PySpark_Desafio_Módulo_2_Cenário_2.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "PySpark-Desafio Módulo 2 - Cenário 2.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyPHFaWtSvl6rhB+GOFpWDdx",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/saurater/ciencia_de_dados_pyspark/blob/main/PySpark_Desafio_M%C3%B3dulo_2_Cen%C3%A1rio_2.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Spark MLLib Árvore de Decisão - Previsão de Derrame - Desafio - Cenário 2\n",
        "\n",
        "Sam Faraday"
      ],
      "metadata": {
        "id": "eFC2FXm6ZKqd"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ukymHWFWWreW",
        "outputId": "98bd619e-09a4-4d63-b11f-a4e704e3b96d"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Requirement already satisfied: pyspark in /usr/local/lib/python3.7/dist-packages (3.3.0)\n",
            "Requirement already satisfied: py4j==0.10.9.5 in /usr/local/lib/python3.7/dist-packages (from pyspark) (0.10.9.5)\n"
          ]
        }
      ],
      "source": [
        "pip install pyspark"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Importando SparkSession e criando o app com o nome de Minha_MLLIB"
      ],
      "metadata": {
        "id": "LuiMi4Ot8nYP"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql import SparkSession"
      ],
      "metadata": {
        "id": "T3lHTojdZNed"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "spark = SparkSession.builder.appName(\"Desafio\").getOrCreate()"
      ],
      "metadata": {
        "id": "cv4RuTTaZ9cf"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pipeline de Machine Learning"
      ],
      "metadata": {
        "id": "LaiKf8xa_PGT"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![image.png](data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAqsAAACHCAYAAAAx6hCVAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAAEnQAABJ0Ad5mH3gAAN57SURBVHhe7P1XkBxJnuYJ/p27B+cELMCBBE1OKiuLdZGsqt3uaTazO3IPOzd9I7cvtyKz8zAiu9PVsicyI7fv87I7+7B9MztE5q77ure7i1fySoJkQIJzBBCcE3cPp/f91FwjHIEIIAI0APgHaJi5mZqamv6JfkpMLVQWbAX4w35bLBatVCq532wLhYItLCxYKBRRCLnj1UmFLFLZq6GGGmqooYYaaqjhSUMoDO8ruf1wOGyRSMSi0ajjhcBvOVd9fCXc9pwI5qpklQBJZesJKluOBcdDViyULZfL6bdPJhz8DbEN9h3KVfs11FBDDTXUUEMNNTzWCIWLonrZYF9kMx6PuwBp5TfB73vC6uMux0rHPFYkq56cemJaLgUJZPM5yxfCls7HbCJdsuGhKcvnSjY7O2sLCwUX5ybUCGoNNdRQQw011FDDE4tIFJKatHg0bG2NddbamrCengZxyClL6lwkVLR4JGrlaMLK2ibogb1Xsup/0ovKPmS1UCjZQrZoubzZXC5mZ68MW//IrIhqyLq7N1lHR5vV1bvLaqihhhpqqKGGGmp4GiDKWCyYzc+bZedzduP6FYuE87a9r8029SZEXstWHyvqWMQsnrKQyGqSqQIirMuxZrLKrg/5fH6RrGYyC7awELXh4QX76tKUiGnS9uxstj19zRYOly0eW/0GNdRQQw011FBDDTU8eYBBhkpmBXFPxtcn5syuXsvY5asT7vyeLSHb2h22cCJudeGY+GLEIrHoivNX10xW3bC/fpZK9KYWrVQoWF772UzCrlxL26lzo7Z9e4sdOthjbU0mpkwCwfA/N7ndjWqooYYaaqihhhpqeHIAhQzBIsPwv7Dl9Ff00aZmzU6cHrXR4TE7uG+7tbdkrD4eslQiZmER1VgsdgtvXBdZLYicFksiqbmclYslS2cLNjnVbO9/fN727D9oLxwxa0wEGSyJTutWlatrqKGGGmqooYYaani6EAzpLw3sF6xYjtpY1uyL43N29fKQvXq03VpbitZQH3EvYMXCsVt6V29HVm+aNAABdSQU0loMlqqaz2Tt/S/6rXNLtz131KxeRBU4Jn3z5TXUUEMNNdRw36Gayf2r4clCTa63wpfJxi4XuF91WAGhojWmTLyxQSQ1ZmeuFWwun7JcOedG7+GXVX2ld8RNPav5fNHy7oWqBcsXcpbOJ+z8lYydOTlk3/+dw7Zti7m3ukC4FLFQOFhbawn8ZlrA0kMUFXys6ke6v/2xy/MR3Mnfb/nZtaI6v7fD3aa/sXHrU4WXlUjJxfHHkLv/HSxN8STi6dUJnsgHgIxXL43lZ7gqcDTV/iHAen3B0yuDR43qEvVS4Fiwz19+qRqymDvC/s3SCmRNrILOxN2vvPZD0qcg5lJ6HvxynSghX7mVjPW9y1Xxwg9hhO/mXK2O6lJ6PLHaE/jjviRuldMSltKorieCv/xGXvADybZUdmt1rm2U9ua73Alrjb3aEz8olKX1Jq4VCsUVREtDIWcr/A0pN6EyVqTf0vMAnKsun2D65RL8k976xLIctw2XK+vhs6yokgokEKBaWgHWW3JB/KWrOE6Iur/9g2Y/+2DUntkdsp2bS24qQCKatGg0YpFIVPIPnk2s0m1Xwk1kNbfAclUFyxV5oSpjY+km++2n07a3p95eEjtW2iq8pcomoKGEShbL2nfnAceiTgABvb25cFbP0t2APFQjuIu/F1ge405w166V9N/fh9kQCBxMsAccUWUpMsqE53XPXNBZT0yffLL6ROtE9XMty3ugAfwlePuGaLgSuQWLR32aSo8rg5/++iUdWU9RPdEy2OBY8gngVtmzNjirHOIVAhq6nKwybYzfVMPoQRCryH45GnxGhjc1HG5NP7i/q8J1Df1OCDiI96DJqrvLY6p31VJbG9QoqGwD8PSEXEV+gBrB71dQXT6LclwJnEPakqOugR/QCSYqxclb070JS51fd4JLZYPI7OY8Ywd5BZp0ymVJJSDjycknerspl7KKJTmEGCanrKrJKql5P1pdVn5/6VhZpetsRdwsJhsLqw4vVszGbSgfJyuVq9t3e0GEZbi13H0+iBv48yA99jin4MpfPFDb/89beT3xDXv+UJOlYmVLxRMWYe5qNGZhXoASbkdWw45pVwKZZlsSqy+WSjY3l7ZCoWxbtjZYLKHTy9LBbdz0UCrYUjkZBBW7fzjXUnC//JEHgeq8FFRGBecEaZ+41lsl1LB2+BJ1CojsUGodoEKiXNU2dNugXFFWTO3JJKpPPG4xDmwVR1QwBlBwcgi/bMwDUqt10daqwTVMrw/2iqqAiqGcs8Oa7T0JQP7BaBk+1ftVDypBjkFWqRvKJXQE319QJUTvaQDSKJTjllcNVizmLKILI67+qURYBaSbU6WbVdSgqkbLbgV5WinUsDKqy8dZuZMbPd7IM+dkGcA3UP3vtUBxve+QjIuuHz1qC+WQLUh81CVRBf/Fy/Wk/PgC+4jZ/ELBMjKYkqtgsROzee3Oa1sox0RUfXmDwBcHwev9aqXF+SBAcOMq20QoZgvhsKV1SZHLdB9i5HVTz5OW7I8IS2ncHMDy30pO6dFXeSvEB3QOLrp1k/KQXrB0WjxRz81UU3rVmRLA9k4IlUo+lshHni9T5d3i/+n0vPUPJ+389YS9+UajdTb5WD5HQUG5v8olL2WVInG7MaJEpeWd7VLtJK0luDLXUIlFle+AzCyWSwVBagEKIsp0C69tWOBWQKOCtjtDRUEbjjuDtbbAg+dyu3fG3WVzgyOQMw5GamGZUsRGJs3SmaA84xJjU6NZg/QCZ0MbMaLyqnw/4q5RrQcr6r7g46x2HqyWzlrSXw1Pqk7wPY/xMclYJgrZgHSoLpFNS6566Lj2kxJwm2waWUd1bOUypXCC4StHSLQ3L+84Ppa39taYNTWoWGgQu2qQq6IqJvrHllK4E1zMp9ouHx2QOeHaDXy0WWuLWbt8ABVVJFS2uVzIhiYUR+d6m+Ub1KYpy1kEvahcGXfaURB5GR03u3Fj3trb6m13d9myElb/UNEta9PTaZYIuEuAihzHZ80u9aetXCzY7p1N1lAf9NuG5ZHYcgew1C8Y1DyAI/go33u33rrlcdY7XwargceaTJtNTJtl5QNiOhCRf29olY+vD/w9fIqamw6rVa11hfLx9QHXu16+UNiuDUtPhmZsc3eT9fYElwWvwpBTwqp3EB7PnlWwpIslSy+E7cSJCYslW+zQnrDq05wjp6evZu3q9XF76Ui3tTfjH/HIHst7V8FSWfk9zjDuAIjP8YmJgn1yesIS4r/PH+qyuAocWaflittkw8RBzqS+lKLSgotJ6GxLIdlbGD63VFglGKrsmS9TsdZqLKB3ArngXOWAsnNJ3PDDj8/a0YO7rLlx2upk5PGYqHQ04npX4Xy361mN/Omf/umfBbtSpVJOjkQsW7VUoZCX8oZtaj5pR56JWVxPsERUfWH4akZVkzI7Nh2yn78/YVfO91t3V7vVN5WcG6EdHlocQlg5OxwLHk+VV6UL17Ft7a7XsaAMDE+nleBVOdaZ6ax1OMGvPa113XF92XtMEJQW+kDn/QXmnLzTb6fPZ+z8+Qm7KqO6dDlrc/Nha2yMuoqJaSdctVb/sBJ8US5p2q3wcW53Hx8HVMe79Th3Yo9QffZW3P7sMqwr8qPFlCqpn781YseO37BzV9J28mLGzl1bsFPn5+3MhbRdvJy2G4OjtmVLo9XLEbhHq3q+pd2gHKGfEIUF/TpzLmvvvH3eQuVZ27a1TToCXcABL5X5zWT19vJYV7GuK3INdwLFWZBYfv7rAfvyxJj8dKtt2azjOkHzZHAyZH/703PyDSHbsy1l9alAmuEyjROujqhCLtt8PmTnzmXs2JeTFo012bbukg1Phe1nvz4nEhu2LZtSqsgUfRn6h8ze/3jMJscL1tfXoPTRHeqWoMkDgg7a4F9GN78yLCI2nVblGHP+iXl7TtsqdcytWK5/we+A5K52zTKsMdrDAjlfDZQF1njifMl+/t5lO30la2cvTdvFS2m7dnnewtGkNTeHLSkmw9xKSOft3theDhq93D+sRirX5UVsjp3I2/GvrqkB22TtXeID5aJF6X0lMrht8r65cWesSwzrinx34BZB3kM2MWN27NicOFbC+jZHrCEVEe8q2+encvbFyXHb1NNobeIsET5l6jgTTTI9Owm4vPInyDQyREu5HsrEZ/DDKLu2iIrOppGxor334YQalWHrk30tlML28WfXRIzT1rO10XVAkJpLq9K2dF/Md2koPf0IiajC57AGPhRVLMieixBVkUzFpZFaUh4iEVIiowoui/qj/1lVCJeujlpbR7fVJxdEUHVdRCnyNStIMPcJLlgRZGdFRCLBME04HDM1doVqhr8ECsl/VpUemWyp2eYWmDPAOainv87FvAlUa/4fvalscTIlkWWIKl844Pd64MpXRUp5n1el+8tfX1RlO+vSWW9aNdCTlrP6SNAfns/HrSy92Ly1yZKNcZtJx+2Lz3J24UzQ+6qCvyd4PSC4NwUr/6qxpjj6Sc+PD/z2Yen4ytc+jVCj1jq74tazqdV6NvdYJFZv03KmDfVNtqU3Ylt6ktbTWS9fkLHKCOEiKL8lecjqygvuGOrAkF9IhCXenLBYQjsq6lwR4iIBuICVBmnU5PD4oHtzr2UkxuHRrKVpkSBpEdCzZ8wyCy3W3FlncXrRIwUrltIStSJXxgiRM5VgJpO0bI55eyWLlSKmutqam+pUade53vuV4OoXVZ7FUFQkiiNKU0ZN7w4VtausVW9g5+DaVRHrnw/b6XOjtuCqIf3hnCdFa4LX1ScTBdenra3q8Fw2aslQvfV1t1gq2SBClbD3356zMyfl31VuQf25/rJwV1QcR0HENC1BzmZ1v2LcpRjxLxE52QS7TzZKNjdvNjozZ0NytJPTOqTnzkyHbGSUz9m3axt1a5U6SF9xrQsypblJs3n5Zj5zDyCn89mSZbOBfLLpss3O5Gwho8u4XtEo+qbOmD3/ynbbvaPTmpNR2a740ZWSDY3V29xcMKpGHc5MP7jmgtKb4V46l+Oz/7oPEiRP6fmS0hfH0jXcZ1b5WdA+vfFRXbw0HYAdnUDIut7xSPTsLmUc+clPfuJ6Vku8kabEIQDO8FU4Y5MRKVXKDuymd5Qbc9dqBOqrS92puWzIzlwOWyEzYzv7mlXZmV3sZ/g4bDOzCbt8I2LXVUhppR2P6+H1YPTAjEpI5/rLdm2kZANqWV9S6254OGN1jUlXkU6oQK5cD3pJh6dUgCqwGHWf7osR9av13K9z14dKNj0XskQSYzA7c7ZsVyT8cKRBLcKwjmXlENXCVnYJlTJcEcuf9LZYV+THC7R2mE4xMi7lPj9rdXXN9ve+32B7+1Iqu4RdGsyr1ZWx3duSNimlPXu1YFP5sF3plzFK2ecWFKsxZBnJfEByuqBwbUyGIJnW10m5JQBaZRclv8uD0oWpkN0YCcmYymqJqUVP74nOh2UBQ9KTSzdCdkXxRhRvPidZ63xcelSUTpy8EOjHsIyftw+n5hakZ1E35DGvSvWCdOjakNmQniVfClkyFnINIket3IuBaMPqDaQnVScYuunpStquHfW2Wa38Sdnq9OScHdzbbN9+OSUHl5ADqrcJNU5mJOPrKsMbE0UrquwWCpIH5ToS1rGITStOLK6WMnY7FbHxiZgl4+22qbPR2pvNJnXdxeGE5BGzKzdiNqDrxiXXkBgKy+IFZS95uO2thfikyuCxQiJk1wayNip5b9uWsIbGsM1lQvbep9OWzuZt364G1xt3w/nlso1PxaUXUQvVyV7pwpGZXR4O2eD4nPV2N5v+W798gqWa5Z+j0kXppPSHOY2DOn7hWmCz/ao7rg3npFtF27OzXn5D9+hn+kBE9UbIpqRHMF3aRbOqZM+czVj/QFmVaL3lCwkrFMLB1ARlQa7Bzind60p/SP5iRvUF19FLFPQdBZ6AOhFUajq3f0dsML1brY4DZDWnAsFf99+Ys56ONvvRd5K2Y1tc9WbSro7mbD49Y20t9daqBgjTeyBXF6mPJV/8LXIKyk7lKZ9/TmU+KFkMSF7UzXMq2zpezpZs6Fm9eC2suiFjW7apMSxZ6zI3vez6CNM8ytY/ERYBkxykZzkR27OS0+hkyFLyKwxj4x3GlIcLN3I2JQLX3rRyf9y6xPAwZCZBcBtGF64OifNcKogk1ltbY8xNfcFevji/oGcPq6xDtn0bC+dHXH328Ym8fXJiQdcs2JWBvE1nE5YQv0rIZ166mLX3Pxu3y0Mx+/LCjJ0+M6n9gs2X6qxNtjUuGb13bNiGJ2UDaji2NMfsNx9O2LC4Vj4Xt4nxkg0PjluLbC+ViNrFS2X77Vezqk/H7ezlebs+XudGxBobVY+m8/bu51P2ycm0XRuM2IkLc3b64jX59qLupQzxjAjUaZ3njGKp+gkhvnBl3Jpa2qylfv09q5E/E4hEHFo9rhtZZDWXLdr4TNSm1QI+sIe5BOSAhKpD1Z7SSEspT12UKuXTtmd3k9VJuf7u3Uk7eT5hF86P2/lrORENFLKojIZtc7fIiZT7g2PT9snxKQlh3s5dLtjIRMpm5nKuAs2JQ7z92Yx9fnzCLkooF/vTdun6jBS50To7lNYVs/c/GrcTp1R4NzKqOEetQ7XitMjMlyembEhOck7sdkQ1bF6t/N07W/Qsosi0rvWfIl0JwdOtEeuK/HiBiRz0hU3joKS48WjKvv5C1A3TTU6X7cxgwZpTRdu3JSXFliw/G7PTcjiXr2bsujxXsq7RNm0y+/L0vH304Zx9enZc8sva1YGcDd+YsN6uJtdr++7nGftARnDpalqNklnJddauXJMequnWtzXpXsD48MtZe/eTSTt/tWznL8/JwU5aqj5ubTIyhhj+8m9n7MzVop28pOuvTdrM/LRt29LiWqS/PZbW9TK+Kzm7fG3eblzPWE6tpo5OPUhYFaAbbqFyovm3Ml19UnWCESMqgVRlPv+1ATm4oQnbu6PZ9m/PW0yO5P2PbthnZwsiADOyMcnm+pTKrlk6ULAPPxiwMyrXi9cl9+sRRzS6O0UiLpfs42NDNjSUt0hRDZrtKTt1akL2XLILV+N24YLk1J+Xzc/JoYZEaIPpJDRhA7pwayE+qTJ4bKAyjUpPJmbU0FCjI5HIy76jagQW7NiJUWtqStqrRxtsUA2a996/YNeuFe3sddncSEaNmzrrUsOVxul5EZOBoRnrlP2zFuO7n47IZos2ly7Zru1q4FAJi+j+8t0B+/IMpDNjN+h1ypWtqTFi+3YnLTOdt7fevmGnzqleuR5WY3pKpDlj3ZvqbFRE6bMvBtQoqrPpmbIN3Bh39druXUk1ophOMGjHvsyqvsqKtM7o3uPSurh1dMRcz1KMOlEPi0cI1Cj4vSasMdrDwmp1HIB8qljsxlDJrg/MWXtryl58JiIyYdbcbnZFBGd0fMa6Whutu0UE9WrJfiti9OEXs/LBedl81m6MzEsP6ixVZ3bsjNlbnwzbOfmCa9fyKtdZ+fMFGxkRmWlTQ1a+4cJVUx0/Y9t2NNlm1eGQ38++yjkdONefswvX03bx6qxkGRVJjtqHImqnzk5ae33SGtUIonfvs69K9u6xa1YMJeyZvqCHdjnWJYYHLLMy3Zwi6qhVIVS042fLavRH1IBKyf8u2Oa+qOtMuax6sagasUENu707o24Q4J3PC7KBWcsWmiwk4xlSA+v6II02+dmOiBpsKqOTRRvL1lu2XGehkuxvsqxGWFkyUyNR9nbynOrj0bjKK29btibs9IUFyzACbglrTCYtKa62Xf55VIT5V+/127XxhEVTTZYuNto51amjo0PWp8aFlXL29gk1OIbrLDsv60jVy27ydnRvgyWSauAoRWc6i1oX1KcAsnrx6oTIcrs1N2TEJe55GgCHgsOOxK4HlfwtjrJom8/nxcZnraUtYV97vsX2bY3bnJpD5y4OujltIuV27krGGhta7aUDDSIvYbWWJi3cIAojg/lMJPTkqaw1NDTYyy922JHdspiFuH16LGv9Em7/jaINigA3diTthVc6bP/+dldRbuk227Ot2Vok0M0dUXv5pQ7b1ddQebKgtXxbK67BwSsIJZaTPsxrj97RYxfNjktu+UzZWlVBJVXhTKuxMqNGTkEyf/5Qnb10OGH7dqrSEfn5+NMpWyjU24t7u+21Z7osmQ+rFVmwT86pVS09KcryisWsbemM2SuHO23/nhZLl0v2xZlpkRndXzXI5t4me+FIl73+XMp2bGq1nBpSZ88t2JzYdF42sVBI2sLCgm3pSdirL6fs4DMJ6Z0MVdefkLNLqPn/7LNJO7yvRe2pgn18OmdnpH/RaNLCZbEkOQn6V2tYQtEyrkym50oqW16SCtlrz/XaG0c6bYdsbHu3bOvZzZJJpx3a2WBzEwt26qtRy8u4t/SGraOtSXLJWzju3je1UKzR5ubKVszN2tEDTfbK83VWlyrYhcuzctQs1wIq9lnDhkS9XPO2nqTFQxm7eGnChlV5MuReUsXT25tyPaU7e8yO7K23o0ca1dhss7mZkJ38csEmJuRT5FR4gapclF5ov05+vndTu+XyGTcFjG+MT2XMPj8Rsslxs719LfbioXrbuTVicTUsQ2VV4vLdyca87Xym3p5/oc2O7kupoVNv167M27Cu6diiinxnpxXj+I2EfevVzfbswWany5+fn7WTV5OWipXstUPtdkR6my802PFTszaoOiWoMAm+gvBe8MnESrMiONQu8imOaoWFnE3PSiby75+dz9vZayrTjri9cqjVdvbW2+BYSPX0mI3PSQd0YbkgktSQspfl/599pkW2HrKL59WQPa9EJYCE6uRwKW9R8TeGhK8M5eyTk2osLGTt+QMd9sKzDSKkIqin+tXoyVl3e9yy6QU7fbEoH6RrRHquqbHDqgXbNzW4/G50wKVK5WBIfCFbtuGRCR0LWypakv2k7RKjVWpgMYoZicdsOpNwjaohepuvlaxFBvPjb5j98e9G7WuvJy29MGnnTo3YjOIUVZYlkcitHRn7+2+a/dGPIqofozY1PGPDQ/PW3mX23EsySpuVzZatWTL97g9arKU9b52tcfvR982+9235Ydnhx6fmbHQuYof3NNvf/2HC/ugHEevritvkbNlOqx4tU0+W81YXnZE95ewPdb8//nGPtbfXO5Ia9Ko+GNycNOMjFVTrr58DVA3OV8dZDdEIramQHT7aaM8dMXvpubBab602NZszNdjUkg6GZNu6k/bc4Sbbs6vZUmL6mfmsc2wXL09aOdpidSrhho6QpdpjFku2W2a2ZKODZUvwJlm8LMKvVr6MYNeeBmuQMHhLvUWBzqJN7Sk7erDR9u5oc72qi0+5lgd46hFMqsdvl8pxya1sP//lsL3/26syuCHb3ltyUwL88Ew5whCdGh5HY/bikTbrlI1ckZLPZ+JqeSUkFxGeVjlC1WiheIddGWDoUDqg5n1Crbu9aoy8/FxEhKjOdm9ttqJYaP9A0b2UEZExlWQN+SzzecJqJabcMhjZeRyX7q3Kp7ExbF97qd5ee7bVDu3rFXGmQZSzTJ4XBZqtvlFG2RpVY6jbskU5BDHsgu4dcrr/AC3tMQRr87HKhyuaUFyVWsKOHuqyF58ze+FoyholD4osK5nwpvaU5BCL1KlCiTqZqrits1uRFCdXKDiigDPm5cctm5vspefNXjwatt07uuTk6lTpzTodwjBrkti4YCSkV43ATtkxvZZnL5n1jyBVkce+eour4ZpXpFj7JpspNYiERuSn1ZBJ50xu3a0ywbdlWGlE/NZVkn3bVUkzJ0jIiZBAgK8Mp62dNb5fTrhGzYE9cVXuUQux7o2ipuJ11tTcIfuNujmpqkItm4/ZlK6tV7pdXXWWLxetWXzm0EGzrZtV30hHrw3M22yp3tq3dFtCdUWTIiTqIMsxkeNM1Zw78vP0VBI0FPziQG4VGO0y4JRwL9aYjYkY9Y8tWC7eYs1drdbUbtbWGbdorM0mJ8s2Kx8AYYlY1nZti9rrL8Ts1Rcitn+fGqwiOBMzvJTDfYijiJSzAmRtvqT6/7kd9vWXxREOJW3P7s3yBSJJk3Hb3ScdSRTt+kjZJkSIIXXD4wVra07ZLjVKHhdQNpRuRv4xw/Q41Vdbe5jvb86GBseK1tISs1b5zfl0yU2fmEsz2pAVGYzbJpHO5njJTRloaEhaOFTnbMVNNFVoVSuyQ1yyRY2MrtakxcJJ1X8iySrvVEx1nrMbeWBFp76MiDzHYrKahGxaJLYoLj2ZKVmkrs66upWO0mpWYAQ7qwbiyBSTYkK6VUmNzaLt2JyyVtl6k66ndx7cPF/1/mLFOsGRk6rtSuDUbU5XIOWX0vN1CnFWU4PBkZqwCjydK5r+u6EDCM7kXNHOXyvYjYG8ZWbSpqK2nFrXWQk1myvZucvz9ou3Ruz9j4dsbHJYpTNj4fyC7dpuIkdNNjsyaR9+MKHz43ZtUAxF1hUKq+Wu0kMoqUjOEqG8MzoHBFzDmuCXH2EnXIpbd0e39fV22msv9Nmb325SBSUZShnQh3Ao64ytQXJOSLvyMoDxyXkpc8TGxWg+/WLG3nqf4f4Jdz6qC7Eh3p6IR8NuqSSmtbWogtmsxgkLIk+pYXJluGTvfTxpxz4bsGuXJ2TsJh2SrkjUbs60QjG0oEpnwVpUAVGh0lfK/O65TFH5Skq3svbb98ft/U8mbXhWLfninAw0LR1VBqSjDPndbhjiyQW2sII9iKWGw1FE4yYRhqNypGoA0isWixXcHOX3jy3YWx/N2Omrc25eMC3rkmQm0bheFGyf6UXMKUaPSqr4WHEzFi04+cdl+6lUSURYhEPXBD5l7cvT1PBwIZE5MEy8uQdjbbAvT0NkytbZkrItmxhdKdgvfztqv/pw0r68kLYZEQz0i9ERevFchSbdiKj2hrhKJVzgZSzmj1OR8mLHlCqABHPhVFlSsaMvaA8LyJdk1198Nm/vfjBjn345av1DWZtVy5SXr3j5JKX7sNJ32S0BwDQfVyW4l0hmRACyUsivLozau5+O2SdfTNrs9LT8wbzuzwuCyq0q5ABs/f6TDTcMW3mzmy8qUQSzE3mVdcEtU8hoSXoh58ru+Llp+/UH4yr7SSvms8ZyYrxHh35EwvLDquv13xGhlsa0FWIZm86pHpDcsmnuFXV+HplMTeYtJzK0SY0JfEuTLuzoDLkOjCn5mCYRoh3bWm1CBI7evauD8AIauR1WR2/UYwCnQfKlNNin0yGR0aK4UMn27YxaMhW1Gzf0TJmC7e2L2s5NEVvIZG18QscK+E8VfCRvMVVoNOOjFJpQ4k18Z09lF6gHaQRyM7gb6xzz03HZPCs6xEVUS86+wmSEF9l5wU5xk8x9EXKFomRfspjKlZemmHIRTcp2wzFXjxLZNTakH0k1Lil+lq0jHw8a6NaDAZmnlHxvrTYIjAVp89GspcRmKPTtUtCW5rhdH0rbLz6ct2v9GWuvj9mRHc3a6ppCztob1Uo7Wm8/+nq7/eG3mu0Pv99lP/p+jz2zVwouR/bG63H78Y+6lGbJLl6VEzo75UhMPhy2Aj1CJd0oK8FQgTrmBZUB5MiHGlYCC8MAlDEZTltP04J973Wz3/t2nX3n2YjJX0hpVYKSKzFjUuSkAvFLkazKX0Yk7Y5LFrt3NNoPvtNkP/xmg/3B99vs979Tb99+scW1zHgBLq+GCTUFKrOg6yfk1DJYnzzf2HjGxgambXNzs65vs1fVAm9uk0OTA8hJvAXFL/CWsGucqKKSSKMKtGZZ2SKlAy/sTdrvfavd/uA7rfZH32u3f/hmr734TJslosySVDOWitA97dOJFa3AfeADu8W7VcpH5ZuRDfXLmZ44OSgCG7cffafBfvf7qmRaJbv8hBqfgXNBHi7dsqiDZFGMwGCZCS0C4WhB8PJmzNKSGyuz1vA4QPWr7dhSL41g1EMyL+Zt/+4Wa1OFOj4ZsaHBkjWI9Xzv1Tr7nW+ZtbZKc0qNrrJGD9xLTIgfoCg65r9+KF7qhohZK4BhXwIfZ8WWWQIxLJ8wN212+uSESEveXn2t0/7ef5m0fftjIqsZYrh59rScoqGY5ZUsPa/6T0LyUVFrTYXt9ec77Uff6LAfvN5oP/5uvf3gdzps+x5GAsgYmkjG2D4dWhmRvOhZpbxnRVS+PGc2MlKylqY629wtN6zyT4RyKruyffPlZvvxd9rtv5Ddv/ntuP3gWz1u6p1zo9S3bBSfhkm5ELFwTA3URNS9WB28XF1w0wBZ0zUVhvKUbGZWfxVfvMzmJd+wCEJcRKlB9cO+PU2ONH1xbs5NK0sp0maROhowj4980HrsQ7WNWvPNbTHb1sfLrSkrqGHWWBey3VvNtnbk9FxZm5ycsbr6qNXHebm1ZAPyqbN6VOaDL0jv4wk1yupUhrGYRWGXsskFBfFSy6sSpP6lbLCzwBerbFVRs+V3pIxtJFxjMqsyp+HYnNKxdM4GRtQwUbwZGQ0vPLN0VWcH9asu1NV0OuTCamTol7M192TColHffzhR3zMq+cMZuG82ixywHl9RmpqnCawnZG4SN6O1FadUVNg8F3yWAoyGEjKKZmtrbrK9u9tt204VnCq+zs46K2dzrqUdjUUslEi5Nwc72kRy26M2IgXnrcSYWn49Wzp0n5TNTM5bRvHjUnZ6V8fV6j9zOeneKk3nICZrhHOgawhPMFgjAtdDRcE8o7po3i3unpAhhItik9BZCZZGiIunFjJDexQL8q6XQW2SU4mFZ9Wa5A1/kVmxWzd8oXjdaqxAdMvSkWK0zq4z9UMt51NqQZ++MqfWd9G62tXiU+ujzFdvFIcv2IyPy5Cms2rMxGmHWFGKFYYY43CVLgZb1LXYcHundC2yYFPSCZqCfI1tXnpDq7y9BRsn5+R2abbaqvAR7hQeOyxZBUQhaNmxjbnSKZcKqlhky9isno/Y9JaWo2GLJJJuyPBSP5VcwfJqIDK0RTIFGgJypmXFYYSRnnR6xvjaHW8Hh1SjReQnIKxh6Q6VFe3922J5Wa8WanigaJVd9rAGZ3nSWhtDtrkr6Yo9k8laWewwasE7AqMiHkMzwdxyiC0fCsAvyyhdDzx1BYMbMf3LSy/QG+yytzFiE+M5O3EqWMnjhnxDviTfIUWa0zVZ1TOheMr5m/FJ3WM07aaq0HMHSZKZK29Fm5gNu+vP39A9VJnT+REWgyVeVJGi8kfFQtSaG1PW3qSKX5USi5+Dima6/Vv0a7XwGIHsUt6RGC9alW1qtmhfXjB795OcvffRdZ2ftx3bm9yUj+aUWVtdVK4h45ZK4kU7luvJ5yPW2x2yrmaVN8VWbLBBkZ3LKvOTIrznr5UsLnm1SJ5FrpGvzunP9Dz+QzrU3WwtKuIvT8w4GZ24qjr92qyxaHzvJnEDxWdloa1dcZufz4pAj9uW3pC18PpKZemtVVEtl9uFhwR69kdnRFrDIWttiYjzlG1LT9TNAd3aGXIfU2qULTXVqfk+O2fd+r19U8Imxmbs12/N2S8/KNix41PylSV3nJcTI7KlUHlB9SgfZtJNFJhmEStLpmpEunfJGY0oMGpQdPJmqk4kWbTRiYy9/WHa3v5gzCYms7Z/e4PVRxfs9Plpe+u30oP3yjY6NGFtjRk7tJd0QrIPGW1pTvVx8GXCam9NA2WpJrm/qFjhvYO33SAtfK2kUNRDwPALasaqhUA/TKXRbGpgWaMKMaqaDrLKGl+lzLyF0iIz43NqUbEI+aj99rhazXrqV57vsZZkyC6cm7CfvzVqf/PrYTv26XUbEJmhQrzYP2s/e3vAfvbemF24NG7x0Lx1tabcPCW+ANHWXLKxyYz9/INp+/D4kHOYFO6dCvRBFfjjBuwYAoGCl4pZNSwyrmJx4/5hmh7MQQzKlF8YBFrl1mzTv7Dc0vPP1ouwhmxs5Kq9994N+9VbEzKOYTt+9pp7kcKRS+lNVozzi3MF+6tfzNk7H83ZpIS1q6/FjuwztexbLN4Qsf7RcfvZO/P22ZdTFtf9i2rITNEiVxrR0rwqmaLIk+tAUb5CburJgQNJa2gqiUxN2U9/PWr/589H7WPp0Llzg65SA2HlXrSLrK8annxAGOk9YZ5yRiRAZSmnimzDhYwl5aBKarBQZlQ+XRCKnkYbGpu0X70za+8fm7EZ2TLzWweHs5Yr8mZr0TLM51FlyFfPYgXZfWFeTk2JOKeK3rCOr+6lf3EVNA0ksLz8CTVsAGAzkmWTyMnB3Q2WKo/Yjh6mBwXy7GpLuRU6MnNZe/vdgr3z24ItlNRYiWREYtJuFCRZx1BjyC5dGrDhkbwbkYmWggpX7SPr6jR7bkez6oaQm9r1lz+7YR8eG7ZZtXZ5ASsq/960qVH6uGAffThtP/vFlE1OLagRnbLJEfkBuaFOEZz2uqINDc/YL98eEvkaVEVrdvhAuzXXiRSd6Ldfvz2pc6P27jvX7drFtPR66bOwQDWV/qoR9RhrX7X9LA8eTIWKMh1vNmO/eH/MvjzFy271tn9/ox06mLS4BMuLc/t2tqsMM/bxZzfsF2/N2s9/M676eNTGh5UeTlemWyok7dyFov3i7by9/fG8m6KxSxcf2B1ngMXq24Lh5qHhtA3LXWzZararPW6j4xH7q9/k7ee/nbZ0Ome9rQnbuiWoe5rka3ZtjVlDIm4p6UffVpao5FzgwG/3bBsBQZ5Krpc/ozqrvr7gGgCpSMi29cqPtmds92Y6d7CruHW00gMqe8jJDg5EbM+2bpuaKNtX5+dsQQ2rI4c22ddeqHMruEQjWWuokw1FlbiUF16VEDltSolYMpKlm8fjRUsl5MMT4mequ5nOte9AjxvhditBDBWtkAvbsyKkX3t+s3tv6My5YNWcrZ0N9q2XN0keeggJo6PRrC4yq/SCjgzshbo+GPnw1nP/pbC4ziprfxXFSphjWhYL5NOrEzMR91bagd2waRf/FpBZV7fo+nwxbBPz09bSoILf2mwdDVGbT8+rlVCQIOqssU6xlc7sbNaaWgvWu6VZxGLeyrkFEdKydfaWraElYeNqPbF+287N9U6Avb0xGUtWlV3OkdAdm+ptT1/SzaNhym9E7auIKsD2poIcUbO9eLjNDU3z1ltLm5xkYUxKsGCdXWXburnFfToyyHQ1bn3AVR75KQNfq5ADkk8olbIq07Dt3tnohuKCAgp6QpjXyNeuYpF527ypyXowNp1zw33hiMhqk2Ss9p7k1Ca96GrD4Fqsp0MVgdI/dyVk0zMZtSzzqkiy1t2St6OHmuxrz9VbowySFmRMhlYfz1ibtrzIcUiONF5Oq/UZkrFHLDtT1H3L1tOmayTjmDLIlIBmGebm7nprkMHWqRXaqhZlX3dKjrjTUqrQmI7CYLVy6v49fcAWoIgh2X/IzQdOpHKyFZarCRzOzFxaDb8627m9XvaM1BVHjrW1JSXfMC+5zNnmzrgdOdCilnnGettCKmMalEVLxAq2tSNs29qTsnXmKSrt7ohtUXzmT6VzfGowY9u3NdkmhprcALB3gzVsRNCD4iQUDVs2N2M7d7bY1jaWoFEFLF/e3NhgITVukqoge3pU4R5Oyf7mrLdpQXKvtzq1XLKZMevtitm+rS2WkEIVlE6b9GRHT737SAAvebD6RFwVY5Mc+qauhHV3Fm3Htqj1qW7YpnohqsZwXWTOWhsKIlUd1tFcsDbVNzt7eaESn1F0X9JKqiLf1Ba3XVvqrKtLxKtXpIf50grNqbBtFxE6sK/ZGpJq4lLr8mwV/fM1xZOojTwbDdRCPqd6MW897Qk1QqN2aE+zvfJsSvYcsVb5SMAoVc/WsDW3qIESVTkVZ0S4QrZvX8r6ehPufZQrA6ytOmKtbTH5izmFjO3Z1WAvHKp3vYRJFSJzL8vFSdvUGZOPiFt7s/x3t4iP6ujmRM5ao3k7vK/eXnyhya3tCtC3URG38+dnVBfE7PVnxSdExhJusuTj4LcDLWJt0e1bg7ftt22KqB4tWWM9qxfJ9/VELSo7gLvs3BpXubWK05gafmbbd0Zt146E7d2TtFeOxO3wTvlfxWMAoFV+9dAzLbZjq2QQKqhODFt3V9j2KO72LdR7sslESL/b7MgzjdLxoDugU/Xmgb2SgRqazx1pUF1ccHbZ3RWxHX1RO7Azbq88H7Nn9wfTLcu5IK99WyXPI93WIiOlc4GmXIAlGQRdHB7BceoV1lltaWkXR4TPycuvY+mqUKny+h89HblC3gpFtVWyBZufX7AL/XG7OtFkf/Rm2M1XWYLPHNxZudXPsshLORRx8xxg9rJ/KXPJivJe4iJuIi6dcfgBet2KKmSc3cBkyT4/Nq5WUqe1yjllMmaffjmr1nfR/svXW+zgtqBlxdtmzB5gyJi0qCz5YoIjSTrOnIugZ1f31jE+wydNMBYuz3I/nXdzZXQ+Rg0pUDS3w+3PPh3grVB6KJmb4uYTqVBwOAzAgJIlJN9gcIGBP2RM7zlDcCgsa7CVJDEVv/sAAB2vgPX2GP4LSYaZhaL99W+KNjAwa19/pd327VHZ6xyL+YsX6b7KgWTp5rApI6TDOTcMTWNSioQzC760xn0D/UAvAVMU0CF6BF189EPHnJ9TWsxFCYWZnUvP6tMo9aCg+EsPEmsb4wSRpXtBW/t8MIXyiisUFYGPLQDkiv1h0zQCic7Qbr32w9KRvDQB3UEfvExYroa0eMGCRnFZN8OuAelznaSpxO5RFsi3hgcCdMXJWn+QN75VnLByNK16pN7ZHD1JzmELnGE+KmJB9kxRZ8uoIv6ANJlYxIuR6Ar2yvQS7Bz9cNON9ButQD9JhzjuJWcurvh3OpN4wcfdUEjrPDqKe6BeInHyxkohBJcfZZ4tU1LwNSuRHy59EuHKmR0e2RWq/quAaMSHmBKhf8TBH1BvU3bInGPIAVfgbFsn3j5m9snn1+y5l7bZS4cDuVLu7g10pY285/Ub1uH8ifZjupDevpIqg7zS5X2DkK5zL3aKB8znIu6lqpOXsnb10qTt395kP3g1IQLGChJE0gWPgXSoS4OGkMgieqb9IOfBWreuXtIBypdyChAocU4xqaGoUcPliOpg1YeFjJtKhRysnBDxI+2lkpBnVdkHx1w9rl1+s8+UD1bWcXLQBW51gNCcuJsKPiSfrbiK4oBPJhVuE+IZ/FpnlQ3b4M0E7sz2Zrg6VWmx+tNP3zpnfdv3iCBPWn0ybHG1gCJStFg05uzO53clLBbJ3SC4WJlT+hQ7tLROpZ2k4mdagP6xpHxSRc05ypJrEvImkBDKYm6SL5mU7Yszs/beu2m3furC7IJbW8993UKRYmFVeao1abmlYqJEagFG5PWCwuet4pxrPbD+X73iRXQRE4Lp1YsqH3VqPTcmlYbSiTlS4sv51oKtwYOyqQQ5jJhaIO4tcBVcSWUOgUXiyBATIvD2bVLyj7mznOF6+lehLByX/FQTJSWLSGRe6UFvFaJpyTHjphhQmdAr2pTIqvJjLk4xeNtQFolTo3eUhcijcnnJ8LzTB0gqw1ToQkwheOeSfGsj4Czj0kEmrSelJ4lYTvGZDxs4ZLa84BXo89MEL+MAPD+NPCdn7TviqIMUo/zKotOKU9hOvmUn65SuqY9JRu53oUJUSVeysazVq+yD/nddQzqquaKuwSiZSngkR9oBUSXdICfuZvcSarjP8PqC5QcegE6MZvyykxtAzgw/yt4kZ94Gp05IhRasMZSRrWL7MET88YKrHyA8wdI36BKvUjF1bEGVl3yK0mfYlxe64vLddbo2GVlQnbIgn6KtfDo9Uf4++B9HVMmmU9GM1UULrn7A9qWILu/oJb2qjcmyKk0aU/gv+TEYlLvUP8+TikCObrF6+fe45EGXAj4c/kevXVBfB77UNe4lJ/wD/rw+pjJNqB5X2eKXA2oUgCKs0/WuzOW7KfaYZIkfhygldX2d7knjBlFBwCK6f0x5qVe9HkvIX8g/uOlIuv/MmNnnvx20qxfnrbcrYi8erXc94J6coY0bG0FZh/Q8DiJ7wcgRdQ55D8rG7elPQFQr/tI5spCTTUJNuaT0NqmyA5A86llWcIipLHCpBO7mznu2KUCMGYHmXQRGz+iWYfqV40Qqd+ahOqLqsKCzsj/ZbCgUvAiHRaIJ5N2xY3quSL5yC3pE8d3B9vY9pHeL+yNlOQjXvVXmbbGwa0VxjFbYUtFJe91+hUio0Oht2bEzZL/zjS57/fmEHd6Ttef35e13v9dhv/OqnCBDAHIoPDfElHKifRGUUDC510GFF1GhRnRP13KRU3PXyUBQkIiu4c1jREu4Pw/99IAGA8PpUTXDghmqUZUm8iQE8sToMAAfAgRHcUnOKCtGJtNy53xIqCW3fXvCdu6MWnOdyBL3Un3mXripyJ5APgCtOCdF9zunBNkh0IfnTMrl0l2naG7f3Yv8Arb03wRnQNDirYEyZsET53gq5Y09u+DO0VChvxz4Asb2KT9CYOfMJSuqjEUf9JvqCs3hXzDB3weO8NdxVydDdkhjZXkEsdcXangwCKyXEkZuUFf+0Syp0xbaQ4xAAtiyFZIVCXt64ZqqwdvLEjdaF2hfXFqGzgiubmHNgSXbdvrB5EfSpHb3JKCSKhN6GE1xkyi5Xsf5yR7a63TXLW6Ob9KNuZ6XKghPGei9LMJEKcuS/LLKE/LIeBgBWQQSDMrf9wUGNupDULo0/Lu7y3bgQJ2bv4yXj4rUhEWESirbUiinkJUuSB4q8sCXEOjhS+kc2hR4BPxHXjmBCvM1s61bYvbGq6324ze73JC0qTFEmoEferywWKdVfvs9ngWfGGghZapIFWBPQVknFQcb4ZqoYiQUaC6o3MIqk3DWXR/UtAFb4n2Qohuz0G+VcU5JO/sAyJ2N0mT5QNcqEYKOn6Au91GRTN7ZFEcU4Fkiu/5uDxp3OQ2gGt7xAO0pNQaFAcpH0UMQcWJB3IqIJC23/IF2yQBugzW98C0M71P3hUu0sLQj5Q7SIh1a3uxRRLzHTZqBwrv7LyK4T/WRIH8KujgwiWrcHBPceuRpQlA+tIQDqK1Ei0ryoKw5SqvOyxoa61tTQbn564JjAYiLpDkrJfeGospiTtbDl0noVXXLkbAYfUTSdZVNAJwcv11LXMGfKoYhz9w3r4CGBen6nATweuqPVcu/Ot7ThOoyCFAtbwdHQiVFL0ZXoYecFeIqXcErGZyfs1kXiXSRk4tRVbpIjnPBUX4BpxHOLlfSmZtxa47vjKX713BvWK30kRd+gfPeN3gE17jhQzVsWLqM85GKT6+oDzEqniH4zX61Lnkvgw9YLlEq/qVhe+yc/j+OaJ9VLRhT1lHS8HnkiBs+FUquVeaJqvcd1WkGeLL0KHj4QBoiJdq44V09JGXE0QCUJmVK+VCGlAklsVQ2wR4eIWJzLEKvX0wJCqZ8MM2nurZFJkEZ05HEceoO5zz024/IBffnri5DwXQS6gZFK4s1QKGDhtHSWi4bF0tPT05v/gWCp/Wv9nFs+fME11AIlN1SCkF9F5QXXplzdBIEZcI+KXEcHob8GMuIqCiXOiNuzp/gKlYFyYmpnYFEgmPsx5EfO46oetya6yUp6rh273UawH0hqwGWF+/6wNU+pdWGYKqdR3WM1R8vgMuZv0CRg/v4u3ncmv97e6InG5ReUNWwT7vs/paWJ021Hs+NhvVY3p2w/rQCvath4+F+6sWDxXLPv1bU9O52uFX+Sw3fu/PjyAku4Ba8p9NK8ASrVi88ClTL+M6432S1Zn81bEjgjGoOqYYaaqjh8YT34ffqxz1RreHpxn0gqyTx8Dkv6uvDmrCuyDXcCYHEg0K9372qNWxkeEO6H8a0/rRqmrZRsX5Z1vAk4f7L3zGL8s3hSQW90D6sBUufA356sKF8v88Mw/0r/auhhho2PtbjdB8UNkIeaqihhntDde9sdXgc4X3SSn5pvc/2uJbBvaDWUbFOrKRsKylfDTU8rXgY9lCzuRpqqOFxgyeZxWLRhWo/Rm/pWv0a6Sy//knHIln1D323Dy8KZ4VC3mZmZ2xqesqmZ/hkWtrt+8C5fD5nmUzGxV8JHJ+fn7PJycnFeP5fNap/r7YP+F2oUgL2fRf6fHrO5adUZipwcJ4thbL4W9etpBTZbDbIn44v75Ln2PL4Hqsd38jwz4M8KYtqcDyfl9xnZtZlbLlczpXhvQ5n+Pstv+9Kx1eS0+2wsLBgU1NTNjfHd5BvjcszjI2NLT4/2/WUwUYBeUaXyT95Ry48G+AYv5E9gfPEJ4yPj7vnXgmcm56efqBl4dNm68NyLD9WHc/rg5eZP169rT5eQwDKA5vAR6Mn1eVDWeIjRkZGnL4sx0pluZYyRg+5n/fFPv7y626Xjj+Hz5+YGF+UP1CK7l91vePhz1WnzbXot7cZAnnDH1SnW72/EUG+8XM8C74Oe0a25Ns/F7iprHSM38vrgpVAWqRdjep0q0GalN9ynVpt/3GDz/v8/LxdunTJ+Uh/jDI6c+aMzc7Out/gds9KWV2+fNldxzXIkPgrXeOPrXZ+rYAlrSfcbyx9bpUbiLQ5pSxCQNb2uVUP3uIaHBqy9z/4wAZu3HBKl5GzOnv2rJ04ftz6r10TYZ22YqFo169ft67OTvepLQ+f/OTklL319tsuDZxdY2Oj9ff3W0NDg4sf0j/IJIsYRGhdKL+0MkrFkvtsl9+6ODxVsNiBS98Jisna2oRCYTt58pTSjFmD7hEsr+ULuJIbXUtaXOe3gPsNDg465eru7nbH/LMQh/OFQsFdA3xrCnDex3lcQF4xhi+//NJSqZTV19cHeiLgrDEW5Nze3m4xPlki3O4Zufa4dOLUqVPWKT1IJpOLjo9r/LV+yzlfltXpsk85A86TLufIE1sfjy1xL1686LZ1dXVuC9gSfFwPnOynn37qnmt0dNSam5vdswMfF+f++eef29atW6VLJ10cdJv0iLs8zY0KyCjyQA7xeNyOHTvmKi10G4d4Q7ZIOfO8PBO/sccvvvjCVezEq35W5EC5kEZHR8ei7axUzmC143eCv44t+YtGo4t6yXGfZnXaxAUcI67XG45X6xDw6RP8fg2B3r/33nt29epVZ5utra3O//mypRzflg9ni31XgzLkGl+u7FNxY5PEJwDS8j6UeNwTfUPX8DFeFmy9/Pw1oDodH9fLe0z2fPKrk7a9r0/1RdGi+G75+mK+YKflk3SRtemZuKqsayKqK1jAyacD0G10HJsfUr1HnrCjTz75xLZt2+Z00aP6uo0G8nZNdfOJEyecD8O28fVtbW3uGXz5svXPwZbnh2zh56rrcVBd5hcuXLCBgQHbsmWLkzWB+F5m1TLG534g/oDPWO6jQfX9q+/xuMDnmTri17/+tSvDHTt2uHMfffSRffXVV9bb22tNTU2O0MJ/KCsC+/haAuVFHQfHQkbUUxB8ygwQl8A54lI/c55r8O93W27SgLv653Gvn1u9b2SV00NqTWczWXvjjW/Y1m1bbfPmzbZ7124Zw1XbvWePHT161IZE8nA8fX3bblJyn/zZc2dtQkbwox/9yOKJhHOIH374YaDQOg9JHBkdcecycg4XL110pLipqdHmJGAIyY2BG5ZIJG0ht+DuPTo2ammdwzDOiHzgWKiYFxaylkgmbFT5Hh0fc06nob7BorGoU4RLFy66ljZChqBRmF7QGODw8LDt3bvX5Q1HdeXKFXc8obwRn96F8+fPu3SJQ9lCflAeiN3jAO+oUH5PLnFQyAUnRznyvDghnolyczLSs3OcsBwYHIQIo8U4cfgY7rlz59wxWos0UigrGioTExNOV3COVGzscz8aCzhW7kEeiUueSIvfxMdZ4iR4jrfeesudQzYYMkSM/HKv6soFcE/u/93vftc5T64hb+guW56be7C/a9cul2f0hEoaXeCZfJyNDuwCR4lcyfO7777rKo7du3e7MgbInTKnjD2xpXx5VhwhMuWZAeVNPMraV0aUDfFxqJQhgbLnGGVIOVHGlCM2g90hF/SIeDQYvNwA8iFwLXnFwaOjyJtryR95ovLFTkmf49goukBa2Cr3J5Au57iGc6SFL+HepI3OPQ6yfBjAD1Dm3/nOd5zMKTvK1usK9kjZQXhaWlqcnSEDZMI55MYx/CC+gkYw8P6Da9Eb0iZd4uJD2adhiJ5yHh9EGt6/InvkBnlGJzlPQOe4N+cgvOge+cd3kxd8iq/s0RfqB6/v3BvZo+foInHIBzrkdRF74TzXcT+25IlnJa8bHZQXtobvfOWVV2zfvn3umZENx9jnObBByoNnxVcjbxqtnKP8vX2wpaw4T1nzm0YG9kZ6AHmQPuVPWUGiKDdkCG9gHz+APlDm+AVk5PPBvvf7jws8WaVMKAfqLp4VPfvss89cmWAD+KD333/fNR4oM8r4448/do0jjoGenh47ffq0i+s7VCgn/DgBOVFWHPvZz37mbAydxh4fVZk9ELJaLpQcWR2fDttkJm4H94TvSFbp4RwVibx0KTBuBNAgpx9SRlBsCqmnu0cV2LTNp+dtm5xOOELWgoTp1aT9NC+HQqWBcmJEM3JG9Mr2SjhRKfCvfvlLrMEaGxqdk4CEnpNSl6QIVIg4RQQHAa2XoH4qQZEXHOAlCZA4g1KAvu19TvAF/T6uVuWM7ofhjIm0Nqty+kgEmV5ghNzV1eWMrbo1h4HhKPvUOsdp0bqh4uRepMux3/72t+4aHBjPjxKhqBgmJA1wvhob0fjIE86B50PhaYXj1DmGwVCpUAlQTjQscOaejNDLgDJWlx1OCvlu2rTJGSMGS48E1/GbiojKhHikwb24BwZMWhgh8TFU8oTRE8gD6eIkSZO0OE7lijPGESIXyCctWdLmHsiRSpD8eXJOHrgvuogeQlj+7u/+zl3DOdLiGPrGMyJ/fuOkcbSUBU5mOTaifClTX97kHdlRIaGzPN8eNTQ5xnnKh3LBHig3yp9jlDXlwPVe533FhN5ThjhVCAz+AFlThjhV4iMzLxdkyLXoF/elFw/5+9EV7kdFhgPnN2ng6CHHyIf7UM48A40HHDbxyD/PiUxJCx3iGDLjvlxDWuSRPOMPvCzZgmo9flqB/nuCR1lRpvhiygXZ4t/wwdgc+oM8uAYfDMGg0YhsuB65oENU0uzTiOUctozc0BHkA9FE3yCYxMPP/tVf/ZWTEffAxpER8kPXaFBxnHjso8tU9vgM9IueK3w3PVzcj+uBb/yi3+yTH7Y8D4QCn4XPQwfxgzRsuA9bdAp9Ik/cBx2nRxH72qhAn3kWAnLwjQHkyLMjI2SAfKjPkCPlTPlhm9gswRN3QN2Ir+YYaSIP7uPLEn3AV2JjlJ3349yDLelh29wDWeFLiIPfQafwB1zv031c7NHnlcYxuux9KPpC+aND6B62hb4+//zzTtcob8rpmWeecfLBx0FW0Vl6ZvF7NDAoG/T/W9/6lit7/Bw8A1m+/PLLt/T4P2zcK1ld1Yr88PlaEa0YZKlUtEKx4IZWGGrXX5cZAnFwDBBEyC33QEigKCNgCB8S+41vftMZySdS7o72Dgmm1/bQg6mHQcCvvPyKW3ttaHBIhlVnKVVU83KYEFgIKkqBA83l8m6Y56UXX3Rp0AP2koTGPbkfhsN9CEeOHrWDBw868svxfD7oEUJ5cMikSfCgYDFcChvgzA4dOmSvvvqqqzj5jRKhaBi/72XgXvzm2scB3hGwpdzIP07j8OHDzgA4hrFRZjwTxkAP+oEDBxbL11/vy49KCz3gWhwbxou8KG8cPkST66hoKD8IMmVKoOXvW9r8pmKkIuM38TF4DNlXfqRJA4V0OQ7xYh/5kFfS4N7E8SCfOEPuhePEyfLM5AUdeeGFF1wFyj25B4FnJ59chwPBSXDvxwHkE7tCFjhCyBnlQu8oQHeRFfF4LsoWW0KPkTN6QJkhUy9jtsRBF9ADyhqZkiYVuk9z//79rjyJ43WBckbGyIXjEIRnn33W7VNhQlSpvCCw2Bv5wdawVcC9eA7yQ36JwxZd4RnQF84Rj/ujy1zr0/JEi/xyvFqO3h6eZlBhUlaU8a9+9StHJiAWvpyQG7ZIGdOI4ThljY1wDbb2ta99zV566SVnT+gbdonsOY/8KWfsC1lRaT/33HNO3hz39sY++sW16BPpITfS4L5U4OgmckTnSIv7cg1+gvtxD/SYe3IeeJ+O7NE9CIAnoPgutugnz8d9yRejENg/Ovb66687neK+Pq2NDPKIjLBH9ilXSCM2zG9k6X0xz055US74aWyc4x5cQ7lSPugIsgGUG2lwHNmhI8TFro8cOeKOEQAkGBKHPJEV9ojPRlYQMOSEXyaf1ba50UF+eWYC9oDOoFuQSRphXu9p5OCPIeOUF+VGfOwE/aacvKzQN3w1uop/JA2uo3ywM9LzdR9xue5xxX2VdFjsuE3OfufOnSqwNneMovECKrq+06AXFsKay0NIRBpzCxaRMChYzvWKWFIJ5mXsCCBfyLvpBfSsQoBTdcHQCgLa1rdNzuHrIoqH7eTJr9xUgE2+YoqErU5OqV7CgtxCbLnHctCDisAhwwiTPKTTQcuR1gzC9+SpGigQTppzACMi+HgoDvnHoHFqVIw4UFr4ODqupVweB/AcOAYCZYsh8NzV+ff7lDHx+M0+jskfZ5+Kw5/H8dFqpDw4B1mgN4L7ebKPM8NoMU7kxDFkT7ni7JARZYkjRX5UblRW3Aenyn0B54gDuC+VLtcTkBmyI1/EwyFQoVGZIn8csL+eLfHIB8/Avnf4gPtxPQ6XMuK3DxsR5Ms7f3pRcIqUDaQQWVNWXrY8J8/m7FI6wHmOLX824iMjAnEIpMtIBWXN/Xw5U6aUE/ZEfO6NDKnwOYbtsOUa5IF+EI9rkRl5IS160JAJxNSTUeKRPypfSBUVbPXzkC/icB3xCeQPnaIygRxDfsijv+ZpB3JHJvhoGhjIARuhvN544w23xRZ8eSIPyvO1115bJJzA6xFxvS5xfvv27fb9739/cZ6jtzviextji30iS+7BPnJm32/RKWyQNDjm9YR7sSUfVOTch3x7YuXBPdBFSBK+iPP0Znmd4XqvEz6f+CcarNXHHkdQXpBvSBRyw/Ygj5QXo0j+2XlGypMy4jfHsSVvz758KC98OA15ev4gT5QNMiUuZc09ic91yAobJm2upV7gevwH+gFp83r0uMCXmS8Lypd9nhu/xHOjW5QNvfb4M/SXOox4nCcQvxqUNySetPGH6CgEGD3EPvx9vSweV6w+DSAXzFmdyiTWNA2Az7RSUFR2vGlJa6itvc0V5MDggHNmrS2tNjk1aZcvXbLrN4JhZF6kYh5rb+8mRy7PnD5jX311QmRjUIbRKIK3z26oopmR0BAwPZ8oK0Yzp2MjIjrDI8FcIgxnWgSTVnBSQkLow0PDtnvPbne8SY4QJZ9QywVCy9zYRgmUXlmcEqKEMLU0t7iueKYx8BykyxAQz4KTBjgvP1eElhDp8uz89vEgTOQL0ouSkSZ5806ZoQ7yiCJ6bEQDxGDIK8B5kUeelaFangMHRLkxzEBcWoWAYzgbhtqowNAtygo9oYeDtHwPB2WEEaIT7OPESItygywQj3PICQMkDVqlVGKUN/dycldlAUklPteTFvIjj1xPDym6yDFkRRzuS2B4ibwjD+7JMDEEh+toaDC9g3sgV+RGXrieyo50cSrkC8dBvnlZgPzgaDcykAvPz7NQSTBCgH3x7FTQ5J9yxNmxz7MiV09SGDGgpxmHy/MCypYywA6QBw1YbAFHSoONIVjIMPegooPQQH7Yp4wJyIH7UJFRtqSHnSFv7odD9pUkZIg88hudQI7kBTlzLTLkXr4CRHbYMM/DfbiWZ0U/cfI8I3KmXHgGKg4qF+69HI9bpXmvgDD4IVzsGtkib8oTu0EfkB8Eh30aCsgF3cJHYO/4QogucqKciY/+IAf00Dc0kR96wjXoAXqDvLwtY3vIFF3kfsiPLXGQN2kjP3SaBi1yJC/4B3SG+3o/gG6RLue8n0eHeC7S41mxbRrJ3IP841d4Lp4H/cT+eUbyzXU8Ezq0kYH+kl+eh/LkmbEpyg+/ij1RJsic58MOkAP2x7NTrjw/jVHsAzkhP84jT8qKMkMWlD3lQlkzyuJ7rSlv0iQPyJ10OY+dc5wyJR/ch55r/JPP++MEfK3jWCpfdB+fiC/CTngW31jj+YlDHUk8dIjyRQ8JnuxzDn+F/uPLKDvskvTwp8iDdLgW/WT/UZXZvU4DCOmBHd0u6l++WLC8HrqULYjALdiF63G7MtFof/+HUUvc5vkgugBDRzkZYqentLOj0w39z87PuV7RVDxh81K4cZFZejEpdOaLbt60WQ5vhwqTuTBzchwsnVF0b+nzwtPs3KxzSE2NTa4XFhIZ3C+v+026fXp06a1NK31A2nxvNpvNOGHS4wrJ4b68mFWnfdKkgEiH+HyjlmOQWKYg7N6x0xks5JJ4fpgQgaMcOEaMHIXA+HCogH2IHE4dR45TZYtjw7AxfIyZsnrxxRdd+hsVPB/PjkMnnxgBlT3PznGeFYXDiWF4lB+OhDgcY/iH58TBeGOhYoDkoKTEw0G98847i0ZK2SAzzhOPNL73ve85J0rZkQ7XUTEC7secOeKTPyq3b3/72y4+Bu2Ps4+zJD55Zp9jOAjkzDMxbMg5nptnRqe5lmv+83/+z87pQqzQA9L1jtxXdvz2FTPOmt4I4m5kIBPAM/C8lDlb5IccKG+eB53nWX0vM3pPGRN4Zu9ISY9r2XLO6wbyQs6/8zu/48rzl7/8pavMfQXJvYhLZUb6VGDIGXmQLuWNznGO0Qn0BvvhHDLyZU98dJVAnpELeUGvAPE5xn3IB2CfSpXKmefieQnE5T48s78XukFePMjT0wJvv9g5oBwpZ8oN30a5oD/IkbLlN3bl/QIyZp/4lCXH2KeskSFlS1l7feA31yNT/DfHvR/hOp8eekJa3Bf5IRP8jvdR3Iv4HPPp+PgcI038APngHNdxT8Dz8pvriEt63IP00GvvV7gH+sZzsCUtriHuRgflx/Ng38iTZ8ImKFvKhnLm2Xku7JJnhWQSDz9H2dEI5VmRmfclpEVcAulxD44Rj/cbuAY5EChvbJgyBaTv782W3kYaPvS689vjcbI/9PZR5Zd7g0dyf916fMbsp2+ds77te6yvZ9Lqk2GLx+KOd8HVGP1m2uhquK9k1S0bFQnmqXowCIJCAxyX68FVPAqMzEEMGaYn074CIL7PuB9E4cxCPudIb/V9qgdZfD6q4fOSp4LRNcyT5T7cgy3pkoavetjPLgQT7mengpe8IFG+AsRI3HNUCZ60OM8xDNQ7Ml4SwLghK7Tiud6D1g+OASPdyFiu4E42VfuA8lhuAFzHs+PEcT44bbBaPAgprWgAIeQYrW7i06om+MoDePn5LfdBZlSYECDKm7T9/fxzkK53qD490sC5ctw7SuIvvx6CBLGiNevBcdJEL3xc0iYfnPNO+HEAFRZleaf8VpedL6PqsiIdynI5aMBBWBniJw49NZQlYSUd8iBtD38vdINyR1e8HgLS8fmrli/Hq8Ex0iGOP4/tkr6/zl+DbmG7XoeJX53X1fL9JIIy8+XF1qO6rClHCJ+H1w0f35crx9EDznl9QXagOj3OE5fr2Pp9wPXe9iA63Jfz1ai+n88Hx7gXW47586D6d3Va/lqfN8Bv4lTn1+fTw6f3uGCteaf8OE9jwhP46viUBXF8mVBGnCcgK0bn6K32duVBfGyb+FzHPunQq40/9T24HrfLYw1L8HJ9JOWlW28IsurBYsvMWwUUCL2qzFP1pBIyyT4L8/tznniyKkA12YRU+swTj99cw7Uo8M1VjxS8svVA4bmef8DH9/fEELwRefi96rxiJD4uz4TACf44AYfJlmMYEeeJ6+O4NHU9x308tsvv/zjBPw/59+WyfAvYry47UP3MPj5YXh789uSSLed8eXIdlSIt9mr4NHy6vjID7EM6OEc80vLx/HXVld+dwDXAy5Nr/HWk6/O+0UFegS8Hfns5+Ofx5ePhyw34cvDnq89Rnr4MfDl5e/HnqtMFt7sXIB2/RZ7+N3G4rvp5qrE8HY/l9/M6QHx/fKVn9FgpzScdXpYelIcPvjyqy8rLmnP89mXOb/ZJjzJnvzpdUJ1m9X2r931Dg3t4faiOB7gfafl9nyfOc63Pb7U++Pj+/tWoPld9jdef5cc3MngW/zyAvBOqn4V9X6b8Bjwf+ys9I+n5eNXw5e2Dj+fzsFI5gpXkAvz5xwXVea/GWp7Dl9VyrJamB9d4WfnrV0trJawn7opQ9jYUWZWquSkAEFZPFDlWDY6thuVxwWrxl5uGJ6v+uP/tsdrxaixPcznuRclqWCq/h11eK8ntdnlYr2F6J+DxOOrDenXbV1wrVVLVlYrHckd5J5AfL4fqa6qPr4bl5zypWW8eargV1XqylvL059AXdAI5VKfhz1cfux2Wx+f3Wq+txu3y/LSCclxJNuB2cqqV5d2Bsqwu17WUIyO22B2jsr7hAGis+elWNOC9bRKfqRxMpfCNRX8f4nBfrqVR4NNj+gq26jt2OM85rr9rSG3ulazeiZ+tGVQI9H4y35ThXF5aAiqaxX93wlrirAYe5L49TA0PBNWGspGx3jw+Ds90J6xHNjg5huSYOoEz4zeB3wyZr0RgGbbnnO/lWgt4mYJrluN2+SQ/TDXAB3EvAvnhGHPHuXY9eajhZviy95Uh5b0aOO91g4rONxg8OM/6uswJrgbHCYxMrAYfp4b7C1+utbJ9sPB24/2n33qf5e3G24v3WdiLXwasGvhe3pFhSgbXID+mGv7iF79wy8v5NXKrQRzmh//85z93W/wkW675m7/5G7fqA2QXP8xc4UeN+8bvYnJEfE2KB2fiNAv1w/SZGsAcUCeAIsuG8NmvYBkoXpbinBdIoZB3L0mxTBU9tDIZd454XEPPK7/Zz+gYUwLYMpeV4f3bgTvcrlcV+DhriVtDDU8KcJT0fBFoRbOldc2+P+edI1uCf0EQR4bzxPGxWgJOjWt94FrASxj+hQx/r+UgXe7Jlvvy9jBzf50PqErP54keg+o8+rxDTCHTOGteyiANXpbD0Qc+KUiD+/jrfL78PsGnR1xCDQEoX96qp1wBlSCy8ZUhWy83ZMiqKcxRBJ4E+WuYx0wFSflS3hzjOl7GQZ98T5I/5+VRQw2PG9BjgB7jj7ANfBQNaxrS6Dz6DfHkPHbmXx7E1xLXdxAQIKbYDw16GoOkSw8ppBM7w0dzPauckDbnvf2RFwKdCKy6wPsDpImdYnNcg43zUQLu6+3wUeK+LV3F6dOnz+j6oluzjiUUWB/1hB74gogr66lOVxwQ32bOSQBnVBAUNuup8ibu51984SqZCxcvuEKlq/v4ieN2VccQXktzs41KgJ8eO2bDao2TPmkg1LbWNotXdVOTHx9wbWH9qT5226A/bJeDPK0UatjYeNAye5BpPwxAHHjZAULJqgg4T7Y4MV5owL4gmrxEwUoLOEPiMF8YZwZhxVZxlLwsASEkLRwhTpYXDEmLFyM4xzXYPQ7Wv2CIo+TjC/4aXmzkGrY4aBynd8wcg3iSF+cXWloCv6FnIK9+GIxzLDlHfnH49ODxHJAj3l4mfYgU+cfxkwbHeE7yCdGlQmEIzb8hXkPwRTJ6cZA3b4bzYqOXG7/RHcqUcqbCJD5xkZuvVNEB5MV5yhsd9I0dgKzo2aFBQdl7PYMEs3QSMvVYbn9rCTXcipXKyQeP252r4fbAx1Fe2Ir/Up9f/hJC6H0lK9vgx7AN/CwjEtgC5/FH2Iu/lnj4La7DZliphqkA3AdbwfbwfaxGhA3is6uH+/FxfHCD+2O72Bv3hKxCoP2KLPhUON294IF9wWq9oGeTT5hSiO+8+45NTapg9bAIgjf+eOOf377SosBbVRBtcmBUOlQOOCME2qrKhzfpcYJnVSm1KA4FN0hlIwdImn3bt9sVCZ0C5w3teHzjv8hSQw0bEdgc9kXrGUKB82OL44J4QPA4DwGkF5VWOtcASB9xcZbEgfThjNnnLX9PfrmWVj6E1FdwrN3re+M4x9qBEBGIKI6WexAYwsJZkjb2jr/AMbMGIXnhHtwPh0saXIsT93nDP0CKuZZ84j/Ik68MILs4bZw5PXj4HdJkn5UluHcNS6AMIfBUZOiKb6xQMaIfyIfzxMPPcw5d8JUkdQCkFLlAXtEztlSG6AMVsr8H19EAojODY5BW30tbQw2PE/B7+DN8Hf4IksiWJS1ZFg8ySgMP3+U/ikKDDt+GfXCM1U9IB7vDL5Mevsr7ZPwsNoTvIg7kFG6FD4ME0/D34F74RGwM4Be5D/bM53Lxnaxxu1Fw38gqQ/x8eerNN990nzz95a9+6ZwSFQHEctvmLW7ZKRbA3bVjpytYPgZw9do1V2Bz83Pus6n7+MqTAl+V4nxBWxwbHx1wDFwVCmn0qlLaKcFRwdIi5/Oo7k1/8lIJK0L3raGGGm4FjokWOE4KB4oTo1WNc/VD5m7kpWJDtOAhGDhc30OKo8RBQvJIC6KCE+Y4JJLGLKQSwgEZIW1A+pBcfAbn8AmAexFwxjhSljAjDRq8fo1WHDDxuR+/uT95JX/sQ6rwHTwfi8NDYv0zkH+u4Tfkl3TJC6QZ30LvBc/sKxofnmZQphBIGgbIkkYEMkGelBPlDWGlMiUeAaJJYwBQeaIXyI/ypucUmVPpUoEiT2SETqGLpOkbFugQsn3aZVDD4wf8CEDfsSHfmOM3NoKe4w9p1KHf+CGOAcgmxz0ZZcu1+C7WncVXY1/EwU9iT9jSN7/5Tecz33rrLZce9wVcDymG1EJi8a/4XuwLfsVa2KRLHjeKz7tvZJVCvTFwwzktv0A/QuBrURTC1ev9bi4qhY7I6BWFydOLSkXC9VQ4fN2K1gRviNGjGotFXWHzmUWG+ik0PrnKigOQ261btzkyy4cGBtUi5+tZKloJNG8DQ0N2nTzNTNt15aGMslSEX8Pa4BUVWdGbRQVO74knEzU8GUDG2KF3qOz73zhSCCfD8Dg1jgMcI84Ox4heABwmDhcSA/n0pJRrsH0cKWQGYslQk1/XFjJCfMiJn1LgHTO6xv3xJ5AYrvFDzBAYnLZ36jhkruN+5I/ryYtPi/x6x0teuI57sSU+1xO4F3nkmfEvHKshACQSmeAT6IzwPaB85Q3ySQMAGQedCEEDh7hsAeWOvji/LTmiL/SYomek53WPa7kOGdC7hF4gp2oZ1lDD4wT0F931xBPfgz2g7xBJ/B+2wzQBRirwkTTqaMAxKsWoA3aAb8I+sCPfGUA6HqSFr8SusFX2OY/P5L7kg0YmHzqig/GVV15Z9OX4V2zR+0O2hEeN+zZnlakGOB8qHFrOfG+9TyQTZ0OFgjAoABxdR2eHCiH46gc9ojioTRII105xLJe3AwcPOIEgpFFdT0FyXVhEtax84hz5ZOuVq1esWy38rWphMI8KgbR3dDhi2q/84AgRzOjwiJvPwfIId0SNzy4CfUB2EFVkgRFAVqmQqo2jhscXOCScFE4SG8UJ0rOI7DmHHWLD2Csy56szxMeWiYcDxAFDWnCk2D/x2dLyp5eW+Azl8pvhKRwleoUjBqRFTxz2io7xIQ5snn38CrbLOXwEvgWSQzpcB0kizzhkegL8sJfv+eM38cg7foN0eFYcMI6ZfJM/el55XgLPQ08f52koQ17xIx7V+08b0BH8OhXlkSNHXG8MZBL/AFmFXHIeuSBfZEVc5EHljI4gfzoxKF/iUAfgYyhXdIoOCuIQOM89kBHXMq0MedRQw+ME/B1+hcY1/gwfhd4T0Gf8I7qPneCn2EJWOY9dcQwfxdY3CvFd+Dy2HMMHYkMErof80lOKD8OnYj++cU5c/CA26PPkA/HIL3FJl8D+vWBDfG4V+C5a3swvSSB8WpV2dFDh8TWKqFvUP1dgod+YKpelDwhAPiGtdFUfPnLYtmze4o4DT5aoQAAPB7FmmSwEzwOCsfEx97LViy++ZKm64IsYxAFuzVcRcCdEpXVHPL310KpADq78FNgHyKWGxx/I09sSYfk+oOWOs/JbD+Jgmz4u9vnBBx8458ucLNLmGMTTjarot7dl7zS5juNen6rvz1DV3/7t37ohetKABP3hH/7hYr6IA6rTB/56Tz4hqlxTrbP+uTm2vGcC+LSq8+Xh7/u0gTIg+Odn35cxcvaVj5eHj0tZeplUyxr4cvZpcR1gn3jV96JR4s/XUMPjCG8r1XqPjq/mZ6rtZSXb8b+xjWof5tPFFr2v9bYEfD6Ajwuq0/THq+PeFfRIG+qjACsBYrkcnmjy8Jz3X7GiB4UXrurE6tcD0oMQ883/hvqGxd5TicZtAasBrBl3+aw11PA4oto5rhXesS2/luMMVeE06SVYCSvdzzvFavAbx0mvHb26kBRa/fSS+vPL4dO+Xf5quDespUzXW+41OdXwtABdf+r0W+a9YT4KsF54YUEoIaqgb8tWS6XqFntE1wr3kJGo1dfVLxLVGmqo4dGA4SeGmNaD1QgPx5k6wDAWvasMV4HVnD3HVztXQw011PCoUfNPd4dHQlYhlNWkEjZNYJH/5efWCq7xpBdU96rWUEMNDx6QS4aiqody7weqnftT2StRQw01PPbwoweMFrHvA6jeB9Xnq4/fCT5u9XV+v/rY44gHTlYhjcv/LQczNwjMaa2Ody//aqihhrXB90auJ3gsP87cKMLtnOLya+4U/DXV27WiOp31XlvDylhLma4lTjXWG7+GGh43VPvE5Xpe/Zt4zKH3pNbvV6M6LToHPIjrt3QcgNv54scJ95msetp5c8FuCISUJx9qqKGGJxw4aELN3muooYZHD//SEkSSFUlYPQMSCtnkN6tdsO8JKi9GEfwxtj6wOgarnABILvtcD9iSHiDuk4L7+IKVrxTUGnCVRNBLWl7Gh1dLxse6t6rl5qv5REDZlt5gi7p8CWXdzZNW9pej1rCvoYbbo2JKga3I4Tprx94D5+g+b1zVc3BvWEs6S3Gw7LD7y7HAF9VQQw01bATwwijrqLIEHEu/8fIoa8vTE8pvlqF65513HNFkjj5LV/H1Kr7qyculLOfH56chvy+88IIjtqTH1Cs+esL6xpBg/8UriO7988V3CbnhB/aClZs7qgdkXdO1gwoiCKXSghIPKi6qCwjsav+AvxJUn1vrv5AEEhbvJrBftrzCgkszpwJgCfub2xi5SqihhhrWhcBkqyArK7NGHvSw8k++Ayd5p7A2BB7ktgFD16bIbmX4zJ2qoYYaatgAwCfhm1hPmrVNWX+aKVMs/g8hhaSeO3fOkVbCyy+/7Mgoy/WxhB/rEnMcosrawxBZ4tOrCrElLoQV8sv6rKyhuj4/u7Gxas9qOp2zC/1xuzrRZH/0ZnhdPatUWvkScyaSNji2YIViXHWZziuEykG85RTYM/+7Lthy+KblqUrhkhXDBVVe9KyGrSkZt952s+jiNIDKtrzCQrePuBFSQw0bGtgZQXZCA7AcKsqaIouNQSwL+16LKd+vBn+scvO8bsz7mVh9YNk+NzXUUEMNjw6eOPIVQL4sxVKd9HxCVv2C/fSaQlw/++wz++53v+tI7c9//nM3rE8PKl+aevvtt91C//SyQk5ZAvTMmTP24x//2JFh0uaDKXxAgJ7ae1of9X5BdcEDW2f1zmT15kpAYtBt+BucKehc/9WsTc2mLRxJ6pjIpOW5o4u/HHxCFfCBgLtBuDKcH1IGGPoPx7Sj2mp6rmgjNyatKdFg3/5Gu8XEaEOqXIO8K7inX4b7VIHWUMOTCswUS8XscIVp2dHoeNYuXuiX/QTOcS0Nz/s1PBVTZorlkBVDwSoE27Z12vbNcUdYA89QQw011PDoAJGEcPJVNuaifvLJJ24tan6zhYD6EanPP//cfZ8fssr+b37zG/va175mBw4ccPt84Y0vXjEVgLmvp06dst/7vd9zadGzylQD1rqG3Pq5so8UqgoeMVldQrHIl2BUSSne+ILZ9QHlrJCyPTtjltRhqi9utBoV9cnfuXpbGYE4SkqnpH9RhbKlc1k7fz5tIwMFSyba7OuvxSwW4esMvDFXqcZWuuH9qT9rqOEJRTDMnpdV4z3K4ZJd71+wkaE5272nUzaGYy5bNPIQDUnmnJbfudaft4vnh23nzk577mhCh5WPmkHXUEMNGwD4TXpU6fmEiNKzCnHlE8V8w581qiGhDP3TMwrhhHyeP3/efeaaa/joCmlAbFmDmjiQVD5vzHV8Dh3wWWI+xQruV6fAXUP1xCMiq55ysoUWqjVQUsVF76hSO3l60GKJlG3Z1mJ1cSqLpZ7Mlbgh8Mmvdn5toJudVz2ixstwZ86P2fWhy9bQ0GPhcrO9/mqTxel6Dfv8cEllW41a3VZDDauC5iBWX5RrKciUBofKNjUzatv6Oq2tLpi7imfAjB6WKdGJe/V6zk5fumL5haRt6WmxZ4806f6Ms9QMuoYa7geqO5tWqDk3CDZmLiGqnjQy9xSS6d/uZ+4qPaD0vvreVXph/XrVHF+ph7Q6vk+/Op4/D/z2kUD++ZF9wQpS6QMVglsjVfc5c3bSIqEm27+7RRlBQCHLl6MKYctLh3K6YHlgNTD/upPbXyHOWkK+GJKAo5ZTIpcvz9ro4LT1dDVZJJJRJnPBwzqB+cfWlp/LgxDQ8AcbaqjhcQQrfDBHlfGJqVmR1YEF27a9y5pEVKt1myZhHrt8gGFBIatw+VLaTh//ypobwtZQH7ZIeEE5pO9XJ2uooYa7BFa+VFsx0Q+LImDfBZ1arIP5re3DCtyfrjZyB1ELUO2BNg7IH8SRwD7kFDCf1JNWjkM0PfmEqPr4y8ExH584bKuv9eAYvx8pUb1PuOue1aA4IH0MvaPQcbtyw+z4iXnr6qmzSBzaGSiNn4+6GpbPU71T/NUQLhUsWgrbyLQqzfyYtTdPq7LK2fxCUjlssW++1GPMVPCEdDU8LFW/u6esoYZHi4Ksn7U2zlxSODdh+/o6raHJTA1kN5e1GvfbR5I8L1JiO3mlLXO30UmziaunrLNFreNI3qazSWtrrbOXDjbL1HWMOQI11FDD2kEF72yXtTyxn7gzPkddZXMT02okZrWvYyXiVUL1S873EwyIgjI2r3viZ+LyNzH97u7Sef12dbvzEASAl9g4texy0vkkEMg1Q4/+yOasVhd7QFajduFqzmbm47Z5s+qMio6sRR7LZHjXFRy3VAPFLl7L2uT4FetqnnPGM7/QoIqtR2S1xdw0ug2gv5jTxjGjGmpYO1geKp0r2YdfDFljS681J7MWiRXlTIoWUU3GaAp4EM7YzUDV/enbLYUKNpMJ2ZXBtNWFJ6wxKeOHrGYarLUlJbLaqjzICTrCWkMNNawfwaQfLC4YUTHr78/a2PiYNTe3uBgBdGaVl6fvF1wuyuQhKp4Stv7BnI0M99s/+C+esVRCtakcQ8ixWvgItasPGwM1srohyCo9rWG7MpCxfDllOyGrlXPgTiK5WYQB/DWcq96/HVBmhgZOnp+zwesXrKc1Z2FVnnO5BstZj31LZDUq3S1WEgzy+GAN7FYExlPU0yAYr8BPleLW8FgDi1lYMPvoxIAdfXGTJYPDrv+FRThuwt0Y/20QVFjBPmlnZfAfHBuzRHRK9jTnXvaaTjdZS1OdvXSoycLhhKLVyOpGxWred+NQjKcVgWRYxYeP6/CvKAsfGirYwI0xO3iwx5pk+AF7CGzS9bA+ILj2b+UeRWXt2mDBjp8Zt0IubT98XQ3mhqTyUpS9K54asSCw+xpZ3RDQoz+yOasrIpRTirzg5OqRxcBNbheq4/JU1dfwu3q/Ou7y4OOhpEGgokq5o4Vw2oo6mZPRMYQJWWQ4EwVaHjiHqT7I4FEjqTU8bqBiikiJo+WCa/BFjPmhfAQkqCQW8QBU2/mB6nSVl5DyUS7nVVEFPSs4PRAOq7Ja6QXKGmqo4Q7wtWnQAZSDqI6ZjYxN2IFDPRYXUc2XRWHLqlHFWCGKpXL2gYQys2HxNcpIXuH6eM6+utBv9XUhi4cyVr6FJS/lfSOBur461LA+3INEV6JeIJjPWh2qY64UPIhLTeSvwUjYh1T6/duFW6FU1MoqhZSnULASQVmVF5S3pKZaQSGoZm8O+TIzXcMPKASDFHzlZ2NjJQnVUIMg1YUPMqICcQ2rIovorzuBWvuwFlTHv4uATTMsGKCiq46g6lg5eJO2hhpqWC+WfH9R9nT5Ss4GRVa3bO2yhoQaqDI86tRiOG7FSMjykYj2kw8kFEJxhahlZe/Xrubs3KlL1pCatbpEUV5n5Zq/hicP9zANYEmZWazGTQMYnLW81dnO3ogqryUsu3QRXE21slzdpqan7Oq1qzY0NGRzc3NuLbFnnnnGmniL4w4gzRPnFxanAUTV+ptYqLNMZLO99myT9V8ct+HRcTf3BdDbuhLKld6Ze0XYvXGydBOGVUB9NGvPH91jSdWn3IsO8FWy8ogQyHYJQe5uPrr0bDeffZhPEsyoKle0iJLk8xM1PFjkFkr2yfFr9tzz243pYswFZ1TiQfYYVOuem7Oqv5m82cefjFksNmrxSFoN07DNzrdYU2O9vXyk08W9NU+PQk9r8FiSI3sry+BRSCbwJAFqPiQoC2xseMLsF785bYm6lLXWq1zys+5DP+VwROeLVgrLCCv16eIUnfsOkWHdY34ma11dSYumRiyX67KxgQn7wbf4nn7MvRUfcstSBqBOrdn4BoH04hHOWfWOhtUA+M2c1XmR1XrbtQnVWsJK6sLVPgUyUBSpu3H9uo2PT9r4xLjt37ffUjIOHmI+nbZUMmnNzY2KGWC1nknSPHku78jqpuaCCiBt07k2G8v1WUdLyFrqCra5K2WpCM+bs1zlazvLcX/IaskNl0KMHTmOhmwyHbKzF4ftRv+QvfntZ62rJafCCowqXNJTbRjboiSXI5hg71G2BSuXgnlBPEFIOuSOr/IQ90xkUJTlzpDP6krrKj+UB0j/hinEJxJMk1nIlezTr67bC8/3uTmrq5a45JUvlEUmlwvOY706EXibkuWsUI5ZPh+2Lz8dtFJkzKIRJviELTdfb62tdXbkcKfixYMKVJeVCsqn64StfgGjhoeNJR/CXiAD/hZENlgC0c/nf/jw43eBN1vMKFl5FNl5lKiYKyPs569kbCZdtj19svRyXkURcXNIgxcdHw6YL3t1KG0jNy5aY4OMODJnhUKnjY/N2ndeP2iNjWE1liU//5KX6tugvqnZ+IaAdGWDvWC1RFbX80oDVUd//3W71t9vO3fusM6ObkuK2GVVyUW1BahcNVFaTQXJ14lzRUdWO1sySjxns9kWG55ptkMHO61vc9jQ9WAOdtkKlfRvwX3T8bKMOgSnsrSy89W5GzY2O2/ZbNy+9uJ2624tVIjxRier/PZkFToIYShLX3ASwVmfd+/IluNeuaonHWy5H46UezmKrG2KslaokdUHi9uR1SWtCfTFH8/opxtkWIb16ASy1q2drFk+p6g2CuHElzKsyJSFIwEJzczFrKkxZYefbXTz3KIVXcTTQVbphWeZm0dDiGpYUoOlRoPXkzwyRb6V3w8LzmdVAjrGJqZMsfXweXwq4G1GBQBZncuaPbef9z90qnIOPLie1JvBMnWX+gt249oZa0zpRyhruVK7I6vf/toBa2oSga6R1Y0L6cnjS1a9N5Iunbt02VgYt3dTr8VdL6NO6/xHH33iPiuWzWbtj//4Dy2RWJqDtlolR75OnC3bwOB562xOW3ouoWdqsN6+rbZrp05Kn3FCVGLAVYDLgIGupxJdCbekq+c5eWLWRiYGLNmYcGtDvvbCdutq2ahkdTmoWLzQom5IZnwqYxcuDtnUtMhCOK7n0Hl6kIWVyGr4PvRWhxzjCeYgM0RVDEVtbr5gucyk/V9+/8VgabIaHihuR1ZdG1DnaTTgE9LzZmfPXbfhiTl3Zjn8y1BrBWk6PVSlVCyGLRZpsMy8Gk6heelXYNSFXMI1cpP1OStneDEDvxEVWU3bt75zyGLKMKuCVGPDmt0TCO9FnB27X5R+2GamS3b27CWbmc+pHgnqgYcF3BXD2mWLW3q6bIlYzr75rT3WIH5W6Wd9unTE+289+FmR1fmFkB3ak3R28yjKAb9ysb9o16+erpHVxxHSp8eLrLquDSVUKFl2eNySDc02mJ22M+fP2nPPPueiXLl0yXp7t+j+aTt/7py99tpr9vY779iePbusvb3NZmZmpJhN1tLS6uLj8BYrPDFEWn3Hz2dFVi9bU4MK6EaLNbR0Wt++qCMyzi1qCyeGUC22EqsfCNwHHSdX3HNBXO7K1bTNz/frGWJWKhRtZCZur77YZ93NKnl3r41OVlXx61+pWLZIOGkj02ZnLg67D0D0bW50Dt33fK6GQNPuHaVS0aISIC+rTYr0f3Vi0m70X7V/8AcHrT61nj79Gu4Gq5FV5vzJg1R+xdRQNLt8Ydgamhpt+846HQtQreK3UZdVcTsT8dRH2XPL2KBz2N/p4+N26dIV++GPX7BUk87J+Jk04nG7NGu4v/BklTIvul8lEdWoXbk8bK1tSevd3Ow+MPEwUO2y8CejafmTzyZtbnrKvv71HdbSHNRlzr8R6WmB99UqnPP9CzabLtvRfUn99G9dPFzkddfL10uLZLVsmUWy+p3XD1tjI3V5jaxuWEifNtbSVXcAL2CAfHbBRq7dMJvL2Fx63nbu2Gl1yaRdOHfeRkdH7cSJEzY7O2uHDh/Wg0TthRdesKKI9OnTZ2xsbNzOnz/vXryilefTdFAF5B41UrRcuaTWYN7ysTlr7w5bOZuxQnbWcguTls5M2PzcvAvZuVnLzM5ZRulVh7TO3Wsg/Rk9x8DwpI1OXLdU44II3bzFQlTo3mV7LP+90UDVnhD5Ttr0uNnI9Unbs02NgO5GHeX7JmWm5Lr3slcKKJo7f4/BNThUcyB1ic7Onu+XX8pZUzO9Zxu9DJ9sYH3MES2pep9RpX/x8oA1tKVsq4gq8vcBSIwuVB9fawik78PNYHoKiMkXlPN5S0tHTp6+ZhOZK5ZqTllZukNjdbU57zU8XBTUQJ8Yj4ioDlp3V4dt3dZsMcnI68eDDugTW9RmZjpnZ08OqEE+Zw0NJVUnwTKMTyVuenARVBmNL6/Az64U6P/0I3CrBR9nfcFZu+p6H6qx/HcwurfkLWp4MvBQpRmiq0OKFUtFbNvOdjGcWdu8uc22bd5UiWH6vdlefPFFO3hwv+VyOXv/vfdsaHDQNm3abM8995wdOHDAIpGoMx5UPxaBi/t/qi55Ite6Ch6td0u9yHDYdvelbFdfnfVtSdq2LXW6Z70LWzc36neD9W1eCvzernP3Grbp3tu3N9qWba3W0BSxuvqiKsqs8kjflIJs7BZD28Cgl2p0QgTx9KS1NrVaR0vYEpUeEP/R3dUCEmGQ9l4D6dDPMTFTtnfe/tgy82mra2BKgFhJDY8UkFWq95mM2RefX7O6pnbr2drkeqaQW3Xw8lxJV5aH5dcGNakPy6Fj2FUoZJlczD7//JLNzMxJR4pWCi246ysqW8MjRla+bz4TsosXZ621o9e6NvF2uWQu+VXb/IMM3It7zklnz56+ZMX5BWuWPwmV5ixSLrjRoqeW8siUAnsDfu/mI0shWEth+T41Hf+WQGn6EiXGzfDXrQiRUHde9TtT9QjVYCUQ+n2DbcAP7mdAX9jW8GjwwKcBVJMxht2ttCAvNWqzo+9azjKW7HrDkvV9Vi4l7MyZM27u6pYt2+yjjz6y6elpa21pcb2sC7mMff3rb1g8HrdLly7Z/v37LZFIuF62oM0Hgnx8eX7BrvdftVQkY62tPXbkcLfFXZRqVbvZBS0njcsN4W5Bxdg/UrQvjl+07s6shUt53avOBiYS9uoLu9w0gBJdPcLGfjGoZOMzYfvs82nraa63zs1RKyjbjLRQVg+Nc+s+mTmzG9f6RfzTlhIJyRfrbHh4zN78zmGrT9HPW8ODxGrTABhGHZOJnfpyzqKmRt9mHZBucA79qJ4H7vfvRm+cH6nC4otXAruuItNNz1+ds1D6ktXFsxaNsqh5yN742nPW0YA234yNbHlPGnzZj0thPvp83Frj7dbbK1+pSgPZMfX9gb+4Q/q6V0GC5xv3V28M2ML8sG1picufxG1qNm0vPHvU2pqVL8Uhz0+bjng5XezPuJeDD+9NqQw4Gpzx55fgjwQlFZBUiL9qZTqqtF9U6yASVaO2SKcN8XXMTS2LySdE3Mt1TCXCp7jeEf0vlaOWK0WUj7Jduf6ZNTRGVO+HrJxvtYnRKfvaS89Yc2PMfdmqevrIuqbUkVUCwIlUb4Ws0qLhnFKektoXb7KQFOOeV7h5WqCy3fBzVqtJIPeI8hp+etKmxo5bojlpkeb9lp6L2MD1cZucnLKDBw/b1atX7eqVK/b1N96wpqYGS89n7PzFc3ZZJPX73/++nb9wwT0cvbDtrc1VRhPk46sLOaVxzeoi89ZWIasMIa8Z903/AmO8MVq0z0VWe9ozrqyKMsXBsfhjRlYLdu5azoZHStbbLpXiS2UVY8Vwi3iKhwBegBifDFl27oY11fNy2rQtFJtEVifszW8fFVkN3lit4cFhNbKKM++fVsV/cdA627tcY4KvWkEuy2KU1Y7d79+PkYUghcB26N8PRUs2qzbxxStz1tsyr4bqvCJF7caYBWS13kVc9Bsb2eqeZJwbzNuFq6O2vafJYmHZclmaJbmEJKsHSlaVPr1vTmPDERuZmLWxyRFra4xaYySrejBpUzNZN5LX1iKy6lpHhKdHU6rr1ICshqvIaoDVPL63x8Ue1UrPD7ZeKKiO1nZmZsqisbKlUnWWTCaX/IDzC3wZj2lnYp7FsPy92aUbkzava2fVoIjJvnlnwcp1Nj8zb72dLRaLqU6q8i+kF6aVsVYsTh0AerLKJ1uDp+RjPgmbnC9aQ6hoLz/XZVER7eDM06MT9wSJ97Egq76X8rMvPrGJ8SH79quv2dXLF+z46ZP2wjd+xyZGpmxSDqOnt8d29u2yP//zP7c333zTurq67PiXX9reffsskYjZv//3/4f98R//sWUyWbt06aIUPWXPHjkcJO7Au6UhOyGy2t/fb8nQ3AYgq6HFntVqsjo0HhNZ3fNYkdWvLk5Zsr7DtnarBasjVb5BJV+tEQ8ORdVmV66bDQ+ct1gkY5FYOiCrQ1Miq8/VyOpDwGpklfbK+JzZ5NSkbd3a6nxGMDkE/xBM1/GgKgKcuR8I3ipnJIN1Os2m02bvf3LF2hrS7pOMEBS1h28iqzU8WlwaKNpMOm0HdydUX+AZeRs/EM2D9YRKXepC7zv3u9Y/Z5f6L1pDKqKm8IItlOoqZPWodTSHn2KyGvx1ZDVrdnhPvUqAY0E5BGdvhbdoL8tcvmSTk5Ou3i4U8jY2Nqa6vadCUEvW2dluyVRchAUfUXmBS/ZaFHEYHS3Y5f5R27az163iEROxiKjyKTKMI4iumKiBibqI/AZbb9rrWhlmlbhwF9o114fNTpwatXI2bX/vd/uUx6y0NaZQ1ZVbw+qQqDf8C1a0dsriwyxBxctT7e099tc/fUcGMG/bdrxig2oxhUJQH7M6tbLopaOlVcjnpdwZu3zlilPuiFpJYbWCmcfKvm9F5YtUTQ8Oiy0+gbx55JW/6nOrYykO8X1YwuPgAIPnZiiHHtRwSE0CtTzJuQ8o2UMJEns4kje+Gc1SRe5b8NIF9wJAhfTX8GiASRZVqZVFZJl2gxuHlNLHAg1Z0pclOVXr0N2HYIZaMONRv+meK8oRhlmlQlqjjDkdcUERlm5fwyNEqCiJZbFr5IZ+BDYe6McdQtDHYmU3oiMt0O8ii7Q6jVgh/k0h2LiYJFOQBi1wb7pYOFrDepEWmx0aGrGpqRnV2f1qAAzY1as37MKVa3b1er/I/4zFkwlL1ddZQ3OTxRIJ/a5zX5McGh2z8xcv2+TMrLQgKtHE3JxzyM3QeMF27G0Tbyi4RmZT3IyVq+olKj772q5jDTrWXKetHE6zjjVVAuc5d8ewQrxGEeM60ZJ6hdlZPvPabw31GYvE552GRPW3uvFdw4PHA7dMCN6Xx7+0gRs3bHffHtu784Dt3HHA9u49YD1dXWJ9BZuVgjc3NtnVy5cdGWV4/9z5825+6h/8/u/b9u19durUKWtsbHDzVG/cGHBp9/Zuci9YPUiU1bS6dvW6ff7553bu3DnXQoSoMgdubWSVNJYT1OVADBvVSXqCrvyptVs9XMKZhx0WsZgP8uRDkK8aHhHku11vlXvBwSOQ0YoyFKqP30sI7qFaRveDusJXy34Zmxo2JJzspCtM7Qn0hKFVP/wKEbhNCIviKoFcoWgLCwuWpxEdJR2wQvybwtIGPaH3DKpc6dMT2FaCc9tE9ueeXtyuBCKqh8PhqGWzOTf/lLKEkLLM5Fw6bYVy0do6262rp8saRVZn5qdtLjNnu/fusa7uXkvWNdn1G+M2mykbTY7+MT5GMG7N7UlrEfOMORJbkYT+0Nu52PCsBFb6wP/cFDh2p7D8GoWFopskYmPjGfvqxGnxjKy1ptQgpoOEPOivotXwEPHAy5se1SGRy66OTisXyvZXf/E3Njszb599dsy++uoL2755q0VDEUvPzbk4c9Mz9vLLL1gmM28ff/yhWmVX7eNjH9uJkyft5ddednxkPjtnB48ccsMHvrKqrpZ4W9Ad8VvA/qqB81VB0FEHSCY9whBnehbfe+895etjt96rP/90AFWBEAa/HoLqrAgnFxFTJyrlJVTSvoIjqzU8cmCfVP7ICWcPcfR9rA8KEI0gBPeFgAB6WwNwb1Uv6IvTX/64HRcjmIrgg7+mhgcOycmRRcmmhG+pIFh79fbgRZ0vvjhu//Sf/jP7H/6Hf2GnTp1xHQt8M2R5uBPc2+NOZwI9cUrsAsdqWAsiIqqEhWzeeFEK8prPF2xifMJN4SuVCjYyMmQL+Zy1d7RZV1enJZIRa2pusImJackprOuTduXqsF3oT9vxE9ctmWpxpHRuymxyIm8T00WFrE1MZmxycl7bBRufKdqYzi+FciXo+HRhjSFXFfR7qmTjkyW7Nrhgx89cdXNkG+Jir/lZKUvR9b/XOkYePiI/+clP/owd5hyWZNn0hJYLJSmaFGMmYtOZhB3YHbrtnE+WjOLv1GxebiZubY1B9QRo9Q4PD9u2bdvs408+tu+++T1r62i13Xv2i/j9VoSzwzLZrCWTKYtGeCMwZB0d7bZ9xw4p8bhdvHSBnNm3vvVNa2trt+mpKRsaHrLu7i6riycqdwngfcvwVMZmZ8eV55ybwN3Z3ax0g1mVK4egQx+/xpZvVDsSqsCU3rHxMdu0qde2bt1qe/bsUQyz48ePu7y2tra67cogdbPpubINjUxaQ13BlRVDHfOZiG3Z1G4NScWqXH+bIn6EoBxkmMrcqJxDOBK31mZcOs+2eo4xZUcO2OqP2yoUVbbz83M2Pj5u/df7zb1Md+2q0xPmILMaxO3AXaenQ2rwjMqJ5FzZF8txpZmxXdu7LB6rkdYHDWTAG7yDIzO2qbfF0Qxky/HZNDaftbZWXsZYkjt67/cXjxSwO/0TYwmO3N0/Zz+LDDVv5aIqyozZ0NUxiydZLo6chW1+vGR7t/dYNLYAQ3EkhbePWUYuV8y4lyYYDA60t4YHjWkRjYXsgm3qSlRKvDJhRI6YKRvUQdeu9dsHH/zWPvnkmJ06edpOnzljnx77zD7/4nP7T//pPzmfwby3n//i527VmDNnzrqRvDOnz9hXX510viWVrLOG+kY3VciDPlx0h97Z8dmSTU2OWzLBWwN5+f+YZRcK1tvbbXXyz0/j7CIsxv+dnOFdFrOednquORYUSHA2QF6cISNZwh9YWrJQLNh8et4aGuqtZ1OPtbQ0WVp+n6ladXWq66Nha2pqdr4ik52zltZGcYFWS+eidvJc1lKJnCUiOjc7Z9OzIqYiqaNT8zYpjjE5m9PvnE3O5aRDIq8zOj6TDsK038+Ij2S1XUtYsIkpbXWP6akFmxR/mJiZsSsDY5YVn2lv4oWvBddjPC2+un9Pd1ACS39qWAMycrsXroxLF9qtpUElK1IZ9MjTUAmmawVTgFbGAyerkA+ICcPmzD3dsnWbHM9H1ljfLGXOWl0qKZKatL0ir4NDw+7TquNj4/ImJdsscnj40GHr6uhSy2pSDuisFD5tjQ1N1tHe6YgJU5eYx0j2fOU1olYYUwviqoRSiWbr7WpUYfiz1f94avp/Ii4dZr/OpyftxPEvbHRkwoaUn2GFkeER27R5szMyWu+trS0q8BadH+IRrZHPZ6yIwJwfa7K61E0lsppbM1nlLJe6UPmdyeZsYGDQlRvDRY2NTdbe3mEtza1S2IhrWKAnleJYEaQzPSNS5MiqylORi+WEyGrWdu2okdWHAWSwKlkVSawmq94P+BgQQxdRDSDXHvSn3J/1h8B2FJQQw3eQHaYxQkKGh4pumRx67UqlhOWycdvS12qxZFTxAnoUkNWwiGpcV3LMp13Dg8bUTFmkMG89IqtBj7dKXkVP4wV5Lizk7N/8r/9GPmMg+J1bcO8x/Nt/929dHfHc88/ZP/yH/9BeeP4Fy+VzIrOnXIdIQ0ODdKAof5OxL774whHbV1595ebPdZOg7kMHxYRIM2Q1laCpkpOGRl2+ent75JNqZHUtZJW3/DOZtOws4BFOntoPi7g21DWqjmeWZ9QSsYRISlxlik2GFSUkjpAQD6h3pGU2XdT9cnb4QJft29Fom3sarLtboafeerW/taveNits6g62m7vq1Njx+5XjPlQdv11w11eu2dSj9LrrrFucoaRGy4L0IBWbx6vomaI2PV+2fSKrjlTdrqKq4RZseLJ65coVu3Dhgu3cudP1oiWSSXvppZeVyZidOHHSDh46IIKYsav91+3AgUMWFbllaGBqdtK+/PwL29G33X79q99YIp7U/g6L6botm7aIsKaUR1VNesDlOjMyUXDTCSCrdckmOcOGlfVKFuWWupERYVw84/Xr12xG127dut05vfr6eucIOzs65biSbliCXldeAuMcc2nZEm5FYM6Pd89qBcrc6GRWzidxR7LKsKqbN6h9H6tULNrJ4185Oezft08kp1sktcnqRVAbVXYtzY2OaCIHKoeg5G4Fx2tk9dECGayVrALOKbr0HM3QTihiC/mQXbo0KrvI2sjYggLbewijORsez2h/TvtZGxzm2IwtFPKWlZfMZERIpnIWT3bZ2NS0TSje2Ni8DctXTEwWLS4ik4yj1X6cpYYHjZvIauUYO/gIQlr1ws9+/jP78Y9+bD/84Q/t1VdetZ27dtpvf/tbR1w7Ozvd2twEOjn4qmFbe5v9k3/yT+wb3/iGPXv0WVefvf/e+27JQzobloC2LierqipDeSuUY5bNFax3c0/Qs+riP12gdPzftZDVkCrGfC5ns3OzIq5F+QcFNRgKuaJNT83