XfzzW9+M5MnT06nTp3yy1/+MltuuWX9/qqqqixYsKDB19Zt94AzAAAASmpybxH/+9//nqOOOiqTJ0/ODjvskDFjxuQzn/nMMsd06tQpc+fOzXvvvbfc6+fMmZNmzZpl6623XlMjAwAAsB5oUoG9aNGinHDCCXnqqafSp0+fjB49Om3atFnuuN69e2fp0qV5/PHHl3v9k08+mS5duqR169ZramwAAADWA00qsH/2s59l2rRp6dmzZ0aOHLnCSD7wwAPTvHnzXHnllfX3XCfJ1VdfnYULF+awww5bUyMDAACwnmgy92C/8cYbufnmm5Mk2267bUaOHNngcccff3w6d+6cY445JiNHjszBBx+cfv365aWXXsqkSZPSq1ev+o/zAgAAgFKaTGA/9dRT9U8Gv/XWW1d43FFHHZVWrVrl1FNPTbt27TJmzJjceOON2WKLLTJ06NCcdNJJK/wILwAAAFhZTSaw991337zwwgv/9PEVFRUZNGhQBg0atBqnAgAAgA80qXuwAQAAYG0lsAEAAKAAgQ0AAAAFCGwAAAAoQGADAABAAQIbAAAAChDYAAAAUIDABgAAgAIENgAAABQgsAEAAKAAgQ0AAAAFCGwAAAAoQGADAABAAQIbAAAAChDYAAAAUIDABgAAgAIENgAAABQgsAEAAKAAgQ0AAAAFCGwAAAAoQGADAABAAQIbAAAAChDYAAAAUIDABgAAgAIENgAAABQgsAEAAKAAgQ0AAAAFCGwAAAAoQGADAABAAQIbAAAAChDYAAAAUIDABgAAgAIENgAAABQgsAEAAKAAgQ0AAAAFCGwAAAAoQGADAABAAQIbAAAAChDYAAAAUIDABgAAgAIENgAAABQgsAEAAKAAgQ0AAAAFCGwAAAAoQGADAABAAQIbAAAAChDYAAAAUIDABgAAgAIENgAAABQgsAEAAKAAgQ0AAAAFCGwAAAAoQGADAABAAQIbAAAAChDYAAAAUIDABgAAgAIENgAAABQgsAEAAKAAgQ0AAAAFCGwAAAAoQGADAABAAQIbAAAAChDYAAAAUIDABgAAgAIENgAAABQgsAEAAKAAgQ0AAAAFCGwAAAAoQGADAABAAQIbAAAAChDYAAAAUIDABgAAgAIENgAAABTw/9q797Aqyz2N4/eCkIOASiIoooaHhWXhTidrUgkkNHeQYukkmlBTOAUSNprNVEO0x8PWa5KCsPbkLpVIJTWZ0Ew6SIaRZXnEnYpQug0V5RjHxfzhBcVm4aklC+H7+Y/nfQ6/xcWl3Dzv+7wEbAAAAAAALICADQAAAACABRCwAQAAAACwAAI2AAAAAAAWQMAGAAAAAMACCNgAAAAAAFgAARsAAAAAAAsgYAMAAAAAYAEEbAAAAAAALICADQAAAACABRCwAQAAAACwAAI2AAAAAAAWQMAGAAAAAMACCNgAAAAAAFgAARsAAAAAAAsgYAMAAAAAYAEEbAAAAAAALICADQAAAACABRCwAQAAAACwAAI2AAAAAAAWQMAGAAAAAMACCNgAAAAAAFjAdR2wf/75Z40YMUJvv/222eubNm3SpEmTNHz4cI0dO1aLFi1SRUVF2xYJAAAAAOgUrtuAXVFRoZiYGJWXl5u9/sYbb+jZZ5+VyWTSjBkz5Ovrq7fffluPPfaYampq2rhaAAAAAEBHd4O1C7gaJ06cUExMjA4cONDq9VdffVV/+MMftHr1atnZ2UmSEhMT9frrr2vdunWaMWNGW5YMAAAAAOjgrrsd7LffflshISHKy8vTnXfeabbPunXrVFdXp6ioqKZwLUmzZ8+Ws7Oz1q9f31blAgAAAAA6iesuYK9atUpeXl5as2aNHnjgAbN9vv76a0nSHXfc0azd3t5ew4cPV15ensrKyq55rQAAAACAzuO6C9gvvfSSNm3apNtvv73VPoWFherZs6e6du3a4pqXl5ckKT8//5rVCAAAAADofK67Z7DHjBlzyT7nz59X3759zV5zcXGRpFYPR7tcpaWl+vjjj3/XHGh79957r7VLADo1/t0E2h7/9wHWxf991x9XV1e5urpe1djrbgf7ctTV1alLly5mrzW2V1dXt2VJAAAAAIAO7rrbwb4cDg4Oqq2tNXut8RVdjo6Ov2sNV1dXjRo16nfNAQCdDTtpAIDOhv/7rj+HDh266rEdcgfb1dW11UPMGtsbbxUHAAAAAMASOmTAHjBggM6ePauqqqoW106cOCEbGxv179/fCpUBAAAAADqqDhmwR4wYIZPJpN27dzdrr66u1nfffadBgwbJ2dnZStUBAAAAADqiDhmw77//ftna2iopKanpmWtJWrFihcrLyzVt2jQrVgcAAAAA6Ig65CFnAwcO1KOPPqq//OUvmjRpkgICAnTkyBF99tlnuv322zV16lRrlwgAAAAA6GA6ZMCWpGeeeUa9e/fWu+++q1WrVsnd3V0RERGKjo5u9RVeAIDOwWQyqbi4WGVlZaqurlZDQ4O1S0IHYDAYZG9vLxcXF7m5ucnGpkPeKAgAuIjrOmCHhYUpLCzM7DWDwaDw8HCFh4e3cVUAgPbMZDLpxx9/VGVlpbVLQQfT0NCgqqoqVVVVqaKiQt7e3oRsAOhkruuADQDAlSouLlZlZaVuuOEGeXp6qmvXroQgWITJZFJFRYVOnTqlyspKFRcXq2fPntYuCwDQhviNAgDQqZSVlUmSPD095eLiQriGxdjY2MjFxUWenp6Sfv1ZAwB0HvxWAQDoVKqrqyVJXbt2tXIl6Kgaf7Yaf9YAAJ0HARsA0Kk0HmjGzjWuFYPBIEkcngcAnRC/XQAAAFhQY8AGAHQ+BGwAAAAAACyAgA0AAAAAgAUQsAEAAAAAsADegw0AQCf02muvKSkpqVmbwWCQg4ODevXqpVGjRikiIkIDBw5sk3qMRqN8fX31wQcftMl6lzJ+/HgdP35cM2fO1PPPP2/tcgAA1wkCNgAAZgxY8KG1SzDr+OI/WnS+cePGaejQoZIkk8mk8vJy5eXlad26ddq8ebMSExN1zz33WHRNc6Kjo9WzZ89rvs7l2LNnj44fPy5HR0dlZGRo/vz56tKli7XLAgBcBwjYAAB0YkFBQQoLC2vR/vnnn+upp55SXFycNm3apP79+1/TOmJiYq7p/Fdi06ZNMhgMevTRR5WcnKyPPvpIISEh1i4LAHAd4BlsAADQgr+/v2JjY1VZWanXX3/d2uW0mZqaGm3ZskU333yzpk6dKoPBoPXr11u7LADAdYKADQAAzJoxY4bs7e21bds21dXVNbuWk5OjyMhIjRgxQsOHD9e0adO0devWFnMUFBQoNjZWAQEBGjZsmAIDAxUfH6/Tp08362c0GvXAAw80aztz5oxefPFFjRkzRn5+fpo+fbq+/fZbRUREKDAwsKnfhg0bZDQalZOTo7feekvBwcEaNmyYgoKClJKSovr6+sv+zJ9++qlKSko0evRoeXp66g9/+INyc3NVWFh42XMAADovAjYAADDL0dFRN998syorK3Xo0KGm9vXr1ysyMlKHDx/WxIkTNW3aNJ09e1axsbFasWJFU7/i4mJFRETo888/1x133KHIyEgNGjRIaWlpeuSRR1RbW9vq2ufOndP06dO1du1aDRkyROHh4aqqqtKsWbN09OhRs2OWLl2qpKQkjRgxoqn/8uXL9eqrr172Z960aZMkaeLEiZKkP/7xj2poaFB6evplzwEA6Lx4BhsAALTKw8NDkpp2nE+dOqWEhAT5+PgoNTVVPXr0kCTFxcUpIiJCiYmJCgwM1JAhQ5SZmamTJ09q4cKFmjJlStOcCQkJSk1N1c6dO1s9QC0pKUkFBQWaP3++HnvsMUkXDmGbO3eutmzZIi8vrxZjCgsLmz0vPnPmTE2YMEHp6emKi4u75GctLi5Wdna2Bg4cKF9fX0nSfffdp0WLFmnjxo2KjY2Vra3tZX7nAACdETvYAACgVY2nZ5eXl0uSNm/erJqaGs2ZM6cpXEuSg4OD5syZI5PJpI0bN0q6EIgl6cCBA81u046Li9MXX3zRariur69XRkaGvLy8FBER0dRuY2Oj+fPntxpyg4ODmx3G1rdvXw0cOFBnzpxRdXX1JT9rZmamamtr9cc//npS+4033qg777xTRUVF+uyzzy45BwCgc2MHGwAAtKqiokKS5OTkJEnav3+/pAvPYP/www/N+lZWVkqS8vLyJF14l3RycrJSU1OVmZmp0aNHa+zYsfL395e7u3uraxYUFKikpER33nlnizDdp08feXp6mh03YMCAFm0uLi6SLhxeZm9vf9HP2nh7+P3339+sPSQkRF988YXWr1+vcePGXXQOAEDnRsAGAACtOnHihCTJ29tbklRWViZJeu+991odU1JSIunC7eXp6elKSUlRVlaWMjIylJGRITs7O4WFhen55583+37pc+fOSVKr78Xu1auXioqKWrSbm8tgMEiSGhoaWq1Xko4ePap9+/ZJurATbs6OHTtUVFSkXr16XXQuAEDnRcAGAABmnT9/XkeOHJGrq6sGDRok6ded7O3btzeF7ovx9vbWwoULVV9fr/379ys7O1sbNmzQ2rVr5eLionnz5rUY4+zsLOnX29L/UeOuuiV98MEHkqS77rpL/fr1a3F9//79OnDggDZs2KDZs2dbfH0AQMdAwAYAAGatW7dOdXV1uu+++5pu1TYajdq+fbv27dvXImAfP35ca9eu1T/90z8pMDBQWVlZys7O1r//+7/L2dlZfn5+8vPz05QpU3TPPffom2++Mbuuj4+PnJyctHfv3hbXSktLlZ+fb9Fd5IaGBmVkZMjGxkZLlixpOtjtt7766is98sgjev/99xUVFdW0Mw4AwG9xyBkAAGghJydHycnJcnJyUlRUVFN7aGiobG1ttXz58mbvsq6rq9PLL7+slStX6vz585KkY8eOKS0tTWlpac3mbrztvE+fPmbXtrOzU0hIiPLz85uNNZlMWrp06UVf73U1du3apZMnT+qOO+4wG64l6Y477lDfvn1VWFioXbt2WXR9AEDHwQ42AABmHF/8x0t36gC2b9/eFHhNJpPKy8t18OBB7d69Ww4ODnrllVeavRJrwIABmjdvnhYvXqz7779fgYGB6tatm3bs2KGjR48qICBAoaGhkqSpU6dq3bp1WrZsmXJzc2U0GnX27Flt3bpVTk5OeuKJJ1qt6+mnn1Z2drbi4+OVlZWlQYMG6euvv9axY8fk4OAgGxvL7RE03h7eWLc5BoNBkydP1muvvab09HTdddddFlsfANBxELABAOjEsrKylJWV1fS1o6OjvLy8NGPGDM2aNcvs88iRkZHy8fHRypUrtW3bNplMJnl7e2vBggUKDw/XDTdc+PWiW7duWrNmjVJSUrRz507t2rVLzs7OGjt2rKKjozV48OBW63Jzc1NaWpqWLVum7Oxs5ebmys/PT6tWrVJUVJQcHR0t8vl/+eUXffTRR7K3t9f48eMv2nfy5MlKTk7Wxx9/rJKSEnXr1s0iNQAAOg4CNgAAnVBMTIxiYmKuery/v7/8/f0v2c/Dw0Px8fGX7Hf48OFmXxcWFsrT01PLli1r1l5TU6PS0lLdcsstTW1hYWEKCwszO+/q1asvuq6jo6P27NlzyfokycvLS4cOHbqsvgCAzolnsAEAQLvz5JNP6u6771ZpaWmz9nfeeUe1tbUaNWqUlSoDAKB17GADAIB25+GHH1ZCQoJCQkI0btw4OTo66uDBg/ryyy9lNBo1c+ZMa5cIAEALBGwAANDuhIeHy93dXatXr1ZmZqYqKyvVu3dvRUVFKSoqSvb29tYuEQCAFgjYAACgXQoODlZwcLC1ywAA4LLxDDYAAAAAABZAwAYAAAAAwAII2AAAAAAAWAABGwAAAAAACyBgAwAAAABgAQRsAAAAAAAsgIANAAAAAIAFELABAAAAALAAAjYAAAAAABZwg7ULAAAAbe+1115TUlJSszaDwSAHBwf16tVLo0aNUkREhAYOHNgm9RiNRvn6+uqDDz5ok/X+0YIFC7Rx48YW7a6urvLx8dGUKVP00EMPyWAwWKE6AMD1goANAIA58d2sXYF58SUWnW7cuHEaOnSoJMlkMqm8vFx5eXlat26dNm/erMTERN1zzz0WXdOc6Oho9ezZ85qvcymTJ0+Wl5eXJKmurk4lJSXKycnRCy+8oCNHjug//uM/rFwhAKA9I2ADANCJBQUFKSwsrEX7559/rqeeekpxcXHatGmT+vfvf03riImJuabzX67Jkydr1KhRzdpqamo0bdo0rVq1SuHh4df8ewEAuH7xDDYAAGjB399fsbGxqqys1OuvKWoTLgAAHS9JREFUv27tcqyqS5cuCg0NVUNDg3bv3m3tcgAA7RgBGwAAmDVjxgzZ29tr27Ztqqura3YtJydHkZGRGjFihIYPH65p06Zp69atLeYoKChQbGysAgICNGzYMAUGBio+Pl6nT59u1s9oNOqBBx5o1nbmzBm9+OKLGjNmjPz8/DR9+nR9++23ioiIUGBgYFO/DRs2yGg0KicnR2+99ZaCg4M1bNgwBQUFKSUlRfX19b/7e2FrayvpQtgGAKA1BGwAAGCWo6Ojbr75ZlVWVurQoUNN7evXr1dkZKQOHz6siRMnatq0aTp79qxiY2O1YsWKpn7FxcWKiIjQ559/rjvuuEORkZEaNGiQ0tLS9Mgjj6i2trbVtc+dO6fp06dr7dq1GjJkiMLDw1VVVaVZs2bp6NGjZscsXbpUSUlJGjFiRFP/5cuX69VXX/1d34eamhplZGTIyclJo0eP/l1zAQA6Np7BBgAArfLw8JCkph3nU6dOKSEhQT4+PkpNTVWPHj0kSXFxcYqIiFBiYqICAwM1ZMgQZWZm6uTJk1q4cKGmTJnSNGdCQoJSU1O1c+fOVg9QS0pKUkFBgebPn6/HHntM0oVD2ObOnastW7Y0HUT2W4WFhc2eF585c6YmTJig9PR0xcXFXdbn3bhxo3Jzc5vWKysrU3Z2toqKirR06dKmzwsAgDkEbAAA0KrGW6LLy8slSZs3b1ZNTY3mzJnTLGw6ODhozpw5ioyM1MaNG/Xss8/KZDJJkg4cOKBJkyY13WYdFxenf/u3f5O7u7vZNevr65WRkSEvLy9FREQ0tdvY2Gj+/Pnatm2b2XHBwcHNDiDr27evBg4cqLy8PFVXV8ve3v6Sn9fcq7qkC39oOHPmzCXHAwA6NwI2AABoVUVFhSTJyclJkrR//35JF57B/uGHH5r1rayslCTl5eVJksaPH6/k5GSlpqYqMzNTo0eP1tixY+Xv799quJYuPLddUlKiO++8symUN+rTp488PT3NjhswYECLNhcXF0kXbvO+nIC9atWqplPEGxoaVFFRoby8PP35z3/Wf/3Xf+nMmTOKjo6+5DwAgM6JgA0AAFp14sQJSZK3t7ckqaysTJL03nvvtTqmpOTCu7o9PDyUnp6ulJQUZWVlKSMjQxkZGbKzs1NYWJief/55s4eGnTt3TpJafS92r169VFRU1KLd3FwGg0HShbB8pQwGg5ydnTVy5EitWLFCAQEBevPNNzVr1qym4A4AwG8RsAEAgFnnz5/XkSNH5OrqqkGDBkn6dSd7+/btTaH7Yry9vbVw4ULV19dr//79ys7O1oYNG7R27Vq5uLho3rx5LcY4OztL+vW29H/UuKveltzc3DR48GDt27dP+fn5uu2229q8BgBA+8cp4gAAwKx169aprq5O9913X9Ot2kajUZK0b9++Fv2PHz+uJUuW6JNPPpEkZWVlKT4+XuXl5bK1tZWfn5+io6OVmpoqSfrmm2/Mruvj4yMnJyft3bu3xbXS0lLl5+db5PNdqdLSUkli9xoA0CoCNgAAaCEnJ0fJyclycnJSVFRUU3toaKhsbW21fPnyZu+yrqur08svv6yVK1fq/PnzkqRjx44pLS1NaWlpzeZuvO28T58+Zte2s7NTSEiI8vPzm401mUxaunTpRV/vda189NFHKigokI+Pj2666aY2Xx8AcH3gFnEAAMyJL7F2BW1i+/btTYHXZDKpvLxcBw8e1O7du+Xg4KBXXnml2SuxBgwYoHnz5mnx4sW6//77FRgYqG7dumnHjh06evSoAgICFBoaKkmaOnWq1q1bp2XLlik3N1dGo1Fnz57V1q1b5eTkpCeeeKLVup5++mllZ2crPj5eWVlZGjRokL7++msdO3ZMDg4OsrG5NnsEv31Nl3ThDwf79+/Xzp07ZWdnpxdffPGarAsA6BgI2AAAdGJZWVnKyspq+trR0VFeXl6aMWOGZs2apX79+rUYExkZKR8fH61cuVLbtm2TyWSSt7e3FixYoPDwcN1ww4VfL7p166Y1a9YoJSVFO3fu1K5du+Ts7KyxY8cqOjpagwcPbrUuNzc3paWladmyZcrOzlZubq78/Py0atUqRUVFydHR0fLfDLV8TVeXLl3k7u6ukJAQPfrooxo6dOg1WRcA0DEQsAEA6IRiYmIUExNz1eP9/f3l7+9/yX4eHh6Kj4+/ZL/Dhw83+7qwsFCenp5atmxZs/aamhqVlpbqlltuaWoLCwtTWFiY2XlXr159ybUlafHixVq8ePFl9QUAoDU8gw0AANqdJ598UnfffXfTwWKN3nnnHdXW1ja9qxoAgPaEHWwAANDuPPzww0pISFBISIjGjRsnR0dHHTx4UF9++aWMRqNmzpxp7RIBAGiBgA0AANqd8PBwubu7a/Xq1crMzFRlZaV69+6tqKgoRUVFyd7e3tolAgDQAgEbAAC0S8HBwQoODrZ2GQAAXDaewQYAAAAAwAII2AAAAAAAWAABGwAAAAAACyBgAwAAAABgAQRsAAAAAAAsgIANAAAAAIAFELABAAAAALAAAjYAAAAAABZAwAYAAAAAwAJusHYBAACg7b322mtKSkpq1mYwGOTg4KBevXpp1KhRioiI0MCBA9ukHqPRKF9fX33wwQdtst7F5Obmau3atdq7d69OnTolBwcHDR06VA8++KBCQkJkMBisXSIAoJ0iYAMAYMat79xq7RLM2jdrn0XnGzdunIYOHSpJMplMKi8vV15entatW6fNmzcrMTFR99xzj0XXNCc6Olo9e/a85utcTE1NjRISErR+/Xp17dpV/v7+Cg4OVnFxsT755BPNmzdPn332mZYtWyYbG24CBAC0RMAGAKATCwoKUlhYWIv2zz//XE899ZTi4uK0adMm9e/f/5rWERMTc03nvxwvvfSS0tPTFRgYqEWLFql79+5N18rLy/XUU0/pww8/VO/evTVv3jwrVgoAaK/48ysAAGjB399fsbGxqqys1Ouvv27tcq65Xbt2KT09XYMHD1ZiYmKzcC1Jzs7OSkxMlJOTk1JTU3Xu3DkrVQoAaM8I2AAAwKwZM2bI3t5e27ZtU11dXbNrOTk5ioyM1IgRIzR8+HBNmzZNW7dubTFHQUGBYmNjFRAQoGHDhikwMFDx8fE6ffp0s35Go1EPPPBAs7YzZ87oxRdf1JgxY+Tn56fp06fr22+/VUREhAIDA5v6bdiwQUajUTk5OXrrrbcUHBysYcOGKSgoSCkpKaqvr7/kZ01PT5ck/eu//qu6dOlitk/37t0VHx+v//7v/5adnd0l5wQAdD7cIg4AAMxydHTUzTffrD179ujQoUO69dYLz6WvX79eL7zwgtzc3DRx4kQ5OTkpKytLsbGxiouL0+zZsyVJxcXFioiI0Llz5zR+/Hj16tVLhw8fVlpamr766itt3ry51aB67tw5TZ8+XQUFBRo9erSMRqN27dqlWbNmqXv37mbHLV26VPn5+ZowYYICAgL04Ycfavny5aqqqlJcXNxFP2t2drYkafTo0Rft949/BAAA4LcI2AAAoFUeHh6S1LTjfOrUKSUkJMjHx0epqanq0aOHJCkuLk4RERFKTExUYGCghgwZoszMTJ08eVILFy7UlClTmuZMSEhQamqqdu7c2eoBaklJSSooKND8+fP12GOPSbpwCNvcuXO1ZcsWeXl5tRhTWFjY7HnxmTNnasKECUpPT79owK6qqtL58+fl7Oxs9YPWAADXN24RBwAArWq8Xbq8vFyStHnzZtXU1GjOnDlN4VqSHBwcNGfOHJlMJm3cuFHShUAsSQcOHGh2m3ZcXJy++OKLVsN1fX29MjIy5OXlpYiIiKZ2GxsbzZ8/X7a2tmbHBQcHNzuMrW/fvho4cKDOnDmj6urqVj9jaWmpJKlr166t9gEA4HKwgw0AAFpVUVEhSXJycpIk7d+/X9KFZ7B/+OGHZn0rKyslSXl5eZKk8ePHKzk5WampqcrMzNTo0aM1duxY+fv7y93dvdU1CwoKVFJSojvvvLNFmO7Tp488PT3NjhswYECLNhcXF0kXXsFlb29vdlzjgWaNQRsAgKtFwAYAAK06ceKEJMnb21uSVFZWJkl67733Wh1TUlIi6cLt5enp6UpJSVFWVpYyMjKUkZEhOzs7hYWF6fnnnzd7oFjjCd2t3a7dq1cvFRUVtWg3N5fBYJAkNTQ0tFpvly5d5OHhoZ9//llFRUXq1atXq33Pnj0rW1vbFqeMAwAgEbABAEArzp8/ryNHjsjV1VWDBg2S9OtO9vbt25tC98V4e3tr4cKFqq+v1/79+5Wdna0NGzZo7dq1cnFxMfs+aWdnZ0m/3pb+jxp31S1pzJgxSk9P186dOzV58uRW+yUlJem9997TSy+9pKlTp1q8DgDA9Y1nsAEAgFnr1q1TXV2d7rvvvqZbtY1GoyRp3759LfofP35cS5Ys0SeffCJJysrKUnx8vMrLy2Vrays/Pz9FR0crNTVVkvTNN9+YXdfHx0dOTk7au3dvi2ulpaXKz8+3yOf7rcZQ/Ze//KXFK8kanT59WpmZmTIYDLrrrrssXgMA4PpHwAYAAC3k5OQoOTlZTk5OioqKamoPDQ2Vra2tli9f3uxd1nV1dXr55Ze1cuVKnT9/XpJ07NgxpaWlKS0trdncjbed9+nTx+zadnZ2CgkJUX5+frOxJpNJS5cuVW1trcU+Z6ORI0dq4sSJOnr0qGJiYppuhW9UVFSkmJgYnT9/XtOnT7+s3XsAQOfDLeIAAJixb1bLHdqOaPv27U2B12Qyqby8XAcPHtTu3bvl4OCgV155pdkrsQYMGKB58+Zp8eLFuv/++xUYGKhu3bppx44dOnr0qAICAhQaGipJmjp1qtatW6dly5YpNzdXRqNRZ8+e1datW+Xk5KQnnnii1bqefvppZWdnKz4+XllZWRo0aJC+/vprHTt2TA4ODrKxsfwewcKFC1VWVqZPPvlE/v7+CggIkKenp06cOKEdO3aooqJC9957r+bPn2/xtQEAHQMBGwCATiwrK0tZWVlNXzs6OsrLy0szZszQrFmz1K9fvxZjIiMj5ePjo5UrV2rbtm0ymUzy9vbWggULFB4erhtuuPDrRbdu3bRmzRqlpKRo586d2rVrl5ydnTV27FhFR0dr8ODBrdbl5uamtLQ0LVu2TNnZ2crNzZWfn59WrVqlqKgoOTo6Wvx74ejoqDfeeEMff/yx0tPTtWfPHhUVFcnR0VF+fn566KGHNHHiRIuvCwDoOAjYAAB0QjExMYqJibnq8f7+/vL3979kPw8PD8XHx1+y3+HDh5t9XVhYKE9PTy1btqxZe01NjUpLS3XLLbc0tYWFhSksLMzsvKtXr77k2r9la2urCRMmaMKECVc0DgAAiWewAQBAO/Tkk0/q7rvvbvFu6nfeeUe1tbUaNWqUlSoDAKB17GADAIB25+GHH1ZCQoJCQkI0btw4OTo66uDBg/ryyy9lNBo1c+ZMa5cIAEALBGwAANDuhIeHy93dXatXr1ZmZqYqKyvVu3dvRUVFKSoqSvb29tYuEQCAFgjYAACgXQoODlZwcLC1ywAA4LLxDDYAAAAAABZAwAYAALCghoYGa5cAALASAjYAoFMxGAySJJPJZOVK0FE1BuzGnzUAQOdBwAYAdCqNh2NVVFRYuRJ0VI0/WxzEBgCdDwEbANCpuLi4SJJOnTqlsrIymUwmbunF79bQ0CCTyaSysjKdOnVK0q8/awCAzoNTxAEAnYqbm5sqKipUWVmpn376ydrloINycnKSm5ubtcsAALQxAjYAoFOxsbGRt7e3iouLVVZWpurqanawYREGg0H29vZycXGRm5ubbGy4URAAOhsCNgCg07GxsVHPnj3Vs2dPa5cCAAA6EP60CgAAAACABXTogF1XV6e3335bEydO1G233aZx48YpOTlZtbW11i4NAAAAANDBdOiAnZCQoEWLFql79+565JFH5OHhoVdffVXPPPOMtUsDAAAAAHQwHfYZ7G+//VZr167V+PHjlZiYKIPBoIaGBi1YsECbNm3Sp59+qoCAAGuXCQAAAADoIDrsDnZqaqokKTo6WgaDQdKF0z3nzp0rg8Gg9evXW7M8AAAAAEAH02ED9u7du9WjRw8NGTKkWbuHh4cGDBigr7/+2kqVAQAAAAA6og4ZsGtqanTq1Cn169fP7HUvLy+VlpaquLi4jSsDAAAAAHRUhoaGhgZrF2FpRUVFGjNmjEaPHq233nqrxfWnn35aW7Zs0bZt29S/f/8rnv+7775TXV2d6uvrLVEu2pCrq6u1SwA6tdLSUmuXAHQ6/N8HWBf/911/bG1tdcMNN2j48OFXPLZDHnJWV1cnSerSpYvZ643t1dXVVzW/ra2tJKlr165XNR4AOit+0QcAdDb833f9qa6ubsp8V6pDBmwHBwdJavV91zU1NZIkR0fHq5r/1ltvvbrCAAAAAAAdVod8BtvZ2Vk2NjYqLy83e72srEyS5OLi0pZlAQAAAAA6sA4ZsLt06aI+ffrop59+Mnv9p59+kpubm7p3797GlQEAAAAAOqoOGbAlacSIETp9+rTy8/Obtf/88886fvy4/Pz8rFQZAAAAAKAj6rABe9KkSZKkV155RSaTSZLU0NCg//mf/5EkTZs2zWq1AQAAAAA6ng75mq5GcXFxyszM1G233aZRo0Zpz5492r17t8aPH6/ExEQZDAZrlwgAAAAA6CA6dMCura3Vm2++qY0bN+rnn39Wnz59FBoaqscff7zVV3gBAAAAAHA1OnTABgAAAACgrXTYZ7ABAAAAAGhLBGwAAAAAACyAgA0AAAAAgAUQsAEAAAAAsAACNgAAAAAAFkDABgAAAADAAgjYAAAAAABYAAEbAAAAAAALIGADaLdKS0v17rvvNn1dUlKiuXPnauzYsXr44YeVk5NjxeoAAACA5gwNDQ0N1i4CAP5RYWGh/uVf/kXnzp3TZ599Jg8PD8XGxuqjjz6Sk5OTqqurZTAYtGbNGg0fPtza5QIAYDGFhYXasmWL8vLyVFZWph49emj48OGaMGGCbrzxRmuXB+AiCNgA2qX58+frww8/1DPPPKPw8HCVlZVp7NixGjhwoNavX6/Tp0/roYce0vDhw7VixQprlwsAgEUkJSVpxYoVqqura3HNyclJ//mf/6kpU6ZYoTIAl+MGaxcAAObk5OQoODhYjz76qCRp8+bNMplMmjRpkhwcHOTt7a3x48dr69atVq4UAADL2Lhxo5KSkuTt7a2oqCjddttt6tmzp8rKyrRnzx6lpKTohRdekIeHh0aPHm3tcgGYQcAG0C6VlJSoX79+TV9nZ2fLYDA0+4XC2dlZNTU11igPAACLe/vtt+Xp6am1a9fKzc2tqd3NzU39+/fX3XffrcmTJ2vFihUEbKCd4pAzAO2Sp6enfvzxR0lSTU2NvvzyS7m7u8toNDb1+e6779S7d29rlQgAgEUVFBQoKCioWbj+LXd3d9177706cOBAG1cG4HKxgw2gXRo5cqQ2b96spKQkHT58WBUVFU3PnP3444/661//qm+//VaPP/64lSsFAMAy3N3ddf78+Yv2KS8vV48ePdqoIgBXih1sAO3SM888o8GDByspKUkff/yx+vbtq9mzZ0uSVq1apXfffVfDhw8nYAMAOoyIiAht3bpVH3/8sdnr33zzjbZu3arw8PA2rgzA5eIUcQDtVuOt4SaTSf/8z/8sBwcHSdLu3bt1+vRpBQUFyc7OzspVAgBgGenp6Xr33Xd16NAhjRgxQiNHjpSHh4eqq6u1b98+bdu2TY6Ojpo0aZIMBkPTOIPBoAULFlixcgCNCNgAAABAO+Dr63tV4wwGgw4dOmThagBcDZ7BBtCu5eTkaPPmzU3PYXfv3l3Dhg3T5MmTNWzYMGuXBwCAxaxatcraJQD4ndjBBtAu1dXV6dlnn1VmZqYaGhpkY2Mje3t7/fLLL5IkGxsbPf7444qLi7NypQAAtK3q6mrZ29tbuwwAZnDIGYB2aeXKlfrwww81atQorV27Vnv37tWePXu0e/duvfnmm/L19dWbb76p999/39qlAgBgES+++GLTH5Jbs3v3boWGhrZRRQCuFDvYANql8ePHy9bWVps2bVKXLl1aXC8vL9ekSZPk6OiojIwMK1QIAIBl+fr6qn///lqyZImGDx/e7FpVVZWWLVumd999V5J08OBBa5QI4BLYwQbQLv3973/X2LFjzYZrSXJ2dlZAQIAKCgrauDIAAK6Np59+WidPnlR4eLheeeUV1dXVSZJyc3MVEhKiNWvW6KabbuJZbaAd45AzAO1S//799dNPP120T3Fxsfr06dNGFQEAcG3Nnj1bAQEBWrBggd544w3t2LFDw4YN0/vvvy87OzvNmTNHjz/+OK+oBNoxdrABtEtz5szRp59+qtTUVJl7kmXr1q366KOPNHv2bCtUBwDAtWE0GpWenq4pU6bo0KFDSk9Pl6urqzZu3Kgnn3yScA20czyDDaBdiI6ObtG2Z88eFRcXq1+/fvLz89ONN96osrIy7du3T3/729/Ur1+/pr/0AwDQEVRVVSkpKUnvvPOO6uvr5eXlpR9//FFGo1Evv/yybrvtNmuXCOAiCNgA2gVfX9+rGmcwGHTo0CELVwMAQNv75JNP9Kc//UknT56Uj4+PFi1apFtuuUUpKSl64403ZDKZNG3aND3zzDNydna2drkAzCBgA2gXTpw4cdVjvby8LFgJAADW4evrK1tbWz366KOKiYlpdtBnXl6ennvuOR06dEi9evXSjh07rFgpgNYQsAEAAIB2ICQkRAsXLtStt95q9np9fX3Tbva+ffvauDoAl4NTxAG0SwcPHtT27dv1448/6ty5c3Jzc1OfPn0UFBSkYcOGWbs8AAAsbsOGDRc9xMzW1lYTJ06Ura1tG1YF4Eqwgw2gXTl69Kiee+65pr/M/+M/UQaDQTfffLOWLFmiQYMGWaNEAAAsYujQoYqOjtZTTz3VrP3777/X999/r0ceeaTFmKSkJCUnJ3P+CNBOsYMNoN34/vvvNWvWLFVVVcloNOree++Vj4+PunfvrsrKSh09elRZWVnau3evHnzwQf3v//6vRo4cae2yAQC4Kg0NDWZfRZmdna3k5GSzARtA+0bABtAulJeXKzo6Wg0NDfrzn/+s0NDQFn2CgoIUFRWlbdu26dlnn9XcuXP1f//3f3J1dbVCxQAAAEBzNtYuAAAkKT09XadPn9aiRYvMhuvfCg4O1pIlS1RUVKQNGza0UYUAAADAxRGwAbQLW7Zska+vryZOnHhZ/YODg3XLLbdoy5Yt17gyAAAA4PIQsAG0C4WFhbr99tuvaMzIkSNVUFBwjSoCAAAArgwBG0C78Msvv8jJyemKxnTp0kW1tbXXqCIAAADgyhCwAbQLHh4e+tvf/nZFYw4fPiwvL69rVBEAAABwZThFHEC7MGrUKG3YsEHHjh2Tj4/PJfsfOXJEX375pWbMmNEG1QEAcG3k5uYqKSmpWdtXX30lSUpOTm7xGq/GawDaJ0ODuZfvAUAbO3r0qEJDQ+Xj46O//vWv6tmzZ6t9//73vysiIkInT57Uli1b1Ldv3zasFAAAy/D19b2qcQaDQYcOHbJwNQAsgYANoN146623tHTpUvXo0UPh4eEaM2aMbrrpJjk7O6uqqkrHjx/X9u3btWbNGpWWlupPf/qTHnzwQWuXDQDAVdm4ceNVj508ebIFKwFgKQRsAO1KamqqlixZ0urhZQ0NDXJxcdFzzz2nsLCwNq4OAAAAaB0BG0C7c/r0ab333nvKyclRfn6+ysrK1K1bN3l7eysoKEiTJk266C3kAAAAgDUQsAEAAAAAsABe0wUAAAAAgAUQsAEAAAAAsAACNgAAAAAAFkDABgAAAADAAgjYAAAAAABYAAEbAAAAAAALIGADAAAAAGABBGwAAAAAACyAgA0AAAAAgAUQsAEAAAAAsAACNgAAAAAAFkDABgAAAADAAv4fMygIE7I91CMAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 380,
       "width": 492
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "table = pd.DataFrame([observed, expected], columns=['Design A', 'Design B', 'Design C'], index=['Obs', 'Exp'])\n",
    "\n",
    "ax = table.plot(kind='bar', title='Number of Items', figsize=(8, 6))\n",
    "ax.set_ylabel('value')\n",
    "plt.grid(color='darkgray')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "tillamook = [0.0571, 0.0813, 0.0831, 0.0976, 0.0817, 0.0859, 0.0735, 0.0659, 0.0923, 0.0836]\n",
    "newport = [0.0873, 0.0662, 0.0672, 0.0819, 0.0749, 0.0649, 0.0835, 0.0725]\n",
    "petersburg = [0.0974, 0.1352, 0.0817, 0.1016, 0.0968, 0.1064, 0.105]\n",
    "magadan = [0.1033, 0.0915, 0.0781, 0.0685, 0.0677, 0.0697, 0.0764, 0.0689]\n",
    "tvarminne = [0.0703, 0.1026, 0.0956, 0.0973, 0.1039, 0.1045]\n",
    "\n",
    "fstat, pval = stats.f_oneway(tillamook, newport, petersburg, magadan, tvarminne)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0002812242314534543871813732085485071365837939083576202392578125000000000000000000000000000000000000\n",
      "7.1210194716424473071469947171863168478012084960937500000000000000000000000000000000000000000000000000\n"
     ]
    }
   ],
