FunctionalAPI.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Week5-KhusheeKapoor-200968052-DeepLearningLab.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU",
    "gpuClass": "standard"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# Week 5 - Deep Learning Lab\n",
        "\n",
        "Author: Khushee Kapoor\n",
        "\n",
        "Registration Number: 200968052\n",
        "\n",
        "Last Updated: 24/8/22"
      ],
      "metadata": {
        "id": "wlLJFXxEVCBG"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Setting Up"
      ],
      "metadata": {
        "id": "_QDjlG5bqahP"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "To begin, we import the following libraries:\n",
        "\n",
        "- NumPy: for data manipulation\n",
        "- Pandas: for data manipulation\n",
        "- Matplotlib: for data visualization\n",
        "- Seaborn: for data visualization\n",
        "- Tensorflow: to build neural networks"
      ],
      "metadata": {
        "id": "_3HDEVY7VLkd"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "q_ajPTJCAtRA"
      },
      "outputs": [],
      "source": [
        "# importing the libraries\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import tensorflow as tf"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Next, we fetch the data using the fetch_california_housing() function from the sklearn library. After that, we split the independent variables (x) and dependent variable (y) using the 'data' and 'target' keys in the data dictionary."
      ],
      "metadata": {
        "id": "73rS3BVSVQWN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# fetching the data\n",
        "from sklearn.datasets import fetch_california_housing\n",
        "df = fetch_california_housing()\n",
        "\n",
        "# splitting the independent and dependent variables\n",
        "x = df['data']\n",
        "y = df['target']"
      ],
      "metadata": {
        "id": "X2_LRib0BJQO"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Next, we split x and y variables into training and testing sets using the train_test_split() function from the sklearn library and fixing the random state for consistent results."
      ],
      "metadata": {
        "id": "Wtm1loMWXRq3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# splitting the dataset into training and testing sets\n",
        "from sklearn.model_selection import train_test_split\n",
        "x_train, x_test, y_train, y_test = train_test_split(x, y, train_size=0.85, random_state=101)"
      ],
      "metadata": {
        "id": "KdW8isbJEqeI"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Sequential API"
      ],
      "metadata": {
        "id": "bd8pDXDcqc1Y"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now, we build our sequential model using the Sequential API from the sklearn library. We add one hidden layer with 30 units and activation function 'relu'. Then we add the output layer with 1 unit and activation function 'linear' for linear regression. We also add a Batch Normalization layer at the very start to optimize the model."
      ],
      "metadata": {
        "id": "knBjkx27XusU"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# building the network\n",
        "sequential = tf.keras.Sequential()\n",
        "sequential.add(tf.keras.layers.Input(shape=[8]))\n",
        "sequential.add(tf.keras.layers.BatchNormalization())\n",
        "sequential.add(tf.keras.layers.Dense(30, activation='relu'))\n",
        "sequential.add(tf.keras.layers.Dense(1, activation='linear'))\n",
        "\n",
        "# printing the network summary\n",
        "sequential.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4oX8SDn-DNQn",
        "outputId": "b767fd6f-ba56-4bd8-97d2-b509d90891fa"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " batch_normalization (BatchN  (None, 8)                32        \n",
            " ormalization)                                                   \n",
            "                                                                 \n",
            " dense (Dense)               (None, 30)                270       \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 1)                 31        \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 333\n",
            "Trainable params: 317\n",
            "Non-trainable params: 16\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "As we can see, there are 333 parameters out of which 317 are trainable and 16 are non trainable."
      ],
      "metadata": {
        "id": "FUMyUGGfYRhs"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "To obtain a pictorial representation of the network, we use the plot_model() function from the Tensorflow library."
      ],
      "metadata": {
        "id": "YzwSkwxrYZu8"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# obtaining a pictorial representation of the network\n",
        "from tensorflow.keras.utils import plot_model\n",
        "plot_model(sequential,  show_shapes=True, show_layer_names=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 422
        },
        "id": "btX0-rwMLLaG",
        "outputId": "31cdecd4-193f-427e-a16b-dd53a6f14ca1"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Next compile the model and set the optimizer to be 'adam', loss to be MeanSquaredError and metric to be RootMeanSquaredError. After that, we train the model on the training data with number of epochs set to 20 and validation split to be 12.75%."
      ],
      "metadata": {
        "id": "V5hADNl4Yg26"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# compiling the model\n",
        "sequential.compile(optimizer=tf.keras.optimizers.Adam(), loss=tf.keras.losses.MeanSquaredError(), metrics=[tf.keras.metrics.RootMeanSquaredError()])\n",
        "\n",
        "# training the model\n",
        "hist = sequential.fit(x_train, y_train, batch_size=32, epochs=20, validation_split=0.1275)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4Q6PJVHKDVO3",
        "outputId": "771afcd0-8bac-4179-b7ab-3043acbdf0d9"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "479/479 [==============================] - 3s 4ms/step - loss: 1.1868 - root_mean_squared_error: 1.0894 - val_loss: 0.5794 - val_root_mean_squared_error: 0.7612\n",
            "Epoch 2/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4698 - root_mean_squared_error: 0.6854 - val_loss: 0.5152 - val_root_mean_squared_error: 0.7178\n",
            "Epoch 3/20\n",
            "479/479 [==============================] - 3s 6ms/step - loss: 0.4441 - root_mean_squared_error: 0.6664 - val_loss: 0.5174 - val_root_mean_squared_error: 0.7193\n",
            "Epoch 4/20\n",
            "479/479 [==============================] - 3s 6ms/step - loss: 0.4308 - root_mean_squared_error: 0.6564 - val_loss: 0.4729 - val_root_mean_squared_error: 0.6877\n",
            "Epoch 5/20\n",
            "479/479 [==============================] - 3s 6ms/step - loss: 0.4302 - root_mean_squared_error: 0.6559 - val_loss: 0.4677 - val_root_mean_squared_error: 0.6839\n",
            "Epoch 6/20\n",
            "479/479 [==============================] - 3s 6ms/step - loss: 0.4284 - root_mean_squared_error: 0.6545 - val_loss: 0.4704 - val_root_mean_squared_error: 0.6858\n",
            "Epoch 7/20\n",
            "479/479 [==============================] - 3s 6ms/step - loss: 0.4226 - root_mean_squared_error: 0.6501 - val_loss: 0.4308 - val_root_mean_squared_error: 0.6563\n",
            "Epoch 8/20\n",
            "479/479 [==============================] - 3s 6ms/step - loss: 0.4248 - root_mean_squared_error: 0.6517 - val_loss: 0.4707 - val_root_mean_squared_error: 0.6861\n",
            "Epoch 9/20\n",
            "479/479 [==============================] - 2s 3ms/step - loss: 0.4178 - root_mean_squared_error: 0.6464 - val_loss: 0.4789 - val_root_mean_squared_error: 0.6920\n",
            "Epoch 10/20\n",
            "479/479 [==============================] - 1s 3ms/step - loss: 0.4151 - root_mean_squared_error: 0.6443 - val_loss: 0.4656 - val_root_mean_squared_error: 0.6823\n",
            "Epoch 11/20\n",
            "479/479 [==============================] - 2s 3ms/step - loss: 0.4091 - root_mean_squared_error: 0.6396 - val_loss: 0.4693 - val_root_mean_squared_error: 0.6851\n",
            "Epoch 12/20\n",
            "479/479 [==============================] - 2s 3ms/step - loss: 0.4126 - root_mean_squared_error: 0.6423 - val_loss: 0.4592 - val_root_mean_squared_error: 0.6776\n",
            "Epoch 13/20\n",
            "479/479 [==============================] - 2s 3ms/step - loss: 0.4112 - root_mean_squared_error: 0.6412 - val_loss: 0.4446 - val_root_mean_squared_error: 0.6668\n",
            "Epoch 14/20\n",
            "479/479 [==============================] - 2s 3ms/step - loss: 0.4084 - root_mean_squared_error: 0.6391 - val_loss: 0.4612 - val_root_mean_squared_error: 0.6791\n",
            "Epoch 15/20\n",
            "479/479 [==============================] - 2s 3ms/step - loss: 0.4041 - root_mean_squared_error: 0.6357 - val_loss: 0.4535 - val_root_mean_squared_error: 0.6734\n",
            "Epoch 16/20\n",
            "479/479 [==============================] - 1s 3ms/step - loss: 0.4037 - root_mean_squared_error: 0.6353 - val_loss: 0.4473 - val_root_mean_squared_error: 0.6688\n",
            "Epoch 17/20\n",
            "479/479 [==============================] - 2s 3ms/step - loss: 0.4037 - root_mean_squared_error: 0.6354 - val_loss: 0.4770 - val_root_mean_squared_error: 0.6906\n",
            "Epoch 18/20\n",
            "479/479 [==============================] - 2s 3ms/step - loss: 0.4015 - root_mean_squared_error: 0.6337 - val_loss: 0.4443 - val_root_mean_squared_error: 0.6665\n",
            "Epoch 19/20\n",
            "479/479 [==============================] - 1s 3ms/step - loss: 0.3986 - root_mean_squared_error: 0.6314 - val_loss: 0.4567 - val_root_mean_squared_error: 0.6758\n",
            "Epoch 20/20\n",
            "479/479 [==============================] - 1s 3ms/step - loss: 0.3962 - root_mean_squared_error: 0.6295 - val_loss: 0.4499 - val_root_mean_squared_error: 0.6708\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We then compare the training and validation losses (mse). To do this, we use the history attribute and extract the losses (mse) of the models. Next, we plot it against the number of epochs using functions from the MatPlotLib library."
      ],
      "metadata": {
        "id": "uggKTPf6nEvF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# extracting the model loss against epochs\n",
        "tl = hist.history['loss']\n",
        "vl = hist.history['val_loss']\n",
        "eps = range(1, 21)\n",
        "\n",
        "# plotting the train and validation loss\n",
        "plt.plot(eps, tl, 'r', label='Training MSE')\n",
        "plt.plot(eps, vl, 'b', label='Validation MSE')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('MSE')\n",
        "plt.legend()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 296
        },
        "id": "d22HC5v9FFEu",
        "outputId": "cea3f7dc-3943-418d-ee39-a252fb99c310"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f73300c3d50>"
            ]
          },
          "metadata": {},
          "execution_count": 7
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "As we can see, the loss, that is the mean squared error decreases sharply and then plateaus in the case of training, and remains more or loss constant with some fluctuations in the case of validation. There is some difference between the two."
      ],
      "metadata": {
        "id": "8gKB4QK0nLTT"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Next, we compare the training and validation rmse. To do this, we use the history attribute and extract the rmse of the models. Next, we plot it against the number of epochs using functions from the MatPlotLib library."
      ],
      "metadata": {
        "id": "oJxWiUL_naUS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# extracting the model rmse against epochs\n",
        "ta = hist.history['root_mean_squared_error']\n",
        "va = hist.history['val_root_mean_squared_error']\n",
        "eps = range(1, 21)\n",
        "\n",
        "# plotting the train and validation rmse\n",
        "plt.plot(eps, ta, 'r', label='Training RMSE')\n",
        "plt.plot(eps, va, 'b', label='Validation RMSE')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('RMSE')\n",
        "plt.legend()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 296
        },
        "id": "jIfu3Yl1FdJ4",
        "outputId": "8bcc621b-9b6d-42e8-9cd4-1b6803fcbfc1"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f72c4152d50>"
            ]
          },
          "metadata": {},
          "execution_count": 8
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "As we can see the rmse decreases sharply and then plateaus in the case of training, and remains more or loss constant with some fluctuations in the case of validation. There is some difference between the two."
      ],
      "metadata": {
        "id": "thdUG_4_njgn"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "To evaluate the model accuracy, we first obtain predictions on the test set using the predict() function. Then, we calculate the root mean squared error using the sqrt() function from the numpy library and the mean_squared_error() function from the sklearn library. To easily interpret the rmse, we calculate the percentage variation by dividing the rmse by the mean of the true values."
      ],
      "metadata": {
        "id": "qMdKrcURptAm"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# obtaining the predictions\n",
        "predictions = sequential.predict(x_test)\n",
        "\n",
        "# calculating the rmse\n",
        "from sklearn.metrics import mean_squared_error\n",
        "rmse = np.sqrt(mean_squared_error(y_test, predictions))\n",
        "print(str.format('Overall RMSE: {:.2f}', rmse))\n",
        "\n",
        "# calculating the variation\n",
        "variation = rmse/np.mean(y_test)*100\n",
        "print(str.format('Variation: {:.2f}%', variation))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4EcliaMlIRgQ",
        "outputId": "8112dba9-b31d-406c-d85a-a4239352d9ca"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Overall RMSE: 0.70\n",
            "Variation: 33.51%\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "As we can see, the rmse is 0.70, which means that it the error is around 33.51% from the mean value. This means that the model performs moderately."
      ],
      "metadata": {
        "id": "Ots3w9GGqM_7"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Functional API"
      ],
      "metadata": {
        "id": "_wqflWOgqfvM"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![image.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "iKyP6ZdTqhTM"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "We build the above model using the functional API. To do that, we follow the steps:\n",
        "\n",
        "- create two input layers using the Input() function from the keras library\n",
        "- link both the input layers to batch normalization layers to normalize the data in turn optimizing the model\n",
        "- link the second normalized input layer to two consecutive hidden layers having 30 neurons each and activation function 'relu'\n",
        "- concatenate the first normalized input layer with the second hidden layer using the Concatenate() function from the keras library\n",
        "- link the merged layer with an output layer having one neuron and activation function 'linear' (since this is a regression model)\n",
        "- connect all the layers together using the Model() function from the keras library"
      ],
      "metadata": {
        "id": "zieiOdGPqlZX"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# creating input layer 1 and normalizing\n",
        "inputA = tf.keras.layers.Input(shape=[5], name='inputA')\n",
        "normalizeA = tf.keras.layers.BatchNormalization()(inputA)\n",
        "\n",
        "# creating input layer 2 and normalizing\n",
        "inputB = tf.keras.layers.Input(shape=[6], name='inputB')\n",
        "normalizeB = tf.keras.layers.BatchNormalization()(inputB)\n",
        "\n",
        "# linking normalized input layer 2 to a deep network\n",
        "hidden1 = tf.keras.layers.Dense(30, activation='relu', name='hidden1')(normalizeB)\n",
        "hidden2 = tf.keras.layers.Dense(30, activation='relu', name='hidden2')(hidden1)\n",
        "\n",
        "# merging normalized input layer 1 and the hidden layer 2\n",
        "concat = tf.keras.layers.Concatenate(name='concat')([hidden2, normalizeA])\n",
        "\n",
        "# linking the merged layer to the output layer\n",
        "output = tf.keras.layers.Dense(1, activation='linear', name='output')(concat)\n",
        "\n",
        "# connecting all the layers in the model\n",
        "functional = tf.keras.models.Model(inputs=[inputA, inputB], outputs=output)\n",
        "\n",
        "# printing the network summary\n",
        "functional.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "D20_JWJ_Fxj3",
        "outputId": "7ed7ba01-4636-4f0e-fc15-2fb3da90e607"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                   Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            " inputB (InputLayer)            [(None, 6)]          0           []                               \n",
            "                                                                                                  \n",
            " batch_normalization_2 (BatchNo  (None, 6)           24          ['inputB[0][0]']                 \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " hidden1 (Dense)                (None, 30)           210         ['batch_normalization_2[0][0]']  \n",
            "                                                                                                  \n",
            " inputA (InputLayer)            [(None, 5)]          0           []                               \n",
            "                                                                                                  \n",
            " hidden2 (Dense)                (None, 30)           930         ['hidden1[0][0]']                \n",
            "                                                                                                  \n",
            " batch_normalization_1 (BatchNo  (None, 5)           20          ['inputA[0][0]']                 \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " concat (Concatenate)           (None, 35)           0           ['hidden2[0][0]',                \n",
            "                                                                  'batch_normalization_1[0][0]']  \n",
            "                                                                                                  \n",
            " output (Dense)                 (None, 1)            36          ['concat[0][0]']                 \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 1,220\n",
            "Trainable params: 1,198\n",
            "Non-trainable params: 22\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "As we can see, there are 1220 parameters out of which 1198 are trainable and 22 are non trainable."
      ],
      "metadata": {
        "id": "Iq2NP2JcsK_t"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "To obtain a pictorial representation of the network, we use the plot_model() function from the Tensorflow library."
