glove.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "glove.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qTZNq4F6P-ud"
      },
      "source": [
        "# Importing Dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "avhgFke8ehEi"
      },
      "source": [
        "import keras\n",
        "\n",
        "from keras.datasets import imdb\n",
        "\n",
        "from keras.layers import LSTM, Activation, Dropout, Dense, Input\n",
        "from keras.layers.embeddings import Embedding\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.preprocessing.text import Tokenizer\n",
        "from keras.preprocessing.sequence import pad_sequences\n",
        "\n",
        "from sklearn.preprocessing import LabelBinarizer\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import string\n",
        "import matplotlib.pyplot as plt\n",
        "import re"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4NGhYtyUe2cq",
        "outputId": "075b53bb-c112-411d-a4be-7f9262fa7103"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount(\"/content/drive\")"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "29Ak2QQ9ejcj"
      },
      "source": [
        "def load_data():\n",
        "    data_df = pd.read_excel('/content/drive/My Drive/dataset/dataset.xlsx', engine='openpyxl')\n",
        "    data_df.to_csv('output_reviews_top.csv')\n",
        "    return data_df\n",
        "data = load_data()\n",
        "# data = pd.read_excel('/content/drive/My Drive/dataset/dataset.xlsx', engine='openpyxl')\n",
        "\n",
        "data['Text'] = data['Text'].str.lower()\n",
        "top_data_df = pd.read_csv('output_reviews_top.csv')"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PmiU6T5ozyzw",
        "outputId": "8afe1ff6-756f-477d-d2d0-a3fddd94cd21"
      },
      "source": [
        "print(\"Number of rows per star rating:\")\n",
        "print(top_data_df['Sentiment'].value_counts())"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Number of rows per star rating:\n",
            "NEG    1000\n",
            "POS     999\n",
            "Name: Sentiment, dtype: int64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "uZz_vgWRzy1_",
        "outputId": "28b68e4b-b937-4b1e-ccbc-12b1a1a35d1e"
      },
      "source": [
        "# Function to map stars to sentiment\n",
        "def map_sentiment(sentiment):\n",
        "    if sentiment == 'NEG':\n",
        "        return -1\n",
        "    else:\n",
        "        return 1\n",
        "# Mapping stars to sentiment into three categories\n",
        "top_data_df['Sentiment_value'] = [map_sentiment(x) for x in top_data_df['Sentiment']]\n",
        "# Plotting the sentiment distribution\n",
        "plt.figure()\n",
        "pd.value_counts(top_data_df['Sentiment_value']).plot.bar(title=\"Sentiment distribution in df\")\n",
        "plt.xlabel(\"Sentiment\")\n",
        "plt.ylabel(\"No. of rows in df\")\n",
        "plt.show()\n",
        "\n"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "j9oSckMR0Rzh"
      },
      "source": [
        "\n",
        "## Selecting equal number of positive and negative reviews for training:\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ay1qAjdJ0Xwo"
      },
      "source": [
        "# Function to retrieve top few number of each category\n",
        "def get_top_data(top_n):\n",
        "    top_data_df_positive = top_data_df[top_data_df['Sentiment_value'] == 1].head(top_n)\n",
        "    top_data_df_negative = top_data_df[top_data_df['Sentiment_value'] == -1].head(top_n)\n",
        "    top_data_df_small = pd.concat([top_data_df_positive, top_data_df_negative])\n",
        "    return top_data_df_small"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cXFzcSQs0ZMX",
        "outputId": "17bc3772-3824-43f8-faa5-723281358b42"
      },
      "source": [
        "top_data_df_small = get_top_data(999)\n",
        "print(\"After segregating and taking equal number of rows for each sentiment:\")\n",
        "print(top_data_df_small['Sentiment_value'].value_counts())"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "After segregating and taking equal number of rows for each sentiment:\n",
            " 1    999\n",
            "-1    999\n",
            "Name: Sentiment_value, dtype: int64\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 359
        },
        "id": "Jz4F7rDz0cUQ",
        "outputId": "a2adae6b-efe4-4c53-d564-6b8f0f656a46"
      },
      "source": [
        "top_data_df_small.head(10)"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Unnamed: 0</th>\n",
              "      <th>URL</th>\n",
              "      <th>Text</th>\n",
              "      <th>Sentiment</th>\n",
              "      <th>Sentiment_value</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>http://www.imdb.com/title/tt0210075/usercomments</td>\n",
              "      <td>Girlfight follows a project dwelling New York ...</td>\n",
              "      <td>POS</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>http://www.imdb.com/title/tt0337640/usercomments</td>\n",
              "      <td>Hollywood North is an euphemism from the movie...</td>\n",
              "      <td>POS</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>http://www.imdb.com/title/tt0303549/usercomments</td>\n",
              "      <td>That '70s Show is definitely the funniest show...</td>\n",
              "      <td>POS</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3</td>\n",
              "      <td>http://www.imdb.com/title/tt0716825/usercomments</td>\n",
              "      <td>9/10- 30 minutes of pure holiday terror. Okay,...</td>\n",
              "      <td>POS</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>4</td>\n",
              "      <td>http://www.imdb.com/title/tt0182225/usercomments</td>\n",
              "      <td>A series of random, seemingly insignificant th...</td>\n",
              "      <td>POS</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>5</td>\n",
              "      <td>http://www.imdb.com/title/tt0347779/usercomments</td>\n",
              "      <td>A very good adaptation of the novel by amrita ...</td>\n",
              "      <td>POS</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>6</td>\n",
              "      <td>http://www.imdb.com/title/tt0095655/usercomments</td>\n",
              "      <td>Ah, Moonwalker, I'm a huge Michael Jackson fan...</td>\n",
              "      <td>POS</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>7</td>\n",
              "      <td>http://www.imdb.com/title/tt0298131/usercomments</td>\n",
              "      <td>Although the beginning of the movie in New Yor...</td>\n",
              "      <td>POS</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>8</td>\n",
              "      <td>http://www.imdb.com/title/tt0088915/usercomments</td>\n",
              "      <td>As many reviewers here have noted, the film ve...</td>\n",
              "      <td>POS</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>9</td>\n",
              "      <td>http://www.imdb.com/title/tt0828154/usercomments</td>\n",
              "      <td>Bear in mind, any film (let alone documentary)...</td>\n",
              "      <td>POS</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   Unnamed: 0  ... Sentiment_value\n",
              "0           0  ...               1\n",
              "1           1  ...               1\n",
              "2           2  ...               1\n",
              "3           3  ...               1\n",
              "4           4  ...               1\n",
              "5           5  ...               1\n",
