Land_data_classification_article_Part3.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Приклад простої нейронної мережі"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:36: FutureWarning: Passing a negative integer is deprecated in version 1.0 and will not be supported in future version. Instead, use None to not limit the column width.\n",
      "C:\\Users\\User\\Documents\\article\\data_print.py:50: UserWarning: You have mixed positional and keyword arguments, some input may be discarded.\n",
      "  plt.legend(fracs, labels=labels, loc='best', bbox_to_anchor=(-0.1, 1.), fontsize=8)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAd0AAAD3CAYAAABYfY6KAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOzdd1ST1xsH8O+bQQYEhLBBNgRxoEYi4Kqj7lWlWkeLtVZbZ6Xu1l1tpYoVV62rLhwVFcX5s1oHwyhSFJGA7C1LdkLG+/sDQwFxVKug3s85nmPfeZPYPLn3vfd5KJqmQRAEQRDE68do6gYQBEEQxPuCBF2CIAiCeENI0CUIgiCIN4QEXYIgCIJ4Q0jQJQiCIIg3hARdgiAIgnhDSNAlCIIgiDeEBF2CIIjXhKKoVIqi+tT5byuKopIpivJvynYRTYcEXYIgiDeAoigTABcBnKNpel5Tt4doGiToEgRBvGYURbUAcAGAFMC0OtsnUBSlpiiq/PEfmqIop8f7Pqco6j5FUWWPe8dTGlxzGEVRf1MUVUpRVBJFUf0fb/+LoqhJj//OoCjqLkVRmW/sxRLPRIIuQRDE66UH4CwAFoCJdP3cuwwA4TRN69E0rdfgvIcABgPQB/A5gPUURXUEAIqiJAD2ApgLoAWA7gBSG7m3LwDD/+6lEK+KBF2CIIjXayuAcgDWALo02KcDoLqxk2iaPk3TdBJd4wpqesrdHu/+AsAumqb/R9O0hqbpLJqm4+ueT1EUF8BiACv/w9dCvCISdAmCIF6veABDAMwDsJOiKF6dfUYAihs7iaKoARRFRVIUVURR1CMAAwEYP97dEkDSc+47C8B5ALJXaTzx3yJBlyAI4vVaRdO0nKbp7QDSUb/n6QIgoeEJFEVxAAQDWAvAjKbpFgDOAKAeH5IBwPEZ9zQCMB3A8ldvPvFfIkGXIAjizfkSwGSKoiQURXUBMBxASCPH6QDgAMgHoKIoagCAvnX27wTwOUVRvR9PlrKiKMq1zv5vAOykaTr39bwM4mWRoEsQBPGG0DSdDGAJgBsA9gGYQ9P0jUaOKwMwE8AR1Aw/jwVwss5+KR5PrgJQAuAKANs6l2CippdMNDMUKWJPEARBEG8G6ekSBEEQxBtCgi5BEARBvCEk6BIEQRDEG0KCLkEQBEG8ISToEgRBEMQbQoIuQRAEQbwhJOgSBEEQxBtCgi5BEARBvCGsZ+2MiooyZbFYOwC0AQnQBPE6aQDEqlSqSWKx+GFTN4YgiNfjmUGXxWLtMDc3b2ViYlLMYDBI6iqCeE00Gg2Vn5/vlpubuwPA0KZuD0EQr8fzeq9tTExMSknAJYjXi8Fg0CYmJiWoGVUiCOId9bygyyABlyDejMf/r5HHOATxDmsW/4Onpqay3dzcWnE4nI5KpfKJ/WVlZQwfHx87iUQi6ty5s8u1a9f4ADB37lyLhQsXmgNAaGioYObMmZYAcOHCBd0ePXo4qdVq2NnZtZFIJCKJRCIaPHiww4u0R61WY/LkydZeXl4uAwYMcACAI0eO6Nvb27cWi8Wixs4JCwvjeXt7O3fu3Nnl0KFDBgDw1VdfWXfs2NFVLBaL7t69y6l7/MGDBw3c3d1dZ82aZQkAt2/f5i5YsMC8sWsvWbLELCwsjHfp0iXdDh06uIrFYtEXX3zRUrv/+PHj+p6eni4SiUSkfW8A4Pfff29hbm7e7nnHvc18fX1bPm3f8OHD7TUazZtsDkEQxDM985luXXKl2p3LZr7w8Y2cr+KymTGN7TM1NVVduXIlYdCgQU6N7Z83b57Fhx9+WDplypSirKws1pAhQ5zCw8Nla9asyenTp4/TyZMnBQxGze+HrKws1ty5c1uePXv2AZPJhFAoVEqlUtm/aevu3bsNW7VqJf/tt98ytdt69uxZERcXF9e1a1eXxs5ZsWKF5fnz55MEAoEGAPLy8ph///03//bt2/EXLlzQ/eWXX0x37tyZoT3+wIEDQqlUKhsyZIgDAKxbt85006ZNmQ2vq1arERERobdixYq89PR0VlhYmIzP59NDhw61l0qlPDc3N/m2bduMr1+/nsBi1f94goODDS0sLKoBoLy8nHracW+zPXv2ZDxtn0QiKT958qRg+PDhZW+yTQRBEE/zwt++XDaTZbfg9EvfKPWnQU+9F5/Pp/l8vvpp+6VSqd7WrVuzAMDKykrVrVu30suXL+sOGDCg/PDhwyldu3YVzZ49O1etVlM+Pj4Oa9asybS0tFS9bFtPnz5tYGxsrJJIJKLRo0cXfvvttwUmJiZPbV9cXJyOQqGgBg8e7MBms+ndu3enmZiYqA0NDVUqlQpFRUUsoVBYrz1sNlujVCpBURQiIiJ4Tk5OCkNDwye6ZZGRkTxLS8tqALCxsam9BovFoplMJn3p0iU9BoOBHj16OJuYmKj27t2bpq+vrzl06JBB7969S/ft22cMAE87Tnu92NhYzpQpU2yqq6sZPXv2LPX3988ZOXKk3fLly3Pi4uI4J06cMAwKCkobP368TVxcHJ/BYNDnz59/wOFwNGPGjLHLz89nC4VC1bFjx1K2bdtm5O/vb5menn63pKSEYW1t7X748OEHADB58mQ7a2vr6tLSUmbfvn0fBQYGZoeGhgq02x9/3rIFCxaYh4aGGqpUKkp7nEgkcnNxcalKTk7mbtmyJa1Hjx6VYrFYFBUVJRs5cqRdYmIiV6lUMr799tucSZMmFffr169s48aNJiToEgTRXDSL4eV/y9raujozM5MNALm5uSxdXV3N0qVLrYOCgoyVSiUVFxfH1R5bWFjI1g4va4einyc/P58tEonkYWFhssOHDwszMjKe+eMkOzubnZyczA0NDU2ePHly/uLFiy24XC7t4OCgcHBwaPPtt9/aTJs2raDuOXPmzMkbNWqU/YgRI4oDAwNNe/ToUT5u3Dib7du3G9Y97v79+1w7O7vquttu3LjBKyoqYonFYnlOTg47Ly+PfeXKlUQvL6/ygIAAEwDYu3ev8KuvvirSnvO047Tmz59vtWfPntSbN2/K7t+/z01KSmIDwK1bt3jbt2832b17dzoA7N+/P/327dvxrq6uVWlpaewNGzaYDB48uCQyMjKhe/fuZb///rshALRs2VJx+vRpwZ49ewzd3d3Ltffx8fEplEqlsrVr19b2UNVqNUaPHl1Yd0QiJCTEKDIyMr7ucbm5uTr79+9PCw0NfbBs2TLLhp9DUFBQyrlz5xIvXbokAIBWrVopEhMTuQ2PIwiCaCpvZdDNzMzUsba2VgLAokWLLLdv3546bdq03A4dOlRcv3494cCBA8Ly8nIKQO3wcmRkpCw6Oppf99mqVCrlNfasV19fX/3hhx+WsdlseHh4lN+/f5+DZzA0NFS7u7tXCAQCzZAhQ8oSEhK4t2/f5t6/f5+XkpISe/DgwaQ5c+ZY1T2nS5cuVWfOnEl2cXGRt23btnLLli0mO3bsyDh58mSLusfRdP15bHl5ecxp06bZ7N27NxUAWrRoofb09CxnsVjo379/aXx8PPfkyZMCT0/Pci6XW3tyY8fVvW5SUhJ37NixDhKJRJSYmMhLS0vTAYC1a9daMBgMcDgcGgAKCwuZbdu2bRUXF8e3s7Orvn//Pnfz5s2mEolEFBQUJHz48CELAAYPHvzo2LFjLcLDw/U8PT0rnvX+PXr0iCkQCOqNJMyePTvH0dGx7Zw5c2qf2drY2CgMDAw09vb2ytLSUmbD64wdO9bew8Oj1ZAhQ0q02yiKetatCYIg3qi3Iuh26tSpYtu2bUZAzTPbq1ev6n/wwQcVV69e5fP5fI1YLJZ37Nixqk2bNpVcLpeeOHFi/k8//WRa9xoMBgP6+vrqysrK2tcskUiqpFKpLDQ0NLnusZ6enuVRUVF8ALh37x7P0dGxXk+zoTZt2igKCgrYKpUKERERfFtb22qapqGvr69mMpkwMzNTNRYkAGDr1q0ms2fPLqisrGQwGAzI5fJ6n4mbm5s8NTVVBwCUSiVGjRpl7+/vn6kdau7WrVuFTCbjAoBUKuXb2dkp7ty5wztz5kyLbt26OScmJvJmzpxp2dhxde/j6Ogo/+OPP5KlUqksNjY2rkePHhWP25cmFosrduzYYQgAQqFQfffu3fu9e/cuCQ4ONhCJRHI/P79cqVQqi4mJiZ83b14+APB4PI1araYa9tIbExUVxW/durW87jYrKyvlJ598UlC3p5uens4pLS1lpKamshsGaaCmp5uamnr3hx9+sASA+/fvc5ycnOQNjyMIgmgqzWJGjUKhoHr27OkcHx/P6969u8uqVauynJycFJs3bzZes2ZN7s8//5w9YcIEmx07dphQFIVffvklncfj0cuWLbP87bff0hpe76uvviry8vISffPNNwXa4WWNRgNHR0d5586dq57XnhkzZhSMHj3aftOmTWa9evUqcXR0VF69epW/YMEC64SEBJ63t7fLxYsXE8+cOSNQqVTUJ598UvL555/nd+7cWcRgMLBv374UV1fXaj09PbVYLBap1Wpq/fr16Q3vc/bsWb3OnTtX8Hg8ety4cYUeHh6iHj161Hv+6OnpWbV06VIdANi1a5fRnTt3dBcsWGANAKtXr87s06dPRbdu3co6deok4vF4mqNHjyabmZmpv//++4cAIBaLRYGBgdkA0PC4uvdZs2ZN1qeffmpXXV1NsVgs+vTp00lATS9+6dKleT179nQaMWJEybhx4+yKiopYTCaT9vPzyxcIBOpx48bZbdu2zZSmaaxevTpLe81du3alUxSFuXPnPjEUrBUUFGRQUFDA9vHxqe2dKhQKas2aNRanT59OunLliq52u7m5efWYMWPsUlNTORs3bnzi/Rw7dqy9RqOhBg0a9AgAzp07pz906NCShscRBEE0Farh8GVdMTExqe7u7gXA6529TDzbkiVLzPr161fapUuX5/5geJdpJ0296PHDhg2zP3bsWAqT2eggQ7MUExNj7O7ubtfU7SAI4vX4N7OXXylgctlvzxdfc7NixYq8pm7D2ygkJCSlqdtAEARR11vxTJcgAODf9HIJgiCaIxJ0CYIgCOINIUGXIAiCIN4QEnQJgiAI4g1pFkH35s2bXG0ifx8fH7uGSerLysoYY8aMsZVIJKKOHTu6bty4UdhETa2nuLiY4eXl5dKpUydRz549nYqLixkAcPLkSUH79u1dO3fu7KLN7NTQhAkTWg4bNsz+edsAYPTo0bbu7u6uV69e5QPAN998YxkXF6fT8DiNRoPhw4fbA8Dp06f12rVr5+ru7u7q7+9vAgATJ05sqc3Opa+v3x4AFi1aZK7dxuPxOuTl5b31M97S09NZ8+fPbzT7WEZGBmvSpEnWb7pNBPEcugDcAPQB0A9AfwADAAwEMAjAYABDUFNredjjP70AtALQAgDJAvOWePElQEq5O9jcl1/Xq5SrwOY2OgO6Xbt2iujo6HgA8PHxsbt27Rq/R48eldr98+bNs+jevXvZwYMH0zQaDc6fP6/30u34D3E4HPrQoUMptra2ynXr1hlv2bLF+Lvvvnu4atUqi8uXLydER0fzli1bZrFv3756a0ozMjJY6enpnLoJHhrbBgC5ublMBoOBEydOJC1btszc0dGxuqKiguHm5vZE0omTJ08KPDw8KgBg3bp15sHBwckODg7VHTp0cJ03b17+rl27MoCaikj+/v7mALB69erc1atX5+bk5LCGDRvmaGZm9tQc028LGxsb1Zo1a3Ib29eyZUtVQUEBu6ioiGFkZERKEBFvCh+ALQB7AHYKldqpUqFuRVFw4LCYliwmxS0oV1Q9LFVoVBoaNE2DBkDTqP930NCu8tTnsmEi4DAM+WwOk0Ex5EpNsVKtyWEwqHh9LusORVFJAB4ASARA1qs3Ey8eRNlcFpYZvPydlpU89V7aFIMAoKOjo7G3t68XUOoWPGAwGBgwYEC5QqGgPvjgA2elUkkJhULV6dOnk1gsFgIDA4UBAQEWBgYGKjs7O0VISEhKREQEb9q0abZqtRpTpkx5OHXq1KLGEuR//vnnLWNjY/kajQaHDh1KcXZ2fmY2JT6fT9va2iqBfwoQlJWVMbhcrsbQ0FDTq1evioULF1o1PO/HH380mzFjxsPff/9d+Kxt2vdGqVRS5eXlDF1dXc3q1avN5s+f3+gSopCQkBYTJ04sBIBWrVpVFRcXM6uqqig+n18vuPzxxx+GI0aMKK677fDhwwYDBw581PCaBw8eNFi3bp25Wq2mvvvuu2wfH59S7XrZ6dOnW7Vq1arqiy++KO7evbtLZWUlo02bNpWHDh1Ke1oBBQAIDg5OPXHihGDkyJEuarU6ys/Pz/L06dMtDAwM1FlZWTqbN29OHT58eJl2O0VRGDhw4KO1a9dm9+vXz7GwsJCdl5fH3rx5cyqLxUJAQICZSqWiqqurqZCQkKRHjx4x582bZxUSEpJiZ2fXxtTUVFlRUcEICgpKdnd3V3h6epafOnVK39fX94nXSxD/ERsAH5TJlQMoiurFYTGMCsurqzKKKzUp+RU6KQUVvMziSmQWVyGzuAoFFQrQNAQvezMemwlTfY6JpQHPxEbIb2dvrDvCxUyvysFYj7ZoweXKlZpsFoP6ny6H9T8AVwGQZYhNpFlkpAKAAwcOGCxbtsza3t5e/iK9LTabTf/vf/9L1NPTo2fOnGl56tQp/Y8++qi0qqqK8e233+b07du3bN68eVYA8N1331kdPHgw2c7OTunh4SH64osvioGatIECgUCzcOFCi0mTJhUHBgZmCQQCzYkTJwSBgYEmGzduzHp2K2qUlJQwdu/ebXLx4sXEgoICpra8HwCo1ep6wz55eXnMgoIClpubm/xZ27QMDQ01bdu2rVy0aJHlN9988zA4OLjF0aNHW8TGxvKWL1+e4+joWFuAODk5mePs7KwAgJEjRz4aPny4EwDMnTs3p+41//zzT4OVK1fW6wmePHnScMuWLfV65Gq1GuvXrzePiIiQqdVqqmfPns4+Pj6lAPDLL78I2Ww2PW3atCKgpjKQRqOBt7e3C/BPAQUnJyflkCFD7LXD7MXFxaySkhLGsWPHDFu3bl2bk3nNmjUZ2kBb9/5r167NoCgKly5dEuTk5LCKi4tZt2/fjq97nFwup8LDwxO3b99uuGHDBhNfX9/aQg/a3Nvbtm0zunHjBt/d3V3h5OSkuHXr1jtRT5hoNlrinyDbGzT0I1MKlX/J8gWRyYVIyi9/paD6PFVKNdIKK5FWWImI5EKg5rtdAAAsBoXWlvp2nR2Ek3q4mIzuaGPIUWk0hQD+FHDZFwBcAfDUEpnEf6vZBN1x48aVjBs3rsTX17fl4cOHDT777LNn9kLKysoY48ePt83NzdUpKChgaYNNUVER09bWtl4PtaSkhCkSibTl8aqzs7NZQE3awNzcXJ0NGzakAcDSpUvNr1y5IlCpVJSzs3O9APjll19ax8TE6M6bNy9HG3iAmueoY8aMsVuxYkWWsbGxmsPh0GVlZbXPyrV1frV++ukns1mzZj183ra6li5d+hAAJk+ebD1z5sz81atXmy9fvjwnICDAdPPmzbU/DGiarg3w8+bNs7569Wq8tbW1qlu3bi6TJk0qEggEmrt373LMzc2r6/4wKCkpYRQVFbFcXV3rvW+5ubmspKQkbpcuXUQAUFhYyNJoNCgqKmLt3r3bRJtuEah5hjx9+nTbrl27lgH/FFAAgNLSUqa2gMLw4cOLN2/eLGQymXTdggyNKSkpYRoYGKi1eautrKxU7du3r2jZsmUbjUZDde/evQwA2rVrVwUAEomk8uLFi/p1r1FYWMj28PAQZWVl6Vy+fFkG1HxmpBAC8Yp0AQwpkyuHUKD6gIL+jeRC1V+yfL2I5EI8eFgOAM2iwpVKQyMmswQxmSXUb1eT9RkUIDIXWHa2F37a3cVkuMTOiA0KZRoN/Zc+j/07gAsAXro0KvFszSLoVlVVUTwejwYAfX19TcPhUIlEUr5161ajr7/+ukij0eB///ufXk5ODsvZ2Vlx6tSplBkzZlhp01lKpVK9WbNmFdQNfAYGBmqZTKZjZ2enTEtL42hr7QYFBaU4ODhUe3h4uPbs2bP8+vXrgqioKNnx48f19+/fb1S3Ddu3b3+iwDwAzJ4929LLy6t86NChZQAgEAg0crmcUVJSwrh9+zbXxcWlXurG1NRUnQULFljL5XJGWloaZ8eOHYaNbZs0aVK94V+ZTKbD4/E0xsbGKrVaDQaDQZeXl9eL6I6OjvLExESOubl5JZPJpIVCoZrL5dIMBoOurq6mAODw4cMthg0bVu8HzR9//GHQp0+fJ575mJubq0QiUdXVq1cTWCwWFAoFxWAwoKurq7lx44Zs6NChDjKZTEckElUPGjSofNCgQfe8vb1dMjIyWI6OjvItW7Zk2NraKlUqFSiKwubNm0169epVPnLkSMcNGzakL1my5Jm9zfv37/PatGkjDw8Pr82/bGBgoN64cWPa1atXa3sNsbGxPAC4efMm38HBoV4hB6FQqLx586YsNDRU8MMPP5jv27cvPTk5mePq6koKIRD/FhNArzK5crIOkzH474xHytA7OYLI5EIkPix/7snNhYYG7ueU4X5OGX4PTxUAgKOJHreLk9BnXGeb/i2N+GqKonbz2MztAO43cXPfOc0i6AYHBxts2LDBDAAcHBzkH330UWl6ejpLW/DA398/Z/LkyS13795tolQqqYkTJ+YPHDiw1N/f3yI6OpovEAjUzs7O8rVr1xpHRkYKhg8f7qhQKKjMzExOeHg4b+XKlVmffPKJg1qtxuTJkx9qnyHXTZBvYmKi1tXVVXt6erq4ubm9UI7j1NRU9pYtW8w7dOhQcerUKcORI0cWzZ8/P3/hwoU5PXr0cOFwOJr9+/enAjUzh3/77beM48ePpwI1QXTevHlWkyZNKtYG2LrbGt5rzZo1ZqtXr842NzdXazQa6qOPPnJsmPR/yJAhJdeuXdPt1q1b5dy5c3N79OjhwmAw0KdPnxKhUKgGgHPnzrU4e/bsg7rnhYSEtFiyZEm9IWgAYDKZmDVrVp63t7eIoijaxcVFvm/fvnQOh6Phcrn0+vXrM7/55hvrDRs2ZE6YMMG2urqa4ejoKG/ZsqXqaQUUAODOnTv3ORwOvWTJkqe+t/PmzbMYNmxYcd3JTnFxcTpJSUncdevW5dQNumw2m+7WrZuzQqFgnDx58kFJSUntDGxtwYuysjLm8uXLMwEgIiJCb9euXU8UTCCIp3CqqlZPo0FPyH4kZx64kaZ3KiabKiivbhY92f9CUn45kvLLqb0RaQJHEz2M9mg5bbRHy68ApAi4rI0MijoM4InvJeLfe+GCB69z9vJ/xc/Pz3LKlCkF2qHkLVu2GNnY2CgHDx5c9rxz3wUajQYjRoywP3HixHuTczg0NFRw4cIFgbaS0vNkZWWxlixZYrFz585m+QyLFDxoNtgAhpRUKecwKarDoZvpzIPSDHZS/tvTo31VTAaFbk7GGOdpW9Hd2ZilUGn+p89jbwJwEcBbv8qhqbx40H0LFBQUMA0MDNRsds3S2NLSUgaLxaL5fP4znx0Sb69/G3SbOxJ0m5yeQqWerdbQfsn5Fczt15IF52JzoVC936vLDHhsDGtvSX/qaVtuZcir5rCYi5kMaieA59bLJup7p4IuQbztSNBtMjoqtWaKUq1ZeSWhgL3+fwl8Wd57MUD2r7lbG+C7QW7lrS31K3Q5rLkAgkB6vi+sWTzTJQiCaCJMAGMqFKp1MZmPdFedvq97L7v0uSe9z2IySzBqW4Sep4OR3uLBbltsjXRX6nFZfgCOAyCjis9Bgi5BEO8jCsCgcrlyQ1pRpemKU3F6N1KKnnsS8Y/I5CIMCryu94HIRG/JYLc9JgLOKgGX/Q1qlhyR4PsUJOgSBPG+6VYqV258VKl0WnHqnu7F+09dIk+8gL9k+biScEVvQBtz1+8GugULeCyZPpc9C8D1pm5bc9QsCh5cunRJV1vw4IsvvmjZcL9KpcKkSZOsJRKJSCwWi4KDg2sTIISEhAg8PDxEnTp1En344YeOubm5zJEjR9q1a9fOVSKRiLp06eL8Im1YuXKlabt27Vzbt2/vevHiRd26+4KCggzc3d1d27dv77p06VIzoGZ5j1AodH/aPfz9/U3qFhsIDg7W37x5s1HD44CaxBvp6emsY8eO6YvFYlH79u1dZ8yYYQXUTBSytLRsK5FIRB999JFd3fOWLVtmJhaLRS/y+t4Gvr6+T3z2QM067o8//tjuDTeHePdYl8qVl/LLFOeWnbzX7oOfL5OA+x+haeDM3Vx0//my7opTcR0KyhXny+TK4wCMm7ptzc0L93QVKoU7h8V56Z6xQqVQcVicRpcMOTk5KcLCwmR8Pp8eOnSovVQq5Ukkktq1suvWrTMxMTFRSaVSWWlpKaNXr17O3t7elQDwww8/WF64cCHR0NBQc+fOHY5CoWAANYkv2rRpo2jsfo0JCgoyjo2NjUtPT2dPnjzZpk+fPrXrSj08PKqioqLiGQwGOnfuLJo5c2YBAHTt2rU0JCSk0eU5f/75p35MTEx83759HefNm5cfFBRkdOTIkdSGxxUVFTHy8/PZNjY2KjMzs7IRI0aUAkDnzp1dtJmzfHx8ChvOzq2qqqLu3LnDe9HX9zbYs2dPo8t4eDwebWhoqLp79y6nbdu2L/yZEoSWRkOPlqvU23deS+Fu/SuJXa1+v2cjvy5qDY2jUZlU6J1s/rz+rgM+8Wj5gK/DmgzgD5AhZwD/IuhyWBxW2z1tX/pGd33vPvVeNjY2tSnHtIUD6u4PCQlpceHChQdATcaqcePGFQYHBxvQNI0xY8YUGhoaaoCaakUv2z47Ozt5VVUVVVhYyDQyMqqXAq1u4QMmk0kzGAwaACIiIgRisVg0dOjQYm2qxrrHKZVKMJlMev/+/S0GDRpUwmQ+WTUvNDRU393dvRL4p/CDSqWCqamp0tDQUA0Ax48fNwoLCxNMnjw5f8qUKUUA8MsvvxhPmDChcPny5ZYNr3np0iXdBQsWWKtUKsrX1zd/1qxZhRKJRBQWFibbsGGD8cOHD9n+/v45ffv2dczLy2MLhULVhQsXHjx8+JA1fvx4u4qKCoazs7N8/1vB91UAACAASURBVP796X5+fpbXr18XSKVSWWxsLKd9+/Zt7t27d/f8+fOCgIAAC1NTU2V+fj5rxowZeX5+fgV1C044OTnJg4ODU8ePH28TGxvLLykpYc6YMSNv0KBBpZ999pmdnp6euqCggH348OFkV1fXam0hBYlEIqJpGmVlZcxffvklvX///uW9e/cuCw4ObtG2bVuSqJ34N1qUy1W7SqqUfb/aH6V7N4sU23kT5EoNVpyK45z6O5uzcWyHXQY89ucCLvtzAI1W/3qfNIvhZa0bN27wioqKWGKxuF6KPoVCwdDT06sNxDY2NtU5OTnsnJwctqWlpfLJK/17PXv2LHNxcWkzcOBAFz8/v0bHnI4cOaJvb2+vMDQ01NjY2CgfPHgQGxkZKbt8+bL+jRs36vU6J02aVDB06FCHyZMnFxw/fryFUChUjR071ubkyZP1kp4nJCRw7O3ta38srF271tjBwaGNoaGhmsfj0d26dat48ODBvcuXLydu3brVNDs7m6VQKKhr164JtKknG1qyZInluXPnEm/duhV/+PBhoVwupwDg2LFjBjdv3tT19/fPAYALFy4kxcTExDMYDLq0tJSxdOlS80WLFuXcuHEjQSAQqLXD7CwWi46Li9PZtWuX0MPDo/aeM2fOzJVKpbLZs2fX/o+kVqup2bNn5xw5ciRZuy08PFxw+/bt+LrH5efns8+dO5e0fv36jBUrVjxR+zY8PFy2adOmtKtXr+oBgIuLi0Imk70zGYCIN6JXZbXqQUhM1oDeAX+RgNsEojMeofe6K7oHItN7VVWrZaipA/xeazZBNy8vjzlt2jSbvXv3pjbcx+Fw6PLy8toM9enp6ToWFhZKCwsLZVZWVqNF4p9m8eLFZhKJRBQYGFhbQq+oqIixb98+4+Tk5NiIiIj72upEdcXFxemsW7fOfNu2bRlAzZCnvr6+hs1mY8CAASXR0dH1gu6IESNKz549m1xQUMAcOXJk8Y4dO4yDgoLSd+zYUe8ZR90iBQAwZ86cguTk5Njs7Gx2WFgYz8DAQMPhcGh9fX2Nl5dX2b179zhbt241GjNmzFOnWsbHx/P79+/v7OnpKcrPz2fn5OSwAODnn382Z7PZtT9eYmJiOCKRyA2oyRmdkJDAXbhwobVEIhFdu3ZNPzMzkw0Ao0ePLty5c6cwOzubbWlp+czF8CUlJUx9ff16Y3cTJkzIt7CwaLd+/fra4Orq6lrFZrPh5eVVmZqa+kQw9fb2Fo0dO9axf//+pY/fJ1KkgHhR3AqFamNxZfWpr/ffFn53PJYrV5Lh5KaiUGnw07l4nfE7b+jnlymCyhWqnaipL/xeahZBV6lUYtSoUfb+/v6ZdYeatYYMGVK8atUqM6Amy9T+/fuFI0aMKBk5cmTJoUOHhMXFxQwAiI2N5aSlpT0zCK9cuTJPKpXKZs6cWajdxmQywePxNFwulzYyMlJXVlbWGwcuLi5mfPbZZ/a7d+9O1QYU7T0BIDw8XM/FxeWJoW2VSoVz584ZjB07tkQulzMAoKqqqt57LhKJ5CkpKZzH+ygAYLFY4PP5Gj6fTxcVFTG014qOjtZ1dnaulslk3G3btpl069bNOTExkbdq1SrTutd0c3OrvHjxYuLjIeE4e3t7JQCEhoY+AIA///xTFwDc3d0VMpkszsjISHXt2jVdJycnRUBAQMbj8+6PGzfuEQDY2tpWR0dH63br1u252QKio6P52qo/WoaGhur58+dn1+3pymQynkqlQmRkJM/Ozu6J9y48PFx27dq1+CVLllgCQGJiIsfFxYUUKSCex71coYqLSC6c2HPtX/wrCflN3R7isai0YvRa+xf/akL+mHKFKg5Au6ZuU1NoFkuGdu3aZXTnzh3dBQsWWAPA6tWrM/v06VPh6+vbcs+ePRlz5szJnzJlSksPDw+RRqPBvHnzcrWVgr777rvsvn37OtM0jRYtWqi0BQb+DQMDA02vXr1K2rdv76pWq6lFixZlA/8UKfD39zfNzMzkTJgwwR4A9u7dmxITE8NbsWKFpY6ODt25c+fyXr16VTS87q+//iocP358EQD07t27tG3btq3Gjx9fL8PXoEGDyr744gtDANi4caPx0aNHjdRqNbp27VrWoUMHeUBAgPHu3btNGAwG7ePjU2RnZ6fcunVrbTk/sVgs+u677+oNhy9fvjyrX79+ThqNBi1atFCfP38+CQCEQqF6/fr1WcOHD3fo1KlT0sCBAx2VSiWjRYsWKk9Pz0pXV1eFr6+vbVlZGZPBYGDXrl217+X58+cfMJlMXLp0qV7pvLrWr19vbGNjUy0Wi+UymUwHqEnNeezYMcM///zzwaZNm2pHF4RCoapv375OhYWFrKCgoCcmo3l7e4vKy8uZU6ZMyQOAP//8UzB9+nTyDUo8VbVKM0Op1qxZEhLLPXY7iwyLNENlChWmHrjN+6iDlc2qj9pEcFlMXwaDOtrU7XqTXjgN5Oucvfy++/LLL60XL16c21gv/12krab0tJnfDcnlcmr8+PG2R48eTX3NTWtyJA3kS2FUVqvWFVcoJ3+yPYKfUfRCRcKIJuZmoY8DX3au4uswF3NYzHVN3Z435d/MXn6lgMlhcV7l9Hfa02r1EjW4XC79PgRc4qXolCtUQWmFFf3H7bjBf1T5n8yrJN6AuJxSDNpwjXdoitdyoa6OvS6HNRPAO//wvVk80yXeLyKRqPpFe7kE8QyCcrnqUlRa8YCRW8N1ScB9+2SXyDF44zXdxIflE8rlqhAA7/wKBRJ0CYJ4G5mXK1TSM7E54om/3+ST2clvr9IqFUZvi9ANSyroXS5XhQFoNHPfu4IEXYIg3jbOFQpV9I5ryY7zjt7hqjUk0dHbTqHS4Kv9UbzDtzLaVChU0QDsmrpNrwsJugRBvE0kVdWqmytD40x/uZj4r9boE80bTQMrQ+N01l2QWVVWq24D6NDUbXodmlXQfVoCf4lEIlIqa57XfPrppzYrV640feLk1+jo0aP6hw4dMgCATZs2CePi4nRmzpz5RPrFV1W3UMOvv/76xBCLr69vS4lEIvr444/tVKp/Jjpfv36dT1GUWPseaR08eNDA3d3dddasWZYAcPv2be6CBQueyP4EAEuWLDELCwvjPa34hEAgaC+RSEQSiUSUl5fHBAA7O7s22m1RUVHvxLOYRYsWmaekpDT6ZT516lSrpKQk8kXfdHpXVqsuTz8YbXDoZkaz+u4i/ju7wlKZ3x6JaVFVrb6Kd3At7wvPXtYoFO4MzssvGdIoFCoG5+kzoF8kgX9AQIBxVVUVY/HixW+0NIiPj09pnb8/+vjjjx369u37WnLKPa1Qw5UrV/hKpZIhlUplS5cuNTt06FCL8ePHPwKAwMBAEzc3t8qG5xw4cEAolUplQ4YMcQCAdevWmW7atOmJmdJqtRoRERF6K1asyEtPT2c1VnzCxcWlSiqVyuqeJxQKlQ23ve1Wr1791Nywn3/+eeGGDRtMGhafIN4Ij8pqVciE3Tf5UlL39p13NjaXYjJidP192l3m67A6AEhv6jb9V1741yKDw2Hdd22Fl/3zvICtTeD/tP0XLlzQ27Vrl8m+ffvStNsmTJjQslOnTqKePXs6FRYWMmUymQ6LxRLfu3ePAwAeHh4iPz8/S6AmYIvFYpFYLBZdv36dDwAikchtyJAh9q1bt2515coVPlCTbEJ7fe3fHyfxN5bJZDpTpkyxuXbtWmJCQgJ32LBh9k87Z8OGDUKJRCJq06ZNq2PHjj01oURdFEXh008/te/Vq5dTQkKCTt19CQkJnLZt21YCQMeOHSvDw8N1AeDWrVtca2trpa6u7hMzSdhstkapVIKiKERERPCcnJwU2uIQdUVGRvK06R1tbGxUfD6fBuoXn0hKSuKJxWLR1KlTrTSamks8evSI1alTJ9HYsWNtKysr6yUjKC8vp4YMGWLv6enpMmjQIAeFQkFp38fExESdbt26ORcXFzN+/vlnY4lEIhKJRG5Xr17lA8CcOXMsJBKJyNPT00Umk+k863Ot2wOv+zlotwsEgvYAcOzYMX1XV1e3usdJJBLR5MmTrd3d3V3Xrl1rDNSMNsTGxnICAwOFdnZ2bdq1a+c6efJk68f3ld+6dUvvRT5L4j8lqqpWX5x5MFqXBNz3R+idHCrgQoJBhUJ1Fe/Q5KpmMUTzvAT+ADBnzhwbJpNJl5aWMoCanl9lZSXj1q1bso8//rgoICDABADEYnHZrl27jO7du8fR0dHRAEBOTg4rNDS0xc2bN2Vnzpx5sGzZMgsAyM3N1dm/f39aaGjog2XLlr3wcHFiYqJORkaGzrOOmThxYrFUKpVduXIlYd26dWYvct3NmzdnREdHx8+fPz935syZ1nX3tW7dWn7t2jUBAFy8eFH/0aNH2nzKZnPnzm205z9nzpy8UaNG2Y8YMaI4MDDQtEePHuXjxo2z2b59u2Hd4+7fv8+1s7Orl1O5YfEJmUx29+bNm7JHjx6xDh48aAAA4eHh8bdu3ZLZ2NgotO+/1oYNG0wGDx5cEhkZmdC9e/ey33//3RAASktLmZ9//rnt3r17Uw0NDTVz584tkEqlsrlz5+ZEREToSqVSXnZ2to5UKpVt2bIlffny5RZA458rAGh74HV73CqVCq6urlVSqVTWqlWrKgA4duxYC39//4yGPXMfH5/iqKio+AMHDhhrC0NozZw5Mzc6Ojo+OTm5dpF5dXU1yXT0ZllVVquuLTkZq0dq375/dlxPYR65lWFeLlddwDuynKhZBN3nJfAHgC1btqTOnz8/Z9GiRZYAIJPJOB06dKgEAC8vr8qkpCQOAJibmyuzs7N1du7cKdReMz4+Xic+Pp7n6ekpGjJkiFNpaSkLAGxsbBQGBgYae3t7ZWlp6ZN1955i5cqV5nPnzq0dhiwsLGRre1v5+flsADh+/Li+RCIR9e/f3yknJ6degN68ebORRCIRLVq0qN7zVTMzMzUA9OvXr1x7HS1vb+8qNze3qs6dO7uUlpYyTE1NlXfv3uXo6+urLSwsGs1k1aVLl6ozZ84ku7i4yNu2bVu5ZcsWkx07dmScPHmyRd3jGmYla6z4hJmZmZrBYGD48OHFd+/e5dVt76hRox7du3ev3qOB+/fvczdv3mwqkUhEQUFBwocPH7IAYMuWLWZyuZwhEAhqA+eIESPslixZYt2nT5+yO3fucMPDwwUSiUT01Vdf2ZaVlTGAxj/XpyksLGTq6emp6277+uuv87/55hubhnMGvLy8qlgsFqytrRVZWVn1RmMCAwPNHR0d2zg5OZEavk3DoEKhurbxzweGf9zKbBbfVcSbtyI0jhOWVOBWLlcdBfDC39PNVbP4h/y8BP4A0LVr18qxY8eWJCQkcO/du8dxdnZW3L59WxcAIiIi+A4ODrVfjF26dCn/+++/+drem0gkqm7Xrl2Ftkd07dq1BABIT0/nlJaWMlJTU9kCgUDd8J6NyczM1GEymbSTk1Ntz1D7bFMqlcpMTEyUAPDzzz9bXLp0KeHs2bMPGIz6b/O0adOKpFKprOHzQ21xg5iYGI6+vv4T7Vm7dm3OjRs3EoRCoXro0KElt2/f5j0uROAsk8l448ePt22szVu3bjWZPXt2QWVlJYPBYEBbfEHLzc1NnpqaqgM0XnyitLSUoZ24df36dT1HR0eFXC6ntAUa/vrrL7267//j91zu5+eXK5VKZTExMfHz5s3LB4Cvv/4678cff8z08/OrreR07Nix1F27dqUEBASYtm7dWt6jR48S7ftZNxNVw8/1aSIiIvgikahecQRHR8dqBwcHeVRUVL2e7o0bN3gqlQqZmZkcKyurej9eZs6cmZuamhobHR3N1wZkHR0dsj7lzWCVy5WnQv7Ostx6JalZ5IgnmgZNAzOConkJeWU9KxSqLQDe6tGmZvGPubEE/unp6azNmzcbr1mzpl5g+umnnzLnzp1rdebMmeRdu3ZpxGKxSE9PTx0cHJxSWFjIBICpU6cWTpkypfD8+fMCALC0tFQNGDCgpFOnTiImk0l379697Oeff84xNzevHjNmjF1qaipn48aN6QBQUlLC0j6rTUlJ4e7Zs6derzA2Nlb3yJEjySqV6pkffL9+/R55enq6dujQoaKxANoYHx8fh5KSEiYA/Prrr2lAzfPkDh06VHl7e1d6eXmJmEwm3aNHj1JtgQVfX99HQM3zyf3796c1vObZs2f1OnfuXMHj8ehx48YVenh4iHr06FFvGN/T07Nq6dKlOkDjxSf4fL5m0qRJdjweT2Nra6sICAjIzsnJYfXr18+Zz+drDAwMVEePHq2XYWr27Nn548aNs9u2bZspTdNYvXp1FgAYGBioP/zww4rdu3cLL1++zD979qzBlStXBNXV1Yyff/45w8vLq+rEiRNKiUQioiiK/vjjj4uGDBlS2tjn2hiZTKazceNG071799Z7L7799lurlStXPjEB6tChQ4Z+fn4tx48fX8jlcusF1MDAQPP9+/cbGxkZKc3NzVVSqZTXsWPHJwpbEP+9CoVq873sUvHikHskfyyBarUGE3ZL+SdndB3HZFApXDbzp6Zu08t64YIHr3v2clMQi8Wihj2fupKSktibN282CQgIeOdnqy5ZssSsX79+pV26dHlvssVLJBJRWFiYjM1+sVVAU6dOtfLz83vo5OT02vINkoIHgFKtmZZXKvcfuOEav1T+XtQAIV6QuT4Xp2d2rTTgs6ewGIz9Td2el/HCQfRVAyaD8/b9YLWyslJ9/fXXBc8/8u23YsWKvKZuQ3O3ZcuWrOcfRbyiD+RK9c9jt9/gkYBLNJRbKseY7ZH8E9O6/MbSYUQDuNfUbfq3XrinSxDE6/ee93QNKqtVD6buv238VxMXn9fIy1F4NhDVBTXLQ40HzoKqrBAl14OgLMyA+WcB4Fg4v/C5HKtWqM5LRuH5zaDV1aAYTBh9+DU4liJUyMJQcu0AGDw9mIz4HkyePpTFOXh0dS9Mhs1/Y6/5bTJG0lLz3UC3B3pcVlsAz5zj0dw0i2e6BEEQ5XLV1jOxOXpNHXABoOjP38B1EMPko0Wg1UrQSgUYHD2YfLQIhec3/etzAaD4r91o0WUMeI6dUJV0E8V/7Yb52J9QJj0O80/XouL+VVTEXYG+eAgeXduHFt3Gv4mX+lY6KM1gDGxrYS22NfyBr8Oa19Tt+TeaxexlgiDee0Mqlaphy0/ea/K1mBpFJeQZ96DXri8AgGKyweDqgW3cEmyh9UudW7u/urL2OKaesGYjxagJzioFKAYT8oxYMHUNwTayeuL6xD/8DsfwVRp6OgDvpm7Lv0GCLkEQTc24slq1d3pQNL+i+oUm+r9Wqke5YPL1UXjmF2TvnonCs4HQVMuff+JzzjXqPRnFl3cjc8sEFF/eCcMevgAAgy5j8PDIEshT/4auWw+UhB+GQZcxr+31vSvyyxWY+8cdXoVCdRTAW5MprlkE3aNHj+prk0uYmJi027dvX71lOiqVCpMmTbLWpvALDg7WB2qSTIwdO9YWqFkqol3qc/fuXY5YLBaVlZUxJBKJyMPDQ6S9/n/R3kWLFpl36tRJJBaLRRs3bhRqt4eFhfG8vb1dPDw8RN26dXO+d+8ex8/Pz1IkErk1LBbwLPPnzzc3NTVt11hRhbKyMsYHH3zgJJFIRL1793bUrpXdtGmT0MvLy0UikYgaJuyfNWuWpbu7u6s2k9TatWuNz5071+g/0o8//tiusrKSioqK4nbs2NG1Y8eOtQUTiouLGb169XLq2LGj66ZNm4SNnf+2CQ8P561fv964sX0RERG8xYsXv1A2MeLllcmVuw7dzOA1lxSPtEaN6twkCDoMhOXngaDYHJRG/vHK55b9fQaGvSfBeurvMOz1JQrPbgAA8Ow7wGLCBpj6LEVlQiR4jp2gKsxE/vHVNUFb+WIB/310/l4uLt7PMyxXqDY3dVte1AsHXZVS7Q5A/LJ/Hp/fKB8fn1JtMgQLC4vqoUOHltbdv27dOhMTExOVVCqVXb58OfHHH3+0yMnJYU2bNq2IwWDQdQNAeXk5NWHCBLvdu3enarMehYeHyxqmCnxZR44c0U9NTeXcunVLFhkZKTt+/LhhVFQUVy6XUzNmzLA9dOhQ8s2bN2UHDhxI0SaUWLNmTYb2/tosTs8yffr0gp07d6Y0tu/YsWP6nTp1qpBKpbJOnTpVBAcHG6SkpLCvXr2qFxERkSCVSmX29vb1lrTEx8fzpFKp7MCBA8Kqqirq1q1buv379y9veO27d+9yDA0NVXw+n964caPJqlWrMm/fvh1/8+ZN3YKCAuaGDRtMRo0aVSSVSuN///33J9Imvo28vb2rZs+e3ehkQS8vr6rIyEg9tbrpe1/vKo2G/uRRpbLXmrPxzWZ5A0tgDKbAGBzLmt/ofFEXVOclvfK55Xf/BN+lZiSU79oVipyEeudqlHKUx/4JQYdBKL66B8KB30DH3AkV9/76j17Zu+n747FcuVLtA2BQU7flRbzwRCoWm8na/NWll77RtF97PfdecXFxOsbGxioDA4N6SflDQkJaXLhw4QEA6Ovra8aNG1cYHBxsMH369MIdO3ZkSCQSkZGRkQoAfH19bSdOnJjfrl2756buk0qlvOnTp9tkZWXp6OnpqQ0MDNRSqVQ2cOBAh/z8fLaOjo7m1KlTSUZGRrXtOXLkiNHSpUtzAIDNZmPWrFl5Bw8eNOzZs2dZ165dS7VZnGxsbFTav/9bLVu2VMXENL5Cy8XFRXHr1i1doKbggImJiSokJERfrVZTXl5eLi4uLlU7d+7MYLH+ebspiqIVCgWlo6OjWbduncmUKVManaly7NgxA29v73IAaN26dVVxcTFT+8OBy+Vqbty4obtjx450FouF1q1bV965c4crkUhq1/UGBAQYHzhwQAgAGzZsyOjatWuldi30yJEj7Xx9fQtbtWqlGD9+vF1JSQlr8ODBxWvXrs25dOmS7oIFC6xVKhXl6+ubP2vWrEKJRCLq2rVrWUBAQHZAQIDx+vXrzTMyMmJHjhxpl5iYyOVyuXRycjL3r7/+im/Tpo1Cu72srIw5e/bs3AkTJhR/+OGHTiqVikpLS+P+9ddf8ZcuXdI7d+6cQWlpKZPD4WhOnz6dfPHiRb0LFy4IAgMDswUCQXtXV9cqpVJJnT179oGFhYXK0dFRERERwe/atesTVZyIV2ahUGl+mxZ0W1eheqIOR5Nh6hmCpW8MZWEm2EJryNNiwDa2eeVzmXpGUGTcBdemXc12w/oDWaU3gqHfaSgoJgu08vGEXIoBWkWykD5LmUKFaQdu83d/7rGPr8NyAdCsV9w0i+FlrUOHDhkOGzasuOF2hULB0NPTq13bZGNjU52Tk8MGaoZb2Ww2/d1331lfvXpVPy0tjSOTyepNxvD29hZJJBLRxIkTW2or5ACARCKpkkqlstGjRxdqe6MAcPjw4dSbN2/KRo4cWbx79+561S3y8vLYtra2tT1JOzu76pycHHZWVhbbwsLitSVN0GrTpo0iKipK18nJqfXff//N79OnT3leXh67urqaioiISODz+ZoDBw7UG54fOnToo08++cRu6tSp+XFxcdwHDx5wxo4dayOVSuvlS05MTOQ6OjpWA0D//v3LFi1a1NLBwaGNRCKp0NPTo0tKSpiGhoZqoCazlDYDGPD0ohJAzfB29+7dy4YOHVrm7OxcfePGjYTo6Oj7N27c0AOAJUuWWJ47dy7x1q1b8YcPHxZqe9CZmZlsjUaDyMhI3brvbVBQUIpUKpX16NGjtryiWq2mjhw5kjx79uxcAIiKiuJaWlpWR0VF1TtOX19fff369URPT8+KvXv31nufXFxcqm7evCkbMGDAo7i4OA4AODg4KGJjY5t8cs87iCqTK4N2Xk/m3sl8LVUyX4lRn69QELoW2bumo/phCvS9RqEyIRyZm32hyI7Hw6PLkXd4MQBAVVaIvD+WPvNcABAOmIHiSzuRvWs6Hl3dC6P+M2rPUZUVojr3AfjOngAAfclHyN03BxWxf0LX7YM398LfUjdSinBQmqFbJlfuaeq2PE+zCrpnz55tMXr06EcNt3M4HLq8vLx2KDM9PV1H+yW8ePFii0WLFuWsWLEiy9zcXBkWFpZw+/Zt3fT09NqunnZ4mcFg0CEhIc8ss6dSqTB16lTrTp06iX777TfT7Ozses9HzczMlGlpabXbUlNTdSwsLJRWVlbK7OzsZ1YeauhphQ+eZcuWLcLevXuXPHjw4F6/fv1Ktm7dKjQwMFB37969DAD69OlTFhcXVy9ITJ06tejMmTPJV65c0ZsxY0b+yZMnW+zcuTPD39+/3vPKumu2FyxYYLVnz57klJSU2Hv37vFkMpmOgYGBuri4mAnUVAsyMjKqHXd9WlGJ+Ph4/sWLFw14PF7tr53t27cbOjk5tfnggw9Ktcf079/f2dPTU5Sfn8/OyclhAUDXrl3LN27cKLS3t3/uT/3S0lJm3RESLy+vKqVSyXBycmp95coVA+329u3b15ZHfPDgQb33KSEhgScWi0V//PGHsF27dnLte0JRb/0oenM0+mGZwmPDn4kvlg7sDdMxc4CF7y+wnLgJpiO+B5OrB76LN6yn7YHtnBNoOWM/zEavBACwBEKYfbz8mecCANe6NSwmbIDlxE2w+CwAHHOn2nNYAiFMff4J3LquXWE5aQvMx/8MJr/2ny/xDP7n4nWqlOoeAHo2dVuepdkE3fT0dBabzdaYm5s/8QBtyJAhxatWrTIDapLv79+/XzhixIiSlJQUdnx8PHfkyJGlEomk0sXFpYrD4dCLFy+urUZUl6Ghobph3deGIiIiaksGfvnllw9pmq53vI+PT/GPP/5oDtQUB9iwYYPZ6NGji3v37l0RFhYm0Ab7rKwsVkxMzDOfUz2t8MGz0DQNbbAzNjZWlZSUMLt3715+584dPgDcvn2bb29v/8Ri8eLiYkZqairHy8urS/a1cQAAIABJREFUSi6XMxgMBqqqqup9/i4uLvKkpCQd7X2MjY3VTCYT+vr66pKSEmbnzp0rQkND9VUqFWJjY/nawAQ8vaiElZWVIioqKn7Pnj3G2p7xl19+WZyenn738OHDQgBwc3OrvHjxYqJUKpXFxsbGaZ9Jjx8/vnjVqlVWX3zxxTNn2CiVShQUFLDqVlvicrm0QCBQHzlyJKluT1f7PkVHR/McHR3rBXMXF5eqqKgo2cyZM3MDAwONASAlJYXTunVrMpPlv8UqV6jWLT4Rq6tUk/oRxH9DodJgxak43TK5ciuaUWxrqNk07PDhwy0GDx5cr5fr6+vbEgDmzJmTn5eXx/bw8BD17NnTed68ebmWlpaqhQsXWi5fvvyJvMj9+vUrz8vLY8fGxnKAf4aXo6Oj+SNGjChteHxd7dq1k6empnK6devmfPPmTd2G+z/55JMSGxsbhVgsFnl6eoqGDRtW7OHhIedyufTGjRvTPvnkEwcPDw/R6NGj7dls9kt9o6xfv954/vz51sHBwcJPP/3UBqiZMZ2SksKeNGlSUXBwsKFEIhEdOnTIaNKkSYXe3t5VPB5PI5FIRFFRUfwJEyY8MUT/008/mfr5+T0EAFdXV7mHh4do7NixhXWP+eijj0rCw8P1AGDhwoW548ePtxeLxSIdHR1aIpFUzZw5M//QoUNGHh4erp999lkBj8erfX11i0p07tzZ5fvvvzcHAIFAoOZyufSSJUty/Pz8LC9fvsx/PMPbdeDAgY8AYPny5Vn9+vVz6ty5s8vQoUMdtNcUCASajIyMuy4uLs/MOPPFF1/YzJ49u14aywsXLugaGBioO3bsWC9gFhUVMbt06eIcHh4u+Oyzz+q9TwkJCTyJRCL69ddfTfv161cGAImJiRwvLy/yPPc/pKHpzxLzyvTDkwqffzBB/Auhd3KQ9ajKCkCzXXP1wmkgVUq1O4vNfOkMViqlWsViM5tVwQPiSaNGjbL9/fff0/l8/jvXBQkMDBSqVCrKz8/vhSZaRERE8EJDQw1WrVr1wiMRr+o9SAPJqVCoMj7bJTWJSnvityFBvDIPO0Ps+VxSwOewWgJodqNU/2b28isFTBb7ra89/F44cuTIE+UB31deXl5VXl5e703VpTdBpdZMisl8xCcBl3hdbqYWQ5paxPN2NJ6pw2L4N3V7GiIFDwiiGXnHe7r8ympV5qhtEYaxWc98ykMQr8TFTA8h07qW8nSYlgCaVQ3sZvNMlyCId1u1SjPtRnKRDgm4xOuWkFeO6w8KmNUqzbSmbktDJOgSBPEmCNQaevGPZ+8/MTmRIF6Hn8/H66o19HcAmtW/ORJ0CYJ47eRKtd9l2UNWQt4T2UcJ4rVorr3dZhN0n5awf+XKlaYSiURkZWXVVls44PTp0822osTWrVuNOnTo4NqzZ0+noqKieu9vYmKijqenp0unTp1EO3bsMARqiiS4uLi4WVlZtW3seqNHj7Z1d3d3vXr1Kh8AvvnmG8u4uLgnknBoNBoMHz7cHgAWLlxo3qlTJ1GbNm1aabMuhYeH8zp27OgqFotFTyt28LYhxQreGgY0jbn+5+J5zz+UIP476y7IdFUazQI0o9rxL17woLr61Qoe1JzfqGcl7F+8ePHDhqkaBw0a1Cx/LisUCmrHjh0mUqk0fuzYsYXr1683qbt/xYoV5j/88ENWZGSkbPfu3cZKpRJubm6KW7duxZubmz+xFjU3N5fJYDBw4sSJpJ07dwqzsrJYFRUVDDc3tyeOPXnypMDDw6MCAJYtW5Z369YtWVhYmGzdunXmAPD9999b/Z+9+w6Pour+AP6dsjtb0zshpJBsChBCSOgEBOmhFylKFwsCKoKdovJaUOwoKMiPIiIgoYmCClICCUUECSEJKZAE0sv23Zn5/REChBogyabcz/P4vL7slLNLzNl7595zfv7557T9+/envPfee563nt8QkWYFDcbooxcLkVFItjsTdev8lXJcKjKwAPrYOpZK1d8yJJWyH48Z9NA3evmnnXe91/0K9t/JG2+84fHrr786SKVSYe3atRmBgYHmWwvWm0wmasyYMX5Wq5UKCQkxrFu3Lmvnzp3qTz75xN1qtVJms5mKi4tLKykpYebNm9csLi4uPTk5WVr575MnT25+9uxZhSAI2LhxY3pgYOA9izScOXOGCwkJMUgkEsTGxpZNnDjRF8D1og2ZmZlcdHS0gWVZuLq6Ws+ePSuLiIi46z4yjuNEi8VCabVaWqlUCkuWLHGfP3/+1TsdGxcX5zBlypTCyvMAQKfT0cHBwQYAKC0tZQICAiwAoNfraa1WS91cz5o0KyBqS6nBMnP9scx69VyNaDr+Lz5DNa9f8Ex7uWS3rWMB6sn08v0K9t8qKyuLPXDggPrkyZPnFy1alLNw4UIP4PaC9R4eHtZDhw5dOHHiRHJ5eTlz5swZDgCMRiP1999/p0ydOjX/s88+c73bfT7//PPsxMTE5Lfeeivn888/v+txlYqKihg7OzseAJydnfnS0tIqm5Nbtmxp/PXXX9Xl5eX0qVOnlDc3DLgTR0dHoXXr1vrXX3/da8SIESWiKGLz5s0OEyZM8ElLS6tSs/bixYtcYGDg9bKGEyZM8AkPDw/t2bNn+bV4rImJibKcnBw2JSVFXlhYeP1bDWlWQNSiQJpC4P7kOza2Iohat+N0DsWx9GMA6kUP8HqRdO9XsP9WKSkpXFhYmAEAunXrpktPT+eA2wvWX716le3fv39AdHS05vjx46qsrCwJALRp08YAANHR0fq0tDQOAA4dOmQXHR2tGT169PUyhAsWLPCIjIzULFiwoFllV6NK06dP946OjtZs3rz5egMFJycnvqysjAGqJuBKixYtyl25cqVLbGysf0BAgNHLy+u+XYkWLFiQt2PHjvSffvrJ8bnnniv4999/FQsWLLjyySefuN183K01otetW5d1/vz5/5YuXeoJAEuXLr08Z86c5pMnT26h0WgMHh4e1+sUk2YFRG0xWvgpW05kM1ah0RU4IxqIMqMVf57P43lBHG/rWIB6knSrU7D/ZoGBgaazZ8/KAeDgwYNKX19fE3B7wfpVq1Y5DR48uCQhISE5MjJSW5mYKs9NTExU+Pv7mwCga9euZQkJCcmbNm26CFQ8Tz106JD6xIkTyQsXLsy5tYjIypUrLyckJCSPHDny+qbD1q1bm86fPy+3Wq3YsWOHXXR0dJVnz82bN7fu27cvbefOnWlSqVQIDg6+5/uslJycLJXL5YKLi4uV53nQNC1qtdoqf3cBAQHGlJQUDgAMBgMFAEqlUlCpVDwAtGnTxnT48OGUH374IbNZs2bmyilogDQrIGoNzQvitE3HLz1Q9y2CqGnrj2UqdCbrLFvHAdSTFV2dO3c2/PDDD8K1ZvSWt95662pWVhb71VdfuXzwwQe31b318fGxdu/evTwiIiJYIpGI69atSwduFKzXarX0ihUrMimKEidPnuy3ffv2KlOREolE7NatW6DJZKK3b9+eeus0MAC4urrySqWS79ixY1BoaGi1SgFyHCdOnjw5PyoqKtje3t66ZcuWdKCiccOaNWsubdy40X7ZsmXuDMPggw8+uEzTNFJTUyVPPfWU34ULF+SdO3cOWr16dYZGo6mSjD/44AP3JUuW5Hh4ePCCIFDDhg0L+OKLL7JuPiY2Nrb04MGDym7duumnTp3aPDU1VW6xWKjKKdtly5a5bNy40UkulwvffPNNlXNvblbAMIzYvXv38o8++ij31mYFkyZNKnzjjTe8jUYjfWuzAkEQ4ODgwP/2229pwI1mBTcn9zt5mGYFHMeJu3btSvvjjz+ur8K+9e8eIM0K6oEe+eUm7lwuKYZB2NaRtEJYeMETQFsA/9gyluo3PDCbw1mp9OEbHpjNVlYqtXnDg507d6orF9/YOpaaJAgChg8f7rdt27Z0W8dSGxpCs4Ka0JjKQJYZLT9/ti9lxPeH0skcP2FzLz0eZJ3a1e87Jcc+a8s4HmT18qM1PJCSGabadG1rUaNMuA+DNCuwORXH0IPi/skmCZeoFzYdv8Q+3d1/AoA5AO671qS21Ivp5bo0aNCg8kGDBpXbOg7iwcyaNYs0X21YRiZkFFkLtNVatkBcQ1GAt6Mc3o4KeDvK4ariQNMUKAC8IKJQZ0ZOiQHpBTpcLibfKR/E5WIDzl8pF9s2d4gFsNlWcTS5pEsQRO0r0Zuf3XAsq1FUPqttPk4K9ApxQw+NW3lUC0epCOjMViGLopCq5Nh0ioKVAiUKoijRm/kWgiAGcCztz4uiLCG9iD9wIV+15+wV5JXbbPDWYPyUmKX2c1GOt5dLSNIlCKLRUCikbNsDF8je3LuRMjQGhXtiWlf/cj9XJXhejFPJ2N0ADgDIUXK3n8OAgr28yqaFFr1C3LtF+TkNen1AyOB/LpVY18Znqvf8dwU82aJ1R/FphWBpqjsACoBNPiSSdAmCqGldU/LKDXozTxZy3IKmgOHtvMXX+gcbWIb+x14u+RDAbkhw3z37d5AJINNOJlkHQNHR33lYqKfd3DcGhgQu2Z2k3HUmF/dYJ9skZRTqYRVEOQA/ABdtEUO92KebnJwsdXZ2Do+OjtZ06dIl8NbXR4wY4dumTZvg6OhoTXR0tCYrK4t8WXgIGRkZko4dOwZFREQEb9u2TX3r6xMnTmweHR2tGTVqlK/VWrFttlevXgFqtbrtnY7/8ccf7cPDw4Nnz57tBQAnT56Uvfrqqx53uvfbb7/tfvjwYXliYqIsIiIiODIyUjNy5EhfQRBgsVgwaNAg/w4dOgQ988wz3jX8tm1m4sSJze/22tChQ/0EQbjbyw2a0cL3/SMpj0wt3yKyhSMOvNJT93Zs6D/OKu5xe7mkC4A44KES7q30ANbbySXtvBzkw/43vHXSvhdjdK2a2d33xKbm6MVCHkAPW92/2klXtAiP1PDg2vl3VVmc4vDhwyl3er2yfGBCQkKyj4+P9U7HEPe2cOFCj3feeSd7//79F5YsWeJ182sHDhxQWCwWOiEhITk0NNSwceNGBwBYtWpV1vTp0/PudL3169c7JyQkJCclJckB4OOPP3abP3/+bcfyPI/4+HhVly5dDG3atDGdOnXq/IkTJ5IB4ODBg4q1a9c6tm7dWn/s2LELBoOBio+PbxTdaNasWXPpbq9FR0drt2/fftsXmcbAZBViD6UW3LPEaVMiYSjM76cxr50aXdrcSfGUnUwSCeBILd1OBLBXLZOE+bkqn/l5Rufylx4PsjA0WURe6UByvqrUYOlvq/tXO+lSEpq9/OpBPOw/lIS+5+g0Pj5eHRkZqVm0aJHbvY6rZDKZqE6dOgW1b99e07dv3wCr1VplxBwdHa0BgMmTJzePiorSREZGalJSUqS//PKLXXR0tMbFxSW8TZs2wYMGDfIHgEmTJjVv3769pmfPni0LCwuZm68VHh4ePGTIED8A+Prrr53Cw8OD27VrF1yZHHx9fVtFR0drwsLCQk6fPs0BwNy5cz2jo6M1HTt2DEpOTpYmJydLO3ToENSrV6+A8PDw4PPnz0vT09MlXbt2DbRarXj++eebrV271sFisSA2Ntavffv2mtjYWD+LxYKdO3eqvby8WkdERATHxsb6AcDmzZvtoqKiNK1atQr58ssvq1VT9Ny5c4pevXrp7O3tBZVKxRcXF1//+79w4QLXunXr62UWjxw5ogSAFi1a3PVbuEQiESwWCyiKQnx8vLxly5YmR0fH24ZvR48elXt5eZmBG80YAEAqlQp+fn7mtLQ0Ljw83AAAbdu2NRw8eLDKKOns2bNcly5dAqOiojTz5s3zBCpmP86ePctt2rTJbty4cS2AinrT7dq1C27fvr2msLCQ0Wq1VGxsrF/Hjh2DBg4c6G8ymajPP//c2dvbu7UgCCguLqaVSmXEzp071ZWfcXR0tCY4ODh01qxZXgBw859X/ky9+uqrHq1atQq5+TiNRhMaGxvrFxYWFnLgwAEFAERGRmoqY23Tpk1wSEhIaGVLx759+5Zv27btnjXGGyhHmYRucSqr2NZx1AvOSiniZnbVTejY4pBCymoAbEXdPEsUaYpaJ5cywVO6+B7fOL2jTsWRCUIAOJZeCJametjq/vVietnHx8eSmpp69ujRo8l//fWX3bFjx24b6YwbN86vQ4cOQSNGjPDVarWURCIR9+7dm3L8+PFkjUZj2LFjhx1wY8SckJCQDNzetGDYsGFllQX3N2zYkL5z586LBw4cUOj1evr48ePJo0aNKvrkk09cb75WZWlIq9WK5cuXuyckJCT/+OOPF19//fVmAODs7GxJSEhInjVr1tVjx44pEhIS5Dk5OdKEhITkr7/+OmvRokWeAJCfny/Zs2dP2rJlyy4tXrzYw8/PzzJs2LDicePGtcjKypI++eSTJWvXrnUMDg42Hj9+PDkkJMT4ww8/OALAyJEjC0+dOnXeaDTSPM+jb9++2sTExORTp04lrV69+o49ZW91rYQkAECtVvMFBQXX/ysMCwszHjx4UA0A+/btsyspKbnvf6Fz5869Onr0aL/hw4cXf/75524xMTHa8ePH+6xcudLx5uOSkpJkvr6+1/eOrF+/3j4wMDCsoKBA4u7uzgcHBxv379+vBoD9+/eri4uLq4yS5s+f32zNmjUZiYmJyUlJSbLKZg/Hjx+Xr1y50nX16tVZQEW96ZMnT54PDg42ZGZmSj777DPXQYMGlR49evRC9+7dyys/y+bNm5t27dqlXrNmjWN4ePj1Up0jR44sTEhISF66dOn1ESrP8xgzZkxh5c8TAMTFxTkdPXr0/M3HXblyRbpu3brMnTt3pi5cuLDKLAJQMVOzZ8+elD///FMNACEhIaaUlJTG2IyhQ1JuucHCk4eJPk4K7JzVVe/nrPxSLZP0xk0dx+pQjkom6RbqZffTjhe66Nzt7rBCq4lJy9eBF0QlAF9b3L9eJF25XC7a2dkJEokE/fv3Lz116tRtSXfDhg3px44duxAREaH/9ttvncvLy+mxY8f6RkVFaXbs2OF4+fJlyZ2ufa+mBZWSk5O5iIgIPQB06tTpehOEW+Xk5LDe3t4mjuNEjUZjrmxuUFhYKImKitIsWrSoWUxMjO7ff/+VHTlyRB0dHa155plnWpSXl9MAEBwcbJBIJOjUqZM+IyNDBgAzZswo3L17t9Ps2bPzACA1NZWLjIzUA0B0dLQuNTWVA4DNmzc7azSaUI7jRIqicPjwYUXnzp2DOnfurElJSanyeVWO5qdMmVLlmSLD3Mhl5eXljLOz8/Vp+s6dOxtCQ0MNHTp0CCorK6Pd3Nzu+5ypS5cuht27d18MCgoytm7dWv/111+7fvfdd5duLbt5a9Wz8ePHl6akpPzn6elp/umnn+zHjh1bYjAYqE6dOgVxHCe4u7tXuXdaWpps3Lhx/tHR0ZqUlBR5ZmamFACWLl3qSdP09dFzYWEh07p165Bz584pfH19zUlJSbKvvvrKLTo6WrNhwwbnvLw8FgAGDRpUsnXrVocjR46oOnbsqLvXeywpKWHUanWVxhUvvvhibkBAQOu5c+de/3x9fHxM9vb2gp+fn6Xy5+Jm48aN84uKigqJjY29Xku6MTZjsPBC5/i0gibfxi/AVYUdM7sYXFTcPLmUeRU2Wil7Da/k2Gme9vL/7Xyhm54kXiAho4gHEGOLe9eLpHvzNOeRI0dUQUFBd91w5ujoaDUYDPQvv/xiFxgYaEpMTEweNGhQyZ3KWd6vaUGlwMBA08mTJ5UAEB8ff70Jwq28vLysly5d4kwmE5WcnCy9qY2fJTExMXnFihUZ7777rkdYWJgxJiamtHLEvXnz5gwASE5OllutVhw9elRe2aThlVde8Zo/f372kiVLPIGKdnTHjx9XAEBCQoKysrD/yJEjC5OTk89xHCccPHhQ8eGHH3qsXr0649ChQxduTQqVo/lVq1ZVeaZ4rTmBsqysjNZqtYyTk1OVqeClS5fmHjt27IKzszM/ePDgUlTT8uXLXV988cUCvV5P0zQNo9FY5ecqNDTUmJGRIQVuNGMAADs7O0GhUAgsy2LNmjWX4uPjLzAMg9jY2CrFegMCAow///zzxcqmCjExMbpr982MjIzUVU7ZOjs782fOnEnq1atX6ZYtW+w1Go3xpZdeupKQkJB8+vTp8/PmzcsHALlcLvA8T908+r6bEydOKG5tmtCsWTPLE088UXDzSDcrK4srKyujMzIyJLf+fQAVXxozMjLOvPvuu14AkJSUxLVs2bLRNWPQGq29T2QWN+l5TDc1h59mdNSrZJLnJQz9la3juUaUSZj31DJ2ycanO+nUTXyqeX9yvqrMYOlni3vXi0/+999/Vy9evNhLKpWKHTp00D722GM64EajAKBipCCTyUSWZcWtW7em6XQ6+sMPP/Q8deqUQq1W84GBgbf9Aqtu04KePXvqV61aJURGRmpUKhW/ZcuW9Dv1umVZFjNmzMiLjo7W0DSNL7/8MhOoGOlGR0drysvLmUWLFl3u1KmTYdu2bZbo6GgNRVHiqFGjimJjY8ucnZ2tffr0aVlYWMhu2LAh/e+//1bk5uZKly9fns3zPPXxxx+7zJw5s3D48OF+7du317i7u1veeeedK3v37lVt3rzZ+fDhw2qWZcU2bdoYBw8eXDx48OCWYWFh+ltbCN7NW2+9dWX8+PF+RqORfvvtt3OAiprGERERhs6dO+s7deqkYRhGjImJKav8O5g0aVLzffv2OezZs8c+NTU1f+7cuVVqH//666+qDh066ORyuTh+/PjCqKgoTUxMTJWKXx07djQsWLBACgBbtmyx/+yzz9wBwN/f3zhs2LCy9PR0yZgxY/xpmhbHjRtXGBAQUGWk+8EHH2Q/+eSTvmazmWJZVty1a1caADg6OvILFiy42rNnz5bDhw8vHT9+vG9RURHLMIz40ksv5avVan78+PG+3377rZsoiliyZEl25TVXrVqVRVEUXnnlldumgitt2LDBvqCgQDJy5MjrX0BMJhP1wQcfeO7atSvtwIED10d0Hh4e5rFjx/pmZGRwtzajACp+fgVBoAYOHFgCAHv27LF7kC82DQSt4Ji2/1wqsXUcNsOxNNZN66BTceyHDE2ttnU8t5JJmCVuaq7F6slR48esOKpoqvt5j10sBE1Rj9ni3tVueCBahPD7LYa6F9EiWCkJbfOGB7aSnJwsnTdvXrO4uLgmWR/57bffdu/bt29Zly5dGmXtusjISE3liuzqGDJkiN/WrVvTb57yBxp8wwNNQbnpePv39jXZ7UKLBoeZRrTz/l0lY4fAtlPK98KUGy0HVx68GPX5H6n1YuBlC0nv9DPLJUwzANVqolJTqv2BP2rCpCT1YiabsJHFixfbYhFJvdVIv3yFXMgrr9asS2PUyd8Zo9p76xVSdhLqb8IFAF4tk4x+NqZl0p/n81Rns5tm68WrpUajr4vSF3WcdEkmrCMajcbcSH/REgAeZJTbiDXPLNA3yVU6HEvj87EReoWUnQCgyNbxVMNlTkLP+GpcO11T3cN7qVgPAC3q+r73S7qCIAhN82+EIOrYtf/WGmyZKpOV98sq0jfGbVD3NaFjC4Fj6XgAu20dS3XRFPWjk1J6fmQ77/o8Kq81F/N1MtTDpHs2Pz/fniRegqhdgiBQ+fn59gDO2jqWh6U380E5pY3ykf09KaQMZvcKNNnJJS/bOpYHJKplkpnz+mkMEqbp/YrPKtJLdSbrbWWHa9s9n+lardZpV65c+e7KlSutQKaiCaI2CQDOWq3WabYO5BH45pQ0vaT7RLSPQFH4C0BDXCh6VMLQZwa09uwQ90+OrWOpU5eLDTBbheA7dXSqTfdMupGRkXkABtdRLARBNGBShvbMLW10W4/v66lOLXRqmeRjW8fxsOzkkk8mdvb9Lu6fnEZZC/xuskv0oGmq3k0vEwRBVAcjkzD2V5pY0g31tIOrijMB2G/rWB7B9lBPO9rbsVH0Gam27GIDZBL6jl3RahNJugRB1ARPrclqsjaxYgtDI5pZWIb6Hg14ARwAIy+Im/u38mhSf3nFegsoUCyAOu1/SJIuQRA1wSevzFgTfWEblN4hbgaOZXbZOo5HpeTY3T00buX3P7JxyS83GVDHK5hJ0iUIoiY0v1RsaFJLYJVSBs2dFHIACbaOpQb83c7HkWtqW3YvF+tFkKRLEEQD5JFTYpDaOoi6FOplB53Jmgbgrg1aGpArVkEo8XNpWhU8y4xWGkCdLiAjSZcgiJpA8YLYpMZJ3o4KAEixdRw1xcKLmc2b2GIqCy9QAO7Y8rW2NNli1wRB1CiBboT9ge+lmYNcVEjZRlP+k6WplGaO8ugHOcdw8QSK/lgBCAJU4X1g33FUldetpXko/PVT8Poy0DIVXAbNBWvnAgC4uultmHKSIfMOhdvIBdfPyd/xESz5mZAHRMExZiIAoOTwj5C6+UER2PFR32YVZmvdJ10y0iUIoiYINF2vi/zXOG9HuVHK0re1cWyo1DI2xdNeXu1V2KLAo2jvcriNWgSvaV9Dd+4AzAVVP47iv76HMqwXvKZ8CfsuY1Hy95rrr9lFD4fLoJeqHG/OqyhP7zXlS5gu/wfBpINVWwRz7oUaT7gAYOEFGkCdPhYhI12CIGqCwFBUk0q6EoYW0Tie5wIAKIoSB7T2oHxdFNU6/vzp49gQosHi5/oDAH42PwEgC6PGDbl+zHM/52PxnHFwcfeCKLbFmC7/w5fjIq69GoEziUewNe8IFlz7s0sXVVibuRuvPhGOV3ZxeHdMBL77aCEGvPcuAkJa1eTbBQC0aWbPoY63DJGkSxBETWhy08tCRS/yxjRbyPmqZZRbevXaMxZfvkI5CaBejO1G8TyPtprWoouTC3oYqeuj5dbNWtDvTBkNOcdRvCDAoNOi9ZVi3tnBCUNnjaXjTydSCrkcPYxUxT29grAsNZkZ06Elolu3F71SMgQPi0DaaAR2AAAgAElEQVTn/7lXdM/KxqCYfjX6xY7jKQpAna4eI0mXIIiaINJ00xrpCqJIoXElXSdaxkKmcaJFMy8KRqsoGKwQDVaR11spUW9lBKMVgtEK0cjDmFGKY/8kYt/LG+DOOaPL0hFUdPM2KNmaylRe0JSjhdIigWDm4evggQw2E7pfsxiGK8I036GI5ILw/fGfr5+TlJeGMJU//nzxBwxfP5N6Y8GbzKsxT2Pm5sVUiGsAsv5Iwri2sTX2hh2GtjRK3BT5NXbBaiBJlyCImtDkWpFdLTNxvCB6NJZ+tBaT0V5fVoZzB/6gBFGgIIqQyhWQq+3AKZWQeajAyRW8lFOLEikHU56UkiplTPs3+osUQ1OtDrWFngK8FneGaOEhmngx55diavuWrXwzR3exNL+E8uoZxDTrGwLBZEWfiP5IjcsFdYqGxEsJwchDquNgtJohiAIKdEUYqOmBr+LXwdveA8uHLsKI9TMxLOxxyCU100GSYikBgLlGLlZNJOkSBFEThKY20r1UrKfLjZYwB0Xj2J4sCEwATcuF1r0GihKOpSUcR1ktZuhLSoSygjyhODebKrmay5QX5KO8sACn449BV1aKt0b1o+zlMiQc+ht+rs5Y9fLTUDm5QOngSDmpOHyx8gNm0tgxmP+/DyGKIi455oqent68VCqDJMWepvYytPOUMIFiacqdaUdtnXsUsXtmgXbkMGbxdPGjZUvhLnGG85Qwnv6NY2R9vQU7Rlll1H39f003/l0033+WnJIwAoA6LRhOki5BEDVBaCQDvmq7XGQAgCBbx1FjRHju+jqJLszWVvx/CnB0V8DVR007ednTDu6ecA+QCJyCEiQymhZ37qAzKQ4/J1+ExWJBh06dBDuFnDqQeYXy0ZoQbDLCRy7FN+s24Ku16+GiUsBOxmHb+wspuVTCfvXnEeSWlsNkscLFzYEeHdUGGg9X+AAIlDNw8vLFnFnTqDlPjsOCb76Ha4Q3Gx4aCqNjCQOFFZxceX3UzbASimFZimYYimJoimJpgKIgWoXro27BeG3K3GgVRZ2VEgwWWtJMJQNJugRBNEAC28RGuuevlEEhZVsCYABUa/FRPWYnkbHejIQCRQGiCEAEiq/oUXxFf/Nx9LV/sPzXL3D+8gl4ufpg7Se/Yetfq2lOKcErL73Mcwpa1Bu19MYzT9PBoaEwm82YOHas8OlXX9K9n5wipl44T6lPJqHUZIYginj+sU5wUiqw/ugpXCktR4inG9QsDWcG+Hb1GogGHRYNegxf/RmPze8vhpRlgIrP/a4YiQQqR6fro26FgwOlUNtDpraDXK0CZ6eCl8KZAkm6BEE0QMXOKq6hJ54HUqy3oEBrsng5yMMBnLR1PI+os1VvFAa/EE4zEobKyyzjs84V0TkpJVReRjl46+3bd3u1HoXLBakwGkz4569MbNm6GZN6vYFNS04xAPDbyfXQ6qWY2edzyFwETFs4gJ409ml4h/bGyx99jsXvfyzIFBy17NNPqXlrN+Ls6X/Eg7ql+L+5LwrjZjzHXL1ShKcH9sGXm7fBUVmxjUmECF4QcJ98CwDgLRaU5l1Fad7Vux4zedm3Wk6hyH7Iz+yhkKRLEERNSPdxUjSmlbzV8uf5POnYaJ++DE016KQrGAz9S1d/j4KvvqKkAf6wHzKUad25K9+2RyglUXJ0yRWdkJVUjOwLxfSVtFKY9FYEeUdgUNQUbDy4DO9umoyOmv7wdPLFzsTV8HHVABSFvNLLeOP7CeBFHhQYaCT98M2COOpKRhlenvkWfbXkEkwWA/xaaPDkkOeo0iIrJPIwpqjcgJdmv8if/O80/X8/b6U+/fxzrD5zBrGjnhAnzJgmFF/JRenVXKa8sOL5cllBPvQlJRDFB+uwKFerWQBk9TJBEA1Ouquaa1qFewHsPpPLDQ73mmYnl7wPNNiKXJTI8yO1f//NAIA57SLyP/kE+OQTBgBoJyc4DBlCB/boIQQ/EcxL7VWMvtQkXjpfLNgdtjL7/9uM10Z8d/1ig6ImAwCCvNri/OXjMFkNMJr1mNL7LXCsDHkllyGXKsHSEpiURmiatcOQDtNA0wz+LT+DXgO7ITpwIPITVcyF//JxQmFEjNPTGDpWBbcWaqo4T8HYuzdHs2AJzykgSjiGlnASmmElMOq00BYVCqX5V4XinBymrOAqVV5QgPLCfJQX5MOo01Z545xCIQNw96FwLSBJlyCImqDlBdHoquJU+dpGU6TpvuIvFsJkFdwBtANwwtbxPKQefHGx2njmzB1fFIqKULR6NYpWr66YyeA42PXrS3n3fpyhYpx5hzUKZurSrshJLeUvJRUxOSklKMrVIelyIrydW2LWoI9RUJaDL3fNQ4Bnawgij7QrZ/HqiG/gqHLHqn3v4OiF39A5eABGdnn++n2/+fUNPNH9Rew5uR7ZhWkI9o5El5CBN4dWZY6ZldFw87GDi7eKdvTwoT0Cg0T/SNrKyWmK5RhawnEUBUBfViqWFxYI5YX5EHjBSjMw1PQHei8k6RIEUSOMFj6nuZM8qCklXVEE1h7NlD3d3f8FFcdOsnU8D4MvK5tX9MOa6ldlMplQFrcdZXHbkW0xM6bLl3H1hWfgOHAg49ExmmeH+NEUy1D9Hv+fOGrMVLhLHSg2k4Gz2gNXSy7BQekKb+eWcLHzAgCE+3ZB+tVzQPCNW/ybcRg+rhqYrUbkFqVj6uNvY1ncHES17AXpXfboWo0Cci6UIOdCSeUfUbglxynspXBrYUc5N1MyngEhEEW64EE+q5rQ5J7BEARRO0Qg7Vq7uyblx2NZDENRowH42DqWhxACmo4p2br1kTZ86ePjceXNN5HZvw+T1i6cyhoxDD6UhbqUfRADZoQKg1/ToIy/in5jOwndenaG0aJFuaEiOSZnn4KH440+8jxvxf4zW9E7fDTMVhNwrbyoCBFWwfooYUJfakbGvwU48Wsmzh3KhdUinHukCz4EMtIlCKJGKKXsfz5Oin6oGGE0GflaE1YfSWef7NjiQ7VM8oSt43kQvFb7btGqVRLR8OAzrHNzspGg16OE59EzLRUznV1gufZY+4mUFEyyWvB6YiLV5vvvKbAM3hw8GB1aWCjmsWB+ZdBXzNyXXobJaIWfZzD6dh4OY1nFIqi//4tDdFAfSCUyNHPyB0QR7/08DWHNo6Hgaq5Msp2rDKyEOV9jF6wmShQb6rN/giDqmRlbT17+5KVNp5vccFfNsYh/rZdBJWO7ADhl63iq6TFrUdGOtMf7KASdrm7vLJPBrm9fqPv0ESWhbQSJsyNjMfHITS3hs84VMblppSjK1dXq0rTek0P1mg4eLwH4tvbucjsy0iUIoqak+7soLbYOwhbKTVa8vf2sbPGQVltVHBuCOi648BCUvE639eq779V9wgUAoxFlcXEoi4ujcG1BlLJrFzgNGMh4dIni2WH+NMUwVF5GKZ/5XxGdm1ZK5WWWQbDWXBZ28VZZAJDpZYIgGqyMZo6K+1ctaKS2nsym+rXycOsS4PKxkmOfv/8ZtqM1a1cIglHl/uYbsBYXQx8fb+uQoDt0GLpDh4FrSZgLDob94FgmvFMXvl3vUEqi4OiiXK1w6dy1/cIXy2A2POQzXgqwd5PLAfxXY2+gurcm08sEQdQQzsIL5a0X/iYxWh6sSEFjYS+X4I+XY/ROCukUmqZ+snU8d2LiTS8VGAreGR43XDE2eCxeCHkaZdu381ff/4ARjfV3gM64usBh6FAouseITIBGkNopGV2xUbh0vli8nFzM5KaWQldSvZXzdi5yjHkzqkgqY51rOezbkKRLEESNKdGbz81YeyLkWHqRrUOxmRcea4kXY5oJNKfqA+APW8dzM0EUnig1la4as3OMPFeXCwBopmyGNY+t4B2NLJ09axZlPHv3wV/vtFQoaQY0BbCg8LOvb5XXS3keb17JxSWzBRxN4V0PTwRy3PXXeVHEqMwMuLMslns3BwC8kpODFJMJMSoVXnR1BQAsLyhAEMehl1p99zcjk8F+wACoevcSJaHhgsTJgTEbrWJuaqlwKamIyUktqagbfYcUF9DOFT3GBR+QqSQ9qvfJ1RwyvUwQRI3hWGZvlK+T5lh6UZPcjvh0dz+81KM5qMPLaHSZEwepsj+Ag7aO65pYvUW/aupvU68nXADI1mWj946BzPz288Sxa9eiaNUqoWD5NzSsd566/aF5cziyd04dKwoLEczJ8EUzb1w0mfBO3lWsbn5jJ9Xa4mIESDlohYoy3cnXRtbb/PwwISsT5TwPoyjijNGAZ11c7v1ujEaUbt2K0ortThXPhbt1o5wHDGA8u0ZZ2WH+DMXQ1NX0Mj7rvyI6J7WEys8qh8CL8AxwsEgV7N5qf3I1iCRdgiBqjFzK/NUtyGXSl3+l2tk6lrq2KDZUfKq9C0WtGw5kxQOXEpUYs+43sNxsMJKVtozNKlhnGKyGZTP2zpCnlKTc8ZgPjn9I7bi4EysnfCn69ekjZM+aRZvTMx7oPmlmE6Y7VczY+nMcciwWFFitcGFZXLFYcECnxQwnZ6wprpgJYSkKJlGAIIqwiCJoisIX+XmY6eL6UO9Td/AgdAcPAtdyGxcSDPshQ5m2HTvx7R4Pq3gunKMVlI4yiqapww91k0fUJL+NEgRRa4608XaQNbXeut9NaCs8FWFHUd8/XpFwAeDiX8CKGDnKr3wKk3YVAFt0u1foLfrv8/R5nzyx8wn5mYI7l3qsdK7oHLpse4w5wKRSvlu2wHH8+CqTsxRFYdrlSxiZkY5NJSW3na/hZNinLQcA/GswIMdiwdVrI+b38/Iw19UNN/9sBHAcPCUSjMjMQD+1GllmM0QAobI7V516UKak88h7/31kDh3CpEe3oy/26gnsWEvLlSwNG5XtJEmXIIialGflhYIQz6Yx0KVpYOdz0Xwvb4Givu0O5CVVPaAwFVjeWYFLR8fAVH4SQFQdhhejt+hTjuQceWLE9hGKrPKsap849+A86rnDL8J+zvOiz/p1AuvmBgBY7+ODLb5++Na7OX4sKcZxfZVeu5ju5IRSXsCwjHSsLylGiEwGBsB+rRZOLIOwOyTT19zc8YuvHyY7OeOLgny84OKCbwoL8GJONn6+Q2J/FNareTCcOAmhrOw/AOU1evFqIkmXIIgaRVPUr10DXRr9Ck0ZS+PQS535MOkVivq2G4XSy3c+0FQGrB+pwJ5XQ2EoOQBT+c8AvGsxtOY6i25dialk9/yD871e3P+iQmvR3v+sWxzJOYKu23rT59wtov+vu2E3cKDoxkoAAM4si14qFf41Vq1kpWIYLPH0xC++fnjfwxNFViu8JRKcNOjxl1aL3mmpeDknB8f0eszLyaly7h/l5QiTyaEXRKSaTFjm1Qzby0phEGp2JbyiYwcrpVDsrNGLPgCSdAmCqFFKjt3eL8zDJqOIuuKikiJhXmfBs+w0qO8fp2EovvcJogicWkdhWZgcid8NgUV/AWbdNwAiUXNlM0NhKt9sFawZm5I3jRqwZYBi/6X9j3RBs2DGlD+mM2+efBfK+XNh99GHPG1vD70g4IhOX2VlMgCU8TzM13bEbC4tRXuFAiqGwUuubvgroCX2BbTEx15e6KBQ4EMvr+vnWUQR60qKMcXJCUZBAIXKessVr9UkVUyMnpZKbbaqnCykIgiipv0V5mUvk0sYGCy8rWOpcf6uSvz6bKQgTdktUnHPMhAe4D2atcC+hRIkrJQgctJURE6aAFZWAKniK9DsDgDJqH7xQwZAB1iMQ8CbRkEUPXBsucTc+QVq/6X90nJLzX3v2XlxJ34/8TulW6llpJQExvw89Fcp0U2pwsaSii8cTzg44qLZjFdzc8BQFAKkUrzj4Vmt6/9YXIwhdvaQ0zQ0HAcRIoakp6O7Sgk7pubqrTAuLpD6+koA2GQRFUD26RIEUQtK9OaTc376J2J/cr6tQ6lRUb6O2Dg5XKQTvhWoPxY9ejagKMCnM9DuSQMCegvg1IBFnwRWlgipIh2A4do/RgBSCLwfTGWhEMVASOQBKL8i4Nw2OZJ/ZXE5ERAFiEO/EU96BAqTfptcK9XBJoVNwpzQZ1G2axd/dcn/mIdplmArjuPGia5zZv/C2NmNsFUMJOkSBFHjeEF87afErAWv/3KWu//RDcPA1h74YlQIqH0LBCphRe08mlN7AO6tALcQQO1pgVTJQ6LgIZELEKwUClIUKM6gUZIFFKYA5Vduv4ZUBcMrKZi0ZzLOFdVOaWFPhSfWPLaCd7HI6MuzZlPGf/+tlfvUNN/NP5fJW7WaAGCHrWIgSZcgiNrgqzNZk9q9s1dmsjb8kpBTu/jhzb6+wLZnRerctnq/IUoYs048onYQnv3juVqthf1y+5fxpP8YFK/5PyH/yy/vWlCjPmDdXBGwd6+e5jgnANWrF1kLyEIqgiBqQ4ZVEE/3DfOwdRyP7M0BIeKbfXxAbRiNhpBwAYDeMZtq79Ge8bf3r9X7fHz8Y4zeOwGSJ4bCL26bKPWv3fs9CnWfPqJoNu+GDRMuQJIuQRC1xF4uWTaxs2+DXsW8fFxbYWp7B4pa1RfIqC/VHKtBXwg244jwfNvna30l24XiC+ga15v+QzwH3y2b4ThxogCq/n03cRg+opxRq1fbOg6SdAmCqC3bwrzsqOZOclvH8VC2zYji+7UARa2IAa6etXU4D4zd/gLd3bs7462qzS3BN7x6+HXq6b9nwm7mDLTYsF5gPerPLAfr7g6pvx8LYJ+tYyFJlyCI2mLiBXHdmPY+9fdB3x1IWRqHXurEhysKKopelFS/klO9UpYNJvuUMKPNjDrbt5VwJQHdtvWm/3XRw3/XTtgNjq0Xi4bsBgwQRItlOwCzrWMhC6kIgqhNrYv15qOR7+xVCA3gV42DgsX+2R0E+6IzoDaMpmHR3/+k+sy5JYzPHsbAXwYiT59Xp7fu59sP70W+LZoSj4u5r71O8zVc0rHaKAotD+zXSdzcHgcQb5sgbiAjXYIgatMZmqIyuwc9XNeYutTCWYGjL3cW7LP+EKm1Qxt+wgWAwlTQeUn81FZT67xKyZ6MPYiJ60Nlh7iI/r/tgbJbt7oOAQCg6t4dtFyeDeCoTQK4BUm6BEHUKnu55NOnOvnqbB3HvbRr7oA/XogSudM/iNSWKQyEBjUjfk/SHbOZ4YHDGAfOoc7vrbVqMeq3cczS5OXw+mwZPN97j6cUijqNwXnaVC2jVi9B9St91SqSdAmCqG0bOwc4My4qW3S2u7++oe7YPD0CzIElIvX7m7W6r9Umck8DxRn8xLCJNtswvT5pPfrtGgJdTAQC9vwqytu2rZP7SgMCIGvVWgDwU53csBqYhQsX2joGgiAaN5PBwocJghhy9GJRvfqiP7FzCywdGgRq+0yROrmmXsVWk9i8JDr4sXeojec3wiyYUfBbAbJXZaPozyLoL+ihDleDYm5s8ynYc+31v4pQllgGZbASjKLi+8jZyWdRdqIMRX8Vofzfcjh0rBhBX/rmEvK25cFSZIEqVAUAyIvLA6/nwXly0Fl0WJfyIy23c6JiprwJWqUS9MePU6jhLkI3c3/tVQMX2HIpxbI2a3Bwq0b7Q0YQRP2hlknemt7d3+qkrD+j3Vf7abCwry+ojWNBnd1S/zaW1qTMw6DKc/lxIeMES7EFhXsLEbAwAIHvBUIURJQeK61yuKyFDAELAhD4biDsouxwZdONcpO0lEbLd1qi5Tst0WJOCwCA8ZIRABD4biD0F/Tg9TwsJRYYLhpg165qb+VPT36KUb+PBzNqkOgft02UtmxZK2+ZdXeHuk8f0Bz3Ra3c4CGRpEsQRF1IE0Wsnd0r0KbVgCp9MaaNOKODC6gfBgAX99s6nDoh3/M6MzlsMi1jZBAFEYJZgMiLEM0iWMeqDedUISrQXEV6UAQoYC26zzNuBhAtIkRBhGgVARrI25oHt+Fudzw8pSQF3eMeZ37jz8D3501wmjK5xgtqOE2ZbIYgrAZQVKMXfkQk6RIEUSeUHPvW6KjmvLejbYtlbH66vTAogAW1okfF886m4sIe0IZifmyHsaJLPxdcePkCzs85D1pOQ91KfdfTiv8uhqqN6vr/FywCUhemIm1xGspOlAEAZF4ySJwkSFuQBvsoe5ivVmyHlbe499/1G0fepKYfeB7qZ6ehxU8bBfamHruPgvXwgOPo0TytUPyvRi5Yg0jSJQiirlylgE/n9wu2SS84KUtj/5yOfKS6BNS3XSkUp9siDJtS7FvMjGk+htKe0iLooyAELwuGYBJQcuTOe2hLjpTAkG6AS3+X63+m+ViDlgtbovkzzZG7IRemvIrJC8/xnmj5Tku49HepGOUOc0Pe9jxkfZWFov13H2wmXk1El2296FMOpfDfuR32Q4c+8ipjt3mvGAB8AeDyo16rppGkSxBEnZFJmA96h7hbgj3uPrKqDXYyFkfnduJbmC5Q1MrHaOgK6vT+9caZTTi493ehpX9LsHYsKJaCXXs76FNv35Os/U+L/B35aDGnBWjJjVQhcZQAAKRuUiiDlTBmGqucV3ayDHI/OQSTAFO2CT7P+6DkSAkE090XTFkFK5756zn6lYS34PLma/Be8S3PODo+1FuUtQqDumdPMy2Xv/tQF6hlJOkSBFGXyliGWvjmoNA627fb3FGOY690EhxzDoD6v1gaZm1d3bpeCirYTxvSDKDMFERRhO6cDpxn1bbHhkwDsn/Ihs9sH7B2N5738joegqUieVrLrdCn6sF53ThXtIoo3FsIl/4uEMwCUPmYVqx47X72Zu5Ft7jHqaxAewT8tgeqHj0e+P15vL1AS0ml8wDUy2YbpAwkQRB1jdOZrFmTVie4JWYU1+qNwr3tsWVaW5E5vV6g9sxnQH7fAQBeM00XV2xYTel4HWQ+MjSb0gz5O/Ih95PDLsIO6R+mw3jZCIl9xahW4ixBizktoE/RI3tNNiiqImE7P+4Mpxin69ct+K0AjJKBY1dHiKKIy99chjHbCHUbNTxGP1gDhLHBYzGv9WxR+/s+4eo7ixlBd/8KYereveH5/v8uMipVEIA6r8JVHSTpEgRhC0+eyyn9esDnh1T3P/ThPBbshu/GhoL6+0OBOvwpmdW7WdeXkdNxqthvS39KrB+Fmu7IVe6KNY+t5D1EOzp79hzKcPLk3Q+WSNBy316dxN19OIDf6yzIB0R+EAmCsIUNPk7KqwNbe9bKxcdF++D7cWGgdr0kkoR7B4c+hj2rEHv69LR1JPeUb8jHgF1DmdW5v1DNv/8ObvPn85REcsdjncaNE2iF4gTqccIFyEiXIAjb6VhutPzZ6+MD8rzymtu++3KfIMzs6gVq05NAar0pRFT/9FqIi60HC0PihjaILyX+9v5YHfMNrywz09kvvECZLqRcf4319ETArp0GWqGIBJBkuyjvr0F82ARBNEpHWZpe9sXYCH1N1UX4ZFRrcWZnN1A/DKrVhHupVEDPNTqEfKVF2NdafHa04ktDkUHE42t1CPxCi8fX6lBsuPugpswkotkn5Zi5+/YdVIN/1KPV1zcWfM3fa0Sb5Vo89cuNY9eeNl+/70P5czHc5a5UJ69OD3+NOnSx9CJitvdhdhmPo8VPP8Fp+jQBdEUKa7b0Ix0YZgnqecIFSNIlCMKG5FJmYZiXffrETr6PXIB349RIflgQV1H0Iucez/5qAEsDH/eRIel5FY5OVeKrRAvO5fN4/5AJvfxYpLygQi8/Fu8funtSfOtPE2Ja3N5fYWuSBTf3hig1ijhymce/z6rAiyLOXOVhsIj44bQFz0U9QllNUYDy303UnIjZNmuE8DAWHF1ITf5rBlTTJ8F300+Cy8yZIqfRZNIc976tY6sOknQJgrAli0rGDpvfL9jY0u3h1lSxNPDH7I58ByctTX3blULRxRoO8XaeahrtPCsSppqjEOJKI7tMRFyyFRPDK545TgyXYFvyncsnnsjhcVUnoE9A1fKLWrOIT+LNeLP7jW04NAWYeRGiKMJgASQM8NERM2ZFSyFhHnGKYM9r8LVrQUW4RTzaderYqfxT6LKtF/2fox5OU6eIjEo1CkCD6MdIki5BELaWImGoF799MlL3oElELWNx9JUugr/1IkWt6ElBm1dLId5dRomAU7k8OngzuKoV4Kmu+LXqqaaRp7t9ECmIIl7+3YiPHpfd9tpbf5rwcicpFJIbn4OaozAiRIKIb3Xwc6Bhz1FIzOExJPjOC4oeCG+GPGknZkfMqpfba+6FoijYKxwNvJR5GcA5W8dTXSTpEgRhcyxDr3RTc/Gv9NWYq3tOMwcZjs3tKDhfPSxSPwygYSqrzRDvSGsWMWKTHp/2k8GOq94Xhq8TLRgQyKK5fdVfv/9c4ZFaLGBYyO3JdF4XDv88o8LHfWV46y8TFvfg8N1JM0b/rMe7fz/aIjRq18tUqHMoE+oU+kjXqWuzI2abPZWeCVJG+pmtY3kQJOkSBFEfiGqZZNyEji30Hfyc7ntwqKca+2d3EOVJm0Xqp/EM+Grn6hpj4SsS7vjWEgy/lijdVTRyyytGt7nlAtyUt/+Kjb9sxZcJZvh+Wo65v5vwf6cteHWfEfGXeJzI4eH7aTm6rtLhQqGAHj9ULdx1KrdiQBrkTOP/TluwaZQCZ/N4pBQ+wkDVrAWX+of4QsTMBjPa7ejZEWOCx+hUUtUYoB5vNL4D9v6HEARB1Il8hZQd//X4dj/3XLpfUWa88yO6HhpXrBoXBurwMpH6+6PbVyLVAVEUMXW7ESEuDF7qdOP56+AgFmtOW/BqVw5rTlswRHP7r9j1wxXX//2Hf8w4nsPj/d4VU83PXlsYlVEiYNAGPfZPUlY5962/TFgRK4NFAPhrqYamAL3l0d4PvWM21f6l/xh/e39cLK39Z+KPwknmhI9iPtLLWfloAPm2judBkZEuQRD1yW5OwqxfPiFSz9C3T9eOae+N1eNbgfr1FZH6+yOb/f46fInH2n8t+DPdirbfaNH2Gy12pw4kX1UAAAjESURBVFjwalcp9l60IvALLfZetOLVrhUJ+XgOj2nbH6250rbzFkR5MfBS03CQUejkzaD1ci0oCgj3eMTvHvpCsBnxwnNtn6vXo105K8fKPit1Ulr6GYB9to7nYZDiGARB1DeScqPlz53/5ka9tvXM9WHk7F4tMSfGG9TPk4CUel10qGGy84Zx1gkMixuGy9p61xEPLMViRZ8V+lDn0G1KiXICGti0ciUy0iUIor6xqGWSQbHhXpef7u5vBYAPh7cS53TzBPV/g0nCrS1ll8Hk/CPMaPN0vRvtUqCwpNsSQ4hTyFGlRDkRDTThAmSkSxBE/dVcb7b+k5lf5hBsz9PUD/2BgpT7n0U8POdAGJ89hIG/DESevu63X93N3PZzzSODRiYpJcrOAO7fbqgeIyNdgiDqq0sKKds72F0lUnvmk4RbFwpTQOed56e2mlpvRrtPhT7FjwwamauUKHuhgSdcgCRdgiDqt1MUK+2D2M/18Gxr61iaBOmOWczwwGGMA+dg61AwwG+AODNiZolSouwOoNDW8dQEknQJgqjv/gSnGo+J2w1wDbZ1LI1f7mmgOIOfGDbRpjWZO3l2wsLOC7VyVt4DQJYtY6lJJOkSBNEQbINUOR1TfzeQEW/tk+2ay4wLHkerJA9XD/tRhTqF4tOen+rlrHwggLM2CaKWkKRLEETDQLPrIbMfi0m79PBpGO3oGqyMQ6DKr/Bjg8fW+Wi3jUsbfNf3O4NCohgP4GBd37+2kaRLEERDEgdONQQTtujQspetY2nU5L+9zkxpNZmWMbc3ZqgtMd4xWNlnpU4tVY8EsK3OblyHSNIlCKKh2Qepsg9GrytH6BCy57G2JP8K2lDCjwwaWSef8cjAkfxHMR+VKiSKxwDsrot72gJJugRBNERHIFV0x9BvyhA9vUE1YW9IFPsWMzPazKAkdA20EbyH59s+b3kl6pU8OStvDyChVm9mYyTpEgTRUP0DqSISvRZmYtg3RrDc/c8gHsyZTZBYjfzggMG1MtqVMTIs67HM8GTokxcUEkUEgNTauE99QpIuQRANWRo4VRsED/oD0//UQe1p63gaHeWBj5jn2z4PhqrZhk7uCndsHLRR18mr0y6lRBkF4GqN3qCeIkmXIIiGTgtOHQungA/wXLwBzTvYOp7GJfE7KECJfX371tgl27q2xdbBWw3eKu93lBLlaACP1oKpASFJlyCIxkCERP4O5I4j8OQ2LaKmkee8NUgZv5yeFfGCSOH2dosPgqZoTAiZIKzos0Jrx9mN4FjuAzTg5gUPgyRdgiAak18hVbRD74UpmLhDT6aba8jBpXCQKMWezXs+9CX87P2wceBG3fNtn/9HzsrbAfi15gJsOEjSJQiisUkBp24N7/bLMDPRgDZjmtRIqrYoj6+hZ7eb9cAzCCzFYkabGdafBv2kC3QMnK+SqqIANNnuFaS1H0EQjVk7mMq3IPOIG+KeU0BXYOt4Gi6Khu7VTPHFv+dS8Tnx1TolxCkEH3b/UOcidzmhkqqeRCOqofywSNIlCKKx42DW/Q+C9RnsekmOM5ttHU/DNfBjnPPrJIzZ9cQ9Z0mltBQzI2aanwh+wsQx3EyaoteiiT27vRuSdAmCaCo6wlj2A4ozmmHnHBWyT9g6noaHkUI/P1185o/nqFN5p+54SIRbBD7o/oFeLVH/pZKqpqKJbAWqLpJ0CYJoSmiIwlOwGD9Byu8y/Pa6HGXZto6pQRGHfSuecA8QJv82pcrGXWeZM2ZGzDQN9B9okLPyqQC22ijEeo0kXYIgmiIVLIbXAXEO4r9mcegTCcw6W8fUMEhVMLySgol7JiGpKAmOnCOmtZ5mHq0ZLQiisEYhUbwGoNjWYdZXJOkSBNGUNYex7FNQVH/EfynBsW9ZGEi+uB9+zHrxuIOreK7wnPWJ4CcEQRTWKyXKhQAu2zq2+o4kXYIgCEADU9lboNkROLWOxuHPpCgl+eOO7JoBPV4z61uPYEWIa5US5dsgq5KrjSRdgiCIG5rBrH8FFDUdybsp/L1Ujrxzto6pfvBuD3R41oDggYAofA+p8kMAl2wdVkNDki5BEMTtHGA1zYTAv4ziDAaJ36nw31aqyU09K12B8LECoqfrIXcsB8t9CUa6HOSZ7UMjSZcgCOLuWACPw1DyHFhZb1z8i8eJ1Uqk7gME3tax1Q6aAVo+DkRN18KvKwurOQ4yu68AHALZa/vISNIlCIKoHkcAo2EomQmKCsB/v1C4sEeG9L/R4Fc+yxyAlr2AkMF6tOxNQ7CmQe7wKYBNAMpsHV5jQpIuQRDEgwuCwA+BsXQ0pMo2uPKvEee2q5G6l0Jekq1jqx5XDRDUX0DYUC3cQjlY9Echd/wRwG6QZ7W1hiRdgiCIR6MC0BMm7VAAseDNSmTF88g8rEb2CSD3X8Cit22EShfAqx3QLFKAb1ctPMOlAKUFEAdOtRXAX2hCPW1tiSRdgiCImkMBaAmgI0zabuAt3cGp/KHLNyL3NJB9QoXiDApl2UDpZaAsBxCsNXNnRgLYNwccfACHFoBjCwHurXRo1o4GZ0fDrDsLqfJPsNxRAIkASCkuGyBJlyAIonZJAAQBCIfVFAGzLgSALxipJyRye5jKTSi/YkHpJQqmchoWHQOzgYFFx8CiZ2A1AVYTIJEDUpUAmZ0VnJ0VnB0PTi1A7gDYNZNAZs/B8v/t3bFKggEUhuEvNew3aZfo/m+m2cW22kMRBzFrcGsIgt/P5XngXMPLWc45fObr+J676Sbzp3Umk02S1yRvSf79lo/xiS7A7cySrJK8JHlO8phkSPKQZMj3ecjpuMz5tMj0fpfZfJtk/2t2uRyn+Egy0trMtYguAJT8+RMRABiP6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQMkPLbMm/rRer4UAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 2874 samples\n",
      "Epoch 1/5\n",
      "2874/2874 [==============================] - 2s 731us/sample - loss: 0.7161 - accuracy: 0.7088\n",
      "Epoch 2/5\n",
      "2874/2874 [==============================] - 2s 594us/sample - loss: 0.3731 - accuracy: 0.8754\n",
      "Epoch 3/5\n",
      "2874/2874 [==============================] - 2s 618us/sample - loss: 0.2374 - accuracy: 0.9294\n",
      "Epoch 4/5\n",
      "2874/2874 [==============================] - 2s 571us/sample - loss: 0.1489 - accuracy: 0.9576\n",
      "Epoch 5/5\n",
      "2874/2874 [==============================] - 2s 583us/sample - loss: 0.1106 - accuracy: 0.9711\n",
      "1416/1416 - 0s - loss: 0.2052 - accuracy: 0.9273\n",
      "test loss: 0.20516728302516507, test acc: 0.9272598624229431\n",
      "[[1195   18]\n",
      " [  85  118]]\n",
      "Train on 2874 samples\n",
      "Epoch 1/5\n",
      "2874/2874 [==============================] - 2s 805us/sample - loss: 1.3955 - accuracy: 0.5612\n",
      "Epoch 2/5\n",
      "2874/2874 [==============================] - 2s 649us/sample - loss: 0.6525 - accuracy: 0.8361\n",
      "Epoch 3/5\n",
      "2874/2874 [==============================] - 2s 628us/sample - loss: 0.4022 - accuracy: 0.9123\n",
      "Epoch 4/5\n",
      "2874/2874 [==============================] - 2s 592us/sample - loss: 0.2985 - accuracy: 0.9315\n",
      "Epoch 5/5\n",
      "2874/2874 [==============================] - 2s 631us/sample - loss: 0.2377 - accuracy: 0.9482\n",
      "1416/1416 - 0s - loss: 0.2796 - accuracy: 0.9301\n",
      "test loss: 0.2796336193963633, test acc: 0.930084764957428\n",
      "[[863   2   5   3   0   0   0]\n",
      " [ 15  93   5   4   0   0   0]\n",
      " [ 18   2 270   0   0   0   0]\n",
      " [ 14   2   0  68   0   0   0]\n",
      " [ 11   0   0   1   1   0   0]\n",
      " [  0   1   0   0   0  12   0]\n",
      " [ 10   5   1   0   0   0  10]]\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import nltk\n",
    "import re\n",
    "import random\n",
    "import tensorflow as tf\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.feature_extraction.text import CountVectorizer,TfidfVectorizer\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.feature_selection import SelectFromModel\n",
    "from data_print import data_class_display,PCA_text_data_display,pp_conf_matrix\n",
    "from EstimatorSelectionHelper import GridSearchCV_Classifiers, Best_Classifiers\n",
    "from ukrainian_stemmer import UkrainianStemmer\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "\n",
    "def ua_tokenizer_sklearn(text,stemmer=True):\n",
    "    \"\"\" Tokenizer for Ukrainian language, returns only alphabetic tokens. \n",
    "    \n",
    "    Keyword arguments:\n",
    "    text -- text for tokenize \n",
    "    stemmer -- if True use UkrainianStemmer for stemming words (default True)\n",
    "          \n",
    "    \"\"\"\n",
    "    text=re.sub(r\"\"\"['’\"`�]\"\"\", '', text)\n",
    "    text=re.sub(r\"\"\"([0-9])([\\u0400-\\u04FF]|[A-z])\"\"\", r\"\\1 \\2\", text)\n",
    "    text=re.sub(r\"\"\"([\\u0400-\\u04FF]|[A-z])([0-9])\"\"\", r\"\\1 \\2\", text)\n",
    "    text=re.sub(r\"\"\"[\\-.,:+*/_]\"\"\", ' ', text)\n",
    "    if stemmer==True:\n",
    "        return [UkrainianStemmer(word).stem_word() for word in nltk.word_tokenize(text) if word.isalpha()]\n",
    "    else:\n",
    "        return [word for word in nltk.word_tokenize(text) if word.isalpha()]\n",
    "    \n",
    "    \n",
    "    \n",
    "pd.set_option('display.max_rows', None)\n",
    "pd.set_option('display.max_colwidth', -1)   \n",
    "\n",
    "land_data=pd.read_csv('land_real_estate.csv').replace(np.nan, \"\")\n",
    "land_data['text']=land_data['description']+' '+land_data['land_types_source']+' '+land_data['land_types_PcmU']\n",
    "def land_types_download():    \n",
    "    data=pd.read_csv('land_types.csv')\n",
    "    data=data[['id','land_types']]\n",
    "    data=data.set_index('id')\n",
    "    data=data.T\n",
    "    data=data.to_dict('records')\n",
    "    return (data[0])\n",
    "\n",
    "land_types=land_types_download()\n",
    "built_up={0:'Не забудована',1:'Забудована'}\n",
    "\n",
    "def Data_Augmentation(number):\n",
    "    unigrams_list=['продаж','земельна', 'ділянка', 'іншої', 'інфраструктура', 'інструменту', 'інвентарю', 'ізумрудє', 'івасюка',\n",
    "                   'яром', 'яворівського', 'шанове', 'чудова','цін', 'цільових', 'цілу', 'цілорічно', 'участке', 'тільки', 'тухолька', \n",
    "                   'турківський', 'твердій','сухий', 'суха', 'сусідніх', 'сусідні', 'судова',  'сторони', 'сто', 'стихії','селі', 'села',  \n",
    "                   'сайті', 'руська', 'росташування','рокитне', 'розташовану', 'розташований', 'розміщені', 'розміщення', 'розміщений', 'розмірі', \n",
    "                   'розділена', 'покупцю', 'показ', 'повідомлення', 'питання']\n",
    "   \n",
    "    random.seed(0)\n",
    "    random_words=[unigrams_list[index] for index in random.sample(range(len(unigrams_list)), len(unigrams_list))]\n",
    "    tsv='для ведення товарного сільськогосподарського виробництва'\n",
    "    df = pd.DataFrame(columns=['text','land_types','built_up'])\n",
    "    for sample_index in range(number):\n",
    "        df = df.append({'text': tsv+' '+\" \".join(random.sample(random_words, random.randint(0,20))),'land_types':6,'built_up':0}, ignore_index=True)\n",
    "    return df.astype({'text':'object','land_types': 'int32','built_up': 'int32'})\n",
    "\n",
    "land_data=land_data.append(Data_Augmentation(34), ignore_index=True)\n",
    "data_class_display(dataframe=land_data,class_column='land_types',expl_lables=land_types)\n",
    "data_class_display(dataframe=land_data,class_column='built_up',expl_lables=built_up)\n",
    "\n",
    "X_train_zab, X_test_zab, y_train_zab, y_test_zab=train_test_split(land_data['text'],land_data['built_up'],\n",
    "                                                                  stratify=land_data['built_up'],\n",
    "                                                                  test_size=0.33,random_state=0)\n",
    "\n",
    "X_train, X_test, y_train, y_test=train_test_split(land_data['text'],land_data['land_types'],\n",
    "                                                  stratify=land_data['land_types'],\n",
    "                                                  test_size=0.33,random_state=0)\n",
    "\n",
    "\n",
    "  \n",
    "def first_model(n_classes): \n",
    "    model = tf.keras.models.Sequential([\n",
    "\n",
    "      tf.keras.layers.Dense(2000, activation='relu'),\n",
    "      tf.keras.layers.BatchNormalization(),   \n",
    "      tf.keras.layers.Dropout(0.50),      \n",
    "      tf.keras.layers.Dense(20, activation='relu'),\n",
    "      tf.keras.layers.BatchNormalization(),  \n",
    "      tf.keras.layers.Dropout(0.50),  \n",
    "      tf.keras.layers.Dense(n_classes, activation='softmax')\n",
    "    ])\n",
    "    \n",
    "    model.compile(optimizer='adam',\n",
    "                  loss='sparse_categorical_crossentropy',\n",
    "                  metrics=['accuracy'])\n",
    "    \n",
    "    return model\n",
    "\n",
    "def fit_print (X_train, X_test, y_train, y_test,n_classes):\n",
    "    t = Tokenizer(num_words=2000)\n",
    "    t.fit_on_texts(X_train)\n",
    "\n",
    "    X_train = t.texts_to_matrix(X_train, mode='count')\n",
    "    X_test = t.texts_to_matrix(X_test, mode='count')\n",
    "\n",
    "\n",
    "    model=first_model(n_classes)\n",
    "    model.fit(X_train, y_train, epochs=5)\n",
    "    results = model.evaluate(X_test,  y_test, verbose=2)\n",
    "    print ('test loss: {0}, test acc: {1}'.format(results[0],results[1]))\n",
    "    y_pred=model.predict_classes(X_test)   \n",
    "    con_mat = tf.math.confusion_matrix(labels=y_test, predictions=y_pred)\n",
    "    print(con_mat.numpy())\n",
    "\n",
    "fit_print (X_train_zab, X_test_zab,np.array(y_train_zab), np.array(y_test_zab),2) \n",
    "fit_print (X_train, X_test, np.array(y_train)-1, np.array(y_test)-1,7)   \n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Реалізація решіткового пошуку для підбору гіперпараметрів нейронної мережі"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\User\\Documents\\article\\data_print.py:50: UserWarning: You have mixed positional and keyword arguments, some input may be discarded.\n",
      "  plt.legend(fracs, labels=labels, loc='best', bbox_to_anchor=(-0.1, 1.), fontsize=8)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeEAAAD3CAYAAAAqsuw8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOzdd3gUVdsG8HvK9k1PCCSUQCAJAQk1QkILRaRIR6qKIqiA5bWAL7yo2BA+QEUQFFCxgCBIERQsIC2Y0GsKhISaRuputk75/gjBAKEnOynP77q4LsjOzHl2N+y9Z+bMOYwsyyCEEEKI67FKF0AIIYTUVBTChBBCiEIohAkhhBCFUAgTQgghCqEQJoQQQhRCIUwIIYQohEKYEEIIUQiFMCGEuAjDMGkMw/Qo9e9AhmHOMgwzR8m6iHIohAkhRAEMw/gB+BPAVlmWpyhdD1EGhTAhhLgYwzCeAH4HEA9gUqmfj2UYRmQYxnz1j8wwTOOrjz3NMEwCwzCmq73n52445gCGYY4wDFPIMEwKwzCPXv353wzDPHv17yzDMMcZhrnosidLbotCmBBCXMsI4DcAPIBn5OvnDmYBxMqybJRl2XjDflkA+gFwB/A0gI8ZhmkNAAzDRAL4FsAbADwBdAaQVkbbTwHwKr+nQh4UhTAhhLjWYgBmAHUBRN/wmBqAo6ydZFneIstyilxsJ4p70p2uPjwOwFeyLP8hy7Iky/IlWZYTS+/PMIwWwAwA75XjcyEPiEKYEEJcKxHAYwCmAFjOMIyu1GPeAPLK2olhmN4Mw/zDMEwuwzD5APoA8L36cD0AKXdo92UA2wAkPUjxpHxRCBNCiGt9IMuyTZblpQDO4/qeaQiA5Bt3YBhGA2AdgLkA/GVZ9gTwKwDm6iYXAATfpk1vAJMBzHzw8kl5ohAmhBDljAcwgWGYSIZhogEMBLCxjO3UADQAsgEIDMP0BvBIqceXA3iaYZjuVwdfBTIME1bq8VcALJdlOaNinga5XxTChBCiEFmWzwJ4C0AcgO8AvC7LclwZ25kAvARgDYpPV48CsKnU4/G4OlgLQAGAnQAalDoEh+JeNKlkmOsH5hFCCCHEVagnTAghhCiEQpgQQghRCIUwIYQQohAKYUIIIUQhFMKEEEKIQiiECSGEEIVQCBNCCCEKoRAmhBBCFMLf7sGDBw/W4nl+GYDmoMAmxJVEURRfa9269QalCyGEVJzbhjDP88tq167d1M/PL49lWZpaixAXKSoq0qakpLwJgEKYkGrsTr3b5n5+foUUwIS4ll6vt8my7K90HYSQinWnEGYpgAlxPYZhgH+XqSOEVFOV5jrvuHHj6rVp0yb06aefrnfjY5GRkaFOpxMAkJSUpB4wYEDD8mp3165d+piYmMYdOnQI2bRpk1t5HddV3nvvvVotWrQIa9myZdiff/5pAIpfr3bt2oVGRkaGlvWcFi5c6NOhQ4eQyMjI0NTUVFXJz9955x3/Nm3ahN64/csvvxwQERERtmrVKg8AmDt3ru/WrVuNZdUzbNiwIIvFwnz22Wc+bdu2DX3ooYeafvTRR34AIIoiJkyYULdDhw4hvXv3bgQA06ZNqx0ZGRkaGRkZqtPpWmVmZnIlxxo7dmy98nyvlXT+/Hl+6tSptct67MKFC/yzzz5b19U1EUKUd9trwqXZnGKEVsXd9fZl7C9oVdzRsh7bs2eP3mKxsAcPHkwaPXp0/Z07d+q7dOliud+27pbVamVef/31emvWrDkbFBTkrOj2KsLKlSt9T5w4cer8+fOqCRMm1O/Ro0cKAMTGxiapVKqbtk9NTVXt2rXLuG/fvusWDrdarcyxY8d0ZbWRmJioi4+PTxoyZEjDgQMHFh44cMDw+uuvX7lxu+PHj2u8vLwEvV4vT5gwIffFF1/MEQQBzZo1C3/zzTezv/76a6+mTZvavvzyy4sl+3z44YcZH374YUZ6ejo/YMCAYH9/fxEoDqbz589r3NzcxAd8iSqF+vXrC7Nnzy5zLdd69eoJV65cUeXm5rLe3t6Sq2sjhCjnrnvCWhXHB725Bff753YBvnv3bkP37t0LAaBnz56Fe/bsKbOXdaNVq1Z5tG3bNrRVq1Zha9eudS/92Pfff+/58MMPhzRp0qTZjz/+6AEAffr0adSuXbvQ6OjoJrm5uezhw4e1+fn53KBBg4Lbt28fcvr0afXSpUu9Zs2a5QcAsbGxuqeeeqoeALz66qsBoaGh4WFhYeEvvfRSAABs2LDBLSIiIiwiIiJsw4YNbsC/vdCwsLDwkt7ije0CQOkeZ8nf//vf/9Zu165daIsWLcL27t1bZiDeKCgoyGa1WpmcnBzO29tbAACWZeXo6OjQfv36NSrdswSAjRs3uouiyHTo0CHkqaeeqicIAgDgk08+8R07dmxOWW0wDCPb7XZGrVZL8+bN83vuueeyy9ru559/9oiKijIDgEajkQHAZrMxTZo0sQHAli1bPE6dOqWNjIwMnTdvnm/pfVevXu3Rp0+f/JJ/z5o1y//FF1/MKqudst73ktdw8uTJgYsWLfK22WxMZGRkaPPmzZuOGDGiAQCcOHFCEx0d3aRdu3ahU6ZMqQMAQ4YMCRoyZEgQUPx+chzXBvj3/Y6MjAwNDAx8qOT9Lf178OqrrwZIkoSePXsGt27dOqxku82bN7t169atcefOnZu0b98+JDMzkyt9BicoKKh5ZGRkaLNmzZoePXpUAwDt27c3//LLL+4ghNQoleJ0dH5+Pufp6SkCgKenp5iXl8fduE1UVFRoZGRk6OOPP94IKD61+fHHH9fet29f0r59+5LmzZt33am+MWPG5MfFxSV/+eWXqTt37jQCwOrVq9P279+fNGTIkLyvv/7aOzc3l/fx8REOHz6cOHPmzMvvvPNO7ZEjR+Zv3brVAwB++OEH7zFjxuQCxT3F+fPnn587d+6Fkjbee++9gB07diTv2LEjeebMmYElP4+NjU1auHDhuV27dpXZ7q1eh2nTpmXt378/aeXKlamzZ88u89TljWJiYkwhISHN+/TpE/Lqq69mAcDGjRtT4uPjk/r165c/ffr0OqW3z8zMVDkcDmbfvn3Jer1e+uGHHzztdjuze/dut/79+5vKaqN///75I0aMCJo4cWL2qVOntGfOnNGMGjWqfnx8/HVfFE6fPq0NDg52lPz79ddfrxMcHPxQq1atigAgOztbFRoaatu7d2/S6tWrfS5cuHDti9mmTZu8RowYkXe1Ru7KlSt8eHi47cZabve+f/LJJz4qlUqeNGlSrlarlePj45OOHTuWkJaWpgGAqVOnBq5YsSJt//79SQkJCdqUlBQVAOTl5fEFBQXszz//7NWsWbOikuPNnj37Qnx8fNLw4cNzSrc/d+7cCyW/B+np6XxeXh5/6NChxNLb2Ww2ZteuXafHjRuX/emnn/qVfg4+Pj7O+Pj4pJdeeikzLi5ODwCNGze2JyQkaMt6/Qkh1VelCGFPT08xPz+fA4CCgoJrgVxabGxsUnx8fNKaNWvOAkBGRgafkpKijY6ODu3cuXNIVlaWSpKuP5M3adKkwOHDhzfu169fgSAImDhxYt22bduGfvnll7UuX76sUqvVUosWLSwA0KlTp6LU1FSN0WiUfX19hdOnT6sPHjxo6N69exFQ/EHt7+8v3FiXt7e35O3tLZUewBYVFRU6atSo4EcffbSwrHYBIDk5WVdyLbRkvyVLlni3bds29JlnngnKzMy87lzyjBkz/CMjI0MXLFjgU/Kz3Nxc9rvvvvM9e/bsiX379iVMmTIlEABKTumOGTMmLyEh4bqg9PDwEDt37mwCgB49ephOnTqlXbx4sffIkSNzb/X+TJw4MffXX389u3PnTuOLL76YvWnTJs/ly5dfmDNnznWjd2X5+jF8c+fOTU9LSzu+YcMG74yMDM7d3V3s2bOnSaVSoV27duaEhAQNABQUFLC5ubl8WFiYAwA++ugj/5dffrnMXvCt3vfc3Fz+66+/9tPpdNd+CbZs2WIMDg5uVtITT0lJ0Y4aNapRZGRk6OnTp3Xnzp1TA8DAgQPzFi1a5MNxnKzVam87ELGgoIDz8PC49vsZGBgotGzZsqhevXrNV69efe29adGihRUAIiMjLSkpKZrSx8jJyVG1a9cudObMmYFdunQpAgBJkkoGYxFCapBKEcKdOnUq2r59uzsA/PHHH+7R0dHmO+1Tu3ZtITQ01FoSzqdOnTrFstc/nUWLFl2KjY1NmD17dp19+/bpLRYLe+DAgaTx48dnybLMhIaGOk6cOKEDik+JBwUF2QFg9OjRuS+99FLd1q1bF7EsC0mScObMGW1ERMR1PTNZlpnc3Fw2NzeXlSTp2idobGxs0u7duxPfeuutgLLaBYCQkBBrfHx8Unx8fFLJfsuWLasVFxeXtHz58rSS7Uq89957mVd7T9d6WxzHQafTSVqtVvb29hYtFgsHFIczAPz555/GkudUonPnzuZjx47pAeDQoUP6hg0bOpKSkrRffPGFX6dOnZqcPn1a98EHH9S68fXOy8tj09LSNB06dLDabDaWZVlYrdbrXvCQkBBbSkqKGig+cwAUn5bW6XSSTqeT27dvbz548KAeAE6ePKkr6TX/9NNPHj169CgoOU5aWpr6zTffrDtmzJiG+/btc1u2bJnXnd53g8EgxcXFJR04cMCQlJSkBoC+ffuaU1NTT6akpGgvXLjABwcH23766aez8fHxSSdOnDhVEoDdunUzr1ixwm/48OF5Nz7vGyUkJOiaN29+3e+Bh4eH+Nlnn50r3RMu+b3av3+/vlGjRte9Bz4+Ps79+/cnffnll2nvv/9+bQA4e/asJiws7KaePyGkervvgVblqWPHjpbly5dLbdq0CW3WrJk1JibGcv78eX7RokW+txrMwnEcXn755cyoqKhQhmHkkJAQ23fffXe+5PH58+f7rlmzxttms7GvvfZaRosWLWxpaWmaTp06NQkMDHQEBAQ4GzRo4IyOjja3bNkyTKPRSN99910aAPTv37/wueeeC3rrrbfSAeCll14KPH36tC46Ojq0sLCQKyws5F5//fWs6dOnX46JiQmRZRlvv/325ZK2o6KiQs1mM/fcc89lltXurV6HVq1aFUVGRoZ26NDhjl9CAMDDw0Pq1q1bQcuWLcNEUWSmTZt2GQA6d+4cqtFoJI1GI/3www9pQPEo5PHjx+dERUVZv/nmGykyMjLU29vbOWPGjEytVnutF9ymTZvQ6dOn39QL/eijj2qVnO4OCwuztWvXLnTq1KnppbcZNGhQwcKFC/3Gjh2bP3369Dp79uxxczqdzIgRI3I8PDykF1988crw4cMbLly40L9bt24FwcHBTgDYuHGjZ8lrDQDr169PA4pHwk+ZMiXw2WefvRaOt3rfNRqNpNVq5Y8//vjiK6+8UvfTTz+9OHbs2AYOh4MNDg621atXT5g9e/alJ554IsjhcDA8z8tbtmxJKTnusWPHEjQajfzWW2/d8vWeMmVKnQEDBuSVHjx16tQpdUpKinbevHnpu3btujYSXaVSyZ06dWpit9vZTZs2nSkoKLh2iSUnJ0cVGRkZajKZuJkzZ14EgH379hm/+uqr8yCE1CjMjacQSzt69GhaRETEFaBiR0dXNoIgoHPnzk1iY2NPA8WDd9atW5dW8viUKVPqjBs3Lic0NNRxy4PUUI8//niDb7755rxer6+x95dv3rzZ7ffff3dbsGDB5TtvDVy6dIl/66236ixfvvxC6Z8fPXrUNyIiIqhCiiSEVAp3HaoPGqBa1U1jrSqlzMxMrn///sHjxo27NgJ44cKFF0tvM3Xq1Cx3d/dqcetMeVuzZs05pWuoagIDA4UbA5gQUjPcdU+YEOJa1BMmpPqrFAOzCCGEkJqIQpgQQghRCIUwIYQQopBKEcJpaWmq8PDwphqNpnXJQg2ldejQISQqKiokMjIydMuWLbec0vLy5cv8wIEDG0ZFRYW8/fbbVW4ZuE8++cQnMDDwoRsXLZAkCaGhoeHz58+/bqrH9PR0vlWrVmHt2rUL7datW2Oz2cwAwPr1693bt28fEhkZGbp792596X3mzJnjFxERETZnzhw/AFi3bp37okWLypzFa/z48XXPnz/PA8D06dNrl7wHoihi7969upCQkPDAwMCHSu9z43ZVHS28QBTAAwgA0BpAJwA9APQFMBjASABPAZgA4FkAwwA8AqA9gKZX9zOCVuCqMu7+liOnLQIq7f3fV+y0CVBpyxxhXatWLWHnzp3Jffv2bVzW4xs2bEjx9/cXCwoK2O7duzfp27dvUlnbPf/88/UmT56c9cgjjxSV9XhlN3z48IIePXqYp0+fHlD65ytXrvT08fG56duJn5+fcODAgUSO4/Daa6/VWbNmjefjjz+e/8UXX/ju2bMnmedvfrv++usv96NHjyY+8sgjwVOmTMleuXKl95o1a9Ju3C43N5fNzs5W1a9fX9i5c6febDazsbGx1xZ9CA8Ptx84cCAxJiamScnPytquqqOFF0gFMQBoCSDC5hSbWB1iU4ZBIw3P+at51mhxCPbcIoezyC5IdkGCQ5AYe8kfp8jYBYnlWEb20KlED51KdtPyMGp51qDmOa2KU7EMGItDzBAkOUWn4o7p1FwCgNMAzgC4AKDqf0OuJu4+VFVaHu943H9L7xTcsi29Xi/r9fpb/lKUTMNoNptZg8EgAsWLHfz5558edrudWbx48bno6GhrYmKiburUqfXeeOMNzJw581KrVq1skyZNqrd58+azTqcTnTp1Cvnnn3+SN2/e7DZhwoSgunXrOkRRxMGDB5NOnz6tvjqRA9u7d+/8Dz74IOPVV18N2LJliyfP8/Lw4cNz3nrrrayy2o2MjAzdu3dvkkqlQsnfN27c6D579uw6VquVff7557MmT55c5uIIpdWpU0coLCy86ezEjz/+6D106NCbZnMqHbKiKDJhYWG27du3G1mWRZcuXZr4+fkJ33777Tl3d/drAcFxnOx0OsFxnPz999979u3bt4Djbr59bPPmze4REREWAFi/fr1nTk4O//DDD4d06tTJNHfu3HQvL6+bQqes7Uo/vmrVKo958+bVFkWRmT59+uWhQ4cWtmnTJvTgwYNJkydPDmzatKl13LhxeZ07dw6xWCxs8+bNLT/++OO5EydOaJ577rn6DoeDjYmJKZwzZ056yaIL69atS9uwYYPbkCFDQkRRPFjynnl4eIiXLl1SL1q0KG3gwIGmkp8zDIM+ffrkz50793KvXr2Cc3JyVJmZmapFixal8TyP+fPn+wuCwDgcDmbjxo0p+fn53JQpUwI3btyYGhQU1LxWrVrOoqIiduXKlWcjIiLsJQsvPPXUU/k3vh6EXOUGoBWA1gUWZ2eGQTudmqt9LsdiOXIhT3Umy6y7mGfFxTwrLudbkVPkgCjJOgB3tYhLWbQqFg28DXWDfPV1g3wMXZr4G60h/m6O+t56lUHDq4rswhk1z+7Uq/ndAPYBSANQY+/tV1KlmDHrTgRBQIcOHUJTUlK0y5cvTwWKFzuYNWtWxokTJzTTpk0L2LRpU2pGRoY6LS3tGAB079495PDhw4kWi4XNy8tjt2/fbuzatWshAFgsFmbkyJFX5s2bl16y+s67775b+5133rn86KOPmjt27NgkLS0tByiexL9Xr17mIUOGNASQVVa7ZdXcq1cv89ChQ5OcTifat28fejchXJZ169a5d+rUycTzvCwIwk2nmHbs2KF/8cUXG2g0GmnGjBmZx48f12VmZqri4uKS5s2b5zd//ny/d955J7Nk+2efffZK//79G02YMOHK999/7/3kk0/mjBo1qv6IESPySi/gkJycrGncuLEdALKyslTe3t5CXFxccr9+/Rrt2bNH37Fjx5uWmrzddqUXXhBFkYmJiWkydOjQQuD6hRcAID4+PkmSJERFRYUA/y680LhxY+djjz3W8G4WXigJ3pKflSy8wDAMtm/f7lZ64YXS29lsNiY2Nvb00qVLvT799FO/p5566tpsYiULL3zxxRfecXFx+oiICHvjxo3tBw4cuO6UP6nxOABtHYLUz+oQh+nUXKPUK0WWg+fytIcv5GlOXipEcqYJgiRX2KpZNqeEpEwTkjKv/Ze+Fup6NYeHAj2atqzvGda+kc/oNvW9eI5lrIIk7/LQqX4F8BuAi7c4NClnVSKEeZ7H/v37k65cucLFxMSEDBo0qHDJkiXeq1ev9mFZFgzDyAAQHBxsKzktWLKgwoABA/JWrVrluWPHDveZM2deBoCcnBze19f3usUY0tLSNFFRURYAaNGihSU5OVkNAFOnTq03ceJE7uWXX84AihdZuLFdoHiqSoZhULJgwt69e/XvvvtugNPpZE6fPn3dN9r169e7z5o1q07z5s0tX3311W0naVi2bJnvunXrUpctW1bmdduYmBjLiRMnEt5++23/hQsX+oSFhdnbt29v5nkejz76aOGsWbOuu545ePDgwsGDBxcuXbrUa8iQIXnLli3z/eWXX1L79+/fsHQIl5672t3dXYyJiTEBQJcuXQqPHz+uLSuEb7dd6YUXSt6D0gsv9O3b91pPcsuWLcbJkyc36Nixown4d+EFACgsLOQeZOGFwsJCDrh+4QVJkpiSRS1KL7zw559/XvchWbLwwqVLl9Q7duxIAmjhBXKNL4BeBVbnMA3Pds822eVtJzO0fyVkqQ6cy4VTlB/gNGL5sjhExKXmIi41l/li51k3AAjy0esebuQzsEuIX8+uIX4LHKKUatTyK3iWXYfiU9ikglSKgVm3I4oisrOzOQAwGo1SSfDdarGDGxdUeOKJJ/LWrl3rnZGRoQoPD3cAwJ49e4ydOnW67rpxgwYN7Hv37tUDwLFjx/RNmjRxAMW9qosXLx5ftmxZrdu1W7KgQNOmTa0AMGfOnNpff/112p49e5JvXJh+0KBBhfHx8Ul3CmAASE1N1fbq1avxwoUL/RcvXux/+PDha8vd2Wy264JSp9PJnTp1KkpKStICQHx8vP7GBRyA4jMLW7du9Rg1alSBzWZjAdy0GENoaKgtNTVVAwDR0dHmI0eO6ADgyJEj+pIe8o1utx0tvECqIR2AkQVW5z6bU7y0Kzl78ezfEgd0m/u3sdOcHW7vb0lQ7TubA6dY+c/ypuVYsHr/BUz84ZChxczftZNWHmr604GL7xRYnccKrc5UhyC9B6AFaMBXuasUPWG73c7ExMQ0SUxM1HXu3Dnkgw8+uNS4cWP7okWLfN9///2MgQMHNpJlmXE6nUzJogFlLXbw3//+93KXLl1CWZaVSxZU8Pb2ljQajRQTE2MGgLVr17qvXr3at6R3mpycrFu7dq37W2+9lTFmzJiG06ZNYx599NH8hg0bOoHinvDbb78ttWvXznSrdsvSv3//vP79+zdu1qyZ5W6nuFy1apXH3Llza58/f17Tq1ev4G3btqUkJiaeAoAFCxb4CILAtGrVyrZgwQKfVq1aWQVBwJQpU+qxLCt7enqKa9euTXVzc5M6depkatu2bahOp5PWrl179sZ2lixZ4lOyTnL37t0LH3rooaZjxoy5bma0vn37msaNG+cFACNGjMgfOXJkULt27UKbNGli69mzZ9GZM2dUTz75ZMPk5GRdVFRUyNdff51W1nYlx6OFF0g1wQBoW2QXXuBYZvjxSwXid/vOuW07mQG7IKmVLq48CJKMvWdysPdMjnba+uNoVc8rqG+LOlMGtgz4j4pjL7vrVLMBrAJw09kwcu/uftrKChwdXdEee+yxhgsXLrzYoEED54IFC3waNWrk6NevnwkAtm3bZkxKStKUXiKQFBs/fnzdGTNmZNSvX/+mdZSro/JaeKG80LSVlYqnIEnjrA7xRatD9P0+7rx27YEL3OWCmnMShGGAzk38ML5zI3PbBl6MJMvf6NX8pygedU3uU7WfO3rkyJENdDqdVHLqt6CggFWpVHLJKj9Wq5VxOByMh4cH3WJSw91rCFc0CuFKIdDiEN5gGWb89sQsrIhN08en5eI2H5s1Ql0vHca0b+Ac/XB9QZJx2KO4d7wZAH2O3qNqH8KEVFUUwopqaLYJ73IsM3T1gQvM0l1nNZfyrUrXVOloeBa9m9fBi90bm2u5aS67aVX/QfHo6hr+NeXu3en0siRJElMy0pgQ4hpXvxzT/zvXa2iyOd/jGGbwin1p/NLdZ1X5lptn8SPF7IKEDUcuYcORS8ZezWqH/K9v0zUeelWKu1b1CoAdStdXFdwphE9kZ2eH+/n5FVAQE+I6FotFyzBM5p23JOXEYHWIb8uQJ6+IPadauvssX2Cl8L0X205m4I9TGYb+EYEt/tsn7BedijvurlP9B8A/StdWmd32dPTBgwdr8Ty/DEBzVIHbmQipRkRRFF9r3br1BqULqeYYAMMtDmHR30nZupm/nNRlFpZ5Bx65BzzLYGibuvKUR8OsKo75202reg40AUiZbhvChBBSjUWYbM6vskz20KlrjxkOnLvj7ebkHml4FpO7NXY+27GRk2OZd9Q8+wkAOsVQCoUwIaSmMRbZhY9lGaNn/ZagWRV/npXoY7BCBfnoMWdoRFF4gHuGUcMPB3BQ6ZoqCwphQkhN8nCRXfj5j1OZ3u/8clJLg65ca1CrQPndAc1sHMss0av56QBq/JBzCmFCSE3A25ziW05Rev2Ntcd0W0+UuTolcQFfoxofDn7IEhXsm23U8H0AnFK6JiVRCBNCqrtGJpvz58QMU+PJKw8ZaOBV5TCsTV155oBmVg3PTeRYZoXS9SiFQpgQUm0JkvSkU5A/n/d7knb53lSOPu4ql1B/N3w1tp3FQ6/aYNTw41ED56OmECaEVEdckV1YmG9xPjFuxX5DYobpznsQRejVHP5vaIS1S6hf+tXT00lK1+RKFMKEkOrG3WRzbkpMN7Ud9+1+Q6G1Rqw/UuWNiqwv/a9fU5uW555gWeZnpetxFQphQkh10rDILvy16ejlgBkbTmgEuveoSmkW4I7vn33YYlDz/1Xz7AKl63EFCmFCSHURbXWIW2ZvTTR+E5vG3XlzUhnV9dJh9YQOFi+DarFezb+Baj6HOoUwIaTKEyRppN0pLZv0wyH938nZSpdDHpCXXoUfnm1fVN9HvxMk/PMAACAASURBVMWo4cegGs+yRSFMCKnSnKL0TJFdWDj8i390SZk0AKu60KpYLH2yraVVPc+DRq2qDwCz0jVVBAphQkiV5RCkiWa7MHfo4ljd2StFSpdDyhnHMpg9pIXt0ea1zxo1fDSAfKVrKm8UwoSQKskuiJNMVmHO4MWx+vO5Ne720hrlvQHN7YNaBSYYtXxHANXq2xYtT0gIqXKcojTebBP+b9DivRTANcBbm05o/kzIDDPbhG0ANErXU54ohAkhVYooyU+ZbcKnQxbH6i7k1vj5/2sEWQZe++mo9p/UnNZmm7ABAK90TeWFQpgQUpV0tziExUOX7NOl5VAPuCYRJRkTvz+kO34pv7PZJqxENcmvavEkCCE1QpjFIawf/+0BXUp2tRwoS+7AIUp4+pv9+pRsc58iu/AFAEbpmh4UhTAhpCrwLbIL29/edNLwz9lcpWshCrI5JYxeFmfINtlHOgTpZaXreVAUwoSQyk5jtgnbfog75/PTgYv0mUVgtgsYszzOYHOKHwLoqXQ9D4J+oQkhlRljtgnfxaXmhM36LVGtdDGk8riYZ8X4bw/orA5xHYDGStdzvyiECSGVls0pvnG5wNpn8srDeprSgNwoLjUXs35LMJjtwlYAeqXruR8UwoSQyqq1IMnvPP31foPVKSpdC6mkvt13jt2ZnB1gtgtfoQoO1KIQJoRURoYiu7Bh2s/HtZfy6V5gcnuvrzmqyzHb+zlF6Rmla7lXFMKEkErHbBc+356Y5bvp6OUq17Mhrmd1inh2xQGDIMqfAqivdD33gkKYEFLZDLTYhaH//fm4TulCSNVxOsuMhTvOaEw250pUodPSFMKEkMokwOIQVjz//SG92S4oXQupYpbsTOHT820tBVGaoHQtd4tCmBBSWTAmm3Plst2p+kPn85SuhVRBoiRj0spDBqcozwPQQOl67gaFMCGksuhfYHW2/Wz76WozOT9xvVKnpVehCpyWphAmhFQGuiK78OWb644bnCLdEEwezJKdKfzlfFsLQZTGK13LnVAIE0IUZ3OK/913Nse458wVpUsh1YAoyfjP6iMGpyj/HwB3peu5HQphQojSgmQZr7+98WSVnPGIVE6n0gvxx6lMldUh/k/pWm6HQpgQoiiTzblk8c4zKpqUg5S3Wb8l6ABMBlBX6VpuhUKYEKKknlaH2OmLnWdpMBYpd+kFNnwTm8abbcI8pWu5FQphQohSGJPN+enMX07p7YKkdC2kmlq044xKgvwYgAilaykLhTAhRCm9c4sc9X49ka50HaQaM9sFzN2WpDHZnItRCW9ZohAmhCiBMdmccz/6LdFISxSSirYy7jxrcYgtAEQpXcuNKIQJIUp4JM/irLf1ZIbSdZAaQJBkLNpxRl9odc5QupYbUQgTQlyu0Op87+M/kqkXTFzmpwMXGY5lugAIVrqW0iiECSGuFukUpWa/HL2sdB2kBrE6RXz3zzmuyC68oXQtpVEIE0JcqtDq/N/CHWe0gkTdYOJaX+1JVXEs8yQAL6VrKUEhTAhxJR81zz6y7tBF+uwhLpdlsuOPU5myQ5CeU7qWEvQfgRDiMqIkj96emCUUWmmtYKKMxX+n6AVJmgJApXQtAIUwIcR1mCK78PL3/5wzKF0IqblOpRci7YqFA9BL6VoACmFCiOu0cgiS/76zOUrXQWq4H+LOuRVanS8oXQdAIUwIcZEiuzDx+7hzGrotiSht87F0RsOz3QF4Kl0LhTAhxBV0HMuMXLP/Ai3UQBRXYHUiNiXHCWCI0rVQCBNCXOHRhPRC4XKBTek6CAEArDt00ZhvcTyjdB30rZQQUuFMNueQX45ediuPYxXu3wDz0d8BBlD5BcG3zyvI2bYItgsnwGr0AADfPv+B2r/RTfsKhVnI+e0zCIXZYBgGtYa9A97D/9rjuX8sgfn4n6j/6tritg7+AvOR38C5+6HW4P+B4VSwXTwJS1IsvLuPL4+nc1sankUDHz2CfAyo5a6Fj0Et8yyDQpvAmGxOmO0CcoscSEgvRJ7FWeH1VCfbE7OgVXFtAPgAUGygAoUwIaSiMRzL9N2emPXAK9gIpisoPPgLAsZ9DlalQfaGj1CUsAsA4NX1aRjCOt52/yub58Ojw3DoGraC5LACzL8l2dNPQ7IXXbe9+eg21HlmIfJ3fw9r6iHogiNRsPdH+A6Y+qBP5ZZC/I2ICa0l936ojim8jrvO5hQzBUk+q+KYNL2av8CxjNPqEH0couQrSbI3GNQxqPmQQptT3HvmCvN3Urb+j1OZMNvpNrDbsThE7D1zxdm9qf9AAMuVqoNCmBBS0SJMNoFPy7GUz9EkEbLggMzxkAU7OKP3Xe3muHIekCToGrYCALBq3bXHZElE3t9fwfexN2BJ3nf9jqII2WkHw/IoOrkduuC24LTG8nkuV3nqVRjzcANxbHSQTafirDKwyajhNwDYoeZZ843b69QcdOBK/4j1NWpCB7QM7No1xG/oh4Meitpw5BK+3puqTc68aXdy1W8nMoxtGngN9tSrKYQJIdWTIEr9tp7IUJfHsXg3X7hHDsKlxU+D4dXQNmwFXcPWKDq1E/m7v0NB7I/QNoiAV5exYPjr52IQci+B1RqQtf4DCPmZ0AW1hGeXp8CwHEyHNkPf+GHwNwS6e+RgpH/3GtS+9aEJbIrs9e+j1rB3y+OpACgO34ldGzueaN9AEiRpo5tWNRvAEQD3OoZcApAAIMFDr14MIHBwq8DnB7QMmHwmy8x9sCXBLS41t9zqri72peRAo+I6oXidYUXG7VMIE0IqlNkuDP8zIbNcQli0mWE5HYfA55eD1RiQvfEjmE/ugGeXp8AZvABRQM62z1AQtxae0SOv21eWRNgunESdpxeAd/fDlY2zYT7+F3SN2sCSuBf+o2bd1J6xeTcYm3cDAOTvWQm3Nv1hPXsQRSf+AufuB69u48Aw9ze+dXDrQPm9Ac1tMrBGp+ZmANyF+zpQ2S5pVNwMAO+2qOs57Oux7T7dffqKfvqG4/orZkc5NlO1Xcq3wmR1MjoV1wzACSVqoNHRhJCK5KNVcSFxZ8unF2ZLOwLewx+c3gMMx0Mf0gH2Swngjd5gGAYMr4LxoR5wpCfftC/v5gu1fyOoPGuDYTnomrSHIzMFjqyzcOZfxqUvxuPi4mcgO+249MX1g64EUw4cGaehb9IeBfuKrwkznAq2tKP3/BzcdTy+eKKN9d0Bzc8ZNHwHo4YfC6A8A7g0J4CVeg0f1CnEd/GO17ta+7WoQ3dql7IzOZsD0E2p9imECSEVqevh8/k2hyiVy8F4dz84LidBctogyzJs545C5VMPgrk45GVZhiX5H6h8G9y0r7pOE0g2M0RLAQDAdu4Y1L71oA9uh3qTv0fdF75C3Re+AqPSIPC5pdftm7/7e3h2GlPchtNRPKCLYSAL9nuqv12QF3a83tXSsbHvt0YNHw7g3lP8/hTp1fzrblpV19lDWmS+9kgIDaW+aldyti7P4higVPt0OpoQUmHsgvhwXGpOuY1i0gSEQh8ajfRvXgHDslD7B8Mt4lFk/vQ2JEsBABnqWo3g3WtScfvpp2E+8ht8er8EhuXgFTMOmT9OB2QZ6tqNYYy48/TBjswUAIDav3gteGOLR5C+fDI4d194Ro+669p7hvvj0xEti/RqfhiA3+75yZePeIOGb/l0dMMddTy0DaeuO64Va/iSkvvO5kCv4joA4ACIrm6fkWkOOUJIBckrcsS9uuZo5I6kLKVLUdSjzWrL84dHmPRqvjuAA0rXA8Botjl/PXwhv82zKw7o7UL5nKmoqvZMjTHV9dJ3B7Df1W3T6WhCSEVh9Gqu+YlLBUrXoajOTXwxf3hEkV7Nd0XlCGAAMBu1qu6t63v9tWBkKyvzwHdwV237U3M5AK2VaJtCmBBSURpanSKyzfd23bQ6CfDQ4vMxbax6Nf8ogMNK13MDp0HDPx4d7Hv6le41+xrxicuFeotdoBAmhFQrbY5fKqix0zaxDLB4TJsiFcu8D2Cv0vXcgs2o5XuN79zQ3LGxr9K1KCY50wS7ILVVom0KYUJIhbA7xfb7U3PLd2qpKmRSTGOhkZ/hpEbFfaR0LXeQoVfzQxeNam1109TMsbpJGSZo1VwTJdqmECaEVAiLQ4w6fqmwRn7GhNdxxwtdg61uWtUQFM9mVdltZ1lsmhTTuEbO5JFlsgMy1ABqubrtGvkfhBBS8XiOaXgup+jOG1ZD/+vXtEjNsdMBXFS6lrvlplW9/lRUkFjHQ6t0KYo4e8VsA9DM1e1SCBNCKgKnV/O+l/KtStfhcq3reyKirqeF59gvla7lHl0EsPDN3mE1700DcOJSoQYUwoSQaiLQZHPaa+L9py92a1KkVXHvAahyw8J1au79R5rVFut56+68cTVzJsustTrEMFe3SyFMCKkI9TML7TXutpdATx06BPuAY5mvlK7lPhXKsvzd423r1bhR7VfMdtgFsa6r26UQJoRUhIDL+dYaNwVEj3B/2SlKmwFU2YvhejX/5cjI+lWuF/+gcosckGTUcXW7FMKEkIpQ52KeVaN0Ea7Wp3ltk5tWtVbpOh7QUS3PmZoFuCtdh0tdMdvBMnD5zdIUwoSQcucQpHrpBTUrhHUqDi3re2oB/KF0LQ9I5jlmdY+m/i5fzEBJOWYH1Dzr7ep2KYQJIeXOLoheJlvNuqzYIdgHVod4EkCVnyxbq+L2tarvWWVPqd+P3CIHNDznBsCll1EohAkh5U6SZENNGxndPMBd1qv5P5Wuo5ycDKvtXqOu6TtECU5REgF4uLJdCmFCSLmTZOjtQo06m4mQ2m4WNc8mKl1HOUn2Nap1Gr5mRYTZLjgAuPSUdM16hQkhLiEDOkcN6wnX99YLAM4pXUc5cVgcYnqwX82a+lsQZRmAypVtUggTQiqCrqadjvYxaBgAmUrXUV4ESc721Ls0jxQnybIMwKWrWNTMJTMIIRVNU9NOR+s1HAfArHQd5YVjGfaVHiEY9bBN6VJcxlOn0oFCmBBS1TGA1u6sWT1hq0MUvfQwKF1HeXGHWd1OOCTBeVFWuhaXYYaxAK92ZZMUwoSQiiCzbI0aXIsiuyADqDYzXDCy5MDeT1ik7Va6FNdp3NMEtcGlt2bRNWFCSLmTgUKDumZ9x796X7Sb0nWUH0YHp0XpIlyL5RkALp3znEKYEFLuGKDATVuzQrjA6mQAeCldRzlhodLXReElpetwLQphQkh1wLFMrlFTs0L4xOUCg1OQHlK6jnISArsJMGUoXYdrsRyFMCGk6lNxbK6hhoXw8YsFnNkudFW6jnISjXN7la7B9dQGNVw87SiFMCGk3OnUXK6blq85o2oBHDqfB72aaw2AU7qWB2Yr6IHUXdVmpPdd0fsAsuiAi28zoxAmhFSEQg+dyqWn9ZR2xexAttkuAGitdC0PjumMC/8oXYRredQFnLYsVzdLIUwIqQi5fm4ah9JFuNqGw5d1RXbhWaXreEB1wKl8kZWgdB2u5R4AyOIFVzdLIUwIqQipDX0NNWvKLAA/xJ3jOZYZA1ThSTsE2ws4sU6CXLMmW4F7IMCpUlzdLIUwIaQipAR66Vw681BlkF5gQ3xqrgRgpNK13CcNJOklxC7QKl2Iy3nWF6BxP+3qZimECSEVIV2n4niDuuqPUbpXy/ekGgutzilw8eLw5UISx+LyYQ7ZSUpX4npeDa0A6HQ0IaRakKwOMaO+j17pOlxu1+lsXDHb6wAYpnQt90gPwTYLf/yvZq1fWMKrgQQKYUJIdSFKckp976p7afR+yTIwdd0xo8UhfA6g6gSaYP8P0vaocemQ0pUowz2AA4UwIaS60Kq440E1sCcMAPvT8vD7yUx9kV34WOla7lJ7iM5p2PpmzfvWBAAsB2i9dAAuurxpVzdICKkZdGouIbS2Ww1bAeBfb206oXMI0mgAfZSu5Q7qwFG0BevG6ZF7VulalOH/EOAsugzA7uqmKYQJIRXlUKv6XoLSRSil0Cpg3IoDOotDWAOgjdL13IIadtNviP3MDclbla5FOfUiAYb9W4mmKYQJIRXlaF0vnU7D19yPmUPn8/DKj0cMFofwJ4BGStdzE7vpS5zf1wQ7P1IpXYqiGnU1Q+P2lxJN19z/HYSQimYrsgtpzQI8lK5DUb+fysSsXxPdi+zCLgB1lK7nKgZO67uw5A7DT0/rIdeoab5vVr8DC0CRFSsohAkhFUbFsTvbNPCq4Z/wwHf/nGMX/51Sy+IQjgFooXA5GthNPyIv7VUs76mH4+7WK