{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Node representation learning with Deep Graph Infomax\n" ] }, { "cell_type": "markdown", "metadata": { "nbsphinx": "hidden", "tags": [ "CloudRunner" ] }, "source": [ "
Run the latest release of this notebook:
" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "This demo demonstrates how to perform unsupervised training of a GCN, GAT, APPNP, or GraphSAGE model using the Deep Graph Infomax algorithm (https://arxiv.org/pdf/1809.10341.pdf) on the CORA dataset. \n", "\n", "As with all StellarGraph workflows: first we load the dataset, next we create our data generators, and then we train our model. We then take the embeddings created through unsupervised training and predict the node classes using logistic regression.\n", "\n", "> See [the GCN + Deep Graph Infomax fine-tuning demo](../node-classification/gcn-deep-graph-infomax-fine-tuning-node-classification.ipynb) for semi-supervised training using Deep Graph Infomax, by fine-tuning the base model for node classification using labelled data." ] }, { "cell_type": "code", "execution_count": 1, "metadata": { "nbsphinx": "hidden", "tags": [ "CloudRunner" ] }, "outputs": [], "source": [ "# install StellarGraph if running on Google Colab\n", "import sys\n", "if 'google.colab' in sys.modules:\n", " %pip install -q stellargraph[demos]==1.1.0" ] }, { "cell_type": "code", "execution_count": 2, "metadata": { "nbsphinx": "hidden", "tags": [ "VersionCheck" ] }, "outputs": [], "source": [ "# verify that we're using the correct version of StellarGraph for this notebook\n", "import stellargraph as sg\n", "\n", "try:\n", " sg.utils.validate_notebook_version(\"1.1.0\")\n", "except AttributeError:\n", " raise ValueError(\n", " f\"This notebook requires StellarGraph version 1.1.0, but a different version {sg.__version__} is installed. Please see .\"\n", " ) from None" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "from stellargraph.mapper import (\n", " CorruptedGenerator,\n", " FullBatchNodeGenerator,\n", " GraphSAGENodeGenerator,\n", " HinSAGENodeGenerator,\n", ")\n", "from stellargraph import StellarGraph\n", "from stellargraph.layer import GCN, DeepGraphInfomax, GraphSAGE, GAT, APPNP, HinSAGE\n", "\n", "from stellargraph import datasets\n", "from stellargraph.utils import plot_history\n", "\n", "import pandas as pd\n", "from matplotlib import pyplot as plt\n", "from sklearn import model_selection\n", "from sklearn.linear_model import LogisticRegression\n", "from sklearn.manifold import TSNE\n", "from IPython.display import display, HTML\n", "\n", "from tensorflow.keras.optimizers import Adam\n", "from tensorflow.keras.callbacks import EarlyStopping\n", "import tensorflow as tf\n", "from tensorflow.keras import Model" ] }, { "cell_type": "markdown", "metadata": { "tags": [ "DataLoadingLinks" ] }, "source": [ "(See [the \"Loading from Pandas\" demo](../basics/loading-pandas.ipynb) for details on how data can be loaded.)" ] }, { "cell_type": "code", "execution_count": 4, "metadata": { "tags": [ "DataLoading" ] }, "outputs": [ { "data": { "text/html": [ "The Cora dataset consists of 2708 scientific publications classified into one of seven classes. The citation network consists of 5429 links. Each publication in the dataset is described by a 0/1-valued word vector indicating the absence/presence of the corresponding word from the dictionary. The dictionary consists of 1433 unique words." ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "dataset = datasets.Cora()\n", "display(HTML(dataset.description))\n", "G, node_subjects = dataset.load()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Data Generators\n", "\n", "Now we create the data generators using `CorruptedGenerator`. `CorruptedGenerator` returns shuffled node features along with the regular node features and we train our model to discriminate between the two. \n", "\n", "Note that:\n", "\n", "- We typically pass all nodes to `corrupted_generator.flow` because this is an unsupervised task\n", "- We don't pass `targets` to `corrupted_generator.flow` because these are binary labels (true nodes, false nodes) that are created by `CorruptedGenerator`" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Using GCN (local pooling) filters...\n" ] } ], "source": [ "fullbatch_generator = FullBatchNodeGenerator(G, sparse=False)\n", "gcn_model = GCN(layer_sizes=[128], activations=[\"relu\"], generator=fullbatch_generator)\n", "\n", "corrupted_generator = CorruptedGenerator(fullbatch_generator)\n", "gen = corrupted_generator.flow(G.nodes())" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Model Creation and Training\n", "\n", "We create and train our `DeepGraphInfomax` model. Note that the loss used here must always be `tf.nn.sigmoid_cross_entropy_with_logits`." ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [], "source": [ "infomax = DeepGraphInfomax(gcn_model, corrupted_generator)\n", "x_in, x_out = infomax.in_out_tensors()\n", "\n", "model = Model(inputs=x_in, outputs=x_out)\n", "model.compile(loss=tf.nn.sigmoid_cross_entropy_with_logits, optimizer=Adam(lr=1e-3))" ] }, { "cell_type": "code", "execution_count": 7, "metadata": { "tags": [ "parameters" ] }, "outputs": [], "source": [ "epochs = 100" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " ['...']\n" ] }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfAAAAEYCAYAAACju6QJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAgAElEQVR4nO3dd3hUVf7H8fd3JpMOCZBCSYAAoXcCYkFZ22IDK6Ji3QXdn67dXXXXXXXXtpYVXazoWhHrKrp27KwgQZASekAIJYQeCOnn90cGjBhKIJmSfF7Pk8e5956Z+c54wyf33nPPMeccIiIiEl48wS5AREREak8BLiIiEoYU4CIiImFIAS4iIhKGFOAiIiJhKCLYBRyqpKQk1759+2CXISIiUi9mzpy5wTmXvOf6sA/w9u3bk52dHewyRERE6oWZ/VjTep1CFxERCUMKcBERkTCkABcREQlDCnAREZEwpAAXEREJQwENcDMbZmaLzGypmd1cw/Z/mtls/89iM9sSyPpERETCRcBuIzMzLzAeOAHIA2aY2WTnXM6uNs6566q1/z3QL1D1iYjI/lVWVrJhwwa2bNlCRUVFsMsJe9HR0aSlpeHz+Wr93EDeBz4IWOqcywUws0nACCBnL+3PA/4aoNpEROQA5OXlYWa0b98en8+HmQW7pLDlnGPjxo3k5eWRkZFR6+cHMsDbAKuqLecBh9XU0MzaARnAZ3vZPhYYC9C2bds6K/Chjxfx4fx1RHg8+LxGhNdD77QEbjulOx6PdlIRkR07dtClSxc8HnWhOlRmRosWLSgoKDio54fq/4FRwBvOuRrPzzjnnnLOZTnnspKTfzG63EFLaRpNh6R4WidGkxgbCcC/p67g1exV+3mmiEjjofCuO4dyBiOQR+CrgfRqy2n+dTUZBVxZ7xXtYfTgdowe3G73snOOUU9N494PFnJC91SS4qMCXZKIiEiNAvln1Awg08wyzCySqpCevGcjM+sKNAO+DWBtNTIz7jqjF0Wl5dz13wXBLkdERGS3gAW4c64cuAr4CFgAvOacm29md5rZ8GpNRwGTnHMuULXtS6eUeK44piP/mbWaqUs3BLscEREJEc899xwREcGbEyygFzKcc+875zo75zo65+7yr/uLc25ytTa3O+d+cY94MF35q060axHLn9+eR3GZbpsQEQlXxx9/PJdcckmdvNa5557L6tV7uxJc/9QT4QBE+7z8/fSeLN+wg398uIg5eVtYsHYbywq2s6WoNNjliYhIHSotPbB/12NiYkhNTa3navZOAX6AhmQmc3rf1jw7dTnD/zWVk8Z9zXEPfsmgu6bwmnqpi4iEvEsuuYQpU6bw/PPPY2aYGc899xxmxssvv8zJJ59MXFwct912G845xowZQ8eOHYmJiaFDhw7ceuutlJSU7H69PU+h71qeOnUq/fv3JzY2lgEDBjBjxox6+TzBO3kfhu4/pw9nD0inuKyCsopKSisqeT07jz+8MYeFawu59eSuRHj1N5GINC53vDufnDXbAv6+3Vs35a+n9Tjg9uPGjSM3N5dWrVoxbtw4ALZtq6r7j3/8I/fddx/jx48Hqu5CSklJYeLEiaSmpjJnzhwuv/xyfD4fd9xxx17fo7KykltuuYVx48aRnJzMddddx8iRI1myZEmdXy9XgNeCz+vhqMykn607pVcr7np/Ac9OXc6S9YU8el6/3feQi4hI6EhISCAyMpKYmBhatmwJQHFxMQCXX345F1xwwc/a33XXXbsft2/fnmXLlvHYY4/tM8Cdczz88MP0798fgNtvv53BgwezbNkyunTpUqefRwF+iCK8Hv56Wg+6tWzKn96ey+njp/L46AF0a9U02KWJiAREbY6CQ9WgQYN+se7pp59mwoQJrFixgh07dlBeXk5lZeU+X8fM6NOnz+7l1q1bA5Cfn1/nAa7zvXVk5MB0XhkzmKLSCk4fP5XXdV1cRCRsxMXF/Wz59ddf58orr+Tcc8/l/fffZ9asWfzlL3+hrKxsn6/j8Xjwer27l3eNtLa/4D8YOgKvQ1ntm/Pfq4dw9SuzuOmNOcxYsYk7R/Qk2ufd/5NFRKTeRUZGHtAsal999RX9+vXj+uuv371uxYoV9VhZ7ekIvI4lN4nipd8exlW/6sRr2XkM/9c3PPnlMuav2UplZUiMTSMi0mhlZGQwc+ZMli1bxoYNG/Z6RN2lSxfmzp3LO++8w7Jlyxg3bhxvvfVWgKvdNwV4PfB6jBt/3YV/XzoQgHs+WMgpj3xD1l2fcvUrs5i3emuQKxQRaZxuuOEGkpKS6NOnD8nJyUydOrXGdpdffjkXXnghl156Kf369WP69OncfvvtgS12PyxERiw9aFlZWS47OzvYZexT/rZipi7dwDdLNzBlwXq27izjtD6tufHEzrRrEbf/FxARCRELFiygW7duwS6jQdnfd2pmM51zWXuu1zXwAEhtGs2Z/dM4s38a24rLeOrLXJ75ZjkfzF3L+Ye15drjO9M8TreeiYjIgdMp9ABrGu3jxl934cubhjJqUDovT1/JcQ9+waTvVuoauYiIHDAFeJCkNI3m76f34v2rh5CZ0oSb35rLWU/8j/lrdH1cRET2TwEeZF1aNuHVywfz4Dl9WLmxiNMe/Ybn/7ci2GWJiEiIU4CHADPjrAFpfHbDUI7tmspfJ8/noU8WE+4dDEWkYdK/TXXnUL5LBXgISYj18cTo/pwzII1HpizhL+/Mp0LXxUUkhPh8Pnbu3BnsMhqMsrKyg57kRL3QQ0yE18M/zu5N87hInvwql81FpTxwTh+N5iYiISElJYXVq1fTpk0bYmJidg8VKrVXWVlJfn4+CQkJB/V8BXgIMjNuObkbzeMiueeDhUzL3cSYIRlcMLgd8VH6XyYiwdO0adVETWvWrNnvuOCyf3FxcSQlJe2/YQ00kEuI+275Jh6ZsoRvlm4gMdbHZUdmcNlRGQpyEZFGYm8DuSjAw8T3Kzcz/rOlTFm4npZNo7l9eHd+3aOlTl+JiDRwewtwdWILE/3bNuOZSwby1v8dQWKsjyte+p7fPp9N3uaiYJcmIiJBoAAPM/3bNuPd3x/FrSd35X/LNnLCQ19p7nERkUZIAR6GfF4PY4/uyKc3HEO/tonc9MYc7v9ooYZiFRFpRAIa4GY2zMwWmdlSM7t5L21GmlmOmc03s4mBrC/ctEmM4fnLBjFqYDrjP1/GNa/Oprhs/xPVi4hI+AtYV2Yz8wLjgROAPGCGmU12zuVUa5MJ3AIc6ZzbbGYpgaovXPm8Hu45sxftWsRx34cLWbtlJ09dlKXZzUREGrhAHoEPApY653Kdc6XAJGDEHm3GAOOdc5sBnHPrA1hf2DIzfje0I+PP78+c1Vs5+/H/sWqTOreJiDRkgQzwNkD13lZ5/nXVdQY6m9lUM5tmZsNqeiEzG2tm2WaWXVBQUE/lhp9Terfi5d8exobtJZz1+P9YsHZbsEsSEZF6Emqd2CKATGAocB7wtJkl7tnIOfeUcy7LOZeVnJwc4BJD28D2zXn9iiPwmDHyyW+Zlrsx2CWJiEg9CGSArwbSqy2n+ddVlwdMds6VOeeWA4upCnSphS4tm/Dm/x1BatNoLnr2O16e/iPlFZXBLktEROpQIAN8BpBpZhlmFgmMAibv0eZtqo6+MbMkqk6p5wawxgajTWIMr19+OP3bJvKn/8zjpHFf80lOvqYBFBFpIAIW4M65cuAq4CNgAfCac26+md1pZsP9zT4CNppZDvA5cJNzTueAD1KzuEheGTOYJ0b3p6LSMeaFbEY++a2ujYuINAAaC72RKKuo5LXsVfzzk8U4B+9cdSRpzWKDXZaIiOyHxkJv5HxeDxcc1o5XLz+c0opKfvt8NjtKyoNdloiIHCQFeCPTMTme8ef3Z3F+Ide9OlvDr4qIhCkFeCN0dOdk/nxKdz7Oyeefny4OdjkiInIQAjaUqoSWS49sz6J1hTz62VI6pcQzou+eY+qIiEgoU4A3UmbG307vyfKNO7jhtR+Ij4rguG6pwS5LREQOkE6hN2KRER4mXJxF99ZN+d1L3/PVYg1LKyISLhTgjVzTaB8vXDaIjinxjH0xW0OvioiECQW4kBgbyUu/GUR6s1gue24GM3/cFOySRERkPxTgAkCL+Che/u1hpDaN5pJnZzB71ZZglyQiIvugAJfdUppGM3HMYTSLi+TCZ6YzN29rsEsSEZG9UIDLz7RKiGHimMNoGu1j9DPTmb9GIS4iEooU4PILac1imTR2MHGRXkZPmK7JT0REQpACXGqU3jyWV8YOJirCy0XPfkf+tuJglyQiItUowGWv2rWI4/nLBrGjpJzLX5xJSXlFsEsSERE/BbjsU5eWTXhoZB9mr9rCbW/PI9ynnxURaSgU4LJfw3q24vfHduK17DxenPZjsMsREREU4HKArju+M8d1TeHOd3M0WpuISAhQgMsB8XiMf47qS9sWsYx9IZsZKzRam4hIMCnA5YDtGjc9KT6K0ROm80lOfrBLEhFptBTgUitpzWJ5/YrD6dqyCVe8NJPXZqwKdkkiIo2SAlxqrUV8FBPHDOaIji34w5tzePLLZcEuSUSk0VGAy0GJi4rgmYsHcmrvVtzzwUI+nLc22CWJiDQqCnA5aJERHh4c2Yc+6Ync+Poccgu2B7skEZFGI6ABbmbDzGyRmS01s5tr2H6JmRWY2Wz/z28DWZ/UXlSEl8cu6I/Pa/zupe8pKi0PdkkiIo1CwALczLzAeOAkoDtwnpl1r6Hpq865vv6fCYGqTw5em8QYHjmvH4vXF3LrW3M1WpuISAAE8gh8ELDUOZfrnCsFJgEjAvj+Uo+GZCZz/fGdeXv2Gl7SaG0iIvUukAHeBqh+z1Gef92ezjKzOWb2hpml1/RCZjbWzLLNLLugoKA+apWDcOWvOnFs1xTufC+HWSs3B7scEZEGLdQ6sb0LtHfO9QY+AZ6vqZFz7innXJZzLis5OTmgBcreeTzGQyP7kNIkmqsmzmLzjtJglyQi0mAFMsBXA9WPqNP863Zzzm10zpX4FycAAwJUm9SRxNhIHh/dn4LCEq57bTaVlboeLiJSHwIZ4DOATDPLMLNIYBQwuXoDM2tVbXE4sCCA9Ukd6Z2WyG2ndeeLRQU8rkFeRETqRUSg3sg5V25mVwEfAV7gWefcfDO7E8h2zk0Grjaz4UA5sAm4JFD1Sd0afVhbZizfxIMfL6JfeiJHdEoKdkkiIg2KhfstP1lZWS47OzvYZUgNdpSUM2L8VLYUlfLRtUfTIj4q2CWJiIQdM5vpnMvac32odWKTBiQuKoLx5/dn285ybntnnu4PFxGpQwpwqVddWjbh2hMyeX/uOt6do/HSRUTqigJc6t3YIR3om57IX96Zx/rC4mCXIyLSICjApd5FeD08cE4fdpZWcOtbOpUuIlIXFOASEJ1S4rnp1134dEE+b32/ev9PEBGRfVKAS8BcemQGWe2acfu781m6vjDY5YiIhDUFuASM12P889y+REV4ufCZ71i9ZWewSxIRCVsKcAmo9OaxvHDZILaXlHPhM9PZuL1k/08SEZFfUIBLwHVv3ZRnLxnI6s07ueTfMygsLgt2SSIiYUcBLkExsH1zHh/dn5y12xj7wkyKSsuDXZKISFhRgEvQHNs1lQfP6cP05Rs57+npbNDpdBGRA6YAl6A6vV8bnhg9gEXrtnHmY/9j+YYdwS5JRCQsKMAl6E7s0ZKJYwazvaScMx+byvcrNwe7JBGRkKcAl5DQv20z3vzdETSN8XHeU9PIWbMt2CWJiIQ0BbiEjIykON644gjioyK45T9zqajUkKsiInujAJeQktwkittO7c4Pq7bw4rcrgl2OiEjIUoBLyBnRtzVDMpO4/6NFrN2q0dpERGqiAJeQY2bcdXovKpzjr+/MD3Y5IiIhSQEuIalti1iuOa4zH+fk89H8dcEuR0Qk5CjAJWT9dkgGXVs24a/vzGdrkYZbFRGpTgEuIcvn9XDvWb3ZuKOE0c9MZ0tRabBLEhEJGQpwCWl90xN58sIBLFpXyAUTFOIiIrsowCXkHds1lScvGsCS9ds5/+npbN6hEBcRCWiAm9kwM1tkZkvN7OZ9tDvLzJyZZQWyPgldv+qSwtMXZbGsYDvnPT1NR+Ii0ugFLMDNzAuMB04CugPnmVn3Gto1Aa4BpgeqNgkPx3ROZsLFVSF+0xtzcE4jtYlI4xXII/BBwFLnXK5zrhSYBIyood3fgPuA4gDWJmFiSGYyfxzWlU9y8nlp2o/BLkdEJGgOKcDNLN7MTjGzzANo3gZYVW05z7+u+uv1B9Kdc//dz/uONbNsM8suKCiodd0S3i47MoOhXZL5238XsGCtJj0RkcapVgFuZhPN7Gr/Yx9Vp7nfBeab2amHUoiZeYCHgBv219Y595RzLss5l5WcnHwobythyOMxHjinDwkxPn7/yix2llYEuyQRkYCr7RH4UGCq//FpQBOgFXA7cNt+nrsaSK+2nOZft0sToCfwhZmtAAYDk9WRTWqSFB/FP0f2ZVnBdu58LyfY5YiIBFxtA7w5kO9/fALwlnMuH5gIdNvPc2cAmWaWYWaRwChg8q6Nzrmtzrkk51x751x7YBow3DmXXcsapZE4KjOJy4/uyCvfreSDuWuDXY6ISEDVNsALgAz/4xOAz/2PY4HKfT3ROVcOXAV8BCwAXnPOzTezO81seC3rEAHghhM706tNArf+Zy7rt6nfo4g0HrUN8NeBl83sU6Ap8Il/fV9gyf6e7Jx73znX2TnX0Tl3l3/dX5xzk2toO1RH37I/Pq+Hf57bl6LSCv7wpm4tE5HGo7YB/gfgYWAecIJzrsi/vjXwdF0WJnKgOqXEc/NJXfliUQETv1sZ7HJERAIiojaN/afBH6ph/QN1VpHIQbj48PZMWbCev7+3gCM7JtE+KS7YJYmI1Kva3kbWx8x6VFs+2cxeN7PbzaxWfwyI1CWPx7j/nN74vMZ1r82mvGKfXTJERMJebU+hPwn0AjCzNOANIB4YA/y9bksTqZ1WCTH87fSezFq5hXFT9tslQ0QkrNU2wLsAs/yPzwRmOOdOAi4Czq3LwkQOxoi+bTh7QBqPfraUKQvy9/8EEZEwVdsAj+SnMcqHAh/4Hy8GWtZRTSKH5O+n96R7q6Zc9+psVm4s2v8TRETCUG0DfBFwtpm1peo+8E/961sBm+uyMJGDFe3z8sToAQBc8dJMiss01KqINDy1DfA7gLuB5cA31e7TPpGfTq2LBF3bFrE8PKovOWu38af/zNP94SLS4NQqwJ1z7wBtgQHAKdU2TQFuqsO6RA7ZsV1Tufq4TN78Po8XNfWoiDQwtb71yz/2eb6ZRZsZzrli59y39VCbyCG75rhM5q/eyh3v5tC2eSxDu6QEuyQRkTpR6/nAzexSM1sKbAe2m9kSM7ukzisTqQNejzHuvH50Tm3CVRNnsXCd5g8XkYahtgO5XAM8RtUsYmf5f94DHjOz39d9eSKHLj4qgmcvySI20stvnstmfaEmPRGR8FfbI/DfA9c45653zr3j/7kOuA64pu7LE6kbrRJieObigWzaUcqYF2ays1Q900UkvNU2wNOp6rC2pyn+bSIhq1daAuNG9WVO3hb+qJnLRCTM1TbA86gawGVPQ/3bRELaiT1acuOJXZj8wxr+PXVFsMsRETlote2F/jjwiJl1Ar72rzuaqlPrf6nLwkTqy/8N7cgPq7Zw9/sL6NkmgUEZzYNdkohIrdX2PvAHqJoT/AKqOq+9B5wP3Oice7DuyxOpe2bGgyP70LZ5LP/38vfkb1OnNhEJP7W+jcw5N9451xZIABKcc22dc4/XfWki9adJtI8nLxxAUWk5//fy95SWa/pREQkv+z2FbmYf72f77sfOuRProCaRgMhMbcL9Z/fhyonfc+t/5vKPs3rj8dj+nygiEgIO5Br46nqvQiRITundisX5mYybsoS4SC+3D+/xsz9KRURC1X4D3Dl3aSAKEQmWa4/PpKi0nKe/Xk60z8vNJ3VViItIyKv1WOgiDY2ZcevJ3dhZVsGTX+USE+nl2uM7B7ssEZF9UoCLUBXidw7vSXFZJQ9/uoTiskquP6EzkRG17ucpIhIQAf3XycyGmdkiM1tqZjfXsP0KM5trZrPN7Bsz6x7I+qRx83iM+87qzaiB6Tzx5TKG/+sb5q3eGuyyRERqFLAANzMvMB44CegOnFdDQE90zvVyzvUF/gE8FKj6RKBq9rJ7z+rNhIuy2LSjlBHjp/LQx4t0m5mIhJxAHoEPApY653Kdc6XAJGBE9QbOuepzPcYBGqxaguL47ql8ct0xjOjbmkc+W8rFz35HSbkmQBGR0BHIAG8DrKq2nOdf9zNmdqWZLaPqCPzqml7IzMaaWbaZZRcUFNRLsSIJsT4eGtmXB87pw7e5G7nlzbmaAEVEQkbI9dDxj/TWEfgj8Oe9tHnKOZflnMtKTk4ObIHS6Jw9II0bTujMW7NW8/CnS4JdjogIENhe6Kv5+ZSjaex7kJhJVE2eIhJ0Vx3biZWbihg3ZQnpzWM5e0BasEsSkUYukEfgM4BMM8sws0hgFDC5egMzy6y2eAqgwx0JCWbG3Wf24shOLbj5zTlMXboh2CWJSCMXsAB3zpUDVwEfAQuA15xz883sTjMb7m92lZnNN7PZwPXAxYGqT2R/fF4Pj10wgA7Jcfzm+Rl8sWh9sEsSkUbMwr1TTlZWlsvOzg52GdKIbNhewkXPfMfi/EIeOrcvw/u0DnZJItKAmdlM51zWnutDrhObSKhLio9i0uWD6d+uGddMmsWL034Mdkki0ggpwEUOQtNoHy9cNohju6Rw29vzeOqrZcEuSUQaGQW4yEGK9nl54sIBnNyrJfd+sJDZq7YEuyQRaUQU4CKHwOf1cO9ZvUltGs1Nr/+g0dpEJGAU4CKHqGm0j3vO7MWS9dt5ZIrufBSRwFCAi9SBoV1SOGdAGk98mcvcPM1gJiL1TwEuUkf+fGp3kuIjufH1HzR7mYjUOwW4SB1JiKk6lb4ov5D7PlxIRWV4j7EgIqFNAS5Sh47tmsp5g9ryzDfLOfXRb5iWuzHYJYlIA6UAF6ljd5/Rk3+d349tO8sY9dQ0fvfSTFZtKgp2WSLSwCjAReqYmXFq79ZMueEYrj+hM18sKuDUR79h+YYdwS5NRBoQBbhIPYn2ebn6uEw+uGYIXo9x2XMz2FJUGuyyRKSBUICL1LP2SXE8eeEAVm/eye9e+l491EWkTijARQJgYPvm3Hd2L77N3chtb88j3GcBFJHgiwh2ASKNxRn90sgt2MGjny2lSXQEvdMT8XkMn9dD+6RYOqU0CXaJIhJGFOAiAXTd8Z1ZsbGICd8s/9l6j8HEMYMZ3KFFkCoTkXBj4X4qLysry2VnZwe7DJED5pwjb/NOSsorKauopKS8kutfnc32knLev2YISfFRwS5RREKImc10zmXtuV7XwEUCzMxIbx5Lp5R4urVqSt/0RP51fn+27CzjuldnU6kR3ETkACjARUJA99ZNuf20Hny9ZAOPfbE02OWISBhQgIuEiPMGpTOib2se+mSxhmAVkf1SgIuECDPjrjN60b5FHFdNnMV3yzcFuyQRCWEKcJEQEh8VwZMXDiAuysu5T33LPR8soKS8IthliUgIUoCLhJjM1Ca8f/UQRg1sy5Nf5jLiX1NZsHZbsMsSkRAT0AA3s2FmtsjMlprZzTVsv97McsxsjplNMbN2gaxPJFTERUVwz5m9eObiLDZsL+W0R7/htrfnsX5bcbBLE5EQEbAANzMvMB44CegOnGdm3fdoNgvIcs71Bt4A/hGo+kRC0XHdUvno2iGMHJjOK9+t5Oj7P+e+Dxeytags2KWJSJAF8gh8ELDUOZfrnCsFJgEjqjdwzn3unNs1cfI0IC2A9YmEpBbxUdx9Ri8+vf4Yft2jJU98uYyj7vuM61+dzUfz17GzVNfIRRqjQA6l2gZYVW05DzhsH+1/A3xQ0wYzGwuMBWjbtm1d1ScS0tonxTFuVD+uOKYjz3yznE9y8nlr1mpifF6O7ZbC7af1ILmJRnETaSxCcix0MxsNZAHH1LTdOfcU8BRUDaUawNJEgq5bq6Y8cE4fyioq+W75Jj6ct47XZ65ixYYdTBo7mCbRvmCXKCIBEMhT6KuB9GrLaf51P2NmxwN/AoY750oCVJtI2PF5PRzZKYm/nd6Tx0cPYNG6Qsa+MJPiMp1SF2kMAhngM4BMM8sws0hgFDC5egMz6wc8SVV4rw9gbSJh7VddUrj/nN58m7uR616dTYXGUxdp8AIW4M65cuAq4CNgAfCac26+md1pZsP9ze4H4oHXzWy2mU3ey8uJyB7O6JfGn0/pxgfz1vHnt+dRWKye6iINmaYTFWlg7vtwIY9/sQyA1gnRZKY2oVurpowZkkELTVUqEnb2Np1oSHZiE5GD94dfd2FwhxbMW72VJfmFLM7fztSluUxduoFXxg4mPkq/9iINgX6TRRoYM+OYzskc0zl597rPFuYz5oWZ/O6lmTxz8UAiIzSKski402+xSCNwbNdU7jmzF18v2cAf3viBSnVyEwl7OgIXaSRGZqVTUFjC/R8tIik+itP6tGbO6q3MWbWFxfmFDGzfnN8MyaBVQkywSxWRA6BObCKNiHOO2yfP5/lvf9y9rnlcJB2S4pi1agsegzP6teHyYzrSMTk+iJWKyC7qxCYimBl/Oa0HfdsmEh3hpVdaAm0SYzAzVm0q4umvc3l1xipen5nHpUdkcMvJXfF5daVNJBTpCFxEfmbD9hLGfbqEF6f9yGEZzRl/QX+SdPuZSNDs7Qhcf1qLyM8kxUfxt9N78s9z+zB71RZOe/Qb5uRtAaCwuIzvlm/ixWk/sji/MMiVijRuOoUuIjU6o18amSlNuPzFmZz9xLe0bBrNyk1Fu7fHRXqZcPFADu/YIohVijReOoUuIvu0aUcpf/9vDsVlFXRv1ZQerRNIbRrNNZNmsXJTEU+MHsCvuqYEu0yRBmtvp9AV4CJyUDbtKOWiZ6ezaF0hj4zqx0m9WgW7JJEGSdfARaRONY+LZOKYwfROS+TKid/z0CeLWVXtFLuI1C8dgYvIISkqLefaSbP5OCcfgL7piQzv05reaQlUVMypK8IAABFDSURBVDoqnMM56JgcT8uE6CBXKxJ+dApdROpV3uYi3v1hLZN/WMOCtdt+sT3G5+VPp3TjgsPaYmZBqFAkPCnARSRglq7fztqtO/GY4THDOcfjXy7j6yUbOLZrCvee1YuUJjoaFzkQCnARCarKSscL367gng8WEhcVwT/O6s3x3VODXZZIyFMnNhEJKo/HuOTIDN77/VG0SohmzIvZTPg6l5oOImasqBosprS8MgiVioQHBbiIBFRmahPe/N0RDOvRkr//dwG3T55PhX9608LiMv789lzOeeJbbnt7HsP/9Q0/rNoS5IpFQpNGYhORgIv2eRl/fn/u+WABT3+9nNVbijknK407Js9n7bZifnNUBgPaNeOOd+dzxmNTufTIDG44sTOxkfonS2QXXQMXkaB6/n8ruOPd+VQ6yEyJ576ze9O/bTMAthWXce8HC5k4fSWpTaMY3qc1p/RuTZ+0hFr1ZN+0o5SvlxTw6x4tifZ56+ujiNQLdWITkZD15eICFq8r5KIj2hEV8cuAnZ67kSe/yuXrJQWUVTjSm8dwXNdU2iTGkNwkiqT4KFomRNMhKQ6P56dg315SzoSvc5nw9XK2l5TTJz2Rxy/oT+vEmEB+PJFDogAXkbC3taiMj3LW8d85a5mWu5GSPTq5NYmKoE96Iv3aJhLt8/LMN8vZtKOUYT1aMqRzEve8v5CoCA/jL+jP4A6ahEXCgwJcRBoU5xyFJeVsKCyhoLCElZuK+CFvC7NWbmHhukIqKh1DMpO48cQu9ElPBKruTx/7YjY/bizi1pO7cdmR7TWojIS8kAhwMxsGjAO8wATn3L17bD8aeBjoDYxyzr2xv9dUgIvInopKyykoLKFdi7hfbCssLuOG137g45x8Bndozp0jetI5tUkQqhQ5MEG/D9zMvMB44CSgO3CemXXfo9lK4BJgYqDqEpGGJzYyosbwBmgS7eOJ0QO464yeLFhbyMnjvubu9xewvaQ8wFWKHJpA3pMxCFjqnMsFMLNJwAggZ1cD59wK/zaN3iAi9cbjMS44rB3DerTkHx8u4qmvcvnPrNVktWtG68QYWiVEk9Yshu6tEkhvHqPT7BKSAhngbYBV1ZbzgMMO5oXMbCwwFqBt27aHXpmINEot4qO47+zejByYzmOfL2VRfiFfLCpgZ1nFT23iIumTnkjvtAQqKx0F20spKCxhc1EpnVPjOb5bKkd2StLtaRJwYTkqgnPuKeApqLoGHuRyRCTMDWjXjGcuGQhUdY7bUlTGqs1FzMnbyuxVW5i9agufLVyPWVWgJ8VH0TTGx+TZa3jlu1XE+LwMyUxizNEdGNi+eZA/jTQWgQzw1UB6teU0/zoRkZBhZjSLi6RZXCS90xIZPbgdADtLK/B5jQjvT12HSsormJ67iU8X5PPhvHWMfPJbLj+6I9edkFnj/ewidSmQY6HPADLNLMPMIoFRwOQAvr+IyEGLifT+LLwBoiK8HN05mTtH9OTzG4cyamA6T3y5jNPH/4/F+YVBqlQai0DfRnYyVbeJeYFnnXN3mdmdQLZzbrKZDQT+AzQDioF1zrke+3pN3UYmIqHkk5x8bn5zDoUl5fRuk8D2knIKi8vZXlJOq4RoDu/YgiM6JnFYh+Y0jfYFu1wJAyFxH3h9UICLSKgpKCzh3g8WsnbrTuKjImgS7SMuyktuwQ5mrNhESXklHoPWiTEkxPh2/8RFRRDt8xAV4SXa58Hn9eAxw2NVp/bTmsVoPPdGaG8BHpad2EREQllykygeHNmnxm0l5RV8/+MWvl22gVWbd7J1Zxlbd5axZP12ikrKKS6vpLisguKyCiprOL5KiPFx9oA0LjisLR2S4+v5k0go0xG4iEiIqqx0OKDSOSoqHbNWbuGl6T/y0bx1lFc6juzUggsHt+P4bqm/uD4vDYdOoYuINBDrC4t5bcYqJk5fyZqtxbRsGs35h7XltD6taR4bSUykl8gIz+7x4gv848VXVDoGtm9OZITCPpwowEVEGpiKSsdnC9fz4rQf+Wpxwc+2+byGmVG6x4xtCTE+Tu7VijP6tSGrXbPd069WVjpKKyrxeT14Pb8cec45R6Wjxm27rN9WTIv4qH22kdpTgIuINGDLN+zgu+Ub2VFSQVFpOTtKK6iodCTFR5LSJJrkJlEUl1Xw7g9r+Gh+PjvLKkiM9WFAUWnFz6Zm9XqMSK8Hn9eoqHSUVVSFe6TXw3UndOaKYzr8YnjZf09dzp3v5dAxOZ7rT+jMsB4tfzY3uxw8BbiIiACwo6Scj3PWMW3ZJiIjPMREeonxeYnyeSivcJSUV1BaXklZhSPCY/giqnrE56zZxqcL8hnRtzX3ndWbaJ+X8opK/vZeDs9/+yNDMpNYu7WYpeu306N1U244sTO/6pJyQGPJl1VUMnn2GtonxTKgnUazq04BLiIih8Q5x2NfLOOBjxfRq00CD43sy93vL+CzhesZMySDm0/qBsDbs1bz8JTFrNq0k9YJ0RzbLYXjuqZyeMcWv7gFzjnHJzn53PvhQnILduAxuOrYTK4+tpM65vkpwEVEpE58mpPPNZNmsaO0Ao/BHSN6cqF/yNldSssrefeHNXycs46vl2ygqLSCaJ+Hri2b0iEpjoykOFolxvDajFV8t2ITHZLjuOGELny2cD1vfp9HVrtmPDyqL2nNYne/5s7SiqrT+zV0wnPOMSdvK/PWbGVLURmbdpSyeUcpCbE+Ljsyg/Tmsb94TrhQgIuISJ1ZnF/IfR8s5MLD2zG0S8o+2+4aM/7zRetZnF/I8oIdrNlaDEBSfCTXHt+ZUQPTdx9xvz1rNX/6z1y8HmNolxTyNhexctNONmwvIcbnZWBGc47q1IIjOyVRVuF4f+5a3p+7lrzNO3e/Z1ykl2ZxkazfVkKFc5zRrw1X/qoTGUk1zxMPVafxi0oriPR6iIyouTNfMCjARUQkZOwsrWDV5iLSmsUQG/nLMcV+3LiDP745h1WbdtK2eSztWsSS3jyWgsISpi7dwJL123e39XmNozolcXKvVhzZKYkW8ZG7J5NZt7WYJ79axsTpKymrqOSwjBZ4PUZpRSVlFZXsLK1gm38wnR2lFT+rwWNVvfaP75bKaX1ac0THFkR4PewsrWDKwnwmz15Dztpt/HFYV07r07revisFuIiINBjrthYzdekGzOC4rqkkxO57XPmCwhImfJ3LtNyNRPh72EdGeImK8PxsONvYSC9lFY6yikpKyytZvWUnn+Tks72knBZxkfRNT2Ra7kZ2lFaQ0iSK5nGRLFxXyB+GdeF3x3Q8oA57taUAFxEROQjFZRV8sWg97/6wltmrtjAkM4nhfVpzWIcWlFVUctMbc3j3hzWMGpjO307via+OO99pLHQREZGDEO3zMqxnK4b1bPWLbV6Pl3Hn9qVt8xjGf76M1Vt28tgF/WkSgJnm1EdfRETkEHg8xk2/7sp9Z/ViWu5Gvlu+KSDvqyNwERGROnDuwLYclZlMm8SYgLyfjsBFRETqSKDCGxTgIiIiYUkBLiIiEoYU4CIiImFIAS4iIhKGFOAiIiJhSAEuIiIShhTgIiIiYUgBLiIiEobCfjITMysAfqzDl0wCNtTh6zVG+g4Pnb7DQ6fv8NDpO6wbh/o9tnPOJe+5MuwDvK6ZWXZNs77IgdN3eOj0HR46fYeHTt9h3aiv71Gn0EVERMKQAlxERCQMKcB/6algF9AA6Ds8dPoOD52+w0On77Bu1Mv3qGvgIiIiYUhH4CIiImFIAS4iIhKGFODVmNkwM1tkZkvN7OZg1xMOzCzdzD43sxwzm29m1/jXNzezT8xsif+/zYJda6gzM6+ZzTKz9/zLGWY23b8/vmpmkcGuMZSZWaKZvWFmC81sgZkdrv2wdszsOv/v8Twze8XMorUf7puZPWtm681sXrV1Ne53VuUR/3c5x8z6H8p7K8D9zMwLjAdOAroD55lZ9+BWFRbKgRucc92BwcCV/u/tZmCKcy4TmOJfln27BlhQbfk+4J/OuU7AZuA3QakqfIwDPnTOdQX6UPVdaj88QGbWBrgayHLO9QS8wCi0H+7Pc8CwPdbtbb87Ccj0/4wFHj+UN1aA/2QQsNQ5l+ucKwUmASOCXFPIc86tdc59739cSNU/mm2o+u6e9zd7Hjg9OBWGBzNLA04BJviXDTgWeMPfRN/hPphZAnA08AyAc67UObcF7Ye1FQHEmFkEEAusRfvhPjnnvgI27bF6b/vdCOAFV2UakGhmrQ72vRXgP2kDrKq2nOdfJwfIzNoD/YDpQKpzbq1/0zogNUhlhYuHgT8Alf7lFsAW51y5f1n7475lAAXAv/2XISaYWRzaDw+Yc2418ACwkqrg3grMRPvhwdjbflenOaMAlzphZvHAm8C1zrlt1be5qnsVdb/iXpjZqcB659zMYNcSxiKA/sDjzrl+wA72OF2u/XDf/NdpR1D1x1BrII5fnhqWWqrP/U4B/pPVQHq15TT/OtkPM/NRFd4vO+fe8q/O33VqyP/f9cGqLwwcCQw3sxVUXbo5lqrruYn+U5mg/XF/8oA859x0//IbVAW69sMDdzyw3DlX4JwrA96iat/Uflh7e9vv6jRnFOA/mQFk+ntcRlLVeWNykGsKef5rtc8AC5xzD1XbNBm42P/4YuCdQNcWLpxztzjn0pxz7ana7z5zzl0AfA6c7W+m73AfnHPrgFVm1sW/6jggB+2HtbESGGxmsf7f613fofbD2tvbfjcZuMjfG30wsLXaqfZa00hs1ZjZyVRdi/QCzzrn7gpySSHPzI4Cvgbm8tP121upug7+GtCWquleRzrn9uzoIXsws6HAjc65U82sA1VH5M2BWcBo51xJMOsLZWbWl6pOgJFALnApVQcp2g8PkJndAZxL1d0ls4DfUnWNVvvhXpjZK8BQqqYMzQf+CrxNDfud/w+jf1F1aaIIuNQ5l33Q760AFxERCT86hS4iIhKGFOAiIiJhSAEuIiIShhTgIiIiYUgBLiIiEoYU4CJS78xsqJk5/5jvIlIHFOAiIiJhSAEuIiIShhTgIo2Amf3ezBaaWbGZLTGzP+0a39rMVpjZXf4ZvLaZ2QYzu9vMPNWe38TMnjSzAjMrMbNsMztxj/dIMbN/m1m+/30Wmdlle5TSzcy+MrMiM8sxs5MC8PFFGqSI/TcRkXBmZrdTNazotcBsoBvwBBAN3OZv9nuqhhEeCAzyb8+nalIVgGf920ZTNWb2FcB7ZtbbObfQzGKAL4GdwAVUDWXaiarhN6t7APgjsIyqIXdfNbN2zrnNdfupRRo+DaUq0oCZWSywATjTOfdhtfUXAY845xL9s6Ctcs4Nqbb9buBC51y6mXUClgCnOOfer9bme2C2c+4yM/sNMB7o5JzLq6GOoVRNinHWrhnrzCyVqrmShznnPqrrzy7S0OkIXKRh6wHEAG+aWfW/1r1AtJkl+5e/3eN5U4FbzKwp0N2/7qs92nwFHO5/PADIqSm89zB71wPnXL6ZVQCpB/RJRORnFOAiDduu69jnAItr2B7omblKa1invjgiB0G/OCIN23ygGOjgnFtaw0+Fv93gPZ53BLDaObfN/xoAR+/R5mhgnv/xTKC77vMWCRwFuEgD5pzbDtwN3G1mV5pZFzPrYWajzOy+ak37mtntZtbZzM4HrgEe9L/GMuB14DEz+7WZdTWzcUBP4H7/81+hat7jyWZ2vJllmNlxZnZuoD6rSGOjU+giDZxz7m9mtha4iqpQ3knV6fTnqjV7FGgHZANlwL/4qQc6wG+pCuuXgKbAXOBU59xC/3sUmdkxwD+ASUA8sAK4t74+l0hjp17oIo2cvxf6BOfc34Ndi4gcOJ1CFxERCUMKcBERkTCkU+giIiJhSEfgIiIiYUgBLiIiEoYU4CIiImFIAS4iIhKGFOAiIiJh6P8BdtvvgACsTBUAAAAASUVORK5CYII=\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "es = EarlyStopping(monitor=\"loss\", min_delta=0, patience=20)\n", "history = model.fit(gen, epochs=epochs, verbose=0, callbacks=[es])\n", "plot_history(history)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Extracting Embeddings and Logistic Regression\n", "\n", "Since we've already trained the weights of our base model - GCN in this example - we can simply use `base_model.in_out_tensors` to obtain the trained node embedding model. Then we use logistic regression on the node embeddings to predict which class the node belongs to.\n", "\n", "Note that the results here differ from the paper due to different train/test/val splits." ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [], "source": [ "x_emb_in, x_emb_out = gcn_model.in_out_tensors()\n", "\n", "# for full batch models, squeeze out the batch dim (which is 1)\n", "x_out = tf.squeeze(x_emb_out, axis=0)\n", "emb_model = Model(inputs=x_emb_in, outputs=x_out)" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Test classification accuracy: 0.7891714520098442\n" ] } ], "source": [ "train_subjects, test_subjects = model_selection.train_test_split(\n", " node_subjects, train_size=0.1, test_size=None, stratify=node_subjects\n", ")\n", "\n", "test_gen = fullbatch_generator.flow(test_subjects.index)\n", "train_gen = fullbatch_generator.flow(train_subjects.index)\n", "\n", "test_embeddings = emb_model.predict(test_gen)\n", "train_embeddings = emb_model.predict(train_gen)\n", "\n", "lr = LogisticRegression(multi_class=\"auto\", solver=\"lbfgs\")\n", "lr.fit(train_embeddings, train_subjects)\n", "\n", "y_pred = lr.predict(test_embeddings)\n", "gcn_acc = (y_pred == test_subjects).mean()\n", "print(f\"Test classification accuracy: {gcn_acc}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "This accuracy is close to that for [training a supervised GCN model end-to-end](../node-classification/gcn-node-classification.ipynb), suggesting that Deep Graph Infomax is an effective method for unsupervised training." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Visualisation with TSNE\n", "\n", "Here we visualize the node embeddings with TSNE. As you can see below, the Deep Graph Infomax model produces well separated embeddings using unsupervised training." ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [], "source": [ "all_embeddings = emb_model.predict(fullbatch_generator.flow(G.nodes()))\n", "\n", "y = node_subjects.astype(\"category\")\n", "trans = TSNE(n_components=2)\n", "emb_transformed = pd.DataFrame(trans.fit_transform(all_embeddings), index=G.nodes())\n", "emb_transformed[\"label\"] = y" ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAcIAAAGRCAYAAAD2ArMcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAgAElEQVR4nOydd5wV1fm4n3fm9u2NhaVXKYoNFWOJRo1do4LGkogttqgx8avGJCYxmqjJT2OKsXeJCooGxQJRFCwoIEhZet1dFra322fO748z9+7dZamCEJmHz2XvnXLOmTPlnfc97/seUUrh4uLi4uKyr2Ls6Qa4uLi4uLjsSVxB6OLi4uKyT+MKQhcXFxeXfRpXELq4uLi47NO4gtDFxcXFZZ/GFYQuLi4uLvs0riDcxxCRR0TkN7u5jukicqXz/WIReW831HGHiDyxq8vdjnrPEZH1ItIqIgd/0/V/E4jI70TkhV1U1nEiUrGV9c+IyN3O92NEZOmuqHdHEJG7RaRWRKq/6bp3J5l967J1XEG4C3AeiqmPLSKRjN8Xi0i+iDwlItUi0iIiy0Tk9oz9lYgsEBEjY9ndIvKM872fs01rp88FO9pWpdQ1Sqk/7JID3776XlRKff/rlNHVw1Qp9Uel1JVfr3U7xV+AnyqlspVSX3ZeKZqfishXIhJ2zvl0Eflhp+1OFpGPnOuhRkQ+FJGznHXjnPN9a6d9KkTkuN15cHsSpdQMpdR+32SdItIH+AUwXCnV/Zuse28i8+X121DPjuIKwl2A81DMVkplA+uAMzOWvQg8CGQDw4A84CxgRadiyoAfsnXyM+tSSr28iw/FZdv0BRZtZf3fgJ+hH65FQE/g18ApqQ1EZAwwAXgO6AWUAncCZ2aUUw/cKiI5u7LxLpvRB6hTSm3a0R1FxLMb2rPby3bZHFcQfjMcBoxXSjUopWyl1BKl1MRO29wP/P7r3gAicoGIzO607GYR+Y/zPdMUVSwib4pIo4jUi8iMlFbqaCSDMsrI3K/A2a9GRBqc77220J5xIjLT+X5rJ402kaH1XiYi5Y6GtEpErnaWZwFvA2UZ+5V1Nt+JyFkissg5lukiMixj3RoRucXR0ppE5GURCWyhvYaI/FpE1orIJhF5TkTyRMQvIq2ACcwXkZVd7DsEuA74oVJqqlIqopSylFIzlVLjnG0EeAD4g1LqCaVUk3NNfKiUuiqjuHLgU+DnXbWzi7r9IvIXEVknIhtFm8CDzrrjHG3yVueYNojID0TkNMc6US8id3QqMuD0U4uIzBWRAzPqKhORV53zv1pEbsxYF3SulQYRWYy+9jPbebBTXouIvAwEMtZ10Py3dd6c49kgIlUicmXmNesc22KnnkoRuaWLPjsRmEr7tfWMs3xb19JtIvIV0CZd3K8iMkJEpjr9ujHVt845+qvT3irnu7/TObpNtIn2admB+2w7+naLZYnIPcAxwD+cfviHs/wh0cMAzSIyR0SOySjvcBGZ7azbKCIPZKwbLSKfOP03Xxwrxpbq2StQSrmfXfgB1gAndlr2BFqLuAwY3MU+ChgMzAGudJbdDTzjfO/nbOPZjvpDQEtmPcAX6IczwDPA3c73PwGPAF7ncwwgGW0alFFG5n5FwHlOXTlo7eb1jG2nZxzHOGBmF+3sDVQBpzq/TwcGAgJ8FwgDhzjrjgMqOu3/O+AF5/sQoA04yTmOW9Eaty/jnHyO1roL0ULmmi303+XOvgPQWvxrwPOdztWgLex7DbBmG+dnqFNG/61sMw6YCRwENACFzvIK4Lgt7PMg8B/n+HKAycCfMvovidY6vcBVQA0w3tl2BBBJtcnp2wQwxtn+FmC1891AX6d3Aj6nn1YBJzv73gvMcNrRG1iYOnfO9muBm52yxjj13J3RzoqMY9rieUNr2NVO20PAC5nnBtgAHON8L0hdS130W+c6t+damuccW7CL8nKcun+BFkQ5wBHOuruAz4BuQAnwCfqFKPMc3Qf4gSDbuM861butvt3uezZj2SXOfh7neKqBgLPuU+BHzvdsYLTzvSdQB5yGvlZOcn6XbKmeveGzxxvwbfvQtSAMAnegHyAJ58Y6NWO9AgY5F89a56LuShA2dvoM20IbXgDudL4PRgvGkPP7mYyb4y7gDbp4sLMVQdjFtgcBDRm/0xc7XQhCpz/mALdtpR9fB25yvh/H1gXhb4BXMtYZQCWO0HDOySUZ6+8HHtlCvf8Frsv4vZ9zzjxd9UunfX8NfNZpWYVzrqJos+pRThmBrRx7us+AV4D7Mso6rovtBf3wHpix7EhgdUb/RQDT+Z3jtOGIjO3nAD/I6NvPMtYZOIIFOAJY16n+XwJPO99XAadkrPsJ7YLwWPTLj2Ss/4StC8IuzxvwFI6gd34PoqMgXAdcDeRu437tXOf2XEuXb6W8C4Evt7BuJXBaxu+TcV6cnHbEt3FddLjPOq3bat/uyD27lfobgAOd7x8BvweKO21zGxkvjs6yd4FLt7eePfFxTaPfAEqbyP6olDoU/Yb1CjBBRAo7bTcF/bC7egtFFSul8jM+5VvYbjz6hgS4CP3mF+5iuz+jhfJ7os2Rt3exzWaISEhEHhVtPmxG3xT5ImJuz/7Ak8BSpdR9GWWeKiKfOeakRvRLQfF2lleGfoEAQCllA+vRb6cpMj0Cw+i32G2W5Xz3oMfxtkUd0CNzgVKqF/o4/GiBVees6sH2cSdwrYhsrf4S9Jv+HMcc1Qi84yxPt00pZTnfI87fjRnrI3Tsk/UZx2Cjr8sytDAvS9Xj1HUH7f1TlrkvHfuyDKhUzhOxi/VdsaXz1rmezO+gtZ/TgLWiHZGO3EY9mW3c1rXUua5MeqMF3jbLdr6XZfyuUUpFUz928D7bat/uzD3rmKXLHbN0I9q/IXVPXoHWnpeIyBcicoazvC8wttP1cTTbf73vEVxB+A2jlGoG/ghkAf272ORX6AdL6GtUMxUoEZGD0AJx/Bba0qKU+oVSagDagefnInKCszrcqQ2ZHnW/QGtKRyilctFvo6Af9FvFEbZD0DdSapkfeBXtkVmqlMoHpmSUpzqX04kq9A2YKk/QD6TKbbVnW2WhnSmSdBQaW+J9oJeIjNrKNkvRD9LztqcxSqklaPPsr7ayWS1akI3IeEnKU9p5a2fpnfoiety4F7pv1qM1zcwXshyl1GnO5hsy90X3Hxnrejrnp6v1O8IGp02btRdAKfWFUupstBnydfTL5/awPdfS1q7H9Whz8TbLRh971VbK3ZH7bFt9u62yOtTtjAfeCpwPFDj3ZFNqe6XUcqXUhej+vQ+YKHo8fz1aI8y8PrKUUvdu4Rj3ClxB+A0gIr8RkcNExCd6sP8mtLlss5gppdR09LjKpTtbn1IqgR4D+DN6bGXqFtp1hogMcm6eJsACbGf1POAiETFF5BT0uF2KHPSDt9HRan+7Pe0SkVOBG4FzlFKRjFU+tMZUAySd7TJDLjYCRSKSt4WiXwFOF5ETRMSLvuljaNPQjvJv4GYR6S8i2eiXlpeVUslt7aiUWgo8CrwkIieJdhwxge9kbKPQDjC/Ee0glCvaQedoEXlsC0X/Hj2+nL+Fem3gceBBEekGICI9ReTk7T7qzTlURM4V7QzyM3R/foYes2sR7dQRdK6P/UUk5RTzCvBLxzmjF3BDRpmfol8qbhQRr4icCxy+k+17BbhMRIaJSAht0gTAuc8uFpE8515opv263p5yv8619CbQQ0R+Jto5JkdEjnDW/Rv4tYiUiEgxWtvfWrzmjtxn2+rbbZW1kY4CPMcprwbwiMidQG5qpYhcIiIlzrXX6Cy2neM5U3R4kCkiAdGOQKmXls717BW4gvCbQQFPo9/cq9ADyKcrpVq3sP2v0QKsM43S0etyax6F44ETgQlbeYgPBqYBregb6WGl1AfOupvQ7vyNwMXot+oUf0WP89WiH47vbKUdmVyANteVZxzDI0qpFrSAfAU9DnER2vEDSGtF/wZWOeaWTHNSSgBdAvzdadOZ6BCW+Ha2K5OngOfRpqPV6LG9G7a6R0euR4dQPIAOgagA/oA+9nVOeyc6vy9HXw8b0WPCb3RVoFJqtdOmrK3UexvazP2ZY/qahtYAdpY3nDY2AD8CzlVKJRzz6hnoMabV6P5+Am02Ay201zrr3nPanTqOOHAuegy03in/tZ1pnFLqbXQ/f4Bz3M6qmPP3R8Aapy+uQV/D21Pu17qWnGv5JGe/amA5cLyz+m5gNvAVsACY6yzbEtt9n21H326rrIeAMaI9Sv+GHtd7B1iGPp9ROpqETwEWifakfgjtjBdRSq0HzkZbtWqcff6PdlnTuZ69gpSHoIuLi8v/LKJDHBYC/u3R3l1cMnE1QhcXl/9JRKe784tIAXqcarIrBF12BlcQuri4/K9yNbAJ7aVpAdfu2ea4/K/imkZdXFxcXPZpXI3QxcXFxWWf5luZ2LW4uFj169dvTzfDxcXFxWUvYs6cObVKqZLOy7+VgrBfv37Mnj172xu6uLi4uOwziEiXmYxc06iLi4uLyz6NKwhdXFxcXPZpXEHo4uLi4rJP4wpCFxcXF5d9GlcQuri4uLjs07iC0MXFxcVln8YVhC4uLi4u+zSuIHRxcXFx2adxBaGLi4uLyz6NKwhdXFxcXPZpXEHo4vJ1UAqq1+jP/9pMLmsWwv8dB2MKYdxAmPL4nm7RXo9SihUr6pkzp4qWltiebo7LLuJbmWvUxeUbYf0Swv/8BRVfLcYwDXoffAD+q/8MPQfv6ZZtmxXz4GejwYoDAm1N8PdrYM0CuO5ve7p1eyU1NW3ccstUystrMAzBMISbbjqCCy7Yf083zeVr4gpCF5edIdzCop+cwUefrsG2tSbo/XgdJy9bTd9nPwN/cM+2b1s8dJUWgqa3fZltwVuPwGV/hGD2nmvbXkY8brFiRT133fUhq1Y10K1bFiJCPG7xl798wqBBhRx6aNmebqbL18A1jbq47ASbJj3Nhx+vIp6waYsp2mKKcMzmnXcWEpn17p5u3rZZtxik0+1vmFoYznt/z7RpD7JgwUauv34K3/ves1x22et88sl6AKZOXcnJJz/PxRe/xuTJS1m7tpHKymaam6P4fAamafDqq+V7uPUuXxdXELq47AQLJkyiJaJoi4GlFJatBWFT2Kb8jTd2rDDbhqVfwNpFO9eY/zwMl/aDH+TApQNgymPb3scX3HxMM/W7pM/OteN/lK++2siVV05m7twNeL0my5fXc+ONb/Pss/P49a8/wDQNsrO9JBI2tbURFi2qYe7caubO3YBhCPX1ka2Wb9s277yzgldfLXfHFfdSXNOoi8tOsKk+QsIGj6kQERDtSBG3oK7Z2v6CZr4KD98Abc2AgpLecPtLMOig7dv/jX/AU7eBaYIvAC218OjNWrM749ot73fy5TDxL1oIG4YWgnYS8ku3v+7/AWbPruJf//qCFSvqOeSQMm699Tv06JHTYZtHHpmNaQpFRSEAfL4gzc1RfvnLabS2xvH7PYiAbStEFImEwrYTtLXFaWiI0r9/Pnfd9SFHHtmL7363Hz6fmS57xoy1XHjhqzQ2RgHweg3uuus4brhh9DfWBy7bxtUIXVx2guyhI1EKhI4fBeSOPGz7Cpk7De69GJo2aUHkDUBtBdx5OsTj21fGhPvA9GgNzzD0X9MDr9y39f2uvB8OOhGwwUpqIZhTBH/+cPvq3UVUsJ7XmMh4XmAuc7DYgZeIbfD66+Wcdda/eemlRcyaVcnf/z6L3r0f5Nhjn2bSpHKUowEvXlxDbq4/Y0/F2rWN1NSEEdFL6uujKKWwMponIkQiSaZOXcWUKcv55S//yw03TCEe1xvFYknGjHmFpqYooZCXrCwfAHfc8T6ff165y47T5evjCkIXl51g8Omn48nKJm4LSUuRsBQJ28BfWEzZoYduu4A1i+D+SxyvTUP/jbSA1w/hJpjxyrbLsG1ortP7ZOLxQ1PNtve/9z14agX85P/BH96GV2qg15Bt77eLmMlHPM1TLGIhK1jOW0zmSR4nSfJrlx2LJfnNbz6goSFCPG4Ri1lYlsKyFHPnVvGHP3zE889/BUDv3nm0tSXS+7a0xKmvjxIMelFKENEfpUBEf7xeE1CYpnamyc310717NrNnV/HuuysAeOWVRbS1JQiFfNpqAPh8Hmxb8c9/fv61j9Fl1+GaRl1cdoIBJ5xA2eGH07h6NdGmJsQwCOTn023ECHoefvi2C3jjb5CItz9ZtW0VYmEQU2uGmdRWakHZfQB4tWaBYUB2PkTD4MsQhskYZBds34F07w/n/Gz7tt2FhGljOtMxMTDRpkSFopoNzOYLRnPk1yp//fpmqqpaSSTstAAD3cXhcJKsLC9PPjmXc88dxtChRbz55jLicQvbtrEsG8tS9O6dRzJp09ISx+MRkkldTk6O7v9YDAzDwLJskkkbv99DMOhl2rRVnHnmftTVbXnssKYm/LWOz2XX4mqELi47gS87m7Mef5y+xx5LwYABFPTrx8CTTuL0f/0Lw7Md75cr5kJhd21LTTmpiEAiASg48Hi9rKkW/noV3HUO/L9x8KuTYc577eWcfSMkE1qoKgXxmP59zs2QTMKGVRBu2cVH//VZyhIUdloIAojzbzE76TSUQX5+gGg0gVIqbQLNxLIUDQ1RTj75ee6//2Pa2uKEwwliMYtkUmHbsHZtEw0NEUpKghQVhfB4BNM0iMUsWlsTeljVtolEklRXtwKQSNjk5OiXklNPHeTUZafr1e2B0077H4g13YdwNUIXl52kYMAAfvDMM4Tr6jBMk0B+/vbvnFsMwRDUrINYxAllUPqz/7Ew9Agt2J68TQe555VA/QaoWgl/PB/O/QX84Ea44HawLJj8d+1wk5XnLEvART20hmmYcNip8ItnwR/YXd2xQ3jxbXGd52s8luJxi/nzq0kkbEpLs1mzpjG9LiUPTVOwLJuKimYKC4OICD6fiW3b2HZHZ9pwOMm6dU307p3Hscf2paYmzJIlNQSDJrGYhW1DdraPqqoWiotDWJbNOecMBWC//Yo5//wRvPzyQkAH4FuWzZAhRVx11SE7fYwuux7p6m3pf51Ro0ap2bNn7+lmuLiglKLy889ZNW0aptfLwJNPpnTkSOTDl2HC/dq0uWEVNG7SY35HngW3/1t7gVavhj9eADkFUP4ZhJvbC/b4YNTJcOuLEHK8IONx8Plg2rPw9+u0CdXj084wsQgcfjr8esKe6YhOJIjz//gzSZJ40UH9NjZJkpzLWEYwYrvLUkrxzjsrePzxuXzyyXqysnwUFgZpaoqyZk0jlqUFHGhrcijkpawsh2TSBoSGhgjhcALbViQS7dqbYehTIgK5uT4WLryOt95awZ13fkBbWxzLshERPB4Dy1L07JnDLbd8h1gsyVtvLccwhB/8YD9M0+C55+bT1pbgnHP24+abjyQU2vKLgMvuQ0TmKKVGdV7uaoQuLrsJpRQz//Qnyl97DTH0KMSiCRMYdfXVHHL55Tqt2bRntYm0qAxOvBRO+0n7gFZbk9bm1pa3C0FxxhKTCVj4McyaDMdfpNf5nIfrhL/op7jH+W16dKabOe9CSyPk7IDmupvw4uM8xjCRCcRp95A9kIN2SAgC/OUvnzB+/AJWr25EKYhEksRiSQYPLqShIUJOjp9oNElLSwzLUgwZUsTYscOZNGkJkUiSurowHo82eaYQAdM0ME3Iy/MTDifJyfGzbl0jzc06FMK2FUrZWJaF3+8hJ8fHXXd9iGEIAwYU4PEYPPLIHEaP7sX06ePSDjMuex+uIHRx2U3ULFrEktdfJ6tbNwxTj4XZySSrHn+Q4ZumEzBsOO0a2P9oKOgOgVDHAlI5Sxuq9d/0g9RxsAk3w4KP2gVhiqZNHVOngRaG8Rg0bNgrBCHAYPbj59zCAr4iSpT9GEYJJTtURmVlMxMmLMbn04+yQMCDUor6+gjl5bV4vSZHHdWb4uIQgYAHj8egoCDISScNYMqUFfj9HiorWxyh1m4dS3mJBoPay9PjMYjHLd54YwmxmJV2wEntEo8nKC+vRSmFz6eD8g86qDtlZTnMnl3F/PkbOeig7rus71x2La4gdHHZTVR+/jm2ZaWFIMBAazUD/MtRH60EvwGfvgH7HQ6/mwwLZ0I8CgMPgrxiCGTBGddpJxnQQ4jgeJo6drvGTXoWib4j2gVlr/1g+RwtDFOyMxnXnqXdB3xTh79d+Akwiu3wst0Cy5bVddC0lFI0N8dIJGyam2OIwIwZ6zj//OE89tgcolELEfjtbz/g+98fSGVlM2VlWVRVteLxGIjYJJMKwxBCIS8ejxCNWowYUcL5509whF2qro5tiUaTKAWGIbS2xqirC1NSkoVtK9asaUwLwk8/Xc9rr5UTDHq47LKD6d9/Oz18XXYbriB0cdkdJBN4SZJpDAuoKIMTK2lLQJYvCEEf2EqP/91wKGTlA46q8d0fQtVyWPq5zhgTC+t1hqmFoOXEvc1+G+ZNg6FHwp2TYMUcqFmvt4+FwfDo/ZUNY27tGGbxLaCwMIhSirw8PyJCNJrAsrQg83gEwzDIzw/w17/OIhDwkJ2tzcXJpMW7765k7NhhTJq0lHA4gWnCyJHdiEYt1q9vJpm0MAyTvn3zSCQs2tpIjzV2RUowhsNJvF6DpqYoJSVZGIbQvXs2SinGjXudiRPL0W81woMPzuK++07guut2/mXA5evjCkIXl12JUjBjIvOefI4HPith5urDKFiR5IQhEU4rWI9t2YhhEgw4pksRra01VLcHs4eb4Zlf6UB523LMnBmBcLYjBD0+CObop/PCGdqxpvxTSEQhmAexVp0xJhmDa/4GZ1zzjXfH7mbkyFIGDChg1aoGevTIYtmyesfEKSSTiuJiP4mERTJpp1OfpZxcLMvm+ecXkJXlpaAgSDxuMW/eJrxeg169cojHbcaOHU7v3nm8+OJXxONWB3NoZzLX6fhFRXV1C/37FzBqVBmTJy9j4kStCZqmHjOOx5P88pf/ZezYEZSUZH0DPebSFW4coYvLThCpr+fTBx/k32eeyasXXcSSVyeiohH4chqL/vVPrp46hJXhQvqWGLRFbF5eUMSXlVob69YtB8NwBFtKs0uN6dk2VCyFeASibVrYpZxdPF4nzMLZVwQSMa35KVs7w0TD2qTq90NukQ67ML1Q0uub7aBdSCJh8dhjszn66Kc48sgneOCBT4nHdfYZEeGvfz0F0xTWrtUORToEQiECzc1R1q1rdJYrGhujNDXFaGqKEY1aKKUIBLwopWhri+v3kqRNYWGIoUOLmTlzPevWNWEYBkqB3+/BMIQt+b2k8yMA8bjNCScM4OGHT8fjMXjhha/Q2WjaH7s+nwfLUrz88tePnXTZeVyN0MVlB4m3tvLG5ZfTXFFBIBQktmQuH057nfohRXxnWDZPLR2NGAaFwSTgZ1DAQ1PY4s3gWG7qthLDSpC+9WzHU7Gkt36Cr1mox/1S69qatGlTDB0GMeRwWDlXb5sKpDdNR1O0tEDMTKWSylqzrhyOOOOb6aCqlTD9JTANOO4iKO2700XZtmLs2AlMnboyPRY4f/5GXn21nA8+uBSfz2T+/GricYsjj+zF6tWNrF7dgGEItq0FTXOznvGhsTGKYQimKWnHGNvWdViWzkBjmlqTjEaT+P0eIpFW5s+vprKymW7dsvD7TUxTiEaTjoONbmdKGzRNHU7h8Rg89dTZHHdcvw7HsiW2Fsb2xReV3HffxyxdWsewYcXccccxruPNLsbVCF1cdpAV775Lc0UFOaXd8G5Yij8ZJjs3m0WrWmirXMfiaiHHl+GKb5rkB2xaJJeWq/6lBUSkVX9sS48N5hbp/KBtjXQ0gyYh2qo/ytZjhP6QI/A6DViJqZfZGZmhUw/YAQfpvKS7O8vMC7+H6w6Cl/8E4++Ga/aHif9vp4ubMWMt06atIhjUSauzsnxkZ/uYN28DL7+8AIBJk5YQDHoJBr2Ew3GysryI6HyjKSEIuissSxGP21gW+Hz68ZdIWB0cYAxDyM31U13dwrJldcyYsY7q6lbmzdMCVymda7SkJERJSQifzyAry0tRUZD8/ACmKQwaVMiAAfmEw+05TC+55ABAZ6NJkUgkMQzh/PO7DhmZPHkpp576Im+9tYzFizcxYcJiDj74Ub7znSepqWnb6X516YgrCF1cdoDlMXimQTHr2DOp8Xj1+J7Xp81lhlBPHr28jSyr9TG7KsSCTUEa2iBqZOkH+QnnwVMrYdw9MOb/4E/T4LYXdR7RTWu1EEtPmNuFlrD8C+0VmjbNKUcDNLTmZZhawFqW1hYjrVrbHH8XXH8I3DAKHv05tDZuXvbXZdV8mPhnZ+wyS89yb3rhxd9pLXEnePvtFY6m1f6oSiXB/s9/lgFa4KXWpwSZnjJJC51UuENHFEOHluDxGE4wvR7T02bRIKGQl/LyGhIJi1DIS2FhkKwsL7GYzlOane3FMKC0NIuePXPwek2i0SSRSJLi4hDV1a0MG/Yw3bv/hbFjX6G1Nc7ZZw/l7LOHEg7rmMbW1hiJhOKee46ntDR7s2NPJm3uuON9IpF4OotNis8+q+CYY552kgK4fF1c06jLtxorHkcMY/vyf26DJxrg0QZoPeC7hLsP57/HncvJL/yFY+e9gzg5LSOFA1jQXExt2IPXhGgC6sNZlJTm8IefHIrHY0Bu4eaJrvuN0PMSrvxST8dUuazrRiTjOtVaz6FQuUQ/+QMh6NYfuvVuH0usXqPHFkccA4tnQO16LSTbgPee1s45t73IFge7doZpz2lbY2Zfe7x6HHPac/Dj3+9wkcXFQbp8IYD0/IGnnDKI++//mJwcHyUlWZSX1yCi84Imk11P62TbsHjxJnr1ymXgwELKy2sIBDyICIWFQdasaSQatfB6TXw+ExFJj+vW1ITx+QwMQ/D5PPTsmUP37jncfvvRrF3byB13/BfDELKyvFiW4q23ljN27Cu8/fYljB9/Hh99tJZJk8oJBr2MG3cQQ4YUddnGmpo2qqtbO6R9y0wevnZtExMmLOLCCw/Y4X516YirEbp8K2lYtYo3r72WJ0aP5onRo/nonnuIt+24KUkpqE3Cgig81gDFJuQbXsyqauyaBl4/92d82NaLmqYEPYFse5oAACAASURBVHrkM2lNb2K2SV9fDaYdx8LEmxUikJvNeecN23IlgSw4/3bI66Y1KU/nMIcMJ5naCti4WsuHvFIYdBgYome493jhvg/g3mlw6d06l2k8psv3BfTHtmHBh9opZ1eSTHS9XKl2p6Ad5NJLDyIY9BKJJJyiVFoDvOYaPd3V2WcPZdSoMqqrW0km7bSw8Hhkq+EOqbHBK644iEGDCkkkbCKRBPX1EZqaYni9emZ6paC1NU5zczxdXjxuE41arFvXzIIFm1izppGVK+sdhxghEPCm06+FQh4+/ng9a9Y0AHDssX158MFT+OMfT9iiEATIzfWnxyyh4ztLyiln/vyNO9WvLh1xNUKXbx2R+nreuOIKasrLidTVoWyb6vnzWfrmm1z16afbXc68KNxdA2viUNEGLUkoaUxSX9lEIFhGVrwWMWHT0OEkZ5tcOMLiqXdtvEYSskP0ToaxrBZKh+9Pq+WnurqVvn3bs7ps+PJLZt76M6LlX1CaLRy4f3e6jTwQ2bBSCzYARGt2Iu2CRgE5hfpTvQoWfaTNkR6v1ghvPd6ZdR5Yt6g9JZvH8Uz1+XWC7qZa6L31PlBKseD9uXzy5uf4Ax5OuOh4+hwwqOuNjz0f3n1Sm3cNJ4mAldRtOfaC7e73TEpKsnj22R9w+eX/oa1Np2Lz+z3cfffxHHRQD0Bnk/nnP0/n00/X8/bbK3j55YUUFYUIhxMsX17fYQwwM8QhGPTQ0hLn5z9/15l4F/x+beIcPLiAL79so61NoZRs1QTZ2pogkbB46KFZtLTEMM2OWrZpmlhWkvvum8ncudXYNpx33lCuvPJQiotDWygVsrJ8nHnmEJ5+el6H5Urp8U0RYdiw4h3tUpcucJNuu3zr+OqFF3j3F78gUlurH8IOyrI49s47Oe63v8WKxzF97ROmdqYiAT+sABRUroENCuJFCm99EyypxGvaWlaV5jNi0odY783jtr4f8nTld5gh55Pw9cajIuRHPqfQXoC/zxDee+8S8vL07A+Vs2cz/vsnkpVswWsKcUthCJx+dHf6nHiqHlub+oyOCTRMZ2IKx8zn8WnNDtGCBqVNqxjQ2gCb1unfBT1g/vs64TY4M9kHdNlWAp5YCkU9ttiPSinuv/xBJk5Z3z5zgwF3/nI0p9+0BcH2wBXw4UsdbXknXwnXPbTtE7cVEgmLd95ZQSJhceKJA8jNbZ9Fw7Js/v3vhY650cPatc3YtqKgIMDnn1dSVxfeTDMMBrUOEIlsPgmwDsSX9FyG24OIDu7XzjGKnJz29kWjCVpa4h3MmoYhDB9ewtSpP+pyfDBz37POeompU1ell5mmni2je/dsFi26jmDQu8X9XTriJt122WeoW75ca4LaTbBDBPSMe+5h0csvY3q9ZJeWcshVVzF87NjNBOIbLRBT4G+DhibIDkJzIo6V7cUI+rBjcSTowY7E8KxqItlzP96sq2Fh/tXEwll47Fb8pk045zsE2gzOPa1nWggCTLvtNjyJMFlBAwuDoBeicYuPv6ylV9+5GGddpz1GZ75GMtzK+iYh32sTiSsMv5CdLYT8gpGIauG2fpm2nEbDWjhWroBAdsb0Tmjh55gYGXakTva9Fea98xkT31pPSa7CdDTUaNzmnvs+5eixx5NX1m3znX7+pM59Ov3fuu4TfgQHHLPD57AzXq/JmWfut9nyRMLihBOeZfbsDR20vR499OwSqbCI1EwSKVLp0KCjlgg4pkiFYbSHU2QKsS2RmvvQsnRMot/vIZm0aGnRmqxh6Ew3oIX30qW1/PWvn/GnP524xTIDAS/vvfcj3nlnGTfd9C7r1zdjmgaHHVbGU0+d7QrBXYQrCF2+deT27o2yrC4dQVQySXNFBabPhzcUYua992JbFgdceGGH7dYlwCfQ4Ez6YFqKwnUrqevbH7s4h2Q4BtiUPfI6LUsr8PUdSku3E5Amgz6hWrLtVnLMGJYyafEN5Xr/MxAZpcf/gE0LF5LjM1C0O3P4vQaWlYC1C7XjjGGilKLCLqC8vo2DC2O0xYFEjEgkQSDgpTSkMGIRCOU68YQ2oCDcpHOXdg6xSLH0C7h2pB5HHH1Wl3310aRPgXYhCBDwGTRHFF++NYPjrjqv67IPPkF/dgGWZfPWW8t49dUlxGJJTjttMGPGDCcU8rJ8eR2XXjqJzz6rdBxXTLKyfNi2zaZNrdxxx9H8+c+fMHhwETU1bWzY0JqRDxRAsCy1ReGWCqAHC8uy8XiMDtM0ZaITepv4/Yr8/ADFxSFWr24gPz+Ix2PQ1BRLC0HAmZtQ8fbbK7YqCFOccsoQli4dkjbzprLkuOwaXEHo8q0jq7R089f8FCJY0SiJtjY2fPklgfx8Pv/b3xhx/vkdkmMfEoD/toE3lQlNKUJN9dgVXnKf/YCaWiF/xTpUOEKTDbddfjDvf9EDz2fVdPM1keWJkVQGHpUg4feyaWkVRZMfhvNvBSCQn0+kuplgxvPMshWnD1OInUwLtmQkQk97Ew0hLxgmhqHd/C1bEY8nSfgUfq+px//aGjses+rsMSmktUNlwca18NydWnMbfeZmXeXzGdCV5VjA6/1m/Oyuu+4tJk1aglLaeWTRok18+OEa7r//JK6++k3Ky2udKZO0KbOlJUZeXoBo1GLJkloKCgIUF2dRUpLF8OHd+OqramprIxQVBamri6QD47dEImGltUrLsp1JfTtubxhCMOghHrfw+Uy6dQtxzz0nMHhwEbW1YS68cCJNTTEn203HBOGtrXEmT17K8cf3T+dB3RperysAdweuIHT51mHHYoS6dSO8adPmwlAp7KQeF7ITCaINDWxatIhoYyOhonYPvtOyYXwTrM4DexNEPQZ2SU/2nzKJobM+oMbKJpGbizJb+P73+nD11YcSDghLKrII1czCVo6/SygHxEdZiWLF2+/xxNxDmDevGn/RpfRc+yQHe9bhM22SNuSYSXIDIL6Q1u4AK5FAgLKsBJYyMb0mVlKhbBvTANv0gkrqYHnlzHCfFngZgo/Uz1TGGUfIGQZMeURnnemkFZ54wTE8PWEN8aTC59HrWqIQ8ioOPeu4XXKutsbDD07n2afnYFgJDAOirT4CWUFmz67i4Ye/oKUlTiDgpbk5jp4BXgfMJxI6J2hZWQ62nRJgWnDrmSX0eJ5hCBUVzVt8Z0oF4KfMo9q0uvmGeqom7WgzfHgJDQ0RHnzwM2prw4gI4bBzvdm6bhHSmuWaNY2cc87LhEJe/vGPU/nxjw/aTb3psjX2qvAJEckXkYkiskREykXkSBEpFJGpIrLc+evOWbKvUrNex6O986ROGbYFm1bRkCHk9+5NyYgReALOuFyGWSrte24YKNvGTibZtGBBhzJyTHi6J1xZDAf2hUAzlE2zOWDCRLKtjZwwYAOn58/lnEFVXPS7n9CwbgNH91pGXlE2zf6eePIKUbml1KlSLug9k01tinHvjuCD91dj24o6KWVa3kXMig2lPiLEEopB/Qsw/H4djJ5uqg6b8BowY0MWXgOCfj0+WBRS+HCcPVImUUQ716Qy0GSqdMr5T3AEoQH+LKjb0DEbjcPgYw7l/64eSlMYNjYpNjYpsG0efOBkAoW79zZsa2zmvt9O0bN4eMAUkGSUcEsbDQ0R5s/XczT26ZMLdMzWotOjmVx22cFccMEIqqtbaW2NE49b6WD7wsIgQ4YUkZ8f6HAZOZdFmtSYX+Y7gmFoh5oUPXpkM3JkKYcc0oOWlji1tWFqa8MUFAQoKQnRr18+gYCJSEpQZ5pXddltbXGuuGIyH3ywepf2o8v2sbdphA8B7yilxoiIDwgBdwD/VUrdKyK3A7cDt+3JRrrsAT6bDOP/0P7AfusROPHHcNZPN9Nkuh1wAH2OOYY106fT8/DDqV+1inBtLVYspoWnk4szvZcIsZbNU48VmnBjEVyXZzOjtoI5/SOU3nU/vWo+oGn1SkoPOIBex5/IPdc8wUfz9du/z++nrO/hVESGUBCK8/Mhb3Bhn4+4461cEv48ujkegn6/B+/wQaz33MJvfj+Ewn59ySkugot76gB0r44j9Pj92IkIa9uCrKkXqtvy6OaLUpAFJ2a3YRherT1acccbxEnLlkrI7fVp5xnbol1LFCfcwq+f6qX9tEdpF4y96xq+96NVzJ3yCf6Al8POPYFgye532Z/16nskExam6SclvEVMxErS0hRlxIhS1qxporg4i75981i3rolkUpuNQyEvjz56Bvn5AX72s9H06aNnj5g1q5KamjZsGz78cC0FBQH69cvH7zepq4tgmoLf78GybFpbUzGDQkGB30nYrdO1iZO/1evVWmgsZhGJJIhEEgwaVMj69U189dVGEgltSi0ry2G//YoZPboXyaTNe++tZN26pnTQP+jTkEhY3HbbND7//Krd3r8uHdlrBKGI5AHHAuMAlFJxIC4iZwPHOZs9C0zHFYT7Fi318O97IDufuDJZs6aOluYweU/+nb7DjsY75OAOm4sIJ957L+WTJrH0P/9Jjxku+89/SMbjWgAaBoZpYni9mB4P3UZsnuuxqSnK+PELeOCBT4lGk+Tm+gmFfOy33wgefvhWcnP9XH/sbcxaEqM0T5tCW6NRqpZ+yItnPcngwCYtkBoU8xsHkNe3Y9BedraPjRtbyRt+MDn5juY67h54/BZItoIIhlIkswuZX9cDMTfQ1hKhMhjgpP7NmOLMgmAnSWuCac3O8Ra1kjqMos9wWPKpnqkilQKtWx/9++wbt9r9RYMHcNJNOz+h7xdfVPLWW8spLAxy6aUHdvCe7UxbW5yNG9toXrUSn2HhMWyStoHH0GpbUgk5PuGnPz2M1asbWLhwE3375lNcHKKysoUDDyxl/PhzaG5O0NwcJTc3wJgxw5kzp4p3311Jbm4Aw9DzFjY2RvF4DP7xj9O4554ZKKVYsaKeSCRJIOAhHrfTZs9YrKPG7PUahEJe4nGLbt1CTJlyMUuW1DBu3Bts3KgTN2iTq7B+fTNtbXHGjBnOL395DKed9iLr1jVtFm9oGLB6deNmY4kuu5+9RhAC/YEa4GkRORCYA9wElCqlNjjbVAOlXe0sIj8BfgLQp0+f3d9al2+O5XOIRGJ88OlyFiyoIJnUHnw5QQ/5l1zKmW+8S06PjvFwps/H/hdcwP4XtMe7vX3TTSx86SWsaBQxtUemFYsx4LTTKBjQ8UFfVxfmssveYPbsKpqbY3g8Bq2tCYYPL2HJkloee2wOF5xYwBdL43TPa5+WJzsgtMXg1ZbDuf3KUTp7S9lAercm+WpxA/GmprRQzcnxEwh4yMrKcIE//Wrotz9M+ivUb0BGHo//vJ9zQUJRU15OvLWV4tUzyJv2MBJ2XFrFAHFCRcCZusnUGl9Wvk7p9ttJWih+/hbMnAgNG6F7fzjjWhj+nV1+ykCbKy+4YCJTpqzQzRT4/e+n8+KL53HKKYM6bat49NE5PP/8fGxbEWsQsswYScNHJOEhYpnYCkyx+eOvR1FSksW//nU6L720kClTVtCjRza33XYU0ajFyJGPUFcXwTCEI4/szUsvncczz8zD5zN1ijsgFPLh9ZosW1bHmWfux2uvlTN+/MK040zKfBmJJMjLC+D3m0QiSQwDJ89oe27TkSO74/OZ/OpX71Nd3dGyYFm6nNraCLNmVbB+fRNjxgzj3XdXOmOGkj5+EYNu3bJcIbgH2JsEoQc4BLhBKTVLRB5Cm0HTKKWUiHQ5MKSUegx4DHRA/e5urMs3x/JPZjN5/CLaItrDz+MRlIKWcAKjrpFPH3iA7//5z9ss57u/+Q0tVVVsnD+faGMjIsLA73+fc55/frNtx49fQFVVC5FIklBIp8tKJm1WrKjnwANLmTx5Gd8fPhDT2Pzt3eeBqo0ROOJ0/QG+t2YBL02cDOi0W5WVLZimwe9/f9zmnoAjjtKfFIk4wVUf0ye8Frr3ha/WQVGZTqjdwUvUGXtKJhzzqKmFXfUquO14LSj7jYAf/V4L2040N8f45z9n8f77a+jdO5ef/Ww0I0duHmto2zZL33iDDXPnUjhoECMvvrjLXK5PPTWPKVOWEwq1C45wOMGll75ORcXNHY57woRFPP74HLp1y8LrNYnleFm6oYZiTzPNZjY5Nph2nEuONRh3ow43yMryccUVh3DFFYcA8Mkn6zjllBfT62xbMXPmOk499UVaWxP4/R37Wb/cxHj11cXMmVOVTtQN7d2qnWsEj8dDVpZFW1uSlhYdiykCvXrlcu21o5gxYy3hcJJEQqWdbzoPYb/99goOOuhRXn/9AoYMKWLp0lps20onEff7TW644bDN+tFl97M3CcIKoEIpNcv5PREtCDeKSA+l1AYR6QFs2mMtdPnG2TB3Lu/963licT2JqmEItqUQ0ZldbG+QtdOnYyeT20ysHSouZsxLL7Fh7lzaNm2iaPBgCgcP7vINfPr0NeTm+p0ZCSxiMRvbtjEMg0gkgWEY9D90OCKvdvCqBIgmhCOP7p/+rZRi8uRl9O6dR21tmHhcJ3POzvalNZQt0lIPD/1Ez0wRj0JbC0RbIJSvs8I0bNSB8pkhAB6vfgrHwlD+ic4kUzZYp1arWgF/u0Yn3M6YJ3DTplZGj36SjRtbERE++UQxcWI5jz56RoekztHGRp4+9lgaVjqzSYgw/be/5cfvv09hJ636mWfmOcmq248xGPTQ1hbnvfdWcvrpQ9LLn312PgUFwbRw9Ad8lA0dyOqlGwhaTfhMmzGnd+fWx27YosZ0330fY1kqHYZgmkJ2to+FC2soK8umurqVUKg9RKG1NYZtw/33f8ySJXVAKuhdnPFGvV3v3rnk5gb48ssN+P0qfc48HqG0NIuDD+7BxImLsW09PtnUZOHxGJulZcvN9ROJJLj++inMn381F1wwkalTV6Wz4Fx77WH85CebJT1x+QbYawShUqpaRNaLyH5KqaXACcBi53MpcK/z94092EyXb5hPH3iA5vUVJFIJYpRy5qi1MbJCJOIJfCmf9O3AME16Hrbtt+7CwiCbNrXh8Qh1ddGMNRZz5mzgppuOIKe0iGt/PIS/PrWcgEfh9UBLVOhbBGf8dEx6j/r6CMuX1zFgQAEDBhSQSGjTbiSSYOp7K7lu4EJ44296Ql5/EPK7QShPmzdrqyDSrE2cFcu10FNKT6OUWwyDDiHe2oyqWkGbN8CmvDJ6Nm8gJxFud5jpPhD8zrhcTqEWnjMmwJhb0m389a8/oLq6lezs9rRz0WiCm29+l7Fjh+PxaAH19k03Ub9iBb7s7PR24ZoaXv/xj7l85swOfajDGDqel9TvzmNutbXhDnk3EwmL5auaiSR9HHj0ISSTireXR2n45Ue0tMT47LMK2toSFBQEOOus/bj22sNYvbpxs3E3ccZRzz13GP/612za2mJ4vSbxeJJYzGbYsCLKynKYP38jlqUcrVDSAe+gY/c2bWrDshS5uX5GjSpzLBMGGze2cu+9M/jiiyrWr2+itDSLlpbYZhPtBgImhmEQCvlYubKBZFIxadIPCYcTbNrURklJiKysbccRuuwe9hpB6HAD8KLjMboKuAw96v+KiFwBrAXO34Ptc9lNxFWceWumU1GxkNxYDh57FK+/tpJPJsbpyzD29yxgZHGEvgUQTkB5jUltUj+cB55ySodg+F3BRRcdwLXXvkVra7xDnFkqBmzt2iYALvnjNQwY8TYvPTGD+qYExx3bm/Nvv4ic0vaYxNQ0Pqkck6msIJZlE6r4Eh7+h64gHgUUbHC0LdOrx/8ME6pXZ9jaHLtbcx1Jb5BVWaV483vREsilNZBLTVYJwzeVUxRpBGwo6GTe9AWgcnmHRVOnrkq3M0Ug4KW1Nc7cudUcfnhPAFZMmYIZCHTYzhMKUT1vHvFwGF+oXZiNHTucO++c3sH5Ix5P4vOZm40RjhpVxpdfVlNcHKS+PkJFRTPhcILi4hA+nxefTwvP556bT3FxkMpKPdPEhg2tLF1ax6uvlnPooT1YubIepfQM89FoMq3ZPfvsfIYPLyEWs6irC5Ofn0c8btG/fyEAxcUhNm5s03Mh2yodOG8Ygtdr0NgYQSnFwIH6Rcbn02PMq1c38vzzX1Famk0w6GHduiaKioK0tMTTeUxDIU9ayKXMrymBHQp56dcvH5c9y14lCJVS84CubAO7Jl+Ty15Js2rmmZV/oSlWD0UQTnipWL6CL6d5iKhCGlUuV47YwIhABUlLUSLQv9BiVnUSa9Qojrz55q2Wr5SiZtEi1kyfjhgG/Y4/npJhW5gSyeG44/oxZEghK1bUOwEHNiY2PsMCj5/586upr49QWBjkqB+dzlE/Or3rghJxclorOfaIbnz42UZKS7UmZduKltpGxhZ8qJ1b4hE6BmuL4wHqaE6ZqdLSklnRLAY3XPcm4966i6MWvkncFyJpmKwsGUJh1Twdh+jtdJvHIjBgZIdFgYC5WYYVnTtTO4eksG0bowstT2UkKkhx442jefXVcr76amMqYgXTNHjooZMJhTrmyLzppiO4+OLXmDZtNbFYex7QcDhBMmnh8ZisX6+dgyorWzrk/7Qs7e15wAHdCAa91NdroZWZW7S+PkIkkqB37zz++c/TmDNnA889N5+Wlhi2rejZM5empijRqOX0g8LnE844Yz+n/bo+nclGHAEpNDZGCYW8jvXAoLAwSDSa5KqrDmHZsjqmT19DVpYv3UeRSILRo3s5qdtc9hbcs+Gyx3mveiJN0Tp8bYJCqGnwkzcQBo5pYs0jBiflLKN7MMLGeC45qhmFIpSXw6nH9yHwz6eRjAD0rpj9yCN8+dRTBPPiFPSJUDHjH/Q7+eccPO7KzbZVSrFkSS1z527AaqrDZ0exlcJEj0kahkk8Fifa0Jj2+OuSRBwe+zm8/wI1DTEOWRVg5qZTKF9aRla3EvwFhYw5OM6ZbUu1xpeMd24JGB4tDDPn+RPR8tIwQSm8rfX86emLyGuro7i1luKWTcQ8AYLJCMqKI/0PhLpq7Tnq9UNzPQRz4OgxHWq77LKD+N3vPkyPg4IOZejbN59hw0rS2/U5+mhWTZ2KPyen/VDDYYqGDCGQm9uhTJ/PZObMy3nppYW8/fYKioqCXHPNqA7lpRg4sJBwOIFlWWltK5m0aWyMsnx5PcOGldDWFse27Q5CMNUlIjB9+loefPBkbrhhCq2t7f1p6q7CtqGiopnLL/8PvXvnsnFjK+vWNeH1mgQCegLeYNDLgQeWEgh4yMnxsWRJHWVludTUtFJTE8a2FdnZfmfWeB0nWFXVTCq1bSCgH6ler8GECedz/PHPsHhxTdoi0KtXLs8/f266bVOmLOfOOz9gzZpGunXL4tZbj2LcODe7zDeNKwhd9igWFmtYjdGYQHx+okkDhZBsUvQ4JcCqfzRzWEkjMdtLXLwUBOIYpoey0Udhhhv1BLVdeECmaFi1inlPP8XIc9oYcNRarFgChSLadCP1ywdQOPh76W03fPklv/2/N5i+wMLwB2jc1EBcBTGw8WCjUCSSNmJ46K3Wkpu9ldvnH9fD9PE0xD1M/spCqRZuyH6FquyhNJjdOPKsH3Dm4SXwtA/imUIwMz0a2jTq9ToaI04yACcrjLLwJeNkhRuIBnNZ1fcQBq2dTW6smer8npSV9tTTOGXn6THHtkY4+EQ44xoo6BiFdMst3+HTTyuZOnVlupri4hATJ57fwQx62t//zlNHH024ri7dWm8oxFmPP95lN5imwcUXj+Tii0d2uT7F3LkbqKxsIS8vgGEY2LaiqSmKZSmqqloYMKAAESEU8hKJWGkNM7Pf4nG9fODAQpYvr6OtLZFel5oZork5Tna2j5YWnWlGjwsmMU3BNAXbtpk1qxIRIRpNIgLDhhWzcWPYyWGaxOcziMcN/H7pMNapVPu0Tg8/PBu/38P771/K559XMmdOFUOHlnDGGYPTLxqTJy/looteQymF36813uuvn0JDQ4Sbbz5yq/3lsmtxBaHLHkeQ9HPfFJU2ECpLMH0+amN+hgYtPAKG6aFgQH9MjxM8nr31VF8Vs2ZRPLCZ/keupaU6ibJ1yjJ/TpSmhT8kv18lhtfLogkTeO7Ox3hv3f7k+hJYzREKLIs4BTSRTRyvk+DEpl+ohdO6LSYZieCLNOgncnGv9idzbaWO1/N4WbAqiWUrCrIMRpQkOSW7HBVYT3j6PBJHPYI3v1R7hBoeZxb3jMFI29Jp0rr1harlOg7QyWqCsiCQjScZo2fNSgTwWAlM2yLqy8Io7IGZWwhKsfKr5cwOmzQ3himqXs8R+1fTs1vHWFvDMJg06QIWLNjIBx+soWfPHM4+e7+0k0yKvD59uH7xYuY89hhVc+ZQOHgwh19/PVklm2t5O0JDQ8RxMBGnPUJOjp+WlhjJpE04nGDcuAOZNm1VehLd9ByJpta2hg8vJh63WLWqgXA4kRaWKaeX1LLW1jjRqA5c93gk7SQzfHgJc+ZswOPRdcfjOhtPeXktoZCHQMDrmDTFiStMdHksuv3wyiuLqKpq4amnzuaEEzZPSPCb3/x/9t47TqryfP9/P6dMn+2dXRZYehdBBCxRsSP2WENiTWKJxhbTPil+YonGFHs0amJiBbuxoyhIld7rwu6yvU2fOeX5/nFmZ9mw2KKJv89vrnnxx+zpZw7Pfe7nvq/reg+QmfqhpqkkkyZ33LGIa66Z2qfbNouvFtlAmMV/FSoqw5WRrM7vwA7ZuFQFn2Yhc1Xq/2FRPukgNq1bycnBDkzLwrRVYu3tBH0aysipUDTgE/eveTwMnNyBmTCQUsvEqkRIIVgaonHFcxSNPYUlv/8921Kj0HQNzQXSTGGlUpS5u3ClUgz37EUXJtViD5OqBUVlRei3ne0Q5jUdaibCxbdD+RBob3B4fIpKWyiBSxNMKLMo8EoSpgDhQiOFePZ2OPtGmPc7R0c13JG+KVpv5lcyEAZPgJqDYPGLzhSqpkOwEFIJFCOBS9eJSYGwLVyWgQeBD2eQtVDxYgAAIABJREFU3rq9hfcW1eGq1HAVlNG+dSuvXXklsx54gIrJ+5TjW+pg3u8Yt3MN4waNgRk3gNZ/A5IrEGDaddf927/9vhg1qpjcXDehUBKPx6kfqqpA11WOP76GefOczHTZsga+9a3n2brVuVc9scLr1TjmmCHcc89SwuFUJkj2oVnK3mlU23ZqiJomMo0r9fUhHB9CJTPt7VApZEaA27Js/H43pmkd0JIJIBo1aWqKMH/+LpYv35tpNtoXu3d34XL1HYJdLpVwOEVHR5yiok+e8s/iy0M2EGbxX8fM0tNpiu6hSTqW8MX+OPUrNOrfq6ag4QOUtm0s0P0cVR1FRSLbG6nzFVL9nd8AZFrV++OXDTzsMLbVWw6/PP03p/FBQSgqHds3IfThjpuDrmayUbcmUYlhkItLsZiormeAqwMrlUJXh3CiezViVSeOCKZ0srq2BrjjXSiqcsSsIx0U52gkUwknCBo4Qt8ILDQUVYHuNrh7EdSudzLA9ibYsRLKa2DiUbDkNXjsR05HqW05x8svh+MugpVvImrX4RHgVsBWBIq0Uc0EtO5BKirLl+1CFxZtW7cRT2xGqCp6IMDiu+/mzCefdC52+yr4yfGQjDrBd+syeO9JuPCXTndpwzYI5MFhZzr/PiVTaWuL8dBDK9i6tZ1p06q48MLxn2oxNHBgLhdffBB//ONSotFUJlgVFHi5++7jM7/tIYcMYNOmK/nd7xZz//3L6epKMGxYATfddBh33rkIv18nENAxDLuP+S44BHohSPP7nAW2bSOEgqoKUikz7fWnZQjuPVOfOTluQqEUUkoqKoJs2pRIP3M9z1Tf61FVp+u2rS3GPfcs7VMX7EF+vpe2tlgfLqlpOh2pnyRDl8WXD/GvfJf/C5g8ebJcsWLFf/s0svgcsLHZ2baePbVrybVzGDf2GFpa4jw6aRy6nUB3u1CEJNdlEY2lMD25XPLRR6x48EF2vvMOQlUZccopHHLVVbj/pWlj6/NXEAw+SLTVmRZFCHyFXmxLEkk9StHIcbwwZw5N+iD+vq6UKcYiiowGbMsiip9NnimclrsYOxGlcto0Zk0tIG/JU05AEEpaRsQCRYdbXnNMaf9yM7z2AF0RiwWb40yrsjFssD0BbBRyBw4krzAHRk6Dy+/q/6Z0t8G3Bzldnorak8o4x7roNlj5tkO1CHcA0nGnT2uNSn8uViLB2xuS7O6A5ojICA7YloU7GOSmtjbnbz+cATtXZ0yDAQh19r452OluUFWDqlFwx3tO800/WLGigZNPfpJIxEg3tzodmQsWfIfy8mC/22SeAVvy+uvbuO++5bS1xTjiiGp+9KMZFBd/cma0e3cXt9zyAa+8soXCQkdz1O1W0zXBZCZzc7LBvn6CPQ0uQjjc0cbGCHl5nowYdiSSJJEwGTGikMrKXEpK/ESjBpWVQebP35XpIv3XjtuCAmcfqZRFZWUOH354Efn53j7r/OlPS7j55ndxu1VcLg3DsEgkTL71rQk89NCsT7zmLL4YhBAfSyn3YyZkM8IsvhZQUBhaNJ6hRU5ThW2a1D5zL4RaMQAroaH7/XRKDcOnYnpSPLjlFtTqEEWjy9GbU2x6/nnatmzhtMcf75MdDj3ld2x75nXyKxsRiobQVKykzaq5w5l513G4gkEKamqwt23nZLGSVKKbsHT4cD4PnJK7lCmXX8LYc891aBcXDcMpGKbf5IUA0g4QW5Y5gfDbt0DRAPJevZ8ZSiNqMkrCdqG4veRXVhGsKIfOJhhxyIFvyvtPOZlgTxAEJ/haNrx8L5xwqUOszy91bKlUFVxeEAIzkcRKpRhTCsv3kLmnip6Wi0skaFi+nKopk2H3eofI3wPTcIKtbactnbTemmXDVnjoWrjxb/2e8sUXv0w0avTJAOvrQ9xww1v84x8HcLTveQYUwcknD++jOPNpWL26iSuvfI329jixmEkiEc44QQghMg4QpinTU6ACVe3xBhSUlQVQFPB6Xfz850fwwgubeOON7dh2r3/ghReO4957T+4jNgCwfn0Tc+a8yPr1LbjdKomE8xKSm+tGUZwgmJPjJhDQ07ZM3j76oj/4waG0t8e5777lRCIpdF3hvPPGcd99J37m68/iy0G2GpvF1xJL77mH9U8/nclipG2TCoUwAoJUsQY1eRg5CqlvlNN4/SDMCi+BsjLaNm2iadWqPvtSdC8DZi5k+5KT2L4gn/Uv5fPxc8dx6I1P487Jcdwq7vodbaMPQtPCaAVu8lxhSt1hKjxhcgtz8JeU9HIPpc1+1u09TSzBdPOOphOedTV1928i9/Uu8m96gIHjRlI5ejg5BTmIziYoHgiHnHTgm2AknWP1p5qTiMDxlzjHsywnYGlupKYTKhzOroiPFgpQFPBoZMQv7bQwt9A04u3tznaq1teP0DLpbdrZR7VHCOf7qncgldzvlFpbo2zf3oHHo/VRVnG7Nd59d3+fvQ0bWvjOd15k2rRHuPLK19i9u+vA96IfSCm59dYPURSFIUPycbmcZhPTtLFtmeki7eETWpZEpvmYHo+G16vR2uo4ReTmuvnjH5dSWOjloIPK8XhUvF6NysocTJOM4kzPcb///VeZPv0xtmzpQNNUdF3j+ONrKC31pzNEm4qKIMOGFaIoCnV13Zx77lwmT/4zJ574D55/fhNSSn71q6NobLyeDRuuoLHxev7yl9n7NShl8dUjmxFm8bVDMhxmw7PPEiwrw0omad+6FWnbSCClm2Bq+L93JImkjdZtYeZpdM/Mp+ipFmyPQndDPeWTJvXZZ6BsANNvfp5EVxe2aeItLMwMbKaEW5RyPpo+i4OWLiGZW4CQksrm3YiuDiJNTYQbGnp3Nu00RxLNchpiACeQaDp84wKSNtzVDq+kjQg8ZpLj1iYYt85kpK+OwgEl5M6+BPWY8/tOR/4rpp0Gj/3YCXQ9yjk9xPrKUVBcCTc/5bhJvHwfVjTEjm4febXbKNUTCBvCgCXJkPAFgO5BcfvIrRnlZJhTToSFz6eVbHqKXgc6qfRy2wTcfZasXNlIMmmRSjnSai6XmibO7y9M/vbbOzj77OcwDAtFUVi7tplnn93Au+/O6Vfouz+EQkl27erMiBQMHJjDunUtKIrIGOzKtBlxz+F7GmQ8Hi2tPCOJRk3cbo1gUGfevE3EYgalpX6qq/MIBt1s2dLGffct5yc/ORyAF1/czN/+tjYdLNWMks3HHzcycmQh8biVIdZ3dcU56aRh/OhH7+L1alRUBInFDH7zmw8wDItzzhmLrqtZdZn/MrIZYRZfO8Tb20FKFE0jf8gQikePRvN6UXw6nDKcsn9cQWD6qMygLaIWoSk57Pn1YFoenMyiI/ewnW397tuTl4evqKjPwDw/Cu/FwFB1dMvAlUqCZVGfX0YqHife0UHX7t29yinn/8zp5uzp7pS2EwTn/Ab8Qe7ugOfDjrFv+eJ3GH72kdQ/8iBLPt7Fcx+289TiGG+8sQ7p661lNq9dy9Jvz2L3sUW0zi6j6/fXOoHu8G+S8RW0TGck9wbh4tucDXOL4OTvwU1P0Bj1UJLYg1c1sG0nGDQlfQTcAk0V2EIjhRfDkGzNO5ML7qxhzWbgqgccpZlE1KkzWpYzVery9KZTGWqDhOFTwNO3bvfmm9s599y5GUd307RJJk0ikSTJpMWJJ/ZKqkkpufrqf2JZNoGAG59PJxBwE4sZXHPNG5/5OfF4NDRNzdT8DMPG79fx+13k5npwubQ+GWEP5cKy7Axh3+NxXjDWrWti6dIGOjocdZm6ujCLFtWxa1cnRUU+XnllaybLffjhlUgpSSYtwmHn+txulVjM4JxzxhII6CxcuIeVKxuprAxSVxciFkuxfXsHH31Ux44dHei6ykMPfZyxacriv4tsRpjF1w7+0lIUVcVKpVBdLvKqq8mrribU3kz0mmPwVFahWALd68WIxTDLgti5GrZpIEoCtAS7eIF5HMcJBMlBQVDBAFz037n4RgQ8AhIFRew+/ESq33sVoWqYbi+m7kKXkpZ161j5l78w+bvfdaYj7/0Ynv8DrHoLcovh9GthwlFEbXg5DCUq6LXbqPrlNZhCJZmTR7ei4N6+ge7aWhp8PprXrqVswgSaVq+m4QcnMcnbDArIJIjX/0Ry1au4H14Ho6fB83dDtBuqx8C3fwPjDu9zDZHS0dyzayI/KqxDSBsLhbpYkBbhZ8TADtbXG0RiOra3iLbxV9E27mqMGFxzK7z2YBD/H5bAhkWwYxUMHOMEwvuvdpporHTLrepy+JKX9W3u6eyM8/3vv4Zp2uTne+jqSmDbpPl5FsOGFXL33cdl1u/uTlBXF9pPZNrr1Vi9uvkzPydut8bs2cOZO3cTZWUBdL33vX7AgCAbNrRmvu8rom3bTkDsUZTRNIVoNEUqLWzQowUqpWTr1g7KygJ9pnobGkIkkyaKomS0ZxMJR5P0wQeXs2NHV6YT9J13dqUFutWMH2I4nCIUamfAgCCRSCrbIfo1QDYQZvG1g+71ctCll7Ls3ntxB4NoHg+Jzk5cbg8jfNPYxE58qp/S8eNpa9hFd6GBEjJhcC6q7gEhiBLlBeZRSikKKooBQ19PYT/zHgV5Hoaf9U38h58Cmo5bOC5GOZUD2HHUbLryi6la9j7CskjlFlAcdOqP6558koMvuwyhKA6dYM4vnX/7IGQ5+9IEeF9+BmGamMEAXiNBkCQelyCRTJDo6qJz507KJkxgzZ/u4ihPM5ZQkJlJGhu9ZRe8cj+cdT3MvuoT79k99yxlQ4uPXYEC2s0ghmkhpUBgU5zrpcs1m7mF9zNyRO9UbE4AmttgyRo4ZhpQOMCxaWreBWMPd2gdm5fBR89DqN1p7DnqPMfBIo2NG1u58ca3qK8PIYTD+yso8JFImKRSJsXFfp544nTy8nqbcVwuFVVVMuLWPbAs9vMM/DRce+2hdHUleO+92owiTVlZIDMlWzpWUjJeYidUGhdrhNtsNE0weXIFjY0R6utDaFqvEa+iiIxIuNNharNjRyff/vaEjCdlD5HeYZGINCXDqbFu395JMOjOzDjYtk1bWxxdl2n1GhVdV0kkTEKh5KfSSrL4zyAbCLP4WmLid76DOyeH1Y89RiIUYuDhhzP5+98nt2AQbpaynnXYLou8wYMQdBHzx1BRcYZ+Z1Cz0x9fws3ejatoLolRunQBO0Mp1rz4GrO/OYOC7/6aU8on8HbUTV5pGTnxBE3xGDunfINARyuX//46CoYORQhBvKODvbEUf4l5mB8Dn4CzXQkmL/4nDR9+iL+4mFHnX0CuWk3UhtzmRmxFpTLaREmyA1IpVLdFUgGXFSNQ6siceXcsQXjZJwgCQsG2LeyPXkA563rnb6F2eO0hh7A/ahocdT4oCuFwkpde2kLKVY0j1m0jZa9Em9ejk6e7uSl1NfltxSwJnM4W96GQ1r+MJ4EPnoO5d6UNfQW8+gCcdDmceBmMO6zf32jjxlYuueQlWlqiGQ+/cNiRMHPMeB13hWHD+lItfD4XM2YMZMGC2kwnpm07U6nnnz+u32MdCF6vzh13HEtjY5jW1hhtbTFuu20h9Q3dHP5zm5qTJZrLUSEyoibvXKeTqPXgcqlUVubQ3h5LBzYn43Pqi710CNuWFBV5ufrqqYCTDfr9blyu3i5RZzvIy3On3Up6g3uPJ6Fh2HR3J9PnrKFpCvn53gxNI4v/LrKBMItPhCmh0YSgAnn/oWY22zRZdu+9bHj2WaRl4crJYdA3vkFBTQ1mMknZR0l8rUF8w6oYNGEqTyh/I0rUkWrD0S+VSFRUEiQwd7cg27tRSrxw2GACC+uIdnXz0QvvMKtrO9MKB/DtCx7gidzRUDUY2xb4G+u56MO5lIwcCTiGtMGaoVze4aHFdOp/srmRVd87k91b1qJJG2maLPjVrzjpqJm8dN1dtBxyBJUfvEpRqo2uiImd6NUUPbookfFF1IvKIbKrbyOqlAhFIAK5zvd1H8CvT3foFFLC24/D3DvhroW0tpq4jBAVBQlebx3CiYXbSCkqlhQENAO/W+FkOZ+2ZDG5sY1MSMznNf/lzIudSrTTJn/7R9irf4uSUwB6OkOxTPjnn2HiMY5aTj946KEVgGDAgBza2mKZ+mA0mgJcmKbNt741fj/+HMDf/346xx33BJs3t6MoznTl9OlVfaZQPw/Ky4MZnuIRR1Tz1rYVvGq/R3drfrpGKPD6uznurhi5r4zl/fl7UFWFgQNzaGuLU1oaoLa2i0DAha6rGU6fx6Px1ltzKChwrqGjI8727e1p+bVeKIqgoMBLJNIruyalzAQ/Zx3nOuNxk+rqXA499JNVkbL4zyEbCLPoFwkb3gjDPe2wKQXtNrgEnBSA35dBhf7p+/iiWH7//az529/wl5Sg6jqpWIz3f/lLpG2z8pFHiDQ1OZw4VaV0wgQm3/t9XvLUY2GhoGQyQhUVHZ1YayuqkJiKQLoUiEfwaTZ7u8GKx1A1nasf+w5n/PAfbA9HUcwGtj5yF627W1nRZaPEOvATRStvpvL73yTnO1cRmXoEVff+L7lb1mGoKiIaz3RldM9/m9kN51L/4