lectures/cap16_2_idmb_classification_self_attention.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyMV18Y7J2VyYU7UoTmyTRPH",
      "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/viniciusrpb/cic0269_natural_language_processing/blob/main/lectures/cap16_2_idmb_classification_self_attention.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Capítulo 16 - Mecanismos de Atenção\n",
        "\n",
        "Para esse notebook Python, utilizaremos a camada Layer do Keras, cuja documentação pode ser acessada [aqui](https://keras.io/api/layers/attention_layers/attention/).\n",
        "\n",
        "## 16.2 Classificação de reviews de filmes utilizando Redes Neurais Recorrentes e Atenção\n",
        "\n",
        "Informações sobre o corpus IMDB Reviews, clique [aqui](https://www.tensorflow.org/datasets/catalog/imdb_reviews)."
      ],
      "metadata": {
        "id": "32i_omQJGKIB"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![class_attention.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "6vVSbWH23-Eo"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Fonte: SUN, Xiaobing; LU, Wei. Understanding attention for text classification. In: Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. 2020. p. 3418-3428."
      ],
      "metadata": {
        "id": "vsfA-rUi3-Ru"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![FeedForwardAttention.png](data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAbkAAAD7CAYAAAD3s+r7AAAACXBIWXMAABYlAAAWJQFJUiTwAAAA8XpUWHRYTUw6Y29tLmFkb2JlLnhtcAAAGJVtUE1vgyAYvvsrCD3Lq67dJkEvM7stMT3p0VbWkgoYYZH9+4q6dtnKifd9vuBhjjrZS24b5GSnDHUZblp94HS6+zVgNFPsJcPVR4ne9MDRjmxJhPMAIcSG9pPui/dVPk0ZPlvbU4BxHMn4RPRwgjhNU4gSSJJwYoTmW9nGhcpsFpMfn4Kb4yB6K7RCfm4O+stmGK+c5SxB3Il7kjJkfjQ5agkegZhEcPP29n5LS+F4VxVCcmWmjDyOo4TBQ+ihtL7jL6/Pv5X1PyWDPz9a6oK1rzxgcOs+D65hZXk5mtOrigAAIABJREFUeJzsnXVcVuf//5/nDjoMBBtsUIdiYYDdgS3bdBabrdhzoWPq3FTsnjWMiU4dDsUEnTHRWQgWBmUBCkjecMf5/eGP+zsGKGV85nk+Hj4eeuK6rnPA63Xe1/UOQRRFEQkJCQkJif8gsnc9AAkJCQkJiTeFJHISEhISEv9ZJJGTkJCQkPjPIomchISEhMR/FknkJCQkJCT+s0giJyEhISHxn0USOQkJCQmJ/yySyElISEhI/GeRRE5CQkJC4j+LJHISEhISEv9ZFO96AP8V7t+/T/Xq1REEoUj3R0ZGUqlSJZRKZQmPTEJC4p+Iooi/vz8JCQnvdBxKpRI3NzfMzc3f6Tj+6whS7sriIYoifn5+GBkZ0bVrVwRB4NGjRwQFBZGeno6HhwcKxeu/Je7du8cff/zBpEmTCnS9hIRE0UhPT6feR44k1miB3NTi3QxCp+PFyd1c++s09evXfzdj+ECQZtNiEh4ezpUrV/j+++/1VpxSqeTOnTscP36ckSNHFqidGjVqULZsWfbu3cvHH3/8JocsIfHBI8gVWPWbiNKq4jvpX9SoSQs99076/tCQ9uSKyZo1axg8eDAy2f+9Smtra3Q6Ha6urgW2ygRBYMiQIfj6+pKUlPSmhishISHxQSGJXDGIj48nIiKCWrVq5Tp35swZXF1dC7VHJ5fLcXJyIigoqCSHKSEhIfHBIi1XvoY7d+5w+PBhlEol9erVQ6fT0b59ewBOnTpFgwYNkMvlOe5JSEggMjISjUbDxo0befjwIYMHD6Z27dqv7a9ly5b89ttv9O3bt8hOLK9DpVLx999/c/PmTTIyMihXrhwtW7bEzs7ujfUpISEh8S6QRO4V3LlzhzFjxrB582YqV67MoEGDsLKy0otcaGgo1atXz3XfxYsXUalUmJqa0rdvX3x9ffnhhx/w8fF5bZ+2trZERESg1WrfiANKdHQ0mzZtwtnZmV69emFiYsKTJ084ePAgRkZGjBw5MpdoS0hISPyvIolcPmRmZjJx4kQGDRpEtWrVEEURtVpN27Zt9dc8fvyYxo0b57r39OnTfPLJJ3Tp0gVBEHjx4gU6na5A/ZqYmJCamvpGRC45OZl169Yxc+ZMSpUqpbfaSpUqhb29PUFBQfj4+DBixAjJopOQkPhPIO3J5cO9e/c4d+4cbm5uCIJAQkICt27dwtXVVX+NKIq5xECj0XDu3Dk6d+6MIAjodDqOHz9O69atC9y3TqfjTUR2/Prrr4wYMYLSpUvnGrcgCHTo0IHExEQeP35c4n1LSEhIvAskkcuHe/fuUaVKFaytrfX/NjExoUyZMty5cweAMmXK8OLFixz3PX36lAcPHtCiRQsAHj58yJUrV3Bzc+P69et6ARs/fjwajSZXv1lZWRgaGubw1iwJ1Go1z58/p0aNGq+8rnPnzpw9e7ZE+5aQkJB4V0gilw+VK1fWL+npdDq2bdtGo0aNePjwIZGRkQDUrVuX+/fv57jv5s2bVK5cGUtLSwDOnz+Pk5MTSqWSixcvIggC0dHRpKWl5bkc+fjxYypWrFjiS5VpaWmYmpq+dr/NyspKCmGQkJD4zyCJXD44OjrSuXNndu7cyZYtW+jUqRNKpZITJ07QvHlzAFq1asWNGzdyLC3evXuXHj166MXEwcEBGxsbfH196d69OyEhIfzwww8YGBhw9uzZXMuSV65coW3btiVuyRkaGpKZmfnavcGMjAyMjIxKtG8JCQmJd4XkeJIPSqWS77//nsTERMzNzVEqlXTu3Bm5XK4XgVq1aiGTyYiJiaFq1aoAfP755wiCoN/zcnR0xNvbG6VSiaGhod5KmzRpEvb29jn6FEWRP//8k2XLlpX488hkMu7evUtycjKlSpXK97rg4GAaNGhQ4v1LSEhIvAskS+4VCIJAmTJl9EmTTU1Nc1g5giAwZswYdu/erbfIDA0NMTAwyNGOmZkZhoaGwMu8eU+fPsXOzo6srKwc150+fZrGjRtTpUqVEnuGpKQk9u7dy7x58+jSpQvr16/Pcy8QICYmhvDwcBwdHUusfwkJCYl3iSRyxaR169bIZDLCwsIKdH18fDxly5YlKSmJkJAQ/fHExEQOHTrEpEmTSsR9PyUlha1bt7JmzRoqVqzI3LlzqVmzJuHh4Xz33Xc8e/ZMf60oipw9e5YVK1agVqtRqVTF7l9CQkLifUBariwmcrkcT09PTp06Rf369V8rUGXKlKFBgwaEhobmCCu4f/8+M2fOxNTUtFjjiY2NJSAggOjoaPr374+DgwNyuZzg4GBOnTqFq6srRkZGbN26ldTUVGQyGRqNhoYNGzJ//nzu3bvHokWLmDZtmlQCREJC4n8eSeRKAIVCQceOHQt0rbm5ORMmTMh1vEmTJsUaQ2xsLHv27EGlUtGtWzeGDx+uF9wzZ85w8eJFevTowbFjx5g6dar+3L9j/erXr49CoWDx4sVMnz4dC4t3VIpEQkJCogSQRO5/GFEUiY6O5sCBA6hUKgYOHIitra3eM1MURYKCgggNDcXDw4MlS5bw5Zdf5hC1vCxPe3t7hg8frrfoSpcu/daeSUJCQqIkkUTuf5T79++zf/9+jIyMcHNzw9bWNodgiaLIkSNHiIiIYPz48WzduhU3NzfMzMwK1H716tX54osv8Pb2ZsqUKVhZWb2pR5GQkJB4Y0gi9z+EVqslPDycAwcOYGFhwdChQ7G2ts5ljYmiiL+/P7GxsYwaNYpr164BhV8StbW1Zfz48SxduhRPT09sbGxK7FkkJCQk3gaSyP2PcPXqVQICArCxscHDw4Ny5crleZ0oiuzbt4/U1FRGjhxJamoqBw4c4JtvvimS12bFihXx9PRk5cqVjBs3jkqVKhX3USQkJCTeGpLIvceo1WquXbvGoUOHqFmzJuPHj8fS0jJfsdLpdPqYvaFDhyIIAps3b2bw4MHFymJiY2PDlClTWLZsGaNHj9YHvktISEi870gi9x6i0+k4ffo0p0+fpnbt2gVy59dqtezYsQMTExMGDBiAIAgEBQVRoUKFXJlVioKVlRUzZsxgyZIlDBs2jJo1axa7TQkJCYk3jRQM/h6hUqkICgpizpw5pKamMn36dNzd3QskcFu3bqVUqVJ6gYuNjeXs2bMMGDCgxMZXqlQppk+fzvbt2/WVGCQkJCTeZyRLrpio1Wp92q+ikpWVRUBAACEhITRp0oTZs2fr04C9Do1Gw8aNG7G1taV79+7Ay325jRs3Mnz48GKP7d9YWloyc+ZMlixZQp8+faQUYBISEu81ksgVEp1Ox/Xr1/H19SU4OBhra2t2796NIAiIokhERASBgYE8fvwYQRCoUaMGHTt2zNMLMjU1lePHj3Pp0iW6dOnCrFmzCixu8FJg161bR926denQoQPwUuD8/Pxo3LhxvntnKSkpBAYGcvPmTdRqNaVKlaJdu3bUq1fvtaV44GUOz2nTprF8+XK0Wi1OTk4FHrOEhITE20QSuUJy7tw5RowYweDBg/n222+pW7cu8HLJcPfu3cTHx+Pm5oadnR06nY5bt26xadMmmjVrRqdOnYCX4vb7778TGRlJu3btmDt3boHE5Z9kZmayevVqmjVrlqNaeVRUFHfu3GHGjBl53nfjxg192Z9u3bphYGDAs2fPOHToECdOnGDChAm5EkznhampKdOnT2fZsmWoVCqaN29eIjk3JSQkJEoSSeQKyeXLl+nZsydeXl45JvUjR46gUCiYOHGiPuOIXC6nfv362Nvbs27dOgwNDYmIiODBgwf07duXjz/+uEjLiZmZmSxbtoy2bdvi7OysP67RaNi6dSsTJkzIUzSfPXuGr68vs2bNypEjs1y5cgwbNowrV66wYcMGJkyYUCDBMjQ0ZPLkyaxevRqtVkurVq0koZOQkHivkBxPColOp0OpVOaYzNVqNX/99Rf9+vXLs9ipQqFg+PDhrFu3jgYNGuDl5UXDhg2LJHDp6eksXryYzp0757CeRFHk119/pVOnTvnG0O3YsYNRo0blmQRaEAQaN26MTCYjIiKiwOMxMjJi8uTJXLp0icDAwFxFYCUkJCTeJZLIFZJ/JzQGCAsLo3HjxigU+RvG5ubmODk5UbNmzSJbO9kC5+bmlmsf7NatWyQmJtKyZcs8783