diff --git a/docs/deep_dives/lv_uode.ipynb b/docs/deep_dives/lv_uode.ipynb new file mode 100644 index 00000000..52d21991 --- /dev/null +++ b/docs/deep_dives/lv_uode.ipynb @@ -0,0 +1,915 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "id": "a0000000", + "metadata": {}, + "source": [ + "# Universal ODE for Lotka–Volterra: recovering the unknown interaction term\n", + "\n", + "This **deep-dive tutorial** demonstrates the **Universal ODE (UODE)** approach for the classic Lotka–Volterra predator-prey system.\n", + "\n", + "**Key idea.** We *partially* know the dynamics:\n", + "- The prey self-growth term $\\alpha x$ is **known**.\n", + "- The predator self-decay term $-\\delta y$ is **known**.\n", + "- The interaction terms ($-\\beta xy$ for prey, $+\\gamma xy$ for predators) are **unknown** and replaced by a sparse polynomial $U(\\Theta, x, y)$.\n", + "\n", + "**Goal.** From noisy observations of prey $x$ only, jointly infer:\n", + "- The sparse coefficient matrix $\\Theta$ of the unknown interaction term.\n", + "- The diffusion $\\sigma_x$ and observation noise $\\sigma_y$.\n", + "\n", + "**What is known.**\n", + "- Known linear terms: prey growth $\\alpha x$ and predator decay $-\\delta y$.\n", + "- Initial state distribution $x_0 \\sim \\mathcal{N}(\\mu_0, \\sigma_0^2 I)$.\n", + "- Observation operator $H$ (observes prey $x$ only: $H = [1, 0]$).\n", + "- Polynomial library $\\phi(x, y)$ (all monomials in $x, y$ up to degree 2).\n", + "\n", + "**What is inferred.**\n", + "- $\\Theta \\sim \\mathrm{Laplace}(0, 0.5)$ — sparse interaction coefficients.\n", + "- $\\sigma_x \\sim \\mathrm{HalfNormal}(0.1)$ — diffusion coefficient.\n", + "- $\\sigma_y \\sim \\mathrm{HalfNormal}(0.5)$ — observation noise std." + ] + }, + { + "cell_type": "markdown", + "id": "d5157a56", + "metadata": {}, + "source": [ + "## Lotka–Volterra Model\n", + "\n", + "### True (fully known) ODE\n", + "\n", + "The full Lotka–Volterra system is\n", + "\n", + "$$\n", + "\\dot{x} = \\alpha x - \\beta x y, \\qquad \\dot{y} = \\gamma x y - \\delta y,\n", + "$$\n", + "\n", + "where $x$ is prey population and $y$ is predator population.\n", + "\n", + "### UODE decomposition\n", + "\n", + "We split the drift into a **known part** and an **unknown interaction part**:\n", + "\n", + "$$\n", + "\\dot{x} = \\underbrace{\\alpha x}_{\\text{known}} + \\underbrace{U_x(\\Theta, x, y)}_{\\text{unknown interaction}},\n", + "\\qquad\n", + "\\dot{y} = \\underbrace{-\\delta y}_{\\text{known}} + \\underbrace{U_y(\\Theta, x, y)}_{\\text{unknown interaction}},\n", + "$$\n", + "\n", + "where\n", + "\n", + "$$\n", + "U(\\Theta, x, y) = \\Theta\\,\\phi(x, y), \\quad \\phi(x,y) = [1,\\ x,\\ y,\\ x^2,\\ xy,\\ y^2]^\\top \\in \\mathbb{R}^6.\n", + "$$\n", + "\n", + "The **true** interaction is recovered when $\\Theta$ correctly identifies the $-\\beta xy$ (prey) and $+\\gamma xy$ (predator) terms.\n", + "\n", + "### Observation process\n", + "\n", + "Both species are observed at discrete times $t_1 < \\cdots < t_n$:\n", + "\n", + "$$\n", + "z_k = \\begin{pmatrix} x_{t_k} \\\\ y_{t_k} \\end{pmatrix} + \\varepsilon_k, \\qquad \\varepsilon_k \\sim \\mathcal{N}(0, \\sigma_y^2 I_2).\n", + "$$\n", + "\n", + "Parameters: $\\alpha = 1.0$, $\\beta = 0.1$, $\\gamma = 0.075$, $\\delta = 1.5$." + ] + }, + { + "cell_type": "code", + "execution_count": 1, + "id": "05966cc0", + "metadata": {}, + "outputs": [], + "source": [ + "import jax.numpy as jnp\n", + "import jax.random as jr\n", + "import numpyro\n", + "import numpyro.distributions as dist\n", + "from numpyro.infer import SVI, Trace_ELBO, Predictive\n", + "from numpyro.infer.autoguide import AutoDelta\n", + "\n", + "import dynestyx as dsx\n", + "from dynestyx import (\n", + " ContinuousTimeStateEvolution,\n", + " DynamicalModel,\n", + " Filter,\n", + " LinearGaussianObservation,\n", + " SDESimulator,\n", + ")\n", + "\n", + "import matplotlib.pyplot as plt" + ] + }, + { + "cell_type": "markdown", + "id": "e1f8ce55", + "metadata": {}, + "source": [ + "## True system\n", + "\n", + "Define the true Lotka–Volterra drift. This is what we are trying to recover — but we only give the model the **linear** part." + ] + }, + { + "cell_type": "code", + "execution_count": 2, + "id": "88b6e4b9", + "metadata": {}, + "outputs": [], + "source": [ + "# Lotka-Volterra parameters (fixed; not inferred)\n", + "_alpha = 1.0 # prey self-growth\n", + "_beta = 0.1 # prey-predator interaction (unknown to model)\n", + "_gamma = 0.075 # predator-prey interaction (unknown to model)\n", + "_delta = 1.5 # predator self-decay\n", + "\n", + "state_dim = 2 # (x=prey, y=predator)\n", + "\n", + "\n", + "def lv_true_drift(x):\n", + " \"\"\"Full true Lotka-Volterra drift.\"\"\"\n", + " prey, pred = x[0], x[1]\n", + " return jnp.array([\n", + " _alpha * prey - _beta * prey * pred,\n", + " _gamma * prey * pred - _delta * pred,\n", + " ])\n", + "\n", + "\n", + "def lv_known_drift(x):\n", + " \"\"\"Only the known linear part: alpha*x and -delta*y.\"\"\"\n", + " prey, pred = x[0], x[1]\n", + " return jnp.array([\n", + " _alpha * prey,\n", + " -_delta * pred,\n", + " ])" + ] + }, + { + "cell_type": "markdown", + "id": "9cf77423", + "metadata": {}, + "source": [ + "## Model configuration\n", + "\n", + "| Symbol | Role | Value | Status |\n", + "|---|---|---|---|\n", + "| $\\sigma_x^{\\mathrm{true}}$ | SDE diffusion (data generation) | `0.05` | **inferred** |\n", + "| $\\sigma_y^{\\mathrm{true}}$ | observation noise std | `0.5` | **inferred** |\n", + "| $x_0$ | initial prey/predator | `(10, 5)` | known |\n", + "| $\\Delta t$ | observation spacing | `0.5` | known |\n", + "| $T$ | time span | `[0, 30)` | known |\n", + "| $H$ | observation matrix | `[1, 0]` (prey only) | known |" + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "id": "809bd8b3", + "metadata": {}, + "outputs": [], + "source": [ + "sigma_x_true = 0.05 # true diffusion — inferred during model discovery\n", + "sigma_y_true = 0.5 # true obs noise — inferred during model discovery\n", + "\n", + "obs_times = jnp.arange(0.0, 30.0, 0.2) # 150 observations (every 0.2s)\n", + "\n", + "# Observe both species: H = I_2\n", + "H_obs = jnp.eye(state_dim)\n", + "observation_dim = state_dim\n", + "\n", + "# Initial condition: near (10, 5) with small uncertainty\n", + "x0_mean = jnp.array([10.0, 5.0])\n", + "x0_cov = 0.5**2 * jnp.eye(state_dim)\n", + "\n", + "initial_condition_kwargs = dict(\n", + " initial_condition=dist.MultivariateNormal(\n", + " loc=x0_mean,\n", + " covariance_matrix=x0_cov,\n", + " )\n", + ")\n", + "\n", + "key = jr.PRNGKey(42)\n", + "key, k_data, k_svi, k_filter = jr.split(key, 4)\n", + "\n", + "# Minimum diffusion to keep the EKF covariance well-conditioned\n", + "_DIFFUSION_FLOOR = 1e-2\n", + "\n", + "\n", + "def make_state_evolution(drift_fn, diffusion_coeff):\n", + " \"\"\"Wrap a drift into ContinuousTimeStateEvolution with isotropic noise.