{ "cells": [ { "cell_type": "markdown", "metadata": { "collapsed": true }, "source": [ "![McLaren Maze Race Banner](media/banner.png)\n", "\n", "# Welcome to Rookie Driver - level 3 of the McLaren Maze Race!\n", "\n", "In this level we look at a major update to our AI algorithm, switching from Q Learning to a model-based method. DRS is also available at certain points around the mazes, which provides a way for the driver to go even faster - if the AI decides it is worth the extra turn to activate it.\n", "\n", "## Model-based Reinforcement Learning\n", "\n", "Let's first look at updating our AI's brain. Up until now we have been using Q Learning, which is a form of model-free reinforcement learning. It is called that because it does not build an explicit model of the environment which can predict what will happen if the AI takes a particular action. Model-free methods generally require fewer assumptions and less knowledge about the environment to do well: the only prior knowledge that we gave the algorithm so far is the choice of state vector and only considering turns at the end of the straight. If you need to train an AI agent to solve a task and you have very little information about it then a model-free method will often get the job done. However, this flexibility comes at a cost: model-free methods usually require a lot more data to learn from than model-based methods, which are able to extra more learning from the same information (although this does depend on the specifics of the algorithm and environment).\n", "\n", "In our maze race, data efficiency is key - our driver only has 24 races in the season and needs to learn as fast as possible! For this reason we will look at a model-based method. The idea is to learn two things about our car:\n", "- the fastest, safe end-of-straight speed\n", "- the effect of each action on car speed\n", "\n", "If we know both of these then we can choose the optimal set of actions to drive the car as fast as possible down the straight and still make the corner at the end. The critical choice in model-based learning is the choice of our model. If the model we choose isn't able to capture the true shape of car's dynamics, then it will make very bad predictions which will lead the AI astray. This is why it is important to know something about the environment when using a model-based method. If you have access to data from the environment, it is a good idea to study it before choosing a model class. In our case we can plot the dynamics functions for a sample car:" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "be9750d37bdc4c61bb9d2c09d6542a1f", "version_major": 2, "version_minor": 0 }, "text/plain": [ "Canvas(toolbar=Toolbar(toolitems=[('Home', 'Reset original view', 'home', 'home'), ('Back', 'Back to previous …" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "%matplotlib widget\n", "from imports import * \n", "\n", "set_seed(0)\n", "car = Car.get_car_for_level(Level.Rookie)\n", "plot_car_dynamics(car)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We can see that the functions mapping \"speed in\" to \"speed out\" for three of the four actions are nonlinear, that is, they are not straight lines. If we chose a linear model our AI would probably do much worse than the Q Learning approach we have been taking so far. We could try and exactly replicate the form of the functions shown above (which look to be piecewise linear) however this would also be dangerous as these dynamics are randomly generated each time and will be different in the online challenge. We therefore need a general nonlinear model that can adapt itself to the shape as data arrives.\n", "\n", "There are two aspects of the maze race that make this modelling task much easier: we know what the car speed is exactly (without any noise) and we know the speed at next turn only depends on our current speed and the action that we take (advance warning - at higher levels this is going to change!). This means we can use a simple linear interpolation algorithm for each action. When we want to predict the next speed for an action, we find the closest two speeds we have seen previously (for the same action), draw a straight line between them and use that to estimate our output speed. With just two data points this would be a simple linear model, however, as we gather more and more data the complexity of the shape will adapt to the data that is observed (this is known as a *nonparametric* regression model). Here is an example of fitting the Heavy Brake curve as we get more and more data:" ] }, { "cell_type": "code", "execution_count": 2, "metadata": { "scrolled": false }, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from scipy.interpolate import interp1d\n", "from IPython.display import display, clear_output\n", "from time import sleep\n", "speed_in = np.linspace(0, 300, 500)\n", "true_speed_out = np.array([car.dynamics_model.heavy_brake(s) for s in speed_in])\n", "fig = plt.figure(figsize=(8,4))\n", "clear_output()\n", "hdisplay = display(fig, display_id=True)\n", "ax = fig.add_axes([0.09, 0.11, 0.9, 0.89])\n", "ax.plot(speed_in, true_speed_out, c='lightgreen')\n", "data_in, data_out = [], []\n", "data_plot = ax.plot(data_in, data_out, 'b+')[0]\n", "interp_plot = ax.plot([], [], 'b')[0]\n", "ax.set_xlabel('Speed in')\n", "ax.set_ylabel('Speed out')\n", "for x in [100, 200, 250, 300, 0, 50, 240, 80]:\n", " data_in.append(x)\n", " data_out.append(car.dynamics_model.heavy_brake(data_in[-1]))\n", " data_plot.set_data((data_in, data_out))\n", " if len(data_in) > 1:\n", " f = interp1d(data_in, data_out, fill_value='extrapolate')\n", " interp_plot.set_data((speed_in, f(speed_in)))\n", " hdisplay.update(fig)\n", " sleep(0.5)\n", "plt.close()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We can estimate the maximum, safe, end-of-straight speed in the same way we estimate the safety car speed: start with an initial guess and then if we successfully make it around the corner we can increase our guess and if we crash we decrease our guess.\n", "\n", "Now we have the building blocks in place for our models we need to decide how to use them to choose our actions. We will start at the end of the straight where we already have a target speed: our estimate of the fastest, safe, end-of-straight speed. We can now work backwards up the straight, using our models to work out how fast we can go at the previous step and still make it below the target speed at the current step. If the models were invertable, then we could easily swap them from mapping speed in to speed out to models that went from speed out to speed in. However, the saturation at 0 and the car's top speed make this tricky (there might be other noninvertible regions at higher levels as well). Therefore, we will instead use a forward simulation approach: for each action we will feed in an array of different input speeds and find the one that gets closest to the target speed. We will then take the fastest input speed over all of the actions and set this to be the target speed for the previous step - we know that if we are travelling at this speed then (if our model is correct) there is an action we can take that will bring us to the target speed. This process in outlined in the figure below.\n", "\n", "\n", "\"Estimating\n", "\n", "Once we have a target speed for each position on the straight we can choose the action we want to take at our current position by finding the action that gets us closest (but below) the target speed. To do this we just need to call our dynamics models with our current speed to get their predicted output speed for each action. Here is the code implementation of this final step:\n", "\n", " def _choose_move_from_models(self, speed: float, target_speed: float, drs_active: bool, safety_car_active: bool):\n", " # Test each action to see which will get us closest to our target speed\n", " actions = CarAction.get_sl_actions()\n", " next_speeds = np.array([self.estimate_next_speed(action, speed, drs_active) for action in actions])\n", " errors = next_speeds - target_speed # difference between predicted next speed and target, +ve => above target\n", " \n", " # The target speed is the maximum safe speed so we want to be under the target if possible. This means we don't\n", " # necessarily want the action with the smallest error\n", " if np.any(errors <= 0): # under or equal to the target speed\n", " errors[errors > 0] = np.inf # at least one action gets us under the speed so ignore others even if close\n", " \n", " # Now we can choose the action with the smallest error score. At the start there will be multiple actions with\n", " # with the same score, so we will choose randomly from these\n", " min_error = np.min(errors ** 2)\n", " available_actions = [action for action, error in zip(actions, errors)\n", " if np.abs(error ** 2 - min_error) < 1e-3]\n", "\n", " return self._choose_randomly(available_actions)\n", " \n", " Now let's test out our new driver!" ] }, { "cell_type": "code", "execution_count": 4, "metadata": { "scrolled": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3oAAAHZCAYAAADQREkRAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8rg+JYAAAACXBIWXMAAA9hAAAPYQGoP6dpAAC1nklEQVR4nOzdeXhM598G8HsSWSQkiKwESexCNbGlltpD0WjS2msppX5oS2kbtW9RWnTRKFWxpVGKtihiCUVsQWsvithCLEkkGJE57x/zZprJLGcmM5PZ7s91zSU5z1m+c+bEzDPnnOeWCIIggIiIiIiIiGyGg7kLICIiIiIiIuNiR4+IiIiIiMjGsKNHRERERERkY9jRIyIiIiIisjHs6BEREREREdkYdvSIiIiIiIhsDDt6RERERERENoYdPSIiIiIiIhvDjh4RERERWa2EhARIJBJIJBJcu3bN3OUQWQx29IiIiCxUSkqK4gOsRCJB+fLl8eTJE9Hlnj59Ck9PT6VlU1JSTF+wlbpz545iP92+fVsxXRAEeHl5QSKRYP/+/aVSS9FOi9gjISFB43pyc3Oxf/9+fPHFF+jVqxeCgoIUy9WoUUPvus6cOYMRI0YgJCQEZcuWhbe3N1q3bo0lS5bgxYsXJXqu165d0/m5anuQYe7du4ctW7ZgypQp6Nq1KypXrqzYt4MHD9ZpHU+ePMHGjRsxcuRING3aFBUrVoSTkxO8vLwQERGBadOmISMjQ+eatm/fjj59+iA4OBhubm5wdXVFYGAgoqKisG7dOshkshI+W7ni/7dqe0ybNs2gbZlTGXMXQERERLrJzc3F5s2b0a9fP63z/frrr8jJySmlqqzfwYMHAQBBQUEICAhQTD937hwePnwIZ2dnNGvWzFzllUiPHj2M1rlftmwZRo8ejefPnyumPXv2DAcOHMCBAwewYsUKbN26FZUrVzbK9qh0+fr6GrT833//jZYtWyI3N1el7eHDhzh8+DAOHz6MhQsXYunSpejdu7fGdUmlUvTv3x+//PKLStvNmzdx8+ZN/Pbbb1i8eDF+++03VKhQwaDabR07ekRERFbA1dUVz549w+rVq0U7eqtXr1ZahrQr7Oi1atVKafqff/4JAAgPD4erq2up17Vjxw6ljmdxVatW1dgmCILi50qVKqFJkyY4dOiQ2g/j2mzbtg3vvfceZDIZfH198dlnn6F58+Z4+PAhli1bho0bN+Lo0aN44403kJKSAkdHR53XXaVKFZw+fVpje8OGDQEATZo0wYoVKzTOFxoaqvOZJ9KuWrVqqFu3Lnbu3KnzMjk5OYrjqmXLlujevTuaNGkCLy8vZGZmYuPGjVi2bBlycnLQv39/eHh4oGvXrmrX9f777ys6eT4+Pvj4448RFhYGJycnnD59Gp9//jmuX7+OP//8E3369MH27dsNfs4//vgjmjZtqrHdx8fH4G2YjUBEREQWae/evQIAAYDQq1cvAYDg6Ogo3LlzR+Myd+/eFcqUKSMAEHr37q1Yfu/evaVXuJVp1qyZAED4/vvvlab3799fACBMmDCh1GpZsWKF4jW7evVqidfz/fffC4mJicKlS5cU06pXry4AEKpXr67TOp4/fy4EBwcLAAQPDw/h8uXLKvP873//U9S7YsWKEterTuF6X331VaOul5RNmTJF+P3334WMjAxBEATh6tWrin0/aNAg0eUPHjwo9OrVSzh79qzGeTZv3ixIJBIBgBASEiLIZDKVeTIyMgQHBwcBgFCxYkXhxo0bKvNkZ2cLNWrUUNR37Ngx3Z9oEUX/b7Xl/xt5jx4REZEV6Ny5M/z8/FBQUICffvpJ43w//fQTXrx4AT8/P3Tq1KkUK7ROT58+xcmTJwFoPqNXfLo1GD58OPr27YuaNWuWeB2bNm3Cv//+CwCIjY1FSEiIyjzz589HxYoVFT+T9Zk+fTq6d+9e4ks4X3nlFaxbtw7169fXOE9UVBSio6MBAFeuXFH8zRV15MgRxb13Q4YMUXvG2sPDA2PHjlX8npqaWqKa7QU7ekRERFbA0dERffv2BfDfpZnqrFq1CgDQr18/nS+je/78Ob777ju0a9cO3t7ecHZ2hp+fH1577TWsWbNG7cAH169fh4ODAyQSCT777DPRbfz000+KwQ22bdumdp7Lly9j7NixaNiwITw9PVG2bFkEBwdj8ODBOH78uE7PRV9Hjx5Ffn4+vLy8UK9ePcX09PR0pKenQyKRoGXLlibZtqXbvHmz4mdNl0a6ubmhV69eAOT3NP7zzz+lUJkysVE327ZtC4lEgrZt2wKQH2fvvfcegoODUbZsWdSoUQNDhw7F9evXlZY7c+YMhgwZguDgYMVgICNHjsS9e/d0qmvz5s146623UK1aNbi6uqJChQpo0qQJpk+fjkePHhn6tK1Ou3btFD9fuXJFpb3oPaDBwcEa11P0C4eiy5AqdvSIiIisxNtvvw0AOHnyJM6ePavSfu7cOZw4cUJpXjHXrl3DSy+9hFGjRiElJQX3799Hfn4+7t69iz/++ANvv/02Xn31VTx8+FBpuerVqys6QNrOMBZau3YtAMDb2xudO3dWaf/iiy9Qv359LFq0CGfOnEFOTg6ePXuGq1evYuXKlWjWrBmmTJmi03PSRN1Ie4Uf/h88eKDouEokElSvXh2A/F63oqMQahrBtGhnw5pH6SvqwIEDAIA6derAz89P43yvvvqq4ufC+x0t1a5duxAWFobvv/8eV69exbNnz3D9+nX8+OOPaNasGS5cuABAfkyHh4cjISEBV69ehVQqxc2bN7FkyRI0a9ZMaXTW4h49eoQOHTrgjTfewIYNG3Djxg1IpVJkZ2cjLS0N06ZNQ926dXH48GGN66hRo4bNjSoqlUoVP6v7EqpOnTqKnwvPJKtTtJNYdBlSxY4eERGRlXj55ZfRoEEDAOrP6hVOCw0NRePGjUXXl5ubiw4dOig+3Pbs2RO//fYbjh8/jvXr1ys+wB84cAA9evRAQUGB0vL9+/cHAFy9ehWHDh3SuJ0HDx4oBnfo1asXypRRHgtu/vz5mDBhAvLz89GoUSPEx8dj165dOH78ONauXYuIiAgIgoCZM2fi66+/Fn1etmLIkCEICAiAs7MzKleujBYtWmDSpEm4deuWybedm5uLGzduAADq1q2rdd6i7efPnzdpXYa4ffs2evXqhQoVKuCbb77BkSNH8Oeff+LDDz+ERCLBvXv3MGzYMBw7dgwDBw5ESEgIfvjhBxw9ehR79+5VfHly/fp1jBs3Tu02pFIpOnbsiD179sDR0RFvv/02fvrpJxw+fBh//vknZs+eDS8vL9y7dw+vvfaayllEW7Zv3z7Fz0XPnhdq2LAhXnnlFQDyL07UdaYfP36MRYsWAZCf9VP3pZG+PvvsM1SvXh0uLi6oWLEiXn75ZYwdO9YsZ6eNztw3CRIREZF6RQcMKBzo4vPPPxcACIGBgUoDGshkMiEwMFAAIMybN08QBOWBPdQNODB+/HhF+6RJk1TaZTKZYkASAMJ3332n1H7//n3ByclJACCMGjVK4/OIj49XrOPQoUNKbWfPnlWsY+rUqWoHaSgoKBAGDBggABDKlSsnPHz4UOO2tMnLyxPOnz+veJw+fVpwdnYWAAg///yzUlvhgA9fffWV0vTz588LeXl5Kusuuq+nTp1aovqKr0fTw9XVVViyZIne69ZnMJbz588rtqfttRUEQcjMzFTM26dPH73r0qRwnWKDsYgNYPPqq68q2mvVqiXcu3dPZZ6ifwve3t7CK6+8ovZ1fuuttwQAQpkyZdSuZ+LEiQIAoUKFCsLx48fV1nvt2jXB399fACD069dP7TyFr5U5PqrrOxiLLk6dOiU4OjoKAISGDRtqnO/ixYtCUFCQAEDw9fUVvvzyS2Hv3r3Cn3/+KcTHxyvaKleuLKSmppa4nqL/t2p6ODg4aPw/yVqwo0dERGSh1HX0bt68qRiZbs+ePYp59+zZo/hwcvPmTUEQtHf0nj17JlSoUEEAIDRo0EB48eKF2hqys7MFLy8vAYBQv359lfYePXooPhzn5+erXUerVq0EAEJwcLBK2zvvvCMAEJo0aaL1A9WjR48EFxcXAYCwdOlSjfPp4+jRo4qOk1QqVUx/+PChYoRAdSP/qWPMjl5wcLAwfvx44ZdffhGOHj0qHD16VEhKShLeeustRV1QM0qoGH06eoX7BoDwySefaJ33yZMninm7d++uV03amKKj98cff6hdx7///quYRyKRCOfOnVM7X+HfGQDh119/VWp7/Pix4OnpKQAQvvnmG601f/fddwIAwcnJScjNzVVpt6WO3rNnz4QmTZoo1vnbb79pnf/+/fvCzJkzhXLlyql0vpycnITx48fr/Hepyd69ewV/f39h1KhRwk8//SQcOXJESEtLEzZt2iQMHTpU8eUTACE2NtagbZkTL90kIiKyIlWqVFEMalD08s3Cn9u3b48qVaqIrictLQ1ZWVkA5ANtaBq4xcPDQ2mwjTt37ii1F16+mZmZieTkZJXl09PTFfdtqcv/+/333wEAMTExWu9HqlChgiJXzVgj7RWOqtm0aVM4OzsrpqempkIQBAQGBmrNqitq8ODBEORfoBt0j94bb7yBy5cvY/78+YiOjkbTpk3RtGlT9O7dGz///DN+++03ODk5AQDGjh2LjIyMEm9Lm6L5i0X3jTouLi6Kn58+fWqSeoyhQoUKiIyMVNsWFBSE8uXLAwAaNWqk9tJCAHjppZcUPxe/j2zfvn3Izs4GALz55ptaa2nTpg0AID8/H2lpaSrt165dUxxP1m706NGKwZQGDRqEHj16aJ3/999/x9q1a9VmPubn5+Pnn39GYmKiQfumadOmuH79Or799lv06dMHzZo1Q1hYGHr27IkffvgBBw4cgKenJwBg7ty5+Ouvv0q8LXNiR4+IiMjKDBw4EADwyy+/4OnTp3j69Ck2bNig1CbmzJkzip+bN2+udd6i7UWXA4DXX39d8QG5cMCVon766SfFB7LCTmGh69evIzMzE4B8+P7iA6UUfxR+WDRW56ZwsJHi8QmF9xuaY7RNT09PrR3e7t27KwalefLkCZYvX26SOooGxIuNbFh0kI2yZcuapB5jqFWrluiXCQBQu3Zt0XkA+f1iRRUdGdbf31/rsRwaGqqY11SddUsQFxeHH374AYC8c7V48WKt83/00UcYMmQILly4gJ49e+LgwYPIzc3F06dPceLECQwZMgTp6en45JNP8Oabb6rcN6wrd3d3xRcm6jRr1gzffvstAPmATIU/Wxt29IiIiKxMdHQ03NzckJOTg19//RWbN2/G48eP4e7ursiqElN0FE0fHx+t8xYdcbH46Jtly5bFG2+8AUA+nPyTJ0+U2gs7f2FhYSqDeug6TH1xxbdRUoVnGot39AqnFw4MYWmGDx+u6LAUHeDCmAo77wDUnlkpKi8vT/FzuXLlTFKPMbi5uWltd3BwEJ2vcB4AKp0Mcx/Plub777/HxIkTAcgH7Nm2bRvc3d01zr9161YsWLAAgPwM+aZNm/DKK6/A3d0drq6uePnll/Hjjz9i8uTJAICNGzfiu+++M1n9ffr0gYeHBwDT/Z2ZWhnxWYiIiMiSlCtXDm+88QbWrl2L1atXK86YvfHGG1o/SGli6BDu/fv3x6pVq5CXl4dff/1Vkfd39uxZnD59WjFPcUU/KE+ZMgVvvfWWTtsryXNs27atxg9r3bp1Uzv9/fffx/vvv6/4fdCgQUhISNB728bm4+MDLy8v3L9/32QjcBa9/PfmzZta5y0cnRMAAgMDTVKPNSh6PJ84cULrGaOidL082Jr89NNP+N///gdAHsWSnJyMypUra12m8MyfRCLBrFmzNM43ceJELFy4ELm5ufjxxx8xZswY4xVeRJkyZVC7dm0cP368VEa6NQV29IiIiKzQwIEDsXbtWkVsQeE0XVWqVEnx8927d7Verlb00rKiyxXq0KEDfH19cffuXaxdu1bR0Ss8m+fg4IA+ffqoLOfl5aX42cnJSelyNtLO1Plq5cuXR2BgIG7cuKGI39CkaLume9vsQdHj2dvb2yY7cLr47bffMHDgQMhkMvj7+2P37t067YvCaA4fHx+t9xm7urqiQYMGOHLkiOixaShrzzHkpZtERERWqEOHDvD398eLFy/w4sULBAQEoEOHDjovX7RTdeTIEa3zHj16VO1yhRwdHRUduZ07d+LBgwcQBEERpN6uXTsEBASoLBccHKwY8MDUQdsrVqzA6dOnFY/CASGGDBmiNH306NEA5GcAi04/ffo0Zs+ebdIadZWZmYn79+8DgNr9aiyFl7RevHhR631kRc+UmuO+Rkvx8ssvK3629OB4U9m9ezd69eqFFy9ewMvLC8nJyQgJCdFp2cJ8zRcvXojOm5+fr7SMKbx48UKRpWfKvzNTYkePiIjIChWGMbu4uMDFxQVvv/220v1DYsLDwxUDS6xcuRIymUztfI8fP8bPP/8MAKhfvz78/f3Vzld4aWbhqHiHDh3CtWvXlNrUPYfXXnsNgLyDaMqw7aCgIISGhioehWcCunfvrjS9cCTFyMhIpemhoaE6jWZaGpYuXaq4XLcw1N4UevbsqfhZ0yWrT548UTo+tJ0ZtnUdO3ZU3N/39ddf28SImfo4dOgQoqKiIJVK4enpiR07dqBBgwY6Lx8UFAQAePDggdb/Cx4+fKgYFKpwGVNYt26dYhRVU/6dmZLFd/SkUimmTZumNKKTpbH0Gi29PsDya7T0+gDLr9HS6wMsv0ZLrw+wjhptyeeff45nz57h2bNnmDt3rl7Luri4YNiwYQDkI2nOnDlTZR5BEDB69GjF2aPCs13qNG3aFLVq1QIgv2QzMTERgPwyq5iYGI3LxcbGwtHRETKZDG+++abW+8EKCgqwdu1a0XvGxGRkZODSpUsAlAdikclkijMxhcPf6yohIUExomJJ4xWuXbuGkydPap1ny5YtmDFjBgD5QDhDhgwp0bZ08cYbbyA4OBiAfPTEK1euqMwzYcIEPHr0SPGzPatQoYLib+TQoUMYO3asxi9QAPkl04X3pRVXo0YNxfFkDU6dOoVu3bohLy8P7u7u2Lp1K8LDw/VaR9HYhQ8//FDtaK8ymQzvv/++oq179+4q81y7dk2x79q2bavS/ujRI6SkpGit5ejRo4rXUiKRYOTIkXo8EwtilvQ+PWRnZwsAhOzsbHOXopGl12jp9QmC5ddo6fUJguXXaOn1CYLl12jp9QmCddRoTdQFputDW2C6IAhCTk6OEBwcrJgnJiZG2LJli5CWliZs2LBBaNu2raItIiJCY6h6oWnTpikCpwuDo998803ROhcuXKjYjqenpzBhwgThjz/+EE6cOCEcOnRISExMFMaMGSP4+/sLAITTp0/rvS+KWrdunQBAqFOnjtL0U6dOCQAENzc34fnz53qt0xiB6YWvd0REhDBnzhxh69atwrFjx4Rjx44J69atUwlMX7x4scZ1Xbp0SVixYoXSozD43svLS6Xtzp07atezdetWwcHBQQAg+Pr6Ct98841w5MgRYfv27UJMTIyillatWokeH/oqXLexAtPF1lMYUi4WEq7tdX727JnQvHlzxTwvvfSS8O233woHDhwQTp48KezZs0f45ptvhKioKMHZ2VkIDw/XWktpfFT/888/lY6F+fPnK7bdsmVLlWOluMuXLws+Pj6KZRYuXCicPn1a6+Pu3bsq65FKpUK9evUU62nYsKGwdOlS4ciRI8Lx48eFVatWCREREYp2X19fITMzU2U9RQPf1b3mhe2NGjUSpkyZIvz666/C0aNHlQLTnZ2dFeuYMGGCMXazWbCjZwSWXqOl1ycIll+jpdcnCJZfo6XXJwiWX6Ol1ycI1lGjNTF1R08Q5B966tatq5hP3aNly5bCgwcPRLd36dIllWU3bdqkU61Lly4V3NzctNYBQHB2dhYuXbqkx15QNWrUKAGAMGzYMKXp33zzjQBAaN++vd7rNGZHT+zh5uYmfP/99zrXo8tD0/EhCPLXpugH3+KPZs2aqf3AbShr7OgJgvwLlOjoaJ32e7t27bTWUhodvUGDBul1rBSn77Gmbd9du3ZNeOmll0SXDwoKEk6ePKl2Hbp29MQejo6OwrRp0wSZTGbA3jUvjrpJRERkx2rUqIG//voLy5Ytw/r163HmzBnk5OSgUqVKePnll9G/f3/069dPp/v/atasiWbNmikGb6lYsaLiHjwx7777Ll5//XV8//332LlzJy5evIisrCy4uLigSpUqaNiwITp16oSYmBjRYdrF/PnnnwCA1q1bq52u72WbxhIeHo41a9YgNTUVx48fx507d3D//n28ePECFStWRIMGDdChQwcMGzZMNPvQmN59911ERETg66+/xu7du3H79m24u7ujXr166N+/P4YNG2bSQTGsTfny5fHLL7/gwIEDWLlyJf7880/cvn0bT58+hYeHB0JCQtCsWTN069YNnTt3Nne5FqV69eo4duwYkpKSsGHDBpw4cQKZmZkQBAGVKlVCo0aN0LNnTwwcOLBEMSuAfGCV9evXIzU1FUePHsWtW7dw//59PHv2DJ6enqhTpw7atm2LYcOGoUaNGsZ9gqVMIgimuVN08eLFmD9/PjIyMvDSSy/hm2++QbNmzUSXk8lkuH37NsqXLw+JRIKcnBzF8L6FoYWWxtJrtPT6AMuv0dLrAyy/RkuvD7D8Gi29PsDyahQEAY8fP0ZAQIBeA5UQERFZO5N09NatW4eBAwdiyZIlaN68ORYtWoT169fj4sWLot9A3bx5U/RGZCKyDa+//rq5S7B6v/32m7lLsHivv/46bty4YbeZVkREZJ9Mcp59wYIFePfddxUjQS1ZsgRbt27Fjz/+iE8//VTrsuXLlzdFSURkgcLCwnDy5EmNQ1Dv3fsI7dqJn4XZu1eGsDDV+WQyGTIzM+Ht7a32bI41tHfo0EHjPrKW0dgsAd9biIjI3hi9o/f8+XOkpaUhNjZWMc3BwQEdO3ZEamqq6PL84EJkPxwdHSGRSDR29MqV84AuKTDlysng4aG+o/Ts2TN4eHho7EhZeru2fcT/L3XHfUVERPbG6B29+/fvo6CgAL6+vkrTfX19FeGkRUmlUqW8pZycHGOXREREdq74e0thyDgREZGtMvud6XFxcfD09FQ8AgMDzV0SERHZmMDAQKX3mri4OHOXREREZFJGP6NXuXJlODo64u7du0rT7969Cz8/P5X5Y2NjMW7cOMXvOTk5HIyFiIiMqvgooDybR0REts7oZ/ScnZ0RHh6O3bt3K6bJZDLs3r0bERERKvO7uLjAw8ND6UFERGRMxd9n2NEjIiJbZ5JRN8eNG4dBgwahSZMmaNasGRYtWoS8vDzFKJxERAAwefJkre0tWrQAcFR0PS1atIAgpKlMd3BwQFhYGE6cOAGZTGaR7enp6WrbAPmXZBKJRGP+G3PhiIiISBOTdPR69+6NzMxMTJkyBRkZGWjcuDG2b9+uMkALEZE2oaF+OH36OWQyZ43zODg8R2ioH5ydw1XaJBIJatasCQAaR600d3tWVhYEQdA46qbY8kRERETqmKSjBwCjR4/G6NGjTbV6IrIDZ8/+gYYN38SZMxkaz4iFhvrh7Nk/NLYD0HpGzdztFSpU0Jqjd/nyZdH1ExERERVnso4eEZGhZDIZnJ0zIAhpEATVjo4gOMDZORwymUzj5Y+CIFh0e+GlmZo6bWLLExEREanDr4OJiIiIiIhsDDt6RERERERENoYdPSIiIiIiIhvDjh4REREREZGN4WAsRGSxNm/ebO4SDDZr1iytOXiFg62owxw9IiIiKil29IiITIg5ekRERGQO7OgREZmQWA4ec/SIiIjIFNjRIyIyIeboERERkTnw62AiIiIiIiIbw44eERERERGRjWFHj4iIiIiIyMawo0dERERERGRjOBgLEZEJieXgMUePiIiITIEdPSIiE2KOHhEREZkDO3pERCbEHD0iIiIyB3b0iIhMiDl6REREZA78OpiIiIiIiMjGsKNHRERERERkY9jRIyIiIiIisjHs6BEREREREdkYDsZCRGRCzNEjIiIic2BHj4jIhJijR0RERObAjh4RkQkxR4+IiIjMgR09IiITYo4eERERmQO/DiYiIiIiIrIxPKNHRERkQWQyGW7fvo3y5cvzPkwiIgsiCAIeP36MgIAAq7h9gh09IiIiC3L79m0EBgaauwwiItLgxo0bqFq1qrnLEMWOHhERkQUpX748APkHCQ8Pj1LZZn5+Pnbu3InOnTvDycmpVLZpLNZcO2Dd9bN287Hm+q259ocPHyIoKEjx/7SlY0ePiMiEmKNH+iq8XNPDw6NUO3pubm7w8PCwug9e1lw7YN31s3bzseb6rb12wHrijYze0Zs2bRqmT5+uNK1OnTq4cOGCsTdFRGTxmKNHRERE5mCSM3oNGjTArl27/ttIGZ44JCL7xBw9IiIiMgeT9MDKlCkDPz8/U6yaiMiqMEePiIiIzMEkHb1Lly4hICAArq6uiIiIQFxcHKpVq6Z2XqlUCqlUqvg9JyfHFCUREZEdK/7e4uLiAhcXFzNVQ0REZHpGv+6nefPmSEhIwPbt2xEfH4+rV6+idevWePz4sdr54+Li4OnpqXhwSGkiIjK2wMBApfeauLg4c5dERERkUkY/o9e1a1fFz40aNULz5s1RvXp1/Pzzzxg6dKjK/LGxsRg3bpzi95ycHJw8edLYZRERkR0rHlXAs3lERGTrTD5KSoUKFVC7dm1cvnxZbTsvnyEiIlMrzagCIiIiS2DyIdtyc3Nx5coV+Pv7m3pTREQWp+hgK+oeRQdbUfcQW56IiIhIHaOf0Rs/fjx69OiB6tWr4/bt25g6dSocHR3Rt29fY2+KiMjiMUePiIiIzMHoHb2bN2+ib9++ePDgAby9vdGqVSscPnwY3t7ext4UEZHFY44eERERmYPRPyUkJSXh9u3bkEqluHnzJpKSkhASEmLszRARWQVDL80UW56MJz4+Ho0aNVLczxcREYE//vhD0f7s2TOMGjUKXl5eKFeuHGJiYnD37l2ldaSnp6Nbt25wc3ODj48PJkyYgBcvXpT2UyEiIjL9PXpERETWoGrVqpg7dy7S0tJw/PhxtG/fHlFRUTh79iwAYOzYsfj999+xfv167Nu3D7dv30Z0dLRi+YKCAnTr1g3Pnz/HoUOHsHLlSiQkJGDKlCnmekpERGTH2NEjIiIC0KNHD7z22muoVasWateujdmzZ6NcuXI4fPgwsrOzsXz5cixYsADt27dHeHg4VqxYgUOHDuHw4cMAgJ07d+LcuXNYs2YNGjdujK5du2LmzJlYvHgxnj9/rnc9GdlPDX5Od7Kf4tCV+7hj0+t6hkvZEtzJfmZhddnDurjvzbcuy9z3ZFlMHq9ARERkbQoKCrB+/Xrk5eUhIiICaWlpyM/PR8eOHRXz1K1bF9WqVUNqaipatGiB1NRUNGzYEL6+vop5IiMjMXLkSJw9exYvv/yyXjV0WrgfH3Z9CZEN/Er0HHaczcDXey5BEACJBHi/fS2N63rx4gVu5QHn7zxGmTKqHw30WZcx69J9XY5YfH6/BdbFfW++urSvS2y/m6su/dZlnH3/zZ5LkAmAgwSIi26I3k2rlWhdZHkkgrqh3MwoJycH+/btM3cZRERGMXPmTKSlpWkcTOXWrVvw8fHROBhLixYttC6/efNmU5RtU15//XVkZ2frlKN3+vRpRERE4NmzZyhXrhwSExPx2muvITExEUOGDIFUKlWav1mzZmjXrh0+//xzDB8+HNevX8eOHTsU7U+ePIG7uzu2bduGrl27qt2mVCpVWm9OTg4CAwMR+OHPcHBxK+GzJiLSn4MESPmoDfw9XU22jfz8fCQnJ6NTp05wcnIy2XZM4cGDB/D399f5PcXceEaPiMiECgdbUafoYCvqFB2sRdPyZFx16tTBqVOnkJ2djQ0bNmDQoEEm//IxLi4O06dP19ju5iigjJ4v9QsZ8KRANX6D6+K6uC6uS9u6ZAKw4re9aOxl+vNAycnJJt+GsT158sTcJeiFHT0iIhNijp51cXZ2Vuzz8PBwHDt2DF999RV69+6N58+fIysrCxUqVFDMf/fuXfj5yS+Z8vPzw9GjR5XWVzgqZ+E86sTGxmLcuHGK3wvP6AHyb9e3j31V72/X72Q/Q9sv90NW5LDRti5t37Druy5j1mUP6+K+N8+6xM4q2cJzLMm6ACDpqhO8g0IwqEV1OOvbe9SBtZ/Rsybs6BERmRBz9KybTCaDVCpFeHg4nJycsHv3bsTExAAALl68iPT0dERERAAAIiIiMHv2bNy7dw8+Pj4A5N9Ye3h4oH79+hq34eLiAhcXF5XpjhIJ4qIbolrl8nrXXa2yE+KiG2LixjMoEAQ4SiSYEx0qui4nJyeVD14lXZcx69K2rtiNp5XuL7KUurjvzVeXPutSt98toS6xdZli3ztIgIAKZXHz0VPM23EJG9JuY3KP+mhXx0fvdetC0763ZNZWLzt6REQmVDQHT52iOXolWZ6MJzY2Fl27dkW1atXw+PFjJCYmIiUlBTt27ICnpyeGDh2KcePGoVKlSvDw8MCYMWMQERGBFi1aAAA6d+6M+vXr4+2338a8efOQkZGBSZMmYdSoUWo7cmJ2jG2N2oG+4jNq0LtpNbSp7Y1r95+gRmU3+HuWtcl1RQRVxM/b9qLXa+1K9GHXVHXZw7q47823LlPte9/yrvjlxE18vv0i/r2fhyErjqFDXR9M7l4fNSq7l3g7ZB7s6BEREQG4d+8eBg4ciDt37sDT0xONGjXCjh070KlTJwDAwoUL4eDggJiYGEilUkRGRuK7775TLO/o6IgtW7Zg5MiRiIiIgLu7OwYNGoQZM2aUqB4/Az4IFvL3LGvQB0rrWJcrankKRhk8wnKfo6Wui/vefOsy3b5/q0kgIkP98M3uS1hx8Bp2X7iHPy/dx9DWQRjdribcXdh9sBZ8pYiIiAAsX75ca7urqysWL16MxYsXa5ynevXq2LZtm7FLIyIqVR6uTvisW330bloN038/iz8v3Ud8yhVsPHETE1+rh9dfCuB94laAN3gQEREREZGKmj7lsOqdZlg2sAmqVXLD3RwpPkg6hbeWpOLMrWxzl0ci2NEjIiIiIiK1JBIJOtX3xc6xbTAhsg7KOjni+PVH6PHtAUzcdBoP856bu0TSgB09IiITKjrYirpH0cFW1D3EliciIioNrk6OGNWuJvaMfxWvvxQAQQASj6Sj7fy9WHnoGl4UyHAn+ykOXbmPO9lPzV0ugffoERGZFHP0iIjIlvh7lsXXfV/GgBbVMfW3szh/JwdTfzuL71Iu495jKYQisQ+9m1Yzd7l2jR09IiITYo4eERHZomZBlbBlTCv8dDQd87ZfwN0cqaJNJgATN55Bm9reRhtplPTHTwlERCZk6KWZYssTERGZi6ODBANaVMeXb72k0lYgCDjy7wMzVEWF2NEjIiIiIqISC63qCQc1dxN89PNfmLjpNDKyn5V+UcSOHhERERERlZy/Z1nERTeE4//fO+4gAer4lkPB/w/Y8ur8vZiz7TwecYTOUsV79IiIiIiIyCC9m1ZDm9reuHb/CWpUdoO/Z1kcvfoQ83dcwLFrj7B0/79IPJKOd1pWR5UCc1drH9jRIyIiIiIig/l7llUafKVZUCX8PCICKf9kYv72izh3Jwdf77kC9zKOyPa6joGvBMHVydGMFds2XrpJRGRCzNEjIiJ7JpFI0K6OD7aMaYVv+r6MGl5uyHshwZw/LqLdFylIOpqOFwUcXMwUeEaPSIOkpCRcvnxZY37ZpEmTzFCVZRHbRzVr1ixxuzHWYSntAHP0iIjIvjk4SNDjpQB0rOOF6at2IOW+G+5kP8OnG0/j+/3/Ylyn2ujW0B8O6kZ1oRJhR49IA+aXidNlH5W03RjrsIZ25ugREZE9KePogAhfAZMGtMK6E3eweO9lXL2fhzE/nUR8yhVMiKyDOn7lcO3BEwRVdmcOnwHY0SPSoOgldaSe2D4ytL00tmHu9qKXZpZkeSIiImvk4uSIoa2C0LtpIH48cBXL9v+Lc3dyMCThmGIeBwkQF90QvZtWM2Ol1otfBxMRERERkVmUcymD9zvUwv6P26Ffc+UOnUwAJm48gzvZT81UnXVjR4+IiIiIiMyqorszujfyV5leIAi4dv+JGSqyfuzoERERERGR2QVVdoe6sViqe/E+vZJgR4+IiIiIiMzO37Ms4qIbwrHYqNJ/XrpvpoqsGwdjISIiIiIii9C7aTW0qe2Na/ef4ODlTHy79wpmbT2PtnV84Ovhau7yrIreHb39+/dj/vz5SEtLw507d7Bp0yb07NlT0S4IAqZOnYply5YhKysLLVu2RHx8PGrVqmXMuokMNmvWLK3D1qenp2sd9r5ly5ZaowE2b95s7JJLnaH7KDMzs8TtxliHtbRrOoaKBqarw3gFIiKyRf6eZeHvWRZNa1TEn5fu46+b2fhs02ksG9iEGbJ60Lujl5eXh5deegnvvPMOoqOjVdrnzZuHr7/+GitXrkRQUBAmT56MyMhInDt3Dq6u7IWT5TBlkLWtMHQfZWdnl7jdGOuwhXYGphMRkb0q4+iAeW++hO7f/Ild5+/ht79uI6pxFXOXZTX07uh17doVXbt2VdsmCAIWLVqESZMmISoqCgCwatUq+Pr6YvPmzejTp49h1RIZkVgQtSFB1rbC0H0kkUhK3G6MddhCOwPTiYjIntXxK49R7Wpi0a5LmP77ObSqWRle5VzMXZZVMOqnhKtXryIjIwMdO3ZUTPP09ETz5s2RmppqzE0RGaxoELW6R9Ega3UPseVtgaH7yNB2Y6zD2tvt4TgjIiLS5n9ta6KuX3k8zHuOab+fM3c5VsOoHb2MjAwAgK+vr9J0X19fRVtxUqkUOTk5Sg8iIiJjKv4+I5VKzV0SERHpyLmMA+a92QgOEuD3v25j51n1/QpSZvbrfuLi4uDp6al4BAYGmrskIiKyMYGBgUrvNXFxceYuiYiI9NCoagW82yYYADBp8xlkP803c0WWz6gdPT8/PwDA3bt3labfvXtX0VZcbGwssrOzFY8bN24YsyQiIiLcuHFD6b0mNjZWZZ64uDg0bdoU5cuXh4+PD3r27ImLFy8qzfPs2TOMGjUKXl5eKFeuHGJiYlTe89LT09GtWze4ubnBx8cHEyZMwIsXL0z6/IiI7MHYjrURXNkd9x5LMXsrL+EUY9SOXlBQEPz8/LB7927FtJycHBw5cgQRERFql3FxcYGHh4fSg4iIyJiKv8+4uKjeyL9v3z6MGjUKhw8fRnJyMvLz89G5c2fk5eUp5hk7dix+//13rF+/Hvv27cPt27eVRqAuKChAt27d8Pz5cxw6dAgrV65EQkICpkyZUirPk4jIlrk6OeLzNxtBIgF+Pn4Tf17KNHdJFk3vUTdzc3Nx+fJlxe9Xr17FqVOnUKlSJVSrVg0ffvghZs2ahVq1ainiFQICApSy9ogsgVg+WeEgGOqI5ZvZCkP3UVRUlNYRI8PCwrSOXCo2j720a3sNtGUdWtJzMGX7xo0b1T53fW3fvl3p94SEBPj4+CAtLQ1t2rRBdnY2li9fjsTERLRv3x4AsGLFCtSrVw+HDx9GixYtsHPnTpw7dw67du2Cr68vGjdujJkzZ+KTTz7BtGnT4OzsbJRaiYjsVdMalTCwRXWsTL2OT385jZ1j28DdRe8ujV3Qe68cP34c7dq1U/w+btw4AMCgQYOQkJCAjz/+GHl5eRg+fDiysrLQqlUrbN++nRl6ZHGYoyfOlPtIIpGI7kOxedjOfWjKLMHs7GwAQKVKlQAAaWlpyM/PVxpZum7duqhWrRpSU1PRokULpKamomHDhkqDkkVGRmLkyJE4e/YsXn75ZZPVS0RkLz7uUhe7zt/DraynmLf9AqZHhZq7JIukd0evbdu2Wj/YSiQSzJgxAzNmzDCoMCJTY46eOFPuo8JlxM5GaZuH7dyHpjqrLpPJ8OGHH6Jly5YIDZV/gMjIyICzszMqVKigNG/RkaUzMjLUjjxd2KaOVCpVGgW0cPTp/Px85OeXzmADhdspre0ZkzXXDlh3/azdfKy5fmPU7uwAzIqqjyEr07Ay9Tq6NPBBk+oVjVWiRta2v3mek+xW0XwydYrmm5VkeVtg6n2kyz40dB323m4JNZTGczS2UaNG4cyZMzhw4IDJtxUXF4fp06erTN+5cyfc3NxMvv2ikpOTS3V7xmTNtQPWXT9rNx9rrt8YtTf3dsCRTAd8sOYoJjQqgLOjEQrT4smTJ6bdgJGxo0dERFTE6NGjsWXLFuzfvx9Vq1ZVTPfz88Pz58+RlZWldFav6MjSfn5+OHr0qNL6Ckfl1Db6dOFtEID8jF5gYCA6d+5cagOU5efnIzk5GZ06dYKTk1OpbNNYrLl2wLrrZ+3mY831G7P2Vk/z0fWbQ7j3WIpLLjUxoXNtI1Wp3oMHD0y6fmNjR4+IiAjyM4djxozBpk2bkJKSgqCgIKX28PBwODk5Yffu3YiJiQEAXLx4Eenp6YqRpSMiIjB79mzcu3cPPj4+AOTfWnt4eKB+/fpqt+vi4qJ2FFAnJ6dS/wBnjm0aizXXDlh3/azdfKy5fmPU7uXkhFk9QzF8dRp+OHAN3V+qgkZVKxinQDWsbV/b9pCBREREOho1ahTWrFmDxMRElC9fHhkZGcjIyMDTp08BAJ6enhg6dCjGjRuHvXv3Ii0tDUOGDEFERARatGgBAOjcuTPq16+Pt99+G3/99Rd27NiBSZMmYdSoUWo7c0REZJjODfzQ46UAyATg4w1/4/kL272lRl/s6BEREQGIj49HdnY22rZtC39/f8Vj3bp1inkWLlyI7t27IyYmBm3atIGfn59SvIOjoyO2bNkCR0dHREREYMCAARg4cCAHKCMiMqFpPeqjkrszLmQ8RnzKFXOXYzF46SbZLeboiTPlPnJwcBDdh2LzWEP7r7/+qnVk0szMTJO1l8Y2LKG9ZcuWGkfl1Cd+QZeoFFdXVyxevBiLFy/WOE/16tWxbds2nbdLRESG8Srngqk96uODpFP4du8ldAn1Qx2/8uYuy+zY0SO7xRw9cczRM7xdbB9lZ2ebrL00tmEJ7dpeA5lMhrS0NJXpRERkW15/KQC//3UHu87fxccb/sIvI19BGUfb/kJeDDt6ZLeYoyeOOXqGt4vtI4lEYrL20tiGJbRrO85MGahORESWQyKRYPYboThy9QH+upmNHw9exfA2IeYuy6zY0SO7xRw9cczRM7xdbB+Zut0SajB1u7bXgB09IiL74evhiknd6uGTX07jy53/oFN9PwRVdjd3WWZj3+cziYiIiIjIZvRqEohWNStD+kKGT375GzKZ7d5iI4YdPSIiIiIisgkSiQRx0Q3h5uyIo1cfYu2R6+YuyWzY0SMiIiIiIpsRWMkNH0fWAQDM2XYev526hTvZT81cVeljR4+IiIiIiGzKwIgaqF7JDU/zZXg/6RRazt2DdcfSzV1WqeJgLGS3mKMnztQ5etoy5grXYe6MNl0y3AzZR1FRUVpH7QwLCytxuzHWYQnt6enpJT7OJBIJCgoK1LYREZHtuvv4GW48eqL4XSYAEzeeQZva3vD3LGvGykoPO3pkt5ijJ87UOXrali9ch7kz2gzJcDN11iCzCJmjR0RE6l29n4fi47AUCAKu3X/Cjh6RrWOOnjhT5+hpW75wHebOaDMkw83UWYPMImSOHhERqRdU2R0OEih19iQAalR2M1tNpY0dPbJbzNETZ+p9ZAsZcObOGmQWIXP0iIhIlb9nWcRFN8TEjWdQ8P9XfDg6SFBgR3ELtn2DERERERER2aXeTavhwKft8NO7zdE4sAJeyATM/eOCucsqNezoERERERGRTfL3LIuIkMqY/UYoJBJgy993cOzaQ3OXVSrY0SMiIiIiIpvWIMATfZoGAgBm/H4OMju4hJMdPSIiIiIisnkfda6D8i5lcPpWNjacuGnuckyOg7GQ3WKOnjhT5+hpWx4AXFxcrDrDzRj7aPLkyWrbbMmsWbNMtg+Zo0dERIUql3PBmA41MWfbBczfcRGvNfRHORfb7Q7Z7jMjEsEcPXHmztELCwuz6gw3YyxvD5ijR0REpWXQKzWw9kg6rj94gu/2XsbHXeqauySTYUeP7BZz9MSZO0fvxIkTin+tMcPNWDl6ts6U+9BeOstERKQblzKO+Oy1ehi+Og0/HLiKvs2qIbCSbWbrsaNHdos5euLMnaMnk8msOsPNGMvbA+boERFRaepU3xcta3rh4OUHmLPtPOIHhJu7JJOwj6+LiYiIiIiIIP8ScHL3+nCQAH+cycDhfx+YuySTYEePiIiIiIjsSl0/D/RrXg2APG6hwAbjFtjRIyIiIiIiuzO2Y22Udy2Dc3dysP74DXOXY3Ts6BERERERkd3xKueCDzrUAgB8sfMiHj/LN3NFxsXBWMhuMUdPnLlz9BwcHLSuY/PmzdqfgIjo6GizPkddlrcHptyHzNEjIiJtBkbUQOKRdPx7Pw/f7rmM2Nfqmbsko9G7o7d//37Mnz8faWlpuHPnDjZt2oSePXsq2gcPHoyVK1cqLRMZGYnt27cbXCyRMTFHT5yl5+gZSpecPubomR5z9IiIyFycyzhgUvd6eCfhOH48KI9bqFHZ3dxlGYXeHb28vDy89NJLeOeddxAdHa12ni5dumDFihWK311cXEpeIZGJMEdPnKXn6BlKl5w+5uiZniXl6Il9mSkIAqZOnYply5YhKysLLVu2RHx8PGrVqqWY5+HDhxgzZgx+//13ODg4ICYmBl999RXKlSunVy1ERFQ62tXxQZva3tj/TybmbDuPpQObmLsko9C7o9e1a1d07dpV6zwuLi7w8/MrcVFEpYH5Z+LMnTGnS46eIXRZP3P0TM+ScvTEvsycN28evv76a6xcuRJBQUGYPHkyIiMjce7cObi6ugIA+vfvjzt37iA5ORn5+fkYMmQIhg8fjsTERL1qISKi0iGRSDC5Wz10uXwfO8/dxaHL9/FKzcrmLstgJvm6OCUlBT4+PqhTpw5GjhyJBw80Z1NIpVLk5OQoPYiIiIyp+PuMVCpVO1/Xrl0xa9YsvPHGGyptgiBg0aJFmDRpEqKiotCoUSOsWrUKt2/fVtwvev78eWzfvh0//PADmjdvjlatWuGbb75BUlISbt++bcqnSEREBqjlWx4DCuMWtpzDiwLr/wLW6IOxdOnSBdHR0QgKCsKVK1cwceJEdO3aFampqXB0dFSZPy4uDtOnT1ea9ttvvxm7LCIismOBgYFKv0+dOhXTpk3Tax1Xr15FRkYGOnbsqJjm6emJ5s2bIzU1FX369EFqaioqVKiAJk3+u+ynY8eOcHBwwJEjR9R2IKVSqVLHs/ALz/z8fOTnl84IcIXbKa3tGZM11w5Yd/2s3XysuX5Lrn1U2yBsPnULFzIeY+3ha+jXTPm9wxJr1sboHb0+ffoofm7YsCEaNWqEkJAQpKSkoEOHDirzx8bGYty4cYrfc3JycPLkSWOXRUREduzGjRvw8PBQ/F6Se8czMjIAAL6+vkrTfX19FW0ZGRnw8fFRai9TpgwqVaqkmKc4dV94AsDOnTvh5uamd52GSE5OLtXtGZM11w5Yd/2s3XysuX5Lrb2jrwS/XHPEvD/OwTnjNNyK9JaePHlivsJKwOTxCsHBwahcuTIuX76stqPn4uLCwVqIiMikPDw8lDp6lkTdF56BgYHo3LlzqdWcn5+P5ORkdOrUCU5OTqWyTWOx5toB666ftZuPNddv6bV3KpDh1OJUXMnMwz9OIZjYtY6iTdvtaJbI5B29mzdv4sGDB/D39zf1poj0whw9cZaeo2cosfUzR690WEuOXuEgY3fv3lV6T7t79y4aN26smOfevXtKy7148QIPHz7UOEiZpi88nZycSv1DkDm2aSzWXDtg3fWzdvOx5vottXYnJ2BKjwYY9ONRrD6cjrcjaiDYu9z/t1levdro3dHLzc3F5cuXFb9fvXoVp06dQqVKlVCpUiVMnz4dMTEx8PPzw5UrV/Dxxx+jZs2aiIyMNGrhRIYydY7erFmzULNmTVy+fFnjNqyhXdNzZI5e6eToJSUllfg11GUeS2nXtg/E9mFCQgI8PT3VtmdnZ6Nu3boq00siKCgIfn5+2L17t6Jjl5OTgyNHjmDkyJEAgIiICGRlZSEtLQ3h4eEAgD179kAmk6F58+ZGqYOIiEzr1dreaFfHG3svZmL21vNYPripuUsqEb07esePH0e7du0UvxdebjJo0CDEx8fj77//xsqVK5GVlYWAgAB07twZM2fO5OWZZHFMnaNXuJxYRpo1tzNHr3Ry9Erabox1WEK72D6USCQa252dnVWmaaPty8xq1arhww8/xKxZs1CrVi1FvEJAQIAia69evXro0qUL3n33XSxZsgT5+fkYPXo0+vTpg4CAAL1qISIi85nUvT7+vLQfuy/cw/5/MtGmtre5S9Kb3h29tm3bav1mfceOHQYVRFRaSiP/TGwea29njp7p95Gh7aWxDXMfZ9ra9b38VduXmQkJCfj444+Rl5eH4cOHIysrC61atcL27dsVGXoAsHbtWowePRodOnRQBKZ//fXXetVBRETmFeJdDgMjauDHg1cxc8s5/PFBa3OXpDeT36NHRERkLcS+zJRIJJgxYwZmzJihcZ5KlSoxHJ2IyAZ80KEWNp28iUv3crH2SDq61ylv7pL0Yh93+hMREREREenB080J4zrLR91cuOsfZD+xrhw9dvSIiIiIiIjU6Ns0EHV8yyPrST6+//OqucvRCzt6REREREREapRxdMDk7vUBAOuO3zRzNfrhPXpkt0ydo6dLRpu1tzNHz/Q5eoa0G2MdltAutg/F2omIiAzRqlZl1PP3wNlrT8xdil7Y0SO7ZcrsrsJ5srOzNc5jC+3M0TM8R0/bcWToa2SMdVhC+/jx47Xm8CUkJGjcx9nZ2SrTiIiI9HEn+ykuZuSYuwy9saNHdsvQ/DNt2V26zGMP7czRM+w4MrS9NLZh6ceRvjl6RERExV29nweZ8b9vNjl29MhumTK7y1jrsPZ25uiZ/zWwhBrMeRyZ6rJfIiKyH0GV3eEgAaztZgC+AxIREREREWng71kWcdEN4SAxdyX64Rk9IiIiIiIiLXo3rYbQyo4IXWjuSnTHM3pEREREREQi/DxczV2CXtjRIyIiIiIisjG8dJPslimzu3SZxx7amaNneAYcj0PD9zEREZE9YkeP7Jah+WcTJkzQmO1VuI6aNWtqzf+yh3aAOXolPY4MfQ10mWfz5s3Izs7WWGNUVJRVH0fM0SMiInvFjh7ZLUPzz7QtX7gOQHtGm723M0dPfHmx+rS16zKPj4+P1hw7XfaRuduZo0dERKSKHT2yW6bOT9NlHntvZ46e6fex2DxFL1/V1Bk193HCHD0iIiL98R2QiIiIiIjIxrCjR0REREREZGPY0SMiIiIiIrIx7OgRERERERHZGA7GQnbLlPlphesQ24a9tzNHT3x5Q9p1mUcsx84SjhPm6BEREemPHT2yW6bMTwOAyZMnG1TfrFmztG6jMP/M09NTY42W3s4cPfHlJ02apN+TKkbsOLp3757WHD2xfWToca6LpKQkjdtnjh4REZF67OiR3TJ1jp6hxPLLxPLPCs8WWWu7LvvAUNaSo2cIQ48jU78GujBkHzNHj4iI7BU7emS3SiNHzxBi+WVi+WeA+HOw9Hbm6BnO0OPI1K+BLpijR0REpD++AxIREREREdkYdvSIiIiIiIhsDDt6REREJrB48WLUqFEDrq6uaN68OY4ePWrukoiIyI6wo0dERGRk69atw7hx4zB16lScOHECL730EiIjI3Hv3j1zl0ZERHaCHT0iIiIjW7BgAd59910MGTIE9evXx5IlS+Dm5oYff/zR3KUREZGd0GvUzbi4OGzcuBEXLlxA2bJl8corr+Dzzz9HnTp1FPM8e/YMH330EZKSkiCVShEZGYnvvvsOvr6+Ri+eyBCmDkw3lKFB1zKZDFFRUVqHpQ8LC7PYdkCeFWhK1hCYbihTB6aXBmsLTH/+/DnS0tIQGxurmObg4ICOHTsiNTVVZX6pVAqpVKr4vTD77+HDh8jPzzd9wQDy8/Px5MkTPHjwAE5OTqWyTWOx5toB666ftZuPNddvzbU/fPgQgGmyfU1Br47evn37MGrUKDRt2hQvXrzAxIkT0blzZ5w7dw7u7u4AgLFjx2Lr1q1Yv349PD09MXr0aERHR+PgwYMmeQJEJWXqwHRDiQVViwVd6xLGbcntpcEaAtMNZehxZOrQel0Yso/NEZh+//59FBQUqHzB6evriwsXLqjMHxcXh+nTp6tMDwoKMlmNRERUcg8ePICnp6e5yxClV0dv+/btSr8nJCTAx8cHaWlpaNOmDbKzs7F8+XIkJiaiffv2AIAVK1agXr16OHz4MFq0aGG8yokMZA+B6bqEcVtqe2lgYDoD0y1BbGwsxo0bp/hdJpPh4cOH8PLyMkpnXxc5OTkIDAzEjRs34OHhUSrbNBZrrh2w7vpZu/lYc/3WXHt2djaqVauGSpUqmbsUnRgUmF74TWnhk01LS0N+fj46duyomKdu3bqoVq0aUlNT2dEji2LrQde61Gjp7abGwHQGpptC5cqV4ejoiLt37ypNv3v3Lvz8/FTmd3FxgYuLi9K0ChUqmLJEjTw8PKzug1cha64dsO76Wbv5WHP91ly7OW9n0EeJq5TJZPjwww/RsmVLhIaGAgAyMjLg7Oys8gbl6+uLjIwMteuRSqXIyclRehARERlT8feZovfEGZuzszPCw8Oxe/duxTSZTIbdu3cjIiLCZNslIiIqqsQdvVGjRuHMmTNISkoyqIC4uDh4enoqHoGBgQatj4iIqLjAwECl95q4uDiTbm/cuHFYtmwZVq5cifPnz2PkyJHIy8vDkCFDTLpdIiKiQiW6dHP06NHYsmUL9u/fj6pVqyqm+/n54fnz58jKylI6q6fpchVA9d6EnJwcnDx5siRlERERqVX8XpDil0oaW+/evZGZmYkpU6YgIyMDjRs3xvbt2y12BGoXFxdMnTrV5PvFFKy5dsC662ft5mPN9bP20qNXR08QBIwZMwabNm1CSkqKyohg4eHhcHJywu7duxETEwMAuHjxItLT0zVerqLu3gQiIiJjMse9IKNHj8bo0aNLdZsl5eLigmnTppm7jBKx5toB666ftZuPNdfP2kuPXh29UaNGITExEb/++ivKly+vuO/O09MTZcuWhaenJ4YOHYpx48ahUqVK8PDwwJgxYxAREcGBWMji2EOOnthzNHf75MmT1baVFuboGZ6jFx0dLZqXuHHjxpI/AVhfjh4REZEl0KujFx8fDwBo27at0vQVK1Zg8ODBAICFCxfCwcEBMTExSoHpRJaGOXrmz9EzN+boGZ6jVxqvs7Xl6BEREVkCvS/dFOPq6orFixdj8eLFJS6KqDQwR88ycvTMiTl6hufolcbrbOs5ekRERKZgUI4ekTWz9fwzXWo0d7u5MUfPODl6pn4O1pajR0REZAn4DkhERGQj5s6dC4lEgg8//FAx7cqVK3jjjTfg7e0NDw8P9OrVSynM/dq1axg6dCiCgoJQtmxZhISEYOrUqXj+/LnWbbVt2xYSiUTp8d5775Vq7QBQo0YNlTrmzp2rdVvPnj3DqFGj4OXlhXLlyiEmJkZlvaauPSUlRaXuwsexY8c0bssY+33atGkq66hbt66iXZf9k56ejm7dusHNzQ0+Pj6YMGECXrx4oXW7Dx8+RP/+/eHh4YEKFSpg6NChyM3N1at2Y9T/119/oW/fvggMDETZsmVRr149fPXVV6LbLcmxZuzaAag9ZsTizoyx7w2tPSEhQeMxf+/ePY3bNcZ+16X+pUuXom3btvDw8FDcFlBcSfajsf+/0Qc7ekRERDbg2LFj+P7779GoUSPFtLy8PHTu3BkSiQR79uzBwYMH8fz5c/To0UNxhvTChQuQyWT4/vvvcfbsWSxcuBBLlizBxIkTRbf57rvv4s6dO4rHvHnzSrX2QjNmzFCqY8yYMVq3N3bsWPz+++9Yv3499u3bh9u3byM6OrpUa3/llVeUar5z5w6GDRuGoKAgNGnSROs2jbHfGzRooLSOAwcOKNrE9k9BQQG6deuG58+f49ChQ1i5ciUSEhIwZcoUrdvs378/zp49i+TkZEVM1/Dhw/Wu3dD609LS4OPjgzVr1uDs2bP47LPPEBsbi2+//VZ0u/oea8auvdCKFSuU1tGzZ0+t2zTWvjek9t69e6sc85GRkXj11Vfh4+OjdbvG2O9i9T958gRdunTR+n9fSfajMf+/0Rcv3SQiIrJyubm56N+/P5YtW4ZZs2Ypph88eBDXrl3DyZMnFfESK1euRMWKFbFnzx507NgRXbp0QZcuXRTLBAcH4+LFi4iPj8cXX3yhdbtubm4ac3JLo/ZC5cuX17mO7OxsLF++HImJiWjfvj0A+YfmevXq4fDhw3qNEm5I7c7Ozko15+fn49dff8WYMWNEBzEyxn4vU6aM2nXosn927tyJc+fOYdeuXfD19UXjxo0xc+ZMfPLJJ5g2bZrae2PPnz+P7du349ixY4qO7DfffIPXXnsNX3zxBQICAkqt/nfeeUdpmeDgYKSmpmLjxo2ikSj6HGumqL1QhQoVdK7DmPvekNrLli2LsmXLKpbJzMzEnj17sHz5ctHtGmO/a6sfgOKMfEpKitr2kuxHY/5/UxI8o0dERGTlRo0ahW7duil1fgBAKpVCIpEo5dW6urrCwcFB6Zvs4rKzs1GpUiXR7a5duxaVK1dGaGgoYmNj8eTJE7PUPnfuXHh5eeHll1/G/PnztV5CmJaWhvz8fKXt1a1bF9WqVUNqamqp117ot99+w4MHDzBkyBDR7Rpjv1+6dAkBAQEIDg5G//79kZ6eDkC3/ZOamoqGDRvC19dXMU9kZCRycnJw9uxZtdtLTU1FhQoVlM5WduzYEQ4ODjhy5Eip1q+Orse8PseaKWsfNWoUKleujGbNmuHHH3/UOmCiMfe9Mff7qlWr4ObmhjfffFN0u8bY79rq10VJ9qMx/78pCZ7RI7vFHD0H/Prrr1pHhMzMzDRpe8uWLbWOXLp582a1042lNHL0oqKiTJoxJ8bUOXpSqVT0df7jjz9K/gTAHD0xSUlJOHHihNr7ulq0aAF3d3d88sknmDNnDgRBwKeffoqCggLcuXNH7fouX76Mb775RvRsXr9+/VC9enUEBATg77//xieffIKLFy/qdUwbo/b3338fYWFhqFSpEg4dOoTY2FjcuXMHCxYsULvNjIwMODs7o0KFCkrTfX19FfnApVV7UcuXL0dkZCSqVq2qdbvG2O/NmzdHQkIC6tSpgzt37mD69Olo3bo1zpw5o9P+ycjIUOrkFbYXtqmTkZGhcnlemTJlUKlSJb32uzHqL+7QoUNYt24dtm7dqnW7+h5rpqp9xowZaN++Pdzc3LBz507873//Q25uLt5//3212zTWvjf2fl++fDn69eundJZPHWPsd7H6y5cvL7p8Sfajsf6/KSl29MhuMUdP/DmKrd/QdlPvQzGlkaNn7ixBU+fo6XIcGoo5eprduHEDH3zwAZKTk+Hq6qrS7u3tjfXr12PkyJH4+uuv4eDggL59+yIsLEzt/rp16xa6dOmCt956C++++67WbRe9L6Vhw4bw9/dHhw4dcOXKFYSEhJRa7ePGjVP83KhRIzg7O2PEiBGIi4tTOqNmTMbe7zdv3sSOHTvw888/i27b0P0OAF27dlX83KhRIzRv3hzVq1fHzz//LPqh2xIYs/4zZ84gKioKU6dORefOnbXOa4xjzRi1T548WfHzyy+/jLy8PMyfP19jR89YjLnfU1NTcf78eaxevVp0XmP9jWurf+jQoXrVby3Y0SO7xRw98ecotn5D2029D8WURo6e2PpNzdQ5eroch4Zijp5maWlpuHfvHsLCwhTTCgoKsH//fnz77beQSqXo3Lkzrly5gvv376NMmTKKe3uCg4OV1nX79m20a9cOr7zyCpYuXap3Lc2bNwcgf7106XAYs/bidbx48QLXrl1DnTp1VNr9/Pzw/PlzZGVlKX3LfvfuXZ3vATJ27StWrICXlxdef/11nbZf/PkCuu93dSpUqIDatWvj8uXL6NSpk+j+8fPzw9GjR5XWUTiKoKZ96OfnpzKy4osXL/Dw4UOD773St/5C586dQ4cOHTB8+HBMmjRJ7+2KHWumrL14HTNnzoRUKlXb8THVvjek9h9++AGNGzdGeHi43ts1xn4vXr8uSrIfjfH/jSF4jx7ZraLZXOoeRbO51D3EljdU8fyy4o/i+WclqVFseVO3m3ofGrqPjXEcWPpzFHudjHEcGsrQ18iWdejQAadPn8apU6cUjyZNmqB///44deoUHB0dFfNWrlwZFSpUwJ49e3Dv3j2lTsWtW7fQtm1bhIeHY8WKFSXab6dOnQIA+Pv7l2rt6upwcHDQOIpfeHg4nJycsHv3bsW0ixcvIj09HREREaVeuyAIWLFiBQYOHAgnJyedtl/8+QK673d1cnNzceXKFfj7++u0fyIiInD69GmlD73Jycnw8PBA/fr11W4jIiICWVlZSEtLU0zbs2cPZDKZorNaWvUDwNmzZ9GuXTsMGjQIs2fPLtF2xY41U9Wuro6KFStqPLtlqn1f0tpzc3MNOotmjP1evH5dlGQ/GuP/G0PwjB4REZGVKl++PEJDQ5Wmubu7w8vLSzG9cIQ3b29vpKam4oMPPsDYsWMV34QXdvKqV6+OL774ApmZmYp1FX7jfOvWLXTo0AGrVq1Cs2bNcOXKFSQmJuK1116Dl5cX/v77b4wdOxZt2rRRihkwde2pqak4cuQI2rVrh/LlyyM1NRVjx47FgAEDULFiRbW1e3p6YujQoRg3bhwqVaoEDw8PjBkzBhERETqPgGeM2gvt2bMHV69exbBhw1S2Y4r9DgDjx49Hjx49UL16ddy+fRtTp06Fo6Mj+vbtq9P+6dy5M+rXr4+3334b8+bNQ0ZGBiZNmoRRo0YpOhtHjx7FwIEDsXv3blSpUgX16tVDly5d8O6772LJkiXIz8/H6NGj0adPH71H3DS0/jNnzqB9+/aIjIzEuHHjFPdKOTo6wtvbW239uhxrpVH777//jrt376JFixZwdXVFcnIy5syZg/Hjxyu2Yap9b2jthdatW4cXL15gwIABKtsw1X4Xqx+Q30+XkZGhOMN3+vRplC9fHtWqVUOlSpV02o+m+P/GEOzoERER2bCLFy8iNjYWDx8+RI0aNfDZZ59h7Nixivbk5GRcvnwZly9fVhkIpPC+yPz8fFy8eFExuqOzszN27dqFRYsWIS8vD4GBgYiJiSnR5W+G1O7i4oKkpCRMmzYNUqkUQUFBGDt2rNI9PcVrB4CFCxfCwcEBMTExkEqliIyMxHfffVeqtRdavnw5XnnlFaXgZk21G2u/37x5E3379sWDBw/g7e2NVq1a4fDhw4pOjtj+cXR0xJYtWzBy5EhERETA3d0dgwYNwowZMxTzPHnyBBcvXkR+fr5i2tq1azF69Gh06NBBsf6vv/5ar9qNUf+GDRuQmZmJNWvWYM2aNYrp1atXx7Vr19TWr8uxVhq1Ozk5YfHixRg7diwEQUDNmjWxYMECpXtqTbXvDa290PLlyxEdHa0yQIm62o2133Wpf8mSJZg+fbpi/jZt2gCQf2kzePBgAOL70Vz/32giEcw1CoIGOTk52Ldvn7nLIDswc+ZMpKWlabzv59atW/Dx8dF431GLFi00Lg/Ih8o2RM+ePREeHq5xGwUFBbh3716Ja9TlOYqt39B2U+9DMWL72NDjwMHBQXT9ph5Z1NDjyNHR0eDjUGwkOzGG/K1mZWWhYsWKyM7OVuSZERER2QPbvnmBiIiIiIjIDvHSTbJbzNETf45iGXBhYWFa29PT0826D8WI7WNDjwNd1j9r1iyD9rG2dl1qEDuOjHEcGoo5ekRERPpjR4/sFnP0TJ8RZ+59KMYScvRM2a7LPMzRIyIisk3s6JHdYo6e6TPizL0PxVhKjp6p2nWZhzl6REREtokdPbJbRbO51CmazVWS5Q1VPL+suOL5ZSWp0dDlLX0fihHbx4Dp95Gp28XmETuOjHEcGsqQ18jWc/SIiIg04TsgERERERGRjWFHj4iIiIiIyMawo0dERERERGRj2NEjIiIiIiKyMRyMhewWc/RMnxFn7n0oxhjP4ddff9U64mRmZqbZ2nVdB3P0iIiIbA87emS3LD0Djjl6pmeMHD2x18Cc7cZYB3P0iIiIrBM7emS3mKPHHD1j5OiJvQbmbDfGOpijR0REZJ3Y0SO7ZekZcMzRMz1j5OhZeruh62COHhERkXXiOyAREREREZGNYUePiIiIiIjIxrCjR0REREREZGPY0SMiIiIiIrIxHIyF7JalZ8AxR8/0Nm/eLDpPVFSU1hEfw8LCLLZdl3mkUqnW10lsHzFHj4iIyDLp1dGLi4vDxo0bceHCBZQtWxavvPIKPv/8c9SpU0cxT9u2bbFv3z6l5UaMGIElS5YYp2IiI7H0DDjm6FkGQ/eBOdt1mUfsOBLDHD0iIiLLpFdHb9++fRg1ahSaNm2KFy9eYOLEiejcuTPOnTsHd3d3xXzvvvsuZsyYofjdzc3NeBUTGQlz9JijpwtD94E523WZR+w4EsMcPSIiIsukV0dv+/btSr8nJCTAx8cHaWlpaNOmjWK6m5sb/Pz8jFMhkYlYegYcc/Qsg6H7wNztYvMYmoPHHD0iIiLLZNA7YOElMZUqVVKavnbtWlSuXBmhoaGIjY3FkydPNK5DKpUiJydH6UFERGRMxd9npFKpuUsiIiIyqRJ39GQyGT788EO0bNkSoaGhiun9+vXDmjVrsHfvXsTGxmL16tUYMGCAxvXExcXB09NT8QgMDCxpSURERGoFBgYqvdfExcWZuyQiIiKTKvGom6NGjcKZM2dw4MABpenDhw9X/NywYUP4+/ujQ4cOuHLlCkJCQlTWExsbi3Hjxil+z8nJwcmTJ0taFhERkYobN27Aw8ND8buLi4sZqyEiIjK9EnX0Ro8ejS1btmD//v2oWrWq1nmbN28OQH4zvbqOnouLC99wiYjIpDw8PJQ6ekRERLZOr46eIAgYM2YMNm3ahJSUFAQFBYkuc+rUKQCAv79/iQokMhVLz4Bjjp5lMHQfmLNdl3nEjiMxzNEjIiKyTHp19EaNGoXExET8+uuvKF++PDIyMgAAnp6eKFu2LK5cuYLExES89tpr8PLywt9//42xY8eiTZs2aNSokUmeAFFJWXoGHHP0gFmzZqFmzZq4fPmy1hpM3a7tOVpyOwBMnjxZ7fRCzNEjIiKyTXp19OLj4wHIQ9GLWrFiBQYPHgxnZ2fs2rULixYtQl5eHgIDAxETE4NJkyYZrWAiY2GOnuXn6Jk7h87a23XBHD0iIiLbpPelm9oEBgZi3759BhVEVFosPQOOOXrGqcHe28UwR4+IiMg28R2QiIiIiIjIxrCjR0REREREZGPY0SMiIiIiIrIx7OgRERERERHZmBIFphPZAkvPgGOOnvlz6Ky9XRfM0SMiIrJN7OiR3WKOnvlz9BISEuDp6am1M5udna1xHrZrbweAqKgorVmB4eHhJs/RS0pKMmmWIXP0iIiIVLGjR3aLOXrmz9ETy28Tm4ft4vvQ0ONIjDGOQ231GXqcMUePiIjsFTt6ZLeYo2f+HD1d8tcMXYe9txvjONLGGMehKY8z5ugREZG94jsgERERERGRjWFHj4iIiIiIyMawo0dERERERGRj2NEjIiIiIiKyMRyMhewWc/TMn6Mnlt+my3Nku/Z9aOhxJMYYx6GpjzMiIiJ7xI4e2S1T5+jNmjVLp3wwTe3WkKM3adIklelFjR8/vsTPX5d52C5BQkKC1gw8Q48jseNYLIdPLC/RGFmCzNEjIiJSxY4e2S1T5+gZIx9MW7ul5OhpY0h+mi7zsF37a1i4rLZ1iB1HxshzNGc7c/RMo23btgCAlJQUs9ZBhrl27RqCgoKwYsUKDB48uETLzp8/H+PHjzdNgSIGDx6MDRs2IDc31yzbF5OQkIAhQ4bg6tWrqFGjhrnLMZuUlBS0a9cOe/fuVfzfMXjwYKSkpODatWtG2QaPBfV4jx7ZraLZXOoeRbO51D3Eli+eX2bs9uL5ZSWp0dDlDd3H5t6HttAu9hoaehwZ4zg0d7stuHLlCkaMGIHg4GC4urrCw8MDLVu2xFdffYWnT5+aZJvnzp3DtGnTjPZBjEpXQkICJBIJjh8/bu5SsG3bNkybNk10vsKaxR7m6DQdOnQI06ZNQ1ZWlkrbnDlzsHnz5lKvyZS0vRaffvqpWbfPY0F3PKNHRERkwbZu3Yq33noLLi4uGDhwIEJDQ/H8+XMcOHAAEyZMwNmzZ7F06VKjb/fcuXOYPn062rZtq/JhaufOnUbfHpW+6tWr4+nTp3BycjLpdrZt24bFixeLdvbatGmD1atXK00bNmwYmjVrhuHDhyumlStXzhRlanXo0CFMnz4dgwcPRoUKFZTa5syZgzfffBM9e/Ys9bpMbcaMGQgKClKaFhoaavLt8lgwDnb0iIiILNTVq1fRp08fVK9eHXv27IG/v7+ibdSoUbh8+TK2bt1a6nXxkljbIJFI4Orqau4yFIKDgxEcHKw07b333kNwcDAGDBhg1G29ePECMpmMx7KIrl27okmTJqW+XR4LxmEb17QQERHZoHnz5iE3NxfLly9X6uQVqlmzJj744AMA8g8rM2fOREhICFxcXFCjRg1MnDgRUqlUaZkaNWqge/fuOHDgAJo1awZXV1cEBwdj1apVinkSEhLw1ltvAQDatWunuEyq8J68tm3bKu61AeT34EgkEvz888+YPXs2qlatCldXV3To0AGXL19W2b66+8GKrxMApFIppk6dipo1a8LFxQWBgYH4+OOPVZ4Tlcy1a9cUgzoVtX79etSvXx+urq4IDQ3Fpk2bMHjwYI2XyS1dulRx3DVt2hTHjh1TtA0ePBiLFy8GAKVL7ozp1q1b6NmzJ8qVKwdvb2+MHz8eBQUFKs/ziy++wKJFixS1njt3DgCwZ88etG7dGu7u7qhQoQKioqJw/vx5xfLTpk3DhAkTAABBQUGK51C43ry8PKxcuVIxXex+xz/++EOxvfLly6Nbt244e/asUfdJaZBIJGrP0mr6Gy8NPBaU8YweERGRhfr9998RHByMV155RXTeYcOGYeXKlXjzzTfx0Ucf4ciRI4iLi8P58+exadMmpXkvX76MN998E0OHDsWgQYPw448/YvDgwQgPD0eDBg3Qpk0bvP/++/j6668xceJE1KtXDwAU/2oyd+5cODg4YPz48cjOzsa8efPQv39/HDlyRO/nLpPJ8Prrr+PAgQMYPnw46tWrh9OnT2PhwoX4559/LOo+GFuydetW9O7dGw0bNkRcXBwePXqEoUOHokqVKmrnT0xMxOPHjzFixAhIJBLMmzcP0dHR+Pfff+Hk5IQRI0bg9u3bSE5OVrkUzxgKCgoQGRmJ5s2b44svvsCuXbvw5ZdfIiQkBCNHjlSad8WKFXj27BmGDx8OFxcXVKpUCbt27ULXrl0RHByMadOm4enTp/jmm2/QsmVLnDhxAjVq1EB0dDT++ecf/PTTT1i4cCEqV64MAPD29sbq1atVLikMCQnRWO/q1asxaNAgREZG4vPPP8eTJ08QHx+PVq1a4eTJkxY3aEt2djbu37+vNK3w+VsaHguq2NEju2XqHD1j5INZeo6eGEOeny7zsF37a6jLOsSOI2Mch+Zut1Y5OTm4desWoqKiROf966+/sHLlSgwbNgzLli0DAPzvf/+Dj48PvvjiC+zduxft2rVTzH/x4kXs378frVu3BgD06tULgYGBWLFiBb744gsEBwejdevW+Prrr9GpUyeVM22aPHv2DKdOnVJcAlWxYkV88MEHOHPmjN739SQmJmLXrl3Yt28fWrVqpZgeGhqK9957D4cOHdKpA0z6iY2NRZUqVXDw4EHF/U8dOnRA27ZtUb16dZX509PTcenSJVSsWBEAUKdOHURFRWHHjh3o3r07IiIiULt2bSQnJxv9kjtAfsz17t0bkydPBiC/vC8sLAzLly9X+XB/8+ZNXL58Gd7e3oppUVFRqFSpElJTU1GpUiUAQM+ePfHyyy9j6tSpWLlyJRo1aoSwsDD89NNP6Nmzp9IH8AEDBuh8SWFubi7ef/99DBs2TOm+2kGDBqFOnTqYM2eOSe63NUTHjh1VpmmKRDI3Hguq2NEjtcSyszZv3mxw9pW528WyvcQy5rQtXxrPwRJy9MSI5ewlJSVpXD9gG8dZ4SUeJc0i1LaPxF5DwPAcPWPkOU6YMMFsWYPW3tEDgPLly4vOu23bNgDAuHHjlKZ/9NFH+OKLL7B161aljl79+vUVnTxA/m10nTp18O+//xpU85AhQ5Tucyncxr///qt3R2/9+vWoV68e6tatq3RGoX379gCAvXv3sqNnZLdv38bp06cxceJEpUEuXn31VTRs2FBxTBbVu3dvRScPUH7NS8t7772n9Hvr1q3Vnj2MiYlR+mB/584dnDp1Ch9//LHigz0ANGrUCJ06dVL8XRlLcnIysrKy0LdvX6Vj2tHREc2bN8fevXuNuj1jWLx4MWrXrm3uMnTGY0EZO3qklrVnZ5m73RJqkMnMn6MnRiyL0BaOM1NnEVpDjp4heYqGthv7XqDS5OHhAQB4/Pix6LzXr1+Hg4ODolNeyM/PDxUqVMD169eVplerVk1lHRUrVsSjR48MqFh1vYUdgJKs99KlSzh//rzSh7Gi7t27p3+BpFXhcVL8OCqcVvj/QVHGfM1LwtXVVeUY0XQsFx89svD51qlTR2XeevXqYceOHcjLy4O7u7tRar106RKA/76sKK7wb96SNGvWzCyDsZQEjwVV7OiRWsWzs4ornl2ljr23W0IN2l5DYyxvKLH128JxZug+NPQ1NPRv2Rj/F4g9B1O2W3tHLyAgAGfOnNF5GV2fr6Ojo9rphl6Spct6NdVYUFCgtLxMJkPDhg2xYMECtfMHBgYaUCkZi6mOJUO3r07ZsmVNWIm4wv+jVq9eDT8/P5X2MmVs42N50cFPShOPBVW2cUQRERHZoO7du2Pp0qVITU1FRESExvmqV68OmUyGS5cuKQ2YcvfuXWRlZam9t0qMqTrJFStWVBs0fP36daXh1ENCQvDXX3+hQ4cOVt1htyaFx0nxkVI1TdOVpb5+hc/34sWLKm0XLlxA5cqVFWdwtD0HXZ9f4cAcPj4+au99szbq/pafP3+OO3fumKcgA9jqscB4BSIiIgv18ccfw93dHcOGDcPdu3dV2q9cuYKvvvoKr732GgBg0aJFSu2FZ8O6deum97YLP9So65QZIiQkBIcPH8bz588V07Zs2YIbN24ozderVy/cunVLMbhMUU+fPkVeXp5R6yIgICAAoaGhWLVqFXJzcxXT9+3bh9OnT5d4vaY6lgzl7++Pxo0bY+XKlUq1nTlzBjt37lT8XQHan4O7u7tOzy0yMhIeHh6YM2cO8vPzVdozMzP1fg7mFBISgv379ytNW7p0qdnO6BnCVo8FntEjIiKyUCEhIUhMTETv3r1Rr149DBw4EKGhoXj+/DkOHTqE9evXY/Dgwfjggw8waNAgLF26FFlZWXj11Vdx9OhRrFy5Ej179lQaiEVXjRs3hqOjIz7//HNkZ2fDxcUF7du3h4+Pj0HPadiwYdiwYQO6dOmCXr164cqVK1izZo3KMORvv/02fv75Z7z33nvYu3cvWrZsiYKCAly4cAE///wzduzYYTX3DpnTjz/+iO3bt6tML8xfLG7OnDmIiopCy5YtMWTIEDx69AjffvstQkNDlTp/+ggPDwcAvP/++4iMjISjoyP69OlTonUZ2/z589G1a1dERERg6NChiiH1PT09Ma1IRlzhc/jss8/Qp08fODk5oUePHnB3d0d4eDh27dqFBQsWICAgAEFBQWjevLnKtjw8PBAfH4+3334bYWFh6NOnD7y9vZGeno6tW7eiZcuW+Pbbb0vrqRts2LBheO+99xATE4NOnTrhr7/+wo4dOyw2fkGMLR4L7OgRERFZsNdffx1///035s+fj19//RXx8fFwcXFBo0aN8OWXX+Ldd98FAPzwww8IDg5GQkICNm3aBD8/P8TGxmLq1Kkl2q6fnx+WLFmCuLg4DB06FAUFBdi7d6/BHb3IyEh8+eWXWLBgAT788EM0adIEW7ZswUcffaQ0n4ODAzZv3oyFCxdi1apV2LRpE9zc3BAcHIwPPvjAqkYCNKf