Manual_feature_extraction_v2.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Manual_feature_extraction_v2",
      "provenance": [],
      "collapsed_sections": [],
      "machine_shape": "hm"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5iZ8_159Fcwr",
        "outputId": "c5ba8ed2-4c11-497b-ac5f-d9c30336d81f"
      },
      "source": [
        "!pip install PyWavelets\n",
        "!pip install pyfeats"
      ],
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: PyWavelets in /usr/local/lib/python3.7/dist-packages (1.2.0)\n",
            "Requirement already satisfied: numpy>=1.17.3 in /usr/local/lib/python3.7/dist-packages (from PyWavelets) (1.19.5)\n",
            "Collecting pyfeats\n",
            "  Downloading pyfeats-0.0.11.tar.gz (32 kB)\n",
            "Requirement already satisfied: opencv-python in /usr/local/lib/python3.7/dist-packages (from pyfeats) (4.1.2.30)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (from pyfeats) (1.19.5)\n",
            "Requirement already satisfied: scikit-image in /usr/local/lib/python3.7/dist-packages (from pyfeats) (0.18.3)\n",
            "Requirement already satisfied: matplotlib in /usr/local/lib/python3.7/dist-packages (from pyfeats) (3.2.2)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.7/dist-packages (from pyfeats) (1.4.1)\n",
            "Requirement already satisfied: PyWavelets in /usr/local/lib/python3.7/dist-packages (from pyfeats) (1.2.0)\n",
            "Collecting mahotas\n",
            "  Downloading mahotas-1.4.12-cp37-cp37m-manylinux_2_12_x86_64.manylinux2010_x86_64.whl (5.7 MB)\n",
            "\u001b[K     |████████████████████████████████| 5.7 MB 8.5 MB/s \n",
            "\u001b[?25hRequirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->pyfeats) (3.0.6)\n",
            "Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->pyfeats) (2.8.2)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib->pyfeats) (0.11.0)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib->pyfeats) (1.3.2)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.1->matplotlib->pyfeats) (1.15.0)\n",
            "Requirement already satisfied: pillow!=7.1.0,!=7.1.1,>=4.3.0 in /usr/local/lib/python3.7/dist-packages (from scikit-image->pyfeats) (7.1.2)\n",
            "Requirement already satisfied: imageio>=2.3.0 in /usr/local/lib/python3.7/dist-packages (from scikit-image->pyfeats) (2.4.1)\n",
            "Requirement already satisfied: networkx>=2.0 in /usr/local/lib/python3.7/dist-packages (from scikit-image->pyfeats) (2.6.3)\n",
            "Requirement already satisfied: tifffile>=2019.7.26 in /usr/local/lib/python3.7/dist-packages (from scikit-image->pyfeats) (2021.11.2)\n",
            "Building wheels for collected packages: pyfeats\n",
            "  Building wheel for pyfeats (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for pyfeats: filename=pyfeats-0.0.11-py3-none-any.whl size=46825 sha256=48af86b763ccf229d247caaa5a6f7448fc7817f94f5e6e0c23db43941d5ec57a\n",
            "  Stored in directory: /root/.cache/pip/wheels/a2/d6/20/ffb98b47157e97a3fe0db4ea16c258986aa0ee386ecc7dbf80\n",
            "Successfully built pyfeats\n",
            "Installing collected packages: mahotas, pyfeats\n",
            "Successfully installed mahotas-1.4.12 pyfeats-0.0.11\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3p1JOXT2Hnne"
      },
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "from skimage.io import imread\n",
        "from skimage.transform import resize\n",
        "from skimage.feature import hog\n",
        "from skimage import exposure\n",
        "import matplotlib.pyplot as plt\n",
        "from skimage.color import rgb2gray\n",
        "from PIL import Image, ImageFilter"
      ],
      "execution_count": 61,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wq7n87wFGdk1",
        "outputId": "0fdf2482-576c-4d83-cdd3-055590cde2be"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "USomhzxmGfAt"
      },
      "source": [
        "!cp \"/content/drive/MyDrive/CS5100 Project/resized_images_v4_224.zip\" -r \"/content/resized_images_v4_224.zip\"\n",
        "!unzip /content/resized_images_v4_224.zip -d /content/resized_images_v4_224"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "27Nu3pEUGiBt"
      },
      "source": [
        "import pickle\n",
        "\n",
        "def write_dataframe(df, extract_path, first = False):\n",
        "  mode = 'a'\n",
        "  header = False\n",
        "  if first:\n",
        "    mode = 'w'\n",
        "    header = True\n",
        "  \n",
        "  df.to_csv(extract_path, mode = mode, header = header, index = False)  \n",
        "  return\n",
        "\n",
        "\n",
        "def read_pickle_file(filepath):\n",
        "  with open(filepath, \"rb\") as f:\n",
        "    return pickle.load(f)\n",
        "  \n",
        "def write_pickle_file(df, extract_path, iteration):\n",
        "\n",
        "  with open(extract_path,'wb') as f:\n",
        "    pickle.dump([df,iteration], f)\n",
        "    first = False\n",
        "    if iteration == 500:\n",
        "      first = True\n",
        "    \n",
        "  extract_path2 = extract_path.replace(\".pkl\",\".csv\")\n",
        "  write_dataframe(df, extract_path2, first)\n",
        "  return"
      ],
      "execution_count": 64,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vo9-WmNEGigh"
      },
      "source": [
        "# add this from GIT\n",
        "filepath = \"/content/class_dict_wo_NF.pkl\"\n",
        "class_dict = read_pickle_file(filepath)"
      ],
      "execution_count": 96,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gwaG8bFUGj9m"
      },
      "source": [
        "#add this from GIT\n",
        "df = pd.read_csv(\"/content/final_df_all_images_wo_NF_labels.csv\")"
      ],
      "execution_count": 97,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 330
        },
        "id": "ihl4xHMAGogg",
        "outputId": "13d160ea-cfed-4aac-b977-25eb841ff32e"
      },
      "source": [
        "df.tail()"
      ],
      "execution_count": 98,
      "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>Image Index</th>\n",
              "      <th>Finding Labels</th>\n",
              "      <th>Follow-up #</th>\n",
              "      <th>Patient ID</th>\n",
              "      <th>Patient Age</th>\n",
              "      <th>Patient Gender</th>\n",
              "      <th>View Position</th>\n",
              "      <th>OriginalImage[Width</th>\n",
              "      <th>Height]</th>\n",
              "      <th>OriginalImagePixelSpacing[x</th>\n",
              "      <th>y]</th>\n",
              "      <th>Unnamed: 11</th>\n",
              "      <th>Atelectasis</th>\n",
              "      <th>Cardiomegaly</th>\n",
              "      <th>Consolidation</th>\n",
              "      <th>Edema</th>\n",
              "      <th>Effusion</th>\n",
              "      <th>Emphysema</th>\n",
              "      <th>Fibrosis</th>\n",
              "      <th>Hernia</th>\n",
              "      <th>Infiltration</th>\n",
              "      <th>Mass</th>\n",
              "      <th>Nodule</th>\n",
              "      <th>Pleural_Thickening</th>\n",
              "      <th>Pneumonia</th>\n",
              "      <th>Pneumothorax</th>\n",
              "      <th>disease_vec</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>112115</th>\n",
              "      <td>00030801_001.png</td>\n",
              "      <td>['Mass', 'Pneumonia']</td>\n",
              "      <td>1</td>\n",
              "      <td>30801</td>\n",
              "      <td>39</td>\n",
              "      <td>M</td>\n",
              "      <td>PA</td>\n",
              "      <td>2048</td>\n",
              "      <td>2500</td>\n",
              "      <td>0.168</td>\n",
              "      <td>0.168</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>[0 0 0 0 0 0 0 0 0 1 0 0 1 0]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>112116</th>\n",
              "      <td>00030802_000.png</td>\n",
              "      <td>['No Finding']</td>\n",
              "      <td>0</td>\n",
              "      <td>30802</td>\n",
              "      <td>29</td>\n",
              "      <td>M</td>\n",
              "      <td>PA</td>\n",
              "      <td>2048</td>\n",
              "      <td>2500</td>\n",
              "      <td>0.168</td>\n",
              "      <td>0.168</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>[0 0 0 0 0 0 0 0 0 0 0 0 0 0]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>112117</th>\n",
              "      <td>00030803_000.png</td>\n",
              "      <td>['No Finding']</td>\n",
              "      <td>0</td>\n",
              "      <td>30803</td>\n",
              "      <td>42</td>\n",
              "      <td>F</td>\n",
              "      <td>PA</td>\n",
              "      <td>2048</td>\n",
              "      <td>2500</td>\n",
              "      <td>0.168</td>\n",
              "      <td>0.168</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>[0 0 0 0 0 0 0 0 0 0 0 0 0 0]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>112118</th>\n",
              "      <td>00030804_000.png</td>\n",
              "      <td>['No Finding']</td>\n",
              "      <td>0</td>\n",
              "      <td>30804</td>\n",
              "      <td>30</td>\n",
              "      <td>F</td>\n",
              "      <td>PA</td>\n",
              "      <td>2048</td>\n",
              "      <td>2500</td>\n",
              "      <td>0.168</td>\n",
              "      <td>0.168</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>[0 0 0 0 0 0 0 0 0 0 0 0 0 0]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>112119</th>\n",
              "      <td>00030805_000.png</td>\n",
              "      <td>['No Finding']</td>\n",
              "      <td>0</td>\n",
              "      <td>30805</td>\n",
              "      <td>27</td>\n",
              "      <td>M</td>\n",
              "      <td>PA</td>\n",
              "      <td>2048</td>\n",
              "      <td>2500</td>\n",
              "      <td>0.171</td>\n",
              "      <td>0.171</td>\n",
              "      <td>NaN</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>[0 0 0 0 0 0 0 0 0 0 0 0 0 0]</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "             Image Index  ...                    disease_vec\n",
              "112115  00030801_001.png  ...  [0 0 0 0 0 0 0 0 0 1 0 0 1 0]\n",
              "112116  00030802_000.png  ...  [0 0 0 0 0 0 0 0 0 0 0 0 0 0]\n",
              "112117  00030803_000.png  ...  [0 0 0 0 0 0 0 0 0 0 0 0 0 0]\n",
              "112118  00030804_000.png  ...  [0 0 0 0 0 0 0 0 0 0 0 0 0 0]\n",
              "112119  00030805_000.png  ...  [0 0 0 0 0 0 0 0 0 0 0 0 0 0]\n",
              "\n",
              "[5 rows x 27 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 98
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iid3_9NCGqu6"
      },
      "source": [
        "df[\"image_path\"] = \"/content/resized_images_v4_224/resized_images_v4_224/\" + df[\"Image Index\"]"
      ],
      "execution_count": 99,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vwz9E2bLGsaU"
      },
      "source": [
        "X, y = df[\"image_path\"], df[\"disease_vec\"]"
      ],
      "execution_count": 100,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uoqlXE45GvGy"
      },
      "source": [
        "# HOG - Histogram of Oriented Gradients\n",
        "from skimage.io import imread\n",
        "from skimage.transform import resize\n",