MzCQjI4OKFSu+so8OHToQHBxcqHEpFArGjx/PnTt3OH78uCR0EhIS7w3ScmUhyUvkHj9+XKAA6XLlyjF//nwsLS0L3W96ejqhoaGUL1+eQ4cOcejQIf25bEuycePGLFy4MM/7VSoVoii+du/P0tKSlJSUQo9PqVQyduxYNm3ahL+/P7169ZKWLiUkJN45ksiVANbW1jx58uS11z1//pzp06cXuk5bcnIy3t7eLFy4EDs7uxznRFFk/fr1LFiwgHr16uXbRmZmJmvWrEGn0+W5pJpNamoqJiYmhRpfNjKZDA8PD7Zv387vv/9Onz59XtmXhISExJtGmoEKSV6WnKOjIxcvXkSn0+V7X3p6OtevXyclJQVDQ8MC/0lJSWH58uWMHz+eOnXq5Dp//fp1zMzMcHJyemU7arWa0NBQ4uPjX/l8AQEB6HQ6tFptkd6PXC5n2LBhZGRk4Ovr+8p3IiEhIfGmkUSukOQlcoaGhjg6OnL06NE896N0Oh07duygX79++Pv7s3DhQh48ePBaAYiPj2fp0qWMGTOGKlWq5DqfkpLCgQMHGDx4cJ5Lg6IoEhsby/r169mwYQNjx45l5cqVZGZm5nntnTt3SEtLo0yZMnz33XecOnUqz2tfhyAIfPLJJyiVSnbs2FFkwZSQkJAoLpLIFZK8RA6gT58+REdHs3XrVhITE/Vi9+jRI7y8vLh69SovXrxgwoQJjBgxgsOHD7No0SLCwsLyFMaHDx+yfPlyPD09qVKlSq4+tVot69atY+TIkRgbG+e6PyYmhtWrV7Nt2zY6dOjAV199hVqtRqfT8eOPP3L79m19vxkZGfj5+fHbb7/h6elJr1699KnFFi5cyB9//FFosZPJZAwcOBBLS0u2bNkiCZ2EhMQ7QRAlV7hC8dNPP5GSksIPP/yQ61x2NpQTJ06QmpqKKIqUK1eO5s2b88cff+Do6MiLFy8YPnw4MpmMhIQE/vjjD6Kjo+nXrx8ODg4oFAoePnzIunXr8PT0xNraOs9xBAUFERcXh7u7u14AdTod0dHR7N27F4VCQd++ffUOMcePHyc8PJzRo0eTlJTEwYMHiYqKAl56R7q6utKyZctcYQ0qlYoLFy5w/PhxmjRpQqdOnfIMQcgPnU7HsWPHePDgAaNGjXqlB6qExNsgPT2d+g0bIZu6DaXVq72N3xSiRk3k17259Mev1K9f/52M4UNBErlC8uOPP5KWlsb8+fMLdd/x48dJSkrCxMSEqKgoRo8erfd0TElJYd++fcTExFCvXj2uXLnC9OnTKVWqVJ5txcfHs3r1ar799lu9KN28eZNDhw5hZGTEgAEDqFChAvDS8jx48CCPHz/m888/L3RmlWy0Wi3nzp3jzz//pHr16ri5uRXKgSYwMJDQ0FDGjx9f4vk0JSQKgyRyHxbScmUhyW+58nW0b9+e+/fvU7duXerUqcOaNWtQq9XAyxi6YcOG0bNnT/bs2UNGRgY3btwgKysrVztqtZr169czZswYZDIZN27cYN68eZw/f56hQ4cyYcKEHALn5+dHfHx8sQQOXjqUtG7dmlmzZuHg4MDSpUvx8fEhKSmpQHFx7du3p0mTJqxYsSLP55KQkJB4E0hrR4WkqIavXC5n3LhxLFy4kNmzZ6NQKFixYgWenp4olUrCwsI4cOAAW7ZsQS6Xc+jQIX788UecnZ1p3769Pp/koUOHcHJyIioqii1btlChQgUmTpyYy+rT6XTs2bMHjUbDiBEjSixmTalU0qhRI5ycnAgNDWXNmjWUK1eOPn365Lu0Ci+dUVxcXFAoFCxdupTJkycXq5CrhISEREGQRK6QFNWSA7CwsKBHjx74+voydOhQDA0NWbp0Ke3bt+fIkSNMnz5dH6PWr18/unXrxtmzZ/n+++9p1aoVNWvWZOfOndSqVQsnJycmTJiAhYVFrvHodDp27tyJgYEBn3766RsJyhYEAUdHR+rVq8f9+/fZsmULpqamDBw4EBsbm3z7dHZ2xsDAAG9vb6ZOnVrkmDwJCQmJgiCJXCEpjsgBtGjRgkuXLnHz5k2aN2/O7du3mTJlCgEBAZiZmeW41tjYmE6dOtGmTRu8vLyYMmUKzs7OTJ06FSsrqzzb12g0+Pj4UKZMGfr06fPGs47I5XJq167Nl19+SVRUFNu3b0epVNK7d2+qVauW63pBEGjUqBFKpVIvdP9+bgkJCYmSQhK5QlJckRMEgS+++IL58+fTrl074uLiWLZsGcuXL8814atUKk6dOsWZM2eIjY1l1apVWFhYsGrVKuzt7enRowfm5ub68Wi1WjZt2kTlypXp0aPHW02rJQgCdnZ2TJ8+nYcPH+Ln50daWhru7u7Y2trmynxSv359lEolixcvZtq0aVhYWLy1sUpISHw4SCJXBIorHsbGxtSuXZsVK1awf/9+lEolpqameHt7M23aNORyOYcPHyYsLIyWLVvi7u5OUFAQnTp1QhAEmjdvzqVLl1i5ciWVKlWiV69eWFhYsH79eurWrUvHjh1L6EkLjyAIVKlShYkTJxIfH8+ePXtITU2le/fu1K9fX//uBEHA3t6ezz77jMWLFzN16lRKly79zsYtISHx30QSuUJSXEtOFEVOnDhBSkoKvXr14sqVKzg7O+Pg4EDfvn0ZNWoUVapUoW/fvvTs2RO1Ws3ChQuZNWtWjn6bNGmCk5MTt27dYt26ddy7d4/evXvrK4S/D5QrV45x48bx/PlzDh48yL59++jfv78+HhCgZs2aeHh4sGTJEiZPnpzvMuz/KiqVivDwO9wIvURaWioVK9rSwKkp5cuXL5a3q4SERMGQRK6QiKJYrKTD/v7+PHnyhHHjxqHT6fDy8qJs2bKcPXuWqKgo3N3duXTpkj5P5ebNm+nXr1+eAdhyuZyaNWtiYGBAz549uXv3LmvXrsXNzY3KlSu/F1UABEHAysqK4cOHk5qayt69ezlw4ADt27fH2dkZuVyOnZ0dY8eOZdmyZUyaNAkbG5t3Pexio9VqOXY0gDMBi7GzuEHNynLiE7WQLsfnqBKtZQ+Gjfq2QNUrJCQkio4kcoXkVZZcdsaTw4cP69NgWVhY6J0w/Pz8SE5OxsPDA0EQSE5OxtDQkNGjR7NkyRKGDBmCQqGgSZMmeHt74+rqioGBAQ0bNsyzv8zMTJYsWULnzp1p0qQJAJGRkezfvx+1Ws2gQYOoXLlyDlEWRZHnz5+zb98+Hj9+jCAICIJA69atcXFxeaOB2mZmZgwbNoyUlBSOHz/OnDlz6NatG87OzlSqVImJEyeyYsUKJkyYQNmyZTl16hTnz5/Xj9vW1pa+ffu+98uaOp2ODWu9MYxfztiuAgYKI+IStKzZm8ramaVpbA/xiX+wZn4wQyZs5yPHBu96yP/zaLVadDodCoXivfi4k3h/kDKeFJI5c+ZgaGjIN998k+N4VlYWW7ZswcDAgEGDBukdSOLj49m+fTtxcXHUq1ePwYMH8/TpU/z8/EhKSqJ///6EhYVhZmZGly5d9O3dvHmTIUOGcODAgTyTM6emprJkyRL69evHRx99lOv806dP2b17N2q1mu7du+Pg4IAgCJw/f54jR44wbNgwqlevDrwMMD98+DAhISFMnjy50KWAikpWVhYHDx4kNDQUZ2dn2rVrR0pKCkuWLAGgTZs2dOzYEYVCgSiKhIeHs3PnTvr06UOjRo3eyhgLiyiK+P3+G6lhk+jc7P+WI+MStMz5+QXrZ5XRH8vMElnlX4FZPx2jbNmy72K4/xn+/vtvvv/+eyZPnkzbtm1fmT6upDKeiFot2tREZIYmyAyNoRDiKmU8eXvIvby8vN71IP6XCAwMxNDQEFdX1xzH9+3bh7W1NQMHDsTQ0FB/3NTUFGdnZyIiIihVqpTeoaRXr1707NmTcuXKUadOHXx9falZsyYWFhaIosi2bdsYM2YMO3fupHbt2jni4dLS0li0aBGDBg2iXr16eX65mpmZ0bx5c6pXr05QUBAHDx4EXua8/OqrryhXrpz+2uwwgPLly7Njxw5atmz5Vr6G5XI5Dg4OODs78/DhQ3x8fPSVEz7//HOaNm2qt0Kzlz2dnZ3ZsmUL9evXfy9j7FJSUvjt51EMapOKTPZ/7zAtQ+Tk5Ux6uvxfMm2FXKBiqRccvaSkqbOLZIEUg7t37zJ79mz27NlDcHAwlStXplKlSnnue6rVatau34DQoi9yk8J/0ImaLJJP/05CwGY0SfGkh50jM/oORjUbFvxnqNORFOjLqE/6vzKJgkTxkdJ6FZKEhIRc+2MqlYqwsDC6du2a5y+5TCbjs88+w8/PD3d3d6ZNm0b16tX1E7hcLufzzz9n06ZNaDQajh8/jq2tLS1atGD69OmsX79en0w5ISGBn376iWHDhuUrcP+kfPnyeHh44OnpqS+3808RzkYQBOrUqUO5cuW4fft2UV9PkciOB5w3bx5xcXFUr149zxg7ABMTE4YNG8aePXve6hgLyt8Xz9OwSiQKecEmu6rlZcTf9yM9Pe0Nj+zDQKVScfjwYTp06EDfvn05ceIEGo2mxNrXqTN5smEWyecOYOMxjzK9RiE3L8vz/avRvnhWYv1IlBzSntxrEEWRmJgYTp06RVBQEH/99RdjxozJcU1oaChNmzZ9peAYGxvj6OiIsbFxnrkby5YtS5MmTdiwYQOxsbHMmjWLrKwszMzMmDJlCt7e3gwcOJA9e/YwYsQIqlSpUqgckDKZjI8++ui1y2IuLi6cO3eOGjVqFLjtkkQURdq2bfvKa+zs7Hj69OnbGVAhuRl2kdZVFWRk5qwVqMoS0erIdRwELGQPSExMQi5XlOiE/CGRkZGR499arZZDhw4RGBhI+/bt+fLLL2nZsmWxqmCIoshzv7WkXQmi6rz9yI3NEHVa0m+cx6C8LXKzvBOqS7xbJJF7DWq1Gnd3d4KDg6lQoQIbN27MtQcWGxurT4r8KsqWLcuiRYuwtLTM87xOp2Pz5s20adOG5cuX5ziXkpJC3759admyJfv37y/0c6hUKoDXeoaamZlx7NgxHj9+XOg+SoKQkBA+/vjjV14jCAJqtZpHjx5RoUIF4uPjsbS0fC9yYaYkJ3EzVc22gJyWWVaWyK1IDWMXJuY4bmEio1UDQ9LS0tiyZQuBgYFvc7j/GeLi4vLMK6tSqQgICOCvv/5i165ddO3atch9qGOjSQzYgmXrfhhUeLnSIMjkVJq6DoSXf5d4/5BELh9EUeTKlSu8ePGCYcOGsWDBAk6ePMnkyZMpU6YMLVq00F9bqVIlIiMj9R6O+bUXHx/Pt99+m6djhyiK7Nu3j59++okbN24wffp0fVLmuLg4li9fTmBgILt27aJHjx44OjoWag8n2xNTq9W+Mj4rKSmJnj178umnnxa47ZLEx8eHuLi4Vzq/ZL/LCRMm8Omnn5KVlUVUVBQqlYpp06ZhaWmJTqcjPT0dpVKJUqksVthHYShV2gr7Ukp+/qpMjuNxiVrmbk5m9fTcnqG+J7SYm5sza9Yspk2b9lbG+V/j9OnT9OzZM5fQmZqaMnjwYCZNmoSDg0Ox+ki/dQFdegpmjTrk+L8nSPGO7zWSyOWBTqfj119/ZdOmTdjY2NChQwfatWtHu3bteP78OdeuXcshch999BF79+6lV69e+S6HpKSkcO3aNcLDw3Fycso16T548ICIiAimTZtGtWrV2LZtGx4eHkRHR7NhwwamTZtG2bJlmTlzJkuWLEGtVr9SVP9NTEwMISEhPHz4EFtb23yvCwwMpEePHgVutyRJSUkhMTGRo0ePMnbs2HxFPCwsDBcXFz777DOCgoK4cOECWq2WpKQk/vrrL1q2bImhoSE7duzg+fPnWFhYMGrUKAwNDdFqtaSmpuaZ2LokaODUihtBS+jULOe+p4FCQCaAgfLfFd5FUqlD6dKlMTAw0H/YSBQOY2PjHP+2tLTE3d2dKVOmYG9vXyJ9qJ9GIsgVGFSpUyLtSbwdJJHLg9u3b3PhwgVGjRqFWq3OIQpyuTzX16JCodAvIw4YMCCXgGm1Wn7++We++OILbt26hZ+fH/369cPR0RG5XI5Go2Hr1q1MnjwZmUxG06ZNuXTpEocOHeL8+fNMnTpVv5dmZGTE9OnTWb58OVlZWbRo0eKVcXuRkZH4+vpSqlQpvvnmG37++We+/vrrXM4zoihy+fJlRFHEzs6uBN5iwRBFkbS0NPbv38+DBw/o2rUrJ06c4MaNG3k61iQnJ+Pr68vMmTORy+V06tSJjh07Ehsby/Hjx1Gr1Zw8eZLevXszZswYRFFErVbr4/+ylwWTkpKoWrUqw4cPRy6Xl1icVUOnxhzaVYc26ohcgpYXd2O0VK474L1Yav0vYGJiwqeffoqnpyd169YtUQteUdoGZDJkRrn/76RePoFxrUYoLKVQkPcNSeTywN7enkaNGvHxxx9z4MABateuneN8Xmv/3bp1Y/fu3axatYru3btTrVo1tFott2/fxt/fH6VSSWhoKOPGjUOj0bB37178/f1p06YNt2/fplevXvqUVoIg0KJFCz7//HMCAgJypboyNDRkypQprF69mqysrFyOGqIoEhISwtGjRzEzM2P06NF6kTQwMOCnn36iS5cuNGzYEBMTE548ecKxY8dITExk4sSJb82VPT4+Hn9/f2JiYujTpw9Dhw4lIiKC9PR09u/fT0hICB06dMDa2prU1FQuX75MYGAgHh4eOfY1BUGgfPnydOzYkdWrVzN06FAuXryIra0tt27dIi4ujg4dOmBlZYWFhQVTpkzRv6dsIiIi8PX1xdjYmKZNm9K6desiPZOpqSkd+8/lwNFxbRR4AAAgAElEQVRhDGire+W7fJGqwz+kPt97e0jhA8WkdOnSeHh4MHXq1BKz3P6NaeMOKA5uJOP2Rcyc2gGgVaWRfOZ3dJkZmDV+dzljJfJHCgb//2RmZnLt2jWaNWumz0ZiYmKCKIo5vu49PT2pVasWEyZMyLOdBw8ecPr0aR4+fIhcLqdatWq0bdsWGxsb/vzzT65cucLEiRNRKpWkpqayefNm/Pz8mD17Nq6urigUCsLDw/n111/p1q0bwcHBeHp65jkJZlcJt7e3p2PHjuh0OsLCwti3bx+1atWia9euWFlZ5WkN/fnnn9y+fRuVSoWVlRWtWrWiXr16bzyfoiiKJCQksGvXLlJTU+nVqxf29vbIZDIiIyPZvHkzU6dOxcLCgpCQEIKDg3n+/DkmJibUq1cPV1fXV5bmSUpKYu3atfTp04d9+/ZRt25d7OzsOHXqFP369cs3NCF7bCkpKWg0GsqUebmndvr0aYKCgrC1taVjx455Bub/G51Ox+5dW4n5ew7u7UQMDSA+ScfcTcmsnlEaURSJjtXhe646E7/aTrXq78aT9b9E9jRWkI+FogaDi6KI6l4IiUd9UJQpj6BQIsgVmNRvhXGthoVyPJGCwd8eksjxclJatmwZLi4uejf/+fPnM2nSpFwlYCZPnkyNGjWYOHFikfo6d+4c586dw9PTE51Ox7x585g1axanT5/m77//pmrVqkRGRvLNN99gZGTEjh07qFKlCm3atMmzPY1Gw4YNG5DJZDx79gxbW1sGDBjw3gVKi6JIdHQ0/v7+pKSk8PHHH+cQnJs3b+Lr68uXX36ZZ57OwvaVzV9//UWTJk1yxAampqYSHh6Oo6NjgVzKdTodUVFRGBgYUKlSJQCOHj1KamoqderUoVatWrliD0VR5OrVq+zbsYCysotUKZtOukoEQSAqoSJmlfvxxdhpby27jMT/UVIZT4qDJHJvD2m5Ejhy5AgNGjSgXLly7N69my5dumBubo5Op+PFixe5lsaK813QsmVLDAwMWLJkib5yd3bF8LJly7Jx40asra05fPgw3bp1Y8CAAfzwww80bNgwV+iBRqPh3LlzPHz4kIyMDBo1asRnn332Xi196XQ6Hj16xM6dOzExMcHNzQ1bW1v9GEVR5ObNm+zdu5cZM2YUW+Ag59d8q1at9P1ke5YaGBgQERHBvn376Nu372sdeGQyWS4LsFmzZkRHR3P79m3KlClDxYoVEUWRGzduYGdnh7GxMU5OTnz00S7++OMPvvpqFqVLWVK33kd8++03VK9e4736OUlI/FeRRI6X+2larZZ58+Yxffp0rl+/zsCBAzl8+HCe4lIckRMEgaZNm3L37l327NnD2LFjgZcWx5kzZ/j555+Ry+WcO3eOH3/8EQcHB9zd3Vm9ejWzZs1CLpejUqk4efIkwcHBNGrUiLlz5yKXy9mxYwe//fZbns4vbxudTsft27c5dOgQSqUSDw+PHKnEsrl8+TKHDx9m1qxZeWZiKSrp6emEhITovWB1Oh2rVq3Czc2NGjVq0L9/f9zc3IiJiSlS+6VLl6Z06dI0aJAzufLjx48JCAhAEAS6d++Ov78/165dY8yYscjlcvr06cMvv/xCnTp16N27dy6vQAkJiZLlg85dqdO9zD4hCALPnj1DoVBQt25dqlatirm5OQcPHqRr1645MvMfO3aMUqVK0bx58yL3m5KSwv79+xkzZgxbtmwhKyuLq1evMmnSJAwMDBAEgapVq+Lq6kpmZiYHDx4kJCQEIyMjbt++rc9z+cknn1C/fn3kcjmCIPDRRx8RFhZGSEhIjgKlb5NscVu7di3JyckMHDiQjh075unN+ffffxMYGMjUqVNL3LvwxYsXHD9+nGbNmgEvrbG6deuyevVqvcONXC7XVzTQarVs27YNc3NzSpcuXaR3JwgC1atX19f5u337Nvb29mRmZpKWlsb58+epXbs2n332GSkpKWzevBkTExMqVaokWXVvkeLmriwRpNyVb40P1pLT6XRs3bqVzz77DAMDA6ytrXNlQ8iu1fZPirtcqdFoWLVqFWPGjKFq1apcv36dH3/8kd9//z2XJaNQKGjatCl16tRh+fLlzJw5k8aNG7NkyRLKly+fa2KUyWS4u7uzf/9+fHx8GDp06FsrzKnRaLh8+TInTpzAxsbmtUuPp0+f5uLFi0yfPr1YqZby4/Hjx7kmD0tLSzw9PXnx4kWu9GZyuZz+/fuzZcsWSpUqxZAhQwr97rKysjh27BgXLlzgk08+QaFQ4Ovry+zZs0lNTWXTpk1Uq1YNhUJBq1atKFeuHPPnz2f37t2MGzculxevhIRE8flgEzTfv3+frKwslEoloijqhevMmTOkpKQAMHDgwFwTXXFF7tSpU9SsWZMqVaoQEBBAZmYmGzZswNvbm6SkJP11oiiSnp7OL7/8wtKlS+nSpQsnT55EJpPh4+PDihUriI+PzzUWQRDo168fVlZWbNq0Ca1WW+SxFgStVktwcDDfffcdUVFRjB8/ns8//zxfgRNFkcDAQK5du4anp+cbETh4Wcevffv2uY5bW1vrSwz9G3NzcyZNmkTPnj0LZVmJokhoaCjff/89BgYGzJ49G2tra3x8fJg0aZJ+JcDAwCBH3xUqVGD06NFUrlyZFStWsGLFClJSUtBqtTl+JyUkJIrOB2vJBQcH06dPHwRB4I8//qBbt24olUpOnDhBy5Yt872vOCIXGxvLmTNnmDNnDvv37ycpKYlRo0Yhk8kYM2YMixYt0sdw+fv7ExUVxYABAxg+fLi+DQ8PD54+fYqLiwubN2/G1NSU7t27U716df3ELAgCvXr14tixY6xbt45Ro0aVeCaNzMxMzpw5w9mzZ6lXrx5eXl4FKrgaEBBAVFQUEyZMeKNW5usC2pOSkrC0tMwlZoIg5LDyDh06hLOzc65YxWxiYmLYvXs3FhYWzJ49GyMjI1JTU/VVzkuVyj9pr7m5Oa1atdI7x9y7d4/Fixfz+PFjqlatSt26dencuXMuD18JCYmC88FacgMGDKB8+fJkZWUREhKCTCZDpVJhaGiIXC4nIiKC58+f57qvqCKXHdM2duxY9u7dS0ZGBsOHD9c7iFSuXJnBgwfj7u7OypUrcXZ2Zs6cObmSQbdu3Zro6Gg0Gg1ffvklbm5uHD16lAULFhAZGZljbJ07d6Zu3bqsWbOmxLLbZxdYnTdvHmq1mpkzZzJw4MDXCpwoihw4cIAnT54watSoNyZwiYmJJCcnv/Y6Pz+/AlUycHBwYNmyZYSEhOQ4rlKp2Lx5M7t27WLQoEF88cUXGBkZkZWVxcqVKxkyZAg2NjaFGnvNmjWZM2cOgwcP5vnz57x48UJviWs0GqKiotBoNJKFJyFRCD5YkTM2NkYQBB49eqQvrqjVahk8eDDwsrhoWlruGl8FFbmsrCzUarX+3wcOHKB58+YcO3YMgMGDB+tThEVFRbF69Wr8/Pzw9vZGo9FgaWmZpxDI5XImTJjAxo0bycjIwNbWlnHjxjF27Fj8/f1ZvHgxISEheqea9u3b4+zszNKlS/WVCLLRarWoVKoCPU9KSgq///47c+fORalUMnfuXLp161ageDxRFPntt99ISkpi5MiRb2yJ8vHjxyxevLhAgm5nZ1egunnVq1fHy8uLxMSX1QPUajXHjh1j/vz5NG7cmBkzZlC1alUEQUCn07Fp0yZat25d5GTACoWCdu3asXjxYuRyORs2bCAiIgJRFAkODmb+/Pn88ssvhSqzJCHxIfNBLldml2iRyWRYW1vj5uYGvEzJlL2XFBsbm+eX+KtEThRFjh8/zpkzZzA2Nkan06HRaGjUqBG3b9/GxsYGKysrevfurR/HL7/8grm5eY74MVtbW7y9vRk7dixVqlTJtaRmampK79692bVrFyNHjkQQBMqUKcPEiROJi4vjyJEj7Nu3j4EDB1KvXj1atmyJXC5nyZIlzJgxgydPnrBjxw60Wi0GBgakpaXh4ODAgAEDcixriqKISqVi//793L17lw4dOjB79uxCLX3qdDp27dqlLxz7pkIbkpOTWb9+vb5KxOuoUqUKN2/eLFDbSqWSNm3aEB4ejo+PD82bN2f27Nk5HIWyhbxChQr65cfiYGBgwPDhw3n8+DG+vr4YGRkxfPhwBg0axIsXL/QfClFRUcTExNCsWTOUSqXkpSkh8S8+SJHbtm0bU6dOxdDQMIew/ZNmzZrlW0E7L5HLyMhg5cqV1K1bFy8vL70VlpmZydatW7l37x6Ojo707NlTHz8ml8sZO3Zsrkm5bNmyfPXVVyxcuJARI0ZQs2bNPMd36dIlQkJCaNiwof64tbU1Q4cOJSMjg927d+Pn56e35kxMTBg9ejS1a9dm4sSJOfZ6Ll68iJeXl77awbNnzzhy5Ajh4eH06dOHTz/9tNATqE6nw8fHB3Nzc/r37/9GJ2Bzc3O8vLwKLKI1atQoUGFYURR5+vQpu3fvRqFQMGHCBHbu3Em7du1y/H6cPHmSlJQU/UdHSVGxYkWmTp3KrVu3WLBgAc2bN6dz587657SxsSEsLIwFCxbQunXrPJ1tJCQ+ZD7IOLmzZ8/qrZt/5rzbtm0b9evXRyaT5euBFxQUhKGhIS4uLjmO79ixAycnJzp06JBjolUoFDRp0oTMzEx9QuKUlBQGDhxIp06d8g0GNjQ0pFmzZqxdu5ZKlSpRtmzZnDWsBAFHR0fWrl2bpyArlUoaNmxIw4YNCQ0N5ddff9VbbjNmzMjVb6VKlahTpw4rVqzgzp07nD9/nlatWjFo0KAixXFptVo2b96MtbW13sGnpNFqtVy+fFk/vpLsQxRFNBoNO3fu5PTp0/Tv358uXbpgZmaGhYUFW7ZsoUWLFsjlcm7cuMHJkycZO3ZsvkuxKpWKkJCQIsdXlitXDhcXF6Kjo/Hx8cHa2hpra2uUSiW1a9emTZs2VKlSRf9xdfToURQKRZ7ONR86Upzch8UHuSf3z7RSPj4+ZGZmotPpiI6Ofu2EkJcll5aWRkxMDK1atcr3/r59+xIaGsqMGTPw8PAokFOCubk5X331Fbt27SI0NDTXeSMjI4YOHcrmzZv1e3D/xtLSkn79+uHl5cXZs2dfORFXrlwZe3t77OzsmDVrFg0