{ "cells": [ { "cell_type": "markdown", "metadata": { "id": "_pTxMoBs4t2B" }, "source": [ "# Getting started tutorial\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "eFfNqb6McHpA", "outputId": "92d2fb87-ba0a-476f-dc02-2873882910f9" }, "outputs": [], "source": [ "%pip install --upgrade --quiet pip\n", "%pip install --upgrade --quiet circuitree==0.11.1 numpy matplotlib tqdm ipympl ffmpeg moviepy watermark" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "\n", "## Problem statement\n", "\n", "CircuiTree solves the following problem:\n", "\n", "> Given a phenotype that can be simulated, a reward function that measures the phenotype, and a space of possible circuit architectures, find the optimal architecture(s) to achieve that target phenotype by running a reasonable number of simulations.\n", "\n", "In order to solve this problem, CircuiTree uses a search algorithm called Monte Carlo tree search (MCTS), borrowed from artificial intelligence and reinforcement learning, to search over the space of possible architectures, or topologies. MCTS is an algorithm for planning and game-playing, so we approach circuit design as a game of stepwise assembly, where each step adds an interaction to the circuit diagram.\n", "\n", "The main class provided by this package is `CircuiTree`, and to run a tree search, the user should make their own subclass of `CircuiTree` that defines (1) a space of possible topologies to search and (2) a reward function that returns a (possibly stochastic) estimate of phenotypic quality." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Creating a `CircuiTree` class " ] }, { "cell_type": "code", "execution_count": 2, "metadata": { "id": "MfzS9AGZMatj" }, "outputs": [], "source": [ "from circuitree import CircuiTree" ] }, { "cell_type": "markdown", "metadata": { "id": "sCOpFJBdMY7D" }, "source": [ "\n", "\n", "Let's consider a simple example. Say we are interested in constructing a circuit of three transcription factors (TFs) A, B, and C that exhibits bistability, where the system can be \"switched\" from one state (e.g. high A, low B) to another (high B, low A). We will allow each TF to activate or inhibit any of the TFs (including itself). Multiple regulation (A both activates and inhibits B) is not allowed. With these rules, we have defined a set of topologies (a design space) that we are sampling from." ] }, { "cell_type": "code", "execution_count": 3, "metadata": { "id": "r_2S8bpkM9wq" }, "outputs": [], "source": [ "components = [\"A\", \"B\", \"C\"] # Three transcription factors (TFs)\n", "interactions = [\n", " \"activates\", # Each pairwise interaction has two options\n", " \"inhibits\",\n", "]" ] }, { "cell_type": "markdown", "metadata": { "id": "ZTbXv0FbMZ8X" }, "source": [ "\n", "CircuiTree explores the design space by treating circuit design as a game where the topology is built step-by-step, and the objective is to assemble the best circuit. Specifically, `CircuiTree` represents each circuit topology as a string called a `state`, and it can choose from a list of `actions` that either change the `state` or terminate the assembly process (i.e. \"click submit\" on the game). The algorithm searches starting from a \"root\" state, and over many iterations it builds a decision tree of candidate topologies and preferentially explores regions of that tree with higher mean reward.\n", "\n", "### 1. Choose a Grammar\n", "\n", "The rules for how states are defined and how they are affected by taking actions (i.e. the rules of the game) are called a \"grammar.\" We will be using the built-in `SimpleNetworkGrammar` class to explore the design space we defined above. (See the grammar tutorial for more details on grammars and how to define custom design spaces from the base `CircuitGrammar` class.)" ] }, { "cell_type": "code", "execution_count": 4, "metadata": { "id": "Of8oXRrDJcCF" }, "outputs": [], "source": [ "# Built-in grammars can be found in the `models` module\n", "from circuitree.models import SimpleNetworkGrammar\n", "\n", "grammar = SimpleNetworkGrammar(\n", " components=components,\n", " interactions=interactions,\n", ")" ] }, { "cell_type": "markdown", "metadata": { "id": "46wtzexDJoF-" }, "source": [ "\n", "### 2. Define a reward function\n", "\n", "The only strict requirement for the reward function is that it should return bounded values, ideally between 0 and 1. __NOTE:__ If reward values have a larger range, you may need to increase the `exploration_constant` argument proportionally. \n", "\n", "For our test case, bistability is known to require positive feedback. For example, positive autoregulation (A activates itself) or mutual inhibition (A inhibits B and B inhibits A). Here we will use a dummy reward function that doesn't actually compute bistability but instead just looks for the presence of positive feedback loops. The reward value will be a random number drawn from a Gaussian distribution, and we will increase the mean of that distribution for every type of positive feedback loop the topology contains. In a real scenario, the reward function might be more complex, possibly requiring multiple simulations. To mimic the computational cost of a costly evaluation, we'll introduce an optional argument `expensive` that pauses for `0.1` seconds before returning the result." ] }, { "cell_type": "code", "execution_count": 5, "metadata": { "id": "-IvOHwxwt5re" }, "outputs": [], "source": [ "from time import sleep\n", "import numpy as np\n", "\n", "def get_bistability_reward(state, grammar, rg=None, expensive=False):\n", " \"\"\"Returns a reward value for the given state (topology) based on\n", " whether it contains positive-feedback loops (PFLs). Assumes the \n", " state is a string in the format of SimpleNetworkGrammar.\"\"\"\n", "\n", " # We list all types of PFLs with up to 3 components. Each three-letter \n", " # substring is an interaction in the circuit, and interactions are \n", " # separated by underscores.\n", " patterns = [\n", " \"AAa\", # PAR - \"AAa\" means \"A activates A\"\n", " \"ABi_BAi\", # Mutual inhibition - \"A inhibits B, B inhibits A\"\n", " \"ABa_BAa\", # Mutual activation\n", " \"ABa_BCa_CAa\", # Cycle of all activation\n", " \"ABa_BCi_CAi\", # Cycle with two inhibitions\n", " ]\n", "\n", " # Mean reward increases with each PFL found (from 0.25 to 0.75)\n", " mean = 0.25\n", " for pattern in patterns:\n", "\n", " # The \"has_pattern\" method returns whether state contains the pattern.\n", " # It checks all possible renamings. For example, `has_pattern(s, 'AAa')`\n", " # checks whether the state `s` contains 'AAa', 'BBa', or 'CCa'.\n", " if grammar.has_pattern(state, pattern):\n", " mean += 0.1\n", "\n", " if expensive: # Simulate a more expensive reward calculation\n", " sleep(0.1)\n", "\n", " # Use the default random number generator if none is provided\n", " rg = np.random.default_rng() if rg is None else rg\n", " \n", " return rg.normal(loc=mean, scale=0.1)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "\n", "### 3. Create a subclass\n", "\n", "Our subclass of `CircuiTree` must define the `get_reward` method. The first argument of the method should be a `state`, or unique identifier corresponding to a topology. For many features, the method `is_success` should also be defined. It should take the name of a terminal topology and return `True` if it is considered \"successful\" overall at generating the phenotype. \n", "\n", "We will say that a successfully bistable circuit should have a mean reward of >0.5, which we will calculate empirically as the cumulative reward divided by the number of samples, or \"visits\" to that state." ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [], "source": [ "class BistabilityTree(CircuiTree):\n", " \"\"\"A subclass of CircuiTree that searches for positive feedback networks.\n", " Uses the SimpleNetworkGrammar to encode network topologies. The grammar can \n", " be accessed with the `self.grammar` attribute.\"\"\"\n", "\n", " def __init__(self, *args, **kwargs):\n", " kwargs = kwargs | {\"grammar\": grammar}\n", " super().__init__(*args, **kwargs)\n", "\n", " def get_reward(self, state: str, expensive: bool = False) -> float:\n", " \"\"\"Returns a reward value for the given state (topology) based on\n", " whether it contains positive-feedback loops (PFLs).\"\"\"\n", "\n", " # `self.rg` is a Numpy random generator that can be seeded on initialization\n", " reward = get_bistability_reward(\n", " state, self.grammar, self.rg, expensive=expensive\n", " )\n", " return reward\n", "\n", " def get_mean_reward(self, state: str) -> float:\n", " \"\"\"Returns the mean empirical reward value for the given state.\"\"\"\n", " # The search graph is stored as a `networkx.DiGraph` in the `graph`\n", " # attribute. We can access the cumulative reward and # of visits for \n", " # each node (state) using the `reward` and `visits` attributes.\n", " return (\n", " self.graph.nodes[state].get(\"reward\", 0) \n", " / self.graph.nodes[state].get(\"visits\", 1)\n", " )\n", "\n", " def is_success(self, state: str) -> bool:\n", " \"\"\"Returns whether a topology is a successful bistable circuit design.\"\"\"\n", " if self.grammar.is_terminal(state):\n", " return self.get_mean_reward(state) > 0.5\n", " else:\n", " return False # Ignore incomplete topologies" ] }, { "cell_type": "markdown", "metadata": { "id": "QsHB9WslSsjq" }, "source": [ "\n", "## Running a tree search\n", "\n", "We can run a search using the `CircuiTree.search_mcts()` method (or `CircuiTree.search_mcts_parallel()` for a parallel search). We need to supply a \"root\" `state` string that is the initial state of the assembly game, in this case a circuit with three TFs (A, B, and C) and no interactions. Using the SimpleNetwork format, this is represented by the string `ABC::`. We can specify any additional keyword arguments for the reward functions using the `run_kwargs` argument." ] }, { "cell_type": "code", "execution_count": 7, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "lqLbva-xTMHj", "outputId": "153dab8d-38ec-42cb-ce26-d3454c8c2f99" }, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "MCTS search: 1%| | 422/50000 [00:00<00:29, 1688.86it/s]" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Starting MCTS search with 50000 iterations.\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "MCTS search: 100%|██████████| 50000/50000 [01:07<00:00, 745.51it/s]\n" ] } ], "source": [ "# Make an instance of the search tree\n", "tree = BistabilityTree(\n", " grammar=grammar,\n", " root=\"ABC::\", # The root state - 3 TFs, no interactions\n", " seed=0, # Seed for the random number generator\n", ")\n", "\n", "# Run the search\n", "tree.search_mcts(\n", " n_steps=50_000, \n", " progress_bar=True, \n", " run_kwargs={\"expensive\": False}\n", ")" ] }, { "cell_type": "markdown", "metadata": { "id": "0GW2Aq1JYQ2n" }, "source": [ "\n", "## Visualizing results\n", "\n", "### The best individual topologies\n", "\n", "To get an initial feel for the results, let's plot the 10 designs with the highest average reward after filtering out the states with 10 or fewer samples." ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import matplotlib.pyplot as plt\n", "from circuitree.viz import plot_network\n", "\n", "%matplotlib inline\n", "\n", "# Top 10 designs with at least 10 visits \n", "def robustness(state):\n", " r = tree.graph.nodes[state].get(\"reward\", 0) \n", " v = tree.graph.nodes[state].get(\"visits\", 1)\n", " return r / v\n", "\n", "# Recall that only the \"terminal\" states are fully assembled circuits\n", "states = [s for s in tree.terminal_states if tree.graph.nodes[s][\"visits\"] > 10]\n", "top_10_states = sorted(states, key=robustness, reverse=True)[:10]\n", "\n", "# Plot the top 10 \n", "fig = plt.figure(figsize=(12, 5))\n", "plt.suptitle(\"Top 10 bistable circuits and their robustness\")\n", "for i, state in enumerate(top_10_states):\n", " ax = fig.add_subplot(2, 5, i + 1)\n", " \n", " # The `viz.plot_network()` function plots SimpleNetwork-formatted strings\n", " plot_network(\n", " *grammar.parse_genotype(state), \n", " ax=ax, \n", " plot_labels=False, \n", " node_shrink=0.6, \n", " auto_shrink=0.8,\n", " offset=0.75,\n", " padding=0.4\n", " )\n", " r = tree.graph.nodes[state][\"reward\"]\n", " v = tree.graph.nodes[state][\"visits\"]\n", " ax.set_title(f\"{r / v:.2f} (n={v})\")\n", " ax.set_xlim(-1.5, 1.5)\n", " ax.set_ylim(-1.0, 1.8)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Recall that our reward function is counting the number of different positive feedback loops. By that standard, our best solutions are great! Most contain 3 or 4 different PFLs. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Overall sampling of the search graph\n", "\n", "To visualize where the search allocated its samples over the whole search space, we can view the whole search graph at once using a *complexity layout*." ] }, { "cell_type": "code", "execution_count": 9, "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 339 }, "id": "JAdrEyttZ8JC", "outputId": "a9931944-efd0-4113-aec3-61c9fb92b3f1" }, "outputs": [ { "data": { "text/plain": [ "(