      ],
      "metadata": {
        "id": "USuvCRL3sSp4"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# obtaining a pictorial representation of the network\n",
        "from tensorflow.keras.utils import plot_model\n",
        "plot_model(functional,  show_shapes=True, show_layer_names=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 644
        },
        "id": "kOj3QmBqK1SO",
        "outputId": "812d3480-74a3-49d1-ede8-a489b5d52733"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA50AAAJzCAYAAACMFXcMAAAABmJLR0QA/wD/AP+gvaeTAAAgAElEQVR4nOzde1iVZb7/8c9CDouFnDyzRTTBszhmcqX8cspd03jYmgcITJvRZnd5qBDD0kzzmGl5qWNKjWXsa7QUSEebyGpqxl3tjKnM0dFyGx7wVAjKQYREvH9/uFnjCpSDLBYL36/rWn94P/fz3N/ne3P6+hxuizHGCAAAAAAAJ/BwdQAAAAAAgKaLohMAAAAA4DQUnQAAAAAAp6HoBAAAAAA4jaerA7ie3bt3a+XKla4OAwDgptLT010dAgAAUCO+0nnixAm9/fbbrg4DNfDFF1/oiy++cHUYbuXkyZN8fbsR5su9MF8AADQulsa6ZEpaWpri4uLUSMPDNWJjYyVxVaE2+Pp2L8yXe2G+AABoXBrtlU4AAAAAgPuj6AQAAAAAOA1FJwAAAADAaSg6AQAAAABOQ9EJAAAAAHAais5rvPfeewoMDNSf//xnV4dSydatW9W5c2dZLBaHj6enp1q1aqX77rtP27Ztc3WYN6Ux5x8AAABA3VB0XqMxv15/7NixOnLkiMLDwxUYGChjjIwxOnv2rFJTU3Xq1CmNHTtWqamprg61zhpz/gEAAADUDUXnNYYPH66CggKNGDHC1aGopKRE0dHR1fYLDg7Wvffeq9///veSrq5P567cMf8AAAAAboyis5HasGGDcnJyaty/U6dOkqT8/HwnRXRrqW3+AQAAAFSNovP/fPbZZwoLC5PFYtHatWslScnJyfLz85PNZtOOHTs0dOhQBQQEKDQ0VJs3b7bvu2bNGlmtVrVp00ZTpkxRSEiIrFaroqOjlZmZae+XkJAgb29vtWvXzt722GOPyc/PTxaLRbm5uZKkxMREJSUlKSsrSxaLRREREdXGv2/fPknS3XffXS/5aGjukP/3339fAQEBev755xsiJQAAAECTQNH5f+666y59/vnnDm3Tpk3TjBkzVFJSIn9/f6WmpiorK0udO3fWo48+qrKyMklXi5mJEyfq4sWLmj59uo4dO6Y9e/bo8uXL+tWvfqUTJ05IulocPfjggw5jrFu3TgsXLnRoW716tUaMGKHw8HAZY/T9999fN+6SkhK9//77mjlzpu6//34lJSXVRzoanDvkv7y8XJJ05coVp+QAAAAAaIooOmsoOjpaAQEBat26teLj41VcXKzs7GyHPp6enurRo4d8fHzUs2dPJScnq6ioSCkpKfUaS0FBgf3ttTabzX4FcPz48fLy8qrXsRqLxpD/4cOHq7CwUPPmzauX4wEAAAC3AorOOvD29pYk+5W26+nfv79sNpu+++67eh3/2rfXlpWV6eTJk5oxY4YSEhLUp08f+22iTZWr8w8AAACg5ig6nczHx0dnz5512vE9PT3Vvn17TZo0SStWrNChQ4f0wgsvOG08d+Ps/AMAAAC4MYpOJyorK1N+fr5CQ0MbZLzIyEhJ0sGDBxtkvMauofMPAAAAoDKKTifatWuXjDEaMGCAvc3T07Pa20Lr6uuvv5YkdevWzSnHdzcNnX8AAAAAlVF01qMrV67o/Pnzunz5svbt26fExESFhYVp4sSJ9j4RERE6d+6ctm/frrKyMp09e1bHjx+vdKwWLVro9OnTOnbsmIqKiioVSiUlJbpy5YqMMTp9+rRSUlI0d+5ctWrVSjNmzHD2qTZKzs7/zp07WTIFAAAAqCWKzv+zdu1aRUVFSZJmzZqlBx54QMnJyVq1apUkqU+fPjpy5Ihee+01+7IkQ4YM0eHDh+3HKC0tVWRkpHx9fTVo0CB17dpVf/vb3+Tj42PvM23aNA0ePFjjxo1Tt27dtHjxYvn6+kqSBg4caF/eY+rUqWrTpo169uypYcOGacOGDYqIiFBWVpYKCgpks9nUrFkzeXh4qHPnznrhhRc0btw47dmzRx07dmyQnNWnxp7/c+fONUgeAAAAgKbGYowxrg6iKmlpaYqLi1MjDa+SKVOmKD09XXl5ea4OpcHFxsZKktLT010Wg7vl392+vm91zJd7Yb4AAGhcuNJZj8rLy10dwi2N/AMAAACND0UnAAAAAMBpKDrrwZw5c5SSkqKCggLddtttevvtt10d0i3lVsn/lClTZLFY7J8JEyZU6vPRRx/pmWee0datW9W5c2d734cffrhS3/vvv1/+/v5q1qyZevXqpT179jTEadyUsrIyLV26VBEREfL29lZQUJB69+6tY8eOSZLeeecdLV++vNJV7+3btzvkrlWrVk6Plflyr/kCAADOQ9FZD5YuXaqffvpJxhgdPXpUMTExrg7plnIr5b9FixbauXOnDh06pA0bNjhsmz9/vtasWaM5c+Zo7NixOnLkiMLDw9WyZUtt2rRJGRkZDv0//PBDpaena8SIETpw4ID69evXkKdSJ3FxcfrjH/+oN998UxcvXtS3336r8PBwXbhwQZI0cuRIWa1W3XvvvcrPz7fv98ADD+jkyZP65JNPNGzYsAaLl/lyr/kCAADOQdEJuBFfX18NGTJEXbt2dXgr77Jly7RlyxalpaXJ39/fYZ81a9bIw8NDkydPVkFBQUOHXG+2bNmi7du3Kz09XXfeeac8PT0VEhKiHTt2qHfv3vZ+06dP1y9+8QsNGzZMly9fliRZLBa1b99egwYNUpcuXRosZubLveYLAAA4B0Un4Oa+//57zZs3TwsXLpTVaq20PTo6WomJiTp16pRmzpzpggjrxyuvvKJ+/fopMjKy2r4LFizQ3r17tXr16gaIrHaYr8oa83wBAICbR9EJuLk1a9bIGKORI0det8+SJUvUtWtXvf766/roo49ueDxjjFauXKkePXrIx8dHwcHBGjVqlL777jt7n+TkZPn5+clms2nHjh0aOnSoAgICFBoaqs2bNzscr7y8XM8995zCwsLk6+urPn36KDU1tVbneOnSJX3xxRfq27dvjfoHBwfr7rvv1urVqxvdshnMV2WNeb4AAMDNo+gE3FxGRoa6desmm8123T6+vr76r//6L3l4eOjRRx9VcXHxdfsuWLBAzzzzjJ599lnl5OTok08+0YkTJzRo0CD9+OOPkqRp06ZpxowZKikpkb+/v1JTU5WVlaXOnTvr0UcfVVlZmf14s2fP1osvvqhVq1bpzJkzGjFihB566CF99dVXNT7H06dP69KlS/r66681ePBghYSEyGq1qkePHlq3bl2Vhcrtt9+uU6dO6R//+EeNx2kIzJd7zRcAALh5FJ2AGysuLtbRo0cVHh5ebd+BAwdqxowZOnbsmGbPnl1ln5KSEq1cuVJjxozRhAkTFBgYqMjISL366qvKzc3V+vXrK+0THR2tgIAAtW7dWvHx8SouLlZ2drYkqbS0VMnJyRo9erTGjh2roKAgzZ07V15eXkpJSanxeVa8eKZ169Z6/vnndeDAAf34448aNWqUHn/8cb311luV9ql4FnD//v01HsfZmC/3mi8AAFA/PF0dQHUsFourQ0ANMVcNLycnR8aYG141u9aSJUv07rvvat26dYqLi6u0/cCBA7pw4YL69+/v0B4VFSVvb29lZmbe8Pje3t6SZL9ydujQIV28eNHhxTG+vr5q166dw+2f1al4CU+vXr0UHR1tb1+4cKFeeeUVrV+/XuPHj3fYpyInFVf7GgPmy73mCwAA1I9GX3TW9lkiNLxVq1ZJkmbMmOHiSNzH7t276+WlKaWlpZLk8GbUG7FarUpJSdFdd92lRx55RMuXL3fYXrFsRfPmzSvtGxQUpKKiolrFV3Fb6Ny5czV37lyHbSEhITU+TkXf3Nxch3Zvb2917NhRWVlZlfbx9fWV9K8cNQbMl3vNFwAAqB+Nvuh88MEHXR0CqpGeni6Juaqt+ig6K/5QLy8vr/E+AwcO1JNPPqkVK1Zo8eLFCgsLs28LCgqSpCqLlfz8fIWGhtYqvtatW0u6+h8TiYmJtdr3Ws2bN1eXLl108ODBStsuX76swMDASu2XLl2S9K8cNQbMl3vNFwAAqB880wm4sTZt2shisdR6PcfFixere/fu+uabbxzae/furebNm1d6aUxmZqYuXbqkO+64o1bjdOjQQVarVXv37q3VflWJi4vTN998oyNHjtjbLl68qOPHj1e5LEdFTtq2bXvTY9cX5su95gsAANQPik7AjdlsNnXu3FknT56s1X4Vt202a9asUntSUpK2bdumTZs2qbCwUPv379fUqVMVEhKiyZMn13qcSZMmafPmzUpOTlZhYaHKy8t18uRJnTlzRpIUHx+vtm3bas+ePTc81pNPPqmOHTtq4sSJys7OVl5enmbNmqWSkpIqX7RTkZOarBPZUJgv95ovAABQPyg6ATc3fPhwHThwQCUlJfa2P/3pT4qIiFBWVpaioqL0xBNPVNpvwIABevLJJyu1z58/X0uXLtWiRYvUqlUr3X333erUqZN27dolPz8/SVfXfax4lrdPnz46cuSIXnvtNSUlJUmShgwZosOHD0u6ehvxjBkztHz5crVs2VIhISFKTEzU+fPnJV29rTInJ0c7duy44XkGBwfr008/VWhoqPr27av27dvr73//uzIyMqpcD/LLL79U+/bt1adPn5qkscEwX+41XwAAoB6YRio1NdU04vBwjZiYGBMTE+PqMNxKXb6+J0+ebNq3b1+p/fDhw8bT09Ns3LixvsJrUOXl5WbQoEFmw4YN9XbM3NxcY7VazYoVKyptmz59umnZsmWtjsd8/UtTnS8AAOA8XOkE3EhJSYk++OADHT582P7ilYiICC1atEiLFi2yr4/oLsrLy7V9+3YVFRUpPj6+3o67YMEC9e3bVwkJCZIkY4xOnz6tzz77TN9//329jVMd5qtmGst8AQAA52gyRecXX3yhHj16yMPDQxaLRW3bttWSJUtcHZaDrVu3qnPnzrJYLLJYLGrXrp0mTJjg6rDgRs6dO6chQ4aoa9eueuSRR+ztzzzzjGJjYxUfH1/rl9S40q5du7R161bt3LmzxmtXVmflypXau3ev3nvvPXl5eUmSduzYofbt22vQoEHKyMiol3FqgvmqXmOaLwAA4BwWY4xxdRBVSUtLU1xcnGob3pAhQ/TBBx/o/Pnz9uUEGpuIiAjl5uba19hzd7GxsZL+tXQKqlfXr+/qfPjhh/rrX/+qZcuW1etx3cWOHTt08OBBPf3005VeunMzmC/ncLf5AgAAddNkrnQ2RiUlJYqOjnZ1GLeEhsi1O8zn/ffff8sWMJL0wAMP6JlnnqnXAsaZmC/3mi8AAFA3FJ1OtGHDBuXk5Lg6jFtCQ+Sa+QQAAABqr8kXncnJyfLz85PNZtOOHTs0dOhQBQQEKDQ0VJs3b7b3W7NmjaxWq9q0aaMpU6YoJCREVqtV0dHRyszMtPdLSEiQt7e32rVrZ2977LHH5OfnJ4vFotzcXElSYmKikpKSlJWVJYvFooiIiDrF/+mnn6pnz54KDAyU1WpVZGSkPvjgA0nSf/7nf9qfDw0PD7cvHD9p0iTZbDYFBgbqnXfekXT1BSDPPfecwsLC5Ovrqz59+ig1NVWS9OKLL8pms8nf3185OTlKSkpS+/btdejQoTrFXBPGGK1cuVI9evSQj4+PgoODNWrUKH333Xf2PjeT64aaz/fff18BAQF6/vnnnZYrAAAAwK257L251ajrK+9//etfG0nm/Pnz9rZnn33WSDIff/yxKSgoMDk5OWbQoEHGz8/PXLp0yd5v8uTJxs/Pzxw8eNCUlpaaAwcOmKioKOPv72+ys7Pt/caPH2/atm3rMO5LL71kJJmzZ8/a28aOHWvCw8MrxRgeHm4CAwNrdD7p6elmwYIF5ty5cyYvL88MGDDAYfmAsWPHmmbNmplTp0457PfQQw+Zd955x/7vmTNnGh8fH/P222+b8+fPmzlz5hgPDw/z5ZdfOuRo+vTp5uWXXzZjxowx3377bY1irMuSKc8995zx9vY2GzduNPn5+Wbfvn2mX79+plWrVuaHH36w97uZXDfEfL777rvG39/fLFq0qFbnz5IO7oX5ci/MFwAAjUuTv9J5rejoaAUEBKh169aKj49XcXGxsrOzHfp4enrar7717NlTycnJKioqUkpKiktijomJ0fz58xUcHKwWLVpo5MiRysvL09mzZyVJU6dOVXl5uUN8hYWF+vLLLzVs2DBJUmlpqZKTkzV69GiNHTtWQUFBmjt3rry8vCqd17Jly/T4449r69at6t69u1POqaSkRCtXrtSYMWM0YcIEBQYGKjIyUq+++qpyc3O1fv36ehvL2fM5fPhwFRYWat68efVyPAAAAKCpuaWKzmt5e3tLksrKym7Yr3///rLZbA63fbpSxZIC5eXlkqR///d/V9euXfXGG2/Y39S4ZcsWxcfH21/OcejQIV28eFG9e/e2H8fX11ft2rVzyXkdOHBAFy5cUP/+/R3ao6Ki5O3t7XD7a31rbPMJAAAANHW3bNFZGz4+PvYriw0tIyND99xzj1q3bi0fHx89/fTTDtstFoumTJmiI0eO6OOPP5Yk/fGPf9Tvfvc7e5/i4mJJ0ty5c+3PgFosFh0/flwXL15suJP5PxVLxTRv3rzStqCgIBUVFTl1fFfOJwAAAHCroeisRllZmfLz8xUaGtog433yySdatWqVJCk7O1ujR49Wu3btlJmZqYKCAi1fvrzSPhMnTpTVatXrr7+uQ4cOKSAgQB07drRvb926tSRp1apVMsY4fHbv3t0g53WtivVTqyounZ3rhp5PAAAA4Fbn6eoAGrtdu3bJGKMBAwbY2zw9Pau9Lbeuvv76a/n5+UmS9u/fr7KyMk2bNk2dO3eWdPXK5s8FBwcrLi5OW7Zskb+/vx599FGH7R06dJDVatXevXudEnNt9e7dW82bN9dXX33l0J6ZmalLly7pjjvusLfVd64bej4BAACAWx1XOn/mypUrOn/+vC5fvqx9+/YpMTFRYWFhmjhxor1PRESEzp07p+3bt6usrExnz57V8ePHKx2rRYsWOn36tI4dO6aioqIbFjZlZWX68ccftWvXLnvRGRYWJkn66KOPVFpaqsOHD1/3ecepU6fqp59+0rvvvqsRI0Y4bLNarZo0aZI2b96s5ORkFRYWqry8XCdPntSZM2dqm6KbZrValZSUpG3btmnTpk0qLCzU/v37NXXqVIWEhGjy5Mn2vjeba2fP586dO1kyBQAAALiBJlN0ZmZmqnfv3vrLX/4iSerRo4eWLl2q5ORk++2qffr00ZEjR/Taa68pKSlJkjRkyBAdPnzYfpzS0lJFRkbK19dXgwYNUteuXfW3v/1NPj4+9j7Tpk3T4MGDNW7cOHXr1k2LFy+Wr6+vJGngwIE6ceKEpKuFYJs2bdSzZ08NGzZMGzZsUEREhLKyslRQUODwfGXFWpHvvPOObDabJCkyMlKzZs3SunXrFBISomeffVb33HOPJOmuu+6yjyNJd955p26//XZNmjRJnp6VL2CvXr1aM2bM0PLly9WyZUuFhIQoMTFR58+f14svvqiVK1dKkrp27apNmzbVy5zcyPz587V06VItWrRIrVq10t13361OnTo5FN1S3XN97tw5Sc6dz4oxAAAAAFyfxVS88rSRSUtLU1xcnBoyvClTpig9PV15eXkNNmZ9Gj58uNauXavbbrutQceNjY2VJKWnpzfouNVpzPPpiq9v1B3z5V6YLwAAGpcmc6WzvlQsReIOrr1dd9++fbJarQ1ecDZ27jSfAAAAQFPEi4Tc2KxZszR16lQZYzRp0iRt3LjR1SEBAAAAgAOudP6fOXPmKCUlRQUFBbrtttv09ttvuzqkatlsNnXv3l333XefFixYoJ49e7o6pEbDHecTAAAAaIooOv/P0qVL9dNPP8kYo6NHjyomJsbVIVVryZIlKi8vV3Z2dqU31t7q3HE+AQAAgKaIohMAAAAA4DQUnQAAAAAAp6HoBAAAAAA4DUUnAAAAAMBpGv2SKWlpaa4OAdU4efKkJOaqNnbv3i2JnLkL5su9VMwXAABoHCzGGOPqIKqSlpamuLg4V4cBAHBTjfTXGwAAt5xGW3QCuL6K/5Th2xcAAACNHc90AgAAAACchqITAAAAAOA0FJ0AAAAAAKeh6AQAAAAAOA1FJwAAAADAaSg6AQAAAABOQ9EJAAAAAHAaik4AAAAAgNNQdAIAAAAAnIaiEwAAAADgNBSdAAAAAACnoegEAAAAADgNRScAAAAAwGkoOgEAAAAATkPRCQAAAABwGopOAAAAAIDTUHQCAAAAAJyGohMAAAAA4DQUnQAAAAAAp6HoBAAAAAA4DUUnAAAAAMBpKDoBAAAAAE5D0QkAAAAAcBqKTgAAAACA01B0AgAAAACchqITAAAAAOA0FJ0AAAAAAKeh6AQAAAAAOA1FJwAAAADAaSg6AQAAAABOQ9EJAAAAAHAaik4AAAAAgNNQdAIAAAAAnMbT1QEAuLGTJ0/qt7/9rcrLy+1t58+fl7+/v+655x6Hvt26ddMf/vCHBo4QAAAAuD6KTqCRCw0N1fHjx5WVlVVp23//9387/PuXv/xlQ4UFAAAA1Ai31wJu4De/+Y28vLyq7RcfH98A0QAAAAA1ZzHGGFcHAeDGsrKy1KVLF93o27VXr1765z//2YBRAQAAANXjSifgBsLDw9WnTx9ZLJYqt3t5eem3v/1tA0cFAAAAVI+iE3ATv/nNb9SsWbMqt12+fFmxsbENHBEAAABQPW6vBdzEmTNnFBoaqitXrji0e3h46M4779Tnn3/uosgAAACA6+NKJ+AmQkJC9P/+3/+Th4fjt62Hh4d+85vfuCgqAAAA4MYoOgE38vDDD1dqM8ZozJgxLogGAAAAqB5FJ+BGYmJiHJ7rbNasme677z61adPGhVEBAAAA10fRCbiR4OBg/epXv7IXnsYYTZgwwcVRAQAAANdH0Qm4mQkTJthfJuTl5aVRo0a5OCIAAADg+ig6ATczcuRI+fj4SJJGjBih5s2buzgiAAAA4PooOgE34+fnZ7+6ya21AAAAaOwqrdOZlpamuLg4V8UDAECTwnLYAIBbnef1NqSmpjZkHECjt2rVKknSjBkzXByJVF5ertTUVD300EOuDuWGdu/erdWrV/PzxEXIv2tV5B8AgFvddYvOBx98sCHjABq99PR0SY3ne2P06NGyWq2uDqNaq1evbjQ5uxWRf9ei6AQAgGc6AbflDgUnAAAAQNEJAAAAAHAaik4AAAAAgNNQdAIAAAAAnIaiEwAAAADgNDdddEZFRalZs2bq27dvfcRTK5MmTZLVapXFYlFpaWmDj98YrVixQm3atJHFYtGrr75qb3/vvfcUGBioP//5z04dv6HGqYlFixapZ8+eCggIkI+PjyIiIvT000/rwoULLo2rMeUIAAAAcLabLjq//PJLDR48uD5iqbWUlBTNnDnTJWM3VjNnztTnn39eqb2hFidvTIug//Wvf9Xjjz+uY8eOKTc3V0uXLtXq1asVGxvr0rgaU44AAAAAZ7vuOp21ZbFYbvoYJSUluvfee6ssmnBzhg8froKCgno9ZlXz5Yxx6qp58+aaPHmymjVrJunq+ppbt25VWlqaTpw4oQ4dOrgkrsaUI77nAAAA4Gz1VnR6eXnd9DE2bNignJycOu1bH0Uvaudm5qshvPvuu5XaWrVqJUm6ePFiQ4fTKDX2OQQAAID7q7cXCX3//ffq3r27/Pz85Ovrq0GDBumzzz5z6PPpp5+qZ8+eCgwMlNVqVWRkpD744ANJUmJiopKSkpSVlSWLxaKIiAj7fhs3blT//v1ltVrl5+enTp06afHixf86CQ8PZWRkaOjQoQoMDFRISIjeeOONWp9DcnKy/Pz8ZLPZtGPHDg0dOlQBAQEKDQ3V5s2bHfoaY7Ry5Ur16NFDPj4+Cg4O1qhRo/Tdd9/Z+7z44ouy2Wzy9/dXTk6OkpKS1L59e02dOlV+fn7y8PDQHXfcobZt28rLy0t+fn7q16+fBg0apA4dOshqtSooKEhPP/10jfNYlc8++0xhYWGyWCxau3atpKvzZbFYqvz85S9/qdN8VTVOTXNVm9zfjFOnTsnX11e33XZbvR2zNqrKUU3Pfc2aNbJarWrTpo2mTJmikJAQWa1WRUdHKzMz094vISFB3t7eateunb3tsccek5+fnywWi3JzcyVd/3vu/fffV0BAgJ5//vmGSAkAAACaOvMzqamppormG7r33ntN586dzdGjR01ZWZn55z//ae68805jtVrN//7v/9r7paenmwULFphz586ZvLw8M2DAANOyZUv79rFjx5rw8HCHY69atcpIMi+88ILJy8sz586dM3/4wx/M+PHjjTHGPPvss0aS+fjjj01+fr45d+6cGTZsmPHx8THFxcW1Oo+fH6+goMDk5OSYQYMGGT8/P3Pp0iV7v+eee854e3ubjRs3mvz8fLNv3z7Tr18/06pVK/PDDz9UOt706dPNyy+/bMaMGWO+/fZbM3/+fCPJZGZmmuLiYpObm2uGDBliJJmMjAxz9uxZU1xcbBISEowks3fv3hrn8fDhw0aSeeWVV+xtJ06cMJLMyy+/bO8ze/Zse47OnDljgoODTXR0tCkvL6/zfP18nLrkqrrc11VxcbHx9/c3CQkJddo/JibGxMTE3HQcVeWopuc+efJk4+fnZw4ePGhKS0vNgQMHTFRUlPH39zfZ2dn2fuPHjzdt27Z1GPell14ykszZs2ftbVXN4bvvvmv8/f3NokWLbvpc6/LzBPWH/LsW+QcA4Kp6u9Lp7++vTp06ydPTU7169dJrr72m0tJSrV+/3t4nJiZG8+fPV3BwsFq0aKGRI0cqLy9PZ8+erfKYZWVlWrhwoQYPHqzZs2erRYsWCg4O1u9+9ztFRUU59I2OjlZgYKCCg4MVHx+vn376SUePHq3z+URHRysgIECtW7dWfHy8iouLlZ2dLenqc3ArV67UmDFjNGHCBAUGBioyMlKvvvqqcnNzHc65wrJly/T4449r69at6t69u729Z8+esr1TyV4AACAASURBVNlsatmypcaNGydJCgsLU6tWrWSz2TRhwgRJcrgqWNs8ViUiIkIvvPCCbDabJGnatGkqKSnRG2+8IQ8Pj3obpy65ulHub8bSpUsVEhKiJUuW3PSxnKUm5+7p6Wm/atyzZ08lJyerqKhIKSkp9RLD8OHDVVhYqHnz5tXL8QAAAHBrq7dnOn8uMjJSgYGB2rdv33X7VDwHWl5eXuX2ffv2KT8/X7/+9a8d2ps1a6bp06dXe9yysrLahl0lb29vh+MdOHBAFy5cUP/+/R36RUVFydvb2+FWx7qMc/nyZXtbTc6lujxWJy0tTX/605+0fPlydevWrV7Hudlc/Tz3dbVt2zalpaXpww8/lL+//00dq6HU9Nz79+8vm83m8B8TAAAAQGPhtKJTulqkXPsHc0ZGhl566SUdOHBAhYWF1f4xXVhYKEkKCgpyZpi1lp+fL+nq21F/LigoSEVFRU4dv7Z5vJG8vDw98cQTioqKUlJSUr2P4+pcSdKWLVu0cuVK7dq1S//2b//m9PFcwcfHp1ZXoAEAAICGUm+31/7c5cuXde7cOYWFhUmSsrOzNXr0aLVr106ZmZkqKCjQ8uXLb3iMigKh4sUnjUVFEVxVwZSfn6/Q0FCnjV2XPN7I9OnTlZ+fr5SUFPvSIvU5jitzJUkvv/yyNm3apL/+9a9NtuAsKytrkFwCAAAAdeG0ovNvf/ubrly5on79+kmS9u/fr7KyMk2bNk2dO3eW1WqtdpmTTp06qUWLFvrwww+dFWad9O7dW82bN9dXX33l0J6ZmalLly7pjjvucNrYdcnj9WRkZOjNN9/UvHnz1KtXL3v7U089VW/juCpXxhjNmjVL+/fv1/bt26u80tpU7Nq1S8YYDRgwwN7m6elZb7eXAwAAADej3orOS5cuqaCgQJcvX9aePXuUkJCgjh07auLEiZJkv+L50UcfqbS0VIcPH670PF+LFi10+vRpHTt2TEVFRfLw8NCcOXP0ySefKCEhQadOndKVK1dUVFSkgwcP1lfotWa1WpWUlKRt27Zp06ZNKiws1P79+zV16lSFhIRo8uTJThu7JnmsicLCQk2ZMkV9+/bV7NmzJUmlpaX66quvtHfv3jrNV1VFjqtydfDgQb344ot67bXX5OXlVWlZmBUrVjhl3IZw5coVnT9/XpcvX9a+ffuUmJiosLAw+/eadPVFUefOndP27dtVVlams2fP6vjx45WOVdUc7ty5kyVTAAAAUH9+/jrburziPSUlxQwePNi0adPGeHp6mpYtW5px48aZ48ePO/SbNWuWadGihQkKCjKxsbFm7dq1RpIJDw832dnZZs+ePaZjx47G19fX3HXXXfblNNauXWsiIyON1Wo1VqvV3H777WbdunVm+fLlxtfX10gyXbp0MVlZWWbTpk0mODjYSDKhoaHmn//8Z43PY926dcZmszkcb/369SYgIMBIMh07drQvAXPlyhXz0ksvmS5duhgvLy8THBxsRo8ebQ4dOmQ/3rXxdejQwWzcuNEYY8zq1avt43Tq1Ml8+umnZtmyZSYwMNBIMm3btjVvvvmm2bJli2nbtq2RZIKDg83mzZurzWNiYqJ9Hz8/PzNmzBjz8ssvm3bt2hlJxmazmZEjR5oVK1YYSVV+hg0bVqf5mjt3bqVxapqr2uS+Jvbv33/d85NkXnrppRofq0J9LJlS1VzU5twnT55svLy8TPv27Y2np6cJCAgwo0aNMllZWQ7j5OXlmcGDBxur1Wpuu+0288QTT5innnrKSDIRERH25VWq+p577733jL+/v1myZMlNnasxLBnhauTftcg/AABXWYwx5toiNC0tTXFxcfpZM3DLi42NlSSlp6e7LIYpU6YoPT1deXl5LouhNvh54lrk37XIPwAAVzntmU4AzlHXpXEAAAAAV2jyRed3331X6Xm+qj7x8fGuDhXVYC5vPR999JGeeeYZbd26VZ07d7bP8cMPP1yp7/333y9/f381a9ZMvXr10p49e1wQce2UlZVp6dKlioiIkLe3t4KCgtS7d28dO3ZMkvTOO+9o+fLlLvuPBvLv2vwDANBUNPmis3v37jLGVPvZsmWLq0NFNW71uZwzZ45SUlJUUFCg2267TW+//barQ3Kq+fPna82aNZozZ47Gjh2rI0eOKDw8XC1bttSmTZuUkZHh0P/DDz9Uenq6RowYoQMHDtjfnN2YxcXF6Y9//KPefPNNXbx4Ud9++63Cw8N14cIFSdLIkSNltVp177332te8bSjk37X5BwCgKWnyRSfQVCxdulQ//fSTjDE6evSoYmJiXB2S0yxbtkxbtmxRWlqa/P39HbatWbNGHh4emjx5sgoKClwU4c3bsmWLtm/frvT0dN15553y9PRUSEiIduzYod69e9v7TZ8+Xb/4xS80bNgwXb58uUFiI/+uzT8AAE0NRSeARuX777/XvHnztHDhQlmt1krbo6OjlZiYqFOnTmnmzJkuiLB+vPLKK+rXr58iIyOr7btgwQLt3btXq1evdnpc5L+yhsw/AABNEUUngEZlzZo1MsZo5MiR1+2zZMkSde3aVa+//ro++uijGx7PGKOVK1eqR48e8vHxUXBwsEaNGqXvvvvO3ic5OVl+fn6y2WzasWOHhg4dqoCAAIWGhmrz5s0OxysvL9dzzz2nsLAw+fr6qk+fPkpNTa3VOV66dElffPGF+vbtW6P+wcHBuvvuu7V69WqnvwmV/FfWkPkHAKApougE0KhkZGSoW7dustls1+3j6+ur//qv/5KHh4ceffRRFRcXX7fvggUL9Mwzz+jZZ59VTk6OPvnkE504cUKDBg3Sjz/+KEmaNm2aZsyYoZKSEvn7+ys1NVVZWVnq3LmzHn30UZWVldmPN3v2bL344otatWqVzpw5oxEjRuihhx7SV199VeNzPH36tC5duqSvv/5agwcPVkhIiKxWq3r06KF169ZVWdjcfvvtOnXqlP7xj3/UeJy6IP+uzT8AAE0RRSeARqO4uFhHjx5VeHh4tX0HDhyoGTNm6NixY5o9e3aVfUpKSrRy5UqNGTNGEyZMUGBgoCIjI/Xqq68qNzdX69evr7RPdHS0AgIC1Lp1a8XHx6u4uFjZ2dmSpNLSUiUnJ2v06NEaO3asgoKCNHfuXHl5eSklJaXG51nxoprWrVvr+eef14EDB/Tjjz9q1KhRevzxx/XWW29V2qdLly6SpP3799d4nNoi/67NPwAATZXn9TakpaU1ZBxAo3fy5ElJfG/Uxu7du2vVPycnR8aYG15lu9aSJUv07rvvat26dYqLi6u0/cCBA7pw4YL69+/v0B4VFSVvb29lZmbe8Pje3t6SZL/SdujQIV28eNHhRTO+vr5q166dw+2i1fHx8ZEk9erVS9HR0fb2hQsX6pVXXtH69es1fvx4h30qclJxddAZyL9r8w8AQFN13aKzqj8gAPC94UylpaWS/lUUVMdqtSolJUV33XWXHnnkES1fvtxhe8UyF82bN6+0b1BQkIqKimoVX8VtpHPnztXcuXMdtoWEhNT4OBV9c3NzHdq9vb3VsWNHZWVlVdrH19dX0r9y5Azk37X5BwCgqbru7bU1WQ+RD59b6RMTE6OYmBiXx+FOn9q+4KXiD/vy8vIa7zNw4EA9+eSTOnz4sBYvXuywLSgoSJKqLG7y8/MVGhpaq/hat24tSVq1alWlc63NVd3mzZurS5cuOnjwYKVtly9fVmBgYKX2S5cuSfpXjpyB/Ls2/wAANFU80wmg0WjTpo0sFkut139cvHixunfvrm+++cahvXfv3mrevHmll8xkZmbq0qVLuuOOO2o1TocOHWS1WrV3795a7VeVuLg4ffPNNzpy5Ii97eLFizp+/HiVy3hU5KRt27Y3Pfb1kH/X5h8AgKaKohNAo2Gz2dS5c2f787M1VXGbZ7NmzSq1JyUladu2bdq0aZMKCwu1f/9+TZ06VSEhIZo8eXKtx5k0aZI2b96s5ORkFRYWqry8XCdPntSZM2ckSfHx8Wrbtq327Nlzw2M9+eST6tixoyZOnKjs7Gzl5eVp1qxZKikpqfLFPBU5qcm6knVF/l2bfwAAmiqKTgCNyvDhw3XgwAGVlJTY2/70pz8pIiJCWVlZioqK0hNPPFFpvwEDBujJJ5+s1D5//nwtXbpUixYtUqtWrXT33XerU6dO2rVrl/z8/CRdXSdy1apVkqQ+ffroyJEjeu2115SUlCRJGjJkiA4fPixJWr16tWbMmKHly5erZcuWCgkJUWJios6fPy/p6m2YOTk52rFjxw3PMzg4WJ9++qlCQ0PVt29ftW/fXn//+9+VkZFR5fqRX375pdq3b68+ffrUJI11Rv5dm38AAJoiizHGYUGytLQ0xcXF6WfNwC0vNjZWkpSenu7iSNxHXX6efP/99+rRo4dSUlI0YcIEJ0bnHFeuXNE999yjiRMn6pFHHqmXY+bl5Sk0NFRLliyxF2I1Qf7dL/8AADRFXOkE0KhERERo0aJFWrRokX09RXdRXl6u7du3q6ioSPHx8fV23AULFqhv375KSEiot2NeD/mvrCHzDwBAU0TRCaDReeaZZxQbG6v4+Phav9TGlXbt2qWtW7dq586dNV7rsjorV67U3r179d5778nLy6tejlkd8v8vrsg/AABNTYMXnVu3blXnzp1lsVgcPp6enmrVqpXuu+8+bdu2zelxTJo0SVarVRaL5Ybrrv083ocffrhSn/vvv1/+/v5q1qyZevXqVe0LLFxtxYoV9rdUvvrqq/b29957T4GBgfrzn//s1PEbahy4t+eff14JCQl64YUXXB1Kjd17771688031a5du3o53o4dO/TTTz9p165dCg4Orpdj1hT5d23+AQBoShq86Bw7dqyOHDmi8PBwBQYG2tdYO3v2rFJTU3Xq1CmNHTu21uv71VZKSopmzpxZq3hbtmypTZs2KSMjw6HPhx9+qPT0dI0YMUIHDhxQv379nBV2vZg5c6Y+//zzSu0N9dwRzzehpu6//34tW7bM1WG4zAMPPKBnnnmm0lthGwr5d23+AQBoKhrN7bXBwcG699579fvf/17S1Rcw1EZJSYmio6OdEZrdmjVr5OHhocmTJ7vVLWc1NXz4cBUUFGjEiBH1dsyq5sUZ49wqGuLrvCHGAAAAwK2j0RSdFTp16iRJys/Pr9V+GzZsUE5OTp3GtFgsNeoXHR2txMREnTp1qkZXSXFz84LKGiKfzBkAAADqU6MrOvft2ydJuvvuux3aP/30U/Xs2VOBgYGyWq2KjIzUBx98IElKTExUUlKSsrKyZLFYFBERYd9v48aN6t+/v6xWq/z8/NSpUyctXrzYvt3Dw0MZGRkaOnSoAgMDFRISojfeeOO68S1ZskRdu3bV66+/ro8++uiG52KM0cqVK9WjRw/5+PgoODhYo0aN0nfffWfv8+KLL8pms8nf3185OTlKSkpS+/btNXXqVPn5+cnDw0N33HGH2rZtKy8vL/n5+alfv34aNGiQOnToIKvVqqCgID399NM1zldVPvvsM4WFhclisWjt2rWSri6d8PNnbys+f/nLX+o0L1WNU9NcJScny8/PTzabTTt27NDQoUMVEBCg0NBQbd68+YZz4So1Oa+EhAR5e3s7PIf22GOPyc/PTxaLRbm5uZKqzueaNWtktVrVpk0bTZkyRSEhIbJarYqOjlZmZma9jCFJ77//vgICAvT88887NV8AAABogszPpKammiqa6114eLgJDAy0//vixYtm586dpmPHjub+++83Fy5ccOifnp5uFixYYM6dO2fy8vLMgAEDTMuWLe3bx44da8LDwx32WbVqlZFkXnjhBZOXl2fOnTtn/vCHP5jx48cbY4x59tlnjSTz8ccfm/z8fHPu3DkzbNgw4+PjY4qLiyvFe/ToUWOMMZ9//rnx8PAwnTp1sse5c+dO88ADDzjs89xzzxlvb2+zceNGk5+fb/bt22f69etnWrVqZX744Qd7v4o4pk+fbl5++WUzZswY8+2335r58+cbSSYzM9MUFxeb3NxcM2TIECPJZGRkmLNnz5ri4mKTkJBgJJm9e/fWOF+HDx82kswrr7xibztx4oSRZF5++WV7n9mzZ9tzcebMGRMcHGyio6NNeXl5nefl5+PUJVcff/yxKSgoMDk5OWbQoEHGz8/PXLp0yThTTEyMiYmJqdU+NT2v8ePHm7Zt2zrs+9JLLxlJ5uzZs/a2qvI5efJk4+fnZw4ePGhKS0vNgQMHTFRUlPH39zfZ2dn1Msa7775r/P39zaJFi2p1/g318wRVI/+uRf4BALjKpVc6CwoK7FfObDab/crV+PHjK72aPiYmRvPnz1dwcLBatGihkSNHKi8vT2fPnq3y2GVlZVq4cKEGDx6s2bNnq0WLFgoODtbvfvc7RUVFOfSNjo5WYGCggoODFR8fr59++klHjx69btwDBw7UjBkzdOzYMc2ePbvKPiUlJVq5cqXGjBmjCRMmKDAwUJGRkXr11VeVm5ur9evXV9pn2bJlevzxx7V161Z1797d3t6zZ0/ZbDa1bNlS48aNkySFhYWpVatWstls9gXcr716Vtt8VSUiIkIvvPCCfemBadOmqaSkRG+88YY8PDzqbZy65Co6OloBAQFq3bq14uPjVVxcrOzs7BqP2RDqcl515enpab+a2rNnTyUnJ6uoqEgpKSn1cvzhw4ersLBQ8+bNq5fjAQAA4Nbh0qLz2rfXlpWV6eTJk5oxY4YSEhLUp08f+y1/VakoSsvLy6vcvm/fPuXn5+vXv/61Q3uzZs00ffr0ao9bVlZ2w9iXLFmibt26ad26dfrss88qbT9w4IAuXLig/v37O7RHRUXJ29vb4dbH2vD29pYkXb58uVYxV5ev6qSlpelPf/qTFi5cqG7dutXrODebq4qcVDdnDc1ZXwM10b9/f9lsNof/iAAAAABcodE80+np6an27dtr0qRJWrFihQ4dOuSwPlxGRobuuecetW7dWj4+PpWeYfy5wsJCSVJQUJBT4rVarUpJSZHFYtEjjzyikpISh+0VL0Jq3rx5pX2DgoJUVFTklLgq1DZfN5KXl6cnnnhCUVFRSkpKqvdxXJ0rZ3H1efn4+NTqijMAAADgDI2m6LxWZGSkJOngwYOSpOzsbI0ePVrt2rVTZmamCgoKtHz58hse49/+7d8k6YZXS2/WwIED9eSTT+rw4cMOLyeS/lXsVlVY5OfnKzQ01Glx1SVfNzJ9+nTl5+crJSXFYb26+hrHlblyJleeV1lZmVvnDgAAAE1Hoyw6v/76a0my38a5f/9+lZWVadq0aercubOsVmu1y5x06tRJLVq00IcffujUWBcvXqzu3bvrm2++cWjv3bu3mjdvrq+++sqhPTMzU5cuXdIdd9zhtJjqkq/rycjI0Jtvvql58+apV69e9vannnqq3sZxZa6cqTbn5enpWa+3B+/atUvGGA0YMMBpYwAAAAA14fKis6SkRFeuXJExRqdPn1ZKSormzp2rVq1aacaMGZKuvjRHkj766COVlpbq8OHDlZ6Ha9GihU6fPq1jx46pqKhIHh4emjNnjj755BMlJCTo1KlTunLlioqKiuxXUOtDxW22114BrGhPSkrStm3btGnTJhUWFmr//v2aOnWqQkJCNHny5HqL4edqkq+aKCws1JQpU9S3b1/7C5NKS0v11Vdfae/evXWal6qKHlfmyplqc14RERE6d+6ctm/frrKyMp09e1bHjx+vdMzr5fPKlSs6f/68Ll++rH379ikxMVFhYWGaOHFivYyxc+dOlkwBAABA3fz8dbbOfsX7tm3bTHh4uJFU6ePj42O6dOlipk2b5rDUgzHGzJo1y7Ro0cIEBQWZ2NhYs3btWiPJhIeHm+zsbLNnzx7TsWNH4+vra+666y77chRr1641kZGRxmq1GqvVam6//Xazbt06s3z5cuPr62skmS5dupisrCyzadMmExwcbCSZ0NBQ889//tMh3latWpnHH3+8yvN66qmnKi2ZcuXKFfPSSy+ZLl26GC8vLxMcHGxGjx5tDh06ZO9zbRwdOnQwGzduNMYYs3r1amOz2Ywk06lTJ/Ppp5+aZcuWmcDAQCPJtG3b1rz55ptmy5Ytpm3btkaSCQ4ONps3b642X4mJifZ9/Pz8zJgxY8zLL79s2rVrZyQZm81mRo4caVasWFHlPEkyw4YNq9O8zJ07t9I4Nc3VunXr7DmpmLP169ebgIAAI8l07NjR/O///u/NfHneUF2WTKnJeRljTF5enhk8eLCxWq3mtttuM0888YR56qmnjCQTERFh/36o6ut88uTJxsvLy7Rv3954enqagIAAM2rUKJOVlVVvY7z33nvG39/fLFmypFbnz5IRrkX+XYv8AwBwlcUYY64tQtPS0hQXF6efNQO3vNjYWElSenq6iyNxNGXKFKWnpysvL8/VoVTCzxPXIv+uRf4BALjK5bfXArh5dV0KBwAAAHA2ik4AAAAAgNNQdAJubM6cOUpJSVFBQYFuu+02vf32264OCQAAAHDg6eoAANTd0qVLtXTpUleHAQAAAFwXVzoBAAAAAE5D0QkAAAAAcBqKTgAAAACA01B0AgAAAACc5rovEoqNjW3IOIBG74svvpDE90ZtnDx5UhI5cxXy71oV+QcA4FZnMcaYaxt2796tlStXuioeADXwww8/6JtvvtHQoUNdHQqAaqSnp7s6BAAAXKpS0Qmg8UtLS1NcXJz49gUAAEBjxzOdAAAAAACnoegEAAAAADgNRScAAAAAwGkoOgEAAAAATkPRCQAAAABwGopOAAAAAIDTUHQCAAAAAJyGohMAAAAA4DQUnQAAAAAAp6HoBAAAAAA4DUUnAAAAAMBpKDoBAAAAAE5D0QkAAAAAcBqKTgAAAACA01B0AgAAAACchqITAAAAAOA0FJ0AAAAAAKeh6AQAAAAAOA1FJwAAAADAaSg6AQAAAABOQ9EJAAAAAHAaik4AAAAAgNNQdAIAAAAAnIaiEwAAAADgNBSdAAAAAACnoegEAAAAADgNRScAAAAAwGkoOgEAAAAATkPRCQAAAABwGopOAAAAAIDTUHQCAAAAAJyGohMAAAAA4DSerg4AwI2VlZXpwoULDm3FxcWSpPPnzzu0WywWBQUFNVhsAAAAQHUsxhjj6iAAXN+PP/6o9u3bq7y8vNq+gwcP1l//+tcGiAoAAACoGW6vBRq5tm3b6pe//KU8PG787WqxWDRu3LgGigoAAACoGYpOwA08/PDD1fZp1qyZxowZ0wDRAAAAADVH0Qm4gbFjx8rT8/qPYDdr1kxDhgxRy5YtGzAqAAAAoHoUnYAbCAgI0NChQ69beBpjNGHChAaOCgAAAKgeRSfgJiZMmHDdlwl5e3vrP/7jPxo4IgAAAKB6FJ2Am/iP//gP2Wy2Su1eXl4aPXq0/Pz8XBAVAAAAcGMUnYCbsFqtGjNmjLy8vBzay8rKNH78eBdFBQAAANwYRSfgRh566CGVlZU5tAUEBOhXv/qViyICAAAAboyiE3Aj9913n1q0aGH/t5eXl8aNGydvb28XRgUAAABcH0Un4EY8PT01btw4+y22ZWVleuihh1wcFQAAAHB9FmOMcXUQAGruf/7nf3TXXXdJktq2bavTp0/Lw4P/PwIAAEDjxF+qgJuJjo5W+/btJUm/+c1vKDgBAADQqFW90nwTkpaW5uoQgHoXFRWlU6dOqWXLlnyNo8np0KGDBg4c6OowXGL37t06ceKEq8MAALihxvz7s8nfXmuxWFwdAgCgFmJiYpSenu7qMFwiNjZWb7/9tqvDAAC4ocb8+7PJX+mUpNTUVD344IOuDqPJiI2NlaRG+0XdGKWlpSkuLk71+X88b7/9tmJiYurteE2ZM/IP56j4+XIra8x/NOBfLBYLf1/UEn8/uBfmy7009t+fPAwGuCkKTgAAALgDik4AAAAAgNNQdAIAAAAAnIaiEwAAAADgNBSdAAAAAACnoegEAAAAADgNRec1Jk2aJKvVKovFotLS0hv2fe+99xQYGKg///nP1+3zn//5n/L395fFYtHevXtvut/NunLlilatWqXo6GinjVEbNckhAMB53O3n8FtvvSWLxdJofo/VB3ebAwCoC4rOa6SkpGjmzJk16luT9f5ef/11vfbaa/XW72YcPnxYv/zlL/Xkk0/q4sWLTh2rplgzEQBcy91+Dr/11lsKDw/X7t279f3337s6nHrhbnMAAHVB0VlHw4cPV0FBgUaMGOHqUKr1j3/8Q7Nnz9bUqVPVt29fV4dj15hyWFJS0qT+5xwAasKdfg7n5eXp4MGDWrhwoSTpj3/8Y0OF5lTuNAcAUFcUnddhsVga9Dj1NV5VfvGLX2jr1q0aP368fHx8nDaOO9uwYYNycnJcHQYA3LKq+zmclpam4cOHa+TIkbJardq4cSNXCesZvwsBOAtFZxU8PDyUkZGhoUOHKjAwUCEhIXrjjTfs2z/77DOFhYXJYrFo7dq19nZjjF566SV169ZNPj4+CgwM1FNPPVXp+DXtV15erueee05hYWHy9fVVnz59lJqaKklKTk6Wn5+fbDabduzYoaFDhyogIEChoaHavHmzE7JSv6rKYU3Pac2aNbJarWrTpo2mTJmikJAQWa1WRUdHKzMz094vISFB3t7eateunb3tsccek5+fnywWi3JzcyVJiYmJSkpKUlZWliwWiyIiIiRJ77//vgICAvT88883REoAoEG5w8/ha7311lsaM2aM/P39df/99+vYsWP69NNPnZWeBuEOc8DvQgD1gaKzCleuXFFQUJC2bNmiY8eOqV+/fpo2bZr9Wci77rpLn3/+eaX95s2bp1mzZmny5Mn68ccf9cMPP2j27Nl17jd79my9+OKLWrVqlc6cOaMRI0booYce0ldffaVp06ZpxowZKikpkb+/v1JTU5WVlaXOnTvr0UcfVVlZWf0nph5V16DjvQAAIABJREFUlcOanlNCQoImTpyoixcvavr06Tp27Jj27Nmjy5cv61e/+pVOnDgh6eov5AcffNBhjHXr1tlvzaqwevVqjRgxQuHh4TLG2J8TKi8vl3T16wEAmhp3+DlcITs7W4cOHdIvf/lLSVJsbKz+P3v3HhdVtfcP/DNc5wIDgyKQXOTiXbRMS0kf89