Model_Training.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Training.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "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/mertbkts/Covid19-Prediction-From-LungCT/blob/main/Model_Training.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Preparation For Training"
      ],
      "metadata": {
        "id": "wHLbRPzFh9Uj"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        " Importing Dependencies"
      ],
      "metadata": {
        "id": "EvjbH5DEY8Cu"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "id": "K-x-uDU7bED5",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3ca1daac-e811-495c-ebff-92976309087e"
      },
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sn6kN0igjHUu"
      },
      "source": [
        "from tensorflow.keras.applications import VGG19\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "from tensorflow.keras.layers import Dropout\n",
        "from tensorflow.keras.layers import Flatten\n",
        "from tensorflow.keras.layers import Dense\n",
        "from tensorflow.keras.layers import Input\n",
        "from tensorflow.keras.models import Model\n",
        "from tensorflow.keras.optimizers import Adam\n",
        "from tensorflow.keras.utils import to_categorical\n",
        "from sklearn.preprocessing import LabelBinarizer\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 imutils import paths\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import cv2\n",
        "import os"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        " Importing Dataset"
      ],
      "metadata": {
        "id": "sxBm19PgY5cF"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YxVYkRSvsjH8"
      },
      "source": [
        "dataset= # Your dataset's path"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        " Define Various Elements"
      ],
      "metadata": {
        "id": "EFbbxcAJxSc0"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QaAWnXl9p5zV"
      },
      "source": [
        "# Configuration of Epoch and Batch Size for training\n",
        "EpochNumber = 100\n",
        "batchSize = 32"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Oj8h0faRp6_c"
      },
      "source": [
        "# Create a list from the paths of the images in the dataset\n",
        "imagePaths = list(paths.list_images(dataset))\n",
        "\n",
        "# Create separate lists for datas and labels\n",
        "data = []\n",
        "labels = []"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Fetch Images and Labels of Them from The Filename They Belong(Either COVID or non-COVID)"
      ],
      "metadata": {
        "id": "vkfsrkqNuz7n"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dtGrTf9Hp8P3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5c256985-75ce-40ca-acc5-72cb04195db0"
      },
      "source": [
        "print(\"Images are loading...\")\n",
        "\n",
        "for imagePath in imagePaths:\n",
        "\t\n",
        "  # Fetch label of the current image\n",
        "\tlabel = imagePath.split(os.path.sep)[-2]\n",
        "\n",
        "\t# Read image\n",
        "\timage = cv2.imread(imagePath)\n",
        "  # Change color channel to RGB\n",
        "\timage = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)\n",
        "  # Resize image to 224x224 pixels\n",
        "\timage = cv2.resize(image, (224, 224))\n",
        "  \n",
        "\t# Append image and the label of it\n",
        "\tdata.append(image)\n",
        "\tlabels.append(label)"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Images are loading...\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Normalization"
      ],
      "metadata": {
        "id": "lUjbQJxUx9TP"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GFS21guyp9p2"
      },
      "source": [
        "# Convert data to a NumPy array and normalize data between 0 and 1\n",
        "data = np.array(data) / 255.0\n",
        "labels = np.array(labels)"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Train-Test Split"
      ],
      "metadata": {
        "id": "xySWkVxszpoq"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GZgdMbjvqAXx"
      },
      "source": [
        "# Splitting data into training and test sets. 80% of the data will be the training set and 20% will be the test set.\n",
        "(trainX, testX, trainY, testY) = train_test_split(data, labels,test_size=0.20)"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "One-Hot Encoding"
      ],
      "metadata": {
        "id": "QD-0cS72rr2V"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Categorize labels into zeros and ones\n",
        "lb = LabelBinarizer()\n",
        "\n",
        "trainY = lb.fit_transform(trainY)\n",
        "trainY = to_categorical(trainY)\n",
        "\n",
        "testY = lb.fit_transform(testY)\n",
        "testY = to_categorical(testY)"
      ],
      "metadata": {
        "id": "z3-H-swixw3j"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Building The Model And Training"
      ],
      "metadata": {
        "id": "q_2ieow2iIgi"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Building and Compiling The Model"
      ],
      "metadata": {
        "id": "-FFgHxzxz8zW"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Building the base of the model using VGG19.\n",
        "vggModel = VGG19(weights=\"imagenet\", include_top=False,\n",
        "                 input_tensor=Input(shape=(224, 224, 3)))\n",
        "\n",
        "# Adding layers of Pooling, Flattening, Dropout and Fully Connected\n",
        "output = vggModel.output\n",
        "output = Flatten()(output)\n",
        "output = Dropout(0.5)(output)\n",
        "output = Dense(2, activation='softmax')(output)\n",
        "\n",
        "model = Model(inputs=vggModel.input, outputs=output)\n",
        "\n",
        "# Freezing layers to prevent modification of it's weights\n",
        "for layer in vggModel.layers:\n",
        "    layer.trainable = False\n",
        "\n",
        "# Compiling the model\n",
        "model.compile(\n",
        "        loss='categorical_crossentropy', \n",
        "        optimizer='adam', \n",
        "        metrics=['accuracy']\n",
        ")\n",
        "\n",
        "# Increasing the size of data with techniques such as rotating, shifting, and flipping\n",
        "train_aug = ImageDataGenerator(\n",
        "    rotation_range=20,\n",
        "    width_shift_range=0.2,\n",
        "    height_shift_range=0.2,\n",
        "    horizontal_flip=True\n",
        ")"
      ],
      "metadata": {
        "id": "7CAM6P5v7WX0",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c631310f-add5-4340-f75f-193919e7d848"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/vgg19/vgg19_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "80142336/80134624 [==============================] - 1s 0us/step\n",
            "80150528/80134624 [==============================] - 1s 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u4MG50zkxwkd",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0f350a00-2b6b-4c15-ffc0-a0445b24811e"
      },
      "source": [
        "# Summarizing the model\n",
        "model.summary()"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_1 (InputLayer)        [(None, 224, 224, 3)]     0         \n",
            "                                                                 \n",
            " block1_conv1 (Conv2D)       (None, 224, 224, 64)      1792      \n",
            "                                                                 \n",
            " block1_conv2 (Conv2D)       (None, 224, 224, 64)      36928     \n",
            "                                                                 \n",
            " block1_pool (MaxPooling2D)  (None, 112, 112, 64)      0         \n",
            "                                                                 \n",
            " block2_conv1 (Conv2D)       (None, 112, 112, 128)     73856     \n",
            "                                                                 \n",
            " block2_conv2 (Conv2D)       (None, 112, 112, 128)     147584    \n",
            "                                                                 \n",
            " block2_pool (MaxPooling2D)  (None, 56, 56, 128)       0         \n",
            "                                                                 \n",
            " block3_conv1 (Conv2D)       (None, 56, 56, 256)       295168    \n",
            "                                                                 \n",
            " block3_conv2 (Conv2D)       (None, 56, 56, 256)       590080    \n",
            "                                                                 \n",
            " block3_conv3 (Conv2D)       (None, 56, 56, 256)       590080    \n",
            "                                                                 \n",
            " block3_conv4 (Conv2D)       (None, 56, 56, 256)       590080    \n",
            "                                                                 \n",
            " block3_pool (MaxPooling2D)  (None, 28, 28, 256)       0         \n",
            "                                                                 \n",
            " block4_conv1 (Conv2D)       (None, 28, 28, 512)       1180160   \n",
            "                                                                 \n",
            " block4_conv2 (Conv2D)       (None, 28, 28, 512)       2359808   \n",
            "                                                                 \n",
            " block4_conv3 (Conv2D)       (None, 28, 28, 512)       2359808   \n",
            "                                                                 \n",
            " block4_conv4 (Conv2D)       (None, 28, 28, 512)       2359808   \n",
            "                                                                 \n",
            " block4_pool (MaxPooling2D)  (None, 14, 14, 512)       0         \n",
            "                                                                 \n",
            " block5_conv1 (Conv2D)       (None, 14, 14, 512)       2359808   \n",
            "                                                                 \n",
            " block5_conv2 (Conv2D)       (None, 14, 14, 512)       2359808   \n",
            "                                                                 \n",
            " block5_conv3 (Conv2D)       (None, 14, 14, 512)       2359808   \n",
            "                                                                 \n",
            " block5_conv4 (Conv2D)       (None, 14, 14, 512)       2359808   \n",
            "                                                                 \n",
            " block5_pool (MaxPooling2D)  (None, 7, 7, 512)         0         \n",
            "                                                                 \n",
            " flatten (Flatten)           (None, 25088)             0         \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 25088)             0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 2)                 50178     \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 20,074,562\n",
            "Trainable params: 50,178\n",
            "Non-trainable params: 20,024,384\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Training"
      ],
      "metadata": {
        "id": "NLUwVuDp7ovh"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qtNd-2XJx3N0",
        "outputId": "360a7b38-e234-4741-f77d-f48bf7b737d2"
      },
      "source": [
        "H = model.fit(train_aug.flow(trainX, trainY, batch_size=batchSize),\n",
        "                    validation_data=(testX, testY),\n",
        "                    validation_steps=len(testX) / batchSize,\n",
        "                    steps_per_epoch=len(trainX) / batchSize,\n",
        "                    epochs=EpochNumber)"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "62/62 [==============================] - 52s 614ms/step - loss: 0.9398 - accuracy: 0.6396 - val_loss: 0.7773 - val_accuracy: 0.6680\n",
            "Epoch 2/100\n",
            "62/62 [==============================] - 35s 564ms/step - loss: 0.5838 - accuracy: 0.7485 - val_loss: 0.3752 - val_accuracy: 0.8249\n",
            "Epoch 3/100\n",
            "62/62 [==============================] - 35s 560ms/step - loss: 0.5065 - accuracy: 0.7742 - val_loss: 0.4900 - val_accuracy: 0.7525\n",
            "Epoch 4/100\n",
            "62/62 [==============================] - 36s 574ms/step - loss: 0.4285 - accuracy: 0.8085 - val_loss: 0.3355 - val_accuracy: 0.8491\n",
            "Epoch 5/100\n",
            "62/62 [==============================] - 37s 591ms/step - loss: 0.4217 - accuracy: 0.8024 - val_loss: 0.3135 - val_accuracy: 0.8612\n",
            "Epoch 6/100\n",
            "62/62 [==============================] - 36s 581ms/step - loss: 0.4476 - accuracy: 0.8044 - val_loss: 0.3366 - val_accuracy: 0.8531\n",
            "Epoch 7/100\n",
            "62/62 [==============================] - 36s 572ms/step - loss: 0.4551 - accuracy: 0.8160 - val_loss: 0.2886 - val_accuracy: 0.8793\n",