\"\"\"\n", + " safe_coeff = jnp.maximum(diffusion_coeff, _DIFFUSION_FLOOR)\n", + " return ContinuousTimeStateEvolution(\n", + " drift=lambda x, u, t: drift_fn(x),\n", + " diffusion_coefficient=lambda x, u, t: safe_coeff * jnp.eye(state_dim),\n", + " )" + ] + }, + { + "cell_type": "markdown", + "id": "73eb3b80", + "metadata": {}, + "source": [ + "## Data generation\n", + "\n", + "Simulate from the **true** Lotka–Volterra SDE using the full known drift." + ] + }, + { + "cell_type": "code", + "execution_count": 4, + "id": "5751c879", + "metadata": {}, + "outputs": [], + "source": [ + "def model_with_true_drift(obs_times=None, obs_values=None, predict_times=None):\n", + " return dsx.sample(\"f\", DynamicalModel(\n", + " state_evolution=make_state_evolution(lv_true_drift, sigma_x_true),\n", + " observation_model=LinearGaussianObservation(\n", + " H=H_obs,\n", + " R=sigma_y_true**2 * jnp.eye(observation_dim),\n", + " ),\n", + " **initial_condition_kwargs,\n", + " ), obs_times=obs_times, obs_values=obs_values, predict_times=predict_times)\n", + "\n", + "\n", + "predictive = Predictive(model_with_true_drift, num_samples=1, exclude_deterministic=False)\n", + "with SDESimulator(dt0=1e-3):\n", + " synthetic = predictive(k_data, predict_times=obs_times)\n", + "\n", + "obs_values = synthetic[\"f_observations\"][0, 0] # (T, 2) — both species\n", + "states = synthetic[\"f_states\"][0, 0] # (T, 2) — full state\n", + "times_1d = jnp.asarray(obs_times)" + ] + }, + { + "cell_type": "markdown", + "id": "d5c23d6f", + "metadata": {}, + "source": [ + "### Data visualisation" + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "id": "09e15d86", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "fig, axes = plt.subplots(2, 1, figsize=(9, 5), sharex=True, constrained_layout=True)\n", + "\n", + "axes[0].plot(times_1d, states[:, 0], label=\"prey $x$ (state)\", color=\"C0\")\n", + "axes[0].scatter(times_1d, obs_values[:, 0], s=10, alpha=0.7,\n", + " label=\"prey $x$ (obs)\", color=\"C0\", zorder=3)\n", + "axes[0].set_ylabel(\"Prey $x$\")\n", + "axes[0].legend(loc=\"upper right\")\n", + "axes[0].grid(True, alpha=0.3)\n", + "\n", + "axes[1].plot(times_1d, states[:, 1], label=\"predator $y$ (state)\", color=\"C1\")\n", + "axes[1].scatter(times_1d, obs_values[:, 1], s=10, alpha=0.7,\n", + " label=\"predator $y$ (obs)\", color=\"C1\", zorder=3)\n", + "axes[1].set_ylabel(\"Predator $y$\")\n", + "axes[1].set_xlabel(\"Time\")\n", + "axes[1].legend(loc=\"upper right\")\n", + "axes[1].grid(True, alpha=0.3)\n", + "\n", + "fig.suptitle(\"Generated Lotka\\u2013Volterra data: both species observed\")\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "id": "fefca0e2", + "metadata": {}, + "source": [ + "### True-model marginal log-likelihood benchmark (for convergence monitoring)\n", + "\n", + "Before fitting the UODE, we compute the marginal log-likelihood (MLL) of the observed data under the\n", + "**true** data-generating model using the same selected filter (continuous-time EKF).\n", + "\n", + "This gives a practical reference for what a \"good\" likelihood looks like so we can monitor SVI\n", + "convergence more efficiently. However, this is **cheating** in real applications: the true model is\n", + "unknown, so this benchmark is only available in controlled synthetic experiments like this one." + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "id": "7ab708f7", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Marginal log-likelihood under true model (EKF): -229.015\n" + ] + } + ], + "source": [ + "from dynestyx.inference.filter_configs import ContinuousTimeEnKFConfig\n", + "\n", + "\n", + "def true_model_with_filter(obs_times=None, obs_values=None):\n", + " with Filter(filter_config=ContinuousTimeEnKFConfig()):\n", + " return model_with_true_drift(obs_times=obs_times, obs_values=obs_values)\n", + "\n", + "\n", + "key, k_true_mll = jr.split(key)\n", + "true_mll = Predictive(\n", + " true_model_with_filter,\n", + " num_samples=1,\n", + " exclude_deterministic=False,\n", + ")(k_true_mll, obs_times=obs_times, obs_values=obs_values)[\"f_marginal_loglik\"]\n", + "\n", + "true_mll = float(jnp.asarray(true_mll).squeeze())\n", + "print(f\"Marginal log-likelihood under true model (EKF): {true_mll:.3f}\")" + ] + }, + { + "cell_type": "markdown", + "id": "70d0ae57", + "metadata": {}, + "source": [ + "## Polynomial library and true coefficients\n", + "\n", + "Library terms (all monomials in $x, y$ up to degree 2):\n", + "$\\phi(x,y) = [1,\\ x,\\ y,\\ x^2,\\ xy,\\ y^2]$.\n", + "\n", + "The **true** interaction term uses only $xy$: coefficient $-\\beta$ for prey and $+\\gamma$ for predators.\n", + "The known linear parts are **not** included in $\\phi$ — they are handled by `lv_known_drift`." + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "id": "5e647a3d", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "True interaction coefficient matrix Theta (rows: dx, dy; cols: library terms):\n", + " Terms: ['1', 'x', 'y', 'x²', 'xy', 'y²']\n", + " Theta =\n", + "[[ 0. 0. 0. 0. -0.09999999 0. ]\n", + " [ 0. 0. 0. 0. 0.075 0. ]]\n" + ] + } + ], + "source": [ + "TERM_NAMES = [\"1\", \"x\", \"y\", \"x²\", \"xy\", \"y²\"]\n", + "N_TERMS = len(TERM_NAMES)\n", + "\n", + "\n", + "def interaction_library(x):\n", + " \"\"\"Polynomial library for the unknown interaction term. Returns (N_TERMS,).\"\"\"\n", + " prey, pred = x[0], x[1]\n", + " return jnp.array([1.0, prey, pred, prey**2, prey * pred, pred**2])\n", + "\n", + "\n", + "# True Theta: only the 'xy' term (index 4) is nonzero\n", + "# Row 0 (prey): U_x = -beta * x * y => coeff at xy = -beta\n", + "# Row 1 (predator): U_y = +gamma * x * y => coeff at xy = +gamma\n", + "true_Theta = jnp.zeros((state_dim, N_TERMS))\n", + "true_Theta = true_Theta.at[0, 4].set(-_beta) # prey: -beta * xy\n", + "true_Theta = true_Theta.at[1, 4].set(_gamma) # predator: +gamma * xy\n", + "\n", + "print(\"True interaction coefficient matrix Theta (rows: dx, dy; cols: library terms):\")\n", + "print(f\" Terms: {TERM_NAMES}\")\n", + "print(f\" Theta =\\n{jnp.round(true_Theta, 4)}\")" + ] + }, + { + "cell_type": "markdown", + "id": "34ceb7aa", + "metadata": {}, + "source": [ + "## UODE model discovery\n", + "\n", + "Build a Dynestyx SDE model that jointly infers $\\Theta$, $\\sigma_x$, $\\sigma_y$.\n", + "\n", + "The drift is:\n", + "$$\n", + "f(x) = \\underbrace{f_{\\text{known}}(x)}_{\\alpha x,\\ -\\delta y} + \\underbrace{\\Theta\\,\\phi(x,y)}_{\\text{unknown interaction}}.\n", + "$$\n", + "\n", + "We use an EKF (via `ContinuousTimeEKFConfig`) to marginalise the continuous-time latent state during inference." + ] + }, + { + "cell_type": "code", + "execution_count": 8, + "id": "b6939763", + "metadata": {}, + "outputs": [], + "source": [ + "def uode_lv_model(obs_times=None, obs_values=None):\n", + " # --- Unknown interaction: sparse Laplace prior on Theta ---\n", + " # Very tight scale (0.005) strongly penalises non-zero entries → SINDy-like sparsity\n", + " Theta = numpyro.sample(\n", + " \"Theta\",\n", + " dist.Laplace(0.0, 0.005).expand([state_dim, N_TERMS]).to_event(2),\n", + " )\n", + "\n", + " # --- Noise parameters ---\n", + " # LogNormal keeps sigma_x bounded away from zero → avoids non-PD EKF covariances\n", + " sigma_x = numpyro.sample(\"sigma_x\", dist.LogNormal(jnp.log(0.05), 1.0))\n", + " sigma_y = numpyro.sample(\"sigma_y\", dist.HalfNormal(0.5))\n", + "\n", + " # --- UODE drift: known linear part + unknown polynomial interaction ---\n", + " def drift(x):\n", + " known = lv_known_drift(x) # (alpha*x, -delta*y)\n", + " phi = interaction_library(x) # (N_TERMS,)\n", + " unknown = Theta @ phi # (state_dim,)\n", + " return known + unknown\n", + " \n", + " return dsx.sample(\"f\", DynamicalModel(\n", + " state_evolution=make_state_evolution(drift, sigma_x),\n", + " observation_model=LinearGaussianObservation(\n", + " H=H_obs,\n", + " R=sigma_y**2 * jnp.eye(observation_dim),\n", + " ),\n", + " **initial_condition_kwargs,\n", + " ), obs_times=obs_times, obs_values=obs_values)" + ] + }, + { + "cell_type": "markdown", + "id": "c14c1b22", + "metadata": {}, + "source": [ + "Build the data-conditioned model (filter + observations) and run SVI for MAP estimation.\n", + "\n", + "**Key numerical choices:**\n", + "- **`cov_rescaling=2.0`** — inflates the diffusion coefficient inside the Riccati ODE, keeping the\n", + " EKF covariance positive-definite throughout optimisation. Without this, numerical errors\n", + " accumulate after ~85 SVI steps, producing non-PD matrices and NaN losses.\n", + "- **Learning-rate decay** — exponential decay (via `optax`) prevents late-stage instability\n", + " when the parameters are near the optimum and the EKF covariance is tight.\n", + "- **`init_to_value`** — initialises at the prior mode/median: $\\Theta = 0$ (Laplace mode),\n", + " $\\sigma_x = 0.05$ (LogNormal median), $\\sigma_y = 0.5$ (HalfNormal scale).