4+Hi10zUFOffo0QM//fQTpk2bhk8//RS1atVCQkICVq5cWeIg5+joaIwZMwZJSUlYs2YNBEGwmI5ex44dsX37dkydOhVTpkyBk5MTXn31VXz++edKA3Y0bdoUM2fOxJIlS7B9+3bIZDJcvXoV7u7uWLBgAYYPH45Jkybh6dOnGDRokNoP9wDQr18/BAQEYO7cuZg/fz6kUimqVKmC1q1bY8iQIaX1tI3i3XffxdWrV7F8+XJs374drVu3RnJyMjp06GDu0krEFo8FiWBhYRg5OTnYt2+fucuwez179kR4eDjS0tLUDnBQUFCAe/fuwcfHR+MADvbcbgk1yGQytGjRQuNr6ODggFu3bhm0vKGBxTNnztS4fsA2jjND96G2fST2GgLym9MN+Vs2dHld9oG29RvaLpFIIAgCsrOzLXJEOyJr0bhxY3h7eyM5OdncpRCRjvQ6oxcfH4/4+Hhcu3YNANCgQQNMmTIFXbt2BSAPKvzoo4+QlJQEqVSKyMhIfPfdd/D19TV64fZu1qxZWocbDwsLK3F74TzaQpJdXFwM2oatt5fGNqRSqcUHposRC0wvOuKjOpZwHKanp5t0H9pCYLrYcygMty2pWbNmaR0V1RrvFyEyl/z8fEgkEqVR/1JSUvDXX39h1qxZZqyMiPSlV0evatWqmDt3LmrVqgVBELBy5UpERUXh5MmTaNCgAcaOHYutW7di/fr18PT0xOjRoxEdHY2DBw+aqn67JRZkbUi7MdZh7+2lsQ1rCEwXI7YPDQ3zLo3jwNT70JDtA5YRmG7O40gmkyEtLc3gbRDZi1u3bqFjx44YMGAAAgICcOHCBSxZsgR+fn4qYdREZNn06uj16NFD6ffZs2cjPj4ehw8fRtWqVbF8+XIkJiYqwv9WrFiBevXq4fDhw2jRooXxqiaTBhAbYx323l4a2zBGULW1B6aLhXmXxnFg6n1oL4HphtC2fksd1p3IUlWsWBHh4eH44YcfkJmZCXd3d3Tr1g1z586Fl5eXucsjIj2UeDCWgoICrF+/Hnl5eYiIiEBaWhry8/OVsiDq1q2LatWqITU1lR09IzN1AHFpbMPW2029DWMEVVt7YHrRywbN9TozMN3w49BQthqYTmQOnp6eWLdunbnLICIj0Lujd/r0aURERODZs2coV64cNm3ahPr16+PUqVNwdnZGhQoVlOb39fVFRkaGxvVJpVJIpVLF7zk5OfqWREREpFXx9xYXFxe4uLiYqRoiIiLT0/uamTp16uDUqVM4cuQIRo4ciUGDBuHcuXMlLiAuLg6enp6KR2BgYInXRUREpE5gYKDSe01cXJy5SyIiIjIpvc/oOTs7K258Dw8Px7Fjx/DVV1+hd+/eeP78ObKyspTO6t29exd+fn4a1xcbG4tx48Ypfs/JycHJkyf1LYuIiEijGzduKMUr8GweERHZOoMD02UyGaRSKcLDw+Hk5ITdu3cjJiYGAHDx4kWkp6cjIiJC4/K8fIaIiEzNw8ODOXpERGRX9OroxcbGomvXrqhWrRoeP36MxMREpKSkYMeOHfD09MTQoUMxbtw4VKpUCR4eHhgzZgwiIiI4EEsJiOXkpaenax3lLjMzs8TtxliHvbcbYx0uLi4mzy+z9hw9sYw3qVRaKseBKfehPeToiTHk/8OsrCyOFEhERHZJr47evXv3MHDgQNy5cweenp5o1KgRduzYgU6dOgEAFi5cCAcHB8TExCgFppP+DM3uEsu10tZujHXYe7sx1lEa+WW2nqOnyz4ydTtz9My7D7Kzs0XXT0REZIv06ugtX75ca7urqysWL16MxYsXG1QUGZadVfgNeknbjbEOe283xjpKI7/M1nP0dNlHpm5njp5594Gzs7Po+omIiGyRwffokWkYmp1laHtpbMPW2w1dR2nkl1ly/hlQOhlvpm5njp5594ExvpAgIiKyRnwHJCIiIiIisjHs6BEREREREdkYdvSIiIiIiIhsDDt6RERERERENoaDsVgoQ/PPoqKitI5SFxYWpnW0Q7F52G74PpRKpWbPL7P1HD2x5V1cXErlONG2D8Uy4sSWt4ccPUOPYyIiInvEjp6FMnX+mVh+mdg8bDd8H1pCfpm95+iJLW/u40iXduboMUePiIhIHXb0LFRp5J+JnY3SNg/bDd+HlpBfZu85emLLm/s40qWdOXrM0SMiIlKHHT0LZer8M11yrQxdh723i81jCfll9p6jJ7a8LjWau505eszRIyIiUocdPSIiIiIie5SXDkjvm7sKMhaXyoB7NcWv7OgREREREdmbvHTg9zqA7Jm5KyFjcXAFelxUdPZ4TQsRERERkb2R3mcnz9bInimdoWVHj4iIiIiIyMZY5aWb2nKnAGDz5s2lW5AJmDr/bPLkyUarldSLjo62+Pwye8/RE1vewcEBv/76q9YRJTMzM83ezhw95ugREREVZ5UdPbHsLVtg6flnJM4a8svMfRxZQ46e2D4Sew3M2a7Lc7SE41AMc/SIiIj0Z5UdPbHsLVtg6flnJM4a8svMfRxZQ46e2D4Sew3M2a7Lc7SE41AMc/SITOPatWsICgrCihUrMHjw4BItO3/+fIwfP940BZaSadOmYfr06cjMzETlypXNXQ7ZsJRzQLvZwPr3gTebm357VvmJv2imkrqHLRB7jkVzo9Q97GEfWbri+WPFH8Xzx4o/DF1el+PA3MeR2PpNvY/EltdlH1l6uzUch4YeJ2LbJ7JHCQkJkEgkOH78uLlLwbZt2zBt2jSd52/bti0kEoni4ezsjKCgIAwfPhw3btwwXaGkkLAPkPQHjv+rvr3tLCD0k9KtyRSm/SJ/noUPhwGA/yig+3zg8CVzV2c4qzyjR0RERESGqV69Op4+fQonJyeTbmfbtm1YvHixXp29qlWrIi4uDgDw/PlznDt3DkuWLMGOHTtw/vx5uLm5mahaskfxQ4ByroBMAG48AJbtBdrMBI7OABrXMHd1JceOHhEREZEdkkgkcHV1NXcZanl6emLAgAFK04KCgjB69GgcPHgQnTp10rhsXl4e3N3dTV0i2ZA3mwOVy//3e88m8jOW649q7+g9ew44lwEs9eIRCy2LiIiIiEzp2rVrkEgkSEhIUJq+fv161K9fH66urggNDcWmTZswePBg1KhRQ+16li5dipCQELi4uKBp06Y4duyYom3w4MFYvHgxAChdjlkSfn5+AIAyZf47TzFt2jRIJBKcO3cO/fr1Q8WKFdGqVSsAwN9//43BgwcjODgYrq6u8PPzwzvvvIMHDx6Ibuv69euoWbMmQkNDcffuXQBAVlYWPvzwQwQGBsLFxQU1a9bE559/zltiilhzAAj/DCg7GKg0HOjzjfwMWVF/XgDe+gqo9j7gMggIHAOMXQ08ff7fPF9slV9KeT1TdRuxSYDzQOBRHjB1A+A0EMjMUZ1v+A9AhXflnTF9+XnK/y1TpKeUck5eU1IqMOlnoMpowO0dIOcp8DAXGL8WaPgJUO4dwGMo0PVz4K/r4tuS5ssvFfUcBhz6Rz5NJgMW/QE0+BhwHQz4jgRGLJc/Z33wjB4RERERAQC2bt2K3r17o2HDhoiLi8OjR48wdOhQVKlSRe38iYmJePz4MUaMGAGJRIJ58+YhOjoa//77L5ycnDBixAjcvn0bycnJWL16tc51FBQU4P59efBzfn4+zp8/j6lTp6JmzZpo2bKlyvxvvfUWatWqhTlz5ihG6E1OTsa///6LIUOGwM/PD2fPnsXSpUtx9uxZHD58WGOH88qVK2jfvj0qVaqE5ORkVK5cGU+ePMGrr76KW7duYcSIEahWrRoOHTqE2NhY3LlzB4sWLdL5uVmT7CfA/ceq0/NfqE6bvRmYvAHo1RwY1k7e+fpmp/wSyJOzgQr/f5J1/RHgyXNgZAfAqzxw9Ip8vpsPgfUfyOfp1Rz4+Cfg5yPAhO7K2/n5CNC5IVDRHXi7FTBjE7DuMDC683/zPH8BbDgKxDQFXHUYk+thrvxfmQy49QiYuQlwdQJ6tVCdd+Ym+Vm88d3knTTnMsC5W8DmNOCtZkCQD3A3G/h+N/DqLODcPCCgovrtPn0ORH0JHL8K7IoFmobIp49YDiT8CQxpA7wfCVzNBL7dCZy8BhycCjjp2IOzyo6etlwrQP6flLWz9PwzEld0kAp1ig5yYYrldTkOzH0cWUOOntg+ioqK0joiZFhYmNnaAfFcUUs4DsUwR4+o9MTGxqJKlSo4ePAgypUrBwDo0KED2rZti+rVq6vMn56ejkuXLqFiRfkn2Tp16iAqKgo7duxA9+7dERERgdq1ayM5OVnlUkxtLly4AG9vb6Vp9erVw86dO9WOpvvSSy8hMTFRadr//vc/fPTRR0rTWrRogb59++LAgQNo3bq12u126NABVapUwY4dOxTPa8GCBbhy5QpOnjyJWrVqAQBGjBiBgIAAzJ8/Hx999BECAwN1fn7WomOc5rYGVf/7+XomMPUXYNZbwMSo/6ZHNwVe/gz4btd/0z/vC5Qt8hIObw/U9AUm/gyk3weqVZY/WtSUd+CKdvSOXQH+vQdMi5b/XtMPiKglP5NYtKO39aT87NfbrXR7nnWKDRxbwQ3YPE75ORZ6lg8cn6X8HBoGAv98oXwJ59utgLoTgOUpwOQ3VNeT+wzo/gVw9iawZ+J/l4geuAj8kAKs/R/Qr8h3Gu3qA10+l3eU+6l+16GWVXb0tGUm2QpLzz8jcdaQX2bu48gWcvTEljdnuy4s4TgUwxw9otJx+/ZtnD59GhMnTlR08gDg1VdfRcOGDZGTo3p9XO/evRWdIQCKztO//2oYrlFHNWrUwLJlywAAL168wMWLFzFv3jx07doVf/75p0on8L333lNZR9myZRU/P3v2DLm5uWjRQn6K5sSJEyodvTNnzqB3796oWbMm/vjjD3h4eCja1q9fj9atW6NixYqKM40A0LFjR8ydOxf79+9H//79DXrOlmjxYKC2v+r0j9YCBUW+R9t4XD6QSa/mymcA/SoAtXyBvef+6+gV7SDlPQOe5gOv1AIEQX7Gqtr/J1z0bgF8uBq4chcI8ZVPW3cYcHECosL/W8fAVsDIFcrzrT0IBHoBr9bT7Xn+8iHgUVZew62HQPxuIGYRsPNT4JXayvMOaq38HAB5TYUKZEBWnnxwlzr+wIlrqtvLfgp0nivvtKZMUu5Qrj8CeLoBnRoq78vwIPk6956z8Y6etswkW2Hp+Wckzhryy8x9HNlCjp4u+8hc7bqwhONQDHP0iErH9evyG4oKv1wpqmbNmor/D4qqVq2a0u+Fnb5Hjx4ZVIu7uzs6duyo+L1Lly5o1aoVmjRpgrlz5+LLL79Umj8oKEhlHQ8fPsT06dORlJSEe/fuKbWp+xKoR48e8PX1xY4dO5Q6ugBw6dIl/P333yodzELF128rmoUATYJVp1d0V+6EXMqQd5JqfaQ6L6B8qWH6fWDKBuC3E6r3nGU//e/nt5oD49bIO3cTo+TrX38E6PoS4FFk0NXeEcCHa+SduynR8stNt5wCxnYBdP1Ouk1d5cFY3mwufy5jVgJps5XnDVJzCMhkwFc7gO+S5ZdZFu0Ee5VTnf/D1fIzgydnq541vJQhfw4+I9XXek/N/YiaWGVHr2hmkq0qmhuljtg+EFueTK94/lhxxfPHjL08YPnHkdj6Tb2PxJYHDN9H5m4XYwnHoRhDXiNbfp8gsgSOjo5qp5f0KgNtwsPD4enpif3796u0FT17V6hXr144dOgQJkyYgMaNG6NcuXKQyWTo0qWL2v9PYmJisHLlSqxduxYjRoxQapPJZOjUqRM+/vhjtbXVrl1b7XR7IZPJO1V/fAw4qvlvt5yL/N8CGdBprvyeuE96AHX9AXdX+Vm0wd/L11MooCLQui7w8/939A5fBtIfyC/9LKqiO9C9MbD2kLyjt+Go/N65ATpetqlOOVegeQjwa5r8rKN7kcFpi5/NA4A5v8rvT3znVWDmW0Ald/llnB+ulp/pLC4qXD6oy9zfgVXvKV/yKRMAHw9g7Sj1tXmXVz9dHavs6BERERGRcRXeg3f58mWVNnXTdGXMW0YKCgqQm5srOt+jR4+we/duTJ8+HVOmTFFMv3RJcwr2/PnzUaZMGfzvf/9D+fLl0a9fP0VbSEgIcnNzlc4y0n9CfOVn3IK81V/qWeh0OvDPHWDle8DAIlfOJp9WP3/vFsD/VgAXb8vP7Lm5AD1eVp1vYGsgaoH8Hr61B4GXa6i/v04fLwrk/+ZKlTt66mw4Kr+Hbvlw5elZecpnCgv1DJcPKDP4e6C8KxD/zn9tIT7ArjNAy9rqO5X64FedRERERISAgACEhoZi1apVSp2pffv24fRpDZ/EdVCYaZeVlWVQfXv37kVubi5eeukl0XkLzzQWP7OobXRMiUSCpUuX4s0338SgQYPw22+/Kdp69eqF1NRU7NixQ2W5rKwsvHihZhhKOxLdVH4mb/pGeYevKEEAHvz/ZZ6FZ/uKziMIwFfb1a835v/X+1Oq/LLN7o3Vd7q6viTvUH3+O7DvPDBAx3vYNHmYCxy6JL/H0MdDdHY4Oqg+7/VH5CN4ajKwNfD1QGDJbuCTn/6b3quF/MznzE2qy7wokHcedcUzekREREQ26scff8T27aqfoj/44AO188+ZMwdRUVFo2bIlhgwZgkePHuHbb79FaGioTmfS1AkPl4+c8f777yMyMhKOjo7o06eP1mWys7OxZs0aAP8NxhIfH4+yZcvi008/Fd2mh4cH2rRpg3nz5iE/Px9VqlTBzp07cfXqVa3LOTg4YM2aNejZsyd69eqFbdu2oX379pgwYQJ+++03dO/eHYMHD0Z4eDjy8vJw+vRpbNiwAdeuXUPlypV13CO2J8RXPuJm7DrgWqY8cLy8q/x+tU3H5SNrju8G1A2Qzzs+Ud4J8igL/HJUcz6cj6f8TNmCbcDjZ/L78dRxKgP0iZBHEDg6AH01zKfJhiPyyzUFAbidJR8p81EesOQd3e7z6/6yPOZhyPfygWVO35BfShrso3250Z3lOXyf/SwfgGVilHwAmRHtgbjfgFPX5Wf+nByBS3flncev3pbfQ6gLdvSIiIiIbFR8fLza6YMHD1Y7vUePHvjpp58wbdo0fPrpp6hVqxYSEhKwcuVKnD17tkQ1REdHY8yYMUhKSsKaNWsgCIJoR+/mzZt4++23AcjPtFWsWBGvvvoqpk6disaNG+u03cTERIwZMwaLFy+GIAjo3Lkz/vjjDwQEBGhdzsnJCRs2bEDXrl0RFRWFXbt2oXnz5ti3bx/mzJmD9evXY9WqVfDw8EDt2rUxffp0eHp66lSTLfv0dfllmwv/kJ/ZA+QjX3ZuCLweJv/dqQzw+0fA+6vkHRlXJ+CNJvIOz0ux6tfbu4X8UsbyrsBrWk7mDmwl7+h1aAD4a8it02Tkiv9+dncBGlUDZveSDwiji4lRQJ4USDwkv8Q0rAawdTzw6Trdls1+8v+dvbLAqM7AkqHyUTa/3yOPnSjjANTwlp+pbKnH7aASwRR3yxogJycH+/bt0zpPkyZNbD5Hb9asWVpHmUtPT9c6yl3Lli21Lr9x40aT1G1JxPahqfPLxOaRSqXIzMzU+Dq6uLgYtLwux4G5jyNtrxFg+n0ktrwl7CNT69q1q9mPQ0P/FrW9RllZWfDy8kJ2drbSUOlEpJ/GjRvD29sbycnJ5i6FjOXhCWB7uPh8VuKv60DjifLBTd5WjUi0H13SgErynrVeZ/Ti4+MRHx+Pa9euAQAaNGiAKVOmoGvXrgCAtm3bqnTSRowYgSVLlhih6v8wR8/8+WfWwNz5ZWLzWEJ+mbmPI3vI0bN0lnAcGtrOHD0i48nPz4dEIkGZMv99RExJScFff/2FWbNmmbEyIu2W7ZVffhnd1NyVWA69OnpVq1bF3LlzUatWLQiCgJUrVyIqKgonT55EgwYNAADvvvsuZsyYoVjGzc1N0+pKjDl65s8/swbmzDfTZR5LyC8z93FkLzl6lswSjkNt69elnTl6RMZz69YtdOzYEQMGDEBAQAAuXLiAJUuWwM/PT20oOZG5/X4COHcLWLpHfgmo2AiZ9kSvjl6PHj2Ufp89ezbi4+Nx+PBhRUfPzc0Nfn5+xqtQDebomT//zBpYQn4Zc/S0s4ccPUtnCcehoe3M0SMynooVKyI8PBw//PADMjMz4e7ujm7dumHu3Lnw8vIyd3lEKsasBO5mA681BqbHmLsay1LiwVgKCgqwfv165OXlISLiv6Ft1q5dizVr1sDPzw89evTA5MmTtZ7Vk0qlkEqlit9zcvSIeyciItJB8fcWFxcXuLi4mKkaIsvl6emJdet0GEGCyEJc+8rcFVguvb/qPH36NMqVKwcXFxe899572LRpE+rXrw8A6NevH9asWYO9e/ciNjYWq1evxoABA7SuLy4uDp6enopHYGBgyZ4JERGRBoGBgUrvNXFxceYuiYiIyKT0PqNXp04dnDp1CtnZ2diwYQMGDRqEffv2oX79+hg+/L84+IYNG8Lf3x8dOnTAlStXEBISonZ9sbGxGDdunOL3nJwcnDx5sgRPhYiISL0bN24ojbrJs3lERGTr9O7oOTs7K0ZACw8Px7Fjx/DVV1/h+++/V5m3eXN5+MTly5c1dvR4+QwREZmah4cH4xWIiMiuGByYLpPJlO6xK+rUqVMAAH9/f0M3o6TwxnxbVji4gDoODg5a94FMJhNd3h6I7QNTtusyT9FBLkyxvC7HgbmPI7F9aOp9pMvrbO59ZGqWcBwa2i62fSIiInukV0cvNjYWXbt2RbVq1fD48WMkJiYiJSUFO3bswJUrV5CYmIjXXnsNXl5e+PvvvzF27Fi0adMGjRo1MmrRzNGz/WwvXSQlJeHy5csan2NCQgI8PT017qPs7GyTtesyjyXklxl6HIm9BjVr1hRt17R+XZ4jc/QMVxrHoan/FpmjR0REpEqvjt69e/cwcOBA3LlzB56enmjUqBF27NiBTp064caNG9i1axcWLVqEvLw8BAYGIiYmBpMmTTJ60czRs/1sL10Yuo/Esr8MaTfGOqwlR8+UWYTM0TO90sjRM2c7c/SIiDRwqQw4uAKyZ+auhIzFwVX+uv4/vTp6y5cv19gWGBiIffv2lbwwPTBHz/azvXRh6D4ydbuh67CGHD1TtzNHz/RKI0fPnO22/D5BRGQQ92pAj4uA9L65KyFjcaksf13/n8H36BERERERkRVyr6bUMSDbwq86iYiIiIiIbAw7ekRERERERDaGHT0iIiIiIiIbY5X36DFHz/azvXRh6D4yZbsx1mENOXqmziJkjp7plcZxaO52IiIie2SVHT3m6Jk2/6xwHl0y0Mzdru05iu2jCRMmmKw+Y6xj48aNatdbqDRy9MydRcgcPdNr2bKlQceh2PKl8X9BQkICc/SIiIiKscqOHnP0TJt/pss81tBuzow4Y61Dm9LI0TNn/pku8zBHz3Bi+8DQ5c39t84cPSIisldW2dFjjp7p88+MsQ5zt5s7I85Y69CkNHL0zN0uNg9z9Aynyz4wdHlz/q3b8vsEERGRNnwHJCIiIiIisjHs6BEREREREdkYdvSIiIiIiIhsDDt6RERERERENsYqB2Nhjp5p8890mcca2s2ZEWesdWhTGjl6lp5FyBw9wxl6HFrD3zoREZE9ssqOHnP0TJt/pss8ltA+YcIEAKbLP5s0aZLKdEuiS46e2HMYP3682bMQtWURbt68mTl6JiaWk2fq5XWRlJQEoGSvEXP0iIjIXlllR485eqbNPzPGOkqjvTSyBi2ZLjl6YkyZJWiMHEFDc/CYo2cbDPlbZ44eERHZK6v8FFI0M0ndwxYUzY1S9xDbB4a2G2Mdpm43dB+JLW/pdNlHYsT2gbnbxZ6j2DoMXd4ejiNrYOhrRET2Yfv27WjcuDFcXV0VZ/uJAODatWuQSCRISEgo8bJffPGF8QszMb4DEhEREdkQiUSi0yMlJcWsdbZt2xahoaFGWdeDBw/Qq1cvlC1bFosXL8bq1avh7u6OOXPmYPPmzUbZhr2wpuOnaD1ly5ZFo0aNsGjRIrN90bpt2zZMmzbNLNtWxyov3SQiIiIi9VavXq30+6pVq5CcnKwyvV69eqVZlkkdO3YMjx8/xsyZM9GxY0fF9Dlz5uDNN99Ez549zVeclbGm46dq1aqIi4sDANy/fx+JiYkYO3YsMjMzMXv2bMV81atXx9OnT+Hk5GTSerZt24bFixdbTGePHT0iIiIiGzJgwACl3w8fPozk5GSV6SX17NkzODs7W9Sl0ffu3QMgH8eBDGNNx4+np6dSXe+99x7q1q2Lb775BjNmzICjoyMA+VlKV1dXg7dnbSznL5SIiIiISkWNGjUwePBglelt27ZF27ZtFb+npKRAIpEgKSkJkyZNQpUqVeDm5oacnBwMHjwY5cqVw61bt9CzZ0+UK1cO3t7eGD9+PAoKCoxW6x9//IHWrVvD3d0d5cuXR7du3XD27FmlmgcNGgQAaNq0KSQSCQYPHgyJRIK8vDysXLlScXmfuudM+rPU48fV1RVNmzbF48ePFZ1/QPM9euvXr0f9+vXh6uqK0NBQbNq0CYMHD0aNGjXUrn/p0qUICQmBi4sLmjZtimPHjinaBg8ejMWLFwNQvvzVnKzyjB5z9Eybf2aMdZRGu73nn+myj8SYO/9MLL9N7DmKrcPQ5e3hOLIGzNEjMr+ZM2fC2dkZ48ePh1QqVYxoW1BQgMjISDRv3hxffPEFdu3ahS+//BIhISEYOXKkwdtdvXo1Bg0ahMjISHz++ed48uQJ4uPj0apVK5w8eRI1atTAZ599hjp16mDp0qWYMWMGgoKCEBISgo4dO2LYsGFo1qwZhg8fDgAICQkxuCbSX2keP4WdOrGzu1u3bkXv3r3RsGFDxMXF4dGjRxg6dCiqVKmidv7ExEQ8fvwYI0aMgEQiwbx58xAdHY1///0XTk5OGDFiBG7fvq32MldzscqOnj3k6Jkz/8wY6zBGe0JCgkH5Zbaef6ZLjp4Ya88KZI6efTDkNWKOHpFxPHv2DMePH0fZsmVVpvfu3RuTJ08GIL90LiwsDMuXLze4o5ebm4v3338fw4YNw9KlSxXTBw0ahDp16mDOnDlYunQpOnXqhFu3bmHp0qXo2rUrmjRpAgCIiIjAe++9h+DgYKNddkglY6rjp6CgAPfv3wcgH5Bn+fLlOH78OLp166ayreJiY2NRpUoVHDx4EOXKlQMAdOjQAW3btkX16tVV5k9PT8elS5dQsWJFAECdOnUQFRWFHTt2oHv37oiIiEDt2rWNepmroayyo2cPOXpiTJl/Zox1GKPd0PwyW88/M0aOnqVjjh4BzNEjsgSDBg3S+MH5vffeU/q9devWRjmjkZycjKysLPTt21fxYR4AHB0d0bx5c+zdu9fgbVDpMNXxc+HCBXh7eytNe/3117F8+XKty92+fRunT5/GxIkTFZ08AHj11VfRsGFD5OTkqCzTu3dvRSevsE4A+Pfff3Wq1RyssqNXNDPJXhXNlTJFe2lsQ6xd7HU29fKWrnhGnC0Se47Fc/CMvTxg+8eRNTDkNbLVvw2i0hYUFKR2uqurq8oH7YoVK+LRo0cGb/PSpUsAgPbt26tt9/DwMHgbVDpMdfzUqFEDy5Ytg0wmw5UrVzB79mxkZmaKDrxy/fp1AP9dMVJUzZo1FV8EF1WtWjWVOgEY5Vg3Favs6BERERFRyWm6tLygoEAxUmFRms7GqJvXWAq/3Fm9ejX8/PxU2suU4cdYc7GU48fd3V0pTqNly5YICwvDxIkT8fXXXxu07uI01arpNihLwL8QIiIiIjtTsWJFZGVlqUy/fv06goODS78gNQoHTvHx8VH6MK8P3ittGpZ6/DRq1AgDBgzA999/j/Hjx6uchStUeA/e5cuXVdrUTdOVpR1vvKaFiIiIyM6EhITg8OHDeP78uWLali1bcOPGDTNWpSwyMhIeHh6YM2cO8vPzVdozMzNF1+Hu7q62Q0KGseTj5+OPP0Z+fj4WLFigcZ6AgACEhoZi1apVyM3NVUzft28fTp8+XeJtu7u7A4DFHHM8o0dERERkZ4YNG4YNGzagS5cu6NWrF65cuYI1a9aUevxAZmYmZs2apTI9KCgI/fv3R3x8PN5++22EhYWhT58+8Pb2Rnp6OrZu3YqWLVvi22+/1br+8PBw7Nq1CwsWLEBAQACCgoLQvHlzUz0du2Epx4869evXx2uvvYYffvgBkydPhpeXl9r55syZg6ioKLRs2RJDhgzBo0eP8O233yI0NFSp86eP8PBwAMD777+PyMhIODo6ok+fPiV+LobiGT0iIiIiOxMZGYkvv/wS//zzDz788EOkpqZiy5YtqFq1aqnWce/ePUyePFnlsWLFCgBAv379sHv3blSpUgXz58/HBx98gKSkJDRu3BhDhgwRXf+CBQsQHh6OSZMmoW/fvoiPjzf1U7ILlnL8aDJhwgTk5eXhm2++0ThPjx498NNPP+H58+f49NNPsXHjRiQkJKBOnTqig7loEh0djTFjxmD79u14++230bdv35I+BaOQCAbcQTh37lzExsbigw8+wKJFiwDI8zA++ugjJCUlQSqVIjIyEt999x18fX11WmdOTg727dundZ4mTZpojVfYunWrXs+DSl90dDTCwsK0Dpmenp6udVj7li1bmnT5jRs3lvDZlY6uXbsiMzPToOco9hoY2m7oPhR7ji4uLlprkEqlBi1vD8eRNZg1a1aJX6OsrCx4eXkhOzubI/QREZGoxo0bw9vbG8nJyeYuxWAlvnTz2LFj+P7779GoUSOl6WPHjsXWrVuxfv16eHp6YvTo0YiOjsbBgwcNLraQPQSm27rSCKq29aBrXQLTxZ6jqdsNxcB0AhiYTkRExpefnw+JRKI0emtKSgr++usvtZcTW6MSdfRyc3PRv39/LFu2TGlHZGdnY/ny5UhMTFRknqxYsQL16tXD4cOH0aJFC6MUzcB061caQdW2HnStS2C6Ls/R1O2GYGA6AQxMJyIi47t16xY6duyIAQMGICAgABcuXMCSJUvg5+enEuJurUrU0Rs1ahS6deuGjh07KnX00tLSkJ+frzQEbt26dVGtWjWkpqYaraNn6yHR9qA0gqptPehal8B0c4feG4qB6QQwMJ2IiIyvYsWKCA8Pxw8//IDMzEy4u7ujW7dumDt3rsYBXKyN3h29pKQknDhxAseOHVNpy8jIgLOzMypUqKA03dfXFxkZGWrXJ5VKIZVKFb/n5OToWxIREZFWxd9bXFxc4OLiYqZqiIjI3Dw9PbFu3Tpzl2FSen3VeePGDXzwwQdYu3ZtiUejKS4uLg6enp6KR2BgoFHWS0REVCgwMFDpvSYuLs7cJREREZmUXh29tLQ03Lt3D2FhYShTpgzKlCmDffv24euvv0aZMmXg6+uL58+fq4QE3r17F35+fmrXGRsbi+zsbMXDEoIWiYjItty4cUPpvSY2NtbcJREREZmUXpdudujQQSUtfsiQIahbty4++eQTBAYGwsnJCbt370ZMTAwA4OLFi0hPT0dERITadfLyGSIiMjUPDw/GKxARkV3Rq6NXvnx5hIaGKk1zd3eHl5eXYvrQoUMxbtw4VKpUCR4eHhgzZgwiIiKMNhAL8N+N+WS9ig6Soald2+ssk8lMvrylM0YGnFgOn6HthmbMFR0sRR2x48jQ5e3hOLIGhr5GRERE9qjEOXqaLFy4EA4ODoiJiVEKTDcm5uhZP+boGc4Y+1Ash8/QdkP3MXP0CGCOHhERUUkY3NFLSUlR+t3V1RWLFy/G4sWLDV21RszRs37M0TOcMfahWA6foe2G7mPm6BHAHD0iIqKSMPoZvdLAHD3rxxw9wxljH5q63dB9zBw9ApijR0REVBJ8ByQiIiIiIrIx7OgRERERERHZGHb0iIiIiIiIbAw7ekRERERERDbGKgdjYY6e9WOOnuGMsQ+joqK0jmaoS06fKfcxc/QIYI4eERFRSVhlR485etaPOXqGK419aO59zBw9ApijR0REVBJW2dFjjp71Y46e4UorA86c+5g5egQwR4+IiKgkrLKjxxw968ccPcOVxj409z5mjh4BzNEjIiIqCb4DEhERERER2Rh29IiIiIiIiGwMO3pEREREREQ2hh09IiIiIiIiG2OVg7EwR8/6MUfPcKWxD829j5mjRwBz9IiIiErCKjt6zNGzfszRMxxz9JijZy+Yo0dERKQ/q+zoMUfP+jFHz3DM0WOOnr1gjh4REZH+rLKjxxw968ccPcMxR485evaCOXpERET64zsgERERERGRjWFHj4iIiIiIyMawo0dERERERGRj2NEjIiIiIiKyMVY5GItYjl50dDTCwsK0jtLGds3tpVUDc/QMYw85ei4uLgYdR8zRsw3M0SMiItKfVXb0xHL0dMnGYrvmdkuogTl64uwhR8/Qv2Xm6NkG5ugRERHpzyo7emI5erpkY7Fdc7sl1MAcPXH2kKNn6N8yc/RsA3P0iIiI9GeVHT2xXCtdsrHYLp79Ze4amKOnnT3k6Bn6t8wcPdvAHD0iIiL98R2QiIiIiIjIxrCjR0REREREZGPY0SMiIiIiIrIx7OgRERERERHZGKscjEUsR08qlSIzM1PrSHls19xuCTUUtpszR2/WrFl2n0Vo7oy5zZs3a22Pjo5mjp4dYI4eERGR/gzq6M2dOxexsbH44IMPsGjRIgBA27ZtsW/fPqX5RowYgSVLlhiyKSViOXoymUxrdhbbtbdbQg26tJs7Q84esggtPWNOLAePOXq2gTl6RERE+itxR+/YsWP4/vvv0ahRI5W2d999FzNmzFD87ubmVtLNqCWWo1f4Lbu2s0Vs19xuCTXo0l4aOXr2nkVYGus3hFgOHnP0bANz9IiIiPRXoo5ebm4u+vfvj2XLlmHWrFkq7W5ubvDz8zO4OE3Ecq10mYft1r8PzZ0hZw9ZhJaeMSeWg1f00kxNHTVLf47EHD0iIqKSKNE74KhRo9CtWzd07NhRbfvatWtRuXJlhIaGIjY2Fk+ePNG4LqlUipycHKUHERGRMRV/n5FKpeYuiYiIyKT07uglJSXhxIkTiIuLU9ver18/rFmzBnv37kVsbCxWr16NAQMGaFxfXFwcPD09FY/AwEB9SyIiItIqMDBQ6b1G03sYERGRrdDr0s0bN27ggw8+QHJyMlxdXdXOM3z4cMXPDRs2hL+/Pzp06IArV64gJCREZf7Y2FiMGzdO8XtOTg5OnjypT1lERERa3bhxAx4eHorfXVxczFgNERGR6enV0UtLS8O9e/cQFhammFZQUID9+/fj22+/hVQqhaOjo9IyzZs3ByC/mV5dR8/FxYVvuEREZFIeHh5KHT0iIiJbp1dHr0OHDjh9+rTStCFDhqBu3br45JNPVDp5AHDq1CkAgL+/f8mrLEYsR08sO4vt2tstoQZd2s2dIaet3RjrMEa7tefoiRGrUew4sobnSMzRIyIiKgm9Onrly5dHaGio0jR3d3d4eXkhNDQUV65cQWJiIl577TV4eXnh77//xtixY9GmTRu1MQwlJZajFxUVhZo1a+Ly5cta88HYrr7dEmrQtR0oef5ZQkICPD09teb4lbTdGOswRrul5+glJSUZdBxaQ46etufIv2XT/60zR4+IiOyVQYHpxTk7O2PXrl1YtGgR8vLyEBgYiJiYGEyaNMmYmxHN0dMlG4vtlp8BZ2i7WP4ZswgtI0fPlMehpeTo8W/ZfH/rzNEjIiJ7ZXBHLyUlRfFzYGAg9u3bZ+gqRYnlWumSjcV2y8+AM3XGG7MIzZ+jZ+rj0BJy9Mzdbgk1mPNvnZfPEhGRveI7IBERERERkY1hR4+IiIiIiMjGsKNHRERERERkY9jRIyIiIiIisjFGHXWztIhlwFlCfpk1t1tCDcZoZxah4VmE5szp07UGS8/R49+y+f/WiYiI7JFVdvTEcvTEsrEmT55s0vrIMowfP575ZSLtYtEnSUlJAMyTVajLPGLtpZGjZ8o8RkPbLaGG0mhnjh4REZEqq+zoGZqjR/bBnPlmpbENY7VrI7YPTZlVaIx1iLUbI0fP3FmI5t6Hlt7OHD0iIrJXVtnRM0aOHtk+c2d7WUINhv4dmDur0NTbMEaOnqW3W0IN5mzX5QsNIiIiW8R3QCIiIiIiIhvDjh4REREREZGNYUePiIiIiIjIxrCjR0REREREZGOscjAWQ3P0yD4wv0y8XYw5swqNsY7SyNGz5HZLqMES2omIiOyRVXb0DM3RmzVrltnzyyy53RJqMFY7oDkfzZTtllCDLu1JSUkG7UNTZhUaYx1i7brk6Jkzj9Ea9qGlt7OjR0RE9soqO3qG5uhZSn6ZpbZbQg3W3m4JNVh7uyXUYO3tllCDudslEonKNCIiIntglR09Y+Tosd36M+Asvd0SarD2dkuowdrbLaEGc7azo0dERPaKN7ERERERERHZGHb0iIiIiIiIbAw7ekRERERERDaGHT0iIiIiIiIbY5WDsRiao8d228iAs+R2S6jB2tstoQZrb7eEGkqjHQ5lNLZLJBIUFBSobSMiIrJlVtnRE8vRu379OrKzs+Hp6al2HplMxnYt7ZZQg7W3W0IN1t5uCTVYe7sl1GDq9jvZzzDkpwvo2MsZI16pAhe3fzH7yGxMajEJjb0bIzs7G3Xr1lW7b4iIiGyZVXb0xHL0ZDIZJBKJxnnYrr3dEmqw9nZLqMHa2y2hBmtvt4QaTN2e8TwLWc8KkC19ig83X4ZnuSz88/wptlbZhs4NOsPZ2VntfiEiIrJ1VtnRE8vR02UetnMfch9afrsl1GDt7ZZQgynbHSTynDxBkP+elVsevpiBXw/+i8aVjiDI86nadRIREdk6q+zoERERqSPB/7V353FRVvsDxz/DqiKgLAoosrhA5gYoXFIUl1xBFNrMBZUrv2vo1TAtuyZW3h9dtVySMq+5VUi55a4/RSAvIC5EZiqpoZipiCWLIho8vz/GeWRkhkWRefCe9+s1r+CceZ7nO2cmnPOcc77HWP3fMhfe//Z3Su6eMHBEgiAIgmAY+m8BC4IgCEIDpenwmUvtDByJIAiCIBiG6OgJgiAITx0JdaZNN3tjA0ciCIIgCIYhOnqCIAjCU0PTwburOs81s7nMG9ncwBEJgiAIgmGINXqCIAhCg6cCJMDU/CphPcw4+NsGKPqNFhYtDB2aIAiCIBiE4jp6kiQxfPhwLl68iJWVVaX68vJy8vPzMTMzqzKdeGFhod7niPqq65UQQ0OvV0IMDb1eCTE09HolxPCk603KSrA2KsWxWRP+1seF3u39MTIy4i0phLtldykpVmfdlDRpOQVBEAThv4TiOnpFRUUAuLi4GDgSQRAEoaE4Aeyror6oqAhra+v6CkcQBOHJiVcZOgJBl1eVd0NRcR09JycnLl26hKWlJSqVisLCQpydnbl06ZLOET4lUHqMSo8PlB+j0uMD5ceo9PhA+TEqPT5QXoySJFFUVISTk5OhQxEEQRCEeqW4jp6RkRGtW7euVG5lZaWILw1VUXqMSo8PlB+j0uMD5ceo9PhA+TEqPT5QVoxiJE8QBEH4bySybgqCIAiCIAiCIDxlREdPEARBEARBEAThKaP4jp65uTkxMTGYm5sbOhS9lB6j0uMD5ceo9PhA+TEqPT5QfoxKjw8aRoyCIAiC8N9AJYmc04IgCIKgWGfPnmXp0qUcPHiQixcvUlZWhp2dHY6Ojvj5+dG3b1/CwsLq5Frjx49n3bp1rFmzhvHjx9fJOetap06duHLlCvn5+ahU6uyDLi4uGBkZkZOTo/e43NxcFixYwP79+8nNzUWSJBwdHenduzfR0dF07dq1vl6CbP/+/cTHx5OamsrVq1cpLS3FxsaGTp06MXToUMaMGYO9vX29x1UTmrY39NfITZs28eKLL7JkyRKmTZsGwJo1a5g4cSJr164lPDy80jFlZWVs3bqV48ePy4/ff/8dY2Nj/vzzz2qvWVRURGxsLJs3byY3NxcLCwv8/PyYMWMG/fr1q/PXWEkVWTddp8HF/KoPXzwGvj0OKach6R8Q2PHxQ1qbAhNWQngArP3bo59n/ApYdwjWRML4Po8f1+MKnF+LdhJZNwVBEARBqKktW7bw6quvUlpaiq2tLT179sTe3p4//viDrKws4uLiSEhIqLOOntIVFBRw6tQphgwZInc0Ll++TG5uLq+++qre4zIyMnj++ecpKiqiVatWDBw4EGNjY7Kysli/fj3x8fHEx8fz4osv1svryM/PZ9SoURw4cAAAV1dX+vbti4WFBVevXiUtLY0DBw4wd+5cDhw4gJ+fX73E1RClp6cD8Nxzz8llqamplcoqKioqeuT3Oi8vj4CAAH7++WccHR0JDg7m2rVr7Nmzhz179rB06VKmTp36SOeuSz07QLuWuus6tlJ39ISnn+joCYIgCIICXbt2jfDwcEpLS5kxYwbz58+nUaNGWs85fvw4mzZtqrNrxsbG8tZbb+Ho6Fhn56xLGRkZSJJUqy/1AJGRkRQVFREZGcny5csxNTUFoLy8nJiYGObPn09kZCTBwcGV2riuFRQU0KtXL7Kzs/H09GTlypUEBARoPae0tJR169YRExPDlStXnmg8DV16ejqNGzemW7ducllaWhr29va0b99e5zGmpqaMHj0aLy8vvL29sbGx0Tq+KpGRkfz888/079+f7du306RJEwB2797N8OHDmT59On369KFLly6P+9Iey18Dqx4R83SC23ehjW3dXG9kD/hLe7BuXDfnE+qG6OgJgiAIggLt3LmT4uJinJycWLRokc7n+Pj44OPjU2fXdHR0VGwnDx6M3vj7+8tlmo5exbKKbty4wYkTJwCYP3++3MkD9ZZO8+bN48MPP+TmzZucPn0aLy+vJxU+AFOnTiU7OxtXV1dSU1OxsbGp9Bxzc3MiIyMJCQnh5s2bTzSehuzu3btkZmbi6+srv6+///47Z86cISgoSO9xFhYWfPnll/LvFy5cqNH1Tp06xbZt2zA2Nubzzz+XO3kAQ4cOZfz48Xz++efExsayYcOGR3tR9aSNXd2ez7qJ+iEoi+KTscTFxeHq6kqjRo3w8/PjyJEjhg5JNm/ePFQqldbD09PTYPF89913BAcH4+TkhEql4ttvv9WqlySJuXPn4ujoSOPGjRkwYABnz55VVIzjx4+v1KaDBw+ut/hiY2Pp0aMHlpaWtGjRghEjRpCdna31nDt37hAVFYWtrS1NmzYlLCyMa9euKSa+wMDASm34t789xoT5Wvr000/p0qWLvI+av78/e/bskesN2X41ic/Q7fewDz74AJVKxfTp0+UyQ7dhTWJUWjs2RJr39FHWaLm6uqJSqbhw4QJbt26lV69eWFlZYWlpSWBgILt379Z5nOZv8Nq1a7XKNf/ezZs3j+vXrxMVFYWzszNmZmY4OzszderUeumQpKenY2xsjK+vr1yWlpaGhYWF3jV2tUkMZGdXx99+H/LLL78QHx8PwEcffaSzk1dRy5Yt8fDwkH+v+D7k5uYSERGBs7MzpqamWmsqt2zZwl//+lc6depE8+bNadSoEW5ubkycOLHSvxkapaWlLFy4EB8fHywtLTEzM8PBwYEePXowa9Ysfv/9d71xbt68Wf6MWVhY0LNnT72fsbqUmZlJaWmp1mhuenp6pVHfurJ161YAevbsiYuLS6V6zfThHTt2cO/evTq/fl0KnA+q0ZB8Srt8/Ap1+doUyMmDsZ+Aw2tgHg5tX4c530Cpjpe2NkV93PgVlesOnITgRdByMpiOg+aToH00jPkEvjutP8baXF/jeA6MjoM2f1cfYxMJgz6A3Vn6j7l0AyauBMcoaDReHds/voGSu/qPaSgU3dH7+uuviY6OJiYmhszMTLp27cqgQYPIy8szdGiyZ599litXrsiP//znPwaL5datW3Tt2pW4uDid9QsWLGDZsmWsWLGCjIwMLCwsGDRoEHfu3FFMjACDBw/WatP6vCuWkpJCVFQUhw8fZv/+/dy7d4+BAwdy69Yt+Tmvv/46O3bsYOPGjaSkpPDbb78RGhqqmPgAJk2apNWGCxYsqJf4AFq3bs0HH3zA8ePHOXbsGP369SMkJISffvoJMGz71SQ+MGz7VXT06FE+++yzSlOADN2GNYkRlNOODVWbNm0AOHnyJImJiY90jmXLlhEaGkppaSlBQUF07NiRlJQUhg0bxscff1zr8126dAlvb282b96Mr6+vvO5t+fLlDBw4UOeXW02nf968ebW+3sM3DPbt20dZWRmWlpZy2bFjx7h16xYmJiZyWXJysnyOpk2bylMj58yZoxVjeXk58+bNo6SkhCFDhuDs7FzrGGtj586dlJWV0axZM4YPH/7I5zl79ixeXl7s3r0bPz8/hg8frtVJfemll9iwYQONGzemX79+DBo0CCMjI9asWYOPjw9paWla5ysvL2fYsGHMmjWLc+fOERAQwAsvvEDnzp25fv06CxcuJDc3V2csMTEx8nq3oUOH0r59e9LS0ggKCpI7RhUlJyfL71NtVTxWpVLJo7j/+te/5DLNSN7s2bPlssDAwFpfS5fvv/8egO7du+us15TfunWr3m+k17WsXOj2NhzKhj6e0NsTrtyEf26DV5bX/DzrvoOBH8CuLHCzh7Ae6nNZNYaEdNhyrO6uv3Qv+L4D8Wlg2xSGe8OzrSH5NAxbCO9tqXzMmd+g+xxYkwIq1Md0cITFe6D//8Ld6nPzKJukYL6+vlJUVJT8e1lZmeTk5CTFxsYaMKoHYmJipK5duxo6DJ0AaevWrfLv5eXlkoODg7Rw4UK57ObNm5K5ubm0YcMGA0RYOUZJkqTw8HApJCTEIPHokpeXJwFSSkqKJEnqNjM1NZU2btwoP+f06dMSIKWnpxs8PkmSpD59+kjTpk2r91iq0rx5c2nVqlWKa7+H45Mk5bRfUVGR1L59e2n//v1aMSmpDfXFKEnKaceGrKioSGrVqpUESCqVSgoMDJTef/99adeuXVJeXl6Vx7q4uMjHffnll1p1CQkJkkqlkkxMTKQff/xRqy48PFwCpDVr1miVx8TESIAESOPHj5fu3Lkj1+Xm5spxxsfHV4qlT58+EiDFxMTUrgEkSVqwYIEUEREhRURESM8//7wESD169JDLNOcOCAiQyyIiIqQzZ85onefMmTOSu7u7BEitWrWSQkJCpNDQUMnNzU0yMzOTxo4dKxUUFNQ6vtoaO3asBEj9+vV7pOMrvg9jxozReh8qSkhIkIqLi7XKysvLpbi4OAmQnn32Wam8vFyuS0lJkQDJy8tLKiwsrHS+o0ePSvn5+VplmjiaNWsmHT58WGecHTp0qHSupKQk+djaOnPmjNb7bGtrKxkbG2uVWVpaShYWFlplCxYsqPK8OTk5EiAZGxtX+Txvb28JkJYsWaL3OVZWVhIg7dy5s9avr8a+Qu/DxU7dtmsi9T9H+gqpzzPq5yX9Q7s8PAD5/flHCNKfXzyo+/EDJAtzdV3aPO3j1kSqy8MDtMvd7NXlh+ZWjuHaJ0iZ/6yb6+99E0mlQrKzREqZo113IhaptY36uOSH6nq4q8tf8kMqWfOg/OJSpLYtH8TycDvpfCiQYkf07t69y/HjxxkwYIBcZmRkxIABA+Q5+kpw9uxZnJyccHd3Z/To0XrveBlaTk4OV69e1WpPa2tr/Pz8FNWeoL5j16JFCzw8PJg8eTI3btwwWCwFBQUA8vSa48ePc+/ePa129PT0pE2bNgZpx4fj