{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Using a custom step method for sampling from locally conjugate posterior distributions" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Introduction" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Sampling methods based on Monte Carlo are extremely widely used in Bayesian inference, and PyMC3 uses a powerful version of Hamiltonian Monte Carlo (HMC) to efficiently sample from posterior distributions over many hundreds or thousands of parameters. HMC is a generic inference algorithm in the sense that you do not need to assume specific prior distributions (like an inverse-Gamma prior on the conditional variance of a regression model) or likelihood functions. In general, the product of a prior and likelihood will not easily be integrated in closed form, so we can't derive the form of the posterior with pen and paper. HMC is widely regarded as a major improvement over previous Markov chain Monte Carlo (MCMC) algorithms because it uses gradients of the model's log posterior density to make informed proposals in parameter space.\n", "\n", "However, these gradient computations can often be expensive for models with especially complicated functional dependencies between variables and observed data. When this is the case, we may wish to find a faster sampling scheme by making use of additional structure in some portions of the model. When a number of variables within the model are *conjugate*, the conditional posterior--that is, the posterior distribution holding all other model variables fixed--can often be sampled from very easily. This suggests using a HMC-within-Gibbs step in which we alternate between using cheap conjugate sampling for variables when possible, and using more expensive HMC for the rest. \n", "\n", "Generally, it is not advisable to pick *any* alternative sampling method and use it to replace HMC. This combination often yields much worse performance in terms of *effective* sampling rates, even if the individual samples are drawn much more rapidly. In this notebook, we show how to implement a conjugate sampling scheme in PyMC3 and compare it against a full-HMC (or, in this case, NUTS) approach. For this case, we find that using conjugate sampling can dramatically speed up computations for a Dirichlet-multinomial model." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Probabilistic model" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "To keep this notebook simple, we'll consider a relatively simple hierarchical model defined for $N$ observations of a vector of counts across $J$ outcomes::\n", "\n", "$$\\tau \\sim Exp(\\lambda)$$\n", "$$\\mathbf{p}_i \\sim Dir(\\tau )$$\n", "$$\\mathbf{x}_i \\sim Multinomial(\\mathbf{p}_i)$$\n", "\n", "The index $i\\in\\{1,...,N\\}$ represents the observation while $j\\in \\{1...,J\\}$ indexes the outcome. The variable $\\tau$ is a scalar concentration while $\\mathbf{p}_i$ is a $J$-vector of probabilities drawn from a Dirichlet prior with entries $(\\tau, \\tau, ..., \\tau)$. With fixed $\\tau$ and observed data $x$, we know that $\\mathbf{p}$ has a [closed-form posterior distribution](https://en.wikipedia.org/wiki/Dirichlet_distribution#Conjugate_to_categorical/multinomial), meaning that we can easily sample from it. Our sampling scheme will alternate between using the No-U-Turn sampler (NUTS) on $\\tau$ and drawing from this known conditional posterior distribution for $\\mathbf{p}_i$. We will assume a fixed value for $\\lambda$." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Implementing a custom step method" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Adding a conjugate sampler as part of our compound sampling approach is straightforward: we define a new step method that examines the current state of the Markov chain approximation and modifies it by adding samples drawn from the conjugate posterior." ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import arviz as az\n", "import matplotlib.pyplot as plt\n", "import numpy as np\n", "import pymc3 as pm\n", "\n", "from pymc3.distributions.transforms import stick_breaking\n", "from pymc3.model import modelcontext\n", "from pymc3.step_methods.arraystep import BlockedStep" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "RANDOM_SEED = 8927\n", "np.random.seed(RANDOM_SEED)\n", "az.style.use(\"arviz-darkgrid\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "First, we need a method for sampling from a Dirichlet distribution. The built in numpy.random.dirichlet can only handle 2D input arrays, and we might like to generalize beyond this in the future. Thus, I have created a function for sampling from a Dirichlet distribution with parameter array c by representing it as a normalized sum of Gamma random variables. More detail about this is given [here](https://en.wikipedia.org/wiki/Dirichlet_distribution#Gamma_distribution)." ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "def sample_dirichlet(c):\n", " \"\"\"\n", " Samples Dirichlet random variables which sum to 1 along their last axis.\n", " \"\"\"\n", " gamma = np.random.gamma(c)\n", " p = gamma / gamma.sum(axis=-1, keepdims=True)\n", " return p" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Next, we define the step object used to replace NUTS for part of the computation. It must have a step method that receives a dict called point containing the current state of the Markov chain. We'll modify it in place.\n", "\n", "There is an extra complication here as PyMC3 does not track the state of the Dirichlet random variable in the form $\\mathbf{p}=(p_1, p_2 ,..., p_J)$ with the constraint $\\sum_j p_j = 1$. Rather, it uses an inverse stick breaking transformation of the variable which is easier to use with NUTS. This transformation removes the constraint that all entries must sum to 1 and are positive." ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "class ConjugateStep(BlockedStep):\n", " def __init__(self, var, counts: np.ndarray, concentration):\n", " self.vars = [var]\n", " self.counts = counts\n", " self.name = var.name\n", " self.conc_prior = concentration\n", "\n", " def step(self, point: dict):\n", " # Since our concentration parameter is going to be log-transformed\n", " # in point, we invert that transformation so that we\n", " # can get conc_posterior = conc_prior + counts\n", " conc_posterior = np.exp(point[self.conc_prior.transformed.name]) + self.counts\n", " draw = sample_dirichlet(conc_posterior)\n", "\n", " # Since our new_p is not in the transformed / unconstrained space,\n", " # we apply the transformation so that our new value\n", " # is consistent with PyMC3's internal representation of p\n", " point[self.name] = stick_breaking.forward_val(draw)\n", "\n", " return point" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The usage of point and its indexing variables can be confusing here. The expression point[self.conc_prior.transformed.name] in particular is quite long. This expression is necessary because when step is called, it is passed a dictionary point with string variable names as keys. \n", "\n", "However, the prior parameter's name won't be stored directly in the keys for point because PyMC3 stores a transformed variable instead. Thus, we will need to query point using the *transformed name* and then undo that transformation.\n", "\n", "To identify the correct variable to query into point, we need to take an argument during initialization that tells the sampling step where to find the prior parameter. Thus, we pass var into ConjugateStep so that the sampler can find the name of the transformed variable (var.transformed.name) later." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Simulated data" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We'll try out the sampler on some simulated data. Fixing $\\tau=0.5$, we'll draw 500 observations of a 10 dimensional Dirichlet distribution." ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "(500, 10)\n" ] } ], "source": [ "J = 10\n", "N = 500\n", "\n", "ncounts = 20\n", "tau_true = 0.5\n", "alpha = tau_true * np.ones([N, J])\n", "p_true = sample_dirichlet(alpha)\n", "counts = np.zeros([N, J])\n", "\n", "for i in range(N):\n", " counts[i] = np.random.multinomial(ncounts, p_true[i])\n", "print(counts.shape)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Comparing partial conjugate with full NUTS sampling" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We don't have any closed form expression for the posterior distribution of $\\tau$ so we will use NUTS on it. In the code cell below, we fit the same model using 1) conjugate sampling on the probability vectors with NUTS on $\\tau$, and 2) NUTS for everything." ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "Sequential sampling (2 chains in 1 job)\n", "CompoundStep\n", ">ConjugateStep: [p]\n", ">NUTS: [tau]\n" ] }, { "data": { "text/html": [ "\n", "
\n", " \n", " \n", " 100.00% [2000/2000 00:13<00:00 Sampling chain 0, 0 divergences]\n", "
\n", " " ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "\n", "
\n", " \n", " \n", " 100.00% [2000/2000 00:07<00:00 Sampling chain 1, 0 divergences]\n", "
\n", " " ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stderr", "output_type": "stream", "text": [ "Sampling 2 chains for 1_000 tune and 1_000 draw iterations (2_000 + 2_000 draws total) took 20 seconds.\n", "The number of effective samples is smaller than 25% for some parameters.\n", "Sequential sampling (2 chains in 1 job)\n", "NUTS: [p, tau]\n" ] }, { "data": { "text/html": [ "\n", "
\n", " \n", " \n", " 100.00% [2000/2000 04:09<00:00 Sampling chain 0, 0 divergences]\n", "
\n", " " ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "\n", "
\n", " \n", " \n", " 100.00% [2000/2000 02:50<00:00 Sampling chain 1, 0 divergences]\n", "
\n", " " ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stderr", "output_type": "stream", "text": [ "Sampling 2 chains for 1_000 tune and 1_000 draw iterations (2_000 + 2_000 draws total) took 420 seconds.\n", "The estimated number of effective samples is smaller than 200 for some parameters.\n" ] } ], "source": [ "traces = []\n", "models = []\n", "names = [\"Partial conjugate sampling\", \"Full NUTS\"]\n", "\n", "for use_conjugate in [True, False]:\n", " with pm.Model() as model:\n", " tau = pm.Exponential(\"tau\", lam=1, testval=1.0)\n", " alpha = pm.Deterministic(\"alpha\", tau * np.ones([N, J]))\n", " p = pm.Dirichlet(\"p\", a=alpha)\n", "\n", " if use_conjugate:\n", " # If we use the conjugate sampling, we don't need to define the likelihood\n", " # as it's already taken into account in our custom step method\n", " step = [ConjugateStep(p.transformed, counts, tau)]\n", "\n", " else:\n", " x = pm.Multinomial(\"x\", n=ncounts, p=p, observed=counts)\n", " step = []\n", "\n", " trace = pm.sample(step=step, chains=2, cores=1, return_inferencedata=True)\n", " traces.append(trace)\n", "\n", " assert all(az.summary(trace)[\"r_hat\"] < 1.1)\n", " models.append(model)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We see that the runtimes for the partially conjugate sampling are much lower, though this can be misleading if the samples have high autocorrelation or the chains are mixing very slowly. We also see that there are a few divergences in the NUTS-only trace." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We want to make sure that the two samplers are converging to the same estimates. The posterior histogram and trace plot below show that both essentially converge to $\\tau$ within reasonable posterior uncertainty credible intervals. We can also see that the trace plots lack any obvious autocorrelation as they are mostly indistinguishable from white noise." ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAABLsAAADTCAYAAABp7hHfAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAA9hAAAPYQGoP6dpAAD0YklEQVR4nOydd5wdVd3/32dm7r1bs9n0QkIKJJGeAKEKCEhRH8HnkSogRRAQwYaKSlFRwEcUBAsiwkMRUX4q0pGuoYQeIARIQtom2exm+969ZWbO748zM3du2WTTSOH7fr2S3b137syZMzPnzvnM5/v9Kq21RhAEQRAEQRAEQRAEQRC2AazN3QBBEARBEARBEARBEARB2FiI2CUIgiAIgiAIgiAIgiBsM4jYJQiCIAiCIAiCIAiCIGwziNglCIIgCIIgCIIgCIIgbDOI2CUIgiAIgiAIgiAIgiBsM4jYJQiCIAiCIAiCIAiCIGwziNglCIIgCIIgCIIgCIIgbDOI2CUIgiAIgiAIgiAIgiBsM4jYJQiCIAiCIAiCIAiCIGwziNglCIIgCFshf/vb35g6dWr0b6edduKggw7ikksuobm5eaNu63e/+x2PP/542esvvvgiU6dO5cUXX1zndW7IZz8MbrjhBqZOnbq5mxHxzDPPcMMNN2zuZmwxVDo+p556KqeeeupmapEgCIIgCFsSzuZugCAIgiAI689VV13FpEmTyGQyvPzyy9x0003Mnj2b+++/n5qamo2yjZtuuokjjzySww8/vOj1nXfemXvuuYcddthho2xnS+K4447j4x//+OZuRsQzzzzDXXfdxVe/+tXN3ZQtlssvv3xzN0EQBEEQhC0EEbsEQRAEYStmxx13ZNdddwVg3333xfM8fvOb3/D444/z2c9+doPWnclkqKqq6vf9uro69thjjw3axpbKqFGjGDVq1OZuhrAObIuiqyAIgiAI64eEMQqCIAjCNkQoPi1fvhyAG2+8keOOO46ZM2cyY8YMPve5z/HXv/4VrXXR5w499FC+/OUv89hjj3Hsscey6667cuONNzJ16lTS6TR///vfo5DJMFSsUijim2++yde//nUOPfRQdtttNw499FC+8Y1v0NTUtN771NzczKWXXsrBBx/MLrvswoEHHsiFF15Ia2trtMzy5cv51re+xX777ccuu+zC0UcfzR//+Ed834+WWbZsGVOnTuWWW27h1ltv5dBDD2X69OmccMIJvP7660XbrBQmN3Xq1IqhhIceeijf/e53i157+eWXOeGEE9h11135+Mc/znXXXcdf//pXpk6dyrJly6LlHnroIc4880wOPPBAdtttN44++mh+/vOfk06no2W++93vctddd0VtCP+F69Fac9ddd3HMMcew2267sffee3PhhReydOnStfZtW1tbUd/uu+++nHjiiTz33HPRMrNmzeK8887joIMOYtddd+WTn/wkl112GW1tbRX7bN68eVx44YXsueeezJw5k6uuugrXdVm4cCFnnXUW06dP59BDD+Xmm28u+nx4Pt13331cddVVHHDAAey2226ccsopzJ07d637UhrGuC7HG+Avf/kLRx55JLvssguf+tSnuP/++/nud7/LoYceutZtC4IgCIKwZSHOLkEQBEHYhli8eDEAQ4YMAaCpqYkTTjiBMWPGAPD6669z5ZVX0tzczAUXXFD02bfffpsFCxZw3nnnsd1221FdXc3hhx/OF7/4RfbZZx/OP/98wDi6+qOpqYmJEyfy6U9/moaGBlpaWrj77rv5/Oc/z4MPPhi1a6A0NzfzP//zP7iuy7nnnsvUqVNpb2/nP//5D52dnQwbNoy2tjZOPPFE8vk8F110EWPHjuXpp5/mmmuuYcmSJVxxxRVF67zrrruYNGkS3/ve9wC4/vrrOeecc3jiiSeor69fp/ZVYt68eZx55plMmDCBa665hqqqKv785z/zz3/+s2zZRYsWcdBBB/HFL36R6upqFi5cyM0338ycOXO4/fbbATj//PNJp9M8+uij3HPPPdFnR4wYAcBll13G3//+d0499VS+9a1v0dnZya9//WtOPPFE7rvvPoYNG9ZvWy+++GLmzp3L17/+dSZMmEBXVxdz586lo6MjWmbJkiVMnz6d4447jvr6epqamrj11ls5+eSTuf/++0kkEkXr/NrXvsZnP/tZTjzxRGbNmsUf/vAHXNflueee4+STT+ass87i/vvv5+c//znbb789RxxxRNHnf/nLX7LTTjtx5ZVX0t3dzY033sipp57KP/7xD8aNG7fOx2Mgx/uee+7hsssu48gjj+SSSy6JtpvP59d5e4IgCIIgbH5E7BIEQRCErRjf93Fdl2w2y0svvcRvf/tbamtrIzfKVVddVbTszJkz0Vpz++2385WvfAWlVPR+W1sbDz74IBMnTizahmVZDBkyZEAhi0cddRRHHXVU9LfneRxyyCEccMABPPDAA5x22mnrtH/XX389HR0d3HfffUyePDl6/VOf+lT0+6233kpzczN//etf2W233QD4+Mc/jud5/PnPf+aLX/xi0T7V1tZy0003Yds2YESj4447jmeffZZPf/rT69S+Svz2t7/Ftm1uu+22SNw75JBD+K//+q+yZUMBEYxDa8aMGUyePJlTTjmFefPmMW3aNMaPHx8JVqXH4PXXX+cvf/kL3/3udznjjDOi1/faay+OPPJIbr31Vi6++OJ+2/rqq69y3HHHcfzxx0evleZmO+mkk4raOH36dGbOnMknPvEJnn32WQ477LCi5U844YSoLfvvvz+zZs3izjvv5MYbb+STn/wkADNnzuTpp5/m/vvvLxO7hgwZwq9//evo3Nxzzz058sgjuemmm7jyyiv73Zf+WNvx9n2fG264gd13351f/epX0ef23HNPjjjiiEhUFARBEARh60HELkEQBEHYiomLFABTpkzhiiuuiMSR559/nptuuok333yTnp6eomVXr15d5PqZOnVqmdC1rvT29vKb3/yGxx57jKamJjzPi95bsGDBOq/v2WefZZ999ikSukp54YUX2GGHHSKhK+S///u/ufvuu3nhhReK9uuQQw6JhA+AadOmAWxQqGWcl156iX322afIxWZZFkcffXRZGOTSpUu57rrreOGFF1i9enVReOnChQujtvXHU089hVKKz372s7iuG70+bNgwpk2bxuzZs9f4+d12242///3vDB48mP3335+dd965zKm1evVqrr/+ep555hlWrVpVFBq6YMGCMrHrkEMOKfp78uTJzJs3j4MOOih6zXEctt9++4p9/pnPfKZIhB07dizTp09f78qdazveH3zwAS0tLZx11llFnxszZgzTp0/faOeFIAiCIAgfHiJ2CYIgCMJWzDXXXMPkyZNxHIehQ4cWuVDmzJnDWWedxcyZM/nxj3/MqFGjSCQSPP744/zud78jk8kUrWv48OEb3J5vfvObvPDCC5x//vnsuuuu1NbWopTinHPOIZvNrvP62tvbGTly5BqX6ejoYOzYsWWvh30RD8kDGDx4cNHfyWQSYL3a1197KoUODh06tOjv3t5eTj75ZFKpFF/72teYMGECVVVVrFy5kgsuuKDs+FQiFMj233//iu+vLezvl7/8Jb/97W+59957uf7666mpqeGTn/wkF198McOHD8f3fc4880xWrVrF+eefz5QpU6iurkZrzfHHH1+xzxoaGor+TiQSVFdXk0qlyl4vFWCBin03bNgw5s2bt8Z96Y+1He/29nag/PiE2xWxSxAEQRC2PkTsEgRBEIStmMmTJ0fVGEt58MEHcRyHm266qUhoePzxxysuH3fTrA/d3d08/fTTXHDBBZxzzjnR67lcjs7OzvVaZ2NjI83NzWtcZvDgwbS0tJS9vmrVqmgdG4NkMkkulyt7PRRL4u2JJ88PKX3thRdeYNWqVdxxxx3MnDkzer27u3vAbWpsbEQpxV133RWJOKVtXhNDhgzh+9//Pt///vdZvnw5Tz75JNdeey2rV6/mlltu4b333mPevHlcffXVfO5zn4s+F+aG2xT013elotXGIjw/Vq9ePaC2CIIgCIKw5SPVGAVBEARhG0UphW3bWFbh6z6TyVRMlL4mksnkgFxGSim01mUCy1//+teicMZ14aCDDuLFF19k4cKF/S6z3377MX/+fN5+++2i1//xj3+glGKfffZZr22XMnbsWN59992i155//vmiyokAe++9Ny+++GJRtULf93nkkUeKlgvFxdL++vOf/1y27XCZ0uNwyCGHoLWmubmZXXfdtexfaUXJNTFmzBhOOeUU9t9//6j64bq0cWPxwAMPFIVzNjU18dprrxUJghuTiRMnMnz4cB5++OGi15cvX85rr722SbYpCIIgCMKmRZxdgiAIgrCNcvDBB3PrrbfyzW9+kxNOOIGOjg5uueWWtbp9SpkyZQqzZ8/mySefZPjw4dTW1jJp0qSy5erq6th777255ZZbaGxsZOzYscyePZt7772XQYMGrdc+XHTRRTz77LOccsopfPnLX2bKlCl0d3fz73//m9NPP53Jkydz+umn849//IMvf/nLXHjhhYwZM4ann36aP/3pT5x00kkbnIcs5JhjjuH666/n+uuvZ+bMmcyfP58777yzrILjeeedx1NPPcXpp5/OueeeG1Vj7OvrA4jEx+nTp9PQ0MDll1/OBRdcgOM43H///WWCGphjAHDzzTdz0EEHYVkWU6dOZc899+SEE07ge9/7Hm+99RZ777031dXVtLS08MorrzBlyhROPvnkivvT3d3Naaedxmc+8xkmTZpEbW0tb775Jv/+97+jRPKTJk1i/PjxXHvttWitaWho4KmnnmLWrFkbpU8r0dbWxle+8hWOP/54uru7ueGGG0gmk3z5y1/eJNuzLIuvfvWrXHbZZVx44YX8z//8D11dXdx4440MHz58gx2PgiAIgiB8+IjYJQiCIAjbKPvttx8//elPufnmmzn33HMZOXIkxx9/fBS6NlC+//3v88Mf/pBvfOMb9PX1MXPmTO64446Ky1577bX85Cc/4X//939xXZcZM2Zw6623rrdQMXLkSO69915+9atfcfPNN9PR0UFjYyN77rlnFNY2ZMgQ/vznP3Pttddy7bXX0tvby3bbbcfFF19cVKFwQznrrLPo6enh73//O3/84x/ZbbfduP7664sqKoJJgP7HP/6Rn/3sZ3znO9+hoaGBz372s+y99978/Oc/j8SxxsZGbrrpJq655houvvhiqqurOeyww/jlL39ZFDIIJmn7q6++yp/+9Cd+/etfo7XmiSeeYLvttuNHP/oRu+++O/fccw933303vu8zYsQIZsyYUZa0P04qlWK33Xbjvvvuo6mpCdd1GT16NGeffTZf+tKXAJNX63e/+x0/+clPuOyyy3Ach/3224/bbrutLBH9xuLrX/86b775Jpdccgk9PT3stttu/OIXv2D8+PGbZHtgKkgqpfjDH/7AV77yFcaOHcs555zDE088wYoVKzbZdgVBEARB2DQoHfeJC4IgCIIgCPz0pz/lvvvuW+8KgJU488wzaWpq4tFHH91o69yWePHFFznttNO4/vrrOeqoozZ3c+jq6uLII4/k8MMP58c//vHmbo4gCIIgCOuAOLsEQRAEQRACVq9ezWuvvca//vUvpk+fvt7rueqqq/jYxz7G6NGj6ezs5P7772fWrFn85Cc/2YitFTYWLS0t/O53v2OfffZh8ODBLF++nNtuu43e3l5OO+20zd08QRAEQRDWERG7BEEQBEEQAp555hl+/OMfs/vuu69TqGcpnufxq1/9itbWVpRSTJ48mZ/97Gccc8wxG7G1wsYimUzS1NTED3/4Qzo7O6mqqmL33Xfnhz/8ITvuuOPmbp4gCIIgCOuIhDEKgiAIgiAIgiAIgiAI2wzW2hcRBEEQBEEQBEEQBEEQhK0DEbsEQRAEQRAEQRAEQRCEbQYRuwRBEARBEARBEARBEIRtBhG7BEEQBEEQBEEQBEEQhG0GEbsEQRAEQRAEQRAEQRCEbQYRuwRBEARBEARBEARBEIRtBhG7BEEQBEEQBEEQBEEQhG0GEbsEQRAEQRAEQRAEQRCEbQYRuwRBEARBEARBEARBEIRtBhG7BEEQBEEQBEEQBEEQhG0GEbsEQRAEQRAEQRAEQRCEbQYRuwRBEARBEARBEARBEIRtBhG7BEEQBEEQBEEQBEEQhG0GEbsEQRAEQRAEQRAEQRCEbQYRuwRBEARBEARBEARBEIRtBhG7BEEQBEEQBEEQBEEQhG0GEbsEQdgiePzxx7nttts2dzMEQRAEQRCEEuQ+TRCErQ0RuwRB2CJ4/PHHuf322zd3MwRBEARBEIQS5D5NEIStDRG7BEEQBEEQBEEQBEEQhG0GEbsEQdjsfPe73+Xvf/87TU1NTJ06lalTp3LqqafS1tbGpZdeyic/+Ul22203Dj30UH7wgx/Q0dFR9PlTTz2VU089tWy9U6dO5YYbbviQ9kIQBEEQBGHbQ+7TBEHYGnE2dwMEQRDOP/982tramDt3LjfeeCMAdXV1dHR0kEql+Na3vsWQIUNYtWoVf/zjH/nSl77Evffeu5lbLQiCIAiCsO0j92mCIGyNiNglCMJmZ/z48QwZMoRkMskee+xR9N4PfvCD6HfP85gxYwaHHHIIb7/9NjvvvPOH3FJBEARBEISPFnKfJgjC1oiIXYIgbLForbnrrru45557WLZsGel0Onpv0aJFchMlCIIgCIKwmZD7NEEQtmRE7BIEYYvl9ttv56c//SlnnHEGBx54IIMGDUJrzfHHH082m93czRMEQRAEQfjIIvdpgiBsyYjYJQjCFsvDDz/Mfvvtx3e/+93otaVLl5Ytl0wm6e3tLXqtNDmqIAiCIAiCsPGQ+zRBELZkpBqjIAhbBMlksuwpYCaTwXGKNfn77ruv7LNjxoxh0aJF5PP56LVnn3120zRUEARBEAThI4bcpwmCsLUhYpcgCFsEkyZNorW1lb/+9a/MmTOHhQsX8vGPf5z//Oc//O53v+O5557jl7/8Jffff3/ZZz/1qU/R3t7OpZdeynPPPcftt9/OH/7wh82wF4IgCIIgCNsecp8mCMLWhoQxCoKwRfD5z3+eV199lauvvpqenh5mzpzJzTffTFdXF7fddhvZbJaZM2fyhz/8gcMPP7zos/vttx+XXnopt912Gw8//DB77LEHv/rVrzjyyCM3094IgiAIgiBsO8h9miAIWxtKa603dyMEQRAEQRAEQRAEQRAEYWMgYYyCIAiCIAiCIAiCIAjCNoOIXYIgCIIgCIIgCIIgCMI2g4hdgiAIgiAIgiAIgiAIwjaDiF2CIAiCIAiCIAiCIAjCNoOIXYIgCIIgCIIgCIIgCMI2g4hdgiAIgiAIgiAIgiAIwjaDsylW2t7eXvZaQ0MDnZ2dm2JzHxmkDzcc6cP1p7e3l3HjxgGwdOlSamtrN3OLtl7kPNxwpA83HOnD/mlsbNzcTdjk+L4vx38rRq7frR85hls3cvy2fuQYbt0M5F7tQ3N2WZaYyDYU6cMNR/pQ2BKQ83DDkT7ccKQPP9rI8d+6keO39SPHcOtGjt/WjxzDbR85woIgCIIgCIIgCIIgCMI2g4hdgiAIgiAIgiAIgiAIwjaDiF2CIAiCIAiCIAiCIAjCNsMmSVAvCIKwtaK1ZlkTvPQyvPeeZslS6OqC7h6wLaiphcENMG4cfGyaYs8ZMHaM2tzNFgRBEITNgrVqLuQz+GNnbO6mCIIgCEKEiF2CIAjA6tWaR/8FDz+i+WCReW3oENh+e5g0CerrwPehNw1tbfCfWXD/AxqA7bbTfPpoxaeOgqFDRfgSBEEQPjpYbR8A4G/mdgiCIAhCHBG7BEH4SLOyWXPnnzQPPgT5POy6C3zz64p9ZsKY0f0LV1prli6DV16Fp5/R3HSz5g9/hKOP1Jx+mmLUKBG9BEEQBEEQBEEQNgcidgmC8JEkndb88TbNvX8DpeAzn4bjP68Yt93ARCqlFOPHwfhx8LljFMuWaf72D80/7oNHHtN87hjNWWco6upE9BIEQRAEQRAEQfgwEbFLEISPHE8/o7nuBs3q1fBfn4EzTlMMH75hotR22ykuvEBx0gma/7tD8//+Dk8+rbnwK3DoJ4w4JgiCIAiCIAiCIGx6ROwSBOEjQzptRK6HHoaPTYOrr1RMm7ZxRajhwxXf+obivz6j+fkvNJf/SPP0s3DxN2FQvQhegiAIwjaK9kFJoXdBEARhy0C+kQRB+Egwb57mzHM0jzwKZ3wRfnvjxhe64kydovjdrxXnn6v493/g9LM0b8zRm2x7giAIgrBZ8d3N3QJBEARBiBCxSxCEbZ4HH9ac91WNm4cbrlOcdYaF42x6l5VtK04+UXHTbxSpFFz0Dc0/HxDBSxAEQdgG8b2yl156WbN0mXzvCYIgCB8+EsYoCMK2idZ43a3887ZFzH9pCd85oJ0jD+oh1doLTymwHLATYCXQVYPQ1UPQNY3omqH49aOhqmGjNWXqFMXNv4UfXqn52c81y5b1cs6X9IciuAmCIAjCh0IFZ1dXF9TUbIa2CIIgCB95ROwSBGHbQGuslnewl7yI3fQy1rJXsLKdnAowPVhkjgPJWvOHlzc35l4eRflTZ50ahN+wHbphHP6QifjDpuANm4Ju3N4IZetIXZ3i6p/ATX/Q/OnuDO/Ph5/8CGpqRPASBEEQtgFKxC6tNb6GfL7w2vMvaCZOgFGj5LtPEARB2LSI2CUIwlaNal+MM+8BEu/cj9WxGIB8w0SebT2cFxZNY68jJ7Lfpyagq4eCk4LSqohaQ7Yb1deG6mtHpVdjda1AdS7F6lyKtXo+9oInUMFNvLaT+EMm4w/bAX/krnijd8cfMQ3s5FrbatuK87+s2GXnGi69rIevfVPzv1dDQ4Pc9AuCIAhbN8r3ih4deUFUYyh25fOanl54ey6MGvWhN08QBEH4iCFilyAIWx9aYy97icQrt+IsfBqtLLxx+5CbeQ7Ngz7O1y8fytJlcMWlin0PquTbiqEUVA0yoYyNEwAoyzri5bDaPsBqfR+r9T2s1vewl71M4p37TXPsJP7InfG2m4m3/f54o/cAp3/x65j/SmFbvVx6heYrF2l++b+miqMgCIIgbLWUOLtKxa5s1vy07Q+xTcI6s2qVxrJh2FC5LxEEYetGxC5BELYq7MWzSP7neuzmN/FrhpLd/6u4u3weXTeCpcs0X/+mprsHrv2ZYsb0jXSjZifxh0/FHz616GXVswprxRvYK97AXvE6iZdvITn7JrRTjTfx4+Snfgpv4kGQqC5b5QH7K37xv/Cd72nOu0Dzy2th3HZyYykIgiBspZQkqO9P7LKkPNYWzRtvmp+fPGzztkMQtNY88yxMmQJjRss9srDuiNglCMJWgdX8Nsl/X4uz5Hn8hnFkPvlj3I/9lwlNBJYs0Xz16xrfhxuvU+y446b/UtR1I/B2/CTejp80L+R6sZe9hL1oFs77j1L9/mPoRA3u5ENxd/k83riZRWGUe+yuuOE6+Oa3NRdcpLnxehG8BEEQhK0I7Rd+93JFb8XFLq01ueBtS5xdWyw9PVI5Uxg46bRm1vMwc69Nk5LDdSHvwnvvwZjRG331wkcAebYiCMKWTaaT1ONXUH3XcVit75E99AekT38Ad9fPR0LXosWar35Ng4YbPiShqyLJWrxJh5A79Pukz36K9HH/h7vTZ3EWz6L63tOpvvO/cd7+O9rNRh+ZsqPixusUvg8XfV3TtFxuNIVth+XLl7Pvvvvyox/9aHM3RRCETUHMzaXcvqK3QrFLYwSvTBjGKLOPLZae3k2z3nxes3SZ3N9sa7S2mp/NqzbN+t0gMlrcoML6Is4uQRC2TLTGmfcgyWeuRvV1kN/zDHL7nV+ophiw8APNRd/QWBbc8EvF+PFbiDPKsvHHzSQ7bibZgy/BefchEq/+H1WPfg/3P78gMeN08nt8ARJVbL+94vpr4atf01z0dePwkkpVWwf77rvvOi3/wgsvbKKWCIIgbAZ0TMCIPciBgtgFZtIahjH6PsIWit5Ex2buO7CqBRoGaQYN2krvb7QuL3K0gcxfoGlthX332Tr7JLz8N3K3ROTXUezK5TTJ5NbZl8KmQcQuQRC2PNJtVD1+Oc78x/HGTCfzP7eU5csCWLDQCF0JB351ndpyQwCdJO7Ox+LudAz2speoeeMOUv/+OYlXbye33/m4O/83kyYluO5auPAbmgu/ofn19ZK0fmvgrLPOKnvtlltuoa6ujhNOOGEztEjYGrjrrru45ZZbaGlpYccdd+R73/see+21V8VlX3zxRU477bSy1x966CEmT54MwN/+9jcuueSSsmXmzJlDKpXauI3fBFgr30Tl0njj99ncTRHWEdW9ovC7myl6Ly525XKQCd52i/PYC1sQ/iYyX4V527yyCkBbB6p9MXbzW3ijdkMPHrfR1vvBoo22qs2Cv4nFLjc4bwYidi1foXl7Luw7U1Nfv5EalG7Dbn0Pf+hkdO3wjbPOLYiuLmMWqKvbducbInYJgrBFYS98itRjl6Ky3WQP+S756aeCKv+WW7jQuKCSKePoGjt2KxiolcIbNxNntyPpfPtJUv/+BVWPX4H32l1kD7uUHXfcm1/+HL72DSPi/eYGGDx4K9ivjzBnn3122Wuh2FXpPUF46KGHuOqqq7j88suZMWMGf/7znzn77LN58MEHGTNmTL+fe+SRR6irq4v+HjJkSNH7dXV1PPLII0WvbQ1CF76H1bFkc7dCWE/s5rcKf5SIXXEHVz4PfX2F34Utk03l7ArFiq3V1acyHeZnvnfNFb4/YoTOLmsLcHa1tZmfvWmor6+wgJtFpVvRg8YOePtW7ypUejUqVb9Nil0vvmR+bsvFKCQCVhCELYNcL6nHLqX6H+ej60aSPuX/kZ/xxYpC15Klmq99U5NMmhxdW4XQVYI/Zjp9x99O37G/Qbl91PzlNFIPf5uPjWvjf69RNK+Cb1+i6euT26ptgXjuqkWLFvGd73yHI488kn333Zfly5evNbfVvvvuy3nnnVf2em9vLzfffDMnnXQSBx98MIcffjhf+9rXeP311wfUriuvvJJ999233+VvvfVW9t13Xx5++OHotfvvv5+LL76YY489loMOOogjjjiCiy66iFdeeWVA2wQ49thjOfbYYyu+d95551UMD9Vac//993P22Wdz6KGHcvDBB3P66adz//33D3i7WwK33nor//M//8Nxxx3H5MmT+f73v8+oUaO4++671/i5oUOHMnz48OifbRdn+VZKFb0/fPjWcWNuL3xqczdB2Ahopwpck4E+l9M88ZSmpbXwfi4P6T4zKdbBMsKWx6ZydoW4W6uzKwzR9TfNDmi9dV4PUbM3sbPLXpeiFv10pd30Cvby16NxakB4QQO20uNjL3ke592HNnczNivi7BIEYbNjLXuZqkcvQXUtJ7fPueT2PQ/sZMVll6/QfO0b5kvnul8oxo7Z+oSuCKXwJn2C9Lj9SL70BxIv3Yy9+Dn2OOJKfnj5wXz/B5rLfqi56kpwnK13P7XWpNPpjbrOZDJJb+/GyaRbU1OD2lQe/BKWLVvGl770JSZOnMinPvUpurq6SCQS5NfD6tDZ2cl5553HwoUL2WOPPdhnn33o7e3l2Wef5Stf+Qo//elPOfjgg9e4jqOPPpoHHniARx55hD322KPs/UcffZTq6uqi9fz85z9nhx12YO+996axsZGWlhaeeeYZvvrVr3L11Vdz0EEHrfO+rA2tNZdffjmPPfYY48eP58gjj8RxHGbPns1PfvITPvjgAy688MKNvt2NTS6X4+233+acc84pev2AAw7gtddeW+Nnjz32WHK5HJMnT64oCKbTaT7xiU/geR4f+9jHuOiii9hpp53WuM7Gxsb125GNiJ9yIGVuR1V9LcqpPPYL5WwRx682yKNZ0wjpdlRDA23tUF2dp7sHwrcXL4Hqahg31mJpk4/nOTQ2SlnGLeEYxmnv8KitNXaahoYk1kay7DQ05MlkfWprt87j7rc4QC1UOdC9ADV2D2DDj19trRHR6uuTJBJb331efb1Lba1HQ4NNY+PGlxXC83HQIIvGxsQalx00KE9Xt09DQ+VzzG9OgFWLaqhHJWui19d0DP3uasjXQn0t1hZ2rQ4Ef2kWamr6bXt4/jU2FlzguZzmX0/kmLl3guHDtn5flIhdgiBsPtwcyed+ReLlP6IHj6fvhLvwx+zR7+KrVpnwvmzWOLrGj9v6bgwqkqgit/8F5KceRdVD36b6H+dx2O4n0v2Ni/nJz6u45n813/suH5ogszHRWnP00Ucze/bszd2Uftlnn3146KGHPpT+nTNnDmeeeWaZ2LF8+fJ1Xte1117LwoUL+cEPfsBnPvOZ6PVzzz2XM888k6uvvpp99913jaFsM2bMYOTIkTz55JN885vfJJEo3EzOmzePRYsWcdRRR1FTU7gxvPvuu8vC7VpbWznjjDO44YYbNonYdd999/HYY4/xX//1X3znO9/BccztSz6f55JLLuFPf/oTRxxxBNOmTdvo296YtLe343keQ4cOLXp92LBhtLS0VPzM8OHD+fGPf8zOO+9MLpfjvvvu4/TTT+eOO+5g7733BmDSpElcddVVTJ06lZ6eHm6//XZOOukk7rvvPiZMmLDG9mwy8mmwEmCveYLixERrt7UZUnVrWFoIaWxs3LTHb4CEx0/7SVRfL+7qVXR2J4k/i7BUwTE0ahQs/MDkKqqu3vq+0zYmW8oxjNPRoaNj19raSyKhyGZNXp8NEWN6esx6V7dCXe3Wd9ztjtUoLwe98wHw/CoGT9hlg49fb6+5MFav7iWV2gr6JdeLs/BpvLEz0PWj6QzOl64uaG/f+O1fvdqsP5Vc+/q7usyyHR2Vxxa7tweVS+O2t0EyFHnWfA1aHW1Yvb34Thf+FnatDoRwfHbb28H3UJ1L0IMnREnWwvOvvb3wQDocA15+ecsvnDAQsXnrl+sEQdgqsVrepfpPx5N8+Rbyu59I+tS/rVHoamvTXPRNTXcX/OLnikmTtuwBeH3QQ3eg7+Q/k9vrLJw37uG/207mW6cv5eFH4Xe/3zot1LB1inSbiqFDh3LGGWds8Ho6Ojp44okn2GuvvYqErnAbX/jCF2hvb+ell15a43qUUhxxxBF0dXUxa9asovfC/E9HHXVU0euV8koNGzaMQw45hKVLl7JixYqy9zeUe++9l+rqar71rW9FQhdAIpHg3HPPBeCxxx7b6NvdVJReE1rrfq+TSZMmcfzxx7Pzzjszffp0rrjiCg455BBuueWWaJk99tiDY445hmnTprHXXntx3XXXMWHCBO68885Nuh9rwlnwFPaStVQfDUJDdGoQAMrLrmlpYUsjHtpjBdell8ML8jINboBJEyHU0Ic0GsHEcbbe3E3bOvGcXeExevY/5l8lFn6wbukW8ltjcQKtjdAVZyPf1mwtiftVtsv87F4JEF3rm+p6DotZDGT94Vdo/8sGC2h/wGGjynejz2zVeHms1Quwm+eiugf2cLWnZxO36UNCnF2CIHy4aJ/EK7eRnHUdumowff99M96EA9f4kc5Ok6OrtRV++XPF1CnbsHhiJ8kd9C287fej6sFv8QV9PPax13DN3QcxerTm2M9uXfuulOKhhx7a6GGMG/OJ+IcZxrjjjjsWuafWl7lz5+J5Hrlcjptvvrns/aVLlwKwePFiDjxwzdfX0UcfzR133MEjjzzCIYccAoDneTz22GMMHTo0cg+FNDU18X//93+88sortLS0kMsVTwJaW1sZPXr0BuxdMZlMhgULFjBs2DBuv/32svfd4G548eLFG22bm4rGxkZs26a1tbXo9dWrVzNs2LABr2f33Xfnn//8Z7/vW5bFrrvuyqJFi9a3qRuFcGLUL74J39XVg82yrohdqms5qms5/sidIFGz9g9sToKJoHaq8BsnYPe2gJePJu4fm2aqfK1q0WRzhbw7liVi15ZKPGdXXICJjpfWoD2WLrf54API5oyTZsb0Na83XNdWWYmzVOiCjX4Cb2m5zGa/pBkxAiZsX3JvFO53kE/X38RiV7hezzMupOdegL32hMY1FG/qVzgM2vziCy75BBy4/wAasJXn7IrwctH3bbRPJSxYqBk2tLCrW/keR4jYJQjCh4bqXkHqkUtwlr6Iu+ORZA6/AqoHr/EzXd2ar1+sWdYE1/5MscvOW5fYs7542x9A+pR7qbr/Ik5sPp9Bh53P5dedy8iRNvtt4bbiUpRS1IaJWzYStbW1ZSLL1kBpBb31pavLiAhz5sxhzpw5/S7XF5Y/WwOTJk1iypQpPPfcc3R3d1NfX8/s2bNpa2vjpJNOKkqEvnTpUs466yx6e3uZMWMGBx54ILW1tSilePXVV3nttdc2+nHp6upCa01LS0uRm6mUgezr5iaZTLLzzjsza9YsPvnJT0avP/fccxx22MDLIb3zzjtrTECvteadd95hypQpG9Te9WagE4PgpltXNaCtBNaquejulfhjpm+6WvZbOFbbQlSmE107HN24/eZuzpoJknX7w3Y0CeoB5eWiHN7h0OHYxX8rtZZTxM2ielahB4/b+G3OdoOdglhuuHff0wyqh9GjN/05196ueflV2H9f2JJSAPX1afr6ikWLSgKM1TIPq20hC1cdRS5vxIOBJA/fqsUuv0KjKwlg67ramLK4pTm7OrvMvwmlQ1DYF8oc9LDdH4bYNX+B+b2lBRoH9/+Z/oVDc31n+jxyoWHLy6M6lvY/1mxDzq4C5YOv1pqFH5gQ8+l7fGitWjuZLpxF/8addAgk128eIWKXIAibHq1x3n2I1BM/BO2TOeoq3I8ds9bJTHe35hvf0iz6AK7+qWL6Hh+tyY8eNJa+E+4i9cSP+PTbv2b4wfP49o+u4ZfX17DjDh+tvtjWsYK62l6FO96eCl7yUDw8+eSTN0pi9qOPPprrr7+eJ598kmOOOSYKYTz66KOLlvvzn/9MV1cXV1xxRVl44zXXXLPWJOshlmX1m5S/tPBAuK/Tpk3jtttuG9D6t2TOOOMMvv3tb7PLLrswffp07rnnHlasWMGJJ54ImFxszc3N/OxnPwPgtttuY7vttmOHHXYgn8/zz3/+k0cffZQbbrghWueNN97I7rvvzoQJE6KcXfPmzePyyy/fLPuIHuDMLbwBd6rwx87AWjUXq3sFvvsxSFRvuvZtyQTuA+Vmt/wn6+FE0HIKudlizq5gWCOMPB6os8te/ioq3YZbOxwSVRu1yc4Hz6KdKrwdjLjsupolxgjLRjSkluF5mrfehtCQ3LUW0+OHzXMvmGMybrvCa5UKD1ptCwFQngsYwXANaSEjQgGiu3sDG7o5qKDMKn8di8r0deAsnoU7fj+oMQ+94sLfliR2ed4aRp5Q9InuWYpf3tiE40S6z/yDgnje37L9O7vMfbOi0Fi94k3slXNxU3VQXUF9joTOLX40XiPKyxWq21fYlSIX5xZ0LlpdywDMw48hE9drHSJ2CYKwSVHdzaSe/BHOgifxxswgc/Q16Ibt1vq5nh7NN76tWbAQfnqlYubeH1Fxx0mRPeJK/BE7sffTP+WmfU7n0ktv5OrrRzBixEe0T7ZB6uvrASomKX/33XfLXttpp51QSvHmm29ulO0fccQR3HjjjTzyyCMcccQRPPvss5HjK05TUxMAH//4x4te931/jQ6zUurr61mwYAGu6xbl4Orr64tCMENqa2uZMGECixYtipxnWzOf+tSnaG9v5ze/+Q2rVq1iypQp/P73v2fs2LGAOQfiec/y+TzXXHMNzc3NVFVVscMOO/D73/++qEJmV1cXl112GS0tLdTX17PTTjtx5513sttuu33o+wcM+DF/mAdH2wmobsQftiN206uVnRRbAl4ecj2VJ0Ubi3DW6K6/UzGd1tg2mz7hdZHYFTilvBwq3UoqD7ZtQnPDSzwUv0qdXV3dmnnzYM8ZYNsK8plwAwD0pT3aV7uMGTcAVWUAKDcT/R6KT8lNXAR06TJYFRvenS1sBlZJKOhPNPB9jet6Uebn5AAi88N19fRC62rNsKFbyf2Lm6ucS9BdN2eXSq8GwOptwd/Cxa6y51DZHqzOJfgN4woPMpQ5gcN2e5tY7IrTn5MwXHbRYhg6RDNkSMk5FohdljYd73k6Oo4q240uHde1jnJ2qa09jNHNFgwGFZTJ8JjHi4lsEYQC3QaIjZKgXhCETYP2cd74MzX/9xnsxc+TPeS79B1/+4CErnRa863vaN5/H678odrqwvY2OkqRn/4FMsf8mh0aFnD9LifzqyveJ53ekr6RhA2htraW8ePH88YbbxSJPb29vfz2t78tW37o0KEcdthhvPnmm9x5550Vk62+9dZbZDKZstcrEebmev3117nnnnvo6+src24BjBo1CoA33nij6PU77riDBQsWDGhbAB/72MdwXZdHH300ek1rzW9+85uK4YjHH388mUyGq666quL7y5cvX6+KlpuLL3zhCzz55JO89dZb/O1vfyvKi3b11Vdzxx13RH+fffbZ/Otf/2LOnDnMnj2bP/3pT0VCF8D3vvc9nnrqKd566y2ef/55brnlFqZPX0sSnU2JHqBYFbojQqHEKriDNhaZjKapSZPPr/94ufADzX+e02Tnv4Kz+LlN++g7nFzlB3btVmLW8/0nFN+YqHDia9nmn1Io36V65YsM73kxmpSW/ix1dr37rgmZKrh+gmPl+6A1zc88RsfsJ8jlNvA7LzZOhsUTwuGkdhOnRysdireoCWWMbEzXKRMaMsaO5rqgYhPmte2K55kQyRFB5PU6fFVsdpz5/4rOFW3FFMqNEMa4JjdNe4cmm908J0lpJgKrYzFW2wdYq+ejopxd5r48v4lTWlUSu+L9ZjW9gr3k+bLXlzVVWluxs+vF2RTcXpnOsqV1/HtoSw5j1H7/ByB0Cue6KVRVKF82PI6qZGz2N/tA1b9AN1C2sOcKgiBsC6j2D6h67DLsppdxt9+f7OE/HJDIBQWh6515Rug6YP+PuNAVw5t0CNmT7mTIX87lx84p/P6n13PWFfviONJH2wInnXQS11xzDWeffTaHHnooWmuef/55pk2bVnH5iy++mCVLlnDjjTfy8MMPs+uuu1JbW8uqVauYN28eS5cu5cEHH6SqamBhQEcddRQvvPACf/jDH7Asq6LY9bnPfY4HHniASy65hMMOO4yGhgbeeust3nvvPQ444ICyio798fnPf54HHniAn/70p8yePZvGxkZef/11enp62HHHHXn//ffLtvvWW2/x0EMPMWfOHPbee2+GDRtGW1sbixcv5u233+ZHP/pRxUqRwmZgoAlcQkHHMY4dHYTCKS+/0YJGPlhkJj55t0L+mQGyfIURRbLtbdQOxghS1gASFQ0Q1b4IXT/a9EPk7BqY2OW6GssCy9oM3wPBDF0H+Xu05YDv4mszhwxbVCp2KVV8ingxzcysKBS7XPDzKO2hNHR1+QwbtgH9HnMMhk6b0F1jbeLH/6XOnS3JyRMnLnZ5PkUPUqxO8yAmnweFx4w94PU5a7/cl5lIJEaOND/TW356xcpYdkGMXtcwxgrE9YnSPFMvv2J+fuJg/aHf4+VKdy0Yi1RPC7ohtECaxofXz6Y6nyudW/G+soKqkKVtqJg6NBR+gocxvWkgFxRPKimO0tGhefUllwOHampqFFtyGKPz7sP4g8fjj9q1/M3g+0RlOtHVxlGofL9sb8L+slSx8Or7m35sXCOhGKn1eh8BcXYJgrDxyPWS/M8vqbn9GKzV75M56ioy//2HAQtdfX2ab1+ieftt+NEVio8fKCJOKf6InXC/eA9e3SguaDiXh3715IBLKAtbNp/73Of45je/SV1dHf/85z95/vnn+fSnP82VV15ZcfmGhgZ+//vfc8EFF5BIJHj00Ue59957efvtt5k0aRKXX345DQ0NA97+IYccQk1NDa7rMn36dEaMGFG2zNSpU7n++uuZOnUqTz/9NA888AD19fXcdNNN/Ypyldhhhx247rrrmDZtGk899RQPP/wwEydO5KabbqKurq5seaUUl112GVdeeSUTJ05k1qxZ3H333cyePZtkMslXv/rVsqqRwmZkgGGIys2grYQJg4PCz40Qxjh/gaZpuY6eWGc3oNBjqCP5XkyE2VjkerGb38Ze8Ua4EfOzUuhUBZ56Bl55deM1Z52IhzGCceZ5eXQgdpE3E8kwesYOZh2lzq7w98LDe9PPSnugNaFe39W5ge6KCu6AjVEFz3XX/h1cmph9S8qLEyeTKYiSvl8iNgT919Nrjk1Vlbk21nYLksuZYz5qpMKyNpJJZiO6P/tlTe7KjbATlQTfUlaVZzYob4rWLFq8Ye7VOPmY8AGg8kadVH4elekIN2qW/RCdXVVBFHPUVyUbjV/LlU3tCq115Eq0vV5TsAKi45lOa3xfM38BWDpfWI+bxVr+2kZx9G1Ugj6wOpZUeK/Qeaqvo9D2CkJtNnir1NnVXZ4y9kMmdHat/wkmzi5BEDYcrXHmPUDy2Z9j9a4iv/N/kzvw6+jaYQNeRV+f5jvf07z5Jlx+meLgj4vQ1R+6fjTOmbfT9vsvc1z+azx761Xsc+ZnNnezhIAXXnih7LUxY8ZUfL2U4447juOOO25A6wSoqqrilFNO4ZRTTln3hlZY15NPPrnW5fbcc09+//vfl70+bdo0zj777KLX1rTfe+21V8XqipXCNkMOP/xwDj/88LW2Udg4eJ42OZTWlYFOBPN9xQnIwyTn/oZPKD5YZH4ODQqg9lMPYUCoIrFLbVylIhSMwolIGBq4Djf3HeUROJsca9U7heMcHrfQ2RW4AVS2B52spfQMUiV5Yfqt6OZ7oP3o87nsBgoMMZEydBFuaDW5bFbz7H9gyg6a7bfv/1rxPBjcYCqdPfVMsbih0qvRytq0ueDWQDwhuetBKmna53ulYpdm+XLN6jYYvr1Hba3CsnRZ36XTGq2httb0h+cXhM7SY79epNtwljyPN24murb/qrQbisoXF0spuibX+YQp3+n4KvK5+OuFZVtaYMxaCiesboP350NnJ+y+PmkaM52mIEgQTh7lbwotMW4GXd2I6ms3ogmA9kzutg24frJZk1twTc41zzfiayoJ+8yE2S+Z9mmtUbnY8fFdfM8m4YCTALfSeK8Uvg8Kl6oUpNoX43kabSeNcyjbw5wnl5MYvh2dvdU42jX5BS0bles120vU4A+fuu47u6lYUzGYsFpu/Wis7hWorqai1+OEzi5Fcf61l1+Bwz6hN51zWPtYq95BJ2vQjeUJ6HVUyExydgmCsJmwmt+m+p5TqHr42+j6UaRPuofskT9ZJ6Gru1vzjYs1b7wBl35fceghInStlerB1J17C4v0nnyi/dssuPeezd0iQRC2ITIZzZNPQ9Pydb/JVFF5+jWM5VqbSX686qJVCGPcWISiQsWwFgquiDW5c8KJXDSh24jOrmhflYXV9CoqmIioLTlHTD6D1bYQq32RCV0MwlCxHVRM7IocBOHD+eDjcXdPW5smE5jYoklWKCpoI3aFn3NzGyp2xSZ5jjnvvHXXFosIz6uVzWtezvPMpD10TcXFLnvJCyYX3GaiVAgOk+eXOrsyGZ/Vbeb34UPD3E3Ffae1ZtbzprpjKKJpXRBONoazSwVuHNW1ifM0ljm7gv2x7HXeifC6jn8uLvrlY0NKvM/T6TWv1/c18+eb3zs61qlJEc6i/2Avfj76OxSwlDLtVW4WXTOseDzXuui8WVexq6/PiMQvvAjWyrf6PZbah+HD4ID9VSSKNa+C554HFS/i4WbxPBgxAkaN7N8p9977ZmytqgLHT+PZtZCqA3xo/YCGzHvkW5bha5PI3tcUckrCh+MoXBdKOl51ryycY8FPXTMUbSUK52Dw/RWPConCGEuuz6Tbgde+lsFtQ8j2YLUvwm6eW/l9SVAvCMLmQrUvIvXgN6m56/OojiVkjryKvpPuxh+9bo+V2js0F31DM+9d+NEPFYcfJkLXQFGpOoZ+5Xe80Xcwuy+5gtUP/GFzN0kQhK2I1tWaJ540Yk8poVuoeX3uc3WYy6n/20zVvsiIYolYZnCljHiyEScU4cStP2dXR4dxRbwzr/91lLp/1MYMY/RjYld3UIVTrX+OmA8loXB8/5O10a+FnF3KhEAFxzEqAhY0LV6Nce47sdV6oNoXF/Ih+SaMMUpj5lYIQ3Q1r76mmTdPrz2kPzwv7UT0+4Y6u0r3rT/cQOxSyoTybarqdetDaYhlKHZ5frFo0NNT2MmGevNGaRhjpQqD8bw/G6XaW5DcLQyvWxOtq3WRc22dKA33ClfjpAgrhQ58XW7xT4pFhfj4FPZnwoHe3tg1nes1oXSxdbS3F0LNNiQkV+UK8WpRZU6fKJeVTlTh149GJ2uixvcndrlu4eHBy69onniqPNl+WyCaZtNZrI7F2Mtfq9iu0pxRofiX7gM/bt/ycpGoatvmHCsbC7UfFFfwSKXA8XtxrVrAoqfbp2O1WZ+lg3FL580xKhK7+gkv1z6qt6Us99cmJx6q2LMKu+kVOhfMNyGt8QIisXG6IHbF1tPRxLCeF1GqeGwa0T0La9nLkN008Yxrz3234QnqRewSBGGdST15JTW3fQZn4dPk9jmP9BkP4+58bEyBHxgtLZoLLtQsWQo/u0pCF9eHZE0VIy+4nqfbPs32711L5tFfbbrkCYIgbFN0dZlJQWeFMLiWIFfMeiWnjWY+/Y/pKsjn5A+bUvyGZWO1L9pokwYvmBf25+wKnTZrKlxaNPmjpEpXQD6v161aYDQ5iiVLCd+ynPUex9dXtOkXrcsnGrEJio5PoiwH/Dy+bwSdcCJTKgjFc3bVxj5Othu7+a3iBPXaj8QRt0L/pvtMGNfSJlgR5qrOpwu5eGJEzgY7EW0jFGfWV4CJTGhr+Xzo7AIT0rclJagv3fewnaVF3np7NBqLceMglQzyqpXk+InvVyWxS62HsyuT0bSujomZ4fnnrlnsymY1r70Or1bWUdZOf+HKVjJm9/RQnU3YS57Hanpl7euKCVXheJJKFo9P4arr6oy+1htE61mr52N1LUd1r6Snx1SZdT1A+4wa4QdOvHU8kSucuOHx8X3Q4Thsp/DHTMeb9Anjxo2JXQmn+Bz4939MqC5Ae4d5r3V18TbaO8zPmtxyXFfTnq6mlLnvaPoyxd9B8dbGnZ7KzUbhsqFYWyrieoEaqPBIJTWOlyZv1aOV4r33fBZ9kI/eByN6+SVil9W90rinSlAdi7GXzsZqfrvsvQHju+suKsXDGHO9uJ7m3XfyvDGH2NMZqyBSUqiiGz/06bYOqvKtWKo8LNl1i8XQjYpXQeWNo9bw3gARsUsQhAGhugsWY2fe/eSnf4H0WY+RO+DCwAK8bjQt15x/oaZ1NfzifxV77yVC1/oyaHCS0edewz+XH8ewt3+LfuIXIngJgrBWvH4MSl1dOgrL6htYUcBiBhLG6OXMDXiY7ylA140yv+R6K3xo3YnCGPt7gBw0sT8xDIJJqTaTgJZWzaxZXpmw9ey/4Zl/V/78kqWa7u7i5VXHEjM5ChIL63hfhX0yUFVAe9hLXsSZ9+BGF1HsRf/Gee/RotfizjZdP6rwhuWA5+J7vtHugonM6FFmQjwqqMYXd3ZVx1K2eTFrSj6vyaQ9+tJetGwlZ1d89hsKls6Cp3A+eLZ82UjsSuIFcWNRGON6ioQDFrvcwiTctjeBKLkBhN1SFwiPShG5z+Iuj+5uTd0gm8ENKrrGB+rsUhvg7Hr9DXjtdeMGUm0LsXoqZ23XWvPOvMK1Fm5/fXPaqVKxKwi51k6yUOWuZyX2itdR6baiyoBlRM6uwjrDcy6VquzsGhLkG2xpLWlXXztNy40b1c3D6K4nGdFu8m2WCjxrp/xgxI+5HwyM2om5m5QC7Udu2FSqxNkVr5YYDGvt7cXbyGYB7dOQmcfyFTB/aS1tbcVtaQqmHXGxa3Cs5o4f31nfxfeMiSkUa0v7ws2HhS98apI5FD55laIvY5y0YZVGS3s4XjcOJoxRl3xHqUwXpVhhGOZ6uH67ujVd3Rqr5V2cD55Zt+++uLPLy6J98K0kPT0UhDBlFzuovXJnl5cPw+fzZWOh57HRvo/LiItdlfpuI0xlROwSBGGNqI4lpB77ATW3Hxu9lj71H+QOuQRdM3S91jlvnub8r2r60nDDdYrddhWha0MZu53NiNOv4C+LT6Z+zh+wn7pGBC9BENZImCemVCAJJ14Ng6Bv7ZFC5UR3y2twdrm54vCQAL9hrHl/TYl314FcHpIJMxnr7a0wJgYv9RfmqAORS+HS1QUrV4KFF+WZitq9huH23ffghdmlrwYl1bNm4mT1rCq8ZYVi19rH8FS+le06HkGlzYx4oGJX03JNa+va16+y3eVKUDBBcScejK6PZc+2HNxsjnQ6EE6CyUtNjeKQgxU1NcE+xwSSon4LwpIWLDSpDV57zWP2S76ZuAFevlwhGkhFu4hgMtvRneDttzUdHTqaEK/v12Xk9htgGCOYn/21VWtNOv3hfneHfVhfb366rnHI+H6xuSnTp6mpCxS7UOwpcXYV6Q8xh5C9Ac6u0NmUd8Fe9Y5xREJZp2cysKwJ42phIwiKunjy7Y2dgTdmuhEOohM4TAI+as3RDcG1EBeKw3O/qoqKIYE11WYMDkP+/Ewv7e0alUubJO2YsD7bz5K0gnDD5ndR8bFkrftYwdkVz8MfPgUoGqutopxd9fWV+1prHe1jaTVc3wdbZ1Hap6/PjPf5fnSiuNg1YzrsspP5PRRoALTvojFji1MhLx5APhg/FD6ObTaW9xLkXQuFH4UvThzcxEEjn6XKW42PZcSi+H5ZxX/je1HifrUeA8mLs82/sEiJs/DpgYfyxy9QNxccO2XGpfA9y0Yna8nnTcqC1pZisWtwAyRsP1pHXOzUWHgexcUANibxMMaKTsoBPk1YAyJ2CYJQEdW2kNTD36Hm1k/hvPMA+V0LFeKKbm7Xkaef0XzlIk0qBb/+lWLKjiJ0bSx22cWi+tjvc8eC06h+/f9IPnmlCF6CIPRLOFkpnRSEw0YqZd5bay6kEpRXHppXhpdDVxC7wonFiqY8b8zZOOPX6NFGWqqURDzctf4mWmHfxMU3pfMb7qBak+stylOy5o1oranJNUW/w8DFrrnvwGtvDGzZMsIJu1VS1N1ySAeCYn09/U7Y4mGMXlD9DwA3R1u7jvLyWLgo7ReWdf2yc7HInVCy72r1/JJ2mwWWr0oAHtnchufsinJBr2mZQDANxS5rDWGMCz+AWc+b0L0Pi9KQ0kymcIyKu1tTVRPGOIYhYSXOrn7CGMPTfX2cXeHyFat1xgh1mdDFGd/OeuWyK51820n0oDGgrMJ4EP5M1pmToZ/Qx3B5lV6NtXR20CbzXipp2hwWA4ly21mQTBb6sXlpL8uaoKsjH43dPT3mzVDgsVrnYy97aeD7GFceg9/j/eznK4hdgWKpNYwfZ7ZdKQddpZDWENeDaicbvKdQ2uv3djUudtm2oiqIeCwSu4K8bLYVc3bFtqm15v33A7FL+yQs86br22itUOhI7Kqvh1RKkaQPTyei7zEdFOIoc/xlY04v7WG1vFuhuMEAKMoNNsCKxPHj52XR2pxr2jdOr7DdOjWIdBq6u6FpmRm/w+M8ehRM392nsRHw8kXHX+EbsWsA+fHWhmpfXKgIGb5WFMZY4dopSba/PojYJQhCEVbLPFIPfN3k5Jr/OPk9Tyf9pcfJHXTxBq1Xa80dd2l+cLlm6hT4/W/UGst0C+vHJz5hkT3oO/zx/bNIvvEnko9fsUFfEoIgbLt4eZdUvqVsIhLe7CaDe+/+XE8VyXRiRSLDGiaYXmVnV5h8esECn1UthYpuG0J1tflXyaUWn2D19WnefKs4oXXYF1bg8qiuNpOJ/tL5lIsx/bR/TePyQBQUTNur3MDRFXPRhGyMviusLF/+e2kIqp2I2mE7/ScfVqpYwEgEq/HdHH190FazG5pQUCjsg9Z+v+cqBBPqWH/bLe+WLGw+3NmbRAXrCh166/tcyB/AXCx0O601jNH3ovx5a6vCtzEJj1