              "6           6  ...               1\n",
              "7           7  ...               1\n",
              "8           8  ...               1\n",
              "9           9  ...               1\n",
              "\n",
              "[10 rows x 5 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lSzCu4dAQMeB"
      },
      "source": [
        "# Dataset cleaning and analysing"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tJx-rX2rfI3u"
      },
      "source": [
        "stopwords = [ \"a\", \"about\", \"above\", \"after\", \"again\", \"against\", \"all\", \"am\", \"an\", \"and\", \"any\", \"are\", \"as\", \"at\", \"be\", \"because\", \n",
        "             \"been\", \"before\", \"being\", \"below\", \"between\", \"both\", \"but\", \"by\", \"could\", \"did\", \"do\", \"does\", \"doing\", \"down\", \"during\",\n",
        "             \"each\", \"few\", \"for\", \"from\", \"further\", \"had\", \"has\", \"have\", \"having\", \"he\", \"he'd\", \"he'll\", \"he's\", \"her\", \"here\", \n",
        "             \"here's\", \"hers\", \"herself\", \"him\", \"himself\", \"his\", \"how\", \"how's\", \"i\", \"i'd\", \"i'll\", \"i'm\", \"i've\", \"if\", \"in\", \"into\",\n",
        "             \"is\", \"it\", \"it's\", \"its\", \"itself\", \"let's\", \"me\", \"more\", \"most\", \"my\", \"myself\", \"nor\", \"of\", \"on\", \"once\", \"only\", \"or\",\n",
        "             \"other\", \"ought\", \"our\", \"ours\", \"ourselves\", \"out\", \"over\", \"own\", \"same\", \"she\", \"she'd\", \"she'll\", \"she's\", \"should\", \n",
        "             \"so\", \"some\", \"such\", \"than\", \"that\", \"that's\", \"the\", \"their\", \"theirs\", \"them\", \"themselves\", \"then\", \"there\", \"there's\",\n",
        "             \"these\", \"they\", \"they'd\", \"they'll\", \"they're\", \"they've\", \"this\", \"those\", \"through\", \"to\", \"too\", \"under\", \"until\", \"up\",\n",
        "             \"very\", \"was\", \"we\", \"we'd\", \"we'll\", \"we're\", \"we've\", \"were\", \"what\", \"what's\", \"when\", \"when's\", \"where\", \"where's\",\n",
        "             \"which\", \"while\", \"who\", \"who's\", \"whom\", \"why\", \"why's\", \"with\", \"would\", \"you\", \"you'd\", \"you'll\", \"you're\", \"you've\",\n",
        "             \"your\", \"yours\", \"yourself\", \"yourselves\" ]"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VL5fXA9cfMwX"
      },
      "source": [
        "def remove_stopwords(data):\n",
        "  data['review without stopwords'] = data['Text'].apply(lambda x : ' '.join([word for word in x.split() if word not in (stopwords)]))\n",
        "  return data\n",
        "\n",
        "def remove_tags(string):\n",
        "    result = re.sub('<.*?>','',string)\n",
        "    return result\n",
        "    \n",
        "data_without_stopwords = remove_stopwords(data)\n",
        "data_without_stopwords['clean_review']= data_without_stopwords['review without stopwords'].apply(lambda cw : remove_tags(cw))\n",
        "data_without_stopwords['clean_review'] = data_without_stopwords['clean_review'].str.replace('[{}]'.format(string.punctuation), ' ')"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "id": "oudl0eDWftJd",
        "outputId": "7789add0-ca49-40b1-b2fb-42a06f1d9d87"
      },
      "source": [
        "data_without_stopwords.head()"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>URL</th>\n",
              "      <th>Text</th>\n",
              "      <th>Sentiment</th>\n",
              "      <th>review without stopwords</th>\n",
              "      <th>clean_review</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>http://www.imdb.com/title/tt0210075/usercomments</td>\n",
              "      <td>girlfight follows a project dwelling new york ...</td>\n",
              "      <td>POS</td>\n",
              "      <td>girlfight follows project dwelling new york hi...</td>\n",
              "      <td>girlfight follows project dwelling new york hi...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>http://www.imdb.com/title/tt0337640/usercomments</td>\n",
              "      <td>hollywood north is an euphemism from the movie...</td>\n",
              "      <td>POS</td>\n",
              "      <td>hollywood north euphemism movie industry went ...</td>\n",
              "      <td>hollywood north euphemism movie industry went ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>http://www.imdb.com/title/tt0303549/usercomments</td>\n",
              "      <td>that '70s show is definitely the funniest show...</td>\n",
              "      <td>POS</td>\n",
              "      <td>'70s show definitely funniest show currently t...</td>\n",
              "      <td>70s show definitely funniest show currently t...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>http://www.imdb.com/title/tt0716825/usercomments</td>\n",
              "      <td>9/10- 30 minutes of pure holiday terror. okay,...</td>\n",
              "      <td>POS</td>\n",
              "      <td>9/10- 30 minutes pure holiday terror. okay, no...</td>\n",
              "      <td>9 10  30 minutes pure holiday terror  okay  no...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>http://www.imdb.com/title/tt0182225/usercomments</td>\n",
              "      <td>a series of random, seemingly insignificant th...</td>\n",
              "      <td>POS</td>\n",
              "      <td>series random, seemingly insignificant thefts ...</td>\n",
              "      <td>series random  seemingly insignificant thefts ...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                URL  ...                                       clean_review\n",
              "0  http://www.imdb.com/title/tt0210075/usercomments  ...  girlfight follows project dwelling new york hi...\n",
              "1  http://www.imdb.com/title/tt0337640/usercomments  ...  hollywood north euphemism movie industry went ...\n",
              "2  http://www.imdb.com/title/tt0303549/usercomments  ...   70s show definitely funniest show currently t...\n",
              "3  http://www.imdb.com/title/tt0716825/usercomments  ...  9 10  30 minutes pure holiday terror  okay  no...\n",
              "4  http://www.imdb.com/title/tt0182225/usercomments  ...  series random  seemingly insignificant thefts ...\n",
              "\n",
              "[5 rows x 5 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RZah_hsvfZKP"
      },
      "source": [
        "reviews_list = []\n",
        "for i in range(len(data_without_stopwords['clean_review'])):\n",
        "  reviews_list.append(data_without_stopwords['clean_review'][i])\n",
        " \n",
        "sentiment = data_without_stopwords['Sentiment']"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6OlKdW-TR6ZU"
      },
      "source": [
        "y = np.array(list(map(lambda x: 1 if x==\"POS\" else 0, sentiment)))"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Crg7mwhnR_Wz"
      },
      "source": [
        "### Data split"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S9B2LohAgl4k"
      },
      "source": [
        "X_train, X_test,Y_train, Y_test = train_test_split(reviews_list, y, test_size=0.2, random_state = 45)"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jpEOD86bSB_Y"
      },
      "source": [
        "### Data Tokenization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8IeQmFRwgqI8"
      },
      "source": [
        "tokenizer = Tokenizer(num_words=5000)\n",
        "tokenizer.fit_on_texts(X_train)\n",
        "\n",
        "words_to_index = tokenizer.word_index\n",
