06_[Case_C]_NNs_with_Season_Gaps_and_LI.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "06_[Case_C]_NNs_with_Season_Gaps_and_LI.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "machine_shape": "hm",
      "authorship_tag": "ABX9TyNTSCE3UfJY01Y0WF0Y/vM5",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/SoumyaShreeram/Microlensing_with_NeuralNets/blob/master/06_%5BCase_C%5D_NNs_with_Season_Gaps_and_LI.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Iv0iBoEj6qK3",
        "colab_type": "text"
      },
      "source": [
        "## {**Case C**}  Deep learning using NNs with microlensing curves ***with Season Gaps and Linear Interpolation (LI)***\n",
        "\n",
        "The following notebook tests NNs on microlensed light curves that do not contain season gaps. The notebook contains the following sections:\n",
        "1. Input parameters\n",
        "2. Loading data <br>\n",
        "  i. Plotting light curves\n",
        "3. Generations of training and testing data sets\n",
        "4. Compiling the 3 networks (can choose one of them for training) <br>\n",
        "  i. CNN <br>\n",
        "  ii. ResNet-7 <br>\n",
        "  iii. ResNet-18\n",
        "5. Training the network on the optimized parameters <br>\n",
        "  i. Plotting loss-accuracy curves <br>\n",
        "  ii. Displaying the confusioni matrix\n",
        "\n",
        "\n",
        "**Author**: Soumya Shreeram <br>\n",
        "**Script adapted from**: Millon Martin & Kevin Müller <br>\n",
        "**Date**: 23rd February 2020"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "deDcrZnEPlzz",
        "colab_type": "text"
      },
      "source": [
        "Import all required python libraries"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hIOkLVF5WNgW",
        "colab_type": "code",
        "outputId": "cfed1f11-1ba6-4919-affe-169a8dfcc24d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "from google.colab import drive\n",
        "import os\n",
        "import pickle\n",
        "import importlib\n",
        "import itertools\n",
        "\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "plt.rcParams.update({'font.size': 16})\n",
        "import random\n",
        "import os\n",
        "from IPython.display import Image, SVG\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import classification_report\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from scipy.interpolate import interp1d \n",
        "\n",
        "%tensorflow_version 1.x\n",
        "import tensorflow as tf\n",
        "\n",
        "from tensorflow.keras.models import Model, Sequential\n",
        "from tensorflow.keras.layers import Input, Activation, InputSpec\n",
        "from tensorflow.python.keras.layers import Conv1D, Conv2D\n",
        "from tensorflow.python.keras.layers import MaxPooling1D, MaxPooling2D, GlobalMaxPooling2D\n",
        "from tensorflow.keras.layers import Dense, Dropout, Flatten, Add, BatchNormalization, Concatenate\n",
        "from tensorflow.keras.optimizers import SGD, Adam\n",
        "from tensorflow.keras.utils import plot_model\n",
        "from tensorflow.keras import backend as K\n",
        "from tensorflow.keras.layers import Layer"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "TensorFlow 1.x selected.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jDf-kPS_OvVQ",
        "colab_type": "text"
      },
      "source": [
        "Mount google drive"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V7aoHjFTP2s9",
        "colab_type": "code",
        "outputId": "e6a398ff-cb3c-4f7f-c924-dfd6aea5df7d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 122
        }
      },
      "source": [
        "drive.mount('/content/gdrive')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/gdrive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CFu89FBF3rYM",
        "colab_type": "text"
      },
      "source": [
        "Directory where the `python` files are stored on Google Drive. The `.py` files that are imported must be in the same working directory (nb stands for notebook). To reload the imported files - if you have made any changes to them - use `importlib.reload(short_name_for_the_imported_py_file); eg. dat or nnet` ."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aQmHoYFN3sX1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "os.chdir('/content/gdrive/My Drive/Colab Notebooks')\n",
        "\n",
        "import nb_02i_preprocessing_data as dat\n",
        "import nb_03_neural_network_designs as nnet"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DJbzcl4A8xJQ",
        "colab_type": "text"
      },
      "source": [
        "Setting up the path to the data directory"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gl3kJxIDwwxH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "current_dir = os.getcwd()\n",
        "data_dir = '/content/gdrive/My Drive/training_set_microlensing/'\n",
        "save_image_dir = r'/content/gdrive/My Drive/Colab Notebooks/Images/'\n",
        "save_model_dir = r'/content/gdrive/My Drive/Colab Notebooks/Models/'\n",
        "save_data_dir = r'/content/gdrive/My Drive/Colab Notebooks/Data/'"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mF4SshagaAw2",
        "colab_type": "text"
      },
      "source": [
        "### 1. Input parameters\n",
        "\n",
        "The scale radius used here is in pixels, where 1 pixel = 0.67 $R_{\\rm ref}$; $R_{\\rm ref} = 0.048 \\ R_{\\rm E}$ is the reference radius for QSO $J0158$ that is borrowed from [Mosquera & Kochanek (2011)](https://iopscience.iop.org/article/10.1088/0004-637X/738/1/96/pdf)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xYQFOUe7Z-98",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "v_t = [500, 300] # transverse velcity\n",
        "v_t_idx = 0 # choose index value for the velocity you would like (0 or 1)\n",
        "r_0 = (2,4,10,15,20,30,40,60,80,100) # scale radius\n",
        "output_class_names = [str(r) for r in r_0 ]\n",
        "\n",
        "# select type of Interpolation\n",
        "cubic, linear = False, True\n",
        "\n",
        "\"\"\"\"select the type of nnet's to use\n",
        "old_version :: borrowed from Martin & Kevins code\n",
        "modified_version :: some minor changes compared to previous version\n",
        "\"\"\"\n",
        "old_version, modified_version = True, False\n",
        "\n",
        "# data loaded with season gaps\n",
        "season_gaps = True\n",
        "\n",
        "# if using the new data set generated by Eric\n",
        "new_data_set = True\n",
        "\n",
        "# generates new interpolated light curves? OR Use previously saved light curves?\n",
        "generate_new_files, use_saved_file = False, True"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ih8HAD2Gw56t",
        "colab_type": "text"
      },
      "source": [
        "If using the `new_data_set`, make sure to have the `numLcurvesPerRadius...npy` file in the `v_x00` folder. This file contains the *number of light curves per radius, total curves, number of samples, etc* such that one does not need to hard-code the parameters when opening the files, whose names contain such variable. See more details on how it is read in the imported file `nb_02i_preprocessing_data.py`."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E9x31ZGmsWsD",
        "colab_type": "text"
      },
      "source": [
        "### 2. Load the data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f58qKFoTr0TS",
        "colab_type": "text"
      },
      "source": [
        "There are 10 categories into which the light-curves are classified. Each category is represented by the scale radius `r_0`. The light-curves corresponding to each `r_0` are collectively stored in the 3D array `l_curves` whose shape corresponds to `= (num_features, num_samples, 1)`"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EwCb_eMa9AlY",
        "colab_type": "code",
        "outputId": "bfa2c2a7-e0e3-438f-e8ec-f7943c7e136f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "# sets the sampling parameters\n",
        "sample_params = dat.setSamplingParameters(v_t[v_t_idx], data_dir, season_gaps)\n",
        "\n",
        "# initializes data arrays\n",
        "class_cat, l_curves, out_catergories, out_radii = dat.initializer(r_0, \\\n",
        "                                                                  sample_params)\n",
        "\n",
        "for idx, rad in enumerate(r_0):\n",
        "  filename = dat.getFilename(data_dir, idx, rad, v_t[v_t_idx], sample_params, \\\n",
        "                             season_gaps, new_data_set)\n",
        "  \n",
        "  # load data from the file\n",
        "  l_curves, _, _ = dat.loadData(filename, idx, l_curves, sample_params, idx,\\\n",
        "                                  season_gaps, new_data_set)\n",
        "  out_catergories[idx*sample_params[0][idx]:(idx+1)*sample_params[0][idx]] = class_cat[1][idx]\n",
        "  out_radii[idx*sample_params[0][idx]:(idx+1)*sample_params[0][idx]] = r_0[idx]\n",
        "\n",
        "print('Shape of light curves vector:', np.shape(l_curves))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Shape of light curves vector: (50000, 955, 1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SrzBq0Rz-k4E",
        "colab_type": "text"
      },
      "source": [
        "#### 2.1 Interpolation of the light curves"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2KU1mbCN_yhc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def interpolateLCurves(l_curves, mjhd, l_curves_interpolated, type_inter='linear'):\n",
        "  \"\"\"\n",
        "  Function for interpolating the light curves 'linearly'\n",
        "  @l_curves :: non-interpolated light curves 3Darray\n",
        "  @l_curves_interpolated :: empty array with shape same as l_curves\n",
        "\n",
        "  @Returns:: linearly interpolated light curves 3Darray\n",
        "  \"\"\"\n",
        "  for i in range(len(l_curves)):\n",
        "    # x array of times\n",
        "    x = mjhd\n",
        "    y = l_curves[i, :, 0]\n",
        "    \n",
        "    # interpolation\n",
        "    f = interp1d(x,y, kind=type_inter)\n",
        "    \n",
        "    # capturing the features using y_inter\n",
        "    xnew = np.linspace(x[0], x[-1], num=len(mjhd))\n",
        "    y_inter = f(xnew)\n",
        "\n",
        "    # setting the shape same as that on l_curves array\n",
        "    y_inter.shape = (len(y_inter), 1)\n",
        "    l_curves_interpolated = np.append(l_curves_interpolated, [y_inter], axis=0)\n",
        "  return l_curves_interpolated, xnew"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gOBzH-tKqJrI",
        "colab_type": "text"
      },
      "source": [
        "Linear and cubic interpolation were tested and it was found that the linear interpolation well approcimated the gaps. Cubic interpolation exaggerated the features in the gaps."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6C1EjlFE-iNW",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# initializing a light curve array to store the interpolated curves\n",
        "l_curves_interpolated = np.array([])\n",
        "l_curves_interpolated.shape = (0, len(l_curves[0,:,0]),1)\n",
        "\n",
        "# time array loaded from a file\n",
        "mjhd = np.load(data_dir+'Light_Curves/mjhd.npy')\n",
        "\n",
        "if generate_new_files:\n",
        "  # linear/cubic interpolated the light curves\n",
        "  if linear:\n",
        "    lc_inter_linear, mjhd_inter = interpolateLCurves(l_curves, mjhd, l_curves_interpolated, \\\n",
        "                                        type_inter='linear')\n",
        "    \n",
        "  if cubic:\n",
        "    lc_inter_cubic, mjhd_inter = interpolateLCurves(l_curves, mjhd, l_curves_interpolated, \\\n",
        "                                        type_inter='cubic')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BtSuikKUfFGa",
        "colab_type": "text"
      },
      "source": [
        "Saves the light curve arrays"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NUkVAet9qVZw",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "if generate_new_files:  \n",
        "  dat.saveFile(data_dir, 'Light_Curves/lcurves_SeasonGaps_LinearInter_%d'%len(lc_inter_linear), lc_inter_linear)\n",
        "  dat.saveFile(data_dir, 'Light_Curves/mjhd_inter', mjhd_inter)\n",
        "\n",
        "  # prints the light curves array shape\n",
        "  dat.printLcurvesShape(lc_inter_linear)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f0ALVXUZqUH4",
        "colab_type": "text"
      },
      "source": [
        "Loads the ligths arrays (if not generated again)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hOLc2Ihge47u",
        "colab_type": "code",
        "outputId": "c6af54ee-e177-4ce1-9cba-0c34e85f2d4a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "if use_saved_file:\n",
        "  lc_inter_linear = dat.loadFile(data_dir, \\\n",
        "                                 'Light_Curves/lcurves_SeasonGaps_LinearInter_%d'%len(l_curves))\n",
        "  mjhd_inter = dat.loadFile(data_dir, 'Light_Curves/mjhd_inter')\n",
        "  \n",
        "  # prints the light curves shape\n",
        "  dat.printLcurvesShape(lc_inter_linear)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "shape of light curves arrays: (50000, 955, 1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c4tKn20MtLo5",
        "colab_type": "text"
      },
      "source": [
        "#### 2.1 Plots of some random light-curves"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dbfCcU01tQAf",
        "colab_type": "code",
        "outputId": "67be1146-d01e-4379-8303-466ce9aeb858",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 313
        }
      },
      "source": [
        "fig, ax = plt.subplots(1, 1, figsize=(6, 4))\n",
        "n = [9900, 2600]\n",
        "colors = ['#7bdbaa',  '#8b88eb', '#888894']\n",
        "\n",
        "mjhd = np.load(data_dir+'Light_Curves/mjhd.npy')\n",
        "\n",
        "for i, num in enumerate(n):\n",
        "  # light curves with cubic interpolation\n",
        "  if cubic:\n",
        "    ax.plot(mjhd_inter, lc_inter_cubic[n[i], :, 0], '.', color='r', label='Cubic interpolation', markersize=4)\n",
        "\n",
        "  # light curves with linear interpolation\n",
        "  ax.plot(mjhd_inter, lc_inter_linear[n[i], :, 0], '.', color='k', markersize=2)\n",
        "\n",
        "  # without interpolation\n",
        "  ax.plot(mjhd, l_curves[n[i], :], '.', color=colors[i], label='Radius: %i'%(out_radii[n[i]]), markersize=8)\n",
        "\n",
        "dat.setLabels(ax, 'HJD (days)', 'Magnitude', ticks=True, legend=True)\n",
        "fig.savefig(save_image_dir+'/interpolated_lcurves.png') \n",
        "\n",
        "print(out_radii[9900], out_radii[2600])"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "4.0 2.0\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n_QsEWkm-3xj",
        "colab_type": "text"
      },
      "source": [
        "### 3. Generate training and testing data sets\n",
        "\n",
        "Proceeding with linear interpolation."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xT004oTHEMmb",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train_l_curves, test_l_curves, onehot_train, onehot_test, train_radii, \\\n",
        "test_radii  = dat.generateTestTrain(lc_inter_linear, out_catergories, out_radii, r_0)\n",
        "\n",
        "trainX, trainy, testX, testy = train_l_curves, onehot_train, test_l_curves, onehot_test"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6Okkl7tfzhu4",
        "colab_type": "text"
      },
      "source": [
        "### 4. Compiling the 3 networks\n",
        "\n",
        "The three networks compiled here are coded in `nb04_Neural_Network_Designs.py`. These `.py` files can be found in the GitHub repository. \n",
        "\n",
        "Note that if you are running this from colab, need to have these `.py` files in your drive folder *Colab Notebooks*. \n",
        "\n",
        "The three networks are the Convolutional Neural Network (**CNN**), Residual Neural Network with 7 layers (**ResNet7**), and the Residual Neural Network with 18 layers (**ResNet18**). Refer to the report for more detailed information about the neural network stuctures."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VqVXaRDVC0uB",
        "colab_type": "code",
        "outputId": "cfe981f0-6a79-4ad5-d140-4f82848f1a79",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "# choose one of these networks to use by setting the boolean to True and rest to False\n",
        "CNN, ResNet7, ResNet18 = False, True, False\n",
        "\n",
        "model_types = [CNN, ResNet7, ResNet18]\n",
        "model_name_list = ['CNN','ResNet7', 'ResNet18']\n",
        "\n",
        "for i in range(len(model_name_list)):\n",
        "  if model_types[i]:\n",
        "    model_name = model_name_list[i]\n",
        "print('Using the Neural Network:', model_name)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using the Neural Network: ResNet7\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TGINq-rrVXgU",
        "colab_type": "text"
      },
      "source": [
        "The following code block uses the methods defined in the file `dr04_Neural_Network_Designs.py`: imported as **`nnet`**\n",
        "\n",
        "Model parameters are defined in the methods in **`nnet`** additionally some more details:\n",
        "* `num_hidden_nodes`: number of nodes in hidden fully connected layer\n",
        "*`dropout_ratio`: % of nodes in the hidden layer to dropout during back-propagation update of the network weights"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zg4jZomdnLWf",
        "colab_type": "text"
      },
      "source": [
        "#### 4.1 Compiling CNN"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KMuhDzuAziUB",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "if CNN:\n",
        "  # reloads the file if any changes are made\n",
        "  importlib.reload(nnet)\n",
        "\n",
        "  # CNN model parameters\n",
        "  num_filter = (32, 32, 44, 44)\n",
        "  kern_size = (21,21,15,15)\n",
        "  maxpoolsize = (None, 15, None, 15)\n",
        "  shortcut_link = (0,1,2,3)\n",
        "  batch_norm = False\n",
        "  dropout_ratio = 0.4  \n",
        "  sampling = 3\n",
        "  num_pieces = 1\n",
        "  num_hidden_nodes = 1500\n",
        "  \n",
        "  # build the model\n",
        "  inputs, outputs = nnet.buildCNN(sample_params[2], num_filter, kern_size, maxpoolsize, \\\n",
        "                                  num_hidden_nodes, dropout_ratio, r_0)\n",
        "  \n",
        "  # optimization, compilation parameters\n",
        "  optimizer_type = Adam(lr=5e-4)  \n",
        "  loss = 'categorical_crossentropy'  \n",
        "  metrics = ['categorical_accuracy'] \n",
        "\n",
        "  # compiles the model\n",
        "  model = nnet.compileDisplayNetwork(inputs, outputs, optimizer_type, loss, \\\n",
        "                                    metrics, save_image_dir+'CNN', True)\n",
        "    \n",
        "  print('Network created')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "e7tQbQzznRy3",
        "colab_type": "text"
      },
      "source": [
        "#### 4.2 Compiling ResNet-7"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-sC8EiXgzTFy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "if ResNet7:\n",
        "  # reloads the file if any changes are made\n",
        "  importlib.reload(nnet)\n",
        "\n",
        "  # CNN model parameters\n",
        "  num_filter = (32, 32, 44, 44)\n",
        "  kern_size = (21,21,15,15)\n",
        "  maxpoolsize = (None, 15, None, 15)\n",
        "  shortcut_link = (0,1,2,3)\n",
        "  batch_norm = False\n",
        "  dropout_ratio = 0.4  \n",
        "  sampling = 3\n",
        "  num_pieces = 1\n",
        "  num_hidden_nodes = [1500]\n",
        "  \n",
        "  # prepare data with sample cuts with the correct shape\n",
        "  testX_prep = nnet.prepareDataSampleCuts(testX, sampling, num_pieces)\n",
        "  trainX_prep = nnet.prepareDataSampleCuts(trainX, sampling, num_pieces)\n",
        "\n",
        "  # build the model\n",
        "  inputs, outputs = nnet.buildResNet7(num_filter, kern_size, \\\n",
        "                                                  maxpoolsize, num_hidden_nodes[0], \\\n",
        "                                                  dropout_ratio, shortcut_link, \\\n",
        "                                                  batch_norm, trainX_prep.shape[1], r_0)\n",
        "  \n",
        "  # optimization, compilation parameters\n",
        "  optimizer_type = Adam(lr=5e-4)  \n",
        "  loss = 'categorical_crossentropy'  \n",
        "  metrics = ['categorical_accuracy'] \n",
        "\n",
        "  # compiles the model\n",
        "  model = nnet.compileDisplayNetwork(inputs, outputs, optimizer_type, loss, \\\n",
        "                                    metrics, save_image_dir+'ResNet-7', do_print_summary = False)\n",
        "    \n",
        "  print('Network created')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0aszD2Huquz8",
        "colab_type": "text"
      },
      "source": [
        "### 4.3 Compiling ResNet-18"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "e9He7fhcqtIR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "if ResNet18:\n",
        "  # reloads the file if any changes are made\n",
        "  importlib.reload(nnet)\n",
        "\n",
        "  # ResNet model parameters\n",
        "  num_filter = (32, 32, 32)\n",
        "  kern_size = (10,20,50)\n",
        "  n_block = 5\n",
        "  maxpoolsize = 3\n",
        "  num_hidden_nodes =[1000.0,]\n",
        "  dropout_ratio = 0.7 \n",
        "  batch_norm = True\n",
        "  sampling = 1\n",
        "  num_pieces = 1\n",
        "\n",
        "  # build network\n",
        "  inputs, outputs = nnet.buildResNet18(sample_params[2], num_filter, kern_size, n_block, \\\n",
        "                                 maxpoolsize, num_hidden_nodes, dropout_ratio, r_0,\\\n",
        "                                 batch_norm)\n",
        "  \n",
        "  # ResNet compilation parameters\n",
        "  learning_rate = 1e-4\n",
        "  max_epochs = 80\n",
        "  decay = learning_rate/max_epochs\n",
        "  optimizer_type = Adam(lr=learning_rate, decay=decay)  \n",
        "  loss = 'categorical_crossentropy'  \n",
        "  metrics = ['categorical_accuracy'] \n",
        "\n",
        "  # compiles network\n",
        "  model = nnet.compileDisplayNetwork(inputs, outputs, optimizer_type, loss, \\\n",
        "                                     metrics,save_image_dir+'ResNet-18', True)\n",
        "  \n",
        "  #model.load_weights('/content/gdrive/My Drive/Colab_Notebooks/Deep_learning_for_optical_IMaging/Resnet_weights.h5')\n",
        "  print('Network created')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-x_2o3VOUtE2",
        "colab_type": "text"
      },
      "source": [
        "### 5. Training the network\n",
        "\n",
        "Constants used for training the network"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4im8AyzDU_YC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "validation_split_ratio = 0.2  # % of the seen dataset to be put aside for validation, rest is for training\n",
        "max_epochs = 50  # maxmimum number of epochs to be iterated\n",
        "batch_size = 60   # batch size for the training data set\n",
        "num_inputs = 1\n",
        "verbose = 1"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gXnVZJPHksoi",
        "colab_type": "code",
        "outputId": "557c3352-d7cf-478b-d5c8-5f367283b725",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "if ResNet7:\n",
        "  # reshaped data is used while training\n",
        "  trainX = trainX_prep\n",
        "  testX = testX_prep\n",
        "\n",
        "# trains the model\n",
        "model, model_history = nnet.trainModel(model, trainX, trainy, batch_size, max_epochs,\\\n",
        "                                  validation_split_ratio, verbose)\n",
        "\n",
        "# evaluates and predicts the results from the trained model\n",
        "results, predictions = nnet.evaluatePredictModel(model, testX, testy, verbose = 1) \n",
        "\n",
        "print('[test loss, test acc] = ', results)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 32000 samples, validate on 8000 samples\n",
            "Epoch 1/50\n",
            "32000/32000 [==============================] - 12s 368us/sample - loss: 1.8865 - categorical_accuracy: 0.2665 - val_loss: 1.6349 - val_categorical_accuracy: 0.3550\n",
            "Epoch 2/50\n",
            "32000/32000 [==============================] - 6s 173us/sample - loss: 1.6020 - categorical_accuracy: 0.3807 - val_loss: 1.5029 - val_categorical_accuracy: 0.4198\n",
            "Epoch 3/50\n",
            "32000/32000 [==============================] - 6s 173us/sample - loss: 1.4529 - categorical_accuracy: 0.4448 - val_loss: 1.3849 - val_categorical_accuracy: 0.4744\n",
            "Epoch 4/50\n",
            "32000/32000 [==============================] - 6s 172us/sample - loss: 1.3658 - categorical_accuracy: 0.4774 - val_loss: 1.3145 - val_categorical_accuracy: 0.4991\n",
            "Epoch 5/50\n",
            "32000/32000 [==============================] - 6s 175us/sample - loss: 1.2905 - categorical_accuracy: 0.5025 - val_loss: 1.2599 - val_categorical_accuracy: 0.5176\n",
            "Epoch 6/50\n",
            "32000/32000 [==============================] - 6s 180us/sample - loss: 1.2291 - categorical_accuracy: 0.5212 - val_loss: 1.2253 - val_categorical_accuracy: 0.5263\n",
            "Epoch 7/50\n",
            "32000/32000 [==============================] - 6s 176us/sample - loss: 1.1867 - categorical_accuracy: 0.5366 - val_loss: 1.2073 - val_categorical_accuracy: 0.5226\n",
            "Epoch 8/50\n",
            "32000/32000 [==============================] - 6s 173us/sample - loss: 1.1484 - categorical_accuracy: 0.5499 - val_loss: 1.1331 - val_categorical_accuracy: 0.5485\n",
            "Epoch 9/50\n",
            "32000/32000 [==============================] - 6s 173us/sample - loss: 1.1080 - categorical_accuracy: 0.5648 - val_loss: 1.1512 - val_categorical_accuracy: 0.5530\n",
            "Epoch 10/50\n",
            "32000/32000 [==============================] - 6s 173us/sample - loss: 1.0866 - categorical_accuracy: 0.5732 - val_loss: 1.0841 - val_categorical_accuracy: 0.5748\n",
            "Epoch 11/50\n",
            "32000/32000 [==============================] - 6s 173us/sample - loss: 1.0567 - categorical_accuracy: 0.5847 - val_loss: 1.0963 - val_categorical_accuracy: 0.5683\n",
            "Epoch 12/50\n",
            "32000/32000 [==============================] - 6s 173us/sample - loss: 1.0443 - categorical_accuracy: 0.5872 - val_loss: 1.0554 - val_categorical_accuracy: 0.5790\n",
            "Epoch 13/50\n",
            "32000/32000 [==============================] - 6s 173us/sample - loss: 1.0229 - categorical_accuracy: 0.5928 - val_loss: 1.0339 - val_categorical_accuracy: 0.5920\n",
            "Epoch 14/50\n",
            "32000/32000 [==============================] - 6s 173us/sample - loss: 1.0042 - categorical_accuracy: 0.5989 - val_loss: 1.0576 - val_categorical_accuracy: 0.5770\n",
            "Epoch 15/50\n",
            "32000/32000 [==============================] - 6s 173us/sample - loss: 0.9874 - categorical_accuracy: 0.6080 - val_loss: 1.0392 - val_categorical_accuracy: 0.5886\n",
            "Epoch 16/50\n",
            "32000/32000 [==============================] - 6s 172us/sample - loss: 0.9759 - categorical_accuracy: 0.6130 - val_loss: 1.0376 - val_categorical_accuracy: 0.5897\n",
            "Epoch 17/50\n",
            "32000/32000 [==============================] - 5s 171us/sample - loss: 0.9608 - categorical_accuracy: 0.6177 - val_loss: 1.0239 - val_categorical_accuracy: 0.5934\n",
            "Epoch 18/50\n",
            "32000/32000 [==============================] - 5s 171us/sample - loss: 0.9467 - categorical_accuracy: 0.6219 - val_loss: 0.9945 - val_categorical_accuracy: 0.5970\n",
            "Epoch 19/50\n",
            "32000/32000 [==============================] - 6s 173us/sample - loss: 0.9334 - categorical_accuracy: 0.6263 - val_loss: 1.0031 - val_categorical_accuracy: 0.6061\n",
            "Epoch 20/50\n",
            "32000/32000 [==============================] - 6s 173us/sample - loss: 0.9165 - categorical_accuracy: 0.6342 - val_loss: 0.9762 - val_categorical_accuracy: 0.6090\n",
            "Epoch 21/50\n",
            "32000/32000 [==============================] - 6s 172us/sample - loss: 0.9107 - categorical_accuracy: 0.6376 - val_loss: 0.9771 - val_categorical_accuracy: 0.6140\n",
            "Epoch 22/50\n",
            "32000/32000 [==============================] - 5s 171us/sample - loss: 0.8978 - categorical_accuracy: 0.6399 - val_loss: 0.9587 - val_categorical_accuracy: 0.6144\n",
            "Epoch 23/50\n",
            "32000/32000 [==============================] - 6s 172us/sample - loss: 0.8848 - categorical_accuracy: 0.6435 - val_loss: 0.9766 - val_categorical_accuracy: 0.6085\n",
            "Epoch 24/50\n",
            "32000/32000 [==============================] - 6s 174us/sample - loss: 0.8763 - categorical_accuracy: 0.6493 - val_loss: 0.9739 - val_categorical_accuracy: 0.6095\n",
            "Epoch 25/50\n",
            "32000/32000 [==============================] - 6s 173us/sample - loss: 0.8675 - categorical_accuracy: 0.6529 - val_loss: 0.9657 - val_categorical_accuracy: 0.6158\n",
            "Epoch 26/50\n",
            "32000/32000 [==============================] - 6s 172us/sample - loss: 0.8494 - categorical_accuracy: 0.6599 - val_loss: 0.9619 - val_categorical_accuracy: 0.6169\n",
            "Epoch 27/50\n",
            "32000/32000 [==============================] - 6s 172us/sample - loss: 0.8575 - categorical_accuracy: 0.6547 - val_loss: 0.9559 - val_categorical_accuracy: 0.6145\n",
            "Epoch 28/50\n",
            "32000/32000 [==============================] - 5s 170us/sample - loss: 0.8426 - categorical_accuracy: 0.6614 - val_loss: 0.9471 - val_categorical_accuracy: 0.6229\n",
            "Epoch 29/50\n",
            "32000/32000 [==============================] - 5s 171us/sample - loss: 0.8307 - categorical_accuracy: 0.6653 - val_loss: 0.9413 - val_categorical_accuracy: 0.6214\n",
            "Epoch 30/50\n",
            "32000/32000 [==============================] - 6s 174us/sample - loss: 0.8287 - categorical_accuracy: 0.6672 - val_loss: 0.9612 - val_categorical_accuracy: 0.6208\n",
            "Epoch 31/50\n",
            "32000/32000 [==============================] - 5s 170us/sample - loss: 0.8152 - categorical_accuracy: 0.6717 - val_loss: 0.9326 - val_categorical_accuracy: 0.6324\n",
            "Epoch 32/50\n",
            "32000/32000 [==============================] - 5s 171us/sample - loss: 0.8089 - categorical_accuracy: 0.6714 - val_loss: 0.9416 - val_categorical_accuracy: 0.6237\n",
            "Epoch 33/50\n",
            "32000/32000 [==============================] - 5s 171us/sample - loss: 0.8035 - categorical_accuracy: 0.6779 - val_loss: 0.9596 - val_categorical_accuracy: 0.6179\n",
            "Epoch 34/50\n",
            "32000/32000 [==============================] - 5s 171us/sample - loss: 0.7984 - categorical_accuracy: 0.6797 - val_loss: 0.9789 - val_categorical_accuracy: 0.6108\n",
            "Epoch 35/50\n",
            "32000/32000 [==============================] - 5s 171us/sample - loss: 0.7911 - categorical_accuracy: 0.6813 - val_loss: 0.9163 - val_categorical_accuracy: 0.6382\n",
            "Epoch 36/50\n",
            "32000/32000 [==============================] - 5s 170us/sample - loss: 0.7819 - categorical_accuracy: 0.6850 - val_loss: 0.9320 - val_categorical_accuracy: 0.6339\n",
            "Epoch 37/50\n",
            "32000/32000 [==============================] - 5s 170us/sample - loss: 0.7723 - categorical_accuracy: 0.6885 - val_loss: 0.9354 - val_categorical_accuracy: 0.6244\n",
            "Epoch 38/50\n",
            "32000/32000 [==============================] - 5s 170us/sample - loss: 0.7633 - categorical_accuracy: 0.6954 - val_loss: 0.9483 - val_categorical_accuracy: 0.6289\n",
            "Epoch 39/50\n",
            "32000/32000 [==============================] - 5s 170us/sample - loss: 0.7618 - categorical_accuracy: 0.6922 - val_loss: 0.9159 - val_categorical_accuracy: 0.6423\n",
            "Epoch 40/50\n",
            "32000/32000 [==============================] - 5s 170us/sample - loss: 0.7540 - categorical_accuracy: 0.6928 - val_loss: 0.9441 - val_categorical_accuracy: 0.6317\n",
            "Epoch 41/50\n",
            "32000/32000 [==============================] - 5s 171us/sample - loss: 0.7527 - categorical_accuracy: 0.6977 - val_loss: 0.9289 - val_categorical_accuracy: 0.6366\n",
            "Epoch 42/50\n",
            "32000/32000 [==============================] - 5s 171us/sample - loss: 0.7462 - categorical_accuracy: 0.6981 - val_loss: 0.9579 - val_categorical_accuracy: 0.6308\n",
            "Epoch 43/50\n",
            "32000/32000 [==============================] - 5s 171us/sample - loss: 0.7416 - categorical_accuracy: 0.7002 - val_loss: 0.9515 - val_categorical_accuracy: 0.6286\n",
            "Epoch 44/50\n",
            "32000/32000 [==============================] - 5s 171us/sample - loss: 0.7331 - categorical_accuracy: 0.7051 - val_loss: 0.9401 - val_categorical_accuracy: 0.6336\n",
            "Epoch 45/50\n",
            "32000/32000 [==============================] - 5s 171us/sample - loss: 0.7255 - categorical_accuracy: 0.7055 - val_loss: 0.9461 - val_categorical_accuracy: 0.6423\n",
            "Epoch 46/50\n",
            "32000/32000 [==============================] - 5s 172us/sample - loss: 0.7254 - categorical_accuracy: 0.7067 - val_loss: 0.9223 - val_categorical_accuracy: 0.6435\n",
            "Epoch 47/50\n",
            "32000/32000 [==============================] - 6s 173us/sample - loss: 0.7105 - categorical_accuracy: 0.7113 - val_loss: 0.9556 - val_categorical_accuracy: 0.6356\n",
            "Epoch 48/50\n",
            "32000/32000 [==============================] - 6s 176us/sample - loss: 0.7130 - categorical_accuracy: 0.7112 - val_loss: 0.9643 - val_categorical_accuracy: 0.6356\n",
            "Epoch 49/50\n",
            "32000/32000 [==============================] - 5s 170us/sample - loss: 0.7058 - categorical_accuracy: 0.7167 - val_loss: 0.9247 - val_categorical_accuracy: 0.6453\n",
            "Epoch 50/50\n",
            "32000/32000 [==============================] - 5s 170us/sample - loss: 0.7015 - categorical_accuracy: 0.7167 - val_loss: 0.9331 - val_categorical_accuracy: 0.6438\n",
            "10000/10000 [==============================] - 1s 109us/sample - loss: 0.9561 - categorical_accuracy: 0.6339\n",
            "10000/10000 [==============================] - 1s 74us/sample\n",
            "[test loss, test acc] =  [0.956053489780426, 0.6339]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mT-IxkOe-QjL",
        "colab_type": "code",
        "outputId": "95d1d892-a60c-4ec0-8ebc-813e269bcdd2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 428
        }
      },
      "source": [
        "# plots the history\n",
        "fig, ax = plt.subplots(1,2, figsize=(14,6))\n",
        "labels = ['training', 'validation']\n",
        "loss_Accuracy = ['loss', 'val_loss', 'categorical_accuracy', 'val_categorical_accuracy']\n",
        "\n",
        "model_history = model_history.history\n",
        "dat.plotLoss(ax[0], model_history, loss_Accuracy, model_name, labels)\n",
        "dat.plotAccuracy(ax[1], model_history, loss_Accuracy, model_name, labels)\n",
        " \n",
        "fig.savefig(save_image_dir+'/LandA_inter%s_epochs_%d.png'%(model_name, max_epochs))    \n",
        "fig.tight_layout()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1008x432 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bdMzy320eeSF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "data = [model_history[loss_Accuracy[0]], model_history[loss_Accuracy[1]], \\\n",
        "        model_history[loss_Accuracy[2]], model_history[loss_Accuracy[3]]]\n",
        "\n",
        "# saves the loss and accuracy curves\n",
        "filename = save_data_dir+'SeasonGaps_LI_%s.npy'%model_name\n",
        "with open(filename, 'wb') as f:\n",
        "  np.save(f, data, allow_pickle=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qr-jbZX8u_bV",
        "colab_type": "text"
      },
      "source": [
        "### 6. Displaying the classification report, confusion matrix"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6Z-OxLmpv_KC",
        "colab_type": "code",
        "outputId": "5ea51db8-2f66-4d3d-badc-9bd6c8a7b0e4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 340
        }
      },
      "source": [
        "y_predict = np.argmax(predictions, axis=1)\n",
        "y_test = np.argmax(testy, axis=1)\n",
        "\n",
        "print('Test set classification report:\\n')\n",
        "print(classification_report(y_test, y_predict, target_names=output_class_names))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test set classification report:\n",
            "\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           2       0.87      0.88      0.87      1001\n",
            "           4       0.82      0.79      0.80       992\n",
            "          10       0.79      0.73      0.76      1019\n",
            "          15       0.72      0.59      0.65       983\n",
            "          20       0.64      0.74      0.69       982\n",
            "          30       0.55      0.52      0.54       981\n",
            "          40       0.61      0.57      0.59      1048\n",
            "          60       0.53      0.73      0.61      1019\n",
            "          80       0.41      0.42      0.42      1026\n",
            "         100       0.44      0.36      0.39       949\n",
            "\n",
            "    accuracy                           0.63     10000\n",
            "   macro avg       0.64      0.63      0.63     10000\n",
            "weighted avg       0.64      0.63      0.63     10000\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tYbr9Vpsv0Zu",
        "colab_type": "code",
        "outputId": "ec17c614-be96-466a-e09d-cca0859c71b2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 716
        }
      },
      "source": [
        "# need this to avoid callback errors\n",
        "from sklearn.metrics import confusion_matrix\n",
        "importlib.reload(dat)\n",
        "\n",
        "# generate confusion matrix\n",
        "cfm = confusion_matrix(y_test, y_predict)  \n",
        "\n",
        "# row normalisation of confusion matrix\n",
        "cfm_probability = cfm.astype('float')/cfm.sum(axis=1)[:, np.newaxis]  \n",
        "# confusion probability matrix\n",
        "cfm_probability = cfm_probability * 100.0  \n",
        "\n",
        "fig, ax = plt.subplots(1,1, figsize=(10,10))\n",
        "dat.plotConfusionMatrix(ax, cfm_probability, output_class_names, 2)\n",
        "\n",
        "# save figure\n",
        "fig.savefig(save_image_dir+'/cfm_%s_epochs_%d_noInter.png'%(model_name, max_epochs))    \n",
        "fig.tight_layout()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAArgAAAK7CAYAAAAZVPyBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOzdd3gUVRfH8e8FQi8JEGrovTepgoCI0gQLIAqCXZoFrOBLE8SCilgQkSo2pCqiAhYQlCJFkCotdFAh1FASct8/ZhNSNgWymw2b34dnn5DZKefsTpKzd8/cNdZaRERERET8RSZfByAiIiIi4kkqcEVERETEr6jAFRERERG/ogJXRERERPyKClwRERER8SsqcEVERETEr6jAFREREcmgjDEhxpj3jDErjTHhxhhrjCntZr3sxpgxxpgjxpjzrvVvcrNeJmPMIGNMqDHmgjFmozHm7rTIJTYVuCIiIiIZV3mgKxAGLE9ivcnAo8BQoANwBFhkjKkdb72RwHDgfaAtsAqYZYxp59mwk2b0QQ8iIiIiGZMxJpO1Nsr1/0eAj4Ey1trQWOvUAv4EHrLWTnUtywJsAXZYazu6lhUCDgCvWWuHxdr+JyDYWlszbbLSCK6IiIhIhhVd3CajIxABzIy1XSTwJXCbMSaba/FtQFbg03jbfwrUMMaUSX3EKaMCV0RERESSUg3Ya60Nj7d8C05BWz7WeheBXW7WA6jqtQjjyZJWBxIRERHJCEzB7JZLKRkY9bIzEVuAC7GWTLTWTryGPeXH6dGN70Ss+6O/nrQJ+1/jr+d1KnBFREREPOlSFDQs5Oso4MdDF6y1N/g6DF9Qi4KIiIiIJCUMCHKzPHpE9kSs9QKNMSaZ9bxOBa6IiIiIpxnj+5vnbAHKGGNyxlteFbjElZ7bLUA2oJyb9QC2ejKopKjAFREREZGkLAACgC7RC1zThN0DLLbWXnQt/gFntoXu8bbvAWy21u5Ng1gB9eCKiIiIZGjGmM6u/9ZzfW1rjPkX+Ndau8xau8EYMxN4xxgTAOwF+gBliFXMWmv/Mca8DQwyxpwB1uMUwTfjTDWWZlTgioiIiHiS4Xp7j3xWvO/Hu74uA1q4/v8g8AowCggENgJtrLXr4237EnAWeAooAuwAulprv/V82InTJ5mJiIiIeJDJl9XSuLCvw4BFB9dpFgURERERET+gFgURERERT/PsLAZylTSCKyIiIiJ+RSO4IiIiIp6mAVyf0giuiIiIiPgVFbgiIiIi4lfUoiAiIiLiUR7/qFy5ShrBFRERERG/ogJXRERERPyKWhREREREPOn6+6hev6OHX0RERET8ikZwRURERDxNF5n5lEZwRURERMSvqMAVEREREb+iFgURERERT1OHgk9pBFdERERE/IpGcEVEREQ8yQCZNITrSxrBFRERERG/ogJXRERERPyKWhREREREPE0dCj6lEVwRERER8SsqcEVERETEr6hFQURERMTT9FG9PqURXBERERHxKxrBFREREfE0DeD6lEZwRURERMSvqMAVEREREb+iFgURERERT9JH9fqcRnBFRERExK+owBURERERv6IWBRERERFPU4eCT2kEV0RERET8ikZwRURERDzK6JPMfEwjuCIiIiLiV1TgioiIiIhfUYuCiIiIiCdpHlyf0wiuiIiIiPgVjeCKiIiIeJoGcH1KI7giIiIi4ldU4IqIiIiIX1GLgoiIiIinaR5cn9IIroiIiIj4FRW4IiIiIuJX1KIgIiIi4mnqUPApjeCKiIiIiF/RCK6IiIiIJ+mTzHxOI7giIiIi4ldU4IqIiIiIX1GLgoiIiIinqUPBpzSCKyIiIiJ+RSO4IiIiIp6mTzLzKY3gioiIiIhfUYErIiIiIn5FLQoiIiIinqYhRJ/Swy8iIiIifkUFroiIiIj4FbUoiIiIiHiSMZpFwcc0gisiIiIifkUjuCIiIiKepgFcn9IIrogfMsaUNsZYY8y0eMuXGmOsj8JKd3FcC2NML2PMJmPMOddj/LSXjzfNdZzS3jyOPzPGtHA9hsN9HYuIpA0VuOKXYhV41hgzP5F1OuuPnlwNY0xTYBqQDXgPGAGs8mVMGYXrZ3Wpr+MQkeuDWhQkI+hkjGlkrVUhAj2BnL4O4jrW1vW1VxqeT4OA14BDaXQ8f7QGqAL85+tAJAPRRWY+pRFc8Xd7AYtTIGR41tr91trtvo7jOlbU9fVoWh3QWnvEWrvdWhuRVsf0N9bacNdjqAJXJINQgSv+bivwJdDcGNM2uZWjGWOaG2MWGWPCjDHnjTGbjTHPG2MC4q0X09tnjGlqjPnRGHPKGBPmuj+6f7KsMeYFY8wu1/7+NMa0ca2TzxgzwRhz1HXfz8aYym5iutMYM9MYs8e1XpjreLdcRV4Jel+NMdlduf1ljDljjDnrOsanxphy8dbNZIx51Biz2rXeWWPM78aYuxI5Xm3X43jWFe9sY0zJlMbr5ri/G2NOu/pftxljxhpjguKtey3P3Q3GmCWu/E8ZY+bF7nmNXhd40LVob3QLTPx9uYk9sX7oisaYT4wxocaYi8aY48aY9caYEfHWc9uDa4zJbYx5xRiz07X9v8aYOcaYWm5iWOraR4Ar372ubf42xvRN7vGPtZ8HXPt5wBjTyRjzhzEm3BizzxjzjGsdY4x5zhXXBWPMVmPM7W72Vc8Y84ExZkus53S9MaavMVeGvmI99uD8HNtYtwcSiet31zn3Z+x9xH5+jDFDXMvGuIntPdd9z6X0sRGR9EUtCpIRDAE6A68aY36w1iZ5cZMxphvwGXAOmAmEAe2B14FmxpiObvZxIzAY+An4CAiOd/9YoB6wAOfnrjvwtXF6OifgvNj8AigN3AEsNMZUtNZejrWP0cBFYBnOCGJR17qLjDFdrLVzU/ZwJDAD5/H5DfgYiAJK4bwd/yWw2/W4GFeMXXFeOEx3bd8emGOMedpaOy56p8aYmsByIAcwCwgFmgMrcB7TFDHGZALmuHLdB3wCnAfKA4+54oh+QXEtz1194HngF5znro7rWDWMMdWttRdcsY9wLa8FjANOpjQHNzkVx3nbPAD42rX/IKAy0BsYlsz2OYClOOfUamA2UALnuWlrjGljrf3VzaZfAA2A74HLrvU/MMZEWGs/vooU7gJaAfNwzpu7gDeNMeFANZzH6Vuc87oHMNcYU9VauzPWPh4FOgC/AguBPMCtwAdARSD64r1QnMd+GM7zPy3WPv6MF9c9wC3ANzjnXlJ/40a7jveM6/fCTwDGmPZAf+Bn4M1kHwmRxGgI0adU4Irfs9buNsZ8DPQF7gU+T2xdY0w+nCLnAtDQWrvNtfwlnKKgA04f6/R4m96C05f5SSK7rgjUstYed+3vB5yiZDHwA9Ajupg1xryH8wf2Ttc60dpZa/fGi7cwsA54A7jqAteV793AfGvtnfHuy4pzMVW0x3AKovHAk7HifR6nGHjDGDPLWnvYtf77QG7g7tjFtzFmOs5jmFL9cQqm74E7rbUX48V/Odb/r+W5awd0s9bOjLXfT4D7Xcf90lobCgx3jaLWAt5xLbtWdwP5gDustV/HvsMYUyAF27+AU9xOsdY+HGvbqcCPwBTXC6SoeNuFANWttadd648DNgPP4Ly4SanbgEbW2g2u/YzBeSH0KnCYuOf6Ipzz+Cmc5zLaaKBv7BiNMVlwCuMnjDFjrbX7Yj32w4BQa+3wJOK6FWhlrV2aXALW2svGmPtxiuTprhdkAcAUnBdGvZJ7MSwi6ZdeX0hGMRJnVG+kifdWdTx3AHmBSdEFEoCr//FF17e93Gy3LoniFmB09B98l3nAJSAQeD7eSG10oVUj9g7iF7euZcdwCtty8d/CTiGLM1vjeTf7vmStPRNrUT/gFDAgdrzW2nM4j29WnJE8jDGlgGY4j0v8wnsIrqI0hfoAkUC/2MWt69inrLVnXd9e63P3a+zi1mWK62v9q4jzWrh73I+7WzGeXjjnz0vxtv0JZzS0HNDUzXaDootb1/o7cEZgKxlj8lxF3J9GF7eu/RzCGZnPB7yayLke/3zeH78At9ZGAhNx/ja1vIp4Yo6VkuI21vFCcc7r4sAkYCpQCHjMWnvwGo4v4jBc+TQzX94yMI3gSoZgrT1qjHkHpyB4DOdtUHei+xeXudnHWmPM2VjrxLY2mRA2xttXlDHmHyCXtfZAvHWjL2AqFnuhMaYIzhX1bXHejs4eb7uiOG/nppi19rRrNPleY0wIMB8n9z9jF7HGmJxAdeAAMNgk/MUZ3ZIR3Ttc0/V1hZtj7jfG7AfKJBefMSa3a5/b3BX48Vzrc7fOzbLo4iYwuRiv0QKc0c55xpivgCU4hXayRZUxJi9OK8tf1lp3F7stxWnLqIXz9n9syeV6xs397mx0s+yIu/tc5/q/JDyfswFP4rQVVMIZ7Y+tKFcvuZ/DBKy1nxlj2gH3uRZNs9bOTmobEUn/VOBKRjIGp79xiIl3wU8seV1fjyVy/zGc/lR3y5Ny2s2yy4ksj3R9jRlpNsbkx+nZDMEpGr/HGU2NAlrg9LZm49p0Bv6H077xlmvZcWPMeGCkawQ0CGdMoiRJ94fmcn3N5/r6byLrHSMFBW6s/RxOci3HtT53ST0HmVNw3Ktmrd1rjGkMDMdp+3gAwBizHmeUdXESmyeX59F468U+rqdyTex8Tuy+SGKdzy5zcdpDtuO0Df3rWq80zgj1tZzPyf0cJmY+VwrcxF78ish1RAWuZBjW2lPGmFdxLhwZgPOHNb7oP86FE9lNYdz/Afd2r97DOKO2L1lrR8e+wxjzIU6Be01cLQaDgEHGmPLAzThv2w7ByWsYV3Jeba1tlILdnnJ9jX+xXbTEHt/E9lMsybUc1/rceUL0W+3ufqfmc7MMa+0m4C5Xr3N9nFHXJ4BvjDG1k5jOLSV5xl4v3THG1Mcpbn8A2sfrw70H960kKXHVP4fGmEI4/eIncS50m2CMaaxp2STVMnaHgM+pB1cymg9w3mZ/DnB3MU/0Vdk3xb/DGFMX523U+Fdup4Xo6bq+ib3QNbNBY08dxFq7y1o7Eaf/MQro6Fp+BucFQdUU9mpucn1N0AdqnGnCUjRVmKu/dhtQIQU9xr587qJnhSju5r46SW3o6nX+zVo7GOdFRTaci7gSW/80zvzOFV0XGcYX/WLHF+dpSkWfzwvdXAh3YyLbROGdEfUpOH23D+Fc+FYPp6dcRK5jKnAlQ3FN+TQc5+1bd3Ncfo0z8vWIMaZi9ELX1d3RHxaR1MVk3rLf9TX+H/+ncd9XmiLGmGBjTDU3dxXC+f1wIday97gywhW//xdjTDXXaBjW2n040zTVMwnnyB3J1RUqH+KMjH7gGu2Mfcy8rj5d8O1ztwOnf7Wjq50k+tiFcdo/4jDG1HX10sYXXbBecHNfbJ/gFMJxCjFjTAuc2SJ241w8ll65PZ+NMY1weuTdOYHTouMxxpkDuD0w2Vo7D3gZpxXoOWPMNb8rIgJAJuP7WwamFgXJiKYDz+J8dGccrjaG3sCnwB/GmNhzqVbDmcLIFwXuDJypod53FTGHgBuARjhXzbe/xv0WBzYYYzYAf+H0uhbCmZHAAm/HWvdDoAnOHL7NjDE/c2U+3ho4I5WNgX9c6z+B0y/8lTEm9jy4ITgjvDVJmQ9w5lztBPxtjFkAhANlcS64a4pzUZzPnjtr7SXX9G6DgfXGmK9xXgzcjnPRW7l4m/QEHjXGLMcpRs/gPB5tcIq/r5I55Os4heyjxpjqrmOE4PTzXgAecjMymp6sxrkgrJvr4sk/cJ7PjjjvUtztZpufga7GmPnABpye329crR5XzRhTBaddaReuOXettZHGmO44o98zjDE1rbXXPN+xiPiOClzJcFzzXw7Gmb7I3f1fGGOO4PSldsWZrWA3zlRTb/tibkxr7QFXYfsGThGUCViFMxVXe669wA3FGdG+GWcO0fw4BeoKYIy1NmYWBFfePYwx3wOP4BScOXEu7NmGM8/wX7HW32iMuQmnGOuIM1XUT0A3rqLQdF2FfzfOyN5DrpvFKQQ/ItbMET5+7obg5PgwzsWMoTgjrAtIWLB9gfMBGDfivEgJwMnnTeBNa22SH4RhrT1vjGnJlTyfwSmSFwIjrLXuZjlIN1w/gx1wzo1bgYY459ADOC+y3BW4T7m+3ozzwiETzgwQV13gut4J+Aznce8Ra6o5rLW7jDFP4UwbNgHnfBWR64zRPNYiIiIinmMK5bDcU97XYcD7m9dZa2/wdRi+oB5cEREREfErKnBFRERExK+oB1dERETEkwyaB9fHNIIrIiIiIn5FI7hJMDmyWPLE/3RJzyqaswBHwo979Rg1Qyp4df8AF8Ivkj3ntX5SbPIyG++fquHnwsmZK6dXj2HS4BX9ubPh5Mrt3TzSgvJIP/whB1Ae6Ula5LAvdD///fefj8ZRDSYtfuEnIyNPI6ACNyl5AqBr/OkrPeuZhn15dvV4rx7jhze+9er+ATav2k71RpW9tv98WfMnv1IqrVqxhkZNG3j1GJmM9980Wbl8FY2bpeTTdNM35ZF++EMOoDzSk7TI4caGCT5IUTIQtSiIiIiIiF/RCK6IiIiIh6lFwbc0gisiIiIifkUFroiIiIj4FbUoiIiIiHhYOuhQyNA0gisiIiIifkUjuCIiIiIeZIBM6WAI97KvA/AhjeCKiIiIiF9RgSsiIiIifkUtCiIiIiKeZNLHPLgZmUZwRURERMSvaARXRERExMM0gutbGsFNpSZl67Co30SOvbqM02+uZt0LX/Fgoztj7h/Wri/2/c1ub+fHrkt2/788NdXttk+16OHRPA4fOsJLzwzl9pZ3ULZgJYrlKsWBfQcSrPfqsDfodnsPqpWoRbFcpZg5Y9Y1HW/f3v0xx9m7OzSV0Sft4IGDdL+nB0ULFKdI/mLc2+U+DuxPmJs7B/Yf4NEHH6NS2SoUyBNMraq1GTH0Zc6dO+fVmBPEceAg93btTuH8RSkUVIR7Ot/L/hTmMPSlYXRoczvFC5UgR5ZczJg+w8vRJu5a81i3dj39evenVrU65M9TkAplKvHA/Q8SujfU+0G7kZrn48KFCwx6fjBlQsoSlLsAzW9syYpfV3g5Yvf8IQ9/yAGUB6SvPOT6pwI3FWoUq8iP/T8mIHMWHv18OHdNepo/9m1mSo+R9G56DwCTfp9Dozfvi3Nr9e7DRFyO4Ju/fknRcTYe3JFgH1+u+96juYTuDmXBnIXkC8xHwyYNEl1vyoRpXLhwgVvatErV8QY9/T/y5s2Tqn2kRHh4OO1u7cCOHX8zccoEJk2byK6du2nbun2yReq5c+fo0OZ2flvxO0OG/4+5C+bQ66FevDv2Pfo82tfrsUcLDw+nbet2/L1jBx9Pncjk6ZPYvWsXbW5pm6JC+8MPJnD+/AXatm+TBtEmLjV5zJo5i61bttG3fx++/nYeI0e/zJ8bNnJjw2YcOHAwjTJwpPb56P1oH6ZOnsaQ4UOY+/VsihQtwu3tOrHxz41pEP0V/pCHP+QAyiNaeslD/INaFFKhW722ZM6Umdsn9OPcpfMA/Lh9JTWLV6Rnw9uZsGImh04e49DJY3G261H/dgIyBzB99TfcnLt6ssc5c/Ecq0M3eSWHaI2aNmRTqDOi/Nm0L1j2069u19txZDOZMmVi7+5QZn0+55qONXfmfDZv3MITz/Zl2AsvX3PMKTF10jT27tnLn1vWU658OQCq16hOzSq1mTxxCk8OeCLRbVf+vopdO3fz9XfzuaW1U9A3b3ETYSfCGPf2u4SHh5MzZ06vxg8wZdJU9u7Zy6atf8bkUKNGdapXrsmkiZN5asCTSW5/7MQRMmXKxO5du/lsxudejzcxqcnjmecHEhwcHGdZ4yaNqVK+KlMnTWXoiCFejT221OSxaeMmZn7xFR9N+pCeD/QEoFnzZtSteQMjh49i9vxre0fkWvhDHv6QAygPSF95eIZRi4KPaQQ3FbJmCSDicgTnIy7GWX7q/FkymcQf2l4NO3L09H8s2vabt0NMsUyZUnYqpHS9xJwMO8WIF0cxdPRL5M2XN1X7SomF335Hg4b1Y37ZApQuU5rGTRqxcMHCJLeNuHQJgLx54o405wvMR1RUFNZaj8frzsIFC2nQsIGbHBrz7TdJ5wCpf848JTV5xC9uAUqVKklwcEEOHz7s6VCTlJo8Fi74joCAADp37RyzLEuWLHTp2pkli3/k4sWLSWztWf6Qhz/k4MSiPNJTHuIf0sdfvuvUtFXzAXi3yyCK5gsmX448PNLkblpVasjYX9z3OYYEFqFlxQZ89se3XI5K2WeM1AmpzMkxK7k0bgMbB83locZ3eSyHtDbqf6MpX6kcne9Lmxy2bd1G1WpVEyyvUrUK27dtT3Lblq1aUr5COYYMHsq2rds5e/YsS39Zxvj3PuSRxx4mV65c3go7jm1bt1GtesIcqlZLPof0xNN5bN+2nX/++ZdKlSt5IrwUS00eW7duo3SZ0glG/qtUq8KlS5fYvWu3J0NNkj/k4Q85gPKA9JWH+Ae1KKTCliO7aDHuIeY9+g79broXgEuREfT+ciQzE+mR7dGgA5kzZWb66m9SdIxfd63lsz++5e9/9hGYIw89G3ZkcveXKZq3IK8smuixXNLC6t/WMPvzuSz+PflRR08JOxFGYFBgguVB+YMICzuZ5LbZs2dnydIldO/anRtq1Y9Z/sBDvXj73bc8HmtiTpwIIzDQTQ5BQYSFhaVZHKnlyTwiIyN5ou9TBAcX5IGHenkqxBRJTR5hJ0643TZ/UFDMvtOKP+ThDzlEH0t5pJ88PEUdCr7ltwWuMaYi0AO4FSgHZAd2A7OAd6y1qb4MvnxwSeY8MpYtR3fTe+ZIzl+6QKeaNzOh2xAuRFzk87UJC7meDTqy/sBW/jr8d4qOMWzhB3G+/+avX5j76Dheuu0x3vllRkzvb3p36dIlnn9yEI/2f5iKVSr6OpwUuXDhAr3u68W///7LpGkfU6JkCGv/WMdro14nS5YsjPvgHV+HmGENeHIgq1auYt43cwhy/QEUERGJ5rcFLvAQ0A/4BvgMiABaAqOArsaYRtbaVFWHo29/iojLkXT4sB+RUZEA/Pz3agrkyse4zi/yxbrv4vRp1i9VnSpFyvLU7NdSc1i+WPsdd9ZqRY3iFVm19/q4uvTj96dwKuwUD/d9kFMnTwFw/rzz8J89c5azZ86SO09ujx83MCiQk25GasNOhBHkZmQ3tulTPuHXZcv5a/tGypYrC0DTZk3Jlzcv/fs8ycOPPUzNWjU8HnN8QUGBnDzpJoewsOuquPNUHv8bNITJH09h0tSJ3HLrLZ4MMUVSk0dgUJDbKZNOuEa38udPu+fTH/LwhxxAeUD6ysMTDJoH19f8uQd3NhBire1urX3PWjvBWnsP8ApQE3g4tQeoUawCGw/tiCluo63Zt5mCuYMolDt/nOW9GnbiUmSE25Hda5FWFzl5ws7tO/nn2L/ULd+AKsVrUqV4TQYPcK58v+3G9tzRuotXjlulahW2bd2WYPn2bdupXKVykttu2byFoKCgmOI2Wr36NwCwY/sOzwWahCpVq7B1S8Ictm1NPof0xBN5vD76Dd4a8zZvvfMm9/W4z9Mhpkhq8qhatQqhe0MJDw+Ps3z71u1kzZo1zsU53uYPefhDDqA8IH3lIf7Bbwtca+1aa+0pN3fNdH1Nfn6uZBw9c5zaIZUIyBx3ILxh6Rqcv3SBE+FXDh+QOQvd6rXl+63L+e9s6nqJutdvT/il8/x1eGeq9pOW+j3Th9nffxnn1m9gHwDen/wOb41/3SvHbd+hHWtW/8HePXtjlu0L3cfK31fR7vZ2SW5buEhhwsLCElzcsHbNHwAUK17U8wG70f729qxZvcZNDitpn0wO6Ulq8/jgvfEMHzqCESOH06dfb2+GmqTU5NGuQ1siIiKYO3tuzLLIyEhmz5rDLa1bkS1bNq/FHZ8/5OEPOYDygPSVh/iHzMOHD/d1DGlqxIgR9YDuwHfDhw9fmuS6Y14ZTrX8id5/9mI4jzftSoPSNTh94Rwl8xdjQMv7ebjJ3by37DO+33rlE1g61byZBxvdweBvxrH92J6Y5beG1GfxIadg2jnsOzrVbMkna5wL0JqWq8v4e4aQPSAbQTnzUjukCq91eppONW/mfwveY+nOP1KU8zOtu6dovW/nLeTv7TtZu2odG9dvolyFshzcf5Dj/52gRKkQAFYuX8XG9RvZsmkrPy9eStHiRTgffp5N67dQv2m9mH11bXcvY197l4f7PghAgYL5KVGqRJzbkUNHWPTtYl4aNYjK1ZK+Ej575hwpyiG+ajWqMfurOcybM59ixYqy8++d9O/zJNlzZGf8xA/ImjUrAPv37adR/cYYY2h2U1PAmYbqk2kzWLhgIXny5CEs7CTz5szj5aEjqVa9GkNHDLnqt6Cu5S2r6jWqMWvmbObNnUfRYsVcOfQne/bsTPh4fEwO+/btJ6RwCbCWDh3bx2y/fNly1q1bz6aNm1j0w2KKhRQnPDycbdu2U6VqlauO51qlJo+vZs6i3+P9ufW21tzf634OHTwUcztz+rTbacTSYx5FihZhx/a/+Wj8RAoUKMDJsDD+N3goa9esZcr0SRQtmjYvmq42j4Y3NMJgaNa8WbrKIyM+F/6SR1qcU1MmTeHxxx4f4dEkU+jlt0YPD6hdCGOMT28Rvx0+Mnz48OvrinQPMdfT29ypZYzJDCwH6gPVrbUJ3mM2xjwGPAaQL39gvSHvvJLkPksXK0mDanUoEJifLJkzc/LMaTbt3MKmnVvjtBB0atGW4oWKMmH2NKKiomKWh+QK5uC5fwF45M4enD57hq+WfA1AYJ683Fy/GQWDCpAjWw6ioqL4L+w4G3ZsYnvorhTnXbN4+RStd2ur29xvX6smb749BoBnBz7Hpo3uP3Ri8U+LYv7/7MDnOHb0GDM+/yTR4y3+YTFvjnmLqZ9MoXjx4knGljnTtbeLHzt2jA8/+JB169ZhraVO3Tr069+PIkWKxKxz9OhRut/bnZ69etLrgStX5YeGhvLJtE/YunUrp06dIrhQME2aNKF7j+7kyXP1n8R2rR1Zx44dY3ysHOomksN993an273deOyxR2OWD3h6IBs3uu/V/u1s2p4AACAASURBVPmXn64xomtzrXm8/trrLFq02O0+a9Wqxdh33k6T+KOl5vm4ePEikydN5qeffubs2bOUK1eOxx5/lNq1a6dpDlebR89ePXkg1s9GeskjIz4X/pKHt8+pZ599lnVr1/ukETZz0Vw2R6+EU6altXOvr11nrb3B13H4QkYrcN8D+gODrbWvJrt+oRyWrt7t+3mzYV+eXT3eq8c4/MZ3Xt0/wOZV26neyHv9oPmyJj6S7imrVqyhUdPEP6bYE5L6ABBPWbl8FY2bNfL6cbxNeaQf/pADKI/0JC1yuLFhU58WuDl7VfPFoeM4+/ofGbbA9dse3PiMMSNxituJKSluRUREROT6lCEKXGPMcOB/wFTAd1eniIiIiIjX+fM8uEBMcTsMmA48YjNST4aIiIj4hObB9S2/HsE1xgzFKW5nAA9Za6OS2URERERErnN+O4JrjOkHjAD2Az8C98V7NXXMWrvEF7GJiIiIiPf4bYGLMxUYQEmc9oT4lgEqcEVERMTj1KHgW37bomCtfcBaa5K4tfB1jCIiIiLief48gisiIiKS5gyGTBrC9Sm/HcEVERERkYxJBa6IiIiI+BW1KIiIiIh4mObB9S2N4IqIiIiIX1GBKyIiIiJ+RS0KIiIiIp5k1KLgaxrBFRERERG/ogJXRERExMOM8f0tZXGaG40xi40x/xhjzhhj1htjHoq3TnZjzBhjzBFjzHljzEpjzE3eeNw8RQWuiIiISAZkjKkJ/AgEAI8CdwF/AJONMX1irTrZdf9QoANwBFhkjKmdthGnnHpwRURERDKmbkBm4HZr7VnXsiWuwrcn8KExphZwH/CQtXYqgDFmGbAFeBnomPZhJ08FroiIiIgHGa6bi8yyAhHA+XjLTwFBrv93dK0zM/pOa22kMeZL4EVjTDZr7cW0CPZqqEVBREREJGOa5vr6rjGmmDEm0BjzKNAKGOu6rxqw11obHm/bLTgFcvk0ifQqaQRXRERExMPSyQhuQWPM2ljfT7TWToz+xlq72RjTApgH9HUtjgB6W2u/dH2fHwhzs+8Tse5Pd1TgioiIiPin/6y1NyR2pzGmAjAHZzS2N06rQidggjHmgrX2s7QJ0/NU4CahZkgFFo/5zqvH+GvVNo52/sGrx6g95j6v7h9gUIXu9HxzlNf2v/3F+V7bdzRrLRFRl7x6jEzG+11BaZNHZq/uH8ACl+1lrx4jcxrkISKSjo3GGbHtYK2NcC37yRhTABhnjPkCZ/S2lJtto0duT7i5z+fUgysiIiLiUQZjfH9LgRrAxljFbbQ1QAGgEM7obhljTM5461QFLgG7UvlgeYUKXBEREZGM6ShQ2xiTNd7yhsAFnNHZBTjz5HaJvtMYkwW4B1icHmdQALUoiIiIiGRU7wOzgAXGmPE4PbgdgXuBsdbaS8AGY8xM4B1jTACwF+gDlAG6+ybs5KnAFREREfEkk25mUUiStXa2MaYd8AIwCcgO7Ab6AR/FWvVB4BVgFBAIbATaWGvXp23EKacCV0RERCSDstZ+D3yfzDrngYGu23VBBa6IiIiIh10HA7h+TReZiYiIiIhfUYErIiIiIn5FLQoiIiIiHmS4Pi4y82cawRURERERv6ICV0RERET8iloURERERDxMLQq+pRFcEREREfErKnBT6fDBIwweOIT2LTpRpkAFiuQswf59BxKsd+HCBUYMHkXNMvUonb887Vt0YuWKVVd9vD9WraVorpIUyVmCyMhIT6TAnJ7jODL0V7e3z+8bE7NexeDSTO4yig0D5rL7xUUs7T2d3o27kdlkTvYYObJk49nmD7Ki32fsGbSEtU/N5t1OgwnJV8QjOaTU3R26EJitAKOGvZLsugf2H6T3w32pXr4mRfIVp161Bowa9grnzp1Lg0jj+nHxT3S4tSPlS1SiYO7CVC5TjV73Psj2rduT3G7pz8t4pNdj1Kxch0J5i1Kzch0G9B/Iv//8m0aRx3Xo4CGeeepZbm7aiuC8hckdkJd9oftStO1//x2nzyN9KVW0DAXzFKJFk5b8uPhHL0fs3oEDB7m3a3cK5y9KoaAi3NP5XvbvT/hzH9++ffvpcmdXKpatTFDuAoQULknrlrfxw3c/pEHUCV1rHuvWrqdf7/7UqlaH/HkKUqFMJR64/0FC94Z6P+h4rjUHcH4vD3p+MGVCyhKUuwDNb2zJil9XeDli91KTx9CXhtGhze0UL1SCHFlyMWP6DC9Hmzh/OKc8JZMxPr9lZGpRSKW9e0L5Zu631Kxdg4Y3NmDpj7+6XW9gn+f48YefGfrKS5QqU5KpH03n3o49GPvuO9RoVCVFx4qIiOC5J14kuFAw/xz7x2M5DPrubXJnyxVn2Q0h1Rhx2xMs+vs3AArnLsDcnu9y5My/DF30HifCT9GsTF2G3NKbAjkDidp/JsljvHn787St3IwxS6ew8cgOiuctzHMtHmTW/WNp9dFDhEec91g+iZk9cw6b/9qSonXPnTvHHW3vIiIigsHDBhFSIoQNazfw6sjX2b1rD1M/m+zlaOMKOxFG7bq1eOTxhykYXJAD+w8wdsw4WjW7lZXrV1CyVEm3202ZOJWz587y/KBnKV2mNLt37Wb0y6/x05Kf+X3dCnLnzp2meezevYe5s+dRp25tmjRtzE9Lfk7RdhcvXqR96w4cP36cUa+OpHCRQkyfOoPOnbryzQ9fc1PzZl6O/Irw8HDatm5HtmxZ+XjqRIwxjBg6gja3tOWPDavJlStXotueO3uWAgULMPzloRQvXpzTZ84wddJU7ux4N1/M+pw77ux0XeQxa+Ystm7ZRt/+fahatQqHDh/mtVde58aGzVi1biUlSoSk+xwAej/ahx++W8To11+hTJnSTPhwIre368TSFT9Tq3atNMkBUp/Hhx9MoGatmrRt34bPZnyeRlEn5A/nlPgPFbip1LhpQzaHbgDgs6lfuC1wt2zaytyZ8xk74U3u7XmPs12zRjSv14rpU6dzZ48OKTrW+LETsNZyb8+ujBvzvsdy+Pu/hCNo3evezsXIS3y9+ScAWldsQoFcgXSc2pc9Jw4C8FvoekoFFadLzduYuX92ovvPkSUbHau1ZPzvX/Dhyi9jlv937gSfd3+TBiWrs3T3Hx7Lx52TYScZ/Nz/GD1mFI/0fCzZ9Vf/vobdu3Yz99vZ3Ny6JQA3tWhGWFgY7439gPDwcHLmzOnVmGPr0q0zXbp1jrPshvr1qFejAV/P/YYnBvR3u93b771JweCCMd83velGylcoR9tWHZg3ez73P9DDq3HH17TZjew9tBuAaZOnp7jAnTd7Hls2b+G7HxfGFLOtb2tNo7pNGPLiEJatXOqtkBOYMmkqe/fsZdPWPylXvhwANWpUp3rlmkyaOJmnBjyZ6LZVq1VlwscfxlnWtl0bKpevyoxpM9K0wE1NHs88P5Dg4OA4yxo3aUyV8lWZOmkqQ0cM8Wrs0VKTw6aNm5j5xVd8NOlDej7QE4BmzZtRt+YNjBw+itnzZ6VJDpC6PACOnThCpkyZ2L1rt08LXH84p8R/qEUhlTJlSv4hXLRwMQEBAXTq3DFmWZYsWejUuSPr1q7j4sWLye4jdE8o77z+Lq+98wpZAgJSFXNycmTJxu1VW7Dk7985ecEZmQ3I7BzzzMXwOOuevnA22bdBMmXKRJZMWRJse+rCWed+4/3TcNjgEVSpWpnO99ydovUvXboEQJ68cUc48wXmIyoqCmutx2O8WvkL5Acgc5bEX6fGLm6j1b2hLgCHDx3xTmBJSMnPiztrVv9Bjhw54ozUGmNo1fpm1q1dz+FDhz0VYrIWLlhIg4YNYv6AA5QuU5rGTRrz7TcLr3p/WbJkIV++vEk+j96QmjziFyIApUqVJDi4IIcPXx/PxcIF3xEQEEDnrldeOGbJkoUuXTuzZPGPKfq97CmpPaeu9efK0/zhnPIY43xUr69vGVn6+Knwczu2/U3J0iXImTNHnOWVqlYkIiKCvbtDk93H808OpsNdHWjctJGXoryibeWbyJMtF19tutIXuGDrLxw/d5LRbZ+mRGBRcmfNSdtKzehc81YmrJqZ5P7OXTrPrI0/8HCDu2lSug45A3JQMbg0Q1v3YfPRnSzfs86r+az8bRVffjaTN8e9keJtWrRqTrny5Rj20sts37ads2fPsuyXX5nw/kQeevSBZN8y9JbLly9z6dIldu3czVN9B1C4SGG6pLBoj7biV6ftpFLlit4I0SsyZ85MgJsXdlmzZQNg65ataRbLtq3bqFa9aoLlVatVYfu2pHuio0VFRREZGcnRo0cZPfJVdv69iz59H/d0qEnyRB6xbd+2nX/++ZdKlSt5IrwUSU0OW7duo3SZ0gneialSrQqXLl1i967dngw1SZ5+LnzFH84p8R9qUUgDJ8NOki8wX4LlQUGBMfcnZfYXc9m0YRPLpyz1RngJdKl1G/+ePcHPO1fHLPvvXBgdpvRh2j2jWfOkU9BG2SjeXDaV8b9/waAK3ZPc59PfvMaoNk8xp+e4mGXrDm6h26cDiYjyzMVy7ly6dIkB/QbyxIB+VKhUIcXbZc+enR9+WUjPbg/QqPaNMct7PnQ/Y66iUPa0m2+8hQ3r/wSgbPmyfLvoa4ILJRz5SMyZM2d48ZnBVKpciQ6d2nsrTI+rULECp0+fZvu2HVSucuWP3ZpVawA4cSIszWI5cSKMwMDABMuDgoIIC0tZHINfeIlxY98FIHfu3Hzy+XRatmrp0TiT44k8okVGRvJE36cIDi7IAw/18lSIyUpNDmEnTrjdNn9QUMy+04onnwtf8odzylMMRtOE+ViGGsE1xuQ0xuwxxlhjjOeaWL0o7EQYw158mUHDXyC4UMK3mz2tcO4CNCtTj7mbl3DZXo5ZXiBnPiZ3HUV4xAUe+WoId01/kneWf8LTzXrSr8l9ye73xZaPcHeN1gxf/AF3TnuC/vNGEpQjH5/dN4YcAdm9ls+4t97l/PkLPPPiwKva7sKFCzzY4xH+/edfPpr6IQt/XMDIV0cwb9Y8nn3yOS9Fm7yJUyfw04olTP7kY/LmyUOndnexL3R/iraNjIzkofsf4cjhI0z9dBJZ0vgt8dToem8XChQswOMPPc7mv7bw33/HGfPam/y23BmNTi9v0aZU/6f6s2LVcubMn82tbW7lgR4P8t233/s6rGs24MmBrFq5iinTJxPkKhBFUkPnlKTW9fMXzjNeBlI+3OUh+QLzcXD/oQTLw1wjt4FBCV/xRnttxBgKFylEx7s7cOrkKYCY3rDTp86QLXs2cuXy3MVOd9e8lcyZMvPVxrjTFvVtch8lAotQf1yXmN7Zlfv+JLPJxPMtH2ba/MQvbKgYXJonmvZg4Dev88WfV/qw1h/axu/9P6d7nQ5MWpP4RWrX6sD+g7z12ljenfAOFy9e4uLFSzH3Xbx4iZMnT5EnT24yZ044zdmMqZ+yYtkKNmxdS5lyZQC4sVkT8ubLy1N9B/DgYw9So2Z1j8ecnEqu0cv6DW6gdZvW1KhQk7FjxvLOB2OT3C4qKorHH+rD0p+WMevrmVT3QeypERgYyOezPuPxh3rTqG5jAMqWK8PgoYMYOWwURYqm3XRzQUGBnDyZ8F2XsLCwFP8hDgkpTkhIcQDadWjLrTe3YdALg2nXoa1HY02KJ/IA+N+gIUz+eAqTpk7klltv8WSIyUpNDoFBQW6nrzrhGmnMnz/tiipPPRe+5g/nlPiP62vYIxWMMXWBp4FhaX3sSlUqsT/0AOHhcafC+nvbTgICAihTrnSi2/69fSdb/9pGlZCaVCpWnUrFqvP+W+MBqFqiJv0efMKjsXat2YbNR3ey9Vjc/rMqhcoSeuJQTHEbbcOh7WTNHEBgnryJ7rNKobIA/Hl4W5zle08c5OT5M1QoWMpD0ccVujeUCxcu8NgDvSlduGzMDeC9se9TunBZtmx237u5dfM2AoMCY4rbaHXrOxdo/b39b6/EfDUCA/NRtlxZ9uzem+y6T/cbyNxZ85j66WRa3Nw8DaLzvBubNuGvHRv5c+t61v21lj+3biAgIIAcOXJQp27tNIujStUqbN2yLcHybVu3U7lK5WvaZ916ddK05xM8k8fro9/grTFv89Y7b3Jfj+TfyfG01ORQtWoVQveGEh4e9+LX7Vu3kzVr1jgXSnmbN84pX/CHc8qTTDr4l5FliALXGJMZ+Bj4AZib1se/td0tREREsGDutzHLIiMj+WbOAurVq0s214Uy7ox8Yzhzfvgqzq1rD+eq31kLv+CFYZ57u7xW0UpUKlSGWRsTTjr/z9kTlM5fnHzZ484qUDfEmcP37PnEP/zgn7MnAKhTPO58v2XzhxCYIw9Hz3jnQwdq1KrBgsVfJ7gBdL2vCwsWf03ZeAVstEJFCnEy7CR7du2Js3zdGueCuKLFinol5qvxz7F/+HvHTsqULZ3keoOf/x/Tp3zC+I/fv676bt0xxlC+QnkqVa5IeHg40yZPo1v3bml60V/729uzZvUa9u658sJiX+g+Vv6+kva3t7vq/UVFRfH7byspW66sJ8NMVmrz+OC98QwfOoIRI4fTp19vb4aaqNTk0K5DWyIiIpg7+8qfhMjISGbPmsMtrVsl+XvZ0zx9TvmKP5xT4j8ySovCAKAycHWXm6fQgnnO2+4bN2wC4OdFv1AguAAFCuanSbPG1KhdnU6db2fo8yOIjIikZOkSTPt4BvtDDzBgwIA4+2pUvSkhJUOY/Z0zX2z1WtUSHO/3X1cCzly6nuyj7FzzNiIuRzLnryUJ7vtk3dfcVaM1X3Z/i/ErvyTs/CmalKpD78bd+G7br5wNv1Lg/t7/cw6eOkbXGU5uq/dvYvPRnQxr3Y982fPEfNDD0816curCmQTtEJ4SGJiPZs2bur2vZMkSMfft33eAOlXq0eP+HtxwYx0A7rv/XsaPG0+XTt145sUBhJQI4c91fzLm1TepXbcWjZo09ErMibmvcw9q1alF9RrVyJM3D7t27uKDcR+SJUtm+j/tzIG74tffuP22Tjz73DPUb1oPgLFj3uH9dz7g/gd6UK5COdasvjLfcMGCBRMt8L1p3pz5ADEXyy3+YQkFgwtSMLggzW5ynpN82YNofWtrZjf9Kma7YS8Np3bd2hQoWIA9u/Yw7u1xZAkIYMQrafumzEOPPMiE8R/R5a6uDHt5GMYYXh72MiElQnjksYdj1tu3bz/VKlbn/vt70LiZM/vJqBGvcCIsjMZNGlGkcGGOHjvG9CnTWfvHWqZ9OjXd5nFLq9a8NGQwg4cMAuCrmbN4buDz3Hpba5q3bM5q18V+AHnz5qFK1ZR9eE1a5hD/uahdpzadu3bmuYEvEBERSenSpZj40SRC94Yy9ZO0/SCX1OQBsHzZcv797z+OHT0GwLp1G8jl+hCXu+6+M13mkV7PKfEffl/gGmPKACOAl621ocaY0p4+xqPd477SfPHplwCnAJ23yJks/J2P3uK14W/w2ogxnD51mqo1qvD51zPImzVu/21k5GUuX75MWsuSKTN3Vr+FX3av5nh4wh6q9Ye2cse0/gy86QFG3vYkebLl5MDJo4z9dToTVn7JgDJd4uwrc6y5baNsFF1nDODJpvfTo25HnssbzInwU6w9uJkxSydz6LTnPpXtWlhruXz5MlFRUTHLSpUuyZLli3ht5Bu8Mnw0x/87QfGQ4vR6uBfPvjgwzS9qqt/wBubNns/777zPpUsRFA8pTrPmNzLw+YGUKl0yXh5X5uhdssj5KNsZ0z5lxrRP4+zzvvvvZcLk8WmXhMv93XrG+X7AE84FgE1vasoPP30HkOD5AGfE+oVnXuTff/4luFAwt3fqwEvDBpM/f/60CdwlV65cfL/kO55/5gUe7vUI1lpa3NyCN99+I+4nw0U/H7HmTK5dtzbvj3uf2TNncerUaQoXKUzNmjX4cekSmtzYON3mERUVFef5WLJoCdZaFi9awuJFcV8QN7upGYt/TpuPHk7NcwEwcfIEhv1vOCOGjuDkyVPUqFmDbxbOp07dOmkSv6fyGDniFZb/ujzm+4/Gf8RH4z8C4Hxk2n20uD+cU56kWRR8y6SHCeu9yRizCCgG1LXWRrgK3L3AB9baBB//ZIx5DHgMoHDhwvU+/cK7n+l9/uwFcuT23iwCAFuP7kl+pVQqkq0ARy8e99r+qxUp77V9Rws/G07O3N79dLK0+HV37mw4ubycR1rMIH7u7Dly5fZu60FaPB9nz54jt5fz8DZ/yAGUR3qSFjk8++yzrFu73idVZraQvLbYEw18ceg4Ql/8aZ219gZfx+ELfj2Ca4zpAbQGbrLWRqRkG2vtRGAiQK26NW2NRt59W+SvVdvw9jHuHzPSq/sHGFShO6/u/Mxr+99+93yv7Tva2t82xLQoeEtafGrbHyvWxbQoeEsmk3DmCU9bveIPGjat79VjZE6DPFYuXxXn7eTrkT/kAMojPfGHHJKjEVzf8tsC1xiTDXgb+A44aoyJHgIs7vqaz7XsP2tt0p+0ICIiIiLXDX+eRSEHzpy37YGdsW5LXff3cH3/iC+CExERERHv8NsRXOAc0MXN8mBgPM6UYZOBTWkZlIiIiPg/dSj4lt8WuK6e2wQfjxVrFoXd1lrPf3yWiIiIiPiU3xa4IiIiIr5gjC4y87UMV+Baa0NJm9mBRERERMQH/PkiMxERERHJgDLcCK6IiIiIdxm1KPiYRnBFRERExK+owBURERERv6IWBREREREPU4uCb2kEV0RERET8ikZwRURERDxMA7i+pRFcEREREfErKnBFRERExK+oRUFERETEw3SRmW9pBFdERERE/IoKXBERERHxK2pREBEREfEgY9Si4GsawRURERERv6IRXBEREREP0wiub2kEV0RERET8ikZwk5DZZCF3QD6vHiOTyez1Y+wa9K1X9w+wesVadnX23nHqjb/Ha/uO9kTRu+j34ZtePcaq3jO8un8Ai+VyVKRXj3Eu6qxX9w8QZS9zJuKUV4+Rx8s/ewAWuGwve/043pQWOVhrvbp/5xgQ6eWfjcwms1f3Hy0tHi9v83YOfvAQSSqowBURERHxMHUo+JZaFERERETEr2gEV0RERMSjjC4y8zGN4IqIiIiIX1GBKyIiIiJ+RS0KIiIiIh6mFgXf0giuiIiIiPgVFbgiIiIi4lfUoiAiIiLiQcaoRcHXNIIrIiIiIn5FI7giIiIiHqYBXN/SCK6IiIiI+BUVuCIiIiLiV1TgesGhg4d45qlnublpK4LzFiZ3QF72he5LdrtPp39G7oC8id6OHT2WBtE75s2Zz31de1ClfHUK5i1MnWr1GPbScM6cOZPstsP/N4KO7e6gZJHS5M6aj08/+SwNIoZP7n6d7U997/b2caeRADQqUZs3bnuOxb2m8Ge/+SzuNYVhLfuTP0e+ZPd/Z5VbEt3/9qe+p2DOIK/ltmLZbwRlD05wK1W4XLLbHth/kD4P96N6+doUDSzBDdUbMmrYaM6dO+e1eAEOHzzCSwOH0KFFJ8oUqEDRnCU4sO9AgvUuXLjAy4NHUatMPcrkL0+HFp1YuWJVio4RHn6eoc8Pp065GygVWI6W9W9hzpfzPJ1KAtf6Mx7t8KHD9HmkL2VDypM/V0GqVajBsJeGey9gN1KTw3//HafPI30pVbQMBfMUokWTlvy4+EcvR5zQ/Dnz6d61B1XLVyf4Kn9PAWzftoP7u/WkVNEyMduPf+9DL0ft3sEDB7nvnu4UKVCMwvmL0q3LvRzYn/DnxZ2cAbnd3jb+ucnLUSeUmjxie/ONt8gZkJtWzVt7Icq0YYzx+S0jUw+uF+zevYe5s+dRp25tmjRtzE9Lfk7Rdre1u42fl8f9I2Gtpcud91CmTGkKFyns+WAT8e7Y9wgpEcLwl4dSLKQYm/7cxOiRr/HrsuX89OsSMmVK/LXRhPETqVGrBm3ateHzT79Is5hH/PIBubPmjLOsdtHKDLrpcX7e6xRM3Wq0I2dAdib88QUHTh2lVGAxnmjUg6al6vL9Tz8kuf+loX9wz8wBcZYZ4MOOwzlw6ij/hYd5NB93Xn97NHXq1Yn5PkuWpH+Ez507x53t7iYiIpKXhr1ISIkQ1q/bwGsj32DP7j1M+XSS12IN3RPKN3O/pWbtGjS8sQHLfvzV7XoD+zzHTz/8zJBXXqJUmZJM/Wg693Xswdh336F6oypJHuPhbo+ybs16Xhj2HOUqlOW7r3+g/0NPYq2l8713eSMt4Np/xgH2he7jlua3Urp0KcaMfYNChQuxP3Qfu3fv8Vq87lxrDhcvXqR96w4cP36cUa+OpHCRQkyfOoPOnbryzQ9fc1PzZl6O/IpxY9+jRIkQhr08lOIhxdj45yZeTeHvqfXr1tPh1o40vakp7094j3z58rJr127OnfXuCz93wsPDaXtre7Jly8rEKR9hjOHloS/TpnU71qxfRa5cuZLdR4+ePXjk0YfiLKtQsby3QnbLE3kA7N2zl9dHv0GhQsFejlj8mQpcL2ja7Eb2HtoNwLTJ01P8hyM4uCDBwQXjLPttxe+cOH6Cl4YO9nicSflq3sw4sTS7qSlBQUE89nBvfl22nBYtmye67eH/DpApUyZ279qdpgXu7hP7EyzrUr0NlyIj+G7HMsApgsPOn4q5/49DfxEadohPu4yhfImykMRAQ9j5U3G2BahXrBpBOfLx3qpPPZNEMipWqkj9hjekeP3Vv69h9649zFnwFTe3bglAsxZNCQsL4/2x4wkPDydnzpzJ7OXaNGrakL9CNwDw2dQv3Ba4WzZtZd7M+Yyd8Cbdet4DQONmjWhRrxXTp07njh4dEt3/6t/XsPTHZbzz0Vvcc39XAFrc0pwjh44w6qVXuLNrJzJnzuyFzK79ZxzgqX5PU6xYUb77cSEBAQHOwpuaeiPMJF1rDvNmz2PL5i189+PCmGK29W2taVS3CUNeHMKylUu9FXIC8X9PNXX9nnr84d4sX7ac5on8noqKiuKxqWpj2QAAIABJREFUB3vTvGVzvph95R2mm1rc5PWY3Zk6aSp79+xl45YNlCvvvCtTo0Z1alSpxeSJU3hywBPJ7qNY8aI0aNTA26EmyRN5ADzV/2nuubcrO//eSWTkZW+GLH5MLQpekNSowdX6/JPPyZo1K126dfbYPlMifqENUPeGugAcOXQkyW09mX9qZM+SjTblm/HL3tWcungWIEGBCvDXsb8ByJUjZaMLsd1R5RYuRUawcMfSVMXqLZciIgDIkzdPnOX58uUjKioKa63Xjp2S82DRwsUEBATQsXPHmGVZsmShU+eOrFu7josXLya67fo16wG4+daWcZa3bN2CY0f/YZ3rfm+41nN8z+49/Lj4J3r3e/xKcesj15rDmtV/kCNHjjgjtcYYWrW+mXVr13P40GFPhZispH5PHU7i99TyZcvZsX0H/Z/u57XYrsbCb7+jQcMGMUUhQOkypWncpBHfLvjWd4FdJU/kMfOLr/hzw5+8/MoIL0WZhpzJcH17y8DSRyUibp0/f555c+bTpn0b8ufP7+twWLF8BQCVqlT0cSQpc0u5JuTOlpP525LuDawfUgOAsNMnr2r/2TJnpU2FZiwNvVJAe9tjD/ahQM7ClC1WkUd6Ps6B/QeTXL/FzTdRrnxZhr/0Mtu37eDs2bP8+styPvpgIg8++kCK3zL0lr+3/U3J0iXImTNHnOWVqlYkIiKC0N2hiW6byTU6G5A1bqGYNVtWALZv2eHZYD1g1e9Oq0z2HDm4vU0n8ucqSEhwSR594DGOHz/u4+hSJnPmzG6L86zZsgGwdcvWtA4pjt9S8Htq5W/O83DxwgVaNm1FUM4ClClejucGPM/58+fTJM7Ytm7dRtVqCdtxqlStwvZt21O0j0kfTSYwV34K5A2mbet2/LbiN0+HmazU5hEWFsYLz77AqFdHpYu/eXJ9U4Gbji34+ltOnz5N9/vv83UoHD50mFEjRtOyVQvq1qvr63BS5I4qrfgvPIxfQ/9IdJ1cATkYfNPj7Dq+nz2HQq9q/7eUa0yebLmYt9X7F9fkzZeH/k/3ZdyHY/nmh7k8N2ggy375P3v3HRbF8QZw/DsUG4KAIiLYKzYEu6IgijXV3lBjS34xpqgp9paoiYklmmg09sQWa2Ji7FGxYOxRsUvRxA72BuzvjwMEqQfsHej7eZ574GZ3Zt/luN25uSk7ae7bkuvXrqeYL0+ePGzYtp6YmBjqeXpTrFApXm/ZhuYtmzFp6kTd405LREQkBeyTDvCzd7CP356SsuUMrUQH9x9OlH4w6CAAkankNZf//rsCwLt9+1O2XBlWr1/F2PFj2LhhI2+0akNMTIyZI0xbufLluHPnDqeCE3+A2L9vPwC3bunfFz0l6b1O/fefoXW3Z9e3aNLUj183rOXDQR+wcN4iegX0MVW48SJuReDgkHSQqoOjQ6rvgTidu3Ri6vTJ/L7xN2bMnM6tm7do6d+anTuS7/eul8yex9BPh1O2XFkCenTTIzwTM/8AMxlkloMopYYAXkANoBQQqmlayVT2rwN8AdQBNGAP8JmmaUf0jzbzlixeglNhJ5q3bGbWOO7du0fHtp2xsrJi1pzvzRpLehW2caReseosPrKOaC35SoOlsuCblp/inL8gnX8ZRHNr4yrub7g3TbMCnVWqVa9GterV4p83aNSA+t71aOLdnB++m8PwMcn30X706BG9uvXlxvUbzJr3PW7FXDl04DCTxn+NpZUVk6dP0j12vfg0bUS5iuUYMXgkBeynUrZ8Gf5Y9ydrf1kHZJ+uMgnFVWAb+ngzZfpkAHwb+2BXwI6eXd9iy6YtNGth3vd7Wjp0bs8XY8fzdq+3+W72dxRxKcL8H+eze5ehxdBcf/d79+7RKfY6NTON61Tc69CxS0eGjx4GQEOfhkRHRzNq2GhOBZ+monsF3WPOKnMXPhss2sAbXnmtNTWr12bMyHFs3bHZjJGl3+7A3Sz5aQl79ge+9BUzkTWy3x0gdeMBP+A8kGozgVKqLrADQ0V4JDAKKAfsUkpV1TnOTLvy3xW2b/2LDp3apzlSXk8PHz6k/ZudCLkYwtr1q3F1czVbLMZ4taIflhaWrEmhe4JCMbHZYOoV86T/+nGcuRFiVPlO+RyoV9yT9af/SrECrTcPTw/KlivD4YOHU9xn8YKfCdy5mxVrl9KxS3saNKzPgI/6M27iWObPWcA/x46bMOKk7O0LcDsyab/ouNZXh9iW3ORYWVkx5+dZ5MuXj1cbv4G7a1Umjv6KIWM/A6BwkcL6BJ0JcV+7Nm6SuN9wE38/ALNM62Qse3t7lvzyMzdv3qKuVz1KupRi8YLFDB05BIAiLkVMHtPDhw/pYMR1KsXXoanhdThm4tfBwcGeiIiktzRDi2jK74GU2Nra0qJlcw4eOJgV4aVbZs5jwLvv0+Ot7ri6uRIZGUlkZCRRUdFER0cTGRmZan98IZKT0yq4ZTRNK6hpmj+Q1kiGb4EnQCNN06ZomjYFaIShJfcbnePMtGVLlhMdHW3W7glPnz6lW6fuHD54mFW/rqRK1cpmi8VYb7g3Jfj6eU7fuJjs9jF+A2hZvhEDN0xkX7jxDfqvVvTDysKStSbonpCW1Fo7Th4Pxt7BnlJlSiVK96plmGrszKmzusaWlgruFQgLCefBg8T9Hs8En8Xa2pqSZUqmkb88W4I2sj94D38d2MLBs0E4x1Zsa9dL/2wTpuJeKfVpz7Jjq3NyGnjX55/TRzly8hAH/znAkZOHsba2Jm/evHh6VTdpLAmvUyt/XUnldFyn0n4dTNuC6F7JneCTSfuongo+RUX3ihku19QtoZk5j1PBp/lx9lyKOrnFP/bu2cv+oP0UdXJjziz9pjTURTYYX/ayN4TnjKtpLE3T0jVRpFKqLFAL+EXTtMsJ8l8GfgGaKqVM38xghCWLl1KlapVEX0ubUkxMDL2692HH9p0sXfkztevUMkscGVGlcDnKFSyR4uCyTxv2oV2V5gzdPJmtF/Zm6BhvuDfh1PULnLph2rlLEzp88Ahnz5yLHzWeHGfnwkRGRHLhuTlW42YYcClq3reBf6umPH36lPWrn42wjoqKYt2q36hRw4vcsQOX0lKsRDEqVDJ8pTxv1gJ8mjaiZOmSeoScKbXr1sK5iDNbN29NlL55o+F/NbXXMrtRSlG2XFkqVCzPgwcPWDB3AZ26djLpwMWYmBh6d+/DTiOvU/4t/MmdO3fS1yF2sYqEc02bQutXWrM/aD8XLzz7QB4aEsrePfto/Wpro8u7c+cOG/74k5q1amRlmGnKzHn8ueWPJI+q1apSuXIl/tzyB2+0fUPv8MULJkf1wTVC3FUuudrLPqAXhn68v+sVwJpVawE4fMjQOrjpz80UcipEIadCNIyd87JAHgf8m/mz0ntForxHDh3h5ImTTPhqvF7hpemj9wexZtVaPv5sMDY2NuwPetbP1NW1KK5uroSFhlG1YnU+G/Ypfr5N4rfv2hnIjes3uHr1GgCHDh6Ov+m9aYKL1OvuTXgaHcVvp7Yn2danRnve8mrLyhMbCY38F48iz1oV7GwST6V1fMB61gZvYfiWqYnSKzmVoXyhUkzcOVufE0hG3x7vUKJkcTw8q1GgQAGOHf2HKZOm4VLUhbf79wUgLDQcr0q16BbQlRr1Da1oXQI68f23M+nwemcGffoRbsXcOHzoCF9P+IbqXh7UrV9H17jXrzG8xY4dNnzlu23jdgo6FcSxkCP1G9ajavUqvN7uVUZ+MoanT6MoXrIYC+csJjwknIEfJV5Uo14Vb9yKu/HLH8vi076dNAO34q4UcSnCpfDLLPhhIZfD/+XXbat1PS/I2HvcysqKsV+M5u3e/+P9dz/k9Tdf5fy5C4wdOY6GPg1TnV86u5wDwKhho6nuVZ2ChQpy4dwFpk2ehpW1NWO+GGXS+AcmuE7lS+U6Ve2561TBgo4M+mQgX47/CltbW3waN+LQwcN8+cVXdAnokmiaK1N4q09PZs38gQ5tOjJy7EiUUowbNQ63Ym70TrB4Q1hoGJUrVCUgoBt1vQ3v3amTp3Hm9Fl8fBvhUrQIYaHhTJs8jatXrjJ/0dxsex5Nm/gzdMQQhg43dG1p5JN0DmJ7+wJERUUnuy27U5i+BV0k9qJWcIvG/ryczLa4tGQ7aSml+gH9AJydnQkKzNgAooBO3RM9/2jAQAA8PDyYPNUwuCQ6Oponj58kOcaM6TOwtLSkbJlyGT5+IhmY6nT9OkPFZNLEr5k08etE27p3706Pnj24cuUK0dHRhIdc4v69+wQFHgDgs0FDOXr0aPz+s2fOYfbMOQBs3Za4xSS9Brikb1UqC6VoV7EZl65epqt9E3iu29cb5VsA0K5yc9pVbp5oW0hoCAPuP1vi08rCksr5SiU5tnf1ekTHxFA4Il+644pzcE/Gxjfa2RZg9S9rmTnjBx4/eoyjoyN169alR8/uhJwJJ+RMePzr8fjRk0THmTptKgsXLmLU0LHcvn0bp8JOtGjRgi7dunB4X8b6Gmrp/Kfq2/WdRM8/+9AwoKeaRzW+mWL4v+rbux9WWPPF8Ancu3ePMmVK88WEz3ErWozj+4Lj8z68/4g7EXcTpYWeCWP+zAXcvHkLm/w21KpVk8GDB3P36gOOXw0mLRYq4wtBZPQ9XrZ0eT4b8hnLli5j8YLF2Nra4tvYlz59+7B/94EMx5MRGT2HE8dOsnDeIiIjI7G3t8e7oTc9evbg7MnzGIZIZICJrlP7Y69Tfr5NuHUjgqU/LWPa5G9xdHSkXft2BAQExO+TERmt03zxxed8/91M3grohaZpeHl5MmToEI4fORG/T9y5PHr8mH2BQQDERMXwd9DfrF65mvv372Njk4/KVaowddpULLGK389U0nseMTExhIWEpxrfndt3iY6ONvk5iBeD0nOidz0ppY4D+ZObRUEpNQIYCzTRNG3bc9v8gK3AR5qmTX0+b0JeNby0XUE7si7oZAQF/k0db52//jfBaxwUeIA63vr1eazxfUfdyo4zwKUN0//Tt+Vv3zuLdS0fDJXouBZcvTyKeaRr+QDH9wWnuVRvZtlaJ52iLKuZ5D2uM1OcgynuRfsDD1Bbx+sUgGUmPjSl177AoPgW3JzKFOfQoE5DDh08ZJZmVJuSDlrF4Y3T3lFnh/quOahpWvYbkGACL2oL7oPYn8l14Mvz3D5CCCGEEFlKuiiYV44aZGaEuBkWkuuGEJeWXPcFIYQQQgiRw72oFdy4zmL1ktlWF0NvL9NOECiEEEIIIUzihazgapp2DjgAtFdKxQ04I/b39sA2TdOumCs+IYQQQrzYzL1M78veRSJH9cFVSgUAJWKfOgG5lFLDY5+HapqWcATPB8B2DCuXTY9NG4ChUj/IFPEKIYQQQgjTy1EVXKA38PxEkeNif+4A4iu4mqbtUUr5Ap/HPjRgD9Be07SjCCGEEELo5CVvQDW7HFXB1TTN18j99wJN0txRCCGEEEK8MF7IPrhCCCGEEOLllaNacIUQQgghsj0Z5GV20oIrhBBCCCFeKNKCK4QQQgiRhRSykpm5SQuuEEIIIYR4oUgFVwghhBBCvFCki4IQQgghRBaTLgrmJS24QgghhBDihSIVXCGEEEII8UKRLgpCCCGEEFlMuiiYl7TgCiGEEEKIF4q04AohhBBCZCUF0oBrXtKCK4QQQgghXihSwRVCCCGEEC8U6aIghBBCCJHFZJCZeUkFNw0WOjdyKxMcI0bF6Fo+ELfwtm7FB749X7ey4xzZe5zA1/Q9To+Nn+laPsBrWkO+3bhM12PM8R+ra/lx9H5v3Hp0TdfyAaJjnup+HMfcTrqWjwZomr7HQO/y446h73E0E5yHZoLjKKRyJnI26aIghBBCCCFeKNKCK4QQQgiRhRRKuiiYmbTgCiGEEEKIF4q04AohhBBCZDFpwTUvacEVQgghhBAvFKngCiGEEEKIF4p0URBCCCGEyGLSQ8G8pAVXCCGEEEK8UKQFVwghhBAiKykZZGZu0oIrhBBCCCFeKFLBFUIIIYQQLxTpoiCEEEIIkdWki4JZSQuuEEIIIYR4oUgFVwghhBBCvFCkgquDS+GX6NKxK0UKFsXZ0YVO7TsTHhae7vyngk/RtVM3ihUpjqNtITwqe/Ldt9/pGHHyLl+6zKAPBuPn3QQnO2fyW9sRGhKa7vz/Xv6X//V5l9JuZXG0KUTlclUZNWy0fgGnIGjPftq/0hH34lUo6VQWv3r+/LxwaZr5vCrUwimvS5LHH79u0DXeKgUrsO61H5M8fm75baL9StkVY1TdD1nWagZLW05nWO33KGJTOM3yi9o406dKJ6b5jmZZqxnMb/Y1w2q/R0k7N71OKV7gjt282uQNSjiWoYJrZfr3HsC1q9fTldc5n2uyj+NHj+sa83+X/2Pk4DG84deO8oWrUNy2LOGhlxLtcynsMr07vk29So0o51QZjxK1aN+iM9s2/pVm+b/8tIritmVTfKT372OsNavW0qVDN9zLVqGQnTOelWswatho7t69m2be0cPH8FqrNyhepCT5cxXgp0U/6xJjWtauWkfXDgFUKlsFJ7sieFauyahhY9J1DuFh4fTr9Q7uZQx5q1eqwdiRn3P//n0TRJ5U3PW2sbcfhewKY2Ntm67r7d27dwno3J2qFT1wKuBM0UJuNKrny9Kfl5kg6qQyc//LZ50/2cfRI8d0jlofSimzP15m0gc3iz148ICWzVqTO3cuZs/7AaUUY0eOpYV/K/Yf2oeNjU2q+Q8eOESrZq1p6NOQ7374jgIF7Dh39jz375n+onv+/AVWr1yDp1d16nvXY+vmbenOGxoSSlOfZpQsWYJJU76isHNhwkJCOX/+go4RJ3Xin5O0a92RGrW9mPzd1+TNl5ff1qznw3cG8uTxE97q1yPV/I39fflk2OBEaWXLl9Ez5Hiz/1nC2ciQ+OcxMdHxv7vYFGaC96eE3rnM5EM/Yqks6FThNSY0+IQN2zfC05TLre5UiaqFKrItfA8XIkOxsc7Hm2Vb8FXDoQwJ/JLzt9P/IcYY+3YH0fHVLvg29WHuktlE3Ipg4pivaNe6I5t3byB37txpltGpWwcC+nRLlFa6nL6vR8iFUNav/oOqnlWoXb8mO7cGJtnn/v37OBZ0YPCIj3BxLcK9O/dYsmAFPdv1YeTo4VSsXTbF8v2a+7J26y+J0jRNo1eHtyleqhiFnZ2y/JwAvp0yHbdiboweO5KibkU5duQY48dNZOeOXWzduRkLi5TbP2Z9P5uqHlVp0aoFS35K+8OiXqZNmU6xYm6MGjsSV7eiHD1yjAnx57ApxXO4f/8+r7Z4g6iopwwfNZRixd04eOAw48dO4Py58yxcMt/EZ2K43q5auRpPL0/qe9dn6+at6cr35MkTLK2sGPzJQIqXLMHjx49ZtWIVfXr25cb1Gwz48D2dI38ms/c/gG7du9Gnb69EaeXKp/z+ESIlUsHNYvN/nM/FCxc5euIwZcoabrxVq1ahqrsHc2fP4/2PBqSYNyYmhr69+uLr58Pylc8+ffv4+uged3K8Gzbg4uXzACyYu9CoCu4H/T+kaFEX/tjyO9bW1obERt56hJmqNb+sJTo6mp9WLSJ/fsPF1beJDyf/CWbFz7+kWcEtWNCRmnVqmCLUJC7d/Y8zEcl/IGhbtiXRWgxj903lftRDAM5EXGRWk/FULVuZFcGbUyx3179/80fI9kRpx26cYk7TibxauglTD8/LupNI4Ovxk3Er7sbCFfOwsjJcespVKEfzhq1YsmApb73dM80yihQtQs3apn096jSozaELQQAsXbA82QpuBffyTPp+YqI0vxaNaVDFl41/bqbPoJ4pll/QqSAFnQomSgva/TcRtyIYOOz9zJ9AClasWY6TU6H45w0beePg4EC/3u+wc8cufBunfN3590Y4FhYWnD933qwV3BVrliU6B+/Yc3i79//YtWMXPimcw749QZw/d561v6+mib8fAI18GxFxK4Jvp0znwYMH5MuXzyTnEMe7YQNCLhve7wvmLkh3BbdgwYIsWJz4PduiZXPOnT3HogWLTVrBzcz9L05RVxdq162td6i6U4DFy92AanbSRSGL/b7+D2rXqR3/5gYoWaok9erXZf1v61PNu3PHTk4Fn+b9D9O+CJhCai04qblw/gJbNm3lnf5vP6vcmsnTJ0+xtrYib948idJtC9gSExNjpqgyr7xDaU7fOh9fuQW4+SiCsLuXKeFSPNW8d5/cS5L2IOoh/96/imMehyyPNc7B/Yfw8WsYX7kFqF7DA8eCDvzx25+6HTezMvo+sLKywtbOFktLS6Pzrlqymly5rHm93asZOnZ6JKwYxvGq6QUYumWkJqN/k6yW2jn8m8o5PHnyBABbO9tE6QXsCxATE4OmaVkYZfpk9d/UsaBjoveaKWTm/idEVsseV6kXyMmTwVSq7J4k3b2SO6eCT6Wad8/uvQA8evQYnwaNsctrT4miJRn04WAePnyYat7sZN+efQDkyZuXV1u8jqNNIdycitO3Zz9u3rxp0lg6BXQEYMig4Vz59wq3I2+zeN5P7NoeyDsD+qWZf+MfmynuWArXAiVo0ai17v1vExpYoy+rX53N4hZTGejVl0J5HeO3xWgxRMVEJcnzNCYKOxtbrC2Mu7Hlt7ahuK0rl+6lXrHJDEtLS6xz5UqSnitXbk6dOJ2uMhb+uIhi9qUoWbAMbVq2Z9/uoKwOM1NiYmKIiori2tXrTJ04nYvnQnj9DeMqqY8ePuL3tRto0sIPe0d7nSJNXuAuQ+t0BffyJj1uVtq9azcAFdwrpLhP4ya+lClbhpFDR3Hq5Cnu3bvHju07mDljFr379UrXV+nZjaZpREVFcfPmTebNmceWTVt574P+Jo0hM/e/OD/+MBd7G0cK2jnR0r8VuwN3Z3WY4iUhXRSyWMStCBwckraCOTg6EBERmWre//41VC66d+nB2+/2Y9z4MRw6eJhxoz/n0qVLibotZGf//XcFgHf79qdT144M+nQgF85dYPTw0ZwKPs2OvdtN1gLkXrkiazeupkfHXsz/YQEA1tbWTJr+JW92eCPVvM1aNcOzpgfFSxTn+rUbzJ01jx4de/H9vOm079xOt5jvP33AmnMbOXHzDA+iHlK6QHHalWvFVwWH8NGOsdx+cpfL965Q0bEMlsqSaM3QNzevZW6K2RZFKUV+axsiHt9O9zH7Ve2MAn69sEWns4Iy5cpwcP+hRGnhYZe4euVqulr623Vug3/LphRxKcKlsEt8N3UmbVt2YMX6pTRoVF+vsI0yfviXzJ4+FwCb/DbMmD+VUi7G9RHeuH4zd+/co22XN/UIMUX/Xv6Xz8eMp3ETX7xqeJn02Fkl8Tl4prhfnjx52PzXn3Tt2J1a1evGp/fo1Z1vpk0yRahZ7ofvZzPoQ8N4AWtrayZN/oquAV1MGkNm7n8Anbt0omXrFrgUdSEsNJyp30ylpX9r1v/5K418GukRso5kkJe55agKrlJqCOAF1ABKAaGappVMYd8FQEodLNtrmrZSjxgzI+4r805dOjJy9AgAGvk0Ijo6mhFDR3Iq+BQV3SuaM8R0iTuPhj7eTJk+GQDfxj7YFbCjZ9e32LJpC81aNDNJLOfPXeCtzr2pUKk8X0//kjx587Dht418POBT8uTOTbvObVPMO3HKF4met369JS0atebzERN0reBevBPOxZPPRh2fuHmGEzfP8HXDYbxSugk/n1rL+otb8Xatxf+qdWPJ6XVYKgt6Ve5AXkvDQK0Y0v8Va9uyLfFxq8u3h+dz5f61LD+fOH3796Z/rwFMGP0lfd7tTWREJIPf+wQLC4t0feD5bu70Z08a1KHFK83xqenHxDFf8dvWtbrFbYze/XvyartXuH71OquWruH93h8xbMSwVAeZPW/lktUUciqIX3Nf/QJ9zr179+jYtjNWVlbMmvO9yY6ble7du0entl2wsrJi5pzUZ5159OgRPbq+xfVr15kz/wfcirtx8O9DfPnFV1hZWTF1xmQTRZ112nZoQ+06tbhx8yZ//PYHgz4cjKWlJb379Uo7czYxd+GP8b838IZXXmtNzeq1GTNyHFt3pDyuQIjk5KgKLjAeuAUcAtL73V1AMmn7syyi5zg42BMREZEk3fDJNvWQCxY0DDTxa+qXKL1J0yaMGDqSo0eO5YgKrqOj4av0xk0aJ0qPG8xx9Mgxk1Vwvxg5AWtra5asXhzfStiocUMibt1i2McjaNPxzXS3JltaWvJam1cZO/xzrvx3lSIuznqGnsiF22H8e/8q5exLAhB86xyzjv1EgHsb/Es0BODI9ZNsC9+Dn1t97j1J36wbLUr40L1SW34KXsPWcH2/CmzXqQ3nTp9j5rQfmPrVtyileL3dazRp7sepk+nropBQftv8NG3RhCULs883Gy6uLri4ugDQtKUfHVp2YfYPc+j5Qdd05b965RqB2/fQ850Ak/WffPjwIe3f7ETIxRA2bPkDVzdXkxw3Kz18+JAO8efwe5rnsGj+YnbtCORo8GFKlykFGAZ52RWw4/3/fUDvvm9R1aOqKULPMk5OTjg5GWbcaNbcnwcPHjD002F0fyvAZGMhMnP/S46trS0tWjZn4fxFWRGeeMnktApuGU3TLgAopY4D+dPKoGnaT7pHlYB7JXeCTybta5Se1lf3Skn7LiVkkUOGZKZ9Hqbr+h18IpjKVSslucB71vRk1fI1XL92A+ciac8d+zxzffWUsF12Q8hfbA4LxMWmMA+fPuTGowhG1vmA6xE34rstpMbXrS5vV+vKmnMb+eXs7/oFncBnoz5hwOD3CL0YSiGnQhR2dsLb04c69WtluMzs/DVgNc+q7N9zIN37r1m2jujoaNp1aaNjVM88ffqUbp26c/jgYX7dsJYqVSub5LhZyXAOPTh88AjrNqyhcjrO4cTxkzg42MdXbuPUjB2gdvrUmRxXwX2eVw0vfl68hGtXr5nsQ0vFzjanAAAgAElEQVRm7n+pyc7v8RQpsMiJcb9ActQgs7jKrTGUgZ1SyiTn2vqV1uwP2s/FCxfj00JDQtm7Zx+tX22dat5mLfzJnTs3WzYl7ge5eZPhq5mc0i+udt1aOBdxTjLNzeaNhvOKG+VsCoWdC3P82In4UdNxDv19mDx58uBgxCCeqKgo1q5ah1sx1wxVijOjbIESFM1fhLMRFxOlR8VEEX73X248iqCErSseTu4Eh6TdGlq3iCfvV3+LzaG7WHDylzT3z0o2NvmoVMWdws5ObNu0nbOnz9GjT3JftKTu7p27bN6wBc+a1XWIMvNiYmL4e+8BihZ1SXeeVUvX4F6lIpWrVdIxMoOYmBh6de/Dju07WbryZ2rXyfiHDHOJiYmhd/e+7DTyHJydCxMREcn5c4lvKX//fRAgvhU+JwvcFUj+/PlxKqzPPMrJycz9Lzl37txhwx9/UrOWeaZqFDlbTmvBzYjbgC3wRCm1ExiuaZpuQ6/f6tOTWTN/oEObjowcOxKlFONGjcOtmBu9E0xeHRYaRuUKVQkI6EZd7zqAoYvC4E8HMfGLL7G1s8O3sQ+HDh5iwucT6RbQNdHUK6ayZpWhb+PhQ0cA2PTnZgo5FaKQUyEaxs5rWyCPA/7N/FnpvQIwTI809ovRvN37f7z/7oe8/uarnD93gbEjx9HQp2Gq82tmtd7v9KJ31750bdudXv16kidvHv5cv4nVK9bwzoB+5Iod1V8kvxv+zfxZvNowwfvq5WvYsH4jTVv44ermyvWr15n7w3yOHf6H2Qtn6hrzQK8+XH1wg/O3w7j/9IFhkFnZltx6GMFvFw0fGgrmcaBlSV+Cb50jKiaKsvYlaVuuJfv+O8yFyxcNkzDGWv3KD2wL38OMowsBqORYjkE1+nHxziW2he+hvEPp+H2fRj/l4p30r7pnjH+OHGfrpm1Uq25oGQvas5/vp87ivYHvUqvus4pJeNgl6lSuT9eArkyqOwGA76fO4tyZ8zTwqU8RF2cuhV3i+2k/cO3qdb6fP0OXeBP6fe2G2HM4AcBfm3fgWMiRgoUcqetdh8njpxEZcZuadWtQ2LkQ167eYPmiFRw5eIzPhn6aqKxS9hVo1+XNJPPm/nPkOKdPnmHE+CG6nw/AR+8PYs2qtXz82WBsbGzYH/R3/DZX16K4urkSFhpG1YrV6RbQjTreNeO379oZyI3rN7h61dBn+9DBw/EzD7zZNvXBm1lp4PuD488hn02+FM+hWkVPPhv2CX6+TQDo2r0LM6Z9T9vX2vPxZ4MoVtyNQweP8NX4SXh6Vade/bopHVJX6bne2uWxx7+ZP6u8DR9M586ex/6g/TRu0hhXN1du3bzJ6pVrWLNqLWPHj4m/xpmCMfe/pk38GTpiCEOHG/7fp06expnTZ/HxbYRL0SKEhYYzbfI0rl65yvxFc012DllFkUNbnl8gyhzz/WWFuC4KqQwymwjkAg4C9wEP4EPABmilaVqyw8WVUv2AfgDOzs41li5bYnRsV69e5fvvZnLw4EE0TcPLy5P+7/WnSJEi8ftcuXKFLp270qlzJ/r16xufrmkaK39Zybp1v3Lt2jUcCzrSvFkzArpnvE9eZl7hJo2bJJvu4eHB5KmT4/fx8/Nj2IhhifbZvGkzy5Yu4/Lly9ja2tLIpxF9+vYhb968RscRo2V8ztr9QftZvmw5oSGhPHnyBJeiLrR+pTWtX2kdP0epv58/fn5+DIm92J48eZL5P84nJDSEu3fukidPHspXKE/7ju2pVSvjLV2hdy6nuU+1clUo41qK/PnyY2VpxYPHD7l09TKHTh3h4WPDdHF5cufBt0ZDCto5Ym1lzZ0HdzkTepYTF4Kx02y4zbO5bnu/3oMzYefYddjQx9azggdeFZNv9bz74B4rNq9KM8YSdkXTc7qJhFwMYeqUaYRcDOHp06cUL16c1998nRYtmyfa78qVKwR06U6nTp3iB8js3bOXZUuXcyn8Evfv3yefTT4qV65M125dMvXVp5bO/6tmTVokm17NoypfT57E3j17Wb1qLSEhITy4/wAHBwdKlylFx04dKFOqLHlscicqy79ZUz7+NPEKed/PmMlvv65nybKfcHA0bj5iS2V8H8sunbtw9erVZLd1796dHj17cOXKFbp26UqnTp3om+A6NfCjgRw9ejTZvFu3pW+BgqSMv1J16dw1lXMISHAO3ejePYB27dpjk9+wgENISCiLFi7k5Mlgbt++jVNhJ+rXq0/Xbl2wtbVNtsx0yUSlJiPX2xPHT/DTTz9x7uw57t69i10BO0oUL0Hbdm2pWy/jFfWMnoUx97/uPbrTs6dhLPiePXtYumQZ4eHh3L9/HxubfFSuUoVu3brhnsH3+KBBgzl08JBZapkFyhbS6k7Sbx7r9NrUZsFBTdNqpr3ni+eFreCmkKcccAT4V9O0cmnt71XDS9sdtCvjQabDvsCg+BZcvcSg/4IGQYF/U8dbv684H0bpv1Txkb3HqV6viq7H6L1puK7lA7ymNeRXpe//7Rz/sbqWD3Bs3wmq1dW3T+jjaP3nlz61/5xRsyhkhGNufb+GDgo8kKgFVw9apj6Kp8/+wAPU1vk8LJTxC3sYa1/gfup667val8pwFTd9THHva1CnoVRwX+IKbo7qg5tZmqadBVYAZZVSOXcmcyGEEEJkaxbZ4JFeSqlWSqmdSql7Sqk7SqkDSim/BNsdlFI/KqVuKKXuK6W2KKWy9UjMFL/zVkqlvt5nKjRNC8toXhMIif1ZCDhjxjiEEEIIIcxKKfU2MCP2MQ5D3bg6kC92uwJ+A0oCA4AIYAiwXSlVXdO0S2YIO02pdeoMIWPdN7U0yjW3uK4JyXfcEkIIIYTIpJwwTZhSqiQwFfhY07SpCTZtTPD7a0ADwE/TtO2x+fYCF4FPgPdNEqyRUquILiJz45PMRillA0RrmvbouXRPoD0QrGnaebMEJ4QQQgiRPfQCYoBZqezzGoaxS9vjEjRNu62U+g14nZxWwdU0racJ40gXpVQAUCL2qROQSykVN2onVNO0xbG/lwM2KKXWAmd5NotCLyCa2FkShBBCCCFeYt7AKaCTUmoEhjpWCDBF07S4Na8rA8eTyXsC6K6Uyq9p2r1ktptVdu5KkJzewPOTqI6L/bkDiKvgXgG2AI2BrkBe4D9gOTBB07SkS60IIYQQQmSBHDQPbtHYxyRgKHAewzfdM5RSVpqmTQMceTZ+KaFbsT8dgBejgquUqgi4Y5ima3Fa+2cVTdN807nfFcD4pZGEEEIIIV4chZRSCdcLn61p2uwEzy0wLIbVU9O01bFp22L75g5RSn1rmjCznlEVXKVUdeBHwDNB8uLYbT7ABqCjpmm/ZVmEQgghhBAiI26kMQ/uTQzdOjc/l74JaAG4YJg1IbnVZxxjf0ZkNkg9pHuatNh5Y/8CKgDTMFRmE9qJobm6XVYFJ4QQQgiR8ygslPkf6XAije0xsfskt+pOJSAsO/a/BePmAR6FYenbOpqmDQT+TrhRMyyJthfQbzkrIYQQQgiRVdbE/mz+XHoL4FJsl89fAdfYb+oBUErZAa/GbsuWjOmi0ARYrWnayVT2CQf8MxeSEEIIIUQOpnLMILM/gO3AD0qpQsAFDIPMmgFvxe7zK4YGzJ+UUh/zbKEHBXxl8ojTyZgKrgOQ1moVCkMrrxBCCCGEyMY0TdOUUm8AE4AxGOp6p4CumqYtid0nRin1CvA18D2QB0OFt7GmaeHmiTxtxlRwrwJl09inMoZWXCGEEEIIkc1pmnYH6B/7SGmfWxjWEuhlqrgyy5g+uNuAV5VSFZLbqJSqhaEbw8bktgshhBBCvAwUhgqWuR8vM2POfwIQBexUSv0Pw8TAKKUqxz7/DbiLoQlbCCGEEEIIs0h3FwVN004rpdoCS4EZsckKOBb7MxJoo2laWJZHKYQQQgiRg6Rzmi6hE6MWetA07U+lVCmgB1AXKAjcBvYB82P7aAghhBBCCGE2Ri/Vq2laJIaFHqZlfThCCCGEEEJkjtEVXCGEEEIIkbocMg/uC8voQXZKqa5Kqa1KqVtKqajYn1uVUl31CFAIIYQQQghjpLsFVyllDawEXsEwqCwauA4UAhoDvkqpDkA7TdOe6hCrEEIIIYQQaTKmBXcIhnWHgzBUaPNomuaCYUULP2A/hsrvp1kdpBBCCCFETqEwzKJg7sfLzJg+uN2Bc4CvpmlP4hI1TYsG/lJK+QLHgZ7A51kYo8gkhWn+yfU8Ti7LPLqVHccCpftxlrScrGv5AEGBB1jire9xPKe317V8gPdd2/HuTH2n1f77vaW6lg9goaywy+Wg6zEexzzStXyNGN2PYaEsdS0fQAOitChdj5HLBOdhChqazuXrfwzxcjOmBdcNWJewcpuQpmmPgXWAa1YEJoQQQgiRU6ls8HiZGVPB/RewTmMf69j9hBBCCCGEMAtjKrhLgHZKKbvkNiql7IF2wM9ZEZgQQgghhBAZYUwf3LFAFWC/UmossBO4CjgDPsAIDAPNxmV1kEIIIYQQOYcM8jK3FCu4SqkYSLYHuAIWp5BeDniYWrlCCCGEEELoKbWK6E6Sr+AKIYQQQgiRbaVYwdU0zdeEcQghhBBCvBCUQroomJnRS/UKIYQQQgiRnUlfWSGEEEKILKakBdesjK7gKqVcgCYYFnTIncwumqZpMpOCEEIIIYQwC6MquEqpMcBnz+VTPBuMFve7VHCFEEIIIYRZpLsPrlKqK4a5bndhWNBBAQuBLsAcIAZYBvhlfZhCCCGEEDmHhVJmf7zMjGnB/R9wCWihaVpUbN+SEE3TlgHLlFJrgN+BpVkfphBCCCGEEOljzCwKVYE/NE2LSpBmGfeLpmkbgY3Ax1kUmxBCCCFEjqOyyeNlZkwF1xq4meD5Q6DAc/scBzwyG1ROdyn8El06dqVIwaI4O7rQqX1nwsPC05U3n3X+ZB9HjxzTOeqkLl+6zKAPBtPY249CdoWxsbYlNCQ0XXljYmKY9OXXuJetjGP+QtTxqsfa1et0jjipLZu28kqz1yhXrAJO+Z1xL1WZHp3f4tTJU+nK/3fQ37R5pR3FnUrgYu9KPc/6rFy+Sueok1qzai1dOnTDvWwVCtk541m5BqOGjebu3btp5n306BHDPhtOmeLlKWTnjF/DpgTu2q1rvIvbf8WZgRuTffzY5otk84xp8j5nBm5kUstP0nWMt7zaMOv1MQT2W8KZgRsZUK9bVp6CUdq+0h773AX5fFTy55aSKZOmYp+7IC0at9IpsrRt/nMLrzR5neIFS1HCqQx+9Zuxc/uuVPNcCrvEu70HUK2sF672JahdpR5fjJrA/fv3TRT1M4E7AnHIXSjJo0Th0mnmHTvic9q0akdpl3I45C7EkkXm+wIyM9fbhH5ZvhIba1vKlaygQ5SpexHuGeLFYUwXhf8AlwTPw4Bqz+1TFIjiJfbgwQNaNmtN7ty5mD3vB5RSjB05lhb+rdh/aB82NjZpltGtezf69O2VKK1c+bJ6hZyi8+cvsGrlajy9PKnvXZ+tm7emO+/YUeOYNvlbRo0biaeXJyuXr6RbpwBWrvuFFi2b6xh1YpG3Iqju5UGft3tTyKkQl8LCmTJpGk0bNmPPoUCKlyieYt6Nf2yka/sA2ndqx4+L5mCdKxeng0/z+PFjk8Uf59sp03Er5sbosSMp6laUY0eOMX7cRHbu2MXWnZuxsEj5s+q7/d5j44ZNfD5xLCVLlWT2zDm80boN23Zuplr159/CWWP01hnkz50vUZqniztDfd9h2/m9Sfb3KlqJ19z9uPs4/RWkDlVbcu/JA7ac30sXj1cyHXNGrVy+iuP/nDA6X8iFEL6eMBmnwk46RJU+C+Ys4tOPhtDnf70YNGQgmhbDP0eP8/DhwxTz3L9/nzat2vP06VOGjPoUt2KuHD54hC/HTeLC+QvM/WmOCc/gmS8nT8Czpmf8cysry1T2Npjz/RyqeFSleatmLPtpuZ7hpSkz19s4kZGRfDLoU5yLOOsQYdpehHuGeHEYU8E9DFRJ8Hwb0E8pFQCsBnwxDD7Tt2kom5v/43wuXrjI0ROHKVO2DABVq1ahqrsHc2fP4/2PBqRZRlFXF2rXra13qGnybtiAkMsXAFgwd0G6L1bXrl1n2uRvGfTJQD4c+AEAPr6NOH/+AiOHjjLpxapdp3a069QuUVqNWjWoWbU261b/yoCP3ks23927d3m373v0eac3E7+ZEJ/euImvnuGmaMWa5Tg5FYp/3rCRNw4ODvTr/Q47d+zCt7FPsvn+OfoPK5b9wsw53xHQo1t83loedfh8zHhWrFmmS7znb4UlSetQtSVPop7w++kdidKtLCwZ2/QDZgUtpWO19LdktlrYDw0NS2VhtgpuZEQkQz8ezvhJn9Onez+j8g4cMJj2ndpx9uw5oqNM3y4QFhLGsI9HMGbCSN4Z8HZ8up9/41Tz7d/zN+fPXWDlb8tp7O8LQENfbyIiIvhuykwePHhAvnz5Ui1DD+UrlqdWnZpG5Qm9fhELCwsunLtg9gpuRq+3CQ3/bARVq1WlSBFntm/7K4sjTNuLcM/ISi/7IC9zM6aLwnqgilKqVOzzicBtYAFwB/gVQ5eP4VkZYE7z+/o/qF2ndnzlFqBkqZLUq1+X9b+tN19gGZBaq2BqtmzawpMnT+jUpWOi9E5dOnLi+AlCLoZkQXQZ51jQEQArq5Q/361dtY4b12/w3of9TRVWqhJWbuN41fQC4L/L/6WY7/f1G7C2tqZt+zbxaVZWVrTt0JYtm7earDU6j1VuWpZryLYLQdx+lLhbRZ+a7bFUFsw9uNKoMrX42QnNZ9TQMbhXqki7jm2NyvfLspUcPXKMUZ+P0CmytP28cCkWFoqefXsYle/J0ycA5LfLnyi9QIECxMTEoGnmf13SK6PXOD1kNpa9u/eybMlypnz7TRZFZLwX9Z4hcqZ0/zdqmrZA07R8mqZdjH0eDtQCZgKbgNlALU3T9ukSaQ5x8mQwlSq7J0l3r+TOqeD09fv88Ye52Ns4UtDOiZb+rdgdmLMaxYNPBpM7d+5ElXyASpUMf5f0/h2yUnR0NE+ePOH82fN88O5HOBdxTrVSsm/3PhwcHTh5/CT1POvjmLcQlUpXZuK4L4mOjjZh5CkL3BUIQAX38inuE3wymJIlSyRpUXOvVNHw9zh3QdcY4/iXrU/+3DasObk5UXpx+6L8r05nRm+bQVRM9vi7ptfe3ftY9vNyvp72lVH54lp9x4wfhYOjg07RpW3fniDKVSjH6hVrqeFem8I2RalZqQ4/zpqXaj4fv0aUKVuascM+51Twae7du8/O7buY/d0cevbtka5uWHro1/MdCuYtTGmXcvTp3o/wsEtmicMcnj59yoB33+fDQR8kue7mBNnxniFyvkwt1Rtb2U3+O96XVMStCBwckt60HBwdiIiITDN/5y6daNm6BS5FXQgLDWfqN1Np6d+a9X/+SiOfRnqEnOUibkVQwL5AkmUK427mt25FmDwmvwZNOXLoCACly5bmt43rUu37+N9/V3j44CF9uvfl46EfU92rOn9t/Yuvxk/i9u3bTPh6vKlCT9a/l//l8zHjadzEF68aXinuFxERgb2DfZJ0x9jXIiLCNK/FG5WacuN+BDsv/p0ofUyTAWw6t5ug8KMmiSOrPHnyhI/6D2TAR/0pV6GcUXlHDBlF2XJl6Nq9i07Rpc+V/65y5b8rjB46luFjhlCydEnWrf6NTz8cQnRUFG+/l3yXizx58vD7tl/p2ak3DTyfXZMC3urKV1MnJJtHT3YF7Hjvw3ep37A+dna2HDvyD5O/mkpznxbsCNpu1j7OpjJ50hQeP37C4E8HmTuUDMmO94zMk3lozS1TFVxTU0qVB7oBzYAyQB7gPPALMFXTtPvP7V8B+BLwAXIBh4BRmqZtM2Xcxpi78Mf43xt4wyuvtaZm9dqMGTmOrTs2p5JTpGb2/FncvXuXkAshfDtlBm+0asOf2zdQomTyg8xiYmJ49OgRI8YOj++m0NDHm1u3bjFn5o98NuJTChR4fhIR07h37x4d23bGysqKWXO+N0sMxihs40j94p4sOryWaC0mPv01dz+qOpen+YLeZowuY6Z98y0PHz5i0GcDjcq3J3Avy35azo6g7WZfp16LieHe3XvMmPMtr77RGoBGjRsSHhrO1Enf0q9/32RjfPToEX269eP69RvMnPcdbsVcOXTgMJPGf4OllRXfTDeuRTuzqlWvlmiwZINGDajfsB5NGjTjh+/mMHzMUJPGY2rnz53nqwmTWLpyCXny5DF3OEJkGylWcJVSKQ8vT4OmaUlHmGSNXkB/DP19fwaeAo2Bz4EOSqm6mqY9BFBKlQH2YJjV4SsM/YX7AhuVUi01TduiR4AODvbJtooZWnaTtqSlxdbWlhYtm7Nw/qKsCM8k7B3suR15G03TEt0gI2I/hTua4WvZCu6GKXNq1q5J0xb+VCtXjSmTpjD1uynJ7h8X4/ODyvya+jFv9nxOnTxFnXp1dI05OQ8fPqT9m50IuRjChi1/4Ormmur+9vb2hIcmnaIurkUkuW8bstpr7k2wtLBkzYlnb7l81nkY4vM2c/5ewZPop9jmNnytbaEssLawwja3TbZt/QgPu8Q3E6fw7aypPH78hMePn8Rve/z4CZGRt7G1zY+lZdJR/B/1H0hAz64UdS1KZORtAKKjooiOjiYy8jZ58+Yhd+7cJjkPh4IOcA58myQeoNi4qQ9bN23jyn9XcSlaJEm+nxYsIXDnHg6cCKJUmZIA1G9YDzs7Wz7qP5i3+vagSrXKpjiFFHl4elC2XBkOHzhs1jhMYfBHn+DTuBG169QiMtLwLeGTp0/RNI3IyEhy585N3rx5zRxl6rLjPSOzlMLsH2Jfdqm14IZAhkZxaGmUmxkrgQmapt1OkDZLKXUWGAb0BmbEpk8A7IEamqYdAVBKLQJOAN8ppSpqOoyGcK/kTnAyc6yeCj5FRfeKGS43J71R3Cu58/jxYy6cv5CoT1VwbD+qzPwdsoK9fQFKlSnNhfMXU9zHvVLSftQJmWNwytOnT+nWqTuHDx7m1w1rqVI17UqEeyV3flu3PsnI9lPBp8mVKxdlyqY9V2hmvVnJn+Br5zl141l/X4e8BSiYz55BDXsxqGHiKfGK2hWmVQUffg/cZFg7MZsJuRjCo0eP6NfznSTbpk+ZwfQpM9i5/y+qeVRNsv30qTOcPnWGeXMWJNlW0rk04yd9wbvvJy1XDxXdK3Ag6GCK21P6Hw8+Hoy9g3185TaOVy1DV5kzp86YvYIbJyddNzPqVPApwkLDcHUqlmSbq1Mx3h3wLpMmf2mGyNIvu98zRM6UWkV0ERmr4OpG07QDKWxajqGCWwVAKWUDvAb8FVe5jc1/Tyn1IzAWwwC5/VkdY+tXWjPk06FcvHCRUqUNE06EhoSyd88+xo0fa3R5d+7cYcMff1KzVo2sDlU3/s39sba2ZvnSFQwdMSQ+fdmS5VSqXImSpUqaLTaAa1evcfb0Wdp3bpfiPq1fa8Xno79g6+ZtVE5QkdyyaQt58uTBPZmBhHqKiYmhV/c+7Ni+k1/WLqd2nVrpyteqdQu+GDueNSvXxvf5jIqKYtUvq2nS1E/31sIqzuUoV6gE4/+alSj9+v1bdFuRdNHDKa2HcOZGCDODluJrnbSCmB1U9ajKb5uSTkD/arPX6dClPQE9u1G6TKlkcpJsviGDhxETHc2XUyammE8PrV9vxU8LlrB983Zea/NqfPrWTdsp6loU5yKFk81X2LkwkRGRXDh/MVG8B/cbKssuRV2SzWdKhw8e5uyZc4nO60W18Kf5PHpuNpRvvprMkUOHWbxsMa6uRc0UWfpl93uGyJlSrOBqmtbThHFkllvsz6uxP6sBuYGkM8pD3CwPulRw3+rTk1kzf6BDm46MHDsSpRTjRo3DrZgbvRMs3hAWGkblClUJCOhGXW/DV91TJ0/jzOmz+Pg2wqVoEcJCw5k2eRpXr1xl/qK5WR1quqxZtRaAw7EDtDb9uZlCToUo5FSIho28AbDLY49/M39Wef8CQOHCTgz48D2+/vIb8ufPT3Wv6qxasYod23fwyxrTzjXZtV03PDw9qFy1MrZ2tpw7e47vp83EysqSAR8axkcG7tzNa81fZ/DHg6jpbfggUalKJbp278L4MROIiYnBw9ODv7b+xaJ5i/lk6Mfkz58/tcNmuY/eH8SaVWv5+LPB2NjYsD/o2WAtV9eiuLq5EhYaRtWK1ekW0I063ob5QD08PWjbvg2fDv6Mp1FPKVGyBD/+MJfQkFDmLdJ/Qv43KjXlaXQUvwYn7vb+JPop+y8lXZ3vcdQTbjyIYP+lY9R1fTY7xMkP/2DNyc0M2/SsS0kV53K42jljoQwtjWUcS9C8nOF/csfFv3kUpc8UaPb2BWjo453stuLFi8VvCwsNx9O9Bt0CulGzgWEBguTyFbAvQHRUVIpl6sW/RVO8fRow8L2PuXnzFiVLlmDd6l/ZvuUvps+eBkB4aDg1KtWhW0BXPOtPAqBzQEe+/3YWnV7vwkeffohbMVeOHDrK1xMm4+HlQZ36pp3Du2+PtylRsgQentUoUMCOY0f+Ycqkabi4uvB2f8NAubDQcLzca9ItoBs1GlSPz7t7525u3LjJ1SvXADh88Ag2+Q3dZV5v85pJzwMydr1Nbs70nxb+RK7cuWnk09BEkT+T0+8ZWSm7drN6WeSoQWbJUUpZAiMw9LVdEpsc95H1cjJZ4tKS7byolOoH9ANwdnZmX2CQ0TF98cXnfP/dTN4K6IWmaXh5eTJk6BCOH3m22tGVK1eIjo7m0ePH8ceIiYrh76C/Wb1yNffv38fGJh+Vq1Rh6rSpWGKVoVggc83w3ToFJHr+4YCPAPDw8GDy1MlA7BRcj5+wL/DZ54UWzVsScTOSKd9MJeJWBG7F3BgxagQOBQom2i/dMtibpIizCyuW/BwrbxYAACAASURBVMK//04jKioKJycnPKp7MHr0KCKv3ebAtYOcOnbK8Fo8fMyBwGdf2QZ0DYBomDHlOyIiInAu4sw7/3sHfz//RPsZJ2MXvPXrfgdg0sSvmTTx60TbunfvTo+ePeL/px4/ekxQ4LMvO3r36gOaYsSQUdy7d48yZcowfsJ4Ht+PSrSfMd53Tbn1O46FUrR3b86lK5fp5uifrnLtrGyomLc477u2wzmXQ/xxrCwsqWJTOtFxm9T2wb3Us0pwqwqNaFXBMKp/4W9LufvgXprHO7A7a/to/ht+Jb7MhK9Hase5e/su0dHRmYwlY++Pjwd/zNwf5/L5yPHcu3uPYsWLMWTYENzLVuLwnmMJzuEJh/c8+0AybdpUFi5cxJih47h9+zZOhZ1o2aIFXbp14ei+4xk8h4y9N+zyF2D1ijXMnD6Lx48e4+joSN26denRswchp8MJOR2e6LU4uDv+Sz2GDh7BsaPPzuvHWXP5cZahMWHLtowN08hMt4iMXm+fd/3ajTT30Uu2uWeIl57KSZNyJ0cpNR3DVGVDNU2bEJsWgKGLRW9N0+Y9t39pDDMvTNM07cPUyvaq4aXtDkp9TfbM2hcYFN+CqxdTTIi/L3A/db31a7mJ0vRf6elA4MH4Fly9WBq1tkrGBAUeiG/B1Yvn9Pa6lg+GSvS3l41b/MFYf7+3VNfywVCJjmvB1Uu0zu+Pw3uO4Vlfn2Wd41iotJfWzayDu48kasHVQy4L/QcJ6n29NQVTnIN3nUYcOnjILM2oThWctTdndjbHoROZ02TaQU3T9L0hZFM5ugVXKTUOQ+V2dlzlNtaD2J/JXWnyPLePEEIIIUSWkg4K5pV91ik0klJqNIZlgecDzw87/jf2Z3LdEOLSkuu+IIQQQgghcrgc2YIbW7kdBSwE+iQz3dc/wGOgXjLZ68b+zFjnQyGEEEKIVChkkJm55bgWXKXUSAyV28VAL01LsDRSLE3T7gG/Ab5KKY8EefMDfYCz6DCDghBCCCGEML8c1YKrlOoPjAHCgC1Al+dGrF7VNC1uPdshQBNgk1JqCnAHw0pmrkBrPRZ5EEIIIYQQ5pejKrgY5q4FKI6he8LzdgCbATRNO6eUagBMBD4DcgGHgBZ6LdMrhBBCCAHSRcHcUqzgKqUaZbRQTdN2ZjRvGuX2BHoasX8w8LoesQghhBBCiOwptRbcv8j4GgH6T2gohBBCCJEtqUwt+iEyL7UK7liSVnDrAC0wLJQQCFwBigDeQBlgAzJ4SwghhBBCmFGKFVxN00YnfK6Uqoth4NYHwHcJZy9QSlkAAzD0dx2rS6RCCCGEEEKkgzHThI0DtmiaNv35qbk0TYvRNG0asA2p4AohhBDiJaYwVLDM/XiZGXP+tYEjaexzhGcLKQghhBBCCGFyxlRwFYZ+tqkpm4lYhBBCCCGEyDRj5sHdA7RVSr2iadr65zcqpV4D2hA7D60QQgghxEtJIbMomJkxFdxhwE5gnVJqR+zvVwFnwAdoBDyM3U8IIYQQQgizSHcFV9O0g0opf2Ae4Bv70DB0XQA4DfTWNO1wFscohBBCCJGjyEpm5mXUUr2apu0BKiql6gNeQAHgNnAodpsQQgghhBBmZVQFN05sZVYqtEIIIYQQQhdKKQcMs3MpYJ+mabfSmzdDFVyllA1QHsivadqujJQhhBBCCPEiUkgXhcxSSvkAa4AYIDcQpZRqp2na1vTkN2oeYKWUm1JqFRABHAC2J9jmrZQ6qZTyNaZMIYQQQgghnjMFGKhpWiHAAVgKTE1v5nRXcJVSLkAQ8DqwHtjLswFmxG4rDHRMb5lCCCGEEC8ipZTZHzmBUmq6Uso2mU0lgWUAmqZFAauBEukt15gW3FEYKrD+mqYlme9W07SnwC6ggRFlCiGEEEKIl1dp4LRSqstz6UHAFKVUJaVUbWBobFq6GNMHtxXwq6Zp21PZJwxoaESZ2Zoy0UTNeh9DYYJzACyUfitfW2Wsu7hxlMJK6Xucu09v61o+QAzR3Iu6q+sx9vX/SdfyAQ7vOcq+N/Q9zrA9n+taPkCd+14M2/O7rscYV2+oruUDKJ1Xto/RonUt30DT/TgRj2/oWj5AdEzU/9m777gojjaA479BFBVRQRAVVIqCYMfeezd2jcYejYmJxpJmyWuNPbEksXeNNcbYoonGGHsXW+yKgAUrFhQLMO8fBygCB6d3YHm+fu6DN7uz+ww33M3Nzs5Y/Dx2abNY9PhoTUTUU8ueAm3R4wvz0Fo3UEo1xdCY7QJ011qfAT4BFgPHo3fdB7zYCE6UKe9YzsDZJPZ5CtiacEwhhBBCiLeMwuo1eLwptNa/Az7AfuCAUmoEcE1rXQHIDGTRWpfVWl9I7jFNaeDeBnInsY8XEGLCMYUQQgghxDtOax2ute4HlIl+nFBKNdRah2mtTb4saUoDdyfQSCmVI6GNSqn8QF2em1lBCCGEEEIIY5RSVkopb6VUUSBAa10T+BaYrpRarZRKqoM1HlMauOOA9MBWpVQ9IGN0ULbRz9dimKvsB1ODEEIIIYR4m6T2DApv0CwKRYBTwEnAH7iklGqqtV4MFAACgGNKqW+USv6NMslu4Gqt9wIfY5i2YR3wZfSme9HP3YEuWuv/kntMIYQQQgjxTpuBoWGbA8gC/AwsUErZaK3va617A1WA94AjyT2oSbeMa63nKKW2A59iWDotG3AX2AP8rLU+bcrxhBBCCCHeNkrJSmYm8AUGaq2vAyilJgKDMMx5ewZAa30EqKiU6pzcg5o8J5LW+izQx9R8QgghhBBCvGA/0E8pdQd4BPQAbgHxZkzQWs9N7kFNWclskFKqchL7VFJKDUruMYUQQgghxDutC2CDoaF7DKgOtIheveylmXKT2RCgahL7VMaw4pkQQgghxDtLvQb/3gRa64ta68pAJsBBa+2ttd76qsc197JNaTHMpCCEEEIIIUSyaK0fAg/NdTxzr73oB1h+nUIhhBBCCCESYbQHVyn1zwtJnZRSVRPYNQ2GVc7yAkvME5oQQgghxJvpTZmH9m2V1BCFqs/9X2OYA9ctgf2iMNzxtgyZYUEIIYQQQqQio0MUtNZWMQ9AAUOeT3vuYa21dtZaf6C1vpEyob++goMv0aZVW5wdcpLdPgfvt2hDUFBwsvIOGjiYhnXfwyV7bjJY27Jw/kILR5u4ly1HYGAQLZu2wsujAPaZsuHqnIda1erw5/o/UyDquC5fuswXvb6kWsXqOGbOjm1aOwIvBiYr748TfqJFk5Z45M6HbVo7RgwbaeFoE7dj604a1WiKm4MnBVwK0qPL59y4lrw/tcCLQXT5oBteOX1xd8xPs7otOXww2XNlm92mP/+mYY1G5MnmRl4nD6qXr8W2LduTzHf61Bk6f9CF/C4FcMmah9KFyzH95xkWjTV/Vk9+rjo+3mNsxRGx+3hlzU8Hn7YMLjOA8ZXGMLjMAN7P35xMaTMlefzsGZxonq8J/Ut+yQ+VRjGi3BA+LvQhLra5LFkswFCnHNI7xXu4OXsmmTcwIJCObTrj5uyJq0NeGtVugv/BwxaPOTGGOtWYPNncyevkSfXytZOsU4EBgXRq0wV35/zkdnCjce2mKVaGK5ev8u0Xg2lUvRn5nHxwzeROcOClOPscOXSUr3v0p0rxGuRz8qF0gQr0+LA3QReT91kCcCf0LoO/HkbpAhXwcPCmpFc5+nz8ZdIZzahZwxZksXFg+OARRvcLCgymTfO2FMpfBOcsuXDPlY/6NRuyccOmFIrUvBQKK5X6j3eZKTeZdQZS7x3sDfHw4UPq1aqPjU06Zs6dgVKKoYOGUrdmPfb778XW1tZo/qmTp1GkaBHqNajLooWLUyjq+F6lHA/CwsjmmI0hwwbh4uLCvfv3mTtrLk0bNWfJr4tp0rRxipXj/PkL/LZiJcX9ilO+Ynk2b9qc7Lzz5szDzi4z7zVqyKwZsy0YpXF7du6l9XttqVqzCrMWzyD0diijh46jRYPWbNy5Hhsbm0Tz3r4VSuMaTclkl4mxP40mY4YMTPtpBs3rtWLDtnV4FcifgiWBeTPn802f/nTt3oUv+vdF6yiOHTlOeHi40Xz+Bw/TpG4zKlauwMSp48mcOTMXzl3gwYMHKRL3r2dXEnj/WcMiSkfG/r9irnLYpLHhr8BN3Ay/jVNGRxq41cXHoQAjD4wzetwCDt54Zc3H3pD9BIddJoN1BmrlrsYXfr2Y4P8TwWGXjOY3h9HjR1K8RPHY59bWxj8Wbt+6Tf0aDcmUKRPjf/6ejBkzMmXSVBrXacKmHRvxLuBl6ZDjmDdzQXSd+jDZdcpQhkbRZRhHhowZmDppGo3rNGPTjj8tXoaL5wNZt/IPChcvTJnypdi6OX5jfM2KtZw5dZYPu3fCyyc/IVevMWn0T9Sv3IifJ/+EL8ZjvBN6l2a1WqKU4utBfXHN48q1q9fZv+eApYoVz4plv3H8WPIWN33wIAwHRwe+HTKQXC65uH/vPvPnLKBlk/dZuGw+jZq8Z+Foxdsm2Q1crfV8Swbytpgzay4BFwI4euIwnvkMPSGFCxeiUIEizJoxm159Pjea/9rtq1hZWXH+3PlUbeC+Sjl8C/oybebUOGn16telQD5fFs5bmKIN3IqVKnDxsmGu6Hmz55nUwD1wZD9WVlZERESkagP3h5ETcM3jwrzls2MbH/m981O3UgMWz1tK5487Jpp3/swF3Lh+k1WbfsPNww2AClUrUKZgecZ99wMzf5mWEkUAIOhiEAO/+h9DRw3mk54fx6ZXr1XdaL6oqCg+7dKDytUqsXD5s7ehSlUrWizWF4U8vMbFewn3/C8/+xthT581tM/dPc/1hzfoU7wHfk7F0HcTn8rx4HV/tl3eESftTOhZhpX9lqqulVh4yvK3NHh5e1GqTMlk7z9nxlyuX7vBuk1rcPd0BwyvhZ9PKUYPH8PcRSn3t/KsTg16oU5VM5pvzox53Lh2g3WbVuPu6QYYylDCpzRjho9jzqKZlgybshVLczjA0NBcPG9pgg3cT/t8QjanbHHSSpUtQbmCldnwxwaqvme8/o8ePJYHDx7w994/sctsF5veuGXKNBRDQ+/Q/6uBjBo3gi4dPkpyfx9fHyZP/ylOWp36tSniXYxF8xdLA1eYzJSFHj5RSp1XSiV47Uwp5RK9vYv5wnvz/LH2D0qXKR3bKARwc3ejXPlyrFvzR5L5razMPbHFy3nVcrzI2tqaLFkykyaJ3iFze5Xf5+vyWhzcd4jK1SvH6VkrVqIoDtns2bDW+LCPg/sO4ZHPPbZxC2Brm5Ey5Uvz94a/iYh4pXm0TbJo/mKsrKzo9FHiDfKE7Ni6kzOnzvDp590tFNmreb5xGyMourc3q00Wo3kfJJD3UeQjroffSDJvajmw7yAe+TxiG7cAtra2lK1Qho3rN6VwnVqClZUyuU4djC2DW2yaoQxlU6QMyXlvebFxC+Cax5Vsjg7cvHnLaN6HDx6yYslK2nR8P07jNiUNHjAEH18fWrzf/KWPYW1tTebMmZO8qvC6Ulil+uNdZkrpPwCuaq2vJLRRa30ZuAS0M0dgb6qTJ05SsJBvvHTfgj6cOnkqFSJ6OeYoR1RUFBEREYSEhDBy+CjOnjlH908/TjqjiCNNmjSkS5c2Xnq6dOk49d/pJPOmTSCvjU06wsMfcfFC8sYjm8OeXfvI752PlctXUcKnFNltc1LStzSzphnv8du7ay8Ajx49onblejhnyoV3bl/69R2Q5NAGc+no044fq3zPmArD6eTTDnubrEb3z5fV8MUw5OE1k8+V0TojOW1zEPLw+kvFaqqPO3fHMaMznrm8+KjDx1wKMj4swlAf08VLt0lnQ3h4OAEXLloo0vj27NpLfu/80XWqNNltc1HStwyzps0xms8qkb8pm3TpUrwMpjh76hw3b9wiT97cRvc76n+MR+GPcMzuSLe23fF0LICXc0G6tO5m0hjel7V75x6WLlrGD5PGmpw35nPjWsg1xowYy7mz5/moe1cLRCnedqY0cL2BpO5MOQoUePlw3ny3b4eSNWv8Dz97e3tCQ0NTIaKXY45yDPhmIHbps+Du6smEHyayYPF8qtUwfulQxOeZ34OD+w7FSQsOusS1kOvcCb1jPK+XBwHnArh969lrFhUVhf8Bw3D6pPKbU8jVEM6fu8CQAUPp9eXnrFi3nCo1qvBN7/5GbxYLuRoCQNf23ahWowq//fErPft+xi9zF9Gt4ycWjTk84hGbg7ew+PQyfjw8lQ2Bm/C2z88Xfr0SvYnMJo0NzfM15uqDEI7ePG7yOVvmb4pCseXStlcN36jMWez4rPenTJo6gdV/ruTL/n3ZumUbdarW48b1xG9gzOeVjwvnLnD71u3YtKioKA4dMNTR0Nsp9z4XcvVadJ0aRq8ve7Ji3bJk1an8Xvm4cC4ggTL4A3AnBcuQXBEREfTrNZBsjtmoW6+u0X2vXTV8Ofpu4CjSpEnD3OUzGfPTSI4f+Y+W9VoTdj/MYnE+efKE3p/1oWefHuT3Nn2M///6DyabbXa88vrw4/ifmfPLLKpWr2KBSC0vtW8we9dvMjOlgZsFSOrT8B5g//LhGKeU8lZKLVJKnVRK3VVKPVRKnVJKjVdK5Uxk/1VKqVCl1AOl1HallPEBf8JsevTqwY492/lt1Qpq161Np3adWb9uQ2qH9cb56LMu+B84zOghY7lx/SZnT5+jZ5deWFlZYWVl/A2sY9f2REVpenbtxcULF7l29RoDv/hfbC9OSr4B6qgowu6H8cPP4+jQpT2Vq1Xih5/GUaN2dSaOm4TWOsF8UVGG9JZtWtB/cD8qVqlAjz6f8dXAL1i/ZgOnT52xWMyXwi7z+/m1HL91gnN3z/PvpW1MPjoDu3SZqOpaKd7+VsqKzr7tyJouC3NPLCRKm7awY+08NSjlXILlZ1dyM9yya+YUKVaE4aOHUrdBHSpUrkD3np/w65qlXL92gxmTEx+D2rlrR6Kiouje5TMCzgcQcjWEfn37E3gxCEjZoT0J16mx0XXqx0TrVKeuHaLHdvck4PxFQq5eo1/fgbFlUK/J8KTnfdt3MAf3HuLH2eOxszM+7CCm3uVxy82U+T9RuXolmrZqzNQFP3M5+Aorl66yWJyTfviR8PBHfNmv70vl/7TnJ2zZtZllK5dQs04Nunboxp9//GXmKMW7wJS/4qtAkST2KQJYcpowVyAn8DvQH+gNbAK6AQeVUtljdlRKeQK7gHLAWOArDOsc/6WUqmmpAO3ts3LnTvzvAaGhodjbW6ztb3bmKIerqwslSvpRv2E9Fi1dSOkypen/zQBzh/rWa966GX2+6cW0H6dT2K0Ylf2qkSNXDmrUqU72HM5G8+Z1z8vkuT9y1P8YZQtVpKhnCQ7uPUS3noZLftlzGs9vTvbZDPWmao2qcdKr1azK9Ws3CLma8OV8e4eYfHF7carVNFwNOHb4mJkjNe5S2GWuP7xBXru4l4kVivYF2uBt78WM43O58uCqScetmKscjTwasPbCevaE7DNnyMlWtHhRPPN7cuigf6L7uHm4MX3eVI74H6FEwdL4uhdm/94DdP/c0JueI4k6aU7P6tSLdaOK0Trl5uHGtHlTOOJ/hJIFy1DQvQgH9h6g++eGIVQpWYbkGDloDIvmLuH7qWOoUqNykvvbOxiuvlWsWj7OYgN+pYpjl9mO40eSN7OBqYKDLvH96PF8O2QAjx8/4c6du9y5cxeAx48fc+fOXSIjI40ew8XVBb8SxanboA7zF8+lVJmSfNtvkEXiFW83Uxq4W4C6SqkEb91USlUC6gHJv03dRFrrzVrr6lrrAVrrKVrrGVrrnhimMMsJdHpu91FAVqCO1nqU1noKUAm4AkxWFlpixMfXhxP/nYyXfvLEKQr4vDmjNyxRDr8SxTl/7vyrhvZO+mbwV/wXfIwt+zZx9MIhps2fTMD5AEqXL5Vk3oZNGnD4/AG2HdrCnuM72LhrAw/CHuLimgvX3C4pEL1BUvUmsd7oAr7eSeRLnd62F/sGW3u1wC97MeaeWMiZO2dNOlYp5xK0yt+czcFb+Cvob/MF+ZKSents1PQ9/rtwjN2Hd3Lwv31s2b2ZB2EPcHF1wTWPawpFCQV8Xr5uNGrakOMXjrDr8HYO/LeXf3ZvIiwVypCUH8f+zJTx0xg2bjAt2jRLVh4vH+NTiFnqb+ZiwEUePXrER50+Jq+ze+wD4KcJP5PX2Z3/jp8w6ZjF/Ipx4fwFS4RrcUqpVH+8y0yp5WOAJ8Df0UMCaiulCkb/nIChJ/Vx9H4pLeZOGXsApZQt0Aj4V2sdO3ev1joMmAV4AUm3DF5Cg/casG/vPgIuBDwL7mIgu3ftpsF79S1xSoswdzmioqLYtXM3Hp4e5gzznWJrmxGfQj44OTvxz8YtnD19jo5d2ycrb5o0afAqkB83DzdCroSw5rc1dPyog4UjjqtBY0O92bJpS5z0zRv/IZdLLpwT6TWrWacGNjY2/JNAPjDMKJGS8ti54pwxO4H3gmLTmno2olzOMvxyaqnJ426LOBamXYHW7L66l9/PrzV3uCbxP3iYc2fO4VfSL8l906RJg3cBL9w93bl6JYTfV6ziw26dLB/kcxKvU1ui61T2hLLFelYGN65eCWHVitV07mbajAyWNHvKXMYO+4FvBn9J50+SH1cul5wU9SvCtn92xBmmcXDvIe7fu0/REkldjH05hYsWZt3GNfEeAO9/0Ip1G9fg8dzsG0mJiopiz649uHskP48QMUyZB/e0UqoVsBjD0IBez21WGMbffqC1jt/tZ2ZKqfQYhhukB3x51qheH/2zCGAD7E4g+57on6UAs18H/LBrZ6ZNmU7LZq0YPGwwSimGDR6Ga25XunZ7NoNaYGAQBb0K0b59O8pVKhubvn3rdm7cvMm1EMOltYMH/bHNZLiZpVnzpuYO1yLl+G7oCG6HhlKufFlyODsTcu0a8+fM58D+A8z7ZW6KlSHG778Zxpv5HzJ819n45yYcnRxxdHKkUmXDBYnM6bNSq3Ytfqv4a2y+QwcOERgYRFSUYTzbqZOnYo9Vp15tMmbMmCLxHzt8nH82bqFwsUIA7Nu1nykTp/FZ3+6UKvts/tLgoEuULViBtu3bMrasYdW1p0+fMnzgCMpVLItdZjtOnTzNT+Mm4+3jzSe9uqVI/DFq1a1JxSoV6dvjS27duoWbW15Wr1zLlr//5acZPxrKEBhMCd/SfDXgC2pWqwWAQzYHen/1Od+PMow9rFS1IocPHeH7kT/Qut37Fv3S1NGnLbce3Sb4/iXCI8JxzeRC7bw1uPP4Lv9eNsxdWjN3dWrkrsquq3u5EX4Tt8x5Y/OHPQmD5+7nmVRlHHtDDrD49DIAPLN40Nm3HZfDrrAnZH+cvBFREVwKu2yxsnXr+Al53fJQpHgRsmTJwtEjx5g4bhI5c+Xk488M85YGBwbj51uKdu3b4le+GGCoU4MHDKVCpfLY2dlx6uQpJo6bRAEfbz7r/anF4k2IoU5VoG+Pr7h163Z0nVoTXacmxZahhG+ZOHXq6dOnDBkwjPKVykWX4TQTx/0YXYaUmY5u3e+Gj6tj/oYvRFs2/ouDowPZHB0oV6ksq39dy5BvhlO1VhXKVynPwX3Pho1cv3gT3zLPemkrFKmKa24Xlv2xKDat/9CvadukI93afkqbTu9z++Ytxgz9gXxenjRpZZm5yLNmzUKlKgnPz5s7T+7YbUGBwdSqWZt+337NNwO/BmDU8NGE3r5DmfKlcXZ25tq1ayyc+wsH9x9i1gLLrlgo3k4mTS6ntf5DKeWBYShAGQxDAO5gaDTO11obn5zPfLoCz88IfRFop7WOmS07Zq7ehD4dYtISvDarlOqGYUwvzs7O7N6+J6HdjBox4jumTJ5Kp3ad0Vrj51ec/v37x76RAYSEhBAZGcmjx4/jnOOrvt9w5MizySqmT5nO9CnTAfhni8VGfyToZcuR3iYDu3euZ8kvS3jw4AH2DvZ4enoyadJEcufK/VK/04RvFUmedq3j9nL27tkHgKJFizJ+4ngAIiMjefL4CXt2PPvOM2b0GDb+tTH2+coVv7Nyxe8ALFqyiBw5cpgcy/MrYCXXxYAg1qxYx4/jfubp06fkyZOHnr16UrdeHY7teXa5L+a1ePLoSWx6ZGQkRw4cY/miFTwIe4CjoyM1alSnTds2nD50zuRYYihe7tLXV19+yexZs/lu0EjC7oeRO09u+g/sj08+H/x3HYktw+XAqzwMC8d/l+FvoWa1Wty9dY/li37l5wmTcXBwoEXL5rRr3y52n5dR5oHxnkrr21aUcSlJ9VxVsE5jTfjjcK5cucKRM0cp9NhwebxsVsPFoPI5y1A+Z5k4+c8Hn+fItSOUCTOcJ41Kg3OEY+zzIrmKkNYqLXnscvOFX9zFU8IehvH7P8m7GejQLtMXmcxsl4WVv65i6s/TefzoMQ4ODpQtW5aOnToQeCaYwDPBsa/H40dPYs8RGRmJ//7DLP1lWWydql69Bh+0bcPxA6ZdfjaHr778KpE65Yv/rqMJ1Kmj0WU4wtJflj/7u4guw38HXm06x+TeWPhJ+8/iPB/Q538AFClahO/Hj2PVkrVorfl301b+3bQ1zr6FChUir9uzL0OPHjzi/p0wTux9dsOlQ0Ynhg4fyoL5C+jyfjfSp09PmbKl+ajbRwQcDSIpaVSaZJUjua4GX+XATsNMGyEhIURFRXHp4uXYNFubTPy5ayPLFi83fG7YGz43Jk6agJuLe+x+bwoV/U+kHpXYXaavM6WUK4bpyDIBxTEMR5intZ4Uvb09sADoorWe80JeD+A8MElr3dvYeUqU9NM79+4wtssr2719T5we3DeVpcth6t3oL2PPjn2UrVjaoue4//SuRY8PcGzPCQqXjT+HsTmltYo/D6q5+e86QvHylh1+MGj3KIseH6BMmB97M1n2w3l4OcvevHlo1+HYHlxL0Vj+b9x/h8xu0gAAIABJREFU11GKl7fM5fkY4REPLXp8gBN7z8TpwbUEu7SWXWjkwM5DlKyQ9FCYV1GlXHX8D/qnSivTxcdFf7wg9ed9H1x68EGtdfKXKnyLvJHLg2itL2FYVAJglVLqN2C/Uiqj1noUEPMOY5NA9vTRPy3/LiSEEEKId49K2WkYRXyJNnCVUjFzkezTWj967nmStNaWnaE8/vmOKqX8gU8xzJ4Qs9paQsMQYtIsN7hNCCGEEEKkGmM9uP9iGProA5x57nlymHfwTvJkAByi/38Mw4wO5RLYL+Y6+oGUCEoIIYQQQqQsYw3cYRgatDdfeJ5qlFI5tNYhCaRXAwphaISjtQ5TSq0Fmimlimqtj0TvlwnDDWpnscAMCkIIIYQQkPR80sKyEm3gaq2HGHueSqZGL8n7D4a5b9MDJYDWwH3gi+f27Q/UADZGz9N7D/gIwxCFBvpNvLtOCCGEEEIk6U27yWwJ0AFoDzhh6FEOBKYD47TWsXOfaK3PKaUqAKOBfkA64BBQV2ud+ksFCSGEEEIIi3ijGrha6+XAchP2PwlYZkZrIYQQQogEKMDKpMVihbkZm0Xhn5c8ptZa13jJvEIIIYQQQrwSYz24VRNJ15Dg8hwx6TK2VQghhBDvMCU3maWyRPvPtdZWzz8w3NC1BggAOgPuGKbmcgc+BC4Aq3m2kIIQQgghhBApzpQBIv8DSgIltdbztdaBWuvH0T/nAWWA0tH7CSGEEEIIkSpMaeC2BX7TWt9JaKPW+jawAmhnjsCEEEIIId5USqlUf7zLTGng5gKeJLHPUyDny4cjhBBCCCHEqzGlgXsJaKyUSpfQRqWUDYYpuS6bIzAhhBBCiDeVFSrVH+8yUxq484F8wD9KqcpKqTQASqk0SqkqwGbAA5hn9iiFEEIIIYRIJlMWehiNYVncRsAWIEopdRtwwNBQVhhmWRht7iCFEEIIIYRIrmQ3cLXWT4EmSqkPMEwTVhxD4/YuhiVw52qtl1gkSiGEEEKIN4SCd/4mr9Rm8lK9WuvFwGILxCKEEEIIIcQrk4WShRBCCCHEW8XkHlylVBHgA8AHsNVa14xOd8Ow0MMmrXWoGWMUQgghhHhzKLCSIQqpyqQGrlJqGDCAZz2/+rnNVsASoDfwk1miE0IIIYQQwkTJHqKglGoNfAtsAooBo57frrW+ABzAMMuCEEIIIcQ7Sr0W/95lpozB/Rw4BzTWWh8l4VXNTgL5zRGYEEIIIYQQL8OUIQqFgXlaa2PL9V4BnF8tJCHiS4lvoioFzpMlnb1Fjw+QRqWx+HnCIx5a9PgxLP16jKwwyKLHBziw46DFz/PRpn4WPX6diPLM2LzCouf43K+VRY8P8CgynJOhxy16jiLZ/Cx6fDD8jWdOl9Wi51AWvwddYWVYL8qCZxDvMlMauAqISmIfZ+DRy4cjhBBCCPFmU4CVkomqUpMpv/2zQPnENiqlrICKwH+vGpQQQgghhBAvy5QG7nLATyn1RSLbBwD5kEUghBBCCPGOU0ql+uNdZsoQhYlAS2CsUqoV0VOEKaW+ByoBJYE9wAxzBymEEEIIIURyJbuBq7UOV0pVAyYBbYGY0eF9MYzN/QXoobWOMHuUQgghhBBCJJNJCz1ore8CnZRSfYFSQDbgLrBPa33DAvEJIYQQQrxx3vV5aFNbshu4SqkLwAat9Wda69vAX5YLSwghhBBCiJdjyk1mThh6a4UQQgghhHhtmTJE4T/A01KBCCGEEEK8HRRW7/gsBqnNlB7cH4H3lFJFLBWMEEIIIYQQr8qUHtxLwN/ATqXUdGA/EEL0dGHP01pvM094QgghhBBvlpRY+l0YZ0oD918MjVmFYWqweA3b51h2gWkhhBBCCGF2Sqk/gTrACK31t8+l2wPjgCZABmA30EdrfSxVAk2CKQ3cYRhv1AohhBBCiDeUUqoNUDSBdAWsBdyAnkAo0B/YopQqprW+lJJxJkeyx+BqrYdorYcm52HJgN8EwcGXaNOqLc4OOclun4P3W7QhKCg4WXkHDRxMw7rv4ZI9NxmsbVk4f6GFo03cy5YjMDCIlk1b4eVRAPtM2XB1zkOtanX4c/2fKRB1fJeCL/HB+23JkS0Xzg45ad2yDcHJfD0yps2U4OPI4aMWjjquV6lTzxs35nsyWNtSvXJNC0Rp3I6tO3FI7xTv4eac9L2rgQGBdGzTGTdnT1wd8tKodhP8Dx5Ogajj+3vjZhrWbkT+3N44ZXLGx70gHdt05tSJU0nmPfnfSdq2bI93Xh9yZnWhTNFy/Dj+JyIiLLs+jq+DF8vqT4v3mFNrfOw+3Yt0THCfZfWn0axaY6PHz2mbnY4+rRhb8Vvm157ItOpj+KpEd/LauZitDFvWbWVglyE0L9mG6u71aFOxI9NGzOJh2MN4+x4/eIK+bfpR17sRNT0a0KFaV/5e9U+S51g67Ve+7jCQxkVbUjFnDWZ/P99s8RsTU6fy5fbGMZMzBZJZp/wPHaZpwxZ4u/niZJeDfLm9ad6oJXv37EuRuF+06rdVtG3VDt98hXDK7EzxgiUYPHAI9+/fN+k4P4wdj126LNSqWsdCkaYMK6VS/WGK6B7aCRiu0r+oEVABaK+1XqK1/jM6zQr4+hV/VRaRrB5cpVQeDAs7aGC/1tr0T9Z3xMOHD6lXqz42NumYOXcGSimGDhpK3Zr12O+/F1tbW6P5p06eRpGiRajXoC6LFi5Ooajje5VyPAgLI5tjNoYMG4SLiwv37t9n7qy5NG3UnCW/LqZJU+MflmYvR+0G2NikY8ac6SilGDZoGHVr1WffoT1Jvh4A7Tq0o+tHH8ZJy++Vz1Ihx/OqdSpGwIUAxowcS/bsThaO2LjR40dSvETx2OfW1sbfhm7fuk39Gg3JlCkT43/+nowZMzJl0lQa12nCph0b8S7gZemQ47hzO5RifkXp+nEXHJ0cuRQUzIRxk6hZqTa7Du0gT948Cea7euUqDWq9R65cORn1/UiyOWZj6z9bGdR/MDdv3GTYKMv3Dcz9bynn7wbGPo/UkbH/X3nuD/4Oinv7hFOGbPQq3pWga8bf8os4+lIwmxfbLu8h4G4QGdNmpJFHbb4r/w2Ddn9PwL2gV459ydRfcXbJzsf9u+CU04mzx88x54f5HNp1mGlrf8TKytBfs+vvPQz4cDC1mlZn8JSBWKe15uKZQJ48fpLkOdYuWk9Gu4xUqluBVQvWvnLMyRX6Qp0Kjq5TNSrVZreROnX3zl08PN1p274NOXLm4Mb1G0z+cQr1azTkr383ULJUiRQrA8CkCT+RO7crg4cNwsU1F0cOH2XU8NFs27qdzds2xb5GxgRcCGDcqO9xSuX3qXfUGOC41nqJUurFBkgj4IrWektMgtb6rlJqLdAY+DwF40yWJBu4Sqnvgd4QO1paK6UmaK2/smhkb6g5s+YScCGAoycO45nP0DNVuHAhChUowqwZs+nVx3gduHb7KlZWVpw/dz5VG7ivUg7fgr5Mmzk1Tlq9+nUpkM+XhfMWpmgDd250OY785x+nHIV9ijJ7xhw+79MzyWPkcslJ6bKlLR1qol61TsX4/LNetP7gfc6cPmvxHkNjvLy9KFWmZLL3nzNjLtev3WDdpjW4e7oDUKlqRfx8SjF6+BjmLpptqVAT1KJ1C1q0bhEnrUSpEpQsXJrVK9fQs0+PBPP9uf4vbt28xcZ//yRf9BekKtUqE3AhgKWLlqVIA/dyWAhn7wQkuO3aw5tce3gzTlphRx8AzgWfN3rcXVf281fgv3HSjt86xc/VRlDfrTqTj8576ZhjjJn/HfaOWWOfFy9fFLusdozoNQb/XUcoUbE4D8MeMrL3OJp2bESv4Z/F7luqsqGhd+f4A6PnWLh1NlZWVkRERKZoA7dl6xa0fKFOlSxVghJJ1Kmq1atQtXqVOGk169TAPWc+li5aluIN3OW/L8PJyTH2ecXKFbG3t+fjLp+wfet2qlSrYiS3QZ+efWnVpiVnz5xL1fepd41SqiLQgQSGJ0QrCBxPIP0/oINSKpPWOsxS8b0Mo1+nosdi9MXQuD0FnI7+f9/obeIFf6z9g9JlSsc2RADc3N0oV74c69b8kWT+5HzDTQmvWo4XWVtbkyVLZtIk0Vtnbn+sW59IOcqybu26FI3lZZnjtVi6ZBmH/Y8wbMSbN4LowL6DeOTziG3cAtja2lK2Qhk2rt/0WnwIOmRzAIz3Rj99YuhBtMtsFyc9S9YsREVFWS64V1DZpSzn7wZy577xNX7uP43fcAyPeMTVB9exT581gRyme75xG8OnmDcAN64aGub/rN3KnVt3aP1Jy5c6x+vy/gvP6pSp75m2trbY2NgkeWXEEp5v3MbwK+kHwJXLV5PMv3zJrxz2P8KQ7wabPbYUp0Apq1R/JCtUpdIB04HvtdanE9nNAcO42xfdjv5pb/LvyMKSKn1XIAKoqbUuqLX2xXBnXRTQxdLBvYlOnjhJwUK+8dJ9C/pw6mTSY/ReF+YoR1RUFBEREYSEhDBy+CjOnjlH908/NneoRp04cRLfgj7x0n18k1+OWdNnk9XWgWyZnahXqz47d+w0d5hGveprERoayjdffMOI0cNxcHCwRIgm+bhzdxwzOuOZy4uPOnzMpSDj9yakSZOGdOnSxUu3SWdDeHg4ARcuWihS4yIjI3ny5Annz56n16d9cM7hTIv3mye6f5PmTcjmmI0ve33NxYBA7t27x9pV61i2aDk9en+WaD5z6lHsQ5bUm8Ksmt/Ts9iHZEuf+GeSt70nOW2zs+3S7pc6l23ajOTOlIvLYSEvG26SDu82jIV3y2+4hH9033Ey29tx4VQAHap1pYprLZqVaM2cHxYQGRlp7FCvhZg6de65OtXSSJ2KERUVxdOnTwkOCuaLXoaLq50+7GDpcJNl5/YdAHj7GB9KFBoaSr+v+jN81LDX4n3qLeKolDrw3KNbAvt8jWFWhBEpHJtFJfUVrwiw+oUxF38rpVYDVS0Z2Jvq9u1QsmaN39Ngb29PaGhCX35eT+Yox4BvBjJpwo8AZMqUiQWL51OtRjWzxpmU0Nuh2NvH/xC3d7AnNPROkvnbfNCaeg3qkjNXToICg5n4w0Tq1WrAuj/XULlKZUuEHM+rvhYDvhlIvvz5ad+xvSXCS7bMWez4rPenVKhUHju7TBw9cowJYydRp2o9/t3zT6Jj7vJ55ePfzVu5fet2bK9WVFQUhw4cAgyvcWqoXqEmhw8ZbnTzyOfB2r9WGx03mN05O5u2beSD5m0p6l0MAKUU/f/Xj95f9rJorA8jwll7YRMnbp8hPOIR7plz08SzLt+V/4Zvdozg3pP4NwFVdilDRFQEO6/sp3yiVy0T96Fva5RSrL+42RxFiOfG1RvMGjePkpX9KBDdk3sz5BaPwh8z9NMRdOzTDu8iXhzYdoj5ExYSdjeMDq06WiQWc6leoSb+z9WpdUnUqRgd23Rm9e9rAHDK7sSKNcso4FvAorEmx5XLV/hu6Eiq1aiKXwk/o/t+2+9/5Mufj3Yd2qZQdJamXpd5cG9qrRMdExZ9j9VADB2aNkopm+c22yilsgL3MfTeJvSNOObbyGvXwEmqB9cew9CEF50CzHPd6RUopTIqpS4opbRS6ucEtnsrpVYppUKVUg+UUtuVUtVTI9Z3UY9ePdixZzu/rVpB7bq16dSuM+vXbUjtsEwye/4sWrRqQYWKFWjTtjV/b91Ezlw5GTpoeGqHliw7tu9k0cLF/Dh5IiqVl40sUqwIw0cPpW6DOlSoXIHuPT/h1zVLuX7tBjMmz0w0X+euHYmKiqJ7l88IOB9AyNUQ+vXtT+BFw41LqXVZecbcaWzesYnZC2ZiZ2dHk/rNYmNKyM0bN2nfqj0ZbTOyYOl81m1ay1f9v2TcqO+ZMG6iRWO9eC+YX079xqHrxzh5+yzrL/7DqP0/kSWdHfXc4n/pTGtlTdmcJTh0/ViCww+S0sSzDhVdSjPnv6Vce3jDHEWI4+GDcPp1GkQa6zQMmPDsBm4dFcWTR0/o1Kc9bT5phV/5YnTr9yHvtW3AynmrefDA9LKkpOfrVGY7OxonUadiDBs1lC27NvPLsgX4FvShVZPWHDronwIRJy4sLIzWzdtgbW3N1JlTjO67c8culvyylIk/jU/196l3kAeQHvgFQyM15gHwZfT/C2MYa1swgfy+QNDrNv4Wkm7gWgFPE0h/Cq/FV5NhQIJfb5VSnsAuoBwwFvgKyAT8pZSy2BxJ9vZZuXMnfs9gaGjCPYmvK3OUw9XVhRIl/ajfsB6Lli6kdJnS9P9mgLlDNcrePmuCvZyGnl3Tv6PZ2dlRt14dDh44aI7wkuVVXouen35Opw874uLqwp07d7hz5w4RERFERkZy584dHj9+bKmwk6Vo8aJ45vc0+mHs5uHG9HlTOeJ/hBIFS+PrXpj9ew/Q/fNPAMiRwzmlwo3D28ebkqVL0qJ1C9b8tZoHYWFMGDch0f0n/jCJoMBgVv7xG42bNaJSlYoMHDKAz/v2ZMSQkdy6eSsFo4eAe8FcfXAdzyx5420rkb0omdLasvXSHpOPWzNPJdp4N2Xp6dX8e2mXOUKN43H4Y77pMJArQVcZv2Q02XM9+wjI7JAZgFJV4t5cVapKSSKeRhCUjMZiavL28aZU6ZK0TGadiuHu4UaJkn40avoev639FUcnJ4YPTr2rzeHh4bRq2pqLARdZtW4lLq7Gp4vr9WlvOnRuTy7XXK/l+9Rb7jBQLYEHGBq91YBzwBrARSkVe6egUioz8F70ttdOckahv5aLOyil/DDM7vA18EMCu4zC0MtcQmt9ODrPAgzfQiYrpQporc1eNh9fH078dzJe+skTpyjgk/qXjJLLEuXwK1Gcn3+c/KqhmcTH14eTCcwleerkq70eKdnL8CqvxamTpzh18hQzp8+Kty2nowtjfxhDz14J36GdkpL6fTZq+h4NGtXn3NnzpEubFndPd77o+RUuri645nFNoSgTlzVrFtw9PbhwPuEZCgBOHD+Ju6d7vC9WJUqV4OnTp1w4f4FsjtksHWo8Cb0JVnEty73H9/G/YdoCRZVylaFLwTasvbCJ38+b/2pNxNMIvv1oKKeOnGHCsrF4+njE2e7u5WY0/5vUO5g1axY8kqhTCUmXLh2FChfk6NHUWVzq6dOntGvdAf+D/qzesIqChRPq9Ivr9KnTnD51mtkz5sTbljt7XkZ/P4rPPv/UEuFajAKT56FNDVrrOxhWqo0j+m8lUGv9b/TzNRhWLvtFKfUVzxZ6UBg6EV87ybm2N0QpFfn8AxgE8GJ69MPitzQrpdIAM4E/gZUJbLfFMGfbvzGNW4DoLvRZgBeGeX3NrsF7Ddi3dx8BF569KQVeDGT3rt00eK++JU5pEeYuR1RUFLt27sbD0yPpnc2oQcPEyrGHBu81MPl49+7dY8P6P1N0+p1XeS3++ntDvEeRooUpWMiXv/7eQNPmTS0dvlH+Bw9z7sy52DutjUmTJg3eBbxw93Tn6pUQfl+xig+7dbJ8kMlw/dp1zp4+i7uHW6L7ODtnJ+B8QLyx3wf2HQAgZ66clgwxHo8seciVyZlzdy7GSc+Szo6ijr7suLKfSJ382R1KOReje5EO/BO8k19O/WbmaA3vIUM/G8nBnf6MmjuMQiXi33hZuV4FAPb+eyBO+t4t+0iXPh153eP3Vr+url+7zpkk6lRCHj58iP8hf9w93JPe2cyioqLo0qEr27ZsY8mKRZQuk7yP2fWb1sV7FC5SGN+CvqzftI4mzVJuakmRMK11FNAQ2ARMAX4HIoFqr+vaCMnpwTX1K0hKfGXpAxQAEru9tAhgg+HbxotirrmVAsy+3MuHXTszbcp0WjZrxeBhgw0LCwwehmtuV7p2ezbxRGBgEAW9CtG+fTvKVSobm75963Zu3LzJtZBrABw86I9tpkwANEvBxsirlOO7oSO4HRpKufJlyeHsTMi1a8yfM58D+w8w75e5KVYGgM5dOzFt6nRaNXufQcMGoZRi+ODhuOZ2pctzizcEBQZRs0YtBvyvPwO+7Q/AxPGTOHP6LFWqViZnrhwEBQYzafwkroVcY+6ClJt79VVei8pV498IlyVLViIiIhLcZkndOn5CXrc8FClehCxZsnD0yDEmjptEzlw5+fizjwAIDgzGz7cUXw34kprVDCOJnj59yuABQ6NvTrPj1MlTTBw3iQI+3nzWO+V7ddq2aEfR4kUpWLggdpntOHf2HFMmTcXaOg09ext6w3ds20mjOo2ZPONn8rvnB+DDbp1ZvuRXmtZvxud9e+KQzYEdW3fw04Sfadi4Ia65LdcT3bPoh1wPv0nA3SAeRITjnjk3jT3rcvvRHf68GHeFr4oupUljlYZtlxOfPWFx3clsvbyH6ccMKy362Ofj82JdCLx/ia2XdpM/67PG1dOoCC7ee/XPv/H9f2TL2q106NWW9BnTc/zgidht2XM6kT2XEx4F3Kn/fh1mj52HjorCq3B+Dmw/xLrFG+jYpx0ZMmSIzVPFtRZ1W9Wm//hnU7qfOnyaq5dCiIoy9GtfPBPIlnVbAShXvQzpM6Z/5XIk5IPoOlXouTo1ObpO9XiuTr33Qp3q9Wlv7O3tKV6iONkcsxEUGMyMqTMJuXqNGXOnWSRWY/p+/gW//7aKr/p9SUZbW/bt3R+7zcUlFy6uLgQFBlGkQDHatW9H6YqGe58qVakU71hZsmYhIiIiwW1vijfpisGLtNbxgtda3wY+jH689ow2cLXWr8+kgNGUUu7AUGCY1vqiUsotgd1yRf+8nMC2mLQEBwVFT6HRDcDZ2Znd200fgzZixHdMmTyVTu06o7XGz684/fv355j/szmSQ0JCiIyM5NHjx3HO8VXfbzhy5Ejs8+lTpjN9ynQA/tlimbuRE/Oy5Uhvk4HdO9ez5JclPHjwAHsHezw9PZk0aSK5c+V+qd+pOcrRuf2Hz8oxoD/HD/8XpxxRUVEEXQxmz469AERFRLF/735WrljJgwcPsLXNSMFChZg4aSJpsI7dLyXL8DJ16kX37t4jMjLylV6HKBN69mJktsvCyl9XMfXn6Tx+9BgHBwfKli1Lx04dCDwTTOCZ4NgyXA68wsOwcA7tOkxkZCT++w+z9JdlPAh7gKOjI9Wr1+CDtm04fuBE0ic24mXucs7hnJPli3/lypVJRERE4OTkRNFiRRkyZDB3rt/lwPWDnDp6isjISM6fvkAuJxcO7DiIwooJE8ezcMEvfNHzSx4+fIizszPt2rWlZauWHNjx8uO660SUN7o9/V1rqrqWo2Hemlinsebh43AuX7nModNHKP847gwJDXPV4Pa9ULxu58aL3ABk0bZxzpHGKg25tXNsWnGHoqRLkxaPLHkZXj7uqp33H4bx69/xLrTFk9QiDLv+Mvy9LZi0iAWTFsXZ1qZ9a9p2+ACAjzp2wy5NFn6d/jt3Qu+Q3Tk7XT7+kMZ1GxERHhV7nsjIKB7dfBLnvEvGr2DzpmcN/i1rt7JlraGBO3vhTJyTMd57fxrTX8eYOjUxkTq1//pBTr5Qp/bvOEg2e0fW/7GBmdNn8yj8EY6Ojvj4FGDylJ9Jiw37X6FOvUxf1brVhnm5x43+nnGjv4+zrUOHDnTs1DH2b/zxo8fs23EgocMAcP/ufSIjI43uI4QxygLDUC1KKfUXhgasn9b6aXQDNwCYrLXuEb1Pe2AB0EVrPeeF/B7AeWCS1rq3sXOVKOmnd+7dYf5CPGf39j1xenDfVJYuR0rU0z079lK2YhmLniMlvtGnRJ0Kj3ho0eMDHNp1GL/yxSx6jjRWlp8M/8COg5SsaNkhLR9t6mfR49eJKM9f1ua/Yex5n/u1sujxwdCIzlooeUtbv6wi2ZIebvOq9u84SCkL1ymVrBGML2/fjgOxPbiWUrlsFQ4d9E+VblS3Qnn1/1b0T41Tx9HVp/tBY9OEvc1SfqmTV6CUagfUAiprrROa3SFGzKevTQLb0r+wjxBCCCGEWVm9FpNNvbvemAZu9OTD44H1QIhSKl/0ppihBlmi024CV17Y9ryYtISGLwghhBBCiDfcazfG1ogMGOa8bQCcfe7xb/T2dtHPuwLHgMcY5sB9Ucy1WxnYI4QQQgjxFnpjenCBB0DLBNKdMExZ8ScwGziqtQ5TSq0FmimlimqtjwAopTJhaACfxQIzKAghhBBCKN7sWRTeBm9MAzd6zO2KF9Ofm0XhvNb6+e39gRrARqXUBOAe8BGGIQoNLLHIgxBCCCGESH1vTAPXVFrrc0qpCsBooB+QDjgE1NVa/52qwQkhhBDiLaZQ6k0aBfr2eeMbuFrriyQyYZ/W+iQgS6AIIYQQQrxD5OuFEEIIIYR4q7zxPbhCCCGEEK8bmQc3dUkPrhBCCCGEeKtIA1cIIYQQQrxVZIiCEEIIIYQZKSXz4KY26cEVQgghhBBvFenBFUIIIYQwMyU3maUq6cEVQgghhBBvFWngCiGEEEKIt4oMURBCCCGEMCslN5mlMunBFUIIIYQQbxXpwRVCCCGEMDNZySx1SQ+uEEIIIYR4q0gDVwghhBBCvFVkiIIQQgghhBkpQCnpQ0xN8tsXQgghhBBvFenBFW+ElJpuRaZ1SZ50Vuksfg6FlcXPkyI9LEphrSz7Vju28hcWPf7pfRcYW9qy56gz+3OLHh+gZ85m/LRspUXPsaf7LxY9PoAGInWkRc9hZfG3Qk0UURY+g3iXSQNXCCGEEMKslCzVm8pkiIIQQgghhHirSA+uEEIIIYSZyZC31CU9uEIIIYQQ4q0iDVwhhBBCCPFWkSEKQgghhBBmJjeZpS7pwRVCCCGEEG8V6cEVQgghhDAzucksdUkPrhBCCCGEeKtIA1cIIYQQQrxVZIiCEEIIIYQZKcBKbjJLVdKDK4QQQggh3irSwBVCCCGEEG8VaeBaQHDwJdq0aouzQ06y2+fg/RZtCAoKTlbeR48e0f/rAbi7emCfKRtVKlRjx7YdFo44YVLoPo+aAAAgAElEQVQOGDRwMA3rvodL9txksLZl4fyFFo42YS9bhoMHDvHZJz0oWrA4DnaO5Hf3plP7zlwMuGj5oF/w+2+r+KBVO3zyFcIxszPFC5Zg8MAh3L9/P8m8jx49YmC/b/HM44VjZmeqV6rJju07UyDqhF2+dJkven1JtYrVccycHdu0dgReDExWXp98BbFNaxfvsXb1WovFe/VyCIO/HEbT6q0okL0IbnZeBAdeMppnyg/TcbPzokWt1sk+z93Quwz9ZgTlfargla0gZb0r8cXH37xq+LEWNB/DqV4bEnzMbDwcgLK5izG2zlds7DiHw5+tYmPHOQyu1gOHDFmSdY6s6TMzomYfdn20lMOfrWLZ+xOomMfPbGVIyI6tO7C3cYz3yJvdI8m8w/73Hc3qt8AjZ37sbRxZvGCJRWM15u+Nm2lYuxH5c3vjlMkZH/eCdGzTmVMnTiUr//69+2nWsAV5nPKSM6sL5YqXZ8Wy3ywctYUohXoNHu8yGYNrZg8fPqRerfrY2KRj5twZKKUYOmgodWvWY7//XmxtbY3m/+Sj7vy5/i9GjhmBu7sb06bO4L36jfl3xz8ULVY0RcoAUo4YUydPo0jRItRrUJdFCxenUNRxvUoZfl32Kyf+O8mnPbrj6+vD5StXGD1iDBXKVGLPwd3kzu2aYuX4ccJPuOZ2ZciwQeRyzcXRw0cZOXw027ZuZ/O2TVhZJf59+9NuPfhrw0a+Gz0MN3c3ZkydSZMGzfhn2yaKFCuSYmWIcf78BX5bsZLifsUpX7E8mzdtNil/zdo1Gfi//nHS8nvnN2eIcVy8EMgfKzdQqHhBSpUvyfbNxr9sBgUE8fO4qTg6ZUv2Oe6G3qVF7TYopfjif71xzePKtZBrHNxz6FXDjzV0y2QypcsYJ61YzgL0r/wx/wTsAaB14fpkTJueafuXEHw3hLxZc9GzbDsq5vVjw+Y/jR4/bZq0zG8+Gvv0mfl+52xuPAilRcE6TG00lC6/D2Df5WNmK0tCxowfRfGSxWOfW1unSTLPzCkzKVS0MHXq12bpL8ssGV6S7twOpZhfUbp+3AVHJ0cuBQUzYdwkalaqza5DO8iTN0+ief9a/xdtW7anZesWzFowk7Tp0nH65GkeP36cgiUQbxNp4JrZnFlzCbgQwNETh/HM5wlA4cKFKFSgCLNmzKZXn88TzXv0yFGWLVnO9FlT6dCpAwCVqlTCr0hJhg/5jhWrfk2RMoCUI8a121exsrLi/LnzqdbAfZUyfPF1X5ycnOKklStfDp98vsydNZdBQ/9n0dift/z3ZTg5OcY+r1S5Ivb29nTr8gnbtm6narUqCeY7duQYy5f+ytSZk2nfsV1s3lJFy/Dd0JEs/31pisT/vIqVKnDx8gUA5s2eZ3IDN1u2bJQuW9oSoSWoTIVSHLiwG4Cl85Yn2cAd2GcIjVu9x4WzAURGRCTrHGOG/MDDBw/5c8867DJnik1v1KLhywf+gvO3g+KltSxUlycRT1l/eitgaASHht+N3b7/8jEuhl7ml5bjyJfbA4xc+KibryLeju50WPF1bGN2e+ABVredwpcVu9BqWW+zlSUhXgW8KFWmpEl5Am8EYGVlxYVzF1K9gduidQtatG4RJ61EqRKULFya1SvX0LNPjwTz3b9/n08/6kHXT7ow+odRsenValS1ZLgWp+QieaqS376Z/bH2D0qXKR3bEAFwc3ejXPlyrFvzRxJ515M2bVpatHr2BmFtbU3LVi3YtPHvFP0mK+UwMNarmFJepQwvNm4B8ubNg5OTI1euXDF3qEnE4hgvza+k4dLv1ctXE833x7oNpE2bluYtm8WmWVtb07xVc/7etDlVenheh3phClPiXb18Lf8d/o9vhnyR7DwPHzxk5ZJVvN+xZZzGraWlt7ahbr5KbAnYy93HYQBxGrcxjl07A4BtBuNXbIrlLED400fxemp3Bh6iSA5vstsmv0c7pbzuddEhmwNg+JtNzKrfVnPzxk169P4spcIS74DX+y/jDXTyxEkKFvKNl+5b0IdTJ42PQzpx4iRu7m5kzBj3EpxPQR+ePHnC+XPnzRmqUVKO14e5y3Dq5CmuX7+BdwFvc4T3SnZsN/Qkevt4JbrPyRMncXPLG78++RaIrk8XLBqjJWz4YwOOmbNjb5uNqhWqWXT8rSnuht5leL+R9Bv+NVkdsiY73zH//3gU/ghHp2x0b9cTb6fC+OYoxketuxN8MXnj3V9GTc/yZLLJyKqTfxvdr5RrYQBC790xul9kVBQRUZHx0p9EPgXAK1vel4w0ebp1+oRsGbLjkTM/XTt0IzjI+Djp11VkZKThb/PseXp92gfnHM60eL95ovvv2bkHewd7Thw/Qbni5XHI4IivR0FGDx9DZGT810OI5JAhCmZ2+3YoWbPG/2Cwt7cnNDTUaN7Q27cTzOtgbx977JQi5Xh9mLMMERER9Py0F05OjnT6sKO5QnwpVy5f4buhI6lWoyp+JRK/iSc0NJSs9gnUJwf72O1vkvoN6uFX0g8397xcv3aD6VOm07rFB8yaN5M2bZN/Q5cljPx2DO753GjZrlnSOz/nWsi12PxValVm1rJp3L55m7FDfqB1/fb8tXcdmezM37PbxKcGNx+Gsu3i/kT3sU2bgQGVP+bcrSAuXL5o9HgBdy5hZ2OLh31uLoQ+a5gXy1kAgCzp7cwS94syZ8lMj96fUr5SeTJntuPo4WOMHzuROlXqsnXvFpyyx78S8zqrXqEmhw8dBsAjnwdr/1pttAxXr4YQ/jCcrh0+4qsBX1HMrxj/bv6XsSPHcffuXUZ9PzKlQjerd/0mr9T2RjZwlVIOwACgCeAK3AeOA4O01tuf268MMAIoA2hgF9BPa304xYMW4jXQ5/O+7Nm9h9/X/IZ99BeO1BAWFsb7zdtgbW3NtJlTUi2O1PDDpO/jPG/U5D2qVqjO4G+HpGoDd9/O/axcspp1O343+YNZR2kAcrvl5ud5E2Pz53HPQ9PqLfl92Rrad/3ArPFmt3WgXO5iLDy8mkgdleA+af7P3n3HRXG0ARz/DdhRpNoAwYqA2Lti11hjYtBooomxvSZqjL33WKKxxURjbzEm9hpjj9EoGmuiYhewlwgqIMrBvn8cIEiXO8r5fP3cB5mb3X0e9g7mZmdmlRkzmg+lYF5bOq4byDvZk14NYfvFP+hbvTNTmw5k1N7ZPAh5zIeezanioO8BjtQ0g+YQrVyFcnEmS9auW5taXjVpVLspC35YxKjxI4xyXGNZuOxHnj17ht91P76b9T3vtWjL7wd24uyS8CSzyMhIwsLCGD1hVMwwBa96dXj8+DGL5i9m2Oih5M+fslUwhIiW5YYoKKWcgZPAp8B64AtgMuAHOMSqVwM4CBQDxgBjgVLAIaWUp7His7a2Iigo/mWwwMDAZBsUVtbWCW77OKqHKrrHKj1IHpmHoXIYNXw0SxYtZcHi+TRu2tiQIabK8+fPafd+B/xu+LF5+0YcHB2SrG9lZUVQYAKvp6grAVnlPCbG3Nyc973f4/at29y9ey/D4hjRbwztP/GmUJFCPAl6ypOgp0TodERERBIcHMyLFy8T3TZ6OEPtejXjNI4rVi1PPsu8XDh7weDxti7TEHMzczYlMjxBoZjadBA1nSrSe/tELj/yS3afz16G0HfHRKxzW7K103x8/vcrbd2b8r3PTwA8DHlsyBSSVL5ieUqWKsHpE6fT7ZiG4urmSpVqVfDu4M3WXVsICQ5m1vRZidaP/pvw+qSyho0bEh4enuJlxoSILSv24P6EPu5ymqYlPjMFvgNeAnU1TbsNoJRaC/gCM4CmxgjOzd2NC+d945X7XrhIGbcySW7r7u7G1s1bCQ0NjTPe8OKFi+TIkSPOJCNjkzwyD0Pk8M3kacyYPpOZc2bwUSfD9qSlRnh4OJ06fMLpk6fZunMzZT09kt3Gzd2NbVu2x389+V6Kej0lv1ZoVpGRlzSvXrrG1UvXWL0k/jqqH7TxZvTUEXTr3SXBbUu7Jb3EmTLCRKj33Brj+/Aalx7dSPD58Q370rx0XfrtmITPzZRftDt55zxNlnfF2aoIZsoMv8DbdKvszfPwMM4/uGKo8FMsq1/mtrLKT7ESxbl+LeHzBPr3eFIy+0S6hCj0H7JExslSrxqlVF2gDjBN07S7SqnsSqk8CdQrCVQF1kU3bgGi/r8OaKyUKmSMGFu2bsnxY8e5cf3Vm9nfz5+jR47SsnWLJLdt0ao54eHhbFy/MaZMp9Oxft0GGjdpRM6cOY0RcoIkj8wjrTn8MHce48aMZ/zEcXzeu5cxQ01SZGQkXT/pzsEDf7Jm/WqqVa+aou1atGxGeHg4m9ZvjinT6XRsWLeRRo0bpuvryRh0Oh0b1m7EqagThQoVzLA41vy2Kt7DzbMMru6lmTbjG1q81yzRbQs7FKJcpbIcPnAELdZl/JPHTvPsaTDlKxn2olnZAqUoZeuc6OSyoV7d8S77DiP2zGRf1PJoqeUfdIcbgbfInT0n7co2Y+vF/TzXpd+KHadPnubK5atUqlox+cqZ2IP7D7hy6QrFirskWqflu/rfY/v27I9Tvnf3XnLlyoWbR9INYCESktV6cKP/mgcopbYBzQFzpdQVYIKmaT9FPR/9lzOh32w+QFegMpD8OlGp1LX7Z/w4bwHt2rZn7ISxKKWYMHYCjk6OdO/ZLaaev38AHqXL0rlzJ2p61QCgQsUKeLf3ZvCAoYSH63BxcWbhgsX43fBj2colhg5V8ngtj8aNmjBy9AhGxFqA/9DBQzx89ChmEs3Jk6exyKufLNP2g/czXQ6vn4u1v65j8IAhNH2nCfUa1OOYz/GY+paW+ZLtOTGk/l8OZNOGzQweNggLCwuOH3s1McjBoQgOjg4E+AfgWaYCw0YOpWH9RoD+Uu0H7doydNAwwnXhOLs4s3jBEvz9/Fm6clG6xf+6TRv0De7TUZNpdv++Bzt7O+zs7fCqWwcAy1xWNGnahA119Gs/r/1lHTu27aBps6Y4Ojny4P4DFs5fyJnTZ1j+0zKjxvvbZv1NDv49cx6AP/b8ia2dDTZ2NtSoU42aXtXjbWOZ35IInY7yFcpT2OFVn0C98o1xcCrCz9tXxpQNHT+IT97rxued+vLhp+14/Ogx306YRYnSxWnTvrVBc2nj1ojwCB3bLh6I91z3yu34rNIHrD+/C/+gO5Qv9Ooqh6VF3Eli5/puZ7PvXkbtnR1TNqBWF84/uErg8ycUtSpCt8re6CJ1zPzLeOenx6f/w9nFmfIVy5E/vyX/nPmXWdPnUNihMP/r3ROAAP+bVHKrwpCRg2hUv0nMtn/9+RePHv3H/XsPADh98gwWefXLobVp+67RYk7Ix96dKF+xPB6eHuSzzMfVK1eZN2c+2bKZ0/cr/Rq4h//8i3ffacOgwQOpUqcyAO5l3fn4k4+YPH4KkZGRlK9Ynj/2/cHKpasYMmIwefOm39JzhqMwy+K971ldVmvgRq9rtAi4gn4cbg5gILBKKZVd07RlQJGoerfj7yKmLMGBf0qpnkBPgIIFC3L0kE+qg5w06Wvm/TCfLp0+Q9M0KlWqyPDhw/n39LmYOvfu3SMiIoKwFy/iHKNb164QqTFy2CiCg4MpUaIEU6ZOJizkxRvFkhZvWx6RkZH4+wXEiW/wgKGcPXs25vsF8xawYN4CAPYfSN3i/umRw+vn4ueVP6NpGrt37WH3rj1x9lm+fHlmzZ75RvG8yVyb7Vv0nyenT/2W6VPjTrT65JNP+LTLpzE53PS7RUhwCMcOnwCgW9fuoClGDx8b83qaPGUyL0J0MXXeSBr+/nTq0DnO91/17Q/of64zo36uERERvHzxEp/D+g8WwU9DuHrlGnv3DOXZ02fkypULV1dXpn4zFSeHojH1UksXGZ5snS86x70hyOj+4wAoV96T6TOnJ7jN82fP9a+pkBdcOv5qObawkDCCg0LilNnlKcT4ieNYuWIlPTt8Qa5cuahWoxo9enbH/9/k11zuWzhlKzeYKYV3mabcun+bj60awWsLbLxXWt/T7O3xDt4e78R5zs/fj74hr24Nnc3MHI88xeIc28uuAh95tCR3ztw8f/Gc67f9OX7+JJ/YpGxU28m/Uj+H2TJvfjau3cT8uT/yIuwFNjY21KhRg0+7fIrfpZv4XboZ89645XeH0ODQmOOMGDSaf87+E7OvxT8uYfGP+o6EvfuTXj4tKW/y1ihUsDBrf17HnTtz0Ol02NvbU75CecaNG0vQgyeceHCSi/9c1L+mnr/gxOGTMdt2/rgzRMD3s34gMDCQgoUK0uvzXjRp2CROPSFSSmlGmhVqDEqpvUAj4Drgpmnay6hy66iyMPQN15HABKCRpmn7X9tHQ2Af0F/TtNkkoXKVStpfx5K+409aHT3kE9PblpWZQh6mkAOkTx4RkSm7u1VaHDt8gup1UndXp9RSyvijtHwOH6dGHePetexhWFLTEdLu0vHruFYz7ljnd5YkfVdBQ+hbuC1z725MvmIa+Hz+U/KV0ujkX2eoXLuCUY9hppK/TXBanDh8MqYH11jq1WjA6ZOnM6QbtVS5ktrs7dMy4tBxtHL+4KSmacb9RZpJZakxuMDzqK9rohu3AJqmBQJbgULoe3lDo55KaHBerqivoQk8J4QQQgiRZioT/HubZbUGbvRtXRJaSye6C8MaiL4eltAwhOiyhIYvCCGEEEKILC6rNXCjB6g5JvBcdNkDIHr2Ss0E6tVAf9MHGdQjhBBCCKNQSmX4422W1Rq4m9HftayTUipmWqVSqjD6u5pd1jTtqqZpV4ETQDulVJFY9YoA7YD9mqZl3IrqQgghhBDCaLLUKgqapgUqpQYBCwAfpdRS9KsofB71tW+s6v2AA+jvXDY3qqwv+kb9wPSLWgghhBBCpKcs1cAF0DRtoVLqETAEmAhEol/v9iNN0/6KVe+IUqo+8HXUQwOOAO00TTsbb8dCCCGEEAagv5NZVrtIblqyXAMXQNO0jUCya71omnYU/bJiQgghhBDiLSEfL4QQQgghhEnJkj24QgghhBCZl6xikNGkB1cIIYQQQpgU6cEVQgghhDAws7f8TmIZTXpwhRBCCCGESZEGrhBCCCGEMCkyREEIIYQQwpAUMsksg0kPrhBCCCGEMCnSgyuEEEIIYUD6O5lJD25Gkh5cIYQQQghhUqSBK4QQQgghTIoMURBCCCGEMDCZZJaxpAdXCCGEEEKYFGngCiGEEEIIkyJDFIQQQgghDEqhpA8xQ0kDV4h0FKlFGv0YWjocR6l0+MWtjH+cZ+FPjLp/gEgtwujHsclpb9T9Z1MBRj/GwZ6LjLp/gH99fDnY2rjHWe673Kj7BygS5mj045SyLm7U/evCFYfuHjDqMYLDnxl1/yJzkwauEEIIIYSBmckkswwl/edCCCGEEMKkSANXCCGEEEKYFBmiIIQQQghhQHKr3ownPbhCCCGEEMKkSANXCCGEEEKYFBmiIIQQQghhYHKr3owlPbhCCCGEEMKkSA+uEEIIIYRBKZlklsGkB1cIIYQQQpgUaeAKIYQQQgiTIkMUhBBCCCEMTCaZZSzpwRVCCCGEECZFGrhGcPPmLTq2/5iCNoUpYF2ID707EhBwM0Xbjhk5llbNWuNQwInc2SxYtWKVkaNN3JvmcfLEKXr36kN5j4rY5LOjVDFXunT+DL8bfsYPOgGmkMetm7f4+MNOFLZ1oJBNETq2+4ibKXxN+d3w4+MPO1HEzhH7/AVp3rgFp06cMnLECbt96zYD+w2iQZ2G2FkWwCJ7Pvz9/FO0rVtJDyyy54v32LZlm9HivXPrDiMGjKJl/XcpZluSQnkcCfCP/3MPCwtj/IiJlCtWCRebErSs/y5HD/uk6Bi//rSObh17UNm1OoXyOPJlz/6GTiNBe3fvo1XTdynp5Ipd3oKUKebBpx0/4+KFi0lu5+8XgGUO6wQfQUFPjBrznVt3GTFgNC3rt6GYbSkK5XFK4nx8TblilXGxKUnL+m1SfD769RyAV8UGlCzoRnF7VxpWb8ri+cuIiIgwdDoxFnyxiP4VBvHb9ztjyi4fu8JPI37m61ZTGFJ9GF+3msK6SRt49vhZsvsLCwlj+eCVTGo9haE1hjO8zihmfTyHEztOGizmkwdPM6LDGDpV6cp7pdrzafXuTP3iWwIuxz0fD+88YnKvabQv24l2Hh8zqec3PLj9MEXHuBdwn8m9pvGhZyc+KNOR4R+O4co/Vw2WgzEpwCwT/Hubvd3ZG0FoaCjNm7Tg8qVLLFq2kCUrFnPt6lWaNW5OSEhIstvP/+FHnj8Po3nLZukQbeLSkse6X9dx4bwvX/T5nC3bNzFx8gTOnD5L7epe3Lx5K50y0DOFPEJDQ2nRtBWXLl1m4dIfWbx8IVevXKN5k5bJ5vDff//RuH5TLpy/wHfz5rBi9TIAmjdpyUXfpBsyxnDt2nU2rN+IlbU1terUSvX2jZs25sChfXEederWMUKkejeu+7F143byW+Wneu3qidYb8PkgVi9bw5DRA1m1YQUFCxWg47sfc+7s+WSPseGXjfhd96deQy/yWeYzZPhJCnwcSIVK5fl29jQ2/7aRsRNH43vhIo28mhLgH5Ds9gOH9Gfvod1xHvny5TVqzHHPR7VE6w34fHCs87E86nx04urVa8keIywsjK69urDop/ksXbMQrwZ1GD1oLGOHTjBkKjFO7TzN7ct34pUfWXeUkCchNOneiJ4/9KBx14ac/+M8szvPJex5WJL7jAiPwNzcnEZdG9JtTlc6T/mYAsUKsHrkGv5Y9adB4g4OCqakZwl6je/OxFVj+XRoJwIu32Tg+8N4cOsBAGHPXzCi41huXbtN/xl9GTirH3f87jKiwxjCwpLO4WngM4Z4jyTg8k16T+7FkO8HADCiwxhuXknfvyMia5IxuAa2dPEybly/wT8XzlCiZAkAPD3LUrZMORYvXEK//l8muf39x3cxMzPj2tVrrF71c3qEnKC05DFwyADs7e3jlNWsVRO3ku4sW7yMMeNHGzX22Ewhj2WLl3Pj+g3OnD8Vk0NZz7KUc6vAkoVL+bJ/30S3XbxgCQ/uP2D3/t8pXqI4APUa1MOjtCeTJkxm1ZqV6ZJDtDpetfG7fR2A5UuWs2/PvlRtb2trS7UaiTdsDK1mnRqc8zsDwOplP/PH3oPx6pz/5wIbf93MrB9n0PGTD/XbedWgXuWGTJv4LYMHDU7yGL9sXY2Zmb6v4cCePwybQBLadfCmXQfvOGVVqlamsmc1tmzcSt/+fZLc3qW4C9WqVzVmiPHUrFOdc36nAVi9bA1/7I3fWHt1Pr597Xw0YsWyFbzfqVWSx1iwcl6c7+s3rsf9u/dZs/JXvv52vIEy0Qt9Gsrmb7fy3qB3WTV8dZznvEe0Ja/Nqw8MJauUwN7Zju+7zefooaO4liiV6H4trCzoPPXjOGXuXm48DHjE8S3Hqd+5bppjr9fGi3ptvOKUlS5fil4N+3L4t6O07dmGXWv2cD/gPj8emEsRl8IAuJRxpmf93uzasYsPKryb6P5/W/U7QY+C+Gbd1xR2LgRA+VqedPf6nNWzfmHYvEFpzkGYNunBNbAd23ZQrXq1mIYIgEsxF2rWqsn2rTuS3T76D11GS0serzcKAZydi2Jvb8edO/F7KozJFPLYsf03qlWvmkAONdixLekcjh87TslSJWIatwAWFhbUrlOLnTt+R6fTGSvsBGWW13dKpSTeXTt2kz17dtp4v/pjnS1bNtp4v8sfew/y8uXLNB8jvdjY2gBgni1z9n2k9XycPHGSFy9epPq41jbWZDPCz2Tb7B0UKlmISs0rxnsuduM2mpNHUQAeP3r8RsezyJ8HM3Pjvd7yWeuvQJhnMwfg2J6/ca1YKqZxC1CoaEHcq5Th+NHjSe7r0unLFHEpHNO4BciVJxfuVd34e98JInTGGzJiEEo/ySyjH2+zzPOb1UT4XvDFo6x7vHJ3D7cMuST8pgydx0Xfizx48BDXMq6GCC/FTCEP3wu+uHvEz8HNPfkczM3NyZ49R7zyHDly8vz5c65fu26wONPDzh07sbMsgLWFLfVrNzDq+NuUuuR7maIuTuTJkztOuau7Ky9fvuTO7fT9UJdaERERvHz5kqtXrtHvi/4ULFSQdh9+kOx240ZNwDq3HY52Rfnw/Y6c/zf54RjpIfHzUZrw8HBuXPNLdh+apqHT6XgS9ITtm39j7er1/K9vd4PGef30DU5sP4n38PdTvM21k/ohFg5FHVJUX9M0InQRhASFcGS9DxePXqJep7T33sYWERFB+Mtwbt+4ww/Df8Ta3op67+qHDQVcuYmza9F42xQt5cTNgKSHGZiZm5EtR/wPFdlzZOdF2Evu+t8zTALCZGXOj+lZ2OPHgVhZWcUrt7a2JjAwMAMiejOGzEOn09H3i37Y29vRpeunhgoxRUwhj8DHgVhZJ5CDjTWBgUFJbluqdCn27z3Af//9h62tLQCRkZGcPHEyZt9ZRYuWzalUpRIuxZx5cP8hC+YtoIP3RyxevoiOH3fIsLiCAoPIb5U/Xrl11Dl79iz5SUEZqWHtxpw+pR+GUbxkcbbv2oJ9gfhXL6LlzJmDrj260LBxQ+zsbbl86QozvplJk3rNOPDXXlzd0vdD7OuSOx9BybxnAPbs3Mcn3p8B+l64vgO/YMDwrwwWoy5cx9qJ62nwST0KuBRI0TZhIWFsnr6FgsULULVmVe5zN9ltDv/6Fxunbgb0varvD25D1dZV0hT76wa2GcbVf/UN78IuhZm8ZgJWdvqfdXBQMHnzx++JzmeVj+BnwUnu16G4A2cOneVp4DMso3qGIyMjuXxWP8nsWVDS2wuR5Rq4Sqm8wJdAR8AFeAFcBhYCKzRN02LVrQ5MAqoDGnAEGKZp2pl0Dvut1v/LAfgc9WHT1g1YW1tndDhvLCvm0b1nN+Z//yM9PuvJt7OmkztPHqZPmR6zEkRmumL7GsQAACAASURBVDyenBlzvo3z/bvvtaZ+7YaMHTUuQxu4Wd3CZT/y9Nkz/K77MXfW97Rp0ZZdB3bi7BK/5w2gUOFCzP5hVsz3terUonHTRlSvUIvpU2eweMXC9ArdaGrUrsbvh7bz7OkzDh04zPw5C1FKMXz8UIPsf//yPwh/EU7j7o1TVD9CF8GqYat58uApXy7vjbm5eYq2q9i0As6ezoQEhXD+j/Ns/GYzZuZm1PKumZbw4xg4qx+hwaHcC7jPxoVbGNVpHNPWT6agU8oa7olp0akp25bvYGb/OfxvfHdy5srBr99v4P7N+wCYmWX2y+9yq96MlnX+ugFKKTNgJzAR+BsYCHwNmAPLgKmx6tYADgLFgDHAWKAUcEgp5WmsGK2trQgKit9DEBgYmGUaRWC4PEYNH82SRUtZsHg+jZum7Je5IZlCHlbWVgn2OgU+DozplUpMseLFWLpyCadPncGzTHlKFi3FMZ/j9OnXG9A3VrIqc3Nz3vd+j9u3bnP3bsZdrsxvlZ8nCSyPFd27ni9f+q2M8CZc3VypWq0K7Tp4s3XXFkKCg5k1fVbyG8bi6ORIjdrVOXXitJGiTLnkzkdCV0NeZ5nfkgqVy+PVoA4jJgzjy8F9mDtjHndvJ99rmpzAu4HsXbyXFl80I+KljudPn/P86XMAdFHfR0ZExtSPjIzk59G/cPnYFbrO6kKR0kVSfKy8Nnkp6uGEW+0yeI/8gCotK7N15nYiwg03ftWplCOuFUtTr40Xk34eT1hoGOvmb9QfP78FwU/i97Q+C3pG3mRW3ChUtBCDZn/F1XPX6VH3Cz6p1p2Lpy7xXrfWAFgXyDp/T0XGyGo9uNWBOsBsTdNiFotUSs0DLgL/A6I/Yn8HvATqapp2O6reWsAXmAE0NUaAbu5uXDjvG6/c98JFyriVMcYhjcIQeXwzeRozps9k5pwZfNTpI0OHmCKmkIebuxu+F+LncNE3ZTm817YNrdu04srlK+TIkYPiJYrTr/dXODo54lTUyRghp7uMnEzh6laanVt/JzT0eZxxn5d9L5MjRw6KOKS8QZLRrKzyU7xEca5fu/FG22eGSS2ubq7s3LorgfNxhezZs1OshEuq91m+UjkiIyMJ8L9JYYfCyW+QhP9u/Uf4Cx0/jYy/Ss6BlQc5sPIgg37pj0MZ/TjbdV9v4Mzus3SZ/gmlqye+ckJKOHk48ve2Ezx7/Ayrgsk39FMrb34LCjsX5q6f/gNn0VJO8dbFBQi4egunoo7J7q92i5rUeKcad67fJVuObBR2LsQPIxdgX8SOAg6JD6PJLDLD++FtlqV6cAHLqK9xZm1omvYSeASEACilSgJVgXXRjduoereBdUBjpZRRuq5atm7J8WPHuXH91R8Ifz9/jh45SsvWLYxxSKNIax4/zJ3HuDHjGT9xHJ/37mXMUJNkCnm0bNWC48f+TiAHH1qk8DVlbm5OGbcyFC9RnLt37rJh3Ua6/6+bsUJOFzqdjg1rN+JU1IlChQpmWBxNWzQhPDycbRu3x4lt64Zt1GtUlxw54k/yy6we3H/A5UtXKFbcJVXb3Qy4ic9fx6hctZJxAkuFpi0aJ3o+KleuRM6cOVO9z6OHfVBKJTpsIzUcXB3ovahXvAdA5ZaV6L2oF3ZF7QDYMmMrxzYdp8P4D/FsWDbNx7524jo58+RMcIUGQwh8GMSta7cp7Kx/P1ZvUpWLpy9zL+DVFZb7Nx/ge+IiVVO43J+5uTlOpRwp7FyI/+4/5tC2v2je6R2jxC9MS1brwT0OBAFDlFJ+wDEgD/ApUBmIboFEL854NIF9+ABdo+onv25XKnXt/hk/zltAu7btGTthLEopJoydgKOTI917vmpQ+PsH4FG6LJ07d6KmV42Y8kMHD/Hw0SPu39OPMzp58jQWefW/jNp+kPLZthmZx9pf1zF4wBCavtOEeg3qcczn1XIwlpb5cHN3y5R5NG7UhJGjRzBi9PBMlcdn3buwYP5C2rftwNgJo6Ny+BpHJ0e69egaUy/AP4CyruXo1LkTNero/3iEh4czcthovOrWJp+lJb7nffl22gzc3MskuyazsWzaoJ/0Ej2xaffve7Czt8PO3g6vqJs2WOayoknTJmyosw6Atb+sY8e2HTRt1hRHJ0ce3H/AwvkLOXP6DMt/WmbUeLdt0jeUzp7+F4D9uw5ga2+DrZ0ttbxq4lmhLG2832XMkHHowsMp6lKU5YtWEuB3kx+WztXPEohSo2xtHIs6sv63X2PKLvle5vLFy4D+JgO3Am7FHLNmnZrY2dsaJa+PvDtRvmJ5ynp6kM8yH1evXOWHOfPJls2cPl/p18A9/OdftH6nDYMGD6RqncoAjBgyisjISKpVr4qdvR1XLl9h5rRZmJkpBg8baJRYY9u2Sf9r++zpf4Do82GLrZ1NrPPRmjFDxqML11HUxYnli1YR4HeT/v3j3iWuRtk6UefjF0A/ueyXVWtp2qIxDk4OhDwLZt/uA/y09Gc6d/uYQkXS3i+S2zI3JauWTPA5m8LWMc/tW7afP1b9SfX3qmFf1A6/f17d8c8sPBvEujAwsPIQqrauQodx7QE4sv4o/v/4U7p6afIXzE/ok1DO7D7L2b3/0KpfC7JlT/uf/q97TqWER3GKubmQJ29ubt+4w5Yl2zHPZsb7PdoA8E7HJmxfsZOJ3afSedBHKAWrZqzBrrAd77R8dRH1wa0HdK/7BR37tadjP30OunAdy6aspGx1D/LkzUPA5QDWzdtI0dJOvN8j8fVzhYiWpRq4mqYFKqXeBRYDa2M99Qz4QNO0zVHfR7/1bxNfdFnK1llJJQsLC3bu+Y0hA4fS7dPuaJpG/Yb1+XbmNPLmjfWpWdOIiIgg8tWcOAAmjp/EoT8PxXy/YN4CFsxbAMBzXfJ3QjOUtOSxZ9ceNE1j96497N61J85+vep6sXv/7+mUReryiIyMJDLy1di3zJKHhYUFv+3eztBBw+jepWdUDvWYNuObODloUeci1jxLlFJcu3qVtb+s5UnQExwcHfikS2cGDxuUYT2LnTp0jvP9V331jQ6vunX4fZ/+