l9kJEimzOXoeeVixiJhJk6qmAHrDU4u+LnZChARDruenS4ny85+CWdHlX1DEW2uOllfWpKlIZVhY4eFavGGOy8doJ4XC/LymZNe0fJ/sUaYwW5+sJVjBljfkYOrnDMUcXXSrhjuYwb7U82yE4fXkPzF5hQ6wETv14WPg1utuh49nVnTSPi17qyUNku6nrfNSGvdmWNLx4aXnoe+R5UJ8wBy1KDwqOnBzpK+43yayVybsWeUPjBBiw7VmAhns4qByoYS8aO1QxrNJ91fQcPI96FYlcortlk8bQTU2qdoAhH8bgWhjXqZC0q04G1ej7Ogicq5vYqJX4dtK32eH/+ujmZlPZJpzWup1HZnsCF7Jk9dcO47hSk6qKRVGm3SPxVCvC9oKhIDt8P+jh8gOKbddlLXoDMesYEA3bzW9jLXy9+cS1hjIWqxOLsEgRhA7GWv0bV38+l5o7P4SyeRX6fL9P7pcfJHfQtdO2wDVp3Nqv5ydWam27WHHkEXHetYvBgEbo2FSeeYNEx4+vc9O65JN/8C6nHLu0/0aogCB9Z6ltnM7xnNtotfopcKnaty0TRWr0ACJ6Cr0EBUF4uqGpWugIzUwnzp/RuhOgJxzaTwUqiXfwW+r33jfsrnl8myvEbtGfoECN89efMKd1Gv46hNU1mwtnWWh5U+D5Y2hy7MP+aX0GoWxNrEsRyfTmWLAkqqwVOA9W9AqtzCdEsN06iJuoXO5azq5R4eftQCFEAbh7PU6RT4/Dsaixt8uqEKPyy/i3LF1XqiCjKaZPH9zW+SqDQLFwYF9L67YY1Eolda/h82Oa42FXx/PHd6Jp75TXWKZxRdTZBEEq1roS3BzWxtD6WKg9jBE0iGVRMDa6HUoNiPA9gpQT1YTjjQJ108Um5LrmPKa2KWhoqtybX34DQPtoqFnnMTxWsXxccXqHY5eZ48y14+RWiioThuvz6UXjj96WtXZPp6IhErNpaGNJYyDkXd3bFrxXH8YNN5CMhKZ/OGC0qJk6sWocoxvgIqPJ9qO7l+J6mtsZsu7ujggNXKVS+j7q++Si8fp2K8eNRegvqxZxdrlWDpV0+WAQvBTn+48e9dN2RmOV60fHRXqE6qF0hjDGTAbRPXR2MGeGjtGm36ztoX2FpLxJWwnPVwcUnLnYlKip7KqpYWVN00lkr3yzvlBLifbRwgUcmE+77wK593/NYsBCWLDFFX3w/5iJ0Q6EyCcoiWzch6GuP3p5CInplmc+EeRY9L+xjs0BPD5DtRqVXY6+aO6B2DZi1hTGG3RAbx/N5U4F3wcKB9ZGIXYLwUUZr7MWzqPrLF6n588lYK98ke+A36P3SE+QOuAiqGzd4E0uXac45X/PoY3D2WYofXKJIJkXo2tSc/SWLVTt9lV/Pu4DE238j9egl65U4UxCEbRPX1VjZQNUpESXCecb6iF0q14NfN9KEu/enAHg5M5Gs6OwyMxkruGHvLi+qt87U1Jib90r7URQFUiGkLXxt0niXCdubdVk6h+peib34OZx5D6KCJPNQLnb1K4L0E6Lo14/CbxgX/tX/TmEmeEr7eFaK9nbjTIsneR+IoPDa6/2/t2pFjs4uM3kOc7bYTa+icmkq5WPTydqCKyWR7DcURykzhwnD+ywLLEuj3RyuNpPXrN1IVb41EhnBODNKw02j7YUTbrc4fEh1r8CZ96CZrPW2kLUGo4NQ2Uyf+fC47dY/jNH3Ae1T3zsX+torLhMKE2G4pm33c035buTQgYLTJyLdVhYGFGKveB1n8ax1anu02dCNZ8Ogethxh0Bk0cWhgAqNk7BAKay2D0DrgrPLy2GteAM3VgkiKkCoC3pBaWXOtVGUy6qSohK7gIvEFV9XvI7XBeW5BUE+puppFROjtW+q3QUJ3FXsnO+OF7ALlvOTdTQ1wduvdOF7Po29b+C0vE0qVWh/6IKygkIBRvvQ2CoUu7zo/LG0GxUUCKsIpio8Q+iXMHRxkMmVaDfPpap7fnQuZHsyBSEv6phQ7DMCtGWZ67k0VDTcnzBnYhzfh5STRSkjdpXmaIwf9/7ELt/1jIUU8LyCUFWpGmMma9pqimT4JtzQgrxn4WNFDw3CdYARgTwKYYxYjvlXNnb7gWJUIqvEKiD2R7wKsKWC/G8+A75AMn3mM3HXctSX+UzBcQj0NexMT2p7AHp7C2GjKtioUuaBgOuacyh80JBOF8YwXZK/rBILF2o6OwfoTIvde5SK18EWzQ/fw2p9H7I9dHSYCrwLPxjQJkTsEoSPJNrHnv841X86ger/9yWszqVkP/F90l96nPzMsyFVv1E28+TTmrPO0axuhWt/pvjiqQq1xjwlwsZCKcVFX1UsmXAe1839Ool37if18LcHnvRSEIRtmkcf64tuYP2SmXc4f1znMEatjSiSrAXLjoUglBCWV09WqOQbuIXsIKdKdoCpS4qbUXyjPWiQcXbVtc4ue9oeXzTcz0rhN3U1HvX1CiuRIOl2klz1GioQN9JtXWXrCOlPROmvb/xhUwqOqbVMeHQwnnsqRUsrLFhYLAgMZILfu4ZwuVSzsVlkspg8QPEqjFUN5R9IVOOFidir6o2zoMJ3TmRc0wXXz9iuf5HsXoyrkzg2ZBIjsHSepGvCZhwH0H6Ywz4inF8nHIJEyj1F71udRhxS6VZUpoMea0RhUqo9amuCsMJ+urq9w1RK6w9fQ112MbV9H2C1L664THgOh9dTIrEGsSuWBq1skr/k+fIwoI2AFxMM95mpmLC9itxnxeevNvdw4QQ62xUVG1AdS7A6l+F0FGag8XDCuIAAAxcX4/2kAyHXb5xQEIRjbpC42GVyDRb/bZbRxY6rNaG9SOzScSEjLnb5nhEAQuE+dr7n42OX1hg5KknersPKtGKnm6nNLcPuWExVlWm/zvZE8/vQ2eV7Zlu+Bl/ZdLS5UTVYpfNYFtiOFY0p6xR2Ho5Dsfx7TrYNK8gXpnO96GRt8Wdi9j1b6SissPSYhscjUUHs8jxIqBxOMoGvnCKxK5fTFat6Ru0LnV1514SUxqokxKsxhp+zWt4jueBf2H6WRDLYZ881orNOoHWxHBKeq5YFrnYI5RJtB2JXqSijfXN+lIhdOrF2sSv+XWMRy5k2wPQj2b5iN5px+wauSzdb5J72/YJYlc94RQ5CdBDG6OfIZAKXp9Ykk+Cr2EOpktB1MHnoVFCNUmvNgg9g9sslC/X3Xebn0UGV0zXl7FJeDqv1Pewlz6/zAzARuwTho4TW2POfoPr2Y6n+51dR2S4yR1xJ+sxHyE8/JSqrvKFks5pfXu9z2RWayZPgjzcrZu4tIteHjWUpLvm2YsGoL/Hzt75N4t2HqXrwmwNPfCkIwjaLFQ8fWJuza6CuiHza3Pgn6wDV7w27yhoBo6JgAnhY6EDViIdF0ddunu6uhfC+evgw2GuGEf8TCXD6WrBiLiwoDhYJJ2dxgS0KyyOYQKRqcfw02vPxxu6JTtYw963YBHfAzq5C3/iDxtDXuBNe9RAjFMYn02siCD/NOw3RtpYvK3RYf8dtQPmSfI98Tw+uXRtNyFS3yfDvjZmON37f8s8oU7nLtkCnBpnXKlTxsoLbAd8nEsds8vga8iRJJsG1zQQ74Rs1znGMVFDk7HJz0QmSSARiV7ZkJhSc5yroq4yuQQfTH4Vv1ltyqvb1aRYv1mSzmpdfCSql9d9N2H6fOc4VJoIQc3YF884iZ1c87KltYdH5OOBKqCXHU3WvwF66hkaXEHfHRW2xCscneg1j0fJG7mK24+UKzi4rwaoWTfPyXOT5i4d4RgLCOji7nnlW88GiWDuDxuiaYYWxI3bg4sKB75c7NHt6NM/+pxAqt1Z8Dx3maSoSu8Kd8AOlT9GTCY597P4qVyR2+dH10ZcYSVW+ldrWV6N+SSU1iVwLvPd05N6z0Ci06SvtgQZfBUnSg/HI0q4RuxIOdbUeVVXm3PLbl0Vjpb34OVQQXt4vsYOvfQ/bhmTCR2f7IFEidvn52Biio4+WilLzg00mEuXJ4n1txlSnKoFWtnERBevMZovH0eqSqYlSRoz1Xc/k0VI2uWwg+ijK2qN6mslnzLnq2EF8rnaxLZOg3i9xqYYP5ZUiCGMM3rCT5iFOmdgVWBdLH+YPwAUVP0eLnF0DDGPsC5xdXqoRf/B43MTgyOGHl0XbgVirNbk8RlhUkM24BbFLAUEYo6VzZHOmzz9xkGbUSPCsZKGdpUVJAHvJC9jLX4Ncb/8PWPqLLPHykXiuOpeVvx9e38F11dORZcEAHV0hInYJwkcEa9nLVN/zBar/eQFKe2Q+fS3p0x/E3eV/KicJXk/mvqM540ua//d3OOkEuOE6xYgRInRtLhxH8aPLFYtHf5Gr5nwfZ/6/qLr/a9EkSRCEjyZh2IZlgS6xmYRmiTDsqtRNA+ahRilRxaZkTfmNf3y5TDfaTlTO2QW4nhOFUMRvnp3Fz2G1vtfvekPCm/jGwdDYaKwnSdVXWfyJ7UYoclXKNWOpQOyqMk/rs6kR6PpRYCWKQu0yJXmD+s/ZVXhDVzfy9FsTeLV932BSHbM+YapYWc1vF3/czUWCYG9ibBSe0tFeWO+Ls6Gtrf+Ez4r+Q1RVpoNcTtNevTNN1R83+Ym6g6f3VQ3lITsBLlVYdkHItHrKkzSr2O75PlhKRw6hvK4mlQLXMjPcBEYscxxQWhedi878f1Hd9DSNvW+QVBkjdmW6ijcWqqVBXp2cl4rcDUobsctShdNgZbPmP8/Be/Mrh9D29mq6unQ02fe1qRrp+7BqZWVLTTZrthGGKCYcM/k364jlTOpZVXS+pPvMa2v9vo45IuwlL5pQ096WAbtDwnPctgvXbChixc+PxkbjToq7mKIE6rZDZwek7Cwfm2beDq83zy/k6hqosyufN5PzFbHTR0dWHTvmfvRinyks63nFbj3XhVUt5veeYvNf/2g3sAolis/3UMTwPdAe7Z02z7+UoKtHofOFASAW0UlXp09Xj8J1oSc1gZ7qHchUjyWfMOlC6mryOH4ffX2g0iZ+NdU8m6ErH0Lle8E3zi5tJRk9GgbVeIwYbnIJWgpQsM+ePsOGBk1f8roZKzNdqL527JZ5/exjOBjELYUmNLLK6sNzNX7gUIrGfC8fjbGWpaNDYQTGWM7A0HWZKBceAWzlkkjZaII8jYGAl80VxuIpO8DkSeXNTiUhn/PAssnkFPPeLTicwvM42qafJ5OBqvDrJnDkKQvyno0fOLuqUtAQe/5inF2JgqvPTgQJ6isUSog5u7Rlo5M1/Yaqx4kn7re0j0YNPIzRy5PpNdtoH7QX/qhd8bEjl5x2c5EA/8Ei808rh0TCFDmIjqEiCmMMk/RXpcC2zLhc5OwqzdMYr9JYoYBIRD/FSvAKzi6rZ1X/Y1YwBq6P01vELkHYxrFa3jOJ5/9yKqpruXFynXYf7tRPlQ9aG0Aup7n5Fp/zvqLJ5uD6Xyi+cp6F44jQtblJJBQ/ukKxYtzJ/PiNy3EWPkXVPy9Yv9LIgiBsE6jgpjaRML9XmqQkk1CbXYTXVzw7XLJE89ajr5NbUpKsNrjx1Ynq4rw2pdvO91YOYQxwdeGGvbJAtZbwvvgTa8BaNZchq54EyhOyV1pV3FET5lSpsjLBOi3zZDwxPGirU5TvpbdkIt2/s0uj7QTu+P1wBxmhqnU1xQ0P+s5ufgurfVHhs/k+9NwHI+FPWwn6EiODz3hMGA+1QQRNJcEmnJCkUuUT1IjgKb1r15Gz6sm7CpUOGriGB2S9iTHYNuhkLTpZi9W2sLgTMp3RhCp039hkI3Elp1OkUqCVCW9K6pjYhV8mzql8ltrcMhpy75v9yvXi1w4vLBA6Cryg8ptXheMoY8TAN8nxIwFGFxUnKBY9PbIrFvHc85oXX4JlQeos7RdCUhcv9iqKi9msuZZCx0hRTqHgszpZh/JdVD6NY8PoUdDT7WIvewm7qSQmqHSyHftbpVsLr7slyms/eLEE8iGWZV4Pz5WJE2D8dqHYVXAxGVeISRaeycKwxhxjx6qipOWhFgADd3YVOXuqTN6nMEG9jotdvnEEqd5WcjmiMFDfLx56fL9QeMIa6G2p7xXyNFUMY9S4OY9FS83feT+Bap3P0B5zvOLOrkWLNO/MM2KXb1XRXTuV7oY9SFeba39QTR4sh64u0ME56/S1ohQ4uQ7QJuxM2ymGDVXM3NNj5IggZ5dl2oL2CvmswjxsK94wTa0eXHkfgwOh486uIJ9V0jIDYU5X8cKLxhXnuhrlx4QSdCFsMJ+HtsU4Xi9TdyxsIhR5w7E37pZNVDkx8TkIXc9CJhiDx441UQqlpFImd5m2bFrbCiGc4WkRhX8Cys0YsStwiKlcD3bLu8Zh6Vv42qx/8mQYP66QXE4pM76r6LzrP4wRVOG8sJMUVexcA/FFLOWhlRNUVFzLBZLtxnn/Mfzl5js4mzc77lEICfVdN7pWm5uD7WEbsauv3Nnl2GD55qRNpUzjlALPij2UKnWRxp20Wvef39Or/Iby88VRRWVCoi4sR+GrccL2/WynAiJ2CcK2SqaL5JNXUn3n57CXv072oItJn/GIcXJVsKFuCLNf0px2pub/7oCjjoLb/6jYc4aIXFsSiYRxeK2ecDyXv3Yl9qL/UHXf+cVPZQRB+MiQcDsAM/EurS5YqP6lGNL3NtUrniv6bOtqqMk14TUvLHo9cnY51WsOxcv1BqGOlfG1HeUdqXjzvJYn5lqb/VOeEUpU57KKFbrMtor/rkpBZ1dBAOpNw6DcB1R1zzcLJKqxLMgpE9qzeFkiEm8Aekqi9iIXlZ/DWjGnMFHyPdNHNUPK9zG4o7e6l+O8+1D5DgYihtVnHCC+ShRC87RH/SDYfz9FwoG+Cs80dEzMhP7yR3n4HlTXmDi/nJ80kz6l+g3XA+hITaN76EyobjR5lbQunAPax1n0H6pXvx61w/MhoTPYNnR1mYTRYbifZ1VFkz7bAkv5hTDGcJIe/GnbCs/1Ufk0VA3GnXBgsB+x/DVA1k+SSNpBQu+CswvMsYpX/wydOdW55VjzHsZf9ha1uaU0pOeS7zIOMs83fT5smBG9QhEsjutSlHjeriR2BWKEynZhWcaVmM+6ZLOa9+f2smhx7EQtnWz3GyK0ZhuE1salls8XnFdRG4Oic65rhKYdJpvQsai6Gya5tGWZc/7l2ebCqkmYbdqFNEpFYtpAnV1xoWjKjqbPWld5uJ4O8jSFYpeLav8Ae+mLqN4WqoJ0YpWcXaF46Wvws2mslW+tWZDwA5XOThQ7VWNi9KpVXnTtea5Jil+db8byMyUhzca1E4XWBavXlulLmzyNQxTtHZDu6MP2elG2g6XAcTuNOO5TuPb8PFVVhTBGFbiVykL4soHTsb+QusjZVXhfeWY9Kct02IIlqSjZfjhWRF2r/IJrr3M5asVbNPTNLRJPk2GTw/MhErs8Roxw2GGKzfbjC0VJPNeMWwmHfh+Yp1Lg5VywHHzfjpzAYVvsQKzFzZHL+eR9Jzo3QizLuJZ1IHZZlsIbPg1dMwSdqMG2jYAZXV+WY/qpkthlxXJ2WYlgudiXTa4X1bG0bD/i14HCRyvb9M9a1GCV6SSb1eRyptpw3rVMeCiWCWPU2ji2w2qi4RgXOLu6uzyag6qdSoHyPRKJmLMr1leeqiq0s+R66V3dRV+fjt7r1ylcydnlu8FDn6TJVUmF/S7ZXuh6Gzmin46pgIhdgrCtoTXO3Puoue3TJN64m/weX6D3rMfI73UmJKrW/vl1oKVFc+kVPt+4WGMpuO5axSXftqitFaFrS8RxFFdcqtB7fI7vv3oV1uIXSf3t3Io5VQRB2LYZ1PMWQHSDG7/p9nUgAGjjfNEliagqRii6OazV7xcWKHEnFVbuotwsHX3V/SaKdn0HhXnSXJSgOrwRrlSinPhyMLJ7FoNWPB22uCiBb8jChTp64h2y3XZgZTtoWbwatE9fHzSqpZErxx+6A52DZpBxhvPv/2haO5xImEsmi8USKEy4U14HVufSKKm9sbuYCWauVJMIJk1hxbvCysx+hzljtAaNhaeSBXcEfhR+mqoqdieFhJPNNYldyndxPaipM8pMzjf3D5UqaHqe5sXZmjfmaLJ5hQ6dVVEm+mCDOZN/K5E3Odt8Mz/HIROrIFc4Vp5VXZQs2rFjYYzB5Ck8veyEVXAWJmujh3pRfp182oh2XhInYdx5YRhjPKyyNw31gQ4bHpeGzHvoIKQv6XVQn/2AquaXon1Q+IwepaitcivOUWtbX2Zk++PR3wVnly6IdoH4a7l9KAWNjUbI6O4x4Yzvz493eGmW/srXg1qL2NXcDC++BE3Ly6NS4zm7QqEYrY3IoZRxuYTOLkD7PkOHwqBUD3i5orxkOiZ22X6W0Z2Po/s61ti2+JBTX2+Svbuux7KlGPtOTExXuV48T2N56Si/U6WcXfl8If0SK9/F6lhswkT7QWkTJqftZHGYXxiu1vQ66dYOautsHBs62tyoQpyl80X7YIQMq1BF0TKHXjlGjFBenskTgwTzne2M7nraCDgKEn43vueaiod2kN/IzVFdbcIYVTBWK10sdumaIfhDJ5vrsd8HBPEs5WGXuqRyLdS1vw7A6s7CNV8QPcwPm9hl7uaMU9PPFgWOJJLFn41yxKk8ViLByFFJBg1SHPJx02Gua1x4peJUHCN25cFK4GsVObvC89FReezelVjti/BcaK/ZlfyEQ/AHj4/WYVvmGvSxCudx7XC88fuhk7Xmfc8qiMZ2wqiulXJ2EfvOs2ywrKIiJPbSF7FXzim7VsMwRgVYePiBs2ttrjCV66HDDKM0DjE7nssVHhSF3+k6EEfD814rh+HDjJs7LnahfZJJsAOXsklQb5xdOaeRzOiZZuFY+z1P8+7rq5m/ANJpXSR2ld0ieEVf5MFrwQViJdBhSoPYfqvuZvPwIob2zbVVU1Oey60/ROwShG0I1bGE6r9+kapHvosePI6+U/4fuU98D6oGbdTt5HKaP/1Zc/Jpmueehy+frbjtFsVee4rItaXjOIqLv6HY7qjP8t1Xfoa17BWSfz2nUB1NEISPFI5jYel8keMprJAXil2loYTRxCA2sQ/DpwqOLat8IYB8H56nmfNuDW/Mqdwmk3fENRMaD8hnyKazvPU2tLXr/iduXh7cHDpQmOLfSKViV1+fqRrVHJvrJhwYN7yb7XOz6H3rBVTXCtK9mhrbKFh+7XCwbPI1o+nuMfm5fJVA6TxTd4Rx2xk3UDwsMDJOhG0OJ07ajzqyVGzS/eTDiiYHgdDjVg1hVf3+TJpkM3mymV1a2o1ErKpUcWn7aDWxMEaoHCrquS4ai7p608blLSkWLNQVxa5MBrq6TU6kXA5GhFGEkfsmmIgGE5dwHWHOLsfvI5GArNNIT2pC1D5PpQohZwocK5agPuiLsKeTtovtp3E9bXLAxBzsnqdZtjhHxk2SyYCTtILz2ytydrmuaX99vfk7l4OE14Xj9eL7kHfB9tJUpYiEWu0XKofayquYsybR14xDQXVMuh0k86uLnF2FPFhGQKqpMeLZ6tVEzqGIUpdE6IgZtVvx62vJ9RXPL2eXGH8sVcjZFTrRjDoTC9UKnF0A1VUeY0YrbFuheluprjbhwL5vBLJwuUS6CdvPYldKRh0jdNVN2QGqq1VQZc4zOXuUbUIZCa6rQERS2i84u0rCGN+bb9ZZE4T3ugSOlxJ3e1d3rPpmEH/pD5+GN7rQtx3dFr29GnpbcTNZqmsskknIZvxIIC0dU804aBWuRxVch1FIaJ5UwiR7D5dR+IEo6+G7vjkeqcFoy8ZaPZ+k42FTOAb4fnQclyyF7qEz8YdPM33V3wOCKB4xJub5HlXZpsiN2JcvFru8kTsVQuDwC2JXPsd770PS66S6bQ4TAl2pEMZY/NPCDcREs4DlG5fa8pXGLbgmMSOVBO3lcXHwsbB0llS+Jdr/utwi6tpewVr9Pp5vkbcHYdcUin/oqsH0jphJLg+trQWxK1TpdM0wbMeEzkbViq0E2g4crkX9GVwX4bUROg9j31PZnoz5XgjGLdWzCnyv6BxVysdXiX4rw8Zx0z00d9RRPaSB6uCcz+fBw+SQjB5ghWJX5OxKUFWlGN7oRuH6YUJ7J6Gi9AaWpSKxSysLr2q4+W6PNTib0aTc1eTtehYshGf+rXn1dYq2F+HFBpvouzC4yO1E4bsinn+w6eVC6HyAr8Ehj+MoDtx/YHNOEbsEYVtAa5w37qbm9mOxWt4jc8RP6DvhTvzh0zbJ5s44W/Ob32n2mgF33qY49QuKZFKErq0FpRSnnaLY+4uf4juv/gK1Yg7qT1+C0sS+giBs89hVVaiSMMYisSuWeyckunGOP3wObly9cXsHC/UTxui7Jm+NStDWTkVc30Zp3+RlccFZ8ATeW/8CTKhbfxM3e9lLOPP/RWLhE0EbiH5ace3NNyJCnAP2g4MPgoTKU1sb7LObwcukSdo+3qjd8MeZp9uWVcjN9bGPOUwc57F9Q1M00e7ogNff0MY1EJsUmp2Li12mUfFE1rmcjjXc0Ndnws0ioSx4Sp4fsQd5p4GGBhjcWJj8R2JXVXnC/KgPgBq9mob0XNxMuSji5T20skkljQjYk02RTkPeLxe7SsW64aHYVZpEPMzvEjzF9/xg8uL3oJwkLfX741QlmTghaIOVQllG7LFU4OwqFbvCfcmvIOF1m65JVEfb9jzNosXmmLz9fhXpPkgmg5BPfBy7cD6HQlrYf7kcjOieBQTCTx6qVA9OAlxPRftgBcc2TFRfRKbLuF1iM67alc8xoucF3Ew2poZaRgQMKhwCDGnIB+JJSbW4EmdJ5PRLllTN8/rP2bVgoS5yi5WmRbLsgrPLiYQwXegsO2GEuXCfqmM77maoqTEuuYKLx/x0sib01kv0H8YMBWfXdtuZn74uiD9FIWNBzi7XNVUxqyuEMaZip2wodjW3BVUNS5wjL842/9x8kKdJ2aZCaqxy7BtzVOTgMvqfEbv8WCigo3OFcVMHVRWVYtHiQn9oDToMM/NyoL0ojx6Yrg5DFJ97zqOjAyO41I5AZTqwOhYx2GkrhMj6buGSw2Jlcyx2tJ8HBFFuqKKcXYCdCPLMmdeiUEQNunEi+dF7Bqsu5Ozy84VxJNW7lB3GdnDoIZTlEQuvYZswJ1pB8HPsQs7EHXeo2GQAEo6LQuP6CTzfIuW2M7xndhSGZysXT1u4Ew+ic8xhuHadOY9Dsat6MMO2NwNVX07FRPXg/cbtcYd+jN7UePKpYea1VB3YYYx1UB2wR9PZEY7lYYJ6yzywCHY4n9e8O9+iaXnwub4O7GUvYbXMi5xdOugPrewBObt62vvI6BoGT9+f/KTDgu0Yscs4u3JRlVSIPewJHIoNdYUvnehBjJPC0h6pRNioMFTcCrRaG6t7BarbVI3I9eVQ2mfEdvXU1lLkZIsPWQs/0GS6YtdZ+PAjdDnHw4TDAb2fME7fB8datyz1InYJwlaO6m2l6m/nUPXEj/C224v0F/+Ju8t/l3vSNwCtNc89Xxh4hjTCr36puOonFqNHi8i1tXL0kYrPf++T/Pi960isnkvuj2fCWkILBEHYNvDsGvoSo3CSDpbOF91bhpXU0Tqa9BYROknir4duk6CMeH9il/I9XNeEU/TbNt/kHUmlCq6jSIhIUFns8nLRzXNYEU3Ftl/k7PLcim41pczTbMsOxC4vj86mzRP+mJCQTBYcRdU1FoMGKazmt6JJ9ZtvQ0urCcUpOLtcU9EwFCBiYYyxOSJvzAFPF3+vzl8Ai5dQELsCccOjEKJiJ01/VudXkkybqomOUzlEMZyE1/fOoz77AXSXh3J5eRetbBwnyHEUVEzzKK+gWZY0Ppy4xKvWefkozNVJmIPx4mzzdsLrpXFkLePHGdExfHhWU5fCFJoz6qul/FgOtKASZeN0sw7LZXDfO6YtTgosh2xW8+57kA7mWb5