avBzDQiZvczNp/Z6zmdm0zUylF48RMumcpbMF1chl29ITF5nJW1jyK0kZuFe4SP2pe9PK1XFygZ64nqbKSCjH/b7LpT9+mxBiW7kuLALNV1tZr3VaFU1ODSZWnXH1nciqsml793QF/CYohhrWiHslqw/cDpubm53injx50p5//nk7fvyEW4LqZz/7mX3jjTekI0U7f/6UJerMGtsStrmv23o2ddmefXv1u8l+8favbWp+2nY/s9/ae9pt265tlmpKBfPhxGr4rr7UaNFwqBRRdj63aqp8cpZ3cZiTtDwUy2EFeaPpol3/5LxYbt7i5TqLJRotHFWLUGQVUkph8lIV86QAhUqAoPIy2MWLF51T9IDM3hpkXJXrCE8qXGOnyAc31WDRPttsJmNXrl6xLVu2OKIahvELiOidd9+1v/jL/599+eUJV76ckkoE6SxKtYbHGVhGKJSTNLHamM2ly3b+7KBlsiHZWJ01NMbvPTRhr+wn1dhNWmd3yo4+v82eObjVDhza7ravvr7XUo1ZNZjVyE2Grb65xWbno3bsWL+NjpFPGlmMr9SwEYDfBLwvUFcvPyyG0N7ean/yJ39i3/vu96yjvWMx0Jlw5OgR++GbP7SOjg5j3U2WQ6JHr66+bjGtGh4cwiIeMdWbLEVEaJZ9JVMp58vn5uZtdnbOraU+RZiedh/9odechgb7vLxcn6w3SdvChZglQkmRwrhFVE/H6IENxyyqupjAi9Vsw6rIGRHxgfdfFkMl7loC6cUikCd5KIWw3EAoL36QE4GtzU/dEHjgPauQPQgrBI2W78uvvGKtbk7qHqtXS/fqtcu2d/8+2733GbViWBRaCqiLQ/rT0d1n3b177MDBwxarCxY4oSohcEdIKPtUgdVheC5uM3NM9I6KdLaI/CYXr6sOjrBqG5MyFnL6IUMLN7Tb+Fza8tl5tfbUApyZtdGRMTckxPybap5Jq4ApAKdPn3at/JsWN3YI9oOe1QlrrOdNQko6anOPZc8qc1bv0LOqQ663e6FkmbFJOa+Ym2M8L9lv3bZFTmrKLojc5xbybi4wL6/t3LnDzWtm/lkmnbXBwUFrbm1SUkrM/V+6D3et9aw+WjjbW2PPqj+OnbEkTVat68tXBixVl7TtO9tFNGLW0hy/p9Ck0FzZb25KBMebSDcsXY3qGKM1OtYSsbaWqI4l3cjMfD5iN6R3xWLUeje36JrgDd+Nvd7xk4PVelY9slKWd959xw4dOmSbNve6xi2uctvWbXb4yBG3aP+zzz1rR48etcOHD9vRZ4+6ETyIKiD+5SuX3VSAH/zgB2vqWQ2H8vInDP8WrHdTrWcVrKVntSR/sCDSyTrt7nPoqscz6bTkVbZWVRjMSW1V6Opqs/aOFvl3fjerYdFqjU117lhTc70asAUbHJi2rT0NlkrqPhWZc69ymAk7VQrioDMcWiGsp1NIzEcpUZ9wJ17442WtjM3NFyyWyAZHKtMA9u6u9azeDTZ8zyoZoUett7fX+m9ct9/85jc2NjZuZ8+csv/z7/7aIiIXW7f1WS4fsvc+/NKOfXrJPjo2bMc+m7RTZ6bs3IVB+/L0gH3y2aAd+1LHK+GT4yN27PMhd/yjTwdceP+Tfvv480G7cHlMldC0jU9l3fajz2fttzq/PHx4bEj3GrFPT1+zkwNX7JPzZ+yzM9ctnU2oMm2xXXv22oEDz1hPb5d7Ftfr5/YC0NvKnNxdu3a5liEosErxMiAEhkmeBkDGnY+YnbaZG/1m8zN26fxZ6+vrc/PNPvnkE7eSAj2tlMsrL79ie/fudRUO85H5mMPE5IRbKowWN6j1sD7eQB1YBXU+Z/bV6SFLyLY2bWtT2zAYqsPx30ug99715CtAnN1WCRPY95bHwh/02rMc58jwgp08eUoNxVmLJ1kzOagQn05qsjHhyUY16WD0hXmR9JhCSgluPmTEV3iViBVwjpE9P5pTw4MBL8jOzc+7LdPmSjIo9lkJoK2jWUS0zoqlBcliwerqE1ZfH7dkKupWBFjIzbv9xqaU6w0PR8oWkW8oFWSsXp5uC6WkBmYEpPJPRr1qWMc/UoSsso8LKIm0kkbRfRJ2CYu6uEzPanjweCgWjIC7u7vtW9/6th04fMB9GWT/4W32/Mv77aBaxDnxu/MXhq2pqcUOHNxpBw912/4DjbZ/X5MdPtBtzx3stWcP9NiRfV129JluF559Zmn/qNvvsucV78j+Tvvh1zrsv/7xAfv+N/bYd9/Yaof21NnhfR12cG+72xLnuf1dSrNN17TYvv09duj5fbb/xX22c/cWGx7N27mrak1KT1kZgN7mUDRobS1WfHomXgYCDDuxRBfkCsfIWrH0sFY7Wfarw5MK9wKNqQnVUbCefSkrJYbstW+86KZMMJyXVmv7yOEj9vXXv+5WV4Dknz1zVpVQ0bb10WNy2Hbu2OkcnS8nelYpd1/2boqS/vj1awk4xxruHl5v0XW/XQ3Iw+txtVzYIoZALujCUlA7RUR1zuL1PdbeVW95qUguHXI9sdNqk7jAfiVAbNcaZkU+CfM+6NicD/ItaR1THemOTypcmlReLpy3FGsnqlKM8DUaVUpJ5ZNXr2rYOHB+NBySThZcZ4F7Ya5ccqNvNDzoPWWf4efgwyBmAwNDdumSGsPaR0fRR4ala3hwYKSkva3NjaLW1dVbXSrl1lhlhYahwUE1DsdtglX+ZWFRCaaQj9jszILFIgmbnc7r/IwNDkzqWEaN2bjqW8ktJMONiEQ6OeJ3gqZnMFsVbxOEYiGomSOVf+zPzMzZxUuX7Ss1SM+du2DZBRm+jgfN46VrfXDvrTD9wOLOYfnP/YaWTQGoXlGihoeLmyXxgNHe2mrP7N4r0hKxrp5e27H7kE3NROyLzy5anY49f2SXtTSaNanWaEqGXGhMKVT2Cc3SJUJjLIjTmAhrG3WhWa2zljrtE0dptDbod31IIWJtDXHraEy4bXt9TMeIG7cmxa/XNpWqt1IuZZel4KlUSaHdogmcY9j1DvO2+vLC8mSKDxVcunTJPvroI/vss8/c9saNG87RPn2IqhIpWX7mso2PnpQjumYhahWBl+ra2zvdsib0nv7d3/6dKzc+z8mLau+++65l0hl3PlVX58qViuahKulTCnR5imXhhoYWRwduR1jXCu/aR8byNj4xbAv5Wbt2fcL6+wlTCuN2zYfrk4vh/OVxOy9WuZZw8eLyML4scGzaLlwctMtXJ+zUuWFjOasG2X0iFlRvIViPaqmn02Y3LhxZlW7G41GbnJi2/ms33OgLZJTtoPQVIjs1NW25fN5GR8ft3/ybfxO8hLWQt3whCPikGh4sKH86bPL5rOtsiMVVv4qwIhN+s8xjfX2DjU3MurmqfjSyu6fbNZAHBwYt1VDvPiaQ53uqTCCtwFtlQC2XQkQeJh6NubmrYGR4zD79+FM7dfyUTY9P2+TYpB37+Jh7ccpdo3jL06gO1fCEdTlUJdXwCLBcPg8WEnJ+Lm0Xzp+1Y5+esrff/cre/e05S9a12r6dncEnGmkpi9uEiyKHRSkWQcdCHA/qk0XN5XcQX4TGBcVTPRsEOblKcOlVrifAnRhOyusa3i/M6dh82uzC6Sn9mlfLkLfeg6IJ6i617spBBb4SqNQZyn7hhRfs2WeDOVTMYWVxZEDPgCe2TzpY2qtUoHWat3hCjYuWNm3r1WKedz2oDXJWkFDI6cFDB+3VV1914eVXXnYvtPEiHg6O+U5URgwp1fDggX6yxByrd5w/f17k7qKNjIy4nhF6XZcDErGc2MmMVgVrKG/f3mGHnmm0nVvbbLtC37YWhXbt+9C6GHZvb7fdO7S9q8C1Qdi5s9127WzVttl27+p1926Q/sXcvLqQFfMF99Zy8Cw3P08Njx7oJaMu2UzOfvXrX9l//8/+e/vTf/Gn9s//+T+3f/bP/pn9z/+v/9n5jJ///Of2H//Df7S/+P/+hdPbb7zxDTdfnmkAhKfF/z4q4CcKhZwlk3HVh2X57ayI67w7R9nncnnn33mhemZmyvVQwkUHBod07bQV5OeZbtfeHnUjcAU1Hukewi4ZoOffnXD58lU7fuK4e+n5xZdedHOaDx48aI0NjZUY9we1ztVHg4dOVhMNdXZg/wF74fkXrb21zzb1brMt29psJm82LV4yIdI4XhUmpe8Tc5Uwy2T4IEwpMKrA8al0yIVpxffBHyO4NHTtFHEVxJtsVsdn5nkpTFulc/LEsAzphjXEg+WwqHqL4qdhlVBCTs/POfUVsq+o2X755ZfW1tbmhrrr6uoc2XK9guWgEqwervbhSQOlk5fD+USt2LPn+i0bfcYuTWy2z87FbEZlf+bMOVep7Nu/z81dZX4qH4qg14MlzZj7u71vu1sbkfllzAM+f/6cTU5MmEoxuEkN9w1eD2loEahQgp6Pelf2vCzIi240GOgFodfVx9WVlVRuRaWDw+mDdy7YjNqFbvSkXucZOamvM0vV65j2VwqpxN2Hm9KKKajtROBlxqTOs/i/G04u8tyy98rkfnJcIzX3D+gX8/rvxd8hDxoTrJHKajL/3f/jv7P/5h/9N/aP//E/tj/5v/2J/Vf/9X/l9LWnt8f+l//1f7H/8B//g3uZCsLC9ACud6HGMB4Y8Ad1dQn3Alu9DLu1tcmamxutsZGpX1HbsWOHRaMRkcmL9tVXx21oaMBaVEdSpzICOT8/ax2drdbb2+nexKdnyr3nhD1qi+RWkx4DIrwMfuHCRfvyiy/tpRdfEqforYyGQpJzNj4x7l7Ue/vtd+2tt962X/ziV4urz9Tw+MDXJw8HYdgfC0uF3SfUbly+bnWJoo2OTNuNgXEp7rhdvz5sA1Xh+o0g3KjaD8KowlAlBMeuuTgcJwxb//Wlc0Hw55biXBkYto8/7bd0ZloVdEatvYKFi7T+ZuzM2evuJYyh4WFXUMxf9fCVGsMbEyJUXV1d7tiTSETvBJxVXmXGvN25+RmbmBxXC/eKiGZBjmLGDd8xl2nT5k0BSRJ5p/UMIAw4rGwm6wg+pDXKeqtyZHUiTvVqjauqcXFruL9AXwOnHnYklDnYTAFgy0uRvIXNi5E0Lpiu4YftGHrzJGA9YAkYR2JlIt7x+IrowQR6Z3LaqhWsAxzjvlSAboiP4Ro2qtTClVGQGu4v7vRFutvB+1KvZzu2b7fXv/6affOb37Rvf/vb9v3vfc9effVlNXzq7Y2vv2H/6l/9K/un//Sfun2IKoDo8tU8fE4NDw6sj15Xn3KhuanBOtpbrbO9zdpbW6xRBLanq8OtmX5g/zO2Z9du1Zmjcv4F2793t+3a0WetLY2mtqzrtWROa2CPGK3zNu7famBFGb5U9sY33rCmpkY3V5kX79i6Zc9SdbZv7z4X9u7b63wbn4FFN2p4fODrjLvG6ip0K0pWML77zLegUcotmzbbkQONdmBvsx3a1+7CwX3dK4YD+7t1fikc3NepbU8l3Hw8OMfvlc/5sP+Zbtu+r8naOxMytrwl60IWCRWl5GXnZNtaW62trdVNFqdCXw6cKcOm7e3tt/mK1ZMNiCrLUn366TFHOp999oht3tJhccl4+5bNtntHrxoRlxzhwXHwgQhIKZ/iw8m0qBX+2muvufJj6NmXM/uUP3HvWUlruAXoLvPLGPpHbn7Yn6VEOAdBgLTSM8Y+L0j6D19A8Rimcz2tld9rQahUWVbM8YalK5HvyiGIw9YF3XNxf80h54JyXPktsr2Mt0Bk1ku8a1iCJ5XVYFiYpQqx+5XOrxXV17JLT1pUOuqG9iVM+Aa95JDTr7/+ulv8H19CrxlxGaVhukcNDxpMlYP8Ldk1CNYxZQlL3gthlZ2EdXS2Wd/2rbZla4/q1yaLSX58hRJ4SS3aozbY7Goo6Z7Dw4N24MB+EdUG9xt4vWFL7yrLJxJYIpF6p6B4IenRzbmtYSPjdnpwW6BKbr6oq0BUFUhH6A0rFlWBFaQAVGTLQ1nnFVd77s3cgohhQg1vlJlqzAWUU78JHEdnUTi2wbBOZatz8lcu8NsHjhPYX+2cDwndr8VSepaIhaNJN5yA82OuKS3FZDIlA6tz+/Q4FYtUeMGzkaf5+Xm3kPGRI0cWSRZGdksQM1/xOIXo/jx+4P1JVxaVnlF6ohmOo3f1448+tlC0bO+8955lJWfWPoQY0bvK52spj6NHjtpPf/pTO3f2gqvY3n77bUeYmGc0NDikMi669BLxhOV4QYJ7KbgKywVVRtK9YLt8f6mCe5rhywLCCfxb/gRGBBjip7GFHtMooNG1e/du16jw5yEb9K4yrMpx9JzSDTGnm/nk+sFsbujgggIEAdAYrQbTAILKTNClgddYHUXlGXvTE+iX4rJkyGzW8pmsO8eHJxZBr9mKPWfcI1kJN99v0QaXZ7SGdQMdY14zwdueIwhDQ06H8Anse+J6U5Dc3CXuz+pwPnuZqPDXHGN77ep1uyxd5gVNRgHQX65h5QBQk/ODg6wo2C4TkCTrpnAhgWg06GEnLv/4EmVcgXrELXdYCasB6622YK9vBEZ8mMJE49r5p2W6xPsn9L4zZ5Z3Jnq6N7l3S8rSCReTP65y4UeApTr65mcKPiernZtvUcNDwM0e/G7ghFYyV3cVSjY3lbWJKbPRifCtYTJk45Nlm5jI23xWlzLsuKgpAbxSLmZMtZwbCFAU/A4O6H4iGBKsJOq2mA/LaszYxUsXHYEaHxtzRhCL8QHXstuHnH311Vdu+SXmqT5tEDV0W+b3JlQuXR3dVpeot5GhUddTmqpLuZemvjpxSs6hV42ZXNDLIUVhvipvgP7hH/yhm0/Emqr0hjDXjB5t5rFCbv1yVLztWcPasdg4VIAg+C2EAceO/kJO0V0WU4egMn+YearERbeJSy8r01uIW10BUAkVygUriCRMO1vngxF52bbZ1CRLYC3Zc7Vt+2H3pSOrGzO9HuFwzEZGxuzatWs2eGPITp06bSeOf+U+38gbxkMDatQUmHfLFUoL+60OLv1KM3jxWAWVvAAaOjXcPdAndAVAUqtJKfpEQ5/zPs5a4dPAb9BTOjo6Jp970oXTZ07bmdNn5KODURt8yJ/92Z/Zv/yX/9L+9b/+127kBrFyf/S3thrA/Ye3MBdcPRqEBwnkCUH1+sQUPN4Rgaje5KO0TzwQV/1Ex8eunTvddu++nWqYt7kXuh0V5Q8PcTMvrWGDocp7rxfe+Blcg7SYNTXUSwGSNjmtymvm1jA5FQvCTNnGp+XQIlGRv3yF9twKX4ekM2n3cs6CWyvtIUAVWUtLi73wwh43qZ8vmfCmI+3EQn7BEViWp2Luy6ZNmxaN4mlCsVRwjoOKilYtLVBIDr95SeqzTz9zlUSEBkmh6IgQKyTwZi89sefOnnPp8P1uvuO9/5n9Njg0aB9++KH7HB/kieE7R4wCNlLDGuEdNT3WDHtB/HHsvARIwwq5ATc/uNIbQYXPFA16wXiZCv1Ht+lNXYlkhKMh697U5V50pPE5MV6yyfGizbOwaZmKg1jch37XCqrJ4irgCgJ5GpA+XDh/wenLwNiITefSFk+lnC+Ympq0keERkdlhF1cK6K53N66qtFbEYj6W9MpdfofLalgZEFRsn1Emhv7p1US3sH8CIyroE8eqCcVa4KdosJzd5198bn/+//5z+9///H+3/+1/+99c+Ou//mvnL/DT5OEv/vIvbMfOHdbW2ubUAN2g8bu8h6yGJaxXJisDm6oOdwK2x33Xcm8fLxglwlcRWLObkSHI6ko+ikYScmeNXbaud96xaTWAGEItZLVVnWVpHXtI3KKGu8ZatGp1uBemULqoU72OVrMjh8K2Z5eZGi/2zLLAsX07orZve9L2bUlZODtnSTVngskBKwMl88OSfPXIz0UisE9A7wgMB/lzPtwNuKebNiD/ViphHGm7dPmSW27pF7/8hZvbxyR/yCrOECe8MoJhBEgXW5zCzUFR3J/HD3ynmWeH+DDsxoR1KiV67qig/v4/+PvuLesWHeNNUBzLsIjF19/4utWrUdOmli1r0lK5XLl8xU0JoExYagRSu503SCvleschvDWQoCcdlB0EFSdOYwrCCTFFRxn+Qkf9hxmc40YvJT+vl1T0fqrG9u3bXQUAqAR8fA+Wz1aK9syOBtvTZ3Zwd8wO70rYgV0R27pdBCXFK5RoP3+DF+lAeQ18gSt4Dl7Yu3Tpsh2WPtBTf/i5o3ZtZMj+4q//yl585WX34QiWPqPH7DJr9U6MO1tCH3ke9gMbW9m+/DmmJtCgYk3OAMEwYvV1y38vx+3OPc7wZemfj9/+GLqCrfv1MjmOnkFKCeyjf3wwhUaS95FelxaDbNuplvtzKyAakE2WoWI5qj/90z+1n/zkJ/aTP/2J/U//z//J/tv/+39rrW2t9vEnH7sh3h/98EfurXAaaiTp8iwdgew+7cDGafgxauIbFm4EqyLTamCHwVQvLyMvn5XldJNMl/1bCSsfDe5LZ1GwGsASGMnjfRc+nfvv//2/t3/37/6d/dt/+2/t7/7u7xw3qNZTQD68rwOBHgS2Sp1ulpFyiaSWeLl3QCnXyOpGh9ON9SNggfwleBVBLdQGNqYt084J3sZdCsH5Cqq1dRWy4VpCisdk+j179jiS6OcioZcsTeG/KMGbfYV8wS0SzfeJibNeVA9heCVnHm5PT6/rGUT5+QrXK6+84gjZ/UF1QTw+wHFg+JQTFdLc/Jybn8ZcVKZH/PKXv3QEf49aLpcuXXRySyVTbjmT/fv3Wq/K9Fvf+pbrveP73QzP7N+33/WIQLAo6xruDJww5G5sbMxVRoChewgCvaaQBE9KvU4vB8chGLz9T0NhOVa6DvMKu5eWKi8whQjBOeCXsFoPeI6FfM71mjJnmakkqVRKula2wclxK6Zitu/QM+4TzfPZjB1TY4eVOtA9CDqk9X4iqNiWWry+rH3F6MOTCvSG56OHikYQxAaCw9Ar+5xnnjMB/cFH0rjxZeKX8/O4nQ6uBclkwvkPQnNLo3uhhs9zQmSZysLaq//jv/gf7R/9X/+RI7lkg/vF4rKBu6kQwBMkXnwqOswcdV4yIvj57CuBGvtu+nsC33D7AFY7fju8/PLL9id/8ieL4Y/+6I/cM62mVyvJPRxW3VLKWW7yoo0NfWFTk5fFH1Yvhxo2BtaiH7dB8LLFzWCOmE926Z1cAnHpQ8UACtKtogLHbwd0DT1kSYpnnnlmccI8PZ9MC5hQCxGSdOLECffVEnromBNJtLv1T8D3AmEEkKdDR444h+jm9QWnFrdrAemVlKGVe5dISS2/xb2NDyok7yR4U/yll16wTHbeGf2RI4ds587tIqA73JzUsfExN9xPPN8bjkyRIR9SeOXlV+zAgQOLxwjMU64tN3N7UNG4xpsKjuAbVBCHaue92r4HsuQ6elVXGk5bHRhYZU6ot2NEVuY3u2jy2nu0yNvk+MQi6S7o+RhipndmfiFjXX1bnC3yzfDjx4+7uaw7du10n2zes2+fTUxPWTa3sKYeXIeqeavVIB/Vcy8haKyawOgAOs9xKnxInI//JILnpIeSnlP0DKJO44f55QR6T9Ebnp9y8gGgR76RdDdYcVQsdHMjgYA8IKyHDx12yxK1trTa1177miO0/jydGL7eWBcec/fDs/uALNFjLw+O0fONPG8vo0q5LXYorRyXWOsJjL0EZHh54CRc4dbK2/fSetkDpqDRu89v9O2OUNr0sJKLUm7KYjZnzY0p6dvd6WkNDw9rkO5aUK2GqwN1oCokRKVrETmQxZebKkCR+GTayZOnVWEU3NC+I6wKnZ3tcpB17tinxz63X/7ql3bs2DE3z47epL7tfY70sBxSNVHlhYwLFy65r6AEiroEdJ4XeQjeCIJA3yHnIU1qpRMchQrWAOXtZnD7p9YTYxzliKpsXs6Kiszp6XUskpdhqjHH4/MdrccRvpKil7mnp8fNM4VEIBMWlP/444/txPETbmif3lda8vSAcZ2XD5UK8kok5DSrZEaZusqu8m8R7IZpBVNwT2/vK5UNjTSG8wC9W5BV3xvGdjXn7cp/hQCqh/39sbWhci93ibcIjgXEdS2A/DF9gbV1+bxxQeToL//yL+3Tzz613Zv77ODm7VYfT7nn/NrXvua+GkcjFdLu50dCKNeXbzWadV/0r5qcQJghqJQzZcyoDpU9ZctxvvLFFlILaT537pw778F13pf43liOPUr4PJAfgs+jB/sQU3rpmevMs1GmEAIaMhBUerohqMv1pDr4tJYfWy/8dci7GqTtn4WtL9+Ym3+9NNrGlgZvc0uzO78muPqoEtZ4yUbB8mfELpAnfpdGh29kAGTKKMpqI4RLmqoyVp1VKiVtXskzu5NVP5YHBtHXE/DgK6VD9RzSvUx15hKW9AedcPVCZUu9gp7iN3zjsRqUAVPU4Ao+/Po3v7bTJ/st2fKSNWdfs9TCHotHmytX1LBREfnJT37yZ+zI/F3r0zmAQskRwomZiE1nEnZgtyjaTf6GH0GQO1CgTYShECqV1ioI2lNlm8+EbGJs2rZvkZIsc2jkg6EmSChOMZFIunmPHpOTU3bt6jV74cUXbO+evbZ121a3xBFfQgqWO2ImbMimJqftlEgvlQpzJN1E7JsfRPcyGxzO2EIuY6lYVkdClodQhptsS29CZBpCHeSPpyObTBcgFXwi25tTvBVz8xDwWUvEeJksatOzBesVud6yJWWhqMi4TJT8VpfdndJ86FCGRicXLBxR5SWR8W110aHgVEV2tNJprTPnkeW8+L17zy5rUWUxMzNtN24MVN46T7hyvB0o65WAE52aKdjc9JjFItI9CaMo5zY3n7HdO7otHgvIUbU+Pe6gEuJ5sE16uOjVwy4Y5qfyoZyX4+E8P/dYCshsNmOWXchYW2vKNUoBx/nqXKRcsLbm2C2y9Tnl+Xhhgo9B8Abv9ev9duTwYUc2+GRiWmTwzJnT1t7RaXFVst29PdbZ3uHIDL4CAskKB/x29lRJ2JUFN9WmoO2Nwax2Z6U/VG5h5TkYHWhtCOIAypx0IL/0SjGkDWFj+JseRU/ekAGVJPIgPs+A74Lccg4ywDVUqP5ltkcFyoFAHjxZZcvzEXjWwN/yNaI69zzolicHa4W/z50wNVOWruStp0v+oHJsSWYiLws598IlI2XUCQMDgzY8NGJDg8MuDA4OBb+H9HtI+ypz4vRLFzh+TY25L49/adf7r9uPfvQjPU81MQsUAn8yMVNUXTFuqYTkI19cVCOYOcy9m7rdV9Aq3HdDA1lCzJAh+oaPQOdoWPhGhu9F9X6DrQfHg87FwI6rURRxHBuftMGRcRsZmrGBoWmV/4wNat+HIbaD02sPw1wzpzB7UxgcDOlc1vo2J93X5nzRwxqCfzeDfKPD+EUIuLcvr4Pz0h104+izR92UszaRW+aoj01MWV/XdksPL9hUOmfJtlYnaJ8+vH98Mm2z6YJ8UVZ6IU0pR21GbH3Prq6ADy3PTA23RUatkQtXxsUH2q2lIasGQsj5G2TmvyJI99RqCEnJnW7KRC1fLLj5nqVsQUqfswv9cbs60WR/9GbYEqumgbkvtcPu1JPihC4MTYTt3Jlr9o1Xt7phOzIK6Plkl3z4lz6oSPgihQfLITGfjc9zQj4ZVuabwMxdYqoA81f5zNvw0KhbwmRT7yY3VLQSxMvt0+MTNjM7YW119FKFLZtLWCm+xV55tsWiZebkyQDCUbUwlbdwVdfsTV5spQLSMZWufKl9/lW/tTZlbCpTto6e7fbM7oQUXvlUGmoiVK5eKrtHV6WtgMCv25lLMxaKNdnOreSveGuvuBwmykdF7ytCtvQA0luzdWufCPqmSjlWLloF1RpVDejF5f68DV49Y6m48hAJ2UKxyYZGRu3N7xyxer63KQT6VJ3KhirRdYEyxKCpfOjlo1LCJihr75SX96p4e3qYwDJuTDCSMWl7drY6skqpI4UL183ipYzt2pa6RbZeMhA8yCqjI1Ss9NAzj9a/7DU0OuLICw0e1n3loxFR5owIEEZ6OCkj5rfz7X/8CnBlwU21ySiTH302Jd0dUONUzFoV0MB4yA4fPmI7u/Wz6s0OrvMVoa8EGXKEyNGTTZl7GVQD+UD+PNljlAFfBuFlBMiD6x62nCgfAnn09yePvtJgH/gG0t2Aa7F3yBBlQFgprcvXizY5PW/PHmxyHQAOikYdgFghjH/zN39rH3zwvvsKUbUN3zqsL9+ji5zv0TnujU4gB6YZ/e7v/u6yOkDXy3/ReLlwPWdXLp2Tf45LZ1U2hQabmp2z5184ZO1qmC/r39hwoEcfH+H9rw+enC4vey9bttWArC7RNV/aZZUI3Sk6L6oQdYQu6MGsfscDrGcJON5iWUkn8qWwnTxntnObWUtKZa+b+K4wYi9pQACegc4oXryig8TPufdp41P+5m/+xn3yFT1k/vtMek4ZKNs3X/2a1YUSOla0fF3M4iHZABfpZujQ2YsTNjCasQbxAhraVkpa/2jZfvj9g5Yk+VuzX8NqUJnyefyfvnXO+rbvsb6eSatPhl2nBPPL4W68yCetqFxwKzYcWV0OT14hN/T2YpCsqUeFxkK/DPXwUtUnH39i3/ve96yuXgqJBxJ8mv56sNyeClL2T4+PO7LaXs/nJMOWzicsE+qw11/ssUQ4J8IasWKlsqp+1mrDXglcUS4V9Kxxe+fj69YQD9nm3pTt3svbsbgAZuwiHi8if49bjXIj4E5kFVDmDDvRy8XQE84EYgHBaGpsviNJ9VgmpkXgNC9fL9ngta8sqTKMRJJWCrfYpcvX7Pd+9JIl1UgBgeyrU9mIJXp7QCggEBAmyI7v7fKVjCdQXs+rK5/V7OlBAtkMjpsIyN2R1d/85jeusQmh49kYiqZHiJ4+ek5Pnj7l5uO+9PLLLj4vXsohLBITwNQTRld6usQ85RsoE1cW3FSblcjq8HTMdu3cY8/0iaitUGy+XPE9yIR7cU+I9XIC6uGv4d5cgwzZ57moQLEReoHpDSI9esAA1/k83w8ZkhbpE+htI01IDGXKlvz4vIL7cU/IPfZPWtyH52frpxN4vV2NrFb7aKaCUO5LtiyfyaLeYAX/E+Dm40GvjZK+6bDSq5DVizfydvniWWtpUJmEC7aQa7DJmWl77oWd1tsuklyRx6MCz18tJ/KCb6WcAbpI+aJDvmyrcbu8V8v+ZrJK3cQ5PD3boEMl6PlaqdyX0lkLkKYfmatGVskcP1u23X0ha0uRJqsABPFWvbOegUbke++9Z6+//rqb184xnhs7+9u//Vv3bgS+hFVGINWN8iPdHZ0WK4t6K9G8At9KdOnrtjWyep+hMr1XsrqS7B8pcFTVwdsZ+/SuTk/xDf+kM06IKuSapY9efOlFd5x5qRgxysa11dcTbgeU2xl2KGalSJtNpkWqJws2Ml2ysamyjU4HYcyHmaUwqjjLw7DC2HTURqZylisvWHdXs+3d1WZJ1eIIBaIa4A4Z20CgOAO3FJRVdQA4CSrwl156yX2/myW+9u3btzhXaq1AMVcKDq6SCuTrhmZE5Dq72uW0V+5JCEAZB+Xs9zZyqVNB+Z4RhvaYe+bJhe85Wf6s/nf1sYcJf9dFOQkcW94DsxqoWJj7jA6xHBrPik2zZBxz0jarEmK+OCSVOBOTEy4Oz+t7kvALjtxUymC1slgqq4Ac3pzrm+Hjci/S9y+m+N5teirJH6NAwFeUBPZ9DyPXso990IAjDUgkveX0DvneTk84uDbI2+3h4/m4bPGB9LgxJYGtn74AYcQWA/8Z3MfnlXAvCEilSrIiN+4FcSAACIUnWLcDpNIHfD7ziclvEFKuYnNB5bdyQBeWAr20NxPVJVQ/MvnP5woigkGveJ3Ia+VMJTx8+AYrW2QISUVXKEfyCPln5IGy8fKrluedZFodr5oo0J/JP57bv61xuyHa+wVV6bqPiCHkcA1l7nWeHlXqG9aKZjUafAl+E6DvvBxNI3bfnj22Z/tO29zRZTHZbpFu8zuRpIoJitfW8Aixigk/WHgVXOkFvJyMMpPJuuB6eYOOXymlgi7k7XIMlAqLcywQfmPghl29ctWt2/n2O2+7SdQsPu9BPNIlPVrqd0LJ4jadjtv1QbOR8ZgNjYVscDxsg2PhYH8sclMYGOXczcd8uDwctjk13Xbv6bID+xosGYOoCmWG26rnGj4aZ7hWrJw7jt4cWMOOOjAWY9mkiFqzDIUGQ4y3LJ53RxB/eajcSbqDJGdm5+W0U3bk6D533yUQy+UoiO9+LQFnfOvRjQVIOMOpVETMA4boVBMT76g3MsghslpLTlnmiGeCANITyUtLDPtDthhJYc7Z0NCgG1nh+EcKn376qfvtiSJEkuvvvl5dPae+4veVO34I4sn9qBg9YeUZPHHz11SDHh4qVypRzvuec4asuZa0/PVrBddBYEiHfJAG6VKmBAgN5AY7XClP9wOkTb7pwQUQLMqGAImkIYIur2h5lSyteK4aiIfGqguV3/cNYZVhXuXVZEcO7rb6BBSGhsd9vcm6QHkiU4InrjQ60Du2lCvy5Pj9ly2SgDRS1rIp99LTTU52EV5uaw2U6ErHCRbSPQnBryoE3rz6H9N2CNF41Nq72u3o80dVLxTt2OfH7MsTX9h7v/3AhkdHpC4V29WNoxGVWUS6Ko7gJevq5Bo2NO7yBatqcKJaqQJl5i9nlgf+yvzd51aZwLx9S6MV5WhZOH9WxOOjDz+yr0RAL1y86BxuJ131cRaGLqjVm3dzUw8eOGhxJsNLAfnkI9+exxEyj3VyYtJ9zYSlriCpIyOj9sEHH7hP80FsWesTQ/cg5/4Fq7o4rX5VHqWIJePN9uwzEettjVhPe9i6WtV602XdLeyHgtBiQdB+R0vIOlcIPe1mmzrCtqk9YQkVCr6EVj4GtlQu1aUVYGlvYwCjHhqZtXqRgcYG5FtchXsuz7mPxPbeQ1GFN6oG88jolGXSZeWnwZ5/dpcqFuWJ8lWsAMTHnanSdEeDHgJyXpIDloZrj8poo5W0ci57oLeOFwMgGJAaKqKV8KCIx3pBOdOunJ8KdhobGNbBq1ReYplWAyZUtNZmEW4uqIJ/AnqNIGz0yFERM1yOD3Cf7xXBo8Kem5t3q35s3bLVbRtExLBnzkOS6FmhJ49yuwncVDfyL1ipCVl5wSpkc1nZeW+3tdbrpFPq1cvUlzdbHzyB8Ms4MT2AHlOIs4e/DlTvcy3yhfiSZwgnPZD0tvoeXHwygbjohr+efcgxZUTZcU/ikA4ElTKhgQNZ5Jrq+64X3MuDdPjtj7Hl3kzboEfL9+Iyrxc9Rm7kBbm6Z3BXSVdmyjYvn9/bW+fssRTKKy3Sk65LaYL9BxCUOEPB3IdVXSYmijY3u2CFXNjqYo126Mhma3Dv/TL0TfC2d/fltxJcXasMLZcLx7EDv2QYeoV+sEUfkOlyed6LbKuh4nAIUsNj4iH1y6cv+8CeiVcd1gtSo6eWbXUg7eHxkrU2RS0V8/VicBf2bgdSozFEQ3DHju3W0tJqfX3b3bz3YL3f4CUyHoXyKlUILGCv8qQOeK7xyYzNpkuWjARrsBZDUZtLF2337m53QXBNDWvFvb5gtXINeI9wiS7XZh8qQGmYFYNyOgJRDJYx+ta3vml/7/d+z37vd3/XXnj+BSlfMJeLF6boZcLpYbSQaXpat2zZbF1dne6Bqah6eEO4s8MRQsgs88ne/OH37e/9/u/a7//B79uevbtcercDKwDELK3KVfeNlN2nZKNKj28Js70lEE/b1QLzfXmKwEiCe0CTgsCkbv8vKDvCxgESCqQUjyfVODBbkKwKpajlipHFsCCCT8iVdVzB/77fQWK3nJQ+l4u7lQn2P7PTvTUaUcWyerlxhrmsSzGqVHHDwS+BxJxGKvz7VRE9aJDNsDhiJi/qoQLOldWY1JYAkb0TGKZjRQ8qagIk7o033lgknrw0FdFNEtGYNYjItTW3OPumjPAJzGtln+vWBDedRJu1ZK4KyMMHD/bJAwFSRv4hGpART0p8PI/l6RCHZ6UHGXLCcXpHfU+rT8cPB0NmfO8lc3kh7dU9qNWovu/9AOlBUOnFJU8EyDVEgcYCfpo8EqqJajUYXI5EG2xebYZsMSa/IrJekO+QjT/YEFZYuk9Z3nlmTv4jkrIDB7daU0V9gtrp5jzfT1AeBE9OCZQnUzeQK+WJPgdEa0lfluN+ypanDYL/V+U33W2Wztzrv5XgbuHs0ocAa31ClZDxIQB6T30j0NlSJQW2ZZUXknVxK8GfXxlBPkpV6zJjzevzGjXcK+7DC1aAHgoPlv8QVpOk0kHkQ1NmZ88N2zdf7pajg2wwbFSUki11yHMcO3Q9kbropz/9mXPG9KAyb+c73/7O4moAxz495t4Y3bd/nzNen46/1mP57+oXrDoaWIw8rBZA0krRzfby82rNhoO25W11eb1ae7u0NhyQLW47bFeulWxoKGvJukBfQqGgjMVJHhpK5aTNT0sfUxnbvztlzapYFl/pQw6LeUHI9KwicAUd1193tKi/RA3cMEc3BqjwISD0qrmRAlXynqAsr+g97mdFda9g/cSR+bJdvDikUm1UQy6lAuZoyWamcra3r876utRsoxVYBf9kkDL/8iQVjJ+nx7NT8UBkf/3rX9urr77qCBHHaMDSi0ePCnPVuI6X+W4Bgtdt/QtWERsMXrASBsbirvdlZ7e0wk1QuzudQE7IA5lB5NiHeLAPifNDtiuBayCmvqfB70Ng6GHnvCew9Kz5HjYf/0HrAc/mwfOQL+BJKvnxz4/Mzp4964gWBHol8gyuDpqdOjdlDc1ME+C8Amts0vBcut0DQFABBNNTwiLU825VmOcPNVlXZblNytvn+e604c5Axn5lBva5H42darmSj9Vk+6Bl/ihAH+aXZ/K2c1vE2up4vqAn924QjOswklupw9cAL2vK/ezFKRsYXbDm+LyVI1lbiJRsdDBib37/sOu4IsUHpRtPHGTPG2A1ALB+sjo4aXb+woh9/aUuN/IGWYVEep/IpHi/hIn76pF+nD592jlEHD8V03d/57uOrH7++ZeuN4rPoLIwdHUawK8osBJ0yi1dNT0ztrgaQDYnkiCy+vzRZquLrUBWV3u2tWKVvGxMVAifkJd88hI1j7+8PFeeEnCfoXu62+gPy+4SWFQlGN7nuOjnYr7QMkipSDW7EOtQ1MmbnwD1eNTOBn3GMVI5sU/l5SuranBuJWykCqug0mbKBZ9hd9mtZJnyvjJg1hAOVgNg6LU6214GPDt2zctHEDPsneWosGtsHmLKEDu/AcPLLG/FJxjpfWKJNCp4/0LPTXA6sDJZHZxI2J7dz9j+bcqJU5a70wpP1qplBRGhdxFCvZJcq+Gv9+BaSCG+jedy8z2lK6RTTRD9/fy1D0onuA8yokHlpxosvydb4hDInyfoK+Upp6LOwU7kPBbPLxXdg0VVdqh3VKTOn0SYjuDk7wep704bvFwI1c/vyxC5elmiE8hzeRnxmzh+fzlWOva4436RVQhqNe6FrDbWzSmBiA2NzlpHW7u9+Nw2i1Z6WTdSZ8eGhsTxWK8GwDA98DZHr2dBNZ0zcukCb/vyNSvANAAWCD965IhrufN2H9dfvdLvnDqf2CMOID3S4NqhwRHnHIAOuXtUBw/eMvbg2mDrNk85lsqFjgZmZbgQM1X2VUHHHnjQfepiJauL5ywe4bv0hSWiukyVcTY4rJnpcctNnZcwZ3WQjy8EJHV1yvBw4Ssc5vox5w+sRGiIt1LYSKB65/3pumjJ6mMFq49rK7nVKdBxxlfibgeemwAJRX70tDI1gLfuKR++fgbxA8zrpBeWYVPichyyWj0ffa0IyvFm/bkbeHlUb+kpI0/4FIg2vebOv1WVBfv0rPEc+CqmD/BcfjoIQ+iQ9f9/e2fW21aS3fHDndqpzba8yW532+nuydjJIAmQQV4CBEEe84lmPkyegzwFyEMQIEASBDNId89MZrrTi9uyJVu2dpEixTXnV8UjFimSIrWSEv/ual5dXt5bdc6pU/86tVxGlbiWnRKItvIbopz85iJtwfKHDsgb+WlHVA1Gpik7JLtT3pLRqkymNCVrai/YDPZS1bp+MSkdJn3WWP2Zk6mypKNlidfKjtT4JtPyDCkhGNMfZaLMlN3qsnU8iLQztxoSjz5JdGRMTmEC4fFNQbvAB6f6SXqXptTumtbUHkxRnHAv8plbuCUvXjx0gbVoTbvlQR0e4eJxdg99zoBwuijq11+7yMre/o7+rb1ddeTwycNiwVV8Ii+r2mCtvV1zezOyOCwERBbGzjw2v5XFxlGktVfcMB9xIvwcLhx3s/O+TDH5CQm8qIFEHjBhP6AfZsQc/Lj23go5JauHSgT1EiKrJN9fb7aZqwBEgAaN/NJo0YgNK9BE29SjT0cOLJaEfBExhazRKUU2ECDqPR1VPmncnz9/7sgsr1zFVywvLzuSdCqQyXM0ZLM/9Gs6xmdBLsk/CQIIKC/EHJIOkYHUUA6uZ6qDDaOTIL689IAoK/eCPHL+okAemXKFLpgXGxKrTrC8drvGo9VSAPW7kc4LdFz5F8L/rc+tqv+oaOKTkUE97ZL7tn+iajC90kFBp8gScoocLTJ+soxuLkzqaMm89UWl44j66W21uOxsMkd+Xp5+tqQcBHsoUsHdVSNcHsxDXDrQdaXiB1xCMGRPD56tqX7KW2W08bIG3CIPuaw69IO8fK2E9qPHH7m5qqwmM1jUdDYz616zxhy2bC6rvzm+xx/5YCeCEDiQkRNpD4Y9WBLmk49Qeqd+OckTUxbckPS45ldIE5Wk4cZGgNOh/kukpiU1/Uhy2xtqcG/cwjnIdnsHdbkg3wxt06BBQohW0chzfhgROv+m1EdVggxBxiChIfHkHeDM4aXhR8dEGvmexh//8Pnnn7uFST2Decy2wOpI3tYsni/QKXlkDiflgLgQBcZeIYEMq0NkIKCUB5JKVNJ8EPmzxDnKzX045j7MD+We4Dz8FvLlWeQNQmxytjyFRLT1eWFeT4ewtnv9nAdcfjSRW+Km1Sqkw8O9tcx1Vup55qKjYnkC2w1hWTmmQ4Wd0vFAr9g0iY7HSdHmEVQVEUYKopIvRySvdbSoqayJlwOcd/Ljc1FhbxjCH+wTQiqqjsq1jOxul2RusSKfPpuVTFL1VmPELqb2c362OUJvGIgFVoDfEBy10fhSxTtmyFAk6uce0ri8X9+SL3/7lTrQCT9EuLDorqPus4cq293cu7ekDUFJktpAeB/lM0O01e5vYP7c//xm2y2wmq2/bpUFVrXEffnZ82k39OzKEpa/U9l6xZD6qbAZ76rjywSy1Hxs72y76JUNT9Ig+ByrbVazsrf6O0nGCpJe+lMpRyb0bFySal1XOecIssJwNsOqRFuGHQ37aLYU4offrogkagfyyTLbFDWjnQaIPDLsD5gGAHHirUgQVnYAgKxBBhhaZ3EURCDEMTLAQ/VUOGc1nfCd13dbCXn69I/k2X31NmeYs9oJIZmBCDLED8mEAFIOiAzwfqw53+FvQ3AeG0dOdO5tugHkkmOew/dcd0wWJ4Df2n2xy3DI39ApXyH6fe5FIcxra54oa3iunzxzX7s3n9RnOg0c0/Eg4ZNG5LR3UPu+/iGrDbNfoFhWUmB9iNag0nmguSOtT3f1v+oW8u5slIRdBT77bFwmJxrBMMDPCIZcXesxZFAdXuyc1fok4lPBa/N4OgFEViGq7livr2hiiQ+zTiDUL9+8lXfvN9wiirt3brkFVmbDOAkibN9998ORc+DjNI4i7GgfA7drl645MBZLgwYaVGswmvXNO+JnZPr2khQO95S3buoZm0JwdbAoMBFCImnAGr4RlNhqRxOSSmQK0kq9ZpeAFy9euNGVly9fOqJKhxXS1zfqG49X673XhpM8f+s2u2SYH6JK2XjJAzs+YAOc75fQ2LXca2lpyRFV5ALBJFLLMyGc/cLywyfyNyI9zEBWligXhJLoJzhrnUPmdKSIpNJRQF5E0Pmk0wBZBTynWxoBGags9P+3FyYlMzMlM1qvZyczkpnUzrymmanMuSfuTzo6N7Eg02O3ZDo9LU8ezcvzF9MyPuEj65BTQ3g8wsnoZ3StE7pHVt/E5dUmkdV4X5FV0M3lQxPC3QBCAkR0lUiMg57ksTFOkk1lsv/wj//khvx+8uln6uDVZPQr8/G0O9tbu/LVb76SF89fKAlo34iF0VW2LGE3ACKrc0FkNRK/73YDSCcZOtIH9CLsXn1OL/caEODcAY0px3weodfyXiTIg8oTN2dz+Bi29I2Tz6DbEYDh/8qO7L/5WlJzDyQ+9VA1zTBz1BEia1QuEsiPhol8rq6uHkXVjHiA8HjY4C0FNI6cfPX//UZWAbJCZmaDzFsFdo7vIWvIrLXBPyZHbqGnjiKr1TWZHFN70Z99+BCVZx8/k4+XsQFsvPuwbwieS/L2dlx3EBlsknySsDM+Ab9joRTXMN+WSBygbL3YQVhuPrFjiCbJdkRAZlzHPZvqbgC7B3bJPfgkLyZvvm/9rf2mG3opw2UDOUAwSRYxDvNpx5QvPG/ltXMQXqbvoE9IKbIyWYN2v+2G8PqbCS8jwlLmH5CImuMRtLm/GNTv20SoyE7977BloJ2BD9ilnXzXCM1Ap0rN5F/OJbLq5m9pcp+XA9sNIHwiRBLjIPltibSYZT2Z16Sd4b/40+fy2bMnPvJpVl0HnDaTmZG7S3fddIBzQ31o4CaBqAwrrHnlJREsGrFOjd2Vom7bOHvyx9AwC0I8mfGOD9dSlQk1qnmZnLkjh9kNiVbe6Td+l4jLKhfPoXGjoYMMsHBm1Eh1BrIxcmekNDwHSTD52XeWToT2r/OFmuwfsANBVW4tKoFkuhGOpQ/wLPRqz4SYUHcgMkQ5ieDRIYEYoe+wHCSmgDCtAXuF1FLP+kF4L8gSw/ZEbIEtRjup7pJnyBsr1ckHeTWiCrh3K8LndkqDCORg0ySQd7d8tpJMZEndtXnmdIqZfsLUlJCotiKUSac0gsGvY3HWquKP6YElTPhCkoo/qm18TNuDmDoGOAcr/j0P8W0/GiJpTXefI1w+VFWDi1iN+YaalJGvf7kipY2CPLwzLzVeo9QB+JdHjx+5t/+cFgTkmqcA+CjOTQEklf0t375966I+NHo490EHDREkEJJA4wvoCVvOq6LkZvxjGU8n5WD9Sz1RdHq9rMYCGSJLm3cZEoIRLhdV7ciUinEplxbkxU8/k6lJXhURjhD1DiM+1BuG4CEyEESmdxA9t7mLrcSEY2wBwgMR4qUH3MfQj8+x+1IHIMU8jwgr9Zh7d6u/9j3ky6K+w1DfTwtkhYzwEUSR28HkiWyos+gVosr16ItOBrprp9cRTgPkxxxQ/aem53ZjIKn/bqTg/Hkn/WdBssbiYR/nHWl2MNCFrOKsenFY/Tg1rsU5+PlCQGmgS/64ntR6SG6T9yobNZZlYjktMrYhufV/llLuf4VXnx2l+vWkGq9HTcYkOa695+B8Uzp6Fr04JSzRolS0hpTV8bCpOflhNwL9zw0NuNmyEQgyv+gCX99OTgMEI3U0rjS2zA8kkkoDyy4KtjoZp0zjGaa2ZbuqxIc2GDQkNEQsXKLBjek/Bm580uN4QqITH0sl+YmUtn4vkdI7Jknrr4uqW7UDtc8TtHxq0Ogd5UvleVOAk0mpSHlNKqi5iPbpiGE/aNRzfVa04BZrlPXRCZV/pZiWg72yvPhsURbm45onrd/4my4uEZtHdyRn/wqO0asRVOqNRVIhfkd1pQusk4XdMjIQkp9+SFB4Lfckwkr9teeHeeaYT+o/8y3JO3XHOlD8/roCOSFvdGZvBzPZhbqFnEJS0S/6RD42FxX5hHrtR08jNGwR2dpeyZxjdJRvAtEeh363vbkt+VyDRxwD6ugnObSr/9e3Hlw6juR8OrTRhJ4KpwK4IfCLgb8zjRZHvvfEsVEL389RjMdl8t6cRMcjkpCsxGueYLUDQfqT/jXQKBvOphaJSq4Yke2DgiwspEU5r0NF88KrYMs1Nrm4XqBxwmmzgIXNxgHz59gGCJnQiNGIOvl09SCDAfLJYhOcH1FhPqNqz5YkooQpMS+TmUdSPMxJZf+tmgFzlT0iF0SiiNDQQEIgbPiadJ1A3fCp8c/A+zpyB2U50H5BvpyUg7LanfLDvFblHOlQj88pcS93P70vzzgoxSVbSbotcFi7mdsvqu5r8vOfP5fF+ajEtCPKXKmYklVnIx2AviApJOyKegPRg/BBZiCG1BW+70e3XItNYBssziG66uy2ThjNVlpTO4Tf0SFiAR95YuiaYX7ySp2mLmOTfNIZxTZBeP8wXSeYH0PWkCVIKbCyIiPkxXnbtouOh+nV5NEqF/uu9fwI7XGgZPP771koSfRfHUONNqb+pUHPHaUAY2Pjasfqq89V1ObBDA0/Zt+0phG6AX5l6exoLLDCeVXZsLoq1UJRDvIHfoEVW1f9bbLLAqvWzNR75u7/reA6vw777ZbIty8/yF/+bFq0n6pn/O/UKuufddDA6LOjmr9IJSel3S+Vb9ySyMSTtk6jP0dRddGWL77akp39PYkntiRXmpWnH38iy+yopVlheQhbuOayB+7eUxPDvzrWgNMmckA0lbKxKIPhLZwyDR3Ompcq0IgSVaABPb2sLw80wpAIyCqv7oQIGMiza6zYy6+yIdkP38lYQss7+3NVNnZ4oFaBRZo9nh48x2REfiANpG4YVJmeFtR1avTKusjrN4cyOZmSspuW2fAZUVflz9/1uwUT5k7YzBsiqmlj/VB+8klKbYNcFFTtfvhRqj7yGy7ADIH9Y1uQST6pI+jTOh+hvlvRi175Pc+gztGJpMPI32eNwnNfRk2Yxw/xgqQB6odFdG8STE/IlqlO6BMfB0nlO/wf5BS5WIdhhPMFsl5ZeeOmVdBh4OUfk5NKQMs1t8MPIMoawtVL6rOmqvY6d3d3HGlNj/uOVhOulxsdUjQUyBj2li2wevxEHt3Zk3EllalkSvmV1rceFlg1k9WKOmM1HsgqFfiblYi83snI3/9NXNId70GGQqvq3Mhbtefq1Q2Rl6+35c+fz0oiynBt4wEMwwP/f7+amG8Z0I3Kph6xittvGM69zuROtPj/+eWWvN86kHjkUD799J7cuT0miRhbBvM9PeWoFEuHTiY4dydg9+PhBo4ZR0EjZmQKJw3B45jhGRYrQWJpkM3JG3ppgK8C5BNdseKefLefv4zlFKVSfC+5vTVJRVWvU3e1UDNSi7EI5uzD9CYvInDMoyXiC+m379phUGV6FjhJ6/8OGbXTTyODR/tqtxfF+UDvffQYPTbxws+S+kW1qp0WVlg4Iqs6tx0k9GIjKhxDTm2Td6KoRlAtddIn6FWndg/qJDZje+9aRO80sHtCzCAI3Afyy3nOheW4aUAGkHhGlPB3dA74tMVSN1EmFw2zQ0by9NDJ/NWPr9wc/szsdNNe6x3Jah2lw5Lkta6Ms8Atobwj/H6kugFAQ4GwqfdK3f7137+Vj548lId3sjKRjrvdAM5GVrVVKZer8vXriLxcH5O/+6u0zE2469UWtAFuuiEZCq2qN7K6riz7D99syp/9sfbyg9vhW20LCYv8291dBEYt2EX+9GY1Rg70XOhT+vEvzHBgDdevv9mVze19+ZOn92V5Wc+5ObU8TJ0W5SEjkZIS1qJk9/JaMRIyO+33xrwOoAHGiUBY2b/y2bNnLgLDxuXIGvIKWp33IDtzykPkiLec8Tre1i2igLerglTLHyS3/q06vYwkZh6roU2q9r0dn7WM5MMicSetGr6+QNJMr0CmVFz9cCJgfrB5hfNG476dn2CexXD8SnRn+mNIGDIDLkOHDEXzvMZWbP0/MyTRHHMP5qQzosIete3qxXUEZbfyAwgTHQJ0yycdkadPnzryDm6CTK4C+EOi2bQ1Hz3+WMYn0vLt/32vRDUji4s+4t8VjgPUD7X67mf3ZWd7R9sopq1FJVqPyjZRlBGuCPhX72Mr6vvX3on82398Iz/56cdye3ZXxllbdBayWimVpVwsyGGhJCtbE/LVtyX56z9PyvJSY6jotGTVw1+7mYvKb/93W5amZuWwWpHyUUSjca/Gvmf+GW7qSjUmMW3w42qYGH5NiRZGaujHybDghmkvu/mc3J6LyGfLCRWayqBWkpp7fWtSS2r3zur1NXn/Th3bQU4eP3oi8c6t4FABJ86QP9MBaLzo7fI3EVeGDIkIho7eMOgOHRLOCmsapsePHx/NywNmZVgdW5ZUS+tS3H4p6WhJJPMzqcV8Z8TKSPlBv2W2OXGtb6gadNmdL5B2a2WBvDb00I1SngWNu+pRvZEz9xVGcczHcJ26FUdO0R3DlNgNJIZRh8vSG89n8Q+fjGyc9tlmtwYjadzbts26KUPdtBfIExmYXkkcs2sCnQJ8H7hZ9fNygPwZYaIThk0zV5o6+KP6aOz7/v0eXpeMOWuqlCtuf878QUFev16R27fvyExm+qhuN1GUEa4I5lM9Wf3+B5Evfv+jPPv0gSzOeLJK/aOu9UVWazUlfWoAvLI0f1iSqpK1tQ2RX3+1KZ9/Mi0/fTarBqUP1uuqkYQbmsfF+Z0sfYY8eiOrLFUqKi+IY3j6N0S0V+BH/Dw3vVv9EzCa1y/qL8tSYSk1dYS5GpDk0In7xhVyu7a2rg4/5+bZRFVOkFZKNYwu3xz4d99954YHIXWQvN/97nfukxcw2NB1K4bBoVM23p3OFkLMSzM064u/DqVa3JPczroqeUum5pZFUkt6nk4aq1RZLc5wrJ8egL12Kz8kgcTqfyKqEH6uPy3pvX4IfUbrX2fH8bp4/Ay2b+AFEuikVC7J3m7WnQuH+8Fl6gw7gVTZIh8i8qcllWZz1AW2s7I3phFV5J4QBdJ1g5UbIEvKi85pIJEnCaBXIs3MFWbKEB12cFNI/EUi9HcQVSKqvHGNukUgiM7imzdrUq1U5eGyf5FFV3A7TdbRZO7q5tami67SHh/N3rrp7nUAwML0mP6/qpwQ9vSr34i8W38jLz6fl8lUXsZSCTdKHdN6FovFTySrsV/84he/9Ic1t79ZRFvhqjrumiYW6B0elOTt2225u7QgY7zNSS/ykdW6A3efYVNzUgX3v+WqhBoWawf47Cvpj1mpzzGflo5d10Oy3zJ6QIXyUVqtRC6F//SiGhuI0xvw8znz+YISuUknbOqQl8hwgTIzLMPcLYYFIVY4aRo2SCrRl04LPC6z8T4tbH4e+mJqg/3d/J5pjrWxjk5INDkmKdmVjfXXMjGW8kaiDRwRf1del2qujvhfNssAuZm8eA4NInK0hs/bWPNvbiaQQSOd97/W+7eD6QI9lbWjvr+flYNc3pEVdGYE8Sp0xvN4PmSZIeowL/0gtEciiPyevyGn3JM56ewSYFMNrhNBozyUH98G6afM6BWZhuXkOr6DzDJ1CJ/HOdIIZwf1Cz0wxQyZmpzp+KfTKVfnCocFmc3MQkP0fP2HXWDXQHCw452dbXcOO+5S5Ue4EMD/EDifIRNS36r/V68jB8pW//uLgsxnynJnviappB/6ZwoARJVPAOvqhKMaG2GBAcPeUd5nnNTHlvSGVbeFU7aQlh/eVORQiRx7jvL4GMTW/fJmgWBMKpWUu3fvuUq3tbV9FJ0dNuBEiCjQWNniAoAjZ/sqG34cdtgOB0xroMydgCONJyYlOv0nsrj4UA4//I+U3/5apPje7UhRjBalFCnocVHrQdhBa8AaQRpHIjU4TyPIIwwWiLKhJ8gMCd0RdaQeDApZIQ/k0fbnNfRqT2Z71HMIA0ScOs05ykvdMDkYqR12MCLEFA50ypQHIqm2uJEyh3I0IGeuwacjq0HQ/XUC8iSaiu1xzDQUptOx+h+7RO6M6p7GTTLvMaN2/G59Xe/TeVvLES4K1CfW+fjRZw+WybMWJiqFkshv/3Aoh8WcLC6MSSquviauHWMlqOxn32tdO4qsehtRJqw/hA9XIzxYz2rPpRKNyw+vVmVyYkGmp7VHXo+18wz/u7DyN3qsF4mLdCXd6gtOHuFS6UhUOiKt9BCHwb3hqK2ho3fLPFWcdOvcNY5ParyGxaGTT8oCecRhQkZMj61geL8W0QqUmpT42KzKaEei5bwkYmUlsmmJuW2tGKwg9o68mu9hMqTx5/48zzAs8rquwPZNBzSQ6Agyg87oVNhw+yCB/BIJxH5d1EiB7VodNnSyLavvkAHs3spHmfkNPoz78gnJ4xzfGaHrdN9BgcnB6jNlwK+hV8oKOYesWr1sRVg+ZIBdEG3md/x+0Ms/DDCbsm3TGL1jOsD8/II7/2b1jezv7fu3vaV7CI60qAQVQVhZHL7PNJfpjJsiMMJlAr35Uaoqbx11iAuz537/SuRXX3wtf/wkLYtzUUmk6q/QJqIa1bY0mL/ZNbKKIbkKqcqtsReh+9Sb6Y0SMW1s0yIPH0wooZmS//rVqqx9UA7N9L1KY77nTQJ7MgJkNjk54SrfN998rQ4y5867zY0HGDgHa4zY2gmHzGr/To7Z7KNdGhaQVxY43b59262EpkJZI9cORE0rbI2WvC/T955JLZaUzdXvRXIb+m1WO5BZibCNhIImPeyqWSOPQ2ZDdsMwyeu6wnRO1I3RBIgNHTVsw6KpgwgcOw09nS3IVDfbbQUkF1LO7yFtoLWckGHqP/4AogaM3J1UV64K1DOrawBdEkkl/+Qd4gPhbNfhpvyWDBzzO+os92J+JdMmBrHsw4RQ1tgxMsaG0R12R2JUb25+zh2fCFMZn0FKpBIu4IKt53J+zvkIlwVIo/9Ez25YPxrXeiTyalXkiy++lcePluThnYiMj1UlGvfzVCGq/SD2SwUH6pKkxvBmpCbRWkxvpD1y9QXRmpLWeE1SYxNyUKjJ7/+QlfzhmExM8n5ltRNnPGFzfTldGvfYC0Kv7on5rWNjaa14fk5kOpV2e7AOeq8OR4FTx7kzT9VW5IVo/XuYYc6SKCcrUXGWDAm2K6OfZcNQEoRV/x+Zkuj4nKRnZiT77kuJZFe1h1eSSIq5UUyX8VPC7U40bkR2IEQ4X3v2CFcPyAc2j/2jfyKKRsoGGWZDRKMog80rD9HJxmz1P7/pdA02a8/Aj5lcwus7/faqQH7IN3WZukbkmHMQVOt4nEa3/A6Cy4gZNtLON45wNtARoEOwdOeu6iiitjnutq7ysj99fYyr3gpqB9Tx2bnZ+tkRLg3ubafaiazEZPO9KFesyH/9pijLD8bkyQORqYmIxJMxtyduQolqTAltGFUF3SKrEa3wjpvpI7SZZs6IVvIa7z0uu0UHxSKkNSKFsjqyg5KsvSvK69dbMjU+LbOL0zKTGZfpKe2da6fd6rR9unx0fvbpMQid3aCs2n7IykrWkZM7d4hS0NP339Oxb+frLrPDzrPIg+VDuZS8ebMtCwuzwo5K+PQgSOH+vjYI5Ewn4v37smzvflCSvuR2cTA9WJnZZcI2q0ckR99zrEIq7m+IlLISje1LZu6u6n5RCspt6SSy5iqiul99W5aYVsjMvL+v7VwRorHjRP9odz+HM9zzOsHsnE92C9GOvjZgStwOadBYhKHnUKh+j34H3fatPMrJlJwRtddytMxW4JpWn8LfylWPfp8IRliPrq8n62AXVEYcx1lvqH+bbJCL3ecqwdvP0KGbiab52dl0Hy6/7OhCHpEN+Q3l0U42baG/x5+/WtmQ8cmo3Jqfc/dvtYuTZNHx/jcRyAJ5aaqoPX33/ZpMTmTk/l0/rQXgP7Ft6qUF3EyGBM1AKPIWjuOfoWk/J/Ljyqosf3RPUmoT2ArP7AXH7jlCb1C5U2e0nyD5PCv+92Tjw6aUS4fy4N6iLN8b105fXtKpmERVIcxTTbJVlVaiVrLadesqI6sgOHTRBxI9+WglIqVKUQ5rBSlUxmRrLyFraxXZ3N6Vre0t16vlt7X6PlLNGTjZUoa25+p6Eh6EtCsqK+TAxGFkwUsEUEwDDY9X3zHswkHPFfmSr4rW2vFxbekU6AxYBIWeboih1UkrtONVqdDS8tpK1VekoHVLvSKo8jrFhn5aKw5Aj+68yqmqn7WKEtXSB3lwpyiZyTHZXC/K/sG0lKtT2pDGJV8qS3p8UmJJ3n3v3znutrxyCxg9athNYDt830B43B71jQgUzXVLc1g/utkIbTczH3NTmDY3d2V7M69qTEoiNqG6Yjjd19fQ7w2i3VueqKPthrWB7TXdtE91wMKp+0StjpdPf6f1IKI9oGispteU1dfn3FZCpVJEkgk/v3sgItBaZ5gPh0u99yAjSa1jb1+XJbt36PbcBJSZMpJCvYLwb2urmhDRtk5vw3D1QWFTEukDiVXvy2GB8RbvL0P5Ne8o0kC15bk3G2o3NXoRqje1rVgCHSLLuPrS1JFtIdeS+s54XY/A9NWqR9Bkx/jPWn1xsD5DYnvOlkuFpPP9RFw9OtmwP990zxH6Q40dNiKSHkvKwuK0LN6alKU7aZmJbysx9fJPJhNqA3GJq779wip3ugndRqU7klWOSZ6squKrRSmrcz/UBjFfYuP8mGT31TBK2igrna5WPdkhCuvvgwG1PrlzTobPUI4TA8q99nZN0ukxmc1knNNqVBSDl0Eo64sGEV/0CEFlmI9nM7m9dcjFdHDdKi2jBSCqZLVcPlCiUnBvPknEpmR6SvVUb+DawfTE99SDalQb8bjqtVqS+fS+TNR+kN3NohQj4xKZuCO7uRnttCRkLKX1QOXOfsWA33lUjyKjDdIZonMdaQdGQUZoBjpznUElN9MZbRRrRTnIFbXPntIv6ZyQvD64NqyLg2j7licfQKjJ1tamm5MZgmvCclhn2Dqr1P/w5Sk2996h6ucKRuPcvyAVte2Dg6w29HHJzLBbACT26u2M8lWU6MQTRW0U01LIs1v3uOaP8jQ63VzH36E8QDv5hIi6iqmkXhvXmuRlP7cuhdy4yuC2kqjmBZnhcTu0PvvmArshKVdQwrqf3ZR8gVXhS0pWfaif4A7y7EhW2+iqKbDgyCp1Wu8TVf1H9uXd+1VJRKclk5n3ttvRTzbOn6TTETqDZo7R9YS2selxtqVizRPEtKo65Timn5BVJaqR2JGsW0V+arJKxXU9dTWWmva0y3pYqlWUtEJWtfIqMa1WvXH5qKJ/cnifsJfJMGo7RAah134qNMpjEQ0IIa8ppTGZnppWxxcPeuCNyEYnGV0EmMxMvpjLQ2JR2ASLD+oOnee36mZ4ddIGOCoXycSetdHWP3d3d2RrY08ePFjWCkRFwtk1w2RjiNRXOVYiWgnVOSYjOYmW32rnTb9T5lmrZWVn+71zkKnErEQSt7SejAuBXZeQM+pXWdMuHiOrYbS1F5yCqF4rvdZB/YKkUDb8EFvg0Emk7tFIxuMQORW8kxejHvX62K+8rwjmMyr1Orqzve222MJuIaG2ONJ8CrJgoQkYn5hw8gnreJMNqEzMxHlMleW7KpdiUUlrmfmEfiQmeqzTfbkol0paLvVlNIpprYda/6pVv1ADuZiPtfraGvXs5mOb5eLbPPYTL5WysvJqU6Ymp2VhcU6vq7mo6wj9wsuVTtCrVy9lampC5hduqY5UmYEPa9Vdq87CUYOw4+Wg9uC/V93FylJQQvx65a3bftHvBtFJb437tNrMCD1CdUjnIcJaj0jR1VHORaMqdz1OxFPOV0FUGYEOp7G1irxnshrCn64bB2RVDcE13uowKtpol5W0umvqDQCvP3N/n6IBGM4eTedysniJSeSPHj1yk/3LlbI6Vd+bsDcgheiggnMFCy1YCc9qV9sc3xA+n+jNQAz5nQtMR40pDjhAnBLbnHz/3Us3EX9h/lZXG0QeVWfXlTrJTEi0qkRIz1Ujh3Ko5CcRyUu8vKGXvJPdrQ2ZnLoj0xP39RdJKYmSimhcykpyK1obiYZGbGJsiNOQp8DZn4Tro9dmYLPYM5vnc0zjxBBxa3TNHw+fDMw2KRtgURFlpR5TJr4PdWvfQzRtBwDQvm5zz3bn8PMVKeSL/m1DXerHeYNnUS6SEW+OvV6TEk9CVpM+MqeNI/C69QiP+4X9ls4n9fH9+pa8W38nTz956p6PnXW6/1mee33R8GnFUsFtlfj48SNJJVnwh91dRH30z3z95pUkkgmZm2UUsQNZDfzndfWPF4nQLzjxuS1POc/sU7+9ZzzWeAsg6OZKzk5W1aAikFVtdokusRCLXm5Fe7Y17q4Kh5A1/6Z3XKYjPDd0IRbIgW1m8gd5t9qeYSXmgPnex3GyetEg+sI78ml0IKvtHG6z0V2XSouOGkQVECFCHiC7f+De3nX37v2m/VBDICcXDXX3Qi6qR4a0aml33u1HrA0mQx67m3lJRrSCRssSl6xU8t9qRS1KNDEhpdiM1FKTEk3NaF2aVrLb2KaF+mSo1rfE6hl9kFVwnRwyZaETRjIi5shMPUpu5G7YYTrDbq2eWseTCCuELtQr5I6/sWm+M1idb/a3Pgjh0WwbELYff1xx/oII1WUBvZFvRoMoMySRRo98k1xZze6dH27ouRpE3/pGS12ikYVgraz86IIOt28vNY3AtPrQEVphuvA2yBviVtdW5eMnT50+bST23IEe1S6y2T23HdvS0m3t2GXqX7aiofPmejFC74DbVBw5dVOsVP7u7yiBobjjPL2KtjNZFfl/kLFDGmEl/GwAAAAASUVORK5CYII=)"
      ],
      "metadata": {
        "id": "4rgkmIgO_IIu"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pipeline: 1 - Ingest - Fase de Aquisição do Dados"
      ],
      "metadata": {
        "id": "o6egFcpQ81-3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame = spark.read.csv(\"stroke_data.csv\", header = True, inferSchema= True)"
      ],
      "metadata": {
        "id": "7JMDe23Eb1hs"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.describe().show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "v0kkUMbcEo0s",
        "outputId": "abbcd122-ab31-49e3-c386-e59c8bd38ce5"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+------------------+------+------------------+-------------------+-------------------+------------+---------+--------------+------------------+------------------+---------------+-------------------+\n",