3hmoxW1/s+E5p/fvP3cWDuYmYW4Yrl5lq1z+RLafGlGyyVmNPvcjCUH/p3N1CHKmPCLFSGfmRG20Ix1p4qnG/8szoHmn5vR5wcLHGLxr8+e8wJe3Wa7n+d7ex51AZVOBODy2bIACmFCSAUyaPg/O4f4unRVmspq4Y4zqqlrj/lYHMI+SZZHKVRGPZvDmSCbMgZjaTcDzJl3vePYlipsHXPzaPcLBRL+OCugvkfZc5PUcWMQ+4wBR543Iu5ZAz7aY8dlU3FYf7DLjloGBskvGnFqkgFdgoond1l22IFjLxjQqjaLbWcEyLKM93bZMaOz5j6e8h3UexhwWuMBKPJlkUKYEFKRdrep76Vmqt7cURXil2PpzJDFsfrMAttSs11YDSDARU2zkiw/aXWIJ1fFX6gvuwfyaNDhng7QuQEPb93Nb+R/ttkwp4f2ltODqTkGGr74UbsgQyoVdV8dceK/HYuDlWUY+Or/jSSnCFicgIpj8N0xJ/o05uFVRvsPLKijHTpPxUalUQgTQirSZUlGfqi/m9J1VBoJ6Sb0mL9TvzLu3ECrQzxjc4rvo2In9ehpsjmTEjNMi0Yu/cdt5uYE7u1fUyAPWwF4BT3QgTclORHoxiKi9u2nJ71QIKHFYjPqfWzG1GgNAtxY5NuK03jGDjtaf2HGsJ8syDQX95Bf76BG++VFyLbIiK7HYcVRJya2q6CpyBt2toNhFbkeDFAIE0IqGMtgY/em/jVsSZ7bK3KI+PDXRHXPj3fqdiRm/cfiEM4LkvQaAL9yaoID0LPQ6tydUWBb/8baY437fLrbeORCPgDgu3/OY2tSoSSP+VmGSndfDVicMj7Ybce7MXc+RVzPg8WxF4w485IRK446kGmWIEgyLhYWh+yh54zoUJfD638U36b7RIQah58z4vvBOszf58BLkWr8dkbA0DUW/GerDVJ5DSRTGwHP+loAik0TRiFMCKlQeg2/elCrALouXIaLeVa88MMh/fAv/vH67XjGuzaneCHf4tgO4CkAte/xcGoAnawOcYHFIWSdyTKv++i3xOiOs7cbtp7IuGnjF1YeYTMkL0ke9MV9LTmZkishNU9GxBIzgj4x4WKhjNZfFCHDfOvvWwFuLJrV4rD7vAgfHQO9ChjUlAcADAtX4VD69aVcNknYf1nEgDAV3t9lx+qhOmh44K+z5bRKZsPOgKPoBBSYpKMEr1TDhJAaY1ddLz1Xx0OL9IIKGN1aDRy/VIAXVx3Wu2l49Aj3jxnUKrBd2yAvFQPGYXOKpzUqdr9ezR8DYEZxYHAANIIoNTTZhEiOZZrp1VztC3lW68Yjl/S/Hc/gkjJNd2y39+L9XPwb3WRV+xck5p/F99Qpe8ifQ9Yb/15mCPrEhAMTDNdd1wWAi4USfHQMdCoGeVYZe8+LeLW9GgzD4LEQHn+niejWkMdfqQLC/a7fd8Z2O9672tO2CgDDACxT3AsvF63GFEHn9WX5HOz+MHJNvz+MEFLhTDbn6rm/Jw9bEZtGQ7TuQS03DcJquyGkthtC/d2sBg0vanhWEiSZcQgSm5ZTpEvOMLGns8xIvVIEu3DvZ/3bBXlhzTMtwfwwBDgXe8vtRq6z4O80EVcsMvwNDGZ21WBc63+v05YO4QOXRSw54MCy/jr8kSLgtd9tYJjixS0mR6oxoU3xfufyJTyx3op8mww/A4OvB+hQ36M4iA+ni1gY78DyAcWnyz/5x46lh5yo585g4wj9tcFe901tBN5IsUOlDQSQ82AHu38UwoQQVxh4+HzeikGfx7orXQi52cSuwXijS20wn7cHTOlKl+MaLYYDvWf/DZ1XjJJl0DVhQogr/B4e4K7x1Nfs2RErq8//TsHeC1ZJHr1WAldBo5Arm9ZPmpQ+FQ1QCBNCXMNic0q7eob7K10HuYUxXx1iczV1Zbnv/HIa9VSJGXyBum1VADYpXQqFMCHEJTx0qoXPRDe882ghopg+iw9wYvggVm71RPW+Thk+SIZg3wqgSOlSKIQJIa6ypYG33tEsgC4LV1aZhXaMX5XAoPccBgGtlC6n4rR5ygStxzKlywAohAkhriPyHPvZ2Kgguk+pEtuRlIUlsemQR68F9D5Kl1P+PBsAPo1ZAL8rXQpAIUwIcSE1z37RLyKAcdPQFAWV2eytiTiWLYnyyB8lsLefkrLKaTlKgCSsBeBUuhSAQpgQ4loZTkH6a3DrwOp9zbEaGPLlfs7sEQq55/vVZ6AWrwXaT3RC4zZH6VJKUAgTQlzKXaea+2ynRooPiCG3J0hAvy8OsXLrp1i52aDq8aWp5WgZkGMBJChdSgkKYUKIq/3tpVfndmzsq3Qd5A7O5Vjw8s9JDAYsYlCrqdLlPBiGBbpMsUDr8bbSpZRGIUwIcTXZqOXfnN63KS3qUAX8cjQdqw5ly/Lon2VoPZQu5/6FDwBUuhQAii1bWBYKYUKIElbX9dLldA0tr5X7SEWatuEkk1KkluRhK0QwVXD6b4YBur9thtZjmtKl3IhCmBCiBMlNq3rtf33DqTdcRfRfsp+z1W7DyF3erHprQ4cPBPTe5wD8qnQpN6IQJoQoZb2/uyajVzOayrIqsDgkDFp2hJWjXmLR5BGly7l7DAv0fLcIWo/XAFS6AWYUwoQQpUhuWtWr0/o0LaqKZzhrosR0E/63OQXy0K8A70ZKl3N3mg8BtB4pqCSTc9yIQpgQoqTN3gb1uX4P1VG6DnKXVsafx5aEAlke87MMlV7pcm5PpQce+cACrcd/UAl7wQCFMCFEWbKbVjXpnf7NLDpVNZuZqRqb/ONR5rLoIcmDv6zcE3nETHdApdsGYLvSpdwKhTAhRGl/q3l22ys9mjiULoTcvd6L9nPOBl1ZucPkyjlQq3YLoO0zdmjdn1e6lNuhECaEKM5Nq5r0ZIcgZ7CfUelSyF0qtAkYueI4g5jpLII6Kl3O9RgWGLy0CLz6FQBZSpdzOxTChJDKIJ3nmOnzH4+gQVpVyMFzeZj91znII1YC7gFKl/Ovh58X4R5wCiz/tdKl3AmFMCGkUlBx7MJGfoazoyLrV87Tm6RMS3aexZ40iySPXieBUytdDuBRF+j2Pwe07mNQSQdjlUYhTAipLEQ3rWrktD5N7bXdtUrXQu7BE98cYnPUAbL82KfKD9QasMgChpsNIFnpUu4GhTAhpDI5ybHM/y0a3bqIpdPSVUrvRfs5Maw/K7ceq1zvM3wgENgmCyrtLMVquEcUwoSQSkWr4t4N8TeeeKVHSKVYdJ3cnWyzA+N+OMng0VkMAlu7vgBjLeCxTy3QuI0GUGVG2lMIE0IqG9FNqxr4bKeGRVHBPkrXQu7BztNXsHjvZcij1gIGFy5VyamA0WuLwGvmA4h1XcMPjkKYEFIZZejV/LDFY9pY/YwapWsh92DOtiQcyRJFeeRqCayLJmDp+7EN3o32QqWrVGsF3w0KYUJIZfWnmmc/WfwEXR+uaoYt3c+Z3JtA7vVhxQ/UajNWQrNBmdC4DQNQ5UbWUwgTQiotnYqbEebvfoquD1ctggT0XXyQlVo+wcrNh1TcQK16kUCvWRZojI8AKKywdioQhTAhpDITjVp+wLOdGhYMaBlQ6e/5JP+6kGfFi2sTGfT/jEGt8PJvwK02MGqNFWr9CFSR25HKQiFMCKns0vVqvuuswQ+ZaaBW1fLr8QysPJhdvOKS1rP8DsypgdHrisDrPgKwpfwO7HoUwoSQquCkXs0/tvTJttaw2m5K10LuwfSNJ5nTJpUkP/6dhPKak/SxBTZ4NdgJlfa98jmgciiECSFVxU6dint65fj2lgAPmlGrKum/eD9nrdUSctfpDz5wqsdMB5o+dgEat+GoAtNS3gmFMCGkymBZZrVBw73z43Mditx1vNLlkLtkEyQMXnqElTtMZBHS6/4P1PNdB9o9ex4aYxQAc7kVqCAKYUJIlaLhuf/zMai/+XF8+yJ3LQVxVZGYacK0X1IgD/kK8G507wd45H0H2j6TBo2xPYAr5V6gQiiECSFVjkHDvxjkY1ixfmJ0kbehEqzcQ+7Kj/svYPOpPFl+Yr0MteHud+z1oQNtxqZC4xYFIKfCClQAhTAhpCqS9Rp+coCnbuHGSdEWf3eaVauqeHH1MeaS012SBy+7u4k8Hv3IjtZPnq2OAQxQCBNCqi5Zp+be9HPTfLhpckdLXS+d0vWQu9T78/2co34nRo56+fYDtXrPsaPVmJIAznVNda5FIUwIqdK0Ku4DL736zY2Toi2NfO/hFCdRjMkmYMTXR1l0fZNFw843b8AwQJ+5drQclXI1gPNcXqSLUAgTQqo8Nc9+5q5TTd4wKdrapoGX0uWQu3D4QgE+/CMN8vDvAY+6/z6g0gMjVlkQMSIBGrdoAPmKFekCjCxX+dusCCGkRG+LQ1jz9qaT+p8OXKRORhWw4qlWUmefQjBfdmZhqAU8sb4IbrV/gcZtLAC70vVVNAphQkh107TILvy55sAFn/e3JGhEiT7jKrv9U6JE39zDDFO3jR28ZgZ47XxUg4k47gZ9UySEVDcJBg3/0LA2dePXPt+hiNYjrvxWHsqGVO9hEVqPgeC181BDAhignjAhpPrirA7xXZsg/mfCtwd0+9Oq7dieKsuo4fHx8JaWDsE+540avg+AVKVrcjUKYUJIddfH6hBX/hB3Tvd/25LUdqHKrfteLUU29MZnI1tZjBp+jUHDvwDApnRNJ7pgFQAABLpJREFUSqAQJoTUBH4mm/PrAquz66QfDhmOXixQup4aS6/mMK1PU/vg1oEWvZp/GsBGpWtSEoUwIaSmYAA8bnWIS1fEpmnm/5GsdojUK3alDsE++HRES4texW82avkXUE0n4LgXFMKEkJqmtsnm/Da3yBE1eeVhw/FL1CuuaAY1hxn9wm39WwZY9Gr+SQBblK6psqAQJoTURIwky6NtTnHhH6cy1R/+mqDLLKz2t6S6HMMAA1sGyjP6hVvVPLvBqOEnoZpPvnGvKIQJITWZu8UhzGDATFq+5yz/+d8pKovj7tYVILfXsbEv3h3QrMjPTZPqplW9AGCP0jVVRhTChBAC1DfZnB9LMnp/9FuCdvX+CwzN8XF/wuu44+3+4UXNAjwKjBr+JQA/owbd93uvKIQJIeRf7Uw255LcIkfIvN+Tjb8eT4dAaXxXgnz0eLVnqLVnuL9TxTHTeI79EoBT6boqOwphQgi5HgOgd4HVOVOU5PDP/z6j/TH+Amu2C0rXVSl1aOSDiTHBRe2CvGVZxgKdmpsNoFDpuqoKCmFCCLm1toU25wyeZR5ZGXeeXb4nVZ1eUCPnlLiOimPwWEQAJsc0Nvu5afINGv59lmG+A2BRuraqhkKYEELurEGRXXiDY5ln/k7KllfGndPvOXMFNe1MdYCHFkPb1hPHdWzoAHDUQ6d6F8A2AHTD9X2iECaEkLvnKcnyEyabMEmW5XprDlxUrT98UZWQblK6rgrjruXxaPM6GP1wfVNobTdOEOWfjFp+HoDjStdWHVAIE0LI/Qm3OsRnJFl+ssDq1K07dFG79UQGfyq9EFX9Y9VDp8Ij4f4Y0qauqVV9T7XNKe3y0KkWA/gVNWCNX1eiECaEkAfDAGhndYhjBEkayjCM1z8pOcL2xCzj3pQrOJdT+S+TGtQc2jX0RsfGvs6uobWs9b31GotD+NtTr/4GxbNbVd+uvsIohAkhpHzVA9C9wOp8TMUx3awOUbUzOZvZe+aKPjHDhDNZZii9klMtNw2a1nFHh2AfISa0lqWhr0FrcQgnDRr+FxXH/gUgDtTjdQkKYUIIqTgMgDAA3fMtjkcYhmlhUHOBV8wOa0J6IY5dzDckZJjYM1lmZJlsKLSW321QLAPUctOiib8RTWoZ0SzAwxoe4O4I8jFoAThsgnjaqOZ/U/HsXwD+AWAtt8bJXaMQJoQQ11IDaALgIacotTDbhA48x4RqeM6LZaEy2QRbXpFDyDbZkWmy8xkFNo3F8f/t3LFOwzAUhtHrNGmCWkCVoH3/F+rIxBuAWGAownEYCLzCrcQ50i95tLx8m2tf2xJ1blHbEptSoislSonYj3073U2Xx9uxPuzH5bAbuvtpGKZhs/2s7ePyNT9v++68G/tzRDyte0l9Af6IMMD1uImI47rTumNETHVu23lZxtZiKCVqKVG7UuZh073HT1Rf1/2e3yLCR9hXToQBIEmXfQEA+K9EGACSiDAAJBFhAEgiwgCQRIQBIIkIA0ASEQaAJCIMAElEGACSiDAAJBFhAEgiwgCQRIQBIIkIA0ASEQaAJCIMAElEGACSiDAAJBFhAEgiwgCQRIQBIIkIA0ASEQaAJCIMAElEGACSiDAAJPkGhFW6YIqS5CkAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Running GridSearchCV for KerasClassifier.\n",
      "Fitting 3 folds for each of 96 candidates, totalling 288 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:  2.6min\n",
      "[Parallel(n_jobs=-1)]: Done 184 tasks      | elapsed: 21.4min\n",
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\joblib\\externals\\loky\\process_executor.py:706: UserWarning: A worker stopped while some jobs were given to the executor. This can be caused by a too short worker timeout or by a memory leak.\n",
      "  \"timeout or by a memory leak.\", UserWarning\n",
      "[Parallel(n_jobs=-1)]: Done 288 out of 288 | elapsed: 25.5min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KerasClassifier\n",
      "          estimator min_score mean_score max_score   std_score  \\\n",
      "0   KerasClassifier  0.729568  0.767888   0.823372  0.0401727    \n",
      "1   KerasClassifier  0.711083  0.759428   0.799853  0.0366701    \n",
      "2   KerasClassifier  0.723602  0.758153   0.802384  0.0328827    \n",
      "3   KerasClassifier  0.740111  0.756638   0.769957  0.0123936    \n",
      "4   KerasClassifier  0.713767  0.75094    0.813934  0.0447837    \n",
      "5   KerasClassifier  0.720611  0.747617   0.761533  0.0190986    \n",
      "6   KerasClassifier  0.709601  0.74715    0.790193  0.0331298    \n",
      "7   KerasClassifier  0.711168  0.744035   0.783422  0.0298556    \n",
      "8   KerasClassifier  0.708643  0.74232    0.802149  0.0424167    \n",
      "9   KerasClassifier  0.713373  0.740036   0.764327  0.0208692    \n",
      "10  KerasClassifier  0.715432  0.739695   0.780584  0.0290807    \n",
      "11  KerasClassifier  0.706017  0.738539   0.796096  0.0408133    \n",
      "12  KerasClassifier  0.717603  0.735187   0.753635  0.0147227    \n",
      "13  KerasClassifier  0.717696  0.730928   0.75733   0.0186696    \n",
      "14  KerasClassifier  0.678253  0.727152   0.790792  0.0471114    \n",
      "15  KerasClassifier  0.699107  0.725325   0.756782  0.0238353    \n",
      "16  KerasClassifier  0.709999  0.719236   0.733204  0.0100467    \n",
      "17  KerasClassifier  0.680095  0.714943   0.75341   0.0300401    \n",
      "18  KerasClassifier  0.688501  0.707824   0.722047  0.0141621    \n",
      "19  KerasClassifier  0.694585  0.706792   0.713836  0.00866617   \n",
      "20  KerasClassifier  0.678323  0.705728   0.724273  0.0197777    \n",
      "21  KerasClassifier  0.655792  0.703054   0.732309  0.0337329    \n",
      "22  KerasClassifier  0.60075   0.702394   0.770743  0.0732841    \n",
      "23  KerasClassifier  0.685916  0.696211   0.715757  0.0138275    \n",
      "24  KerasClassifier  0.646875  0.69608    0.723383  0.034863     \n",
      "25  KerasClassifier  0.669639  0.69393    0.708359  0.0172774    \n",
      "26  KerasClassifier  0.662191  0.693603   0.731009  0.028413     \n",
      "27  KerasClassifier  0.681056  0.68712    0.697616  0.00745167   \n",
      "28  KerasClassifier  0.672318  0.681621   0.691919  0.00803319   \n",
      "29  KerasClassifier  0.670231  0.68104    0.690901  0.00846501   \n",
      "30  KerasClassifier  0.660827  0.680846   0.713266  0.0231343    \n",
      "31  KerasClassifier  0.659068  0.675592   0.68959   0.012588     \n",
      "32  KerasClassifier  0.630444  0.674152   0.712346  0.0336628    \n",
      "33  KerasClassifier  0.623718  0.661435   0.720381  0.0422211    \n",
      "34  KerasClassifier  0.626947  0.656005   0.683487  0.0231093    \n",
      "35  KerasClassifier  0.639167  0.654034   0.671477  0.0133156    \n",
      "36  KerasClassifier  0.594666  0.648829   0.687682  0.0394867    \n",
      "37  KerasClassifier  0.619968  0.632912   0.650708  0.0130099    \n",
      "38  KerasClassifier  0.59991   0.625795   0.641019  0.0183982    \n",
      "39  KerasClassifier  0.608211  0.618735   0.631977  0.00989093   \n",
      "40  KerasClassifier  0.591843  0.617323   0.632454  0.0181224    \n",
      "41  KerasClassifier  0.575359  0.613964   0.634776  0.0273259    \n",
      "42  KerasClassifier  0.58376   0.606219   0.640425  0.0245796    \n",
      "43  KerasClassifier  0.56793   0.591148   0.607356  0.0168417    \n",
      "44  KerasClassifier  0.57096   0.583801   0.599827  0.0119984    \n",
      "45  KerasClassifier  0.543137  0.577086   0.617361  0.0306305    \n",
      "46  KerasClassifier  0.549695  0.555164   0.564163  0.00641267   \n",
      "47  KerasClassifier  0.483342  0.54863    0.664971  0.0824705    \n",
      "48  KerasClassifier  0.47184   0.536059   0.574132  0.0456705    \n",
      "49  KerasClassifier  0.507303  0.520444   0.536172  0.0119271    \n",
      "50  KerasClassifier  0.484833  0.512942   0.540867  0.0228759    \n",
      "51  KerasClassifier  0.444737  0.508692   0.578135  0.0545979    \n",
      "52  KerasClassifier  0.418552  0.499036   0.62834   0.0923411    \n",
      "53  KerasClassifier  0.475456  0.492024   0.513207  0.0157533    \n",
      "54  KerasClassifier  0.424048  0.489859   0.529218  0.0468326    \n",
      "55  KerasClassifier  0.427347  0.484655   0.585273  0.0713769    \n",
      "56  KerasClassifier  0.453752  0.4762     0.517271  0.0290835    \n",
      "57  KerasClassifier  0.431816  0.462373   0.496548  0.0265505    \n",
      "58  KerasClassifier  0.406917  0.459995   0.500766  0.0392896    \n",
      "59  KerasClassifier  0.38489   0.454986   0.497651  0.0499537    \n",
      "60  KerasClassifier  0.418343  0.452848   0.496199  0.0323941    \n",
      "61  KerasClassifier  0.366607  0.437407   0.500779  0.0550268    \n",
      "62  KerasClassifier  0.418993  0.428011   0.436349  0.00710166   \n",
      "63  KerasClassifier  0.320162  0.409572   0.513705  0.0796963    \n",
      "64  KerasClassifier  0.342952  0.403866   0.444949  0.0439375    \n",
      "65  KerasClassifier  0.380738  0.394822   0.412383  0.0131506    \n",
      "66  KerasClassifier  0.277604  0.38474    0.589439  0.144797     \n",
      "67  KerasClassifier  0.28271   0.382976   0.442141  0.0712817    \n",
      "68  KerasClassifier  0.347797  0.379213   0.415134  0.0276741    \n",
      "69  KerasClassifier  0.340074  0.374849   0.430865  0.0399927    \n",
      "70  KerasClassifier  0.301149  0.369955   0.430976  0.0532868    \n",
      "71  KerasClassifier  0.338707  0.344961   0.352189  0.00554701   \n",
      "72  KerasClassifier  0.164322  0.252048   0.305625  0.0625365    \n",
      "73  KerasClassifier  0.186102  0.23874    0.273396  0.0378378    \n",
      "74  KerasClassifier  0.196617  0.234188   0.258593  0.0269599    \n",
      "75  KerasClassifier  0.165258  0.208695   0.257876  0.0380287    \n",
      "76  KerasClassifier  0.161364  0.194025   0.244808  0.036396     \n",
      "77  KerasClassifier  0.16805   0.189135   0.2139    0.0188986    \n",
      "78  KerasClassifier  0.16535   0.182183   0.215021  0.0232225    \n",
      "79  KerasClassifier  0.164149  0.179712   0.189853  0.011172     \n",
      "80  KerasClassifier  0.16263   0.177992   0.200394  0.0162006    \n",
      "81  KerasClassifier  0.153744  0.176376   0.191241  0.0162634    \n",
      "82  KerasClassifier  0.14645   0.175753   0.200512  0.0223034    \n",
      "83  KerasClassifier  0.158059  0.170982   0.187794  0.0124472    \n",
      "84  KerasClassifier  0.134296  0.170489   0.189035  0.0255949    \n",
      "85  KerasClassifier  0.156314  0.169929   0.186934  0.0127282    \n",
      "86  KerasClassifier  0.135951  0.167576   0.186274  0.0224863    \n",
      "87  KerasClassifier  0.145235  0.166644   0.180404  0.0153426    \n",
      "88  KerasClassifier  0.137722  0.165694   0.186681  0.0205885    \n",
      "89  KerasClassifier  0.147095  0.16517    0.194809  0.0211258    \n",
      "90  KerasClassifier  0.11696   0.164426   0.194809  0.0340002    \n",
      "91  KerasClassifier  0.148417  0.164275   0.183057  0.0142922    \n",
      "92  KerasClassifier  0.124237  0.157903   0.180895  0.024331     \n",
      "93  KerasClassifier  0.147975  0.150147   0.153742  0.00256049   \n",
      "94  KerasClassifier  0.141743  0.14623    0.150036  0.00341982   \n",
      "95  KerasClassifier  0.113372  0.13814    0.170533  0.0239507    \n",
      "\n",
      "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              Vectorizer  \\\n",
      "0   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "1   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "2   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "3   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "4   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "5   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "6   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "7   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "8   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "9   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "10  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "11  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "12  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "13  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "14  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "15  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "16  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "17  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "18  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "19  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "20  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "21  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "22  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "23  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "24  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "25  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "26  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "27  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "28  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "29  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "30  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "31  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "32  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "33  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "34  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "35  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "36  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "37  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "38  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "39  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "40  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "41  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "42  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "43  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "44  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "45  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "46  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "47  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "48  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "49  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "50  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "51  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "52  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "53  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "54  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "55  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "56  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "57  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "58  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "59  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "60  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "61  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "62  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "63  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "64  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "65  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "66  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "67  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "68  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "69  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "70  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "71  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "72  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "73  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "74  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "75  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "76  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "77  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "78  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "79  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "80  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "81  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "82  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "83  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "84  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "85  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "86  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "87  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "88  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "89  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "90  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "91  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "92  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "93  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "94  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "95  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "\n",
      "   Vectorizer__ngram_range  \\\n",
      "0   (1, 3)                   \n",
      "1   (1, 3)                   \n",
      "2   (1, 3)                   \n",
      "3   (1, 3)                   \n",
      "4   (1, 1)                   \n",
      "5   (1, 1)                   \n",
      "6   (1, 1)                   \n",
      "7   (1, 3)                   \n",
      "8   (1, 1)                   \n",
      "9   (1, 3)                   \n",
      "10  (1, 3)                   \n",
      "11  (1, 1)                   \n",
      "12  (1, 1)                   \n",
      "13  (1, 3)                   \n",
      "14  (1, 1)                   \n",
      "15  (1, 1)                   \n",
      "16  (1, 1)                   \n",
      "17  (1, 3)                   \n",
      "18  (1, 1)                   \n",
      "19  (1, 1)                   \n",
      "20  (1, 3)                   \n",
      "21  (1, 3)                   \n",
      "22  (1, 3)                   \n",
      "23  (1, 3)                   \n",
      "24  (1, 3)                   \n",
      "25  (1, 3)                   \n",
      "26  (1, 3)                   \n",
      "27  (1, 3)                   \n",
      "28  (1, 3)                   \n",
      "29  (1, 3)                   \n",
      "30  (1, 1)                   \n",
      "31  (1, 1)                   \n",
      "32  (1, 1)                   \n",
      "33  (1, 3)                   \n",
      "34  (1, 1)                   \n",
      "35  (1, 1)                   \n",
      "36  (1, 1)                   \n",
      "37  (1, 3)                   \n",
      "38  (1, 1)                   \n",
      "39  (1, 1)                   \n",
      "40  (1, 1)                   \n",
      "41  (1, 1)                   \n",
      "42  (1, 1)                   \n",
      "43  (1, 1)                   \n",
      "44  (1, 1)                   \n",
      "45  (1, 3)                   \n",
      "46  (1, 3)                   \n",
      "47  (1, 3)                   \n",
      "48  (1, 3)                   \n",
      "49  (1, 3)                   \n",
      "50  (1, 3)                   \n",
      "51  (1, 3)                   \n",
      "52  (1, 3)                   \n",
      "53  (1, 3)                   \n",
      "54  (1, 1)                   \n",
      "55  (1, 3)                   \n",
      "56  (1, 3)                   \n",
      "57  (1, 3)                   \n",
      "58  (1, 1)                   \n",
      "59  (1, 3)                   \n",
      "60  (1, 3)                   \n",
      "61  (1, 1)                   \n",
      "62  (1, 1)                   \n",
      "63  (1, 1)                   \n",
      "64  (1, 1)                   \n",
      "65  (1, 1)                   \n",
      "66  (1, 1)                   \n",
      "67  (1, 1)                   \n",
      "68  (1, 3)                   \n",
      "69  (1, 1)                   \n",
      "70  (1, 1)                   \n",
      "71  (1, 1)                   \n",
      "72  (1, 3)                   \n",
      "73  (1, 3)                   \n",
      "74  (1, 3)                   \n",
      "75  (1, 3)                   \n",
      "76  (1, 3)                   \n",
      "77  (1, 3)                   \n",
      "78  (1, 1)                   \n",
      "79  (1, 1)                   \n",
      "80  (1, 1)                   \n",
      "81  (1, 1)                   \n",
      "82  (1, 1)                   \n",
      "83  (1, 3)                   \n",
      "84  (1, 3)                   \n",
      "85  (1, 1)                   \n",
      "86  (1, 3)                   \n",
      "87  (1, 1)                   \n",
      "88  (1, 1)                   \n",
      "89  (1, 3)                   \n",
      "90  (1, 1)                   \n",
      "91  (1, 3)                   \n",
      "92  (1, 1)                   \n",
      "93  (1, 1)                   \n",
      "94  (1, 1)                   \n",
      "95  (1, 3)                   \n",
      "\n",
      "                                    Vectorizer__tokenizer clf__Dropout  \\\n",
      "0   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "1   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "2   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "3   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "4   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "5   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "6   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "7   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "8   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "9   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "10  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "11  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "12  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "13  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "14  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "15  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "16  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "17  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "18  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "19  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "20  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "21  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "22  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "23  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "24  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "25  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "26  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "27  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "28  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "29  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "30  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "31  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "32  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "33  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "34  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "35  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "36  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "37  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "38  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "39  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "40  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "41  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "42  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "43  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "44  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "45  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "46  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "47  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "48  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "49  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "50  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "51  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "52  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "53  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "54  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "55  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "56  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "57  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "58  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "59  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "60  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "61  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "62  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "63  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "64  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "65  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "66  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "67  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "68  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "69  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "70  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "71  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "72  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "73  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "74  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "75  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "76  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "77  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "78  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "79  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "80  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "81  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "82  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "83  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "84  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "85  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "86  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "87  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "88  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "89  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "90  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "91  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "92  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "93  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "94  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "95  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "\n",
      "   clf__epochs clf__n_classes clf__n_layers clf__units  \\\n",
      "0   5           7              2             500         \n",
      "1   5           7              2             1000        \n",
      "2   5           7              2             500         \n",
      "3   5           7              2             1000        \n",
      "4   5           7              2             500         \n",
      "5   5           7              2             500         \n",
      "6   5           7              2             1000        \n",
      "7   5           7              2             500         \n",
      "8   5           7              2             1000        \n",
      "9   5           7              2             1000        \n",
      "10  5           7              2             500         \n",
      "11  5           7              2             1000        \n",
      "12  5           7              2             1000        \n",
      "13  5           7              2             1000        \n",
      "14  5           7              2             1000        \n",
      "15  5           7              2             1000        \n",
      "16  5           7              2             500         \n",
      "17  5           7              2             1000        \n",
      "18  5           7              2             500         \n",
      "19  5           7              2             500         \n",
      "20  5           7              3             1000        \n",
      "21  5           7              2             500         \n",
      "22  5           7              2             500         \n",
      "23  5           7              3             500         \n",
      "24  5           7              2             1000        \n",
      "25  5           7              3             500         \n",
      "26  5           7              3             500         \n",
      "27  5           7              3             1000        \n",
      "28  5           7              3             500         \n",
      "29  5           7              3             1000        \n",
      "30  5           7              2             500         \n",
      "31  5           7              3             1000        \n",
      "32  5           7              3             500         \n",
      "33  5           7              3             1000        \n",
      "34  5           7              3             1000        \n",
      "35  5           7              3             500         \n",
      "36  5           7              3             1000        \n",
      "37  5           7              2             1000        \n",
      "38  5           7              3             1000        \n",
      "39  5           7              3             1000        \n",
      "40  5           7              3             500         \n",
      "41  5           7              3             500         \n",
      "42  5           7              3             1000        \n",
      "43  5           7              3             500         \n",
      "44  5           7              3             500         \n",
      "45  5           7              2             1000        \n",
      "46  5           7              2             1000        \n",
      "47  5           7              2             500         \n",
      "48  5           7              2             500         \n",
      "49  5           7              2             1000        \n",
      "50  5           7              2             1000        \n",
      "51  5           7              3             500         \n",
      "52  5           7              3             1000        \n",
      "53  5           7              2             500         \n",
      "54  5           7              2             1000        \n",
      "55  5           7              3             500         \n",
      "56  5           7              3             1000        \n",
      "57  5           7              2             500         \n",
      "58  5           7              2             1000        \n",
      "59  5           7              2             500         \n",
      "60  5           7              2             1000        \n",
      "61  5           7              2             1000        \n",
      "62  5           7              2             500         \n",
      "63  5           7              2             500         \n",
      "64  5           7              2             1000        \n",
      "65  5           7              2             1000        \n",
      "66  5           7              2             500         \n",
      "67  5           7              2             1000        \n",
      "68  5           7              2             500         \n",
      "69  5           7              2             500         \n",
      "70  5           7              2             500         \n",
      "71  5           7              2             500         \n",
      "72  5           7              3             1000        \n",
      "73  5           7              3             500         \n",
      "74  5           7              3             1000        \n",
      "75  5           7              3             500         \n",
      "76  5           7              3             1000        \n",
      "77  5           7              3             500         \n",
      "78  5           7              3             1000        \n",
      "79  5           7              3             500         \n",
      "80  5           7              3             500         \n",
      "81  5           7              3             1000        \n",
      "82  5           7              3             1000        \n",
      "83  5           7              3             500         \n",
      "84  5           7              3             1000        \n",
      "85  5           7              3             1000        \n",
      "86  5           7              3             1000        \n",
      "87  5           7              3             500         \n",
      "88  5           7              3             1000        \n",
      "89  5           7              3             1000        \n",
      "90  5           7              3             1000        \n",
      "91  5           7              3             500         \n",
      "92  5           7              3             500         \n",
      "93  5           7              3             500         \n",
      "94  5           7              3             500         \n",
      "95  5           7              3             500         \n",
      "\n",
      "   feature_selection__threshold  \n",