KMWY0jBtvJm9hgnrLMjnePnjbnD+pmNgVTuBcF1a3QZWVxXEg75o2am0SbYPJKVY6R3MW/dsIYjaFczHMH9e9qjCpC3JhKS9v8ulluqhJBSJWyf2cal9cvJGoZGhxInV71Tv9humFYk1IpQT1kbMrFsYYzmK1U4VyM9EYUFPloS0HbdkoN0Ntjem/piCHWSoYFuxsh/n8WuwruZzZbrxCZG12idl6mCgczFjk5Uw7dSYSnH2vMG4lYmJX6BZqWm5cZx2tfYXw6Bgvvxz2afm9dPx46MC21tgILfUH0Fllcg8l7EJouBcNlFZUTU4FzjkrkQRlodxMFEoWYooogBltCqKUP3p3AOyWd5kwAUaNjtSmSFTN23U4oQimKg3i4Q6U5+zyfSPWKaUi92coEoaLhxUMlS64yeKlUpUCK9uFbauyfInLmoyAWpUwObfi7ra4w3BNzi5HBQ9XcPB1LN9YlLtL4/s2pOpxtenURCK2n5ZDfb1i9CgLE4IYriB4X1n4QyehlUO2dgLujp+ERE3BgRnkAHz+RZg/3w+/QArrUFbUWWZ8tEinwV7+Cs7iWdE6ig6L76Gxg4/1f334vqavx0Mpm0ENFk6g8Obyhe8E208XhTEWOsi0rXFQYdC2VNAIp5ppU2G/vcMvJF14SBRbhd30CgD5jDkGDcNS1NUWL1Vwy2oWLIR3386wslnT0lpwZlsdi43InxpUfr8Qr1Ycq9qsNTiI2CUIHxnspbNNAvqml8kc/kMyn7sJXbcOWfvWgu9rnvm35pzzNZf/qDCI/eqXihnTReTaFthlZ8U5PzuU33bcQG3v+3TfeBpdy5vX/kFBELZqWoZ8ktV1e2InTYL1uHBVGsZY5uwKxJZ4iXTl5Uy4TJSkN3ycX55M24hddrCO8pt6T9tY2uSecl3zZDx0ejhO7El0vE3xqlDhawqcdx+OEmkX2lweblYaCqQ15LN5lA4mYLFJQzgBtCxwRgSxOlUN0UQ7noMrcnbpQKgJJkkqFsYYCnl1ddDRCd3dhdtzb+wMVg46CICeDrNi5eUBFQmGlgXJqgQaxajqZuwVr6N6TUiPMb4U93F4rBN2WLWrvD+1a/LHOAlz/HNOIxqLTG15GayKFTOheALjZqKqZkmncE4kEzCoqheruo6pU1SUkPoTB8MuexQmOZaCBJnC5DBwduWdQWgrUciDpRPRdj/o2j46zrvsDHvuU8XOOylS1Y7J36R9qqrNYRjZPYvkMlOIIXTK5HOFdoZVLatrQhExyOHmg2UFE+x+whij6ylosx2UDPRy+SJnF45xdg3reg5n0b+pVsVu6/BcsnpKvqPD68EqhPf5g8aAUqjO8qTYHR3l15xToj1bltme68bfizm7EtWobDfJxc8AUF8bFFxwUuBmo1DQ9+ZDTTUmoXS+D8s3578uEV9aWzX/maVZscK0ra/PhOFG7fEzJLzuQl9FbqAsysviemD7majinucFoj2w806F9YTVGpX26eiApndW0vxueUhluscPrpviqbLWuiis1Di7bHaYrDjkiEH0VE0EIGnno+PluWZfGxsLnwtdc5YFOlEVXR9VKeiummTaqELBS0djnmVZ5jgHBQ2sqhr0iCCKw3dRStFZPYW22j0Kx80qDqkDaGvTtHcUcsbFk2xpDamsEcyrhzaigMGDg02ExptAHVJWYez0Yqq97xMJ+6EAN/tl6OrSdHTCqJEmhFjHxS6/IHbZdkw0r0BN5zsAuH4iEt4ALBWGFPv4wXFyY98dEcEYMahemeMZXhKxfgjPlbxLJHJpJyzOEBvkA5eujsQuVSR2aU30nsrFHE52oqDpaG0SzKsEzasgn+/f2fX8C7CqRVNVY2FZKhqv5rwJfTnzh+OnzThvFYcxgunz6mRhbFNBBUadrCGRUCSsoKhH2H4s087SfGOBJTlRlSSVirkEY0R52rJpWlpg5UpiiqlrxinbjNktrZqO9rCEa2xbdqpofQlVIdn9GhCxSxC2RrRP4oXfUnXvGZBqoO/kv+DudnyFIOn1I5fT3P+A5gtf1Hz/Uk13F3zz64V1r+kLSNj6GDxYcfoPD+KRYb+nXi9H3fYFZj+6qOLTTkEQtg0GN4RxRUbs0iViV5i0tpLYFSZk7+mhIGZ5scpPFJ4oqwrJtD3P5OUKt1WK61vYlkciad6f9y60thavo4xKpc1jX1VFoZdevqIoESwRTd5a3l1MY3qOmSTFctqEoVJVKcBO4teNAN+LJVk35N2YloEXbDsexmj6oKvLrCuclIchcmDWX1tn+rK7K5azy3aKTEFOwmLaFM24ccFrncui8LOyCpRhNbRAWNIlC1gt7+H1daOxsW1oGAR9ydE0DT6KXKIxWi6d1sybp4vEvShfF0QTR6W9Qtih5ZCwC8dq332gJpmPQhtDHEdhJauiEM9EAqqzy9BuDrTGXv6a2ReStNbPjCayeV1YjxpaEOaUUoVJUzjxxKeulrIcOWEer9qON6PXQkfOqFFEoWazX9K0tprE0mB+xr82MxnNsiYdJXl35j8OuV6TD4dysUtbCZSXxVJmJTWucWWF4kroUDEfzqM6lqBWLygc0FiVQt2wnXFMlDgrcznNa69TRm1JBGQoUORyJWJX+FvgdExh4nlrq31zjQSOryFDFIMCwcuEwylUX3shWskrbldHp6nAtyTQ5vr6Cm4i1b2C+sxCANoagkTZThJtJ1DZXnBzZvjR+Ugge+ddk/tKWUbQOOQgmDgBRo4M1olXSKC/7A2gVHgPKn/GM61jCj6UuVJjTi9lO2gsklYuOiyR2DVEMT64PqMQUQdwqk0lW+2TTFlkHHMRGc3EwkhLQXGKMKdUmBsxUVs4n4OxuDs1CdeuLxx6y46E5nRak05rXnnV4+VXKISqheF74cOLYBuj9juAww9TDBsW9FUU4Va4hsLzI5fO0ZcYiRoyjkENVhSybce6a+kyc05VJwsCbZjvSsWcXU5xt5eRSK8ATO41D1PBdfgwSNim/Zby8YM2up7ZHyOk+0X7W1NjhKlIsI8dy9AZ58a/wqLcejEXW1Bps+COU0UCo/mOUxXyWOViCerN92L4EKinu/L9dzarSfcB2mfsGLPCuBuws8f84fh9QRhjUKgjvm07VmESsMLvpOCaVvHvKDBjky78bULDNZl0IN5XpWhoUOw8rXBNlxYjsP1MIcw3vDDCisSYkWXlSnhjToUwRmWxfLnmnXmm2qmtxNklCNs2+TRVD3yd1HO/wp32GdJf+Av+sB03yqpbWzW33a457iTNNT/X1NXClT9U3HW74qgjRODalrFtxWFn7MOKw/+PGqeP3V89hV9+by5LlorgJQjbIjP3TnDg/mA5Dkq7RRN+3/ONW0XrKGeX7+toMhgm7O3pgdUtodiVK3I/xZ/WFxFU+QsnjJVEJ8+3sS2f7cbCrjtrRo401emg4MwqIxDg/KGT8VKN9CVGVU5Qr4340l8eMmJhOR0dJim2Y1M06Q1DsqJqYZYDvkmaG88PlM8VO7s8D1QsZ5dWCt/XtK42ea7CCV5c7NJ2EitImBM+SSffZ0S2mNgFJvxPKWWSKLu5ovChbNakI5i/QNMbmAscx3SKH++MXC/W6vfJ9/SglU3Cgel7wO67mg3FJ34fLIKlTcHTemCfmcZBFbU9yqvkFURPpyoShyAIGYwJf3F0so6clyCTGI4zepIJjXFzprIi4NeNwLOSeInB2EHH52KCkFVt1JJQyInCYVRBhKuqUmUVKy0rEKc8U5igrXYPE9KowLEVjYNh+HATnlVfD8OGFHJ21XTPi9Yz9x0TyhquE8DqWBI5wfx8nkhACsIYrXxP4Xj6xc4uz9XGCWM+jL3yTeyWebEwRjua3Go7Rd5V+PniE703bSb/U0tuG8eOLemD4HDki8IYIVKQg+t7/DiY+bEWHL/XTIrtFCrdBm42CkMLnScq2x2IN6DcdJEQF06Ku7ohl/VIp3X0ebvpVRNuC2T9WGxbsg6V6wYvi+tCwsrjOGB7vTT2vkG614/CqRIJxQ6TVSTMRE5LCmG4+bwJZw3fz+coEj/AiHGuXUdb7R5RoQUVi72zLfCtBI6KhTEGxQzshGLqFMWE7Y270g1EUJ2oNte09khWWVFeJWVU7Kg9ZluBODDIHDC/bnihjZEAY/6O9MTAZeT7mlnPw2vPNLFdxyPYXi+uGxNKg+OwunaG0Z9yhcJFofgZiRjBeWChcRxFMgmZ3hyuVUPjLrsZ4SRwdg2qV0zfA4YOgeVGo6I6WajEBwT56mJiV/9R7ia8PmiPqxP4vkV1NYwapaLvBlv5eEF4Yz5fOAeLnJRAIqkil1rQ6dGvRc6uqCNCsaswaCg0nm/FQiEDZ5cfikNGsA7f9rbbG109GBULYwzPx8FDzEa7+xG7eoJDsuMkn8FDgnEs1mZXO6aqop+J8mxByUMlO4Hy8lEeNkuHzq5AQA3Po6CfNcp81wQCWVeXqQ7btNQcLythvgiHDjWO3IkTgtxuWheKEWgX3wq+MMPv71jBi75smGzOL17GdDBz3Y/TnJqO7ytxdgnCtozqXkH1Padgz3+c7CGXkD3qahNDvgF4nub5FzWX/MDnf47X/OGPmh0mm1DF3/9WccjBqihngrBtM3rGzqgz7iRVk+Ibjadz3Tdf4ne/9/v94hUEYeskkVBUVyuwEzh+H6qvI7g51QxtfpRhnbOM8OOYe9J//weeedZ81o8lafKC35WfLxa7YkmHAVau1EF5cq8ojLGS2OX6Fo7l4TiKUcM9RgxXjB2jghtvKtvBAmeX3ziRzNj90coqfppd08jq2umRs6s8z1Twi+8XJROHYLIfc3aFE7LIDWMlou2H4VtgcqjEc3a5LoVJUhD6ksuZ3amvL0zwXC/WADtp8rgoCzebh0ynCWOzndJ5G97wqeaXqgYjhoTrc80kqTdtBKr5C8zrjhNWFPTBzWGtfBO76VUAVqwwyYwTCXOuNDQU1hUSTpY6Os3P+rriHEuFClt+QfRMVBe78ILXtV1hduukWD74CFrrZpIaPDiYP3qoTIf56LCpphS9ZRw1lgU5P57fxWx/wvaF9YUdNnwYDG4wHbhdTOhR2iT5tm2z/1lnCOnkWFOZMRRtlGLsWNhzhknpMGxoIS9cTe+CaF19fYWk5VGYrJtF+Z5xTObzKO3T26vp7FLgVBXy5WH6fehQGDs6EMc834Q6QrGzJNdjEp3biYJKZSd54y2Lt+cWXyth19fEnFyfPEwxqL74Pi+esL6g5RQap2uN+8hxFEO7XkL1dRixK0jkbnUsJun4WH62IFxkOtHJOrSysLuXFeUeCy9px+uh+amHGdoxizq7I3o/LHrQFxO7dLIOst2o4HpOWnksS9HY9za1uWXo3raylFvJZJgLyo9cg+E57a1axMjuWaTyrSg841gsEbtCYTedHIuuHmraEXMl2ja4Vg0peiOxyQ9+sYOBxQ50EM8NBG6n2uTs8n1SVRZaFcZRZSdQ2o+cXeG4qWuH4U45Ct04kUhl8fJBKHlwjELdIMwNmDXrqM4ZxSnh99DbE4qtobMLMs5Q3IaJeGP3LLQj7owFQglBWSYUcmR+Dpm0h1aOEYmcFMotCELDhirGjy/0Y5UTlJwM+85OYHUuLVQ2XZOzS3vR+6528HRsrPfDcE+fXE6xYKFmaTxKVRc/ITD5FyvLIcmk6dkwYX7YTihxLGsfz1dRn5jYTjsKPTWVRF1QRtTT1Y3oRE1Rzq7Q+eskjVjV2VH+Had1odphVUqXnZsAOT9BImGKNXge0fdW0feslQA/Hzkn/SBHQJSX0ssFsZfhfb8yX7yxsEwwTsohQyhU6NW+ET5j1TcLBk4dubmtDmPfDCupAixbHpxPYSGX+He8MtU0+5JjyOtkUWXbgSBilyBsJVir3qH6T8djdS4j87mbyM84bYPCFltajIvr+JM1F39HM3cunHwy3PMnxbU/s5gxXUm44kcUa/hE1Jl/IjVsJL/e5xyWPPEUx52oufX/ND09InoJwraECpwDqaZZzJ8PTz5tJmdJrwt0IaQvly/kZoqLXU5fkADby6PtJB0dJvdIpLQET4PffBteegXS3a5J3L6mMEbPOLtU+wfF4RYquN+u5OwKl7Mc0MGT9PhXWON2ZJ0h5vNudo3OrtIJspOwiiYWw4cZAWXypOAF24lCiIrErrizCx/XpTAB1D5YdjTRTjj9i12eB75K4GVzRlQA1MidCvO2sOlDd8Cd9mkTQurnI6dYc3Mhj9iuO8P03WGfvSHhhCKKh0q3YnUsQWULbqKGwTaplFl7lMcqplPF+9CplGMnXo0xDGNMVIP2mLojTNmBWAheZSvHrrsY95CybGOYyOewl79u3kzVoQmOnbJJJiHrFpxdGmgedHAUNhhN5iyLUaMUO39MB/umGB6EaVl+NkiMHky8Yo4zK7T0EWs3RCFiCiKXSG+vJhskuE8nR0fikXL7wiaYSab2WfgBvDbHwh9sSg8qBd6YPQAYM1pRVxMTu4JQTOXlyec1b76l6Vm6FIIcTlg2WmtemZNAKxu3JPdPeA2HobhxJ2IcUwXQHDMrdm0UnF1JvFG7lnzIRg+dbNxnbo7GrtmM6XycKm3sbSrfC1XGPeL7FFVCDM/lhNdFTw8kvU7G9s2K3h81UjFyBEXJyElURaJDRleb81n7kXiRy3hlt8lKKXb+mDkuGb+WntT4IERLs3CeyQnm+MZ1tqwJ2juLB4Mod6ANOm1UXr+qELvb2Ah5u4EqvxPtmZ0KwxitUOwKdiFnopFNzi5Mf1i2zU67ODQODvvUATQq6KBcPrZDobAVjE3Ky4Oy+cTB5u3o+gwGtHwuitWLOr2vr1j88X0jjuWH7YSuH1XYVMwZa5YLnF1ao/raaPCXRrmpnASmUEFfO6p7RbSOYUMVo4NV1gRil06E9j1zIlZ5baZ/1+Ts8s0YnXMG41q1+NoqXF9hbjPl4/pWVIihLiz+qYvFvWSi0H+l2Laivr7gzgzRlgPai5zOCt84uyKKc3Z5roelc/TW7IC34yeNYBaE+xacXYUiE0OHmtxmnZ3F99vhuTeoHmw7popjxnMAXyXNAwqvx+Qsix1XMOHy2nLAy7PrLjBuO6irCr4cAvOEynThvPsQqstUl9BBGGP4HReOIXvsnGPUSFV4yKXDfjM/PS8+TGry9iB8lcTqXGqKYQWOXt/XNC0vhMWaxeNf0IX9zPg15vpcB0TsEoStAHvpbKr/chpYCdIn3o034cD1Wo/nmTCG737f539O0Nxyq2bC9vCTHyv+318UX/6SFcWACx9tdP1IsifdiTVmGtfvexFf2e9v3HKrCXG9/c7AoSEIwlaPCkQp7ZsQE8vPmOTdCkCXPWF//Q1NuteNJiPJtndRbQtR2W66MzW89AosWkxQGU6hvHw0KcjlYMliF1fbjNuu+CY8jqctE0rWPLfYwRIkzcb3yOc1z/5b07o6mHBEYpcdTNiKb3GVkzShf4HYVRq6Vpg3+EUTrWHDzNP2OLat2HEHRSIRzrBs04HaZ9x2JkQs4Rh31Ntzg0UCR5txObmEle1CsctxwLKMey0fz9mlFL5vJjFeLh+4eBzUoFGFhOWld/OB0yw8dgs+gO6ewv4MG6YYNEiZxNcKkt2LUdkeSqmpjQk9wWxy/gITEglEzhjTJ2UfL0wifT8WxphC+R7jh3ew/Xiwl7wQrKDy7HbUSBP6pZWNpUyIGoDfMA6UVUj+bhmxK50tqDe+D16irrCy0AkfhjGGM7FYWJStc9gW1LnLSLntWEHlyGWNn6Zn7GH4dUHSJ9+DfAb7vUfNn0MmYlnQp+t57nnNcy8YkUOhyVv16PqRJrw03Rb0J7Q052hqCtx1KEjUsHrwAbQPORA9aCzulCPxG8ZFYZ++56PDaqd+nkxgjmlq8pm/0sT5asuhN+PQ1mEBVlkxBy92vu2xG+y9V8Vup6pjLmM7HqMqt7KgBejiCXbc7RjvVxyTcD2VX236NB+cW24OJ5UCHQgxMedPKMzYQQL7hoYwxxLGtQb0DjeNDceT0FHS26vpzNSainBenkk7FERWq8J5aYRMj5yrcK06Myb4Ln0ZFbQhQ9LrAOCNtwoXl+tqfE1UqVKP3gXXrsVLNUTL7PQx2HlGAwnHxfbTxi0bhjEG+xO/VmybQgXbXC9YFiNHJdguGB+V45jrNDgHsvkKOxTm7Eq3gu2YfHcqlosuOC6h8KljxzCbKXd2gSobU6Jca6VhjMqHfF8khmtskgnQg8aYz/UVK0U77wT77wspFdilgn33Rpr4Z0eZE3TNYYwelgU9yfG4LqStIUUFRky7jPOpcTAcchDsvlv42WI7rHF29T/vaWyEzs6SF4MQxXzWA+0XwhhjIck6SLi+qtnDbjWhzX5M0Deiv1+oHBxcp8q2o8qGfZnizYbfldttV+yKgpgYqRI4jvn+jKe29DwYNhT22xcjLHo5qqoU06YqlJ8vFApQFipjdtjqXhFswipKcxDmIKytcs21WVJJMSpY4BFzrvl4VooVg4wSq9Kro76Kf18r7eO6uijEOax8C5D2aqlyV2O//y9zvQwAEbsEYQvHfv8xqv52Nn79aPpOvBs9dPI6r2PVKs0fbzNCxbcv0bwzD045Gf4SuLgO/nihApIgRFQPpu/zf8Tbfn9OqrmUR75xHdN39/n9HzSfP9E4AyW8URC2buwwXwfmxj/h9UT5iZTvlYX0tbRCwuulpkaxqv4AkzB2lamMla4y2Ze7AnNQ+AQ5Lmj5rkdtnRMliq4cxmhHN8zxKotKBZMx7ZFOm6Thr71uJrv4bjQp1phJV5Grw0pg2xZZ12Hh/DxdJcUbC84uXTTRG9xAWZLqUqLy7r5LKqUYP16x264xNwEUJcTuXJ1h/vsubR2FxMjh5M5xwPOKO93XwSQmvRKrfVHMMVXS9hDbQcXCGMG4E2yb4u96bcQu7fuo9oWF7YUzm34EqFA4K6o6WUG0LIQxeuDn0VYiOkb2ijdMSF/o7ikVTsrWZSZwuIEjpGFsuAtBMQVNKgnpbCJKsl2qzUT9Fs3cQ5uKGx3zYT0vkWp5ncbeNwBIJmPV2RLgb7cX/tAdUb6L6muLRFadqCZXOxZLu1FONKVdsyk7gT1xL3TtsGhdnmucUwsWhM62wI1kD8YPxRPLAdvBVqa0oO9TCPuKicCuC/M7J5HLabASdKaNgGAmkJqubhOeDDFRyTZ5x6qrFSq9GtWzqqi7E3kjUiT8ntgYYATaiHjYcrxfnSRWb0s0B1ZeBnwX5btYqSos5Zv9j+U90toIxLY2xzcudijt4g/dAb/GDBrRuRb0xbImcK1aBg0y/TJkiINjm9Axq8KtrW2bSXUuZ6OVjfbDcGxFVRUMyrxPY/rtYOOFazE836tSQU6iujGsHHRIJASDEYVrB1ehFFTlm/FcXRbGGC8G4DgFd5PKp801ExtvlJPEOLtMEvRcvsLUPS7sxyq0FpxdxWGMIVVJTTYbJfYq9K2yyhLEhwJd2PdhgnqURuXTheWVZfZp0FgjRpYMDEopamsVuBnzfnRNhiGeYV+V72aY8B7fM32ubLq7odcZS/fowM7mZsDLk+3z0ChGjjTOzWjcC7cXjDe2rYrCRktJJQmqksbudS3HjBnvP86o7mdB+7i+8XUGO4lWNitXwptzXFYtNaJMtmpMYR2B8G65ZrAIHU3KcYKKw4UKxCFRsncL4vmuIPbAQylyw3cxIf8xhcr3Tfiu4yjzwMHLQ187qrPJhIWH1SbtZEk+smCvNFG4rudC0vHMQxvbKRe7YufKqnBYCURUbSWN6y9vxE4dPJwCK7gufHP/ERO7Ivch4Nq1Jp+clytyDa6JtXyzCIKwOXHmPUjq4W/jj5lO3zG/Nnk4Bojral6cDf98QPP8C2Zc3GcmfO2riv33Q8QtYWAkasgc+xtST/2EUW/czLX7LuXNL/yUW+9K8Yc/au6+B/77c5rjP69oHCznlCBsdVSZuDutIZHwwDNKlVKAlyt7wj64ASY2dlGXrGVu82C6xhzGoNGm7KDbZmZx0SQrCKeLnu76Oaoyi7GrU4VQh4phjFY0wVZ9bdHr2eqxVOkVKN8tmkM1N0N9tVsQZzRlObu0nSDvQlt3ilw6ByXV56Kq8b5HIqHoS4ygOt8cuIbWcrsclnfvakI3jAfLZsgQxbBhOhKGkrYfhQC2teRI9/rQYRGk/SmIXXYQxuiAH+RF8jyiMEgAv3GC+dmP2KWtBGiNozwIwkW7uqGmmpIF/Sg0NHQ5aSeFzhrBQSWLO2nyJFiwEIK3yeehcbBp+5DGSv0SmwQFBQx0w3bojkXG1RMLUdVr7WPbTHJCsSvo88jZpX2SSXD7kmSzpniA1kHoa/XgQhXMEGUVQmV84xRxrWq0snF6mqLjkUjZEBYrC90r4XlWIjIo24kqlYI5ZjvuAM72NrpWQcwhkvctHJ1HEXN2ERRQiB9P5WApz+T00YCdpK9Pk16dIxEc/6zTCMrCdcEfPo1VK/KQCUJ5tceLs00+tX33KeSnKnIXBe46d9qnC6/5pp8tPxdzdlHUOG0ViwQ6VRf0T6posmp5uYKLy6ki4QThsDFnl++bKpd2p3ktGne0D1qbCXJ4OgX7rYOQTs+DQaPqqKpSuH4eLBsnYXIK6f7ELjx8ZZLB+xpyfS4aRSJB5JgLWh/9FooPVVWQ7iuEc5WJzYEwO7jvHVSrh+ePMGuqIHYVObuCzwK4Ew40+76iCaWNs6vUrRr7UKy5dvQjOgSRs6s4jLGqCryeLqgvbDccV/tzdkVhjEG/WGjI98X0d6swvwjG/4q4mYr7HYZWZ/p80LEQzc5l2CvewJ14ULRjtmPTvCrcRQtcsJvfQnevZPgwn1zeZszo4s36w6eiEzXo0KGJCYf0rBS6ujwHcngeRpUzw7b6efBcHM/Fs1J4nkVnh6a6T5NC0eeZayHhdZHOgnIa0U4831xQ+TCfBhoLzi7LwQ5c1bmSIStK9h5Ua42fD/Fz0EklAvehWa51tTYhs+ExCsRVZ/FzhX6pM+codqLsAVNYjdHb/gCs5rdwvWZSjmvGb8uhcP4FDsagWW1tpoBJ4b1QDE8VHnLEnF22DcrV5rpSBWG2t7twDuXshoL4nij9QquMiF2CsIUSCl3euJlkjvnNgC/qVas09z+oefAhWNViQhZOOwU+8yllKpUIwrpiOWQPvQx/8PYkn/kZu3ev4Orv/Zr5K4dw+52au/4Ef71Xc8xnNSedoBg2VM4zQdhqaBhDZ1Uvo/V7pOwclmvEJdsG3EzZpGfXXaC2qQcv2QDN4JGEpJl05vPmZjcSoqxkEMZo/qzNLzdJeMkWhTqU4vl24YF/EFLhD5lIV8cUans7IdtD3PyUd4ndeBM9KS4iTNSrklH1qTgq0wHJumgmvbp2BkPSr+M4KwYgdpnG2s1z8bDQjSYjerwkfNLxyHs24ONmssbppe2oumG8CpnrKdwph0STQe0HCYd9cMfvBzVDotehUhhj4FpQLqHYBUZMKEL7hdBQAKXwJh6M/8FsoB2vdlTR4hMnwMKFkAk0ilzOiF3TpvUz5sfDBX3XTH7tBP6wKdjLX4+ObbzN/RKIl5afCXLCBS6+aA7lYwVJ5cNzKqgBgLf9AWWr0yqmBgTOrs7qaVjaRak3C+Jj0kblzFQtER7PsK0xd5WuH4lyelGB2DVsKLSv8kgE+YugINSBceop7UYJucMcdmHC/Wi9lo1taeNu8o3QtvADi45knrpsAos8bTW7A4GQVVNFdy5FfR2oPityjISia1gBsN+crL6H6lyC5QVil87H2lPi5A6PgZ3A2+HwgvgXuERqqs1+NtRmCmKjkzTuRZciATIULS2/xNkV9rFllxnyQmeX70PVoEA88PLgu4wZDdmeHMmCplFoduDs0srGVwkT7tq2HFuXi/tOzNkXig/hdRSGJ5ZqUHGhXWe6wQ8