hjqUdTwcS0J618qV28G2pqXsvbIz5eyOPJ6BwtBdSjJllmPWq+Ul521DTMayreQRS5wwjf3x/9mKdxABmcYWbg83695g/2Xnuv716LWWuvvfesDTj+I7aOUAfsC4nIEjjorEJUVBS8vLyg0+kQGxuLkpISXLp0qcr0RUVFiI+PxwsvvIDJkyfD29sbKpUKPj4+tUpXXFyMhIQEDBo0CNHR0fD29sasWbPg6uqKxMREk1i1Wi18fX0RGxuLgoICZGRkWK4wbKAmx+Ti4oLWrVvD3d0dbdq0QUJCAvLy8kzKp7b69euH3NxczJ492yL7IyJyJPbQDlf48ssv8de//hXOzs4AgAEDBsDd3R0pKSkoKiqyaF72xB7qgH0hEVkCB5014OrqCgDV3j28cOECCgsL0atXr2r3VdN0Z8+eRWFhIdq1a2dYplKp4O/vjzNnzlS5nZub2yNjdTQ1PaZOnTpBrVZXWz5ERGQ+W7fDFY/WVtBqtejduzdyc3Oxc+dOi+Zlr2xdB0REj4ODTgu5du0aAMDX19ci6QoKCgAAs2bNgkKhMHyuXLliN688sUfu7u7IysqydRhERA2WpdvhX3/9FadOnUL//v2N+sOK91o6+iO21sC+kIjsDQedFqJUKgEAJSUlFklXMSiNj4+HiBh9Dh8+bIGI6x+9Xo+cnBwEBgbaOhQiogbJGu3wF198gWHDhpn0hXfv3oVKpcLevXtx69Yti+Xn6NgXEpE94qDTQtq1awcnJyccOHDAIumCgoKgVCpx4sQJS4ZZr+3fvx8igi5duhiWubi41KtHjYmI7Jml22ERwZYtW/DOO++YrNPpdBgyZAjKysrw5Zdf1jrm+oZ9IRHZIw46LcTX1xfR0dHYunUrNmzYgNzcXJw8eRLr16+vVTqlUolRo0Zh8+bNSEhIQG5uLsrKynDt2jXcvHmzLg/NbpWXl+PevXt48OABTp48iYkTJyI4OBgjR440pImIiMDdu3exY8cO6PV6ZGVl4cqVKyb78vHxwY0bN3D58mXk5eVBr9djz549nCaeiKga1m6Hf/rpJ2i1Wjz33HOV5j9u3DgADfsRW/aFROQQpJ4DIElJSTVKu2TJElGpVAJAmjdvLhcvXpRNmzaJTqcTABIYGCi//vqrrF69Wvz9/QWAqNVqGTBggIiI5OXlyVtvvSWNGjUSDw8P6datm8yZM8ew7S+//GJWupKSEomLi5Pg4GBxcXERX19fiY6OlvT0dFm7dq2o1WqjWNevXy9arVYASEhIiJw7d05ERA4fPizPPfecBAQECAABIP7+/hIVFSUHDhwwu0xjYmIkJibG7O3+rLIyNOeYxowZI66urtK0aVNxcXERrVYrAwcOlIsXLxrlk52dLT179hSlUimhoaHy3nvvybRp0wSARERESEZGhoiIHDt2TEJCQkSlUkm3bt3k1q1b8vXXX4unp6csXLjwsY5VRCQpKUkawNfNbrH8HYcl2hdHVpfHb+/t8F//+lfRaDTi4uIiHTp0kGPHjhntd8GCBUb9WtOmTWXt2rV1UnYi5p1fVMXe68DSfWFD/347GtaXY7H3+lKIiNTR+NYmFAoFkpKSTN5RRbVX8X60lJQUm8UwduxYpKSkIDs722YxmCM5ORlDhw5FPf+62S2Wv+Owh/bFlhzp+B2tHbY0ezi/cLQ6cKT/b2J9ORp7ry8+XksOq+KF1UREZBtsh22PdUBEjoCDTiIHsG/fPsyYMQPbtm1DWFiY4ZUBr732mkna3r17w9PTE87Ozmjbti2OHTtmg4jNo9fr8dFHHyEiIgJubm7w9vZGu3btcPnyZQDArl27sGTJEpudXNXn8l+yZAlatWoFlUoFjUaDVq1aYfbs2cjNzTVJe+jQITz33HNQq9UICAhAXFyc0Uzctq4nIqp/xo4da/SqnBEjRpikqc9t9MKFC42Ov+Lz5/e4V9X27tixw2ibxo0bWz1e1pdj1Vdd4qCTHM7MmTORmJiI+/fvIzQ0FFu3brV1SFb14YcfYtWqVZg5cyaio6Px+++/Izw8HI0aNcKmTZuQmppqlH7v3r1ISUlB//79kZ6ejo4dO9oo8pobOnQo/vnPf+KLL75AYWEhfvvtN4SHhyM/Px8AMGDAACiVSvTq1Qs5OTl1Glt9L/8ff/wRo0ePRkZGBm7fvo0FCxZgyZIliImJMUqXnp6O3r17o1evXsjKysL27dvx2WefGSZyAWxbT1S3Glo7bI8aUh34+Phgz549OHv2LDZs2GC0rr630TVRVdv78ssv49q1azh48CD69u1bZ/Gwvqpnb/VVVzjoJIfz0UcfoaSkBCKCS5cumZwc1yeLFy/Gli1bkJycDE9PT6N1q1atgpOTE8aMGYP79+/bKMLHt2XLFuzYsQMpKSl49tln4eLigoCAAOzcudPoyuCECRPQoUMH9O3bFw8ePKiT2BpC+bu5ueGdd96Br68vPDw8MGTIEAwcOBDfffed0UzZCxYsgL+/P+bNmweNRoOuXbsiLi4On3/+Oc6cOWNIZ4t6orrXkNphe9WQ6kClUuGll15CixYt4O7ubljeENpoANi4caPJe2p//fVXozSVtb0KhQJNmzZF9+7d0bx58zqLl/XlWPVVVzjoJLJTFy5cwOzZszFv3jwolUqT9VFRUZg4cSKuX7+OqVOn2iBCy/jkk0/QsWNHREZGPjLt3LlzceLECaxcudLqcTWU8t++fbvJ8TVt2hQADHeaHzx4gNTUVPTo0QMKhcKQrk+fPhAR7Ny502j7uqwnImqYGkobbQ57bntZX6bsub6sgYNOIju1atUqiAgGDBhQZZqFCxeiRYsW+PTTT7Fv375q9yciWLFiBVq3bg13d3fodDoMHDjQ6C5VQkICNBoN1Go1du7ciT59+kCr1SIwMBCbN2822l9ZWRnmzJmD4OBgqFQqtG/fHklJSWYdY2lpKY4cOYInn3yyRul1Oh169OiBlStXWn0m2oZQ/lU5f/48vL29ERISAgD4/fffkZ+fj+DgYKN04eHhAICTJ08aLa/LeiKihqkht9FVsee2l/Vlyp7ryxo46CSyU6mpqWjZsiXUanWVaVQqFT7//HM4OTlh9OjRKCgoqDLt3LlzMWPGDHzwwQfIzMzEwYMHcfXqVXTv3h23b98GALz99tuYNGkSioqK4OnpiaSkJFy8eBFhYWEYPXo09Hq9YX/Tp0/H0qVLER8fj5s3b6J///549dVX8fPPP9f4GG/cuIHS0lL8+9//Rs+ePREQEAClUonWrVtj7dq1lTbCTz31FK5fv45ffvmlxvnURkMo/z/T6/W4fv061qxZg3379mH16tVwc3MDANy6dQsATB6HUiqVUKlUhvj/rK7qiYgapobURs+YMQM6nQ5ubm4IDQ3FwIEDcfTo0UrT2mvby/pyrPqyBg46iexQQUEBLl26ZLiTVJ2uXbti0qRJuHz5MqZPn15pmqKiIqxYsQKDBw/GiBEj4OXlhcjISKxbtw537tzB+vXrTbaJioqCVquFr68vYmNjUVBQgIyMDABAcXExEhISMGjQIERHR8Pb2xuzZs2Cq6srEhMTa3ycFY9v+vr6YtGiRUhPT8ft27cxcOBAvPvuu/jyyy9Ntqn4ncOpU6dqnI+5Gkr5/1lQUBACAwMxd+5cLF26FEOHDjWsq5ih1tnZ2WQ7V1dXFBUVmSyvi3oiooapIbXRr7/+Onbt2oWrV68iPz8fmzdvRkZGBnr06IH09HST9PbY9rK+HKu+rMXF1gHUhfj4eLt9UaojOnLkCID/ewktPdq1a9fMSp+ZmQkRqfaK4J8tXLgQu3fvxtq1a40GCxXS09ORn5+PTp06GS3v3Lkz3NzckJaWVu3+K+54VVwVPHv2LAoLC40m+lGpVPD39zd6tOVRKiYYaNu2LaKiogzL582bh08++QTr16/H8OHDjbapKJPK7q5ZSkMp/z+7evUqcnJycPz4ccyYMQPr16/HDz/8gCZNmhh+f1PZxEClpaVQqVQmy+uinuqrI0eOsH11EDy/MM+RI0fQpUuXx95PQ2qjg4KCEBQUZPi7S5cuSExMxJNPPom1a9ciISHBKL09tr2sL8eqL2vhnU4iO1RcXAwARrO+VUepVCIxMREKhQJvvPGGyZ2niim5PTw8TLb19vZGXl6eWfFVPPIya9Yso3dKXblyBYWFhTXeT0BAAADgzp07Rsvd3NwQEhKCixcvmmxTMcCpKCNraCjl/2eurq7w9fVF7969sWXLFqSnp+Ojjz4CAPj7+wOAybs7CwsLUVxcbKjHP6uLeiKihqkhttF/FhkZCWdnZ5w7d85knT22vawvx6ova2kQdzonTZqEV155xdZh1BsVV+B5dbfmkpOTK71aV5WKRujhFwdXp2vXrpg8eTKWL1+OBQsWGE364u3tDQCVNsQ5OTkIDAyscT7AH4/DAn9c5Z84caJZ2/6Zh4cHmjdvjtOnT5use/DgAby8vEyWl5aWAkCld9cspaGUf1UiIiLg7OxseBQoNDQUnp6euHLlilG6CxcuAADat29vso+6qKf6qkuXLmxfHYBCoeD5hZksdQe/obfR5eXlKC8vr3QQZ49tL+vLserLWnink8gONWnSBAqFwux3VS1YsACtWrXC8ePHjZa3a9cOHh4eJj+IT0tLQ2lpKZ5++mmz8gkKCoJSqcSJEyfM2q4yQ4cOxfHjx/H7778blhUWFuLKlSuVvkalokz8/PweO++qNJTyz87Oxquvvmqy/Pz58ygrKzM8IuTi4oK+ffvi4MGDKC8vN6Tbs2cPFApFpbMR1kU9EVHD1FDaaAB48cUXTZYdPXoUIoKuXbuarLPHtpf15Vj1ZS0cdBLZIbVajbCwMLN/C1rxSMrDE74olUpMmTIF27dvx6ZNm5Cbm4tTp05h3LhxCAgIwJgxY8zOZ9SoUdi8eTMSEhKQm5uLsrIyXLt2DTdv3gQAxMbGws/PD8eOHat2X5MnT0ZISAhGjhyJjIwMZGdnIy4uDkVFRZVOIlBRJjV5r2dtNZTy12g02Lt3L3744Qfk5uZCr9fj+PHjeP3116HRaDB58mRD2tmzZ+P27dv48MMPUVBQgMOHD2PZsmUYOXIkWrZsabLvuqgnImqYGkobDQDXr1/Hli1bkJOTA71ej8OHD+Ott95CcHAwxo0bZ5LeHtte1pdj1ZfVSD0HQJKSkmwdRr0SExMjMTExtg7DoSQlJYm5X7fx48eLq6urFBYWGpZt375dwsPDBYA0btxY3n333Uq3nTZtmrz88stGy8rLy2XZsmXSvHlzcXV1FZ1OJ4MGDZKzZ88a0qxdu1bUarUAkObNm8vFixdl/fr1otVqBYCEhITIuXPnRESkpKRE4uLiJDg4WFxcXMTX11eio6MlPT1dREQGDRokAGTOnDmPPNarV6/KsGHDRKfTibu7uzzzzDOyZ8+eStP269dPmjZtKuXl5Y/cbwWWf9UGDBggoaGh4uHhIe7u7hIeHi6xsbFy6tQpk7QHDhyQZ555Rtzd3SUgIECmTZsmxcXFle63NvUkwvaloR+/I+H5hflq8/89ZswYadq0qcnyhtJGT5kyRcLDw0Wj0YiLi4sEBgbK6NGj5caNG5Wmr6rtnTBhgjRq1KjavB7G+qr/9VWXOOgks9n7P7U9qs2g5/z58+Li4iIbN260UlTWVVZWJt27d5cNGzZYbJ937twRpVIpy5cvN2s7ln/dqm09ibB9aejH70h4fmE+Sw5i2Eabqq7ttfWgk/Vlyh7qqy7x8VoiOxUREYH58+dj/vz5hvdZOoqysjLs2LEDeXl5iI2Ntdh+586diyeffBLjx4+32D6rwvKvvbqsJyKq/4qKivDtt9/i/PnzholX2EaberjtFRHcuHEDhw4dMkz8VhdYXzVjL/VVVzjoNMO2bdsQFhZmNJ2yQqGAm5sbmjRpgueffx7Lli3DvXv3bB0q1RMzZszAkCFDEBsba/YP8G1p//792LZtG/bs2VPj93I9yooVK3DixAl8/fXXcHV1tcg+H4Xlbz5b1FNDduTIEbRu3RpOTk5QKBTw8/PDwoULbR2WkYf7Tn9/f4wYMcLWYZEDuXv3Ll566SW0aNECb7zxhmE52+j/U1nbu3PnTjRt2hTdu3dHamqqRfKpCdbXo9lTfdUVhYiIrYOwJoVCgaSkJItOaR4REYE7d+4gJycHIoL79+/jxIkT+Mc//oF//OMf8Pf3x65du0xeWltf8JUp5qt4ZUptv24Vk70sXrzYwpE5hp07d+L06dN4//33TSYUqAmWf9143HoC2L7U9vhfeuklfPvtt7h3757hdQL25s99Z31gjfOL+s5a3++G3kZbou2tDOvLOhytviyFdzofk0KhgLe3N55//nkkJiYiOTkZt2/fRr9+/RzqKo6jKSoqQlRUlMPnUVO9e/dusI0zALz88suYMWOGRRtnczT08q8pW9cT2Qd7ajvrs4bWD1anobfRjtb2sr4cq74shYNOC4uJicHIkSORmZmJdevW2TqcemvDhg3IzMx0+DyIiOobtp11g/0gETkSDjqtYOTIkQD+eHF6hbKyMsyZMwfBwcFQqVRo3749kpKSAAAJCQnQaDRQq9XYuXMn+vTpA61Wi8DAQGzevNlo3wcOHMAzzzwDtVoNrVaLyMhI5ObmPjIPWxMRrFixAq1bt4a7uzt0Oh0GDhyIM2fOGNKMHz8ebm5u8Pf3Nyx75513oNFooFAocOfOHQDAxIkTMWXKFFy8eBEKhQIRERFYtWoVlEolmjRpgrFjxyIgIABKpRJRUVFIS0uzSB4A8M0330Cr1WLRokVWLS8iosdR036lrtpOc/34449o06YNvLy8oFQqERkZiW+//RYA8NZbbxl+HxoeHm54cfyoUaOgVqvh5eWFXbt2Aai+X1y6dCnUajU8PT2RmZmJKVOmoGnTpjh79mytYn4U9oNE1KDZbN7cOgIrTGkeHh4uXl5eVa7Pzc0VABIUFGRYNnXqVHF3d5etW7fKvXv3ZObMmeLk5CRHjx4VEZEPPvhAAMj3338v9+/fl8zMTOnevbtoNBopLS0VEZH8/HzRarWyZMkSKSoqklu3bsngwYMlKyurRnlYSm2mZJ4zZ464ubnJxo0bJScnR06ePCkdO3aUxo0by61btwzphg8fLn5+fkbbLlu2TAAYjlNEJDo6WsLDw43SjRkzRjQajZw+fVqKi4slPT1dOnfuLJ6enpKRkWGRPHbv3i2enp4yf/58s46/Nq/sIMth+TsOe5/y3dpqe/wvvviiAJB79+4ZltWkXxGpm7ZT5NF955+lpKTI3Llz5e7du5KdnS1dunQxen1AdHS0ODs7y/Xr1422e/XVV2XXrl2Gv2va906YMEFWr14tgwcPlt9++61GMZp7ftHQ+0ERfr8dDevLsdh7ffFOpxV4enpCoVAgLy8PAFBcXIyEhAQMGjQI0dHR8Pb2xqxZs+Dq6orExESjbaOioqDVauHr64vY2FgUFBQgIyMDAHD58mXk5uaibdu2UCqV8PPzw7Zt29C4cWOz8qhrRUVFWLFiBQYPHowRI0bAy8sLkZGRWLduHe7cuYP169dbLC8XFxfDVeQ2bdogISEBeXl5FiuDfv36ITc3F7Nnz7bI/oiIrK26fqWCtdtOc8XExODDDz+ETqeDj48PBgwYgOzsbGRlZQEAxo0bh7KyMqP4cnNzcfToUfTt2xeAeX3v4sWL8e6772Lbtm1o1aqVxY+H/SARNXQcdFpBQUEBRARarRYAcPbsWRQWFqJdu3aGNCqVCv7+/kaP1TzMzc0NAKDX6wEAYWFhaNKkCUaMGIG5c+fi8uXLhrS1zaMupKenIz8/32Q2386dO8PNzc3osR9L69SpE9Rqtc3LgIjIHjzcr1TF3trOilcKlJWVAQD+8pe/oEWLFvjss88Ms1Jv2bIFsbGxhsk57KlfZD9IRA0dB51WcO7cOQAwXC0tKCgAAMyaNcvo/Z5XrlxBYWFhjferUqnwww8/oFu3bli0aBHCwsIQGxuLoqIii+VhDRXT43t4eJis8/b2NtwRthZ3d3fD1XEiIqoZW7adqampeP755+Hr6wt3d3e8//77RusVCgXGjh2L33//Hd9//z0A4J///CfefPNNQxp76hfZDxJRQ8dBpxV88803AIA+ffoAAHx9fQEA8fHxEBGjz+HDh83ad9u2bfHVV1/hxo0biIuLQ1JSEpYvX27RPCyt4p1xlXWqOTk5CAwMtFreer3e6nkQEdU3dd12Hjx4EPHx8QCAjIwMDBo0CP7+/khLS8P9+/exZMkSk21GjhwJpVKJTz/9FGfPnoVWq0VISIhhvT31i+wHiaihc7F1APXNrVu3EB8fj8DAQLzxxhsAgKCgICiVSpw4ceKx9n3jxg3k5OSgTZs28PX1xccff4y9e/fi9OnTFsvDGtq1awcPDw/8/PPPRsvT0tJQWlqKp59+2rDMxcXlkY99mWP//v0QEXTp0sVqeRAR1Td13Xb++9//hkajAQCcOnUKer0eb7/9NsLCwgD8cWfzYTqdDkOHDsWWLVvg6emJ0aNHG623p36R/SARNXS801lLIoL8/HyUl5dDRJCVlYWkpCQ899xzcHZ2xo4dOwy/6VQqlRg1ahQ2b96MhIQE5ObmoqysDNeuXcPNmzdrnOeNGzcwduxYnDlzBqWlpTh+/DiuXLmCLl26WCwPa1AqlZgyZQq2b9+OTZs2ITc3F6dOncK4ceMQEBCAMWPGGNJGRETg7t272LFjB/R6PbKysnDlyhWTffr4+ODGjRu4fPky8vLyDJ1neXk57t27hwcPHuDkyZOYOHEigoODDa+xedw89uzZw6niiajesXbbWRW9Xo/bt29j//79hkFncHAwAGDfvn0oLi7G+fPnq/zN47hx41BSUoLdu3ejf//+RuvsqV9kP0hEDV4dz5Zb52DBV6bs2rVL2rdvL2q1Wtzc3MTJyUkAiEKhEG9vb3nmmWdk/vz5kp2dbbJtSUmJxMXFSXBwsLi4uIivr69ER0dLenq6rF27VtRqtQCQ5s2by8WLF2X9+vWi1WoFgISEhMi5c+fk8uXLEhUVJTqdTpydneWJJ56QDz74QB48ePDIPCypNlMyl5eXy7Jly6R58+bi6uoqOp1OBg0aJGfPnjVKl52dLT179hSlUimhoaHy3nvvybRp0wSAREREGKZ8P3bsmISEhIhKpZJu3brJrVu3ZMyYMeLq6ipNmzYVFxcX0Wq1MnDgQLl48aLF8vj666/F09NTFi5caNbx85UdtsXydxz2PuW7tZl7/EeOHJG2bdsa+iN/f39ZtGhRjfsVEbF62/nJJ59IeHi4AKj2s337dkNecXFx4uPjI97e3jJkyBBZs2aNAJDw8HCjV3+IiDz11FMyY8aMSsunun5xyZIlolKpDK8427hxY43LXcT884uG3g+K8PvtaFhfjsXe60sh8v+nfaunFAoFkpKS8Morr9g6lHpjyJAhAICUlBQbR2Js7NixSElJQXZ2tq1DMZGcnIyhQ4einn/d7BbL33HYa/tSV2xx/PbcdtZEv379sGbNGoSGhtZpvvZ4fmHvddnQv9+OhvXlWOy9vvh4LdUrFdPpExFRzTlS2/nnx3VPnjwJpVJZ5wNOe+ZIdUlEDQcnEiIiIiKHERcXh3HjxkFEMGrUKGzcuNHWIRER0SPwTifVCzNnzkRiYiLu37+P0NBQbN261dYhERHZPUdsO9VqNVq1aoUXXngBc+fORZs2bWwdkl1wxLokooaDg06qFz766COUlJRARHDp0iXExMTYOiQiIrvniG3nwoULUVZWhoyMDJMZaxsyR6xLImo4OOgkIiIiIiIiq+Ggk4iIiIiIiKyGg04iIiIiIiKyGg46iYiIiIiIyGo46CQiIiIiIiKrUYiI2DoIa1IoFLYOgYiIzBATE4OUlBRbh2ETQ4YM4asuiIioVuy5/3SxdQDWlpSUZOsQiCzu8OHDWLlyJf+/qV4KCgqydQg2M3nyZAwZMsTWYRDZpfj4eADApEmTbBwJkX2y5/6z3t/pJKqPkpOTMXToUPDrS0REDcUrr7wC4I8+kIgcC3/TSURERERERFbDQScRERERERFZDQedREREREREZDUcdBIREREREZHVcNBJREREREREVsNBJxEREREREVkNB51ERERERERkNRx0EhERERERkdVw0ElERERERERWw0EnERERERERWQ0HnURERERERGQ1HHQSERERERGR1XDQSURERERERFbDQScRERERERFZDQedREREREREZDUcdBIREREREZHVcNBJREREREREVsNBJxEREREREVkNB51ERERERERkNRx0EhERERERkdVw0ElERERERERWw0EnERERERERWQ0HnURERERERGQ1HHQSERERERGR1XDQSURERERERFbDQScRERERERFZDQedREREREREZDUcdBIREREREZHVcNBJREREREREVsNBJxEREREREVkNB51ERERERERkNRx0EhERERERkdW42DoAIqpeVlYW/vWvfxkt+/nnnwEA69evN1ru6emJYcOG1VlsRERE1pCWloZffvnFaNnvv/8OwLTv69ChA5599tk6i42IzKcQEbF1EERUtZKSEjRp0gT5+flwdnYGAFR8bRUKhSGdXq/H66+/js8//9wWYRIREVnM7t270b9/fzg7O8PJ6Y8H8x7u+8rLy1FWVoavvvoKf/3rX20WKxE9GgedRA7gzTffxKZNm1BaWlptum+//Ra9e/euo6jxoQTQAAAgAElEQVSIiIisQ6/Xo3HjxsjNza02nVarRVZWFtzc3OooMiKqDf6mk8gBvPrqq48ccHp7e+Mvf/lLHUVERERkPa6urhg2bFi1g8mapCEi+8BBJ5ED6NmzJ3x9fatc7+rqihEjRsDFhT/TJiKi+mHYsGHVXnDV6/V49dVX6zAiIqotDjqJHICTkxOGDx8OV1fXStfr9XpOIERERPVK9+7d4efnV+V6X19fdOvWrQ4jIqLa4qCTyEEMGzYMer2+0nVPPPEEunbtWscRERERWY+TkxNee+21Sh+fdXNzw8iRIw2TDBGRfeM3lchBPPPMMwgJCTFZ7ubmhtdff91oJlsiIqL6oKpHbEtLS/mED5ED4aCTyIG89tprJo/YsuMlIqL6qmPHjoiIiDBZHhYWhqeeesoGERFRbXDQSeRAhg8fbvKIbUREBCIjI20UERERkXWNGDHC6IJrxRM+ROQ4OOgkciCtWrVCmzZtDI/Surq6YtSoUTaOioiIyHoevuBaWlqK2NhYG0ZERObioJPIwfzXf/0XnJ2dAQAPHjzgo7VERFSvRUREoEOHDlAoFFAoFOjQoQNatGhh67CIyAwcdBI5mGHDhqGsrAzAH791CQ0NtXFERERE1lVxwdXZ2Rn/9V//ZetwiMhMHHQSOZjg4GA8++yzAMDftBARUYMQGxuL8vJylJWVYejQobYOh4jM5GLrACzl8OHDWLFiha3DIKoTJSUlUCgU2Lt3Lw4ePGjrcIjqREpKitX2vWLFChw+fNhq+yeix+fj4wMAmDhxoo0jIaLqdO3aFZMnTzZaVm/udF69ehVbt261dRj1ztatW3Ht2jVbh+FQjhw5giNHjlg1j8DAQPj5+UGpVFo1n/qO/9+O4dq1a1Zv3w8fPmz17y2RI7KndjI4OLjS91Xbm7o4D6Cqsfxt68iRI5VexK03dzorWPNKeEOkUCgwadIkvPLKK7YOxWEMGTIEgPX/Fy9cuFDpu8uo5vj/7RiSk5Pr5HG6Ll26sA8heog9tZN3794F8H93PO1VXZ0HUOVY/rZVUf4Pq3eDTqKGggNOIiJqSOx9sElEVas3j9cSERERERGR/eGgk4iIiIiIiKyGg04iIiIiIiKyGg46iYiIiIiIyGoa7KBz1KhRUCqVUCgUKC4urjbt119/DS8vL3z11VdVpnnrrbfg6ekJhUKBEydOPHa62po/fz7atGkDrVYLd3d3RERE4P3330d+fr7F8zJHTcqQiMhRdO7cGc7OznjyySfrPG9z+q+GYvny5WjSpAkUCgXWrVtnWF5XfY899nHl5eWIj49HVFSUrUMBYJ9lRER1p8EOOhMTEzF16tQapRWRR6b59NNP8fe//91i6Wrrhx9+wLvvvovLly/jzp07+Oijj7By5coqpy+uKzUpQyIiR3H06FH07NnTJnmb0381FFOnTsVPP/1ksryu+h576+POnz+P//iP/8DkyZNRWFho63AA2F8ZEVHd4itTaqBfv364f/++rcOoEQ8PD4wZMwbOzs4AgFdeeQXbtm1DcnIyrl69iqCgIJvEZU9lWFRUhF69elV6gkJEZA6FQvHY+2CbZD3W6Hsqqy976uN++eUXzJ8/H+PGjUNBQYHdDPbsqYz4nSOqew32TuefWeKkwZz9WCq/yuzevdsw4KzQuHFjALCbq522tmHDBmRmZto6DCKqB1xdXR97H4/TJlmzP6HK2Xsf0qFDB2zbtg3Dhw+Hu7u7rcOxS/Zeh0T1UYMfdDo5OSE1NRV9+vSBl5cXAgIC8NlnnxnWHzp0CMHBwVAoFFizZo1huYhg2bJlaNmyJdzd3eHl5YVp06aZ7L+m6crKyjBnzhwEBwdDpVKhffv2SEpKAgAkJCRAo9FArVZj586d6NOnD7RaLQIDA7F58+ZHHuP169ehUqkQGhpamyJ6bJWVYU2PadWqVVAqlWjSpAnGjh2LgIAAKJVKREVFIS0tzZBu/PjxcHNzg7+/v2HZO++8A41GA4VCgTt37gAAJk6ciClTpuDixYtQKBSIiIgAAHzzzTfQarVYtGhRXRQJEdUTFy5cQKtWraDRaKBSqdC9e3ccOnTIKM2PP/6INm3awMvLC0qlEpGRkfj2228BVN0mAcDGjRvRqVMnKJVKaDQaNGvWDAsWLDCsf1T/VVPm9DEighUrVqB169Zwd3eHTqfDwIEDcebMGUOapUuXQq1Ww9PTE5mZmZgyZQqaNm2KcePGQaPRwMnJCU8//TT8/Pzg6uoKjUaDjh07onv37ggKCoJSqYS3tzfef//9GpdjZSrrey5cuACFQlHp57vvvqtVfVV3nvCosnrc/t1R8DyAiCD1RFJSkph7OB988IEAkO+//15ycnLk7t270rdvX3F3d5eCggJDuqtXrwoAWb16tdG2CoVC/vu//1vu3bsnhYWFsnbtWgEgx48fNzvd1KlTxd3dXbZu3Sr37t2TmTNnipOTkxw9etQk1vv370tmZqZ0795dNBqNlJaWVnmMBQUF4unpKePHjzerbCoAkKSkpFpt+2dVlWFNjmnMmDGi0Wjk9OnTUlxcLOnp6dK5c2fx9PSUjIwMQ7rhw4eLn5+fUb7Lli0TAJKVlWVYFh0dLeHh4Ubpdu/eLZ6enjJ//vzHPtaYmBiJiYl57P2Q9Vnq/5usqzbtu7lq873t1auXhIWFyaVLl0Sv18uvv/4qzz77rCiVSjl37pwhXUpKisydO1fu3r0r2dnZ0qVLF2nUqJFhfWVtUnx8vACQjz/+WLKzs+Xu3bvyt7/9TYYPHy4iNe+/aqqm7fGcOXPEzc1NNm7cKDk5OXLy5Enp2LGjNG7cWG7dumWyvwkTJsjq1atl8ODB8ttvv8mHH34oACQtLU0KCgrkzp078tJLLwkASU1NlaysLCkoKJDx48cLADlx4kSNy/H8+fMCQD755BPDsof7nvPnz8v06dMNZXTz5k3R6XQSFRUlZWVlta6vyvo4c8vK3P69Jp599lnp0KHDY+2D5wHm43mAbbH8bauq8m/wdzoBICoqCl5eXtDpdIiNjUVJSQkuXbpUZfqioiLEx8fjhRdewOTJk+Ht7Q2VSgUfH59apSsuLkZCQgIGDRqE6OhoeHt7Y9asWXB1dUViYqJJrFqtFr6+voiNjUVBQQEyMjKqjPWjjz5CQEAAFi5cWIuSqRs1OSYXFxfD1eI2bdogISEBeXl5JuVTW/369UNubi5mz55tkf0RUcPg6emJZs2awcXFBW3btsXf//53FBcXY/369YY0MTEx+PDDD6HT6eDj44MBAwYgOzsbWVlZle5Tr9dj3rx56NmzJ6ZPnw4fHx/odDq8+eab6Ny5s1Fac/uvR6muPS4qKsKKFSswePBgjBgxAl5eXoiMjMS6detw584do2OusHjxYrz77rvYtm0bWrVqZVjepk0bqNVqNGrUCMOGDQMABAcHo3HjxlCr1RgxYgQAGN0VNLccKxMREYGPP/4YarUaAPD222+jqKgIn332GZycnCyWT23Kytz+vT7heQBR/ceJhB5S8fscvV5fZZoLFy6gsLAQvXr1qnZfNU139uxZFBYWol27doZlKpUK/v7+Rh3uw9zc3KqNdfv27UhOTsbevXvh6elZbQz24lHHVKFTp05Qq9XVlg8RUV2LjIyEl5cXTp48WWWain6mrKys0vUnT55ETk4OXnzxRaPlzs7OmDBhwiP3+6j2s6Yebo/T09ORn5+PTp06GaXr3Lkz3NzcjB51rE0+Dx48MCyrybE8qhwfJTk5Gf/617+wZMkStGzZ0qL5PG5Z1bQvrI94HkBUP/FOZy1cu3YNAODr62uRdAUFBQCAWbNmGf2+5MqVK7We/GfLli1YvHgx9u/fj2bNmtVqH/bO3d3drCvPRER1wdXV1eiEOTU1Fc8//zx8fX3h7u5u8lvFh+Xm5gIAvL29rRqnuXJycgD8MUv6w7y9vZGXl2fV/M0tx+pkZ2fjvffeQ+fOnTFlyhSL52PrsmooeB5A5Dg46KwFpVIJACgpKbFIuopBaXx8PETE6HP48GGz41u9ejU2bdqEH374AU888YTZ2zsCvV6PnJwcBAYG2joUIiKDBw8e4O7duwgODgYAZGRkYNCgQfD390daWhru37+PJUuWVLuPina7YuITe1ExCK5swGTt9rg25VidCRMmICcnB4mJiUYzvlsqH1uWVUPB8wAix8JBZy20a9cOTk5OOHDggEXSVczWd+LEiceKS0QQFxeHU6dOYceOHZVeYa0v9u/fDxFBly5dDMtcXFwa5KNIRGQ//vd//xfl5eXo2LEjAODUqVPQ6/V4++23ERYWBqVS+cjXnDRr1gw+Pj7Yu3dvXYRcY+3atYOHhwd+/vlno+VpaWkoLS3F008/bbW8a1OOVUlNTcUXX3yB2bNno23btobl06ZNs1g+tiyrhoLnAUSOhYPOWvD19UV0dDS2bt2KDRs2IDc3FydPnjSZGKCm6ZRKJUaNGoXNmzcjISEBubm5KCsrw7Vr13Dz5s0ax3X69GksXboUf//73+Hq6moyHfzy5cstcvy2UF5ejnv37uHBgwc4efIkJk6ciODgYIwcOdKQJiIiAnfv3sWOHTug1+uRlZWFK1eumOzLx8cHN27cwOXLl5GXlwe9Xo89e/ZwqnQiMltpaSnu37+PBw8e4NixYxg/fjxCQkIMbVPFHc99+/ahuLgY58+fN/k938NtkpOTE2bOnImDBw9i/PjxuH79OsrLy5GXl4fTp0/X9SEaKJVKTJkyBdu3b8emTZuQm5uLU6dOYdy4cQgICMCYMWOslndNyrEmcnNzMXbsWDz55JOYPn06gD8m8/v5559x4sSJWtVXZYMcW5ZVfcXzACIHV5dT6FqTuVPqL1myRFQqlQCQ5s2by8WLF2XTpk2i0+kEgAQGBsqvv/4qq1evFn9/fwEgarVaBgwYICIieXl58tZbb0mjRo3Ew8NDunXrJnPmzDFs+8svv5iVrqSkROLi4iQ4OFhcXFzE19dXoqOjJT09XdauXStqtdoo1vXr14tWqxUAEhISIufOnZNTp04JgCo/y5YtM7tcYYGp0isrw5oek8gfU6W7urpK06ZNxcXFRbRarQwcOFAuXrxolE92drb07NlTlEqlhIaGynvvvSfTpk0TABIREWGYVv3YsWMSEhIiKpVKunXrJrdu3ZKvv/5aPD09ZeHChY91rCKcqtuRWOL/m6zPXl+ZkpiYKD179pQmTZqIi4uLNGrUSIYNGyZXrlwxShcXFyc+Pj7i7e0tQ4YMkTVr1ggACQ8Pl4yMjErbJBGRNWvWSGRkpCiVSlEqlfLUU0/J2rVra9x/1ZQ57XF5ebksW7ZMmjdvLq6urqLT6WTQoEFy9uxZw/7+HF9QUJBs3LhRRERWrlxpyKdZs2by448/yuLFi8XLy0sAiJ+fn3zxxReyZcsW8fPzEwCi0+lk8+bNjyzHiRMnGrbRaDQyePDgSvue5cuXV9lH9u3bt1b1NWvWrErPE2pSVuaUfU0dPnxYnnvuOQkICDAcm7+/v0RFRcmBAwfM2pcIzwNqg+cBtsXyt62qyl8hImLhcaxNJCcnY+jQoagnh2M3FAoFkpKS8Morr9gshrFjxyIlJQXZ2dk2i8EcQ4YMAQCkpKTYOBJ6FHv4/6ZHq4v2nd9bosrZQzvJ8wAyB8vftqoqfz5eSw6htlPiExERkePjeQCRY+Ogk4iIqJ45c+aMye/6K/vExsbaOlR6BNYlEdUHHHSSXZs5cyYSExNx//59hIaGYuvWrbYOyer27duHGTNmYNu2bQgLCzOcULz22msmaXv37g1PT084Ozujbdu2OHbsmA0irrklS5agVatWUKlU0Gg0aNWqFWbPnm14L+GfHTp0CM899xzUajUCAgIQFxdn9PqhXbt2YcmSJTa7+s16+oO911ND1apVK5NXcFX22bJli61DpUdo6HXJ84D61b8sXLiw0osm7dq1M6Sxdb/B8rdS+Vv1l6R1qC4mmmiIwIlWzPY4P2CfM2eO9O/fX3Jzcw3LwsPDpVGjRgJAdu/ebbLNnj175OWXX651vHWpX79+snz5csnMzJS8vDxJTk4WV1dX+c///E+jdL/++quoVCqZPXu25Ofny08//SSNGzeWUaNGGaVbuXKl9OjRQ+7du1ereGr7/816+kNd1ZO9TiRE1BDwPMB8PA+o2oIFCyqdyKtt27ZG6R6n32D5V82W5c87nUR2YvHixdiyZQuSk5Ph6elptG7VqlVwcnLCmDFjcP/+fRtF+Pjc3NzwzjvvwNfXFx4eHhgyZAgGDhyI7777zuj1QAsWLIC/vz/mzZsHjUaDrl27Ii4uDp9//jnOnDljSDdhwgR06NABffv2xYMHD+rkGFhPjlFPRESOpiH0LwCwceNGkzv1v/76q1Ea9u/WY6vy56CTyA5cuHABs2fPxrx586BUKk3WR0VFYeLEibh+/TqmTp1qgwgtY/v27SbH17RpUwBAfn4+AODBgwdITU1Fjx49jF7K3qdPH4gIdu7cabT93LlzceLECaxcudLK0bOeAMeoJyIiR9NQ+hdzsH+3LUuXPwedRHZg1apVEBEMGDCgyjQLFy5EixYt8Omnn2Lfvn3V7k9EsGLFCrRu3Rru7u7Q6XQYOHCg0d2nhIQEaDQaqNVq7Ny5E3369IFWq0VgYCA2b95stL+ysjLMmTMHwcHBUKlUaN++PZKSkh7voP+/8+fPw9vbGyEhIQCA33//Hfn5+YaXtFcIDw8HAJw8edJouU6nQ48ePbBy5UqrvzKJ9eQY9URE5Ggacv9SFfbv/6c+lD8HnUR2IDU1FS1btoRara4yjUqlwueffw4nJyeMHj0aBQUFVaadO3cuZsyYgQ8++ACZmZk4ePAgrl69iu7du+P27dsAgLfffhuTJk1CUVERPD09kZSUhIsXLyIsLAyjR4+GXq837G/69OlYunQp4uPjcfPmTfTv3x+vvvoqfv7551odr16vx/Xr17FmzRrs27cPq1evhpubGwDg1q1bAGDyaItSqYRKpTLE/2dPPfUUrl+/jl9++aVW8dQU68kx6omIyNE0pP5lxowZ0Ol0cHNzQ2hoKAYOHIijR49Wmpb9+x/qQ/lz0ElkYwUFBbh06ZLhDlF1unbtikmTJuHy5cuYPn16pWmKioqwYsUKDB48GCNGjICXlxciIyOxbt063LlzB+vXrzfZJioqClqtFr6+voiNjUVBQQEyMjIAAMXFxUhISMCgQYMQHR0Nb29vzJo1C66urkhMTKzVMQcFBSEwMBBz587F0qVLMXToUMO6iplPnZ2dTbZzdXVFUVGRyfLmzZsDAE6dOlWreGqC9eQY9URE5GgaUv/y+uuvY9euXbh69Sry8/OxefNmZGRkoEePHkhPTzdJz/69/pR/vRt01uRdVvzU/AMAQ4cOtXkcjvQxdzr3zMxMiEi1V9f+bOHChWjZsiXWrl2LQ4cOmaxPT09Hfn4+OnXqZLS8c+fOcHNzQ1paWrX7r7iTVXGF7ezZsygsLDSaTlulUsHf39/oMRFzXL16FZmZmfjyyy/xj3/8A0899RQyMzMBwPBbisp+uF5aWgqVSmWyvKLsKru7ZimsJ8eop8exdetWm7cf/PBjbx+A5wHmfngeULWgoCA89dRT8PDwgJubG7p06YLExEQUFRVh7dq1JunZv9ef8nd57D3YGWs/39zQDB06FBMnTkTXrl1tHYrDiI+PNyt9cXExAMDd3b1G6ZVKJRITE9GtWze88cYbWLJkidH6nJwcAICHh4fJtt7e3sjLyzMrvorHR2bNmoVZs2YZrQsICDBrXxVcXV3h6+uL3r17IzQ0FC1atMBHH32ElStXwt/fHwBM3glZWFiI4uLiSvOsGOBUlKU1sJ4co54eR5cuXTBp0iRbh0FkV3geYD6eB5gnMjISzs7OOHfunMk69u/1p/zr3aDzlVdesXUI9crQoUPRtWtXlqsZUlJSzEpf8YU25yW8Xbt2xeTJk7F8+XIsWLDAaDIXb29vAKi0UcvJyUFgYKBZ8fn6+gL4oxOdOHGiWdvWREREBJydnQ2PdYSGhsLT0xNXrlwxSnfhwgUAQPv27U32UVpaCgCV3l2zFNaTY9TT4wgMDGRbR/QQngeYj+cB5ikvL0d5eXmlgz727/Wn/Ovd47VEjqZJkyZQKBRmv/dpwYIFaNWqFY4fP260vF27dvDw8DD5cXlaWhpKS0vx9NNPm5VPUFAQlEolTpw4YdZ2D8vOzsarr75qsvz8+fMoKytDUFAQAMDFxQV9+/bFwYMHUV5ebki3Z88eKBSKSmeWqyg7Pz+/x4qxOqwnx6gnIiJH01D6FwB48cUXTZYdPXoUIlLp3XT27/Wn/DnoJLIxtVqNsLAwXLt2zaztKh7veHgiF6VSiSlTpmD79u3YtGkTcnNzcerUKYwbNw4BAQEYM2aM2fmMGjUKmzdvRkJCAnJzc1FWVoZr167h5s2bAIDY2Fj4+fnh2LFjVe5Ho9Fg7969+OGHH5Cbmwu9Xo/jx4/j9ddfh0ajweTJkw1pZ8+ejdu3b+PDDz9EQUEBDh8+jGXLlmHkyJFo2bKlyb4ryi4yMtKsYzMH68kx6omIyNE0lP4FAK5fv44tW7YgJycHer0ehw8fxltvvYXg4GCMGzfOJD3793pU/lJPJCUlST06HLsBQJKSkmwdhkOJiYmRmJgYs7YZP368uLq6SmFhoWHZ9u3bJTw8XABI48aN5d13361022nTpsnLL79stKy8vFyWLVsmzZs3F1dXV9HpdDJo0CA5e/asIc3atWtFrVYLAGnevLlcvHhR1q9fL1qtVgBISEiInDt3TkRESkpKJC4uToKDg8XFxUV8fX0lOjpa0tPTRURk0KBBAkDmzJlT7XEOGDBAQkNDxcPDQ9zd3SU8PFxiY2Pl1KlTJmkPHDggzzzzjLi7u0tAQIBMmzZNiouLK91vv379pGnTplJeXl5t/g8z9/+b9WSbeqqL9r0231uihoDnAebjeUDVpkyZIuHh4aLRaMTFxUUCAwNl9OjRcuPGjUrT16bfYPlXzZblX29GaRx0Wgc7G/PVprE7f/68uLi4yMaNG60UlXWVlZVJ9+7dZcOGDXWe9507d0SpVMry5cvN3tbc/2/WU+09Tj1x0ElkOzwPMB/PAyyjtv0Gy98yLF3+fLyWyA5ERERg/vz5mD9/PvLz820djlnKysqwY8cO5OXlITY2ts7znzt3Lp588kmMHz/e6nmxnmqvLuuJiMjRsH8xxf69Zhyl/DnorMK2bdsQFhZm8u4lNzc3NGnSBM8//zyWLVuGe/fu2TpUqidmzJiBIUOGIDY21uwfs9vS/v37sW3bNuzZs6fG77iylBUrVuDEiRP4+uuv4erqWid5sp7MZ4t6sjdV9SkuLi5o3LgxXnjhBWzfvt3qcYwaNQpKpRIKhaLaKfAfjve1114zSdO7d294enrC2dkZbdu2feRviWxt+fLlhglD1q1bZ1j+9ddfw8vLC1999ZVV86+rfMhxsX/5P+zfa85Ryp+DzipER0fj999/R3h4OLy8vCAiKC8vR2ZmJpKTkxEaGoq4uDi0bdvWZHYqotpatGgRxo8fj48//tjWodRYr1698MUXXxje21hXdu7ciZKSEuzfvx86na5O82Y91Zwt68meVNaniAiysrKQlJSE69evIzo62urvmk5MTMTUqVPNirdRo0bYtGkTUlNTjdLs3bsXKSkp6N+/P9LT09GxY0drhW0RU6dOxU8//WSyXETqJP+6yoccG/sX9u/mcpTy56DTDAqFAt7e3nj++eeRmJiI5ORk3L59G/369XOoKyKOpKioCFFRUQ6fhzl69+6NxYsX2zoMu/fyyy9jxowZJrPG1RXWU83Yup7snU6nQ69evfA///M/AIDk5GSztq+L9mvVqlVwcnLCmDFj6mVfV9GH9+/f32L7rKxerJFPQ8DzgIbH1v0Gy9865c9B52OIiYnByJEjkZmZafSoDlnOhg0bkJmZ6fB5EBFVp1mzZgD+eHG4OR6n/VIoFDVKFxUVhYkTJ+L69es1uktK7FcsiecBRPUDB52PaeTIkQD+eCF6hbKyMsyZMwfBwcFQqVRo37694ZGphIQEaDQaqNVq7Ny5E3369IFWq0VgYCA2b95stO8DBw7gmWeegVqthlarRWRkJHJzcx+Zhy2JCFasWIHWrVvD3d0dOp0OAwcOxJkzZwxpxo8fDzc3N6PHAN555x1oNBooFArcuXMHADBx4kRMmTIFFy9ehEKhQEREBFatWgWlUokmTZpg7NixCAgIgFKpRFRUFNLS0iySBwB888030Gq1WLRokVXLi4gIAE6ePAkA6NGjh9HyH3/8EW3atIGXlxeUSiUiIyPx7bffAqi6/QKAjRs3olOnTlAqldBoNGjWrBkWLFhgWO/k5ITU1FT06dMHXl5eCAgIwGeffVZlfAsXLkSLFi3w6aefYt++fdUeS036gaVLl0KtVsPT0xOZmZmYMmUKmjZtinHjxkGj0cDJyQlPP/00/Pz84OrqCo1Gg44dO6J79+6GF6V7e3vj/fffr3F5VebQoUMIDg6GQqHAmjVrAAAXLlww+e1txee7776rVb1Ulk9Ny8qc8wZ7wPMAIqrUY86mazesNaV+eHi4eHl5Vbk+NzdXAEhQUJBh2dSpU8Xd3V22bt0q9+7dk5kzZ4qTk5McPXpUREQ++OADASDff/+93L9/XzIzM6V79+6i0WiktLRURETy8/NFq9XKkiVLpKioSG7duiWDBw+WrKysGuVhKTBzqvQ5c+aIm5ubbNy4UXJycuTkyZPSsWNHady4sdy6dcuQbvjw4eLn52e07bJlywSA4RhFRKKjoyU8PNwo3ZgxY0Sj0cjp06eluLhY0tPTpXPnzuLp6SkZGRkWyWP37t3i6ekp8+fPr/GxV+CrFxyHuf/fZBv16ZUpD/cphQ6X7HIAACAASURBVIWFsmfPHgkJCZHevXtLfn6+UfqUlBSZO3eu3L17V7Kzs6VLly7SqFEjw/rK2q/4+HgBIB9//LFkZ2fL3bt35W9/+5sMHz5cRIz7oJycHLl796707dtX3N3dpaCgwCTeS5cuiYjITz/9JE5OTtKsWTNDnHv27DF5P11N+4GKOCZMmCCrV6+WwYMHy2+//SYffvihAJC0tDQpKCiQO3fuyEsvvSQAJDU1VbKysqSgoEDGjx8vAOTEiRM1Lq/z588LAPnkk08My65evSoAZPXq1YY006dPN5TFzZs3RafTSVRUlJSVldW6Xh7OpzZlVd15g7XwPIDnAY6G5W9bfGWKlXh6ekKhUCAvLw8AUFxcjISEBAwaNAjR0dHw9vbGrFmz4OrqisTERKNto6KioNVq4evri9jYWBQUFCAjIwMAcPnyZeTm5qJt27ZQKpXw8/PDtm3b0LhxY7PyqEtFRUVYsWIFBg8ejBEjRsDLywuRkZFYt24d7ty5g/Xr11ssLxcXF8NV1DZt2iAhIQF5eXkWO/5+/fohNzcXs2fPtsj+iIgq3L9/33DnTK1WG+5cDR8+3GSWwJiYGHz44YfQ6XTw8fHBgAEDkJ2djaysrEr3rdfrMW/ePPTs2RPTp0+Hj48PdDod3nzzTXTu3NkobVRUFLy8vKDT6RAbG4uSkhJcunSpyri7du2KSZMm4fLly5g+fXqlaWrTDyxevBjvvvsutm3bhlatWhmWt2nTBmq1Go0aNcKwYcMAAMHBwWjcuDHUajVGjBgBAEZ30Mwtr8pERETg448/NswC+fbbb6OoqAifffYZnJycLJZPbcqquvMGe8DzACKqCgedj6mgoAAiAq1WCwA4e/YsCgsL0a5dO0MalUoFf39/o47xYW5ubgD+OGEAgLCwMDRp0gQjRozA3