VqURERFxzoWLizMPHjxk5LBRBD4OxMLCgoqVK7J5xyaaGHlMdI+P4/bcD/tqBKC/ecCmXesBmL1gBlPHTWPq+Ok8ffIUd083ft6yinIVPfnX50LMtjpdRLxbvm7dsI0Zk1+NeT3y51GO/Kn/bL7h97XY2af+jm8pUbV6FTat38z3s7/n5ctwHBwd8KpXmwFDBsT0VEa/piIjX72m3NzLsHjBUn5e+TPBwSHY2NpQt74Xw0cNpZRr2i6hp0T88zESiD4f+g9EiZ0PyxxxL8u/fj5cijujRUbyzfjpPHr4H5ZWlhQvUYy5i2fzfvs2Rs4sLt/D+iUAj20+zrHNcdeMrde4Lu9VftXIi4yIjDNut3DJwpw7cJ4ts7YR+iSUvFYWFChekO7fdcWjbvwlB99EmYqlObT9CJsXb0X3UoddETs8a3jQ7osPYiaY5cqTi0lrxrN4wjJm9J8Dmkb52uXoMaYruUNyo5/7DZoWlUOs97xSijs37nJwyyGCn4ZgV8iWJu0b0a7PB2TPkd0gORibTDLLWEp7rYGV2SmlKgKjgOvoV0WwAXoDZYA2mqbtUUqNBiYAjTRN2//a9g2BfUB/TdNmJ7D/nkBPgIIFC1Ze84tx7yYWHBxC3rwWRj1GejCFPNIjh/R4t4UEh2CRxc8FpE8ekZrxF4t/HhxG7ry5jHoMY99zPiQ4FIu8eYx6jEgik6+URulxLh6HvdlNGFIje3gOwrMnfQORtMppnvqhHKkSBhj3VDBo0CCu/HM1Q1qZbhXKaCv3LM6IQ8dRrYDXSU3TDLs2XBaRpXpwo1Y/OIK+cfpjrPI1wDlgkVKqBBAa9VRC79Dot1RoAs+hadpC9EuOUblKJS328AFjOHrIB2MfIz2YQh7pkUOkZvw/4j6Hj8cMUcjK0iOPZ+HxZ9sb2r8+F/CsYZhes8TkyWbcDwJ/Hz4ZM0TBWEJ1xl/X9F8fXzxrGHdo0ZrLvyZfKY2K3HHkTpGkb5SQVqWsiydfKQ10FxXZymStDjaRtWS1SWb90TdQ18Uu1DQtFP14Wmf0a+NGT0JLaBhCdFlCwxeEEEIIIdJEEb0Sbsb+e5tltQZudOM0oVWus8X6+nfU/xNazboG+ivFJw0bmhBCCCGEyAyyWgM3erZGl9iFSikroA0QCFzVNO0qcAJop5QqEqteEaAdsF/TNLmRtRBCCCGMQ6mMf7zFstQYXGA28AkwNWo87l/oJ5n1AAoDvTUtZtZIP+AA+juXzY0q64u+UW/8tWyEEEIIIUSGyFINXE3T/JVS1dDfercR0AF4DpwBBmqatjFW3SNKqfrob+X7NfphCUeAdpqmnU3v2IUQQgghRPrIUg1cAE3TrqG/sUNK6h5F3xAWQgghhEgnMskro2W1MbhCCCGEEEIkKcv14AohhBBCZHbqLZ/kldGkB1cIIYQQQpgUaeAKIYQQQgiTIkMUhBBCCCEMTCaZZSzpwRVCCCGEECZFGrhCCCGEEMKkyBAFIYQQQggDkyEKGUt6cIUQQgghhEmRHlwhhBBCCANSyDq4GU16cIUQQgghhEmRBq4QQgghhDApMkRBCCGEEMKglEwyy2DSgyuEEEIIIUyKNHCFEEIIIQxMZYJ/ycaolLdSaoNSyl8p9VwpdUkpNUUple+1etZKqcVKqUdKqRCl1F6llKfRfngGIA1cIYQQQoi30yAgAhgBNAPmA58De5RSZgBKvxzEtqjn+wIfANmBA0opx4wIOiVkDK4Q6Sg9xmSpdDiOhmbU/aeXHGY5jX4MhTL6cTTN+OfD2MeI1CKNun/Q52Ds4+TJnsuo+wcwU2ZGP86TF0+Muv9cWj5CXjwz6jEitAij7t9EtNY07WGs7w8qpR4DK4D6wH7gXaA20FDTtAMASqmjwA1gCPBlukacQtKDK4QQQghhSEq/Dm5GP5LzWuM22t9RXx2ivr4L3Ilu3EZt9wR9r26bNP6kjEYauEIIIYQQIlq9qK++UV89gHMJ1DsPFFVK5U2XqFJJGrhCCCGEEKbJTil1ItajZ1KVlVIOwARgr6ZpJ6KKbYDABKo/jvpqbbhwDUfG4AohhBBCGFgmWQf3kaZpVVJSMaondgugAz4zalTpQBq4QgghhBBvMaVUbvRjaosD9TRNuxXr6UAS7qW1ifV8piMNXCGEEEIIA1KQoklemYFSKjuwHqgCNNE07d/XqpwHmiawqTsQoGlasJFDfCMyBlcIIYQQ4i0UtdbtaqAh8J6maT4JVNsKOCil6sXazhJoHfVcpiQ9uEIIIYQQb6cfgHbAJCBEKVUj1nO3ooYqbAWOAj8ppQajH5IwHH1H9bR0jjfFpAdXCCGEEMKgMvomvSm7VS/QPOrrSPSN2NiP7gCapkUCrYA9wDxgE/q7nzXQNO2mYX9uhiM9uEIIIYQQbyFN01xSWO8x0DXqkSVID64QQgghhDAp0oMrhBBCCGFgmWQd3LeW9OAKIYQQQgiTIg1cI7h58xYd239MQZvCFLAuxIfeHQkISNk47LCwMIYPGUExx+JY57WlXu0GHP7zsJEjTtib5nHyxCl69+pDeY+K2OSzo1QxV7p0/gy/G37GDzoBpnA+bt28xUcffkwh2yIUtClMh3YduZnCHPJkz5vg4+yZf4wcdXy3b91mYL9BNKjTEDvLAlhkz4e/n3+KtnUr6YFF9nzxHtu2bDNy1Anb8/seWjZqjZOtM0Xti9GwVmP+PHAo0fqnT57hqy8GUL1cTRysi+JZsgI9P+2F/42U5W9sbVt5kz+nDRPHTkq2blhYGKOGjaG0sxsF8xehcd2m/HXoiNFjvHP7LiMHjqF1g/cobudKEQtnbvrHfR+cPfUPg/sMw6tiQ4rbuVLFtSa9P/uSAL+AFB9n9bI1eFVsiIt1KepUaMDKxT8ZLIfzhy8w87PvGFx3OL0rfMXQhqNYOGAJd67ejVPv0rHLTOs0kz6V+tO/5hCWDlvB00dPU3SM/3n0SfBx0/dW8hun0D+H/mXSJ9/wRa1+fOrRgz5eA/iu3zxuXb0dr+6ZP84y4aMpdK3Qi24VP2dU2/H8e+b1pVbje/kinJ+/+ZXetb+ii2dPxrb/Gt+/LxksB2NTSmX4420mQxQMLDQ0lOZNWpAzZw4WLVuIUorxY8bTrHFz/j59DAsLiyS379Xjc37/bReTv5lEsWIu/Dh/Ia1btOGPw/spX6F8uuQAactj3a/ruHDely/6fI67uxu379xh6qRvqF3dC5+TR3FycswSeUDmOB+hoaE0b9qSnDlzsHDpApRSTBgzgWZNWnD8lE+yOQB0+qQT3XvEnRtQqnRJY4WcqGvXrrNh/UYqVqpIrTq12LdnX6q2b9y0MSNHD49TVsq1lCFDTJHli1YwpP8wun/ejUHDBxKpRXLu7DmeP3+e6DYb123i4oWL9OzdgzJurty9c4/pU2bQsHYTDh47gKOTQzpmENf6Xzdw7t/zKa7f539fsnvnbiZMGY9LMRcW/biEtq282fPnLsqV9zRanH7X/Ni2YQflKpaleq1qHNz3Z7w6m9dt5ZLvZbp93gVXt9LcvXOP2d98R3Ov1sz9YS5QJsljrF62hiF9h9N30Bd4NajD4T/+YvhXo9A0jU97dE5zDiFPQinq4US9Dl7ks8nL47uB/L54N998NIMxm0dgW8SGKyevMrvn93jUduN/s7sTEhTClu+2MavbXEasG5Ki49R8rzp129eJU1bQpUCa448W/CSEYh4uNP6oIZY2+Xh05zHbFu5gbLuvmbp9IvYOdgDs++UAKyaspkmnRrzf+120yEj8fQN48eJFssdYNGIpZ/44S8chH1LAyZ49q/fxTdcZjPt1FC7uRQ2WizBN0sA1sKWLl3Hj+g3+uXCGEiVLAODpWZayZcqxeOES+vX/MtFt/zn7D7+uWcuCxfP5pMsnAHjV86JSuSpMHPc16zevS5ccIG15DBwyAHt7+zhlNWvVxK2kO8sWL2PM+NFGjT02Uzgfy6JyOHv+dJwcPN3Ks2ThUr7s3zfZfRRxKEy1GtWMHWqy6njVxu/2dQCWL1me6gaura1thucR4BfAiMGjGD9lLJ/37RVT3qhJwyS36zewL3b2dnHKqtesRoUylVm5dBUjxg4zSrzJCQwMYvjgkUyZPolun/RItv6//5xj3S/r+WHhXDp9+jEAderWpnqFWkweP4VfNv5stFhr1KnOP34nAVi9fE2CDdw+Az7H1t42Tlm1mlWo7l6HnTt20vDduonuX6fTMXXcdLw7tmXYOH1Dsna9Wty7e59pE2fwUZcOZM+ePU05VGtZhWotq8Qpc/F0ZmyriZzafZomXRqxfd5ObAvb8Pl3PTHPZg5AoeIFmfLhdP7acJS2tb2TPY5VASuKly+WpliTUqtVDWq1qhGnrES5YgxuNoLjv5+gZbdmPLz1iFWT1tBxSHuad3l1I6xyXp7kup6PMJ4lun9/3wCObPOh55Su1PvACwC3aq4MbTmKDd9tYuCP/YyTmDAZMkTBwHZs20G16tViGiIALsVcqFmrJtu37khm29/Inj073u1f/fLKli0b7dp7s2f33hR94jWUtOTxeuMWwNm5KPb2dty5c8fQoSbJFM7Hju2/JZJDDbZv254uMRiKmVnW/5Xz04qfMTMz47MeXVK13euNWwAnZyfs7O24e+duAlukj7EjxuHm7ob3hx+kqP7O7TvJnj07bdu9H1OWLVs2Pmj/Pvv27Dfq+yIlr5/XG7cAjkUdsbWz5dGj/5Lc9uSxU/z36D/adng/Trl3x7YE/hfI8SN/py7gFMprpb8KY2auz+/G2Ru41SoT07gFcCnrjIWVBWf2nTVKDIaQzyovAObZ9Hkc3HAIMzNFo44NUr2vU/vPYJ7dnBotXn2gNc+m//6fQ+cIfxlumKCNKKPXwH3bJ7ll/b82mYzvBV88yrrHK3f3cOOi78Ukt71wwReXYi7kyZMnTrmbhxsvX77k2tVrhgw1SWnJIyEXfS/y4MFDXMu4GiK8FDOF83Hhgi/uHm7xyt3cU34uFi9YgpWFDbaW9jRv0oK/Dv9l6DDTxc4dO7GzLIC1hS31azfIkPG3x44co5RrSTau3UQlt6rYWxSisntVFv+4JNX7unTxMg8fPKR0mdJGiDR5R//y4ZfVvzJjTspvRuR74SLOLs7x3hdl3Mvw8uVLrl+7bugw0+zKxSs8eviIos5OSda75HsZgDLucc+Hq5v++8sXrxgspsiISHQvddz3f8BP49ZgaWdJ1Rb6nl1lbka27PEvsGbPkY3bV1L2YejPXw/Ru8JX9Kncn5mffceVk1cNFnts0Xnc87vHkjErsLLPT82ont1LJy9TuHhhju44Rv9GQ+js1o0BjYey+6fkr9zcunKbAo725MydM065YykHdOH6n5sQSZEhCgb2+HEgVlZW8cqtra0JDAxMctvAx48T3NbG2jpm3+klLXm8TqfT0feLftjb29Gl66eGCjFFTOF8BD4OxDrqmLFZ21gTGBiU7PYdP+pA85bNKFykMAH+N5k9YzbNm7Rk++9bqVsv8cu1mU2Lls2pVKUSLsWceXD/IQvmLaCD90csXr6Ijh93SLc47t29x9279xg7Yjyjxo+kWHEXtmzcypCvhqHT6ejV538p2o9Op2Ngn0HY2dvRucvHRo46vpcvX/JV7/707d8nVeOYAx8HYmWdP1559Gs08HHyr8n0pNPpGNpvJLZ2tjRr3izJukFR76f8r+VnZWMV53lDmNJxOgHn9RPkChS1Z8DSL7G0zQdAIZeCXD97I079/+485snDp3F6dRNTvXVVPOuVxapAfv6785g9S/cxs+t3fLWoD67VDPthaky7idw45wdAQecCjFgxhPy2lgAE3g8i8EEQa775lfYDvClY1J5jv59gxYSfMOuVncYDEv/9E/wkhDyWeeKV582v7+0ODgoxaB6GppBlwjJahjdwlVLDgUpAZaAY4J/UnTWUUtXR3zO5OqABR4BhmqadSaBuEWAq+lvR5QXOA99ompZ+g1kF/b8cgM9RHzZt3ZBgQ00Y15IVi2P+X7sOtHq3JVUqVGP8mInsO7gnAyNLnRlzvo3z/bvvtaZ+7YaMHTUuXRu4kZGRBD8L5odFc2n9XisA6jbwIsA/gNnT5/C/3j1TNHt5yFfDOO7zN79s+hkr6/gfpIxtzozveP48jEHDBqT7sdPTyAFjOOFzkpUblpEvb76MDidG1ymfEhYSxsObj9izfB+ze3zP4FX9sXOwpWHn+iwduoLNc7bRsFN9Qp+EsGrcGpSZQpkl/9rqOvVVR0KpylChYTnGt5nElu+2M+Qnw57vz6f14HnIcx7cfMiOJb8z5bNvGbtmBPaOdmiaRlhIGL2m9qbqO/reaY+a7jy89YhNv26iUX+vt36mvzCezDBEYTLQELgGJNklppSqARxE3xAeA4wFSgGHlFKer9W1AQ4DbYH5QD8gGFirlPrMwDnEsLa2Iigo/qf8wMCEe+Fis7K2TnDbx1E9jTY26dc4TEsesY0aPpoli5ayYPF8GjdtbMgQU8QUzoe1tVWCvc36nt3UN4zy5ctHs+bvcPLESUOEl2HMzc153/s9bt+6zd2799LtuNa2NgDUb1Q/TnmDxg14cP8h9+7eT3Yf40dNYMWSlcxdMIeGTVI/PjGtbgbc4tupMxk1bgQvXrwkKOgJQUFPAHjx4gVBQU+IiIhIcFsrayuCAp/EK49+jVrbpH9jPTGTRk/lp6U/M3P+dOo3Tv5qRX4rfc/tk9fyC4rqlTbkB5HCJQpRrJwL1VpWof+SvrwIfcGuxbsBqN6qKi3+14y9K/YzuO5wxr07CesCVpT1cie/vWWqj5XLIheedcvify7lS6WllEPJIpQsX4JarWowYsUQXoSGsXWhfn5D3qgxuWVre8TZplwdD4ICgwh6kHiPuIVlHkKfhsYrD34SErXv5FePEW+3zNDALaFpmq2maU2A5GYgfQe8BOpqmjZL07RZQF30PbkzXqs7DH1DuKOmaWM0TVsINAL+Br5VSuU1aBZR3NzduHDeN16574WLlHFLenkad3c3/G74ERoa90198cJFcuTIEWeSkbGlJY9o30yexozpM5kx+1s+6vSRoUNMEVM4H27ubvheiD/W9qJvys9FQkyp5yQ9cynjlvQ4crNkethmTJ3JnG/nMnXmZD78uL0hQ0sxvxt+hIWF0aPL/3AuWCzmATB31vc4FyzG+XMXEtzWzb0M/n7+8d4Xl3wvkSNHDoqXKG70+FNizrS5/DBzPhO/HYf3R21TtE30WNtLvnHH2kaPvS1dxjhL0uWxzIN9UTseBDyKKWvzZStmHJ7KmE3DmfbHJLp/+xkPAh5SslIafu8Y+W1iYZmHgs4FY8bHOpYsknQ4SUwadCzlwINbD3nxPO6kxdtX75AtezYKOhtuyTPjyPg1cE3pd/ybyPAGrqZpKZqRoJQqCVQF1mmaFrOSdNT/1wGNlVKFYm3yEXBN07RtsepGAHMBG6CFAcKPp2Xrlhw/dpwb11+Nn/L38+fokaO0bJ30IVu0ak54eDgb12+MKdPpdKxft4HGTRqRM2fOJLY2rLTkAfDD3HmMGzOe8RPH8XnvXsnWNxZTOB8tWyWWgw8tW7dM9f6ePn3Kzt9+p0rVyoYMM93pdDo2rN2IU1EnChUqmG7HbdVG/7rZv2d/nPJ9u/dRxKEIBZOIZcEPC5k0bgqjxo+gx+fdjRpnUjzLe7J999Z4D4APP2rP9t1bKV4i4SWmmrVsRnh4OJs3bIkp0+l0bFy3iYaNG6Tr76nELJ63jG/Gf8uwcYPp2qtLirerXL0SNnY2bPx1c5zyDb9swtrGiqo1qySyZdo8ffSU+9fvY+8Ud6WNnHly4lDaAUs7S84dusC96/fjrW2bEs+Dn/PPwXO4eDobKuQEPXn0Nc5MCQAAGoBJREFUhDvX71KwqH4lnSpN9L9j/jl8Lk69s4f+xdbOFiv7+GO5o1VqWIGI8AiO7Xy1ckWELgKf347jWceD7DnStlybMH0ZPgY3FapGfT2awHM+QFf043h3KKUKAw7A6kTqRu9vraGD7Nr9M36ct4B2bdszdsJY/aL8Yyfg6ORI957dYur5+wfgUbosnTt3oqaXfsZphYoV8G7vzeABQwkP1+Hi4szCBYvxu+HHspWpn6GdUXms/XUdgwcMoek7TajXoB7HfI7H1Le0zIebe/wVATJDHo0bNWHk6BGMiLqRQGY5H59178KP8xfQvu2HjJkwBqUUE8dOxNHJkW6xbt4Q4B+Ah6snnTt3okad6gDMnjmHy5euUK9+XQoXKUSA/03mzJzD/Xv30/01FW3TBn3j4fQp/bD53b/vwc7eDjt7O7zq6v94W+ayoknTJmyoox8uv/aXdezYtoOmzZri6OTIg/sPWDh/IWdOn2H5T8vSNf4mzZrgVa8OA/oM4vF/j3F2cWbLxq0c2PsH3y/8DoCb/jep5F6VwSMG0biBfmjOhrWbGDFoFI2aNsSrvhd/HzsRs898lvmS7Rk2JCur/HjVS7ih5FTUKea5AP+bVHCrRKfOnahSuxIA5SuUo2279xk2aATh4eE4uzizZOFS/P0CWLRiodFj375Jf/n739P6RtP+3X9ga2eDrZ0tNb1qsHndVsYOGU+DJvWoXa8WJ4+fitn2nt9DytZ4ddWjlmddHJ0cWPvbGgCyZ8/OkNEDGf7VKAoXKai/0cPBI/yyci1fzxhPjhw50hz//C8XUtTNCQdXB3Jb5OK+/wP2rjyAWTZzmnTRr6Uc4HuTc4cuUNRNv+rD1VPX2LNsH027NqZExeIQNdLgvzuPGdVsHC17NafVF80B2L1sL/dvPMC1WinyF8jP4zuP2b18H08fPaXbNMNN8p31xVxcPJxxcnUkd97c3PO7x85luzE3N6NFV/1kvgr1y+FeowxLx6zgWWAwBZzsObbzb/49fJ7eA3rH7Ovh7UcMaDyU93u/S9s+bQBwcXemRotqrJq8hghdBPaO9uxdc4CHtx7yxYyeBstDmK6s1MCNvtYR/z6Ar8oc3qBuHEqpnkBPgIIFC3L0kE9C1ZI0adLXzPthPl06fYamaVSqVJHhw4fH/EIGuHfvHhEREYS9eBHnGN26doVIjZHDRhEcHEyJEiWYMnUyYSEv3iiWtHjTPH5e+TOaprF71x5274o7ial8+fLMmj0zU+YRGRmJv19Apjwf0Tl81rnrqxxGDOfcmVd3n4p9LnwOHwMgUhfJ38f+ZuP6jYSEhGBhkQePsmWZPWc25mSLqZdaWhpy6dQh7t2gvurbH9C/NmZGvTYiIiJ4+eIlPof1H46Cn4Zw9co19u4ZyrOnz8iVKxeurq5M/WYqTg5FY+qlOg8t8o22GzRoEEsWL2HimEkEPwvGqagTw0cOp0xJN04dORNzLm773yE0+Dmnjpxh3ep1aJrGvt372bc7bu9vufLlmDnr9VFWKWfI2dp3b97lxF/6RmF0Hi/CXsSUAXTv2h2lmTF2xPiY98XkKZPQhUbEqZcaEVrCY35f17PTF3G+H/7VKED/M/x25nQ2/bwVTdM4sOcgB/YcjFO3bNmyOLu86sV8HhLG06Bgzvm8GgJUqWwV+vbry/rVG5g3awH2Bezp3bc3VctXj1MvMTYvk7507lG0LEd+P8reZQfQ6XTY2tniWc6T9z98jwJmBSAAQh694OKezexZvI/w8HAcnRzp2acHDZo2gAAwf5kNm4AC6O7rl+nKFZQbmwD9cUtalOacry9n9/xLaEgoufPkxtXdlT69+1LKvmRM4zg5ZirpC7xlirpxZMdRflu8C51Oh52dHWXLedL2w/cpoBWAqGuzwwePYPWy1WyctYWQ4BCKOBbhq6H9qFerPpHX9ec8573nREZEYv5fdnJdfzUR8Mv/9ePnnGtYP2MzIcEhuBR3ZtTEUbjlLhuz/8zt7R4ikNGUpqXlT5VhKaXOAXkTWkVBKTUamAA00jRt/2vPNQT2Af01TZutlPIC/gQmapo25rW6ZkAEsEXTtPeSiqdylUraX8cOpyWlZB095BPT85mVmUIe6ZFDerzffA4fi+nBNRYtTU3clPE5fJwadYx717IXEWFG3T/AqSNnqFSrglGPYa6SXzoqLU78dSqmB9dYQnSJ39XKUM75XIzTg2sM2/yMvzazTUABHhc17jqwebMbdxJXruv5CCtu3HM+qu14rv97I0NamZ4Vy2qbDmb8gk2l8ruf1DTNOGNrMrms1IMbPaMhoQFeuV6rk5q6QgghhBCGo0xrIm9WlOGTzFIheoWFhIYWRJfdfoO6QgghhBDChGSlBm70VMqaCTxXA/2wwJMAmqbdRd+ATeh6c3TZiQSeE0IIIYQQWVyWaeBqmnYVfaO0XdQdyoCYu5W1A/ZrmhZ7tfc1QAmlVOtYdc2BvkAQ8Fu6BC6EEEKIt47KBP/eZhk+Blcp1RmIntZqD+RQSo2K+t5f07RVsar3Aw6gv3PZ3Kiyvugb6gNf2/VU9A3fn5VSM9H36HZEvzxYd03TjD+jQQghhBBCpLsMb+AC3YB6r5VNjPp6EIhp4GqadkQpVR/4OuqhAUeAdpqmnY29A03T/lNK1Ubf0O0N5AUuAB00TfvVCHkIIYQQQohMIMMbuJqm1U9l/aPob7mbkrq3gc7JVhRCCCGEMKC3fYhARssyY3CFEEIIIYRIiQzvwRVCCCGEMCUKJevgZjDpwRVCCCGEECZFGrhCCCGEEMKkyBAFIYQQQggDk0lmGUt6cIUQQgghhEmRHlwhhBBCCAOTHtyMJT24QgghhBDCpEgDVwghhBBCmBQZoiCEEEIIYWCyDm7Gkh5cIYQQQghhUqSBK4QQQgghTIoMURBCCCGEMDBZRSFjSQ+uEEIIIYQwKdKDK4QQQghhQAolk8wymPTgCiGEEEIIkyI9uEk4dfL0o9zZLPyNfBg74JGRj5EeTCEPU8gBJI/MxhTyMIUcQPLITNIjB2cj719kYtLATYKmafbGPoZS6oSmaVWMfRxjM4U8TCEHkDwyG1PIwxRyAMkjMzGFHJIjk8wylgxREEIIIYQQJkV6cIUQQgghDE56cDOS9OBmvIUZHYCBmEIeppADSB6ZjSnkYQo5gOSRmZhCDiITU5qmZXQMQgghhBAmo3ylctrOw9szOgwcLJxPmvpY58TIEAUhhBBCCAOTAQoZS4YoCCGEEEIIkyINXCGEEEIIYVKkgZvOlFKllVITlFI+SqmHSqlnSqkzSqmRSimLjI7vTSml8iilriulNKXU9xkdz+uUUsOVUutixeiXTP3qSqm9UefnqVLqd6VUhXQKN6m4UpyHUmp5VJ2EHt7pGPbrcaXqPaCUclVKbVZKBSqlQpRSh5RSDTMi9gTiWq2U8lVKPVFKhSqlLiqlZiqlCidSP9Pl8brk3suZOQ+llI1S6lul1FWlVFjU6+uAUsrrtXqZ8v0NoJTKq5QaoZT6Nyq+R0qpI0qpLuq1e79mhjyM+btVKVVEKbUy6jw+V0qdUEq1M0oiRqCUyvDH20zG4Ka/rkBvYCuwGggHGgBfA+2VUjU0TXuegfG9qQmA0W+MkQaTgcfAKcAqqYpKqRrAH8BtYExUcR/gkFKqlqZp/xoxzuSkOI9YOidQdtxgEaVeit8DSqkSwBFAB0wDngA9gF1Kqeaapu3NgPijOQKFgU3ALfQxegI9gQ5KqQqapj2ATJ/H6xJ9L2fmPJRSzujft3mBJcBlID9QDnCIVS/Tvr+VUmbATqAWsAKYC+QBOgLLADdgaFTdzJKHUX63KqVsgMNAAWAm+vfYR8BapVRXTdOWGTgPYWJkFYV0ppSqAlzRNO3Ja+VfAyOBvpqmZboe0KQopSqhbzANAWYAP2ia1idjo4pLKVVc07TrUf8/B+TVNM0lkbrHgTKAm6Zpt6PKHPh/e3ceLFlZ3nH8+xOBYVFWo7KOIopLBhGJTjAKxqjBEhKpMS4gamFlMURxQy0U1DIqVglZqlSwFIuIFRaLgKCA4pSKEAQhRkRBdESIQIgTlgHUyJM/zmnouvSduUvPPef2/X6orkOfpe/z3J6379Pvec974Drg8qp68cJEPTK22eRxKnBEVfXqa/xs2kCSM4BDgX2r6pp23dbAtcD9wF7Vsw+xtofpDOCYqjqhXbco8thQW+5zHkm+BSwH/qCqfrme/frcvlfSfIE4qaqOHlq/GfAjYPuq2rZd14s8NtZna5ITgHcCB1fVee26TYDLgD2A3avqno2W2Dzt/awVdeGlF3QdBo/fctclO4uCQxQWWFVdOfUPe+tf2+UzFjKe+Wo/cE4Bvgp8qeNwpjX4AN6QJE8C9gPOHHwAt8ffApwJvCjJ4zZOlBs20zyGpfHotneoczNtA2mGKxwMrB4UU+3x9wCfAZ5M8171zc/b5XawePLYUFvucx5Jng88Dzihqn6ZZNMkW47Yr9ftG3h0u/yv4ZVV9RvgDmAd9CuPjfjZ+hrgxkFx2+77O5pe7e2Bg8YQviZYL/7gCWhOdwLc1mkUs3c0zTfyXvXYzsPgD/RlI7ZdTjPzy74LF85Y3Nk+7ktycZLndB3QNKa2gRXA5kz/XkA/CsNlSXZMskuSFwOfbjcNum8WRR5suC33OY9BsXNTkvOA+4B1Sa5PctjQfn1v31cA/wu8K8mqJLsl2SvJR9q4jm/363seo8w45jRj2HfmoX9XU/cdfj1pJMfg9kDbc/I+mnFtp3cczowleQLwAeCDVbUmyfJuIxqLndrlLSO2DdbtPGJbH90KnAhcRdPzszfwVprxbgf1adznNG1gsbwXR9L0Kg2sAQ6rqm+1z3ufxwzbcp/zeEq7PAW4ATgC2Ax4O3Bakk3bMZt9zoGqWpvkYJoe8TOGNt0NHFpV57TPe53HNGYT82LM72F6NTZsCbLA7YeTgJXAe6vqx10HMwufAn5KcwHApBic1vz1iG33T9mn16rq3VNWnZPkdOAa4JPAngsf1bRGtYHF8l6cQzM+cmtgH5rT+DsObV8MecykLfc5j0e1y7uBA9tT+iQ5hyavv0/yefqdw8A9wA9oLsL8Ds3p+DcDpyc5pKouZnHkMdVsYl6M+alnLHA7luRDNKcET66qj3Qdz0y1p/3+BHh+Vf2263jG6N52ufmIbcum7LPoVNUN7YVCr0/y5Kq6vuuY1tMGFsV7UVU301zhDc2XiLOB7ybZss2n13nMoi33OY/BzDNfHBS38GCP6LnA62h6efucA0l+n6aoPbqqPjW0/os0Re8p7UwWvc5jGrOJeTHmp55xDG6HkhwPHEsz/ctfdRvNzCXZnKan5wLg1iRPai8g2L3dZZt23UynseqTwcUdo05/DdaNOm22mKxplzuub6eFsIE2sCjfi6r6PnA18Dftqt7mMcu23Ns8eOgLxq0jtg1mVNiOfucAzTjoZTQXXT2oqu4Fzqd5X5bT/zxGmU3MizG/KdKTx9JlgduR9g/7cTRzHR7ZhymCZmELmnkyX0Yz3m3wWN1uP6x9fmQXwc3Td9vlyhHbngsUzZjWxWwwNKHTCxpn0Ab+k+YU5XTvBcCVGy3A+dmC5tQy9DuP2bTlPucxmNd5lxHbButup//te1C8bTJi2yOHln3PY5QZx9xO83YLD/27mrov9LftqycscDuQ5P00f9hPA95YVQ90HNJsrQNWjXgMeqy+2j4/t5Po5qGqfkLzwbkqyeBCB9r/XwVcUlWjeol6JclWSZaNWL8PTR7XVdWNCx/Zg3FssA2000+dBxyQZO+hY7emKbhuoMMbVkw3DVOSA2mmOrscep/HjNtyz/M4h2b87WFtPIPYHg/8GXB9Vf1kEbTvH7bL1w+vbHvQDwHWAoshj4eZQ8xfBPZI8vKhfTcBjqKZaaL7SWbXI6Hzu5jFO5lpISV5M83VyjcBXwNeM+Uf4W3tRQS91Y7TO2vq+qErr2+sqodt71KSw3notOtjgM2SHNs+/3lVnTa0+1uAb9DMNjC4Ov4omi+Eb1+IeKczizz2BL7SXmRzAw/NovBG4Hc0d9vqxCzbwHuAPwYuSnIicBfNnbN2Bl7W8ZmPT7YF1CU0c98uo5nm6FU0xdbwv5Ve5jGHttzXPNYmeQfNFG2XJ/kszSwKf90ujxravbftm+Ziy9cBH23H415KcybgTTR3zXtzOxcs9CSPjfjZ+lGawvf0JJ+g6dF9Nc30YEdW1d1jT0aTpap8LOADOJXmVMx0j9VdxziP3Ja3Ofxz17GMiG31bH7nNKfRvk5zRfPdwIXAsxZLHsDjaHpHf0RThPyWpqD8PM3dprrMYVZtgOb2pP9G02tzL83tO1/Ug/filcCXgV/QXNl9X/v7/idgtxH79zKPaXKbti33OQ/gFTQ95+vadnsRsP+I/XrZvtvY9mjb6c1tu70L+Cbwij7msTE/W2m+OJ1Gc5OL+2luB/wXXb9HM3ns/awVddt9t3T+AK7s+nfR1cNb9UqSJI3RM/fduy669Ctdh8Fjt9jZW/VKkiRJk8AxuJIkSWOWJT5NV9fswZUkSdJEscCVJEnSRHGIgiRJ0pg5RKFb9uBKkiRpoljgSpIkaaJY4Epa8pKsSbJmDK9TSVbPP6L1/oyxxCpJk8wCV9KctMXcbB6v7zpmSdLS4EVmkubqAyPWvRXYBvgHmlu5Drtmo0ckST2ReJFZlyxwJc1JVR0/dV3bS7sNcFJVrVngkCRJAhyiIGkBJFndDlPYLMn7k/w4ya+TnNpuP77dfsCIY5e3204dsW3LJO9Jck2SdUnuSXJZklePIeZtkrwzySVJbk7ymyT/neTcJCs3cOxOSU5LcnuS+5JcleQ169n/JUkuSHJH+3u5McnHk2w73zwkaSmyB1fSQjob2A/4CnAOcPtcX6gt/i4B9gG+B3yW5kv7S4DTkzy9qo6dR6xPBT4MfBM4H1gL7AYcDPxpkpdX1VdHHLcd8B2aIRqfA7YFXgl8IcnOVfXxKXkcBxwP/Ar4Ms3vZAXwDuCgJCur6q555CFJS44FrqSFtDvwjKq6YwyvdRJNcXtMVZ0wWJlkGU3x/N4kZ1XVXMf+XgfsNDXWJLsAVwAnAqMK3BXAmcCrquqB9piPAlcBH05ydlX9tF1/IE1xexlwUFU9OG65He7xOZqxzkfPMQdJWpIcoiBpIb1vHMVtkh2Aw4Arh4tbgKq6HzgGCDDtsIANqao7R8VaVTcDZwF7JdltxKG/oym6Hxg65mfAPwKbAocP7ft37fJNw8Vte8ypNBfmvXauOUjSUmUPrqSFdMWYXmc/YBOgkhw/Yvum7fKp8/khSfYH3gKsBH4P2GzKLjsDN01Zd1Nb0E61GjiOptd5YCXwW2BVklUjjtkMeEySHarqf2afgaRuxFv1dswCV9JCunVMr7NDu9yvfUxn67n+gCR/TtNTez9wMXAjsA54ADgAeAGw+YhDb5vmJQe5bzO0bgeaz+HjNhDO1oAFriTNkAWupAVTVTXNpsHp/FGfSaNmErizXZ5YVW+bd2CjfQj4DfDsqrpueEOST9MUuKM8dpr1j2uXdw6tuxN4RFVtP59AJfWRPbhdcgyupD5Y2y53HbHt2SPWXUFTFP/RRosIngT8cERx+wjgees5brcky0esP6BdXj207nJguyRPn3uYkqSpLHAl9cFgbO4bkjzYi5tkV+D9U3euqtuBLwDPTvK+JJtM3SfJHkmeMI+Y1gB7Jtlp6DVDM+vB09Zz3CbAx9pCeHDcE2guKPs/4F+G9j2xXZ4y/HOGjtsqyXPnmoAkLVUOUZDUuar69yTfBJ4PXJHkEppT/S8HLmR0z+7fAnsCHwQOT/JtmvGvO9FcXLYf8Gpg1AVfM3Ei8Cng6iRn01wMtj9NcXteG9so3weeA1yV5CIemgd3W+BdVXXjUN5fT/Ju4CPADUkuaOPdmmZKtRcA3wZeOsccJHUgOECha/bgSuqLQ4DPALsAR9HMNvAumim/Hqa9+cEL2n3vAA4F3gYcCNxNM3fsxXMNpqo+DbwB+CVwBM10Xb+gKV6/t55D1wJ/CFzbHn8ETdH62qk3eWh/zsdoCvvzaQrotwKraGZoOBmYz80qJGlJyvTXfEiSJGm29tn3mfWNy77WdRhst/ljrqqqUdcxTDx7cCVJkjRRLHAlSZI0UbzITJIkaey8zKxL9uBKkiRpoljgSpIkaaI4REGSJGnMHKDQLXtwJUmSNFHswZUkSRo7+3C7ZA+uJEmSJooFriRJkiaKQxQkSZLGKiQOUeiSPbiSJEmaKBa4kiRJmigWuJIkSZooFriSJEmaKF5kJkmSNEYB4jy4nbIHV5IkSRPFAleSJEkTxSEKkiRJY+cQhS7ZgytJkqSJYg+uJEnSmNl/2y17cCVJkjRRLHAlSZI0URyiIEmSNGaJgxS6ZA+uJEmSJooFriRJkiaKQxQkSZLGKjiPQrfswZUkSdJEsQdXkiRpzOy/7ZY9uJIkSZooFriSJEmaKA5RkCRJGjsHKXTJHlxJkiRNFHtwJUmSxineyaxr9uBKkiRpoljgSpIkaaJY4EqSJC1RSXZNclaSO5PcleRLSXbrOq75ssCVJElagpJsCVwC7AUcARwO7Al8I8lWXcY2X15kJkmStDS9CXgi8JSq+glAku8DNwB/CXyiw9jmxR5cSZKkMQqQHvw3AwcDlw+KW4Cq+hlwKXDIxvntLAwLXEmSpKXp6cAPRqy/FnjaAscyVg5RkCRJGqPvXXX1hVs8cqsdu44DWJbkyqHnJ1fVyUPPtwfWjjjuV8B2GzWyjcwCV5IkaYyq6qVdx7DUOURBkiRpaVrL6J7a6Xp2Fw0LXEmSpKXpWppxuFM9DfjhAscyVha4kiRJS9O5wHOTPHGwIslyYP9226KVquo6BkmSJC2w9mYO/wHcBxwLFPAh4FHAiqq6p8Pw5sUeXEmSpCWoqtYBLwSuB04DvgD8DHjhYi5uwR5cSZIkTRh7cCVJkjRRLHAlSZI0USxwJUmSNFEscCVJkjRRLHAlSZI0USxwJUmSNFEscCVJkjRRLHAlSZI0Uf4fAX6Ys/V9WpwAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 720x720 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}