cfVFS9IILyHFMP2orYYp8hCsLHm7b9nJAM2aUT+diO8p5VUSRql8QHIqqMQJezrwwdgw0KljqplGti2G0inSIyD1U4qyK69bBSoP/TVh2KEDoeFXVoOpfJZTvFTtHgzaOGuHzTjNMzjyMvWgQ3sSDzNvdpjKAyvVEobzKsSEaA2NuuHQrjQ2DGTzEwi8tsGU56CETi15KOJCbcCje+PJrwglzFeZj46dlo9LthfRfWrNipaIr49PQBztvp+jMDwJgSHUXbsZFqxKncTCns/I9Jj9ezNmlfEXSCaqBxigIkcWhmOa1wnJ2MlnIbwksX25+hnkJ4yKjDgoJ6JrgqYudAmK2Z2UETF+bNuu6UXheMwnHN8cwVhAh7JCwLem++Gp0IVzUqYoq6oIJRdeYB1wK34Rf2gWxPNvRSaLafM9nnWFRYv3KduJyROwShC0Q550HSD3ynQELXVpr3nob/vr/NM88Y25c99sHvvE1xb77iItL2AgoRX7P0/EbxlH10MVU330iOxz7O3542WTOOkNz55809/4/+PvfNZ/+tOYLJ4q4KghbA0pB3hlknF29TTj5ZvJ2PY7Tjcr3FU16pu7oU7vkcRPCFTwJjjuzssHNeehg0nYCq7cFK/8aMD2aWNpW4Ya4qwvmvKnZZ28YNMiMGXnPwgqEhUjsGroj9hKbvFWPynXhxkwl+TworyB2aUArVTy5CCbmvkpi+/HyWqC0V/SUG2VRXWPhZupMmMvawhirBqGTtahcLyqfjiSBZKyNyaRHLpMC+shlPCzt43p2JMzEwxhdF4i5qjwftHKALDpZV3iO3l8UebCvNnnAzNAmT4QhQ0obrguhoRDlbcmN3otVC3oZlKwqWlwpRSqlI+dLPl+YDFZEKYI4STPpDcNMw0TIudikai19TJizSwcVzGJil6UwecEwxzecA/m6yPNSjFVIIB2KXVo5+EG+ovC8t2wLyzaibJiXKGprINa44/czAo/tmEmr9piyo031FA9ricILBAC/YTx2ug1vxE5YK7PYPcuoTprO9K0kWutyZ5edwLIUYzqfwE+B5ylc7WAFTpDu1EQ825wr+bxxc6TTMGECDG+wWfZ24QI11VbXXOlO9a7Cbp4bXZ+WzsdCAUsaFx6DQWOLFZ9A7KquVuy0Ww1OMocfuLi0ncIOznHl5SNFUge5sMKw6igsLhK7ElE7ovPeTuL72kR0JWvAA7vpFdA+tbWKnca4+KPLz4Cw+IBPAicRjBnN71OTAycIoKithWyWIDzNkIs5u6Ag1JdtQRUSppNuQydMeLcVE16SSSOeRcciWQOZTuIFEQCqqhVTdtSkRmhUg8LbscIZHe/7WN6kwkOHwNnleihgzBiF1aWo6snSE4zVeVeBq/t1dkUPJ8JLJtRbgmsvHL+K8l9ZiSJBuAjfKw6DDc6rREJz2Ccg+Z6CmMOocK36UeiwZRfErr5cod+0kyoo3QPgkIMV8YcCccLzsKh6r2WbPFdRzkRNT9oCKzwxLbL5JK5VzZCqTlbrPK4y43A6bfJgplIWumYoNb3zGe700I1RZZXjQM70Q66k6yIhUpVbCuPHy0okIJazrbsbRgyHwUH0hY7NKf2Ru6Cy3ejB4817TrLofDayrCro3JZxkCacIPekioUxUix2RX1WUgHTt1LYQZi0KahhtmHZ5qFOYtnzMGZUsA5NVzrF6EmwZKlZR3La/rDyeSoWqqmA5OwShC2MgtC1z1qFrnxe8+hjmrPP1Zx3geall+HEE+Cvdyt+drXFgQdILi5h4+LtcBh9J9wBboaaP5+Evfg5xo9TfO87Fn++U/GpT8EDD8IJX9Bc9TOfpcskp5cgbMlYFvgqhfYh2WtiDtprdjEJ09sXFW6itY9DLppoqGDyGX+4Gob5ROEXgfPC6lpuBA/t4mtID90jmuSFYSirWsxP39d4OMXulmStcQkoyNv1qFwap9XkCUsmg+36+YLjRlMxZ1fjYOM0sUucXUYUim1PWczcC3baJYpbWXMnJmrwJh2CTtVDrpDoPREXuxyPnGfW5+ZMH3q+Im8q0mMHE+FI7Irh+9A+aG/aanbDo7DS/hPUm+04yuzX5EkwaZKKJjshSntFDoDIGaeS5JzGsnxtYCb62awpeKNZSyJpApHOd4NqnUHbw3Ce+GR2re45C2WZ8DqNVeTiM2GMXnAuJwrina7QN9HOF8QuFYhdvnLwlYOKC6XKifohETYxDFsNnWmBuGMFThVLmwIOlip2YOiGsaZa5pCJTN2tjl12skn5Habamfr/7Z13fBzVvbefM7NVq5W0qu694N4wYGyKIZDEhJ5ceguvIRBIQgqpFy4tBBISIIFggykBQkIgAXIhkMJNoQZMC8ZgGzDutmzLKitpy8x5/5iys0WyZKvrPJ/PWt7ZMzNnzjlTznd+xUc6nW/Z5QgCjniRTGtIoTGsJm0FLfcM8lTaOhckVlypaKnmBjYH+Ovz1rmZG48pCzuCvSyuIaVHLbHLsbLMiROEHiA9fjFm9dSsTUiPAqoXRcBIIOK11rr+sOWq64xxJw5TqpVQagc1lWn8RSHCzqOvY/3lcWN0rzm6n3Qwxu6imeiuj6nngpTrumqjacJ2GxMMG27Hs7IvAVE7zl5FueUZkTZ1NyGP68ZoW/i89Iq1IE9T8VrxpBKEa1daiz0Fy5yoJM5YtcVtmRsHDUE4KC1xqI1zRO5F7HKsrYykaVsnWpULaq1ICf95V/L3fwlWv28LCuQLorkB6t2A4pgIM5PB1mvZJdtzY5RG9ksEN76fzIqB5v7snNTpVvfAfH6PG593W1J2Suxqj0yGXG9lshtn7GgDf0BgCNsS1x/GMKysnEWiASENNzj9iy/DP1+w1jNjY5ASIultBN0slI47J/lujG7Mroyo5lYpS2O03BilHWusuSU7S3CWZVcwilk5MTO29Oz0rJYbo8gIzHZSEL9ugmn3Yc7gcLoiU//suHCfbA1mCWBOnDhdh5LWdYjm3Wh7NgCwOzqf7dGF1FTDgXOtT6CkNGt/e0OJXQpFHyJb6LqjTaErnZY8/SfJmedIrvuhpDUB3/qG4A+/E1xysaYsahTdilkzjZYzH8WMDiP0+6X4V94PUjJ0qOCbV2g8+mvB50+Bv/4NzjpXcu31Jlu2KtFLoXB4+OGHOeqoo5gxYwannHIKr7/+eptlX331VSZPnpz3+fDDD7PKPffccyxZsoTp06ezZMkS/vKXv3SoLkIISyAARKqJlF5MWou4E2Jn4mFN4D1WIpqWbT1A5uHWncg6QW+lFbtESNPKLBipzkpPbtUje13vZEuGyuy6QEvQDvLbagkl4ZA9CU23uu4tVp1y7oNCMGO6ZUGjySTlMRhne7NopPPKBgKCkko7nlmopI3Wy0YGihEesSvLjVEkSBqW9U4qaR102rQEDq9gpOvZrp07dtiuUsEIzcGRWdZcH6xxq5tTD9ulKxnnmKMF48a28UyQ58ZoWyDle8nk1W/TJut7u8IJQCBixYAxUxlLDlsEzcoAuVc3RktA0GQa0xNQ2hugXghL7HImhXkB6rO2Z4tdRgptx2oroLcWzto2ZMY5eN0Y7bhIToBkJxue3ZE+07KIdOOgFTg2vagEv19QlNqOKawNb9psuYhmzfUdIc0+1wzDijMVClhB66XHIiWd8rg66eDz6UyaIBkzKrO5pqb2Lbsca4n0kJmktQiaTGZbduWeV/6i/EbWPANfD4CRRjRsxYwOBd1PIGids+m0RKvfgKjfTFndy5TVv8awGoPJUwJuplPXDU7z5Vt2AYlhh9IcHGlZs+QgzHTeMvc3TKKlOoGwtZ4jZoTDVsbA0lKB32e5V734MjQ0StcSK9eaMT9mVyaboaW7WBvXPOrxAZNhxHAr259TI8CKL5e1LQ1LvW9HvMmqQCbAe26A+nQ6bQm2jjul5g1OpuXHEsvahbAsJx1xzrZ4s+JHmZ5wiVkmRm1bduVkE3Rjc5ltWOvYwqVIt7pjVLMvPppmifnupsw0YLZz8nccV+zyHIYj0mXcGE00XaM5OIKdxfORZSMxDDACJQRkE7qZsKxGc7QZGR1iJzgAn51l1h+wxKNQ0KSh0cp8K6Vk1y6ZidlVwI3R21+a33oxlE5n4s+FvUa6mo5ZOREZKMpkqHU3lC12gceNEUDotnW2CTJtuzFqbjtAJmZXi21AnRuXcPMOb6w2x7IrE6fTm/mxhVIMLUw4DLGYIBbzxHFr5/z2osQuhaKPoK/7616FLsOQPPtnydnnSW68SRKLwU9uEjx4n+DE4wWhkBK5FD2DjA6l5Yxfk554LMF/3ETwT9+GlHVXraoSXP5ljcd+IzjrDPjXC5bodecyk6YmJXopBjfPPPMMN954I5dccglPPPEE8+bNY+nSpWxxAmu0wbPPPssLL7zgfsaMGeP+9uabb3LFFVdw4okn8uSTT3LiiSfyta99jbfffrtjlfJbll2mmZnw58WuMeP4hOdpXbPiannnJo7YZZqWy5T0xNTxG40IW1TSfRnLrdy5TTptPRx7J/wybL3J1TRIEkGGY26Gs1AIUkkDkU64901rSm5mXjgHLbFK18EQVvDsipa3Mq5audYHtugji6tJT/4sZvWUvbWghT+MSGUmkI7YVV20m4CWoNVXTjoNhm3Z1ZLQ2LgpJ1h4jtj19n+sv05/OBMmJztYpKhAqIJABKn5smNi5WLP1LxujM6MNSvrVw66bk1i1qyz69WeGyO2UNjaYIkWjtglBNIXcoMUmxUT9m6F4XEN8wpSrmVXqMQVuzweT+2LXaaBaNiMSDQgohUYWtjNzOYE8y8tyVjoOcJennjl1N1vTeLKmt+1+lS2oxrawdy9P62z9WvvZM/NMCqtgM9NTZbVYsiftuLg5Fh2ed3QpKYRCkJZaWaDzS17scbzBP82tEB2zC5ZQOwqhLd97Mygwky5E2vHeqqpCbSda9G3voVINVuxi8xUlpuVI3JIzefu2it2OeJ4QS+GtoQTYOY0g0mTNYTdZxkhQbgCf0kJTJpsdXoyYQl0AT95GR7zDJE8ll2miWu9JTwnVDAomHKAcC06ZdhSvWRRJdnYJ6g02nb19TaImW3ZlUxKtmyz9pFOmtnWpiJzrSopEVY2S+eYCgxZJ+i9aUriLR7XNdPMxOzySgu6HYespS5/Y7mWankB2XKws/dZiS2sznLiqY0akZtl1rI6yrXA2hdcscurqziWmh7vPEfIbPXb7v222GV5FJqW1WgBQ6QUIUvsMi2xS/drgGDYUEvcqq2Fj9fDG2/Bzp3WOrrIv654Y/D5/FY249ZWePlVe52cpjArJ2GMW5x3gZQescssHWn/rGWsZbHGlU8zLStQb5KKHEtjN9FDznXQ0IJuxlyEZr2cEFomyYGJm62xNaUhyH5xZLnHe9zQ94ISuxSKPoC26TVCT38Dc+isNoWulW9IvniR5PofSooicPONgmV3Cg45WLQdaFSh6E78RSSO+ymJRV/H9/7/Ev7tWYiGzIQ9FhNcvFTjkYcFnz4GHvkNnH625Jk/ycyNTqEYZNx3332ceuqpfOELX2D8+PF8//vfZ8iQITzyyCPtrldRUUFVVZX70T1Prw888ACHHnooF198MePHj+fiiy/mkEMO4YEHHuhQnYTmw7QDxaa1ME42MQefzxKrnGDa1kpWDCVvzC5vjBFrcmBP5CREkhvxGc1INDSfNaEU5Aeot77LbDdG27LLEWak5kMaafch2LTTAzqxoKTEsloSYFZOxBh7GOC4bNpiRmJzZpKa+4a4gMVBR3BjxdjWDMGgYM4sk9mlr+DTodVXTWMjCJm26mZb5VR65reaVjhovzcrGGQmX6NH5ZcFIFTavthFAbFLZItdhR4tdD0Tk81brzb3EiyxJrzSzBaJ/Jm3+2bFhPY3YldG2IPC9FgOOXU3hs8jNWKB62bj/NZ2Ajtr5i7sWFL+cfOYOQOmz7DqGIkIpk6BsjKYOQPGjCbjWteW2BUux9BCgERr2o6+6TW7fIFJt9AQYw+1g+4nsn6qb/B8sSefTr9/tC6BFDpBnzXGci27DK9QadfLp2cPqBpv0Hbv2Jema22n+63g7ZqZcrO/WfXOP5RcvK54UvO5neRksXRiXiW8blre+GqhMszSEXb98i27vKKB0y4FrdXacqFr2YNuJhCaD80WSQwjf7xrmqA8Zk/QTUvM9/vzuzNvjGlalnbjxmFr51oiY6NJjz8Kcq1I7e0Is23LLuGNXeSxBGxshPdWw+o1OomExEgZ1vlqiwR+mRG7Zs+CoNcgr0B7anYcqDVr4aP19r6FJcQ5op23jtJnpUT1ffIStOzJ3lih4xEaYGaJGCK+E9JJ6xoCCCPhWnbptvqsF7gGCSPZpW6M3mueGxvSI3Z5d2WaVkZB6Y8QCtnXWNsiKhdT6uALoNvXgEBQRwqNcHIL0cR6WhNWP0LmZZLrxtjG3E/ToKVkErXFB+cdx94P2CN2DZ2JccAShMeN0TA1uw5mtgDr9B0ZscvNFmxvz7mKGCJEayvU7pSkTeF2tzOGTPv+DZBI6gQC5M1zpch509YOSuxSKHoZrfZ9wk9cilk2mpaTfpkndG3eLPnu902++nVJSwtcf41gxTLBoQuUyKXoAwhB6qCltJ58F1r9RsIPfwHNecC3qawQfOdKjXuXC0aPgh/eJLn8a5JPPlGCl2JwkUwmWbVqFYsWLcpavnDhQt5888121z3ppJNYtGgR5513Hq+88krWb2+99VbeNg877LC9btNBaGBiBY934nn4PROf8eNhbPh9omHP7FTo6Hq+G6MTOscwcB9YpYRguo5IchPSXg+sCZWrs2BNErbvsEQhs2I8RtUB1sTZY5llmoDmw0wbRIxtlLS8j0w28/F6ialnshc62a288Ww0TWBqdhwxDYT9hlxIWzjQMpZH+4TPSdOWaafKSAO6BrJyAmk9wuYtltuk35+xghgxPLOJXMsu73LITLAcy65CEz2w3KFEoj7v7bdWuwaScXei4I3Z5bjneF3hcsl1Xd2rG6P3mcZjBWBUeazl9hac3saxjLEEJdy6ahrWJD9S7i4DMsHrC6FpltuNkbRESj1ATbWgqsYTb0oXCCkpKhJMnOB55soTu6zl/oAg4YtZVoVxKwidWToyK9lAFuEYpSUZN58D51qLI17PIls4qql2d4YpfPg0e8x6xncimWNE4YpdmQ7TNBg5wtMoRmYWr21fhVa/0T52SOplCEz0RJ1XES18LF687ZNlvWO7H9luft5xZHqESan7kJEqq6wtRqL5s1wD3eq3K3YVngxrDVZsQhmpss43LYhhZFszGVWTMYYf6AaVNw3Lssvvz7d4zGsRobvb+ni9JVbsTewC2ghd4qhmRpvrO+7bMhjFGHEgYLnDpg2o3WldZ9JpMNK2ZZfdgIGApKwMJk6whPmQx7usYNwszUoosnET7pgXjoslMGQIzJ3rEbvKRpEefSgIgda0LXtbMp13HrkChkeA1Te+6sZBlLrfitklsy9QzjXILB2RcQNt14e54+i6QBPtW3ZBxg0RrLKGAfjDaJolmscqfFnXdSuxgpWQQvMkArGs1QRaupWq1CpaW6S7bzcgfgHLruw6Q6JkIgl/ZdayjiBDsewFdmZRV+yStsCoebIxYl+HHKvQvH1lXztMLcRHH8O2bfDCyxp77HcyPt0ql/VMkdLyswiDfSPqmNilsjEqFL2I2LOR0ONLkaESWk+52001DNaD7CO/hfsekPh8cOmXrDhIVspihaJvYYw9nOYzHyX85GWEH/siycO/RWrOOVkPGxMnCn5+KzzzLNx5l+S8CyVnnyk592yhxrViUFBXV4dhGFRUVGQtr6yspLa2tuA6VVVVXHfddUybNo1kMsmTTz7J+eefz4MPPsj8+fMB2LlzZ942Kyoq2tymQ8wOGFMSTRJu1mjxBxk/dSijp8fQ3it2y0VKhhAr24YImkg7sIuIxSgpKaaoSBCL+TEMSSiUJFYmqNsjKS72U1w5HSlaSKZ2EAxaD6aGHqa8vJRYTKesLElLq/WAWxzViTcLdtSmiRRD7IBDiEY1pDnHDT5cWpIilZJEy2KEAs2MiK9iXLHOjmCMVFMQX9EQSmLFxOMGRaEAxcURgqUxRCYwDuFIgmAqSKRYI1BaTCQiCOsBIpGI5VqWaIJwKVos56G/A0h/GlkfQRSHEBHbLSm1AxmJYI6YTdEOCLYEkUEfUX+IPXoULRChsjJAUZGw+yTNjlqD0tIAmiaIRKzJflGRoDUhKSnxU1KisWePSSQSobLCTyyWP+mR2ihk6zZEWEfYE0DZ2oBs3QJ6EjH6IGQkgl41nBYzRiS6GkpiaLEY6bRJJJKivNxPWWn2tmOxNHvqM5OMqio/xcXtWKyEdWSdPWbKKxFlMWdDmLveAehwW4toCcFgnKJoBWX2OpFIkkjEGoOBgCQSSRKN+ojFdCKRFMGgJBbLj0Nj1pVYkzNfEPTyrDqYnuBFIhrNGj8A0jSQWz1lYhUI3Uc0KtkcjhBIJYgGBFSPRht/WLvHNHJcNcNmj+LQ4ZZQV1MjCQaz3bLMSISxY6F0SJRXd00l1LiS0kiKYDBNuKgYQlZdNE0Qjfqsvov5KRUNyKYIvooSIhHrHDrm6EDW/Va2NrrnNDKOE7hJKy/HVxogaKwiFkwQLStDRiKI0tK89shFJoPIWuc6UYFsyvxfxGKYkQhFRWkCAeHWKxhIU1QEkYgPUVYBuh9ZH4GiECQiiIoq0ppGJJKipCQz5lsTBpFImspKP5HGHFFR96PFYsj4LuTOdYhRB1kxCvcIqBqBNnIyxi6TdTWfJtbyPJqWJhKxpsaieiSiZAihVmtMRYp9BAIG5eWCklJrbAFEIhHKYtnnoJSS1miEYNAO9i8hFIayimqEL38sttuWqVJkS8Qyh/OH2zhXYsjyU6yMlDZzZhu88aajkoQJhYsIxP3EykuIFhWBabXVAVP9iIlHIQJFmGXFBIMpIpEIsVi+wlASta7XkQggDYItQcpKivGJCPitoSNGVSJC3mjo5Zgtm4CUW3cpJbKoCFFaljWWzGgUosWI0pLMmARo+sjaeOkwaNyOKC5CxiOU6GVE6k1KS63zndiRyPotyE8s373c7bdFbC9lysuTBIPWNQZAJsqRyVqSKZNg0KC4WCdoVBNJW3UuLg5QVJTG75fWfQWI+cqoq4u4bt9FRQECAev6VVo1gVTjKgCqq2PIXVFIJ4hG05hCQKiYSETa54skVgZaXQQRK0dEM3V37hWVlQG27zBoaMxcpysq/ESjHbNxMlsmgi/o9le4KMGOWpg6NUAgCsFgkPLSEJFEBFFWbp3TxcXutdI0rXPGQTN0gq1B91olzAD+RNDedhRNixKJmBRHiwgGDUJhvzXGdD/+PcVEo5m2d+tYUgrhNl4i5KDELoWilxDxWsKPX4iQBs2n3oOMZuzKV70nufknkg8/gqMXw1cuE1RUKDFA0beRsTE0n/EbQs99l+Dfb0Tf/Aatx17vxiYB62H8c0tg4aFw5y8l9/8KXnhRcvV/w9gxaowrBgd5JvlStmmpO27cOMaNG+d+nzNnDtu2bWPFihWu2NXZbTrU1VmxVJqbJTt2tJBOg2b4aGzcgy/ueVNdNRRf/EPMXdvQ7OVGQxPNzWGSSairE7S0SOJxKIlCPG7FF9mchF27plGyZT0J20Ajpftpaqqnrk6QTFrrWHWx1ovHYe5sSKcFdTmhXuJxSUMjNDY109zYRCqVJp0WjCpdz7ZdSXbuSWBoKbZtl+wyYjQ3v0eLGcK7ocaWJJFEgmQSmhvriMeDiNYG4vE4UgYRzXGk4cfI3XlHSLTgi8cxdu1AJi03OW37BkRS0tTSQjwuSSQSJMwmTK2VeKKVRCpOPB4nkRDuMVrtF8fvF8TjGXfDeNyK22QYgrRRSjwep6mpjXhFSWnVZftGpB2zSTTvQo/HkTKEsXs3vnicRiPATqOCxkA1kiJkXR27d1t1aKjHDUTt0NSU6TPrO6RS7YwzI+mOJaOpGSkz7aqnQCTjpDvY1qlEnEQiwc6GJNJep7HROra6OkFrq1W3uj0QDgsaGqSVybIuv35avBktvtOKKSb0rP72jn2zsQEzmF+/rPOjvt61smhuTSJTcZp2NSGjQzHbObZYLEb9kIOwD8BdnpuFzdmXWTGB+MYWRpSkiDfsIZGQNPsSNNtWJYlWqK62xkljI+g0ocfjJHZtQN9Zh+kLE4+Pzuo/Wurc7UufYQUAB9J1dTS1SIpTOi112zDKaqzx1NCI9O+lv4xUps8b4+ju/5uQWPtLpSRNTfDx9jKqIntoabVckuNxgdHYBJoPPR7HpBYtHifd0EhDvd2/deBYiezaZS1raoLsA8M9Dn3tXxFGinR4JPjD6Lu2I8NlmHV1NDZKGlthVzJKJL2VJqoxqg6AdADq6kilrO3vqYPddVBUBLt2WruKRCLE43EaGkDPsYTypZsZPUqyZq31fXd0MaMb40B+HdtDNFp9KFMSGZTtjifiGXfYUBCKwpJaO85To2gm3hAn1NJAU7oB4fR5qAwjnoB4wu2XeDxOXV1z3uabm63MfkNqYNs261rW2FiPv7HetcBL19dDS7ZbuNZqIFp2Z84xM22NpaZmZMBzPYg3I9mDtuHxnOOKI/1hzJSO3tiAqe1AtCRopJF4HOrrM+e4aGp0x5sZrMcMtT9WY7GYex9sC9OUbN+e2YfW2IQWj9NolLHZP54hNaXU7gkSt/dbWxtnzx7L+KgppCESjTQlkzQ2Zfp+R22ccMgaR3uGDSGReAOAPXv2oMfjiHSCdFrSuGMzzcHhtLRYbsqJJDQ2WOeQUd+ATGfEU+deUV/fTFNj/nU6ne7gM3bJJOuv3S7Odt99N87Q8iYSiQTNjbuIp+IYjXGkr87qO60eM2w/U8StbL2jR8Hooc189GyCZr3FvVaV2Q8EzS0tNDdb/dhcopFMNrM9NJuiaCsyXEbd+ibKSvOv4Xq8BZnaQyl7R7kxKhS9QaKR0O8vQjTvouXk5cjYWMCy5lpxn8kll0ma4lZcrmuu1pTQpeg/BItpPf52Ekd8G33dXyl6+POW20wOsTLB97+r8eObBLt3w4UXSX7/hIrlpRjYxGIxdF1npxNp1mbXrl1UegM37YVZs2bxySefuN8rKyvztrl79+4Ob1PToElYbkMl5XbA5uEHZgrodva8dItnJd1Nbw52RkQycY3SBqxbB5u3ZGKOAFlZl7xBZ007QL6ANu95wvFc0PxIT3DcqNiJoYVJ2Q/zH30MKV8ZyUnH5bmQpbUIaS2MroEmUyAlmrQtMPbXjdFuJ+yJn7Z5JVrTdmQgYrkPCYFEIGQaf8B22yE/QD3kuzI6sbmcOCj19ZavR5uxWJz08mlPxjW7XlYQYie+ihXjyqyajIwOtfbhxlApcIh6+9/zV/DE1wpmT02MMYeRnnDMXjaQwYlT1ZLMWJ248ZDI/HW8W6QsHGgbsFyopIlo2ZMVlNmLDJVh2s9n7eJxJxpSo1FdnkIYKTeG3P7ijMtAyM/hi2DMmEynS48TXdqwBC+3Sk7Wvfp1jA6tY27Vf/Ky4+0to1laC9vnfftxgrLwxOwq5NIoI1VoGqzWPs3KPQvYtVtYrlrCU85xkzOSbgww182xDTdGs9gbjMzGNDIZHQ0rdpxINbvXBWd8pA0rBqHUdCtulr1/Z3ynbNc0vx8qKizBx6GtFgkGrV+2lRy5H2PB3rqR7nQMKq9xVCKtA4btxpi5uMiijGWQMeoQxi0+iAWZUE9ZOG1REvW+XMm4MVrVLXBB8BdZAcedjnNO0Nzj0doOOm4Mm+vG+ROte8AXKpid03HpBCCdoxrvI0VF0OzR/pzrhUgnafXXIPwhhEfsNAzro+tWLEFj+DyMcPa9OJX0XmeFm7QBcH/w+cFIJNwxnkza7vdOeIB2xoOZ8yjdUTfG