      "0   0.2                          \n",
      "1   0.2                          \n",
      "2   0.2                          \n",
      "3   0.2                          \n",
      "4   0.5                          \n",
      "5   0.2                          \n",
      "6   0.5                          \n",
      "7   0.1                          \n",
      "8   0.1                          \n",
      "9   0.1                          \n",
      "10  0.1                          \n",
      "11  0.2                          \n",
      "12  0.5                          \n",
      "13  0.1                          \n",
      "14  0.1                          \n",
      "15  0.2                          \n",
      "16  0.2                          \n",
      "17  0.5                          \n",
      "18  0.5                          \n",
      "19  0.1                          \n",
      "20  0.2                          \n",
      "21  0.5                          \n",
      "22  0.5                          \n",
      "23  0.2                          \n",
      "24  0.5                          \n",
      "25  0.1                          \n",
      "26  0.2                          \n",
      "27  0.1                          \n",
      "28  0.1                          \n",
      "29  0.1                          \n",
      "30  0.1                          \n",
      "31  0.5                          \n",
      "32  0.5                          \n",
      "33  0.2                          \n",
      "34  0.2                          \n",
      "35  0.2                          \n",
      "36  0.5                          \n",
      "37  0.5                          \n",
      "38  0.1                          \n",
      "39  0.2                          \n",
      "40  0.1                          \n",
      "41  0.5                          \n",
      "42  0.1                          \n",
      "43  0.1                          \n",
      "44  0.2                          \n",
      "45  0.2                          \n",
      "46  0.5                          \n",
      "47  0.5                          \n",
      "48  0.2                          \n",
      "49  0.2                          \n",
      "50  0.1                          \n",
      "51  0.5                          \n",
      "52  0.5                          \n",
      "53  0.5                          \n",
      "54  0.5                          \n",
      "55  0.5                          \n",
      "56  0.5                          \n",
      "57  0.2                          \n",
      "58  0.5                          \n",
      "59  0.1                          \n",
      "60  0.1                          \n",
      "61  0.2                          \n",
      "62  0.2                          \n",
      "63  0.5                          \n",
      "64  0.2                          \n",
      "65  0.1                          \n",
      "66  0.2                          \n",
      "67  0.1                          \n",
      "68  0.1                          \n",
      "69  0.1                          \n",
      "70  0.5                          \n",
      "71  0.1                          \n",
      "72  0.5                          \n",
      "73  0.5                          \n",
      "74  0.5                          \n",
      "75  0.5                          \n",
      "76  0.1                          \n",
      "77  0.1                          \n",
      "78  0.5                          \n",
      "79  0.5                          \n",
      "80  0.1                          \n",
      "81  0.2                          \n",
      "82  0.1                          \n",
      "83  0.2                          \n",
      "84  0.1                          \n",
      "85  0.5                          \n",
      "86  0.2                          \n",
      "87  0.1                          \n",
      "88  0.2                          \n",
      "89  0.2                          \n",
      "90  0.1                          \n",
      "91  0.2                          \n",
      "92  0.2                          \n",
      "93  0.5                          \n",
      "94  0.2                          \n",
      "95  0.1                          \n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "________________________________________________________________________\n",
      "Running GridSearchCV for KerasClassifier.\n",
      "Fitting 3 folds for each of 96 candidates, totalling 288 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\joblib\\externals\\loky\\process_executor.py:706: UserWarning: A worker stopped while some jobs were given to the executor. This can be caused by a too short worker timeout or by a memory leak.\n",
      "  \"timeout or by a memory leak.\", UserWarning\n",
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:  1.6min\n",
      "[Parallel(n_jobs=-1)]: Done 184 tasks      | elapsed:  8.4min\n",
      "[Parallel(n_jobs=-1)]: Done 288 out of 288 | elapsed: 11.6min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KerasClassifier\n",
      "          estimator min_score mean_score max_score   std_score  \\\n",
      "0   KerasClassifier  0.79723   0.826307   0.84281   0.0206233    \n",
      "1   KerasClassifier  0.771555  0.805384   0.838992  0.0275313    \n",
      "2   KerasClassifier  0.748641  0.787086   0.816258  0.0283728    \n",
      "3   KerasClassifier  0.752167  0.784266   0.823064  0.029329     \n",
      "4   KerasClassifier  0.711212  0.781563   0.82033   0.0498318    \n",
      "5   KerasClassifier  0.741909  0.773901   0.825407  0.0367753    \n",
      "6   KerasClassifier  0.745009  0.771084   0.79261   0.0196972    \n",
      "7   KerasClassifier  0.723101  0.7503     0.775632  0.0214863    \n",
      "8   KerasClassifier  0.73416   0.745092   0.760888  0.011441     \n",
      "9   KerasClassifier  0.69644   0.739778   0.774253  0.0323795    \n",
      "10  KerasClassifier  0.688804  0.736822   0.770581  0.0348745    \n",
      "11  KerasClassifier  0.703836  0.729421   0.762676  0.0246261    \n",
      "12  KerasClassifier  0.696026  0.718749   0.744245  0.0197829    \n",
      "13  KerasClassifier  0.631478  0.712942   0.790795  0.065091     \n",
      "14  KerasClassifier  0.599582  0.708079   0.770295  0.0769941    \n",
      "15  KerasClassifier  0.657307  0.701578   0.7511    0.0384701    \n",
      "16  KerasClassifier  0.629372  0.696636   0.743293  0.0487371    \n",
      "17  KerasClassifier  0.64053   0.695602   0.733686  0.0398829    \n",
      "18  KerasClassifier  0.662472  0.694164   0.735465  0.0305641    \n",
      "19  KerasClassifier  0.647349  0.686575   0.714277  0.0285125    \n",
      "20  KerasClassifier  0.597597  0.683241   0.786043  0.0778835    \n",
      "21  KerasClassifier  0.634488  0.679308   0.739879  0.0444439    \n",
      "22  KerasClassifier  0.597673  0.673403   0.7511    0.0626516    \n",
      "23  KerasClassifier  0.595422  0.673131   0.727946  0.0564729    \n",
      "24  KerasClassifier  0.588959  0.665723   0.731731  0.0587808    \n",
      "25  KerasClassifier  0.613871  0.664941   0.71089   0.0397731    \n",
      "26  KerasClassifier  0.601028  0.660045   0.725778  0.0511499    \n",
      "27  KerasClassifier  0.63206   0.657798   0.671109  0.0182029    \n",
      "28  KerasClassifier  0.606958  0.649771   0.684297  0.0321125    \n",
      "29  KerasClassifier  0.589751  0.64921    0.71931   0.053425     \n",
      "30  KerasClassifier  0.589751  0.63604    0.675291  0.0352745    \n",
      "31  KerasClassifier  0.536276  0.634344   0.703201  0.0712088    \n",
      "32  KerasClassifier  0.605413  0.630312   0.672277  0.0298452    \n",
      "33  KerasClassifier  0.552503  0.620663   0.715106  0.0689346    \n",
      "34  KerasClassifier  0.554964  0.619434   0.702511  0.061656     \n",
      "35  KerasClassifier  0.509401  0.617665   0.802116  0.131082     \n",
      "36  KerasClassifier  0.577886  0.612578   0.63463   0.0248297    \n",
      "37  KerasClassifier  0.56352   0.612344   0.659804  0.0393199    \n",
      "38  KerasClassifier  0.58297   0.611733   0.629656  0.0205433    \n",
      "39  KerasClassifier  0.575115  0.604129   0.623402  0.0208817    \n",
      "40  KerasClassifier  0.464804  0.60375    0.837046  0.165966     \n",
      "41  KerasClassifier  0.53094   0.601943   0.645517  0.0506375    \n",
      "42  KerasClassifier  0.568468  0.601247   0.661955  0.0429725    \n",
      "43  KerasClassifier  0.541032  0.59642    0.705943  0.0774464    \n",
      "44  KerasClassifier  0.472393  0.593747   0.771202  0.128277     \n",
      "45  KerasClassifier  0.536276  0.592612   0.674223  0.0590847    \n",
      "46  KerasClassifier  0.478791  0.592579   0.751701  0.115935     \n",
      "47  KerasClassifier  0.512147  0.58713    0.634187  0.0535934    \n",
      "48  KerasClassifier  0.549065  0.586017   0.629762  0.0332927    \n",
      "49  KerasClassifier  0.465467  0.584106   0.684684  0.0904015    \n",
      "50  KerasClassifier  0.519261  0.580026   0.671752  0.0659917    \n",
      "51  KerasClassifier  0.510027  0.569263   0.617177  0.0444704    \n",
      "52  KerasClassifier  0.48598   0.559401   0.669595  0.0793422    \n",
      "53  KerasClassifier  0.518488  0.55891    0.604269  0.0351934    \n",
      "54  KerasClassifier  0.495291  0.553513   0.666583  0.0799642    \n",
      "55  KerasClassifier  0.492645  0.547442   0.594526  0.0419486    \n",
      "56  KerasClassifier  0.472393  0.543054   0.667365  0.0881751    \n",
      "57  KerasClassifier  0.528375  0.540348   0.56173   0.0151555    \n",
      "58  KerasClassifier  0.529115  0.5401     0.56173   0.0152952    \n",
      "59  KerasClassifier  0.47923   0.539378   0.603385  0.0507595    \n",
      "60  KerasClassifier  0.525851  0.535816   0.553221  0.0123508    \n",
      "61  KerasClassifier  0.492645  0.535171   0.616253  0.0573565    \n",
      "62  KerasClassifier  0.496369  0.528384   0.554964  0.0242284    \n",
      "63  KerasClassifier  0.496615  0.527146   0.554964  0.0238983    \n",
      "64  KerasClassifier  0.492645  0.52684    0.561283  0.0280218    \n",
      "65  KerasClassifier  0.509401  0.526383   0.553221  0.0191995    \n",
      "66  KerasClassifier  0.509401  0.52162    0.537645  0.0118403    \n",
      "67  KerasClassifier  0.511509  0.519962   0.529115  0.00720442   \n",
      "68  KerasClassifier  0.496369  0.517967   0.53094   0.0153751    \n",
      "69  KerasClassifier  0.460586  0.510564   0.605639  0.0672576    \n",
      "70  KerasClassifier  0.460586  0.508999   0.548597  0.0364672    \n",
      "71  KerasClassifier  0.465467  0.508746   0.548949  0.0341507    \n",
      "72  KerasClassifier  0.48598   0.50842    0.542666  0.0246016    \n",
      "73  KerasClassifier  0.480949  0.49878    0.512147  0.0131214    \n",
      "74  KerasClassifier  0.482344  0.498306   0.526593  0.020057     \n",
      "75  KerasClassifier  0.464804  0.495027   0.510876  0.0213792    \n",
      "76  KerasClassifier  0.482344  0.494231   0.511509  0.012502     \n",
      "77  KerasClassifier  0.472938  0.494102   0.516724  0.017905     \n",
      "78  KerasClassifier  0.465467  0.489917   0.523334  0.0244603    \n",
      "79  KerasClassifier  0.458451  0.48122    0.496369  0.0163908    \n",
      "80  KerasClassifier  0.458451  0.476268   0.489403  0.0130625    \n",
      "81  KerasClassifier  0.458451  0.474334   0.496615  0.0162236    \n",
      "82  KerasClassifier  0.458451  0.471244   0.482344  0.00982747   \n",
      "83  KerasClassifier  0.464804  0.470796   0.475189  0.00438748   \n",
      "84  KerasClassifier  0.464804  0.470796   0.475189  0.00438748   \n",
      "85  KerasClassifier  0.458451  0.46886    0.475189  0.00741706   \n",
      "86  KerasClassifier  0.458451  0.468533   0.482344  0.0101043    \n",
      "87  KerasClassifier  0.464804  0.468487   0.475189  0.004747     \n",
      "88  KerasClassifier  0.464804  0.468487   0.475189  0.004747     \n",
      "89  KerasClassifier  0.458451  0.466148   0.475189  0.00689912   \n",
      "90  KerasClassifier  0.464804  0.46607    0.467937  0.00134804   \n",
      "91  KerasClassifier  0.458451  0.463731   0.467937  0.00394641   \n",
      "92  KerasClassifier  0.458451  0.463731   0.467937  0.00394641   \n",
      "93  KerasClassifier  0.458451  0.463731   0.467937  0.00394641   \n",
      "94  KerasClassifier  0.458451  0.46128    0.464804  0.00263987   \n",
      "95  KerasClassifier  0.458451  0.46128    0.464804  0.00263987   \n",
      "\n",
      "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              Vectorizer  \\\n",
      "0   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "1   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "2   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "3   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "4   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "5   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "6   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "7   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "8   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "9   CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "10  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "11  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "12  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "13  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "14  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "15  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "16  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "17  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "18  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "19  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "20  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "21  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "22  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "23  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "24  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "25  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "26  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "27  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "28  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "29  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "30  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "31  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "32  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "33  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "34  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "35  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "36  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "37  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "38  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "39  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "40  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "41  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "42  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "43  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "44  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "45  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "46  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "47  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "48  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "49  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "50  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "51  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "52  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "53  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "54  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "55  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "56  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "57  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "58  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "59  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "60  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "61  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "62  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "63  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "64  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "65  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "66  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "67  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "68  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "69  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "70  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "71  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "72  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "73  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "74  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "75  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "76  CountVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',\\n                lowercase=True, max_df=1.0, max_features=None, min_df=1,\\n                ngram_range=(1, 1), preprocessor=None, stop_words=None,\\n                strip_accents=None, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, vocabulary=None)                                                                                    \n",
      "77  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "78  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "79  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "80  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "81  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "82  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "83  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "84  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "85  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "86  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "87  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "88  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "89  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "90  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "91  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "92  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "93  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "94  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "95  TfidfVectorizer(analyzer='word', binary=False, decode_error='strict',\\n                dtype=<class 'numpy.float64'>, encoding='utf-8',\\n                input='content', lowercase=True, max_df=1.0, max_features=None,\\n                min_df=1, ngram_range=(1, 1), norm='l2', preprocessor=None,\\n                smooth_idf=True, stop_words=None, strip_accents=None,\\n                sublinear_tf=False, token_pattern='(?u)\\\\b\\\\w\\\\w+\\\\b',\\n                tokenizer=None, use_idf=True, vocabulary=None)   \n",
      "\n",
      "   Vectorizer__ngram_range  \\\n",
      "0   (1, 1)                   \n",
      "1   (1, 1)                   \n",
      "2   (1, 1)                   \n",
      "3   (1, 1)                   \n",
      "4   (1, 1)                   \n",
      "5   (1, 1)                   \n",
      "6   (1, 1)                   \n",
      "7   (1, 3)                   \n",
      "8   (1, 1)                   \n",
      "9   (1, 1)                   \n",
      "10  (1, 1)                   \n",
      "11  (1, 1)                   \n",
      "12  (1, 3)                   \n",
      "13  (1, 3)                   \n",
      "14  (1, 1)                   \n",
      "15  (1, 3)                   \n",
      "16  (1, 1)                   \n",
      "17  (1, 1)                   \n",
      "18  (1, 1)                   \n",
      "19  (1, 3)                   \n",
      "20  (1, 1)                   \n",
      "21  (1, 1)                   \n",
      "22  (1, 1)                   \n",
      "23  (1, 3)                   \n",
      "24  (1, 3)                   \n",
      "25  (1, 1)                   \n",
      "26  (1, 1)                   \n",
      "27  (1, 3)                   \n",
      "28  (1, 3)                   \n",
      "29  (1, 3)                   \n",
      "30  (1, 3)                   \n",
      "31  (1, 3)                   \n",
      "32  (1, 3)                   \n",
      "33  (1, 3)                   \n",
      "34  (1, 3)                   \n",
      "35  (1, 3)                   \n",
      "36  (1, 3)                   \n",
      "37  (1, 3)                   \n",
      "38  (1, 3)                   \n",
      "39  (1, 1)                   \n",
      "40  (1, 3)                   \n",
      "41  (1, 1)                   \n",
      "42  (1, 1)                   \n",
      "43  (1, 1)                   \n",
      "44  (1, 3)                   \n",
      "45  (1, 3)                   \n",
      "46  (1, 3)                   \n",
      "47  (1, 3)                   \n",
      "48  (1, 1)                   \n",
      "49  (1, 3)                   \n",
      "50  (1, 1)                   \n",
      "51  (1, 1)                   \n",
      "52  (1, 1)                   \n",
      "53  (1, 3)                   \n",
      "54  (1, 3)                   \n",
      "55  (1, 1)                   \n",
      "56  (1, 1)                   \n",
      "57  (1, 3)                   \n",
      "58  (1, 3)                   \n",
      "59  (1, 3)                   \n",
      "60  (1, 1)                   \n",
      "61  (1, 3)                   \n",
      "62  (1, 1)                   \n",
      "63  (1, 1)                   \n",
      "64  (1, 1)                   \n",
      "65  (1, 3)                   \n",
      "66  (1, 3)                   \n",
      "67  (1, 3)                   \n",
      "68  (1, 1)                   \n",
      "69  (1, 3)                   \n",
      "70  (1, 3)                   \n",
      "71  (1, 1)                   \n",
      "72  (1, 3)                   \n",
      "73  (1, 1)                   \n",
      "74  (1, 3)                   \n",
      "75  (1, 3)                   \n",
      "76  (1, 3)                   \n",
      "77  (1, 3)                   \n",
      "78  (1, 3)                   \n",
      "79  (1, 3)                   \n",
      "80  (1, 3)                   \n",
      "81  (1, 1)                   \n",
      "82  (1, 3)                   \n",
      "83  (1, 3)                   \n",
      "84  (1, 3)                   \n",
      "85  (1, 1)                   \n",
      "86  (1, 1)                   \n",
      "87  (1, 1)                   \n",
      "88  (1, 1)                   \n",
      "89  (1, 1)                   \n",
      "90  (1, 3)                   \n",
      "91  (1, 1)                   \n",
      "92  (1, 1)                   \n",
      "93  (1, 1)                   \n",
      "94  (1, 1)                   \n",
      "95  (1, 1)                   \n",
      "\n",
      "                                    Vectorizer__tokenizer clf__Dropout  \\\n",
      "0   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "1   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "2   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "3   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "4   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "5   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "6   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "7   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "8   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "9   <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "10  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "11  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "12  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "13  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "14  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "15  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "16  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "17  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "18  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "19  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "20  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "21  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "22  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "23  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "24  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "25  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "26  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "27  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "28  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "29  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "30  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "31  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "32  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "33  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "34  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "35  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "36  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "37  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "38  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "39  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "40  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "41  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "42  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "43  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "44  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "45  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "46  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "47  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "48  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "49  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "50  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "51  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "52  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "53  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "54  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "55  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "56  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "57  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "58  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "59  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "60  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "61  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "62  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "63  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "64  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "65  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "66  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "67  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "68  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "69  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "70  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "71  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "72  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "73  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "74  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "75  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "76  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "77  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "78  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "79  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "80  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "81  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "82  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "83  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "84  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "85  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "86  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "87  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "88  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "89  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "90  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "91  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "92  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "93  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "94  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.4           \n",
      "95  <function ua_tokenizer_sklearn at 0x0000026CD9E183A8>  0.5           \n",
      "\n",
      "   clf__epochs clf__n_classes clf__n_layers clf__units  \\\n",
      "0   5           2              2             500         \n",
      "1   5           2              2             1000        \n",
      "2   5           2              2             1000        \n",
      "3   5           2              2             1000        \n",
      "4   5           2              2             500         \n",
      "5   5           2              2             500         \n",
      "6   5           2              2             500         \n",
      "7   5           2              2             500         \n",
      "8   5           2              3             1000        \n",
      "9   5           2              3             1000        \n",
      "10  5           2              3             500         \n",
      "11  5           2              2             1000        \n",
      "12  5           2              2             1000        \n",
      "13  5           2              2             1000        \n",
      "14  5           2              3             1000        \n",
      "15  5           2              2             1000        \n",
      "16  5           2              3             500         \n",
      "17  5           2              3             1000        \n",
      "18  5           2              2             1000        \n",
      "19  5           2              2             500         \n",
      "20  5           2              2             500         \n",
      "21  5           2              3             500         \n",
      "22  5           2              3             1000        \n",
      "23  5           2              3             1000        \n",
      "24  5           2              2             1000        \n",
      "25  5           2              2             500         \n",
      "26  5           2              3             500         \n",
      "27  5           2              3             500         \n",
      "28  5           2              2             500         \n",
      "29  5           2              3             1000        \n",
      "30  5           2              2             1000        \n",
      "31  5           2              2             1000        \n",
      "32  5           2              3             500         \n",
      "33  5           2              2             1000        \n",
      "34  5           2              2             500         \n",
      "35  5           2              3             1000        \n",
      "36  5           2              2             1000        \n",
      "37  5           2              2             500         \n",
      "38  5           2              2             500         \n",
      "39  5           2              3             500         \n",
      "40  5           2              2             1000        \n",
      "41  5           2              2             500         \n",
      "42  5           2              2             1000        \n",
      "43  5           2              3             1000        \n",
      "44  5           2              3             500         \n",
      "45  5           2              2             500         \n",
      "46  5           2              3             1000        \n",
      "47  5           2              3             1000        \n",
      "48  5           2              2             1000        \n",
      "49  5           2              2             500         \n",
      "50  5           2              2             1000        \n",
      "51  5           2              2             500         \n",
      "52  5           2              2             1000        \n",
      "53  5           2              2             1000        \n",
      "54  5           2              3             500         \n",
      "55  5           2              2             1000        \n",
      "56  5           2              2             500         \n",
      "57  5           2              2             1000        \n",
      "58  5           2              2             500         \n",
      "59  5           2              3             1000        \n",
      "60  5           2              3             500         \n",
      "61  5           2              3             1000        \n",
      "62  5           2              2             500         \n",
      "63  5           2              2             1000        \n",
      "64  5           2              2             1000        \n",
      "65  5           2              3             500         \n",
      "66  5           2              2             500         \n",
      "67  5           2              2             500         \n",
      "68  5           2              2             500         \n",
      "69  5           2              2             500         \n",
      "70  5           2              3             1000        \n",
      "71  5           2              2             500         \n",
      "72  5           2              3             500         \n",
      "73  5           2              3             1000        \n",
      "74  5           2              3             1000        \n",
      "75  5           2              2             1000        \n",
      "76  5           2              3             500         \n",
      "77  5           2              3             1000        \n",
      "78  5           2              3             500         \n",
      "79  5           2              3             500         \n",
      "80  5           2              3             500         \n",
      "81  5           2              3             1000        \n",
      "82  5           2              3             1000        \n",
      "83  5           2              3             1000        \n",
      "84  5           2              3             500         \n",
      "85  5           2              3             500         \n",
      "86  5           2              3             1000        \n",
      "87  5           2              3             500         \n",
      "88  5           2              3             500         \n",
      "89  5           2              3             1000        \n",
      "90  5           2              3             500         \n",
      "91  5           2              3             1000        \n",
      "92  5           2              3             500         \n",
      "93  5           2              3             1000        \n",
      "94  5           2              3             500         \n",
      "95  5           2              3             500         \n",
      "\n",
      "   feature_selection__threshold  \n",
      "0   0.1                          \n",
      "1   0.1                          \n",
      "2   0.2                          \n",
      "3   0.1                          \n",
      "4   0.2                          \n",
      "5   0.1                          \n",
      "6   0.2                          \n",
      "7   0.1                          \n",
      "8   0.1                          \n",
      "9   0.1                          \n",
      "10  0.1                          \n",
      "11  0.2                          \n",
      "12  0.1                          \n",
      "13  0.2                          \n",
      "14  0.2                          \n",
      "15  0.1                          \n",
      "16  0.1                          \n",
      "17  0.2                          \n",
      "18  0.5                          \n",
      "19  0.1                          \n",
      "20  0.5                          \n",
      "21  0.2                          \n",
      "22  0.5                          \n",