0/jqq6+ws7OjU6dOzJ49m9u3b9d7bKBOYZ2QkMCtW7fw9/dXXPs9HJ+GEtovKiqKYcOGabUVKOszqC9GDSW0Y0PWtGlTEhMT8fPzQ5IkkpOTeeeddxg2bBgtWrTAy8uLFStWUFZWpvccISEhjB49Wqvs5ZdfJjQ0lD///JNly5bVKqbWrVsTFxenNR1SM3UTkLNIVtSmTRs8PDweaVrkzJkzWbVqFatWrZJH5ebPny+XaUZQFi9eLJetWrVKa7ojgIeHB+np6QwcOJDLly+zbds2tmzZQk5ODu3atSMwMBArK6tax1db169fB6BFixaPdR4bGxuWL1+ud1rqyy+/jIWFhVaZSqXitddew9/fn59++onTpx/MmdNMEw4ICMDS0rLS+bp3746tre6sHe+9916lrKCzZ8/G2tqan3/+mUuXLmnVNWnSBA8Pj0rvUU14eHjI73FcXBy3b9/G19dXLvvwww+5ffs2ffv21fo8zJw5s9bX0qWoqAigUttW1LRpUwAKCwvr5JqPasJK9VTKhx+B82t2vI8bvP8iGFfoKXRyhrG91D8fOFmz81wrVK/d66Xj7W5hDV6udXP9mE0gSbBiIvR+Rruucxv4aIz654/3PShPzYajv4CFOXwyARqZPahrYweL9CfybTAUm4wlPz+fsrIyWrbUzg3bsmVLzpw5Y6CotPn5+bF27Vo8PDy4cuUK7777LgEBAZw8eVLnH0pDunr1KoDO9tTUKcHgwYMJDQ3Fzc2N8+fP8/bbbzNkyBB5XUZ9Ki8vZ/r06fTs2ZNOnToB6nY0MzOjWbNmWs81RDvqig/UawRcXFxwcnLixIkTvPnmm2RnZ7Nli445C0/Ijz/+iL+/P3fu3KFp06Zs3bqVjh07kpWVpYj20xcfKKP9EhISyMzM5OjRo5XqlPIZrCpGUEY7Pg08PDw4fPgwR44cYdeuXWRkZJCZmcn169fJyspi8uTJbN68mV27dmFmZlbpeF17iGnKN2/erDXFsSb69++vlYBC45ln1N+sLl++XKlu/fr1tbqGPvv378fc3FwrQ2ViYiLNmzevNoFKamoqoaGhmJiYEB8fT79+/TAzMyM1NZXo6GgiIiJITU3l888/r5NYn7QBAwZgbW1d5XPOnTvH3r17OXfuHEVFRfINAU2nLjs7W/675+3tjbGxMatXr6ZDhw6EhobWOClPcHBwpTJzc3Pc3d35/vvvuXz5staUWF9f3zr5HpeamkpJSYnWjabk5GTKysro37//Y5+/odO3vYKnU82OD/ICXbNrn7l//GX9Sza1+Lqrp06O+xSmDQYvFzCqwTBTba6fXwRHfoHGZhDsrft8gfc7f2kVZtQm37/XMbgL2Or42h7io+6kFjTge5SK7eg1BEOGDJF/7tKlC35+fri4uPDNN98QERFhwMgarldeeUX+uXPnznTp0oW2bduSnJxc73+4o6KiOHnypEHXXVZFX3yRkZHyz507d8bR0ZH+/ftz/vx52rZtWy+xeXh4kJWVRUFBAZs2bSI8PJyUlJR6uXZN6IuvY8eOBm+/S5cuMW3aNPbv3//E07w/qprEaOh2fNr4+vrKCUgkSeL7779n4cKFJCQkcODAAZYuXapz1MLNzU3n+TTlv/76a63i0KwbfJhmNOxJrfm+desWhw8fplevXjRurM7fnp+fzw8//MDIkSMxquKb482bNxk5ciT5+fmkp6drjT5p1i127tyZ1atXM2bMGPr27ftEXgM8SKzzuLkGXF1d9daVlZUxZcoUPvvssyo3M6844tS2bVsWL17MzJkzmTJlClOmTMHFxQV/f3+CgoJ48cUXdd5IAMN9JjSjxxU7epq1rE/q+4LmJv7D6+IrKi4uBqiXEeKqVLe9QnX0bbtgdX/7hDs1zDXzyQQIWgRf/Ef9sGwEPdpCv47q0Tl92T9rc/2cPPVoXslddQKWqlwvevDzr/c7i256BthVKnC1gx+UOVmvRhQ7ddPOzg5jY+NKmeSuXbuGg4ODgaKqWrNmzejQoQPnzp0zdCiVaNqsIbUngLu7O3Z2dvXeplOmTGHnzp0kJSXRunVrudzBwYG7d+9Wyi5X3+2oLz5dNF9q6rMNzczMaNeuHT4+PsTGxtK1a1eWLl2qmPbTF58u9d1+x48fJy8vD29vb0xMTDAxMSElJYVly5ZhYmJCy5YtDd6G1cWoayqhIT6HTyuVSoW3tzcbNmyQE3o8nMG4pqrqCOhSVYeqrlVMutG0aVPu3btHUlKSXGZvb48kSWzZskXruQ/btWsX169fx93dXefG4xXLdU09rUuarTAyMzOrnHJbHU1nV5elS5eyYsUKWrZsSXx8PBcuXKCkpARJkpAkiVGjRgGV3/upU6dy8eJFVq5cybhx4zA2NiYhIYExY8bQsWNHvfv51ddn4uFM57GxsQD06dNHLtMkGOrSpYtc9ihJgPTRdLD1LdMpLCyUO9BVdcYbAqPa58rR6ZlWkL0Ids2EGUPV0y8PnYE5G6H9DPhSz7302ly//P5HuWkjCA+o+jG67pOxKppiR/TMzMzw8fEhMTGRESNGAOqpaomJiUyZMsWwwelRXFzM+fPnGTt2rKFDqcTNzQ0HBwcSExPlTUELCwvJyMhg8uTJhg2uCr/++is3btyot32dJEli6tSpbN26leTk5Ep3w318fDA1NSUxMZGwsDBAPf0lNzdX7x5O9RmfLllZWQAG3RurvLyc0tJSg7dfdfHpUt/t179/f3788UetsgkTJuDp6cmbb74pp1I3ZBtWF6OuadZK+Bw+jQYOHMj27dvJz8/XWZ+Tk6Nz2wHNvmHV3SgypIrTTtPS0jh79iwvvPCCvD7q4MGDXLp0iVdeeaXKLRQ0X8qrGmHRTIOsaguBuhAUFER0dDQ3b95k+/btjBw5ss6v8c033wDw2Wef6czsWVU2yJYtWzJp0iQmTZoEwJkzZ5g4cSLp6em89dZbrFu3rs7jralu3brJn4nbt2+zceNG2rVrR8+ePQH1e7djxw48PT21OvQ13Qi9Jry9vdmyZQvHjulOFakpt7CwoEOHDnV23YbOxBiGdlM/AApvw0d74N0t8D+rYWR3sHiMCSzO90f/VMDqyJpNDQVo1Vz93wvX9T/nou4/rQ2GYjt6ANHR0YSHh9O9e3d8fX1ZsmQJt27dYsKECYYODYA33niD4OBgXFxc+O2334iJicHY2Fi+W1bfiouLte6W5+TkkJWVhY2NDW3atGH69OnMnz+f9u3b4+bmxjvvvIOTk5PckTZ0jDY2Nrz77ruEhYXh4ODA+fPnmTVrFu3atWPQoEH1El9UVBTx8fFs27YNS0tLec2TtbU1jRs3xtramoiICKKjo7GxscHKyoqpU6fi7+/PX/7yF4PHd/78eeLj4xk6dCi2tracOHGC119/nd69e+tMf/8kzJ49myFDhtCmTRuKioqIj48nOTmZffv2Gbz9qotPCe1naWmpteYS1F8abG1t5XJDt2F1MSqhHZ8GkiRVm4Je04nR12H74osvdP6N16ybq6u0809Cxb38fH19adasGV9//bU8gtS2bVtatWpV7RY8rVq1AtSdloKCgkpr2+7du0dmZiagf6prXWnbti2jRo3iq6++YsaMGfTp06fKvfTy8vL4448/apW4RNNZ1bXP208//STfdKkJzc2bESNG1Oq4J2HEiBHyZ3n37t1s3LiRyZMnEx0dDag/6zt27CA6OlruqD6JGObMmUNqaiq5ubmVpq1q9kgMDg6WN3AXKrNqAvPC1Nsh3LwNP1/Vn5SlJpyaQ5c2cCIX9p540KGsTp/76/b2noDfi8GmqXb99uPq+BoyxU7dBHXWqEWLFjF37ly6detGVlYWe/furZRQxFB+/fVXRo0ahYeHBy+99BK2trYcPnz4kTa3rQvHjh3Dy8tLXpQeHR2Nl5cXc+fOBWDWrFlMnTqVyMhIevToQXFxMXv37q3XdUBVxWhsbMyJEycYPnw4HTp0ICIiAh8fHw4dOlSrDW8fx6effkpBQQGBgYE4OjrKj6+//lp+zuLFiwkKCiIsLIzevXvj4OBQbwkmqovPzMyMAwcOMHDgQDw9PZkxYwZhYWHs2LGjXuID9ReTcePG4eHhQf/+/Tl69Cj79u3j+eefBwzbftXFp4T2qwlDt2F1Gko7Kt0nn3xCeHh4pT3PAHnK4vLl6g2lKq5vrmjr1q0kJCRolW3atInNmzdjYmIiZ8t8ksaNG4enp6cca20VFBSQmZlJ79695U5ebm4uv/zyC/369av2+CFDhmBhYUFJSQmTJk2S11CBOsP366+/Tm5uLqamprzwwguPFGNtfPzxx7Rr146cnBx69eqlcx343bt3Wb16NV5eXlrZMWtCkxgnLi6O8vJyufzKlSuMGzeOP/+svDHYwYMH2b17d6V9ECVJYufOnYDujuOjOHLkCJ6ennh6ej7yOQ4ePAigtZ4yKSkJeHLr80C9d3JISAhlZWVERERQUlIi1+3Zs4e1a9diZGTE7Nmzn1gMDcntUvhoN1zXkYD00Bl1J8rYCFrrv9dRY/PVWzky4TPYkVm5XpIg4xz834kHZQGe4O0KxXcgaq32RuyXbsAb8Y8fl6EpekQPkBcFK9HD/3gaWmBgYJXrLVQqFe+99x7vvfdePUalrboY9+3bp7euPtRkvUqjRo2Ii4urctP3J6W6+JydnQ2e9KS6rHWGbD+oOj4ltJ8uD2dGNHQb6lIxRqW2Y0Nz79491q9fz/r167G3t8fLyws7Oztu3rzJqVOn5OmXY8aM0ZsAbNq0aYwaNYqPPvqI9u3bc/78eTIyMgBYtGhRvYyw5ubmkp2drXd6aXVSUlIoKyt75C/19vb2rFixggkTJrBx40aSk5Pp0aMHpqamHDt2jMuXL2NkZMSyZctwd3d/pBhro3nz5qSmpvLyyy+TnJxMQEAAbm5udOnShSZNmnDt2jWOHDlCcXExVlZWODnVME3ifW+//TZ79+7l3//+N0lJSXh7e1NYWEhKSgru7u6MHDmy0kbmmlF3KysrvL29cXJyoqSkhMzMTC5evIi1tXWdfXe4ffs22dnZj3WOgwcP0rx5c61pyUlJSbi6utb4PXzttdfkkVzN1P2ysjKtmRHDhg3jnXfe0Tpu5cqVnDp1igMHDtC2bVsCAgLIy8sjJSUFSZJYunSpmLlw390/YcZXMDMeOjtDewcwNYYL+XD4/uSuf4SAfR3krQn2hqVjYUY8DP9QnXHUw1GdNfN6oTqhSl4hvBkMAyu8PV9MhsB/qjdv/+4M9OoAt+/CwVPQxRns2kO6/tnOiqf4jp4gCIIg/DeKiIjAzc2NxMREMjIyOHXqFNeuXcPExAQnJydGjRrFuHHjGDx4sN5zTJs2jeeee47Fixezfft2JEkiICCAWbNmERQUVI+v5tFpOnUVp5lqymoyogfqznDnzp1ZsmQJ3333HYmJiUiShKOjI6NHj+bvf/+7nNW0PrRo0YKkpCT27t3Lhg0bSEtLIzExkdLSUmxtbfH392fYsGGMHTu2yqmduvj5+XHs2DHmzJnD0aNH2b59u7zX4Zw5c3SO4gYHB1NQUMChQ4c4e/Yshw8fpnHjxjg7O/PWW28RFRWlmPWcf/zxBz/88APBwcHyCO+FCxe4cOECEydOrPF5Tp06Jd/0qKhima5RxxYtWnDs2DFiY2PZvHkz27Ztw8LCgkGDBvHGG2+IrR0qaNpIva9dymn4/iLsP6nu/Dk1h9Du8Nrz0O/Zurve3werz/fx/0HSKUj8Sb1ez+H+fn3DukHYQ/+bd2wNx96HmM2w5wf49rh6hHHqQJg7EoYurLv4DEEl1TblliAIgiAIiubq6srFixfJyclp8Nn/BEF4SHwdpcQU6taryutSKXqNniAIgiAIgiAIglB7oqMnCIIgCIIgCILwlBEdPUEQBEEQBEEQhKeMSMYiCIIgCE8ZTUZOQRAE4b+XGNETBEEQBEEQBEF4yoiOniAIgiAIgiAIwlNGdPQEQRAEQRAEQRCeMqKjJwiCIAiCIAiC8JQRHT1BEARBEARBEISnjEqSJOVt4y4IgiAIgiAIgiA8MjGiJwiCIAiCIAiC8JQRHT1BEARBEARBEISnjOjoCYIgCIIgCIIgPGX+HzTsIpw8PCb0AAAAAElFTkSuQmCC\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "\tCrashed! We targeted 350 speed and were going 275\n", "New target speeds: mid-straight->[350. 350. 350. 350. 350. 265.]<-end\n", "\tCrashed! We targeted 265 speed and were going 263\n", "New target speeds: mid-straight->[350. 350. 350. 350. 350. 253.]<-end\n", "\tCrashed! We targeted 253 speed and were going 163\n", "New target speeds: mid-straight->[350. 350. 350. 350. 350. 153.]<-end\n", "\tCrashed! We targeted 153 speed and were going 151\n", "New target speeds: mid-straight->[350. 350. 350. 327. 234. 141.]<-end\n", "\tCrashed! We targeted 141 speed and were going 123\n", "New target speeds: mid-straight->[350. 350. 350. 299. 206. 113.]<-end\n", "\tCrashed! We targeted 113 speed and were going 262\n", "New target speeds: mid-straight->[350. 319. 286. 256. 206. 113.]<-end\n", "New target speeds: mid-straight->[350. 331. 297. 266. 218. 126.]<-end\n", "\tCrashed! We targeted 126 speed and were going 262\n", "\tCrashed! We targeted 126 speed and were going 121\n", "New target speeds: mid-straight->[350. 324. 290. 260. 203. 111.]<-end\n", "New target speeds: mid-straight->[350. 324. 290. 260. 204. 111.]<-end\n", "\tCrashed! We targeted 111 speed and were going 250\n", "New target speeds: mid-straight->[350. 324. 290. 260. 204. 111.]<-end\n", "\tCrashed! We targeted 111 speed and were going 240\n" ] } ], "source": [ "from drivers.rookiedriver import RookieDriver\n", "\n", "driver_name = 'Ricorris'\n", "driver = RookieDriver(driver_name)\n", "set_seed(1)\n", "# Turn off DRS and safety car for the moment to focus on new driver performance\n", "race(driver=driver, level=Level.Rookie, track_index=1, plot=True, use_safety_car=False, use_drs=False);\n", "plt.close()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Well it crashed quite a lot but this is the first maze race that the driver has seen. It also hit a lot higher speeds than we have seen before - reaching \"Vmax\" (or maximum speed) several times. Also many of the crashes happen at dead ends, which should be eliminated as the driver sees more tracks.\n", "\n", "We can see how the driver adjusts the targeted speeds at the end of a straight depending on what happened. The target speeds can change for two reasons: either because the driver crashed and has lowered the end-of-straight speed or because the dynamics models have updated and the AI now has a better estimate of how the speeds will change each turn as actions are applied.\n", "\n", "There are two aspects of the training, which we introduced in the previous levels, that we haven't talked about so far in this level. The first is the random action probability. We still need a way of encouraging the driver to explore new actions rather than purely exploiting the existing learning and so we make use of the same random action probability as before. There is an opportunity here though for you to improve the AI if you want to: we have a much better idea of where we need to explore in this algorithm as we can check the data we have already collected and see if the action we are about to take is likely to teach us something new or not. For example, if we see that we already have a data point in the dynamics model at a very similar speed for an action then we won't learn much new by taking that action; it might be better to take an action that doesn't have any data points nearby and hence learn a lot. If all actions have nearby data then we know that exploitation is the best approach.\n", "\n", "The second feature that we haven't spoken about is the learning rate. We don't have the speed quantisation in this algorithm that we had previously, however, the maximum, safe, end-of-straight speed that we learn is strongly affected by dead ends. At the start of training the driver is likely to come across several dead ends as the AI hasn't yet learnt the correct turns to make. This means that the driver learns a much more conservative end-of-straight speed than it needs if a turn was always available as it has to be able to brake to a halt at a dead end. As learning progresses the driver should eliminate dead ends and so it is safe to raise the end-of-straight speed. One way of doing this is to use a learning rate parameter either directly on the end-of-straight speed or perhaps by modelling dead end speeds and cornering speeds separately and blending between them. We leave this up to you!