9pASDPvlh1/Y55TjR2+lxM2Utf8GAmRi/tljxuu2L9Hc40uNXMC22DEk/JXIspEQjGKk23CV1/S2M3fmFu1QKYVC0XWkE4SfuBRt94e0nvgLzCHTAdiyVXLZVyX3PQCf+TQ8eJ8Vl0uh6HcIQWre+bT81wOQaiH8yGn4Vj1RsOiCgwUP3CuYfyD89FbJt78r2b1bCV6KgUkgEGDatGm8+OKLWctfeukl5syZ0+HtrF69mqqqKvf77Nmz87b5wgsvdHibQsDOyIGkJhxLrMy673itjd2gtamM2CXtuDTOA7tjjeKIXR9+aGV+y0UK3X0+9opd3glCW/h8Vhmp6RhGJvC2FXcnRCptZR9zCBTSMIRgT3gqug7FW/5BVdPLaDJlCQrO5GFfAxvrfiuQui0waU07rM1J6TkugSbThIKZmF1eC7xcscunW0KXkx2wqQlqayUfrDHcNimIplsZGdMZaw/RtN3eScCdIWq6yEoyAJ7AwoUC1Oc0TUGrsrbIdeHS9Pxl7SB8PgJ+iCczE1ppZgSL3MDI7YXtMWume+qV7bIlbXcwY8j0bOHGu37F+ILLR4/RGTbECa7TBTNMyLSR7rMEFHu7NdVQXJy9D+ecswLU278lmxkxQhCJiPyg7Ub7goChhdHN5uxo3J3Am5TAEXeNEQeyq/pYpPCB0Ni8zVruiheaL3MOphOuqO0GqPc8IqRSlhzk84E54kDSBxxn19NaX9v1oVtWmClX7JO2MO2eb9jZHnPOfSeRhpNZLuC3xIkZ04WnTNvHLwSk9ci+h49yNSWzsJDUDt5rQ2tSR0gTvydAPZCV/VEWVeCPVVNcXLiyUds7MRCAeXOhrEzg08gRuwq4VPtzhHdHbMs7P7SMMJlLMJp56ZJoBH+YUls7L/YKRaESjKGzrXJ7GdsdpShsJV9JpexECyXWSwHDjv+Y27emlaDSGluBCDI6BJ8/u12c+x1Y42fEwoOYutiKY+gE49c0CLd8gmZYJ7WTCTETmK/tQZUbDL8rxC7ThNZW+wWNtO8rwhG7CguVgUDm/JAIjjnaykI6dYqzLCN26YEAwh/IqnvaaCMupdCzRNv2UGKXQtGTmGlCT38DbfNKWj97M8aoBQA892fJ+RdK1n8C1/6P4Hvf1tz4HQpFf8UcPo+Wc36PMWyO5dr45x9AqjWvXCwm+NENgm9eIVj5Jpx3oeTlV5TgpRiYXHDBBTz22GM89thjfPjhh/zwhz9k69atnH766QDccsstXHnllW75+++/n7/+9a+sX7+etWvXcsstt/Dcc89x9tlnu2XOPfdcXnzxRZYvX86HH37I8uXLefnllznvvPM6VCfNDmYdjmRP7KUn05LU/G46cGslK9B82skOaP8N2bpBg23NFSmC+vCkzDaF7go3hcSu9vSBYMB64E+mdZJJCAYyAXHTWhFGGtddctaMfNdObx103ZorBNN1BIw9EIxkAg/vq9glhGXBkcx2VTLLx7oBn6WwLLtCoczk34sjJjkClGlip7a3vr//Abz1TqZ8u1m2fEHXmovWBjQnW66moW/7j7U/v0Y6RZZVremZhOXVz1PlhQva2bcHY+isbEvBfcQcOptkbBIbdpezZ48dzNgThN6dVOW0XUF0v2sxJAPZVjdm2Wi3TJt1KWsjDaa3TzspTrRJjhjnDIbqasHESdmd5Ix/IXAbxMpM51gI5YgJe7F+SWthNCMBKce0pWPPpjJUilkxsbCVl9DQA5nleeeBprvnoDCS7jZETv+CnRQjJ1tb+oDjMIbMtDa1a22msJHMBD+3Z9B+b/U0CiYT0PRMuxYS0NsTskTO2Ow8toWb2XnLLu+1oW6PhsAgFMZVQsyKiW4igI4wYTwMGwpVVdZz28iRAm3X2mwLm0IH6gTet+8fosV2GyyU1tIjzjsYVQdYYyJUmrlG636qqwWHLcy3BJZO6IwuErucPnesl9EDpMceQUullVEiT+wq8OLGa80EVhe4+rEGRKuROQK6JkA3monV/zt75awVC5PrJFEo4UBn2bQZVr1vZ2OUjpWwxxq6DbHLzR7puXZo4WJ3PScAvxDZCVqklBnRMAfpmnnvHRWzS6HoKaQk+Nf/wffh32g9+iqMSZ+hpUXyk59Jnvuzlfb3v78vqKlWIpdi4CCLKmg95W4Cr9yJ/5Vfom1fRevnbkXGRmeVE0Jw0okwezZce73kW9+RnHKS5MuXCNcVQKEYCCxZsoS6ujruvPNOduzYwaRJk1i+fDnDh1tPw7W1tWzdutUtn0qluOmmm9i+fTuhUIgJEyawfPlyjjjiCLfM3Llz+elPf8qtt97K7bffzsiRI/nZz37GrFmzOlQn5+GyKMfTxhh7BMLOhIUvYL3edhBaVkIkJ6mbN6MXQGkpbGmeiG4mKU6sR/fprjWQN8tSRyy7nPLNCR/pdGYSousCqQVIecSughmcbGROJslAeg8yMjljudaOyLE3ZCCCSDa7kzajZiqyuNpTQkOTBoFAxrLLS65llyktAaytuUq7b+x9QYSRACmzrRyMNKJ5FwChoIYprXZz+s6xnCk0b/XOT/0dbCZZOqJjBfe2neJqqqZWseltS0wtK7PnVzmCgiMUtuvGiDXZ13auRZaOzN5P+TjSJcPzRSYvbQhZ0rvDLrLsMqoOQN/0GjIYtfeRmb75A9kH6FggWZlOM9n9ZLAE0VpvuVc61at9H9G8u919J3wV1iR0/Qv2ko7dj40xdnZYr3jhaQ/vuJVYX1KOR6Xmy4hy0nRFskKWXY7YlYe37R2rEyOVEbuEI3YJfLrEZ8YRvnxXW7DOvxb70lBozLcndm2rPA5S++4ZnSVodFIx87p/SaEjpGG5y+02MaNDMasmtbluIYJBwbSp3m1qiFzrmgLnheOGK8w0MtWCvuWtwmU1Le+FqFkxISMC6X6MEQeib/y3awEWChVo2GAUMzoEs2JCh4+tPQq6lgeLMbRMPMWsOsv8e5nfLwgGJAn7MpxOZ65TudayDm7XewTqcWPBcUFvT+zKtezqKqR9sD7s89q9VxYe4EW+jPuq1+XaGHkQ433r2bQ2nGXhpnusxb3WzXloOiJe36E6K7FLoeghAi/8DP+7j5NYcBnpWWewYYPk+1dJPtkASy8UnH2m9cCuUAw4NJ3koZdbFl7PfIuihz9P67HXY0z6dF7RMaMFy+6Eu++VPPIbePs/kuv+B0aNVOeGYuBw1llncdZZZxX87Uc/+lHW96VLl7J06dK9bvMzn/kMn/nMZ/apPqNHQ22tFYsmC38Y6byV1wNkBVe2BaPmFnjrbUm5Hfold+IZi8GWrWDaExt/MDNbdNxiwAoynDZyXFJyCNpizJ4maycBr1WG38fu3VY6c8gX3bxINHRf9ttvs7garX6TvbH9eDz2FyHiO9Fsyyn8uZYilmWXEKKwZZdnYmXas3qhtT3PbS8JgfSFEK178H3wTMZKDxAtu9z/B8PWhltaPGKXPdnYm2VXV7jGdJbKSoFAulYWXkFLCOs3p1/bc2MEkOVjMcrbiMnVntBl7ayN5d1g2RWOYUw8NvNdywx8n30+jRoJGzZCqz0H1TTAl1GvZTCKaK2HZKPVRv6iLBc/u1TerpO+cswR89C3WgG0O63aeOOhec6rLLHLnrSn0hpmdIhlCeSx2nBjdhWI0dSm2OURrI3Kiei1H9iWXfkudOEwJJtK0UStZT2Ug6Z53Bh7ybLL+u++uzFKLGHKetHQnsljJ2pWKMZqoe16TPJE3JMhOHd9oVnifDvbk0UVmCXDMMvH0SZCwxw+r52adw5nqOQKSNLzUsBbzXSq8Iub6mrYaN9iDKN9C1rvckNkrkXjxgKNexe7uiv8rUSnuBh8yTpLfXbjXGa7Mc6YBn7ZRMnWf2I6Lzu8jeQvQtZMgbW4ll2afZ9zRC7XvbHA7ViWDMdsI75bLsqNUaHoAfyv30vgtbtJzj6L1CGX8vzfJRdeLNlTD7feIjjvHKGELsWAxxiziOZzfo9ZMYHw/36N4J//2+MakcHvF1x6scZPfyzYtcsKXv+355Vbo0LRXQwbKpg1UxR+S27jil4OthsjQO3OjBuj9wH/wLm4gXc1aQVHLh2RsXKKlQk+dRRUV2XWL/gW1yZoTzR3t1gT0mAw42opfH721GdiFhWcADvHInR0LWOlYWhBCJV6Zqb7LnbJokqk0NDs+Fgyx0pMCsGI4dZD+syZGnNnZ6+fLXbZy3LELk3A1Ck+aqppHz1gWZlBlquRswwg5LNMDVrsyfyOHZK16+z9FBK7sgxNeue5xR+wAjxDdoB6AK93S7tujPtLW8KDd4f7rnC0iywZhjFsDsbIg9GKSvnUUTB5krVfxwLJcU121wmVAKBvfw9947/xffT3/A234RYkSod6v3WusrmdY5MtdtnnW2wU5vC5Vrksi6Ycscszx020IXbJcAyjeipmbAyydCRS0xGpFtciyGsdFw5DQ2gSdTVHuUHQveh6xpqss5ZdGSG27TLt0kb7dQRH7NI0mDDRx+QJXv/eLhib3thfgeI8d2AXr/+pd51g7psNLU+lyXshIDTMYXOs63UPkWvZtXadpKFBZo1Dr7D40cfWeImVZW9n8iQ46MDMtmQ7FrRS82USbmgZsUsIkbkx5PRhRTmU2c1S4nmRFNqLbt8Zamo0xo4RmWu/17LL03dDhggqSu3YYw1bqK6G6VOz6+t6Vruin8hyY3REsIJujNEhmCPm5/9QAGXZpVB0M75VfyD4zx+Tmnwc8UXf5c5fSH73OMyYDtdeLaiqUiKXYvAgo0NpOe1BAi/fgf/VZehb3qB1yU8wq6fklZ1/oOD+e+DqayVXXyt56x3J5ZcKAgF1zigUPY4vZxLoEbvAejC13nBnsjKFw9YkIBwGX8l4Jk6IIMuGZm1GCEEwmHlIzs1C6CUQsKbaexoDlOul+IoNwHpF7g/6IQEjR0A4tBeLJ3SEEHb9Ja0+O26NM3nYH7ErWoOhz8e34WVrQV6mP0F5zKpbVbUvbxZcSOwSOZYDkyfB2DE6ZaV7uRYWyDIo9UCWS2OgtARBRiTZYnvQVlcVbsNu0m86hd+X8aj1BqgH2w3GseyiO8WutkztMieF3B8LwfbQdGTJsMwu2+knqfks97GS4bD9PWuZv8jKSJiDk5lRFlkmnpMmWBN3sN0gEw37We/Cll0TJugEUuAb6rWK8xyTbWWXm4AALPfbYCFhW2jI8rEZWzU9iLZnA+zZkFeXcNjanwiEc7eSVxW/v3PnxH5bdnl2LjvpFutaxEiIluiAiQH7FOx+b5ixMXnhKVy8mRvtHklPPCbv+iQ1zZVSzfKxYKSQbcXG60Ecgd80rSD16z+BzZthzBhruabBlAOsQPbrN1gCbFlpfiwxIQSlpaBp0nLbb8OCNj3xWBACI/kusIlQqpZQchta6RCrgOM6mtOHc+dk9jd2LFRWWfeRcDtWzp0lECDbANQZk7qv4DXFqq9JTbXAqBDZqzo5KIyMi7LPlwlF0K4bYyfoA7cshWLgoq/7G8E//zfpMYexce4NXH6F4HePw2n/BT+/VQldikGK5iO58Ku0fOF+SDYTfuQ0/G88UNDuurJScNtPLTffPzwBX7pMsnmLsvJSKHoar2WXGRtjBZn2PIQmk5m320PtZ/JAwMrWt+hQwZz5YWRsbEH1ocgzx2wv1ogjUKUN2FV2KIw/HOfJ+4ApPg46EA6YLBg9uv17a661QKvfNpFyUrrvr0iRFZQ78/8FB8PEiZ5H7wKCSSGxS9OyxQFfR2NlFRC7vBPM9OTPooVLCIUyYldLiyV0zZpZuA17w3UxF7/fChYtpcwTtIQn5kuuENaltKWiae33b0/hCGDGmEWkxx5u9bu9zMkm5yCDJe7/zUgVxsiDABg9WrD4SHsdOztrVpKKDpIevdAKVu9pM+daUVYK0VKdcFige2e1Xqs0X3bWO8eiJpGwAliHCmtUWRjD52YHY89xY4SMJUkuTbb3dkV54d87YtnVK26MjvudtC1MnWQE0uiSE8PIeknZznOZ1//ULdaOuyNgxsZhDp21X/ETuwpvLEDHndXJuAjW4QUCgrEej+hoO+74Ph9s2QZvvpW9fRfdD5qPVNVMkrp1bs4qW8mi2XZ8PecEaEf8FEJQEhXEytq32O4sbb2MMqPDrCyZrfWQakarXVMgnlv2gbrtamR+0vXMeZjOziWxzyixS6HoJrSN/yb09Ncxh87kX0Nv5Ytf8vPRx3DDtYLLL9U6l65boRiAmCMPovmcP2CMPYLg339E6A8XI+I788r5fIIvXaRx842CbdvgwqWSvz3fNVl2FApFB7Etu2Q4hlkzDSBP7HK+T50Chy/quJtb2DNZbc+yy7vPUMh2dbJF8kDYT+neLJ2wXCunTctUfE94Ki1+ayLfkSxXHcIrMnkma8XFgqKI49dUeB/O8ZlesUtkT4j8HX349+WIXUJkzxzsOnjFruaWbPExl3DYsqSp7ngSty4nYOdKcLvL0+3eAMe5Lo49QnfE7OogB0wusDAQASewvZPJzhOPzBh1CMaQ6Znyur/g2DTLx1mxkkqH5/22V8JlecHQndh8DQ1kRDhRWOwiV+zCsy7ZLlttEiq14oE55Fp2kbEoaYup+QboWfXyYgydhTFkRpdadnVW+MnK1OoLuckChGngBOjfH2T5OEwneHy7mfEcM8OMZVfB61+W62ofUNVtvC8g3AQQGplDsfvI28fRdsakrlv3S4c2Y3b5BKZmXcM1Adq2/6DtXGPF3oMebyO/z8rGWQgZLgNApFrRN620sqAmmrLL5Aic3ucDr2VXXswuZdmlUPQ9tB3vEX7yUsyy0dxZfwdXfDdEZSXcs0xwxOFK5FIoXMJltB5/O62f+h/0Ta9R9KsT0Nf+uWDRQxcI7l0uGDUKvnJFIz+/0ySdVlZeCkVPIANWoHXTk1mvLbFL0zqXRdUbc6fjYpdbM+tfrWMTwVhMMGRoZqI37fCxHDTf8VNx3EP2V+zy1CVvJmxP7NuYqDjxO9NesUvPnhj4OjpPzXMT8rkCTG7MopaWjKVMuB2xKxAQHH6YaNPyqyfw+60g0F4x0EF4Q8f0itjVexP2kSP25tZqi1wel2SpB3Lq3Mbg0nxWrKRAbsKFfcOxkiovB/x2rCdvXECv+55dX9c10+73hkbrbGovqYWXrODz3vFvN0dyL+/Q2rKQKTTGZOkIZNmo/Y/Z5RUIOmlxmhWg3hdCmCnXuquzLpFtYl9PRHtil9ckT5rZy7x4j6+7XID3Ac1ze3BiGwphienea0/WNbqd21GueNOW2GW5TzoqUACRjKPtXIvWsKVHrUbDISgtgSOPEMRiORdbB7u/RKLBzeC8t2QDkGk/pw28Mbu6yo2x74wkhWKAIOrWE3p8KWl/GV9/Yzn/eruUk06Eyy/t3MO/QjFoEIL0zNMwRswn9KfvEP7jV0lNOYHE4u9DqCSr6JAhgjtuh3sfCPLgQ62sWiW55mqoqVbnlkLRrQSjpMcfnRXAOVfsak8kaY+SEuuN8ZatHRe7XIHMETY64+vgWJKEY5REPdcO2UYQlc7S3kTE+a0dqx/ngd+JTZRrIFcoSHYh8twYNZ9n/5k6hsNWnJkm+0V8e5ks+wJ+f45lV46+5IwhszvdGNvCW5ketuzaK/6QlRQmS1Twg5nKfO+hSbQQgsMXSSv4uz4Vs3x8dnB4b9vlWHY550VLi+XC2OEET0GPuY1n4u2M97bG/YzpkGht71j2/ltXuDHmJrvYG5b4IqmqJNO2Kdsns4vELkeIbFc8cxvI9GRwLCR2eYT4Hlep28a556xZl3kfIrGvP21Usz1rpFzxpk2xS3f2BLK4Gtjs/tZlYmUHWLSw8EGmvRli7euGtnNNZlmey3P+doQGGJ7QX3omWY2TWbbDL3faQIldCkUXIho2E378QlJpwfkv3M2GpmquuVpw9OK+c9FWKPoqsnwcLWf8Gv+/lxN45ZfoG/9N4tM3YIw+NKuc3y/4zrciTJ6Y4MabJV/8f5L//j4ccrA6zxSKbiUnU5k3M18y2XELi1yEEEybClu2Sior2y/rTCIygo+jCHXukTY97sj8oPuu1UE3Tvg7MPvVdWtS1dycXbSi3A7I3dHsWrlil5RIYQeB9kyWqirhw4/gjbes7/s7uehuAn4nWLT13SsG9rYbY9YktC9E8/dgVB2Avul1ZNBj4aT7MwGvoUet0bJeAOdmQcxyY8y27DLdmF1tBKdvizZdhwXz5kgibRitDalpfxD1WDbGfbB2OnyRfT4nrDYUSVvs6iIhVpaOwJAmsmxkO6W8/qcFfI8dnOPrQy6MkO3GGCmy+rSpyRbT90Hsyv2trbKaRibuVagUWjNiV28K6UbNVPAV5Yjm+fUR6RyFuA3LLoNMGzhC4CuvShqbLIuy/U1K1beuwgpFP0Y0bCH06PkkGpo49y/LMEtHc+9yJXQpFJ1C85E65FJazvgNMhgh/PiFBJ6/3nobncORRwhWLBNUV8O3viO5e4WJYSi3RoWip/A+pEvaf8DvCIcvgunTOrZPJ26VWXWA9Z8OujG6BCL5D+gBy51K6l2Yqz0X54F/L5ZdaQP+86713Zkwz50jWHCI6LglS64liBCZY/bsPxoVWTG4Omo51ls4LkKO21mWLuLLBDaWZi+7MfbChHTWDJg8sY0fwzGMiceAL4DpZHPU9GxRtBeD6mchhBUsf+gst05u1kn7Nt8p4dfGLK4p6PJcXt557wundE9lY+z0NQ5LUNR14VpguWJXl7kxCisLY0esWaXZbhx717W6r4xBGyGEK2rFYjB8uHUYqVTb15eOiF0jhztiZOGNaBq4+QuDOUpsLwqCMjbWTVjhUkiITeWaQ7adydR5cea0TaNtZVxTve/1dOjj724Uiv6BaNyK/zfn01LfwP97YQUzj5nCJReL/VajFYrBilkzjZazHifw4m34V96P7+N/kPjU/2CMXphVbsQIwV13wG2/kDzwIPznXcnVP8hP+axQKLqe3Af6/bUI6sxk09mXjI0m3VbK+05iVk5ChsuhqI20a53AGHWIGxA8i064MTrs+0RZsyzYAG37KmR0KKJ5l73R7P2PHQM7aq3/9wfLLoCNm6y/WZZduiWCmKbsEgG203gnfb3gilXdQZd+c+gsN9GEVxQtmMGzlzDtrJBeNJFxY0wkoKKik9sccWAX1Mxi9mwfb73tEeEKkIljuK978Q7u/TgxHStWO55Sj8bEcsRKadJuIL2+5vZbgFAoY32USrWty7UXZypabF1rNb39e56u41obC39uspE+1lYFGkLkvqQuZNmVI3Ll3ntKsiOZ7BN9SzpVKPohonEr/Oo8EnX1fGXlCs795jS+ermmhC6FYn/xBUkecSUtpz0IeoDw4/+P4LPfgZa6rGLBoODKb2j89/cE762GLy6VvPGmsvBSKHqaQA9YBDkhX9oLALzPCM2OjbL/yKIKCMcK7cT6087s16dnXPTAE3B9XwhEIBDBHHmQ5WrkWnZl799rzdXXxS6nrtu2W3+9h+L3W5ZdjnVXdx6LMWw2xvAc8cQJtt7XEVq+myu48bH6LMKa/6fTkrTRecuurmTYUJ0FHQyfsM/jcD/dGDPr6kjd73Fj7EEJwD0GJxtjW+ZQzvH1vfmTI7CGQ5m+TKXarml7InuRbaTVkhvSKgfLsssSu/SOBvrqA5gV4zPx5bIsXQsIYjm3w9zD2tfQCF76bkspFP2AVO0mEnefh9FYz4+23cP3bp3G4Yf1vYu0QtGfMYfPo/nsP5A85Mv43n+GyP3HYb79eN4M8NPHCu6+SxAtga99Q/KrhySmqUQvhaK7yH2gz3353B048Xr2N0NTr9EByy5NywTnhX0P/N+Z/XvFrg67SfYSgVwjB091HTdGR+zqTssuWTI8351HCGR/EbwK0Ycsuwqha2CYGRfW3LHQ13AeU/z7qFMVtA7dV3whV+zqyQDnLlKCNNs+JuG4MfZclTpLJJIjdu2DG2NlBcTKYNzY9veladDiHwoUsuzqu28kzNKR7nVE7sUCzRuYHiBlX7dLojB5khWjd39RYpdCsY/sWLUWee9ZaMkGnojezbd+PJ0hQ/rwFVqh6M/4AiQPvYzmc36PGRuD8diXCf3+IkT95qxiY8cI7v6l4JhPwfJ7JFd+V1JfrwQvhaI7KC4WTJmc+d4TsZ7cyWMfjyvVFtLJBNnOJFbXodV+6z9rRufcO/eKvd/cyW5bcWP6Irl973Vj9NnxztJdlLZ+XzDGHmZlLu2HdGu8ui7A77dEBkf07nE31U7iuCPvjwdiVyF9IU/G2R6ukNAAuZcUhn2gkfZCNCrcarbnxthuNkaf4MB5gmi0/WuurgsaQhPZXHoMmj+AMeoQzGJLXJd92LILzZexcNXat+xyrt1Oe1WUW6NjyhQYNbJr7kl9uKUUir6JlJJ//e4dyv94LoYJ7x30IKd8eWa/elBUKPorsmICLac9hHb8j9C3vknRA8fjf+0eMJJumXBY8IPvCr79TcEbb8AFSyXvrlKCl0LRHQwdmvl/pzKj7SPuJLfvz4vawPHbaD9ml3PF6nJRz3Vj7L/PLLkuYVkB6u32SrQWLtsjaL787IL9BV/fNpUKBCyRwRWR+rjY5Vyv9tWyq0vNnPweE9EeDgIvhebJdttWGacz+9616aADYcHB1v99HiuktrIxthfHrTNMmCAYMy5AICCQRRWYVfbbpZ4WKzuDpmesW7Puc3sPUB8KCT51tKBkL0Jgp6rTZVtSKAYBW7dK7v3+3zn84wtIalFaTnuYmUe1lfZGoVB0C0JDP+h8ms97GmPMIoL/uoWiB09G3/BypogQHP85wbI7BQE/fPlyyYr7TNJpJXopFF2J1+WtRyy77L9tTTL6PK4bYfuWXQ5d3aYyGLV2n8zPcNtfEEK4E0/Iseyy54COG2hfF0P6HH3cjdHvt1wYXdG7j89kXQvD/YzZ1RWusTJUmvnS42KJsN0Y27Hs6sMCTmmpoLjYqre3L7v7ncHYMYIJ470XONvyso9lrMxC8yEdYdX7yN1OgPru9Krtwy2lUPQdDEPy2O8lT139EJdVXEZrZDxFlz5E+bgRvV01hWLQIqM1tJ5wOy2n3A2mQfixLxL83ysQDRnXxokTBSuWCz79abjvAfjSlyWffKIEL4WiO+gJsWvKZCgrhaL+GhbJeeBvJ5ZJxHNsXS52hcqs3Sca8n4LBKwAzP0BZ+IJOZZddrNusjM19vVg+30Fo2YqsguykHY3fj80NsFuO09NXxczTUfs2sfzWBh2pgr//gfuk96EGT0ds0sIN2ZX29kYHavXvn3SZl1Tevqlix6wLnh9vI0IFLpB5zeWyHFj7A6U2KVQ7IW1ayVfvixJ8Pnr+cYBP6Rl9KcIXvQAdFHGJoVCsX8YYxbRfO5TJBZdge+jv1N03xIC//wxtFqTuUhE8L1va9x4vWDbdsut8dHHVPB6haKrmD0TSkustOzdTVmZYP6BAq3/mnZZf9qJuTJsmFXKp3eDgOgPI30hjJppeT8dvggWHtrF++tGnOyf3pHgtFeTnXSur4shfQUZG4sxakFvV2OvOH3+0cfW377ev8Z+xo6T4TJkUUXB87XTBIoxo0MxS4b1TswuaQKybaukQASzYgLGsDk9WrXOomkiY5HUC7chY9gczLLRPb/jTiAj1RhVkzGHzswsLCByOgkmuvM87uOyoELRezQ3S+7/leSvT9by44O+yazS10kceCEc9vW+bT6qUAxGfAFSB11EesoJBF76Of7X78P/7uMkD/4SqVlngi/AYYsE06fBzT+R3P4Lyb9egCu/CSNH9NdJs0LRN6iqElRV9XYt+gkdyMbo9wuOPMIS47tD1DMmFA6e3lVxZnqKykrYsjU7MW9u5kpl2TWwSKayv/flON3gyR67r+NQ82GMOqRrKiME5vC5XbOtfdj3XgPUQyYmVR/H57PcaXPH34zpmUyh3YWMDuneHXQFmo6smGD937HqK3B/CdpembIb3z338UuEQtHzGIbk6T9JzjhHsuq5N3nsmP9iRsUqWpf8hNTh31RCl0LRh5HRISQ+fQMt5zyBMWQmwX/cRNEDx+F7/2mQJrGY4IfXC773bcG6D+G8CyS/ekiSSikrL4VC0QN0IGYXWBm7VOKb9jlgspWevqwssywUym6zvm75o+gc0Wj2977evzHbc3DQi65CeCy7ersy+0+gDYvbITWiy7II9jfS448iPfbwvOVSOIO/gNhlW3Z1p0CoZu0KhYc33pT8v4slN9+c4pIpv+S+w8+nuCxM65m/JX3Acb1dPYVC0UHMqkm0nrKcls/fiwyWEHrmm4R/fRr6xn8jhGDJZwUPPyA47DBYfo/kwotUxkaFQtH9mLHRmBUTMEtVzM/9RdetiWV7Fmn9zVpN0T4jR0CsLPO9r4td06fBwgXdY6HZv+hAgPp+hGORpC4vHvxhCEbzFhujDsGsGF8wTlwkYq/ajWLwYNeZFQoA3lstWXGf5NV/w4Fj1vPX075DrPk/pKacSGLx9wuevAqFou9jjFpAy1m/w/f+0wRevJXw784jPXohyUMupXz4XK65SvCZYyU/+Znkksskx39OsvRCQaxMPcEoFIpuIBjtN646/ZUZ06E5jnKtHYAIISiOSOr2WN/7uhujrov+m0yjKxEabmq+AeAh47hL57pNKwoQKsEMlRT8qaJCMHumpKKi+3avxC7FoOb99yUr7pe8/ArEykzuOOc3HBq/BcwQLZ+7DWPSsb1dRYVCsb8IjfSU40lPPBb/24/gf20FRb89i/SoQ0gecikLDpnPg/fBfQ9IHn0Mnn9ecsH5cMpJVuwchUKhUPQfhtSo6/ZARrdnr5qmLPf6DW6AesFAsOxyRFYlZO4/VVXdOx6U2KUYdJimJW49+phk5RtWBqkffPF9TtSuwb/jHdJjjyBx7HXIiHolqFAMKHxBUvPOJzXzdPz/eRT/a/dQ9Oi5pEcchH7IpXz5Swdx/OfgF3dKfn6H5Imn4NKLYdFC9UCtUCgUCkVfwHFdHPSegf0KgZASKRgQvn+O250Su/o+SuxSDBoaGyXP/KmFXz0k2bQZamrgiovq+ULVLwm/+zCyqJzWJT8hPXnJgLgQKxSKNvCHSM09l9TM0/D/5zH8r91N+LHzMYbPY+z8pdz8w8N49TXBz++QfPcHkqlT4OKlMG+uui4oFAqFQtGbOMHeVZTNfoSdkc96iur/z1KjR0MwBNXKLqLPo8QuxYAmnZa88m947s+SF1+EZKqZGdPhkgsTHFX0EKGVd8P2OKmZp5Nc9DUVm0uhGEz4gqTmnEVqxufxrfo9gX/fTfiJL2HGxrJw7rnMu+t4/vz3IlbcL/nq1yVz50jOOUtw4Dxl6aVQKBQKRW/gil1K7eo/CA0wQWoDwqBA1wXDh/V2LRQdQYldigFHU5Pktdfh5VclL70Ee+qt1L8nnwSnnSAYvu1e/G/8Ci2+g/T4o0ksugJZMb63q61QKHoLX5D0rDNIT/88vnV/wb/yAUJ/u4bgC7dy0sz/4phfnslTf6/hoV9Lrvim5IDJcM7ZcNhClWFJoVAoFIqexKfcGPshworZJeSAELsU/Qcldin6NVJKtmyFNWtgzVrJf96F/7wLhmHF4jrkYPjUpwQHTd5B+J0HCTz1KCQaSY9eSOvnfoY5fG5vH4JCoegr6H7Sk5eQnrwEbcub+N/4Ff7X76Vs5X2cNf4oTvnRqTzzwaE8+IjG9/9bMmY0nPYFOOZTEAqphzeFQqFQKLobJzj4iOG9Ww9Fx5FCINIJRKoVGY71dnUUgwgldin6DaZpxdr6YA2sWSNZs9b6f1OT9bvfDxMmwDlnwSEHC6ZMMghsehn/f36H/vL/Wb7iM04kPvNszOopvXswCoWiT2MOm0Ni2BySDVvwv/VrfO89QXTtn/lCdCgnfukk/tl0Cvc8PoybfiK54y447rOSk04UjByhRC+FQqFQKLqLigqYMR1qqnu7JooOIzREstn50qtVUQwulNil6JOk05ING7EErQ8sYWvNWmhpsX4PBGDiBMuiYvIkweSJMGaMlR1D2/Eevg+ewXf/n9Aat2IWVZCaex6pWWdQNmY6Zl1drx6bQqHoP8iSYSQP/ybJhV9F/+jv+N99nOBryzhG3sVRnz6E9Z8/ngdeW8zjfyjht7+TzJsrWfIZwRGHK2svhUKhUCi6GiEEQ2p6uxaKTqFcFxW9hBK7FL2OlJbF1urV8N5qyfsfwNp1kEhYv4dDMHEiHLckI2yNGgU+n33hNNNoW9/G9+q/8K35E9qeDUjNjzFmEYkjvoMx/kjQA712fAqFYgCg+zEmHoMx8RhE4zZ87z2B/93fM2HD97i22M/3LlrAyw2fZtk/F3PdD0u55VY4erHkmE8JZs7wXK8UCoVCoVAoBhNC83wxe60aisGHErsUPU4iIVn1Hrz5luS91bD6fWhosH4Lh+GAyXDyiTDJFrZGjLCyXriYBtquD9G2volv/YvoG19BJBqRmg9j5CEkD7qY9ISjIVTaOweoUCgGNDI6hNTBXyJ10MVo21fhW/MswbXPcVTD91k810fd0Qv4x46juffFI/jj09WUlcJhiySHHy6YNwcCASV8KRQKhUKhGCT4w5n/m0rsUvQcSuxSdDvptGWt9cabsPINK4h8Mgm6BuPGwZFHwNQpgqlTYPSoHGELoLUefcPb6FveQtv6Fvq2dxDJOABmyXAroPSYRRgjD4FgcS8coUKhGJQIgTlkOskh00ke9g3LhXrNc5SteY6T9f/h5MNhT9E0Xqs/gt+sPJIrn55CUURj/jzJ/PmCg+fDkCFK+FIoFAqFQjFwkaEyzxcldil6DiV2KbocKSUffgSvr4Q33pC89Q40N1vu2hMnwCknwby5glkzoagoZ6InTbSda9G2vIW+9W30rW+h7f7I+knzYVZNJjX1JMyhszGGzUaWDFd+4AqFovcRArNmGsmaaSQXXYHY/RG+j/5O9KP/41Mtd3HMQXeSCFTxXmIhf/nwEO77xSH8OFHFyJGS2bNg5gzBzOkwbJgVj0ShUCgUCoViIJCVgVEavVcRxaBDiV2KLmHHDsnrK+G1ldZfJwb8mNHwmWMtcWv2LCgtzZnEtexB3/YfS9Ta+hb61ncQSSu9ollUgTl0NqmpJ2EMm41ZMz3bDFahUCj6IkIgK8aTqhhPav6F0LIH3/p/oX/0d2Zv+AdzJjzBlRNgj388bzYs4M+rDua2Z+fTlI5SUQ7Tp0smTxJMnGi9IKisUOKXQqFQKBSKfooviDF8LvrmN0DK3q6NYhChxC7FPlFfb7kjvr5S8trr8MkGa3lFBRw0H+YfKDhwLlRWeiZpRgpt+xpb2HoHfds7aHXrAZBCt6y2ppyAOWw2xtDZyNIRympLoVD0f8JlpKccT3rK8SSkibZzDfonL1O84WWOlI+zeMZDyJk6O4Mz+E/Twfz5owXc98IsUqaVWKM8JpkwASZNhIkTBGPGwIjhEAyq62N/5eGHH2bFihXU1tYyceJEvve973HggQfudb2VK1dyzjnnMHHiRJ588kl3+e9//3u++93v5pV/5513CAaDXVp3hUKhUCg6i9Tte5FyY1T0IErsUuwVKSUbN8I778K771oilyNuhcMwZzaceIJg/oGWJZcQAqRENG5F/+BttG3voG99B237KoRhpVg0I9WYQ2eRmn4qxpCZmEOmg7+o9w5SoVAoegKhYVYdgFl1AKkDLwAjibb1HXwbXqZ8w8ssTtzDUZOWIaeGqS+ZxXpzHm/smsv/rZvFb38XJpWy3ogKAUNqJCNHwqiRMHKksP9CdRVomhLC+irPPPMMN954I1dffTVz587lN7/5DUuXLuXpp59m2LBhba7X2NjIt7/9bRYsWMDOnTvzfi8uLubZZ5/NWqaELoVCoVD0CZyMjErsUvQgSuxSZNHQIFn/Caz/BD7+2Pr/mjVQb2dLLCuDGdPguCWCGdNhygHg8wlIxtG3vYv22jvoW99G2/Y2Wtx6GJe+EGbNNFKzz8IYOgtz6ExkdEjvHaRCoVD0FfQA5ogDSY44EA69HBJN6JtfQ9/wCtFNrzOr9pfM1k0umOLDOHwqu6PzWG/O5Z36eazdXMqGjfDsc9AUz7gFBAKWEDZkCNTUwJAawZAaGDLE+lRWFEgEougx7rvvPk499VS+8IUvAPD973+fF154gUceeYRvfOMbba531VVX8bnPfQ5d1/nrX/+a97sQgqqqqm6rt0KhUCgU+4xmyw5K7FL0IErsGmAkk5J0muyPAS0t0NQEjU3W36YmaGiA2lrJjlrYsQN21EI8ntlWUZFlqXX44TBjmmDGDMt1RmveiVb7gfX58wdotavRdn+EsC9eZvk4jNGLSNrCllkxEXR/L7WIQqFQ9COCxRjjFmOMW2x9TzRZyTo2v46+eSVVH/+aauM+DgLMIaMwZk3DrJrKnqKprE9M4aOtZWzcJNm2HbZtgw8/hF27s+Nj6BqUlEoqyvdQXGxSVma9yCgrhbIyQVEYgiEIhyAUgmDQ+vh8lpA2bKgSyvaVZDLJqlWruOiii7KWL1y4kDfffLPN9R5//HE2bNjAj3/8Y375y18WLNPc3MzixYsxDIMpU6bw1a9+lalTp7Zbn1gs1u7vir6N6r/+j+rD/o3qv44jkwHkjggIDa0PtZvqw4GNErsGEK+8KrnyuxKzg4K5EFBeDtXVMHoUzD/QsgAYMwbGDm+h2r8JrX4j2p4NaHs+QbzyiZUpsXmXuw2zZBhm5WSSkz+LOWQWxpAZECrpngNUKBSKwUawGGPMQowxC63v6STajlXom1eibV+Fvm0V/g/+RA1QA8wvGY45ZCLm1HGY5danNTKG7Y1lbNsG27fD9h2Suj3Q0qKzY4fBhg3wzjtQXw+GuffAsdf+Dxx1pBK89oW6ujoMw6CioiJreWVlJbW1tQXXWb9+PbfccgsPP/wwPl/hx7Zx48Zx4403MnnyZJqamvjVr37FGWecwZNPPsmYMWParY+ifxKLxVT/9XNUH/ZvVP91knQSn21Vke4j7ab6sH/TEaFSiV3djWlYH5m2/59GmOnMcjMN0kAYaauMk6FCSkBm/TWbStAa6gv/DswqlVx/gfVd1yS6LtF10HVJMABFYUkoJIn4mwlrTYRoQks1IZJNiPguRLwWLb4D8e9aRKIx+zDC5ciykRjjFpOqmoxRNRmzcrISthQKhaIn8QUwh83BHDYns6y1AX3HarQd71mfXR/i3/AKIt0KQBFQFi5nUukIZHQo5tghyOgQioZOpFGGkKEyZDiGGSihqUWnpQVaW6A1Aa2t9idh3bKEgAWH9M6hDyRETvIVKWXeMgDDMPjGN77B5ZdfztixY9vc3uzZs5k9e7b7fe7cuZx88sk89NBD/OAHP+iyeisUCoVCsU9oem/XQDEI6bdiV/jh/0LbtQ40zQp4Z3+k0K2nceevpgNaphwCKYT7f2s97O+ava73d2H5FjvClC1cCa9YZS8XWaKWLWzRdelVDaxJS1sUAUv2YbsyEEEWVSAjVRiVk5CjFyIjVZhlI5GlIzHLRkEwum+VVigUCkX3EirBGHUwxqiDM8ukiWjcirb7Y7TdH6Lt/hjRsAWx+0P8n7yASMYL3lOiwVJkuBQZiIAvjPSHrL++IPhCSF8IXtKz75MIz31U2Pe/lOflThqMNJipnO9phF0O07Tvn6b1Asg0QDovi+x7sLTUtsRnbsIYfWhPtW6XEovF0HU9L8D8rl27qKyszCsfj8d59913Wb16Nddddx0ApmkipWTq1KmsWLGCBQsW5K2naRozZsxg/fr13XIcCoVCoVB0CidAvULRg/RbsSs16zS03R9blk3Sfhi2P8L5v2kAMvOwLCVg/RVS2gHyvNZTZqacpyxCt0QzzQeajnT/73N/k/ZvCB/o1nLpWccpK3VnncxyqemZZcKZODhveIW7rDgapampKVPG87uEnHXt3wtsB38YGYwiA8UQKFIXH4VCoRhoCA1ZMhyjZDjGmEX5vycaKSNO4/b10LoH0VJnf/ZYn1QzpFst67CWPWjpVut7qsUSpPDeNz0WydK0721+636nZT7Wd3/OMj/4wu59E6F57omafV+0/kqhWQH9y0b3YEN2LYFAgGnTpvHiiy9yzDHHuMtfeukljj766LzyxcXF/PGPf8xa9utf/5pXXnmF22+/nREjRhTcj5SS1atXM2nSpK49AIVCoVAo9gV7TioDxb1cEcVgot+KXenpp/Z2FXocLRbDUH7FCoVCodhfglFEbBRGUGXG7WkuuOACrrzySqZPn86cOXP47W9/y9atWzn99NMBuOWWW9i+fTs333wzmqblCVYVFRUEg8Gs5b/4xS+YNWsWY8aMcWN2vf/++1x99dU9emwKhUKhULRFeswi6wWXQtFD9FuxS6FQKBQKhaK/sWTJEurq6rjzzjvZsWMHkyZNYvny5QwfPhyA2tpatm7d2qltNjQ0cNVVV1FbW0s0GmXq1Kk89NBDzJw5szsOQaFQKBSKzhMq7e0aKAYZQkrZdUGlbAplNVDZDvYf1Yb7j2rDfScejzNy5EgANm7cSCQS6eUa9V/UONx/VBvuP6oN22awpCJX/d9/Uedv/0f1Yf9G9V//R/Vh/6Yjz2oqWJNCoVAoFAqFQqFQKBQKhWLAoMQuhUKhUCgUCoVCoVAoFArFgKFb3BgVCoVCoVAoFAqFQqFQKBSK3kBZdikUCoVCoVAoFAqFQqFQKAYMSuxSKBQKhUKhUCgUCoVCoVAMGJTYpVAoFAqFQqFQKBQKhUKhGDAosUuhUCgUCoVCoVAoFAqFQjFgUGKXQqFQKBQKhUKhUCgUCoViwKDELoVCoVAoFAqFQqFQKBQKxYBhn8Suhx9+mKOOOooZM2Zwyimn8Prrr3dovZUrVzJ16lROPPHENss8/fTTTJ48mUsvvXRfqtZv6I42bGho4JprrmHRokXMmDGDz372s/zjH//o6qr3GbqjDe+//34+/elPM3PmTI444gh++MMfkkgkurrqfYbOtOGrr77K5MmT8z4ffvhhVrnnnnuOJUuWMH36dJYsWcJf/vKX7j6MXqWr2/DRRx/lzDPPZP78+cyfP5/zzz+fd955pycOpdfojnHoMFjuKdA97TjY7iuDgX29dyq6l2XLlnHqqacyZ84cFixYwKWXXspHH32UVUZKyc9//nMWLVrEzJkzOeecc1i7dm1WmWQyyXXXXcfBBx/M7Nmz+dKXvsS2bdt68lAUWP05efJkbrjhBneZ6r++z/bt2/nmN7/JwQcfzKxZszjxxBN599133d9VH/Zd0uk0P/vZzzjqqKOYOXMmRx99NL/4xS8wTdMto/pvECI7ydNPPy2nTZsmH330Ublu3Tp5/fXXy9mzZ8vNmze3u15DQ4M8+uij5Re/+EV5wgknFCyzadMmedhhh8kzzzxTXnLJJZ2tWr+hO9owkUjIU045RS5dulS+/vrrctOmTfK1116Tq1ev7s5D6TW6ow2ffPJJOX36dPnUU0/JjRs3yn/9619y4cKF8oYbbujOQ+k1OtuGr7zyipw0aZL86KOP5I4dO9xPOp12y7zxxhtyypQp8q677pLr1q2Td911l5w6dap86623euqwepTuaMOvf/3r8qGHHpLvvfeeXLdunfzOd74j582bJ7dt29ZTh9WjdEcbOgyWe4qU3dOOg+2+MhjY13unovv54he/KB9//HG5Zs0auXr1annRRRfJI488UsbjcbfMsmXL5Jw5c+Rzzz0nP/jgA/m1r31NLly4UDY2NrplrrrqKnnYYYfJF198Ua5atUqec8458oQTTih4jVR0D2+//bZcvHixPP744+X111/vLlf917fZs2ePXLx4sfzOd74j3377bblx40b50ksvyU8++cQto/qw73LnnXfKgw46SP7f//2f3Lhxo/zTn/4kZ8+eLe+//363jOq/wUenxa7Pf/7z8qqrrspa9pnPfEb+5Cc/aXe9r33ta/JnP/uZvP322wuKXel0Wp5++uny0Ucfld/+9rcH9MSkO9rw17/+tTz66KNlMpns8vr2RbqjDa+55hp57rnnZi278cYb5RlnnNE1le5jdLYNnclxfX19m9v86le/Ki+88MKsZV/84hflFVdcsf8V7oN0Rxvmkk6n5Zw5c+Qf/vCH/alqn6W72nAw3VOk7J52HGz3lcHAvt47FT3Prl275KRJk+S///1vKaWUpmnKhQsXymXLlrllEomEnDdvnnzkkUeklNYLvWnTpsmnn37aLbNt2zZ5wAEHyH/+8589ewCDlKamJnnsscfKF198UZ599tmu2KX6r+/z4x//uN1nftWHfZuLLrpIfve7381adtlll8lvfvObUkrVf4OVTrkxJpNJVq1axaJFi7KWL1y4kDfffLPN9R5//HE2bNjAZZdd1maZO+64g/Lycr7whS90pkr9ju5qw+eff57Zs2dz7bXXcuihh/K5z32Ou+66C8MwurT+fYHuasN58+axatUq12Vs48aN/OMf/+DII4/ssrr3Ffa1DQFOOukkFi1axHnnnccrr7yS9dtbb72Vt83DDjtsr9vsj3RXG+bS0tJCOp2mtLR0v+vc1+jONhws9xTovnYcTPeVwcD+jBNFz9PY2AjgXvs3bdpEbW1tVv8FAgHmz5/v9t+7775LKpVi4cKFbpmamhomTpyo+riHuPbaazniiCM49NBDs5ar/uv7PP/880yfPp2vfOUrLFiwgJNOOolHH33U/V31Yd9m3rx5vPLKK3z88ccAvP/++6xcuZIjjjgCUP03WPF1pnBdXR2GYVBRUZG1vLKyktra2oLrrF+/nltuuYWHH34Yn6/w7lauXMljjz3GE0880Znq9Eu6qw03btzIK6+8wvHHH8/y5cv55JNPuPbaa0mn0+2KjP2R7mrD4447jt27d3PmmWcipSSdTnPGGWdw0UUXdfkx9Db70oZVVVVcd911TJs2jWQyyZNPPsn555/Pgw8+yPz58wHYuXNn3jYrKira3GZ/prvaMJdbbrmFmpqavAfngUB3teFguqdA97XjYLqvDAb2ZZwoegcpJTfeeCPz5s1j0qRJAG4fFeq/LVu2ANY92O/3570cqaysZOfOnT1Q88HN008/zXvvvcdjjz2W95vqv77Pxo0beeSRR7jgggv40pe+xDvvvMP1119PIBDgpJNOUn3Yx1m6dCmNjY189rOfRdd1DMPgiiuu4HOf+xygzsHBSqfELgchRNZ3KWXeMgDDMPjGN77B5ZdfztixYwtuq6mpiW9961tcd911lJeX70t1+iVd2YbO+hUVFVx33XXous706dPZsWMHK1asGLCTkq5uw1dffZW77rqLq6++mpkzZ7JhwwZuuOEG7rjjDr785S93ef37Ah1tQ4Bx48Yxbtw49/ucOXPYtm0bK1asyBJqOrPNgUB3tKHD3XffzdNPP82vfvUrgsFg11a8D9GVbThY7ynQ9WNxMN5XBgOD7RrdH7n22mtZs2YNv/71r/N+K9R/e6MjZRT7x9atW7nhhhu49957271fq/7ru0gpmT59Ol//+tcBmDp1KuvWreORRx7hpJNOcsupPuybPPPMMzz11FPccsstTJgwgdWrV3PjjTdSXV3NySef7JZT/Te46JTYFYvF0HU9T9nctWsXlZWVeeXj8Tjvvvsuq1ev5rrrrgPANE2klEydOpUVK1ZQVlbG5s2bueSSS9z1nKwJU6dO5dlnn2XUqFGdPrC+Sne04YIFC6iqqsLn86HrurvuuHHjqK2tJZlMEggEuvfAepDuasPbbruNE044wXV7mjx5Ms3NzVx11VVccsklaNo+JS/tk3S2Ddti1qxZPPXUU+73Qm8+du/e3alt9he6qw0dVqxYwbJly7jvvvs44IAD9ru+fZHuaMONGzcOqnsKdN9YHEz3lcFAV40TRfdy3XXX8fzzz/PQQw8xZMgQd3lVVRVgWR5UV1e7y739V1lZSSqVor6+PssyYdeuXcyZM6eHjmBwsmrVKnbt2sUpp5ziLjMMg9dee42HH36YZ599FlD915epqqpi/PjxWcvGjRvHc8895/4Oqg/7KjfffDMXXXQRxx13HGDN47Zs2cKyZcs4+eSTVf8NUjo1ew8EAkybNo0XX3wxa/lLL71UcAAUFxfzxz/+kSeeeML9nH766YwdO5YnnniCWbNmMW7cuLwyRx11FAcffDBPPPFE1o1+INAdbQgwd+5cNmzYkJVedf369VRVVQ24CUl3tWFra2ueoKXrOtJK5NB9B9QLdLYN22L16tXuzQNg9uzZedt84YUXBuQNorvaEOCee+7hzjvv5J577mHGjBldUt++SHe04WC7p0D3jcXBdF8ZDHTVOFF0D1JKrr32Wv785z/zwAMPMHLkyKzfR4wYQVVVVVb/JZNJXnvtNbf/pk+fjt/vzyqzY8cO1q5dq/q4mznkkEPy7j3Tp0/n+OOP54knnmDkyJGq//o4c+fOdeM9Oaxfv57hw4cD6hzs67S2tuZZbTnzOFD9N1jptBvjBRdcwJVXXsn06dOZM2cOv/3tb9m6dSunn346YMWX2b59OzfffDOaprmxBhwqKioIBoNZy3PLlJSUFFw+UOiONjzjjDN48MEHueGGGzj77LP55JNPWLZsGeecc06PHltP0R1tuHjxYu677z6mTp3qujHedtttHHXUUVmWDQOFzrQhwP3338+IESOYMGECqVSKp556iueee46f//zn7jbPPfdczj77bJYvX87RRx/N3/72N15++eWCrhgDge5ow7vvvpvbbruNW265heHDh7sxBoqKiohEIj1/kN1MV7dh7nkNA/+eAt0zFgfbfWUwsLdxoug9rrnmGv73f/+XO++8k0gk4l77o9EooVAIIQTnnnsuy5YtY8yYMYwePZply5YRCoXcmDTRaJRTTz2Vm266iVgsRmlpKTfddBOTJk0akHEf+xLFxcV595iioiLKysrc5ar/+jbnnXceZ5xxBnfddRef/exneeedd3j00Ue59tprAdQ52MdZvHgxd911F8OGDXPdGO+77z5OPfVUQPXfYKXTYteSJUuoq6vjzjvvZMeOHUyaNInly5e7qndtbS1bt27t8ooOJLqjDYcOHcq9997LjTfeyAknnEBNTQ3nnnsuS5cu7Y5D6HW6ow0vueQShBDceuutbN++nfLychYvXswVV1zRHYfQ63S2DVOpFDfddBPbt28nFAoxYcIEli9f7mY5Aeut2E9/+lNuvfVWbr/9dkaOHMnPfvYz13puoNEdbfjII4+QSqX4yle+krWvyy67jMsvv7xnDqwH6Y42HIx0RzsOtvvKYGBv40TRezzyyCMAeWLyjTfe6LrGLV26lEQiwTXXXEN9fT2zZs3i3nvvpbi42C3/ve99D5/Px9e+9jVaW1tZsGABP/rRjwbkS7v+huq/vs3MmTP5xS9+wU9/+lPuuOMORowYwfe+9z1OOOEEt4zqw77LD37wA2677TauueYadu3aRXV1NaeddlpW3GXVf4MPIQeaf5ZCoVAoFAqFQqFQKBQKhWLQMnAibisUCoVCoVAoFAqFQqFQKAY9SuxSKBQKhUKhUCgUCoVCoVAMGJTYpVAoFAqFQqFQKBQKhUKhGDAosUuhUCgUCoVCoVAoFAqFQjFgUGKXQqFQKBQKhUKhUCgUCoViwKDELoVCoVAoFAqFQqFQKBQKxYBBiV0KhUKhUCgUCoVCoVAoFIoBgxK7FAqFQqFQKBQKhUKhUCgUAwYldikUCoVCoVAoFAqFQqFQKAYMSuxSKBQKhUKhUCgUCoVCoVAMGJTYpVAoFAqFQqFQKBQKhUKhGDD8f5M7ZZnqe+5eAAAAAElFTkSuQmCC\n", "text/plain": [ "