LlzcfnyZUPa2uZhbenp6cjPz0enTp2Mlnfu3Blubm5Gj71YWqdOnaBWq216/ERENfHn2Wv1ej2uXbuGSZMmYfz48Wjfvr3hsb/KVAxKy8rKKl1/8uRJ5OTk4MUXXzRa7uzsjAkTJjxyvxV9UFUWLlyIli1bYu3atTh06JDJemv1AxV95IMHD8yK+VHl9SjJycn417/+hXnz5qFly5YWzedxy+rh8wZ7wPMAIqoKB52P6dy5cwBguDpbUFAAAJg1a5bR70CuXLmCwsLCGu9XpVLhhx9+QLdu3bBo0SKEhYUhNjYWRUVFFsvD0iomwPDw8DBZ5+3tbbgbbC3u7u5mXWUmIrI1FxcXNG3aFKNGjcLy5ctx9uxZo6n6U1NT8fzzz8PX1xfu7u4mv2F8WMXv/r29va0Sr1KpRGJiIhQKBd544w0UFRUZrbd1P2BueVUnOzsb7733Hjp37owpU6ZYPB9bl5U12PqYeB5AZL846HxM33zzDQCgT58+AABfX18AQHx8vOFKdsXn8OHDZu27bdu2+Oqrr3Djxg3ExcUhKSkJy5cvt2gellRxklNZp5KTk4PAwECr5a3X662eBxGRNUVGRgIATp8+DQDIyMjAoEGD4O/vj7S0NNy/fx9Lliypdh9PPPEEAFR7t/Rxde3aFZMnT8b58+eNJicCbNsP1Ka8qjNhwgTk5OQgMTHR6NUBlsrHlmVlLTwPIKKqcND5GG7duoX4+HgEBgbijTfeAADDrHonTpx4rH3fuHHDcOLh6+uLjz/+GB07dsTp06ctloeltWvXDh4eHvj555+NlqelpaG0tBRPP/20YZmLi4tFHwnav38/RARdunSxWh5ERNb073//GwAMj3GeOnUKer0eb7/9NsLCwqBUKh/5mpNmzZrBx8cHe/futWqsCxYsQKtWrXD8+HGj5eb0A5ZWm/KqSmpqKr744gvMnj0bbdu2NSyfNm2axfKxZVlZC88DiKgqHHTWgIggPz8f5f+PvTsPa+rK/wf+DksIBBBQFCogAq4Vq3UZpPBTv61a9cEVlKrT2sXBZcSFWkRRca1oC2qFsXQc5qlOLVD9ulTttNax7ox+3bEqogiuIIosggRyfn84ZJqCSjDJBfJ+PQ9/eHPuPZ+ck0De3pt71GoIIZCfn4+UlBS88cYbMDc3x/bt2zXf6VQoFHj//fexZcsWJCYmoqioCFVVVbh58ybu3LlT5z5v376NyZMn49KlS6ioqMDp06dx48YN+Pn56a0PfVMoFIiIiMC2bduwefNmFBUV4fz585gyZQpcXV0RFhamaevj44MHDx5g+/btUKlUyM/Px40bN2oc08nJCbdv30Z2djaKi4s1fzzUajUePnyIyspKnDt3DjNnzoSHh4dmCZuX7WPv3r28VToRGUxZWZnmb8rt27eRnJyM6OhotGjRArNmzQLw9KY5ALBv3z6Ul5cjMzOzxnfifv/7y8zMDPPmzcPBgwcRHh6OW7duQa1Wo7i4WPMfmfpQfZnt7xcP1+XvgL7VZbzqoqioCJMnT0a3bt00N0wqLy/HyZMncebMmXrNS23BR8qxMhR+DiCiZzLCnXONQt+31N+5c6fo2rWrsLGxEXK5XJiZmQkAQiaTCQcHB9G7d2+xZMkSUVBQUGPfJ0+eiMjISOHh4SEsLCyEs7OzGD16tMjIyBAJCQnCxsZGABDt2rUTWVlZIikpSdjb2wsAok2bNuLKlSsiOztb+Pv7C0dHR2Fubi5eeeUVMX/+fFFZWfnCPvQJOt4qXa1Wi9WrV4t27doJS0tL4ejoKEaOHCkuX76s1a6goED0799fKBQK0bZtWzF9+nQxZ84cAUD4+Phobnl+6tQp0aZNG2FtbS0CAgLE3bt3RVhYmLC0tBStW7cWFhYWwt7eXowYMUJkZWXprY89e/YIOzs7sWzZMp3HjLfqbjx0fX2TNJrCkinbtm0T3t7eAkCNHysrK9GuXTsxdepUreUehBAiMjJSODk5CQcHBxESEiLWr18vAAhvb2+Rk5NT6+8vIYRYv3698PX1FQqFQigUCtG9e3eRkJAgYmNjhbW1tdbfoM2bNwtHR0cBQLi5uYkLFy5o1duiRQvx5z//udbnNWfOnBpLptTl78Bv63B3dxebNm0SQgixZs0azd9IT09PcejQIbFy5UrRrFkzAUC0atVK/OMf/xDffvutaNWqlQAgHB0dxZYtW144XjNnztTso1QqxahRo8QXX3whXFxcBABhY2Mjhg0bJj777LNa5wmAGDJkSL3mJTo6ukY/dR2run5uMBR+DuDngMaG4y+tZ42/TAghDBlqjSU1NRVjx45FE3k6DYZMJkNKSgrGjBkjdSkakydPRlpaGgoKCqQupVYhISEAgLS0NIkroRdpiK9vqskYv9/5viWqXUP8PcnPAfQ8HH9pPWv8eXktNUr1vf09ERERNX78HEDUuDB0EhERERERkcEwdFKjMm/ePCQnJ+PRo0do27YtvvvuO6lLIiIiIiPh5wCixslC6gKIdLFixQqsWLFC6jKIiIhIAvwcQNQ48UwnERERERERGQxDJxERERERERkMQycREREREREZDEMnERERERERGUyTu5FQamqq1CU0OceOHZO6hEbl5s2bAPhaNBS1Wg0zM/39fxlf3w2fsebo5s2bfN9So6Dv34Mvwt+TuuHnAGlx/KV18+ZNuLm51dguE0IICerRu9TUVIwdO1bqMoiIyEAM+ecqJCSESy8QERHpQXBwMNLS0rS2NZnQSUSmYdeuXVi8eDFOnTqFoKAgLFq0CK+//rrUZRERGczRo0exdOlS/PDDD/Dz80NMTAwGDRokdVlERHXG73QSUaMSFBSEkydP4scff8Tdu3fRo0cPDBgwAP/+97+lLo2ISK+OHDmCoKAgvPHGGygqKsLOnTtx7NgxBk4ianQYOomoUXrrrbeQnp6On376CUVFRfjDH/7A8ElETcLhw4cRFBSEgIAAPHz4EDt37tQEUCKixoihk4gatd+Gz+LiYk34TE9Pl7o0IiKdHD58GG+99RYCAwM1YbM6gBIRNWYMnUTUJLz11ls4fvw4Dh06hMrKSvj5+SEgIAAHDhyQujQiouc6fPgw3nzzTQQGBqK8vBz79u1j2CSiJoWhk4ialICAAPzrX//CoUOHIJfL0b9/f802IqKG5PDhw+jfvz8CAwPx5MkT7N+/XxNAiYiaEoZOImqSAgICsH//fhw6dAhWVlb4n//5H4ZPImoQ9u3bBz8/PwQGBsLCwgLHjh3TBFAioqaIoZOImrSAgAD8/PPPNcLn/v37pS6NiEzMvn370Lt3bwwYMAB2dnaa76P7+flJXRoRkUExdBKRSfht+FQoFHjzzTc124iIDEUIgV27dqFXr14YOHAgWrVqhX//+9/46aef0Lt3b6nLIyIyCoZOIjIpAQEB2LdvHw4dOgRHR0e89dZbmm1ERPry27A5fPhwuLi44MSJE5ptRESmhKGTiExSQEAAdu3apQmfAwYM0GwjIqovtVqNXbt2oUePHhgxYgRcXV1x8uRJzTYiIlPE0ElEJq06aB4+fBiOjo4YNmwYwycR6UytViMtLQ1dunTBiBEj0Lp1a03YfP3116Uuj4hIUgydREQA3njjDezatQtHjhzRhM/qbUREz1IdNl999VWEhoaiS5cuyMjIwK5du9C9e3epyyMiahAYOomIfsPf3x+7du3C0aNH4eTkhGHDhmm2ERFVqw6bnTp1QmhoKHx9fXHx4kWkpqaiY8eOUpdHRNSgMHQSEdWiT58+mvDZvHlzhk8iAgCoVCp8/fXX6NixI8aPH48//OEP+PXXX5GamooOHTpIXR4RUYPE0ElE9BzV4fPYsWNo3rw5hg8frtkmhJC6PCIykuqw2blzZ3z00Ufw8/NDRkYGvv76a7Rv317q8oiIGjSGTiKiOvDz88OuXbtw+vRpuLu7Y/jw4ejevTvS0tIYPomasIqKCnz99dfo1KkTJk2ahD59+uDXX3/F119/jXbt2kldHhFRo8DQSUSkg9deew2pqak4ffo02rdvj7FjxzJ8EjVBFRUVSEpKgre3NyZNmoQ333wTV69exddffw1vb2+pyyMialQYOomI6qE6fJ45c0YTPrt168bwSdTIPXnyBElJSfDy8sL06dMxZMgQZGVl4csvv4S7u7vU5RERNUoMnUREL6Fr166a8NmhQweGT6JGqrS0FGvXroWXlxfCw8MxdOhQXLt2DV9++SXc3NykLo+IqFFj6CQi0oPq8Hn27FlN+HzttdcYPokauOqw2a5dO8yfPx/BwcGasNm6dWupyyMiahIYOomI9MjX1xepqak4d+4cunXrhtDQUHTt2hVff/011Gq11OUR0X+UlJRg7dq18PHxwfz58xESEoKrV69i7dq1eOWVV6Quj4ioSZEJ/hc8EZHBXLhwAatWrcI333yDTp06Yc6cORg/fjzMzc2lLo3IJJWUlGDjxo349NNPUVpaig8++ADz5s1Dq1atpC6NiKjJYugkIjKCjIwMxMbG4ptvvkHHjh3xySefMHwSGVFxcTESExOxatUqqFQqTJ06FZ988gmcnJykLo2IqMlj6CQiMiKGTyLjKioqwl/+8hfExsaisrISU6dORWRkJBwdHaUujYjIZDB0EhFJ4OLFi1i5ciW++eYbdOjQAZGRkQyfRHpUUFCAL774AmvXroWZmRmmT5+OmTNnwsHBQerSiIhMDkMnEZGEfv31V3z66af45ptv0L59e8ydO5fhk+gl3L9/H+vXr8eaNWtgYWGBP//5zwybREQSY+gkImoArl27htjYWPztb3+Dh4cHIiMj8cEHH8DCwkLq0ogahfz8fCQkJCA+Ph5yuRzTpk3DrFmz0KxZM6lLIyIyeQydREQNyPXr17Fy5UqGT6I6ys/Px+eff44vvvgCSqUSU6dOxezZs2Fvby91aURE9B8MnUREDdBvw6e7uzvmzp3L8En0G3l5eYiLi8O6detgZ2eH2bNnY/r06bCxsZG6NCIi+h2GTiKiBiw7OxuffvopwyfRf9y7dw/x8fFYt24d7O3tMWvWLISHh8Pa2lrq0oiI6BkYOomIGoHfhk83NzdERUUxfJJJycnJweeff46kpCQ4OzsjIiICf/rTnxg2iYgaAYZOIqJG5MaNG4iLi8OXX34JFxcXzJo1C5MnT4aVlZXUpREZRPVrPikpCS1btsTs2bMRFhYGhUIhdWlERFRHDJ1ERI0Qwyc1ddnZ2YiPj+drnIioCWDoJCJqxH57ySHPAlFT8NubaFVfSv7+++/D0tJS6tKIiKieGDqJiJoAhk9q7H67Vi1vmkVE1LQwdBIRNSG5ubn47LPPtG62wvBJDdnFixexcuVKfPPNN2jTpg3XpiUiaoLMpC6AiIj0x93dHWvXrsWVK1cwcuRIzJ07F+3bt8fatWtRXl7+wv2PHDmCkydPGqFSaooqKiqwcePGOrXNyMjAu+++i65du+L//u//8Le//Q1XrlzBn/70JwZOIqImhmc6iYiaMF3XNAwICMD58+fxyy+/oFu3bkaulhozlUqF4OBgfP/99/j111/Rvn37WttduHABq1atwjfffINOnTphzpw5GD9+PMzNzY1cMRERGQvPdBIRNWGtWrXCypUrkZ2djYkTJ2Lx4sXw9PREbGwsysrKtNr+8ssvOHLkCEpKStCvXz9cuHBBoqqpsamsrMQ777yDPXv2wMzMDMuWLavR5vz58xgzZgy6du2KM2fO4G9/+xvOnj2Ld999l4GTiKiJ45lOIiITkpeXh7i4OKxbtw52dnaYPXu25sxnYGAgjh8/jsrKSlhYWMDOzg6HDx9G586dpS6bGrCqqiq8++67SElJQVVVFQDAzMxMc7bz7NmzWL58Ob777jv4+voiOjoawcHBkMlkEldORETGwtBJRGSC7ty5g1WrVuHLL7+Eg4MDxowZg7Vr12q1sbCwgL29PQ4fPoxOnTpJVCk1ZGq1Gu+99x62bNmiCZwAYGlpiSFDhkClUmHPnj3o3bs3FixYgKFDhzJsEhGZIIZOIiITdvfuXaxatQqbN2/Gw4cPUVlZqfW4hYUFmjdvjqNHj8LLy0uiKqkhEkJg8uTJ+Otf/wq1Wl3jcTMzM7z22mtYvnw5Bg8eLEGFRETUUPA7nUREJszFxQXBwcHIz8+vETiBp9/Ve/DgAQICAnD9+nUJKqSGSAiBadOmPTNwAoC5uTlef/11Bk4iIuKZTiIiU9evXz8cOXKk1tBZzdLSEs7Ozjh69CjatGljxOqooRFCYPr06fjLX/7yzMBZzdzcHJmZmWjbtq2RqiMiooaIZzqJiEzYsWPH8Msvvzw3cAJPl8PIz89Hv379cPv2bSNVRw3R3LlzkZiY+MLACTy9xDY2NtYIVRERUUPGM51ERCZs4MCB2LdvHywsLKBSqV7Y3tLSEp6enjh8+DBatmxphAqpIfnkk0/w+eef1ylwymQymJmZwczMDFlZWXB3dzdChURE1BAxdBIRmaiqqiqkpaUhOzsb2dnZuHbtGq5du4bc3FxUVFQAeHqmSi6XQyaTaa3r2alTJxw6dAjNmzeXqnwysvnz52PFihWaf5uZmcHCwgJqtVrrTLmDgwM8PDzQrl07eHp6ok2bNhg8eDB8fHykKJuIiBoAhk4iPUhNTcXYsWOlLoOIiIj0JCUlBWPGjJG6DKImwULqAoiakpSUFKlLoJcUHx8PAJg1a5bElTRsDx8+RH5+PqytrXHz5k2sWbOGr/8mqqysDFevXkWLFi3QokULWFpaSl1So8XfL40H/yOZSL8YOon0iP8j2vilpaUB4FzqIjU1FWvWrOGYEb0Af780HgydRPrFu9cSERERERGRwTB0EhERERERkcEwdBIREREREZHBMHQSERERERGRwTB0EhERERERkcEwdBIRGcCePXvQrFkz7Nq1S+pSiIiIiCTF0ElEZABCCKlLICIiImoQGDqJSFJlZWXw9/eXugy9Gzp0KB49eoSgoCCpS2myY0xERESNA0MnEUlq48aNyMvLk7qMJo1jTERERFJi6CRqgjZt2oSePXtCoVBAqVTC09MTS5cuBfD0ss+4uDh06tQJVlZWcHR0xIgRI3Dp0iXN/omJiVAqlbCxscGOHTswePBg2Nvbw83NDVu2bNGpv0OHDqFz585o1qwZFAoFfH198c9//hMAMHPmTERERCArKwsymQw+Pj5GGB3DO3z4MDw8PCCTybB+/XoAdR/TdevWQaFQoGXLlpg8eTJcXV2hUCjg7++P9PR0Tbvw8HDI5XK4uLhotk2bNg1KpRIymQz3798H8Owx/uGHH2Bvb4/ly5cbY0iIiIjIhDF0EjUxa9aswbvvvovg4GDcvn0bN2/exLx583D58mUAQExMDKKiojB//nzk5eXh4MGDyM3NRWBgIO7duwcAmDp1KmbNmoWysjLY2dkhJSUFWVlZ8PLywqRJk6BSqerc37179zB27FhkZ2fj9u3bsLW1xfjx4zX7BgUFwdvbG0IIXL161cijZRgBAQE4evSo1ra6jml4eDgmTpyIx48fY8aMGcjOzsapU6dQWVmJAQMGIDc3F8DTcDpmzBitPhISErB48WKtbc8a46qqKgCAWq02yBgQERERVWPoJGpCVCoVFi9ejP79+2Pu3LlwcnKCo6MjPvzwQ/Tq1QtlZWWIi4vDqFGjMGHCBDRr1gy+vr7YsGED7t+/j6SkpBrH9Pf3h729PZydnREaGorS0lLk5OTUqT8ACA4OxqJFi+Do6AgnJycMGzYMBQUFyM/PN+rYNCTPG9NqFhYWmrPRnTt3RmJiIoqLi5GcnKyXGoYOHYqioiIsWLBAL8cjIiIiehaGTqIm5Ny5cygsLMSgQYO0tpubm2PGjBnIyMhASUkJevbsqfV4r169IJfLtS7frI1cLgcAzVm5F/VXG0tLSwD/PdNm6n4/ps/Ss2dP2NjYaF0GTURERNQYWEhdABHpT1FREQDAwcGh1scLCwsBALa2tjUec3BwQHFxsV77A4Ddu3dj9erVyMjIQFFR0QvDFT2blZWVSZ8hJiIiosaJZzqJmpBXXnkFADQ3kfm96nBYW7gsLCyEm5ubXvvLycnByJEj4eLigvT0dDx69AixsbE69UFPqVSqes0RERERkdQYOomaEE9PTzg5OeHHH3+s9fEuXbrA1tYWJ0+e1Nqenp6OiooK9OjRQ6/9nT9/HiqVClOnToWXlxcUCgVkMplOfdBTBw4cgBACfn5+mm0WFhY8c0xEREQNHkMnURNiZWWFefPm4eDBgwgPD8etW7egVqtRXFyMixcvQqFQICIiAtu2bcPmzZtRVFSE8+fPY8qUKXB1dUVYWJhe+/Pw8AAA7Nu3D+Xl5cjMzKzxvVEnJyfcvn0b2dnZKC4uZoj6D7VajYcPH6KyshLnzp3DzJkz4eHhgYkTJ2ra+Pj44MGDB9i+fTtUKhXy8/Nx48aNGseqbYz37t3LJVOIiIjIKBg6iZqYiIgIrF+/HgcOHICPjw+USiX69u2LAwcOAAAWLVqEFStWYMmSJWjRogX69u0LT09PHDhwAEqlEsDTNSXj4+MBAF27dsW1a9fw1VdfISIiAgDw9ttvIzMz84X9+fr6IjIyEgkJCXB1dcX8+fPRr18/AE+XFcnNzcWUKVPQsmVLdO7cGUOGDMGDBw+MO2AGsH79es3deyMjIzF8+HCdxhQAysvL4evrC2trawQGBqJ9+/b417/+BSsrK02bqVOnon///njnnXfQoUMHLF26FNbW1gCAPn36aJZXaYpjTERERI2HTAghpC6CqLFLTU3F2LFjwbdT4xcSEgIASEtLk6yGyZMnIy0tDQUFBZLVoAu+/onqpiH8fqG6kclkSElJqbEeMhHVD890EhE1QFxShoiIiJoKhk4iIjK6yZMnQyaTaX4mTJhQo82+ffsQFRWFrVu3wsvLS9P2j3/8Y422AwcOhJ2dHczNzfHqq6/i1KlTxnga9RYbG4uOHTvC2toaSqUSHTt2xIIFCzTLEP3W4cOH8cYbb8DGxgaurq6IjIzEkydPNI/v3LkTsbGxevuPiqY87suWLdN63VX/dOnSRdPmWeO5fft2rX1atGhh7PKb9NzU9T3xMnNIRBISRPTSUlJSBN9OTUNwcLAIDg6WrP+oqCghl8sFAOHp6SnS0tIkq6Wu6vP6DwsLE05OTmLv3r3i8uXLory8XOvxhQsXiqCgIFFUVKTZ5u3tLZo3by4AiO+//77GMffu3SuGDx9evydhZEOHDhWfffaZyMvLE8XFxSI1NVVYWlqKAQMGaLW7cOGCsLa2FgsWLBAlJSXi6NGjokWLFuL999/XardmzRrRt29f8fDhw5eqq6mP+9KlSwWAGj+vvvqqVrvaxlOtVoubN2+KgwcPiiFDhojmzZvr3P/L/H5p6nNT1/fEy8yhLgCIlJSUej8fItLGM51ERA3IihUr8OTJEwghcP36dQQHB0tdksFYW1vj7bffRvv27bVukLRy5Up8++23SE1NhZ2dndY+69atg5mZGcLCwvDo0SNjl6w3crkc06ZNg7OzM2xtbRESEoIRI0bgp59+wp07dzTtli5dChcXFyxevBhKpRJ9+vRBZGQk/v73v+PSpUuadjNmzMBrr72GIUOGoLKysl41mcK4A8CmTZsghND6uXDhglab2sZTJpOhdevWCAwMRLt27YxasynMTV3fE0D955CIpMPQSUREDcbVq1exYMECLF68GAqFosbj/v7+mDlzJm7duoWPP/5Yggr1Y9u2bTWeX+vWrQEAJSUlAIDKykrs3r0bffv21VrfdvDgwRBCYMeOHVr7x8TE4MyZM1izZo3O9ZjKuOviZcZTn0xlburyntBVQ5lDImLoJCKiBmTdunUQQmDYsGHPbLNs2TK0b98ef/3rX7Fv377nHk8Igbi4OHTq1AlWVlZwdHTEiBEjtM4SJiYmQqlUwsbGBjt27MDgwYNhb28PNzc3bNmyRet4VVVVWLhwITw8PGBtbY2uXbsiJSXlT1Y1EwAAIABJREFU5Z70f2RmZsLBwQFt2rQBAFy7dg0lJSWa9W6reXt7AwDOnTuntd3R0RF9+/bFmjVrdL6TsCmP+7O8zHjqkynPze/fE7pqKHNIRAydRETUgOzevRsdOnSAjY3NM9tYW1vj73//O8zMzDBp0iSUlpY+s21MTAyioqIwf/585OXl4eDBg8jNzUVgYCDu3bsH4Ol6p7NmzUJZWRns7OyQkpKCrKwseHl5YdKkSVCpVJrjzZ07F6tWrUJ8fDzu3LmDoKAgjBs3DidPnqzX81WpVLh16xbWr1+Pffv24YsvvoBcLgcA3L17FwBqXE6pUChgbW2tqf+3unfvjlu3buHs2bM61WFK4x4VFQVHR0fI5XK0bdsWI0aMwIkTJ2ptW9/x1CdTmhvg+e+Jao1tDomIoZOIiBqI0tJSXL9+XXMm73n69OmDWbNmITs7G3Pnzq21TVlZGeLi4jBq1ChMmDABzZo1g6+vLzZs2ID79+8jKSmpxj7+/v6wt7eHs7MzQkNDUVpaipycHABAeXk5EhMTMXLkSIwePRoODg6Ijo6GpaUlkpOT6/Wc3d3d4ebmhpiYGKxatQpjx47VPFZ9h1pzc/Ma+1laWqKsrKzG9urvGp4/f77ONZjSuL/33nvYuXMncnNzUVJSgi1btiAnJwd9+/ZFRkZGjfb1GU99MqW5qfa89wTQ+OaQiJ6ykLoAoqYkNTVV6hLoJd28eRMA51IXx44d08tx8vLyIIR47hmd31q2bBm+//57JCQk1PhgCgAZGRkoKSlBz549tbb36tULcrkc6enpzz1+9dmV6rM6ly9fxuPHj7WWZrC2toaLi4vWpYm6yM3NRWFhIU6fPo2oqCgkJSVh//79aNmypeb7bbXdBKWiogLW1tY1tlePXW1nQZ/FlMbd3d0d7u7umn/7+fkhOTkZ3bp1Q0JCAhITE7Xa12c89cmU5qba894TQOObQyJ6iqGTSI9q+yNPjRPn0vjKy8sBQOtOts+jUCiQnJyMgIAAfPDBB4iNjdV6vLCwEABga2tbY18HBwcUFxfrVF/1JYvR0dGIjo7WeszV1VWnY1WztLSEs7MzBg4ciLZt26J9+/ZYsWIF1qxZAxcXFwCosU7h48ePUV5eXmuf1UG0eizrwhTH/bd8fX1hbm6OK1eu1HisPuOpT6Y4N897TzxLQ55DInqKl9cS6dHvb+HOn8b3ExwcjODgYMnraEw/+rppSPWHQ10WdO/Tpw9mz56NzMxMLF26VOsxBwcHAKj1g3RhYSHc3Nx0qs/Z2RkAEB8fX2MM