        "#type(words_to_index)"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "isPi5LYDSXGr"
      },
      "source": [
        "# Making Global Vector Map"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wYrnkzA-g2if"
      },
      "source": [
        "def read_glove_vector(glove_vec):\n",
        "  with open(glove_vec, 'r', encoding='UTF-8') as f:\n",
        "    words = set()\n",
        "    word_to_vec_map = {}\n",
        "    for line in f:\n",
        "      w_line = line.split()\n",
        "      curr_word = w_line[0]\n",
        "      word_to_vec_map[curr_word] = np.array(w_line[1:], dtype=np.float64)\n",
        "  return word_to_vec_map"
      ],
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LlwtMOkouiKZ"
      },
      "source": [
        "Download and mount the file glove.6B.50d.txt\n",
        "\n",
        "https://www.kaggle.com/watts2/glove6b50dtxt\n",
        " \n",
        "https://nlp.stanford.edu/projects/glove/\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dwnyzSlwu6Eb"
      },
      "source": [
        "# !unzip '/content/drive/My Drive/dataset/glove.6B.50d.txt.zip' -d '/content/drive/My Drive/dataset/'"
      ],
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Av9P7mU_g9kB"
      },
      "source": [
        "word_to_vec_map = read_glove_vector('/content/drive/My Drive/dataset/glove.6B.50d.txt')\n",
        "maxLen = 150"
      ],
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_uLsjb2pz2BX"
      },
      "source": [
        "vocab_len = len(words_to_index)\n",
        "embed_vector_len = word_to_vec_map['moon'].shape[0]\n",
        "\n",
        "emb_matrix = np.zeros((vocab_len, embed_vector_len))\n",
        "\n",
        "for word, index in words_to_index.items():\n",
        "  embedding_vector = word_to_vec_map.get(word)\n",
        "  if embedding_vector is not None:\n",
        "    emb_matrix[index - 1, :] = embedding_vector\n",
        "\n",
        "embedding_layer = Embedding(input_dim=vocab_len, output_dim=embed_vector_len, input_length=maxLen, weights = [emb_matrix], trainable=False)"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "u4EykN1tUECO"
      },
      "source": [
        "### Defining Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uwEC4-pH1i5g"
      },
      "source": [
        "X_train_indices = tokenizer.texts_to_sequences(X_train)\n",
        "X_train_indices = pad_sequences(X_train_indices, maxlen=maxLen, padding='post')"
      ],
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7aEkYD8B8HZS"
      },
      "source": [
        "X_test_indices = tokenizer.texts_to_sequences(X_test)\n",
        "X_test_indices = pad_sequences(X_test_indices, maxlen=maxLen, padding='post')"
      ],
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8G1BRSsla1RY"
      },
      "source": [
        "## Helper Functions\n",
        "\n",
        "from sklearn.metrics import accuracy_score, confusion_matrix, classification_report\n",
        "from sklearn.metrics import precision_score\n",
        "from sklearn.metrics import recall_score\n",
        "from sklearn.metrics import accuracy_score\n",
        "import pickle\n",
        "\n",
        "def generate_report(pred_list, custom_model, model_name,y_test):\n",
        "  print(f'Confusion Matrix:\\n{confusion_matrix(y_test,pred_list)}')\n",
        "  print(f'\\nClassification Report:\\n{classification_report(y_test,pred_list)}')\n",
        "  acc = accuracy_score(y_test, pred_list) * 100\n",
        "  print(f'Accuracy: {acc}%')\n",
        "  prec = precision_score(y_test, pred_list, average='macro') * 100\n",
        "  print(f'Precision: {prec}%')\n",
        "  rec = recall_score(y_test, pred_list, average='macro') * 100\n",
        "  print(f'Recall: {rec}%')\n",
        "  pickle.dump(custom_model, open(model_name + '.sav', 'wb'))\n",
        "  print(model_name + ' trained Model Saved')\n",
        "  return acc, prec, rec"
      ],
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nPn-gx6sdAF5"
      },
      "source": [
        "\n",
        "## Bernoulli Naïve Bayes (BNB)\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SUbqypskU70P"
      },
      "source": [
        "from sklearn.naive_bayes import BernoulliNB\n",
        "classifier = BernoulliNB()\n",
        "classifier.fit(X_train_indices,Y_train)\n",
        "BNBpred=classifier.predict(X_test_indices)"
      ],
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3-rpC8BnU728",
        "outputId": "54b1716b-e9e0-4038-a0b9-235cae1267a4"
      },
      "source": [
        "bnb_metrics = [None, None, None]\n",
        "bnb_metrics[0], bnb_metrics[1], bnb_metrics[2] = generate_report(BNBpred, classifier, 'BernoulliNB_model',Y_test)"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Confusion Matrix:\n",
            "[[ 96  91]\n",
            " [109 104]]\n",
            "\n",
            "Classification Report:\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.47      0.51      0.49       187\n",
            "           1       0.53      0.49      0.51       213\n",
            "\n",
            "    accuracy                           0.50       400\n",
            "   macro avg       0.50      0.50      0.50       400\n",
            "weighted avg       0.50      0.50      0.50       400\n",
            "\n",
            "Accuracy: 50.0%\n",
            "Precision: 50.081300813008134%\n",
            "Recall: 50.08159473776706%\n",
            "BernoulliNB_model trained Model Saved\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BaYTqlCOdYpH"
      },
      "source": [
        "\n",
        "## Multinomial Naïve Bayes (MNB)\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tCUGwgh7U75x"
      },
      "source": [
        "from sklearn.naive_bayes import MultinomialNB"
      ],
      "execution_count": 26,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1AGHc8g5U78u"
      },
      "source": [
        "classifier = MultinomialNB()\n",
        "\n",
        "#train the model\n",
        "classifier.fit(X_train_indices,Y_train)\n",
        "MNBpred=classifier.predict(X_test_indices)"
      ],
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sng-giaDU7_s",
        "outputId": "7dbf9b04-7a4f-40dd-bdb7-f00e1f4a469d"
      },
      "source": [
        "mnb_metrics = [None, None, None]\n",
        "mnb_metrics[0], mnb_metrics[1], mnb_metrics[2] = generate_report(MNBpred, classifier, 'MultinomialNB_model',Y_test)\n",
        "\n"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Confusion Matrix:\n",
            "[[ 83 104]\n",
            " [ 92 121]]\n",
            "\n",
            "Classification Report:\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.47      0.44      0.46       187\n",
            "           1       0.54      0.57      0.55       213\n",
            "\n",
            "    accuracy                           0.51       400\n",
            "   macro avg       0.51      0.51      0.51       400\n",
            "weighted avg       0.51      0.51      0.51       400\n",
            "\n",
            "Accuracy: 51.0%\n",
            "Precision: 50.60317460317461%\n",
            "Recall: 50.59626923752856%\n",
            "MultinomialNB_model trained Model Saved\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cmF4bU7heeaV"
      },
      "source": [
        "\n",
        "## Maximum Entropy (ME)\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "frXdHU6zU8C_"
      },
      "source": [
        "from sklearn.linear_model import LogisticRegression"
      ],
      "execution_count": 29,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "T5n3mcScU8GO",