zso9kNXUxzZQ+KW4PUINtYlGbt5FYFxh1B2xf+Quv0kdMXEEukGHNtCceuIWVc63393keNo0dNlKiU0bEP+7WfsTQ7l4C33oCqCLiPOu+2C4tI8LMVD2FvCLDbSJXKIoWPauahmiiMj9/FY6ximqO8RfvSvdOa2kjvhkN4BQdWcJqCNiw4YCFevbsI0bRTFEehuaYlhWXa6axTOOWcMP/vZEf1uW1joY8WKy/noozo2b25jypSKz9wt+mnQNIWtWgcyFUQBFEUikSTjeZRVqdxw+0TOXj6W117bhmXZHH/8UKZPr+TSS19h7tyNGTUZl0vlkUdmZ4IgOI0yPU4S0OuIZZqSjo44fr9GLJbC69X72EmpqmPdpKpOlhgKJbn88oO/lOvN4t9HNhBm0Qc7knBWHaxPQY+7XJouTkw63ZCL47B2COR9BU+PmUyy4Zln8JeUYFsWHdu2EWtvR1oWL19+OTkDBpBTUQE4b9yNK1dS+fdVjL50DFvZgo3EjYckSQSCIDnENRVbVRA2eNc0g21io6AqtmPo6/ZCaz0DbpjCANMAIRjnc3OfOgzLDuG2k3RpRUS6fHgb91J1y3XU3XwHOQvfAQFaNNLbmghgGBDqYtITf2TGMeN5sW4vhunQBHJyPPhz/ES6o2x/610mjjuEwcccS93OWyma9zNU2yGCK24XyunXODzDPRsd8v2+nZpCgObCfP9ZNm6voKSqmD07WtHQaemOs607wabAFO46aCk2KqmUTUleEqF209kcxiU6ODXyU/y5Nu78AqzEXvas3UKXXsL2aAHNFDHQZfGNhEpuzzGlhOZaUDVeWRxl06Y24nEDTXO0Oaurc9MZkOSxx07jmGMG9yt713sJghkzBjJjxsADrvNFsXGHwDdUoOalxbMlqJqCKT0oQmHq1EqmTq3ss83jj5/GD34wlRdf3ExOjps5c8ZTUtJbU02lLJYta8DtVolE+tI7VVUQjRr84hdHZpRxDMN5IXCmf51MUEqZuVcHH1zxpV93Fl8M2UCYRQbbU3DoLuj4l47unn45DbCBNgv+1AH/U/Lln0MqHMYynGDUvHo1qUgE2zSxbZtUJIIQgmB5eaaZwZufz5aXXuKsS59nMR+xmU3Y2FhY2FgkSaAPLSNab1D0yCq0tihSSqJJybhKDVFY7pDDG3c4Ed8bxLYt9FSEq4as59buwSSVcoTmgYRFLK7gzvFR/NTDIARKKoWw92+Bj7e3E+voIP6NmQT88/Hn7sMXtEwUVSOc6N2u6rIbqZt5Mc3zHiPHZTLo7AtQy6vSSw9UR5IYsRiqy0V1qgm3r536ZC4J00OZ0sr3hj7PRL2bBrOIoDdAlbUHPdZNIJXC7VIJGjo7dwUI6DkMckM03MZtTUeyvisfW4LPXUReKML9Q1oZ6Wmi+bfXEq3bQyplYYfzmDLwTBZtVzMuCXv3hqmuzuWGG6Yzc2b/WeR/Co0rhjNs5HZUTeJVnWFO8SSId/vIs4oOOPJNmlTOpEnl/S5TFJHWK3Wevd5GH4daIiWUlATYvPkqVqzYS0NDiEsueTnTHOR2O0ID0ajBeed9vlpoFl8tsoEwiwx+2wadtvNQOAY0vbDoOxwvin015+AtKMBbWEioro5kJII0TboqqonmF5OzexuitZVYW1tf2x8h0NE5giOZxnRMTNy4aaWVWnZBvqTjry9Q/8JOomYKmZQMLlI55OCBkFfiiErbljMdGA+DBLewcUmDgvgemjzOoO5WBfFIknhpIb6GWjqPPIGyv97T73UYsRhWIkHpqRciH3gEO5nISGtJBHZhBRXpGqGUkj/9aSl///s6IBfDMLEefpWLLprIpZdOIqdqBBRUQGdjb1YoJZgGjZ5hKJqGCLVT7olR7o2DZRE1JNV+HUUIKt3dVKkxQglI2Apu1SYUtYk315Ev/GzZWkYyUIQvmMfGrjzK3FFAUGeX4LFUfnbzm/ze+zvaWiIkVB/JpEUpbVzneoKOYT9g554YyaSJbUtmzx75uRtevgqM9Q1i50djqZy+wRFFR5CMuBCLT8B18BdrUNE0hZNPHsYjj6zcxzLKcbHQdUF+voeKigAej8ZhhzlZ7vz5tfz972syrvamaVNdncdVV035XMe20opAqppVnfkqkA2EWWSwJO781xY4QW/fPEfSaxcrgOqvqOtbKArTr7+eF+bMIabqzL/+TurGT0WxLKSiMva1f3Dce/PwFxcjpSTe1cXY887LbK+nPwCl6U+ss43NeSMwTzwTj0gxptJFReNSR5cz0uWIVguc4JLmhtmWRBU2BV6LFjOJrXmQtmRgjgsjHqalqoatV/6MiucehUh4/wuRktDevXRs387Yc89j7bwXcekqQneRFDol4ydSfeSRACxaVMcTT6yluNjHzp2d1NZ2YxgWN930Nv/7vwu4445juezGv8IvT4V4tFd2beBojCOuJ3XHXbg1HZEEbIukJXFrggKv84sZ7nz0RAdIHSuRpDMueHcr6FgY0sAX2kODVcRrsaOJaUFMNZeI9BEzFQa4FGo3N7CtOEUgEEQTkErZdNl+irQIo9VdFE45CMuStLdHOfLI6k+cDv2ikEg6aMfApIiiTIfwgXDFeYLv/ORImhaPoXBoE/GIm/iuah766b/34F577aHs2NHJs89uyDTMaJpCaamfwYPz95vmfeihWRx+eBV//vNKwuEks2YN50c/Ogyf77Odx4YNLZx66tPs2tUFSMrLgzz55BkcccSgf+s6suiLbCDMIoNyFTbiBMD+3pltQMfhyP2wf/OBLwVDZs5k/Jw5/CbqY8/EaQTbm1EVFVtRWTv7W5S2NRLYux2hKFRMnsy4Cy7odz8mJlubVvLBXbfDqkZcSRUrmaRhq5/Z/3MrBXs/doLgqEPh+d/3KtXgGAQLCZOqYNe2CBHDRNG8lAdshJXk0puupHpskPmH1rBk/mrsfYxiAYSqEsgN8v5lF3DasYMpGetlY20Yo3oUNWd8k1FnnonmdiTKXnllC7qu0t2dpLa2G8uy0XUF23aaMG666W2mvPdtJj62A954GNr3wpgZMO00hhoG2+cvoO7dTtREJ0iJqgiOHqHyROR8/hq7mBB5DFG3c4L8M9auBaSETtxOYEsLpIWuSLr1fHKGBLn6rPeYMaUWw1R59e1RfLz5DLqSSXRhZWpiPTw4RUjUWHeab+fUQA80rfhJMDBYzjI2swkdnYOZwkhGoqSfwm66eJM36KCDBAliRPHio5JKpjODAVTut8+qcnjmbnjxnSLWbi1i+CA44zKo+Den84NBN489dipnnDGKW2/9gEjEwOvVGDAghzvumJnmL/bFhRdO4MILJ3zuY6VSFlOnPkIsZmQEw/fuDXPccX+ntvZayso+QZ4vi8+FbCDMIoOL82FlordGqNHbMKMCuoBiFe4th1FfQAwj3NhIeO9e8qqr8RUVfeK6E6/9ITs+rCPY1e5MB0kQloEnlWT35TdwU2oX+TU1lE+a5AzEhkGyuxtPXh6KprGW1bzFm0SD3cjrB6FHqyma20rOom6ira0smfc6J913n3OwaAhefRBS8YzLggLYQjCi0ORoK8XizjzyggpFNdVMufJKBs6YDgueYXJZig1BL/FEKpMhqJqK4vHhT7WTArY1G0w7dCg1NV2gxeHcbzryZWmkUhaKIqivd4Jgz6AHEl1XSCYtHnzwYx58cBaceX2f+6S53XjOuI5XPghyVP6rHOXfzNAiydPGxfw58n28WjdaZBs7lBzu0u7kROtiyo31jswYzmG8ebl0lk3k97+ci1sP09LuR8HmrFnrmTAmypOpkygVUbql42yh6wput4K0YGOshGRTBCHgxz8+jLy8z/5gmJi00Mw8nqOTrkx+t5taJjCRWcxGInmNV2mlhQhRUjh0BAOTOvbwAk2cyTcpZ/8AXJQPl579mU/nM0MIwezZI5g1azg7djgGwUOHFnzpmfAddyxMNyMJRNrpRFEcTuJPf/ouf/nLqV/q8f7/jGwgzCKDs3Og2YA726HFcuqCVRo8VgEjXBCVMEz/VF/W/WAmErz7k5+w9ZVXQAg0j4ex553HjJtucro2+4FWXEpggIRQB9JypgIVXcfj1vGMHseY4b31tdWPP86KRx9n6d48VnZXwYBSCs83mXyJQFcNCOoYPmg7qxh3bQKfXUj9kiVI23ZUYvw5cPT5sOw16HZ0ToXmQlVV/IUDOHp8BTNVleTNz7Fh7jyW/uEPrHz4YY7KqyO3tJAjjtBYuGg7lhV3ZLeQ5BYEcYkopq0Qi6b5g/486GyGTUtgwjeore1iw4YWKitzeP/92n188nrUTUDTVAzDpqUlmrk3sbY2jFiM4IABxOIWv/jVQrwV4/jAO4m48TEXyGd4sP0iOjrWEAq19yjlYITSAAAgAElEQVSEorn3UqtOoCCyClVIbJcXv8eDFY9y2WVxCvPj1O3NwTQtNFWjs9vLiCFbOfGkAI2bfIzp3EF9RzGppEWBMHg7MYGZF86kvDzAzJlDqKn57NMEW9nCh3xAJx2ECKGi4sKNQGBhsZ51TMEhsTfTRIwYKVLIdOXawqSLLhQU5vIs3+OKzJT4Z4VEkiSJlv58XiiKYNiwwn6XxeOOsPbTTztTqMceO4Rbbz2G0tLPnsWtWdOMlGSCYOa8pWTjxtYDbJXFF0E2EGaRgSLgB0VwUT7sNaFMg9wvoTb/zo9/zMqHH0aoamZQXn7vveRWVzP+/PP73SZPgfFVpexwH4zcU4uVSuHNzyNRNYRjC3pdD9Y9/RQfvv1X1s0eR1PMj/uDFC1Ld9P0SBH1i7ycfp9A99lIl4IVVIkeFEDdGUIPBJzO1LVrWf/UU4R3dVLZlcsYTxKvEksLXJuIjkaEkcT05fPKnAvpaGjCnZOD3dzMXrERyoMcdNBYSkqCzJ23ElVVCLgFrsJcZHsYyxYMrN53sJTIRJTf3rGQuXM30iPR1d2dRFUdRRMpLWc/AYeULoTghBOGEu/o4IOf/5TEqvdRBURzKtFPux7TtPB6Ha7bcv1gmo1SGjvbiHW14fX5ER4fUkIi1EV9MsZkRSDcXlTNhTRN/MXFFCjrGT68mMrqfIyUTVdnC3vqEthSsHx1gHtXnkKV3sBpbCKWcPFSy0g+DA9ldk0rv/jFkZ8rG2qmmfm8ixs3KZyXBBubFCncuFFQSJGijt0UU5IRS5B92rfIbNdKC2/wT07hs2dIjezlAxbQSScSmyoGcjCTKaTogJq0nxVSSk4//Wk++GBPpsv0qafWs2DBblav/t5nllWbPLkibddk7xcMJ0z4cjiXWTjIBsIs9kNQhRFfUnOaEY+z+tFHUVQVoWmZDDAZDrPigQcOGAiFgJ8WC64wC0nmF+IRkJQwQINvpx1rpJS81/EK7T8bhxLzMRCbQbN1ds/zUvdmF6bPRe0SN8OOiiMsie1RsF2CWHs7E+bMYdPzz7Po9ttBVTFshR2dgg2m4KzhFn5NOE0pyRik4iRaW+hoNAlWDcqcY5MxiKK2TZhVMQYMyGfmMaNY9OFmDNWLLV0YMYuKAYUMGZyeBrZMQPBuQzGPP74M07TRdZWKimA68Onk53vYtKkNVVUwDMcl4eCDyzn77NEsufocxjW+hyvHMeq1zb28+uefYdkX9bl3tVoV7d3NuNz5CL8TIAUgfAU0JErIm3AwyVZH3NpfWoq3oIDWLTtQhCTgdyH9sGNbGE2VaJpk/aZCtu7wsCYxnMX5lYQiAaR0dD1ffXUrb721g+OPH0p/SCRMbr/9Q557bhNCwLnnjmHKTQbC5XT6in0+PZJ4Pd89eAgQyPz9k7CWNUznMIIE2c42Wmkln3yGMRwvfZVtOunkVV7BxiZMmBhRmmhiBcvJp4CpHMohTM3UKD8vli6tZ+HCOvx+V6ZT2O1WaWyMcP/9y7npphmfaT833jiN2277kHA4hao6129ZErdb49Zbj/lC55ZF/8gGwiy+UuxZtIhUNIq0LEjzrzSvF6GqhOrrP3Hb8R54phJeDMPuFBzkhZMDkJMO0i1GI51H5pLoNEiGJKoiUdySkdf6qT7XxJAmnhwBPjfSTCGSEt7fg7ewkPVPPknrpk0oLhcRVwntMR2ERtwwWOdNMXWQ6kyb4tQMPTLJQH+Kzn3Or1krIyfeRE5bA1owyJgKjaLTp7A5bzJJQzJIjmJIfAtqpN0JqraNPP5Srr15BVu3dmT2s2NHB2PGFGNZNi+9dB4ffeR0kcZiKU4/fRRz5kzA3ruNEY3voeoukiLNi1MtTs7bzNyGWsJhP8GgkylbpokuQyhqIabpWBkaJiiKgu73oeouSsc5FIeQL4dXh0+hZcp4SlO3M8Tag2uNpHR5O1qZzRurhrFhawGK4mRjXSEnqPR4/pmmzbPPbug3ENq2zZFHPs66dc0ZN4Zbb13IkLfc/OB9PyiQSy4JEn22MzHx4GEYI/bTj+0PPd2jG9nAbnYTIYxAYGIwn3eooJIqqhjLWILksIH12FhEiJAgTrpaio1NB+0s5ENcuJjEF1N+WbiwLnOPMueYzpgXLKj9zIFQVVVWr/4ep5zyFJs3twEwZEg+zzxzVh+1myz+fWQDYRZfKd7/xS+cIJiGtG2MaBTF5SJ/8OBP3X6ADlceoPS0Sd+C9KroHSnS/qsEa1Q0D0hbIdEl8BVIpFeg2F5GxKooqnSzc/0b6IEAQlVJxg3szlrcuYNoooyUUFnTtp0pA2OoIk0aUTUENlVaZ59AaAMr4xUM+s6vcHssCBZQOnwKpVq6VmXbyA2LMBb/E83vQxxyEq9u9rNly3Npqx8nOFiWzfr1rYwbV4JlSc47b9x+hOvmNxeSDATZU1qDpWr44lGKO5vJMRPcMGEzv2oaTXNzJF1TglknDmDNui6ihk48KfEHDXyeEFNGDEbZ/R7R1laM4jLuP+27hHOCVFYkubXtSm576joqOxvJS9iE1roY2drG6EAba7uKEEIikJgWCOEEDymdDKU/zJ27ifXrmzPC2j3r7/g4wZo3FKad5CeHHOLECRECnA5SDx7O4hw8eAjRjY6OgdHvMQAUFNy42ckO4sTw4KWTDrrpBiBBgg7a2cQGTuNMuugEBAmS6QArM9OuEkmEMO8zn4kc9IWywqqqnAMskVRWHmhZ/xg8OJ/166/43OeQxedDNhBm8ZUhtHcvzWvWOClJjwlcGtI0mX7jjf/W/pMijlISRN/bhEdJYegu9IDANsEWGqoqCBapKEiGKEM4zXcGf3/3OPwlJcj0+SRTkrAI8nr3IYSVXDTb4IPUZFp3beGnY9chFAGKikhE0RWINuzBU1CE4vYQbW2lcMQICo44oa/eVhpL7rmHRXfcQaKrC93nY9IlER7dehCqCpbVUy0VqKqCadpYljygU/lHeQMYnFeELQSqZRLz+tnjGczIXV2MHl/JPx+9gMWL60kmTSZNKsflUrn00pfZUd9GWHRjS4marzDmThgYuQoeW8O84ACS5XmMyO/CFfBw9HsLMNtdLMuZgndvN+2tUXKUONeVL+SyyCxAw7J7dTZt28bl0rj00oP6Pef339+FbdOnfiiEQEjB7nc1xp4URkcnQBA3bqoZxBBqGMHITPOKho6FdcAaITjcUR8+YkRx4aaRRpIkMuvHiBEkSApYxhLKKGcXuxCAvU8Q7EHPlOkSFpNDkCqqye0VmvtUnHnmKG644W06O+MZOkUiYaLrKldc8fmI9Fn8Z5ANhFl8ZejY5rjEu4NBUpFIb2YoJa6cHIaecEKf9a1Uind+8hM2PvMMZjJJ5fTpnHD33eQP6V+uayCDWRZcjjKxkrJdXXTGAcWLbQuQCgPGa7hdKrlmEGNXK6/+4VoiLc24g0FUlwvd50PpbuU9jiOGG5+aQFFMyrwG8xpHMb2si6NL6sFIIlHo6IzQvnkDlmXj9vsYee4FHPWbW9M+dVba8dwZ+D5+5BHe/clPUHUdVyCAZRgsvece6kuvQ1V9KIpMe+D1DMKS2bOH95lO60HEhocGzOAuTcWVimOqLlRpYAmFWG4hA8+4CN3v2k/W7K9Pzubn8x+hcaOP8kE6Y4/z484V7CTEqb/+Lo9/HMWLRdz0oUubiZuXE0uLeg8cphDqdtMRFwx2deJOmfiIY9gKSbwYho2mKdx44zQmTepfM7O0NNBvh7EQghllE5hBHrXswoefsYylrB8KhA8fLlxY2JgYfUTVwckGc8ljJKOoo4522rHZZwYifX9bacWDh60kGcVoAgToomuf+98XFhZv8M/MtGslVZzPhfjx97v+vtA0lddfP59vfnMue/Y4WWlurofbbz8m2+TyNUU2EGbxlaFo5EhUXYe0JqhtGI5VTzJJ9eGH79dp+I+TT6Zu0SJUlwuhKOx86y0ePewwvrdmTR9JNYlkO9tYy2pAkgqCMr6AQgSmNMGCwnFODcVMpWjZswX10XVoK+sxWhuJdrThcvswEwmiSpAOuwAfEaT0kvCX06HE8YsIL+6t4eiKZmxFo7GpjW3tASor/ZimJB6JYmxeQUx6OPvkf/Dhh3uwLJuamgLuu+8kVt1+O4qqZkjzmsvpFKxuWcBGTsDv17EsiWna2LaNomhcfvnkPvejk06aaWJ+xMNadyUPHncz33/7t4hUtyPxpessO+ZSakbun2VIKVleu51Kq5nvWh9SumAP5lIPtd+Yyt4BecT+8HsKR1+GPW4mNHSyTcsnjhtvrBvp9pITbuDQCcNYuqIZpKRQi+MXCpGkQmWwjeEnnsjPf34Eo0YV73fsHlx22ST+8IclxOMGHo8z1PRYG333kikU4GM8n040H89EVrICO22vJdL6RioKoxjNNKYzgEpW8nE6eCn71RRtbGLpz195jCBBqqmmll2YGbbsgVFPPf/gCS7ne5+6LsDYsaWsX38FGza0EI0ajBtXgt+fNeH9uiIbCLP4yhAoLWXEaaex8bnnkJaFUFVs00TzePjGr37VZ929K1ZQ/9FH6H5/RsRZ0TQS3d0svvtuZt52G800s5UtbGULLbRgkOrjQQiAcAjtRjp7MDq6kds6cG2LohblIYZLUuvrMSNxPLm5KD4/MiKwhQsLlQgBUlouNcU2CZ8Xc0iYzY27eH9kEZ4qBd/aVlyKRM8L0LppC6dNv42V9d60Ea3Gzp2dnHrq09xgtmaCXw9UXWdccjVNE+ewak1rxg1dURSuvnpqpn4kkSxiIevlOnakoEuDoyq8PKWcwjtlkzhn+9tUkuSDmiMZPeag3mlZIwVt9bQ0hfnhDe/Q3lDP/5bPI+WBvdV+8nTBsNfmM7I7zG5ZxJHrFvLeuGOJe/xUJduYXz6D81qfpYg23NEYdmgto30mC8IjkC4vhS6T2eWtjBDbOPvXN1Mw9MBBEKCiIocnnjidyy9/hVDIoUnk5Xl47LFTKSjwfebnaDrTaaKRbrpIkcLAwIWLEzmZ0YzJrDeBiSxmEe20f+L+VFQiRDAwOZ8LeYp/ZJ6Xfw2gPc+XRNJME220UsQnX3cPFEUwblzpZ77OLP57yAbCLL5SzHrgAfwlJWx4+mmMRIKCYcM47s47KT+ob12pbvFihzS/z1xaj8p//ZIlrGMti1hIlCjddPWug8g0U0gkSvpjYeHCReovq1E+qkcoPkBBq8jD2N2G7I4DkK90E1TjJG0XbitGdV6K8pGDaW2NMfG747h7/kck14ewJeBR8Zw1jMoXdqCkIJTSaN/TSCBnZCa79ftdRCJJYq5c/MlOFF/vgG8ZBr6AlzffnsOTT65n7txN5OS4uPzyg/tMa+6mlnWsJWz62ZtSUAV4lBgzS95gXuoCflM8jPFu8CpwS66T/W14eh7rn3yOPLuTp9flsi1azCUD11BQlCRWGURiERUCVbUpDMcwFRcDU/Vc8/Kd/PWYS+gKFvDKiJM5rH0px2z+AIQLacQRUkGUlPKz0q2Z8ws3SppWr6ZgaP+UiT6//6wR1NcPY+HCPSiKwvTplX1+48+CIDmcw3lsYTMttJBPPiMZRYC+5PQkSYZQQxttn1hTtLBw4yZJgjbamMAE1rEekd5Hf9v1BMROOj5zIMzi/zvIBsIsvlJoHg/H3nEHR99yC2YigSsY7Jd8XTBsWP87kJKc8UNZzCJ0dMLp7sLMYiSpdGaooCKx0zUlJ0MUa1ogaqIEBdgSMxZHagJcKimfRIZTHJ2ziNfCRxOxfMQamoh0R5kwNp/QCw9jt3eg5fsxkilIWSRe2kLXlAryF+wmLl10KYX7XY8Qgi3lszhoz98cmyS3G9swsAyDaddfj9/v5rJLD+KyC7dC7C9g/xa6J4D/WtDHsZnNqKjsNZ2AIQDD9pKjhSh0tdOaKkIX8OdyqFJtbvrOo7wydxVJswhF5tNpBRiX38Xg6m5koQuZkthIDNFzx0BN2JgKTNi9mT8++gMK/CHaYx4mTSpBGT4dUnGSbe1E6powRK9ii9OVKnAFg5/5GVAU5d8WifbgYQIT+10mkbzFm6xgWR/1mQNh3+VddDGFqTSwFxWVKBF6eoN7+Iw96ysolJN1lf+/iGwgzOI/AtXlQnUduEYy9LjjyKmspLuuDt3nQwiBEY+j6DpDrruQVjYSJfKJg5zTHyowsZDYJEignDcecctCzHAXqV0tkDTBskFTsDsjUOCltDTOOQ0vszNWRSzho0aLMrQ9RvfaWjxTa1A0BTPVgXSpEDfpEhItkqJ65imEXs/DL2WfYCilpPL42ZxyyOG8/8tfEmlsxJufz9Qf/pBp117rrBR7GGJ/BiUPlFIwN0P3dyHvb9ianXH56IFAoAhBriIJA+flwhgPvPDkcp57Zg05dgivpmAaFnFbsrPbQ9vYYkbubUd43QgpEbYCngC2bKch6SbgUvGoFkiFklALVmAYilrgCMu6vbhTSUIiSKw7zLLuQSxqKCSShEGBcg4tHtXn3u9ui/PHJc1sj1hMLXFz1bQycr3/meFlHWtZyhI01E8NgkCal+isN5jBlFPBURzNIhbiwYuLaCagOi9UAg2VCUzcLwvN4v8GsoEwi68FhKJw4Vtv8fz559O0ejVSSgKlpZx47734q2uAjaTSXLL+pr16mijyKcCLJyPLdfypv+DFNy8j/reljqWGKiDoQRo28swRyIsnOM08mmDMskaUPyxDr4vjGzaO7t02Vn0HSk0p7vwCzGgEQzNQ4pIZt9zOqIu/xwNH/o3Vq5vweFSEUEgkDPx+F9dfP42GUAvBB55gbEGQI8eP7O0IlTHs6GNEWk1irdtRXTrBsnI8wSTE/86w4PnsZhcDdEmLKZCAS0mQtLy0JQtxKXBOmo725P3voMkUmu6oDOiKxI1B3FR53zecKa4dBKJRwh4vbizmby7m/k3HEOtQwZYcW1LLBVWbWOqazGFFEUjGHUHwUDuqEFTc/EfevuoVFjfk49eS5Hk0IiVjueKad3nqqTMdndQtnZyxLkbS40MEBO/GJA+8spePji1hYP4XUGf/nFjC4sxLwydNie4LgxTlVDCM4QCMZBQ1DKWbLlRUdrCDxSwiTBg/AaZyKFM59Ms/+eT7ELoBzK0gvOC9AIK/BSU7NP8nkb3bWfxHICXsNpw4VKn1S7sjb+BALl64kFh7O0YsRm6V49BuYWV4YtDPYCdBxEz4oA5jQTvuOTPQR+cwhakMV0YyMFbBrupiDAw0W0U1BbHZVcg5YyGccjroVeDoauzmKF1/bSJVXoO6cT1mWxi9phQpBCQFSkIy/dvXMPaESwB4661vcc01b/Dyy5tJpWwOOWQAv73zKK79YAUrcspRpA87ZjN04wKeOvEgivNzMWJ1hHZuJ9JsIFQnIMVa2ygcNoBA+WaGUEMNQ9mp7qTKbdNhCUyps6DlOCyhcGcJlKdnK6MtLQhNB5lw1NAVBb9mEDc9tL7UyJM5boJbu6gIhrGGD+OX245GVnsorQkQzi3hr4UnsaHcxeM/PBzXxy/Aqw/AzjWOEk5eGdqa99kVmMzQSY7LusvvRygKzc0Rnn12A9ddN43vLu/EyPUTSPSS3jv8Xq5d0MTzpw3q8xtvje5id+RNhIxQ4pvB2MDUz10z/FckSCDSxPfPEgQ1NIZQw6mc1ocwr6Nn6n+TKWAyXzHnL7USOk8DaQJekCmIPQhWHRQ899UeO4s++NoFQiGECqwAGqSUs4QQg4GngULgY+BbUsrUf/Mcs/h82JCAn7ZCY3qcrHHBrSUw6AAzpb7CQijsFapWUTmJWbzKS4QIZQjWCgpCCsQ7u5FXvYZV101IQPi+d6j59eVUzDyMF+++iF0vv4miKGimicvrdeYcZw2FhAlm2uRWgmyJkzpjAs/X/AqZW8SwqtEc8uyfULbuxdrVDpYFisqyb/+EkgcLCFVM4Xe/W8yqVU0MGpTHpZdO4tJLJ/GrVz9kWW4FhfFwRmR8W04JN721gsfOOYad76wmN5BA83rpjrvp/n/snXeYVOX5/j+nTd+d7YWyS+9NQVRQsPfYosYoYsMWY4wlBmNPNLYYjBoxMRr5WbChYAURsICA0ntZ2F2W7W12duqZc877++PMzu7SRKN+E937uuaCnTnlPfV5n3bfugpCIGoqceSehQOFEzmZaqmKKnk3pbtaWWsOYkpWNyb7oWeH8zbxEC8bK6MISUayLJBkdFRy5WZGrVmA2TedYK8sWuMGzUsbcReGEdl5bDr3FERWPgjBZ6bglw0Onht7Lr0WvgT5vSEzFwTUrNuAUeNlZ9pg6urjIELk5zkpyFHYurWB2kCM3ZlpeKOdH0l3XOdz2dnpu0UNH5Abu49ukh2OlYKvsjhyIkfnPZCks/t2KKaYtazBOAgaNoHAwmIbW3mCxzmV0w+qheN7Qeg+EAmQ28KtDrAU0OeCsRvUvXUWu/D94D+bin0/uBHY3OHvh4FpQoh+QDNw5f/JqLrwrVBvwPm7YUkEKhIQt2BnHK6rBv3rJ+8pZJPNJVzGOfycAgrIIBM/GeStMJHPn4WrCTxpmbh9GTg0Dzv/8C9mjJ/Aptdfx4hEiAUC6D5BfGgGVq4bHCrURUC3kDQFFBldOLHcTgrKN5EebqHsqJNYdu+ziF1BJCQcael4MrMwolFmTb6Ua3/5PB9+WEJdXYjNmxu4+eZ5/PKXs5gTc5KmRzqE6yAjHuYzdx56wqB8yUq2f5JP1AHNpkLUlGn1Z7AzVsijf8mzJZiQMJaUseVn99Jy8R0UX3IuYx66hdxoS6fzcsntFzLcX0dAzqRZsgtl4i433osGMf/OaWw//iwSmgs53Y2IC4Y2rqP50BHEszNxtcZwheJoLVECCcFDO+ogUAdZ+SDZ3mV6Xg7bAtlU7KwiHm4kHm6ifGcNa9YFGDAgB6emIFkCsYeHbyKhdWASakiEyIo9QFxKJ0whYSmfVimfHGM+W8NffLubK4mJHPu1fKSiA4NMAgMTkwgR3mYWbzOrvf3mh4Sxhb18EVmxwyfG+h9+PD9h/Fd5hJIk9QBOBx4AbpbsCoTjgDaJghnAvcD0/5MB/ohhxGJUr1qFZZoUjBqF8xtUBe4PQsCVVVCWALcEcaDUsCWWTGBZBCZ8PVFHCjIyIxjJUIbRRCMyCh8/fDWNCQvZ2VZhKSGpGvFYAFMIXH4/piyI/noo1km9iJsCXVNgdTXy06uQhITQDazTB8IFw5G2NTD6uXfZcv6VxM+5iMyVnyJkGU96O+Gp5nYTbQxQWLmQXb7TUoTSiYTJ/Pk7SRuSTvrwzjp1shAYioJpWfjy8vjw371o3Opl7EU6r/S7nh1WHxpa0lB7tHL81iYGOlr46JZb0NzuFCXcrs8+4+OpUzljevvt323UCJ6ceTVv3v0kJZUJynsWs/7qyVQpAsUyKT3kSDaOPoWzXrgNLdNNWmMNwYG90SI6oGKaNkNMvlNmRdRDRFbp2OG3IjCUhCWBMJFlWy3CNE3isRhofjK8GiPDIdb4fHhjiZQHrDs1LnK0e4llkXW4JIM47TlDCRkThUDkE0g76uBvhD3gwEEmWcSJESFKgrZxtFd7djaSIuUhWgg2s4ne9GEU+6aK+96g9AVzD+J5y7JnTuqQH3YsP3H8VxlC4HHgNqDtLZwNBIQQbdQPu6Grfvm7xs6FC1kwdSpmIoEZj6NoGsc//DD9O1CgBSsr+ervf6d+40ZyBg9m+EUXkT9ixAG3uyEO62MCVQgU7IIYRUCLBR4ZAgdW1tkvFBRyyQNAlvfWi7KSVG6yZifSzEmD4Iz+UBlEEgLV48E6shg5IsF7JSRaI0jvb0MaWkjjYgEFPRj26XuUXDoFR21VR4rU9n1Y4JVDKSMIoGkKoZBOxsrdtIzpQXasXUy3xelmZKQRt3MIA886i389vYSPlo1kxvmXEY9lY2kKpAPjC7g4bDLjvX8jLAuHzw6bSbKMNz+f6pUrCZSVkdGrV2rbvSdO5NZPJtDU1MzRq5vxNzdAoBkkCVc4SENhH7aNOo4hwblUxQowIyYOj4tEwsSyBAMH5dg9nA4nsiSTYu4Gltdko6gWTo+bWDRs84u63CA7Wba8EhjCyycWcvyCemo8rpQhHKnH+MtJ7Y+qJrvaqFU7QcZCkb/BbGg/UFHxk4+BQTVVyMhEiNjXah9N8m15ZiVpJDew/oc3hL47oHkxWBHAjT09jIJ2DKjFP+xYfuL4rzGEkiSdAdQJIVZKknTMt1j/auBqgKKiou94dD9OGPE4i+6+mxXTp6f63GTVviVeO+ccRkyahK+gICWka+k6kqJQ/vnnbP/wQ07+61/pc/z+ddFWllWh11oYXj+xhI6sqjh8PgQKMQuGOvdYYdNS+OfNsGszOD1w3CS44kGbtHs/GDF5Mtvffx/LNFNah2bCTka2MbvoZ/VF1EfASr7gLYEjrsKEPkSf34TQJERMRr9/FUaPIQinm0R9E9qSuTQOHsJAQ8cyEshJVQkhBLIEpXL/pJJ8kt3GtHA4FI7pX8iyeJg6lxcpaVR8iRh/PsLulczu359xV17EO+sM4gVZCE1BMUxkVcXUTSp9Tt5IK2Kk1llxXZIkJEUh2tzcyRC2/bbeK4gXOlG7eZASLkRFELkmgqM1QkXOKMZKCzn/2euINiRYlOYh3ZWgZ480/H43NSaclOXB1WsIbF8JvgxAUCy2gzQAV3oGrvTM1P7CrRH8Pvu4e2W72X5eT97a1ExJQOfwHl6OKc7rVBA1yDOCtcEsnKKZuGRvRxI6kiRR5Dt1v9f3YODBQz751FOPGzdSUth3fxAdLLKMjIYDnfh/NIZvBed48L8IrbeCWQmSA5zng/8fP/xYfuL4rzGEwHjgTEmSTgNc2PPjvwEZkiSpSa+wB1C5r5WFEP8E/gkwZsyYb5B9+uniy6eeYsvbb2NZFpZhpHhA7eIRwernniO9qIhQVRWWaaI6HLaRkSRayl5zVb4AACAASURBVMpY8vDD9DrmmJQB6oh4MMi2xx7Bfepkeix6n8K1y0i43Gw/7iwqx5/IBK9M347FMmUb4K7TwEjY5ft6FN59CmpL4c7OFXQ75s9n4R130FRSgtPvJ2fIEBo2bUpKMQmEELiys9FDIVSfD8spQdAOlykOzT7GoIRc6Ke2TENvldAkQURyUOHIpDirFgOZmvxeHO8vo6jAy+7qALLLmxTENfD3H0iksReWHkNxujFNC103yc/3cfEvRvDAiFxeX7GJtU0RcvxeynsPZ5LuoLAcbs6GX/zmTP70zCaavS4U3URSVYRlF35qisTqkeMZGn+k03G3Gfisvn33Ot/NopmPxDwizuNRTS+qbKD1VRCajNGikB2Nct7jz1F89NGcLuCOWvgkAvVLS1n3zy+QG1rZfUQ3dtx6H30Hz4Vl74AkM+m60dy9USYSjuN2uxBAXLdwO00mX9xOgiDLEucN249eFuBUVLIzH6e5+UY8oo628pWE5zYK3YMOfKMeBI7leN5hNmHCONCIJr3BthBoxxyi3WojUlqHLpz0Yt/E7t873GfZH8sOVe+TpbwL3zsksa+4z/8xkh7hrcmq0TeAWUKIVyVJegZYJ4R4+kDrjxkzRqxYseKHGOr/LMxEghcmTsSIx6nfuBHLNG11iA73gyTLyG3GzzRBlm2jlzQ2ad27M2XpUtK67a0+sGXOHD6+74986c3DX1lKwunGkmUU02DXmRfz2oO34etoPx+4AJa/D54ODcuWCXoMnt0MuXYrxc6FC3n9nHOwTBPF4cCIRjETCTJ69yZ7wADqNmzA6fcjyzINW7Zg6jrmY8fDkd2QGmxmGyMWw8p2ItbEKLtxDY54GLeIsMU1krXySCZkrUQ7vDe/OjqX8Q2bEabBsoXrWN/gQLh8DBnZm/GZdTQ3hCgpD/OvykOZFRhNXp6Pm246gmuvHZPyEst1OLLUDgOrkGzPhgfzIFob5A9xNyRza4oi4/U5SEgyE50JJv3uQlrKynCkpWElEhixGGNvuIFRl1221/l+KvopW8RGPm44kaZ4Bk4iSJioiQSuoJ9Xh2Uxwtc+8xAC7n96JX/+/TwkAYosYVkWWVluFiyYzIABOallF39RxyWXfUBDYwIQZPklplwxmDvvGL9PlqCO+4DOrTKGlWB7eDWWiNHLMwKvurfslGVZbN/ehMOh0Lt35l6/7w8JEuyinCqqWMGXKY1D2+vTUkr3CRIoKCn5pkyyOIdzSeObaQV24X8PkiStFEKM2fP7/yaPcH/4PfCqJEn3A6uB5/6Px/OjgKnrmLqOOysLIQTC2IOBX5JAkmzdPsvq8HXye9Mk2tiI5ts300akvp5Y5W56BzbS0HsgCa+d9lVbA4x95gHM31wEPTqUh5etB3UPz1JWAAl2rE4ZwkV33ollWWgeD/FgMCXtFCgtJVxbS1phIZnFdn5F83rZvWwZylMrsQ4pgAIfesJAzvBBU4jQn78gzbJAMwkbaVQoxWhGnHXyKH5/dA5jd86n2XKxrC6HgOcQxt92DsN6aUhvTwNPJrndHWQXZjKscSvXjz2JwkNGUkg1NFZBjp0f+0OdbQTTOkz04xb8sQFK+qfzRKmg2angACRZwkAiTYJzMjXOev55Nr3xBqWLFtkE4RP7sDh7LQs+v4EioycnjJpMZmYBcQvWmfW4FY0JWatY0nQoDXomIJBcCe7p6e5kBAEMw+SJexfiVOVOwrpNTVHuuecTZs48L/XdUePy2LD6IpYsqaC1Nc6IEfn079+5GKgjNpXAtffB6s22g3PMYfDsHyEvG1RZY3Da2P2u+/HHO7j66veorbXzq717Z/DKKz9nxIivJ6/W0OhLP7rTg01sQEfHwkr1CgrAg5f+9KeQbjTRRD75DGAgbroU33/K+K80hEKIT4BPkv/fCez/yenCt4Lm8ZDVr5+tGbifWb0kyyiahpXorA4uhLA1BffDGwqQO3Qo8dZWFNOgsHQLZjK/JutxZIeD8k8+YcSkSe0rFPaFul2dN9Im5tuznc6reccOFM0ObwrTTPWfWYaBZZqE6urI7NMHWVVpraxEkmXEzmaUC+dgndEXMSgba1M90rwKSuv745ViVCsFbBpxKrrqIRY1UKUgRa88x+yYk2mlRxCTXEhKLmpdKyd51/KnoTHk2lUgQEaQ5vRy6ILfw0dt50mCcWfDrTNYHJHZs13SKUPYgi0JeK27xGVVEo2m/TBmKTDaBeeng0vxc+iUKRw6ZQpzFkxjXeEOJBNkU7DFsYvSrfdz7dB7aXbmENRzSfPW4SbOCblLaTW8RE2FAmc9Z7snAxBpbGTdnLnM2dDIW74RNIZ0HF4HsgAteRk1TWHp0j0qGbHJxE86ae+Q7J5oaILjLofWMLhd9uWb/wUccymsn3PAdC+VlUHOP/9NDMPC67Xvl9LSAKec8hI7d96YknL6OrhwMYpDWcoSQoSwkvK7IOHDxziOIpv9G/Iu/PTQFZD+iUKSJI6aOpVgVdXe3iCAEMgOB8KyUgU0EqRCqM6MDNIKC9lfaL372LF4srMRlmUXlyR0ZD1uaw1KEr49w6m/uN32AOPRZB+VAbEwDDocurfnorwFBanCnjYDLizLDuMmw7aJiJ0fSokBSxJSYxR5xgbk2xYi/3sdPuFmZ8Z4VvlOoMHVj7zaauRAArM0wKE1qylQgjxVNhbdUshXWsh0xsnrkU1DTZDW6hqbAktzgOqAQC2EA3Zu0+UFhwMWz4KZ95Mus1d3m5U8ZbkKjPfCp73gkXy4Phv+VgAzukNaB4MRbKlnfW4JWgQcUQlVl3GGJOLp8Pn6N8hSoKZ1JJZQUeQwIEjXAuS7duNJDMOJk+rVq5lxyhlMf3wGL4wdT2O6fY3juknQEqmeTtMUpKfvWcXUjq/Ww6/ug5OmwG8fhB17zF2emmkbQZ/XNnqqCj4P7KqGdxbtd7P2uk99ia4beDxaSnnE53MQDMZ5+eV1B155D4zlcE7hNIooxoUbP37GMIbz+UWXEezCXugyhD9htPULyqqK6nKhOJ2dkvVGJGIbsmRoVEgSDq8XzecjFghQt3EjjxcXM/emm9jx8ce8femlvHnhhax96SUkSeK0p59OEW3LioLq8SAAX2EhvY85pvNgho6HW14Af65tAC0DDj8D7pnTabHxv/udHcpNjsuyLIRporndqG43pmEgt+0zacBVl8s24IaRSlw509M4s281cVMmbLkJNYWJVDTjNOJsi+Xy513HUqWnISSZeuGnMQzCNDgjbxulIovtOX0Jax57e5aZDCUnz52sgqrBvOe5JtPOCxpJQyMEhAUMdkCfNnsT0cnYUctp4SCnp4Frj6dyV+UGu73B6ux9y7qg3CojTYHTvBmsqD+HkN4NVQ5jWAqlLUfxS8fhWIbBgttvp7ImzKqfXwaKRlo6qDlOiBtYCYuIsMOllmVx7bV7pVAAWLAMptwNb3wEn30FT70CI86GFztcog3b916vrSNj3ba9f+uI8vKWfX4vBCml94OFhMQgBjOZy7iF33EDv+UUTiOTg885duGng//K0GgXfjjImobicKC57RyJmUigh8NYuo6vWzeEaZIIhRCAEY2ih8MI00T1eHB4vViGwZdPPcWKZ55BVhSEZbHt3XdZ+cwzTJo3j+Puv5/P//xnTF1HGAYZRUVc8Pbb+6bUOupc+xMOgsMNe7QPAAy/+GKiTU18dv/9RBoaUoZND4XQIxFUp5NoYyNxTUPzeOyWhHjc5vQUIpVTjDU1MaR/iGtGlrGkws/uBoGe7ibbCkJcYl24OzW6h6o4FHjjpMshEpbJM1fezeJDjsdUNdxGlKu+fIEH3r0NaY/+O2QVoiGuz4K1MXg1CHFh56n6ajCrJ1iW4Jp/rOPVZ1eAEFiWoP/Ynnzw6LF0y2j3yjL8BYiA3fwtd2jEs1RSBR43ZYO7KZfXGs4iLqC7CrdlwzAXNGzZRqy5maaQRHPf/jhCISRZJm/KIGr+sQWzWSeRMEmoEhdcMJTrrtvbEAoBj/0bKmog2Gp/J0kQjdvGcWBvGDsChvWHDz/be10kGP418oXjx/dkzpytnVpS2iIO48f3PPDK3xN27IKnXobKOjvXecW54PoWPOJ1dWE+/HA71dUhxozpxoQJxTgcB4gTd+EHxX9l1eh/iq6q0YND3aZNfHD99VQsWYIjLQ1hmilDhxBoHg9mIoEzPR0jFiMRjaYKZxzp6WguF2YiQazZ1m/z5tlN7sKySEQiHH3XXRw9dSp6OMyuxYvx5OZiGQYr//lPYk1NFB19NL78fOKtrRSMGkX3ww47aM7JlspKnh46FD0YTLV7gJ3X9BUWMviccxh7ww28MGEC4fp6+zdJAlnGSiTQfD56HnkksixTW7OT6Om9aemfRbjWxfpXZRKVLhrCKgiLbGcMU9Go/fMN1IzphdeMogkTXVbRVQf3f3gXN3/2N0jLajeE0TAMPAweseOBu3X4IgrFGox124s98VE5U2+ehyfbg6IpCEsQrg8x4tg+LH3ihE7H+/dPbqEpL44WskPUpiqwnBIX+afQp0e7Tp8uIGpButw+lMbt23lr0iTW74jz8e1/pKVbD5zhEJJloOOiqUKhONPN7J/1YED/fbdABIIw/mIoKU/2xbcJaWAXFJ94JMx9FhqbYeiZyRyh2+bujsahqAA2vnvgHGEsZjBixHR27w7idCpYls3WM2pUAUuWXHHACtXvA2/Ph8vusKP0bSjuBktngv8bEC+tWVPNpZfOobq6FVmWyMx0cdhh3Zk+/XS83v1Lk3Xhu8f+qka7QqM/QQghePeaa3juiCPYvWwZlmEQa2oi3tpq5wZVFVnTkDUNyzSJBQIY8c4Nx0bUVnjfs5AGbGMkKQqb33wTAIfXS7+TT6Zs0SJePOEENs6cyfYPP+SjW2/lnSlT+Orpp/ng179m3s03p3rlvg4bXnkFkUjgyc219QtlGSkZCjViMXYtXkz1qlW4srNtI5kMp1qGgSc3FyyLcF0dVdXbaP3rEYhJA9D6peE/VuOoZxRuPOVLHhvwIVlqlLghkxAqVSOKUYWEQ7KNqhMLzTJ56ugbQdEwIyEigVaigQAhQ6H8xD+kxtvDARf44XBPuxH5x8sbUF0qSlJCSZIlPLle1n1ayq6mWKfjveTQqeQ1pJHwQsIDiqRwsnxaJyMI4JDAr3Suf8rq14/07t0pzBAMmD0Lw+Ui4XIiIYj4csg+Yxi3XzoEc/0nvHv11Xxwww2ULlrUKf/rdYMstXOUpyBAkWHzTvvP7Ez4+HkYNRhiMduIHDcWPplxYCMI4HKpfPHFlVx88XB8PidZWS5+9avDWLjw0h/cCFoWXH+/fev4vPbH64ayKrj3qYPfjhCCc855jQ0bamlqilJfH6GkpIn583fw1lubv34DXfhB0BUa/Qli1XPPse6ll1CcThRAmKZt0CwLLS0Nd3Y2seZm++WT7Bm0GVnayz6sRMIuRkn+Le35lpOkTt5dqLaWxQ8+iCTLaG43sUAASZYx43GMaBR/z56Uf/45JXPnMvBnP0utFygvZ9Vzz1GxZAlp3box5ppr6Dl+PC3l5SlP0IzHbUMoSXZhihBYhsHHU6faTfUeD0pbmFWSMONxcocO5aipU3mv/EXU7ploARNV0Yi0JnC5dYJT+tHvni+ZefwXbJUHUqtlcJ1Twud0gKRBIg5GAlWWacrszoqfP0dsxp/prjZRKhXxz5ZTqbi3gueLGujbO50V06ez+a230NxuDp0yhSHnnUewOZoygm1okyRqCOoUZbkIhXSamqLk52dyzYQHCQRrCcUCFOT2RZX2/fhu3lzPK6+sZ3NjnAGju3PtGX058ZFHeO+6X3HIpi+RnnqI1b+8gqbcYtIzsri7l4PMu37DZ8uXo7g91DZLfPb6F7QMuogJt93Cz0+0638u/hnc8bh9G8jJSLBl2b8V5rbvf1h/WPJy++3yTXrEc3I8PPvsmQe/wveEr9ZDMASeDmFQSQZNhfc+hWm3H9x2Zs3axK5dLciylLy2AsuC+voIr7++iUsu+X6UL3YtXszihx6iubSUgpEjmXjPPeQMHPi97OvHgC5D+COFZVmsf/ll1r/yCpKiMPKSSxh6wQVIksTK6dOxTBOjpaWTcQM7Z5g7ZAiRxkYat261v0x6VCiKHQdLIhGNtock1fZbyUwkQAiGX3RR6ruyTz/FiEZx+Hx24YplIcsylhCEamrIGTgQze2m5MMPU4awbsMG3rjQbipvG8fWOXM4/MYbKTjkELTXX7c91aShFthhWSuRoKW8nEQ0mjK2KT9TtrNsseZmyj75hMR52ShhK/mTXaWYZoRJFKYz+ORheCMmfUggqOf2RCsRh4pLVuwKUYcb3YJ8VXD9ukIU/z1km8k4mhPM+jD/fGY5g5f+ibr161Nj3PX55+xcuJBxR1/OuzNW4+gQHouHdLxZHgYXeHjssS94881NCAEOh8J1143hwguHkZG+/566d9/dytT7F1N/0eFEj8znQ8Pi6S9D/GV4Ny57710qly+npryWWIaHbqPz6VOYRuWSxby3chXLz53CwgFHE9Nc5O4o4fCX/8WTT+1m2doePH47/O4KeO0DWLstecplcGjgcsCNk/Yey/8ySYrL2YlbIgVh2ZWwB4vZs7ckb892LRJZhkTCorY29J2MdU9seO013p0yxa6mVlUCpaWUzJ3LJR99RLcx+y6E+qnjf/hW7cL+IITgtbPP5r1rr6X8s88oW7SIOVdcwVtJwxSur8fS9X1Sf8RbW6lZs4aGTZtSnhZJjsvUsqkEkbDZZjQNSZaJh0LEQyGEZdH7+OM5dMqU1HbbinH2hZTnKESKKBvg84ceIrhrF6rLhcPrtQmoJYlVzz5LVr9+5A4ebM+yJSlVFaq6bBowSZbtHOceIV2S7RxOv5+dH32EuTuAtceLzZBlFCFw6B2IBITg1tUvYpom0UgYI9RCKB4jKiwsw2TjMYPYdN1xlI7rn/KS/X4XtfNmUbt+PZrXi8Prxen1orrdrPt//4/fTXCTXphGa20r0ZYYofoQZjzBn+44iuefW81LL65FNFcjaneiN9bw6KNLWLBg5z7PYWtrnJtumssll7zN1pOG09gjG6kpjE83sCI691aarLWcFE+YwOGXnM/En42lf7c0FAmqV63indMuZWnRMEQkhq+5iebiYubfdgd5mTv4Yg2s32bPg5a/BpPPsnNk6V7bE7zzWrjw9P1e3u8MQgjmzi3h0UeXMG9eCZa1Z2PKd4cRA+1ji3SIUAsLDBMmnfFNtiQhy1JyLmnfGW3/Dh+e990NOAnLsvj4tttAknD4fGguF860NMx4nI9uvfU739+PBV0e4Y8QO+bNo3TBAhzJ3BnYD8jWd96hYulSPDk5BCsq7IXbjFpbGNQwCFVV7bVNYe6t1+bKzETRNIx4HF9+PsXHHkvjli1YlkXe8OEkkh4gQO/jjsOVmUmspcUel6JgJdsZ0nv2tD3UWIxBZ58N2Mw31StWtPcHJqE6nSTCYapXreLcV15hxfTpbJo1i+Du3bgzM9FbW22j6FVt1fI9q+6TxynLMhl9+9Ly+jaMEbn2PnQDPRIhke/m8BW7UcNxcLptgx+o4+ba50lLhHns0MtocPnRDJ2MugrSaquhOU6oNMqOVicJE3pbJpFIgsLQxuRu2ycbbceTWPIRX732Kx54fStffFlJj+7p3HLhYCYOzGT8be8S3rSaqNlOHp1wpPGvZ9I54YS9G9vvvfcT3ntvGy2KDEO7QV2QVgGRcAKvz0E4GOf1YBqj95iPNDQ08dKiLynetpkiSUZ3+1j182uQ6EMk3ceuUcVIX8K2Mts4OBzwwp+hpRVqG21jkfafi0d8LRobIxx77Ax27GhO1Ub165fJJ59cRmbmd88KI0kw8zH42XUQipDksYWjRsPUqw68bkNDhNmzt7BqVRW1tSFEsiLYrtWyjaKmydx559Hf+bgjDQ2EGxpweDtfFNXlonZ9l8bh/tBlCH+E2DJ7dsorakNb7mnL22/Tc/x46jZssPOCHeM/HaovvxYdvURAj0TY+s47du+haVKzciUrnn6an8+cSd8TT8Th9XLO//t/vHnhhejhcGpsTr8fxekkXF/P0F/8gl7J/sK23sY9x9PW4C+rKundu3Pc/fdz3P33EygrY/3MmSz/13SMPAUpy0NiSw0oEpgdtpEMo+rhMN68PLIrNOTPDKoPtzDlOKQriHdLaHxuKVXdLXIHDsDhckF+L6TKbVxdu5yr535JiSePU4fcgLu2CnDS8GIJ8ZoIlqpSbgjqW+N0755Gv6E9aPxswz5PodPvp2eWi2euHQnXtueKWlvj1GzdgcvQkbW2R1QgxcNs+nw17fKcNqqrW3nttY3U1ITsypVHPoKJ/aF/HqZp0RqMk5Huon4fvAl/v2kqUskWIpk5SEgo0QhHvPgYn153H4rXS1OPIvJWQM4e7Xf+tG9WObknGgyYEYBFEZt+7sJ0+FmaXZCzL0yZ8g4lJU34fDISMYSVYPu2GFde9gRvzb7Zztt+xzhsGJTNh5ffh6o6OPZwOOrQA69TUdHC5ZfPoaysmcrKVnTdnkC2ZRcAnE6FqVOPonfv/ZOUf1s4fD5kWbbDoh2eT8swcKZ3canuD12G8EcIxwFEdZ3p6Qw880xKFywgUF6OGYshKQqSomDGYvtdD+yCmDbPUJJsXn9T11EcDqJNTTZxtNOJ6rR74PRwmPevu44rly7Fm5tL7+OO48ayMta99BK1q1eT2bcvad264UhPJ3/YsE7SQpIsM+qKK1j4hz9gJhIommarY+g63rw8+pzQub3A36sY3+0nE+++EuPxJYAG6U67J82uUWiv+xfCDqEmPcPTx17DB7fdREPrbkRDFLXVZBc+Xq4NURyKcco//s6WWbOo+Hgh3swWhg/rzu6cLIKLd9O6sZpoXZxEbRwty4UhySjRBFaLoEePdE67+1ZePHEeRiKBmgz7JqJRVJeLYb/85b6vHzq+WB1RxYNCm/WSiMtuilr3rjT829+W09Bgs+mQMCGiw+y1cNFh0D0DIQRha28R5NqdZQTWrYHcAiQLMAwMlxtHJET3DctYfd7xZKyUycuGcaP22u1+UfnVV6x+/nkCZWXkDx/OoVddRXb/dnagVhOuqIKqpEhzqwn31cMOHW7O2Xt7QggWLSrD41aQRCtg9xm6XSYLPwkiWh9BSr/j4Af4DeBywZU/P/jlp09fwbp1tTQ2Rjql31VVJifHjWFYTJt2CpMmHVjL89vC4fFQfMwxdkTI57MLyEwTM5Fg5OTJ38s+fwzoyhH+CDH6qquQFQVD7xBWi8eRVZVRl11GjyOPpPdxx5HRqxfOjIykDl4H/bb9lKp3DI8KyyLa3IywLNK6dUvNQJUOOT5ZVYnU17P+lVcQQrB7+XLm3ngjn/3xj5TMm8e6l19m6bRpVC5bRnpHAu4kDrvuOkZOnpxq8k9EIrgzM5lw553kDh7cadmtbGUJnyMuHIJy0gCkujAEIrZHKMvgcbR7vJKEEILm0lLc2dnEg0HiVQ2wtQktZKXovYTmYldFgFev+S1rP1xIOBKnqirAW+9v5b4XVIKvb6VlW4jQhgB6fRRiJpKm0EOJMW5cT+rrI2SNGM3Ee+5J9Wjq4TCqy8XZr8zkr2EPA7dDj23wywoojbedfomjteUkUAlbLnSh0mq60TAY69zY6bhDIZ1Fi8pQ1Q6NfXWtoMqwrhJ8Tsj1YVUGOCYR6bxucwBLVlBkCZcsQbJlxnS6cQZbMGNOjjLhH/faIdGDQenChbx/3XXUrVtHvLWV9TNn8ty4cbx//a8pW78TXYcPQrYRLFTBLUO6AgUqvBa0PcV9wTAsIG4foGTnhWVZwjRliL8DVsPBDbADli/fzd13L+Lxx5cRCBx4EniwmD9/B01N0U4ZB0myx5+f76OoKANN+35fu+e++CLdxowhEQ6jRyIY8TgDzzyTY/70p+91v//L6PIIf4TIHjCAEx55hAXJZnawq0FPeeIJu4fwqquoXb8eYVl48/MxIhFCNTV2G0UHGrJ9okOhjBCCeChE47ZtOP3+lEcphEAPhWzjKkksmzaN7R98QKS+nmBFBUY8jiRJGLEY+cOHs/2DD+g2diyDzmwvm4+1tNBSXs7Eu+9m/O9/z47583H6fBRPmLBP2afVrERFQ3ZoKA+djLXzMBL/+hLe3gxxE+rD9stTVbFMk9q1a1HdbmRN4/3rrqO1qgrLMGyS8WRVq7AsrKTKRv7w4SDFcTTXsDQ6kJ1BJz0SpVT3GQLOOCJhEa2JkCnpFCQJoxVFIibJ1F/zOxJnXUnOx+8yOtvL6HPO5MIaBx802A+gDLwTgs8j8FUfKPT5GD0iC+f691kvj6LZTKe/VsFwazVjTjmu03E3N9sv3T59sti4sT75ZQSiybB3ST2sqSBtVTmcdjnkelLr9hjQD7ciEY7reJ0OFAtiSGiJGJ4jjmLJOBh+RufI8oEghGDZ44/jTEtDALVr1yIsi3LlEP4+bzKNi7Jw+AxyR6ukXQ0dKT8VyQ6L7tAhZ4+3kiRJDB+ex7p1Ffjah080KjNiWARJUmxhW3kf7uQ+YFkW55//BnPn7khuH/74x095+eVzOfXU/l+z9oERDieSNWQyktS5mKemJkRRkZ+srO9X6cKVkcHln31G/ebNNG7bRuGhh+Lv+X/DzPO/gi5D+CPFYddey7ALLmDz228jSRKDzzsPzeXitXPOIdLQgC8/Hys3l12ffooRiyE7HAdXhdfBSCoOh60Ekcw1mokEsqbZGoG6jgA0lwtPTg7ln36KJzsbU9dTFaSJcJhIYyMOn4+ts2cz6MwzEUKw8tlnWfP886mwTu/jjmPiPfccsPI0TBg3LiRkW3qnTxbaDePQF5WieT1kDRpGZO0OYsEgIhpFCIERiVC/YYPtESePwYjFUj2QwjRRNA2n32/vpHt/eCGakQAAIABJREFUcHnYvKEYF3EyNTfazk3UKVkEIwpSzKRQi+AtHkF9fZijju/LFQ0qVQYoahbmKZfyhgy/i8O8MHiwU3oADgHNFjzSCNMK4KwXXiB4wgnktyxAyDZHWXpRD0752986HXd+vg+XSyUtzUF9fZiaGnviQywBOxqQH/gQEKT39NOjR+cckTPNx8Tf/Ib5f/kLpqLhcDhwhUOYvfvy5KWnMceEM7ZBowmZCtySBb89gK3RQyFaq6vx5efb7SJC0OQewqvpM4hIWaiYJKJRSpeloZRC7hOgutpvK1NA/n7eSNOnn8GJJ/yTUDjZ0S/A67GY/mgpYILSff8D2wP/+tdq5s4twePRUrnzaDTB5ZfPYffum1HVb++xHXlkD7ZsaehUWN0G07TIyHBx1FFF33r73wS5gwfvFTnpwr7RZQh/xHBnZXHolVem/i7/7DPC9fX48u0+tHBNDUYshpCkVF7P1PW9eguBvQtpFAXN7UbRNLsC1DRxpqcTrq1NLack2x7qt2zBTCQI1dbam1JVFFVFkmWiTU04vF5iLS0s/etf2fnxx9SuX09Wv344fT6EZbFz/nyc6ekc/Yc/7D2uJLrRjW1mDVWRQ6i3YuQ568gtSKA8eCLaXUsxquqJtbSkPD1ZVVM6jGZSzinFpCJs0R7V6QRJIt7S0n4OcnrgyvRjNmlkDRqEZ/s63C272G1mEhReVM1BQ0OUwYPz8E8ZR2UyBNiGBgMebLAjmEqH960kgSxgaTJ6mTdkCDeUlLDm3/+meedOCkaMYPjFF6eIxNvgcCj8+tdjefDBzyksTKOlJU4sZrR1tqCqEoqicPnlI/f5gj918oV069eHeTPfINjQSM9jjuXcX5zNLMPLrXWgAG7JzuP9od4mD781d6/NAHaLjMPrxdR1Yi0tKA4Hy50XE5aycBBFkVWEaZCeLmiuldixBAYcZ5OS15twpBt67ScEO2pUAevWTOJvj/6JdRvdjBhq8NtrKijIrQXX+QftDQLMmLEGSZJSRhDA7dYIhXTmzt3OGWd8+8bzRx89kZkzNxCPG8iyhGW1PzOHHtqN6dNPx+3+7gt7uvCfocsQ/oQQaWxMKUkAhOvq7HBhMmfmSk/H0HUS4TCSotgvXSHa6dU6kFa3CfS2JeLRdYxoFMXtxkzKIJm6TjQQsA2rZdm1KopCPBjEkxQEbqsYDdXWEty1i0B5OUY0St26dRSMGoXm8eDNy2Pru+9y5M0325Wk+0BabDzTa6LELJWY6UAXEt1dNTwyPsgxc6fx2R/uZtVzz6U8v700FpPnRUoesysjg4xevZAUhcYtW9DDYZtk3DQZ5SqhwnckoqYMr0vG481EDmsMyQ1wcfF28k7vy8irz+asz3cy9KM5uBtrCY0ZT/Dok8hyutiud6jd6TgGoEeHJ9Lh8TD2+uu/9rqed94Q8vK8vPDCahoaIui6QUaGm1jMwDQtiov93HTTuP2uP3LcWEaO6yz5+dA2++XQpoThlAALpjXt3xDKqsqoSy9l+ZNP2sVNlkWNMhAZC0VRAGFrXMoSThnEVqidaIeGT/fB777GlhX26MNDj90F4ScgsQIkP7hvAPclX3uOOsLON+4buv6f9SZmZXl49dXzuOyy2USjBpJkF2T9+teH8eCDJ/zgVHFdODh0GcKfELIHDEipzkuynGphEEKkettUhwNT1/H36IFlGAQrK1Mk3KnMf9KTaiPbTm4ESVXxZGWhYxtByzSx9mhob9MHjIdCdpVpMmfnyszEm5NDoLwcxeXCMgyay8rIGzIk1XNoxGL7NISWgIfrMsmwfJQZBiFLASGxM9yf35fLvFRgsPvLLwHbgIv9tYkkK2EVt5vsAQPwZGdjxGJk9uuHrKqE6+oQwGkXnERv3zhenPYhkuZBWBbDc0M8elIlwYpadr05jY/LShj0/ny7ihRBxrzZxIufYduTr+J3p4MKlQZ4JNsYxoVtEG7tYAyEEOyYP5+1M2bQWllJ8cSJHHHjjbiz9i67nzChmAkTigmHdZ588kvmzNmKYZiMH1/ELbcceUCNwX2hwbQ9wY5wSHb4to1ibV8YeemlCMvii8ceo6W8nG6edZQ5xyEwbQo/r9teX4LfDIPJxfZ23QcbjVT7gP/xAy6SCLfSsPk9ZFkhe8gZqC5Pp98vuGAod921qJPKRTxu4HQqnHbaf5YjBDjzzIFUVNzEW29tJhJJcPbZg8jP9/3H2+3C94cu9YmfEIQQzL/tNkoXLMCZlka4oYHmHTsQlmUTV0sSpmHgzszkZ88+y6wLLyQWCOy1HUnTEB08qra2CkmWcfr9yIpCLBi02Ws6rdhugNKLi8kfPhx/jx6UfPQR2QMGoKgqjSUlhKqrU0UrRePHEwsE8BYUcP7rr+9zRr01DpMrwQBKdHAlF9GFbWgmlKzi2HuvIbh7N7HmZjv8u+d9L8uoTidGPI7mdlN89NFIikJrdTUjJk3i8BtuoLW6Gmd6Os70dBpWLGXnHVPYsNsg0tiAU7GIROKEQ3GQZUxTYDpdhAp74mmosYsdjQTB3gPInPE2k4b34heVsCVue4J+GR7Oh0sy2of01TPPsPjBB4nU12MZhu21Z2YyZelSsvoeWC3esuwm7m+b7yraBi1mZ23EuGUbrOqDiBwaus5Xf/87i19dyP2tzxPDh9sJstOFnpDolmcTcXf7jslVqpa/gNV4Mw6PDhLEwy5cPaaTO+L81DK6bnDssTNYs6Y2aQxBUWSefPJULr30G/SJfA2EECxfXsmiRaU4HCqnnNKXoUO/ezaZLhw89qc+0WUIf2IwEwm2vvMOW2fPxkgkiDU10VpdjR4MpozgWf/+NwAvTJxohz8NI+VJiqSEUSLUgSdRUZAlCcuy0FwuHD4fQoh2vcAkO4yUXEYYBrlDh6K63UiyTP2mTWgeD4WjRiGEoGbNGoxoFFnTyEp6Y6c99RSFh+67m3lbHC6ptEvyIxaoHQxhpgz91yzhzMduwZ/hp37z5nZWnQ6QZBnN67Ub7QsKSMvPR1IUsvv359g//YndS5cSrq8nZ+BA1s+cScPmzTRvXo+pJ0j3e7Asi+bmKLIEDp8XPRrHsiwsSSLht+WZJCOBomoMGtqbX946Gan/aMp6HUYrMkMcnXOGsUCA58aNo3nnzk5tK5Zpkt6zJzfu2LFXvvBgsXNnM2++uYny8gCjRhVwwQVD8fs7e9pPNsDUejtk5JRBt0AH7sqB2/cTGt0XEpEIX33ZxM3TC9hcqqIpMKgPPHkHHDLkWw1/vwjX7SS4bpTNL6vbHrDiiCEsiayxm3GmF6aWtSyLN97YxAcfbCc7280114xh4MCDzzN+HYQQ3Hffp8yZs4VIJIFlCbxeB7feOu576yHswtejyxB2YZ+wTJPK5cupWrkST24ufU84AU9ODp898ACLH3ywE1VTLBjcZ9N9dn/wZLlo2R1DCBXL8GPEDVvb0DDshv2OQqtCkDN4MJm9ewO2wkRTSQkZxcVk9etHrKWFhi1bSO/Zk8HnnMOISZM6Nds3l5Wx6K67qFiyBHd2Nof/5kbumjiJL6O28dOSjmccWwlebm3hkqtPIS0jnWhzM/WbN2MaBlY8jqwoaMkCD81jh9BcWVkpY93vlFNY//LLmLqOJEkEdu0CIcjs04eGzZtQjDimYWAYAisZPlZdbnQ9QdzlRtXjuH1eUFSIR/FKOh6nwrkn9iI70wPDjoIrH7EV7TugetUqXjz5ZFsFJKmsAaQqey98+20GnPGNSC8Bu3fu17/+gJqaEAHdIjGwgLRMNy/dfgQnH9G58vLxBvhLEzSbtsd6QxZMzdlvm+nXIhiCWBxys77dNuo3b6Z0wQKEZdHrmGPIGz68U4Sg/OPbcWtPkIh2DkNqnhCGdifdjjxIyYjvAKtXV3PhhW9SXx/pVDCTk+Nh6dIryc39AXjpurAX9mcIu3KEP0GYus6G115j05tvYuo6/U4+mVGXX46rrU0A7P/vg94MaNfgAXqNF5wwFYQMG2bL7FpuEg83YzZ4UB0OEslGe1nTkLANoTBN0ovaS8j9PXuih8O0VldjWRbV0QRbz7uKhsGHUNBYzU1V9RyRNITNZWU8d/jhxEMhFE0jXFfH+9dew8XXb6TimgfZpttl+JJkN2kLAQNy/Ey87VaWPPwwkcZGO1fldOLMtV2beDCYYq3JGTSIYGUloZoaAqWlbH//fVSXi8LRo3F4vTRu346wLCL19YCEx+vCaUYwDYOwLtEaE4SRqB06GkuS8TXUsLPfUPqt/BQtEsKb4wVVxfDlQGYarPsUVn4Eh3dmrfbk5qYmHR1f9pL9BeWLF3cyhAED/twA74fAK8OUDPvTMZcnhOCBBz6nujpEc89s9N8cB5pCoyU4q0rm2V1RLilqb1H5bQ7cmG3nL53Svo1XnBib2YQABjMEF/uXb0/32Z9vgzUvvMDSadOIBwJYhsFXf/87o6+7jiN/+9v24zNbkJx7T+wlQBh7h/i/TyxaVEZlZStOp4rTaeffLUtQXR3iww9LmDz5+5Ff6sK3Q5ch/Ali4Z13svPjj3FlZCDJMutefJFdS5ZwxE03UbViBQ6vl6KjjkJ1uVJ0YG09fZCkV5NlwODoG2DHYtj5qU4iqlI8Pp+G7Y3Iqgdft2E0bN1KuLYWKx5H83rx5OYiq2qnaklJlskoKiKzd2/KTZm3f/MQcY8PR0JnV79hXFIf5/WmMIdkefn0vvuIh0I4k2TeOJ2YiQSl/3ia+VOnMjXm55OwnRt0y3brwgN50P3nPyd3yBBWP/88619+GX9REZ7sbCRZJhGLUbN6Nf6iIqKNjYRra5EVBSHLdnuFrlO7di2Fo0fb41UU9EgYLzG8wkRINtlJhlMgCWgwTXJ3bCbu9uCIRjArSgjkFJK3q5WGQIzcHB85ObaSBk4PrJy3lyH09+xJ9sCBVK9ahWjjjEySGGhuNxnFxallQxYcXgoVhv1AW8Bva2FJFGZ0cPLq6yOUl7fQaoF+4/FIhokU1pGEwFAV7q42ObYQenRwTiWpPee6J9awig94HxP7vviA9zmN0zmEryHk/IYIVlay+OGHCVVXp77TQyE+f+AB+p54InlDhwKQ3utUjNoZCCybcB0QtkIl/j7f3Hv+T1BXF0rVl0UiCfu6aQpCCDZtqvtBx9KFr0eXIfyJoXH7dkoXLSKtsDBVNSrJMiVz57Lt3XeRFQVnRgbunByKJ06k7JNP7ApPSUJRVYQkoTidmLpORk+Vsi8MNswGV4aGO9NHyfwmglUxuo9JT+X9os3NBCsqGDF5MiMnTWLls89S/umn+AoK7CpOyyLW0kKPI47gleLD0N1eMlrtGbwHCGpOHtgR4M0sL7uWLNkrN6ZoGrquE1ixnH+ddBK7ErApbucHV0XhrApoteCEjMHc9cAjuPx+tsyeTbSpyVYGMAw82dm4MjJo3LbNzue1hXKT+zBiMfRwGM3tRg+HyfeYaMKgrtXCkfS6dBNy0+x+wLpgDG8sgpSXw5DYbhQEaoZMIGSRleVFSSYEI5E4H720gJI/FoAQ9DnpJE6ZNg1PTg7nz5rF9KFDbd1HbLYS1eXCX1TEgNPbDef0JthtgK+D12YKmBWEP+TAwGTBqNutYhgW5tBC0GSkUHtFr2KaROMWi8KdC3b2hwAB3uc9ABzYzX8GBh/wHr3oTSaZB1q9HUJA4iuILwUlHZwngNKZBWX38uW0VlaiaFrq2rexF6169llOedyuIs3sfzoVpeNwe5ZgGvb5VTSBbpxEbvf/UOnBsiD2CsTeBrzgvQqc+9/myJEFGIZBU1MiWWwtEYsZSBJdFaT/hegyhD8xNO/c2bl1wrKoXrmSRDhsK7k7nYTr6ghVVVG/cSOZxcW4JYm+p55Kv5NOYs7ll2PG47j8foRksvWjZlwZEr58P5IEqlvBsgSttTpZafYLwJOVRSwQYMPMmSyfNg0hBKrbbQv1pqVhmSb9Tz8df1ERO7JH4ot0Fiz1hINsUXuhC/Dm5u4lE9WWd2zjKy3S7M9VlfBK0G5LkIB/B+DDVokVv7+D4okTKZk7F4D+p53G1nfeYef8+Z1kAkTblN7+AyyLzL59qVqxAl8iyuhiizVVUJXst++VC8N6aCwMZRPMSqNw9w5yrAAJVSOhOHEocXJ9Bs31/5+98w6To7i+9lvdPXlmc17lnBEgIXKWEFmATU7C5GRjom1s+OGAMR852YDIJhgQYKJFzkISSCignLW72pwmT3fX90f1hA3CwgkBe/TsM6ue7p6antm6fW/dc047lmUjpMUjzy6lNWJj+PxIYPnzz1Mzbx4XLFlC4cCBnLtgAS+cfjptGzZg+HzkVVVx0B//SH5OafkfYfX+ckuXugAkvBnOBsJQyMMhhwzh3sWtyExOLjPGvz6fQXI7Wwa+YAEWNh6yDHgDgyRJPmcBBzP1n59EWtB+McSfBxkBoYMogby/gC97fLylJSOCkEa6KtG4PCtCLjSN/tPeoHnpA5j1TyE0DXfJ6VSPOnP73tS2YNvQehAk52W3JWZD4ArIu77H7uFwgmeeWUYymf4eQfqWSgh48MEvGDeujKlTv77ztw//O/QFwh8IkuEwa/7xD9bOmUO0oQFvfj6Gx0OspUXZIjnE+mRHB9J0lI+lJFxfT/GIEdTNm8e+v/wlh917L29dfTXRpiaSEYtQhUFetS8zrRoegdAg0ZljAWNZtKxahdD1jBVMMhIhXF/P0TfdRPn48RQMHEjTypX4528lFcpHtx37GtvG0l0UBP24gN0vu4yXzjgj4+YgpSQZiVA6ejRlY7JtiBuS8FSHKusZOQFiqwW3tQp+u88+DNxnHywJ65JQXFrF+g8/pDO/mJbCUnQzRaC1iSA2tDaDpmGbJqlIBH9JCVqJjq5vZnSlzshyy+FRahg6NLlCiITEi0lCukEzsCXEbQ8uyyRoxFj/xUqiKYu2iIU7ryCTgeouF51btrD4scfY5eyzKRk1irM/+4yWNWsw43FKRo5E76aAXWFkM9fuqOj2F37DDQfy8Y+eZYFlY2qgmcobzxdwU1jgZS9/7+fp8X0iSW+vKpGkSPY8oDfE/w6xR1HEFxukDnIrdJwD3uUg1Hpl9W6K7G+ZJroTDNNNQ7muFgBC0ymZcD5wfnajbUHsWTA/A20A+H8C2jewJIo/6gTBQNYnyk5B9FaVGebIu5mmzcknP8+cOet6dTVzu3XC4SS/+tU7lJcHmTChfPvH0Yf/GvoC4Q8A0aYmXpw5k/aNG0HTiLW1seHdd9HcbhVonI5IyzQz64BpCF2nY8sW8vr1Y/MnnzDu+OMZeeSRin8oJS+fPRNkI9AOUuDL19BdfqStiPJIScNXX2WUa9ITvicYJBkO07JqFSOdpo+SkSM5afFG7nV5EdEwum0jNQ1GjuOUYhdCwNgf/YimFSv49JZbMoLiZWPHcuJLL3UZ99uO5KbRbX1LA951nlsYg182QEdbG4Wvvk2odBiF7QncsQhtlQOZe+pPSRSWsN8Tt7P5Z9fRNmIc/SOtDL/9euLxFhLNtfhdkkhKJTc+3WZZjSRct5E8wLIlkeYY3iIdr6ZhSwhbHkZXB7mj9TCGezchtfquzTBCIIGaefPY5eyzM9u6T/i5uKQIXupUFAe308cUlVCgw4xujlzBoJtPXj2J82ev45mKfKQQ5IXclJYGOKtYY9R28u5HMpL5zMNGojm3QTYSgWAE2ylRFr4JiKOmIaU8A0mw6yHxEXhVVlg2fjwD9t6bzR9/3EUZKVRVxfiTT+7lxDmwO6F5HzBXk9HyCd8IRa+Ce/L2jTP2nHrMNUvUXGAnIf4iBLLqP59+upkvvtiKy6VjWYqjmEqpMWuawDA0fD4XQgieemppXyDcQdAXCH8A+PS229j86acq05OSRGenkhkzs543EnpSI5xJx0om1QTk3N66fD7Kxo0DYNSM41j27LMESgahGZCImhQOiVM9ZQq1DoUlVFlJpKGhBxle2jata9d22XbVj6Zhr2zgqXbVY+/Oy2NGsYezc5ac9rv2WvZwGntCVVW9Bol0JpRb3Uy/zwpdaX5eshU8zQ1MvPDHGE31JDxe9ESCgvoa6g79EdawkbRXDuSN+1+lIuQjKAQrXV4+PfOXnPG3O2j8RGfv/glKg4AOa5th/kaJy0hhB3z48goQLTFizWGMgAdN0xg5MI+O0rEsie+C2RGhfy+tmAIoHDKk9w+zvQlevx8WvgVuH+zzY6YccDJ/KnNzbSNEbPUeyw14rp9yY+oOXdd44MfD+GVS3RSkpPIqHLWdVksAgxnKMIazmlWQU2YdyjCGMmz7TmKudo5NHy9QtyopkNkuTyEER82axasXXkjL6tUAuPx+dj33XPrtscfXv0bHNWCuAoI52VwE2k6HsmxZdWUC7m6Bz+NQqsMZBXBMyPnuaF/jFiG6ptDLljU6BH3haI3aOd8/9fdTXh5ECKit7fj6sffhf4a+QPg9h5SShbNmIVMpDK+XeEdHD/FscoJcLoRhkIpG8QSDaG53r5POHldcgTsUYukzz2DGYhQOGcJeV11F1aRJmexyyZNP8ur553eRtFKGuRoVO+/c5XyagF+OKuMiCzanVEDrbssD4A4EGLTfftt839OCUKIrqbD0VJWSaqr9WbHKGOMShj5xL0bTVuKFpZgIEnmF+MLtjHrqz6QuvoYaEcStQbFT6a3M89MpJH/faX+Of+VZXl4p8WgmSJuYKcBt4B5TDWV+7EiKSllPo20zeGAxQ8rduA3B3fo0TNNG7nQIntrXSHR2ZjiMqWgUw++nbPx4Vr78MoMPPDDL5YxH4NazoLkW8oohFYeX7oSaVVww8w/MLFSBrUCHKd5ty6ClMditfv5VnMBJLOZLvmQRADuxExP4BsoswuXEhlzVVUeF1b1vl13zqqs58YUXqF+8mERHByWjRxMo3Q5mf+JlwN01m8MP1iYwN4AxiA1J+EmtEhQfq69mAv9gU2uUl5L7M6N4Mvh+AvE3wDZBc76MdhyERwl+56CyMkhenpfOziTBoBLyNs30+qBg2LAi8vI81NV1MmVKTw/OPnw76AuE33M0LF2K5ZjyAkr2LGfxQtN1tf6VSqmZ07YzwtNIqeTXgkH2uvJK8qp7Wt3oLhe7XXwxk84/X8mTOVJtmXMD408+mQ9//3vaN23KaIWa8Tih6mombMM1O19XP9uDDgs+iynNyt19SgnFJeCl/vCjLVDnJL4eATeWwt4BuLdFbQt+8i5Jf0iN2bkXMP1BvA01+Bd+RmzngxiUMw4hNAqLCtk0bBxRbwCv0LBSSdwlBWjRDgi6oDKIFw+e4ABaBpRjRpZiuDTWG8P4h34Q66xqYrEwJ5w2mTGXzeGF00+nZdUqNZ7KSlKxGM+dcIK6hi4X+113neLLffEmtNRBcVohxQ1ur8oODz0Xb8UgDu1WCv1vQkNjIjszkZ3/+c69wTUJku+gpA9yAqE2FPSeJUOhaVRM7BlopW2z7o0b+eL+P9NR20HZ2IHsdsmNlO9yOD2lzbscCcAT7UqI4RjXs5yk3UChqMNHGDspMJumYBTNAf/ZEHsQ7ISTxLoh/0HQunaAHnjgYCoqAkSjKVpbY3i9BlKaWJZkxIgiCgt91NV1Ulrq5/jjx/5r160P/3H0BcLvORLt7fhKSog2NCB0XQXBbjZL6XUXzTDQdB13IKAcJ5zuzlNef51ypxTasHQpb159NRvffx8rlSJQVsa+v/41k849F/c2JL80Xeesjz/mtUsuYf2bbyKBEUceyaF33onRi/V5R20tK2bPxuX3M+b447OcwV7waCtc0wAdtpqfinV4rBr2C8BOPlgxDObH1PN7+SDgBLVRbqhJQXsgj1BHO2bagQmBOxYBodHkDRLQwJ+TWdWasDJUim2EePmGB3AlYkxdMZdxG5fTsnYtiViUwkQ5tcLHahOE5UHvP5rbr32Sgr9+TnBzM6UtH/PjgrXU3/E6wUMOYeb772MmEtiWxYOTJ5Nob8/cUJjJJO9eey1Vu+7KwC3LQOt2dyA09VO/Hlk+8F92N1jGMj7iAyKEqaCSAzmYCir+pXMlk/B/98Izr4NpwaH7wI2XQUH3/pTA5WBtVO7yshMQIIog755v9Horn7uMD37/AO6AC2/ITeOy9bx8zokc9dBsygYeDrHHwHblZIVR0PqBoZSNliWgQmvhx9ofKBUb8aDcUzQkuvkBNE+B0vngvxgSL4IIgu/kXhtuQiEPDzxwFP/3f++zcGEt0ajJ6NGlnHHGBD74YBP19WGOOWYUJ500jpKS7exM6sN/HX0Sa99zxFpaeHz6dOItLUSbmrAcuyRQd9hd1GJQQatil13whkJEmpsZPWMG+/3mN4CSuHr6mGNoXbMmoyEqbRvdMNjn2mvZ79e/7vLa0ZYWPrvtNmrmz6dg6FD2/PnP/6lY9LvXXcfcW2/NeAPqHg9Hz5rFqBkzeuy7Mg5TNigCeVrPJCYVkX7dMAh+TUb50zrVVdrv9efY885fEy0swTZc+FubKNiyjkhZFZH3v2KPPIN7WxUnUaKCqpQWxWuXU5SIkDQ8JN0eTr3t97yx0EuxO8aQxDys0iA+lwYuNzVX/YGaKfuTiqU4/7Zf0vz2i8qayONBd7koGT2aYx59lLVz5vD8ySd3yaoBEuEww6ZP54RLfgzP3wJFOQFKSupq2rnLewlvL+jA49GZMWMUp58+gQUL6qip6WDYsCL23LM/LlfvF+QTPuZt3lSfv2NsrGMwk7OopOprP6/esN/pMH8JuBxifsqEARWwaDb0MA+JvwPRu8HaAPogFWy8B273a9lmhL9O7Ye0BC5/Vgkg2hShercJTL/7DWjaE6z1ZLJD4YfCl8CjrKmua4DG6Dtc7zqDIlGPouGrvweXsBQ5P+9hCJySfeHkIog9CcIA3zngGtxjbC0tMXRd9NBx7cO3hz6JtR8ofEVFTD7/fObfey/e/HysVIr2LVucOauEAAAgAElEQVRItLdnm9+FwB0IqFKmEDQtX07p6NGMP/FEpvz0p5lzLZw1i84tW0DKLLHZUZz57Pbb2euqqzIGv+vffZdnZsxQjTkOFs2axRF/+QsTzzij17Fu/PBDPr3lFnS3G92ZRc14nJfOOovBBx2EJ9S17veXNrXuF8zJ2PxCNYz8tR3O6+lWBEBtCj6Nwa4eaDriODauXMSQV5/GHenEHYsQKyjhxduehYTBTUHI0+DPrbDeMdwY4dHxBn10drSgJ1PE2i3+sngghX43RROG8d7E09DMLeR73RTOOAyzqIQ8ILJ1E1tefBQtmcwU7NIWU+vffZfw1q3Ox9E1qxNCKEm3SdPhjQehrVGtEUqbSF0dZ7+zNw1yKSUV+dieYh5+eCF33PEZZWUBbFuiaTB8eDF//vMRPeyYTEw+4D00NIyc6SBJkjnM4QzO7P0ibgPvzIUFSyHoV4kqgMcFm7fCQ7Phwu5Nnt4D1U/3ribASnSweNYFLHv+Dcy4yeADdmHSxfcQqMjSZJLt64i3pQiWd9XudIfcNC5fp7K2kkUQvQ9ij4C1VZVerVqQNgiNU/Ph5qgXt4iCcxsAiosp0g08iReygbDtbJVl4tTcIzeD72oo+F2XMRQVfU2TTR92KPQFwh8AJs6cSemYMSyfPZtEZyf7T51KxcSJvHjWWYRra/GEQuhut/K5E4LOujqOmjUrI12VRuPy5crCKKcLI+3vl4rFaF23jtLRo0l0dvL3s88m0dmp5NQc41/bNHn90ksZ86MfdRHzTuPz++8HKTNBEMDweklGIix58kkmnXdel/0bzN75cxKoN3t5wkGzpZr1DQ0q3ILwVb/jr0fPZNBnbxMvraJ530OwXS7CFtzaAr8vh2Py4P5WeLAVqlzA0CH4igqJ1DdQ99pGAkVlDB5aCkIQL6mifuQkaI0w2Z+fyVYrX3wcs60ta26saegeD521tWxduDBDBbAtK7O+Kp112oFHTmNTsAPXFTdR/swstOXzMGNh3lxdREPEoCLYAvWN0FZHOF5Fc3OcIUMKCAY9SClZubKJhx9eyE9/unvXa0EzKVIZdZg0NHTqqaM7tqRgfqQFt13DXn4XBZ6RXQLYh5/jvLfsMenfP/qil0CY2albSVdK3rn6ANa9sxxfgRfD7WLly/OoXXAAxz27BHdI2Rm5Q/1xB3XMmInhy05nqUiKsvHpLK0DYg873aM6WIug/SwwP4e8GxnugYvLdyXRmk+IVkBHF6o0qr5NerY7NPa2w320gfT31ILYH8F/ArjHb+MN9mFHxr9mVtaH7xSEEPTbfXem/ulPHHHffYw+9lgKhwyhetdd8RcXE6qqwl9SklGb0XQdf3Fxj/MUDR2aca1PI63qorvdeAuUNlftggWKLpHjmpB+NONxVr78cq/jTEYi2ySGJ8PhHtsOCDiCylIJQ7dZKsilJOz2NTfjg91q3k2rqLRY0DZgKF/++Fw2H3gEthOIczmHhoBDguoxJQEEvsIigiNGoXdCeVEoM5mXrt6q1qOEIJFQQS9sQ/UrT2fFClAmxWYshunouRYMGsSEU08lFYuRjETUYziM54zdWP3zIt7gVf5eMo9HzxjPs4GJPPRaLU8u8hBO2Ni6AS4vdjxCa2MYXRfEYmbm2hcV+XnttdU9rkWQIAKB3e3KS+weAtoPt9ocu7GW3zXUc32TxuGbEnzc8DuwWzL79KvYtrNE/8ret/eG1jXvsOH9FYTKg7j8LnSPTqgySHhrmHWvZ415NXcBu8w8imhrjFQ0hbQliY4YVkqyy7m/VDtFHgBzpSLoa16HDqFD9F6w6gGY4PNQXvQgmjBwCRMdC8XodIMoBO+P1LmidwMWXXMIhwMZ/u32v8E+7FDoC4Q/YIw5/nisVEpleaigFqmvp2rSJIIVPRsldj7rLILl5Sq7s231aFkIXWfg/vsTqnRmOim32acH2W7S7hjtrAPmrlunuY6jjz22x/6n58Nwt3JNb3f88izn55p6iFk9DiHW0Unnhg2c443TZKkgCNneQk/OwG26ClAPdcOFhdBiSpLvvUXJVecw5JKTOD35KTJHFq5gUzPlX24iFfLS7nZRZ4IZTxCs2dB1ME73rmWaDJmqyOOH33cfh997L+U77UT+wIGM/+1PCdx1HB7di58A3oSL2vVfseYwCPh0+vnDxKIpmprSry/ASoJt4fNmJ2spZUbfNBcBAvRnICmSmWBoO/92Z8/MfisScF9zO0VyPVVaBxV6BI+AazqOINLx/zL7nXI45AUgElWVR2lDLAZuF1zULRuMbF1O/RfPEW/pyiUFaFu3EKEJhN6tTKxD41cLu2wbd+ZD7H3VaQgdOrd2EijROeQP46kavQhSyyH5hrouuWmq5gaZgraLoHkqNB8C5kJE/pMIUYrK9nygVYDvRPAelh51j7FmYLdu+7kdHAticMwmGLsGjt8Mi3u6rX2v0Vca/QGj35Qp7H7ZZcy/++5MCa5swgQO+G3vd7YVEydy7FNP8co559C4fHnGiHfI1KnMePTRzH5VkyYRKC9XprJOVpjOHF1+P8MPP7zX848/9VQWPvwwNfOUpmPacmjX887LeBfmwqXBU1UwYX1WT9QD+ASsSql1vcscr1UzZXLvTXeyYfaz2LbE5XZx0hnnsHbGaTSagloTIk4lTKIyTA34ebd1xjMLofyRu/ny0UfQfD4KPC7MTSvYrd7DPO008ovzSKVs/E/P4+JzJzNx7zL8OozdsJanFNO6izoKKDHtpU8/zVfPPsuwQw9lp9NPz6yjfsLH1LIos34Xqa9HtCWQ5QGSo4oYv6qWdztj1Id9hDxhNDuJlyC2aRHcuhS845GGi5aWGDNnTiQeN/F6u/7Zn8AJPMWTbEEZFmto7MYUdmNKZp93ImDbrbi0rLJ3QEvQYOXzebSZffM6QMvD64VX/gwnX6HWBQGKC+Av/wcDnPskMx7hw+uPZM2ceWi6QNqS8Sfux26Xv4DQVIk2VD0GaUukLRE5HEDbgsKhXR19heZi7Kl/YcwJ1yKbT0XTw4rjl5gDyTkg3fQootsxIAbJVwAXiBDYW0EfCmXLIPW5KqVKG4zRIBNOY8zpkHwLdZuUDqzO55nOGr9jeL0TTqxRQu0asCEFcyLw9/5KaOGHgB0mEAoh+gOPAeWob+39Uso7hBBFwDPAIGADcLyU8rt767WDYcIppzDyyCNpXr0aX2EhBYMHf20Lfv/dd+eCJUtIhMN0bN6Mt6Agmwk68OTlcdh99/H8CScQb3MUQpwu08qdd2bVyy8z5sc/RuvG+NY0jdPfeosvH3uMr557Dpffzy7nnMOwadO2OZ4PYiqLC3RLdjQbXg5nA+Hddz7IxqefwC4uR3MZJBJJau67g6P6lXHo0dNZnYDjtsDapAqofg3+UAq7B1Q29eyzy7jppk9o2LCFY6KPUTWsmn7FjkZoKMgoaigZ3cEHdUGK82DmtHym7q5THlLrfe0FIdzBIKloFJlKZaZlaVkIw2D9228TteHDv7+GPOYkDrrycvb0QUIkMpZCyaRFY20zyUgKEXCxRSsm1VjLga5PWWIPoSnan5DbxaW71vBFawmbmgUyvoZ4qIpk0uKRRxbxxBOLmTp1KFdeuWemm9GLj5n8hDbaaKOVCiq34Sto0xsnT72XrL7oLmNgxWuwcj0kEjBuRFdy/4I7zmb163MJlgUQuoadsln0+Lvk9f8Zo0+8F4DiMdOp2mUQWxasJ1DiQ9M0Ii0xvAU+hh1xWa/fBRF/HKFHQU93uoaUxBpRNUo7hsr0UkC6iSuo0kyiYLcBm5UThrUWoveA3a6e1yoh/yHwnQLh/wfWYlTtwbke2nDwn9rruHZ0XF6vgmDu31DEhp9vhQU/EF3wHYY+IYSoBCqllF8IIULA58AM4EygRUr5RyHENUChlPLqrztXH31ix0CkoYHP77+f5bNn07BsWbYkKgQVEydyxrvv9soj/CZ4pl2pgvi7GceGbZgehOf7Qzxlcs1eB4LXh8h5Pauzg0BlFTe++FRm2/I4tNmwq5eMvdIf//gR11//HomERX82MY1/EBMBBg0sYNw41bQRbWpi6CGHUD1lCh/+/veqpCsl/rIypt9+O0VDh/LMsceyds4cZfOkadjJJGY8TsXOO9NRVs36FNQOHMnyPaZSu9t+9At4uba8jo7QbDxJP4sWbkWEmwnQjOX3UHfIh3jiSdyYuOwIQ0cM5pL9E1DSH0vC5zU+NtdFuK1xOpbLR3GxH9uWNDREmDixggceOHK7eYfL4nDmlgZK5CoMJ2uLSTcxW+MfxbcQLHpk24uDObBNk0f2KsEd0NHdOV2q4STeQoMTXm7Mbuts4LNbzmDVqx9hpWz67TaSPa/5MwWDJys6hGwFfRho+Uo2reVgkHHQyhThHcBKqKYYnAywC9KSbi6nGcYCfTC49oT4cyBjKgtEgjTBGKg6UBHQ+XuIPgrCVplg6BegbaNNeQeGaUPeyq4WXoASiZfQOfrbG9t/A9uiT+wwa4RSyjop5RfO753AcqAaOBpI190eRQXHPnwHECgrY4/LL6d940ZcPh+eUAhPKIQ7EGDrokXMu+uuf+m8ZjzOV7Nn88IZZxC87jL6rV1GLOuelJFSu9DRJ22NJRGxKCKnGxVA83hJNtQjpaRu4UIWPvQQ8tXnmRBtygTBpqYoN9/8CYmEhRAQw4eGMsjduKmNSFRxKmzTRHO7ef/663H5fPhLSvCXlRFvaeGNSy/FNk2OvP9+hk6bhicvD93lwuX3E6yuxltVzYYUNA4ZxafHn0e0sIS8jla2pOC39RW0hCdR096E9Ccx+oWQhT7iNy7CiKWwpU3AYxEODebxlr2p8QwBAboGu/WLsl/hBtzCpKQkoDwldY2KiiCLF9ezcmXzdl/zMR74SWGIZllBjRmizgoStTV+H3yAYN6V2xUE09fJjKfQuomg6i6DeGuiSyOWO1TGPte/zsxPWzl7fhuH3f8FBYOGQ/t50HaCsnBqOQQ6r4OWw8Bco35S88CqU2uA5lygla6apmmku0Ith9Dv/G5+BTKsmmuESwVV4VOSbMkPVYaY9xuoWAvl6yH/5q8PguZGxZd0GnN2JGgoRSarWz5kOnzcHwp2mNJoLoQQg4Cdgc+Acilluo97K6p02tsx5wLnAgzI8Wrrw7eLtW++iZlIdKFLpL3klj3zDHtefvk3Ol8qGuXO8y/ltZKhNO5yFGXrlnPYFSfx5UkX8dnx54FU/K8ri+AgR5CmJOgjNWAIrsatkJefPVlHG7699+edX/2KtXPm0NHWTqSxCWmmKJw8hcN/cy2bGERHR7ZzoJkSGimlhEaidoCtdZ1Ul+joHg8CRX7vqKkh3taG0HXy+vXDSqXY+uWXVO26K8c//zyt69aRaG+nefVqPrn5ZtotNQUv2fswjGQCdyxKXFct/JYUzG+ZTPUza+goqsNle1g724XnndUMFE0kpIeN/acSDfaHjY0srvdRnecQHjtbqNOriepBcpd6hFCC0A0NEUaNKtmu6y4EnF/sY1qwP/M7l+G21rGP36TE/2vQy7b78zO8XkpHD6B1/RZ8hVlllVhblGFTx/QIqLZpsmXuXOq++AJ/aSmjDngPl/EFCKc91Y5D+FYwhoExHMxlIDWw1oDVimpu8ZIt63avgNmork9LrQNqxdldcoeSluAzV4HnALVNSki+B/FnVQnWczB4j8vKrtkRaD0Wkh8759DBcxwUPOCUY799aBocGVImzgGpmp1tqUTvTv4GTlXfdexwgVAIEQSeB34mpezILd1IKaUQotdarpTyfuB+UKXR/8VYv28w43GklLh8/zkicHc3+QykVJJv3xCzX36TOw47B2kYeOJRVvUbzOo9Dua4Gy/lvFOPI1pUwvQgVOQkfy5NsNcVVzL3skvQmxqQfj+EI0ifn313Hc/ae26nBYNwfQPSMMDloWHhQh776RVM+MmVXQcgBP+Qh7Af79OPzciwgW/ECPb7zW9Y8tRTtK1fnymLSilpXrECT35+xjJKCJFR1ykcMoT5d99NLNwJ7hDtJZV4O1qRhkG4oJiorayVOmyNUyP9mP9slPLyIJ31br70HcSXHoNEwmS3ftV4DBB1nYQSjdDapiZpX5CGqZdhrdjc5S3YtsSyJEOHdnWRl7ZN+6dvkazdSHDSvvgH97RTGuLxMMSzC7BLZlubqTKLvO2cTfa65lZeOe9Ewg1hDI9OKmbhDrkw8vbl8WnTsJJJhh5yCBPPPJP3b7iB2vnzEULgCZlU9FtDqP9EvHnpeSGiApbdBq5+qtnF2qCyQWpRa4JuIMy2g6GptmvlkD8LoverrLKLMbOpuk6NCdnDon9WFAwZU/um5kPiH1DwEAgvtJ2hMkgCjo6vCfGnoXOQyih3EPylEupT8EkMhNMsdpAfbvsBOUTtUIFQCOFCBcG/SilnO5vrhRCVUso6Zx2x4dsb4fcT4YYG5vz856x/5x1s26bf5MlMv+suCgcN+rfPPXTqVNyBAKlYLBNg09SLnU477Ruf7564D+GyCLa2YdsWRriDaGEJHx97Nie/9XwP0n0aJx0wmdCsR5jz2BOYG9aTd8AEjp95Mk1/vgPpctO+Zj26YWT0WEUySQeC+vefJxjclba2RKZqF8PPGxxKoTfJi/cewz6H74oQgnn33JOhomi6ruZn2ybR3k64vmdZzFtQwPQ77mDO1Vfja2okr3Yj4dJKOgaPIIY63u3wHT/ZZyzWa6sJh5OUlPhZv76NSCRJRUUQj8egtTVG/qjRTL7qINiyBFlcTb2rnII1Wxjt2cjyLRXkF+djmjYdHQmOOWYU1dXZW/54zQYaLp2Gt60GKQThv0D9hOkM+n9/y/BLu2NeFM6uyzYYTfDA4/0UzeTrUDbxUI575l2+evoPtKxdRfm4MdR+qfHVC+/hLyrC8HhY/vzzrHzpJcx4nLx+/RBC4C+KIeUamlesomryZGd9U4IUIJxmHb1aBTRzsUNnSEAXE+HcdE9DBcdK0AogeCEkP4Dkp4CprKCkR2VvMqX0Sa11YA1QXamRO1UJNh2TbQnyHYjPAc9U1bGKP9sppBlK8zQ2a4cKhH4N5gyC1QlYmoCJ3n/PleS7iB2pWUag1gBbpJQ/y9l+M9Cc0yxTJKW86uvO1dcss/2wTZMHpkyhacUKpegiBHYyib+sjAsWL+4ha/avYPXrrzP7lFMyvoZCCAYddBAnvvhij87R3hC14cE2mN0Oy1atoWD9KjSZXt8R2EIQzy/iyQ3vss8vfvGNxvbWL37B4jlv07Z0McJWNUppGEjDRdvYnSlNxSiceSdXXfU28XiWDG8YgqOOGslf/3pcho7w1IwZrHntNWUx5ewnUUFx3IknMuORR3odg22azFv0FVd3+JlfMYSEY3WbbgAa7Vb9iQev2Miiuz6ipSVGOJwkGk1RUuJH05TH3f/7f9MYMaIYK5nkzSuvZPMnn2CbJrbQaYkaLKg8EaOkmuOPH8Mxx4zGyFmn23jKJHxbl5NyBZwyoI07FSF5wq+pPrfnNW0wYexaiNlqnBJlBlyqw+phitrS8422KcNdvQrc2YyyYdkyXjrzTALl5V2adzbPnUuwoiJzQyaEZN+L5yGJUzZmsrKukkkVvAiokqRWClYzSIe7QYIMvaELcuuebhABcE8Ha7la77M7VZkVS51bmKD1d9YLDXAfAZE/5nzKLhR5JwLeGRC8AZomqPN26UIxVdpV0UIf/vf4LmiN7gWcBiwRQixytv0S+CPwNyHET4CNwPHf0vi+l1gzZw7NK1eiVeSRqvQi3QISNuFVLXzxwAPs8fOf/9uvMfzQQ7l07VoWPfQQ0ZYWhk2fzsB99tmuY6WEK+vh06jyF9SDIVIeD0YyiS6VC63p8eJvbcpodXY9XnL3nE3c9eiXtDdGGL97f/54zk5MGqAC/IgjjmD+rFkIyyl9aQJhphCWhZaI4xk6jIsvnsKAAQX89rfvs359G4GAi+LiABs2tLH//o9w9NGj+NnPplA8fDjrXC4MtxszmVTNKW43tmluM6sCVT7efdIE3rDh+ga4r0UNxS9goAtKDUX8t3YZyCuv9KehIUIw6MbrNVixognD0Bg5sgTN4dstf+EFNnz4IVZ5FQhBvgbu5ibOGLiIGY/8rMfrJ2o34qtfTtKVI/YtNCxhkHr9EeglEN7brG5Q0jqvAgig/B+f7oDTCrod0P4riN2teHnYoA+BolfAGEj7xo0qC+9FYzUVyRLYpRQsfXUYOx2zGKE1q/VBqwZFh2hRCjf2Jmc0VUAHKuvrToLPbZzRVKanFUDiJXBNdhRoPOAaD6lVKgjqE0B31pdlQpVFiQHp8n4KFXQtiL8KqRXOvlEVDLNXG4xv4NnYh/8JdphAKKX8iN5ISgoH/S/H8kNC7YIF2AEde6APYUlESiINgVXuZcVXH7IH/34gBPAVFrLHN2yMAViRVI4PlTo02RAvLKV5wh74WxopqlmPnkyQyCtg/7efo6lhTY/jL3t4Gfff+jGGz4XhNvj0xeVMe2897z59LDtVBwhVVeHxeIi5vYhETN29axq27sLbWM/+t92CEIKjjx7F0UePYtOmdk466Xl0TTI2rxlhmbz4XJL29jhXX3ghC2fNwjZN3IEA0raxkkm8BQUEph/Jsx1KwHtvf1fOVtxW7uizOxV/y6WpLLAw568zKWGER7nLV1Zms/Tx43su5Mx98WVWukIkk0okTANGFhRjL1tGpLGxh6GtFVZO6aLbn5/UdLRk70oqq5I9V9rSNperk92eiD0D0dsBr1MelKrE2HIElC0hVKV4f7nGzQCeggJkKoVlmujOWvOmzyzy++9M2Tg/EHPk3fwqU5MpVBk0nQU6a3+ZYKWhMjyXs08QjJ1VgLPb1HPmYsCngl36XNICawmIEYqagUuVTXODKZApwYpC0CtVaVZuVCoAGOr8wgN5N/d6Tfvw7WGHCYR9+HZQPGwYVqEbYclMBUnYIG1JePy375e2KaWmmzoLahpbGPrFRzTkl9IwYhw1RSWUb1nHtE9eYZ9NS/H36+r43REzefi+efgK/Rge9VV3+VyE6zu5/ollvHD1boTr6sjr1w//sBFs2FSD3tKIME1sl4sB++3LlAP27nLO5577igqzlquKZlNgdSgVmlIPd35wBPGfTuGoWbN47aKLSHR0oBkG/tJS2vY9hF9NOByzSU2ZIQ3uqYTRjhHE75rg9bAqKxZp0GLCogTsLFTgbLZU5nX0dlSpIzZ8HrMJdLSRV1+D1HQSxWWsCOYxVvbMugB8w8bS4s5HT0WwjSyR3rASxHfpXQVoVx+81E3+Nb3KMql7r1X4TlQzijPdaALsgAqGyUWU77QTpWPH0rB4MYGyMiTQvHIl8fZ2XF4vNXPnEigvx1ccYvr1m6ia0IlICVRQSwD5ah1P6EpFRiZQDeaQLYtqqO7RdHlbqOfstaCNBrtDnUvaqOwugpoehXqUNqSWgcub7fgUQcVbxCZLxTDAGKBeR7icu4WUM9Y8yLtDNdwkPwJRDK7vGVHvO4q+QPgDx+gf/wgx6zfI+k7I9yKEwI4m0Qwd7bgx//wE/2VUG4rjlJj/KTP+dAVGKok0TRIeHysOP4mSijIO/vJ9IuEwE7o133yxrh3btDNBMA2X382iBbUAFAwejLQs8vxeJowZTqc9HFuCaNjK3kcdRndsXtvANaGn8EuTelcpNSX9yI+3cqn5HK3rTmPMMccwYO+9WTtnDtHGRurGTuLxgbtT7jEyvrBtFlxdDy/2V6XEOWGo0LO+sRN8isBek4KYAXvGWzji05dZ8OXnFA4bxk6nnYavsLDH2AA+DdsQieJfvwqpGyDA1biVeGERiX0PwF/Sky4hNA3vRbdg3XYeejKMFBqatIl5i6j82R97eRU4rxDuaFHjT4fOOEr79YjuslyymWxW5kATYAuw6xB2FYfecilz73mR1a/+g/DWrZixKOVjS3H7wsQ63MTbwuz9s12o3mkZItOFGUcFwg6QpWlNPrJBKTfo26hgJAEfKntzq9Jl6kuHR4izT8zZPze1TahjUwucUmdKqcnYteo5aarjRIEqrSYXAmmepktlgloFdP4G2i/Nnk8UKjK/bFdk/tC14Dl8u3mZffjPoC8Q/sBhuNwMuv8iNl/6INaGJqQQGCEf+bceS1m//4y+UjwFDy+ApVtgeAWcMwUC29mVNtYDo+0Y7luuwXJ5SOQXqSkuGWfXp+9j8U8uJ9HRweQLL2TY9Oldjh1c5kfaEtuy0XIEp81EiqoBar0nVFnJiCOPZNnf/kYqFiMaiSKFoGjYcIYf1jMQHjygGd+KKK+PPYonDjgTS9OxNI3RNcu4tX0JMI5AaSkTTlHedT/fCu5Yjjk6UKCrTsvLtqoMLmJDWc5fogb0d6kuzN+0LOOF009jwerV2KkUtmXx1lVXMfaEE/Dm59O6fj2lY8aw81lnUTJyJM2fLyC0fBHYNloqhkAihYbHlnjOvmib17nk8JPoHDicpsdvRdZvwjV+bypmXoG7qHeeYVCHDwfBT7fCRzE15h8F4K7KrnJqALj3g9jjkCvbZqcAHeKvQed1eITGfhd42POin/PE0XdiuDZguOoAA3+eBDPM2tf/zsi99ewLCI/qGEWSDWwJ5wV0Mllfl6xQB2EBBaqUKiXQTrbhBVQG1x3pQnBUBU/8YK0GUar4i3ZY/V8foTwPCZNZixQedW57kyLq43OyyU6QNc57KABzCbSdCqHbIXDmNj+rPvzn0RcI+8A+Q4/m9VdciNWtuJICOaII0y2ZnCO6/K9iSwfs8yto2KSWWoQOfyqDd/4Ao0r/+fFCwNV1C/lzMk5nkeNBJ8Dv86IF/EzqqOO0OXNwB4M9jh1Y6mP3acP45PXV+EsC6IZGIpwABFedNi6z37gTT+SLvz5JS0MTpq5TP2ICyw48kVe/6uCeXfIYkBO0992tmHdXjNcKlfEAACAASURBVGLWwecQinfiTiWxJKzrP5ob8lw82M1ftnteAqrcuyGlLJ48Qv0es2GMN7tvXMJEr+TtX/6C1jVrkLatdEkBK5lk8eOP4ysspGLiRDa+/z4bP/iAox54APcrz+JpacQK5qFJGywTYdlIl8GIzqavvdahcZMI3fTkP/9Q0tfXDS9uj3ZF8DpIvOrQGXQypURjIqQ+VXQHoYGMk2r8A1a8A68/DGQviNsvaNvSSpdgKgTIINAJMuIEwbQYdgEZtwkpUeoyEfX6ohz0sSDrwFxONsilM8beuIa529IBLgSyEWyPIuLrh4C9Eew0w8sGDLV+iXC0Tsm5Bumg7WSwWlDtE7lJeRuKPmPf/xX6AmEfGMwQDhOHM3/EPFpooYgidmMKAxn0b5/7gsehfj0Ei7PzUns9nHU/fPKr7TtHvgGVBlhCBUFNKBk1DcGeRYFeg2Aas2/Yh5M8Bh+8uhJpSwKlAX53wwEcPTEbhT+//wG2+vJoOGBX3jv1Z0SD+RjxKOvaExy+GZ6ohp2c+Tc0YTIL9qnDbYDbTKJpAr/XoNhsY0nermxykolXwtBoQrmhgpytGlJJSFiXVM0yAxRbhaiE9SkoTEGxrmylCjSYGq3n5ZUrAbAdQ2QBYFkgJalYjFQ0SqC0lGhzM5/ddRfxrVsxBCQ0DQsNdOUf6YlHKY+0fcNPb/sQt+GhNnitU2W2FxZ1Wyc0qqB4nvLrS74PWgn4ZirX+HQQBJBRvMEG3L5WUlGBK6AWUTVDMnxalFHTBKpBxnbKkxYquGnOT9JZ73N8F9NqMBkyvU/tZ3eAnKvuzDBzBpouqfaG7ttTan1QH6IywoJHVeCK/Q3C/wdWGEXkDwLpcTgi3cJFz6Dr+IEJXWWK1hZ13j78T9AXCPsAwCAGM4jB//HzfvQh+PLpYgXnz4fF8yBpgns7voEVO+9MVdCLbkaocQVISgjYJtUu2Gv61zcU5/kMXv3dPrReszuN4SRDS/3o3Tzu1s39jHBeIat3P5hoMJ9AZytI8ITbaRk4iBubdJ6qdjK94ipaR+9FKGJSGHTMn6QJ+aVoXj/vReC+VjWt6UCTCQ2W6qTM15VFlADGOEnClpQKmLaEWlMFyCODcHYh5LfpSqHGsW3K5CSO4omUknh7O/6SErwFBTQsWULZuHG4NQ0XFimnqcNlmWiGTsmI//zEGrVhj/WwJmc57dkOlfVfkOvtbFRBwX3Z/5vrIPaX7BfDXA32VjTNZreZkvdvS2CZ4PIHGH14J/0mmviL/c4ViCr+YKYnVqLKmQLsZWQoE9LtNL8kUBw/Z22PuJM9pr98aROvXgwstwknW7UbwfJB+0Wg5Sn6Rehu6PwFyA6QDU5l1lTH4JSEMyo36bKtU3aQtqJufAcFvL/L+AHJqvbhuwqXz8dBN95ImZVgfHsdO7XWMCLcxN5nnUnZ+PHbdY7CoIsRFYEeQRDAXVKClkhQM2wcnphqhUy53LRXDKDd1vgsqoJVGvuNGUesajjkl0FeMQwYQ7xqFDqCR9uVUkeFw/1bb0KnVJlgo6kCxzA3hHQVHNel1FToEeAV6tifFkOVCwKlpVRNnoy0rKxZsfMohEAzDAyvSlVT0SjBigpGHn00waoqdNvGaybxmkkMQ6d41CjKxo2jC8x1EL5ZTeLRR/4lY9mbG9X7CAjV2RrUVHj5VaN6r9uE3s9pEokp8rq9FfCAEIw6bAhTf+0hrzKGy9vMoN2T5FUaSq/WvR8YO6GCShAVQDTUq+o5P0Kt35Hn/D+JaoKJOz/pLMxFl4xsu+A0v0hLlUatGki8ANFZ0H4utM0Auw7VOVoJWgCMMZD/CIgKtU4oU2SnX5fKBO2katn2zlCl1j78z9CXEfbhv4r99oU5r0OwKHvzH+2AXaZsXzaYxoA99+SkV15h00cfYcbjVO26K4VDhvxHxrj7Gaex4tobcEU7SRWUEPX4aC2rJuX1A4KNJkxeD/dUqKk0KQUDgwHWekYStyEmwWfCxUXw1w5Fg0hIFeR0Mi0ahDRotVQQ3GpCvamCn+WIHQ9zqy7MV8JwisPdPuSWW2hauZL6RYuwUqlMAMR5DJSWYsbjJDo72euqqxhy8MEMOeggtsydi5VSzSCuQICDb7oJPdfyKjmfr1pu5e7ocSwyp1OhNfIT/20cVnYJwtiOxVsHL4UdkkHO/YVbUw1Ab4bh6G7CzVLCB1F4qt1NS+qv7C8e4yTX0xRiOuuBXtCLGLL3Jobs7VAXsHBuF4CYKq2yAefToOc05mRewgtsUb+nS6cZD0FJ10aa7YVThpVJVFnW7zTApKkcTnlVdoA2EIxRkHcPGA7f07MvdPxKiXULHxAAe6U6h/Ar49/Q777hmPrw72KHkVj7T6JPYm3HQV2napap2+hUfTQorIB3/wDD/s2b3mfb4ZZmVXqc5IUbymCU55ufR0rJS/c+wMNfrOKT486hvaQC0xtAOrqjThUTIRQ9IGErN/ukrdbHks7zBbrKiEa6YbMJyxLZIphLqEAYsVVZ1KPBwrjarqGOKTYUZ3B/P/whhydvWxaLHn6YubfdRrS1FV9hIbrHg6ZpaLqO4fcz6fzzGXv88QghsFIpNn34IZs+/hhfUREjDj+cglzdWClZ3XghZ7ZdiC0MCkSEmHTTKV1ckb+Wk8qP2O5rt9d6+DLeVSBASnV9XuwHU3O5j1LycJvg7lZVInYJ6LCi9GMZT/h/TEjPVxQDc5FDuUgHv/QVBhXk0hmdRnaNL93xKVFZYgJIl1JzCY+9NcL0hrQOafr39KNX8f80l8oEtX5gryMbWHNeR1QqObmCB8G107Zfyo6qjFgrV9ljH/5r2JbEWl8g7MN/HSkT/voFLHboEzMng9f1z4/7OvyxEX7blKE7k0JNrh8MUt2X/wq2tkc4ZHWcxb4iJflFVoxLooJaUFPTXcTOTsFphpxANfP0cymj0y8T2RWsoDPxh204yKEZHLEJCjXIy+EQbjXhgkI4q3eaINGmJsxEglBVFXYqRbytDW9hodKJ3V7YTfxq0995K7kXZXp7ZnPc1kji5c1h43BvJ41tVitculVd+3TVOWqr9dBNw0CTrdBxASTmYEuTJdZ47rfuYpPIao3WpUwud/+Ok7xvqatofkZ2Dc1F1lC3exAz6JrV6ahAGUM5z6fJ8O1km1PS58ldn/tncMx50wR7fSDoO4H1OVgNZN3uu8MH2gDIvwW8vQsT9OF/ix3emLcP31+4DDhzN7j1WLhgz68PgraEX9dD2QrwfwWjVsOb3eaZuA03N6v7/oCmsqugpkqUv/w3vE+L8gIYhcX4NIGOmvbSU2BaOyRmq6k5dwrNbdmQEjotFdjS7RcuVGdoi63ylMEuFSynB9V203nfjaZ6H0d+jYKMv6SEvOrqjI5poKzsmwVBAHwsM4cQEtEuW70iQRwfLd9guWxmPhybp+geaU6kX4Onq0ETpnKNj/8dpMDCw3DxJb9zHUoxm7KvSzufmaNVJmh+QlaJxSGpZ5AbBAUQxJZe5pm782jidF5LHUJYGijJtUJAc1Rg0gXqNNJdpj66klvSnzrd9lWjRHhAlEDh++DZRfkN0k1epwsE0Ap6/6+9hv9trF7dzNVXv8m0aY9zzjl/57PPtnyr49kR0bdG+B3CuqT6KTdgnOf7KT5xWo0SbQY1jaxKwWGb4eX+MN0JEEsSag0u0O02zg187lCzpFRmo4+0q6aVKT7lWD/8a0qnYdvx1RPQ1EuhRKIyvrjsGgjTOYlATeEVOjxQBR9H4NcN0J5zrgEGfByDe1rgd2XqcXanej+TvHBFiRLZ/q9CCzDc7eLDuI5PJkm7TSSlwK0VUehc1xQp3uQfLGUpFib96M+hHEYJ2TVETYPHq2FZMbwRhhIDTsgDrwbE3wJzJWk/Pk1CTPjxE2aGdiuz7NvBbiJpbWWQaysqMEV7GXAu0l96SUJKLovewnxrEhY6Om4KtA7+7L+AoUZc6YFK6fD90tqg6cYYP+rTyv2g03cAuWXRYnWccIM+QO2TfBVizyqptNSHdA3WuUiAGAjG9jV0/TewenUzJ574PFu2tBOLmcydu4U33ljDffcdwRFHjPjWxrWjoS8QfgeQknB9o5LiSmceYz1we4UqQX1f0GjCc04QTH8xddR0dVEdrHUCYZXzpC1B2BaWmVJByHBTqGtICefVqoBqo8p2bSYsiMGT/ZRqS28o0BQPziUgmlDrXLll0XQJNrWN1YT0FOvTYSeP8nXbmIIXOsGnqRsYr8OBfKZDUSQuL4HLip3M8Rvc2LRYsCSuzruz95sdC3Bm6c58sKWRNtsiX0SI46ZVDOOCwgI8TiB8gsfYzCZ0dDQ0NrCeWTzIuZzPBtazlCWkSDGEoUz27sbl3m7rW+ZSdVUcJRhdgEsIbCkYIJYC0JmqR0dwgvtVRYrPCGN/HdQHMDs5nbnWZKpELUKoztEm2Y/rYtfxuP8shIiQLVt2W7/LBFyDrlzC3G020AjoIDUwHRk0c6naprlBHwXWl3SVdUs/GhC8/lu9Y73jjs9YvboZXRd4PDqWJWltjXPJJa9z2GHDM44lP3T0lUZ3cNg2nLwZ/tSkmiuWJ9RN7pKEKg9+n7Awrqaf7rFdB2pM1aQCUO2CSV5JZzJFrL2dVCRCLBYnlUhwQusajt0MD7WrUmlCKoL6mhTUmfBUO9uEJuDyYhXsRrtguEs1uJTocHq+CpJxJwj2Nn0I1B9UhZ6d+zakVOAd6FJBEFTQsiWZEqQmtj+QmTb8pAaGrIYZm+HULXDkJmWq+k0w1ufnnn4D6e8bSa22K8K1C5eXlvCTQjWQWmrZwmZcuDEw0NBw4yZBgr/xFHP5BBMTHY0VfMULzCZBt0EYo9VVyelDCAowhGSVNYYG0yRP1HNr8G4qtVqy3aFfh+y5Xk4dSZ7oQAidNA2imFpWWUNolGnaRO5aoB8oJRlx07wW4pnvQvdvXPdu0nSQCyuKiVaUHYde6Zw3PZUK1DqlD4yx4JvxT95PFpYFH30Of3oQ7v8bbKrd7kO3iffe24CUEsuS/5+9846Tqr7a+Pd37526s70Au3RFqoiI2BVbMGoixBKjiTUxMdEoqaa9URP1taWYWBKNNbHF2BU1oChYEETpHemwvc3u1Ht/7x/nzs7s7myjKPr68FmR2ZnbZuZ37jnnOc9DQ0OM5uY4sZjNtm2NLFu2G32Ezxm+yAj3cVxdCU+H0+39Fg0fxmCiT+jpvy5xy1B7GVVUspSlhGlmEIMZzRgC7FkJqOHe9vfTKTiIN1/mCODdzcv5yroEW4aNQmkwtM1Bi+dx7YRjSIbbbyHF3NycgA+i3R/DCSH4mwkPN8CmJJzvhwsLRAVmdjOcuc3NCHXnIp4JDLPaJwAT/PBsc/sybsyRvmZZH7P5mAOTN8BSd3jdAJodaHLg6p3w3GCw+nCDPykAjw4KktRuyMh47Ra3h2d0CPkKqKKaoQxts20KkUuYZtazjjGMTT/Zd7JkUPZ60DKKYtCKTwU4PeTjq7GhBKhE4YAdpD2hpWeY2Oh2x5dikcYx2gbYUwLaoLXNB/9qZfETErC1A6NOsTnieyZmuypBxwwxdUw+0PXgnQLRZ0VIQVkyIpFcKVfHGAy0ilxawb9cebWekUzCT2+DuQslgdYa7n8abrwaTji815ekEyzLIJGwsW2xuFJK4TgOtq158cW1jB/ff9c3/jnCFxnhPowWGx5uTLf2U2PDGhlitunbGPCuYj3reIanWcdaqqlmPu/xH56itcd+Tt+wv1eysMy8IHV+0/LaL/LVs/7LMY/9lUFL36dg28eMmfM8iw88gqTp6bSWppbXFCmlJ0wMwJ8GwDOD4DelEgQBJgfFJ3CCH0b5JbvJUZKLlLo+gwUWHJjRh/xmgWSCVUnxFGyyodaB7xXQVoIEWLFhCz/5z+uc89jr3DrzbWoaOzMR762HFXEJWh7SFJBaW4L8sj5mhSlYqnP1rgQR29YdLqaDxsLs5F2oUFRRRfsHvVA8C7wnInl2RLLE4JWEYncTJIwiZfWVmsXrDSTTm+55lrCTi6PTdKZqXcyB5hJKjJgotLQtcZrVryX44OEIvlwIFkKgAJa/qFn0aE/7dbM8FZLtqSIIXAS6Buwd8ntzOPiOk3nBwHQofAo8o3p5PjDnfXhrIfQvhX4l8ndOEK67C6K7+L4CTJs2kkTCkZw2MJhk4RTsouMxQ0N5881Nu77hzxm+yAj3YayPQ1LLQhrV6RzHAMJaekMdCSN7GjY2c3kLDx4sLOLECRKkmSaWs3SPCHNn4q2hcMRG0d50kCXvcD/8tcON69/yhjPrgqmYaLzJOOsnHEXcn4NKJtBZ3OBTZJZhvXS9yIZcEy4ogPsapGQ6zAPrE8KSHOuTsQGvkuekMNgDD1bAPfXSoyz3CNPy5Ax51P8uXMb3t9kkfP2xPEnm2xaPz1rJc1MOYGBxemNPN3ceIEgxU5ucdOk4E+sSTcxObCSpbQ73DOIQb3Y3iY4Yzn4UUEA99XjwukSgBCYGueSi0e2CoUZTRBZZMLMfFL8g2RNKWJzVh7hnkcrW+tanatQFzEycxianP2PN5ax0xqC1iYFNmarhusCt6U22GfbabJir8eWK/CqAYUJOISx7VjPpW6AMj3ssqf4gpOUQYqCVKL5YZeC9AvxfgeSHMgjvOVKywF3E7PfA62rPJhJQ0yB/OxpWrIOJY3veRjZceeVh3HzLO0RzDofQaFHDUZAM7c+SbRt3+Xg/b/giEO6DCBPmNV5hmXcLSfV9vBpM5cXW8u12kEzkF71b03YLzTQRI0YrLTTSiOMuEAECbGDDHg+EpR5YNwI+jMCqGBzolyCTmbG0OjDz4BPJXbtcyqVK4YtKdqoNU6zusmzbC0zahYH7THyvUATAH3H7S2N8Uq5uduDogPy+Y7Ad7oVbOhvJA+A4mt+srUcF8ymKpqn4NYEQt7+9nD9+9ai2xwJKluQk7cvHKTbruA7zk4+1ruFdYzYYDijNasfgjfAh/CTU83umUFzIJTzFk2xD6Pa55HIqp7OaVWxlK0GCGBhEaMWPjxF0w0JUGUuNXUXa60+R1B6W2mNIYHCguYyASinAdEZSm6xKDufO6HeJ4uUYcw43+R+iiVJKVAOH++vwakeCn3bQeGhUIRKGycYGA9PfgpeIe+08GF5Noj6JHVdY/lzAZWvhIe1w7w7PKD8EvuEq2wDWEPnZA8gJCB+gKQzL1rbpqpN04I5/wr3XQ5+nZIDi4iDHnHAws1aORifCKAWGUmhHURMfytz5YY45bNcD+OcFXwTCfQwxotzLPTTTjGmajAqtZHnzGDxE8Sk/MS0zbncNkAV2b8OLjzDNNNKeZRIhwna29WobWms2z5tH4+bNDD76aAqG9Lx4HByQn2zYloBYMIec0jI827cCmkAsCoMPcBmKnTMME8msy3dzkF8pKdNOy+v5ub3Blh2V7MwrpjDSfh4tJ9bKW6p91P5OIbzZkmanpv42gZvK2lcHquwo7xqvo7UPA0tUv7TDZusD3osP43BvWY/Hlkcel/BtYkRJkCSELJhDGcZ7vMtqVpLEZhCDOIpjet8zNodCchvgYaU9ihmtt1CvC1A4eElwnf86jvfOoSPTU2sp1PY3djLN+wzbnXK+5/s7DbqIo6y1BKgD25VcM0eBrqZWOSSVl7BZTGyYl8SmjfiLHDwYKBJEGzUlIxSWHyTweZBpT1fEOyWGjQE5V0Lop707xz7i9Cnw7GxYuV7eV59PMkK/D5asgVfflufsCsYeMp5ZKxWWZcikjHZbLZbJw09WfhEI+SIQ7nN4n/cJE8aLF4XinPKZPLczyZKmMcS1psxS/LZUek+9gUazgfUs5iNaaGEQg5nIRPLI79XrgwSJkJ1h0kADcWJ46TrNaty8mX+ecgqNmza1pXVjzjqLr95/P0aWEmZv0N+CoKEIlw8jWVyG2dwIhoFXaeJKdcoGLWR5sxGbo30JucEAynHQClRGzdNWJiGnfXPoq7miOPNQg7BhU5WBvw6Aszq8nW/Ft4FyJAi6UMpAo3kn/nGvAmEKPvzt3mEvXo7lOI7iaDQaq6/LSPBCaHqXqDa5svUPxPBSZlQBBhHt4xfRG/iPdRYVRuaNlhRjtZuhTTQ/YISxjgS5DDU24aWGNLszAfYyktZYbFVPzCxlcd43KbgiSdXCe6mtdcj3xXCiHgwTjrw8BDRJ1qqjtA/A7mc052cQvBqiz8k+vEft0UH5CaPh3NPg2r+KAEU8LqXScSMgloCX39r1QFje34ffp1GOiW07WKaBP2ARTUB1ZR2PP5nA6/Vw8okDyc3dzZLJZxRfBMJ9DJvYCNDWfzENzdfKX+PU/i9yOCcw1Ti2V9txcNjBdpawmNWsJoAfH35WsZKP2cBZnEMu3UiYtNtW9jKVRlNNDRVUdPnax6dPp2HjRryhkDDWkkmWP/kk/Q8+mMOvuqpX+++IXBO+mS/kEceXg+nPIeH2Um8vgbvqYWVC7uVzkSQxjjg8bE/C6F3aa++wkY3M4XVqqaGAQqZwPPuxf5fPLyrM58jwAubllVMUCbd1qCI+Pxfktb9RMBTcMQCuKoJZLXIdvpabnTVstPGMO0PtIY6c2WnsII1mG1bERClnTEfxh8A0aPkTC2ODada5lBk1SF5rE1ARGnQ+sxIncKHvkQ4HrrG1okA1MsJYjzK8OFqRp+pIT3tCqrtc6xnK3JzTiFr7gzLwDIV+j32fusfnYi/ewehhAxl3zgkUli+F6OtimUSSNik1bIQbrKHlevlJKc8a+RD8MYR+tOtzgk6D7E8Vg1Kc9SV49EXweaQ36PNKOTQaB2sXbuDCLdDcCtO+lMP1dzbi8QXxmMKITTgWiXiMOfPq+e+HQ0Apcq/bxD035DPt1C7q+J9jfBEI9zFky9Q0GsOA/r3M4lpp5SVeYCc7qEFcyRswyCVECaVEiLCMpRzBkb3anh8/YcIYGQtoqkfV3RJQt2EDtatWtQVBAMOyMCyLRX//+y4HQoCb+4kCzCNN0jMc5IHrS8XtIGSJFmmj6wyhEEeIXENIRh2hNdxVB/9bK+4Q5Rbc2q+zc0JPWMtqnuBxeb8w2M42HuNfTONMxjGuy9f9+cuTufiVRSwLlaK0g1YG59h1XHxk9vdnP5/8dIcpvgrmxE00cZRb3tPaBqU4ziuuHQ4OS/iI1azCT4DJTGZANzc1bYjPg9aHwNkOnskQvKRddvSfJhFDd9ysdbBHxB8GpsrSRgmEfkxrYiG6rfSYutlSaAxadOeLrwAfCSzlUKctAjpBqdqBSQLVtpSlPpke/In1+OxBDIm9Sr1nCFW+MVgDCsi56jhOblnGsJbngZkQLZT+n7k/JFYK29RJAnV09iqMAUXiFtFyq7hJeA/t+Zplwq6C8PWQmA8obA7AyP8fhpQfwLAKmL8EwhHaypi5OXDdD3q/+WgMbn8QXnhDXp+fG+S0Izfw8jv5JJI+lAInEcGp+win/HCCVhw71kRzi+KiaxKsnmzTr2QfK53sZXwRCPcxHM7hLGMJCRJYeNBokiQIEmQsvaOOvc08qqiiyW38KxQahxZacFx2X2/7ewATmMDbvN3GFNTun3wKKMzGFHQRqa3NeresDIN4uDuNxp7hVXBdP/hFqRBVijOEqyf6JVsa7hFX+9RSVm3DhCwB5Noq+H2GOMG6BJy1VZzpv967ew8AXuVVNBov6eZtSqasu0BYUpDHC+dOYcnHW9lS38xBQ8oZWLyLNEEXRYaPL+mpvKZeQaf0MBWMto9ior+YJEke4B/sZEfbiMRSlnAiJ3d/gxR5Bpp/B7oVCEN8kZQLC58HayDz5n/E0/98hkl2Fbkn92Pn6Iksso/hih2FPDMo4+MQvJgJeiAq7JDQfjyqFRDVGQPNYZ6lnXat3P9EnCBb9UA8eAmqBnJVHIVNWxlThUAnyU+u4cv1a3DcgmqT1Z/ni3/IhPBsBre8iRTMPW4mGAc9DNDg2IgijasOQyLjKBx5Ln7QzRB5qm+BUNvQ9ANIbmLesuP405PT2LC9lH5F9Xz3vDDDB4V46S0Rqk+J8ngtCPdhUunW++HZWVBWIplkSwTWNY7l1h9v5PU3NtDSkiAWbeUdJuHRDTStm0ciUgfaptXw8sOfHMETDx7R+x1+DvBFINzHUEY/zmA6M3mZiDunV0gRX+fcXvVikiTZwLq2YJVCKoAliBMhQl/c6I/iWD5mI9VUkySJhUUOORzBkd0SJPoffDCeYBA7HsfypSOQHY8z+LjjOj3f0ZLBBbPMtXUFv9G5NHigD04NwUvhtO1qEvhWPgzpQDCyHbitTv4/8+omgB/t7FsgbKAei/ZsHAuLZprarlt3GD9sIOOHQaMtrNliS7KpmiQsicl1mRig184QpweGMNm+kNnxzSS1zVHegQz3Szl8AfPZyQ48eNrK8EmSvM4sJjCBQNt8XwZ0HFr+BHon6BY0JhoHlVyFbrqKFf89h9n/extHHtTMKRdsw8QhUR+kvnw/bohdx+stUzkxg5dRZrbw/eA8/ho5F+U0u3mXyamemUw0FpG+hUlLlykg14gyzqzDMYcSUJNRiTmk5c19Minv9rVNCnAMm6STIC+5g1Pqn6A0sRqTgMxPAJAr5sA4Mgqhw0gJtKPSTepYbDddU3TpPOHEoOUuiD0lxxW4UHqjiQ8huYkFa45mxl8uozi3nqmHziae0Nx2X38WrwuRTNGCke9EUxj+/E+YdlLP73ljM7w4B8qKoTUiTNTcHGg1Fetrh/HEw/K9v/SHS3lno0HTxjkkW927QGWAneDpx97k9Qv6c8IJvV8jPuv4IhDugxjLOEYyihpq8OKhkKJOA8zdIUGSBuqJuxT1VEBUKBz3z3jG93p7OeTwDc5nIQvYxEb8+DmICYxmTLevMy2LE268kddmzCAeDqMMA8dx8BcWcuLv0+aj2tXe/Fs9rI7JzOTBPhk5OCTL7Q32wQAAIABJREFUetwTlILflord0SthGRg/NSTC2x2x05b9dfwimIjPYUfEHHi7VbLL0T4Juqmg7cFDkmQ7goqD0yZR1hO0hn80wH31sk1bQ54BDXZaVafQhDv6wwG95DSUmX6+Eeg81rCc5QDtPlepOdE1rOEgJnTemFMN9k5wWnCUl5iKS5VAK1TsFebfuw7fwCKmXrKOSNRHMm5iNsRwcmL8svBa7q8/mBNDmSQdPxf6Z3Kwp5ZXWguJYnCiZw5HmO+gVMoFstzN8LYjZckYSuVT6BnoxiWvuL87UXBcE16V72asuSjDwIuBNjzgeBgQXyG7Njp6dbn76/cxtPwDwreB3px+nBbSE5weiTAA/umdr5N2XTcSHyCfJA2JBRB7CQKXgoK/PTeVQ0a8z5XTbsMyRfz73OMf5lf/uJ7/fjAV02iTaCVpw+LVssue+GX1TRCLw4crpbfoOPJZGjxAGKkr18MBQ+GML5fx6Is12BH3w2bmgJULKGwnyrcuW8CqxcPI/X9ij/hFINxNtDrw450y7JzQcGhAFqqRu0m+srDoT9/ljxwcokTagmAmNJo4cfrTnyB9+4TnkccJnNjn4znk29+mZORI3r39dpq2bmXwscdy9M9/TqhfuiH/RBPcWCM6qilxq9cjcORG6S1d3nX1tUsYCo7NkZ/uUGSkJ8Uy1xgNdJxO2RwXfc/1cckYfQq+lAOPDJQs7SAO5n3ew3D/ODgkSXIQE3oVCOe0wl/rpJcZNMTy6d0I9DNhrLtu19swoxKeH9Recq6v6C477ZjVtkEVAC1oVDoIojAUJBMKq6Ka0eUeDNMhmRC9PG0ojLownkKTkP02jp7eVsLGexQoP+OtZYzPMyDxPjJkn8rhFbANtAmqH6gSkTgzR0t26sSBZvBMgLx7IbkQ7C1glEPdKWRqiLrUM/dfOktUSYocnPJD6AcQ/BZUjwWnlrSOT6r8Ggdi4PsyeLOkadF/Q2IRKdcNQPYXewUC54LW7KzN4eZv/5yk7SEczaWpJYDC4fqLfsuSDQdRVd8frcE0XRqQgtoGKO3hu1BWBFt2SjCMJyU51lpmE6tq4aJfQnEB3DSjjNHDNvHhRhvMAFh5oB2U0himh+rKOn53N9zyk+7393lBn+ljSqmTlVL3KqUmuP++bM8f1mcHUzfBQ43iVq6Aea1w3EaR1Po0sJGPMbNIYKXgwUMtNTzH0yTa9T72HoYccwznPvssly1cyCl/+EO7IOhoYX9ujqeXvtSRxxF/wb744/UVKaeIVBEsJdblAKd1GK+6aDusjotAdsjtPb4QhtuEj8SXmMpYxmFjEyeBjc0BjOQ0vtLjcWgN/1MlNlvLYjA/khZY35KEOWFY1AoJByqTsHw3ZLcAJnIIIJJpKSRI4MHLAV0Nxhs54DkUTUwWTRRK26Btok6QaFjjsyJtBUTtnljS48WnNAFly3urNcTegKYfgVEKugloFZFqcwwYqf5orht8TdCVEPg6BH8lAS/xBiTngL0Rgt8Xx3jvERA4B3xHS5bYSQKwFTzjwXsc0CKEGEeLQ7xSkPurjHPNg+K3JdgpjwRh8zAwDxGCUP7dUPDv7DX86AvuNjKWV8P1jUl8CP5TmXroTJRKEkv4SCQNbNsingxgmQmOPnCufA61DNYrJaXOUC+qIys3yPNaIzKHmLSl/A/Q3AIlhdIz/OENisf/Nha/zwBT7rJMEywL0EmCoSBz3oe6hp73+XnArmSElwCXA79WShVBthrK/w+81wqLorIopr4PISXkjdtr4OZPQc+2mWa3PCcZSUetSJAh+Saa+ZgNHMDIPbLfqiS80Cx0+Wm57XU0u0NEC7uzuQsN0GYNy6NwzF4s0bw6GA77WES2U4JfB/vgoQwCZXVCJNIyndgtwNJwfwP8shQMDM7kbFpooZpqSihpG0LvCS+FxVEkZfekNdTrNFexFTG93R4DH3BjNfx5gFg77QrGcSDrWMdy0qQUC4uvcSaerjJCgNxbSCbex3BqQRu02hY1TgUrnDF80NpC/J1NjJ+q8JgOTkJUTLxlRTjEGJ5zGAoNDZdB9DHaicX5ToXQj8EcAdXDgRAY7smpPDHBjT4Dzk7kKhXK73QL1J8Dpeszen5A/l2SFTph2np7KiiPmyOh8TsQexF0BFQ/dPACwtFHqYn+mSr/JMq85zLUGo4q+k/fL67RTdpmFEHOVZx23P8SbVTEkx5sx4uDlCaUAp8nfZdjmBAKwPmnQ6BjNTcLdlRBVV126fJoHLZVwqABsLMGVm/J4YILDuQfD69HKxtDKRzbxlQJho09EqWgqQWKejmz/FnGrnyNmrXWDcBPlFL/C/SRO/z5wYKIm8F0uClUwMIeXA72Fkopw8RqK1ul+oJyXJn/1tRRt0f2eX2VEE5SgtZXGvCvivZ6ml0hqGRAPpskWiqzyN3LTO4SV9btnVaRdZsQEOZp5vva4galjuVIEzH0zUSO+6cjtiTgpWaotGGyH07ISd8wPNwo16JK1NA6+TBk7sJGbJcu3wFPDuyb40QKBgZf40yO5Ch3ztTHQUzARw+rrWcE0YJ7aWq+inCLIqKCrLRH8evW31NzQz6b/v4HRvznccadUY2TA7GyAahAjMXmlZxXNAjicyH6OOBLBzrHhvhM0D+RPqROZunh+ST7Q7XX9NQ54NQIKSXwdSmZxhcBESiaC5H7xD/QOlBm/sxyeV3hP2W/uhXdejtN0X/SpGJYwKD4e6z1/5cFoe9xpjq723lJtAb7Y7dkO0IyyeDlEHlYSreGV56jo65E27cAh+GDgiTzthCJbWfhmoN5Zu5ZbKupYMyQ5by7PC2rpx0491T40UW9eFORjDDeuSvSdqibd8Cg/pIpNoXh7rtPY2PVW7wxazUGDj6fyZAxh1E4cByGAQP7MFK4vUp+xu0P/l4E7X0JuxIIX0r9j9b6GqXUlXvweD5TOMDrloB0+0VTu7/7NDCQgQxmMMtZho3dqUSa8pWLYFOYuqveDbzVArfUSpZiuYt6xIHztsLmA8Q4tjsoJQPir7dIdpha/FNHnW+0d3PoiKgDP64UQ9+YhvGuYfHEPjpEKQVH5chPNgz2QIEpg+KZ5xQHDu3Fl35+q/T26m2oSUimGzBEGODqYtgQE7Yo9GxGpIB+HtiagPcjcOQuEIpS6O/+6QvyfKfy5PoqHjCGUqX7s9UZhKE02tK8cdn/UFB5NneWVUH/bcSUSYH/WM4LyOwi0ScAJx0EQdIeR0uWmHsTKF86iLQhhnzKOqzySrktxQ0Q/xAazgN7G8LsDIkxbv7t2U/EMCGxlXjsGaoMA1RQvi9aMyK2jhX+eazyjGNsV6MvTh00/RSSS2nrPwYvh+AFkHs7NP9cMllws9F/0BIro/Ljn5Lvm008Ppx1W3P58V03E0968ZhxFqyeRE1jGV6PlDgHDYDxB4jUWm/wxnwpb8azdD0MJQEwlpDv6kGjwDAUjz94HOf/5Ai2bG0hryCHpOMl3Ao3znBLpT2grgHOvBoWuMUFnxd+din8/Nu9O+Z9AT2eplLqIeA7Wus4gNb6uczfa63/speObZ/HyTkw1CNOCSlrzogW4sRPPgFB7GwwMDiD6RRTzFzmYmd4q3nwkE8+ESKECDGM4bu9v7vdMoyVERwChpTxnmwSL7+OiNTXU7d2Lf7CQor235+TQ4q7+sNlO9ITWymyyqMV3ZNCztgCc1uEuGIq8Rv80iZYMHz3nCY6wlBwfQnMqBKCVIpgEzLgxtLuX5vU8NtquUHYnoBWd9A85ogKzj8ahGiV0jTpiFRmnPp/D/JZc+hbLzruwGNNkvWO98HX83pmIXaF4NoBrCgbTTQUwtI2KqlRCQcMzVsDRpI3eiJeBZ2/BllWdCcJJCC5SX7vvxgi97jMTFf7UynwfU2yycw7T8ftRnoOgfqzwKkEFQAMycKafwbeCdI/zIbkUqK6RcTaU7dfSmE4DmXJShZ7Puw6EDb/BpJLXCKPcsdL7gDrAMj5jgh0x2bKOfi+zIZtPq6/Yy0/PXMODzz/df495wxqGwsxDIdQoJVtzRU0hKWs6jgwcYz0+jb2fuQXj0dmB5N2mtiagmFIhlldD1+ZAmNdwaPCfHjsD16emeXl3Y9gQCmcPRXGjujdPqddCQuXiW2UYUhGev3dMHwgnH1K74/900RvMsItwLtKqTO11htTDyqlxgNXa60v2VsHt6/DMGD2ULhgq7D7HMSJ/K4BsN+nlBGCBLzjOZGjOZalLGEjH9NII620YmExnOEczpHtBr93FY3dGIo3diC5aK358L77WHTffWAYaNumbNw4vnTbbVxUVMSpuUI8mR+BMV64tkwyn66wPCrkpBwjPUwfUlKqvKkG/l7e+/OY3wp31wtBZaQPflEMoztkepcWSRn3ljrYkRDpsN+Wdi0OnsLmhGSClcl0wEvFdge5eUrBpn3go8O/NRIEE65izv69fAs3xYXEVZ3xnlxbDXOHQdku1IUmjNyPeJMfK5LE0OkPgUbh93qoTkJFtvcucBG0/h2chKvt2QQpZ/v4K1C9PxS9BGY+tN4rM33GEMj9PfhOgvhscKpA+2kbbvdMlHKqUymZVypIKr/0CMN/gqIuAqHKIxtnUGMQUz40XdyF2ZUyEqHKABvsahmw10lR3fEeLiXcwNmyPQ2/+QuUBD9m3fYDeOKN6YT8zawL7w9aU9sEuAxcw5SApZRkbxP6oAn4leNh1rviZtHiEnA1slaFcmDsfnD9lXDcoe2rWPm5cNF0+ekLNm6FRSvSQRDA6xVBgNsf/BwFQq31r5VS7wGzlFJXIbdoVyMyjn/ey8e3z6O/Ba8NlRJd1BFj1n0BNjZNNDKCEW0Mwcx5wj2Fr+bCm63tb9KT7iJ9Rgcp081z57LwnnsIlpZiejxoralaupQ5117Ll++4gzILbulDlW5hVPZjdDgdE/go0vvtzA7D93ZIoFIII/PFZum/ndShz3lanvz0BQElS3bYaVvys5Y+U6eR0jJJPafjvUYYeDsClxaIRVVvcNE2mYtMEbtSbNTLd8B/dkE7evTI/dhv7hZWhUrwJuIorbFNC59lUur3UtBVW817IORcA63/6w6up2oAOUBQgln9WVC2HEK/xlWJTX+4it+Bph9D/A1hc/rOgrwbIfKY/L5Dw76u3sf2yqUMqDie4pJSCF4KvhNps4XyHo3fKMHvbCNm+FEofE6UuOFhi3cwp3fFBdQtkFKGTXzkzi2K7yHRxyFxBXjS6kBVtbB2IxSMHsDri47BUA4ey0YpB0crTEOTtE20dlCOvEnV9VBRBqd11p7oEhd8VZztl60VRZp4QjLEnABc8jX43ZUQ7GPboDtsdn2JO1YWLBMqa7O+ZJ9Eb5ftt4BXgBeAKuAcrfVbe+2oPoPIpnDyaWED63mTOcSJoYFyKjiJkwlmUwvZTVxaAA80COVfZfT4vlvQWcVl2RNPYPn9mK6xmlKKnLIytrz7Lq01NQRL+lZPHuvL3qO16f0cp9bw+xrYmZSSdkyLk3xEwze2weYRPfc5e8IAj7jar4mnCUXZ4I5et+Xp2TgPCvnSJjWcndc7BR7HgQVRcapoS5YUBLT0ZncFSikenDSI6eujNODFUVBqKfKDfr6R34NhdN6vIPgNqDkCdAOoXFBu5HRywNkIiWXgGQcdyTtWORQ91nmbvuPderUNholtK/54z2j+/cL+GAY4jsEZU9fx0x98C0/eeZB3u+zTCOHLu5ec5u9gOpsBTYuRw+u5JzPCOJhRXUm0m4Mlm0yuc7PaVI1aS1CsPxtKl7QF3NTNx6aqcfjZit8TxVCa4rxaqhtKcRyNdjToBLY2yfOFOf+0Ai6arsjrg0uS1wvP/kVk1l57Wx6bNA5mXADjurGL3FVMHCNtkURSXDNSiCdh0u6pBH6i6E2P8C7gNOAxRLj/t8APlVILtdZ9UMD7Ap8EaqjmGf5DCy1opEwaJ8ZrvMIZTN+j2SCAx4C5Q0UV5ulmyTi+WwSnZzG2iDU1YXRwF1WGgWEYJFp7/ihtSUCtLRqieSZMCsA4n4wdBJBA0qplzu/nvYypYUcG5A0lfc1UadJA5he/tx0eGpj9tTt3rGX9mgfwqqW0xIZRUnEJ48ZkzyBuKBM3hncjXRNhUqa7KSOg1Exl5vPlPZXfr4nD0b0YK+myeq26J+X0hIlBxaPDA9xWC1uT4v5xXr74JrY/gDpILBYSjOcQ+dsaDkYu2NF0EHSPCa3kNX2BNQx8Z6IjTzF7zjBuv+cQNm8PMai8EY9lYTsG/3l5NKXFUb5z4UuQOAu8R8trPWMpLpxHvb2A1ayixezPaWoMAxnY9fdFWRD6DdSdjtyypCK/BeSCvUmyVt/JgMwBjh0BqzYoLPM46sINDA5sYL/ydURiPppapZluGCYFvmoGMZMJg6dQUrhf366Du69bfwo3Xi0BuDeEl11FKAcu/zrc8S+ZW7RMyUKDfinBflbQm3vdxcAorfU1WuvVWuvzgHeB95RSe+EeozOUUqcopVYrpdYppa75JPb5WcXTPEUjjdg4ONjEiFJPPVvYQj31e2WfXgOuLIY3hsILQ7IHQYDhJ55IvLm9NmOsuZlAcTF5A7uINkCTDT/YAWdukVLe1M3wQL3cYb86GE4OSs+sRYvDwRMVnd3au0LAcHtuTlo3JLX0mcD7UViXJTXbuWMV1RsvIs87G0NFKMpZCE2X8d6CeVn3U2bB3CFwvJ+sZHwvUjZVSOZmIoxZdww7PaDuPs8BBvZygbNc5m1rh6gX0XDEbpbJjsqBpwfB60Pk/b+8qMM4R+TfUHcqNF8DTTOg7jRIiLybqLIk5I1MQceF7OI5rN1+onaMOVUP8d7WM3h/26nMqbqT5kQHnc+C+7n1H7/m5zeeyNJVJeTnxlBKkbQ1pqkpKozy6LOj0U5YHDQyoQwKrcM43LqQE9VUBjGo6yCotatgUwrW/kjw84LKkSxRASiIL2z3suuvgOJ8iMaCROLlLFhzFAtWH0lzYxKVrKfMWsKxg2dy5MA3ycs1eOSRzsLjmYdwfz1MWg/7r4VvbxPD6kx4PHs3CKZw04/hjl8JOSYnCCcdAW8+DKP7HsM/NfSmR/i3LI/drpT6EHgZujFb2wNQSpnAncDJwFZggVLqea31ir25388immhiO9tF9srNJ0QpxSFMM1H60DjbCxh95pmsnTmT+nXrMH0+nEQCZVmcdPPNqG7oizfUCJmlvyXBan1cRhGeaYYZxfDsYNHljCIszr7AUnBuPvyhtj0hRSMZTkjBx/HOpJQt6+/Gb9hEEqKdaTsBfFYTifAfSSaOwvJ0XkQNA54eAkduEKYxSFfJQEqJliMD8v09sCMpGqNe0kW3VHbouMeWlYzSBW4sg3O2Sgas3W0Vm3DPAPc4tPRI842+z20q1cVrkmsgfCsYhaDcC+g0QtPVQogJXSeyY06VyKjhgDIg9xaxQnKhHc07lT9jgDOXCAVobVAWv5+l1e8wuf+DWIZciE2bm/n38yX0qxjMtp0b8HpFFsVxNI7WeD02dfUBwHIVa3YBsXeg6buQ3CxkHG0ik/A56fpnSu/UbN/wHlwOT/8F3v1I+mfhFsWyNQ5PPLaWEf1qyQuk69Rer0lNTdd166t2CtvYQH4ea4JXWmDhcPmefBJ4fwk8/rKIes+4EC4985PZ797ALl8yrfXrSqnj9+TBdIHJwDqt9QYApdTjwBnAF4GwAzazGWhPhhGavyZBguIsZHaAOHE+ZgONNFBMCUMY2nfX8V7Al5vLtAceYP2rr7J1/nxyy8sZecYZFAwZ0uVrGm3pY9nIzFxtxkL+SljKjd8vhB+VkFXDZQc7WMj7tNLKSEYxjgM7ndv1pTCzGZbF06VRDzDAkJ5heZZLEbAW0xprv5jGkrmU5G2kpj5G/7LsKWm+CT8uhj/USXk3R0lv0tHSx+vvkQDlUcI0TWmgKuTYLMRbsb/VPWO37fwT8KsqKR8PtsSPcaRXSqojfVKq/Xs9PNEoKp+5BpyXJ6Xl3qoDdYnoq4iiS8ZdhJEvQ/OJj8A7GUo/hPAfhQ1qlENoBnjb20Ctjqygn/MOzaqcVL4exk++XsOK8LuMzxOz6hUrqkU/2jQI5uRRVZPLkIG1KCX9t/pGP5MO2okyi8DXBzqj1hB9FsI3QPJD9xhykZJoErBJO1YABMGsAN/UTpvyeWHK5MxNW2xauJXa2ggE0sG/sTHG9Omjsh5OZQIebJSboczsu84WxaE7BvT+1HYFWsNXvg+vzEuPaNx0L/zh53DF+Xt333sLu7Xaaa237KkD6QYVyAhHCluBwzo+ydU8vQxg8ODBn8Bh7XsIEsDExHYLo5kBsZgSPHioogqbJKWUYWHRQAP/5nGaXeNdH14KKeYMpnVrsZRCkiRb2Mx2tpNDkP05oFtZMU8wyKjp0xk1vXc87bAtg+NhR0qfmVmbhVgUPdgodkkdM6QFvM9rvIrjCgusZhULeJ+LuKSdjJjXkLLeyZtgW1wCQquGdTZYNvy0Ep4aCMGMrCeaKMdjbCPhpLdjGjFaovkMCXU/0zAtHx5vlkBnGFKibLBFiWdNDLQhrhPjfGn2a4EB5R4JygZQY8PwbnajNfytDn5WJdsPKRjlg/4KtiVF4eapJsk8tyXlWgaVXO8/10ly85se5iN7RgS6Ki+mqEBGIeRdD1zf5VbqYxsQom7mthRKOzTHVgMSCAsLA21koKFDC1ixXLL1gvww8bhJMJDkR5dvh9zbwBra+9MI/x7CN5MuSgM0gi6QIK8VGBVuU9cHqghyf5dWsekGSimuueZofvjDmVRWJggEPLS0xCksDHDxxdn7zXMjcuNkdbhRsRAG997GvU/BzLnuDaMbQWwbZtwMXzsJyj+DBvf7CNl/96G1/jvwd4BJkybtDgdgn4aNzXKWUkkVIxnFYNJBfxCDKaCQeurdTFC+tCYWR3E0j/MvmgmjUJiYHMtxvMor1FPnuiMookSwcfiAhRzNMd0ey3rW8yLPUU992748eDiSozmBE7P2WOqopZZacsihPwN6dGVocITEkqk6k0ICMDU0JiUzzAyEUaLM4r+QYZSr0exgB+/zHkd1OLciC54ZDGdvkcxMIYHBi7hCXLodHssYMwiVXopu/BlOzIOtAxhGnIC3jk0NVxEMdn9ORSY8UC4EozdbJeidnwdHBuBP9aJf62gh8EwMiFBAkwOFhozoNGk4J6/7Etivq6TcG0fOpUHD/CgMNuX1RaZkhEui7fuQftf14p+N8KPiHtifPcF7HESedIfi3A3pKGCAdVCvNxOy+pEtoGoMAlY62EyaVE6/fiGqqlooKQky7sCBbNyYQ019I9++pJjzvzGKiqFHgvZB5DlILgZrNPinyShGNthV0HInslSmhl9SReowUCCP+6ZC6ErQMbBGdL29LJg8uYJHHpnOE08sZ+PGBg45pJyzzhpNaWl2JlSFmZ0t7fDJlEXvfBTQYGbsyzSFOXrbA/CHzyCL47MQCLcBmZNOA93H/t+hkkru42/EiKFQzOMtSinjcn6AiYkHD2dyFs/wNI00orDx4OE4jmcJi4kRbdPATJDgBZ6nmaa2YGS6Qt2ttLCOtRzNMWg0tdQQIUIxJQQJ4uAwl7eYy1vE2xYHKckmSfIO86igoh313MbmOXs2M6OVmEYDKE0smcuUQCEnmZOoIDtZZl1CGKHZ/Ow1knNstSUQZmqbbmQDSRJ48GC7f1KvWsayToEQREatMgm5qv0ojHLg5bAowaTKhaNGnsCiRb/GSNyJ36okYfvZ0ng5J075Zg/voqDCA9eXSen329vhBzvT3MODfHBRgRzP8TlSfHuoHma1QrEBV+V1ntHMRL0N9zakSTkpBqoGNtnyeK4hsm6ZPKCIlmEFS0mG2mTvZiD0TAb/6RB9Mf2YMiD0W2GM9hKjQ4cwv3kYBWyghVI0iqCuIaKKGZc3pe15lmVw552n8otfzGbNmlqUgjFjyvj9789l/Hg3TbFroXaSq1vqwiyHotlgZakkJZfLzGBbPzBTykecN0CBdwpYu84OGTGimF//+thePfcIV9FqY0KmL5US1SCAq3fBsqyviHXhfKKAuqa9v/+9gc9CIFwAjFBKDUMC4LnAeZ/uIX06eIj724Jgyl2iikqe5inO5uuAzAx+l8vZznYcbAZQThVVLOajdkLQHjw009SujJoggRcvcRJoNM008xRPUEkVCkUAP4dzJIUUsZgPcdqCSxoKhY3NAt5vFwiX6KW8FNuMz2wh7nhAg99q5M1YnObAC0xVJzIii/1PiQG9+W79rgbOzUsr0RhuGEi4dkgpCHEoW1gVNDlZTHqVmPfW29A/FRiUYuIh00kmvkJtfQO5oTwmBPum1KM1XLwdnmuWRSQ1R/hRTGYzF2WsqzNKYEYvt7sxLlmd15W665hL2UhJtKOmaRI5T62ll1nah9Wh2ZYS9c4kHB90hQiUkqDnO11YmipHxgn6UpYELMNkVOmdLK29hRJnDgpNnXkoIwp/gd9snzUNGpTPI49MZ/v2ZhIJhyFD8lGZaVPjZSKSTUhSbq3B3g4Nl0DJrM47N/Lc7M4RNqtO0H4gpVWUbQLpnmP1ypWseeEFog0NDD3+eIZOmdI2O9sGrSH6PET+IY4a1oGQc5XYRCHSajfdC8/MkvLjt74KV54vTFCAlwaLvOC6uMzvehVcVwJf7v39xS7jnFPg93/roHTnKuF849S9v/+9gX0+EGqtk0qpK4BXkbXifq318k/5sD5x1FBNCy1tQRBEV9TGYTWr2j3XwmpXMo0RazOJbaaJGHEMN2B1LF/GiGFicgAjeZgHqKUOUWFUNNHEm7xBBRWYWT46mco1Edrbb7wQW4fPaCXheNtW5oT24TOibI2W8E5gHsPZr5PSv8+t26Vm+zqGXhMoULLwP9QAP3P7WsPZDwuLCBEM9/hTAV/jUE99VtHxUW65MPPsYkhZsjjLt8XyWPRnEpUSAAAgAElEQVQr2zVh2bVxeLMlfR4p2Ii6zfsRmLwL4w2llpBtEu56na1PkBoVyRzLUEgv1oNQQQ7ZIJqkvygVObmuMK8Vpm8RDVaAP9WKu8ZrQ8BjKPBOkp/dQLG3hCkDbiHuRHG0zViz6wFKpRQVFVnkf3RcZvsIpOWIlAIdEI9Dp0kCXyasg8DcH5Ir5HUq1yXG2EAOBL4JebdKxgisfOYZ5t10E2iNYVmse/VVBh1xBFP/+Mf2wTDyKLT8wR23KIXkKgnSBQ8Sjo7igNPEKkkhZcdf/hlmvwsvuRz+oV5YvJ/IDFYl4bCgmDl/Evj19+CR52HTjvRjCjj6EDjx8E/mGPY09hEtlO6htX5Za32A1no/rfUNn/bxfBpopSWrt6CwCbumD27kY95mLlVUsoH1VFNNE400INRy7f5x3D8aTZAgJRRTRx1ePFhYmJhYeIgSo5IqvHi7tKdRKEbQXrG30bFBabTq+FyHiDaJEqW1k5GqEGWGerPfsVkIw9JQch22JDN/ZzGBg93Ap9t6mAEC5BBiJzuybBFuLpPSYIsjjNFWR/p115RIcMlEzIHnGuHibXDldnivjyotNbbso2PGlipjrtxFK6+BHjg20NmWNrXtYMY+jA6/S7FT1ydlbOTpZjh+o4yvZIPjwLe2yY1IyJCfHCX9yJv3gsSW1/B3ygJ7j5T1ckeorn+nLCj8txj9qqhoiaocCFwO/aqg4C4xLEZmYt+59Vb8BQWE+vcnWFJC7oABbH3vPTa9lSHEpeOit2oUix6pMto8DN9+93WGT4Ud1ZJxpXqBThLeWADzFrU/vLF+OD70yQVBEPWata/AL74DBwyB0cOFMfrafZ/M3OLewGciEH4BqGCQ6zPYHhpNKdnpfTvZySvMJEKkLRD0BhYWa1gDtB/FSIdLmzhxiinJSnbJp4AjaE+B38/o3z79aNumIdJfGPiyOBNUeKSfdVgAysy0d4Ei3b9KIKWhIztkT0MZRimlFFFMPvmUU0GFqxbizeaCgBgAzxwMhwdkv6O8cN8A+EGH3ktcw/nb4JvbhFhyVwMcvQmO+ljmHHuD/b2iQ9r5PZUvZk9i3tngaBms/mYB5Kl0kFVIrzWVWZe71zIzCLuTfDI/aYjYQMiQLPHnldn394GbkQQzNqSU3KQ82tj3499l6Bgk1wq5pSsov0vSiXa46K1gjpbRjmywhkDJfCieD4XPQdlmNwC2H5GpWbkS7ThYvvRnSymF6fGwcc6c9BOdBjkG1f4zuKVqGD+8/WTqm1KvBbTcbGggmYRZ73R/GToi3AJ/ehAmnwOHng23/gOie8Ar1bLghqth1cuw/AW46gIZDcmE1u21EvZlfEbj9/8/mJicwPHMYla7npeJyXSyT7Iu5kMUihixbh3rgbbMyYuPKFEWsQgHu+2xOHEct5+YIIlDCwECFFNCM80kiBMixDjGM4UpnQxeL/BN5ketG/CaTcQd+dh5zARJO4cBgVrGMD6rG8ZEP4zwwuq4lOlsDe9FZbwhNUSvtbArT+vQHxnCUPwEREjZDXxRIvjwtSsdd8QRQZg1tMtfA/DvRuntZfbZFPBeBC7cBs8OgpIevl1llsiR3VIr/bmUtRKI4sv4Ppqbvtsq4gPVScmkW7RkZzHtlpSVsGwLTSk5t3aoM2cKfWfCC6yIS/D3dkhfM500MqGQ96cnzGmB55ogYMKF+b3XiG2H1oeg+TrQjRJcvCdD/l86lzkB8u+GuhNlsL/tjQtB/t+734dS4BkjP13AEwyiO6z8tjZYHD2U55aezV+uhCmHwmVnFRDAL8E7Ixi+9O4EIrE8EbBOvTcqnRmiRT6tt6iqhZMvhWXr0gFp0Uq47z+w4AnI20v9xDffh4t/BZt3gmnAkRPgqT9D8T7sdP9FIPwM4Rim0I/+zGYWzTRTTgWncjpFZKeK1VOPBw9hmtsc66WvaLcFvhRS/58gjo3h/rHayCap15uYDGAAceJYmIxhLEFy2I/9uxX1DqkA1/rP5o7ImzRZH6NxSCSKGB2Ic7Aaz+Fkt8kxFdw5AP5YC6+2yOJ6WYEombwcFkWUabki7ZXXoVIbJMipnMZ/eY0WpG6ZQw5TOaXdHOGu4Prq9vzBzGCwNQHPN8Mlbgsy7sBbrZL9HRFor9R/Qxn0M+HGWhmIDio4J1fOuS/YEIcZO4XtWmLCsqgEtpRdkw/J+Co8MoKxKYtxa+rTENG0eydTWWK2K3aYX8Y/wo6cH8iimwCmd7PQOo5k1M+H0wIJf6mFm0rhir60XKMvQ9MVMsunPOC0QvQpCTJFj3d+vmc0lCyF1rsgsVRKnqErwNh9A9HSMWPIq6ggvGMHgeJiQPFk1Tm8FzuW1s3l6C3w8lvw2zu9/M+lt3HZ1BmEQkkgCLqBqvpC/IE8fB7R60yRUbR7vYJ+cZfIRDwO73wEG7bA4AHSp0uZ+N7xT1i+njZrJ5B5vw1bhIhz0492+5Q7YdV6+NJ3ZJTCNGR/by6EQ86EDf/ddf/LvQ3V8Q7m84BJkybphQsX9vzEzzle4xUWsgAbmyTJLrPBFEzMtp6hgYGJSTyDYG9hUUEFPvxoNK20ci7fIJ++3eo5WmNjE1Wt+PB3ygTjxPmID1nNKjSakYziYCZiaS8OHbQsewEbmxpqMFBdlnPDToIHwtt5tcUmz9CclxfitEBZ1lnImiSMXAd1WRiZAEMskW27qR/8s0ECVGoOssSExwdK1rkncVuNGCH3s2BFFDZkMexVwGivBMuPY1Cv04+nyrGpjLTYkJsQW0v2PaMIbuhiUPrFZikRJ3Q6qI3wwtvDuh6/eLpJeotBleatxB1wFKzdvw8eiTVHiaC3kdE3dJIy81KyCqxPdrq7cfNmXpkxg6YtW9iu9+feyHU0eobg8Ynre6rM6fdppkzcwUPXzKAkbx1YY3ltyW/45V9G0BCG7VUQyShhej3wxO0w7aT0Yw1N8N1rYdUG8S30eWFIOdx3vWSOB34VVmwQIexMJJMwchiseJE9ji9fBq++3d6JQmsJiM/+VfwSP00opT7QWndibu2j8fkL7C5Sii8J1++tp8H1FDLJMxYWOeS4uaEHEwurXV6gswaKnmAohUdZ5JLXKQg6OLzMi3zAApIksbFZxAe8xAugnD4HQZAA349+lFKW9Tq06iTTd27l1poAq6MB3m3J4Qc74HcNG7NuL+y4JB3atz1Tf4cMsYhaHoXv75QAkeOSSOpsmLZFsrI9iS2ujVTYEaWYbNAIkcarpBSZIsdk/t5AssGYFsJQVMMZIbi2G5WZ03Nh6XD4YaG43t/ZHz7oJggC/MvtH2Z6SXoNCbxP9qW3mFwHJMCpl3Kn4w7skwBnex82tGeQP3gw5zz1FNMefJCyC3+HUbofpsfbFtRMMyVJqli/rZz75zwBJQuh8GGmHDGCAw8Q7c4BpZAblAA4rAIOHAE/ux3O/ymsEyVF/voo/PcdWPmxZHlrPob5S2SoHSDQVWldde7n3f0YDD0JQpNg6Ilw16O7dv7L1mbZnfsev7UP5yZflEY/h6ihmkd4qM1tIlUK9eBpC4wmZqeeYUfjXimfynOVmzdEiBAiRIQIRZSQSx9daoG6devaWHRDjj2Wov3Tuu072E4lOwmR23YcFhaVVLKNbQxqp62wZ/BkSyVrojkUmbF25sIP1AW5NLeVCrN9+jbQI0zWejutfZq6irmuFNrpufCLSlnYU36GSkkG1OIIkeSSztMbXSLqwCvNMlRfYcFZeTAio582OQBvt8o8X8cYm8r4QI5nU1zmJbORdCzgln5waEAGticH5Fx7wiAv3NQHU+Xu7mc6Gi13ieRGhBubJE31aUEabCZY6bnUyhp4Yqb0yIZVwDdOgwOG9v54+wKlFKVjxjCsFnjGJbrY6bKgQj4LOUEZifjZpQrbhvVb4KpvwZpN8nhujmSFDz0HVfVSanx2tjjQz31EFF4awy5rWsk+autFCPumGfDtM+GjVZKNmW5WmLRlOxdnKBze9Sj89DZRign4oL5Z/m07cGXv9CHaMGiAZLOZSPU4D+m6vfqp44tA+AlgO9t4n/nUUE0+BRzKZIYyDJCB71QZck8gSZJH+VcnyyWNJkkSDx6SJLskzphYJEm0BUoHh3zycdA000SEiMtADHISJ/c6I9RaY8difPDkIyz6610YtkKhWHjPPUy+4goOuuACAKqdeppxMLQi4F6S1AxgA/V7JRDOicjsZObctaVAKc38WBNfC7YPhJaCX5WIBml+UobIYwiF/bICuLoYCkyotLOTT0Be01u0OjBts4wkpAgod9XB38rhVLcP99WQEHg+inYdZAyEdORR2Q2CPcAoD1xUKNnc5M5P6TPqkvCdHTDb7e8e5BfHi2/mw8ywPGZmlEZNJfJxvULkEVD9QW8kLU8OEAHfxW3l0m2VcME1EjRyArBivWhl3vFLmDx+988xEknwwAMfsWpVDaNHl3DxxQfj91sccwj0L4bahjSDMtX385iSseUEYc58uOxamRs0DSlvPngjjBoGFccLQ7Mtg/NCcytc80eornP7fxk3WrYtjysFF02HNz+AJ2dKz04hzz312P9j77zD5CiPrf/rMHk2J2mlVUQClAgCiRxENDYZYww2TmQwcG0csDHgjOGCTTDGGGzAYLjG5ByMhABJSAihjHJkgzbv7OTufr8/qnvCBgW4sn2/R7XPPpJGPd09PTPv6VN16hRcWWBJ8hu37cGrLQYDIkK79cFdB8JfXgMnXiTHMw03LerI+Kkv/Qc32+8Bwt0cTTTyAs+hY+DHTyedvMgLNDCCbTSTJEmQIBOYxHQO+cwijq1soZOOfmIYEDD0GOFA4eCgo+fOQUOnkgpKKMmB6HQOIUkCA5ME8QGb0gsj2dHB4sceY+ULz9ByfIT0X+dj1JRiRkJUUEEoG2L+Pfcw6thjeSjUwAPxKIfU6iQsEWEcFBKFo47OlnSUhzqh1YajwsKKyv8X7h+qDQ27D3zIoqVRqQ98gKMj8NgweComLi2HhoQFFo6BOiECr/X2ceBQ+f/b2bivA95LimAlqMs+uhy4oglWRyXVWWLAX4bBve3w3x3STO+lPr26nQ9J6Q73waIUpK28/ZqJ6zhjw1VN4jM6eRdVq33DceCYTeJ+EkJuIBalYMYmWDIaziqVWqH3MTWA22p2rLbNhbUMzAZwwmCtQyQ6phh5l/wot9mDT0FPr6QbAUqj8u9bHoCn7iz269zVWL++g2OOeZj29iRKKTRN4ze/eY+33/46I0eW88gtcObVsG5zno0FAjB6uLjHHH8InHUtZDMQDMpnZe1mec4DP4d0RsAyF5qkS+d9JAIYNUCKXdMF8Px+eOxW+Mll8NAzApLnnQJTJxVv39YlNwiFEfBBR7e8h7sicDl2Otz9Y/jBHdK6gQZ7NUiP4We5zrs79gDhbo75vI+GRooUTTRiuVrDbbTkGtUzZFjCYnqJcTJy25QhQxONgEY99TsNkL305hSeHpPalaimBg0YQj0rWE4zzbTQTJAQxzKDObxHt9uMDzCcBr7ChQOe3+KHH+a9226ja+NG7NoATpMJiSxOKo1jGHQEOzB9tSjb5rlZ7/PjgxoI0EDKKiVq9tBjhZmfhAPCCWJWCTd9MgITCGhwX1pEGg8PEzuwXQlbQaMlKsdqEy6MVvJUV4qUoxPUHZSCbtvHEF+aowKDM9BxAfjhduT+3yyHP3SIX6rppocc4MQoHLyTYpm4A3d3SOuCQtSfPuTc22xYmMwLbyoM+HEtjA/Aj7aJMtRT4VfroqodYgrwJR0x3lYqbysX0eT8Vmfgsib467DtT7jYUbwWl57KqJZfBKNuHfPuDtn/FRUChiEdvlEOo3fleMZYSY8aI8AYLobeSgMSoOdFMnMWQXkfllkSgS1NAohln6GN4KKLXqCtLUFJiXwQlFK0tsa55JIXeO21r7LPGFj+vKRlf/0naG2HqgoRk5x2LKzfKt6d0YIbo3BIWh+WrpJ/9wUj25Z9WDZ0xfKjkLw7ikl7FSPOPmPglu8O/hrKS6E3kWeEIKrV8pL+IKgUPPka3P2oMN0jp8JNV0J9bX6by86Di86BzU1QFpVz/U+PPWKZ3RyttJIgSQftORD0Qiy/dNc/FNaznnba2cgGHuEhXuUVXuVlHuEhtrizBncUw2lwewblS9FXHLKjOYOeeGUFy7CwXKg2SJPmTV6nkw58+PDjx4ePLWzmTV7vt5/GhQt595Zb6Ny4EdvO4jREoScNrQnQNOxkCk2JbRuaxvNJAVJTN5ndcgZNyVGUmAmUniBgjeLlxjOp0E2qTWE/Q03Yakkv367E+wn4wmY4ZwucshmuboLhRhm31Ak36rJ8dNl+GgIpHh5Sgf4ZbmMDOswZDf9VAQ0+UVLeUgtPDdu556ccuKQRmt26X1YJiKQKGKY1QIrzi2Wi2HywHq6vhj8NhffHwNSQACAUG3E7uIOBXal+pSHH+qwN8YtT+XpYv/9zjZsPDcNtQ8SAfJdAECB8oWsK0+UCoA6qA4LnF6lIayqFWRWGZYtvZ2g7NzI7CsdRLFjQSDicvwnUNI1QyMe8eVtzPYVKiVNMKABDaoTRn3g4/PBiYYp9c9ne9epNwuTxwhwdB2wLYr2QysDnjhR2F/BBwKfQNRvbzqKcDBXWe6xY0Vq0T6WkdeMr34fTroTfPixABvCdr0kqM50WhplJQ8YaOC160z3wzRvgo5WwtQX++gIc9EXY2ly8nWnCmIb/GyAIe4Bwt0cJJfTQPSAzE48WUUZuYxvttPMID/Eoj7CNFnqJ5YQqr/EKKXZsCVFOOROY6JpGFdcByyijiqrt1vU66XQb5LME8Od+fPhyPqSFYhoDg2Us67efuXfcQffmzdjpNMrQUI6CqiBYNiqWljSSgmwyie7zsXr/w3NV0pQd5f3WU3huyyX8fdMlpHpOIeVE+w2KDWkwZwfz15QSz86btwnoXdwkrQw1pgy4nZuE65rh3PBQPhxZx1+H+XlueJBZ9SMZ7+tPFTKOMNEXY3nH/+1F1IBfDoHle4mJ9rerdj7V9Gav1PTqc7VSl80paYaP6CJqGSh0YHMW3knA//TAG3G4sFSANOlIG0fKcZvkkVpTVknqFKTeuWqQKQNe9NJLM00DWuOBGAIoBnYXmVwAQJYD97TD0RvgpE0yJHinurrMfaDsXjAawGkCLIh8GyKXF2321dMgnhSWA/k62jknSvpwe7F+swyhrToU6o6Ei38C8Z2Y+Vd4/k+8DI++AJXlMKwO6mvghZni1zl1IlDowKIglRZg6umF+24WC7NYHDp7BNAry+CteZJmveoC8BspHCtF2OzlgNq5xJo/5pJLXmDLlvydzD2PwQ13CkuLJ+V8vvEj2e93vwE3Xi41y96EpG5vvFyAujA6uuDex6W+GQ5Jijbol/P6yV07viYgr21z485dw39l7EmN7uaoYwjLBwAKLwprdg6KLjpRKLdNIU0LzdRQi4PDZjYxnr0H3E+CBB/wATG6Gc94yilnIQvJkCFMmJGMpJVWLKwB64eF+/HhKwI8AN39e19A19Cx+zJd22bze+/imFJnI5OFrI0K+dBqIpDMgqWwshAxS5jx858zu76adX3YXdoxcZBp6n/pzs/o8yKjRKHp/f3RLumlSzhwTAQurYDfdYgZd9xd9C0kvRjQZdjtaB8sS0sda1zA4IjA4HNsnumBS5ryrCqkw/1D4cxdFM6uSktq8IOUANLXy+Ds0v7M6f2UpEHHBqQuGld5V0xdgy+WwPEbYbMlbPP6Kji+BBbE4Yytkjr1ARUm3Nkudcnrq+HeTqm/dmiicm23hGUG3esBcg0nDlIjtLCYxUzWsArNNTSfjJgiFGYgTo7AaJ/F+qxOyN0y4baSXOP2rzuO1AwXFoh83k2IkOahnWHOvqlQ8Tfx78Q3IP084TBhZPc/KSCoFJxxHFy1gxk2XT1w5IWy0IcC0rz+2IuwdA3MewJ0XWPatHrmzfuEqDuMWSlFIpHlyCNH5KZePPoCVJQJs1u9CZJJOYebfw9vPAB3Pyb1OE/s4iiIhgXs5n4Ev7serviZsKyaCgFvpeCdD+GnlyV4/fG/Uza8hLA/i6YpIExLSy9PPrmC73znUBYth5/fJ/suL5H6ZH0tNLbCS28Ls/zhxfJrWYP7hS5eJUAWDgqAZt2vvaPgube2/1ylBCzv+x9Rwgb8MsXinhv+M/xJ/wNO4f/fyJBhDatzU+N3FKoAZES5KYtMF12UuoKVgWITG/kbj2G545MWspBqqrmaa9HR8eMnQYIHuJ9uujDxFc0RLAwfJmPZi9V9JlowCBDaWIygeA5bqrOTtD+LjYMW9KFiNqzvgpFlqJoQ+i0nEIqZlEarOGfatVSVDePGtNSUeh1hKBay4H+rXEQbh4bhvbg0jOuaLNQK+KILQj9rhZdjssCDTGB/vdcdSeTecXtXL42AYY8DS9Ky+HfugN01ZmVkklL5/riUI49ND+UBeUexMgWnbxU1ZViDtAG/apPm/Ev6pJHqTBe4NQGl5a61nI5MyXiwK2+L9oklzPfycri/S4bveg3ybbYwyLcTcEWlTIXosCUVOjsuNxAfpiR9a2hi0RbQ4PxBrDdnM4sFvC/vLxp+fCzkA0opZRKTZSO7CT32M2aVruFb3T9iljUdmyBTAgZ/GJofIPt4j4BgpKCOaCt4skcEOzttM6cNTu00Db52BnzxJFGQVlcIMO0o7npUwLCkoH7nM2HpanjrfZgxHR588DSOOuqhIrFMbW2E++8/FYDumIhfHCWqUJRkBAwdWtrh2G9APOUychekIyE47ABhXF09cN2tksqtqSxmsJ09cPmNvbSvsymvyDCmAUqiOslsBM2nsXp1O7MXwBeuEKanaVJT3NoCh0wRMPpgmQChF9sDpaE18lmLJwUEvcyGsgUg//4qnP+FgZ/7yz/CLX/KGwok0yJi0jW496Ydvxe7O/YA4W6MLWzBwaaSSlpp3fET+oSkJwMuaJUwjP63yArFU/wDy50l6D3WRiuzmMnnXPGND1/OMaYvg/Ma5sOEGMZwZjCDdawhSxYTM9d6ESFKlgxpMm59SREkmDuGF+kShV0VxEiFcbqTaEEfdKdhZRt840BGT5zOPkMPZiKTcrZsewXgnVEi8pifFDXolRVwqUvQfl4DP0OsynRNRBe/qoV9A7AlC6/2CjBtcAm2hiyovQVuJ4WRcERtmVawLjtwO0FhPNApqcNCVWjQtRZ7oBNurB38uV6sSMPRGwX0DCCtQbeCoQY81AUXlBU3oZ8ahce6xMlmVUaAK4KkLRemct1yOQ4WV/DbTnms8HFLQUaDbkd6A8cF8s4tZ7hDfp/qgYe7BTSnheCqShkM3DcypJnHXBQKn7t8ZN2fxXwkQKgs6L4CrC1UqzaeKzkLsHEcDT36XQjegDd46iV3NGQhkTM00JTY1O2q3+r2IhyCcaN2fvsFy+hfv9PlXD9YKkA4alQFq1d/m4cf/oiVK/PtE36/wZYmuOgnkuLs7CluYfBaKZpaRVDiDwvL8vuk7SAWF/a2tQWaW+U5jduk5rb3KGGWW5uhrqIUv0+RSDp8uH4Yvop90XSDTNYh0OFwyU1uH6GeP65lw/K1MLoBRrlLSjwBazYJ6I9pGLiuu88YmTaxeJX0JeamYwDDh0gKeCAgVAp+db+0URQ+Ztnw4NNw14///axwDxDuxhCGBhWfEggdt93BwGAqBw9oZbaNFuL0Fqk2PU/RlazIgdSHLCRGDwYGOrorkrDd4ziECRMmxCEcSg21nM25vMTzOY/OGmo5i3PQ0JjDu7TTznAaOIzDiRItOqdMwCZ0wTRSD87FqCvN1QQJGOz1lbM5ZeiFrGA5M/kntdSyLxOJEKEy0M6dDSmqqSmaRBHfto2lDz3EgbNnM7WsgoYvX8D0U06SWXeIOrLDhnZbmIyGLP49CnxampGRjYSMBK2pWprT9e57A0pZ2JjEHI2Lm+DOOphR/FJy0TlIT6By/2+w8Hw3GzNw7hZhnh5Apdz63MaspD7vbIPLqkSsAjDSD7fXCevMKAHCckPAuNk9pjsfPRcWkg7te0q2EmZcP8A3XtPgnDL53V5sYTOv81quVi1iKh8mBlms3GivRNMbpBvXUTo0jWF0AX7QNHQtBclHaGQf5qSibMu20Zk+AkeNoS/iKKDs36xgGD9S0pOF4bUrjB2ZfywYNLn00oNwHAEUb1G/4yFpTp+4l/htQt431Pu75oKJ3yfgksnK87tj8tvRLaKegE8YY0u71DfjSdlHVzxIoHIyiY7NpCKTsFMZgv4UQUOnI1nPui0CqErlWZyhQ1cv+E1JET/1upyr7ci57TsWfn2t1DSDfcRE/3M77HemiGk8IVTIL4De0i5imuu+UcyiF38sjBGKAVYpyGZh0Uo4ePKneYf+92IPEO7GqKdemAk2AQKkB0lHDhYa4CfAURzNwYO2N+9Y1Zgly1KW5EBQfEcd99kaOgZpUoxiJB+ykFm8RZIUIcJUUMk0pjORSTg4bGUrNdTSwAj2Yq9+IAhQTgXRy47Eh0H88QVopRp6eQT/fx1F/fTp/J0nyJAhQ5olLOafvEk55TgFxt6HchiTmUKys5PHTzuNjnXrUI6DLxKhc/Eigp9s4qBLL5XrbLjpPPJftKSCSn8bpwx5Dr+RFh2t0tkQH8VbrScCOkEjKTM57CifZE1+1QZHRQb2Mj0hKinHwp5A5dbsThwEPN/qhdvbpVWj15F0qNev5yk2vVRnzIGft4n59gWl4lNaacJhEZgRkXpiuSG1zI2DjHnygNogn1r2jmMjI6W2N2B3e/Eus5nJW0U1bdv9yboW7aXpMK/f8D02zXwOTX2CL5TmiKvCjD3WvWAabHVqaOm+g301hymazQTf46zK/IE2ZyQhN9eWdOSG5qs7kb7cHbGlSdKYXzsD/vIsJBIiJPHSgvW1cOZx+e0dB679FTz6oopG8BsAACAASURBVCg6K0rghxfJ7MDaKqmd+X15MCi8oVIIeGpIHbInLvvz+0VUYtsiSLEdSW/a7h2OrosFm+1AmzYNIqPBgXTWwiHM5InlVFcarNwgzCsahkRKxDaOEtZ5zw3Q1ikpy4oySZValoD/lDOkrWLUMBm3NN01Hhg/Wljf+0vB1GHjJ3IutgN1VVJzbG4VkY/3PfGMAqwBbhg1N0X87449QLgbo4RSDmY683mfEkpJ7yIr1ND4HJ9jCvsPuk0ttUSIkiCeY4WeV+i+7AtAjBgKhwoqi9o4vF7DBhpooonFLM49H8iJdN5hNqWUspxlbGQjCkWKJG/yOmPZi8M4nKHU584pQID9zYOYd1WWsoum4Ys5ZCtNokYp7bSRJesO4o3neiw95aHh/vyTN6miiiU/vY2WpUsxAwF0wyDT00MmFmPhffcx6bzzCJaVMdwvwpW0I3fYGpBWilNq3sTUs8QtD6kUY6LraUytYG1sbwJ6mqCRpja0jY2J8TRZAloDpQRPjkjT/Jwk6O5K5iAzEE/q0xzvOIrr31rEXxN+HE1D8/lRwTC9tcNyty0O+ZqlU/CbUSIMejMBH46RHsn9gwKEcQe2Zel3O1W4sBqIOKbTllSuB4gNJjwybICU1/svwmM/g9bNUDcaLrgRDv5c0SbttDOTmYOaMXh1Y98tC9g48wOitUPR7Bay8TRv/SZJdIiPun1MUNBhO5hkiWvDAQgbipuDv+D7qVvodaTjPaSJOUB5wepkWQ66rqFvx39NKVieho8zwqoPC0n6emcjkYQb7oJ3FgpQOA6c9zkZK7SxUa7dQRPh0VuLlb9X/hweflbYUzQkbQ8/+C0Mq3XbB1wRjHdT0jdsW1xvykvybizKFkZoOwJgmlaQVnXywJJMIwcI5nPzlgYrNsChpQKAqbQAcSQsYBtPSl3wgAlw492yT8+5ZtVGaatIZ4WdNrfJKKcHfgbnuh+LH14El9wk18QTyAT9App+nzC8tZvyaegxDTKCqaU9f/PoXYeqchg+iC96V4+43rz8DpRF4NLz4IJBapCfNfa0T+zmOJCpnMnZHMhUIkR3ujFeQyNKlH2YsN12Bw2NszgbA9PlWGmyZKmhhhmIVX2YMAoZQVRdMMRXR6eUUhLESZMqasQHcipTA53ZvM0GNhAmTJxeEiTIkGElK3iWZ1jGUkBAdBEf8hEySrsr2EusxmaiMZkzOZsmmvDhI0E8ZytXqGCVeYdZOungHedtVjz1FIbPh+HzYVf4sfYtJzOxlI50K59syqtxJwfEeSXj1gUjZowyXwcpO+ReJ4WBjeUY7F2ykpCZoi7YzDE1MxkS3oKjhDWVDPKN0HV4dYTUJScE5PdXtfJY33aIB159h4dTIYxUknCil0BnO3Y8RnRbEzbuxIU++zcoru01W3CHe6d8dqkoXNdk3PmLg3wWDOCsCIzyiVFAQIMgcE5U+gqH9/3ozXwcfn0ebFoh1iYblsIvz4V3nyra7CM+xNqOIxGAP2vSPXMx0bo6simD7sYAyW6FnbVZ+nQMx4qj0mH8doxkKgypuDTGoXGAfy33Rq7n9jr4cz1sGS8uPQBLl7Zw4ol/ZfjwOxg79k5uuOEtEon+52Ip+H6LpJF/3QY/bIEzt/Rnz71xEW4ccp4s8E++lv+/3z0Cby+Qxdkw5H2dswh+fjVsnQkts2H2X2XckRepFDz+srjC+HzCcIIBYUs9vZLGTKcFyHR9cHcVpUTI0jAEnr4TTF/x/ylHmJzjuPs3RH2Zb6Yv3j6Vhg9XiItORZnYssUT8qffL8xt71Ng5vv5VG4yBW0d+VqmzwVPFPzwDmmfAAG4J38H++0N5WUwdgQcOEHA1HudhSxv3Ej40ufkvHUtX7MPB+GEQ2G/ffq/ht44TD8P7nwUNmwRj9hLboSrfzHw9fussYcR/guijjoqqGARH5IhnQOa7Q3JDROhhFI6aC9iWwPFKEZzDf/FQhYQo4cRjGQCk3JS9iBB9mFfVrCMMBEiRMiQQUOjnAqakW7Y/upWRQvNlFJKjF7XbDtBmnSuMd/GxkBnLu8xjvE00cQ85ro1xwgVVBCnlyQpokQJECBJMvc6+18Dkdnb2KxnXW42RbbaR7bOD45CZRUZI8vMmvexs/vxnaYoLZaIRVLuHWdAyS20t3ddczA0G1O3qA20cHr9k5KOMjJs6i3HAk6MiDvLYGHq4iN67XaGo9rZLA90OBiRNCaSR9UNjVBvNwnDzDnjKAS0QwjD8w6rcMUiwJtx+CkCag8Ng2kb3PMgP8jXY5WlwBAf/L5eehcXpoQRHhjcjvPOwz92V263GdH0QToBD/0YjsgPe+50W3oGiwBBkuuHsiB9AlVNSyhtmYeuLJyshqY5bP0gS9eaLOHX4vi+GZZ5l9k0ZNMQiqJMRUBz+GaF+KB6sX59JzNmPEwikcXvN0kmLX73u3msX9/J3/5WPIz6hRi8FRc1qkca2yy4qVXch0AW1yO+IkITj1nNWSTgd8f34fmZwmgWLC2o4ym46zE468Ti17y1WdiSrgtwRPs4Bfl8wpamT4FZ8/MMU+/DDnVdHnMccbj5yqnCAldtEOBIpPKiFO8t8OqIg1JMN7p7Jb1ZWwkVpXDwJHj6TanLaRo0bZPUZmlUzjOdEZbpnZc3vsk0BSTnLYZTjpbHqiskdfxJi5gEeOE4cv6jh+cf0zS47ToxO7/LdaQpLxVWev3FA98c3PEwbGkWhq25d42WBX95Bq77VvHNyP9G7AHCf1E8z7O0sq3f4wOBQQklVFNDihRB+ndMt9LKK7xMI1sx8TGF/TiWGRzJ0YMe/3COwMRkBcsJECRLlhJKXRcaO1eb69vmoWPQQ4waapAZhMl+DNXAdGf+tbKMJTnrOO/1RYiykfUkSTKZKczh3dxz+8NgwQ2CphGe0ED3gpVkQgot64Ct0NrjmIeNwRoa5K6uZbRZhzDaL4rLxa6a0qeitKdrqPS3k7DD2EonasaJmr2k7ACGbmE5PjLKz8exvTkyLENyP2ukOjvpjJYTtCzsgnqi4ThoiACk3pSRQyEdNmdgk1V8HYKaAHpNAYAN88EIE3ptAQsLSZV6zHJcAH5RCzUukzhiRxZujgMdzRDqU+D0BaFta9FDtdSioRe193ihKY2Fjx7Bkuf3I9S+gKO3PkOnuS+1bMKX7AENmpdpND9qMLTJwlqbJTK+h1i2BAMHLdmLL5pEBc8oAkGAn/50FolElmg0X9jMZm1eeWUt69Z1MHZsvufzuZi8/4WZ0yoDVqahxZJWlD/9Q5SR0XCBM48tac1LzhEg2NQowqJMWv5PA+YvlakOVRUCCD+5S8QvpiGgopT0GBa2NqQzMt3i9z+BhcvF9LuiXNjhynV5p5to2M0oKNh/H2nLePg5dxah5gJnwYfDmzSRzuxYHeA3BVzrayXN+tq7AsI5OzdfXpiztQl8fjE3UA5EouQO4DhSI+2rrD7xMLl2W5qlyd9y07lnHi/GAUXn4odrvya/mYxrPL6d5Ngbc1wGXZA6MU153W/MgW+dPfhzP03sSY3+C2I2s1ji1t/6hvRh+dHRCREi4Pq5pEgxlPp+ptZddPJn/sQmNuDgkCbFPObyd57Y7l27icnhHME3+BYXcynf5CLGMhYDgwZGDDr9QnN/RzMmlzr1jiJs0JDWDOXgVwESJHL7yq5rIf7KYtILNqBsYZcWFmWUo1BkyOSUq144rjmchkZIC7HfOecT3G8UWmUI2pPQnUI7chTVvzwX5ZjE9UaqC5xXQrqkN5WmMbv1eFJ2kIgZo8SMYSuTFd2TeWbreazvHUeZGcOfOIJ/1Ed5daRYt33WCJaX09C4HickSOQ5hmT8QSKJGH+uhzljxHrthRHSjO8xO4Us5rYroLm2T2//9JCMYMqllnRpI5kcgFdGwMmDeGZmlbSY9BReal2HYATsPmlG2+oHjgdyEAECA89y3DyMRc9PZnRlCROjKzFVBtBoskbgYAAamYRO0zsaTR2ltNzlEG6JUeHvpCTQTYnZxSexwzipqv9ogvnzGzEMOWYikaW9PUFXV4re3jSzZm0o2tYTBg0U3uf15dlura1gQ9OQBX7WB8JSUmkRx1i2bOe4ysZf/FG2v/sxmDlfWFZ1hYhhSiLC3FJpcGypNaLBzVfKPvYeDafPgO5uOV59bbHoCmDCOGF6aza5Ijn/wPZ0hiEpSF2XdpDBLOIMN0XrPb00Ch09+XFMXvh8woLPOQmOOggOneIeWxcATKflnCvL8oIZLyJhePAXcP7n3W1KpX54fR9Hmr7h928fBAGqy/sbinv/rh3c8+JTxx5GuJsjQ4Z5iAZ7IMbl1eQqqCBJ0h18pBjBCI7lOHGQSIjbxkgfhEpnkdEzRQNtdRw2soEWWhjC9ofCme5PiBCnIJVnB4d/8HdWsqJoWz9+QoTx42MkI6mhhnnMwcEmi8LEoFLVsN5K0piu4sGWKo7f5lC5cQ689DHZlU1gCOPVRpTx8r296LWyyJZS6o52ivVjxQoHP34cFO9c1IP9pS/gJOIYH7bhi4ap3Hc84ZpqOp1eklb+RsFLOWpK0o5xu4J/bP0K9aFNRIxeOrOVtKVrqfC1s7DzEEYGYvx3VQOlOxbe7nQYfj9XVsD3s1nskigqlSZl+tCAK4cGOKXAhSaqwX8PgTNKZFr7NltqXQENbqqGY/qQtUsrYX5KUp+GJuIgXYPbCphg33g5Bv/dLuwR4PNR+H61KyKZ8RV4+Y+ALrfbliXpys9fVrSPKFFO5TTecFsnFPI5qqGGNbNOpUqrJGjoBCoqMUMhDM0g6+jESoZRZrWjZdOkHZ0UAbKtGh0/DRKapEhXBwmti/F05lDU5SaWJfUiz8C5ri7C1q3d9PSkSKXy3xvLUlx55cuUlQU5++wJaJrGaVFR3JYUAEyHDeP9UOcu/pVlA2QS3VR6ZRmcdowoPR0lQKKUAIfPJ2wqkYBn3pTGdq8ubBgwYayYaccSIvAYMRR+cTWcNgN+cqeAp+UIG6sohZOPkNe5aIXUI6vK5ZibGoXBDa2WtgnlSMuEF5om51lRJgyuJCqCnA2f5H1DvXCUgLOhS9N8Z48ArVIueHpszxbA++hj2NQkxz72YJjzkdwAlERgSDXccJk8tzdebBBeVS4Wbd/9xsCfv08bV38V3pgr5+z3ybVIpOS1f37wxNenjj1AuJtDJjXkJ7lLKjJ/q6NQhAjlprWXUMqpnE4ZZaQdmLFZGsy9+HpgK3VBrej212uJaKZph0A4UOjonMt5rGMtz/A0OjpVVOYEOA4OoxhNiBD7sC8f8gFLWYqhdNZnszSna2ltP46Rf76Tl6qHUdo4nqmzZxEiQ8m4KFrYj725k9TPXqfmngtJksDGppdeTEzXcGBb0XXJkMHCwq/5cUp9qJIw2sljGaqNQNd0UqQIaBp6ejJdjtT2SnQo0aBdiVOJAuLKZFNiDAE9leOymqbhxyGQPpDS6K4PFt5RnHXemfiffon718TYXFXPiHSMb08ewYkHjxtw+yMisH4cfJCUPsPDQgJ2C5Lwi1ZpxB/rhxuqRUzyp06xhRsZgIsqpAF+oHg1Blc0C7usMaBCh+fcBvYba4FLbodED7z7D6EzugnHfxW+/ksAmrIwMy5p2kPCU7jQX89qbRVp0oxiFCMYyV0YuY9ibOgxlC65h5TPT9bw0VEzhpg5FF8ywQ0//A3nffw8U+bOJJ4K0jNfUWd2Mat7b+5f5/DIClnk6qrgki/CxV+Ea66Zzle/+gyplO1Odc+/tkzG4fLLX6arK81FFx3IaaXwblKGE3szDssNMfP2gPHK80Ukks26jESJq0soAEdNFfC4+fd5mb/pCmaSKWhqE6PqZEoW/8LwmVBXDcv+Uvz4Yy9KrSsQgEhAFvPuGKxYC7MegV/9EV57T0QySsHIoaJY3dwo52ea8htPChgcur+kUjc3iTI1lRFgMg1RbaYyeXWpo+TcP/oY8BSnmtQzWzvFvcZnCrj4Tal31lbCtg7Z/z5jYK8Rsr85i+DL3xMmWhYVm7fpU4RFHjNNtnn+LfjB7fLcUBDOPwXu+OGnb5Q/drow6l/cJzcgCjm/p+/etbFQOxuaUv3ukf7Px0EHHaQ++OCDf/dpABCjh7/wZ9poRXct07z2BoCjOIY2V505jvHsx/4EETuNH7fAbztkUfdqH8fVP8LI8AbKjTwF8OYMfo1vMoIRn+l8G/mE13iVDFLE8OHnJE5iGMOLtrOwmJ9p5/pmPyWqHN6byYeRGpRhcPyt3yXS0UomEmVox1Ya9qmn0+mAriTmW98kGcj0s4vT0QkSdNO9GRQOJmausd57vJYadAzKKeNIjqYnPZwrmqRPz1GSIjM1YQPNlqQaex1hzUoTC7q9Qm1krXKurAhzVeXg36qEI7ZuFabMRdzVUEqhbBv9U6wGb8VF9WgpaZDPIoKaR4aJG8yOYmESTt0sNnI+LZ92neSXWYavjywQ0SR6oXk9DBkDYaGhb/TCT7bJ8T0Bz5dKxfasMF23eBWc+yNIRyGjQ+3GZzh0yS2gHLTKFOmSEt685lc0TziQqJVkSuNSfvKn71Ea72FubC9uWPcFkpG9KNvrc2Sywj5KIvDHm+HACYovf/kf/M//FGcqTFMTZhMymTixlldf/Qrl5UGUknmHXvvEkeG8U49S4qZy5yMyEslRAgpKieqxNCLpy3/Og3lLBCQsS9oT/O5g3DENAhTDaqC22nuTBUROmyGsyZvCMHwITDsXVm4onvXnOAKmi54S9WVbpzSj/+xeGRacTOc1MKYhQ3rTGQHASEiY4vqt7nvgtfG4rNY0RYmplLQ/WLaw0ELhi2c87qWIK0qE4Y4aJoC7dLWIXXRdroHHNDVNgFwhoFtRJubh550C0yaLjZvHNrOWnPPpM+Dx23f8Wd1e9MaFGVaVwRFTPzsIapq2UCl1UN/H9zDC3RwllDKOcSSIkyBRlAY8kIM4nhMGfe7jPbIIFgoAlnccyfDwRtLKwq+ZKBwsLKqp+dTT27dkhXXGHaj3DeEY/wUEfO0oFHUMGXB0k4lJb6aOlAWlJmxMK4hCKNGLkcmgKzCVor2kjpEZUD6wDh9KRk9iD1DLdNzXYeLLiTLk32bOCMBAZxqHMInJBAigoTEsAM82CHNptWFKQECrxxYp/cas2K4ppaMTYKwfahhOhwbHbWc47t1tcFNbfg5ggwnPNMC+u2D5pWka2qe8Jb6uRZiNZ+kWQID5uhaxRNvedChHwc2teZ9SzyCg15H0a0ATgPSAcHEsytzuKYwDZoyW7W5ulWOH3OPbCp7ogeOi0tcIsvA9VQ7J46DjNUnVNkfPZN0hRzLp4Pl01GtsmTSNdLSUUCpBj17B26OOZM3Rd6NuX0RjMgLKwSyfmhsku61DUm8vz4apEzWuumo6L7ywGttW2LaD7q6Etu0QCJhoGqxd28FBB9WjaXBgSH4Lo6sHrr1FmJiuC6D5fLLI7zVCQC6Zgjv/KrUx5Qiw2LaAj6YJUAyrkwW+vUdSlh3dAmSGIf13X7hCUqTe9l4jeWF4C3ljqwBhdYWAz5yPBKxKwtI4r5SA1rrNMLYBfvAt2LpNWj+CfrlGti1WbI4lQ3rDoYJJ9hqQEeD1Gu+9fsJsVq5BW4coS7NWXgGracLokilhrx5PKuRLli1MsqxE/EWf+aeci2mIIQAIiL74tsxVrN2OynpHEY2I+GZ3xx4g/BfEDI5Hx2A1H5MggYnJkRzNNKbntokRYy1rCBNhHOMk3an6q5makqOZ2fx5zhr6JllNcqbDaeBczsuB7IcsZAHvkyLFKEYzg+MoZWCbjr90wu87oMm2SJLAbyTwoXFESYzf1dRhaoN/RIab8gXLONBRW4+/txs0na0HHMbE155EBYNkTB9KKbId3dgn7icTagcJy/3JhYIUKXwE8aGDBtVU5xizbKLYZqxFlS6lDp16pmLQQIUpsvmneuDv3WKuXWbI+tBuw+UV4lM6UMyKw/WtchMScWtFWyz4/BZYO7b/XemqFNzZIdscE4GLy2UI7s6GpcTzNOzepSsFa9OiHi2MkAZNlrjRRLYDhJ9YopSsMeCTbB4ITeT5EwLSZpCx4LTX4R0faA6oVhi1FH4+Xc4pVPAaDHcfM3vzQLgyA6/EYfLnIT4N5i4Fywe9e1czP3JKTsCiA4HSMlLdCuU4fDJtf0iuAn81VE7DMmpJuilKz7zSU1UedlgDo0dXsHp1uzvfT2G7MsrRo8uxbUVl5fbp+q0PwrI1wjY1t1XhvQ+hvk6AI5uV1gCPjUXD8ljGEnC2HTHObu+W+uU+ddDSJj6g9bWSKnz+LQG96fsJIGxpFoDJZvsPvPWMu3/0WzlmZXnems0w3EZ5VzUa8AtY3v2YKC79pnz+Mq5DTMA9P68Fw5uuoWlynKwl23TH8p8tTRMVrAd8WUt+LRfMImHZV9Ya+Ho67nvkpZDXbJIbg0w2f4OWcnsnV67/bED4r4o9QPgviAABTuJkjuQoMqQpobRIpfkar7CA+TmmGCLE+VzA8ZHhPNEDvgIBQFJBS3wq12n70UO36xCaXwhe5kU+ZGGuJrmUJaxlDZdyBSUUywpXpWUkT0opMlqMcrMLhYmj4L1YFd83V/DHyqk54FmSgn/0wDbXzuythJhBOxkwho0h2rKVkq52Nh96PCM+XkR08zr8uk6HSuArCcGBw7ALFvC+/ZRFtVOVN/S1lUxCrFaVjNRH5bdB8Q+e5GNW5BSti/mIQziUEziJMgO+WSG/HbbUjzJKamoN21Gt3dnu+j+6gKdpYnbdYslcv5MKLuOzPfD1xvyA3LficH8HvDtqcAGLFwkHrmyEF+Py/PEB+F2dTNrwnHKMgutlIUbhgR2Ie0JuKnSoKe+Vt5+MkgXzu1WSLv3eXJjth0hawEoBG8Pws7kQ2K//fpUCX8FNwNJUfiRUcxlY+fu63DupkIU6lgLH0WSYXSaKNvpbRSwjkZT9GIY86cTD5XFd13jppfM59dTHWb58G5blYJo6Y8ZUoJTigAOGMnp0OZYlDOTZN2WB/sIxMGVvcTh5fqakKr3L5hk+N7VKXW7pGgEkTyDjubtouFMbADRodA2wh9cKg0ql4JOEWLI5Sphae5cAblW5pPXSXh3PzIPUkGr4zm/c9KMhgGM7EHZbKTLZvKm1YYjisycuMwz9LpMt8ux0hMlaVv58NU2mRdi222qhg5N1U5sB+dPngmoqnRcReS0QpX2yJQWZ2ByYhoPy3MoyAdbC5IdS+df7fyH2AOG/MMLuT2GsYAXzeR8T0zXDViRJ8jf+xq9rr+PthE6zp61HFrB7h4JPM6mi+FYrRoxFfJjblxcJErzD2zmVqBez4rLbVjtD2NeL8j4OGoS0DPN6RrK2cg2TmMwLMRl1pAPdtkxssMl/gDJ+P11DR2DaFtFsinkXX0/JJxs4N7EcY1I7FccfSKLMponG3PG9aRh92z4cpZG0QgSMFLomvmmxdCUtnV9HG5pfAdayho9ZUfR6bRzmMZcDmEo11bltKw04dZD2gr7RbPVn4poGOLA8A4fZ0mrhOHCluwgW1qK2WHBzG/x+B02/J28S67QAcrzlKTh9C7w3SvxG7+sSAm1okppMOXBO6cBeqIVRbQrYL0gKe2u05EbA0OCnNfB59zo80SNg771WDQhnYG0Y9ktDr55PzWbdxe/4ggWytOC5W6w8+ytSvSsg7aofvQ3eUPgMRdbWZF4lcg3jSamJfe4oOOLA/C5GjixnyZLLeeqp5dx++zwyGQtd1zn88AZuvPFoQOPGu+HVd4XNKQXX/FoAorpCGug/Xi/G1cGAML2sJYv4nEVyXG8ahK6LwtNRAk4aeZcXpQlQLF4lSk0PcGzHXfCV2zrhhs8nrjTvL5G5gtUVArJzP8qzM8cRJhxLCLDqWv5cHEdA7N0PpY5XVlLs1+mxWzQBIS+taRrSflBWAhecCvf+DaysnGdJRJxe1m1xJ9IX3GQXgpaX4ix8G3Ovy4SGodKMbxgwfTKs3phnpkrJ9SuN9lez/qfGHiD8FNFCCy/wLM00o6ExklGcwVkDGlDvKOYzT5wc3CVFQ8OHjyQJYr61LB4znj90wjtxmUZwdaU0T1tYOZuzNClKKWUta3NT6TU0TEy8SRQb2TjwCSiwUGiaQ+Hyb2gOcSdAh+okpeDWNrH5CmiyYHtZk3wXmoYyTRIjxpKuq6MGiyuPn8oxNcfykvY8GkG2sb7f4T0WWKim7c2W0J2twKc7GFj4jCTtvYeyJlnG5mzeC3QZS3NgmjtvdGxgGUs4hhm7/H4AHB+GxX0WibgNSeCudni4O8nplVs5KKCRYDhBLZ+q1TRh8K/0Dr5/peCJbmmFCJNnWSZSn/tvF0QbLXdMkcuOj43A/fUyW/DaFng2JinV/QJw5xA4oCBD+NMauLoJVmekTSPiExPrywraUtM6GH3u2DVAGXBDJdySFBas3MevqZQ5iF4cHpJ0bns2z4hV4X4UkAGWAKuBMhddXgMzmMJvOliOH8sOEAlJmvHRW+GAfQeugZ599kTOPHMCjY0xwmFfLiX68XoRVAytESD7pEVAzrYlRYl72ERafjVcCzV33FEhG9KVpCm9kUF+f97oWtMgHJA2B6XyDKgwvRhLwPuL5bGSiIxquuoC2e7196SxHi3fz6cUJDPCJkMBV+Di5D97jpsm3twojLW+1hXcuK/PMKCyRED/gAn5HkSADVvhmTcERFs7Jd2aSkvNNJsVb9Nk2gV/z8xbK26c76vW1TU5h9Ub5TkBn9Qxvd7GlMusG+oE+Gt2Q8/f7og9QLiLkSDBQzxIhow7RV6xnnX8mT9xFdcM2HTsNY8XOq54kSKVm/7uhZcyTJAkasAp1Uupq/6ANBk2MIoPaGMda3P1tAgRokRJkcpNoNfQpP2AAAo1IEgfHYEHuqDCUPQ6AUJ61j0+pJ0gY0ON1Go1rEsL8AV1SNoQxLjm0gAAIABJREFU77cn73VqxHWTkvJyjojAuTWAVouBSTvtODj92ke8KHwsaCZI2GHAj40Pn7Jx7Gp0zW0Kd4FwoGvthT6IQcDOxLer4E/dUkv0IQwnjXxZSoOrOaByNnOTw/hnsorJ5ZtpT42mNTk6BxgKCro8iyPhwHXN8HKvpCptwHSk9cMjTMsyYuf2RIMMBF6akvpmVBfA+fxmWOAySVMTO7UTN8P80TDaPfBq9z1LuIKbi8rh4opigJluwD81KCnw40yZUJuAk2vhKAWzesW1pcOGNWkx/a4xZXDuA10iUPLGUBWb87kX7Q+guaLPoJVFy1hkDINIME5teTvRUC9jRpSwass+fPsC8azcXui6xvDhxbLZVRvybA4ECNMZqaMp+oOql67VlLC2bLZP2q/gY2VoUFqa7y1MpsFJCyty3JSk9/VVSH9fro+xW+qTD/1a/v3qO3mQ8oAux+p0merwna/B5T+F1+fK8Twg0nWpSY4bKXXIrh55DaVREaRs6ygQyrj7b2p1NTMu+1XuDcGWZlF6JlPyvMKJFromvYUgzNLvy6doh9bAtReKzVlzq/RE6rp4MHT2SC3w2GmyfVuntDscUtCEv60dVqwTAJ6y9+5pg/i0sQcIdzHmMZcMhQ3t4gzTTTcfs4IJTCravpFPmM0sWtiGhsYkJnMUR+eUmOPZm21sy7E4EMcWHZ2x7MVTPMkKlrtH0tjKFhQKAwPl/sSJu+0FaQyMHCsESJPCxOQwDuv3WvYJwKUVcHdHiJhKkHAXB79mEDDSnFa1gTGcwjbDldErYQjbCx+y8L6ZgKk98KUyH8dxPH/j0dw5iQLUwMTM+Y56oObgYGATMWPErXKCegbHqqE9OQK/BnsVfNkP4ECWsDjncANgYaOhs992JnZsL3psuLpF6msOoq60lHxR6n29HFD5Lm+1nkDcjqArGRxlOQZJx8Yp6Km7YJARQg92wbwkDDFgq5tOtJQAZESXY04seI1hHZ7vlTSnoQkzXZ0R4PTUxFFNmORtbXBvPSxKwjUtUisc45M+wD92Sd3xqwU9cHdMhaPehx4/6A44OvhsuMtlZN0WXNUsSlOFCGN+3ylZiedi0rs5xgc9hhiCJxwxPM/d0jwNLAClg5ZVKEMRDFhoGpx40EwOnfAh0/edx6pPppJVd3B2Hz/PnY3Ksvy1cNxevb4Mp7DGBfn0Y6ave4mSWpuuC2NMZkQ8Yni1PHekUSqTHz+kkRem+Hz5aQ6mIabe3/4KTJ0oj1WWitDGa2vwjqnrcNX5MizXUcK0vNdj+iTlmUrJ422d+fYJzX2NFSXu8d0VPZMVoDNd8Y0HZrqbXVi3RUAtHIQ5i+V7XzgqyWcKQ1XIfrKWAOTMebB0lZxjT68AZTAg6tqe3rzZ9gH7wk1XCPArJcKgPz8t5xUOwZTx4krjmSf8u2MPEO5ibKOlX03LA7AWWoqAsJ12nuFpuunKgda7zGYLW7iQr6GhcRiHs4wldNGV2w/A4RxJC82sZAUGppvys3PH9sCysK7ow4eGGFbLzEHZto46yvtYtXlxUQWcEIHX42W8l9lKk9NLfbCTM0oMTjBPxI+f4T6YGhTm4XpZD9AAIeHXhKWUafBUDL5UBqMYzQQm8hGLcvW8Yjan5a6ijg6aQ0DPkNU0OuJT2NhzCGnH5Kaa4rE6IxjJNA5hPvNyjj0aOidwAmWUoZQAWUDb+XE8D3XJoj7cJ/6eCumpU0B1eCOrYnsTtyKEjKTLLEJ0O1GCRoyImSLrmDhOCYeG/BS5HrjxTA9UG3KdwhkxCtcRsMKRx3+QL21yU6u0ttQZsvh12ZKOtimuY+rAR+58pge7pJbstUeE3PfswS74clm+xjiuChYdBbd9BAu6YUwQvrc/THRNlK9thha7eCJH3IHftOUFPQlXiDPShDVZWVB1wGkC3nQvQQi0cshgYid0AlqWb566gFF1q7Atk2MO3MgZZ7iTDj5FTJ8ibGRbhysucsVCmsuyveG3OwqPfXnKSW8GYHevAINpCJNauU4a2r1311N66prYtBXvFP7whIDKqo2SoqyvkQb9rFv7Nwy460ew9xh5yqjh8O6i4uG2jptuOGiSTILPKTo1mY147DQ5jmkKoMWTAjhZ1yyokH3pmqR/f/sw7DtGTAxa2vNG47ourz+dlv1VlgmANQyF2QulR9EU9zxSaRfcggKIr93v9icW3Aje8zcx2/b55HnxpNQ9r/oFPH3XTrwx/4LYA4S7GHXUsYbVRY8VAk5hLOEjF+Dki+iB03rW8h7vcARHESTIpVzOHN5jDWsIEWI6hzCevXmB51z2pxcdp+9x8x6g8mctdTnAFmB0eJqnOJ0zqBvAeWakHy72m1zMKJazjHl8zBYSzKKDoziGKFF+VScN/vOSgwNhEGkbiLgLZOHd9pEcxQqWF6RHFVm3ad/MsVgJXdOJ6j72T1/Ee2kfe0fgzBKYNEAP30mczAEcyHKWYmAwhf0pp5yFSfjRNkktRnQ4rQT+qyrfFzdYvNwrKcfVaUn7ZVW+DtqRCbMtPRIbA0sZBPUE3dkIVf42AnoahU6ZmcJ20tzeHeD4kup++7eU1AQ1DaYHpRbZ6TYqj/ZJbXCsW4fbZsE7iSx+o4tGLY0fPxG9HPCTdERx6IUD7OOyiHWZfHtFryPg5dcEbHscMf5emhb2OyEItx8x8LWYFYe+lzwIdCDMdH5SUq9eKtjGvenQILUAEjoQBYZ4i6yOU65Tkkyy17B2RgwB7BgEj4FPCYIgrOMPN8p09NkfyELv98k1TqX61C23t6OCDX0+UBmZmBAJwtfPkr//+k8wcZyk+Lpikha0lTjLdHb332U2C488J8f2wMvzCR1SJYKWB34uE9off0lAo6VNWGkiIeOdHEfApqoU5i6GYUNElBMKCOCs2QS3XidTIK7+paQ+dV1Sl4lkcY3PY8KpTN69pq4qzxwTSQH/Hrc27TXHG7q83lzDv5IbDC+1m0jBSYdLrfaR56CtC6ZOgG9fAL97WPbvpW4NQ/Y59yNJY/c16P53xB4g3MWYzqHM5/2iGmGWLOWUsw/FBY5mmlHuAl84DFeY4TuMYSz1DCNAkGM5jmM5DgCLLPOYyxpW52p+psv2ChvyPZbp7dfGxsGhicZczU1SpTIqaR5zOZ0zB31tb/IGc3kvt7+FfMBKVnApV1BpRPlDvfSl3d0O93TmQULEPmJpNSEgX5IuW5xIvKiimjM5m1d4mV5iuXSmTMLI4MOPB+0Kh1FaA+eV+ThvJ6aU11JLrXvtAP7SAZc15wU9IZdJdTvwm+186RYkxZmky3ZbFSiue21OjsDUsiilk1BhUnYQU7MJ6CksJSiUsMGvKdZmk7zIC9jYjGY0+7AvfvycFBWRyxATggZMD8s1PSIsgpfCetZGp40ubKJaEh2dBAmUkSCoNZBWRk6ll1QCdN93mdyEgHjTbshCm/sClLvNqhTc2i5CHI8BXVIB55dKa8jMhChszyjJKwALQ9NkAfwgLeDrYXEaV5rvCYx6Qa+WehpZ8OkaStk4vQp7HCzIjOTue2awbMM4xo6eyDfP2XF9cHvRMBT+/Et4cw587zbpzVuyStJ5afdOZkes0Ksn2rYwpkhIWOHKDfDTe4Wh9fTKHL7xo0Q16TnPnHcK/PHvxVMo4nEBEr/bb1toeF0alYb6F+8VC7bv/Abue0JAVXPBKum2Nfh9kkKcvh88+Yrsx6sfptLCQt/7EK7/rbR1lEYATVLEGfdLYLnvt+mmcEN+qfkFA9JaYhoweW+pY3oN/Uq59UK3v3DNpnxfYjqTFxQBhH3CVn/zgJxPVZnMQzz7Gvhkmxw/Ywl4+3yuvVtSzv8/IfZYrH2KENXoczTThIbGKEZzOmf2E6S8zmu8z9yCtJ2WA68wESYyiS9wam57G5t5zOEt3sIii4ZWBGjeLL+BpkwUqk4HMvYOEaaCci7kG/jx9xPtJElyB7fhjWPyIk2GqUzl8wXnCbAsBQ92ipDjgxR0u/WKMmCMH/YNwX1D8/J7L7awmWd5Gh9+woTppZcWmt0F108AjXKtjC/x5aIhwjsba9Kw33pRU3qvwkHu+CYE4KURMHSA/r6tWfjSVmmd2JiV5zrkpxp43/mgHsdWJoZmYzk65f4uTM1C11QuVZZx/IwIb+Fr9e8SIpwblHw6Z9Jj+bi4Sdx8vKg14YGh/c/rafUcP9o0FR2dgG67r8WhLVVGyi6n0/0Y1OnwgDtz75VeYYH/jEOHazFnI8Be4wo7qg0RvIDczGyzZDRUky1sznLrfGU6zElKDdIDzbiSf7c6ck09luVdqzpdhKKpDyDxPjAE9NluGleBVmkz8opWsgugbm6W0tIq4qkwWQt++wOx0fosoRRcdjM8+888SHmtALqWB8W+oRX+RQkTC/pdZoSkCj3LMp8789BrSE9lYOJYOPdk8ca0HWFJqQw5ya1S+T5JxxFgG1INT/7WtW87sbgh3YvTjoUzjhNQufcJYZiGmyaPuunbmgoxon7wH8I0C8U+PTGpTy5elVeZZi1Jd07ZO19TXL9F2JmuC7vzmQKI3nUDAcOMJcDsM/NgqGtiU1deKrVKn0+EOms3y/E9EZlpynWIRuQaREOw+a3+EzF2Z/xHW6xpmnYbcCryHVoHfEMp1eX+3/XAt5Dv89VKqdcG3dFuChubFpqxsaljCHXUcRGX7PB505nOAuaTJZtLB+KmKyOE6aSjaPt/8gbzCoCzEPBs7NxAXA/sCgHPwcmpR1toKdqvhkaCOEkS3M6tlFDCIRzGwUzLAehGNqBQ/ezUdDQ2DND2MCkogpDDOmWRNZE3qAtJv/2lnn7z5QCWs4wAAUKE6aCDmIqxumcy8zoPImFF8eHjrEgUX60fPsUX5O4OWdzdEgYgf88iTK/VHhgIX4yJinOIAZuzFM1jL0ytZZwIYGMrAwWU+zroylaDUpj/j73zjpOkqvr+91Z17sl5ZmdnZwObYZfdZReQjOSkpIegqKioICIKiPqoKCYEA+pjQIQHFBFUEJC0gOQcdheWTWyOs5NnOndX1X3/OFVd3RMIvo/K876c+fRnUtftW7e676/OOb/zO8qi4AQxlGZO5UpixAgTIUyYHnpYYa/nscFZDNoy1vQQnFYFR1TA0iRcs12UXxZEpPyhN7KLkxpW8ufuReQcTUA5pO0APXacAJoKd9fMIh3Zg0rINQ4S9jSVy/ZVsEcQQhpeykNnyfkHXQLOSxYsivgbcc6RetE9QrDe3dAVrrdYIYxR7RJRTCWh2CH3rmFJBJKL4fUtMLAQYlOAHiACe+xtsj3fQoUNTS5RIhIRwPnxzfC+BbL2eS1jvpmU3FimFHz7InjkxQIdc9fTNHsbVqKKlQ/OpHtLDabpsyRLzTD8+jcQgkjGresLhQSsQg4MJAQAgkEB2WBACCzZvIT4Ni6Fvzwk+bqd3VJHpymvDQQfzHIFePg5N/xolp9vwZIaxE07RKHFcktZPELLwLCAtVcC4uVFR65HVRz2nQ8r14mXW7BgzjQfgPoGRL/Ucdwei26Bvxcu9dbEC+3mC/55ePPtHZD3R1Od/Lx+i08oAhnPy6MmkuKJfu0z/1oQfDN7VwAh8BDwZa21pZS6Cvgy8CWl1GzgDGAO0AY8rJSarrUe4638z7FeeriPe8vYjYdwKHsw/S2PraaGUziV27i1jMmpUAwwQAUVxTzeAANuGyRV/AKPRWkSIsQkJjOD6TzP8+TIlQGpNM1NlxFuPPO8Sk+cu59+/s7DKFRR5q2CyqK3WjqGdMcYO4FzkRt+9PZVr5h6owVrsrDnGMpXQwxhEmCIQQboZ6BQw8b0RPJ2kIiRw8bihsFqVuakX198nJzekA3XDUidXQA4sQo+XgNbXG/Oq98utayWVlZjWZcl4xjKD6mONI2MbblZWw3szk3g4MaHWTM8m4RVTXOkhz2rX6E6mCHkrptCYWiT36a28MLQLOrdThmrcvB4Gjbm4es9/vj3JUWh5oud7XREd3NwzRusSrdga4esE0FTRVgpQkrWO2GL0PT7on5PRU9sfEFE8qK7CvKcnBYPvj0AnW7Pu5QLcqWbcNiQEN1vm0Wy7aWsMETPqoYnUnDDkAtUJczHAPIYdsAIQvvJQAIiUYi3SSguHIM31sPMET2qK+Oy4X93J9ybFyCcHoLLG2Cvd6DvCuCYOQ756p1UT+hDOybKcJj2/uX89Yrj6V7VXmRgKpdY47g5Wo9d6vXiAx8EUhnxxjwQLRTkATCxRTzGb/8KLrnaLbGwBRj7hylql3qvYSgJHdZUwpR2CUeOZ72DolITMMEy/NIGz7J5IaYtmiMeoacy45ntSLNf0xSPbNjVFo2ERW9UKcl3BgMSpsy6+qTe+ZeaUjJ3z+M1DPEOYxHxIr2xt3UJ49ayysOnHnNVGXD1pfDpM972Jf2n27sCCLXWS0t+fQ441f35JOCPWuscsEkptR5YDDz7r5iXhcW9/I0kSTKkyZJDAX/iNlpopZlm9mYBbUwYd4yZzOJ4TuRe7gEoelwaSJJkh9tlfic7cHCKm7cHSF44NUyYmcxkClN5nucYZACgyBrV7jFZsm96Tl4dX548L/ICC1hIgADttFNNNUMMlbBP5V28L/uNOdaaPBholNKuQogqguHS1NhA2Eobz/Os9F7UmpwdZGHt83TEN/Jy34EEVIGokefVbJQHk3DyGJ0WClr69y3LiFcSU/CbfpGAOygquS5HCxh6YTuA2SEBzmZTcnLhkg1jnyj8LQkb3yJnYSFg5XmJCaua5YOLOKjxEcJGDlNpkoVKglYrSg8X0XjQdujJV9IWBHCoiqxjRux1hrTNc4npBNUcgspH6aQDt+88lJQxiOWEyDuKAprhQgWOViS0H3LyznFNTgSnTSUd2QfzUu4SVKIE5IGViSjBDDvi+Q26pRtZ7eubag27C3D2Dkg4MCcMJzTJmh1RKYo1yzIytoHkCDuC8KMmuC8lQHZcndQ1XrZbxsgif5//OqhuKM0kZPPQH4E/Z0A/DzuXwpoBeKQTbvokHL3nm1+XUttetZK6SX0kuisJuUsaCuc47MJHePC/PkzkMIMuDcE1kH8MnKy/8YdCErIbGPZBMFciQVZqnhRbV6/UyNmOhCsdLYX1G7cL2A0O+7lmraGxXjzMKz8n4LRkntsO0vZzf56eZ0XULeVwHyOyHiiE1eoV2+/oljycUuLR2g5EghD3GKhhkXwbTkqdYzYn3q1jQ06NxXMW88K6Xq4xGIAqV/xdO/L6LY2yDumM/G6PAaYa2GcOfOr0t3Up/2X2LippLNq5wP3uzxOAbSX/2+7+bZQppc5TSr2klHqpp6fnf2Qiu9hJH3300kOCBAXy5MiRJctOdrCD7dzNXWOGDkuthRbi7qc+Rw6NpoEGbGxu5zbu5k6e4WlSpIrM0tIvgAhR9mA61VSzF/PLvDyAkOuXjWxvVGqep6dcMMySLbZbUijO4sNUU02BAnm3FdJ+7M+cEbWRNjab2cSc6peJmwOYKotpZFHKKoZo9h7nLn4KU4prAAYZJ0rGidIY7qUhLCHdgNIUEPLKSOu3pV7u70nZXNOO5Lh2WvBCGhbHpFOEzLNc7uuxjKitXLYbTt0ujFLPDo+LxzMwxod3pHm5MRBQ2ZGZxG1bP8LdO0/l9m1n8trgQvLmTpZlNQUtvRWzWtGTnglAc9XjtFY/TDjQR8gcYm7tU7y/5R5XE0csBKzKxqjSVVSaaUKBJHEzTUGH3EpLObfSc0w4QpDxWKIFLTWHy3MCejbCLC1oyDhSx5jWUg+Y0/Cyq8WqtRyzzRIiT0bDsxk4eis87+aNHpkE59fKsXETDovBkXEh0XyuDm6YAKdUCzDf2wE/boHvN8MDHXDFYsk/Zdx7tlweugfBOAycx2HTbaJfG6uD1Hb41DdFPebt2ia1gY76kOQE8/JID4fJNNeQOc+hvwOCrZDuBqsOAm5IWCEeX8YVGPAK3sdjUXhhQ8uSEGdFrCSM6hbqR8Mwf6Z4w5VxmD9DvLePnyq5O4AFs+ThFenb7h1cTZW0e6qt8msVR1osKiB1633wxM2iZGO5TM9FcwWsY9HRxyTT8OTN8PcbBKwtR3oNhkKjw5VK+aFjr7O8o2We6Yx4vdmchGiPO9glFaXLj/celi1s2XdTMT38C4FQKfWwUmrlGI+TSp7zVeSm+5Z3Or7W+jqt9SKt9aLGxndOsig1C4seetjIBoYYLNbvlebsBLAgSJBneHpMAotna1nDMEPFnF6ePF3sYoghFBAjXgI+/pcHdjXU8kFOKZJx9mIeIcIECBAkSIQIQYKjCDAjTZX4myDi3qWdHBpo4Iz+k5lxp2LCTX0c+ereHKoPLxtjgAFu5RZu5zYObPob5075NUc030Ol2U9TeBuxwCAtATis5G5fa9HJdDRkyFJHHVVUodAElI3WMu+6cA8ag5wTIYzU8hXXW2u+2ZtmwcYCX+/RpJBN3EC8FBvxfnpteKxTPCIvTxjETzemXMJHtwXf6/XHjxjwk5aSHn1jWBzJIX6kRsKQcQX1pjS8NTDpyzcybNWSzU/GtlopqBTrCykUiin5o0hb9YTMAaqjqyk4lThOBFuHSVmV1IW7aI5uLb5WAcnPtpmVxOyJDOXayeQnosqyn+VmITWQL2UklGlSHvLJAGtcz9C7XZoYkBrROSF5zfV5qR3sseR8427NY9wQsPxaj79e17TC5ukCgo+l4eYh+GEfHLQZft7nv27YgP1jIhFXbcLh+8HXPyMb4+5e2bRPPwMa9oftD0KoCkwXnEJV8t757zvHvy4jLUyYcNRh0VwpeWhpgGmdsCq4N7PbYEIMKoYh3AfhRmiYKixOZch7NJuVjboy/tYbticO7+USQeYdj8pl6umXzb+myg1DbpRu95/5Fsz7oOTTgkH4zZWw/3yYPklClntNhys+Ky2UAq62Z6n0mTe/eEzCkMNJaG2Cv/0Shl6E4Rfh0f+WEOhIz8yypEA/FIL6WtF3tSyZb2VcADHmkoUaa+VcKmJS6oH2853DSbmZ8TRcHUdISj/5MrQ1yJoETP8Y05C8aiLNu87+ZaFRrfWbdpVSSn0UOB44XPtU1h1Q1mSv3f3bP81WsJyHeJAMGQpllIlyc3BIk6KBRhIkyJMvNpEtNQuL53keoCzUaWOjUFRQSRddDDO6CMnAYC/2Ik4lL/ECU5jKHkynkkomM5mNbCiCm4XlFtmbRN0Gt15bo5HlFp4yzSEcWlbYvuzle7n/8xehM3m0o9ll3s3yY/bnY1fcTMAIotEs5QGGGSZNiqAJMWUzu3ols6tWgpJQaUxFWcFx7MU8nkkpftQvubsqA86sj2BUBGhU1URVjHwoyc5cjWxAVpThXDOWNugISe0fwBuFIS4fXM0TQ1MJm2mU1QY6QAHFgCMEjhACcBOCsqG3BYT9mNdeY173umkBzMkRqYlMO0IwAWgLCpGk3x4hHO2ZEg/oKw1SnP6S230h6ZTnFXfmoyzr+SBNoSGGnAJfbq9FR0x+b0LG6HXJE4qcFmZmwY7xXP8cNmUayNjyoXQQ0HglC8O2QU4bRQauF3723ycCjUX9VzfXVUDq/rL4ZIp8yTEG0GXLB6w5KCHVqSERQThjm5R3lFoIYQuX2p+GRDghquR4rQUwv9kjYe22cXKyJx0uHSKGErL5Prwcrvk+JNZBMA6xCfLd0VBXAWs2jT3OJjaygfVUUMF8FhAhwlz2ZDvbCQcdJrRI6mBnQWMNV9IUDdA0SchEr0bBDIOKwH7z4NV1AlzK3cCVcjsrDLq54xFhSa/XoOMIGNkeGSTtlwVYtoRGU1kJSSo3r+pYUpj/4cvgQyeKZNmiucIOrYzDeaeJ+PjEZvj+9QI2k1pF49M0obJCZPgKrojBoYvL5+Z5YCcdKuo2cbezhdeN4pwP+M8950SZS9q9toYBMyaLVmlTHQwm/dpC25b+iP3Dcl5oIc8E3Ca+Xb3w+e/B7Gmuzis+A9djok6bNPa1/HfauyJHqJQ6GrgMOFhrXXq/cDfwB6XUjxCyzB7AC/+seWxlK3/jbpnTuNFy31KkqKGWIMExm9cCbGMbFgXChIsen2caTQ/d4+b1IkTYyEaibpnBMpZRRRVncjYncyp38Ge2sqXYwLaRpiKBxsAg5H6lSZfkEjUhQhzL8cxmTvG1ClaOpV++DGWaBFulq4V2HLrve5rnDr2NAw79EAP0M8hgsYTDwCBoaByPDSALR5Y0d3EnW/KaK3fvTUyJJ5XT8OvuNk4JV6BCSWwsGgIOEdXLrnwFrw0vAB1g74jmRy2KiUHYaTn8JH8PLyUPwnT9JFPZONrAwcBBkbXlQ19piHcz6EqCWbrcd/LIKKV5w5F2dIW0lko55c9RSI3XniHJi51dDT/uE9D0QEbWXcgp6wtQZ1aTdOWtKk24vg1+OBSloDVZbVMfsJloRnmge396CgH3Bkm824km9LvLGjUg69ZBgni4HjM/gNSiJ5GyhkpDOkrsLkjeLslo/9G7vatUck7bClLCUdDiuS2O+vJhpY1lLfyyC89ucu/fvHZRSokgQUrDg0n42BiCRlrDUy+LQkr/kGy2d/0dMgMCELlByA9DbCpEayCWhZmTy8ewsPg9N7MN34t+lEc5gzPpZDILWcQyXnFzqJp2o41qqrA9lms7GAGw8lARkY1875mwfhu0N8PLqwRgBhM+ezIU8OvyvPNQChrqBJA8KTKPXamUeEoDw5LHM4ySsKMpALL0WWnSu2WnHNdQK7m+r/1UwL+7X1RzKmJw6D6wowd+casArWUIuEzvhHNPlmGTLoPUU6X59RWiZvPESxRl2Y45EH54qX8eJx4Gv79HnufVL/YPwmH7ime4ZqOQhbbugrmzXUC0/LIi24GqqPyezsL2rvKcYjQs76N0VkLGF58z+j3x77Z3BRACP0c+Pw8poa49p7X+tNb6daXU7cAq5HN4wT+TMfo/O4LtAAAgAElEQVQ0T2LjECZEmjESVCNMoxlikAM4aNywpJe7MzCJEHEJKLoIimPl9LxSizRpaqmjx9UiBeijl+v4JR/iHM7mwwwxSJYs1dTwKit4kPvLmJ82NgGCnMKpWNhUEKeTyaMEq9eveRZrOE2o0RejVIaBCgZY9cBdHHDoh4qh3Ry5omfr4JTttKWC38+pJwmyJ1WmvM0iCuqVyZM9h7J/22/JK9c/MaE5muRLU25BO3H2CjQxXx0FmPw+sQszmiBnRzBUXkLJZpqkrgANDiY5oN6AKxplk6sPwPtikgNMOBKaS7vTDLhgMeAIYSZWvgycWS2sTcv1HIcc8SoB2kwhnRyzVUKwI3OQADVKXiOvxdvaO+IzOduCNkc3vMZLegCHHgxl8OTQHIbtAM2hLHVOAxlHAOrFnDBcd1oyVkRJrs7z6jzWqg0Mu78HFcwIy7n2WBQb2bpdA3E5kUXQHnbH25yXkpGwgijQa8EBEXg8AzEXDPPujcEXRjRZzTqjgdYruh/vZuPmu0RyKxKWjfyex/zC8LThCkQ7kN0Ms2bJNf3YyeVjPMezbGULQUJF0fo8ef7Cn/gCl7KYJcxlT/roJUqMerOeNyoU9yTcsHkIOk6Ddb8Xqbv+rBB2Zk6WGrhwUDZ/2w0ZRkKyDgXbryE0XQ9rYEjyXj25chCsjkMoLOOAnJM2/JpMcBma7s/xmHjIE5oFRP7zWgnZehqiazfBH6+BPfeA6/8sjNJ954l6SygEF34Htm3byISGDbQ2N/HRU/eivVVx369hyw54fb2A5nMr4CNfEam3c0+GWVPhgrPhvG9I13pvPjt7YO4eAqZdvfCdX8u8d3RLHjSTkxsXrcU79DzOcEhA74CF0okjkZL5tzTCL772nkc4rmmtp73J/74DfOdfMY9BBjHcGj097sdYzAOCiXSwgPGrgFtopZIqEgwTIkQAs5hzDBMZEwhVSQl3mlTR+wKK4Hkv93ABn2MjG3mSJ0iTooJK4lSQIolytwcHhz2Yzixmv6mXa5qusGaRi+ia1hhuv5k66gkTFkAhiDUydKxl9to9B1tlmF77KDWRrVhOlN7EvpCfQkv1k1jYRImWecMFs48Os4LNbGQTm5jGNFZbWaqBlkgP27NNRFWWmJlGA4lCFSbQFsxzTO1m2mrfYA1TmMo0vtEYpN+Cu5K+Z+iBR40hodNSTU/PZofh4jqpR+wICvhsyQuFv9aU4x9Lyd9rDReY3PE9JRpLy6PSkDCqZ+t5g01sYIJqpY8+cmTpydeisWimhbBhUuHWARa0iBHsExWvc8iWer/djoRTU9oHaO9qVSp57g4LLO1gUMCSopARF0mO8N7hFkLgzGj4YrcIm08IwJIovJyRzTqs4Et1cE41vJIRIk17AE6uFCJNqfeYd+Rm4INj9H9MpOBXt4nnEwqKp2W5XnXBEu8saLhthiyY2CjizTNGeIQrWOYq1vrvVWlflmErW+mk0+3+2VH8/5ca5PosTcoKVO0D350OmRckV9c/BI+9KHlLTxczHvVbIcWiApD5gl8X57F2d/S4ebegz6z04/H+yhcsnxkK8vyhpF9DqNy84mDCbWlkunWKOQmLXvlL+N0P4NSj/LXI5+HUiwoct/AKzn//UpShsG1Yt2IGDTU/JRKtZdIEaYl08Dmw8g2f/HLTX+HqS+C6P8l5GYYL2FrmunwNnHA+PPsHea2d3e783dIJD/hLSy7CQbfRcAQOWiQ6pn/9mdsc+a0Dbf8We1cA4bvF2plID91jtgkqtShRbGyqqWEhi960HZCJyamczh/5Axkk6uv0pQkFgzRTQ74SulRX2TGOq2dSRSUJEmOO2U8/T/I4j/Fo0QsboB9P6aaffkxM5jGffdnvLUO9U6cvIdxYQ34wQbBGdjDHttGWzfxjTy2+7uEcwZ+5jWGGywlCGhxUsbM3ShMwUrRVP43AokN1dB3bBw+kPrqRoOspe3P3SjpEei3AetYxjWnU6yYKaOZXrWZ3ro60HcVUFgFVoMK0mBvdzomtD1MdLNCFwXY2s5pVnGCexE3tAVZmpVi+1xLvb1JIisRbAyJDFlRwQFS8SM8+VANHVojc2t8S8AwCgiC5paz2w0JR5SuxKCUg0mVJiPE3E8rJN+tY69KbgrTQgo3F9BCsJ1i2oSuNK94nH9BdBekA4bh/y7nkgwqjXDhguw07bAHkZrqwcNhBm3vtdFHfaOTNjoOAYVzJ96gST3RGGP7QJp7tzJB4yHtvEGZq1BCFmnkRKcR/OQvK8WXXrmqCujF2lw1uJNMrayiCgvbr2yJuKM004aufgnkzR48z1mfUe4+PdxMbNeDKJtGcHbAFyMMGME88lwu+7ZNePEDzGvh694ihkHiFti3PCQVdz8gpb8qLlt8NQ3ogllopeaW+xmeLFs/DEM/QI+Fk8z7x5KERhWOr1sOHvwQzWv7EPtPvZ9OuVgzDoCKmaalZza7NP2DyLOkD9ZlvSQ7ScJPMphLP7YtXC0gWC+W1f02UkjDpVTeI93j5j3wmbmaMUiOvJ+Fkl69YsIR8M/EtmlT/u+09ICwxEYZeSZrxaU0mgaJwdBNNTCy54xzP2mnnc3yeNXo1a35zCz03Pszgz5ewu9BHKBCmcW4HvZFBNA6exNkS9mOAftawuozsInMQkeqneQqvAS9IjWKePMMMcxEXA9KG6VEeYQ2rMVxg3I/9R+U0TTPAB67+BX/57MfJ7x4oJkCmnvkBFuzv65O20kob7aRY6yvbaE89Q2oK5W/ySbIcVyNVAdi0VT9NWBlvemcoZB6Z31mVlXwnsTcdVa9wQstSViem0leoJqQcFhlVTGt6gP5CiH47TlsAagOaLrrYwHpmMJO5kdFi3b8ZgEt2u7qOSliR32+Cg0rU/psCcFSFCFg/V7LBDY/YY4MuozLhyHgG8NFa8QRHhl1HBhFNAsyv6OLxwWl0W4qozpB0smR0gBmhMOvyIVa49XelQ+W0eDaVuLJhjlvojIBQHkipGPMC6xgs1BUFALKY6DFC+J4XG3VdZhtoNKUuMYc0/H0tC/tv8rtlpGw555yGc2rg/DrRUK03pfnvWALpIBu/bfv5NaUkJ7dhG0V2oe16iR0tooAyls1mDk/zVFkaoOD233yrz2SdKY9Se+ApAd9oBLSU6BIISN4vGBQvKRIRFuX2LpFtq4wJaKU9NhKSR7QcXyh7aFjOKRJyb2JKej8ahnihkayMb7lhxsZaUWYBn2HplXc42l+7x1+ED14ooPmV0//CYLIaxzHQQDKtcJwmZtc/AjrD1l1Rbn8AIbS53qDtgHbBPh6VOTh6tNemFDz0jNT/ae2TasYy7Ug7r0ltsgZ9g/DJU8d/vmfdffKY3umzb/+V9h4QllgNNZzLJ7ie64pKMuCHQUEo7QYGi1jMgWPkBgsUsLCIECnzwkKEqH02xeB1f6emsZnAwwn6TmkkN5SFdTto2XMKGZVhDnNZzL7UUssww+xkJ4MMuMX20r3ewmYi7WxlK17TXw9ATUyGGASEUHA9v6GPXrdRreZRHmED6/ko5446/2mz9uNzf3ueFc/cQyY5xB7zDqC9s7ySeQtbGKCPDiaxla0oIOcYFHQe09BorSk4YZQOYxgZQu6nykLypCGVp9mop4duQoSK3iBAsOT3mYgbsCAK59n7ct1AK/HI6+wR6+VIajkv08AXVj7K+iqbtG2A0mwrKNqDiqkRky1sZga+K5F1hESzIQ+/6heg89oRpR3pVPHApNHaqIfH4cZB8SjX56Vkw0tSG65jFUJChx+shJ+1jt5kPZvBDLazFV3y3lDmMJ9sXsFvt0ZZa8UxCKFw2G3laDZzrMePL3rkFq8xsheStZFwpEJAOawg54RZZ3VgY7gNu2z0m0QuHHxd1UrXSzNw85IaPrJDQLBUXzSnRbLu/gT852Rhm76VTWyVTgvPr5CODYYhNXZemUEqLcDT1ghXXybe4Vh2IAezjrX00lsEQwODEzhxXOLam5njnnxzg2huFizxwGxH5thYL2LbSklOMOGKNngygl5sJOLeAKTSclw0LGNVukLYg8MCkODXGiYsyQValtwoJFK+gstIKWhP9zOZgk99w/fKwqEstiOyfzhyH1qwDCJhQbv7nsD39PBvQjylnMQYrCpvrLwj4P/pb44tTwdyA2C6gt4VMbczhwNH7A/nnjL+ug8l4LSL4Zll7vqF4cufhC9+bPxj/hn2HhCOsEaaaKWNbWwj4IKc434FCXIKpzGFqaMAME2Kv3Inm9iExqGWWo7jBDrxkxur77gDMxhEBQKYd28i8Pp6sqdOIdNi0dEf57D602nE71RZRRXn81lu4Hp66SmCRB11HMBB/I6bis/1coc2NjUI4WUFy+mnrwg4ci6abWxlM5toZyKv8DKrWU2EMItZwuToFPY9/Kxx12cLm0mQpJ9+CuTdsHCMYauK9UML6c83YukwBzf+nXh4C0qLZ2Rp+RA7wEIW8QxPk2DYBwQUVVSRJcMCFtJeUjVzVIXi8HgnPVYnlSZE8lmuvvwrPHLBxZzABgzHkk++odhOgKagQ9QUeTNbw68H5LGtIKCHkk19ilsyEDMgaUk+7OB4+fnOCMPpVQKUHlDgfu/32CqIWs0vxqhDfD4N3+6RXn2zQntwUvNuUuGVxc07QoT+4emE6eWQwEYyKsIr1lS01lh6kJiKkdXS8Fe5oJvVfvf5ZhMKjtQI5rUvdxciwG5di0ITxCZPAIXjeoSj3XEL2fD2CMsNgldWMjUkRflbCj7dxjt/hdxgZPSo4d7Uvnex5Loef9EvUbjrv2S8x16EhhopHZjcPv4YIUJ8ivN5leVsZCNxKljMEmrH6bv5VnbUAXDvEwJOc/eAdVukDZJpiEj07Ck+iExsk7yXR2IJuGQipeT5jiPEF8/ryuXFu/RA0bNISPJoQ0kBzllTJd+3dedoAPQsnRWyUUOteKWeCs0DLx7Nx476b3b2RYUe5UBTTR+xyvlgVNI3JC2bdvf5gXHb9jVFqyog5daIlrVtwgU5V4NUM1o1xnTfGPGohIOv+KwIBExsgUnji24B4tE+96ocaxhy/l//OXRO4B9u1PyP2HtAOIZVUkEN1SRJ4bj1fjXUECBIGxNGgaBGcxM30kMPAYIYbg7vVm7hPD5NPcKYyCUSGIEAw9u2MbhpE6wLoNbsxLYLrPmlplBlcEzweOrxqXkRInyGC9jKFnayk0aamMpU/sztZZ5q6VzmsTcgNVYjtUO9XNQ61vEA99PNbrzsyjrWsj8HcnhJS6NSc3DYwHrSpIp5URsbjARhAxJ2nJwOUks1PYl9iYe34OAwaMtzDVVA6QitzOWzgYUsZzm72EE1NbQiSYSYrual3CB/1GtZHJjI1KAgU0D5Ytkbn3qKxzrnMZBvYChbQ3V4iIwVkbxHIE+/DTPNWQDcMAi/6BcVFQMJHWa0dJuwNMwq8TjGi9Y+nBKACQrWkncYFTzfbQub9I4OabEE8LdhOGuHYGUA2FYweHTTgdzUPodpFbuIEKWDDo5J7KBOdWEYBklbXIqwsuh3aqhWw+SpFQKNI2FPGwHEgBIvV2n5uSXgsmQBTRALh0qSNKkBtuhWnDE8QuXOrc6EWgV9FqzNyTpdUi/nnXYkfBzE8+zFvPuAI+LvjARRVSFakwNDwqhsbfTLCg7a5+2PY2AwnwXMZ8HbP2gc228+nHok/GWpAMuEJgGICz8kbMmnl0kotKlB+kB+4jS473HoG5J1CAahtd4nkezY7fYcbJAavZRLOvJAxjBcDzjpAiYiiu2RZsa7t7Bt+OGNApqRkK93+vuHz+GAuU8xtXUjGoVpaCa2VWFUfgWQYv0/P+D3G/TAUinpQtHaKN762s0iKZfNi4caDsn1Gkz4Hey9ThqlBf5ay/NbGiUUGn0b+rBbdsILr/kgCG4O1oKrb3wPCP/tNpPZdNPNBCYU84EZMjTTTJxyl0GjWc4yuunGxCzy2EKEyJPnaZ7iRKR6deqRR7Lj+ecZ3LwZIxwms1cFumBBIIIzrYYNvMHvuIlP8RmiiC7S+jzcPKgo6E7+o6qTaa5c0i52EiQ4qguFgcFE5Fa6kspxSTL99JWFJ0GA7lmeYglLRrWUei0LNyV34VRaKCNG3ExhuHEarTTxYArIY1mN5IizLllDRXgx9RUvEjDkNth2wrzacwJXEefnrbCIRYDfEeWJwhvc6dwpJSZK86xt0Jk9hIsrZpVttIObN6Ns8Ume3HoIB3Q8Tm1kQD6MKkQheTiNdY1YGn43JDWBOae83CGnpQTC+3tQwcKxRMJtOfewkvo8rSmjL3mcTBshjFy0C/7oejKXdMuGUxpuTTmKK3bXs6yi3v0dEk6IsA4SwSak7LIrZipFxPXQsvhh2aAS8AsAtQEhujSZsDIv4UobeS/MCMcZdoJELBO0WSyo98KbJtJH8pY2+FyX3CAoZOzv9ckNxI9bJJScdHNIXhhVA60mXDYG+/btWG11eSfz/wmzHLn5uSsh3vmna8tzv+OZUvClT0jLo+VrxFOLReHMS3wQ6xmUUN53vwAXfVj6DZ50oZQjFCzYvltAw3Tr++buIV7g7l63qa0n9O3mDDO5cu/LMH2w9EosSgFHKQHKtiYRyvbCxkrBcLKaj1x1Mwft9QTzprxGItPBr753BJiywAculA4UGgHDrj55/WkdMh5I+Up7M1xwptQU/uC3Aup1bri2YJXMy1O2R57TVC/n/tEPvD0QBCnpgNHqPUFTBAb+lfYeEI5hM5nFDraz0fWoDBRxKoqNcz1zcLiHu3mV5UVAKlAgTKTYVb4X/4pOP+44Xv397+nbvIFcq4EezEDQwPn2IRA0UFmHRHCYV1nBEvblh71wRY+/6fysH86tgZ+2SreIIYaIEKFU/s3Cos71KPdgOs/xLGnSmJgECGBjEyRIEmm0Vu4tGlhYrGU1C/FvzV/KwAW7oDk2zFQNIaNAxgkRUQVM5YY1VQWfbN5IamgxA7bU6P1+8FheHNyXJl4m0T3Mjh2NBCoCdE0s0NcYLGNqJnWKu5xHyDtBAgTRWrP5wQQP3HoXt/Y/x4kHT+HjH19AW1slNZMns9/dv+Sxo84iYVWzdMOxVIeHUEaBgUAHT0+RT2LGgW152cy9iJR3thYCgjss2TC/3zR2p4u0I39Pa++aj1Z18b5rpKPEurzUAG4viCB4qUWU3NxoDbcPw0/7Ybeup8eqpdEeYkZgB2GVJ+lEaDb6mRhuZUVOxsm4RJUaw89vZhwB04AhWqpeS6xq13tdnTeZFooWZdUUfojTcedTacBWCzZbUhLi3XTYDtwyBBfVw/eaRKd1t+2zZNuD8MQk6ZRxX1KIMufVSjj532F5R+TdXithM96dgMvq4etN4x5WNKVg5hR5rNsEh3xUiuHjbtkESsDsN3+Cz54Fn/2OdFkvhvUKoiBTXyVNbivcpi17zhCvsHdARLTzlqtTOoJ45dgU36ABQ0poymTVlISNg66k2czJUuRu2eKNWVaY51YfQW/yCCZNAFUSuAoE4NovC+t06dMyl2WrpY9gqWktMmlPviyeo2FIUb/Xid6bRymAW7aMB+Jdvl3be7YrVmDJ/DzLWxJa/Vfae0A4hpmYHMFR9NJDr1uQ2077qCT833mEFSwvAxONJkeWmNuCpwWfNxyIRHj/D67ijQcuxulPQEscfcxUaK+Sd3xextnONprz+3JFjxuSc3dbW8vd7mlVsG98f+7kzxQoYGLiyba10kY9DexmN4/wEFVUM8RgUW4tRowzOJuneWLc8w9RTtv6UZ+ExiK6GkNp0ahRJo4TIW5qbBziVGCaGS6s84/7Tb+me10fG7vC9LUvIddaCdoh0t3HeidF/aypxec+b23F0jYBHQMFK29IsPL6BMFKh0Igw913r+WJJ7Zy662nMOnAA5n1y19y1N038OCJH6OggmSdCAYGl9WH2dO9I+22pLOC95n1rpJXluCB45SgeHxjWXNA8mSDGcnNBUo2gNJDvLxLUEl4MqTkZ6uEUOFdw4ghTNSr+wQ8ZoVDFOwUPbqKtBVkgtFNv6oiqWt5JWcyMQCX1sEXXA8zUDJeAPEAGwx5zZCWdGnIgLlhuYnZnBOFGW/qpvuoDcAEE6aGBTCgPMRpujVlf0vApQ1SGnLHsOQLD4vDMRUixL0s66/Hbwfh2mb4yDtM1eXz8OWfwO33i4D1++bDj74Ekye+9bGeXT8oIFjax9By4Jp++HSdeLVvZVpLO6WrrvcL4ZMpyAbE4wmFpLh8y07469/F+zFcwPHYjom0HFsRk03+9fWSb6yICSty/VZAQ78b1jQM3xPU2s07uuzTYncHt6bRq6c0lIDxhCa48U7xXlsbJJ/ZNzhagMCb33EHy2PDVim7GBjyvfKBYZFF27DNb7GUyfmMVY9taiiorRHPctpEAf++IfGWP/pluOOnsHivt17ryriEmH95K5huw+F8AcJh+Nbn3vr4/0kbn0b2/7kpFI00MYvZdNI5CgRz5FjOK5gYhAiW1RJ67ZCCBHkfB5Ydl9kjiv3JvdCX7oc+b28BQQWYCjsiOb96Grh1SO66gyVXyHTJJjcPwixmsT8HFD1ChUErbfwHZwLwFE8wyBApkmU1V2nSPM6jxTxi6f8KFAgSZFaJ9JrW0m6p2oBMoZVMvg2Q/hU2Gls75OwAr2fhzr52ztsB3e4HfEnfZnqzBXZNm0shGidQEGWYXDDM11/cWLYulleAqCCfcFj9uyTRRoNwlcIMm7S0VDI4mOEvf1mFGQpxwnXX8cXgIN+48qMc++CtfHjrcl7ssPhms7+TP5WRcKEHABr/Zwkdijbp5gJc0AWPpxhlhoJvNEJnCOoMd4PAJ4t44zpIni1iwIyQHHdihYCnd1dvaylHOK1SvMGQkkdAKfaOVzAvbFBphjkymsJUFWSIE9IwaMG1/VLL6LjjgCsQjRveDLilHKaEYnNaPMhmU5RnQvhEmjziBc4Kga3gY9Xy+1im8QlA08NweSP8sg1Oq4b/GhAt1LiS14y7G+Wl3eKlvhM77tOyISZSAh4PPg3v+5Aonbxd+4vb9aoUzAOGrNPdw29vjAeekpBgpqREwOsw4SneKOXKrRXKJejAFZlGPMjdfRIuTabkfdPWJGBTVy2h15BbtO+p1HjvzWBAhLcnNMFBC0UYe2ILdLRKSPSxF6TsorYavvN5aXDb2ih5vQ1bJS+J9vOHpdbVA+d+FU6/WLy41RulO31PvwBbfbXMJxxyWz+5a2kY4s3utxd0tsOHTxBQ7uoTJR7LEg9yZw8cfi584j8lJPxWds2l8MPLJD8ZdQvwH71Rwsr/SnvPI/wHbdAtUfC8wQgRcuSK+boGGjmZU6imPAmylS2Y4RBGqoDluRHuSNrQLtllPk+/yWtr93UP5TAWsYgeeokRpYlmLCxe41VeZ6W7QY/ekbawhWqqWcRiXubForpNkCCncnoZ6Cs395TREFOKHYPH4zgh6ipexsBh2IozaIXJOyFe6l9EvwUPp2HFFFjw3EMEauZht3ViWnmsYAhDOzStf501Vc2sTeaYUSFxtMXBiTyQMyhoi8Q2mbMRBI2izpDkXTQaZNkyER+I1tVx0Fe/ykHemmjos6Xgvcr01ylsyCaddnzwcxBQqHJLDcKmHHdtHxwUG0382C8meb87hoV5Oj8CDyTgr0nfq6xSolZzcZ0vqfbrNimEfzIthBaAI2KScztjB7yRg1e0ANnEAEwLRUgBv0gtxHDvwodcIO3NS3i02pRcXcrxFXMq8QgyJfklBAy3WW6o1J1T2mV59jnQb8lafGynjG0jzNNQSdg1ouBMtwPEk2nYZYmHvDACvxscDTwRQ+b2WAqOG0NZZix7aSU8s1w8Jq+pbDAoHsaPb5bN/u1Y/E1IJnHX67r/CWEpdk6ADx0v3kepXXODeHMBE7TthwAdLUBoO3D4EmGOtrn996IlueVMDvaYBN+5CP68VOTj4lFZo9JWUo6W9ke27eYK8T2vyjgcsAA+f46AzeoNcPZlIrEGAk49/XDsp+Av18Jnz5Yu9Jf/WHKRXb3CyJ3RCX/5qZwryPl/7rtSVN83CCgBsK5e+O7Fojhz58MSCvW8U8MV5w2HJIdYVytr0NEqKjO9AxTbK5W2snruVXmtP/7wrYlUnz7j39+k9z0g/ActRpQwYbJkMTCLVPgCBeLE+TTnj6k/Gka6ZgYqotja68uHFLpiUEMtXezizKpavt8r4ZFASWhUIV3CPaukikqq0Ohi54w06TICTakZGJgYrGMtn+eL7Mf+rGE1ESLMYQ7BEWFRkM7v3++TDTtshNgwcDyrhxdzdN1yXrEGGMi1sSGxFwWnmkpDFEh+1g/HxSJMWfYU2aoaOed8jorBPsxchmxFJT3axEsp1KpKDlYH8KjxJNFmG+1YaG0QcOpocpugZbMWkyfXlM1Na7hpEK7sFXmxSkPUTppN8WQ9BqSDfI8ijM8QojJTXEcDtri5xLEaJkwNSXjQs3NrJf9486DUJraHRE3m3oSEkqeF4NJ6uH+SNPtdmYM9I9KotqsAjyUF5Ezc9kkFAfIK14MJK/m/cuftgXjKVbax3XmGlZROdNtyw7LblvAoWnJmFiLG7VnMkDUYdmBV3m807BF+MhoK7vOjCm6aIKHe83ZJqNlyPc0+W9RZLGQTrCrxHqA8HPxW9uxyAUA1wrtSSogob9fOq4NH0jLH0hxqzIAjTVjyH+JRyeDw9Z/BQ7+F2X6EnnWbfXD38lceuDpannvjdwRUvn4+XHCleHzBgE8mCYfg0mvkvVnnNpbetotiXz+QQvahpLA5PbHuqgoB2wvPllCnB8LtLeI9Tm2HzbskhGga4omdfjHc/mP43m/EI6uI+fm8tZtlfh85CV5+XdZ3+Rq/9ZNhiGebTMNPbha2p1JSmtHTL6+Zy/tCAI4Du7oFqCe3i1daei/vzdc0ZQ02bYfX1r2zvMpFjIgAACAASURBVOG/y94Dwn/QKqliGtN5ndeKnes9UDuaY8cV4Z5IB5WedJoLfl6JQweTKFCgiy4ODc/msnq4qk8+zCBvuDOrynv9ebaeN3iMR8mRx3T1TMcyw6VKODgUKFBDzZgd6C0sdujt7MhHSTnN7BeVMNiQK1V2flUTXZkjub9nNMlEIQSKC9//fqY88l3Wp1NUDfYACq0dHGUQaWxkaqT87feB8J7MtSaytHYLmw5bxZalA0xprgCtGRzKEQgYnH76nLJjftEPl3XLTYKNiGRvTwqhJO/IBu+BWwEpRq8EWl2ZNc8yWiTD3skHYmIIvuqSMFZl4ZAt4g0ZwKs5uDsJt06AYythSonn8cM+8daCyJy9jaTfkXxlny3/p+R/ADFTvDgL2egN5FpEkG4XC8Pyt+2WeGZzwuLBDQsJtwhOBcdtwOuui0cA8sKn32+SUOsHK8WjvsgFweaArNMzKQnxenPLIyLm9aa8VysNODT29tdxWoeEHLVTDoZay+b/du34SvhUDVw36K9fRMHv2uAbP5FcXanXOTgMZ10Ky+8oWWO3E4XjFtIHAi5T1obLPw7f/jz86o+SRxxMCHDUVrlA1eEDSHUlRYHq7V3ys1ci4jgCFJYNm3dIqLHCEA+4Z0Ck0Lx83JK94JrL5LidvRJy9XKSkbDk5n58k3z3Si/AD2s+/qKAcDwm+b9d3XJOHrnHawK8diMsmO2uD3ID0DfkK9mYhgDfwDB85j/k5qW+RsK/ZfJwrlcYjYBVkGP+N9h7QPh/YYfzfkIEWcsa8uSJEuVwjmQa42qI00QT89ib5SwrKsAECFBDTbHkwiuI/1oTfKBKvI4k/UyqeYxwZBM3UsOBHMQ0/ED6Ml7BwnKLN/y+h6XmybHZ2NRTXyT0jLQVLON+fT9pnccKQiZWzWuJs3B0M19rgKMrZSO+w827eOE4zzSi41jZ1sbFHz2NFzb3M1jTSCQxgBOIoCZ08LEpjcW6wFKbatZwUqiGw746h7sbXuCvd66mUHCYPLmGyy8/gMmTfRZGyhHvK6/LXxtk8w8j4OeBjYnU2l1YC7cNy/FxQzbvQRu+0vjO6uFK7eLdMl5pqUTakZKEY0tChAkbHk+7rYCUbNgF7eZ/tfQF3GULAaboibjfDQSAosqXb0tpOaeIIfm+zhB8tFrynfckZewCcoMQRsKeXhcLTyVGuetWQH7ZNwbz3XBf1oFn3FwrCBM2h9/P0ChZ4yFb5nVDW3lu+63sqAPE6/E2bIWEC4MB+OK5b38cgB+1Csv1rwkpCzmtSub08UeELFIKtPEovLFZQnxeCcHxh8J1t/sg4V2DiS1SKH77/XDJ1QIm8ahEbPqH4IxjRQTgB9e7DWzBl4/bAsmMDxihoBybysDRB/rhUQMJn1bEffHrZ5fDd38tHlk+L4zKfAFww5YtjUK+KWr8lljBljGqq2DNBl8NJ1/wBbY9Mo5KClO0pkpCpOmMALVhQDQkYzTWSl7y8ZckZDu5XcYZTgGWS/rRAtCVMQm/jmyf9W6194Dw/8JChDicIziAg8iTJ078TQW4QcDoEA5lMpO5j3tJkqCGWqJESZMmRJDp+LGEPSPw2ZbN3MLvKLitd5Mk+CN/4FiOL3a+SJIkgEkOXaxndHBGgaGFTYggR3PMmDWGO9nBPdyNgyLnhDCUJhYYYt/Wm3hxxyVc3W/w/goBwg9UiBfVY1Osrsy54b7Pu5oAex56MA+nsly7YgtPqBYaaqs4p7WCD1SNXpsnV/Vx8a1r2dGTpnKfDqZ9YDHXfmoJM5VFRUUINQKlthZkg3fwm9l6z7AZXfSukXIJB/h2k4Rvd7klA5c3SCeF8WxzXkS6NVIaMnlEBPmljHhmpRZVMn6PJQD0vV54NAXrc1KOEVKi7xlFNrIU4oXtLsgGu1v7IN4eEAZsgNEhyO4CTAvDn9vhoSR8sQv63I0xrEQKroAAWExJT8VVOb9JL+5ruM7BqGa6HlCC5FK9v4EQeLzmx4fE4Jb2cgHzt2OGISHKMy+BFWtkLZob4L/+c3TXiaQDjyQkrDs9JF5geMRHblJIwLDUbGd8wYRStZfPnS1yXzt2C/h4Ht81l0o48TvXyXy9Gr6gASoCdzwkIUBHjx6/sQ5yu4VsYpoCxgWXaHPuyeIRN9ZB4/tEos0LnypDfn/oWbjoHLjyF+VlC7YNu3skZ9ndJ8SVQMAPjVqWhFhfXSNAPPImz3Ek/BsIyGfiaz+VFkn11eIt27b8PxgUEG1tFABPJP1axgWzYON2v4wjFITONskznnOSAPX/BnsPCP8HLOx+vV3zOkScyyd4midZz3pSpGikkYM5ZFTR/oPcj41dVtZQwOLvPMw85mNi0kZbmVi41wjYk4aroRYDRQut7M8BNNM85tye5Rnpbq+981E4BAmYadriq1ifmMuKLCxxO17f1wGnbRfmJW74bc8I/KAPTqiUjb0lHuF7+795ouCBpRs487JHyTuaUNAg88xGhu9fxUXfPp67poWpHGMXqzfF4/G8mfGIEuCroABc3QP3dcK9HQJIUVUOLiPt+gH4bo/k5iJKNv8v1MPZJenKsAFpu/wD5QFzVMEXd0vBfZMJwTA8mxVgq1WyQaW1hHP/q196HlpaGu46yM1GnYJuN/zYbQsTFOT1ggoOicPqHFy2W0DQW4usFqANK18QuwD0DUjtoINfA+kA+0T8MoMhGzblYb+oeIUtAfF4u2x5rlcjGVQC6h+ueecg6FlHKzx9i3Q6SGXFAxtp2wtw5nZ4PeeHPlsCcO9EmPkWRdzv309UY4IB3ytMZ6X336Q2/3mTJsCt18CNd8BLr8v/P/ZBUZ4Bv0VTqQUCApqtjT4IDSflPOJR8bKmTBSA9aInDlDXDMd/U8Bzz1kSQYiOZKEakMqJ4LWhfJ6Al2+0HQmXfv+LcP63BLBAnhOPyfkOJ8u7fJTJo5lS3B8wxUM1DSmf8Mg7BVvGjEclxNvSIH8/dLGEWV/fIMX01ZUS2p06UYD9lCPh0CVvfk3eTfYeEP4bLUKEwzmCgzgEG5vIKJ9CWJ+99I3KOQYwyZAlwTA11LIPS9jOdgrESLk9DBWKCbRzCqdRR13Z8YMM8CIvsI2tRIkxj/nMYjbDJfqfvsnvQVO2X7Pk37MjsHIqvJ6Fq/vh1ax4hDsLcHUvPJGCn7eWHzPSslmLr3zrCZx4iIqo7DJaazJrd7P78fUsbZlZBjqeBZXfcX6keezQsSyDhDL/1C7lE29mz6bhC11uHZULWEkN1/SJLmm7e/wZVfCrAQhqv+A4raXFU7ct+dUWN4dTG4B5YSHQJNxc4YSAm8/L+yFLW8FZVULEWZqEClvCpqXnayEe3yer4TNd4iEXyQvuOBmE1ZrTEi48KC51g1pL2ybPE5xgwj0TZaP8yE7JcTpa1rLRlLEiBsV3YgC5EbG05A+PHiN3/U6troYR71TfruyBFTl5XS8E3WXJjdiyqeX1lSPth5dJmHFXj78ukTBcf+Xo53ZOgG9eOPY4nRMk1xgs2TkLBfGEjj8I7npECtJz7p2Zdl9nSjss3lMIMfU1cO+r0I+oqASi8PJqKLSBs9VnPYOETDvaxPsKh+T943WwUEped/tuOPkI2G8e/O5u8dCmd8LTr8hcPPOK1otkoIAAfUcrvLxS3rfrt/j5TG2AU5DfcwUJxQ4mhCxzyGJ5LH1a5Oea6uDEQ2GPzvGvwbvZ3gPCf9AcHNayhtd4FROTvVnAFKa+9YFjWND9GssUihBBcpQ3/3KQZr1hFzzrqecUTmMFy9nBDsKEmMksZjG7DES11mxa/QL3Nf4dJ2JSHW8gY0htYYIEU5nKNrYSVNqlAIG39XalplJhSJPasjkquXNdlZMSAq0skqQIGxbPZCt4IRNkv9E9iYq2fn0/mZyNURUpGVNhRoIkn91Ez6ljNKRDGJptQUjny8N8GvGmkuO8XlDJBn5/Ej7xFoXfl+/2PUEDAYGcI8zY5zJwqnvZrmqS838yLcDkIPWEt7TDxrxs0qWhqQlBAcD2EFxQC1/tlh6ApcSjnCONhS+vlzxme0DKKIprhIxhKvjFAKzO+0X8XkjVA03LgVhQ+ix2hkQp5ru90GhJCHn/GFzTLJvwZ3ZJ/jdqyHW1HJnbITE4vkrYsj/rh0E3DD4hCNc2je0Nbs4LOWhdHhZH4HP10PgP7Dp5DY+kXNKPu0aOy45dm4dvdUv3i40F8cD3i5WvZVM9vPZX+O0dwkSdPAHOPxNaRyjODGfhptehPwfHTYJFI0Sjv3khnPZ58Z4iri5mwRKm51AK1mwS4AgFxHuzLMnt9Q6I19TVC3UToc9lkVqIhxpvBycPdEJygwCRbQtY/fhLUiOIcqXfIm5pgyGMz6mu6EBrE1z2CfHc3ne2EG9KQ7WW5eZJXTBtqIWZncIuTWflNZMZt3TCzfl5dZH5vPQVrIiK0IEXGj7xMHn8b7f3gPAftNv5I2+wruh5reJ1FrKIYzl+1HM10vEhRYpOJhd1RMczCwsHp6gDujcLeJqn3I7cwjK1sJjGHsWxXuR5nuRJ0qSpoIJDOIS5lLdQsrJZll5yCWs6d5E6thWzL08uuoOmPfckFo6zguWcwZm8zEskSBA2TApao5RDd2oGabuZa1v8OrNSW+vSrPMqw052YmOhFSR0jt/mNrJPbN9x2+PE40HCaByt0cr3R52Cg1EZYcE4Ya8X0kLtrzYh6cY9A0o2wIvq4aoeGBgnXqq0MCHfzO4cErDzOj0oLSHGCPK6pT560JA2Tiuz8EJGyjccpKnu1JBblF1C6x92YE1BmKJf7hZ1lpFybGG3Ju9PwxIe3WoJuHn5UBChg6yGF9zC9gQChlqXe8QBBVc1CwiCMI8PjEtdZKXhg5PjwG1DAvzeXAMGhB14IA2/dVmcH6mV81TAPtHRXTcAHk3Cydv9dlFPpUX95YlO6XLxTkwh6+eZpeUaeOf46wG4dgA6XZWgqAHXtkjNp2fRqNTcffbssV/jwQ1w1lrIBeS9fPXrcMyrcNtRft7u6AMkdPrVn8DmnVILeNZxwp48+BxhgSrlhwo9BmoiLV5kJgu33SU1iqa7DWgHklsh2AwfOB3080JYmd4pjYkXzJYx2puE+Rlz6xLTWdEy/fw55efx5R8LCFbG3H6VefHyHS0eqGUL8aWpDlZtlPxiU72EQL0WURphygYDAvgAv/kWHHNguRza/yv2/+Ap/fPtDdbyBuvcThOyWzg4vMxLLGSfsvxbN7v5A7eQcLM6Bib78z4OZfRtVJ48z/Esa1iNg00jTRzIQRzCYQwxxCpex8ZGo5lEB6cgHS+f4Ske4WEUBiYmSRLcw91oYO8SZf5Xb7mFbc88gz5hMQHbxAyHKWQy9G/YQONsEffLkec8PsOj6hHeUG+AEyCeX8jBan8O6BhbjxMk/2QgTXE99qpCYSjIBLbwBDkOG6erRWdnDXvOrKd/VR+DVXEChkLnbfKOw/tPmMX+Y5Bbt+ThobQQUcLK97hmBiWkeFgMri51iUosq4VtueRN7kdezsAndvm5R68O0SMDmQa8b4x5zQnDr/vhpiFfI3ZmGE6Mw++GfYm1ra7nMi0kczeVhEk91RqQzd50yS5KSTjQ+5+X18OdW2cQDozBL/sFCJU7VwcRxr5/kgDRCVvlNU6ugtOrYcoI0k/WXZ+KEaAcVMJ49azahCPeIhR6fpd43qVM2iEHLuqC+ya9+bEjzcuD/mnYDzt7OVgDCf8qLXPvCIL1xlquvu0pPkkvRqFArKGBqUcdRe3ksWmMlg0fWQOFAMRdMHCA+0Jw46vw8fn+c48/RB4gYHfKRQI8/4e9846Po7y+/veZ2b7qxZYlyw0bYzDFYIrB9N4DpoUECBASICEhkEACAVJ+EJJAKIEUCL2TQAJJ6D00Y3oxuNtyk6wurbbPPO8fd0azu2qWMc2vjj/+2Frt1JWeO/fec85tafeYpm0d9Jpkg7z+8RLJ4JJpscGzs2A4PUs7C1YM9t0eTjtG3F56EhIMQQLxs7fBNy+At+bLPaivgRsv6esX+sIbUjJ1M0h3YoQNVFVIifaWX8s5H/0DyQzjCbmWXJPvoF+2NU3JhA/fe3if2VcJI4FwA/Ch49pi4K0WLlt0Ph/2BkKN5h7uppuu3uzOwuJlXqKOWjYnv+T3HM+wjGVEiaJQtNHGozzCcZzA0RzDARxII41UUpU3d+1lJ1t0My4DPxkyvMDzeYFwwb/+Rbi8nPTaNKkJIUiAGQiQaG3FsrJoUxMlSoQIh3GEe2F96ZD9YKcwlPvSrM36iZoZNIqEHSJkpJgUaeA9GgcMhAmV4Pzfb03ivLf4aGE3bVqhleK083bjiv1q+u39/LVDSnYZZPEDCRjzU+IreUtH/wQYhZRRSw3JiPpD0hGQp7Rndm3jmVYngO+X9e9deVM73JqTUWktJJYlKQlqsZw0bVpI3mNrGGPAckuCTZHyxO37ROXe/jsmRJuQcz4Kyczdsu25FcLevaVdsk0DqFJQF4S/jYErW2TifNwWssvjPeJ1et/Y/FFUYSX9wA5L9tv7GWmYOEQ/NRftlgT7aMFnEFIwN9H/NkPhqtESzNdkvSxTIQOLs0iwbM1C1UM3U3XvTaQSKZ5Zsxx/PIbh9xOIRplx2mnsd+WVGAVpzZPLIOb3giA45uQabl+ZHwhz8ezr0NIh5Ug7h6Tklm17S9OWU97EKTcD2Tj4nIcpOwtlE2F2VDxAFy73BO8/+65kYmNr4IU7hfySSIqbzM3/EPed0ZXwg28KczUUlPeABNnSIsn0Eim4/mdw6F4SpMMh6Zk2tXrnqvBs48IhySSnjBdf000ZI4FwA+DDz0AcRTPnli5nKTG680YdmZhksXiN1/ICYSedLGc5YcLY2PjwESZMjBgfM5+d2YUiiplMPsc/Q4YkiT5G2SY+egq6ZLYlPkgl/+sgtnMJVpGJ6s6ifYoeYmzFNgNqC4dCQMFVY3o4tbmJxuQYFFAVaGO/6v8RMjNknOtvygrrr8SAbUOat9TrYlxeo9j3HosDF9awVdcObL3FGIqLB66fPRuTBd2FywrtQbK5eUlvyoKbPbm/6H5g36K+Jd62rASyF3qkZGgqyTZt25kE4Oxrr4hkX3stk3OYFfb8SG9sk/e4wdtlEXYCxU52lNASZD9Jyf6bLe8800g5VyEB8fU4PNkjxy4C6k1YZokUIqmFBPPramGafmeNF6yzQFyJAXarJT26uC1km4DzeS1KwZENkh1uFxTNnWnAr6rFezVuy71KO+f2m/6JxoA43sxNCotzy6Bs5ybkubdZazEE3xBU+eD9SRLAr26VoDreL5l0o8OArVy1lOp7byIZKaLHH6aoOYBy6pNWJsN7d9xB7Y47Mv344/P2nRyAWaW0p1ONpeDS5+CR1aC6xUJuVFYyKcsSMovPKT26cLWDSkk2iENEKYk6/TjACEFRKdx4JvzyaskGqyuE2bt0NZx6EVx5HnzjMGGslhTJ/mYeJ6VY90Y/+Dj88WIZhXT5TeB3bM+0Fv3h1IleEAS46BrpWeZCI8GvphKmTZbZhJPH9x2VtKlhJBBuALZne97jnd5ZhSB9PYXBdo6ZNYi2Tzl/cmGgiJPv8NxCM5100IbliJ19VFGFiUk77QOeiw8fAQJYWHk9OBurTy9yyqGH8s4tt1AcClHzp9W0HlVNYoxJyF/G9uaObME0UqSGlIK00srbvEmWLFswjYlMAmC6v5wTam+mNavwESBqxkFBBptaXcef2+HWDpsMaTJkiJhJ9qpuYHo4jKkMtNL0TG0kxVKKmdB7vDaHxl/tk7IXSGaV644in4Eshi8kvMAHnpjeXZQDSticuXg7DoetlNJdVsvir5CsstiQkmbCFkLIzBCc5MxSM4EHMzKG6NWJIokobJe5/BbTyRLDeH6fhe4y2wfFGq7KELu03OSpG1hsQa2C+qBcR0zDvZ1i0QaeebZ2yr9Xt0oPrzkj8osgnvA5hTjSuNNFfroOnhwn/b9yE/6vRSQLWwXEhGBpGv7YCieU5hNerlgnjGFXyjI1AI/Uw8yw9BGL8BbklIbj19ODtD9ETLhslPQAH+iUrDxkSy82C2zx/qtg26QsG2VbhJI9GEpha41hmmSSST64994+gfDgiRBYAimnHwpOSdyAY2skUM5+ARb6wD8BMOCWDNQ/47E440nZxp1Sr3Ou29X/RcLSe+uJOXP8ykResf/OsE0V/KFBvp73gefKYmvx7Xz9Pbj+IiG8XPFXsTDLdcpJpuD838PKZ+Gxl4Ql6/4OhALwx4vkfFY3wYVXwwOPD3yfTZ/0Hzu6ZEbjpo6RQLgBqGMsuzKbV3m517BaYXAQB1OCpxSf5DjMWFi9o5JsbCys3u+BlFBf5iUSJNCOIB40TTRRSik19COqcqBQzGRHXuVVINtrr2ZhM4vd8t673SmnsHruXJo//pjE2h6yc5P4aoupuelM3uddXudVFAZTmcre7NtvQJzLXJ7mCdwZiG8yj82ZyrEcj4HBwRzKv3wPYxPHckT9EaKUJA7mDx02ttlImjQJrWnOhnlo3Q684ktxVPUixgS7iBBlEQuYze74tJ8b22S4roEEjF0icMUokRt8ks6flu4+1LuBz3a2cR/QbSQDmx6EvXP6W1pLKbTLllKerSQgWkigiWpvUftuGfyhTQJYbmbTbcPFTZId/jtG753LDcjuL5vr7emWW134lQQ/hWSFCfKlEIb7fkMkGCBGBsvT8rBQlnM+Sonn6P8Sknmus6SM7M5hRHv3xcIz6D60ARZNgSNK5C/ANS3w7Uavd/TzZmHJnlkB/+qCK1oly3SnT3ySloeK/9bDfivE8s0tvc0Iwx8G/nFeb5xTIf3hF+NyX8KGPBh1G37SGmxlULV8IcpJydz7qJQi09N3zEjED1fVww9XQ8znPTRtk4IfbA+3r4BFJhSlckwbfLBqTxi7QLLCtDPUGCWZWybjZcWmIQErlYbyYofoErLYa7sGTj50EXvsNJb3lk5DKcWSBgmCynDK+7Zs9+TL8NxccaN59AVHPJ/zmYeC0BOHh54Wg+5Q0NuHZcPZv4Y3HoDv/kLmLRbCDdogLNF1bbD/rnD4Xvnve/oV+NtDcg3HHAAnHfHVzxhHAuEGYh/2ZQbb8xEfYmCwDdv2meoeJcpMdmQeb/SK2100s44ECcKE+Zj5NLCy1xbNHfLrMkSn0r98wDuX/QB4k3mkyRAkwO7sxq4FgTBQVMSRt97KM6/fxoeLX6S4tprsXnUsCqwCXENukzeZRzcxjuW4vO176OEZnkSheuUeNpqFLGA+85nOdLZiOmWU8zIv0k4HtdQyi135fXcVabrIkiJm+ZysN0N3tpjNoou4dc3OnDfueUJmBtthxT4T83Nbp/TG3H7bq3H4XQscVSKLvDskNrey5ZYVY9rTyPkQFuYuIfjjmPyyaEMGlmUkSCrlmIsrz4psjClB5MAo7BuVye2FpKEg8FoCnh0PL8QlY/XhlSr95BguO9vklmxN5xtdzoW4gvH+0F5QxnOnLhSWIW3EV7TUD+GsZGwK+bc/J1oFLM3C3R1wktOC/iQFlxbMxUzb4u96UBFc2yqvud9TSjLAxWkp+c7fTPqRCxz5xOz1mBa/Pggb8PsaWJWGy1skY0tpaJy5B9Nvv4bNulvIxDrBYSK7MHw+phx6aL/7PG06zKqBP82HtjQcXAsnbiGf2xNdYGTz76+ZBR2AXQ+Goudk0K1hSOkz5ZYUHFg2KFsCUzINQX+MI2Y9yc3nX4lhWBAzmT5mDnAhjS2yodae+D0QECszNxCGAvkuMyDsU61l4kUiBcUFD3vLV8MN9whLtKIMVq1zvEad8zSdSoFpwDknwZz9YNpm+bKf838Lf3mQ3h/i516HOx8RZ6CvcjAcCYSfAuWUM7tg3mAhDuRgAoR4kecAKWX68LGUJTzAfXyL03iXdwFNkCBZsr3MUIDNmTqk3MLAYD8OYB/2I0MaP4EBrd6yPpsVs20qZ+9LnDjtNPZ+zw3Ufvws5BM6aKcsh5TzER/2yjq8Yys08D7vMp3pANRRx/Hkd9fjNqRUgrRtOtsJc05rRXmglfmWnwXx0WxRvJhyygkR4r4uWeRz+22jTHgyBn+vF4bk4oxkIyktjEcDCWJ+A0q1BKRyEy6pgllR2CbYl0TTHxkn6qSgthKN3dElcGARrM442afO308WsWqbHIS5E+FXzSK9qPGJ/u/pHrkHhhKSj7uG5f6bRgT5FpLpJQveYxd83fu5KdFvNlteGdJ2yrsVSqQXkTC8k4CuAQKs+8AAYhbgBsJb2vvOxQwYcl9v75BMs7AUrJQQl1ZnJQM8tBj6Dz2fHg1Z6QfX+R0jg/rRrDjjx/j+/BtGRyKku2XisOHzoW2bup13Zsappw64v2lV8Mc9+r4+KiL3ORfuw8e2k+Cak2DWCSJHyLVsK/Tg1TZYVpaLT/wd3z7ydQy/kx5ri6D1D37xnb054tyd89xflJMVZi3Px/TUo+CCqz1zcJAMbVSFDNsttFNzv/5wsQTXyjJHB5nxepquDd2c/eHn3+27jyUN4sMaDkjpFOd6Xn8P7n8MTuyrHPvK4Cscw7842Nh8yAc8zn95lVdI9S5ZfaFQrGEVPnxEiRIkiIlJgAArWckCFrCKlVhYvdPmQ84fjaadVm7iL9zBbbzJPDL0M23TgSuwH8zvtIcebCx6iLGOpn7nFVqOS+kaVhdcy4bjgCikbF/eU2zG9uE3MhT7u4jbflqyNu90jee2hq8xa5nipZ6+mZGbRUWU6PZ2C8sCGDEccTkeydXSEuQurYazKkVT1h+TtNYnQ2qTOb0ct9S6cwhuHCOLuU+Jj+U2QaHvu+/NOmXU7zuWKBMCcGsdzJ8Mz02Am2rhO+Xy+hif50aTe2luVpvRkjnV+fMF8bmIafggKT3LdVkpqT5QB+N8ki22WtKrnB2SQKyRwL5bBKYHhKjkkFWmuQAAIABJREFUPso4a2xeP3VRRkg64NyTfs4BJLDPjuSXd0GydENJmXh98UESjl4JkxfBXsvhye712+6ZHqfUrCAdi7H6jTdYUT6axy+6nvg2M5l27LFsedxxTJszh8P/9jdO+Ne/CJX1Y1M0BE4fC9EiSGQ9dmjcDyWd8PWZEpQ+6afc2J/36FYT3mH/HZ6iojSn9aCkYL7T1KeJOD/T7igonylBLhoW9xYQM4DD9oREQkg4PXFxfHngD7DtFs7PZs6x3Wxxz5le1jp9ipB23CG8RRH41Tlw71V9gyDIrEKNFwTBK80+/PTw7ueXDSMZ4TCRIsnfuIlWWnvF9C/xAt/kFMaSPzPGxmYVq2hkbR/CjJRBbR7nP9hOoSrj/PG+r1nIwt4SaSstrGE1h3NkPzZoQ0NrWJiM0BTowjaSaOy86Obu08bCxEdRTr8TYCu25imeJEu2l5hjo1HAtgzAL3dwUDFs3W0zLxHE0o7KUNnsWv4Kq+MTMQngz2zOK211lJs+KgxoRGzIxttCNHF7YOP8oltUCp6eIAzPZWl4Lga3dUKLJWUoAzitbGj3GKXg9jo4oEFYre4cP4BXEjB9sZiIn1UhC8HD9XD4Sikb4iz6Z5bCtytk+z+0wDspCa4/qpQA+MtRIvJvycLZa0RMn9v/A1nQx/jgmfEyxumqFri/O7/sG0aCbkNWsrGokhJlSksAdElBBvB+GvaMiA/saMeMuc4vAf3cSjhzTb4xuVtSNhXc3Sn92KNL4NaO/AzYck54Ton0av/TLYE34JybDZw1DN/ReXG592ktDzONWZizCv5YA6cO8dm516u1Tedrr1C2ZgXZsgrWTt6K4GZT6PjwTY666y6qthi8vTAUZobg8inwa59kXDZCWrp7F8kCH3rKIW758jPCQvhMCPkNLFvTGRNBfi90jGzPI9xz0Xs8/86e3PXU12npqkRp+eyO3h+23lzeahgyh/D9T+CpV0UQf/xBMmi4bpRYrXXHvRJqMg3TJ8Npc2Qm4SvvSHY5fXMplW41Ge76rUgmBkJkkAeb6DAeer6MULqw0LwJYObMmfrNN9/8TPb9Tx7iXd4BvMChUJRTzg/4EZ108h7vsoylrGIVVk6p00+AQG9vzSZJklrqyJJlHU0A5E6LMDAIEeoNilksSillDsdQS4H30xCI2XBuIyy2mplZcxt+I46pLAzlBcPc4FpFFWdzTh+P03m8wZM83kuWUSimsgXHcNyQkzcSdpZvtM5jrWUTMJKMC68kZYd5eu3hzA5HaLSEANFuCzW+woCFmfxspcyAx8fBTgOoPBoy4m8a13BwVGYGri+SFuy+HN5OeQJ6kKBSacC4gPQKKwxZoLcNSEDaJQwVPvFb3WeF3Gs3mwspeHScZGO2hsub4RctfYk9OO+fUwwP1HvntNsSOZ8gzjxBJUHHQo5f65eAuCItLNFc8XrMFinDpIAQatxjHFUCF1bCQ13wDSfp9ylHKoLoBcf55by1hhNXie9oLo4tkYcHgIY0/KIZXopLGfr75XByWf9ZRX/Ya5mUN3PPPWlD1IRVkwfvPb2VgO+u0Wx+y1VU/ONOME2UZdE9fjJ7+dNkmtcx/cQTmfWjH63fyQyBdktmTxYp2MZxeLnkerj+bsnKTFM+Zyubn0kbhmOPFgafmebuCw5hyoQMY2vkYbOtrY1UsoFgIMWSNRMpjXTS1FHP92+4n7RVKUzRc+HQPaWcedPf4cEnZJ+nHgXHHZx/nu/MF6bpR4vlnA7YFf50qQS/VBr+/gQ8It0ajtgbjj3Is00bCJ3dMGE/b9QSyMzBVAaeuBlmbz/49l8GKKXe0lrPLHx9JCMcBjSaD/mgNwC4QcvGpoMOlrOcZ3maNCnWsQ4Lq5dYkiZNhrQjAhZhvR8/cXpkSG8O3IDkHsd9zUCRJEE77cMOhDe3iY3XFkUtaLuIjB1G+VsxcvIN93jllDOH4/odLrwjOzGZKbzNW2RIswVbMoEJ63UOYcPH9WW7cEFLC6szbcyLbcGy5BgUBi/HxbtSIYEnhizwOF8Hke/12DJrbqBAOM4P3xx+5QuQkuCSjOMcgxfMbER6EEuKE8xHGRlWe2IpXFvjZUrnNErwyV3Qe2w4ey28Owm+t1YcZ3IXSI0nnzCAHzvjg1od0XjQcIKU2weyvYzP7zACqw1424bSgsATRkgqz4+H/8aEiLNf1BshdVQJbNsKKzKSUbq6PAuY4WQGSsG9Y+Hf3R5796QyODSHiDHOKQVvKD5M9/VsCCqZfrEqK/sfCNuH4LC3nqbpX/fRWVWD9vlRWlOz8AM6TZtodXW/LNGhMH8JXHodvDkfqsrgR6fANw6XQL9bAeEnFJSyYsxJrw0F2vEKdWEaUoIESKYCXHrHVTz8fz+kuTXFL247nXkfiwRpbHUTX9vtIaaMXUJN+UpOP/ROHpn7IxJJmWZvWbD3t4SY47I8X3pTssJcA/EZW8L/nODs9wnZpvfeBuCbR8jf4aC0WI5xxqWyX5BzOO+Ur0YQHAwjgXAY6KC9d/J74Zw/C4v3eZesU9q0sXslE1myvUN33SkTlVSylrX9agT9+GUUkuM5mptpaTTRAnbq+uDRmJQTs3YUUGTsUtLJEpTZTU2gg7TKUEopOzCTmew4KEGnnHL2dZiqw8VYv+KemmreTlZzwiqnv4eIouX6vKDghuhe2YDz/7+2iTn0tAAcXrTx2GoL0vlDfiG/h5ZEWJU+JFjc2iH39DLHuPntpDiz5CKihEH5RkImPgSUlBYLx0aZSB91fEAC57yEcw+0837tHFd72slqU/pxrpi8W0PIkqBpukxSDUeslACKkizwqtEwPSTHPL7E0xHWOEEwpODUnIeJxqxMrfjrGMl8NzaKlUxiKBxhZaihj6eBpnvvRClFqEceKJVh0j56LNHlnxAsTTNhr72GdT4fL4U9ThL3Fr9PhO3fuUzIIpd+r+/7D9kd/vGkvC9jOeORnCBYHPWG3HbFJIOzLFjauB1XP/IYb37QRUc3jC57F4XJuo4q/vzo2Vz0zcsJ+OOMq3qdZAp+8yMZinvXoxIEo2GvP2dZQlb50SlS4sxFdMP8MQbEnANg/1nwwBNyf47aD+rHbNxjfBEYIcsMA2nSAxpHA7TQQohQr7bQ1Q3a2L29vyBBAgRZw5o+wTT3OO4+kiRJkXICo0U55dRT3+92g8GdYxZP15GxSvAZ8pScyhZTQSX11HMW32N39hiSpToYkjY82Al3dIhTS39QChamZIEtUmI5lUvnd2RnvchlS6aANg2/XAcnrobNl8CS3NETQ+DpbrhknTBO4wU8oSpTSCUuIceFm4FpJDN1y4g+4J/dXtnRr/qSI1w3mveSEmQNJGP04f3yKYSAc1cdfH+tDPkdZUqgCxjyt8uGFttjkm4blPv0VkKG1Lr3LYaUThOWN/xXIwL4alOu+ZxGYZAe1CDDgm0kI1yUkkB+R630NZM2HLkCNl8MhzTAhMXw7dX9E0A+Db5T5mgcnf3aWs59n3B+dt0fXnrzPRLrmtCWJdUWLXVJbVn0BMOMnz2b+l13Hdb5XHyNLPJFUem5RSKSUV17l5BTCjFjS5k+scUksTOzndJ43WiYvQPsvK0ELq2d4KVk4vtjL0d57YMxtHZGUUqjlKa8qJOsbdLaOZbJtauZMWUVj/+5mz13lGM9+rzzsJhzX0xTXnv0uWFd5gajpBjOOBZ+cNKmEQRhJBAOCxVUDjguqYgifPjoIe5YsHlyBPAyyDhxOgZxiimERlxrkiSppILj+Xq/JcuhcFCRkCnAZGXbEcTTdWD0UB3ooVJVcSRHDWu4cH94vBvGLYLT1kgZcOJiuLq5//e+kcjXvQ32g+isHfmkEUOIImuzcPzKoRfntC29qK+tkgzoJ00wcZFMsHAxKyyuKIbyxO6uDhAkU3L7XraWrMuvREAOcGhUAlUv89RZ0GdHoMKURd1GAmmp83UQqPfBBVXCmnwsJtfUbcux3CkLCiGj+JEgui4rmWaH7bnguLCQgFhpQr3fGaablSxzQVpmRu65XBiXPbbYo2mEEJJwzhng22vE2Dyo5AHBD9zbBb8c4DPdUFxYBd8okWAYt+Ucdg3DPUM876W6unjqZxexesvtCSR6em+8Mgz8iRjmZptz4LXXooZZMnjzI8e0OgcBvwjm31vQ/zanzYGn/iaszZ+dAdtsATOmSUm0ohT23BGmToC6apgyDmqdae9+P7R3F2NZAeQOZFBkaWqPUhptZHz1G5TEx0DLbMgso3SAYpDCk1aMYPgYKY0OA378TGIzPmY+4AW3ECH8+GmjjW66elmeA2Gw73mQOYQGBhFnYv2BHEwJpRt07mdVwLtJYRBaupjVzUdQ70/w4zE29UQ2iIWai5gN31wtJTxXbJ6x4bIWGQS7Y0GJZmJQFtgsssCGDVnwQXpbUSU9Qzc7dNmI4LjMaOmRhZAZdB+lYOtBGG9XNMuooqKcYNZjS1a5eIp8HTRkfuCPGsWxpNuSQDcrLF6dLU7AsJzz2syROJQ713vDGJkJ+GFKPCo1MMkPd9TJOY/2QXdGJAk+JJgGFHytGH7bIvv0KTmvD1KSJb6f9LLKsAE+LcE2g5ThLLw5ie7DgoUMAZ5TIsShtxKeY43b88TZptd5x5YhvREFbyVh8wD8JyafhTtU2afEreaWDmHBbiwYBtxcB1dk5Xo3C9A7LmowfPzwwxS9O5c1R5/J2PffoKRpFZbPj2FlyUaK2O78n2CYw39orCqHti7IPQXXM7RuEL/VUZXixBIMwJOvyDZtnWJqrbVMk2jvlAG3BpI9alscYuKpKCWRFFprbG2ww5S3nL1GQfsh8w607cuZx83n/sdCpDNesE4kJWP95ldYx/dFYyQQDhN7sTdNNJEmDWjChEmRpoceSiihjHK66NwIR9K91mxVVNFFF7EBR80OjXIT7h4Lr8RFajDeD7Mj4V4SxqfFPR1SXst1XPEbUl77c3vfQHhUMdzaLv6VNvKD6NqigdD6a0wJTl22BNpkDpmmUA+YO9WhP9zX5Ux+z9ku4mSUHyS9IFrvh3/Ui8wBxOgZZFL9nJVynIiSAGcqCW47OJXkIhPmToIXuuHxuIyEOqXM62H+cQz8pFGszLosKVf+ZhTc3C6fDwhxxYcEug+TXinUdq7RjzxAJDVMCUBLMl8H6Nq3KSSY/btbgqCBZ+rtQufc8wxi01bkF51h3JaHmkJjcp/qe69bs2IrF1XSsw0NP/YAcj/2HUb7e/mLLxIpLmafO67iibN/waiP36F68UfEKkcRrarmqP0HN7sYCOeeDGf9UtiZfr8EtJ4E7DR9/UqBs7aFyePgiZdlxp92sntXJG9qefho7oBIELSdIp4KoYw6unv8bDX+ffbY5iV6vYmMIOgo2E3MnPww//eDE7nsRo+wEg7C7VdIyXIEG4aRQDhMVDOKgziE//EiadJ00UWCOBpNJx0oFEUUY2H1kmQyZNYzC8yHS5TRyDT6XJeXDUFAwd5R+bux0WXT7xUqxLezEPV+uKUWftYko5NSGvaOiPi9ISPf3zkE31wjWWyVKZlNi6MP9CNP2RktC+hWQ1R1c0uchefXn51ZVcFvxqwIPDVe7MZWOZ4GWwbh16PynWme7oYz1orLC8Dv2uCeOtg+LObXT0+QsUw+JWXYmC1OLjXO8Wp8EpzRHoFIIb+oGglYppbs0M3u3ExP4WXYQUOIMNe09XWmcaELtjOQ8u+eURlTVWkKczN3HFNSy3W7+M06uLLN0xZGDel1HjB8PtewEamqIlxeTv3iDznh0tNYscMexMNFbDHvBb519mn4fSZ2Nttn5NJQOPlIIca4kgiAHafDw9ev3/Z+v4wueugpQDvT5p2b795K05QAG0/CDtOamFK7CkuX8vX93ubIXW7AxsSybUzD3VCJFVPmE849BU75mvQLQ0H42j7SyxwOOrpES5hKy7VtKr2+DcWXSkeolDofuAqo1lq3KKUUcB1wCJIkfEtr/fZQ+/ksdYQuLCye4kle5zVwNHWQry3sz7VlOFAowoQppYxKKpnDsRvUH/w88HESdlwmi6ZbSnMnIPxpjGRG/UFrCRphBUWm5mNrNW9aS4n4utnSmEBFdio3tPl4OiaL8LK051yikezs2hr4xhCSiXPXwk0d+aXRhC3s04YhtGqF57vKcXOpKVhfV6Vh66Vyfi57NK5lIsSyKTKQF+DvnXBnp5z7hZXw43Ve6VNr6f8tdEqoAeQeuqVPV2pxaim8nZD3dRf8CvuB22vhxDLYe5nM/1PKC/qunVoQr+ysEV3izXVehntnu4xj0nhZqk/Bv8aKafnLPUK4CeBZsCVsuc6lkyVD/iyxet48/nvWWZihEO2LF5Noa0PbNoGiIopqasgkEkQqKphy6KHMOu88whUVw9p/PA7vfAJjR8P4YchDkimYcTSsafb0dq3tjoWZ8ub9gQTD752Y4rrv7s/a5gBLVoWpLmmgrmopAGmrjMrStGMNk4Cy+7ACh7N0pQTTiWPXX6/p4pLr4Zo7pOcZCcmcw3NPhm8fM7z9fBXxpdcRKqXqgQOAhpyXDwamOH93Bv7s/PuFYxnLeIPXKcyDCjWGhTCciQwu+ntvrqm1QjGdrdmRnT63INhFF8tZRilljGf8em0zLSS6uns6pT8GkoFsF4STSgbeTikZpZPV8KOW1TzS5UOxOYbSzCqbz5yyhfxi1OH8ypn7023DDW1ivl3vhzPLYbv1ILn+qlrIIcsyDnsPyZBvGjNM+YXSBPxNpEiRoiaPYPSnNmGG5jIdo0rO+c4OOL0Udl4uAnkX/+iGQyOw2pb7FXE0g7U+CX7VPpFlxGwvCE7xiWXbOSnYOSwZdbvlZYf7RCQIAhxSLKzShLMI+5HszUDYqMqRchxbDDfX5k/UOLncKd+2SJa+RVDuo6vhvK5VPreAE8CVkmAes+W6vrWBes71Rd2OO7LbhRfyzIUXkuzoQNs22rJIdXaS6urC73iNLnjkEVoXLWLOPfcMKzuMRGC3DdDHtbQDKj9A9WaFWkgyPseBxrbhx6cFaej+ORf9qY2n39qHcCDBfT8/gcl1i2hoqqUovJSgrwd803h32UFcdB00t0Frp1ikXX0B7Dtr/c7t59fCb26W/xuGuM8sWCaBcdftYMvJg2+/qeJLEwiBa4ALgEdyXjsSuFNL2vq6UqpMKTVGa732CznDHDyLmOu58watHAHAYGVQNwjmCu0LYWAwjvH0EGMP9mIrx8z6s4ZG828e4X3ec65AU0oZJ3EK5etRlr2pVrRwN3dIb/CYEjirfOhAEyfOtYlX6Yku5ICoojExkeXd0/hf23aU+F5l6+LlbKbkN7TUhIurh39tJT54e5KIwl/skYkSO0VgURru6pBS3ughfhuaWcf93NfL+g0SZG/2ZUd2AmBFtv/ysEb6gte2SxB0+3ggwe2xuGj0Hu6Swbm1JhxSJAN5E7aUGzPaIUUqCJpC3nH9KLcICrkmqSXILcvC8zHJ2k4tk2kSH1oes1YBowyoDUhg/H4FfKO0b2bxVEwcY5ampUx6VIkQeP7YKkF/UcYh5zis1mJH/K+QCfefBzY/7DCevegibMtC2zkVGK3la61JdnbSsWwZq+fNo37WekaMT4GqcmFwrmuVPqPPJ44y6Qy9ATJrOeL4nWHcGJh++AF8vNQm6E8TT0WZc9k/OXG/ezh0l8dpbBvD+HEzadNXc84Vfjq7YcFyCaQKOOAMOOVIuOX/Bs8O43Ep9yolgRkcja4lMwqfff3/30D4pZBPKKWOBFZrrd8r+FYdsDLn61XOa/3t4ztKqTeVUm82N29kfncBEiR62aHg9lnyfwLFCWbg21uYGeZuY2HRQgsTmMQWTNvo5z8Q3mQe7/KOYwrux4+fTjq4j7vXex/HlMKT4+HFiXBOpWgEB0MbrfxV/4XV5gfOSm8xvuhjti5/lZCRZm7nNBryigQbjoABp5fD3+qgzYbzmuDXzXBeoxg9vxofeFsLizu5g3banDsjDzFP8QSrnB/R2RGnX5cTDd3/7xGG29o9so8Ll6yyICXWauN98GJCRgu9FpdssCkr+wkaoh+sMeHGdjizTOzNXolLRpi2RZw+ypQhu8vTUpptsZwxSkg5NIjMO5xgShA8qqTvAvp4Nxy7Sti4FrAmCz9aK5rCC5qkJOuWVN1/u21hCoOQZj4PxBob6WpoyL/pDrKpFIbfT6qzE9uy6F6z5nM5p1AQvj1HprybpgTAbFZ8PEdXCqu0KAwH7CYklzc/hMUNgDKwdIhMNkpbrIZrHzqfE694hg/a34fyW3luXjk9CVi0QoKo3+fMJATu/S/c8+/Bz+vDxU7wLPidNJQzVPjL0yX73PG5ZYRKqWeg3wmzFwMXIWXRDYbW+ibgJpAe4afZ11AwMQkSIkkSH75e8buLMGGSJPP6hoX/d+3XXEiGGEBjo1BMYQoHcXDve7rpIkmSLrp5mzdppZUyStmRndmMjfMYN4+5GM4f75z8tNJKM81UswGp2CBIkuR+7qObLuLZsRgqgwYSlklFsInyQCttmTIibDx7DK3hx41idZZrbRZPwwmrYOFkr5eXiw/1At6O1dGWHE15oJMpRcuJmEnSpHmd1ziGek4tk3JhQ1ZkBhrJ5GaEYf9iyJl41fe8EOLQkz0Q1MLuzHWfSQM+J/MKGrA8Bb9z7NEyeBMYbCXn32PB4w7JuMXyiDIZnBKqhrlJWG6JTOKW2nzG76XrJIN0X/MrkZOstMQJyNU1pvFIOxoJvGeUw2afTpK63sjE49jZrMc+yZ0u62SEyjRRhkHZhAmfz0kBp8+RsuVt/4RVjSK2v/i7MKleAll5idffe+Axr3TqkpdcxJMwy/Gzb+sUd5p0Vsy7Xbj937/+fXDbtLrRTnXGkQH1amJtIdvss8tGvQVfKXxugVBr3a8nl1Jqa2Ai8J5wYxgLvK2U2glYDXk2KmOd175QBAgwhc35gPdJOcHQcmbPhwgTJUICz4Iit1Sa6x/qvS5fZ0j3vlZKKQpFkiTP8jRLWUqcHpIkKaaYKqrpopsneYL9OYApbP6prytFqt/MFqCH2EYPhItYSJweTAxCRpaUNlDYvQurYcSZFE4POZg4F1pLjypkOCbVBXgtAfc6QdBlYrrHa7Xg4W7pdebi4xQc1jCRpqw8cARUmspgO8fW/pfSYGOvV2zIgNcnws/WibenT0nvzdXcnVIOF67zdIHg6QC/Wwb7NDjnZEhJtPBpLoEYPgeAltxrdvYRUdKjXJuRc2m1RAaSgV4bCDf4u3rIWp+Uhx/vlmwe5/Xlmb73z8o5nvstVwbud67/x5Xw8437YzIoAkVFKL8fnc2iDAOda/CpFNlUimBpKTXbbsuY7T8/Q0ylxAi70AwbYIet8r+eMl7eHwk5mZnzutbi4Vnp9FpnTHNGQPVTcTCM/l1vclE3WlxuXnm79xlB/gLHHdTXnu3/J3zhpVGt9Qda61Fa6wla6wlI+XN7rXUj8ChwshLsAnR+GfqDAHuwJxOZSCllhIlQTBE7sTPHcwJbsy31jBuQ3KKRIbxVVDuBRvf+cfEy/6OdNl7gOT7hY7roJE4cC4sOOljCYhpppIN2nufZDZJnFGI8E7EKyrXuyKW6YZp8rw9aaRXLOgWjAjHQJlnbh6XB0oqsHebi8mrKKGNRCn6xDk5cKTZklzdLQFuXk4zf0g4TFsGYhVC3EM5bKxlZLh7q8hZyC8faC9HrWVoCTS7iNhy7ElqzQfwqjV+lsTBpTlXy9LrZ2FoziUm97y/zwZ9roWFzWDoFflvjZZjnlcOMoHdcN+M7txyq/cLoNIBUP5mB5ZxnF9CMJ6L35Xw/7mzf7hBrdo947NVcraH7rxvowgr+l7OIGji+tAN8bi6hJvcxbqwPpgXhnOERM/OQtD3h+vqitL6e8okTUT4fyjTzaryGz0ekqopdzz+fg667DjVceuVGxEeL4evnw1aHwxFnyzBbF/vOgvG1YpBQUuROqBCf0mt+6r1vh63goN0cTaIl98rWIqbXwOF7DX0eD18Hs2eIGN80pYx7zolSov0Cb88Xji8TWaY/PIZIJxYj8olTv9jT8RAhwhyOpYkmeuihkopend8kNmMv9uYObmUpS/vdXgMJ4pRQSicded9zM8H/8G+66CRJstfA29tekyWDjU0jjaxkJeMY96muaV/2YwmLSJDIc8fZh33xsx5WH8NEFVUECJIkScBMMD5k0Z4NY5Oh2K7hwVHTGO/38U5CjKizGhotEX4rDWMDcG0r/HaUMCMvaJIf6IiSAPiXDglkf8mJ4S2WBKa2fhbcFKKfy8VLcemPRZQio0zH5MAia5usTo0mna7m722zuDYrDM6TS2U0Un8wTXhrM3igw5FPGPDTKpEraC1ONe+n+ifcuMgNPjbepAwLyQZN5NpnhWU0080drk1fPhTCCMXZblRBqe3cSji/SYKzaxQOnn7TLYu6GWaRCT+okD5hkeFJaNYHz8bgB43C6PUp2D8Ct9UKwWkoKMPgkBtu4NHTTyfR1gaBAMowKK6r45gHHmD01luv/4l8Rnj9XTj4u9Ir9JmwfA28MA/u/A18bT/J5p6+Bb5xAcxzWuVja+CGn8PUid5+DAOuv1gC5u3/lNf8pigrNp8A554y9LmUlcCztws5ZnWjzDcMf8VnCW4MfKl0hBsLn4eOcH1wI9fT7kysyGWVgpRXFYoQYTrpcPpyXmHSnVIRJUqMGAZGXrm1cF/bMYPDGOZclX7QTTf/40WWs5wiitmV3Zi8nj3ID5NivxW3xTbsoKLBnzKTJHmQ+53+Z6rXmKCGMZzMt3qlCSeuEvp+VouFmbKhh5yhrEgWk0GYi0HnmG42uGIKFJvwaJcs7qszDHAnYU9H9O5mS7e3C6lGAT6le4cnZ2yTIjPDRD+YupisFo/MahPuGwtj+gmGbyUkI22zxNTg8GIJhi7mxuHQBjGlr9A3AAAgAElEQVQgyA1crqA+12auv++FlLjCXFYlUyyOWyX3JqnzTc1NpM+3a1TuaZctNnDTcvp6WsNvW8WXtceW+3FwVEwBLm9xsjekL3lg1JmJmJHj1/ukPDp7PYwbPkrCbsvlPCJKzjOhhRT02qShtvaw7sMPef+ee+hcvpy6WbPY+oQTiI7aiD5wnwK7nigDdKM59yOeEOLM4ifz3xvrkfLoqMrB9/nG+3DjfdDcDgfPhtOO3viTJjZFDKQjHAmEnyHu4S6WsqSXHJOb0fnxo9HMYAfmMTePZeqWSYMEKaKINtr6lE5zESbMOMbzDU76XK6rP1zTIq4rWQAtC+QBUXho7ODyiW66eJ3XWMYyTEy2Yjo7MLNXR5nWMGuZyB3mp6XnlXBuQ2GW49qIFTv9QVfQ/8Ekp2fjjCNamvGsywrhQzR619VIdvJaHE5eLczNsFsSdPZbpOS8lmQl6EqwFG3j7wtoYQ90ioMMWtxamiwR8/9hNBzr9OY+SEr599GY54STmwH6kEwrpfPt0bTzvekB+F0N7F8kLNg3E0J2SWq5bve9FUqmx5ebkh3/tBJ2j8JzMbk3kwMSqCOGlN9WZMXZp9jJGt9NiB9rQ0ZKc42OHZt7rsVKLOgeqIcpQ5BmTl4lmsPinJ8R9/6+OAFmbgLZStlOYPryCS7asW1rfAmKPyeG7Qi+AoL6TRGz2YNlLO2XPWphUUwxO7AD7/Mu6RyijEuoGUMtMWJDOtRoNGOo/WwvZhA0ZeGSZglavQu3hidi8PduOH4Qn/BiStifAwf8vg8pteWW/cDLcHKDhUtCSWoJhBkkS6r1w+9bpOxoQUFunr8fjcwN/HqpyCF2CsMeEZkK4U6et4DRpsgQPs545UGc87u5HS4f7fl0dltwXZuUXT9IuVNAoDsrRuX/7ILzK+HsRpkSEVbeDEbTuR6f81rQKVO6kgU3QF5SJVMc3P7d4rR3TSElr2ecfd1eJ0zWblum17dZ8LWVsMYhyAQU/LUd/lYrZgcTc6riHVk4abWQaWyH3eoGbb9zvG4tGsP7u+CSIYgzn6T6LkJKyc/PR6mNHwhjtvjitlhwYNHnE2jDIUgmIZcyYNkSGIeaCj+CzwdfOFlmU8Z4xnMQhxAm3EdTqNHEiXMrtzCVLfLmHBoYhAgxitEE8FNLLUYB8cYV8gOECDODL25E9L0dkqW5JUoD+X8aEat/GhhKxN4tlhA4bPKJH5CvzbOQMlvKlszptDIJHn9tl0DgkkwGQpEhmrxXnKHmphKz7MtHwa4R2C4EP6uENyZJn9H16HT/gpBW5uZoEhem5Xw7sk5/E4/oYgCPxIRp6nfE1n4lEomAkqAYcrK6Hi1EGAMZSrxVUPqKhxWJP2izJWXMHzeKqXePlmtJOBm66fT7dotAnV+E+AbwtQaZUrE0LbMd2y0pH9/Y1vf+nLracarRklW7Wbn7EOF+FmktpeChsG2474OJ7fws7TDINJENwQs9MHGhlMd/0yKjqI5dOXyCznBx4qGQyoDtXKhty8SIA3YTX9IRfPEYCYSfMWayI9/nh5zEt5jJzF4NYoRIb59wIQuYw3FMZDOqqGZrtuU7nEUnnQQIUkQxk5hEmHCel6lCESXKHI5dL+eXzwrLM14QdOEGw/4Mt4eL08rgm05WWWJ4VmIuXKeWMM6ECaQkeFGlEGk+SXsBJEvfPhvONiVOf1GT37uLGHBauRgF3FknAfW0NV7wd8/F/b+JBD8XJYZkcU1W/n3SeJne20nJGIsNLzD4cKzVnKkTYeUJ47tsyUrbbLFYu2CdZHWHN8DzPRIo3R5q3IaYJX3RY4qhPCcFO79RxlO5rNSElpJnXIslHUipst2C9xLwhPuAQN/FQxf8fyinHhCyUEg516nFFKDHFrLP9I0YCG0bvrFaHiiihvyNKMn0b/6UD2tD4Xc/hoN3h0RKeoOJpMgYbrv8sz3uCNYfI6XRzwFhwkxiEi/wHAYGvpycxIdJ2vlzCt/K266MMlaxEr8zl7CWWpppJkGCCiopppi92Yf6T8kW/bTYJuSxGHMXeZAF7dPCp+C8KhFqL89Iz26+Y1WWdo4bBrSCmQG4rx7G+T2iToczVzCj88uYLgykb2YoWYgDCo7uZ6TN8rSUMhszkm26sxLdrFDhET62ySl5TQ5I9rUyJzi698dlXwaVBM+khqztEWOSzntLlLOPjGR6bZaUesNIVmgY4uryv4SMZ6r1w7bIIN6klntzdjlcmTNPr9MSYwE3m/WIWjJhY7QpBKj/a5FMsdER77ufdX8lZjdbDysZtTUUNgvIA8YPG2WOY9CAE4rhmv6sNz4Fnu2R643kfPiGU4K9vQO++ylkH0PBNOGh62H5anj3Y5g2CaYOgwg0gs8eI4Hwc8Tglmt96ZUzmEEDK0iTJkAAnAxwJjuxLdtRSumg+/wsobVm4X/+w3t33MFqfxGVP/w9bSWVWBi9i3yFgpM2bI5wvyg2YWsT/jsODmmQzCWMI4jXwmi8s64vY3NqUIbn9theWdQt5bmMy04tpSsfcEkl1PWjFrmpXcT1Se2V70AW/yJDAkRWw9bB/N6TUnD1aDjLgn/GvN5f1AmappPprclKYDQUJDNpzHic4mQCu6QEIxJlQVrYlBphaWotPT6FZFPayS6bnDFWY/wyIaPL8Su9qiC4vJ+U7UKIdCT3JzCpYa8onLXWye5MKe1qvIy68CfW/brMkG33W08SyMwwvDJx6Pd9Grg+54UsZoP+x3B9FphQJ39H8OXDSGn0c8QMZgDkkV9EsO7v1xlmDLUcyEEE8BMjRoY02zOT3dmDcsq/sCAI8O7tt/PCLy6juWUF/vY3mfrKPylet5IiEow2NZv74YyK/PJWS1bMmq9rzc+OhotaP/y8SsqILltxrB9+P7p/2UKFKUQSl8zhIqRgy4BHOAkgTNCHu6VfVoj/dEup180GXVcYhcgRSgwJxqeUSobz62bJqEA0e/8YB1dUSwB0yTx+JXKDGkeQ7lMQal4jTinaprirhYqP3yXaKH6mqx32arEh59+QEY/UV+JyLD/5jFilpMy5Tz9ShpDhaP4MOa7bf7URlmilKQGx3NGpj/J7U9tzSUogDyS1PqkA/KxKLNt8fZ/tvjDsF5WsO53vy00WkfqM4P9vjMgnPkdoNA9yPwtZ0PuaicnBHDoo2UWjSZLEjz+PVPNFIZNIcNf++5OJKtoCHayrmcqjh19Ia3ktPaWVGH4/B0ZM7h0rQ2Y/TMnsur91esJsA/hJ1dCswv6wKgPHrJRF3JU0dDrlz/+O698rVGsxvf7JOlnAR5licP1G0snwyGdi7hWGZwqylFELhPGZ+wm4wfD6Gtg2JI4385JCFvEjMoXfj4bDchbbpiw82OkNIz5jLYzxOaXZZIq31jTTVimTUkOpBNq2qFyxiPiUrVChMBMCcE45/LYFFmS8gOSK66tMeVgIKukPprWUiotNKfnOKZEgldGw73IpsaLzhff/HAsvxOVvpendw/dSMo/RhVsOBinJvrPZ+n2GXwTu6BBXotwe8RYByUb7+5kZwaaHEfnElwAKxXGcQAMr+Jj5BAmxHdtRzuANCndA75cFPU1N2NkssUAG2/Tzn8N+TEfZaJKREvyZNPjhmR6DXZYpQoYwOOc7dP4yx3Uko+F3LSLG3mmYQuBnemTBDhvSK1uRkQXfROb+faef26kUnFYhWc3VreLl2epYVLm9PrdsagEvJGSyw7icEqk7Z9ENlrlMyX2jMkbp6R6vV+pXkMzARU2SkbiL7WifTOYACS6lpgSrkILG7h56ikoBjaE1fiuDZZh01k0glOxhUkmYP9TAOJ8E0NzHWLeveHgRbB6Uvt47KUha8kDSZMGVzfBeEq4YLed3U6249riTKsIKLquGA4qlpPp0T/49nOKXQK6QXmpQedrG5qxkqONysnJbi0wjamzcDHFuXCQpa7OiVz2nMn8OZH84pQxmhoQN68onTikdekLKCDZ9jATCzxkKxXgmMJ4JX/SpbDAiVVUopchkUzSO3ZZ4uIR4tAxlZbECIbRTtF2QFgp8TMvCqRB2YKkpi3BKw60dww+EceeRvjUrARZHY5hAgty2Idh5gH0eViyzB1dlRPg9falMasiVVBhIQLm7Ey7KyVgnBKA7lS8eD+C5qVzZ6vX/NHJ9fmQW4aIUbN3Ps4xSoln8SzvUKGjzBVEpC58l+altmBi2TToYpl4nuXesEEwubpLA55Z73YBsIxneX+pEIP/oKgn4LrElYkiJ99QyEbtvERSyyrtJOd9tgiL0BwkUd3VKObbMKUN325K9tliej2nW+XyrTLmvbiB8oFMy5BZLGLHnVMp8yk/rafnXVvjxOo89PDchP0fzJorX62DYKgR/+uIktyP4kmLkWWgTR4+1hNbM8yTtdRttn4GiIrY67jjMdUmylk3GHyKrfKRCRWQDASxtkkIWqk5LFkpXNmDhlUfBC2rDwS5hWQAXpSSwdmmZr5fQ8mR3Q/sQ56+EZFLtF0ZnIe3fRvZTOFz25DLpgxU71PuoksDy/Upha8ZtL+C7eroMEmDMQX7TTi0TWUOrBZlgCMv0UdXWyNh1KzFti6xpEkgnuXyUBEGQrIyc47hyBo1kaTEbLl4Hq2x5QMgg977LhsUZWJjyju9X0tucHfGCIEiGeHIZ7BqSfdb5JVu8pEoCWwZhvPqdXqvf8ILg3zvhjDXC8k1oCaY/bYI/9qNNHA6SNlzU7BktRA3p667OSk92BCPYEIxkhJsoMrqLld3fpig9F41JJ7A2fDwTIpejCidzbgB2OuccMsWK+IO3QSqN5Q+iDO0szAY2Co0svmP8sDyb35txp6UP5jozEGaExPfy1wU+aQpxNAmuh5DbxWXVcFhDvqwigJQx9ysgmJxeJo4tL/V4Djd7RWX00A1tEhS7CyZH2EjJc/NBPMv9SjLP75TDw10m1zRY+JtWoWyb4uZG4pFixtWN5mv1Y7x7EAZ/p9ejdI+lgCOLJeh8kBPs3ADvOtWsHWi0hINb2yVLda+jxJR7NTUoDwh/7xbyU5mT/rbbcEixkJZAJtvn6TGVBMSrWmUYsLGBWeH/euTBInd2olLg0/BYD1y9Ybsdwf/nGAmEmyhWxs6jOPUaKaNCRlLrLEXxu1ljbkZd6Nufev+Gz8fu3z6XrU4/mYamZm5v83IhjSKAZ71VakCtCaudDMv1Cj0wCgdvgM+iUlJydTOgXFcX17TZhW3DHV1wU5vIJw4rEhG3O9nggCI4vkSYou4V+BXsFRG3llyEDNG3LU6Lnm+sTzJKpSQAjPbJhAHXas4NiN8p79sfs53+W5cN+0elpFflg9PLYVW2mP9GdyYV68FAU1sU4YZxgbyS4mmlYhu3NusdCyQ7ei8Br6W8GYSQM4MQz/llILyTgD+3C0nGNR9vt8R8/NF6CX631kqv7fm4HPPsCskeXazMyAMAeDPvglqy3oQN0cHsfQZBuc+5Xp1fYrWRDH0EI9gQjATCTRAZ3U009Swpo0yCIIDykVUhsok7YCMEQhcVqoLTSip4PSaLctLp14WUY7OmRR83yie9wbF+KWmdUAJHl2x4v2h+RhZ6N5PLJa9U5iyy314LD3R5JcRr2+Gf3TIOyc1Wbq+Dw7vg9k7JVOcUwwmlIu7uD5MD8jcXBxaJjZuhxeos6Sz+04NweQEz9vU4HLvKs1vzKbi4En5SLcSTX1bDCaU+PkyWUmLCbmHpg13RIl6kM0NiLvDSBDhzrQwbTmvJlNIa7u32epUu3LjnsjynDeJx+VjMI/skbOiwPX3i/BRsHZIy6RWjB95HiSkeq8opWbtC+xCeB+uGYGZYytNrsyJZUUpK7TZyT0Ywgg3BSCDcBJHVMZS2vSDowFZ+TN210Y+3Y1jo+uWmMDndHmDchl+NksW73BTz6tINzAQKsV1QgkZQ5YwEQsqa+zhZ5pI0PNglpA4zJztZkZVs5idV8pqp4LhS+Qvi09lhS4Yx2OSMXFSY8Ncx8KtmWJCS89k9ApdW5/cHszYcvdJhUipZyNM2/LJFCD57RJ1p5UoC5ktxpw9qeRMjnusR3eDdY+Gx8dCcgcmLnbmBeIEntxTtPijYSBDcbZARSUmH/NKQlh5fVtPb8z1zDdxaJyXS/tBhSf/yuBK5x67rDs72EUOMCb43xJihwfBQvYyranf1Logk5IyyQTcbNpI2/K1dJCS1PvhexcDXPYKvNkYC4SaIkBpNs1mD327DUl7t0W/HiIUP2ujHCyi4oUZKZy1ZWciDCn47um95cWPhwCKY7BfiR1g5GjwtDM6znMzgyZj8mzskVinJ2p6MeYHQRUcWTlwtxBdbQ4UBV9aI6ff6YGoQ7q6Tfpkfb2xRLh7ulnJoRHnZcMCQYHhDmwTCuXE4ZbWwTdEykRpgrShTKDdkYb6rAy6uhoecDDCjhcBSiNwScjGi3RwsK9vPkYKsdMq8Oa1GXkvCHsvhP/WwYw4zN6uFsfuPTnG9abXyg7GJnPOWAbi3S0rAG6rd2yYEKyaLWfnarJSWhxr3NFx0ZGX8V0NOL/WOTrhlDByzEd2SRvDlwEgg3AShlEG4+BfYnd/DsNrRyoehM2SMUkaHf/qZHHNqEB6pl9JZWktJ8LMUKRsKHh8npc+3k7LQTw7ADWM8d5lRA2SfGinVFuLwlfBWUoKUoSRgnbkGxpqw53r2MpWS7DBri2bvPofQckwx/LTaG8FUWBJWiMzA1vDLZmh05AmFvbws0GJDJiuSBxduEHR3m7tZELkntT7RdI7rx30nF7uFZX8xnb8fn7PfbhsubIJnJ3jXcWsH3N8pD0KNlpd9BpHS79SgHB+gMyv39tP8fBgGHFWy4dsPhcuapXJQlPPAkrDhnEYxJVjfSsEIvhoY+Tg3UVT7DyZU9k96woeQ9G9OPHoK5eWPEzU/O7dfn5Kn9Znhz8epozYgpcF3JsHcifD6RJkf6OJrxVKK7bGlJAqSeSngnALR/cdJeCcp5UrTWfxCzqSLK1uHf24HNMCvWmBZRkTmv2uTTOpAJ4vKlZC4ZJJDi6SMvCrjTLFQfQcIu73OTlsyVvc6+zPAxnlvqQlbBqXEuigj5zFmAZy/VgJ2IZ6PS1k7N8lyNYru8RdlZH/u+d/bKftaY3lZoEaySVOJ2B4kmBQb8rCQi49SMifw3WEwfj9LPBaT7Dv3gSWk5CFg7kBTnUfwlcVIRrgJo9S3LaVFf/qiT2PYyNrw75gEsEOKoGKIn9L+/EVBHEMeqRc7thann+RXcGUVzCoQ3C9ynG8Kaf0+xLkGYEVaJkRU+0R4PpAE4OluIbfkZhNay2I/LyVuJrd3CpHEHQ01yQ9nVUgm5lMeM3KgkVEakZGAsDjHGLDazs/g3JIoiHPMJ2m5nogh/a8/d8CSDPyrYHjJQ10SPNsdhkuuFMRl6AaV97BjIT3MlRkvWOaed8IpFXdYok38RbXHok3acEQDvJrwtpsRkmy/v9Ly5wW/yr+XLjSekcAINh2MZIQj+FLhuR4Yu0imoJ+5FiYuht99CqH0zDAsnQxPjIP7x8LqzeH7Bb3BpA0fOJ6j3Vb+Ip5FiDmXNMkg12NWwW7LYOrigbOXZ5ymXm42oZyF9dkY3Fgr8oOdwlIy/FklzJ0kAarUlKkNblm0MBa4u4woONjxLw0ZsH+xlHD9SCZXhjA0R5tC2KnxyeJeYkoQChiS/T7TI6QiF8/HhGC0JC19P8iXZwQdQs43Szxphc9hoTot2TwdJcj9LDWlj3ddDRye47v6w0Z4OSHXE3HO6e0knL6m/3u7IchmZQ7gcGyVTy51Zlfmmj9oKe9u9+VxOxzBRsJIRjiCLw2SNhy/Ml8wnbGlxLhbZHCm42AwDJg9wLZrMxLYmpwAmAASltivWUg5bNcw/LJVyoXuyJ4lGSFT/KMeDi2YXjDK7DuiCOS10U72enzpwGYCF1cJI/I/3ZCyvG0jSgKjhegvx+dIOC6sgsUpCU7tjtF2mQm318L+RfDntr6/7O48vncS4ljzw7VwY7sndXBZuNmcrw3g2P/X3plHyVVXif9zq6q7eu90upOQpLOHhMWEDCRsAkrYkSFz+AGGfVNGRh0UFESQkZGABn4KHhXEgCPIKkRQFIGMiDoDgSQQ9oQOwexLJ+k1vVXXd/64r6jX1VXd6Sxdr7vu55x3qupV1avb36569929FG5IKQm5Zig85ktI9pdrhIGXxqePTS5sVIWduGgQgUIHzzfrQOKeOvL0Rls7/OxRePIFaGuDKRPg+ivgkAN6f+91lfCyl7Wb6DFbGoJHq3dfHiO4mCI0AsPjDZr67+8akuc17b5nR98V4cft8NsGbb81q1AVlr8xc7uDy9ZrnWOpV7LQ7DQu1+x0MsSdI+DGrerea3PJsUshtL3YtZv12P7kmyuGaM1fSzxZVN7quWX/1UvxX9Gm7sTDCronXpSH4Rej1BW7ql1HQv1kuyah5IvGEuen1PAdEIUnxsAz3gipT0XhzDKv8wuaSPRRSkppwvV6YBQW1msdpEOtyg70uQ7UwqzO07jquWWwXxqFNqNQ+60mCvwhqTwPimZO0Glz3TNYw6Lfg066W8R94bb74Nm/QFUFVJTBmo3wb/8JD9/R+1zAUAieGwev7FQvxag8OK/MplQMVkwRGoEhtbennx09PJeO11rUwomhJ9qXdqqifWCUZizeXQtPN2n8zwGNTpVkaQhKnCrCR6s1Hrgjlox9+Tu0xIF17XD/jq4W0pAIPFENF2/QXqugx713pCa5nFKT/NyyEPxwBJyXpgZuXL5us4HLK7Q+L+E+TcfIPPhShkEmN1SpGzShnBP1hocVaCPqqzclY4CQbFYQRzsD3TIMzu+hTi8scH45/KZBE0panMZIy8LaSzUTB0d1QHCJTxnudDAhT923u0vtDnjurzCiKmlVVpTB5lp48nn4xuW7dpyjirrHk43BhylCIzCcUaqNomMumUyRiOuc2YfhqXGnhe35ogp0rVcUvrINrtqgSSMr25NuRlDLpzGuJ/0EiR/H8cVaVpEoCfC36YwBC+o01neWL53/hBJYO1mTQGLAcV5caUKNJu4k2oHtjMO/blRlNN03xDiVkGROCnq3TS3MmYWZLa9ZRfDQaPjGZrWAwwKnFqvrFFRBR+hah5jQt0VhOHoXlME1lZp09I8OzYoNiSq6y3vo+HLXCDhtjWagJv4feaKxxD1hcy1Ewt1dqwVRWLV2z45tDD5MERqBYVK+Jin8qr7r7L8Doz1bFalsimnsb02H1twlaHfweKPO8hPU6kuQcAN2OHV5HhzVvpYA11XBf9VpaUAqHehnXb5BldoF5cl4V2ps8qkGrSNMdJQBtc6a4nDXNu3YAjrE+PdNakVePkTdcumoj8EZa7XsI2Gpfq4EHh2dvs5tTpluDTF18fktrtNLVKGGXbJtXSJWOLdUR1D1RlVErehXW2BDh75nZkHPDbaPKNKyl9tr1TI8IF/rLaf1cFGwK4wZ6U2gj0HEd5ZraYXpU/fs2LtMeysgkG/taIKOKUIjUPxsFMwuhl/UafnEWWXwbxV9c5O1xLUxdkOaLEGH9gJtcV0H7Caea3SaYfmgb2ZdeVibbV+2Qd12fuJorVxbXMcO/c9O+NnI9D1UN3jmVrpi+rUd2oj7nHXwJ98w3Du2wb37pXedXrReLdVEmUanUwX6na0wr6c+oGl+9ddWqut4TYdehCSK8+eWwq09HCuViOg4p74wJQq/7CVm11fKSuCiObDgSb0fzYPtDVBeCmefvBc+oHE7fLgM8vJh6uGQ79PcH70FP7oC1ryn/5jJh8E3H4QR4/bCBxv7AnF9ySkeIMycOdMtWbIk22IYWeL2Wphfq/1C05FQfv4uLIn7I8OwajIUpMThvr0ZXmqG9R3q2oyRdCMm9EpCIf33OB2TlMpbrZppmuhcA2q1NDmd8Tc6D67a2PX51ri6MddM6Zro0xKH4Ss1kcXfQq41rnG5tVO6frZzOih5SYvGGY8rUgvOT1OnWqZ/aNKC969U7t50kKDgHDzzZ3jkWdheD0fPgCvPheo9dLvy96fg8R/AznroaIOiUrj6F3DgkdBUB1+YCi1NEPWuCNqaoXw4PFDT1Tw1+h0RWeqcm5m63/4rxqBjUZMWqC9ry/wavyWIdz+KKoB0UydKw+oynBCFhtZkh5pEjV2BaGyr2WnKfTpFOL0ATi7W0oCIV6jegVqgX66A/7dOX+d3JRaE1DJ+ugEu9FmFjXGNhXZrACD6ej/O6wP6eEMy6edOgfnDu7puS8Jw03DdBgMi8C8n6LbX2FADj94KtRsg1qb//IZtcPMZ8JOl8NLDqgQLfVcQBSVQXwt/eQROvHgvCmPsLSwZ2Bh0FHmTyzOl3pdJ+i9+G3BIQXq35hklSYV3YH6ya4ugCrRIkoNvS3v4VT1VrRmY1XmqdC8q1xhZWSR9F5lMVIVgWFjLMvy0Ou1642dpq9b4DQtrQfjIiP4dN25VC9LoA8sWwY7NagnmRTX+V1Ck8cAHvwNrV3Sv3BfvkmntB1kR2egdU4TGoOPzZepuHBvpqgxDwOgwLKzumq6fQIDlGazIaQUaR6v3Cu/H5+nxSlBLyqHJOFVeZ5hMhEI69eK9ybB6Cvx8VLIG8XyvwN7fzaQ9rlbemSXdj3PnCP3cpri6Spu80og7U2J6i5pVVr8Ltdirz1xufTP7Rkcb7GyEvJQMpnAEVrwOE6fr484YxLzAr/MumSYd2u/iGruGuUaNQcfnyzUe9scmtfJ2dCYtuW8Pg+NKNPkm6munli9QiBbhb+/s3hQadFjvKSXwjjeh4oVmmL9NXZGC1hz+eETv0x0ycVm5Zpa+vNObwI4qr7v3S5/gcna5ZpTeXgur22FGAdxUBQekZFxG0PKRmnZt6N3hNS2ozNABx+iBTx3rpdT6+t+1t0LrTti6BpYuAheHph2+odghGDkJjjkrW1Ibve20uJsAABRCSURBVGDJMsagZU2HKraCkLoSq/NV4TkHo1d6Exb8Q3O9n8LrE2FiPnx3C9y7QxXd9CjcMxKmp8T+muLwlyY9zpGFXRtFv9EC/7VD44BnlyUHBvdEPK4xxGcboTyiXWom7ULpQk/8sVGzUVMzXiPAT/eDKzMU4RtpcA7+45/hzZcgHIZYu1qJIhAthlBYY4eRqCbJOCBaCBNnQGsTTD4UPn8DjJqU7b8kJ8mULGOK0MhJvrheB8Qmptd3OnVtjs+DJZPgnDXw++ZkN5k4qkSXTtDi9964ZYuWPiSSaQSdov7gXi4T2BW+tB4W1Hcf1RRBrdf3J/c8qDfBkhbNmv10mozTPaUpru3w/tys3XbOKde5iOnitVmnpQnuvBSWvQAtjd2fF4Ehw2HKLI0nfrhUfdn5hRDv1NvbX1ClaPQrmRShxQiNnOS24bB/niq/FqflECPC8L3hsCUGf2hWz1cEvc1DX/uVjb0f+6N2uHObJtEk2rYViro9/5TmvLkv6XBal5iuQ51D5wT21NoOYE07TKvR6RsXroeJNfDNTXtPxpY4XPg23Pw0vPIn+OtH2vLtl3V77zP2KoUlcNY1UDmKtM5l5zRLdO0HUPOGPg5F1DIsLIH2Frj36/0utpEZixEaOcmwPHhxPDxcr+UO1WG4qEIbaP/Us+TSjUB6s4eSjASP16sFGfFdZoY9l+xjDXBqH9rF7SniuvZp9ddOJjrHlPdyOXzmWp22kaiTjMXhZzs0gejiPnT8ScU57Ql77R/h1YcgL64lKKHfw6hT4L5TdAp9RRbnEmakZqneivePTdwmiHfC9k0Q9xryxTp0C4UhrwBWv9X/MhsZCYxFKCJfFZEPRORdEZnv23+DiNSIyAoROSWbMhqDi2ER+FolLBwDPx6lShC0BhG6D2bdFaUB3oT7DM/t6Q+uPQ7PNcLzjemny6fS4BsdBV1nCzrgM8Xp6yYTvNuq/UP9Q4YjIf37frJ99/6GBA/UwRUfwisPQWchtA6B5nJwJbDheWhZp58dSEoq1N0Z8rR02hCTf+E7YWcDNNdBa7N2pDECQyAsQhE5HpgDHOKcaxOR4d7+g4C5wMHAKGCRiExxzvVxFoFh7DqnlsKQkM71i6An/cQX7huVvb//gjK4tVaVVqI1XMwb4XTpHlhQTzVo55mdngIsCcGCUdqsPBPnrOvaSNtPlcDPe+mysinmxUlTRyXR94kgfmpjcN8O2JIou8tLxmKbBQrjUP8OVByx+5+xT1jxGjzxA1j1plp8PdHuJc3EPDeC8/r6uTbIL9L4YUUf+tcZ+4ygWIRXAd93zrUBOOe2ePvnAI8559qcc6uBGuDwLMlo5Agi8NI4zTTtJDlt4rIybTvWG6PzYV6VvrfJq+9rc/CFCjguZaZiaxweqYNfbNfYZCbWtevU9ra4lj4Uh1QhXrw+8/veb9MG2GXeQN8wyR98MfDCeJW1J44oUndle4r12QF8djcHJYM2+G4H6sPgpgFTwBWpgJ2Vah2OiuocxcDwwq/g+tnwv0/D5o+howU6M11moG7ReCxZRuHiahlKGOq3wHfPhI+W94voRs8EwiIEpgDHisg8dC7qN5xzrwOjgVd9r1vn7TOMtKxog+9uhb/v1JP+qSVw0zAY08favumFsHmqToL4Rwec4Rtyuyt8tUoH4z5Yr0k255Z1n6jwXKMqshbPg3btFrhuaPoWZ/fVaeKLv99oodd+7YEdOrEhlbe9qRT5IRjitOtMDDVMxuSlbwOXSkkIrq+E79VCR1zXtBO1mP+jatfWIpVlLdoL9p1WiE0FRqC/+gifzLmKxWFbNSxrhcN2Qc59TiwGC74JnZ2q2EIhVWydPVy9gMYKP8k9RvuPFhZDR7v+Ix76Lty8MKDpsblDvylCEVkEpHPE3OjJMRQ4EpgFPCEiE/t4/CuBKwHGjh27Z8IaA5JH6uALG/ScGgLygV/Xa0LG78dqW7O+IALHlsCxvn3vtqqifaMVRkfg+io4PYNrclw+fCeNggK1Ei9ar0qy2FNuHXH4/ja1tI5Jsba2xbrHLEH1xrYMLsqpUe1FGotrXC8xA7Ep3jfl8q1hcFCBNuTeHNOG3d8Zlnk8VCac02za22rVEo6B/qPKUT2RuFBYA3nDYX0YTl0Dyyf2/bP2Oitfh7adQKLBq/OUXCqpXWzRxxLSraNNlWhePpRVwbYNulXZ9X026TfXqHPuROfcp9Jsz6CW3kKnvIb+vquA9cAY32GqvX3pjn+fc26mc27msGEZzj7GoOWB7XDJBmghOUG+BbWiVrfDr/dCKv6bLXDsx9qxZmtMe3ieuw7u2db3Yz1erxZaoe8XmBdS2e/d0f31J5Z4p1h//oXXfebkDC7K6VE4vEAvDNriqmib4xAV+HYfrbkzS+HP4+HdyXDPqN1TTK+3ws936N+YH/IlFIVRZZjnbZOguDLp/p1f2/fP2uuUDEGVX1ytwlis9wSZLrvj6iZ1cS2fGDk5qVDzbF5htglKjPBp4HgAEZmCXszXAr8D5opIVEQmAPsDr2VNSiOQbIlpckq6JMoOtO/oKy17/jnXbdZYX0lIu9UkGnvfUqvnx77Q2MPr09X1zSmBwwr0b9kZ163JwdGFcFIGi1QEnhgDZ5Wqwg0JTMmHZ8aohdffPN+kFyafyIdPGYZQv1BES+7yvCfCqHs061RPVbdmp6fMMim8BOI7tYYiSdeni+txyqugbitMman3jawSlBjhA8ADIvIOGkO/xGnLm3dF5AngPdST8mXLGDVSea2le/swP3EyXLz3keVtWiTvJyqq1FZ1wP59uLD/51K4aYtmk0Z8swlBa+dSCYVg0Ti4ezs8Vq8K5IIhcHVFz58zNAwPVWuGZ5vTpgHZCkc59AoX1PALkaxl9DsUiyXZzjMO7B+EhJkVr2mGZ2uzWnTpCIXVXVpaqY25O9uT8URC+lzcizHWbYWRE+HCW/r1zzDSEwhF6JxrBy7M8Nw8YF7/SmQMJCKiTbV7Il0T7VTWd8CGGIzLS06E8FMWgs1xPYkniKMn9GF9LPqelA9fHKJJMLikIpgehUvK078n35tc8c3dMCCCUJR+UrG2UdsRV9d1oTe/UdA1TcxKDIvGTjud/t+u7EXZ9wvv/l0tuemfhbdfhraEE95HvBPKh8HkmbB6uZZHJNypiRoUCcGJl8Ipl8GEaZYkExCC4ho1jN3mqML0iguSJ9meMiRbOuHstTBtFcz+WG+v2aQnYj9XDVW3RKI5d9ypJXpMEQzZjUvKH42Ex6vhxCI4ohDuGAF/G9+1I81g4shCneAxOk+vwDvRC5gTi+BXo2HFJDi7VK3swpCWTvx4hJZwZJ2CYnVrhkKkT4jx6OyAd16Gus1aKuHiSSuQkCrTGbO94xhBwZpuG4OCvzXDnDVQ55KnqHzUlZkfgiUTYXwGF9t56+DJhmSMMYxamfOGw9d9dYPxOHxlE/y6ITkzcFYB/HbM7inCXMQ5WNmuWbd5ohmow1LWrq5Tt1F5u9YMvF/Y+BHcfh4Ul8EHi9M3294VwnmQXwBllXDI8fClu6GoH3vu5Tg2fcIY9LzYCFdtgu0xtdqcaILIHcMzjxpqiUPZB2rp+ftwCtpqrWb/7u/ZHoOlLTAxuucjkowBxOI/wGPz4ON30ihCv5WYwWKUkLpCi8o06aakAk77Ipx/076V2/gEmz5hDHpOKoWHR2sh+0EFWkz/pzE9z9tbvDPZOcaPAzZmaBoyNKKfNVCU4Kp2tXrHrYQZNXB/mvIMYxc44nMw73k49CSoHK3u0oJiKCzV0Urid5umm0rhRUFj7ZAf1d6jr/5OyzGMrGIOHWNQcURR32JKb6ZJzU+cygLjltsDVrfD0as1szUKbO+Ef9+kLc5+2EufUSMNRaVwwc1w15Vq0TXVaRZpUYHWGq5530v/zeRpiye70zinStHF6T7rxOhPzCI0cpqqSPJq0KXczgpCa6895D+3aieZ0pDGSotC2sBlQZ26eI3dYOIh8OWfwKQZUDEcDjxSH1/5Q4j04iaIx3WyfTyuCvDgYyCS7bY5hlmERk5zeilUhbXOrgNVgiE0czFTe7SBxCst3X/kkZB2mlncAqdZnsbuMXUWXPdQchZhgvNuhEdv1VZqn+CPPns3bTth9P5w7vX9JLDRE2YRGjnN0DD8aASMiGhdW4GoYvz3Ci2LGOiMCHefTp/Ijxtrhsiek1oHeP5N8NM3tEYwWgTF5epOLSxJZotWT4FzroM7/wbDxqQ/rtGvmEVo5Dxzh8DhRfBik3Zf+UyxFrYPhlrnayrh/PXJ2YiJ2scDo3BwFtqs5QRjD4TbXoS7vgjbNybrD4ePh6vvg9IgdAgw/Fj5hGEMcv7/VrhtW3JS/bQo/HYs7GeXwfuW9lZ4+6+w5R8wajIc9GmbTJ9lMpVP2E/BMAY51w6DL1dqEfuoiI6HMvqB/AI47ORsS2HsAqYIDSMHKAjBUYMg5mkY+wJLljEMwzByGlOEhmEYRk5jitAwDMPIaUwRGoZhGDmNKULDMAwjpzFFaBiGYeQ0pggNwzCMnMYUoWEYhpHTmCI0DMMwchpThIZhGEZOMyibbovIVuAfe3iYKqB2L4jTX5i8+5aBJi8MPJlN3n3LQJMX9r7M45xz3SaNDkpFuDcQkSXpupQHFZN33zLQ5IWBJ7PJu28ZaPJC/8lsrlHDMAwjpzFFaBiGYeQ0pggzc1+2BegjJu++ZaDJCwNPZpN33zLQ5IV+ktlihIZhGEZOYxahYRiGkdOYIjQMwzByGlOEPkRkhoi8KiJvisgSETnc2y8i8mMRqRGRt0Tk0GzLmkBEvioiH4jIuyIy37f/Bk/eFSJySjZlTIeIXCsiTkSqvMeBXGMRucNb37dE5LciMsT3XCDXWERO9WSqEZFvZVueVERkjIi8JCLved/bq739Q0XkRRH50LutyLasfkQkLCJviMiz3uMJIrLYW+fHRSQ/2zL6EZEhIvKk9/19X0SOCvIai8jXve/DOyLyqIgU9NsaO+ds8zbgBeA07/7pwF98958DBDgSWJxtWT25jgcWAVHv8XDv9iBgORAFJgCrgHC25fXJPQZ4Hm16UBXwNT4ZiHj3fwD8IMhrDIQ9WSYC+Z6MB2VbrhQZRwKHevdLgZXees4HvuXt/1ZirYOyAdcAjwDPeo+fAOZ69+8Frsq2jCny/gr4gnc/HxgS1DUGRgOrgULf2l7aX2tsFmFXHFDm3S8HNnj35wAPOuVVYIiIjMyGgClcBXzfOdcG4Jzb4u2fAzzmnGtzzq0GaoDDsyRjOn4EXIeud4JArrFz7gXnXMx7+CpQ7d0P6hofDtQ45z5yzrUDj6GyBgbn3Ebn3DLvfiPwPnoinIOevPFu/yU7EnZHRKqBzwELvMcCzAae9F4SNHnLgeOA+wGcc+3OuToCvMZABCgUkQhQBGykn9bYFGFXvgbcISJrgTuBG7z9o4G1vtet8/ZlmynAsZ7r4GURmeXtD6q8iMgcYL1zbnnKU4GV2cflqNUKwZU3qHKlRUTGA/8ELAZGOOc2ek9tAkZkSax03IVevMW9x5VAne8iKWjrPAHYCvzSc+cuEJFiArrGzrn16Dl3DaoA64Gl9NMaR/bFQYOMiCwC9kvz1I3ACcDXnXNPici56NXUif0pXyq9yBsBhqKuxFnAEyIysR/FS0svMn8bdTcGhp7kdc49473mRiAGPNyfsg1mRKQEeAr4mnOuQY0sxTnnRCQQtV0icgawxTm3VEQ+m215dpEIcCjwVefcYhG5G3WFfkLA1rgCtVYnAHXAb4BT++vzc04ROucyKjYReRC42nv4Gzw3CLAejWslqPb27XN6kfcqYKFTB/prIhJHm9RmTV7ILLOITEO/6Mu9k141sMxLSgrkGgOIyKXAGcAJ3lpDlte4B4IqVxdEJA9Vgg875xZ6uzeLyEjn3EbPLb4l8xH6lU8DZ4rI6UABGj65G3XfRzyLJWjrvA5Y55xb7D1+ElWEQV3jE4HVzrmtACKyEF33flljc412ZQPwGe/+bOBD7/7vgIu9zMYjgXqfeyGbPI0mzCAiU9CAeC0q71wRiYrIBGB/4LWsSenhnHvbOTfcOTfeOTce/bEe6pzbREDXWERORV1iZzrndvqeCuQaA68D+3vZdvnAXFTWwODF1+4H3nfO/dD31O+AS7z7lwDP9Lds6XDO3eCcq/a+s3OBPzvnLgBeAs72XhYYeQG839RaEZnq7ToBeI+ArjHqEj1SRIq870dC3v5Z42xnCwVpA45B/dLL0ZjFYd5+AX6KZuO9DczMtqyeXPnAr4F3gGXAbN9zN3ryrsDLhA3aBnxMMms0qGtcg8bc3vS2e4O+xmgG7kpPthuzLU8a+Y5BE6Xe8q3r6Wjc7b/RC9BFwNBsy5pG9s+SzBqdiF781KAepGi25UuRdQawxFvnp4GKIK8xcAvwgXc+ewjNyO6XNbYWa4ZhGEZOY65RwzAMI6cxRWgYhmHkNKYIDcMwjJzGFKFhGIaR05giNAzDMHIaU4SGYRhGTmOK0DAMw8hpTBEaxiBCRL4kIvf4Ht8qIg9lUybDCDpWUG8YgwgRKUI73UxDO7h8DzjaOdeSVcEMI8CYIjSMQYaIzAeKgdOAk5xzq7IskmEEGlOEhjHIEJED0GG3c5xzgWq4bRhBxGKEhjH4uBkdyvrJmDURmSgi94vIk5nfZhi5iSlCwxhEiMi16My8c0nO1sQ595Fz7oqsCWYYASbnBvMaxmBFRGYDlwFHOecaRaRMRGY4597MtmyGEWTMIjSMQYCIjAUWAOc45xq93XcDX8ueVIYxMLBkGcPIAUSkEpgHnAQscM7dnmWRDCMwmCI0DMMwchpzjRqGYRg5jSlCwzAMI6cxRWgYhmHkNKYIDcMwjJzGFKFhGIaR05giNAzDMHIaU4SGYRhGTmOK0DAMw8hp/g8+I4vurZUg/wAAAABJRU5ErkJggg==\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "alpha = 0.7\n", "\n", "fig, ax = plt.subplots(figsize=(7, 7))\n", "ax.scatter(\n", " emb_transformed[0],\n", " emb_transformed[1],\n", " c=emb_transformed[\"label\"].cat.codes,\n", " cmap=\"jet\",\n", " alpha=alpha,\n", ")\n", "ax.set(aspect=\"equal\", xlabel=\"$X_1$\", ylabel=\"$X_2$\")\n", "plt.title(\"TSNE visualization of GCN embeddings for cora dataset\")\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Comparing Different Models\n", "\n", "Now we run Deep Graph Infomax training for GAT, GCN, APPNP, and GraphSAGE. Note that switching between StellarGraph models only requires a few code changes." ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [], "source": [ "def run_deep_graph_infomax(base_model, generator, epochs):\n", " corrupted_generator = CorruptedGenerator(generator)\n", " gen = corrupted_generator.flow(G.nodes())\n", " infomax = DeepGraphInfomax(base_model, corrupted_generator)\n", "\n", " x_in, x_out = infomax.in_out_tensors()\n", "\n", " model = Model(inputs=x_in, outputs=x_out)\n", " model.compile(loss=tf.nn.sigmoid_cross_entropy_with_logits, optimizer=Adam(lr=1e-3))\n", " history = model.fit(gen, epochs=epochs, verbose=0, callbacks=[es])\n", "\n", " x_emb_in, x_emb_out = base_model.in_out_tensors()\n", " # for full batch models, squeeze out the batch dim (which is 1)\n", " if generator.num_batch_dims() == 2:\n", " x_emb_out = tf.squeeze(x_emb_out, axis=0)\n", "\n", " emb_model = Model(inputs=x_emb_in, outputs=x_emb_out)\n", "\n", " test_gen = generator.flow(test_subjects.index)\n", " train_gen = generator.flow(train_subjects.index)\n", "\n", " test_embeddings = emb_model.predict(test_gen)\n", " train_embeddings = emb_model.predict(train_gen)\n", "\n", " lr = LogisticRegression(multi_class=\"auto\", solver=\"lbfgs\")\n", " lr.fit(train_embeddings, train_subjects)\n", "\n", " y_pred = lr.predict(test_embeddings)\n", " acc = (y_pred == test_subjects).mean()\n", "\n", " return acc" ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " ['...']\n", "Test classification accuracy: 0.448318293683347\n" ] } ], "source": [ "gat_model = GAT(\n", " layer_sizes=[128], activations=[\"relu\"], generator=fullbatch_generator, attn_heads=8,\n", ")\n", "gat_acc = run_deep_graph_infomax(gat_model, fullbatch_generator, epochs=epochs)\n", "\n", "gat_acc\n", "print(f\"Test classification accuracy: {gat_acc}\")" ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " ['...']\n", "Test classification accuracy: 0.4376538146021329\n" ] } ], "source": [ "appnp_model = APPNP(\n", " layer_sizes=[128], activations=[\"relu\"], generator=fullbatch_generator\n", ")\n", "appnp_acc = run_deep_graph_infomax(appnp_model, fullbatch_generator, epochs=epochs)\n", "\n", "print(f\"Test classification accuracy: {appnp_acc}\")" ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " ['...']\n", "Test classification accuracy: 0.7227235438884332\n" ] } ], "source": [ "graphsage_generator = GraphSAGENodeGenerator(G, batch_size=1000, num_samples=[5])\n", "\n", "graphsage_model = GraphSAGE(\n", " layer_sizes=[128], activations=[\"relu\"], generator=graphsage_generator\n", ")\n", "graphsage_acc = run_deep_graph_infomax(\n", " graphsage_model, graphsage_generator, epochs=epochs\n", ")\n", "\n", "print(f\"Test classification accuracy: {graphsage_acc}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Heteogeneous models\n", "\n", "Cora is a homogeneous graph, with only one type of node (`paper`) and one type of edge (`type`). Models designed for heterogeneous graphs (with moer than one of either) can also be applied to homogeneous graphs, but it is not using their additional flexibility.\n", "\n", "HinSAGE is a generalisation of GraphSAGE to heterogeneous graphs that can be trained with Deep Graph Infomax. For homogeneous graphs, it is equivalent to GraphSAGE and it indeed gives similar results." ] }, { "cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " ['...']\n", "Test classification accuracy: 0.7223133716160788\n" ] } ], "source": [ "hinsage_generator = HinSAGENodeGenerator(\n", " G, batch_size=1000, num_samples=[5], head_node_type=\"paper\"\n", ")\n", "\n", "hinsage_model = HinSAGE(\n", " layer_sizes=[128], activations=[\"relu\"], generator=hinsage_generator\n", ")\n", "hinsage_acc = run_deep_graph_infomax(hinsage_model, hinsage_generator, epochs=epochs)\n", "\n", "print(f\"Test classification accuracy: {hinsage_acc}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "RGCN is a generalisation of GCN to heterogeneous graphs (with multiple edge types) that can be trained with Deep Graph Infomax. For homogeneous graphs, it is similar to GCN. It normalises the graph's adjacency matrix in a different manner and so won't exactly match it." ] }, { "cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " ['...']\n", "Test classification accuracy: 0.7333880229696472\n" ] } ], "source": [ "from stellargraph.mapper import RelationalFullBatchNodeGenerator\n", "from stellargraph.layer import RGCN\n", "\n", "rgcn_generator = RelationalFullBatchNodeGenerator(G)\n", "\n", "rgcn_model = RGCN(layer_sizes=[128], activations=[\"relu\"], generator=rgcn_generator)\n", "\n", "rgcn_acc = run_deep_graph_infomax(rgcn_model, rgcn_generator, epochs=epochs)\n", "print(f\"Test classification accuracy: {rgcn_acc}\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Overall results\n", "\n", "The cell below shows the accuracy of each model." ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
Accuracy
GAT0.448318
GCN0.789171
APPNP0.437654
GraphSAGE0.722724
HinSAGE0.722313
RGCN0.733388
\n", "
" ], "text/plain": [ " Accuracy\n", "GAT 0.448318\n", "GCN 0.789171\n", "APPNP 0.437654\n", "GraphSAGE 0.722724\n", "HinSAGE 0.722313\n", "RGCN 0.733388" ] }, "execution_count": 21, "metadata": {}, "output_type": "execute_result" } ], "source": [ "pd.DataFrame(\n", " [gat_acc, gcn_acc, appnp_acc, graphsage_acc, hinsage_acc, rgcn_acc],\n", " index=[\"GAT\", \"GCN\", \"APPNP\", \"GraphSAGE\", \"HinSAGE\", \"RGCN\"],\n", " columns=[\"Accuracy\"],\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Conclusion\n", "\n", "This notebook demonstrated how to use the Deep Graph Infomax algorithm to train other algorithms to yield useful embedding vectors for nodes, without supervision. To validate the quality of these vectors, it used logistic regression to perform a supervised node classification task.\n", "\n", "See [the GCN + Deep Graph Infomax fine-tuning demo](../node-classification/gcn-deep-graph-infomax-fine-tuning-node-classification.ipynb) for semi-supervised training using Deep Graph Infomax, by fine-tuning the base model for node classification using labelled data." ] }, { "cell_type": "markdown", "metadata": { "nbsphinx": "hidden", "tags": [ "CloudRunner" ] }, "source": [ "
Run the latest release of this notebook:
" ] } ], "metadata": { "celltoolbar": "Tags", "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.6.9" } }, "nbformat": 4, "nbformat_minor": 4 }