        "from skimage.feature import hog\n",
        "from skimage import exposure\n",
        "import matplotlib.pyplot as plt\n",
        "from skimage.color import rgb2gray\n",
        "\n",
        "def hog_features(path):\n",
        "  # reading the image\n",
        "  img = imread(path)\n",
        "  img = rgb2gray(img)\n",
        "  resized_img = resize(img, (224,224))\n",
        "  exposure_balanced_image = exposure.rescale_intensity(resized_img, in_range=(0, 225), out_range=(0, 225))\n",
        "  fd = hog(exposure_balanced_image, orientations=4, pixels_per_cell=(8, 8),\n",
        "                \tcells_per_block=(1, 1), visualize=False, multichannel=False)\n",
        "  return (fd , exposure_balanced_image)"
      ],
      "execution_count": 73,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PwW0hf57J7xr"
      },
      "source": [
        "# LBP- Local Binary Pattern\n",
        "from skimage.feature import local_binary_pattern\n",
        "from skimage import data\n",
        "\n",
        "def LBP(exposure_balanced_image, radius = 3):\n",
        "\n",
        "  n_points = 8 * radius\n",
        "  lbp = local_binary_pattern(exposure_balanced_image, n_points, radius, method = 'uniform')\n",
        "  n_bins = int(lbp.max() + 1)\n",
        "  histogram_2 = np.histogram(lbp.ravel(), density = True, bins = n_bins, range=(0, n_bins))\n",
        "\n",
        "  return histogram_2[0]"
      ],
      "execution_count": 74,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DqdryOPSNvss"
      },
      "source": [
        "# Gray level co-occurance matrix\n",
        "\n",
        "import numpy as np\n",
        "from skimage.feature import greycomatrix, greycoprops\n",
        "from skimage import io, color, img_as_ubyte\n",
        "\n",
        "def GLCM(image):\n",
        "  image = img_as_ubyte(image)\n",
        "\n",
        "  bins = np.array([0, 16, 32, 48, 64, 80, 96, 112, 128, 144, 160, 176, 192, 208, 224, 240, 255]) #16-bit\n",
        "  inds = np.digitize(image, bins)\n",
        "\n",
        "  max_value = inds.max()+1\n",
        "  matrix_coocurrence = greycomatrix(inds, [1], [0, np.pi/4, np.pi/2, 3*np.pi/4], levels=max_value, normed=False, symmetric=False)\n",
        "  # plt.imshow(matrix_coocurrence[0], cmap = 'gray')\n",
        "  \n",
        "\n",
        "  # GLCM properties\n",
        "  def contrast_feature(matrix_coocurrence):\n",
        "      contrast = greycoprops(matrix_coocurrence, 'contrast')\n",
        "      return contrast\n",
        "\n",
        "  def dissimilarity_feature(matrix_coocurrence):\n",
        "      dissimilarity = greycoprops(matrix_coocurrence, 'dissimilarity')    \n",
        "      return dissimilarity\n",
        "\n",
        "  def homogeneity_feature(matrix_coocurrence):\n",
        "      homogeneity = greycoprops(matrix_coocurrence, 'homogeneity')\n",
        "      return homogeneity\n",
        "\n",
        "  def energy_feature(matrix_coocurrence):\n",
        "      energy = greycoprops(matrix_coocurrence, 'energy')\n",
        "      return energy\n",
        "\n",
        "  def correlation_feature(matrix_coocurrence):\n",
        "      correlation = greycoprops(matrix_coocurrence, 'correlation')\n",
        "      \n",
        "      return correlation\n",
        "\n",
        "  def entropy_feature(matrix_coocurrence):\n",
        "      entropy = greycoprops(matrix_coocurrence, 'entropy')\n",
        "      return entropy\n",
        "\n",
        "  # print(contrast_feature(matrix_coocurrence))\n",
        "  # print(dissimilarity_feature(matrix_coocurrence))\n",
        "  # print(homogeneity_feature(matrix_coocurrence))\n",
        "  # print(energy_feature(matrix_coocurrence))\n",
        "  # print(correlation_feature(matrix_coocurrence))\n",
        "  return np.concatenate((contrast_feature(matrix_coocurrence),dissimilarity_feature(matrix_coocurrence),homogeneity_feature(matrix_coocurrence),energy_feature(matrix_coocurrence),correlation_feature(matrix_coocurrence)), axis = 1)\n"
      ],
      "execution_count": 75,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7nTdaj8WO4Kd"
      },
      "source": [
        "import pywt\n",
        "import pywt.data\n",
        "def wavelet_transform(image):\n",
        "\n",
        "  coef = pywt.wavedec2(image, 'db1' , level=4) # level3 decomposition #use db1 or haar\n",
        "\n",
        "  cA = coef[0]\n",
        "  (cH1,cV1,cD1) =  coef[-1]#level1\n",
        "  (cH2,cV2,cD2) =  coef[-2]#level2\n",
        "  (cH3,cV3,cD3) =  coef[-3]#level3\n",
        "  (cH4,cV4,cD4) =  coef[-4]#level4 output of level4 decomp and furthur level dont offer better features\n",
        "\n",
        "  return np.concatenate((cA.reshape(1,-1),cH4.reshape(1,-1),cV4.reshape(1,-1),cD4.reshape(1,-1)), axis = 1)\n"
      ],
      "execution_count": 76,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M0m6Bce0Y_N3"
      },
      "source": [
        "import tqdm\n",
        "import gc\n",
        "def extract_features(dataset, extract_path, name, chunk_size = 500, start_index = 0):\n",
        "  return_df = pd.DataFrame()\n",
        "  extract_path = extract_path + \"_\" + name + \".csv\"\n",
        "  for i in tqdm.tqdm(range(start_index,dataset.shape[0])):\n",
        "    lst = list()\n",
        "    path = dataset[i]\n",
        "    hgf = hog_features(path)\n",
        "    fd, exposure_balanced_image = hgf[0].reshape(1,-1) , hgf[1]\n",
        "    hist = LBP(exposure_balanced_image).reshape(1,-1)\n",
        "    covariance_matrix = GLCM(exposure_balanced_image)\n",
        "    wt = wavelet_transform(exposure_balanced_image)\n",
        "    lst = np.float32(np.concatenate((covariance_matrix, hist, wt, fd), axis = 1))\n",
        "    # return_df = np.concatenate([return_df, lst])\n",
        "    df = pd.DataFrame(lst)\n",
        "    return_df = pd.concat([return_df, df], ignore_index = True, copy = False)\n",
        "  #   if i!=0 and i % chunk_size == 0:\n",
        "  #     print(\"chunking\",i)\n",
        "  #     header = False\n",
        "  #     mode = 'a'\n",
        "  #     if i == chunk_size:\n",
        "  #       header = True\n",
        "  #       mode = 'w'\n",
        "  #     x = pd.DataFrame(return_df)\n",
        "  #     x.to_csv(extract_path, mode = mode, header = header, index = False)\n",
        "  #     del return_df\n",
        "  #     gc.collect()\n",
        "  #     return_df = np.empty((0,830))\n",
        "  \n",
        "  # if return_df.shape[0] != 0:\n",
        "  #   x = pd.DataFrame(return_df)\n",
        "  #   x.to_csv(extract_path, mode = 'a', header = False, index = False)\n",
        "  #   del return_df\n",
        "  #   gc.collect()\n",
        "  return_df.to_csv(extract_path, index = False) \n",
        "  # return_df = pd.read_csv(extract_path)\n",
        "\n",
        "  return return_df\n",
        "  # return_df = pd.DataFrame(return_df)\n",
        "  # return_df.to_csv(extract_path, index = False)\n",
        "  # return return_df"
      ],
      "execution_count": 113,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "extract_path = \"/content/drive/MyDrive/CS5100 Project/manual_features_v2\"\n",
        "x2 = extract_features(X, extract_path, \"all\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 470
        },
        "id": "woeGVCev0XPZ",
        "outputId": "931a129c-5f93-4430-b0c1-4fa0a9ef79c0"
      },
      "execution_count": 114,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "  0%|          | 0/112120 [00:00<?, ?it/s]/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:12: FutureWarning: The behavior of rgb2gray will change in scikit-image 0.19. Currently, rgb2gray allows 2D grayscale image to be passed as inputs and leaves them unmodified as outputs. Starting from version 0.19, 2D arrays will be treated as 1D images with 3 channels.\n",
            "  if sys.path[0] == '':\n",
            "  0%|          | 12/112120 [00:01<3:36:54,  8.61it/s]/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:12: FutureWarning: Non RGB image conversion is now deprecated. For RGBA images, please use rgb2gray(rgba2rgb(rgb)) instead. In version 0.19, a ValueError will be raised if input image last dimension length is not 3.\n",
            "  if sys.path[0] == '':\n",
            "  5%|▍         | 5374/112120 [10:21<3:25:37,  8.65it/s]\n"
          ]
        },
        {
          "output_type": "error",
          "ename": "KeyboardInterrupt",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-114-379166ea3273>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mextract_path\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m\"/content/drive/MyDrive/CS5100 Project/manual_features_v2\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mx2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mextract_features\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mextract_path\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"all\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m<ipython-input-113-847723c88679>\u001b[0m in \u001b[0;36mextract_features\u001b[0;34m(dataset, extract_path, name, chunk_size, start_index)\u001b[0m\n\u001b[1;32m      7\u001b[0m     \u001b[0mlst\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      8\u001b[0m     \u001b[0mpath\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdataset\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 9\u001b[0;31m     \u001b[0mhgf\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mhog_features\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     10\u001b[0m     \u001b[0mfd\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mexposure_balanced_image\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mhgf\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreshape\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m,\u001b[0m \u001b[0mhgf\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     11\u001b[0m     \u001b[0mhist\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mLBP\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mexposure_balanced_image\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreshape\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<ipython-input-73-fb4a9065d9dd>\u001b[0m in \u001b[0;36mhog_features\u001b[0;34m(path)\u001b[0m\n\u001b[1;32m     14\u001b[0m   \u001b[0mexposure_balanced_image\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mexposure\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrescale_intensity\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mresized_img\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0min_range\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m225\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mout_range\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m225\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     15\u001b[0m   fd = hog(exposure_balanced_image, orientations=4, pixels_per_cell=(8, 8),\n\u001b[0;32m---> 16\u001b[0;31m                 \tcells_per_block=(1, 1), visualize=False, multichannel=False)\n\u001b[0m\u001b[1;32m     17\u001b[0m   \u001b[0;32mreturn\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mfd\u001b[0m \u001b[0;34m,\u001b[0m \u001b[0mexposure_balanced_image\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/skimage/feature/_hog.py\u001b[0m in \u001b[0;36mhog\u001b[0;34m(image, orientations, pixels_per_cell, cells_per_block, block_norm, visualize, transform_sqrt, feature_vector, multichannel)\u001b[0m\n\u001b[1;32m    279\u001b[0m             \u001b[0mblock\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0morientation_histogram\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mr\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mr\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mb_row\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mc\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mc\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mb_col\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m:\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    280\u001b[0m             \u001b[0mnormalized_blocks\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mr\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mc\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m:\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m\u001b[0;31m \u001b[0m\u001b[0;31m\\\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 281\u001b[0;31m                 \u001b[0m_hog_normalize_block\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mblock\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmethod\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mblock_norm\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    282\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    283\u001b[0m     \"\"\"\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/skimage/feature/_hog.py\u001b[0m in \u001b[0;36m_hog_normalize_block\u001b[0;34m(block, method, eps)\u001b[0m\n\u001b[1;32m     13\u001b[0m         \u001b[0mout\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mblock\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msqrt\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mblock\u001b[0m \u001b[0;34m**\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0meps\u001b[0m \u001b[0;34m**\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     14\u001b[0m         \u001b[0mout\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mminimum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mout\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0.2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 15\u001b[0;31m         \u001b[0mout\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mout\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msqrt\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mout\u001b[0m \u001b[0;34m**\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0meps\u001b[0m \u001b[0;34m**\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     16\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     17\u001b[0m         \u001b[0;32mraise\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Selected block normalization method is invalid.'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<__array_function__ internals>\u001b[0m in \u001b[0;36msum\u001b[0;34m(*args, **kwargs)\u001b[0m\n",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "fd.shape[1] + hist.shape[1] + covariance_matrix.shape[1] + wt.shape[1]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "msgNz_EGCUTz",
        "outputId": "ad8efcfc-f3e9-4fb2-f637-4143798f0b50"
      },
      "execution_count": 111,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "3966"
            ]
          },
          "metadata": {},
          "execution_count": 111
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "from skimage.feature import greycomatrix, greycoprops\n",
        "from skimage import data\n",
        "\n",
        "\n",
        "PATCH_SIZE = 21\n",
        "\n",
        "# open the camera image\n",
        "path = '/content/00000075_000.png'\n",
        "image = io.imread(path)\n",
        "\n",
        "# select some patches from grassy areas of the image\n",
        "grass_locations = [(200, 250), (570, 170), (275, 300), (200, 600)]\n",
        "grass_patches = []\n",
        "for loc in grass_locations:\n",
        "    grass_patches.append(image[loc[0]:loc[0] + PATCH_SIZE,\n",
        "                               loc[1]:loc[1] + PATCH_SIZE])\n",
        "\n",
        "# select some patches from sky areas of the image\n",
        "sky_locations = [(600, 500), (900, 700), (800, 200), (800, 450)]\n",
        "sky_patches = []\n",
        "for loc in sky_locations:\n",
        "    sky_patches.append(image[loc[0]:loc[0] + PATCH_SIZE,\n",
        "                             loc[1]:loc[1] + PATCH_SIZE])\n",
        "\n",
        "# compute some GLCM properties each patch\n",
        "xs = []\n",
        "ys = []\n",
        "for patch in (grass_patches + sky_patches):\n",
        "    glcm = greycomatrix(patch, [5], [0], 256, symmetric=True, normed=True)\n",
        "    xs.append(greycoprops(glcm, 'dissimilarity')[0, 0])\n",
        "    ys.append(greycoprops(glcm, 'correlation')[0, 0])\n",
        "\n",
        "# create the figure\n",
        "fig = plt.figure(figsize=(8, 8))\n",
        "\n",
        "# display original image with locations of patches\n",
        "ax = fig.add_subplot(3, 2, 1)\n",
        "ax.imshow(image, cmap=plt.cm.gray, interpolation='nearest',\n",
        "          vmin=0, vmax=255)\n",
        "for (y, x) in grass_locations:\n",
        "    ax.plot(x + PATCH_SIZE / 2, y + PATCH_SIZE / 2, 'gs')\n",
        "for (y, x) in sky_locations:\n",
        "    ax.plot(x + PATCH_SIZE / 2, y + PATCH_SIZE / 2, 'bs')\n",
        "ax.set_xlabel('Original Image')\n",
        "ax.axis('image')\n",
        "\n",
        "# for each patch, plot (dissimilarity, correlation)\n",
        "ax = fig.add_subplot(3, 2, 2)\n",
        "ax.plot(xs[:len(grass_patches)], ys[:len(grass_patches)], 'go',\n",
        "        label='Grass')\n",
        "ax.plot(xs[len(grass_patches):], ys[len(grass_patches):], 'bo',\n",
        "        label='Sky')\n",
        "ax.set_xlabel('GLCM Dissimilarity')\n",
        "ax.set_ylabel('GLVM Correlation')\n",
        "\n",
        "# display the image patches\n",
        "for i, patch in enumerate(grass_patches):\n",
        "    ax = fig.add_subplot(3, len(grass_patches), len(grass_patches)*1 + i + 1)\n",
        "    ax.imshow(patch, cmap=plt.cm.gray, interpolation='nearest',\n",
        "              vmin=0, vmax=255)\n",
        "    ax.set_xlabel('Grass %d' % (i + 1))\n",
        "\n",
        "for i, patch in enumerate(sky_patches):\n",
        "    ax = fig.add_subplot(3, len(grass_patches), len(grass_patches)*2 + i + 1)\n",
        "    ax.imshow(patch, cmap=plt.cm.gray, interpolation='nearest',\n",
        "              vmin=0, vmax=255)\n",
        "    ax.set_xlabel('Sky %d' % (i + 1))\n",
        "\n",
        "\n",
        "# display the patches and plot\n",
        "fig.suptitle('Grey level co-occurrence matrix features', fontsize=14)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 539
        },
        "id": "PmUp8l86v5Hd",
        "outputId": "f76181fb-a142-4280-8a3e-6d1de0d76836"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeQAAAIKCAYAAADyCDyWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOy9eZxkZ13v/35q36t6nV5m6XQmCZCEkCE3gwmGQPQSIIpcuEiISwwwuKBXlJ+KQQ3oIPyQJT8UNYA3ooMIysUEkKgQ8Eog+zKZCVlmpqdnpvet9r2e3x/nfJ85VememSQz090zz/v1qld3nao656lT3fU5311prbFYLBaLxbK6+FZ7ARaLxWKxWKwgWywWi8WyJrCCbLFYLBbLGsAKssVisVgsawAryBaLxWKxrAGsIFssFovFsgawgmw5IZRSWin1ltN8zFuUUo+fzmNazjxOxt+uUuqNSqmnlVINpdTtJ2lpFksbVpDXIEqpDUqpT7pfABWl1IxS6h6l1K8rpRKrvT6LZTVRSt2olCo8h5cMAne+wMN+HvhnYAvwv17gvgBQSo24FwuXnYz9WdY/gdVegKUdpdQI8H0gB/wB8BhQBi4E3gnMA19c4bUhrXXttCzUclpZ7rNVSvkApbVurtKy1jRyzrTWUy9wPxmgB7hLa33k5Kzu5GL/988QtNb2toZuwL8Ch4D4Co8rz+8a+DXgq0AR+DN3+08BDwIV4ACwEwi5j/0h8Pgy+/0+8P8dY10aeIvn/jDwJWDRvX0DOM997Hz3+Rd37GMHMAcE3fsvcV+XB2aAfwAGPM+/Zbm1duxzCNiFc6FSAh4BXu15/N3AM0DN/fmuE/gMfDgXQ4eAKrAbeONzPO7rgXtxLqbmcSy0iPvYGPC+jv19F/hzz/0x9/3/DbAEfAW4ESi4+34caAAXASHgo8Bhdy33A6/17Otq9/O4xl1TCXgA2NaxhlcA33H/lrLu70Pydwf8DrDPfU+7gZ87znm8Hfg68LvAlLvPj7jn9xb3M58Cfrfjdb+FcyFaBI4AnwMyHe/Fe7tlpXPW+bcL/Lz7/l/kOd6fup911zLvYbnjXe0+dgXwPXd/R4C/BFKe114L/F+c/48F4C7gxR3/U97bd73nrWMdt+D5X+g4t4eBmeP9X7qPbwL+xV1PCfgR8LbV/t6zN/fzWe0F2Jvnw3CuwlvA753g87X7pfZOYBQ4B3gtjnX9S8C5wKuBJzkq1htxvsgv9+znAndflxznWPKlFgOecr8UXgq8yP3SPAjE3OfcB3ykYx/fAz7j/j6II84fBV7s7udOHMHwuc9p+xJaZk1x4Gmci4kfd9/v/8AVRuBNQB14D85Fwq+793/qOOf1ve45fLv7ug8BTeBlJ3jca91z/Cc4Fx0vBd7nOTdjnJgg53BEcCtwHo4gN4AfAFe6a0viXBj8ELjK/Tt4D84FyCXuvq52P7/73L+HF+GIwxO4F3jAJThCexvwMvczeTew2X18J87f0bU4f2dvxxHMNxzjPN7uvoe/co95Pc7f97dwRPB84Jfdtb3c87rfBF4DjACvwhHnv3MfC+G4jIvAgHtLrHTOOv923ftfBB5y93W1+zfx6hXeQ8j9DLX7GQ+42y7GuTj6bfez2e5+Lv/kee2b3dt57t/Al3EuCuXi+L+5+32tu99uz3k7EUHOu5/9Re56TuT/8k7g393P+xz387x2tb/77M39XFd7Afbm+TCcf2oNvKlj+2H3n78A/JVnuwY+3fHc/wT+oGPbz7ivlS/fr3fs56PAA8dZm1eQb8IRJK+17sexBN/q3v8N94tAjrkZ58v4Cvf+h4Bvdxyjyz3O5e79ti+hZdb0LvdLqXeFx78P/E3HttuB/zrOez0C/GHHtu8Cf/8cjvulY+x/jBMT5Ds7nnMjzxavc93zurnjuV/j6MXP1e7rvFbzle62je79XcAPVlhvHEesf7xj+6eAbx7jfd6OY3n6PdseAB493vnoePxaHE+FXKjdCBRWOK93LrO9U5DT7nNvc9f30eP8PfTisYzdbV8APt/xvJe5z+s/xnlsAq9074+4z79smfN2IoI8C4Q9207k//Ix4I+O9X7tbfVuNqlrffDjOP/s9wGRjsce6Lj/cuBmpVRBbjgWQRznKhzgs8DblFJRpZQfx433+eewnpfjXF3nPcfI4gjque5zvoTj1v1x9/71wAGt9T2efVzVsc5D7mOyj+NxKfCY1npuhcdfjCOOXv4Lx+JBKXWD9/hKqR9XSqXcda/4uhM47qXAt0/wPRyLzs8WHAv5Ec/9bTju5L0d5/INPPs8Pub5fcL92e/+vBTHRb0cL8H5u/tWxzF+ZZljdLJXt8e4p3Hc7XRsk3WglHqNUurflVKHlVJ5nJBMiKN/v8diuXPWhtY6C/wizoXVHPCBE9hvJy8Hfq7jfMjfzLnu+zhXKfVFpdQ+pVQO5336cC5OTwaPa62rHWs63v/lrcAHlFI/UEr9iVLq5SdpLZaTgE3qWls8g3PF/CLvRq31AQClVGmZ1xQ77vuAD+LEHDuZdX9+Ayd+9Gacf9gMKySKrYAPRxTetsxjC+6aZ5RS/w7cgGO134BjhXn38Q0cV24n089hLc8H7f68A8dFLhwBgifwuhdKC0dEvSx33M7PFqDaIXA+d13/Dcf16qXccd/7uLyXE7kol+f8FDB+jH0uR+fjeoVtPgCl1Bacv4vP4uQ7zONcdPwDjigfj+XO2XJchWOt9gMp9zjPBR+OO/iTyzwmiV9fx/Fuvdvd1gD2cvz38Xz/Pk7k//LzSqm7cPIQfgK4Ryn1p1rrW46zJstpwAryGkJrPa+U+jfgPUqpT2utn0tph/AQTsLKM8c4jtRS3oQjyF91rYbncozrgTmt9dIxnvf3wJ8rpW7DiXF5a0EfAt4KHNRaH+9LfSUeBn5eKdW7grX6BI5r1mv9vxLnSxGtdR7H9eylrJSacF/37eVedwLHfRgngeqzK6x7FieGDoBSKoJzEfbwCs8/Fg/jfHkPaK3vfh6v9+7nNSs8thfHZbxFa72SFX2yuAxHsN4rFx5Kqes6nlPDccU+L5RS23GS9t6EI/qfxYkPPxceAi5c6f9MKdWD85n+qnwuSqlttH/nSlZ053uZxfGIeem8v9Kajvt/qbU+jOOuv00p9bs4MflbTmD/llOMdVmvPX4V53N5UCl1vVLqJUqp85VS1+MkYhyvxOVDwNuVUh9SSl2klHqRUuotSqn/t+N5n8NJmLmO5+auBsfSnQb+RSn1KqXUOUqpq5RSH1dKned53tdwruw/D9yvtX7K89hf4MTy/lEptV0pNaqU+gml1G1KqeQJruOLOElt/+K6m0eVUj+tlHq1+/jHcITz15RS5ymlfh3HUu88F518DHife/7PV0p9CMf1/mcneNydwP90XYIvUUpdqJR6r1Iq5j7+HeAGpdTVSqkLcbKCn9fFsXtOdwG3u5/zqFLqMqXU+5RSz0VkPgZc6p7/S5RSFyil3qmU2uxeuPwZ8GdKqZuUUluVUi9TSv2yUmrH81n3MXga5+//N92/q+txkry8jAERpdRPKqV6Pef1uCinjv/vcXIo7sRJTvvvSql3PMd1fhS4XCn1V0qpS91zcp1S6q/dxxdx3OHvch97FU5yW8OzjxkcL8ZrldN7IO1u/w7OZyHn+ndwLhCPx3H/L5VStyqlrnX/Tl6GE5/fe6ydWk4jqx3Etrdn33BiZbfiuLCrOAlZ9wPvB5Ke57Ulq3i2/3eccosSTtbpA8B7lnned3DKWNQJrKkzMWYD8L9xvlSqOOVVf0NHohNO8osGfmOZfZ4H/BPOl1cZJ4v30xzNQr2F45c9bQT+EafMpYRj6V3tefyX3fNY5/mVPdVwSnx+5jke96dxSs+qOF/Md3C07CmF44LN4rgyf5Xlk7o6E79uZPlkpqB7rva7651yj/dy9/Gr3c+g1/OaEToSinC8AP/pfhZLwH8Ag+5jCidLXazlWZxs3Z88xnm8nWcnJ30duL1j2w9xqwDc+7/hnpcyjpfire5aRzzP+Uv3vGray56elRxGe0Li54E98lm4296B8z+2dYX38aykLnf7ZTgZ4zkc9/Fu4EOex1+DEy+vuD9f6x7nRs9z3okTBmjilj15/vYn3b+RzwAfZpmyp2XWesz/S5z/r6fdNc3i5HoMn4zvLXt74TfJgLWchSil9gK7tNY7V3stFovFcrZjY8hnIUqpPpx47gjw18d+tsVisVhOB1aQz05mcNx979Yrl+5YLBaL5TRiXdYWi8VisawBbJa1xWKxWCxrACvIFovFYrGsAawgWywWi8WyBrCCbLFYLBbLGsAKssVisVgsawAryBaLxWKxrAGsIFssFovFsgawgmyxWCwWyxrACrLFYrFYLGsAK8gWi8VisawBrCBbLBaLxbIGsIJssVgsFssawAqyxWKxWCxrACvIFovFYrGsAawgWywWi8WyBrCCbLFYLBbLGsAKssVisVgsawAryBaLxWKxrAGsIFssFovFsgawgmyxWCwWyxrACrLFYrFYLGsAK8gWi8VisawBrCBbLBaLxbIGsIJssVgsFssawAqyxWKxWCxrACvIFovFYrGsAawgWywWi8WyBrCCbLFYLBbLGsAKssVisVgsawAryBaLxWKxrAGsIFssFovFsgawgmyxWCwWyxrACrLFYrFYLGsAK8gWi8VisawBrCBbLBaLxbIGsIJssVgsFssawAqyxWKxWCxrACvIFovFYrGsAawgWywWi8WyBjgjBVkpda1S6kml1DNKqd9b7fVYLBaLxXI8lNZ6tddwUlFK+YGngJ8EDgP3A9drrfeu6sIsFovFYjkGZ6KFfDnwjNZ6v9a6BnwJeOMqr8lisVgslmMSWO0FnAKGgUOe+4eB7d4nKKV2ADsAQqHQy30+H9VqFa+3QCmFUuo0LNeh89ida5DffT4fPp+v7XeAYDCIUopGo0EgEMDv99NsNtFao7XG7/eb35VSaK3Na5VSNJtNms0mjUYDrTWtVss8f7lbq9V61ns4nefrZNFqtea01n2rvY5TiVLqWuBWwA98Tmv9kY7HNwN/C2Tc5/ye1vqbx9pnb2+vHhkZOTULtljOIB588MET/o45EwX5uGitbwNuA9i4caNOJpPs27ePZrMJOMLi9/sJBAJGtF7AsYxQecXQ+7j8lN+DwSA+n49gMIjf7ycYDBIIBAiHw4TDYSKRCNFolHA4TCwWIxQKkU6n6enpoVqt0mw28fv9RCIRlpaWzHFTqRTRaJTp6Wm01kSjUbq7u8nlcuRyOUqlEpOTkzQaDSqVCvV6nWq1Sq1Wo9Fo0Gg0qFar1Ot1Wq0WjUaDer1uLhY6L2LWikAfKyxTLBYPnsalnHbcEM5f4AnhKKXu6AjhfAD4stb6L5VSLwG+CYwca78jIyM88MADp2jVFsuZg1LqhL9jzkSX9RFgk+f+RnfbsniF0MvJEpPO/Sx3rE6B9vv95iZiHAgEUErRarWMqPv9flqtFvF43Dw/HA5TrVZJpVJGqJvNJul0mi1bthAOhxkcHCQQCFCpVGi1Wvh8PmKxGJFIhFgsRjAYJBQKtR1brHGxvuVCJRAIrPheVnq/ltPKiYRwNJByf08DE6dxfRaLxeVMFOT7gfOUUucopULA24A7TuSFp0I8vG7e4x1P3NBincstFArh9/sJhUKEQiGANhd0oVCgXq+Ty+WoVqts2LABn8/H0tIS0WiUVqtFLBaj2WzS1dVFKBSiXC4Tj8dpNpsUCgUWFxdZWFgwFq9Y6bKGSCRiRDocDpsLgOXeY+dPy6qyXAhnuOM5twA/p5Q6jGMd//rpWZrFsj7ZtXsXI58awfdBHyOfGmHX7l0nZb9nnCBrrRvAe4C7gCdwXHF7jve6Fxoz7nQ9LyfCKz3mdfd648Net3k4HMbn8xEKhczjss9Wq0UgEDAu6Hg8TqPRIJ1OU6lUzDH6+vooFApks1mGh4fp6urC5/ORTCZJpVLG2m21WjSbTVqtlrHSAWM5i5UsN+95W06MVxJsy5rheuB2rfVG4PXA3ymlnvXdoJTaoZR6QCn1wOzs7GlfpMWyFti1exc77tzBwexBNJqD2YPsuHPHSRHlM06QAbTW39Ran6+1PldrvfMEnv+cRWKlOPBK+17JLe5NzhKBq9frbdsCgYCxTAGTsCWx7mAwSKvVolKpUCqV0FoTj8eNGEejUebm5njkkUfQWhOJRIjH45TLZarVKuVymVarRSQSMWKrlDK/B4NBIpGIiUuLpS4XC94YeedP68JedU4khPMO4MsAWusfABGgt3NHWuvbtNaXaa0v6+s7o/PgLJYVufnbN1Oql9q2leolbv72zS9432ekIJ9qjic4y1mE8vtKIiy3TstYKUU4HDauY3nMa5UGg0EajQbNZpNAIEA+nycSiZBMJk0yVi6Xa0saa7VahMNharUaPp/PWMRese2MHYdCIZNsJvFl4JiWcuc5OdY2yynhREI448A1AEqpF+MIsjWBLZZlGM+OP6ftz4WzMsu6k2O5qo8nGCdiAXZmIXtd03LzliGJUHlF2CvS3u3hcNgcU8qcIpGIKV+q1WpUKhWKxSLNZpNMJkMymaRWq1EsFgmFQlQqFROTlgsCrzUuYu9dm/cCIhAIGNe5lEMt5zXozDDvtKw7n9eZnX48rLg/G611QyklIRw/8Dda6z1KqQ8BD2it7wB+G/isUuq9OAleN2p7Mi2WZdmc3szB7LMTpzenN7/gfZ/1guyNHXeKwHKcqAvWWz/sjQsDxur0inDna+SnV4QlmUpeI0lZtVoNpRSJRIJQKEQsFjPJXNVqlWq1SqPRIBqNEovFyOfzhMNh+vr6yOVyhMNhstmsEXA5D/F4nFKpZNzoEquWWmdwrG1vvbPUL69Ux+x9r8uJs/ecer0KJ5uzSW+0U1P8zY5tf+j5fS9w5elel8WyHtl5zU523LmjzW0dC8bYec1xo6PH5awX5GNlCJ9IZnSnqHgt304Blt+9z5N9eB8HTGa1CF8kEjGuZr/fT19fH6VSybiRRawLhQKAiQ1Ho1EymQyFQgGfz8fc3By9vb0Ui0WKxSLlcplisWiEO5FIUKvVjGUssWhpHiJrqVarbS53sY69lnLn788lsWsl97f3sU7LWcrCjvf64223WCyWlbjh4hsAJ5Y8nh1nc3ozO6/Zaba/EKwga91m4ck2b82v97levC5nKQESAe10SXs7bHVahyK0InBaa9N5S1zQPp+PVqtFMpkEYGFhgVgsZtYnjTrC4TD1et00DCmXy8b69fv95PN54vG4WafEkcvlMpFIxMShxWVeq9VMQxKtNaVSqc2Klv1orWk0GmadIsJybmWb97HlzulyLFfL/Vw8GVaQLRbLyeSGi284KQLcyVkvyNJuUqw/QbKYve5jsWS9Att5E3FeqZTJ69r1ZjPD0expb7MNieVqrclkMsaVXK/XCYVCbTFkEUMpj5LyJWkWEovFSKVSNJtNI/6NRoNSqWRizuFw2LxnKXdqNBrUajVjGct6vGVS8n6870suEuR8yvO8vz+f5K7OCxrBJolZLJb1zFmfZX2sL3ERDnncm+Us2cjSylJqdGVb53ZpgykCnEgk2up6vclbIurRaJRgMEgikWDjxo3E43FTLxyNRtvWKEJYrVaJx+NG9AEKhQKVSoVyuWx+B0fAE4kEQ0NDxONxU+MsiOjH43HzXrzH93oF5L3Lzds5zHtOQqEQkUjE/O7N2Pa68Tvd3svFo4916/yMP/nJT9rM7rOMXbtgZAR8PufnrpPTu8FiOWWc9RYysOygBKGzXaT8FJeyuHdFgMTC9VrG4mIVV7TP52Pr1q1MTU2ZoRZi2YpVnUgkyGQyJBIJEolE2/69FwmAiSM3Gg1isRiVSoVAIEC9XjcJWMlkkkAgQHd3tymFkrWKhSzIsSSWLGIpFxA+n4/5+Xm01lSrVWNpy7kSy1guEprNprG2vUIrVrm4tTuHW3iFGJZPevNyLLGV45/Icy3rn127YMcOKLl5NwcPOvcBbjj5nkaL5aRw1gtyKBTikksuodVqceTI0X4JPp+PaDTaZhV39pSGo4MoOmt1JSPZm0UtrxkeHqZerxvBa7VaRjB8Ph8DAwPGakwmk2afkUjE7EMsaxHoRCJhBE1ELZlMkk6nyWazxiU/OTnJ0NAQWmtCoRBaa2Mxx2IxI+jRaJR8Pm9c36VSyawzEomwYcMGkskkuVyO+fl509JTzgk4lnMmk2FhYcEMvGg2m4RCIXMfMBcD9XrdZG13urjlfucxOmP83oufzuetlNVtOfO4+eajYiyUSs52K8iWtcpZL8jhcJiBgQH6+vrI5/P4fD66u7uJxWIkEgnTJQswHbQkCUu6V4lINxoN0xfaGweWxhvxeJyuri5e8YpXcOeddxKLxYxl3Gw2jSCKhay106c6lXL6/jcaDRYXF41FHo/HzRqz2SwA6XTadNeqVqvGipeypVgsRqFQIJFIkE6nmZmZAZxuXhJPrtVqZv3FYpFarWbc1qVSyTxXLkzi8TgLCwtUKhUj4MFgEK2djmGFQsGIrljQch7lvHoFW7a1Wi1zzkVQj5UU1pmU94EPfIDZ2VkOHDjwgqd2WdYX4yv0aFhpu8WyFjjrBTkQCDA6Osrg4CAHDhyg2WyytLREvV431qskNHn7PQNG4MSKy2QyDA4Otk1qkpitZDNXKhXuuusu4/IWS623t9ccp1Ao0Gw2qdfrlMtlFhcXKZfLwNGLAiEajTIyMkJvby/RaNQIZrFYNIMkRNhkpGK5XKanp4fFxUUT063X60bQQqEQk5OTZpRjNps1AhyNRlFKmelQAwMDLC0t0dfXh9/vZ3Z21iR8VatVQqEQV1xxBQ888IBJlPPSaDQIhUJmxKNSinq9bmLkktkt9c+dWfHLxZdFvP/oj/6Ivr4+BgcH+dznPvesMinLmcvmzY6berntFsta5awX5HA4zIte9CK++93vsrS0RLVaJRwOt8WCRTxrtZoRA3EpiwWaTCaN5RgKhWg0GsbFnMlkiEQiHDhwgGq1ahpySGa2V+TEGjxy5IjJep6YmDBCLOIqbvJQKES1WmViYoLu7m5GR0fbmoSIu9fn89Hb20t/f7+xhMUizWQyzM/PE4lEyGazjI+PMz8/T6lUMq5qEUoR4Xq9TjabNS06xQoXgW42myQSCUqlEnv37uW6667j0KFD7Nu3r014u7q6TMKbWMr1et3MYhb3ucSk5f2LlS0WvLzPzrjwrbfeyu7du/n4xz8OLD+T2nLmsXNnewwZIBZztlssa5WzXpBFCAcHB40g/ehHP8Ln81GtVo1YyAhCwLhXK5WKcXOLaMdiMbq7u6nVaqTTaWKxGACLi4sMDQ2ZsYgyd1hEKBgMUiqVmJycBGB2dta4dyXGKyVHgHGZg2NlLi0tAU4cOJPJUK1W2bRpE/F4nGq1Si6XM9Z/V1eXsTQBstksPp+P2dlZZmdnmZ6eNl4Cr9XZaDQoFoscPnyYWCxGo9Fgy5YtZLNZMpkMmUyGUqlEIpFgdna2Ldnrscceo6+vj3PPPZd0Om3W5W1kIudCkAsHae8pcWTvc6SGOpfLtV24yAVVIpHgK1/5inkPKyWEWc4sJE58882Om3rzZkeMbfzYspY56wVZkpoqlQpjY2NMT08bsRRBlszoWq0GYCw4GfogZT6JRIJsNouMppPWlSL4kn3c09PTZr3W63UOHTpEsVg0Qijx2UKhYHpTe4dKdJYaAWbaU6vVYmBggJmZGSOcYtHXajVyuRzpdJru7m6mpqaIRqMsLi6SzWaZmZkx1r2Itlj80hBEYtPBYJB9+/YRDodJpVIsLCyYjPDh4WEqlUpbgpuUY4lLPRQKsWnTJnNBUi6XqVQqtFotisUiWmtisRi1Wo1gMEg+n6fkmjxyLnp7eymXy3R1ddHb22suAPbu3UuhUODNb34zl1xyCWNjY8/qkW05s7nhBivAlvXFWS/I4FhiYoWKper3++nt7TXCFAgEKBQKRojAsbKkdldisn6/n5GREX70ox+Rz+fNMebn540bORgMEo1GCYfDtFotDh06xMTEBD6fj0wmY9YiAigWufwuVruIr9/vN25xSQLL5XJtJU5ey1BiudlsllgsxuzsLIuLi6ZcSjwCMspRRFBivZKwJXHder3O7t278fl8bN++3aylXq9TLBbp6ekxlqvE08G5gHj66aepVqskk0lzMeJ1Pcs0KondJ5NJsw+x/MPhsElGkzKs3t5e4vE4tVqNxx57zIqwxWJZ81hBBlNaVK1WTYxS2k4CdHV1EYlETJaxxC+9HbjC4TDbt2/n8OHDjI2NmZIi75jFQqFAMBhkYGDA1C/v2bOHUqlEd3c3wWCQSqViXieWqt/vN+MTxTr3Znhrrc00p0QiYQS4VCqZ54vLNxwOE41GqVarRCIRE38FTKw5HA4zNzdnPAQivJLEViqVTImVNPyQbOt9+/Zx7rnncsEFFxjRlGlS4HgXJiYmyGazVKtV+vv7aTabJsFM3NyAOaZ4C7xlU3C0PWk4HCaZTDI8PMzS0hLPPPOMuVDwWumAdVtbLJY1y5oWZKXUJuALwAacsXC3aa1vVUp1A/8IjABjwFu11ovK+Ya9FXg9UMIZI/fQsY7hnfc7MDBAPB6nWCyaIQtiKfb29ho3c7lcNtnJ0iyjXq8zMzOD3+83vaeLxaIRABE/2W8mk+Ghhx4in8+beLNYd93d3eZYMzMzRhABI0w+n880/pDHIpEIlUqFubk5hoaGTEKZuIm11vT39+Pz+czQCJkK5RXlhYUF06JTrHVxmUvsOpVKGXe4DKFIpVKEQiFKpRJPP/0027ZtM68pFouUSiVzTsSNvbi4iFKKc845h4WFBdM7WxLXBG+DkXA4TKVSoaurC4ClpSUWFxeZmJigWq0yPT1t6rzFRS0Z8xaLxbJWWdOCDDSA39ZaP6SUSgIPKqX+HbgR+LbW+iNKqd8Dfg/4XeB1wHnubTvwl+7PY9Lb28uWLVvI5XLGulRKsWHDBtP+sVKpEI/HjbvZ5/ORTqfNJCSxOEulkhGUSCRiLDgRUGmWcffdd5PNZgmHw6TTaWM9i8DJ8TZu3GjiyJFIhOQvFHYAACAASURBVJmZGVMiJEItFuPS0hI9PT34fD5jyYuFKvFvEWGxjEW4ZH3FYhE4mmQlQiaNRMRqlnUmEgm6urqMG1zc+aVSiccff5yXvOQlhEIhZmdnKRaLxvsgrn7xAkiG+dDQEPv376darVIsFk38Xi46ZCSk1FB7u4KJ1Z5Kpcz7ECvYO8/ZYrFY1iJrWpC11pPApPt7Xin1BDAMvBG42n3a3wLfxRHkNwJf0M637g+VUhml1KC7nxUR4fR245Je1NKzWb7YY7EY8XgccDKnk8kkSinTFKNQKLQ9v1wuEwwGKRaL9PX14fP5uO+++/g/5/4fqsHqs9YSqoV43Z7XtdXlptNpIzCSRS2WqrjPBakdlv7RlUrF1EP7fD6TACXu5nK5TCKRMAlhEouVzmDhcJhgMMjj1z7OdHR62fVedd9VJmlLhFnizWNjYwwNDdHV1WUSthqNRltf797e3jY3/PDwMI1Gg1wuZ+rAxUKWGLNcpHjrsr3nwdvVrONvymxfT+KslLoSuAXYgvN/qwCttR5dzXVZLJaTx5oWZC9KqRHgUuBeYINHZKdwXNrgiPUhz8sOu9vaBFkptQPYAbB58+a26UbiJk0mk20DIXw+Hxs2bKBQKJDNZkmlUmQyGWZmZozbVyzRaDRqSoYkJistJx955BEn+WgZMQaohWosLS0ZF7NYkYlEgkgk0pb0Ja5rcDpqxWIxk7kdDoeNeEsmtXfCk4iztKqU+Hl3dzeAac4hiV7NaHPF9UrLy4WFBQBSqRRXXnkl+Xye8847j927d5vOZwsLC6buWsY9inscnLGSIr6jo6NMTk6aemZ5z+VymXQ6bWLD1WqVWq1mWnzOzs62ZYp7+1aLGK9D9/XngfcCDwLLfxgWi2Vdsy4EWSmVAP4Z+E2tda6jf7FWSj0nU0drfRtwG8Bll12mW62W6RcdDAbJZDKmD7WMI5Qxh95t09PTpr2l7AOODmcol8ukUinTE/vAgQMARlxWwpuRHYvFjNWezWbN0AdJpNLamZc8ODhoYqbBYNDUHYvQNZtN06lLa006nTZWbLPZJJ/Pt10EiNdALNdjEQwGmZmZMVnb4NQ2Ly4u8vKXv5ytW7eyb98+M1FqaWnJZGrLGmTYRDKZJJFIMD8/z+zsrDl38j4lni/nUVzgsm7JB8jlckxPT7e5qr0ivA6HS2S11v+62ouwWCynjjXf4FcpFcQR411a66+6m6eVUoPu44PAjLv9CLDJ8/KN7rZj4p3kJIjLGDBJWtLTeevWrRw5csQkJ0mMuNVqmZGJ4MQ7xdLOZrMUCgXTCONYeNtszs3NmdaZUvrkjeEmk0mTaCa1ySLmvb29hMNhc4HgFSApPZK1ymQpiVWff/75bN++nVe96lVs3br1mOudm5sz7udarcbs7Czj4+Ncf/313HvvvQQCASYnJ9sGRHgvNKS9p8/no1Ao8Mwzz5BKpfD7/SSTSQqFAhs2bGhzo0sTFu9FkKzB5/MRi8Xo6ekx71+E2ju7ep1xt1LqY0qpH1NKbZPbai/KYrGcPNa0hexmTX8eeEJr/QnPQ3cAvwh8xP35L57t71FKfQknmSt7vPixexxTJtMpGBK/FKGVEilJSvL5fKYmV+KZ0ilK4rjFYpG5uTkT+5R48EpI8pd3ypJkKotFGAgE6OrqYnp62swnrtfrpnxKsrrFwpVyqEAgQH9/P/F4nFwuRzQapbe3l8HBQXp7e7nnnns4dOgQjzzyCLFYzJR+sXHl9YqXAI5mej/xxBPMz89zzz33sG/fPiPUQ0NDxioWy1ZrbeLWcs4mJiZMi07xBsjkKG/XLb/fbz4LqV+Wz0Kywb1tN0WI16GFLMmJl3m2aeA1q7AWi8VyCljTggxcCfw8sFsp9Yi77fdxhPjLSql3AAeBt7qPfROn5OkZnLKnXzqRg0ibRenKVavVTEML+RKXbOP+/n5TZyyuc/myl4ESksksVt7s7CytVovZ2VkTbz0WMh1J3LqJRIKpqSlj5YlruNFomHaVcHRQg9QES2KT3+83Yt7V1WUuPBKJBAMDA6TTaZ555hkefvhhcrkc2WzWxGRnZ2fbWlWuhJyDcrlMNBqlv7+fj3zkIxQKBUqlEqOjo6bvtFi3sg5xSUsNuFyMLC0tsWXLFtMERWLjIraSsCaJX3C0V3Vn2085d/JYZzLcWkdr/erVXoPFYjm1rGlB1lr/F0426XJcs8zzNfBrz/U44m4WQajX62aKk7THlIzrZrNpMq8lU1msO9mXiCI4rm+pNRZ3c7FYJFgNUg/Xn72Wks+4o6XuVgRSSn68btpSqWQuAkKhENls1jQyCYVC9PX1sbi4aJLNZCKV9HmWi4Wenh6q1SqlUonh4WEOHDhgLiCazSb+sn/ZxC5f2WfmSPf29poGJFJOJUlX4+PjbN26lZmZGc4///y22me/329qsGVClIjoU089xbZt25icnDQiLa5viemLtd3ZsERKycRlL/OjxbJeTxayUioN/BFwlbvpe8CHtNbZ1VuVxWI5maxpQT5dSBmON1Na6o/FBZxOp0kmk+TzeRMzlpizxChFSH0+H8lkknq9zsLCgkma8mZsb/+/2ykUCgQCAWZmZow41Ot1ivWiEV0RXCGRSJgxiBJXDYfDJt4tJU4y5CGbzZokrp6eHprNJt3d3aZ5hghVpVIhFovR399PoVBgYGCAcDjM4cOHqdVqDH1xyFx01Ot1LrjgAnw+H48++qgpc5qbmzO1xPPz88b6Bacca//+/WzZsoVSqWSsZFmfPE/iyjLzOZVKmfMtDVnEFe+dl1woFIyVLcLrjRt3dvnyXkStE/4GeJyj3qCfB/438D9WbUUWi+WkYgWZ9naKMvJQ4r3Sc3phYYFCoWD6Q0vfaHGLehOFxOUqzSsmJibI5/PEYjEqlUpb1vLExIRZA9DWjEPaXDabTZLJJLVazVid4orN5/OEQiHT4xmO1uDKcaSbWF9fX9vwC4m7SntJiccODQ2xsLBgBmx4y68k+zqRSJjYdaPRoFwuG0GcmpoyXc8k5i1x3NnZWTKZDKOjo8b1LMMqpI+2uKO93gZpoSn9qsXyFtezN3wgLTtlzd6ELxncsc7EGOBcrfWbPfc/6AnjWCyWMwAryGA6bYlbV2LAUrokzUDEHQoYwRB3t0wnEgstHA4zPz/P+Ph4mwtcXNkzMzOmvKhcLpvY6eI7FmnFnl0j6y/72fKVLSYrWcY4SqlSrVYjFosRjUa55ppr2LNnj4mLSwLX2NgYPT09xooXt7h3gIU02+jp6SEQCFAul1laWjLPDQQC9PX1mTi19N6WecZSnrS0tEQ+n2dgYIBGo2G6gzUaDebm5hgdHV32IkYmW8k2KfPq7+9namoKONoApNVqtXVXk5BAvV435V7S7Uvc6F5vg8Sd1wllpdQr3TCONAopr/KaLBbLSWTdmQmnArGYpIZWelQDJj4rrlQZ4iDZ1dJUQ9zX3v1Iba3cL5VK5HI55ubmTAxVWnBGo1GAZcUYoBltmpaYxWKRQqHAwsKCaXkp7lulFI8++iiXX345mzZtYnh4mMXFRZNo5XXzeicoZTIZ051MXL5yYZFIJMy+vR20xsfHiUajJkFLXOkDAwNUq1UajQbT09OmflhixIVCwbj+pSypVCq1vRdxMYu3QpqBpNNpk3gnvcZlhKZY/HLxIK5ur+hLeMFrVa8TfgX4C6XUmFLqIPDnwC+v8posFstJxFrILuJyle5RkUiEDRs2mPivDGnwZunK7/W6k5zVarUIBoOkUinK5TKHDx82yUU+n49sNmus5EqlQk9PD0opFhcX2wRyJSTeWqvVTD1uNuvk9AwNDRnXtLhs5+fn8fv9DAwMMDw8bNYq71UsRhFTr6tXXPEizOIaF5ex1pqpqSmzlmg0SjKZpFqtsrS0BBx1wx86dMi4qBOJBNVqlSeffJIrr7ySaDRq1iPxcDia+S6WrSRsyUANscoliUysYG8ZVbPZJJVKmQsq8WCId8Bbd77W0Vo/AlyilEq593OrvCSLxXKSsYIMZmyh9I2enp6mVCoxPj5OJBIhHo9TrVaNhSUuXynhgaPTiCT5SCxgcQWPjY0Z4anVarz4xS9my5YtNJtNent7+eIXv2gs6ZXwCqd04Wq1Wmb+sWQTX3zxxUxOTuL3+7nggguIx+NtWeFPPfUUgIntSk/rcrls3MqtVovFxUXy+Tzj4+NGxCSuLpaylB15hVR+Fw8BOKK8adMmlFJGuCWxSqxV2SbvSy5wxLUsDVpkbrIkg8k8ZDmmEAgESCaTzMzMmG3eMqj10D5TKfVzWuu/V0r9Vsd2ADrq8y0WyzrGuqyhza28sLBALpcjl8uZL26vZSYlN94pQ2ItBwIBUqmUaaspFtuhQ4eMVddqtdiwYQOvf/3r6evrY3Z21rSSzOWObfSIm1wywWu1mhlHmEgkjBg/+OCD5thSd5xIJIjFYsYqFOtTYtDyHmVMopQ7ZbNZcyEgCWVKKXK5nBlkIRZ1uVwmEomQzWaJx+PGZQzOxcTk5GRbOdfY2JiJuXvj6yLO8h68jVHEC5DJZIw3QGLH4mGQOcjFYpFsNmsy3mWfcox1Qtz9mVzmllitRVnOLnbt3sXIp0bwfdDHyKdG2LV712ov6YzECjKY/sqSOS0xTBGgZrNpkoe8X+rNZtPUwEr3LREbsTgnJyfbjhMIBJibm2PPnj0sLS1RrVa56qqrTqgmVqxJsZDluJKEFo/H2bt3L36/3+z74YcfJhwOMzc319ZRSzLLpQmKCLJYkCJoY2Njxp0spWASu5UErGKxSLHolGql02kAc+68faZlZrS8DxlVKXFrcCxYicmLBSzCDDA7O0sul2NpaYlQKERXV5dppSkJZhISqNfr5PN5EomjuiUXJFKLvNbRWv+1++t/aK0/6L0B317NtVnODnbt3sWOO3dwMHsQjeZg9iA77txhRfkUYAUZRwRKpRITExMmY1fEU+LF3m5QIkjiwvY2ufB2xhJrUJ4vVnU0GuUb3/gG//qv/8rTTz/N+9//fi666CJHWIrLW26+ko+enh4ymQwDAwP09PQQi8VIpVIMDg6SSCT4iZ/4CVP6JLW++XyeRx99lIMHD3L48GHm5uaM9SuPA2Y0oiRFjY2N8fjjj1Mul40lLdOkJHlK4rHy2MLCgolpi+g2Gg1T/iQXM4VCwby+VCqZNYs1LXFh6efdOXZRkstmZmZM85KBgYG2enC52ABnTvRFF130LHFfT41BgE+f4LZnoZS6Vin1pFLqGeXMD1/uOW9VSu1VSu1RSn3xBa3UckZx87dvplRvH4hTqpe4+ds3r9KKzlxsDBlM1rC4p2VoRGfMeG5ujoGBAeM69pYsRSIRkskk0WjU7E+SmyRGC06P7KmpKWNtS03xD37wA4aGhih/oWwEXNzo0t2rGnZGNi4sLBCLxUydcb1eZ3h4mMnJSZLJpBk6Ie7latV5nc/nY/PmzeTzedPistlsmn7Z8rwjR44wPz9Ps9mkq6uLVCpFf38/4+Pj5n1LDTJgGpxID20ZsRgMBo1lLEMitNYUi0Uz01ksWBFaucnaG42G8UKUy2XzGclnIvs+cuSIaTUq58Xn85mLgOnpaUZGRvjRj35kPvf1UIuslPox4AqgryOOnAKOa+IrpfzAXwA/iTOO9H6l1B1a672e55wHvB+4Umu9qJTqP5nvwbK+Gc+OP6ftlufP2v9GOsWIW1SsJensBI4lLEIqrmhon6Urbk9vLFNGB3qbjIRCIarVKtPT08YCl4QoSRA7dOgQ6XSaSCRiErQAuru7jQvX7/fT19dnyoVECLdu3Uo2mzUWq9RRSxMOsfbHx8cpl8smQctr5UsDjVKpZLKnU6kUPp+PiYkJgsEgXV1dhEIhU8rkTVRrNBrk83nq9bopy/LGztPpNNVqlWq1alzX8pjEeTs/G7FopdyrWq2auHkoFKLZbLKwsGCyqmUIhZxbmYxVq9UoFAqcc845bYld64AQTqw4QHv8OAe85QRefznwjNZ6v9a6BnwJeGPHc94F/IXWehFAaz2DxeKyOb35OW0/mZxtsWtrIYOxxuDZXbtEGL1lTiK8UpoTiUTMbF+xAMER6aWlJbTWTE9Pm/2LqEiylHS+qtfrJnO5Wq0ad/LCwoKJscrUpOHhYePKTaVSHDhwgFKpZEYSJhIJMpkMsVjMZDVrrenu7ubIkSNUq1VT4hWPx407t1qtkkql6O3tZWlpycTCpcGIDIiQ10uSGWAsYm+GtCRViUs/FAqZbmXiVk6n0yam652WJXFr2b83ZCDZ295wgTRxyefzxk0uE668SWObNm3i0KFDp/Av6uShtf4e8D2l1O1a64PPYxfDgPfNHubo5CjhfACl1PdxrO5btNbf6tyRUmoHsANg8+ZT/2VsWRvsvGYnO+7c0ea2jgVj7Lxm5yk9rsSu5bgSuwa44eIbTumxV4t1YSKcSkRAgLaBAxIzliQjoVqtGleplDVJa0tpWTk/Pw84XaKkVlgSj8SCFctRREyOHwgEaDQadHd3m6Sp+fl5crmcOaZ0z5Kxis1mk6WlpbYSIhH8QCDAwsIC4XDYNDkZHR01pUmSQCWTrsLhMPF4nFgsZiZNxeNxU4MsQjg7O2sSp5LJJPF4nFQqRT6fN6VZcpEizxMRTiaTaK3NjORYLNbWfUuaigQCAXMh4fVKiDDLa+Siye/3k81mCQQClEol42mQiycJE8RiMQYGBk7Z39QpoqScecjfVEp9R24nad8B4DzgauB64LNKqUznk7TWt2mtL9NaX9bX13eSDm1Z69xw8Q3c9lO3sSW9BYViS3oLt/3UbadcFM/G2PVZbyGLy9rbhEKaS8j4RMDUE4uAedtmijs4Go0a606EW/pGi0tVhCmTybTFQb11sRKvFhEV61xENhKJMDg4aEqAtm3bZixnbxmRdByTftBihYqoyyQkbwvJWq1mRhzKhcbg4CBLS0tMTU2ZemfJkAZM8pq4smU/3oS3UChELpcz84m7urrMhY4kzslFgZxrQYZHiEdBjun1bIiVrLUmn88bj4K3JlpKxSQ23mq11o2lDOwC/hG4DqdD1y8CsyfwuiPAJs/9je42L4eBe7XWdeCAUuopHIG+/4Uu2nJmcMPFN5x2q/RsjF1bC9nNGBYRFUtKtgMm8cr7GrH6JAs5kUigtWb//v3med5xgJK17W16IdOhpJtXuVw2bTHByTbu7u42AiV1vr29vXR3dxOPxwkGg+b3aDRq3MDiCpeMZjhaiiRxZ2+2calUMl26pIOXDMNYXFw0/bOj0Si9vb0kk0mTjCU10NItTNqHitXqjb1LzLxSqRCPx5mamjLHEotYkrG8/bbFUhbPgjdbujMRTOqn5bPxDp6QmcjFYpF1ZuX1aK0/D9S11t/TWt8EvOYEXnc/cJ5S6hylVAh4G3BHx3O+hmMdo5TqxXFh78diWUVWM3a9Wpz1giyIZewduABHE7hEmObn55mYmDDtMMUqLJfLJkFK9iXdowKBgGni4W0PCY7LWFpiikvbK0STk5Ns2LCBTCbDxo0bjTvYm1Gcz+dNsxCxgqPRqLH8xVUr65BGKBKrlmOJOzoUCtHf329cyTInub+/n61bt7K4uMj09DSpVIorr7zS1Dw3m00uvfRSXvWqV6GUMlaotOyU8+SdFy1dzWTdsl65AADM+5LXyVrlMxIPg7jrxeuRz+fJ5XLG25FMJs15r1QqJka/TpCMt0ml1BuUUpcC3cd7kda6AbwHuAt4Aviy1nqPUupDSqmfdp92FzCvlNoL3A38P1rr+ZP/FiyWE2fnNTuJBdu7F56O2PVqsi5c1m7pxgPAEa31dUqpc3CyRXuAB4Gf11rXlFJh4AvAy4F54Ge11mPH279YXN5eyNKxS8TPO57Qm+AlsVJxVRcKBZPMJRanlO1IgpKIqpT6wNFkJ5mgFIlESKfTzM3NMT8/TzweJ5/P09XVRTqdNpaez+dj//79DA0NmQQzmSwla4Kj2eCxWMwcV5K/JGlK4uUi0IuLiyQSCZPIFY1GmZmZMQ1UpqamzDEHBgaYmJggm83S29vLtddey7e+9S2TnCUiGggEWFxcZHR0lGq1SjKZNBcKkkQHtGVNy/kFjNteXO/yOrGivbFn+VmtVunq6jIeCimfkguQdcKfKKXSwG/j1B+ngPeeyAu11t8Evtmx7Q89v2vgt9ybxbImEBf5zd++mfHsOJvTm9l5zc4zNqEL1okgA/8L5+o+5d7/KPBJrfWXlFJ/BbwD+Ev356LWeqtS6m3u8372eDuvVqttpUAybjAQCJDJZIzoSNKXiKkkPoklKNZnNps1sdhsNmt6RovlKu5UibGKOMjzJHNZSq0KhQKhUAilFH19ffT09JgEqZ6eHjM5SURf5iOLhSzC7W3/6bXYJc4sbvVIJML09LTJsPYmXJ1zzjlMT0+TzWZN0prWmgMHDhAOh5mdnaW7u9u0EhWrVWqKJZ48Pj7O0NBQW1xektDEKhf3c6lUMjXO3uxqeT8ST5bz6XW9y+c4NzdHV1eX6dol4r1e0Fp/3f01C7x6NddisZwuViN2vZqseUFWSm0E3gDsBH5LOSbla4C3u0/5W+AWHEF+o/s7wD8Bf66UUvo4LZmk/nZpaYlDhw6ZsYZdXV0mU7her5u6V697VcQqlUqZRiDeoQmJRMJ88ft8PhYXF9sSk0SsRBy9YtJqtUgkEsYivvzyy013rnq9TjqdZmBggGKxaFzh4kIXcZb1ijUvSVOxWIxqtWp6WUsjD8C4eM8//3wOHz5ssr5lzKJMfmo0GkxNTXHjjTfyhS98gVqtxo033sj4uJN0sbi4aMYlyvsXIZWMdMmM9pZayXmRxDLp2CXnN5PJmPPvHbdYr9dNgppYzXKx02g0WFhYMHOcveVtaxml1KeBFf9+tda/cRqXY7FYTiFrXpCBTwG/g9MMARw39ZIbGwMnQ3TY/d3UXGqtG0qprPv8Oe8OO+sp8/k8ExMTPPHEExw+fNhkTIulKolY3tm96XS6LYkIMM0warUasVjMWIXSESufz7e5YKUvtFiv4rKViwDAWLKZTIZarcbo6GibIIplK67u7u6jYcVarcbIyIjJUBZXeDQaNeU/UoNcKBSM1Q1OydbMzIx5juwvmUwyOjrKgQMHqFarDA4Osn//flKpFJs3b+b+++/nySefZNu2bVx99dV85zvfaUtMk05i0jxEXNTexidyrGazSaVSMS51sWpzuZx5r53DP7w3yayW67Fms2lK0qLRaFt2+RrmgdVegMViOT2saUFWSl0HzGitH1RKXX2y9qu1vg24DWDbtm16//797Nmzh4mJCZMEVCqVTGmPWI/9/f2mDabMR261WqaFpLTZjMfjxtIDp7GH1xKWJCdxf4vb2htPBsy0JKmblUESUqo0NTVFOp02buVarUapVDJxYZnLDEcTo8S1LvOXJalKrF8ReIkJ79692zQYEQtfktUA9u7dS6PR4KUvfSl33XWXaZZy6NAhzjvvPJrNJrOzs2za5FTeyMUOOKI4Pz/Ppk2bTHKaJLe1Wi0TRpA1S5tSsYrz+Tx+v5/e3t62GLKcXwkheLPOA4GAie+vhyxrrfXfeu8rpWJa69JKz7dYLOuXtZ5lfSXw00qpMZwkrtcAtwIZpZRcTHjrKk3Npft4Gie5a0Wq1apJRpI2kIKIg/xcWlriySefNEIrj0nmrzS6kNhtNBqlUqkYsfSKg7TOlOfHYjESiQQvfelLGRgYMHHp0dFRtmzZYub+Siy00WiY9pWZTMasR6xFcZV7LyzEwpf5yHIRIKInIiau+Q0bNpBKpRgbG6NYLBrLPpfLGRF8wxvegN/vZ9OmTaa2GZwY79atW43rXeYzi2ve655XShGNRo0bu1qtGsu4UChQKBTI5/NGlBcXF1lcXGRmZobp6WkeffRRarWaWX+5XKZSqZjWntJZTCxn2a+4wNcDSqkfc7Ogf+Tev0Qp9ZlVXpbFYjmJrGkLWWv9fpym97gW8vu01jcopb6C08f3SzgNEv7Ffckd7v0fuI9/53jx41qtxuTkJPPz88Y1KnjjjFKmIxnBYqmJyAWDQePSXVxcNM0sBGm2Ia5ZsVDd92laWE5NTZna4VAoRDQa5WUvexmNRoMtW7Zw9913MzY2xsDAAJVKxbjEa7UaPT09xo2ezWbbpjMJ0s5SErjkYkHird6xiuVymS1btrC4uEg2mzXHkt7TfX19/Nu//RsXXnghe/fu5Wd+5me4+OKL2bdvH9lslk9+8pO84hWv4PHHHycejzM9PW0S0tLptLnQmZiYYHh42CSIeZt9yEAKGQcp06EkAU/i+Y899hgbNmwwvcNlprO4pb2NRuRPQtzX64RPAa/FrSHWWj+qlLpqdZdksVhOJmtakI/B7wJfUkr9CfAw8Hl3++eBv1NKPQMs4DRBOCbNZpOnnnrK9Jr2IoIsWcpiVUoZkM/nM0MRpBFGPp83jS9kahM4wi9Wq4ikxI7FalRKMTU1hc/nIxKJsHXrVgYGBoyoHD58mHQ6TTQaJZFIcM4553D48OG2+l6JP4fDYRKJhFmnCLxcQHiTq6QUSzKuZdxid3c3559/PgBPPPGEsTiTySTbt2/nH/7hH+jt7UUpRSQS4VOf+pRJ1KrX63ziE5/gnnvu4Rd+4RcYHR1lamqKUqlEb2+vmUglzUOUUqaRhyTKVatVarUaY2NjxnKWMY7ipvdeNBWLRRKJBLlczjQ/8XoxvKVOsn09obU+1JGI1lyttVgslpPPWndZG7TW39VaX+f+vl9rfbnWeqvW+n9qravu9op7f6v7+HG7DbVaLSYmJkxiluAVY4k9ypehz+djamqqbTygTEgyoxLdTlNwtDVkpVIxQuHtCAbOCMjFxUUikQg+n49SqcTu3bupVCocOHCA66+/3sRuxQJMJpNGhKX1pNcVLSIEmIxruRiQWmhJtpKs7whlcgAAIABJREFUZbGW/X4/sViM3t5eRkZGuPDCC0mn0/h8Pq677jq+/OUv87M/+7NMTk7y0EMPceuttxo3vCRnBYNBtm/fzoMPPsj+/ftNrDqXy5lJT94mIUop4wqXOPD09HTbZyMXEnJe5cJHhHpmZqZNtKV8TD4n+V0ujNYRh5RSVwBaKRVUSr0PpxTQYrGcIawbQT5VeCcFefF2ffJmUsPREY3eblEidrVazVjQ3qEIMghCnislUd4sY4lFFwoFYw3edddd7N+/nz/+4z/moosuIh6Pm0lG4v6W+cTeOl5JDJNji6UubnkpGQLMuEbv5CoRcqnDHhoaYnBwkNtvv53LLruMSqXCJz7xCUZGRnj00UcJh8OmVajP5zPJYrLP1772tW3jDuXiRcT38OHDRlz9fj+FQoHJyUkj7iKgK7XMlIYucgyv9SstO0WovaK8jvhl4NdwKgmOAC9z71ueA7t2wcgI+HzOz11n9jQ/yzpjvbqsTxreL/flXJiSgCQWpAibdOCq1+v09fUxMTFBpVJhamrKCJ0ISDgcNuMKxUUtWcLyXBGueDxujunz+VhaWuKCCy4glUrxwx/+0EyNSqfTPP3002Z+sljaEquWARPSY1tixYBJngqHw23WojQVEfe1ZJtL1rbWmv/4j/9gdHSUD3/4w3zuc5/jT//0T80YR7m4kJ7X0prT5/Nx77330mq1TO9rb+2wlHBJ0xQpQ5OLCzn/8nl53bYiwLJmL+IN6BRnuRhYL7id6m7VWp89HRJOAbt2wY4dIEUCBw869wFusGfWsgY46y1kaE/4EeSLXBKuOkVAOmiJxSmlSOKeFiR7GDCNOESMJIlKxC+XyzE7O2sSwiqVCtu2bTNNLZrNphHg0dFRotEog4ODpoQqkUiQTCaJxWLmGBs2bKC7u9s0BkkkEsaClbV7LU9Zk0xe6u7uNq975zvfyX333cfBgwf52Mc+Rrlc5oILLjDDNryNUrx9wQOBgGlU4s3sXlhYMOdN1lyr1Xj66afbPhNJoOv8nLzxfW882XuT9pty6/RorAe01k1gi3KGQ1ieJzfffFSMhVLJ2W6xrAXOegsZjlpXnV/QIsheMfaWQrVaLTKZTNvAA6BtipNYliLE0qRD6o9lpKA0tpAYdHd3t3GjN5tNzj//fGNRl0olHn74YdLpNEtLS0bsRPCkxCkWizEzM2Nqo/P5vHF1y2hGsdgDgUBbvFX2I4Mo8vk8d911F6985Svp6+vjqquu4tprr+X+++/nwgsvbOvyJfuXLPNGo8GFF17IAw880DZKUrwE3d3dFItFIpEIi4uLz/oMJJPa2+TD+5l4E7uW83TIMcW9LffXWVLXfuD7Sqk7gKJs1Fp/YvWWtL4YX2Fq30rbLZbTzVlvIS/3Bb/c41KrK5OEtNam/WSr1WJqaqptFnE6nQYwGdRwtKWmt1e1iHEwGDTWq9Q8Dw0Nkclk6O/vN5nMkUiEm266if7+fjP7VwRVrFFx8coUJBE/sX5FlEulkhE7wJRZibtZ+kCHw2H6+/sZGBjg0ksv5cEHH2RpaYnPfOYzLC0t8Za3vMXEj+V9eoda/Mqv/Ar33nsv4Ay3qNVq5HI5k3jlTboqlUptvbPlc/FeCHmFtFOY5Wdna0yvhSy/rxcL2WUf8HWc/9mk52Y5QTavMLVvpe0Wy+nmrBfkYwmxfHmLK9fbTUusS+keJf2kxTrstDyXlpZMswvJYAaMBSqJS41Gg56eHgKBAH19fWQyGdO6M5VKcdNNN/HDH/7QlCyJgIqoiStaLG7vQIlgMNhWaiTxaG8Cm7ymVCrRarVMdzCfz8fb3/529uzZw7ve9S6Gh4d505vexBNPPMGFF17I7OwsF154IU8++aRx2weDQSqVCuPj4yQSCVPaJMltgGn8IfFloO2cez8nb0JX52e40ufo/Ty9rmtxZa8H3Bjy+VrrD3beVntt64mdOyHWPs2PWMzZbrGsBc56QQbahHY5xJKSGluxKAcHB80MZJnmBEdd1tJ3WR4TUWw2m6YtZDAYNDW1IpAAF198MS960Yt4yUtewvDwML29vfzSL/0Sd9xxhxmbKEMW6vU68XjclB1J9rXEZb2Ctbi42ObClWQv7yxluZDwToqqVCrcf//9xONx+vv7SaVSjI6O8rrXvY5kMsmnP/1ptm3bxiOPPMIVV1xBNpul1Wpx1VVXkc/nKRaLBINB04EsFosRCARIp9PUajWy2ayxsqXzmJxzuTjyhg9O1N3cGW7oFOX1gI0hnxxuuAFuuw22bAGlnJ+33WYTuixrh7M+huy1gpfb7n1MXL4SAy2VSia+KnN3ARKJhOllHQqFTHZzPB43bmKxpqVVpbTBlKlH0t/Z23f6a1/7mqnBbbVaRuhlVrE0+JAuXOVy2dQOS8azNBDxjiwUN7s3VivPEYGXJLZXv/rV3H333Vx//fXk83k2btzIhg0b8Pl8XHXVVTz++ONs2bKFPXv28Pu///sMDAzwzDPPEIlEWFpaQillxlaK1Vur1Th8+DBbt24lnU63lZPJe5f39kJjv/K52hjy2ckNN1gBtqxdznpBhmNbyPIFLjFYeb6UDklGtHTrKpVKphezN5YaCARM/2lx54o4KaVMVy1JthKXaqVS4bzzzmNiYsKsU5K3xB0tYu4dNSiDK6SFpnTzkhjz4OAgk5OTbedApiN5s66lKYp0KNuzZ4857ve//30uueQSbrrpJubn5/nwhz9MNps1Qzjy+TxjY2NtbUO9zUokPi3JbfF4nGKx2CbInRby8crUlqPzs/V6CNYR+9ybxJAtFssZhnVZw4qxRK+bVMRK3MJiYcZiMdMEQ1plSjlRZ5mNCEAikTBxVLFec7mcSdqSMYk+n4/e3l62b99u5gZLty6JX3vbbnq7iUk9sVKKrq4uM3FKYuHSolPET6xV2Ye4vaWUS2LW4+PjDA0Ncc899/CKV7yCzZs3c99993H33Xfz7ne/m1gsRl9fHzt37mybliVrqdfrxlIOBoOEQiHTc1usYzlfkkAHywvwSoJ6IkK7nlzWAJ6Y8ceBj9sYssVy5nHWW8idWbreL32xjkWMJEYsLR4l8ahSqdDb28vc3Jx5jrc5hwxDkKQqSXoSsZIJRL29vQCk02kz8hDgn//5nykWi229qMPhsCk1knivJHl5Lw4CgQAHDx40tdBi7csYxEAgQLVabYvPygWH3PfWE7daLfbu3UulUiEUCvGDH/yAmZkZ/H4/X/3qV+nq6uI///M/qdfrDA8PUywWyefzxhMgGeqFQoFMJkOj0aBYLJo1y4QqEWTxFIh4LhdPXu7zPJ4F7e2ith5QSl0E/B3Q7d6fA35Ba71nVRdmsVhOGme9IMuX9nJf3p0WstddKq7lpaUlWq1WW1YyYGp4wSn1ETe1WM3pdJpms8nc3BzVarUt2Skej5uWmBLzBYybO5lMGrGSRDApg2o2m8bS9vv9VKtV4vG4sU7D4TCBQIBkMkmlUjHuZK+VLaVcYsWm02lyuRzNZtNYs41Gg7GxMY4cOWJaXc7NzVEsFunp6WFsbMzE0L0XHlKf7ff7yeVypvGItN70WsReIfb5fMvG+o9lDa/DOPGxuA34La313WCmn30WuGI1F2WxWE4e1mXt0vnF7hVq7/ADybIWcZDWjzL4XmK53iQpKXOKRqOEQiFSqRSNRoP5+XnTSMOLCK4kaJXLZSNi6XTaWMfe7ldeF6+3r3YoFDJTmqSuuNFomFnAxWLRlF0BxoUtGeEywlBaesr+JIYdjUYBpyNZPB4nmXTCm93d3YBjkUtTEhFXbzLX1NSUmbUsrv6+vr62eL0k13k/l2N9dif62DoT67iIMTjDVoD46i3HYrGcbKwgnyDeWLDEkaUblTfuKVYmHLVoAePaFoEVUSuVSvj9fjKZDMlkkmq1alzNfr/fiLjP5zNu31Kp1NYZSxqEVKtVI/7yuLjbZR/iFhbRlQYhYrmKRevt5CXrkQsFKbXyXhTIPOd0Os25557bZt3L/GQZjCHnzpvo5vf7TWcyeR9yTsVSXqkOebnPap2J7YmwXyn1B0qpEff2AZzMa4vFcoZw1rusYWVLyfvFLgImQim9qKXcSEYnfvey79KIPHuKUKAS4Ip7nPpcaY/p9/uJx+NGaPn/2Xvz8Eiu8tD792paarXU2mczs3jGZGIyNmDsAWOSSxy2OGyGkAUCN5DwZL4kELLcJDjxFxIIcOGSeyEEQj4ngOGLIXFYTXDYDA4BYmMDBu9jZ+xZPOPZtG/dWs79o+o9UyqV1Ec9LanVen/Po0dSdS2n6nTVW+8Ovm70xo0bfcqUph5t2rSJM2fOICK+SUOpVPI+3y1btvia2JqjnH5JUIGarCGthT/grADUKHGNutYGE2qG7uvr49ixY17AalERDXbr6+vzZUHVfK+FQVTQa2Wzo0ePcsEFF/jiJ+oGSM5BMne4krBN5x5XWmeN8OvAW4HPAA74j3iZYRgNgmnILP5wTlaNSja9h7OpTFp9S0QyhTHAdGtUx7q5uZlyuUxXVxfNzc10d3d7wa5CS9OoNP9XI6lVaKqJVzVJjZ4eGxvz/m01Y6uw1LaGpVLJB3WpwN66dasXqHpMzU/Wcx0cHPQFQwBGRkbo6Ojw7RpbW1vJ5/O0t7dTLpcplUp0dHR4rTeXy9HW1sYTnvAEP2Z9EVAzuzaRSFbpWkrgVbJoSCWhvVY0aBFpFZFNzrkB59ybnHOXOucuA94JTFTa3jCMtUPdC2QR6RaRT4nIAyJyv4hcISK9IvJVEXko/t0Trysi8n4ReVhEfiQil4YcY6GI3Xifc4KD1ISr/lznHENDQ95HW4nJyUkf4KWm4Hw+7wWsaqSqParZur29nYGBAW8CVqGrfuOWlhampqbo6Oggl8v5KOtyuUyxWKSvr89r01pZTIPFRkdHfdS4jlH9vOqjhuhFQE3gWi7UOUdfXx+dnZ3eb97d3e1bSeq6Gj2tudDT09N+G/VHJ1spZvWJ1usSIkwbKKDr/cB/y1j+k8B7V3gshmEsI3UvkIG/Br7knHsS8FTgfuAa4Bbn3B7glvh/gJ8D9sQ/+4EPVdr5YlHWujxZREIjj1Wr02CrEBOoCpru7m7y+bxP+0lX+ZqZmaFUKvlKYM3Nzd6/q8dWbVpNxhoANjAwQEdHhw+GyuVytLe3c+LECS/4NBI7mS89MDDApk2bvCBXgZYMEkt2pFIftbaJVD9xa2srbW1tbNmyhdnZWbZu3epTvNrb2/1Lgp57sVj05mgtiALM0eCzTNULFfZIz2cDBHVd5pz7THqhc+6zwLNXYTx1zw03wK5d0NQU/b7hhtUekWGEUdc+ZBHpInrovA7AOVcGyiJyNXBlvNrHgFuBNwNXAx930ZP2tli7Ps85d5wqSD7ck1qbaplTU1McO3aM/v5+r7UuhgZCbdq0iZMnTzI1NeVTj9T0m8vlGB0d5bHHHqO9vZ18Pg/Ali1bfIEOEfER0zo2FcpNTU0+jUrNzKdOnaJYLPq+yS960aWcOdM8b3y9vWW+850mHnvsMR9N7ZybE0CmpmYV5vqZpil1dHQwPDxMT0+Pr1i2detWXzJ0ZGSEQqHgq52pL7mpqYlDhw6xd+9eurq6OHr06Jzc47RmvJhVQ8eTnMM1TNsin62FF+oV5YYbYP/+s32PDx2K/gcrmWnUP/V+Q+8GTgEfFZEfiMg/iEg7sCUhZB8HtsR/bwOOJLY/Gi+bg4jsF5E7ReTO8XTH8gVQLVj9yF1dXV7L1OCoSg9/1SyPHz/ui2Fs3LjR+401j3l0dJRyuezNu/l83mvMKog0+EsraiX7GE9OTnLy5ElfrUv93KVSiZaWlkxhDNDf38LJkyfZvHkzgPcFJ33o2glKfdH6YqA51EkNXrX9mZkZxsbGfC6zrq9tHnO5HGNjY96kntSG0y0T05rvQgFcixUPWWOcFJFnpBeKyNOJ7g0jwbXXnhXGyvh4tNwwQrnh7hvY9b5dNL21iV3v28UNd6+MmaXeBXIOuBT4kHPuaURF9a9JrhBrw0tSg5xz1znn9jnn9mk1rIUe3MmHfNLHOTU1RVdXly9zqUKkpZzdkCdXyvnt1F+q2rBqx0eOHOHMmTOcOXOGqakpzpw547Xc5HG0GEiy7rRGU6uA1CIkGiFdLBa9z3oxCoUCR48e9eZrfRHRADGt7qUlL5MBVFqnu6OjwwtQPZ4KaS0ios059CVCm1kkzwvOVhZLz0WakPlbo/wRcKOI/IWIvCT+eStwY/yZkeDw4aUtN4w0N9x9A/u/sJ9DQ4dwOA4NHWL/F/aviFCua5M1kYZ71Dl3e/z/p4gE8gk1RYvIecDJ+PPHgB2J7bfHyxYkHbS10DqA98FqsNWjjz5Ka2srQ0NDPk/4yjuvpLm5mZGREVpaWhgeHgaial0aaa15wOqvPX36NNPT04yMjDA1NcUVV1zhBZYGUk1OTtLd3e3TpbThg5qptWOURklrtS2tea0+2m3b5hkM5qDrPfzww+zevZvbb48u/YYNGxgZGWHTpk2cOnXKF/ZQAar50SqEC4WCr7zV1NTkq4Ulu1+1trZSKpXo7u725TWHhoZwcWMMrXqWjppezHfcaDjnvhtryG8gdt0A9wKXO+dOLrhhAhG5iigWYwPwD865dy2w3iuI7rGnO+fuPNexrwY7d0Zm6qzlhhHCtbdcy/jUXDPL+NQ4195yLa9+8vL6PepaIDvnHheRIyJyoXPuQeC5wH3xz2uBd8W/Px9vchPwRhH5J+ByYCjEfxya16opON3d3czOznL48GGvLapfVYOf1D+6bds2RkdHGRgY8JHJra2tbNu2zWuF+XyeM2fOeEEGUXOIrVu3IiL09/fT3d3N8PCwN/NqgQ/1a6uvWUR8PWs19aqZuFQqce+995IdtBuR1LTvuusuLrnkEr773e96wa/CvVQq+X7LcFa4qparwWQTExN0dnbO6Sylvm81qWuLyaamJoaGhsjlcmzcuJGTJ0/O0Y41inu9EQveP69mWxHZAHwQeD7RC+4dInKTc+6+1HodwO8Ct8/fy9rhHe+Y60MGaGuLlhtGCIeHss0pCy2vJXUtkGN+B7hBoubsB4FfIzK13ygirwcOAb8Ur3sz8ELgYWA8XrcmqDaqAkH9o8k2i0mfqmq0x44dI5fL0d3dTU9PD52dnXPqNRcKBcbGxigWi7S0tDA4OEhPTw+9vb0+t1fN3JqGNDo66iOWVVBq+0LVijU3GvAlPjVNajHUVKwm77vuuos9e/Zw4MABCoUC/f39AHR2dnqTswpVFZ6a76ymcx3v0NCQN7kDXhBrFPbMzAyDg4Ns27aNxx9/3BcZSZYuNZbMM4CHnXMHAeKX1auJXmqT/CXwbta4GVwDt669NjJT79wZCWML6DJC2dm1k0ND880sO7uW38xS9wLZOXcXsC/jo+dmrOuITHtLotKDPqlpanS1+lSTfYQ1Ent6epq2tjYGBwd58pOf7P236n/VaGz10fb29lIulxkeHqazs5MtW7ZQLpe9FpkMdtJjau5y0pybbEShZm2N4tb9Oefo7S3T3z/f193bW/Z5v5ri1NTUxCOPPOLzkVUrV3Py7OysL79ZLBYZHx/3pvjx8XE6OzsZGhqiv7+frq4uxsbGvD9co6s1yEurhxUKBV8lLKnpJ+dqobQnYx5ZgY6XJ1eI8/V3OOe+KCJrWiBDJHxNABvV8o7nvoP9X9g/x2zd1tzGO567/GaWuhfIy01IviqcDejS4Kuenh4vfEZHRwF86cfh4WH27t3rzc99fX1z8pVVqMPcQiHqZ9VcXq2epfspl8s+H1mF2YYNG7yA0/PQyOVkdLKahGdmZrj55h/4ZVrnemBgAICpqahgyNDQkE990jrWp0+fnpMznCzyoWZujeju7Oz0tbq1+5T6kLVutda71pxk9TsPDAx4jTupIRu1R0SagP/DWf/0YuvuJ8rvZ6c5ZY0GRf3E195yLYeHDrOzayfveO47lt1/DCaQK+arasCXRvt2dHQwOjrKmTNn/Dr5fJ7BwUHa29spFos89alP9YU/koFWhULB+2A1YjpZMhOY01RBzdFa+1qFHETCf2RkxLdt1CIaWm9a85VVY9YgLNVctXjIzMwMmvo1NjZGU1MTw8PDTE1N+YC0crlMX18f+/bt49vf/ra/XmqqT3aVUq1ctezZ2VmKxSLDw8NzzlvXV6Hc2trqtW5tlFEsFv2y9YqI3LTY5865l1bYRaVAxw7gYuDW+Du4FbhJRF6aDuxyzl1H1AaSffv2NWYUnWEQCeWVEMBp1r1AhvCgLsALCg3eGhkZ8Z2bfuInfoItW7bQ29vrWzKqqVr9uKrJJtslqml2w4YNtLS0eCGnTRi6urp8wQ1NndLPteCGtoFUX6zm4KqgU404n897zTpp7k76s2dnZ32DB9VOh4eHeeCBB3yEuQpfDRjTylqa76xm9kKhwMDAAG1tbT6/WgPM1ASutb0HBgbYsGEDJ0+e9NdDx5Yea6hlowG4gsjk/EmigKulnvAdwB4R2U0kiF8J/Ip+6JwbAjbq/yJyK/CHazXK2jDWMiaQWfihvlBKlGqsKvRmZ2fZt28fGzduZMuWLeTzed+zWLVBjRDWzkfqawZ8VHSyUUShUPD71iAp1S61fGXSHF0qlebUetZa1WoaVx/w2NgY7e3tmZqnCnm1Buixent7OX36NCLChRdeyAMPPODXTUZ6qx9dK5JpZa6kH7ivr48TJ07MKckJcPLkSe/v1rHri0XyvNLzsQ4E8laiCOlXEQnSLwKfdM7dG7Kxc25aRN4IfJko7ekjzrl7ReRtwJ3OuUU1cMMwVg4TyIEkq0dNTk7y6KOP+taFl112mRfGra2tPjVJA7A0mrijo8NriIqWntR8XTX/6jqaGtXU1ORNudrbeGJiwgswFeIq1FVQaXCWCk0VhCrc1AcMeN+2asXqDz516pQveXngwAFGRkZ8XnJbW9ucYid6rlNTU17j1Trd09PTnDp1iunpadrb2xkcHMQ556O1S6USXV1dPuJ8YmJi3svQOtKMAXDOzQBfAr4kInkiwXyriLzVOfeBwH3cTJSBkFz2lgXWvfLcRmwYRrXUe6WuZafSAz7pQ3bO+XSjRx55hMHBQXbt2kVvby+bNm3ygUutra00Nzf7AKZkJLGmMqkA1aIZKswKhQLFYtEHbyV9zB0dHfT19fl9aZGNZPENFYzd3d1+DKqha9EONTvrcdPpUPpZsVhkz549PmhsdHSUoaEhxsfHfQQ14LdXwTw9HRVA0QA2fRnQl4l8Ps/p06d9PjXA448/7k3/6RKYi81boxYESSIieRH5eeAfibII3g98dnVHZRhGrVn3GnI6qEsFcFpAqylZlw8PDzM4OMiznvUsuru7fX6tluJUAa75uOniIi0tLd6HrA0cVPOdnZ31gkojptVHqy8FWsFL99Pc3Mzw8LAPAtPULBHxPYmdc4yOjvqCHPpSAGe7K+nfGow2OjpKb2/vvJQtPYb6wCcnJxkbG/PCOp/PMzExwcDAgC+YAvgCKupv1gA51ao19Uu1/KamJt7+9rfzzne+c04g3XrRkEXk40RBVzcDb3XO3bPKQzIMY5lY9wI5XZZxIY1L/akTExO+VOaOHTvo7+9n9+7dFItFv60KYPXfqhDO5/Nz8mlVwCZfCrQ5hAropDBT7VfbJ6ofWgWylpwEvPas+9ccZV1P/btJC4D6gFWA6/lt3ryZwcFBJicnGRgY8IFq6ifO5XL09vbS1tbGyMgIXV1dbNiwgdOnT1MqlRgbG6O/v39Ow4hcLsfRo0e9P7xUKvnrAZFf/fTp0/zRH/2RLxyyXoRwitcQ1XD/XeBNiWsgRKn3nas1MMMwasu6N1mHFJhQgdXa2srw8DAnT56kWCyydetWxsbGeOCBBxgdHfX+WC0tmQ7UUuGvKUqqwSajmVVA6XoarKXaqZakVFMzQHt7uw/C0n2qME62N1SBrilRGi2tQWA6Rj3XiYkJLywnJiZ8WpT6fVUrnpyc5MyZM0xOTuKc81Hlra2t7NmzhwsuuICenh6Gh4cZHx9neHiYAwcO+DHreLTblZLL5di6dSutra284hWvyAywa3Qh7Zxrcs51xD+diZ8OE8aG0Visew0Zzhb9ADIf+nBWKKuvd/fu3RQKBR98dPDgQQqFgtdcVZOdmpryZmEVtio0tZhGssiHRjhrswYV5jpGDcDSiG3A+5vb2tp8sQ41oSfrP6tGnewGBXiBr9HMTU1N9PX1cerUKa9JJ9Or1OSsZm7V5tXMrmbx1tZWyuUyPT09XHTRRb4+tlY8Sx4vqT3rNQHo7+/n1KlTfPrTn55nTVjIvdBIiMgXgU8An3POre+kbMNocEwgc9aknEUytUl/2traKBaLPopatc5HHnmEXC7Hjh07fB9jzbXVQCqNQlbztJqck7Wn8/n8nEIZcFaYqtBWYavHUQGnZl/N9U0GVWlEd1Jr1jGqhq9VtoaHh8nn84yOjjIyMuLrUKtAbm5u9sJYBbKazHVcGmk9Pj7OyMgIxWKRZz3rWRw4cIAzZ874a6vnnn5JGBsb421vexubN2/2XaCyfPtpkssaIOjrOqLc4feJyDeI8pG/6Jwrr+6wDMOoNSaQK5CsowyRYNy4caOPoNYykprzm8/nOXLkiDcxNzU10dXV5dOYNGhLhaMKTC3ooRHRavJN1tBWX7LE/YlVQ9aOU1rxSj9XbVxfCFRwDw8P+2AyiGpgq+96eHjYd6hSE7uWxBweHqa9vZ3JyUmmpqZ8VLRqulr+M2nu1gCyqakpenp6GBkZ4cILL2RoaIj777/fa+cQCeGmpiYOHTrEj//4j/uXi0OHDnntW1ksKj79/1oWys65zwOfF5E24CXArwIfEpF/Az7hnPvqqg7QMIyase59yJVIp9eoBqgpSc3Nzd5MrWZfNVmrsJ2YmGB6epru7m6uvPJKXvDcrHBgAAAgAElEQVSCF/h8ZRVo6tMFfAEQfRHQnGMNoErmHGueskZzq0k7GbClqEDXv1VQ6rFmZ2fZvn07k5OTc4R+W1sb5513HiLCxRdf7MegLwRabUtfJDRtSk3jEPWDbm1tpbu7m0KhQGdnJ5deeikbN27040lqwMlWlCFa8WLLGwHn3Lhz7p+dcy8HXgBcQpSfbBhGg2AaMmdbK2ahWqKarVXA6HYqeJImZdVYVTjlcjm6urq46KKL2LRpE0eOHOGKK67gW9/6ls9bVoHU3t5OoVDwKUzqa9aqWOo7VpO3VsXSsSWFaVIA6/iSgl9rX6vZWlOYklq8FjQZGRlh8+bNPPbYY2zevJlyucxVV11FZ2cnra2tfO5zn+PYsWO+qIeOcXR01GvqWj5TS4+KCLt376alpYVTp075lxlN69Jr393dzalTp/w8JH/WCyKyhajN6CuB84AbCWgIYRjG2sE0ZM7mDKdRgZCMhlYTLpw1NyeFHDBHUCYLcQC+b7EuT2uEWqdax6OCuqWlhWKxSEdHh9esVWgmx6mmbTUV63iSpnD1UaupXKOdVVDqsTXPWI/9zGc+079c7Ny5k2KxyM6dO9m1axcvfOEL2b59+5x63RpopsVJCoUCbW1tvsa1Bqpt3ryZnp6eOYVEtEmGfg6sS2EsIr8hIl8Hvg/sAf7IOXeBc+4aoHV1R2cYRi0xDZm5UdZKMopX/1aNVTVhFTSqiSZNtJqipGbf6elpX6Eqn8/zne98Z14ZzdnZWTo6Opienva+YPVL6zH1+PqSAGejpDXATDXudGtE9SG3tLQwNjbmS3Smuy2pcFetWYuNjIyMcPHFFzMxMUGhUKCvr8+/aPT09LBp0yY6OzuZnp5mZGTE1/RWX7MGxMFZM7e+uPT29tLf3z+nA5a+HGjUdVoQp4Vyg0ZcXwH8T+AW51w6+vBfAOuDaBgNQt1ryCLy+yJyr4jcIyKfFJFWEdktIreLyMMi8s8i0hKvm4//fzj+fFfIMSr121VhoO0TtT9wfEzfblCFdjJgSoWZc47bb7+dH/7whxw/fhzAC0T1Oeu2Wn5S06TU7J3UZpNFOZJVxnK5HG1tbRQKBR8Nrc0nNmzY4AOu9NhtbW0+UEuPr4If8PnJu3bt8kFs2rf46NGjPrhLx6r70JQwzUvWaGzVgvV89OUhn8+zbdu2OdYGDah76KGH0t+JRV0MjYRz7tedc1/NEMaw9M5PhmHUMXUtkEVkG/AmYJ9z7mKibjWvBN4NvNc592PAAPD6eJPXAwPx8vfG61UkS0NOP/BVaORyOS6//HJ+9md/lhe/+MVzOhsBXmACXiPM5/NeyJbLZY4cOeLLZCpqwtZ0pObmZh9QpUJGWxaqpq4CSzVxEaGrq8sLbM1LVm032ftYi4Ro+pYKaO2lnMwLbmpq4vDhw77QSblcplgsMjAwwDe+8Q2+8pWv8IMf/IDzzz/fB3ol63nruWmgm744qFldfcwqxNXkPTw87LV1XdaAGvC50FhvH4axzlkLJuscUBCRKaANOA48h7M9XT8G/AXwIeDq+G+ATwEfEBFxi6hN6hteDN28XC7T2trK9u3bOXDgADt27PCCLen/1KjlDRs28IY3vILh4cK8ffb0lPjkJ2/1plwVOmNjY3PycdPpPurXTVbfam1t9eUzterV9PQ0z3nOXvr7mzOOXeaLX/yeP7aaqpP5yvqyMDs767VnzUNWjVstAjrOxx9/nO7ubl9MRK9b8rpo3rVeM8191haUGiQ2NDTkj6952h0dHf6FotJ8NYrgFpEvkC14Behb4eEYhrGM1LVAds49JiJ/BRwGJoCvAN8DBp1z0/FqR4Ft8d/biJq5ax/YIaKH1unkfkVkP7AfoLu7m+7u7uQxM8cyOzvL6OgoMzMzHD16lMsvv5yRkRGe+MQn0t/fTz6f90JITbUzMzOZwhhgYCDvo4lVmCd7ACe7MKVrbas/WM3QKrTK5bJvhzg7O5spjKNjt3jT8sjIyJziJeoD17rWGzdu5PTp0/Nyscvl8hxTe9LMruZy1bQ1klxre6tZXoPSVIMul8t0d3f7QiHJADPAv6wko8cr0QAm7L+q8jPDMNYYdS2QRaSHSOvdDQwSBbFcda77dc5dR1QBie3bty+oQCeDulRTPHXqFKOjowB87WtfY8OGDXR3d1MulznvvPMAfJ/iZOR1FirIVZAm05K0dGZSYE1OTnphpqbyUqnkNWTVRLXy1mK0tLQwPj7uBfj09DTj4+NeuIsI7e3tnDx50v+v+1d/sWrr2nVKg6/UnJ0U3Ek/t75wJLfV8ep5qjBO5yerTztgjr2veS0LZefcv6/2GAzDWBnq2ocMPA94xDl3yjk3BXwG+EmgW0T0ZWI78Fj892PADoD48y7gDBVY7AGvD/VkxO+jjz7KjTfe6AXHxMQEs7OzHDt2jMHBQbq7u3niE5/oC1tUOu75558PwPj4uDd3q09Vfa+6vmqVyTrQyUYSWnNao5kXOy/1GWurRK02pgJauz5pEJaif+sLR7o5xvHjx70vPVmnulAo+MYYGgSn0eDNzc0Ui0X/2ejo6Jz9JlnLAnapiMjVIvKGxP+3i8jB+OcXVnNshmHUlnoXyIeBZ4pIm0Qq0nOB+4BvAPowei3w+fjvm+L/iT//+mL+YwgLEkqmFwE+8ElzhnWdfD7vI5aPHTvmBdpCqDZ74sSJOVHGeizVOFW4JQVjspSm+mNbW1uZnJz0rRIrnZOmP6mmDFAsFsnlcjzlKU/xZmU4K3STTS5U802akVWTTkZla/crDTxLBnm1t7f79LGkP1rN32mTfTLCutLcNYjg/mOi77WSB54OXAn81moMyDCM5aGuBbJz7nai4KzvA3cTjfc64M3AH4jIw0Q+4g/Hm3wY6IuX/wFwTchxKjWXSOcka4MFYE4EdbKYR6lU8j7ThUgW7nDO+e5Kzc3N7N6923dx0kpa6XaNSZ+v+oOBOYVFFqJcLjMxMUGpVPLCUn82bNjAXXfd5U3nyajrZFCW+nq17nYyT7q3t9ebtJOBc+r31sj0dGlPTSfTMSX7R2fNzTqgxTl3JPH/t5xzZ5xzh4H21RqUYRi1p659yADOuT8H/jy1+CDwjIx1J4FfXOoxFtO0khqZCsWZmRlfYlKDn9Snm0yDiqKGxxkenm8+7u6enFP4Y3Bw0OcAl0olDh065POXu7q6fEtDwAeNqXk3+TKgTSqmpqbo7Z3KDOzq7S37+tjJtou6z/PPP5+DBw964ZksLarCWc3R+mJQKpW8D1z9wHp+qllrmU5dV48NZ1+Kkn7spD97HdOT/Mc598bEv5tWeCyGYSwjdS+Ql5vFgn7SplEVWFpFK+lb1lQl1TDL5TLNzc28+90f94JWA722bdvGmTNngFafC5xsrQj4IiDNzc2MjY1RKpV8BauWlhZ6e3t9wJWasbW7lJbK/PrX7/M5yVqPWoXn7KxQLBa90NVo65mZGQ4fPuxfLABfYEQDxyYnJ+dU0NIXBcC/DBSLRX/tkqZvNcUXCgUfyDUzM8OmTZsYGxtjZGRkjkDOmo+lzuMa53YR+Q3n3N8nF4rI/wN8d5XGZBjGMrDuBfJCGthCZRlnZ2d9a8Fk20Bt7KBCTCtsqeBS7XlmZobTp0+zdetWBgYGcM75QCzVNlXrVW1XI5JVAI6NjXntVs3oqjUno62T/ZLVfN7c3MzExIQ3M6vw3rhxo9dc+/v7KRQK7Nixg/vvv9/7dnXf+vKh4yyVSv466QtGqVSit7fXR4rr+Y2OjtLe3k5nZyeHDx9m7969AGzbto3777+fb33rW76YSaX88Kz5SfubG4DfBz4nIr9C5LoBuIzIl/yyVRuVYRg1Z90L5MXKMCrJwCIVEsmqVl1dXb4e9OzsrC/OoaZeFcYqAKenpzl27Bhbt26lv79/XhS1NoBQLVGbLCQFjq6rY9LUKY3MVl9uS0sLjz/+uG8Tefr0aV8hS33Qp06d8sfQSlsaZKWa+vnnn88jjzziTfTafSrZAnJ6epqBgQH/oy8B2pqxXC77Hw1kO3z4MPv37+f6669nz549vjJZ0u9cjWBtEGGMc+4k8CwReQ5wUbz4i865r6/isAzDWAbqOqhrJckydyaFtQqJcrnszbwqcJKanK6nGp5GSKtA0vWdc14T1ahiLc+pWrUK6qRpXAPAkn5j9TsDvvykavCazgR4TVuDw9TPq5q3jlf9y0eOHPHavVbI2rt3L21tbfT09HjhOTk5SalU8i0Tn/WsZ81Je9KXh2QKVC6XY2hoiMHBQW677TYuuOACDh065F861FednpeFBO1CJutGMWM7577unPub+MeEsWE0ICaQA1FtVU3BKgDVL6wCTlOh0sFKScGXLHpRKBR8jWqtNZ0uKJI0ZetY1KeraUadnZ0+GltN4/oSoJqyaqeasqXHKxaLOOfYsWOHH5sWG1Gz+eDgID09PV6gaxS0jkFfEFpaWrjjjjvmBL+pX11fcNJ+6FtvvZUTJ05w77338tBDD/mXg4XSnZIkl6VfoCwgzDCMtcS6N1knWUibSpqr1TSrQgbwwk39uCq4NYJaBVCxWPSFPVTrnZycZOPGjd6fDGfToLSGtL4EqLk7KYzhrLAfGxujUCh437UKPy1eAswRtNrEQoXro48+ypYtWzhx4oRPncrlcvT09HD69GlGR0e9wOzq6kJEOHr0KNPT07S1tfnCJrlcjre+9bcYHZ2fldPePsKv/dqfehO5Bq396Ec/8mlbydaLlapyJeesUapzGYaxPln3AnkpGpQKEdVyOzo6fJMF1SQBX5dZTbqlUsn7ZlXAqEBvb29nYmLC+5tVmCQjrnXdpNlXA7f0ZSCpNbe0tHhztBYA0SjvZG9hHa8K5OnpaR5//HEuvfRS7rnnHpqamhgeHmZsbGzOC0gul+Po0aNea08WJZmcnGR6ejpTGAOMjXXMMbur2X9iYsK3cawm3UkFcFIoG4ZhrCXMZM3CpTOTD3XVfJMdj9TsrGUmdXlSm00KVuecj3BO5vVqjnGyoYQKKhW+SX9qMv9Yj5NMvVIfd7KKmHZwUhO2mpJVqBaLRfL5PMVikQMHDuCcY/v27b6PsbZFVCGq/vSWlhae/OQn+1rfmuK1GCrE1ZKQ1JaTjSNCK3LpOskXmiz/s2EYRj1jApnFK3UlfZJJjbVUKvmHvgqipEDR3yq8dF0tjDE1NeVTiDSaOSmQ1aesPls1c7e0tMzpmaxCp7293ZuuR0dHKZVKPpjLOcfY2Nic7kxqyn7sscd8BHhzczOjo6O+A9XBgwf99m1tbVxwwQVcdtllbN261Wvlam6emJjg0ksv9S8si6HXNOlfV3+2XuelCGMdY3Kb5LwZhmGsBUwgQ0WNTkkHGsHZyGU4K4RVWKrgS/Y8Tppjky0GNY0o3eZQq1+pRpvsv6zmal13YmLCB21pMRHVipNNKuCs5r1lyxaam5s5ceIEY2NjvojH1NQUO3bs8Cbw8fFxDh8+zL333svp06fnBHNpLvLdd9/N5ZdfXtHvm0zvGh0dnfNiU6lcZsj8JAO6zHRtGMZaYd0L5KQvdTGSD3oNdtKcXBWQqtGqwNGfpPlZ6zsnTbUwV5iomTjZohDOvjgkq2SJCH19fd4UrsFbSfN5Pp/32m97e7s3TWsutTZ5SKZZlctlHn74YXp7e5mZmfE+YhXAuVzO+6aTUeO33XYbl1xyyaLXMqoUNustDYAPVtPzS5brDJ3HheZsvSMiV4nIgyLysIjMq+8uIn8gIveJyI9E5BYROX81xmkY6511L5DhbDvBSqjGpVW4ND9XhYuaptNtA5NCX9dVklq3BmqplqspTHo8LeaR3F9vby9nzpzx5Sg191d90irci8UixWKRjo4OWltbGRwc9C8WuVyO1tZWX7gk2QDizJkz5HI5H6CmRUW0jKea0VVAO+d44IEH6OgYz7yGhcKwN9mrEFYfuPqVkwFuSRYSsMn85qw5W8+IyAbgg8DPAXuBV4nI3tRqPwD2OeeeQtTM5X+t7CgNwwCLsgYW9yEraZ9ksuOT7kOFsZqqk9upJqlCK2lynp6e9v5TNWVrMJhGb2szi2QUdrFYZHh42LdlTGuumj6k5HI5jhw5QkdHx5zqYL29vZTLZZ+21N4eRUhrhS2twOWcY8uWLRw7dsyfb1KDV419amqKb3zjfr72ta8xNTXltWA1qU9O5rymrcdRVEtOB9QtFHWdrOi10Jytc54BPOycOwggIv8EXE3UxhQA59w3EuvfBrxmRUdoGAZgGrInpMpTMi822f9Xc3ZVGCU/h7PCOLmtpjrBWQ1dzb5aXES1adV4df8tLS0+XUrTrkqlEq2trRQKBfL5PK2treTzed/C8fTp0wwPD5PP573wV0E+OjpKa2srLS0tbN261Y+jubmZtrY28vm8r4l95MgRZmZm2LZtG8Vi0b9AJKt9lctlbr31Vnbu3OmviwrlpIaf9CWrhp401ydraKd/9NovZpI2kzUA24Bk+8aj8bKFeD3wb1kfiMh+EblTRO48derUoge94QbYtQuamqLfN9ywtEEbxnpk3QtkffBnmUeTv5OoZqdCC86apVW4quaW1OCSwUp6zGRLwrT5Ojmu5GeqhSdLUzrnfIepfD7v61rPzMwwMjLiXwo0ACtZNSyXy3ntOJlXrVq+muOTmuyRI0d8sNj5558/J5K8paWFlpYW77NWU/vU1BTj4+Nea9brkHy5UQ1/IYGcnB8TtrVFRF4D7APek/W5c+4659w+59y+TZsW7vx4ww2wfz8cOgTORb/37zehbBiVWPcCGfB5vkkqaWQqIJOR1YoK3qSwUW1UhVByvWQnJ/1cG1KoANVjqbDS1CYVpJoOpcdQP3SyI5NWBtMXCt1ex9HW1sbk5KQPREsWOtFgNU3P0vzlqakpHnroIaamprj00kt96peI0N/fP8c/n0y5amtr8y8c6rvW4+Tz+XlpT8kgOcszXhKPATsS/2+Pl81BRJ4HXAu81DlXSn++FK69FsZTIQTj49FywzAWpi4Esoh8REROisg9iWW9IvJVEXko/t0TLxcReX8cMfojEbk0sc1r4/UfEpHXBh7bRwknSQZnJf3DycCtZGEODXBSIajm7XK57P3BqqEmBfzQ0JDfjwqvZKERjboulUo0NzfT1dXlhXcy91nNy5pKBGcrcCXLZBaLRW9iVs1ax6qmbDWZJ83QmkOd1Ph1WVtbGx0dHfzwhz+kqamJZz/72V6b1nElXzbU/67COFnxS83zhULBC/CLLrpozlyYUF4SdwB7RGS3iLQArwRuSq4gIk8D/j8iYXzyXA94+PDSlhuGEVEXAhm4Hrgqtewa4Bbn3B7glvh/iKJF98Q/+4EPQSTAgT8HLicKZPlzFeKLsVjgT7qEo2q5ahbW4Klk0JF+ribnZPMJYE4xj9nZWfr6+nwvZRVI6rNVnypETSy6uroYHh72PtdcLkehUGDjxo00NzczODjI1NQU3d3d/hgaId3S0sLMzAybN2+ms7OTvr4+n+7U2dnJ+Pi4176TJmuNtk6ehwrmYrFIW1sb7e3tXutta2vjBz/4Ae3t7fT29lIqlXj2s59Ne3u770LV0dHhfdPNzc20t7fT3t5OX18fL3nJS3ja057Gli1b/DHvu+8+byHIEsgmlBfGOTcNvBH4MnA/cKNz7l4ReZuIvDRe7T1AEfgXEblLRG5aYHdB7Ny5tOWGYUTURZS1c+6bIrIrtfhq4Mr4748BtwJvjpd/3EVP4dtEpFtEzovX/apzrh9ARL5KJOQ/udixVYhW+rytrY2+vj46Ozvp6urywqe1tZWpqSmKxSLt7e1zhI3+qIk62RYxqQlu2LCBnp4en0qkGq9qwZoHPDExgYjQ2dmJiNDT08PY2Jj35ba1tfkAqvb2dh+oNTU15Rs3aDqVRlNLXNxDfbcquDWNSjV/FXy9vb1MTEzMKQ+qAnlycpLW1lb27t3LXXfdxczMDLt27fIpVzqWyclJCoWCtzLomGdnZ3nwwQcZHR2lu7sbgDNnzjA2Nsb4+LgPnlss6tqYj3PuZuDm1LK3JP5+Xi2P9453RD7jpNm6rS1abhjGwtSFQF6ALc654/HfjwNb4r8XihoNjiYVkf1E2jWbN2+mu7t7Ts3pJK2trb7/7+bNm+nt7aVYLPp2h6pFtrS00N3dTaFQoLOz02vPzc3Nvga0CuhyueyF9/j4OL29vQwPD5PL5SgWi0xMTHgBrBpk2p+rNbA1Knl2dtZrwvqCoU0mtKa2NphQs7Nq8c45X2pTXwDUjF8sFr3pWgV0oVCYkwY2MzPjTeHFYpETJ06wadMmSqUSIyMjFAoFtm/fzvDwsE/h0lxkwBdLGRsbo6uri8nJSQYGBujs7PRR4ENDQ4yNjTE2NrbU75Gxwrz61dHva6+NzNQ7d0bCWJcbhpFNPQtkj3POiUjN7JLOueuA6wAuvPBCp4IyXYNZhWxPTw+bNm1i69at9Pb20tbW5jVi1WLz+TwdHR0Ui0UKhYLXhlUoqwDM5/M+f7ezs9ObcFUwasqSNnxwcR1p9bNqsJb6k1V7VqGphT3y+Tyzs7Neg1eTskZBJ6uEzczM0N3d7ZercFStWf3BaglIRl1r8JlWD9N0Ky31qb/PO+88nv70p/PII4/Q39/vg9E00jyXyzExMeEjvXt6ehgaGqK3t5dTp05x5swZBgcHOX36NIODgwvObZbWbCbtlefVrzYBbBhLpZ4F8gkROc85dzw2SWuwyUJRo49x1sSty2+tdBARobu722tmKqhEhCc84Qls376dTZs20dfXx44dO7zg6u7u9tqkVq7q7Oyko6MDiCKWAV8IJFnlSjXqyclJent7vdlaTcPaqlEFpeYsT0xMUCgUKBQKcwLDIPJHq7lZ05V0mQZMqfDUoC7VnJMFR3S7ZNpTMihLU62SUeBaQlSFam9vL3A2Pcw5x+7duxkaGmLnzp3Mzs7y7W9/2wfHqUau20xMTDA5OcnExARDQ0OcOHGC/v5+BgcHOX78OA8++KBv2pGMWk/OaZKkeTt93UzjNgyjXpB60R5iH/K/Oucujv9/D3DGOfcuierv9jrn/lhEXkQUpPJCogCu9zvnnhEHdX0P0Kjr7wOXqU95keOOAA8uxzmtATYCp1d7EKtA8rzPd84tnFRrZCIip4BDqz2OFI34fW7Ec4L1dV7Bz5i60JBF5JNE2u1GETlKFC39LuBGEXk90Y3/S/HqNxMJ44eBceDXAJxz/SLyl0RpHgBvqySMYx50zu2r1bmsJUTkzvV47uv1vGtJPb7ENOK8NuI5gZ3XQtSFQHbOvWqBj56bsa4D3rDAfj4CfKSGQzMMwzCMFaFe8pANwzAMY11jAjmOtl6nrNdzX6/n3eg04rw24jmBnVcmdRPUZRiGYRjrGdOQDcMwDKMOMIFsGIZhGHXAuhbIInKViDwYd466pvIWawsReVRE7o4bBtwZL1tyF621wGp2DDNWDxFpFZHvisgPReReEXnrao+plojIBhH5gYj862qPpVZkPZcagbivwqdE5AERuV9ErljqPtatQBaRDcAHibpH7QVeJSJ7V3dUy8LPOOcuSeTGLamL1hrielapY5ixqpSA5zjnngpcAlwlIs9c5THVkt8l6tLVaKSfS43AXwNfcs49CXgqVczbuhXIRA/ch51zB51zZeCfiDpJNTpXE3XPIv79ssTyj7uI2wDtorUmcM59E0gXglnquf4scccw59wAoB3DjDolnsPR+N/m+KchIlVFZDvwIuAfVnssxuKISBfwbODDAM65snNu4aL7C7CeBXJwd6g1jAO+IiLfiztcwdK7aK1llq1jmFE/xGbdu4jq3X/VOXf7ao+pRrwP+GNgfhu6tU3Wc2mtsxs4BXw0djH8g4i0L3Un61kgrwd+yjl3KZGJ9g0i8uzkh3HVs4bQJiqxns51veGcm3HOXULUUOYZInLxao/pXBGRFwMnnXPfW+2xLAOLPpfWKDmiPgofcs49DRjjrIssmPUskBfqGtUwOOcei3+fBD5LZKY/oabowC5aa5mlnmsjXoN1Q2wi/AaN4Wb4SeClIvIokTvtOSLyj6s7pNqwwHNprXMUOJqwznyKs42OglnPAvkOYI+I7BaRFuCVwE2rPKaaISLtItKhfwMvAO4hOkeNHn4t8Pn475uAX40jkJ8JDCXMvWuVpZ7rl4EXiEhPHMz1gniZUaeIyCYR6Y7/LgDPBx5Y3VGdO865P3HObXfO7SJ6Nn3dOfeaVR7WObPIc2lN45x7HDgiIhfGi54L3LfU/dRFc4nVwDk3LSJvJHrgbgA+4py7d5WHVUu2AJ+VqO9vDviEc+5LInIHS+iitVZY5Y5hxupxHvCxOGuiCbjROdcwKUINSOZzaXWHVDN+B7ghVvAOUsUz1EpnGoZhGEYdsJ5N1oZhGIZRN5hANgzDMIw6wASyYRiGYdQBJpANwzAMow4wgWwYhmEYdYAJ5BVGRLaLyOfjbkL/JSJ/HYfJZ637BBH5VMA+b9ZczCrG8xci8oehyw1jPSAiW0TkEyJyMC7x+J8i8vL4syuzui+JSLOIvCu+t78fb/Nz8WePish/pNa/SxLdyRLLd4nIRFyC8f64m9XrEp+/VKroTici31ni+r8pIr8a/329iPzCOWz/OhF5wlK2X4+s2zzk1UCi5LvPEJVXuzrOnbwOeAfwR6l1c865Y0DFm8A598LlGK9hrEfi+/RzwMecc78SLzsfeGmFTf+SKC/6YudcSUS2AD+d+LxDRHY4546IyE9U2Nd/xSUYEZELgM+IiDjnPuqcu4kqihg55561xPX/bqnHUOLnV3L71xEVADlW7T7XA6YhryzPASadcx+FqAYv8PvAr4tIW/wWeZOIfB24JX5Tvgcg/vxGEblPRD4rIreLyL74s0dFZGO8/v0i8vcS9Yb9Sly9CBH5DRG5Q6K+sZ8WkbbQQYvIrXZc+JAAACAASURBVCLyXhG5M97/00XkM7Em8PbEep+LtYl7JVE0XkReLyIH4jf9vxeRD8TLN8VjuSP++clzv8SGcc48BygnBYpz7pBz7m8W2iC+n34D+B3nXCne5oRz7sbEajcCvxz//SrgkyGDcc4dBP4AeFN8rNcl7qFfFJF74vv6m/Gyi+J77S6J+n3viZePxr+vFJF/jy11B2Ot/tXxNneLyBPj9Raynr0lvl/vEZHr4hcYfU68T6Iex7+r28ea9T6iohl3iciLRORzif09X0Q+G3ItGh0TyCvLRcCcYvHOuWHgMPBj8aJLgV9wzv10atvfBgacc3uBPwMuW+AYe4APOucuAgaBV8TLP+Oce3rcN/Z+4PVLHHs57l36d0QlKN8AXAy8TkT64nV+3Tl3GdHN9yYR6YvNVH8GPJOoPu+TEvv8a+C9zrmnx+O0NnNGPXAR8P0lbvNjwOH4fl6ITwM/H//9EuALS9j/95l77yhvAX42vq9Vg/9N4K/jhhv7iOosp3lqvN5PAP8d+HHn3DOI7sHfqTCWD8TPkouBAvDixGctzrl9zrn/rQucc58C7gReHY/pZuBJIrIpXuXXgI9UOOa6wARy/fHVBco1/hRRkXmcc/cAP1pg+0ecc3fFf38P2BX/fbGI/IeI3A28muihsxTURHY3cK9z7nisCRzkbEOGN4nID4Hb4mV7iArH/3vcY3gK+JfEPp8HfECi1nk3AZ0iUlziuAxjWRGRD8Ya6B2V116UM8CAiLyS6KV4fCnDWGD5t4HrReQ3iEoAA/wn8Kci8mbgfOfcRMZ2dyTu4f8CvhIvv5uzz4yF+JnYQnc3kTUh+Sz550onEnde+/+B10gU+3IF8G+VtlsPmEBeWe4jpdmKSCewk6iuMkRtu86FUuLvGc7GCVwPvNE592TgrUBrlfudTR1jFsiJyJVEAvaK+G39BwHHaAKe6Zy7JP7Zlmg2bxirxb0kOvU4595A1Cxg04JbRPfvzvh+Xox/Bj5IoLk6wdOIhPgcnHO/Cfy/RC/A3xORPufcJ4i05QngZhF5Tsb+0vdw8v5eMLZIRFqBvyWy4j0Z+Hvm3uehz6+PAq8hMt3/i3NuOnC7hsYE8spyC9AmZyMPNwD/G7jeOVfpbfnbxM0RRGQv8OQlHrsDOC4izUQacq3pIjKpj4vIk4hM1BA1avhpiToo5ThrQofordybx0TkkmUYl2Esla8DrSLyW4lli8ZcxPfvhwGfNRHHSPxiatXPAv+LJXQRE5FdwF8B83zYIvJE59ztzrm3AKeAHRIFgR10zr2fyL30lNBjBaDC93RszQqNvB4hegYBEAesHiN6mfhoDce3pjGBvILEppqXA78oIg8BB4BJ4E8DNv9bYJOI3Ae8negtfmgJh/8z4HYiwb4c7em+RKQp30/UZek28L1P3wl8Nz72o5wd95uAfXHgyX1EPi3DWFXi+/RlRC+Sj4jId4GPAW9OrPZcETma+LmCSLicAu6TKBjzX4Hh1L5HnHPvds6VKwzjiRKnPREFg71fg0FTvCcOxLoH+A7wQ6IX93tiV9DFwMeXeg0WIu45/fdEEdNf5mxntEpcD/xdHNRViJfdABxxzs3T/Ncr1u1pjRBr083Ouck4CvJrwIUBN/aqIyJF59xorCF/lqjVpUVVGsY6Jo4U/4Fz7sOrPZZ6wfKQ1w5twDdik7MAv70WhHHMX4jI84jMXV8hyvE0DGOdIiLfI/I3/4/VHks9YRqyYRiGYdQB5kM2DMMwjDrABLJhGIZh1AEmkA3DMAyjDjCBbBiGYRh1gAlkwzAMw6gDTCAbhmEYRh1gAtkwDMMw6oBzEsgicpWIPCgiD4vINbUalLF62Jw2FjafjYfNaeNSdWGQuJTjAeD5RP027wBe5Zy7b6FtWlpaXKFQWOhjAELHcw7jXtF9V3u82dnZRf9faLtKxy+Xy0xPT2cOaqlzmsvlXD6frziGSmPKYrm/B2lCxhR6vKx9pbcLWSfk+FNTUzWbT4Cmpia3YcOGhT5eErWa52rnJnS7asZUS9LjnJ6eZnZ2tiZz2tzcXNU9mkX6umRdp2qveQjLue/Q/VczhsnJScrlctCG51I68xnAw865gwAi8k/A1UQtBjMpFApcccUVi+50ZmYmaFn6y5AlsJqa5hsA0g+brO2mpqYq7itrYrKOF3IzlEqlecvK5blVMcfG5nc1m56u3LEsfb4HDhxYbPUlzWk+n+dJT5rbMz19DbKuSday9HxmzXnWXIW8uGSRnr/m5uag7dLfjdDvQXpcWetkzWd6/+njHzx4cOHBVnGPbtiwgZ6ensX2GXzOuVzlx0vIdzhr31mkr3HWOLNeNtLftdCX32pfstLnk/7unThxYt42CZZ8jz71qU+dsyzku5hFeq7Sz6iF9pVeFnItYf73J+S5kUXo9zXk2RXyspoe92233VZxG3/M4DXnsw04kvj/aLzMWLvYnDYWNp+Nh81pA7PszSVEZD+wH6C1tVK/eqPeSc5nS0vLKo/GqAXJOQ3Vloz6xe7Rtcu5COTHgB2J/7fHy+bgnLsOuA6gs7PTVTIxVmtyrNZ/G2IOh8pmpqx1svaVZXoeHR2tuF3WmELMb0uk4pwm57O9vd1VMvNkzUGIKTNru8nJyXnLQsxvWQ+ltLkt1B9Wy+9UmhBhuESBueR7NJfLzbtH0+ccYpKH+dch1HSY/l5nbRfynAg1IdcqZqSW57cIS7pHOzs75/mQQ0z7IfMZch/D/PMNvd7VmtbT5xP6nMzy5YcQYpIP5Vxeh+8A9ojIbhFpAV4J3HQO+zNWH5vTxsLms/GwOW1gqtaQnXPTIvJG4MvABqKm8/fWbGTGimNz2ljYfDYeNqeNzTn5kJ1zNwM312gsRh1gc9pY2Hw2HjanjcuyB3WlSdvlQ1IHqk1DCkmhCvU5hvhBsnzBab9nll8ia5wh5xfi607vu9Y5lunjpf0+Wf6brDGkffJZPqysfaX9WKF+rbRfrZY5lVnf1zShx0uvlz6/5c7NzDpGSHpWFlnzV8uc9GrTrEL2H5JWGZriV8mHvNxzmv7uZ40xKxUzfY+GpgqmCUlfzBpXqC84JIUyhKzvU0iQ3LnE8VhIpWEYhmHUASaQDcMwDKMOMIFsGIZhGHWACWTDMAzDqANWNKjLOTcv4CWkNmhIEEBoEYd0IEKW4z4ryCFdSGJkZGTeOlmFKyoF5SxESMGUeqiqVCl5PytgJivoKaSecEiRkdCiFel9hRbzSG8XGpyV/t5lbXeORV2WjVo1xsgKqMqar/S1Cj1eSLGQkBri1Rb4yFonZOwhRYDOhWoCPbOey+mApqz5DAn4q2XjmBD5kRV8FvJdCX1Wp885ve+lzOfqP9ENwzAMwzCBbBiGYRj1gAlkwzAMw6gDTCAbhmEYRh2w4pW60oR01MhywIc43NMVaSCs6lBW4+10wFZWYFJIgE9otaJqK15V2nctqwCJyLx5qFSJbaFlIVWPsqrkpANUQrq4hBIy9qxxhnw3Q65B1rLQDjTVIiIVA/NCg6xCgtRCAhNDKzSF7Csk6KjaimJZhFalSlLroK5KAYyhY6w2+C1N1v2RNS/p9UKr94VU+Mq6xpWCsxZaVukZZEFdhmEYhrHGMIFsGIZhGHWACWTDMAzDqANWvDBIJf9JSAJ+1rLQAh9pP0FWgY+QwhWh46xlN5s0Wdey2kIk1ZI+XkjCfUhMQNa4Qwp8VNsRJosQf1i1nYhC/OEw//xCuoCdK5U6eIUWpAm5NiG+52p9lVmEFBAK8Z9Wu+8sau0zTu97ueIOatkhLet+D+kmFRKnUG3xnqzvWNbx0tc3RFYshGnIhmEYhlEHmEA2DMMwjDrABLJhGIZh1AEmkA3DMAyjDljRoK6sogMhASOhAVRpJiYm5i0bHR2d839oB5o01XZfCgl6yiK0E1A6gCGkUEe1zM7OUiqV5iwLCToKKdgSGliTDgYJLbgRQkjxiax1sgJUQs4vpMhBSAGVcyU91vT5hAYvhYyt2g5eIV29Qgs7hBR7qba7VMh21T4TQhCRioGBoZ3V0kFPWd/9kKCnrGuSdc+EFAbJIuQeyXqWhBRMCSkEFHK/LIRpyIZhGIZRB5hANgzDMIw6wASyYRiGYdQBJpANwzAMow5Y8aCukOorIaSraY2Njc1bJ8RxH9L1A8ICVEKCB0I7j1Rb8WaluwNVqtRVy845WdWtQvYVsk5ocFZIEFno/tOEBDgtZwCQkj6n9ByGdrgK6RIVUmEvtPpa+tpU22kstHNVyNyHBJFVs9+lUO13Nk36mRt6fdOEBgWGBPxW001rKeNKExJMnFXpMRTTkA3DMAyjDjCBbBiGYRh1gAlkwzAMw6gDVtSHDPNt8On/s3yek5OT85aluzSVy+V562T5HNJ+glB/Rq38MKH+4mo75awkzrl5vrW0nzfEN5RFaPeuELLiFtL7b21tnbdOSFGH0IIR6euQ9T0P+W6k75dad3sSkYqFOUKPmV4vax6yfHIh1ziLkHs067qnjxfqL672u5BmOTp2Jfcd8sxLE1IUJ7S4TboQUOgzN6RrUsh9VG0XrqxzyTpepc6DVhjEMAzDMNYYJpANwzAMow4wgWwYhmEYdYAJZMMwDMOoA1Y8qCvt4E4HY42Pj8/bJt1RCOY7/EMKE2QREviVRVbQQzXdes6FaoNIakVTU9O8IK504E5oQENI4FBIsFTWdlnzkvV9SROS4J8VqJQ1zvR3OKvISdZ26fsja7taklW8J6Q7UUi3p9COUCHBQyH3UWiHtPR61RYBCQ0+W+1gzJDzDbmPQu+1SseH6jukhTwnqn1Oho6z0v2xlPk2DdkwDMMw6gATyIZhGIZRB5hANgzDMIw6oKJAFpGPiMhJEbknsaxXRL4qIg/Fv3uWd5hGLbE5bSxsPhsPm9P1SUhQ1/XAB4CPJ5ZdA9zinHuXiFwT///mSjuanZ1leHh4zrJ0sEtod6JadfnJCg5ZzmCprEChkGCwLELGtMB+rqcGcyoiFAqFRY+fdS1DujaFXqf0dlnnG1IhKut4IUFIIR3FoHJXLMgOGEnvPx3kFe/nemp0j2ZRyw4+IfsOqfCVte/0HGbNQ1ZFv/R6ofd6rQK9Fljnemowp1nV9EICIUOCrEIDGrOueZpqg8FCvj+hz/OQuQoJFFzWoC7n3DeB/tTiq4GPxX9/DHhZ8BGNVcfmtLGw+Ww8bE7XJ9X6kLc4547Hfz8ObKnReIzVw+a0sbD5bDxsThuccw7qcpE+vqBOLiL7ReROEbnzXBo3GyvHYnOanM9Q94KxuizlHg3N3TVWl9B71J65a4tqC4OcEJHznHPHReQ84ORCKzrnrgOuAygUCm5oaGjxAQUUbFjgOPOWhSSNZ3WSquUYQvwJIX7s0O1CijgsQNCcJuezo6PDVXut0oQ8OLL8TJW6h0H2NQjp8FONv2ihZWmq9WstQWBWdY+2tLS4aoRy1vmE+M2zSH+vQ4v+pMcdsk7WuEJ9ftV0UQo53iLHX/I92tnZ6dLXs9rCJOn9ZPmGQ+6HrOdWSLe+kBiSrO1CC4qEkHXt0mM/l+JP1W55E/Da+O/XAp+vegRGvWBz2ljYfDYeNqcNTkja0yeB/wQuFJGjIvJ64F3A80XkIeB58f/GGsHmtLGw+Ww8bE7XJxXtjc65Vy3w0XNrPBZjhbA5bSxsPhsPm9P1iVXqMgzDMIw6YEW7PTnnKgZ2ZEXuZgV1hDjOs/YVEliStV26mEVWUEBI4npoIZL0stBgpUpjqnWnmZCOLJW2gbDOQCHXKXReQgIxQgKAsggNJqrmeNUGoyyFSkE4oQUaQoqvhHRbC+26k75vs7arNoAyi2oLD1UqXFHLORaRivOXVain2sCkrGuXVUAkTcj9UW3BpmqL1oQGkZ1DIO08TEM2DMMwjDrABLJhGIZh1AEmkA3DMAyjDjCBbBiGYRh1wIoGdUFlh/dyB62EVHGpNoAqi5DzyQoiCwkGCwlsqWY8S6FSAExIFxeYf34hQUJZhFblSQeRZAWVhHwPQoJ2stbLGmdIKdJaVV1aiKzuQOlxhQbXpIO4QoNrQjpqhVbvCtmu0vEhLJgni2rmp5b3aEggbWjwW0iAYT6fr7hdrTrnQW27RKUJLQ2c/k6lx7SU+TQN2TAMwzDqABPIhmEYhlEHmEA2DMMwjDpgxX3Iad9EtQnoab9AiC8qZDyQ7TuodpyVigBA9cVKsqh0fWvtQ67U8Sbr3LKuZdrvElpAJX1dQv1T6eOF+pnS/qLQeanVdV/u+XTOUSqV5i2r5pghBT5CfIChhVZCni1ZY0+PIaTrVta+QovZhIypllTjw622eE+1HdJC1svad1bRkZDzDYlLqLbIybnEdZiGbBiGYRh1gAlkwzAMw6gDTCAbhmEYRh1gAtkwDMMw6oAVD+pKU6uCCSEJ/1nHCw3ECAlWCB1DCOnrEFoIoVJ3p1oXkqgUoJH1eUhwT2gRiWoDYqoNtgsJ7gsZZ7VBgunvwXIEBFX6joR+h0LWC+lwlXWtQgpXhARiZS0LDTpKb5cVoLbahUFEpOK9FTrG9Hev2uCsLEKDI9OE3LfVdgsMDSxNf4ervbfBNGTDMAzDqAtMIBuGYRhGHWAC2TAMwzDqABPIhmEYhlEHrGhQl4jQ0tIyZ1mtOn9kOdtDHP6hjvsQQjoNhe47JNitHqjUfSgk+AbmB0ZkBciEBH5kVdIJqcYWUoEH5o89dLuQoMBqA01qTbWVuSoR2omr0ngW2q7afYUEPlbbfWy1cc7NG3u11dFCuneF3A+hHfZCguZCjlculyvuG8KqP2YFiNUy0NI0ZMMwDMOoA0wgG4ZhGEYdYALZMAzDMOqAFS8MUqkTR7VJ+SE+Olg+/xhU7+cN9VmlqWWhjGqp5DMOvSbp7UK6+UCYPyzL95zef6gvOKTDTwhZ55KOr4D517ceYglCzzmkUEeoPzFNSDxB1n5q1bUti5BuZFlUisM4F1paWti2bducZenvfto3vBDpLmBZ51bN+UJY8Y5qu2mFUm1xovQY0tdgKWM0DdkwDMMw6gATyIZhGIZRB5hANgzDMIw6wASyYRiGYdQBKx7UVanrTWh3oHQQQKjjvNpCHSGE7Cs0sCVNrTqP1Pp8KwVxhBZ1SBPa4Sck6KmWgW7VBgWFbFdtgFMtySokkbVONcuqnYdqO4ZljSkkoCg0eCjkuxfyna11B7b08YvF4qLrZF3LagNN04FfWcuyCnVMTk7OWxZSxCnr+RPyPWttbZ23LH1+WWMKKTISEhy2EKYhG4ZhGEYdYALZMAzDMOoAE8iGYRiGUQeYQDYMwzCMOmDFg7oqBXqEBvOsNCHBZ7WsJFNtBbOV7hKVPr9qO6ZU+z0I2S6rElF6nNUG5GQFXWUFyVR7XSp1l6p1ZTYRqbjP0ICfkIDCWnZ7Sl+r0GuTDgwKrdAUEiCatV36OxNS3apacrkcPT09c5aFBJVlBS+lxxlaTa+9vX3O/1n3Y9YY0sFg4+Pj89YZGxubtyykamPIOqGV3tJzfC7P3NWXdIZhGIZhmEA2DMMwjHqgokAWkR0i8g0RuU9E7hWR342X94rIV0Xkofh3T6V9GauPcw6bz8bC7tHGwu7R