        "outputId": "bd8adeaa-a3aa-47af-9a1e-0704038ee5e3"
      },
      "source": [
        "classifier = LogisticRegression()\n",
        "\n",
        "# Train the Model\n",
        "classifier.fit(X_train_indices,Y_train)   \n",
        "LRpred = classifier.predict(X_test_indices)\n",
        "\n",
        "lr_metrics = [None, None, None]\n",
        "lr_metrics[0], lr_metrics[1], lr_metrics[2] = generate_report(LRpred, classifier, 'LinearRegression_model',Y_test)\n",
        "\n"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Confusion Matrix:\n",
            "[[ 94  93]\n",
            " [ 96 117]]\n",
            "\n",
            "Classification Report:\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.49      0.50      0.50       187\n",
            "           1       0.56      0.55      0.55       213\n",
            "\n",
            "    accuracy                           0.53       400\n",
            "   macro avg       0.53      0.53      0.53       400\n",
            "weighted avg       0.53      0.53      0.53       400\n",
            "\n",
            "Accuracy: 52.75%\n",
            "Precision: 52.59398496240601%\n",
            "Recall: 52.59847857196655%\n",
            "LinearRegression_model trained Model Saved\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ac6kN1cgfEdM"
      },
      "source": [
        "\n",
        "## Decision Tree (DE)\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P0jNbtqyU8I-"
      },
      "source": [
        "from sklearn.tree import DecisionTreeClassifier\n",
        "from sklearn.model_selection import cross_val_score\n"
      ],
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cX3junW9fLRk",
        "outputId": "7ff4d8ae-a489-40dc-92e0-6abd391c9a62"
      },
      "source": [
        "dt_clf = DecisionTreeClassifier()\n",
        "\n",
        "# Train the Model\n",
        "dt_clf.fit(X_train_indices,Y_train)   \n",
        "dt_pred = dt_clf.predict(X_test_indices)\n",
        "\n",
        "dt_metrics = [None, None, None]\n",
        "dt_metrics[0], dt_metrics[1], dt_metrics[2] = generate_report(dt_pred, dt_clf, 'DecisionTree_model',Y_test)\n",
        "\n"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Confusion Matrix:\n",
            "[[ 94  93]\n",
            " [ 97 116]]\n",
            "\n",
            "Classification Report:\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.49      0.50      0.50       187\n",
            "           1       0.56      0.54      0.55       213\n",
            "\n",
            "    accuracy                           0.53       400\n",
            "   macro avg       0.52      0.52      0.52       400\n",
            "weighted avg       0.53      0.53      0.53       400\n",
            "\n",
            "Accuracy: 52.5%\n",
            "Precision: 52.35852601518074%\n",
            "Recall: 52.36373678792901%\n",
            "DecisionTree_model trained Model Saved\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Vp4DydoOfkYg"
      },
      "source": [
        "## Support Vector Machine (SVM)\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "psMhtltUfLT4"
      },
      "source": [
        "from sklearn.svm import SVC\n",
        "from sklearn import svm\n"
      ],
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gRteWC42fLWf",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0bbe8a31-7dd5-4a7e-eb46-46be575c7902"
      },
      "source": [
        "%%time\n",
        "SVMclassifier = svm.SVC(kernel='rbf')\n",
        "SVMclassifier.fit(X_train_indices,Y_train)\n",
        "SVMpred = SVMclassifier.predict(X_test_indices)"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CPU times: user 831 ms, sys: 17.2 ms, total: 848 ms\n",
            "Wall time: 856 ms\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wNw4h6frfrAM",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5593ae7e-7c80-4e16-c16a-ea0fc8b9b07c"
      },
      "source": [
        "svm_metrics = [None, None, None]\n",
        "svm_metrics[0], svm_metrics[1], svm_metrics[2] = generate_report(SVMpred, SVMclassifier, 'SVM_model',Y_test)"
      ],
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Confusion Matrix:\n",
            "[[123  64]\n",
            " [113 100]]\n",
            "\n",
            "Classification Report:\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.52      0.66      0.58       187\n",
            "           1       0.61      0.47      0.53       213\n",
            "\n",
            "    accuracy                           0.56       400\n",
            "   macro avg       0.57      0.56      0.56       400\n",
            "weighted avg       0.57      0.56      0.55       400\n",
            "\n",
            "Accuracy: 55.75%\n",
            "Precision: 56.54712691194709%\n",
            "Recall: 56.36187893851523%\n",
            "SVM_model trained Model Saved\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eaAgPvxNrzsn"
      },
      "source": [
        "\n",
        "## Results comparision plots and tables\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sFbcz-faTFhb",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a61110a8-db6e-41eb-e824-7e0dd2d5f6b6"
      },
      "source": [
        "from tabulate import tabulate\n",
        "table = [['Model Name', 'Accuracy', 'Precision', 'Recall'], \n",
        "         ['Multinomial Naïve Bayes', mnb_metrics[0], mnb_metrics[1], mnb_metrics[2]], \n",
        "         ['Bernoulli Naïve Bayes', bnb_metrics[0], bnb_metrics[1], bnb_metrics[2]],\n",
        "         ['Support Vector Machine', svm_metrics[0], svm_metrics[1], svm_metrics[2]],\n",
        "         ['Maximum Entropy', lr_metrics[0], lr_metrics[1], lr_metrics[2]],\n",
        "         ['Decision Tree', dt_metrics[0], dt_metrics[1], dt_metrics[2]]]\n",
        "\n",
        "print(tabulate(table, headers='firstrow', tablefmt='fancy_grid'))"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "╒═════════════════════════╤════════════╤═════════════╤══════════╕\n",
            "│ Model Name              │   Accuracy │   Precision │   Recall │\n",
            "╞═════════════════════════╪════════════╪═════════════╪══════════╡\n",
            "│ Multinomial Naïve Bayes │      51    │     50.6032 │  50.5963 │\n",
            "├─────────────────────────┼────────────┼─────────────┼──────────┤\n",
            "│ Bernoulli Naïve Bayes   │      50    │     50.0813 │  50.0816 │\n",
            "├─────────────────────────┼────────────┼─────────────┼──────────┤\n",
            "│ Support Vector Machine  │      55.75 │     56.5471 │  56.3619 │\n",
            "├─────────────────────────┼────────────┼─────────────┼──────────┤\n",
            "│ Maximum Entropy         │      52.75 │     52.594  │  52.5985 │\n",
            "├─────────────────────────┼────────────┼─────────────┼──────────┤\n",
            "│ Decision Tree           │      52.5  │     52.3585 │  52.3637 │\n",
            "╘═════════════════════════╧════════════╧═════════════╧══════════╛\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "v_Mvrk61rb2V",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 346
        },
        "outputId": "fcc11290-b720-4b87-8690-73ae5ae21e0c"
      },
      "source": [
        "  \n",
        "N = 5\n",
        "ind = np.arange(N) \n",
        "width = 0.25\n",
        "  \n",
        "acc = [mnb_metrics[0], bnb_metrics[0], svm_metrics[0], lr_metrics[0], dt_metrics[0]]\n",
        "bar1 = plt.bar(ind, acc, width)\n",
        "  \n",
        "prec = [mnb_metrics[1], bnb_metrics[1], svm_metrics[1], lr_metrics[1], dt_metrics[1]]\n",
        "bar2 = plt.bar(ind+width, prec, width)\n",
