Supervised_simple_features_10k.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "upU5iWemHtCV",
    "outputId": "4fc31e68-a2aa-48b8-b124-d7ff2678439a"
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import random\n",
    "\n",
    "import pickle\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "import re\n",
    "from tqdm import tqdm\n",
    "from nltk.corpus import stopwords\n",
    "import random\n",
    "\n",
    "import nltk\n",
    "from pre_processing import load_df_and_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "id": "WNufNhiOH19v"
   },
   "outputs": [],
   "source": [
    "df_raw = load_df_and_features(\"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['original_text', 'label', 'aoa', 'concrete_score', 'verb2',\n",
       "       'conc_unknown', 'aoa_perc_known_lem', 'conc_total', 'syllable_per_word',\n",
       "       'conc_mean_score', 'dale_chall_score', 'conc_subtlex_score'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_raw.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "icnh7aF8ILwN"
   },
   "source": [
    "### Consider whether to scale the features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAvMAAAKjCAYAAACQi4AgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOzdd3gUVdvA4d/JphfSQyhJ6CSQ0EtCaEpHX7EhAhYURVEBlSZNRJEiIopiQYogIEKkSRWld0JNIHQTQgnpCYSQtvP9sUuSTcDyprHv99zXlUt25szuc/acmTn7zJlRaZqGEEIIIYQQwvxYVHQAQgghhBBCiP+ODOaFEEIIIYQwUzKYF0IIIYQQwkzJYF4IIYQQQggzJYN5IYQQQgghzJQM5oUQQgghhDBTMpgXQgghhBCihJRSC5RS8UqpyPusV0qp2UqpC0qpk0qpZqXxuTKYF0IIIYQQouR+ALr/xfoeQF3j3yDgm9L4UBnMCyGEEEIIUUKapu0Ckv+iSC9gsWZwAHBRSlUp6efKYF4IIYQQQoiyVw2ILfT6inFZiViW9A3Eg83Ot69W0TGU1OQNL1d0CCXSpnJ2RYdQYpuv2FR0CCUW4pVT0SGUiJOVvqJDKLGMXFXRIZRY7C1dRYdQYsFmvi9MPelU0SGU2MigmxUdQok1dnv0gdihy3Occyd2+WsYpsfcNVfTtLnl9fn3I4N5IYQQQggh/oZx4F6SwftVwKfQ6+rGZSUi02yEEEIIIYQoe+uAF4xPtQkG0jRNu17SN5XMvBBCCCGEMEtKPTh5aaXUT0BHwEMpdQWYCFgBaJr2LbAR6AlcAG4DL5XG58pgXgghhBBCiBLSNK3v36zXgDdL+3NlMC+EEEIIIcySkhnj8g0IIYQQQghhriQzL4QQQgghzNKDNGe+osg3IIQQQgghhJmSzLwQQgghhDBLkpmXzLwQQgghhBBmSzLzQgghhBDCLCmlKjqECieZeSGEEEIIIcyUZOaFEEIIIYSZkry0fANCCCGEEEKYKcnMCyGEEEIIsyRPs5HMvBBCCCGEEGZLMvNCCCGEEMIsSWZeMvNCCCGEEEKYLRnMCyGEEEIIYaZkmo34r3074zV6dGpKQlI6LbqMquhw7in22Gn2LQxD0+vx79SGJk90NVl//fQF9i0MIznmGp3eeYlaIU3z122cPIf4c9F4+9ei+9jB5Rq3pmks/WI1Jw9EYW1jzStj+1KjfvVi5aLPxjJvyk9kZ+XQKDiA/sOeQCnF8jnrOL7vNJaWOryquTNwTF8cnOzIzcnlhxkriT4bi1KKfsOeIKBpnTKty/Xjpzi62NAGtR4KpUEv0zaIjzrPscW/kHr5Km2GvoRP62b5637u9xbOvlUBsHd3o/3I18s01qhDUayaswq9XiO4ZzBd+nY2WZ+bncuS6UuIPXcFh0r2vDjhRdy93QHYumwrBzYdxMJC8eRbTxLQMgCA7WE7OLDxACioWrMK/Ub1w8rail1rdrPzl50kXkvk41WTcXR2LNO6aZrGstmriTD2qYFj+uJ3nz41f8pP5GTnEBQcQL+hhj614uuCPuVZzZ2B7/XF3smu1OM8fSiKX75ajV6vEdKzNV37mbZBTnYuP05bmt8GL73/Iu7ebgD8tux39m80tMHTQ54koKU/KfEp/DhtGTdTbgIQ+mgIHZ/qAMCxHcfZuGgzNy7HM+Lrt/Gt71vq9bl05DR/zFuFlqenUdcQgp/uYrI+NyeHDbOWcONCLHaVHHhs5ACcK7tzasdhDq/ell8uPvoaL84aSeVaBW32y+S5pMUl8fJXY0o97sI0TWP+Z2s4uj8KGxtr3prwLLX9i/edi2di+fKj5WRn5dAsJICB7z6OUoo/z13l2+lh5GTnotNZMGjkU9RtWPBdnz99mTGvfsm7Hz1Hm4cbl2ldANIiI7n8889oej2ebdtSpUcPk/VxW7eSsGcPysICSycnar74Ijbu7vnr8zIziZg4EdcmTfDr16/M4wVDGyyctYZj+6KwsbXmjQnPUuse+++lM7HMMbZB0zYBvPSOoQ1mjV/MtcsJANy+mYm9kx0zFg/n5KGzLP16I7k5uVhaWfL8W48S2KJuudSptCjJS8s3UF6UUrf+Zn0NpVTkv3zPH5RST5cssv/ejyt30uuFaRX18X9Ln6dnz7wV9Bj3Br1njefCniOkxF43KePo4UrHN5+nTtsWxbZv/FhnHhr6QnmFa+LkgShuXElk+k9jGTCqN4tnht2z3KKZYQwY9QzTfxrLjSuJRBw8A0Bgy/p8vGgkkxeNxNvHkw1Lfgdgx68HAJi8aBQjZ73O8q/Wodfry6weer2e8IUr6DD6TXp8OoHL+8JJu2LaBvYebrR+/Xn8Qou3gc7aiu7TxtJ92tgyH8jr8/SsnB3Ga1NfY8yC9zi67Shx0XEmZfZvOoCdoz0TfhxPx6c68uv3vwIQFx3H0e3HGDP/PV6f9jorvwhDn6cnNSGVXat3Mfybdxkz/z30eo2j244CUKthTd6YMRi3yq5lWq+7Iox9auqysbw4sjeLP7t3n/rR2KemLjPtUw1a1OejH0by4Q8j8a5e0KdKkz5Pz8ovfmHwtEGMWziaI9uOcf0ebWDvZMfEJeN46OkOrJ1raIPr0XEc2XaMsQtGM3j6a6z43NAGFjoLnnj9McYtfI/hc95m19q9+e9ZpWYVXpn0MrUb1Sr1utytz+/fraT3xNcZOGcsUbuOkHjZtP9HbD2AraM9g+a+T4vHOrJj0ToAGnZsyYAvRjPgi9E88s7zuFR2MxnIn9t3AmtbmzKJu6ij+89wPTaROSvH8PqY3sz95Jd7lvvuk18YPOYZ5qwcw/XYRI7tN/SdxV+tp8/Arnz243CeHdSdxV+tz98mL0/Pj3M20KRVvXKpi6bXE7NsGXWHDiVw0iSSDh8m89o1kzL2Pj40GDuWwIkTcWvWjNhfTOt7Ze1anOqVT7x3Hdt/hrjYRGavHMOg93oz7z5t8P0nv/DamGeYvXIMcbGJHD9gaIN3Jr/AjMXDmbF4OK0fakTrDkEAODk7MHrGy8xcOpI3JzzLl5OWlVudROmRwbz4r+09dIbk1L/8jVKhEi5E4+ztQaXKHuisLKkd2ozowydNyjh5ueNeoxrKovj/Drpao/pYldPJsqhjeyIJ7d4CpRR1Gtbg9q1MUhPTTcqkJqaTmZFFnYY1UEoR2r0FR3dHABDYqj46Sx0AtRv6kZyQBsC16BsENDNkXSq5OmHvaEf0mdgyq0fyhWicvD1xrOyBztIS35DmXA03bQNHT3dc/KpBBf8vuWPOxOBZzQOPqh5YWlnS7KGmROyLMCkTuS+CVl1bAtC4Q2POHT2PpmlE7Iug2UNNsbS2xL2KO57VPIg5EwMYBnQ5WTnk5eWRfScbZw9nAKrXrZ6f1S8Px/ZE0qaboU/V/qs+dTuL2sY+1aZbC47do0/VauhHirFPlaaYM5fxKNQGzR9uSsQ+0xxHxN5IWndtBUATkzaIpPnDTbGytsSjijse1TyIOXMZZ3dnfOr5AGBrb4u3b2XSEg2xe/tVprKvV6nX467r52NwqeKJi7fhGBTQrhkXDpr2qfMHIwh82FCf+qFNuHziHJqmmZSJ2nUE/3bN819nZ2ZxeO12Qp4xvcpVVg7tiqRjz+Yopagf6EfGrUySi/Sd5MR0MjPuUD/QD6UUHXs25+AuQ9spBbcz7gBw+1Ymbp6V8rfbuHIPIQ8F4exatlem7sr4809svLyw9fTEwtISt5YtSTlxwqRMJX9/dDaGY79DrVrkpKQUbB8TQ256Os4NGpRLvHeF74qkfQ9DG9QztkFKkTZIMbZBPWMbtO/RnMM7TfcfTdPY/8dxQrsarkLXrF8dN0/DMcmnljfZWTnkZOeWT6VKiVIW5fb3oHpwI/sfpZRyVEr9oZQ6qpSKUEr1KrTaUim1VCkVpZQKU0rZG7dprpTaqZQ6opTaopSqUkHhm5WM5DQcPAqyng7urmQkl/4ApCykJKTj5uWS/9rV04WURNPYUxLT8g/C+WUSTA/uALs2HKJRa38AfOtU5djeU+Tl5pFwLYnoc7EkxaeWUS0gMyUVe/eCNrBzdyEz5Z9/Xl5OLlvGTmfrhBlcOXzi7zcogbTENFw8C2J18XTJH/TdlZqYhquXoYxOp8PWwZaM9Ixi2zp7uBiXufBQ74f4oO8kJvR+HztHO/xb+JdpPe4nJdG0T7ndp0+5FupThjLF+9SejYcICi79eqQmpuJaKEYXD2dSi/xoSEtMw8VYRqfTYWdsg9SENFw9C23r6UJqomlfS4pL5sqFK/gF+JV67PdyKykVJ4+CmJw8XLiZlFakTBqVjGUsdDpsHGzJvJlhUubMnqMEtC+YfrZ76QZaPv4QVjbWZRh9geSENDwKtYu7l3N+gqBwGXfPwmVc8su8/PbjLP5qPa8+9iGLvvyV/oN7ApAUn8bBnRF0e7JNOdTCIDs1FWs3t/zX1i4uJoP1ohL37ME5MBAwZPVjV67Ep3fvMo+zqOSENDwqF/p+Pe/TBl73boO7oo5fwtnNiSo+nsU+4+D2k9SqXx0ra5mBbW5kMF/+7gBPaJrWDHgImKlUfkqyPvC1pmkBQDrwhlLKCvgSeFrTtObAAuDjCohbmKF1i7ei01kQ0tWQ1WvXsxVuns588Oosln25hrqBNbCweHAPA//58iO6TRlNyFsvcXRxGDdvJFR0SP/K7Zu3idwXycSl7/PRig8NGdWt4RUdVon8ungrFjoLgrs0//vCD5CszCzmT1zIk288gZ2DbUWH849dOxuNpY01nn6Ge0duXLpCalwi9ULKfm55adm8ah8vDevF9+ve56Vhvfj64xUALPh8Dc+/+egDewxKPHCAjJgYvLsaroDE79iBc2Ag1q7lMzWuLOzdeozQLk2LLY+9FMfSrzfw6ugKm7n7X5PMvNwAWxEUMEUp1R7QA9WAysZ1sZqm7TX+ewkwFNgMBAJbjWN+HWA66bLoByg1CBgEYOnaAkvHsr3B8UHl4OZMRmKhy6NJKTi4Of/FFhXr91V72Gmc017T34fkQhnzlIRUXD1MY3f1MM3MpCSk4lro8vXujYc4se80oz4fzN3fizpLHf2GPp5fZvLg2XjfI0NTWuxcXbidVNAGmUmp2Lm6/MUWpuzdDGUdK3vg1aAuqdGxOFUum3idPZxJTSiINTUhNX9KzF0uHs6kxKfg4ulCXl4edzLu4FDJodi2aYmGbc8ePYebtxuOLoYpBI3aNeLP03/Sskvx+wPKwh+r9rBr/b37VPJ9+lTh6TOGMgV9as+mQ5zcf5oRswr6VGly8XAhpVCMqYlpuHiaxujs4UxqfCquxjbINLaBi6czKQmFtk1IxcWY8c7LzWPexIW06NycJu0blXrc9+Po7sLNQlcHbiam4uTuXKSMM+mJqTh5uKLPyyMr4w52Tg7566N2HyWg0BSba2f+JO7CZb595QP0eXncTrvFT2Nn03fK0FKNfVPYHrauPQhAnQAfEgu1S1K86VVBADdPZ5ISCpdJzS+zY2M4A981HHfadGrM11MMg/mLUVf4bPyPANxMy+DI/jPodBb587nLgrWLC9nJyfmvs1NTsbrH4Dzt9Gmub9yI/4gRWFhZAXDr0iVunT9P/M6d6O/cQZ+Xh4WtLT5PPlkmsW4O28Mf6wxtUDvAh8Qbhb7fhPu0Qfy92wAM+8GhHRFM++Edk+2S4lP59L2FvDmhL97VPcqiKqKMPbg/M/539Qc8geaapjUBbgB300RakbIahsH/KU3Tmhj/gjRN+8uJkpqmzdU0rYWmaS3+vw7kATzr+JF2PYH0G4nk5eRyce9R/FqW34n83+r8ZFs+WjiCjxaOoFm7IPZuDkfTNC6cisbO0RaXQoMqABePStg52HDhVDSaprF3czhN2xouB588GMWmZdsZNnUgNrYFl+Kz7mSTlZkFQOThs1joLKhW07vM6uRW24+bcfHcik8kLzeXy/uPUK35PztRZ9+6TV5OjiHu9FsknrtEpWplN8PM19+XhKuJJF1PIjcnl6PbjxHYJtCkTGBIIId+OwzAiZ0nqNu0LkopAtsEcnT7MXKzc0m6nkTC1UT8/P1w9XIhJiqG7DvZaJrGuaPn8fatfK+PLxOdnmzLpAUjmLRgBE3bBbFvi6FPXTwVjb3DffqUvQ0XjX1q35aCPhVh7FNDivSp0uTr70PC1QQSjW1wZNsxgkIampQJahPIwd8OAXB85wnqNa2DUoqgkIYc2XaMnOxcEq8nkXA1AT9/X8OToWYsx9u3Mg/37lgmcd9Plbq+pFxLIDUuibycXKJ2H6VOa9P+X6dVIJHbDPU5u/c4vo3q5v9Q0vR6zu45ZjLFpmnPdrz5w2Ren/cB/ae9jVtVr1IfyAP0eLotn/04nM9+HE6rDoHs2HgETdM4GxmDvaMtbkX6jptHJewcbDkbGYOmaezYeIRW7Q19x9WjEqeOXgQgIvx8/hSPb1eP47s14/luzXhCHmrEoJFPlulAHsChRg2y4uPJSkxEn5tL8uHDuDY2vcqRcfkyMUuWUPfNN7GqVFDP2q+8QuPp02k8dSo+vXvjERxcZgN5gO5Pt82/abVV+0B2bTK0wbnIGOwdbE1+aIPhe7ZzsOWcsQ12bTpCi/YFx7CIw+ep6udlMhUn42Ym04bPo98bj+DfuGaZ1aUsSWZeMvMVwRmI1zQtRyn1EFB48qavUipE07T9QD9gD3AW8Ly73Djtpp6maafKP3RTi74cQruQADxcnbhw8Cs++iyMRT/vqOiw8lnodIS+8gybJs9Br9eo/3Awbj5VCF++Ho/avtRo2Yj4CzFs/eR7sjJuExMewZGfN9D78/EArBs/i9RrN8i5k8XSQeNp/0Y/fJqUz01PjUMCOHkgilHPTsHG1oqBY/rmr5vw0qd8tHAEAC+8+3ShR1P60yjY8DjEJbNWkZuTx4x3vwUMN8EOGNGb9JRbzBz+HcpC4erhzKDxZftYNQudjuYDnmHn1Dno9XpqdQzB2acqESvX41bTl2otGpF0MYY9n80lO+M2145GErFyAz0/nUD6tTgOz/sJpRSaphHwWFecq5fdYF6n0/HUkKf4ZvS36PV6gnu0pkqNKmxcuBGf+r4EtQkkuGcwS6Yu4aPnJ2PvZM+L4w1PO6pSowpNOzZhystT0ekseHrIU1joLKgRUIPG7Rsz4/VPsdBZUL1Oddo8YpgfvHPVTv74eRs3k28y/dVPaNCqAX1HPFtm9WsUHMDJ/VG813cK1jZWvFyoT018+VMmLTD0qefefZoFUw19Kqi1P0HGPrX081XkZOcx826fauDHCyNKd+6wTqej95Cn+Hr0d2h5xjaoWYUNCzfhW8+HoNBAQnq2ZvGUpUx67mPsnex5acLzgOHJNM06NmHKS9Ow0FnQe+jTWOgsuBhxicNbw6laqwrTXp0BwH8GPkLD4Aac2H2SsC9XcSvtFt+O/Z5qtavx5iel99QkC52Ozq89zcoPvkbT6wnqHIyHbxV2L92Adx1f6rYOolGXEDZ89iNzB32IrZM9j40ckL997KmLOHm44OJdsdnS5m0COLovijeenoqNrRVvjS/op+8+P5PPfhwOwKCRTxV6NKU/zUIM91W8MaY382etJS8vD2trKwaPqbipHEqnw7dvX85+/jno9XiEhmJXtSpX167F3s8P1yZNuBIWRl5WFhe++w4AGzc36r71VoXFDNDU2AZDe0/F2saKNwq1wcgXZjJjsaENXhn5FF9PNrRBk2B/moYU3Nuy9/fiU2w2h+0h7koSYQu2ErZgKwDjPx+Es5tTOdRKlBZV9K55UTaUUrc0TXNUSnkAvwKOQDgQDNx9yO1m47LmwGngeU3TbiulmgCzMfwQsAQ+1zTte6XUD8B6TdPu/Yw5wM63r9k38OQNL1d0CCXSpnJ2RYdQYpuvVMxTfUpTiFdORYdQIk5WZfcI0fKSkVuxTywqDbG3dBUdQokFm/m+MPWk+Q80RwbdrOgQSqyx26MPxA7tWf+dchvnJJyd9UDUuSjJzJcTTdMcjf9NBELuU+yej4fQNO040P4eyweUVnxCCCGEEML8yGBeCCGEEEKYpQd5Lnt5kW9ACCGEEEIIMyWZeSGEEEIIYZYkMy+ZeSGEEEIIIcyWZOaFEEIIIYRZksy8ZOaFEEIIIYQwWzKYF0IIIYQQwkzJNBshhBBCCGGmJC8t34AQQgghhBBmSjLzQgghhBDCLMkNsJKZF0IIIYQQwmxJZl4IIYQQQpglycxLZl4IIYQQQgizJZl5IYQQQghhlpTkpeUbEEIIIYQQwlxJZl4IIYQQQpglmTMvmXkhhBBCCCHMlmTmhRBCCCGEWVJKVXQIFU4y80IIIYQQQpgpycz/j5u84eWKDqHExj+yoKJDKJFzUf0qOoQS61/7TkWHUGJHE837cNe9et2KDqHEtl27UNEhlNjnB+0rOoQSO/KcXUWHUCL+lVIrOoQSC3KtVdEh/M+QOfOSmRdCCCGEEMJsmXeqSgghhBBC/L8lz5mXzLwQQgghhBBmSwbzQgghhBBCmCmZZiOEEEIIIcyS3AArmXkhhBBCCCHMlmTmhRBCCCGEWZLMvGTmhRBCCCGEMFuSmRdCCCGEEGZJHk0pmXkhhBBCCCHMlmTmhRBCCCGEeZI585KZF0IIIYQQwlxJZl4IIYQQQpgleZqNZOaFEEIIIYQwW5KZF0IIIYQQZkkpVdEhVDjJzAshhBBCCGGmJDMvhBBCCCHMkjxnXjLzQgghhBBCmC3JzAshhBBCCLMkT7ORwbz4C7HHTrNvYRiaXo9/pzY0eaKryfrrpy+wb2EYyTHX6PTOS9QKaZq/buPkOcSfi8bbvxbdxw4u79D/kW9nvEaPTk1JSEqnRZdRFR3OfWmaxpwZazm0JwobW2tGTepD3YDqxcot+GoTWzeEczM9k/V7p+Qv37LuMHM/X4+HlzMAvfqE0vOJ1uUa/9yZaziyzxD/sPefpY5/8fgXf72R7RvDuXUzk5U7p+Yvjzx6ke9nrSX6wnVGTX6O0E6NyyXu8+FRbPpuFZpeT7NuwbR7povJ+tycXFZ9uoTrF2Kxc3Kg95gXca3snr8+NT6ZOa9PpWP/HoQ+9TCJV26wctqi/PUp1xN56PmehDzesVzqo2kaH388l507j2Bra8O0acNo2LBOsXIDB04kISGZvLw8mjdvyMSJr6PT6Xj77en8+edVAG7ezMDJyYG1a2eXSZwrvlzNqYNRWNta8cLovvjW8ylWLuZsLIun/0ROVg4NWwfwzJAnUEqRkZ7BvA8XkxSXjLu3G69MfBEHJ3tO7Ing14WbUEphobOg91tPUCeoFmePnSdszpr89427HM/A91+gSdugUq9baFVXRreshU4pVl2IY37klXuW6+zrzqyODeiz4Rink27hbGPJZx0CCHR3Yu3FG0w5dLHUY/unNE1j2pTF7N51AltbayZPeY0GDWsWK/f6q9NJSEglLzePZi3qM27CS+h0FmzZfJBvvvqFS5eu8dOKD2kYWKvc63D1+CnCFxnObXUeDiWwl+m57UbUecIX/ULK5au0G/oSfsHNALiVkMTOmd+jaXr0eXn4d+tIvS7tyj1+TdOY8vF8du0y7MtTpg6hYcPaxcq9+sqHJCSkkJuXR4vmAUx4fxA6nQ6AJT9uYNmyTVjoLOjQoTkjR75Y3tUQpUgG8+Ke9Hl69sxbwSPvv4WDmwur35uBX4sgXH2q5Jdx9HCl45vPc3LdH8W2b/xYZ3Kzs4n6bU95hv2v/LhyJ98u2sK8WW9UdCh/6dDeM1y9nMCite8RFXGZL6b+wleLhxUrF9y+Ab36hPLi49OKrevYtTFD3nuyPMIt5si+M1yLTeS7X8ZwNvIy30z/hZkLi8ffql1DHn2mLa89NdVkuae3K2+//yyrl+wop4gN/X/D1yt54eM3qOThwty3Z1I/OAgvX+/8Mke37MfO0Y5h8ycQsfMoWxf8yjNjBuSv3/L9Guq0aJD/2qN6ZQZ/NSr//We+8D4BIY3KrU67dh0hOvoav/32HSdOnOWDD75h5cqZxcp98cVoHB3t0TSNoUOnsnnzXh55pD2ffz46v8y0afNxdLQvkzhPHYwi/moCk5aM5c+oGH6aFcbob94pVu6nz8PoP+IZagb48dV7czl16AyBrQPYsuwP/JvVpVu/zmxZ9ju/LfuDJ177D/Wb16NRaCBKKa5cvMa8SYv4YPEY6jety7h5IwHISM/g/eem0KBF/VKvl4WCca1rM2hrJHG3s1jeswnbY5O5lHbbpJy9pY7nAqpxIiE9f1l2np6vjsdQx8Weui4OpR7bv7F71wliYuLYsHkmJ09cYPKHC1n284fFyn06a0h+P3p32Bf8tvkgPR4JoW7d6sz68m0+nLigAqIHvV7PoQUr6DxuCPbuLmwa+wnVmwfhUr3g3Obg7kabwc9zev3vJtvauTrT/aPh6KysyLlzh19HfEz15kHYu7mUax127TpKTMw1Nm/5mhMnzvHhpO/4ecUnxcrN+nxEfhsMG/oJmzfv45FH2nHwQAR/bDvEmrWzsLa2IikptVzjF6VPrk2Ie0q4EI2ztweVKnugs7Kkdmgzog+fNCnj5OWOe41qKIvij4Wq1qg+VrY25RXuf2XvoTMkp96q6DD+1r4dp+jyaAuUUjRo5Metm3dIKnSiv6tBIz/cPStVQIR/7cCuSB7u2RylFP5BfmTczCQ5sXj8/kF+uHkUj79yVTdq1q16z35WVq6ei8GtqiduVTywtLIksH0zzuyPMClz5kAkTTq3AqBB28b8eeIcmqYBELXvJC7e7iaD/8IunTiHq7cHLpXdyrYihfzxxwEef/xhlFI0aeJPenoG8fHJxcrdHaTn5uaRk5Nb7LFvmqaxadMeHn20Q5nEeWJvJMFdW6KUolaDGtzOyCQtKc2kTFpSGncy7lCrQQ2UUgR3bcmJPYb2ObEvkuBuLQEI7taS43sNy23tbPLrkn0nm3s9ze7ozhM0bOWPta11qdcryN2JyzfvcOXWHXL1GpuiE3jIp3j7v9XEj/mRsWTn6fOXZebqORafbrKsomzfdoTHerVDKUXjJnW5mX6bhPiUYuWK9yPD8lq1q1GzZtXyDJCw9c8AACAASURBVNlE0oVonLw9carsgc7SEr82zYkNNz23OXq54+pXjaKdRGdpic7KCgB9Tm7+/l7etv1xiF69HjLuy/X/9b68fPlmXn31SaytDXVxdy/fHyOlTqny+3tAyWD+AaCUWqOUOqKUOqWUGmRc1lcpFaGUilRKTS9U9hulVLix7KSyiikjOQ0HD9f81w7urmQkp/3FFqKsJMan4Vm54GDr6eVMYsK/a4vd2yJ49ZmZTBq5iPi48s3CJMWn4VEofncvZ5LiH+y+lJ6UhrNHQczOHi7cLDKgvJmUSiVPwz6i0+mwsbfldnoGWZlZ7An7g479ut/3/SN3HiWoY7OyCf4+btxIwtvbI/+1t7c7N24k3bPswIHv06bNczg42NGtWxuTdeHhp3B3d6FGjbIZkKUmpuHqVfDdu3q4kJqYVqyMi6dz/msXT+f8MjeTb+LsblhXya0SN5Nv5pc7vvskH7wwlTljvuf5UX2LfXb49mO07FQ27eJlb0NcRlb+6xu3s6lsb5rwCHBzwNvBht1Xiw+OHxTxN5Lx9i6YTlbZ2434ewzmAV57ZRod2g7G3sGWLt3Kb2rfX7mdnIqDe6Fzm5sLmcn//JiYkZjCr6M+5pc3xxP4WJdyz8qDcV+uUtAG3t7uxN8oPpgHeGXgJNqGDjDuyyEAREdf40j4afo8M4rnnxtHRMT5colblB0ZzD8YXtY0rTnQAhiqlKoGTAceBpoALZVSjxvLjtM0rQXQCOiglCq/6/TCLAW3b8CS9eP4fsVwmreuxyfv/1TRIf1P27F0EyGPd8TG7t5XpnJzcjl7MJKGbZuUc2T/3Pz5H7Jnz2Kys3M4cMA0a7l+/S4efbR9BUX276gi2bQm7RrxweIxvP7Ry6xbsNGkbFpSGtcuXadBS//yDhMABYxsUYtPwy9VyOeXhe/mvcf2XXPIyc7l4IFTFR1OqXDwcOU/n4zj8c8/4OKug2SmFr/K+CCZN38iu3YvMO7LhqtUuXl5pKXdYvnP0xk56kXeefvTCrvKUCosyvHvASVz5h8MQ5VSTxj/7QO8AuzQNC0BQCm1FGgPrAGeMWbvLYEqQAPA5GxrXD8I4On3hxH89CP/OiAHN2cyEguyLRlJKTi4Of/FFqI0rf15LxtXHwSgXkMfEm4UZI4S4tPw8PznbeFcaI5tjydaM3f2htIL9D42rNzDljWG+Os28CGxUPxJ8Wm4ez3YfamSuzNpiQUxpyWm4uRuGrOTuwvpCSk4e7iQl5dH1u072Fdy4MrZGE7vOcHWBeu4k5GJUgpLa0ta/8cwAL4QHkWV2tVxdC37KVFLl25gxYotAAQF1SUuLjF/XVxcEpUL3bBblI2NNZ06BfPHHwcJDTXc3J6bm8fWrftZtWpWqca5Y/Ue9m7YD4Cfvy8p8QXffUpiKi4ept+9i4czqYWuTqUmpOWXcXJzIi0pDWd3Z9KS0nBydSz2eXUb1ybxehK30m7h6GxYf2T7cZq0DUJnqSvVut0VfzsLb4eCH3iV7a25cbsgU+9gpaOOiwMLuhnyMx521nz5UAOGbD/N6aSKnQ7409Lf+CVsOwCBgbWIiyu4onMjLhkvL9f7bYqNjTUPPdyc7duO0Ca09G8q/rfs3VzISCp0bktOxe6/yK7bu7ng4lOF+DMX8m+QLUtLl24kbOVWAAKD6hB3vaAN4uKS8PqLKXs2NtY83KkV2/44RGhoE7wre9ClSzBKKRo1qoeFhSIlJR03OcebrQf4d8b/D0qpjkBnIETTtMbAMeD4fcrWBEYAnTRNawRsAGyLltM0ba6maS00TWvx3wzkATzr+JF2PYH0G4nk5eRyce9R/FrKRYDy0qtPKN8tf5fvlr9LaMeGbF0fjqZpnD4Zg4Oj7b+aG194fv3+nafwreFVFiGbeKR3W2YvHc7spcMJ7hDIto1H0DSNMxEx2Dva3nNu/IOkaj1fkq8lkBKXRG5OLpG7juIfHGhSpn7rQI7/fgiA03tOULNRXZRSDJwxjHd+mMg7P0wkuFcH2vXpkj+QB4jYeYSgDuUzxaZ//0dYu3Y2a9fOpnPnYNas2YamaRw/fgYnJ3u8vEwHABkZmflzb3Nz89ix4zC1ahU8eWjfvuPUqlXNZLpOaej4RFvGzRvJuHkjaRwayIHfDqNpGpdOR2PnYJc/beYuZ3dnbB1suXQ6Gk3TOPDbYRqHGtqnUZtADmw5DMCBLYdp3MawPP5qQn728fK5WHJz8nCoVPBD9/C2Y7Qooyk2AJFJN/FzsqWaow2WFooeNTzZEVswNeJWTh7tVxyg+6rDdF91mJMJ6Q/EQB6gb/+uhK2eStjqqTzcqQXr1u5G0zROHD+Po5MdnkUG87cz7uTPo8/NzWPXzmPUrFVx8+QLc6/tx824eG7GJ5KXm0vMviP4NP9nPzIyklLIzc4GIOvWbeLPXKJS1cplGW6+/v17snrNLFavmUWnTq1Zu3a7cV8++4/25Z07j+Tvy506t+LgIUOW/s8/r5KTk4trOSQXyswDNGdeKdVdKXVWKXVBKfXePdb7KqW2K6WOKaVOKqV6lsZXIJn5iucMpGiadlsp5Q8EA/YYptB4AClAX+BLoBKQAaQppSoDPYAdZRGUhU5H6CvPsGnyHPR6jfoPB+PmU4Xw5evxqO1LjZaNiL8Qw9ZPvicr4zYx4REc+XkDvT8fD8C68bNIvXaDnDtZLB00nvZv9MOnSYO/+dTytejLIbQLCcDD1YkLB7/io8/CWPTzjooOq5jWbQM4tOcML/Saho2tFSM/6JO/7rVnP+O75e8CMPfz9WzbfIysOzk82/0jejzeihdf78bq5XvYv/MUOp0FTs72jJr0bLnG3yI0gPB9UQx6cio2tlYMm1Dw+UP7z2T20uEALJz9Kzt/M8Q/4NEP6fpYa/oN6sa505eZMuoHbqVncnj3aZbO3cLXP5fto0R1Oh09Bz/Fj+O/Qa/X07RrMF5+Vdj240aq1vXBPziIZt2CWfXpEr4Y+BF2TvY8PfrvH+2WfSeLi8fO8p8hff62bGnr0KEFO3eG06XLIOzsbJgypeCJQr16DWXt2tlkZt5h8OCPyM7ORdP0tG7diGef7ZFfbuPGXTzySNnc+HpXYHADIg9G8f5zH2NtY80Lowv6y8evzMh/8kzft59i0bSfyMnOoWGrABq2DgCgW99OzJu0iL0bD+JW2ZVXJxra5diukxzcchidpQ4rGyteef+F/BsCk+KSSUlIpW7j4o/3Ky15Gkw5dJFvOweiU4rVF25wMe02bzb241TSTXZcufec57s2P9kSRysdVhYWPOzjzqDfI4s9Cac8tOvQhF27jtOz27v5j6a86+knxhC2eiq3M7MY8uZnZGfnoOk1WrZuwDN9OgHwx9bDTPl4ESnJN3nj9Rn4+/vx3bxiY54yY6HT0eqlZ/hjyhzDoykfCsHFpyrHV6zHvZYvPi0akXgxhp0z55KVcZsrRyM5EbaBxz6dQNrVOI4sWYVCoaHR4NFOuPpWK7fY7+rQoTm7dh2hW9fBhkdTThmSv+6Jx99h9ZpZZGZm8eYbU8nOzkGv6WndKog+z3YD4MknOzF+3Ff85z9DsbKyYuq0ocVudBf/nlJKB8wBugBXgMNKqXWapp0uVGw8sELTtG+UUg2AjUCNEn+2Wc+T+h+glLLBMH2mBnAWcAE+wDCFZiyGqZQbNE0bbSz/A9AGiAXSgHWapv1wv/efGbHV7Bt4/CMV8wiz0nIuql9Fh1Bimbnmf6A/mmjeuYtna5f/87hL27ZrFyo6hBIb9rtTRYdQYkees6voEErkkxPm/yjFsU0ejCsVJWGhGjwQJ4Z6bb4tt3HOuX2v37fOSqkQ4ANN07oZX48B0DRtaqEy3wGXNE2bbiw/U9O0Nvd8w3/BvM9u/wM0TcvCkGG/l2J3KmqaNqBMAxJCCCGEEP9WNQyJ1ruuAEUf4/QB8JtSagjggGGadYnJnHkhhBBCCGGeyvFpNkqpQcbHg9/9G/Qvo+0L/KBpWnWgJ/CjUqrEY3HJzAshhBBCCPE3NE2bC8y9z+qrGJ5IeFd147LCBgLdje+1XyllC3gA8SWJSzLzQgghhBDCLGlKldvf3zgM1FVK1VRKWQPPAuuKlLkMdAJQSgVgeCJhQkm/AxnMCyGEEEIIUQKapuUCbwFbgCgMT605pZT6UCn1mLHYcOBVpdQJDPdFDtBK4Uk0Ms1GCCGEEEKYpwfimToGmqZtxPC4ycLL3i/079NAaGl/rmTmhRBCCCGEMFOSmRdCCCGEEObJ4gFKzVcQycwLIYQQQghhpmQwL4QQQgghhJmSaTZCCCGEEMI8/f0jI//nSWZeCCGEEEIIMyWZeSGEEEIIYZ4kMS+ZeSGEEEIIIcyVZOaFEEIIIYR5kkdTSmZeCCGEEEIIcyWZeSGEEEIIYZ7kaTaSmRdCCCGEEMJcSWZeCCGEEEKYJ0nMS2ZeCCGEEEIIc6U0TavoGEQZ2h+/wewbuLqDvqJDKJF6AcsqOoQSW7brhYoOocS6Vneo6BBKZPThvIoOocR6VL9T0SGUWBU78z4eAcRmmHcez83G/NvgRqauokMosadrdn8gcuJ1uy8ot3HO+c0vPxB1Lsq892ghhBBCCCH+H5M580IIIYQQwjw9kLny8iWZeSGEEEIIIcyUZOaFEEIIIYRZ0uQ585KZF0IIIYQQwlxJZl4IIYQQQpgnC8nMS2ZeCCGEEEIIMyWDeSGEEEIIIcyUTLMRQgghhBDmSWbZSGZeCCGEEEIIcyWZeSGEEEIIYZ7k0ZSSmRdCCCGEEMJcSWZeCCGEEEKYJ3k0pWTmhRBCCCGEMFeSmRdCCCGEEOZJEvOSmRdCCCGEEMJcSWZeCCGEEEKYJ3majWTmhRBCCCGEMFeSmRdCCCGEEOZJMvOSmRdCCCGEEMJcSWb+byilBgC/aZp2raJjKQ+aprH0i9WcPBCFtY01r4ztS4361YuViz4by7wpP5GdlUOj4AD6D3sCpRTL56zj+L7TWFrq8KrmzsAxfXFwsiM3J5cfZqwk+mwsSin6DXuCgKZ1yqU+c2as5dCeKGxsrRk1qQ91A4rXZ8FXm9i6IZyb6Zms3zslf/mWdYeZ+/l6PLycAejVJ5SeT7Qu87j/qW9nvEaPTk1JSEqnRZdRFR3OfZ09HMWv365Cy9No2SOYjn06m6zPzc5lxYwlXD1/BftK9vQd+yJu3u7Enolh1Rc/A6Bp0Pn57gSGNir3+DVNY8bU5ezZFYGtnTWTPn6JgAZ+JmUyM7MY/e53XIlNwMJC0b5jY4a++xQAn077mfBDZwC4cyeb5OSb7Dowu9zrcVfiyVOcXbYCTa+nWvtQaj7a3WR9zObfubprD8pCh7WTIw0GvoCdh3u5xxl1KIo1X69Cr9cI7hFMp77F+82y6UuIPX8Fh0r2vDDe0G8y0jL44cOFxJ69TMturXhqyNPF3nv+hO9Jup7EqHnvlWkdNE1j0eerOb4/CmtbawaP60vNexxTL52J5duPDcfUJiEBvPi24ZgaNn8z29YdoJKLIwB9XutJ0zYNAIi5cI35n6zkdsYdLCwUk+e9g7WNVanX4ezhKNZ+swpNr9GqezAPPVu8HZbf3X+d7Ok/ztAO546cZdP8X8nLzUNnqeORVx+jTtN6AHw74kvSk9OxsjbE++rUwTi6OpV67HdpmsZPs1cTcdBwbnt5TF/86t373LZw6k9kZ+cQ1DqAvkMN7RC+/TjrftjC9Zh4xn37NjX8fQx1z81j0Sc/c/ncFfLy9LTp1oKez3Uu9r4ldS48ig3frEKv19OiezAd+nQxWZ+bnUvYp0u4ej4W+0oOPDvmRVy93Yk9G8Ma4zEUTePh57rTMLQxCbE3WD51Uf72KXGJdHq+J6FPdCz12MucpKVlMA+glLLUNC33PqsHAJFAhQ/mlVI6TdPyyvIzTh6I4saVRKb/NJaLp2NYPDOM9+e+XazcoplhDBj1DLUb+PHZyO+JOHiGRsEBBLasT+/XHkFnqWPFN7+yYcnvPDP4P+z49QAAkxeNIj3lJjNHfM/E79/GwqJs98JDe89w9XICi9a+R1TEZb6Y+gtfLR5WrFxw+wb06hPKi49PK7auY9fGDHnvyTKN87/148qdfLtoC/NmvVHRodyXPk/P2jlhDJw6GGcPF74a8hkBwYFU9vPOL3N4ywHsHO0Z+cN4Tuw4yub5v9Jv3AAq16jCW18NR6fTkZ6UxheDZxAQ3BCdTleuddi7O5LLMfGs3fQxEScvMfXDpSxePrZYuecHdKVla39ysnN5beBM9u6OILRdECPe65NfZvnSPzgTFVue4ZvQ9HrO/PgTzUYOw9bNlYOTpuLZtBGO1arml3Hy86H1xLHobKyJ3baT8ytW0eiNV8s1Tn2enlVfhvH69ME4e7ow683PaNgmEO9C/ebgpgPYOdkzbvF4jm0/yvrvf+WFCQOwtLakx4CexEVf53r09WLvfXL3CaxtbcqlHsf3RxF3JZFZP4/lwqkY5n8axuTvix9TF3waxqujn6FOQz+mj/ieEwfO0CQkAICefTrwaL+HTMrn5eYx58OlvDmhH351q3EzLQNLy9LfL/R5elZ/Fcar0wz775dDPqNBiOn+e2izYf8d/cN4jm8/ysb5v/LcuAE4ODsw4KNXcXZ3Ju7P68wb+y3jf5qUv13f957Hp55vqcd8LxEHo4i/ksiUpWO5dDqGJZ+FMe7b4u2w5LMwXhj5DLUa+PHFqO+JPHiGoOAAqtaswhsfvcTimStNyh/ZfpzcnFwm/TCKrDvZvP/idFp1aoZHFbdSi12fp+fXOSt5acobVPJw4ZuhMwkIDsKrUBuEb9mPraMdwxdO4OSOo2xZ8CvPjh1AZb8qvPFlwTH0qzc+wT84EE+fygz5elT++09/7n0atCn/RIkoHQ/87xml1AtKqZNKqRNKqR+VUjWUUtuMy/5QSvkay/2glJqtlNqnlLqklHq60HuMVkpFGN9jmnHZDqXU50qpcGCYUqq5UmqnUuqIUmqLUqqK8T1aAEuVUseVUnb3KvcXsQ9VSp02xrrcuMxRKbXQGM9JpdRTxuV9jcsilVLTC73HLaXUTKXUCSBEKfWcUuqQMZ7vlFKlevQ+tieS0O4tUEpRp2ENbt/KJDUx3aRMamI6mRlZ1GlYA6UUod1bcHR3BACBreqjM55Qajf0IzkhDYBr0TcIaFYXgEquTtg72hF9puwHNPt2nKLLo4b6NGjkx62bd0hKSC9WrkEjP9w9K5V5PKVt76EzJKfequgw/lLs2Rjcq3rgXsUDSytLGndsyun9ESZlTu+PoFmXlgAEtmvMhePn0TQNa1vr/IF7bk5uhU2N3LHtOI8+FoxSikaNa3Pz5m0SElJNytjZ2dCytT8AVtaWBDTw40ZcSrH32rzxMN17tiqXuO8l7VI09pW9sPfyxMLSEu/WLUk4dtKkjFtAfXQ21gA4167JneTi9Shrl8/G4FHVA/eqhn7TtGNTIvea9pvIfRG07GroN43aN+b8MUO/sbGzoVZQLSyti+ersjKz2Bm2gy7PdS2XehzZE0k74zG1bmANbt/MJKXIMTXFeEytG2g4prbr3oLw3RH3eUeDk4fO4lu7Cn51qwHg5OyAha70T+mxd9vh7v7boSmn9hXff1sY99+g9o25YGyHanWq4+xuuKpZuYY3Odk55GbfL29Wto7viSSkm6Edat89tyUVObclpXPndha1jee2kG4tOLbHUNeqNSrj7etV/I2VIiszm7zcPHKycrC0tMTWoXR/KF45G4NbFU/cjG3QqEMzooocQ6P2R9Kss+G40rBdYy4eP3fPY+i9nsl+8fg53Kp44Fq59H6AlCulyu/vAfVAZ+aVUg2B8UAbTdMSlVJuwCJgkaZpi5RSLwOzgceNm1QB2gL+wDogTCnVA+gFtNY07bbxPe6y1jSthVLKCtgJ9NI0LUEp1Qf4WNO0l5VSbwEjNE0LN5b7smg54OX7VOE9oKamaVlKKRfjsglAmqZpQcY6uiqlqgLTgeZACvCbUupxTdPWAA7AQU3ThiulAoDRQKimaTlKqa+B/sDi//Y7LiolIR03L5f8166eLqQkpuHiUTDQTUlMw83T2bTMPQbIuzYcovXDTQDwrVOVY3tPEdy5KcnxqUSfiyUpPpVaRaYqlLbE+DQ8KxfUx9PLmcSEtH81cN+9LYKTR/+kup8Hg4f3wsvb5e83EvnSk9Jw9nTNf+3s4ULsmRjTMolpuBjL6HQ6bB1suZ2egYOzI5fPRBM2czmp8ck8M+q5cs/KA8THp1DZu+DQ4VXZlYQbqXh63rsv3Ey/za4dJ+j7XCeT5deuJXHtSmL+oL8iZKWkYONW0B42ri6kX/rzvuWv7dqLR6PA8gjNRFpiGi5eBXG6eLoQU6TfpCUV7zcZ6Rk4Ojve9303LdxIh94Plcl0lHtJTkjHvdAx1c3LheSENFwLHVOTE9Jw8yo4prp7upBc6Ji65Zc97NocTi1/H5576zEcK9lzPTYBpRRT3/mO9NRbhHRuymP9Hy71+NMSi+y/nsX338Jliu6/d0XsPkG1OtVNfmCt/PQnlIUiqG1jOvXviirDwVJqYvFzW2pCGi7uBe2QmpCGa5FzW9FkVlHNOzbm+N5Ihj/5AdlZOfR5sxeOlRxKNXbDMbQg9koeLsSeLXIMTUq9bxvEnolm1Wc/kRqfzNMjix9DT+48SqOOzUo1ZlG+HvTM/MPASk3TEgE0TUsGQoBlxvU/Yhi837VG0zS9pmmngcrGZZ2BhZqm3S70HncZJ5JRHwgEtiqljmP4AVF8Mt0/L3fXSQxZ/eeAu+mIzsCcuwU0TUsBWgI7NE1LME73WQq0NxbJA34x/rsThgH/YePndwJqFf1QpdQgpVS4Uip8zeLNfxFe2Vm3eCs6nQUhXZsD0K5nK9w8nfng1Vks+3INdQNrlPkUm9IQ3L4BS9aP4/sVw2neuh6fvP9TRYf0/46vfw3e/f493vryXXYs/52c7JyKDukv5ebmMWbk9zzbvxPVfTxN1v228RCdujZDVwYZ1LJwfd9B0v+8TI0eXf6+sBm4euEKSdcTadTWfKYTdH4ilC9WjGPaD8Nxda/Ekq/WAYapEWdP/smbE/vzwTdDCN8ZQWT4uQqO9t7ioq+zcf6vPDXsmfxlfd97nnfnjmbwZ0P5M/ISR38/XIER/vf+jLqMhYUFn676gGnLx/Hbih0kXEuq6LBM+PjXYNjcMQyePZydP5seQ3NzcjlzIJKgdk0qMEJRUg90Zv6/kFXo3//kJ35GobKnNE0L+Zvy/7TcXY9gGJT/BxinlAr6h9sVdqfQPHmF4arEmL/aQNO0ucBcgP3xG7S/+4DfV+1hp3FOe01/H5LjC6YPpCSk4urhbFLe1cM5f/pMfplCme7dGw9xYt9pRn0+OD/TorPU0W/o4/llJg+ejXeRgU5pWfvzXjauPghAvYY+JNwoqE9CfBoens7327QYZ5eCDEuPJ1ozd/aG0gv0/4lK7s6kJRRM00hLTKVSkT5VycOZ1IQUnD1dyMvL407GHeyLZLe8fL2xtrPhRvR1qpfDPNufl21nddguABoG1uRGXEEeIP5GiskVn8Imf/Ajvn5e9H+h+E1wWzYd5r3x/com4H/IxtWVrELTZrJSUrFxdS1WLulUFH/+uokWY97Fwqp8stiFOXs4kxpfEGdqQmr+lI38Mu6GfuNSqN84/EVWNPp0NLHnYvmo/yT0eXpupd5izrtf8uZnQ0o19t9+2cO2dYZjaq0AH5IKHVOT41NNrmwCuHk6kxxfcExNSkjFzXhMdXEruCn04ceC+WTkPMM2Xi74N66Vf2Nsk5AA/jx7hcAW9Uq1Ls4eRfbfhFQqFW0HYxmXe+y/qQmpLJ60gGdH9ce9qkehbQz7j629LU0fbkbs2cs071K608+2rd7D7vWGdqhRv/i5zaVIO7h4OpNS5NxW+Kr0vRz6/SiBrfyxtNRRydWJOoE1iT4Ti2fV0rth3HAMLYg9PbH4vlDJ3YW0f3AMtSlyDD0XHkXVOtVxdDW/aab5HtzZL+XmQU8PbQN6K6XcAYxTZPYBzxrX9wd2/817bAVeUkrZF3qPos4CnkqpEGMZK+MUH4CbgNM/KGdCKWUB+Giath3D1BhnwNEYz5uFyrkCh4AOSikP4xz4vhim/RT1B/C0Usrrbl2UUiWep9L5ybZ8tHAEHy0cQbN2QezdHI6maVw4FY2do22xg5mLRyXsHGy4cCoaTdPYuzmcpm0Nl+FPHoxi07LtDJs6EBtb6/xtsu5kk5Vp+K0VefgsFjoLqtX0piz06hPKd8vf5bvl7xLasSFb1xvqc/pkDA6Otv9qik3h+fX7d57Ct8Y95kyKv1S9vi9JVxNJjksiNyeXEzuO0SDYdNpGg+BAjm41ZOYid5+gduO6KKVIjksiL8/wWzblRjIJsTfKbV5nn34PsXzVRJavmkjHTk1Yv+4AmqZx8sRFHB3t7jnFZs4Xq7l1M9Pkhte7/rx0nfT02zRqUrs8wr+vSjX9uH0jnsyERPS5ucQdPIxnU9NMdXrMZaJ+WErjYYOxrlQxJ3mf+r4kXE0k6bqh3xzbcYzANqb9pmGbQA7/Zug3J3edoE6Tun85VSP0sbZ88POHTFg6kSGfD8WzumepD+QBuj7VlmmLRjBt0QhatA9it/GYej4yGntHW5MpNgCuxmPq+UjDMXX35nCaG4+phefXH94ZgU8tw3GzUav6xF66TtYdw3ztqOMXy+SYWr2+L4lXE0k2tsOJncdoEFJk/w0JJNy4/0YUaofMW7dZOGEuPQY+So2GBReR8/LyyEgz3OuTl5tH1IHTVK5x39vP/msPP9GWifNHMHH+CJq2C2L/FkM7XDwVjZ2DrckUGwAX90rY2ttw0Xhu278lnCZt/3qKmVtlF6KOngcM92NcOh2Dt1/pnieq1fcl6VpC/jH05M6j+Bc5hgYEB3L090MAc9Tr5AAAIABJREFUnNp9glr/8Bh6cscRmWLzP+CBzsxrmnZKKfUxsFMplQccA4b8H3t3HV9l+T5w/HPvrDuBERsNKxq2UYJB2AqSCigWJiqdKioNoqiESJeUGCAG0s2owWgYzToY692/P85cMEqW5/u73q/XXnDOc53nXPeT97nO/TwHmKuUGghEAi/fYx6/K6UaAPuUUmnAOmDYLTFp2Re7fqWUcsK4XL4EjgLzgBlKqWSMQ3zuFHcrA7AoO04BX2mt45RSnwHfKKVCMQ6h+URrvVopNQT4Jzv2N6312tu05ZhSagTGMfVmQDrGDwbht8Y+qPrBPhzeFcagbl9gZW1B36Hdc6aNfHkSY+YOAKDXh53z3JqyLvWCjHddWDR1NRnpmUz8cAZgvAi2z4AXSIi9weSPZqLMFC7uTrxeQtXJwJY+7Nl2nF7PjMPK2oKBH+d2st7oNoWZyz4EYNaXv7Lx9wOkpqTTrcMYOj7bjN5vtmfNsm3s3HwUg8EMBydbBn3S7U5vVSrmf/0urYJ9cHdx4PTu6YyZspL5yzeVdlr5GAwGnn67Ez8Mm2G8rVq7QMpX9eSP+euoXNsL32B/mnQI4scJi5jY5zNsHGzpPqwXAOdDz7Jp+d8YzM1QZmY8+27nfONwS0rL1gFs23KEZzoOx9rako8/65Mzrdvzn7Bs9WiuX4thzqx1VK1egR6dxwDQtcfDPNe5FWCsyrfv2LRYxwXfDzODgTovdiVk0lforCwqtmqOfaWKnF79M47VvCnXsD6nlq8mMzWVw9/MBsDazZWG/Uv2jkkGg4Hn3+3ErCHG7aZZh0AqVPVk/bx1VKnthX9zfwI7BrFk3CI+7/UZtg629BreK+f1Y3p+QsrNVDLTMwjdfoQ3xvfLdyecktIw2IeDO8Po38V4TH1jWO4xdUjvSYybbzymvvxR59xbUwbVzbmTzZJvfyH81GVQCo8Krrw66AUA7B1tebzbQwzvOxWlFA2CfWiUfcvKomQwGHjmnU58n73/Nm1vXA8bsvdfv2B/mnYIYtn4RYzvY1wPPbL33x1rtxF1OYq/Fm3gr0UbAOMtKC2tLfl+6AwyMzPRWZqaDWsT2PF+v/B+MAFBPhzZFcawHl9gaWXBy0Ny18MnfScxeo5xPbz4QWd+GLeU9NR0/APrEhBoXA8hWw6z9Ks1JMbdYNqQ2XjVrMQHk96g7bMtmTtuGaN6j0draNGxKVVqVLxtDg/KYDDw1FudmDf8O3RWFo3aBVG+qid/LVhHpVpV8AkOoHGHIFZOWMTkl8dg42BLt6G9AQgPPcuWH//CzNyAUoqn33kh5xialpLK6ZATPPteweKDKdFmUppXWt9zFIYwYfczzKasq2yXVdopFEptnyX3Dirjlmzpde+gMq5d5aK9KK2kDd5brHelLREdK6eUdgqF5mlj2scjgItJZf1L+btztTL9dXA9ueQv5C9qnat1KBO96Brdl5RYP+fM0h5los23KtOVeSGEEEIIIe6oDN8ysqRIZ74IKKW+AVrc8vQ0rfXc0shHCCGEEEL8/yCd+SKgtX773lFCCCGEEKJISWG+zN/NRgghhBBCCHEHUpkXQgghhBCmSe5mI5V5IYQQQgghTJVU5oUQQgghhGmSu9lIZV4IIYQQQghTJZV5IYQQQghhmqQwL5V5IYQQQgghTJVU5oUQQgghhGmSu9lIZV4IIYQQQghTJZ15IYQQQgghTJQMsxFCCCGEEKZJhtlIZV4IIYQQQghTJZV5IYQQQghhkrQU5qUyL4QQQgghhKmSyrwQQgghhDBNMmZeKvNCCCGEEEKYKqnMCyGEEEII06SkMi+d+f9xv1+yKu0UCq1njZTSTqFQlmzpVdopFFqP1gtKO4VC+333S6WdQqG0LGf6X6TaGHRpp1BocWmm33Fo5pFR2ikUynPrXUo7hUJr651a2ikUWudqpZ2B+Jd05oUQQgghhGmSMfMyZl4IIYQQQghTJZV5IYQQQghhmqQsLYtACCGEEEIIUyWVeSGEEEIIYZrkbjZSmRdCCCGEEMJUSWdeCCGEEEKYJjNVcn/3oJTqoJQ6oZQ6rZQacoeYLkqpY0qpo0qpJUWxCGSYjRBCCCGEEIWglDIA3wCPAZeAvUqpn7XWx/LE1AKGAi201rFKqXJF8d5SmRdCCCGEEKJwmgGntdZntdZpwDLgmVtiXgO+0VrHAmitI4rijaUzL4QQQgghTJJWqsT+lFKvK6X25fl7PU8qlYCLeR5fyn4ur9pAbaXUdqXULqVUh6JYBjLMRgghhBBCiHvQWs8CZhViFuZALaANUBnYopQK0FrHFSYv6cwLIYQQQgjTVHbGmFwGquR5XDn7ubwuAbu11unAOaXUSYyd+72FeeOyswiEEEIIIYQwTXuBWkqpakopS6Ab8PMtMT9hrMqjlHLHOOzmbGHfWCrzQgghhBDCNN3HLSNLgtY6Qyn1DrABMAA/aK2PKqU+BfZprX/OntZOKXUMyAQGaq2jC/ve0pkXQgghhBCikLTW64B1tzw3Ks//NfBh9l+Rkc68EEIIIYQwTapsVOZLk4yZF0IIIYQQwkRJZV4IIYQQQpimMjJmvjRJZV4IIYQQQggTJZV5cUdXDx4lZMFKdFYW1du2wPeZdvmmR4Sd4sCCVcRduEzz916mSmCjnGnLe7yDk1dFAGzdXGk98M0Szf1fWmtmTf6J/TvCsLK25P1R3ahZt3KBuAXfruOfdfu4kZjMis1jc54PDTnD7KlrOX/6KoM+e5EWj9QvyfQBOLE3jF9mrEZnapp2DKJN10fzTc9Iy+DHiYu4fOoSto62dB/WG9cKblw8Hs7qacsB0BoefakD/i3qlXj+9zJj4ht0fKQhkdEJNHlsUGmnk0NrzfKv13BkVxiW1hb0GdId79pVCsSFn7jI3HFLSU9NJyDIh67vPodSin2bDvLLvN+5Fh7B0O/6U7WuFwDH9p1g9axfyUjPxNzCQOc3n6Zuo1rF0oZT+8JYP3M1OiuLRu2DaNXlsXzTM9IzWD1pEVdPX8TGwY4XhvbGpbxbzvS4iBi+eXMsbXp2pEWnh4m6dJ0V4+bnTI+9GkXblx4n+Nk2RZaz1pofv17D0d3G5d5rcHe87rDcF4w3Lne/QB+6ZC/3pIQkvv90AdHXYnCr4Mqro3tj52DLoW1H+GXuepRSmBnMeOGd56gZUJ3oazHMHPUDOkuTmZFJm+db0frpFmWyDScPnua7EXNwr+AKQINW9Xiid3sANq7czLbfdoHWtHgymEc6P1SoNtytbV9PWMuu7WFYW1sy5JOu1PYpeEz9fvp6Nvy6j8SEZH7f8UXO89MnreXA3jMApKakERtzg9+2flYsud5JUHlnPqhfHTOl+PncdRaevJRv+nPVKtCphidZWpOckcnYkNOcT0wGoKajLYMb1cTOwkCWhlc2HiQtS5do/tcOHeXwwhXoLE3VNs2p83T7fNOjwk5xaNFKEi5cptk7r1Apz7n5ZlQMIbMXkRwTCyiaD3obOw83TJoU5qUzXxYppW5ore1v8/xE4CkgDTgDvFzYXw27k6ysLPbN/ZG2w97Fxs2ZP4dPoFLjAJwqe+bE2Lq7EvjmSxz/7a8CrzdYWtBh3LDiSO0/2b/jOFcuRjFz1VBOhF7gu/GrmDz3/QJxzVr58WSXlrzRaWy+5z0quNB/VDfWLNpUQhnnl5WZxdpvVtJ3bD+c3J2Z/u4UfIL8Ke9dISdm74Zd2NjbMnDeCA5tCuH3Ob/QY3gfylf15J3pH2EwGEiIjmdav4n4BPlhMBhKpS13snDFZmbM38D3U98q7VTyCd0dxvVLkXy2eBjnjoWzeOpKhn33QYG4xVNX0mtAF6r5evPV4FmE7jlOQKAPlap50u/TV1g0+cd88fZOdrzzxas4uztx+exVpg2ayYSVHxd5/lmZWfz27Qp6ff4Wju7OzOo/mTpBAZTzyt12QjbsxMbehvfnjOTI5hD+/OEXugztkzN9w+yfqNnEN+exe+Xy9Js+KGf+k3uNwie4aD8gHt0dRsTlSD5ZNIxzYeEsnbqSwbdZ7ku/XEnPAV2o5uPN9CGzOLrnOP6BPmxY8jd1G9WifY9H2bDkL/5Y8jfPvfEUdRrXpl4Lf5RSXDpzhe8/mc/HC4bi5ObIwOn9sbA0JyU5lTEvj6dec3+c3Z3KXBsAagZU5+2xr+Wbz+VzV9n22y6GfPcBBgsDXw+aSUCwL+UqeTxwG+5k97bjXLoQyeK1Qzh25AJTv1jFdwsLHlODW/vyXNcW9HxmXL7n3xnwTM7/Vy/dxqkTt/6mTvEyAwY0qMF720KJuJnG3IcbsPVqdE5nHWDDxUjWnLsGQCtPV96vV50Pth/FoODjZnX4eO9JTscn4WhpTkYJd+R1VhaH5i2n5dD3sHF15p+R4/FsVA/HPOdmG3dXmrzxEqduc27eN2M+dZ7pQPkAHzJSUkDJAI3/BbIWyxBldLd18ifgr7WuB5wEhhZXLjGnz+NQwQP78u4YzM3xCm7M5X2H88XYe7jh7F2pTF9JvmtLKA8/3hilFHUDvElKTCYmKqFAXN0Ab1zdHQs8X76iK9VqVUSV0pi8iyfCcavojpunO+YW5tRv05BjO4/kizm28wiNHmsKgH+r+pw+eAqtNZbWljkd94z0jDK7mrbvOU5M3I3STqOAg9tDCW7fFKUU1f2qknwjmbjo+HwxcdHxJCelUN2vKkopgts35eA24/rx9C5PBa9yBebrVatyTkexYrUKpKWmk56WUeT5Xz4ZjmtFD1yztx3/1o04fsu2c3xXKA0ebQaAb8v6nDt0EuOd0yBsx2GcK7jl6/zndfbQSVwquONc3rVI8z60PZSgdtnL3bcqN5OSib9lucdHx5OSlEJ1X+NyD2rXlEPZy/3QjlCC2hv3h6D2TTm43fi8tY0VKnsnSEtJy9kfzC3MsbA01rUy0jJy2l8W23An18KvU83HO2efr12/Jge3HL7rax7U9s1Haf9kE5RS+NXz5kZiCtGRBY+pfvW8cfMoeEzN6+/fD/BIh4bFkued+Lo6cCkphStJqWRozZ+XImldMX9l+mZGZs7/rQ1mgHGbaFbehdPxSZyOTwIgIS2DrBLL3CjmzHnsyntgV84dM3NzKgc15ur+Q/li7DzccPKqXKCjnnDpKjozi/IBPgCYW1tjbmVZYrkXF22mSuyvrJLKfDFQSo0DLmqtv8l+/DFwA+OXQV0AK2CN1nq0Uqoqxh8R2A00Bh7Pfs1UoB1wDeimtY7UWv+R5212AZ2Lqw3JsXHYurnkPLZxcybm9Pn7fn1megYbho3HzGCGz9PtqNy05IenAERHxONe3jnnsVs5J6Ij4m/bcS+LEqLjcfLIXQ9O7s5cPB6ePyYqHufsGIPBgLWdNTcTkrBzsufC8fOsnLyMuIgYugx6scxV5cuyuMh4XDxytx0XD2fiIuNxdnO6JcYpT4wTcZH5O213E7L5EF61KuV0JotSQnQ8Tu65+Tu5O3PpRP5tJzE6Dsc8246VrXHbMbe0YNvKv+n1+VvsWLXxtvMP3RxCQJtGt51WGHFR8biUy7Pc3Z2Ji4rHKe9yj4rHOc9yd/ZwIi7KuNwTYxJzYh1dHUmMScyJO7j1MD/N/o3EuBv5qtsxEbF8O3Q2EZejeP6NpwtVlS/uNpw7dp7P+k7Eyd2RTm8+TcVqnlSs5snPc9ZxIz4JSysLQncfw7tOwWE9RSEyIh6PCrlt8yjvRGRE/D077re6diWGq1diaNi0ZlGneFceNpZE3EzNeRyRnIqfq0OBuE7VPeleqyIWZma8s9X4YcrL3gat4cuWfrhYWvDnpUgWnSzZbxZSYuKwyXtudnUh5sz5+3rtjWvXsbC1YdfUmSRFRlPOvy7+3Z5FmUld19TJGiweyzF22v/VBYgEagHNgAZAY6VU6+zptYBvtdZ+WutwwA7jr4X5AZuB0bd5j1eA9cWUf6E99fUY2n8xmOB3XiZkwUoSr0eWdkr/L3nVrcqHs4fwztcfsmnZX6SnpZd2SiLblXNXWTXrV178qMu9g0vYpsXrCX62DVY2VrednpGewYndofi1bFDCmf03Sql83xw2aFWPjxcM5c0xr/DzD7m/6+JazoURcwbx6aLh7PpjLwl5Os+lLW8bqtSqzGfLRjFizkDaPteKGSN/AIzfArXr9jBfDZzB14NnUrlmpTLfQdu44SAPPVIPg6Fs5rnq7FU6b9jPN6Hn6VPX+MHIoBT13R0ZvecEr28+zEMV3WjiUbgPfiVJZ2YRdeI0AT070XbMYJIiogjfsrO00yo8M1Vyf2WUVOaLgdb6gFKqnFKqIuABxAIBGCvtB7LD7DF24i8A4VrrXXlmkYXxAwHAImB13vkrpYYDGcDi272/Uup14HWAJ4b3p/HzT/znNti4OHMzOjbncXJ0HDYuznd5RX62rsZY+/LulPOtRdz5iziUL/rxm7fz24ptbPhpNwC1fKsQdT33soLoiHjcypnOwdfRzYn4yNz1EB8Vh+MtVUNHdyfiImNx8nAmMzOTlKQUbB3t8sWU86qApY0V189fpXJtrxLJ3RT9s2YbW381ntyq1vUiNjJ324mNjMtXSQVjNTU2TyU+NjK+QMztxEbE8e3IubwytAflKrkXUfb5Obo5ER+Vm398VBwObvlzc3BzJiEyFid347aTetO47Vw6Ec6xbYf484efSUlKRimFuaU5gU8Z6w+n94XhWaMy9i5F8w3XpjXb2P6bcbl71/UiNiLPco+KK1Apd3bP/w1IXGR8ToyDqwPx0cYqeHx0PA4uBS4/olb9GkRdjeZG/A3snezzzbdi1QqcPnKGRg/9tw8qJdEGGzvrnHj/IF+Wfrkypw0tngiixRNBAPw0+7d83xgV1prl2/l1tfGYWtevCpHXctsWeT0ejwc4pm7ccJD+Q54vshzvV2RyGuVscz+klrOxIjI57Y7xf16MZFDDGozhFBHJqRyIiic+e1jcjmux1HGxZ99/+DausKxdnUnOe26OicXG5f6Wv42rC87elbErZzzmVGxcn5jT56BNcWQqSlLZ/Ej8v2EFxmEwXTF2zBUwVmvdIPuvptZ6TnZs0j3mlTOIUynVB3gS6KnvMLhTaz1La91Ea93kQTryAK41vEm8FsGNiCgyMzK4sHM/lRoH3Ndr027cJDPdWAFOTbhB1MmzOFbyvMeris4TL7Tkq8Uf8dXijwh6yJ+N6/ajteb4kXBs7a1NZogNQOU6XkRfjiLmWjQZ6Rkc2nQA3yD/fDG+Qf6E/LkXgNCth6hRvxZKKWKuRZOZaRz7GXs9hsiL13Ep4vHN/2vaPteSUXMGMmrOQBq09Gfnhr1orTl79Dw2djb5htgAOLs5YWNnzdmj59Fas3PDXhq08L/D3I1uJibz9dDZPP/6k9QMqF5sbalY24uYK5HEZm87oVtCqHvLtlMn0J+Df+0B4Ni2Q1SrZ9x2+k58nw/mjeaDeaMJeuYhWnV9LKcjD3Bk834CHiq6ITZtnmvJ8O8HMvz7gdRv4c+uP7KX+zHjcne6Zbk7uTlhbWfN2WPG5b7rj73Uz17u9Zr7s2uDcX/YtWEv9Zsbn4+4HJkzHv7CyYtkpGdi52hHbGQcaanGzlxS4k3OhJ6jfJWC1zqUhTbExyTktOF8WDhaa+yyP7gnxBq/TYi5HsvBrYdp+mjj/9yGO3muawvmLP+QOcs/pGVbPzb8ug+tNUcPh2Nnb/2fh9iEn4sgMSEZv/reRZbj/QqLTaSKvQ2etlaYK8VjlT3YeiUmX0wV+9wPTS08Xbl4w3hx7O7rsdR0tMPKYIZBQSMPJ84l3CzR/F2qe3PjWgRJEVFkZWRwadd+PBvf30XoLjW8Sb+ZTGqCcVuJOHYChxI8N4vio4riYh9RkFLKD5gNuAMPYazMjwEe0VrfUEpVAtIBW+BXrbV/ntdqoLvWeplSagRQXmv9rlKqAzAFeEhrfV/jVkaH/PXAK/jKgVAOLFhFVlYW1dsE4/dcB46s+BXXal5UalKP6DPhbJsyi7SkmxgsLLB2cuDxSSOJOnmWvd8vRSmF1praHdtSo23zB02DnjVSHvi1WmtmTFxNyM4TWFlb8P7IbtTyNX5l+l7PyXy1+CMA5n71C5v/OEBMZAKuHo60ezqQHq+35+SxC3wxaB43EpKxtDTH2c2Bb5f/t9snHo0t3Dj143uO8euMNWRlZdGkXSAP92jHH/PXUbm2F77B/qSnpfPjhEVcOX0ZGwdbug/rhZunOyF/7WXT8r8xmJuhzMx4pGc7/Jo/2J1HerReUKg23M38r9+lVbAP7i4ORETFM2bKSuYv31Tk7/P77pf+U7zWmqXTVhG65ziWVpb0Gdwt5/aSn/adyKg5AwE4f/wC88YtJS0tHf9mPnR//3mUUhzYepil01ZzI/4GNvY2VKlZif4T3+S3BX+wfsnf+Sry/Se9iaNLwXG7eV29+d9rLyf3HuX3mcZtp2G7IB7q1o6NC9dRsVYV6gYFkJ6WzupJi7h25hI2DrZ0HtwbV8/83xT8s2g9ljZWtOj0MABpKalM6f0x/X8YhbWdzX/Kp5zNvS8X1FqzbNoqju01Lvdeg7vhXce43D9/dSLDvzcu9/ATF5g/binpaen4NfOh63vG5X4jPonvP5lPTEQsruVdeG10b+wc7diw9G92b9iLwdyAhZUFz7/5NDUDqhO27wSrvluLsd6ieejZlrR66sGPV8XZhk1rtrJl7XbMDMY2dH7rGWr4VwNg0ntfkZRwE4PBQOe3nqFu49p3zM/HOfOO0+6nbdPGrWHPDuMxdfDHXanrZzym9u06hTnLPwRgxpe/8tf6A0RHJuDm4cgTzzXj5TeNt1CcO2MDaakZvPH+gxWbnlvvcu+guwiu4MIH9apjpuDX89eZd+ISr/l6cTz2BluvxvBB/eo0LedERpYmMS2DSQfPci7R2GnvUMWDXnUrozXsvBbL9NDzD5RDW+/UewfdwbWDoRxeaLxttPdDwdR9tiPHVv6CczVvKjauR8yZ8+yaOov0m8Zzs5WTI49NGAnA9SNhHFm8CjQ4V/Oi0as9MDN/sEEaY5s8UibGnVQdsb7EOrLnP+tYJtp8K+nMFyOl1BEgSmvdNvvx+8Cr2ZNvAC8CmRTszN8AZmEclhMBdNVaRyqlTmO8eDY6O3SX1vquN3AvTGe+rChMZ74sKGxnviwozs58Sfmvnfmy5kE682XN/XTmRfErTGe+LChsZ74sKExnvqyQznzZIWPmi5HWOuCWx9OAabcJ9b8lruAgT+PzJXvZvxBCCCFEWWb6dY5Ck0UghBBCCCGEiZLKvBBCCCGEME1l9RcRS5BU5oUQQgghhDBRUpkXQgghhBCmqQz/mFNJkcq8EEIIIYQQJkoq80IIIYQQwjRJZV4q80IIIYQQQpgqqcwLIYQQQgiTpOVuNlKZF0IIIYQQwlRJZV4IIYQQQpgmKUvLIhBCCCGEEMJUSWVeCCGEEEKYJhkzL5V5IYQQQgghTJV05oUQQgghhDBRMsxGCCGEEEKYJvnRKKnMCyGEEEIIYaqkMi+EEEIIIUyTVOalMi+EEEIIIYSpksq8EEIIIYQwTVKYl878/7rgcumlnUKhhUSZ9mb6lLdtaadQaL/vfqm0Uyi0DoELSzuFQtl5oGdpp1Bok0IdSjuFQute/WZpp1BoR2MNpZ1CoXwalFDaKRSajUGXdgrif4hp95KEEEIIIcT/W1rGzMuYeSGEEEIIIUyVVOaFEEIIIYRpUlKZl8q8EEIIIYQQJkoq80IIIYQQwjTJmHmpzAshhBBCCGGqpDIvhBBCCCFMkxTmpTIvhBBCCCGEqZLKvBBCCCGEMElmUpaWyrwQQgghhBCmSjrzQgghhBBCmCgZZiOEEEIIIUyS/GaUVOaFEEIIIYQwWVKZF0IIIYQQJkkq81KZF0IIIYQQwmRJZV4IIYQQQpgkJaV5qcwLIYQQQghhqqQyL4QQQgghTJIU5qUyL4QQQgghhMmSyrwQQgghhDBJUpmXzrwAwvaEsfqb1WRlaYIeD+Kx7o/mm56RlsGi8Yu4ePISdo629B7ZG7cKbgD8ueRPdq3fjZmZ4vl3nsenqQ8A/6zcxK51u0BBxWqe9BjUAwtLC7b8tJXNqzYTdSWKz1d/hr2TfZG359S+MNbPXI3OyqJR+yBadXksf3vSM1g9aRFXT1/ExsGOF4b2xqW8W870uIgYvnlzLG16dqRFp4eJunSdFePm50yPvRpF25ceJ/jZNkWe++1orZk4dhnbthzB2saSTz5/GR9f73wxycmpDP5wJpcuRmJmpmjdpj7vfdgJgEnjlrNvz3EAUlLSiIlJZMuur4o95+Vfr+HIrjAsrS3oM6Q73rWrFIgLP3GRueOWkp6aTkCQD13ffQ6lFPs2HeSXeb9zLTyCod/1p2pdLwCO7TvB6lm/kpGeibmFgc5vPk3dRrWKtS33Y8bEN+j4SEMioxNo8tig0k4nh9aaeVN/4sDOMKysLek3ohvV61QuEHf2+EW+/WwZaanpNAz2oc8Hz+ZcVLZ+xVb+WLUdM4MZDZv78OLbT5GRkcnMsT9y7sQlMjOzaN2xCc/1eqTY2xMfGsqF5cvRWVl4tGyJZ8eO+aZf+/NPIrdtQ5mZYe7gQLXevbFyy923M5OTOTJ6NC4NGuDdo0ex5wvG4+tP32YfXzsG8chtjq9Lxi/i4inj8bXXiN64VnAjKT6JeZ/O5eKJCzRt34xO73YGIOVmCtM/yN1/4yPjafRoY5576/kizVtrzYrpazi627gPvzSoO1632YcvnLzIwvFLSUtNxy/QhxfeMe7DSQlJ/DBmAdHXYnCr4ErfUb2xdbBlz1/7+XPZ36DBysaKbh90pnKNSgCM7P4p1rbWKDOFwWDG4BkfPXD+x/aEsWr6GrKyNMGPB9KuR/7lnp6WwcJxi3POay+P6o1bBVcA/ljyFzvXGc9rnd99Hp+mdQG4eSOZpZMB6nbDAAAgAElEQVSWceXcNZSCngO7U82vKuvm/c6O33Zh72wHwFN9n8AvyPeBc78drTVLv1rDkd1hWFpZ8srQ7njXLrgvnz9xkbljl5KWlk5AoA/d38s+pv5zkJ/nbeBqeATDZ/Snal3jutz15342LPsn5/WXzlxl5OwP8apVqUjzF8VLOvNFQCl1Q2td9L3SEpCVmcWKr1by1oR+OHs4M/mtKQQE+1OhaoWcmJ3rd2Fjb8vIhSMI2RjCL7N/oc/IPlw7f42Qfw4wdM4Q4qPj+Wbgt4yYP5yEmAS2rNnC0B+GYGllydxP5xGyMYTADoFU96uGX5Av0z+cXmzt+e3bFfT6/C0c3Z2Z1X8ydYICKOeV256QDTuxsbfh/TkjObI5hD9/+IUuQ/vkTN8w+ydqNsk9ELtXLk+/6YNy5j+51yh8gusVS/63s31rKBfCI1i7/nOOHD7L2E8Xs2DZsAJxL/VpR9PAuqSnZfBG38ls33qEFq0CGDCka07MssV/czzsYrHnHLo7jOuXIvls8TDOHQtn8dSVDPvugwJxi6eupNeALlTz9earwbMI3XOcgEAfKlXzpN+nr7Bo8o/54u2d7Hjni1dxdnfi8tmrTBs0kwkrPy729tzLwhWbmTF/A99Pfau0U8nn4M7jXLsUxbQfh3Lq6AXmTFzF59+/XyDu+4mreH1IF2r5eTHuo+85uOs4DYN9CN1/mn1bjzJhwQAsLM2Jj0kEYNfGQ6SnZTBp0UBSU9L4qMcEWjzWkHKersXWFp2VRfiSJdT+4AMsXVw49sUXONevj03FijkxtlWq4DtsGAYrKyI2beLiqlXUfP31nOmX1q7FoXbtYsvxVlmZWaz+eiVvju+Hk4czU9+egl9zfyp45x6Pdq/fhY2DLcMXjODAPyH8OvsXeo3sg7mlOR37PM6181e5ev5qTry1rTUDZuZ+YJzSbxL1WtYv8tyP7g4j8nIkHy8cxvmwcJZ9uZJB3xbch5dNXUmPj7pQ1cebb4fO4tie4/gF+vDH0r+p07AW7Xo8yh9L/uKPpX/z7OtP4V7BlQ+mvoOtgy1Hd4exZPKP+eb7/pS3Cl3kycrMYsW0Vbw98U2cPZyZ2G8qAc398bzlvGbrYMPoRcPZvzGEtbN+4ZVRvbl6/hr7Nx5g2A+Djee1Ad8xcsEwzAxmrJq+Gp+mPvT9+GUy0jNIS03PmV/bzg/xSNe2hcr7bo7sDiPiUhRfLB7G2WPhLJqykuEz+heIWzRlJb0GdqG6rzfTBs0mdPdxAoJ8qFjNk7fGvMyCySvyxQc91pigxxoDcOnMFb4ZMdfkOvJKBozLmPn/78KPh+NRyR33iu6YW5jTqG1Djuw4ki8mdMcRmrVrCkD9h+pzMuQUWmuO7DhCo7YNMbc0x83TDY9K7oQfDweMB9P01HQyMzNJS0nDyd0JgMq1KudU9YvD5ZPhuFb0wNXT2B7/1o04vjN/e47vCqXBo80A8G1Zn3OHTqK1BiBsx2GcK7jl6/zndfbQSVwquONcvvg6LbfatPEgTz4dhFKKevVrkJh4k8jIuHwxNjZWNA00Vo8sLM3x8fXm+rXYAvP6fd1eOjzerNhzPrg9lOD2TVFKUd2vKsk3komLjs8XExcdT3JSCtX9qqKUIrh9Uw5uM64rT+/yVPAqV2C+XrUq45y9LVWsVoG01HTS0zKKvT33sn3PcWLibpR2GgXs3RpK6w6NUUpR29+bpBvJxEYl5IuJjUogOSmF2v7eKKVo3aExe7eEAvDnmh0889LDWFga6z5Org4AKCA1JY3MjEzSUtMxtzBga2ddrG1JOncOq3LlsPbwwMzcHNemTYk9dChfjGPduhisrACwq16d9NjcfSApPJyMhAScfIu2Yno3F06E417RHbfs42vDNg0J3V7w+No0+/har3V9Th0wHl+tbKyoHlAdc8s719wiLkVwI+4G1QOqF3nuh3eEEviYcR+u5mvch+Nv2Yfjo+NJuZlCNV/jPhz4WFMOZbfv8PZQAtsb2xXYvimHsvft6v7VsHWwBaCarzdxkfnnWRTCj1/APc95rfHDDTmyIzRfzJHtoQS2Mx4LG+Q7r4XS+OGGWFia4+7phnsld8KPXyD5RjKnD58l+PFAAMwtzLG1tyny3O/k4LZQgts3QSlFDb+q3LyRTFx0/n05LjqBlJup1Mg5pjbhQPZyr1j19sfUvPb8fYCmDzcstjaI4mMSnXmlVC+l1GGl1CGl1EKlVFWl1Mbs5/5WSnllx81TSn2llNqhlDqrlOqcZx6DlVJHsucx7i7vtUkp1ST7/+5KqfPZ/++jlFqtlPpdKXVKKTXhNq91V0rtVEo9oZRqkz2vlUqp40qpxSr7e2ul1CNKqQPZ+fyglLJSSjVVSq3Onv6MUipZKWWplLJWSp3Nk9t4pdQepdRJpVSrwi7b+Kh4nD1cch47ezgTH3VLpysqHpdyxhiDwYC1nTVJCUkFXuvk7pz9nDNtX2jLx90/YeQLo7Cxt6Fuk7qFTfW+JETH4+TunC+nxFtOQInRcTh65LbHytaamwlJpCansm3l37Tp0eGO8w/dHEJAm0bFk/wdRETEUr5C7oeHcuVdiLwed8f4xISbbNl0iGZBPvmev3IlmiuXonI6/cUpLjIeF4/c9eDi4VzgpG2MccoT4/SfTuwhmw/hVatSTkdTFBQbGY9b+dz14ObhRMwtyzgmMh7XcrkxruWcic2OuXoxkuOHzjL81Wl8/NY3nD52AYDAh+tjZW3JG09/wtvPfcaT3dtg72hbrG1Ji4vD0jV3P7B0ds7XWb9V1LZtOPn7A8aq/sUVK6jywgvFmuOt4qPicS53y/H1Nh1iZ4+Cx9f7ceCfEBq0aVgs99k25p67XTh7OBN3m3ODc5592NnDKef8kRibiJObcZqjqyOJsYkF3mPHut345TkeKaWYPnAG496YzLZfdzxw7nFRcbjkzd294LElb/sMBgM22cv91mOXsd1xRF+Lwd7JnkUTljL+9UksmbSM1OTUnLgtP21l7KsTWDxhKTcTbz5w7nduU0K+/fT+jqnOxN3y4f1u9v5zkMBHTK8zr1TJ/ZVVZb4zr5TyA0YAD2ut6wPvA18D87XW9YDFQN4BwJ5AS+BJYFz2PDoCzwCB2fMo0BG/Tw2ArkAA0FUplTOAUClVHvgNGKW1/i376YZAf8AXqA60UEpZA/OArlrrAIxDnfoBB7LnD9AKCAWaAoHA7jw5mGutm2XPd/TtklRKva6U2qeU2rdu8foHbOqDu5l4k9AdoYxePIoxP35KWnIqe//cV+J5/FebFq8n+Nk2WNlY3XZ6RnoGJ3aH4teywW2nlwUZGZkMHTibbj0foXIVj3zT/li3h0faNcJgKPO7/T1dOXeVVbN+5cWPupR2Kv/TMjOyuJFwk89mv8eL7zzFlyMXorXm9LELmBkUM34ezdcrh/Hrss1cvxxd2unmiNq1i6TwcCq0awdAxKZNOPn7Y+nico9XmpaD/xygYduSLS48CHWbntDJA6fYsX4Xz7z2VM5zH057lyGzBvD2uNfZ8tN2Th06U9Kp3lFWZiaXTl2i1dMtGDxrAJbWlvy59G8AWj7dgtGLRjB41gAc3RxZ893aUs72vzt7LBxLKwsqVfcs7VTEAzCFktbDwAqtdRSA1jpGKRUM/Hu1z0Lyd85/0lpnAceyO9gAjwJztdY3/53HA+byt9Y6HkApdQzwBi4CFsDfwNta68154vdorS9lxx8EqgKJwDmt9cnsmPnZr/tSKXVGKeUDNAOmAK0BA7A1zzxXZ/+7P3t+BWitZwGzAH6/tF7frUFO7k7EReZWt+Ii43KGxPzL2d2J2IhYnD2cyczMJCUpBTtHuwKvjY8yvvZEyElcK7hi72wc91ivVT3OHTtH08ea3C2VIuHo5kR8VG7VOj4qDge3/O1xcHMmITIWJ3dje1JvpmDraMelE+Ec23aIP3/4mZSkZJRSmFuaE/hUawBO7wvDs0Zl7F0ci70dy5f8w5qVWwDw86/G9Wu5m2zE9Vg88lRb8/rs44V4eZejZ69HC0zbsH4vQ0YU30V//6zZxtZfdwJQta4XsXmGAsVGxuWr4IGxihebp7IUGxlfIOZ2YiPi+HbkXF4Z2oNyldyLKPv/HRtWbePvn42f/2vUrUJ0nm9xoiPjcb1lGbt6OBETkRsTExGXU91zK+dEs4fqoZSipq8XZkqRGJfE9j9CaBBYF3NzA06uDtQJqMrZ4xcpX6n4htBZOjuTFpO7H6TFxWFxm855/LFjXF23jroDBmBmYQHAjbNnuXHqFBGbN5OVkkJWZiZm1tZUeb5oLxq9lZO7E3ERtxxfbzkeObkZj6O3Hl/v5fKZy2RlZlHlNhelPqjNP21j+2/Gfdi7jhdxebaLuMi4nCFu/7q14h0XGZ9z/nBwcSA+Oh4nNyfio+NxcM4dB3/5zBUWT1rOW+Nex94pt63O2RVxBxcH6rcMIPz4BWrVr/Gf2+Hs7kxs3tyjCh5bjOsmDpfs5Z6cvdyNx6Vb2+2Ms4czzh5OVPUx3nygQev6OZ15x+zhZwDNnwhm5rDZ/znn29m4Zhtbf90FQNU6VfLtp/d3TI3D2f3+zld7Nh6g2SNl/4Ph7ZiV4Yp5STH9El1BqXn+/yCrOIPc5XLrINC8884k98NQBsbOdfv7jL+TLUBHIB34C+M3DC3J35n/d573M7978qrrReTlKKKvRpORnkHIPwfwb+6fL8Y/2J89f+wF4NDmQ9RqWAulFP7N/Qn55wAZaRlEX40m8nIU3nW9cSnnTHhYOGkpaWitORlyigpe5W/39kWuYm0vYq5EEnvN2J7QLSHUDcrfnjqB/hz8aw8Ax7Ydolo9Y3v6TnyfD+aN5oN5owl65iFadX0spyMPcGTzfgIeKpmDXdcebVm2ejTLVo+mzSMN+PXnXWitOXzoDPb2Nnh4FOzMfzNtDTcSk/Nd8Pqvc2evkpBwk3oN/vuJ8X61fa4lo+YMZNScgTRo6c/ODXvRWnP26Hls7GxwvqUT4+zmhI2dNWePnkdrzc4Ne2nQwv8Ocze6mZjM10Nn8/zrT1KzGMYJ/y9o36klE+Z/xIT5H9G0tT9bft9v3A9Dw7G1s8bllpO7i7sjNnbWnAwNR2vNlt/307SVcT00be3PsZDTAFy5EElGRgYOzna4l3chdL/x+ZTkVE4dvUBF77uPxy0su6pVSY2IIDUqiqyMDGL27sWlfv4LP5MuXCB80SJqvf02Fo657azx6qvUHz+e+mPHUuWFF3APCir2jjxAlTr5j68HNhU8vvo192dv9vH18JZD1GxQ676GzRzYGELDh4v2ePTQsy0ZNnsgw2YPpH5Lf3b/adyHzx0z7sO3+yBibWvNuWPGfXj3n3upl92+gOb+7N5gbNfuDXupl71vx1yPZdboufQe2pPyVXK3mdTkVFJupuT8P2zfCTyr3f7apXvxqluFyMuRRGUv9/0bDxAQ7JcvJqC5P7v/MJ4HDm4+RO2GNVFKERDsx/6NB0hPyyDqajSRlyPxruuFo6sjzuWcuX4hAoCTIafwzL6QOe/QqUNbD+NZrWiq2w8/15LRcwYwes4AGrYKYOeGfWitOXP0PDZ21ji75d+Xnd0csba14kzOMXUfDVre/ZgKkJWVxb5/DtLMBIfYCCNTqMxvBNYopaZoraOVUq7ADqAbxqp8T/J3dm/nT2CUUmqx1vqmUsr1LtX580BjYA/Q+Q4xt9LAK8AKpdRgrfX4u8SeAKoqpWpqrU8DLwH/VvO3AguABVrrSKWUG1Ae45CbYmEwGOj0bie+GzyDrKwsgjoG4lnVk3Vz11GljhcBzf0JejyIRWMXMealz7B1sKX3iF4AeFb1pGGbBnzxylgMBjM6v9sJM4MZVX2qUr91fSa+OQkzgxmVa1am+RPNAdi8ejN/L99IYkwi41+bgG8zX7oP6Fak7Xm8XycWjviOrKwsGrYLopy3JxsXrqNirSrUDQqgUfsgVk9axLS+Y7BxsKXz4N73nG9aSipnDpzgqXcLdpSLW8vWAWzbcoRnOg7H2tqSjz/rkzOt2/OfsGz1aK5fi2HOrHVUrV6BHp3HANC1x8M819l4WcWG9Xtp37FpsYytvZ2AIF9Cd4cxvOfnWFpZ0mdw7jr+tO9ERs0ZCECP/p2YN854GzX/Zj74BxrH+R/Yepil01ZzI/4GXw+dTZWaleg/8U3+WbOViMtR/Dp/A7/O3wBA/0lv4ujiUDCJEjT/63dpFeyDu4sDp3dPZ8yUlcxfvqlUcwJo2NyHAzvDeP+FsVhaW9BveO56GNR7MhPmG2/913dAJ779bBnpqek0CK5Lg2DjOOa2Tzbju8+X81HPiZhbGHhrRHeUUrTv1IJvP1/GRz0noDW0eaIp3jUr3jaHoqIMBry6d+fEl19CVhbuLVpgU7Eil9euxdbbG5cGDbi0ciWZqamcnjkTACtXV2q9806x5nU3BoOB59/txKwhxuNrsw6BVKjqyfp566hS2wv/5v4EdgxiybhFfN7LeHztNbxXzuvH9PyElJupZKZnELr9CG+M75dzJ5yDmw/y2hev3+mtC80v0Jeju8P4+MXPsbS25MVBudvOF69NZNhs4z7ctX8nFo433l7Wt5kPftn7cLvujzDn0/nsWL8b1/Iu9B1lPM6uX7iBpIQklk1bmb2MjLegTIxNZNaouQBkZmbS9JHG+DXLf93P/TIYDLzwbie+HTwTnZl9XqvmyW9z1+NVuwoBLfwJfjyQBV8s5pMXP8fWwZaXR74EgGc1Txq1acAXL4/DzGDGC+91xix7aOIL73Zi/hcLyczIxM3TjRcHdQdg7cxfuHTmCkqBa3lXun1Y9NdmBAT5cGRXGMN6fIGllQUvD+meM+2TvpMYPWcAAC9+0Jkfsm/36x9Yl4Ds9RGy5TBLv1pDYtwNpg2ZjVfNSnww6Q0ATh46i2s5ZzwqFt83a6J4qX/v4lGWKaV6AwMxVqMPYBwrPhdwByKBl7XWF5RS84BftdYrs1+Xc8tIpdQQoBeQBqzTWhe8t58xri7wY/Z7/Qa8qLWuqpTqAzTRWr+THfcrMElrvenf91FKWQE/A2uBY8AArfWT2fHTgX1a63lKqUeASRg/TO0F+mmtU5VSNkAc8JTW+g+l1Cyggtb66ex5bMqe5z6llHv2/Krebdnda5iNKYhLNe3v0J7yLt4LA0vCvsiyd6eW/6pD4MLSTqFQdh7oWdopFNqk0NL90FUUulcv+osbS5qVmWmfFrJKO4EiYGMw7XUA0KrCE2Xi5Oz7w5YSW5jHXmldJtp8K5PozIsHJ5350ied+bJBOvOlTzrzZYN05kufdOaLjnTmTWOYjRBCCCGEEAWU5VtGlpT/t515pdQ3QItbnp6mtZ5bGvkIIYQQQgjxX/2/7cxrrd8u7RyEEEIIIcSDK6kbO5Rl/4u3phRCCCGEEOL/hf+3lXkhhBBCCGHalJSlpTIvhBBCCCGEqZLOvBBCCCGEMElKldzfvXNRHZRSJ5RSp7N/3+hOcZ2UUlop1aQoloF05oUQQgghhCgEpZQB+AboCPgC3ZVSvreJcwDeB3YX1XtLZ14IIYQQQpikMlSZbwac1lqf1VqnAcuAZ24TNwYYD6QU1TKQzrwQQgghhBD3oJR6XSm1L8/f63kmVwIu5nl8Kfu5vK9vBFTRWv9WlHnJ3WyEEEIIIYRJKsnbzGutZwGzHuS1SikzYArQpyhzAqnMCyGEEEIIUViXgSp5HlfOfu5fDoA/sEkpdR4IAn4uiotgpTIvhBBCCCFMklnZ+QHYvUAtpVQ1jJ34bkCPfydqreMB938fK6U2AQO01vsK+8ZSmRdCCCGEEKIQtNYZwDvABiAM+FFrfVQp9alS6unifG+pzAshhBBCCFFIWut1wLpbnht1h9g2RfW+0pkXQgghhBAmqSQvgC2rZJiNEEIIIYQQJkoq80IIIYQQwiRJZV4q80IIIYQQQpgsqcz/j3OwyCrtFAqtQ+VapZ1Cobyz8/K9g8q4luVM/3P/zgM9SzuFQgluuLi0Uyi0tTtfKu0UCq2CjekfU88lGko7hUIp9z+wDq7dNP1jalmhytC9KUuLbE1CCCGEEEKYKKnMCyGEEEIIkyRj5qUyL4QQQgghhMmSyrwQQgghhDBJUpmXyrwQQgghhBAmSyrzQgghhBDCJEllXirzQgghhBBCmCypzAshhBBCCJMkt5mXyrwQQgghhBAmSyrzQgghhBDCJMmYeanMCyGEEEIIYbKkMy+EEEIIIYSJkmE2QgghhBDCJCkpS0tlXgghhBBCCFMllXkhhBBCCGGS5AJYqcwLIYQQQghhsqQyL4QQQgghTJKS0rxU5oUQQgghhDBVUpkXQgghhBAmSQrz0pkXd6G1ZslXaziyKwxLK0v6Du2Od53KBeLOn7jInC+Wkp6WTkCQDz3eew6lFD9++zMHdxzD3NyARyU3+g7pjq2DTYm34fPPZ7F5836sra0YN+59/PxqFojr23c0kZExZGZm0rixH6NHv4nBYKB///GcO3cZgMTEJBwc7Fi79qsSbUNeUYePcmLJj+isLCq1bkG1Jzvkmx7++19c3rINZWbA0sEe3769sHF3K/E8T+0LY/3M1eisLBq1D6JVl8fyTc9Iz2D1pEVcPX0RGwc7XhjaG5fyuXnGRcTwzZtjadOzIy06PUzUpeusGDc/Z3rs1SjavvQ4wc+2KbY2aK2ZN/UnDuwMw8rakn4julH9Ntv/2eMX+fazZaSlptMw2Ic+Hzyb87Xv+hVb+WPVdswMZjRs7sOLbz9FRkYmM8f+yLkTl8jMzKJ1xyY81+uRYmvH/Zgx8Q06PtKQyOgEmjw2qFRzObYnjFXT15CVpQl+PJB2PR7NNz09LYOF4xZz8eQl7BxteXlUb9wquALwx5K/2LluN2Zmis7vPo9P07oAjO7+KVa21piZKcwMZgya8VHO/Dav3sKWtdsxM1P4Bfny7BtPF2l7tNbMnfoTITuM29HbI2+/HZ05fpFvxhi3o0bNfXg5ezuaMmIBVy5EAnAzMRlbBxsmLfiIU0cvMHP8in/fhBf6tiewTUCR5v6vE3vDWPvdanSWplmHINp2y79OMtIyWDZxEZdPXcLWwZaew3vjWsGNk/tPsH7OL2RmZGIwN/DEa09Ts2FtAGYM+JqEmAQsLC0AeG1sP+xdHIolfzCuh8XT1nA4+3z26rDuVL3D+ez7L5aSlppOvSAfer5vPJ8t+yb3fFaukht9h3bHzsGGHX/sZ/3Sf3Jef+nMVT6e8yHetSoVaf4n9obxy4zV6ExN045BtOlacB38+O86cLSl+zDjOji1/wS///ALGRmZmJsb6Pja09RsYFwHhzaF8M+yP8nK1PgE+tLx1aLd9kXJkc68uKMju8K4fimKsUuGcfZYOAumrGTkzP4F4hZOXkmfQV2o7uvN1EGzObL7OPWCfPBtUodOrz+BwdzAiu9+4bdFf/FCv6dKtA1btuzn/Pkr/PHHTA4dOsHHH3/HihWTC8RNmzYYe3tbtNa8995Yfv99O0880ZovvxycEzNu3Bzs7W1LMv18dFYWxxcupdHA97F2dWH3J2PxaFgP+0oVc2IcvKsQOHoYBitLLm7czKkfV1PvrddKNM+szCx++3YFvT5/C0d3Z2b1n0ydoADKeVXIiQnZsBMbexvenzOSI5tD+POHX+gytE/O9A2zf6JmE9+cx+6Vy9Nv+qCc+U/uNQqf4HrF2o6DO49z7VIU034cyqmjF5gzcRWff/9+gbjvJ67i9SFdqOXnxbiPvufgruM0DPYhdP9p9m09yoQFA7CwNCc+JhGAXRsPkZ6WwaRFA0lNSeOjHhNo8VhDynm6Fmt77mbhis3MmL+B76e+VWo5gHHdrpi2ircnvomzhzMT+00loLk/nlVzt52d63dh62DD6EXD2b8xhLWzfuGVUb25ev4a+zceYNgPg4mPjuebAd8xcsEwzAzG0aTvTXkLeyf7fO938sApDu8IZcjsgVhYmpMYm1jkbTqw8zhXL0bx9QrjdjR7wirGzim4Hc2esIo3hxq3oy8+zN2OPvysV07M/K9+xtbOGgCvGhUY/0N/DOYGYqMSGNBrMk1a+mIwNxRp/lmZWayZvpLXxvXDyd2Zr9+dgm+wP+W9c9fJnt93YWNvy+B5Izj4Twjr5vzCi8P7YOdkR58xr+Hk5sS1c1f5ftgMRiz9JOd13Ye8RJXaXkWa750czj6fjV86jDPHwlkweSWjZhU8n83PPp/V8PVmysDc85l/0zq88IbxfPZj9vmsS7+naN6uMc3bNQbg4pkrfDVsbpF35LMys1j7zUr6jjWug+nvTsEnKP862LvBuA4GzhvBoU0h/D7nF3oM74Otkx29P30NRzcnrp2/yg/DZjBsySckJSSx7vufeXf6AOyd7flx4mJOHziZ82HLlEhl/v/JmHmlVFWlVGhp53ErpVQfpdT00s7jTg5sC6V5+yYopajhV5WbN5KJi0rIFxMXlUDyzVRq+FVFKUXz9k04sPUIAP7N6uScWKr7eRMbGV/ibfj77108++zDKKVo0KAuCQlJRETEFIj7t5OekZFJenpGgQtqtNasX7+NJ598qETyvp34s+exLV8O23IemJmbUyGwKZEHDueLcfWpg8HKEgCnGtVIiYkt8TwvnwzHtaIHrp7umFuY49+6Ecd3HskXc3xXKA0ebQaAb8v6nDt0Eq01AGE7DuNcwS1f5z+vs4dO4lLBHefyxdv53bs1lNYdGqOUora/N0k3kom9ZfuPjUogOSmF2v7eKKX4P/buPCyq6g3g+PcwCAgiw+aaoOaGgoqpgEuaa7a7ZGlmmW1W9rPcSitTyzQt27PcdyvTtNTUNBdccEMFRXPFXdlRkHXO748ZgUFcis2p9/M8PM7cOffOe+65y5n3nnu99/572LnJfKhZu3Qrjz7djjIO5pyJm4c566iA9LQMsrOyyUjPxLY44Y4AACAASURBVL6MIaeDVlq27DhEfOKVUo0BIPrQKbyqeuFVxbzt3NMukIit1ofuiC2RBHUybzuN2zTirz1H0FoTsTWSe9oFUsbBHq/KnnhV9SL60Kmbfl/o8i107NU+p41ciyEzvHNTJG26/L3tqE2Xe9ix0breWmu2rdtLq06BADg6OeQcXzMyMimu/szpw9F4VfHC07I/N2oTyIGt1vvzwW0RNO3YDICAextxNNzcJlVr3YWbpxsAFatXIjMjk6yMrGKK9ObCQyNpeb/5fFbrZuezlHRqWc5nLe9vyp4Czmd3N/AlvoDzWdgf4QS1Dyzy2E8fjsYzbxu0DeTgtuvboImlDfxbN+Lo3tw2KH+tDXwrkZluboP483F4VfWmnNH8A7dWYB0iQ/cVeeyiZNhsZ14pJVcVillCbDIeFYw57z28jSTEJuUrk4S7t1u+MtYHSIDQlTsICK5XfMHewMWLcVSq5JXzvlIlTy5ejCuwbP/+79GiRR9cXMrSuXMLq8927TqAp6eR6tWrFDhvSUhPSMDRwz3nvaO7kfSEG3fWz23agldD/5IIzUpyXBJuXrnbjZuXkctx1tvN5bhEynub62IwGHB0diI1OYX0q+mELl5H297Ww4fyity4h4C2TYon+DwSYpLwrJhbD09vt+tO4PExSdb7SAVjzo/W86djOLTvOCOf/5z3X/maowfNHcugdo1wdHLgpUdG82rXD3ioV1vKlS+9Kz53ksTYRNzzrE+jlxuJ+dZ5UmwSRksZg8FAWRcnUpJTSIxJwt07z7zeRhJjE81vlOLroVP4+KVP2PLb1pwyl87EcCziOJNemczng766Zef/n4i/ze3IM0+9PSsYrysTtfc4bh6uVK7mnTPtyIFo3uj9MYP7TOKFYT2KPCsP5vXt5p173HHzNpIcd32buOXZn51czPtzXhGb91G11l3YO+Seun+atJDJL3/MH/NW5/yYLy4JMdbnM/cbnM888pzP3L2NJMRcfz7btGIHDYOuP5+Frd9LcIei78wnx+VrAy8jyfliT45NwniLNogMzW0DzypexJy5RPyFOLKzszmwNYLEmMQij70kKFVyf3eq2+rMK6V+UUrtVkodUEq9aJnWSykVoZSKVEpNyFP2W6XULkvZ0TdeKiilTiqlPrYsZ4dSqpZlurdS6mel1E7LX0vL9PeVUnOVUluAuUqpikqppUqpfZa/Fjf7PssyaiqlwpVSzSyZ8SVKqd+VUkeUUh/nKXdd/ZRSjyulPrW8/p9S6nieZW7JU6fRSqk9lvlvqwd7izrPVkptVkpFK6W65VlnvyulytzO8kvTr3PWYmewI7jjPaUdyk1Nnz6G0NA5ZGRksn27dcb7t9828dBD95ZSZH/f+a1hJJ84RfUuHW9d+A6yYf4qQh5ri2NZxwI/z8rM4nBYJA1aNS7hyP6+7CwTV5JT+WDq6/R57WE+e3cuWmuOHjyFnUExZfkovlw8gt8WbeTi2YJ/YIqi8cbnAxn+/RAGjH+RTb9s4ei+Y4B5+EJqciqDvx7Eoy89zIwxs4u9U/lPha4Np1VH645i7Qa+TF4wjPEzBrF0zjoy0jNLKbqbu3DyPCun/0r3//XMmdbrrad58/vhDPj0dU5EHmfPHztLMcLbt3zOWgwGO0I6WZ/Pjh2IxtGpDHfVrFxKkd3cxZPnWTX9V7pa2sDZ1ZnHBj7OwnGz+W7wF7hX9MDO7g7urYqbut3s9nNa63ilVFlgp1JqBTABuAdIANYopR7TWv8CjLSUNQDrlFINtdb7b7LsJK11gFKqL/AZ8BDwOTBZax2qlPIBVgN+lvL1gVZa66tKqR+AjVrrrpbvK1fQF1yjlKoLLAKe1VrvU0o1ABoDgUA6cFgp9SWQXVD9gM3AtbvDWgNxSqmqlteb8nxVrNa6iVLqFWAI8PzN4rK4WZ3vBu6z1H0b0F1rPUwptRR4EPglXz1fBF4EGDrxNR59+sZZzvzWLQll02/bAahRrxrxl3J/qcfHJOLu5WZV3t3LzWr4jLlM+Zz3oat2sH/bQYZMHlBiz4KdP38FP/64GoCAgNpcuBCb89mFC3FUrHjjG0IdHR1o3z6YdevCaNnSfOLMyspm7dptLFkyuXgDvwVHd3fS8wybSU9IxNHd/bpycQeiOPHrKpq+/SZ2ZUr+t155TzeSYnO3m6TYRFw9rbcbV08jyTEJuHkZyc7OJj01DefyLpw5HM3B0H2snbGctJSrKKWwd7An6GHzD6mju6KofPddlHMvT3FY/XMo65aHAXB3vWrEXcytR1yMddYOwMPbzXofuZSYc6XKs4Ibzds0NF/Wr++DnVJcTkxhy5o9NA6qh729ATcPV+oGVOf4odNUrFryNyrfaYxeRhLyrM/E2CSM+da5m5cbiZcScfc2bztXU9JwKe+C0duNhDyZxcSYRIyWK0RGS8be1d2VRq0CiD50ilqN7sbobaRRa3MbVffzxU4priSl4Gq86ankln5fHMoflu2olt/tbUdxeeoddynRqkx2VjY7NkQwYdYbBX7fXdUr4uTsyOnjF7jbr1qhYs/PzcuNpJjc405STGLOsI38ZYyWNklLMe/PYG6HOaNn8OSwp/Cs4pVnHnObODk7EdiuCacPn+Kejs2LNPY/loSy8deCz2cJNzif5b0ikhCTiLt37rFm88od7Nt6kGGfXX8+C1sXTlD74rliWN4zXxvEJlI+X+zlvdxIjEnArYA2SIpJZO6YGfQcat0G9YP9qR9svnobtnJrzv0ltkZ+g9z+MJvXlVL7gO1ANcyd0w1a6xitdRYwH7iWtuyplNoDhAMNMHdAb2Zhnn9DLK87AF8ppfYCy4HySqlrR9flWuurltftgG8BtNbZWuubDcr2BpYBT2mt8w4MW6e1TtJapwEHAV+gWUH101pfAMoppVwt62GBpd6tMXf0r1li+Xc3UP0W9b/mZnVepbXOBCIAA/C7ZXpEQcvXWn+vtW6qtW76dzryAO27tWL0jCGMnjGEwNYBbF29C601xw6cxNnFCaOXdSfK6FWess6OHDtwEq01W1fvIrCV+eAQERbFqgV/MvCj/jg6OfytOArjqaceZNmyL1i27As6dAjml1/Wo7Vm795DuLo6U6GC9VjrlJSrOePos7Ky2bBhJzVr5j7lYOvWvdSsWdVquE5pKF/Dl9SLl7gaE4spK4sLYTvxDrS+CTQ5+hRRs+bT6H8DcChfPB3eW6lSx4f4czEkXIgjKzOLyE17qBdsPdynbpA/e//YAcDB0H3UaFgbpRT9J/6PN2aN4o1Zowh+tA2tn+iY05EHiNi4m4A2xTfEpnP3Vnw8ezAfzx5Ms3v92fT7brTW/BUZjbOLk9UPVQB3r/KUdXHir8hotNZs+n03zVqb69rsXn8O7jkKwLlTMWRlZeFqdMGrojuRu83T066mc+TAKar4Vii2OtkSn3rViDkbQ+x587aze304ASENrMoEtPAnbI1529m7cR91AmuhlCIgpAG714eTmZFF7Pk4Ys7G4FvPh/Sr6aSlpgGQfjWdQ7sOU7mG+X6Mhi39ObLX3BaXTl8iKyubcm4uha7H/T1aMWnOYCbNMW9HG1f9ve1o46rdNLs3d5/Zv/MIVXwrWA3FuXgujuysbABizsdzLvoS3pWv/3FfWHfV9SH2bCzxljbZtzGc+iHW+3P9EH92rTVn1iM27aNWY/P+fPVKKjPf/Z4u/R+ieoOaOeWzs7NJSTLfo5GdlU3U9oNUrF70Ge0O3VoxduYQxs4cQpPWAWz53Xw+O3rgJGXL3eB85uLIUcv5bMvvueez/Zbz2f8KOJ+ZTCZ2/LmXoGIYYgPmNog7G0u85Zi6b0N4Tif8mvrB/uyxtEHk5n3c3ci6De5/zroNAK4kmm/4Tr2cyvZfQ2l2f3CxxC+K3y0z80qptpg7miFa61Sl1AZgL+Zscf6yNTBnoptprROUUrOAW93ZpQt4bQcEWzrYeZcPYD0I7PYlAaeAVpg77dek53mdza3XyVagH3AYcwf+Ocw/QgbnKXNtmbezvGtuVud0AK21SSmVqXOvA5v+xvL/tobBfuzfFsVbvcbh4FiG597ulfPZqOcmMXrGEAD6vNmDGR+ZH+UVEFSPgGDzBYX5ny0hMyObT96cAsDd9X3pO+Tx4gq3QG3aNGXjxl107PgiZcs6Mm5c7lMkHn30dZYt+4KrV9MYMGAsGRlZaG0iKKghTz7ZJafcypWbePDB0rvx9Ro7g4G6fZ5gz6Qv0CYTVVq3oFzVKhxdspzyNXypENiIIz8sITs9nf1fTwXAydODwEEl+4QSg8HAAwO6M/edbzGZTAR2CqaCb2XWz11JldrVqBccQJPOwSyZNI/P+4+lrKszPYY/c8vlZqSlcyz8MA8PfKIEagGBLfwI3xbF/x7/CAenMgwY+WTOZ8Oe+YSPZ5t3+f5DuvPNB4vITM+kcUg9GoeYR9bd91Bzvv3wBwY/NRH7MgZeeacXSik6d2/JNx8uYvBTH6M1tH2wGb61Su9eDIDZXw6kdYgfXu6uHA37irGfLmb2DxtKPA6DwcDjA7vzzfDv0NkmgrsEUblGZVbMXIVPnWoEtPQn5IEg5oybz+g+H+Ls6ky/d58GoHKNyjRp25hx/cZjZ7Dj8dd7YGew43LCZaa+NxMAU3Y2TdvfQ/3m5mNUcJcg5k9cxLjnJmCwN9BneO8iv4LYpIUf4VujGPj4Rzg4luHVd3K3oyF9P2HSHPN29MLQ7nxtecRp4+B6BIbkjtDc8sf1Q2wO7TvBL3PXY7A3YKcUzw/pRvlCXlEoiMFg4NHXujNtxBRMJhPNOgdRqXplVs9eyV11fGgQ4k+z+4NZNGEeE579AGdXZ3qPMD+BZ+uyUGLPxvLHvNX8Mc98tfSFjwbg4OTAtLenkJ2djTZpagXWIahLyM3CKLRGIX7s3x7FsCfH4ehUhv55zmfv9pvE2Jnm81nfN3vkeTRlPRpazmfzJi8hKzObidfOZw18edZyPju87zgeFYxUqFI8V9cMBgOPvNqdGZY2aNopiIrVK7PG0gb1Q/xpen8wP348j4nPfkBZV2d6XWuD5aHEnYtl3fzVrJtvboP+Hw2gnNGVX79dwvnj5wBo/1RnvO+yzaSCZOZB3Wp8oFLqUeB5rfXDlvHfezF3ZseTOwxlNfAlcBKYg3nYijewHxiutZ51g2WfBKZorccrpfoAT1i+ZwEQrrWeaCnXWGu9Vyn1PnBFaz3JMn0RsF1r/dm1YTYFZeeVUtWB34AgS6zfaK0XKKWeBZpqrV+zlPsNmIS5o749f/201sss84yx/M0EIoGrWusmeerUVGsdq5RqCkzSWre9Qf1zvv9v1PmK1rqc5bXVZwXZcnHFnTkA9G9oWbF2aYdQKK9tO1vaIRRaqwrpty50h6tnLJ2naBSVkMD5pR1CoS3b9nRph1BolcqaSjuEQjtxuehvlC1JFf4FbXAh1TaHtOTVtXqXO6Ib3fH3LSXWz1l7f8s7os753U5W93fgZaVUFLmd3PPAW8CfmJ+0tkJrvQxAKRUOHAJOA1tuY/nuSqn9mLPP134qvw58bZluj3k8+ssFzPs/4HulVH/MWfABmMeUF0hrnaKUeghYq5S64XPYtNbnlVIF1g9zNr4asElrna2UOm2pb2Hdbp2FEEIIIQRgp2w+Z1lot8zMF+uX58lil1oQ/3KSmS99kpm/M0hmvvRJZv7OIJn50ieZ+aLTeXVoifVzVndudUfUOT/b35qEEEIIIYT4jyqR/3jJ8gjFGvkmD9daVy/i7/EE1hXwUXutdak9yFkp1Q/zkKC8tmitXy2NeIQQQggh/g3kBtgS6sxrrbuW0PfEYX5u/B1Faz0T882yQgghhBBCFJkS6cwLIYQQQghR1GS8uKwDIYQQQgghbJZk5oUQQgghhE2SR1NKZl4IIYQQQgibJZl5IYQQQghhk+RpNpKZF0IIIYQQwmZJZl4IIYQQQtgkyUrLOhBCCCGEEMJmSWZeCCGEEELYJBkzL5l5IYQQQgghbJZk5oUQQgghhE1S8px5ycwLIYQQQghhqyQzL4QQQgghbJKMmZfMvBBCCCGEEDZLOvNCCCGEEELYKBlm8y+XkmX715/Wnzta2iEUSpe7bL8Nyhps/wajSZGupR1CoSzb9nRph1Boj4bMLe0QCu3f0A5eTqbSDqFQUjJt/5hq621wJ5GstKwDIYQQQgghbJZk5oUQQgghhE2yk0dTSmZeCCGEEEIIWyWZeSGEEEIIYZPk0ZSSmRdCCCGEEMJmSWZeCCGEEELYJMlKyzoQQgghhBDCZklmXgghhBBC2CQZMy+ZeSGEEEIIIWyWZOaFEEIIIYRNkufMS2ZeCCGEEEIImyWZeSGEEEIIYZNkzLxk5oUQQgghhLBZkpkXQgghhBA2SbLSsg6EEEIIIYSwWdKZF0IIIYQQwkZJZ14IIYQQQtgkO6VL7O9WlFL3K6UOK6WOKqXeKuDzN5VSB5VS+5VS65RSvkWyDopiIUIIIYQQQvxXKaUMwNdAF6A+0EspVT9fsXCgqda6IbAY+Lgovls680IIIYQQwibZqZL7u4XmwFGt9XGtdQawCHg0bwGt9Z9a61TL2+3AXUWxDuRpNoKDO6L4+aulmEyakAeC6NS7g9XnmRlZzB0/n9N/ncGlvDP93nsGz0oeAKxZ8AfbVoZhZ6foMbAbfs3qkXApgbnjF3A54TIALR8KoW33NgCEb9jLytm/c/HUJYZ8Mwifuj5FUgetNT9+uZQDYVE4OJWh7/Be+NSpdl256MOnmTNhIZnpmTQI8qPnwK4opUhJTmHamDnEXYjHs5IHz496BhdXZ/aFRvDrzFUopbAz2PH4a12pFVCTw+FHWPz1LznLvXDqEv3f60vjVgFFUp+oHVH88s0STCZNcJdg2veybpOsjCwWTJjH6SPmNun7zjN4VPIkJSmFWWNmcvrwKZp1bk73gT2uW/b0d6cSdz6OYdOuuwJYKCXdBnEX4vnuvRlokyY7K5u23Vpz7yMti7RO1yRFRnLqhx/QJhPerVpRuUsXq88vrF1LTGgoys4Oe1dXajzzDI6enjmfZ1+9SsSoUbg3boxv797FEuM1Rb0/A4zqNQZHZyfs7MxtMGzK4JzlbVyyiU3LtmBnp2gQXJ/HXnqkWOt3I1MmvkSX9oHExCXTtOOwUonhmuJoAwBTtomJAz7FzcuNl8e9AMDGpZvZ8PMmYs/F8tHSsZRzK1esddNas+CLpURsj8LB0YH+b/fCt+71/ZGTh08zfdxCMjMyCQj2o/fr5v38x2+Ws3frQeztDXhX9aT/W71wdi1bZLH99FXuMejpYQUfg079dZq5ExaSYTkGPf5a7jFoxtjcY1D/957B2dX5hsuNuxDP1PdmYNKWY1DX1rS2HIOWT19B2JpdpF5OZfLKCUVSt4VfLCUizLzen3u7F751Cl7vMz9aSEZGJgFBfvSyrPddf+5l+azVnI++xMgpg6hez7xetq/dzepFf+bMf+bYed6d+iY+tasWOub/oKrA6TzvzwBBNynfH1hVFF8smfn/OFO2iZ8+/5kB419k5Mzh7F4fzvmTF6zKbFu1HWfXsoyaN5L7erRh2fe/AnD+5AV2rw9nxIzhDJjwEj9+thhTtgk7gx1dX36EkTPfYvDXg9i0bEvOMivXqMzzo5/j7oY1i7QeB8KiuHQ2htHzRtB7cE8WTl5cYLmFny3mqSE9GT1vBJfOxnBgxyEAVi9YR70mtRkzbyT1mtRmzYJ1ANS9pw4jpw1l5LShPD2sF/Mm/mCeHlg7Z/qgT1/BwcmB+k3rFkldTNkmlny5mBfHvcTw6W+x5889XIi2bpOwVdsp6+rMyDnv0KZ7W36bam4Tewd7ujz7AI+89GhBi2b/5n04ODkWSZz5lXQbuHmWZ+hXgxg5bSjDvn2D1QvWkRibVOT10iYT0QsWUPv11/EfPZq4nTu5eu6cVRnnatWoP2IE/qNG4dGkCad//tnq8zPLluFap06Rx5ZfcezP17z+6Su8NXWoVUf+r/Aj7N8ayVtThzJy5lu073lfsdfxRub+tJFH+44vte+/pjjbYMOSTVT0qWi1rJr+NXht0gA8KroXf+WAiO1RXDwTy0cLRvDM0MeZ82nB+/ncTxbz7LCefLRgBBfPxBIRZt7P6zety9hZQxkzayiV7vJmxbw/iiy2A2FRxJyN4f25I+j9Zk8WfVZwbIsmL6b34J68P3cEMWdjOGg5Bq1ZuI66gbV5f+5I6gbWZs3CdTddrptneQZ/NYgRU4cy9Js3WLMw9xgUENKAYd8MKrK6RYRFcelMLOPmj6DvkMeZd4P1Pu/TxfQd2pNx80dw6UwskZb1XqVGZV4Z24/ajazPvcEd72HU9CGMmj6E/iN641XZw+Y68iWZmVdKvaiU2pXn78V/ErNSqg/QFJhYJOugKBYibo9S6lmlVJXbKDdIKeV8G+U2KKWaFiam6EOn8KrqhVcVL+zL2HNPu0AitkZalYnYEklQp+YANG7TiL/2HEFrTcTWSO5pF0gZB3u8KnviVdWL6EOncPN0o5olG+Lk7EQln4okWQ5wlXwrUtGnQmFCLtC+LZEEd2qGUoqa9auTmnKVpDjrjl1SXBJpKWnUrF8dpRTBnZqxLzTCPP/WSII7NwMguHMz9m4xT3cq64hS5mtrGWkZqAIus+3ZuI8Gzevh4ORQJHU5dTgarypeeFraJLBtIJGWeK6J3BpBs07meBve24gj4eY2cSzrSM2Amtg7XH/RLf1qOhsXb6Bjn05FEmd+Jd0G9mXsKWOpZ1ZGFlrf+uakfyLlxAkcK1TAydsbO3t7PJo1I2HfPqsy5evVw+Bo/pHkUrMmmQkJufNHR5OVnIxb/fxDJ4tecezPNxO6fAsde7XPaQdXd9fiqdht2LLjEPGJV0rt+68prjZIiEnkwPaDhDwQbLWsarXvysnql4Tw0EhadG6KUoq7G1Qn9cpVEmOTrcokxiZzNTWduxuY9/MWnZsSvtm8P/s3r4vB3gBAzQa+JMQU3Q/w/VsjCepoPgbVqF+dq1ducAxKTaOG5RgU1LEZ+yzHmv1bIgmyHIOCOucem2603Jsdg2rUr46bp1uR1W1vaCQh+dd7XL71HpdMWp71HtK5KeGWOlSpXpFKtzj37lgXTrN2gUUW87+R1vp7rXXTPH/f5/n4LJD3UtBdlmlWlFIdgJHAI1rr9KKIS4bZlKxngUjg3C3KDQLmAam3KFdoibGJuFcw5rw3erlxMsr6BJ4Um4TRUsZgMFDWxYmU5BQSY5KoUT/3Rmyjt5HE2ESreeMuxHPm6Bl8/Yrkhu0bSoxNsqqHu5eRxNgkq4NpYmwSRu/c90Zvt5wsyuX4yzlly3uU53L85Zxyezfv55epK7iceIVXP3rhuu/e9Wc4HR5vW2R1Ma/v3Cyb0dtI9KFo6zJxSRi9zWUMBgNOlja52SX2VTNX0ubx+3BwLFNkseZVGm0QfymBb96eyqWzsXR76RGMXkV38rwmIzERB4/czpKD0UjKiRM3LB8bGoqbvz9gzuqf/uknavbvT3JUVJHHll+x7c9K8fXQKSilaPlwCC0fagHApTMxHIs4zm/TV1LGoQyPvfwIvvWKZuicrSquNljy9VIefelh0lOL5Nz/jyXEJuORp34e3kYSYpMwepXPUyYJ9zz7ubmMdccTIHTlDpq3a1xkseVdr3Bt/d36GHQt2XQ5Id8xyDJU9GbLTbiUwDcjphJzNpauxXQMMsdtvd7dvY0kxiRh9Mxd74kx1uvd3dt43Q+tm9n5515e+/C5ogm4BN1BWemdQG2lVA3MnfgnAatxlUqpQOA74H6t9aWi+uI7aB2UDqVUX8sjgvYppeYqpaorpdbneWyQj6XcLKXUF0qprUqp40qpHnmWMVwpFWFZRoHXeS3lmwLzlVJ7lVJllVLtlVLhlnlnKKUclVKvA1WAP5VSf1rm/dZyOeeAUmp08a+VopF+NZ3po2bS7ZWulHVxKu1wbptSirwp+MatG/L+nLd5eexzLJ+x0qpsUlwS546fp36eca13orNHzxB3PpaGrRqWdii35XbbwKOCO+9MH8aYeSPZvmYnyXl+AJSG2O3bSYmOplIn89WPSxs24Obvj4N7yQyBKC5vfD6Q4d8PYcD4F9n0yxaO7jsGmIeUpCanMvjrQTz60sPMGDO72K6Q/JdFbjtAOaNrgeO/bdWvc9ZiZ7AjuOM9pR1KgfIfg27EvYI7I6cN4/25IwlbXfrHoH/q+MFoHBzLULVm5dIOxWZprbOA14DVQBTwo9b6gFJqjFLq2s1EE4FywE+WvuDyovju/3RmXinVAHgHaKG1jlVKeQCzgdla69lKqeeAL4DHLLNUBloB9YDlwGKlVBfMdysHaa1TLcu4jtZ6sVLqNWCI1nqXUsoJmAW011r/pZSaAwzQWn+mlHoTuE9rHWuZfaTWOt7y2KN1SqmGWuv9N6nXi8CLAP8b/xoP9Olyo6IYvYwkXMrNpufPWgC4ebmReCkRd28j2dnZXE1Jw6W8C0ZvNxJi8swbk4jRy5w5yM7KZtqomTTtcA+N7y2eDuSGpaFsWbENAN96Plb1SIhNvC5DYvRyIzHPJd3EmKScMq4eriTFmTMtSXFJuLpfn+Gu3ehuYs/HcSXpSk4GfPefe2ncKiDnsnFRMK/v3GEaiTGJ112udfN0IzEmAaOlTdIsbXIjJw+e5PRfpxn71GhM2SauJF7h6ze/5NVPBxYq1juhDa4tt0r1ShyNOEaTNkWX6QNzJj4jPj7nfUZiImUK6JwnHTzI+ZUrqTdkCHZlzFc/rhw/zpUjR7i0cSOmtDRM2dnYOTlRrVu3Io3xmuLan43e5n9d3V1p1CqA6EOnqNXobozeRhq1bohSiup+vtgpxZWkFFyNxXsT5p2sONogYmskkVsjORh2kMyMLNJS05g9bh7PjOhTInVatySUTb9tB6BGvWrE59m8/gAAIABJREFU56lffEwi7vn2c3cvN6vhM+YyuRnk0FU72L/tIEMmD8gZQvdPbfwlzzGorg+Jl/Kvv1sfg9yuHYPc8x2DLNvxtfa61XIr1yjaY9D6paFstqz36nWt13tCTOJ125V5+0myLpNnvd/MjvXhNG/fpAiiLnm38/z3kqK1XgmszDftvTyvO1w3UxH4r2fm2wE/Xes0a63jgRBggeXzuZg779f8orU2aa0PAtfuQuoAzLz2qCHLMm5HXeCE1vovy/vZwL03KNtTKbUH8/NJG2B+fukN5R3TdbOOPIBPvWrEnI0h9nwcWZlZ7F4fTkBIA6syAS38CVuzA4C9G/dRJ7AWSikCQhqwe304mRlZxJ6PI+ZsDL71fNBaM3/iIir5VKRdEQ4/ya9t11Y5N0Y2aunP9jU70Vpz/OBJyrqULbAD7OTixPGDJ9Fas33NThq1NA+HaNjCn+2rdwKwffVOGrUwT790NiYn03jqr9NkZWZbdZp3rg+naREfAKvV9SHmbCxxljYJ3xCOvyWeaxq08GfnGnO8+zfto1bj2jc9KbZ8pBXv/zCGd+ePYuBnr+N9l3ehO/JQum2QEJNIRnoGACmXUzkWeYKK1Yr+fgyX6tVJv3SJ9NhYTFlZxO/ciXujRlZlUk6dInrePGq/+iplyueePO9+/nkaTZhAo48+otrjj+MVHFxsHXkonv05/Wo6aalpgPlq26Fdh6lcoxIADVv6c2TvUQAunb5EVlY25dxu/KPyv6A42uCRFx5i7I/vM3rhe/R7ty91AmuXWEceoH23VoyeMYTRM4YQ2DqArat3obXm2IGTOLs4XddhNHqVp6yzI8cOmPfzrat3EdjKvD9HhEWxasGfDPyoP45FcJ9Rm8daMWLqUEZMHUqjVv6ErTUfg07c7Bjk7MQJyzEobO1OGlqONQEt/AmzHIPCVu+kYc6xqUGBy817DEq9nMrxIj4GtevaKufm1MDWAWzLs97LujhZDbEBMHqWxynPet+2eheNW/nfYOm5TCYTu/7cS/P2Ml7eVv2nM/P/QN7BioVLJ9wmy9irIUAzrXWCUmoWUGRjVgwGA48P7M43w79DZ5sI7hJE5RqVWTFzFT51qhHQ0p+QB4KYM24+o/t8iLOrM/3efRowP5mmSdvGjOs33vzIwNd7YGew41jEcXau3UWVmpUZ/4L5Ru2H+z9Ig+D67Nu8n8VfLuFK0hWmjJhK1bur8urHLxe6Hv7B9YkMi+K9Ph/i4OhA3+FP5nz24fMTGTltKAC9BnVn9njz49IaNPejQZAfAJ17tWfa6NlsWRmGR0V3Xhj1DADhm/YTtnonBnsDZRzL8Px7fXM6zXEX4kmISaR2o7sLHX9eBoOBbgO78/1bUzCZTDS/P4hK1SuzatZKqtXxwb+FP0Fdglkwfh4f9v0AZ1dn+o7smzP/2KdGk5aaTnZmFpFbInhpwgAq+VYq0hgLUtJtcCH6Ij9/uwzzrqjp0LMtVWve8v7yv00ZDPj06sXhzz4Dkwmvli0pW6UKZ5ctw9nXF/fGjTmzeDHZ6ekc/e47ABw9PKj92mtFHsutFMf+fDnhMlPfmwmAKTubpu3voX5zc5sFdwli/sRFjHtuAgZ7A32G9y50pvWfmv3lQFqH+OHl7srRsK8Y++liZv+wocTjKI42uJkNSzaxbtF6kuMv89HzE2kQ5EfvIU/edJ7CaBjsx/5tUbzVaxwOjmV47u1eOZ+Nem4So2cMAaDPmz2Y8ZH58Y8BQfUICDZvM/M/W0JmRjafvDkFgLvr+9J3yONFEluDoPocCIvi/T4f4uDkQJ9hueth3AsTGTHVfAx6YlB35loej1s/zzGoU6/2TB8zm62rzMeg/u89c9PlXoi+yJIpy1AoNJr2eY5BS79bzq51e8hMz2Rkz/dp8UAwDz57/z+uW0CwHxHboxjR27ze+72Vu95H95/EqOmW9f5GD2aMN9fNP6geAZa67dm0n4VfLOVy4hU+f2sqPrWq8saklwD4a99xPCoY8a7ief0X24DbeP77v576L49vtAyzWQqEaK3jLENkZmHO1s9VSj0LPKq17mrpRP+mtV5smfeK1rqcUup+4D2gw7VhNjfKziulfgU+1Vr/aRlm8xfQTmt91LL8cK3150qpCMx3OZ9QSjUC5gCBgDewHxiutZ6llNqAZdjOjeq45uxKm29gexvfUa9m23gFgLIGm9+MmHHEtjPGfWullHYIhfZoyNzSDqHQlm17urRDKDQXe9ven69m2f4x1fFfcExtXenBO6Ih3gxbX2Ir89OgdndEnfP7T2fmLTcmfAhsVEplYx7GMhCYqZQaCsQA/W6xjN+VUo2BXUqpDMxjpUbcoPgsYIpS6irm4Tz9MN8EYY/5LugplnLfA78rpc5pre9TSoUDhzD/ZwRb/nmNhRBCCCH+Pf7r48XhP96ZB9Baz8Y8Xj2vdgWUezbf+3J5Xo8Hbvm/lWitfwby/m8y6zBn3POX+xL48kbfnWd621t9pxBCCCGE+Pf6z3fmhRBCCCGEbZIx89KZLxZKqa+Blvkmf661nlka8QghhBBCiH8n6cwXA631q6UdgxBCCCGE+PeTzrwQQgghhLBJ6g76T6NKi9wELIQQQgghhI2SzLwQQgghhLBJcgOsZOaFEEIIIYSwWZKZF0IIIYQQNkmy0rIOhBBCCCGEsFmSmRdCCCGEEDbJTp5mI5l5IYQQQgghbJVk5oUQQgghhE2Sp9lIZl4IIYQQQgibJZl5IYQQQghhkyQzL5l5IYQQQgghbJZk5oUQQgghhE0ylHYAdwDJzAshhBBCCGGjJDMvhBBCCCFskjxnXjrz/3qnr9j+BajPwpxLO4RCmf1AcmmHUGiJGbZ/h1GvmqmlHUKhVCprKu0QCm3ZtqdLO4RCezRkbmmHUGgXjz1X2iEUyjcHSzuCwnuu7tXSDkH8i8gwGyGEEEIIIWyUZOaFEEIIIYRNkkdTSmZeCCGEEEIImyWZeSGEEEIIYZMkMy+ZeSGEEEIIIWyWZOaFEEIIIYRNMkhmXjLzQgghhBBC2CrJzAshhBBCCJskY+YlMy+EEEIIIYTNksy8EEIIIYSwSXZKl3YIpU4y80IIIYQQQtgoycwLIYQQQgibJGPmJTMvhBBCCCGEzZLMvBBCCCGEsEmG0g7gDiCZeSGEEEIIIWyUZOaFEEIIIYRNkjHzkpkXQgghhBDCZklnXgghhBBCCBslw2yEEEIIIYRNkv80SjLzQgghhBBC2CzJzAsrx3cfZN20JehsEw07hRDco6PV51mZmayYPI+LR09TtrwLjwx9FreKnhzYsJOdS9fnlLt08hzPTB5KxZp35Uz7+YPvSboQx3NfvV1i9WlZxZ3hzWpiUIolRy8wPfJMgeU6+HgyuW19nlgRzsG4K7g52vNpGz/8PV1Zduwi43YcK7GYAbTWzP5sKXu3ReHg5MCAkb2oUfeu68odP3SaKR8uJCM9k8YhfjwzqCtKKRZP/531y7dT3lgOgCdeeoDAFvUBiD56jukf/0RqShp2dooPpr2Bg2OZIon5xy+XciAsCgenMvQd3gufOtWuKxd9+DRzJiwkMz2TBkF+9BxojjklOYVpY+YQdyEez0oePD/qGVxcnflr71G+fWc6XpU8AGjcuiEPPtMZgPWLNxK6YjtoTcuHQmjfo02h63FN1I4ofvlmCSaTJrhLMO17dbD6PCsjiwUT5nH6yBlcyjvT951n8KjkSUpSCrPGzOT04VM069yc7gN7AJCWmsZXb3yRM39STBJNOtxD11e6FVnM+WmtmTn5F/ZsjcLRyYFX332SmgVsR8cOnebrsYvISM+kSQs/+r3xGEopPn1nDudOxQCQevkqzq5lmTRnMEcOnOK7CT9d+xIe79+ZoLYBRRLzwR1R/PzVUkwmTcgDQXTqbb3eMzOymDt+Pqf/Mq/3fu89g6dl21iz4A+2rQzDzk7RY2A3/JrVy5nPlG1i4oBPcfNy4+VxLwCwcelmNvy8idhzsXy0dCzl3MoVSR3+qSkTX6JL+0Bi4pJp2nFYqcZyI1prPvloMVs2H8DJyYFRHz5NvfrW+3na1QzeenM6Z87EYmenaN02gIFvPGpVZv3acIa/MZ3Zi4ZS39+3JKvAmb0HCZu5GG0yUad9Cxo+1snq8wsHjxI2ezEJ0edoO6gf1YMDAYg7eYZtUxeReTUNZWdHw26dqdninhKNHcxt8PmEZWwPPYSjUxlGjH2Cun7X79fff7mK1b/u5nLyVdZs/zBn+sXzCXz4zg9cuXyVbJOJl//3ACGt/UqyCkXKIDfASmde5DJlm/jju5/oOeZVXD2NzBk8iVrN/fHyqZxTJmLtdpzKOfPi9+8RtWk3G2Yv59Fh/WjQthkN2jYDIObkOZaOm2rVkf9r6z4cnBxLtD52CkYG3c2LayO5kJrOogca8+fpeI4npVqVc7Y30MevKvtiknOmZWSb+GpvNLWMztQ2upRo3AB7t0Vx4Uwsk38YwdED0UyftJgPpg66rtyMSYt5YXhPajXwZcKQqezbfojGIeaD8gNPtOGh3vdZlc/OyubrMfN59d3e+NauyuWkFOzti+YpvQfCorh0NobR80ZwIiqahZMXM/zbN64rt/CzxTw1pCc1/Hz56q3vObDjEP5BfqxesI56TWrTuXcHVi/4gzUL1tH1pYcBqBVQk1c/esFqOWdPnCd0xXbe+vYNDGUMfDnsOwJC6lOhqneh62LKNrHky8W8PGEAbt5GJr/6KQ1a+FPJt1JOmbBV2ynr6szIOe8Q/ucefpv6K33ffRZ7B3u6PPsAF06e5/zJ8znlnZydGPJdbgft0wGTaNiqUaFjvZnwbYc4fzqWL396myMHTjH145/5aPr/ris39eOfefntntRu4MO4N6exd/shAkP8ePODvjllZn+xHGcXJwB87q7EhBmDMNgbSIhNZkjfT2jaqj6GQm5LpmwTP33+M69OfBmjt5GJAyYT0MKfytVz1/u2Vdtxdi3LqHkj2b1+D8u+/5Xn3nuG8ycvsHt9OCNmDCcpLomvh3zLu3NGYGcwX4DesGQTFX0qkpaalrOsmv418A9pwBdvfFWouIvK3J82MmX2aqZNfqW0Q7mhrZsPcupUDEtWjiJy/0nGj13ErIVDryvXp197mjavQ2ZmFq/0/5Itmw/QsnUDAFJS0lg0bwP+DauXcPRgMpnYPv1HOr/zGs6eRn59eyI+TQMw3pV7nnPxcqf1K08T+es6q3ntHcrQ+rW+uFWuQGp8Isvf+piqjfxwdHEu0TpsDz3EmVOxLPx1OAcjTvHJB0v4fv7r15Vr2aY+3Z5sSe+HJ1hNnz11Hfd1bkjXni04cewiw16bzk+rbLczL+7AYTZKqZNKKS/L6yu3KFtdKRV5g882KKWaFkeMdwql1CylVI+iWt75I9EYK3tjrOSFoYw9fq2bcDQswqrMkbAI/Ns1B6Buy8ac2vcXWluPV4vatJt6rXOzFRlX09m57E9CelpnP4pbgKcrpy6nceZKGlkmzaqTMdxXzeO6cq819mV65Gkysk05065mmQi/lGw1rSTtDo2k9f1NUUpR2786qZevkhCbbFUmITaZqynp1PavjlKK1vc3ZdfmiBss0Wz/jsP43F0Z39pVAXB1c8np7BTWvi2RBHdqhlKKmvWrk5pylaS4JKsySXFJpKWkUbO+OebgTs3YF2qOed/WSII7m38QBnduxt4tN6/LheiL1PDzxcHJAYPBQJ1Gtdi7aX+R1OXU4Wi8qnjhWcUL+zL2BLYNJDJfPJFbI2jWyRxvw3sbcST8CFprHMs6UjOgJvYON86VXDpziSuJV6gZULNI4r2RnZsiadPlHpRS1PH3JeXKjbajNOr4+6KUok2Xe9ix0fqwqrVm27q9tOpkzlA6OjnkdNwzMjIpqsRY9KFTeFX1wsuy3u9pF0jEVutYIrZEEtTJfAxq3KYRf+0xr/eIrZHc0y6QMg72eFX2xKuqF9GHTpnrGJPIge0HCXkg2GpZ1WrflZPVvxNs2XGI+MSbnvZK3cY/9/PgI81RShHQqAaXL18lNsZ6P3cq60DT5nUAKFPGnrp+1bh0MTHn8ylf/kbf5zricJN9pLjEHj2JayUvXCt6YbC3p2aLJpzaaX3ccK3giYdvVZSy3rLdqlTErXIFAJw9jDi5uZKWXPLtFfrnAe5/2LxfN2joy5XLacTGJF9XrkFDX7y8y183XQGpV9IBSLlytcAytsROldzfneqO68zbMqVUsR2ZinPZ11yJS8TVy5jz3tXLyOV8nbErcUmUt5SxMxhwdHHi6uUUqzKHQvfgd2+TnPeb56+g2WP3UcbRoRijv14FZ0cupKTnvL+YmkFFZ+urA34eLlRycWTz2YQSje1W4mOS8ayQ2xYeFYzE5zthxsck4VHBLee9p7eR+DwH9NU/hzKs70SmjFvElWTz1Yjzp2NQSvHRG9/xdr9PWD5/PUUlMTYJ9zwxu3sZSYxNuq6M0Ts3ZqO3W06Zy/GXcfM0f1beozyX4y/nlDtx8CQf9J/Il8O/49wJc7a7So3KHI04zpWkFDLSMogMO0hCTCJFISk2CWMF9zxxGgv8YWL0NpcxGAw4uTiRkmy9L9xI+J97aNw28LrOQlGLj0nCs2Jum3h6uxW4HeXd1jwL2Nai9h7HzcOVytVyr3ocORDNG70/ZnCfSbwwrEehs/IAibGJVtuQ0cuNxHyxmNvGXMZgMFDWst4TY5Jw984zr7eRxFjz9rDk66U8+tLD2N3JZ2MbEXMxkYqVcveNChWNVh31/C4np7J5YwTNguoCcOjgaS5eSKBVG/9ij7UgqfFJuHjmxu/s6U5KfNJN5ihYzNGTmLKyKF/RqyjDu73vvpRMhTz7tXdFN2Iv3X4d+g3oxJoVe+jW8QOGvjqDQW89VhxhihJUIp15pZSLUmqFUmqfUipSKfWEUuqXPJ93VEotvcn85ZRS65RSe5RSEUqpvIPv7JVS85VSUUqpxUqp6653KaU6KaW2Web/SSl1w4GRlisDH1u+Z4dSqpZlurdS6mel1E7LX0vL9PeVUnOVUluAuTdY5gqlVEPL63Cl1HuW12OUUi8os4mWdROhlHrC8nlbpdRmpdRy4KCl3FdKqcNKqT+ACjde66Xj3OGT2Ds64O1bBYCLx8+QeCGWOiHFO5zgn1DA0KY1mbTreGmHUuQ6dG3J5z+OZPyswbh7lmfeV8sB8zCGw/tP8Oqop3j/24Hs2hhB5K6/Sjna6ymlwNLRrVb7Lj5Y9B7vTB/KfV1bM+XdGQBU9q1Ipyfb8cXQKXw5/DvuqlUVZWcb+Ym9f4YTeF+TWxe8Q4SuDadVx0CrabUb+DJ5wTDGzxjE0jnryEjPLKXobi5y2wHKGV0LvH9DFK+srGxGDpvFE0+15a5qXphMJiZ//DODhhbffSIlITUhiU1fzqHVgD42c8zJ649V4XR5pClL1r7DxK+fY+zIhZhMpXMVuihIZr7kxszfD5zTWj8IoJRyA0Yrpby11jFAP2DGTeZPA7pqrZMtQ3C2Wzq4AHWB/lrrLUqpGcArwKRrM1rKvwN00FqnKKWGA28CY27yfUla6wClVF/gM+Ah4HNgstY6VCnlA6wGrg0yqw+00lpfvcHyNgOtlVLRQBbQ0jK9NfAy0A1oDDQCvICdSqlNljJNAH+t9QmlVDdLfesDFYGDBa03pdSLwIsAT49+nTZPPHCTquYq52nkcmxuhuVybCKunm75yriRHJuIq5c7puxs0lPSKOuaO6Y8avMe/PIMsTl36AQXjp5iyvPvY8rOJjXpCgtHfEGvcdeP7ytql1LTqeSSm4mv6OzAxdTcTL1LGQO1jC7M6NwQAK+yDnx5X30G/nmQg3Elf+l0zc+hrF++HYCaftWIu5TbFvGXEvHwtm4LD2834vNkY+JiEvGwXC41erjmTG/3SDAfD51mnqeCkXqNaubcGNs4xI8Th8/g37TOP4p5w9JQtqzYBoBvPR8S8sScEJuI0cs65vyZ1sSYpJwyrh6uJMUl4ebpRlJcEq7u5hjLWsZpA/gH12fhZ4u5knSFcm7laPlgMC0fNA+d+GXqCtzzraN/ys3LjcRLuVdrEmMSc64a5JTxdCMxJgGjt5Hs7GzSUtJwKX/r+yvOHjuLKdtEtWLqXP6+OJQ/locBUMuvGnF5sqZxMUkFbkd5t7W4fNtadlY2OzZEMGHW9fc/ANxVvSJOzo6cPn6Bu/0KVyejl9FqG8p/JQeutU0i7pb1ftWy3o3eblZXZhJjEjF6GYnYGknk1kgOhh0kMyOLtNQ0Zo+bxzMj+hQq1v+SHxdu5JfFWwGo7+/LxQu5+8ali4lWWeK8xr2/EB8fb3o/bb53JzUlnWNHz/Nyv88BiItNZvDA7/jky5dK7CZYZw83UuJy40+NS8DF4/aPGxmpV1k7/lvu6fUwFerUKI4QC7Rk0RZ+XWLer+s1sB62FHMxCa8Kt1+HFUt3Munb5wHwb1SdjPQskhJScfcs3RvAxT9XUj8pI4COSqkJSqnWWuskzFnsPkopIxACrLrJ/AoYp5TaD/wBVMXcmQU4rbXeYnk9D2iVb95gzJ3fLUqpvcAzwK2OGgvz/Btied0B+MqyjOVA+TwZ/uU36ciDuTN/L+ZO/AqgnOUKQg2t9WFLzAu11tla64vARqCZZd4dWusTltf35il3DihwjITW+nutdVOtddPb7cgDVK7tQ8K5GBIvxJGdmUXU5j3UCrJ+QkWt5v5Ert8BwOEte/FpWDtnqIA2mTgcGm41xCbwgda8OusDXp72Pk+NH4RHlQol0pEHiIy7jK+rE1XLOWJvp+hS3ZsNp+NzPr+Smc29P27n/iU7uX/JTvbHJJdaRx6gU/dWjJ89hPGzh9D03gA2/74LrTVHIk/iXM4Jdy/rcY3uXuUp6+LIkciTaK3Z/Psu7mllvnSdd1z0zo0RVKtpvoGwYfO6nD5+nvS0DLKzsonae4yqNSrxT7Xt2oqR04YyctpQGrX0Z/uanWitOX7wJGVdyhbYAXZyceL4QXPM29fspFFLc8wNW/izffVOALav3kmjFubpSfHJOfdlnIyKRmud02lOTjAPxYm/mMDezftp1qFonixRra4PMWdjiTsfR1ZmFuEbwvFvYT0soEELf3auMce7f9M+ajWufVvDZsLX7yGwXfFl5e/v0YpJcwYzac5gmt3rz8ZVu9Fa81dkNM4uN9qOnPgr0rxuN67aTbN7c+u6f+cRqvhWsBqKc/FcHNlZ2QDEnI/nXPQlvCu7U1g+9aoRczaGWMt6370+nICQBlZlAlr4E7bGfAzau3EfdQJrmcdvhzRg9/pwMjOyiD0fR8zZGHzr+fDICw8x9sf3Gb3wPfq925c6gbWlI/839ezVhgU/v82Cn9+mbbuGrFi+w3yfwr4TlCtXFq8CfkR/+8WvXLlylTff6p4zrZxrWf4IncDyNWNYvmYM/g2rl2hHHsDrbl+Sz8dw+VIs2VlZHN+6h2pNG97WvNlZWayfNJVa9wblPOGmpHR7siUzf3yTmT++Sev7/Pn9V/N+fWB/NOXKOf2tce8VKxvZHXYEgJPHL5KRkYXRo+Qf9FBUJDNfQpl5rfVfSqkmwAPAB0qpdcA04FfMWfeftNZZN1nEU4A3cI/WOlMpdRK4lq7L/78F5H+vgLVa615/J+QCXtsBwVrrtLwFLSfvWw2U3Qk0BY4DazFn318Adt9GLLc3CLcI2BkMdHipBz+9/w3aZCKgQzBePpXZPH8FlWr5UDsogIYdQ1jx6Vy+f3EMTq7OPDL02Zz5Tx84hquXEWOlkh9DWJBsDeN2HGNKB38MSrH06EWOJaXyaiNfDsRdZsOZ+JvO/3u3ZpQrY6CMnR3tqnny4h+R1z0Jp7gEhvixd1sUg3qOw9GpDC+NyN1833pmEuNnDwGg3+AeuY+mDK6X8ySbBd/8SvSRs6AU3pU8eH7Y4wCUK+/MA0+2YWT/ySilaBziRxPLIysLyz+4PpFhUbzX50McHB3oO/zJnM8+fH4iI6eZn3jRa1B3Zo9fSGZGJg2a+9EgyBxz517tmTZ6NltWhuFR0Z0XRj0DQPjGfWxatgU7g4EyjmXo/27fnE7z96NmkpKcisFg4Mn/dce5XNkiqYvBYKDbwO58/9YUTCYTze8PolL1yqyatZJqdXzwb+FPUJdgFoyfx4d9P8DZ1Zm+I3Of/DL2qdGkpaaTnZlF5JYIXpowIOdJOHs37uWFcS8WSZy30qSFH+Fboxj4+Ec4OJbh1Xdy22RI30+YNGcwAC8M7c7XHyzK2Y4CQ3If6bjlj+uH2Bzad4Jf5q7HYG/ATimeH9It52pPYRgMBh4f2J1vhn+HzjYR3CWIyjUqs2LmKnzqVCOgpT8hDwQxZ9x8Rvf5EGdXZ/q9+zQAlWtUpknbxozrNx47gx2Pv97jljd3b1iyiXWL1pMcf5mPnp9IgyA/eg958qbzFKfZXw6kdYgfXu6uHA37irGfLmb2DxtKLZ6CtLy3AVs2H6Brl9E4lS3De2Nzfxj17v4RC35+m4sXEpjx/Wqq16hIn8fNT1Lp2asNj/VoUVph57AzGAh+ridrPvwabdLUvi8Y92qV2fPDb3jd7YNP04bEHI1m/aSpZKSkcnp3BOE/rqDrp+9wcuseLkQdJf1yCkc3mK+itnr1aTyrX/9YyOIU0roe20OjePKh8Tg5OfD2mJ45n/Xr+Skzf3wTgG8m/8YfK/eSlpZJt44f8FC35jw3oBOvDn6Yj8f8xI/zNqMUjBjTs9jv3xHFS+V/EkmxfIlSVYB4rXWaUuoh4Hmt9WNKqV8xDyPpoLWOspQ9CTTVWscqpa5orcsppf4H1NJaD1RK3Yc5I33t+tYJoIXWeptSahoQpbX+RCm1ARgCRGPuNLfTWh9VSrkAVbXWBQ4Utnz/FK31eKVUH+AJrfXDSqkFQLjWeqKlXGOt9V6l1PvAFa31pIKWl2dlkuguAAAgAElEQVS5G4C7gADgEcxDgSZprT+3DJ95CfOPHQ9gFxAE1AOGaK0fsiwjb7kKmIfZvKC1/j979x0eRdU9cPx7khBSCOl0CB1C6L0KykvTVxGlCGJBLK9dEZBiQVGKIthQBBEQBKRKEUREBRIg9B6QIr2kJ5CE1Pv7Y5eQCkjq+juf58mT3Zk7M+fO3Jm9e/bu7JLctjvz6Dqb/2m0T4ML97Zf+W3OvdnvMmBropNs/0KfkGrbdajsmlrUIeTZpQTbG1+cVc82OX41yqZcPvFUUYeQJ18dLuoI8u6pOjf7MN82lHF6oFhcVOef+KXQ+jkDanQvFnXOqrDGzDcAPhaRNCAZeN46/QfA93pH/iZ+AFaJyAEsHd0jGeYdBV60jpc/DHydcUFjTJiIPAksEJHrA6jfAm72rT9P65CeROB6SvQVYKp1ugOwCct499u1GehsjEkQkc1YOvabrfOWYxnOsw/LJwHDjTGXRKRulnUsB+6x1vMMsPUfbF8ppZRSSv3LFNYwm3VYvjCaVXtgRpayVTM8LmX9H86NsetZZe3wXl+2U4bHv3NjDPrt+NgY82aW9YUD/XLYzpjbWaEx5m3gbevjC3Dj1szG8vHIMOtfxmX+BP7MUu6l26uCUkoppdS/m+1/3pd3RfYLsCKyC8t48DeKKgallFJKKaVsWZF15o0x+XPbiTtkva991vtKvZnxk4E7WGc3YGKWyX8bY3rd6TqVUkoppVTOivNdZgpLkXXmi1pBdLBvMpxIKaWUUkqpfKdDjZRSSimllLJR/28z80oppZRSyrbpMBvNzCullFJKKWWzNDOvlFJKKaVskr3Y/G9j5plm5pVSSimllLJRmplXSimllFI2ScfMa2ZeKaWUUkopm6WZeaWUUkopZZM0M6+ZeaWUUkoppWyWZuaVUkoppZRN0sy8ZuaVUkoppZSyWZqZV0oppZRSNsleM/OamVdKKaWUUiqvRKS7iBwVkeMiMiKH+SVF5Efr/GARqZof29XMvFJKKaWUskl2xeQXYEXEHpgKdAHOATtEZKUx5nCGYoOBKGNMTRF5BJgI9MvrtjUzr5RSSimlVN60BI4bY04aY5KAhUDPLGV6AnOsj5cAnUUkzwOFNDP/L9e6THJRh5BnuwY6F3UIebL27NWiDiHPWvqmFHUIeXYoyr6oQ8iTv6/YdvwAPk5pRR1Cnl0+8VRRh5BnZWt8V9Qh5MmhQwOKOoQ8u5ps+wO9yzgVdQQWxSgrXRE4m+H5OaBVbmWMMSkiEgN4A+F52XAx2gdKKaWUUkoVTyLyrIjszPD3bFHHBJqZV0oppZRS6paMMdOB6bnMPg9UzvC8knVaTmXOiYgD4A5E5DUuzcwrpZRSSimbZCeF93cLO4BaIlJNRByBR4CVWcqsBJ6wPu4N/G6MyfM3eDUzr5RSSimlVB5Yx8C/BKwD7IHvjDGHROR9YKcxZiUwE5grIseBSCwd/jzTzrxSSimllLJJxelHo4wxa4A1Waa9k+HxNaBPfm9Xh9kopZRSSillozQzr5RSSimlbFJx+dGooqSZeaWUUkoppWyUZuaVUkoppZRNuo27zPzraWZeKaWUUkopG6WZeaWUUkopZZM0M6+ZeaWUUkoppWyWZuaVUkoppZRN0qy07gOllFJKKaVslmbmlVJKKaWUTRIdM6+ZeaWUUkoppWyVZuaVUkoppZRN0sS8ZuaVUkoppZSyWdqZV0oppZRSykbpMBuViTGGmZN/YvfWEEqWdOSltx+hRt1K2cqdOHKWL8YuJCkxmaZt/Bk85EFEhL//Os+0iUtITkrB3t6OZ4c9TK2AKunLHTt8hpHPfMGQsQNpe0+jQqnPhHHfs3nTPpycHPlg3HPUC6iWrdz/nplIWFg0qSmpNG1eh9FvD8Le3o51vwTz9ZdLOXnyAgsWvU9A/eoFHjPA0R0hrPh6GSbN0LJ7a+5+5D+Z5qckpbDw43mcP3YOFzcXHh39BF7lvPlr11HWzlxFakoq9g723PfMA9RsUhuAaUO/IDYylhKOJQB4ZvzzlPJ0K/C6GGP44qMVbAsKwcnJkRHv9aO2f/Y29e2Xa1m3eidXYhP4Zcu49OlfTlrBnh0nAEi8lkRU5FV+3vxBgcW6+MvlHAoOwdGpBI8N70+V2pWzlTvz11nmTlxAUmIyAa386fNSL0SEuNg4vhv7PRGXIvEu58Xgd57Axc2F7b/tYv3CDWCgpHNJHnm9N5VqVATg7f7v4+TihNgJ9vZ2vDntjXyrz7+pHYHl+Mz/fDkHtoXgWNKRwSP741cne1s6dfQsM8ctIDkpmQat/RnwiuX4LPpqJXu3HMbBwR7fit4MHtEfFzfnQon9evyfjF9C0OZDODk58u6Hj1G3Xub2dS0hiRFDZnLuXDh2dkKHTg14+fWemcr8vn4Pb74+kzkLh1Gvvl+hxX8r0z5+jh6dmxAWEUvzLsOLOpxcGWOYNmkFO4JCKOnkyBtj+lEzh9e52VPXsmHNTq7GJrB8841r0s9LtrB68Rbs7O1wcnbkldG98ateTuMvIvoFWM3Mqyx2bz3CxbPhTF08kv+N7MP0j5bmWO6bj5by/Mi+TF08kotnw9mz9QgA33+5mn6DuzJ57hs88mx3vv9ydfoyqalpzJ36M41b1i6UugBs3rSP06cv8fMvn/Due4P54P1ZOZabNOVllv40nuWrJhIVeYVffwkGoFatSkz54jWaNa9baDGnpaax/MslDP7wOd6YMYK9f+7m8ulLmcps/2UbzqVceHP2W3R4qBNrZq4CwNXdlSfHPsOQ6W/Sb9ijLPzoh0zL9R/xGK9PG87r04YXWgcsOPAI586E8cOKEbzxVm+mjMu5TbW5qx7T5r6abfpLQ3sy88chzPxxCA890p67OjcosFgPBYcQdj6MMXNHMWBIXxZ+uiTHcgunLGHAG30ZM3cUYefDOLzd0v5/XbCBOk1qMWbuaOo0qcWvCzYA4FPOi9envMTomcPp8VhX5n+yKNP6Xp38AqNmDMvXjvy/rR0BHNgWwuVz4YyfP4onhvXh+8k5H5+5nyzhyeF9GT9/FJfPhXMg2HJ86jWvw9jZw3h/9jDKVfLl53m/FVrsAFs2H+bMmTCWrXmXUWP6M2HswhzLDRzUmSWr3uaHJSPYv+ckQZsPpc+Li7vGwnl/Ur9h1UKK+vbNXbyRno9PKOowbmlH0BEunA1j5vIRvDK6N1+Oz/ma1Oquenw2J/s1qVP3pnz941Cmzh9Cn8fvZsaUVQUdcia2Hr/Kf9qZtyEiUuCfpGzfdJBO9zZDRKhT34+4qwlEhsdmKhMZHktC3DXq1PdDROh0bzOCNx20xgjxcdcAiL+agJdv6fTl1iwOpM3dDXD3LFXQ1Uj3x++7eKBnB0SERo1rcSU2nrDQqGzlSpVyASAlJZXk5JT0d/rVa1SkWrUKhRYvwNmjp/Gp4IN3eR8cSjjQqGMTDm05kKnM4a0HaN6lBQAN7mrE8T3HMMZQsWYl3L3dAShbtRzJScmkJKUUavxZBW08RLf/NkdECGjox9Ur14gIi81WLqChH94Z2ktONvyyh87dmxRUqOzfcpBWXVogIlSrV5WEqwnERMRkKhMTEcO1+GtUq1cVEaFVlxbsC7Icn/1BB2nVzXJcWnVrwb5Ay/Tq9avh4mZpY9Xq+REdlnmdBeHf1o4A9gQepG03S1uqEVCV+KsJRGe5PkWHx5IQn0iNAMvxadutOXs2W+pdv2Ud7B3sAage4EdUIRyHjDb+sZ/7HmiJiNCgUTWuXEkgPEsMTs6ONLcmPEqUcKCOf2VCL0enz5/2xWoef6oLjo7F74P1oO1HiIy+WtRh3NK2jYfofK+lHfk3sFyTsr7OAfg38MPLJ/s1ybWUU/rjawlJhZ4ZtvX485tdIf4VV4Uam4g8LiL7RWSfiMwVkaoi8rt12gYRqWItN1tEPheRLSJyUkR6Z1jHmyJywLqOXFMAIvKniEwRkZ0iEiIiLURkmYgcE5EPMpQbKCLbRWSviHwjIvbW6V9blz0kIu9lKH9KRN4Tkd3WOHJN2YpIR+t694rIHhFxy60OItJYRLZZ98VyEfHMUI9PRWQn8KqINBORjSKyS0TWiUj5Oz4gOYgMi8GnjEf6c+8y7kRmebGJDIvB2zdjGY/0Mk+99iDff7maZx54nzlfrOLR5+8FICI0huCNB+j2UNv8DPeWQi9HUq6cd/rzsuW8CM2hMw/w3NMT6Nj+eVxcnejSrVVhhZhNTHgM7r6e6c/dfT2IzdqhzFDG3t4eJ1cn4mPjMpU5sHkfFWtWwiHDi/7iSQuY8r+P+G3eOowxBViLG8JCY/Atd6O9+JZ1Jyz0n3eiLl2I5OKFSJq0qJmf4WUSEx6DR4b27+HrQXR45lijw2Pw8HXPUMadGGuZK1FX0jvBpb1KcyXqSrZtbFkTTECrG5cNEeHLYdOY8NwnBK7ekq91+Te1I4Co8Fi8MhwfL18PorIcn6jwGDwzHB9LmewdncA122nQuvA+cQMIuxxN2XI3jkmZsh6ZOupZXYmNZ/PGA7RoVQeAI4fPcvlSFO071i/wWP/NIsJi8MlwTfIp6074P7wmrVoUxKCe45n5xWr+N/TB/A7xpmw9fpX/Cq0zLyIBwFvAPcaYRsCrwBfAHGNMQ+AH4PMMi5QH2gP/Ba53eHsAPYFW1nV8dIvNJhljmgPTgBXAi0B94EkR8RYRf6Af0M4Y0xhIBR61LjvaumxDoKOINMyw3nBjTFPga2DoTbY/FHjRuu4OQMJN6vA98KZ1XxwA3s2wHkdrLJ9b91lvY0wz4Dvgw1vsg0L1y7ItDHq1JzNWvsOgV3vy1YeW4QTfffoTj734X+zsiu9722++HcEfm6aSnJRC8LZDt16gGLt06iJrZq7i4Vf7pk/rP+Ixhkx/k+cnv8LfB0+y+7cdRRjhP/f7ur107NwQe/vi24YyEpFsgzn/2nOMLWu30fOZ+9OnDfnsZUZMH8qLE55l009BHNt3orBDzdW/sR0BrPp+PXb2drTu0qyoQ8lVSkoqo4fPpt+jnahU2Ye0tDSmfLSU14Y9VNShKeD+vu2YtWIkT718HwtmFu5wrfxg6/FnJGIK7a+4KszP6e4BFhtjwgGMMZEi0ga4fmWaS+bO+U/GmDTgsIiUtU77DzDLGBN/fR232OZK6/8DwCFjzEUAETkJVMbyZqEZsEMsL7rOQKh1mb4i8iyWfVQeqAfst85bZv2/K0P8OQkCJovID8AyY8w5EclWBxFxBzyMMRuty80BFmdYz4/W/3WwvBlZb43XHriYdaPWuJ8FeHfyi/R5svtNQoS1SwJZv8IyRrymf2XCQ29kiiJCY/DKkOUC8PJ1JyIsY5no9DJ/rtnJ4CGWd/ltOzfiq3GWzvyJkHNMfmsuAFdi4ti19Qj29na06pj/458X/PArS5f8AUD9+tW5dCkifd7lS5GUKeOZ26KULOnI3fc044/fd9G2XcGNzb4Zdx93YsJufHoQExZNaW/3HMt4+HqQmprKtbhruJR2BSA6LJrv3/uOR4Y/incFnwzLWDI5Ti5ONLmnKWePnqFZl5YFUoflPwaxepmlTdUNqEzYpRvtJexyDL5l3HNbNFe/r9vLayPyvyOz8adAgn7eCoBfnSpEZ2j/0WHRePhkjtXDxz3TMJnosBjcrWXcPN2IiYjB3dudmIgY3DxuDCk7f+ICP0z6kRcmPEspd9cb67N+yuXm6Uaj9g04feQMtRrVyHO9/g3tCGDDskA2rd4GQLW6lYnMcHwiw6LxzHJ8PH3cMw2fsZS5MdQgcO129m89zNApzyOFML5g0YKN/LTE8olLvfp+XL5045iEXo6mTFmPHJcbN2YBVar4MuCxuwGIj0vkxPGL/G/QZwBEhMfyxsvf8MkXzxWrL8EWV6sWBfHLT5ZrUu16lQnPcE0KvxyDzx1ckwA6dm3Ml+OX3bpgHtl6/KpgFb9BdzckZnh8p1fc6+tIy7K+NCx1FyyfDIzMuJCIVMOSVW9hjIkSkdmAU4Yi19eVyk32oTFmgoj8DNwLBIlItzusx/XPvQXLm5I2NytsjJkOTAc4FLX6lm8le/RuT4/e7QHYGXSYtYuDaN+lCX8dOoNLKadsY+68fErj7OrE0YOnqR1QhT/X7OLevpblPX1Kc2j3Ceo3q8mBnccoX9kXgGnLR6cv/8X7C2jWvl6BdOQB+j/alf6PdgVg0597mD//V3rc24b9+45Tys0Z3yyd+fi4a8TFJeBbxpOUlFQ2bdxD00L8wmtWlepUIfx8OJEXIyjt486+jXvoP+KxTGXqtanPzvU78KtXjQOb9lGzcS1EhISr8cx6ezo9Bv+XqgE37ryTmprKtasJuLqXIjUllZBth6nZtOC+iNyrXzt69WsHwNbNh1m+MIh7ujfm8IEzuJZyuuXY+KxO/x3KldgEAhrlf6el44Pt6figpf0e3HaIjT8F0uyeJpwKOY2zq3P6sJnr3L3dcXJx4u/Dp6jq70fw+h10fLADAA3a1id43Q66DvgPwet20LCdZThE5OUopr87iydGPkrZymXS15WYkIgxBicXJxITEgnZeZQej3fNl3r9G9oRQOeH2tP5Icvx2bf1MBuWBdKqcxNOHj6Ni6sTHlmuTx4+pXF2KcmJQ6eoXs+PLet28h/r8geCQ1g7/w/e/OJFSjo5Fmjc1/Xt35G+/TsCELjxIIsWbKJrj2Yc3H+KUqWc8fHN3gn7+vNVXL2awFvvD0ifVsrNmd8CJ6Y/f+7JT3l1aC/tyN+m+/u24/6+lmvS9sDDrFoURMdujTly0HJNymlseW7OnwmjYhVf67pCqFjF5xZL5J2tx1+QbHzIf74ozM7878ByEZlsjIkQES9gC/AIlqz8o8DmW6xjPfCOiPxgjIkXEa/byM7fzAZghYhMMcaEWmNyA0pj6UDHWD8V6AH8+U9XLiI1jDEHgAMi0gKom1sdRCRKRDoYYzYDjwEbc1jlUcBXRNoYY7aKSAmgtjEm38aENGvrz+4tIbzQezwlnUrw0luPpM8b8tgnTJ5rudvGs8MeznBryro0bWPp/L4wsg8zp6wgNTUVR8cSPD+yd47bKSwdOjZm06a93NttSPqtKa/r3WskS5aPJz4hkZdfnExSUjImzdCiVT369usMwIb1Oxj34RyiIq/wwv8+pm5dP775dkSBxmxvb0/Plx7m21HTSEtLo0W3VpSrWp51c9ZQqXYVAtrUp0X31iycOI+JT36Ai5sLA0Y9DsCWFYGEnw/nt3nr+G3eOsBy60BHJ0e+HTmN1NRUTJqhZpPatOpx0/eE+aZ1e3+CA4/w6AMTKOlUgjfH9EufN7jfZGb+OASAaZ+u5re1e0i8lkzvbmO5r1dLBv3P8v7393V7uKdb4wLPpAa0qseh4BDGDPwQRydHBg6/0f7HPfMxo2YMA6Dfaw8zd+ICkhOTqdfSn4BW/gB07d+Zme/PYcvaYLzKejL4nScAWDt3HXGxcSz8zHL3leu3oLwSdYXp71jusJSamkqLzs0IaOmfL3X5t7UjgIat/dm/NYQR/cfhWLIET43snz7v3acm8d53llGPA4f05rvxlluHNmhVlwatLfv0h0+XkZyUyidDpgFQo54fjw/tU2jxt7srgKDNh+jV4z2cnEvwztiB6fMGPDye+UtHcvlSFN9NX0fVamUZ2MfSee/bvyMP9i7c7xvdiTlfvEyHNv74eLpxPPhLxk5ewpwf/yzqsLJp0c6fHUFHeOrBCTg5leD1d29ck14cMJmp8y3XpJmfreaPdZZr0sB7x9K9Z0sGPteNVYuC2LP9GA4O9pRyc+aNMY/ktimNXxUKKcwvL4nIE8AwLBntPVjGhc8CfIAwYJAx5ow1E77aGLPEutxVY0wp6+MRwONAErDGGDMql239CQw1xuwUkU7Wx//NYV4/YCSW7w8kYxnjvs0aQ1vgLBADrDTGzBaRU0BzY0y4iDQHJhljOuUSwxfA3Vg+CTgEPGmMScypDiLSGMvYfhfgpHVfRGWM1brOxljGzrtjeTP2qTFmRm77/HYy88VdLXfbvf8twNqzYUUdQp619C36O5nk1aEo+6IOIU/iUmw//+TjlFbUIeRZA6/CyegXpLI1vivqEPLk0KEBty6kClx1t/uLxUVpX2Th9XMaef23WNQ5q0LtzKvCp535oqed+eJBO/NFTzvzxYN25lV+0M588VGcx8wrpZRSSimVq2LZuy5kNt+ZF5GpQLsskz8zxuT8U58FE8MgLLfazCjIGPNiYcWglFJKKaX+/7H5znxx6DBb3zgU2psHpZRSSikFdpqaL9a/TquUUkoppZS6CZvPzCullFJKqf+fNDGvmXmllFJKKaVslnbmlVJKKaWUslE6zEYppZRSStmkAv5hcJugmXmllFJKKaVslGbmlVJKKaWUTdLEvGbmlVJKKaWUslmamVdKKaWUUjZJM/OamVdKKaWUUspmaWZeKaWUUkrZJDtNzWtmXimllFJKKVulmXmllFJKKWWTNDGvmXmllFJKKaVslmbmlVJKKaWUTRIxRR1CkdPMvFJKKaWUUjZKM/P/cuP3uxV1CHlWt3R0UYeQJx3LpxV1CHnWa61nUYeQZ++3ji3qEPKkjLPtt6O4ZNsf3frV4aKOIO8OHRpQ1CHkSUDA/KIOIc827hpY1CHkWfVi0r2w/atK3mlmXimllFJKKRulmXmllFJKKWWTRFPzmplXSimllFLKVmlmXimllFJK2STNSus+UEoppZRSymZpZ14ppZRSSikbpcNslFJKKaWUTdIvwGpmXimllFJKKZulmXmllFJKKWWTNDGvmXmllFJKKaVslmbmlVJKKaWUTdIx85qZV0oppZRSymZpZl4ppZRSStkkTcxrZl4ppZRSSimbpZl5pZRSSillk+w0Na+ZeaWUUkoppWyVZuaVUkoppZRN0sS8ZuaVUkoppZSyWZqZV0oppZRSNknEFHUIRU4z80oppZRSStkozcyrXMUcPMiZH3/EpKXh27495Xv0yDT/0vr1hAUGInZ2OLi5Ue2JJyjp7Z0+PzUhgQPvvotn48b4DRhQ2OEDcH7vIXbOWYJJS6PmPe2o37NrpvmXQ46xc85Sos6cp8Mrg/Br3RSAq2ERbPxkBsakkZaaSt1unajdpUOhxW2MYcHnyzkQHIJjSUeeGtkfv9qVspU7dfQss8YvICkpmQat/On/Si9EhJ1/7GXl7HVcPB3K6GmvUbVuZQBSUlKZ89GPnPnrHKmpabTt1px7B/6nQOvSuqwHrzeqjp0IK/++zNy/zmWa36taOR6uUZ40Y0hISWX87uOcupIAQM3SLrzZtCauJexJM/DU73tJSiu4LMzh7SEs/XI5aWmGNve2ouuAzPsmOSmFuRN+4Oxf53At7cKgd57Au5wXAL/O/42ta4KxsxN6v/wQ/i3qAhB/NYEFkxZy4e9LiMCjw/pTLaAqa2b/wpaft1HKwxWA+wffR0DrevlaH2MMP3y2nP3bLO3o6VH9qVon53b07bgFJCUm07C1P4++amlHC6euZO+Wwzg42FOmojeDR/bH1c2ZLb/uYu2CP9KXP3fiImNmDsGvVsU7jnPxl8s5FByCo1MJHhvenyq1K2crd+avs8ydaIkzoJU/fV6yxBkXG8d3Y78n4lIk3uW8GPzOE7i4ueS63ohLkcx45zvSjCE1JZVOvTrQ4YF2AKyc+TPBv+4k/ko8U9ZMvKP65Obc3sMEz7Jcj2p3bkvDBzNfjy4dPk7wnCVEnb5Ap9cGUbV1EwAiTp1j64yFJCdcQ+zsaPhQN6q3bZavsd0uYwzTJq1gR1AIJZ0ceWNMP2rWzd6mZk9dy4Y1O7kam8DyzePSp/+8ZAurF2/Bzt4OJ2dHXhndG7/q5QqzCjc17ePn6NG5CWERsTTvMryow0lnjGHuZ8vZt9Wy35/N5Vz++8hZplvP5UZt/HnMei4vmbGW3YEHERFKe5bi2dH98fRxJy42nhnjFxJ6IYISjg48PfIRKlcvXwQ1vHM6Zl4z8yoXJi2N0/PnU+uVV6j/3ntE7NhBwoULmcq4VK5MvVGjqP/uu3g1bcrZpUszzT+3YgVutWsXZtiZpKWlsf27Rdwz4kXu/+RtTgXtJPrcxUxlXL29aPv8Y1Rr1zzTdGdPd7qPfYP/ThxFjw+GcXDFr8RHRhda7AeCQwg9F864H0bx+NA+zJu8JMdy8yYv4fFhfRn3wyhCz4VzMPgIABWqleeFsYOo1ah6pvK7/thLSnIK780eztszhrBx1VbCL0YWWD3sgKGNa/B60CH6/7qbrpV9qermnKnMurNhDPxtD49v2Mu8v87zakNLzPYCY1rWYeKeEwxYv4cXNh0gpQA78mmpaSz+bCnPT3iW0bPeZNfve7h46lKmMlvXbsPFzZl3543m7t4dWTF9FQAXT11i1+97GPXdmzw/8TkWfbqEtNQ0AJZ+uQz/Fv68PWckI2YMo6xf2fT13d27IyNmDGPEjGH53pEH2L8thMvnwpm4YBRPDu/D95/k3I7mfLKEJ4f3ZeKCUVw+F84Bazuq36IOH84ZxgdzhlGusi8/z/sNgLZdmzF21lDGzhrKs28NwKe81x135AEOBYcQdj6MMXNHMWBIXxZ+mnOcC6csYcAbfRkzdxRh58M4vN0S568LNlCnSS3GzB1NnSa1+HXBhpuu1927NG98+RqjZgxj2Fev8+uCDUSHxwDQoE0Aw7967Y7rkpu0tDS2zVxE11Ev0GvKW5wM2pX9euTjSYcXHqN6+8zXIwfHEnR46XF6TX6LrqNeYPvspSTGxed7jLdjR9ARLpwNY+byEbwyujdfjl+aY7lWd9XjszmvZpveqXtTvv5xKFPnD6HP43czY8qqgg75H5m7eCM9H59Q1GFks29bCJfPhjNp4SieGtaHWZNyPkdmf7KEwcP7MmnhKC6fDWf/Nss5ct+Auxk3Zxgfzh5K47b1+GnWr1mMrFwAACAASURBVACsnPsbVWpVZNycYTz31gDmffZTodVJ5Z9i15kXkTEiMvRO5//DbVUVkYP/cJn07YvIbBHpnR+xFDdxf/9NyTJlcPL1xc7BAa8WLYjaty9TmdJ162JfsiQArtWrkxwVdWP506dJiY3FvV7+d1BuV8TxU7iV88WtrA/2Dg74tW3G2Z37M5UpVcYbT7+KIJnf29s7OGBfogQAackpGFO4Y/L2Bh6kTbfmiAg1AqoSfzWB6IjYTGWiI2K5Fp9IjYCqiAhtujVnT+ABACpULUu5KmWyr1iExIQkUlNSSU5MxsHBASfXkgVWj3pebpyLu8aFuERSjGH9uTDuquCdqUx8Smr6Yyd7O8Cyr1uW9eR4TBzHY+IAiE1KIa3AIoXTR87gU9EHnwo+OJRwoNk9TTiwJfPl4UDQQVp1bQlA446N+Gv3MYwxHNhykGb3NKGEowM+5b3xqejD6SNnSLiawPH9J2lzbysAHEo44FLKOdu2C8qewIO0625pRzWvt6PwLO0oPJaEuERqWttRu+7N2b3Z0o7qt6yDvYM9ADUC/IgMi8m2jeDf9tCqc5M8xbl/y0FadWmBiFCtXlUSriYQE5F5WzERMVyLv0a1epY4W3Vpwb4gS5z7gw7SqlsLAFp1a8E+63mQ23odSjhQwtHywXRKUubzu1q9qrh7u+epPjkJP34Kt3I+6dej6m2bcmZH5uuRWxlvvPwqIlmuR+4VyuJe3nI+u3h54OTuxrXYq/ke4+3YtvEQne+1tCn/Bn5cvXKNyCxtCsC/gR9ePqWzTXct5ZT++FpCUtZLb5EL2n6EyOii2bc3s3vzQdpfP5fr3+Jcrm85R9p3b84u67ns7HpjvydeS0p/zTt/6jIBzWoBUMGvLOEXI4mJvFJItVL5RYfZ2AixXN3FGFOQ/Zl0SdHROHp5pT939PAg7u+/cy0fHhiIe/36gCWrf3bxYqoPHkxsSEiBx5qb+MhoXL0905+7enkQfvzUbS8fFx7F7x99xZVLYTR7tBcuXh4FEGXOosNj8SpzY3uevh5Eh8Xg4X3jxTE6LAZPX/fMZXJ4Uc2oWadG7A06yBsPjSEpMZl+L/akVGnX/K+Ala+zI6HxienPQxMSCfByy1bu4erl6V+rAiXs7HjJ+uJTpZQzxsCn7QPwdCzB+nNhzPvrfIHFGh0ejWeGfe7h486pkDOZysSEx+BhLWNvb4+zqxNxsXFEh8VQrZ7fjWV9PYgOj6ZEyRKUci/FvI8WcOHEBSrXrsTDL/aipLPlDdSmnzazff0OqtSuTK/ne+Li5pKvdYoKy96OosJj8MjQyYoKj8ErSzuKCsvejjb9vJ1W9zTONj349728Ov6pPMWZcb/C9f0Xk6lTHR0eg0eGOD183YmxZtOvRF1JL1vaqzRXoq7ccr1RoVF8NWoGYefD6fXcA3j45H8HPqP4yJhM1yMXb0/Cjp36x+sJO36KtJQUSpf1ycfobl9EWAw+5W7sU5+y7oSHxuTYcc/NqkVBLPthEykpKUz4+n8FEea/TlSW1wSvMh5EZjmXI7Ocy15lPIjK8Jqw+Js1BK7bibOrE6M+fwGAKjUrsGPjfuo0qs6Jw6cJvxxFZGg07jlcp4ur4vaGsCgUi8y8iIwWkb9EJBCoY532jIjsEJF9IrJURLK9yolIDRH5RUR2ichmEal7k22UFZHl1vXtE5G21ln2IjJDRA6JyK8i4ny727+Nek0QkcMisl9EJt0sDhEZIiIHrX+vWadVFZGjIvI9cBCoLCLDrHHtF5H3ctnusyKyU0R2HltV8B9hhm/bRtzp05Trahn/Gfrnn7jXr4+jp+ctlizeXH08uf+j0Tz46RhObAomIfrmHWVb8HfIGezs7Ji0bAwTFo7m10V/EnYhoqjDYunJi/Ret4upB0/xpHV8v70IjXxK8+72ozy7cT8dK3jT3LdgO1z5LS01lXPHztHhgXa8OX0ojk6OrLcOAWn/QDvenfcWb04fSmnv0iz/ekURR5u7ld+vx97ejjZdM4/TPnHoNCWdSlCpGI2xFZHbenX3LOPJ6G+HM2buaILX7SDWBrKR8VExbPrie9o/PxCxKxYv33fk/r7tmLViJE+9fB8LZv5W1OH8v9HnuXv5bNk7tO3alPXLAgG4f2Bn4q8mMPrJSaxfGohfrYrY2dtu2/r/qsgz8yLSDHgEaIwlnt3ALmCZMWaGtcwHwGDgiyyLTwf+Z4w5JiKtgK+Ae3LZ1OfARmNMLxGxB0oBnkAtoL8x5hkRWQQ8DMy7ze3frF7eQC+grjHGiMj1t9TZ4rDug0FAKyzf5QgWkY1AlDW+J4wx20Skq/V5S2u5lSJylzFmU8ZtG2OmW/cNAzduvKPxIY4eHiRF3hhLnRQdTYkcOucxhw9zcc0a6g4dip11WMrVkye5euwYoRs3knbtGmmpqdg5OVH5oYfuJJQ75uLlQVxEhqE/kdE430F23cXLA4/K5Qk9cjz9C7IF4fflgWxevQ2AqnUqExl6Y4x+VFh0pqwkWDKTURmGPUSFRWfK0uRk+2+7qd+yLg4O9pT2dKNm/WqcOnIW3yxDX/JLWEISZVxuDOMp41ySsISkXMuvPxvG8CY1GMsxQhMS2RMeQ0xSCgBbLkVRx7MUO3MY6pEfPHw8iMqwz7NmggHcfdyJDo3G09eD1NRUEuKu4Vra1XosMiwbFo2Hjwcevh54+LpT1d+StW98V6P0znzpDJmvtve14ZtRM/KlHr8tC2TjKks7qlY3ezvyzJKB9vRxzzR8JiosGk/fG+1o85rt7NtymOGfPp9t+Efwhj206nxn58TGnwIJ+nkrAH51qhAdmnX/ZWnvPu5EZ4gzOiwGd2sZN083YiIsGfeYiBjcPEoBN47XrdZbvlo5jh84QdOO2T95yC8uXu6ZrkfxEVG4et3+m9Ok+ATWT/iaZv3vp0ztagURYq5WLQril5+CAahdrzLhl27s0/DLMfiUubM32R27NubL8cvyJcZ/o/VLA/nTei5X9898LkeGRuOVpS17ZTmXI0Oj8czhNaFtl2ZMGjaDhwd3x9nViWdH9QcsX7Id0ucDyhTQ60FB0cR88cjMdwCWG2PijTGxwErr9PrWbPsB4FEgIONCIlIKaAssFpG9wDfAzdJD9wBfAxhjUo0x11v838aYvdbHu4Cqt7P92xADXANmishDwPVvK+UUR3vrPogzxlwFlmHZLwCnjTHbrI+7Wv/2YHnTUxdL5z7fuVatSmJoKInh4aSlpBC5YweejRplKhN35gyn582j1osvUqL0jQtGjaefptHEiTQaP57Kffrg07p1oXfkAbxr+HHlUihXQsNJTUnh9JZdVG7W4LaWjYuIIiXJ0ulMvBpP6JGTlK5Q9hZL5c09vdrz7syhvDtzKE06NGDrup0YYzhx6BTOrk6ZhtgAeHiXxsmlJCcOncIYw9Z1O2ncvv5Nt+FV1oOQ3ccASExI5OTh05Tzy2FsfT4JibpC5VLOlHcpiYMIXSr5svlC5i/cVs4whrZdeS/OXrXcySb4chQ1S7tS0t4Oe4Gmvu78HVtwX/qrUrcyYefDCL8YQUpyCrt+30ODNplP+wZt6xP863YA9m7cR+0mNRERGrQJYNfve0hOSiH8YgRh58Pwq1uF0l6l8SjjweUzoQD8tfsY5f0sd+7IOCZ83+b9lK+WP9nt/zzUPv3LqU07NCDoF0s7On7oFM6lnLK94fPwKY2za0mOW9tR0C87aWJtR/uDQ1g7/w9eHT+Ykk6OmZZLS0tj+x97afWfOxsv3/HB9oyaMYxRM4bRqH19gtfvwBjD34dP4ezqnG3curu3O04uTvx92BJn8PodNGxribNB2/oEr9sBQPC6HTRsZ5nesG1AjuuNCosmKdFyfsdfiefkwb8pW7ngzgMAnxp+xF4MS78endyym8rNG97WsqkpKfw+aQY172qVfoebwnR/33ZMnT+EqfOH0KZTABvWWNpUyIHTuJZy+kdDbM6fCUt/vD0whIpVima4kC3o8nB7Ppw9lA9nD6VZhwYEXj+XD57C5Wbn8kHLORL4y06adrCcC5fO3tjvuwMPUsF63Y+7kkBKsiVh8ueqbdRpVCPT+HqVf0TES0TWi8gx6/9sGVIRaSwiW62jRfaLSL/bWXeRZ+ZvYjbwoDFmn4g8CXTKMt8OiDbG5DWVkpjhcSpw/dtpt9r+TRljUkSkJdAZ6A28RO6fGtxMXIbHAow3xnxzB+v5R8Tenir9+3P0008hLQ2fdu1wrlCB8ytW4OLnh2fjxpxbsoTUxESOf2MJp6SXF7VeeqmgQ7ttdvb2tBzUlw3jplpuTXl3GzwqV2DvotV4V69C5eYNCT9xmo2fTCcxLp5zuw+yb8nPPDDpbWLOX2LXvGUIgsFQ77+d8axy53fr+KcatPbnwLYQRg0Yh2PJEgwa0T993nuDJ/HuTMt3wAe+3pvvJiwgOTGZ+q3q0qCVPwC7N+1nwefLuRJ9lc9GzKBKzYq8Puk57n6wPbMmLOSdJyZiDLTr0YLKNSoUWD1SDUzae4LP2tfHTmD1qcv8fSWeZ+pV4UjUVTZfjKR3jQq0KONOSprhSlIK7++wvNm4kpzKgmPnmXVPI4yBrZei2HIp6hZbvHP29vb0eflhvnrzG0xqGq17tKJ8tfL8PGstVWpXpkG7+rS5txXfj/uB9wZ+iIubC4PefgyA8tXK07RTY8YNmoCdvR19Xumd/lF1n5cfZs64uaSmpOJd3puBwy3HcsU3qzh34gIi4FXWi0eG9Mn3OjVq48/+bSEMf2QcJZ1KMHjkjXb09qBJjJ1laUePD+md4daUdWnY2tKO5k1ZRkpyKh8PmQZYvgT75FBLnEf3ncSrjEe+ZPECWtXjUHAIYwZ+iKOTIwOHP5I+b9wzHzNqxjAA+r32MHMnWtp7vZb+BFjbe9f+nZn5/hy2rA3Gq6wng9954qbrvXT6MsumrUg/vzv37UTF6pbzYPk3K9m5YTfJicmM7juGtve25r4nu+e5jnb29rR+qi+/fjgVk2aodXdrPCuXZ/ePq/GpUYUqzRsSdvw0v0+aQVJcPGd3HWDPop/pNfktTm3ZzaWQ4yReieP4n5bcTvsXH8O7avZbExa0Fu382RF0hKcenICTUwlef/dGX+PFAZOZOn8IADM/W80f6/aQeC2ZgfeOpXvPlgx8rhurFgWxZ/sxHBzsKeXmzBtjHsltU0Vizhcv06GNPz6ebhwP/pKxk5cw58c/izosGrXxZ+/WEIb2G4ejUwmeGXXjXB795CQ+nG05l594ozfTP7ScIw1b16WR9Vz+cdpqLp4Jw85O8C7ryaBhlnt3XDh9mekfLACBStXK8fSI2+o7FivFISt9m0YAG4wxE0RkhPX5m1nKxAOPW0ecVAB2icg6Y8xNb6cnhX2XjmwBiDTF0nFuxY1hNt9gqWQ9LENN1gDnjTFPisgY4KoxZpKIbAGmGGMWW78g2tAYsy+HzSAiC4FtxphPswyzWW2MqW8tMxQoZYwZIyLht7H92dbls90jyvrJgYsxJlRE3IGTxhjvXOKoYd0HrbEOswEes247Y3xdgbFAZ2PMVRGpCCQbY0Jz2793OsymOKlbOrmoQ8iTjuUTb12omBsWZFtj1XPyfmvb/s6DWwmbP5WJS7b9D8R3hpco6hDyrG/1a0UdQp4EBMwv6hDybOOugUUdQp619L2vWJzQEddWFtrF0dvpgTuus4gcBToZYy6KSHngT2NMnVsssw/obYw5drNyRf6GxhizG/gR2AesBXZYZ72NpVMbBBzJZfFHgcHWyh4Cet5kU68Cd1uHzezC0lG/mdvZ/s24AatFZD8QCAzJLQ7rPpgNbLdu81tjzJ6sKzTG/ArMB7Zal19i3Y5SSiml1P8717/zXhh/eVTWGHP9xyUuATcdu2sd3eEInLjlPijqzLwqWJqZL3qamS8eNDNf9DQzXzxoZr7oaWY+/0QmFmZmvudzwLMZJk233nQEABH5DcjpJ41HA3OMMR4ZykYZY3K87d/1zD3WG6DcKq7iPGZeKaWUUkqpmyi89xQZ7xaYy/z/5DZPRC6LSPkMw2xyHCItIqWBn4HRt9ORh2IwzCa/We9ZvzfL3+gC3ubyHLbZrSC3qZRSSimlbMZK4Anr4yeAbD8uIiKOwHLg+5y+j5mbf11m3hjzIfBhIW+zV2FuTymllFJKgdjOneYnAItEZDBwGugLICLNsfxm0tPWaXcB3tY7KQI8meEW6jn613XmlVJKKaWUKk6MMRFYbleedfpO4Gnr43lYfrj0H9HOvFJKKaWUskki/7oR4/+Y7gGllFJKKaVslGbmlVJKKaWUjbKZMfMFRjPzSimllFJK2SjtzCullFJKKWWjdJiNUkoppZSySTZ0a8oCo5l5pZRSSimlbJRm5pVSSimllI3SzLxm5pVSSimllLJRmplXSimllFI2SX80SjPzSimllFJK2SzNzCullFJKKRulY+Y1M6+UUkoppZSN0sy8UkoppZSySXqfeRBjTFHHoArQvsjVNn+AG3hWL+oQ8mTZqTNFHUKe7YooUdQh5Nm9la4VdQh5En7N9j9I9XFKK+oQ8qyOR2pRh5BnV5Ntu/PzbzgXOjabV9Qh5FnCmQXFoiFdSd5QaP0ctxKdi0Wds9LMvFJKKaWUskmamdcx80oppZRSStkszcwrpZRSSikbpXlp3QNKKaWUUkrZKM3MK6WUUkopmySiY+Y1M6+UUkoppZSN0s68UkoppZRSNkqH2SillFJKKRulw2w0M6+UUkoppZSN0sy8UkoppZSySfqjUZqZV0oppZRSymZpZl4ppZRSStkozUvrHlBKKaWUUspGaWZeKaWUUkrZJB0zr5l5pZRSSimlbJZm5pVSSimllE0S0cy8ZuaVUkoppZSyUZqZV0oppZRSNkoz85qZV0oppZRSykZpZl5lYoxh1pSf2LMlhJJOjrzw9iNUr1MpW7mTR84ydexCkhKTadLWn0GvP4iIMOWt77lwJgyA+CsJuLg58/H3b7B/+1F++GoNKckpOJRw4LGX/kv95rUKpT7jPpzJpk27cHIqybjxLxMQUCNbuWeefp+wsChSUlNp3syft995Fnt7ewDmzf2Z+fPXYmdvR8eOzRg27IkCj/uvnSH8/PUy0tLSaN69NR37dck0PyUphSWT5nH+2FlcSrvyyMgn8Cznzdmjp/npsx+vV557BnYnoF0jws5eZuH4OenLR10Kp/Nj99KuV6cCr8ulfYfYP3cxJs1QtVNb6jzQLdP88JBj7Ju3hNgz52n50lNUbNU0fV58eCS7Z8wjITIKENoOfxFXX+8CjxksbWfB58s5EByCY0lHnhrZH7/a2c+FU0fPMmv8ApKSkmnQyp/+r/RCRNj5x15Wzl7HxdOhjJ72GlXrVgZg2/pdrFv4R/ry505c5O0ZQ6hSq2K+1+HojhBWTVuGSTW06NGaTv3+k2l+SlIKiz6ex/lj53Ap7UL/UU/gVc6bY7uO8st3q0hJScXBwZ4ezzxAzca1Adj3527+WLietFSDf6t69Hj6gXyP+7p/wzHIWp/PJq5gW+ARSjqVYNTYftTxz16f6V+sZd2qXVyJTeDXbR+mT798MYoP3/qRq1cSSE1L43+v3kubDv4FGnNWxhimTVrBjiDLa8QbY/pRs272OsyeupYNa3ZyNTaB5ZvHpU//eckWVi/egp29HU7Ojrwyujd+1csVStxzP1vOvq2WuJ8d1Z+qOby2/X3kLNPHLSApMZlGbfx57FVLW1oyYy27Aw8iIpT2LMWzo/vj6eNOXGw8M8YvJPRCBCUcHXh65CNUrl6+wOtzM9M+fo4enZsQFhFL8y7DizSWwiCal9bOvMpsz9YjXDobzueLR3Ls0Bm+/Wgp42a+mq3cjI+W8tzIvtQKqML4Id+yd9sRmrTx5/UPHk8v8/3nK3FxdQLAzd2VNz9+Ci9fd86cuMiHr03nm1XvFnh9Nm3azenTF/hl3Vfs2/cX77/3DT8u+ihbuSmfDqVUKReMMbz6ykf88ssW7ruvA8HbDrDh9+38tGIKjo4liIiILvCY01LTWDV1MYPGvUBpHw++fuUT/Fs3oIzfjRe8neu24lTKmTdmvc3+P3ez7rtVPDLqScr6leeFL97A3t6e2IgYvnzhI+q2ro9v5bK8/NXw9PVPHPgO9do2LPC6mLQ09s3+kfYjX8HZy4M/3p5I+aYNKV3pxouds48XzZ97jGM//5Zt+Z3T5lCnZ3fKNvAn5do1kMK7aB8IDiH0XDjjfhjFycOnmTd5CaOnvZat3LzJS3h8WF+q1/Pjs+EzOBh8hAat/alQrTwvjB3E958szlS+dZdmtO7SDIBzJy4w9a1ZBdKJTEtNY8XUJQwe/zzuPh58+fJk/FvXp2yGdrRj3TacS7kwbPZb7PtzN7/MXMWA0U/i4u7KE+8/Q2lvdy6dush3o6Yxav57xMXGsebblbz85VBKeZRi0cc/cHzPX9RsUjvf4wfbPwZZbQs8wrkz4SxY9SaHD5zhkw+WMf2HV7KVa9exHg890o4B90/MNH3OjA3c3a0hvfq25e8Tlxn+0kwWry3czvyOoCNcOBvGzOUjOHLwDF+OX8qnc7K/RrS6qx4P9GvH4F4TMk3v1L0p9/VuC8C2jYeYMWUVH3zxTIHHvW9bCJfPhjNp4ShOHDrNrElLeG9G9rY0+5MlDB7elxoBfkwaOoP9247QqI0/9w24m97P9ABg3eJN/DTrVwYN68PKub9RpVZFXhv/FBdOX2bO5GWM/Oz5Aq/PzcxdvJFpc9bx7ZQXijQOVXj+9W9nRKSqiBzMZd5rIuKS4fkpEfEpvOiKn52bDnJXj2aICLXr+xF3NYGo8NhMZaLCY0mIu0bt+n6ICHf1aMaOjZl3sTGGrRv20q5rEwCq1amEl687AJWrlyMpMZnkpJQCr8/vG7bTs+fdiAiNG9chNjaO0NDIbOVKlbI0g5SUVJKTU9K/Hb9w4S8888xDODqWAMDb26PAYz539DRe5X3xKu+DQwkHGnZsSsjWA5nKhGw9SNP/tAQgoEMjTuz9C2MMjk6O6Z8opCSn5DiU8MTev/Aq74NnWa8Cr0vkiVO4lvXFtYwPdg4OVGrdjIu79mUq4+rrjXuVStk66rHnLmJS0yjbwNJZcXBywqGkY4HHfN3ewIO06dYcEaFGQFXiryYQHZH5XIiOiOVafCI1AqoiIrTp1pw9gZZjVaFqWcpVKXPTbWzfsIcW9zQpkPjPHj2NdwUfvK3tqFGnJhzO0o4Obz1A0y4tAKjfoRHH9x7DGEPFmpUo7W05X8v6lSM5MZmUpBQiL0bgU9GXUh6lAKjZpDYHAzMfz/xk68cgq8A/DtH9fsv1NaChH1evXCM8LDZbuYCGfvj4ls42XYD4q4kAxF1NyLFMQdu28RCd77UcE/8GljpEhmevg38DP7x8ssfnWsop/fG1hCQK60YkuzcfpH13S9w161vbUpa4o8NjSYhLpGZ9S1tq3705uzZb2pKz6424E68lcT3w86cuE9DM8ilzBb+yhF+MJCbySuFUKhdB248QGX21SGMoXFKIf8XTv74zfwuvAS63LFUMiEihfIoSGRaDT9kbHVZvX3ciw2KylfEuk6FMGY9sZUL2nsTdy43ylX2zbSP4j/1Ur1OJEo4FX6XLlyMoV/7GsIxy5bwJvZy9Mw/w9OD3aN/uSVxdnenWrQ0Ap05dYNfOw/TrO5zHBo7mwIFjBR5zbEQM7r439m9pHw9iImKylInG3dcTAHt7e5xcnYiPjQPg7JFTfPbseL743wR6vtw3vXN/3f6Nu2nYqSmF4VpkNM7enunPnb08SYiKuckSN1y9dJkSLs5sm/ING0aN48D8ZZi0tIIKNZvo8Fi8MrRzT18PorO08+iwGDytb1LTy+TQscnNjj/20qpzwXQkLe3oxr539/EgNjxLOwqPwSOXdnTdwcB9VKxZCQdHB7wr+BB2LpTISxGkpqZyaMsBosMK7tMqWz8GWYWFxlImw/XVt6w74aG3dz4ADHq+K7/+vJuHunzAsBe/47URDxZEmDcVERaDT7kbdfD5h3UAWLUoiEE9xzPzi9X8b2jh1CEqS1vyKuNBZJbzITI8Jj3pdL1MxmTW4m/W8OpD77Pl1908PLg7AFVqVmDHxv0AnDh8mvDLUUSGFvwnuEpllG+deRF5XET2i8g+EZlrzYj/bp22QUSqWMvNFpHPRWSLiJwUkd4Z1vGmiBywrmPCTbb1iogctq57oXXaGBEZmqHMQRGpan3qICI/iEiIiCwRERcReQWoAPwhIn/ksI2BIrJdRPaKyDciYi8iLazbdBIRVxE5JCL1c4mxvIhssi5/UEQ6WKd3F5Hd1jpusE7zEpGfrOveJiINM9RprogEAXNFxFdElorIDutfu39yjApT0Po9tOuS/QXy7MlL/PDVzzzzZu8clipa3858l02bvyMpKZlt2yzZmJTUVGJirrLwx4kMG/4Er782CWNMEUd6c5XrVuXV6SN5/vM32PjjbyQnJafPS0lO4ci2gzTo0LgII7w9JjWN8KPHafDow9w99k3iQsM5vWlrUYeVb04ePo1jyRJULOLxtTdz+dRF1s5cRa9X+wLg4ubCgy/3YcG4OXzzxud4lvXCzq74ZqtuxRaOQUa/rd1Djweas2z9W3w89SnGjl5AWiG+wc0v9/dtx6wVI3nq5ftYMDP78Lriqs9z9/LZsndo27Up65cFAnD/wM7EX01g9JOTWL80EL9aFbGz//+eJ1WFLV9SoyISALwFtDXGhIuIFzAHmGOMmSMiTwGfA9ffgpcH2gN1gZXAEhHpAfQEWhlj4q3ryM0IoJoxJlFEbmfcQx1gsDEmSES+A14wxkwSkSHA3caY8Cz18Qf6Ae2MMcki8hXwqDHmexFZCXwAOAPzjDE5DuEBBgDrjDEfiog94CIivsAM4C5jzN8Z6vgesMcY86CImx/jiQAAIABJREFU3AN8D1zvbdUD2htj/q+9O4+3sirfP/65IAQVERxwQjGtHEJxFnKeSiutHErTVKxvk6Wp+f2l2eScY6Zl2bdwrpxTc8YBcRZwttEhyxRFcESZrt8f69mwwQOcfeCctdc+9/v1Oi94nn2OXps9nLXXc697TZF0KXCm7THVh6ObgfcVTEr6KvBVgGPOOJg9D9h5vv84N10xhlHXPgDAmuusyqsvz55VmPjKnDMVAMssvzQT62YeJk6YPMf3zJg+gwfvfJyTzz9sjp+bOGEyp31vJAf/YB9WHNR51UyXXHIDV1x+KwBD1vsQL/134qzbXnppIgPnU17Su/dibL/DZtw+6kG22GIDVlxhOXbaaRiSWH/9j9Cjh5g06Q2WWWbpef43Fla/ZZfm9brZzjdenczSyy491/f05/VXJrH08v2ZMWMG7779Lkv0W3KO7xm42or0Xrw3Lz/3XwZ9ZDUgLaxd+UOD6Dugay7P91mmP1MmTpp1POW1SSw+oH3/dosvM4D+gwex5MD0XFl546G89o9nYdvOSJrcfvUY7r7+fgBWX2vVOWbYJr0ymf5zvRb6L780k+pmiie9Mpn+bZQWtOXB28ez2Q6dd4UkPY9m/9u//upk+i031/NouaWZPI/n0euvTOaiY3/H54/cl2VXnv16XXfYENYdluYwHrjh3kU+cGmlxwDgqj/cw3VXpffXtT+6KhPq3l9fefl1lhvY/veSP1/9EKed+xUAhgxdnanvTef1Se8wYNm+izb0XK677B5uuibdh4+suyqvvjT7Prza4H2ot83HN+Cck65aJBnbcuuVY7jzuvRcWmOdOZ9Lr02YzDJzvR6WWW7OK9GvTZjMgDaeSx/baWNOO/I37PHlnVl8yT589eh9gFReevhexzNw5a5ZpB+S2DRq0c3Mbw9cXhsU234NGA5cWt1+EWnwXnON7Zm2nwJWqM7tCIy0/U7df2NeHgMukbQf0J7C6xds31P9/eK5srRlB2Bj4CFJj1THa1S3HQvsBGwCvH8l5WwPASMk/RhYz/abwDBgtO1nYY77uCXp3wjbtwPLSqq9g1xre0r19x2Bc6pM1wL9JL3vXdz2ebY3sb3JggbyADvvuSWnXngEp154BJttPYTRN47FNn974nmWWLLP+97MBizXj8WX7MPfnnge24y+cSybbD37AsXjD/2dlQcPnKMU5+03p3DyEf/HF7/5KdYe+sEFZloY++77Sa6+5kyuvuZMdthhc/70pzuwzSOP/JWlllqCgQPnHMy//faUWXX006fP4K67xrLGGqnLwQ47bsYDD6ZZ+mef/Q/Tpk1nQCcPhFdZazUmvvgKr700kenTpvPYXeNYe9icF4DWGTaEcbc9CMCTdz/KGkM/jKRZ5Q8Ak15+jVdeeHmO2vjH7hzbZSU2AAPWGMxbL03g7QmvMnP6dP59/1hW2rh9C28HrDmYae9M4b03Uv3phKf+ylKrdO4M6vaf25If/fa7/Oi332XDrdbjvpsfxjb/fPI5Fl+yD/2XnfOx779sP/os0Zt/PvlcWidy88NssGWbF+vmMHPmTB6+4xE268TyjkFrrcbE/7w663n06J3jZw3Ca9YdNoRxtz4EwBN3P8qa1fNoylvvMPIH57HzQZ9m9Y+uMcfPvDU5PR7vvPkO9183hk13HrZIc7fSYwCw+95bMPKywxl52eFstd0Qbrouvb8++djz9O3bp6G69xVW6s/YB1Kp33PPvMzUqdPpv8ySC/iphbfr57fgF5cezi8uPZzh236UUTekx+Tpx59nyb592qyNn5f/VN3OAB4c8zSrrNZ5Ezs77bElJ5z/XU44/7tsvNV6jLkp5f7HE8+xRN8+7/vQ13+5fiy+ZG/+8UR6Lo256WE22io9l156YXbucWOeYOXBaS3G229OSeuTgDuvu5+1hq45R319CF0hVzeb9+r+3pGPVJ8CtgZ2Bb4vaT3SoL7+w0n9q2nuuogF1UmIdFXhqDZuWxboC/Sq/h9vt/E92B4taesq6/mSzgAmtfW9C1D/3+8BDLP9bgf+O+2y4cfWYdy9T3PIXiexWO9efPOYvWfdduT+p3PqhUcA8JUj9+CXx6fWlBsMW5sNh6896/vuue39JTY3XTGGl/49kSt+dytX/C7Nmh/zs6+y9DJLddZdAWCbbTZm9OixfOLj30itKU/89qzbPvfZw7j6mjOZMuU9Dv7mSUydOo2Znsnmm63HF/ZO7RN3330Hjvn+Oey66yH06tWLk04+pNNnAXr27Mmu39yD879/Lp45k40+PowVVl+J2y68gVU+vCrrDF+PjXcexhWnXMzpI45j8aWWYO+jUrvM5594htGX3UaPD/REErt9ay+WXDp93pv67nv8Y9xf+ewhX+jU/PV69OzJBgd+gXt+eg6eOZPB2wyn36CVeeqK6+j/wcGsvPH6vPbP57j/zPOY9s47vDT+cZ668s/sdMoPUI8eDPni7tx94llg6P/B1fjg9l1XWbbesHV4/P6nOfqLJ7JY716M+N4+s277yZdP40e/TVV9+x22J787+fdMe28aQzZfm/U2TxfLxo1+jN///GrenPwWZ33vN6z2oVU47LSvAfC3R59hmYH9Wb4TZ/B69uzJbgfvwe+O/lVqcfrxzVlh9ZW45YIbGPSR1Vh3+BA22XkYl51yMaceeDyLL7UE+xydulHde+0YJr74KqMuuZlRl9wMwJdP+gZ9+y/FdedexX+feRGAHfb9BMsPmv8C04VR+mMwt+Fbrc39Y55m70+fTJ8+i3HUsZ+fdduIz5/ByMsOB+CXZ17PbTc8wrvvTmP3nY7n07tvxkHf+DgHH7Erpxx7OZddfDcSHH3s57t8VnLTLdbhoXv+wkGfPZk+fXpx2I9mv58c/MUz+MWl6T789qzruePm8bz37jT2++Rx7PyZzdjva5/gusvuYfyDf+cDH+hJ36UW54gf7z2v/9UiNXT4Ojxy39N89wsnslifXvzP0bOfS98/8DROOD89lw44Yk/OOyE9l9YftjZDh6Xn0h9/dT3//dcr9Oghll1hACOOTKWiLz7/Mucd/3sQDPrginzle133/jovF5z9bbYavg7LDViKfzxwDsedcQUX/PHO3LE6UczMa1HU/1ZlNlcDw21PrMpHzifN1l8k6UDgM7Y/J+l84HrbV1Q/+5btvpJ2Bn4I7Fgrs2lrdl5SD2A1289J6gU8TypF+TTwadt7S9qINDNeayj+LKkE6D5J/wc8bft0SY8Du9VmyiU9R5pxHwj8iVRmM6G6P0vZfr4qs/kD8EFgJdvfmse/yWDg37ZnSPoW8CHgBGAcdWU2tl+T9HPgFdvHSdqWVEqzYTWr/5bt06r/5qWkcpxTq+MNbD8yv8fm0deub+4C73ZYb8AaC/6mJnbVc//KHWGhjZ3YK3eEhfbJQZ32GbhLvPpu+XW4y/Upr757bmv1n5E7wkJ7a1rZg59WeC1ss/HFuSMstCn/+n1TPJGmzhzbZeOcxXps3BT3eW6LZGbe9pOSTgDukjQDGA98Gxgp6UjgFWDEAv4bN0naAHhY0lTgBuDoNr61J3CxpKVJH8d+bnuypCuB/SU9CTwA/K3uZ/4KHFzVyz8FnFudPw+4SdKLtrery/KUpGOAW6oPD9Oqn98GmGb70qoO/l5J21elMXPbFjhS0jTgLWB/269U9exXVf/dCaSSnR8Dv5P0GPAOMK9diQ4BflF93weA0cDX5/G9IYQQQggtLTaNWkQz86F5xcx8fjEz3xxiZj6/mJlvDjEzn1/MzC8602aO77JxTq8eGzbFfZ5b7AAbQgghhBAK1ZTj6y7V1IN5Sb8A5l7xdpbtkTnytKVafHvRXKffs715jjwhhBBCCKH7aOrBvO2Dc2dYENuPM7snfAghhBBC6CKKmflYNRBCCCGEEEKpmnpmPoQQQgghhHmJHWBjZj6EEEIIIYRixcx8CCGEEEIoVMxLx79ACCGEEEIIhYqZ+RBCCCGEUKToZhMz8yGEEEIIIRQrBvMhhBBCCCEUKspsQgghhBBCoaLMJmbmQwghhBBCKFTMzIcQQgghhCLFplExMx9CCCGEEEKxYmY+hBBCCCEUKual418ghBBCCCGEQsXMfAghhBBCKFJsGhUz8yGEEEIIIRRLtnNnCIWT9FXb5+XO0VGl54e4D82i9PtQen6I+9AMSs8PcR9CWWJmPiwKX80dYCGVnh/iPjSL0u9D6fkh7kMzKD0/xH0IBYnBfAghhBBCCIWKwXwIIYQQQgiFisF8WBRKr8krPT/EfWgWpd+H0vND3IdmUHp+iPsQChILYEMIIYQQQihUzMyHEEIIIYRQqBjMhxBCCCGEUKgYzIcQQgghtBBJS+TOELrOB3IHCGWRtPv8brd9VVdlWViShgDrAn1q52xfmC9R+0kSsC+whu1jJa0GrGj7wczRQmEk9QQ+BaxO3e8E22fkytRekjaa3+22x3VVlo5qhfdUSSsAJwIr295F0rrAcNu/zRytYZK2BD5se6Sk5YG+tp/Nnau9JH0M+D+gL7CapKHA12x/M2+y0JliAWxoiKSR87nZtg/qsjALQdKPgG1Jg/kbgF2AMbb3zJmrvSSdC8wEtre9jqQBwC22N80cbYEk9QOOAgYBN9q+tO62X5b2S0fSFsCPgcGkwbBIr4U1cuZqL0k3AO8Cj5OeUwDY/km2UO0k6Y753Gzb23dZmA6qe08dCHwMuL063g641/answRrgKQbgZHA920PlfQBYLzt9TJHa0j1e2ETYC3bH5G0MnC57S0yR2s3SQ8AewLX2t6wOveE7SF5k4XOFDPzoSG2R+TOsIjsCQwl/cIZUc0sXZw5UyM2t72RpPEAtidJWix3qHYaCfwduBI4SNIewBdtvwcMy5qsY34LHAaMBWZkztIRg2yvnztER9jeLneGhVV7T5V0C7Cu7f9WxysB52eM1ojlbF8m6SgA29Mllfha+BywITAOwPaLkpbKG6lxtl9IF29nKfGxCA2IwXzokBa4rDrF9kxJ06uZ4gnAqrlDNWBaVR5hgOpy8Mz5/0jTWNP2HtXfr5H0feB2SbvlDLUQXrd9Y+4QC+FGSR+3fUvuIAuj5LK5yqq1gXzlZWC1XGEa9LakZZn9fjQMeD1vpA6ZatuSavdjydyBOuCFqtTGknoBhwJPZ84UOlkM5kNHnU91WbU6/hvwR9IsZQkeltQf+A1pRvUt4L68kRryc+BqYKCkE0hXGo7JG6ndekvqYXsmgO0TJP0HGE2q8yzNHZJOBa4C3qudLKFeu3I/cLWkHsA0ZpcJ9csbq/3mVTYHlDSYHyXpZuD31fEXgNsy5mnE4cC1wJqS7gGWJ70nleYySb8G+kv6H+Ag0u+IknwdOAtYBfgPcAtwcNZEodNFzXzoEEkP2d5U0vi6urxHbG+QO1ujJK0O9LP9WOYo7VINuoYBrwE7kAZfo2wXMfsi6RRSff9tc53fGTjb9ofzJOuYedRtF1GvDSDpWeAzwOMu9BeCpMeZXTY3tFY2Z3unzNEaIulzwNbV4WjbV+fM0x7VFcJDgLOBtUjvR3+1PS1rsAZVTQUGAWsDHyfdj5tt35o1WAOqx+JC2/vmzhK6VszMh44q/rJqVdZR+8V5F1DEYL4qD/pF9SHqL7nzNMr2/87j/E1AUQN5aIm67ReAJ0odyFeKLpurBmFP2l6bdMWtGLZnSNrH9pnAk7nzdFRVXnNDtWi3mAF8veqxGCxpMdtTc+cJXScG86Gjir6sKulkYFPgkurUIZKG2z46Y6xGjKoWjl5V6iCs+jD4Y2AL0ofCMcCxtifmzNUoSf8klarcDdxtu7QBzTPAnVVHkvoyoaZvTVmn6LK5ahD2V0mr2f5X7jwdcI+kc0illm/XThZUalYzTtKmth/KHWQhPEN6PK5lzseipNdzaFCU2YQOq9qPFXlZVdJjwAa1uu1qZmx8KV09JL0JLEnqUvBudbq0OudbSXXytS5C+wLb2t4xX6rGSeoNbA5sRfpgshbwmO3PZQ3WTlW9+fuU0JqyLaWVzdVIGk3qpPIgcw7Cmn5heOmlZjWS/gJ8CHie9BjU1o8U8XsBWu/1HNonBvOhQyT1Ab4JbEmaVb0b+JXtd+f7g02iGsxva/u16ngZ4M6S3rRL11bvY0mPF9ib+gOkqzzbkF4Py5IG81/LGqxBkpaw/U7uHB0haZTtHRZ0rplJ2qat87bv6uos3ZWkwW2dt/18V2dZWJL6Ath+K3eW0PmizCZ01IXAm6RFTwBfBC4C9sqWqDEnAeOrGSWRaue/lzdSY+aq+b/T9vU583TALZL2Bi6rjvcEbs6Yp6PeIG24dAbwmwLLhIaTulAVt2NkNamwBLBctXFarbl2P1I3j2LYvqtauFvb+O1B2xNyZmovSUsDP2LONUjH2i5qHZXt56vn/1bVqbttP5ozU6OqFq0XActUx68C+xdY/hcaEDPzoUMkPWV73QWda2bV7n5fAcaTSlZetD06b6r2aaPmfx/gYdtH5UvVPlWJkEkDr1qpEEBP4K2SSoUAJH2GNCO/GTAVuJfUiWRU1mDtVPKOkZIOBb4DrAy8WHfTG6QPVudkCdYBkj4PnArcSXptbAUcafuKnLnaQ9KVwBPABdWpLwFDbe+eL1XjqufT/5DazELaROo822fP+6eai6R7STvx3lEdbwucaPtjWYOFThWD+dAhki4GzrF9f3W8OXCw7f3zJmsfSV8hbaYxCHiE1OrxvlJqPFug5l+kTXJKXOzXJklrk/qbfwcYaHvxzJHaRdIDtjefq83so7aH5s7WXpK+XdKAqy2SHgV2qs3GVxvB3VbC49BWW+ISWxVX76vDbb9dHS9J+r1QxPsqtP3aLe31HBrXI3eAUBZJj1dveBsD90p6rupTfR+wSd50DTmUNLP9fNVacENgct5IDetf9/els6XogKoDz59z51gUJF0p6R+kjVqWBPYHBuRN1ZA5doyU9F3K2zHy15IOkXRF9fUtpd0vS9JjrrKaiZTzO3qKpC1rB5K2AKZkzNNRYvaVQqq/ax7f26yekfQDSatXX8eQOtyEFhY186FRn84dYBF51/a7kpDU2/ZfJK2VO1QDiq/5pzXawEH1WNiescDvbE6tsGPkL4Fe1Z+QyjzOJZXRleImvX8H2Bsy5mnEN4ALqtp5gEnAgfnidNhI4AFJtV7/n6WcXc1rDgJ+QioVqjWnOChrotDposwmLBRJA4E+teNSyiaqN+sRpJKI7Um/fHrZ/mTWYA2QtBJzLpZ7KWeeRrVCGziAagb4G8y5+O9XJbVqLV2rlBZI2p20/gLS4suiNpCqNuzC9hu5s3SUpI2Y8zEYnzNPCO0Rg/nQIVUnldNJC88mAIOBp21/NGuwDqhawi0N3FTKrnlK277fXusWUW2Ys63ta/Ima79WaQMn6f9Is8L1i/9m2G7qWWFJZ1Pt4NwW24d0YZyFImkcsJftf1bHawBX2N4ob7L2k/Rl0sLpv+fO0ihJJwKn2J5cHQ8AjrB9TN5kjVHayfxJ229Wx/2AdWw/kDdZ+1X7d+w112PxB9ufyJssdKYYzIcOqRZrbU9aoLWhpO2A/Wx/OXO0bmEeC85mLWAsRVVn+2HbI6sFf31tP5s7VyNKnRWWdMD8brd9wfxubyaSdiCVSDxDusIzGDjI9u1ZgzVA0k9IHWxWJ+1iO5o0M/xIzlzt0dZ7j6RxJX2YgnQ/gI2qNT1I6kHqElbM/ZjHY1Hc74bQmKiZDx01zfZEST0k9bB9h6Sf5Q7VjbS1MK6o17PSToWbkHZMHUma3b6YtItqSWZIWnOuWeGmr59v72Bd0tm2v93ZeRbSGODDpOcSwF8zZukQ2z8CkLQ4qT3ikcDPSC1bm13Pau3RezDrPvTOnKkj5LoZTtszlTaFK8lMSavVSl6rK6Axa9viSnuShuYxWWmHudHAJZImULcFeeh0D0s6A/hFdfwt0mxeST5H6iI0DsD2i5KWyhupQ44E7pBUPys8Im+kRaqED1f3VbOnj9VOVKU3Jc2oHkP6t+5L2vviu6TFiyW4BBglaWR1PILZZWcleUbSIaTF05B2OS+tE8z3gTGS7mL2fgVfzRspdLYoswkdUvXffZf0ZrEvqeb8ktJ2vyxV9e//A2DH6tStwPG1/sglkPSg7c1ql+NL7OlcI6k3dbPCtRnKVtDM5RKSViR14bmYtAt1/Q6wv7K9dq5sjao+fEwntWy9i/RaKOZ5JGln6t6PbBe3m3PV0OHnpBJSA6OA75SyE2+NpOVIe6cA3G/71Zx5QueLwXwIhas2jFqytA4SVT/zDwM7kdo7HgRcWuLmP1Wf9tWpu9pp+8JsgRahJh/MH0BqgbgJ8BCzB/NvABfYvmoeP9qUqgWXW5C6qewFTLC95fx/Kr/qg/iUqixlLdIH2xujo1PXq3r8P2L7bUn7ka5OnVVaY4HQmBjMh4ZIepO26+9qbQX7dXGkbknSpaT+4DNIg5h+pDfsU7MGa4CkI0idkGoz8bfYvjVjpA6RdBGwJmkn4VqtvEvqBjM/JSyek7SH7Svnc/sBzb6gV9IQUknENqQPJy+QFsD+MGuwdpA0lpR9AGn9wsPAVNv7Zg3WIEmnAMeTNry6ifTedJjti7MGa0C1qeNQUvaRpD75n7e9TdZgoVPFYD6EAtW62UjalzTz8j1gbEklKtUC2M8DrwF/BC63/XLeVI2T9DSwrlv0zVTSgbbPz51jYTTz1YUaSdeT1iCNAR4qaVa7rlTu28Ditk9pq+NWs6t7X/0caYPEw0ntQpu6M1W9usfih8B/bP+2hOd/WDixADaEMvWqNiv6LHCO7WmSihpM2v4J8BNJ65N2u7xL0r9t77iAH202TwArAv/NHaQjJH2EtIh3MHOWCW1f/Xl+nmSLlBb8LXnZnu/u2pKutL1HV+VpkCQNJ62fqrUnLqELz9xqz/9PkSYXXpea/qkztzclHQXsB2xdtdfslTlT6GQxmA+hTL8GngMeBUZX7ceKqpmvMwF4CZgIDMycpSOWA56S9CAwa8Gi7d3yRWrI5cCvgN9QQEvNDirqg+48rJE7wHwcChwFXG37yao96x2ZM3XE9dXO1FOAb1R7X7ybOVOjvkBaDP5l2y9JWg0opvwydEyU2YTQApSmj3ranl4dl1Aj/E1Smc3ypAHlZbafypuqcdUOwu9j+66uztIRksba3jh3js5UQt3/gpRcKlHIXgUASFoGeN32jGph71K2X6pu26nEdT31JN1ne3juHGHRipn5EFpAVa89ve7UoTR/n+dVSW3fmn6HywVYwvaN9SckfZ3UXrAE11UfrK5mzisLr+WLtMjdkztAN1fCXgXAnM/7qtVvfbvfn5LaAJesT+4AYdFraxfJEEL5mr7Q0/ZRLTCQB/iBpO1rB5L+F/hMxjyNOoBUM38vaeOxsaRuJMWQdKKk/nXHAyQdXzu2/a08yRappn9NdwOt8BhEOUYLisF8CK0p3rC7zm7AiZK2knQCsDkFDeZtf7CNr2auz27LLrYn1w5sTwI+mTFPQyT1lHTJAr7t/3VJmDA/8b4amlKU2YTQmlphBqkItl+VtBtwG2lWe8+S2lRKGkMqCbobuMf2m5kjdURPSb1rO6ZKWhzonTlTu1X12YMlLWZ76jy+55auzrUIxftR84jHogXFYD6E1hQ1wp2sjQ3UFiN1HNlTUkkbqH2JtOHPHsCpkt4jbVZ0WN5YDbkEGCVpZHU8guZfMzK3Z4B7JF1LXZ227TPyRVpkzsodYBF5LneABZG07tyNBCRta/vO6vBLXZ8qdLboZhNCgSStAJwIrGx7F0nrAsNt/zZztFAgSSuRdh7dCtgO+JftnfOmaoykXYAdqsNbbd+cM0+jqk3U3qfaj6GpLWivgmYnaff53W77qq7KsrAkPQFcBJxCWux6CrBJdLBpbTGYD6FAkm4kbdX9fdtDJX0AGG97vczRuiVJq/D+gczofInaT9I/gVeBS0mlNo/Ynpk3VfclaQnb7+TO0QhJj5L2KhhL3V4FtsdmC9WAuis6bbHtg7oszEKq2mn+FNgYWIp01eqn8ZpubVFmE0KZlrN9WbXTH7anS2rVDX+amqSfkjZqeYrZAxkDRQzmgZ8DWwL7ABuSduIdbfufeWO1XzWz+lPSpmOqvkoqdaLaQfW3QF9gNUlDga/Z/mbeZO0y3fa5uUN0lO0RuTMsQtNIm14tTpqZfzYG8q0vBvMhlOltSctS1WxLGga8njdSt/VZYK3a4svS2D4LOEtSX1Kt+Y+BQUDPnLkadAqwq+2ncwdZCD8DPgFcC2D7UUlb543UbkXvVSDp8PndXti6hYeAPwGbknan/pWkPWzvlTdW6EwxmA+hTIeTfumvKeke0i6q8WadxzNAL+oGMSWRdDppZr4vcB/wQ1K5TUleLnwgD4DtF9JmzrOUcrXtgOrPI+vOmbQgvARL5Q6wCH3Zdm2fiP8Cn5EUi15bXAzmQyjTk6QFi2uRSgr+Suwbkcs7wCOSRjHnrOQh+SI15D7gFNsv5w6yEB6W9EfgGuZ8DIpZuAi8IOljgCX1Iu3iXMQHFNsfzJ1hYZSwyLgBYyXtB6xh+1hJq5F+P4QWFgtgQyiQpHG2N1rQudD5JB3Q1nnbxbRGrPrk10o67rJ9Xc48jZrHAsbSFi4uR2rhuCPpg/nNwKG2J2YN1k6ShgDrkuq0AbB9Yb5EjZPUB/gy8FHmvB8lPY/OBWYC29teR9IA4Bbbm2aOFjpRzMyHUBBJKwKrAItL2pDZG4D0A5bIFqwbK2nQ3hZJJwGbkbpeABwiabjtozPGakgrLGC0/Sqwb+4cHVG11dyWNJi/AdgFGAMUNZgntXT8C2ntwrGkx6OIqyN1Nre9kaTxkHZDlrRY7lChc8VgPoSyfAI4kLRAsX5R1htAMYOvViLpWdrY5t12KfXCnwI2qHW8kHQBMJ6Cnk+SBgFnA1tUp+4mzWpJB37xAAAQ9klEQVT/O1+qxkhagzQzP4z0fLoPOMz2M1mDtc+ewFBSe9wR1T4YF2fO1BEfsr2XpM/YvkBSrV1rSaZJ6sns5gjLk2bqQwuLwXwIBalmgS+ouhNcmTtPAGCTur/3IS1EXiZTlo7qD9Q6jyydM0gHjST1ya8tAt+vOrdTtkSNuxT4BfC56nhv4PfA5tkStd8U2zMlTZfUD5gArJo7VAdMq/6cXJUNvURqd1qSn5O6Cg2UdALpg9YxeSOFzhY18yEUqCq3OYHYAbYpSRpre+PcOdpD0j7AycAdpLKtrYHv2f5j1mANkPSI7Q0WdK6ZSXrM9vpznXvU9tBcmdpL0i9JV3L2Bo4A3iJtPlZU+ZOkrwBXAuuTPgz2BX5o+1dZgzVI0tqk3ZAFjGqFTk9h/mIwH0KBYgfY5iGpftFxD9JM/TdKGITVSFqJ1Jca4EHbL+XM06iqk9BI0kw2pA2wRtjeIV+qxlSbj00C/kAqkfgCMAA4FYrq2b460M/2Y5mjdCuS5ns1sJTnT+iYGMyHUCBJD9neVNJ42xtW54qaiWwVku6oO5wOPAucbruYdnCSVgEGU1d6abuUHWyRNJhUMz+cNBC+F/i27ReyBmtAtfZiXtzMazCUmuPvy5ztEFe0/WDmaA2R1BvYA1idOV8Lx+bK1F51a3dqTRFqg7vabshN+/wJCy9q5kMoU+wA2yRsbze/2yUd0Mwdb6oZ4S+Q9i6oLZQzUMxgntR55ADbk2DWLOVpQDEtBRfUq13STrZv7ao8DfolVTtE0mPxJqlcpbR2iH8ivY+OpbBN4Erv9R8WTszMh1CgqrTjbGAI8ARpB9g949J282n2/v+S/gqsb7uowUu9+itU8ztXsmZ+HtWyzXWlsIh6/3qSnrA9JHeOhSFp1NzlZW2dC60lZuZDKEzVdmwb5toB1va0+f5gyEUL/pasngF6UdhM5Fx6SBow18x8q/1+a+bnUau0Q7xX0nq2H88dpFHVhldLAstVG0XV70GySrZgoUu02ptdCC3P9gxJ+9g+k1QaEZpbs1/+fAd4pFpEOmtAb/uQfJEadjpwn6TLq+O9SN2eWkkzP4+Kboco6XHSv+8HgBGSniG9Fmr15uvP7+ebxNeA7wArA+Pqzr8BnJMlUegyUWYTQoEknUmaTf0j8HbtvO1x8/yhkEWzl3tIOqCt881c59+Wqj3r9tXh7bafyplnUWvmMhsoux1itYB6nmw/31VZFpakb9s+O3eO0LViMB9CgebqoFJj29u3cT5kJOkc29/KnWNeJG1se+xc5z5t+/pcmcL7SbrK9u65c8xLVdqxKnN2gSlqcqFqJPCk7Ter437AOrYfyJus/SQtDnwD2JJ0teFu4Fe2380aLHSqGMyHEMJCkHQicIrtydXxAOAI20WUGUgaB+xv+4nqeB/gO7ZL2Hm0ZUhagrTh0mq2/0fSh4G1SvhQJek44EDgn8wuBypuckHSeGAjVwMjST2Ah5v5isjcJF1G6iZ0cXXqi0B/23vN+6dC6WIwH0KBSh9AtpJ5dFJp6pKIepLWAK4g/dLfCtgf+LTtaHXahST9kdQScX/bQ6rB/b0l7B1RdURaz/bU3FkWxjx2En7fzrzNTNJTttdd0LnQWnrkDhBC6JBdagN5gKqLxycz5unOelabzQCzLnP3ns/3NxXbzwB7A1eRNsz5eAzks1jT9inANADb79DcHWzqPQH0zx1iEXhG0iGSelVfh5K6PZVkXFUuBICkzYGHM+YJXSC62YRQpp6Setd6g5c2gGwxlwCjJI2sjkcATb94tK6DR80yQE/gAUmUNBvZIqZWr+NaicealNMu9CRgvKQnmLMj0m75InXI10mdeY4hPQ6jgK9mTdROda/nXqQWm/+qjgcDf8mZLXS+KLMJoUCS/h+wK1A/gLy2mtkLXUzSLqROHgC32r45Z572aG8Hj/r+7aHzSNqJNIhcF7gF2AI40PadOXO1h6QngV8Dj1PXX972XdlCdQJJR9k+KXeOtrRSR57QuBjMh1CoEgeQoTwl1f+XTtKywDBSec39tl/NHKldJD1ke9PcOTpbCa8FSau1dd72v7o6S+g6MZgPIYSFIGl34KfAQNIgrLbRTL+swRaRZu+TXzpJ8x0cltDeUdIZpPKaa5mzzKbpszeihNdCXbmNgD7AB0k7hH80a7DQqaJmPoQCtfoAsjCnALuWtElOg2LGp3OdPp/bzOyNsJpZbYA7rO5cKdkb0fSvBdvr1R9XHxa/mSlO6CIxmA+hTK0+gCzJy/E4hI6yvV3uDAtrQfdB0gGl7Sg8D6V0F5rF9riqo01oYTGYD6FMMYBsHg9XPcKvYc4Sg6vyRVqkihvAlEhSH9IMaivu3HkoBXR4aofLcwdYEEmH1x32ADYGXswUJ3SRqJkPoUCSzgJWpHUHkMWoa0lZz7YP6vIwHbCgLewlLWP7tZwZu4NW3rmzhFpzAEkfAc4FVqg27lof2M328ZmjtZukH9UdTgeeA65skQ+FYR5iMB9CgUofQIbm0Qpb2LeCVt65s4QuMACS7gKOBH5d+/Ah6QnbQ/Im65jqtdzX9hu5s4TOFWU2IRTI9ojcGUIiaRBwNqkvOKTyiENt/ztfqobIdbM6tmdKit8NXW+cpGG274eW27mzlFKtJWw/KM0Rd3quMB0h6VLS5lczgIeAfpLOsn1q3mShM/XIHSCE0DhJgyRdLWlC9XVlNagMXW8kqSXfytXXdczezKsErbCFfbEkPS7pMVJt872SnpP0LHAfsEnedIvMPbkDtNOr1c67tatUewL/zRupYetWM/GfBW4ktab8Ut5IobNFmU0IBZJ0K3ApcFF1aj9gX9s75UvVPUl6xPYGCzrXrCQNJG1hvz2zt7D/ju0JWYN1E62wc6ekE4FTbE+ujgcAR9g+Jm+yxkhaAzgP+BgwCXgW2M/2czlzNaLajXcD0u+Hc2zfJelR20MzRwudKGbmQyjT8rZH2p5efZ0PLJ87VDc1UdJ+knpWX/sBE3OHag9JPYEzbe9te6DtFWx/MQbyXcf28/VfwBTSh6raVwl2qQ3kAWxPAj6ZMU+H2H7G9o6k99K1bW9Z0kC+8mvSotclgdHVh8WomW9xURcZQpkmVoPG31fH+1DIALIFHUSqmT+TNPi6FzgwZ6D2sj1D0mBJi9memjtPdyZpN9IGUisDE4DBwNNACTt39pTU2/Z7AJIWB3pnztRuc7VzrD8PgO0zujTQQrD9c9KVNgAk/QvYru64VXr+hzoxmA+hTMUOIFvQscAB1WwkkpYBTiM9RiV4BrhH0rXA27WTJQ1gWsRxpB1Ub7O9oaTtSOVzJbgEGFXXZWsEZfWVXyp3gM5SLW6vX8TbKj3/Q50YzIdQptIHkK1k/drjAGD7NUlN31O7zj+rrx608KCmANNsT5TUQ1IP23dI+lnuUO1h+6fVIt4dqlPH2b45Z6ZG2P5J7gxdqJTOQqEBMZgPoUylDyBbSQ9JA+b6YFXMe2ttICNpCdvv5M7TjU2W1BcYDVwiaQJ1V0qane0bSd1TilXtwvtlUmlTn9r5Ftu/o5R1GKEBsQA2hDL1qDpGAOUNIFvM6cB9ko6TdByp5OmUzJnaTdJwSU8Bf6mOh0r6ZeZY3dFnSItfDwNuIl0t2TVronaStLukv0t6XdIbkt6UVOKiy4tIO2t/ArgLGETalbeVxMx8C4rWlCEUSNL+wNHA5dWpvYATbF80758KnUXSuqTWjgC3234qZ55GSHoA2BO4thV2vQxdT9I/gF1tP507y8KQNL5ar/CY7fUl9QLutj0sd7ZFRdI5tr+VO0dYtGImL4QC2b5Q0sPMHkDuXtIAstVU//bF/vvbfmGuXS9n5MrS3Uh6k7ZLH0Rav9iviyN1xMulD+Qr06o/J0saArwEDMyYp2EL6vkfA/nWFIP5EApV+gAyNI0XJH0McDUTeSipJWLoArZbYdHxw5L+CFwDvFc7afuqfJE65Lxq8HsMaVfnvsAP8kZq2C62j64d2J4k6ZOk+xRaVAzmQwihe/s6cBawCvAicDNwcNZE3Ui13mWebL/WVVkWQj/gHeDjdecMFDGYn6vP/Ijqz19Ufy7ZxXEWVtE9/0PHxGA+hBC6MduvAvvmztGNjSUNfNtamGhgja6N0zjbIxb8XU2tdnVkLWBT0qw8pAXID2ZJ1HGl9/wPHRALYEMIoRuTtAZpZn4YafB4H3CY7WeyBgvFkDSItIndFtWpu4FDbf87X6rGSRoNfMr2m9XxUsCfbW+dN1ljJO3C7J7/t5bU8z90TLSmDCGE7u1S4DJgJWBlUoek32dN1A1JulLSJyWV+Ht5JGk2e+Xq67rqXGlWAKbWHU+tzhXF9o22v1t9xUC+GyjxTSOEEMKis4Tti2xPr74upm7DnNBlziWVO/1d0smS1sodqAHL2x5Z9xw6H1g+d6gOuBB4UNKPJf0YeAA4P2uiBrVQz//QgCizCSGEbkzST4FJwB9IZTZfAAYAp0IxCzBbhqSlgX2A7wMvAL8BLrY9bb4/mJGkUaSZ+NoVnX2AEbZ3mPdPNSdJGwFbVYejbY/PmadRrdLzPzQmBvMhhNCNSXp2PjfbdtMvwGwVkpYF9gO+ROosdAmwJbCe7W0zRpsvSYNJNfPDSR8I7wW+bfuFrMG6IUn32N5iwd8ZWkkM5kMIIcyTpJ1s35o7R6uTdDWpm8pFwPm2/1t328O2N8kWbgEkXQB8x/ak6ngZ4DTbB+VN1v1IOgtYkfJ7/ocGxGA+hBDCPEkaZ3uj3DlanaTPAzfZfkPSMcBGwPG2x2WOtkCSxtvecEHnQuera0lZz/HBqrVFn/kQQgjz01b/87DoHWP7MklbAjuS1iycC2yeN1a79JA0YK6Z+RhfZNACPf9DB0Q3mxBCCPMTl2+7xozqz08B59n+M7BYxjyNOB24T9Jxko4j1cyfkjlTtyRpkKSrJU2ovq6s9gEILSwG8yGEEEJ+/5H0a1I3oRsk9aaQ39G2LwR2B16uvna3fVHeVN1Wq/T8Dw2ImvkQQgjzJOkq27vnztHqJC0B7Aw8bvvvklYidbG5JXO0UBBJj9jeYEHnQmuJwXwIIXRzkoYA61K3WVQ12xpCKEgr9fwP7ReD+RBC6MYk/QjYljSYvwHYBRhje8+cuUIIjYue/91TEfV4IYQQOs2ewA7AS1UnjKHA0nkjhRA66FjgANvL2x4IHAT8JHOm0MliMB9CCN3bFNszgemS+gETgFUzZwohdMz6tRahALZfA6Lff4uLPrAhhNC9PSypP/AbYCzwFnBf3kghhA6Knv/dUNTMhxBCAEDS6kA/249ljhJC6ABJ+wNHA5dXp/YCTohWoa0tBvMhhNDNSdoN2Lo6vMv2dTnzhBA6TtK6wPbV4e22n8qZJ3S+GMyHEEI3JulkYFPgkurUPsBDto/OlyqEEEJ7xWA+hBC6MUmPARtUi2CR1BMYb3v9vMlCCCG0R3SzCSGE0L/u79GWMoQQChIrnEMIoXs7CRgv6Q5ApNr57+WNFEIIob2izCaEELo5SSsDXwHGA0sCL9oenTdVCCGE9oiZ+RBC6MYkfQU4FBgEPAIMI/WZ335+PxdCCKE5RM18CCF0b4eSutk8b3s70m6Rk/NGCiGE0F4xmA8hhO7tXdvvAkjqbfsvwFqZM4UQQminKLMJIYTu7d+S+gPXALdKmgQ8nzlTCCGEdooFsCGEEACQtA2pNeVNtqfmzhNCCGHBYjAfQgghhBBCoaJmPoQQQgghhELFYD6EEEIIIYRCxWA+hBBCCCGEQsVgPoQQQgghhELFYD6EEEIIIYRC/X+04fQDXZPZ1AAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 864x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.figure(figsize=(12,10))\n",
    "sns.heatmap(df_raw.corr(), annot=True, cmap=\"YlGnBu\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df_raw.sample(10000, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "num_features = ['aoa', 'concrete_score', 'verb2','conc_unknown', 'aoa_perc_known_lem', \n",
    "                'conc_total', 'syllable_per_word','conc_mean_score', 'dale_chall_score', 'conc_subtlex_score']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "id": "Laz2lFWYH2Ab"
   },
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "# for mlp & svm\n",
    "sc = StandardScaler()\n",
    "df[num_features] = sc.fit_transform(df[num_features])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "id": "84HI4VKfH2DH"
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "\n",
    "X = df[num_features].to_numpy()\n",
    "y = df['label']\n",
    "\n",
    "Train_X, Test_X, Train_Y, Test_Y = train_test_split(X,y,test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(8000, 10)"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Train_X.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "zuc4BRcoITGC"
   },
   "source": [
    "# MLP"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "-i14KcfpH2H9",
    "outputId": "01c67c89-ce1d-468c-ad95-4649f813be21"
   },
   "outputs": [],
   "source": [
    "from sklearn.neural_network import MLPClassifier\n",
    "from sklearn.metrics import f1_score, accuracy_score\n",
    "\n",
    "mlp = MLPClassifier(hidden_layer_sizes=(150), random_state=1, max_iter=300).fit(Train_X, Train_Y)\n",
    "mlp_predict = mlp.predict(Test_X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.712375"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mlp.score(Train_X, Train_Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(65.55, 66.79518072289157)"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mlp_acc = accuracy_score(Test_Y, mlp_predict)*100\n",
    "mlp_f1 = f1_score(Test_Y, mlp_predict)*100\n",
    "mlp_acc,mlp_f1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "id": "CUBaxxhGISc7"
   },
   "outputs": [],
   "source": [
    "import pickle\n",
    "\n",
    "pickle.dump(mlp, open('trained_models/mlp_10simplefeat_1w.pkl','wb'))\n",
    "# reloaded_clf = pickle.load(open('drive/MyDrive/Milestone2/trained_models/mlp_simple6feat.pkl','rb'))\n",
    "# mlp_predict = reloaded_clf.predict(Test_X)\n",
    "# accuracy_score(Test_Y, mlp_predict), f1_score(Test_Y, mlp_predict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "CC0_yqu6Ii-O"
   },
   "source": [
    "# SVM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "id": "KVxhMqm_Ij8_"
   },
   "outputs": [],
   "source": [
    "from sklearn import svm\n",
    "\n",
    "SVM = svm.SVC(C=0.5, kernel='linear', degree=3, gamma='auto')\n",
    "SVM.fit(Train_X, Train_Y)\n",
    "predictions_SVM = SVM.predict(Test_X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "phs9ho64Zy8v",
    "outputId": "1daf4761-be94-4a3b-c8a0-36882808db66"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(65.2, 63.900414937759344)"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import f1_score, accuracy_score\n",
    "\n",
    "svm_acc = accuracy_score(predictions_SVM, Test_Y)*100\n",
    "svm_f1 = f1_score(predictions_SVM, Test_Y)*100\n",
    "svm_acc, svm_f1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.657125"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SVM.score(Train_X, Train_Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "id": "E4MGOvk3IodN"
   },
   "outputs": [],
   "source": [
    "import pickle\n",
    "\n",
    "pickle.dump(SVM, open('trained_models/svm_11simplefeat.pkl', 'wb'))\n",
    "# reloaded_svm = pickle.load(open('drive/MyDrive/Milestone2/trained_models/svm_6simplefeat.pkl', 'rb'))\n",
    "# predictions_SVM = reloaded_svm.predict(Test_X)\n",
    "# accuracy_score(predictions_SVM, Test_Y)*100, f1_score(predictions_SVM, Test_Y)*100"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### RBF kernel doesn't increase overall performance compare to linear kernal.  It has higher F1 score but lower accuracy, which is our main evaluation metric. Moreover linear kernel have better generalization ability."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(64.7, 65.04950495049505)"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SVM2 = svm.SVC(C=0.5, kernel='rbf', degree=3, gamma='auto')\n",
    "SVM2.fit(Train_X, Train_Y)\n",
    "predictions_SVM2 = SVM2.predict(Test_X)\n",
    "svm_acc2 = accuracy_score(predictions_SVM2, Test_Y)*100\n",
    "svm_f12 = f1_score(predictions_SVM2, Test_Y)*100\n",
    "svm_acc2, svm_f12"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.677"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SVM2.score(Train_X, Train_Y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ZM_eWW4lI4Ps"
   },
   "source": [
    "# Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "id": "oU-uZPENQzbX"
   },
   "outputs": [],
   "source": [
    "# without feature scaling\n",
    "\n",
    "df = df_raw.sample(10000, random_state=42)\n",
    "X = df[num_features].to_numpy()\n",
    "y = df['label']\n",
    "\n",
    "Train_X, Test_X, Train_Y, Test_Y = train_test_split(X,y,test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(8000, 10)"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Train_X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "id": "wP2J_T7VI51G"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(max_depth=70, min_samples_leaf=4, min_samples_split=10,\n",
       "                       n_estimators=800)"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "rf = RandomForestClassifier(bootstrap= True,\n",
    " max_depth=70,\n",
    " max_features='auto',\n",
    " min_samples_leaf=4,\n",
    " min_samples_split=10,\n",
    " n_estimators=800)\n",
    "\n",
    "rf.fit(Train_X, Train_Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.920875"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_training_score = rf.score(Train_X, Train_Y)\n",
    "rf_training_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "JGSe28yWepVV",
    "outputId": "623bc80e-6b81-44ed-dc86-13059c862830"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.671, 0.6821256038647343)"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import f1_score, accuracy_score\n",
    "\n",
    "rf_preds = rf.predict(Test_X)\n",
    "rf_acc = accuracy_score(Test_Y, rf_preds)\n",
    "rf_f1 = f1_score(Test_Y, rf_preds)\n",
    "rf_acc, rf_f1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "id": "XtNQtV6xJA2Y"
   },
   "outputs": [],
   "source": [
    "pickle.dump(rf, open('trained_models/randomforest_simplefeat_1w.pkl', 'wb'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The model is overfitting in this case, try smaller parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "max_depths = [5,10,15]\n",
    "\n",
    "training_scores = {}\n",
    "testing_scores = {}\n",
    "for depth in max_depths:\n",
    "    rf2 = RandomForestClassifier(\n",
    "     max_depth=depth,\n",
    "     max_features='auto',\n",
    "     min_samples_leaf=4,\n",
    "     min_samples_split=10,\n",
    "     n_estimators=100)\n",
    "\n",
    "    rf2.fit(Train_X, Train_Y)\n",
    "    rf2_training_score = rf2.score(Train_X, Train_Y)\n",
    "    rf_preds2 = rf2.predict(Test_X)\n",
    "    rf_acc2 = accuracy_score(Test_Y, rf_preds2)\n",
    "\n",
    "    training_scores[depth] = rf2_training_score\n",
    "    testing_scores[depth] = rf_acc2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max depth training scores:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{5: 0.684625, 10: 0.7815, 15: 0.880125}"
      ]
     },
     "execution_count": 126,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('max depth training scores:')\n",
    "training_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max depth testing scores:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{5: 0.673, 10: 0.669, 15: 0.669}"
      ]
     },
     "execution_count": 127,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('max depth testing scores:')\n",
    "testing_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [],
   "source": [
    "estimators = [50,100,150,200]\n",
    "\n",
    "training_scores2 = {}\n",
    "testing_scores2 = {}\n",
    "\n",
    "for est in estimators:\n",
    "    rf2 = RandomForestClassifier(\n",
    "     max_depth=10,\n",
    "     max_features='auto',\n",
    "     min_samples_leaf=4,\n",
    "     min_samples_split=10,\n",
    "     n_estimators=est)\n",
    "\n",
    "    rf2.fit(Train_X, Train_Y)\n",
    "    rf2_training_score = rf2.score(Train_X, Train_Y)\n",
    "    rf_preds2 = rf2.predict(Test_X)\n",
    "    rf_acc2 = accuracy_score(Test_Y, rf_preds2)\n",
    "\n",
    "    training_scores2[est] = rf2_training_score\n",
    "    testing_scores2[est] = rf_acc2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "n estimators training scores:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{50: 0.783875, 100: 0.787125, 150: 0.785875, 200: 0.787625}"
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('n estimators training scores:')\n",
    "training_scores2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "n estimators testing scores:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{50: 0.671, 100: 0.6705, 150: 0.674, 200: 0.668}"
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print('n estimators testing scores:')\n",
    "testing_scores2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### With this small number of data points, the grid search results do not seem to be realiable and scalable. More research need to be done on this part."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "rf_acc= rf_acc*100\n",
    "rf_f1 = rf_f1*100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "id": "QhX8wihUJA5J"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    accuracy\t f1 score\n",
      " mlp: 65.55 \t 66.79518072289157 \n",
      " SVM: 65.2 \t 63.900414937759344 \n",
      " rf:  67.10000000000001 \t 68.21256038647343\n"
     ]
    }
   ],
   "source": [
    "print('    accuracy\\t f1 score')\n",
    "print(' mlp:', mlp_acc,'\\t', mlp_f1, \n",
    "      '\\n SVM:', svm_acc,'\\t', svm_f1, \n",
    "      '\\n rf: ',rf_acc,'\\t', rf_f1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "id": "34QzogcYJA8w"
   },
   "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>accuracy</th>\n",
       "      <th>f1 score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MLP</th>\n",
       "      <td>65.55</td>\n",
       "      <td>66.795181</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SVM</th>\n",
       "      <td>65.20</td>\n",
       "      <td>63.900415</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>67.10</td>\n",
       "      <td>68.212560</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              accuracy   f1 score\n",
       "MLP              65.55  66.795181\n",
       "SVM              65.20  63.900415\n",
       "RandomForest     67.10  68.212560"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "score_df = pd.DataFrame([[mlp_acc, svm_acc, rf_acc],[mlp_f1, svm_f1, rf_f1]]).T\n",
    "score_df.columns=['accuracy', 'f1 score']\n",
    "score_df.index = ['MLP','SVM','RandomForest']\n",
    "score_df.to_csv('scores_1w_simplefeats.csv')\n",
    "score_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "name": "Milestone2_simple_features.ipynb",
   "provenance": []
  },
  "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": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}