      "23  0.1                          \n",
      "24  0.2                          \n",
      "25  0.5                          \n",
      "26  0.2                          \n",
      "27  0.1                          \n",
      "28  0.1                          \n",
      "29  0.1                          \n",
      "30  0.1                          \n",
      "31  0.1                          \n",
      "32  0.1                          \n",
      "33  0.2                          \n",
      "34  0.2                          \n",
      "35  0.5                          \n",
      "36  0.2                          \n",
      "37  0.2                          \n",
      "38  0.2                          \n",
      "39  0.5                          \n",
      "40  0.5                          \n",
      "41  0.5                          \n",
      "42  0.5                          \n",
      "43  0.5                          \n",
      "44  0.5                          \n",
      "45  0.1                          \n",
      "46  0.2                          \n",
      "47  0.5                          \n",
      "48  0.5                          \n",
      "49  0.2                          \n",
      "50  0.1                          \n",
      "51  0.1                          \n",
      "52  0.5                          \n",
      "53  0.5                          \n",
      "54  0.2                          \n",
      "55  0.2                          \n",
      "56  0.2                          \n",
      "57  0.5                          \n",
      "58  0.5                          \n",
      "59  0.2                          \n",
      "60  0.5                          \n",
      "61  0.5                          \n",
      "62  0.1                          \n",
      "63  0.2                          \n",
      "64  0.1                          \n",
      "65  0.5                          \n",
      "66  0.5                          \n",
      "67  0.5                          \n",
      "68  0.5                          \n",
      "69  0.5                          \n",
      "70  0.5                          \n",
      "71  0.2                          \n",
      "72  0.5                          \n",
      "73  0.5                          \n",
      "74  0.2                          \n",
      "75  0.5                          \n",
      "76  0.2                          \n",
      "77  0.2                          \n",
      "78  0.5                          \n",
      "79  0.2                          \n",
      "80  0.2                          \n",
      "81  0.5                          \n",
      "82  0.1                          \n",
      "83  0.1                          \n",
      "84  0.1                          \n",
      "85  0.5                          \n",
      "86  0.2                          \n",
      "87  0.5                          \n",
      "88  0.2                          \n",
      "89  0.2                          \n",
      "90  0.1                          \n",
      "91  0.1                          \n",
      "92  0.2                          \n",
      "93  0.1                          \n",
      "94  0.1                          \n",
      "95  0.1                          \n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import nltk\n",
    "import re\n",
    "import random\n",
    "import tensorflow as tf\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.feature_extraction.text import CountVectorizer,TfidfVectorizer\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.feature_selection import SelectFromModel\n",
    "from data_print import data_class_display,PCA_text_data_display,pp_conf_matrix\n",
    "from EstimatorSelectionHelper import GridSearchCV_Classifiers, Best_Classifiers\n",
    "from ukrainian_stemmer import UkrainianStemmer\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "from sklearn.svm import LinearSVC\n",
    "\n",
    "def ua_tokenizer_sklearn(text,stemmer=True):\n",
    "    \"\"\" Tokenizer for Ukrainian language, returns only alphabetic tokens. \n",
    "    \n",
    "    Keyword arguments:\n",
    "    text -- text for tokenize \n",
    "    stemmer -- if True use UkrainianStemmer for stemming words (default True)\n",
    "          \n",
    "    \"\"\"\n",
    "    text=re.sub(r\"\"\"['’\"`�]\"\"\", '', text)\n",
    "    text=re.sub(r\"\"\"([0-9])([\\u0400-\\u04FF]|[A-z])\"\"\", r\"\\1 \\2\", text)\n",
    "    text=re.sub(r\"\"\"([\\u0400-\\u04FF]|[A-z])([0-9])\"\"\", r\"\\1 \\2\", text)\n",
    "    text=re.sub(r\"\"\"[\\-.,:+*/_]\"\"\", ' ', text)\n",
    "    if stemmer==True:\n",
    "        return [UkrainianStemmer(word).stem_word() for word in nltk.word_tokenize(text) if word.isalpha()]\n",
    "    else:\n",
    "        return [word for word in nltk.word_tokenize(text) if word.isalpha()]\n",
    "    \n",
    "    \n",
    "land_data=pd.read_csv('land_real_estate.csv').replace(np.nan, \"\")\n",
    "land_data['text']=land_data['description']+' '+land_data['land_types_source']+' '+land_data['land_types_PcmU']\n",
    "def land_types_download():    \n",
    "    data=pd.read_csv('land_types.csv')\n",
    "    data=data[['id','land_types']]\n",
    "    data=data.set_index('id')\n",
    "    data=data.T\n",
    "    data=data.to_dict('records')\n",
    "    return (data[0])\n",
    "\n",
    "land_types=land_types_download()\n",
    "built_up={0:'Не забудована',1:'Забудована'}\n",
    "\n",
    "def Data_Augmentation(number):\n",
    "    unigrams_list=['продаж','земельна', 'ділянка', 'іншої', 'інфраструктура', 'інструменту', 'інвентарю', 'ізумрудє', 'івасюка',\n",
    "                   'яром', 'яворівського', 'шанове', 'чудова','цін', 'цільових', 'цілу', 'цілорічно', 'участке', 'тільки', 'тухолька', \n",
    "                   'турківський', 'твердій','сухий', 'суха', 'сусідніх', 'сусідні', 'судова',  'сторони', 'сто', 'стихії','селі', 'села',  \n",
    "                   'сайті', 'руська', 'росташування','рокитне', 'розташовану', 'розташований', 'розміщені', 'розміщення', 'розміщений', 'розмірі', \n",
    "                   'розділена', 'покупцю', 'показ', 'повідомлення', 'питання']\n",
    "   \n",
    "    random.seed(0)\n",
    "    random_words=[unigrams_list[index] for index in random.sample(range(len(unigrams_list)), len(unigrams_list))]\n",
    "    tsv='для ведення товарного сільськогосподарського виробництва'\n",
    "    df = pd.DataFrame(columns=['text','land_types','built_up'])\n",
    "    for sample_index in range(number):\n",
    "        df = df.append({'text': tsv+' '+\" \".join(random.sample(random_words, random.randint(0,20))),'land_types':6,'built_up':0}, ignore_index=True)\n",
    "    return df.astype({'text':'object','land_types': 'int32','built_up': 'int32'})\n",
    "\n",
    "land_data=land_data.append(Data_Augmentation(34), ignore_index=True)\n",
    "data_class_display(dataframe=land_data,class_column='land_types',expl_lables=land_types)\n",
    "data_class_display(dataframe=land_data,class_column='built_up',expl_lables=built_up)\n",
    "\n",
    "X_train_zab, X_test_zab, y_train_zab, y_test_zab=train_test_split(land_data['text'],land_data['built_up'],\n",
    "                                                                  stratify=land_data['built_up'],\n",
    "                                                                  test_size=0.33,random_state=0)\n",
    "\n",
    "X_train, X_test, y_train, y_test=train_test_split(land_data['text'],land_data['land_types'],\n",
    "                                                  stratify=land_data['land_types'],\n",
    "                                                  test_size=0.33,random_state=0)\n",
    "\n",
    "\n",
    "from tensorflow.keras.wrappers.scikit_learn import KerasClassifier\n",
    "from sklearn.base import TransformerMixin\n",
    "def skl_model(units=2000,n_classes=10,n_layers=1,Dropout=0.5): \n",
    "    model = tf.keras.Sequential()\n",
    "    \n",
    "    model = tf.keras.models.Sequential()\n",
    "    for n in range(1,n_layers+1):\n",
    "        model.add(tf.keras.layers.Dense(units/n, activation='relu'))\n",
    "        model.add(tf.keras.layers.BatchNormalization())\n",
    "        model.add(tf.keras.layers.Dropout(Dropout))\n",
    "    model.add(tf.keras.layers.Dense(n_classes, activation='softmax'))\n",
    "    \n",
    "    model.compile(optimizer='adam',\n",
    "                  loss='sparse_categorical_crossentropy',\n",
    "                  metrics=['accuracy'])\n",
    "    \n",
    "    return model\n",
    "\n",
    "class ToarrayTransformer(TransformerMixin):\n",
    "\n",
    "    def fit(self, X, y=None, **fit_params):\n",
    "        return self\n",
    "\n",
    "    def transform(self, X, y=None, **fit_params):\n",
    "        return X.toarray()\n",
    "\n",
    "    \n",
    "y_train=np.array(y_train)\n",
    "y_test=np.array(y_test)\n",
    "y_train_zab=np.array(y_train_zab)\n",
    "y_test_zab=np.array(y_test_zab)\n",
    "models_params_typy={\n",
    "        'KerasClassifier': [Pipeline([('Vectorizer',TfidfVectorizer()),\n",
    "                                             ('feature_selection',SelectFromModel(LinearSVC())),\n",
    "                                              ('ToarrayTransformer',ToarrayTransformer()),\n",
    "                                             ('clf',KerasClassifier(build_fn=skl_model, verbose=0))]),\n",
    "                                      {'Vectorizer':[TfidfVectorizer(),CountVectorizer()],\n",
    "                                       'Vectorizer__ngram_range':[(1,1),(1,3)],\n",
    "                                        'Vectorizer__tokenizer':[ua_tokenizer_sklearn],\n",
    "                                      'feature_selection__threshold':[0.2,0.1,0.5],\n",
    "                                         'clf__units':[1000,500],\n",
    "                                        'clf__n_classes':[7],                                         \n",
    "                                         'clf__n_layers':[3,2],\n",
    "                                         'clf__Dropout':[0.5,0.4],\n",
    "                                         'clf__epochs':[5],\n",
    "                                        },\n",
    "                                     ],   \n",
    "\n",
    "    \n",
    "}\n",
    "\n",
    "  \n",
    "GridSearchCV_Classifiers(X=X_train, y=y_train-1,\n",
    "                         models_params=models_params_typy,scoring='f1_macro',cv=3)\n",
    "\n",
    "\n",
    "models_params_zab={\n",
    "        'KerasClassifier': [Pipeline([('Vectorizer',TfidfVectorizer()),\n",
    "                                             ('feature_selection',SelectFromModel(LinearSVC())),\n",
    "                                              ('ToarrayTransformer',ToarrayTransformer()),\n",
    "                                             ('clf',KerasClassifier(build_fn=skl_model, verbose=0))]),\n",
    "                                      {'Vectorizer':[TfidfVectorizer(),CountVectorizer()],\n",
    "                                       'Vectorizer__ngram_range':[(1,1),(1,3)],\n",
    "                                        'Vectorizer__tokenizer':[ua_tokenizer_sklearn],\n",
    "                                        'feature_selection__threshold':[0.2,0.1,0.5],\n",
    "                                         'clf__units':[1000,500],\n",
    "                                        'clf__n_classes':[2],                                         \n",
    "                                         'clf__n_layers':[3,2],\n",
    "                                         'clf__Dropout':[0.5,0.4],\n",
    "                                         'clf__epochs':[5],\n",
    "                                        },\n",
    "                                     ],   \n",
    "\n",
    "    \n",
    "}\n",
    "\n",
    "\n",
    "GridSearchCV_Classifiers(X=X_train_zab, y=y_train_zab,\n",
    "                         models_params=models_params_zab,scoring='f1_macro',cv=3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Нейронна мережа із підібраними гіперпараметрами"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\User\\Documents\\article\\data_print.py:50: UserWarning: You have mixed positional and keyword arguments, some input may be discarded.\n",
      "  plt.legend(fracs, labels=labels, loc='best', bbox_to_anchor=(-0.1, 1.), fontsize=8)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "KerasClassifier\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\_base.py:947: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x1008 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "________________________________________________________________________\n",
      "KerasClassifier\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\_base.py:947: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  \"the number of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import nltk\n",
    "import re\n",
    "import random\n",
    "import tensorflow as tf\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.feature_extraction.text import CountVectorizer,TfidfVectorizer\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.feature_selection import SelectFromModel\n",
    "from data_print import data_class_display,PCA_text_data_display,pp_conf_matrix\n",
    "from EstimatorSelectionHelper import GridSearchCV_Classifiers, Best_Classifiers\n",
    "from ukrainian_stemmer import UkrainianStemmer\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "\n",
    "\n",
    "def ua_tokenizer_sklearn(text,stemmer=True):\n",
    "    \"\"\" Tokenizer for Ukrainian language, returns only alphabetic tokens. \n",
    "    \n",
    "    Keyword arguments:\n",
    "    text -- text for tokenize \n",
    "    stemmer -- if True use UkrainianStemmer for stemming words (default True)\n",
    "          \n",
    "    \"\"\"\n",
    "    text=re.sub(r\"\"\"['’\"`�]\"\"\", '', text)\n",
    "    text=re.sub(r\"\"\"([0-9])([\\u0400-\\u04FF]|[A-z])\"\"\", r\"\\1 \\2\", text)\n",
    "    text=re.sub(r\"\"\"([\\u0400-\\u04FF]|[A-z])([0-9])\"\"\", r\"\\1 \\2\", text)\n",
    "    text=re.sub(r\"\"\"[\\-.,:+*/_]\"\"\", ' ', text)\n",
    "    if stemmer==True:\n",
    "        return [UkrainianStemmer(word).stem_word() for word in nltk.word_tokenize(text) if word.isalpha()]\n",
    "    else:\n",
    "        return [word for word in nltk.word_tokenize(text) if word.isalpha()]\n",
    "    \n",
    "    \n",
    "    \n",
    "land_data=pd.read_csv('land_real_estate.csv').replace(np.nan, \"\")\n",
    "land_data['text']=land_data['description']+' '+land_data['land_types_source']+' '+land_data['land_types_PcmU']\n",
    "def land_types_download():    \n",
    "    data=pd.read_csv('land_types.csv')\n",
    "    data=data[['id','land_types']]\n",
    "    data=data.set_index('id')\n",
    "    data=data.T\n",
    "    data=data.to_dict('records')\n",
    "    return (data[0])\n",
    "\n",
    "land_types=land_types_download()\n",
    "built_up={0:'Не забудована',1:'Забудована'}\n",
    "\n",
    "def Data_Augmentation(number):\n",
    "    unigrams_list=['продаж','земельна', 'ділянка', 'іншої', 'інфраструктура', 'інструменту', 'інвентарю', 'ізумрудє', 'івасюка',\n",
    "                   'яром', 'яворівського', 'шанове', 'чудова','цін', 'цільових', 'цілу', 'цілорічно', 'участке', 'тільки', 'тухолька', \n",
    "                   'турківський', 'твердій','сухий', 'суха', 'сусідніх', 'сусідні', 'судова',  'сторони', 'сто', 'стихії','селі', 'села',  \n",
    "                   'сайті', 'руська', 'росташування','рокитне', 'розташовану', 'розташований', 'розміщені', 'розміщення', 'розміщений', 'розмірі', \n",
    "                   'розділена', 'покупцю', 'показ', 'повідомлення', 'питання']\n",
    "   \n",
    "    random.seed(0)\n",
    "    random_words=[unigrams_list[index] for index in random.sample(range(len(unigrams_list)), len(unigrams_list))]\n",
    "    tsv='для ведення товарного сільськогосподарського виробництва'\n",
    "    df = pd.DataFrame(columns=['text','land_types','built_up'])\n",
    "    for sample_index in range(number):\n",
    "        df = df.append({'text': tsv+' '+\" \".join(random.sample(random_words, random.randint(0,20))),'land_types':6,'built_up':0}, ignore_index=True)\n",
    "    return df.astype({'text':'object','land_types': 'int32','built_up': 'int32'})\n",
    "\n",
    "land_data=land_data.append(Data_Augmentation(34), ignore_index=True)\n",
    "data_class_display(dataframe=land_data,class_column='land_types',expl_lables=land_types)\n",
    "data_class_display(dataframe=land_data,class_column='built_up',expl_lables=built_up)\n",
    "\n",
    "X_train_zab, X_test_zab, y_train_zab, y_test_zab=train_test_split(land_data['text'],land_data['built_up'],\n",
    "                                                                  stratify=land_data['built_up'],\n",
    "                                                                  test_size=0.33,random_state=0)\n",
    "\n",
    "X_train, X_test, y_train, y_test=train_test_split(land_data['text'],land_data['land_types'],\n",
    "                                                  stratify=land_data['land_types'],\n",
    "                                                  test_size=0.33,random_state=0)\n",
    "\n",
    "from tensorflow.keras.wrappers.scikit_learn import KerasClassifier\n",
    "from sklearn.base import TransformerMixin\n",
    "def skl_model(units=2000,n_classes=10,n_layers=1,Dropout=0.5): \n",
    "    model = tf.keras.Sequential()\n",
    "    \n",
    "    model = tf.keras.models.Sequential()\n",
    "    for n in range(1,n_layers+1):\n",
    "        model.add(tf.keras.layers.Dense(units/n, activation='relu'))\n",
    "        model.add(tf.keras.layers.BatchNormalization())\n",
    "        model.add(tf.keras.layers.Dropout(Dropout))\n",
    "    model.add(tf.keras.layers.Dense(n_classes, activation='softmax'))\n",
    "    \n",
    "    model.compile(optimizer='adam',\n",
    "                  loss='sparse_categorical_crossentropy',\n",
    "                  metrics=['accuracy'])\n",
    "    \n",
    "    return model\n",
    "\n",
    "class ToarrayTransformer(TransformerMixin):\n",
    "\n",
    "    def fit(self, X, y=None, **fit_params):\n",
    "        return self\n",
    "\n",
    "    def transform(self, X, y=None, **fit_params):\n",
    "        return X.toarray()\n",
    "\n",
    "    \n",
    "y_train=np.array(y_train)\n",
    "y_test=np.array(y_test)\n",
    "y_train_zab=np.array(y_train_zab)\n",
    "y_test_zab=np.array(y_test_zab)\n",
    "models_params_typy={\n",
    "        'KerasClassifier': [Pipeline([('Vectorizer',CountVectorizer()),\n",
    "                                             ('feature_selection',SelectFromModel(LinearSVC())),\n",
    "                                              ('ToarrayTransformer',ToarrayTransformer()),\n",
    "                                             ('clf',KerasClassifier(build_fn=skl_model, verbose=0))]),\n",
    "                                      {'Vectorizer__ngram_range':[(1,3)],\n",
    "                                        'Vectorizer__tokenizer':[ua_tokenizer_sklearn],\n",
    "                                      'feature_selection__threshold':[0.2],\n",
    "                                         'clf__units':[500],\n",
    "                                        'clf__n_classes':[7],                                         \n",
    "                                         'clf__n_layers':[2],\n",
    "                                         'clf__Dropout':[0.4],\n",
    "                                         'clf__epochs':[5],\n",
    "                                        },\n",
    "                                     ],   \n",
    "\n",
    "    \n",
    "}\n",
    "\n",
    "  \n",
    "Best_Classifiers(X_train, X_test, y_train, y_test,\n",
    "                         models_params=models_params_typy,expl_lables=land_types)\n",
    "\n",
    "\n",
    "models_params_zab={\n",
    "        'KerasClassifier': [Pipeline([('Vectorizer',CountVectorizer()),\n",
    "                                             ('feature_selection',SelectFromModel(LinearSVC())),\n",
    "                                              ('ToarrayTransformer',ToarrayTransformer()),\n",
    "                                             ('clf',KerasClassifier(build_fn=skl_model, verbose=0))]),\n",
    "                                      {'Vectorizer__ngram_range':[(1,1)],\n",
    "                                        'Vectorizer__tokenizer':[ua_tokenizer_sklearn],\n",
    "                                        'feature_selection__threshold':[0.1],\n",
    "                                         'clf__units':[500],\n",
    "                                        'clf__n_classes':[2],                                         \n",
    "                                         'clf__n_layers':[2],\n",
    "                                         'clf__Dropout':[0.5],\n",
    "                                         'clf__epochs':[5],\n",
    "                                        },\n",
    "                                     ],   \n",
    "\n",
    "    \n",
    "}\n",
    "\n",
    "\n",
    "Best_Classifiers(X_train_zab, X_test_zab, y_train_zab, y_test_zab,\n",
    "                         models_params=models_params_zab,expl_lables=built_up)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Приклад згорткової нейронної мережі"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\User\\Documents\\article\\data_print.py:50: UserWarning: You have mixed positional and keyword arguments, some input may be discarded.\n",
      "  plt.legend(fracs, labels=labels, loc='best', bbox_to_anchor=(-0.1, 1.), fontsize=8)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAd0AAAD3CAYAAABYfY6KAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOzdd1ST1xsH8O+bQQYEhLBBNgRxoEYi4Kqj7lWlWkeLtVZbZ6Xu1l1tpYoVV62rLhwVFcX5s1oHwyhSFJGA7C1LdkLG+/sDQwFxVKug3s85nmPfeZPYPLn3vfd5KJqmQRAEQRDE68do6gYQBEEQxPuCBF2CIAiCeENI0CUIgiCIN4QEXYIgCIJ4Q0jQJQiCIIg3hARdgiAIgnhDSNAlCIIgiDeEBF2CIIjXhKKoVIqi+tT5byuKopIpivJvynYRTYcEXYIgiDeAoigTABcBnKNpel5Tt4doGiToEgRBvGYURbUAcAGAFMC0OtsnUBSlpiiq/PEfmqIop8f7Pqco6j5FUWWPe8dTGlxzGEVRf1MUVUpRVBJFUf0fb/+LoqhJj//OoCjqLkVRmW/sxRLPRIIuQRDE66UH4CwAFoCJdP3cuwwA4TRN69E0rdfgvIcABgPQB/A5gPUURXUEAIqiJAD2ApgLoAWA7gBSG7m3LwDD/+6lEK+KBF2CIIjXayuAcgDWALo02KcDoLqxk2iaPk3TdBJd4wpqesrdHu/+AsAumqb/R9O0hqbpLJqm4+ueT1EUF8BiACv/w9dCvCISdAmCIF6veABDAMwDsJOiKF6dfUYAihs7iaKoARRFRVIUVURR1CMAAwEYP97dEkDSc+47C8B5ALJXaTzx3yJBlyAI4vVaRdO0nKbp7QDSUb/n6QIgoeEJFEVxAAQDWAvAjKbpFgDOAKAeH5IBwPEZ9zQCMB3A8ldvPvFfIkGXIAjizfkSwGSKoiQURXUBMBxASCPH6QDgAMgHoKIoagCAvnX27wTwOUVRvR9PlrKiKMq1zv5vAOykaTr39bwM4mWRoEsQBPGG0DSdDGAJgBsA9gGYQ9P0jUaOKwMwE8AR1Aw/jwVwss5+KR5PrgJQAuAKANs6l2CippdMNDMUKWJPEARBEG8G6ekSBEEQxBtCgi5BEARBvCEk6BIEQRDEG0KCLkEQBEG8ISToEgRBEMQbQoIuQRAEQbwhJOgSBEEQxBtCgi5BEARBvCGsZ+2MiooyZbFYOwC0AQnQBPE6aQDEqlSqSWKx+GFTN4YgiNfjmUGXxWLtMDc3b2ViYlLMYDBI6iqCeE00Gg2Vn5/vlpubuwPA0KZuD0EQr8fzeq9tTExMSknAJYjXi8Fg0CYmJiWoGVUiCOId9bygyyABlyDejMf/r5HHOATxDmsW/4Onpqay3dzcWnE4nI5KpfKJ/WVlZQwfHx87iUQi6ty5s8u1a9f4ADB37lyLhQsXmgNAaGioYObMmZYAcOHCBd0ePXo4qdVq2NnZtZFIJCKJRCIaPHiww4u0R61WY/LkydZeXl4uAwYMcACAI0eO6Nvb27cWi8Wixs4JCwvjeXt7O3fu3Nnl0KFDBgDw1VdfWXfs2NFVLBaL7t69y6l7/MGDBw3c3d1dZ82aZQkAt2/f5i5YsMC8sWsvWbLELCwsjHfp0iXdDh06uIrFYtEXX3zRUrv/+PHj+p6eni4SiUSkfW8A4Pfff29hbm7e7nnHvc18fX1bPm3f8OHD7TUazZtsDkEQxDM985luXXKl2p3LZr7w8Y2cr+KymTGN7TM1NVVduXIlYdCgQU6N7Z83b57Fhx9+WDplypSirKws1pAhQ5zCw8Nla9asyenTp4/TyZMnBQxGze+HrKws1ty5c1uePXv2AZPJhFAoVEqlUtm/aevu3bsNW7VqJf/tt98ytdt69uxZERcXF9e1a1eXxs5ZsWKF5fnz55MEAoEGAPLy8ph///03//bt2/EXLlzQ/eWXX0x37tyZoT3+wIEDQqlUKhsyZIgDAKxbt85006ZNmQ2vq1arERERobdixYq89PR0VlhYmIzP59NDhw61l0qlPDc3N/m2bduMr1+/nsBi1f94goODDS0sLKoBoLy8nHracW+zPXv2ZDxtn0QiKT958qRg+PDhZW+yTQRBEE/zwt++XDaTZbfg9EvfKPWnQU+9F5/Pp/l8vvpp+6VSqd7WrVuzAMDKykrVrVu30suXL+sOGDCg/PDhwyldu3YVzZ49O1etVlM+Pj4Oa9asybS0tFS9bFtPnz5tYGxsrJJIJKLRo0cXfvvttwUmJiZPbV9cXJyOQqGgBg8e7MBms+ndu3enmZiYqA0NDVUqlQpFRUUsoVBYrz1sNlujVCpBURQiIiJ4Tk5OCkNDwye6ZZGRkTxLS8tqALCxsam9BovFoplMJn3p0iU9BoOBHj16OJuYmKj27t2bpq+vrzl06JBB7969S/ft22cMAE87Tnu92NhYzpQpU2yqq6sZPXv2LPX3988ZOXKk3fLly3Pi4uI4J06cMAwKCkobP368TVxcHJ/BYNDnz59/wOFwNGPGjLHLz89nC4VC1bFjx1K2bdtm5O/vb5menn63pKSEYW1t7X748OEHADB58mQ7a2vr6tLSUmbfvn0fBQYGZoeGhgq02x9/3rIFCxaYh4aGGqpUKkp7nEgkcnNxcalKTk7mbtmyJa1Hjx6VYrFYFBUVJRs5cqRdYmIiV6lUMr799tucSZMmFffr169s48aNJiToEgTRXDSL4eV/y9raujozM5MNALm5uSxdXV3N0qVLrYOCgoyVSiUVFxfH1R5bWFjI1g4va4einyc/P58tEonkYWFhssOHDwszMjKe+eMkOzubnZyczA0NDU2ePHly/uLFiy24XC7t4OCgcHBwaPPtt9/aTJs2raDuOXPmzMkbNWqU/YgRI4oDAwNNe/ToUT5u3Dib7du3G9Y97v79+1w7O7vquttu3LjBKyoqYonFYnlOTg47Ly+PfeXKlUQvL6/ygIAAEwDYu3ev8KuvvirSnvO047Tmz59vtWfPntSbN2/K7t+/z01KSmIDwK1bt3jbt2832b17dzoA7N+/P/327dvxrq6uVWlpaewNGzaYDB48uCQyMjKhe/fuZb///rshALRs2VJx+vRpwZ49ewzd3d3Ltffx8fEplEqlsrVr19b2UNVqNUaPHl1Yd0QiJCTEKDIyMr7ucbm5uTr79+9PCw0NfbBs2TLLhp9DUFBQyrlz5xIvXbokAIBWrVopEhMTuQ2PIwiCaCpvZdDNzMzUsba2VgLAokWLLLdv3546bdq03A4dOlRcv3494cCBA8Ly8nIKQO3wcmRkpCw6Oppf99mqVCrlNfasV19fX/3hhx+WsdlseHh4lN+/f5+DZzA0NFS7u7tXCAQCzZAhQ8oSEhK4t2/f5t6/f5+XkpISe/DgwaQ5c+ZY1T2nS5cuVWfOnEl2cXGRt23btnLLli0mO3bsyDh58mSLusfRdP15bHl5ecxp06bZ7N27NxUAWrRoofb09CxnsVjo379/aXx8PPfkyZMCT0/Pci6XW3tyY8fVvW5SUhJ37NixDhKJRJSYmMhLS0vTAYC1a9daMBgMcDgcGgAKCwuZbdu2bRUXF8e3s7Orvn//Pnfz5s2mEolEFBQUJHz48CELAAYPHvzo2LFjLcLDw/U8PT0rnvX+PXr0iCkQCOqNJMyePTvH0dGx7Zw5c2qf2drY2CgMDAw09vb2ytLSUmbD64wdO9bew8Oj1ZAhQ0q02yiKetatCYIg3qi3Iuh26tSpYtu2bUZAzTPbq1ev6n/wwQcVV69e5fP5fI1YLJZ37Nixqk2bNpVcLpeeOHFi/k8//WRa9xoMBgP6+vrqysrK2tcskUiqpFKpLDQ0NLnusZ6enuVRUVF8ALh37x7P0dGxXk+zoTZt2igKCgrYKpUKERERfFtb22qapqGvr69mMpkwMzNTNRYkAGDr1q0ms2fPLqisrGQwGAzI5fJ6n4mbm5s8NTVVBwCUSiVGjRpl7+/vn6kdau7WrVuFTCbjAoBUKuXb2dkp7ty5wztz5kyLbt26OScmJvJmzpxp2dhxde/j6Ogo/+OPP5KlUqksNjY2rkePHhWP25cmFosrduzYYQgAQqFQfffu3fu9e/cuCQ4ONhCJRHI/P79cqVQqi4mJiZ83b14+APB4PI1araYa9tIbExUVxW/durW87jYrKyvlJ598UlC3p5uens4pLS1lpKamshsGaaCmp5uamnr3hx9+sASA+/fvc5ycnOQNjyMIgmgqzWJGjUKhoHr27OkcHx/P6969u8uqVauynJycFJs3bzZes2ZN7s8//5w9YcIEmx07dphQFIVffvklncfj0cuWLbP87bff0hpe76uvviry8vISffPNNwXa4WWNRgNHR0d5586dq57XnhkzZhSMHj3aftOmTWa9evUqcXR0VF69epW/YMEC64SEBJ63t7fLxYsXE8+cOSNQqVTUJ598UvL555/nd+7cWcRgMLBv374UV1fXaj09PbVYLBap1Wpq/fr16Q3vc/bsWb3OnTtX8Hg8ety4cYUeHh6iHj161Hv+6OnpWbV06VIdANi1a5fRnTt3dBcsWGANAKtXr87s06dPRbdu3co6deok4vF4mqNHjyabmZmpv//++4cAIBaLRYGBgdkA0PC4uvdZs2ZN1qeffmpXXV1NsVgs+vTp00lATS9+6dKleT179nQaMWJEybhx4+yKiopYTCaT9vPzyxcIBOpx48bZbdu2zZSmaaxevTpLe81du3alUxSFuXPnPjEUrBUUFGRQUFDA9vHxqe2dKhQKas2aNRanT59OunLliq52u7m5efWYMWPsUlNTORs3bnzi/Rw7dqy9RqOhBg0a9AgAzp07pz906NCShscRBEE0Farh8GVdMTExqe7u7gXA6529TDzbkiVLzPr161fapUuX5/5geJdpJ0296PHDhg2zP3bsWAqT2eggQ7MUExNj7O7ubtfU7SAI4vX4N7OXXylgctlvzxdfc7NixYq8pm7D2ygkJCSlqdtAEARR11vxTJcgAODf9HIJgiCaIxJ0CYIgCOINIUGXIAiCIN4QEnQJgiAI4g1pFkH35s2bXG0ifx8fH7uGSerLysoYY8aMsZVIJKKOHTu6bty4UdhETa2nuLiY4eXl5dKpUydRz549nYqLixkAcPLkSUH79u1dO3fu7KLN7NTQhAkTWg4bNsz+edsAYPTo0bbu7u6uV69e5QPAN998YxkXF6fT8DiNRoPhw4fbA8Dp06f12rVr5+ru7u7q7+9vAgATJ05sqc3Opa+v3x4AFi1aZK7dxuPxOuTl5b31M97S09NZ8+fPbzT7WEZGBmvSpEnWb7pNBPEcugDcAPQB0A9AfwADAAwEMAjAYABDUFNredjjP70AtALQAgDJAvOWePElQEq5O9jcl1/Xq5SrwOY2OgO6Xbt2iujo6HgA8PHxsbt27Rq/R48eldr98+bNs+jevXvZwYMH0zQaDc6fP6/30u34D3E4HPrQoUMptra2ynXr1hlv2bLF+Lvvvnu4atUqi8uXLydER0fzli1bZrFv3756a0ozMjJY6enpnLoJHhrbBgC5ublMBoOBEydOJC1btszc0dGxuqKiguHm5vZE0omTJ08KPDw8KgBg3bp15sHBwckODg7VHTp0cJ03b17+rl27MoCaikj+/v7mALB69erc1atX5+bk5LCGDRvmaGZm9tQc028LGxsb1Zo1a3Ib29eyZUtVQUEBu6ioiGFkZERKEBFvCh+ALQB7AHYKldqpUqFuRVFw4LCYliwmxS0oV1Q9LFVoVBoaNE2DBkDTqP930NCu8tTnsmEi4DAM+WwOk0Ex5EpNsVKtyWEwqHh9LusORVFJAB4ASARA1qs3Ey8eRNlcFpYZvPydlpU89V7aFIMAoKOjo7G3t68XUOoWPGAwGBgwYEC5QqGgPvjgA2elUkkJhULV6dOnk1gsFgIDA4UBAQEWBgYGKjs7O0VISEhKREQEb9q0abZqtRpTpkx5OHXq1KLGEuR//vnnLWNjY/kajQaHDh1KcXZ2fmY2JT6fT9va2iqBfwoQlJWVMbhcrsbQ0FDTq1evioULF1o1PO/HH380mzFjxsPff/9d+Kxt2vdGqVRS5eXlDF1dXc3q1avN5s+f3+gSopCQkBYTJ04sBIBWrVpVFRcXM6uqqig+n18vuPzxxx+GI0aMKK677fDhwwYDBw581PCaBw8eNFi3bp25Wq2mvvvuu2wfH59S7XrZ6dOnW7Vq1arqiy++KO7evbtLZWUlo02bNpWHDh1Ke1oBBQAIDg5OPXHihGDkyJEuarU6ys/Pz/L06dMtDAwM1FlZWTqbN29OHT58eJl2O0VRGDhw4KO1a9dm9+vXz7GwsJCdl5fH3rx5cyqLxUJAQICZSqWiqqurqZCQkKRHjx4x582bZxUSEpJiZ2fXxtTUVFlRUcEICgpKdnd3V3h6epafOnVK39fX94nXSxD/ERsAH5TJlQMoiurFYTGMCsurqzKKKzUp+RU6KQUVvMziSmQWVyGzuAoFFQrQNAQvezMemwlTfY6JpQHPxEbIb2dvrDvCxUyvysFYj7ZoweXKlZpsFoP6ny6H9T8AVwGQZYhNpFlkpAKAAwcOGCxbtsza3t5e/iK9LTabTf/vf/9L1NPTo2fOnGl56tQp/Y8++qi0qqqK8e233+b07du3bN68eVYA8N1331kdPHgw2c7OTunh4SH64osvioGatIECgUCzcOFCi0mTJhUHBgZmCQQCzYkTJwSBgYEmGzduzHp2K2qUlJQwdu/ebXLx4sXEgoICpra8HwCo1ep6wz55eXnMgoIClpubm/xZ27QMDQ01bdu2rVy0aJHlN9988zA4OLjF0aNHW8TGxvKWL1+e4+joWFuAODk5mePs7KwAgJEjRz4aPny4EwDMnTs3p+41//zzT4OVK1fW6wmePHnScMuWLfV65Gq1GuvXrzePiIiQqdVqqmfPns4+Pj6lAPDLL78I2Ww2PW3atCKgpjKQRqOBt7e3C/BPAQUnJyflkCFD7LXD7MXFxaySkhLGsWPHDFu3bl2bk3nNmjUZ2kBb9/5r167NoCgKly5dEuTk5LCKi4tZt2/fjq97nFwup8LDwxO3b99uuGHDBhNfX9/aQg/a3Nvbtm0zunHjBt/d3V3h5OSkuHXr1jtRT5hoNlrinyDbGzT0I1MKlX/J8gWRyYVIyi9/paD6PFVKNdIKK5FWWImI5EKg5rtdAAAsBoXWlvp2nR2Ek3q4mIzuaGPIUWk0hQD+FHDZFwBcAfDUEpnEf6vZBN1x48aVjBs3rsTX17fl4cOHDT777LNn9kLKysoY48ePt83NzdUpKChgaYNNUVER09bWtl4PtaSkhCkSibTl8aqzs7NZQE3awNzcXJ0NGzakAcDSpUvNr1y5IlCpVJSzs3O9APjll19ax8TE6M6bNy9HG3iAmueoY8aMsVuxYkWWsbGxmsPh0GVlZbXPyrV1frV++ukns1mzZj183ra6li5d+hAAJk+ebD1z5sz81atXmy9fvjwnICDAdPPmzbU/DGiarg3w8+bNs7569Wq8tbW1qlu3bi6TJk0qEggEmrt373LMzc2r6/4wKCkpYRQVFbFcXV3rvW+5ubmspKQkbpcuXUQAUFhYyNJoNCgqKmLt3r3bRJtuEah5hjx9+nTbrl27lgH/FFAAgNLSUqa2gMLw4cOLN2/eLGQymXTdggyNKSkpYRoYGKi1eautrKxU7du3r2jZsmUbjUZDde/evQwA2rVrVwUAEomk8uLFi/p1r1FYWMj28PAQZWVl6Vy+fFkG1HxmpBAC8Yp0AQwpkyuHUKD6gIL+jeRC1V+yfL2I5EI8eFgOAM2iwpVKQyMmswQxmSXUb1eT9RkUIDIXWHa2F37a3cVkuMTOiA0KZRoN/Zc+j/07gAsAXro0KvFszSLoVlVVUTwejwYAfX19TcPhUIlEUr5161ajr7/+ukij0eB///ufXk5ODsvZ2Vlx6tSplBkzZlhp01lKpVK9WbNmFdQNfAYGBmqZTKZjZ2enTEtL42hr7QYFBaU4ODhUe3h4uPbs2bP8+vXrgqioKNnx48f19+/fb1S3Ddu3b3+iwDwAzJ4929LLy6t86NChZQAgEAg0crmcUVJSwrh9+zbXxcWlXurG1NRUnQULFljL5XJGWloaZ8eOHYaNbZs0aVK94V+ZTKbD4/E0xsbGKrVaDQaDQZeXl9eL6I6OjvLExESOubl5JZPJpIVCoZrL5dIMBoOurq6mAODw4cMthg0bVu8HzR9//GHQp0+fJ575mJubq0QiUdXVq1cTWCwWFAoFxWAwoKurq7lx44Zs6NChDjKZTEckElUPGjSofNCgQfe8vb1dMjIyWI6OjvItW7Zk2NraKlUqFSiKwubNm0169epVPnLkSMcNGzakL1my5Jm9zfv37/PatGkjDw8Pr82/bGBgoN64cWPa1atXa3sNsbGxPAC4efMm38HBoV4hB6FQqLx586YsNDRU8MMPP5jv27cvPTk5mePq6koKIRD/FhNArzK5crIOkzH474xHytA7OYLI5EIkPix/7snNhYYG7ueU4X5OGX4PTxUAgKOJHreLk9BnXGeb/i2N+GqKonbz2MztAO43cXPfOc0i6AYHBxts2LDBDAAcHBzkH330UWl6ejpLW/DA398/Z/LkyS13795tolQqqYkTJ+YPHDiw1N/f3yI6OpovEAjUzs7O8rVr1xpHRkYKhg8f7qhQKKjMzExOeHg4b+XKlVmffPKJg1qtxuTJkx9qnyHXTZBvYmKi1tXVVXt6erq4ubm9UI7j1NRU9pYtW8w7dOhQcerUKcORI0cWzZ8/P3/hwoU5PXr0cOFwOJr9+/enAjUzh3/77beM48ePpwI1QXTevHlWkyZNKtYG2LrbGt5rzZo1ZqtXr842NzdXazQa6qOPPnJsmPR/yJAhJdeuXdPt1q1b5dy5c3N79OjhwmAw0KdPnxKhUKgGgHPnzrU4e/bsg7rnhYSEtFiyZEm9IWgAYDKZmDVrVp63t7eIoijaxcVFvm/fvnQOh6Phcrn0+vXrM7/55hvrDRs2ZE6YMMG2urqa4ejoKG/ZsqXqaQUUAODOnTv3ORwOvWTJkqe+t/PmzbMYNmxYcd3JTnFxcTpJSUncdevW5dQNumw2m+7WrZuzQqFgnDx58kFJSUntDGxtwYuysjLm8uXLMwEgIiJCb9euXU8UTCCIp3CqqlZPo0FPyH4kZx64kaZ3KiabKiivbhY92f9CUn45kvLLqb0RaQJHEz2M9mg5bbRHy68ApAi4rI0MijoM4InvJeLfe+GCB69z9vJ/xc/Pz3LKlCkF2qHkLVu2GNnY2CgHDx5c9rxz3wUajQYjRoywP3HixHuTczg0NFRw4cIFgbaS0vNkZWWxlixZYrFz585m+QyLFDxoNtgAhpRUKecwKarDoZvpzIPSDHZS/tvTo31VTAaFbk7GGOdpW9Hd2ZilUGn+p89jbwJwEcBbv8qhqbx40H0LFBQUMA0MDNRsds3S2NLSUgaLxaL5fP4znx0Sb69/G3SbOxJ0m5yeQqWerdbQfsn5Fczt15IF52JzoVC936vLDHhsDGtvSX/qaVtuZcir5rCYi5kMaieA59bLJup7p4IuQbztSNBtMjoqtWaKUq1ZeSWhgL3+fwl8Wd57MUD2r7lbG+C7QW7lrS31K3Q5rLkAgkB6vi+sWTzTJQiCaCJMAGMqFKp1MZmPdFedvq97L7v0uSe9z2IySzBqW4Sep4OR3uLBbltsjXRX6nFZfgCOAyCjis9Bgi5BEO8jCsCgcrlyQ1pRpemKU3F6N1KKnnsS8Y/I5CIMCryu94HIRG/JYLc9JgLOKgGX/Q1qlhyR4PsUJOgSBPG+6VYqV258VKl0WnHqnu7F+09dIk+8gL9k+biScEVvQBtz1+8GugULeCyZPpc9C8D1pm5bc9QsCh5cunRJV1vw4IsvvmjZcL9KpcKkSZOsJRKJSCwWi4KDg2sTIISEhAg8PDxEnTp1En344YeOubm5zJEjR9q1a9fOVSKRiLp06eL8Im1YuXKlabt27Vzbt2/vevHiRd26+4KCggzc3d1d27dv77p06VIzoGZ5j1AodH/aPfz9/U3qFhsIDg7W37x5s1HD44CaxBvp6emsY8eO6YvFYlH79u1dZ8yYYQXUTBSytLRsK5FIRB999JFd3fOWLVtmJhaLRS/y+t4Gvr6+T3z2QM067o8//tjuDTeHePdYl8qVl/LLFOeWnbzX7oOfL5OA+x+haeDM3Vx0//my7opTcR0KyhXny+TK4wCMm7ptzc0L93QVKoU7h8V56Z6xQqVQcVicRpcMOTk5KcLCwmR8Pp8eOnSovVQq5Ukkktq1suvWrTMxMTFRSaVSWWlpKaNXr17O3t7elQDwww8/WF64cCHR0NBQc+fOHY5CoWAANYkv2rRpo2jsfo0JCgoyjo2NjUtPT2dPnjzZpk+fPrXrSj08PKqioqLiGQwGOnfuLJo5c2YBAHTt2rU0JCSk0eU5f/75p35MTEx83759HefNm5cfFBRkdOTIkdSGxxUVFTHy8/PZNjY2KjMzs7IRI0aUAkDnzp1dtJmzfHx8ChvOzq2qqqLu3LnDe9HX9zbYs2dPo8t4eDwebWhoqLp79y6nbdu2L/yZEoSWRkOPlqvU23deS+Fu/SuJXa1+v2cjvy5qDY2jUZlU6J1s/rz+rgM+8Wj5gK/DmgzgD5AhZwD/IuhyWBxW2z1tX/pGd33vPvVeNjY2tSnHtIUD6u4PCQlpceHChQdATcaqcePGFQYHBxvQNI0xY8YUGhoaaoCaakUv2z47Ozt5VVUVVVhYyDQyMqqXAq1u4QMmk0kzGAwaACIiIgRisVg0dOjQYm2qxrrHKZVKMJlMev/+/S0GDRpUwmQ+WTUvNDRU393dvRL4p/CDSqWCqamp0tDQUA0Ax48fNwoLCxNMnjw5f8qUKUUA8MsvvxhPmDChcPny5ZYNr3np0iXdBQsWWKtUKsrX1zd/1qxZhRKJRBQWFibbsGGD8cOHD9n+/v45ffv2dczLy2MLhULVhQsXHjx8+JA1fvx4u4qKCoazs7N8/1vB91UAACAASURBVP796X5+fpbXr18XSKVSWWxsLKd9+/Zt7t27d/f8+fOCgIAAC1NTU2V+fj5rxowZeX5+fgV1C044OTnJg4ODU8ePH28TGxvLLykpYc6YMSNv0KBBpZ999pmdnp6euqCggH348OFkV1fXam0hBYlEIqJpGmVlZcxffvklvX///uW9e/cuCw4ObtG2bVuSqJ34N1qUy1W7SqqUfb/aH6V7N4sU23kT5EoNVpyK45z6O5uzcWyHXQY89ucCLvtzAI1W/3qfNIvhZa0bN27wioqKWGKxuF6KPoVCwdDT06sNxDY2NtU5OTnsnJwctqWlpfLJK/17PXv2LHNxcWkzcOBAFz8/v0bHnI4cOaJvb2+vMDQ01NjY2CgfPHgQGxkZKbt8+bL+jRs36vU6J02aVDB06FCHyZMnFxw/fryFUChUjR071ubkyZP1kp4nJCRw7O3ta38srF271tjBwaGNoaGhmsfj0d26dat48ODBvcuXLydu3brVNDs7m6VQKKhr164JtKknG1qyZInluXPnEm/duhV/+PBhoVwupwDg2LFjBjdv3tT19/fPAYALFy4kxcTExDMYDLq0tJSxdOlS80WLFuXcuHEjQSAQqLXD7CwWi46Li9PZtWuX0MPDo/aeM2fOzJVKpbLZs2fX/o+kVqup2bNn5xw5ciRZuy08PFxw+/bt+LrH5efns8+dO5e0fv36jBUrVjxR+zY8PFy2adOmtKtXr+oBgIuLi0Imk70zGYCIN6JXZbXqQUhM1oDeAX+RgNsEojMeofe6K7oHItN7VVWrZaipA/xeazZBNy8vjzlt2jSbvXv3pjbcx+Fw6PLy8toM9enp6ToWFhZKCwsLZVZWVqNF4p9m8eLFZhKJRBQYGFhbQq+oqIixb98+4+Tk5NiIiIj72upEdcXFxemsW7fOfNu2bRlAzZCnvr6+hs1mY8CAASXR0dH1gu6IESNKz549m1xQUMAcOXJk8Y4dO4yDgoLSd+zYUe8ZR90iBQAwZ86cguTk5Njs7Gx2WFgYz8DAQMPhcGh9fX2Nl5dX2b179zhbt241GjNmzFOnWsbHx/P79+/v7OnpKcrPz2fn5OSwAODnn382Z7PZtT9eYmJiOCKRyA2oyRmdkJDAXbhwobVEIhFdu3ZNPzMzkw0Ao0ePLty5c6cwOzubbWlp+czF8CUlJUx9ff16Y3cTJkzIt7CwaLd+/fra4Orq6lrFZrPh5eVVmZqa+kQw9fb2Fo0dO9axf//+pY/fJ1KkgHhR3AqFamNxZfWpr/ffFn53PJYrV5Lh5KaiUGnw07l4nfE7b+jnlymCyhWqnaipL/xeahZBV6lUYtSoUfb+/v6ZdYeatYYMGVK8atUqM6Amy9T+/fuFI0aMKBk5cmTJoUOHhMXFxQwAiI2N5aSlpT0zCK9cuTJPKpXKZs6cWajdxmQywePxNFwulzYyMlJXVlbWGwcuLi5mfPbZZ/a7d+9O1QYU7T0BIDw8XM/FxeWJoW2VSoVz584ZjB07tkQulzMAoKqqqt57LhKJ5CkpKZzH+ygAYLFY4PP5Gj6fTxcVFTG014qOjtZ1dnaulslk3G3btpl069bNOTExkbdq1SrTutd0c3OrvHjxYuLjIeE4e3t7JQCEhoY+AIA///xTFwDc3d0VMpkszsjISHXt2jVdJycnRUBAQMbj8+6PGzfuEQDY2tpWR0dH63br1u252QKio6P52qo/WoaGhur58+dn1+3pymQynkqlQmRkJM/Ozu6J9y48PFx27dq1+CVLllgCQGJiIsfFxYUUKSCex71coYqLSC6c2HPtX/wrCflN3R7isai0YvRa+xf/akL+mHKFKg5Au6ZuU1NoFkuGdu3aZXTnzh3dBQsWWAPA6tWrM/v06VPh6+vbcs+ePRlz5szJnzJlSksPDw+RRqPBvHnzcrWVgr777rvsvn37OtM0jRYtWqi0BQb+DQMDA02vXr1K2rdv76pWq6lFixZlA/8UKfD39zfNzMzkTJgwwR4A9u7dmxITE8NbsWKFpY6ODt25c+fyXr16VTS87q+//iocP358EQD07t27tG3btq3Gjx9fL8PXoEGDyr744gtDANi4caPx0aNHjdRqNbp27VrWoUMHeUBAgPHu3btNGAwG7ePjU2RnZ6fcunVrbTk/sVgs+u677+oNhy9fvjyrX79+ThqNBi1atFCfP38+CQCEQqF6/fr1WcOHD3fo1KlT0sCBAx2VSiWjRYsWKk9Pz0pXV1eFr6+vbVlZGZPBYGDXrl217+X58+cfMJlMXLp0qV7pvLrWr19vbGNjUy0Wi+UymUwHqEnNeezYMcM///zzwaZNm2pHF4RCoapv375OhYWFrKCgoCcmo3l7e4vKy8uZU6ZMyQOAP//8UzB9+nTyDUo8VbVKM0Op1qxZEhLLPXY7iwyLNENlChWmHrjN+6iDlc2qj9pEcFlMXwaDOtrU7XqTXjgN5Oucvfy++/LLL60XL16c21gv/12krab0tJnfDcnlcmr8+PG2R48eTX3NTWtyJA3kS2FUVqvWFVcoJ3+yPYKfUfRCRcKIJuZmoY8DX3au4uswF3NYzHVN3Z435d/MXn6lgMlhcV7l9Hfa02r1EjW4XC79PgRc4qXolCtUQWmFFf3H7bjBf1T5n8yrJN6AuJxSDNpwjXdoitdyoa6OvS6HNRPAO//wvVk80yXeLyKRqPpFe7kE8QyCcrnqUlRa8YCRW8N1ScB9+2SXyDF44zXdxIflE8rlqhAA7/wKBRJ0CYJ4G5mXK1TSM7E54om/3+ST2clvr9IqFUZvi9ANSyroXS5XhQFoNHPfu4IEXYIg3jbOFQpV9I5ryY7zjt7hqjUk0dHbTqHS4Kv9UbzDtzLaVChU0QDsmrpNrwsJugRBvE0kVdWqmytD40x/uZj4r9boE80bTQMrQ+N01l2QWVVWq24D6NDUbXodmlXQfVoCf4lEIlIqa57XfPrppzYrV640feLk1+jo0aP6hw4dMgCATZs2CePi4nRmzpz5RPrFV1W3UMOvv/76xBCLr69vS4lEIvr444/tVKp/Jjpfv36dT1GUWPseaR08eNDA3d3dddasWZYAcPv2be6CBQueyP4EAEuWLDELCwvjPa34hEAgaC+RSEQSiUSUl5fHBAA7O7s22m1RUVHvxLOYRYsWmaekpDT6ZT516lSrpKQk8kXfdHpXVqsuTz8YbXDoZkaz+u4i/ju7wlKZ3x6JaVFVrb6Kd3At7wvPXtYoFO4MzssvGdIoFCoG5+kzoF8kgX9AQIBxVVUVY/HixW+0NIiPj09pnb8/+vjjjx369u37WnLKPa1Qw5UrV/hKpZIhlUplS5cuNTt06FCL8ePHPwKAwMBAEzc3t8qG5xw4cEAolUplQ4YMcQCAdevWmW7atOmJmdJqtRoRERF6K1asyEtPT2c1VnzCxcWlSiqVyuqeJxQKlQ23ve1Wr1791Nywn3/+eeGGDRtMGhafIN4Ij8pqVciE3Tf5UlL39p13NjaXYjJidP192l3m67A6AEhv6jb9V1741yKDw2Hdd22Fl/3zvICtTeD/tP0XLlzQ27Vrl8m+ffvStNsmTJjQslOnTqKePXs6FRYWMmUymQ6LxRLfu3ePAwAeHh4iPz8/S6AmYIvFYpFYLBZdv36dDwAikchtyJAh9q1bt2515coVPlCTbEJ7fe3fHyfxN5bJZDpTpkyxuXbtWmJCQgJ32LBh9k87Z8OGDUKJRCJq06ZNq2PHjj01oURdFEXh008/te/Vq5dTQkKCTt19CQkJnLZt21YCQMeOHSvDw8N1AeDWrVtca2trpa6u7hMzSdhstkapVIKiKERERPCcnJwU2uIQdUVGRvK06R1tbGxUfD6fBuoXn0hKSuKJxWLR1KlTrTSamks8evSI1alTJ9HYsWNtKysr6yUjKC8vp4YMGWLv6enpMmjQIAeFQkFp38fExESdbt26ORcXFzN+/vlnY4lEIhKJRG5Xr17lA8CcOXMsJBKJyNPT00Umk+k863Ot2wOv+zlotwsEgvYAcOzYMX1XV1e3usdJJBLR5MmTrd3d3V3Xrl1rDNSMNsTGxnICAwOFdnZ2bdq1a+c6efJk68f3ld+6dUvvRT5L4j8lqqpWX5x5MFqXBNz3R+idHCrgQoJBhUJ1Fe/Q5KpmMUTzvAT+ADBnzhwbJpNJl5aWMoCanl9lZSXj1q1bso8//rgoICDABADEYnHZrl27jO7du8fR0dHRAEBOTg4rNDS0xc2bN2Vnzpx5sGzZMgsAyM3N1dm/f39aaGjog2XLlr3wcHFiYqJORkaGzrOOmThxYrFUKpVduXIlYd26dWYvct3NmzdnREdHx8+fPz935syZ1nX3tW7dWn7t2jUBAFy8eFH/0aNH2nzKZnPnzm205z9nzpy8UaNG2Y8YMaI4MDDQtEePHuXjxo2z2b59u2Hd4+7fv8+1s7Orl1O5YfEJmUx29+bNm7JHjx6xDh48aAAA4eHh8bdu3ZLZ2NgotO+/1oYNG0wGDx5cEhkZmdC9e/ey33//3RAASktLmZ9//rnt3r17Uw0NDTVz584tkEqlsrlz5+ZEREToSqVSXnZ2to5UKpVt2bIlffny5RZA458rAGh74HV73CqVCq6urlVSqVTWqlWrKgA4duxYC39//4yGPXMfH5/iqKio+AMHDhhrC0NozZw5Mzc6Ojo+OTm5dpF5dXU1yXT0ZllVVquuLTkZq0dq375/dlxPYR65lWFeLlddwDuynKhZBN3nJfAHgC1btqTOnz8/Z9GiRZYAIJPJOB06dKgEAC8vr8qkpCQOAJibmyuzs7N1du7cKdReMz4+Xic+Pp7n6ekpGjJkiFNpaSkLAGxsbBQGBgYae3t7ZWlp6ZN1955i5cqV5nPnzq0dhiwsLGRre1v5+flsADh+/Li+RCIR9e/f3yknJ6degN68ebORRCIRLVq0qN7zVTMzMzUA9OvXr1x7HS1vb+8qNze3qs6dO7uUlpYyTE1NlXfv3uXo6+urLSwsGs1k1aVLl6ozZ84ku7i4yNu2bVu5ZcsWkx07dmScPHmyRd3jGmYla6z4hJmZmZrBYGD48OHFd+/e5dVt76hRox7du3ev3qOB+/fvczdv3mwqkUhEQUFBwocPH7IAYMuWLWZyuZwhEAhqA+eIESPslixZYt2nT5+yO3fucMPDwwUSiUT01Vdf2ZaVlTGAxj/XpyksLGTq6emp6277+uuv87/55hubhnMGvLy8qlgsFqytrRVZWVn1RmMCAwPNHR0d2zg5OZEavk3DoEKhurbxzweGf9zKbBbfVcSbtyI0jhOWVOBWLlcdBfDC39PNVbP4h/y8BP4A0LVr18qxY8eWJCQkcO/du8dxdnZW3L59WxcAIiIi+A4ODrVfjF26dCn/+++/+drem0gkqm7Xrl2Ftkd07dq1BABIT0/nlJaWMlJTU9kCgUDd8J6NyczM1GEymbSTk1Ntz1D7bFMqlcpMTEyUAPDzzz9bXLp0KeHs2bMPGIz6b/O0adOKpFKprOHzQ21xg5iYGI6+vv4T7Vm7dm3OjRs3EoRCoXro0KElt2/f5j0uROAsk8l448ePt22szVu3bjWZPXt2QWVlJYPBYEBbfEHLzc1NnpqaqgM0XnyitLSUoZ24df36dT1HR0eFXC6ntAUa/vrrL7267//j91zu5+eXK5VKZTExMfHz5s3LB4Cvv/4678cff8z08/OrreR07Nix1F27dqUEBASYtm7dWt6jR48S7ftZNxNVw8/1aSIiIvgikahecQRHR8dqBwcHeVRUVL2e7o0bN3gqlQqZmZkcKyurej9eZs6cmZuamhobHR3N1wZkHR0dsj7lzWCVy5WnQv7Ostx6JalZ5IgnmgZNAzOConkJeWU9KxSqLQDe6tGmZvGPubEE/unp6azNmzcbr1mzpl5g+umnnzLnzp1rdebMmeRdu3ZpxGKxSE9PTx0cHJxSWFjIBICpU6cWTpkypfD8+fMCALC0tFQNGDCgpFOnTiImk0l379697Oeff84xNzevHjNmjF1qaipn48aN6QBQUlLC0j6rTUlJ4e7Zs6derzA2Nlb3yJEjySqV6pkffL9+/R55enq6dujQoaKxANoYHx8fh5KSEiYA/Prrr2lAzfPkDh06VHl7e1d6eXmJmEwm3aNHj1JtgQVfX99HQM3zyf3796c1vObZs2f1OnfuXMHj8ehx48YVenh4iHr06FFvGN/T07Nq6dKlOkDjxSf4fL5m0qRJdjweT2Nra6sICAjIzsnJYfXr18+Zz+drDAwMVEePHq2XYWr27Nn548aNs9u2bZspTdNYvXp1FgAYGBioP/zww4rdu3cLL1++zD979qzBlStXBNXV1Yyff/45w8vLq+rEiRNKiUQioiiK/vjjj4uGDBlS2tjn2hiZTKazceNG071799Z7L7799lurlStXPjEB6tChQ4Z+fn4tx48fX8jlcusF1MDAQPP9+/cbGxkZKc3NzVVSqZTXsWPHJwpbEP+9CoVq873sUvHikHskfyyBarUGE3ZL+SdndB3HZFApXDbzp6Zu08t64YIHr3v2clMQi8Wihj2fupKSktibN282CQgIeOdnqy5ZssSsX79+pV26dHlvssVLJBJRWFiYjM1+sVVAU6dOtfLz83vo5OT02vINkoIHgFKtmZZXKvcfuOEav1T+XtQAIV6QuT4Xp2d2rTTgs6ewGIz9Td2el/HCQfRVAyaD8/b9YLWyslJ9/fXXBc8/8u23YsWKvKZuQ3O3ZcuWrOcfRbyiD+RK9c9jt9/gkYBLNJRbKseY7ZH8E9O6/MbSYUQDuNfUbfq3XrinSxDE6/ee93QNKqtVD6buv238VxMXn9fIy1F4NhDVBTXLQ40HzoKqrBAl14OgLMyA+WcB4Fg4v/C5HKtWqM5LRuH5zaDV1aAYTBh9+DU4liJUyMJQcu0AGDw9mIz4HkyePpTFOXh0dS9Mhs1/Y6/5bTJG0lLz3UC3B3pcVlsAz5zj0dw0i2e6BEEQ5XLV1jOxOXpNHXABoOjP38B1EMPko0Wg1UrQSgUYHD2YfLQIhec3/etzAaD4r91o0WUMeI6dUJV0E8V/7Yb52J9QJj0O80/XouL+VVTEXYG+eAgeXduHFt3Gv4mX+lY6KM1gDGxrYS22NfyBr8Oa19Tt+TeaxexlgiDee0Mqlaphy0/ea/K1mBpFJeQZ96DXri8AgGKyweDqgW3cEmyh9UudW7u/urL2OKaesGYjxagJzioFKAYT8oxYMHUNwTayeuL6xD/8DsfwVRp6OgDvpm7Lv0GCLkEQTc24slq1d3pQNL+i+oUm+r9Wqke5YPL1UXjmF2TvnonCs4HQVMuff+JzzjXqPRnFl3cjc8sEFF/eCcMevgAAgy5j8PDIEshT/4auWw+UhB+GQZcxr+31vSvyyxWY+8cdXoVCdRTAW5MprlkE3aNHj+prk0uYmJi027dvX71lOiqVCpMmTbLWpvALDg7WB2qSTIwdO9YWqFkqol3qc/fuXY5YLBaVlZUxJBKJyMPDQ6S9/n/R3kWLFpl36tRJJBaLRRs3bhRqt4eFhfG8vb1dPDw8RN26dXO+d+8ex8/Pz1IkErk1LBbwLPPnzzc3NTVt11hRhbKyMsYHH3zgJJFIRL1793bUrpXdtGmT0MvLy0UikYgaJuyfNWuWpbu7u6s2k9TatWuNz5071+g/0o8//tiusrKSioqK4nbs2NG1Y8eOtQUTiouLGb169XLq2LGj66ZNm4SNnf+2CQ8P561fv964sX0RERG8xYsXv1A2MeLllcmVuw7dzOA1lxSPtEaN6twkCDoMhOXngaDYHJRG/vHK55b9fQaGvSfBeurvMOz1JQrPbgAA8Ow7wGLCBpj6LEVlQiR4jp2gKsxE/vHVNUFb+WIB/310/l4uLt7PMyxXqDY3dVte1AsHXZVS7Q5A/LJ/Hp/fKB8fn1JtMgQLC4vqoUOHltbdv27dOhMTExOVVCqVXb58OfHHH3+0yMnJYU2bNq2IwWDQdQNAeXk5NWHCBLvdu3enarMehYeHyxqmCnxZR44c0U9NTeXcunVLFhkZKTt+/LhhVFQUVy6XUzNmzLA9dOhQ8s2bN2UHDhxI0SaUWLNmTYb2/tosTs8yffr0gp07d6Y0tu/YsWP6nTp1qpBKpbJOnTpVBAcHG6SkpLCvXr2qFxERkSCVSmX29vb1lrTEx8fzpFKp7MCBA8Kqqirq1q1buv379y9veO27d+9yDA0NVXw+n964caPJqlWrMm/fvh1/8+ZN3YKCAuaGDRtMRo0aVSSVSuN///33J9Imvo28vb2rZs+e3ehkQS8vr6rIyEg9tbrpe1/vKo2G/uRRpbLXmrPxzWZ5A0tgDKbAGBzLmt/ofFEXVOclvfK55Xf/BN+lZiSU79oVipyEeudqlHKUx/4JQYdBKL66B8KB30DH3AkV9/76j17Zu+n747FcuVLtA2BQU7flRbzwRCoWm8na/NWll77RtF97PfdecXFxOsbGxioDA4N6SflDQkJaXLhw4QEA6Ovra8aNG1cYHBxsMH369MIdO3ZkSCQSkZGRkQoAfH19bSdOnJjfrl2756buk0qlvOnTp9tkZWXp6OnpqQ0MDNRSqVQ2cOBAh/z8fLaOjo7m1KlTSUZGRrXtOXLkiNHSpUtzAIDNZmPWrFl5Bw8eNOzZs2dZ165dS7VZnGxsbFTav/9bLVu2VMXENL5Cy8XFRXHr1i1doKbggImJiSokJERfrVZTXl5eLi4uLlU7d+7MYLH+ebspiqIVCgWlo6OjWbduncmUKVManaly7NgxA29v73IAaN26dVVxcTFT+8OBy+Vqbty4obtjx450FouF1q1bV965c4crkUhq1/UGBAQYHzhwQAgAGzZsyOjatWuldi30yJEj7Xx9fQtbtWqlGD9+vF1JSQlr8ODBxWvXrs25dOmS7oIFC6xVKhXl6+ubP2vWrEKJRCLq2rVrWUBAQHZAQIDx+vXrzTMyMmJHjhxpl5iYyOVyuXRycjL3r7/+im/Tpo1Cu72srIw5e/bs3AkTJhR/+OGHTiqVikpLS+P+9ddf8ZcuXdI7d+6cQWlpKZPD4WhOnz6dfPHiRb0LFy4IAgMDswUCQXtXV9cqpVJJnT179oGFhYXK0dFRERERwe/atesTVZyIV2ahUGl+mxZ0W1eheqIOR5Nh6hmCpW8MZWEm2EJryNNiwDa2eeVzmXpGUGTcBdemXc12w/oDWaU3gqHfaSgoJgu08vGEXIoBWkWykD5LmUKFaQdu83d/7rGPr8NyAdCsV9w0i+FlrUOHDhkOGzasuOF2hULB0NPTq13bZGNjU52Tk8MGaoZb2Ww2/d1331lfvXpVPy0tjSOTyepNxvD29hZJJBLRxIkTW2or5ACARCKpkkqlstGjRxdqe6MAcPjw4dSbN2/KRo4cWbx79+561S3y8vLYtra2tT1JOzu76pycHHZWVhbbwsLitSVN0GrTpo0iKipK18nJqfXff//N79OnT3leXh67urqaioiISODz+ZoDBw7UG54fOnToo08++cRu6tSp+XFxcdwHDx5wxo4dayOVSuvlS05MTOQ6OjpWA0D//v3LFi1a1NLBwaGNRCKp0NPTo0tKSpiGhoZqoCazlDYDGPD0ohJAzfB29+7dy4YOHVrm7OxcfePGjYTo6Oj7N27c0AOAJUuWWJ47dy7x1q1b8YcPHxZqe9CZmZlsjUaDyMhI3brvbVBQUIpUKpX16NGjtryiWq2mjhw5kjx79uxcAIiKiuJaWlpWR0VF1TtOX19fff369URPT8+KvXv31nufXFxcqm7evCkbMGDAo7i4OA4AODg4KGJjY5t8cs87iCqTK4N2Xk/m3sl8LVUyX4lRn69QELoW2bumo/phCvS9RqEyIRyZm32hyI7Hw6PLkXd4MQBAVVaIvD+WPvNcABAOmIHiSzuRvWs6Hl3dC6P+M2rPUZUVojr3AfjOngAAfclHyN03BxWxf0LX7YM398LfUjdSinBQmqFbJlfuaeq2PE+zCrpnz55tMXr06EcNt3M4HLq8vLx2KDM9PV1H+yW8ePFii0WLFuWsWLEiy9zcXBkWFpZw+/Zt3fT09NqunnZ4mcFg0CEhIc8ss6dSqTB16lTrTp06iX777TfT7Ozses9HzczMlGlpabXbUlNTdSwsLJRWVlbK7OzsZ1YeauhphQ+eZcuWLcLevXuXPHjw4F6/fv1Ktm7dKjQwMFB37969DAD69OlTFhcXVy9ITJ06tejMmTPJV65c0ZsxY0b+yZMnW+zcuTPD39+/3vPKumu2FyxYYLVnz57klJSU2Hv37vFkMpmOgYGBuri4mAnUVAsyMjKqHXd9WlGJ+Ph4/sWLFw14PF7tr53t27cbOjk5tfnggw9Ktcf079/f2dPTU5Sfn8/OyclhAUDXrl3LN27cKLS3t3/uT/3S0lJm3RESLy+vKqVSyXBycmp95coVA+329u3b15ZHfPDgQb33KSEhgScWi0V//PGHsF27dnLte0JRb/0oenM0+mGZwmPDn4kvlg7sDdMxc4CF7y+wnLgJpiO+B5OrB76LN6yn7YHtnBNoOWM/zEavBACwBEKYfbz8mecCANe6NSwmbIDlxE2w+CwAHHOn2nNYAiFMff4J3LquXWE5aQvMx/8MJr/2ny/xDP7n4nWqlOoeAHo2dVuepdkE3fT0dBabzdaYm5s/8QBtyJAhxatWrTIDapLv79+/XzhixIiSlJQUdnx8PHfkyJGlEomk0sXFpYrD4dCLFy+urUZUl6Ghobph3deGIiIiaksGfvnllw9pmq53vI+PT/GPP/5oDtQUB9iwYYPZ6NGji3v37l0RFhYm0Ab7rKwsVkxMzDOfUz2t8MGz0DQNbbAzNjZWlZSUMLt3715+584dPgDcvn2bb29v/8Ri8eLiYkZqairHy8urS/a1cQAAIABJREFUSi6XMxgMBqqqqup9/i4uLvKkpCQd7X2MjY3VTCYT+vr66pKSEmbnzp0rQkND9VUqFWJjY/nawAQ8vaiElZWVIioqKn7Pnj3G2p7xl19+WZyenn738OHDQgBwc3OrvHjxYqJUKpXFxsbGaZ9Jjx8/vnjVqlVWX3zxxTNn2CiVShQUFLDqVlvicrm0QCBQHzlyJKluT1f7PkVHR/McHR3rBXMXF5eqqKgo2cyZM3MDAwONASAlJYXTunVrMpPlv8UqV6jWLT4Rq6tUk/oRxH9DodJgxak43TK5ciuaUWxrqNk07PDhwy0GDx5cr5fr6+vbEgDmzJmTn5eXx/bw8BD17NnTed68ebmWlpaqhQsXWi5fvvyJvMj9+vUrz8vLY8fGxnKAf4aXo6Oj+SNGjChteHxd7dq1k6empnK6devmfPPmTd2G+z/55JMSGxsbhVgsFnl6eoqGDRtW7OHhIedyufTGjRvTPvnkEwcPDw/R6NGj7dls9kt9o6xfv954/vz51sHBwcJPP/3UBqiZMZ2SksKeNGlSUXBwsKFEIhEdOnTIaNKkSYXe3t5VPB5PI5FIRFFRUfwJEyY8MUT/008/mfr5+T0EAFdXV7mHh4do7NixhXWP+eijj0rCw8P1AGDhwoW548ePtxeLxSIdHR1aIpFUzZw5M//QoUNGHh4erp999lkBj8erfX11i0p07tzZ5fvvvzcHAIFAoOZyufSSJUty/Pz8LC9fvsx/PMPbdeDAgY8AYPny5Vn9+vVz6ty5s8vQoUMdtNcUCASajIyMuy4uLs/MOPPFF1/YzJ49u14aywsXLugaGBioO3bsWC9gFhUVMbt06eIcHh4u+Oyzz+q9TwkJCTyJRCL69ddfTfv161cGAImJiRwvLy/yPPc/pKHpzxLzyvTDkwqffzBB/Auhd3KQ9ajKCkCzXXP1wmkgVUq1O4vNfOkMViqlWsViM5tVwQPiSaNGjbL9/fff0/l8/jvXBQkMDBSqVCrKz8/vhSZaRERE8EJDQw1WrVr1wiMRr+o9SAPJqVCoMj7bJTWJSnvityFBvDIPO0Ps+VxSwOewWgJodqNU/2b28isFTBb7ra89/F44cuTIE+UB31deXl5VXl5e703VpTdBpdZMisl8xCcBl3hdbqYWQ5paxPN2NJ6pw2L4N3V7GiIFDwiiGXnHe7r8ympV5qhtEYaxWc98ykMQr8TFTA8h07qW8nSYlgCaVQ3sZvNMlyCId1u1SjPtRnKRDgm4xOuWkFeO6w8KmNUqzbSmbktDJOgSBPEmCNQaevGPZ+8/MTmRIF6Hn8/H66o19HcAmtW/ORJ0CYJ47eRKtd9l2UNWQt4T2UcJ4rVorr3dZhN0n5awf+XKlaYSiURkZWXVVls44PTp0822osTWrVuNOnTo4NqzZ0+noqKieu9vYmKijqenp0unTp1EO3bsMARqiiS4uLi4WVlZtW3seqNHj7Z1d3d3vXr1Kh8AvvnmG8u4uLgnknBoNBoMHz7cHgAWLlxo3qlTJ1GbNm1aabMuhYeH8zp27OgqFotFTyt28LYhxQreGgY0jbn+5+J5zz+UIP476y7IdFUazQI0o9rxL17woLr61Qoe1JzfqGcl7F+8ePHDhqkaBw0a1Cx/LisUCmrHjh0mUqk0fuzYsYXr1683qbt/xYoV5j/88ENWZGSkbPfu3cZKpRJubm6KW7duxZubmz+xFjU3N5fJYDBw4sSJpJ07dwqzsrJYFRUVDDc3tyeOPXnypMDDw6MCAJYtW5Z369YtWVhYmGzdunXmAPD9999b/Z+9+w6Pour+AP6dsjtb0zshpJBsChBCSOgEBOmhFylKFwsCKoKdovJaUOwoKMiPIiIgoYmCClICCUUECSEJKZAE0sv23Zn5/REChBogyabcz/P4vL7slLNLzNl7595zfv7557T9+/envPfee563nt8QkWYFDcbooxcLkVFItjsTdev8lXJcKjKwAPrYOpZK1d8yJJWyH48Z9NA3evmnnXe91/0K9t/JG2+84fHrr786SKVSYe3atRmBgYHmWwvWm0wmasyYMX5Wq5UKCQkxrFu3Lmvnzp3qTz75xN1qtVJms5mKi4tLKykpYebNm9csLi4uPTk5WVr575MnT25+9uxZhSAI2LhxY3pgYOA9izScOXOGCwkJMUgkEsTGxpZNnDjRF8D1og2ZmZlcdHS0gWVZuLq6Ws+ePSuLiIi46z4yjuNEi8VCabVaWqlUCkuWLHGfP3/+1TsdGxcX5zBlypTCyvMAQKfT0cHBwQYAKC0tZQICAiwAoNfraa1WS91cz5o0KyBqS6nBMnP9scx69VyNaDr+Lz5DNa9f8Ex7uWS3rWMB6sn08v0K9t8qKyuLPXDggPrkyZPnFy1alLNw4UIP4PaC9R4eHtZDhw5dOHHiRHJ5eTlz5swZDgCMRiP1999/p0ydOjX/s88+c73bfT7//PPsxMTE5Lfeeivn888/v+txlYqKihg7OzseAJydnfnS0tIqm5Nbtmxp/PXXX9Xl5eX0qVOnlDc3DLgTR0dHoXXr1vrXX3/da8SIESWiKGLz5s0OEyZM8ElLS6tSs/bixYtcYGDg9bKGEyZM8AkPDw/t2bNn+bV4rImJibKcnBw2JSVFXlhYeP1bDWlWQNSiQJpC4P7kOza2Iohat+N0DsWx9GMA6kUP8HqRdO9XsP9WKSkpXFhYmAEAunXrpktPT+eA2wvWX716le3fv39AdHS05vjx46qsrCwJALRp08YAANHR0fq0tDQOAA4dOmQXHR2tGT169PUyhAsWLPCIjIzULFiwoFllV6NK06dP946OjtZs3rz5egMFJycnvqysjAGqJuBKixYtyl25cqVLbGysf0BAgNHLy+u+XYkWLFiQt2PHjvSffvrJ8bnnniv4999/FQsWLLjyySefuN183K01otetW5d1/vz5/5YuXeoJAEuXLr08Z86c5pMnT26h0WgMHh4e1+sUk2YFRG0xWvgpW05kM1ah0RU4IxqIMqMVf57P43lBHG/rWIB6knSrU7D/ZoGBgaazZ8/KAeDgwYNKX19fE3B7wfpVq1Y5DR48uCQhISE5MjJSW5mYKs9NTExU+Pv7mwCga9euZQkJCcmbNm26CFQ8Tz106JD6xIkTyQsXLsy5tYjIypUrLyckJCSPHDny+qbD1q1bm86fPy+3Wq3YsWOHXXR0dJVnz82bN7fu27cvbefOnWlSqVQIDg6+5/uslJycLJXL5YKLi4uV53nQNC1qtdoqf3cBAQHGlJQUDgAMBgMFAEqlUlCpVDwAtGnTxnT48OGUH374IbNZs2bmyilogDQrIGoNzQvitE3HLz1Q9y2CqGnrj2UqdCbrLFvHAdSTFV2dO3c2/PDDD8K1ZvSWt95662pWVhb71VdfuXzwwQe31b318fGxdu/evTwiIiJYIpGI69atSwduFKzXarX0ihUrMimKEidPnuy3ffv2KlOREolE7NatW6DJZKK3b9+eeus0MAC4urrySqWS79ixY1BoaGi1SgFyHCdOnjw5PyoqKtje3t66ZcuWdKCiccOaNWsubdy40X7ZsmXuDMPggw8+uEzTNFJTUyVPPfWU34ULF+SdO3cOWr16dYZGo6mSjD/44AP3JUuW5Hh4ePCCIFDDhg0L+OKLL7JuPiY2Nrb04MGDym7duumnTp3aPDU1VW6xWKjKKdtly5a5bNy40UkulwvffPNNlXNvblbAMIzYvXv38o8++ij31mYFkyZNKnzjjTe8jUYjfWuzAkEQ4ODgwP/2229pwI1mBTcn9zt5mGYFHMeJu3btSvvjjz+ur8K+9e8eIM0K6oEe+eUm7lwuKYZB2NaRtEJYeMETQFsA/9gyluo3PDCbw1mp9OEbHpjNVlYqtXnDg507d6orF9/YOpaaJAgChg8f7rdt27Z0W8dSGxpCs4Ka0JjKQJYZLT9/ti9lxPeH0skcP2FzLz0eZJ3a1e87Jcc+a8s4HmT18qM1PJCSGabadG1rUaNMuA+DNCuwORXH0IPi/skmCZeoFzYdv8Q+3d1/AoA5AO671qS21Ivp5bo0aNCg8kGDBpXbOg7iwcyaNYs0X21YRiZkFFkLtNVatkBcQ1GAt6Mc3o4KeDvK4ariQNMUKAC8IKJQZ0ZOiQHpBTpcLibfKR/E5WIDzl8pF9s2d4gFsNlWcTS5pEsQRO0r0Zuf3XAsq1FUPqttPk4K9ApxQw+NW3lUC0epCOjMViGLopCq5Nh0ioKVAiUKoijRm/kWgiAGcCztz4uiLCG9iD9wIV+15+wV5JXbbPDWYPyUmKX2c1GOt5dLSNIlCKLRUCikbNsDF8je3LuRMjQGhXtiWlf/cj9XJXhejFPJ2N0ADgDIUXK3n8OAgr28yqaFFr1C3LtF+TkNen1AyOB/LpVY18Znqvf8dwU82aJ1R/FphWBpqjsACoBNPiSSdAmCqGldU/LKDXozTxZy3IKmgOHtvMXX+gcbWIb+x14u+RDAbkhw3z37d5AJINNOJlkHQNHR33lYqKfd3DcGhgQu2Z2k3HUmF/dYJ9skZRTqYRVEOQA/ABdtEUO92KebnJwsdXZ2Do+OjtZ06dIl8NbXR4wY4dumTZvg6OhoTXR0tCYrK4t8WXgIGRkZko4dOwZFREQEb9u2TX3r6xMnTmweHR2tGTVqlK/VWrFttlevXgFqtbrtnY7/8ccf7cPDw4Nnz57tBQAnT56Uvfrqqx53uvfbb7/tfvjwYXliYqIsIiIiODIyUjNy5EhfQRBgsVgwaNAg/w4dOgQ988wz3jX8tm1m4sSJze/22tChQ/0EQbjbyw2a0cL3/SMpj0wt3yKyhSMOvNJT93Zs6D/OKu5xe7mkC4A44KES7q30ANbbySXtvBzkw/43vHXSvhdjdK2a2d33xKbm6MVCHkAPW92/2klXtAiP1PDg2vl3VVmc4vDhwyl3er2yfGBCQkKyj4+P9U7HEPe2cOFCj3feeSd7//79F5YsWeJ182sHDhxQWCwWOiEhITk0NNSwceNGBwBYtWpV1vTp0/PudL3169c7JyQkJCclJckB4OOPP3abP3/+bcfyPI/4+HhVly5dDG3atDGdOnXq/IkTJ5IB4ODBg4q1a9c6tm7dWn/s2LELBoOBio+PbxTdaNasWXPpbq9FR0drt2/fftsXmcbAZBViD6UW3LPEaVMiYSjM76cxr50aXdrcSfGUnUwSCeBILd1OBLBXLZOE+bkqn/l5Rufylx4PsjA0WURe6UByvqrUYOlvq/tXO+lSEpq9/OpBPOw/lIS+5+g0Pj5eHRkZqVm0aJHbvY6rZDKZqE6dOgW1b99e07dv3wCr1VplxBwdHa0BgMmTJzePiorSREZGalJSUqS//PKLXXR0tMbFxSW8TZs2wYMGDfIHgEmTJjVv3769pmfPni0LCwuZm68VHh4ePGTIED8A+Prrr53Cw8OD27VrF1yZHHx9fVtFR0drwsLCQk6fPs0BwNy5cz2jo6M1HTt2DEpOTpYmJydLO3ToENSrV6+A8PDw4PPnz0vT09MlXbt2DbRarXj++eebrV271sFisSA2Ntavffv2mtjYWD+LxYKdO3eqvby8WkdERATHxsb6AcDmzZvtoqKiNK1atQr58ssvq1VT9Ny5c4pevXrp7O3tBZVKxRcXF1//+79w4QLXunXr62UWjxw5ogSAFi1a3PVbuEQiESwWCyiKQnx8vLxly5YmR0fH24ZvR48elXt5eZmBG80YAEAqlQp+fn7mtLQ0Ljw83AAAbdu2NRw8eLDKKOns2bNcly5dAqOiojTz5s3zBCpmP86ePctt2rTJbty4cS2AinrT7dq1C27fvr2msLCQ0Wq1VGxsrF/Hjh2DBg4c6G8ymajPP//c2dvbu7UgCCguLqaVSmXEzp071ZWfcXR0tCY4ODh01qxZXgBw859X/ky9+uqrHq1atQq5+TiNRhMaGxvrFxYWFnLgwAEFAERGRmoqY23Tpk1wSEhIaGVLx759+5Zv27btnjXGGyhHmYRucSqr2NZx1AvOSiniZnbVTejY4pBCymoAbEXdPEsUaYpaJ5cywVO6+B7fOL2jTsWRCUIAOJZeCJametjq/vVietnHx8eSmpp69ujRo8l//fWX3bFjx24b6YwbN86vQ4cOQSNGjPDVarWURCIR9+7dm3L8+PFkjUZj2LFjhx1wY8SckJCQDNzetGDYsGFllQX3N2zYkL5z586LBw4cUOj1evr48ePJo0aNKvrkk09cb75WZWlIq9WK5cuXuyckJCT/+OOPF19//fVmAODs7GxJSEhInjVr1tVjx44pEhIS5Dk5OdKEhITkr7/+OmvRokWeAJCfny/Zs2dP2rJlyy4tXrzYw8/PzzJs2LDicePGtcjKypI++eSTJWvXrnUMDg42Hj9+PDkkJMT4ww8/OALAyJEjC0+dOnXeaDTSPM+jb9++2sTExORTp04lrV69+o49ZW91rYQkAECtVvMFBQXX/ysMCwszHjx4UA0A+/btsyspKbnvf6Fz5869Onr0aL/hw4cXf/75524xMTHa8ePH+6xcudLx5uOSkpJkvr6+1/eOrF+/3j4wMDCsoKBA4u7uzgcHBxv379+vBoD9+/eri4uLq4yS5s+f32zNmjUZiYmJyUlJSbLKZg/Hjx+Xr1y50nX16tVZQEW96ZMnT54PDg42ZGZmSj777DPXQYMGlR49evRC9+7dyys/y+bNm5t27dqlXrNmjWN4ePj1Up0jR44sTEhISF66dOn1ESrP8xgzZkxh5c8TAMTFxTkdPXr0/M3HXblyRbpu3brMnTt3pi5cuLDKLAJQMVOzZ8+elD///FMNACEhIaaUlJTG2IyhQ1JuucHCk4eJPk4K7JzVVe/nrPxSLZP0xk0dx+pQjkom6RbqZffTjhe66Nzt7rBCq4lJy9eBF0QlAF9b3L9eJF25XC7a2dkJEokE/fv3Lz116tRtSXfDhg3px44duxAREaH/9ttvncvLy+mxY8f6RkVFaXbs2OF4+fJlyZ2ufa+mBZWSk5O5iIgIPQB06tTpehOEW+Xk5LDe3t4mjuNEjUZjrmxuUFhYKImKitIsWrSoWUxMjO7ff/+VHTlyRB0dHa155plnWpSXl9MAEBwcbJBIJOjUqZM+IyNDBgAzZswo3L17t9Ps2bPzACA1NZWLjIzUA0B0dLQuNTWVA4DNmzc7azSaUI7jRIqicPjwYUXnzp2DOnfurElJSanyeVWO5qdMmVLlmSLD3Mhl5eXljLOz8/Vp+s6dOxtCQ0MNHTp0CCorK6Pd3Nzu+5ypS5cuht27d18MCgoytm7dWv/111+7fvfdd5duLbt5a9Wz8ePHl6akpPzn6elp/umnn+zHjh1bYjAYqE6dOgVxHCe4u7tXuXdaWpps3Lhx/tHR0ZqUlBR5ZmamFACWLl3qSdP09dFzYWEh07p165Bz584pfH19zUlJSbKvvvrKLTo6WrNhwwbnvLw8FgAGDRpUsnXrVocjR46oOnbsqLvXeywpKWHUanWVxhUvvvhibkBAQOu5c+de/3x9fHxM9vb2gp+fn6Xy5+Jm48aN84uKigqJjY29Xku6MTZjsPBC5/i0gibfxi/AVYUdM7sYXFTcPLmUeRU2Wil7Da/k2Gme9vL/7Xyhm54kXiAho4gHEGOLe9eLpHvzNOeRI0dUQUFBd91w5ujoaDUYDPQvv/xiFxgYaEpMTEweNGhQyZ3KWd6vaUGlwMBA08mTJ5UAEB8ff70Jwq28vLysly5d4kwmE5WcnCy9qY2fJTExMXnFihUZ7777rkdYWJgxJiamtHLEvXnz5gwASE5OllutVhw9elRe2aThlVde8Zo/f372kiVLPIGKdnTHjx9XAEBCQoKysrD/yJEjC5OTk89xHCccPHhQ8eGHH3qsXr0649ChQxduTQqVo/lVq1ZVeaZ4rTmBsqysjNZqtYyTk1OVqeClS5fmHjt27IKzszM/ePDgUlTT8uXLXV988cUCvV5P0zQNo9FY5ecqNDTUmJGRIQVuNGMAADs7O0GhUAgsy2LNmjWX4uPjLzAMg9jY2CrFegMCAow///zzxcqmCjExMbpr982MjIzUVU7ZOjs782fOnEnq1atX6ZYtW+w1Go3xpZdeupKQkJB8+vTp8/PmzcsHALlcLvA8T908+r6bEydOKG5tmtCsWTPLE088UXDzSDcrK4srKyujMzIyJLf+fQAVXxozMjLOvPvuu14AkJSUxLVs2bLRNWPQGq29T2QWN+l5TDc1h59mdNSrZJLnJQz9la3juUaUSZj31DJ2ycanO+nUTXyqeX9yvqrMYOlni3vXi0/+999/Vy9evNhLKpWKHTp00D722GM64EajAKBipCCTyUSWZcWtW7em6XQ6+sMPP/Q8deqUQq1W84GBgbf9Aqtu04KePXvqV61aJURGRmpUKhW/ZcuW9Dv1umVZFjNmzMiLjo7W0DSNL7/8MhOoGOlGR0drysvLmUWLFl3u1KmTYdu2bZbo6GgNRVHiqFGjimJjY8ucnZ2tffr0aVlYWMhu2LAh/e+//1bk5uZKly9fns3zPPXxxx+7zJw5s3D48OF+7du317i7u1veeeedK3v37lVt3rzZ+fDhw2qWZcU2bdoYBw8eXDx48OCWYWFh+ltbCN7NW2+9dWX8+PF+RqORfvvtt3OAiprGERERhs6dO+s7deqkYRhGjImJKav8O5g0aVLzffv2OezZs8c+NTU1f+7cuVVqH//666+qDh066ORyuTh+/PjCqKgoTUxMTJWKXx07djQsWLBACgBbtmyx/+yzz9wBwN/f3zhs2LCy9PR0yZgxY/xpmhbHjRtXGBAQUGWk+8EHH2Q/+eSTvmazmWJZVty1a1caADg6OvILFiy42rNnz5bDhw8vHT9+vG9RURHLMIz40ksv5avVan78+PG+3377rZsoiliyZEl25TVXrVqVRVEUXnnlldumgitt2LDBvqCgQDJy5MjrX0BMJhP1wQcfeO7atSvtwIED10d0Hh4e5rFjx/pmZGRwtzajACp+fgVBoAYOHFgCAHv27LF7kC82DQSt4Ji2/1wqsXUcNsOxNNZN66BTceyHDE2ttnU8t5JJmCVuaq7F6slR48esOKpoqvt5j10sBE1Rj9ni3tVueCBahPD7LYa6F9EiWCkJbfOGB7aSnJwsnTdvXrO4uLgmWR/57bffdu/bt29Zly5dGmXtusjISE3liuzqGDJkiN/WrVvTb57yBxp8wwNNQbnpePv39jXZ7UKLBoeZRrTz/l0lY4fAtlPK98KUGy0HVx68GPX5H6n1YuBlC0nv9DPLJUwzANVqolJTqv2BP2rCpCT1YiabsJHFixfbYhFJvdVIv3yFXMgrr9asS2PUyd8Zo9p76xVSdhLqb8IFAF4tk4x+NqZl0p/n81Rns5tm68WrpUajr4vSF3WcdEkmrCMajcbcSH/REgAeZJTbiDXPLNA3yVU6HEvj87EReoWUnQCgyNbxVMNlTkLP+GpcO11T3cN7qVgPAC3q+r73S7qCIAhN82+EIOrYtf/WGmyZKpOV98sq0jfGbVD3NaFjC4Fj6XgAu20dS3XRFPWjk1J6fmQ77/o8Kq81F/N1MtTDpHs2Pz/fniRegqhdgiBQ+fn59gDO2jqWh6U380E5pY3ykf09KaQMZvcKNNnJJS/bOpYHJKplkpnz+mkMEqbp/YrPKtJLdSbrbWWHa9s9n+lardZpV65c+e7KlSutQKaiCaI2CQDOWq3WabYO5BH45pQ0vaT7RLSPQFH4C0BDXCh6VMLQZwa09uwQ90+OrWOpU5eLDTBbheA7dXSqTfdMupGRkXkABtdRLARBNGBShvbMLW10W4/v66lOLXRqmeRjW8fxsOzkkk8mdvb9Lu6fnEZZC/xuskv0oGmq3k0vEwRBVAcjkzD2V5pY0g31tIOrijMB2G/rWB7B9lBPO9rbsVH0Gam27GIDZBL6jl3RahNJugRB1ARPrclqsjaxYgtDI5pZWIb6Hg14ARwAIy+Im/u38mhSf3nFegsoUCyAOu1/SJIuQRA1wSevzFgTfWEblN4hbgaOZXbZOo5HpeTY3T00buX3P7JxyS83GVDHK5hJ0iUIoiY0v1RsaFJLYJVSBs2dFHIACbaOpQb83c7HkWtqW3YvF+tFkKRLEEQD5JFTYpDaOoi6FOplB53Jmgbgrg1aGpArVkEo8XNpWhU8y4xWGkCdLiAjSZcgiJpA8YLYpMZJ3o4KAEixdRw1xcKLmc2b2GIqCy9QAO7Y8rW2NNli1wRB1CiBboT9ge+lmYNcVEjZRlP+k6WplGaO8ugHOcdw8QSK/lgBCAJU4X1g33FUldetpXko/PVT8Poy0DIVXAbNBWvnAgC4uultmHKSIfMOhdvIBdfPyd/xESz5mZAHRMExZiIAoOTwj5C6+UER2PFR32YVZmvdJ10y0iUIoiYINF2vi/zXOG9HuVHK0re1cWyo1DI2xdNeXu1V2KLAo2jvcriNWgSvaV9Dd+4AzAVVP47iv76HMqwXvKZ8CfsuY1Hy95rrr9lFD4fLoJeqHG/OqyhP7zXlS5gu/wfBpINVWwRz7oUaT7gAYOEFGkCdPhYhI12CIGqCwFBUk0q6EoYW0Tie5wIAKIoSB7T2oHxdFNU6/vzp49gQosHi5/oDAH42PwEgC6PGDbl+zHM/52PxnHFwcfeCKLbFmC7/w5fjIq69GoEziUewNe8IFlz7s0sXVVibuRuvPhGOV3ZxeHdMBL77aCEGvPcuAkJa1eTbBQC0aWbPoY63DJGkSxBETWhy08tCRS/yxjRbyPmqZZRbevXaMxZfvkI5CaBejO1G8TyPtprWoouTC3oYqeuj5dbNWtDvTBkNOcdRvCDAoNOi9ZVi3tnBCUNnjaXjTydSCrkcPYxUxT29grAsNZkZ06Elolu3F71SMgQPi0DaaAR2AAAgAElEQVTn/7lXdM/KxqCYfjX6xY7jKQpAna4eI0mXIIiaINJ00xrpCqJIoXElXSdaxkKmcaJFMy8KRqsoGKwQDVaR11spUW9lBKMVgtEK0cjDmFGKY/8kYt/LG+DOOaPL0hFUdPM2KNmaylRe0JSjhdIigWDm4evggQw2E7pfsxiGK8I036GI5ILw/fGfr5+TlJeGMJU//nzxBwxfP5N6Y8GbzKsxT2Pm5sVUiGsAsv5Iwri2sTX2hh2GtjRK3BT5NXbBaiBJlyCImtDkWpFdLTNxvCB6NJZ+tBaT0V5fVoZzB/6gBFGgIIqQyhWQq+3AKZWQeajAyRW8lFOLEikHU56UkiplTPs3+osUQ1OtDrWFngK8FneGaOEhmngx55diavuWrXwzR3exNL+E8uoZxDTrGwLBZEWfiP5IjcsFdYqGxEsJwchDquNgtJohiAIKdEUYqOmBr+LXwdveA8uHLsKI9TMxLOxxyCU100GSYikBgLlGLlZNJOkSBFEThKY20r1UrKfLjZYwB0Xj2J4sCEwATcuF1r0GihKOpSUcR1ktZuhLSoSygjyhODebKrmay5QX5KO8sACn449BV1aKt0b1o+zlMiQc+ht+rs5Y9fLTUDm5QOngSDmpOHyx8gNm0tgxmP+/DyGKIi455oqent68VCqDJMWepvYytPOUMIFiacqdaUdtnXsUsXtmgXbkMGbxdPGjZUvhLnGG85Qwnv6NY2R9vQU7Rlll1H39f003/l0033+WnJIwAoA6LRhOki5BEDVBaCQDvmq7XGQAgCBbx1FjRHju+jqJLszWVvx/CnB0V8DVR007ednTDu6ecA+QCJyCEiQymhZ37qAzKQ4/J1+ExWJBh06dBDuFnDqQeYXy0ZoQbDLCRy7FN+s24Ku16+GiUsBOxmHb+wspuVTCfvXnEeSWlsNkscLFzYEeHdUGGg9X+AAIlDNw8vLFnFnTqDlPjsOCb76Ha4Q3Gx4aCqNjCQOFFZxceX3UzbASimFZimYYimJoimJpgKIgWoXro27BeG3K3GgVRZ2VEgwWWtJMJQNJugRBNEAC28RGuuevlEEhZVsCYABUa/FRPWYnkbHejIQCRQGiCEAEiq/oUXxFf/Nx9LV/sPzXL3D+8gl4ufpg7Se/Yetfq2lOKcErL73Mcwpa1Bu19MYzT9PBoaEwm82YOHas8OlXX9K9n5wipl44T6lPJqHUZIYginj+sU5wUiqw/ugpXCktR4inG9QsDWcG+Hb1GogGHRYNegxf/RmPze8vhpRlgIrP/a4YiQQqR6fro26FgwOlUNtDpraDXK0CZ6eCl8KZAkm6BEE0QMXOKq6hJ54HUqy3oEBrsng5yMMBnLR1PI+os1VvFAa/EE4zEobKyyzjs84V0TkpJVReRjl46+3bd3u1HoXLBakwGkz4569MbNm6GZN6vYFNS04xAPDbyfXQ6qWY2edzyFwETFs4gJ409ml4h/bGyx99jsXvfyzIFBy17NNPqXlrN+Ls6X/Eg7ql+L+5LwrjZjzHXL1ShKcH9sGXm7fBUVmxjUmECF4QcJ98CwDgLRaU5l1Fad7Vux4zedm3Wk6hyH7Iz+yhkKRLEERNSPdxUjSmlbzV8uf5POnYaJ++DE016KQrGAz9S1d/j4KvvqKkAf6wHzKUad25K9+2RyglUXJ0yRWdkJVUjOwLxfSVtFKY9FYEeUdgUNQUbDy4DO9umoyOmv7wdPLFzsTV8HHVABSFvNLLeOP7CeBFHhQYaCT98M2COOpKRhlenvkWfbXkEkwWA/xaaPDkkOeo0iIrJPIwpqjcgJdmv8if/O80/X8/b6U+/fxzrD5zBrGjnhAnzJgmFF/JRenVXKa8sOL5cllBPvQlJRDFB+uwKFerWQBk9TJBEA1Ouquaa1qFewHsPpPLDQ73mmYnl7wPNNiKXJTI8yO1f//NAIA57SLyP/kE+OQTBgBoJyc4DBlCB/boIQQ/EcxL7VWMvtQkXjpfLNgdtjL7/9uM10Z8d/1ig6ImAwCCvNri/OXjMFkNMJr1mNL7LXCsDHkllyGXKsHSEpiURmiatcOQDtNA0wz+LT+DXgO7ITpwIPITVcyF//JxQmFEjNPTGDpWBbcWaqo4T8HYuzdHs2AJzykgSjiGlnASmmElMOq00BYVCqX5V4XinBymrOAqVV5QgPLCfJQX5MOo01Z545xCIQNw96FwLSBJlyCImqDlBdHoquJU+dpGU6TpvuIvFsJkFdwBtANwwtbxPKQefHGx2njmzB1fFIqKULR6NYpWr66YyeA42PXrS3n3fpyhYpx5hzUKZurSrshJLeUvJRUxOSklKMrVIelyIrydW2LWoI9RUJaDL3fNQ4Bnawgij7QrZ/HqiG/gqHLHqn3v4OiF39A5eABGdnn++n2/+fUNPNH9Rew5uR7ZhWkI9o5El5CBN4dWZY6ZldFw87GDi7eKdvTwoT0Cg0T/SNrKyWmK5RhawnEUBUBfViqWFxYI5YX5EHjBSjMw1PQHei8k6RIEUSOMFj6nuZM8qCklXVEE1h7NlD3d3f8FFcdOsnU8D4MvK5tX9MOa6ldlMplQFrcdZXHbkW0xM6bLl3H1hWfgOHAg49ExmmeH+NEUy1D9Hv+fOGrMVLhLHSg2k4Gz2gNXSy7BQekKb+eWcLHzAgCE+3ZB+tVzQPCNW/ybcRg+rhqYrUbkFqVj6uNvY1ncHES17AXpXfboWo0Cci6UIOdCSeUfUbglxynspXBrYUc5N1MyngEhEEW64EE+q5rQ5J7BEARRO0Qg7Vq7uyblx2NZDENRowH42DqWhxACmo4p2br1kTZ86ePjceXNN5HZvw+T1i6cyhoxDD6UhbqUfRADZoQKg1/ToIy/in5jOwndenaG0aJFuaEiOSZnn4KH440+8jxvxf4zW9E7fDTMVhNwrbyoCBFWwfooYUJfakbGvwU48Wsmzh3KhdUinHukCz4EMtIlCKJGKKXsfz5Oin6oGGE0GflaE1YfSWef7NjiQ7VM8oSt43kQvFb7btGqVRLR8OAzrHNzspGg16OE59EzLRUznV1gufZY+4mUFEyyWvB6YiLV5vvvKbAM3hw8GB1aWCjmsWB+ZdBXzNyXXobJaIWfZzD6dh4OY1nFIqi//4tDdFAfSCUyNHPyB0QR7/08DWHNo6Hgaq5Msp2rDKyEOV9jF6wmShQb6rN/giDqmRlbT17+5KVNp5vccFfNsYh/rZdBJWO7ADhl63iq6TFrUdGOtMf7KASdrm7vLJPBrm9fqPv0ESWhbQSJsyNjMfHITS3hs84VMblppSjK1dXq0rTek0P1mg4eLwH4tvbucjsy0iUIoqak+7soLbYOwhbKTVa8vf2sbPGQVltVHBuCOi648BCUvE639eq779V9wgUAoxFlcXEoi4ujcG1BlLJrFzgNGMh4dIni2WH+NMUwVF5GKZ/5XxGdm1ZK5WWWQbDWXBZ28VZZAJDpZYIgGqyMZo6K+1ctaKS2nsym+rXycOsS4PKxkmOfv/8ZtqM1a1cIglHl/uYbsBYXQx8fb+uQoDt0GLpDh4FrSZgLDob94FgmvFMXvl3vUEqi4OiiXK1w6dy1/cIXy2A2POQzXgqwd5PLAfxXY2+gurcm08sEQdQQzsIL5a0X/iYxWh6sSEFjYS+X4I+XY/ROCukUmqZ+snU8d2LiTS8VGAreGR43XDE2eCxeCHkaZdu381ff/4ARjfV3gM64usBh6FAouseITIBGkNopGV2xUbh0vli8nFzM5KaWQldSvZXzdi5yjHkzqkgqY51rOezbkKRLEESNKdGbz81YeyLkWHqRrUOxmRcea4kXY5oJNKfqA+APW8dzM0EUnig1la4as3OMPFeXCwBopmyGNY+t4B2NLJ09axZlPHv3wV/vtFQoaQY0BbCg8LOvb5XXS3keb17JxSWzBRxN4V0PTwRy3PXXeVHEqMwMuLMslns3BwC8kpODFJMJMSoVXnR1BQAsLyhAEMehl1p99zcjk8F+wACoevcSJaHhgsTJgTEbrWJuaqlwKamIyUktqagbfYcUF9DOFT3GBR+QqSQ9qvfJ1RwyvUwQRI3hWGZvlK+T5lh6UZPcjvh0dz+81KM5qMPLaHSZEwepsj+Ag7aO65pYvUW/aupvU68nXADI1mWj946BzPz288Sxa9eiaNUqoWD5NzSsd566/aF5cziyd04dKwoLEczJ8EUzb1w0mfBO3lWsbn5jJ9Xa4mIESDlohYoy3cnXRtbb/PwwISsT5TwPoyjijNGAZ11c7v1ujEaUbt2K0ortThXPhbt1o5wHDGA8u0ZZ2WH+DMXQ1NX0Mj7rvyI6J7WEys8qh8CL8AxwsEgV7N5qf3I1iCRdgiBqjFzK/NUtyGXSl3+l2tk6lrq2KDZUfKq9C0WtGw5kxQOXEpUYs+43sNxsMJKVtozNKlhnGKyGZTP2zpCnlKTc8ZgPjn9I7bi4EysnfCn69ekjZM+aRZvTMx7oPmlmE6Y7VczY+nMcciwWFFitcGFZXLFYcECnxQwnZ6wprpgJYSkKJlGAIIqwiCJoisIX+XmY6eL6UO9Td/AgdAcPAtdyGxcSDPshQ5m2HTvx7R4Pq3gunKMVlI4yiqapww91k0fUJL+NEgRRa4608XaQNbXeut9NaCs8FWFHUd8/XpFwAeDiX8CKGDnKr3wKk3YVAFt0u1foLfrv8/R5nzyx8wn5mYI7l3qsdK7oHLpse4w5wKRSvlu2wHH8+CqTsxRFYdrlSxiZkY5NJSW3na/hZNinLQcA/GswIMdiwdVrI+b38/Iw19UNN/9sBHAcPCUSjMjMQD+1GllmM0QAobI7V516UKak88h7/31kDh3CpEe3oy/26gnsWEvLlSwNG5XtJEmXIIialGflhYIQz6Yx0KVpYOdz0Xwvb4Givu0O5CVVPaAwFVjeWYFLR8fAVH4SQFQdhhejt+hTjuQceWLE9hGKrPKsap849+A86rnDL8J+zvOiz/p1AuvmBgBY7+ODLb5++Na7OX4sKcZxfZVeu5ju5IRSXsCwjHSsLylGiEwGBsB+rRZOLIOwOyTT19zc8YuvHyY7OeOLgny84OKCbwoL8GJONn6+Q2J/FNareTCcOAmhrOw/AOU1evFqIkmXIIgaRVPUr10DXRr9Ck0ZS+PQS535MOkVivq2G4XSy3c+0FQGrB+pwJ5XQ2EoOQBT+c8AvGsxtOY6i25dialk9/yD871e3P+iQmvR3v+sWxzJOYKu23rT59wtov+vu2E3cKDoxkoAAM4si14qFf41Vq1kpWIYLPH0xC++fnjfwxNFViu8JRKcNOjxl1aL3mmpeDknB8f0eszLyaly7h/l5QiTyaEXRKSaTFjm1Qzby0phEGp2JbyiYwcrpVDsrNGLPgCSdAmCqFFKjt3eL8zDJqOIuuKikiJhXmfBs+w0qO8fp2EovvcJogicWkdhWZgcid8NgUV/AWbdNwAiUXNlM0NhKt9sFawZm5I3jRqwZYBi/6X9j3RBs2DGlD+mM2+efBfK+XNh99GHPG1vD70g4IhOX2VlMgCU8TzM13bEbC4tRXuFAiqGwUuubvgroCX2BbTEx15e6KBQ4EMvr+vnWUQR60qKMcXJCUZBAIXKessVr9UkVUyMnpZKbbaqnCykIgiipv0V5mUvk0sYGCy8rWOpcf6uSvz6bKQgTdktUnHPMhAe4D2atcC+hRIkrJQgctJURE6aAFZWAKniK9DsDgDJqH7xQwZAB1iMQ8CbRkEUPXBsucTc+QVq/6X90nJLzX3v2XlxJ34/8TulW6llpJQExvw89Fcp0U2pwsaSii8cTzg44qLZjFdzc8BQFAKkUrzj4Vmt6/9YXIwhdvaQ0zQ0HAcRIoakp6O7Sgk7pubqrTAuLpD6+koA2GQRFUD26RIEUQtK9OaTc376J2J/cr6tQ6lRUb6O2Dg5XKQTvhWoPxY9ejagKMCnM9DuSQMCegvg1IBFnwRWlgipIh2A4do/RgBSCLwfTGWhEMVASOQBKL8i4Nw2OZJ/ZXE5ERAFiEO/EU96BAqTfptcK9XBJoVNwpzQZ1G2axd/dcn/mIdplmArjuPGia5zZv/C2NmNsFUMJOkSBFHjeEF87afErAWv/3KWu//RDcPA1h74YlQIqH0LBCphRe08mlN7AO6tALcQQO1pgVTJQ6LgIZELEKwUClIUKM6gUZIFFKYA5Vduv4ZUBcMrKZi0ZzLOFdVOaWFPhSfWPLaCd7HI6MuzZlPGf/+tlfvUNN/NP5fJW7WaAGCHrWIgSZcgiNrgqzNZk9q9s1dmsjb8kpBTu/jhzb6+wLZnRerctnq/IUoYs048onYQnv3juVqthf1y+5fxpP8YFK/5PyH/yy/vWlCjPmDdXBGwd6+e5jgnANWrF1kLyEIqgiBqQ4ZVEE/3DfOwdRyP7M0BIeKbfXxAbRiNhpBwAYDeMZtq79Ge8bf3r9X7fHz8Y4zeOwGSJ4bCL26bKPWv3fs9CnWfPqJoNu+GDRMuQJIuQRC1xF4uWTaxs2+DXsW8fFxbYWp7B4pa1RfIqC/VHKtBXwg244jwfNvna30l24XiC+ga15v+QzwH3y2b4ThxogCq/n03cRg+opxRq1fbOg6SdAmCqC3bwrzsqOZOclvH8VC2zYji+7UARa2IAa6etXU4D4zd/gLd3bs7462qzS3BN7x6+HXq6b9nwm7mDLTYsF5gPerPLAfr7g6pvx8LYJ+tYyFJlyCI2mLiBXHdmPY+9fdB3x1IWRqHXurEhysKKopelFS/klO9UpYNJvuUMKPNjDrbt5VwJQHdtvWm/3XRw3/XTtgNjq0Xi4bsBgwQRItlOwCzrWMhC6kIgqhNrYv15qOR7+xVCA3gV42DgsX+2R0E+6IzoDaMpmHR3/+k+sy5JYzPHsbAXwYiT59Xp7fu59sP70W+LZoSj4u5r71O8zVc0rHaKAotD+zXSdzcHgcQb5sgbiAjXYIgatMZmqIyuwc9XNeYutTCWYGjL3cW7LP+EKm1Qxt+wgWAwlTQeUn81FZT67xKyZ6MPYiJ60Nlh7iI/r/tgbJbt7oOAQCg6t4dtFyeDeCoTQK4BUm6BEHUKnu55NOnOvnqbB3HvbRr7oA/XogSudM/iNSWKQyEBjUjfk/SHbOZ4YHDGAfOoc7vrbVqMeq3cczS5OXw+mwZPN97j6cUijqNwXnaVC2jVi9B9St91SqSdAmCqG0bOwc4My4qW3S2u7++oe7YPD0CzIElIvX7m7W6r9Umck8DxRn8xLCJNtswvT5pPfrtGgJdTAQC9vwqytu2rZP7SgMCIGvVWgDwU53csBqYhQsX2joGgiAaN5PBwocJghhy9GJRvfqiP7FzCywdGgRq+0yROrmmXsVWk9i8JDr4sXeojec3wiyYUfBbAbJXZaPozyLoL+ihDleDYm5s8ynYc+31v4pQllgGZbASjKLi+8jZyWdRdqIMRX8Vofzfcjh0rBhBX/rmEvK25cFSZIEqVAUAyIvLA6/nwXly0Fl0WJfyIy23c6JiprwJWqUS9MePU6jhLkI3c3/tVQMX2HIpxbI2a3Bwq0b7Q0YQRP2hlknemt7d3+qkrD+j3Vf7abCwry+ojWNBnd1S/zaW1qTMw6DKc/lxIeMES7EFhXsLEbAwAIHvBUIURJQeK61yuKyFDAELAhD4biDsouxwZdONcpO0lEbLd1qi5Tst0WJOCwCA8ZIRABD4biD0F/Tg9TwsJRYYLhpg165qb+VPT36KUb+PBzNqkOgft02UtmxZK2+ZdXeHuk8f0Bz3Ra3c4CGRpEsQRF1IE0Wsnd0r0KbVgCp9MaaNOKODC6gfBgAX99s6nDoh3/M6MzlsMi1jZBAFEYJZgMiLEM0iWMeqDedUISrQXEV6UAQoYC26zzNuBhAtIkRBhGgVARrI25oHt+Fudzw8pSQF3eMeZ37jz8D3501wmjK5xgtqOE2ZbIYgrAZQVKMXfkQk6RIEUSeUHPvW6KjmvLejbYtlbH66vTAogAW1okfF886m4sIe0IZifmyHsaJLPxdcePkCzs85D1pOQ91KfdfTiv8uhqqN6vr/FywCUhemIm1xGspOlAEAZF4ySJwkSFuQBvsoe5ivVmyHlbe499/1G0fepKYfeB7qZ6ehxU8bBfamHruPgvXwgOPo0TytUPyvRi5Yg0jSJQiirlylgE/n9wu2SS84KUtj/5yOfKS6BNS3XSkUp9siDJtS7FvMjGk+htKe0iLooyAELwuGYBJQcuTOe2hLjpTAkG6AS3+X63+m+ViDlgtbovkzzZG7IRemvIrJC8/xnmj5Tku49HepGOUOc0Pe9jxkfZWFov13H2wmXk1El2296FMOpfDfuR32Q4c+8ipjt3mvGAB8AeDyo16rppGkSxBEnZFJmA96h7hbgj3uPrKqDXYyFkfnduJbmC5Q1MrHaOgK6vT+9caZTTi493ehpX9LsHYsKJaCXXs76FNv35Os/U+L/B35aDGnBWjJjVQhcZQAAKRuUiiDlTBmGqucV3ayDHI/OQSTAFO2CT7P+6DkSAkE090XTFkFK5756zn6lYS34PLma/Be8S3PODo+1FuUtQqDumdPMy2Xv/tQF6hlJOkSBFGXyliGWvjmoNA627fb3FGOY690EhxzDoD6v1gaZm1d3bpeCirYTxvSDKDMFERRhO6cDpxn1bbHhkwDsn/Ihs9sH7B2N5738joegqUieVrLrdCn6sF53ThXtIoo3FsIl/4uEMwCUPmYVqx47X72Zu5Ft7jHqaxAewT8tgeqHj0e+P15vL1AS0ml8wDUy2YbpAwkQRB1jdOZrFmTVie4JWYU1+qNwr3tsWVaW5E5vV6g9sxnQH7fAQBeM00XV2xYTel4HWQ+MjSb0gz5O/Ih95PDLsIO6R+mw3jZCIl9xahW4ixBizktoE/RI3tNNiiqImE7P+4Mpxin69ct+K0AjJKBY1dHiKKIy99chjHbCHUbNTxGP1gDhLHBYzGv9WxR+/s+4eo7ixlBd/8KYereveH5/v8uMipVEIA6r8JVHSTpEgRhC0+eyyn9esDnh1T3P/ThPBbshu/GhoL6+0OBOvwpmdW7WdeXkdNxqthvS39KrB+Fmu7IVe6KNY+t5D1EOzp79hzKcPLk3Q+WSNBy316dxN19OIDf6yzIB0R+EAmCsIUNPk7KqwNbe9bKxcdF++D7cWGgdr0kkoR7B4c+hj2rEHv69LR1JPeUb8jHgF1DmdW5v1DNv/8ObvPn85REcsdjncaNE2iF4gTqccIFyEiXIAjb6VhutPzZ6+MD8rzymtu++3KfIMzs6gVq05NAar0pRFT/9FqIi60HC0PihjaILyX+9v5YHfMNrywz09kvvECZLqRcf4319ETArp0GWqGIBJBkuyjvr0F82ARBNEpHWZpe9sXYCH1N1UX4ZFRrcWZnN1A/DKrVhHupVEDPNTqEfKVF2NdafHa04ktDkUHE42t1CPxCi8fX6lBsuPugpswkotkn5Zi5+/YdVIN/1KPV1zcWfM3fa0Sb5Vo89cuNY9eeNl+/70P5czHc5a5UJ69OD3+NOnSx9CJitvdhdhmPo8VPP8Fp+jQBdEUKa7b0Ix0YZgnqecIFSNIlCMKG5FJmYZiXffrETr6PXIB349RIflgQV1H0Iucez/5qAEsDH/eRIel5FY5OVeKrRAvO5fN4/5AJvfxYpLygQi8/Fu8funtSfOtPE2Ja3N5fYWuSBTf3hig1ijhymce/z6rAiyLOXOVhsIj44bQFz0U9QllNUYDy303UnIjZNmuE8DAWHF1ITf5rBlTTJ8F300+Cy8yZIqfRZNIc976tY6sOknQJgrAli0rGDpvfL9jY0u3h1lSxNPDH7I58ByctTX3blULRxRoO8XaeahrtPCsSppqjEOJKI7tMRFyyFRPDK545TgyXYFvyncsnnsjhcVUnoE9A1fKLWrOIT+LNeLP7jW04NAWYeRGiKMJgASQM8NERM2ZFSyFhHnGKYM9r8LVrQUW4RTzaderYqfxT6LKtF/2fox5OU6eIjEo1CkCD6MdIki5BELaWImGoF799MlL3oElELWNx9JUugr/1IkWt6ElBm1dLId5dRomAU7k8OngzuKoV4Kmu+LXqqaaRp7t9ECmIIl7+3YiPHpfd9tpbf5rwcicpFJIbn4OaozAiRIKIb3Xwc6Bhz1FIzOExJPjOC4oeCG+GPGknZkfMqpfba+6FoijYKxwNvJR5GcA5W8dTXSTpEgRhcyxDr3RTc/Gv9NWYq3tOMwcZjs3tKDhfPSxSPwygYSqrzRDvSGsWMWKTHp/2k8GOq94Xhq8TLRgQyKK5fdVfv/9c4ZFaLGBYyO3JdF4XDv88o8LHfWV46y8TFvfg8N1JM0b/rMe7fz/aIjRq18tUqHMoE+oU+kjXqWuzI2abPZWeCVJG+pmtY3kQJOkSBFEfiGqZZNyEji30Hfyc7ntwqKca+2d3EOVJm0Xqp/EM+Grn6hpj4SsS7vjWEgy/lijdVTRyyytGt7nlAtyUt/+Kjb9sxZcJZvh+Wo65v5vwf6cteHWfEfGXeJzI4eH7aTm6rtLhQqGAHj9ULdx1KrdiQBrkTOP/TluwaZQCZ/N4pBQ+wkDVrAWX+of4QsTMBjPa7ejZEWOCx+hUUtUYoB5vNL4D9v6HEARB1Il8hZQd//X4dj/3XLpfUWa88yO6HhpXrBoXBurwMpH6+6PbVyLVAVEUMXW7ESEuDF7qdOP56+AgFmtOW/BqVw5rTlswRHP7r9j1wxXX//2Hf8w4nsPj/d4VU83PXlsYlVEiYNAGPfZPUlY5962/TFgRK4NFAPhrqYamAL3l0d4PvWM21f6l/xh/e39cLK39Z+KPwknmhI9iPtLLWfloAPm2judBkZEuQRD1yW5OwqxfPiFSz9C3T9eOae+N1eNbgfr1FZH6+yOb/f46fInH2n8t+DPdirbfaNH2Gy12pw4kX1UAAAjESURBVFjwalcp9l60IvALLfZetOLVrhUJ+XgOj2nbH6250rbzFkR5MfBS03CQUejkzaD1ci0oCgj3eMTvHvpCsBnxwnNtn6vXo105K8fKPit1Ulr6GYB9to7nYZDiGARB1DeScqPlz53/5ka9tvXM9WHk7F4tMSfGG9TPk4CUel10qGGy84Zx1gkMixuGy9p61xEPLMViRZ8V+lDn0G1KiXICGti0ciUy0iUIor6xqGWSQbHhXpef7u5vBYAPh7cS53TzBPV/g0nCrS1ll8Hk/CPMaPN0vRvtUqCwpNsSQ4hTyFGlRDkRDTThAmSkSxBE/dVcb7b+k5lf5hBsz9PUD/2BgpT7n0U8POdAGJ89hIG/DESevu63X93N3PZzzSODRiYpJcrOAO7fbqgeIyNdgiDqq0sKKds72F0lUnvmk4RbFwpTQOed56e2mlpvRrtPhT7FjwwamauUKHuhgSdcgCRdgiDqt1MUK+2D2M/18Gxr61iaBOmOWczwwGGMA+dg61AwwG+AODNiZolSouwOoNDW8dQEknQJgqjv/gSnGo+J2w1wDbZ1LI1f7mmgOIOfGDbRpjWZO3l2wsLOC7VyVt4DQJYtY6lJJOkSBNEQbINUOR1TfzeQEW/tk+2ay4wLHkerJA9XD/tRhTqF4tOen+rlrHwggLM2CaKWkKRLEETDQLPrIbMfi0m79PBpGO3oGqyMQ6DKr/Bjg8fW+Wi3jUsbfNf3O4NCohgP4GBd37+2kaRLEERDEgdONQQTtujQspetY2nU5L+9zkxpNZmWMbc3ZqgtMd4xWNlnpU4tVY8EsK3OblyHSNIlCKKh2Qepsg9GrytH6BCy57G2JP8K2lDCjwwaWSef8cjAkfxHMR+VKiSKxwDsrot72gJJugRBNERHIFV0x9BvyhA9vUE1YW9IFPsWMzPazKAkdA20EbyH59s+b3kl6pU8OStvDyChVm9mYyTpEgTRUP0DqSISvRZmYtg3RrDc/c8gHsyZTZBYjfzggMG1MtqVMTIs67HM8GTokxcUEkUEgNTauE99QpIuQRANWRo4VRsED/oD0//UQe1p63gaHeWBj5jn2z4PhqrZhk7uCndsHLRR18mr0y6lRBkF4GqN3qCeIkmXIIiGTgtOHQungA/wXLwBzTvYOp7GJfE7KECJfX371tgl27q2xdbBWw3eKu93lBLlaACP1oKpASFJlyCIxkCERP4O5I4j8OQ2LaKmkee8NUgZv5yeFfGCSOH2dosPgqZoTAiZIKzos0Jrx9mN4FjuAzTg5gUPgyRdgiAak18hVbRD74UpmLhDT6aba8jBpXCQKMWezXs+9CX87P2wceBG3fNtn/9HzsrbAfi15gJsOEjSJQiisUkBp24N7/bLMDPRgDZjmtRIqrYoj6+hZ7eb9cAzCCzFYkabGdafBv2kC3QMnK+SqqIANNnuFaS1H0EQjVk7mMq3IPOIG+KeU0BXYOt4Gi6Khu7VTPHFv+dS8Tnx1TolxCkEH3b/UOcidzmhkqqeRCOqofywSNIlCKKx42DW/Q+C9RnsekmOM5ttHU/DNfBjnPPrJIzZ9cQ9Z0mltBQzI2aanwh+wsQx3EyaoteiiT27vRuSdAmCaCo6wlj2A4ozmmHnHBWyT9g6noaHkUI/P1185o/nqFN5p+54SIRbBD7o/oFeLVH/pZKqpqKJbAWqLpJ0CYJoSmiIwlOwGD9Byu8y/Pa6HGXZto6pQRGHfSuecA8QJv82pcrGXWeZM2ZGzDQN9B9okLPyqQC22ijEeo0kXYIgmiIVLIbXAXEO4r9mcegTCcw6W8fUMEhVMLySgol7JiGpKAmOnCOmtZ5mHq0ZLQiisEYhUbwGoNjWYdZXJOkSBNGUNYex7FNQVH/EfynBsW9ZGEi+uB9+zHrxuIOreK7wnPWJ4CcEQRTWKyXKhQAu2zq2+o4kXYIgCEADU9lboNkROLWOxuHPpCgl+eOO7JoBPV4z61uPYEWIa5US5dsgq5KrjSRdgiCIG5rBrH8FFDUdybsp/L1Ujrxzto6pfvBuD3R41oDggYAofA+p8kMAl2wdVkNDki5BEMTtHGA1zYTAv4ziDAaJ36nw31aqyU09K12B8LECoqfrIXcsB8t9CUa6HOSZ7UMjSZcgCOLuWACPw1DyHFhZb1z8i8eJ1Uqk7gME3tax1Q6aAVo+DkRN18KvKwurOQ4yu68AHALZa/vISNIlCIKoHkcAo2EomQmKCsB/v1C4sEeG9L/R4Fc+yxyAlr2AkMF6tOxNQ7CmQe7wKYBNAMpsHV5jQpIuQRDEgwuCwA+BsXQ0pMo2uPKvEee2q5G6l0Jekq1jqx5XDRDUX0DYUC3cQjlY9Echd/wRwG6QZ7W1hiRdgiCIR6MC0BMm7VAAseDNSmTF88g8rEb2CSD3X8Cit22EShfAqx3QLFKAb1ctPMOlAKUFEAdOtRXAX2hCPW1tiSRdgiCImkMBaAmgI0zabuAt3cGp/KHLNyL3NJB9QoXiDApl2UDpZaAsBxCsNXNnRgLYNwccfACHFoBjCwHurXRo1o4GZ0fDrDsLqfJPsNxRAIkASCkuGyBJlyAIonZJAAQBCIfVFAGzLgSALxipJyRye5jKTSi/YkHpJQqmchoWHQOzgYFFx8CiZ2A1AVYTIJEDUpUAmZ0VnJ0VnB0PTi1A7gDYNZNAZs/B8v/t3bFKggEUhuEvNew3aZfo/m+m2cW22kMRBzFrcGsIgt/P5XngXMPLWc45fObr+J676Sbzp3Umk02S1yRvSf79lo/xiS7A7cySrJK8JHlO8phkSPKQZMj3ecjpuMz5tMj0fpfZfJtk/2t2uRyn+Egy0trMtYguAJT8+RMRABiP6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQMkPLbMm/rRer4UAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pad sequences (samples x time)\n",
      "x_train shape: (2874, 400)\n",
      "x_test shape: (1416, 400)\n",
      "Build model...\n",
      "Train on 2586 samples, validate on 288 samples\n",
      "Epoch 1/6\n",
      "2586/2586 [==============================] - 18s 7ms/sample - loss: 0.4355 - accuracy: 0.8546 - val_loss: 0.3997 - val_accuracy: 0.8576\n",
      "Epoch 2/6\n",
      "2586/2586 [==============================] - 18s 7ms/sample - loss: 0.2997 - accuracy: 0.8813 - val_loss: 0.2049 - val_accuracy: 0.9340\n",
      "Epoch 3/6\n",
      "2586/2586 [==============================] - 17s 7ms/sample - loss: 0.1293 - accuracy: 0.9509 - val_loss: 0.1607 - val_accuracy: 0.9410\n",
      "Epoch 4/6\n",
      "2586/2586 [==============================] - 17s 7ms/sample - loss: 0.0781 - accuracy: 0.9718 - val_loss: 0.2130 - val_accuracy: 0.9444\n",
      "Epoch 5/6\n",
      "2586/2586 [==============================] - 17s 7ms/sample - loss: 0.0625 - accuracy: 0.9811 - val_loss: 0.2324 - val_accuracy: 0.9306\n",
      "Epoch 6/6\n",
      "2586/2586 [==============================] - 17s 7ms/sample - loss: 0.0363 - accuracy: 0.9857 - val_loss: 0.2167 - val_accuracy: 0.9340\n",
      "1416/1416 - 2s - loss: 0.2101 - accuracy: 0.9308\n",
      "test loss: 0.21011744961563478, test acc: 0.9307909607887268\n",
      "[[1163   50]\n",
      " [  48  155]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pad sequences (samples x time)\n",
      "x_train shape: (2874, 400)\n",
      "x_test shape: (1416, 400)\n",
      "Build model...\n",
      "Train on 2586 samples, validate on 288 samples\n",
      "Epoch 1/6\n",
      "2586/2586 [==============================] - 18s 7ms/sample - loss: 1.1947 - accuracy: 0.6121 - val_loss: 0.7535 - val_accuracy: 0.6562\n",
      "Epoch 2/6\n",
      "2586/2586 [==============================] - 18s 7ms/sample - loss: 0.6555 - accuracy: 0.8005 - val_loss: 0.5115 - val_accuracy: 0.8507\n",
      "Epoch 3/6\n",
      "2586/2586 [==============================] - 18s 7ms/sample - loss: 0.4232 - accuracy: 0.8720 - val_loss: 0.3215 - val_accuracy: 0.9236\n",
      "Epoch 4/6\n",
      "2586/2586 [==============================] - 18s 7ms/sample - loss: 0.2397 - accuracy: 0.9393 - val_loss: 0.2479 - val_accuracy: 0.9375\n",
      "Epoch 5/6\n",
      "2586/2586 [==============================] - 18s 7ms/sample - loss: 0.1604 - accuracy: 0.9551 - val_loss: 0.2623 - val_accuracy: 0.9340\n",
      "Epoch 6/6\n",
      "2586/2586 [==============================] - 18s 7ms/sample - loss: 0.1293 - accuracy: 0.9640 - val_loss: 0.2831 - val_accuracy: 0.9340\n",
      "1416/1416 - 2s - loss: 0.2571 - accuracy: 0.9350\n",
      "test loss: 0.2570587779674153, test acc: 0.9350282549858093\n",
      "[[857   2   8   5   0   0   1]\n",
      " [  2 108   1   6   0   0   0]\n",
      " [ 24   4 262   0   0   0   0]\n",
      " [  9   2   2  70   0   0   1]\n",
      " [ 10   0   0   3   0   0   0]\n",
      " [  0   0   0   0   0  13   0]\n",
      " [  8   4   0   0   0   0  14]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import nltk\n",
    "import re\n",
    "import random\n",
    "import tensorflow as tf\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.feature_extraction.text import CountVectorizer,TfidfVectorizer\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.feature_selection import SelectFromModel\n",
    "from data_print import data_class_display,PCA_text_data_display,pp_conf_matrix\n",
    "from EstimatorSelectionHelper import GridSearchCV_Classifiers, Best_Classifiers\n",
    "from ukrainian_stemmer import UkrainianStemmer\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "from tensorflow.keras.preprocessing import sequence\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Dropout, Activation\n",
    "from tensorflow.keras.layers import Embedding\n",
    "from tensorflow.keras.layers import Conv1D, GlobalMaxPooling1D,MaxPooling1D\n",
    "from tensorflow.keras.datasets import imdb\n",
    "    \n",
    "def ua_tokenizer_sklearn(text,stemmer=True):\n",
    "    \"\"\" Tokenizer for Ukrainian language, returns only alphabetic tokens. \n",
    "    \n",
    "    Keyword arguments:\n",
    "    text -- text for tokenize \n",
    "    stemmer -- if True use UkrainianStemmer for stemming words (default True)\n",
    "          \n",
    "    \"\"\"\n",
    "    text=re.sub(r\"\"\"['’\"`�]\"\"\", '', text)\n",
    "    text=re.sub(r\"\"\"([0-9])([\\u0400-\\u04FF]|[A-z])\"\"\", r\"\\1 \\2\", text)\n",