\n", + "- **Dense observations + both species** — observing both prey and predator at $\\Delta t = 0.2$\n", + " (150 points) provides enough signal to break library-feature collinearity and recover the\n", + " correct sparse $\\Theta$ via SVI/ELBO." + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "id": "078b13c9", + "metadata": {}, + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "100%|██████████| 4000/4000 [15:11<00:00, 4.39it/s, init loss: 376523.5938, avg. loss [3801-4000]: 210.3712]\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Inferred sigma_x = 0.0593 (true = 0.0500)\n", + "Inferred sigma_y = 0.4182 (true = 0.5000)\n" + ] + } + ], + "source": [ + "from dynestyx.inference.filter_configs import ContinuousTimeEnKFConfig\n", + "\n", + "def data_conditioned_model(obs_times=None, obs_values=None):\n", + " with Filter(filter_config=ContinuousTimeEnKFConfig(\n", + " record_filtered_states_mean=True,\n", + " record_filtered_states_cov=True,\n", + " )):\n", + " return uode_lv_model(obs_times=obs_times, obs_values=obs_values)\n", + "\n", + "num_steps = 4000\n", + "optimizer = numpyro.optim.Adam(step_size=1e-3)\n", + "guide = AutoDelta(data_conditioned_model)\n", + "svi = SVI(data_conditioned_model, guide, optimizer, loss=Trace_ELBO())\n", + "svi_result = svi.run(k_svi, num_steps=num_steps, obs_times=obs_times, obs_values=obs_values)\n", + "\n", + "\n", + "median = guide.median(svi_result.params)\n", + "Theta_inferred = median[\"Theta\"]\n", + "sigma_x_inferred = float(median[\"sigma_x\"])\n", + "sigma_y_inferred = float(median[\"sigma_y\"])\n", + "\n", + "print(f\"Inferred sigma_x = {sigma_x_inferred:.4f} (true = {sigma_x_true:.4f})\")\n", + "print(f\"Inferred sigma_y = {sigma_y_inferred:.4f} (true = {sigma_y_true:.4f})\")" + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "id": "7afae164", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "fig, ax = plt.subplots(figsize=(8, 3))\n", + "ax.plot(svi_result.losses, lw=0.8, color=\"steelblue\")\n", + "ax.set_xlabel(\"SVI step\")\n", + "ax.set_ylabel(\"ELBO loss\")\n", + "ax.set_title(\"SVI convergence\")\n", + "ax.set_yscale(\"log\")\n", + "ax.set_xscale(\"log\")\n", + "plt.tight_layout()\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "id": "75c18e87", + "metadata": {}, + "source": [ + "## Parameter recovery\n", + "\n", + "### Interaction coefficient matrix $\\Theta$\n", + "\n", + "The learned $\\Theta$ (rows = state dimensions, columns = library terms) should be sparse — only the\n", + "`xy` column should be non-zero, encoding the predator–prey coupling." + ] + }, + { + "cell_type": "code", + "execution_count": 11, + "id": "49bb4601", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "TERM_NAMES: ['1', 'x', 'y', 'x²', 'xy', 'y²']\n", + "\n", + " x-row (prey drift) y-row (predator drift)\n", + " true inferred true inferred\n", + "--------------------------------------------------------------------\n", + " 1 0.0000 -0.0002 0.0000 0.0002\n", + " x 0.0000 0.0001 0.0000 0.0000\n", + " y 0.0000 0.0000 0.0000 0.0000\n", + " x² 0.0000 0.0003 0.0000 0.0001\n", + " xy -0.1000 -0.1008 0.0750 0.0749\n", + " y² 0.0000 0.0012 0.0000 0.0003\n" + ] + } + ], + "source": [ + "import numpy as np\n", + "\n", + "print(\"TERM_NAMES:\", TERM_NAMES)\n", + "print()\n", + "print(f\"{'':>12s} {'x-row (prey drift)':>22s} {'y-row (predator drift)':>24s}\")\n", + "print(f\"{'':>12s} {'true':>8s} {'inferred':>10s} {'true':>8s} {'inferred':>10s}\")\n", + "print(\"-\" * 68)\n", + "for j, name in enumerate(TERM_NAMES):\n", + " t0, t1 = float(true_Theta[0, j]), float(true_Theta[1, j])\n", + " i0, i1 = float(Theta_inferred[0, j]), float(Theta_inferred[1, j])\n", + " print(f\"{name:>12s} {t0:>8.4f} {i0:>10.4f} {t1:>8.4f} {i1:>10.4f}\")" + ] + }, + { + "cell_type": "code", + "execution_count": 12, + "id": "f58e1d71", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "vmax = max(float(jnp.abs(true_Theta).max()), float(jnp.abs(Theta_inferred).max())) * 1.1\n", + "\n", + "fig, axes = plt.subplots(1, 2, figsize=(10, 4), sharey=True, constrained_layout=True)\n", + "\n", + "im0 = axes[0].imshow(\n", + " jnp.asarray(true_Theta),\n", + " cmap=\"RdBu_r\",\n", + " aspect=\"auto\",\n", + " vmin=-vmax,\n", + " vmax=vmax,\n", + ")\n", + "axes[0].set_xticks(range(N_TERMS))\n", + "axes[0].set_xticklabels(TERM_NAMES, rotation=45, ha=\"right\", fontsize=11)\n", + "axes[0].set_yticks([0, 1])\n", + "axes[0].set_yticklabels([r\"$\\dot{x}$\", r\"$\\dot{y}$\"], fontsize=12)\n", + "axes[0].set_title(\"True coefficients\")\n", + "axes[0].set_xlabel(\"Polynomial term\")\n", + "\n", + "im1 = axes[1].imshow(\n", + " jnp.asarray(Theta_inferred),\n", + " cmap=\"RdBu_r\",\n", + " aspect=\"auto\",\n", + " vmin=-vmax,\n", + " vmax=vmax,\n", + ")\n", + "axes[1].set_xticks(range(N_TERMS))\n", + "axes[1].set_xticklabels(TERM_NAMES, rotation=45, ha=\"right\", fontsize=11)\n", + "axes[1].set_yticks([0, 1])\n", + "axes[1].set_yticklabels([r\"$\\dot{x}$\", r\"$\\dot{y}$\"], fontsize=12)\n", + "axes[1].set_title(\"Inferred (MAP)\")\n", + "axes[1].set_xlabel(\"Polynomial term\")\n", + "\n", + "cbar = fig.colorbar(im1, ax=axes, shrink=0.6, aspect=20)\n", + "cbar.set_label(\"Coefficient value\", fontsize=11)\n", + "plt.show()\n" + ] + }, + { + "cell_type": "code", + "execution_count": 13, + "id": "524c0f0b", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "fig, ax = plt.subplots(figsize=(5, 3))\n", + "\n", + "labels = [\"$\\\\sigma_x$ (prey noise)\", \"$\\\\sigma_y$ (obs noise)\"]\n", + "true_vals = [sigma_x_true, sigma_y_true]\n", + "inferred_vals = [sigma_x_inferred, sigma_y_inferred]\n", + "\n", + "x = jnp.arange(len(labels))\n", + "width = 0.35\n", + "\n", + "ax.bar(x - width / 2, true_vals, width, label=\"True\", color=\"steelblue\", alpha=0.8)\n", + "ax.bar(x + width / 2, inferred_vals, width, label=\"Inferred\", color=\"darkorange\", alpha=0.8)\n", + "ax.set_xticks(x)\n", + "ax.set_xticklabels(labels)\n", + "ax.set_ylabel(\"value\")\n", + "ax.set_title(\"Noise parameter recovery\")\n", + "ax.legend()\n", + "plt.tight_layout()\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "id": "2006029a", + "metadata": {}, + "source": [ + "## Filtered trajectories\n", + "\n", + "Using the inferred parameters, run the conditioned model one more time to extract the EKF-filtered\n", + "state means and covariances." + ] + }, + { + "cell_type": "code", + "execution_count": 14, + "id": "205ceecd", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "# Run the conditioned model with MAP params to extract filter means/variances.\n", + "# Important: AutoDelta stores unconstrained optimizer params in `svi_result.params`;\n", + "# for posterior predictive/filtering we must pass latent-site values explicitly.\n", + "map_params = {\n", + " \"Theta\": Theta_inferred,\n", + " \"sigma_x\": jnp.asarray(sigma_x_inferred),\n", + " \"sigma_y\": jnp.asarray(sigma_y_inferred),\n", + "}\n", + "\n", + "predictive_filter = Predictive(data_conditioned_model, params=map_params, num_samples=1)\n", + "pred_samples = predictive_filter(k_filter, obs_times=obs_times, obs_values=obs_values)\n", + "\n", + "filtered_mean = pred_samples[\"f_filtered_states_mean\"][0] # (T or T+1, 2)\n", + "filtered_cov = pred_samples[\"f_filtered_states_cov\"][0] # (T or T+1, 2, 2)\n", + "filtered_std = jnp.sqrt(jnp.maximum(jnp.diagonal(filtered_cov, axis1=-2, axis2=-1), 1e-10))\n", + "\n", + "# Some filters include a prior state at t0 (length T+1). Align to observations for plotting.