        "  \n",
        "recall = [mnb_metrics[2], bnb_metrics[2], svm_metrics[2], lr_metrics[2], dt_metrics[2]]\n",
        "bar3 = plt.bar(ind+width*2, recall, width)\n",
        "  \n",
        "# plt.xlabel(\"Dates\")\n",
        "plt.ylabel(\"Percentage\")\n",
        "plt.title(\"Comparison\")\n",
        "  \n",
        "plt.xticks(ind+width,['Multinomial NB','Bernoulli NB','SVM','Logistic Regression', 'Decision Tree'])\n",
        "plt.xticks(rotation = 45)\n",
        "plt.legend( (bar1, bar2, bar3), ('Accuracy', 'Precision', 'Recall'), bbox_to_anchor = (1.05, 0.6))\n",
        "plt.show()\n",
        "\n"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 346
        },
        "id": "R1VQAq8K_voi",
        "outputId": "9babb4e9-1549-44b8-e6c0-c37864196145"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "models = ['Multinomial NB','Bernoulli NB','SVM','Logistic Regression', 'Decision Tree']\n",
        "recall = [mnb_metrics[2], bnb_metrics[2], svm_metrics[2], lr_metrics[2], dt_metrics[2]]\n",
        "precision = [mnb_metrics[1], bnb_metrics[1], svm_metrics[1], lr_metrics[1], dt_metrics[1]]\n",
        "line_chart1 = plt.plot(models, recall, 'o-g')\n",
        "line_chart2 = plt.plot(models, precision, 'o-y')\n",
        "plt.title('Recall vs Precision')\n",
        "plt.ylabel('Score')\n",
        "plt.xticks(rotation = 45)\n",
        "plt.legend(['Recall', 'Precision'], bbox_to_anchor = (1.3, 0.6))\n",
        "plt.show()"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAd4AAAFJCAYAAAA1/Uw4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXxU1f3/8ddnJvu+EsK+E/Ydq+KOKCIuuFCXVlFAcbetWlur/Wnbr1at1qooCmhb9wW1gIg7KlIIQtiEAAECYSb7vs/M+f0xExsxQAKZuVk+z8cjj0zu3Ln3fYeQz5xzz71HjDEopZRSKjBsVgdQSimlOhMtvEoppVQAaeFVSimlAkgLr1JKKRVAWniVUkqpANLCq5RSSgWQFl7VYYnIFyIy2/f4WhH52upMx0tEThGRHc1Y73ci8mIgMimlWkYLrwoIEdkrItUiUiEiThF5SUSirM7VmnyFvsZ3jAUi8q6IpLbmPowxXxljBjdjvb8YY2a35r6VUq1DC68KpOnGmChgNDAGuNfiPP5wi+8YBwFxwBOHriAiQQFPpZRqM7TwqoAzxjiBj/AWYABE5GcislpESkQkQ0ROb/RcgogsFpGDIlIsIu/5lseLyFIRyfctXyoiPVqaR0Q+FJFbDlmWISIzxOsJEckTkTIR2Swiw5txjEXAO8Bw3/b2isg9IrIJqBSRoGM85tNF5ECj9e4RkRwRKReRHSJylm/5H0Xk343Wu0BEtvr29YWIDGn03F4R+Y2IbBKRUhF5Q0TCWvo+KqWaRwuvCjhfcZwK7PL93B1YBvwJSAB+A7wjIsm+l/wLiACGAV34XyvSBiwGegO9gGrg6WOI9BpwRaN8Q33bXAZMAU7F24KNBS4HCptxjEnAJcCGRouvAKbhbQmncGzH3Hgfg4FbgAnGmGjgHGBvE+sN8h3jHUAysBz4j4iENFrtcuBcoC8wErj2aMeolDo2WnhVIL0nIuXAfiAPeMC3/GpguTFmuTHGY4z5GEgHzvOdI50K3GiMKTbG1BtjvgQwxhQaY94xxlQZY8qBPwOnHUOuJcBoEent+/kq4F1jTC1QD0QDaYAYY743xjiOsK2nRKQEyAAcwK8aP2eM2W+MqT7WYz6EGwgFhopIsDFmrzFmdxPrzQSWGWM+NsbUA48B4cBJh2Q76Gup/4dGvRFKqdalhVcF0kW+ltnpeAtZkm95b+AyXzdoia9wTQJSgZ5AkTGm+NCNiUiEiDwvIvtEpAxYBcSJiL0loXxFexnwc9+iK4BXfM99hrcV/QyQJyILRCTmCJu7zRgTZ4zpboy5yhiT3+i5/Y0eH9MxH5J7F95W7B992V4XkW5NrNoN2NfodR5flu6N1nE2elwFdKiBb0q1JVp4VcD5Wm8v4W15gbcI/MtXsBq+Io0xD/ueSxCRuCY29WtgMHCCMSYGb5cwgBxDrNeAK0TkRCAM+LxR3qeMMeOAoXi7nO86hu0DNJ4K7FiP+ccbNOZVY8wkvIXcAI80sdpB3/MAiIjgLe45x3gcSqnjoIVXWeVJ4GwRGQX8G5guIueIiF1EwnyDiHr4unU/BJ71DaYKFpGGAhuN97xuiYgk8L+u62OxHG9xehB4w9cqREQmiMgJIhIMVAI1gOc49tPgWI/5ByIyWETOFJFQX67qw2R7E5gmImf5juPXQC2wuhWOQynVQlp4lSV8XbD/BO43xuwHLgR+B+TjbfHdxf9+P3+B91zrdrznhu/wLX8S77nKAmANsOI48tQC7wKTgVcbPRUDvAAU4+2uLQQePdb9NNrfsR5zY6HAw3iP34l3ENZPLtEyxuzAe075H751p+O9tKvueI9DKdVyYow5+lpKKaWUahXa4lVKKaUCSAuvUkopFUBaeJVSSqkA0sKrlFJKBVC7uFl7UlKS6dOnj9UxlFKqXVm/fn2BMSb56GuqQGoXhbdPnz6kp6dbHUMppdoVEdl39LVUoGlXs1JKKRVAWniVUkqpANLCq5RSSgWQFl6llFIqgLTwKqWUUgGkhVepY5Cb+wrfftuHL76w8e23fcjNfcXqSEqpdqJdXE6kVFuSm/sKO3bMxeOpAqC2dh87dswFICXlKiujKaXaAW3xKtVCWVm//6HoNvB4qsjK+r1FiZRS7YkWXqVaqLY2u0XLlVKqMS28SrVQaGivFi1XSqnGtPAq1UL9+v0ZkdAfLRMJol+/P1uUSCnVnmjhVaqFUlKuosbWA48Bj4Fqt+AxLsLDB1odTSnVDmjhVaqFXst4Dlv9bpbkwFmr4NJvDYW1wpqMy/B46q2Op5Rq47TwKtVCn27+HcE2+NDp/bnKDX/faQhxZ3PgwBPWhlNKtXlaeJVqAWMMJ8YVs6Mcdlf+b/nXhfB1Aezd+0eqq/dYF1Ap1eZp4VWqBSoqvqN/FCx3/PS5d3K7I2InM3MexpjAh1NKtQtaeJVqAYdjIYZgVhX8+KZvgnDbSQ/Rt+9fKC7+iLy81y1KqJRq67TwKtVMbnc1ubmvktLlciLCuhBiD0EQukR2wWDYXrCd7t1vIjp6Irt23UF9fZHVkZVSbZAWXqWaqaDgXdzuUoqCTuBg+UGePe9ZPA94yP1NLrPHzOaxbx9jbU46gwcvoL6+kKyse6yOrJRqg7TwKtVMDsciwsL68uK2dKJCopg5fOYPzz025TG6RXfjug+uIzgsjZ49f4XD8SIlJassTKyUaou08CrVDNXVWZSUfEZc0pW8ufUtrhpxFVEhUT88HxsWy4LzF7AtfxsPfvkgffo8QFhYX98sRrUWJldKtTVaeJVqBqfzJUBYVRhOtauaOWPn/GSdqQOncu3oa3nkm0fYmLudQYPmU129g+zshwOeVynVdvm18IrIXhHZLCIbRSS90fJbRWS7iGwVkb/6M4NSx8sYN07nYuLjz+HZDW8xNnUs47qNa3Ldv035G10iuzDr/VlExZ5Bly5XsG/fX6is3B7g1EqptioQLd4zjDGjjTHjAUTkDOBCYJQxZhjwWAAyKHXMios/obb2AFWhp5KRm9Fka7dBfHg8z5//PJvzNvPnVX9mwIAnsNsjyMy8AWM8AUytlGqrrOhqngc8bIypBTDG5FmQQalmczgWEhSUyEuZO4kIjuDKEVcecf3pg6dz9cir+cvXf2FbkYN+/R6ltHQVTufiACVWSrVl/i68BlgpIutFZK5v2SDgFBH5r4h8KSITmnqhiMwVkXQRSc/Pz/dzTKWaVldXQEHBeyQkXc4rW97k58N+TkxozFFf9/dz/05ieCKz3p9FUpdfEBt7Crt330VdnX7OVKqz83fhnWSMGQtMBW4WkVOBICAB+BlwF/CmiMihLzTGLDDGjDfGjE9OTvZzTKWalpf3CsbUs7Ysmcr6SuaMO3w3c2MJ4QnMnzafjc6NPPLNXxk06Hnc7kp27brTz4mVUm2dXwuvMSbH9z0PWAJMBA4A7xqvtYAHSPJnDqWOhTEGh2Mh0dHjeSZjOSO6jOCE7ic0+/UXD7mYmcNm8tCqh8iqcNGr173k5b1KUdFHfkytlGrr/FZ4RSRSRKIbHgNTgC3Ae8AZvuWDgBCgwF85lDpW5eXrqazcTH3EFNIPpjNn7Bya6Jw5on9M/QdxYXHMen8W3XveRXj4YDIz5+F2V/kptVKqrfNnizcF+FpEMoC1wDJjzApgEdBPRLYArwPXGJ3KRbVBTudCbLYwXstyEhYUxtUjr27xNpIjk3nmvGdY71jP498+xeDBz1NTs4e9ex/0Q2KlVHsQdPRVjo0xJgsY1cTyOqDlf8GUCiC3u4rc3FeJT7yIxd+8xWVDLyM+PP6YtnXZsMu4ZOsl/PHLP3Jh2ga6dr2O/fsfIyXlCqKifvJfRCnVwemdq5RqQn7+u7jdZWyp6kV5XTlzx809+ouO4JnzniE6JJpZ78+id9//Izg4gR075mKMu5USK6XaCy28SjXB6VxEWFg/nt60iiFJQzi558nHtb2UqBT+MfUfrM1Zy9PpLzNgwJOUl68lJ2d+KyVWSrUXWniVOkR19W5KSj7HFj2N1QfWMHvs7BYPqmrKz4f/nAsHX8gfPv8DxbaxxMdPYc+e31Fbm9MKqZVS7YUWXqUO4Z0QwcY72RWE2EP45ahftsp2RYT50+YTERzB9f+5nv4DnsYYFzt33toq21dKtQ9aeJVqxDshwkvExk9mwaYlzBgyg6SI1rvMPDU6lb+f+3dW71/NC5uW0qfPAxQULCE//71W24dSqm3TwqtUI0VFH1Nbe4Cs+sGU1JQwd+zxDapqytUjr2bawGn8/rPfUxt5AZGRI9m58xZcrrJW35dSqu3RwqtUI07nQoKDk3h683cMSBjA6X1Ob/V9iAjPn/88IfYQrv/gBgYOeo66uoPs2XNfq+9LKdX2aOFVysc7IcL7BMecxxfZ3xzTnaqaq3tMd5445wm+yv6Kf21Pp3v3m8nJeZqysrV+2Z9Squ3QwquUT27uvzGmnuUOCLIFcc2oa/y6v2tHX8s5/c/ht5/+FomfTUhIN3bsmIvHU+/X/SqlrKWFVym8EyI4nQuJjBrP0xuXcVHaRaREpfh1nyLCC9NfwC525i67kwED/05lZQYHDjzp1/0qpaylhVcpoLw8ncrKLRw0oyisLmTO2OZN/3e8esb25PEpj/P53s95d18+iYkXsnfvA1RX7wnI/pVSgaeFVynA4ViIzRbO89t30ieuD5P7TQ7YvmePnc3kfpO56+O7iEi5BxE7mZnz0LlDlOqYtPCqTs/triIv7zXCYs7hw6xVzB4zG5sE7r9GQ5ezMYZ5Hz1A375/prj4I/LyXg9YBqVU4GjhVZ1efv47uN1lfF4Qjl3szBozK+AZ+sT14dGzH+XjrI9ZkRtGdPREdu26g/r6ooBnUUr5lxZe1ek1TIjwt42fcP6g8+kW3c2SHDeMv4HT+5zOr1b+hthuD1JfX0hW1j2WZFFK+Y8WXtWpeSdE+IIi+wnkVeYHbFBVU2xiY+EFC3EbN7d+9iQ9etyJw/EiJSWrLMuklGp9WnhVp+ZwLAZsLNp5kB4xPTh3wLmW5ukX34+Hz3qYFbtW8E35AMLC+viu7a21NJdSqvVo4VWdVsOECOExp/HuzlVcP+Z67Da71bG4eeLNnNLrFG7/6B4Suj9EdfUOsrMftjqWUqqVaOFVnVZR0Urq6nL4b0kCIsJ1Y66zOhLwvy7nWnctv/76Dbp0uYJ9+/5CZeV2q6MppVqBFl7VaTkcCwkKTuKxjNWcO+BcesX2sjrSDwYmDuQvZ/6FpZlL2Vh3MnZ7BJmZN+q1vUp1AFp4VadUV5dPYeEHVIacxP4yh1+m/ztet51wGyf2OJFbPvoDSd3vo7T0S5zOxVbHUkodJ78WXhHZKyKbRWSjiKT7lv1RRHJ8yzaKyHn+zKBUUxomRHhtTwmpUalMGzTN6kg/YbfZWXThIqrqq7h37VfExp7C7t2/oa4uz+poSqnjEIgW7xnGmNHGmPGNlj3hWzbaGLM8ABmU+oExBodjIaERo/nX9q+ZNXoWQbYgq2M1KS0pjYfOeIj3tr/PLrkQt7uCXbvutDqWUuo4aFez6nTKy9dRVbWVzVU98RgPs8fOtjrSEf3qxF8xsftE5q38PxJTbyMv71WKij6yOpZS6hj5u/AaYKWIrBeRxifRbhGRTSKySETi/ZxBqR9pmBDhsYzvOLvf2fSN72t1pCOy2+wsvnAx5XXl/Ckji/DwwWRmzsPtrrI6mlLqGPi78E4yxowFpgI3i8ipwHygPzAacACPN/VCEZkrIukikp6fn+/nmKqzaJgQoS70JHaW5DB3XNsbVNWUoclDeeC0B3hj2xIcoVdQU7OHvXsftDqWUuoY+LXwGmNyfN/zgCXARGNMrjHGbYzxAC8AEw/z2gXGmPHGmPHJycn+jKk6kfz8t3G7y1mSU0tyRDIXDL7A6kjNdvfJdzMudRxzP36GuKQr2b//MSoqNlkdSynVQn4rvCISKSLRDY+BKcAWEUlttNrFwBZ/ZVDqUA7HIoJD+/L8ltXMGj2LEHuI1ZGaLcgWxOILF1NSU8KTmdUEByewY8ccjHFbHU0p1QL+bPGmAF+LSAawFlhmjFkB/NV3idEm4AxAh2iqgKiq2kVp6ZfsqhuIux0MqmrKiJQR3Hfqfby8eQklEVdRXr6WnJz5VsdSSrWAtIc74YwfP96kp6dbHUO1c1lZvyc7+2F+tbU7cVED+Oyaz6yOdEzq3fVMfHEijvKDLD1rGNUV6Uyc+D2hod2tjqbaGBFZf8ilnKoN0MuJVKfg8bhwOl/CEzaBDQX7LZ3+73gF24NZfOFiCquLWJwdizEudu681epYSqlm0sKrOoXi4pXU1R1kZa6NxPBELh5ysdWRjsvorqO5d9K9PLvxPaojL6egYAkFBe9bHUsp1QxaeFWn4HAsxB6UyFOb1/HLUb8kLCjM6kjH7b5T72NElxHM+XIlYRHDyMy8GZerzOpYSqmj0MKrOryGCRFyPEOpdrvadTdzYyH2EBZfuJiDFXm8m9eXurqD7Nlzn9WxlFJHoYVXdXi5uf/CGBfzt+9jUq9JDEkeYnWkVjOu2zjuPvluHl2/FFfkNHJynqasbK3VsZRSR6CFV3VoDRMiEDKEVY7sDtPabeyB0x5gaPJQ5q3+juCQVHbsmIvHU291LKXUYWjhVR1aeflaqqq2sao4mriwOC4bepnVkVpdaFAoiy9cTFapk09Lh1JZmcGBA09aHUspdRhaeFWH5nAsRGwRPLFpA1ePuJrw4HCrI/nFxO4T+fWJv+aBtZ/gCT+RvXsfoLp6j9WxlFJN0MKrOiy3u5K8vNcpZBgldfXtZkKEY/X/Tv9/DE4czK/T9wF2MjPn0R5ukKNUZ6OFV3VYDRMiLN6dxwndT2BEygirI/lVeHA4iy5cREaBg/XVoygu/oi8vNetjqWUOoQWXtVhORyLILgHS7P3dfjWboOTep7EHT+7g9+s+QYTMphdu+6gvr7I6lhKqUa08KoOqapqJ6Wlq0gvSyY6JJqZw2ZaHSlg/nTmn+iXMIA/bi6nvr6QrKx7rI6klGpEC6/qkJzOxYCNJ7Zs46oRVxEZEml1pICJCI5g0QWL+MrpILN+FA7Hi5SUfGV1LKWUjxZe1eE0TIhQZhvKwapa5ozreNfuHs0pvU/hlom3cPua7yCoK5mZc/F4aq2OpZRCC6/qgIqLP6KuzsGb2WWMTR3L2NSxVkeyxP+d9X+kxvTliUyhqmo72dkPWx1JKYUWXtUBORwLwR7P61nZzB3bOQZVNSUyJJKFFyzkg/0ODrgHs2/fX6is3G51LKU6PS28qkOpq8ujsPA/bKvqTmhQBFeMuMLqSJY6o+8ZzBs/j9v/uwMjoWRm3qjX9iplMS28qkNpmBDhH9t2c8XwK4gJjbE6kuUemfwI0RG9+ef+CEpLv/QNPFNKWUULr+owGiZEqLb1ZXtZdYecEOFYRIdG8+IFL/LSrlyKPD3Yvfs31NXlWR1LqU5LC6/qMMrK/ktV1ff856CHkSkjmdh9otWR2ozJ/SYze+wcfr0+B5ernF277rQ6klKdlhZe1WE4nQtBwnhp1z7mjJ2DiFgdqU159OxHcQV1Z1leHHl5r1JU9JHVkZTqlLTwqg6hYUKEPXW9MRLG1SOvtjpSmxMbFssL01/g79sLqDQJZGbOw+2usjqWUp2OXwuviOwVkc0islFE0g957tciYkQkyZ8ZVOeQl/cWbncFz+3I5vJhlxMXFmd1pDbp3AHncvWoWfxhUwk1NXvYu/dBqyMp1ekEosV7hjFmtDFmfMMCEekJTAGyA7B/1Qk4nYuos6WwtlAHVR3N3875G7nurnxbEsf+/Y9RUbHJ6khKdSpWdTU/AdwN6AWF6rhVVWVSWvoVn+aFMCRpCCf3PNnqSG1aXFgcz5//PP+3tYQ6E8qOHXMxxm11LKU6DX8XXgOsFJH1IjIXQEQuBHKMMRlHeqGIzBWRdBFJz8/P93NM1Z55r0u18+LO/TqoqpnOH3Q+Fw39BY/vqKG8/L/k5My3OpJSnYa/C+8kY8xYYCpws4icCvwOuP9oLzTGLDDGjDfGjE9OTvZzTNVeeSdEeJmDrp5UuEP45ahfWh2p3Xjy3CfZVJnM9ooo9uz5HbW1OVZHUqpT8GvhNcbk+L7nAUuA04C+QIaI7AV6AN+JSFd/5lAdV1HRCurqHLyclcclQy4hMSLR6kjtRkJ4AvOnPcdDWyuoc1ezc+etVkdSqlPwW+EVkUgRiW54jHcw1TpjTBdjTB9jTB/gADDWGOP0Vw7VsTmdC3FLDJ86q5g7rvNOiHCsLkq7iNMGXMHiPR4KCpZQUPC+1ZGU6vCC/LjtFGCJ73xbEPCqMWaFH/enOpm6ulwKC5fybXEK/eJTOK33aVZHapeemvoUI+d/wnndKgjJvJm4uDMICtJ7XHdm69ev7xIUFPQiMBy930NLeYAtLpdr9rhx45q8N6vfCq8xJgsYdZR1+vhr/6rjczq9EyK8sDOHW076qw6qOkZJEUk8NfVZ7l9xGc+MPciePfcxcOBTVsdSFgoKCnqxa9euQ5KTk4ttNptefdICHo9H8vPzhzqdzheBC5paRz/JqHbJGIPTuZAiTyqOmmCuGX2N1ZHatUuHXsrwnpfxn4NCTs7TlJWttTqSstbw5OTkMi26LWez2UxycnIp3t6CptcJYB6lWk1Z2Rqqqrbz+t4yLkq7iC6RXayO1O49fd7TvO2Io9QVxI4dc/B46q2OpKxj06J77Hzv3WHrqxZe1S45HAvxEMrSg5V6p6pW0iWyC3895xke215PZeUmDhx40upISnVIWnhVu+NyVZCf/wYZFfF0je7LWf3OsjpShzFz2EySky9mdaGNrD33U129x+pIqpOy2+3j0tLShg4cOHDYmWeeOaCgoMDemtvv3r37CIfDEQQQERExpjW3fTTNLrwiEi4ig/0ZRqnmyM/3ToiwaKeT2WNnYxP9/NhaRIRnpz3LS/ujqXXXkZk5D2O0x1Ed2XPrnkvo9ni3Ebb/ZxvX7fFuI55b91zC8W4zNDTUs3379m07d+7cGhcX53r00Uc7zJ2UmvUXS0SmAxuBFb6fR4vIB/4MptThOJ2LqDAJfF9uY9boWVbH6XC6RnXl/jOfZkGWh+Lij8jLe93qSKoNe27dcwl3rryzt6PCEWIwOCocIXeuvLN3axTfBj/72c8qc3JyQgC2bt0aesoppwwcNmzYkHHjxg3esGFDGMD+/fuDzj777P6DBw8eOnjw4KEff/xxJMDkyZP7Dxs2bMiAAQOGPfbYY21iNrzmXk70R2Ai8AWAMWajiPT1UyalDquqagelpV/z3oFIpg++gNToVKsjdUhXjbiKt7a+wY7yZdh23kpCwjkEB7fa31HVjlz3/nU9t+RtiTjc8xnOjMg6T92PruWrcdXYbl9xe59FGxc12Uod3mV41aILF+1vzv5dLheff/559PXXX18AMHv27N4LFizYN2LEiNrPPvssct68eb3WrFmTeeONN/Y65ZRTyu+///7dLpeL0tJSO8Arr7yyNyUlxV1RUSFjxowZevXVVxd37drV0llBmlt4640xpYdcJ6n9TyrgHI7FGGy8u7+Sf1+mg6r8RUSYf/7zTHtpMI9FFbJ7992kpb1odSzVBh1adI+2vLlqa2ttaWlpQ3Nzc4P79+9fc9FFF5WVlpbaNmzYEHXZZZf1/2E/dd79rF69Ovrtt9/eAxAUFERiYqIb4JFHHklZtmxZHIDT6QzeunVrWNeuXSuPJ9vxam7h3SoiVwJ2ERkI3Aas9l8spX7K43GRm/syO6sTiQoP45z+51gdqUPrFt2N20/5B29tnsXPZSFdu15DXNwpVsdSAXa0lmm3x7uNcFQ4Qg5dnhqVWrd2ztodx7rfhnO85eXlttNPP33gww8/3OWmm24qiI6Odm3fvn1bc7axdOnS6C+//DI6PT19e3R0tGfixImDq6urLR8U0twAtwLDgFrgVaAUuMNfoZRqSlHRh9TVOfnn7nyuH3M9dlurDnJUTbhm1DU47GfjrBG2fD8Lj6fW6kiqjbn/1PtzwoLCPI2XhQWFee4/9f5Wme4qOjra89RTT2U/++yzKdHR0Z4ePXrULVq0KB7A4/Hw7bffhgOcfPLJ5Q0DsFwuF4WFhfaSkhJ7bGysOzo62rNhw4awjIyMyNbIdLyOWnhFxA4sM8b83hgzwfd1nzGmJgD5lPqBw7GQWhPJ2mLhujHXWR2nUxARnp62kOf2hOOq3c2+fQ9bHUm1MTdOuLHoiSlP7EuNSq0ThNSo1Lonpjyx78YJNxa11j5OPvnk6rS0tOoFCxYkvPbaa1mLFy9OGjx48NCBAwcOe+edd+IA5s+fn/3ll19GDxo0aOjw4cOHbtiwIeySSy4pdblc0q9fv2F33XVX91GjRlnaxdxAmnOpgIh8CswwxpT6P9JPjR8/3qSnp1uxa9VG1NY6+fbbHvzHGcYOz+ksvXKp1ZE6lRe/e5F9u+ZwRpcgTpiwmcjINKsjqWYQkfXGmPEtfV1GRsbeUaNGFfgjU2eRkZGRNGrUqD5NPdfcruYKYLOILBSRpxq+Wi2hUkeRm/svwM1b2XqnKitcP+Z6ttSfSpXLxaZt1+q1vUodh+YW3neBPwCrgPWNvpTyO2MMDsdCcuricdlTmTZomtWROh0R4clp/2TxvlBqK/+Lw7HI6khKtVvNGtVsjHlZREKAQb5FO4wxegd1FRBlZd9SXb2DV/cI1435HUE2f04jrQ6nd1xvzh75NzJybsadeRtJSdMJCdHJKZRqqebeuep0YCfwDPAskCkip/oxl1I/cDgW4iaYz/MM14+53uo4ndrc8TfyTeVEPJ4qNn1/g9VxlGqXmtvV/DgwxRhzmjHmVOAc4An/xVLKy+WqIC/vDb4uDOaUvlPoG683TLOSTWw8ct7rvJUTTEXxexQWrrA6klLtTnMLb7Ax5ocLoY0xmUCwfyIp9T/5+W/i8VTyVnYVc8fOtTqOAvrG92XUoEfIroLvtvwCt7vK6khKtSvNLbzpIvKiiJzu+3oB0Ot7lN85HIsockWR705m+uDpVsdRPvMm3s6nZSMJNgVsybzL6jiqA2o8LeDUqVP7lZeXH/cdp+64445u7yDPCgkAACAASURBVL33XvThnv/rX/+a/PTTTyce736OprkHMg/YhvdWkbf5Hs/zVyilACort1NW9g1vZ1cya/R1hNh/clc6ZRGb2PjDlHf4KNdOoXM+5eUZVkdSFsrJeS5h9epuI774wjZu9epuI3JyWndawODgYPP444//aMKF+vqWj+998sknD1500UXlh3v+7rvvzr/lllsKjyFuizS38AYBfzfGzDDGzACeAvR+fcqvnE7vhAgrnIbZY2dbHUcdYkDCALr1+n+U1htWb7wEYyyd8EVZJCfnuYTdu+/sXVfnCAFDXZ0jZPfuO3u3RvFtMGnSpIpdu3aFLl26NHrcuHGDzzzzzAEDBw4c7nK5uOGGG3oMHz58yKBBg4Y++uijP0z79/vf/77roEGDhg4ePHjoTTfd1B3gkksu6bN48eJ4gJtuuql7//79hw0aNGjo3LlzewD86le/6nb//fenAKxevTp81KhRaYMGDRp69tln98/Pz7cDTJw4cfC8efO6jxgxYkifPn2Gr1ixIqqlx9Pc6zI+BSbjvZEGQDiwEjippTtUqjk8nnqczpfZUBrKmB4nMiBhgNWRVBNuOuG33Pr2y1wespPvsx5haP/fWR1JtbLt26/rWVl5+GkBKyoyIo358UxEHk+Nbdeu2/s4nU1PCxgZObwqLa150wLW19fz0UcfxUyZMqUMYNu2bREbNmzYmpaWVvfYY48lxcbGurds2fJ9dXW1TJgwIW369OllmzZtClu+fHnc+vXrt0dHR3tyc3N/1FB0Op325cuXx2dlZW2x2WwUFBT8pCF57bXX9n3iiSeyp02bVnHHHXd0u+eee7otWuTN7HK5ZPPmzd+/8cYbsQ8++GC3c889N7M5x9KguS3eMGNMQ9HF9/iw/xANRGSviGwWkY0iku5b9pCIbPItWyki3VoSWHUORUUfUl+fyzv7q/VOVW2Y3WbntjM+YH2xjf37HqCm5oDVkVSAHVp0j7a8uRqmBRwxYsTQHj161N1+++0FACNHjqxMS0urA/jkk09i3nzzzcS0tLShY8aMGVJcXBy0bdu2sI8//jjm6quvLoiOjvYApKSk/Kg7JjEx0R0aGuqZOXNmn5dffjkuKirqR5M8FBYW2svLy+3Tpk2rAJgzZ07hmjVrfmjZXnbZZcUAJ510UuWBAwdafA6suS3eShEZa4z5DkBExgPVzXztGcaYxvf8fNQY8wffdm4D7gdubG5g1Tk4HAupdIexsyqCi9MutjqOOoK05DQ+7nIX1DzCF9/N4NyT1lodSbWio7VMV6/uNsLbzfxjISGpdePGHf+0gIcuj4iI+KFIGmPk8ccfz77kkkvKGq/z4Ycfxhxp28HBwWzcuPH7Dz74IObtt9+Onz9/fpc1a9Y0u9UaFhZmwDvvr9vtbvEHjOa2eO8A3hKRr0TkK+B14JaW7gzAGNP4DYoE9Kav6kdqax0UFi5j6cFafjHqWkKDQq2OpI7iphP/zGfFPQmrW8eu/S9bHUcFUO/e9+fYbD+eFtBmC/P07t060wIeydlnn106f/785NraWgHYtGlTaFlZme2cc84p+/e//53UMBL60K7m0tJSW1FRkX3mzJmlzz333P7t27f/qAc3MTHRHRMT4244f7tw4cLEE088sYJWcsQWr4hMAPYbY9aJSBpwAzADWAHsacb2DbBSRAzwvDFmgW+7fwZ+iXde3zMOs++5wFyAXr16Ne9oVIfQMCHCMgd8dK4OqmoP7DY7vzz1P/x33RjKM+fRJ3UGQUGHvWpDdSDdu3un/9u378HudXXOkJCQrnW9e9+f07Dcn+68886CvXv3ho4YMWKIMUYSEhLqly9fvvvSSy8t++677yJGjx49JDg42EyePLn06aef/uGDQElJif38888f0FCwH3rooZ+06hcvXrxn3rx5vW+77TZbr169al977bW9rZX7iNMCish3wGRjTJHvFpGvA7cCo4EhxphLj7hxke7GmBwR6QJ8DNxqjFnV6Pl78Z4/fuBI29FpATsPYwxr16axqSCbf+VNYNWsVUd/kWoznv7qJoa65lMTMZXzTlhudZxOT6cFtM7xTAtoN8Y0fGqZCSwwxrzjO0d71GGmxpgc3/c8YAkw8ZBVXgEuOdp2VOdRVraa6upM3t1fo4Oq2qEbT36Kb0uTCav6kH25H1kdR6k26aiFV0QauqPPAj5r9NzRuqkjRSS64TEwBdgiIgMbrXYhsL1lkVVH5nAspM4EsbE8lkuHHrFDRbVBQbYgpp/wAUV1sG7zz/F4dBIzpQ51tML7GvCliLyPdxTzVwAiMgDv+dkjSQG+FpEMYC2wzBizAnhYRLaIyCa8xfj24zkA1XG4XOXk5b3Bp7keLhv+S8KDw62OpI7ByNSfkR/2c5KCSljxnZ6jb6c8Ho/nuC4H6sx8753ncM8fsdVqjPmziHwKpAIrzf9OCNvwnus90muzgFFNLNeuZdUk74QIVSx1wOtnaDdze3b9yS+z4MMPGej+J47i20mNH2t1JNUyW/Lz84cmJyeX2mw2vfKkBTwej+Tn58cCWw63zlGv4zXGrGliWYvu0qFUczgci3DUhhATM4YRKSOsjqOOQ0hQCCePfgPHjnP5JH06V08+gIg2oNoLl8s12+l0vuh0OofT/MtOlZcH2OJyuQ7b3dPcG2go5VeVld9TVraa9w7AHJ3+r0MY0/Mc0nefy0BWsCLjbqaOftTqSKqZxo0blwdcYHWOjko/yag2welcjMcI3xZHMnPYTKvjqFbyy0nvsq86nLr8v5FfnmV1HKXaBG3xKst5PPU4HC+xpkiYlvYLIkMirY6kWkloUDhD0xZSvvdK0tf2J9QGRfV2bPFzmTHhWavjKWUJbfEqyxUVLcflymepw8PccdrN3NHsL/gKj4FwO9gEkkLcRJTN5911N1kdTSlLaItXWc7hWEipKwhXyEjGpI6xOo5qZZ7iBQQdcgv9MDtUFz/HR9snkBIzhG6x/UgMT8Ru02m+VcenhVdZyjshwnKWHXQze9wNVsdRfpAQ7G5yeXyIAed1lDghxwXFdVDuDqHWROC2xSJBSYSEdCUyrCdxEX1JjB5E19g0UqO7ExUSpaOkVbulhVdZKjf3n4CbLwrCeXz4FVbHUX5QVG8nKeSnxbe03kZCt3spq9pLdW0OQfY8EtxFBJlyImwHCLfv865Y5/3ylMD+bNhSDyX1QqUnjHqiMbZ47MHJhIV2Iyq8N/FR/ekSPZjUmAGkRKUQbA8O7AErdRRaeJVljDHkHHyRLaU2ThtwFdGhOptNR2SLn0tN2XzCGvUi17jBHXcDpwz702Ff53ZXU1vnIL9sO/llmZRWZVFRvR+X7SAR9gKiTAmhlBJpyyPItsM7F1qV96syD7a4YVUdlLmCqDERuCQGCUogJLgr4WE9iY3sQ2LUQFJi00iN7kl8WLy2olVAaOFVlikt/Ybaml0sdcCD0/VOVR3VjAnP8u46qCheQEKwu9mjmu32cCLC+9E7vB+9U8477HrGGFyuIiqqsskr/57C8l2UVe2hquYAdnsuca4igigjQpxE2A8Am8CF96a3pZCfAzvroLheqPSEUmci8djjsNmTCQ1NJSq8F3GR/egSk0bXmIF0jeqqtzNVx0ULr7KM07mQGo+NIhnKhG4TrI6j/MhbZP1z+ZCIEBycSHxsIvGxRx6c5/HUUlubS3HlLvLKdlBSuYvy6mxcdgfhQflEeooJoZxIWxHBtt3eF9V4v2oLYacH/lsHZS471Z5w6iUa7AkEB6cQHtqd6Ig+JEQNICUmjdSY3iRFJB3zgLF3192Ep4UfVlT7oIVXWcLlKseZ+zqf5Hq4ZsyN2sWnAsJmCyU8vBfh4b3olnTmYdczxuB2l1FVc4D8su0UlO+ktGoPVTX7sdmdxLoKSTClhEkBETYHNtnqvVFghferxAn766GoHipcIdQSidsWi82eREiob8BYZH+SogfRNWYQqdHdfjRg7N11NxFRNp8w32jwpBA3NWXzeXcdWnw7AC28yhJ5eW+AqeHTvFA+v/gqq+Mo9SMiQlBQLDFRscREDaP/Edb1eOqpr8+npDKL/LIdFFfsorx6H3W1OYQF5RPuLiKYciJs2YTa9npf1DBgrBiyPZBRDyX1Nqo8YdQTxZCIPMIO+escZvd21/ur50AFjhZeZYkcxwtkVwmjes8kLizO6jhKHTObLZjQ0G6khHYjJWHSEdd1uSqorTtIQXkmBeU7KKncQ1V1NtQ6iQkqIN6UEEoJEYfpnT7cpVmqfdHCqwKusvJ7KsvXstQBt5+td6pSnUdQUBRBQYOIjBhE75TzD7ve2yuDmrwEq6hebzDSEegtI1XAOZ2LcBthn2sQJ/U8yeo4SrU5tvi51BxSd2vc3uWq/dPCqwLK46nnwMFFfFNguGLUPB1UpVQTZkx4lqqYeRTU2fEYKKizUxUzTwdWdRDa1awCqrBwGcZdxCd5QSyZ9gur4yjVZvnzEixlLW3xqoDKObiAojqhV+qlJEYkWh1HKaUCTguvCpja2oMUF6/gQ6dhztgbrY6jlFKW0MKrAsbp/CeCYXtNH07tfarVcZRSyhJ+PccrInuBcsANuIwx40XkUWA63kvIdwOzjDEl/syhrGeMITvnOTJK4MLhN+ugKqVUpxWIFu8ZxpjRxpjxvp8/BoYbY0YCmcC9AcigLFZa+jXuun2szLVzzahrrI6jlFKWCXhXszFmpTHG5ftxDdAj0BlU4B04uIAqN8QmXEByZLLVcZRSyjL+LrwGWCki60WkqSu/rwM+bOqFIjJXRNJFJD0/P9+vIZV/uVxl5OW9xWe5cN24m6yOo5RSlvJ34Z1kjBkLTAVuFpEfRtSIyO/xzor5SlMvNMYsMMaMN8aMT07WFlJ7lpf3BjZq2ViZypl9Dz8jjFJKdQZ+LbzGmBzf9zxgCTARQESuBc4HrjLGGH9mUNbbs/9Z9lbCWYNvwSY6kF4p1bn57a+giESKSHTDY2AKsEVEzgXuBi4wxlT5a/+qbais3EZ99UY+dArXjplldRyllLKcPy8nSgGW+C4bCQJeNcasEJFdQCjwse+5NcYYvZtCB5Vz8AVcHgiKnkpqdKrVcZRSynJ+K7zGmCxgVBPLB/hrn6pt8XjqOHBwEasL4eoxt1gdRyml2gQ94ab8prBwKTZTRnp5MlP6T7E6jlJKtQlaeJXf7M5+mvxaGN9vHnabTuCtlFKghVf5SW1tDtVlX/BxLlw3do7VcZRSqs3Qwqv84qBjMSKG6rAz6BGjNydTSqkGfp0kQXVOxhiy9j/L5hK4bNQdVsdRSqk2RVu8qtWVln6F3e3g25I4zht4ntVxlFKqTdHCq1rdruynqHRBWq8bCLJpp4pSSjWmhVe1KperjNLC9/ksD2aN1fuiKKXUobTwqlblzH0Fu7gosp9An7g+VsdRSqk2R/sBVavasffvZFfC9BF3WR1FKaXaJG3xqlZTWbkVe/0OviqKYvrgC6yOo5RSbZIWXtVqdu57inoP9Ei9jmB7sNVxlFKqTdLCq1qFx1NHft6/WV0I14y9zeo4SinVZmnhVa0iv+B9gqnioBlJ/4T+VsdRSqk2SwdXqVaxJetRimrhzCG/tTqKUkq1adriVcettjYHqV7HqsJwLhoyw+o4SinVpmnhVcdtV/Yz2ATik64kNCjU6jhKKdWmaeFVx8UYD/tzFrChBK4c+xur4yilVJunhVcdl+KSLwmjkKz6waQlpVkdRyml2jwdXKWOS8au/6PSBT8bpHeqUkqp5tAWrzpmLlcprorP+KYwhBlDr7Q6jlJKtQtaeNUx273/BYLFTUjsxYQHh1sdRyml2gW/Fl4R2Ssim0Vko4ik+5ZdJiJbRcQjIuP9uX/lX7uynyKrAi4d83uroyilVLsRiBbvGcaY0caYhiK7BZgBrArAvpWfVFRsJtLsZ1tNH0akjLA6jlJKtRsBH1xljPkeQEQCvWvVitIzH6LeAyP63ml1FKWUalf83eI1wEoRWS8ic1vyQhGZKyLpIpKen5/vp3jqWHg8dVSXfMDa4iAuHXG91XGUUqpd8XfhnWSMGQtMBW4WkVOb+0JjzAJjzHhjzPjk5GT/JVQttvfgq4TbanFHTiEyJNLqOEop1a74tfAaY3J83/OAJcBEf+5PBcaWrL+SXwvTRz1odRSllGp3/FZ4RSRSRKIbHgNT8A6sUu1YTc1+otzfk1GZypjUcVbHUUqpdsefLd4U4GsRyQDWAsuMMStE5GIROQCcCCwTkY/8mEG1snWZf8Im0L/nLVZHUUqpdslvo5qNMVnAqCaWL8Hb7azaGWM8lOS/RlaFjetOvNXqOEop1S7pnatUs+XkLSfaXk5lyKlEh0ZbHUcppdolnSRBNdv6zIewu2DycB1UpZRSx0pbvKpZ6utLCK9fR0Z5IhN6TLI6jlJKtVsdtsX77rqb8BQvICHYTVG9HVv8XGZMeNbqWO1W+s5HCLEZUlOv17uOKaXUceiQLd53191ERNl8kkLc2ASSQtxElM3n3XU3WR2t3TroWMieSmHGqHutjqKUUu1ahyy8nuIFhNl/vCzM7l2uWi6v+L8k2vPJt00gLjzO6jhKKdWudciu5oRgd5PLE4PdvPNpN0IiRtEzeTIDu11ARPgA7To9itXf30ekB04e8oDVUZRSqt3rkIW3qN5OUshPi2+VG3KrcunPCkoOrGDdgd9Q5Q6iUnoSGTWOvl3Po1/KVEJDu1qQum3yeGoJrv6CzRUx3Nl7qtVxlFKq3euQhdcWP5easvk/6m6ucUNt7DzmnvkU2/I2senABzgKPsdTu5Uk+x4SZA+5WW+TmwVlrnBqg/oSG/MzBqVOp0fyWQQFdc7rVtN3PUWk3UVc0hXaM6CUUq1AjDFWZziq8ePHm/T09Ba9piWjmmtcNWQ4/su2A++RX/I1trpMuoWU0S3c+7zHQLE7Bk/wQBLjJzGk28WkJJyIzRZyvIfW5r32SW+C3NmceUoeiRE6S5RS7YmIrDfGjLc6h/qxDlt4j1d5bTnrD3xGpuM/lJauIcS9h15hVcT7am29Ryj2JGILHUrXxNMZ2n0G8TEjEOk449WKKzLZsG4wGTXDuXPqZqvjKKVaSAtv26SFtwUKKgtYv38ZWbkfUlmRTqRnP30i6gj3dWlXu22UmFRCI0bSI2kyQ3tcQlREb2tDH4cla2YQX7MEur/K6QOvsDqOUqqFtPC2TVp4j9OB0my+y17C/vyPqa3aSKw46R3uJsjX8C11BVMpPYmIHEffrucypPtFhAQnWBu6GYzxsOSTCPJqbdwwrVLP7yrVDmnhbZs65OCqQOoR24seI24HbgfAGMOuwq1kZL+Do+gLPDXewVtx9iyK977F6r3XU1AfQa29D7ExP2Ng6nQGdD0Huz3c2gM5xMa9L5MQXEtJmA6qUkqp1qQt3gBwe9xszV3D1v1LKCj5CqnLJDWkhETf+WKXBwpcMbhDBpEUN4khPS6mZ+LJiNiPvGE/+tenacSbHUw4YR8p0b0sy6GUOnba4m2btMUbAHabnZGpJzMy9eQfltXU15CR8wmZjg8oKV1DiDuL7rZ0QkvSySp5kq1uodDtHbyVkng6I3teRlLMsIC0PitqnHSRHeys68/5WnSVUqpVaeG1SFhwGCf0OZ8T+pz/w7Ly2jK+y36fLOdyKqrTiTT76RG0CnvBKrYWPEiZy06JpyvBESPpmTyZkT0vJyaiR6tn+2TTvcTZYEjfX7f6tpVSqrPTruY2Lr/iIBuy3yE7fyW1VRnE4qBbmAubr+GbXxdCpfQgImocfbtMZUTPGYSFxB7XPl/5KAaPp5Yrz6nCbrOuu1spdXy0q7lt0hZvG5cc1Y0pQ28Fbv1h2f6SHWTsewtH4We4XVtJtO8huTaLyv1v8U32deTWeQdvxcScwMDU6QzrPg17M2/2seXAB3QPLWev7Xwtukop5QdaeNuhnnGD6Rl3H3Af4B1JvTPvv2w98A75xV9hc+8gxbaN6KptFO1ezCeZ4KyPxRM8gMS4SQzpMYP+yZOw2f53s4+GO30l+iaYCLFHWnFoSinV4WlXcwflcrvY6ljJjpz3KSlbQ4hrN6khlYT6GrFl9UKBOxEJHYrbXUlPWf/Dc+C9t3VVzLzD3mZTKdX2aVdz2+TXFq+I7AXKATfgMsaMF5EE4A2gD7AXuNwYU+zPHJ1RkD2IUT3OY1SP835YVl1XwaYDvsFbdeuIIJuurMLexG9BmB0qihcAWniVUqo1BaKr+QxjTEGjn38LfGqMeVhEfuv7+Z4A5Oj0wkOiOKHfVZzQ76oflpVW5/Ldmq40dZXS4eY1VkopdeysuKP/hcDLvscvAxdZkEH5xIanUFjf9CCqosMsV0opdez8XXgNsFJE1ovIXN+yFGOMw/fYCaT4OYM6Clv8XGoOadzWuL3LlVJKtS5/F95JxpixwFTgZhE5tfGTxjuyq8nRXSIyV0TSRSQ9Pz/fzzE7txkTnqUqZh4FdXY8Bgrq7DqwSiml/CRgo5pF5I9ABTAHON0Y4xCRVOALY8zgI71WRzUrpVTL6ajmtslvLV4RiRSR6IbHwBRgC/ABcI1vtWuA9/2VQSmllGpr/DmqOQVY4rupfxDwqjFmhYisA94UkeuBfcDlfsyglFJKtSl+K7zGmCxgVBPLC4Gz/LVfpZRSqi2z4nIipZRSqtPSwquUUkoFULu4V7OI5OM9H3wskoCCo66lGuj71TL6frWMvl8tdzzvWW9jTHJrhlHHr10U3uMhIuk6nL759P1qGX2/Wkbfr5bT96zj0a5mpZRSKoC08CqllFIB1BkK7wKrA7Qz+n61jL5fLaPvV8vpe9bBdPhzvEoppVRb0hlavEoppVSboYVXKaWUCiAtvMdJfDejVk0TkU71O9bZjlcp1XJ6jreFRGQyMBrvdMKPW52nrRGRs4GxeN+fv1qdJ1BEZIgx5nvfY5sxxmN1JqXA2zgw+oe+TdFP5y0gIlOAp4Fi4AYRud3iSG2Kr+g+CVQBk0TkjkbPddjfNREZB2wWkYUAxhhPRz7e9qShR0pEghots1uXKPAaiq6I3Cgid4nII53tPWhr9I9DMzUqKr8wxiwE/gBEisgMa5O1DSJyHrAIuNAY8w+88yzHicjlDS3ADtwtnwtsBk4XkVdAi29bYYwxIjIdmC8iz4tIiDHG3dkKj4jcDMwEVgJXAr+1NlHnpn8Ymm8UEGeMWSciccDDQHfgXhFZLCLR1sazXBe870e+7+d7gG7ANcAqEQnuqN1dxpgDwOPAI0CFiCzxLdfiazERGQX8EfgQiADWi0hoZyq+vt/BwcD5eKdk3Qg8IiLhlgbrxPSPwlGIyGgRiTbGPAa8JCJbgc+Ah4wxNwMn4f2lvsXKnFYRkUkiEmmMeQnve3BARHYC9xlj5hpjpgHlwINW5mxtInKWiNwnIiG+P2y7gZ8DfwXyROQd0OJrJREZDtwKvG+MedcY8wtgPbCmofham9A/muhZsgHJwBvAROAyY4wLuE5ELgl0PqWFtznuBd4UkShjzO/wdqcmAg1/WOuBt4EO+Z+4Ga4DdopIhDHmWeBmoDfeT9UNPgFKrQjnD77zhX/H+2HiAby/Izvx/h5c6vu5SkQ+Am/xtSiqghhgmIgMBTDGXIv33ypDfKwM19oaD6QSkXNEZCzenqi/423tLjTG1IjI1cBNwAbr0nZeWniPwhgzE+9gqn/6Wr6PA68CX4hIjIhcDMwC/mNlTqsYY64DlgMbfS3ffwJ3AmtFZICInAn8AvjAypytRUROxduyPQfYBHQFduE9vjTgNGNMEd7W/z4R6WFV1s6m0UCqUb4u5gpgDlANXCgiaQDGmMuBmcbHssB+0Kjo3oZ3HMqpwArAgfe9eE5EFuP9/bzcGJNlVdbOTC8naoKInAVEAvuMMRm+ZS/i7a652hhTLiJ/wtu62wdc0XApSWfgu6QqGdhijNnsW/YUMA0YaYypFJF5wDNAFjC9o7w/vpHttxpjpovIGOAt4G7ge+BM4ATgQWPMLr2MI3AaDeA7F3gM+Bo4DfgzsAx4FDgAvGmM2daR/m18HzikoWdFRPoC840x5/r+To0EZhhjXCLSHV/vnDHGaVnoTk4L7yF8g6T+A5wC7MHbum34/m+8Xaa/MsaU+UYKftZRikpziEgs3mJzFt6RvBvxDlx5D/gb3nPeJ/q6s2biLc5brcrb2nx/1J4FnjDGrBSRE/H+XvzZGLNIRIJ8589UAIhIIlDqKyrxeP/vPmCM+dTXvfw5cD3eD0YPAH/saK08EQk3xlT7Hp8OePCe8nAAJ+M9p1vt+//4lTHmoGVhFaCFt0kicjIwHUgCtgF9gV7AXuAGvJ+mLzTGVFqV0Uq+/9yX4G1BJACVwEXAYv5/e+cdZldZrfHfSwghUkSQLiIQyhVQQIo8CChCIKEktGDoESMSQpEEAtwLUjQgCEgxNBELJAqEEpQWC1U6CoIiTQHpVQg9yXv/WN/Bw0ixZM6e2Xv9noeHmcOeZM3h7L2+1d4VtaQHgE+3Hga9nVLfn9b2/SjiYb657cclrQH8GPh+iqp0DklzEXXK82w/XF77CXCI7b+W77cBPm97tKR5bb9UmcHdgKRNiSzc8PK7foW4Ny8gsk+Ll+t2JfoxtrL9bFX2JkHWeNtojRfYvoFoCHqMGCHai5h7uxy4DhhApFobRdv7czUwlWhceRoYD+xKRMCXAH2JUaJej6QBwBGlGaXFqcAthIIZtm8l6vzbl1GzpDO8QTQ7Tpd0cEm5vgSc3uW6Bcpnd1rXP6A3I2lj4JvAMZLWBbYimqdeAcYS3dsXS/oG0d29ZzrdnkFGvICkBW0/U77u0xozKJHdUCK9fKLt5yXNC8xoUrTb5f15Ww6xnLY3JjIBk2w/IWlOoK/tlyszeBYhaRCwM3Az0YwyBbjd9rmSxhAp9W3aru9n+41qrG0WXT6HGwE7ADfbPrWMQmnn1QAAFLJJREFUcn0EuIZIuY6zfVl11s56Sq/BT4GTbX+jNHkeDlwIjLf9ZrkXRxEp59tt31edxUk7jXe8pev0MOAy2xeW17o630FAP+Bw2y9UZGolSFqaOEl/r62O1P7+DCbqvS8AZ9t+rDJjZyGSNiOiiUNtT5G0ELAjMQc5PzFK9EPgINvnl5+pTcNOT6b1PpeO+YUJB7QJMAS4y/YESUMJwYzHbF9Tp/835cD7LaJ+3Re42vYFpbHsQOAk4nn2eoVmJu/D7B98Se2ZCdwJbCRpuu0pLVUb2zNsXy2pL9GW36j3q6RNHwbOAJaVtLTti7u8P5dJmoNwSHWp6S4CjAG+avsWxYxySxTjFCL63RFYmvjcXFjei1o82Hs6xekOJQ7MB5XvpwJvAttIOhA4zjFj//bPVGPtrKXca6sQaeMbJO1OjErNtH2hpH5EWnmO8rl8s1KDk3elUY6kndYJuDTHTCZu2i3K65d0cS5TJV1fl2ahfwVJHydSVxfY/kVJZS0o6XXbV3R5fy6WNLVG6fc3gLeA10q67oBSQ+sH3A3sC4iYX77fNVVA6qmUzvqvEg2QT0haC1jL9knlkDyUaIZ8sEIzu4WSQj6mdaiwfbqkmcBQSRTnOwdRIvk58VxLehiNTDW3p53a5v/mIoQeVgcutX1J+e99mvZglTQ3ccPuBSxHdOzeQdSLPgZMbdXM6vj+lCad/YCBwIpEo931RPPYaKKLtpGCKVXRll7+MNEkdTXwZ2AOoqt+Y+Bi2/tJ+mgTmoi61LlHEqN8V9me1LUTP+lZNK6ruYvT3QM4o3T9LQWcBdwGDJa0LUDdnMoHUSLd84CVgJOJCG834DOEIMYTwAYl1VfL96d8Pk4nUpn7Abvb/r7tm4kDSXYud5A2p7sJ/9D8HgG8CJxpew9gS6J7uV8TnC68Uwfc9pnE4Xh9hcJeOt0eTOMcb5vT3QsYRjjbdYja3ZrAmcC9wLol8msarxIR3qHAJwkH9DtiBnB1onHjZWDNkiWoJban2b7R9nmtTuVyGPs08NtqrWsWbY1UJwI/K+WNB2yPtX1dKYP8BJhct67ykn1pfd2n9X3baN/b6zYd6zgPrMNEQd1pTKpZ0orEQPkkhcLNvsQmmZFEmuoyYvD8QOBWYB7bL1Zlb6fpkgnoD3yN6Fb+P0JEZHdCem4i4Xjmtv1cReZ2FEmLErtMRxIav3dXbFJjKA5GxIHvTuJgPBzYEHikjNJcCPywdJ/XqXu5/Z7chyh7vEps/prWZbqgNr93E2hExFvSMQsBUyUtVUaCjgaWIUYQtiQkD/sTwuJ9m+R04R2ZgGHAArZPAH5FjNS0It/7icPJbE1xuoUXid99SDrdztAW6c10SHBOAfYhVnKuSIzSrFayUjvUzenCO+7JDYjFHFOIZ9QVpYY7Q7EpqzZd202hMREvvB25HEsMk58gaVki6t0K2B74LLFn9+kKzawUSccBg4ENHIIYXyfE5o8kmovmcsNmmZPO0mVOdzCRgforMSs+3fZDklYmVKuGuMbaw5KGEB3Kk21PLK9NIHowNs20cu+k1hGv/nkB+fOEhulykva0fT8xx3shEdmd0VSnW+b/sD2GWIJwhaTFSuR7C7A/cVBLp5t0K8XpDiSa+24muuu3Jka3HlJIJZ5HHJJr5XTba7qFxwh52rUVqnnYHkUcRC54l+uTXkAjIl5JuxEqNn+zfZFC+WUr4De2zymCCTMb7HQ3J9bZndBKIUs6moh0tyqR7/yOPbNJ0i20Rbr9iW7yiwg98AnEQorHFEpzqwBvlPn62qSXu9R01yeaGJ8gAqQfE6nms10WPUhaxLnar1dSS8fb5QO8AaG8dDLhbK+0PV6hw7sr8MvSit8Yuj6sJC1PjApdDUxwaFIvQyw8eAoYWMexoaTnodBd7g8sT0S5ArZ0CN1sRoiYXOIar15ULLHfCbiLWMjyXSLdfhZR456QKebeTe1SzV2c7qcJsfSdbJ9IdOZuKulA25cTH+SfV2dt5+ny/mykUGSaDmxL1Lj3LBmANYl03q7pdJNOUO7XLxNR3hSiqe2nxemuRSy4/3vdnK6kxVpNUpKWJBY+bE6oc40B/pfY9nUgIZLRWMXBulCriLeLUxlFKC31I1JWx9p+RtIKwGTgB27w7tS2U/UdhDrVJOBnROQrYrZ5kO0/VWZk0ghKnXIB4FFCNW5YUajaBtgAWByYh1hiXyvFsCJYMww4jRgVmo848A5ykYWUtD+RWj9J0pzO5Qe9nlqdnLqMxKxGnA43Kf8MlHSF7XvLwH1jNUxLRDsM2KLUbwcQNaRHiEXa8wB9sn6UdCdtB+X5bD+r2Hk8SdIg25crltpPJHSXX7X9aJ1quoXHiVLYMsBqts+W9CTRBDqkXNOfOHxAg59bdaJWjhegqCntDixSmhDOU4iGbwT0V2zsaNReyvd4WE0nFgFg+4HykPuU7SuAPFEn3UpbI9VawJGSTrI9WdIOxD27bfksQmgyA/WZV239/iVt/pKkdYBVJL1ApJjPkPRbQkVuS6IUhIs2c9K76fU13vaRIUlzODbk7AI8L+kMANvnANcRkoe1qg99EF3S7ysohNWfBH4PTFZsc4HoHh3wLiNYSTLLKU53EHAQsehggqSBjt3GOwOXKbSZa0e5B1v35HaloewMQid+Q2Cw7R2Jpqp7gK1t31uZwckspzY1XklfJtLLzxNzqC8SncyPl7k3JM3basVvAl2c7mhCJvNm4Eqizv1N4kb/BbAZsE3e4EknkLQ40UC1p+2byufzS8BRjjWU2wIv2b6yUkO7EUljic7tr9i+pxyCW0I+NxC61G+935+R9E5qEd2U9NR+xPjLEkTT0MrAnsCqko4vlzaqBb/N6Q4h3o8vAlcRc5AjbH+d6Jq8FhiaTjfpIE8RKeSWwP8pwE3A6ZLWsn2+7SvrJBAhaWlJS5SvBwCb2V4beLiIguxu+0fEUpKVidpuUkN6fcSrEFEfD9xl+1zFAoQRwADboxQykX1tP1KpoR2kS6S7JKG5fIvt7cv7tSXwOaKx49ScCUy6m7aa7oLE/fi4pBOIhr5Jtp+U9FkivdofWLdO2SlJ8xCH3BOA14j+mquIA8hzRBC0DnAucBQNW9LSNHpdc5Xalj9D7IOV9DdgiKRrS+fjmcAvJH28SQ4X/snp7kWMJ+wLnCZphO2zCam5foTYfN/3/tOSZNZQnO4WhPTo3yXdSYzNHAAsI2k6MTo0jHBQiwC1cLzlnnxZ0lHEPbc1sYZ0F6KR6ke2/1imLVYtc/PpdGtMr0o1lw/wzPL1ZpJGSlqA0Fp+CNhFocK0PpHCalwk12WkanXgJ7Z/Tqh07SNpRLnuXKKeljKQSbcjaVVgLDCU6NQdavtGYDQwlYj8hgMLE9mYOjndVlqxP/FM+gRxP06zPa443dHAYcRhJKk5vSLV3HUcRqG9PIaohXyGSJ3OT+zVXR94AzjA9u8rMLcS2lJ5IkRDJgMr2V6y7ZoNgLOB/y2d3knSLUianzj8vmb7VUmrE70F04lxvx0cCw9WaPUWSFqbWGg/1DVbvyhpJLCy7b0Vm5X2I2rclxDR7UnAEbb/UKGZSYfoLY53UdtPlK/XIZzuqFIXOgQYCOxh+25JHwXerFN96IPokl6e1/ZLpZY2EXiqjCa0rl0feNT2QxWZm9QchTrcJOBeYFFgO0K6dQIwJ7CtY+HBQCLK26bUfPsS8/ePVmN59yBpZ8LRbm37wfLap4C9iT6LM4n7NMUxGkKPTzUrVJbGSZpboWc6kBAOHwxg+0jgcqJuubLtZ5vkdOEd6eU9gBMlHUbIQO4IzC3prLZrr0mnm3QXkpYDzgFOBL5OzIsfVKLa3xAR72qSdgWOB8YXpzub7bfq5HQVzE6MOX7D9oOS+pWD8l1Eo9UCRFYgnW6D6PERr0KJCqK9/mNEPXcssaPy145lB0gaA1xo+y+VGFoxZaTqa4Sz/RVwju3DSuT7U+DPrXnmJOkOipP5MTC37S3KaysSMqQH2H5L0k6EPOLiwPm2r+paSurNvNvvIulg4nl1mO2/l9eGEBuHnnbNlj4kH0yPjXgVIukUJaqFifnc3Qnd5RMJoYwNygcY28c1zem2ZhxLh/JSRIfoekSj2bfKZa8TnaLjq7AxaQ7FgRwJ9JU0rry8CbAHoUR1G/HMudz2SNtXlZ+rndOVtImk4ZKWAu4k6t2bSlq4iIPsT+ihp9NtID1unKg4kz7ARpI+AUwjnO7JwExCFEPAccAhwBqSflkcdCNQSO0NAmZIOsax6OCvwKnAs7Y3LNeNJerdJ1VnbdIUSrr4T5L2A06StAawJJGteoZYd7cSUfutHV3G+IYT3dqHExKYvwfWBnYjnm+j6pRWT/49elyqWdInS3v9QkTK9KPAmmU+9yPErN/ORErrUmLQ/LnqLO4sCl3X7xBdkJ8CXrc9TtIngaOBywgpvnWJ/Z072P5jVfYmzaI1Zy9pWaKGeY/tcW3/vZ/tN6qzsHtomypYnpBiHQaMBIbb/kK5ZnZgIWLTUs7pNpge43hLpCvgPuB627tKmkSMxtwDHG57uqQ5iXVZQ4Cv2p5WmdEdpowDXUIM2T9QZnU3A24ELiYii42AVYn1YeNyPCHpBJL6FOGHdue7AuF8bwJOruPMeAkG3iTq2k9J6g+MIxpAFyQWHswoI5BTbD9ToblJD6FH1XiLOMYqwJqSjrY9HNiLGDj/drlsRSJtNbJJTrfwLPAh4qYGOJgYyF8OuAZ40Pb+xALxYel0k+6krcdgPWDLciimON3ZSifzWKLvYL7qLO0eJA0GfkgchieXJqq3CE2BpYnAYIak7YF9SO3lpNBjIt52JM0N3AH8yvYeJY16CFHr7Qds1dT6SKmbXQXMIOpE55XXjyWa0Ea0Io8k6S7aotqBwGnAzravf49r5qpbD0b5vY8nZnGfBuYisk4nEeNUxxOazH2IuvYOdRMFSf5zelRzlaTtgGVsj1dIzP1O0gTHsoOvEDJrU5vqdAFs31oijGt5p87yw8Re01yUnXQbkha0/UxxqPMQHcujbV/f5mhbS95bn8VXKzR5llNKPpOBVcpsbt8yKvU5ouzzOKHDvDJxGD7IDdOMT96fSlPN+uel638BdpS0bzkhrwqsL2mi7Vdsf8/2fZ23tGdRUsgDgVMUi7S/QBxKzqzLaEbS8yhja6MlLQ3g2Gr1FP847PUp/x5QnDLlurp9Jlsln9XK99MlzVFUqb5EzNLPbvtm21PS6SZdqdTx+h8LD1aUtIDtW4gP7S6SDijO97PEjbxoq6aURORLNFJNIgQydspUVtLNzACOBV6TdFx57Wlgm+J43iqZqhOAD1dlZHdTVKfWAs6QtEc5WExXrNx8lUgx1yq1nsxaKkk1S9oQWNv2kZIWIxowbpM0yfYdCkHxX0r6UFFfWquGp+b/Gtu3SVoJmGH7z1Xbk9ST0qk7p+0XFFroiwEflzTO9qGSJgKTJL1ALC05wvbfqrS5uyn33kbA1JJanwBv61S/QpR9XqvSxqTn0tHmqhKxzk4ouSwBfMf24ZKGAl8E/gBMtv2cpO8RN/HGLZm1JEk6S7ln1yQ65e8j5sMPITSGxwB32f62YrPQQsDjpQ+hNjKQ74di69JUYq/uC8SM/Y6ZfUrej0q6miVtSgyYAzxje6ykzYm65UvAc0SKeWzWR5KkWhR66T8g7s/Rts8to0MrEV29T9o+oEobq6Q431uItPsXbP+pYpOSHk7HarySFm779mFgXqIzcGaRPbyUqFdOJ/Rdj0inmyTV0dZT8RpwF7GgZLCk5W2/bvs2YtXf/JL+pyo7q6a8DysB66fTTf4VOhLxlpruJOLUfDLwGKFlugUx97YbsaXjoHJ9f9tZH0mSimiTQBxK6ILvTezS3ZcYk9mNmF1dD7giJRCT5F+nUxHv00T7/W7ApsTi5xeImu7DRF1kGUlHl+tf75BdSZK8C8XpDgK+Qazve4MoA50O3E7UNa8FHkunmyT/Hh2r8ZZU1NVEaupyQtllReBA26cr9nY+b/uJjhiUJMk7kLQosJ3t75bvv0XoLP+OGJ8ZTiwnuY5ouJpm+4aKzE2SXkunu5o/Q2wcGmH7IknrEGvrbu2YEUmSvCtls85swIuOVZN7AxsS2sOXAouWr0c1UCc9SWYZHZ3jtX170Ti9StL8ts/q5N+fJMn7ch8hQ3q2pMdt7y/peuBl2/dLGgCcS8ggpuNNkv+QjitXFXWqLwJnShrR6b8/SZL35BO23yQ2gS0s6SjbdxSnuyWxBGB8kUZMkuQ/pBLJSNu3E+IYv63i70+SJGhb7bcscHlRo7qLcL6LSzqmXNof2Nv2JSndmiT/HT1yLWCSJJ1D0hBgJ0LqcDXgZ7a/WdZxHgHcZ/vgKm1MkjqRjjdJGoyk+YjRoP2AG4gZ3VOBi2wfK2lloK/tOyo0M0lqRY/ax5skSceZQay5e6js0r2bWOQ+RtIrLfH/JElmHZWuBUySpLO01XQXk9Sv7NS9CZhcFONmAI8AlwGDynx9kiSzkIx4k6RBFEWqTQhFqvvLDtmDAQN3SDqLkIfcCdgByEaqJJnFpONNkgZRupe/C4wEngKGAhOJxSStOd5BwDzA6oRMZJIks5B0vElSc7rsxp0B3Gj7Okmz2T5G0pLAFrbPLdevQTjnEbkhLElmPVnjTZKaIunD8HZ6uZUyfhlYT9KetmeW154DFmn70aeBobbv7Jy1SdIcMuJNkhoiqR9Rsz3F9gnF+fa1/YykbYFJZSnCbcR6zn1bP2v74YrMTpJGkHO8SVJTJK0NXAIcavu08lpf229JWgI4ktiNfZPtSys0NUkaRUa8SVJTbN8oaTAwVRLF+bbSy30J4YyJrVS08xSeJB0ha7xJUmNs3wZsBBwlaZTtGZI+D9wCPNVytul0k6RzZKo5SRqApNUJUYyLgPWBg21fWK1VSdJM0vEmSUMoY0K/Br5s+/xWp3NGu0nSWdLxJkmDkDS37WlZ002S6sgab5I0i1eqNiBJmk5GvEmSJEnSQTLiTZIkSZIOko43SZIkSTpIOt4kSZIk6SDpeJMkSZKkg6TjTZIkSZIO8v/T9u4Qu9G5cAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}