,\n", " )" ] }, "execution_count": 9, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAABCkAAAHHCAYAAACSgAIyAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8fJSN1AAAACXBIWXMAAA9hAAAPYQGoP6dpAAD9h0lEQVR4nOzdd3xUVfo/8M+d3mcy6SGQQGjSFRUUAVEQUYooYtm1oH7XVbHrqutakHUt2LuLLqg0AVGxIDYUBLHTZEFaqCG9T597f3/wO2dnSIAQAjPA5/165aW5c2fmzCRkzn3Oc55H0TRNAxERERERERFRgukSPQAiIiIiIiIiIoBBCiIiIiIiIiJKEgxSEBEREREREVFSYJCCiIiIiIiIiJICgxRERERERERElBQYpCAiIiIiIiKipMAgBRERERERERElBQYpiIiIiIiIiCgpMEhBREREREREREmBQQoiIjoqKIqC8ePHJ3oYSae4uBhjxoxBamoqFEXBc889l5BxXH311XA4HAl57n0pLCyEoiiYOnXqEX3eb775BoqiYO7cuQc89+qrr0Z+fn7cMUVR8PDDD8vvp06dCkVRUFhY2LIDJSIiSkIMUhARHWdWr16NMWPGIC8vDxaLBa1atcKQIUPw4osvJnpo1Ay33347Fi5ciPvuuw/vvPMOzj333MP2XD6fDw8//DC++eabw/YcB/Lwww9DURT5pdPpkJ2djeHDh2P58uUt8hyvvPLKEQ9sHKyjYYxERETNYUj0AIiI6MhZtmwZBg0ahDZt2uD//u//kJWVhe3bt2P58uV4/vnncfPNNyd6iHSQvv76a4waNQp33XXXYX8un8+HCRMmAADOPPPMw/58+/Pqq6/C4XBAVVVs374dkydPxoABA/Djjz+iV69eAIC8vDz4/X4YjcaDeuxXXnkFaWlpuPrqq1t+4HuZPHkyVFXd7zlXXHEFLr30UpjNZnnsSI6RiIjoSGKQgojoOPLoo4/C7Xbjp59+gsfjibutpKTkiI+nvr4edrv9iD/vsaSkpKTBz/JQBAIBmEwm6HTJnWw5ZswYpKWlye8vuOACdOvWDXPmzJFBCkVRYLFYEjTCpmlKAEWv10Ov1x+B0RARESVecs9AiIioRW3atAldu3Zt9KI2IyOjwbFp06ahd+/esFqt8Hq9uPTSS7F9+/a4c5YsWYKLL74Ybdq0gdlsRuvWrXH77bfD7/fHnSdqFmzatAnnnXcenE4n/vSnPwEAVFXF888/j+7du8NisSA9PR3nnnsufv755wZj+uCDD9CtWzeYzWZ07doVn332WZNe+4svvoiuXbvCZrMhJSUFJ598MmbMmCFvF9sI1q1bh7Fjx8LlciE1NRW33norAoFA3GNNmTIFZ511FjIyMmA2m9GlSxe8+uqrjT7vggULMHDgQDidTrhcLpxyyilxzwsAP/zwA84991y43W7YbDYMHDgQS5cu3e/rEXUKNE3Dyy+/LLc/CJs3b8bFF18Mr9cLm82Gvn374pNPPol7DFE7YdasWfjHP/6BVq1awWazoaampsHzFRYWIj09HQAwYcIE+XyxtRMAYOfOnbjgggvgcDiQnp6Ou+66C9FoNO4cVVXx3HPPoWvXrrBYLMjMzMT111+PysrK/b7m/cnKygIAGAz/W39prCbF7t27MW7cOOTm5sJsNiM7OxujRo2S9R7y8/Px+++/49tvv5WvUWSNVFRU4K677kL37t3hcDjgcrkwbNgwrFy5stExRaNR/P3vf0dWVhbsdjtGjhzZ4N9PYzUp9rZ3TYp9jXHz5s1QFAXPPvtsg8dYtmwZFEXBzJkz9/tcW7duxciRI2G325GRkSG3EymKErfN52D/3W/btg3Dhw+Hw+FAq1at8PLLLwPYs/3srLPOgt1uR15eXoN/GwBQVVWF2267Da1bt4bZbEb79u3xxBNPHDADhYiIjk7MpCAiOo7k5eXh+++/x5o1a9CtW7f9nvvoo4/igQcewNixY3HdddehtLQUL774IgYMGIDffvtNBjrmzJkDn8+HG264Aampqfjxxx/x4osvYseOHZgzZ07cY0YiEQwdOhRnnHEGnnrqKdhsNgDAtddei6lTp2LYsGG47rrrEIlEsGTJEixfvhwnn3yyvP93332HefPm4cYbb4TT6cQLL7yAiy66CNu2bUNqauo+X8vkyZNxyy23YMyYMTLosGrVKvzwww+4/PLL484dO3Ys8vPz8dhjj2H58uV44YUXUFlZibffflue8+qrr6Jr164YOXIkDAYDPvroI9x4441QVRU33XSTPG/q1Km45ppr0LVrV9x3333weDz47bff8Nlnn8nn/frrrzFs2DD07t0bDz30EHQ6nQyCLFmyBKeeemqjr2nAgAF45513cMUVV2DIkCG48sor5W3FxcU4/fTT4fP5cMsttyA1NRVvvfUWRo4ciblz52L06NFxjzVx4kSYTCbcddddCAaDMJlMDZ4vPT0dr776Km644QaMHj0aF154IQCgR48e8pxoNIqhQ4eiT58+eOqpp/Dll1/i6aefRkFBAW644QZ53vXXX4+pU6di3LhxuOWWW7Blyxa89NJL+O2337B06dImZRdUVFQA2BPw2LlzJyZOnAiLxYKxY8fu934XXXQRfv/9d9x8883Iz89HSUkJvvjiC2zbtg35+fl47rnncPPNN8PhcOD+++8HAGRmZgLYE/j54IMPcPHFF6Nt27YoLi7G66+/joEDB2Lt2rXIycmJe65HH30UiqLgnnvuQUlJCZ577jkMHjwYK1asgNVqPeBr3Jd9jbFdu3bo168fpk+fjttvvz3uPtOnT4fT6cSoUaP2+bj19fU466yzUFRUhFtvvRVZWVmYMWMGFi1a1ODcg/l3H41GMWzYMAwYMABPPvkkpk+fjvHjx8Nut+P+++/Hn/70J1x44YV47bXXcOWVV+K0005D27ZtAezZYjRw4EDs3LkT119/Pdq0aYNly5bhvvvuQ1FRUcIKxRIR0WGkERHRcePzzz/X9Hq9ptfrtdNOO03729/+pi1cuFALhUJx5xUWFmp6vV579NFH446vXr1aMxgMccd9Pl+D53nsscc0RVG0rVu3ymNXXXWVBkC799574879+uuvNQDaLbfc0uBxVFWV/w9AM5lM2saNG+WxlStXagC0F198cb+ve9SoUVrXrl33e85DDz2kAdBGjhwZd/zGG2/UAGgrV66Uxxp7zUOHDtXatWsnv6+qqtKcTqfWp08fze/3N/q6VFXVOnTooA0dOjTutfp8Pq1t27bakCFD9jtmTdvzvtx0001xx2677TYNgLZkyRJ5rLa2Vmvbtq2Wn5+vRaNRTdM0bdGiRRoArV27do2+pr2VlpZqALSHHnqowW3i5/vII4/EHT/xxBO13r17y++XLFmiAdCmT58ed95nn33W6PG9iZ/T3l8ej0f77LPP4s7dsmWLBkCbMmWKpmmaVllZqQHQJk2atN/n6Nq1qzZw4MAGxwOBgHzvYp/DbDbHvW7xvrZq1UqrqamRx2fPnq0B0J5//nl57KqrrtLy8vLiHnPv93jKlCkaAG3Lli0HHOPrr7+uAdD++9//ymOhUEhLS0vTrrrqqv2+7qeffloDoH3wwQfymN/v1zp37qwB0BYtWiSPH+y/+3/961/yWGVlpWa1WjVFUbRZs2bJ4+vWrWvw2idOnKjZ7Xbtjz/+iHuue++9V9Pr9dq2bdv2+5qIiOjow+0eRETHkSFDhuD777/HyJEjsXLlSjz55JMYOnQoWrVqhfnz58vz5s2bB1VVMXbsWJSVlcmvrKwsdOjQIW5lNXZFuL6+HmVlZTj99NOhaRp+++23BmOIXVEHgPfeew+KouChhx5qcG7s9gUAGDx4MAoKCuT3PXr0gMvlwubNm/f7uj0eD3bs2IGffvppv+cBiMuEACCLiX766afyWOxrrq6uRllZGQYOHIjNmzejuroaAPDFF1+gtrYW9957b4O6COJ1rVixAhs2bMDll1+O8vJy+T7X19fj7LPPxuLFi5uV0v7pp5/i1FNPxRlnnCGPORwO/OUvf0FhYSHWrl0bd/5VV111SCv7sf7617/Gfd+/f/+4n8+cOXPgdrsxZMiQuN+t3r17w+FwNLpq35j33nsPX3zxBT7//HNMmTIFHTt2xEUXXYRly5bt8z5WqxUmkwnffPNNs7aWmM1mWasjGo2ivLwcDocDnTp1wq+//trg/CuvvBJOp1N+P2bMGGRnZ8f9LrW0sWPHwmKxYPr06fLYwoULUVZWhj//+c/7ve9nn32GVq1aYeTIkfKYxWLB//3f/zU492D/3V933XXy/z0eDzp16gS73R6X+dKpUyd4PJ4Gvy/9+/dHSkpK3O/L4MGDEY1GsXjx4gO8I0REdLThdg8iouPMKaecgnnz5iEUCmHlypV4//338eyzz2LMmDFYsWIFunTpgg0bNkDTNHTo0KHRx4hNx9+2bRsefPBBzJ8/v8GFn7hgFwwGA3Jzc+OObdq0CTk5OfB6vQcce5s2bRocS0lJOeAF5z333IMvv/wSp556Ktq3b49zzjkHl19+Ofr169fg3L1fc0FBAXQ6nawHAABLly7FQw89hO+//x4+ny/u/OrqarjdbmzatAkA9rutZsOGDQD2BAn2pbq6GikpKft9fXvbunUr+vTp0+D4CSecIG+PHZdIrT9Uop5IrL1/Phs2bEB1dXWjNVCAphdwHTBgQFzhzDFjxqBDhw64+eab8csvvzR6H7PZjCeeeAJ33nknMjMz0bdvXwwfPhxXXnmlrGmxP6J2yiuvvIItW7bE1dpobLvR3r9LiqKgffv2cb9LLc3j8WDEiBGYMWMGJk6cCGDPVo9WrVrhrLPO2u99t27dioKCggbBwfbt2zc492D+3Tf2e+F2u5Gbm9vgudxud4Pfl1WrVjW4v5CIgr9ERHR4MUhBRHScMplMOOWUU3DKKaegY8eOGDduHObMmYOHHnoIqqpCURQsWLCg0a4CDocDwJ7V5CFDhqCiogL33HMPOnfuDLvdjp07d+Lqq69ukAUQuxLdHPvqcKBp2n7vd8IJJ2D9+vX4+OOP8dlnn+G9997DK6+8ggcffFC21NyXvS+iNm3ahLPPPhudO3fGM888g9atW8NkMuHTTz/Fs88+e1CZD+LcSZMmyY4UexPv9eHUUlkUTelAoaoqMjIy4lb6Y+3rYvRAHA4H+vTpgw8//HC/XWNuu+02jBgxAh988AEWLlyIBx54AI899hi+/vprnHjiift9jn/961944IEHcM0112DixInwer3Q6XS47bbbkqqI45VXXok5c+Zg2bJl6N69O+bPn48bb7yxxTq2HOy/+339XjTl37OqqhgyZAj+9re/NXpux44dm/kqiIgoWTFIQUREsjhlUVERgD3ZA5qmoW3btvu9CFi9ejX++OMPvPXWW3GFG7/44osmP3dBQQEWLlyIioqKJmVTNJfdbscll1yCSy65BKFQCBdeeCEeffRR3HfffXHbMTZs2BCXWbBx40aoqio7MHz00UcIBoOYP39+XGbH3tsUxLaUNWvWNLoSHXuOy+XC4MGDW+R1AnsKpK5fv77B8XXr1snbm2PvgE1zFBQU4Msvv0S/fv1aLDgiRCIRAEBdXd1+W9sWFBTgzjvvxJ133okNGzagV69eePrppzFt2jQA+36dc+fOxaBBg/Dmm2/GHa+qqorL6hBEpoygaRo2btwYV2y0ufb3szj33HORnp6O6dOno0+fPvD5fLjiiisO+Jh5eXlYu3YtNE2Le/yNGzfGndcS/+6bqqCgAHV1dS3674OIiJIba1IQER1HFi1a1GjWgdgj36lTJwDAhRdeCL1ejwkTJjQ4X9M0lJeXA/jfSmjsOZqm4fnnn2/ymC666CJomtZoRsOBMiSaSoxXMJlM6NKlCzRNQzgcjrtNtEYUXnzxRQDAsGHDADT+mqurqzFlypS4+51zzjlwOp147LHHGrQwFfft3bs3CgoK8NRTT6Gurq7BuEtLS5v8GmOdd955+PHHH/H999/LY/X19fj3v/+N/Px8dOnSpVmPK7qxVFVVNev+wJ6aCdFoVG5FiBWJRJr92BUVFVi2bBmysrL2uZXE5/M1+FkUFBTA6XQiGAzKY3a7vdFx6PX6Br+Tc+bMwc6dOxt9vrfffhu1tbXy+7lz56KoqEj+Lh2KfY0R2LOt6rLLLsPs2bMxdepUdO/evUmBkaFDh2Lnzp1x9WkCgQAmT54cd15L/LtvqrFjx+L777/HwoULG9xWVVUlA1NERHTsYCYFEdFx5Oabb4bP58Po0aPRuXNnhEIhLFu2DO+++y7y8/Mxbtw4AHsu3P75z3/ivvvuQ2FhIS644AI4nU5s2bIF77//Pv7yl7/grrvuQufOnVFQUIC77roLO3fuhMvlwnvvvXdQRQkHDRqEK664Ai+88AI2bNiAc889F6qqYsmSJRg0aBDGjx9/yK/7nHPOQVZWFvr164fMzEz897//xUsvvYTzzz8/rrAhAGzZsgUjR47Eueeei++//x7Tpk3D5Zdfjp49e8rHMplMGDFiBK6//nrU1dVh8uTJyMjIkJkowJ7siGeffRbXXXcdTjnlFFx++eVISUnBypUr4fP58NZbb0Gn0+GNN97AsGHD0LVrV4wbNw6tWrXCzp07sWjRIrhcLnz00UcH/XrvvfdezJw5E8OGDcMtt9wCr9eLt956C1u2bMF7773X7LR/q9WKLl264N1330XHjh3h9XrRrVu3A7azjTVw4EBcf/31eOyxx7BixQqcc845MBqN2LBhA+bMmYPnn38eY8aMOeDjzJ07Fw6HA5qmYdeuXXjzzTdRWVmJ1157bZ9ZBn/88QfOPvtsjB07Fl26dIHBYMD777+P4uJiXHrppfK83r1749VXX8U///lPtG/fHhkZGTjrrLMwfPhwPPLIIxg3bhxOP/10rF69GtOnT0e7du0afT6v14szzjgD48aNQ3FxMZ577jm0b9++0UKUB2tfYxSuvPJKvPDCC1i0aBGeeOKJJj3m9ddfj5deegmXXXYZbr31VmRnZ2P69Oky00i8ry3x776p7r77bsyfPx/Dhw/H1Vdfjd69e6O+vh6rV6/G3LlzUVhY2GgWCxERHcWOZCsRIiJKrAULFmjXXHON1rlzZ83hcGgmk0lr3769dvPNN2vFxcUNzn/vvfe0M844Q7Pb7Zrdbtc6d+6s3XTTTdr69evlOWvXrtUGDx6sORwOLS0tTfu///s/2RpUtH7UtD2tCO12e6PjikQi2qRJk7TOnTtrJpNJS09P14YNG6b98ssv8hw00mpT0zQtLy/vgK0VX3/9dW3AgAFaamqqZjabtYKCAu3uu+/Wqqur5TmiteXatWu1MWPGaE6nU0tJSdHGjx/foIXo/PnztR49emgWi0XLz8/XnnjiCe0///lPgzaR4tzTTz9ds1qtmsvl0k499VRt5syZcef89ttv2oUXXijHl5eXp40dO1b76quv9vu69ve+bNq0SRszZozm8Xg0i8WinXrqqdrHH38cd45olTlnzpwDPo+wbNkyrXfv3prJZIprF7mvn694X/f273//W+vdu7dmtVo1p9Opde/eXfvb3/6m7dq1a7/P31gLUrvdrp122mna7Nmz487duwVpWVmZdtNNN2mdO3fW7Ha75na7tT59+jS43+7du7Xzzz9fczqdGgDZ6jMQCGh33nmnlp2drVmtVq1fv37a999/rw0cODCuHah4X2fOnKndd999WkZGhma1WrXzzz8/rj2neN+a04J0X2OM1bVrV02n02k7duzY73saa/Pmzdr555+vWa1WLT09Xbvzzju19957TwOgLV++XJ53qP/uBw4c2Ghb4Ly8PO3888+PO1ZbW6vdd999Wvv27TWTyaSlpaVpp59+uvbUU081aJ9MRERHP0XTWiiXloiI6Cj28MMPY8KECSgtLeXKLB0TTjzxRHi9Xnz11VeH9DjPPfccbr/9duzYsQOtWrVqodERERE1jjUpiIiIiI4xP//8M1asWBFX2LIp/H5/3PeBQACvv/46OnTowAAFEREdEaxJQURERHSMWLNmDX755Rc8/fTTyM7OxiWXXHJQ97/wwgvRpk0b9OrVC9XV1Zg2bRrWrVu3z5axRERELY1BCiIiIqJjxNy5c/HII4+gU6dOmDlzZlx73aYYOnQo3njjDUyfPh3RaBRdunTBrFmzDjrYQURE1FysSUFERERERERESYE1KYiIiIiIiIgoKTBIQURERERERERJgUEKIiIiIiIiIkoKDFIQERERERERUVJgkIKIiIiIiIiIkgKDFERERERERESUFBikICIiIiIiIqKkwCAFERERERERESUFBimIiIiIiIiIKCkwSEF0HFAUBQ8//LD8furUqVAUBYWFhQkbExEREVFjCgsLoSgKpk6d2uRzn3rqqcM/sCR2MO9ZrDPPPBNnnnnmYRkTUXMxSEF0lHvllVegKAr69OmT6KEQERFRE4jFAkVR8N133zW4XdM0tG7dGoqiYPjw4QkYYfL59NNP4xZcWtKjjz6KkSNHIjMzs8HCzt527tyJsWPHwuPxwOVyYdSoUdi8efNhGVci7Nq1Cw8//DBWrFiR6KHQccyQ6AEQ0aGZPn068vPz8eOPP2Ljxo1o3759oodERERETWCxWDBjxgycccYZcce//fZb7NixA2azOUEjS6y8vDz4/X4YjUZ57NNPP8XLL798WAIV//jHP5CVlYUTTzwRCxcu3Od5dXV1GDRoEKqrq/H3v/8dRqMRzz77LAYOHIgVK1YgNTW1xcfWVI29Z03x+eefx32/a9cuTJgwAfn5+ejVq1cLjpCo6ZhJQXQU27JlC5YtW4ZnnnkG6enpmD59eqKHRERERE103nnnYc6cOYhEInHHZ8yYgd69eyMrKytBI0ssRVFgsVig1+uPyPNt2bIFRUVFmDZt2n7Pe+WVV7BhwwZ8/PHH+Nvf/obbb78dn3/+OYqKivD000/v974PP/ww8vPzW3DU8Zr7nplMJphMpsM0KqLmYZCC6Cg2ffp0pKSk4Pzzz8eYMWNaPEhx9dVXw+FwYNu2bRg+fDgcDgdatWqFl19+GQCwevVqnHXWWbDb7cjLy8OMGTMaPMbmzZtx8cUXw+v1wmazoW/fvvjkk0/k7cXFxTAYDJgwYUKD+65fvx6KouCll16Sx6qqqnDbbbehdevWMJvNaN++PZ544gmoqhp331mzZqF3795wOp1wuVzo3r07nn/++ZZ6a4iIiA7ZZZddhvLycnzxxRfyWCgUwty5c3H55Zc3ep/6+nrceeed8nOwU6dOeOqpp6BpmjynW7duGDRoUIP7qqqKVq1aYcyYMXHHnnvuOXTt2hUWiwWZmZm4/vrrUVlZGXffn3/+GUOHDkVaWhqsVivatm2La665Zr+v74477kBqamrc2G6++WYoioIXXnhBHisuLoaiKHj11VcBNKyvcPXVV8u5h9gmoyhKg+f797//jYKCApjNZpxyyin46aef9js+oanBg7lz5+KUU07BKaecIo917twZZ599NmbPnt2kxzgY4XAYXq8X48aNa3BbTU0NLBYL7rrrLgCN16TYvXs3xo0bh9zcXJjNZmRnZ2PUqFFxNclia1J888038rWNGzdOvs/iMTds2ICLLroIWVlZsFgsyM3NxaWXXorq6uoWf+10fGOQgugoNn36dFx44YUwmUy47LLLsGHDhiZ/IDdVNBrFsGHD0Lp1azz55JPIz8/H+PHjMXXqVJx77rk4+eST8cQTT8DpdOLKK6/Eli1b5H2Li4tx+umnY+HChbjxxhvx6KOPIhAIYOTIkXj//fcBAJmZmRg4cGCjH+7vvvsu9Ho9Lr74YgCAz+fDwIEDMW3aNFx55ZV44YUX0K9fP9x3332444475P2++OILXHbZZUhJScETTzyBxx9/HGeeeSaWLl3aou8NERHRocjPz8dpp52GmTNnymMLFixAdXU1Lr300gbna5qGkSNH4tlnn8W5556LZ555Bp06dcLdd98d9zl4ySWXYPHixdi9e3fc/b/77jvs2rUr7rGvv/563H333ejXrx+ef/55jBs3DtOnT8fQoUMRDocBACUlJTjnnHNQWFiIe++9Fy+++CL+9Kc/Yfny5ft9ff3790dFRQV+//13eWzJkiXQ6XRYsmRJ3DEAGDBgQKOPc/3112PIkCEAgHfeeUd+xZoxYwYmTZqE66+/Hv/85z9RWFiICy+8UL6GQ6WqKlatWoWTTz65wW2nnnoqNm3ahNra2hZ5LsFoNGL06NH44IMPEAqF4m774IMPEAwGG/09ES666CK8//77GDduHF555RXccsstqK2txbZt2xo9/4QTTsAjjzwCAPjLX/4i3+cBAwYgFAph6NChWL58OW6++Wa8/PLL+Mtf/oLNmzejqqqqxV4zEQBAI6Kj0s8//6wB0L744gtN0zRNVVUtNzdXu/XWWxucC0B76KGH5PdTpkzRAGhbtmzZ73NcddVVGgDtX//6lzxWWVmpWa1WTVEUbdasWfL4unXrGjzPbbfdpgHQlixZIo/V1tZqbdu21fLz87VoNKppmqa9/vrrGgBt9erVcc/fpUsX7ayzzpLfT5w4UbPb7doff/wRd969996r6fV6bdu2bZqmadqtt96quVwuLRKJ7Pf1ERERJYL4HP7pp5+0l156SXM6nZrP59M0TdMuvvhibdCgQZqmaVpeXp52/vnny/t98MEHGgDtn//8Z9zjjRkzRlMURdu4caOmaZq2fv16DYD24osvxp134403ag6HQz7XkiVLNADa9OnT48777LPP4o6///77crwHo6SkRAOgvfLKK5qmaVpVVZWm0+m0iy++WMvMzJTn3XLLLZrX69VUVdU0TdO2bNmiAdCmTJkiz7npppu0xi5dxLmpqalaRUWFPP7hhx9qALSPPvqoyeMtLS1tMJfZ+7ZHHnmkwW0vv/yyBkBbt27dPh/7oYce0vLy8po8FmHhwoWNvo7zzjtPa9eunfx+7/essrJSA6BNmjRpv48/cOBAbeDAgfL7n376qcF7r2ma9ttvv2kAtDlz5hz0ayA6WMykIDpKTZ8+HZmZmTKdU1EUXHLJJZg1axai0WiLPtd1110n/9/j8aBTp06w2+0YO3asPN6pUyd4PJ64CteffvopTj311LiCYA6HA3/5y19QWFiItWvXAgAuvPBCGAwGvPvuu/K8NWvWYO3atbjkkkvksTlz5qB///5ISUlBWVmZ/Bo8eDCi0SgWL14sx1hfXx+XPktERJSMxo4dC7/fj48//hi1tbX4+OOP97nV49NPP4Ver8ctt9wSd/zOO++EpmlYsGABAKBjx47o1atX3OdqNBrF3LlzMWLECFitVgB7PlfdbjeGDBkS97nau3dvOBwOLFq0CMCez1UA+Pjjjw8qMyE9PR2dO3eWn89Lly6FXq/H3XffjeLiYmzYsAHAnkyKM844o9EtHE11ySWXICUlRX7fv39/AGixzht+vx8AGi1marFY4s4BEPd+lpWVwefzQVXVBseDweB+n/ess85CWlpa3M+ysrISX3zxRdwcaW9WqxUmkwnffPNNg607zeF2uwEACxcuhM/nO+THI9ofBimIjkLRaBSzZs3CoEGDsGXLFmzcuBEbN25Enz59UFxcjK+++qrFnstisSA9PT3umNvtRm5uboPJhNvtjvsg3Lp1Kzp16tTgMU844QR5OwCkpaU12M/57rvvwmAw4MILL5THNmzYgM8++wzp6elxX4MHDwawJx0VAG688UZ07NgRw4YNQ25uLq655hp89tlnh/I2EBERHRbic2zGjBmYN28eotFoXM2IWFu3bkVOTg6cTmfc8b0/V4E9F+1Lly7Fzp07AeypN1BSUhJ3YbthwwZUV1cjIyOjwWdrXV2d/FwdOHAgLrroIkyYMAFpaWkYNWoUpkyZcsALbGBPsEBs51iyZAlOPvlknHzyyfB6vViyZAlqamqwcuVKGVRorjZt2sR9LwIWLXGBDkAGdhp7zYFAIO4cAA3ez0mTJmH79u0Njsdu9WmMwWDARRddhA8//FA+97x58xAOh/cbpDCbzXjiiSewYMECZGZmYsCAAXjyyScbbAFqqrZt2+KOO+7AG2+8gbS0NAwdOhQvv/wy61HQYcEWpERHoa+//hpFRUWYNWsWZs2a1eD26dOn45xzzmmR59pXleh9HddiimMdjEsvvRTjxo3DihUr0KtXL8yePRtnn3020tLS5DmqqmLIkCH429/+1uhjdOzYEQCQkZGBFStWYOHChViwYAEWLFiAKVOm4Morr8Rbb73VrPEREREdLpdffjn+7//+D7t378awYcNk5sKhuOSSS3Dfffdhzpw5uO222zB79my43W6ce+658hxVVZGRkbHPwttikUJRFMydOxfLly/HRx99hIULF+Kaa67B008/jeXLl8PhcOxzHGeccQYmT56MzZs3Y8mSJejfvz8URcEZZ5yBJUuWICcnB6qqHnKQoqXnJXvzer0wm80oKipqcJs4lpOTI4/tnc359ttv4/PPP2/QQaRr164HfO5LL70Ur7/+OhYsWIALLrgAs2fPRufOndGzZ8/93u+2227DiBEj8MEHH2DhwoV44IEH8Nhjj+Hrr7/GiSeeeMDn3dvTTz+Nq6++Gh9++CE+//xz3HLLLXjsscewfPly5ObmHvTjEe0LgxRER6Hp06cjIyNDVrqONW/ePLz//vt47bXX4iL6iZCXl4f169c3OL5u3Tp5u3DBBRfg+uuvl+mMf/zxB+677764+xUUFKCurk5mTuyPyWTCiBEjMGLECKiqihtvvBGvv/46HnjgAbRv3/5QXhYREVGLGj16NK6//nosX748Lq1/b3l5efjyyy9RW1sbl03R2Odq27Ztceqpp+Ldd9/F+PHjMW/ePFxwwQVx2xUKCgrw5Zdfol+/fk2aM/Tt2xd9+/bFo48+ihkzZuBPf/oTZs2aFbctdG8i+PDFF1/gp59+wr333gtgT5HMV199FTk5ObDb7ejdu/d+n/tQtoK0BJ1Oh+7du+Pnn39ucNsPP/yAdu3axf1M9p6rfPfdd7BYLE2aw+xtwIAByM7OxrvvvoszzjgDX3/9Ne6///4m3begoAB33nkn7rzzTmzYsAG9evXC008/vc92qwd6n7t3747u3bvjH//4B5YtW4Z+/frhtddewz//+c+Dfl1E+8LtHkRHGb/fj3nz5mH48OEYM2ZMg6/x48ejtrYW8+fPT/RQcd555+HHH3/E999/L4/V19fj3//+N/Lz89GlSxd53OPxYOjQoZg9ezZmzZoFk8mECy64IO7xxo4di++//x4LFy5s8FxVVVWyz3x5eXncbTqdDj169ADQeJomERFRIjkcDrz66qt4+OGHMWLEiH2ed9555yEajca15gaAZ599FoqiYNiwYXHHL7nkEixfvhz/+c9/UFZW1mB7wNixYxGNRjFx4sQGzxWJRGTXhsrKygYZCb169QJw4M/Vtm3bolWrVnj22WcRDofRr18/AHuCF5s2bcLcuXPRt29fGAz7Xzu12+0AkNBOEmPGjMFPP/0UF6hYv349vv76a9mJ7HDQ6XQYM2YMPvroI7zzzjuIRCL73eoB7OmIJrahCAUFBXA6nfv9me3rfa6pqZHzLKF79+7Q6XScW1GLYyYF0VFm/vz5qK2txciRIxu9vW/fvkhPT8f06dMP+AF2uN17772YOXMmhg0bhltuuQVerxdvvfUWtmzZgvfeew86XXyc9JJLLsGf//xnvPLKKxg6dGiDdNe7774b8+fPx/Dhw3H11Vejd+/eqK+vx+rVqzF37lwUFhYiLS0N1113HSoqKnDWWWchNzcXW7duxYsvvohevXrJfbtERETJ5KqrrjrgOSNGjMCgQYNw//33o7CwED179sTnn3+ODz/8ELfddhsKCgrizh87dizuuusu3HXXXfB6vQ1W8QcOHIjrr78ejz32GFasWIFzzjkHRqMRGzZswJw5c/D8889jzJgxeOutt/DKK69g9OjRKCgoQG1tLSZPngyXy4XzzjvvgOPu378/Zs2ahe7du8taESeddBLsdjv++OOPfRYKjSUyLW655RYMHToUer1+v+03D8Y777yDrVu3yoKQixcvlpkBV1xxhcxQufHGGzF58mScf/75uOuuu2A0GvHMM88gMzMTd955Z4uMZV8uueQSvPjii3jooYfQvXv3A85n/vjjD5x99tkYO3YsunTpAoPBgPfffx/FxcX7fd8KCgrg8Xjw2muvwel0wm63o0+fPli5ciXGjx+Piy++GB07dkQkEsE777wDvV6Piy66qKVfLh3vEttchIgO1ogRIzSLxaLV19fv85yrr75aMxqNWllZmaZph9aC1G63Nzg+cOBArWvXrg2O790qTdM0bdOmTdqYMWM0j8ejWSwW7dRTT9U+/vjjRp+vpqZGs1qtGgBt2rRpjZ5TW1ur3XfffVr79u01k8mkpaWlaaeffrr21FNPaaFQSNM0TZs7d652zjnnaBkZGZrJZNLatGmjXX/99VpRUdF+Xy8REdGRENuCdH8a+1ytra3Vbr/9di0nJ0czGo1ahw4dtEmTJsn2nXvr16+fBkC77rrr9vk8//73v7XevXtrVqtVczqdWvfu3bW//e1v2q5duzRN07Rff/1Vu+yyy7Q2bdpoZrNZy8jI0IYPH679/PPPTXq9okXnDTfcEHd88ODBGgDtq6++ijveWAvSSCSi3XzzzVp6erqmKIpsRyrObazV5t7zn30ZOHCgBqDRr0WLFsWdu337dm3MmDGay+XSHA6HNnz4cG3Dhg0HfI7mtiAVVFXVWrdu3WgLWk1r+J6VlZVpN910k9a5c2fNbrdrbrdb69OnjzZ79uy4++3dglTT9rRv7dKli2YwGORjbt68Wbvmmmu0goICzWKxaF6vVxs0aJD25ZdfNvs1Ee2LomktVE2GiIiIiIiIiOgQsCYFERERERERESUFBimIiIiIiIiIKCkwSEFERERERERESYFBCiIiIiIiIiJKCgxSEBEREREREVFSYJCCiIiIiIiIiJICgxRERERERERElBQYpCAiIiIiIiKipMAgBRERERERERElBQYpiIiIiIiIiCgpMEhBREREREREREmBQQoiIiIiIiIiSgoMUhARERERERFRUmCQgoiIiIiIiIiSAoMURERERERERJQUGKQgIiIiIiIioqTAIAURERERERERJQUGKYiIiIiIiIgoKTBIQURERERERERJgUEKIiIiIiIiIkoKDFIQERERERERUVJgkIKIiIiIiIiIkgKDFERERERERESUFBikICIiIiIiIqKkwCAFERERERERESUFBimIiIiIiIiIKCkwSEFERERERERESYFBCiIiIiIiIiJKCgxSEBEREREREVFSYJCCiIiIiIiIiJICgxRERERERERElBQYpCAiIiIiIiKipMAgBRERERERERElBQYpiIiIiIiIiCgpMEhBREREREREREmBQQoiIiIiIiIiSgoMUhARERERERFRUmCQgoiIiIiIiIiSAoMURERERERERJQUGKQgIiIiIiIioqTAIAURERERERERJQUGKYiIiIiIiIgoKTBIQURERERERERJgUEKIiIiIiIiIkoKDFIQERERERERUVJgkIKIiIiIiIiIkgKDFERERERERESUFBikICIiIiIiIqKkwCAFERERERERESUFBimIiIiIiIiIKCkwSEFERERERERESYFBCiIiIiIiIiJKCgxSEBEREREREVFSYJCCiIiIiIiIiJICgxRERERERERElBQYpCAiIiIiIiKipMAgBRERERERERElBQYpiIiIiIiIiCgpMEhBREREREREREmBQQoiIiIiIiIiSgoMUhARERERERFRUmCQgoiIiIiIiIiSAoMURERERERERJQUGKQgIiIiIiIioqTAIAURERERERERJQUGKYiIiIiIiIgoKTBIQURERERERERJgUEKIiIiIiIiIkoKDFIQERERERERUVJgkIKIiIiIiIiIkgKDFERERERERESUFBikICIiIiIiIqKkwCAFERERERERESUFBimIiIiIiIiIKCkwSEFERERERERESYFBCiIiIiIiIiJKCgxSEBEREREREVFSYJCCiIiIiIiIiJJCQoMUixcvxogRI5CTkwNFUfDBBx8kcjhERER0HOO8hIiIKPESGqSor69Hz5498fLLLydyGETUQur8YVTVBVHnDyd6KEREB43zEqJjS50/zDkJ0VHIkMgnHzZsGIYNG5bIIRBRC6nzh/HVrzvk92eflAuH1ZjAERERHRzOS4iOHbHzEs5JiI4urElBRC0iElX3+z0RERHRkRI7D+GchOjowiAFERERERERESUFBimIiIiIiIiIKCkwSEFELWLbtm2JHgIRERERAOCNN96Q/+/3+xM4EiI6WAxSEFGL+GH593HfG/T880JERESJMWP6NPn/9XW1CRwJER2shHb3qKurw8aNG+X3W7ZswYoVK+D1etGmTZsEjoyIDtb6tavw0EMPwWqzY9CZAzGqH1v4EdHRhfMSomNDZWUldmzdhBsuPQs6nYJff1qe6CER0UFIaJDi559/xqBBg+T3d9xxBwDgqquuwtSpUxM0KiJqjsLCQhTtKAQAdCrgZJ6Ijj6clxAdGzZv3oxwOIzS0kLo9XpUVVUhIyMj0cMioiZKaJDizDPPhKZpiRwCEbUAVVWxY8cO+X11dXUCR0NE1DyclxAdGzZv3oxQKCS/37ZtGzp27JjAERHRweCmcSI6ZJWVlaiqqpLf19TUIBqNJm5AREREdNzauHEjVFUFACiKguLi4gSPiIgOBoMURHTItm/fjpqaGvl9IBBAWVlZAkdEREREx6vCwsK4rKjt27cncDREdLAYpCCiQ7Zu3TqEw2EAe1YswuEwJwRERER0xGmahh07dsgghaIo2L59O7dyER1FGKQgokP2xx9/IBKJQFEUKIqCaDTKIAUREREdcXV1dSgtLQUAOS8pKytDXV1dgkdGRE3FIAURHbKtW7ciGo1Cr9dDr9cDAHbv3p3gUREREdHxpri4GDU1NVAUBTqdDjqdDlVVVSgpKUn00IioiRikIKJDEg6HUVRUBADQ6XRyUrBr1y5ZtIqIiIjoSNi0aRP8fj90Oh30ej0URUEwGGSGJ9FRhEEKIjokNTU1KC8vl8EJg8EARVFQWVmJ2traRA+PiIiIjiPr169HOByGoigwGo3Q6/WIRCLYuXNnoodGRE3EIAURHZKdO3fKfZ5Go1FmU9TU1Mg9oURERERHwqZNm6CqqtyCqtPpoKoqdu/ezeKZREcJBimI6JBs2rQJgUBATgQMBgN0Oh3C4TB27dqV6OERERHRcSISiWDXrl3QNE1u9xBbPsrLy+H3+xM9RCJqAgYpiOiQbNy4EdFoFDqdDkajUQYpmFpJRERER1JdXR3KyspkwUy9Xg+j0QgAqK6uRlVVVWIHSERNwiAFER2SwsJCGaQQgQqj0QhVVVFSUsLimURERHRE7Ny5E7W1tVAURWZRxBbPZIcPoqMDgxRE1GyBQAC7du2SkwGDwSALZ+p0OlRWVsLn8yV6mERERHQcKCwsRCAQgKIock4SWzxTdCMjouTGIAURNVt5eTkqKytlUEKkVYqABVMriYiI6EjZsmULIpGIzKAQQQpN06AoCmtlER0lGKQgombbvn076uvr47IoYrMpIpEIO3wQERHREbFlyxZEo9G4eYmol6VpGioqKhAIBBI9TCI6AAYpiKjZtmzZgnA4LLMoRE0KMTkIhUIoLi5O9DCJiIjoGBcIBLB7924oihK33UMsnBiNRlRXV6O2tjbRQyWiA2CQgoiabevWrQiHwzIoEZteqdPpoCgKioqK2JeciIiIDqvS0lJUVlYCQNy8xGQyyeLegUAA5eXlCR4pER0IgxRE1CyapmHbtm0AEBegMBqN0Ol08nhVVRWLZxIREdFhtXv3btTU1MQVy9y7w0c0GuU2VKKjAIMURNQslZWVKC8vl6sTsSsW4v8VRUFNTQ3q6uoSPVwiIiI6hhUWFiIcDstsztjCmSKbIhqNYvfu3YkeKhEdQEKDFI899hhOOeUUOJ1OZGRk4IILLsD69esTOSQiaqLi4mJUV1fLWhSx3T1i/z8YDDK1koiSHuckREe3rVu3IhKJyIUT8V+DwQAAMBqNUBQFlZWVCAaDCR4tEe1PQoMU3377LW666SYsX74cX3zxBcLhMM455xzU19cnclhE1ARFRUXw+/2NdvYQQQpFUaCqKlMriSjpcU5CdPRSVRU7d+6EpmlxGZ4iszN28aS2tpbFM4mSnKIlUUW70tJSZGRk4Ntvv8WAAQMSPRwi2o9Vq1Zh3ocfIxxR5TGdTgen0wlVVdGqVSuYzWaoqooTOndCz26dEzhaIqKDwzkJ0dEjFAph+vTpWLdhC/QGIwBAURSYzWYYjSY4HHakp6dDURTY7Xac3PskZKWnJHjURLQvhkQPIFZ1dTUAwOv1JngkRHQg7TqcgBPPcjbp3MJqoMAfhsNqPMyjIiJqGZyTEB09TCYTLr70z/jq1x0HPDcK4Ic/qnC2w8F5CVGSSprCmaqq4rbbbkO/fv3QrVu3RA+HiA4gElUPfNIhnE9ElCickxAdfTgvITp2JE0mxU033YQ1a9bgu+++S/RQiIiI6DjGOQkREVHiJEUmxfjx4/Hxxx9j0aJFyM3NTfRwiKgJHnvssYM6X1W5YkFEyY9zEqKjzw8//IDbb7890cMgohaS0CCFpmkYP3483n//fXz99ddo27ZtIodDRE00efJkvPP2Wwd1n88XfnaYRkNEdOg4JyE6OqmqipEjR2LhZwsO6n5+Hzv3ECWrhHb3uPHGGzFjxgx8+OGH6NSpkzzudrthtVoTNSwi2o9du3ahffv28Pv9yM7Nh9VmP+B9/L56pKU4sWDBArRq1eoIjJKI6OBwTkJ0dPr73/8uszubOi+JhEO48k+XYOLEiYd7eETUDAkNUiiK0ujxKVOm4Oqrrz6ygyGiA4pEIjjnnHOwaNGig76vy+XChRdeiP/85z/7/LdPRJQonJMQHX1++OEHDBgwAKFQ6KDuZzAYkJOTg2nTpqF///6HaXRE1FwJDVIQ0dFl0qRJ+Nvf/hZ3TKfTNanehMvlgsViwXPPPYfLLrvscA2RiIiIjgNVVVU47bTTsG7duoO+r8FggMPhwAknnIDPP/8cDofjMIyQiJorKQpnElHyW7p0Kf71r381OG4wNK1JUCQSQTgcxsSJE7Fz586WHh4REREdJyKRCG6//fYGAYqmZmpGIhGYTCasW7cOf//73w/HEInoEDBIQUQHVFxcjJtvvhlVVVXNfoxAIACbzYaioiKMHz8eTOIiIiKi5pg6dSrefffdBseNRmOTH8Pn80Gv12PevHn49NNPW3J4RHSIGKQgov0KBoP4+9//jtWrVze4Tadr+p8QVVVRV1cHs9mMpUuX4vnnn2/JYRIREdFx4JdffsEjjzwCv9/f4DZFUZqcTREMBmE2m1FXV4f7778fxcXFLT1UImomBimIaJ80TcN//vMfzJ49G5FIRB4XwQmDwQCDwXBQEwJFURAOh/Hyyy/j119/PSzjJiIiomNPSUkJ7rzzzrhto3q9Xs5DjEZjkxdQVFVFfX09bDYbtm7dirvuuqtJNbaI6PBjkIKI9mnJkiV4/PHHUVdXJ4/FplKaTCZomtbkrRvRaBShUAhmsxmlpaW4++67D2kLCRERER0fQqEQ/vWvf+H777+PCyYoigJN06DX62GxWJq8cBKNRhGJRBCNRqHX6/HFF1/gtddeO1zDJ6KDwCAFETWqsLAQ//jHP+JWK0SAInYyoKpqk1ctwuEwVFWFoijQ6/VYtWoVJk6cGJelQURERBRL0zS88847eOedd+LajYp5iKIoMJlM8v+bKhgMyjmN3+/HSy+9hOXLlx+Ol0BEB4FBCiJqoL6+Ho8//jh++uknRKNRedxoNMqghMlkgk6ng06nO6gJgc/nA7AnCyMUCmHevHmYPXs2C2kSERFRo5YsWYJJkyahurpaHjMajdDr9QD2bD8VAYuDEYlEEAwGodfrodfrsWvXLkyYMAG7d+9u0fET0cFhkIKI4qiqildffRVz5sxBIBCQx00mE4xGIzRNk/+/v72fex8X30ciEdTV1cFisUCv16OyshLPPfcc61MQERFRA4WFhfjXv/6FLVu2yIUTnU4Hs9mMcDgsF06MRiMURWl0XiIyOPemaRr8fj/C4TBsNhtUVcVvv/2GSZMmyUUVIjryGKQgojiffvop/v3vf8etVhgMBthsNoRCITkZMJvNMqti7ywInU4HvV4Pg8Egj4lsC1VVEQ6HUVtbi5SUFKiqis2bN2PSpEnYtWvXkXmRRERElPTq6+vx1FNP4bvvvpPbPHQ6HQwGAxwOB6LRKCwWCwwGg1xIaSy7s7HMz9g5Sl1dHUwmE6xWK+rr6zF//nzMnDmT21GJEoRBCiKS1qxZg2eeeQZbt26VKZOig4fdbkc4HIbFYoHZbIbZbJYf3nunV4oWYHuvZphMJnl7fX09VFWFy+VCIBDA0qVL8eqrr6K2tvYIvFIiIiJKZqqq4vXXX8fcuXPj2o3qdDpYrVb4/X6ZUSEWTsS21Fix85HYwITBYIBer0c0GoWiKCgrK0N6ejr0ej1KSkrw5ptvYsmSJdyOSpQADFIQEQCgvLwcTz/9NH788UeEw2FomgaDwQCdTgebzQa/3w9FUWC1WmG32xEKhWC1WmXqpVidEJMB0ZpUpFeKyYPBYEAkEpEdPrxeL0wmEyoqKvDhhx9i7ty5cUWxiIiI6PjzySefYMqUKSgrK5OBArPZDL1eD5fLhWAwKBdRxCLIvoIU4it2y4eqqjCbzQD2BD4ikQjKy8uRlZWFSCSC9evX49VXX8X69euP0CsmIoFBCiJCOBzGiy++iE8++UQGI/R6vSxK5Xa75WQgLS0N0Wg0bj+oXq+PS6OMnRCIVQtN06CqKiwWCwDI84uKipCbmwu9Xo/t27dj+vTp+Pbbb9mrnIiI6Di1Zs0aPP/881i/fr3cwiECEXa7HYqiIBqNwm63w+PxwO/3w+l0wu/3Q1VVGYwQGRRi8SR2rhKNRmU2RTAYhN1uR01NDRRFgcfjQSAQwPLly/HGG2+guLg4MW8E0XGKQQoiwqxZszBjxgxUVFQAgJwMaJoGh8MBRVEQDofhdDqRkpKC+vp6uN1u1NfXQ9M0udcTgJwYiErZsVs+xKqFTqeD3++Hy+WCz+dDXV0dMjIyEAqFsGbNGkydOhWrV68+8m8EERERJVR5eTkmTZqEH3/8US6KiJoTer0eXq8XVVVVMBgMaNWqFYLBIADIBRWx0CIyO0WdLLFwIuYpIvhhNptlpobFYsGuXbuQkZEBq9WKiooKLFy4ELNmzeJ2VKIjiEEKouPc999/j1deeQWFhYUAICcCOp0ORqMRGRkZqKyshE6nQ35+PiorKwEAKSkpcsVCrG7EFssUBTVFoEJ8LyYJqqpCVVXY7XaUl5fDarXC7Xajrq4Oy5cvx1tvvYUdO3Yk5D0hIiKiI09kdi5YsAD19fUy80HMMzweD3Q6HUKhEBwOB7KyslBeXo6UlBTU1dXJ7M7YL+B/GRVGoxHA/7aoiu2nBoMBfr8fDocDAFBcXIyMjAzodDps374d8+bNw2effcbtqERHCIMURMex7du348knn8TKlSuhqqpMhRRBitTUVPj9fgQCAXg8HqSkpKCyshJut1u27DIYDHLyEFs9W7QCs1gscuuHpmmIRCIwmUzQ6XSoq6uD0+mEXq9HaWkpUlNTYbFYUFpaikWLFmHmzJkyKEJERETHtpkzZ2LatGmoqKiQQQZRN8JkMsHj8aCyshKKoqB169ZyG0ZaWhpqa2tlm3QRhBDzmdhtqOKYqEMhCoSrqopQKASXy4X6+npEIhGkpaUhEolg3bp1mDlzJr7//ntuRyU6AhikIDpO1dfX44knnsC3336LYDDYYDLgcDhgNptRX18Po9GIvLw8lJSUQNM0eL1e1NTUIBKJyGDE3umVgpgsiEKamqbBbDbLVmHBYFCmaNbX18Pr9UKn02Hbtm349NNPZZ0MIiIiOnYtXboUzz//PLZv3y63eMQunogMTrFwkpaWhrKyMrjdbkQiEQQCAeh0urjaV+K/IrNT0zQ5b4mtmSXmKoFAACaTCRaLBVVVVbBarXC5XPD7/VixYgWmTZuGtWvXJuw9IjpeMEhBdBxSVRWvvfYa5s2bh9raWrnFQwQSLBYL7HY7gsEgQqEQPB4PLBYLysvL4XK5oKqq3ANqsVjiilqJvZ8AZOAidi+oSMUUE4JQKAS9Xg+HwwGfzwedTicnHH/88Qdmz56NxYsXyy4iREREdGzZtm0bHnnkEaxbt052FxOZmmKOIOYMBoMB2dnZKC8vBwBZI0tkd8bWxor9XqfTyS2qYn6i0+lkAU3RGSQUCsHpdEJRFNTW1sLlcsFqtaKyshLLly/H9OnTuR2V6DBjkILoODR//ny8+uqrKC8vh8FgkPsxxYTAarXKmhFms1mmUSqKAqfTiWAwiGg0KicAYoVCBDnERMBgMCAajcpVC/Ff4H+VtqPRKKLRKKxWKxRFgc/ng9lsht1uh8/nw9q1azFt2jT8+uuv7FVORER0jKmrq8ODDz6IH374QS5ciOCEaH1uMpnkvMTtdsNqtaK6uho2mw2qqiISicjgxt7dxcTWDjFfEQsqYhuICFyI+0YiEQCQ7daj0SicTicMBgOKi4uxaNEizJkzB1VVVQl814iObQkNUrz66qvo0aMHXC4XXC4XTjvtNCxYsCCRQyI65q1atQoTJkzArl27ZAqlaDcqAhSxWzjEB3NdXZ2cDGiaJicDsXs9xf/HTgxEUarY2wHICQGwpw2Ypmmw2+2IRqNQVVVOSmpqarBy5UpMmzYNGzduTNj7RkTHNs5JiI48VVXx+OOPY8GCBfD5fDCZTDJAYTQa5VwgtvC2aDkqim+LxQ4xpxH2rpcVW8RbnCceOzaQITI3xSJOMBiETqeD3W6HqqrYuXMnPvroI3z88ccIBAJH/k0jOg4kNEiRm5uLxx9/HL/88gt+/vlnnHXWWRg1ahR+//33RA6L6Ji1e/du3H777di0aZNcTRDZE0ajUdajEGmQRqNRriTodDqYzWYZpBCThdj/j+3kIYISsbfHdvYQ5+r1ejkh0Ov1sFgsUFUVOp0ONpsNwJ52ZMuWLcO7777LXuVEdFhwTkJ05E2dOhVvv/02amtrZYDCaDTCaDTCYrHIuYNYUHE4HNDpdAgGgzIDE0DcvEIsooh5jHgMMbdQVVUWzxRbQQDI+yiKIuclFosl7jHNZjNCoRC2b9+O2bNn46uvvuJ2VKLDQNGSLH/a6/Vi0qRJuPbaaxM9FKJjzowZM/DiK68jEt2zPcNkNEFv0MNoMMJoMsoPYfHhDUAWuBQTANGho6KiQq5cxK46RCIR6HV6qJoKnaKDBg0KFESiEblaYTAYYLPZEAgE5KRB7DcF9kw2RJFNcY7NZkO7du1wwagRGNCvb8LeQyI6fnBOQnT4+Hw+3HTLHVj733Vyy4UofKlAgcVqQSQcgcVqiescBvxvniDmCrW1tfDV+/Y8sAJEwnu6dmiIv8wR8xAxl9Hr9VBVdU/rUQ17im/q9yzKiJpbAGQXkEgkAr/fD6PRCLfbjb59+2LMRaPRKiv9iL53RMc6Q6IHIESjUcyZMwf19fU47bTTEj0comPSyNEXw553dP/7qgRQ5w/DYTUe8FwioubgnITo8FMVIy4Ydw8uSPRADtHPm+rgdns4LyFqQQkPUqxevRqnnXYaAoEAHA4H3n//fXTp0iXRwyI6JkWix0Zv72PldRBRcuGchOjIOZY+y4+l10KUDBLe3aNTp05YsWIFfvjhB9xwww246qqr2H+YiIiIjjjOSYiIiBIv6WpSDB48GAUFBXj99dcTPRSiY8qmTZvw2KTnMOKKOxI9lEN2cns3WmV6Ez0MIjrGcU5CdPg89+LraHvSOYkeRosY2DMHHoc50cMgOmYkPJNib6qqIhgMJnoYRMeMJUuWoH379mjfvj0+/eTjRA+nRYwdcxHef/99JFmMlYiOMZyTELWsaDSK2267DSaTCU8++Xiih9Ni9Dol0UMgOqYktCbFfffdh2HDhqFNmzaora3FjBkz8M0332DhwoWJHBbRMWHGjBm46667UFRUJI8V7SjEDZeeBavNnsCRHRq/rx5FOwpx8fdLcNppp+Hpp5/GKaecIjuDEBE1B+ckRIdPTU0Nrr32WnzwwQeIRCIAjo05idvtRru2+dhwyom44oorkJmZmeghER0TErrd49prr8VXX32FoqIiuN1u9OjRA/fccw+GDBmSqCERHdUikQgmTZqEp556ChUVFYkezhFhNBoxdOhQTJgwAT179ozreU5E1FSckxC1vM2bN+Paa6/FkiVLZLvyY41er0ebNm3QrVs3jBw5EhdddBFSUlISPSyio1rS1aQgooNXU1ODe+65B2+//TZ8Pl+ih5MQNpsNgwcPxt13341TTz0VJpMp0UMiIiI6Li1evBjjx4/HmjVrjputmWazGa1atcKJJ56I4cOHY9SoUQxWEDUTgxRER7EtW7bg9ttvx4IFCxAKhRI9nKRgs9nQv39/3HrrrRgwYADs9qM3jZSIiOhoEY1GMW3aNDz44IPYtm1booeTMCJY0atXLwwbNgzDhw9HVlZWoodFdFRhkILoKLRs2TLccccd+Omnn6Cq7M29N0VRYLPZcMopp+C6667D0KFDkZaWluhhERERHXN8Ph+efvppPPfcc8fNVtOmsNlsyM7ORqdOnTBkyBCMHDkS7dq1S/SwiI4KDFIQHSWi0Shmz56NBx98EJs2bTpu0icPhaIosFqt6NatGy6//HKMGDEC+fn50OmSrrERERHRUWX37t24//778e6776K+vj7Rw0laTqcTGRkZaNOmDfr164cLL7wQvXr1YsFvov1gkIIoydXW1uLFF1/E888/j5KSkkQP56hltVpRUFCAUaNG4cILL0SPHj1gMCS0wREREdFR57fffsO9996LRYsWIRwOJ3o4RwW9Xg+HwwGv14vs7Gz06NEDo0ePxqBBg2A0GhM9PKKkwyAFUZLavn07Jk6ciFmzZqG2tjbRwzkoOp0uabehmM1m5ObmYvDgwRg9ejTOOOMM1q0gIiLaj2g0ik8++QQPPvggVq9enbSf8fuSLPMSvV4Pj8cDt9sNj8eDdu3aYdiwYbjooovgdrsTPTyipMEgBVES0TQNv/76K/7xj39g0aJFCAaDiR7SMUmn08FoNCI9PR19+/bFyJEjMWTIEBa2IiIiilFfX48333wTTz/9NLZv337UbjVNliCFYDAYZKDCZrMhMzMTAwcOxGWXXYZ27dpxKwgd9xikIEoCkUgE8+fPxz//+U+sWrUqqXuJ2+12+P3+Q/qw1+v1SfEaFUWByWSC2+1G9+7dZWGrzp07c4JARETHraKiIjzxxBN45513kqYYZmNzB71eD7PZfMjt1w0GAyKRyCE9RnPo9Xq4XC54vV45H+nVqxcuvfRSnH766dwKQsctBimIEqimpgaTJ0/GCy+8cNS06zIajfvcg6ooStwqy76CESIAkCx/fhRFgdFohN1uR/v27XH66adj9OjROO2002AymRI9PCIiosNO0zSsWLECEyZMwJdffnlUFMPU6/VQFAWRSKTBHATYU4/K7/cD2JNNoShKo/MSg8GAaDSa0HmJ1+uFx+OBwWCA1WpFhw4dMGLECJx//vlITU1N2LiIEoFBCqIE2LZtGx5//HHMnj0b5eXliR5Oo/YVSDCbzQiHw9DpdPtddVAUBU6nEzU1Nfs8Z38BjyNNURTodDro9XqZennSSSdhxIgROPfcc5GSkpLoIRIREbW4SCSCzz77DI888ghWrVqVtFtN95dJEQqFoChKgzmF1WpFMBiEqqry4n9fdb4URYFer09IRkUsp9OJlJQUGI1GmM1mZGdnY+DAgbjgggvQpUsX6PX6hI6P6EhgkILoCNE0DT/99BMmTJiAb7/9NulXKGw2GwKBQINtHeIDf1+rEWazGZFIBNFoFDabDX6/Py7QEbsvNPbcZCKCFRaLBV6vF+3bt8fQoUNx4YUXcq8oEREdE+rq6jB58mS89NJL2Lp1a9J9FsfS6/WwWq2oq6uTx8SWTYfDgfr6ekQikQYBBqPRKIMXer0eBoMBgUBA3r53rYpkWjwxm83wer0wGo0wmUxIS0vDiSeeiPPOOw8DBgyAy+VK9BCJDhsGKYgOs3A4jPfffx+PP/441qxZkzQffvuj0+lgs9niJgMA5CpEXV1doymRiqLAbDbLIIbRaIxbkTGZTNA0Tb4HIisjmYpZxRLBCrPZDJfLhZycHPTr1w8XXXQR+vbty72iRER01NmxYwcef/xxzJkz56hpbW42m2EwGOIWeHQ6HTIyMuD3+2WQQhwX8woRpIhEIjITI3bOYbPZ4upZGI3GBuckml6vh9vthtlshtlshtvtRkFBAQYPHoxBgwahU6dOXDyhYw6DFESHSXV1NV566SVMnjwZO3bsSOoVir2ZTCbo9XoEAgHodDo5dovFAlVVEQqF5LkGgwEA5ARApEoajcYGWRJ2ux2aph1ygasjTafTQafTwWw2w2azITU1FV27dsUll1yCIUOGwOPxJHqIRERE+6RpGn755RdMmDAB33zzTYNFiGQmikuGQiGZnalpGmw2GxwOB2pqahAKheSWDp1OJ+cpsYsJYl4jGAwGOBwOVFdXN1h4SZYC33tzOBwyWOFyudCmTRv07t0bgwYNwumnn86W6nTMYJCCqIVt3LgRjz32GD744IOkqYh9MK23RDqkKCAlViDEbSJooaqq/F6n08mMCYPBALPZ3CAYIWpU+Hw+OZZkWqloKpFZYbVa4XQ60bp1a4wcORKjRo1C+/btuZpBRERJIxwOY968eXjyySfx+++/J029iabOSxRFgc1mi6sVIbaiWiwWmZEZCASgaZoMUsQukogL90gk0iC702q1orq6Gnq9HpqmyTElW8vSvZlMJlgsFrndJScnByeccAJOP/10DBo0CPn5+ZyP0FGNQQqiFqCqKhYvXoxHH30Uy5YtS7pMgcYqXu/L3lsYVFWVH/RiX2QkEpFFqmK3hogJQlpaGmpqamRapviwt1qtCAQCMJlMUBQFBoOhwWqOOPdgxpwIIrPCYrHAYrEgLS0N/fv3x2WXXYaTTz4ZFosl0UMkIqLjVE1NDZ5//