\n", + "n = min(obs_times.shape[0], states.shape[0], obs_values.shape[0], filtered_mean.shape[0])\n", + "t_plot = obs_times[:n]\n", + "state_plot = states[:n]\n", + "obs_plot = obs_values[:n]\n", + "mean_plot = filtered_mean[:n]\n", + "std_plot = filtered_std[:n]\n", + "\n", + "fig, axes = plt.subplots(2, 1, figsize=(10, 6), sharex=True)\n", + "\n", + "labels_state = [\"Prey ($x$)\", \"Predator ($y$)\"]\n", + "colors_true = [\"tab:blue\", \"tab:orange\"]\n", + "colors_filt = [\"steelblue\", \"darkorange\"]\n", + "\n", + "for i, (ax, label, col_t, col_f) in enumerate(zip(axes, labels_state, colors_true, colors_filt)):\n", + " # True latent trajectory\n", + " ax.plot(t_plot, state_plot[:, i], lw=1.5, color=col_t, label=\"True state\", zorder=3)\n", + "\n", + " # Filtered mean ± 1σ\n", + " mu = mean_plot[:, i]\n", + " std = std_plot[:, i]\n", + " ax.plot(t_plot, mu, \"--\", lw=1.5, color=col_f, label=\"Filter mean\", zorder=4)\n", + " ax.fill_between(t_plot, mu - std, mu + std, color=col_f, alpha=0.25, label=\"±1σ\")\n", + "\n", + " # Observations (both species)\n", + " ax.scatter(t_plot, obs_plot[:, i], s=12, color=\"black\", zorder=5, label=\"Observations\")\n", + "\n", + " ax.set_ylabel(label)\n", + " ax.legend(loc=\"upper right\", fontsize=8)\n", + "\n", + "axes[-1].set_xlabel(\"Time\")\n", + "axes[0].set_title(\"EKF-filtered trajectories (MAP parameters)\")\n", + "plt.tight_layout()\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "id": "ad5cabdc", + "metadata": {}, + "source": [ + "## Phase-space tracking with uncertainty ellipses\n", + "\n", + "The same filtered posterior can be viewed in $(x, y)$ phase space. We overlay true states,\n", + "noisy observations, filtered means, and axis-aligned 1-sigma ellipses from the EKF covariance." + ] + }, + { + "cell_type": "code", + "execution_count": 15, + "id": "d5fa65a1", + "metadata": {}, + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "/var/folders/fx/6n6838cn51v4zyntllq4rxq80000gn/T/ipykernel_50406/3954310149.py:28: UserWarning: Legend does not support handles for PatchCollection instances.\n", + "A proxy artist may be used instead.\n", + "See: https://matplotlib.org/stable/users/explain/axes/legend_guide.html#controlling-the-legend-entries\n", + " ax.legend(\n" + ] + }, + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAscAAAHuCAYAAABzvaFNAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjgsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvwVt1zgAAAAlwSFlzAAAPYQAAD2EBqD+naQABAABJREFUeJzsnQV4FFcXhr+VuHtCAhESHEJwd3cvUGgpFC+U0kIF/kIpUKQ4pRRoKW7F3d0hIXhIIO7usvY/5y67bAwCJJBN7vs8Q9jZ3ZnZO3fufHPuEYFCoVCAw+FwOBwOh8PhQMjbgMPhcDgcDofDUcLFMYfD4XA4HA6H8xIujjkcDofD4XA4nJdwcczhcDgcDofD4byEi2MOh8PhcDgcDuclXBxzOBwOh8PhcDgv4eKYw+FwOBwOh8N5CRfHHA6Hw+FwOBzOS7g45nA4HA6Hw+FwXsLFcSnk33//hUAgwJ07dz72oXA4ZRYXFxeMGDHiYx8GZs+eza53beDEiROoW7cu9PX12TEnJSWxNqS21ITeo9/1oWjTpg1bShN520A1rgcFBZX4vvOeE9on7fv3339HeejToaGhrI9evXr1ox1DeSc+Ph5GRkY4duwYtBEujj8gqsFRtdDFW6VKFXz11VeIjo7+kIfCeU+2b9+O5cuXa2070oD1vuJl/vz5OHDgQLEdE+ftWbNmDRtXPtTNbtCgQTAwMMAff/yBLVu2sJtfUbh27RrrbySmOUUnIyODtduFCxdKXbOV5mObM2cOGjdujObNm3/sQ9Ea/vzzTwwcOBCVKlVi+uR9DQdWVlb48ssv8b///Q9aiYLzwdi4caOCmnzOnDmKLVu2KNavX6/4/PPPFUKhUOHq6qpIT0/P9bnbt2/zs1NK6d69u8LZ2VmhrUycOJH1sffByMiI9V9tJSsrS5GTk/OxD0Mxa9asdz4XNWvWVLRu3VrxITh+/Dg7ztOnT+daT21IbakJfY5+l4rFixezdYGBgSVybNnZ2WwpTeRtA6lUqsjMzFTI5fIibyM2NjbfdopC3nNC7U7bofNQXLzu2CQSCfutH4OYmBiFjo6OYvv27R9l/9oK3c8sLS0VXbp0UYjF4mIZ2x8/fsz6yNmzZxXahvhji/PySNeuXdGgQQP2f3qyoiespUuX4uDBgxgyZMjHPjxOMZOVlQVdXV0IhXyipjShp6f3sQ9Bq4iJiWF/zc3Nc63X0dH5qNZLQ0NDdn2VdkQiEVtKkvT0dGbN/5jnhBCLxWz5GGzdupXtu2fPnihvkLWXXGjexZp/8eJFtdXY2Ni4WI6nevXqqFWrFpvdateuHbQJfrcuBag6TWBgYK712dnZmDp1KmxsbNiA17dvX8TGxub6DAnq7t27o0KFCuxmX7lyZfz666+QyWS5Pufv74/+/fvD3t6euXM4OTlh8ODBSE5Ozjew1K9fn02dWlpass+Q/9abSE1NxZQpU5ifGx2Hra0tOnbsCG9vb/VnyCeQLpS7d++iWbNmbB+urq5Yu3Ztrm3l5OTg559/ZsdhZmbGfnvLli1x/vz5fPuVy+VYsWIFateuzX4XtVWXLl3y+WsX5XfRjfbp06eIi4t77W+l33H06FEEBwerXWRU/n00KNHrnTt3YubMmXB0dGQ375SUlEL98ArzRTx+/Dj73fT7TUxM2Hl+9OgR3oREIsEvv/wCDw8P1ib08NWiRQucPn1aPYDStDih6eajgvwS6fzQ96i9qN3++++/XPugz9ONeNOmTerva07DhYeHY+TIkbCzs2P9oWbNmvjnn3/wrtC2acCm7fbp04f9n871d999l6+v03F9++23qFixItt31apV2W9SGvMK9zl+U7tt3LiR/U4fH58CXUxI+NDxvY4rV66gYcOGbPt0rf71118Ffo72ReMCXUf0G2rUqMGmPfMeP/UHuqmpzoHK7zYhIYG1DV0X1FampqbsodzX1xfvAm33888/Z/+n49c83wX5HGtC/X7atGns/3S9q45Vs78X5frUHD9atWrFrquffvqpQJ9j1XW4e/duzJs3j4131Obt27dHQEBAvmOk68HNzY3tv1GjRrh8+XKR/ZhpnP7mm29Yf6TrtFevXggLCyvSdU7jVOfOnWFtba0eD+m6IehztE2C+qWq3VTuUKpr4vnz5+jWrRvb96effvrGc7Js2TI4Ozuz/bVu3RoPHz7M184F/W7Nbb7p2Aoa66RSKbs3Ub+nPk3bovNH7acJre/Rowe7Vuhc0Hmjc7N582YUBXL1IpeKvAJP1X/u37/Pfjf1H3d3d/XYRtcRfY/ahcaMM2fO5Nt2Uca1ot6/NP3A161bp24Xur5u376NDw31iaL6if/999+oU6eOOvZAc8nb70gHHD58ON/4W9rhluNSAA1uBN2MNZk0aRIsLCwwa9YsdiGRjyv5J+/atSvXgEuDAIlo+nvu3Dl2YZIYW7x4sfpipQGYBiHaJglkusiPHDnCfADpAiboJkL+QeRXSBZtEuKrVq1iNyISBHktRpqMGzeODTJ0fHQjJ/9EGtyePHmCevXqqT+XmJjIBnLaB1nJ6eY1fvx4ZvlR3RTo2Dds2MDeHz16NBPedDHSb7h16xYLCFIxatQo1gZ046djpgGYbmw3btxQW+eL+rto223btmXt/Tp/3BkzZrCHCroB0o2GyDsQ002AfhMJFGr3t7VskT8niRH6zQsXLmTCncQRiTU65jeJkd9++439Vrq5UHvSTZgeVGigGjt2LCIiIpjoo/3khR426AZPN1rqOyT0yReN+gsJdNXxqbY/ZswYto4Gd4L855s0acIGSuoPdBMloU/nio6FHqLeBRLB1B50A6MbCt28lixZwvZLfYigAZiOnW5EtD/qKydPnmTijPq86ny9S7sNGDAAEydOxLZt2+Dl5ZXru7SObr70MFQYDx48QKdOnVh70L6or1JfoxttXuhc042XfgtZwejmMmHCBPYwSMdA0HhA1zP1PeqThGpbL168YCKBzhsJLjonJMRJFDx+/Jg9TL8NtH0SDHQTJ39O2qbqfL+Jfv364dmzZ9ixYwdrfxKChEpcvc24Q+MKXesknocNG1Zg22myYMECNmND1yFds4sWLWL9+ubNm7namvopCRgSuTTW0gMYjb0kqt8EHTOJ+6FDh7KHShqDVdfJmyzxqv7www8/sN9J+963b5+6fejYqG+TYYTakSBRooL6EF0TNC7QNUGC73WQwKTxlPoQzWjRtU4PYdQ339SWmhTl2ApqJ3qYpuuIHl7pHND1RveI/fv35/osPcDQ5+gapnGQBCiJcxKcdF0UBj3gkrBUjQd5ofsPCW/qP3Rt0G+g/9P1S+MS3cfoPNK9k/ZPD2j00PE249rb3L9U8Sv0GRqXadvUR6k96Rp+3QwAjQX0EKwJ3WuoDfIaeOgeX1yzCRs3bmTnku7zc+fORWZmJot9iIqKYj7etF4TOmd03dODPD2caA0f26+jPKHyJT5z5gzz1woNDVXs3LlTYWVlpTAwMFCEhYXl+lyHDh1y+ad98803CpFIpEhKSlKvy8jIyLefsWPHKgwNDdU+Zz4+Pmx7e/bsKfTYgoKC2LbnzZuXa/2DBw+Y/1He9XkxMzNjfqyvg3wj6TiWLFmiXkd+gnXr1lXY2tqq/T/JNy+v/2BiYqLCzs5OMXLkSPW6c+fOse1Nnjw5375U7fY2v+v8+fNF9u8rzOdYtQ03N7d856Yw31LV+Vb5Y6ampirMzc0Vo0ePzvW5qKgo1s551+fF09OTHd+7+hznPW46L7Vq1VK0a9euSD7Ho0aNUjg4OCji4uJyrR88eDA7/oL67Jug/aj89TXx8vJS1K9fX/36wIED7HNz587N9bkBAwYoBAKBIiAgQL2Ozp/m8Rel3YYMGaKoUKGCQiaTqdd5e3uzfdJ5fB19+vRR6OvrK4KDg3P55FH/zHsuCmqjzp07s35VFJ9juvY1j5Gg/qWnp5evDYtKYbEQ1IZ5r4Wi+hy/zfWpGj/Wrl2b79joPc12UF2H1atXzzWWrFixgq2n7RP0Ho2/DRs2ZH6yKv7991/2uTf5c9+7d499bsKECbnWDx06NF8b5L3O9+/f/8bYktf59aquiR9++KHA9zTPicrnWPM+Q9y8eZOtp3tLYW1Z2DZfd2x5xzpVO3355Ze5Pvfdd9+x9TSWq6B90LpLly7l8iOmvvvtt98qXgdd3/TdVatW5XtP1X80fZGfPn3K1lHcz40bN9TrT548me+aLuq4VtT7l+qcUP9LSEhQrz948CBbf/jw4df+VtX3i7LQ9VBc8SQuLi4Ke3v7XDqExjFqw4LOz7Vr19gx7Nq1S6FNcLeKj0CHDh3YUydN+9JTK1l+6Mk5r9WJLHKa0xxk2SDrGU3nq6ApIBX09ElPjPQ5lYsAobIMkwWN1hcEWSvoSZSsN7QN1UJWZppmLsilQROyepAlgCySr4OsYPSErIIsqvSarCg0XUrQ9LTK0qp6OiYLCVmCNd009u7dy9qHrG95UbXb2/wusvzRfb04UlCRtUPz3LwNZNEliz5ZHjSPmdqFrKZFORf0lE6uNO+C5nGTpYUsbtSnNNu+MKj96LyQvx/9X/P4yXJC2yrKdgqDLDua0HGRhUUzCwe10+TJk3N9jixVdDxk6Xmfdvvss89YH9c8B2R1ojYjt6XCoOuWrj+ySJJfn6ZPHrXL684BtRm1H1l96bfmdYUqCJqeVfm4077J4krjDFl/36f9i5u3HXfod33xxRdF3j59VnPWhvoLoeozNDNAbUMWPk0fWbIuk+X4TajSVOXtb0WZHVFZxGlGhqx970phVtKCoP6neZ+hGRIaU0o63ZZq+zTDmfe6JMhNTROyPqrOFUH3S+q7mtd6QdC5JAo7d3QN0D1XBW2TzgNdh9QOKlT/V+3vbca1ot6/VHzyySe5jjdvHy0MukboXqG50EwEWe/zrvf09ERxEBQUxBay4qt0BUHtR/vWnNVWofptb3JXLG1wt4qPAPm3UQo3GoxpKosu0IKCtTRvopqdjASLCrqZk28rTeXRdI4mqpsoTYHSoERBf3Qjp4uPpmtpWlLVwUkQ0EVPN6SCeNOUDE0FkSAkwU/TKOQ6QUKCfMU0oencvOmfqC0Iuuho2oqg6TeaMieBr3njoN+i6Y5C2yMfxcJ439/1rmge59uiEmeFBTCQ/+jroGnv3r17s3alaSzywR4+fPhrpzw1oZs1TZfdu3cvlz9gUfzRaEqchD1Nv9PyusCut0XlU573mtC8HujBkfqEaipUc/BWvf8+7UbuFQ4ODuw6Iv9VuvmRuwB9L+8+87YLTT8W1A/p+s8rTig/Kz30Xb9+Pd8DLV3XmjemglD54tN0J8UyaPpl53Xf+pi87fVJwu5tXJTeNIaq+gP5nmpCY/PrXJdU0Pdp7M7rYkLn9E3Qww49UJHPLk0708M5iVea1i9qsCgdZ1FcP1QU1M7U38m9rSRRtVPediaBR+I073WZ97wVdK2/jsL8W6mt8o5jdC3RfSvvOkK1v7cd14py/3qb+3xh4yEZ2jQh9x4as/OuLy4iXhq/CurfNMaSAYD2r9l/VedCW3K5q+Di+CNAT+sqf9jXUVhks6qz0cVKAyyJJbqx0wBNFww9nX7//ffsBqmCLlR62qMAvlOnTjFLB/l7kW8uDRj0Weq8ZFkraL9vil4lyw+JbrKA0/bJZ4t8ZckyRD6CbwNd4HSsdKMgX1EKSqJjouNV+WcXlff9Xe9KQVbjwgaHvAFlqvNGfr1088jLm6LAyVeT2kl1rsn/jW6+FPhIvmKvg/y16cGJtkHCioQgCRTyMyPfuDehOnZ68FIFcOWlqCI9LyUd6V+UdqNjIPGyfv161j4kYumGQb+3uKBjIOFdrVo19kBLN24ShCSg6Xg0r+vCoABB8uMlP37yf6cHSBInZNEsyvc/FG97fb7tbMybxtCPCf1uitOgMZh8yklY0PmisZrWFWVs0pwhKM7jKqh98o5T77rtkjxvqge/woRlYdt90/7eZlx72/tXae6jhR1rQX2B1tEx531PdS5UsQbaAhfHWgxFZNM0EglQurGryJv1QgVFrtNClmZKyk/O83TjJyshCWvq2PRkq7Lkvi0kpChoiBZ6iqZAPAq20RTHJCRU6YZUULAOobLU0A2DLM70uzQH07zuE3TMdEOhaavCrMfF8bsK4l2eglUWAXqo0Qwyyms1UVmhaFB9VwsAtQdNKdOSlpbG+ge5i6hEXmHHT1OH9IBF7ar59E/iOC8FbUMVsU8DZElZL94UcU2BeuRipGnJVbkY0fvv024EzYiQgCFBQ6KOfnNBrhGa0GdI2BXksuHn55frNW2XrC+HDh3KZVUqyJ2msPNI1xAFl1IgkCbU9z7GTaqw4yyp67OoqPoDBYBRe6mgaXCayXrTgxx9n4QTiR5Na1rec/o6aLaMFhor6QGUXDooCJb6XHFb2wrqfzT+alrJaZwqaEo/7zj1Nsemaifav2oWRxXkRn3yTddlUaHrha6zwu6B78rbjGtFvX9pI24vZ4JV46kmtI7uWXmDQlXnQvO8awPc51iLUT3FaT5hUnYBsmhpQu4WNNhrQiKZLA6qaXOKjqXt0RRf3idWeq3y5SoIGjDy+kHSRULT23nT9NBxaKavouOl1zT4kDtGYb+L/JlpilkTmpKkz9Ax50X13bf5XUVN5UaQuC+K