   "source": [
    "print(format(pval, '.100f'))\n",
    "print(format(fstat, '.100f'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### two-way 카이제곱검정"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 컨틴전시 테이블 제작\n",
    "\n",
    "male = [269, 83, 215]  # 디자인시안 A, B, C 순으로 클릭 량\n",
    "female = [155, 57, 181]   # 디자인시안 A, B, C 순으로 클릭 량\n",
    "cont_table = pd.DataFrame([male, female], columns=['Design_A', 'Design_B', 'Design_C'], index=['Male', 'Female'])\n",
    "\n",
    "chi2, p_val, df, expected = chi2_contingency([male, female])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7.0942644148\n",
      "0.0288071342\n",
      "2\n",
      "[[250.425   82.6875 233.8875]\n",
      " [173.575   57.3125 162.1125]]\n"
     ]
    }
   ],
   "source": [
    "print(format(chi2, '.10f'))\n",
    "print(format(p_val, '.10f'))\n",
    "print(df)\n",
    "print(expected)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### two way ANOVA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv(\"two-way-anova-example.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Unnamed: 0', '체류시간', '성별', '디자인시안'], dtype='object')"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "예시 데이터 셋에서 필요한 칼럼들만 아래\n",
    "\n",
    "formula 스트링 안에 넣어줍니다. R 언어는 모르지만 칼럼(C) 파트에 잘만 집어넣어주면 됩니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "formula = '체류시간 ~ C(성별) + C(디자인시안) + C(성별):C(디자인시안)'\n",
    "model = ols(formula, data).fit()\n",
    "aov_table = sm.stats.anova_lm(model, typ=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sum_sq</th>\n",
       "      <th>df</th>\n",
       "      <th>F</th>\n",
       "      <th>PR(&gt;F)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>C(성별)</th>\n",
       "      <td>205.350000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>15.571979</td>\n",
       "      <td>2.311828e-04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C(디자인시안)</th>\n",
       "      <td>2426.434333</td>\n",
       "      <td>2.0</td>\n",
       "      <td>91.999965</td>\n",
       "      <td>4.046291e-18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>C(성별):C(디자인시안)</th>\n",
       "      <td>108.319000</td>\n",
       "      <td>2.0</td>\n",
       "      <td>4.106991</td>\n",
       "      <td>2.186027e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Residual</th>\n",
       "      <td>712.106000</td>\n",
       "      <td>54.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     sum_sq    df          F        PR(>F)\n",
       "C(성별)            205.350000   1.0  15.571979  2.311828e-04\n",
       "C(디자인시안)        2426.434333   2.0  91.999965  4.046291e-18\n",
       "C(성별):C(디자인시안)   108.319000   2.0   4.106991  2.186027e-02\n",
       "Residual         712.106000  54.0        NaN           NaN"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "aov_table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0002311828\n",
      "0.0000000000\n",
      "0.0218602690\n"
     ]
    }
   ],
   "source": [
    "print(format(aov_table.iloc[0,-1], '.10f'))\n",
    "print(format(aov_table.iloc[1,-1], '.10f'))\n",
    "print(format(aov_table.iloc[2,-1], '.10f'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "위 2-way 아노바 표의 해석은 다음과 같다\n",
    "\n",
    "- PR(>F)는 F값을 바탕으로 도출한 p-value이다. \n",
    "    + 성별과 디자인시안 사이에는 0.02의 p값이 있으므로 유의미한 관계가 있다고 검증"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAwoAAAILCAYAAABIE2swAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAAWJQAAFiUBSVIk8AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAgAElEQVR4nOzdeZyNdf/H8dcsZsFYxpppNJEsg8hSQslPmxbpzloMwriVpdtyVyQ3ihZKlog7IpRbiooS2SOyjGgQpjANTcYyxpnlzDm/Py5nzhyzmxnnnDnv5+PR45HvXNe5vhfHcb3P57t4Wa1WKyIiIiIiIpl4O7sDIiIiIiLiehQUREREREQkCwUFERERERHJQkFBRERERESyUFAQEREREZEsFBRERERERCQLBQUREREREclCQUFERERERLJQUBARERERkSwUFEREREREJAsFBRERERERyUJBQUREREREsvB1dgc80S+//EJ6ejr+/v7O7oqIiIiIlGApKSn4+PjQqFGjAp+roOAE6enpWCwWZ3ej2JhMJgACAwOd3BMRKUn02SIixaGkf7YU5plTQcEJbJWE+vXrO7knxSM6OhooufcnIs6hzxYRKQ4l/bPFdn/Xw+2CQnx8PDNmzGDz5s2cO3eO8uXL06pVK4YNG0ZoaCgA7du3JzY2NtfXmTx5Mk899VSux4waNYrVq1dn+7MBAwYwcuTI67sJEREREREX51ZBIT4+ni5duhAXF0fr1q3p2LEjMTExfP3112zdupXPPvuMsLAwevfuTWJiYpbzk5OT+eijj/D398/XOK3Dhw9TuXJlunfvnuVnzZo1K5J7EhERERFxRW4VFGbMmEFcXBwvvfQSffv2zWhftWoVo0ePZsqUKcyZM4c+ffpke/6kSZOwWCyMGTOGOnXq5HqttLQ0YmJiaNeuHUOGDCnK2xARERERcXlutTzq+vXrCQ4OJiIiwqG9U6dO1KxZk23btuU4YePnn3/mk08+oXXr1nTp0iXPax0/fpy0tDTq1q1bJH0XEREREXEnblNRSE9PJzIyEl9fX7y9s+YbPz8/0tLSMJvN+Pn5Zfn5m2++iY+PD2PHjs3X9Y4cOQKgoCAiIiIiHsltgoKPj0+WSoLN8ePHOXHiBDVr1sw2JKxbt44DBw7QtWtXatWqla/r2YJCTEwM3bt358iRIwQEBNCuXTuGDx9OtWrVrv9mRERERERcnNsEhZxYLBYmTpyIxWKha9eu2R6zYMECvL29ee655/L9uragMHv2bB544AGaNGlCVFQUK1euZPv27Sxfvpzq1atfd79NJlOhlqtyZbb1iEvq/YmIc+izRUSKQ0n/bDGZTNe9R4RbBwWr1cq4cePYsWMHDRs2zLbi8Ouvv7J3714efPBBwsLC8v3aAQEBhIWFMXPmTIeJzx988AHvvfcekyZNYubMmUVxG7myWq2kp6eTnp6O1Wot9usVBVs/k5OTndyTksPLywsfHx98fHzw8vJydndERETEA7htUDCbzbz66qusXLmS0NBQZs+ene2woy+//BIgx2pDTmbNmpVte2RkJCtWrGDjxo0kJSVRpkyZgnceY/e/vDb2sFgsnDp1iitXruDr6z5/VLYJ5dnNJZHCKV26NKGhofq9FY9U0jdFEhHnKOmfLR614RoYJZRhw4axefNmwsLCWLBgQY5zBjZu3EiFChVo1apVkVzb29ubevXqcfr0ac6cOUPt2rWL5HWzk5CQkBESqlevTpkyZdziAbGkb4V+o1ksFpKSkjhz5gxXrlwhISGBypUrO7tbIiIiUsK5/lPnNS5evEhERASbN2+mQYMGLF26lBo1amR77IkTJzh58iTt27cv0DfyJpOJ/fv3c/jw4Wx/bhtS4+/vX/AbKADbpnHVq1cnKCjILUKCFD1vb2+CgoIy5sRkt5mgiIiISFFzqyfPlJQUIiMjiYqKomXLlixevJhKlSrleHxUVBRQ8F2U//77b7p168aoUaOy/MxkMvHrr78SHBxMSEhIwW6ggFJSUgCue3iTlCy294HtfSEiIiJSnNwqKEybNo19+/bRtGlT5s2bR9myZXM9/tdffwUgPDy8QNcJDQ0lPDyco0ePsnr16ox2q9XK1KlTSUhIoEePHsU+qdQ2KViVBAEy3m/uMqldRERE3JvbzFGIj49nyZIlANSqVYt58+Zle9zAgQMzhgSdOnUKgKpVq+b62gsXLiQxMZGIiAjKlSsHwIQJE+jVqxejR49m3bp1hISE8PPPP3Pw4EFatGjBoEGDiurWRPJFqx2JiIjIjeQ2QSEqKoq0tDQAPv/88xyPi4iIyAgKFy5cACAoKCjX1160aBGxsbF07tw5Iyg0bNiQFStW8P7777Nz5042bdpESEgIQ4cOZcCAAdmusCQiIiIiUlK4TVDo0KFDxiZo+fXpp5/m67gffvgh2/batWszffr0Al1TRERERKQk0OB3ERERERHJQkFBRERERESycJuhR1KyrF+/nh9//DHP4+655x46dOhQ6PNEREREsuNz7pyzu+CyFBTEacaNG5fnMevXry+y80REREQc/PUXt/TpY/z/jz9CHitlehoNPRIRERERz/PXX3D//fgfP47/8eNw//1Gm2RQUBARERERz3I1JHB1c17A+H+FBQcKCiIiIiLiObILCTYKCw4UFERERETEM+QWEmwUFjIoKIiIiIhIyZefkGCjsAAoKIiIiIhISZdDSDDjw3sM4z2GYcbH8RyFBS2PKiIiIiIlWC6VhKX05EXeAyCYBHqz2PEAW1jYuNEjl05VRUFERERESqZcQoIZHybyasavJ/Jq1qoCeHRlQUFBREREREqePOYkLKUnx6iT8etj1GEpPbN/LQ8NCwoKIiIiIlKy5BESrq0m2ORYVQCPDAuaoyBOcfbsWSZMmJDncbVr1y6S80RERMRD5GN1o2urCTa2qkKWuQo2HjZnQUFBnOKZZ565oeeJiIiIB8hHSMipmmAzkVfpyVJ8Sc/+AA8KCxp6JCIiIiLuL5/7JORUTbDJda6CjYcMQ1JQEBERERH3ls+QkFc1wSbXuQo2HhAWFBRERERExCPkVU2wyVdVwQMoKIiIiIiIe6ta1Zgz0KBBlh9ZgWPUZj79eIGZ+X7JPKsKDRqU+HkKmswsIiIiIu7valiwtGvPwWhvttKWLdzLVtoSR40Cv1yuKyB5QEgABQXJzDbGroS/6UVERKTkSEuDPXtg61bYsqUq2+MOcL6IBs1kuwKSh4QEUFAQG9skIPCYN7+IiIi4nytXYOdOWzAw/v/KlcxHZA0JAVwhmdIFvlaWqoIHhQRQUBDIulKAh6wNLCIiIq7vwgXYvt0IBVu2GNWDtLTcz6laOZ22vju498xyWvEjPVmWr0nM2cmoKjSo63HPRwoKni675cQ8aCMRERERcS1nztirBVu3woEDYLXmfk5YGLRtC/feC20vr+X2ib3wOnMOgEX0uu6QAFerCjVG0Xvjix73XKSg4MlyW3NYYUFERESKmdUKMTGOweC33/I+r0GDTMGgLYSGAvHxMHgwrFiRcZwZHyYGvQWJhevnRP+J9Az29bgHZ0+7X7HJz8YkCgsiIiJShCwW4/EiczCIjc39HG9vuPNOezBo0wYqV77moBUrjJAQH29vCw1labdvOPZO9UL3+1iML0uXQu/ehX4pt6Kg4InyuXshoLAgIiIi1y0tDfbtsweDbdsgISH3c/z94a677MGgVSsICsrh4L//hhdegM8+c2zv3x/zlHeYeHf5IrkPgIkToWdP8PWgp2cPulUBChYSbBQWREREJB9MJti1yz7xeMcOSErK/ZygILjnHvswohYtICAgHxf74gsYNMi+vDtASAjMnw8PP8zSRXDsWKFux8GxY3hcVUFBwZNcT0iwcYGwsHLlSubMmcOff/5JuXLlWLp0KWFhYU7pS25Onz7N//3f//F///d/zJ4929ndERERKTYXL8KPP9qHEe3alfeKRJUrO84vuOOOAn5Lf+4cDBkCy5Y5tvftC9OmQYUKmM1GBaCoeVpVwUNuUwoVEmycGBaOHz/OmDFjKFu2LD179sTb25saNQq+y6KIiIhcv7/+cpxfEBVlzDvITWioEQpswaBePfDyus4OrFoFkZFw9qy9rUYNmDcPOnbMaFq6tGirCTaeVlVQUPAERRESbJwUFqKjo7FYLPTs2ZMXX3zxhl1XRETEU1mt8McfjsHgyJG8z6tXz7FicMstRdCZhAQYNgw++cSxvXdveO89qFgxS3OeD/PXPh952GZq+aGgUNIVZUiwcUJYSE1NBaDiNR8EIiIiUjSsVoiOdgwGp07lfo63NzRp4rgiUZE/Gnz9NQwcCHFx9rbq1eHDD+Hxx6//datWhY0bSbnnHgD8FRKyUFAoyYojJNjcwLDQvn17Yq+unTZ58mQmT57MCy+8wJAhQzh06BCzZs1iz549mEwmbr31Vrp370737t3xylTXbN++PTVr1uTll1/mrbfeYu/evfj5+fHQQw/x8ssvc+nSJSZPnszWrVvx8/OjdevWvPLKKwQHB2e8RlpaGkuXLuWbb77h+PHjpKSkUKVKFdq2bcvw4cMdjs1OamoqCxYsYNWqVZw6dYqyZctyzz33MHz4cEJDQ4vnN09ERCQHZrMxdMg28XjbNmMRodz4+RmTjW3VgnvugfJFt7CQowsXYPhw+Phjx/Znn4Xp0yGPf3fzpWpV/li4EIDbFRKycLugEB8fz4wZM9i8eTPnzp2jfPnytGrVimHDhjk8bP3vf/9j7Nix2b7GHXfcwfLly/O8lslkYu7cuXzzzTecPXuWm2++mWeeeYaePXs6PIS6pOIMCTY3KCz07t2bXbt2sWHDBtq0aUOTJk1o2bIlmzdv5oUXXqBUqVI8+OCDBAcHs3XrVsaPH8+vv/7KxGtmMZ0+fZoePXrQpEkTunfvztatW/nss8+4cOECBw8epHLlynTt2pV9+/bx1VdfYTKZmDVrVsb5I0aM4LvvvqNZs2Z07dqV1NRUtm3bxmeffcahQ4f4/PPPc7yHtLQ0BgwYwM6dO2ncuDHPPvss586dY+3atWzbto3Fixdz++23F9vvoYiISHIy7N5trxZs3w6XL+d+TpkyjisStWwJgYE3oLNr1sCAAfDnn/a2atVg7lzo1KlIL5VeqVKRvl5J4lZBIT4+ni5duhAXF0fr1q3p2LEjMTExfP311xkPfbZVcI5cHUQ3YMAA/P39HV6nevW8N95IT09n2LBhbN68mfvuu4+HHnqILVu2MGHCBE6fPs2///3vIr+/InMjQoLNDQgLffr0oVy5cmzYsIG2bdvSp08fTCYT7du3JygoiOXLl3PzzTcDMHLkSIYPH87y5cvp0KED9913X8brnDp1it69ezNmzBgA/vnPf3Lvvffy3Xff8