\n", "\n", "### DRS\n", "It is time now to talk about the drag reduction system - DRS. Just like in F1, DRS is available at various points around the track (indicated by cyan triangles in the track plots). When the driver reaches one of these points the `drs_available` flag will be set in `track_state` and the driver can choose to use the `OpenDRS` action (attempting to use this action when DRS is not available will result in a `Continue` action being applied instead). If the driver chooses to open the DRS then the car will proceed for one turn at its current speed and then open the DRS. With the DRS open, the car's top speed is increased and so the driver can reach speeds not otherwise possible. Again, like in F1, the DRS will automatically close as soon as the brakes are applied, but be aware that braking from the higher speed is more risky! \n", "\n", "It won't always be quicker to open the DRS when it is available. This is because opening the DRS fixes your speed for one turn, which is obviously bad if it is at the end of the straight and you need to be braking for the corner, but might also be bad earlier in the straight as it means that you miss a turn's acceleration. The optimal choice depends on your current speed and how much of the straight is left. The AI needs to learn to make this trade-off.\n", "\n", "In a model-based algorithm, we can tackle this problem by simulating the rest of the straight using our learnt dynamics models, both with, and without, DRS open. We can then compare the two simulated outcomes and apply the action that we believe will lead to the faster time. This is very similar to the approach a human would take to choosing the best way to get home from school/work!\n", "\n", "Let's turn on DRS and watch our new AI at work." ] }, { "cell_type": "code", "execution_count": 5, "metadata": { "scrolled": false }, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Opening DRS\n", "Opening DRS\n", "Opening DRS\n", "Opening DRS\n", "New target speeds: mid-straight->[348. 312. 280. 250. 181. 88.]<-end\n", "Opening DRS\n", "New target speeds: mid-straight->[348. 312. 280. 250. 181. 89.]<-end\n", "done\n" ] } ], "source": [ "season = Season(Level.Rookie)\n", "set_seed(0)\n", "print('Running races...')\n", "# Get some learning under our driver's belt\n", "driver, *_ = season.race(driver=driver, track_indices=range(14), plot=False); \n", "driver, *_ = season.race(driver=driver, track_indices=[14], plot=True, use_safety_car=False);\n", "plt.close()\n", "print('done')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "One thing we need to be careful of is ensuring the AI explores DRS enough. There are only a few chances in a race to open DRS and many of these might be in places where it is slower to active it. This means that we might struggle to get data to populate the dynamics models with DRS open. To try and counter this we add in a heuristic that will open DRS if any of the actions have fewer than 3 data points recorded under DRS. This is a very basic step to encourage exploration - you might like to come up with something better!\n", "\n", "An alternative is to model DRS as a modifier on top of the the non-DRS data, which would allow us to use a single pool of data rather than having to learn two separate models for each action. The downside is that this is a more restrictive model and, if it doesn't match reality, could lead to much worse performance. We also leave this up to you to explore." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Safety Car\n", "\n", "We have kept the same mechanism for dealing with the safety car that was introduced in level 2: we keep a running estimate of the current safety car speed, decreasing it each time we get a penalty. This estimate can be naturally incorporated into the framework discussed above by capping the target speeds at each point on the straight to our current estimate of the safety car speed. Unlike our previous approach our new AI algorithm won't keep accelerating back above the safety car speed because the target speed is also taken into account when accelerating from a slower speed, not just when we are above it. We can see this behaviour in the safety car sim." ] }, { "cell_type": "code", "execution_count": 6, "metadata": { "scrolled": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Running safety car sim...\n", "Safety car deployed for 50 turns at 100 speed\n", "\tDecreasing estimate of safety car speed to 72.0\n", "\tIncreasing estimate of safety car speed to 121.8\n", "\tDecreasing estimate of safety car speed to 58.4\n", "\tIncreasing estimate of safety car speed to 103.8\n" ] }, { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "cd6f6f28e5af4948bb42baf8f9623d01", "version_major": 2, "version_minor": 0 }, "text/plain": [ "Canvas(toolbar=Toolbar(toolitems=[('Home', 'Reset original view', 'home', 'home'), ('Back', 'Back to previous …" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Complete\n" ] } ], "source": [ "# Run the safety car sim\n", "set_seed(0)\n", "driver = RookieDriver(driver_name) # always best to reset driver for sim\n", "car_speed, safety_car_estimate, safety_car_true, safety_car_penalty = safety_car_sim(driver, level=Level.Rookie)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We can see how, once the safety car speed has stabilised, the driver stays below the estimated speed and doesn't keep accelerating above it. This is much better than we saw before.\n", "\n", "If you want to investigate some advanced behaviour then you might want to look at the safety car speeds and durations as you might find they are not totally random and there might be some predictions you can learn to make that will allow your driver to save some more time! Remember that in the online challenge the general form of any relationships you discover will stay the same but the exact parameters will change so you need to develop a learning algorithm rather than hardcoding in some values.\n", "\n", "# End of level and the Online Challenge\n", "This is the end of the Rookie Driver level. We now have a driver that is able to efficiently learn from just a small amount of data to control the car, navigate the mazes, and deal with the safety car and DRS. There are a number of improvments that were suggested above that can be made to our algorithm, which you might like to try. If you think you have improved the driver then why not test it out by entering it into the Rookie Driver challenge on the [McLaren Maze Race website](https://www.mclaren.com/mazerace)?\n", "\n", "In the next level we will introduce tyres, pit stops, and rain. If you master that, your AI will be a pro driver!\n", "\n", "### We hope to see you in Pro Driver, level 4 of the McLaren Maze Race!" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Playground\n", "If you want to play around with the driver AI here, then it might be helpful to use this class." ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "class MyRookieDriver(RookieDriver):\n", " def __init__(self, *args, **kwargs):\n", " super().__init__(*args, **kwargs)\n", " \n", " def make_a_move(self, car_state: CarState, track_state: TrackState):\n", " return super().make_a_move(car_state, track_state)\n", " \n", " def estimate_next_speed(self, action: Action, speed, drs_active: bool):\n", " return super().estimate_next_speed(action, speed, drs_active)\n", " \n", " def _choose_move_from_models(self, speed: float, target_speed: float, drs_active: bool):\n", " return super()._choose_move_from_models(speed, target_speed, drs_active)\n", " \n", " def update_with_action_results(self, previous_car_state: CarState, previous_track_state: TrackState,\n", " action: Action, new_car_state: CarState, new_track_state: TrackState,\n", " result: ActionResult):\n", " super().update_with_action_results(previous_car_state, previous_track_state, \n", " action, new_car_state, new_track_state, result)\n", " \n", " def _update_safety_car(self, new_car_state: CarState, result: ActionResult):\n", " super()._update_safety_car(new_car_state, result)\n", " \n", " def update_target_speeds(self):\n", " super().update_target_speeds()\n", "\n", "my_driver = MyRookieDriver(driver_name)" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Safety car deployed for 12 turns at 100 speed\n", "\tCar speed of 106.9 exceeds safety car, penalty is now 1\n", "\tDecreasing estimate of safety car speed to 63.5\n", "\tCar speed of 106.9 exceeds safety car, penalty is now 2\n", "Safety car speed estimate of 63.5 already below car speed of 106.9\n", "\tCar speed of 135.3 exceeds safety car, penalty is now 3\n", "Safety car speed estimate of 63.5 already below car speed of 106.9\n", "\tCar speed of 135.3 exceeds safety car, penalty is now 4\n", "Safety car speed estimate of 63.5 already below car speed of 135.3\n", "\tCar speed of 121.3 exceeds safety car, penalty is now 5\n", "Safety car speed estimate of 63.5 already below car speed of 135.3\n", "\tCar speed of 108.7 exceeds safety car, penalty is now 6\n", "Safety car speed estimate of 63.5 already below car speed of 121.3\n", "Safety car no longer active\n", "\tCrashed! We targeted 350 speed and were going 292\n", "New target speeds: mid-straight->[350. 350. 350. 350. 314. 282.]<-end\n", "\tCrashed! We targeted 282 speed and were going 291\n", "New target speeds: mid-straight->[350. 350. 350. 350. 314. 281.]<-end\n", "\tCrashed! We targeted 281 speed and were going 271\n", "New target speeds: mid-straight->[350. 350. 350. 325. 291. 261.]<-end\n", "\tCrashed! We targeted 261 speed and were going 256\n", "New target speeds: mid-straight->[350. 350. 346. 310. 278. 246.]<-end\n", "\tCrashed! We targeted 246 speed and were going 239\n", "New target speeds: mid-straight->[350. 350. 336. 301. 270. 229.]<-end\n", "\tCrashed! We targeted 229 speed and were going 238\n", "New target speeds: mid-straight->[350. 350. 335. 300. 269. 228.]<-end\n", "\tCrashed! We targeted 228 speed and were going 169\n", "New target speeds: mid-straight->[350. 338. 303. 272. 234. 159.]<-end\n", "Safety car deployed for 12 turns at 135 speed\n", "Opening DRS\n", "\tIncreasing estimate of safety car speed to 84.0\n", "Opening DRS\n", "Safety car no longer active\n", "Safety car deployed for 12 turns at 173.5 speed\n", "\tIncreasing estimate of safety car speed to 147.7\n", "Safety car no longer active\n", "New target speeds: mid-straight->[350. 341. 306. 274. 236. 159.]<-end\n", "New target speeds: mid-straight->[350. 342. 307. 275. 237. 159.]<-end\n", "New target speeds: mid-straight->[350. 342. 307. 275. 238. 159.]<-end\n", "Opening DRS\n", "\tCrashed! We targeted 159 speed and were going 114\n", "New target speeds: mid-straight->[350. 319. 286. 256. 196. 104.]<-end\n", "New target speeds: mid-straight->[350. 320. 287. 257. 196. 104.]<-end\n" ] } ], "source": [ "race(driver=my_driver, level=Level.Rookie);\n", "plt.close()" ] } ], "metadata": { "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": "3.8.10" }, "pycharm": { "stem_cell": { "cell_type": "raw", "metadata": { "collapsed": false }, "source": [] } } }, "nbformat": 4, "nbformat_minor": 1 }