nm8+eab2L59e1JccIuLZkVRmhygiF340Ol0MBqNqKqqgqZpMBqNsFqtck4i6k4YjUaoqopwOAxN0+ByuWCxWFBeXi7nM2JLKrAneGEymWAwGBAOhxsEcsR8pLHARaLnKjqdDiaTSXYoy8jIQEFBAXr16oUBAwagb9++nI/QUYlBCqJDEAgE8Pbbb+OFF17A+vXrE14RWoj9ID2Y1QpxvqgcLapgi6wJ8SEYjUZleqTRaITT6URdXR1UVYWqqnC73VBVFTU1NfJxRRpmJBKB1WqFwWCApmkIBAL7fN/2NSEAkq99qVjRcbvdOOGEE3DppZdi8ODByMjISPQQiYjoOLFhwwY8+uij+Oijj5ImmxOIb/HZ2Gd77DFFUWSnDZGtKYIINpsN27dvl7c7nU7odDr4fD74/X4oigKHw4FoNIpwOCwDEHa7HT6fD4FAQD5+bPDBarXKz/Oqqqq4rR5i7rKvLSAtsTVEvM7mziPF+yECNykpKcjPz5dt1c8880y0bt2a2RV01GhWkKK+vp57nui4VlJSgqeeegrTp09HUVFR0lwwA/EfljqdDkDDbRXiHPEBHZvmqNfrYTKZZIaEaJGqKIqM0ldWVsrjLpcLBoNB7hMVH+SiQrZ4XNHDXFTjFi3DxPhiVy5iAyaNffDvr9d5IomOIOKrTZs2OO+88zB69Gh07tyZbcOIDgPOSeh4p6oqvvzySzzxxBP4/vvv4ff7Ez0kSafTyRoSQOPdM8RnOgAZNBD3UxRFLgK4XC5UVlaisrISBoMBiqIgPT0dRqMRO3bsQCQSgcFggMfjgd/vl5mge9eqiK1ZodfrZdFvq9WKaDSKUCiEYDAYV4hTZH/sa17SUl1B9i4w3hxisclkMsHlciE7OxsdOnRAz549MWDAAJx88skwm82HPFaiw6lZQQqHw4GxY8fimmuuwRlnnHE4xkWUdETRqcceewxffPHFPvtsJ9LeFa0jkUiDAIqYCOj1eqiq2mBVw2AwwOVyyYh+dXU1DAYD9Ho9vF4vzGYzduzYgVAoBIPBgIyMDNTV1ckP7Wg0ikgkIoMcABAMBuWkSaQdigmISN/0+/1xgYq9K2zvnVIpLvgPNlBxJFIzxWsSe0bT09PRt29fXHrppejbty+cTudhfX6i4wnnJHS8CgQCeOONN/Dqq6/ijz/+SJpsTmHvDIPGMg5EwW2xXTQUCsFkMgGAXExxu92w2+1QFAXbt29HNBqVWRWpqamoqqqSCycpKSkAEDefEF3ELBaLDEqILFGDwSCfz2azya2sRqMRZWVl8jFEUEQED0R2hSACA43Nu/anscUk0W41toZXc4iFIZFdkZaWhvz8fHTs2BFnnHEGBg4ciOzsbGZXUFJqVpDigw8+wNSpU/Hpp58iPz8f11xzDa688krk5OQcjjESJVQkEsHcuXPx1FNPYdWqVUndQlRcgDe2JUKsBMSmOEajUXmxL+5ns9nQpk0b7Nq1C/X19XEf7E6nExUVFTJrwm63yxoVIrXS4XCgqqoKOp0ObrcbVqsV9fX1qKysBAA4nU6ZVRE70QiHw6iqqpLj3XtCICYRe3cWEemjzXmf9nY4CmWJCZDJZILT6US3bt0wevRoDBkyBHl5eZwcEB0izknoeLNjxw489dRTmDFjBkpLSxM9nH2K/azdO0AhLp4bm5eYTCZZmwoAWrdujUAggMrKSvh8PhiNRhgMBtkxrKysTC56uN1uOU8IhUIwGo2ykLfFYoHb7YZOp0NpaamsYWGz2aBpmpyviMWa3bt3x2VTiMUTcczlcsmtruI1Ag0XT5oyt9g76AHsWdQRW1kOhXh/xWtwu93IyclB+/btcfLJJ2PgwIHo1auXnGcRJYNDqklRWlqKd955B1OnTsV///tfDB06FNdccw1Gjhwp2xISHa0qKyvx7LPPYvLkydi9e3eih7Nfe6dT7n0RLvZbBoNBuaUjtmZFbAAjNzcX1dXVsue4zWaTk4VoNCrbd1ksFhnkMJvN8oM6JSVFrj54PB5Zr0Icc7lcsnimCDDYbDbodDpUVFTIcVksFrl1RKyIpKWloba2Nm6F5FD2gu5r4nA4si3E5EAEZfLy8nDWWWdh9OjR6NWrFwtbER0izknoWKZpGpYtW4aJEyfi22+/jWulmYxiP0djt26IY3a7XW7HiG1lrtPpZPZmMBiE1WqFx+NBRUUFQqGQnDOEw2EYDAb4fD75GCIjQ2Rn1NfXw2azIRKJwOfzwWQyIT09HaqqygUXsYggsj9ramoAAOnp6SgtLZUBArFwEts5RNS6qK6ujgtUxBYGjT2+d5Bm73mGwWCQtb0EsW0jFAod1FynsccX8zzxuKLod35+Pnr06IH+/fvjjDPOQEZGBhdQKOFarHDmiy++iLvvvhuhUAhpaWn461//invvvVemex/I448/jvvuuw+33nornnvuuZYYElGz/Prrr3jwwQfx1VdfJf0kAGh4ob33vk4AMkAhPrzF7WIlQ1TCNhqNMl1RZDuID2ARpBCTjdgVB5vNhqqqKhmwEB/qNpsNdrsdgUBAFtqMTacUdSlin0esgIi2YmKMgUAARqMRDocD9fX1cRkVjWlqoCERlblFm1ez2Yz09HSceuqpuOCCCzBw4EBODohawKHOSQDOSyg5BINBvP7663juueewZcuWRA/ngMSFfGwXDZH1KOYfRqNRzhVEAF/MOUSxTLHwYrFYEA6HZR0J0R2svr5ePo9YaBGPabfb5aKMmAeJOURKSgqi0Shqa2vh8/lgNpvlY4rMUJ/PB6/Xi2AwCJ/Ph2AwKOc3oVBIPo+maXC73bIjiBhL7EJQbBv3vTNE936vhNjtuOI8keVwMC1kYzNUAMQ9npg7iqCQ3W6Pa2M6cOBAdO3aldkVlDCHFKQoLi7GW2+9halTp2Lr1q0YPXo0rr32WuzYsQNPPPEEcnJy8Pnnnx/wcX766SeMHTsWLpcLgwYN4mSAEubue+/HJ59+ljR9xIXYbRyxKxPiYlZ8+McGJvb+r1ihiE37i23BFfscsZOBcDiMcDgsJxGiRZeI7ItiVMFgUD62WAnYU4jKhkDAL3uYiwmIqqoypTMcDss2YmIrSezKighYqKoq0zvFuBp7b2KzRGIzTMSxxqqK7/0hDuy7bsf+NOW82IkYALma0bNnT5x55pkY0P8MuOwsakV0MFpqTgJwXkLJYf369bj2Lzdid3FJUhXojrX33EF8H/t5L24XWQZiHgHEZ3PGzmEEMR+I/VwXNa5ii4CL5xELLiLTIhQKQa/Xy/mC0+mEqqoIBALyNpFNIbI36+vr5TG/3y8zOMR2D2DP4k9dXR0MBoPM/BSLQeI1iSCAKCouAhhiW4sYq9FolEEV8RrFXEqcL95Po9EosypiF6XEV2zR0dj3HfjfNhTxHsd2XxPzN4fDgfT0dHTs2BH5+fk4b9h5yMnJgsMa356e6HBrVpBi3rx5mDJlChYuXIguXbrguuuuw5///Gd4PB55zqZNm3DCCScccLWzrq4OJ510El555RX885//RK9evTgZoISo84fx1a87Ej0MIgDA2SflclJA1AQtOScBOC+h5FFd58c3K5N7uykdHzgnoSNNd+BTGho3bhxycnKwdOlSrFixAuPHj4+bDABATk4O7r///gM+1k033YTzzz8fgwcPbs5QiFpMJNqyBROJDgV/H4mapiXnJADnJZQ8tOZN04laHOckdKQ1q5JUUVHRAfd1Wq1WPPTQQ/s9Z9asWfj111/x008/NWcYREREdJxrqTkJwHkJERFRMmhWiNbpdKKkpKTB8fLycrnv6UC2b9+OW2+9FdOnT2dVe0oKRUVFiR4CEREdpJaYkwCcl1DyWbRoUaKHQESUEM0KUuyrjEUwGGxyFdhffvkFJSUlOOmkk2AwGGAwGPDtt9/ihRdekFWAiQ4nVVUxe/Zs5OfnQ1EUnH3WoEQPiUjq2+cUXHbZZfjuu+/495BoP1piTgJwXkKJV15ejquuukp2gbjpxhsSPSQiAIBBz61HdGQd1HaPF154AcCearBvvPEGHA6HvC0ajWLx4sXo3Llzkx7r7LPPxurVq+OOjRs3Dp07d8Y999xzUKsfRE21e/duPPTQQ5g2bZpskykU7SjEDZeeBavNnqDREe3h99WjaEch1q9djVmzZsFms6Fv3764+uqrMWrUKLhcrkQPkSjhWnJOAnBeQkeeqqr47LPP8MADD+DXX39tcDvnJU0n2po6nU54vV6ccMIJOPHEE9GuXTtkZ2fDbrc36BxysCKRCGqqa1BdU42amhpEwhEYjAa4XC643W54PJ5Dfo5kZNDrWDSTjriD6u7Rtm1bAMDWrVuRm5sb94FtMpmQn5+PRx55BH369GnWYM4880xW0aYWFYlE8Omnn2LChAmNTgCIjjaKoqBjx4645JJLcNlll6FDhw68eKLj0uGekwCcl1DL2717NyZNmoS33noL5eXliR7OMcVoNMo2n3q9HhaLBRaLBV6vFxkZGWjdujXatWuHE044Ad27d0dmZiacTqdsxdmY+vp6lJeXo7y8HDU1NdA0DQ6HA6mpqUhNTT3g/YmoeZrVgnTQoEGYN28eUlJSWnQwnAxQSygsLMSTTz6Jd999FxUVFYkeDtFh5fV6cc455+Cqq65Cv3794HQ6Ez0koiPqcM1JAM5L6NCFw2F8+umneOKJJ/DDDz9AVdkl4XAzGAwwGo0wGo2wWCzQNA06nU4GMAwGAywWC9xuN1JTU+OCF23btoXdbkcoFEJlZSUCgQD0ej1SUlKQmpoKr9cLs9mc6JdIdMxrVpCCKJkEAgHMmzcPzz//PH7++WdOAOi4ZTQa0a1bN4wbNw7Dhg1Dfn4+DIZmNXEiIqJm2rRpE1555RW88847KC0tTfRwjkk6ne6A8z1FUWA0GmGz2WAwGGSGhdFolN9rmoZIJAJVVRGNRqHX62G1WuHxeNCmTRt07doV3bt3R15eHnJzc+H1evm5SnQENDlIcccdd2DixImw2+2444479nvuM8880yKDI2qMpmlYs2YNXn75Zbz33nsoKytL9JCIklJGRgZGjRqFsWPH4pRTToHb7U70kIhaBOcklEx8Ph/ef/99vPzyy/j5558RDocTPaTjkiiUG41GYTTuqaEQiUSg1+tlJoXRaISiKNA0DYqiwGAwwGq1wmKxwOFwQK/XQ1VVhEIhBAIBGQgRt3s8HuTm5qJdu3Zo164d8vLykJ2dDa/XK4MhRHTomvwv6bfffpN/dH/77bd9nsd9WXQ4VFVVYdasWXjjjTewcuVKRCKRRA+JKOmVlJRg8uTJmDx5MiwWC3r27InLL79cZlmISRzR0YZzEkokVVXx22+/4fXXX8eHH37YaAtcall6vf6AHXZCoRCcTqfMjLBYLMjJyUEkEkEkEkEgEEA4HIaqqjJoEQwGEQgEoNPpUFFRAYvFAqfTifT0dHg8Hrm1o6amBjU1NSgqKsLWrVuxePFi6HQ6mEwm2O12pKSkICMjA7m5uXL7SOvWreHxeGCz2WC1WvmZS3QQuN2DklI0GsXSpUsxefJkLFiwgMWliFqQoijIzs7GsGHDMHr0aJx22mlISUnhBR0R0T6Ul5dj5syZePvtt/Hbb79xsSSJGY1GmRGh1+tht9vhdDpl0MHv96O+vh7RaFQGK0KhEPx+P0KhUFyWhdlsht1uR2pqKtLT0+V2j3A4jFAohOrqalRXVyMQCCAYDELTNLmtxGazwe12Iz09HampqcjKykLbtm2Rk5ODjIwM2O12WK1WWK1WmEwmfgYTxWhWkKK0tBTp6emN3rZ69Wp07979kAdGx5/t27dj2rRpmDlzJtauXcue9ERHiM1mQ/fu3TF69Gicf/756Nixo0ybJUp2nJPQ4RAKhbB48WL85z//weeff87FkqOIoijQ6XQQlzh6vR4mkwlOpxMulwtt2rRBbm4uIpEIysrK4PP5oNPpYLPZoCgKSktLUV5eDp/Ph1AoBFVVoSiKfByr1QqXywWv14u0tDSkpKRAp9MhGo0iGo2ipqYG5eXlCIVCCIVCiEQiMvBhNBphtVpht9tll5D09HTk5OQgKysLGRkZSElJgd1uh8VikVtRGMCg402zghRZWVl48803cf7558cdf+qpp/DAAw/A7/e32ADp2OXz+bBw4UJMmTIFS5YsQVVVVaKHRHTc0+l0yMzMxKBBgzBq1CiceeaZSE9P5wSJkhbnJNQSNE1DYWEh3nnnHcyZMwf//e9/uViSAE3Z1tEUIhNCp9NBURQZZDAajdA0DRaLBVlZWWjTpg1sNpusQWE0GpGRkYFWrVrBaDRi06ZN2Lx5M8rKylBbW4twOCzHJ55DFOdMSUmRXyLjIhqNIhAIwOfzobq6WgY+otGoDFyI1y2CElarVW45ycjIkNkYHo8HHo8HTqdTnmexWKDT6Q75/SJKNs0KUjz55JN48MEHMW7cODzzzDOoqKjAlVdeidWrV+P111/H6NGjD8dY6SgnCl6+/fbb+OCDD7B582Z24iBKchaLBZ06dcKIESNw/vnno1evXrBYLIkeFpHEOQk1V319PebPn4+3334by5cv52JJEhB1G1qq+Gjs55XJZEJKSgrMZjMqKysRDAZhMBhgs9ng8XiQmpoKq9UKYM+c1ePxoEuXLujVqxfMZjP++9//4vfff8e2bdtQVlYGv98PVVVlpoQIPBiNRphMJpjNZng8HqSkpCAtLS0ucCG2itTU1MitJ+FwWAZVdDqdDLIYDAY5TpGBkZaWBpfLJbND3G43PB4PXC6XrIFhtVqh1+tb5H0kOtKaXZPit99+wxVXXIFgMIiKigr06dMH//nPf5CVldXSY6SjWGVlJebMmYPp06fj119/RV1dXaKHRETNpCgKUlNT0bdvX4wYMQLnnnsucnNzuYpDCcc5CTWFqqpYsWIF3nrrLXz88ccoLCzkYkkSEls1DAYDFEVBKBQ6pMczmUwyOAAAqamp6NSpE6LRKEpKSlBfXy8DDaIFqcPhkPezWq3Izc3Fqaeeir59+yIrKwubNm3CDz/8gNWrV6OoqAg1NTWIRCIyM0Js/YhEInG1L0R7U4/Hg4yMDKSlpcFisSAcDqO+vh61tbWoqalBbW2trHEhtokAkHUyRGaITqeDxWKBzWaDw+GQY3c4HLIjicvlks8ZWweDnUgomTU7SFFbW4v/+7//w3vvvQcAeOONN3DVVVe16ODo6BMOh7F06VJMnToVX331FXbs2JHoIRHRYWI2m9GmTRsMGTIEI0eOxOmnnw6n05noYdFxiHMS2peSkhLMnj0bM2bMwKpVq1BfX5/oIR3XRPtPYE8wIjZIJLIGRHcOALBarQgEAmjO5crez2Wz2eIe1+v1wuv1ykCA6Pbh9/vh9/sRiURkEU4xNrfbjQ4dOmDAgAHo27cvCgoKUF1djd9++w0///wz/vvf/6K0tBTBYFAGLEwmE1RVRTQaRTAYRDAYlAENRVFgs9ngdDpljYv09HQ4HA5omoa6ujpUVVWhvLwc9fX1CAQCCIVCUBQFZrNZjs9oNDYYq9lshtlslgELRVFgsVhgsVhkV5LYAEbsNhLWpaJEa1aQYunSpfjzn/8Mr9eLadOmYenSpbjjjjswbNgwvPbaa0hJSTkcY6UkVVhYiGnTpmHevHlYu3YtgsFgoodERAngdrvRs2dPDB06FKNGjUKnTp24UkOHHeckFCsYDOLbb7/FW2+9hUWLFqGoqCjRQ6IYIvNOVVUYjUZEo1EZOBCZC8FgMK4uhV6vl58lYo65d4CjKRRFgclkgqZpMJlM8Hg88Hq9sFqt8Pl8iEajcLlccLlcAPZkQ1RXV6Ourk5u5RABFIPBAJfLhQ4dOqB///445ZRTcMIJJ8But2PTpk1YsWIFVq5ciW3btqGyslJmUxiNRrmlQ1VVhMNh+P1+GXzQ6XSyM4jT6URKSgpSU1ORkZEhC3vW19ejuroa5eXlqKyshM/nk0EQEawwGAxyy4l4TLPZDJPJFBeoMJlM0Ov1Mpgj7mO1WuF2u+U2ErfbDZvNBovFIjM5iA6nZgUpzGYzbr/9dkycOFGmH23atAl//vOfsX37dq6eH+N8Ph8++ugjvP322/j+++9RWVmZ6CERUZIxGAzIzs7G6aefjuHDh+Occ85BRkZGoodFxyDOSY5vmqZhw4YNmDZtGt577z1s3LjxkLcH0OFlNpvl9goRaBCr/FarFTU1NQD2BCJEbQqLxYJIJIJwOCy3PxiNRtkKtKnEfUXQwOl0Ijc3F7m5ufD5fKiqqoKmaXC73bBYLPJi3+/3o6qqSgZQRFaEyGoQXUN69+6NPn36oKCgAKmpqQgGg9i6dSvWrFmDDRs2oKysTG4vEe+F6Bqi1+sRDAZRU1ODuro6mXEhbhctTb1er/yv1+uFXq9HeXk5amtrUVFRgcrKStTU1CAcDss2rOIxRKFP8XwiIGKxWOB0OuX/iyAKsCdLOhKJyKCGxWKRGRgiiCG2kbATCbWUZgUpvv32WwwcOLDBcVVV8eijj+KBBx5okcFRchB7ON98800sWLAA27ZtY8VrIjooVqsVHTt2xNlnn40RI0agT58+skAZ0aHgnOT4U1tbi/nz52Pq1Kn48ccf5UUtJZ/YLRcHIrYgqKoKVVVl+06bzQav14uKigrU1tYC+N/WEJPJhFAohHA43OTnERkVIlBiNBqRlZWF9u3bw2g0orKyEqFQCHa7HXa7HcCez7DU1FSEw2GUlJSgsrJSZlWoqopAIAC/3w+dTgeHw4GcnBx07twZJ510Etq0aYO0tDTodDqUl5ejqKgIhYWF8vWILAoROLFarbDZbDCbzQgGg6iqqkJVVRV8Pp8MkohWqHa7HR6PR7ZCTUlJgcPhgF6vh9/vR1lZGcrLy1FeXi6DLLEBEpF1IbaMiG0vZrMZTqdTPr7L5ZKdRDRNk2MWgZrYYqGxhTzdbnfcNhLWsKKmanZNCgDYuHEjNm3ahAEDBsBqtca10qGjW0lJCWbOnIlZs2Zh5cqVbOFGRC1Gr9fD7Xbj5JNPxuDBgzF69Gi0bduWVcjpkHBOcuyKRqP46aef8MYbb2DhwoXYuXNns2oUUMs4mDah4qJ2Xz+vxupSKIqCvLw81NXVobKyEiaTCV27doXD4cAPP/wg64qITACxZUT8m9/7ufa1NURkXonX4nA40KNHD7Ru3Rq7d+9GRUWFrEMhLq7dbjdycnKQkpKC0tJSbN++XQY1NE1DKBSSF++iDkZOTg4KCgrQtWtX5Ofnw2azQafTob6+HmVlZSgtLZUBiEgkImtiBAIB+Rgulwt2u112BamsrERlZSUCgQCCwSBUVYXZbIbdbofT6URaWhoyMzPhcrmQmpoKh8MBg8Eg39PS0lLs3r0bVVVVcZkdOp1OZkuI1yy2g8QGL1JTU2WrVfHYInhRV1cnX09s/QxRHyM2gOHxeGQGBreHUqxmBSnKy8sxduxYLFq0CIqiYMOGDWjXrh2uueYaeL1ePPXUU4djrHQYBYNBfP3113jzzTfxzTffoLy8PNFDouOI6DXeUi3H6OhiNBqRnZ2NAQMGYOTIkTjrrLOQmpqa6GHRUYJzkmPTzp07MW3aNMycORPr1q1jvaskIALJIiAgtlvEOlCtiNj7iNaagUCgwXkGgwFer1dmDmRnZ6Njx44oKSnB77//Lrd4GI1GaJomO2HEEkFKUQtC0zS5BWLvMYlAh16vR3p6Ok4//XR4vV5s2rQJNTU18sJdbPVwOBxo27YtCgoKoCgKtm7diuLiYnnRH41GZXaFCAA4HA5kZGSgffv26N69O/Lz8+FwOBCJRFBZWSnv7/f75fYXAAgEAvJxDQaDDES4XC5Eo1HU19fLoIUIXIj3x2QyycBAVlYWsrKy4Ha7kZ6eDqfTCYPBgFAohPLycuzcuRMlJSUoLS1FVVUVQqGQrL8hCnGKzAvxmlRVjQtepKWlITs7G6mpqTITQxQkrampkTU+RABDBFdEgEVkYaSkpMDj8ch2qiKgRMePZgUprrzySpSUlOCNN97ACSecgJUrV6Jdu3ZYuHAh7rjjDvz++++HY6zUgjRNw/r16zFlyhTMmzcPhYWF8g8OEVEiWSwWdOrUCUOGDMEFF1yA3r17x/W6J4rFOcmxwe/3Y8GCBXjzzTexdOlSVFdXJ3pIFEN0jYhGo3JLxt6dUkTdCDGfFFkNIrghggbidnHhLy7qG2M2m2GxWOK2M6xbtw4VFRXQNE1e4Pp8PtTW1jYaqDAYDLIFqE6nQyQSaTToFZuFYTAYkJGRgW7duiEtLU1mS5hMJtku1O/3w2azIS8vDx07dkRqair8fj/Ky8tRUVGBiooK2ZHD7/ejrq5OZlyYTCakp6ejffv26NWrFzp16oTMzEwAQHFxMbZu3Ypdu3ahuroaBoMBDocDTqdTFs6sqamBz+eTNSVcLpd8LwKBgMy2qKiokOeK4IwICKSkpCAzMxNZWVnwer1o1aqVDFxEo1HU1taiqKgIW7duRVFRESorK1FfXy8LfMbWqBBZEKJ+hfi5iyCJ1+tFTk4OWrVqhZSUFNjtdhn0qa2tlVtaROtVkVUinkPU4xDZF2JbCzuRHLuaFaTIysrCwoUL0bNnTzidTjkh2Lx5M3r06IG6urrDMVY6RJWVlZg7dy7+85//YMWKFY1GromIko3H40GPHj0watQojBw5kltDKA7nJEcnTdOwatUqTJ48GfPnz8fOnTsPulsDNZ/ZbJYXzPvjcDjg9/sRjUahKIpcHRcXyoLYxicuxIH/dfIQgQKj0SgvlvV6PRRFgdvtRjgcjnus2GCCkJ6eDovFIi9SN2zYIDMOPB4PzGYzioqKZLvSvRfeDAYDbDab3LYQDAbj5sFirLFbU0TdC5EhkJqaCrfbLYMJPp8PO3fulAG17OxsdOnSBTk5OQD2ZECIIEV1dbXcAlFbW4va2lp5wS/eu1atWqFnz57o3bs32rZtC7vdjkAggA0bNuCPP/5AUVERgsGgbJ/qdDqhaRp8Ph/q6upkhorNZoPNZpMBDGBPEFAU1ayqqkJ1dbX8uep0OlgsFtjtdqSmpiIzMxMZGRnwer3Izs6G1+uNe5zi4mJs375dBi9EYEG0bBXBEqvVCkVRZL0NEcASnUYcDgdSUlKQk5OD3NxcpKamyu0jOp1OFjIV4xVjDgaDCIfD8nGsVmtcHQyxlcRqtbITyVGsWUEKp9OJX3/9FR06dIibEPz8888YOnQotwokiUgkgu+++w6vvfYavvzyS/5cqEU1p/0X0aHS6XQyFXfs2LEYMmQIvF4vJyHHMc5Jjh6lpaWYPn06pk6dirVr13KLX4J4vV7U1NTsN4PWZDLB6XQiEAjA5/PJ7R0iW0JceAL/y7Iwm80yKJiTk4PS0tK4bRQZGRnyYl3UkhDbBOrq6uKCBqKOgdjGYTQa5UVvSkoKwuGwrOsgMg0URUFJSYm8be9WpsCevxfAns8SUT9CnCM6U8QeExfworikXq+H3W5HRkYGOnfujLZt28Ln82H9+vUoLCxEMBiE0+lEhw4d0KFDBxgMBiiKIrM3Yi+yRaBEtBMVARyr1Yr09HS0a9cOPXr0QM+ePZGXlycDMWvXrsXmzZtRVlYGALDb7XC73bJmhdhiIjJfxDYN8TpE1kNVVZUMWohAgKhvodfrZWHO9PR0pKeny0BNRkaGzGRwOByIRqOyPkdhYSF27tyJ0tJS2boVgCysKQIHwJ4gTyQSQTQalT8zs9ksC6Xm5OQgJycHmZmZSE1NRWpqqiyUKra4iODF3ttIRKaH2WyWGRixQQzRxYSFPJNXs4IU5513Hnr37o2JEyfC6XRi1apVyMvLw6WXXgpVVTF37tzDMVZqgk2bNmHy5MmYNWsWtm/fzotIanEHU6n7aHgeOrrp9XpkZ2djyJAh+POf/4y+ffvKFR86PnBOkrzC4TAWLlyIl156CcuWLZOdGShxxHaF/X2+KooiixlWVlbKbhiapskVcQAyO8JoNMpVeVVVZYBDdN0QK+jighbYExAQF+qiQ4boMCFajYrtAOICVgQRRA0rcSHqcDhktw6/3y8fN7ZWRWyrU6fTKcehqqp8DrENw+12w+fzyeAMsCcQIJ5HFMY0Go2yhWnv3r2RlZWFzZs3Y9WqVSgvL4der0dOTg7at28Pj8cDIL6uRyAQQCAQkB01otEoysvLUVZWJreX6HQ6uFwuZGVloXPnzujYsSM6duyINm3aIBqNYuvWrVi/fj127doFv98vgwsia0RVVVnEU3yJ57PZbDL4AkDWxhAX/SJwITJuDAaDbFWalZUlt+B4vV6kpqbC4/HIbSnid2Lnzp3Ytm0btm3bJreviN8V0Z1FBJn0ej0ikQh8Pp+sVyFazorOLykpKcjOzkZmZiZycnKQnp4uC4MCkEE1EYARryW2i4r4OYsAWWPbSCwWCzM2E6xZQYo1a9bg7LPPxkknnYSvv/4aI0eOxO+//46KigosXboUBQUFh2Os1Ii6ujq8++67eO2117Bq1Sr2BqdjUmyRJqL9MRgMaNOmDYYPH44rr7wSPXr0YMGtYxznJMnl999/x/PPP4/58+ejuLg40cOhgyS6NBiNRnmRaLfbZYAiHA7LC0xxIazX6xEKhWQmhMfjQX19vcy6iL3YFI8vWlXqdDoUFRXJC/K9L6IDgQCi0ajcFmG322WGhciUMBqN8Hg8yMjIQFVVFerq6mSnDfH8wJ7AgMgwEAUdY7ta1NfXy0KQGRkZsFgsKC4ulltIxMV0WloarFYrampqZG0LnU4Ht9uNjh07olu3bqipqcGGDRuwfft2RKNRuN1u5ObmIiMjQ9ZQENkN4nWIzCKz2QyDwYCamhoUFxejrKwMNTU10DQNFosFHo8H2dnZ6NChA3Jzc9GhQwd4vV7U1dVh586d2LFjByorK2VNCPGeiwyO2PfYZDLFFb8UrcEVRZEBDhGsqK2tRWVlJWpqauT1hvg9cLlcyMjIiAtciNalTqdT1o+IRCIoLS2VW0a2b9+O8vJyuV1FvM8mk0m2YxW/X2J7jKh5oaoqjEYjbDYbUlJSZGHQrKws5OTkyOCF6E4iaoOIAIx4XTU1NTLDRRQKFYE2t9sdV8xTbCNhJ5LDr9ktSKurq/HSSy9h5cqVqKurw0knnYSbbroJ2dnZTX6Mhx9+GBMmTIg71qlTJ6xbt645QzouRKNRLFmyBJMmTcLixYu515aOOfsKSDCzgppDBC1Gjx6Nq666Cl27dmV65zGIc5LEqaysxOuvv44pU6Zg48aNzOA8jMTFeUsSF6viQlVVVRlMEFsy9Hq9vHgT6fmiGKWqqrIDQ0lJiVz5r6mpgaIo8kI8HA6jrq5OZkuI7hViW0ZsfQRRb0BcoKampmLnzp0IBAJyBVwEn8XWBnEBb7fbYbPZUF9fL1P/xQq9qIMhCn5aLBbZ2cNkMsFoNMZlDpjNZqSnp8NoNGLHjh1xq/putxtpaWmyaGc0GpWBANF2tF27drBYLHILhCi0KQpV2u12uR1EXOyLbBTxcxC1JUKhECoqKmTQQtQIsdvt8Hq9sgBm69at0apVK5jNZtTU1KCkpAS1tbUyYCEyTmK3RsQG8sXWSfF+iDGKMUWjUfh8PtTU1MjaGmVlZTLIoNPpZDDJ4/EgMzMTXq8Xbrdb1tFwuVwy40JkPlZUVKCkpARlZWXYsWOH7DAiAjjicUXwIrblaSAQkNtpxN8fEYjyeDxIT0+XwYtWrVohIyMDTqczLlNCbJOpqamJq9tRU1Mjf5fEOMSWFJfLFVcDIyUlBTabjQsjLaTZQYqW8PDDD2Pu3Ln48ssv5TGDwYC0tLREDSlpffvtt7j/wUdQXFLKCQAlpZasUaEoioziE7Uki8WCLl264IorrsBZg86Ew8rJBO3BOcnBueuev2Ph51/C5/MleijHhZbIKNw72C8CDZFIJO5zV1w4iyCFuGAWq/5iZToSichVZ71eD7/fLy8mY7MfMjIyUFtbC7/fD+B/RSotFgsCgYBc2ReZDiIIIi6Mc3JyoNPpZHFVUdPAZDJBUZS4jhKapslimrFtLsXrjB2rCMCI5xIp/uJCXGRdWCwWuFwuWexSZI+I4o+pqamy7oR4nlAwBL1BL1f1rVYrioqKsH37dvj9fhgNRqR4U+Bxe+D2uKHT6eT4RXBGfIVCIdmGMxqNQq/bE3TZsXNPxoRoqyoKaooLZZFd4nA4YLPZZJ0IRVFkhoPYilNZWQm/3y9/1iaTSb5HFrMFesOeuiMpKSkySBAIBOTFeyAQkIEoscVCbLsRv2tpaWmyxoXT6YTH44HVYgUUyC0q4oI/FArJgEFlRSUqq/YEDkT2ividFT9P8d5EIlFEoxEZUAkEAvL3So2q0Ol1MqgmAiedOnVCTk7OPreKiq05ooOM+Kqrq5MZGLG1NywWiwxi5OTkIMWTApv9wNtQDXod5yQxmhykWLVqVZMftEePHk067+GHH8YHH3yAFStWNPmxj1dz3/8IxoxuiR4GEdEx5eyTcjkpOApxTpJYdf4wvvp1R6KHQUR0TOGc5H+avKGmV69eTUq3VhRln72OG7Nhwwbk5OTAYrHgtNNOw2OPPYY2bdo0+f7Hi8FDzsG3K3clehhERMeUSJTZOkcjzkkSi/9uiIhaHv+2/k+TgxRbtmxp8Sfv06cPpk6dik6dOqGoqAgTJkxA//79sWbNGtkiiIiIiCgW5yRERETHroTWpNhbVVUV8vLy8Mwzz+Daa69N9HCSSlVdkJkUREQtbGDPHHgc5kQPg5IQ5yT7xjkJEVHL45zkf5pd4nz9+vUYP348zj77bJx99tkYP3481q9ff0iD8Xg86NixIzZu3HhIj3Osufvuu3HmmWcmehhERERJiXOSI2fXrl246KKLEj0MIiI6hjUrSPHee++hW7du+OWXX9CzZ0/07NkTv/76K7p164b33nuv2YOpq6vDpk2bDqpl2LFs9+7dOPHEE/HUU0+hpHh3oodDRHTM0euURA+BDhHnJEfOm2++ifbt2+O/a39P9FCIiI45Bj1bpAvN2u5RUFCAP/3pT3jkkUfijj/00EOYNm0aNm3a1KTHueuuuzBixAjk5eVh165deOihh7BixQqsXbsW6enpBzusY4amaXjnnXcwfvx41NbWyuPZufmw2uwJHBkRUXIzmUwwGAyyTZ7D4YDT4YTD6QAABPwBeFI8iEQicLuc6Nf3ZFx88cVwu90JHjk1F+ckh19FRQUuv/xyLFy4UB7jnOTI0el0stWiXq+XbT5F20tgTztPk8kEs9kMRVHg8/lgsViQnZ0Np9MJl8sFg2FPKbpQKASHw4FuXbuh3xn95PGmUlUVoVAIgUAANTU1qKmpQXV1tWzLWFdXJ9t/apoGg8EAo9EIu90Ol8sFl8sFvU6P7Tu2w2AwIDMzE2pURSS6px2mqqryfK/XC6fTiV27dmHHjh2IRqPw+/yorKpEKBSSLUItFguMBiMUnQKdTgej0ShbhNrtdjgcDtl2U7TzVBQGqSl5sAVpvGYFKWw2G1atWoX27dvHHd+wYQN69uzZ5J7Zl156KRYvXozy8nKkp6fjjDPOwKOPPoqCgoKDHdIxo6SkBNdccw0WLFgAVWWFVyKig2EwGKDX6wEADocDXq8XqqoiHA6jTZs2KCsrg9VqRX5+Pmpra2EymdC9e3dcdNFFOPnkkzlpPQpxTnJ4vf/++/jrX/+KkpKSRA+FYuj1ehgMBtnhxmw2y2CEzWaDwWCA3++H2WyGx+OBTqeDwWBAeno6UlNT0bdvX0QiERQUFKBbt26wWq1xf/80TUMwGITf70d9fT2qqqpQXl6OqqoqVFdXw+/3IxAIQFVVRCIRhMNhqKoqAwwWiwVOpxM2mw12ux12ux1msxlWqxX19fVYv349FEVBbm4u9Ho9jEYjbDabDLZEIhEUFhZizZo1WLFiBaqqqmCz2WCz2eByuZCamoqMjAykpKTA7XbD7XbD4/HA4/HA7XbDZrPBZDIl6sdDRIfo4EKn/9+ZZ56JJUuWNJgQfPfdd+jfv3+TH2fWrFnNefpjkqZpmD17Nm6//XYUFRUlejhEREcNvV4Ps9mMYDCISCSCaDQKnU4Hn8+H1NRUuFwuVFZWorCwEHa7HRUVFTCbzWjdujXC4TDWrFmD0tJSnH766Rg9ejS8Xm+iXxIdBM5JDo+qqirceOONmDt3LsLhcKKHQ3sRrXVjA7PBYBDV1dUIBAIwGAxQFAVVVVUoKSmBxWJBamoqamtrsW3bNuzYsQOqqkKn06Fz587IzMyE3W6HXq+XmRKapiEajcqMBUVRZCBC07S4v7ciMOB2u5GamorU1FTY7XZYLJa4gMfWrVuxZs0a6PV65OXlIRAIIBqNwufzobS0FLt370ZpaSlqa2sRCAQQDoeRnZ2NYcOG4YQTTkBubi48Hk9cRgQRHXualUnx2muv4cEHH8TYsWPRt29fAMDy5csxZ84cTJgwATk5OfLckSNHttxoj1Hl5eX461//ig8//JATASKiZlAUJW5VUdM0aJom04utVqtMIw4GgzCZTEhNTUX37t0RDocRDAbhcDjQunVrDBs2DKeddpqc+FNy45ykZWmaho8//hi33HILCgsLEz0cagKdTgedTgez2YxoNAqDwQCLxQKdbs/+dlVVZRvdSCQCh8MBRVEQjUZRV1cHAPKCX2Q+iG0S4v+NRiNMJpM8LjIy0tPTkZ6ejpSUFFitVuj1epSXl6O4uBilpaUoKipCVVUVfD4famtrUVxcDJvNhtTUVITDYfh8PtTX1yMcDkOn08Hj8SAnJwdWqxUOhwM9e/ZE586d5WshouNDs4IUTf1DIf4AUuM0TcN7772Hu+66C1u3bk30cIiIjgli33Y4HIbNZpN7uE0mE9xut6z1o2kanE4nTj31VJhMJtTW1srVwE6dOmHMmDHIyspK8KuhA+GcpOVUVVVh/PjxmDdvHvx+f6KHQ4dABC7E3z/xu68oitwqYjabYTQaZd0Ig8Ega1yIWg8mk0lusXA6nfB4PEhPT4fb7YbX64XL5YKmaaivr0dtba3M5AiHwzAYDHvqAjmdCAQCKCsrk9kWmqZBVVXY7XakpaUhLy8Pbdq0gd1ux4YNGxAIBNCpUydkZGQk8m0kogRpVpCCDl1FRQX++te/4qOPPkIgEEj0cIiIjklmsxl2u12mG4tibOFwGIqiIBwOw26348QTT0R2djZKSkqQnp4Oh8MBg8GAIUOGYODAgdzbTMc0TdPw0Ucf4Y477sDmzZvBqeGxSVEUGI1GGdjTNE3WoVAUBRaLBVarFRaLBXq9HtFoNK5YpyiwKbLSotGoPMdkMslCl61atUJWVhays7PhdrtRU1ODXbt2xRXyzMjIQOvWrZGWlgaXywVFUVBWVoZ169bBZDKha9eusNtZmJXoeMUgxREmsifuvvtubN26lRMBIqLDzGg0yqJw4XAYJpNJTtZDoRAikQisVis6deqE7t27Y+fOnfB6vWjdujWqqqqQl5eH0aNHo23btiysScec6upq3Hjjjfjwww9RX1+f6OHQESSCD4qiQNO0uECECGCIv3mim4gocimKYNpsNpjNZvl3NRqNykKakciebh2apiEtLQ3t27dHfn4+2rdvj7Zt26JVq1bIzMyEwWDAli1bsG3bNqSnp6NTp04H3XGEiI4tzQ5S/PTTT1i0aBFKSkoadKF45plnWmRwx5ry8nLccMMN+OSTT5pcbZyIiA6dmHyL1nSKosjsCL/fj2g0CpPJhDZt2uCUU05BaWkpTCYTTjzxRPh8PgSDQfTv3x+DBw+Gw+FI8KuhvXFOcvA0TcP8+fNx5513YsuWLewoRlJsm1MRvBDHFUWJ20Yi2j6LL5F5If6+Kooiu29omiZrAAGQdS4AwG63o1OnTujWrRvy8vKQlZUFt9sNi8XCgAXRcahZQYp//etf+Mc//oFOnTohMzMzbmVJURR8/fXXLTrIo53o3HHPPfdg+/btnAgQER1hsSuCosimxWKBy+VCJBJBTU0NQqEQ9Ho9MjMzcdJJJyEYDCIcDuOkk06C1+vFrl27kJmZifPOOw/dunVjYc0kwTnJwauqqsJNN92E+fPny8KJRI0Rf+dEYEL8LVVVVWZgiKwLg8EAk8kEq9UKs9kMTdNgtVpht9thMplgsVhkMU6j0Qi/34+qqioAe7qUiOwLALKIp+gW0qpVK7Ru3VoGMBwOBywWi9yaQkTHlmYFKTIzM/HEE0/g6quvPgxDOraUlpbir3/9KxYuXMg0SiKiJCBWAoE9E2GPxwOTyYTKykq5wme321FQUIDU1FT4/X60a9cOPXr0QGVlJerr63HiiSdi6NChSE9PT+RLIXBOcjA0TcOHH36IO++8E9u2bZMXhET7Ehvcjf1/VVXj6lWI20RBTFGM02g0yi0iIoghtoZEIhFYLBZ4vV54PB7ZWlQU8/T7/aitrUVNTY0MGgN7Aieic1NKSgoyMzNlDYzs7GykpKTA6XQyiEF0FGtWkCI7OxuLFy9Ghw4dDseYjgmqqmLmzJm4//77mT1BRJSkxORa7K0OhUJyO57oBpKbmwudTofU1FT06tULbrcbu3fvhtvtxplnnok+ffqwsGYCcU7SNOXl5bjpppvwySefMHuCmkX8vdw7oyK2rkVsVoUIBsdmN4lzVFWVgYvYLSIikGGxWGCz2ZCSkgKv1wu32y0DGNFoFNXV1TJo7PP5EI1G5ZjMZjMcDgc8Hg/cbjfS0tKQkZGBzMxMpKSkwOVyyW4lVquV7U2JklCzghRPPvkkdu3aheeee+4wDOnoV1RUhBtuuAFfffUVJwJEREcBkcpsMpmgaRoikYisX2EwGOB0OuF0OpGWlobOnTsjPz8fwWAQ9fX1aN++PQYPHszCmgnCOcn+aZqGOXPm4L777sPWrVvZhpVazN71KYD/FdzU6XRx9SxES1RN06DX62WxTVFoMxKJxAUaYoMd4vFFwU6n0ym3gaSmpsLr9cJms0FRFPj9ftTV1cmvQCAgi9SrqioLfoqtJw6HA2lpaUhPT5eZHC6XC263WwYyWBOD6MhrVpBCVVWcf/75+OOPP9ClSxcYjca42+fNm9diAzyaqKqKt99+Gw8++CB27tzJ7AkioqOM6PohVvt0Op0slCkmyV6vFx07dkRBQQE8Hg/q6upgtVrRu3dvDBgwAE6nM8Gv4vjCOcm+FRcX46abbsJnn33GLadHmdjtFeLv0dFCBCTElyDqWIgghvibKgIGVqtVbs0IBoPySwQwxDaSxp5HtE91u93wer0ygOF2u2EwGKCqqtwyEgwGEQgEEAwGZTBEBE9EUERkclitVpnJ4fp/7d15eFzleTf+78ycmTmz7yNptFuWLXlfWAI4NktYHEigpQ1J+r5vm6ZZ4E1I0l7NLyVJS5smTduk6ZulgSSEzQQCSUkCJUAg4GAIYGOIMUbebUnWLs2+z5zz+0N5nhzJC7axPWP7+7kuX9jSLGdGQnrO99zPfXu9cLvdMrQWYQa3lBCdeMcVDd5888145plncMkllyAUCvHKEYD+/n7ceOONePbZZzm5g4joNKXrOkqlkvy32WxGOp2Wjd8qlQqGhoYwOTmJ/fv3Y9GiRWhtbUWlUsGGDRuwe/duvPOd78TSpUtZQnyKcE1yME3TcN999+GLX/wiBgcHWT1xmprd6+FwPUTESXa90DTtkBfqxOsxm82yaqJUKiGbzWJqakpOYRIVEy6XC9FoFB6PB263GyaTSfaoECGDaLaZyWSQTqcxOjo6o5pDNEl2u92yh0UgEIDf70ckEpHVF8bqDfF7IJVKIZVKYWJiQh6/+LkuKjLEf0VVhjHAEP+12+2w2+0HBahEdHjHVUnh8XjwwAMP4Oqrrz4Zx3RaqVQq+N73vocvf/nLGBkZYfUEEdEZyuVywel0ygV2tVqFqqro6OjA4sWLEQgEYDKZ4HQ6MXfuXFxyySVobGys9WGf8bgmmenAgQNyyykvmpx44iT1RK/3jiZoENslgOnmkSK0MI4LPZ4gwxiEnGqismJ2uGgMaETjTTE5xOv1IhgMyv4SAFAul5FMJpFOp5HP55HL5VAqlVAqlVAsFmVwInpoiD4YDodDBhgiDPH5fHC5XDP6bIjwRByXsSJEVN0ZPyZGXVutVjgcDthsNtkrY3aQoarqjCCDQSvRcYYU7e3teOKJJ9DT03Myjum0sXv3bnz84x/Hhg0bUCgUan04RER0CjidTtjtdiiKIq8GOhwOdHZ2oqenRy46w+Ewli5diosuugh2u73Wh33G4ppkWrVaxR133IF//ud/5pbTU0D0SxDbEN6K1WpFpVI5bBAgRnYaK7ne6vkBzAgt7HY7LBYLNE1DuVw+5GMZm1we7rjFSbymaSe1CuetgpnZjTfFf42BgKIosuoiGAwiFovJHhVWqxXlchm5XA7ZbBbZbBb5fF7+11iNYQwwzGazDBVEFYYYnSoe1+FwzBi7amz6KbZ+iCkmAA6agmL8+on7G4MMEZyIUasixBC9PFipR2e64wop7rzzTjz++OO488474XQ6T8Zx1bVKpYJvfvOb+Ld/+zeMj49zIUBEdJYRi1jjHmaTyQS32422tjbMmTNHNnSLRqNYvXo15s2bxytkJ8HZviYBgL179+Kmm27CM888I8fo0lvzeDzyKnulUjnh2ybMZvNBa0Rx4ny0txfEia/4GXK4igmxxcFY9SVOxmcT40FF88pDvXYxzUNM1SiXyye84sJ4wm187CNVfhibcBq3aRi3edhsNrjdbkSjUTQ2NiIYDMLr9cLhcKBcLssAQ1RbZDIZZLNZpFIp5PP5Gf0wjGGQ1WqVf0QvDKfTKUMGsQVEBBeqqsqJJeJ3hbidoihyEooIhozfi6KPh2jsLP6IEMPlcsnHNwYZIrAiOl0dV0ixfPly7N69G7quo6Oj46A9Vps3bz5hB1hvtm3bho9//ON46aWXjjrtJiKiM5e4qmXcT+3xeNDU1ISOjg60trYiGo2is7MTl156Kfx+f60P+YxyNq9JyuUy/uu//gv/8i//grGxsbrqS1DvFEVBMBiEzWZDPB6XJ4cADgoRxBXw2Y0bD2f2VXJd12c85qHCCONIz8M9nmjqazabYbfbUSgUZLgw+7jEybTJZILP54PVapVbIUql0ozbiylG4rUbT8pnv2fiduI1iNc1+72Z3ePhaAMg8bN09vshTtiPFFyIP+JkXzBOGRFNMV0uF0KhEBoaGhAMBhEOh6Gqqgx2RINNMZY6l8shk8nIvxsrMMQxGRt5GntriP+KigjR40iECeLjIlgRHxPBhqqqsmpH9OAQ742YRCWe3xhkOJ1OuN3ug8ILUY3BPhlUz46rceZ11113gg+j/pVKJXz1q1/Ft771LdlAh4iISATWYo94oVBANptFMpnE0NAQotEourq6kEgkcODAAZx77rk499xzeZXrBDkb1yTA9EWTj33sY3jppZcOe2WeDk+MrHQ6nahWqyiVSjKsAKarLPL5PCqVyowTXpvN9pYXqRRFkV+TUql0VCMsD9WXQRAnwZVKRQYP5XIZiqLMCCjEzyDRM0d8bGpqasbUCofDIaddlEolWVUgbiOmYYjnEa9fjAoVjyu2WiiKIm8r3i/jNhFjgGDc7nCo4GF2001RGQFA9oQwEs8n3gfjcxufT7weTdNkk82RkRFs27ZtRqWC0+mU403D4TBCoRAikQhcLpd8PvGeiHGnouqiUCigUCigVCrJr0M2m5Whz+ytH6KSRVEUWY1h3NohKmKMgYXb7ZYjVEV/DrGtRGwXEl/bTCaD4eHhg953EVAYt5gc6Q/7ZFAtHFclxdlm48aN+PjHP44tW7YctryOiIgImLmoNplMcmHZ2NiIBQsWYN68eejo6MCll16K1tbWGh8tnW5KpRK+8pWv4Fvf+hampqZqfTh1zThJApgOD8Q6TpTeOxwOOBwO5PN5TE5OyhNmRVHg8Xig6zrS6fQhT7oPtzVDXPWeXVkg7nOoRpUiGBCPKU5gxZhMI3Gl3m63w+l0Ip/PI5/PQ9M02WAyl8vNGPlpt9tnNN4UnzObzahUKiiVSrLHjgg/xPOIK/aiZ4PxOI0nv+LE29iDRwQXxhN1cQzG7RnG9+RIvTBmBx5i+4R4DPE8xgBk9tdJPO/hmnYaPy9CBPFz3O/3w+/3y8adbrdbbtcQr9PY70K8ryIMMoZIohrD+LqN41lF400RJIh+GGKroQhVRMWEy+WC1+uVo1NdLtdBAYP42omwpVqtzqhcEcTziADKWH1xuMoM9smgE+lthRSvvPIK3nzzTQDAwoULsXz58hN2YPUgl8vhc5/7HH74wx9yvjgREb0tojN9U1MTli1bhsWLF2P58uW4+OKL4XA4an14p70zfU0CAJs2bcJHPvIRbNmyhf2wfk/s5Xe5XEgkEvLjJpMJHo9HXtm2WCxwOp3IZrPyZF6EE+L25XJ5xom8zWZDZ2cncrkcxsbG5AmnkbHHhPEE3LjtQNM0WTklggBgZkghTu7FFg5x0mg8kdd1XT6/8cRS9J/I5XLQdV0GHMZwAZh5BR/AjGOzWCzyucUxGgNXccIvTtjF1XrRo8J4W9ErwjhFw1hJIF6HMUwwhjfi9sb3SLzuQ4Ujh/u78T0yvv8iBDFWccx+XOMxGE/exd9F1Yk4QRdNLv1+P6LRqNxmIQIM0fdCbCMxBj5iG4f4u5hKIipcxNQWY+hkfH+N20rE+22svBBTRAKBAHw+nww1jNNKjO+3eD4RYswOgsTtxHsgzA4x2CeD3o7jCinGxsbw/ve/H88++6zcW5tIJHDJJZfggQceQCQSOerHOnDgAP6//+//wy9/+UvkcjnMnTsXd955J84555xjPawT6rHHHsONN96I/v7+mh4HERGdeaxWK8LhMJYtW4Y1a9Zg7dq1WLJkSa0P67R0NqxJcrkcbr75Ztx7771nVT+sIzWSFJ8XUzFUVZ0RUgDTWzpE+bvNZkNDQwOGhoZkqCBOIIE/nJyJEzbjFWhRVaCqKtLp9IyvgXHihKggMI5+nX3imE6nZTAwezKIKMU3BiHGSgBVVWV/C3G8xqvX4kRSnLgaQxfxmsSJsLGPgfivCC8KhcKM55hd/SFOksWJMoAZvRIOVRlhDDGMvRtEpYuoMjFu3TC+B8a/i+M3himzv1bGY539PSOOXZx8i2MUvTqMVRjG74vDbXkwBjrGnhSigaYYaxoKheB0OuFyueQJvvF9E99X4v0TxyhuI74uIkAQYYaxL4kxEBJ/xPeg+B6Y3XzT5/MhEAggEonA6/XC6/XKyhtjWCbee/H9KP4+u4+I8Wsx+zRTvC9HqshgnwwCjjOkuOGGG7Bnzx7cc8896O3tBTC9N/LP//zPMXfuXNx///1H9TjxeBzLly/HJZdcghtvvBGRSAQ7d+5EV1cXurq6jvWwTojR0VH8yZ/8CTZs2FCT5yciorOLoihobW3Fddddh7/5m79Bc3NzrQ/ptHImr0kA4J577sGnPvWpg07AzwaHa7ZonK4gtjWI24sTpEOdQHV2dmJsbEy+l8aQAwACgQAmJiZQqVRk5ZM4MdQ0DQ6HA3a7HZlMRgYd4uRPBA+qqs4Y/2kyTTeuFNUcol+FeMzZUzVmn8yLfgMi+BDBijiJNp7gG6+Ki2aY4vHFNg4xAUdsORDBgAhGxNVvccW/UqnMuGIuTpJnT7sQW0REtYk4YRbbBcRWDPG1MgYXxq+bMeAQt5kdUhi3ccxukClCitmVIOLxjLedvbVC/F18TcXrMJ6UG4OL2QHT7OOc/bgiMBBTQZxOJ7xeL4LBoNzO4XQ65feGsSpEBELGySU2mw2VSuWgRp7FYlH+W1RpGN8H8T1jDGDE948ICcSxieNrbGxEQ0MDPB4P3G73QWGUcZLM7PdFfM/Ovr3JZJJbYozE/5eH+yOen85sxxVS+Hw+PPXUUzj33HNnfPzll1/GFVdccdS/SD/3uc/h+eefx3PPPXesh3DSbNu2DX/8p+9HuXL45F4r5eFGBhm4YbaxRJeIiN4eu92OtrY23PJ3n8Pqi86v9eGcVs7kNQkA/K//9b+w/rnfwqYeer1xrGuSw534HyvjVVPjFgQxAUKMvBTbA8RzlkoleUKjaZrseSAmVwDTJ8Jut1teKa5Wq7J5oNVqhdvtRiAQADB9cSmRSMDn9cEf8AMAJiYmYDKZkMlkEIlEZGl6OBxGPp/H8PAIdF2Dx+NBtVqFz+dDMpHEnK45mJiYwNjYGKrVKtxutzxZzefziEQi8sQwk8kglUpBURS4nC6UK9PjLAOBgAwsEokEbDYbfD4fXC4XJicmYbVZ4XK5UCgUYLFYMDY2Jr8moVAIk5OT0DQNdrtdhgENDQ1QFAXxeFxWdORyOdhsNjlho1gsyqoJ1a6iqlXlNjLxOZNpukdOS0sLLGYL4om47NFhMpkwNTWFaqUKt8eNpsYmjI6NIplMyq0ykXAEVa2KoaGhGSfw4gTY5XLB5XJhanJKNtk0madDGr/fD4vFgng8jkwmI6tbXC6X7M9QKpXkSa240m+xWJBKpZDL5WSzUxHCiKakJpMJ+dx0E8tiafr7ToQnYjSoyWSS21lsVtsfesyZpt8fMVFD3Fd8L5rNZnmybxw37fP5UCwWMTExAcWioKpNvwZR+VAoFOT3nSBel3EbkMfjQSQSQbVaRSAQQKwphnQmDbPZjHnz5snArFwuY2pqCvF4HLlcDpFwBJdfcTmy2Szi8TimpqaQyWQQDofle+n3+1EoFJDP5+FwODAxMYFEIoFMJiODg0AgMCOEqlar8nmymSwq1QrsdjsCgQCWL18Oj8eDlpYW+RzFYhHQgbb2Nvn/muhxYryN+HupVIKu6fD6vOju7gYAWR0kwr1yqYxS+Q9bXsT98PsfW8uWL+O2kVNIsZjhdpz66pbjCik8Hg+ee+45LFu2bMbHX331VaxZswapVOqoHmfBggW48sorMTg4iPXr16O5uRk33XQTPvKRjxzrIZ0wmXwZT28ePOznf33HLbhKfxprYgWsH1LxuOkyXPrhr5zCIyQiojPZZStaarIgOF2dyWsSABifSuGFNycP+TmuSYiI6GSrxbrkuNqwXnrppfjUpz6FoaEh+bEDBw7gM5/5DC677LKjfpw9e/bgu9/9Lrq7u/HEE0/gxhtvxM0334y77777eA7rhKhUD19Bsa/vNVylP4217UU4rSasbS/iSv1p7Ot77dQdIBERndGO9HuIDnYmr0kAwGqzH/LjXJMQEdGpUIt1yXGFFN/+9reRSqXQ0dEh92p2dnYilUrhW9/61lE/jqZpWLFiBb7yla9g+fLl+OhHP4qPfOQjuO22247nsE66ga3PYU2sMONjF8cKGHyD/SuIiIhqgWuSP+CahIiIzgTKW9/kYK2trdi8eTOeeuop9PX1AQB6e3vxrne965gep6mpCQsWLJjxsd7eXvz0pz89nsM66VoXvRPrN6zD2vai/NizQypaVq2q4VERERGdvc70Ncn69esBf/dBH+eahIiIzlTHVEnx61//GgsWLEAqlYLJZMLll1+OT37yk/jkJz+Jc889FwsXLjymhlMXXXQRtm/fPuNjO3bsQHt7+7Ec1gm19fUth/1cR88yPG66DI/ttyNX1vHYfjueMF2Gjp5lp+4AiYjojKZYjqvI8axzNqxJAODnP3v4kB/nmoSIiE6FWqxLjqlx5nvf+15ccskl+MxnPnPIz3/zm9/EM888g4cfPvQv1Nk2btyICy+8EP/4j/+I973vfXj55ZfxkY98BN/73vfwZ3/2Z0d7WCfcjt378bm/uwW/+93vDvl5TvcgIjozzB5BKDqTizntolu+oihwOJyoVP4wfk9VVQSDQbhcLhSLRVitVjQ1NaGpqQk2mw3JZBJjY2MoFArwer0IhUKyU7nH40F7ezuWLl2Kjo4OOU6tVl20T0dny5oEAL7/w3vw7e/8FzKZzEGfO1PXJGJihHHKgnHEpMfjgcvpQr6Qh8vlkpMK2tvb4fP5kE6nMTIygt7eXpRKJbS3t6OtrQ2vv/465s6dC13XsWfPHixbtgypVArJZBKxWExOEIAJUFX1938ccKgqVMf0eNGBgQGUSiXEYjE0NDQc9TjEXDaHA0MHsHfvXkxNTgEA3B433G43rFYrisUi8vk88vk8dF2Hx+OBx+NBpVJBuVyWky5GR0cxPj6OdDotbxcOh9HS0oLW1lb5c8npdMJut3NcIxEdt9Niukd7ezsef/xxOYd8tr6+PlxxxRXo7+8/6gN49NFH8Xd/93fYuXMnOjs78dd//dc176QtbNiwAR/72Mewbdu2Wh8KERG9DWazGRaLRY49BCDHGYrRfCKgEKPrxFx6RVHgdDoBAE6nE4FAAI2NjQgGg8jlckin0wiFQli6dClisRiGhoawb98+JBIJNDY2IhwOY3x8HGNjY/D5fOjp6cFFF12E+fPn8+ThbTjb1iSlUgmf//zncdtttx0yrDjbiFGQVqtVjsFsampCW1sbbDYbBgcH5YhKt9uNcDiMarWK1tZWXHbZZVBVFZs2bUIwGMT8+fNRrVaxYMECdHR0IJfLIZvNIpfLyb/n83kcOHAAU1NTCAaD6OnpQSgUgtPplIGAcdykkM1mMTw8jO3bt2NwcBDlchlerxd+vx9WqxWZTAa6rsufLdFoFA0NDahWqxgYGMDk5KQ8jqmpKeTzeTidThlGlMtlDA0NYXBwEIlEQo7BdLlcCAaDCIVCaGhoQDgcRiAQgNvthsvl4ghHIqprxxRSqKqKrVu3Yu7cuYf8/K5du7B48WLk8/kTdoC1pus67r77bnz+85+f0TmciIjqj/GkX1x5FR/TNE1elfV4PGhoaJCL/Hg8Pj2HHZBXZVVVhdlshtfrhdvtRiAQQLk8XUkRCARgs9kAAPPmzcOyZcswOTmJPXv2IJVKwev1wufzYf/+/RgcHITb7caKFStw0UUXobu7W14lpuN3Nq5JAGBychIf/ehH8dhjj6FQKLz1Hc4SFosFdrsdTqcTDocD2WwWTqcTPp8P1WoVqVQKmqahubkZTqcTHo8HmqZBVVU4nU6Uy2U4nU6sXLkSjY2N8v9hr9crA4NyuYxYLIZAIIB8Po9sNjvja2Cz2eByuWAymZDP5zE+Po7x8XHkcjkZlKiqimQyiUwmA5vNhnA4jKamJvnzyGw2o1qtYmpqCpOTkxgcHMTQ0BAymQx8Ph8aGhrg8/mQy+VgNpvR0NCAWCwGt9uNZDKJAwcOYGRkBIODg5icnJTVFhaLBaqqwu/3w+VyIRAIIBwOIxqNwuPxyGoOIqJ6cEyNM5ubm4+4INiyZQuamppOyIHVC5PJhL/4i7/A+9//ftx66624/fbbkUgkan1YREQ0iwgNNE2DzWaToYPFYpEl4qqqYsWKFVi6dCmSySR2796NyclJqKoKq9UKu90uKykKhQIcDoc8USmXy2hsbMS5554Ls9mMVCqF5uZmVKtVbNmyBZlMBi6XC3PmzMHu3bvx+uuvw2q14sILL8SaNWswd+5cXr08gc7GNQkAhEIh/PSnP8Xvfvc7fOxjH8Mrr7yCSqVS68OquWq1ilwuh3w+D5PJBJPJBE3TUCqV5HYJp9OJbDYLk8mEQqEgt1X4fD5UKhXkcjns2rULnZ2dsnoqn89D0zREIhF0dHQglUrB5XLB6/XKIEPTNORyOQwPD2N8fBzDw8NIJBKw2+3w+Xzw+XxIJBIYHh6G2WxGY2Mj5s6di46ODkSjUSiKglKphJGREUxOTiIej0PTNLhcLnR1deG8886D3W7H0NAQDhw4gFwuB5/PB0VRMDk5iaGhIXi9XjQ1NWH+/PmyAWw2m0U8HsfQ0BDGxsaQSCSQTqcxPj6O0dFRmEwmmM1m2Gw2+bMuHA4jEonIqguH48zZQkREp49jqqT45Cc/iWeffRYbN26EqqozPpfP53HeeefhkksuwTe/+c0TfqD1YnR0FB//+Mfx5JNPIpfL1fpwiIjOOGazWW7JeKvbiZN+m80Gp9OJTCYDRVFgtVqRTqehqiosFgs0TUM4HMayZcugqipGR0eRTqeRTCbh8XiQSqVQLpcBAHa7HXa7HbFYDCtXroTf74eqqliyZAkaGhrw5ptvYnh4GMViEZVKBSaTCT6fDxaLBXv37sW+ffsAACtXrsQVV1yBuXPnHrIMnN4erkmmqz0feugh3HLLLdizZw+OYUl31rBYLPLnhPiZIf4fFwGm2HpRLpeRzWbhcDjgdrvlybvX64XL5ZrxmKqqwmazwWQyyZBB0zRYrVYEAgEEg0FkMhmMj48jlUrBbrfD7/cjGAwiGAzC4XDIHhvVahW6rsPr9SIajaKlpQXhcPiQAUG1WpWVEvl8Hj6fDy6XC/l8HvF4HIqiyOoK4zHruo5sNotEIoF4PI7JyUlkMhkUCgVomoZyuYxisQhN01AoFGAymaCqKjweD4LBIKLRKCKRCDweD5xOJ6vBiOikOqaQYnR0FCtWrIDFYsEnPvEJzJ8/H8D0vs/vfOc7qFar2Lx5MxoaGk7aAdeLl19+GZ/4xCfw2muvyYUtERG9fYqiHHRl2GazyWoIh8OBarUqKyWsVqssZzabzYhGo5iampJ7v7PZLHw+H/x+PwDIUu/+/n6USiWUy2WUy2W43W40NDSgtbUVXV1duPjii6EoCorFImKxGCqVCjZt2oQdO3bAZDKhubkZ0WgUxWIRu3fvRn9/P6rVKnp7e3HNNdegp6eH5dMnEdckf1AqlfAv//Iv+M53voPx8fFaH05NHU3IKYILEVJYrVYoiiJ/9lSr1d83y3UgHA7LigIRCIjKCdHkslgswmQywWazwW63y61lqqqisbER8+bNQ3t7OxoaGlAoFDAxMYHR0VEkk0lUKhXYbDYoiiKDD5vNBrfbLRtnimoun88nq76sViumpqYwMDCAVCoFp9OJUCgEXdcxOjqKcrkMn8+HWCyGSCRyUKig6zoymQzi8TgSiYQ8FvHagenvq0wmg1QqhVwuh3K5DFVV4XA4EAwGEYlEEI1GZdUFw1giOlGOKaQAgP379+PGG2/EE088IRN7k8mEK6+8Et/5znfQ2dl5Ug60HlWrVaxbtw7/9E//hL179/IKBhHRCaCq0x30q9WqLEUWV/L27NmDUqkkG2Da7XaEQiEZRIiGlyMjIzCbzcjn81BVFQ0NDXC73WhubkYmk8GOHTug6zqCwSAKhYJsnrdw4UIEAgFcfPHFyOVymJycRHNzM3bu3Imnn34ayWQSnZ2dWL58OeLxON58802MjY2hVCqhra0Na9euxYoVK2C322v9Np4VuCaZaXx8HDfffDP+53/+B+l0utaHU1dEbxrj94n4t9j2YOxpI0JP0ZxTVGGIiUAOh0P2dhC9JPL5PDKZDCwWC9xuN5xOJ0qlEvL5vAwAbDYbvF6vDERbWloQCoWgqipMJhOKxSIKhQKKxeKMIKRarQKYDmxVVYXdbofb7YbX6wUApFIppNNpOByOGT0qcrkcVFVFc3MzYrHYYbdv6LqOdDo9I7TQNE02HhU/W0V1yOTkpOzJYbFYZIgjggtRdcGfhUR0PI45pBDi8Th27doFXdfR3d2NQCBwoo/ttJHL5XDrrbfirrvuOuuvYBARHS0xbQOYefXTZrOhWq3CZDLJLvSi+d3ExIQ8cahUKvD7/bBYLHKfud1ul5URpVIJNpsNCxculGFHPB5HKpWC2+1GS0sLRkZGEA6HceGFF2LJkiWw2WxYvnw5hoaGkMvlYDKZ8Mwzz2D79u0Ih8O45JJLUC6XsXnzZkxMTEDXdTQ0NOCiiy7C6tWrZ5RX06nDNclMr7zyCm6++Wa88sorKBaLtT6cuiK2fwGYUV0gJvyYTKYZP5dmhxoi0BCPZTabYTabYbVa4fF40NjYiJaWFthsthmfF9tMrFYrqtUqisUiisWiHC0qeuK43W74/X4ZfjgcDjgcDlgsFlQqFXk/YwUHADnlRDTltFgsaGxsRCAQkNs8TCYTgsGgHJ0qjkdUb4iJR4qiQNM0pFIpJBIJJBIJ2XjUarXC7/fD7/fLqhHRIHRiYgKJREKGKmLbjOhzEYlEEAwG4XQ6OdmIiI7ouEMKOtj+/fvxmc98Br/61a84HoyI6C2IcmuxbWP25zweD2KxmFzwlstlOWGjv78f5XIZiqLIkwFRkp1MJmWzu66uLiiKIk8QxHM6HA7k83m0tLTgXe96F+bPn498Po9oNIpkMomRkRHs378ffX190DQN55xzDoLBIDZv3ozx8XFZ3bFw4UKsWbPmjGzQSKc3TdPwwAMP4NZbb8WePXvkiTcd2uwtIOL9EuGF+CP+DWBGBa3YcmZsRqmqKrxeL4LBIAKBgBxZKnpdKIoyYzyyrusyuKhUKqhUKjK8sNvtcLlccLvdcLvdUFVV9mLRNA2apqFarcqtcCJcAKabrcZiMVitVqRSKZRKJRk2eDwe2XRYEMdvDDAURZFVHeKPxWKB0+mUoYro7ZHNZjExMSGDC1F1IUbCiq0pIrgQVRdsLExEAkOKk2D9+vX4m7/5G2zdupVXMIiIjkBcZZzdg8JkMqGrqwuRSAR79+5FNpuF2+2G2WxGNptFOp2WDeui0SgcDgfK5TLGxsbkCUYwGERbWxs8Hg+8Xi8ymQyy2Sz8fj/a2toQCARwzjnnoLGxUQYlw8PD2Llzp2yk2dLSAlVVMTAwgImJCbhcLkQiETQ3N+O8885Db28vG8hRXcvlcvinf/on3HnnnRgfH+fW1KMkwgbRV0a8b2IbmsVikQGpCBRE0GAcdyyIsFSMPBVbNUS/CbfbDbvdLivCjBUbs0MTALLiQYQVomrB4XDILRai/4UYg2o2m+Hz+WC321EoFFCpVKCqKiKRCFpaWtDY2DijykHXdZRKJdm7RzQHFe9DNpuVWz5ERYjT6UQ4HEYwGEQ4HIbX65XVI5lMZkZ4kUgk5AhXh8MBv98vqy6i0SiCweBBAQoRnR0YUpwklUoF3/ve9/C1r31NNlMjIqKD2e122Gw2lMtluWAVAYSu61AURXbXT6VSaGxsRDKZxNTUlGymKTrzixJju92O+fPnH9TErrW1FWvWrMHo6ChCoRDa2tpgNpuxdetW7N+/H6VSCT6fT5ZI79u3D8ViEX6/Hy0tLfD5fJg3bx5WrlzJrR10WhkYGMCnP/1p/OpXv2K/iuNkMpnk1X4x5lQ03FRVVQYRoupChBdiDKqodBD3BTAj8FBVVY43dbvdcLlc8o/oJWG8r3geRVHkxxwOh6yAEJNJ7Ha7bIKZz+dht9sRCASgqiomJiYwNjaGTCYDs9ksJ4yI5zQ27lRVVW63MwYoJpMJ+XweU1NTcoSq2B4imoCKRqAul0tWaZjNZpTLZaRSKaRSKSSTSaTTaeTzeflz3ev1IhwOzwgvuF2E6MzHkOIkSyaT+OIXv4gf/ehHckFNRETTxIJblElXKpWDOvO7XC709vbKiRxmsxlTU1MoFArI5/MApid2qKqKarUKn88nu+yL3hMtLS2IxWK48MIL0dfXh0qlgvb2doyPj+P555+HpmmYM2cOWltb0d/fj76+PuRyOTQ2NqK7uxsejwfRaBRLlixBc3MzF8h02lq/fj0++9nPYsuWLTIUpLfHYrHIgEKEFmKLhPi46HUhtmWIkZ8iyBCBAzDz56IYqWzc7iEqMZxOp9w2IraqiFBWhBfiMcTPRJPJhFwuh1KpBKfTiUgkgsbGRiiKIoMCTdNkY1BRlSEqJUSfDREGizHP4hjF9hDxPCKAyOfz0HUddrtd/rwW21VEiANAbnkRvTXEn1wuB2C6IsXlciEajSIajaKpqQnNzc0Ih8OsaiM6gzCkOEXeeOMN/O3f/i2ee+459qsgIjoC40IbmF6Utra2wu12Y3BwUO6nVhQFiURCVklEIhHE43G5aF66dCmWLVuGtrY2OBwO9PT0YHR0FPv27QMAHDhwAPv27UN7eztWrVqFkZERPP/885iamkIoFMLKlSsRjUahqio6Ozsxb948uagmOp2Vy2V873vfwze+8Q3s37//oO1W9PYZt4uIgECMDjWbzXJcp7HHhbEXRbValSGSCDfESbixGad4bLGNxOPxyIoM8Vzi56XYiiKqPSqVCsrlsqygaGxshMvlklNFRFghxo2KQETcd3Y1iTEsEU1HRehcLpeRzWaRy+VQLBblqFNRKREMBuH1euX2DrHVRGwzEdOWRJNOEX6IbdVWqxXBYFBOgmpsbJSvZ3ZjUPa+IKp/DClOIV3X8cgjj+Dv//7v8eabbx6yWRwR0dnObDbD6XQil8vJBa5Y0KuqimAwKLfQ5fN5JBKJGYvxYDCIrq4uXHPNNQiHw2hoaIDT6USlUsHjjz+OiYkJmEwmaJqGJUuWIBwO47e//S327NkDu92OlStXYtGiRbBYLPD7/eju7kZDQ0PN3g+ik2Vqagr/8A//gB//+MdyWg0dntjqMLvayzid6HCNNY23FZUN4qRZhAjGrRTi3+IxRUNM8V8RaIimmeJ2xioM8dgOhwNOp1NWRojARAQdlUoFuVwO1WoVLpdLBgbAdF8LMSnJ7/fD5/PJxp3iNYrjFwGAsfLD5XLJz4kQw2w2I51OY3JyElNTU5iampJVcaK3htfrRSAQkI1GRRgj3jMhm81iZGQEw8PDstdFPB6XQY+qqvB4PHIiiXgNsxuDHu7f4ncPEZ1aDClqoFAo4Otf/zpuv/12DA0NsV8FEdEsYo91uVyWHzObzfB4PJg7dy7GxsaQSqVgt9uRSCRQLpdhtVqxYsUK2XTz2muvxeLFi9Hf34/R0VE89dRTyOfzaG9vl4vRQqGA3bt3I5/Po7W1FZdccgmCwSDMZjMaGxsxd+7cGQtiojPR1q1b8dnPfvaMrvY0hgcn4rFEMCH+bqwaMFYTAH+YviG2dRzq8cR9ZocXwB/6Vohww1iRIfpciKoIUaUgGl4aG3kan0eECTabTfafEFvuAMiRqqLHhXGss5jo4ff75f3E8RhDEvGcIqRwOByyyaeYBuLxeOB0OqHrugwuJiYmkMlkUK1WYbVa5dYQMZ56NmOwIJqZZjIZOTpV9MkQ22sURYHH45FhiOj/IRqfGh1q0snh/s3fFUQnDkOKGhoeHsbnPvc5PPLII4jH47U+HCKiuib2RquqCl3XUSgUZEnz5OQkbDYbli5dCqvViksvvRQ33XQT/vu//1uGFHa7HS0tLRgaGkI6nYbFYkEymYTT6cTSpUvxjne8A9VqFXa7HfPmzUMwGKzxKyY6dXRdx8MPP4wvfelL2LZt2xlf7Wmsfng7RPhhbCQpPm4MK4zNgI0jQ42VEMagw1gVYay8MFYjiFBDURS55cO47UKEGKVSCYVCQf4RJ+vipLxSqcgtJCJ0EX8XIYax748IMkR/CBE2iC0ndrtdBs2iN4aiKDPCEvH+i6ajourCWEEBQB672F4iAgXj8ZRKJRSLRRnMiD+zv76FQgHpdFpOiMpkMigUCvJ9DAQCskmnqLhQFEUGNIebdGL8XjhUiHG4QIO9jYgOjyFFHXj++efx+c9/Hhs3bpSNgYiIaCbRq0L8V1EU2fBtfHwchUIBra2tuPTSSxGLxWSZcjQaRblcxujoKIaHh2E2m2XgEYvFcMEFF6C5uRn5fB4tLS3o7OzknmU6a+XzeXzta1/D9773PRw4cOCM3wIigoATEVgc7vGNPSREgABAhgjGppniJN4YIMweaTo7RJgdVoiTYuM2DFVVZ2z1sNlsyOfzsqmlCC+KxaLcSlIul1Eul2U/DGOlh2gGauy9IUIK8VwOh0P+EfcTE0LEaxXvhfE1is+Lkaai2sK4BUZs4xAVGX6/H16vd0bzzNnhwqGCDFG9kUgkkEgkZMNOETK53W45VaShoQGxWEyOvRbVfsbnOVygMbsCRFSfvFWVhvg7Aw062zCkqBPVahV33303/vVf/xW7du06ab8siYhOZ6KEWNM02O12RCIR5PN52Gw2JBIJuFwurFy5Eg6HA729vVi0aBG2bNmCN954QwYbolR56dKl6OnpgclkgtPpxPz58+U+bKKzXX9/P77whS/g0UcfPeuqPU/k1pDZjyvCBdG/QQQXAA4KJUQYISoRzGYzisWiDGCNlRfGxzc2yxRbLGw2m3ws8TlVVeH1euHxeGC322dUXhSLReTzeZRKJeTzeVmBIEILYGawYOwfZGzmKY7bbDbLbRsul0v+XVReiNsYm4Pqui5P7o1NP429OkRwIrZuNDU1ob29Hc3NzfD5fEd1cm/cHiNe7+joKEZGRjA2NoaJiQlMTk7KC4kivPD7/QiHw4hGo4hEIrLXhfGP6KMxuxHooUIN8fdDNbI9lkCDU07oTMCQos6kUil86Utfwj333IOxsbFaHw4RUd0Ri1ex6BVX+RwOB0qlEubPn48FCxYgHA4jk8nIaR6KoiAQCGDRokWYP3++3Bvd3t6OtrY2LuyIDmH9+vX44he/iJdeeumM3wJyOMbeEif6cUWwIE62jdM/xKQPsTUDgKyWEJUKJpMJxWIR2WwW+Xx+xkhTADO2Vxi3I4jxpVarVVY0iPGgXq8XXq8Xfr8fiqIgn8/LLSK5XA6ZTEZ+L4jtEOVyWY4cFX0xjGNUjY1BxceM4YQIL8RoVRFyiD4P1WpVVm+IcER8PcQWlGKxKKtSrFYr/H4/mpqa0NbWhubmZrS1tck+GsdaLafrOjKZDA4cOIChoSEZYCQSCbn1QzQLFUG4z+eD0+mc0U/EGFwcKswQ22LeKtAQ/zb2bRKMvUveqo8GqwapXjGkqFPbtm3DLbfcgieffFJ2PCYiomnGEm1xNdLY/Kyzs1NexSuVSnA4HOjo6MCqVatgMpng8XjQ1NSE+fPnw+Vy1frlENW1arWKH/zgB/j3f/937N69u9aHUzPGXgonevk8uzrAeGJrt9vlzztxpd14tV00pRRTN6xWK6rVqty6UCgUUC6XZQ8MYGZ4YexpYTx5FRUZPp9P/hHPkc/nMTExgVQqJasyPB6PnBYiGldmMhl5vMZjMDYZFa/beDzihF28NtGzQmwnMY4qNX49jM8l+lmUSiX5mKL6oampCc3NzYhGowiHwzJQEFtVjnZ7RalUwuTkJMbGxmZMFxHTUsTvJhH8GCejHGobiGjU+VZhhuhpcTQVGsZAY/b3rTEc46QTqicMKeqYruv4+c9/jr//+7/H1q1bz/h9oUREx8O4cBILq0gkAovFApfLhfb2dixduhTz5s1DLpdDe3s7ent7EYvFuM+X6BjE43H8wz/8A+655x4kk8laH85JIa5kv1XVxOypHid6i4gIKwDICgLjFXLxMXHiL7ZhiEaVFotFBhdiBGm1WkUul5NbN2ZXXsxu+il+norHM54kO51OBAIBeDwemEwmlEolGShYrVZZaVGpVORJsnGbi5hEUigUkMlkkMvlZpy0ixBFHM/sAEP0qxCjVUUFhnh8Y4PPSqWCQqGAfD6PXC6HQqEg31dRweH1emVY4fP55PjqcDgst6eIrTNvdbJerVaRSqUwPj6OsbExObFEBBdie0ooFEIwGITP54PH44GiKAdtPRF9NGZXTBh7WhwpzDBu/9B1/aAeGkcKODjphGqJIcVpIJ/P4+tf/zq+9a1vcQsIEdEsYiEvFrdi4en1etHZ2YmrrroKoVAIqqpi+fLl6O3tlVeyiOjYvf766/jMZz6D9evXH3L//OlOnKQfTVhh3CYmwgrxGEcTXhxtuCGOybg9ZPaED1HlIU7QRdWCsbmlOLEXDS0ByBNhMfEin8/L8MMYXBi3ahi3cYjQQgQEHo8HoVAIDQ0N8udzpVKRUzXy+TwURZF9HURjzWq1inQ6jVQqhVwuh1wuh3w+L49HnGSLAMN4bGLLiJgsYtw+IgIMcR8R1uTzeWSzWVlpIrYRil4eYtuJy+WCz+eD3+9HIBBAMBhENBpFMBiU76MIMQ5XhaHrOnK5nKy6mJiYwMTEBNLpNIrFogyVxISRSCSCSCQip6aIionDNQA1fmz295OxIudIYYYxfBH9UQ7XN4OTTuhkq2lI0dHRgf379x/08Ztuugnf+c53anBE9W1gYACf/vSn8eijj561+0KJiI6GyWSCqqro7OzEpZdeilWrVuHiiy9GQ0NDrQ+N6hTXJMdG0zQ89NBD+NznPif7vpxJjKNFjVUGb3Uf431nP45xq8Wh7muszDgS49YQ43jT2VtFxN+NDTnFfwHIE0Rjc01RKVCtVuW0j0wmg2w2K/tSiPdDhAXG4xBBiuhxIQILsd1BVAyIio5qtSqPQRzz7AoBs9mMcrkst4+IKhBRGVEsFqFp2oygQRyfeA9E6GCcPCK2j9hsNlltIUIa0d9CfE2NjTzFCb2owBDTRfx+P0KhkHy9ImAwhhizqzCKxSJSqZQMLsbGxpBMJpHP56FpGlRVhdvtlqNRo9EofD4fXC7XIftJGMOFtwozDrfV5K3CjEOFCscSaHDSCR2NmoYU4+PjM75Rt27dissvvxzPPPMMLr744lodVt178sknceONN2Jg3x7MaQqgaA3AbHMc9vZaKQ+1HIfZbELO4ofZ5oBWysONDDJwH/G+R+NEPhYR0Ykgroapqoo/++D7cfP//bjcw0x0KFyTHJ9sNotbbrkF37/tv9AW8Zz0NYnxZF/8e3aIcLjHEydDxivq4qRW/F3cRlQJ6LqOQqEgRxeLfgdia4UICADICULiJEqc0BvL5EWDR1EFYewVIBoAezyeGeNAAcwIL0QoIKZGiPdhOpx1wKoogAkoFAqy0sLpdMLtdsNkMsHhcCAejyORSKBSqcir3na7HWaTGQ7n9Am83W5HuVxGIBBAsViECSbANP01Hx8fRyqVggkmaLom3xNRvWE2m1GpVGRgMmfOHPk+O51OtLS0IBAIIB6PI5PJyGlNbrcbqVQKmUwGPp9PTm9SVRWlYgnhSBjlchmhUAilUgm5XA7JZBITExMwm81yAof4eCqVmtGLobe3F9Ax3aujWIDJZILP60NbexvsNju8Pq9sDioCEYfDgWg0CrvdDl3X5chSEZSUSiUEAgHY7XZUq1W0trYiEolAVVVEo1FUK3/4ueJwOrBgwYIjfo+LSo9kMim/TslkUvbXaG1thdPhxNzuufD5fEd8rCMRwY4IECrlCsqV8syPVSqolGdVS5mAxsZGNDc3v63nNY65Nf57Rg+TSvWg+ztdzumv42lMsZjhdnBLzJHU1XYPUSWwc+dOpmRv4Wuf/TAWjP0cFzeXsH5IxeOmy3Dph79y0O1+fcctsGz/BdbO0XFRm4Jf7wPu2d+AD3UnsSZWOOJ9j8av77gFV+lPn5DHIiI6WS5b0cIFAR0TrkmO3jf+7qNYMPow3tlUrOmaRDwH1yVEVO+4Ljmyupm3ViqVsG7dOvzlX/4lFwNv4bWNL2HZ1KN4d0cZTqsJa9uLuFJ/Gvv6Xptxu319r2H+xON4dxdweZcVTqsJ13Sb8OdtQ2iy545436Oxr+81XKU/jbXtxbf9WEREJ1OlemJHB9KZjWuSo/faxpewePznuLKtVNM1iXgOrkuI6HTAdcmR1U1I8bOf/QyJRAJ/8Rd/UetDqXuvbHgSFzbMHEt6cayAwTc2zPjYwNbnYKrmcWHrzD1rl3RasHm4esT7Ho2Brc9hTazwlsdBRER0OuGa5OjVy5pEPAfXJUREp7+6CSnuuOMOrF27FrFYrNaHUvdWrroCL4zO3LP57JCKloWrZnysddE7oVsceGFg5n6uZ/ZWsaLJcsT7Ho3WRe/E+qGZHfKP97GIiIjqBdckR69e1iTiObguISI6/dVFSLF//3489dRT+Ku/+qtaH8ppYdm55+P1yLX45T4rcmUdj+234wnTZejoWTbjdh09y7A9fBUe2w08uauMXFnHIzt13N0fw1DRecT7Ho2OnmV43HQZHttvf9uPRUR0MimWuvh1R6cBrkmOjViTPDmg1nRNIp6D6xIiOh1wXXJkddE489Zbb8Xtt9+OgYGBg0bz0OG9tvEl/PJnD2DPWBavbe3D1NTUIW8nOmlbzCZkOd3jrCFmjs8e9XS2O9RM+qOdU0/1w2w2y074gUAATU1N6O3txYIFCzB37lzEYjFYrdMNqdhFm44F1yTH57WNL2HDrx9DUfHh+ZdewZtvvnnIcekne00inqPW6xKz2Qyr1QqLxSJ/F4vpFgBgsVjgdrsBTE8BURQFixYtQjAYxPj4OJxOJ6LRKCwWC+LxOBwOBwYGBtDT0wPVriKeiKNUKmHhwoVobW0FAIyMjCAcCmNu91w4nU709/djYmJCjukcHx/Heeedh56enhnjK/P5PHbs2IGdO3eit7cXCxcuPOJrS6VSGBgYQKFQQCQSQXNz84zHGxkZwYYNG7B3717Y7XbEYjG8973vlROWNE3D8PAw3nzzTYyNjcHlcqGnpwednZ3YsWMHLBYL5s+fL0etHi1N05BOp+U0jHKpDB06hoaGpqd/mMxIZ9KYmJiAyWRCW1sbVqxYgUgkgtHRUfT392N0dBSFQgE2mx2hUBDtbe1obWuF3++H0+mUz5XNZrF3716USiW0tLQgEokcsX9NuVyeMTEkm80COmC2mOF2u+FyueB2u+F2u4/5ddPph+uSt1bzkELTNHR2duIDH/gAvvrVr9byUE57hUIBP//5z/HNb34TmzdvRqFQeOs70RlNLI6ITndizJ/H40FTUxM6OjqwbNkyXHjhhVi0aBFisRhPKOlt45rkxNq7dy+++93v4r777sPQ0FCtD6dmFEWB3W6XgXixWJRjU8U4UV3XYbPZ0NraKsMJj8eD5uZmGTIUCgW4XC68733vw/DwMB599FH4fD4sWbIEF110EUwmEw4cOAC/34+uri7MmTMHY2Nj2LFjBzZu3Ih8Po+GhgYsWrQI559/PkwmEyYnJ7Ft2zYMDw8jHA7jne98pwx4ZysUCti9ezfGx8fh8/nQ3d0tgxYAGBsbw1NPPYVXX30VVqsVq1evRktLCzKZDN7xjnegUChg3759eO211zA2NoZAIIAVK1agp6cHZrMZW7duRTKZxDnnnANVVQ95DIc6psnJSTkWVNM0Oep1dHQUr7/+OqampqCqKjRNg6qq6OzsRFtbG4rFIvbu3SvHD3u9XsRiMSxYsAAtLS0IhUKw2+0znk/TNOzfvx/9/f3weDzo6emZEV4A0+NixQhR8UesyVVVhc/ng9frhc/ng8vlYnNeokOoeUjx5JNP4sorr8T27dsxb968Wh7KGUXTNGzduhW33XYbHn74YYyMjNT6kIgOiVUMBBz8fWAymWCxWOBwOBAOh9He3o6lS5dizZo1WL58OWKxmLwqR3SicE1y8mQyGTzyyCO47bbb8PLLL5+1F1IsFosMAcRFhEqlIn/+Wa1W2O122Gw2WK1WOBwORKNRebV9bGwMq1atwp/8yZ/glVdewUsvvQRN0+DxeHDFFVfA4/Egn89DVVU0Njait7cXAPCLX/wCExMTqFQqyGQyuOCCC9De3o49e/bAarWiUChg4cKFaGxsPOiYNU3DwMAA+vv7YbFY0NXVhYaGBvn5RCKBDRs24OWXX0axWMSyZctwxRVXIBQKYfPmzahUKrDb7di2bRsmJycRiUSwbNkydHV1yXC5v78fe/bsweLFixEKhQ77/um6jlQqJYOJbDYLk8kEh8OBXC6HqakpjI+PY3x8HJOTk7BarfB4PNA0DXa7HaqqIp/PI51Ow2w2IxQKob29HfPnz0dDQwMCgcCMqhCjbDaLN998E9lsFh0dHWhra4PJZEK1WkU6nZaBRCqVQqVSgclkgtvtnhFKzA49iOjQah5S0KkxOjqKn/zkJ7jrrruwZcuWQ5ZgEtWC2WyGpnEM09lGXDky/goym81QFAU+nw9tbW1YunQpVq1ahYsuughtbW1HfWWNiOpbpVLB5s2bcccdd+CRRx7B8PBwrQ+pJkRZv8VigaZpMrSwWq1QFAWKokDXdVSrVXnCazab4XQ6ceGFF2LhwoWIRqPIZrPYunUrMpkMFi9ejIaGBkSjUTgcDrjdbixatAibNm1CNBpFOp3GCy+8gDfeeANz5szBNddcg0wmA1VVsWzZsoOOcWpqCjt37kShUEBLSwva29tlsJDL5fDb3/4WmzZtwtTUFLq6unDppZdizpw5MJlMGB0dxc9+9jNUKhVYrVZEIhEsWbIEHR0dMyrf4vE4tmzZgra2NnR2dh50DJVKBVNTU5icnMTU1BTK5TKsVitUVUU6Pb19Y3R0FOVyGX6/HwCwb98+xONxVCoVAJABhaqqiEajstIkEonA4/EcsZpB13UMDg5i7969cDgc6OzshK7rMpTIZDLQdR2Kosgwwuv1wuv1HjbwIKIjY0hxFsrn83jmmWdw11134emnnz5sLwsiohPpUNUSZrNZXvFbuHAhVq9ejVWrVmHBggXweDw1PFoiOhV0XUd/fz8eeOAB3H///ejr60OxWKz1YdWc2WyGxWKB0+mEzWZDNpuFpmkytLDZbPD5fAiFQli8eDHmzp2LUqmEVColQ46lS5eira0NmqYhl8vhggsugNvtxrPPPounn34amUwGXV1daG9vx9q1a+FyueTz5/N57N69GxMTEwgEAuju7pbbGkqlEjZu3IhNmzZhfHwcDQ0NOP/887Fo0SJYrVYMDw+jv78fO3fuxPbt27F8+XKsXLkS7e3tB20lKRaL2LRpE9xuN5YsWSLDglwuJ6slkskkdF2Hy+WC2WxGIpHA2NgYxsfHAUD2xbDZbHj22WdlRYfX60VDQwMaGhrQ2tqKOXPmoK2tDeFw+KhD73w+j82bN2N4eBgejwcej0de6BNbN0Qowa0bRCcOQ4qzXLVaxRtvvIF7770XP/vZz7Bv3z6ZOhOdabi1pPZMJpMsv+3o6MDq1auxevVqnHfeeRz3SERIpVJ44okncPfdd+P5559HIpGo9SHVnMVigcVigaIosFqtKJVK0DRN/jxVFEU24mxoaEBjYyNMJhNyuRx6enpw7rnnYmJiApdddpkMO7q7u/Hkk0/iiSeewNy5c/Gud70LK1asgM1mQ39/P/r7+2G1WjF37lxEIhEA0xUNW7ZswcaNGzE6OgqPx4PFixdj0aJF8Hq9GBwclE0qy+WyrJK8/vrrDxkKaJqG1157DcViEStWrJgRTOTzeZjNZvh8PllJceDAASQSCVgsFjQ1NaGhoQFerxfDw8N44YUX8OKLLyKTySAWi2Hx4sVYsGABuru70dzcjEAgcFR9i6rVKlKpFFKpFHbt2oW+vj6YTCa0t7ejsbFxRijBrRtEJw9DCppheHgY//3f/40f/ehH+N3vfjfdfZiI6G0wm82w2WwIh8NYsmQJrrzySqxZswaLFi1iKSwRHValUsHLL7+MdevW4bHHHsPg4CCbQf+eqLQwmUyIRqPQdV2u2VRVhdPphK7rv59UYUM0GsXChQtxwQUX4IMf/CBcLhc2btyIjRs3yu0kdrtd9mVobW1FW1ubbMC9fft2vPzyyxgbG4PNZkNXVxfmzZuHQCCAkZERTExMIJVKQdM0eL1etLW1IZ/Po1qtYsWKFYd8Ddu2bUNfXx+am5tRKpVQrVZht9vh8/mQzWYxMTGBgYEB5HI52O12NDY2IhQKwel0YmJiAnv37sXWrVuxf/9+JBIJxGIxXHnllbjooovQ1NQEr9f7lpUNxWJR9pEQWzdKpRKGhoZQqVTQ2dmJJUuWHLFXBRGdeAwp6LByuRyefvpp3H333diwYQPGx8fZO4CI3pLYxuF0OtHc3IxVq1bh2muvxerVq+H1emt9eER0GtJ1Hfv378f999+PBx98EDt27EAul6v1YdUNh8Mhqw4AwOv1wu12I5PJYHR0VDav9Pv96O7uxrJly+Dz+bBq1Sokk0lZBQEAa9aswcUXXwyr1Yq9e/fipZdewsjICGw2G2KxGJqbm+H3+5FKpZBOp1EsFlGtVuFyudDU1IT29nY4HA689NJLCAaD6O7ulseZyWQwOTmJnTt3YuvWrWhubkZXVxecTidSqRRGRkYwODiIUqkEl8uFUCiEQCAAm82GyclJjI6OYnh4GAMDA0gkEtB1HaFQCBdffDGuvvrqI/6OEUGOMZQQDVwdDge8Xq+cCuJwONDT03PEJp5EdPIwpKCjUq1W8frrr+Ouu+7CY489hn379qFcLtf6sIioThhHhM6bNw9XXXUVrr/+eixYsIAz34nohEsmk3jkkUdw33334aWXXpInrGc7k8kkg2KXy4VwOAyz2YxCoYBSqYRCoYBKpQKLxYJgMCgDAIfDgXPOOQednZ2YmpqSTYzL5TJsNhsikQh8Ph8cDgd0XUe5XIbJZJJNLKPRKDo6OmTfinK5jOeffx7z5s2D3W6X2ziKxSLK5bIceRoIBDAwMIDh4WHoug63241gMCgbYMbjcUxMTCCTySCVSiGRSCCXyyEQCMjnXLFiBTo7Ow+qmhBbN0QoYZy64fF4ZJNLn88Hs9mMnTt3YnR0FJFIBPPmzTvsKFYiOvkYUtAx03Udw8PDuO+++/DTn/4UW7du5bYQorOQ2WyG1WpFKBTCihUr8Ed/9Ee45pprEI1Ga31oRHQWKZfLeOmll/CDH/wAzzzzjCzVp2lmsxl2ux1OpxMulwvlchnpdBrAdAWBaPioaZqcIJLL5eBwONDb24sLL7wQwWAQoVAIHo8HNpsNxWIRuq4jEomgo6NjRtPNYrGInTt34uWXX0Zra6scpRoKhVAqlfDcc89hZGREPpfX64Xf74fX64Wu64jH40gmk0in09A0TR5PuVxGMBiUAYKqqujt7UU4HJbPaxwDapy6YRwD6vF4ZmzdiMfj6OvrQ7VaRXd394zxqkRUGwwp6G3LZrP45S9/ibvvvhu//e1vMTU1xasZRGcosdiNxWJYvXo1rr32WlxxxRVwOBy1PjQiIui6jt27d+Ouu+7CL37xCzk+k2YS/SzEBJFAIIBisYh8Pi8bdBaLRVQqFTidTrS3t6O7uxvRaBTNzc3o6elBT08PYrEYrFYr0um0rJbIZDIYGRlBoVDAJZdcglwuh6GhIezbtw87duxAJpPBwoULEYlE4Pf7UalUZLiQy+Wg67rcrlIoFJBMJuFyudDd3Y2WlhYMDw9DURR0dnbK+6ZSqRlbN4yhhNPpPGRvimq1ij179uDAgQMIBALo6elhM0yiOsGQgk6oarWKTZs24Qc/+AGefvppDA4OclsI0WlOlA13dHTg8ssvx3XXXYdzzz33qEe4ERHVSjwex8MPP4x7770Xr776KpLJZK0Pqa7ZbDbY7XbY7XY5OaRUKqFUKsFisSAQCCAUCiEcDiMcDsPv9yMQCKChoQHhcBixWAyNjY149dVXMTw8LKeMANMBUiaTwZIlS2C325HJZGRfCLPZDL/fj7a2Nng8Htl3wmq1Yv78+ejt7cXQ0BDeeOMNKIoiqyfE1g1jKGGz2d7ydaZSKfT19aFQKKCrqwuxWIzjQ4nqCEMKOmnE7PO7774bDz/8MJtcEZ1GxOi3rq4uvOc978F73vMe9Pb2MpggotNWsVjE+vXrcccdd+C5557D6OgoG4Ifhslkgs1mk6NPBYvFArvdDpvNJrf8ud1ueL1eOBwOOBwOmM1mZLNZNDc3o7u7G42NjfB4PNi5cydUVYXdbkexWISiKIhEIpgzZw7a2tpQKpWwZcsW7Nu3DxaLBS0tLWhqakK5XEZfXx/i8ThisRh6enrk9hCv13tMfY80TcP+/fvR398Pj8eDnp4e2UeDiOoHQwo6ZZLJJH72s5/h3nvvxebNmxGPx2t9SERkYLFY4PP5MG/ePFx77bVYu3Yty1+J6IykaRp27NiB2267DY8//jj27t2LUqlU68OqS2JbiK7rMJlMss+Dqqpy2wjwh7GoNpsNiqKgsbERgUAALpcLmUwGTqcT8+fPR1dXF3p7e9Hb2wuPx4MDBw7ghRdewO7du6FpGpqamtDU1CS3aYyNjUFRFCxfvhzt7e3HXfGQzWbR19eHTCaDjo4OtLW1sXqCqE4xpKCaqFQqeP7553H77bdj/fr1GBkZ4dUMohoQDcV6enpwzTXX4Oqrr0Z3dzcrJojorDIxMYF7770XDz74ILZu3YpMJlPrQzptmM1mWVUh/q7rugwrTCYT3G432tvbZXVEIBBAoVDA2NgY0uk0vF4vuru7sXz5cjQ3N8Pn86FarWLnzp2wWq1YtGjRjOacx0LXdQwODmLv3r2y2abH4znB7wIRnUgMKajmdF3Hrl278P3vfx+PPPII9uzZw6sZRCeRoihyQfjud78b1157LebPn89ggogIQKFQwBNPPIHvfe97ePnllzExMVHrQzqtiJDCYrHI0MJmsx1UeWG1WmUfit7eXnR1dSEYDCISiaBarSKRSCAWi2HlypXHPQ60UCigr68PiUQCra2t6Ozs5FhsotMAQwqqO/F4HD/60Y+wbt06bNmyhX0siE4Ai8UCj8eD9vZ2rF27Ftdddx2WLl3KYIKI6Ag0TcOrr76K2267Db/61a8wODiIarVa68M67YhgwrhdRIxFVVUVbrcbgUAAgUBANu9sa2tDT08PvF4vIpEIGhoa5AhRh8Pxlls1hoeHsWvXLlitVtnHgohODwwpqK4Vi0X86le/wve//31s2LABU1NTtT4kotOGyWSC0+lEU1MTLrvsMrzvfe/DBRdcwHGhRETHaXBwEHfeeSceeughbN++nZWfx8FkMsmwwmw2y3+LKgtFUeByuRAMBuH3+9HQ0IDGxkZ4vV6Ew2FEIhH4fD40NDQgGo3C6/XC4/HI0L1UKmH79u2YnJxEU1MTurq6ZjT/JKL6x5CCThuapuH111/Hd7/7XTz22GMYGBio9SER1SW73Y6Ghgacd955+OAHP4grr7yS3cuJiE6wbDaLhx56CHfeeSdeeeUVZLPZWh9SXTOZTDO2WohwQtM06Lou/1gsFgCQAYaYKOJ0OuF2u+Hz+dDc3CxHoAaDQUSjUYTDYdhsNqTTabjdbixbtgyxWKxWL5eI3gaGFHTaGh4exg9/+EP86Ec/wvbt21l+SWc1RVEQCASwcOFCvO9978MHP/hB+Hy+Wh8WEdFZoVKp4LnnnsN3v/tdPP3006z8nEUEDuLv4vRDfEzTNNm/Qny8Wq3K282+vQguVFWFy+WC3++HqqpwOBwIBoPo7u5GLBZDY2MjWltb0dDQICsujre/BRGdOgwp6IyQy+Xwk5/8BN///vfx0ksvoVwu1/qQiE4Jl8uFrq4uvPvd78ZNN92E1tbWWh8SEdFZTdd17NixA7fffjt+/OMfY2hoqNaHdMzEdoyT9RizKyqMtxeBhhhpKm4vHqtarcoAQ9M0eT+TyQSbzQaHwyGrLgKBAEKhEDweD4LBIBobG9HS0oLOzk60trYiEAjA7XZzOwhRnWFIQWecarWK9evX4//9v/+HJ598EoVCodaHRHRCWSwWNDc3Y82aNbjppptw/vnnc9Y7EVGdmpycxD333IPvfOc72L17d60P55gYAwNd12dUNxwtsa3D+HvKWP1q/LwIG0SIYRxPbzaboSgKFEWR/SvEMWmaJreNiL8bgw4RXjgcDrjdbgSDQXg8Hvj9fkQiETQ3N6OzsxMdHR1obm5GKBSSAQkRnXo1DSmq1SpuvfVWrFu3DiMjI4jFYviLv/gLfOELX+CCm06Yvr4+fP3rX8f999/P/aJ0WvN4PFixYgU+9KEP4QMf+ABsNlutD4nojME1CZ0KxWIRDz/8MP7t3/4Nr776aq0P55gpijJjK8bxnEaIk39jTwoRMIjPi7BC3O5Qz2Mcc2q322Gz2WSAoes6KpUKyuUyKpUKKpXKjMcwm82wWq2w2Wyw2WxwOp3weDzwer3w+XwIhUKIxWJoaWlBa2srmpubZeUFfx4QnXw1DSm+8pWv4D/+4z9w9913Y+HChdi0aRM+9KEP4ctf/jJuvvnmWh0WncEmJyfx1a9+Fd///veRTCblx5taOuBwuk7Y82ilPNzIIAM3zLZDT1I40m2O5v6nQr0cx9kqn8tieHAf2tracMMNN+Dmm29GS0tLrQ+L6IzENQmdapqm4Te/+Q0+//nP48UXX5xRNXA6rEtMJhOsViscDgfK5TIKhYIMF2ZXQIhgQlQ9AJBVEYqiyAoIY6WGuI/xsbRSDo5qCnmLFyarY8bnLRbLjLHaonLCYrZAURRouiYDi3K5jGKxeFDYYrFM39bv98PpdCIUCslxqW63GzabDW1tbWhra8M555xznF+N05diMcPtYE8POvlqGlJcc801aGhowB133CE/dv3118PhcGDdunW1Oiw6S+TzeXz1q1/F/Q/+N/79B784YY/76ztuwVX601gTK2D9kIrHTZfh0g9/5ahvczT3PxXq5TjOdkp6F666/FKWnRKdZFyTUK299tpr+MQnPoH+A6P41ronT9jjcl1CJ9JlK1oYVNBJZ37rm5w8F154IZ5++mns2LEDAPC73/0OGzZswNq1a2t5WHSWcDgc+Md//Ee8vHHTCXvMfX2v4Sr9aaxtL8JpNWFtexFX6k9jX99rR3Wbo7n/qVAvx0HARatWM6AgOgW4JqFaW7ZsGTZs2IAtr289YY/JdQmdaJWq9tY3InqbahpSfO5zn8P73/9+9PT0wGq1Yvny5fj0pz+NP/uzP6vlYREdt4Gtz2FNbGajzotjBQy+seGobnM09z8V6uU4iIhOFa5J6EzEdQkRnY5qGlI8+OCDuO+++/CjH/0Imzdvxt13342vfe1ruPvuu2t5WETHrXXRO7F+SJ3xsWeHVLQsXHVUtzma+58K9XIcRESnCtckVC+Mky/eLq5LiOh0VNOQ4m//9m/llYvFixfjf//v/43PfOYz+Jd/+ZdaHhadZUw4cWVrHT3L8LjpMjy2345cWcdj++14wnQZOnqWHdVtjub+p0K9HAdNN6kiopOPaxKqFzu3952wx+K6hE40rkvoVFBq+eS5XE7OQRYsFsuMLr5EJ5tLtcJdPoD/fvhneOWVV5BMJlGpVOSMbQAHja0SnzPO9BZ0XcejZS88JgVZkxsm6zb88sXr5H1Ef4HHy0G49CzyFg+gbMeTG6+XM72fLIbhqKaRNblgsvbh6c3vg81mg8ViQbFYlPPBxeitSqWCarUKm80Gs9mMYrEoO2ZXq1X4fD5Uq1Wk02mUy+XpDtVWG3RMd9O22+yoVCvyNRWLRQQCAXxlbD7+7fUpVB1BQJlE31c+A03TsKB3AX635XcoFotwqA54fV6UiiWYzNPHPz4+Dq/XC7PZDLfbjZGREQQCAQBAMpGEokz/6HF73EjEEyhXKnA4VFQqFaiqinw+DxNMKFcqsFim32+LxQITTEhn0lBVB0qlouzILUaiVSoV+TUS/y6XywAAq3W6yZNx1JnoJi7mv+u6jnK5PGP+OvCHMWniva5UKvI5DjcWzfg9IX6mzf4+MRJz3BsbG7F69Wq85z3vQffcLjanIjpFuCahejF3Ths2PfQwnvzVU9i/fz/y+bz8fTV7/KWu6/J32KE+Jz72P2UfXFCQ0Z2w2Lfhsd9eK28nfic/XgrChSyycMJsO3hdolZSyJpcsNi345nX3g+73Q5N0+TvTbG+sVgsKJVKMJvNsNlsqFarKJfLcDgcqFQqsFqt8Hg8yGVzSGfSsFgs8Hq808eM6d+NNpsNhUIBqqqiWCzCbDbD4/HgyyPz8e9b48iZ3PAEM9j2z59CIBCAqqrYsWMHrFYr3G43rFYrisUiVFVFIpFAtVqFy+WCy+VCJp1BqVxCJBLB+Pg43G43CoUCGhsbMTY2hlwuJyeXQAdUVUVDYwMSiQScTicsFguSyaR8vW1tbdA0DYODg1BVFTabDT09PYhEImhpacGKFSug6zr27NmDbDaLtrY2RCIRANPrg/HxcYyMjKBSqSAcDqOpqaluR4xzugedKjUNKd7znvfgy1/+Mtra2rBw4UK8+uqr+I//+A/85V/+ZS0Pi84yiqLgsotX4bKLV+HAgQNYt24dHnzwQezcuRP5fF6e+B7uhFT8Ahefmz03XIzDEmOuxH//MNs7IX8Zi+cSJ6y6nkK5XJYn2g6HA8FgEIlEApVKBYqiQFVVWK1W5PN5VKtVWK1WmM1mlMtlKIqCUqkEp9OJuXPnolJIY9/evdB1HS6XS4YaJpMJdrsd+XweDQ0NGBgYgKIoaGhoQF5XkRwehq7rcLvdiMfjeMc5S5FLTeLNN99EJBJBc3MzxsfHEY1GMT4+jkQigXA4DJPJBLfbjQMHDsCid2BkZAS5XA42mw3RaBQH+nchkZh+/Q6HY3pEmKYhnU7L91JRFFitVui6jng8DqvVCkVRkM1moes6VFWFw+FAOp2WiwpN01AoFFAul2EymeByueRCqlQqyUWSCCgAyGDKGD6IvyuKIr8O4uPi+Iy3r1arM+a+Hw2bzYZgMIjzzjsPf/mXf4krrrgCDgfHvRKdalyTUL0Ih8P4xI0fwcc/8iH85je/wQ9/+EP85je/wdjYmLwocShiLSLCNeM4UPF7ymazyb/rui7XHeL+iqLA6XRC16dP1sXvRfG7V9NScnSnxWJBKBSCyWRCKpUCMP07Tawt8vm8vK9xtGilUkFLSwtCoRASE8MYHx+H3W6Hy+WSgYnVakW5XIaqqrBYLBgfH0c4HIbb7cZQqoxqdQLpTAqlUgmJQADvete78MT/bEO5XMaCBQvkOiIQCGDH9u1wu90AgPb2duzZswdOpxNaKYt9+/ZBVVVEIhFMmavY+tprAKbXbmazGT6fDwldB6p5jI+PIxKJYGxsDD6fD7quw+fzITk5gj179sh10tKlSzFvTisaGhqwcuVKFItFvPHGGzCZTDh3+SJ4vV5Uq1UMDQ1hYGAA5XIZjY2NaG9vnzFClehsVtMRpOl0Gl/84hfx8MMPY2xsDLFYDB/4wAfw93//93WbINLZoVwu45lnnsGdd96J5557DuPj4yiVSke8jwgrjCenxivxIjwAgFKpNON2iqLA4/HAYrEgnU7LK/XixL1YLMqTa4vFgqamJpRKJWQyGbnocLvdKJVK8jgtFosML8rlMpqamjB//nzs378fQ0NDcmHidDphtVrl1Y6GhgbE43GMjY2hsbERDQ0N2LdvH5xOJ1RVxcjICN71rndhcHAQr7zyCsLhMJxOJ4rFImKxGHbt2gWr1Qqfz4dSqSSv4oiKikqlgkAggGAwiG3btsFischZ5na7HePj4wfNUDebzZiampJhSjabRbVahd1uRyQSkVdJrFYrKpUKMpmMfA8cDgc0TUOpVEKhUJBfD1GJIuayGwML8TlRiSLuJz5nfAyx0DuWH6Vmsxkulwtz587F+973Pvyf//N/EIvFjvr+RHTicU1C9WxoaAgPPPAAfvzjH6Ovr0/+HjwUY9AOYEZYISiKIisBRAWlcc1it9vh9/uRy+WQzWYBTAcH0wGGjnw+L9cbLpcLoVAI8Xhcrl+MF1DEMYh1UCqVgqqqmDdvHgKBAF577TV5ocZut8PhcMh1k8ViQSwWwxtvvAFd17Fs2TLs27cPmqZNBwTJ6erMq6++Gj/96U+RTCaxYMECDA0NIRgMwmKxYN++fZg7dy7Gx8cRDAYxMTGBaDSKZDKJTCYDp9OJVatW4be//S2Gh4ehqircbresihAXOsTrdTgcaG9vx+Dg4HTQkkhA0zRUKhW8853vxKWXXgpVVbF8+XLE43Hs2rULPp8PCxYsgMViwYEDBzAwMIBKpcJwgugwahpSEJ0OBgYGsG7dOjz00EPYvXv3ERcGwuzKCiNx4m2xWJDP5w9aNIRCIZRKJaRSKflLXZRIlstlFItFWR3gcrlQKpXkx5xOpww1CoWCDC/Gx8dhNpvR2tqKpqYmDA4OwmQyYWhoCJqmweFwyKsknZ2dKBQK2LlzJ6xWKxobG6FpGlRVRaFQQDqdRmNjI0qlEiYmJuDxeJBKpWCz2eD3+zExMQGbzYaGhgYMDw/DbrfLcs2pqSn4fD6oqopSqYSRkREoigKv1wu73Y5qtYpEIiEDH7FIyefzKBQKsjJE0zQoioLm5mbkcjmk09PlouVyWZZsi9ej6zqKxaKsghABhLiyI947sZgT99U07aBFm/F2YtvIsbDb7YhGo1izZg1uuukmnHvuufKqERER0VupVCp45plncNddd2H9+vWYmJhAsVh8y/sdbl0i1iMilDf+3gOmLzB4PB7E43GUy+XfV4DaZYWhsarC7/cDgKxaFKGGqG6sVCrw+/3I5/PIZDLweDzo7e1FJpNBoVDAxMQE0uk07Ha7fA6Px4Oenh5s2rQJ6XQa0WgUVqsVqqpC13WkUikoigK/349EIoFMJoNgMIjR0VHEYjG5bmlvb0cymUSxWITNZoOqqhgfHwcwXV0RDAbxyiuvIJfLweVyyXWJWGfl83nYbDY4nU6cf/75eP3111EqlWCz2WC325FKpXDOOefg+uuvh67rWLx4MYaGhjA6OoqWlhZ0dHTIyolKpYKmpia0tbUxnCA6DIYUREepVCrhySefxF133YUXX3zxqKsrjrRFRCwOjFs6gOkKgEAggEwmI/d6ip4K1WoVpVJJlnSKqxXValVuN/B6vSgUCshms3C73VAUBfF4HBaLBY2NjQCAnp4e7NixAwcOHJA9H2w2G9ra2tDc3IyXXnpJPl57ezt6e3vlx1RVRSaTQWNjI9LpNNLpNEqlEoLBIMrlMpxOJ2w2GxKJhNyeIl5jS0sLhoaGoKoqJicnZdWFeG2ZTEa+bxaLBS6XCxMTEzPeH1FBYbPZMDw8jGq1ikqlIntxWK1W2ZNC7IUFIK/uiK9npfKHPhwmk+n3W2z0GZUuYmFn7HNxLEwmEzweDxYsWIAPfehDuOGGG+TrJSIiOl4DAwO477778JOf/AQ7d+5EOp0+pqo+I7EF1WazyVBhdlhhtVqRTCYBQF4IAIBisSh/n4oLExaLBYVCQa4ZXC6XrDgIh8My9FBVFbFYDJqmYc6cOXjhhReQy+UAQFZVXHLJJXj11VcxNjYmt3ledNFFmJqawrZt2xCJRBCPx2UvrEKhgMnJSbjdbrhcLhSLRTQ1NWF0dFRWV4o1ktfrlVtW4/E4stksfD4f3G43mpubMTIyIvuBBINBLFq0CMViEb/73e/g9/tlmOHxeHDDDTfA4XBg3rx5GBwcRD6fx9y5c1EulxlOEB0jhhREx2Hv3r24++678fDDD2P37t3I5XJva2EgAgJjXwoA8oqFaBolFgTipFyEHcbKAVFxIco0RaPMSqUiT74dDgdaWlrQ0tKC119/HVNTU7JSwe12Y+HChejv70cymUSpVEIoFMKHP/xhPPTQQxgdHZXH29HRgd27d8Nut2NiYgIulwtWqxWtra3o7++H3W5Ha2sr3njjDXn8oomV0+mUfSScTqd8nWIfqWh+lc1mMTU1Jbd9eDweuN1umM1mxONx5HI5uV3DWCEhSi/F10VcmRHvsbGZpWiUaby9eP/FbY51JJyqqohGo1i7di0+/elPY968eQc15SMiInq7SqUSnn76adx555144YUXjuoiyuEY1xQmk0k27RREs8xqtSovtBh/54o+Tn/oraXLHllOp1Nuz3A4HPL3sej9sGLFChw4cAD79++XgYvNZkNLSwv8fj/6+/vlumH16tVobm7GL37xC3lho6mpCYVCAZlMBsViUVZdii2yU1NT6OzsRDwex/j4uKzUUFVVrnfy+TwcDgfmzJkjq1rF2qO7uxuhUAgbN25EqVRCd3c3vF4vbDYbent7sWTJErS0tGB4eBgmkwnBYBBTU1OoVqtobGxkOEF0DLhiJjoOnZ2duPXWW/Hiiy/innvuwXvf+140NjYe10moqIwQ2xREsyZgekuB2BIith8YgwmxKBB7JcXnFUVBOp2WDbLEFhWbzQar1YpMJoN9+/Yhk8nA7XYjGAzKk/t8Po8tW7ZA0zQ4nU4A03u19+zZIyd0ZLNZeDweua1ENOLMZDJyukg+n5dbWkTAkMvlpid3mEyyukFMyxDNMMvlsmy6Ja6GiG0fXq8XXq8XuVwOU1NTSKVSsu+FqHQQ/TvE9g4A8Hg8MJvNsuLDuE1DTOwwbgcx9psQi7GjYTKZ4PP58M53vhN33XUX+vr6cNttt6Gnp4cBBRERnRQ2mw1r167Fgw8+iA0bNuCWW27BsmXLZLPIYyHWFcViEcVicUbFBAC5fVL8Hq9Wq/L3r6g8FJWd4raimXcmk4HNZpvRI8rhcKBarWJiYgJbt26Fw+FAIBCA0+mUFw8OHDiAvXv3wul0ytvv2bMHAOD1epHNZlGpVODz+WasIwAgn8/L6s9qtQqn04lcLge73S63iIpjERWcVqtVXuTRdR12ux3hcBiKoqC/vx+ZTAaLFi2SlRTz58+Hz+eDx+PBvn375PGIhuLnn38+5s2bx4CC6BiwkoLoBNB1HTt27MCdd96JRx99FHv37pXlisfKbDbDYrEc1HXbSIwjLZVK8gRaVCqIhYJYOFgsFnmlQgQP4uqF2+2G2+2WnbNF0yoRmIiTe1GyKPpK5HI5BAIBWf0htlmk02lZXjk1NSWvaphMJsTjcRQKBTl6TPSZEN2xTSYTxsfHZYMqscCpVqvTI1N/v4dULCjElhfj+NDZPSJE89FCoSA7nRuJMEI4nlGDory1qakJf/zHf4xPfvKTaGtrO2gUHBER0alSLBbx2GOP4a677sLLL7+MsbGx4/odJwIIsZY41GMYR40aKxKNzTnFx4yTPkTfLUVR5LojFApBURQUCgUkk0l5+1KpBFVVZYNOh8OBzs5ODA0NyeqKhoYGGS6I4ygWi4hGo7LReCwWw/j4OFRVxfDwMBRFQTAYhNlsRiKRkNtJQqEQLBYLMpkMotEo/H4/IpEIdu7ciaamJtnv4vLLL8fAwADa29tlc/NoNIpYLIa2tjbY7fa38VUkOnsxpCA6wbLZLH7xi1/grrvuwquvvoqJiYnj2goiQgdjRYCR6KFQLpdnLBpEN+zZIzHFYxmnkLhcLnmVQ1Q0hEIhZDIZpNNpOBwO+Hw+2WdC9Iswzi2PxWIYHBycsX1CjA5zuVyy6mNqakr2yPD7/bIcVfSvEJUgIhQRWzDsdvuM4EVUYojnmz1Zw9jPQryuQzUvnT169FiJqonly5fjr//6r3H55ZdzMUJERHVF13Xs2rULd9xxBx555BHs3r37qBptzibWEQBm9IgyEhdKjJ8XAQaAGWsSYHq7q6jWENWeYtuqqGaw2+0IBoNynSECBQCyYaWYwpHJZBCJRFCtVjE1NSUrOERfClGdoaqq7KcFTDfOLBaLGBsbg6IoiEQicLlcssLC5/Nh0aJFcvpaMBhEOp3Ge9/7XlQqFbz++utoa2tDR0cHenp6GE4QnQAMKYhOEl3X8cYbb+D222/HE088gf379x/XHlGxL/Rw9xVXKYxbEkSfi9mdvN1utyy/rFQqcLlcCAQCmJyclGGIy+VCLBbDjh07AExvlRCNpjKZjGx6JbZNeDwe5HI5+bjiqojo9q2qKjRNk9s2GhoaMDk5KRdJgUAAuq4jk8nIyhAxOlQ03jIGIKLC5FA/uoxVFYdqWqooilwsiSqNYyECmNbWVtxwww346Ec/ivb29mN6DCIiolrIZrP4+c9/jjvvvBObN29GPB4/5osookJT9KE4FPF5I7GWEf2jRBAh+lEVCgWYzWY0NjYil8shl8vNGLueSCRkICEqQMUWEqfTiUAgICeIiUpMsU4R08NEiOJwOJBOp5HL5RAMBuF2uzE2NoZyuQyv1yv7VuTzeTidTixduhThcBj5fB7FYhHJZBKXXHIJnE4nXn/9dfh8PnzgAx9Ad3c3wwmiE4QhBdEpkEwm8dBDD+Guu+7Cli1bZHp/LA51hcJo9qLAuN1A/N1kMsHv98NisWBsbAwAZLVEPB6XPS3EZBFN0+S2D0VRkEgkUK1W5X0sFgsSiQQ8Ho/8nGiWBUCGFpVKBblcTv6SFzPXxRWLRCKBbDYLTdPkyFWbzYapqSl5DOIKzuHCGlEdcrj3RjQPFaPRjoVo6rVy5Up86lOfwqWXXir7dRAREZ1OxEWU2267Db/85S8xMDBw2LXF4YgKidkjS4VDXSgQHzNWRzqdTvj9ftnAGwCi0SgqlYoMUVRVlVtAgD9M/dA0DYlEQo46DQQCSCaTqFarcLlcmJycnDFiVdM0WVkxNTUFh8OB1tZWOeVM13WEQiF0dnZi//79KJfLWLBgAebPny+3m27ZsgULFixAKBQCAPj9flx99dXo7e09ni8FER0GQwqiU6harWLz5s34r//6Lzz11FNyfOaxEM2ejmeLAgC5XzKXyyGdTssQIBgMYnJyErquQ1EUuWVCbMVwuVzI5XLyKoiiKGhoaMDw8LCcSS5uL65GpFIpAJC9LkTzTFHpIRpminGgooICADKZjDxmq9V6xB4dR+LxeABAXlU5FmIk6w033IA///M/R1dXFxtgEhHRGSORSODBBx/EPffcg9dee01eRDhaorLicCHH4UaxGzkcDoTD4RlrIqfTKZuHi1BDfLxUKsn1RjqdlmuWSCQif9dbrVakUilZQWq1WuV2UkH0p1AUBT6fD5OTk4hEIohGo0ilUtA0DYsXL8aCBQtgt9vx05/+FIFAAAsWLEBLSwsuv/xyJBIJrFixAl6v95jeNyI6MoYURDUyPj6OdevW4d5778Wbb74prxAcLREeHA+LxQKXyyV/WZfLZVgsFnklwXi1w+/3I5vNym0Uolv47L4RYkyX0+mUJZKiwaVouiXKLcUViVKpJKsuRHPP2UGEaNp1rGGO0+mU+1GPZf+t2WyG3+/Hueeei4997GO4+OKL5VQTIiKiM5Gmadi0aRO+/e1v46mnnsLo6OgxXQwxjvI+Hg6HA2azWW771DQNqqrKtZFYl9jtdnmRQ0zx8Hq9sjGmqPAol8soFApwu90oFAoHbfUU1ZCiWafb7UZ/fz+8Xi9aW1thMpmQTCZxzjnnYO7cuXj00UeRz+dx/vnn4x3veAeuvfZaJBIJDA0N4cILL2SzbKITjCEFUY1VKhX85je/wbe//W08++yziMfjx3T/I21zOBJRzQDgoOacovGluJ042TduF6lWq3KLhggSRNAgFiti4gcA2azKarXK8WNib2i5XD7sVRhjn4mjfT/EiLJjCX7sdjs6Ojrwx3/8x7j++uuxePFiuaghIiI6W4yPj+Oee+7B3XffjR07dhxT0H80lROHY2ywOXv7qmj+LW4nbmO322XFZDablZO/xDpGXIAR/bLEiHKr1Qqr1YpoNAq32w2n04ldu3YhFAohFArJkeYLFizA5OQkhoaGMGfOHPzJn/wJrrvuOjgcDrzyyitwOBxYsGDBcb1eIjo8hhREdWRgYADf//73cd9996G/v/+4r0gcC1EmaVwQiGaXsys1TCaT7I4tRpWK/aiiWaamaVAUBXa7Xc5aF/d1OBzQdV02sxJVF4cLWY51sSMaVh3LgioYDOKCCy7A+9//fqxevRotLS3c0kFERGe9crmMZ599Ft/61rewfv16uYXzZBITwTRNm7E28Pv9SCaTB60JrFYrbDabDC6MFZzAHxqH2+12uN1u2bPCYrHA6/Wiu7sbTqcTe/fuRalUwrnnnovt27dD13W0t7fDbDYjmUxCVVX86Z/+Kf7qr/5KBiDPP/885s+fj6amppP+vhCdbRhSENUhMd/8P//zP7Fp0ybkcrmT/pyHqsiw2WzQdX1GlYPoRyEqJ8R9Zt9XjBEzcjqdskQTwIxRom/32I+0J3Y2i8WC7u5uvPe978UVV1yBlStXwu/3v+3jICIiOhP19/fjtttuw7p162SjyZNJjCo1Po+olJzdN0OMKhfVD4KxClNsExGjzm02GyKRCObNm4f+/n5MTU2hubkZwWAQ+/btg9PpxJw5czA0NASTyYQ1a9bgy1/+MqxWKwBgbGwM27ZtwwUXXMCJHkQnAUMKojrX19eHb3zjG/jpT3+KycnJU/rcovphdkhi3IIhSikBHHSFwxh8iHDjeOaznyhutxurV6/G1VdfjVWrVqGnp4dbOoiIiI5SoVDAI488gq9//et49dVXj7s31vES40SNF0GM00KAQ18kAabXIU6nU4YaPp8PbrcbBw4cgKIomDdvHqrVKnK5nAxIVFVFJBLBP/zDP2DFihXysfr6+pBOp3Huueee/BdNdBZiSEF0mkin07j//vvxH//xH9i5c+dJv4ohvJ39pcaO3Mfa+PJE6ujowHvf+16sXr0aK1euRFtbG7d0EBERHScxxvTf//3f8Ytf/AKJROKUPK+xQeXh1iaHW7coijJja0axWITFYoHVaoXJZEJvby/y+Tzi8TjK5TKamprg8/nQ2dmJm266CV1dXfK+L7zwAhoaGmZ8jIhOHIYURKcZXdfx4osv4ktf+hKeeeaZY54K8naIKR1H2yvjeJt6ngiKouD888/H1VdfjXe84x1YtmwZp3QQERGdYKlUCj/84Q/xn//5n+jv7z/uCxvHQzTyPhptbW0oFApIpVJwOp2w2WwIh8Ow2+0YHx+H1+uVDb8DgQAuvvhiWCwWuN1uXHLJJbjooosATI9I37RpE5YuXcp1BdFJwpCC6DQ2NjaGr33ta7jzzjsxMTFxSp7z7Yw+PRXcbjeuuOIKXHHFFTjnnHOwaNEi7hclIiI6yTRNw3PPPYcvfOELePHFF09J8+9juRji8XjkVhGHw4FoNIrFixdj586dGBoagtlshtvtRmNjI4LBIK666ipEo1FomoZkMok//dM/RTAYxMDAAPbu3YtVq1axKpPoJGFIQXQGqFQq+MlPfoJ//Md/RF9fX60PpyYaGxtx9dVX4/LLL8c555yDOXPmcG45ERFRDQwPD+PWW2/F/fffL0eR1wur1YrW1lZ0dnZiYmIClUoFO3fuhMPhQFtbG7q7u+FwOHDZZZehu7sbixcvxgMPPIBYLIZrrrkGW7ZsgclkwpIlS2r9UojOWIz/iM4AiqLg/e9/P95880309fXhqquugsViqfVhnRIdHR34v//3/+LOO+/Ev/7rv+KGG25AV1cXAwoiIqIaaWpqwu23346pqSl897vfRUtLS60PSapUKkin0xgZGUE2m8X4+Diq1Sp0XUcgEICmaQgEAqhWq4hGo/D5fDj//POxZ88ebN++HclkEsFgsNYvg+iMVtOQIp1O49Of/jTa29vhcDhw4YUXYuPGjbU8JKLT3vz58/HLX/4SqVQKn/3sZ+F2u2t9SCec2WzG3Llz8alPfQr3338/vva1r+Gqq65CKBSq9aER0WmM6xKiE0tRFHz84x/HwMAANm7ciPPOO6+mx2OxWGA2m5HP57Fv3z5MTEwglUrBbDbDarViamoKyWQSbrcbU1NTiEajAIAFCxagqakJzz//PCYnJxlSEJ1kNd3uccMNN2Dr1q347ne/i1gshnXr1uEb3/gGtm3bhubm5lodFtEZRdd1PPTQQ/jkJz+JsbEx+fGmlg44nK4aHtmxyeeyGDmwH+3t7fijP/ojfOQjH8H8+fO5H5SIThiuS4hOvkQigZtuugk//vGPZT+JU7UmEZM8qtUqLBYLdF2HpmlwOBzw+XwolUqwWq0477zzYLfb8eEPf1heANm9ezdefvllWK1WvHvtu+H1uuF2WE/6MROdjWoWUuTzeXg8Hvz85z/H1VdfLT++cuVKrF27Fv/8z/9ci8MiOqP19fXhuuuuQypbxHcf+HWtD+eYPf/oD/HJmz6K1tbWWh8KEZ1huC4hOrU0TcNXvvIVfO+Ou/GtdU/W+nCOy2UrWhhUEJ0ENbsEWalUUK1WoarqjI87HA5s2LChRkdFdGbr6elBX18ftr15ejbXvOXzX2BAQUQnBdclRKeW2WzGF77wBWx5fWutD+W4Vaq1GbNOdKarWUjh8XhwwQUX4Etf+hKGhoZQrVaxbt06/Pa3v8Xw8HCtDouIiIjOQlyXEBER1Yeabua+9957oes6mpubYbfb8c1vfhMf+MAHuMeciIiITjmuS4iIiGqvpr91u7q6sH79emQyGQwMDODll19GuVzGnDlzanlYREREdBbiuoSIiKj26uLSgMvlQlNTE+LxOJ544glce+21tT4kojOaYqmL//WP2el63ER0euG6hOjUOZ1/t5/Ox05Uz2o6gvSJJ56AruuYP38+du3ahb/927+Fqqp47rnnYLWyUy7RyZTJl0+rhk+KxcwO2kR0UnFdQlQbp9uaBOC6hOhkUmr55MlkEn/3d3+HwcFBBINBXH/99fjyl7/MhQDRKcBfrEREM3FdQlQbXJMQkVFNKymIiIiIiIiIiARupCIiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoLDCmIiIiIiIiIqC4wpCAiIiIiIiKiusCQgoiIiIiIiIjqAkMKIiIiIiIiIqoL/z+z7fsBjpIJjgAAAABJRU5ErkJggg==", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from circuitree.viz import plot_complexity\n", "\n", "# Plotting options\n", "plot_kwargs = dict(\n", " tree=tree,\n", " aspect=1.5,\n", " alpha=0.25,\n", " n_to_highlight=10, # number of top states to highlight\n", " highlight_min_visits=10, # only highlight states with 10+ visits\n", ")\n", "min_visits_per_move = 10\n", "\n", "## Plot\n", "fig = plt.figure(figsize=(13, 5))\n", "plt.suptitle(\"Search space for the Bistability game\")\n", "\n", "ax1 = fig.add_subplot(1, 2, 1)\n", "plt.title(\"All moves\")\n", "plot_complexity(fig=fig, ax=ax1, **plot_kwargs)\n", "\n", "ax2 = fig.add_subplot(1, 2, 2)\n", "plt.title(f\"Moves with {min_visits_per_move}+ visits\")\n", "plot_complexity(vlim=(min_visits_per_move, None), fig=fig, ax=ax2, **plot_kwargs)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "\n", "In a complexity layout, terminal topologies are arranged into layers based on their complexity, or the number of interactions in the circuit diagram. The width of the layer represents the number of topologies with that complexity, and topologies within a layer are sorted from most visited to least visited during the search. A line from a less complex topology $s_i$ to a more complex one $s_j$ indicates that the assembly move $s_i \\rightarrow s_j$ was visited at least once (left) or at least ten times (right). Finally, we use orange circles to highlight the top 10 topologies shown above.\n", "\n", "The graph on the left shows that the overall space is quite well sampled. In all the layers, even the least-visited states (on the right of each layer) have many incoming and outgoing edges, showing that many options were explored. If we only look at the moves with 10+ visits, the graph on the right shows that the search favored a subset of the overall graph that has a higher concentration of top solutions. This is great! It means that our search struck a good balance between exploring the overall space and focusing samples on high-reward areas." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Animating the search\n", "\n", "To make a video of the search process, we will re-run the search, this time saving the tree object every 1,000 steps. To do that, we'll create a *callback* function that saves the tree to file. A callback is a function that is passed as an input to another function. If you supply the `callback` and `callback_every` arguments, `search_mcts()` will call your callback periodically during search. We can use callbacks to perform periodic backups, save progress metrics, or end the search early if a stopping condition is reached." ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "MCTS search: 0%| | 206/50001 [00:00<00:24, 2055.06it/s]" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Starting MCTS search with 50001 iterations.\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "MCTS search: 100%|██████████| 50001/50001 [01:21<00:00, 614.11it/s]" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Search complete!\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "\n" ] } ], "source": [ "# # Remember to delete the backup folder before re-running this cell!\n", "# # Otherwise, the video may contain multiple runs\n", "# !rm -r ./tree-backups\n", "\n", "from pathlib import Path\n", "from datetime import datetime\n", "\n", "today = datetime.now().strftime(\"%y%m%d\")\n", "\n", "# Make a folder for backups\n", "save_dir = Path(\"./tree-backups\")\n", "save_dir.mkdir(exist_ok=True)\n", "\n", "## Callbacks should have the following call signature: \n", "## callback(tree, iteration, selection_path, simulated_node, reward)\n", "## We only need the first two arguments to do a backup.\n", "def save_tree_callback(tree: BistabilityTree, iteration: int, *args, **kwargs):\n", " \"\"\"Saves the BistabilityTree to two files, a `.gml` file containing the \n", " graph and a `.json` file with the other object attributes.\"\"\"\n", " gml_file = save_dir.joinpath(f\"{today}_bistability_search_{iteration}.gml\")\n", " json_file = save_dir.joinpath(f\"{today}_bistability_search_{iteration}.json\")\n", " tree.to_file(gml_file, json_file)\n", "\n", "# Redo the search with periodic backup\n", "n_steps = 50_001\n", "tree = BistabilityTree(grammar=grammar, root=\"ABC::\")\n", "tree.search_mcts(\n", " n_steps=n_steps,\n", " progress_bar=True,\n", " run_kwargs={\"expensive\": False},\n", " callback=save_tree_callback,\n", " callback_every=500, \n", " callback_before_start=False,\n", ")\n", "print(\"Search complete!\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "\n", "Then, we can make the video using `matplotlib`'s `animation` interface. This might take a few minutes to run." ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "1 / 101\n", "2 / 101\n", "3 / 101\n", "4 / 101\n", "5 / 101\n", "6 / 101\n", "7 / 101\n", "8 / 101\n", "9 / 101\n", "10 / 101\n", "11 / 101\n", "12 / 101\n", "13 / 101\n", "14 / 101\n", "15 / 101\n", "16 / 101\n", "17 / 101\n", "18 / 101\n", "19 / 101\n", "20 / 101\n", "21 / 101\n", "22 / 101\n", "23 / 101\n", "24 / 101\n", "25 / 101\n", "26 / 101\n", "27 / 101\n", "28 / 101\n", "29 / 101\n", "30 / 101\n", "31 / 101\n", "32 / 101\n", "33 / 101\n", "34 / 101\n", "35 / 101\n", "36 / 101\n", "37 / 101\n", "38 / 101\n", "39 / 101\n", "40 / 101\n", "41 / 101\n", "42 / 101\n", "43 / 101\n", "44 / 101\n", "45 / 101\n", "46 / 101\n", "47 / 101\n", "48 / 101\n", "49 / 101\n", "50 / 101\n", "51 / 101\n", "52 / 101\n", "53 / 101\n", "54 / 101\n", "55 / 101\n", "56 / 101\n", "57 / 101\n", "58 / 101\n", "59 / 101\n", "60 / 101\n", "61 / 101\n", "62 / 101\n", "63 / 101\n", "64 / 101\n", "65 / 101\n", "66 / 101\n", "67 / 101\n", "68 / 101\n", "69 / 101\n", "70 / 101\n", "71 / 101\n", "72 / 101\n", "73 / 101\n", "74 / 101\n", "75 / 101\n", "76 / 101\n", "77 / 101\n", "78 / 101\n", "79 / 101\n", "80 / 101\n", "81 / 101\n", "82 / 101\n", "83 / 101\n", "84 / 101\n", "85 / 101\n", "86 / 101\n", "87 / 101\n", "88 / 101\n", "89 / 101\n", "90 / 101\n", "91 / 101\n", "92 / 101\n", "93 / 101\n", "94 / 101\n", "95 / 101\n", "96 / 101\n", "97 / 101\n", "98 / 101\n", "99 / 101\n", "100 / 101\n", "101 / 101\n", "Saved to: animations/240513_bistability.mp4\n" ] } ], "source": [ "from matplotlib.animation import FuncAnimation\n", "\n", "# Load the saved data in order of iteration\n", "gml_files = sorted(save_dir.glob(\"*.gml\"), key=lambda f: int(f.stem.split(\"_\")[-1]))\n", "json_files = sorted(save_dir.glob(\"*.json\"), key=lambda f: int(f.stem.split(\"_\")[-1]))\n", "iterations = [int(f.stem.split(\"_\")[-1]) for f in gml_files]\n", "\n", "# Make an animation from each saved time-point\n", "anim_dir = Path(\"./animations\")\n", "anim_dir.mkdir(exist_ok=True)\n", "\n", "fig = plt.figure(figsize=(13, 5))\n", "ax1 = fig.add_subplot(1, 2, 1)\n", "ax1.set_title(\"All moves\")\n", "ax2 = fig.add_subplot(1, 2, 2)\n", "ax2.set_title(\"Moves with 10+ visits\")\n", "\n", "def render_frame(f: int):\n", " \"\"\"Render frame `f` of the animation.\"\"\"\n", " ax1.clear()\n", " ax2.clear()\n", " \n", " tree = BistabilityTree.from_file(\n", " gml_files[f], json_files[f], grammar_cls=SimpleNetworkGrammar\n", " )\n", " \n", " plt.suptitle(f\"Iteration {iterations[f]}\")\n", " ax1.set_title(\"All moves\")\n", " ax2.set_title(\"Moves with 10+ visits\")\n", " plot_complexity(fig=fig, ax=ax1, tree=tree, aspect=1.5, alpha=0.25)\n", " plot_complexity(\n", " fig=fig, \n", " ax=ax2, \n", " tree=tree, \n", " aspect=1.5, \n", " alpha=0.25, \n", " vlim=(10, None),\n", " )\n", "\n", "# Make the animation\n", "anim = FuncAnimation(fig, render_frame, frames=len(gml_files))\n", "anim_file = anim_dir.joinpath(f\"{today}_bistability.mp4\")\n", "\n", "# Save the animation\n", "anim.save(\n", " anim_file, \n", " writer=\"ffmpeg\", \n", " fps=10, \n", " progress_callback=lambda i, n: print(f\"{i + 1} / {n}\")\n", ")\n", "print(f\"Saved to: {anim_file}\")\n", "\n", "plt.close(fig)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "\n", "Now let's watch the video!" ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
" ], "text/plain": [ "" ] }, "execution_count": 12, "metadata": {}, "output_type": "execute_result" } ], "source": [ "import moviepy.editor\n", "\n", "moviepy.editor.ipython_display(str(anim_file))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "\n", "---" ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Python implementation: CPython\n", "Python version : 3.10.8\n", "IPython version : 8.24.0\n", "\n", "circuitree: 0.11.1\n", "numpy : 1.26.4\n", "matplotlib: 3.8.4\n", "tqdm : 4.66.4\n", "jupyterlab: 4.1.8\n", "ipympl : 0.9.4\n", "ffmpeg : 1.4\n", "moviepy : 1.0.3\n", "watermark : 2.4.3\n", "\n" ] } ], "source": [ "%load_ext watermark\n", "%watermark -v -p circuitree,numpy,matplotlib,tqdm,jupyterlab,ipympl,ffmpeg,moviepy,watermark" ] }, { "cell_type": "markdown", "metadata": {}, "source": [] } ], "metadata": { "colab": { "provenance": [] }, "kernelspec": { "display_name": ".venv", "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": "3.10.8" }, "orig_nbformat": 4 }, "nbformat": 4, "nbformat_minor": 0 }