72dBovfSpUvqdap0aJqQ0KLZALL+FeSLmDedX17yniuyQNF0pua5UD2c5K1YRm1FA7rm0z+JhIIq4dE2Cvo+nRcS2XlTRBXl2N8XcuehY1+9enWu9WRxpd/1ulmMorQbQYKJFrIs0+8kH8Y3WfOpXei8UjuGhISo11OkPj2I5P0sodlfqa8V9IBS0DlQbSNvf9+zZ88bU82VFIX1t/cZd4oDmsEjayPNBGiOkeQ2U5QpfFV/WrlyZa71RameSdvP+5tVmQxUfU4lNIqrsiD1P80+QNkTaGzVvC5onKJxUPNapRSAeUsxv82x0XVZULvQzAhRlOweRYFmueic5k3l+b68zbhW1PtXSUCZm0qyYqGNjQ17kKN7lmamDEqPR0aJgs4jxRKRm8rrsoyURrjlWIuhtEH0lE/TPOQmQTd/morPO+CSPzKlnqHUNWSdoZsAfU51wasGRLIg//jjj+pURvSkTE995CpBwYGUDqkgyEpHrhmU+oYc/0lU0IVCKXXIwqYJCWZyt6B90LGQAz/5tpIfl8q/kFLt0FM3pQiii42OgSzcFKRB1jwVZOkhn1C6MZFFgnxEyTpBrgH0Hv3mt/ldRU3lRpCQp2MnX27KS0m/+U1J5ylggSwblPqHptuo/SlFEQ04moKJhDGlGKLfRtZ3El+qz1DgCln884o/TaidyH+RjpEsoXSjUKXZ0zx+gvoNiTY6FtoPtTfdsKgtyX2AZgDIR55EIg2AeduAzjN9ns4rWf8okIXSZ5GVk/5PgU50PDSQkrsPfV5zUKXjpPyixTWFSOeAziGlHqPzTf2RXCTIVYJcCl6Xfqwo7aZpPVb1m6K6VJAAPHHiBHM/otkVug4pZRndNDTblvoJuVHQb6FgVerzJN7ogTMyMjLXNulYqa9QH6dzRJ8hX3W6hsjViizgNE5Qqi4SfHljAEriHBSEqr/ReaF+Rtc6/b73GXeKA2pnutYpJR61G7mH0XGQyKBje5N1lMQsBc6SQYIeYKitz549W2Au5byQwKDv0ThH+6JxlM4zXf8qMUlWUOqXNNbQeEn9kvzh3zUlFvURSvtGQXwkwEms0sPB9OnT1Z8h1w66pmlcoLGKxgAaf6mfasa1vM2x0XVI9yka51XugDTeUhvQOde02r8v5P9P/YyO9U3xGW9DUce1ot6/3hcyrORNgVcYFCvxplR9NGOlyoMukUjYmETXJkGudqpZFEoZSNcK9XWa3aBYCroH02xoQfdNCgika13bfI55KrcPSFHLQhf2OVV6Is20LFevXlU0adKEpeihFFPTp09Xp6FRfe7FixcshUzlypVZKikqEdm2bVuWUi4ve/fuVbRo0YKlcqGlWrVqLO2Xn59focdLaWumTZvGUmGZmJiw79H/16xZky+VDqWeunPnjqJp06bsWChtz+rVq/OlYZs/fz57j9L3ULquI0eOFJgyitLmUJooOk5dXV2FjY2NomvXroq7d+++9e96m1RuaWlpLF0TpVyj76iOS7WNwtLm0XE1btyYHWulSpUUS5cuzZfiSfN4KH0XpQmitqLzN2LECNZ+r4PSmDVq1IgdG/UL+q2UEkuzVDK126RJk1h7UYozzaHg77//Vnh4eLC2p+/S8RWUho7SILVq1Yrtg97TTP0THR3N2rdixYqslCul/mnfvr1i3bp1ubZBadjovTdB26bzlpeCjotS4VFqKroeaN/0W6iP5C3bmzeVW1HaTUVkZCRLQValShXF23Dx4kX2m+n8U1o2SktW0G84dOiQok6dOuy8U+qkhQsXKv755598/YTS+1H6ObruNFOPUSo3SqtEqafotzRv3lxx/fr1AtN0FfUcvE8qN+LXX39VODo6spRPeX9HUa5P1fhREIWlcst7HarSX+VNu7dy5Ur1eEN9gMZVahcqpfsmqEwypZOklFx07D179mRpOt+Uyo1SAFJqQBoHaL+UzrJHjx75rm9KhaXqM5rbLOyaeF0qN7oOKJUmXZe0z5YtWyp8fX3zfX/r1q2sf9I+KdUm3VMKOs+FHVtBfZpS5f3yyy8KV1dXdl3SMfz444/5So/TPgpKqVhYirm80NhDaQC3bNmS7/sF9Z/C9kfHnzc9aVHGtaLev15X0rso96HiTuWmSg1Y0LIxz/VC4xhdrzS2mJqaKnr16qV48uRJvm3SOlX6Wm1DQP98bIHOKR+QhYpcFgqaluKUP8hSRtYmsl6pCltoC9SPyceeCu5Q4Ju2os3noCShGSiarSG3D7LmcrQLsniTLzXNInI+HlOmTGFuhORaoW2WY+5zzOFwPgo0aFJqLpqi1DZo2p18m8n1RZvR5nNQXFCluLw2IqokR9PkRSkfzSl9kGscufXl9ZPmfDji4+NZXAa5ZmibMCa45ZjzweCWY462Q/77VH6ZrMXkJ6kq9cvRXiiAicpGU0wG+d+SDyll+aDoerJ4vW3pdw6Ho/3wgDwOh8MpIhTkpkqDSMF0HO2H0phRLmkKKlKlhaSASwrA4sKYwymfcMsxh8PhcDgcDofzEu5zzOFwOBwOh8PhvISLYw6Hw+FwOBwOpzz5HFNaHipbTMnltTFqksPhcDgcDofzflBmGkphSYWrqEpwuRbHJIwp4ILD4XA4HA6HU74JDQ1llX3LtTgmi7GqMYqznGR5gyzwVEOekuO/7omLw9u4tML7MG9jbYf3Yd7GZQH5R9ITVFacjKUqXViuxbHKlYKEMRfH79eZKWE+tSEXxyUDb+OShbdvycPbmLevtsP7cNlv4ze52HLzH4fD4XA4HA6H8xIujjkcDofD4XA4nPLkVsHhcDgcDufDZQSQSqWQyWRaO+UvkUjYtD93IdS+NtbR0YFIJHqvbXBxzOFwOBwOp1jIyclBZGQkMjIytFrck3ijlF88/av2tTFtjzJRGBsbv/M2uDjmcDgcDofz3pDYCQwMZFY7yiOrq6urleJSZfkWi8VaefzluY0VCgXLghEWFgYPD493tiBzcczhcDgcDqdYrMYkkClVlqGhoda2KBfH2t3GlB4uKCiIuW28qzjmAXkcDofD4XCKDe6ny/mYFIfY5uKYw+FwOBwOh8N5CRfHHA6Hw+FwOBzOS7g45nA4HA6HU+4ZN24cvv/+e9YO5LNK7iFJSUkftV3q1q2Lf//9t9yfmw8ND8jjcDgcDodTbmjTpg2uX7/O8uGqWLRoEdauXasOFsvLiBEjYG5ujuXLl3/QY+V8HLg45nA4HA6HU65YuHAhpkyZ8sH2R5kZKHMCTw2nHXBxzOFwOBwOp0T4ctNtBMeXfEEQZytDbPi84XttQ2UdXrZsWa71K1euxLZt25iw3bBhA5ydnfHo0SOWKuzXX39l75H7RfPmzZn1mXI8E/T5VatWsXX+/v6Ii4tDTEwME+U3btxg6e5Gjx6Nn376SZ3hY/Xq1Uy4UxEVcvN4HeRuQZbsXr16Yc2aNSwt2ooVK1gBjLFjxyI0NBQDBgzA+vXr1dv39vbGt99+C19fX1haWjI3EjoGwsfHB5MmTcLjx4+ZkO/QoQM7HisrK7XFvWnTpmwb165dY3mEN23ahNq1a6OswX2OORwO503kpAO3/waOTVP+pdccDqdcMHnyZHz66aeYMGEC0tLSmDAmZsyYgatXr+LKlSusKmCVKlUwePDgXN/dvn07Tp06hZSUFCY427dvz5bw8HBcvnwZO3fuxMaNG9lnz507x7a5e/dutj3i4cOHrz02et/a2hpRUVGYN28exowZwwTyxYsX8eTJExw5cgQHDhxgn6XPdOzYEePHj2eFMmj9rFmzcPbsWfY+CegFCxYgOjqabZeO8Ycffsi1vy1btjAXlMTERDRo0ICJ6bIItxxzOBzOm/DdCdzfBYj1gUhf5bqGo3i7cThv4H2tuSXFjz/+iNmzZ6tfkxB8G8gvmay1JI4dHBzYurlz58LIyIhZbKkQCjF9+nS1JXnPnj2wsLBQu3NUqlQJX3/9NRPQo0aNYhZoEuFknSXo+Mhy+6aCFyTeiSFDhuDLL79k21JZe1u3bs0svf369WPCtlWrVhg0aBB7r1atWvjiiy/Y/kmwe3p6qrdrZ2eHqVOnYtq0abn2N2zYMPXnPv/8c3Tp0gVlES6OORwO503EPlUKY0tXICEQiPXjbcbhaDG//fbbe/kck4tEeno6E5uafsRUMltTHJMAVkEZMMgiS64bKlQVBYmIiAjmuqCCAgZVwrswSMSqUFUlzLuOrN2q/R87dizX/mUyGVq2bMn+HxAQwFwubt++zb5Dx6YZtEjY29ur/08PAqptlzW4OOZwOJw3YVNNaTEmYSzNAmyq8jbjcMpx1T+yzJLwvHnzJqpVq1ak75EIrl+/PvM3LgiyMAcHB6tfk0+zyr2iOKD99+3bl7lyFAT5OJNryKZNm5iAJrcL8sMuj3CfYw6Hw3kTnoOBOp8oRTL9pdccDqfcQNbYFy9eqNO8keglMUmWVrIUE/Hx8di1a1eh2+jRowfz5yV3jKysLGa19fPzw4ULF9RuEeRaQYI7JycHc+bMYdbp4mL48OHMr3nv3r1MeNNy7949ZikmyC/axMQEpqam7DctXrwY5RUujjkczttRHoPTdI2UPsbdFin/0msOh1NuIF9e8kumDA916tRRu2aQf3C7du2YqCSrMAXfFYaxsTHOnDnDAuBcXFyY9Xno0KEsUI6g7BCU/aJ///7MnYLcGsgvuLhwdHTEyZMn8ddff7Htk+CfOHEiE8XE0qVLWQCfqakpevfuzY6jvCJQFJTtuoxBJ97MzAzJycnspHPeDbpQKQ2Nra1tvikmTjlqYxLEquA0cjEgS6qWBKdpRftqObyNy2/7kjU0MDAQrq6u0NfXh7ZCsojyElNqNJ6XWPva+HX9sKh6sHRdWRwOR7uC0+hvMQenSWRyZEtlyJHKi3W7HA6Hw+EUBR6Qx+FwSk1wWnq2BA/DU+ATkggdsRBWRrqoYG4ISyNdtpgZ6EAkLF4rA4fD4XA4mnBxzOFw3g5VMBpZjEkYv2dwWmJ6Di4HxOGiXywuPotBXFpOoZ+l2TdzAx1YGOky4WxhqBTN9Nry5f9zvTbWhZEuL9nK4XA4nKLDxTGHw3m34LR3RCqTwzcsSSmG/eNwPywJqsgHD1tjdKvlgEauliADcVq2DGnZUiSk5yAhI4cJafp/YkYOAuPS4Z2RBJn89WETuiIhLIx0mJCmxUgsh4NlLCyN9NRCOrfQ1oGeWMR7BYfD4ZRTuDjmcDglTmRyJi49I8twLK74xyElS8rWm+qLmRhuVcUararYwMHMgAVqkN4lcfymQA25XIHULCkTzgkq4awhpOPzvH4Ukfxy34mv3S5Zm8nqTNZnlRWa/VVZptVWah0msrm7B4fD4ZQduDjmcDjFTpZEhttBCcw6fMk/Fs+ilVWUSOvWcTJH6yo2aF3FGp5O5hCLcscFkyAWvU4TU+o4Kucc+xRCm2ow8xwMM2sjuFobFSnSPyIyGjrG5kjKlKqt0GoR/fK1SmjHpmbjaVQqst8QHKjp7lGge0ee12SdNtYrJZHwGu3J/MnJTYanquNwOOUYLo45HM57Q9beF3HpajF840U8siRKQWljoocB9Z2YZbiluzUTiEWFUv3cv38fV69exZMnT2BhYQH77EA4pD+CvaUxHIyvwz4zG0atJhR5m2KRgB2TnZmBeh0l2g8LC0N4eDTCksIQHhPOXuuEhyMnLAzpsfGQ6xpBoGcMPVMrVHSvDntnd5jZOcLAzAbZAh0mrklYB8Wnwyf07dw9CnbvyCusS8jdg4SxKjUfBVoSWpKaj8PhcEoCLo45HM47kZolwbXn8cxVglwmwhIz2XodkQANXSyZGCYLcTV7kyJbSKli1KNHj5gYvnbtGqsUlZamtDqryU4B5FJAKH759x5MrFfA3t6eLZTcXvP/lOiecmlS7ksSvFSRinJcRkREsKT+tI5evwkdyCFWZCEjIQTPgnzgr/GetbU1vLy80LVuXXi18EIdzwYQ6RkX2d3jYbjK3QNvdPdQi+jicvfQSM0nTwiEPPoJ5FI5O4+lwrLN4XA4HxgujjkcTpEg/97HkSlMDNPiHZwI6UvrqLOVIT5r6oxWHjZoWtkKRnpFG1rIzYHEKolhWq5fv66u1qSCErZTFap69eohNTUVUb5nEfn8IaJSJIhMkiNdLmbrafH39y/SPgsrnuDm5gZ3d3c4OTmxalKqv3YOFZAlMkRkUhYevwhBQuhzRD/zxpMHvsyiHRcXh9OnT7NFBVXAIsFMx123bl20qlMLenp6r83vTJbnxHTJB3P3YEJbWgcWaVEwzYqDhUIKXXMPSHKkEEAAIz1RPrcXDofDKetwcczhcAolLi2bBdCRGL7sH6tOs2aoK0KbqjbMOkyC2KUI/r4q94uAgIBcYjghISFfiVUSw82aNUPz5s2Z9TcyMvKl20M4cqo0Rma2BGlhoTCUGyI9/c0W10IRCAF9UwgMzQADcwSJzBCZZAZDoSFEKUIoAtMgEYUgUxEBBTStqMYwNmyHugP64ZtKZrCSJyIj7Ake+vrAx8cHL168QFBQEFv279/PvkHW6xo1ajDBTGKZ/pIQVwl1HZEQtib6bCkqmTkyxKdnKwW1RjaPgrJ7FO7uYQngVTo+nVABGj7xRu+6FdCxhj0T0xxOWYHGFxWZmZnsutTR0WGvW7ZsiePHj3+U42rTpg369OmDKVOmFOnzNLZQBbjExESYm5uX+PGVN7g45nA4uayXPiFJ6swSD8JfuRuQe0T/+k7MVaK+s0WR/F9JDNMgrimGqfRtXgwMDJiltVKlSrC0tERUVBS2b9+OxYsXIyMj4w17EcDQ0FBt5aWlgqMTTG0doWdhB5GRBbIE+giJSUJwbDLC4lKQnAO2ZMrFSvOqBjJyGaFjl0mBjGQoMmOBjCQoMpOhoNdZKRCZ2SOrQjVczXTClYA49j0hbFHRrhPqjuiDzxyMYKVIRmjAEyaWaSHrMvlP07Jp0yb1jZqEskos00LuIEXFQFcEJ11DOFkU7fOFZfegh6DolCwkZkjY31uBCcxlZv6xp+jr5YiBDZxQs4JZkY+LwymtaLppvU6QUryDSMRzpJdXuDjmcMo5YYkZuPSMrMMxuBYQj9RspSXW3FAHPT0roJWHMs2anWnRLJqhoaHMX5jE8KlTp/K5SRQEWXDIPYGWgrCxsWHi18HRCVb2jjCWJUJXLIDAxh1Sx0ZIlusiJiULManZ8EvJRnxUNhSR9E0S97n9iXVEusw6W81OD3Ym+rAz1YOtKVls9WAslkOYlQJpWjxS46IRFZXAhDpZrqMSohAZFYW4+DhIFApI75IiFkNg7QKBrTuEth4IsqmM4DQhjgRksX0JUs1gLK4Ou8YV4ZYZjcQwf0RHRanbhG7UV65cYYsK8pHWdMfw9PSEiYkJigOhUAAzQx225M3uQRlGsiVyKKBAfFo2jj+MwgGfCPx7LYgtNSuYYlCDisyibG7IrckcFOnhmK7tDwU9ZL+Pnzx9d9WqVVi7di1z0YqNjWVuXfRwS9cisXz5chw4cAAXLlxgr+lh/5tvvsG5c+fY9wcNGoSFCxcW6EJFs2SjRo1i36W2qVy5Mvbt24eVK1fi8uXLzHgwc+ZMtQV76dKl+PPPP9kYZGtry/bz1VdfsW01atSI/aVxkfjrr7/w6aefwtvbG99++y18fX2ZoeH777/H6NGj37lNyitcHHM45QwSQZRNQhVI9zw2na2nuK26FjlobR+OVm6mqNO6H0T6r6Yg8xbyyJHKkBgfB997Pti9ezcTwu8CTWlWcKoIu0ruMHNwhrF1BeiYWgMGZsgRGSBdJkJsmgTRqdl4nJ4NRY6GiTeMlmh19gfKQlHJ0gANXSyY2FWJXhL2tqZ6sDHWRU5qIhOghfkdvwmJRMJuiGrRHBX18v9hiAy5g9BkKeJgCqmFMxPMqTa1mSWaUFglQ+EQAGGMPxQxAVAkhJGJ+tXGFQpEhwXiRIg/ThzeB+hQRg0BnJ2d0bhxY7WFmdwzVFPBxYW+jgh6YiHLMW2qr4Ov2nlgYlt33AtNwu47YTjsG4FZhx5h3tEn6FTTjgnl5u7WvJw3p1BIGJPr0IeCXLZoFul9oBmrkydPMlFMYvt1kMDt1asXc/96/vw5+70DBgzA3Llz8euvv+b7/O+//84s0uQeRuL5wYMH7MF3yZIluHv3bj4rNl33JLpJAJOg7tatG7v+aX+3bt1ibhXkbqZyq6BxqGPHjkxQ9+/fnxkbOnXqxGIp2rdv/17tUt7g4pjDKeMwP9+YNHUgHU2ZqwK57E318UmDiswy3CL1OMye7lRmLgjLQvbdDIRU6Kz29VX99bl3D8+iMyCs5AWIdF56JZhAVK9//p0LRRAZmcPA0oGJXYnYEFLkd8cgIy8z9BJkVGWG1eyXS8GYiSWwFabCzlAAW9dazBpKQWRk+VRB7gK0PIp4ab1WKJCRmQlDg/h87hSa7ZVvXYGf1INC4QJY0wJY1lJ679agTyuAnJwc5hISlpAOvwQpMmUCCAzMIHCuD6Fz/fz7SI2BPNofyE6FIiEE8uC7QHYaoGOI4OBgttBDiCYkBLp27coW8tOmm+T7Ws40c0zTa69KFmz5X4/qOPEwCrtuh+LI/Ui2VDDTZ2n6BtSviEpW7ydKOJzSwPTp01GhQgUmYt/0AH3nzh1mYaaZMvosXY8//fQTxo0bV6A4pgfa+Ph49h2aEVJZowuDBK6Ktm3bonPnzkwkkzguiC1btqBVq1bMek3UqlULX3zxBRP8XBy/HVwcczhlkORMCa4FxLGcw5R7OCI5S21dbexmiZYe1qhnrw+97AQmeMNv38by8/8gPCwMYWkChMckIjb9GqA359VGhWIIXRpAWKUPdBu/nTWoJCZWk6U6SIYl/MmS7BOO0s2bBavAxBYiE9tXK1oCgpx0SJ9dhjz8ARQxz3NbmQEmvvfu3cuWvLRo0QK1a9fOl+KOrObvYnU21BWjXz0ntgTFpeO/u2HY6x2GlecC2NLUzQqDGjqhU3W7t942p2xClley5n7I/b0vFPdQVCieIikpibkvaD5cU0rKgpg2bRpLKUnildJHfvLJJ1iwYEGhx71t2zZmVab9UJYdut7JWvy64zl27FiuAD06FnLT4LwdXBxzOGUACrSi4DlVIJ1mVgIbfQXqGibDJC0UOaGPEXwlGNfDw/MHukkyAFkOYGgBgWFFCGztIDCxY369IrfGb31M5noCWOgLYWEghKW+CJYG9H8RrAxFsDAQw1RPBJFIyKy9ZHURvMzHS/8XkcVGoFwvfGkJZf+nRZYNgf8pCJNCoLBwhqBqF4gNjCEUCJmlU/DyO/R/5Xde/Z9uXPHxCbC1tWHBNoXK1gJW0nEW+NECP1vQ5wr+vlQuZ5bt24EJuPXgMe7Ev/JVVOgaQVSrC1sIRyMBGpunI83/Js4fP8iKlxRGXl9m5MnLrBLMKtGcNz80TfcWdsyUneS7zlXxTccqLCBx951QnH4Ujesv4lnlv45VzPFZC13UrWTBcyWXY6j/vK+bw4cmr7XYyMgo11hJrlQqKlasyHyBNde9DgrAJX9kWgIDA9GzZ0+sWbOG+Qjn3W9ISAg+//xznDhxggUOUlYNcrtQzWwVZNWm4+nbty927tz51r+bkxsujjkcLSUmNQuXWSBdLBMolHmAIHHSrqoN9BJf4MKONQgP9YfKrqoQ6UHoUA1wagSRoTn0KZuDsSWyBHpQ6JlAYGD6xv26WhnAw9YIKdFhePHIBxHPn0CRmfQys0MSkKO8kVBOivx5KYoPhfqfBYV5SLwxz/ErAf1SWOd5XdzrNf/m/Sz7Cznqp0YgR6CHNKsaSLOuiVgTD/Wxh6crsC/dEDBsC/Rvi/Ye1pjc3gN1HIzx4MF9diMlyxFZkF4HZc6g5eHDh4V+hqxZeUUz+UDS1C4JAoIKjChLgduwrBcH74Vj151Q7H8Qx5YqdsbMN7mPlyOsjQvP8czhlFYoMJbcFSgAjq4X+n+VKlXYew0bNmSClILoKPCNxC+J2sePHzNXp7wcOXKEfZf8sE1NTdkMDolegmZ0yG9ZBQXrkhCma43GB7quKa5jzJgx6iBlWk/fqV9f6aY1fPhwFsRHM0nkC01QUSWKk6Bj5RQdLo45HC0hRyrH3eBEtasEFeRQQZkEBjdU+g7HP72FBb9NVQ+0lCKtfuOmiDGrgQcSW2TIXlkcJGy7lNMsCYrUaCiSIyG0r5prv5Us9DG9a3V0rWkPX997bKrv4F8H1dYUGtyrV68OgcAacrklG9BJgNJf1f9Vrwv6SwsFkhAkEOmmQYN+QduQ02tDKyis3dgisHSGQi6DIi0OSIuDIj0eilT6S68TlJbw16CaAi1sGvTjEgjgqPK/+ibMgm9crTmyHTzVn7jsH8cWFf3r9cSuCVNR0dKQFUWhtHH37t1Tp5MrqoWLoOAism7Rogn5VHbo0AFDhw5lfpCqmzsVFBnR3JUVg7n8MAhnX2TgoG8E5h59ggXHn6J9dVsmlElI88IiHG2BsleQBZdcFcjXl/5PWSUImn0iwUvCmMZAykJDbhljx44tcFvkYjJ58mRER0czIU0+xePHj2fvUSDeiBEj2H7IJYq2O2PGDLRr146NTyR2VYJX9fA6a9YsJsIpvoEs0HRNUjAhHQ8dA42ZdFxz5mi4x3GKhEBRUPRJGYM6LEWeko8P3Xg57wZdaBSlr3qS5ZR8G4fEZ+DiSzF8/Xkc0nOUIo4KM6hSrLX0sGFZGijamYJAKIqZsLKywuQpUyF3a4Y/LzxHUrYCYkkacu4fhzQuBMhKZvl7IcmEXa1mMPbsigixHWQKAbPyDWlUEUMaVYIhspklgoI6NFOtUQQ0pQ4aOHAgm6YvMjnpyLi5CVcvncc5/3SceRiNsNhkCCycAGk2e99YRwALEwOYWVhC164yJGaVkG5gh3iBGTIVr57pTURSSGUyZCp0C/RvMBLLYWckZmLR1daUBZAZKjJR1ckaDqZ60BUJChTrhQl7WlTnqajfySfw86x70zYoyf/FixeZi0R29ssARZEudByqwLzpQCQbVCi0qSe2rYzBDSvByUKZ4opuyiqhTAuleyIRna/djIxyWYxpIV9JShNFiwrqp3T+Bw8ezNJS5e3DOTIFTj+OZm4XNLtBdxvqq/3rObHcyZVtCs6GwtHOcZj6CD1MkV+svn7Ri9mUNui6o4A8evDjJdS1r41f1w+Lqge5OOaUiUG5rJCWlYOTPoG4HyPBJf84BMalq6ev61UyfzmFbcssxeRLS9AU+m+//YbDhw+z1zTYNGjUBAY12+J2uhVyxEZQZCRB9vA45M8uAXIpEz6Nm7eCXtUWuJ9hjufxyoC9xq6WGN7UGR2r2+Hu7ZvMSnz06FFmmSAo/RD5yZEopmnGtxnU6DjPnj2Ls7vX4Zr3Q0is3CF08oSgUn0IyAr8Fn1KnhgOefh9lgpNkZ0OAVmIxboQUOozXUMIjCwBE2sIjKwgMLaGwLDgClJGQinsjHVYpgV3B0u42ZkxIe1kYYgK5vpFKnTyoaABn6LiWRuePcumb9UIBLCr1xHymt2RjIJ9PMndZlI7d7TwsEY1e1PWp+iapmp+JJRVFmbVNGxeVOWwSTg/ffqULZoFFWjalixXlG6KfKHzjhPhSZnYezeMCeWwRGWIZgNnC2ZN7lbHgR0fR7vHYS6OOUWFi+NSALccl/1BWZsHiGfRlGYthvkOU1AWWdsIR3MDZhluXcUazdytWe7ZvAnlly1bhn/++edV+jGBEEK3JhDV6QmBiTUUWalMFFvGP0LzJg3ZtGCF6vVxPlSG/T7hSMuWMlHSv54jPm3iDHNBJvbs2cNEsabfKuXVJUFM04BFnX0hK+eNGzfUYi4wJhnCCjUhtKsCgWMtCPRKxmqoUMiVfs/Z6Uw4kyUa5Dqio69cxPoQ6Bqy/wt0DV67HUFmMgwVWbA1EsHFxhTVnW1RpYIVszy7V7CC7kcSz3S+yW1G1bY3b97MJWj1jUxRucMQhFnUQ6a84Gu1so0R+tR1RAMXS9StaM6q7anOGwlkTXcMEtBFhcQzpY0aOXIkE8x5H6AoePRGYDz23AnDsQeRLK0glSPvXtsBgxpWZIKZW+u0cxzm4phTVLg4LgVwcVz2B2VtgTJIpGZJcOZJNK4GxOP683hEpSittlSAoYmbJbzs9dG9nivc7QrOFkC+vlOnTsWhQ4deraRgLpdGEHn2hMDUjolDd0kgBnrZol3L5qjk4oZTj6Ox9UYwbgYmqMtBk5W4Z2173L5+hblNnD59mk11EeQTR5HPZA2sU6dOkQQL5UKmpPUk2C5fv4Vss4pKQVyhJgRmDq/9roWOHCbZMcgIeYioh1chT6AwwldeX2QRr1azDmp6NYRrtVqoWLkq9EwtkZwpRXxqFmKS0hCbnI6EtGwkZuSwMslpEgUyZUJIUTL9VSGTQBEfzHIUGwtyUIm5b5jB0VwfFa1NYGVhDgsLC/VC/oTvW8WrIMg1gtwuVGKZ3Cc0cavbDNb1u+KepPDS1DQbQenYSCw3cLHIFUBHU5CaYllVDvtNQY/Uh8iXklwvKICooJSDR+5HsCIjvqFJbB1V7iOXC3K9KGpVxvJEaR6HuTjmFBUujksBXByX/UFZWwiKS8PU3b7wDklSW+/aVLVl7hKNXC2ZD2zeNqZBxM/Pj5UXpQCM3AggqOQF3Xp9oTC1h55QjoG1rfBd7/owN9RDRFImdtwKwY5boawYBuU57lbbnoliW2E6du3ahR07duQK1KLIZ7ISk/sEWQFfB1krb9++zQTxmbNn4R+TDsFLMSx0qF7o90SQwyonCoKEYMQHPkFG2GMgM3eZZ0dHRxYprlooZ++7+jFSVUASYiScgyJiINA3Zq9JVEclpiIuOQMxyelMYFPWhQwZIBXqQUCW5mJCTiKaloQQWCAVlsJsWJmbMMGcV0Br/lUtBZWjLQjqLxQtrxLK5Iuu8pUmTCxtUK1NXwgq1sW9xMLzHbtZGzGRzMSyswUTrSpRT/uIiIjI57+sKhWsmRFEky+//JKVwKXfkxe/qFTsoUwXPuGIT89hFRvp2hjUwAntqtlBV8zHnNI+DnNxzCkqXByXArg4LvuDsjZARTkm7fBhN34KdPuimTMq25rkKr9LbRwdFozU2ztw/eplXAvKwLVncYhPSMy3PYGTJxw7j0GMRJdNS49o5oIxrdyY+wUFP225EYyzT6JZOWBy0fi0SSX087TH3WsXmdsEBXmp3DFIiJF1b8iQIahWrdprfwdZJc+fP8+E14Ubd5Fp6swEsahy00K/oy8kq3AsMkMfIdH/LhRxQbkySVAuVFVpZBLn9JdSG33sPpwtVYrq5AwJEtKzERqdgODIWARFxiI4Mg5RCSmIT82ERNcUAqtKEAjf329WnhShFNFJ4axCHnMNyU6HLqQwN9SBuaEurC3McoloWqgQQUHCmvyC6VzT+aKHGArwe4UAVZq0h23ddkjQr4CAhPy+xiqsjHSZWG7ILMuWzNKsI3rVhjTjQJW/KGCPfKPpL1X1Kww6z+Smk7ccNmVlOfc0hgnl834xrP9SAGpfL0fmn1zV3gTlmdI8DnNxzCkqXByXArg4LvuDcmmGfCz/OB+AZWeesSpjiwbUQbfaDrkGCRIRJChoavzymaOIj4lWVpIg8UpBZjqvgqwa9R4BafUueBiRylwxKHXWuNaVmcimqmXkOhEUn8GSN7SpYsOsxE6iVOzetZOVH9acDqeUQWQl7tKlS6GWSUojRJZBEldnzp3H45gsCCt6QVSjQ6G/mTJc6CYGMiEsifQDkilVm1KIk/XRw8NDbREmkUSvVSnBtLkPk78uWeGDQkLxPCQSgRExCIuJR2hiFqJzdJEiMoPAyrlYLNKKjEQostKUPtXkX83+ZkBBJafVr9NhIJTDWF/MRLWlmQnzGTc3NcHTx49w3+c2C9Ck6ofUz8ztK6Fq675AhZoISNNBanbhKe70dYTwqkhiWWld9qpkDhN9nVxtTGKJcsNSUYI3FSagvqB6QKKFIs1jUrOxzzucCeUXL4NTPZ3MMLBBRfT0rAAzg7ev9KftlOZxmItjTlHh4rgUwMVx2R+USytUmOObXfdYsB35+K75tB7cNNJXUSDVd999lyv5uzw7DUJ5jlKwqISLnil6jvoWUbYN4RueytwjhjauhAltKiMyOYtZiQ/7RrDgJrKykYVtQF073Lt6lvkSU2CcCjp/VLaUrMSUgaAg4uPjceHCBaW18fYDpNvUhqhO90J/p540HfLgO8xfWBH7HMh6lR6MUsppukd4enp+tJSKH7sPq1KzUcluckugv2Fh4QiKjEFIXBoiU6VIEZtBYFlJuegXr5VUkRYPWeBtyINuorGtFC1s0mBooA+foARcDJIiVfKqTYQiMaq16AaLmi0RI7ZBcFLhVmX2eQH5sZsyFwwPCyHa13FBBQvDfG445MpDeVjflEWU+ohKKJNoFtq641RAKo7cj0RGjow9GHatZc/6ehM3K3X2lrLOx+7Dr4OLY05R4eK4FMDFcdkflEsdOenwPrcXX103QkS2PgbVs8ecvnWhryNSiwTKNLFy5UrWrmQ1JRFAfrW3zh/Fw0dP1Jbjlt0HILvxaNyLSIdYKGAR/V+2cMWd4ERmJb4fpvTVpVRvZCV2FSdhz84dLDcx9X2CzhdlEKDgOvqb10pLx/DgwQMmhk9fuIKHOdYQ1e1d6M8T5aQh59EpyKP9lS4SJOIBNjVOv0HTPYKS4peW7APa0Iepb5D1mYSzSjwHRkQjMDYVkcnZiM8GJDqmyhR1xlaAsRUE4revPidPDIXc/xKEgTfQpoIEPVp5wbL9ZHWGEfJz18TGpSoqt+gFiU0V+CcLkSV95cessiTrCIVIzVb2BYLyKyvdMJTuGO42xmoRS+45lBmFHt7eVNFP0w+9tlcD6Hk0xQvYwS9eqt7PwPoV0b++I0vDV5YpzX2Yi2NOUeHiuBTAxXHZH5RL20W/ccdOzL9vBJFAgV9tzmNQy9pAw1HsfXKh+Oqrr1igFEFW3P/973/Yv38/K/0ZHxcHoTwbtes1gEHLkbiXZsascv3qObGp5EvPYtk0c0qWFAY6Ilaat19ta/jdPMt8iSkwSoWTkxMTxLQPym2sCWUgIF/Uk+cu4kxQNrI8OkAgKmSaOjsN0vtHIA9/CKS8yoRAwlfTKlyrVi3o6uqitFIW+jD1L8pOoRTOYUw8B4RFs1R5lD84LlOBNCqKwnI8k3i2Lvy8am43NQayR6egiAvEqMF90alNc+ZDTDMIFAyamaXMqkLSVkffEFVb9oBJ1aaIUFggMlVpVSZrLvm364sUrIhLVHIWnscq3SEIcoOo72yhFsu1Hc3Yd6hwDQWGUq5uVVCfCvKlJt9pSieX19osNLOHdaOeyLL3ZCXQ6dgojzO5XXSqYad+GC1LlOY+rO3ieNy4caxAxIIFC1g1Oyr1TDM95MP/saBZE1X1vI9NmzZt0KdPH3Y8NC7Q/5OSlMHlVKmPgrgnTJhQJsQxz7rO4RQjlKZt+n/3cfyhKVx1k7CmxmNUz34BxCoFIwlgmlKmICm6MBcuXMhucD169FBbzxxrN0WFzmPwIF4BQTrQo44Dajma4WpAHD7/R1n9zt3WGN90rITKojgc2L0RQ6YfUpdzJust+RCTKG7ZsmWurBdUuOHAyQs4cj8cYeZ1lXl+9dsB1ZSiR4VCkgXZvUOQv7jOhDFBabnIElyv3mAmhOn/b1UZj1MsqEps00KlYQuCbjpkmSUBHRoWDv/QEDyPTERoYgZi0mVIloog0beA0P5V8KXAxBbiJsPY/zenAZuPJEOemANIq8CiTRWIYkKREfUC8sgnyMnKwMPTuwFaqM9Wq4eKjbsiw9QNAQkZkFIUXXQGPGyNMbC+E0wNdFjp73uhSbjsH8sC7ghyD6rjZMZ8lnuO/RHf/fQzLp4+zoSyqhIf5fOmGyiV7aXAPbqpaZbDjjm9nuX3pqBQcZWWuKLwZCW1TfXF7MGR3C4oeLC0zF6URw7eo7SMJUvvuo5vJfKoBLQqCJRYtGgR1q5dy/5fkMuPqrTz8uXLUVqg/k9lou/cucP+T9cEiemPwfHjx1GW4OKYwykmHkekYMK2uywYrrtjJhbo7YBJthCQZiHV0Bk/TZrEXB0IKo5AN/v169