9HG218fHB+bm5ppLBNu2bQs7OzvcuHFDq93Vq1cBAF27dq1xjIqKCgCo9Szos5j6uKvVaqjV6lqDXX3GU59MfW5+/554loY8h0T0FEMnERE1CC1btoRMJtN5rcGlS5eiY8eOOH36tNb2Ll26wNbWtsYNTdLT01FRUYEePXro1I+7uzsUCgXOnDmj036/V1BQgHHjxtXYnpmZiaqqKs2lgxYWFhgyZAgOHjwItVqtabd3717IZLJa72ZaPXatWrWqcz2mMu4AMGjQoBrbTpw4ASEE+vTpU+Ox+oynPpnK3NT1PQE0vjkkoqcYOomIqEGwsbGBl5eX5nu1dVV9SeHvb7ijUCgQERGBbdu2YfPmzSgqKsL58+cxZcoUuLq6IiwsTOd+3n//fWzZsgWJiYkoKipCVVUVbt68qVm8PjQ0FK1atcKpU6eeeRylUokff/wR+/fvR1FREVQqFU6fPo333nsPSqUSs2fP1rRdsGAB7t27h0WLFqG0tBTHjh3D6tWrMXHiRHTo0KHGsavHztfXt871mMq4A8CtW7fw7bfforCwECqVCseOHcNHH30EDw8PTJkypUb734+nsZnK3Ojynmhsc0hE/yGI6KWlpKQIvp2ahuDgYBEcHCx1GY1KfV7/YWFhonXr1jW2h4eHC0tLS/H48WPNtm3btglvb28BQLRo0UL8+c9/rvWYc+bMEcOHD9faplarxerVq0W7du2EpaWlcHR0FCNHjhSXL1/WtElISBA2NjYCgGjXrp3IysoSSUlJwt7eXgAQbdq0EVeuXBFCCPHkyRMRGRkpPDw8hIWFhXB2dhajR48WGRkZQgghRo4cKQCIhQsXPvf5Dxs2TLRt21bY2toKKysr4e3tLUJDQ8X58+drtP3ll19E7969hZWVlXB1dRVz5swR5eXltR536NChonXr1kKtVutUj6mMe0REhPD29hZKpVJYWFgINzc3MWnSJHH79u06jWe1GTNmiObNmz+3r9rU5/eLqcxNXd8T+prDFwEgUlJSdNqHiJ6Nn5KJ9IChs+lg6NSdPkNnZmamsLCwEJs2bdJXeUZVVVUlAgMDxcaNG43e9/3794VCoRCfffaZzvVw3GuqbTyrGTN0cm7q73lz+CIMnUT6xctriYhIEmVlZfjnP/+JzMxMzc0+fHx8sGTJEixZsgQlJSUSV6ibqqoqbN++HcXFxQgNDTV6/zExMejWrRvCw8N1rofjXtPvx1MIgdu3b+Pw4cOamzkZA+em/n4/h0QkHYZOIiKSxIMHD/D222+jffv2+OCDDzTbo6KiEBISgtDQUJ1voCKlAwcOYOvWrdi7d2+d11XUl7i4OJw5cwZ79uyBpaVlverhuP9XbeO5Y8cOtG7dGoGBgdi9e7de+qkrzo3uaptDIpKOTAghpC6CqLFLTU3F2LFjUZ+30+XLl7F+/Xrs378fOTk5KCsrg1KpRKtWrdCuXTtER0fXekc+MoyQkBAAQFpamsSVNB4v8/p/nuobi6xcuVKvx21qduzYgYsXL+KTTz6pceOY+jD1cdf3eP7Wy/5+MfW5qSt9zKFMJkNKSgrGjBmj5+qITJOF1AUQmbKNGzdiypQp6NOnD+Li4vCHP/wB1tbWuHXrFk6cOIF169bh/PnzDJ1kkgYOHIiBAwdKXUaDN3z4cAwfPlxvxzP1cdf3eOqTqc9NXTXkOSQyVby8lkgix48fR1hYGAIDA/Hzzz9j0KBBcHBwgJWVFby8vDB27FgsXLhQ8123hqisrAz+/v4m17ehGeO5NeXxIyIiooaFZzqJJLJs2TJUVVXh008/hYVF7W/FQYMG1boQdkOxceNG5OXlmVzfhmaM59aUx4+IiIgaFp7pJJJARUUFfv75ZzRv3hy9e/eu835CCMTFxaFTp06wsrKCo6MjRowYgUuXLmnaJCYmQqlUwsbGBjt27MDgwYNhb28PNzc3bNmypcYxN23ahJ49e0KhUECpVMLT0xNLly4FABw6dAidO3dGs2bNoFAo4Ovri3/+858AgJkzZyIiIgJZWVmQyWTw8fEB8PRuhQsXLoSHhwesra3RtWtXpKSk6Fybvvs2pLrMS3h4OORyOVxcXDTbpk2bBqVSCZlMhvv37z/zua1btw4KhQItW7bE5MmT4erqCoVCAX9/f6Snp+ulDwD44YcfYG9vj+XLlxt0vIiIiMjESLhcC1GToes6hVeuXBEAhJ+fn079LFy4UMjlcrFp0yZRWFgozp07J15//XXRokULcffuXU27+fPnCwDi559/Fo8ePRJ5eXkiMDBQKJVKUVFRoWkXHx8vAIhPP/1UFBQUiAcPHogvv/xSjB8/XgghRFpamoiJiREPHjwQBQUFws/PT2ttutGjRwtvb2+tGj/++GNhZWUlvvvuO/Hw4UMxb948YWZmJk6cOKFTbYbouy7qs45eXedl/PjxolWrVlr7rl69WgAQ+fn5z31uYWFhQqlUiosXL4ry8nKRkZEhevXqJezs7EROTo5e+vj++++FnZ2dWLJkiU7Pn+vUEtUN1wFuPMB1Oon0imc6iSRQVFQEALC1ta3zPmVlZYiLi8OoUaMwYcIENGvWDL6+vtiwYQPu37+PpKSkGvv4+/vD3t4ezs7OCA0NRWlpKXJycgAAKpUKixcvRv/+/TF37lw4OTnB0dERH374IXr16gUACA4OxqJFi+Do6AgnJycMGzYMBQUFyM/Pr7XG8vJyJCYmYuTIkRg9ejQcHBwQHR0NS0tLJCcn17k2Q/etT/WZl/qysLDQnE3t3LkzEhMTUVxcrLfnN3ToUBQVFWHBggV6OR4RERERwMtriSRRHTYfP35c530yMjJQUlKCnj17am3v1asX5HK51mWWtZHL5QCehk0AOHfuHAoLC2t8Z9Tc3BwzZsyo9RjVa51VVVXV+vjly5fx+PFjdOnSRbPN2toaLi4uWpeavqg2Y/b9sl52Xl5Gz549YWNjY9DnR0RERPSyGDqJJODp6QmFQoErV67UeZ/CwkIAtZ8ddXBwQHFxsU41VJ9tdXBweGab3bt3o1+/fnB2doaVlRU++eST5x6ztLQUABAdHQ2ZTKb5uXHjhk4BW+q+daHvedGVlZXVM8/+EhERETUEDJ1EErCyssKgQYNw//59HDly5JntHjx4gI8++gjAf8NhbSGmsLAQbm5uOtXwyiuvAIDm5jK/l5OTg5EjR8LFxQXp6el49OgRYmNjn3tMZ2dnAEB8fDyEEFo/x44dq3NtUvatK33Piy5UKpXB+yAiIiJ6WQydRBKJiYmBlZUVZs+ejbKyslrbXLhwQbOcSpcuXWBra4uTJ09qtUlPT0dFRQV69OihU/+enp5wcnLCjz/+WOvj58+fh0qlwtSpU+Hl5QWFQgGZTPbcY7q7u0OhUODMmTM61dKQ+taVLvNiYWHx3EuIdXXgwAEIIeDn52ewPoiIiIheFkMnkUS6deuGf/zjH7hw4QICAwOxZ88ePHr0CCqVCtevX8dXX32FDz/8UPNdRoVCgYiICGzbtg2bN29GUVERzp8/jylTpsDV1RVhYWE69W9lZYV58+bh4MGDCA8Px61bt6BWq1FcXIyLFy/Cw8MDALBv3z6Ul5cjMzOzxvcTnZyccPv2bWRnZ6O4uBjm5uZ4//33sWXLFiQmJqKoqAhVVVW4efMm7ty5U+fapOxbV7rMi4+PDx48eIDt27dDpVIhPz8fN27cqHHM3z+36hCpVqvx8OFDVFZW4ty5c5g5cyY8PDwwceJEvfSxd+9eLplCRERE+ifZfXOJmpCXWTIiJydHfPzxx8LX11fY2toKc3Nz4eDgILp37y4+/PBDceTIEU1btVotVq9eLdq1aycsLS2Fo6OjGDlypLh8+bKmTUJCgrCxsREARLt27URWVpZISkoS9vb2AoBo06aNuHLliqb9+vXrha+vr1AoFEKhUIju3buLhIQEIYQQkZGRwsnJSTg4OIiQkBCxfv16AUB4e3uLnJwccerUKdGmTRthbW0tAgICxN27d8WTJ09EZGSk8PDwEBYWFsLZ2VmMHj1aZGRk6FSbvvuuq/osaVCXeRFCiIKCAtG/f3+hUChE27ZtxfTp08WcOXMEAOHj46NZ+qS25xYWFiYsLS1F69athYWFhbC3txcjRowQWVlZeutjz549ws7OTixbtkyn588lU4jqhkumNB7gkilEeiUTQgiJ8i5Rk5GamoqxY8eCb6fGLyQkBACQlpYmcSXaJk+ejLS0NBQUFEhdSg18/RPVTUP9/UI1yWQypKSkYMyYMVKXQtQk8PJaIqJG4lnLxRARERE1ZAydREREREREZDAMnUREDdy8efOQnJyMR48eoW3btvjuu++kLomIiIioziykLoCIiJ5vxYoVWLFihdRlEBEREdULz3QSERERERGRwTB0EhERERERkcEwdBIREREREZHBMHQSERERERGRwfBGQkR6VL3wNzVex48fB8C51MXNmzcBcMyIXoS/X4jIVMmEEELqIogau2PHjiEuLk7qMojIwE6fPg0A6N69u8SVEJGhzZ49G3369JG6DKImgaGTiIiojsaMGQMASE1NlbgSIiKixoPf6SQiIiIiIiKDYegkIiIiIiIig2HoJCIiIiIiIoNh6CQiIiIiIiKDYegkIiIiIiIig2HoJCIiIiIiIoNh6CQiIiIiIiKDYegkIiIiIiIig2HoJCIiIiIiIoNh6CQiIiIiIiKDYegkIiIiIiIig2HoJCIiIiIiIoNh6CQiIiIiIiKDYegkIiIiIiIig2HoJCIiIiIiIoNh6CQiIiIiIiKDYegkIiIiIiIig2HoJCIiIiIiIoNh6CQiIiIiIiKDYegkIiIiIiIig2HoJCIiIiIiIoNh6CQiIiIiIiKDYegkIiIiIiIig2HoJCIiIiIiIoNh6CQiIiIiIiKDYegkIiIiIiIig2HoJCIiIiIiIoNh6CQiIiIiIiKDYegkIiIiIiIig2HoJCIiIiIiIoNh6CQiIiIiIiKDYegkIiIiIiIig7GQugAiIqKG6PHjx3jy5InWtoqKCgDAw4cPtbZbWVnBxsbGaLURERE1JjIhhJC6CCIiooYmMTER06ZNq1PbhIQETJ061cAVERERNU4MnURERLXIz8+Hq6srqqqqntvO3Nwcd+7cgbOzs5EqIyIialz4nU4iIqJaODs7480334S5ufkz25ibm+Ott95i4CQiInoOhk4iIqJnmDBhAp53QZAQAhMmTDBiRURERI0PL68lIiJ6huLiYjg7O9e4oVA1uVyO/Px82NvbG7kyIiKixoNnOomIiJ7Bzs4OQUFBsLS0rPGYhYUFhg8fzsBJRET0AgydREREzzF+/HhUVlbW2F5VVYXx48dLUBEREVHjwstriYiInqOiogItWrRAcXGx1nZbW1vcv38fVlZWElVGRETUOPBMJxER0XPI5XKEhIRALpdrtllaWmLs2LEMnERERHXA0ElERPQC48aNQ0VFhebfKpUK48aNk7AiIiKixoOX1xIREb2AWq2Gi4sL8vPzAQAtWrTA3bt3n7uGJxERET3FM51EREQvYGZmhnHjxkEul8PS0hLjx49n4CQiIqojhk4iIqI6eOedd1BRUcFLa4mIiHRkIXUBRGRajh07htzcXKnLINKZEALNmzcHAFy/fh3Z2dnSFkRUD+7u7ujTp4/UZRCRieF3OonIqEJCQvDdd99JXQYRkUkKDg5GWlqa1GUQkYnhmU4iMjp+6Hm2kJAQAOD46CA1NRVjx46FMf4P9eLFiwCAzp07G7yvpoivb2lVjz8RkbExdBIREdURwyYREZHueCMhIiIiIiIiMhiGTiIiIiIiIjIYhk4iIiIiIiIyGIZOIiIiIiIiMhiGTiIiIiIiIjIYhk4ioiZoz549aNasGXbt2iV1KURERGTiGDqJiJogY6xZSURERFQXDJ1EZHLKysrg7+/f6Pt4nqFDh+LRo0cICgqSrIZqUo8FERERSYuhk4hMzsaNG5GXl9fo+2gsOBZERESmjaGTiBo8IQTi4uLQqVMnWFlZwdHRESNGjMClS5c0bcLDwyGXy+Hi4qLZNm3aNCiVSshkMty/fx8AMHPmTERERCArKwsymQw+Pj5Yt24dFAoFWrZsicmTJ8PV1RUKhQL+/v5IT0/XSx/GdPjwYXh4eEAmk2H9+vUAgMTERCiVStjY2GDHjh0YPHgw7O3t4ebmhi1btmj2NdZY/PDDD7C3t8fy5cuNMSREREQkIYZOImrwYmJiEBUVhfnz5yMvLw8HDx5Ebm4uAgMDce/ePQBPw9KYMWO09ktISMDixYu1tq1ZswZBQUHw9vaGEAJXr15FeHg4Jk6ciMePH2PGjBnIzs7GqVOnUFlZiQEDBiA3N/el+zCmgIAAHD16VGvb1KlTMWvWLJSVlcHOzg4pKSnIysqCl5cXJk2aBJVKBQBGG4uqqioAgFqtNsgYEBERUcPB0ElEDVpZWRni4uIwatQoTJgwAc2aNYOvry82bNiA+/fvIykpSW99WVhYaM6mdu7cGYmJiSguLkZycrLe+mgI/P39YW9vD2dnZ4SGhqK0tBQ5OTlabQw9FkOHDkVRUREWLFigl+MRERFRw8XQSUQNWkZGBkpKStCzZ0+t7b169YJcLte65FPfevbsCRsbG63LeJsauVwOAJoznc9iCmNBREREhsHQSUQNWmFhIQDA1ta2xmMODg4oLi42aP9WVlbIz883aB+NBceCiIiI6oOhk4gaNAcHBwCoNVwWFhbCzc3NYH2rVCqD99FYcCyIiIiovhg6iahB69KlC2xtbXHy5Emt7enp6aioqECPHj002ywsLF54maguDhw4ACEE/Pz8DNZHY8GxICIiovpi6CSiBk2hUCAiIgLbtm3D5s2bUVRUhPPnz2PKlClwdXVFWFiYpq2Pjw8ePHiA7du3Q6VSIT8/Hzdu3KhxTCcnJ9y+fRvZ2dkoLi7WBCe1Wo2HDx+isrIS586dw8yZM+Hh4YGJEyfqrY/GwtBjsXfvXi6ZQkREZCIYOomowVu0aBFWrFiBJUuWoEWLFujbty88PT1x4MABKJVKTbupU6eif//+eOedd9ChQwcsXboU1tbWAIA+ffpolvuYMmUKWrZsic6dO2PIkCF48OABAKC8vBy+vr6wtrZGYGAg2rdvj3/961+wsrLSWx/GsH79evTq1QsAEBkZieHDhyMxMRHx8fEAgK5du+LatWv46quvEBERAQB4++23kZmZqTlGUxkLIiIikp5MCCGkLoKITEdISAgAIC0tTeJKtE2ePBlpaWkoKCiQtI6GMD4NZSzqKjU1FWPHjgX/nDV8DeH1bco4/kQkFZ7pJCL6j6qqKqlLaDA4FkRERKQvDJ1ERGTS9u3bh6ioKGzduhVeXl6QyWSQyWT44x//WKPtwIEDYWdnB3Nzc7z66qs4deqUBBXXXWxsLDp27Ahra2solUp07NgRCxYsQFFRkVa7ZcuWaZ73b3+6dOmiabNz507ExsZK+h8STXmuqqnVasTHx8Pf37/GYw1hDoiI6oOhk4hM3rx585CcnIxHjx6hbdu2+O6776QuSTKmNhaLFi3CunXrMG/ePIwePRrXrl2Dt7c3mjdvjs2bN2P37t1a7X/88UekpaUhKCgIGRkZeP311yWqvG4OHTqESZMmIScnB/fu3cPSpUsRGxuL4OBgnY81bNgwKBQKvPnmm5r1c42pqc8VAGRmZuL//b//h9mzZ+Px48c1Hpd6DoiI6ouhk4hM3ooVK/DkyRMIIXD9+vV6fSBvKkxpLFauXIlvv/0WqampsLOz03ps3bp1MDMzQ1hYGB49eiRRhS9PLpdj2rRpcHZ2hq2tLUJCQjBixAj89NNPuHPnjlbbTZs2QQih9XPhwgWtNjNmzMBrr72GIUOGoLKy0mjPwxTm6uzZs5g7dy6mTJmCbt26PbOdVHNARPQyGDqJiMjkXL16FQsWLMDixYuhUChqPO7v74+ZM2fi1q1b+PjjjyWoUD+2bdtW4/m1bt0aAFBSUlKvY8bExODMmTNYs2bNS9dXF6YyV6+99hq2bt2K8ePHa90lujbGngMiopfF0ElERCZn3bp1EEJg2LBhz2yzbNkytG/fHn/961+xb9++5x5PCIG4uDh06tQJVlZWcHR0xIgRI3Dp0iVNm8TERCiVStjY2GDHjh0YPHgw7O3t4ebmhi1btmgdr6qqCgsXLoSHhwesra3RtWtXpKSkvNyT/o/MzEw4ODigTZs29drf0dERffv2xZo1a4xyx2BTnqtnMfYcEBG9LIZOIiIyObt370aHDh1gY2PzzDbW1tb4+9//DjMzM0yaNAmlpaXPbBsTE4OoqCjMnz8feXl5OHjwIHJzcxEYGIh79+4BeLqu6axZs1BWVgY7OzukpKQgKysLXl5emDRpElQqleZ4c+fOxapVqxAfH487d+4gKCgI48aNw8mTJ+v1fFUqFW7duoX169dj3759+OKLLyCXy7XaREVFwdHREXK5HG3btsWIESNw4sSJWo/XvXt33Lp1C2fPnq1XPbowtbmqK2POARHRy2LoJCIik1JaWorr16/D29v7hW379OmDWbNmITs7G3Pnzq21TVlZGeLi4jBq1ChMmDABzZo1g6+vLzZs2ID79+8jKSmpxj7+/v6wt7eHs7MzQkNDUVpaipycHABAeXk5EhMTMXLkSIwePRoODg6Ijo6GpaUlkpOT6/Wc3d3d4ebmhpiYGKxatQpjx47Vevy9997Dzp07kZubi5KSEmzZsgU5OTno27cvMjIyahyvXbt2AIDz58/Xq566MsW5qitjzQERkT5YSF0AEZme48ePaxYpJ23Hjx8HAI6PDm7evKlT+7y8PAghnnvm7LeWLVuG77//HgkJCTXCGgBkZGSgpKQEPXv21Nreq1cvyOVypKenP/f41Wccq8+eXb58GY8fP9ZarsTa2houLi5al4DqIjc3F4WFhTh9+jSioqKQlJSE/fv3o2XLlgCehlJ3d3dNez8/PyQnJ6Nbt25ISEhAYmKi1vGqx676zKChmOJc1ZWx5oCISB94ppOIiExKeXk5ALzwZi3VFAoFkpOTIZPJ8MEHH6CsrEzr8eqlK2xtbWvs6+DggOLiYp3qq740NDo6WmvNzBs3btS6jEZdWFpawtnZGQMHDsS3336LjIwMrFix4rn7+Pr6wtzcHFeuXKnxmLW1NYD/jqWhmOJc1ZWx5oCISB94ppOIjM7Pzw9paWlSl9EgVZ/h5PjUXWpqaq1ntZ6l+sN6VVVVnffp06cPZs+ejc8++wxLly6Fh4eH5jEHBwcAqDWwFBYWws3Nrc79AICzszMAID4+HjNnztRp37rw8fGBubl5rZfN/pZarYZara418FVUVAD471gaiqnP1fMYaw6IiPSBZzqJiMiktGzZEjKZTOc1HZcuXYqOHTvi9OnTWtu7dOkCW1vbGjeOSU9PR0VFBXr06KFTP+7u7lAoFDhz5oxO+/1eQUEBxo0bV2N7ZmYmqqqqtC6nHTRoUI12J06cgBACffr0qfFY9di1atXqpWp8EVOZq/ow1hwQEekDQycREZkUGxsbeHl56fxd0OpLN83NzWtsj4iIwLZt27B582YUFRXh/PnzmDJlClxdXREWFqZzP++//z62bNmCxMREFBUVoaqqCjdv3sSdO3c581lcAAAKMUlEQVQAAKGhoWjVqhVOnTr1zOMolUr8+OOP2L9/P4qKiqBSqXD69Gm89957UCqVmD17tqbtrVu38O2336KwsBAqlQrHjh3DRx99BA8PD0yZMqXGsavHztfXV6fnpitTmav6MNYcEBHpA0MnERGZnKFDhyIjI0PrO3//+7//Cx8fH2RlZaFXr16YPn16jf38/Py0wlq1RYsWYcWKFViyZAlatGiBvn37wtPTEwcOHIBSqQTwdO3H+Ph4AEDXrl1x7do1fPXVV4iIiAAAvP3228jMzAQArFmzBrNmzUJsbCyaN28OV1dXzJw5Ew8fPgTw9NLKvLw87Nix45nPUaFQ4I033sBHH32E1q1bw87ODiEhIfD09MTx48e1bn7z9ttvIzo6Gm5ubrCxscGYMWPwxhtv4Pjx42jevHmNY584cQKtW7dG165dXzjWL8sU5gp4ehOxgIAAvPLKK0hPT8fZs2fh6uqKN954AwcPHqzR3phzQET0smSCqwoTkRHxO4vPx/HRXfV3OnX5c3b16lV06tQJycnJmDBhggGrMwy1Wo1+/fph4sSJ+OCDD4zad0FBAdzc3LBs2TJNCKur+ry+OVc11XcO+PuFiKTCM51ERGRyfHx8sGTJEixZsgQlJSVSl6OTqqoqbN++HcXFxQgNDTV6/zExMejWrRvCw8ON0h/nqiZjzwER0cti6CSiBm/r1q3w8vLSWpJAJpNBLpejZcuW6NevH1avXq25nI2oLqKiohASEoLQ0FCdb1QjpQMHDmDr1q3Yu3dvndev1Je4uDicOXMGe/bsgaWlpdH65Vz9l1RzQET0Mhg6iajBGz16NK5duwZvb280a9YMQgio1Wrk5eUhNTUVbdu2RWRkJF599dUad6Ukep7ly5cjPDwcn376qdSl1Nmbb76Jf/zjH3BxcTFqvzt27MCTJ09w4MABODo6GrVvgHMFSD8HRET1xdBJRI2STCaDg4MD+vXrh+TkZKSmpuLevXsYOnRoozoT0hCVlZXB39+/0fdRVwMHDsTKlSulLqPBGz58OKKiomrcEdaYTH2uGsIcEBHVB0MnETUJwcHBmDhxIvLy8rBhwwapy2nUNm7ciLy8vEbfBxERETUMDJ1E1GRMnDgRALB3717NtqqqKixcuBAeHh6wtrZG165dkZKSAuDpsghKpRI2NjbYsWMHBg8eDHt7e7i5uWHLli1ax/7ll1/Qu3dv2NjYwN7eHr6+vigqKnphH8YghEBcXBw6deoEKysrODo6YsSIEbh06ZKmTXh4OORyudZlftOmTYNSqYRMJsP9+/cBADNnzkRERASysrIgk8ng4+ODdevWQaFQoGXLlpg8eTJcXV2hUCjg7++P9PR0vfQBAD/88APs7e2xfPlyg44XERERGRdDJxE1Gd26dQMAXLt2TbNt7ty5WLVqFeLj43Hnzh0EBQVh3LhxOHnyJKZOnYpZs2ahrKwMdnZ2SElJQVZWFry8vDBp0iSoVCoAQGlpKYYNG4bg4GA8ePAAmZmZaN++PSoqKl7YhzHExMQgKioK8+fPR15eHg4ePIjc3FwEBgbi3r17AIB169ZhzJgxWvslJCRg8eLFWtvWrFmDoKAgeHt7QwiBq1evIjw8HBMnTsTjx48xY8YMZGdn49SpU6isrMSAAQOQm5v70