9UuID3ka+B/Oue+LSAfwPRH5KvA64Bbn3LtE5BrgGuDN5zqg0EIIIf6bkCIcWfb+ajsthawXWqwgfR1Cu1kF+pCXbT5DiixkzXHINc/aLuT7EuLrCo1TCPFThowpa50QH/IC3aZqeo9W45eupY81TbVFf0LPo9qCPtUWPkl/Hxc4fs3ms1JHttDOY5WKDi20LO0Lziq4kXXt0r7mtC8csp9vg4ODc/4fHh6uOKasMRQKhXnrhBQnOZeYgIpPIefccefc9+O/R4D7gW3A1cDH4tU+Brys6lEYK4aIYPPZWNg92ljYPbp+WZIPWUR2AU8Dbge2OOeOxx89Dmyp6ciMZcfms/GwOW0sbD7XF8ECWUSKwKeB33POzbEBuMgmkml3EZH9InKniNwZanY1lp9azGdoY3NjZajFnC53mpwRTi3mMystyKhfggSyiDQTfTFucM59Jl58QkTOiz8/DziZta1z7jrn3D7n3L6VLpRvZFOr+UznxBqrR63mtB5y/o3azWc6B9iob6RSQINEnu6PAf3Oud9LLH8PcCYRYNDrnPvjCvs6BRwCNgKnz3Xwq8BaHTecHfv5wL9R2/lM7n+tsdbHfT6wGbtHlUYY93Lco2v1usDaHbu/R51zm0I2CBHIPwX8B3A3oPasPyXyadwI7CSa8F9yzvUHHVTkTufcvpB164m1Om44O/blmM/k/ms87GWnEcZt9+hZGmHcNp9zWatjr2bcFdOenHPfAhaK53/uUg5mrD42n42HzWljYfO5fjGHkWEYhmHUAaslkK9bpeOeK2t13LD8Y1+r18bGvTr7Xy5s3Kuz/+VkrY59yeOu6EM2DMMwDGP5MZO1YRiGYdQBJpANwzAMow5YcYEsIleJyIMi8nCcS1eXiMhHROSkiNyTWFb33VZWuvPPWplPWJtzutLzGe97TczpWpxPsHt0MdbinNZyPldUIIvIBuCDwM8Be4FXicjelRzDErgeuCq17Bqibit7gFvi/+sN7fyzF3gm8Ib4Gtd87GtsPmFtzumKzSesuTm9nrU3n2D36GJcz9qb09rNp3NuxX6AK4AvJ/7/E+BPVnIMSxzvLuCexP8PAufFf58HPLjaYww4h88Dz1+Osa+1+WyEOV3O+VyLc7rW53O553StzWcjzOm5zOdKm6y3AUcS/x+Nl60V1lS3FVn+TjFrfT5hDc3pCswnrP05XTPzCXaPBrJm5vRc59OCuqrERa89dZszJlV2ilnP1PN1sflcOvV+XWxOl049X5dazOdKC+THgB2J/7fHy9YKQd1WVhs5h04xS2StzyesgTldwfmEtT+ndT+fYPfoEqn7Oa3VfK60QL4D2CMiu0WkBXglcNMKj+FcuAl4bfz3a4l8BXWFiAjwYeB+59z/SXy0HGNf6/MJdT6nKzyfsPbntK7nE+werYK6ntOazucqOLxfCBwA/gu4drUd8IuM85PAcWCKyO/yeqCPKFruIeBrRO3PVn2sqXH/FJFp5EfAXfHPC5dr7GtlPtfqnK70fK6lOV2L87kac7pW5nOtzmkt59NKZxqGYRhGHWBBXYZhGIZRB5hANgzDMIw6wASyYRiGYdQBJpANwzAMow4wgWwYhmEYdcC6EMgiskVEPiEiB0XkeyLynyLy8hU69pPi45VE5A9X4pjrgVWe01eLyI9E5G4R+Y6IPHUljtvIrPJ8Xh3P510icqeI/NRKHLfRWc05TYzh6SIyLSK/sJLHrZaGF8hx0vbngG865y5wzl1GlBy/PWPd3DIMoR94E/BXy7DvdUkdzOkjwE87554M/CVw3TIcY91QB/N5C/BU59wlwK8D/7AMx1hX1MGcaqerdwNfWY79LwcNL5CB5wBl59zf6QLn3CHn3N8AiMjrROQmEfk6cIuIFEXkFhH5fqwBXR2v1y4iXxSRH4rIPSLyy/Hyd0nUB/NHIjJP6DrnTjrn7iBKdDdqw2rP6XeccwPxv7eR8ZAxlsRqz+eoO1uQoZ06rZW8xljVOY35HaJylnVXanMhluXNpM64CPh+hXUuBZ7inOuP39Ze7pwbFpGNwG0ichNRj85jzrkXAYhIl4j0AS8HnuSccyLSvYznYZylnub09cC/ndPZGKs+n7Ep9X8Cm4EX1ea01jWrOqcisi1e52eAp9futJaX9aAhz0FEPhi/bd2RWPxV51y/rgK8U0R+RFTubBtR26y7geeLyLtF5L8554aAIWAS+LCI/DwwvnJnYiirNaci8jNEAvnNtT+r9ctqzKdz7rPOuScBLyNyQxg1ZBXm9H3Am51zs8t1TsvBehDI9xK9iQHgnHsD8FxgU2KdscTfr44/uyz2KZ0AWp1zB+L93A28XUTe4pybBp4BfAp4MfCl5TwRw7PqcyoiTyHyNV7tnDtTqxNbp6z6fCaO/U3gglhLM6pnted0H/BPIvIo8AvA34rIy2p0bsvGehDIXwdaReS3EsvaFlm/CzjpnJuKNaDzAUTkCcC4c+4fgfcAl0rU/7LLOXcz8PuARduuDKs6pyKyE/gM8N/jB4Zxbqz2fP6YiEj896VAHrCXrHNjVefUObfbObfLObeLSHD/tnPuc7U4seWk4X3IsY/hZcB7ReSPgVNEb2YLmRlvAL4gIncDdwIPxMufDLxHRGaJArR+C+gAPi8irUQmlz9I70xEtsb76QRmReT3gL0u1cDaCGe15xR4C1Enl7+Nn+PTzrl9NTm5dUgdzOcrgF8VkSlgAvjlRJCXUQV1MKdrEuv2ZBiGYRh1wHowWRuGYRhG3WMC2TAMwzDqABPIhmEYhlEHmEA2DMMwjDrABLJhGIZh1AEmkA3DMAyjDjCBbBiGYRh1gAlkwzAMw6gDTCAbhmEYRh1gAtkwDMMw6gATyIZhGIZRB5hANgzDMIw6wASyYRiGYdQBJpANwzAMow4wgWwYhmEYdYAJZMMwDMOoA0wgG4ZhGEYdYALZMAzDMOoAE8iGYRiGUQeYQDYMwzCMOsAEsmEYhmHUASaQDcMwDKMOMIFsGIZhGHWACWTDMAzDqANMIBuGYRhGHXBOAllErhKRB0XkYRG5plaDMlYPm9PGwuaz8bA5bVzEOVfdhiIbgAPA84GjwB3Aq5xz9y20TXd3t3vCE56w5GOFjFFEgvY1Oztbk31lrZM1ztBxVUPI8dLrHD9+nMHBwcxBLXVO29vbXXd396LHa2oKe+dLb1ft9Q39Pi/nvIRQ7TjT2w0ODjI2NlaT+fy/7Z1vrGXlVcafJfMPpkNmYJAZ6LRgJCV88E8dW5tgGm2btI0JNG20TSM0JQHUahs1ztQaNfgFNVE/2BIJ0hmioVatMsaSFommaqthbJDSVqBqG6hQBrjDTAfmpBeXoQAAF+5JREFUMgPLD+fAnLvOOmc/5z1777P3uc8vmcw9++797rXf9b77vWevZ68FANu2bfOdO3dOtYGxcx6YsbDMHDlyBMeOHavFp9u3b/ddu3bFNtZ8zsZi6brQF5h7CXsPquKJJ57As88+Sw3iDTO3foY3APiGu/8PAJjZpwBcBWDiZL/oootwxx13zHyi06dPj22LN/oNG8YvJeu8559/fmo7k9oqdVbcLztf9kcCM2my46LtL7744prP11xzzdgxI8zk0+3bt+OGG26YauemTZumne8Voo8zH5w6dWps28aNG9d8jtcL5H1+1llnrfmc9W92XGyf/YMjkl1LZkOVP2+55ZZpp5l5ju7cuRM33XTTmm3MGN68efPYtujT0j+ao48B7sbILjSMXdlci/1Q1x8O+/btm/brmXy6a9cu3HbbbWu2xbGfjUV2HjHEPmcXtXg+xgdZ+9m1ZDbEccbe4+M4jzbdeOONY8dMYp5H1hcDeHTk82PDbWsws+vN7LCZHV5ZWZnjdKIFKn066s8TJ060apyYmZnn6PHjx1szThQx0xw9evRoq8aJ+Whc1OXut7r7Xnffu2PHjqZPJxpm1J9bt25dtDmiBkZ9um3btkWbI+Zk1J8xpCS6zTyPrL8NYM/I51cPt03kpZdewgsvvDC2bRT28VTVYwIgf+SQPQqNZG0x58uI18c8YsnOl/VLnY+Vhszs06pH+aWx0uzaMt/Fx2+Zz7NHXaUx6+iHOJ4ntcUQrwUY7wdGAzFCkT9jP8fryWxYXV0d2xavJzuOme/so8r46DXzQ9bHsX02PMQ8ys/aymyY1m6g6L477fxM6GvSfiX7sMS+y/qtNM7LrAPZo3xmTWHCm5OY5+59H4DLzOxSM9sE4L0ADs3Rnlg88ulyIX8uH/LpElP8DdndT5vZhwB8DsBZAG5396/WZploHfl0uZA/lw/5dLmZ55E13P2zAD5bky2iA8iny4X8uXzIp8vLXAvyrJjZ2PN0RmrOxGayV6OymEM8jo1ZM+fLXvGJMZUs1nbOOeeMbYuxySymmvVV3I/p31LcvVIDkPnu5MmTY9tiHIbxHcDF6DNi++yrEXW9k8vGw0qvrxR3r3xdiYm/AVysMuv3GLvL5lVJbHbSPnFbpgsofb0mIx6XvTJWF2Y21n9ZbDTC3AOZ/s72y8ZBdj9lYvRZnzO6kuy4uC0bd88999zYtmwNGWWWd5eVOlMIIYToAFqQhRBCiA6gBVkIIYToAFqQhRBCiA7QqqgLGBeExM9ZAJwRNLHJA2KgvjTgz4qOmDyuzPVl+7DX3BRmRuWbjTDiN1Z8ViqsyUQkESaJBJtDPQppWIFK2zDCS1ZgGPsqOy4TxJSKckoTdcSxkO3D5GTPBJtMju94XJ2FHdy9UsRVmlQp8x2To51NaMQkiGG2MQl+2OPOPvvssW3xmlmxW4a+IQshhBAdQAuyEEII0QG0IAshhBAdQAuyEEII0QFaz9RVJQJiKvNM2i/CZHEpLV5dWsWIzTrEHJcJkxhhQp0w4rMSSgUcmTiktHJNlrEpkp2PEXVk52eOm6eSDEu8pihaYfuTEXUxY58RSwJcRqjSewlT7YkVn0U7m56jVe1nY5gRHTLiV2C8XxixZEY2H5nMa6xfmGyMTYtm9Q1ZCCGE6ABakIUQQogOoAVZCCGE6ACtJwapiruUJs7I4gRZfIFJZJHZwCQwyaoYMdWWmDgak3gho8mYsplVxj3ZWHDsXzb5BJOwIbORSQySxUqrKiGxNpT6oY14Y+z7eI1Z3zHagcxXWf8xY5+ZD0wFIWD8npDtw1RIypJNMOOsaaqqc5XGvpkkIJO2MftEv5RW+Mpg7vvsPZe5d7HoG7IQQgjRAbQgCyGEEB1AC7IQQgjRAbQgCyGEEB2gdVFXJAbus4B4VjGl9CV1pvoSk1wiOy6zMx7HvmzOCAwyQUOVwKDORBKZAIh5cZ5JRMAI1oDx/swELIzIIrOJqabFVHYCOPEJI6RpIhFIbL9KBMQKd5ikMaVJVJg5w95LmHsQU92NrQAXx0esIFS3j2P/MWMqu5bYL4z4NTuOndsRdo5GH7MixHhcaZWx0usD9A1ZCCGE6ARakIUQQogOoAVZCCGE6ABakIUQQogOsHBRFyMUYCqtZGICRrDFZs5iAvWMgIPJ9AKMCxGYykMAsLq6OvX3dWd6iv1ZKtJhMoplvop9x/qTqfqVXUtpBjVGhMiIgpj50jSsaC1m5srGcNZ/ceyzoq6YKSubV0wVI8amzC6mSlxmQ5yz84iCSmCzI0bYe27sz8wHWZ8z9zzm3slm+GKEpcwcnacilL4hCyGEEB1AC7IQQgjRAbQgCyGEEB1g4dWemJfrsyoxTOIMphpJFuPJqjYxFW+YmFUW483aiv3AxjOY+EmdVCUdYF/mZ/Zh4sPscYyWgIGtFsbYxNBGzDjGD+P8y8ZdNo9iDJBNnBG3sbFgpm1mfJTGHFn9ApOIpE6q7pVMEhBgvO/YWCkzR9n7foTpO+ZaAE67wIyNeRK76BuyEEII0QG0IAshhBAdQAuyEEII0QG0IAshhBAdoFVRl7uPBc6jGKT0JXVWMBKFAqwIKO6XCQ6y46KIKxO/ZDBVjDIxGGNTXZjZWL8zoiMmMQArXopjg0nmwdqUEccBk6wEqK/qVtPVntx97BqjOKt0jk46X4RJJMFUe2Kp8/7C7FOV1KhOH5vZ2L0qXm8mfs3uLcy9i7nnZufLiPdORpw5ab9IafIn5j4xjz/1DVkIIYToAFqQhRBCiA6gBVkIIYToAJULspndbmZPmtmDI9vOM7N7zOyR4f87mjVT1Il8ulzIn8uHfLo+YURdBwD8MYA7RrbtB3Cvu99sZvuHn/dVNWRmlGAjkokumKo3TOacrKIIIxRgs/Js3rx5zedMOJCJHJjKNVm/kIKRA6jJp1WiLrZyTWn1LibrEJNVLRsHTOUqtsIPA5PVbMJ8OYAa52jVnCwVxLDCL0YsVSoiY45jK3gxYzYbe7F/m/ZpVTWirE9Kx3B2LyutfhT7kxXgMj7ORLnRTibLGns+lsqecvcvAHgmbL4KwMHhzwcBXF2bRaJx5NPlQv5cPuTT9UlpDPlCd398+PMTAC6ctKOZXW9mh83s8MrKSuHpRAtQPh3154kTJ9qzTsxK0Rw9duxYO9aJEmaeo0ePHm3POjE3c4u6fPC9fuIzJne/1d33uvveHTsU8ugD03w66s+tW7e2bJkoYZY5eu6557ZomSiFnaPbt29v2TIxD6WJQb5jZrvd/XEz2w3gSeag7CV1Jr6QxQlKK2wwMbksJhBjwWzMOpKdL7YNjMdP2NhMVQWaKTYW+TTCnI/ZxvZv9FXmuyy2V1WlapINjE1MPDWLLTLVbWaoUlWLP4Hx68muL+tjJk7PxOBZHUL0YTZnGF0AW40o+oKpLpftt2XLlkqbhhT5tCqGXHovy2Aq87EJXJhYe+arupLwsJReX0bpN+RDAK4d/nwtgLuKLRBdQT5dLuTP5UM+XXKY157uBPAlAK8zs8fM7DoANwN4m5k9AuCtw8+iJ8iny4X8uXzIp+uTykfW7v6+Cb96S822iJaQT5cL+XP5kE/XJ8rUJYQQQnSAVqs9AdUB9tJEEqUJN9jqQEzCi2xbJtiKZAKfeD620klVO/MIDjKiXaUVqJgkC7H6S0apP9lEL0yflwpNMkFabKtpf2bVnpi+yq45E3FFMj8zFcSYbdncY+4vrAiIEUcx8yHu0/QcZZIqZWMx+qVUZMWKyJiKd4zAsFR4ySY0KU18krZVW0tCCCGEKEYLshBCCNEBtCALIYQQHUALshBCCNEBFi7qigF+tmIKk2mpVIBTKqDKKqREYUkmTGKESKxwgBXF1UU8X2lVLEakx4hIWFFXJvirajuzgRUAMdmnSjOD1U3JGKpz3MV+yOZ/NteqxEtAuQgoI9qVjc/MztKKaKVU3UuyFLil4iym8lHWT1mGL+ZewhzHVMUDuDnKjKlsDWPRN2QhhBCiA2hBFkIIITqAFmQhhBCiA7QeQ66CiUEA48/tmSojwHiMsTShARPDArhYZQbTD1m8tCqmWncMsiqJRBYvYuxmY8iRrN+YvsziPkzMOrMzGxvMOMj6JW5jEk3MS1ViDjYhBFOtJ7se5hqzPo7Hlcae2aQtpZWjIqUxbAZ3H7O9VH8T+5ONzdZVAa60X5j7DTCu72HvXXWib8hCCCFEB9CCLIQQQnQALchCCCFEB9CCLIQQQnSA1kVdUQgQA+eZEIN5iTtLuJEF5RlxDSMeYF/mjyKATBTAiFiy4zLRWhSI1F05porSxCCMACjzHVOBhqn2kvmAaavUTqa6DVA9X5qgSrjGip6YBA3MWMj2OXny5Ng2RnCTCTajnaxAlElm00Yil2mYWeUcYcc+M9eYe1kmoMzmDNN3jMCQFWeVClLjeGEqu01C35CFEEKIDqAFWQghhOgAWpCFEEKIDqAFWQghhOgArYq63L1SVMVmQmGqRGVB+RhwZ7PNRLECW1WIqezCXDMrVornYzOflWBmY9WsYj+VVnsqFVk1ncmKqXhTmmWMETg1nSnI3SkfRphsSKxYihHclVZIYirxZMKvOM4z2Ex90adRoFqnEMzMKqsYZTZm27Zs2TLWNkP0H1NxCxgfG2zGrTjOsmvJzsdkxWOyv83jP31DFkIIITqAFmQhhBCiA2hBFkIIITpA6zHkGBeoim8AeRw0PsvPYgJMvCiDiU+xsVGmLSYhBBN3zfZj2iklizcySR2YWFB2bVkclrke5jg2Jsmcj411lVBXO5PIEklEX2R9wPRLqZ4hm8el/ZDFE7MKUBHmHpTB9FWTczRrj0luwyTOYOdojOmyCVSYZEEZsT9ZvU88Lht3TFWqeSrs6RuyEEII0QG0IAshhBAdQAuyEEII0QG0IAshhBAdoFVRl5lVVhrJAvDZtijOyEQXTCKJ7DhGlMO88J+dLxM0MIItpkpN1lapsI2l6vpYQUM8jq1uxSTKyHwct7HJPJjryfq8Svw2qe14zXHcNVHNK/ZpaWUsRuCX+Sb2X+ZjxqdZIhImWVAm8mJ8yiR2AThRVZ3E88Xry66XrUYWyeZR6T0h9jkrIovjpVSEmO0Tk6MA1QJDibqEEEKInqEFWQghhOgAWpCFEEKIDqAFWQghhOgArYq6gPEgPBPwZ7LrlIqHMrK2olCAzTpUlZkMyK/v5MmTaz5nApXMzij+ip/rFAFlIr3Yv4z4BhgfF6yohMnKw4hmsv4tPV92XF2ZwZh+mpeqMcKKZEqzqEVYQSPTFiM+ywRc2RxlKqsxosMmhHmjMOK6EpisVcD4Pa+0oh9bma80i1vsJ7YKVhwvjBhtog30nkIIIYRoDC3IQgghRAeoXJDNbI+Z/aOZfc3MvmpmHx5uP8/M7jGzR4b/72jeXDEvp06dgvy5XGiOLhenT5/WHF2nMDHk0wB+xd2/bGbbAPyHmd0D4AMA7nX3m81sP4D9APZNa2jTpk3Ys2fP2sZDvO3EiRNjx62uro5ti3ECNs4U4xCbN28e2yeLAca2suOyGAeTVIE5H1tBKMZdJlQeqcWfmZ0xfsJWwGLiPkyCCDa2XxX7nnS+uB8bj2OSQTB9MOH8tc3RKed4BbZ6DhPTZaq7lVb5YhORREp1AZkOISPGISdU02rMn4xfsntuhNW1xLayeydzT2CrRDH3INbHEUZLEPeZJaZd6Rl3f9zdvzz8+TiArwO4GMBVAA4OdzsI4Gr6rGJhbNy4EfLncqE5ulxs2LBBc3SdMlMM2cwuAfDDAP4dwIXu/vjwV08AuHDCMdeb2WEzO/z000/PYaqom3n9mT3NEItlXp8eP368FTsFx7z+XFlZacVOUQ/0gmxmrwLw1wA+4u7HRn/ng+/s6TMkd7/V3fe6+97zzz9/LmNFfdThz61bt7ZgqWCpw6fbtm1rwVLBUIc/d+xQmLlPUAuymW3EYGD8ubt/Zrj5O2a2e/j73QCebMZEUTfy5/Ihny4X8uf6xKrEETaImh8E8Iy7f2Rk++8DeHpEYHCeu/9aRVtHAHwLwE4AT81r/ALoq93AGdtfC+Bu1OvP0fb7Rt/tfi2A74Xm6Mssg91NzNG+9gvQX9tfmaPufgFzALMgXwngnwF8BcDLcrFfxyCm8WkAr8HA4T/t7s9QJzU77O57mX27RF/tBs7Y3oQ/R9uv2ezGWQa7NUfPsAx2y59r6avtJXZXvvbk7v8CYFLur7fMcjKxeOTP5UM+XS7kz/WLMnUJIYQQHWBRC/KtCzrvvPTVbqB52/vaN7J7Me03hexeTPtN0lfbZ7a7MoYshBBCiObRI2shhBCiA7S+IJvZ283sITP7xlC630nM7HYze9LMHhzZ1vnk7m0XGuiLP4F++rRtfw7b7oVP++hPQHN0Gn30aZ3+bHVBNrOzAHwcwDsAXAHgfWZ2RZs2zMABAG8P2/ZjkNz9MgD3Dj93jZcLDVwB4McA/MKwj2u3vWf+BPrp09b8CfTOpwfQP38CmqPTOID++bQ+f7p7a/8AvAnA50Y+fxTAR9u0YUZ7LwHw4MjnhwDsHv68G8BDi7aRuIa7ALytCdv75s9l8GmT/uyjT/vuz6Z92jd/LoNP5/Fn24+sLwbw6Mjnx4bb+gKV3L0rWEFi+hnpuz+BHvm0BX8C/fdpb/wJaI6S9Man8/pToq5CfPBnT2cl6laYmH490+V+kT9np+v9Ip/OTpf7pQ5/tr0gfxvAnpHPrx5u6wu9SO5u7SWm77s/gR74tEV/Av33aef9CWiOzkjnfVqXP9tekO8DcJmZXWpmmwC8F8Chlm2Yh0MArh3+fC0GsYJOYWYG4E8BfN3d/2DkV03Y3nd/Ah33acv+BPrv0077E9AcLaDTPq3VnwsIeL8TwMMA/hvAxxYdgJ9i550AHgdwCoO4y3UAzsdALfcIgH/AoNrKwm0Ndl+JwaORBwDcP/z3zqZs74s/++rTtv3ZJ5/20Z+L8Glf/NlXn9bpT2XqEkIIITqARF1CCCFEB9CCLIQQQnQALchCCCFEB9CCLIQQQnQALchCCCFEB1jXC7KZfWxYneMBM7vfzN443P5NM9tZ0N7lZvYlM1s1s1+t32IxjQb8+f5hW18xsy+a2Q/Wb7WYRgM+vWqkrcNmdmX9VotJ1O3PkXZ/1MxOm9l76rO2fTYs2oBFYWZvAvBTAF7v7qvDwbBpzmafAfBLAK6e1z4xGw35838BvNndV8zsHQBuBfDGOdsUJA359F4Ah9zdzewHAHwawOVztikIGvLnyxWtfhfA5+dta9Gs52/IuwE85e6rAODuT7n7/43uYGZnm9ndZnbDsKblBcPt32OD2qIXjO7v7k+6+30YvNQu2qUJf37R3VeGH/8Ng7SDoj2a8Ol3/Uzyha3oaF7kJaV2fw75RQzSVnYupeasrOcF+fMA9pjZw2b2CTN7c/j9qwD8HYA73f1PAPwZgPcPf/dWAP/p7kfaM1dU0LQ/rwNwd91Gi6k04lMze5eZ/ReAvwfwwebMF4Ha/WlmFwN4F4BbmjW9Hdbtguzu3wXwIwCuB3AEwF+Y2QdGdrkLwCfd/Y7h59sBXDP8+YMAPtmSqYKgSX+a2U9gsCDvq9lsMYWmfOruf+Pul2MQWvqdBkwXCQ35848A7HP3lxoxumXW7YIMAO7+orv/k7v/FoAPAXj3yK//FcDbh4nD4e6PYlC94ycBvAH6ttQ5mvDnMM54G4Cr3P3pRi9AjNHkHHX3LwD4vnnERGI2GvDnXgCfMrNvAngPgE+YWW81POt2QTaz15nZZSObfgjAt0Y+/yaAFQAfH9l2GwaPUf7S3V9s3krB0oQ/zew1AD4D4Gfd/eH6rRbTaMin3//yDd/MXg9gMwD9odUCTfjT3S9190vc/RIAfwXg5939b2s3viXW7YKMQbzioJl9zcweAHAFgN8O+3wYwNlm9nvDz4eGx6WPwsxsl5k9BuCXAfyGmT1mZuc2Yr2I1O5PDG4Q52PwV/f9Zna4frPFFJrw6bsBPGhm92Nw4/+ZEZGXaJYm/LlUqNrTDJjZXgB/6O4/vmhbxPzIn8uHfLpcrDd/rtv3kGfFzPYD+DmcUf2JHiN/Lh/y6XKxHv2pb8hCCCFEB1jPMWQhhBCiM2hBFkIIITqAFmQhhBCiA2hBFkIIITqAFmQhhBCiA2hBFkIIITrA/wM1ZopQb8F4CAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 576x576 with 10 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}