            "Epoch 8/100\n",
            "62/62 [==============================] - 36s 576ms/step - loss: 0.4308 - accuracy: 0.8206 - val_loss: 0.2952 - val_accuracy: 0.8612\n",
            "Epoch 9/100\n",
            "62/62 [==============================] - 35s 564ms/step - loss: 0.4114 - accuracy: 0.8256 - val_loss: 0.2739 - val_accuracy: 0.8913\n",
            "Epoch 10/100\n",
            "62/62 [==============================] - 35s 560ms/step - loss: 0.4345 - accuracy: 0.8196 - val_loss: 0.2601 - val_accuracy: 0.8974\n",
            "Epoch 11/100\n",
            "62/62 [==============================] - 35s 559ms/step - loss: 0.4010 - accuracy: 0.8392 - val_loss: 0.2485 - val_accuracy: 0.9034\n",
            "Epoch 12/100\n",
            "62/62 [==============================] - 35s 557ms/step - loss: 0.4558 - accuracy: 0.8317 - val_loss: 0.2680 - val_accuracy: 0.8773\n",
            "Epoch 13/100\n",
            "62/62 [==============================] - 35s 559ms/step - loss: 0.4194 - accuracy: 0.8246 - val_loss: 0.6239 - val_accuracy: 0.7565\n",
            "Epoch 14/100\n",
            "62/62 [==============================] - 35s 562ms/step - loss: 0.4473 - accuracy: 0.8206 - val_loss: 0.2479 - val_accuracy: 0.8934\n",
            "Epoch 15/100\n",
            "62/62 [==============================] - 35s 561ms/step - loss: 0.4033 - accuracy: 0.8397 - val_loss: 0.3073 - val_accuracy: 0.8652\n",
            "Epoch 16/100\n",
            "62/62 [==============================] - 35s 563ms/step - loss: 0.4074 - accuracy: 0.8342 - val_loss: 0.2658 - val_accuracy: 0.8954\n",
            "Epoch 17/100\n",
            "62/62 [==============================] - 35s 566ms/step - loss: 0.4209 - accuracy: 0.8377 - val_loss: 0.2900 - val_accuracy: 0.8773\n",
            "Epoch 18/100\n",
            "62/62 [==============================] - 35s 569ms/step - loss: 0.4312 - accuracy: 0.8327 - val_loss: 0.2534 - val_accuracy: 0.8974\n",
            "Epoch 19/100\n",
            "62/62 [==============================] - 36s 571ms/step - loss: 0.3813 - accuracy: 0.8543 - val_loss: 0.2828 - val_accuracy: 0.8793\n",
            "Epoch 20/100\n",
            "62/62 [==============================] - 36s 571ms/step - loss: 0.3741 - accuracy: 0.8518 - val_loss: 0.2570 - val_accuracy: 0.8913\n",
            "Epoch 21/100\n",
            "62/62 [==============================] - 36s 574ms/step - loss: 0.4017 - accuracy: 0.8412 - val_loss: 0.2575 - val_accuracy: 0.8873\n",
            "Epoch 22/100\n",
            "62/62 [==============================] - 35s 568ms/step - loss: 0.4420 - accuracy: 0.8311 - val_loss: 0.3517 - val_accuracy: 0.8571\n",
            "Epoch 23/100\n",
            "62/62 [==============================] - 35s 563ms/step - loss: 0.4861 - accuracy: 0.8322 - val_loss: 0.2166 - val_accuracy: 0.9135\n",
            "Epoch 24/100\n",
            "62/62 [==============================] - 35s 557ms/step - loss: 0.4194 - accuracy: 0.8417 - val_loss: 0.2374 - val_accuracy: 0.9074\n",
            "Epoch 25/100\n",
            "62/62 [==============================] - 35s 559ms/step - loss: 0.4127 - accuracy: 0.8427 - val_loss: 0.2261 - val_accuracy: 0.9034\n",
            "Epoch 26/100\n",
            "62/62 [==============================] - 35s 559ms/step - loss: 0.4009 - accuracy: 0.8558 - val_loss: 0.2302 - val_accuracy: 0.9014\n",
            "Epoch 27/100\n",
            "62/62 [==============================] - 35s 564ms/step - loss: 0.4152 - accuracy: 0.8382 - val_loss: 0.2500 - val_accuracy: 0.8934\n",
            "Epoch 28/100\n",
            "62/62 [==============================] - 35s 561ms/step - loss: 0.4159 - accuracy: 0.8523 - val_loss: 0.2208 - val_accuracy: 0.8994\n",
            "Epoch 29/100\n",
            "62/62 [==============================] - 35s 557ms/step - loss: 0.4490 - accuracy: 0.8382 - val_loss: 0.2157 - val_accuracy: 0.9074\n",
            "Epoch 30/100\n",
            "62/62 [==============================] - 35s 563ms/step - loss: 0.4455 - accuracy: 0.8281 - val_loss: 0.5191 - val_accuracy: 0.7968\n",
            "Epoch 31/100\n",
            "62/62 [==============================] - 35s 556ms/step - loss: 0.4396 - accuracy: 0.8448 - val_loss: 0.2135 - val_accuracy: 0.9115\n",
            "Epoch 32/100\n",
            "62/62 [==============================] - 38s 611ms/step - loss: 0.4530 - accuracy: 0.8382 - val_loss: 0.2352 - val_accuracy: 0.9054\n",
            "Epoch 33/100\n",
            "62/62 [==============================] - 35s 556ms/step - loss: 0.4154 - accuracy: 0.8488 - val_loss: 0.2113 - val_accuracy: 0.9115\n",
            "Epoch 34/100\n",
            "62/62 [==============================] - 35s 558ms/step - loss: 0.4064 - accuracy: 0.8483 - val_loss: 0.2134 - val_accuracy: 0.9195\n",
            "Epoch 35/100\n",
            "62/62 [==============================] - 39s 632ms/step - loss: 0.3995 - accuracy: 0.8614 - val_loss: 0.2608 - val_accuracy: 0.8833\n",
            "Epoch 36/100\n",
            "62/62 [==============================] - 35s 569ms/step - loss: 0.3815 - accuracy: 0.8543 - val_loss: 0.3684 - val_accuracy: 0.8592\n",
            "Epoch 37/100\n",
            "62/62 [==============================] - 35s 570ms/step - loss: 0.3622 - accuracy: 0.8604 - val_loss: 0.2063 - val_accuracy: 0.9115\n",
            "Epoch 38/100\n",
            "62/62 [==============================] - 34s 553ms/step - loss: 0.4769 - accuracy: 0.8397 - val_loss: 0.2403 - val_accuracy: 0.9034\n",
            "Epoch 39/100\n",
            "62/62 [==============================] - 35s 567ms/step - loss: 0.4420 - accuracy: 0.8548 - val_loss: 0.2068 - val_accuracy: 0.9135\n",
            "Epoch 40/100\n",
            "62/62 [==============================] - 35s 560ms/step - loss: 0.4443 - accuracy: 0.8483 - val_loss: 0.1865 - val_accuracy: 0.9235\n",
            "Epoch 41/100\n",
            "62/62 [==============================] - 35s 559ms/step - loss: 0.4588 - accuracy: 0.8337 - val_loss: 0.2120 - val_accuracy: 0.9074\n",
            "Epoch 42/100\n",
            "62/62 [==============================] - 35s 557ms/step - loss: 0.4222 - accuracy: 0.8478 - val_loss: 0.2082 - val_accuracy: 0.9215\n",
            "Epoch 43/100\n",
            "62/62 [==============================] - 35s 558ms/step - loss: 0.4016 - accuracy: 0.8599 - val_loss: 0.2958 - val_accuracy: 0.8773\n",
            "Epoch 44/100\n",
            "62/62 [==============================] - 35s 565ms/step - loss: 0.4000 - accuracy: 0.8569 - val_loss: 0.2211 - val_accuracy: 0.9135\n",
            "Epoch 45/100\n",
            "62/62 [==============================] - 35s 566ms/step - loss: 0.4150 - accuracy: 0.8498 - val_loss: 0.2210 - val_accuracy: 0.9115\n",
            "Epoch 46/100\n",
            "62/62 [==============================] - 34s 551ms/step - loss: 0.4024 - accuracy: 0.8619 - val_loss: 0.2083 - val_accuracy: 0.9276\n",
            "Epoch 47/100\n",
            "62/62 [==============================] - 35s 561ms/step - loss: 0.3754 - accuracy: 0.8649 - val_loss: 0.1956 - val_accuracy: 0.9215\n",
            "Epoch 48/100\n",
            "62/62 [==============================] - 35s 559ms/step - loss: 0.4203 - accuracy: 0.8528 - val_loss: 0.2330 - val_accuracy: 0.9014\n",
            "Epoch 49/100\n",
            "62/62 [==============================] - 35s 558ms/step - loss: 0.4346 - accuracy: 0.8493 - val_loss: 0.3760 - val_accuracy: 0.8632\n",
            "Epoch 50/100\n",
            "62/62 [==============================] - 35s 557ms/step - loss: 0.4328 - accuracy: 0.8599 - val_loss: 0.2226 - val_accuracy: 0.9195\n",
            "Epoch 51/100\n",
            "62/62 [==============================] - 35s 557ms/step - loss: 0.4318 - accuracy: 0.8473 - val_loss: 0.2159 - val_accuracy: 0.9195\n",
            "Epoch 52/100\n",
            "62/62 [==============================] - 35s 559ms/step - loss: 0.3863 - accuracy: 0.8579 - val_loss: 0.2228 - val_accuracy: 0.9135\n",
            "Epoch 53/100\n",
            "62/62 [==============================] - 35s 557ms/step - loss: 0.4281 - accuracy: 0.8543 - val_loss: 0.1995 - val_accuracy: 0.9195\n",
            "Epoch 54/100\n",
            "62/62 [==============================] - 34s 553ms/step - loss: 0.3858 - accuracy: 0.8548 - val_loss: 0.2622 - val_accuracy: 0.9034\n",
            "Epoch 55/100\n",
            "62/62 [==============================] - 34s 549ms/step - loss: 0.4802 - accuracy: 0.8322 - val_loss: 0.2381 - val_accuracy: 0.9175\n",
            "Epoch 56/100\n",
            "62/62 [==============================] - 34s 551ms/step - loss: 0.4397 - accuracy: 0.8508 - val_loss: 0.2417 - val_accuracy: 0.9054\n",
            "Epoch 57/100\n",
            "62/62 [==============================] - 34s 551ms/step - loss: 0.4045 - accuracy: 0.8579 - val_loss: 0.2118 - val_accuracy: 0.9054\n",
            "Epoch 58/100\n",
            "62/62 [==============================] - 34s 549ms/step - loss: 0.3831 - accuracy: 0.8548 - val_loss: 0.2630 - val_accuracy: 0.8954\n",
            "Epoch 59/100\n",
            "62/62 [==============================] - 34s 552ms/step - loss: 0.3941 - accuracy: 0.8624 - val_loss: 0.3347 - val_accuracy: 0.8793\n",
            "Epoch 60/100\n",
            "62/62 [==============================] - 34s 549ms/step - loss: 0.4361 - accuracy: 0.8533 - val_loss: 0.2989 - val_accuracy: 0.8974\n",
            "Epoch 61/100\n",
            "62/62 [==============================] - 34s 549ms/step - loss: 0.3931 - accuracy: 0.8664 - val_loss: 0.2238 - val_accuracy: 0.9175\n",
            "Epoch 62/100\n",
            "62/62 [==============================] - 34s 550ms/step - loss: 0.4346 - accuracy: 0.8548 - val_loss: 0.3272 - val_accuracy: 0.8773\n",
            "Epoch 63/100\n",
            "62/62 [==============================] - 34s 550ms/step - loss: 0.5183 - accuracy: 0.8397 - val_loss: 0.3175 - val_accuracy: 0.8793\n",
            "Epoch 64/100\n",
            "62/62 [==============================] - 34s 550ms/step - loss: 0.4562 - accuracy: 0.8488 - val_loss: 0.2486 - val_accuracy: 0.9074\n",
            "Epoch 65/100\n",
            "62/62 [==============================] - 34s 549ms/step - loss: 0.4363 - accuracy: 0.8523 - val_loss: 0.3062 - val_accuracy: 0.8893\n",
            "Epoch 66/100\n",
            "62/62 [==============================] - 34s 550ms/step - loss: 0.4552 - accuracy: 0.8548 - val_loss: 0.2224 - val_accuracy: 0.9034\n",
            "Epoch 67/100\n",
            "62/62 [==============================] - 34s 549ms/step - loss: 0.4355 - accuracy: 0.8538 - val_loss: 0.2246 - val_accuracy: 0.9074\n",
            "Epoch 68/100\n",
            "62/62 [==============================] - 34s 548ms/step - loss: 0.4468 - accuracy: 0.8498 - val_loss: 0.2091 - val_accuracy: 0.9296\n",
            "Epoch 69/100\n",
            "62/62 [==============================] - 34s 548ms/step - loss: 0.4634 - accuracy: 0.8478 - val_loss: 0.3126 - val_accuracy: 0.8712\n",
            "Epoch 70/100\n",
            "62/62 [==============================] - 34s 553ms/step - loss: 0.4199 - accuracy: 0.8639 - val_loss: 0.3763 - val_accuracy: 0.8652\n",
            "Epoch 71/100\n",
            "62/62 [==============================] - 35s 556ms/step - loss: 0.4447 - accuracy: 0.8543 - val_loss: 0.3160 - val_accuracy: 0.8853\n",
            "Epoch 72/100\n",
            "62/62 [==============================] - 35s 558ms/step - loss: 0.5545 - accuracy: 0.8377 - val_loss: 0.2460 - val_accuracy: 0.9014\n",
            "Epoch 73/100\n",
            "62/62 [==============================] - 35s 559ms/step - loss: 0.4285 - accuracy: 0.8609 - val_loss: 0.2040 - val_accuracy: 0.9135\n",
            "Epoch 74/100\n",
            "62/62 [==============================] - 35s 557ms/step - loss: 0.4207 - accuracy: 0.8710 - val_loss: 0.2415 - val_accuracy: 0.8913\n",
            "Epoch 75/100\n",
            "62/62 [==============================] - 35s 559ms/step - loss: 0.4928 - accuracy: 0.8508 - val_loss: 0.3710 - val_accuracy: 0.8672\n",
            "Epoch 76/100\n",
            "62/62 [==============================] - 35s 563ms/step - loss: 0.4602 - accuracy: 0.8558 - val_loss: 0.2621 - val_accuracy: 0.9014\n",
            "Epoch 77/100\n",
            "62/62 [==============================] - 35s 565ms/step - loss: 0.4318 - accuracy: 0.8624 - val_loss: 0.1988 - val_accuracy: 0.9235\n",
            "Epoch 78/100\n",
            "62/62 [==============================] - 35s 556ms/step - loss: 0.4865 - accuracy: 0.8513 - val_loss: 0.3488 - val_accuracy: 0.8793\n",
            "Epoch 79/100\n",
            "62/62 [==============================] - 34s 553ms/step - loss: 0.4219 - accuracy: 0.8705 - val_loss: 0.1995 - val_accuracy: 0.9336\n",
            "Epoch 80/100\n",
            "62/62 [==============================] - 35s 557ms/step - loss: 0.4444 - accuracy: 0.8569 - val_loss: 0.1976 - val_accuracy: 0.9256\n",