            "|summary|                 0|gender|               age|       hypertension|      heart_disease|ever_married|work_type|Residence_type| avg_glucose_level|               bmi| smoking_status|             stroke|\n",
            "+-------+------------------+------+------------------+-------------------+-------------------+------------+---------+--------------+------------------+------------------+---------------+-------------------+\n",
            "|  count|             67135| 67135|             67135|              67135|              67135|       67135|    67135|         67135|             67135|             67135|          67135|              67135|\n",
            "|   mean|           33568.0|  null| 51.95950845311693|0.16410218217025396|0.10142250688910405|        null|     null|          null|113.41439606762462| 29.16154047813857|           null|  0.600089372160572|\n",
            "| stddev|19380.349498052576|  null|23.413054156327913|0.37037102916366954| 0.3018896147748789|        null|     null|          null| 51.25881719094036|7.1020570070927205|           null|0.48988334555668295|\n",
            "|    min|                 1|Female|              0.08|                  0|                  0|          No| Govt_job|         Rural|              55.0|              10.1|formerly smoked|                  0|\n",
            "|    max|             67135| Other|              82.0|                  1|                  1|         Yes| children|         Urban|            291.05|              97.6|         smokes|                  1|\n",
            "+-------+------------------+------+------------------+-------------------+-------------------+------------+---------+--------------+------------------+------------------+---------------+-------------------+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pergunta 1\n",
        "Quantos registros existem no arquivo?\n",
        "\n",
        "**Resposta: 67.135**"
      ],
      "metadata": {
        "id": "HU5fMiO63C3B"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pipeline: 2 - Prepare - Análise e Preparação dos Dados\n",
        "\n",
        "1. Análise de Missing Values (Valores Faltantes)\n",
        "2. Feature Selection (Seleção de colunas): adição ou remoção de colunas - features de acordo com suas relavâncias\n",
        "3. Análise dos tipos das features:\n",
        "    Integer - String - Categórico - Númerico"
      ],
      "metadata": {
        "id": "J8f7fEei_kU9"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Analisando as Features e seu Tipos"
      ],
      "metadata": {
        "id": "28dIrmIdAcde"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.printSchema()"
      ],
      "metadata": {
        "id": "IYCxcPKCaHRE",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c8cd7b20-7b05-498c-9421-7b159042930f"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "root\n",
            " |-- 0: integer (nullable = true)\n",
            " |-- gender: string (nullable = true)\n",
            " |-- age: double (nullable = true)\n",
            " |-- hypertension: integer (nullable = true)\n",
            " |-- heart_disease: integer (nullable = true)\n",
            " |-- ever_married: string (nullable = true)\n",
            " |-- work_type: string (nullable = true)\n",
            " |-- Residence_type: string (nullable = true)\n",
            " |-- avg_glucose_level: double (nullable = true)\n",
            " |-- bmi: double (nullable = true)\n",
            " |-- smoking_status: string (nullable = true)\n",
            " |-- stroke: integer (nullable = true)\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pergunta 2\n",
        "Quantas colunas existem no arquivo? Quantas são numéricas? \n",
        "\n",
        "Ao ler o arquivo com spark.read.csv, habilite inferSchema=True. Use a função printSchema() da API de Dataframes.\n",
        "\n",
        "**Resposta: 12 e 7**"
      ],
      "metadata": {
        "id": "mBt0YXU539Rd"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "total_colunas = 0 \n",
        "qtd_string = 0\n",
        "qtd_nao_string = 0\n",
        "for colunas in df_derrame.dtypes:\n",
        "  total_colunas = total_colunas + 1\n",
        "  if colunas[1] == 'string':\n",
        "    qtd_string  = qtd_string + 1\n",
        "  else:\n",
        "    qtd_nao_string  = qtd_nao_string + 1\n",
        "\n",
        "\n",
        "print(\"Total de \" + str(total_colunas) + \" Colunas \\n\" + str(qtd_string) + \" são Strings \\n\"  + str(qtd_nao_string) + \" são Numéricas\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "quz_eJ7t3Qa8",
        "outputId": "5002a401-de80-450b-9929-93e9ce8c662d"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Total de 12 Colunas \n",
            "5 são Strings \n",
            "7 são Numéricas\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Verificando Missing ou Null Values\n",
        "Null, None, NaN"
      ],
      "metadata": {
        "id": "b2KIM8pKet_x"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.sql.functions import col,isnan, when, count"
      ],
      "metadata": {
        "id": "1_WvsWdCeClE"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.select([count(when(isnan(c) | col(c).isNull(), c)).alias(c) for c in df_derrame.columns]).show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "L9ebUH3teari",
        "outputId": "f8454fde-5028-45ac-9d93-c6487a18767a"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+---+------+---+------------+-------------+------------+---------+--------------+-----------------+---+--------------+------+\n",
            "|  0|gender|age|hypertension|heart_disease|ever_married|work_type|Residence_type|avg_glucose_level|bmi|smoking_status|stroke|\n",
            "+---+------+---+------------+-------------+------------+---------+--------------+-----------------+---+--------------+------+\n",
            "|  0|     0|  0|           0|            0|           0|        0|             0|                0|  0|             0|     0|\n",
            "+---+------+---+------------+-------------+------------+---------+--------------+-----------------+---+--------------+------+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Análise Visual dos Dados\n",
        "\n",
        "Aqui já identificamos:\n",
        "1. Missing Values como pontos de interrogação nas features Tipo e Ocupação e deverão ser removidos\n",
        "\n",
        "2. Também percebemos, por conhecimento de negócio, que a feature Peso Final não tem importância para nossas análises e deverá ser dropada"
      ],
      "metadata": {
        "id": "OBsV7Fz3AmiJ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.show(5)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "omvwAvpQYnU6",
        "outputId": "656b47da-b611-4867-8c36-1ce95d57b35f"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+---+------+----+------------+-------------+------------+-------------+--------------+-----------------+-----+---------------+------+\n",
            "|  0|gender| age|hypertension|heart_disease|ever_married|    work_type|Residence_type|avg_glucose_level|  bmi| smoking_status|stroke|\n",
            "+---+------+----+------------+-------------+------------+-------------+--------------+-----------------+-----+---------------+------+\n",
            "|  1|Female|18.0|           0|            0|          No|      Private|         Urban|            94.19|12.12|         smokes|     1|\n",
            "|  2|  Male|58.0|           1|            0|         Yes|      Private|         Rural|           154.24| 33.7|   never_smoked|     0|\n",
            "|  3|Female|36.0|           0|            0|         Yes|     Govt_job|         Urban|            72.63| 24.7|         smokes|     0|\n",
            "|  4|Female|62.0|           0|            0|         Yes|Self-employed|         Rural|            85.52| 31.2|formerly smoked|     0|\n",
            "|  5|Female|82.0|           0|            0|         Yes|      Private|         Rural|            59.32| 33.2|         smokes|     1|\n",
            "+---+------+----+------------+-------------+------------+-------------+--------------+-----------------+-----+---------------+------+\n",
            "only showing top 5 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Análise Estatística"
      ],
      "metadata": {
        "id": "ZzvOsfqSBPya"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.select('age').summary().show()"
      ],
      "metadata": {
        "id": "C_18RjW6Ju2K",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "da7d3dfa-0cc3-4016-a2d8-ecd18059ed2b"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------+------------------+\n",
            "|summary|               age|\n",
            "+-------+------------------+\n",
            "|  count|             67135|\n",
            "|   mean| 51.95950845311693|\n",
            "| stddev|23.413054156327913|\n",
            "|    min|              0.08|\n",
            "|    25%|              35.0|\n",
            "|    50%|              56.0|\n",
            "|    75%|              73.0|\n",
            "|    max|              82.0|\n",
            "+-------+------------------+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.groupBy('hypertension').count().sort(\"count\", ascending=False).show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ohEAFgZdZooE",
        "outputId": "95d083ed-1bdc-42df-a741-08295f8658d1"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------------+-----+\n",
            "|hypertension|count|\n",
            "+------------+-----+\n",
            "|           0|56118|\n",
            "|           1|11017|\n",
            "+------------+-----+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.groupBy('gender').count().sort(\"count\", ascending=False).show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "InwcVaS9DmPS",
        "outputId": "2615697a-72e6-4646-e57d-bbe72b29e14e"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------+-----+\n",
            "|gender|count|\n",
            "+------+-----+\n",
            "|Female|39530|\n",
            "|  Male|27594|\n",
            "| Other|   11|\n",
            "+------+-----+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.groupBy('ever_married').count().sort(\"count\", ascending=False).show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ox03ttyPZ9dP",
        "outputId": "a98c616b-f44f-4571-d3aa-e3d6e3c25249"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------------+-----+\n",
            "|ever_married|count|\n",
            "+------------+-----+\n",
            "|         Yes|49323|\n",
            "|          No|17812|\n",
            "+------------+-----+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.groupBy('work_type').count().sort(\"count\", ascending=False).show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6NFHamkgdJC7",
        "outputId": "d704b812-459c-49d6-fe2f-52011730f974"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------------+-----+\n",
            "|    work_type|count|\n",
            "+-------------+-----+\n",
            "|      Private|37806|\n",
            "|Self-employed|14736|\n",
            "|     Govt_job| 8260|\n",
            "|     children| 6156|\n",
            "| Never_worked|  177|\n",
            "+-------------+-----+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.groupBy('Residence_type').count().sort(\"count\", ascending=False).show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ck9Q_gaLdSDz",
        "outputId": "87209fdc-e386-4c5b-98ae-7ff024f15fd3"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+--------------+-----+\n",
            "|Residence_type|count|\n",
            "+--------------+-----+\n",
            "|         Urban|34023|\n",
            "|         Rural|33112|\n",
            "+--------------+-----+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.groupBy('smoking_status').count().sort(\"count\", ascending=False).show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "n_-4ZlGTda1L",
        "outputId": "a9525b7c-72fb-4c1c-ff3e-3db54402f20d"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+---------------+-----+\n",
            "| smoking_status|count|\n",
            "+---------------+-----+\n",
            "|         smokes|35963|\n",
            "|formerly smoked|17880|\n",
            "|   never_smoked|13292|\n",
            "+---------------+-----+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.groupBy('stroke').count().sort(\"count\", ascending=False).show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VxasXWxPdkFa",
        "outputId": "fa924823-0c4f-4681-c02f-b55a7933e904"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------+-----+\n",
            "|stroke|count|\n",
            "+------+-----+\n",
            "|     1|40287|\n",
            "|     0|26848|\n",
            "+------+-----+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pergunta 3\n",
        "No conjunto de dados, quantos pacientes sofreram e não sofreram derrame (stroke), respectivamente?\n",
        "\n",
        "**Resposta: 40287 e 26848**"
      ],
      "metadata": {
        "id": "vDbgZyBC71fL"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pergunta 4\n",
        "A partir do dataframe, crie uma tabela temporária usando df.createOrReplaceTempView('table') e a seguir use spark.sql para escrever uma consulta SQL que obtenha \n",
        "\n",
        "**Quantos pacientes tiveram derrame por tipo de trabalho (work_type)**. \n",
        "\n",
        "**Quantos pacientes sofreram derrame e trabalhavam respectivamente no setor privado** \n",
        "\n",
        "Resposta: 23.711\n",
        "\n",
        "**Quantos de forma independente** \n",
        "\n",
        "Resposta: 10.807\n",
        "\n",
        "**Quantas no Governo?**\n",
        "\n",
        "Resposta: 5.164\n",
        "\n",
        "**Quantas são Crianças**?\n",
        "\n",
        "Resposta: 520"
      ],
      "metadata": {
        "id": "zhSOzjia8Vss"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#Criando uma View Temporária para trabalhamos com SQL\n",
        "df_derrame.createOrReplaceTempView(\"derrames\")"
      ],
      "metadata": {
        "id": "O-dyULKn8XXj"
      },
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.printSchema()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CKTOWYws9B0z",
        "outputId": "d4e0180f-f38d-4336-c8f4-84f0c80bafb5"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "root\n",
            " |-- 0: integer (nullable = true)\n",
            " |-- gender: string (nullable = true)\n",
            " |-- age: double (nullable = true)\n",
            " |-- hypertension: integer (nullable = true)\n",
            " |-- heart_disease: integer (nullable = true)\n",
            " |-- ever_married: string (nullable = true)\n",
            " |-- work_type: string (nullable = true)\n",
            " |-- Residence_type: string (nullable = true)\n",
            " |-- avg_glucose_level: double (nullable = true)\n",
            " |-- bmi: double (nullable = true)\n",
            " |-- smoking_status: string (nullable = true)\n",
            " |-- stroke: integer (nullable = true)\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Quantos pacientes tiveram derrame por tipo de trabalho (work_type).\n",
        "spark.sql(\"select work_type, count(work_type) as Total from derrames where stroke = 1 group by work_type \").show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MUCnCuaD9Hox",
        "outputId": "4f4cb3a4-9d33-4d4e-e66c-51890a435b0f"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-------------+-----+\n",
            "|    work_type|Total|\n",
            "+-------------+-----+\n",