    "    text=re.sub(r\"\"\"([\\u0400-\\u04FF]|[A-z])([0-9])\"\"\", r\"\\1 \\2\", text)\n",
    "    text=re.sub(r\"\"\"[\\-.,:+*/_]\"\"\", ' ', text)\n",
    "    if stemmer==True:\n",
    "        return [UkrainianStemmer(word).stem_word() for word in nltk.word_tokenize(text) if word.isalpha()]\n",
    "    else:\n",
    "        return [word for word in nltk.word_tokenize(text) if word.isalpha()]\n",
    "    \n",
    "    \n",
    "    \n",
    "land_data=pd.read_csv('land_real_estate.csv').replace(np.nan, \"\")\n",
    "land_data['text']=land_data['description']+' '+land_data['land_types_source']+' '+land_data['land_types_PcmU']\n",
    "def land_types_download():    \n",
    "    data=pd.read_csv('land_types.csv')\n",
    "    data=data[['id','land_types']]\n",
    "    data=data.set_index('id')\n",
    "    data=data.T\n",
    "    data=data.to_dict('records')\n",
    "    return (data[0])\n",
    "\n",
    "land_types=land_types_download()\n",
    "built_up={0:'Не забудована',1:'Забудована'}\n",
    "\n",
    "def Data_Augmentation(number):\n",
    "    unigrams_list=['продаж','земельна', 'ділянка', 'іншої', 'інфраструктура', 'інструменту', 'інвентарю', 'ізумрудє', 'івасюка',\n",
    "                   'яром', 'яворівського', 'шанове', 'чудова','цін', 'цільових', 'цілу', 'цілорічно', 'участке', 'тільки', 'тухолька', \n",
    "                   'турківський', 'твердій','сухий', 'суха', 'сусідніх', 'сусідні', 'судова',  'сторони', 'сто', 'стихії','селі', 'села',  \n",
    "                   'сайті', 'руська', 'росташування','рокитне', 'розташовану', 'розташований', 'розміщені', 'розміщення', 'розміщений', 'розмірі', \n",
    "                   'розділена', 'покупцю', 'показ', 'повідомлення', 'питання']\n",
    "   \n",
    "    random.seed(0)\n",
    "    random_words=[unigrams_list[index] for index in random.sample(range(len(unigrams_list)), len(unigrams_list))]\n",
    "    tsv='для ведення товарного сільськогосподарського виробництва'\n",
    "    df = pd.DataFrame(columns=['text','land_types','built_up'])\n",
    "    for sample_index in range(number):\n",
    "        df = df.append({'text': tsv+' '+\" \".join(random.sample(random_words, random.randint(0,20))),'land_types':6,'built_up':0}, ignore_index=True)\n",
    "    return df.astype({'text':'object','land_types': 'int32','built_up': 'int32'})\n",
    "\n",
    "land_data=land_data.append(Data_Augmentation(34), ignore_index=True)\n",
    "data_class_display(dataframe=land_data,class_column='land_types',expl_lables=land_types)\n",
    "data_class_display(dataframe=land_data,class_column='built_up',expl_lables=built_up)\n",
    "\n",
    "X_train_zab, X_test_zab, y_train_zab, y_test_zab=train_test_split(land_data['text'],land_data['built_up'],\n",
    "                                                                  stratify=land_data['built_up'],\n",
    "                                                                  test_size=0.33,random_state=0)\n",
    "\n",
    "X_train, X_test, y_train, y_test=train_test_split(land_data['text'],land_data['land_types'],\n",
    "                                                  stratify=land_data['land_types'],\n",
    "                                                  test_size=0.33,random_state=0)\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "def plot_graphs(history, metric):\n",
    "    plt.plot(history.history[metric])\n",
    "    plt.plot(history.history['val_'+metric], '')\n",
    "    plt.xlabel(\"Epochs\")\n",
    "    plt.ylabel(metric)\n",
    "    plt.legend([metric, 'val_'+metric])\n",
    "    plt.show()\n",
    "\n",
    "\n",
    "\n",
    "def fit_print_conv(X_train, X_test, y_train, y_test,n_classes): \n",
    "\n",
    "    # set parameters:\n",
    "    max_features = 5000\n",
    "    maxlen = 400\n",
    "    batch_size = 32\n",
    "    embedding_dims = 150\n",
    "    filters = 500\n",
    "    kernel_size = 3\n",
    "    hidden_dims = 250\n",
    "    epochs = 6\n",
    "    \n",
    "    \n",
    "    t = Tokenizer(num_words=max_features)\n",
    "    t.fit_on_texts(X_train)    \n",
    "    X_train = t.texts_to_sequences(X_train)\n",
    "    X_test = t.texts_to_sequences(X_test)\n",
    "\n",
    "    \n",
    "    print('Pad sequences (samples x time)')\n",
    "    X_train = sequence.pad_sequences(X_train, maxlen=maxlen)\n",
    "    X_test = sequence.pad_sequences(X_test, maxlen=maxlen)\n",
    "    print('x_train shape:', X_train.shape)\n",
    "    print('x_test shape:', X_test.shape)\n",
    "    \n",
    "    X_train, X_val, y_train, y_val=train_test_split(X_train,y_train,\n",
    "                                                  stratify=y_train,\n",
    "                                                  test_size=0.10,random_state=42)\n",
    "\n",
    "    \n",
    "    print('Build model...')\n",
    "    model = Sequential()\n",
    "\n",
    "    # we start off with an efficient embedding layer which maps\n",
    "    # our vocab indices into embedding_dims dimensions\n",
    "    model.add(Embedding(max_features,\n",
    "                        embedding_dims,\n",
    "                        input_length=maxlen))\n",
    "    model.add(Dropout(0.2))\n",
    "\n",
    "    # we add a Convolution1D, which will learn filters\n",
    "    # word group filters of size filter_length:\n",
    "    model.add(Conv1D(filters,\n",
    "                     kernel_size,\n",
    "                     padding='same',\n",
    "                     activation='relu'))\n",
    "    model.add(Dropout(0.1))\n",
    "    model.add(Conv1D(filters//10,\n",
    "                     kernel_size,\n",
    "                     padding='same',\n",
    "                     activation='relu'))\n",
    "    model.add(Dropout(0.1))\n",
    "   \n",
    "    # we use max pooling:\n",
    "    model.add(GlobalMaxPooling1D())\n",
    "\n",
    "    # We add a vanilla hidden layer:\n",
    "    model.add(Dense(hidden_dims))\n",
    "    model.add(Dropout(0.5))\n",
    "    model.add(Activation('relu'))\n",
    "\n",
    "    # We project onto a single unit output layer, and squash it with a sigmoid:\n",
    "    model.add(Dense(n_classes))\n",
    "    model.add(Activation('sigmoid'))\n",
    "\n",
    "    model.compile(loss='sparse_categorical_crossentropy',\n",
    "                  optimizer='adam',\n",
    "                  metrics=['accuracy'])\n",
    "    history = model.fit(X_train, y_train,\n",
    "              batch_size=batch_size,\n",
    "              epochs=epochs,\n",
    "              validation_data=(X_val, y_val))\n",
    "    \n",
    "    results = model.evaluate(X_test,  y_test, verbose=2)\n",
    "    print ('test loss: {0}, test acc: {1}'.format(results[0],results[1]))\n",
    "    y_pred=model.predict_classes(X_test)   \n",
    "    con_mat = tf.math.confusion_matrix(labels=y_test, predictions=y_pred)\n",
    "    print(con_mat.numpy())\n",
    "    plot_graphs(history, 'accuracy')\n",
    "    \n",
    "\n",
    "\n",
    "    \n",
    "    \n",
    "fit_print_conv (X_train_zab, X_test_zab,np.array(y_train_zab), np.array(y_test_zab),2) \n",
    "fit_print_conv (X_train, X_test, np.array(y_train)-1, np.array(y_test)-1,7)   \n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Приклад рекорентної нейронної мережі"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\User\\Documents\\article\\data_print.py:50: UserWarning: You have mixed positional and keyword arguments, some input may be discarded.\n",
      "  plt.legend(fracs, labels=labels, loc='best', bbox_to_anchor=(-0.1, 1.), fontsize=8)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAd0AAAD3CAYAAABYfY6KAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOzdd1ST1xsH8O+bQQYEhLBBNgRxoEYi4Kqj7lWlWkeLtVZbZ6Xu1l1tpYoVV62rLhwVFcX5s1oHwyhSFJGA7C1LdkLG+/sDQwFxVKug3s85nmPfeZPYPLn3vfd5KJqmQRAEQRDE68do6gYQBEEQxPuCBF2CIAiCeENI0CUIgiCIN4QEXYIgCIJ4Q0jQJQiCIIg3hARdgiAIgnhDSNAlCIIgiDeEBF2CIIjXhKKoVIqi+tT5byuKopIpivJvynYRTYcEXYIgiDeAoigTABcBnKNpel5Tt4doGiToEgRBvGYURbUAcAGAFMC0OtsnUBSlpiiq/PEfmqIop8f7Pqco6j5FUWWPe8dTGlxzGEVRf1MUVUpRVBJFUf0fb/+LoqhJj//OoCjqLkVRmW/sxRLPRIIuQRDE66UH4CwAFoCJdP3cuwwA4TRN69E0rdfgvIcABgPQB/A5gPUURXUEAIqiJAD2ApgLoAWA7gBSG7m3LwDD/+6lEK+KBF2CIIjXayuAcgDWALo02KcDoLqxk2iaPk3TdBJd4wpqesrdHu/+AsAumqb/R9O0hqbpLJqm4+ueT1EUF8BiACv/w9dCvCISdAmCIF6veABDAMwDsJOiKF6dfUYAihs7iaKoARRFRVIUVURR1CMAAwEYP97dEkDSc+47C8B5ALJXaTzx3yJBlyAI4vVaRdO0nKbp7QDSUb/n6QIgoeEJFEVxAAQDWAvAjKbpFgDOAKAeH5IBwPEZ9zQCMB3A8ldvPvFfIkGXIAjizfkSwGSKoiQURXUBMBxASCPH6QDgAMgHoKIoagCAvnX27wTwOUVRvR9PlrKiKMq1zv5vAOykaTr39bwM4mWRoEsQBPGG0DSdDGAJgBsA9gGYQ9P0jUaOKwMwE8AR1Aw/jwVwss5+KR5PrgJQAuAKANs6l2CippdMNDMUKWJPEARBEG8G6ekSBEEQxBtCgi5BEARBvCEk6BIEQRDEG0KCLkEQBEG8ISToEgRBEMQbQoIuQRAEQbwhJOgSBEEQxBtCgi5BEARBvCGsZ+2MiooyZbFYOwC0AQnQBPE6aQDEqlSqSWKx+GFTN4YgiNfjmUGXxWLtMDc3b2ViYlLMYDBI6iqCeE00Gg2Vn5/vlpubuwPA0KZuD0EQr8fzeq9tTExMSknAJYjXi8Fg0CYmJiWoGVUiCOId9bygyyABlyDejMf/r5HHOATxDmsW/4Onpqay3dzcWnE4nI5KpfKJ/WVlZQwfHx87iUQi6ty5s8u1a9f4ADB37lyLhQsXmgNAaGioYObMmZYAcOHCBd0ePXo4qdVq2NnZtZFIJCKJRCIaPHiww4u0R61WY/LkydZeXl4uAwYMcACAI0eO6Nvb27cWi8Wixs4JCwvjeXt7O3fu3Nnl0KFDBgDw1VdfWXfs2NFVLBaL7t69y6l7/MGDBw3c3d1dZ82aZQkAt2/f5i5YsMC8sWsvWbLELCwsjHfp0iXdDh06uIrFYtEXX3zRUrv/+PHj+p6eni4SiUSkfW8A4Pfff29hbm7e7nnHvc18fX1bPm3f8OHD7TUazZtsDkEQxDM985luXXKl2p3LZr7w8Y2cr+KymTGN7TM1NVVduXIlYdCgQU6N7Z83b57Fhx9+WDplypSirKws1pAhQ5zCw8Nla9asyenTp4/TyZMnBQxGze+HrKws1ty5c1uePXv2AZPJhFAoVEqlUtm/aevu3bsNW7VqJf/tt98ytdt69uxZERcXF9e1a1eXxs5ZsWKF5fnz55MEAoEGAPLy8ph///03//bt2/EXLlzQ/eWXX0x37tyZoT3+wIEDQqlUKhsyZIgDAKxbt85006ZNmQ2vq1arERERobdixYq89PR0VlhYmIzP59NDhw61l0qlPDc3N/m2bduMr1+/nsBi1f94goODDS0sLKoBoLy8nHracW+zPXv2ZDxtn0QiKT958qRg+PDhZW+yTQRBEE/zwt++XDaTZbfg9EvfKPWnQU+9F5/Pp/l8vvpp+6VSqd7WrVuzAMDKykrVrVu30suXL+sOGDCg/PDhwyldu3YVzZ49O1etVlM+Pj4Oa9asybS0tFS9bFtPnz5tYGxsrJJIJKLRo0cXfvvttwUmJiZPbV9cXJyOQqGgBg8e7MBms+ndu3enmZiYqA0NDVUqlQpFRUUsoVBYrz1sNlujVCpBURQiIiJ4Tk5OCkNDwye6ZZGRkTxLS8tqALCxsam9BovFoplMJn3p0iU9BoOBHj16OJuYmKj27t2bpq+vrzl06JBB7969S/ft22cMAE87Tnu92NhYzpQpU2yqq6sZPXv2LPX3988ZOXKk3fLly3Pi4uI4J06cMAwKCkobP368TVxcHJ/BYNDnz59/wOFwNGPGjLHLz89nC4VC1bFjx1K2bdtm5O/vb5menn63pKSEYW1t7X748OEHADB58mQ7a2vr6tLSUmbfvn0fBQYGZoeGhgq02x9/3rIFCxaYh4aGGqpUKkp7nEgkcnNxcalKTk7mbtmyJa1Hjx6VYrFYFBUVJRs5cqRdYmIiV6lUMr799tucSZMmFffr169s48aNJiToEgTRXDSL4eV/y9raujozM5MNALm5uSxdXV3N0qVLrYOCgoyVSiUVFxfH1R5bWFjI1g4va4einyc/P58tEonkYWFhssOHDwszMjKe+eMkOzubnZyczA0NDU2ePHly/uLFiy24XC7t4OCgcHBwaPPtt9/aTJs2raDuOXPmzMkbNWqU/YgRI4oDAwNNe/ToUT5u3Dib7du3G9Y97v79+1w7O7vquttu3LjBKyoqYonFYnlOTg47Ly+PfeXKlUQvL6/ygIAAEwDYu3ev8KuvvirSnvO047Tmz59vtWfPntSbN2/K7t+/z01KSmIDwK1bt3jbt2832b17dzoA7N+/P/327dvxrq6uVWlpaewNGzaYDB48uCQyMjKhe/fuZb///rshALRs2VJx+vRpwZ49ewzd3d3Ltffx8fEplEqlsrVr19b2UNVqNUaPHl1Yd0QiJCTEKDIyMr7ucbm5uTr79+9PCw0NfbBs2TLLhp9DUFBQyrlz5xIvXbokAIBWrVopEhMTuQ2PIwiCaCpvZdDNzMzUsba2VgLAokWLLLdv3546bdq03A4dOlRcv3494cCBA8Ly8nIKQO3wcmRkpCw6Oppf99mqVCrlNfasV19fX/3hhx+WsdlseHh4lN+/f5+DZzA0NFS7u7tXCAQCzZAhQ8oSEhK4t2/f5t6/f5+XkpISe/DgwaQ5c+ZY1T2nS5cuVWfOnEl2cXGRt23btnLLli0mO3bsyDh58mSLusfRdP15bHl5ecxp06bZ7N27NxUAWrRoofb09CxnsVjo379/aXx8PPfkyZMCT0/Pci6XW3tyY8fVvW5SUhJ37NixDhKJRJSYmMhLS0vTAYC1a9daMBgMcDgcGgAKCwuZbdu2bRUXF8e3s7Orvn//Pnfz5s2mEolEFBQUJHz48CELAAYPHvzo2LFjLcLDw/U8PT0rnvX+PXr0iCkQCOqNJMyePTvH0dGx7Zw5c2qf2drY2CgMDAw09vb2ytLSUmbD64wdO9bew8Oj1ZAhQ0q02yiKetatCYIg3qi3Iuh26tSpYtu2bUZAzTPbq1ev6n/wwQcVV69e5fP5fI1YLJZ37Nixqk2bNpVcLpeeOHFi/k8//WRa9xoMBgP6+vrqysrK2tcskUiqpFKpLDQ0NLnusZ6enuVRUVF8ALh37x7P0dGxXk+zoTZt2igKCgrYKpUKERERfFtb22qapqGvr69mMpkwMzNTNRYkAGDr1q0ms2fPLqisrGQwGAzI5fJ6n4mbm5s8NTVVBwCUSiVGjRpl7+/vn6kdau7WrVuFTCbjAoBUKuXb2dkp7ty5wztz5kyLbt26OScmJvJmzpxp2dhxde/j6Ogo/+OPP5KlUqksNjY2rkePHhWP25cmFosrduzYYQgAQqFQfffu3fu9e/cuCQ4ONhCJRHI/P79cqVQqi4mJiZ83b14+APB4PI1araYa9tIbExUVxW/durW87jYrKyvlJ598UlC3p5uens4pLS1lpKamshsGaaCmp5uamnr3hx9+sASA+/fvc5ycnOQNjyMIgmgqzWJGjUKhoHr27OkcHx/P6969u8uqVauynJycFJs3bzZes2ZN7s8//5w9YcIEmx07dphQFIVffvklncfj0cuWLbP87bff0hpe76uvviry8vISffPNNwXa4WWNRgNHR0d5586dq57XnhkzZhSMHj3aftOmTWa9evUqcXR0VF69epW/YMEC64SEBJ63t7fLxYsXE8+cOSNQqVTUJ598UvL555/nd+7cWcRgMLBv374UV1fXaj09PbVYLBap1Wpq/fr16Q3vc/bsWb3OnTtX8Hg8ety4cYUeHh6iHj161Hv+6OnpWbV06VIdANi1a5fRnTt3dBcsWGANAKtXr87s06dPRbdu3co6deok4vF4mqNHjyabmZmpv//++4cAIBaLRYGBgdkA0PC4uvdZs2ZN1qeffmpXXV1NsVgs+vTp00lATS9+6dKleT179nQaMWJEybhx4+yKiopYTCaT9vPzyxcIBOpx48bZbdu2zZSmaaxevTpLe81du3alUxSFuXPnPjEUrBUUFGRQUFDA9vHxqe2dKhQKas2aNRanT59OunLliq52u7m5efWYMWPsUlNTORs3bnzi/Rw7dqy9RqOhBg0a9AgAzp07pz906NCShscRBEE0Farh8GVdMTExqe7u7gXA6529TDzbkiVLzPr161fapUuX5/5geJdpJ0296PHDhg2zP3bsWAqT2eggQ7MUExNj7O7ubtfU7SAI4vX4N7OXXylgctlvzxdfc7NixYq8pm7D2ygkJCSlqdtAEARR11vxTJcgAODf9HIJgiCaIxJ0CYIgCOINIUGXIAiCIN4QEnQJgiAI4g1pFkH35s2bXG0ifx8fH7uGSerLysoYY8aMsZVIJKKOHTu6bty4UdhETa2nuLiY4eXl5dKpUydRz549nYqLixkAcPLkSUH79u1dO3fu7KLN7NTQhAkTWg4bNsz+edsAYPTo0bbu7u6uV69e5QPAN998YxkXF6fT8DiNRoPhw4fbA8Dp06f12rVr5+ru7u7q7+9vAgATJ05sqc3Opa+v3x4AFi1aZK7dxuPxOuTl5b31M97S09NZ8+fPbzT7WEZGBmvSpEnWb7pNBPEcugDcAPQB0A9AfwADAAwEMAjAYABDUFNredjjP70AtALQAgDJAvOWePElQEq5O9jcl1/Xq5SrwOY2OgO6Xbt2iujo6HgA8PHxsbt27Rq/R48eldr98+bNs+jevXvZwYMH0zQaDc6fP6/30u34D3E4HPrQoUMptra2ynXr1hlv2bLF+Lvvvnu4atUqi8uXLydER0fzli1bZrFv3756a0ozMjJY6enpnLoJHhrbBgC5ublMBoOBEydOJC1btszc0dGxuqKiguHm5vZE0omTJ08KPDw8KgBg3bp15sHBwckODg7VHTp0cJ03b17+rl27MoCaikj+/v7mALB69erc1atX5+bk5LCGDRvmaGZm9tQc028LGxsb1Zo1a3Ib29eyZUtVQUEBu6ioiGFkZERKEBFvCh+ALQB7AHYKldqpUqFuRVFw4LCYliwmxS0oV1Q9LFVoVBoaNE2DBkDTqP930NCu8tTnsmEi4DAM+WwOk0Ex5EpNsVKtyWEwqHh9LusORVFJAB4ASARA1qs3Ey8eRNlcFpYZvPydlpU89V7aFIMAoKOjo7G3t68XUOoWPGAwGBgwYEC5QqGgPvjgA2elUkkJhULV6dOnk1gsFgIDA4UBAQEWBgYGKjs7O0VISEhKREQEb9q0abZqtRpTpkx5OHXq1KLGEuR//vnnLWNjY/kajQaHDh1KcXZ2fmY2JT6fT9va2iqBfwoQlJWVMbhcrsbQ0FDTq1evioULF1o1PO/HH380mzFjxsPff/9d+Kxt2vdGqVRS5eXlDF1dXc3q1avN5s+f3+gSopCQkBYTJ04sBIBWrVpVFRcXM6uqqig+n18vuPzxxx+GI0aMKK677fDhwwYDBw581PCaBw8eNFi3bp25Wq2mvvvuu2wfH59S7XrZ6dOnW7Vq1arqiy++KO7evbtLZWUlo02bNpWHDh1Ke1oBBQAIDg5OPXHihGDkyJEuarU6ys/Pz/L06dMtDAwM1FlZWTqbN29OHT58eJl2O0VRGDhw4KO1a9dm9+vXz7GwsJCdl5fH3rx5cyqLxUJAQICZSqWiqqurqZCQkKRHjx4x582bZxUSEpJiZ2fXxtTUVFlRUcEICgpKdnd3V3h6epafOnVK39fX94nXSxD/ERsAH5TJlQMoiurFYTGMCsurqzKKKzUp+RU6KQUVvMziSmQWVyGzuAoFFQrQNAQvezMemwlTfY6JpQHPxEbIb2dvrDvCxUyvysFYj7ZoweXKlZpsFoP6ny6H9T8AVwGQZYhNpFlkpAKAAwcOGCxbtsza3t5e/iK9LTabTf/vf/9L1NPTo2fOnGl56tQp/Y8++qi0qqqK8e233+b07du3bN68eVYA8N1331kdPHgw2c7OTunh4SH64osvioGatIECgUCzcOFCi0mTJhUHBgZmCQQCzYkTJwSBgYEmGzduzHp2K2qUlJQwdu/ebXLx4sXEgoICpra8HwCo1ep6wz55eXnMgoIClpubm/xZ27QMDQ01bdu2rVy0aJHlN9988zA4OLjF0aNHW8TGxvKWL1+e4+joWFuAODk5mePs7KwAgJEjRz4aPny4EwDMnTs3p+41//zzT4OVK1fW6wmePHnScMuWLfV65Gq1GuvXrzePiIiQqdVqqmfPns4+Pj6lAPDLL78I2Ww2PW3atCKgpjKQRqOBt7e3C/BPAQUnJyflkCFD7LXD7MXFxaySkhLGsWPHDFu3bl2bk3nNmjUZ2kBb9/5r167NoCgKly5dEuTk5LCKi4tZt2/fjq97nFwup8LDwxO3b99uuGHDBhNfX9/aQg/a3Nvbtm0zunHjBt/d3V3h5OSkuHXr1jtRT5hoNlrinyDbGzT0I1MKlX/J8gWRyYVIyi9/paD6PFVKNdIKK5FWWImI5EKg5rtdAAAsBoXWlvp2nR2Ek3q4mIzuaGPIUWk0hQD+FHDZFwBcAfDUEpnEf6vZBN1x48aVjBs3rsTX17fl4cOHDT777LNn9kLKysoY48ePt83NzdUpKChgaYNNUVER09bWtl4PtaSkhCkSibTl8aqzs7NZQE3awNzcXJ0NGzakAcDSpUvNr1y5IlCpVJSzs3O9APjll19ax8TE6M6bNy9HG3iAmueoY8aMsVuxYkWWsbGxmsPh0GVlZbXPyrV1frV++ukns1mzZj183ra6li5d+hAAJk+ebD1z5sz81atXmy9fvjwnICDAdPPmzbU/DGiarg3w8+bNs7569Wq8tbW1qlu3bi6TJk0qEggEmrt373LMzc2r6/4wKCkpYRQVFbFcXV3rvW+5ubmspKQkbpcuXUQAUFhYyNJoNCgqKmLt3r3bRJtuEah5hjx9+nTbrl27lgH/FFAAgNLSUqa2gMLw4cOLN2/eLGQymXTdggyNKSkpYRoYGKi1eautrKxU7du3r2jZsmUbjUZDde/evQwA2rVrVwUAEomk8uLFi/p1r1FYWMj28PAQZWVl6Vy+fFkG1HxmpBAC8Yp0AQwpkyuHUKD6gIL+jeRC1V+yfL2I5EI8eFgOAM2iwpVKQyMmswQxmSXUb1eT9RkUIDIXWHa2F37a3cVkuMTOiA0KZRoN/Zc+j/07gAsAXro0KvFszSLoVlVVUTwejwYAfX19TcPhUIlEUr5161ajr7/+ukij0eB///ufXk5ODsvZ2Vlx6tSplBkzZlhp01lKpVK9WbNmFdQNfAYGBmqZTKZjZ2enTEtL42hr7QYFBaU4ODhUe3h4uPbs2bP8+vXrgqioKNnx48f19+/fb1S3Ddu3b3+iwDwAzJ4929LLy6t86NChZQAgEAg0crmcUVJSwrh9+zbXxcWlXurG1NRUnQULFljL5XJGWloaZ8eOHYaNbZs0aVK94V+ZTKbD4/E0xsbGKrVaDQaDQZeXl9eL6I6OjvLExESOubl5JZPJpIVCoZrL5dIMBoOurq6mAODw4cMthg0bVu8HzR9//GHQp0+fJ575mJubq0QiUdXVq1cTWCwWFAoFxWAwoKurq7lx44Zs6NChDjKZTEckElUPGjSofNCgQfe8vb1dMjIyWI6OjvItW7Zk2NraKlUqFSiKwubNm0169epVPnLkSMcNGzakL1my5Jm9zfv37/PatGkjDw8Pr82/bGBgoN64cWPa1atXa3sNsbGxPAC4efMm38HBoV4hB6FQqLx586YsNDRU8MMPP5jv27cvPTk5mePq6koKIRD/FhNArzK5crIOkzH474xHytA7OYLI5EIkPix/7snNhYYG7ueU4X5OGX4PTxUAgKOJHreLk9BnXGeb/i2N+GqKonbz2MztAO43cXPfOc0i6AYHBxts2LDBDAAcHBzkH330UWl6ejpLW/DA398/Z/LkyS13795tolQqqYkTJ+YPHDiw1N/f3yI6OpovEAjUzs7O8rVr1xpHRkYKhg8f7qhQKKjMzExOeHg4b+XKlVmffPKJg1qtxuTJkx9qnyHXTZBvYmKi1tXVVXt6erq4ubm9UI7j1NRU9pYtW8w7dOhQcerUKcORI0cWzZ8/P3/hwoU5PXr0cOFwOJr9+/enAjUzh3/77beM48ePpwI1QXTevHlWkyZNKtYG2LrbGt5rzZo1ZqtXr842NzdXazQa6qOPPnJsmPR/yJAhJdeuXdPt1q1b5dy5c3N79OjhwmAw0KdPnxKhUKgGgHPnzrU4e/bsg7rnhYSEtFiyZEm9IWgAYDKZmDVrVp63t7eIoijaxcVFvm/fvnQOh6Phcrn0+vXrM7/55hvrDRs2ZE6YMMG2urqa4ejoKG/ZsqXqaQUUAODOnTv3ORwOvWTJkqe+t/PmzbMYNmxYcd3JTnFxcTpJSUncdevW5dQNumw2m+7WrZuzQqFgnDx58kFJSUntDGxtwYuysjLm8uXLMwEgIiJCb9euXU8UTCCIp3CqqlZPo0FPyH4kZx64kaZ3KiabKiivbhY92f9CUn45kvLLqb0RaQJHEz2M9mg5bbRHy68ApAi4rI0MijoM4InvJeLfe+GCB69z9vJ/xc/Pz3LKlCkF2qHkLVu2GNnY2CgHDx5c9rxz3wUajQYjRoywP3HixHuTczg0NFRw4cIFgbaS0vNkZWWxlixZYrFz585m+QyLFDxoNtgAhpRUKecwKarDoZvpzIPSDHZS/tvTo31VTAaFbk7GGOdpW9Hd2ZilUGn+p89jbwJwEcBbv8qhqbx40H0LFBQUMA0MDNRsds3S2NLSUgaLxaL5fP4znx0Sb69/G3SbOxJ0m5yeQqWerdbQfsn5Fczt15IF52JzoVC936vLDHhsDGtvSX/qaVtuZcir5rCYi5kMaieA59bLJup7p4IuQbztSNBtMjoqtWaKUq1ZeSWhgL3+fwl8Wd57MUD2r7lbG+C7QW7lrS31K3Q5rLkAgkB6vi+sWTzTJQiCaCJMAGMqFKp1MZmPdFedvq97L7v0uSe9z2IySzBqW4Sep4OR3uLBbltsjXRX6nFZfgCOAyCjis9Bgi5BEO8jCsCgcrlyQ1pRpemKU3F6N1KKnnsS8Y/I5CIMCryu94HIRG/JYLc9JgLOKgGX/Q1qlhyR4PsUJOgSBPG+6VYqV258VKl0WnHqnu7F+09dIk+8gL9k+biScEVvQBtz1+8GugULeCyZPpc9C8D1pm5bc9QsCh5cunRJV1vw4IsvvmjZcL9KpcKkSZOsJRKJSCwWi4KDg2sTIISEhAg8PDxEnTp1En344YeOubm5zJEjR9q1a9fOVSKRiLp06eL8Im1YuXKlabt27Vzbt2/vevHiRd26+4KCggzc3d1d27dv77p06VIzoGZ5j1AodH/aPfz9/U3qFhsIDg7W37x5s1HD44CaxBvp6emsY8eO6YvFYlH79u1dZ8yYYQXUTBSytLRsK5FIRB999JFd3fOWLVtmJhaLRS/y+t4Gvr6+T3z2QM067o8//tjuDTeHePdYl8qVl/LLFOeWnbzX7oOfL5OA+x+haeDM3Vx0//my7opTcR0KyhXny+TK4wCMm7ptzc0L93QVKoU7h8V56Z6xQqVQcVicRpcMOTk5KcLCwmR8Pp8eOnSovVQq5Ukkktq1suvWrTMxMTFRSaVSWWlpKaNXr17O3t7elQDwww8/WF64cCHR0NBQc+fOHY5CoWAANYkv2rRpo2jsfo0JCgoyjo2NjUtPT2dPnjzZpk+fPrXrSj08PKqioqLiGQwGOnfuLJo5c2YBAHTt2rU0JCSk0eU5f/75p35MTEx83759HefNm5cfFBRkdOTIkdSGxxUVFTHy8/PZNjY2KjMzs7IRI0aUAkDnzp1dtJmzfHx8ChvOzq2qqqLu3LnDe9HX9zbYs2dPo8t4eDwebWhoqLp79y6nbdu2L/yZEoSWRkOPlqvU23deS+Fu/SuJXa1+v2cjvy5qDY2jUZlU6J1s/rz+rgM+8Wj5gK/DmgzgD5AhZwD/IuhyWBxW2z1tX/pGd33vPvVeNjY2tSnHtIUD6u4PCQlpceHChQdATcaqcePGFQYHBxvQNI0xY8YUGhoaaoCaakUv2z47Ozt5VVUVVVhYyDQyMqqXAq1u4QMmk0kzGAwaACIiIgRisVg0dOjQYm2qxrrHKZVKMJlMev/+/S0GDRpUwmQ+WTUvNDRU393dvRL4p/CDSqWCqamp0tDQUA0Ax48fNwoLCxNMnjw5f8qUKUUA8MsvvxhPmDChcPny5ZYNr3np0iXdBQsWWKtUKsrX1zd/1qxZhRKJRBQWFibbsGGD8cOHD9n+/v45ffv2dczLy2MLhULVhQsXHjx8+JA1fvx4u4qKCoazs7N8/1vB91UAACAASURBVP796X5+fpbXr18XSKVSWWxsLKd9+/Zt7t27d/f8+fOCgIAAC1NTU2V+fj5rxowZeX5+fgV1C044OTnJg4ODU8ePH28TGxvLLykpYc6YMSNv0KBBpZ999pmdnp6euqCggH348OFkV1fXam0hBYlEIqJpGmVlZcxffvklvX///uW9e/cuCw4ObtG2bVuSqJ34N1qUy1W7SqqUfb/aH6V7N4sU23kT5EoNVpyK45z6O5uzcWyHXQY89ucCLvtzAI1W/3qfNIvhZa0bN27wioqKWGKxuF6KPoVCwdDT06sNxDY2NtU5OTnsnJwctqWlpfLJK/17PXv2LHNxcWkzcOBAFz8/v0bHnI4cOaJvb2+vMDQ01NjY2CgfPHgQGxkZKbt8+bL+jRs36vU6J02aVDB06FCHyZMnFxw/fryFUChUjR071ubkyZP1kp4nJCRw7O3ta38srF271tjBwaGNoaGhmsfj0d26dat48ODBvcuXLydu3brVNDs7m6VQKKhr164JtKknG1qyZInluXPnEm/duhV/+PBhoVwupwDg2LFjBjdv3tT19/fPAYALFy4kxcTExDMYDLq0tJSxdOlS80WLFuXcuHEjQSAQqLXD7CwWi46Li9PZtWuX0MPDo/aeM2fOzJVKpbLZs2fX/o+kVqup2bNn5xw5ciRZuy08PFxw+/bt+LrH5efns8+dO5e0fv36jBUrVjxR+zY8PFy2adOmtKtXr+oBgIuLi0Imk70zGYCIN6JXZbXqQUhM1oDeAX+RgNsEojMeofe6K7oHItN7VVWrZaipA/xeazZBNy8vjzlt2jSbvXv3pjbcx+Fw6PLy8toM9enp6ToWFhZKCwsLZVZWVqNF4p9m8eLFZhKJRBQYGFhbQq+oqIixb98+4+Tk5NiIiIj72upEdcXFxemsW7fOfNu2bRlAzZCnvr6+hs1mY8CAASXR0dH1gu6IESNKz549m1xQUMAcOXJk8Y4dO4yDgoLSd+zYUe8ZR90iBQAwZ86cguTk5Njs7Gx2WFgYz8DAQMPhcGh9fX2Nl5dX2b179zhbt241GjNmzFOnWsbHx/P79+/v7OnpKcrPz2fn5OSwAODnn382Z7PZtT9eYmJiOCKRyA2oyRmdkJDAXbhwobVEIhFdu3ZNPzMzkw0Ao0ePLty5c6cwOzubbWlp+czF8CUlJUx9ff16Y3cTJkzIt7CwaLd+/fra4Orq6lrFZrPh5eVVmZqa+kQw9fb2Fo0dO9axf//+pY/fJ1KkgHhR3AqFamNxZfWpr/ffFn53PJYrV5Lh5KaiUGnw07l4nfE7b+jnlymCyhWqnaipL/xeahZBV6lUYtSoUfb+/v6ZdYeatYYMGVK8atUqM6Amy9T+/fuFI0aMKBk5cmTJoUOHhMXFxQwAiI2N5aSlpT0zCK9cuTJPKpXKZs6cWajdxmQywePxNFwulzYyMlJXVlbWGwcuLi5mfPbZZ/a7d+9O1QYU7T0BIDw8XM/FxeWJoW2VSoVz584ZjB07tkQulzMAoKqqqt57LhKJ5CkpKZzH+ygAYLFY4PP5Gj6fTxcVFTG014qOjtZ1dnaulslk3G3btpl069bNOTExkbdq1SrTutd0c3OrvHjxYuLjIeE4e3t7JQCEhoY+AIA///xTFwDc3d0VMpkszsjISHXt2jVdJycnRUBAQMbj8+6PGzfuEQDY2tpWR0dH63br1u252QKio6P52qo/WoaGhur58+dn1+3pymQynkqlQmRkJM/Ozu6J9y48PFx27dq1+CVLllgCQGJiIsfFxYUUKSCex71coYqLSC6c2HPtX/wrCflN3R7isai0YvRa+xf/akL+mHKFKg5Au6ZuU1NoFkuGdu3aZXTnzh3dBQsWWAPA6tWrM/v06VPh6+vbcs+ePRlz5szJnzJlSksPDw+RRqPBvHnzcrWVgr777rvsvn37OtM0jRYtWqi0BQb+DQMDA02vXr1K2rdv76pWq6lFixZlA/8UKfD39zfNzMzkTJgwwR4A9u7dmxITE8NbsWKFpY6ODt25c+fyXr16VTS87q+//iocP358EQD07t27tG3btq3Gjx9fL8PXoEGDyr744gtDANi4caPx0aNHjdRqNbp27VrWoUMHeUBAgPHu3btNGAwG7ePjU2RnZ6fcunVrbTk/sVgs+u677+oNhy9fvjyrX79+ThqNBi1atFCfP38+CQCEQqF6/fr1WcOHD3fo1KlT0sCBAx2VSiWjRYsWKk9Pz0pXV1eFr6+vbVlZGZPBYGDXrl217+X58+cfMJlMXLp0qV7pvLrWr19vbGNjUy0Wi+UymUwHqEnNeezYMcM///zzwaZNm2pHF4RCoapv375OhYWFrKCgoCcmo3l7e4vKy8uZU6ZMyQOAP//8UzB9+nTyDUo8VbVKM0Op1qxZEhLLPXY7iwyLNENlChWmHrjN+6iDlc2qj9pEcFlMXwaDOtrU7XqTXjgN5Oucvfy++/LLL60XL16c21gv/12krab0tJnfDcnlcmr8+PG2R48eTX3NTWtyJA3kS2FUVqvWFVcoJ3+yPYKfUfRCRcKIJuZmoY8DX3au4uswF3NYzHVN3Z435d/MXn6lgMlhcV7l9Hfa02r1EjW4XC79PgRc4qXolCtUQWmFFf3H7bjBf1T5n8yrJN6AuJxSDNpwjXdoitdyoa6OvS6HNRPAO//wvVk80yXeLyKRqPpFe7kE8QyCcrnqUlRa8YCRW8N1ScB9+2SXyDF44zXdxIflE8rlqhAA7/wKBRJ0CYJ4G5mXK1TSM7E54om/3+ST2clvr9IqFUZvi9ANSyroXS5XhQFoNHPfu4IEXYIg3jbOFQpV9I5ryY7zjt7hqjUk0dHbTqHS4Kv9UbzDtzLaVChU0QDsmrpNrwsJugRBvE0kVdWqmytD40x/uZj4r9boE80bTQMrQ+N01l2QWVVWq24D6NDUbXodmlXQfVoCf4lEIlIqa57XfPrppzYrV640feLk1+jo0aP6hw4dMgCATZs2CePi4nRmzpz5RPrFV1W3UMOvv/76xBCLr69vS4lEIvr444/tVKp/Jjpfv36dT1GUWPseaR08eNDA3d3dddasWZYAcPv2be6CBQueyP4EAEuWLDELCwvjPa34hEAgaC+RSEQSiUSUl5fHBAA7O7s22m1RUVHvxLOYRYsWmaekpDT6ZT516lSrpKQk8kXfdHpXVqsuTz8YbXDoZkaz+u4i/ju7wlKZ3x6JaVFVrb6Kd3At7wvPXtYoFO4MzssvGdIoFCoG5+kzoF8kgX9AQIBxVVUVY/HixW+0NIiPj09pnb8/+vjjjx369u37WnLKPa1Qw5UrV/hKpZIhlUplS5cuNTt06FCL8ePHPwKAwMBAEzc3t8qG5xw4cEAolUplQ4YMcQCAdevWmW7atOmJmdJqtRoRERF6K1asyEtPT2c1VnzCxcWlSiqVyuqeJxQKlQ23ve1Wr1791Nywn3/+eeGGDRtMGhafIN4Ij8pqVciE3Tf5UlL39p13NjaXYjJidP192l3m67A6AEhv6jb9V1741yKDw2Hdd22Fl/3zvICtTeD/tP0XLlzQ27Vrl8m+ffvStNsmTJjQslOnTqKePXs6FRYWMmUymQ6LxRLfu3ePAwAeHh4iPz8/S6AmYIvFYpFYLBZdv36dDwAikchtyJAh9q1bt2515coVPlCTbEJ7fe3fHyfxN5bJZDpTpkyxuXbtWmJCQgJ32LBh9k87Z8OGDUKJRCJq06ZNq2PHjj01oURdFEXh008/te/Vq5dTQkKCTt19CQkJnLZt21YCQMeOHSvDw8N1AeDWrVtca2trpa6u7hMzSdhstkapVIKiKERERPCcnJwU2uIQdUVGRvK06R1tbGxUfD6fBuoXn0hKSuKJxWLR1KlTrTSamks8evSI1alTJ9HYsWNtKysr6yUjKC8vp4YMGWLv6enpMmjQIAeFQkFp38fExESdbt26ORcXFzN+/vlnY4lEIhKJRG5Xr17lA8CcOXMsJBKJyNPT00Umk+k863Ot2wOv+zlotwsEgvYAcOzYMX1XV1e3usdJJBLR5MmTrd3d3V3Xrl1rDNSMNsTGxnICAwOFdnZ2bdq1a+c6efJk68f3ld+6dUvvRT5L4j8lqqpWX5x5MFqXBNz3R+idHCrgQoJBhUJ1Fe/Q5KpmMUTzvAT+ADBnzhwbJpNJl5aWMoCanl9lZSXj1q1bso8//rgoICDABADEYnHZrl27jO7du8fR0dHRAEBOTg4rNDS0xc2bN2Vnzpx5sGzZMgsAyM3N1dm/f39aaGjog2XLlr3wcHFiYqJORkaGzrOOmThxYrFUKpVduXIlYd26dWYvct3NmzdnREdHx8+fPz935syZ1nX3tW7dWn7t2jUBAFy8eFH/0aNH2nzKZnPnzm205z9nzpy8UaNG2Y8YMaI4MDDQtEePHuXjxo2z2b59u2Hd4+7fv8+1s7Orl1O5YfEJmUx29+bNm7JHjx6xDh48aAAA4eHh8bdu3ZLZ2NgotO+/1oYNG0wGDx5cEhkZmdC9e/ey33//3RAASktLmZ9//rnt3r17Uw0NDTVz584tkEqlsrlz5+ZEREToSqVSXnZ2to5UKpVt2bIlffny5RZA458rAGh74HV73CqVCq6urlVSqVTWqlWrKgA4duxYC39//4yGPXMfH5/iqKio+AMHDhhrC0NozZw5Mzc6Ojo+OTm5dpF5dXU1yXT0ZllVVquuLTkZq0dq375/dlxPYR65lWFeLlddwDuynKhZBN3nJfAHgC1btqTOnz8/Z9GiRZYAIJPJOB06dKgEAC8vr8qkpCQOAJibmyuzs7N1du7cKdReMz4+Xic+Pp7n6ekpGjJkiFNpaSkLAGxsbBQGBgYae3t7ZWlp6ZN1955i5cqV5nPnzq0dhiwsLGRre1v5+flsADh+/Li+RCIR9e/f3yknJ6degN68ebORRCIRLVq0qN7zVTMzMzUA9OvXr1x7HS1vb+8qNze3qs6dO7uUlpYyTE1NlXfv3uXo6+urLSwsGs1k1aVLl6ozZ84ku7i4yNu2bVu5ZcsWkx07dmScPHmyRd3jGmYla6z4hJmZmZrBYGD48OHFd+/e5dVt76hRox7du3ev3qOB+/fvczdv3mwqkUhEQUFBwocPH7IAYMuWLWZyuZwhEAhqA+eIESPslixZYt2nT5+yO3fucMPDwwUSiUT01Vdf2ZaVlTGAxj/XpyksLGTq6emp6277+uuv87/55hubhnMGvLy8qlgsFqytrRVZWVn1RmMCAwPNHR0d2zg5OZEavk3DoEKhurbxzweGf9zKbBbfVcSbtyI0jhOWVOBWLlcdBfDC39PNVbP4h/y8BP4A0LVr18qxY8eWJCQkcO/du8dxdnZW3L59WxcAIiIi+A4ODrVfjF26dCn/+++/+drem0gkqm7Xrl2Ftkd07dq1BABIT0/nlJaWMlJTU9kCgUDd8J6NyczM1GEymbSTk1Ntz1D7bFMqlcpMTEyUAPDzzz9bXLp0KeHs2bMPGIz6b/O0adOKpFKprOHzQ21xg5iYGI6+vv4T7Vm7dm3OjRs3EoRCoXro0KElt2/f5j0uROAsk8l448ePt22szVu3bjWZPXt2QWVlJYPBYEBbfEHLzc1NnpqaqgM0XnyitLSUoZ24df36dT1HR0eFXC6ntAUa/vrrL7267//j91zu5+eXK5VKZTExMfHz5s3LB4Cvv/4678cff8z08/OrreR07Nix1F27dqUEBASYtm7dWt6jR48S7ftZNxNVw8/1aSIiIvgikahecQRHR8dqBwcHeVRUVL2e7o0bN3gqlQqZmZkcKyurej9eZs6cmZuamhobHR3N1wZkHR0dsj7lzWCVy5WnQv7Ostx6JalZ5IgnmgZNAzOConkJeWU9KxSqLQDe6tGmZvGPubEE/unp6azNmzcbr1mzpl5g+umnnzLnzp1rdebMmeRdu3ZpxGKxSE9PTx0cHJxSWFjIBICpU6cWTpkypfD8+fMCALC0tFQNGDCgpFOnTiImk0l379697Oeff84xNzevHjNmjF1qaipn48aN6QBQUlLC0j6rTUlJ4e7Zs6derzA2Nlb3yJEjySqV6pkffL9+/R55enq6dujQoaKxANoYHx8fh5KSEiYA/Prrr2lAzfPkDh06VHl7e1d6eXmJmEwm3aNHj1JtgQVfX99HQM3zyf3796c1vObZs2f1OnfuXMHj8ehx48YVenh4iHr06FFvGN/T07Nq6dKlOkDjxSf4fL5m0qRJdjweT2Nra6sICAjIzsnJYfXr18+Zz+drDAwMVEePHq2XYWr27Nn548aNs9u2bZspTdNYvXp1FgAYGBioP/zww4rdu3cLL1++zD979qzBlStXBNXV1Yyff/45w8vLq+rEiRNKiUQioiiK/vjjj4uGDBlS2tjn2hiZTKazceNG071799Z7L7799lurlStXPjEB6tChQ4Z+fn4tx48fX8jlcusF1MDAQPP9+/cbGxkZKc3NzVVSqZTXsWPHJwpbEP+9CoVq873sUvHikHskfyyBarUGE3ZL+SdndB3HZFApXDbzp6Zu08t64YIHr3v2clMQi8Wihj2fupKSktibN282CQgIeOdnqy5ZssSsX79+pV26dHlvssVLJBJRWFiYjM1+sVVAU6dOtfLz83vo5OT02vINkoIHgFKtmZZXKvcfuOEav1T+XtQAIV6QuT4Xp2d2rTTgs6ewGIz9Td2el/HCQfRVAyaD8/b9YLWyslJ9/fXXBc8/8u23YsWKvKZuQ3O3ZcuWrOcfRbyiD+RK9c9jt9/gkYBLNJRbKseY7ZH8E9O6/MbSYUQDuNfUbfq3XrinSxDE6/ee93QNKqtVD6buv238VxMXn9fIy1F4NhDVBTXLQ40HzoKqrBAl14OgLMyA+WcB4Fg4v/C5HKtWqM5LRuH5zaDV1aAYTBh9+DU4liJUyMJQcu0AGDw9mIz4HkyePpTFOXh0dS9Mhs1/Y6/5bTJG0lLz3UC3B3pcVlsAz5zj0dw0i2e6BEEQ5XLV1jOxOXpNHXABoOjP38B1EMPko0Wg1UrQSgUYHD2YfLQIhec3/etzAaD4r91o0WUMeI6dUJV0E8V/7Yb52J9QJj0O80/XouL+VVTEXYG+eAgeXduHFt3Gv4mX+lY6KM1gDGxrYS22NfyBr8Oa19Tt+TeaxexlgiDee0Mqlaphy0/ea/K1mBpFJeQZ96DXri8AgGKyweDqgW3cEmyh9UudW7u/urL2OKaesGYjxagJzioFKAYT8oxYMHUNwTayeuL6xD/8DsfwVRp6OgDvpm7Lv0GCLkEQTc24slq1d3pQNL+i+oUm+r9Wqke5YPL1UXjmF2TvnonCs4HQVMuff+JzzjXqPRnFl3cjc8sEFF/eCcMevgAAgy5j8PDIEshT/4auWw+UhB+GQZcxr+31vSvyyxWY+8cdXoVCdRTAW5MprlkE3aNHj+prk0uYmJi027dvX71lOiqVCpMmTbLWpvALDg7WB2qSTIwdO9YWqFkqol3qc/fuXY5YLBaVlZUxJBKJyMPDQ6S9/n/R3kWLFpl36tRJJBaLRRs3bhRqt4eFhfG8vb1dPDw8RN26dXO+d+8ex8/Pz1IkErk1LBbwLPPnzzc3NTVt11hRhbKyMsYHH3zgJJFIRL1793bUrpXdtGmT0MvLy0UikYgaJuyfNWuWpbu7u6s2k9TatWuNz5071+g/0o8//tiusrKSioqK4nbs2NG1Y8eOtQUTiouLGb169XLq2LGj66ZNm4SNnf+2CQ8P561fv964sX0RERG8xYsXv1A2MeLllcmVuw7dzOA1lxSPtEaN6twkCDoMhOXngaDYHJRG/vHK55b9fQaGvSfBeurvMOz1JQrPbgAA8Ow7wGLCBpj6LEVlQiR4jp2gKsxE/vHVNUFb+WIB/310/l4uLt7PMyxXqDY3dVte1AsHXZVS7Q5A/LJ/Hp/fKB8fn1JtMgQLC4vqoUOHltbdv27dOhMTExOVVCqVXb58OfHHH3+0yMnJYU2bNq2IwWDQdQNAeXk5NWHCBLvdu3enarMehYeHyxqmCnxZR44c0U9NTeXcunVLFhkZKTt+/LhhVFQUVy6XUzNmzLA9dOhQ8s2bN2UHDhxI0SaUWLNmTYb2/tosTs8yffr0gp07d6Y0tu/YsWP6nTp1qpBKpbJOnTpVBAcHG6SkpLCvXr2qFxERkSCVSmX29vb1lrTEx8fzpFKp7MCBA8Kqqirq1q1buv379y9veO27d+9yDA0NVXw+n964caPJqlWrMm/fvh1/8+ZN3YKCAuaGDRtMRo0aVSSVSuN///33J9Imvo28vb2rZs+e3ehkQS8vr6rIyEg9tbrpe1/vKo2G/uRRpbLXmrPxzWZ5A0tgDKbAGBzLmt/ofFEXVOclvfK55Xf/BN+lZiSU79oVipyEeudqlHKUx/4JQYdBKL66B8KB30DH3AkV9/76j17Zu+n747FcuVLtA2BQU7flRbzwRCoWm8na/NWll77RtF97PfdecXFxOsbGxioDA4N6SflDQkJaXLhw4QEA6Ovra8aNG1cYHBxsMH369MIdO3ZkSCQSkZGRkQoAfH19bSdOnJjfrl2756buk0qlvOnTp9tkZWXp6OnpqQ0MDNRSqVQ2cOBAh/z8fLaOjo7m1KlTSUZGRrXtOXLkiNHSpUtzAIDNZmPWrFl5Bw8eNOzZs2dZ165dS7VZnGxsbFTav/9bLVu2VMXENL5Cy8XFRXHr1i1doKbggImJiSokJERfrVZTXl5eLi4uLlU7d+7MYLH+ebspiqIVCgWlo6OjWbduncmUKVManaly7NgxA29v73IAaN26dVVxcTFT+8OBy+Vqbty4obtjx450FouF1q1bV965c4crkUhq1/UGBAQYHzhwQAgAGzZsyOjatWuldi30yJEj7Xx9fQtbtWqlGD9+vF1JSQlr8ODBxWvXrs25dOmS7oIFC6xVKhXl6+ubP2vWrEKJRCLq2rVrWUBAQHZAQIDx+vXrzTMyMmJHjhxpl5iYyOVyuXRycjL3r7/+im/Tpo1Cu72srIw5e/bs3AkTJhR/+OGHTiqVikpLS+P+9ddf8ZcuXdI7d+6cQWlpKZPD4WhOnz6dfPHiRb0LFy4IAgMDswUCQXtXV9cqpVJJnT179oGFhYXK0dFRERERwe/atesTVZyIV2ahUGl+mxZ0W1eheqIOR5Nh6hmCpW8MZWEm2EJryNNiwDa2eeVzmXpGUGTcBdemXc12w/oDWaU3gqHfaSgoJgu08vGEXIoBWkWykD5LmUKFaQdu83d/7rGPr8NyAdCsV9w0i+FlrUOHDhkOGzasuOF2hULB0NPTq13bZGNjU52Tk8MGaoZb2Ww2/d1331lfvXpVPy0tjSOTyepNxvD29hZJJBLRxIkTW2or5ACARCKpkkqlstGjRxdqe6MAcPjw4dSbN2/KRo4cWbx79+561S3y8vLYtra2tT1JOzu76pycHHZWVhbbwsLitSVN0GrTpo0iKipK18nJqfXff//N79OnT3leXh67urqaioiISODz+ZoDBw7UG54fOnToo08++cRu6tSp+XFxcdwHDx5wxo4dayOVSuvlS05MTOQ6OjpWA0D//v3LFi1a1NLBwaGNRCKp0NPTo0tKSpiGhoZqoCazlDYDGPD0ohJAzfB29+7dy4YOHVrm7OxcfePGjYTo6Oj7N27c0AOAJUuWWJ47dy7x1q1b8YcPHxZqe9CZmZlsjUaDyMhI3brvbVBQUIpUKpX16NGjtryiWq2mjhw5kjx79uxcAIiKiuJaWlpWR0VF1TtOX19fff369URPT8+KvXv31nufXFxcqm7evCkbMGDAo7i4OA4AODg4KGJjY5t8cs87iCqTK4N2Xk/m3sl8LVUyX4lRn69QELoW2bumo/phCvS9RqEyIRyZm32hyI7Hw6PLkXd4MQBAVVaIvD+WPvNcABAOmIHiSzuRvWs6Hl3dC6P+M2rPUZUVojr3AfjOngAAfclHyN03BxWxf0LX7YM398LfUjdSinBQmqFbJlfuaeq2PE+zCrpnz55tMXr06EcNt3M4HLq8vLx2KDM9PV1H+yW8ePFii0WLFuWsWLEiy9zcXBkWFpZw+/Zt3fT09NqunnZ4mcFg0CEhIc8ss6dSqTB16lTrTp06iX777TfT7Ozses9HzczMlGlpabXbUlNTdSwsLJRWVlbK7OzsZ1YeauhphQ+eZcuWLcLevXuXPHjw4F6/fv1Ktm7dKjQwMFB37969DAD69OlTFhcXVy9ITJ06tejMmTPJV65c0ZsxY0b+yZMnW+zcuTPD39+/3vPKumu2FyxYYLVnz57klJSU2Hv37vFkMpmOgYGBuri4mAnUVAsyMjKqHXd9WlGJ+Ph4/sWLFw14PF7tr53t27cbOjk5tfnggw9Ktcf079/f2dPTU5Sfn8/OyclhAUDXrl3LN27cKLS3t3/uT/3S0lJm3RESLy+vKqVSyXBycmp95coVA+329u3b15ZHfPDgQb33KSEhgScWi0V//PGHsF27dnLte0JRb/0oenM0+mGZwmPDn4kvlg7sDdMxc4CF7y+wnLgJpiO+B5OrB76LN6yn7YHtnBNoOWM/zEavBACwBEKYfbz8mecCANe6NSwmbIDlxE2w+CwAHHOn2nNYAiFMff4J3LquXWE5aQvMx/8MJr/2ny/xDP7n4nWqlOoeAHo2dVuepdkE3fT0dBabzdaYm5s/8QBtyJAhxatWrTIDapLv79+/XzhixIiSlJQUdnx8PHfkyJGlEomk0sXFpYrD4dCLFy+urUZUl6Ghobph3deGIiIiaksGfvnllw9pmq53vI+PT/GPP/5oDtQUB9iwYYPZ6NGji3v37l0RFhYm0Ab7rKwsVkxMzDOfUz2t8MGz0DQNbbAzNjZWlZSUMLt3715+584dPgDcvn2bb29v/8Ri8eLiYkZqairHy8urS/a1cQAAIABJREFUSi6XMxgMBqqqqup9/i4uLvKkpCQd7X2MjY3VTCYT+vr66pKSEmbnzp0rQkND9VUqFWJjY/nawAQ8vaiElZWVIioqKn7Pnj3G2p7xl19+WZyenn738OHDQgBwc3OrvHjxYqJUKpXFxsbGaZ9Jjx8/vnjVqlVWX3zxxTNn2CiVShQUFLDqVlvicrm0QCBQHzlyJKluT1f7PkVHR/McHR3rBXMXF5eqqKgo2cyZM3MDAwONASAlJYXTunVrMpPlv8UqV6jWLT4Rq6tUk/oRxH9DodJgxak43TK5ciuaUWxrqNk07PDhwy0GDx5cr5fr6+vbEgDmzJmTn5eXx/bw8BD17NnTed68ebmWlpaqhQsXWi5fvvyJvMj9+vUrz8vLY8fGxnKAf4aXo6Oj+SNGjChteHxd7dq1k6empnK6devmfPPmTd2G+z/55JMSGxsbhVgsFnl6eoqGDRtW7OHhIedyufTGjRvTPvnkEwcPDw/R6NGj7dls9kt9o6xfv954/vz51sHBwcJPP/3UBqiZMZ2SksKeNGlSUXBwsKFEIhEdOnTIaNKkSYXe3t5VPB5PI5FIRFFRUfwJEyY8MUT/008/mfr5+T0EAFdXV7mHh4do7NixhXWP+eijj0rCw8P1AGDhwoW548ePtxeLxSIdHR1aIpFUzZw5M//QoUNGHh4erp999lkBj8erfX11i0p07tzZ5fvvvzcHAIFAoOZyufSSJUty/Pz8LC9fvsx/PMPbdeDAgY8AYPny5Vn9+vVz6ty5s8vQoUMdtNcUCASajIyMuy4uLs/MOPPFF1/YzJ49u14aywsXLugaGBioO3bsWC9gFhUVMbt06eIcHh4u+Oyzz+q9TwkJCTyJRCL69ddfTfv161cGAImJiRwvLy/yPPc/pKHpzxLzyvTDkwqffzBB/Auhd3KQ9ajKCkCzXXP1wmkgVUq1O4vNfOkMViqlWsViM5tVwQPiSaNGjbL9/fff0/l8/jvXBQkMDBSqVCrKz8/vhSZaRERE8EJDQw1WrVr1wiMRr+o9SAPJqVCoMj7bJTWJSnvityFBvDIPO0Ps+VxSwOewWgJodqNU/2b28isFTBb7ra89/F44cuTIE+UB31deXl5VXl5e703VpTdBpdZMisl8xCcBl3hdbqYWQ5paxPN2NJ6pw2L4N3V7GiIFDwiiGXnHe7r8ympV5qhtEYaxWc98ykMQr8TFTA8h07qW8nSYlgCaVQ3sZvNMlyCId1u1SjPtRnKRDgm4xOuWkFeO6w8KmNUqzbSmbktDJOgSBPEmCNQaevGPZ+8/MTmRIF6Hn8/H66o19HcAmtW/ORJ0CYJ47eRKtd9l2UNWQt4T2UcJ4rVorr3dZhN0n5awf+XKlaYSiURkZWXVVls44PTp0822osTWrVuNOnTo4NqzZ0+noqKieu9vYmKijqenp0unTp1EO3bsMARqiiS4uLi4WVlZtW3seqNHj7Z1d3d3vXr1Kh8AvvnmG8u4uLgnknBoNBoMHz7cHgAWLlxo3qlTJ1GbNm1aabMuhYeH8zp27OgqFotFTyt28LYhxQreGgY0jbn+5+J5zz+UIP476y7IdFUazQI0o9rxL17woLr61Qoe1JzfqGcl7F+8ePHDhqkaBw0a1Cx/LisUCmrHjh0mUqk0fuzYsYXr1683qbt/xYoV5j/88ENWZGSkbPfu3cZKpRJubm6KW7duxZubmz+xFjU3N5fJYDBw4sSJpJ07dwqzsrJYFRUVDDc3tyeOPXnypMDDw6MCAJYtW5Z369YtWVhYmGzdunXmAPD9999b/Z+9+w6Pour+AP6dsjtb0zshpJBsChBCSOgEBOmhFylKFwsCKoKdovJaUOwoKMiPIiIgoYmCClICCUUECSEJKZAE0sv23Zn5/REChBogyabcz/P4vL7slLNLzNl7595zfv7557T9+/envPfee563nt8QkWYFDcbooxcLkVFItjsTdev8lXJcKjKwAPrYOpZK1d8yJJWyH48Z9NA3evmnnXe91/0K9t/JG2+84fHrr786SKVSYe3atRmBgYHmWwvWm0wmasyYMX5Wq5UKCQkxrFu3Lmvnzp3qTz75xN1qtVJms5mKi4tLKykpYebNm9csLi4uPTk5WVr575MnT25+9uxZhSAI2LhxY3pgYOA9izScOXOGCwkJMUgkEsTGxpZNnDjRF8D1og2ZmZlcdHS0gWVZuLq6Ws+ePSuLiIi46z4yjuNEi8VCabVaWqlUCkuWLHGfP3/+1TsdGxcX5zBlypTCyvMAQKfT0cHBwQYAKC0tZQICAiwAoNfraa1WS91cz5o0KyBqS6nBMnP9scx69VyNaDr+Lz5DNa9f8Ex7uWS3rWMB6sn08v0K9t8qKyuLPXDggPrkyZPnFy1alLNw4UIP4PaC9R4eHtZDhw5dOHHiRHJ5eTlz5swZDgCMRiP1999/p0ydOjX/s88+c73bfT7//PPsxMTE5Lfeeivn888/v+txlYqKihg7OzseAJydnfnS0tIqm5Nbtmxp/PXXX9Xl5eX0qVOnlDc3DLgTR0dHoXXr1vrXX3/da8SIESWiKGLz5s0OEyZM8ElLS6tSs/bixYtcYGDg9bKGEyZM8AkPDw/t2bNn+bV4rImJibKcnBw2JSVFXlhYeP1bDWlWQNSiQJpC4P7kOza2Iohat+N0DsWx9GMA6kUP8HqRdO9XsP9WKSkpXFhYmAEAunXrpktPT+eA2wvWX716le3fv39AdHS05vjx46qsrCwJALRp08YAANHR0fq0tDQOAA4dOmQXHR2tGT169PUyhAsWLPCIjIzULFiwoFllV6NK06dP946OjtZs3rz5egMFJycnvqysjAGqJuBKixYtyl25cqVLbGysf0BAgNHLy+u+XYkWLFiQt2PHjvSffvrJ8bnnniv4999/FQsWLLjyySefuN183K01otetW5d1/vz5/5YuXeoJAEuXLr08Z86c5pMnT26h0WgMHh4e1+sUk2YFRG0xWvgpW05kM1ah0RU4IxqIMqMVf57P43lBHG/rWIB6knSrU7D/ZoGBgaazZ8/KAeDgwYNKX19fE3B7wfpVq1Y5DR48uCQhISE5MjJSW5mYKs9NTExU+Pv7mwCga9euZQkJCcmbNm26CFQ8Tz106JD6xIkTyQsXLsy5tYjIypUrLyckJCSPHDny+qbD1q1bm86fPy+3Wq3YsWOHXXR0dJVnz82bN7fu27cvbefOnWlSqVQIDg6+5/uslJycLJXL5YKLi4uV53nQNC1qtdoqf3cBAQHGlJQUDgAMBgMFAEqlUlCpVDwAtGnTxnT48OGUH374IbNZs2bmyilogDQrIGoNzQvitE3HLz1Q9y2CqGnrj2UqdCbrLFvHAdSTFV2dO3c2/PDDD8K1ZvSWt95662pWVhb71VdfuXzwwQe31b318fGxdu/evTwiIiJYIpGI69atSwduFKzXarX0ihUrMimKEidPnuy3ffv2KlOREolE7NatW6DJZKK3b9+eeus0MAC4urrySqWS79ixY1BoaGi1SgFyHCdOnjw5PyoqKtje3t66ZcuWdKCiccOaNWsubdy40X7ZsmXuDMPggw8+uEzTNFJTUyVPPfWU34ULF+SdO3cOWr16dYZGo6mSjD/44AP3JUuW5Hh4ePCCIFDDhg0L+OKLL7JuPiY2Nrb04MGDym7duumnTp3aPDU1VW6xWKjKKdtly5a5bNy40UkulwvffPNNlXNvblbAMIzYvXv38o8++ij31mYFkyZNKnzjjTe8jUYjfWuzAkEQ4ODgwP/2229pwI1mBTcn9zt5mGYFHMeJu3btSvvjjz+ur8K+9e8eIM0K6oEe+eUm7lwuKYZB2NaRtEJYeMETQFsA/9gyluo3PDCbw1mp9OEbHpjNVlYqtXnDg507d6orF9/YOpaaJAgChg8f7rdt27Z0W8dSGxpCs4Ka0JjKQJYZLT9/ti9lxPeH0skcP2FzLz0eZJ3a1e87Jcc+a8s4HmT18qM1PJCSGabadG1rUaNMuA+DNCuwORXH0IPi/skmCZeoFzYdv8Q+3d1/AoA5AO671qS21Ivp5bo0aNCg8kGDBpXbOg7iwcyaNYs0X21YRiZkFFkLtNVatkBcQ1GAt6Mc3o4KeDvK4ariQNMUKAC8IKJQZ0ZOiQHpBTpcLibfKR/E5WIDzl8pF9s2d4gFsNlWcTS5pEsQRO0r0Zuf3XAsq1FUPqttPk4K9ApxQw+NW3lUC0epCOjMViGLopCq5Nh0ioKVAiUKoijRm/kWgiAGcCztz4uiLCG9iD9wIV+15+wV5JXbbPDWYPyUmKX2c1GOt5dLSNIlCKLRUCikbNsDF8je3LuRMjQGhXtiWlf/cj9XJXhejFPJ2N0ADgDIUXK3n8OAgr28yqaFFr1C3LtF+TkNen1AyOB/LpVY18Znqvf8dwU82aJ1R/FphWBpqjsACoBNPiSSdAmCqGldU/LKDXozTxZy3IKmgOHtvMXX+gcbWIb+x14u+RDAbkhw3z37d5AJINNOJlkHQNHR33lYqKfd3DcGhgQu2Z2k3HUmF/dYJ9skZRTqYRVEOQA/ABdtEUO92KebnJwsdXZ2Do+OjtZ06dIl8NbXR4wY4dumTZvg6OhoTXR0tCYrK4t8WXgIGRkZko4dOwZFREQEb9u2TX3r6xMnTmweHR2tGTVqlK/VWrFttlevXgFqtbrtnY7/8ccf7cPDw4Nnz57tBQAnT56Uvfrqqx53uvfbb7/tfvjwYXliYqIsIiIiODIyUjNy5EhfQRBgsVgwaNAg/w4dOgQ988wz3jX8tm1m4sSJze/22tChQ/0EQbjbyw2a0cL3/SMpj0wt3yKyhSMOvNJT93Zs6D/OKu5xe7mkC4A44KES7q30ANbbySXtvBzkw/43vHXSvhdjdK2a2d33xKbm6MVCHkAPW92/2klXtAiP1PDg2vl3VVmc4vDhwyl3er2yfGBCQkKyj4+P9U7HEPe2cOFCj3feeSd7//79F5YsWeJ182sHDhxQWCwWOiEhITk0NNSwceNGBwBYtWpV1vTp0/PudL3169c7JyQkJCclJckB4OOPP3abP3/+bcfyPI/4+HhVly5dDG3atDGdOnXq/IkTJ5IB4ODBg4q1a9c6tm7dWn/s2LELBoOBio+PbxTdaNasWXPpbq9FR0drt2/fftsXmcbAZBViD6UW3LPEaVMiYSjM76cxr50aXdrcSfGUnUwSCeBILd1OBLBXLZOE+bkqn/l5Rufylx4PsjA0WURe6UByvqrUYOlvq/tXO+lSEpq9/OpBPOw/lIS+5+g0Pj5eHRkZqVm0aJHbvY6rZDKZqE6dOgW1b99e07dv3wCr1VplxBwdHa0BgMmTJzePiorSREZGalJSUqS//PKLXXR0tMbFxSW8TZs2wYMGDfIHgEmTJjVv3769pmfPni0LCwuZm68VHh4ePGTIED8A+Prrr53Cw8OD27VrF1yZHHx9fVtFR0drwsLCQk6fPs0BwNy5cz2jo6M1HTt2DEpOTpYmJydLO3ToENSrV6+A8PDw4PPnz0vT09MlXbt2DbRarXj++eebrV271sFisSA2Ntavffv2mtjYWD+LxYKdO3eqvby8WkdERATHxsb6AcDmzZvtoqKiNK1atQr58ssvq1VT9Ny5c4pevXrp7O3tBZVKxRcXF1//+79w4QLXunXr62UWjxw5ogSAFi1a3PVbuEQiESwWCyiKQnx8vLxly5YmR0fH24ZvR48elXt5eZmBG80YAEAqlQp+fn7mtLQ0Ljw83AAAbdu2NRw8eLDKKOns2bNcly5dAqOiojTz5s3zBCpmP86ePctt2rTJbty4cS2AinrT7dq1C27fvr2msLCQ0Wq1VGxsrF/Hjh2DBg4c6G8ymajPP//c2dvbu7UgCCguLqaVSmXEzp071ZWfcXR0tCY4ODh01qxZXgBw859X/ky9+uqrHq1atQq5+TiNRhMaGxvrFxYWFnLgwAEFAERGRmoqY23Tpk1wSEhIaGVLx759+5Zv27btnjXGGyhHmYRucSqr2NZx1AvOSiniZnbVTejY4pBCymoAbEXdPEsUaYpaJ5cywVO6+B7fOL2jTsWRCUIAOJZeCJametjq/vVietnHx8eSmpp69ujRo8l//fWX3bFjx24b6YwbN86vQ4cOQSNGjPDVarWURCIR9+7dm3L8+PFkjUZj2LFjhx1wY8SckJCQDNzetGDYsGFllQX3N2zYkL5z586LBw4cUOj1evr48ePJo0aNKvrkk09cb75WZWlIq9WK5cuXuyckJCT/+OOPF19//fVmAODs7GxJSEhInjVr1tVjx44pEhIS5Dk5OdKEhITkr7/+OmvRokWeAJCfny/Zs2dP2rJlyy4tXrzYw8/PzzJs2LDicePGtcjKypI++eSTJWvXrnUMDg42Hj9+PDkkJMT4ww8/OALAyJEjC0+dOnXeaDTSPM+jb9++2sTExORTp04lrV69+o49ZW91rYQkAECtVvMFBQXX/ysMCwszHjx4UA0A+/btsyspKbnvf6Fz5869Onr0aL/hw4cXf/75524xMTHa8ePH+6xcudLx5uOSkpJkvr6+1/eOrF+/3j4wMDCsoKBA4u7uzgcHBxv379+vBoD9+/eri4uLq4yS5s+f32zNmjUZiYmJyUlJSbLKZg/Hjx+Xr1y50nX16tVZQEW96ZMnT54PDg42ZGZmSj777DPXQYMGlR49evRC9+7dyys/y+bNm5t27dqlXrNmjWN4ePj1Up0jR44sTEhISF66dOn1ESrP8xgzZkxh5c8TAMTFxTkdPXr0/M3HXblyRbpu3brMnTt3pi5cuLDKLAJQMVOzZ8+elD///FMNACEhIaaUlJTG2IyhQ1JuucHCk4eJPk4K7JzVVe/nrPxSLZP0xk0dx+pQjkom6RbqZffTjhe66Nzt7rBCq4lJy9eBF0QlAF9b3L9eJF25XC7a2dkJEokE/fv3Lz116tRtSXfDhg3px44duxAREaH/9ttvncvLy+mxY8f6RkVFaXbs2OF4+fJlyZ2ufa+mBZWSk5O5iIgIPQB06tTpehOEW+Xk5LDe3t4mjuNEjUZjrmxuUFhYKImKitIsWrSoWUxMjO7ff/+VHTlyRB0dHa155plnWpSXl9MAEBwcbJBIJOjUqZM+IyNDBgAzZswo3L17t9Ps2bPzACA1NZWLjIzUA0B0dLQuNTWVA4DNmzc7azSaUI7jRIqicPjwYUXnzp2DOnfurElJSanyeVWO5qdMmVLlmSLD3Mhl5eXljLOz8/Vp+s6dOxtCQ0MNHTp0CCorK6Pd3Nzu+5ypS5cuht27d18MCgoytm7dWv/111+7fvfdd5duLbt5a9Wz8ePHl6akpPzn6elp/umnn+zHjh1bYjAYqE6dOgVxHCe4u7tXuXdaWpps3Lhx/tHR0ZqUlBR5ZmamFACWLl3qSdP09dFzYWEh07p165Bz584pfH19zUlJSbKvvvrKLTo6WrNhwwbnvLw8FgAGDRpUsnXrVocjR46oOnbsqLvXeywpKWHUanWVxhUvvvhibkBAQOu5c+de/3x9fHxM9vb2gp+fn6Xy5+Jm48aN84uKigqJjY29Xku6MTZjsPBC5/i0gibfxi/AVYUdM7sYXFTcPLmUeRU2Wil7Da/k2Gme9vL/7Xyhm54kXiAho4gHEGOLe9eLpHvzNOeRI0dUQUFBd91w5ujoaDUYDPQvv/xiFxgYaEpMTEweNGhQyZ3KWd6vaUGlwMBA08mTJ5UAEB8ff70Jwq28vLysly5d4kwmE5WcnCy9qY2fJTExMXnFihUZ7777rkdYWJgxJiamtHLEvXnz5gwASE5OllutVhw9elRe2aThlVde8Zo/f372kiVLPIGKdnTHjx9XAEBCQoKysrD/yJEjC5OTk89xHCccPHhQ8eGHH3qsXr0649ChQxduTQqVo/lVq1ZVeaZ4rTmBsqysjNZqtYyTk1OVqeClS5fmHjt27IKzszM/ePDgUlTT8uXLXV988cUCvV5P0zQNo9FY5ecqNDTUmJGRIQVuNGMAADs7O0GhUAgsy2LNmjWX4uPjLzAMg9jY2CrFegMCAow///zzxcqmCjExMbpr982MjIzUVU7ZOjs782fOnEnq1atX6ZYtW+w1Go3xpZdeupKQkJB8+vTp8/PmzcsHALlcLvA8T908+r6bEydOKG5tmtCsWTPLE088UXDzSDcrK4srKyujMzIyJLf+fQAVXxozMjLOvPvuu14AkJSUxLVs2bLRNWPQGq29T2QWN+l5TDc1h59mdNSrZJLnJQz9la3juUaUSZj31DJ2ycanO+nUTXyqeX9yvqrMYOlni3vXi0/+999/Vy9evNhLKpWKHTp00D722GM64EajAKBipCCTyUSWZcWtW7em6XQ6+sMPP/Q8deqUQq1W84GBgbf9Aqtu04KePXvqV61aJURGRmpUKhW/ZcuW9Dv1umVZFjNmzMiLjo7W0DSNL7/8MhOoGOlGR0drysvLmUWLFl3u1KmTYdu2bZbo6GgNRVHiqFGjimJjY8ucnZ2tffr0aVlYWMhu2LAh/e+//1bk5uZKly9fns3zPPXxxx+7zJw5s3D48OF+7du317i7u1veeeedK3v37lVt3rzZ+fDhw2qWZcU2bdoYBw8eXDx48OCWYWFh+ltbCN7NW2+9dWX8+PF+RqORfvvtt3OAiprGERERhs6dO+s7deqkYRhGjImJKav8O5g0aVLzffv2OezZs8c+NTU1f+7cuVVqH//666+qDh066ORyuTh+/PjCqKgoTUxMTJWKXx07djQsWLBACgBbtmyx/+yzz9wBwN/f3zhs2LCy9PR0yZgxY/xpmhbHjRtXGBAQUGWk+8EHH2Q/+eSTvmazmWJZVty1a1caADg6OvILFiy42rNnz5bDhw8vHT9+vG9RURHLMIz40ksv5avVan78+PG+3377rZsoiliyZEl25TVXrVqVRVEUXnnlldumgitt2LDBvqCgQDJy5MjrX0BMJhP1wQcfeO7atSvtwIED10d0Hh4e5rFjx/pmZGRwtzajACp+fgVBoAYOHFgCAHv27LF7kC82DQSt4Ji2/1wqsXUcNsOxNNZN66BTceyHDE2ttnU8t5JJmCVuaq7F6slR48esOKpoqvt5j10sBE1Rj9ni3tVueCBahPD7LYa6F9EiWCkJbfOGB7aSnJwsnTdvXrO4uLgmWR/57bffdu/bt29Zly5dGmXtusjISE3liuzqGDJkiN/WrVvTb57yBxp8wwNNQbnpePv39jXZ7UKLBoeZRrTz/l0lY4fAtlPK98KUGy0HVx68GPX5H6n1YuBlC0nv9DPLJUwzANVqolJTqv2BP2rCpCT1YiabsJHFixfbYhFJvdVIv3yFXMgrr9asS2PUyd8Zo9p76xVSdhLqb8IFAF4tk4x+NqZl0p/n81Rns5tm68WrpUajr4vSF3WcdEkmrCMajcbcSH/REgAeZJTbiDXPLNA3yVU6HEvj87EReoWUnQCgyNbxVMNlTkLP+GpcO11T3cN7qVgPAC3q+r73S7qCIAhN82+EIOrYtf/WGmyZKpOV98sq0jfGbVD3NaFjC4Fj6XgAu20dS3XRFPWjk1J6fmQ77/o8Kq81F/N1MtTDpHs2Pz/fniRegqhdgiBQ+fn59gDO2jqWh6U380E5pY3ykf09KaQMZvcKNNnJJS/bOpYHJKplkpnz+mkMEqbp/YrPKtJLdSbrbWWHa9s9n+lardZpV65c+e7KlSutQKaiCaI2CQDOWq3WabYO5BH45pQ0vaT7RLSPQFH4C0BDXCh6VMLQZwa09uwQ90+OrWOpU5eLDTBbheA7dXSqTfdMupGRkXkABtdRLARBNGBShvbMLW10W4/v66lOLXRqmeRjW8fxsOzkkk8mdvb9Lu6fnEZZC/xuskv0oGmq3k0vEwRBVAcjkzD2V5pY0g31tIOrijMB2G/rWB7B9lBPO9rbsVH0Gam27GIDZBL6jl3RahNJugRB1ARPrclqsjaxYgtDI5pZWIb6Hg14ARwAIy+Im/u38mhSf3nFegsoUCyAOu1/SJIuQRA1wSevzFgTfWEblN4hbgaOZXbZOo5HpeTY3T00buX3P7JxyS83GVDHK5hJ0iUIoiY0v1RsaFJLYJVSBs2dFHIACbaOpQb83c7HkWtqW3YvF+tFkKRLEEQD5JFTYpDaOoi6FOplB53Jmgbgrg1aGpArVkEo8XNpWhU8y4xWGkCdLiAjSZcgiJpA8YLYpMZJ3o4KAEixdRw1xcKLmc2b2GIqCy9QAO7Y8rW2NNli1wRB1CiBboT9ge+lmYNcVEjZRlP+k6WplGaO8ugHOcdw8QSK/lgBCAJU4X1g33FUldetpXko/PVT8Poy0DIVXAbNBWvnAgC4uultmHKSIfMOhdvIBdfPyd/xESz5mZAHRMExZiIAoOTwj5C6+UER2PFR32YVZmvdJ10y0iUIoiYINF2vi/zXOG9HuVHK0re1cWyo1DI2xdNeXu1V2KLAo2jvcriNWgSvaV9Dd+4AzAVVP47iv76HMqwXvKZ8CfsuY1Hy95rrr9lFD4fLoJeqHG/OqyhP7zXlS5gu/wfBpINVWwRz7oUaT7gAYOEFGkCdPhYhI12CIGqCwFBUk0q6EoYW0Tie5wIAKIoSB7T2oHxdFNU6/vzp49gQosHi5/oDAH42PwEgC6PGDbl+zHM/52PxnHFwcfeCKLbFmC7/w5fjIq69GoEziUewNe8IFlz7s0sXVVibuRuvPhGOV3ZxeHdMBL77aCEGvPcuAkJa1eTbBQC0aWbPoY63DJGkSxBETWhy08tCRS/yxjRbyPmqZZRbevXaMxZfvkI5CaBejO1G8TyPtprWoouTC3oYqeuj5dbNWtDvTBkNOcdRvCDAoNOi9ZVi3tnBCUNnjaXjTydSCrkcPYxUxT29grAsNZkZ06Elolu3F71SMgQPi0DaaAR2AAAgAElEQVTn/7lXdM/KxqCYfjX6xY7jKQpAna4eI0mXIIiaINJ00xrpCqJIoXElXSdaxkKmcaJFMy8KRqsoGKwQDVaR11spUW9lBKMVgtEK0cjDmFGKY/8kYt/LG+DOOaPL0hFUdPM2KNmaylRe0JSjhdIigWDm4evggQw2E7pfsxiGK8I036GI5ILw/fGfr5+TlJeGMJU//nzxBwxfP5N6Y8GbzKsxT2Pm5sVUiGsAsv5Iwri2sTX2hh2GtjRK3BT5NXbBaiBJlyCImtDkWpFdLTNxvCB6NJZ+tBaT0V5fVoZzB/6gBFGgIIqQyhWQq+3AKZWQeajAyRW8lFOLEikHU56UkiplTPs3+osUQ1OtDrWFngK8FneGaOEhmngx55diavuWrXwzR3exNL+E8uoZxDTrGwLBZEWfiP5IjcsFdYqGxEsJwchDquNgtJohiAIKdEUYqOmBr+LXwdveA8uHLsKI9TMxLOxxyCU100GSYikBgLlGLlZNJOkSBFEThKY20r1UrKfLjZYwB0Xj2J4sCEwATcuF1r0GihKOpSUcR1ktZuhLSoSygjyhODebKrmay5QX5KO8sACn449BV1aKt0b1o+zlMiQc+ht+rs5Y9fLTUDm5QOngSDmpOHyx8gNm0tgxmP+/DyGKIi455oqent68VCqDJMWepvYytPOUMIFiacqdaUdtnXsUsXtmgXbkMGbxdPGjZUvhLnGG85Qwnv6NY2R9vQU7Rlll1H39f003/l0033+WnJIwAoA6LRhOki5BEDVBaCQDvmq7XGQAgCBbx1FjRHju+jqJLszWVvx/CnB0V8DVR007ednTDu6ecA+QCJyCEiQymhZ37qAzKQ4/J1+ExWJBh06dBDuFnDqQeYXy0ZoQbDLCRy7FN+s24Ku16+GiUsBOxmHb+wspuVTCfvXnEeSWlsNkscLFzYEeHdUGGg9X+AAIlDNw8vLFnFnTqDlPjsOCb76Ha4Q3Gx4aCqNjCQOFFZxceX3UzbASimFZimYYimJoimJpgKIgWoXro27BeG3K3GgVRZ2VEgwWWtJMJQNJugRBNEAC28RGuuevlEEhZVsCYABUa/FRPWYnkbHejIQCRQGiCEAEiq/oUXxFf/Nx9LV/sPzXL3D+8gl4ufpg7Se/Yetfq2lOKcErL73Mcwpa1Bu19MYzT9PBoaEwm82YOHas8OlXX9K9n5wipl44T6lPJqHUZIYginj+sU5wUiqw/ugpXCktR4inG9QsDWcG+Hb1GogGHRYNegxf/RmPze8vhpRlgIrP/a4YiQQqR6fro26FgwOlUNtDpraDXK0CZ6eCl8KZAkm6BEE0QMXOKq6hJ54HUqy3oEBrsng5yMMBnLR1PI+os1VvFAa/EE4zEobKyyzjs84V0TkpJVReRjl46+3bd3u1HoXLBakwGkz4569MbNm6GZN6vYFNS04xAPDbyfXQ6qWY2edzyFwETFs4gJ409ml4h/bGyx99jsXvfyzIFBy17NNPqXlrN+Ls6X/Eg7ql+L+5LwrjZjzHXL1ShKcH9sGXm7fBUVmxjUmECF4QcJ98CwDgLRaU5l1Fad7Vux4zedm3Wk6hyH7Iz+yhkKRLEERNSPdxUjSmlbzV8uf5POnYaJ++DE016KQrGAz9S1d/j4KvvqKkAf6wHzKUad25K9+2RyglUXJ0yRWdkJVUjOwLxfSVtFKY9FYEeUdgUNQUbDy4DO9umoyOmv7wdPLFzsTV8HHVABSFvNLLeOP7CeBFHhQYaCT98M2COOpKRhlenvkWfbXkEkwWA/xaaPDkkOeo0iIrJPIwpqjcgJdmv8if/O80/X8/b6U+/fxzrD5zBrGjnhAnzJgmFF/JRenVXKa8sOL5cllBPvQlJRDFB+uwKFerWQBk9TJBEA1Ouquaa1qFewHsPpPLDQ73mmYnl7wPNNiKXJTI8yO1f//NAIA57SLyP/kE+OQTBgBoJyc4DBlCB/boIQQ/EcxL7VWMvtQkXjpfLNgdtjL7/9uM10Z8d/1ig6ImAwCCvNri/OXjMFkNMJr1mNL7LXCsDHkllyGXKsHSEpiURmiatcOQDtNA0wz+LT+DXgO7ITpwIPITVcyF//JxQmFEjNPTGDpWBbcWaqo4T8HYuzdHs2AJzykgSjiGlnASmmElMOq00BYVCqX5V4XinBymrOAqVV5QgPLCfJQX5MOo01Z545xCIQNw96FwLSBJlyCImqDlBdHoquJU+dpGU6TpvuIvFsJkFdwBtANwwtbxPKQefHGx2njmzB1fFIqKULR6NYpWr66YyeA42PXrS3n3fpyhYpx5hzUKZurSrshJLeUvJRUxOSklKMrVIelyIrydW2LWoI9RUJaDL3fNQ4Bnawgij7QrZ/HqiG/gqHLHqn3v4OiF39A5eABGdnn++n2/+fUNPNH9Rew5uR7ZhWkI9o5El5CBN4dWZY6ZldFw87GDi7eKdvTwoT0Cg0T/SNrKyWmK5RhawnEUBUBfViqWFxYI5YX5EHjBSjMw1PQHei8k6RIEUSOMFj6nuZM8qCklXVEE1h7NlD3d3f8FFcdOsnU8D4MvK5tX9MOa6ldlMplQFrcdZXHbkW0xM6bLl3H1hWfgOHAg49ExmmeH+NEUy1D9Hv+fOGrMVLhLHSg2k4Gz2gNXSy7BQekKb+eWcLHzAgCE+3ZB+tVzQPCNW/ybcRg+rhqYrUbkFqVj6uNvY1ncHES17AXpXfboWo0Cci6UIOdCSeUfUbglxynspXBrYUc5N1MyngEhEEW64EE+q5rQ5J7BEARRO0Qg7Vq7uyblx2NZDENRowH42DqWhxACmo4p2br1kTZ86ePjceXNN5HZvw+T1i6cyhoxDD6UhbqUfRADZoQKg1/ToIy/in5jOwndenaG0aJFuaEiOSZnn4KH440+8jxvxf4zW9E7fDTMVhNwrbyoCBFWwfooYUJfakbGvwU48Wsmzh3KhdUinHukCz4EMtIlCKJGKKXsfz5Oin6oGGE0GflaE1YfSWef7NjiQ7VM8oSt43kQvFb7btGqVRLR8OAzrHNzspGg16OE59EzLRUznV1gufZY+4mUFEyyWvB6YiLV5vvvKbAM3hw8GB1aWCjmsWB+ZdBXzNyXXobJaIWfZzD6dh4OY1nFIqi//4tDdFAfSCUyNHPyB0QR7/08DWHNo6Hgaq5Msp2rDKyEOV9jF6wmShQb6rN/giDqmRlbT17+5KVNp5vccFfNsYh/rZdBJWO7ADhl63iq6TFrUdGOtMf7KASdrm7vLJPBrm9fqPv0ESWhbQSJsyNjMfHITS3hs84VMblppSjK1dXq0rTek0P1mg4eLwH4tvbucjsy0iUIoqak+7soLbYOwhbKTVa8vf2sbPGQVltVHBuCOi648BCUvE639eq779V9wgUAoxFlcXEoi4ujcG1BlLJrFzgNGMh4dIni2WH+NMUwVF5GKZ/5XxGdm1ZK5WWWQbDWXBZ28VZZAJDpZYIgGqyMZo6K+1ctaKS2nsym+rXycOsS4PKxkmOfv/8ZtqM1a1cIglHl/uYbsBYXQx8fb+uQoDt0GLpDh4FrSZgLDob94FgmvFMXvl3vUEqi4OiiXK1w6dy1/cIXy2A2POQzXgqwd5PLAfxXY2+gurcm08sEQdQQzsIL5a0X/iYxWh6sSEFjYS+X4I+XY/ROCukUmqZ+snU8d2LiTS8VGAreGR43XDE2eCxeCHkaZdu381ff/4ARjfV3gM64usBh6FAouseITIBGkNopGV2xUbh0vli8nFzM5KaWQldSvZXzdi5yjHkzqkgqY51rOezbkKRLEESNKdGbz81YeyLkWHqRrUOxmRcea4kXY5oJNKfqA+APW8dzM0EUnig1la4as3OMPFeXCwBopmyGNY+t4B2NLJ09axZlPHv3wV/vtFQoaQY0BbCg8LOvb5XXS3keb17JxSWzBRxN4V0PTwRy3PXXeVHEqMwMuLMslns3BwC8kpODFJMJMSoVXnR1BQAsLyhAEMehl1p99zcjk8F+wACoevcSJaHhgsTJgTEbrWJuaqlwKamIyUktqagbfYcUF9DOFT3GBR+QqSQ9qvfJ1RwyvUwQRI3hWGZvlK+T5lh6UZPcjvh0dz+81KM5qMPLaHSZEwepsj+Ag7aO65pYvUW/aupvU68nXADI1mWj946BzPz288Sxa9eiaNUqoWD5NzSsd566/aF5cziyd04dKwoLEczJ8EUzb1w0mfBO3lWsbn5jJ9Xa4mIESDlohYoy3cnXRtbb/PwwISsT5TwPoyjijNGAZ11c7v1ujEaUbt2K0ortThXPhbt1o5wHDGA8u0ZZ2WH+DMXQ1NX0Mj7rvyI6J7WEys8qh8CL8AxwsEgV7N5qf3I1iCRdgiBqjFzK/NUtyGXSl3+l2tk6lrq2KDZUfKq9C0WtGw5kxQOXEpUYs+43sNxsMJKVtozNKlhnGKyGZTP2zpCnlKTc8ZgPjn9I7bi4EysnfCn69ekjZM+aRZvTMx7oPmlmE6Y7VczY+nMcciwWFFitcGFZXLFYcECnxQwnZ6wprpgJYSkKJlGAIIqwiCJoisIX+XmY6eL6UO9Td/AgdAcPAtdyGxcSDPshQ5m2HTvx7R4Pq3gunKMVlI4yiqapww91k0fUJL+NEgRRa4608XaQNbXeut9NaCs8FWFHUd8/XpFwAeDiX8CKGDnKr3wKk3YVAFt0u1foLfrv8/R5nzyx8wn5mYI7l3qsdK7oHLpse4w5wKRSvlu2wHH8+CqTsxRFYdrlSxiZkY5NJSW3na/hZNinLQcA/GswIMdiwdVrI+b38/Iw19UNN/9sBHAcPCUSjMjMQD+1GllmM0QAobI7V516UKak88h7/31kDh3CpEe3oy/26gnsWEvLlSwNG5XtJEmXIIialGflhYIQz6Yx0KVpYOdz0Xwvb4Givu0O5CVVPaAwFVjeWYFLR8fAVH4SQFQdhhejt+hTjuQceWLE9hGKrPKsap849+A86rnDL8J+zvOiz/p1AuvmBgBY7+ODLb5++Na7OX4sKcZxfZVeu5ju5IRSXsCwjHSsLylGiEwGBsB+rRZOLIOwOyTT19zc8YuvHyY7OeOLgny84OKCbwoL8GJONn6+Q2J/FNareTCcOAmhrOw/AOU1evFqIkmXIIgaRVPUr10DXRr9Ck0ZS+PQS535MOkVivq2G4XSy3c+0FQGrB+pwJ5XQ2EoOQBT+c8AvGsxtOY6i25dialk9/yD871e3P+iQmvR3v+sWxzJOYKu23rT59wtov+vu2E3cKDoxkoAAM4si14qFf41Vq1kpWIYLPH0xC++fnjfwxNFViu8JRKcNOjxl1aL3mmpeDknB8f0eszLyaly7h/l5QiTyaEXRKSaTFjm1Qzby0phEGp2JbyiYwcrpVDsrNGLPgCSdAmCqFFKjt3eL8zDJqOIuuKikiJhXmfBs+w0qO8fp2EovvcJogicWkdhWZgcid8NgUV/AWbdNwAiUXNlM0NhKt9sFawZm5I3jRqwZYBi/6X9j3RBs2DGlD+mM2+efBfK+XNh99GHPG1vD70g4IhOX2VlMgCU8TzM13bEbC4tRXuFAiqGwUuubvgroCX2BbTEx15e6KBQ4EMvr+vnWUQR60qKMcXJCUZBAIXKessVr9UkVUyMnpZKbbaqnCykIgiipv0V5mUvk0sYGCy8rWOpcf6uSvz6bKQgTdktUnHPMhAe4D2atcC+hRIkrJQgctJURE6aAFZWAKniK9DsDgDJqH7xQwZAB1iMQ8CbRkEUPXBsucTc+QVq/6X90nJLzX3v2XlxJ34/8TulW6llpJQExvw89Fcp0U2pwsaSii8cTzg44qLZjFdzc8BQFAKkUrzj4Vmt6/9YXIwhdvaQ0zQ0HAcRIoakp6O7Sgk7pubqrTAuLpD6+koA2GQRFUD26RIEUQtK9OaTc376J2J/cr6tQ6lRUb6O2Dg5XKQTvhWoPxY9ejagKMCnM9DuSQMCegvg1IBFnwRWlgipIh2A4do/RgBSCLwfTGWhEMVASOQBKL8i4Nw2OZJ/ZXE5ERAFiEO/EU96BAqTfptcK9XBJoVNwpzQZ1G2axd/dcn/mIdplmArjuPGia5zZv/C2NmNsFUMJOkSBFHjeEF87afErAWv/3KWu//RDcPA1h74YlQIqH0LBCphRe08mlN7AO6tALcQQO1pgVTJQ6LgIZELEKwUClIUKM6gUZIFFKYA5Vduv4ZUBcMrKZi0ZzLOFdVOaWFPhSfWPLaCd7HI6MuzZlPGf/+tlfvUNN/NP5fJW7WaAGCHrWIgSZcgiNrgqzNZk9q9s1dmsjb8kpBTu/jhzb6+wLZnRerctnq/IUoYs048onYQnv3juVqthf1y+5fxpP8YFK/5PyH/yy/vWlCjPmDdXBGwd6+e5jgnANWrF1kLyEIqgiBqQ4ZVEE/3DfOwdRyP7M0BIeKbfXxAbRiNhpBwAYDeMZtq79Ge8bf3r9X7fHz8Y4zeOwGSJ4bCL26bKPWv3fs9CnWfPqJoNu+GDRMuQJIuQRC1xF4uWTaxs2+DXsW8fFxbYWp7B4pa1RfIqC/VHKtBXwg244jwfNvna30l24XiC+ga15v+QzwH3y2b4ThxogCq/n03cRg+opxRq1fbOg6SdAmCqC3bwrzsqOZOclvH8VC2zYji+7UARa2IAa6etXU4D4zd/gLd3bs7462qzS3BN7x6+HXq6b9nwm7mDLTYsF5gPerPLAfr7g6pvx8LYJ+tYyFJlyCI2mLiBXHdmPY+9fdB3x1IWRqHXurEhysKKopelFS/klO9UpYNJvuUMKPNjDrbt5VwJQHdtvWm/3XRw3/XTtgNjq0Xi4bsBgwQRItlOwCzrWMhC6kIgqhNrYv15qOR7+xVCA3gV42DgsX+2R0E+6IzoDaMpmHR3/+k+sy5JYzPHsbAXwYiT59Xp7fu59sP70W+LZoSj4u5r71O8zVc0rHaKAotD+zXSdzcHgcQb5sgbiAjXYIgatMZmqIyuwc9XNeYutTCWYGjL3cW7LP+EKm1Qxt+wgWAwlTQeUn81FZT67xKyZ6MPYiJ60Nlh7iI/r/tgbJbt7oOAQCg6t4dtFyeDeCoTQK4BUm6BEHUKnu55NOnOvnqbB3HvbRr7oA/XogSudM/iNSWKQyEBjUjfk/SHbOZ4YHDGAfOoc7vrbVqMeq3cczS5OXw+mwZPN97j6cUijqNwXnaVC2jVi9B9St91SqSdAmCqG0bOwc4My4qW3S2u7++oe7YPD0CzIElIvX7m7W6r9Umck8DxRn8xLCJNtswvT5pPfrtGgJdTAQC9vwqytu2rZP7SgMCIGvVWgDwU53csBqYhQsX2joGgiAaN5PBwocJghhy9GJRvfqiP7FzCywdGgRq+0yROrmmXsVWk9i8JDr4sXeojec3wiyYUfBbAbJXZaPozyLoL+ihDleDYm5s8ynYc+31v4pQllgGZbASjKLi+8jZyWdRdqIMRX8Vofzfcjh0rBhBX/rmEvK25cFSZIEqVAUAyIvLA6/nwXly0Fl0WJfyIy23c6JiprwJWqUS9MePU6jhLkI3c3/tVQMX2HIpxbI2a3Bwq0b7Q0YQRP2hlknemt7d3+qkrD+j3Vf7abCwry+ojWNBnd1S/zaW1qTMw6DKc/lxIeMES7EFhXsLEbAwAIHvBUIURJQeK61yuKyFDAELAhD4biDsouxwZdONcpO0lEbLd1qi5Tst0WJOCwCA8ZIRABD4biD0F/Tg9TwsJRYYLhpg165qb+VPT36KUb+PBzNqkOgft02UtmxZK2+ZdXeHuk8f0Bz3Ra3c4CGRpEsQRF1IE0Wsnd0r0KbVgCp9MaaNOKODC6gfBgAX99s6nDoh3/M6MzlsMi1jZBAFEYJZgMiLEM0iWMeqDedUISrQXEV6UAQoYC26zzNuBhAtIkRBhGgVARrI25oHt+Fudzw8pSQF3eMeZ37jz8D3501wmjK5xgtqOE2ZbIYgrAZQVKMXfkQk6RIEUSeUHPvW6KjmvLejbYtlbH66vTAogAW1okfF886m4sIe0IZifmyHsaJLPxdcePkCzs85D1pOQ91KfdfTiv8uhqqN6vr/FywCUhemIm1xGspOlAEAZF4ySJwkSFuQBvsoe5ivVmyHlbe499/1G0fepKYfeB7qZ6ehxU8bBfamHruPgvXwgOPo0TytUPyvRi5Yg0jSJQiirlylgE/n9wu2SS84KUtj/5yOfKS6BNS3XSkUp9siDJtS7FvMjGk+htKe0iLooyAELwuGYBJQcuTOe2hLjpTAkG6AS3+X63+m+ViDlgtbovkzzZG7IRemvIrJC8/xnmj5Tku49HepGOUOc0Pe9jxkfZWFov13H2wmXk1El2296FMOpfDfuR32Q4c+8ipjt3mvGAB8AeDyo16rppGkSxBEnZFJmA96h7hbgj3uPrKqDXYyFkfnduJbmC5Q1MrHaOgK6vT+9caZTTi493ehpX9LsHYsKJaCXXs76FNv35Os/U+L/B35aDGnBWjJjVQhcZQAAKRuUiiDlTBmGqucV3ayDHI/OQSTAFO2CT7P+6DkSAkE090XTFkFK5756zn6lYS34PLma/Be8S3PODo+1FuUtQqDumdPMy2Xv/tQF6hlJOkSBFGXyliGWvjmoNA627fb3FGOY690EhxzDoD6v1gaZm1d3bpeCirYTxvSDKDMFERRhO6cDpxn1bbHhkwDsn/Ihs9sH7B2N5738joegqUieVrLrdCn6sF53ThXtIoo3FsIl/4uEMwCUPmYVqx47X72Zu5Ft7jHqaxAewT8tgeqHj0e+P15vL1AS0ml8wDUy2YbpAwkQRB1jdOZrFmTVie4JWYU1+qNwr3tsWVaW5E5vV6g9sxnQH7fAQBeM00XV2xYTel4HWQ+MjSb0gz5O/Ih95PDLsIO6R+mw3jZCIl9xahW4ixBizktoE/RI3tNNiiqImE7P+4Mpxin69ct+K0AjJKBY1dHiKKIy99chjHbCHUbNTxGP1gDhLHBYzGv9WxR+/s+4eo7ixlBd/8KYereveH5/v8uMipVEIA6r8JVHSTpEgRhC0+eyyn9esDnh1T3P/ThPBbshu/GhoL6+0OBOvwpmdW7WdeXkdNxqthvS39KrB+Fmu7IVe6KNY+t5D1EOzp79hzKcPLk3Q+WSNBy316dxN19OIDf6yzIB0R+EAmCsIUNPk7KqwNbe9bKxcdF++D7cWGgdr0kkoR7B4c+hj2rEHv69LR1JPeUb8jHgF1DmdW5v1DNv/8ObvPn85REcsdjncaNE2iF4gTqccIFyEiXIAjb6VhutPzZ6+MD8rzymtu++3KfIMzs6gVq05NAar0pRFT/9FqIi60HC0PihjaILyX+9v5YHfMNrywz09kvvECZLqRcf4319ETArp0GWqGIBJBkuyjvr0F82ARBNEpHWZpe9sXYCH1N1UX4ZFRrcWZnN1A/DKrVhHupVEDPNTqEfKVF2NdafHa04ktDkUHE42t1CPxCi8fX6lBsuPugpswkotkn5Zi5+/YdVIN/1KPV1zcWfM3fa0Sb5Vo89cuNY9eeNl+/70P5czHc5a5UJ69OD3+NOnSx9CJitvdhdhmPo8VPP8Fp+jQBdEUKa7b0Ix0YZgnqecIFSNIlCMKG5FJmYZiXffrETr6PXIB349RIflgQV1H0Iucez/5qAEsDH/eRIel5FY5OVeKrRAvO5fN4/5AJvfxYpLygQi8/Fu8funtSfOtPE2Ja3N5fYWuSBTf3hig1ijhymce/z6rAiyLOXOVhsIj44bQFz0U9QllNUYDy303UnIjZNmuE8DAWHF1ITf5rBlTTJ8F300+Cy8yZIqfRZNIc976tY6sOknQJgrAli0rGDpvfL9jY0u3h1lSxNPDH7I58ByctTX3blULRxRoO8XaeahrtPCsSppqjEOJKI7tMRFyyFRPDK545TgyXYFvyncsnnsjhcVUnoE9A1fKLWrOIT+LNeLP7jW04NAWYeRGiKMJgASQM8NERM2ZFSyFhHnGKYM9r8LVrQUW4RTzaderYqfxT6LKtF/2fox5OU6eIjEo1CkCD6MdIki5BELaWImGoF799MlL3oElELWNx9JUugr/1IkWt6ElBm1dLId5dRomAU7k8OngzuKoV4Kmu+LXqqaaRp7t9ECmIIl7+3YiPHpfd9tpbf5rwcicpFJIbn4OaozAiRIKIb3Xwc6Bhz1FIzOExJPjOC4oeCG+GPGknZkfMqpfba+6FoijYKxwNvJR5GcA5W8dTXSTpEgRhcyxDr3RTc/Gv9NWYq3tOMwcZjs3tKDhfPSxSPwygYSqrzRDvSGsWMWKTHp/2k8GOq94Xhq8TLRgQyKK5fdVfv/9c4ZFaLGBYyO3JdF4XDv88o8LHfWV46y8TFvfg8N1JM0b/rMe7fz/aIjRq18tUqHMoE+oU+kjXqWuzI2abPZWeCVJG+pmtY3kQJOkSBFEfiGqZZNyEji30Hfyc7ntwqKca+2d3EOVJm0Xqp/EM+Grn6hpj4SsS7vjWEgy/lijdVTRyyytGt7nlAtyUt/+Kjb9sxZcJZvh+Wo65v5vwf6cteHWfEfGXeJzI4eH7aTm6rtLhQqGAHj9ULdx1KrdiQBrkTOP/TluwaZQCZ/N4pBQ+wkDVrAWX+of4QsTMBjPa7ejZEWOCx+hUUtUYoB5vNL4D9v6HEARB1Il8hZQd//X4dj/3XLpfUWa88yO6HhpXrBoXBurwMpH6+6PbVyLVAVEUMXW7ESEuDF7qdOP56+AgFmtOW/BqVw5rTlswRHP7r9j1wxXX//2Hf8w4nsPj/d4VU83PXlsYlVEiYNAGPfZPUlY5962/TFgRK4NFAPhrqYamAL3l0d4PvWM21f6l/xh/e39cLK39Z+KPwknmhI9iPtLLWfloAPm2judBkZEuQRD1yW5OwqxfPiFSz9C3T9eOae+N1eNbgfr1FZH6+yOb/f46fInH2n8t+DPdirbfaNH2Gy12pw4kX1UAAAjESURBVFjwalcp9l60IvALLfZetOLVrhUJ+XgOj2nbH6250rbzFkR5MfBS03CQUejkzaD1ci0oCgj3eMTvHvpCsBnxwnNtn6vXo105K8fKPit1Ulr6GYB9to7nYZDiGARB1DeScqPlz53/5ka9tvXM9WHk7F4tMSfGG9TPk4CUel10qGGy84Zx1gkMixuGy9p61xEPLMViRZ8V+lDn0G1KiXICGti0ciUy0iUIor6xqGWSQbHhXpef7u5vBYAPh7cS53TzBPV/g0nCrS1ll8Hk/CPMaPN0vRvtUqCwpNsSQ4hTyFGlRDkRDTThAmSkSxBE/dVcb7b+k5lf5hBsz9PUD/2BgpT7n0U8POdAGJ89hIG/DESevu63X93N3PZzzSODRiYpJcrOAO7fbqgeIyNdgiDqq0sKKds72F0lUnvmk4RbFwpTQOed56e2mlpvRrtPhT7FjwwamauUKHuhgSdcgCRdgiDqt1MUK+2D2M/18Gxr61iaBOmOWczwwGGMA+dg61AwwG+AODNiZolSouwOoNDW8dQEknQJgqjv/gSnGo+J2w1wDbZ1LI1f7mmgOIOfGDbRpjWZO3l2wsLOC7VyVt4DQJYtY6lJJOkSBNEQbINUOR1TfzeQEW/tk+2ay4wLHkerJA9XD/tRhTqF4tOen+rlrHwggLM2CaKWkKRLEETDQLPrIbMfi0m79PBpGO3oGqyMQ6DKr/Bjg8fW+Wi3jUsbfNf3O4NCohgP4GBd37+2kaRLEERDEgdONQQTtujQspetY2nU5L+9zkxpNZmWMbc3ZqgtMd4xWNlnpU4tVY8EsK3OblyHSNIlCKKh2Qepsg9GrytH6BCy57G2JP8K2lDCjwwaWSef8cjAkfxHMR+VKiSKxwDsrot72gJJugRBNERHIFV0x9BvyhA9vUE1YW9IFPsWMzPazKAkdA20EbyH59s+b3kl6pU8OStvDyChVm9mYyTpEgTRUP0DqSISvRZmYtg3RrDc/c8gHsyZTZBYjfzggMG1MtqVMTIs67HM8GTokxcUEkUEgNTauE99QpIuQRANWRo4VRsED/oD0//UQe1p63gaHeWBj5jn2z4PhqrZhk7uCndsHLRR18mr0y6lRBkF4GqN3qCeIkmXIIiGTgtOHQungA/wXLwBzTvYOp7GJfE7KECJfX371tgl27q2xdbBWw3eKu93lBLlaACP1oKpASFJlyCIxkCERP4O5I4j8OQ2LaKmkee8NUgZv5yeFfGCSOH2dosPgqZoTAiZIKzos0Jrx9mN4FjuAzTg5gUPgyRdgiAak18hVbRD74UpmLhDT6aba8jBpXCQKMWezXs+9CX87P2wceBG3fNtn/9HzsrbAfi15gJsOEjSJQiisUkBp24N7/bLMDPRgDZjmtRIqrYoj6+hZ7eb9cAzCCzFYkabGdafBv2kC3QMnK+SqqIANNnuFaS1H0EQjVk7mMq3IPOIG+KeU0BXYOt4Gi6Khu7VTPHFv+dS8Tnx1TolxCkEH3b/UOcidzmhkqqeRCOqofywSNIlCKKx42DW/Q+C9RnsekmOM5ttHU/DNfBjnPPrJIzZ9cQ9Z0mltBQzI2aanwh+wsQx3EyaoteiiT27vRuSdAmCaCo6wlj2A4ozmmHnHBWyT9g6noaHkUI/P1185o/nqFN5p+54SIRbBD7o/oFeLVH/pZKqpqKJbAWqLpJ0CYJoSmiIwlOwGD9Byu8y/Pa6HGXZto6pQRGHfSuecA8QJv82pcrGXWeZM2ZGzDQN9B9okLPyqQC22ijEeo0kXYIgmiIVLIbXAXEO4r9mcegTCcw6W8fUMEhVMLySgol7JiGpKAmOnCOmtZ5mHq0ZLQiisEYhUbwGoNjWYdZXJOkSBNGUNYex7FNQVH/EfynBsW9ZGEi+uB9+zHrxuIOreK7wnPWJ4CcEQRTWKyXKhQAu2zq2+o4kXYIgCEADU9lboNkROLWOxuHPpCgl+eOO7JoBPV4z61uPYEWIa5US5dsgq5KrjSRdgiCIG5rBrH8FFDUdybsp/L1Ujrxzto6pfvBuD3R41oDggYAofA+p8kMAl2wdVkNDki5BEMTtHGA1zYTAv4ziDAaJ36nw31aqyU09K12B8LECoqfrIXcsB8t9CUa6HOSZ7UMjSZcgCOLuWACPw1DyHFhZb1z8i8eJ1Uqk7gME3tax1Q6aAVo+DkRN18KvKwurOQ4yu68AHALZa/vISNIlCIKoHkcAo2EomQmKCsB/v1C4sEeG9L/R4Fc+yxyAlr2AkMF6tOxNQ7CmQe7wKYBNAMpsHV5jQpIuQRDEgwuCwA+BsXQ0pMo2uPKvEee2q5G6l0Jekq1jqx5XDRDUX0DYUC3cQjlY9Echd/wRwG6QZ7W1hiRdgiCIR6MC0BMm7VAAseDNSmTF88g8rEb2CSD3X8Cit22EShfAqx3QLFKAb1ctPMOlAKUFEAdOtRXAX2hCPW1tiSRdgiCImkMBaAmgI0zabuAt3cGp/KHLNyL3NJB9QoXiDApl2UDpZaAsBxCsNXNnRgLYNwccfACHFoBjCwHurXRo1o4GZ0fDrDsLqfJPsNxRAIkASCkuGyBJlyAIonZJAAQBCIfVFAGzLgSALxipJyRye5jKTSi/YkHpJQqmchoWHQOzgYFFx8CiZ2A1AVYTIJEDUpUAmZ0VnJ0VnB0PTi1A7gDYNZNAZs/B8v/t3bFKggEUhuEvNew3aZfo/m+m2cW22kMRBzFrcGsIgt/P5XngXMPLWc45fObr+J676Sbzp3Umk02S1yRvSf79lo/xiS7A7cySrJK8JHlO8phkSPKQZMj3ecjpuMz5tMj0fpfZfJtk/2t2uRyn+Egy0trMtYguAJT8+RMRABiP6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQInoAkCJ6AJAiegCQMkPLbMm/rRer4UAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pad sequences (samples x time)\n",
      "x_train shape: (2874, 400)\n",
      "x_test shape: (1416, 400)\n",
      "Build model...\n",
      "Train on 2586 samples, validate on 288 samples\n",
      "Epoch 1/5\n",
      "2586/2586 [==============================] - 24s 9ms/sample - loss: 0.4483 - accuracy: 0.8531 - val_loss: 0.4096 - val_accuracy: 0.8576\n",
      "Epoch 2/5\n",
      "2586/2586 [==============================] - 22s 9ms/sample - loss: 0.4058 - accuracy: 0.8561 - val_loss: 0.4192 - val_accuracy: 0.8576\n",
      "Epoch 3/5\n",
      "2586/2586 [==============================] - 22s 8ms/sample - loss: 0.2882 - accuracy: 0.8948 - val_loss: 0.2273 - val_accuracy: 0.9236\n",
      "Epoch 4/5\n",
      "2586/2586 [==============================] - 23s 9ms/sample - loss: 0.1716 - accuracy: 0.9490 - val_loss: 0.2336 - val_accuracy: 0.9167\n",
      "Epoch 5/5\n",
      "2586/2586 [==============================] - 22s 8ms/sample - loss: 0.1186 - accuracy: 0.9640 - val_loss: 0.1900 - val_accuracy: 0.9306\n",
      "1416/1416 - 4s - loss: 0.1975 - accuracy: 0.9237\n",
      "test loss: 0.19753522800523682, test acc: 0.9237288236618042\n",
      "[[1142   71]\n",
      " [  37  166]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pad sequences (samples x time)\n",
      "x_train shape: (2874, 400)\n",
      "x_test shape: (1416, 400)\n",
      "Build model...\n",
      "Train on 2586 samples, validate on 288 samples\n",
      "Epoch 1/10\n",
      "2586/2586 [==============================] - 104s 40ms/sample - loss: 1.2290 - accuracy: 0.6145 - val_loss: 1.1404 - val_accuracy: 0.6181\n",
      "Epoch 2/10\n",
      "2586/2586 [==============================] - 106s 41ms/sample - loss: 0.9951 - accuracy: 0.6323 - val_loss: 0.7462 - val_accuracy: 0.8194\n",
      "Epoch 3/10\n",
      "2586/2586 [==============================] - 108s 42ms/sample - loss: 0.5626 - accuracy: 0.8561 - val_loss: 0.4532 - val_accuracy: 0.8889\n",
      "Epoch 4/10\n",
      "2586/2586 [==============================] - 110s 42ms/sample - loss: 0.4085 - accuracy: 0.9045 - val_loss: 0.4153 - val_accuracy: 0.8889\n",
      "Epoch 5/10\n",
      "2586/2586 [==============================] - 110s 43ms/sample - loss: 0.3263 - accuracy: 0.9172 - val_loss: 0.4220 - val_accuracy: 0.8819\n",
      "Epoch 6/10\n",
      "2586/2586 [==============================] - 110s 42ms/sample - loss: 0.2639 - accuracy: 0.9312 - val_loss: 0.3875 - val_accuracy: 0.8924\n",
      "Epoch 7/10\n",
      "2586/2586 [==============================] - 109s 42ms/sample - loss: 0.2154 - accuracy: 0.9428 - val_loss: 0.4007 - val_accuracy: 0.8889\n",
      "Epoch 8/10\n",
      "2586/2586 [==============================] - 112s 43ms/sample - loss: 0.2233 - accuracy: 0.9404 - val_loss: 0.3754 - val_accuracy: 0.8958\n",
      "Epoch 9/10\n",
      "2586/2586 [==============================] - 111s 43ms/sample - loss: 0.1627 - accuracy: 0.9578 - val_loss: 0.3836 - val_accuracy: 0.9028\n",
      "Epoch 10/10\n",
      "2586/2586 [==============================] - 113s 44ms/sample - loss: 0.1575 - accuracy: 0.9586 - val_loss: 0.3572 - val_accuracy: 0.9132\n",
      "1416/1416 - 8s - loss: 0.3083 - accuracy: 0.9181\n",
      "test loss: 0.3082659235066276, test acc: 0.9180790781974792\n",
      "[[845   3  14  11   0   0   0]\n",
      " [  3 108   0   4   0   0   2]\n",
      " [ 14   3 266   7   0   0   0]\n",
      " [  5   5   2  71   0   0   1]\n",
      " [  8   0   0   3   2   0   0]\n",
      " [  0   8   0   0   0   5   0]\n",
      " [ 13  10   0   0   0   0   3]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import nltk\n",
    "import re\n",
    "import random\n",
    "import tensorflow as tf\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.feature_extraction.text import CountVectorizer,TfidfVectorizer\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.feature_selection import SelectFromModel\n",
    "from data_print import data_class_display,PCA_text_data_display,pp_conf_matrix\n",
    "from EstimatorSelectionHelper import GridSearchCV_Classifiers, Best_Classifiers\n",
    "from ukrainian_stemmer import UkrainianStemmer\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "from tensorflow.keras.preprocessing import sequence\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Dropout, Activation\n",
    "from tensorflow.keras.layers import Embedding\n",
    "from tensorflow.keras.layers import Conv1D, GlobalMaxPooling1D,MaxPooling1D\n",
    "from tensorflow.keras.datasets import imdb\n",
    "    \n",
    "def ua_tokenizer_sklearn(text,stemmer=True):\n",
    "    \"\"\" Tokenizer for Ukrainian language, returns only alphabetic tokens. \n",
    "    \n",
    "    Keyword arguments:\n",
    "    text -- text for tokenize \n",
    "    stemmer -- if True use UkrainianStemmer for stemming words (default True)\n",
    "          \n",
    "    \"\"\"\n",
    "    text=re.sub(r\"\"\"['’\"`�]\"\"\", '', text)\n",
    "    text=re.sub(r\"\"\"([0-9])([\\u0400-\\u04FF]|[A-z])\"\"\", r\"\\1 \\2\", text)\n",
    "    text=re.sub(r\"\"\"([\\u0400-\\u04FF]|[A-z])([0-9])\"\"\", r\"\\1 \\2\", text)\n",
    "    text=re.sub(r\"\"\"[\\-.,:+*/_]\"\"\", ' ', text)\n",
    "    if stemmer==True:\n",
    "        return [UkrainianStemmer(word).stem_word() for word in nltk.word_tokenize(text) if word.isalpha()]\n",
    "    else:\n",
    "        return [word for word in nltk.word_tokenize(text) if word.isalpha()]\n",
    "    \n",
    "    \n",
    "    \n",
    "land_data=pd.read_csv('land_real_estate.csv').replace(np.nan, \"\")\n",
    "land_data['text']=land_data['description']+' '+land_data['land_types_source']+' '+land_data['land_types_PcmU']\n",
    "def land_types_download():    \n",
    "    data=pd.read_csv('land_types.csv')\n",
    "    data=data[['id','land_types']]\n",
    "    data=data.set_index('id')\n",
    "    data=data.T\n",
    "    data=data.to_dict('records')\n",
    "    return (data[0])\n",
    "\n",
    "land_types=land_types_download()\n",
    "built_up={0:'Не забудована',1:'Забудована'}\n",
    "\n",
    "def Data_Augmentation(number):\n",
    "    unigrams_list=['продаж','земельна', 'ділянка', 'іншої', 'інфраструктура', 'інструменту', 'інвентарю', 'ізумрудє', 'івасюка',\n",
    "                   'яром', 'яворівського', 'шанове', 'чудова','цін', 'цільових', 'цілу', 'цілорічно', 'участке', 'тільки', 'тухолька', \n",
    "                   'турківський', 'твердій','сухий', 'суха', 'сусідніх', 'сусідні', 'судова',  'сторони', 'сто', 'стихії','селі', 'села',  \n",
    "                   'сайті', 'руська', 'росташування','рокитне', 'розташовану', 'розташований', 'розміщені', 'розміщення', 'розміщений', 'розмірі', \n",
    "                   'розділена', 'покупцю', 'показ', 'повідомлення', 'питання']\n",
    "   \n",
    "    random.seed(0)\n",
    "    random_words=[unigrams_list[index] for index in random.sample(range(len(unigrams_list)), len(unigrams_list))]\n",
    "    tsv='для ведення товарного сільськогосподарського виробництва'\n",
    "    df = pd.DataFrame(columns=['text','land_types','built_up'])\n",
    "    for sample_index in range(number):\n",
    "        df = df.append({'text': tsv+' '+\" \".join(random.sample(random_words, random.randint(0,20))),'land_types':6,'built_up':0}, ignore_index=True)\n",
    "    return df.astype({'text':'object','land_types': 'int32','built_up': 'int32'})\n",
    "\n",
    "land_data=land_data.append(Data_Augmentation(34), ignore_index=True)\n",
    "data_class_display(dataframe=land_data,class_column='land_types',expl_lables=land_types)\n",
    "data_class_display(dataframe=land_data,class_column='built_up',expl_lables=built_up)\n",
    "\n",
    "X_train_zab, X_test_zab, y_train_zab, y_test_zab=train_test_split(land_data['text'],land_data['built_up'],\n",
    "                                                                  stratify=land_data['built_up'],\n",
    "                                                                  test_size=0.33,random_state=0)\n",
    "\n",
    "X_train, X_test, y_train, y_test=train_test_split(land_data['text'],land_data['land_types'],\n",
    "                                                  stratify=land_data['land_types'],\n",
    "                                                  test_size=0.33,random_state=0)\n",
    " \n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "def plot_graphs(history, metric):\n",
    "    plt.plot(history.history[metric])\n",
    "    plt.plot(history.history['val_'+metric], '')\n",
    "    plt.xlabel(\"Epochs\")\n",
    "    plt.ylabel(metric)\n",
    "    plt.legend([metric, 'val_'+metric])\n",
    "    plt.show()\n",
    "\n",
    "\n",
    "\n",
    "def fit_print_rnn(X_train, X_test, y_train, y_test,n_classes,units,epochs): \n",
    "\n",
    "    # set parameters:\n",
    "    max_features = 5000\n",
    "    maxlen = 400\n",
    "    batch_size = 32\n",
    "    embedding_dims = 100\n",
    "    kernel_size = 3\n",
    "    hidden_dims = 250    \n",
    "    \n",
    "    \n",
    "    t = Tokenizer(num_words=max_features)\n",
    "    t.fit_on_texts(X_train)    \n",
    "    X_train = t.texts_to_sequences(X_train)\n",
    "    X_test = t.texts_to_sequences(X_test)\n",
    "\n",
    "    \n",
    "    print('Pad sequences (samples x time)')\n",
    "    X_train = sequence.pad_sequences(X_train, maxlen=maxlen)\n",
    "    X_test = sequence.pad_sequences(X_test, maxlen=maxlen)\n",
    "    print('x_train shape:', X_train.shape)\n",
    "    print('x_test shape:', X_test.shape)\n",
    "    \n",
    "    X_train, X_val, y_train, y_val=train_test_split(X_train,y_train,\n",
    "                                                  stratify=y_train,\n",
    "                                                  test_size=0.10,random_state=42)\n",
    "\n",
    "    \n",
    "    print('Build model...')\n",
    "    model = Sequential()\n",
    "    model.add(Embedding(max_features,\n",
    "                        embedding_dims, \n",
    "                        input_length=maxlen\n",
    "                       ))\n",
    "    model.add(Dropout(0.2))\n",
    "    \n",
    "    model.add(tf.keras.layers.LSTM(units,  return_sequences=True, dropout=0.4))\n",
    "    model.add(tf.keras.layers.LSTM(units//2,  dropout=0.4))\n",
    "    model.add(Dense(n_classes, activation='sigmoid'))\n",
    "\n",
    "\n",
    "    model.compile(loss='sparse_categorical_crossentropy',\n",
    "                  optimizer='adam',\n",
    "                  metrics=['accuracy'])\n",
    "    history = model.fit(X_train, y_train,\n",
    "              batch_size=batch_size,\n",
    "              epochs=epochs,\n",
    "              validation_data=(X_val, y_val))\n",
    "    results = model.evaluate(X_test,  y_test, verbose=2)\n",
    "    print ('test loss: {0}, test acc: {1}'.format(results[0],results[1]))\n",
    "    y_pred=model.predict_classes(X_test)   \n",
    "    con_mat = tf.math.confusion_matrix(labels=y_test, predictions=y_pred)\n",
    "    print(con_mat.numpy())\n",
    "    plot_graphs(history, 'accuracy')\n",
    "    \n",
    "fit_print_rnn (X_train_zab, X_test_zab,np.array(y_train_zab), np.array(y_test_zab),2,64,5) \n",
    "fit_print_rnn (X_train, X_test, np.array(y_train)-1, np.array(y_test)-1,7,196,10) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Приклад побудови моделі word2vec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\User\\Documents\\article\\data_print.py:50: UserWarning: You have mixed positional and keyword arguments, some input may be discarded.\n",
      "  plt.legend(fracs, labels=labels, loc='best', bbox_to_anchor=(-0.1, 1.), fontsize=8)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Word2Vec\n",
      "Слово 'ділянка' -  [('мрію', 0.12444563210010529), ('присвоєний', 0.1221877932548523), ('горы', 0.1125452071428299), ('господарства', 0.10289157181978226), ('неї', 0.10067432373762131), ('пилипец', 0.10029172897338867), ('зовсім', 0.09704037010669708), ('потічок', 0.09689418971538544), ('широка', 0.09640874713659286), ('проходить', 0.09575922787189484)]\n",
      "Слово 'земельна' -  [('увазі', 0.1161714568734169), ('хуст', 0.10643313825130463), ('зведення', 0.10264638066291809), ('початкова', 0.1005157008767128), ('зведено', 0.09715737402439117), ('гакадастровий', 0.095176562666893), ('тзов', 0.09422482550144196), ('колії', 0.09348714351654053), ('суховолі', 0.09305611252784729), ('електричка', 0.09153789281845093)]\n",
      "Слово 'будинка' -  [('різного', 0.11177098006010056), ('садочка', 0.10531207919120789), ('приватизований', 0.10071966052055359), ('облаштування', 0.0977017879486084), ('станция', 0.09768658876419067), ('плай', 0.09451328217983246), ('жилыми', 0.08689279854297638), ('спарку', 0.08635914325714111), ('тихо', 0.08573484420776367), ('грушів', 0.0851108729839325)]\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import nltk\n",
    "import re\n",
    "import random\n",
    "import tensorflow as tf\n",
    "from sklearn.model_selection import train_test_split\n",
    "from data_print import data_class_display,PCA_text_data_display,pp_conf_matrix\n",
    "from EstimatorSelectionHelper import GridSearchCV_Classifiers, Best_Classifiers\n",
    "from ukrainian_stemmer import UkrainianStemmer\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "\n",
    "\n",
    "from tensorflow.keras.preprocessing import sequence\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Dropout, Activation\n",
    "from tensorflow.keras.layers import Embedding\n",
    "from tensorflow.keras.layers import Conv1D, GlobalMaxPooling1D,MaxPooling1D\n",
    "from tensorflow.keras.datasets import imdb\n",
    "\n",
    "from gensim.models.word2vec import Word2Vec\n",
    "from sklearn.preprocessing import scale\n",
    "\n",
    "    \n",
    "def ua_tokenizer(text,ua_stemmer=True,stop_words=[]):\n",
    "    \"\"\" Tokenizer for Ukrainian language, returns only alphabetic tokens. \n",
    "    \n",
    "    Keyword arguments:\n",
    "    text -- text for tokenize \n",
    "    ua_stemmer -- if True use UkrainianStemmer for stemming words (default True)\n",
    "    stop_words -- list of stop words (default [])\n",
    "        \n",
    "    \"\"\"\n",
    "    tokenized_list=[]\n",
    "    text=re.sub(r\"\"\"['’\"`�]\"\"\", '', text)\n",
    "    text=re.sub(r\"\"\"([0-9])([\\u0400-\\u04FF]|[A-z])\"\"\", r\"\\1 \\2\", text)\n",
    "    text=re.sub(r\"\"\"([\\u0400-\\u04FF]|[A-z])([0-9])\"\"\", r\"\\1 \\2\", text)\n",
    "    text=re.sub(r\"\"\"[\\-.,:+*/_]\"\"\", ' ', text)\n",
    "    for word in nltk.word_tokenize(text): \n",
    "        if word.isalpha():\n",
    "            word=word.lower() \n",
    "            if ua_stemmer is True:      \n",
    "                word=UkrainianStemmer(word).stem_word()\n",
    "            if word not in stop_words:\n",
    "                tokenized_list.append(word) \n",
    "    return tokenized_list\n",
    "    \n",
    "    \n",
    "    \n",
    "land_data=pd.read_csv('land_real_estate.csv').replace(np.nan, \"\")\n",
    "land_data['text']=land_data['description']+' '+land_data['land_types_source']+' '+land_data['land_types_PcmU']\n",
    "def land_types_download():    \n",
    "    data=pd.read_csv('land_types.csv')\n",
    "    data=data[['id','land_types']]\n",
    "    data=data.set_index('id')\n",
    "    data=data.T\n",
    "    data=data.to_dict('records')\n",
    "    return (data[0])\n",
    "\n",
    "land_types=land_types_download()\n",
    "built_up={0:'Не забудована',1:'Забудована'}\n",
    "\n",
    "def Data_Augmentation(number):\n",
    "    unigrams_list=['продаж','земельна', 'ділянка', 'іншої', 'інфраструктура', 'інструменту', 'інвентарю', 'ізумрудє', 'івасюка',\n",
    "                   'яром', 'яворівського', 'шанове', 'чудова','цін', 'цільових', 'цілу', 'цілорічно', 'участке', 'тільки', 'тухолька', \n",
    "                   'турківський', 'твердій','сухий', 'суха', 'сусідніх', 'сусідні', 'судова',  'сторони', 'сто', 'стихії','селі', 'села',  \n",
    "                   'сайті', 'руська', 'росташування','рокитне', 'розташовану', 'розташований', 'розміщені', 'розміщення', 'розміщений', 'розмірі', \n",
    "                   'розділена', 'покупцю', 'показ', 'повідомлення', 'питання']\n",
    "   \n",
    "    random.seed(0)\n",
    "    random_words=[unigrams_list[index] for index in random.sample(range(len(unigrams_list)), len(unigrams_list))]\n",
    "    tsv='для ведення товарного сільськогосподарського виробництва'\n",
    "    df = pd.DataFrame(columns=['text','land_types','built_up'])\n",
    "    for sample_index in range(number):\n",
    "        df = df.append({'text': tsv+' '+\" \".join(random.sample(random_words, random.randint(0,20))),'land_types':6,'built_up':0}, ignore_index=True)\n",
    "    return df.astype({'text':'object','land_types': 'int32','built_up': 'int32'})\n",
    "\n",
    "land_data=land_data.append(Data_Augmentation(34), ignore_index=True)\n",
    "data_class_display(dataframe=land_data,class_column='land_types',expl_lables=land_types)\n",
    "data_class_display(dataframe=land_data,class_column='built_up',expl_lables=built_up)\n",
    "\n",
    "X_train_zab, X_test_zab, y_train_zab, y_test_zab=train_test_split(land_data['text'],land_data['built_up'],\n",
    "                                                                  stratify=land_data['built_up'],\n",
    "                                                                  test_size=0.33,random_state=0)\n",
    "\n",
    "X_train, X_test, y_train, y_test=train_test_split(land_data['text'],land_data['land_types'],\n",
    "                                                  stratify=land_data['land_types'],\n",
    "                                                  test_size=0.33,random_state=0)\n",
    "\n",
    " \n",
    "def dataset_to_Word2Vec(X,model_dir='word2vec_gensim.bin',ua_stemmer=False):\n",
    "    print('Word2Vec')\n",
    "    try:\n",
    "        model = Word2Vec.load(model_dir)\n",
    "    except IOError:\n",
    "        X=X.map(lambda x: ua_tokenizer(x,ua_stemmer=ua_stemmer))\n",
    "        model = Word2Vec(X, size=1000, min_count=10, workers=-1)        \n",
    "        model.train(X, total_examples=model.corpus_count, epochs=10000)\n",
    "        model.init_sims(replace=True)\n",
    "        model.save(model_dir)\n",
    "    finally:\n",
    "        return model\n",
    "\n",
    "model = dataset_to_Word2Vec(land_data['text'],model_dir='word2vec_gensim_all_corpus.bin')\n",
    "\n",
    "print (\"Слово 'ділянка' - \", model.wv.most_similar('ділянка'))\n",
    "print (\"Слово 'земельна' - \",model.wv.most_similar('земельна'))  \n",
    "print (\"Слово 'будинка' - \",model.wv.most_similar('будинка'))  \n",
    "\n",
    "\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}