            "Epoch 81/100\n",
            "62/62 [==============================] - 35s 557ms/step - loss: 0.4455 - accuracy: 0.8473 - val_loss: 0.2513 - val_accuracy: 0.8934\n",
            "Epoch 82/100\n",
            "62/62 [==============================] - 34s 551ms/step - loss: 0.4074 - accuracy: 0.8594 - val_loss: 0.2884 - val_accuracy: 0.8974\n",
            "Epoch 83/100\n",
            "62/62 [==============================] - 34s 553ms/step - loss: 0.5083 - accuracy: 0.8473 - val_loss: 0.4449 - val_accuracy: 0.8491\n",
            "Epoch 84/100\n",
            "62/62 [==============================] - 34s 552ms/step - loss: 0.4829 - accuracy: 0.8498 - val_loss: 0.2282 - val_accuracy: 0.9054\n",
            "Epoch 85/100\n",
            "62/62 [==============================] - 35s 555ms/step - loss: 0.3978 - accuracy: 0.8639 - val_loss: 0.2024 - val_accuracy: 0.9135\n",
            "Epoch 86/100\n",
            "62/62 [==============================] - 35s 555ms/step - loss: 0.4207 - accuracy: 0.8710 - val_loss: 0.3616 - val_accuracy: 0.8712\n",
            "Epoch 87/100\n",
            "62/62 [==============================] - 35s 557ms/step - loss: 0.4658 - accuracy: 0.8548 - val_loss: 0.1931 - val_accuracy: 0.9115\n",
            "Epoch 88/100\n",
            "62/62 [==============================] - 35s 566ms/step - loss: 0.4579 - accuracy: 0.8594 - val_loss: 0.3068 - val_accuracy: 0.8873\n",
            "Epoch 89/100\n",
            "62/62 [==============================] - 35s 567ms/step - loss: 0.4036 - accuracy: 0.8725 - val_loss: 0.2154 - val_accuracy: 0.9195\n",
            "Epoch 90/100\n",
            "62/62 [==============================] - 35s 566ms/step - loss: 0.4244 - accuracy: 0.8695 - val_loss: 0.4775 - val_accuracy: 0.8270\n",
            "Epoch 91/100\n",
            "62/62 [==============================] - 35s 564ms/step - loss: 0.5808 - accuracy: 0.8427 - val_loss: 0.2353 - val_accuracy: 0.9135\n",
            "Epoch 92/100\n",
            "62/62 [==============================] - 35s 563ms/step - loss: 0.4864 - accuracy: 0.8533 - val_loss: 0.2053 - val_accuracy: 0.9095\n",
            "Epoch 93/100\n",
            "62/62 [==============================] - 35s 564ms/step - loss: 0.4228 - accuracy: 0.8664 - val_loss: 0.2849 - val_accuracy: 0.9074\n",
            "Epoch 94/100\n",
            "62/62 [==============================] - 35s 563ms/step - loss: 0.4412 - accuracy: 0.8639 - val_loss: 0.2213 - val_accuracy: 0.9155\n",
            "Epoch 95/100\n",
            "62/62 [==============================] - 35s 558ms/step - loss: 0.3980 - accuracy: 0.8674 - val_loss: 0.2257 - val_accuracy: 0.9115\n",
            "Epoch 96/100\n",
            "62/62 [==============================] - 35s 558ms/step - loss: 0.4620 - accuracy: 0.8584 - val_loss: 0.1968 - val_accuracy: 0.9095\n",
            "Epoch 97/100\n",
            "62/62 [==============================] - 35s 556ms/step - loss: 0.4424 - accuracy: 0.8634 - val_loss: 0.2903 - val_accuracy: 0.8893\n",
            "Epoch 98/100\n",
            "62/62 [==============================] - 35s 558ms/step - loss: 0.4402 - accuracy: 0.8674 - val_loss: 0.2234 - val_accuracy: 0.9115\n",
            "Epoch 99/100\n",
            "62/62 [==============================] - 35s 555ms/step - loss: 0.4369 - accuracy: 0.8604 - val_loss: 0.1902 - val_accuracy: 0.9296\n",
            "Epoch 100/100\n",
            "62/62 [==============================] - 34s 553ms/step - loss: 0.4521 - accuracy: 0.8609 - val_loss: 0.3089 - val_accuracy: 0.8833\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Saving The Model"
      ],
      "metadata": {
        "id": "LUNBj4m6Cw7N"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XC2EhC6tCW6B"
      },
      "source": [
        "# Saving the model and it's weights\n",
        "model.save('model.h5')\n",
        "model.save_weights('weights.hdf5')"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Testing And Statistics"
      ],
      "metadata": {
        "id": "wpsGmaphiP-Q"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Prediction Test"
      ],
      "metadata": {
        "id": "hf9Y4p3aC2bB"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "630UsYI0qMBi"
      },
      "source": [
        "# Predicting on test set\n",
        "predTest = model.predict(testX, batch_size=batchSize)"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tTrPtvjPqOEk"
      },
      "source": [
        "# Converting results to binary classes for confusion matrix calculation\n",
        "predTest = np.argmax(predTest, axis=1)"
      ],
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Statistics"
      ],
      "metadata": {
        "id": "sJXEJvyVCraJ"
      }
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QPfgR4JbqPNU",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "61e6fb9d-07db-45f6-b76c-618efbff4385"
      },
      "source": [
        "# Creating classification report for performance evaluation\n",
        "print(classification_report(testY.argmax(axis=1), predTest,\n",
        "\ttarget_names=lb.classes_))\n"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "       covid       0.83      0.98      0.89       252\n",
            "      normal       0.97      0.79      0.87       245\n",
            "\n",
            "    accuracy                           0.88       497\n",
            "   macro avg       0.90      0.88      0.88       497\n",
            "weighted avg       0.90      0.88      0.88       497\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "I2Nyt-AUqQnG"
      },
      "source": [
        "# Calculating confusion matrix, accuracy, sensitivity, and specificity\n",
        "cm = confusion_matrix(testY.argmax(axis=1), predTest)\n",
        "total = sum(sum(cm))\n",
        "acc = (cm[0, 0] + cm[1, 1]) / total\n",
        "sensitivity = cm[0, 0] / (cm[0, 0] + cm[0, 1])\n",
        "specificity = cm[1, 1] / (cm[1, 0] + cm[1, 1])\n"
      ],
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LOWVHRSyqSfi",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "af979263-9c51-444a-f0c7-93bc1694bc63"
      },
      "source": [
        "# Printing calculations\n",
        "print(cm)\n",
        "print(\"acc: {:.4f}\".format(acc))\n",
        "print(\"sensitivity: {:.4f}\".format(sensitivity))\n",
        "print(\"specificity: {:.4f}\".format(specificity))\n"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[246   6]\n",
            " [ 52 193]]\n",
            "acc: 0.8833\n",
            "sensitivity: 0.9762\n",
            "specificity: 0.7878\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Visualizing model's training history\n",
        "N = EpochNumber\n",
        "\n",
        "plt.plot(np.arange(0, N), H.history[\"loss\"], label=\"train_loss\")\n",
        "plt.plot(np.arange(0, N), H.history[\"val_loss\"], label=\"val_loss\")\n",
        "plt.xlabel(\"Epoch\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.title(\"Model Loss\")\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.savefig(\"plotloss.png\")\n",
        "plt.show()\n",
        "\n",
        "\n",
        "\n",
        "plt.plot(np.arange(0, N), H.history[\"accuracy\"], label=\"train_acc\")\n",
        "plt.plot(np.arange(0, N), H.history[\"val_accuracy\"], label=\"val_acc\")\n",
        "plt.xlabel(\"Epoch\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.title(\"Model Accuracy\")\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.savefig(\"plotaccuracy.png\")\n",
        "plt.show()\n"
      ],
      "metadata": {
        "id": "GMohSMZXFIv0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        },
        "outputId": "b07946a8-0c8e-4f7e-d07f-0db44ca4b20e"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOydeXxcVd3/32eWZLLvSZO0Tdp0TXdoWVqWUlnKIos8IiguPALiig/KD1BERVH0UUR98FFARERBHgRBKatCKS2lLV2ge9M1aZo9aZZJZj2/P+69k9mSTJK52ea8X6++krlzM3OS3rmf8/l+v+d7hJQShUKhUCQultEegEKhUChGFyUECoVCkeAoIVAoFIoERwmBQqFQJDhKCBQKhSLBUUKgUCgUCY4SAoWiH4QQ5UIIKYSwxXDu54QQ74zEuBSKeKKEQDFhEEIcEUK4hRD5Yce36Tfz8tEZ2eAERaEYaZQQKCYah4HrjAdCiAVA6ugNR6EY+yghUEw0/gR8JujxZ4Engk8QQmQJIZ4QQjQKIY4KIe4WQlj056xCiJ8JIZqEEIeAS6P87O+FECeEEMeFED8UQliHM2AhRIkQ4kUhRIsQokoIcVPQc6cJIbYIIdqFEPVCiAf04w4hxJNCiGYhRJsQYrMQomg441AkLkoIFBONjUCmEGKufoO+Fngy7JxfA1nAdOBcNOG4QX/uJuAyYAmwFPiPsJ99HPACM/RzLgRuHOaYnwZqgBL9/X4khFilP/dL4JdSykygAnhGP/5Z/XeYAuQBtwDdwxyHIkFRQqCYiBiu4AJgD3DceCJIHO6SUnZIKY8APwc+rZ9yDfCglLJaStkC/DjoZ4uAS4CvSym7pJQNwC/01xsSQogpwArgDillj5RyO/Aova7GA8wQQuRLKTullBuDjucBM6SUPinl+1LK9qGOQ5HYKCFQTET+BHwS+BxhYSEgH7ADR4OOHQVK9e9LgOqw5wzK9J89oYdj2oDfAYXDGGsJ0CKl7OhjPJ8HZgF79fDPZfrxPwGvAk8LIWqFED8VQtiHMQ5FAqOEQDHhkFIeRUsaXwI8F/Z0E9psuizo2FR6XcMJtHBL8HMG1YALyJdSZuv/MqWU84Yx3FogVwiREW08UsoDUsrr0MTmJ8CzQog0KaVHSvl9KWUlsBwtnPUZFIohoIRAMVH5PLBKStkVfFBK6UOLs98nhMgQQpQBt9GbR3gG+JoQYrIQIge4M+hnTwCvAT8XQmQKISxCiAohxLmDGFeynuh1CCEcaDf8DcCP9WML9bE/CSCEuF4IUSCl9ANt+mv4hRDnCSEW6KGudjRx8w9iHApFACUEigmJlPKglHJLH09/FegCDgHvAH8BHtOfewQt5LID2Eqko/gMkATsBlqBZ4HiQQytEy2pa/xbhVbuWo7mDp4HviulfEM/fzWwSwjRiZY4vlZK2Q1M0t+7HS0PshYtXKRQDBqhNqZRKBSKxEY5AoVCoUhwlBAoFApFgqOEQKFQKBIcJQQKhUKR4Iy7Toj5+fmyvLx8tIehUCgU44r333+/SUpZEO25cScE5eXlbNnSV1WgQqFQKKIhhDja13MqNKRQKBQJjhIChUKhSHCUECgUCkWCM+5yBNHweDzU1NTQ09Mz2kMxFYfDweTJk7HbVZNJhUIRPyaEENTU1JCRkUF5eTlCiNEejilIKWlubqampoZp06aN9nAUCsUEYkKEhnp6esjLy5uwIgAghCAvL2/Cux6FQjHyTAghACa0CBgkwu+oUChGngkjBAPR5fJSd7IH1W1VoVAoQkkYIXC6fTR09OA3QQfa2tr4zW9+M+ifu+SSS2hraxv4RIVCoTCRhBECI6riN8ER9CUEXq+3359bs2YN2dnZcR+PQqFQDIYJUTUUCxZdCMyIDN15550cPHiQxYsXY7fbcTgc5OTksHfvXvbv38+VV15JdXU1PT093Hrrrdx8881Ab7uMzs5OLr74Ys466yw2bNhAaWkpL7zwAikpKfEfrEKhUIQx4YTg+//Yxe7a9ojjXr/E5fGRkmTFMsika2VJJt/9aN/7k99///3s3LmT7du389Zbb3HppZeyc+fOQJnnY489Rm5uLt3d3Sxbtoyrr76avLy8kNc4cOAATz31FI888gjXXHMNf/vb37j++usHNU6FQqEYChNOCPpiJOttTjvttJBa/1/96lc8//zzAFRXV3PgwIEIIZg2bRqLFy8G4NRTT+XIkSMjNl6FQpHYTDgh6Gvm3t7j4UhTFxUF6aQlm/trp6WlBb5/6623eOONN3j33XdJTU1l5cqVUdcCJCcnB763Wq10d3ebOkaFQqEwSJhksfGLmpEjyMjIoKOjI+pzJ0+eJCcnh9TUVPbu3cvGjRvjPwCFQqEYBhPOEfSFsRhLEn8lyMvLY8WKFcyfP5+UlBSKiooCz61evZrf/va3zJ07l9mzZ3PGGWfE/f0VCoViOAgzF1gJIVYDvwSswKNSyvvDni8DHgMKgBbgeillTX+vuXTpUhm+Mc2ePXuYO3duv2Ppdns50NBJWV4aWSnjt2lbLL+rQqFQhCOEeF9KuTTac6aFhoQQVuAh4GKgErhOCFEZdtrPgCeklAuBe4EfmzgeALWyWKFQKMIwM0dwGlAlpTwkpXQDTwNXhJ1TCfxb//7NKM/HDUtgQZlZ76BQKBTjEzOFoBSoDnpcox8LZgfwMf37q4AMIURe2DkIIW4WQmwRQmxpbGwc0mCUI1AoFIrojHbV0DeBc4UQ24BzgeOAL/wkKeXDUsqlUsqlBQUFQ3oj5QgUCoUiOmZWDR0HpgQ9nqwfCyClrEV3BEKIdOBqKaUpXdiUI1AoFIromOkINgMzhRDThBBJwLXAi8EnCCHyhRDGGO5CqyAyBWNlsXIECoVCEYppQiCl9AJfAV4F9gDPSCl3CSHuFUJcrp+2EtgnhNgPFAH3mTUeIQQWIUxxBENtQw3w4IMP4nQ64zwihUKhiB1TcwRSyjVSyllSygop5X36sXuklC/q3z8rpZypn3OjlNJl5niEAL8Jr6uEQKFQjGcSZmUxoDkCE2JDwW2oL7jgAgoLC3nmmWdwuVxcddVVfP/736erq4trrrmGmpoafD4f3/nOd6ivr6e2tpbzzjuP/Px83nzzzbiPTaFQKAZi4gnBy3dC3YdRnypze7FYBNisg3vNSQvg4vv7fDq4DfVrr73Gs88+y6ZNm5BScvnll/P222/T2NhISUkJL730EqD1IMrKyuKBBx7gzTffJD8/f3BjUigUijgx2uWjI8sI9KJ+7bXXeO2111iyZAmnnHIKe/fu5cCBAyxYsIDXX3+dO+64g3Xr1pGVlWX+YBQKhSIGJp4j6GfmfryhA5vFwrT8tD7PGS5SSu666y6+8IUvRDy3detW1qxZw913381HPvIR7rnnHtPGoVAoFLGSUI7AgjlVQ8FtqC+66CIee+wxOjs7ATh+/DgNDQ3U1taSmprK9ddfz+23387WrVsjflahUChGg4nnCPpBCHPWEQS3ob744ov55Cc/yZlnnglAeno6Tz75JFVVVdx+++1YLBbsdjv/+7//C8DNN9/M6tWrKSkpUclihUIxKpjahtoMhtqGGuBIUxcen5+ZRRlmDc90VBtqhUIxFEalDfVYxCxHoFAoFOOZhBICs1YWKxQKxXhmwghBLDf48e4IlIgpFAozmBBC4HA4aG5uHvBGOZ4dgZSS5uZmHA7HaA9FoVBMMCZE1dDkyZOpqalhoE1rTnZ76HR5sbSnjNDI4ovD4WDy5MmjPQyFQjHBmBBCYLfbmTZt2oDnPfjGfh584wAHf3QJVssILDNWKBSKccCECA3FSrLeY8jtNaMHqUKhUIxPEkoIHHbt1+3xROyGqVAoFAlLQgmB4QhcyhEoFApFgAQTAu3XdXmVI1AoFAoDU4VACLFaCLFPCFElhLgzyvNThRBvCiG2CSE+EEJcYuZ4HHbNEfR4lCNQKBQKA9OEQAhhBR4CLgYqgeuEEJVhp92NtpfxErTN7Ye232OMKEegUCgUkZjpCE4DqqSUh6SUbuBp4IqwcySQqX+fBdSaOB7lCBQKhSIKZgpBKVAd9LhGPxbM94DrhRA1wBrgq9FeSAhxsxBiixBiy0CLxvoj2a4cgUKhUIQz2sni64DHpZSTgUuAPwkhIsYkpXxYSrlUSrm0oKBgyG/msClHoFAoFOGYKQTHgSlBjyfrx4L5PPAMgJTyXcABmLaLu3IECoVCEYmZQrAZmCmEmCaESEJLBr8Yds4x4CMAQoi5aEIw9NjPAASSxcoRKBQKRQDThEBK6QW+ArwK7EGrDtolhLhXCHG5fto3gJuEEDuAp4DPSRPbgwaSxcoRKBQKRQBTm85JKdegJYGDj90T9P1uYIWZYwjw/uPkv/1zkrlXOQKFQqEIYrSTxSOHuwvryWMk4VWOQKFQKIJIHCGwJQOQjEc5AoVCoQgicYTAqglBmtWnms4pFApFEIkjBDZti8d0u0+1oVYoFIogEkgIkgBIV45AoVAoQkgcIdBDQ+k2Hy7lCBQKhSJA4giBnixOt3iVI1AoFIogEk4I0mwqR6BQKBTBJJ4QKEegUCgUISSOEOg5ghSLTzWdUygUiiASRwj08tE0q1e1oVYoFIogEkgItPJRh/AqR6BQKBRBJJAQaI4g1aIcgUKhUASTOEJgVY5AoVAoopE4QqA7AofwKEegUCgUQSSOEOiOINmiHIFCoVAEkzhCYLGAxa61ofb6MXEjNIVCoRhXmCoEQojVQoh9QogqIcSdUZ7/hRBiu/5vvxCizczxYHOQjBcpwe1T4SGFQqEAE7eqFEJYgYeAC4AaYLMQ4kV9e0oApJT/FXT+V4ElZo0HAFsSSXgAcHn9JNuspr6dQqFQjAfMdASnAVVSykNSSjfwNHBFP+dfh7aBvXnYHAEhUP2GFAqFQsNMISgFqoMe1+jHIhBClAHTgH/38fzNQogtQogtjY2NQx+RNQk7bgC1XaVCoVDojJVk8bXAs1LKqNN0KeXDUsqlUsqlBQUFQ38XmwO7NEJDyhEoFAoFmCsEx4EpQY8n68eicS1mh4UAbEnYpBEaUo5AoVAowFwh2AzMFEJME0Ikod3sXww/SQgxB8gB3jVxLBo2Bza/C0C1olYoFAod04RASukFvgK8CuwBnpFS7hJC3CuEuDzo1GuBp+VIFPZbex2B2q5SoVAoNEwrHwWQUq4B1oQduyfs8ffMHEMINgdWfzugHIFCoVAYjJVk8chgS8bq16qGVPmoQqFQaCSWEFiTsKgcgUKhUISQWEJgc2Dx6esIVPmoQqFQAAknBElYfJojUOWjCoVCoZFgQuBA+JUjUCgUimASSwisSeBVjkChUCiCSSwhsDkQXhdWi3IECoVCYZBgQpAESNJsUjkChUKh0EkwIdD2Lc60+ZUjUCgUCp3EEgJrMgDpNp9qQ61QKBQ6iSUENk0IMmw+etSCMoVCoQASVAjSrT7VdE6hUCh0EksIrEkApClHoFAoFAESSwj0ZHGa1Ts2HEH9btjx19EehUKhSHASTAh0R2Dxjg1H8P4fYM03R3sUCoUiwUkwIdAcQepYyRF4usHbM9qjUCgUCU5iCYFePppq8eIeC47A2wM+N/jHwFgUCkXCYqoQCCFWCyH2CSGqhBB39nHONUKI3UKIXUKIv5g5HqNqKMXiGxsb0xhuQO+IqlAoFKOBaVtVCiGswEPABUANsFkI8aKUcnfQOTOBu4AVUspWIUShWeMBeoVAeMbGxjQeXQi8PWBPGd2xKBSKhMVMR3AaUCWlPCSldANPA1eEnXMT8JCUshVAStlg4niCHIF3bDkCr3IECoVi9DBTCEqB6qDHNfqxYGYBs4QQ64UQG4UQq6O9kBDiZiHEFiHElsbGxqGPSM8ROMaKI/D2hH5VKBSKUWC0k8U2YCawErgOeEQIkR1+kpTyYSnlUinl0oKCgmG8myYEycKH1y/x+kZZDJQjUCgUY4CYhEAIkSaEsOjfzxJCXC6EsA/wY8eBKUGPJ+vHgqkBXpRSeqSUh4H9aMJgDrZeRwBjYAN7QwCUI1AoFKNIrI7gbcAhhCgFXgM+DTw+wM9sBmYKIaYJIZKAa4EXw875O5obQAiRjxYqOhTjmAaPHhpKYowIgUc5AoVCMfrEKgRCSukEPgb8Rkr5cWBefz8gpfQCXwFeBfYAz0gpdwkh7hVCXK6f9irQLITYDbwJ3C6lbB7KLxITVhsIK8m6I+hyeU17q5hQOQKFYtzS0NFDY8fEmMTFWj4qhBBnAp8CPq8fsw70Q1LKNcCasGP3BH0vgdv0fyODLZksu+YEqludTMlNHbG3jkDlCBSKccttf92BzSp4/IbTRnsowyZWR/B1tHr/5/VZ/XS0Gfz4w5ZMdpIE4HBT1+iORTkChWLcUtvWTVVD52gPIy7E5AiklGuBtQB60rhJSvk1MwdmGtZk0ixeHHYLhxpHUQj8Pq29BChHoFCMQ1qdbtp7vHh9fmzW0S7AHB6xVg39RQiRKYRIA3YCu4UQt5s7NJOwJSN8bsrz0jjUOIpqHnzzV45AoRhX+PySk90efH5J/QTIE8QqY5VSynbgSuBlYBpa5dD4w5YM3h4qCtI5NJqhoeCbvxIChWJc0d7twa9FmKlpcY7uYOJArEJg19cNXIle9w9I84ZlIrZk8LmZXpBGdYtz9LqQhgjB+J9RKBSJRKvTHfi+prV7FEcSH2IVgt8BR4A04G0hRBnQbtagTMWqOYLpBWn4JRxrGSVXoByBYoLy9v7G0V+1bzKtTk/g+4QRAinlr6SUpVLKS6TGUeA8k8dmDjYHeN1My08H4OBoJYw9yhEoJh5769r5zGObeGNP/WgPxVTaQhxBgoSGhBBZQogHjMZvQoifo7mD8YctKeAIgNGrHFKOQDEBqW7RZse1bRP7mjYcQX56EsfbEsQRAI8BHcA1+r924A9mDcpUrMngc5HpsJOfnjx6lUMqR6CYgNSd1G6KTZ0T+5pu7dIcwfzSrAkRGop1ZXGFlPLqoMffF0JsN2NApmNLDtx4pxekjd6iMuUIFBOQunbtWp4orRf6otXpxmoRzC3O5J0DTfj8EqtFjPawhkysjqBbCHGW8UAIsQIYnzIYJAQVBWmjV0Iaso5gYn9oFIlD3UntWm6c6I7A6SEn1c6UnFS8fkl9+/iezMXqCG4BnhBCZOmPW4HPmjMkk9HLRwGm5afR0uWmzekmOzVpZMfhCdJR5QgUE4S6du26nuiOwLhnTM7Rtpitae2mJHv8bjcba9XQDinlImAhsFBKuQRYZerIzEIvHwWYPpqVQ4YLSM5UjkAxYag7mTihoZxUO6UBIRjflUODapAhpWzXVxjDSHYMjSe2ZPBqjqC3cmgUEsZe3RE4spUjUEwY6ts1AWjucuP3j881p7HQ5vSQnZpEqe4Cjo/zhPFwOiWNz8yIrdcRTMlNxWYRo5MwNlyAI2tiOQKfZ+BzFBOSjh4PnS4vJVkOfH4Zsvp2otHSpTkCh91KQUbyuK8cGo4QjE+5tzlA+sDnxW61MDUvdXTWEhg5AkfWxHEEtdvgvmJoqx7tkSiC+O4LO/njhiOmv4+RMJ1fqqUSJ2rCWEpJm9NDTpqWV5yck0JN2wQODQkhOoQQ7VH+dQAlIzTG+GLVk8I+vYQ0P41DTaMRGpqAjqDlMPg90B6+NbVitJBS8retx3lum/n/J0bF0AJDCCZonsDp9uH2+cnRC0xKs1MmtiOQUmZIKTOj/MuQUg5YcSSEWC2E2CeEqBJC3Bnl+c8JIRqFENv1fzcO55eJCX0D+961BOkcaXbiG+l4prdbS1zbHRPHERguxzO+Z0cTieYuN50uL3tPtJve/+eEvphs/uTRFwKfX5oW8jVCXjmpdgAm56RS29Y9rnMipu2mIISwAg8BFwOVwHVCiMoop/5VSrlY//eoWeMJEC4E+Wm4vf6RT/Z4XZoI2BwTxxEYAuAZ37OjicTRZu1m6PL6Tc+FGaGhseAIXttVx3k/e4v3j7bG/bXb9PYSRsn55JwUPD5Jwzh2QGZuq3MaUCWlPCSldANPA1eY+H6xYdWFwNfrCICRDw95ezQRCEpej3sCjkAJwVjhSFOvO9t9wtyGwXXtPWSn2slLSyLFbh3VNhMH9UrA3609GPfXbukyHEGvEMD4LiE1UwhKgeCsYY1+LJyrhRAfCCGeFUJMMXE8GmGOYFq+VkI64pVDHkMIJpIjUKGhscbR5i4sApJsFnbVmiwEJ3uYlOlACEF+RtKoOoLjetO71/fUx31fYSM0lJvWGxqC8d2OerQ32vwHUC6lXAi8Dvwx2klCiJuNzqeNjY3De8cwIchPTyLDYRv5yqEJ6QiM0NAE+X0mAEeanZRkpzC7KIPdZgtBew+TshwAFKQn91s11NzpMnVTqNq2bqbmppJktfDI24fi+trhoSFjLYFyBNE5DgTP8CfrxwJIKZullMbV8ihwarQXklI+LKVcKqVcWlBQMLxR2bQL1RAC4fPw2YwtHB7pRWXeHk0EbA6t5YV/AmzkoRzBmONoi5PyvDTmlWSy+0Q7UpqX0Kw76WJSpi4EGcl9OgK/X3LRg+v4ySt7TRtLbVs3c4sz+PjSyTy/7XhcewEZjiA7RXMEKUlW8tOTlCPog83ATCHENCFEEnAt8GLwCUKI4qCHlwN7TByPRlj5KPtf5psdP8Xe8IHpbx2CtwfsKb0OxTeyNnp7dVvcLbNKFo89jjZ3UZaXSmVJJi1d7kB30Hjj9vpp6nRRFIMQ1J7spqnTxXNba/D0U8m05UgL7x5sHvRYpJTUtmm9f246ezpev5/H1h8e9Ov0RZvTQ4bDhs3ae/sszUkd1/sSmCYEUkov8BXgVbQb/DNSyl1CiHuFEJfrp31NCLFLCLED+BrwObPGEyDgCPQPxEnNpMiuBrrdPtPfPoAnyBEEj2cEkFJy8xNbuPefu+P7wsbvoBzBmKDN6abN6dGEoDgTwLTwUEOH9n9fHAgNOWh1eqLe6I0JSKvTw9v7+w71fueFXXznhZ2DHkt7j5cut4/S7BTK8tK4eEExf9l4jI6e+Kx61/oMhTapLMsdpYWpccLUHIGUco2UcpaUskJKeZ9+7B4p5Yv693dJKedJKRdJKc+TUprnFQ1s+n+g3m+IjhMA5NA5sgljbw/YUiJyFiPBgYZOGjpc7Il3FYmqGho07x5s5rqHN9Ljif8k5GizJshleWnMKc5ECPOEwAi9FOlCkJ+hfc6aOyPbTBhCkJFs4/k+Fro53V721bVzqLFz0H+bWn1mbnQDvW7ZVDpcXrYdaxvU60TF1UlHZ2dgDYHBnOIMjrd1c9I5PlusjHayeOQxZuBGKMYQAjHSQuAaNUfwzoEmQKvzNkrh4oLhBLzmCcH3XtzFTU9sMe31R5q3DzTy7qFm3jvcEvfXPqKvISjPSyM92UZ5XppplUMn9K6jgRxBujbBiRYeOlDfSV5aEledUsrru+ujztQ/rDmJX4JfaucPBkMIDHdSUahVBh5tiYNT/fPHubrxfwPtJQzm6o5rT525CXmzSDwhMHIExgy8ow6AHNHB4ZFcS+DtDs0RjKAjWF/VhNBbBu6N54U7Ao5g27FWXt9dH1jFOt4xEoxv7m2I+2sbjmBqrlbeWFmcadpaAqP9dCA0lKELQWfkBKeqsZOKwnSuXFKKy+vnlZ11Eedsr+6dvQ/25lqrj8Wo5inKcJBss3A0HhO9tmPkek5EhIaM0FvcXfYIkXhCEH7j1R1BaVL3yMb4RskReHx+Nh5q5vy5RQDsr+uI44ubnyw22hyv+TDy5jEeMUoO39pnjhBMynSQkmQFoLIkk2MtTtrjFCsPpr69h2SbhSy9kiYgBGGOQErJgfoOZhams2RKNmV5qbywvTbi9XbUtFGanUKK3Trom2ttWzd2qyBfdyUWi2BqbuqAjqC/xHXvSV3Yfd1kh4WGCjOSyU1LUkIwbggrHzUcQWmSc2S3rfR0964jCB6PyeyobqPL7eOqJaXkpNrZVx9PITC3fNTnl4Ha9Jc+iLx5jEdqWrtJtlk40uyMe2jSqBgyqCzRZ61DDA+1dLl5bmtN1BLUEye1NQRCt5rGTbgpLEfQ2OmivcfLzMJ0hBBcubiU9QebIso7tx9r45SyHGZPymDvicFdo7Vt3UzKcmAJ2kO4LC+VY819X5cnuz0su+8NPvPYpn7dpnR34ZDdEY5ACEFlcSZ7YhyrlHJki1MGIPGEILh81NUBbi0cVGDr4lBjp6l11iF4Xb0ri2HEHME7elhoeUWe9iGLqyMwNzTU3OnC55eUZqew9Vhbn+V6Lq+P6njEg02mx+OjscPFRxdpjXzjHR460qytITCYZ1QODWHW6vdLvvrUVm57ZkfUPEN9e0+gdBTAYbeS4bBFOIIqPd4/ozADgCuXlCIlvLC9N2nc0N5D7ckeFk3OYm5xBnvqBrf+obatm5Ks0G0jp+amcbSlq8/X2XiomTanh/VVTVz4i7d59v0ogufzIHxu0uiOSBYDzC3OYF99R0zN/V7ZWceSH7wW39DsMEg8IQi+8Xb0hhey6KS9xxtT8nTPifYhV3nsr+9g3f4GLUcQIgQj4wg2VDWzoDSL7NQk5kzKZH9dR/y6JgZCQ+bchI2w0OeWlwPw8ocnIs7x+vzc+MctfOSBtaPa6yYWjKTm8oo8KgrSeDOO4aFOl5emThdTgxxBQUYy+elJ7DweefM56fT060geW3+Y9VVaTf/6qqaI5+vaewL5geD3CxeCA3rF0MwircfXtPw0TpmazdObqwM33m16fmDJ1GzmTMqkzekJ/N/HQm1bTyA/YFCen0qPx99nY7h3DzaTYrfyyq1nM2dSBt/8vx08+d6x0JPc2t8nTbgiksWgJYzdMTb3e+9wCz0eP/e8sGvkJp/9kHhCYNWV3OsO5AfILiPddxKg3/BQU6eLrz61jYt/uW7INfg/XrOHrz75nvbAHhwaMt8RdLm8bD3WyooZ+QDMnpRBl9sXv4UwAUcQ+rs0dPTw/57dMexZuhE+OH16LvNKMvnnB5FC8NNX97HuQBNur583dtcP6/3MxkgUl2ancN7sQt471ILT7Y3Lax8NqhgyEEJwZkU+L+88EUjugiaen3x0I5f+anOXPmkAACAASURBVF3IcYPdte389JV9XFBZxKyidN4JEwIpJfVBq4oNCtIjhaCqoZOMZBuFeg4B4PozyjjU2BVYPLa9ug2bRTCvJKu3GidGF+PzS+raeyI2kjcS5kf7CA+tr2pi2bRcZhZl8PTNZzI9P423wh2aPsHRHEF0IYDYHNeeE+0kWS1sOtzCizsGDnP6/JLvvrDTtMrGxBMCIfRGb0GOoGgeSe42QHK4j4Tx37cd5/wH1vLqzjoqizN5dktN1A9Nf0gp2V7dhs+t/1yQI/C4uk0vX910uAWvX3JWkBAAwwoPHWnq4ra/bufPG48go4SGnG4vn398C89sqeHJjUeHPnigXl+0VJTp4NKFxWyvbgsRl79vO87Dbx/i02eUMSU3hVd2je2EsiEEk3NTOW9OIW6fnw1Vg19JG43eNQSpIcdvv3A2Xr/kxy/3LuJ/4t2j7KrVXO6P1oQu7u/x+Pj6X7eRlWrnJ1cv5KwZBWw63BLiiFu63Lh9/pDQEOiOoDPcEXQwoyg9kEsAuGRBMTmpdv6kXx/bj7UxtzgTh90auEZjrRxq6OjB55cUZ4eOpUwXREMgw3/mQEMnyyvyALBaBKeU5bC9ui10th7kCLJTrBGvU1GQjt0qBswTSCnZc6Kdj51SyoLSLO57ac+Ai90efGM/f3z3KFuOxL/MGBJRCEBrRe0LcgRF8xA+F1lWNwejlJB+UNPG1/+6nYqCdNbceha/+/Sp+KTk4UE2szra7KTV6cGBHn4KSha/vP0Iqx9825SKDoN3qppIslk4tSwHgFlF2ods3zDilI++c4jnth3n3r9vQ+i7l7p6OvH5JV6fn6/+ZRu7ak8yJTeFV3fVDcsG17e7sAjIS0vi0gVad5J/fFDLvroOntp0jDv+9gGnTcvlno9WsnreJNZXNZn69xwuNa1ObBZBUUYyS8tzSEuyxi081JcQTM1L5QvnTOeF7bVsOtzCiZPd/Py1faycXcBXVs3kxR21gZm52+vn609vZ399Jz/7+CJy05I4e2Y+Lq+fLUd6+/wbbStiCQ1VNXQyszA95JjDbuWapVN4bXc9tW3dfFDTxuIp2QBkpdgpzU4JSRgfburivpd2819/3c6nf/8et/11e+C6MsJtZSnd8NZPAj28SrNTsAg4FsWVGr/vior8wLHFU7Jp7nKH9g9y94pIrj0yNJxkszCjMGNA93LiZA/tPV7mlWRy7xXzaOhw8at/Hejz/Df3NvDrf1dxzdLJfHypOQ2aE1MIjI6fHXWQlA7ZUwGYl+OL6ggeWXeYjGQbj9+wjBmFGUzJTeXKxaX8ZdPRQcWhd9Rosc95hdrN32vtLR/dcbgel9fPxiH0VomV9VVNLCvPwWHXZjPpyTYm56QM2RF4fH5e+uAEly0s5uUvaf0C3djwu5xc9ODbfOnPW/nX3gbuvWI+N59TwZFm57D6GzW095CfnozNaqEsL40FpVn89JV9XPTg29z13IeUZKfwm0+dgt1qYfX8SXh80pT6/HhxvK2b4mwHNquFZJuVFTPyeWtfY1xixkebu/TOupFJzS+tnEFJloN7XtjJd1/Yhdcv+cEV8/nSygom56TwvRd30eXycsuT7/PKrjq++9FKzp2lNXs8fXoudqtgXVVva4h/7dH+xrP02btBfnoynS5voDqmtctNU6ebGWFCAPDJ06fil5IfvrSbLrcvIASgJWGNm6vfL/nik+/z+IYjbD7SwvHWbp7bdjyQwK7V20/PaF4Lb/0ImvYD2k26NCeFI1FCQ+urmsh02AJVVUDg/bceC9rYJij3lW2L/rmfW5wREhp6/2gLv3mrKuQc43eZW5zJkqk5fGLpFH7/zmGu+J93uOu5D3hy49FA4Up1i5Ov/3U7lcWZ3HvF/KjvGQ8SWAjc0F4LGcWQkgtAZZYnIkdQ3eJkzYcnuO70qSEfqi+dV4HL6+exd2JvZrXtWBspdiv/efokAKqaPSE5AouInoiLBydOdrO3roOzZ4Z2b50zKYN9QxSCdQcaaXV6uHJxKdOzNHGxZxSQItxYkLy2u55bzq3g+jPKuEBft/DaMOL2dWGVKXdfOpebz5nOg59YzBu3ncMbt50bKFtcMiWHgozkqIuVxgo1rd1Mzu6dsZ83p5Djbd2DEubdte24vJGz0yPNXYFwSDgpSVa+fWkle+s6eG13PV/7yEym5KbisFv5zmWV7Kvv4PwH1vLvvQ3cd9V8blgxLfCzqUk2TpmaE1id7nR7+cP6w6yaU0hFQegN3lhLYEyWqvQOvzMLQwUDtNDNubMKAutDFk8NFoJMDjV10ePx8dKHJ9hb18F//8ci3rljFc/cciZCwOv6dWU4ghyhTzjcvROPstw0jkUJDW042MyZFXlYg8pN50zKwGG3hCxsC3YEKTJ6WLiyOJPGDhdNnS46ejx8+c/b+Okr+0LKYw0hmKPnFL592VxuOmc6qUk21nxYx91/38mqn6/l7J++yacefQ+/lPzv9acEJnBmkMBCoDuCjEmQqsUGZ6S7OdrcFbJ/8R/WH0EAN6woD3mJioJ0LplfzBPvHo3aX+TOv33Az17dF3Jse3UbCyZncfpU7QO65Xg3Ut8xbWqWhbNnFkQk4uLFm3u1GdyqOYUhx2dPyuBQU1fUm8lA/H1bLdmpds6ZVRDICwj9b/nKl09jzdfO5o7VswGYlOVg0ZRsXhtG3L6+3UVRZm+S8fTpeXzrkrlcuaSUGYUZIR9ki0Vw0bwi3trXGPd67W63j88/vpnzH1jLzU9s4f6X9/LkxqO8svMEm4+0xByOqml1UprTm9S8oLIIi4A1UaqholHb1s1lv17HH9YfCTkupbZfb1luavQfBC5ZMIlVcwpZUJrFTWdPDxy/sLKIc2cVUN/ew88+vohPnV4W8bNnz8xnV207zZ0unt5UTavTw5dWVkScZwiBUaljuMFojgDg02do75XpsDEtSMTmTMrE55fsq+vgF6/vZ1ZReqDkNj89mVOn5oQIQYbDRrJbn8m7emfnU/MiF5Uda3ZS09odKKAwsFktLCzN7lMIggUmmOAVxj99ZV8gbGYIp/ZcB1NzU0lPtum/r527Lp7LUzefwfZ7LuCtb67kB1fOZ25xJj0eH7+4ZnGfoh4vElMIgnMEGcWQqjmCshQXHp8MrPY86fTw9OZjXL6ohOKwumSAL583g06Xl9+/E5oreP9oK09vruax9YcDNyG318/u2nYWT8kmWWo5go3Hunj3iDb7W1qaytkz8znY2BVT+4Tnt9Ww9Iev80FNbI203tzXQGl2SkR8drb+IRvsquoul5fXd9dzyYJikmyWXtucouUfLL4eKksyQ5KCF1YWsaPm5KCT7AYN7T0UhiUk+2P1vGK6PT7ePjDMzYyCkFLyzWd38O99DUzJSeFgYyePrjvE3X/fyS1PbuXjv32Xm6P0Qup2+0Li5S6vj/p2V2CbQ9BuaqdPy+OlD0/EFB5au78Rv4xcf3Ck2Ul9u4slQbPqcIQQPPqZpTz3peXa/1/Q8Yc+dQov33oO/3Hq5Kg/e5buKtfub+SRdYdYVp7D0vLciPN6+w1p/98H6jtx2C0RpZ0GK2cXMiU3hWXluSGLweYUaw7ixy/v4VBTF7ddMDtE9C+oLGL3iXaOt3Vz3CgddRpC0OuuynJTaXN6QiZuGw5qN2gjURzM4qnZ7Ap2XMFC4IouBEbl0F/eO8aT7x3lc8vLyUtLYl3QNbjnRDtziyNdEWh///L8ND59RhmPfGYpm759PudXFkU9N54kphDYkqI6gmlp2gX7mcc28fruev686ShOt48bg2ZMwVSWZHLZwmJ+9/ahkFWLD76xH7tV4HT7eGOPNlPZc6Idt8+vxR71UtGGbsG3X9hFD3Zm5ycFZiXr+6kckVLy0JtV/Ndfd9DU6Q7MhPrD5fWxvqqJ8+YUhNyYQbPAwKDDQ6/vrqfb4+PKxfruo0alkP63jLaW4KJ52gX9+p7Bh4fcXj/NXW6KMmIXgtOn55KVYufVOFYP/frfVbz0wQnuXD2HP9xwGv/6xkr2/mA1733rI7z0tbP4+KmT2XKklU5XaBnoD1/azaW/Whdwmyf0WLaxzaHBJQuLOdTYFdOK77X7tJvL1mOtdAW9n9Gu4txZhVF/zsBiEditkbeA9GRboFonGgtKs8hKsfPjl/dy4mQPX1o5I+p5k3NSSLJZuG/NHjYcbNIqhgrTQ27ywVgtgme+cCY//Y+FIcfL89Jw2C1sPNTCwslZgevI4AL9RvmG3oOqJDsFuvXqmmAhMCqHWnpv6OsPNlOYkRwR1gJYMiUbt9ffWwUUfE27o0+cctKSmJTp4OWddRRnOvjmRbM5a2Y+71Q14/dLnG4vh5u7mDMpM+rPjxYJKgQO6KjXVhdnFINDmzlNsjl54j9Pw261cNMTW/j5a/s5e2Z+SBIpnLsvrcRmEXz/H7sAbTONdQeauO2C2RRlJgdqhA2LGSwEfmsyh5u6kFYHNr+b2UUZ5Kcn9Zkn8Pkl3/77Tv771X1cubiEucWZbI6hnGzT4Racbl9EWAi0BT12qxh0wvjv249Tmp3CUr0CKfAh0d1VtNXFFQXpTMtPG1J4yChDDA4NDYTdauH8uUW8sbt+SKGvcF7ZeYIHXt/Px5aUcvM5vZMDm9VCUaaDeSVZXLG4FK9fhvy/SCl5Y089DR0uPjyurVcJlI7mhM6OV8+bpIWHoqyRCMbj87O+qonp+Wl4fJL3DvdOHtbub2R6flrIYrJ4YrUIllfk0djhYm5xJitnR981MDs1iT/feDo2i4VPPvIeGw81MyPKDTeY4qwU8tJD/4+tFsFsvcLtGxfOjpjMTC9IZ3pBGq/rVUfFWQ5wRhOC0LUEUkrePdjE8oq8iNeE3jzFdj1h7HIGVQO5+/68GLP9H141n/RkG2fNyKep08Xeug7213ciZa9zGCskphBYk6BNr2nPmARWmyYG3S2cM6uAl289m+99tJKyvFS+fv7Mfl9qUpaDr58/i3/tbeD13fX84o395Kcn8dnlZVy2sIS1+xo56fSwo7qNgoxk7SLVhaByaiFCQFJyipYstgiWV+TzTlVTRGhASsl3X9zJX947xhdXVvDANYs5Y3ou26vbBtz79c29jSTbLJw5PT/iObvVQkVBOq/sPBFzmKm508W6A01cvrikd3ZnLIjrxxEIIbiwsoh3DzZzsntwZZ3h/e5j5epTS2nv8fJ4WBx9sLi9fm5/9gMWT8nmRx9bEPXGAXBqWQ5JVkvIzlr76jsCK2ONjViM8GN4mKQgQwsP/XOA8ND26jY6XF5uPX8myTYLb+/XJg89Hh8bDzVreRsTMYoOvriyos+/BcCy8lzWfO1sbjxrGl6/DKkGGgxXLinlY6eUcs7MyGsYNFew8VAzrU5PmCPoDeEYi8qMEtJ3DzXT1OkOhLrCKc5KoSgzOTCJe/9ATe+TfTgCgJvOns63L5nLqjmaUzH+Vu9UNQYSxZVKCMYANkdvEilD3y0zNRec2ofXbrXwuRXT+Pc3VnJqWWTsM5zPrShnVlE6tz+7g/VVzdxybgWpSTYuX1SC2+fnlV0n2F6t1UYLIQLtJP7z3Dk8+InFWO2OwLGzZuTT2OFif1gP9offPsSTG4/xhXOmc8fqOVgsgmXlufR4/OyqPdnv+N7c18CZFXmBLpTh/L/Vs+l0ebn8f9Zz21+3R43hP/t+Def+95ssvvc1TvvRv/D5JVcsLuk9IeAIDCGInge4cF4RXr/kd2sPDqpMssEQgkGEhgCWV+Szak4h//Pvqj5LfU+c7OZLf36/XyHcUdNGR4+XW86t6Ld6IyXJypKp2YHYM/SGcCbnpARixTWt3VgtIqL2HmILD63d14jVIlg5u5DTp+cFXneT3rrg3D5m6fHi6lNL+c2nTuGyBcUDnpuSZOXuyyrZ9K3zuf6MyORzLNywYhoPXLO4T9G5sFK7rkAX14Aj6J3FpyXbyE9P5miz1nPoF6/vpygzmcsW9v07LJ6SzbbqNmpaneyrDnKy/QjB8hn53BTkGCdlOZhZmM66A03sOdEeKNseSySoEAQtD8/UL4KU3N6LZ5DYrRbuvWI+bU4PBRnJgYt94eQsyvNSeXLjMQ41dfXOhvSwSfmkPK5YXNpbxQSs0Gc8wdVD//yglh+/vJdLFxZzx+o5geNLy7WwTPDinnAON3VxuKmL82b3HS9eNaeIN7+5klvOreCfH5zgsl+v40DQTWjDwSbu+NsHZDrsXL6ohFvOnc7PP74oNM5phIJSjNBQ9KX8S6bk8NFFJfzmrYN84U/vx1xhY4jTYEJDBt+6ZC7dHh+/eH1/xHMur48vPrmVNR/W8ZnHNoX83sEYezicMX3gicHyCq2qpk3f5Hzt/kbmTMrgisUlbD3WRnuPh5pWrUW0LUqMPpbw0NsHGjllajZZKXbO0YsMjrd1s3Z/I0k2C2dMi0x+xpNkm5VLFhT3Ge+PRkFGctTfNx4snpJDfrr2uS7JSo6aIwAoz0vlSLOTdQea2Hykla+cN6NfYV8yNYejzU6+/fxOUnDhT9Y/w30ki/vi7JkFvHe4he3VbcyZlDGov9tIYKoQCCFWCyH2CSGqhBB39nPe1UIIKYRYauZ4AtiCZmHpWk0/qXkBRzAUzpiex31XzeeXn1gcuLCEEFy+qCQQF15iCIHRYM6W0jse/VhpdgrT8tNYX6XNHu5/eS+3PbODZeU5/Pzji0IuoMIMB+V5qf3mCYyKkv6EACDDYefOi+ew5tazEEJw3SPvUdXQSXWLky//eSvlean8+abTufeK+dx+0RyuDq8oiUgWR698slgEv7p2MXdfOpd/7W3go79+h23H+hYyg/oOF3ariNrjZSBmFKZz/RllPLXpWES3x3v/sZvt1W3cfelckqwWrv/9e1F7Im2oamZ+idasbyCWz8hDSth4qIUul5fNR7SQ4zkzC/D5JRuqmjne1t3nrNAID/VVPdTcqeUaztFDDoHQw4FG1u5v5Izpfbu/iYrVIviIHoopcbhB6uHSMCGYqrejfuD1/ZRmp3DNsv5X6hqTt7X7G1lYaMeSmqOFlvsoH+2Ls2fm4/b6+aDm5JjLD4CJQiCEsAIPARcDlcB1QojKKOdlALcC75k1lgj02n1ScrTGb6CFhroHviH1x6dOL2N5WD3y5Xr4RAhYMDlLO2hs5WgsJgtyBAArZuTx770NXPzLdTyy7hDnzMzn4U8vjTpzWVqey5ajrX2GWd7c10BFQeyJwxmFGTx10xkAXPfIRv7z8c34/JJHP7uMzCirVANEJIv7bjAnhODGs6fz9M1n4PL4ueo3G7j9/3ZE3dbQoL69h8IMx5BnUrd+ZCYZDjv3/mM3R5q68Pj8/N+Wav783jG+cO50bjx7On/6/On0ePxc//v3QsbidHvZVt3K8hmxzbIXTc4mxW7l3YNNbDzUjMcnOXdWAaeU5ZCebOPtA43aYrKcfur8FxZzsLGLHTWRYT8th0Qg/DOrKJ3CjGSe3lxNVUNnYBVwonHTOdO48axplCQFTULChKAsN4269h62V7fx1VUzSLb1L5gLSrOwCK20d1aOgKQ07V8/oaFoGCuyobccdixhpiM4DaiSUh6SUrqBp4Eropz3A+AnwMht2mvcgDOCYoOpeUMODfXHjMIM5pVkMqswo3dlcsAROHq/BrWhvnbZVM6fW8gPrpjHpm99hEc/uyxq21uAZeU5tHS5o3ZNbep08d6hlqjVQv2POZ2nbjodKSUHGzv5n0+ewrT8ARa0BBxB31VDkWPP5Y1vnMsXzp3O37cfZ9XP3uqzoqih3UXhEMJCBjlpSfzX+TPZcLCZlT97iznfeYU7n/uQ5RV53H6htuht9qQMHr9hGcdbu0PaAmw+0orHJ1leET1RGU6SzcKyablsONjM2v2NpNitLC3PwW61cGZFHm/tbaCuvSdkMVk4Vy4uITctiZ++sjdC5NfuayQ3LYn5JdrEQgjBWTPzA5uzJ6oQzCjM4O7LKrH0BE3owoVAnxBNzU2NdLVRSEu28Y0LZ/PfH1+IzdetC0H6oB1BapIt0ONrLDoCm4mvXQpUBz2uAU4PPkEIcQowRUr5khDi9r5eSAhxM3AzwNSpU4c/soAQTOo9lpIDni4tyWkfXEJyIH7zqVNCt8HzdIPFplUrGePp6U1Uzi/N4tHPLovptY2FPFuOtETUQj+96Rhun59PDGB/ozGzKIPnv7SCuvYelkVZLBSBx6k5rSR9DDFuTpOebOOui+fyiaVT+NrT27TKnKnZFIYlhevbe6LWeg+Gzy4vZ+GUbKoaOjnW7KTH4+OLKytC4tZLpuZw6cJi/m9LDbddMIsMh50NVU3YrYJlek4mFpZX5HH/y3tpdbpZXpEXmHmeMzM/sPajv4RhhsPO11bN4Hv/2M3a/Y2s1EN7fr/k7QNNnD0zP8QdnTOzgOe2aiW9FQXmrkId8xgTurTCiDJPYzZ+2wWzoq6hiMaXz9PXSazr0q7vIQgBwEXzJrHreHtg7c5YYtSSxUIIC/AA8I2BzpVSPiylXCqlXFpQEIfZTlRHoN/suuPvCsry0gI7MgH67mRBN4EwRzAYpuenkZuWxOawhLHH5+fJjcc4a0Z+6HsPgim5qbGJAGg3fnuK9g96w18xMr0gnV9eu4Ruj4/vvrAr4nltB6yhOwLQZs6nTM3hmqVT+OZFs7n7ssqImnXQKlQ6XV6efV8rF9xwsJklU3NITYp93mSsVG3qdIeUcgZ/P1DlyCdPL6MsL5X7X96Lzy+RUvKjNXto6nRFuLwVM/IRQgsX9VfOmRAYn+HsqRGOYM6kTNbfuYorl5QO/nXdzt7Q0CCTxQCfPbOcd+5YNajraKQwUwiOA8FT0cn6MYMMYD7wlhDiCHAG8OKIJIytURyBkeQcRsI4Zrw9vWIEETmCwSCEYGlZTkSf8td21VPX3sNn9d28TMfjBHtqr8ANYbvKioJ0vn7+TF7eWRey+1i320d7j3fQawiGyuIp2SyZms0fNxyhpcvNztqTIS2KY2FeSRYZDu0DHxyqKctLC4QnpvSTIwAtxPT/LprD3roOnn2/mrue+5BH3znM55aX89GFJSHnFmQk84fPLRtw3UtCYDiCnLIIIYDItRsx4+7URCA5fdA5AtAKJbKibHE5FjBTCDYDM4UQ04QQScC1wIvGk1LKk1LKfClluZSyHNgIXC6ljGzUEm+M8tFgR2CUPZqQJ4jA29M7c4ZhOQLQYu1Hmp00dPSKyR83HGFyTsqg8wNDxgipWSza7zPE7SpvOns680oy+c4LuwLll/VDXEMwHG5YMY0jzU7uf3kPUhJzotjAahGcO6uAGYXplIflV1bOKiDJZmFSDMJ2yYJJLJqSzV3PfcjTm6v56qoZfPejlVGT5itnF0aE1BISZzMIC2RNjioEQ8bjDMoRmLuJ1EhjmhBIKb3AV4BXgT3AM1LKXUKIe4UQl5v1vjFhJGnDk8Uw7hwB9K4n+MXrB+hyedld286mIy185syykOZcpuLp1hwBaCI3xA3s7VYLP7l6Ia1Od2C3rIAQDKLh3HC5eP4kJmU6eGZLDalJVhZNHvyK2PuvXsgzXzgz4vhtF87m/75wZkwxaiEE37l0Lg67lTsvnhO1xYIijO4WLefnyNKaS8ZrP3B3l3aNJ6X122JiPGJqsEpKuQZYE3bsnj7OXWnmWEKwRnEEJuYIIvD0hK5lGKYjWDQ5m0+dPpU/v3eMt/Y16L3lLVxj0m5GUfE4e12OLWVYG9jPL83ixrOn8bu1h7hm6RTqOwbfZ2i42K0WPn1mGf/96j5Om5Yb0qETgD3/hJxymNT3ZiHpyTaIMuSsFDuLBtFqYWl5Lju+e2HMyc0RoasZ3v0fWHlX6ALNsYCzRXP4yXp1jqsjdOI1FPz+Xkfg6VaOYEKQXaYpe05577FAaGh4awliwhsuBMNzBBaL4L6rFvC3L55JVoqdTYdbuGpJaUyLn+KGkSwG3REMrxr41o/MpCTLwbef38lxvUHbYFpQx4PrTptKdqqdCysnRT75j1thw69GbCxjSgQAdj0H7zwAxzYM7efbTwzZNQ5Id4s2sTMq2Fzt/Z8fC8bEZojrCMY6Yy99PRLMvABur9L+Qw1sSZCUMYKhoTBH4HODlNrKsyFyalku//zqWbyxpz5qgzlT8TghRZ/l2lOH/SFPPfIvfna2jU/+s4Pfv3MIh91CpmNkL9fctCQ2fev8wEKgAH6fdp2MxLUyVqnfqX09sQOmrxzcz0oJvzsHlt4A530r3iPTJnNZkyFZr5YbQoVPBIYQ2FO11/X2gM/bWwI+zhlj04wRQohQETBIzRmZ0JA3bK1CYLvK4ccybVYLq+cXj3x1QoQjGHpoCIB/fJ0za//IqjmFNHW6Kcp0jEpsPMlmiXzf7lZAjkxhwVilLkgIBou7E7oaAvsJxx3DEQSEIA7xfMMBJKX33juGsJZgrJKYQtAXw+w3FDNeV6QjgGGFh0adOCWLAW2m1VmHcLbwvY/OI9lmGdGKoQHp0hsCJqoj8PugYbf2/VCEoFPfUa09ti05B41TTxYPJAReF/ztJmg5FP35YAJCkDohhWBi+Jp4MYwOpIPC0x2ZI4D4VTeMBsHJYnvK8Po2ddZrTcO6W5mal8pvP30qqSZu3D1onLoQDLM31bil5bD2/51TDs0HtRtt8iAWLXbp2za218Z/bG6ntpgxNS80WRyNpgPw4TNQdibkRt+FMEBwjsCnd8ydQHkC5QiCSc0bodDQKDmCnpPmiU3w2gh7yvB+F+MGod9oz9N77o8ZDEfgageve3THMhrUf6h9XfwpQPaGiWLFcAQdJ7RqnHhifH5DQkN9JIuNc2MRdGP2b0/rTUJPIEeghCCY1BFyBN7uyHUEYL4j+P2F8OZ98X9dKbUZk7GqeLjJ4nZ9AfpYjcE7g7YSTURXULcThBUWXqM9Hmx4qEsXAr8n9G8ZTvNBaNw3uNc2rpmUXG0FbDEIMgAAIABJREFUMPTtCJyDEYKwqiGITxJ6jKCEIJiUXG324BvcNoqDxuuKXFkM5joCrwsa90JT1cDnDvq19XHHK1lsOAJvt3klhsOhKyg3kIh5gvqdkD9LCw2lFQ5eCDobe7/vLzz00jfgxa8N7rWDHYE9VVth3NfMPeAIYtiiNZAjSOsVGBUamqCkjlCbCU+4IzCEwERHcFLfb7Wrsf/zhoJxs45Xsrg9qCXVWJxxB89iE1EI6nb2LqQrXjQEIajv/b4/Ieish87obcn7JNgRCKGFh/pyBMa1Fcs15tFv+vbUoNCQEoKJSX+ri9tPaNUSw8XnBekL6z5qhIZMdARtR7WvpgiBUWMdtrJ4EHsShxB8cxiLQtDVBOglpSORUxpLOFugvQaKgoSgce/gFhB2NfYu4OzoRwiczYOflAU7AtASxgOGhgbpCAJCMHHaTCghCMboNxQ8IwVo2AMPzodtfxr+e4TvTgYj4wjajmlfR8oRwNCFrb1W268BxmaewNkE2Xr7jpF2BF6Xtqq59cjIvq9Bvd4iPNgRSB80RLYO75POBiiap/0f9+UIpNT+toNNyAc7AtAdQV/J4kE4gmg5AuUIJiglp2gX0LsPhR5/68fg90LVv4b/HsbN3j7SjkAXAndn70UdL8IdgSEIQw0PtddCvrZr2Nh0BM1ajBxGXqhqt8H7j8OmR0b2fQ2MFcVFC7SvxQu1r4MJD3U1QHqRtl94X2sJXO3aZw4G57qcLVqHAKP/UX+hocEkiz1d2oTNYlXJ4gmPIxPO/gYc/DccWqsdq/sQdr+gXQRHNww93GFg3OyjOgIzhSBos7j+KjWGQsARpIR+HYoQ+P1aWWHRPO3xWAy9OJsgs0QrJRxpIWjcq33d/WL/1+K6B2Dbk/F//7oPIa0AMrSN4sku07p8DkYIOhs1IcgsiXTfBs4hJuS7W7QOAQZJ6f3kCPT/u54YQ0PGBMdi1b5X5aMTmGU3QuZkeON72gftrfshOQtW3a3dAIa7LN6IpY70grK2YwTi2vEODwX3YQn+OhQhcDZpZYVG6GGsOQIjZJGaP3LrToIxyilPHtPcQTT8fnjzR/DCl7XKm3hWwdV92JsfAC0hW7wITnwQ28+7u7TZdXoBZBZroh+NYIHtGsTExeg8apCc0ffM3XgPj3PgHIfb2ZsbgAnXeE4JQTh2B5x3F9RuhX//APb+E878Msy+RHv+6Prhvb43mhCMhCM4BgV6uGUwH6xY8ISXj+q/z1BKSI0ZYm6FtpPcWMsR9LRpIYu0fG3mOdI5gsa9kDNNi6/vfiH6OV0NmpgWzYfNj8KfrorP39Hn0d4/vPV28SItdxCL4BiLydIKIbO07xzBsBxBuBD0UzVk5KIGcgXuTq29hMEQ9y0eqyghiMai66BgDqz7OTiy4YxbtCXo6ZPgiBlCYLIj8Lq0mVfJKdpj48MYLyJyBMMIDRk3hqxS7QM91hyBsYbAcAQjHhraB1NOg2nnaEIQLTx0UhfTVXfDlb+F6vfgtbuH/95NB7QuuUZ+wGDSIvC5Ylv8ZVx76YXafiDuTuiJkswdqhBEOII+qoak1K6t7Kna44GuM2MvAoMJtkuZEoJoWKyw6jva98u/qsVAhYDyFZojGE6eILD4agQdwckaQEKpLgRxDw2F5wj0mdMgN7AHeoUgs1RrHDbWhMDIr6Tl6b2pRtAR9LRrjqlgNlReAa2He5O3wbTra0YyS2HxdVCxCo5vHf77h1cMGRiPjUZ0/WGsKk4r0HIEEN0VhAjBIMS2u6W3+g80R+DuiGxl0XNSq3bKrdB/boDrLDhHAHpoaAQdgZTxb8cRhKlCIIRYLYTYJ4SoEkLcGeX5W4QQHwohtgsh3hFCVJo5nkEx51K44WVY8fXeY2XLtZl16+Ghv24sOYLORq0yZLiJaQOjYqhgjjaTiXtoqI/y0SE5guNgsWsz7pSx6Aj0v91oOIKmA9rXgjkw5zJt1Wy08JDhCLIma18L50LzgcgyzJ72wY3fKFkNb9CWW6GFWBr2DPwaAUdQ1CsE0dYSOJu16yA5K3ax9Xm1G3x4aAgib9pGbsf4XcKvs5otoX8bd1eoI0hOH9mqoe5W+GGBdl8wAdOEQAhhBR4CLgYqgeui3Oj/IqVcIKVcDPwUeMCs8QwaIbQbf/DGE2VnaV+HEx6KFhoSQouHG89teUyrFR9Ki99onNQrhrKnarFt05LF4eWjQ8kR1GpJRItF2+hmrAlBwBHoQuA6aX5LEgOjYqhgjvb+ZStg198jJwztx7VFfSl69UzBXC2v0XIw9LwXvgxPXRf7+7fXaL9zcOkzaKWauRW94+sP49pLyx/YEaTmac4rViEwrpWQ0FAf/YaMnQjzDEcQlCPweeAPl2hbcRqEC8FIJ4s76rT/w8F0eR0EZjqC04AqKeUhKaUbeBq4IvgEKWVwcDANiNMU2CQKZmsX59Ehbs8H0YXAeGw4AkMADsZh3QJojkBYtVBBWoF5oSFjtbTxuw01R5Ch3yBGqgngYAhxBMZK9BESq8a92oQhu0x7XHmFNtMPvwGfrNFyLMaGOoVzta/BM3Ypteu4fmfszvPkce0aikbhnNiEoLNBu1Fb7b17hkdbS+DUQzyD2SMkfFUxxOAIooSGOhu0nEdzUF8ujzMsNJQxsqEho9VGepRtU+OAmUJQCgQVr1OjHwtBCPFlIcRBNEcQtcOUEOJmIcQWIcSWxkYTVsbGiuESjr6jPZYSjm0MbUI2ENHWERiPjecMITjwxvDGa9B2TPsAW20mCYFTX2yjX07DKR9tr+2dKRo5gniFyNxO2Pqn4b2es1kLr9kdI9ebyqBxH+TN6HWpcy7TvoYvdGwPu2Hnz9LCSMFCcLJGczfuzthDhe3He8NN4RTM1fcpGOD/vKtBSxSDds2n5kdfS+Bs1v6+gxGCwKrioHUEfe1JYJybU6ZNkoKFoEO/6RohVeijfHQEhaBD78+UMf6EICaklA9JKSuAO4CopQ1SyoellEullEsLCgpGdoDhlK3QLpD9r8ITV8BjF8GzN8R+cwnv1GlgOIKuJs2Cp+Ro1R49J4c/5rZjvS0RTAkNdYf+PkPNEUgZJgS52sxsuNteGux6Hl78yvASp11NvclIIwQxUgnjpn29JcCghdBS87XjwZwMu2HbHVosPDiZG7wGIZYduozX7c8RIAdeZ9PZqE1GDDKL+w8NDSYPE3AEYcliiGwzERxGcmSFCYE+nla9P5eUUcpH9dBQvCYpA2E4gnEoBMeBKUGPJ+vH+uJp4EoTxxMfylZoX/9yDdR9oM3KDq+Nvf2EZwBH8P/bO/P4qsoz8X8fQgIhbCGBsISdoOxbQKRUKzIWtFWrtcWlWmtrx9ZqtfNTW3/TmXbaOlWni9SxtaPWqrXWpVNGO1TrhoqIgCgqKKgBwhowYd8C7/zxnDf33HPPuUu4N4Hc9/v55HNyzz25ec8957zP/rybluvrqd/QrAZb4Qz68P58TOatpBvWxdLkSnrpZJbNDITGffFmc3MFwb56/Sw72VjNLluuFxvkD06cmbB3uwpTiE04LVFUdnCvTkw9T4zfXz48FkSGpmU+EybsXiPiXTeZCoIDuzQe0i1CEPS07qcU7qHdW2IWAeg4o4LFnco892CGFkGYayhoEez7GBCNQwWz06xFsO9j/bvDB/VZDAaLjzRmJ+V7w1L49wGxDsFh7Nqs7qiwtdazQC4FwetAlYgMFpEiYA4wz3+AiFT5Xp4FrOZYp2IUDJ+tE/W3lsHn79O+7M98P73upE2uoQiLwLqFqq/QjIk1PvfQ8z/RitIVf0r83KhJt/GgZjo1CYKeelOnU1afLkGLoF2B+rIz1eSbUkd9MQLInuvFZr0cTXX4Xp9F0KkFLYLtqwETbxEAlFfFn8+uTbrMZ3DC7jlCJ3yriGx8Q/dJu/QEgc1E6hrhGirzMofqUmQO7fHaS1i6hFgERw7rxGwtgsZ96fXH2vK2PkclPkETJQj2fqyWQLsCFQT+58E/nvq1saBwYaCOAMIDxusWwbLfpx6vpXapWv7rF0cfs2tzrK1HDsiZIDDGNAJXA38DVgJ/Msa8IyI/FJGzvcOuFpF3RGQ5cD1wWa7GkzXaFcBFf4RZt+hE0L4ITv8X7b745sOp/75xPyAaLPNjLYKNy7VytKQMhpyqloYxun/10+rPXPk/8X9bvxZuHQJvPpL4/3Zu0ImhSRB42mw2i8qCggDUHZGpReCvIYAcWAQ1ut12FPrGHq+9BLSsa8gWa4VZBHu3x+JUOyMm7F4j9D7Y9r53P72hhWnd+qcnCOznRlkEBYUav0hmERzcqy6WONdQPx2/v8XD/h06VisIIPV3bAysegqGnh5fo1MUkTXkr0COsghA27f7W1A3fW6SBewXzoW/3pB+23rbIj5ZHcbuLbHgeg7IaYzAGPNXY8xwY8xQY8yPvX3fN8bM836/1hgzyhgz3hhzmjEmg162xxCjPgf9quG5H6fWXOzavjajw+K3CPqM033DZmq8oG4VvHS7Wgin3qg3jN89tOx+1b5fuj3R5WMDXn6LALIbJ/AvU2kp7NQMi8BOYr4YAWTP9fKxdQ010yIwRi2CEm9yKuqk59kSweK6VapxB3P4bRfU7Z5ws+6F4ITtzxyqr1ENuO8E/bx06mKark2EIAAVUskyh/b4qootXb3Jzd9zyE76cYIgRUB703JNkz7xrPj9ySyC4ihBsEmVMdDnp0kQBFpMQLgg2P6BWjHbP0h8LwwrCLYkEQS7NsdbUlmm1YPFbQIROOPf1Nf5yi+SH9t4IDE+ALpv1ya9KfqO133DTtftwrlqBZz0dRh/ke5b5VkFhw9pl8lOZTrBrX46/nNbRBCEWQTNWMB+50Z1VdgbPpsWwYFdOpm0L1YNuDm5/wf36DlZiwC86uKWEATv6aRt2ytbyj3vqhVuURN2j6FaoLX1Xe2jBVpp3mNIBq4hiQnpMHqNUCETpQzZJSr9rpumorIwQdAjfYtg5ZNqLZ8wO35/QaFe86QWQfdEQVAxSl1B9Wt9q5Ol4Ro6cjj2fW5ZkXzMFvuMRlkExniuodwEisEJguwxcBqMnQMv3prouvFzaF9iDQHoPlvwYy2CbpXqx13+kN54U6/S7J++E2L/4/35ajZ+5hfqDlg4N/5zG9bp5GonBquNZbO6OJhjDc1bwH7nRs2TtumRVhBkY6K1GSBDTvWKq5pRHe4vJrN06tEyweK69xLjA6ACvqBDTBDs2KApkx27xh/XvkhdN3Wr1C1U0EHvrR6DdRJM9R3vrFUBHXRp+umZInOoySLwuYa6hBSVhVoEKca36kl9Bv2BYktY47l99bH7q7hU3VHWlbNrkwqo0oHRrqGoQrUdtZrpBrqkZzrYzsD1NeExhwM71cJwFsFxwmd/AZXV8PjXwlsEG6MPYoeuie/5rYQ+42O/V83U7eQrYjf5iM9qpsGOWi0579JXu6NOvUprHPzpkQ3r9H37ABeXqmBoCYugOa4hv8ZZ2FEFSjYsAhsfqDpDt81xD/kbzlkyyWppLo0HVMsMxgdAY1Zlw2Jxj2ANgZ9eIzyLYLn2B2pfFHM1pXIP7dgQHR/wfz5Eu4f8nUctYdXFoYIgyXe8bY3+zxGfDX8/TBDsrY93DYEKg4N7dduljxbu1a/1rU4WSB+FxInbFqFJQXgfqCD7d+r9XTkZMOExlqYaguM0RpB3FBbDnD+o++UPc2KZFpYVj0Ht6zphB7FWQrcB8VrNuAu1tcXJ34rtG+HF2hfdpcHkiV9SLXripSpk/KXxO9bH3EKgE0ensph2lg0O7Uu0CNo3I1i8a1PMZ2zJVr8hO9EdjSAItQia0W9ow1J47kfp56BvX6OZXuUhFgF4mUO+GEHUhN1rpCoGG5aqVQkxQZDKQkomYCw9hnjup4jMoSZB4LMIOnZVazdKEHTsropLMkFg3aTB+IAlKAgaD2ojuk4BQbCvPuai6tJHn5uGtbE4QLCgDBIFgXULDZquazekwrZ/OWGWbsOW/GyqIXAWwfFD515w0SN6gzx4Xuzm378Tnr5ZW0FPvDTx76xF0Hdc/P6KUXD5U/HmdHmVmvV2wp/wJd127AqTvqz9Z5b+TnsW1a2KFwTgVRdn2zV0lMHiw4dU++oWGGu2OpDW12i6YPf+aiE1J3Ooqb2Er2CpOR1Il9wHC25Lv/akdolu+44Pf798uJ5f44EUFoFnURzaGxMEpYN0m0wQGJNYpBaGzRyKsgj2bNXrGYxzlA6Kb+ewd7v69Ys6eT2nUnzHK5/U84kaX1AQNBWTlcZv9zXEMoa69FbX0MHdsQB8sMUEJAaLt69RgTFspgqVVF0HbHxg0CkagwgLGNsx5ai9BDhBkBsqRmqKacN6+N1ZeiFf+HcVCmfdrlp5EGsR9BmX+F4Y1gyu+odY1TDASf+oD+T/XAtPXqcPUGV1/N9mu7r40P4I11AGFsHG5eoH7T8lfn+n0izFCGpik14w9z5doiyC/Tu0kCtdrKa44Nb0rIL1i3UyLBsW/n75cG8B+ZV6XaMmxF6+no9WEBQWq+BIFjDe36AB01QWASTvObR7a7xbyFIxOt6NYvsMWZK1mdi5ETYsibYGIEQQBArPOnb39vssgq59Yz2drIWTTvro9jVqGdnW3KkCxlYQlA7S7y7MItjlLILjl0HT4ZLH9Ea95wx47deqrfebFH68tQj6TEjv80efDwVFmknkp1s/uGY5XL0Url8FN62HKV+LP6akV+aCoL5GLY0gR44kVhaDZxFkkDVk+zfZym1LNi2CJkHgVeNm2h5gzzYNsvpdBJ3KAJN+gd7hQzqxdOmjLURqXkr9N+sXQf+TElOOLTZz6MMXdBs1YZcOUoWjfXG8m6l0cHJBsCNFDYGfniPi/ep+9tTFp45aeo/2tGdP0No+Q5Zk7rf3/qrbEyPiAxBbk8DS1JMomWvIswggls3jFwSFxeqySogRfKAC2y7ekypgXL9Wn5WSchXUYRbB7i16TFhsMUs4QZBLBk6DSx6PVTGe/v3oY61Gna5F0OtEuGmdmqBBuvaB8mG6DWaPQOauoSNH4NHL4dHLEnOjo3onZRosrnlZJye/Cwy8GMFRWgRHDqvmZXPDy4dru4RMi+psewn/hJxpdfG21ZpVctr3NAtkwW3Jj9+zXbXMASdFH2MthQ+f123UhN2uQCebfhPj26v3SCEIoorUwmjqORTSxmP31vj4gKW3nTQ97TlBECRxDa1dqIIvLKPKEuUaCsYI9nuuITvpWoug7j2t4SjwubREVCHwr0nQeFBjCmXD9D7u3Dt1wLhhrbpuRdQNvHdb4n1pawiiFIEs4ARBrhkwFb7+Inxlfnhqm2XchXDuXYkTYTKCk2+6lJRrSlq6Gvvyh2K558HS+eDqZP6xpesaOtyoZfmDPpH4XjY6kO7cqP1i/K4hyLznkL/hnCXTNhh2squcDNOugY8WwLrXdN/uOg3k+lnvvdc/iSDo0Fkn6bWv6utkE/bn74HP/Tp+X48h6r+PWmglqkgtDLuwfVhjvyiLwGrPdtK0fYYsyVxDtUv0u0w2SUa5hppiBD7X0M6Naq2JqBJVXKqCu7Ak8X8EO5DW12hFtF3joPfo1BaBvw+YzbraEnAP5biGAJwgaBnKhibXWEDNUFsslmuCRWUH90YLhX0N8Pd/hf5T4YSzVCj4V7oKLkpjsRZBOhP45jf1gRo0PfG9Tj00798+yE//f/hzSNZVMmzqqN81BJnHCfZui48PQOZtJrasUPdSWRVUX66T3FPXwz2fhtur4Lcz4oXB+kWaidM3hcuwvCqWv56s6KvHkMTkgVQppDs3qEacTh57jyHatuKD5+L379mmykdYCmRJmQbw4yyCEEEQvJd2b1WNunJy8jF16KKKgG0QF3QNFRRq8HdfvTfp+sZov6uigOsTEhewtwFva6FVjNZ4SXBlOD9xgmCUboNZV7udIHDkAr8gOHIYfncm/GpyuHvghVtUgzrzNqj+iv7Nqidj7weXqbQUFgNGH77GA9owz9/DxY9d8W1giCDw+28P7oHX79GeTlGfFUZQEHTtqw9xpplDe7bF1xBA5h1IN6/QZIKC9qpRTr9ONeFDe+HUG/R79C9HuH6xZgulsv6scCvuET5pJaMphTTCPbRjg06OYUkOQUS0Iv7DF+Ort9+fr9uhp4X/ndWeDx/ylpsMCIIjjYmtpG02VTAZIkhwTYJ9H6ubx+/zt9XFuzbGT7rWPRTW9bOoJN6KsgWh9vvsPQaOHIpWOPbvUHeUFQSde+qzGQwY79qS04whcIIgP/FXFy//gxa/7dmqy/P5J8cNS2Hxb2HS5dBnrD7E3QbET1SRFoFvAfvFd8OLP9Xc+TBqXlYtKiwrwt9v6P353v8z8PYT6Z9vfY0W+HTzsqtEMs8cOnJYhWDQIsgkRmCMCgLrEwc4+Wq4sQb+8SWNG4w6D1Y8rpNW4wF1sSRzC1msuysd902QHl7sJEoQBAv9UjFspgZn/d00Vz2l33/vseF/UzFaXXVWwAcFASTGtWpfV0slVVytqfGcJ0hsnyG/q6dJEGyOr2WxAeOgogM64W9cFtP4t6/Rz7X3hHWTRdUTNPiWkLUEA8YHdut3mcOMIXCCID+xk1n9R/Dcv0HlFPjac6p13TcbXrkD7p2lboriUpjhrRfUrgAmXarrL9igcbIYAeiDteB2dW+8+XD8qk+gE+y6V8PdQhBvEax4XDWjijHw9mPpn2/9R5pi6w+QBvv4p2LdIhVCwfTWwk6aiZNOjGDXJhUYFT5BIBK/otakyzRV850/w6a31N2TliDwLIJ0ArpBOnRRTTTSIqhNL3XUMvgUFbx2qdWDe9RVdOJZ0b783mP0/rPLwAazhiDxO659Xf8ulbVkG8/ZRZ721SfG64pLvVqM/QHXkLUIOpPA+Iv1er73lL62GUOWsmF6b9jYx5Ej8XGzpj5gA2P7KkapO8k2j9ztVRU7i8CRdaxraMFteqPNusUrXPtfnbCf+Wd9WGb8M3x9QfxDM/4SfcitVdAY4Rqy3Uif/4k+gHMe0nS7l38ef9zmFaqphbmFIPa/62tgzTMw+jwYe4FaK+murOVPHbWUV2lVZ7oLkK96Sn37wSwtkejGc9s/iG8ZYDVDv0UQpHKytpJYer/GByAzQdAciwBUu12/GP7+A3joAvjTpWqR2FXjMvncjt10zHYtjTXP6gSbLNfffic2BdZvEdhur36r68hhtZZSxQfsZ0s7eNdbDsXfedRSXBpTDPyuIXvfhLnbhp6mgtcmUAQFQUF7DQCvXQgv/BTmToD/ODF2r9iuo35B0GukKhw2XuMvcMshThDkI0UlOnHvqYMxF8R8rOVVmuF01avwjUVwyj8lTgBd+2iHx2W/166nNhUvyiJYOQ/GfgGGf1o1qDcejG+9sdaLD4RlDEFMW172gAb8Rp+v7hNQCyEdQgWBN3H+chz8bKTGSOwkFMQYjYsM+VRMu/RTNhRWPKpB9f07VfDN/x7cOUWXMt3vuSQ2v6XbilHRYxWBiZdpkdQbD+q403ELdOmtq+VVfTr1sWH0Gqma6MI7VFN99y8amN+zTa2STC2NYTO0pfruOhWixaUwYFr08T2GqPLwkbciX5hryC8Itq5UyykdQVA6UAswl9yjLrd99bFMIUtxqRblQawRHviCxSExgnYFMOES+OB5Hc+ujVAWaBNeMVrdRy/copbG/obYuiUN6zQbya9o9Zuo248W6DbHS1Q2nUpOP91x7FJSrmbr6f8Sv79zLw1mJkvHO+1mvTH/8k144krdF1ZQBhqUO+1m/X36dZpe98ovY8fVvKz5/VE+aCsINi5TzanfJHXzDJim7qFUWUn7d+oEYmsILENnwNRvqlAbepqO6+GLYgFIP1veUe0tSqM977e6JsXLP4e5E2HuJFj0nzox72+Axb/R4zav0HGE1Xb4GftF/d7qVqVnDYBerzkPwfAz0js+yBk/UgXgexvhm6/pd7P4blh0p76fqaVhLafVT2tsZ/jseNdckHYFKiBtzUIqQbAhzUCxZdo1KqDfeDC+BbWlo08wxAWLPUFQGCIIACZcrNtnf6jbYPX39Ovg07fAt1doCnnlFG39YkwsY8j/rPUaqUrKCs/1aRvO5bDzKDhBkL9M/irMvjW+PUW6VIxUi+HSv2gTt+4DEzUWa0pP/mos4FY6EMbN0YV0nvsxPHKJ+o6j4gMQS+0DtQbsQzPmfJ0ogznXoA/Yorug7v2Y+R20CDp0gVk/gbPnwjl3wpefUiH44PmJ6XurngQksde9pWsfOO83GmepGKUP85XPwxfu166wC3+lk1AwUBxFSVmshUi6guBo6dBZr6utcJ/5r6ptW1deJjECgN7jNMPqxZ+qMEzmFmr6m9Gx3/2um6LOKhj9gqD2dRUQQQEfRWW1Kg+v/me0a8jiv5cLi/Wa2dqAIN0HqFJhK5yDgqBsKJz8jdhzVv0VDSrXvBQrJvMjolb62lc0NrNrk7ok/ePLATkVBCIyS0TeE5E1InJTyPvXi8i7IvKWiDwrIgPDPseRAz5xrQYmm4uIukoufBi+/Vaiy6TfJI0xfCpw2T/5HdWGFtym2RFVZ2h/pGR08h6C0efH9o08V2MVKx5NPH7+d2H+TXDnZPj9ubovKAiCdOkNl/63WkkPfC62fgGoIOh/UngxlJ9+k1Q4XjYvlvd/6o06Eb70H9rYLR1BAPqddCrTSaY1aF+k63HbCShVw7kg7drp2BvWqssnnfOwWTZFneOXmxRJLCqrXaKrAmZSbTvtW7rm95FD4cFiuw26Oa9coG7SKPzPUXAFuSCjzlXrY8m98TUEfkafj2bGPe4tUZnbqmLIoSAQkQLgTmA2MBK4UERGBg57A6g2xowFHgNuzdV4HC1M+w768HTsFr+/xxC47h11QVyzDL74QLwmGEZJT20/4fetl5Tr5PLmH+MDvttWq096ypVw5u3m5klMAAAJOElEQVT6N+UnxNIrk1E6CL70Z83seOBc9W/Xr1VNPh2NNoy+4z2rYC5g0hcE/afADR/GUjtbg+794QsPQPUV4a0hUmFX2Bs6I73aBptaGlaB7+83tK9BrcF04gN+hs+Kr7fwYwVBWMFbuxTT5PDZav106RseS/BTWKxxhXfnqZUYJgjKhqqQW/Go114it/EByK1FMAVYY4z50BhzEPgjcI7/AGPM88YY25BmEdCM3DfHcUfnnpkVPZ31M7jgvkSt6JPf0WCaP+awcK66EU65QZvtXTYPrl6c+gG1VIyEi/4EOzdpG/G3HtH9zRUEoFaB8dIB0xUExwqDPwmf+VnzNNJhM1URGH9hesdXeHpisI0HqHDYtVGFtG13km58wNKundZtQEiFeBJBkIr2RZp5N/269I6f9OVYYLo0wgky5gJVQDa+kfMaAoAk0Zujph+w3ve6Fkjm8LwC+N+wN0TkSuBKgAEDQiSoo20T1Yd/4MlqRr/yS9WyCjpoRsaESzLr2RRkwElqqTw8RzN9eo2M9hGnQ9/xGjiuXZK5r/14pqQcblybvhDp0MVr2BYy8XWthDf/AD/u41UKSyzDJhPGX6xxp6CrymYRNXcVsLFfSP/Y8iqttfhoQbhFAJp88Lfvamp1C1gEuRQEaSMilwDVwKlh7xtj7gbuBqiurj6K7mOONsfMH8Cqv8Iz39fA4eFDMa3vaKj6Bzj31/DEVzUecbR87jeatphjX+8xR6bn+/l7w6t4Z/1EM6Lq3lO3ULfKRLdjOhS0D+/p1WQR5H7SBWD69dpZtizCZdmlAgafqh1lW2BMuRQEGwB/Skqlty8OEZkJ3Aycaow5kMPxONoi3fvD9G9rnnb7Yhh59tFp737GXqBaZ/cs5DB06Bxb8NwRTVS7iOJS1ZJzRecKGHCyauotwdDT4BsLkx8z5oI2IQheB6pEZDAqAOYAcaJYRCYAvwFmGWOyuIiuI6+Ydo0WnO2shWnXZvezsyVUHMc2BYWa538sMepcjRGErTmSZXImCIwxjSJyNfA3oAC41xjzjoj8EFhijJkH3AZ0Bh4VNSHXGWPOztWYHG2Uok6ax79uEVRGrADncBxvFJXo0rYtgJijWfCjFaiurjZLloRUfzocDocjEhFZaowJTbVylcUOh8OR5zhB4HA4HHmOEwQOh8OR5zhB4HA4HHmOEwQOh8OR5zhB4HA4HHmOEwQOh8OR5zhB4HA4HHnOcVdQJiJ1wNqUB4ZTDmzL4nCOF/LxvPPxnCE/zzsfzxkyP++BxpjQtrzHnSA4GkRkSVRlXVsmH887H88Z8vO88/GcIbvn7VxDDofDkec4QeBwOBx5Tr4JgrtbewCtRD6edz6eM+TneefjOUMWzzuvYgQOh8PhSCTfLAKHw+FwBHCCwOFwOPKcvBEEIjJLRN4TkTUiclNrjycXiEh/EXleRN4VkXdE5Fpvfw8ReUZEVnvb0tYea7YRkQIReUNEnvReDxaR17zr/YiIFLX2GLONiHQXkcdEZJWIrBSRk/PkWl/n3d9vi8jDItKxrV1vEblXRLaKyNu+faHXVpQ7vHN/S0QmZvr/8kIQiEgBcCcwGxgJXCgiI1t3VDmhEfiOMWYkMBX4pneeNwHPGmOqgGe9122Na4GVvtc/BX5ujBkG1ANXtMqocssvgfnGmBOBcej5t+lrLSL9gGuAamPMaHQZ3Dm0vev9O2BWYF/UtZ0NVHk/VwJ3ZfrP8kIQAFOANcaYD40xB4E/Aue08piyjjFmkzFmmff7LnRi6Iee6/3eYfcD57bOCHODiFQCZwH/5b0WYAbwmHdIWzznbsApwD0AxpiDxpgG2vi19mgPFItIe6ATsIk2dr2NMQuAjwO7o67tOcDvjbII6C4ifTL5f/kiCPoB632va719bRYRGQRMAF4DKowxm7y3NgMVrTSsXPEL4AbgiPe6DGgwxjR6r9vi9R4M1AH3eS6x/xKREtr4tTbGbABuB9ahAmAHsJS2f70h+toe9fyWL4IgrxCRzsDjwLeNMTv97xnNF24zOcMi8hlgqzFmaWuPpYVpD0wE7jLGTAD2EHADtbVrDeD5xc9BBWFfoIREF0qbJ9vXNl8EwQagv+91pbevzSEihagQeMgY84S3e4s1Fb3t1tYaXw74BHC2iNSgLr8ZqO+8u+c6gLZ5vWuBWmPMa97rx1DB0JavNcBM4CNjTJ0x5hDwBHoPtPXrDdHX9qjnt3wRBK8DVV5mQREaXJrXymPKOp5v/B5gpTHmZ7635gGXeb9fBvylpceWK4wx3zXGVBpjBqHX9TljzMXA88DnvcPa1DkDGGM2A+tF5ARv1+nAu7Tha+2xDpgqIp28+92ed5u+3h5R13YecKmXPTQV2OFzIaWHMSYvfoAzgfeBD4CbW3s8OTrH6ai5+Baw3Ps5E/WZPwusBv4O9Gjtsebo/D8FPOn9PgRYDKwBHgU6tPb4cnC+44El3vX+b6A0H6418ANgFfA28ADQoa1db+BhNAZyCLX+roi6toCgWZEfACvQjKqM/p9rMeFwOBx5Tr64hhwOh8MRgRMEDofDkec4QeBwOBx5jhMEDofDkec4QeBwOBx5jhMEDkcAETksIst9P1lr3CYig/wdJR2OY4H2qQ9xOPKOfcaY8a09CIejpXAWgcORJiJSIyK3isgKEVksIsO8/YNE5DmvF/yzIjLA218hIn8WkTe9n2neRxWIyG+9nvpPi0hxq52Uw4ETBA5HGMUB19AXfe/tMMaMAX6Fdj0FmAvcb4wZCzwE3OHtvwN40RgzDu0D9I63vwq40xgzCmgAzs/x+TgcSXGVxQ5HABHZbYzpHLK/BphhjPnQa+632RhTJiLbgD7GmEPe/k3GmHIRqQMqjTEHfJ8xCHjG6OIiiMiNQKEx5ke5PzOHIxxnETgcmWEifs+EA77fD+NidY5WxgkChyMzvujbvur9vhDtfApwMfCS9/uzwFXQtKZyt5YapMORCU4TcTgSKRaR5b7X840xNoW0VETeQrX6C71930JXCvt/6Kphl3v7rwXuFpErUM3/KrSjpMNxTOFiBA5HmngxgmpjzLbWHovDkU2ca8jhcDjyHGcROBwOR57jLAKHw+HIc5wgcDgcjjzHCQKHw+HIc5wgcDgcjjzHCQKHw+HIc/4PFThROHyJTJsAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOydd3yU9f3A39/L3nsQQggj7L0REFQEJ1r31mrVarW2Vau21lE7tbX+tNVWrVr3QFREUAFB9p4hjAAJkIRssvfd9/fH93nunrvcJZdxBPB5v155JXnWfe/ueb6f/fkKKSUmJiYmJiauWHp6ACYmJiYmpyamgDAxMTExcYspIExMTExM3GIKCBMTExMTt5gCwsTExMTELaaAMDExMTFxiykgTH7wCCHShRBSCOHvxbG3CSHWnIxxmZj0NKaAMDmtEELkCiGahBDxLtu3a5N8es+MzGks4UKIGiHEkp4ei4lJVzAFhMnpSA5wvf6PEGIkENpzw2nFlUAjcL4QIvlkvrA3VpCJibeYAsLkdOQd4BbD/7cCbxsPEEJECSHeFkKUCCGOCCEeF0JYtH1+Qoi/CSFKhRCHgYvdnPtfIcRxIUS+EOIPQgi/DozvVuDfwC7gJpdrTxdCrBNCVAghjgkhbtO2hwgh/q6NtVIIsUbbNksIkedyjVwhxGzt76eEEPOFEO8KIaqA24QQk4QQ67XXOC6E+KcQItBw/nAhxFIhRLkQokgI8RshRLIQok4IEWc4bpz2+QV04L2bnEGYAsLkdGQDECmEGKpN3NcB77oc8xIQBfQHZqIEyo+1fXcClwBjgQnAVS7nvgW0AAO1Y+YAP/FmYEKIvsAs4D3t5xaXfUu0sSUAY4Ad2u6/AeOBs4BY4NeAzZvXBC4D5gPR2mtagV8C8cBU4DzgXm0MEcAy4GsgRXuPy6WUhcBK4BrDdW8GPpRSNns5DpMzDFNAmJyu6FbE+cBeIF/fYRAaj0kpq6WUucDfURMeqEnwBSnlMSllOfBnw7lJwEXAL6SUtVLKYuAf2vW84WZgl5QyC/gQGC6EGKvtuwFYJqX8QErZLKUsk1Lu0Cyb24EHpJT5UkqrlHKdlLLRy9dcL6X8XEppk1LWSym3Sik3SClbtPf+H5SQBCUYC6WUf5dSNmifz0Zt3//QLB7tM7we9Tmb/EAx/ZUmpyvvAKuAfri4l1CacwBwxLDtCNBb+zsFOOayT6evdu5xIYS+zeJyfFvcArwGIKXMF0J8j3I5bQf6AIfcnBMPBHvY5w1OYxNCDAKeR1lHoajnfKu229MYAL4A/i2E6AcMBiqllJs6OSaTMwDTgjA5LZFSHkEFqy8CFrjsLgWaUZO9ThoOK+M4aqI07tM5hgowx0spo7WfSCnl8PbGJIQ4C8gAHhNCFAohCoHJwA1a8PgYMMDNqaVAg4d9tRgC8Jpmn+ByjGtL5leAfUCGlDIS+A2gS7tjKLdbK6SUDcDHKCviZkzr4QePKSBMTmfuAM6VUtYaN0opraiJ7o9CiAjN9/8rHHGKj4GfCyFShRAxwKOGc48D3wJ/F0JECiEsQogBQoiZtM+twFJgGCq+MAYYAYQAF6LiA7OFENcIIfyFEHFCiDFSShvwBvC8ECJFC6JPFUIEAQeAYCHExVqw+HEgqJ1xRABVQI0QYghwj2HfIqCXEOIXQogg7fOZbNj/NnAbMA9TQPzgMQWEyWmLlPKQlHKLh933o7Tvw8Aa4H3UJAzKBfQNsBPYRmsL5BYgEMgCTqACwL3aGosQIhgV23hJSllo+MlBTbS3SimPoiyeB4FyVIB6tHaJh4DdwGZt318Bi5SyEhVgfh1lAdUCTllNbngIFe+o1t7rR/oOKWU1Km5zKVAIZAPnGPavRQXHt2lWmskPGGEuGGRiYmJECPEd8L6U8vWeHotJz2IKCBMTEztCiIkoN1kfzdow+QFjuphMTEwAEEL8D1Uj8QtTOJiAaUGYmJiYmHjAtCBMTExMTNxyxhTKxcfHy/T09J4ehomJiclpxdatW0ullK61NcAZJCDS09PZssVTxqOJiYmJiTuEEB7TmU0Xk4mJiYmJW0wBYWJiYmLiFlNAmJiYmJi45YyJQbijubmZvLw8GhoaenooPic4OJjU1FQCAsy1XUxMTLqHM1pA5OXlERERQXp6OobWzWccUkrKysrIy8ujX79+PT0cExOTM4Qz2sXU0NBAXFzcGS0cAIQQxMXF/SAsJRMTk5PHGS0ggDNeOOj8UN6niYnJyeOMFxAmJiZnMM0NsOVNaPF2dVaTjmAKCB9TUVHByy+/3OHzLrroIioqKnwwIhOTM4jDK2DRL2DdSz09kjMSU0D4GE8CoqWlpc3zFi9eTHR0tK+GZWJyZlBTrH6v+QdUF/bsWM5ATAHhYx599FEOHTrEmDFjmDhxIjNmzGDevHkMGzYMgMsvv5zx48czfPhwXn31Vft56enplJaWkpuby9ChQ7nzzjsZPnw4c+bMob6+vqfejomJoqUR5t8ORVk9O47aEsd4vnumZ8dyMlnzD9j6ls9f5oxOczXy9Jd7yCqo6tZrDkuJ5MlL217L/i9/+QuZmZns2LGDlStXcvHFF5OZmWlPR33jjTeIjY2lvr6eiRMncuWVVxIXF+d0jezsbD744ANee+01rrnmGj799FNuuummbn0vJl5wfBdYmyF1fE+PpOcp2A6Zn0LKOEga1nPjqCuDwHAYfxus/xdMugt6jW73tNMaawus+jskDVfv24eYFsRJZtKkSU61Ci+++CKjR49mypQpHDt2jOzs7Fbn9OvXjzFjxgAwfvx4cnNzT9ZwTYx89lP436VQcqCnR9Lz5G9Tv+vKenYctSUQFg9nPwyhsfD1b+BMX+Pm+E5oqoYa37vUfjAWRHua/skiLCzM/vfKlStZtmwZ69evJzQ0lFmzZrmtZQgKCrL/7efnZ7qYeoLKPCjeo/7++Ba48zsIDO3ZMfUkBaeKgCiF0HgIiYZzfgNfPQhH10Pfs3p2XL4kd5X6XV2khKEPU9xNC8LHREREUF3tfvXGyspKYmJiCA0NZd++fWzYsOEkj+4HQH0FLLgbKo517ToHl6nfc/8MJftg8UNdH9vpzMm2IIqyYNGvwGZ13l5bCmHaUgZDLtGO3XNyxtRT5K5Rv1vqoaHSpy9lCggfExcXx7Rp0xgxYgQPP/yw074LLriAlpYWhg4dyqOPPsqUKVN6aJRnMJtfg10fws4PWu87usH7B+zgMohMhSn3wMxfw473YPu73TvW04X6Cig/pP6uKz85r7njPdjyX2XJGakrhTAtZheeBAFhUHbo5IypJ7A2w5H1ymoCqCny6cv9YFxMPcn777/vdntQUBBLlixxu0+PM8THx5OZmWnf/tBDPxDNVfcjd8V8bm6AjVpm2MFlamLXqTgKb1wAs5+E6b9s+zrWZjj8PQz/kRrPzEcgZxWs+BOMbSdZoKURpE39bQkAPzePnI/dBN1OwXb1OyTm5FkQusVSXQgxfdXfUjpbEEJAbH8oP3xyxtQTFOyA5loYc4NSfqoLIWGwz17OtCBMTk0WPwxvz+vaNXZ9BLXF0Hca5G121nazFgJSCYr2OLYRGqsg43z1v8UPhl8BVfmtNVojWQvhD0nwx2T189wA9YAbOfw9/KUv5G/t8NvrMXQB0f+ckyMgbFYVmAWoLnBsb6gEW7NDmwaI7XfyBMSWN+A/M0/Oa+no8YeRV6nfPrYgTAFhcupRsF1pR0c3gM3WuWvYbLD+n5A8CmY/pbT4wysc+/cuVL+9Ka46uAws/tDPMBn0maR+H20jblS4W/0+70k1hsAw+ORW5aLRX/vTO6CxEvZ87jJ+K3z7OBTva3987XF4JWx6revX0SnYpjT1uAFQX97578hbSvYrrRmcv6/aUvU7zLCcctwAOJGrUkF9Tc5qOL7D9+/f9TUTh6kfgOrjPn05U0CYnFpIqVIVAaxNndeQsr+F0gMw7QHoPV65Q7K1QHPVcWUVAFQVeL6G/VrLoM8UCI50bEsaAQGhcGyT5/PqylTq5YxfKTfWVW8qi+OLn6kJbP4d0FgD8YMdQXCdo+tV+4hVz3bsfbtj9d+7t4gsfzukjIXQOCV4G3zcEka3WMD5+9KL5MIMdUOx/ZVVUdWGZddd6JaKLrw6Sv7WjvWQamlS9236DAiKUPdftWlBmPyQyPoCjq6DwRer/ys8rqfeNutehKg+MOwy5RIacK6ahG022LdIHZM6qX0Louo4FO2GjNnO2/38leDRBY076krVJKqTNhlmP61e/7/nw5E1cMnzKo5RnAWV+Y5jszQLZ+8iqD/h/ft2xdoCeVuVO6axpvPX0akpVpNvyjjHe/N1oLpgGwRGQFSa8/dV58aCiO2vfvvazSSl4zU687lmfQGvnQvvXOH991uwDZrroN8MFW+JSPZ5LYQpIEyc2f81vDSh62mhnaG5AZY+AYnD4dzH1TZvYgSu5G+FI2tVxpGftsLewNkqHlG0Wz2c8YOh/yy1zdUd8eGNKoC953PI/sZxvitpU5QbqcmDBllb5uwfB5j6MyX8CrbB2JtVsFG/tm5F2GzKBRaXAdZGVbHcWYoyHRpuVX7bx3qDHizuPU5ZR9C5OMRXD8LCn3v/miljIDLF2aVityCMAmKA+t3RTCabDd69Clb+1bvj68pUXApaf/85q+GFkSqRwR3NDfDt7yCyt1Iw3rjAu/s8ZzUgVEwNIDzZ5/2nTAFh4kz+VijLhvk/ViatL9jxARxc3nr7hpeVxTD3jyrYCG1bEBXH1ANtbXbevu4lCIqCcbc4tumT8M6PlPAYNk9pYNKmhISOzQoHvlFB7U9uhS9/oR7EpBGtX7/PZJBWx6TpijEFU0cI+NErcMkLcNFzalviUDVZ6AIif4uaCM9+WAnL7e95/gzaw+gCayug7i0F20BYVDsLuwVR2rFr1JaqPkJ7Pm+/6rmlSQm5lLEQ2ctFQGiCyWilRSSDfwiU53RsTLs/gYNLIXd1632Hv4c9nzlvMwqgJpc6p6I9asJ/5wrY9Unr6238t7qvL/sX3LxAWamvn68+j7ZiJ7mr1X2oC+aIJFNAnO50tt03wAsvvEBdXV03j6gd6kpB+KkJcvnTvnmNlX+Cj26C4r2ObflbVdrokEtgwDkQEAJhiXCiDQGR9YW6lrFpWXmO2j7hx8pPqxOeqCa1Tf9RQmHoPKWRgvOkU12ofNgX/AWue1+NZeq97tNQUyeo38c8BKr1Kl9XgqPU+AJC1P9CKAF2eKUSdllfqJTYQXNh7I1qUjZ+Vh3h2EbwD1Z/d5cFkTBEBdztAqKDFsSuj8HWooLz7bmCijJVLKr3OIjopSZTXajUlihFwN/RaaBTqa5NtbDsKfW3u89o9d9hyaPO24zXd3UxNWoCI3UiLPgJrH7eUeBXUwyr/gaDLlD3Vr+z4favHQkML46Fdf9sHZuwWSFvC/Sd6tgW0cvMYjrdOe0ERG0pxA1UTc/W/xP2ftn9r9Fcr3ypH9+qHq76E/DxbUr7m2fo6x+d1rbprbsYVvzR4cfd8IoScJN/2vr4gbPVxBSTDskj1euBmnR0dIslth8MuRhu/kwFut0REqMmS3eBaptNZfiEuREQ7sg4X7ksjm1S7qX+s1T7iFHXqgyqzhblHdukWU+i6xaElEpYpYxT/3dGQEipit50wenJ+tLRW3qkjFXfV3OtYwJ2Z6EBxPV3FPJ5w7qXVPpsnynOAkinKl/5+msMlqZRQLi6mJqqwS8IbvkcRlypFK0Xx8L6l2Hpk6oCes4fHMcnDYP7NsO170F0H/j2tyqGZkTP5OptaBQZngRNNY7PwweYAsLHGNt9P/zwwzz33HNMnDiRUaNG8eSTTwJQW1vLxRdfzOjRoxkxYgQfffQRL774IgUFBZxzzjmcc845J2/AdWVqUpvzB/VQfv6ztjN1PNFUC29eBPu+ar2vuV7d6GXZsOiX8Nk9Sou/+i2H+QyqIKpNAVGqMjnqK+D7Z1WwdPs7MOoa5Y5wZaBWxzB0nhbkc2NB6K8Xne7d++wzSX0+rqmODRXKUgl1M4G5o99MJQhW/02NYZhWAxIWr7TNXR+3dqW1R1UBVB5VPuvwJOcguCtHN8Jr57U92Rxeoe6PFNU4koBQZZ10REAU7lJWwdkPK1dQQXsCYjuExEJ0X8P3pblVakuc4w86sf1VqqtrWw53VObDmhdUEeSwy9TkbQwaS+kQrIW7HNvLDylFBNQkbaSxGoLClWVzxetwzdtK2//mMdj5Pky8E+IznM+x+MHQS+DHi1Vqdo6Lq0vP5NKFMzgUHB9mMv1wKqmXPOrIS+8ukkfChX9p8xBju+9vv/2W+fPns2nTJqSUzJs3j1WrVlFSUkJKSgpffaUm08rKSqKionj++edZsWIF8fFeaqHdQV2Zunn9g9SN/b95qoPplf9VN7C3rHtJ+fr7TFKauI6Uynrof46a+Fb8UW2/4K8Ol41OdJrK5rFZ1QPkSm2xGmuvMbDpVZWp01wHU+9zP6Y+k5XgG3mN+j8sXj3k7gREVKp377PPFNj2thJ2xopWPUffnYvJHcGR6lqHvlNjGmz4zMbcqDKfPrtbmxAFjL5WCfC20AV72mTI7N126mf2Nyr2kbsWBl/Qev/u+fD5PSq4P/wKtU0IJQA7ksW0/T2lXY++FvYsaN+CyN+u3Et61g4obT9hkIpBxKS3Pie2v3JLVeY5qq498d0zSpDPftoxCVcVGALw5dCiNdAs3O2IZZUfhvhBULLXjYCocbg3LRYleIZdprLJDi5VyRNt0Weyag1jbXFU3uuZXHEDHcfpn0dNIcQPbH2dbsC0IE4i3377Ld9++y1jx45l3Lhx7Nu3j+x9exk5ciRLly7lkUceYfXq1URFRfXcII1+8+g0+Mky1Xf+o5tgw7+9a6VcmQ9r/0/93eTiItNbTwSEwIyH1GQ9/jaYfHfr60SnqXiAp2IgXYM893Glje54T1kJntYnsFjgrPtVcA+U0IlwyQQ5cUQFpQOC23+foB5maJ3uak/B9NKCAEcqbfp05/Myzlea48FlauLY/Dp8+UD738WxTepzSR6lguBtWRAl+9Vv1yCtlGpxmk/vUGnBd3zjPLbQWO8tiJZG2P2xUhhCYtR7Or7Tc2C2qVZNwLrWHOnOgnAjgPVMJm/iEPsWK4szpq/6jMC51sIoVI9rFoSUUHYYeo1S/7uLQQRG0IrU8TDrURWDaos+k5XQKTYsxqRnclkMU3a4LjB9F6j2qQUhhLgA+D/AD3hdSvkXl/19gTeABKAcuElKmaftuxXQch35g5Tyf10aTDua/slASsljjz3G3Xdrk2HVcSX9E9PYtm0bixcv5vHHH+e8887jiSeeOPkDdOc3D4uHWxepCeLrR5S2POUeGHm150l0+e+V1h8YrjR6I/r/gWHqZr+yjQrfaE37qzjqXqOvLVUxgPBEOPshWPYkTPMydVInItl5Qqg40r7WaSRugHKBHNvonDXVUQsClEW17Cnl7jDiFwB3GarAt76lBETuahXk9MSxDUr79gtQn9/BZZ77PnkSEDnfqzGNuBIuf8U5IAzq/XkrIPYvUe6bsTeq/3uPg42vqO64yW6yxI7vUspEb01A2GNGBepe1d2hrhhrIQa04Z5taVKBcv0+092SxnYeulCNSHF4IOrK1XnJI1U7F1cLoqnGOUGio+hV+sc2KiGkZ3K5xtV0RceHAsJnFoQQwg/4F3AhMAy4Xgjhqtr9DXhbSjkK+D3wZ+3cWOBJYDIwCXhSCBHjq7H6EmO777lz5/LGG29QU1MDDVXkZ++iuLScgiOHCA0N5aabbuLhhx9m27Ztrc49KdSf0PzmLg9dYChc+y5c9rKaXBbep/K83WXW5G9V3VOn3qv83q0EhLaWhZ7B0xZGAeGKlCpoqPugz/o53N3OhOmOiF7OD1jFUWW5eIsQSuM76mpBuEnBbI/EoXDPOhh3a9vHjbpOve91L3k+prleaef6ZBOVqr4Ld0VZLU1qMg0IVZOy8ZjMBUrQX/Zya+EAmovJSwGx8wOlpffXJm3dMjBWShtx9bsHhqmspepCLcZjdR+DiOilpbq2Y0Ho4zZ2gxUWFwtCExCDL4Syg8pa0K8bl6E+s1YupioVg+gs0WnqPeguQmMml5HgaOWu82GxnC9dTJOAg1LKw1LKJuBD4DKXY4YB32l/rzDsnwsslVKWSylPAEsBN47RUx9ju++lS5dyww03MHXqFEaOHc9Vdz9CdU0tu3fuZNKkSYwZM4ann36axx9XhtNdd93FBRdccPKC1Ha3iButzOKnNL+froFbFirXz5JHnN0cNht8/Zh6aKf/SgkWVxeTXUB4sdiObjW4S3VtrFZFZPoEYbE4TP6OENHLoTFaW9SE0BEBAcqlVX7YOSja1mfZ5rWGO7sR3BEQDJPuVu1EPKW/FmxXGVu6C8zuPnHjZio/pCbbUdcAUrWTBvV+9n2l0m09WYveCoimOji0AoZd7ognxfZXE76nQHXpAeWK0jVl0FyCBW1baBaLd0379O9IF+J+AUpIGD+jyjyVcjzgXECqGge7gBighFYrF1MXLQghtOQHLX3ansk1rvVxEcmnbZC6N2Asx81DWQRGdgJXoNxQPwIihBBxHs7t7foCQoi7gLsA0tI6+FCfRJzafUvJAzdeqCbK+MFwIocBg4OYu2tXq/Puv/9+7r///o6/YHO90oTcaXxtYdd6Yz0fIwT0nwmzfgNLHlZugyEXqX3r/k+ZxfP+qYKuAaGeXUzeWBABwWoCd2dBuKui7QwRySq43VSnJgxbi8Ny8ZaoVDXBVh93CLXaMuWH7uh34C0T74A1z6uc+cv/1Xq/HhNJNVgQoFwmySOdj9XdS2NuhJ0fKjfTkIvgyDr1mQxto6tuaJz6/KzNjqp1d+SuUQLd2LLEYlF+dU+B6hO5ENPPeVukZvHZv38PAji2v9L4deorlO/f6F5zJ2QiU5zTnqvy1TZ9nevCXeq1hUUpEoHhHrKYuiAgQAn2rC/UWPK3q8/ZneISkezThn09HaR+CJgphNgOzATyAS9y0xRSylellBOklBMSEro4UZwsqgtU8C2qj5oA/YNVal138vbl3rcxMNIRv/mEH6ssjm8fVy6KI+tg+TNKQ9TXSHArIDrgYgKtFsKNBaFPEOFd/N71wGdNoSHFtYPKRpR2vLHOwFOOfncRGqs+510fufdBH92oMl70MdgtCDeZTCX7AaGqdPtMcqRY7l2oXDV6m3NP44D2+wkdXKruhzSXpUB7j1NaubumdSdyW2cp6cVy7SkIsf1V0aTNpoTecwNb15LYXUyG+z2il7OLqTJfCdeoVGXNFO5SFkRUqhL+QeFu6iBqlODoCrrll7fJUXviLnYUnuTTYjlfCoh8oI/h/1Rtmx0pZYGU8gop5Vjgt9q2Cm/OPS1pqFR+89A4x4MVEKL8i97kbLtScQwaqlxeo0ppj9nfdLwNcUfcIn4BMPdPyj2x8s8w/3YV3J33kuNGDgh1CAQdvS9QQBhe0Z6A6A4LAtSk01kBEa3dqsb+VZ6qqLuTKfcoy2XTq87bbVYlsPWePaAC+RZ/95lMpfs1bTgU0s9WPu/aMlUkOfA85UbxhLfFctlLVRdSV1dVyljlrizMdN5ubYHKY46WKzp6gzpvLAhrIyz5tUoPtjU7WxTGMTtZEL1bZzFF9lb3dPJIFaguO+TIlAoMd64dsbYopSjI0Pm3MySPUsrjoRUqiO8af9CJ6OVTF5MvBcRmIEMI0U8IEQhcByw0HiCEiBdC6GN4DJXRBPANMEcIEaMFp+do2zqM9CYt82TQ0qh86f4haulKHb0Ngp5r7Y76itYCxNaiJnTtQbG/z/wtgFQaXZHLQ9ce7nrbtEXG+TDgPOXqqCtXdRPGltiBoa21qw5bEH3VpOaaCqlXtYYlencdTxiL5U4cAYSy7jqC3X1jEBB1ZR0LUHeG2P4q4Ju10Hl74S6VZWMM2Fv81Ht1V01dst9Rw9FvBiBh7QvqMxnmGjZ0wRsBUXYITuS4t0TsgWoXN1NVnqPq3UhEitpess/59V3RM5k2v6ZSqd21pagtBYSqWNeJTFGfXWONeuaqCiBKs76SR6m1scsOOa7v6mLS/+5KkBrAP1B9Nrs+VokjrvEHnYgkNV7XWF834TMBIaVsAe5DTex7gY+llHuEEL8XQuhOzVnAfiHEASAJ+KN2bjnwDErIbAZ+r23rEMHBwZSVlfW8kJA2ZS6D0oiMQUh9onTVtHWsTerhqnd5+7pJ3lSLtNkoKysjODjYuepZX9xcx2Zt26qoK1WaT0f85nP/pCbpS/7R2rcdEOLGguhAkBqUZiutzqmHYHCHdXESthdfaRZEZIp6ODtCYJhKdXUVEB0NUHeGjPNVkZ5+f4HDRZQ+3fnYqN6tg9Q2K5QaivxSxqnvZsMrjn5QbeGNgNCbELrriBuVqqxA1ziE/n5aCQjt+yrcrWXxeIh7JI9Umv+MB+GKV9X36iog6kqVJW8swjT256opVsIo0iAgrI1qQrYLiDBnJUi3JroagwDl7tMtbk9FkeGGYjkf4NM6CCnlYmCxy7YnDH/PB+Z7OPcNHBZFp0hNTSUvL4+SkpKuXKbrNFYrjT4sHipcMiukhKpSCKyHEDfjtDZDdTEE1UGIoWtmU63joSyTBIdFkpqaCqs3KF9yc50KNk6913HOlw+oFsQ3fKRSKl3RF7jpCIlD4MH97jNvAsK6FqQGh7vnxBFn109tiZogOjqZuxIcpS28UtjxFFcjUakOF5O+VrKvLQjQ2oc8qlw4k+5U23LXqBRMfTLVieytmjAaOZGrJr14TUD4Byr/9+EVkDG3/aIubwRE9lLlknF1F4Fy3aSMa53qahcQrkFqbQIvzHT87XZcsfDLPQ53Z3hS62QHd25A/ZpV+Y44gm4hGrPk4jQXU1C4cxaTLiC6GoMA5ww0YyaXEWO7DV1odSNndKuNgIAA+vVzc1OebD67Rz1wD3pYPvK1+9Qkddui1vuO74T510Df6fBjQ1+j5c+oLpNIOP8ZGPFzR8fHUdcon2vWF442FXXlKqBpbYL/zoXr3tPcCQY66zf3lJapu5iMxVm6BdGWX9tIjIdaiNpi5VfvKq1mHM4AACAASURBVHqqYFWBinX0Pav9c9wRneZoAd1Uo6XgngQLIm6A0rIPLlcCwtqi4g+jrm59bFRv7Z6wOb6z0gPqd8IQx3Hp09X9OqyN7CUdXaGo9SAgmuuVojL+Ns/X6DVaWRnN9Q7FoTxHWTCuQkCfEJtr2/98jUHd8MTWwrHOTTNFu4AocFi5ugURl6EllTQYLIgIDy6mLsYgwFHD0lZLlQjfWhA9ncX0w6As27mHiiuJQz3ns+uupBIX4VKWra0LnOGofi3eqzpJ9pmsgo0NlY7qz92fKOFw3Qfqpnr3itY97utKu3dSCwhR7iFjkzndHPfWgohMBYR77a+rAWqdiBTlHqrK73iKq05UH3UN3XoA3wepwdEqPGeVulcKd6p7wNW9pI/R1uwI8ILjvkoY5Ng26lpVKd9WequOf5CaJD1ZEEfWqgl1YBuZUL1GqfvE2FriRK4Suq49uMKTAG3i78i9Gp6kvhdjLEt3MRmJ0Kqpqwoc7jjdgvDz19aCFo77JDBMCQXdja0vItTVGASo9zftFyql2RM+brdhCghfI6Xy8bp2bzSSOFzdrMZ2wjq6xl1X6qyllR5UQqffDFXYZG1x5L73meSYIPQ4xPZ3lQ91yEWqn07SCNVJ1RifcbcCWlfQM5WMa/bq78ffQ+GVK/6BSoNzzWSqKe4+YRaR7Gjr0BUXU1ONqvDVm9edDBcTqMm3uVZZDvb4w4zWx7lLdS05oCZFoyspug9c+bpzwkFbtNWPKXuZ+q7Tp7nfD47Y1XFDLZC7FFdQMQddMejIvRqeCEhn4ejOYg4IUfGkqgKVHBEQqtJbdfrPVBaPno0VFK7uG/2+1t1N3RGDADj/aa1IzwOhscrSMgXEaUpdmZo04toQEHpzuaI9rfcZ88NLtYImm02ll8ZnqImgqVq5oo5tUgHjmHRVUBQ3UFkXhbtVZotenxASo9xQ9Scc2q6U3Z+77y4A31ynHjp3Od2ecLcuRG1J1zOYdCKSlUtIf63OYEx17WwVdWfpNwP8ApWbJne1iie4c7/p2TjGVNeSfc5daDtDW9XUh5YrZaUtizG6r6qoNnZbPpHrPmYBDrdKRyxIezBXC1S3tV6HnupqTHHVOfcJuPM7x/96rEF3LXVnDMIb7GtT+ybV1RQQvkbPvW7TxTRc/Taa2DrGIjq94rUqT5ntcQMNlsIqVZqfNtlxQ6fPUFrltrfVBDLS4JfWBVZZtvrdVKNcUN2p9epxBmMKntHP7C3Rac7tNlqalNDtLheT0c/dkUZ9Ruyprnndl2HlLYFhKnZy4Gs4uqF1bElHT6/WXSdSqhhEvI8ERG2Zur47d5cRe42BZkHUn1DfrzsLAhzfV0ddTOCw0u3rdbgTECnqM6rMdwhVHYvF2e3lKiCautmC8IbwJJ9VU5sCwteUahNwW/3awxPUjepOQDQb6iN0AWG/ZobSFBOGqKZqJ3IdmQ+gHszGKtjypmo2ZvS36uPRr+ULv7ndgnAVEF4GqHXiBqoH1riSGHSviwlU+4TI3m0f6wl7NXUPWBCg3ExlB9UE5c69BOr79w921EJU5avju8WCcJOFnqelXPeZ0v41kkcqC9pm9ZziqmO3IDrqYsKhabclxPW1r6vynWuW3KHHGhpdLIiTKSB6jXIIwG7GFBC+pixbae/tBT+ThqkiHFf0ArqwRIeLyW6VaFZA+gyH9uUkILSJwtYMY25yvm5UH9UJUrcg3LUd6Cp6FoiTgKjtuAWRPBJ7ozQwtNnoLheTppFG9m67n1BbhMWrybfiqJp8/AJPnpsBnIvQPGnsQmjrQmgCwh6g9pEFcXSD8o/rK9C1Ra9R6j4pO+SFgNC+rw7FIHQLQvPVt7VeR2RvdY9VF7a2IFxx52LyD+78fdQZLvmHqvXwAaaA6E5WPw8f3ui8rfSgyjZytyKakcTh6oF1LWTTBUSv0c4WRGCEY4LUJwS/QEdTMVC50/GDlf/VNdBl8VMpknpqpi8sCLuLySVI3VEBoeef6z7q7mqzoaNrpJ3NYAI1+Ualqsm3rlx9jh2Js3SV+EHKikkc1raQj9IExKHv4Pvn1DZjimtnCI1Vgt+1KPLYJi2g68X3rQeqC3e1LyB0F1NHFISAYBWI111Mbd3vdpejbN+itAsI7R7vjkZ9pxBndB3ESWf/YpVrXVvm0EzKstXD2x5Jw5QGdSLHUYQDzgLi4FLVa6nsoHIRGWMNoPKlXaug572ofK1+br7quAFQrGmRnVkBrT3cBqnrva+i1onopbTU4zvV/zXdLSC01MbOBqh19FTXsATfNupzhxBwxX9Uv6W2iExV6yK/8yNllV74XNetRnuxXLlD425pUu0zJrSRomkkfrBScAp3qdYyofGeJ9rhPwJkxwWbsbFdWxazfj9A+xaE3cWkuZa6o1HfKYQpILoLm81Ry3BkrSoysraogp/BF7V/vq69VhU4C4hmg4AAZT2UHYS0qY5jwuJg7M2Q5sbX626bTlyGatdtbe7cAjftYU9zNbqY6lQFdEcwNkqD7rcgAoJVvx7j2tmdISpVrdGAODk1EK54U+Q35CKVMjzmRhh5Vfe0IzdWU+sTauFupdzoxV7t4R+oJvzC3SoO4cl6ADUpG1fv85bwJIcF4boWhBGj1dBeDMKdi8m0IExaUXnUcZPkrlYCouKI8v+3VQOho7tjXJv2tTQoP26ilgp7fLvSUuNdHpDL/tnxMcdnqF4zJ45ofvOg7tV+PAWp22qR4InkUbDx30qY1RarsXbng9jW0qfeEp2mNFTh13bef08y9FL10524a7dhr8lxXQKmDXqNgv1fqwr8VC8FS0cIT3S09GhrvQ7j/dluDMLFjdrVxYJOMcwYRHehWw+h8Y5iJddgclu4m0xBCQj/YKVR+QWqRdah7bRZb7Gnuh50NJfrTr95oOZKMqa5NtV23MUEWqO0JpU2WVuqHvaT6eP3Bj3Vtbrg5KW4ngrobhonAbFBCczIXu7PcUfyKKXZVxz1XAPRFcKTHK2x26r5CY5UilJwVJuTfVlNI2uOaQqdPYupyhQQJm7QM2zG3wole5Wf3JiO2h56ZbFroK+lQblA/PyVUMhZ5f0120N3ZZVl+6a5nD2LqYtBanAEqo9rK3qdzBRSbzG2Ce8JF1NPod83xqLLY5u8th7yTtSRVVDl3A24LRdTZwlPUvdiY43WmNLzd9QUmkxjWNuW7h++2stNb26nmQBqqiu0E8+sGIQpILqL4iylMQ26UP1/ZI2aeENiveuQap9MXdtjNziER8Jg5bKC7uncGBqrHu7SbK0vTTcLCL9A5W7papAalHD0D1E+6pri7os/dCfRBgFxsoPUPUlIjPLVb/qPSqKoPKbqCLwQEFJK7n1vG7e/tRmZNNyxw0sBkZlfycgnvyG7qLr9g+2prkVtKkRlNY28WzGCBbWe1zivb7LyzZ5ChvaKpEYGs2jzAT7bnnfGxSBMAdFdFGWpOEHKGKVB5Kx29EvyBk/rQrTUOwSEXvEamep9N9T2iMtwdjF1J0Jozcw0F5OUWquNTlgQFj9IGq6tCVzafW02upOIFOyN5H5IFoTFT8VwThyBhfer5U7BqwD1xpxyduVVUljVQF5dgKO9t5cCYtneIqobW1iw3YsFJ+3FcsVt3u9Pf5nF7xuu4bGKyyiucr+Q17K9RdQ1WfndJUOJiIomOcTKgx/vRDZWd0+jPjf0xLo2poDoDlqalLWQOEwVyKRNVU3yyg567wryGINodLYgoO2q7I4SP1CNs7sb9ekEhDjek7VZde3sjAUBjnYMp6qLyT/QkSJ5Co2vtKaRg8VeaNhdoe9ZcN4TkPU5LH9aZbAlDm/3tNdWHSbQT01D246eUK5Ev0BHMVw7bM5VFdxf7Tre/gRqLJbzYEEs31vEwp0FXDhC1cZ8f8D9WjILdxaQGBHE5H5x+AdHMjzegkW2IFoaPLb6llKy8XAZDc3Oq0Mu3n2ciX9cxqJdBW7PA8gtreWCF1bzxBcdXCWyi5gCojsoy1bZQLqJnD5dVT3XFHpvQfgFqlYP7lZgC3AREN4Evb0lbqAyuZuqfeMWCQh1CAj7etSdFBC9RqkW5rbmHncxFVY2cLTMzTKPeqD6FApSP/rpbq54eR31TZ1Y97wjnPVzGHSBcjGljndfe2PgYHE1y/cV89OZ/QkL9GPrkRNw1gNw4bOe1xgx0Gy1se1IBQkRQRwtryMzv6rtE/SCyLJDbtfrqG5o5refZTI4KYIXrhtDQkQQK90IiMr6Zr7fX8Ilo1LwsygrOdLSSBiateHBxbQ0q4hrX93AeX//ns+359PYYuWZRVnc+942ymoa+evX+2i2tl7xcfvRE1zxyjoOltTw9vojHoWWLzAFRHegt8jQU1GNzdK8tSCEUBNnqzRXgwURl6HadA84p2vjNWIUNj6xIEIdQq+j61G7kmzwCXdXm40OYrNJ3lqbwzl/W8lV/17X+oHW4xCniIvpRG0TK/cXU9XQwuLdvmnoZsdigctfoSw8g6PJbaz/oPH66hyC/C3cNq0fY9KilYBIHQ8TfuzVy+0pqKK+2covZmcQ4CdYtNuzBg6oeKDwc844NPDs1/sprm7gr1eNIsjfj5mDElh9oIQWl+/4m8xCmqw2LhujWTlB4QRa60gK1uKDHoLUn+/IJyY0gOjQAH7x0Q4m/mEZ/12Tw21npfPyjeM5Vl7Pp1ud1wxfllXE9a9tIDzIn8U/n8GAhDB+s2A3NY0tbl+juzEFRHdQvEdVsOrWQvJoh5nZEW3f6I7RMcYg/APhnrVdL+gyYhRgvtB69VXloOPrUbuSOExZWdAjLpyjZXVc9+oGnvoyiz6xIRRXN7LKVZuLG6itIdDBYsBu4OvM43y3z7nt85LMQlpskqiQAD7cfNTDmd1HUUso40uf5q69Y9p0+RRXN7BgWz5XT0glNiyQ8Wkx7D1eRa2HiS+ntJb/fH8Iq81xzc05yr10/tAkpg+Mb9/NZLEoxUJvimm43+uaWvhk6zGumdCHMX3UdzdrcAJVDS3szKtwuswXO/PpGxfKqFRtDY3AMERTLSNitfiTGwuiuqGZ5XuLmTc6hS/vm84/rh3NwMRwXrx+LE/NG87c4UmM7hPNS98dpKlFCaSV+4v56btbGZwUwYJ7z2JwcgR/vXIUBZX1/O2b/Z7fZzdiCojO8P1zsPZFx/9FWaqdhr4+sp+/ikMIS8fyuf1D3ASpGx0uJl8Q08+3k65R6Om/AzsmIN7dcIQrXl5LPUEOgRuWSIvV5l32Sjdx7/tb2VtYxXNXjWLR/TOICwvkky3OGh9T74OfLGu/91Y3c7C4hvs/2M7PP9hBRV2TffvCnfn0jw/j3lkD2Jx7wuexCN39sa+wmu/2uVkAS+Od9Udottm4Y7rKxhvXNwabhJ3HKlod29Bs5a63t/DnJfucBPLGnHLS40JJjAzmopG9yDtRz868yrYHGJ7oWGbVcL+v3F9CQ7ONy8Y4CuNmDEzAItQ+neLqBtYfKuOy0SkIvQ5HW3Y0I0b9L91YEEuzimhssTFvTAoWi+BHY1NZcO805o1WVogQgl/OziC/op75W/PYfvQE97y7jUFJEbz7k8nEh6uCvgnpsdw6NZ3/rc/ljTU5vLrqEE9/uYeXlme3/b47iSkgOkpdOax6FpY95WieV7zX4V7Smf4LOPd3HWtlEOBGQDTXe7/6WmfwN3Sa9YmLKcwgIDruYpJS8saaHLYdreD5pfsd9RBhCTy2YDdzX1jF8cr6ti/SDlab5KbXN/LI/F1U1jW7PWZPQSWZ+VU8eP4grp7Qh0B/C5eP7c3yfUWU1zomZIIj7bGopVlFXPLSao9acXchpeSphXsI9LNQ09jC66tzABUn2ZhTzqWjU7hyfCr+FsGHm475dCzf7y8hMSKI1JgQ/rnioEeNftneYs4aEEe/eJWNNzZNrdq29ciJVsc++/V+sotrCAv0450Nal0Qm02y5Ug5E9NVCvmcYckE+Am+chPolVLyn+8Pcay8TgWqbdr3YbAglmQWEhcWyKR+jpT0qNAAxqbFOAmIz7fnY5Mwb4whiB4UDo3VDIxUmv8JW+tn/osdBaTGhDAuLabVPp2ZgxIYmxbNi8uzuf2tzSREBPHW7ROJCHbuDPvw3MH0jg7h94uy+NPifXy8+VgrK6e7MAVER9HXdvYLhG8f1/K+jzpWhdPpexbM+FXHru1OQBhjEL5CdzP5woIIDHWkuTZ1PEi9p6CKw6W1pMaE8N81ORyJOQti0pm/t45PtuZhk7Alt/Wk0hHWHixlzcFSPtpyjDkvfM/yva1X55q/NY9AP4uThnnV+FSarZIvdrROsZRS8sKyA2TmV/FtlvNykFUNzby88iClNY2tzvOGjzcfcwpUfrX7OGsOlvLIhUO4eGQv3lybw4naJhbtKkBqk1l8eBDnD0vi0215NLb4JljdYrWxOruEmYMSuHvmALYfrWDD4dbrRDQ0W8kuqra7cgCiQgIYlBTO1qPO3+Xag6W8sTaHW6b25fbp/Vixv5hj5XUcLKmhoq6ZidqEHhUawIyMBLdupk055fx5yT5eWJbtvG6Cdr83NFv5bm8Rc4YnqaCzgVmDEtidX0lpTSNLs4p47pv9TO0fx8BEgxspMAyaakmLUAIit8p5Wi2raWTNwVIuNVodblBWxCAKqxrwswjevn0SiRGtn/2wIH+++Nk0vv3l2ex6ag57fn8Br9860eN1u4IpIDrK9ndV47xzfqMas63XeiC5WhCdob0YhK9IGKJeowNN9PJOuMngcYdR6HXCgli4swB/i+CDO6eQEBHEXTsHknnVKh7/ch+T+8USHGBR6ZFdYP7WPKJCAvj0nqlEhwRyx/+28H/LHCZ7U4uNL3YUMHtYIjFhgfbtQ3tFMqJ3ZGs3E7DlyAn2FFQhBCzY5ixA/rs6h2e/3s/FL65mS66bhXbaYOX+Yn796S5ufWMTv/p4B/kV9TyzKIvhKZHcOLkvD8zOoK7ZymurD7NwZwEjekcyIEG5PK6blMaJuma+3eOb5Sl3HKugqqGFWYMTuXp8KvHhQfxrxcFWx+0vrKbFJhnZO8pp+/i+MWw7cgKbFmeorGvmwY93MiAhjMcuHMr1k9IQwAebjrJJiz9MSndo/BeP7EVBZQNbXKyQT7TA7+Ldx2kK0bLfDOt1rMkupbbJygUjWrcFmTVYJUP8YVEW9763lWEpUfz75vHOBwWGg62Z3oHqmThQ6SwEFmcWYrVJuzupLWZkxPOHy0fwwZ1TSI/3XOsUFx7EoKQIIoN9u+6EKSA6gr6285ibYPLdyn///bNqX3cJCNcspuaGLsUgahpbqGpw7zaxM/2XcOsir1ILATYeLmP6X1ewJru0/YMDwhzprbrw89KCsNkkX+4sYOagBPrEhvLHy0eyv6iaK19ZR3hQAC/dMJZRqdFsP+psXtc3WXl11SGvXDuV9c18s6eQy8akML5vLF/eP515o1N48btsdmv+7O/2FVNe28TV4/u0Ov/q8X3IOl7FngJn3/eba3OIDPbnjmn9WHuw1F5w1dhi5b2NRxjTJ5rgAD+ue3UDr6w8xJLdx3ljTQ7/WHqAE0aXlYGaxhZ++1kmAxLCuO+cgXyxo4CZz66gqKqRZy4fgZ9FMCgpgotH9uK/a3LYlVfpNCnNGBhP7+iQDgWrj5bV8dLybL7Ykc+x8ro2g8DfHyjBImD6wHiCA/y4c0Y/1hwsZYdLXGF3vvqshqc4C4hxaTFUNbRwqKSGphYb932wjdKaRl64diwhgX6kRIdw3tAkPtp8jLUHS0mICKJvnONemjsimaiQAF5dddi+rbZRZW+N6B1JfbOVzErtWTKs17Eks5CIYH+m9m+dpDE8JZL48EA+31HA8JQo3rljElEhLpOyFpQOb9TiL2XOn9HCHflkJIYzJLn9CmshBDdN6UtG0qlRje1TASGEuEAIsV8IcVAI8aib/WlCiBVCiO1CiF1CiIu07elCiHohxA7t59++HKfXbH9PW9tZa5M85xlAqiBVV9cSAOeUUFCVxy0NKnjdSe57fxtT/rSc/1uWTV2ThwkzNBb6eG+i6hrxiv2eg5B2AkIcLqYOZjFtzi3neGWD3d87e1gSl45Ooclq48XrxpAYEczYtGj2FFQ6FR99ubOAPy3ex1ML97T7Gl/uLKCxxWaf/AP9LTxz+QjiwgL59ae7aLbamL/1GIkRQczIaO2Cu2xMCoF+FuYb0hPzK+r5Zk8R109K4/rJadik8kGDKugqrWniwTmDWHjfdGYNTuSvX+/jnve28ftFWfzf8mw+81AV/NzX+yiorOfZq0bx0NzBfPGzaYxKjeLus/s7+bYfOC+DJqsNIeBSg4CwWARXjU9l3aGydt1bFXVNPLMoi/OeX8nflx7ggQ93MOPZFUz583K+zix0e87K/SWMS4shKlRNoDdO6UtksD9vr8t1Oi4zv5Lo0ABSY5zv6wmaNbApt5yHPtnJ6uxS/nzFSEamOgTJzVP6UlbbxJLMQialxzq5bMKD/LntrHSWZhWxv1AF4xfvPk5dk5WnLh1O/4QwVugfrVbz02y1sWxvEecPTSLQv/V0aLGoCfvcIYm8c8ck9xq73tWgWn0uu0oc92J+RT2bc09w2Zi23UunKj4TEEIIP+BfwIXAMOB6IYSrmv048LGUcixwHfCyYd8hKeUY7eenvhqn17Q0we6P1doOem+lIZfAgPMgbXL3dBb1D3Z2MVmbANnpnv0tVhsbD5cTGujPP5YdYOZzK1nSxVz4xhYrSzLVNdYdcl5msqHZyq1vbLJXtwLq4bE1qyrqDgqIL3YWEBxgYfZQh9/471ePZvmvZnLWQDVZj0uLodkq2VPgKJJapsUQPtma53Ey05m/NY8hyRGM6O2ofo0KCeCZy0ew93gVf/xqLyv2l3DFuFT8/Vo/LtGhgZw/PIn3Nx5lwTYlJN5ZfwQpJTdP7cuAhHBGp0axYHs+UkreXJvLwMRwpg+MJyokgNduGc+n90xl0f3T2fa780mKDCIzv3Umzpbcct7ecIRbp6Yzvq+6/0b0jmLBvdN47KKhTsdmJEVw69R0Lh2VQq8o50l4zvAkpKTNDKNFuwqY+dxK3lybwxVjU1n76Ll89fPpPHP5CJIig7nnva28tTbH6ZyS6kZ251cyc5CjgDE8yJ9zhiSyKrvE7jYCyCyoZERKVKsJMz0ulNiwQP66ZB8LdxbwyAVDuHqCs9U2fWC83WowBpR1bjsrndBAP15ZqVxbn2zNo398GOP7xijhWKRN8FpCxvpDZVTWN3OBVjntjl/MHsQbt7UOFtvRs5aqj9Msgthf7LC0Ptp8DCFg3uhOrnXew/jSgpgEHJRSHpZSNgEfApe5HCMB/cmMAtqpdOlBDnyt+reMMSwpKgTc8BFc/1H3vIZrkFp3N3WysOxgSQ31zVZ+e/EQPr1nKgnhQTz4yU4q69txObXB9/tLqGpoYVJ6LHuPV1Fm0ERXHSjh+wMlvGXUGI0tROyV1O2/n2arjSW7j3P+sGTCghwVuYH+FvonONIIx6apuMl2LQ7R0GxldXYp103sw4jekfzms90UV7vvp3OwuJodxyq4anxqq8lq7vBkLh7Zi7fW5WK1Sa6e4HnhmKcuHc6YPtH86uOdPLZgNx9uPsrc4cmkxqiJ7Edje7P3eBXvbzrK7vxKbjsr3f56QgjG941lRO8oYsMCGdk7il0uAkJKyWMLdpMSFcLDc71bP/qpecN58fqxrbYP6xVJSlQwy7JaxyGaWmw8tXAP972/nQEJYSx+YAZ/vWoUvaNDGJ4Sxc1T+vLRXVOZPTSJp77M4o9fZdkn/tXZyr2i++x1zs5IoLSmib2FSoA3tljZX1jNCJf4g/5Z6G6m26f146czWzektFgEN09RWXdT3LiEYsICuXFyGgt3FrA6u4RNOeVcqX2/V4xNpQwtzqZlMC3JLCQ00I+zB3WhMl/vvVRdSEtAOLVNVvIr6mlssfL+xiOcOziRtLhO1v70ML4UEL0BY05dnrbNyFPATUKIPGAxcL9hXz/N9fS9EGIGbhBC3CWE2CKE2FJS4uPy8x3vu1/b2S+g3ZYCXuPqYtJXk+ukBaHnlI9OjWZ831ievWoUdU1WPtnS+VTHhTsLiA0L5CFtojJmqeja+op9xQ6Xj7FLrf7ePATdd+VV8N81OSzefZz3NhzhRF1zu4G9xIhgUmNC7IHq9YfKqG+2csGIZF64dgy1jS08Mn+XW9/5J1vy8LcILh/rXrt7at5wokMDGN83xh7odUdCRBDv/WQyd8/szwebjlJR18yPpznqXy4dnYK/RfDUwj1EBvtzxTjP2uTI3tEcKqlxqpQ9VFJDdnENPztnoJOw7AxCCM4dmsjq7FInt1xZTSPXvbqet9blcsf0fnx091SGJLfuKRQS6Me/bxrPzVP68trqHGY//z3vrM/lmz2FxIcHMjzF+RzdLbfqgIpXZRfV0GyVThabkbvO7s+vzh/E4xcP9eiS+fG0fnz+s2kM9uDT/8mM/vhbLNz77jYsAq4cp4R7clQwgwaoFveNQTH8Zck+Pt2ax3lDkwgO6ELdisGCEFo84kBRNYt2KnfibdPSO3/tHqang9TXA29JKVOBi4B3hBAW4DiQprmefgW8L4RodUdJKV+VUk6QUk5ISPBxb57C3dB/VvcJA3d4siA6GYPYcaySyGB/0uOUj3RE7ygmpsfwv/W5ThWp3lLb2MKyvUVcNDKZcWnRhAf5s+6QevCbWmws3VtEelwodU1WR0GTccWt5jr1XtwEw602yX3vb7f3pnnqyyyiQgI4e1D7qbfj0mLsgeqle4sIC/Rj6gCVivjYhUNYsb+EZXudXSpWm2TB9nzOGZJoL0JyJSEiiAX3nMU/b2itibvi72fhsQuH8votE/jl7EFMTHfEBOLCg5g5KIFmq+S6SWmEBnq+h0alRiElan0EDV0InzWgeyrdZw9Nor7ZyvrDHoTkrQAAIABJREFUDhfhc9/sJzO/in/dMI7fXTKMADfuNB0/i+D3lw3nXzeMIyIkgN99sYdv9hRxdkYCFpc00cTIYIYkR9gtDD1A7ZrBpDOpXyw/Py+j1XVcX9+YIutKUmQwV09IpbqxhRkZCSRHORSSeZMyeL/lXH66OZl/f3+Ii0f14olLuphgoguI+nICQtU0ta+wmjfX5ZChuRNPV3wpIPIBowMxVdtm5A7gYwAp5XogGIiXUjZKKcu07VuBQ8AgH461fZpqVBGULwkIUU3EbJpm19J1C2J0n2inh+3H0/pxrLy+TR+0J5ZmFdmrTf39LEzqF2uPQ6w7VEp1QwuPXjiUqJAAh+/f2Ma8ud5jFfV3+4o5Wl7Hs1eOYvHPZ/DGbRP44M4pBPm3r9mNTYvmeGUDBRX1LN9bxIyMBPt5N03pS1RIQKvYy6acckqqG/mRB+tBp39CeCs/flvMHpbEA7MzWmm/N0/tS0xoALdM7dvm+brrZZeh8GlTTjmJLhk7XWFK/zhCA/3sbqZj5XXM35rHDZPTuHiUdyvACSG4eFQvPr/3LD69Zyo3TE7j9unuuwacPSiBLbknqGtqITO/kohgf9Jifety+enMAcSEBnDbWelO288flsRr0Q/Q2Gc6i+6fzj+uVU35uoShvbdfSBQpUcHM35pHZn4Vt01LPy2D0zq+FBCbgQwhRD8hRCAqCL3Q5ZijwHkAQoihKAFRIoRI0ILcCCH6AxnAYXqSptruW4PBE/pkqguGLsQg6pus7C+qZnSqs6Y1Z1gSKVHBvOkSZPSGhTsLSIkKZryWMXPWgDhySmspqKjn68xCLSiZwOyhSSzdW6R6ygRon1lznX2xoH8sPcCHm5xTLd9al0OvqGB+NK43w1IiOXdIEsNSvBPIegbPOxuOUFTVyOxhjqC2v5+F84YmsnxfsVNjva8zjxMcYGHW4JPTFXbW4ES2PzHHHpfwREJEEL2igu2atpSSjTllTO4f120TTXCAH2dnJLB8bzFSSl76Lhs/i+CeWQM6fC09hvKnH410G1cA5WZq0hImMvPdB6i7mz6xoWx/Yg7nDHGOiQT5+7HioVm8f+cUj+PtMMbWGoHhZCRFcLiklshg/3YVkFMdnwkIKWULcB/wDbAXla20RwjxeyHEPO2wB4E7hRA7gQ+A26RyFp8N7BJC7ADmAz+VUnasoqg7aWlUmTg+FxAuq8p1IQaxp6ASq00y2sUU9/ezcPPUdNYdKrOnArZHQ7OVjYfLWHWghEtHp9gtkrMGKNN5TXYp32YVce6QRIL8/bhwRDLVDS3K/aQLt6ZaaKrF5h/MyysP8vjnmfY6gwNF1aw9WMbNU/u26drwxNBekQT5W3hzbQ5CwDkuk/6cYclU1jfbm7vZbJKv9xQyc1BCm+6enmJE7yi7gDhSVkdRVSOT3WTsdIXzhiZSWNXA4t2FfLotnxsmp5EU6ZuCzInpsQT5W/huXzF7C6s9xh9OW4wCIijCHhu5vh134umAT2MQUsrFUspBUsoBUso/atuekFIu1P7OklJOk1KO1tJZv9W2fyqlHK5tGyel/NKX42wXvUWEr9eata9LraW6diEGscMeoG6tJV03sQ/BARbeWte2FXGopIYrX1nHyKe+4dpXN+Dvp/LodYYkRxAbFsjLKw9SXttkX2RlekY8YYF+ys0U6BykrrEF0myV+FkEv/hoO/VNVt5cm0uQv4XrJ3auliTQ38LI3lE0NNsYnxZDnEtM4exB8QT5W/hWc6lsP1ZBUVUjF7qpnD0VGNU7isMltVQ3NLMxR7nwpvTvXgFx7pBEhICH5+/E3yK4Z2bHrQdvCQ7wY3L/OD7dlkdTi637NPdTBf8g1c0ZICicyf1iiQjy5+Z23ImnAz0dpD49aKpRv31gQThl17guO2p3MTk0uw2Hy7xKU92ZV0lKVDCJbrTCmLBAfjS2Nwu25TulqbrylyX7OFBYze3T+/Gfm8ez+tfnOlV4WiyCqf3jyC2rIzjAwkxNcw8O8OPcoUl8m1VEi8Ug9JrrONHkR3CAhVduGsehklp++/luPtuex4/G9nZqY9FRxvVVbqbzDDUTOqGB/szISODbPYVIKfk68zgBfiqb51RELwzLzK9iY045cWGBbWZRdYa48CDGpcVQ12Tlxsl93d4n3cnZGfHUaQsWnXECQl9aFyAogvOGJrH9ifPbdSeeDpgCwht8ZEFsOFzGmN8vZe9xLWOllYvJOS20uKqB61/bwB1vbW7VcM01jVMPUHvijun9aGyx2btjurK/sJqlWUXcPr0fj104lLnDk90G86ZqmTWu7poLRyRTXtvEjmJNmGkxiKIGP6b0j+PcIUn8eFo6C7bl09Bs63Iq4MxBCQT6WzwWPM0ZnkRBZQOZ+VUsySxk+sB4n/ex6Sx6hs/u/Ao2Hi5nUr9Yn/js541OITLYn5/Oal1v0N3odQbhQf70i/Oxq7YnCIxw+u2uqPJ05Mx4F77GRwJi+d4iKuub+d3nmWqC1y0FuwWhafeagNiUW46UqhGc/RyUMJj1t5X86qMd2GyS8tomjpbXtSkgBiZGMHtoIm+vP9JqjVyAV1YeJDTQr1UWiCszByUQ6GfhinHOhWSzBicQGujH55la47SmOpoaailv8ufsDDVZPHLBEIb1iuS8IYluc+47wrSB8ex6co69fbQr5w1JxCLgH8sOkHei/pR1L4HS7ntHh/B1ZiH5FfXdHn/QuWVqXzb+ZrbbjqHdTUZiOMmRwQxLiWwzhfW0Rc9k8rDc6OnK6R1BOVn4yMW05cgJQgL82HLkBJ9uy+eqBN2C0GMQzhbE5pxyQgL8+PG0dF5eeYjByZEEB1h4emEWIYF+LNieT0Swv72a1TWDyZU7Z/Tn2lc3MH9rHjdNcfhLj5bVsXBnAT+Z0b9dt0+f2FC2PXE+4S4FXKGB/lwzoQ/zN2bzhwD1nhrra6gn1l7bEBzgx8L7ptHxigz3tFXsFBcexMT0WL7bV4yfRXD+sNauqFOJEb0j+UbrujrZTcVwdyCEICTw5CxsJITg5ZvGEXqSXu+kY3cx+ThOeZIxLQhvsFsQ3Scg6pusZOZXcsvUvoxLi+bPi/dSbdVcHvY0V82C0GITm3JPMK5vNA/NGczc4Uk8syiL336WydQBcax8aBZ3zujH/9Yf4ZlFWQiBU5Mzd0zqF8vo1Cj+uybHqXDule8P4W+x8BMPee2uuAoHndun9aPO5o9EQHMdtsZaLAGhTv50fz9LpzKXOsOc4cr9NKV/bJfiHSeDUZpwjwoJYPAp0tmzq4xLi+mypXjKEnhmWhCmgPAGH7iYduZV0GyVTEyP5ZnLR3Cirok3NmnFZa1iEEFU1jezr7CKiemxWCyC568Zw5xhSTw0ZxBv3jaRmLBAHrtwKFeM7c3h0loyEsM9Ttw6QgjuPLs/OaW1LNtbhJSSrIIqPt2ax9UTUrscuEyLC+XCESnUE0hjXTUWaz1xMdE9Vjg0d3hSq0V/TlX0QK7+fZuc4uhzg68zHU8ypovJG3zgYtKXVhzfN4aYsEBumZrOJ+u38EAQBheTHoMIYdvhUqR0dLAMC/Ln1VsmOF3TYhH89apRIFRTNm+4YHgyqTEh/PazTB5bsJvy2iaC/C3cfXb3pD3+ZEY/ag8EsS87nymyieR43/jTvSE1JpS1j55LfPipbT2ASk82ZoaZnOLYYxBnloVkCghvaOx+AbE5t5yBieF2V8ev5gxi3e790AwtjXXqi2mpV+tPWCxsyi0nwE8wto/nNW0BAvwsPH/NGK/H4e9n4eG5g/n394cZkRLJ2LQYpg+M77buk2PTYij2D6GivIQAPyu9E3tOQABdb6twkogODWTVr88hPuz0GO8PnjPUxWQKCG/o5hiEzSbZeuQElxj63kQGB/DIpeNgAWzcn8e0qTitR705p5wRvaN8ElS8bExvn7pdQsMiiWlSVdshoWfWA+RLTkZ2kUk3YQapf8A01agaBUv3TM4HiqupbmhhQl9nbfq8EaqSeNuhArKLqlUMwj+YhmYrO/MqnNbfPZ0IC4+kb3DH16M2MTlt0C2HMywGYQoIVw6vhFrnldK62qjv/Y1HueG1DVTWqaKxLbkq/jAh3cVdZLEg/YOJ8GvmkU93YdMExI5jjoD26YgIDKVPkL5Y0BlYJGViMuwyOPvXEHxmVYmbAsKItQXevRI2v+68vQsCYuHOAn77+W7WHSrjofk7kVKyJbechIggty2PRUAIM/qFs+1oBTtzipD+wfYmc60EyulCQKhajQ9MC8LkzCRhMJz72+5ZevgUwhQQRloawNbimMx0mmo7ZTquyS7lwY93MLFvLA/PHczSrCL+uyaHLUdOMKFvjPt0T/8Q+kdZePD8QZRWVHKkysqq7BKGJEcQHXrqZ9+4JSBEfa7g9XrUJiYmPY8ZpDZibVK/G1wWjW+q6bAFsSuvgrvf2cKAhHBeu3XC/7d399Fx3fWdx98fayRLcuLYTpwnO3HMYshD08TgZvPAcmgC1FBK2AUap01hs1ncspDStAsNPT00TZddtrSFUlI2D2Rh99AYcAvrti6BJlDakgQbEgJ2cDAmje3aiRw/xLakGc3ou3/c38jX8sge2boaa/R5naPjuXfmzvyurs989Hu4vx+zu0t8b+te/sff/ZDacBy2JOVhOntQdYBbr1vCc091s/35/ax7Zs/IOrxTUv535xqE2ZThGkRe/Q7mIwJifE1M39+2j5vue4y5s7r47H+6gtN6OpHER99+GefOyUam/MxYzUW5danP6g1ecs7pLJzb0/RKXyelfCg4IMymDNcg8uo3pjUKiNnNfUF/b+tebvr0Y5zW08mqlVcetgjLaT2d3PuOZaxev41Lzh2jMyu/LnV1kDmz5/NP7752vGdycsk3KxW96JKZTRgHRN7RAqKJPognt+3lpvseY86sTlatvIoFc478a/nCs2fzu0dbJL2z+/AV5UptMBbeTUxmU9Ixm5gk/YKk6dEUVasHxN7D91f2NxUQd6zZwCndJT4/Rjg0pbP38BXl2iEgDmticie12VTRzBf/DcCPJP2hpAuLLlBLHbUGcfSmkaef2893n93LLa9azLnHGw6QfZmOzOY6eNhqclNWPhRcgzCbMo4ZEBFxE7AU+DHwGUmPSFopqf3mTKgHROVAdk8EQLWSjW46Rg3igW8/23DhnHEr9Rw+m2tb1CByAXEc62ubWWs01XQUES8Cq4FVwDnAvwe+K+nWAss2+aq59ZnLaRnQoWPPwzQ4VONLj2/n9ZecxbwTXWegs+fw2VzbISC6UkCUemDG9GitNGsHzfRBvFnSl4BvAJ3AFRHxBuAy4LeKLd4kq+UCot4PkSbq2zHYwY59Aw0Pe3DDTvb2D3HjFeefeBk6e7LO6YhsNtd2CIh6DcLNS2ZTSjN/zr0V+FhEXBoRH42I5wEioh+4pdDSTbZ8DaLeD5EC4u5HnuOGux/lxcGhIw574NvPcv68Xq6aiKUh6zWIkdXk2ikg3EFtNpU0ExB3AN+ub0jqkXQBQEQ8VEipWqVhQGRrQWw9KJ7d3c8HvvgkEYeW5/zJroM8umU3N/zMeROz8ldnDxCHPr8d2uy7XIMwm4qaCYgvAsO57Vra137yTUwDhzcxHYxulp4/h69s2MlnvvUMALsOlPmjr26iY4Z4+ytPsHO6rv5X9kA24yulNlgwxk1MZlNSMwFRiohKfSM9bqonVtJySZskbZZ0e4Pnz5f0dUmPS3pS0htzz30wHbdJ0s8183kn7ChNTAeim99540W89qKz+O9rn+J9qx7n6o88zN8+uYObr77ghNdvHlH/Eq0HRDt8qdYDwndRm00pzQREn6Q31zckXQ/sOtZBkjqAu4A3ABcDN0oafQvx7wJfiIilwArgz9OxF6ftS4DlwJ+n9ytWo4BIy432082ieb388dsv46zZ3Ty4YSdve+VC/v43X330O6PHqzQqIFyDMLMWaWaqjV8DPifpk4CArcA7mjjuCmBzRGwBkLQKuB7YmHtNAPVVvk8D/jU9vh5YFRFl4CeSNqf3e6SJzz1+IwGhI/oghjp6mX/qTCTxt7/+77IC93ROfBlG1yDaqg/CndRmU8kxAyIifgxcKemUtH2gyfdeQBYmdduAfzvqNXcAX033U8wCXps79tFRxx6xaLKklcBKgPPPn4AhprUyoGxVqFFNTPPmHlq/oZBgqBvpg8gWCWqLGkQ95FyDMJtSmpqsT9LPkzX3dNe/JCPizgn4/BuBz0TEH0u6Cvi/kn6q2YMj4h7gHoBly5bFMV5+bPW5j3rmHBEQZ86bpOU+68Na26kPYsaMLCRcgzCbUo4ZEJL+F9AL/CxwH/A2csNej2I7cF5ue2Hal3cLWR8DEfGIpG7gjCaPnXjVCpS6oHvOyI1yUTlAOTpZeMYkzSxSD4T+eg2iDe6DALjgVXDu0laXwszGoZlO6qsj4h3Anoj4feAq4GVNHLcOWCJpsaQusk7nNaNe8yxwHYCki4BuoC+9boWkmZIWA0toLpROTC1NbZFrYho8+CIH6W64fnQhjhjm2iYBcdNqWHZzq0thZuPQTBNTmlqUfknnAi+Qzcd0VBFRlfRe4EGgA7g/IjZIuhNYHxFryKbquFfSbWQd1v8xsrvQNkj6AlmHdhV4T0TUxnty41YtQ8fMLCB2PQdA/4EXORiTGBCl0U1MbRIQZjblNBMQfy1pDvBR4LtkX+T3NvPmEbEWWDtq34dyjzcC14xx7IeBDzfzOROmWk5NTIdqEOV+1yDMbHo6akCkhYIeioi9wF9K+hugOyL2He24Kat6ZBNTdWA//XRz8dzJCoj6MNd0J7cDwsxa5Kh9EBExTHazW3273LbhAFkfREfqpB7qh2qFKB9gqKOXnq7i79MDGtwH4YAws9ZoppP6IUlvVX18azur1yB65mTbg/vQ0EE0mVNEdHTCjFK2zCk4IMysZZoJiF8lm5yvLOlFSfslvVhwuVoj3wcBMLiPUrWfGd3HXo96QtX7ITq6vMCOmbVMM3dSt9/SomOpDkLvvJGAKB/YzcwYoKtnkn8FnT3ZinbtMM2GmU1Zzdwo9+pG+yPimxNfnBarVVIfRBYQL7zQxzwG6Zk1+xgHTrB6s5KHuJpZCzUzzPX9ucfdZJPmfQe4tpAStVJ9qo0UEHt27eBcDTFr9pzJLUe9iakd5mEysymrmSamX8hvSzoP+HhhJWqlkak2soAYfGEbALMnPSBS05KbmMyshY6nB3QbcNFEF+SkMDLVRhYItb3Z9E+nnOoahJlNP830QfwZ2d3TkAXK5WR3VLef+lQbnT0wo5PSgWx5Cs2c7FFM9T4I1yDMrHWa6YNYn3tcBR6IiH8uqDytVS1nf7UrWxNiVjmbj2nSl8ocaWJyDcLMWqeZgFgNDNYny5PUIak3IvqLLdokGx6G4aGRL+XoPo15B/uyNfQmPSDqTUyuQZhZ6zR1JzWQ/6bqAf6+mOK0UC0tN9rRBUCldCqnk2YVmeyA8DBXMzsJNBMQ3fllRtPj9lsarJpmNU9fzvt1CjOUul66WnQntafZMLMWaiYgDkp6RX1D0iuBgeKK1CLVSvZvKatB7K7lvpwnPSDqfRAOCDNrnWb6IH4D+KKkfyVrkT8buKHQUrXCqBrEc5XuQ8vmtayT2gFhZq3TzI1y6yRdCLw87doUEUPFFqsFaqkG0ZF1Um8b6Dr0XKsCwn0QZtZCx2xikvQeYFZE/CAifgCcIum/FF+0STZSg5jJQKXG1oHObLtjZjYF92RyDcLMTgLN9EG8K60oB0BE7AHeVVyRWmSkD2ImTz+3nxcjdRRPdu0B3EltZieFZgKiI79YkKQOoOsor5+a6sNcSzPZtHM/+yIFw2R3UMOhYHBAmFkLNdNJ/RXg85LuTtu/CvxdcUVqkXoTU8dMntr5IoMdaQ2IVtYg3AdhZi3UTED8NrAS+LW0/STZSKb2kmti2rRzPwtOnw97aVFAeDZXM2u9YzYxRcQw8BjwDNlaENcCTxVbrBZINYjo6OKHO/dz9vwzs/2TPVEfeC4mMzspjFmDkPQy4Mb0swv4PEBE/OzkFG2SpWGuuysz2H2wwoJzzoYf0Zo+iDmLoPd0OGPJ5H+2mVlytCamHwL/CLwpIjYDSLptUkrVCqkGsfmF7BaPCxYuyPa3oonp1LPgA1sm/3PNzHKO1sT0H4AdwNcl3SvpOrI7qZsmabmkTZI2S7q9wfMfk/RE+nla0t7cc7Xcc2vG87nHpZqNYno6BcTLF5yRjSJqRUCYmZ0ExqxBRMSXgS9LmgVcTzblxpmSPgV8KSK+erQ3TsNh7wJeR7YK3TpJayJiY+4zbsu9/lZgae4tBiLi8uM4p+OTAuKpvkHOmj2TubO64FW3waKrJ60IZmYnk2Y6qQ9GxF+ktakXAo+TjWw6liuAzRGxJSIqwCqyoBnLjcADTbxvMdJ9EBufL/Pys2dn+15zOyx+dcuKZGbWSuNakzoi9kTEPRFxXRMvXwBszW1vS/uOIGkRsBh4OLe7W9J6SY9KessYx61Mr1nf19fX5FmMIQ1z3dhX5sKzTz2x9zIzawPjCogCrQBW11etSxZFxDLgl4CPS/o3ow9KYbUsIpbNnz//xEpQHSQ6uqhUg9Nntd+N4mZm41VkQGwHzsttL0z7GlnBqOaliNie/t0CfIPD+ycmXq1CpJlce7o6Cv0oM7OpoMiAWAcskbRYUhdZCBwxGilNJT4XeCS3b66kmenxGcA1wMbRx06oVIMA6Ol0QJiZNTPVxnGJiKqk9wIPAh3A/RGxQdKdwPqIqIfFCmBVRETu8IuAuyUNk4XYR/KjnwpRrTCcahC9XYX9WszMpoxCvwkjYi2wdtS+D43avqPBcd8CLi2ybEeoDlKbkdUget3EZGZ20nRSt16tPBIQ3W5iMjNzQIyolqnJNQgzszoHRF21TNUBYWY2wgFRVy0zpGztaQ9zNTNzQBxSKzMkD3M1M6tzQNRVK1TSoC4PczUzc0AcUh2kQicSdHf612Jm5m/CulqFcnTS09mBNK5lL8zM2pIDoq46SJmS+x/MzBIHRF21wuBwp0cwmZklDoi66iADUfI9EGZmiQMCIAJqZfqHS/R4BJOZGeCAyNSy1eQGh0v0eASTmRnggMhUs/WoDw53+B4IM7PEAQGHAqLW4U5qM7PEAQFQqwdEiV4PczUzAxwQmVSDOOAahJnZCAcEjATE/qoDwsyszgEBI01M/bUOejvdSW1mBg6ITKpBlOnyjXJmZokDAkYCohIluh0QZmaAAyJTDwg6PYrJzCxxQMBIH0SZTjcxmZklDgjI1SBKHsVkZpYUGhCSlkvaJGmzpNsbPP8xSU+kn6cl7c09905JP0o/7yyynPlOaq8HYWaWKWxMp6QO4C7gdcA2YJ2kNRGxsf6aiLgt9/pbgaXp8Tzg94BlQADfScfuKaSw1UEAylHyXExmZkmRNYgrgM0RsSUiKsAq4PqjvP5G4IH0+OeAr0XE7hQKXwOWF1bSNJtrGS8YZGZWV2RALAC25ra3pX1HkLQIWAw8PJ5jJa2UtF7S+r6+vuMvaX4UkwPCzAw4eTqpVwCrI6I2noMi4p6IWBYRy+bPn3/8n54LCPdBmJlligyI7cB5ue2FaV8jKzjUvDTeY09crUxNHQwzw01MZmZJkQGxDlgiabGkLrIQWDP6RZIuBOYCj+R2Pwi8XtJcSXOB16d9xaiWqWkmMwQzSydLpcrMrLUKG7ITEVVJ7yX7Yu8A7o+IDZLuBNZHRD0sVgCrIiJyx+6W9AdkIQNwZ0TsLqqsVMtU1UlvVwlJhX2MmdlUUuiYzohYC6wdte9Do7bvGOPY+4H7CytcXnWQoRlddLv/wcxshNtTAGoVhjyCyczsMA4IgOqgh7iamY3igACoVqjQ6SYmM7McBwRArUyZkmsQZmY5DgiAapnBcBOTmVmeAwKgWqYcJXo8UZ+Z2QgHBEC1zMBwiZ5O/zrMzOr8jQhQKzPgqb7NzA7jgACoDtI/7NXkzMzyHBBAVCsMDpfo9TBXM7MRDgggqoPZcqOuQZiZjXBAAKqWqeAmJjOzPAcEpBvlfB+EmVmeA6JWRTFMJTrp6fQoJjOzOgdELVtutEynm5jMzHIcELn1qN3EZGZ2iANiRol/eemvsHF4ET0e5mpmNsIB0T2bxy/5IOviQtcgzMxyHBBAf6UG4D4IM7McBwQwMJQFRK9HMZmZjXBAAAOVKuAahJlZngOCrImpNEN0lfzrMDOr8zciWUB4BJOZ2eEcEMDgUM3NS2ZmozggyGoQHuJqZna4QgNC0nJJmyRtlnT7GK/5RUkbJW2Q9Be5/TVJT6SfNUWWs79S83rUZmajFPatKKkDuAt4HbANWCdpTURszL1mCfBB4JqI2CPpzNxbDETE5UWVL29gqOr1qM3MRinyW/EKYHNEbImICrAKuH7Ua94F3BURewAi4vkCyzOmgUrN61GbmY1SZEAsALbmtrelfXkvA14m6Z8lPSppee65bknr0/63NPoASSvTa9b39fUdd0GzJib3QZiZ5bX6z+YSsAR4DbAQ+KakSyNiL7AoIrZLegnwsKTvR8SP8wdHxD3APQDLli2L4y3EwJA7qc3MRiuyBrEdOC+3vTDty9sGrImIoYj4CfA0WWAQEdvTv1uAbwBLiyqo74MwMztSkQGxDlgiabGkLmAFMHo00pfJag9IOoOsyWmLpLmSZub2XwNspCCDbmIyMztCYU1MEVGV9F7gQaADuD8iNki6E1gfEWvSc6+XtBGoAe+PiBckXQ3cLWmYLMQ+kh/9NMHlpN9NTGZmRyi0DyIi1gJrR+37UO5xAL+ZfvKv+RZwaZFlq6vUhqkNh0cxmZmNMu0H/w+ktSC63QdhZnaYaR8QQrzpp8/hpWee0uqimJmdVKZ9u8ppvZ188pde0epimJmddKZ9DcLMzBpzQJiZWUMOCDMza8gBYWZmDTmCtzJSAAAFlElEQVQgzMysIQeEmZk15IAwM7OGHBBmZtaQsumQpj5JfcC/nMBbnAHsmqDiTBXT8Zxhep73dDxnmJ7nPd5zXhQR8xs90TYBcaIkrY+IZa0ux2SajucM0/O8p+M5w/Q874k8ZzcxmZlZQw4IMzNryAFxyD2tLkALTMdzhul53tPxnGF6nveEnbP7IMzMrCHXIMzMrCEHhJmZNTTtA0LSckmbJG2WdHury1MUSedJ+rqkjZI2SHpf2j9P0tck/Sj9O7fVZZ1okjokPS7pb9L2YkmPpWv+eUldrS7jRJM0R9JqST+U9JSkq9r9Wku6Lf3f/oGkByR1t+O1lnS/pOcl/SC3r+G1VeYT6fyflDSu1dGmdUBI6gDuAt4AXAzcKOni1paqMFXgtyLiYuBK4D3pXG8HHoqIJcBDabvdvA94Krf9P4GPRcRLgT3ALS0pVbH+FPhKRFwIXEZ2/m17rSUtAH4dWBYRPwV0ACtoz2v9GWD5qH1jXds3AEvSz0rgU+P5oGkdEMAVwOaI2BIRFWAVcH2Ly1SIiNgREd9Nj/eTfWEsIDvfz6aXfRZ4S2tKWAxJC4GfB+5L2wKuBVanl7TjOZ8GvBr4NEBEVCJiL21+rcmWUO6RVAJ6gR204bWOiG8Cu0ftHuvaXg/8n8g8CsyRdE6znzXdA2IBsDW3vS3ta2uSLgCWAo8BZ0XEjvTUTuCsFhWrKB8HPgAMp+3Tgb0RUU3b7XjNFwN9wP9OTWv3SZpFG1/riNgO/BHwLFkw7AO+Q/tf67qxru0JfcdN94CYdiSdAvwl8BsR8WL+ucjGPLfNuGdJbwKej4jvtLosk6wEvAL4VEQsBQ4yqjmpDa/1XLK/lhcD5wKzOLIZZlqYyGs73QNiO3Bebnth2teWJHWShcPnIuKv0u7n6lXO9O/zrSpfAa4B3izpGbLmw2vJ2ubnpGYIaM9rvg3YFhGPpe3VZIHRztf6tcBPIqIvIoaAvyK7/u1+revGurYn9B033QNiHbAkjXToIuvUWtPiMhUitb1/GngqIv4k99Qa4J3p8TuB/zfZZStKRHwwIhZGxAVk1/bhiPhl4OvA29LL2uqcASJiJ7BV0svTruuAjbTxtSZrWrpSUm/6v14/57a+1jljXds1wDvSaKYrgX25pqhjmvZ3Ukt6I1k7dQdwf0R8uMVFKoSkVwH/CHyfQ+3xv0PWD/EF4Hyy6dJ/MSJGd4BNeZJeA/zXiHiTpJeQ1SjmAY8DN0VEuZXlm2iSLifrmO8CtgA3k/1B2LbXWtLvAzeQjdh7HPjPZO3tbXWtJT0AvIZsWu/ngN8DvkyDa5vC8pNkzW39wM0Rsb7pz5ruAWFmZo1N9yYmMzMbgwPCzMwackCYmVlDDggzM2vIAWFmZg05IMzGQVJN0hO5nwmb8E7SBfkZOs1arXTsl5hZzkBEXN7qQphNBtcgzCaApGck/aGk70v6tqSXpv0XSHo4zcX/kKTz0/6zJH1J0vfSz9XprTok3ZvWNfiqpJ6WnZRNew4Is/HpGdXEdEPuuX0RcSnZnasfT/v+DPhsRPw08DngE2n/J4B/iIjLyOZJ2pD2LwHuiohLgL3AWws+H7Mx+U5qs3GQdCAiTmmw/xng2ojYkiZF3BkRp0vaBZwTEUNp/46IOENSH7AwP+1Dmob9a2nRFyT9NtAZEf+t+DMzO5JrEGYTJ8Z4PB75eYJquJ/QWsgBYTZxbsj9+0h6/C2ymWQBfplswkTIloV8N4ysmX3aZBXSrFn+68RsfHokPZHb/kpE1Ie6zpX0JFkt4Ma071ayld3eT7bK281p//uAeyTdQlZTeDfZSmhmJw33QZhNgNQHsSwidrW6LGYTxU1MZmbWkGsQZmbWkGsQZmbWkAPCzMwackCYmVlDDggzM2vIAWFmZg39f3R/E+jWStQ1AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}