n0AT8M78HSJCSIiImo6GDqJqMmws7ODTCZDcXExAKC8vByJiYkYOXIkRo8eDQcHB0RHR8PS0hLJycla+/r7+8Pe3h7Ozs4IDQ1FaWkpcnJyAADZ2dkoKirCq6++CoVCgVatWmHr1q1o0aKFTn0YQllZGeLi4jBq1ChMmDABzZo1g6+vLzZs2ID79+8jKSlJb31ZWFhozqZ27twZiYmJKC4u1tvzHDp0KIqKirBgwQK9HI+IiIgaBoZOImoySktLIYSAvb09AODy5ct4/PgxunTpomljbW0NFxcXrUtPf08ulwOA5kynl5cXWrZsiQkTJiAmJgbZ2dmatvXtQ18yMjJQUlKCnj17am3v1asX5HK51uWv+tazZ0/Y2NgY5XkSERFR48XQSURNxpUrVwAAHTt2BPA0hAJAdHS01vqeN27cwOPHj+t8XGtra+zfvx8BAQFYvnw5vLy8EBoairKyMr31UV+FhYUAAFtb2xqPOTg4aM76GoqVlRXy8/MN2gcRERE1bgydRNRk/PDDD8D/b+9+QmHr4ziOfzTK+FPEpdSMKFH+FptBkh2yJNlhY8iCLJSNokZKZmdjYaPETFIWlmYnKyKTUiIaSv5baBjnWaipm6fnudd17pnh/arZTKfz/fSbqenTmXN+kpqbmyVJ2dnZkiSv1yvDMH56bW5u/ta5S0tLtba2plAopJGRES0tLWl6evpTZ3xERkaGJP1ruby9vZXD4TBt9vPzs+kzAABA/KN0AvgSLi4u5PV65XA41NPTI0lyOp2y2+3a2dn5o3OHQiEFg0FJb0V2cnJSVVVVCgaDnzbjo8rKypSWlvbuoUVbW1sKh8Oqrq6OvpeYmBj9y/BnCAQCMgxDLpfLtBkAACD+UToBxBXDMPT4+KjX11cZhqHLy0stLS2prq5ONptNq6ur0Xs67Xa7uru7tbi4qNnZWd3f3ysSiejs7Ezn5+e/PDMUCsntduvg4EDhcFjb29s6OTmRy+X6tBkfZbfbNTw8rJWVFS0sLOj+/l57e3vq6+tTbm6uent7o8cWFhbq+vpaq6uren5+1uXlpU5OTt6dMzMzU6FQSMfHx3p4eIiWyNfXV93c3Ojl5UW7u7saHBxUXl5edKuaP52xvr7OlikAAHxBlE4AMW9tbU2VlZU6Pz/X09OT0tPTZbPZZLPZVFRUpJmZGXV1dWl/f/+nK3vS2/YcQ0NDmpqaUlZWlnJzczU4OKibmxvNzs7K6/VKkioqKnR0dKS5uTkNDw9LkpqamnR4eKjs7GxFIhHV1tYqJSVFra2tcrvdGhgY+N8Zf8PY2Jg8Ho/Gx8f148cPNTQ0KD8/X4FAQKmpqdHj+vv71djYqM7OThUXF2tiYkLJycmSpJqamujWJ319fcrJyVFJSYlaWlp0fX0t6e1pwOXl5UpOTlZ9fb2Kioq0sbGhpKSkT5sBAAC+ngTDMAyrQwD4Ptrb2yVJPp/P4iSxKVbXx+12y+fz6erqyuoo7ywvL6ujo0P8nMW+WP1+fxesPwCrcKUTAPBLIpGI1REAAEAconQCAAAAAExD6QQA/KfR0VHNz8/r7u5OBQUF8vv9VkcCAABxJNHqAACA2ObxeOTxeKyOAQAA4hRXOgEAAAAApqF0AgAAAABMQ+kEAAAAAJiG0gkAAAAAMA2lEwAAAABgGp5eC+Cv8/v9SkhIsDpGTGN9fh9rFj/4rKzT1tZmdQQA31CCYRiG1SEAfB+bm5s6PT21OgYAfEtOp1M1NTVWxwDwzVA6AQAAAACm4Z5OAAAAAIBpKJ0AAAAAANNQOgEAAAAApkmU5LM6BAAAAADga/oHmPj4hHFDwKIAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Next compile the model and set the optimizer to be 'adam', loss to be MeanSquaredError and metric to be RootMeanSquaredError."
      ],
      "metadata": {
        "id": "hrKUdaEvsafW"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# compiling the network\n",
        "functional.compile(optimizer=tf.keras.optimizers.Adam(), loss=tf.keras.losses.MeanSquaredError(), metrics=[tf.keras.metrics.RootMeanSquaredError()])"
      ],
      "metadata": {
        "id": "fmwlJqkyJc0e"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "We then create features sets by passing features 0 to 4 in the first set and features 2 to 7 in the second set."
      ],
      "metadata": {
        "id": "2ZoLUo2Hsc7_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# creating feature sets\n",
        "inputA = x_train[:, :5]\n",
        "inputB = x_train[:, 2:]"
      ],
      "metadata": {
        "id": "K9wSFXY4Qbnn"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "After that, we train the model on the training data with number of epochs set to 20 and validation split to be 12.75%."
      ],
      "metadata": {
        "id": "HXO2rctys0Km"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# training the network\n",
        "hist = functional.fit([inputA, inputB], y_train, epochs=20, batch_size=32, validation_split=0.1275)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1zW-vJTvMoS2",
        "outputId": "07695ebd-3d7b-47bb-a737-12731d9afa75"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "479/479 [==============================] - 3s 4ms/step - loss: 0.8455 - root_mean_squared_error: 0.9195 - val_loss: 0.5767 - val_root_mean_squared_error: 0.7594\n",
            "Epoch 2/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4659 - root_mean_squared_error: 0.6826 - val_loss: 0.5201 - val_root_mean_squared_error: 0.7212\n",
            "Epoch 3/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4512 - root_mean_squared_error: 0.6717 - val_loss: 0.5007 - val_root_mean_squared_error: 0.7076\n",
            "Epoch 4/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4586 - root_mean_squared_error: 0.6772 - val_loss: 0.5165 - val_root_mean_squared_error: 0.7187\n",
            "Epoch 5/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4487 - root_mean_squared_error: 0.6699 - val_loss: 0.5090 - val_root_mean_squared_error: 0.7134\n",
            "Epoch 6/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4369 - root_mean_squared_error: 0.6610 - val_loss: 0.4758 - val_root_mean_squared_error: 0.6898\n",
            "Epoch 7/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4335 - root_mean_squared_error: 0.6584 - val_loss: 0.4611 - val_root_mean_squared_error: 0.6791\n",
            "Epoch 8/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4355 - root_mean_squared_error: 0.6599 - val_loss: 0.4652 - val_root_mean_squared_error: 0.6820\n",
            "Epoch 9/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4375 - root_mean_squared_error: 0.6615 - val_loss: 0.4600 - val_root_mean_squared_error: 0.6783\n",
            "Epoch 10/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4333 - root_mean_squared_error: 0.6583 - val_loss: 0.4887 - val_root_mean_squared_error: 0.6991\n",
            "Epoch 11/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4285 - root_mean_squared_error: 0.6546 - val_loss: 0.5034 - val_root_mean_squared_error: 0.7095\n",
            "Epoch 12/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4236 - root_mean_squared_error: 0.6508 - val_loss: 0.5465 - val_root_mean_squared_error: 0.7392\n",
            "Epoch 13/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4271 - root_mean_squared_error: 0.6535 - val_loss: 0.4775 - val_root_mean_squared_error: 0.6910\n",
            "Epoch 14/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4287 - root_mean_squared_error: 0.6548 - val_loss: 0.5372 - val_root_mean_squared_error: 0.7329\n",
            "Epoch 15/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4201 - root_mean_squared_error: 0.6482 - val_loss: 0.4570 - val_root_mean_squared_error: 0.6760\n",
            "Epoch 16/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4195 - root_mean_squared_error: 0.6477 - val_loss: 0.4896 - val_root_mean_squared_error: 0.6997\n",
            "Epoch 17/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4287 - root_mean_squared_error: 0.6547 - val_loss: 0.4649 - val_root_mean_squared_error: 0.6818\n",
            "Epoch 18/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4097 - root_mean_squared_error: 0.6401 - val_loss: 0.5086 - val_root_mean_squared_error: 0.7132\n",
            "Epoch 19/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4042 - root_mean_squared_error: 0.6358 - val_loss: 0.4175 - val_root_mean_squared_error: 0.6461\n",
            "Epoch 20/20\n",
            "479/479 [==============================] - 2s 4ms/step - loss: 0.4018 - root_mean_squared_error: 0.6338 - val_loss: 0.4590 - val_root_mean_squared_error: 0.6775\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We then compare the training and validation losses (mse). To do this, we use the history attribute and extract the losses (mse) of the models. Next, we plot it against the number of epochs using functions from the MatPlotLib library."
      ],
      "metadata": {
        "id": "XhjUIKW0s6lI"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# extracting the model loss against epochs\n",
        "tl = hist.history['loss']\n",
        "vl = hist.history['val_loss']\n",
        "eps = range(1, 21)\n",
        "\n",
        "# plotting the train and validation loss\n",
        "plt.plot(eps, tl, 'r', label='Training MSE')\n",
        "plt.plot(eps, vl, 'b', label='Validation MSE')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('MSE')\n",
        "plt.legend()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 296
        },
        "id": "TBfxRTQZNX3M",
        "outputId": "c6dc24b9-250b-4a4b-c9c1-57bec0573e1c"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f72ad688d90>"
            ]
          },
          "metadata": {},
          "execution_count": 15
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "As we can see, the training mse decreases sharply and then plateaus, whereas the validation mse fluctuates. There is some difference between the two."
      ],
      "metadata": {
        "id": "MqLrTRI_tECG"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Next, we compare the training and validation rmse. To do this, we use the history attribute and extract the rmse of the models. Next, we plot it against the number of epochs using functions from the MatPlotLib library."
      ],
      "metadata": {
        "id": "QPVSGtX6tVkS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# extracting the model rmse against epochs\n",
        "ta = hist.history['root_mean_squared_error']\n",
        "va = hist.history['val_root_mean_squared_error']\n",
        "eps = range(1, 21)\n",
        "\n",
        "# plotting the train and validation rmse\n",
        "plt.plot(eps, ta, 'r', label='Training RMSE')\n",
        "plt.plot(eps, va, 'b', label='Validation RMSE')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('RMSE')\n",
        "plt.legend()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 296
        },
        "id": "o6vu758KTpmL",
        "outputId": "98ce3165-f724-4df2-ff8d-430059c235c7"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f7330212b10>"
            ]
          },
          "metadata": {},
          "execution_count": 16
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "As we can see, the training rmse decreases sharply and then plateaus, whereas the validation rmse fluctuates. There is some difference between the two."
      ],
      "metadata": {
        "id": "8VEKvzlhtZXb"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "To evaluate the model accuracy, we first obtain predictions on the test set using the predict() function. Then, we calculate the root mean squared error using the sqrt() function from the numpy library and the mean_squared_error() function from the sklearn library. To easily interpret the rmse, we calculate the percentage variation by dividing the rmse by the mean of the true values."
      ],
      "metadata": {
        "id": "oV_b493evvTX"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# obtaining the predictions\n",
        "predictions = functional.predict([x_test[:, :5], x_test[:, 2:]])\n",
        "\n",
        "# calculating the rmse\n",
        "from sklearn.metrics import mean_squared_error\n",
        "rmse = np.sqrt(mean_squared_error(y_test, predictions))\n",
        "print(str.format('Overall RMSE: {:.2f}', rmse))\n",
        "\n",
        "# calculating the variation\n",
        "variation = rmse/np.mean(y_test)*100\n",
        "print(str.format('Variation: {:.2f}%', variation))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BXVH9zqnT-Kb",
        "outputId": "e1a81158-cb56-479f-9fe9-15b7593bc689"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Overall RMSE: 0.75\n",
            "Variation: 35.73%\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "As we can see, the rmse is 0.75, which means that it the error is around 35.73% from the mean value. This means that the model performs moderately."
      ],
      "metadata": {
        "id": "Bpt7T0BKwgZp"
      }
    }
  ]
}