            "| Never_worked|   85|\n",
            "|Self-employed|10807|\n",
            "|      Private|23711|\n",
            "|     children|  520|\n",
            "|     Govt_job| 5164|\n",
            "+-------------+-----+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pergunta 5\n",
        "Escreva uma consulta com spark.sql para determinar a proporção, por gênero, de participantes do estudo. A maioria dos participantes é:\n",
        "\n",
        "**Resposta: Feminino**"
      ],
      "metadata": {
        "id": "WcBeDg_WBpL_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Com SQL\n",
        "spark.sql(\"select gender, count(gender) as Total from derrames  group by gender \").show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fSVrxZTMB8Ld",
        "outputId": "94a4b4ed-e9d8-4587-8607-bc648c5137a3"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------+-----+\n",
            "|gender|Total|\n",
            "+------+-----+\n",
            "|Female|39530|\n",
            "| Other|   11|\n",
            "|  Male|27594|\n",
            "+------+-----+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Com DataFrame\n",
        "df_derrame.groupBy('gender').count().sort(\"count\", ascending=False).show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LjZSVZEEBcaI",
        "outputId": "6ea560b1-cacd-4b29-8909-6e247b2fcfc9"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------+-----+\n",
            "|gender|count|\n",
            "+------+-----+\n",
            "|Female|39530|\n",
            "|  Male|27594|\n",
            "| Other|   11|\n",
            "+------+-----+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pergunta 6\n",
        "Escreva uma consulta com spark.sql para determinar quem tem mais probabilidade de sofrer derrame: hipertensos ou não-hipertensos. Você pode escrever uma consulta para cada grupo. A partir das probabilidades que você obteve, você conclui que: **A hipertensão, neste conjunto de dados, diminui a probabilidade de derrame.**"
      ],
      "metadata": {
        "id": "1hlXe0-kCh-F"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "spark.sql(\"select hypertension, count(hypertension) as Total \\\n",
        "from derrames\\\n",
        " where stroke = 1\\\n",
        " group by hypertension\").show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mRzCerDmCnjf",
        "outputId": "ec29838a-4acc-4773-afd4-e75f59ee58fd"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------------+-----+\n",
            "|hypertension|Total|\n",
            "+------------+-----+\n",
            "|           1| 8817|\n",
            "|           0|31470|\n",
            "+------------+-----+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pergunta 7\n",
        "Escreva uma consulta com spark.sql que determine o número de pessoas que sofreram derrame por idade. Com qual idade o maior número de pessoas do conjunto de dados sofreu derrame?\n",
        "\n",
        "**Resposta: 79**"
      ],
      "metadata": {
        "id": "bZ67zkOKFuM_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "spark.sql(\"select age, count(age) as Total\\\n",
        " from derrames\\\n",
        "  where stroke = 1\\\n",
        "   group by age\\\n",
        "     order by count(age) desc\").show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LY_KjiC9FxBx",
        "outputId": "779d4f76-56e4-4c1d-e378-962405ea18c6"
      },
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+----+-----+\n",
            "| age|Total|\n",
            "+----+-----+\n",
            "|79.0| 2916|\n",
            "|78.0| 2279|\n",
            "|80.0| 1858|\n",
            "|81.0| 1738|\n",
            "|82.0| 1427|\n",
            "|77.0|  994|\n",
            "|74.0|  987|\n",
            "|63.0|  942|\n",
            "|76.0|  892|\n",
            "|70.0|  881|\n",
            "|66.0|  848|\n",
            "|75.0|  809|\n",
            "|67.0|  801|\n",
            "|57.0|  775|\n",
            "|73.0|  759|\n",
            "|65.0|  716|\n",
            "|72.0|  709|\n",
            "|68.0|  688|\n",
            "|69.0|  677|\n",
            "|71.0|  667|\n",
            "+----+-----+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Pergunta 8\n",
        "Usando a API de dataframes, determine quantas pessoas sofreram derrames após os 50 anos.\n",
        "\n",
        "**Resposta: 28.938**"
      ],
      "metadata": {
        "id": "7w-eR8OiHey2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.select('age').where((df_derrame.age > 50) & (df_derrame.stroke == 1)).count()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "t3M1oCHUHikY",
        "outputId": "8311b72d-dfb2-4080-b881-b0ccbb994ec3"
      },
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "28938"
            ]
          },
          "metadata": {},
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "spark.sql(\"select count(age) as Total\\\n",
        " from derrames\\\n",
        "  where stroke = 1 and age > 50\\\n",
        "     order by count(age) desc\").show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oPogutIRJ3F8",
        "outputId": "34db06e9-3702-485d-a0d8-889143a65ea8"
      },
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+-----+\n",
            "|Total|\n",
            "+-----+\n",
            "|28938|\n",
            "+-----+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pergunta 9\n",
        "Usando spark.sql, determine qual o nível médio de glicose para pessoas que, respectivamente, sofreram e não sofreram derrame.\n",
        "\n",
        "**Resposta: 119 e 103**"
      ],
      "metadata": {
        "id": "sq3F-XeAKVVN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "spark.sql(\"select stroke, (avg(avg_glucose_level)) as Total from derrames group by stroke\").show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bEneAAaNKaUW",
        "outputId": "f9824388-4e23-458b-b54c-ee7319afe65c"
      },
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------+------------------+\n",
            "|stroke|             Total|\n",
            "+------+------------------+\n",
            "|     1|119.95307046938272|\n",
            "|     0|103.60273130214506|\n",
            "+------+------------------+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pergunta 10\n",
        "Qual é o BMI (IMC = índice de massa corpórea) médio de quem sofreu e não sofreu derrame?"
      ],
      "metadata": {
        "id": "XdpLi2joMKVm"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "spark.sql(\"select stroke, round(avg(bmi), 2) as Total from derrames group by stroke\").show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "B4iFObHhMOSB",
        "outputId": "0a700251-6f85-4340-bb6e-eb3ac26b2fe6"
      },
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------+-----+\n",
            "|stroke|Total|\n",
            "+------+-----+\n",
            "|     1|29.94|\n",
            "|     0|27.99|\n",
            "+------+-----+\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Dados Categóricos x Dados Numéricos\n",
        "Ao aplicar algoritmo de Machine Learning sempre estaremos trabalhando com números, assim, precisamos separar as features categóricas e as numéricas, aplicando StringIndexer nas categóricas e OnehotEnconder na Numérica\n",
        "\n",
        "Por exemplo, a coluna sexo, como só tem dois tipos, sseus dados serão convertidos para 0 = Masculino e 1 = Feminino. Note que essa ordem pode mudar"
      ],
      "metadata": {
        "id": "GCcB_afAE1xA"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Importando StringIndexer e OneHotEncoder\n",
        "from pyspark.ml.feature import StringIndexer, OneHotEncoder"
      ],
      "metadata": {
        "id": "zb3001iugp9S"
      },
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.printSchema()"
      ],
      "metadata": {
        "id": "JfIrDn53JLqK",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5d080893-aa6f-43ed-f937-6cda6599e3ae"
      },
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "root\n",
            " |-- 0: integer (nullable = true)\n",
            " |-- gender: string (nullable = true)\n",
            " |-- age: double (nullable = true)\n",
            " |-- hypertension: integer (nullable = true)\n",
            " |-- heart_disease: integer (nullable = true)\n",
            " |-- ever_married: string (nullable = true)\n",
            " |-- work_type: string (nullable = true)\n",
            " |-- Residence_type: string (nullable = true)\n",
            " |-- avg_glucose_level: double (nullable = true)\n",
            " |-- bmi: double (nullable = true)\n",
            " |-- smoking_status: string (nullable = true)\n",
            " |-- stroke: integer (nullable = true)\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#criando a lista colunas_categoricas\n",
        "#colunas_categoricas = [\"gender\" , \"ever_married\", \"work_type\" , \"Residence_type\", \"smoking_status\"]\n",
        "colunas_categoricas = [\"gender\" , \"smoking_status\"]"
      ],
      "metadata": {
        "id": "FwofBwx6kqbp"
      },
      "execution_count": 32,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#verificando o tipo\n",
        "type(colunas_categoricas)"
      ],
      "metadata": {
        "id": "HfwIvbngIz2L",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d6eea573-0b93-4e7d-b84d-42d0d4dcc00c"
      },
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "list"
            ]
          },
          "metadata": {},
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#vendo os dados da lista criada\n",
        "colunas_categoricas"
      ],
      "metadata": {
        "id": "QK91mpndI0LJ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "35dc6916-2dce-4cf2-fde7-a9ea03132b87"
      },
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['gender', 'smoking_status']"
            ]
          },
          "metadata": {},
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YGqu3Ir5B9Rz",
        "outputId": "d5352034-0363-45a7-edbf-cb2d5dc6d225"
      },
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+---+------+----+------------+-------------+------------+-------------+--------------+-----------------+-----+---------------+------+\n",
            "|  0|gender| age|hypertension|heart_disease|ever_married|    work_type|Residence_type|avg_glucose_level|  bmi| smoking_status|stroke|\n",
            "+---+------+----+------------+-------------+------------+-------------+--------------+-----------------+-----+---------------+------+\n",
            "|  1|Female|18.0|           0|            0|          No|      Private|         Urban|            94.19|12.12|         smokes|     1|\n",
            "|  2|  Male|58.0|           1|            0|         Yes|      Private|         Rural|           154.24| 33.7|   never_smoked|     0|\n",
            "|  3|Female|36.0|           0|            0|         Yes|     Govt_job|         Urban|            72.63| 24.7|         smokes|     0|\n",
            "|  4|Female|62.0|           0|            0|         Yes|Self-employed|         Rural|            85.52| 31.2|formerly smoked|     0|\n",
            "|  5|Female|82.0|           0|            0|         Yes|      Private|         Rural|            59.32| 33.2|         smokes|     1|\n",
            "|  6|Female|82.0|           0|            0|          No|     Govt_job|         Urban|            234.5| 24.0|formerly smoked|     0|\n",
            "|  7|Female|33.0|           0|            0|         Yes|Self-employed|         Urban|           193.42| 29.9|         smokes|     0|\n",
            "|  8|Female|37.0|           0|            0|         Yes|      Private|         Rural|            156.7| 36.9|         smokes|     1|\n",
            "|  9|Female|41.0|           0|            0|         Yes|     Govt_job|         Rural|            64.06| 33.8|         smokes|     1|\n",
            "| 10|Female|70.0|           0|            0|         Yes|Self-employed|         Rural|            76.34| 24.4|formerly smoked|     1|\n",
            "| 11|Female|25.0|           0|            0|          No|      Private|         Urban|            91.15| 28.7|         smokes|     1|\n",
            "| 12|Female|43.0|           1|            0|          No|Self-employed|         Rural|            60.12| 34.2|formerly smoked|     0|\n",
            "| 13|  Male|72.0|           0|            1|         Yes|      Private|         Rural|           235.22| 40.3|formerly smoked|     1|\n",
            "| 14|Female|20.0|           0|            0|          No|      Private|         Rural|           106.47| 33.7|         smokes|     1|\n",
            "| 15|  Male|20.0|           0|            0|          No|      Private|         Urban|           104.78| 20.3|         smokes|     1|\n",
            "| 16|  Male|41.0|           0|            0|         Yes|Self-employed|         Urban|            159.3| 34.6|         smokes|     1|\n",
            "| 17|Female|23.0|           0|            0|          No|      Private|         Urban|           116.95| 23.8|         smokes|     1|\n",
            "| 18|  Male|22.0|           0|            0|          No|Self-employed|         Rural|            72.05| 31.9|         smokes|     1|\n",
            "| 19|  Male|69.0|           0|            0|         Yes|      Private|         Rural|            64.06| 35.1|formerly smoked|     0|\n",
            "| 20|Female|44.0|           0|            0|         Yes|Self-employed|         Rural|           135.03| 36.1|         smokes|     1|\n",
            "+---+------+----+------------+-------------+------------+-------------+--------------+-----------------+-----+---------------+------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Indexando as Features Categóricas - StringIndexer\n",
        "Note que além de indexar (transformar em índice / número), também vamos dar um novo nome para as colunas indexadas com esse código \" *x + \"_Indexada\" for x in colunas_categoricas* \", elas aparecerão com o sufixo _Indexada. Assim, teremos a coluna tipo e a tipo_Indexada. Veremos mais à frente"
      ],
      "metadata": {
        "id": "SzEdoAhpLrM6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "indexador_de_strings = StringIndexer(inputCols = colunas_categoricas, outputCols = [x + \"_Indexada\" for x in colunas_categoricas])\n",
        "#stringIndexer"
      ],
      "metadata": {
        "id": "BuzUdFuQk9v9"
      },
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Transformando em Binários - OneHotEncoder\n",
        "Transformação que fazemos nos dados para representarmos uma variável categórica de forma binária"
      ],
      "metadata": {
        "id": "ChAswNXxh7mA"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "meu_codificador_numerico = OneHotEncoder(inputCols = indexador_de_strings.getOutputCols() , outputCols= [x + \"_OHE\" for x in colunas_categoricas])\n",
        "#oneHotEnconder"
      ],
      "metadata": {
        "id": "7YU-aNXInBbS"
      },
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Separando nossa coluna alvo, stroke (derrame) e a indexando"
      ],
      "metadata": {
        "id": "eMeP1cMKNyuv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "coluna_rotulo_indexada =  StringIndexer(inputCol=\"stroke\", outputCol= \"stroke_rotulo\")\n",
        "#labelToIndex"
      ],
      "metadata": {
        "id": "-ndT43m1nulc"
      },
      "execution_count": 38,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Criando o Montador de Vetores - VectorAssembler\n",
        "\n",
        "VectorAssembler é um transformador que combina uma dada lista de colunas em uma única coluna de vetores."