fDDD/Pee+/h5eVFeno6jzzyCOvXr8dkMhEYGMj+/fv57rvvePzxx3nnnXcyXtNsNtO5c2cOHjxITEwMt956a7b38PHHH7Nz50769+/PqFGjMtp79epFjx49eOWVV1iRaSdJERGRwkpMdFyR6Kef4OpI3hxVqmQMH7IFg6ZNb/AKPxcvwosvwoIFju09esCMGUYH5YZxq6AwY8YM4uLieOmll+jbt29G+6pVqxg9ejRTpkxhzpw5gBEUKlSowMiRI6/rWmvWrGHz5s3069cvIxQMGzaM/v37s2DBAp588knq1q1b+JsqjKlTYfz4vL8OKG6//mqk/KscvmgoW9bo44gRRXrJH374gYSEBEaPHp0REgC8vb0zvvn//PPPHYICGKHDply5ctSuXZuDBw/St2/fjCqRj48P4eHh/PHHH/z555/Url2b6tWrM2XKFJo3b+7wer6+vjRr1oyjR49y7ty5HIPCihUrKFeuXJaJ2I0aNeLhhx/mq6++4rfffqNOnTqF+W0REREPFh9vDB+yBYN9+/JekSgkxHFFovr1jXkHTvHdd9C/P5w+bW+rUgXmzIGnnnJSpzybWwWF9evXExwcTEREhEN7p06dmDlzJtu2bcNiseDt7c3Ro0cLNZRjyZIl+Pr6MmjQoIy2UqVKMXz4cLp3786KFSsyvpl2mqlTnR8S8nL5stHPIg4KBw8eBODQoUPMmDEjy899fHw4fPiwQ1upUqUICQlxaCtdujSAQ9gAMqpQtknU1atXp3PnzpjNZg4dOkRMTAwnT54kOjqaH3/8EQBLDp/GSUlJxMTEUKVKFT744IMsP//76oDQ6OhoBQUREcm3kycdJx5HR+d9Tp06jsEgLKwQS5UWlUuXjOeE+fMd27t2hZkzjbAgTuE2QSE9PZ3IyEh8fX3xzibq+vn5kZaWhtlsJiEhgQsXLlz3N/6pqan88ssv1KtXj/LXzNBp3LgxgYGB7N69+7peu0iNGOEaFYXclC1b5CEBIDExEYBvvvkmx2MuXrzo8OuAXLZ59PPzy/Oan376KbNmzeKvqztAlitXjjvuuIPatWsTFRWF1WrN9rzLV/984uPjmTlzZr77KyIiYmO1wtGj9onHW7caS5fmxssLGje2h4K2bY3FglzK99/Dc885Lq9UuTLMng1dujivXwK4UVDw8fHJUkmwOX78OCdOnKBmzZr4+fllzE9IS0tj8ODB7Nu3j+TkZO68806GDRtG48aNc71WbGwsZrOZmjVrZtuP6tWr8/vvvxf6ngptxIicH8Jv1DyFbNYcNplMAAQW42wnWyVg4cKFtGrVqtiuY7N27Vpee+016taty2uvvUZ4eDg33XQTAK+99hpRUVF59rV58+YsWbKk2PsqIiLuLz0dDhywh4KtW41/2nNTqhQ0b24PBq1bQ4UKN6a/BZaYCKNGGZOTM/vHP4yQoBWIXILbBIWcWCwWJk6ciMVioWvXroB9IvOnn35KmzZteOqpp/jjjz/44Ycf+Omnn/jggw9o27Ztjq954cIFAIKCgrL9eVBQEDExMZjNZnyvc4aPyWQiOo8aYXJyMn5+fhkP3gUSFARr1uD/yCN456cWeR0s9euTsmaNca1MfbQNwbmufufANgQoLS0tYxlUgH379tGkSROHYy9evMjcuXMJDw/n0UcfzeiT1WrN0qf09PSMvpYqVSqj3Ww2A8afgclkYtWqVQC8/vrr3HbbbQ7399tvvwGQkpKCyWQiOTk547VNJhO+vr7cdNNN/Pbbb5w/fz5LZeOrr77i9OnTPPHEE1mGRl3LYrGQmpqa53tHpCSy/Z3T+19KotRULw4eDODnn0uzZ09p9u0L5PJln1zPCQy0cMcdJpo3v0KzZldo3NhEYKC9uh0X57j1gKsovWMHN736Kn6ZVjQyV6jAmVdfJfHhh+HcOeO/G6Skf7bYFma5Hm4dFKxWK+PGjWPHjh00bNgwo+JgsVgICQlh+PDhPPHEExnH79q1iz59+vDyyy+zYcOGLKsh2dgeEnMajmJrT0lJue6gcENUrUrK2rXFEhYs9euTsnat0xJ/+/bteeedd1i4cCHt27fnlkzbPr733nt88cUXREZGFtn1bO+VhIQEh/avvvqKPXv2AEaIycnjjz/Ohx9+yPTp0xk1alTG8Lnjx48zZcoUwFgBSUREPENSkhdRUaWvBoNADhwIJCUl91nE5cql06zZlYz/GjRIJtN3XC7PKymJqtOmEbxsmUP7pQ4dODNuHOmVKzupZ5ITF37KzZ3ZbObVV19l5cqVhIaGMnv27IwH+EGDBjlMQrZp2bIljz/+OF9++SW7du3KsapgeyjM6cEvNTUVLy+vQg2tCQwMpH79+rkeY0u2hRrCc8stsGlT0Q5DatAA740bCcwhJBTH0CPbn22pUqUIDAwkMDCQSZMmMXLkSLp3706HDh2oWrUqu3fv5sCBAzRq1IjIyMiMPnh7e2f7Z+bj45PR18w/swXAgIAAAgMD6dy5M99++y3/+te/ePTRRylbtiy//PILu3btolKlSpw7dy4jsdsqBj4+Phmv+fzzz/PTTz+xbNkyoqKiaNmyJZcuXeLbb7/FZDLxzjvvUCUfk7W8vb0JCAjI870jUhLZPhP1/hd3dO6c44pEe/caw4tyc9NNjhOPw8N98PYOArIf8eDSNm2Cfv0gJsbeVrEizJxJuR49KOfEGdUl/bOlMJUStwwKJpMpY4+DsLAwFixYQLVMy3PmpkGDBnz55Zeczrz01jVsE5gv5zBJODExkdKlS2c7qdolXd2ivEjCQjZzEpzlkUceoXr16sydO5etW7diMpkICQlh8ODBPPfcc5QpU6bIrtWuXTveffdd5s2bx1dffUVAQAChoaGMGzeOpk2b0rlzZzZv3sxjjz2W7fkBAQEsWrSI+fPns2bNGpYuXUpQUBB33nknkZGRtGzZssj6KiIiznf6tH1uwZYtcOhQ3ufUrm0PBffeC7VqucCKRIWVlAQvvWSsXpTZE08Yy55ene8nrsnLmtNSLS7q4sWLDBgwgKioKBo0aMD8+fOpdM3mG4cOHeLKlSu0aNEiy/kffvghU6dOZfLkyTyVw5q8ZrOZJk2aEB4ezmeffebws/T0dJo1a0bt2rVz3Yk3N/lNrkWecAs7wTmfIeFGTGb2VCX9Ww+R3Oj9L67KaoVjx+zVgi1bHL84z46XFzRs6LgiUY0aN6a/N8yWLdC3L5w4YW+rUMHYOO2ZZ1wmBZX0z5bC3J9bVRRSUlKIjIzMGLrxwQcfULZs2SzHPf/885w9e5bt27cTHBzs8DPbePKGDRvmeB1fX1/uuOMODhw4wOXLlx2uceDAAUwmE02bNi2iu7qBClNZcKFKgoiIiDOlp8PBg47B4OzZ3M/x9YVmzRxXJLrmEaXkuHIFXnkF3n/fSFE2jz1mrHJU4hJRyeUmY2cM06ZNY9++fTRt2pR58+ZlGxIAHn74YSwWC++++67D2vZr165l06ZNtGjRIs/N2J588klSU1MdNvNKS0tj+vTpAHRx17V9bWGhQYP8n6OQICIiHiw1FXbsgDffNJ51K1WCJk1g6FD43/+yDwmBgcb3cq+9BuvXw4ULsHMnvPUWPP54CQ4J27bBHXfA9On2kFC+PCxcCKtXKyS4GbepKMTHx2esQV+rVi3mzZuX7XEDBw5k8ODBbNmyheXLl3PkyBGaNWtGTEwMmzZtokqVKkyePNnhnIULF5KYmEhERATlypUD4KmnnuLzzz9n4cKFHD16lPDwcLZu3crhw4fp16/fdW/m5hIKUllQSBAREQ+TlGQ81NsqBjt3OqwCnq3y5aFNG3vFoFkzyMdeniWHyQRjx8K77zpWER55BObNgzyW/xbX5DZBISoqKmMVotzmBtge9j/99FNmzpzJ999/z+LFi6lQoQJPP/00Q4cOpeo1D72LFi0iNjaWzp07ZwQFHx8f5s+fz4wZM1i7di179uyhZs2ajBs3jh49ehTfjd4o+QkLCgkiIuIBzp83vgi3DSPasweurpSeo2rVHCceN2wIPrlve1By7dgBffoYW0fblCtnhIa+fV1mLoIUnNtNZi4JnDaZOTs5TXAuREjQZObiU9InXInkRu9/KSp//um4ItHBg45fgmfn1lsdg8Ftt+n5F5MJxo2DadPg6marADz4IMyfD6GhzutbAZT0zxaPmcwsxSC7yoIqCSIiUkJYrcaiO5knHh8/nvd54eGOKxLdfHPx99Wt/PSTUUU4fNjeFhRkhIbnnlOKKiEUFMQxLIBCgoiIuC2LxdizIHMwiIvL/RwfH7jzTnswaNPGmLAs2UhOhvHj4e23HasIHTrAf/8LNWs6rWtS9BQUxGALC7b/FxERcQNpacYux7ZgsG2bMecgN/7+cPfd9mDQqhXksJCiZLZ7t1FFyDxcuWxZeOcdGDhQVYQSSEFB7BQQRETExV25Yox6sVULduww2nITFGRUCWzzC5o3N8KC5FNKCkyYYKwPm55ub7//fvjoIwgLc1rXpHgpKIiIiIjLunABtm+3B4OffzaqCLmpUsVx4nHjxh68IlFh7dljVBEOHrS3lS5tDD0aNAi83WpLLikgBQUBjGXgZs40/v+FF4wdJEVERG60M2ccVyQ6cCDvFYluucUxGNx+u0bBFFpqKkyaBG+84VhFuO8+o4pQq5bz+iY3jB4HBYClS+HFF43/Dw6G3r2d2x8RESn5rFb4/XfHice//Zb3efXrO65IpPmzRWz/foiIMFKaTWCgMfTo+edVRfAgCgqC2QwTJ9p/PXEi9OypqoKIiBQtiwWiox2DQWxs7ud4e0PTpvZqQZs2xtAiKQZpaUYFYdIkxx3n2rSBBQuMzSPEo+hRUFi6FI4ds//62DGjzVOrCu3bt+fSpUv8/PPPzu6KiIhbM5th3z57MNi6FRIScj/Hzw/uusseDFq1Mjb5lWIWFWXMRdi/394WGGgEh6FDVUXwUAoKHu7aaoKNqgoiIlJQJhPs2mUPBj/+CElJuZ9Ttiy0bm0PBi1aQEDAjemvYFQRpkwx/uHPPEv8nnuMKsLttzuvb+J0egz0cNdWE2w8vaogIiJ5u3jRCAO2YUS7dxtzYHNTubJ9bsG998Idd+hLKac5eNCYi7B3r70tIABefx2GDdNSUaKg4MlyqibYqKogIiKZ/fWX44pEUVGOm/NmJzTUcUWievW0IpHTmc3w1lvGDsuZqwh33w0LF0Ldus7qmbgYPQJ6sJyqCTauVFV46aWXWL16Ndu3b2fq1KmsX7+elJQU7rzzTl577TWqV6/O9OnTWbVqFVeuXCE8PJwxY8ZQr169jNfYuHEjS5Ys4eDBgyQmJhIUFMSdd97JkCFDqF+/fp59WLt2LR9//DFHjhzBy8uLRo0a8c9//pO77767OG9dRMRp/vjDceLxkSN5n1O3rj0U3HuvsXSpuJBffzWqCJnn4fn7G98O/utfqiKIAwUFD5VXNcHGlaoKVquV3r17Y7FY6Ny5M0ePHmXbtm1ERkZyyy23cPToUR5++GHi4+P59ttvGThwIN999x2BgYF88sknTJw4kZo1a/LYY49RqlQpfvnlFzZs2MDOnTv59ttvqZrLztTTp09n9uzZhISE0LlzZ7y8vPj222/p27cvU6ZMoVOnTjfwd0JEpOhZrXD4sGMwOHUq93O8vKBJE8cViapVuzH9lQIym2HqVBg3znF8WLvq5NQAACAASURBVMuWRhUhH1+Yiedxgcc/cYa8qgk2rlRVsFgsGQ/9fn5+AHTv3p19+/aRmprK6tWrKVu2LAAvv/wyK1euZNeuXbRq1Yp3332XsLAwvvjiC0qXLp3xmuPHj2fZsmVs3LiRbt26ZXvdAwcO8MEHH9CyZUs+/PBDAgMDAXjhhRfo1q0br732Gm3btiU4OLiYfwdERIqO2WwMHcq8ItHff+d+TqlSxnOlLRjccw+UL39j+iuFEB1trGi0a5e9zc8P/vMfGDnSNb4NFJekta7c2NSpEBRkfKNT0P8iIvJ/nYiIgr126dKBlC4diJeX0b+pU4vunnv06JEREgCaNm0KQLdu3TJCAkDjxo0BiI2NJT09nYkTJ/L66687hASAli1bAnDu3Lkcr7lixQqsViujR4/OCAkAFStWZMCAAZhMJtauXVv4mxMRKUbJyUYYeOMNePhhY3PN5s2N0SZffJF9SChTBh54ACZMgE2bjMnL27bB5MnwyCMKCS4vPR3eftvYiCJzSGjWzJjA/NJLCgmSK7073NjUqXD5srN7kbvLl41+jhhRNK9X85rtN20P/jfffLNDu7+/PwCpqakEBgbSsWNHAGJiYjh+/DgnT57kt99+Y8eOHYBRrcjJoUOHAFi3bh2bNm1y+NmZM2cAiI6Ovs47EhEpHomJjisS7doFKSm5nxMc7LgiUZMmRhVB3NCRI9C3L1z9dw4w/jDHj4fRoxUQJF/0LnFjI0YYf99dOSyULVt0IQHIUhGwyVxlyM7u3buZPHlyxkO/v78/9erVIzw8nLi4OKxWa47nJiYmAvDhhx/meMzFixfz6rqISLGKjze+7bcFg3378l6RKCTEceJx/fraV8vtpafD9OkwZoxRRrK5805jLkKjRk7rmrgfBQU3NmJEwR/CFy0q2LCjzD7+OH9zFUwmE4DDMB1nio2NpX///gQEBDBx4kSaNWtGWFgYPj4+rFmzhvXr1+d6funSpfHx8SEqKopS+mpNRFzEqVOOE4/zU9isU8ceDNq2hVtv1VKlJcpvvxlVhO3b7W2+vsYE5pdeUnlICkxBwYPkd6WjnLjSCkgFsX79epKTkxk9ejRdu3Z1+Nnx48cBcq0o1K1bl+joaKKjozPmPtjs37+f9evX065dO5o3b170nRcRwViR6OhReyjYssVYujQ3Xl7QuLFjMKhe/cb0V24wiwVmzICXXza2x7Zp0sSoItxxh9O6Ju7NzR75pDDyu9JRTlxpBaSCsM1X+PuamXqHDx9m0aJFAJjN5hzP79y5M19++SVvvPEG8+fPz5g0ffnyZcaPH090dDRt27Ytpt6LiCdKT4cDB+zBYOtWY7Oz3Pj6QosW9mDQujVUqHBj+itOdOwY9OtnvElsfH2NoUevvGKsbiRynRQUPERhqwk27lhVuP/++5k6dSpz587lxIkT1KxZkz/++IONGzcSFBQEwIULF3I8/+6776ZXr14sXryYRx99lPvuuw8/Pz/Wr19PXFwc3bt356677rpRtyMiJVBKirH/lS0YbN8Oly7lfk5goLE8qS0Y3HUX5DCNS0oiiwVmzTKGFF25Ym9v1MgYK3x1VUCRwnCjxz0pjMJWE2zcsapQrVo1FixYwLRp09i5cyfbtm2jRo0a9OrVi8jISB588EG2bt2K1WrFK4fBumPHjqVRo0YsW7aM1atX4+Pjw6233sqQIUPo3LnzDb4jEXF3ly8bi9HYgsFPPznOO81OhQrGhma2icd33qkh5x7rxAmjirB5s73Nx8eoIIwdqyqCFBkva26Ds6VY2JbSrJ/HLoj5PS4vZrOxkkVRBAWA224zJs3lVFVwtcnMJUlRvSdE3JE7v//PnXNckWjvXmN4UW5uuslxfkHDhlqRyONZLDBnjrG8aVKSvT083KgiNGvmvL65MXf+bMmPwtyfKgoeoKiqCTbuWFUQEbmRYmMdJx5fXZk5V7VrOwaD2rW1IpFk8vvv8Nxz8MMP9jZvb2Po0bhxcHU+nkhRUlAo4YpqbsK13HGugohIcbBajS9QMk88PnEi7/MaNXIMBjVqFH9fxQ1ZrfDhhzBypOPGSQ0aGCsatWjhtK5JyafHvBKuqKsJNqoqiMiNZDbDokUVAeOLCmd+SWGxwC+/OAaDq5u058jHB5o3d1yRKDj4xvRX3Ngff0D//pB5vx9vbxg1ythxNSDAaV0Tz6CgUML17q2HeRFxf0uXwpQpxiYA9evf2M+11FTYs8dxRaJcFkoDjOe3u++2VwvuvtvYqV4kX6xWmD/f2FU1MdHeXq+eUUXQSntygygoiIiIS7t2CGVxD31MSoKdO+3BYOdOxz2sslO+vFElsAWD5s218Ixcp1OnYMAA+O47e5uXlzH06D//MdbFFblBFBRERMSlXTuEsqiHPp4/b1QJbBOP9+wxwkluqlVznF/QqJExvEjkulmtsGABvPii4yYat99utN9zj/P6Jh5LQUHETWglY/FEOS3IUJiqQlyc4/yCX34xntFyc+utjsGgTh2tSCRFKDbWqCKsXWtv8/IyQsOkSaoiiNMoKLgwLy8vrFYrFosFby2e7fFsQSGnTeFESqKcFmTIb1XBajVWIMocDPKzwEN4uGMwuPnm6+u/SK6sVli0CIYNg4sX7e233WZUEdq0cV7fRFBQcGn+/v4kJyeTlJREUFCQs7sjTpZ0dXMdf62VLR4ir+Wds6sqWCzGngWZg8Gff+Z+HR8faNrUHgratIHKlYvmHkRy9OefMHAgfPONvc3LC4YOhTfegNKlndc3kasUFFxYUFAQycnJnLm67l6ZMmXw8vLSN8oexGq1YrVaSUpKyngfKDSKp8hreedjx4wvY8PD7cFg2zZjzkFu/P2NRWNswaBVK9BfK7lhrFb45BMjEGRePqtWLaOKcO+9zuubyDXcLijEx8czY8YMNm/ezLlz5yhfvjytWrVi2LBhhIaGZhx3+fJlZs+ezffff09cXBxlypShWbNmDBkyJN9bWI8aNYrVq1dn+7MBAwYwcuTIIrmnnAQHB5OUlMSVK1c4ffp0sV6rKFksFgANlyoGpUuXJliLr4sHyO9mkf375