bNiySg4hGo2H9+vVUrVqVXr16vZFJ9v79+3z33XdFvj8uLi7fvTuNRsOpU6fw8vLCwcGBr7/+Wv/Rkx10P2zYMJKTk4mJiWHPnj1Mnjy5QN6lxcHAwIAuXbrw3Xffcf78eby9vYmOjtaP65/9Ozk5ceTIERYsWKDfT5SQ+BD5IEXu008/1U8Iz549Q6vVkpmZqc82cuvWLR4+fFjg9kJCQvSZNfLDwMCAmjVrotFo0Ol0Bf5jZGTEjBkz8PPz48qVK7nO165dG0tLS4KDg1/ZTlZWFtWqVXttQmBXV1eePXtWZGFSq9WsXr0ae3t7unTp8sasiOvXr+Pp6UnTpk2LdP/GjRtzxRCKokhERATff/89iYmJfPvtt/m2b2dnh6GhIevWrWPSpElvtR6coaEhX3zxBe7u7vj6+rJx40YyMzNzfHxZW1szZswYZsyYoV+WTkpK4t69e1IMnsQHxQe5J/dPTE1NUalUlCpViqFDhyIIAn/99RedO3fO8/q8LLnExMQCLTlYWlqyYMGCImWeF0WR+fPno1Ao8sxavmvXLurWrZuv04VKpUImk712z8rY2DiXF2ZByczMZNWqVTRv3jzXcm5xEUWR5ORkfUXvvn37Frmt9PR0gBxLvHFxcezZswe1Ws2XX375ynI+2W14e3szbNgwwsLCaNOmzVtdFsxO/TZjxgxCQ0OZO3cubdq0oV27djksun86CSmVSs6ePcv27dvp2rVrjur2EhL/VT5Ykbt27Rp16tTB3d0dc3Nz5HI55cuXB15ad/kF/+YlclWrVuXmzZuvtCpEUSQuLg4vL68iZ9rPjsFq1qwZrq6uOSbViIgIdu/ezYwZM/JcYlSr1SxatEifiT8/nj179soA5vzIzMxk6dKltGvXrlh5Pf9NdvWAbdu2IYqivpJ3cXj8+DGdOnUCXi5N7tu3j7t37zJw4EBq1679WrESRZE1a9bQp08fatWqRUBAADY2NkUOGygO2XuzderU4fjx43h5eTFs2DBq1aqVp1fu8OHDUalU+uw6//xdlvbuJP6LfJDLlfDS5fzw4cOULl1a/+Wr0+kQRZGOHTvm6yafHU/3T+rVq8e1a9fIzMzMt7/nz59z/fp1goODi2wpGRgYMHnyZK5cuUJgYGCOCapatWo4ODhw5MiRPO9VKpVYWVlx9+7dfNsXRZGjR4/i6upaqHGlp6ezaNEiunbtWqIClz2mjRs30qBBAzw8PEokiLxmzZo0btyYc+fO8d1331G5cmW++eYb6tSp89qJPtuhpnHjxjRq1EjvIbt///5890XfBoaGhvTs2ZNvv/2Wo0ePsnz5cp48eZLnsqSRkZH+gw5eehuvXLmSe/fuvdNnkJB4E3ywIte8eXOuXr1Kamoqoiii0+nYsmULoijSuHHjfCdTY2NjkpOTc0weMpmMrl276q2Nf6PT6Vi/fj0zZ85EpVIxb948Dh8+XKS8ktlu7Hfu3OHIkSM5+uvVqxdXrlzJdz/x008/ZevWrbx48SLP84GBgVhZWVGxYsELSaanp7Nw4UL69OmDk5NT4R6mAGSnPGvcuHGJtfnw4UPmzp1LTEwMX3/99Ssdhv7NkSNHMDIyom3btvpjxsbGTJky5b2whIyNjZk4cSK9e/dm69atbNu27ZXB8YIgMHToUPr160dwcDAZGRlvcbQSEm+eDzKEAF5aZI6OjhgZGbF48WJatGjB0aNHadWq1Sv3rUxNTZk3bx5Xr14lOTkZAwMDrKyssLW1JT4+nj179mBsbIxSqSQ5OZkrV66wadMm+vXrR4MGDahVqxbt27cnKiqKrVu3kpqaSqVKlQpVS00mk9GkSRP+/PNPHj58qF9iEwSBunXrsn79elxcXHIJtaGhIfXq1WP58uVkZGRgbGxMVlYWUVFR/Prrr2RkZDBs2LACx5olJyfz008/8dlnn+Hg4FAik7xGo+HEiRMcOXJE78xTUuKRkJDA0qVLCQsLY9KkSTRu3LjAge2iKHL+/Hlu3LjB559/nusdvS4Qu7ghBIWldOnSuLi4oNFoWLt2LTKZjCpVquT5sxUEAUtLSz766CP9c0RGRvLkyZNcoSv/BaQQgg8LQZTcrNi9eze1atXiyJEjjBo1ipCQEDp06JDntdlpuAICAjh37hwGBgZs2bJFPxHExcVx7tw5njx5gkwmw87OjpYtW+aZZFetVvPXX39x8uRJatWqxYABAwoldjqdjm3btmFqasqAAQP0YwgKCiIuLo6PP/44z/syMzO5ePEid+7cITMzk9KlS9OiRQvs7OwKPKFlC9zIkSPzDFYvKitXrqRq1ap07969xJJK63Q6Dhw4wOnTpwFYunRpoSfue/fusWvXLmbOnPnKn9GBAwdwc3PL1X5iYiLbtm3D09Oz8A9QTFQqFQEBAVy9epURI0bkGwP6TxISEti7dy+xsbGMGjWq0Hk432fS09Op37ARsqnbUFoVfNWiJBE1aiK/7s2lP37N05FMouSQRA54+vQpP/zwA/PmzSM5OZnjx4/j4eHx2vv+GUheHERR5OrVq+zfvx8HBwe6du1KmTJlCtSuKIrs2rULAHd3d32A+/Lly+natesbcYZ49uwZS5YsYdy4cVSuXLlEv/R1Ol2JBXbrdDouXbrEgQMH6NChA48ePaJp06bY29sXqp2nT5+ycuVKvvnmm9c6Dfn5+VGqVKkcy5nwbkUOXv6epKamsnnzZhQKBYMGDXqtBSGKIi9evECtVudZ1f1/FUnkPiw+2D25f2JjY0OTJk04fPgwmZmZBY55KqnJWBAEGjVqxNy5c3FwcODnn39m3bp1+e6d/fveTz75BKVSyfbt2/UiMXz4cHbu3FniiXzj4+NZsmQJEyZMyDOvZmGJi4vDx8dHvz8pk8lK5J1mf7jcuXOHWbNm0b59e7p160atWrUK1U5aWhqrVq1i4sSJBfKK7dy5M3/++ed7l0BZEATMzc3x9PSkU6dO/Pzzz/j6+r52v65UqVJ6gYuNjWXXrl05Eo9LSLzvSJbc/0cURf744w9q1KiBpaUlVapUeadjefDgATt37sTGxoYePXq8NrWWKIr4+/sTGxvLiBEjUCgUXLlyhYsXLzJ69OgSEY5Hjx6xevVqpk6dipWVVbHaFEWRs2fPcuLECSZOnFhiez9JSUns37+fp0+fMmrUqGK1m5mZyaJFi/jkk0+oUaNGgS3r9PR0TExMclz/ri25f6PT6Th//jwBAQF0796d5s2bv9ZzNXvFYe/evbi7u9OgQYO3NNqSRbLkPiwkS+7/IwgCrVq14scff3znG8GCIFCjRg1mz55Nu3bt+O233/D29ubJkyevvMfNzY1KlSqxfv16tFotjRo1QhAELl++XOwxxcTEsGbNGqZOnUq5cuVKRJCsrKz49ttviy2Y8HLSDggIYPny5Tg5OTFr1qwc7SYkJBSqPa1Wy8aNG+nQoQM1a9Ys8PgEQcDU1PS9d9aQyWS0atWKb775hpiYGObNm0dMTMwr78lecZgzZw4VKlR4SyOVkCgeH2wweF6ULVsWGxsbLl68yLNnz7C0tKRt27ZvrDzM6xAEgdq1a1OrVi3i4uLYvn07MpmMHj165Bu03K1bN06dOsXKlSsZN24cn332GQsWLKB27dpFrjB97949tm7dypdffomlpWWxnik9PR2lUolSqcTBwYGsrCwuXrzIo0ePsLS0xMXFpdClfK5fv87+/ftp0aIFs2fPzmWRJCYm8ssvvzB16lS0Wi1hYWFERERgZGSEq6trrmVIURTZs2cPVatWLXJWkMzMTGJjY6latWqutqOiovDx8aFs2bLExcXh5ORE7969SU1NRaFQFKhGX0lhYmKCu7s7L1684Oeff8bCwoJBgwa9slyRkZGRfkk/IyMDf39/unXrVqBMPhqNhtDQUCIjIzExMcHV1fWtPq/Eh4dkyf0DQRDo1KkT27dvp1u3bsTHx7Ns2bJ3vr8iCAI2NjZMnz4dd3d3/vzzT+bPn094eHiuuDxBEGjXrh2NGzdmxYoVGBgYMGTIEDZu3FikfIXh4eFs27aNmTNnFlvgEhISWLhwoT6tVnp6OhMmTCAsLIwWLVoQGhrKjz/+CLyMR3tVcD28dID56aefuHz5MtOmTaNbt255LrlduHCBZs2aodFo9MHSzZo148WLF3rhO3/+PHFxcQCcOHGCjIyMYiWX1mg0/PLLL7n2r2JiYhgyZAg9evRgwoQJTJkyhQULFvDw4UOOHTv2TpIpZ++9zZgxA1dXV1atWlXg4HZDQ0Ps7OxYuHAhV65ceeW1WVlZfP3115w6dQpnZ2fi4+OZMWMGoihy6tSpAu1BS0gUFknk/oVKpWLs2LEYGhoyaNAgRo8e/cazyxeGSpUq8cUXXzBp0iROnjzJggULuHLlSq7AcldXV9q3b8+iRYuoWrUqlSpVIigoqMBCJ4qivkRPSVhwKSkpeo9MS0tLtFotM2bMwMTEhBEjRlC+fHkGDx7M9u3biY6O5v79+/ladCkpKWzfvp1169YxcuRIRo4c+crxZWRkUL9+fby9vblz5w7Tpk2jYsWKDBgwgDNnzhASEsLly5exsrLi6tWrhISE6POYFhUTExPs7e1zTfwzZ86kRYsW+uB2S0tLrKysCAwM5NmzZzkykbxtBEGgXr16zJ49m9KlSzN79mwuXrz4SrGTyWQ0a9aMOXPmvHKZX6fTsXDhQiIjI/H09KRixYr079+foKAgbty4QWhoaJFXGiQkXoW0XPkvdDqdPiBWpVLl2F85ePCg3hPzXe65ZAfvjh49moSEBA4fPsyBAwfo0aMHTZs21Xt9Zo/T29ubKVOmsHjxYpycnApUOTskJAR/f39mzJhR5Fyb/0QulzNx4kR9vFVUVBQ7d+7kwoUL+ndZtmxZ1Go1v/32W66Ex1qtFplMRmBgIH/++Se9evVi8ODBBVpK7tu3LwkJCaxdu5bNmzfrrT25XI65uTknTpygfv36PHr0iN9//52ZM2cWO0u/IAj07NmT8PBw/bH4+HhOnz6Nr6+v/tlkMhmGhob4+/vz7bffIpfLUavVnDlzhoSEBKpXr06jRo2KNZbCIpPJaNeuHU2bNuX333/n2LFjeHh4vHIfzsDAgMqVKwMvLfTr16/nCHxPTk5m9erV7N69W/8zMzY2xsTEhEOHDun3j58/f05ERESu9uVyOQ0aNHhnWwcS/7tIvzH/wsXFBTs7O/7++2+Cg4NZtWqVfrmyc+fOhIeHM2/evPcm/VGZMmX49NNPmTVrFhEREXh5eXH69Gn9mBs1asTAgQPx9vZm8ODBrFmzRl/uJykpibi4OFQqlX5SF0WRCxcuEBAQwMyZM0tE4OClZfNPK+X69etUqFAhhxerIAio1Wri4+P1KcLCw8NZvXo1Q4YMwcvLi7S0NL777juaNWtWqAkvJiaG5OTkXIIhk8kICwujQYMGrF+/nnbt2vH777+zZMkSZs2a9dol09c98z+rticnJ6NQKLCzs9Mfy37vderU0T/z9u3befjwIT169GD8+PHcunWryGMoDmZmZgwZMoTRo0fj4+PDli1b9ImdX4WRkRGPHz9mwYIFemepu3fvIpPJqFu3bq7rw8PDad26NQBbt24lNjYWhULBmDFjiI+PR6FQsGrVqvfemUfi/USy5P5FtqWRlZVFfHw89evXZ82aNXh6emJgYMDgwYNJSkoqVGaSN40gCBgbG+Pu7k5aWhonTpzghx9+oGXLlnTq1AkHBwc+/vhjduzYQf369Vm2bBlpaWlYWlpiZGREQkICFhYWDBs2jNDQUM6dO8e0adNK5BkzMzOJjIykTp06OUxMpRgAACAASURBVI6XLVsWExOTHEL17NkzRFGkd+/eemvL0NCQmzdvkpCQgKenZ4Gs0H8iiiJ3797FzMwMS0vLHHt2aWlpJCcn07VrVzZs2MCoUaP4+uuv+eqrr/jkk0/o0aMHFy5c0E/ARUWr1SKKIlWqVNGnf6tatSo6nY4TJ06QlJSElZUV9+7do1KlSjRs2JCMjAwMDAzQarXvfNWgXLlyfPnll1y7do0VK1bQpEkTunbtmm/IgUwmo2/fvsTExPDo0SMqVKhA6dKlMTU1zVGwNzk5mfT0dNzc3DA0NCQ1NRUHBwd69OhBfHw8MpkMFxcXzM3N6d69uyRyEkXig81d+ToqVarEvn37GDRoEPXr18fY2Fj/n8zIyEj/99u3b7N69WqUSiU2Njb5Vt1+WxgYGGBvb0+rVq30+TFVKhUNGjTA3t6eFStWUKNGDTw9PXFxcaFp06a4uLhgbW3NwoULefjwIbNmzSqxlFoPHjzg77//xtHRMcdxa2trrl69ilKp1MfMXblyRb88fPbsWaKjozl+/DjOzs4kJiYycODAIo1hy5YttGnThoSEBBISEjA2NubSpUucPHmSihUr4uPjQ7ly5Rg4cCBmZmZUr14dMzMzfH19admyZQ7LqyicO3eOv//+m6ysLGbOnImfnx/37t3j0qVL1K5dm/79+3P69Gm0Wi2Ojo76LDI7duygYcOGdOvW7Z1P8IIgUKFCBVxcXIiMjGTHjh2ULl0aa2vrPMeWvaSenezb1NRUX/TXyOj/tXef4VGV+f/H31PTGyEEkhhCCIRQJJTQpKOAVAFpImBbe9/VxXX96fpfd1FXcVUWEHVZVkFExAKIFEFASiAUQwmEICWhpbfpc87/QTYDoYa0CZPv67p4wCQz557JnPM5d/dm586dbN68mfDwcPLz80lNTSUpKYkWLVqg1+vZuHEjOTk5jB8/Ho1GQ1RUVI3dWMralQ2LTAa/xNatWwkNDSU+Pp7s7GwCAgIqrIBSXFx82VDpkpIS1q5dS/Pmzeu8/+R6bDYbW7ZsYfPmzfj5+eHr68ujjz56xaa+3NxcZs2axWuvvVZjYb1582bsdjsDBw687GcWi4X9+/djt9uJjo4mIiKCkpIS5syZQ2ZmJhMmTKBXr14kJyezYMEC5s6dW6UyzJ8/n8GDBxMZGcn+/fsxmUw0a9aMmJgY5s2bh5+fH4MHD66wPuOePXt49913mT9/frV3/S4fZNOsWTNeeOGFSj3HYrGQk5PDc889xxtvvEHr1q2rVYaaVlRUxNKlSzl//jwPPfTQdZf9slqtvPPOOxQWFjJy5EgiIyOJiYnBarWyd+9eYmNjK1zsZ8yYQevWrXnggQdqvOwyGbxhkebKSxQXF7sC4NIT1+Fw8M477zBo0CB69+7tuoP19/dnzJgxFQYrfPnll3Ts2JG4uLga2QOtqoxGIwMHDqRv377MmDGD11577ap9WaGhoXTp0oVdu3ZVa7V8h8PBoYMH2f7LalL3/oLZXMqhX38msesgunTt5goNb29vunbtClxoVlyyZAlt2rTh2WefrVZt0ul0cjQ9na2bfyB1188kb/qC9h27c2vngXTr3gNfXz+WL19OaGhohcWtoWxT1TVr1vDhhx9W62+nKArHjx9ny88/4DT9xs8//ojOmU27xL706NnniiNCFUXh3//+N927d6d9+/YEBASwdOlSXn755SqXozYEBgbywAMPcO7cOebOnUtMTAx33XXXFefKqapKfn4+cbFRLPtiFfOObaH9rV1oldCD2/oMuuy7pigK27dvZ/r06XX1doQHk5C7hI+Pj2seV7l9+/YRFxeHn58fL730EosWLUJRFPr161fh9y6+UCYlJbFhwwZWr17N448/7vZpCFarlfDw8OsOJOncuTPr16+vcshlZ2cze9afCVNW0K2Ng2GjdGg1kF2wn73r57JiaXemPfpmhYWji4qK+PTTT12DDaq7GHBRUREfvvcXfAq/pGc7G7eP0GHQQ27hAX7d9jGvL+tAXOJErFYbjz32WIW/W2FhIRs3buSpp54iLy+PEydOXNbUWhlms5k5H76FPWsBvdqW0m+YHi8DFBQv5MC+fzPzu9YMmfBX+vXrX+H4qqqyfft22rdvj6IoZGVlMXHixGp9HrVFo9HQtGlTXn75ZXbt2sXbb79N7969ueOOO1zvyel0svjz/3A0+R16xWfzt+k6/H20lJiOkX5qMf96PZyE3jMYP2Gy64YiKysLk8l0w+uMCnEl0lx5iZKSEhwOB8HBwa7Hjh8/zqJFi3jxxRfR6/WoqorT6axUk57D4UCn06HRaDh58iQ5OTkkJCTg4+NTm2/jMgUFBSxYsIBnnnnmmv07p0+fZsWKFTz88MM3fIyzZ8/y1itjmHzbMW4J113xOPnFCv9Z48eEx5bQJqEtP/zwA3v37uW+++6jRYsWlz3n4MGDLFq0iD179vDwww8zfPjwa37u+fn5/P2Vexl56y5a3XLlMpSYFD5dqTJg0gIG3T6kwu/MmDGDH374Ab1ej81mY9WqVTe8jqnJZOKvf36Q/i3Wc2uc/oplsFhVvvgJEgb8k1F3VexrzM7O5ueff6a0tJQmTZowZMiQm2LovN1uZ/Xq1ezatYsJEyYQHx/Ph+/9lcam2QzsYkCnu/xzcDpVftxhp7TR0zz+9MusXbuWbdu2sX37dh555BGGDBlS4yuiSHNlwyIhV0krV67E6XQyatSoCo+fPXuW9evXM2nSpOs2bRUUFLB27VoOHz7MwIED6dWrV20WuQKHw8HMmTOZMWPGNUMiJSWFEydOMHbs2Bt6fbvdzv975RGGxq3k3UXFvHx/ABFhFY9jsij8eW4hD4z0Y9m2SBrFjGDQoEH07dv3qp+dzWZzrRpSPor0WiH95l//QM+whXy4tISnJvjTOrpiDdpqU3n9k0JG9/Vh65FmPPv6ugp9cWazucLk50sXWr4eVVWZ8+HbxKrvsHRdCeMG+tKlTcVmV6ei8u7nxdzaysBv2Y2Y8twPxMV5Tq0lPz+fL7/8kl/37aFr2LccyywlMkzHmP6Xh9VX603kFCpENPGlRf+F9Op1W4UtrLy9vWs84CXkGpb6f3tYx1RVveL2I3feeSeDBw++7PEmTZoQGxvLq6++yvHjx6/52sHBwYwfP56XXnqJpKQk1/F+/vln0tPTa3X5ML1eT3R0NKmpqVf9HUVR+PHHH+nTp88Nv/6+vbuJ0q0gNlLHQ6P9eOGDQk6evfA5FpUqvPhBIbd19KJdrIHxt50mNiaC/v37X/PmwGg04ufn5xo0c63AOXLkCIb8JbSO1vPUBH/+76Mi0k5cWFbLYlP5y8eFxETo6ZJg5O7bslny+UcVXsPHx8d1vKostHzu3DlyDn9EYisdT9ztz7uLitl50Oq6cNvsKm//txg0cHuSNxP6lfLlZ+9Xagmtm0VISAjTp0/HX93PHUkapg/3Y12yha83mFCUss9BUVS+WGNi0x4r04b5MrS7ho3fv41Wq63w974ZarCifpNv0CVKS0tZsmTJZSteaLXaK46y02q19OzZkxdeeKHSTZA6na5CH11UVBRbtmzhrbfeorS0tHpv4BomTJjA0qVLyczMvOxniqLw1Vdf0aZNmyr1iW3e8B29by3bC65TvJFnJvrz0r8KOZ3txGRReHlOIX0SvRjTv6wm1uoWPSfTVl+2HFl1/LT+RwYk2ssWto428NpDgfxlfhFHTtrLanD/C7gHR/qh02qIaqIjM+3bGg2YHds20a1VIRqNhsgmev7+eBDvLS5hz2E7TkXlvcXFADw7KQCdTkNIgAZH3kaKi4trrAz1wYkTJ7gl4BAGvYYAXy1/fyKYtckWvt1UtojC1xvM/LLPyhuPBeHrrcVo0BBu3HfdnRCEuFEy8OQS3t7enDhxAlVVr3gXr6oqH3/8MUOHDq3QVxMUFOQaLVdQUMCWLVsYOHDgdfsTyrfVadmypav/DsqafJYuXUqnTp1cOwhUd66Ut7c3f/jDH/jwww9p0aIF3bp1w8fHh9OnT7N+/XratWt3WXNsZeWcOUijDhfumZLaGvn9PQH8/p8F+PloGNrTm3EDLjQ1GvQaNNajWCwW/P39q/W+yuWd2Utw5wufUZsYA6/9LpBXPyoiNEhDx1ZGHhjlh1Z74XeaBORz7ty5Gts65uiRfYyKu3Ba3RKu482ngnjxg0LCgrVENtHxwpQAV/+URqMh2JBFfn5+tdcHrU/SDu4jPvpCDT3QT8tbTwbzx9kFbP3VhtWm8uZTZQFXLjZS5Wh6mgw4ETVKanKX0Ov1KIpy1bt7jUbD+PHjmTNnDllZWVf8ncDAQHx9fZk5cya7du26oWOXh0BwcDAjRowgKyuLdevWVVh2qzoaNWrEn//8Zzp16kRKSgrr1q0jNzeXRx99lLvuuqvqzUOKpUIIazQa2sSUXex/y3LQs73xspDWqpYarUWpDgvaS44RE6EnJEDDnsN2+nTyQqet+HOjgRptJrZZrRgNFT+HZo11xDTTsWWflf6dvS4bgGE0qG7f6aKmWS0mvC4ZUBzgp6Vney9Wb7fQ81ZjhYADMOo12G2WOiylaAikJncFTz311DX7iYKDg/n973/P8ePHiYyMvOznWq2WgQMH0rt3b9f2IRd3pleGRqMhIiKCu+66q8LjO3fuZMWKFbRo0YJBgwZdtl9ZZWi1Wtq3b1+jHd5eAdFY7fvw+t8Fvqi0rIly3AAfYiL0/PHDQv7+RBC3hJd95VRVxaGPqrGVVQC8g2Kx2H/C26usDOWDTBLjjTw5IYC/zC/itd8FEt/8wtU3u1BfoytONItoztlcJzHNyt6n3aEya1ExwQFalr3ZmJfnFPL8ZA2d2xhc34V8k3+F0byeIKp5HCd+UohvXvZ/RVX5ar2Z5IM21n0Qxl8+KcTbqGF0Xx9XzfpsrpM23WPcV2jhkaQmdwVBQUHXDaPyidPXYjQaXf1bdrudOXPmsGfPnmrVxrp168Zrr7122XqKycnJHD9+vNo1vapqn9iPAxllgzzKR1H2utXI2AE+dE0w8vREf16aXdZHB5BXpOAfllSja4B2SOzNgWNl799qLwu46KZ6HhrlR9sWBv7vwUBem19E+qmyclpsKqp/Uo1O5+jcrS8ph8vKUN4H51TguckBxDTT87fHg3h3UTF7j5SVwWZXMes7elzIJSS0Iy0r2PV9XL7BzKY9ZX1w4aE6Zj4RzJrtFr7bXNZHp6oqx3Ka0rJlrDuLLTyQhNwVKIqCxVL5ZpOff/6Zffv2XTNgDAYD06dPZ//+/fzrX/+q1oALrVZLy5YtK9TiQkJCWLduHa+//nqFZtQrjRStaTabDbTeLP3ZSIlJ4U//KhtkMvEOX9ddelJbI89ODuDFDws4n+fkh2Qvho66r0bXZOw/YABbj8Vjtii88WmRK+DKy5DQwsCrDwXyf/OKOHHGwaqtCsPHPlpjxwdo0yaBM/Z+5BcrvP9FMYoKL0wNwKAvK0N0uI6ZTwbxj8+K2Z9hY9sBha79HqjRGm19EBwcTEjsZH47o/Dtz2Y2plh588kgAv3KLjmBflrefKos6FZtNXPstEJqZgQFBYVuu1ETnknmyV2B2WzmP//5D48+WrkLoMVi4YMPPiAxMZHbb7/9uhfu8r3RyvfPCgoKqpWFnZ1OJ++99x6KopCYmMiAAQNq/DipqaksW7aMLl26oNPCgTW/o020g64JxgoDPModPGbnaKYDZ+SfePSJP9RoWQB27dzBqn+Pp11zEz07GC/rgwM4ctLOibMqZwwP8PyLf6vxZdfS0g6x8N3RdGmZR/d2RlfAXSwr28HBYw7SLXfxyv+b5/aFvWtDXl4ub8wYTc8WaXRsbXAF3MWKShVS0mzsPNGOR1/4hJUrV2I0Gpk6deoVlwirCSaTiXaJndE8MRdDqJvmyTkdnHz1bnat+FLmydUyzzuzaoC3tzdnzpy56gjLK/3+s88+y+bNm1EU5boXzYt/fvToUb777juSkpIYOnRotRcDvvQ4zzzzDKWlpZw4ccL1XnJyctizZw+tW7emSZMmVWquO3PmDIsXL8bf35/f//73rhGSDvsHpKx9mZhmRYSFaF3HVFWVYpPK/hNeaCOf4+GHn6mx93mx+DZt+VgZiuXYFmKa5RERpq0wGKXErHD4lJ4876k8/dxfamVd0fj4Nox9+DO+/OgxIsOyiG6qrRC2ZqtK2gktGba7+cOf/uGRAQfQqFEoz76ymH/+7UEC/H6lXaxaIfBtDpWjmQq7z/Tk+VfnEhkZxdNPP83hw4d566236NGjB4MGDarRcwLKBnhFhDUm490H0Gjc1Zil0lix1/nKRw2R1OSu4v333+d3v/tdlb+ElQ1IKKtx7dy5k5YtW1Z73cbKMJvN7N69m0OHDtGoUSPX6ialpaV4eXld86JrsVhYsmQJZ8+eZeLEiTRv3vyy95mRkcHyL+dRlPUTRucxtBoFqyYSQ0gPBo98mKRu3Wpl6xin08nbb7/NsGHDaNSoEcuWfEzu8dUYHUfRahzYNE3QBnaj35AH6dO3X61PNM7OzuarLz7ldPoKjPbD6DQ2bISCf2e6D5jO4MF3unXx7rpSWlrKN18vIS3lCwzWAxi0JdiVABzeHWjTdSJjxk687DxzOp2uJb5GjRrl2jm8phQWFrp942ODwUCjRo3cvo2Sp5OQuwqbzYbBYKjSF1BVVWbOnMmYMWOIj4+/4dfIy8tj/vz59O/fn44dO9b4nezFLg7j1NRUli9fjqIoDB8+3LUqi6IoOJ1Otm7dypo1axgzZgydOnW65gVaVVXMZrNr13EvL69aXcHC6XTy0Ucf0b59e9cOEaqqYrFYsFjMOJ2Kqwx1GSyqqmK1Wjl79iyLFi3iySefxNfX12Nrb1ejqip2ux2z2YTd7sBgMLg+h2udH6WlpSxdupRz584xZcoUIiMjJRTEDZGQqyWFhYXMnz/f1U9XleevW7cOnU532TSC2uZ0OrHZbPj4+JStxThnDj/88AMBAQHMmjWrwlqP9cWyZctQVZVx48bVy4tgfn4+Cxcu5JlnaqeZ1tNlZWWxaNEigoKCuPfee2t80WbhuSTkriE7O5vGjRtX+aLpdDopKSmp1koWF9e01q9fj7+/Px06dLjuQsXVpaoqubm5fP7552i1WiZMmEBeXh5NmzYlJCQEh8PB999/T0BAAK1ateKWW25xyzqDqqqyefNmDh48yMMPP1xv1zqUkKs+RVE4cOAAS5YsoW/fvvTv39/jRqWKmlc/rwj1xLx587BarVV+vk6ncwVcUVERO3bsuOHh0RcHWdeuXcnKymLWrFlkZ2dXuVzXY7PZWLx4MfPmzWP48OE89dRThIeHk5CQQEhICFD23nr16oXRaGT9+vWuz8lsNrNx40ZOnjxZJ4sOHzp0iF9++YUHH3yw3gacqBlarZYOHTrw6quvYrVa+dvf/savv/7q7mKJek5qctfw/fffExYWVq1dsss5HA6WLVtGZmYm9913H6GhodV6La1Wi1arpaSkhC1bttClSxdCQ0OrdaF3OBwkJyezcuVKRowYQVJS0g33HdntdlJSUti/fz8Wi4VHHnkEg8GA2WwmOzubkJAQ/Pz8aiSQsrKymDt3LjNmzLjuZrDuJjW5mqWqKsXFxSxZsoSioiLuueeeGlt/VHgWCblrKC4uZsOGDVVetPhKjh49SmlpKR07dqyR13M4HKSkpPDLL78QHh7O5MmTqxQgGRkZLF68mLi4OO66664aH+ySl5fH119/TX5+Ps2bN2f8+PFoNBpMJtNVd3i4loKCAt5++22ee+45GjduXKNlrQ0ScrVDVVVOnDjBkiVLaNq0KRMmTJBh+aICCbnruJGpADfKbrdz/vx5IiIiqn2M8tFr5SNC9+3bR0lJCR06dCAgIOCqr5+bm8uXX36JzWZjypQphIaG1npfn9PpdO2WnpaWxvLly7FYLPTp08c1SMdkMrk2zby0PFarlZkzZzJ16lRiY2+OZaAk5GqX0+lkz549LF++nMGDB9OrV68K21mJhktCzo3MZjNz5szB39+/xkeMFRYWsmXLFvbv38+wYcPo0KFDhZ/b7Xa+++47Dh06xNixY0lISHD7qERFUVy10M2bN/PLL78AMGrUKNq2bQuU9Rd++OGH9OvX77prh9YnEnJ1w2638+2333Lw4EEmTJhQpSk8wrNIyF2Hw+Fg8+bN9O/fv1ZOFkVR2L9/P/n5+fTr169WXl9VVdfcsE8//RQvLy927NjB2LFj6d27d72ds1U+106v12M0GlFVlSeffJLi4mLatm3LtGnTiIgoW5aptLQUb2/veju5WkKubhUUFPD5559jt9u59957a72FQtRf9fPqVo/odDq2bt1K165da2UtPa1Wy6233ur6v81mIysri5iYmBo5KS/unzt58iTHjh0jJCSENm3a0LVr13obcFA2srS8dquqKqtWraJHjx7ce++92Gy2Cu9t0aJF5OTkEBgYyNSpUwkMDERVVVRVlVGXDVBwcDBPPPEEGRkZzJs3jxYtWnD33XfLlIMGSGpylZCSksKxY8cYP358rR/LbrezZMkSTp06xT333EN0dHS1wu7iUWgFBQVMmzatwv5pGo2GrKwsPvnkE1q1akW/fv1ctaP6JDk5mW3btvHEE09cMZhVVcXhcGA2m/Hz80On02Eymfjggw8wmUyEhYXx2GOPodPpXL/n4+Pj6husbVKTcx+Hw8GuXbv4/vvvGTZsGN27d6/XN3eiZknIVYKqqmRkZBAXF1dnxzx37hwHDhxgwIAB1ZqMvnLlSnbv3s3YsWMr1Bgv5XA4OHToEHl5ea5m07y8PHx9fWt1WbHKOHr0KIsWLeKPf/xjlfefs1qtrueeP3+eRYsWYTabiYmJYfLkyUBZH6miKLUyHUFCzv0sFgvffPMNGRkZTJw4sU7PZ+E+EnI3iU2bNhEfH0+TJk2uG3qKopCamsrSpUvp378/ffv2rVIzzb59+/juu+/QaDRMmjTJLReF8+fP8/777/Piiy8SGBhYY69bPsrTbr+wEvzx48dZunQpJSUldOrUybWcWmFhITabDV9fX3x8fKrU/CkhVz+oqkpOTg6LFy9Gr9czadIkGjVq5O5iiVokIXcDioqKrjkcvzalp6ezfPlyQkJCmDZt2lVrNOVr/DVq1IiJEye6tsCpDrPZjNPpxN/fH0VR+P777wkPD6djx461OieptLSUv//97zz++ONubUI9cOAAGzduxGQy0atXL2677Tag7G9it9uJioq67vdCQq5+UVWVtLQ0vvrqK9q2bcvIkSOlv85DScjdgEWLFpGYmOgazl7XFEXh3LlzhIWFodfrK2y+WlxczPLlyzlz5kytrtauqirnz59n586dpKam8vTTT+Pn54eiKNjtdoxGY40c12q18u677zJmzBjatGlTAyWvvksHshw+fJi9e/eSmZnJ4MGD6dChA6qqsn37dgCio6Np0qQJBoNBQq6ecjgcbNmyhbVr1zJ27FgSExPr7QhdUTUScjegtLSUt956i1dffbVejNjbs2cPGzZsIDIykrS0NIYPH06XLl3cUtMsKipi3rx56HQ6EhMT6d+/f5U/o/KdDzp06ECfPn1quKS1S1VVjh07Rnp6OqdOnWLkyJE0bdqUvLw8nn76aYYOHUp0dDQ9evSQmkM9Ur5P4unTp5kyZQrR0dHuLpKoITLE6Ab4+fnRq1cvCgoK3N6OrygK3t7eZGRkcPr0ae655x46d+7stvIEBgby/PPPU1hYSG5uruvxkydPkpmZSZs2bQgMDLzuqDZFUVi6dCmRkZH07t27totd4zQaDS1btqRly5aXPd62bVt69OjBqVOnXItXm81mZs+ejVarJTY2lmHDhrnmBFqtVoxGY724ofJ03t7eTJs2jfPnz/Pf//6XwMBAJk+ejL+/v8yvu8lJTe4mdP78eT777DO8vb2ZOnVqhfl75bt+JyUl1YuaQmFhIVu3biUtLY2OHTsycOBAANfSXpfasGED6enpPPTQQx51cb9Wc6WqqhQVFZGdnU1MTAx6vZ6SkhJmz56NqqrccsstTJw4Eb1ej8VioaCggLCwMGlWq0WpqaksW7aMTp06MXz4cJlycBOTkKuiq12ka5PZbGb58uUcO3aMqVOnXnEOncPhYNOmTWzYsIGkpCRGjBhRb8Li4nVAV61axbZt24iIiGDMmDGEh4ezf/9+Vq1axfPPP+9x6w5WpU9OVVVsNhs2m81Vo8jOzuarr74iJyeHFi1aMGXKFDQaDWfPnuXUqVNERkYSHBwsm4rWAJvNxqZNm9i4cSPjx4+nQ4cO9eZcEpUnIVdFH3/8MZMnT66TLV5UVWXDhg1s3LiRO++8kx49elRqGkF2drZryoHJZKr1jVZvlKqqnDlzBj8/PwoLC/nkk0+IjY0lIiKCDh06EB4eXq/KWx21PfAkOzub5ORkMjMziYuLY9CgQQDs3r2blJQUwsPDSUxMlL6mKigtLeWLL74gLy+PyZMnExUV5e4iiRsgIVdFycnJpKWlMXXq1Fq7EKuqytGjR/nss8/o1KkTQ4YMqfKQ/Y0bN7J+/Xr69+9Pnz596kVTZrmLt81RFIW0tDSOHj3K2LFjCQ4Oxm63k5+fT2BgIF5eXjdl8LlrdKXJZCI/P5/z588TEhJCTEwMUFaTTk5OxsfHhzFjxtC6dWsASkpKADAaja4dLUTZuZiZmcnixYsJDw9n3LhxNTI9R9Q+CbkqUhSFjz76iPvvv7/Kq3BcS05ODl988QWKojB16lTXjtzVYbfb2bRpEwkJCfVm6S6z2cybb77J/fffT/Pmza/4OyaTic8++4zc3FzCw8N54IEH6riU1VdfpxCYzWYA183T999/z+HDh3E4HEyaNMkVilu3bsVoNNKkSRMiIiIabB+VqqqkpKSwYsUKevTowe23395gP4ub5IxFjQAAFXBJREFUhYRcNVw8T62mWCwWVq5cyYEDB5g8eTJxcXG1djddWlrKrFmz6NGjBz179qzz3bWdTifvvfceAwYMoFOnTtd9n+W7EpT3Nx05coTPP/+c8PBw+vbtS/v27eui2FVSX0Puai6dE5iamsqJEyfIy8tj+PDhhIaG4nA4mDNnDj4+PkRFRTFw4ECMRiOKomCz2TAYDDV+ftQXVquVdevWsX37dqZMmULr1q2lv66ekpCrJxRFYdu2baxZs4YBAwbQr1+/Ork42Gw2fv75Z3Jzc5k0aVKtH6+coigsWLCA6Oho10apVaGqKmfPnqWkpIRWrVoBsGvXLjIyMkhISKBVq1b1Yqfomy3kKstut3Pu3Dmys7Np3749BoOB4uJiPv30U0wmE8HBwTz66KOufuGMjAwaNWpE48aNa6UFpK6VlJSwcOFCLBYLU6ZMITw83N1FEpeQkKsBixYton///lVqAlRVlVOnTrFgwQJat27NqFGj3D4ybv369eTm5jJw4EAaNWpU43eo5dvmFBQUMHny5Bp//ZKSEk6ePMnBgweJj493rUSSnp5OcHCwW/r2PDXkrqZ8bVCLxeLquyoqKmLjxo2cPXsWo9HIfffdB8CpU6dYvXo1TZs2JS4ujoSEBKCspq8oCnq9vl7XBlVV5eTJk/z3v/+lRYsWjBs3zu2LmosLJORqQH5+Pu+88w6vvPLKDd2dFhQUsHjxYkwmE/fddx+hoaG1WMrKUxSFffv28dNPPzF69OgaX5h527Zt7Nq1iyeeeKLOmnjK+1K2b99OcXEx06dPJyIiAkVRKCoqIigoqFYvpA0t5G6Ew+Hg/PnznDlzBq1WS6dOnYCyNUO//fZbDAYDXbp0cc2xPHfuHIWFhYSGhhIcHFxv5guqqsqOHTtYtWoV/fr1Y8CAAdKEWQ9IyNWQbdu24e3t7TpBr8Vut/PDDz+QkpLCpEmTiI+Pr7cnQ/ncNpPJxOeff05SUhJt2rSpck3oyJEjfPnll7zwwgtua64q/8prNBpsNhtLlizh6NGj+Pj48Mwzz+Dj4+Pan66mahESclWjKApms7nCBrpHjx4lOTmZnJwcOnfu7FoZZ8OGDfz22280bdqULl26uJoObTYbGo2mzmqEZrOZH3/8kT179jB16lRatmxZr2uink5Crg6pqkpycjKrVq2id+/eDBo0qN6G25UUFBTw448/kp+fz4MPPnjDE7bPnTvH+++/z4wZM2pll/XqunTB69mzZ+Pt7U2rVq0YOnRotWoMEnK1z2KxcP78ec6ePUtUVBQRERGoqsq3337LwYMH0ev1TJgwgZiYGFRVZf/+/QCEhobSpEmTGh8lWVBQwMKFC9FqtUycOJGwsLAafX1RORJyNax8Nf6LaymqqnL69GkWLFhA8+bNGT16dL28yFfWxTWh1NRUtm7dyuDBg4mKirpq8BUXF/Pmm2/y5JNP0rRp07osbpU5nU4KCwsxmUxERESg1WrJyMhg8eLFtGrViq5du162RuXVSMi5l6qqWCwWDAYDer0eRVHYvXs3x48fJy8vj3HjxrlGjM6dOxeHw0FERAQjRozA19cXVVUxmUzodDr0en2ld5Qv7wtesmQJCQkJjBgxQvrr6piEXA0rLCzkn//8Jy+++CLe3t4UFRWxdOlScnJyuO+++zxu9FV5p/uPP/5Iu3btXHutXcxms/HWW28xfvx44uPj3VDKmmWz2UhPT8dms7map/fu3UtqaiotW7YkISHhsnmNEnI3D1VVKSwsJCcnh+joaIxGI3a7nblz51JaWoqfnx8PPPAAfn5+WK1WvvnmG5o0aULTpk1p3br1FWv8qqqyadMmfvrpJwYPHkyPHj3qTV+ip5OQuwKLxcLs2bMxmUxVev7BgwfRaDS0bt2aFStWEBsbS/v27W+4XT4yMpLp06e79WT45JNPOH369A0/7+DBg7Ro0QIfHx/S09MpLS0lMTGxSmXw9fXl8ccfd+tUgJUrV7J79+6r/txsNnP69Gl+++03oqKiLtsDz2KxsHfvXnr06FHlMuh0OqZPn05kZGSVX0NUX/nqJwsXLsThcHDkyBFycnJQVZVevXqh0+koLi52LacWERFBUFAQUPY9WbduHbGxsbRr167aZenWrRtDhgyp9ut4Mgm5Kzhy5Ai39u6O19290FSxz0wptYBei9arastnKUUmvNYe4ETGMbdOKQhp0hjHHe3RhVSteVUpKVtRQ+tftYBS7Q5KvtrCoW0prqWn3GHAwIEkB5gw3HLj/SoqoBaZUB1OdI2q3kxdujqFz96YxcSJE6v8GqJmLFy4kAdfexG/YUk3/FzV4UQpMaOWWNBFhlZ5UIr95Hn62EJYvXp1lZ7fUMh6NFehDfEn+KmRaPTuqUU5zuZj35bhlmNfTOvrRdBDQ9BHNXbL8RWLDfOWA245dgVaDf7jbsOnp/t2KbcfP+e2Y4vLeSXGEvzMKLcd37zlACxJc9vxbxY3z9A+IYQQ4gZJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYencX4GZlz8zBvDEVJa8YXZMg/Cf2RaPRuLtYdU4pMlH6wy6cOUUABP1uKBpjw/taqTYHpT+m4Dydh7OwlKBH7kQX5OfuYgk3cZzOxfTTryj5JWhD/AmY0r9BXh/qA6nJVZEG0Bh0FM79AXv6aXcXx21UQONloGTpFizb00DbcE9kjVGPacOvmL5PBrmgNWwqoKoULViH/UiWBJwbSchVkT6qMd49E0CjwatbfIP9EusCffEblgR2J95dW4GuYX6lNEY9voM7o/U2YmzfHK2/t7uLJNxIHxmKz21tAfDuHu/m0jRsDfOKVEOsezLQ+hoxtolyd1HcypZ+GqXYjFfnuAYb9gBKsRlbWiZeXeLQaOXUauhsaadQrXa8EmPdXZQGTc7EKlJVFeuOw+gjG6OPDHV3cdzKuucoGl8vDA097A+dQrXY8Ooc5+6iCDdTVRXLjiPobwlDF9HI3cVp0BreCIEaolpsWFNP4JXUCtuhU1i2paGYrPj0a493p5buLl6dUVUV6850DC3CUU1Wiv69FmehCa/2zfEZ2BFNA+qjs+3JQOPrhS4skOJFG3HmFqNrGoz/qB5ovAzuLp6oSw4n1j0ZeHdthe3ACSzbDqOYLPjc1rasWV/UGanJVZHzTD6OE+dwns7FdiSLgEl98buzC9mP/wtbWqa7i1dnlGIztgMnURUF09o9+I3sTuDUAeT/fSmmNbvdXbw6ozqdWHamo/XzpmTZL3j3aUfgA3dgWrmLwnk/oKqqu4so6pAzrxj7kSyc5/KxpWXiP6E3voMSyXlufoO6PtQHUpOrIuv+46gWOz4DO+I/thcajQaDrxdqqQXboZMNpp/OeToPR2YOAf06EHj/HWj+N/BEG+KHNeUofkO7uLmEdUMpNmM7dApj60iCHr4Tra8XAPpbwrCkHCVIUUCnc3Mpq+fkyZN89913GI1GxowZQ1hY2A093+FwsGHDBnbv3k2XLl0YOHAg2hvsuzx37hzLly/H6XQyevRooqLq53lmTTkKKnj3bU/A3b3LHlQBjQbrr781mOtDfSAhV0WW7YcxtIvGf0xP12ALx6kcVKeKNtjfzaWrO5ZdR9D4eRP0xHBXwKk2B46sXHwGdnRz6eqOPeMsSm4Rwc8/4Qo4VVWxHz+HLsQPbvKBKEeOHGHQoEFkZpbVQj744APWrVtHeHh4pZ6vKAqvv/46M2fOxG63YzQaeeWVV3j55ZcrPVgpMzOTIUOGcPDgQQD+8Y9/sHbtWuLi6l8fqGXHYQwJUfiP6+V6zJlThFJiQRss8yfr0s195rlJeT+Ud9dWoL9wd27ZcRhtgA/G+MgrPke1O+qymHXCmpyOsV1ztIG+Fx7bfxzVZMW7y+UXH8VsRbXZ67KIdcKSfARd0xAMLS5c9JXcIuyHM684xUS1O1BM1pumGfPf//63K+AA9u/fzzfffFPp5xcXFzN79mzs9rK/vc1mY/78+eTl5VX6Nb7++mtXwAEcP36czz77rNLPryuq3YF1dwZeiS3R+ni5HrftOwYOJ14X9dk7C0sx/3Lw6v+2HkIpNrvjbXgMqclVgeNUDo7TeXh3a+26eKkOJ6Y1uzHeGoOuWaOyQNPrUEst2DPOYlqdgibAh+DHh7u59DVHtTmw7j6K/4Q+rvlxqqpiWrMHXZNgjB1iUB1O0GpRTVZKlv2C6lSwp2eh9fUm+NnRaAN83Pwuqq98pK2xXXM0fhfmx5m3HER1Kvj0bovqVMp/m9KVu1DNVlS7E+uudIKfGYWhRVP3FL6SLBZLpR67GlVVL/t9m82Gw1H5Gz+z+fKL/ZUeczfHqRwcZ/IJ6tba9ZjqVChdnYJXlzh0jQNR7Q40Bj2l32zDcSoHn0Ed0XgZyP3TQnxuS8B3RDdUk5XCD1cQ+sY0jzhP3EVqclVg3X0UrZ8XhvgL7eqOzByse4/hN7IbGo2Gwk/WgArOglKUwlKcecWopspfFG4GtvQslCJz2byw8rAvsWDemIrPHYlofb0oXrIZpdhEybfb0bcIJ/D+2wl9/V5saac8ZkBG2fy4U3h3beWaH6c6nJhWp+B1awz65k0wb/wVe8YZ7L+dp/CD7/Hu3Y7AewegMejIn/UtqlK/P4fJkycTEBDg+n+zZs0YOXJkpZ8fEBDAlClTXP/XaDSMGzeOxo0bV/o1Ro8eTdOmF24GgoODmTRpUqWfX1dcU0k6XZgf58wpwrL9MH6jugFQ+PEaVIcTxWwj5E8T8emZgD4yFLXEgu+Ibnh3aonPbW3xGXhrg5+iVF1Sk7tBqqpiST6CPjIU/UXzX6y7M0Crxbf/rdiOZKH1MaLRajBENcYQ1RjzLwev8ao3J+vuDDR+XhhaX2ietZ/KxvHbWfxHPYyzsBTn6Vy0gb7Y9h3DvGEfPr0S0Bj0eHVrjWXLQbA74SZf67LsombHeNFFTSk2Y91zjJAZd4NTwbLjMCG926GUWPDp2x6ttxEAVVHL+jLr+UyLpKQk1q5dy2effYbRaOTBBx8kNrbyk5x1Oh3vvvsunTp1YufOnXTr1o3p06eju4HBOPHx8axZs4ZPP/0Up9PJtGnTSExMrMrbqTVl8+MOo48OQ9c0xPV4eVeGd6+22I9kofXzxpGV+78bo7I/vv3IaVS7A6+20a7n6UIDG+RasDVJPr0bpFrt2A6cxLt7azQX9cfpW4SjC/HHdvAkpnV7CXqq8ne5Nyvr7gwMcc3QNQ50PaZrHIiuaQjO8wWUrkjGb9xtaDQagp+7C9VidzVrOs8XogsP9ohlwGy//lbWF3tR2Gt8jBhim6KUWilasA6fPu3ReBnQeRlo9EpZ7cORlYvjt3M0+suUer9SjEajoXv37nTv3r3Kr+Hv789jjz3GY489VuUydOjQgVmzZlW5DLXO4cSWehzv7hX7YQ0tm6IN8MGefhrzlgME/W4I2pAA9NFlI1RVVcWSko7x1pgKcyr9RlX98xZlJORukMZoIOSl8RUuaABeHWMJ++BRHGfzCHpyRINoQw98aDBaP+8KJ7MuLIjwz1/AnpZJwL0DyoIM0De7UOt1ZOViTTlK45n3uUZk3sx8h3Ypu6j5XhhkoPEyEPavx7HuPoo+JhxDbMU+t9JVO7HuycDnjkQMLZvVdZFFbdHrCHlpAobYiqNOjW2jCXv/ERxZuQQ9NuzyHSoUFeuuo/j071DhYU84P9xNQu4GabQafHq0ueLjxjZRDWr+i1e75pc9ptFcaKK9EsVkpeCD7wl5abzHrOlnuCUMbqk4Z0yj0aAL8cd30JWb0/yGJeF7Z1dy//xf8v6yiNC/TkNjuLnn0Ymyv7v3RQNOLn7cGB+FMf7K1wfVZMV24CQhfxhb20VscOQ2QdQZ1e6gaME6Aib1xadvexznC+r9gIuaZks/TemPu1GdStkFsUscpSuSceYWubtowo2sqcfRBvm6mi9FzZGQE3VCdTop+W4HvoM6ltXgnAqmH3eDB4yuvBElSzZR8I+vUc1WAJRSC1pfbzQGaVRpyCzJRzC0imxQC0nUFTmzapmzoATr3mPYDp5Eo9Nh2piKV6fYBrdrdMmyreS/vQytjxHUslGFfqO7N7hNVn2HdUUb5IdqtuEwWTFv+JXAR4aiDZGLW0OjKiq21OM484ox/7QPQ0w4lh1pGOOj0DUKuP4LiEqRkKttGg1aPx+Cnx5d4bGGxqtzS5r864kKj+mjQuv9qMKa5tWpJbpGAZi3p6Fa7AQ9Nvx/+881rM9BAKig1aD196HRK5PLHtLQIK8PtUlCrpbpgvzQJcnWGsa4CHcXoV7QaDQYYsIxxFRuzUfhuTRaLV4dYtxdDI8nfXJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8loScEEIIjyUhJ4QQwmNJyAkhhPBYEnJCCCE8lt7dBai3VFAdTvcd3+kE1X2Hv5jqVNz3WTjqz+eAOz8HALW+fBACAFV16/dBdSpuO/bNRELuCgICAohQvLAMmwka95UjNjIKrda9le248Ciy7p+L042fQ6ROj5+fn/sKALSIbk76/33l1s8hFA2NGzd2XwGES5MmTQjZlYlzyBtuK4NRhZjhw912/JuFRlXl9vBKsrOzcTrdeNcOeHt7ExQUhEbjvitrQUEBFovFbccH0Ov1hIaGuvVzKCkpoaSkxG3HB9BqtYSGhqLT6dxaDgFOp5OcnBzcffkMCAhw+w1gfSchJ4QQwmPJwBMhhBAeS0JOCCGEx5KQE0II4bEk5IQQQngsCTkhhBAe6/8Dxa+NyRlN+ikAAAAASUVORK5CYII=)"
      ],
      "metadata": {
        "id": "VG5t1OjU5NP8"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Fonte: Raffel, C., & Ellis, D. P. (2015). Feed-forward networks with attention can solve some long-term memory problems. arXiv preprint arXiv:1512.08756."