      ],
      "metadata": {
        "id": "o9Bgk3HJN_29"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#importanfo a Lib VectorAssembler\n",
        "from pyspark.ml.feature import VectorAssembler"
      ],
      "metadata": {
        "id": "ShsHTsEToV06"
      },
      "execution_count": 39,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.printSchema()"
      ],
      "metadata": {
        "id": "zytLzgvIO0AV",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "47b7e14c-3a2d-4762-9e36-a483dcb050eb"
      },
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "root\n",
            " |-- 0: integer (nullable = true)\n",
            " |-- gender: string (nullable = true)\n",
            " |-- age: double (nullable = true)\n",
            " |-- hypertension: integer (nullable = true)\n",
            " |-- heart_disease: integer (nullable = true)\n",
            " |-- ever_married: string (nullable = true)\n",
            " |-- work_type: string (nullable = true)\n",
            " |-- Residence_type: string (nullable = true)\n",
            " |-- avg_glucose_level: double (nullable = true)\n",
            " |-- bmi: double (nullable = true)\n",
            " |-- smoking_status: string (nullable = true)\n",
            " |-- stroke: integer (nullable = true)\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Criando a Lista com as Features Numéricas"
      ],
      "metadata": {
        "id": "oEBEynfHP6f-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "colunas_numericas = [\"age\", \"hypertension\", \"heart_disease\", \"avg_glucose_level\" ,\"bmi\"]\n",
        "#numericCols\n",
        "colunas_numericas"
      ],
      "metadata": {
        "id": "n_UcrRiBobDK",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "518d968d-0b52-4bdd-d9ee-3e8479798ca0"
      },
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['age', 'hypertension', 'heart_disease', 'avg_glucose_level', 'bmi']"
            ]
          },
          "metadata": {},
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Juntando as Colunas Categóricas e as Numéricas em uma Única Lista\n",
        "### E acrescentando o sufixo _OHE (que vem de OneHotEncoder) \n",
        "### apenas para fins didáticos\n",
        "\n",
        "o sufico é acrescido com *c + \"_OHE\" for  c  in colunas_categoricas*"
      ],
      "metadata": {
        "id": "adIW6POyPRFN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "todas_as_colunas = [c + \"_OHE\" for  c  in colunas_categoricas] + colunas_numericas\n",
        "#allCols\n",
        "todas_as_colunas "
      ],
      "metadata": {
        "id": "kbypuDx1op3c",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a42dd38f-54d5-43c6-c595-69d62383a1dc"
      },
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['gender_OHE',\n",
              " 'smoking_status_OHE',\n",
              " 'age',\n",
              " 'hypertension',\n",
              " 'heart_disease',\n",
              " 'avg_glucose_level',\n",
              " 'bmi']"
            ]
          },
          "metadata": {},
          "execution_count": 42
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_derrame.printSchema()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3sSiK9KPrPs2",
        "outputId": "730f20f0-c665-4ae0-d86f-64c1bd30252e"
      },
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "root\n",
            " |-- 0: integer (nullable = true)\n",
            " |-- gender: string (nullable = true)\n",
            " |-- age: double (nullable = true)\n",
            " |-- hypertension: integer (nullable = true)\n",
            " |-- heart_disease: integer (nullable = true)\n",
            " |-- ever_married: string (nullable = true)\n",
            " |-- work_type: string (nullable = true)\n",
            " |-- Residence_type: string (nullable = true)\n",
            " |-- avg_glucose_level: double (nullable = true)\n",
            " |-- bmi: double (nullable = true)\n",
            " |-- smoking_status: string (nullable = true)\n",
            " |-- stroke: integer (nullable = true)\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Montando o Vetor com todas as Colunas (Features)"
      ],
      "metadata": {
        "id": "wdAdFnMOQVo0"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#montador_de_vetores = VectorAssembler( inputCols = colunas_numericas, outputCol = \"minhas_features\")\n",
        "montador_de_vetores = VectorAssembler( inputCols = todas_as_colunas, outputCol = \"minhas_features\")\n",
        "#vecAssembler"
      ],
      "metadata": {
        "id": "MadYU013o4pY"
      },
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pipeline: 3 Treinando o Modelo"
      ],
      "metadata": {
        "id": "KUTjNJlMRETU"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Importando a Lib DecisionTreeClassifier\n",
        "Classificado Árvore de Decisão - um dos vários Algorítmos de Machine Learning"
      ],
      "metadata": {
        "id": "tcEpX11qRYPr"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.ml.classification import DecisionTreeClassifier"
      ],
      "metadata": {
        "id": "fShofQm9pfyE"
      },
      "execution_count": 45,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Criando o classificador\n",
        "Informando qual a coluna rótulo (labelcol) - a que queremos prever: derrame_rotulo, bem como a lista de features"
      ],
      "metadata": {
        "id": "4K60h-7gRpYp"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "classificador_arvore_de_decisao = DecisionTreeClassifier(labelCol=\"stroke_rotulo\", featuresCol=\"minhas_features\")\n",
        "#dtc"
      ],
      "metadata": {
        "id": "Dm5_eSZOpnLO"
      },
      "execution_count": 46,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Criando o Pipeline do Spark\n",
        "Aqui passamos para ele tudo o que construimos lá até aqui"
      ],
      "metadata": {
        "id": "9x37IdFASWEC"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#importando a lib PipeLine\n",
        "from pyspark.ml import Pipeline"
      ],
      "metadata": {
        "id": "HMLDoCq8p1Pr"
      },
      "execution_count": 47,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "meu_pipeline = Pipeline(stages = [ indexador_de_strings, meu_codificador_numerico,coluna_rotulo_indexada, montador_de_vetores, classificador_arvore_de_decisao])\n",
        "#meu_pipeline= Pipeline(stages = [ indexador_de_strings, meu_codificador_numerico,coluna_prevista_indexada, montador_de_vetores, classificador_arvore_de_decisao])"
      ],
      "metadata": {
        "id": "gjFE1foEqMpa"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Separando os Dados em Treino e Teste\n",
        "Geralmente pegamos 70% (0.7) dos dados para Treino e 30% (0.3) para Teste"
      ],
      "metadata": {
        "id": "TOU3vzZCSri4"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#Vamos relembrar quantos dados gerais temos no df_derrame\n",
        "df_derrame.count() "
      ],
      "metadata": {
        "id": "I3ctdlPr7wMn",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "7f712511-0904-45c5-b1a0-de0f83a0c242"
      },
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "67135"
            ]
          },
          "metadata": {},
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#df_derrame_numerico = df_derrame.select(\"age\", \"hypertension\", \"heart_disease\", \"avg_glucose_level\" , \"bmi\", \"stroke\")\n",
        "df_derrame_cenario_2 = df_derrame.select(\"gender\", \"age\", \"hypertension\", \"heart_disease\", \"avg_glucose_level\" , \"bmi\",\"smoking_status\", \"stroke\")\n",
        "\n",
        "df_derrame_cenario_2.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xrhty8E8Sm7r",
        "outputId": "ce2a3c25-9568-4bad-bdf3-4d0923bcff61"
      },
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------+----+------------+-------------+-----------------+-----+---------------+------+\n",
            "|gender| age|hypertension|heart_disease|avg_glucose_level|  bmi| smoking_status|stroke|\n",
            "+------+----+------------+-------------+-----------------+-----+---------------+------+\n",
            "|Female|18.0|           0|            0|            94.19|12.12|         smokes|     1|\n",
            "|  Male|58.0|           1|            0|           154.24| 33.7|   never_smoked|     0|\n",
            "|Female|36.0|           0|            0|            72.63| 24.7|         smokes|     0|\n",
            "|Female|62.0|           0|            0|            85.52| 31.2|formerly smoked|     0|\n",
            "|Female|82.0|           0|            0|            59.32| 33.2|         smokes|     1|\n",
            "|Female|82.0|           0|            0|            234.5| 24.0|formerly smoked|     0|\n",
            "|Female|33.0|           0|            0|           193.42| 29.9|         smokes|     0|\n",
            "|Female|37.0|           0|            0|            156.7| 36.9|         smokes|     1|\n",
            "|Female|41.0|           0|            0|            64.06| 33.8|         smokes|     1|\n",
            "|Female|70.0|           0|            0|            76.34| 24.4|formerly smoked|     1|\n",
            "|Female|25.0|           0|            0|            91.15| 28.7|         smokes|     1|\n",
            "|Female|43.0|           1|            0|            60.12| 34.2|formerly smoked|     0|\n",
            "|  Male|72.0|           0|            1|           235.22| 40.3|formerly smoked|     1|\n",
            "|Female|20.0|           0|            0|           106.47| 33.7|         smokes|     1|\n",
            "|  Male|20.0|           0|            0|           104.78| 20.3|         smokes|     1|\n",
            "|  Male|41.0|           0|            0|            159.3| 34.6|         smokes|     1|\n",
            "|Female|23.0|           0|            0|           116.95| 23.8|         smokes|     1|\n",
            "|  Male|22.0|           0|            0|            72.05| 31.9|         smokes|     1|\n",
            "|  Male|69.0|           0|            0|            64.06| 35.1|formerly smoked|     0|\n",
            "|Female|44.0|           0|            0|           135.03| 36.1|         smokes|     1|\n",
            "+------+----+------------+-------------+-----------------+-----+---------------+------+\n",
            "only showing top 20 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#df_dados_de_treino, df_dados_de_teste = df_derrame.randomSplit([0.7, 0.3])\n",
        "#df_dados_de_treino, df_dados_de_teste = df_derrame_numerico.randomSplit([0.7, 0.3])\n",
        "df_dados_de_treino, df_dados_de_teste = df_derrame_cenario_2.randomSplit([0.7, 0.3])"
      ],
      "metadata": {
        "id": "n6739VrA7QUf"
      },
      "execution_count": 51,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Verificando se temos 70% dos Dados do DF_derrame no DF_DADOS_DE_TREINO\n",
        "df_dados_de_treino.count()"
      ],
      "metadata": {
        "id": "ApRjWZIx7d74",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "7a1ecc22-d78f-4812-bb3f-2db66b40526d"
      },
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "47167"
            ]
          },
          "metadata": {},
          "execution_count": 52
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_dados_de_treino.show(10)"
      ],
      "metadata": {
        "id": "r-gf_4TRYcb9",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a7e6255e-d89e-4386-cda6-85f9cee1a323"
      },
      "execution_count": 53,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------+----+------------+-------------+-----------------+----+--------------+------+\n",
            "|gender| age|hypertension|heart_disease|avg_glucose_level| bmi|smoking_status|stroke|\n",
            "+------+----+------------+-------------+-----------------+----+--------------+------+\n",
            "|Female|0.08|           0|            0|            67.55|31.2|  never_smoked|     0|\n",
            "|Female|0.08|           0|            0|           108.82|16.7|  never_smoked|     0|\n",
            "|Female|0.08|           0|            0|           111.09|12.2|  never_smoked|     0|\n",
            "|Female|0.08|           0|            0|           135.52|12.8|  never_smoked|     0|\n",
            "|Female|0.16|           0|            0|            60.67|13.8|  never_smoked|     0|\n",
            "|Female|0.16|           0|            0|            70.87|14.7|  never_smoked|     0|\n",
            "|Female|0.16|           0|            0|            81.42|15.4|  never_smoked|     0|\n",
            "|Female|0.16|           0|            0|            89.07|15.7|  never_smoked|     0|\n",
            "|Female|0.16|           0|            0|            90.15|14.7|  never_smoked|     0|\n",
            "|Female|0.16|           0|            0|           123.85|17.9|  never_smoked|     0|\n",
            "+------+----+------------+-------------+-----------------+----+--------------+------+\n",
            "only showing top 10 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Verificando se temos 30% dos Dados do DF_derrame no DF_DADOS_DE_TESTE\n",
        "df_dados_de_teste.count()"
      ],
      "metadata": {
        "id": "0CIPLJBb7ilv",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "26ec88b5-7fdc-497e-e717-54c8c2e8d210"
      },
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "19968"
            ]
          },
          "metadata": {},
          "execution_count": 54
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_dados_de_teste.show(10)"
      ],
      "metadata": {
        "id": "p7hpettyYYeW",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "31591dc8-80f7-479d-a22e-8496a89ff550"
      },
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------+----+------------+-------------+-----------------+----+--------------+------+\n",
            "|gender| age|hypertension|heart_disease|avg_glucose_level| bmi|smoking_status|stroke|\n",
            "+------+----+------------+-------------+-----------------+----+--------------+------+\n",
            "|Female|0.08|           0|            0|            93.97|13.1|  never_smoked|     0|\n",
            "|Female|0.08|           0|            0|           106.31|15.8|  never_smoked|     0|\n",
            "|Female|0.08|           0|            0|           125.11|12.1|  never_smoked|     0|\n",
            "|Female|0.08|           0|            0|           139.67|14.1|  never_smoked|     0|\n",
            "|Female|0.16|           0|            0|            98.52|13.1|  never_smoked|     0|\n",
            "|Female|0.16|           0|            0|            99.25|12.0|  never_smoked|     0|\n",
            "|Female|0.24|           0|            0|            57.62|14.1|  never_smoked|     0|\n",
            "|Female|0.24|           0|            0|            60.64|14.8|  never_smoked|     0|\n",
            "|Female|0.24|           0|            0|            89.24|13.8|  never_smoked|     0|\n",
            "|Female|0.24|           0|            0|            90.47|15.5|  never_smoked|     0|\n",
            "+------+----+------------+-------------+-----------------+----+--------------+------+\n",
            "only showing top 10 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Verificando se a soma dos 2 DF's é igual ao do DF_derrame \n",
        "if (df_dados_de_treino.count() + df_dados_de_teste.count()) == df_derrame.count():\n",
        "  print(\"Deu legal :) \")\n",
        "else:\n",
        "  print(\"Deu pau. Dá uma olhada no que pode ter acontecido aí\")  "
      ],
      "metadata": {
        "id": "2v7sYAOD70K3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6ab8e664-eed7-4f42-8e37-0508090357b7"
      },
      "execution_count": 56,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Deu legal :) \n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Ufa, Finalmente hora de Treinar o Modelo\n",
        "\n",
        "Fazemos isso com o fit (ajustar) do Pipeline"
      ],
      "metadata": {
        "id": "p6Gyo1HkUguY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "meu_modelo_pipeline = meu_pipeline.fit(df_dados_de_treino)"
      ],
      "metadata": {
        "id": "68x6kKje8pdj"
      },
      "execution_count": 57,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "minha_arvore = meu_modelo_pipeline.stages\n",
        "display(minha_arvore)\n",
        "len(minha_arvore)"
      ],
      "metadata": {
        "id": "IMUZkpNywcli",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        },
        "outputId": "f7ddc087-162a-4036-9938-8a139e6af70f"
      },
      "execution_count": 58,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "[StringIndexerModel: uid=StringIndexer_256454996623, handleInvalid=error, numInputCols=2, numOutputCols=2,\n",
              " OneHotEncoderModel: uid=OneHotEncoder_401d1cbed3cc, dropLast=true, handleInvalid=error, numInputCols=2, numOutputCols=2,\n",
              " StringIndexerModel: uid=StringIndexer_71d885d2bafa, handleInvalid=error,\n",
              " VectorAssembler_5cd644e6f249,\n",
              " DecisionTreeClassificationModel: uid=DecisionTreeClassifier_32f25071617c, depth=5, numNodes=31, numClasses=2, numFeatures=9]"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "5"
            ]
          },
          "metadata": {},
          "execution_count": 58
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Blz, Vamos, agora, aplicar nos dados de Teste o que foi Aprendido no Treino\n",
        "Fazemos isso com o *transform* do Pipeline e jogamos em um novo DF, o DF_DADOS_PARA_PREDICAO"
      ],
      "metadata": {
        "id": "0LDbYNptXWVI"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df_dados_para_predicao = meu_modelo_pipeline.transform(df_dados_de_teste)"
      ],
      "metadata": {
        "id": "--RkCdcNUVzI"
      },
      "execution_count": 59,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Agora, vamos analisar o DF com os dados de Predição DF_DADOS_PARA_PREDICAO"
      ],
      "metadata": {
        "id": "HR5Ejx6TYCMl"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df_dados_para_predicao.show(10)"
      ],
      "metadata": {
        "id": "A4KfVQfiWJc9",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "bffae3b8-736d-45e3-cd8d-1edc7d47f516"
      },
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------+----+------------+-------------+-----------------+----+--------------+------+---------------+-----------------------+-------------+------------------+-------------+--------------------+-------------+--------------------+----------+\n",
            "|gender| age|hypertension|heart_disease|avg_glucose_level| bmi|smoking_status|stroke|gender_Indexada|smoking_status_Indexada|   gender_OHE|smoking_status_OHE|stroke_rotulo|     minhas_features|rawPrediction|         probability|prediction|\n",
            "+------+----+------------+-------------+-----------------+----+--------------+------+---------------+-----------------------+-------------+------------------+-------------+--------------------+-------------+--------------------+----------+\n",
            "|Female|0.08|           0|            0|            93.97|13.1|  never_smoked|     0|            0.0|                    2.0|(2,[0],[1.0])|         (2,[],[])|          1.0|(9,[0,4,7,8],[1.0...|[43.0,8263.0]|[0.00517698049602...|       1.0|\n",
            "|Female|0.08|           0|            0|           106.31|15.8|  never_smoked|     0|            0.0|                    2.0|(2,[0],[1.0])|         (2,[],[])|          1.0|(9,[0,4,7,8],[1.0...|[43.0,8263.0]|[0.00517698049602...|       1.0|\n",
            "|Female|0.08|           0|            0|           125.11|12.1|  never_smoked|     0|            0.0|                    2.0|(2,[0],[1.0])|         (2,[],[])|          1.0|(9,[0,4,7,8],[1.0...|[43.0,8263.0]|[0.00517698049602...|       1.0|\n",
            "|Female|0.08|           0|            0|           139.67|14.1|  never_smoked|     0|            0.0|                    2.0|(2,[0],[1.0])|         (2,[],[])|          1.0|(9,[0,4,7,8],[1.0...|[43.0,8263.0]|[0.00517698049602...|       1.0|\n",
            "|Female|0.16|           0|            0|            98.52|13.1|  never_smoked|     0|            0.0|                    2.0|(2,[0],[1.0])|         (2,[],[])|          1.0|(9,[0,4,7,8],[1.0...|[43.0,8263.0]|[0.00517698049602...|       1.0|\n",
            "|Female|0.16|           0|            0|            99.25|12.0|  never_smoked|     0|            0.0|                    2.0|(2,[0],[1.0])|         (2,[],[])|          1.0|(9,[0,4,7,8],[1.0...|[43.0,8263.0]|[0.00517698049602...|       1.0|\n",
            "|Female|0.24|           0|            0|            57.62|14.1|  never_smoked|     0|            0.0|                    2.0|(2,[0],[1.0])|         (2,[],[])|          1.0|(9,[0,4,7,8],[1.0...|[43.0,8263.0]|[0.00517698049602...|       1.0|\n",
            "|Female|0.24|           0|            0|            60.64|14.8|  never_smoked|     0|            0.0|                    2.0|(2,[0],[1.0])|         (2,[],[])|          1.0|(9,[0,4,7,8],[1.0...|[43.0,8263.0]|[0.00517698049602...|       1.0|\n",
            "|Female|0.24|           0|            0|            89.24|13.8|  never_smoked|     0|            0.0|                    2.0|(2,[0],[1.0])|         (2,[],[])|          1.0|(9,[0,4,7,8],[1.0...|[43.0,8263.0]|[0.00517698049602...|       1.0|\n",
            "|Female|0.24|           0|            0|            90.47|15.5|  never_smoked|     0|            0.0|                    2.0|(2,[0],[1.0])|         (2,[],[])|          1.0|(9,[0,4,7,8],[1.0...|[43.0,8263.0]|[0.00517698049602...|       1.0|\n",
            "+------+----+------------+-------------+-----------------+----+--------------+------+---------------+-----------------------+-------------+------------------+-------------+--------------------+-------------+--------------------+----------+\n",
            "only showing top 10 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Observe que as colunas do Tipo String foram Triplicadas\n",
        "Por exemplo, agora temos as colunas gender, gender_Indexada e gender_OHE\n"
      ],
      "metadata": {
        "id": "ygazQkD-ZNAb"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#Criando uma View Temporária para trabalhamos com SQL\n",
        "df_dados_para_predicao.createOrReplaceTempView(\"predicoes\")\n",
        "#Dando Select nas colunas tipo\n",
        "spark.sql(\" select *  from predicoes order by 2 desc\").show(10)\n",
        "# Altere esse selec para ver outras colunas como educacao, educacao_Indexada, educacao_OHE"
      ],
      "metadata": {
        "id": "MrPzz0ZYaARC",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1b0152c6-995a-40fa-ec2e-311576152748"
      },
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------+----+------------+-------------+-----------------+----+--------------+------+---------------+-----------------------+-------------+------------------+-------------+--------------------+--------------+--------------------+----------+\n",
            "|gender| age|hypertension|heart_disease|avg_glucose_level| bmi|smoking_status|stroke|gender_Indexada|smoking_status_Indexada|   gender_OHE|smoking_status_OHE|stroke_rotulo|     minhas_features| rawPrediction|         probability|prediction|\n",
            "+------+----+------------+-------------+-----------------+----+--------------+------+---------------+-----------------------+-------------+------------------+-------------+--------------------+--------------+--------------------+----------+\n",
            "|Female|82.0|           0|            0|            56.54|20.4|        smokes|     0|            0.0|                    0.0|(2,[0],[1.0])|     (2,[0],[1.0])|          1.0|[1.0,0.0,1.0,0.0,...|[6576.0,294.0]|[0.95720524017467...|       0.0|\n",
            "|Female|82.0|           0|            0|            59.32|33.2|        smokes|     1|            0.0|                    0.0|(2,[0],[1.0])|     (2,[0],[1.0])|          0.0|[1.0,0.0,1.0,0.0,...|[6576.0,294.0]|[0.95720524017467...|       0.0|\n",
            "|Female|82.0|           0|            0|            58.02|37.2|        smokes|     1|            0.0|                    0.0|(2,[0],[1.0])|     (2,[0],[1.0])|          0.0|[1.0,0.0,1.0,0.0,...|[6576.0,294.0]|[0.95720524017467...|       0.0|\n",
            "|Female|82.0|           0|            0|            59.32|33.2|        smokes|     1|            0.0|                    0.0|(2,[0],[1.0])|     (2,[0],[1.0])|          0.0|[1.0,0.0,1.0,0.0,...|[6576.0,294.0]|[0.95720524017467...|       0.0|\n",
            "|Female|82.0|           0|            0|            59.32|33.2|        smokes|     1|            0.0|                    0.0|(2,[0],[1.0])|     (2,[0],[1.0])|          0.0|[1.0,0.0,1.0,0.0,...|[6576.0,294.0]|[0.95720524017467...|       0.0|\n",
            "|Female|82.0|           0|            0|            59.32|33.2|        smokes|     1|            0.0|                    0.0|(2,[0],[1.0])|     (2,[0],[1.0])|          0.0|[1.0,0.0,1.0,0.0,...|[6576.0,294.0]|[0.95720524017467...|       0.0|\n",
            "|Female|82.0|           0|            0|            59.32|33.2|        smokes|     1|            0.0|                    0.0|(2,[0],[1.0])|     (2,[0],[1.0])|          0.0|[1.0,0.0,1.0,0.0,...|[6576.0,294.0]|[0.95720524017467...|       0.0|\n",
            "|Female|82.0|           0|            0|            59.32|33.2|        smokes|     1|            0.0|                    0.0|(2,[0],[1.0])|     (2,[0],[1.0])|          0.0|[1.0,0.0,1.0,0.0,...|[6576.0,294.0]|[0.95720524017467...|       0.0|\n",
            "|Female|82.0|           0|            0|            59.32|33.2|        smokes|     1|            0.0|                    0.0|(2,[0],[1.0])|     (2,[0],[1.0])|          0.0|[1.0,0.0,1.0,0.0,...|[6576.0,294.0]|[0.95720524017467...|       0.0|\n",
            "|Female|82.0|           0|            0|            59.32|33.2|        smokes|     1|            0.0|                    0.0|(2,[0],[1.0])|     (2,[0],[1.0])|          0.0|[1.0,0.0,1.0,0.0,...|[6576.0,294.0]|[0.95720524017467...|       0.0|\n",
            "+------+----+------------+-------------+-----------------+----+--------------+------+---------------+-----------------------+-------------+------------------+-------------+--------------------+--------------+--------------------+----------+\n",
            "only showing top 10 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pipeline 4: Validar"
      ],
      "metadata": {
        "id": "aZsHwx7wcUDY"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Vamos Analisar as Predições?\n",
        "Como temos apenas cerca de 83% de acurácia (veja próxima seção) , teremos alguns imprecisões, ora Maior que 50 Mil será previsto corretamente, 0, ora com erro, 1"
      ],
      "metadata": {
        "id": "JpXudtQ4ayyT"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "spark.sql(\" select stroke , prediction from predicoes order by rand()\").show(10)"
      ],
      "metadata": {
        "id": "-4c9WDflsgMH",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f7e76c1c-7f96-492a-ca8a-882d16125e2b"
      },
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "+------+----------+\n",
            "|stroke|prediction|\n",
            "+------+----------+\n",
            "|     1|       0.0|\n",
            "|     1|       0.0|\n",
            "|     0|       0.0|\n",
            "|     1|       0.0|\n",
            "|     1|       0.0|\n",
            "|     1|       0.0|\n",
            "|     0|       1.0|\n",
            "|     1|       0.0|\n",
            "|     1|       0.0|\n",
            "|     0|       1.0|\n",
            "+------+----------+\n",
            "only showing top 10 rows\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Importando MulticlassClassificationEvaluator para Avaliações"
      ],
      "metadata": {
        "id": "ns5YLPT7cb6a"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from pyspark.ml.evaluation import MulticlassClassificationEvaluator"
      ],
      "metadata": {
        "id": "VriPknYgYFS7"
      },
      "execution_count": 63,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Vamos Criar um Avaliador usando a Métrica Acurácia"
      ],
      "metadata": {
        "id": "sl62NQUyclgQ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "avaliador = MulticlassClassificationEvaluator(metricName=\"accuracy\",labelCol=\"stroke_rotulo\")\n",
        "#evaluator"
      ],
      "metadata": {
        "id": "dfWfARigYXmL"
      },
      "execution_count": 64,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **Pergunta 11 - Cenário 1 - Veja o Notebook Cenário I**\n",
        "Crie um modelo de árvore de decisão que prevê a chance de derrame (stroke) a partir das variáveis contínuas/categóricas: idade, BMI, hipertensão, doença do coração, nível médio de glicose. Use o conteúdo da segunda aula interativa para criar e avaliar o modelo.\n",
        "\n",
        "Qual a acurácia de um modelo construído?\n",
        "\n",
        "Resposta: **Acurácia** : 68.71336313959264 %"
      ],
      "metadata": {
        "id": "Fc5T-jWDKtE9"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pergunta 12 - **Cenário2**\n",
        "\n",
        "Adicione ao modelo as variáveis categóricas: gênero e status de fumante. Use o conteúdo da aula interativa para lidar com as variáveis categóricas.  A acurácia (qualidade) do modelo aumentou para:\n",
        "\n",
        "Resposta: **Acurácia**: Acima de 80 (83.21726675556873 %)"
      ],
      "metadata": {
        "id": "EF5325AoLAdR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(f\"Acurácia :  {  (avaliador.evaluate(df_dados_para_predicao) * 100) } %\")"
      ],
      "metadata": {
        "id": "Nh9LIGNTYqtp",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d34feb58-abcf-4089-8a95-d57cf199feb2"
      },
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Acurácia :  83.87920673076923 %\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pergunta 13\n",
        "\n",
        "Qual dessas variáveis é mais importante no modelo de árvore de decisão que você construiu na questão (12)?\n",
        "\n",
        "1. Nível médio de glicose.\n",
        "\n",
        "2. **Ter hipertensão.**\n",
        "**Nota: A resposta correta deveria ser Idade, mas não aparece na lista**\n",
        "\n",
        "3. Status sobre fumo.\n",
        "\n",
        "4. Ter doença cardíaca."
      ],
      "metadata": {
        "id": "WkWuPbp-MLhB"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "va           = meu_modelo_pipeline.stages[-2]\n",
        "minha_arvore = meu_modelo_pipeline.stages[-1]\n",
        "\n",
        "importancia_de_features = list(zip(va.getInputCols(), minha_arvore.featureImportances))"
      ],
      "metadata": {
        "id": "-qIEL8IqMM7_"
      },
      "execution_count": 66,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "colunas = [\"Feature / Coluna\", \"Importância\"]\n",
        "importancia_de_features = pd.DataFrame(importancia_de_features, columns = colunas)\n",
        "importancia_de_features.sort_values(by=['Importância'], ascending=False, inplace=True)\n",
        "importancia_de_features"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        },
        "id": "9NR8P0t4Nmhq",
        "outputId": "61763703-1245-4439-a428-15e9336f0ee6"
      },
      "execution_count": 67,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "     Feature / Coluna  Importância\n",
              "2                 age     0.486544\n",
              "3        hypertension     0.330430\n",
              "4       heart_disease     0.173199\n",
              "5   avg_glucose_level     0.000489\n",
              "0          gender_OHE     0.000000\n",
              "1  smoking_status_OHE     0.000000\n",
              "6                 bmi     0.000000"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-928e2698-8596-41e8-be54-bfc4a342a141\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Feature / Coluna</th>\n",
              "      <th>Importância</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>age</td>\n",
              "      <td>0.486544</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>hypertension</td>\n",
              "      <td>0.330430</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>heart_disease</td>\n",
              "      <td>0.173199</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>avg_glucose_level</td>\n",
              "      <td>0.000489</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>gender_OHE</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>smoking_status_OHE</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>bmi</td>\n",
              "      <td>0.000000</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-928e2698-8596-41e8-be54-bfc4a342a141')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-928e2698-8596-41e8-be54-bfc4a342a141 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-928e2698-8596-41e8-be54-bfc4a342a141');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 67
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Pergunta 14\n",
        "Qual a profundidade da árvore de decisão da questão (12)?\n",
        "\n",
        "1. Entre 2 e 7.\n",
        "\n",
        "2. Menor que 1.\n",
        "\n",
        "3. Menor que 2.\n",
        "\n",
        "4. Maior que 5."
      ],
      "metadata": {
        "id": "6RCI7EEqVYQO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "minha_arvore =  meu_modelo_pipeline.stages[-1]\n",
        "display(minha_arvore)\n",
        "print(\"\\nProfundidade: \", (minha_arvore.depth))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        },
        "id": "1LRvwdekPzMd",
        "outputId": "86301bca-08d4-4469-b4c2-ceecc2fa7768"
      },
      "execution_count": 68,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "DecisionTreeClassificationModel: uid=DecisionTreeClassifier_32f25071617c, depth=5, numNodes=31, numClasses=2, numFeatures=9"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Profundidade:  5\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "kPsieJViVs-c"
      },
      "execution_count": 68,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Pergunta 15\n",
        "Quantos nodos a árvore de decisão possui?\n",
        "\n",
        "1. 1\n",
        "\n",
        "2. 3\n",
        "\n",
        "3. **Mais que 5.**\n",
        "\n",
        "4. 2"
      ],
      "metadata": {
        "id": "FAF-FldGXLQG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "display(minha_arvore)\n",
        "print(\"\\nNúmero de Nós: \", (minha_arvore.numNodes))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        },
        "id": "UN24w92TXYDF",
        "outputId": "7cb64fe8-45d2-442c-fd1f-09ad872e9958"
      },
      "execution_count": 69,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "DecisionTreeClassificationModel: uid=DecisionTreeClassifier_32f25071617c, depth=5, numNodes=31, numClasses=2, numFeatures=9"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Número de Nós:  31\n"
          ]
        }
      ]
    }
  ]
}