z2/ICjIcUWiFi20ia04yZkzEBkJ1z5bDBkCkydDmTLO6ZdIDtwqKMTHx9OlSxfi4uJo3bo1HTt2JCYmhq+//pqtW7fy2WefERYWhslk4plnnuHw4cM0bdqUDh06cObMGdatW8e2bdtYsGABzZo1y/N6hw8fpnLlynTv3j3Lz/JzfmF5e3sTGhpKQkICiYmJpKSkuMWE1tTUVAACtBFMkfDy8sLf35+goCCCg4MVwMQj5HezyOw+EqtUcZxfcMcdWpFInMxqhWXLjECQkGBvDwszqgjt2jmrZyK5cqugMGPGDOLi4njppZfo27dvRvuqVasYPXo0U6ZMYc6cOSxevJjDhw/Tq1cvxo4dm3Hcrl276NOnD+PHj+err77K9VppaWnExMTQrl07hgwZUmz3lBdvb28qV65MZTcaMBsdHQ2Q78qNiEhm+a0m2Pj6QteuxrNW27ZQt65WJBIXcvYs/POf8MUXju2DB8Obb2onPnFpbvXV5Pr16wkODiYiIsKhvVOnTtSsWZNt27ZhsVhYt24dXl5eDB8+3OG4li1b0rJlS44ePcrZs2dzvdbx48dJS0ujbt26RX4fIiKSs/xWE2zMZnjoIWN1yXr1FBLERVit8NlnxiSazCHhlltgwwaYNUshQVye21QU0tPTiYyMxNfXN9uhF35+fqSlpWE2m+nevTvnzp2jbDZ/Af2ubpVpW0EmJ0eOHAFQUBARuYEKWk2wKe7dmkUK5K+/jIrB5587tkdGwttva/a8uA23+Uj18fHJUkmwOX78OCdOnKBmzZr4+fnx9NNPZ3tcQkICP//8M6VLl+bmPBbFtgWFmJgYunfvzpEjRwgICKBdu3YMHz6catWqFe6GREQkizlzClZNsCnq3ZpFrtv//meEhL//treFhsJ//wsPPOC8folcB7cJCjmxWCxMnDgRi8VC165dcz327bffJikpiR49emRUFnJiCwqzZ8/mgQceoEmTJkRFRbFy5Uq2b9/O8uXLqV69+nX322QyZYzlL2lMJhNAib0/ESke331XlhEjbgaub+zQq6+m0rTpcVUVxCl8zp+n+sSJlPv2W4f2808/zV+jR2MpWxb076JLKunPLSaTicDr3N3brT9OrVYr48aNY8eOHTRs2DDHigMYD/wrV64kJCSEF198Mc/XDggIICwsjJkzZ1KnTp2M9g8++ID33nuPSZMmMXPmzCK5DxERT3bxojdvvFGdr74qX6jXOXnSj2++KU+nThfzPlikCAV9/z3VJ0zA99y5jLa0atWImzCBpLZtndgzkcLxsrrDepvZMJvNvPrqq6xcuZLQ0FCWLFmS43Cg6dOnM3v2bCpUqMAnn3zi8OBfUBaLhQceeIAzZ86wa9cuylzHmsclfVWgkn5/IlJ0vv0Wnnsu792T8+u224wvbVVVkBvi3DljydNlyxzb+/aFadOgQgXn9EsKpKQ/txTm/txq1SMbk8nE4MGDWblyJWFhYSxatCjbkJCens6YMWOYPXs2lSpV4uOPPy5USABjudJ69ephNpszdsoVEZGCuXTJWKXokUeKLiSAfa6CSLFbtcpY0ShzSKhRA775Bj76SCFBSgS3CwoXL14kIiKCzZs306BBA5YuXUqNGjWyHJeamsrzzz/PihUrCAkJYenSpdSrVy9f1zCZTOzfv5/Dhw9n+/Pk5GQA/LW1p4hIgW3cCI0bw/z59rai3BBt4kRj9SSRYpGQAM8+C08+aeyRYNO7Nxw8CB07Oq9vIkXMrYJCSkoKkZGRREVF0bJlSxYvXkylSpWyHGe1WhkxYgQbN26kTp06LFu2jLCwsHxf5++//6Zbt26MGjUqy89MJhO//vorwcHBhISEFOZ2REQ8SlISDB0K7dvDH3/Y21u0gPT0oruOqgpSbL7+2qgiLFlib6teHVavho8/hooVndc3kWLgVkFh2rRp7Nu3j6ZNmzJv3rxs90kAWLx4MevWreOWW27JcVhSbkJDQwkPD+fo0aOsXr06o91qtTJ16lQSEhLo0aMHXtrVR0QkX378EZo0gRkz7G0VK8Inn8D580V/PVUVpEhduAB9+sDjj0PmYcfPPguHDhntIiWQ20z3io+PZ8nVBF+rVi3mzZuX7XF9+vRh9uzZgLFZ2pLMqT+T7t27U6VKFQAWLlxIYmIiERERlCtXDoAJEybQq1cvRo8ezbp16wgJCeHnn3/m4MGDtGjRgkGDBhX1LYqIlDjJyTBuHLzzjrFRrc1jj8GHH8L331/fvgl50b4KUmTWrDEm1GSeTFOtGsydC506Oa9fIjeA2wSFqKgo0tLSAPj82p0OM+nQoQPnr349tW7dOtatW5fjcbagsGjRImJjY+ncuXNGUGjYsCErVqzg/fffZ+fOnWzatImQkBCGDh3KgAED8tyHQUTE0/38M0REwK+/2tvKlYP33jO+nPXyMh7k8/swX9JXJhEXc/EivPgiLFjg2N6jh1Eay2bos0hJ4zZBoUOHDhmboOUlv8fZ/PDDD9m2165dm+nTpxfotUREPF1qKkyaBG+84Tj3oEMHY3PamjWd1zeRfPnuO+jfH06ftrdVqWJsHf7UU87rl8gN5jZBQUREXN+BA0YVYf9+e1uZMvD22zBokFFFEHFZly7BiBGOS3IBdO0KM2caYUHEgygoiIhIoZnN8NZbMH48XB0lCkDbtsbIjdq1ndY1kfz5/ntj979Tp+xtlSvD7NnQpYvz+iXiRG616pGIiLiew4ehdWsYM8YeEgICjI1pN21SSBAXl5holLsefNAxJPzjH8aKRgoJ4sFUURARkeuSng7TpxsB4eo+lAC0bGksKZ/PPS5FnGfDBqOKkHljj0qVYNYsY7iRxsqJh1NFQURECuz4cWjXzhjObQsJpUoZE5i3b1dIEBd3+TI8/7wxwz5zSOjc2agidOumkCCCKgoiIlIAFoux8MuoUXDlir29SROjitC4sfP6JpIvmzZBv34QE2Nvq1jRmKzco4cCgkgmqiiIiEi+nDwJDz1kfBFrCwk+PsaGaj/9pJAgLi4pCYYMgfvvdwwJTzxhVBF69lRIELmGKgoiIpIrqxUWLoThw43VI20aNDCqCM2bO61rIvmzZQv07QsnTtjbKlQwNk575hkFBJEcqKIgIiI5ioszvnDt188eEry8YPRo2LNHIUFc3JUrRsJt184xJDz2mFFFePZZhQSRXKiiICIiWVit8OmnxjCj8+ft7bfdZlQR7rnHeX0TyZdt24wqwrFj9rby5Y2lunr3VkAQyQdVFERExEF8vLEyZM+ejiFh6FCIilJIEBdnMhnLcd17r2NIeOQRo4oQEaGQIJJPqiiIiEiGL76AyEgjLNjccouxu/L99zuvXyL5smMH9OkDR4/a28qVg3ffNaoLCggiBaKKgoiIcP489OoFTz3lGBIGDIBfflFIEBdnMhlr9rZp4xgSHnwQDh40JtkoJIgUmCoKIiIebu1a6N8f/vzT3hYSAvPnw8MPO69fIvny009GFeHwYXtbUBBMm2bsuqyAIHLdVFEQEfFQly4ZFYOOHR1DQq9eRhVBIUFcWnIyvPSSMWkmc0jo0MGoIvTvr5AgUkiqKIiIeKAffjCGbJ88aW+rWhXmzoUnn3Rev0TyZfduo4rw66/2trJl4Z13YOBABQSRIqKKgoiIB7FtTvt//+cYEp5+2vgSViFBXFpKCowZA61aOYaE9u2NMlhkpEKCSBFSRUFExENs3258CZt5xcjgYJg1C7p10/OVuLg9e4w38MGD9rYyZeCtt2DQIPDWd58iRU1/q0RESrjkZGNBmLZtHUPCY48Zz1zduyskiAtLTYVXX4W77nIMCffdBwcOwODBCgkixUQVBRGREmz3bmN/qehoe1u5csbmtNp3Slzevn1GFeHAAXtb6dLw5psKCCI3gP6GiYiUQLYvYVu1cgwJDzxgfCnbp49Cgriw1FQYPx5atnQMCW3bGr9+4QWFBJEbQBUFEZES5sAB6N0boqLsbWXKGAvCaK6nuLyoKCPJ7t9vbwsMhMmTjZn4CggiN4z+tomIlBBmM7z+OjRv7hgS7r3XCA+DBikkiAtLS4OJE403cOaQ0Lq18YYeNkwhQeQGU0VBRKQEiI425hzs3m1vCwgwvoQdOlTPV+LifvnFqCLs3WtvCwgwku+wYeDj47SuiXgyBQURETeWnm5MTH7lFWOJeZu77oKPP4a6dZ3XN5E8mc3GxOT//MeoKNjcfTcsXKg3sIiTKSiIiLip48eNL2G3bbO3lSoFEybAyJHgq094cWWHDhlv4J9/trf5+8OkSfDii6oiiLgA/TMiIuJmLBaYM8fYG+HKFXt706ZGFaFRI+f1TSRPZrMxs/6114zVjWxatjSqCPXrO61rIuJIQUFExI2cPAnPPQfr19vbfHxg7FgYM8aoKIi4rOhoo4qwa5e9zc/PKIONGKEymIiL0d9IERE3YLXCggXGiIxLl+zt4eFGFaFZM+f1TSRP6ekwbZqxuUfmyTTNmxtVhPBwp3VNRHKmdTBERFxcXBw8/rhRSbCFBG9vGD3aGN6tkCAu7cgRaNPGeMPaQkKpUsaKRjt2KCSIuDBVFEREXJTVCsuWGZvQnj9vb69Tx6gitGrlvL6J5Ck9Hd57zxgXl5xsb7/zTqOKoMk0Ii5PFQURERcUHw9dusAzzziGhKFDjb2oFBLEpR09auz0N3KkPST4+hpzEXbuVEgQcROqKIiIuJiVK41dlOPj7W233GLMUbj/fuf1SyRPFgu8/z68/LJjFaFJE6OKcMcdTuuaiBScKgoiIi7i/Hl49ln4xz8cQ8KAAcbGtQoJ4tKOHYN27YwZ95mrCK+9Bj/9pJAg4oYKVVHYu3cvCQkJeHl55Xms1WolNTWVjh07FuaSxMfHM2PGDDZv3sy5c+coX748rVq1YtiwYYSGhjoc++WXX7Jw4UJ+//13ypUrxyOPPMLQoUMpU6ZMvq5lMpmYO3cu33zzDWfPnuXmm2/mmWeeoWfPnvm6ZxGR/FqzBvr3NyYu24SEwPz58PDDzuuXSJ4sFpg1C/79bzCZ7O2NGhmTaZo2dV7fRKRQChUU1qxZw9ixY/N9/KRJkwoVFOLj4+nSpQtxcXG0bt2ajh07EhMTw9dff83WrVv57LPPCAsLA2Du3LlMmzaNunXr8uyzz3L06FEWLlxIVFQUixYtws/PL9drpaenM2zYMDZv3sx9993HQw89xJYtW5gwYQKnT5/m3//+93Xfh4iIzaVL8K9/wX//69jeqxdMnw4VKzqnXyL5cuIE9OsHmzfb23x84JVXjEnMefxbKyKurVBBwWq1Fuvx15oxYwZxcXG89NJL9O3bN6N91apVjB49milTpjBnzhxiY2N5//33adq0KYsXL6bU1R2Ipk+fzuzZs1m+fDnPPvtsrtdas2YNmzdvpl+/fhmhYNiwYfTv358FCxbw5JNPUrdu3ULdj4h4tg0bjGeskyftbVWrwocfQqdOzuuXSJ5s24OPHg1JSfZ2bewhUqIUao5CQYffFHa4zvr16wkODiYiIsKhvVOnTtSsWZNt27ZhsVhYvnw5ZrOZyMjIjJAAMGjQIMqWLcv//ve/PK+1ZMkSfH19GTRoUEZbqVKlGD58OFarlRUrVhTqXkTEcyUlGUuedujgGBK6dIFDhxQSxMX9/js88AA8/7w9JHh7G1WEPXsUEkRKELdZ9Sg9PZ3IyEh8fX3x9s6ab/z8/EhLS8NsNrN7924AWrZs6XCMv78/TZo0Ydu2bSQmJhIUFJTttVJTU/nll1+oV68e5cuXd/hZ48aNCQwMzLiGiEhBbN8OERFw/Li9LTgYZs+Gbt2c1y+RPFmtRrlr5Ei4fNne3qCBsaJRixZO65qIFA+3CQo+Pj5ZKgk2x48f58SJE9SsWRM/Pz9OnjxJ5cqVs520HBISAkBMTAyNGzfO9vViY2Mxm83UrFkz235Ur16d33///fpvBmOidHR0dKFew1WZrk5mK6n3J3I9UlK8mD69Ch9/HIzVaq+u3n9/IuPHx1GlSjr6K5M7fbY4j29sLDeNG0fZHTsy2qze3pzr14+/n38eq78/egOLuyrpny0mk4nAwMDrOtdtgkJOLBYLEydOxGKx0LVrVwAuXLjAzTffnO3xtirC5czfhlzjwoULDsdm9xoxMTGYzWZ8fd3+t1BEitkvvwTw8ss1OHHCP6OtbNl0XnnlLJ06XUSLqInLslqpsGIFVd96C59McxFSatXiz9dfJ1lLnoqUaG79lGu1Whk3bhw7duygYcOGGRUHs9mc46pGtvaUlJQcX9dsNjscm9trXG9QCAwMpH79+td1rquzJfKSen8i+ZWaamxEO2UKpKfb2x94AP77Xx9CQ2sANZzWP3ejz5Yb7NQpYxOP776zt3l5wciR+P/nP9x6nd9Qiriakv7ZUphKidsGBbPZzKuvvsrKlSsJDQ1l9uzZGQ/wAQEBpKWlZXteamoqQK4lGH9/41u/3F7Dy8vruss4IlLyRUUZcxGiouxtZcrA1KkwcCCqIojrslqNbcBffNFYv9fm9tuN9nvucV7fROSGKtSqR4Vd7vR6mUwmBg8ezMqVKwkLC2PRokVUq1Yt4+flypUjMTEx23Nt7TkNKwIyJjDnNDwpMTGR0qVLZzupWkQ8m9kMr79uzOvMHBLuvRcOHIDISIUEcWGxsfDoo/Dcc/aQ4OVlbPaxf79CgoiHKVRFITw8nA0bNuTrWKvVSsUi2Dno4sWLDBgwgKioKBo0aMD8+fOpVKmSwzFhYWHs3r2b5ORkAgICHH4WGxuLt7c3t9xyS47XCAkJoVSpUpw+fTrLz9LT0zlz5gy1a9cu9L2ISMkSHW1UETIvihYQAJMnw9ChxgqSIi7JaoVFi2DYMLh40d5+221GFaFNG+f1TUScplBB4amnnirQ8R06dCjM5UhJSSEyMpKoqChatmzJBx98QNmyZbMc16xZM3766Sd+/vln2mT6cEtJSWH//v3cdttt2Z5n4+vryx133MGBAwe4fPmyw7EHDhzAZDLRVFvSi8hV6enw3nswZgxknv50993GqpHam1Fc2p9/GuPhvvnG3ublZaTbN96A0qWd1zcRcSq3+n5r2rRp7Nu3j6ZNmzJv3rwcH/Yfe+wxfHx8mDlzZsacBIA5c+Zw+fJluuVjsfInn3yS1NRUZsyYkdGWlpbG9OnTAejSpUsh70ZESoLjx6FdO2NpeVtI8PMzqghbtyokiAuzWmHxYmM35cwhoVYt2LTJSL8KCSIezW0mM8fHx7NkyRIAatWqxbx587I9buDAgdSuXZt+/foxb948nnzySe6//36OHTvGpk2buPPOOzOWUbVZuHAhiYmJREREUK5cOcColnz++ecsXLiQo0ePEh4eztatWzl8+DD9+vWjrv71F/FoFgvMmQOjRsGVK/b2pk2NERwNGzqvbyJ5OnPGmDCzerVj+5AhRsrNZh8iEfE8bhMUoqKiMlYh+vzzz3M8LiIiAn9/f0aMGMFNN93E0qVLWbRoEVWqVKFPnz688MILWZY9XbRoEbGxsXTu3DkjKPj4+DB//nxmzJjB2rVr2bNnDzVr1mTcuHH06NGj+G5URFzeyZPQrx9knqLl4wNjxxrDj0qVcl7fRHJltcKyZUYgSEiwt4eFGXMR2rVzVs9ExAV5WQuxdNHevXtJSEjAKx9LeFitVlJTU+nYseP1Xq7E8JT1ekvq/Ynnsq0aOXw4ZF5YLTwcPv4YmjVzXt88gT5bCunsWfjnP+GLLxzbBw+GN9+EXObuiZRkJf2zpTD3V6iKwpo1axg7dmy+j580aZKCgoi4pT//NPaeWrPG3ubtbcxN+M9/jNWNRFyS1QrLl8Pzz8O5c/b2W26Bjz6C9u2d1zcRcWmFCgoFWhgTWgAAIABJREFULUY4a98FEZHrZRup8cILcP68vb1OHaOK0KqV8/omkqe//jIqBtcO2Y2MhLffhlz2FBIRKdSqR/kZclSY40VEnOmvv+Dpp+GZZxxDwtChxt5TCgni0v73P2NcXOaQEBoK69YZM/EVEkQkD24zmVlE5EZauRIGDYL4eHub5nuKW/j7b2OY0fLlju39+8PUqXB10Q4Rkby41T4KIiLFLSHBqCD84x+OIWHgQDhwQCFBXNzKlUYVIXNIuPlm+PZbmDdPIUFECkQVBRGRq9asMb50jYuzt4WEwPz58PDDzuuXSJ7OnTOWPF22zLG9Xz+YNg3Kl3dOv0TEramiICIe79IleO45ePRRx5DQuzccPKiQIC5u1SqjipA5JNSoYey2/N//KiSIyHVTRUFEPNqGDcaXridP2tuqVoUPP4ROnZzXL5E8JSQYM+uXLHFsj4iAd9+FihWd0y8RKTEKVVHQcqci4q6SkowlTzt0cAwJXbvCoUMKCeLivvrKqCJkDgk33WS0L1yokCAiRaJQFYXw8HA2bNiQr2OtVisV9cElIi5g2zbo0weOH7e3BQfD7NnQrZvTuiWSt/Pnja3BFy1ybH/2WZg+3Xgji4gUkUIFhaeeeqpAx3fo0KEwlxMRKZTkZBg71pjbmbkg+sQTMHcuVK/uvL6J5GnNGmN78D//tLdVq2a8eVUCE5FioDkKIuIRdu82hm5HR9vbypc3voTt3Ru0H6S4rAsX4F//MjbxyKxnT3j/fahUyTn9EpEST0FBREq01FSYMAGmTIH0dHv7gw8aC8LcfLPz+iaSp2+/NdbsjY21t1Wtauys3Lmz8/olIh5BQUFESqyoKKOKEBVlbytTxticduBAVRHEhV28CCNGGGk2s27dYOZMqFzZOf0SEY9SqKCwd+9eEhIS8MrHv7ZWq5XU1FQ6duxYmEuKiOTJbIY334T//AfS0uzt990HH30EtWo5r28iefr+e2Njj1On7G2VK8MHH8DTTzuvXyLicQoVFNasWcPYsWPzffykSZMUFESkWEVHG1WE3bvtbQEBxtCjIUPAW9tMiqtKTISRI41NPDL7xz+MJbmqVnVOv0TEYxUqKBR0HwXtuyAixSU9Hd57D8aMgZQUe/vddxvLytet67SuieRtwwajivDHH/a2SpVg1ixjcw+NkxMRJyjUd2v5GXJUmONFRPLj2DFjWNHIkfaQ4OcHkyfD1q0KCeLCLl+GwYONnf8yh4TOnY2d/7p1U0gQEafRZGYRcVsWizFse/RouHLF3t60qbEfVcOGzuubSJ42boR+/eD33+1tFSsak5V79FBAEBGn02hdEXFLf/wBDzwAL7xgDwm+vvDaa/DTTwoJ4sKSkowJM+3bO4aEJ54wqgg9eyokiIhLUEVBRNyK1WqsXPTii8bcT5vwcPj4Y2jWzHl9E8nTli3Qty+cOGFvq1ABZsyAZ55RQBARl6KKgoi4jT//hMceM/afsoUEb2/4979hzx6FBHFhSUkwbJgxmSZzSHjsMaOK8OyzCgki4nJUURARl2e1wtKlxmiN8+ft7XXqGFWEVq2c1zeRPG3bZlQRjh2zt5UvD9OnQ+/eCggi4rIKVVHQcqciUtz++stYRv7ZZx1DwrBhsH+/QoK4sCtX4F//gnvvdQwJjzxiVBEiIhQSRMSlFaqiEB4ezoYNG/J1rNVqpWLFioW5nIh4mM8/h0GD4O+/7W1hYbBgAbRr56xeieTDjz9Cnz7w22/2tnLl4N13jeqCAoKIuIFCBYWnnnqqQMd36NChMJcTEQ+RkGAMM1q61LE9MhLefhuCgpzTL5E8mUwwbhxMnWqMmbN58EGYPx9CQ53XNxGRAtIcBRFxKd98AwMGQFycvS0kBP77X3joIef1SyRPO3caVYQjR+xtQUEwbZqx67KqCCLiZrTqkYi4hEuXjGepxx5zDAkREXDwoEKCuLDkZGPprdatHUNChw7Gm7d/f4UEEXFLqiiIiNNt2GBsUHvypL2tWjX48ENjDyoRl7V7t5Fmo6PtbWXLwjvvwMCBCggi4tYKFRT27t1LQkICXvn4ILRaraSmptKxY8fCXFJESpCkJOOL2FmzHNu7dYOZM6FyZef0SyRPKSkwYQK8+Sakp9vb27c3xsmFhTmtayIiRaVQQWHNmjWMHTs238dPmjRJQUFEAGNp+T594Phxe1ulSjB7NnTt6rRuieRtzx7jzXvwoL2tTBl46y1jmS5vjeoVkZLhhu6joH0XRMRkgpEjjaXlM4eEJ54wnrsUEsRlpabCq6/CXXc5hoT77oMDB2DwYIUEESlRClVRyM+Qo8IcLyIly65dxnDuw4ftbeXLw/vvQ69eGs4tLmzfPqOKcOCAva10aWPokQKCiJRQ+mQTkWKXmgpjx8I99ziGhAcfNL6Y7d1bIUFcVGoqjB8PLVs6hoS2bY1fv/CCQoKIlFhuverR2bNn6dixI0OGDKFPnz4Z7e3btyc2NjbXcydPnpznhnGjRo1i9erV2f5swIABjBw5ssB9FvE0UVFGEMj8jFWmjLEflRaFEZcWFWVUEfbvt7cFBsLkycaOgAoIIlLCuW1QSEpKYsiQIVy+fDnLz3r37k1iYmKW9uTkZD766CP8/f1p1KhRntc4fPgwlStXpnv37ll+1qxZs+vruIiHMJthyhRjYZi0NHv7fffBRx9BrVrO65tIrtLS7G9es9ne3ro1LFgAdeo4r28iIjeQWwaF2NhYhgwZwqFDh7L9eebqQmaTJv1/e3ceZ2Pd/3H8PbuxDJHlbqwJY8l6kz1uuutXKdwJyRLKkrHcZG9BQt0VmexlmmgREkm32xZxy5LsM24imkJIDMes5/fH5cyZ0xlmu2auOWdez8ejx+/uc865rs/0G8d5n+/2qlJSUjRhwgRVy+CNPjExUSdPnlSbNm0UHh6e05aBAuXIEWMtwp49zlqhQsZnL76IRb528KAxivD9985aoULS1KnSsGGSn59lrQFAXvO4v64jIyPVoUMHRUdHq2nTppl+3Z49e7RkyRK1aNFCXbp0yfD5J06cUGJiomrUqJGTdoECJTnZOGeqYUPXkNC0qTF7Y9gwQgLyqaQkIww0auQaEhy/vP/8JyEBQIGTp9ujmiEqKkqhoaFasmSJHn/88Uy/bsaMGfLz88v0uQ8xMTGSRFAAMun4cWNa0QsvGGdRSVJgoDGK8O23En+UkG8dPiw1a2asuHfMkwsKkt54g19eAAVajqYe1a5dWxs3bszUc+12u+64446c3E6SNGnSJDVv3lx+fn46depUpl6zfv16HThwQE8++aTuzuTEaEdQOHnypLp166aYmBgVKlRIbdq00fDhw1W2bNns/giSJJvNpqNHj+boGvmVzWaTJK/9+eAqJUX6+OM79NZbZWSzOb97qFnzhqZP/0XVqsXr2DELG4TXSLy5SYVp7yxJSSq1eLHujIiQb5qFNLZ779Uvr72mhKpVxS8v4P28/XOLzWZTcHBwtl6bo6CQ0a5Bf9a+ffuc3E6S1KpVqyy/ZvHixfL19VW/fv0y/RpHUJgzZ44eeOAB1a9fX/v379fKlSu1fft2LVu2TOXKlctyL4A3iY3118SJd+m774qk1vz97Row4IKee+6CAgIsbA5exe/iRVUeMECSdOaDD5RcqlSOrhd44oTuGj9ewQcPptZSAgJ0ITxcF/v0kfw9cgkfAJjK698Jjxw5ou+//15///vfVbly5Uy/rlChQqpcubIiIiJcFj7PnTtXM2fO1KuvvqqIiIhs9xUcHKyaNWtm+/X5mSORe+vPB8lul957z5i2nXaDsdq1pagoHzVsWFpSacv6g5c5f1564gnpxx8lSdUHDJA2b5bKlMn6tZKTpbfeMk5YdsyRk6S//lW+kZEqU7u2snFVAB7M2z+35GSkxOuXFa5atUqS9OSTT2bpde+++67+/e9/u+2ONGDAAJUvX16bN2/WtWvXTOsT8BS//CI98oj07LPOkODrK40dK+3dayxkBkxz/rzUtq2xlZbDkSNG7fz5rF0rJkZq2VIaPdoZEgICjEXM//2vkXQBAKm8Pihs3rxZJUqUULNmzUy5nq+vr8LCwpSUlKSzZ8+ack3AE9jt0pIlxmepdeuc9erVpe3bjTOogoKs6w9eKL2Q4JCVsJCcbJzwV7++tHOns96woZFux49nqhEApMOrg8KPP/6o06dP629/+5v8s/CXgM1m0w8//KDo6Oh0H79x44YkKYhPRSggzp+X/vEPqWdP6fJlZ334cGnfPmMHScBUtwsJDpkJC8eOSa1bS6NGSTffuxUQIE2ZYoSGTBy+CQAFlVcHhf3790vK+inKFy5cUNeuXfXCCy+4PWaz2XTkyBGVLFlSoaGhpvQJ5GcrVhijCJ9/7qxVqSJt2SK9/bZUuLBlrcFbZSYkONwqLKSkSDNnSvXqSTt2OOv16xuHfEycKFbbA8DteXVQOHLzL5naWZx3WqFCBdWuXVvHjh3T6tWrU+t2u11vvvmmLl26pO7du8vHx8fUfoH85NIlqUcPYw3phQvO+oAB0v79xpkJgOmyEhIc/hwWjh+X2rSRRoxwjiL4+0uvvCLt2iXVrWt21wDglbI1KfPxxx9Xp06d1KdPH0nS7t27FRoaqrvuusvM3nLszJkzkqQyGeyMERkZqatXr6p3794KCQmRJE2ePFk9e/bU6NGjtX79eoWGhmrPnj06dOiQGjdurIEDB+Z6/4BV1q41Fiv/+quzVr68sdPR3/9uXV/wctkJCQ6OsNCjh/Tqq9LNfdElGcHggw+M0QQAQKZla0Th5MmTunTpUuq/9+rVSytXrjStKbNcvjmZulixYrd9XlRUlCIiInTlypXUWp06dbR8+XI9+OCD2r17t5YuXaq4uDgNHTpU77//vgIDA3O1d8AKf/wh9esnPfqoa0jo3Vs6eJCQgFyUk5DgcOSINGGCMyT4+RnboO7eTUgAgGzI1ohCSEiINmzYoObNm6tEiRKy2+26cOHCLRf/phUWFpadW6arc+fOtz307ZNPPsnUdTZt2pRuvWrVqpo1a1a2egM8zYYNUt++0s2BOElS2bLSggXSY49Z1xcKADNCwp+FhRnbdGVxjRoAwClbQaFDhw5avHixnnnmGUmSj4+PPv30U3366acZvtZbj8cGPFVcnDRmjDRnjmu9a1cpIkK6805r+kIBkRshQZJ8fKQKFcy9JgAUMNkKCqNGjVKlSpV05MgRJSQkaNWqVQoLC/PaE+0Ab7Vtm9SnT+qBt5KkUqWM0JDFMwqBrMutkCBJR48a187uCc4AgOwFBT8/P3Xr1i3131etWqX27dtryJAhpjUGIPfYbMbukG+/bRyk5vDYY9L8+VK5ctb1hgIiN0OCg2OBM2EBALLFlKMoN27cmLpbEID8bdcuY3Fy2iVFxYtL77xjHKjGrr/IdXkREhwICwCQbaYEBcfBY6tXr9by5csVExMjm82mEiVKqFq1aurYsaM6dOhgxq0AZFN8vDR5sjR9unEWlcPf/25se1q+vHW9oQDJy5DgQFgAgGwxJSjY7XaNGjVKX331lex2u0JCQlSxYkX98ccf2r59u3bs2KFvvvlG//rXv8y4HYAs+uEHYxThwAFnrWhR6c03jfMSGEUAAAB/ZsrJzB9//LHWrl2r++67T19++aV27dqlL7/8Utu2bdO///1vNW/eXGvXrtWKFSvMuB2ATEpMlKZMkRo3dg0J999v/PtzzxESkMfKlDG+2a9VK+/uWasWowkAkA2mBIXly5erYsWKmjdvnu655x6XxypVqqSIiAiVL18+U9unAjDHkSNS8+bSSy9JSUlGLThYmjVL2rRJqlLF2v5QgOVlWCAkAEC2mRIUTpw4oVatWqlQoULpPh4cHKzWrVvr+PHjZtwOwG0kJ0tvvCE1bCjt2eOsN21qTEEaOlTyNeVPPpADeREWCAkAkCOmfFzw9/fX9evXb/uc69evy5dPJ0Cu+t//pNatpdGjjcXLkhQYKM2YIX37rVS9urX9AS4cYeFPI9GmICQAQI6Z8sm9bt262rhxo2JjY9N9/MyZM9qwYYPuvfdeM24H4E9SUoxTlOvVk3bscNYbNpT27jWCg5+fdf0B6UpIkBYtks6cMfe6hAQAMIUpQaF///66cuWKevXqpc8//1xnzpzR77//rujoaC1dulQ9evRQXFyc+vXrZ8btAKRx6pTUvr0UHm4cpCZJ/v7SpEnSzp1SnTqWtgekb8cOqVEjacIE5/CXGQgJAGAaU7ZHbdGihSZMmKAZM2Zo/PjxLo/Z7Xb5+/tr/PjxatmypRm3AyDjROX33pNGjJDi4pz1OnWkDz4wRhOAfOfyZWncOOMI8LTHgjdoYBzyMWJE9s9YICQAgKlMCQqS1LNnT7Vp00arV69WTEyM4uLiVKRIEYWFhemxxx5ThQoVzLoVUODFxhrnH6xb56z5+hpTjF55RQoKsqw1IH12u7R8ubGa/uxZZ71wYWMP36FDjaGwzZuzdyAbIQEATGdaUJCkChUq6PnnnzfzkgDSsNulpUuNaUaXLzvr1asbowhNm1rXG3BLP/0kPf+8tHata/2RR6R335UqVXLWHAucsxIWCAkAkCvYhgjwEOfPS//4h9SzpzMk+PhIw4dL+/YREpAPJSVJb71lfJBPGxLKlZOWLZPWrHENCQ5Z2TqVkAAAuYagAHiA5cul2rWlzz931qpUMT4fvf22MXsDyFf27pXuu08aOVJKu332wIHS0aNSly63PxY8M2GBkAAAuYqgAORjly5JTz1lfKa6cMFZHzBA2r9fuv9+63oD0hUXZyxIbtJE+v57Z712bWn7dmnuXKlEicxd63ZhgZAAALmOoADkU19+aXy2+vhjZ618eenf/5bmzZOKFbOuNyBda9YYH+BnzjQO95CkQoWkqVON0NC8edavmV5YICQAQJ4wdTEzgJz74w/jC9nFi13rvXsbn78y+2UskGd++cXYtWjFCtd6+/bGCEJOT16+GRbibwaNIEICAOQJggKQj2zYIPXt63pQbdmy0oIF0mOPWdcXkK7kZGN4a9w46epVZ/3OO43FMz163H4dQlaUKaOfIiMlSdUJCQCQJ0wLCjt27NCKFSsUGxurhIQE2dMepHOTj4+PVq5cadYtAa8RF2ecgTB3rmu9a1dj98hSpazpC7ilAwek556TvvvOtf7MM9Ibb+TKL20yfxAAIE+ZEhTWr1+v4cOHK8UxJ/UWfMz6ZgnwItu2SX36SD/+6KyVKiXNmSM9+aRlbQHpu35dmjxZevNNY/tTh2rVjNOW27a1rjcAgKlMCQrz5s1TQECApk6dqvvvv1/FWGUJZMhmkyZMMNYdpB2Ae/xx4/NW2bLW9Qaka/16adAg11QbECCNHSuNH28sXAYAeA1TgsLx48f12GOP6dFHHzXjcoDX++47Y3FyTIyzVry4NHu29PTT5k3rBkxx/ryxwv6jj1zrrVoZqbZmTWv6AgDkKlO2Rw0JCVFwcLAZlwK8Wny88cVr8+auIeHBB6VDh4xTlwkJyDfsdum996SwMNeQUKKEtHChtGULIQEAvJgpQaFdu3batGmT4uPjzbgc4JV++EFq3FiaNs25xXzRosaORuvWGWckAPlGdLTUpo3Uv7/0++/OevfuxmP9+0u+HMUDAN7MlKlHI0eO1KFDh9SrVy89/fTTqlSpkgIDA9N9blhYmBm3BDxGYqI0fbqx/jPt2s82bYyzEipXtqozIB3x8UaanTZNSkhw1qtUMVbYP/SQdb0BAPKUKUGhSZMm8vHxkd1u14EDB2773KNHj5pxS8AjHD5srEXYu9dZCw42gsOQIXwhi3zmm2+kAQNc58X5+UkjR0ovvywVLmxdbwCAPGdKUOjYsSNbnwJpJCdLb70lTZzo+qVss2ZSZKRUvbplrQHuLl2SXnhBev9913qTJsbcuHr1rOkLAGApU4LC9OnTzbgM4BX+9z/jXIQdO5y1wEBpyhTji1k/P8taA1zZ7cYi5REjpN9+c9aLFZNee83YCpVfWAAosEw7mTkzbDYbuyPBa6WkGKcojxljnJHg0KiR9MEHUu3a1vUGuDlxQho82DgbIa1OnYx9ekNDrekLAJBvmBYUoqOjtW7dOl26dEnJycmypzlBKjExUZcvX9bevXu1b98+s24J5BunTkl9+0qbNztr/v7Siy9K48YZZ1IB+UJionGq8qRJ0o0bznr58lJEhHHiHwAAMikofPfdd+rXr19qQHAsbHZwrF8we8ejc+fO6eGHH1Z4eLj69Onj8thnn32miRMnpvu6evXqadmyZRle32azaf78+Vq7dq3OnTun8uXLq0ePHnrqqadYkwFJzm3mR4yQ4uKc9Tp1pKgoqUED63oD3OzcKT33nHTwoLPm4yOFh0uvvmpMOQIA4CZTgsL8+fOVnJyskSNHqkmTJho9erTuvfde9ezZU8ePH1dERIRu3LihhQsXmnE7SdK1a9cUHh6uuLSfztKIublrx7PPPqugoCCXx8qVK5fh9ZOTkzVs2DB98803uv/++/Xggw9q69atmjx5sn7++WeNGTMm5z8EPFpsrPTss8YZCA6+vtLo0dIrr0h/+rUDrPPHH8ZJf3PnGunWoX59Y7Fy48bW9QYAyLdMCQqHDh3S/fffr/79+0uS7rvvPh04cEB169ZV3bp11bRpUz366KOaM2eOXnrppRzfLzY2VuHh4Tp8+PAtnxMTE6MSJUpo1KhR2brHV199pW+++UZ9+/ZNDQXDhg1T//79tXjxYnXs2FE1atTI1rXh2ex2aelS40vYy5ed9erVjbUITZta1xvgwm6XVq40fll//dVZL1zYONhj2DBjjhwAAOkwZRf369evq1q1aqn/fs899+j48eNKunm61F133aV27dpp165dOb5XZGSkOnTooOjoaDW9zSeyY8eOqXoO9qBcunSp/P39NXDgwNRaQECAhg8fLrvdruXLl2f72vBc585JnTtLPXs6Q4KPjzR8uLRvHyEB+cjp08Z6gyeecA0JDz9sHPAxciQhAQBwW6YEhRIlSujatWup/16xYkUlJSXpxx9/TK395S9/0S+//JLje0VFRSk0NFRLlizR47dYdHf27Fldvnw529/4JyQk6ODBgwoLC1Px4sVdHqtbt66Cg4O1e/fubF0bnmv5cmPtwapVzlqVKtKWLdLbb3MWFfKJpCRp5kypVi1pzRpnvWxZ6dNPpS+/5DhwAECmmBIU6tevrw0bNujSpUuSpGrVqslut2tHmo3ko6OjVdiET1KTJk3SqlWr1LBhw1s+x7E+ITExUYMHD1azZs3UoEED9evXL8OToyVjalNSUpIqVqzo9pifn5/KlSunU6dOZftngGe5eFHq3l3q0kW6cMFZHzhQOnBAat3aut4AF99/L913n7G6Ps2XNxowQIqOlp580hgCAwAgE0wZd37mmWfUq1cvPfLII5oxY4Zat26txo0ba+bMmbpw4YIuXLigbdu26YEHHsjxvVq1apXhcxxB4ZNPPlHLli3VuXNn/fTTT9q0aZO+++47zZ0797bXuXxzTkmxW+wAUqxYMZ08eVJJSUnyz+bQvc1m09GjR7P12vzOdvMQAW/4+bZsKaqXXvqLLlxw/v+5XLlETZnyq1q0uKYzZyxsDrjJ59o1lY6IUMkPP5RPSkpqPb5qVf06aZJsDRsa04/STkHyQN703gIg//D295acnGNmSlBo1KiRZs2apTfffFMJCQmSpIkTJ6pv375atGiRJCk0NFQvvPCCGbfLUEpKikJDQzV8+HA99thjqfVdu3apT58+GjdunDZu3Oi2G5KDY21FYGBguo876vHx8dkOCsjfrl711bRpZbVqVQmXeseOlzV27DmFhKTc4pVA3iq6ZYvKTZmigDQhICUwUBcGDtTFvn2NY8EBAMgG0z7ltm/fXu3bt089P6FGjRpav369du7cqaCgIDVq1CjPTmUeOHCgyyJkhyZNmqhDhw5atWqVdu3adctRBUeASExMTPfxhIQE+fj45OjnCQ4OVs2aNbP9+vzMkcg99ef7z3+kfv3kMlpQtqy0cKHUoUMJSSVu+Vogz/z6qzR0qLF4Jq2//U2+8+apTLVqKmNNZ7nG099bAORP3v7ekpORElPWKKTlOIjs+vXrKlKkiNq1a6eWLVvmWUjISK1atSRJP//88y2f41jAfKszGq5evarChQvL19f0/3ywUFycNHiw9Pe/u4aEbt2MTWI6dLCuNyBVSopxHkJYmGtIKFXK2J93wwYpzS50AABkl2mfdO12uz7++GN16dJF9957r/76179KkpYsWaJx48bpQtpVoLns8OHDt9yVKD4+XpJuOe1IMqZJBQQEpBsmkpOTdfbsWVWpUsWcZpEvbN0q1atnfP5yKFVKWrZM+vhj438Dljt0SGrZ0ki0V6446717G4uVe/VisTIAwDSmTD1KSkrS4MGDtW3bNvn7+6tIkSL6448/JBnf3H/++efau3evPvnkE5UsWdKMW97W888/r3Pnzmn79u1u99u7d68kqU6dOrd8vb+/v+rVq6cDBw4oLi5ORYsWTX3swIEDstlsatCgQe40jzxls0kTJhi7SaY9sLZjR2nePGPKEWA5m02aMkV64w1j+1OHatWMX9S//c263gAAXsuUEYX3339fW7duVZ8+fbRr1y716NEj9bFRo0YpPDxcp0+f1vz58824XYYeeughpaSk6O23305dMyFJ69at05YtW9S4ceMMD2Pr2LGjEhISNHv27NRaYmKiZs2aJUnq0qVL7jSPPPPdd1KDBsYZCI5fk+LFpago4zBbQgLyhQ0bpHvvlaZNc4aEgABp4kRjf15CAgAgl5gyouA412DMmDGSnOsUJOPb+eeff167d+/Wli1bNG7cODNueVuDBw/W1q1btWzZMsXExKhRo0Y6efKktmzZotKlS2vatGkuz4+MjNTVq1fVu3cPclc3AAAgAElEQVRvhYSESJI6d+6sFStWKDIyUseOHVPt2rW1bds2RUdHq2/fvtk+zA3Wi4+XJk2SZswwpns7PPSQtGiRFBpqXW9Aqt9+k/75T2nJEtd6ixbSggXGgWoAAOQiU0YUzpw5k7om4Vbq1Kmjs2fPmnG7DIWEhOiTTz5R79699dtvv+nDDz/UoUOH9MQTT2jlypWqUKGCy/OjoqIUERGhK2nm/Pr5+WnRokXq06ePTpw4oaioKCUnJ+ull17Ks21eYb59+6TGjY0vZx0hoWhR43PXV18REpAP2O3S4sXGYuW0IaFECeMXdetWQgIAIE+YMqIQEhKi2NjY2z7n9OnTtzzALLs6d+6szp0737Kn8ePHa/z48RleZ9OmTenWixYtqnHjxuXJKAhyV2KiEQ6mTHGd4t22rfT++1Llypa1BjjFxBinKH/zjWu9Wzdjjly5ctb0BQAokEwZUWjWrJn+85//3HKf1h9++EGbNm1S06ZNzbgdkCWHD0vNmkkvv+wMCcHB0jvvGNO/CQmwnGM+XN26riGhcmVjqOvjjwkJAIA8Z8qIwtChQ7VlyxZ1795dTzzxhH766SdJ0ueff66DBw9q+fLlCgwM1KBBg8y4HZApycnSW28Zaz5vHhguyQgNH3zAVvPIJ7ZuNUYRoqOdNT8/Y33Cyy9LRYpY1xsAoEAzJShUrFhRH3zwgcaOHaslaebUjh8/Xna7XeXLl9eMGTNUtWpVM24HZOh//5P69JF27HDWAgOlV181Pn/5+VnWGmC4dEkaPVp67z3XeuPGxlqE+vWt6QsAgJtMCQqSVLt2ba1Zs0b79+/XoUOHUk8vrlGjhho3bswpxsgTKSnSu+9KY8YYW887NGpkjCLUrm1db4AkY7Hyxx9LI0ZI588760WLSq+9ZhymRpIFAOQDpgUFh3r16qlevXpmXxbI0KlTUt++0ubNzpq/v/Tii9K4ccbW84ClfvzRCAL//rdrvWNHafZsqXx5a/oCACAdpgWF+Ph47dq1S7GxsUpIOyH8T3r16mXWLQFJxhe0ixYZU4ri4pz1OnWMw9M4RBuWS0w0FsxMmuQ61BUaKkVEGEEBAIB8xpSgEB0drYEDB+rcuXOS5HIaclo+Pj4EBZgqNlbq31/6+mtnzdfXmHr08stSUJB1vQGSjCPAn3vOOEXZwcdHGjLEWDRz85BHAADyG1OCwmuvvaazZ8+qU6dOqlevnoL4dIZcZrcbZ1ENHSpdvuys16hhrEW47z7regMkSVeuSOPHS3PmGL+wDvXqGYuVmzSxrjcAADLBlKBw+PBh/d///Z+mTZtmxuWA2zp3zthN8osvnDUfH2n4cGnqVOOMBMAydrv0+edSeLj0yy/OenCwMfVo+HAWzAAAPIIpQaFw4cIqXbq0GZcCbuuzz6RBg6SLF521KlWkyEipdWvL2gIMZ84YU4pWr3atP/SQMbJQpYo1fQEAkA2m7Fn62GOPaePGjbKlXaQHmOjiRal7d+nJJ11DwqBBxtRvQgIslZwszZol1arlGhLKlpU++cQ4XZmQAADwMKaMKAwbNkwnTpzQY489pq5duyo0NFSBgYHpPrddu3Zm3BIFyJo1xlrQs2edtfLlpffflx54wLq+AEnSvn3GL+iePa71556Tpk+X7rjDmr4AAMghU4LCuXPndPr0aZ05c0Zvvvlmus+x2+3y8fHR0aNHzbglCoA//jCmc0dGutafeUZ6+22peHFL2gIMcXHG1lozZxon/TnUrGksVm7Z0rreAAAwgSlB4ZVXXtGPP/6oBg0aqEGDBipcuLAZl0UBtn691K+f9PPPzlq5ctLChdKjj1rXFyBJWrvWODjt9GlnLShImjBBGj2afXkBAF7BlKCwb98+tWzZUosWLTLjcijA4uKkF16Q5s1zrXfvbhxcW6qUNX0BkqRff5WGDTNW1afVtq3xS1u9ujV9AQCQC0wJCkFBQapRo4YZl0IB9s03xrSikyedtTvvlObOlZ54wrq+AKWkGNOJxo415sQ5lColvfmm1KuXsUcvAABexJRdj9q1a6etW7cqMTHRjMvBgyUlSVFRdygq6g4lJWXuNTabNGKE8aVs2pDQsaN06BAhARY7fFhq1crYYittSOjVS4qOlnr3JiQAALySKSMKo0aNUq9evdSzZ09169ZNlSpVUvAtTr0KCwsz45bIpz76SJo+vZwkY01nr163f/533xmfs2JinLUSJYxpRj168PkLFrLZpFdflV5/XS6p9557jGlG7OAGAPBypgSFFi1aSJKSk5O1f//+2z6XXY+8V1KSNGWK89+nTJGeekryT+e3LD7eOKR2xgzXDWMeekhatEgKDc39foFb2rhRGjhQOn7cWfP3l8aMMRYsc/w3AKAAMCUodOjQQT589VvgffSR6+eq48eN2p9HFfbtM0YRDh501ooWNbY87dePUQRY6LffpJEjpQ8/dK03b26sUahd25q+AACwgClBYfr06WZcBh7sz6MJDmlHFRITpWnTjFramRxt2xqHp1WunGftAq7sdumDD6RRo1yP/i5e3Bj2evZZydeUJV0AAHgMU4IC8OfRBAfHqEKjRsYowt69zseCg43PYM8/z2cwWOjYMWOa0ebNrvUnnzQOU/vLX6zpCwAAixEUkGO3Gk1w+Oc/pStXjBEFh2bNjC9wq1XL/f6AdCUkGEl16lRj0YxDpUrSnDnSww9b1xsAAPkA3+Mix241muBw8aIzJAQGGpvIbNtGSICFvv1Wql9feuklZ0jw8zPWJxw+TEgAAECMKCCHMhpNSKthQykqivWgsNDvvxs7Fy1c6Fr/61+NxcoNGljTFwAA+RAjCsiRjEYT0nr+eUICLGK3S598YhzukTYkFC0qzZol7dxJSAAA4E8ICsi2rIwmSMaOR5k9rRkwzcmTxlSi7t2lc+ec9ccfl44ckYYONaYdAQAAFwQFZFtWRhMk5w5IQJ5ITJTeeMMYxvr6a2f9rruklSulVaukChWs6w8AgHyOoIBsyepogsOfz1AAcsWuXVLjxtLo0ZLNZtR8fKQhQ6SjR6VOnaztDwAAD0BQQLZkdTTBgVEF5KorV6TwcKlpU2n/fme9bl3pv/+VZs+WQkKs6w8AAA9CUECWZXc0wYFRBeSKzz+XatWSIiKMxcuS81S/PXuk++6ztj8AADwMQQFZlt3RBAdGFWCqn3+WOnaUOneWYmOd9QcflA4dMqYfBQRY1x8AAB6KoIAsyeloggOjCsix5GTpnXeMLU+/+MJZL1PGSKLr1kl3321dfwAAeDiCArIkp6MJDowqIEd++EFq1kwaNkyKi3PW+/c3Fit3724sXgYAANnm0UHh3LlzatSokSIjI90ei4uL0+uvv64HHnhAderU0X333afBgwfr6NGjmb7+Cy+8oBo1aqT7z7/+9S8TfxLPYNZoggOjCsiya9ekF14wTlLevdtZDwuTtm41DlMrWdK6/gAA8CL+VjeQXdeuXVN4eLji0n6beJPNZlOPHj0UHR2tBg0aqH379jp79qzWr1+vb7/9VosXL1ajRo0yvEd0dLTuvPNOdevWze2xzLze25g1muDgGFXo1cu8a8KLrVsnDRok/fSTsxYYKE2YII0ZIwUFWdcbAABeyCODQmxsrMLDw3X48OF0H//www8VHR2tnj17auLEian1Xbt2qU+fPnrllVe0Zs2a294jMTFRJ0+eVJs2bRQeHm5q/57I7NEEhylTpKeekvw98jcReeLsWWn4cOnTT13rbdpI8+ZJNWpY0hYAAN7O46YeRUZGqkOHDoqOjlbTpk3Tfc769evl4+Oj4cOHu9SbNGmiJk2a6NixYzp37txt73PixAklJiaqBh9CJJk/muDAWgXcUkqKtGCBsVg5bUgoWVJavFjatImQAABALvK473GjoqIUGhqqSZMm6dSpU9q5c6fbc7p166aLFy+qaNGibo8FBgZKMqYu3U5MTIwkERRu6tUr81OEHOtAatasmYsdwasdOSI995y0fbtrvWdP6c03pdKlrekLAIACxOOCwqRJk9S8eXP5+fnp1KlT6T7niSeeSLd+6dIl7dmzR4ULF1b58uVvex9HUDh58qS6deummJgYFSpUSG3atNHw4cNVtmzZHP0cANJx44Y0dapxSFpiorNetaoxzah9e+t6AwCggPG4oNCqVatsv/aNN97QtWvX1L1799SRhVtxBIU5c+bogQceUP369bV//36tXLlS27dv17Jly1SuXLls92Kz2bK0A5MnsdlskuS1Px9yR+GdO1Vu0iQFpVmsbPf318W+fXVh4EDZCxUytj5FgcV7C4Dc4O3vLTabTcHBwdl6rccFheyaM2eOVq5cqdDQUI0YMSLD5xcqVEiVK1dWRESEqlWrllqfO3euZs6cqVdffVURERG52TJQIPj9/rvKvP66SqQ9NE3S9fr1dXbSJMWn+fMHAADyToEICrNmzdKcOXNUokQJzZ8/X8WLF8/wNe+++2669QEDBmj58uXavHmzrl27piJFimSrp+DgYK+dw88aBWSK3S5FRUkjR0oXLzrrxYtL06er8HPP6W5fj9tvAbmI9xYAucHb31tyMlLi1X8LJycna8KECZozZ45KlSqlDz74wGV0IDt8fX0VFhampKQknT171qROgQLmf/8z1hv06eMaErp0MaYXDRwoERIAALCU144oJCQkaOjQodq8ebNCQ0P1/vvvq3Llypl6rc1mS128HBYW5vb4jRs3JElBHPAEZE1CgvT669Krr0rx8c56xYrSnDnSI49Y1xsAAHDhlUHBbrdr5MiR2rx5s6pVq6b33nsvS7sUXbhwQV27dlX16tXdDmaz2Ww6cuSISpYsqdDQULNbB7zX9u3GlqdHjjhrvr7GYWqTJknpbGcMAACs45Vj+x9++KHWr1+vSpUqKSoqKstbmVaoUEG1a9fWsWPHtHr16tS63W7Xm2++qUuXLql79+7y8fExu3XA+/z+uzRggNSypWtIaNRI2r3bOBeBkAAAQL7jdSMKCQkJmjNnjiTjsLSlS5em+7xu3bqp9M1DmyIjI3X16lX17t1bISEhkqTJkyerZ8+eGj16tNavX6/Q0FDt2bNHhw4dUuPGjTVw4MC8+YEAT2W3S8uWScOGSWlPQi9SxJh6NGSI5O91b0EAAHgNr/tb+sSJE/r9998lSevXr9f69evTfV779u1Tg0JUVJRiY2PVqVOn1KBQp04dLV++XO+884527typLVu2KDQ0VEOHDtWzzz6b4TkMQIF26pQ0eLC0bp1rvUMHKSLCWJMAAADyNR+73W63uomCpqBsw+WtPx9uIylJmjlTevll6fp1Z/2uu6TZs6VOnSSm7CGbeG8BkBu8/b0lJz+f140oALDI7t3GYuUffnDWfHyMkYWpU43zEQAAgMcgKADImatXpYkTjSlFKSnO+r33SgsWSE2bWtcbAADINq/c9QhAHvniC6lWLemdd5whoVAhafp0ae9eQgIAAB6MEQUAWffzz1J4uLRqlWv973+X5s6V7r7bmr4AAIBpGFEAkHnJycai5Fq1XENC6dLS0qXS118TEgAA8BKMKADInP37jcXKu3a51vv1k15/XSpZ0pq+AABArmBEAcDtXb8ujRljnKScNiTUqCF98420aBEhAQAAL8SIAoBb+/pradAg4wA1h8BAafx4aexYKSjIstYAAEDuIigAcHfunDR8uPTJJ671+++X5s2TwsKs6QsAAOQZph4BcEpJkRYuNIJA2pBwxx3Se+9JmzcTEgAAKCAYUQBgOHrUWKz87beu9R49pLfeksqUsaYvAABgCUYUgILuxg3ppZekevVcQ8Ldd0vr10tLlhASAAAogBhRAAqyzZulgQOlY8ecNX9/adQo6cUXpcKFresNAABYiqAAFEQXLxphIDLStd60qbRggXTvvZa0BQAA8g+mHgEFid0uffihsSA5bUgICZHefVfavp2QAAAAJDGiABQcx48b04w2bnStP/GENGuWdNdd1vQFAADyJUYUAG+XkCC99poxUpA2JFSoIK1ZI332GSEBAAC4YUQB8GY7dhhbnh4+7Kz5+krDhkmTJ0tFi1rXGwAAyNcICoA3unxZGjfOOEU5rYYNjcXKjRpZ0xcAAPAYTD0CvIndLi1bJtWs6RoSihQxDk377jtCAgAAyBRGFABv8dNP0uDB0ldfudYffVSKiJAqVbKmLwAA4JEYUQA8XVKS9OabUq1ariHhL38xFiqvXk1IAAAAWcaIAuDJ9uwxFivv2+es+fgY26BOmyYVL25dbwAAwKMRFABPdPWq9OKL0uzZUkqKs16njrFYuVkz63oDAABegaAAeJrVq6UhQ6QzZ5y1QoWkl16SRo2SAgKs6w0AAHgNggLgKWJjpaFDpZUrXevt2xs7HFWtak1fAADAK7GYGcjvkpOld981tjxNGxJKl5aWLJHWryckAAAA0zGiAORnBw4Yi5W/+8613rev9PrrUqlS1vQFAAC8HiMKQH50/bo0dqxxOFrakFCjhrRli/Tee4QEAACQqxhRAPKb9euN7U1PnnTWAgOlceOMf4KCrOsNAAAUGAQFIL84f14aMUL66CPXeuvW0vz5UliYNX0BAIACialHgNVSUoypRGFhriHhjjukRYukzZsJCQAAIM8xogBY6ehRacAAads21/pTT0lvvy2VKWNNXwAAoMBjRAGwwo0b0ssvS/XquYaEKlWkr7+Wli4lJAAAAEt5dFA4d+6cGjVqpMjIyHQfX7VqlTp27Kj69eurdevWmjZtmq5du5bp69tsNs2cOVMPPPCA6tatq4cfflhLly6V3W436SdAgbRlixEQJk+WEhONmp+fNGaMdOiQ9OCDlrYHAAAgeXBQuHbtmsLDwxUXF5fu4/Pnz9eYMWOUkpKip59+WmFhYYqMjFS/fv2UkJCQ4fWTk5M1bNgwzZ07V1WqVFGvXr3k7++vyZMn6/XXXzf7x0FBcPGicf5B27bSsWPO+n33Sd9/L02fLhUubF1/AAAAaXhkUIiNjVXPnj21f//+Wz7+zjvvqEGDBlqxYoVGjRqlBQsWaPDgwdq3b5+WLVuW4T2++uorffPNN+rbt68WLFigUaNGacWKFWratKkWL16smJgYs38seCu73ThBuWZNafFiZ71YMSkiQtq+Xapb17r+AAAA0uFxQSEyMlIdOnRQdHS0mjZtmu5zli1bpqSkJA0YMEABAQGp9YEDB6po0aL67LPPMrzP0qVL5e/vr4EDB6bWAgICNHz4cNntdi1fvjznPwy834kTxlSinj2l335z1v/xD2Mh8/PPG9OOAAAA8hmPCwpRUVEKDQ3VkiVL9Pjjj6f7nN27d0uSmjRp4lIPCgpS/fr1FR0dratXr97yHgkJCTp48KDCwsJUvHhxl8fq1q2r4ODg1HsA6UpMlKZNk+rUkf7zH2e9QgVp9Wpp+XIpNNS6/gAAADLgcdujTpo0Sc2bN5efn59OnTqV7nNOnz6tO++8U0WKFHF7LPTmh7OTJ0+q7i2me8TGxiopKUkVK1Z0e8zPz0/lypW75b0B/fe/0nPPGQuTHXx9paFDjQXMxYpZ1xsAAEAmeVxQaNWqVYbPuXz5ssqXL5/uY8Vufki71SJox+vTPje9a5w8eVJJSUny98/ef0KbzaajR49m67X5nc1mkySv/fluxffqVZV5+22V+PRT+aTZGetGzZr6ddIk3ahTR/r5Zws7BDxbQX1vAZC7vP29xWazKTg4OFuv9bigkBlJSUkKDAxM9zFHPT4+/ravT/vc210ju0EBXsRuV7H161X2tdcUkGYdQkpwsH4LD9elp5+W+D0BAAAexis/vRQqVEiJjv3p/8SxNertklVQUJAk3fYaPj4+2U5njvvXrFkz26/PzxyJ3Ft/PhenTxsLkr/80rX+8MPyffddla1cWWWt6QzwOgXqvQVAnvH295acjJR43GLmzAgJCbnlYmVH/VbTiiSlLmC+1fSkq1evqnDhwvL19cr/fMiMpCTprbekWrVcQ0K5ctKyZUatcmXL2gMAAMgpr/ykW7lyZV28eFE3btxweyw2Nla+vr6qVKnSLV8fGhqqgIAA/ZzOfPLk5GSdPXtWVapUMbVneJC9e41D0kaOlNKe9D1woLHlaZcuko+Pdf0BAACYwCuDQqNGjZSSkqI9e/a41OPj4/XDDz/onnvuUdGiRW/5en9/f9WrV09HjhxxG1U4cOCAbDabGjRokCu9Ix+Li5NGjJCaNDFOUnaoXds4NG3uXKlECev6AwAAMJFXBoVHH31Ufn5+ioiISF2TIEnz5s1TXFycunbtmuE1OnbsqISEBM2ePTu1lpiYqFmzZkmSunTpYn7jyL++/NIIBDNnSikpRi0oSJo61QgNzZtb2x8AAIDJvHIxc9WqVdW3b18tXLhQHTt2VNu2bXX8+HFt2bJFDRs21JNPPuny/MjISF29elW9e/dWSEiIJKlz585asWKFIiMjdezYMdWuXVvbtm1TdHS0+vbtqxo1aljxoyGv/fKLNGyYcUBaWu3aSfPmSffcY01fAAAAucwrg4IkjRw5Un/5y1/00UcfKSoqSqVLl1afPn00ZMgQt21Po6KiFBsbq06dOqUGBT8/Py1atEizZ8/WunXrtHfvXlWsWFEvvfSSunfvbsWPhLyUkmIEgXHjpCtXnPU77zQWMT/9NOsQAACAV/Ox29OcDIU8UVC24fLYn+/gQeNk5Z07XevPPCO98YZUqpQ1fQEFnMe/twDIl7z9vSUnP59XrlEAssVmM0YQGjZ0DQnVqkmbNknvv09IAAAABYbXTj0CsuQ//zG2N/3xR2ctIEAaO1YaP14qVMi63gAAACxAUEDBdv689M9/SkuXutZbtpTmzzcOVAMAACiAmHqEgsluN6YS1azpGhJKlJAWLJC++YaQAAAACjRGFFDwxMRIAwYYYSCt7t2lt9+Wypa1pi8AAIB8hBEFFBzx8dKkSVLduq4hoUoVad066aOPCAkAAAA3MaKAgmHrVmMUITraWfPzk0aOlF5+WSpc2LreAAAA8iGCArzbpUvS6NHSe++51ps0MdYi1KtnTV8AAAD5HFOP4J3sdmMqUc2ariGhWDFp9mxpxw5CAgAAwG0wogDv8+OP0qBB0vr1rvVOnaR33pHKl7emLwAAAA/CiAK8R2KiNGOGVKeOa0goX15atUpauZKQAAAAkEmMKMA77NwpPfecdPCgs+bjI4WHS6++akw5AgAAQKYRFODZ/vhDGj9emjvXWJfgUL++sVi5cWPregMAAPBgTD2CZ7LbpRUrjNOT58xxhoTChaU33pB27yYkAAAA5AAjCvA8Z85Izz8vrVnjWv+//zNCQ+XKlrQFAADgTQgKMJ3fxYu5c+HkZGNr04kTpWvXnPWyZY3djLp0MdYlAAAAIMcICjDX+fOq1KeP8b937JDKlDHnut9/byxW3rvXtT5ggDR9ulSihDn3AQAAgCTWKMBM589Lbdsq6MQJBZ04IbVta9RyIi5OGjnSWG+QNiTUqiV9+600bx4hAQAAIBcQFGCOmyFBR444a0eO5CwsrF0r1a4tvfWWlJJi1IKCjO1O9+2TWrTIed8AAABIF0EBOZdeSHDITlj49VfpySelRx+VTp921v/2N+OchAkTpMDAnPcNAACAWyIoIGduFxIcMhsWUlKMqUQ1a0qffeaslyolffCBtGGDVK2aOX0DAADgtggKyL7MhASHjMLCoUNSq1bSoEHGIWoOvXtL0dFSr17saAQAAJCHCArInqyEBIf0woLNZpys3KCBsUuSQ7Vq0saNUmSkdOedprUNAACAzCEoIOuyExIc0oaFDRuke++Vpk2TkpKMxwMCjHMSDhww1iQAAADAEpyjgKzJSUhwOHJEql7ddYqRZOxitGCBsfUpAAAALMWIAjLPjJDgkDYkFC8uzZ8vbd1KSAAAAMgnGFFA5pgZEtIKCTEOTqtTx9zrAgAAIEcYUUDGciskSNKVK1LXrjk/wRkAAACmIijg9nIzJDjk9ARnAAAAmI6ggFvLi5DgQFgAAADIVwgKSF9ehgQHwgIAAEC+QVAAAAAA4IaggPSVKSNt3py325XWqmXcs0yZvLsnAAAA0kVQwK3lZVggJAAAAOQrXnuOQo0aNTJ8TlRUlO67777bPuepp57S3r17033slVdeUffu3bPVn8dwhIXcXK9ASAAAAMh3vDYoDBkyJN36xYsX9fHHH6tUqVK6++67M7xOTEyMqlSpokceecTtsToF5ZCw3AwLhAQAAIB8yWuDQnh4eLr1QYMGycfHR2+88YZKly5922v8/PPPiouL0xNPPHHL6xUYuREWCAkAAAD5VoFao7B69Wpt2rRJXbp0UYsWLTJ8fkxMjKTMTWMqEMxcs0BIAAAAyNcKTFCIj4/X22+/rWLFiumf//xnpl5DUEiHGWGBkAAAAJDvFZig8NFHH+mXX35R//79dccdd2TqNTExMfLx8dHevXvVqVMn1a9fX61bt9bUqVN19erVXO44H8tJWCAkAAAAeAQfu91ut7qJ3JacnKz27dvrjz/+0JYtWxQSEpKp1z300EM6efKkihUrpgcffFCFCxfW7t27dfToUVWvXl0ff/yxihYtmuV+jh49KpvNpuDg4Cy/Nj/xu3hRlfr0UdCJE5l6fnzVqvopMlLJpUrlcmcAvJHNZpMkj3/vBJC/ePt7i+MzZ82aNbP82gIxorBp0yb98ssvevLJJzMdElJSUhQSEqKaNWtq7dq1mjp1qiZMmKCVK1eqa9euOnbsmGbPnp3LnedvyaVK6afISMVXrZrhcwkJAAAAnqVAjCg8//zz2rBhg9atW5epLVEzcu3aNbVo0UJFixbVt99+m+XXHz16VJKylezypfPnb78bEtONAJjA6947AeQL3v7ekpOfz+tHFOLj47Vjxw5Vr17dlJAgSUWKFFHlypX122+/KT4+3pRrev57UrYAAAl4SURBVLTbrVkgJAAAAHgkrw8Ku3bt0vXr1/Xggw9m6XVXrlzR999/r5MnT6b7+I0bN+Tr6yt/f689iiJr0gsLhAQAAACP5fVBYf/+/ZKkRo0aZel1hw8fVvfu3TVjxgy3x86fP6+ff/5ZNWvWlJ+fnyl9eoWbYSG+alVj3QIhAQAAwGN5/dfhR27Om69du3aWXteoUSOVLl1aW7du1e7du9W4cWNJUkJCgqZMmaLExET16NHD9H49Xpky+ikyUpJUnZAAAADgsbw+KJw5c0aFChXKcLcjxw5G4eHhkqTAwEBNmTJFQ4YM0TPPPKOHHnpIJUqU0I4dO3TixAk98sgj6ty5c67374nY2QgAAMDzeX1Q+P3331WsWLEMnxcRESHJGRQkqW3btlq6dKnmzJmjLVu2KD4+XlWqVNGLL76op556Sj4+PrnWNwAAAGAlrw8Kmd2+NCYmJt16/fr1tWDBAjNbAgAAAPI9r1/MDAAAACDrCAoAAAAA3BAUAAAAALghKAAAAABwQ1AAAAAA4IagAAAAAMANQQEAAACAG4ICAAAAADcEBQAAAABuCAoAAAAA3BAUAAAAALghKAAAAABwQ1AAAAAA4IagAAAAAMANQQEAAACAG4ICAAAAADcEBQAAAABufOx2u93qJgqaH374QSkpKQoODra6lVxhs9kkyWt/PgDW4L0FQG7w9vcWm80mX19f1a9fP8uv9c+FfpABPz8/q1vIVd76Bw2AtXhvAZAbvP29xdfXN9ufPRlRAAAAAOCGNQoAAAAA3BAUAAAAALghKAAAAABwQ1AAAAAA4IagAAAAAMANQQEAAACAG4ICAAAAADcEBQAAAABuCAoAAAAA3BAUAAAAALghKAAAAABwQ1AAAAAA4Mbf6gbgffr3769t27apdevWWrhwodXtAPBQs2fPVkREhFu9SJEiqlixoh555BH16dNHAQEBFnQHwNMdPnxYn376qXbt2qVff/1Vfn5+ql69ujp06KCuXbvK35+PyT52u91udRPwHr/99pvuv/9+BQYGKj4+Xps3b1a5cuWsbguAB3IEhXbt2qlmzZqSpOTkZMXFxWnPnj06evSo2rdvr3fffdfiTgF4kpSUFM2ePVtz585VQECAWrdurYoVK+rq1av69ttv9euvv6pJkyZauHChChUqZHW7liIqwVRr1qxRcnKy+vfvr9mzZ2v58uUaMmSI1W0B8GDt27dX586dXWp2u12DBg3Shg0b9N///lfNmjWzqDsAnmbevHmaM2eO6tevr3feeUdly5ZNfSwhIUHjx4/XmjVrNHbsWM2cOdPCTq3HGgWYatWqVSpevLj69++vYsWKaeXKlWLQCoDZfHx8UsPD7t27Le4GgKc4efKk5syZo5IlS2rhwoUuIUGSAgMDNW3aNIWGhurrr7/WiRMnLOo0fyAowDTR0dGKiYlRs2bNVKhQIbVv316xsbHavn271a0B8EJ+fn6SjL/YASAzVq1apcTERPXo0UMhISHpPicgIEAvvviiXnvtNd1xxx153GH+QlCAaVatWiVJevjhh13+72effWZZTwC8k91u1+effy4/Pz+1b9/e6nYAeIht27ZJklq1anXb57Vt21adO3dWyZIl86KtfIs1CjBFcnKyvvzySxUpUkRt2rSRJDVv3lylSpXSxo0bdenSpQL/hw1A9mzYsEGxsbGSjIBw7do17dq1S//73//04osv6p577rG4QwCe4uzZs5KkypUrW9uIhyAowBTbt2/Xb7/9po4dOyooKEiS5O/vr4ceekhLly7VF198oWeeecbiLgF4oo0bN2rjxo1u9eLFi+vKlStKTk5OnYYEALdz5coVScY2y8gYU49gii+++EKS9Mgjj7jUO3ToIElavnx5nvcEwDtMmzZNMTExqf/s27dPK1asUP369fXWW29p4sSJVrcIwEOUKFFCkjMw4PYICsixuLg4bdiwQZL07LPPqkaNGqn/dOvWTZJ0/Phxff/991a2CcBLFC5cWHXq1FFERITKli2rlStX6scff7S6LQAeoEKFCpKkn3766bbPu3r1qs6fP58XLeVrBAXk2Ndff60bN27o3nvvVdeuXd3+adKkiSQWNQMwV2BgoBo0aCBJiomJsbgbAJ7AsYg5ox0ZP/30U7Vq1arAn6PAGgXkmGPa0dixY/XXv/7V7fFffvlF7dq109dff60JEyaoaNGied0iAC/lmD5QrFgxizsB4Ak6dOigOXPmaMmSJerdu3e67x02my31y80WLVrkdYv5CiMKyJHY2Fjt3r1boaGhatSoUbrPueuuu9S0aVNdv35da9euzeMOAXir/fv3a9euXSpevHi6X1IAwJ9VqFBBffr00e+//67+/fu7TS+6evWqRo0apVOnTqlt27Zq3LixRZ3mD4woIEe++OIL2e12dejQQT4+Prd8XufOnbVjxw599tln6tq1ax52CMDTpd0eVTK2Yz5+/Li2bNmi5ORkjR8/XoUKFbKwQwCeZMSIEbp48aJWrlypdu3aqU2bNqpYsaLOnTun7du369KlS2rYsKFef/11q1u1nI/dbrdb3QQ814MPPqhTp07pq6++UtWqVW/5vBs3bqhly5a6evWqVq9erRo1auRhlwA80ezZsxUREeFWDwgIUMmSJVWvXj317NkzdR0UAGTFt99+q08++UQxMTE6d+6c/P39VaNGDT3++OPq0qUL2y6LoAAAAAAgHaxRAAAAAOCGoAAAAADADUEBAAAAgBuCAgAAAAA3BAUAAAAAbggKAAAAANwQFAAAAAC4ISgAAAAAcENQAAAAAOCGoAAAAADADUEBAAAAgBuCAgAAAAA3BAUAAAAAbvytbgAAkL9t2LBBO3bsyPB5zZs3V/v27XP8Ok99LQB4G4ICACBDL730UobP2bBhg2mv89TXAoA3YeoRAAAAADcEBQAAAABuCAoAAAAA3BAUAAAAALghKAAAAABwQ1AAAAAA4IagAAAAAMANQQEAAACAG4ICAAAAADcEBQAAAABu/K1uAACQv507d06TJ0/O8HlVq1Y15XWe+loA8DY+drvdbnUTAAAAAPIXph4BAAAAcENQAAAAAOCGoAAAAADADUEBAAAAgBuCAgAAAAA3BAUAAAAAbggKAAAAANwQFAAAAAC4ISgAAAAAcENQAAAAAOCGoAAAAADADUEBAAAAgBuCAgAAAAA3BAUAAAAAbggKAAAAANwQFAAAAAC4ISgAAAAAcPP/xylNSG5s0KEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 261,
       "width": 389
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 아쉽지만 한글은 지원하지 않는 statmodels의 interaction_plot\n",
    "\n",
    "fig = interaction_plot(data.디자인시안, \n",
    "                       data.성별, \n",
    "                       data.체류시간, \n",
    "                       colors=['red','blue'], \n",
    "                       markers=['D','^'], \n",
    "                       ms=10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "위 인터랙션 플롯의 해석은 다음과 같다\n",
    "\n",
    "- 디자인 C가 남녀(hue) 공통에게 가장 오래 홈페이지에 있게 만든다\n",
    "- 디자인 A와 B는 남자가 더 체류시간이 길다"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9304905807\n",
      "0.0000938553\n"
     ]
    }
   ],
   "source": [
    "ad = np.array([2, 5, 1, 3, 4, 1, 5, 3, 4, 2])\n",
    "amount = np.array([50, 57, 41, 54, 54, 38, 63, 48, 59, 46])\n",
    "pearsonr, p_value = stats.pearsonr(ad, amount)\n",
    "\n",
    "print(format(pearsonr, '.10f'))\n",
    "print(format(p_value, '.10f'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_pearsonr = pd.read_csv(\"pearsonr.csv\")\n",
    "data_pearsonr = data_pearsonr.set_index('캠페인_주차')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "corr_table = data_pearsonr.corr(method='pearson')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TV_광고횟수</th>\n",
       "      <th>매출액</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>TV_광고횟수</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.930491</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>매출액</th>\n",
       "      <td>0.930491</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          TV_광고횟수       매출액\n",
       "TV_광고횟수  1.000000  0.930491\n",
       "매출액      0.930491  1.000000"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "corr_table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9304905807411791\n"
     ]
    }
   ],
   "source": [
    "corr = lambda p : p['매출액'].corr(p['TV_광고횟수'])\n",
    "pearsonr = corr(data_pearsonr)\n",
    "print(pearsonr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}