      ],
      "metadata": {
        "id": "TIL27nc24-yK"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install tensorflow-datasets\n",
        "!pip install keras\n",
        "!pip install tensorflow\n",
        "!pip install -U gensim"
      ],
      "metadata": {
        "id": "lBhH-LrUBPQC"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "KdvNPSjGbgpV"
      },
      "outputs": [],
      "source": [
        "from keras.models import Sequential,Model\n",
        "from keras.layers import Dense,Embedding,Activation,SimpleRNN,BatchNormalization,RNN,Flatten,Input,LSTM,Bidirectional,Attention\n",
        "from tensorflow.keras.preprocessing.text import Tokenizer\n",
        "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
        "from keras.utils.np_utils import to_categorical\n",
        "from sklearn.metrics import classification_report\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.model_selection import train_test_split\n",
        "import gensim\n",
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "import tensorflow_datasets as tfds\n",
        "import pandas as pd\n",
        "from tensorflow.keras.optimizers import Adam"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "ds_train = tfds.load('imdb_reviews/plain_text', split='train', shuffle_files=True)\n",
        "ds_test = tfds.load('imdb_reviews/plain_text', split='test', shuffle_files=False)"
      ],
      "metadata": {
        "id": "LR_xXvK8BRWR"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_train = tfds.as_dataframe(ds_train)\n",
        "df_test = tfds.as_dataframe(ds_test)"
      ],
      "metadata": {
        "id": "V6WDpjh1BS8J"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_train.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "fecoQgq2jmrF",
        "outputId": "e7a42eec-b0da-451c-f12d-bf3295180bec"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   label                                               text\n",
              "0      0  b\"This was an absolutely terrible movie. Don't...\n",
              "1      0  b'I have been known to fall asleep during film...\n",
              "2      0  b'Mann photographs the Alberta Rocky Mountains...\n",
              "3      1  b'This is the kind of film for a snowy Sunday ...\n",
              "4      1  b'As others have mentioned, all the women that..."
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-06ee720f-018a-4c24-a791-4d38f88442c1\">\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>label</th>\n",
              "      <th>text</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>b\"This was an absolutely terrible movie. Don't...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0</td>\n",
              "      <td>b'I have been known to fall asleep during film...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0</td>\n",
              "      <td>b'Mann photographs the Alberta Rocky Mountains...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1</td>\n",
              "      <td>b'This is the kind of film for a snowy Sunday ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1</td>\n",
              "      <td>b'As others have mentioned, all the women that...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-06ee720f-018a-4c24-a791-4d38f88442c1')\"\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-06ee720f-018a-4c24-a791-4d38f88442c1 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-06ee720f-018a-4c24-a791-4d38f88442c1');\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": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def preprocessDataFrame(df):\n",
        "\n",
        "    dic = {}\n",
        "    dic['text'] = []\n",
        "    dic['label'] = []\n",
        "\n",
        "    review = [x.decode('utf-8') for x in df['text'].values]\n",
        "    #class_label = [x.decode('utf-8') for x in df['label'].values]\n",
        "\n",
        "    for idx,sentence in enumerate(review):\n",
        "        dic['text'].append(review[idx])\n",
        "        dic['label'].append(df['label'][idx])\n",
        "        \n",
        "    return pd.DataFrame.from_dict(dic)"
      ],
      "metadata": {
        "id": "fjJTCUMjBUtR"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_train = preprocessDataFrame(df_train)\n",
        "df_test = preprocessDataFrame(df_test)"
      ],
      "metadata": {
        "id": "7m7IV2NMBWHA"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_train.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "bIavxdu1kOBh",
        "outputId": "e21b94b7-41b9-4e65-b110-19cf0880ad81"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                text  label\n",
              "0  This was an absolutely terrible movie. Don't b...      0\n",
              "1  I have been known to fall asleep during films,...      0\n",
              "2  Mann photographs the Alberta Rocky Mountains i...      0\n",
              "3  This is the kind of film for a snowy Sunday af...      1\n",
              "4  As others have mentioned, all the women that g...      1"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-db2ab7ad-dd86-4b2c-94f1-e5e8481fa4cb\">\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>text</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>This was an absolutely terrible movie. Don't b...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>I have been known to fall asleep during films,...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Mann photographs the Alberta Rocky Mountains i...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>This is the kind of film for a snowy Sunday af...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>As others have mentioned, all the women that g...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-db2ab7ad-dd86-4b2c-94f1-e5e8481fa4cb')\"\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-db2ab7ad-dd86-4b2c-94f1-e5e8481fa4cb 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-db2ab7ad-dd86-4b2c-94f1-e5e8481fa4cb');\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": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_train['label'] = pd.Categorical(df_train['label'])\n",
        "y_train_int = df_train['label'].cat.codes\n",
        "\n",
        "df_test['label'] = pd.Categorical(df_test['label'])\n",
        "y_test_int = df_test['label'].cat.codes\n",
        "\n",
        "y_train = to_categorical(y_train_int)\n",
        "y_test = to_categorical(y_test_int)"
      ],
      "metadata": {
        "id": "_7_5cN9NBXyh"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_reviews = df_train['text'].to_list()\n",
        "\n",
        "vocabulary = {}\n",
        "\n",
        "for i in range(0,len(train_reviews)):\n",
        "    train_reviews[i] = train_reviews[i].lower()\n",
        "    for word in train_reviews[i].split():\n",
        "        if word not in vocabulary:\n",
        "            vocabulary[word] = 1\n",
        "        else:\n",
        "            vocabulary[word]+=1"
      ],
      "metadata": {
        "id": "NsIFIfO5HJhn"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "trunc_type = 'post'\n",
        "padding_type = 'post'\n",
        "max_length = 32\n",
        "vocab_size = len(vocabulary)"
      ],
      "metadata": {
        "id": "j1Pgp7vvHFmo"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_reviews = df_train['text'].to_list()\n",
        "\n",
        "vocabulary = {}\n",
        "\n",
        "for i in range(0,len(train_reviews)):\n",
        "    train_reviews[i] = train_reviews[i].lower()\n",
        "    for word in train_reviews[i].split():\n",
        "        if word not in vocabulary:\n",
        "            vocabulary[word] = 1\n",
        "        else:\n",
        "            vocabulary[word]+=1\n",
        "\n",
        "tokenizer = Tokenizer(num_words = vocab_size)\n",
        "tokenizer.fit_on_texts(df_train['text'])\n",
        "\n",
        "word_index = tokenizer.word_index\n",
        "print(f\"{word_index['blue']} is the index of word 'blue'\")\n",
        "print(f\"{word_index['launch']} is the index of word 'launch'\")\n",
        "\n",
        "train_sequences = tokenizer.texts_to_sequences(df_train['text'])\n",
        "test_sequences = tokenizer.texts_to_sequences(df_test['text'])\n",
        "\n",
        "train_padded = pad_sequences(train_sequences, maxlen=max_length, padding=padding_type, truncating=trunc_type)\n",
        "test_padded = pad_sequences(test_sequences, maxlen=max_length, padding=padding_type, truncating=trunc_type)"
      ],
      "metadata": {
        "id": "4j2yXN8zGpz6",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "7e091609-fc84-4e16-efb2-8cdf68cf9348"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1331 is the index of word 'blue'\n",
            "7292 is the index of word 'launch'\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = Sequential()\n",
        "\n",
        "model.add(Input(shape=(None,),dtype=\"int64\"))\n",
        "model.add(Embedding(input_dim = vocab_size, output_dim = 32))\n",
        "\n",
        "model.add(Bidirectional(LSTM(100)))\n",
        "model.add(Dense(2, activation='softmax'))\n",
        "\n",
        "adam = Adam(learning_rate=0.00002)\n",
        "\n",
        "model.compile(optimizer=adam, loss='categorical_crossentropy', metrics=['accuracy'])\n",
        "\n",
        "history = model.fit(train_padded, y_train, epochs=6, batch_size=32, validation_split=0.2, verbose=1)"
      ],
      "metadata": {
        "id": "x2T2s75Prm8h",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "43c79d0a-e7ad-48c9-af7e-ca1dc7734af3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/6\n",
            "625/625 [==============================] - 84s 126ms/step - loss: 0.6929 - accuracy: 0.5181 - val_loss: 0.6925 - val_accuracy: 0.5454\n",
            "Epoch 2/6\n",
            "625/625 [==============================] - 78s 125ms/step - loss: 0.6844 - accuracy: 0.5963 - val_loss: 0.6415 - val_accuracy: 0.6304\n",
            "Epoch 3/6\n",
            "625/625 [==============================] - 78s 124ms/step - loss: 0.5831 - accuracy: 0.6898 - val_loss: 0.5800 - val_accuracy: 0.6932\n",
            "Epoch 4/6\n",
            "625/625 [==============================] - 77s 123ms/step - loss: 0.5134 - accuracy: 0.7477 - val_loss: 0.5564 - val_accuracy: 0.7160\n",
            "Epoch 5/6\n",
            "625/625 [==============================] - 77s 123ms/step - loss: 0.4651 - accuracy: 0.7814 - val_loss: 0.5508 - val_accuracy: 0.7196\n",
            "Epoch 6/6\n",
            "625/625 [==============================] - 77s 123ms/step - loss: 0.4259 - accuracy: 0.8071 - val_loss: 0.5531 - val_accuracy: 0.7288\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.title('model loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['training set','validation set'], loc='upper left')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "hZdWpHyKgOm4",
        "outputId": "96e80a86-1520-42f7-c656-d1f42350de08"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "y_prob = model.predict(test_padded)\n",
        "y_pred = np.argmax(y_prob,axis=1)\n",
        "print(classification_report(y_test_int,y_pred))"
      ],
      "metadata": {
        "id": "axfJ9F9ggVnY",
        "outputId": "3a628d94-1075-46bc-8076-9644bb8297a6",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "782/782 [==============================] - 11s 13ms/step\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.72      0.67      0.70     12500\n",
            "           1       0.69      0.74      0.72     12500\n",
            "\n",
            "    accuracy                           0.71     25000\n",
            "   macro avg       0.71      0.71      0.71     25000\n",
            "weighted avg       0.71      0.71      0.71     25000\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Com Atenção"
      ],
      "metadata": {
        "id": "u_3NiJG-pS70"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "input_layer = Input(shape=(None,),dtype=\"int64\")\n",
        "\n",
        "embedding_layer = Embedding(input_dim = vocab_size, output_dim = 32)(input_layer)\n",
        "\n",
        "bilstm = Bidirectional(LSTM(100))\n",
        "\n",
        "query = bilstm(embedding_layer)\n",
        "values = bilstm(embedding_layer)\n",
        "\n",
        "query_value_attention_seq = Attention(use_scale=True, score_mode=\"dot\")([query, values])\n",
        "\n",
        "output_layer = Dense(2, activation='softmax')(query_value_attention_seq)\n",
        "\n",
        "model_att = Model(inputs=input_layer,outputs=output_layer)\n",
        "\n",
        "adam = Adam(learning_rate=0.0002)\n",
        "\n",
        "model_att.compile(optimizer=adam, loss='categorical_crossentropy', metrics=['accuracy'])\n",
        "\n",
        "history = model_att.fit(train_padded, y_train, epochs=6, batch_size=32, validation_split=0.2, verbose=1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5tRYgwSzpVO0",
        "outputId": "26246420-a90f-4098-fde6-8bc6545708a8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/6\n",
            "625/625 [==============================] - 107s 163ms/step - loss: 0.6933 - accuracy: 0.5009 - val_loss: 0.6931 - val_accuracy: 0.5062\n",
            "Epoch 2/6\n",
            "625/625 [==============================] - 99s 158ms/step - loss: 0.6932 - accuracy: 0.5001 - val_loss: 0.6932 - val_accuracy: 0.4938\n",
            "Epoch 3/6\n",
            "625/625 [==============================] - 99s 158ms/step - loss: 0.6932 - accuracy: 0.4974 - val_loss: 0.6931 - val_accuracy: 0.5066\n",
            "Epoch 4/6\n",
            "625/625 [==============================] - 100s 160ms/step - loss: 0.6932 - accuracy: 0.4992 - val_loss: 0.6932 - val_accuracy: 0.4938\n",
            "Epoch 5/6\n",
            "625/625 [==============================] - 99s 158ms/step - loss: 0.6931 - accuracy: 0.5051 - val_loss: 0.6933 - val_accuracy: 0.4938\n",
            "Epoch 6/6\n",
            "625/625 [==============================] - 99s 159ms/step - loss: 0.6758 - accuracy: 0.5522 - val_loss: 0.6113 - val_accuracy: 0.6788\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "y_prob = model_att.predict(test_padded)\n",
        "y_pred = np.argmax(y_prob,axis=1)\n",
        "print(classification_report(y_test_int,y_pred))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tMQ_SXTZ1U4H",
        "outputId": "9437c8ae-c8af-4f2a-df2f-4c9277189798"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "782/782 [==============================] - 20s 23ms/step\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.67      0.69      0.68     12500\n",
            "           1       0.68      0.66      0.67     12500\n",
            "\n",
            "    accuracy                           0.67     25000\n",
            "   macro avg       0.68      0.67      0.67     25000\n",
            "weighted avg       0.68      0.67      0.67     25000\n",
            "\n"
          ]
        }
      ]
    }
  ]
}