erRYCla01U6jYej1P1EB+vYIFOlOf1bnAi87MklwoqlNCnliWCbp3B5pmzc019U/leCq4bMGCAWrQmpaRi98krOHLbHw+k9lDoUmq2SoBtpdxiWCaFzPcQ5E/PsRLOJKirVq2K+gP7qK3C7u7upc5SxSkYEpPkgkBLo0YFf4asz+T3HBoWhgeBUTgdIsXjnNwPO0Iq6W3hBOa4Y1sXOrUARWYKZC9uQBFwFYrkCPa58KfebCH0TMxRuWk3mFRpgpBUMfbcDWPrTfTFaOVhw/IXWxjqIjAujbkG3Q1KZH/XXlTu08PWCQ1G/oau47MQdPssju/dxiyl//77L3ufKjOSv/yKFSvYQ6Yq/zIFtN45vwbQN4GYMqd4dcXm61Jsvh6M6g6mLCUcFT2xMCq9MxucDwcZJsgC+qGgB1aRSFSsD2k0HpMhZObMmWjcuHGxbZfDxTGHUyzsvh2K/x18yKxjs3vWwOcNbCC4nwzE+uFunB4mfr8JIaFhbDDr3bs3u6mTawVhYOuCyr2/gn+2GR6nAka6QtiaGsAnJIlZ4OxN9TGlvTsqC6JxfN8GjJ5xTF3OmaaMVOWcSXCT5fr4zSc4fOM0bsWJWOlohrAykEcTkBiW+V1gwXM0RcsyR/T+jlmEKWjuTTmOOdqNiYkJewCihfKOfKMRRHrUNxxbLvvhWeKrbBXypEggPV5pYa7ZCajZCfLYF5AHXIU86DYgUbpDZKcm4dGJrRCe2s6cMJy9WsChXgck61fEsYeROPogkmVSqeNohrbVbPF1ew+Ww9g7OBG3g2hJYLm5ldSG3eBlqKqfjaQAb/hdOYZbt28zNwwSBL169cIng4dg6nfTME0ohPfdO8w16eLFU4h7dAoiGzdU7jgMIXHO+OXwY/x27Ck61rBjRUZaetjkSqPI4aiswxQPognFhpDLGgnbDRs2wNnZWV1m/ddff2XvkXsBVSAl63OFChXY9+jzq1atYuvIRYkyBdGDHoly8u2nFJajR4/GTz/9pDY8rF69mgl3mgkkN4/XQekuX+fGQMf3ww8/sCBYmklSFXgiCguIpawy8+fPZ+XpKYsQPYQ2a9bsnVwuqG1oW7QvsnDPnj2btQm5PNDvphz59MBAM2BUgIrag+6NdMxU5IrcI0j8UxuSKwRlAxo/fjx7j35bxYoVsXHjRnbvK2645ZjDeQ8yc2RMFFMKNfK1XD3UC16VlAUOZPVGsIHu999/Z+nQSMi6ubmxC5teC01t4d57EkKE9niW/Wqb6TlyBMalo4W7NXpWN0PU3dPYNet/uYKWatasiaFDP0XTDt3wIDIdf97yw4i9+5AhUPlXmea7uqmMtDzwNnQyYuFZpzbqNamHehN+Z6KYBnM+7cwhKNvJ8GaubAlLzMAB71DsuBaAcDgA5g5QpCdCThlJRDoQWjhC2HQ4FA0HQR7sDXnAFSiin6kbkkRwsM9lthDGVg7waNUboop18Dw2Db5hyVh+xp9V42tT1Qada9pjbt9arLrhnaAE3Hkplr2js6Ewqwdx93rQFcghSAhCevAD7Djng+2798LNpRL6DRiIvv36Y+u27bjn480EDuV1frZ9DoQ6emjQdzTg2oSJc1rooXNAfScmlJ2t+IMgp3AmT57MZvjyulXMmDGDxY5QCk7KykMid/Dgwbh06ZL6MxRweurUKfY+jfsUEE0CkmYRo6Ki0K1bNxYPMmrUKNZfaZsnTpxgAc2//PILS4X4rqxZs4YFvZLwpJnFOXPmYN26dex1QRw7doxlTzp06BBzzzhw4AAzvjx79owd/9tAM1PUZpSJiYR2x44d2f2PZkzpN9Lx0H5olovcQlRB4lR2nv5///595vZChYPIkLRlyxZs2rSJxdSQPziJZXrgoIqjJQEXxxzOO/IiNg0TtnnjaVQqu7EvG1RXPWVL09V0QWumUKOnYKpcBiNLOHcdgxijygiih3eNB3jyk+xazQK19BNx/tBf+O6X05DKZOwzRqamaNd3GCrVa4uAJDkWPYtHpt/tV19WC2MlZBWWh91HJf1sNPSqA6/+9VG//hj2lK7pa8fhFAZlfviqfVW2sAp2twKx944CiUbKB0BFegKQmQqIdVgRGFoUKTGQkkh+cQOmOnJm3aHS0XSzi4+PhM9+pV8n+QlXbtwR1nXaIF5ozx4waSH3IQrYa1fNluXwntO7Jgs+9Q5JZIKZrMu+IjHEFm7K7chlCI0PxtI9F7BsyRK0bdsGnw37lLlhkLAgkUxi4Nbu1RAI/kCH3p/AqdUAnA/MxOrzAWxp7GrJfJO71rZnucg5ZZ8ff/yRWTJVkH/+20DWUBKfV69eVQc7z507l824hYaGsn5PTJ8+XW1JJqFKwaUqdw4KaP7666+ZgCZxTBZomgVs2lRZUImOjwws7wqJ20mTJqljE37++WdmhVVlMcrLH3/8gWnTpjGDCdGvXz8sWbKEiebhw4e/deAoWcDJGkyFpVQCl8Qx3X/owYAMPnRsKst0bGwse2ggC7sqCJIEPRmD6Hqm75HofvLkCXMjoYDJkoKPAhzOO3DkfgS+/+8+MiUyTOtcFeNbV1anqDpy5Ah7+s43ABmYwa7tZ0i0rgNWrFlDFFPUfvcqJoi/dwo7523Glvh49XvuLXtDUL0jwrPEOAUR4K2MDgZyC1xmuYt4hNq2YnRo4gWvTl3h5fVTgaV6OZy3hSrTzexVBzN61mZCddeNFzhyX44M2cuAz5wMlpNbYGoLnXr9oKjbB+mRj/DI/wrE/jfQpVMH5q5D08VkISNf4ec3TrKFMKvgisotekNuXx33QpJw40UC5h97ytK0kVBuW9UWk9p5sAwUVLnw5osE+IQk4nZgHG6JBcixGQlFai+ce3QCZ78YBVtrS1b1kUqu0/VIIpkscqcP7ITg4C5079EDEweNwY0YIU4+isLNwATMOvQIPT0dWLYLr4rmfDalDPPbb7+9l88xCbj09HS0atUqVz+hbD2a4pgEsAoSg/TAppn9gkSk6rNkVCH3BBUkBvNmGXobSIBq7p+OjbZHwXsFQcdH1u9Zs2ap15H7wts+OBA0U0rueirIFUW1ncWLF7N9UIYLajtyZyHhTvun9qAsE5qQywn9FhLodOzkbkJtTG5VNDNbEoHhXBxzOG8Blbadf+wJ/r0WxKaC13/eAM0qKy9MuunTBU5WgFzom8C86UBkVGwKzcK9lMKqe217VJaH48qhP/DrgkvMGkGDA01hNWnRBgGWjRCssAaUWbRyQf6eiohHqG4BfNG7HRqNHsoGFR40xylJ6GZW39mSLfMGyFmp8p3XAnDumRwSxauATYFQCIFjbQgda0ORlYpjL27iyOWNcLPSZzc58mWkQDoSyuSjmBwRCO/dyilrcoNwb94N5tWaIVIiZkF1tFAuZbre2la1YT7Dnk7m+LyZC1KyJNh1OxSbrwqR2mQYBHV7Ie7BCfzx13pmDSMrE013k78iTSsfPXoURw4fZgvdoDdP/BrPsk2x+04odtxSLpQRhoL4+no5wcbk7fNIc8oWecdVGqPJKkqFnMgyWpTvkQgmdwnNGUVNyMKsWXJdlfP8XSFhTCkQVZDPLglUSvFZEHR8ZGke9wZf56JAM6WqlIMEuVZQgDBB68jqTj7QZAXu1KkTy49PPtvUXvSQQG1bECTeaSEfagrMJdcTsoYXN1wcczhFhPwvJ273gW9oEhq5WmL1EC/YmipdGaiABt14NQci6BnBsF5vSD3aQplkTYmzlSG6uhsh7cEZHJy7Azs1rMRkKTBzrons6t1w1sI1l3WZ+XpGPIKLXga+HdYDndt+WWL+VhxOUdARCZlFl5aMHCnOPInBtit+uBWaBgU0hLK+CSs3TktwXBDmbD+P+YuXoVfXjmw6mW5u5J945swZJpbphvnswn6AFgDWbrXg3LQbss08WJ5vVSq4KnbGLANGqyo2mNDGnVWMJGG7/tILRDccBHG93sh+cAI3vc8yEUNT3hQQS5Yryt1MeZUpBRUtdIOeP3UqRNZ1sOdOGA7cC2eW60Un/FjgILldkPsU/WZO+YOC3ygIjwwY9IBIIo5E5LfffssC7khYUmVR6sOUV74gKGUnuXOQMCTfWhrvyX+WBDBZjEns9e/fn7keUGA0BbSRdfpNIlQFBWrTa7IQ0/GRRZZmTchvmILrSFTScRaW2WLixInMmt6wYUPmWkE5xykLDIn/wgR1YdD+6beSWwgJY3pIVbmx7N69mx0DWbHJik5BeeRnbG9vzwL5yAWDUuuRRZgsxpR2r2/fvmxsoLznlE+frmWyTuctaFVccHHM4RSB809j8M3ue0jKkGB8m8r4tmMViEVCZuX966+/2CCmRscAug36Ax6t8So2GGhbxRqVFWG4c2QDVi67+eoNPWMIK9RgeVpRqR5SdHL7Dkvv7IazXgYObV4LKytlHloOp7RBvrq9PCugR217+AdH4HaUFNuv+eNxrDKzigqhtQtbFNJB2B/ijb1jv0cVc+Cz4cOZ1YqCdchyRFX6aCERG/fiIVsIsYExPFr2gqF7Q4SliLDhSiBbVKniSMjun9gc5/1i8OeF5wir2xsGXj2hG3wdCdf/YzM7tJBYIBFDhVRIAFDQFC0dOnTAN998gx+7tceZxzHMmnzmSTROP45ms0X96zmyID532+It7V2WeZscxKUVCgwbNGgQE2ckMMmHnlwzSMS1a9eOiTiyJlPAXWHimHLFk3gmP2TypSUhS4Fp5OdLUN+jewkJZBKmJL5JCL4OTQOJSvSeP39eLbbpuOiBkPKF0/sUBEditCBIRNMxjR49mhl69PT0WOpEErbvkg2HgvooCI/uk2PGjGGin6AgxqlTp7ICK+T2Rw/I5CJBkG8xuVyQQKeHDXooofYkcUzWYhLw5FJBv5vaS9MFpDgRKN5U4L4MwCvklf3KTCWFVCbH0tPPsObCc1bda+kgT7SvbsfeowuVnszVFenEehC3GgOhU51c2xhUywzZj87g5P6daj9kBUQQ1ewIoUsDCK2cC9x3zsH/wcFQwAZfGnA570957MMfu40jkjJZEYid1/wRnJK75LQKRWosZAFXoRvug76dWzO3C/JPJgsdWcNoGppEBYllSgP1CgHsqzeEU6POSDVxRlCKgk22qFLFkUWZMl+QtflFXDr0RIBjdjACj69HdqKy0iNZnsgdiY6ZCoyoaNu2LbuB0zQ4/YZ93mHYfScMIQkZ6nLtZE3uXscBJvo6pbcP56QDvjuB2KeATTXAczDA8p0XP9peIe9DVG8rD1zIUz1PGyvkcXHMKTLlTVjEpGZh8g4fFhjk6WSG1UProaKloTrojp6EGfom0OnyPQSmStGsYoBjGh4c24QHvvdyb9jIEuKWoyG0dYciKw3GWTHIMXeGhILtpDmQ3NkDxbOLGD58GMvlSk/gnOKhvPXh0tbG/tGp2Hc3lGW9iCvAj16hkEMR+YSlhKtuKsXnw4Yyi5Fmzm266dH0KollsviSX6YKXVMrlipO19kLwdmGSMtRinErI11kSWRIz1HmbdYRCVDXNAvJN/bg4c2X1UcKgQKuSCSTBU0uV+BWUAKzJh97EIksiZyVcO9W24H5J5O7VUmLqbfuw7f/Bu7vAsT6rCAR6nyiLmVf3HBxzCG4ONYSuOW4eChPwuLGi3hM2uGD2NRslk5qRvfq0BOL2FQXWZLoiVhhZA29/r/l+p5YALTMvoFLB7ezzxLkV0ZBP/SEfPhuENBkOAR6xmhqlYUBLWph2cUwhCVlQR7tD+m1jahgosMi66n0c5H4gJYhbac89eHS3MZkNfIJTcJ/t4KYVTldml9QKrLTIH9xkxUZGd6zLT4bMgA1cnxz9fMMqYCl0iKLMollcsdQIxDCyasVbOu2Q5KBI8JSC7Zat69sAtOw6zi9dwubei4MChYikaxKs0UFd6hyJQllKthDuFgZskwXVAXQ3ky/dPThY9NYMSJYugIJgcq267aoRI6Ni2MOwcWxlsDFcfFQHoQFWYb+vPgcS075MYvQgv510NNTmaPyn3/+YZZcCEUQ1RvAgotUWOtIYHD9Lzx//NLFgsrgenhg6NChynQzS5dh99NsiGp2gliejXl96+BZfA7+uRIIhUwCqfdeKPzO48tRo5hflouLS9Hb+ANahrSd8tCHta2NyXXp6vN47L75AqefxCBHnv878vhgyP3Owz3pBsa3sEWv2uYwaDA0Vz8nwU3l1MmqTGKZquhR0QUVhjZOqNyiFwQVaiEwUxfZ0twehQY6QkyoIcft47vYNuh3FESTJk2YSCaxrLISk0WcymST60VcWg4oqyO5dJDbRfvqtuzBurjgluOSh7tVlDzcraIUwMVx8VDWhUVSRg6m7vZlkfAUBf/nsPqobGPMbrjkf0iWKGHlphA3G6H+jpU4B7Hbp0GapfRDJP8mCn4gUdygQQMWODBi4lQE2LaG0MYNDjqZ+Kl/Uyw++RQhiVms0pj06kZ42JmysrcUofzWbfwBLUPaTlnvw9rexlRx8uzTaOy6/hxXA5MhR36LsizgCuRPz2J4DSFGzt1caCEAGvcvXryoFsuUl1aNSAcujTrBsmYLPMzJnyO1SzVLVEy8h6N7tuSqTKkJBQxRpgKa4VGJZIlMzoJ3yTeZAgJlcgUsDHXQx8uRCeXqDoX7OJa8z7EfYFOV+xwXAS6OSx4ujksBXBwXD2VZWFB6Nqp2F56UiX71HDG3Ty2kJsazGu8nT52C0KURRJ49c/kVtxD64dy/v7P/U45GqmxEQQgqJ3+a5h0/by1y6g6CQM8InVx0UNHJCf9ceQGFTArpvQMQ+J3DVxMnsuh4Sr/zTm3MLcdFpiz34bLWxskZEhx/GInt1wJwP0rpoqSJIi0e8udXWSDfyt9ms0h7iq4v7Jgo3SKJZBLLlF9ZMxbd1KkKrFoMQbhu/nRVPd31EffgEq4f3obslw/BmpKdAgcp2wA9QGv6G1PMwn7vcOZ28Tw2XV3sh3yTe3k6wsxQp8z1YZVbBRV8KCxPrTbAxbF2tzG5NNJDLQ/IewNcHBcPpXlQfp8LlIoLzD36mF2gc3rVRPfqFiw929KlyyBwrgexZy8IzJWuFYRIIIBX4gVcO7SVvaZUMmPHjlW/T1O5i35fijVXI5S5XaWZ+KK5K876JyEwIQvyuEBmLa7pZMl8izVT9bxTG39Ay5C2Uxb7cHlo48jkTOavv+PSAwRm5e/b8sjHkPtfRZOK+li84Ld8FbbyQimiKN0ViWVWgEQjS4XQ2hW6nb+DXKQsBa/CSFeESrrpyLy7H343zrGMGJrQQzHla6b0Upo3expjvEOSsOdOKA77RrCgQF2xEF1q2jNrcrPKVurqmtreh+nY/P39WaowKhlOD/zamO2Bi2PtbWPaLpWhpqJc5NqYN20dz1bxDo3B0d5B+V1Iy5bih733WVANFeZY+Ukd3Dt/BPN/+w1pZq4QefaG0LIiFJIsGBoaIFOigK2JLizubcODq6dYoN3KlSuZG4XmTXfk5OnwMajHcrlaIRWt6lbBwXsRkMtlkN47DNGzs/h26jesaAhtoyy3cWmDt2/JB3iWdBsHxKTigHcYdt4MRFxm/kyksqfnWBDfD2M/ZXli815jeaEbNFmSVaniHj9+rHzD2Bqimp0hqvqqnK+K9m5GsAy7gm0b1hS4TcoLS7NIeW/6VCjl2IMoZk2+FagM/nM0N8CA+k5sUWXD0eY+TKn3qKgFiRNtRVWplNpXG8V9eW9jgUDAipZQXum8cHH8Do3B0e5B+W3wi0rF+G138SI2HV1q2qGLZRx+XzAPgZkGENXt/bJIQQ6qi2NQt0Fj7LwTjormepCdXY6gR96sH1GAXrNmzdTbpMTmn89chtTqvSHQNUQ1MzkyhQYITsxmAUWyqxtRr7I98y2mJ9qy3salEd6+Je+m86HamG6uvmHJ2HsnBFtvhkCRxz9ZkZkC2YOjEIffw3/b/mUxAEWBhJ3KT/nS6SPIkAmVD8o1OkGgUaCH0sFNbFMZLlkBWPfnH7h3L0/KRgATJkxgVckKaofAuHT8dzcU/90NQ3RKNlvX3N2KWZM717SHvo5Ia/uwyiqoGRCpTVAbk6GDinqU1jbWduQl2Mb0QFxYoRMujt+hMTivRxsG5aJAN6OZBx5AKlNgeG1j3Nm5FHdDkiH26qPMPSyTQBx0A/OHt4WvrAK23wyBh5UuonfPQlx4ECtxSRW2qKSm6kaw4Z9/la4ZVdpAIM1CNQcz+MVmsTaT3j8CHf9z+OmH71nJ0MIu2rLUxqUV3r4lH+D5MdqYMl5cfxGPPbeDceh+dP5jCn8I2ePTaF/LCcuWLmE35KKQc3Aqbt26iTMvpDj1MAqhDu0hrNWVPfyqMBTJ8UufOmjupIt1f/2FdevW5dsOleKlh2lqk7xQ0N4l/1jmdkFV+CQyBav217tuBSaUyU9Z07LG+3DJw9u47LYxF8fv0Bicsj1gUBGAWQcfYdedUNiZ6MIl4hwu798EgUcr6DQdDoVcyvwW6+nHYPXvv2H++QgcfxiFmlYi+P3zHTKSE5gg3rZtG6sJT1Dd+wnTZ+F8jhurdKcjz4GRoT6SsuSQJ4ZBdvUfNKteCYsXL2bp2cp6G5d2ePuWHcvx665zyjiz+UoAbgQrK1JqInt4ArJnF/Hz1AksbaJm+d03tUWQdXscj7XDjjvhCDGsAmiIZHtDAZZ92ghN3Kxw4sQJVhK3IKg8LpUbJl/LvCSk5+CAjzKI72lUKltXzd6E5U7uU7cCrIz1Pnr7lgd4G5c8XByXArg4Lh60ecAIikvH+G3eeBKZgmau5ojavwBP79+F0LUxRC1GAqlxkF9YjekTRuLzUaMxbqs3rj2Ph6eFFD5/fgNpThZzoSDrj+oBKyAgAEOmL0aMS0cIdA0gYIVrle0kf3gc+s8vYNb/ZrC0bkVtL21uY22At2/JB3iWpjZOzpTgxMNILD72sED/ZOmVf2CS5I+Vy5agTZs2+f2TX9MW8clp+GnDEZyM0gM0gvdq2+ljxbDGcLMxRlhYGMuiQaXmC3K5oAw3BQUP0mzUw/AUJpKpSEpKlpS5cXSobocB9R1R1UyBCvZ2H719yyqlqQ+XVeTccvzx4eK4fA8YdHOctuc+0nKk+KqlE64t+RLX7z2BwLUxxO2+BjKTYf9oO9YtWwDHytXwxcbbeBCeDE+jZNxeO43uVCzobsWKFSz6mth38DC+23YdcrfmufYlT4pgvsXt6lbGwoULUaHCqywXZbmNtQXevuW3jaNTsrDvTggWnvLP9548Lgiym9vQqWE1jBs7luUwLuqxxyUk4esVO3A1u1Ku9Z2qWmLBoPqwNNJl96Dp06fj0KFD+b5PFTeHDx+OHj16FJj+jCzhJx9FYc+dMFx9HkfDEWyMdDCgQUUMalgJrtZG6gJGOTI55AoFhAIBdEXCt8qCwSn9fbgsIefi+OPDxXH5HDBypHIsOP4U/1wNhJWRLpYNqoNN0wbg+KVbEDh5Qtz1JyAnA53Fj7Fszo9IyBbgs39usUCZWgiG9+a5bDuUVWLGjBnsN0skEkyb8zv2RVuwTBYqFGQtfnwSxkGXMPeXWejXr987ReBqWxtrG7x9eRsTz2PTsOLYPRx68iqFmwqZ3wXI7h3E+JHD2XVco0aNIl3LAS8CMWXpFjw0rp9r/WeNK2JGz5qsSl5qairzSV6yZEm+71OO5gEDBmDw4MHMR7mgfYYmZLAgvl23QhCVmsPWNXKxRL/6jmjtYQMD3VfxDAIIYKQngljEx5G3hY8TJQ8Xx6UALo7L34ARkZSJr7Z7s/yiDV0ssHKwF5bO+RHb/lkLgV0ViHvMhkAuwQ/GpzDu17/xNCoFn/19C7Gp2fBI8cbDA3+ym9Mvv/yCL7/8km0zKioKg6YtQrBt81xR64rkSEiv/ovuTWpg3rx5LL9neWhjbYS3L2/jvO4Ld17EYtr26whKz+8DLL22CeaJTzFq5Bfo27dvkeIGbty8ie+Wb0WYc5dc66d3rorxbSqzcYViFTZt2oS5c5UP4Hmhqn8kkkksW1tb5+vDUdHReJ4qwn/e4SwuggwBhroidK5phz51HeFZ0Zx9ViQUwET/3YqNlGf4OFHycHFcCuDiuHwNGBefxWLKTh8kZkgwtpUbvutcFb/M+hn//LUaAlN7iHvNgUgkxFqj9ejUYxDu2PTFyH9vIyNHBofgU3h+fjdzn1i9ejWb6iTOX7qCMWtPQ1KpkXo/CgVZi8/AIuwKfps7B927dy83bayt8PblbVwYlDViy5m7mHM2AnJB7owyisxkSM+tgmdFCyaSe/XqBTs7u9f2s/37D2DW+n1Iqzsk13uL+tfBoIbKWSfKBbxlyxaWFzlXeeuXUNBex44dMWTIEOYTTa/z9uHE9BzsuRuKA/ci8DhCGYDYtZY9fuxaDaYGOjAz0OG5et8SPk6UPFwcv4HffvsN+/btw9OnT1nUMAU9ka9m1apVc5WkpBr2O3fuRHZ2Njp37ow1a9a8dnDShIvj8jFg0M1txZlnWHU+AMZ6YiwZ6IlONe2ZW8TBgwcBAxPo9JgNsZ4BNpuvR3MH4GyDNZiw6zHo1xj5bEOEzzmW2YQiyhs3bswsS7+uWI+/nwACC0f1vhQpMZBe24gBrb2YddncXGmpKettrO3w9uVtXBQSklMw499TOB6ZP5OFPNSXXfstGnoxodytWzc2ZhRWxvavv9Zh5f7LkDf9Itd7f35aD11rO6g/R1lwSCQXFLxH0P1u0KBBbDEyMlKPETRGpWRKoYAC/tGp+OP8c1x4FgsbEz380qsmur3cB6fo8HGi5OHi+A106dKFTR9RAAQlDaeE6Q8fPmRVimgAIEjcHD16lAkWGoS++uorNihcvXq1SCeBi+OyP2DEpWXj650+uBoQj1qOplgztD4qmOmibdu2eP78OWBsA53O0yA2MsV6p1NoZ/AMe80+x3RvCxjpCCA7twqJz+/B0dGR5TCmIh3Ubz75fgkeG3lCINZT70v25AxsIq7j94XzWUqm8tLGZQHevryN35art33wwx4fhArzG2OkPgcg8juD9u3asop4ZOUtKDUcXdOLFi3CrltBEDXLLZI3jmiIttVs1YagHTt2sDLU5MZVGHXr1sVnn33GLNgUxEdBe7QQJJYP+UZi0YmnrFT1p40r4adu1WGkl99thFMwfJwoebg4fkuoJjYJg4sXL6JVq1YsNzH5cJJgIf8rgqzM1atXx/Xr19GkSZM3bpOL47I9YNwOSmD+xVRlim4E/+tRA/ExUeyBi2FoAZ3O0yEytcaqRonorncf6xO9MO+BKaz0BUg6MBeZMUEs8IasN2ShuXvvAYYtP4xMe0/1fhSpccxiNLxzExagZ2JiUm7auKzA25e38btC95G1Ow5i/RMBJPoW+d6XnFsNg8QAdOvalQnlli1b5ksN9+TJE8yePRtX43Qhbjg413sbv2iINlVsmAsElWBWieSIiIgC+zGND1QelzLpkNtFjdp1WAERuRygoSMuNQc/7X/AiqNUsjTEkkGeaOhi+QF6gPbDx4mSh4vjt4Ryx5LV7sGDB6hVqxYr49m+fXskJibmmrp2dnbGlClT8M033+TbBrle0KI5qFWsWJFtgxcBeb/OTA8v9LBSGoQbWUjWXw7E4lPPWNqi+X1rsapSlC6Jcogy9E2g02kaBOYOWNCvFgbWd8LCk35YdykQtnoyRO6YAVlaPJo3b44NGzYwwbty0x4sv5UKmL2ajpT5nYdj7G1WYlpdMppyoN7fDcQ+VVYRqzPorfLBakMblzV4+/I2Lo5xx9vbGwu2n8Jdvbr5+1hCCGRX/oGFMJPFLJBQptLVquuZvk/3tTm//oog41oQ1ckdq7BueD20r2bLRDJlx9m1axeLf6CcySrIWkyLpp8yBfF98skn6N+/vzqIj9K7bb4RjIUn/Fiat9EtXPFNBw/oFVKamvPyHPJxuMy2MelBCwuLNxaFEyjoSi1FjUXTRElJSbhy5QpbRxbjL774IpfYJRo1asSmzMk/OS/0ZE5+oHl59uxZiVj7ygt0fqhDkWvLxxZulBT/11NBuPwiGS6W+vituxvsDBRYsGDBq1yiugYw7jkDUhMHfNO6Ivp72mDBmWAceRwPG2EaQrdSKrd0FkhHfUYmk2HSsu14YKBhLU5PgOTqRnzavj4T3LmmTP1PA0GXAaEOIJcALi0Bj45lpo3LIrx9eRsXJ3St7j90BFt8EpBe6eVDswZSv4uQ+h6Gvakei5Uh/2QSsSrhS/E2f/y1HhnVukNcpVWu7/7a1RXtPCxYxgn67JEjR9gDfHh4OOvHZCwiAxKlgLt27Zr6Hknl6Vu3bs1EOT3IUxBfUEIW5pwMxOPoDLhZ6WNWZ1dUtc2fU5nDx4myPhanpqaya1CrxDH5Fh8/fpwJYycnp3cWx9xyXHafpqm7+oYm4etdvghNzGSW4rm9ayLg6SPmix4YGKj8oFgPpr1+QpZxBXzb0QOjWrhi8s57OPMkBvbyeITs/BmQ5mDixIn44Ycf4PciCJ8s2o9Uq+rqfcn8L8El0RtLF85nifrzcfx7IO4ZYOEKJAYC1lWBrgu0vo3LMrx9eRuX1Lh048YNrNu2FxcyK7J0kXmR3t4F+fNrqOLihB49e6JXr95wc3VFWloqc5/YsHU3FA2HQuhUR/0dfR0hC6qj0tE6IqFaUFOeZJW7BYlkcquwtLRksTn37t1Tf59EAH3Wy8sLUpkcay++wMpzAey9ye3cMa61G8+DXAB8nCh5uOW4iJCwoYwCly5dylVO813cKvLCfY7Lhh8WTRGeexqD8dvustc/96iJoY2c8NdffzGLMQV0MoRimPT4HtnmLhjb2g0T2rhj9KY7uBWUANuMIITtXQAB5Cwn8YgRI7DpwGnMupICiJW5ixXSHMgvrcWkAe1ZH1NVxcvH7b+B+7uU35NmAXU+ARqOes/fyH2OSxLeviVPeW/j+Ph47Ny5C38fu474an0h0HkVzEsoslIhvbmDZb2oU7sGs/D279uHpXWbP38+jl7xgajZCAitX+VUNjfUwdSOVTCoQUXoigRMGFPMzcqVK5UBxwCzgo0ePZr5Oh87dgy7d+9ms7B0Dsjw9N133zEr88PwZHy72xd+0amoW9Gc+SJXtjH+4O1UminvffhDwH2Oi/DEPWnSJOzfvx8XLlxg/saaqALyKDiBfKkIPz8/VKtWjQfklbMBIz4tG11XXEZathT/jGgIZ4McfP3112oXHIZABKMuUyCxqYZhTSphcjsPVvXuaVQqLGPvIer4Gujp6eLPP/9Ehw4d8OVvG3Eh/VWJZ1nANVRJ98XK3xeiZs2arz8g8jn23QnE+gE2VQHPwcXic8wH5ZKDt2/Jw9v4VTvQ2PTP1p04F5wDoVfffG0lvbsXsocnWN71YcOGsWxNFHBObl6+MRKImw6HwMhK/Xmq9DmmlSs6uhrAxcmB3T8PHz6MZcuWwd9fWRabXAepcNHAgQPx+++/M0sz4e7ujuXLl7Pqe9lSGZaefoZ1l15ATyzE912q4fOmLrzcNO/DHww5Lx/9esiPk1wnyGqsmduY/FBU/p301EtPwpTKjZ6OSUwT5GdVFLjlWPtvenQTGL35DnOLWDrIE8Zxj9msAc0oUEQ4TTdCIIBB2/GQOXmhn5cjJrX3wOf/3EJIQgZMAs8h/vIONvuwefNm2Dk6o9uCQ0g1sFfvQ3JuFX4Y3h3jxo3LF2X+oeDCgrevtsP7cH4od/HORVPw79mHiHPvAZF7i1zvS+/uYQWFbKyt8Ouvv7JAvgMHDmDu/N8QY+QGUYNBuapymuqLMLK5K75o7gYzQx3W5uSTTOKXxDVBmSxGjhzJqvrRzBqN3TRu0/g2bdo0ZkWmTD9kRaYxslllKywe6AlH8/yp6MobvA+XPFwcv4HCatZv3LiRTXlrFgEh67FmERB7+1fC5nVwcaz9A8ama0GYdegR+no5ol+FVJYbmwQzWXcpAOX+/fvQa/EFFG7NWAnViW3dMfLfO0hIz4aO716k3jvBMpbQg9i9oDh8e/pVlLc87D5qpt3FysW/5Zu5+NDwQZm3r7bD+3AhHJsGafRTnA0VYfMpX1xS1IGo5ZhcH5E9PgPZswtoW78GK5BFs6br1q3Dyj/XIcetNYS1u0IgeDX2UrGj4U2dWUyFtbEea3uK21m6dClLG0dQVgty3SBXDJqdVVmRydpMsRTp2VLMP/YE226GwERPjJ971sCA+k7luqoe78MlDxfHpQAujrW7M1NJ1D5rrsLBTB//DqmOvj26sOPo168fs46QJVin0ScQVOuAVlVs8GULV0zc5o1MiRTyy+uR9fwWateuzT43bdN5XEp4lbFEcmENZo/qwywsJLI/NnxQ5u2r7fA+jCLFKEQ4dMb2xwpsPOOL9Lq5cx7LY/yBF9fxTf/W+GrcaDZDRkVEdh46CYFnb4gqN831eQrcG9ywEouxcDAzYOfg9OnTLBiPimoRNBPr4OCAoKAgdZ7ksWPHMiuyvr4+LvjF4Pu991m++A7V7fBbv9qsyl55hPfhkoeL41IAF8fa25kzcqToueoKguMzsGdcEyz6fiIrEEMuOFRdccWKFRB59oTIsxcauVhiaONKmL73PgRyGdJPLIE8yg9t2rTB3IVL0G71XeZ6QShSY1El+CA2/LGcBXeWFvigzNtX2+F9+O1iFCiQ+OSp01h46C5CzL1yfUUhyYI4whdzRnTF0E5NmMvErFmzcPlRMHQafgKh/StXREJHJGBW33GtK8PZyojNrp05c4ZZkn19fQs8rMqVKzN3DLIiJ2Xk4OeDj3DINwKWRrqY16eWusR1eYL34ZKHi+NSABfH2tuZf9x3HztuheLHrtUge3QSc+fOZVYOCl6hnJ/CGh0hbjAItR3NWFo3mh7UhxTJh+ZDkRiKgQMHoV63oZh1Pla9Ten1zfjtyx5sG6Vt6pAPyrx9tR3eh9+d6w/8MX3nHYTK8udfFaXH4qvuDTCkiRsunz6KlatW4UW2EcT1B0Jgmru0tVAA9PKsgAlt3VHFzoSJ5PPnzzORTAVM8kLjIFmRp0+fzsbXo/cjMfPAAyRmSJgr2+xeNWFm8HHiMD4GvA+XPFwclwK4ONbOzkwD9MTt3mjpYY3JniL079ePWVnatWun9J2r3Bzipp+hip0xWlexYdXyjBQZSNw/F0iLxcRJX+NIaiVEyl/daPSOz8KFE4fZ9GJphA/KvH21Hd6H3w8Ssod8wvC//b5IkQigyE6HIu4FBFYuEOibAAo5GjkZ4fNWHojyPoslS5YixbYOm0ET6ClTsomFAkjlyhIGXWra46t27qjlaMa2TelSyd3izp07+fbt5ubGfJEbNmyImNQs/Lj3Ac4+jYG9qT4WDajD3NbKA7wPlzxcHJcCuDjWvs4clpjB0rZRWeg9o7wwuE83Vj6VAlQovZ/UsS7ELb5EJSsj1HEyw5H7kTDKSUTiwXkQZqdixuy5WPjMguU8Jsjq/FO3Ghg3bmypsxZrwgdl3r7aDu/DxUNypgSLTz7FthshUMhlkD0+BUViOISujSCoUAsCoRAmOkD32vZIf3AGB7ZvhKJGJ4iqtYNApANLQzHMjfTwIjadbY8MCCSSG7pYMpF89epVZkmm4iV5GTNmDL7//ntmRd5zJwy/HH6E9BwZhjdxxo/dqsFQVzmullV4Hy55uDguBXBxrF2dmSo5fbLuBu4GJ2LjiAbYvmQmq/ykQlDRE7ptJsDKRB+u1ka4HZQIg7RwJB9eCH2RAmN+mIe1L15Zi3MOz8Gd0wdQoULptBZrwgdl3r7aDu/DxQuNgzP2P2C52kUZ8ci68i9kSREQu7eAyL05BKa27HMtK+pB7v0fzl26BnH9/hC5NIAQCtRztoCejgjXnseD6uE2drVkIrmFuzUzFFAxERLJJJY1oQDlvXv3smq0oQkZ+G6PL24GJsDFypAVDqnvbImyCu/DJQ8Xx6UALo61qzMvPeXHSpxS1olKCXdYeWcVAofq0G3/NQz0dVnqIsrPqRfnh9STK2BhZoLGw77HuaRXg7bk1BI8v3Yc+vraEXXNB2XevtoO78PFj0Qmx99XArH8zDNkSeQQhtxB5vVtQFYaBHYeENfpAWGFGtCTZaK9RRwCLx/E/Yg0iBoMhNCmMjMa9K5XCZk5Mhx7EMlcLjydzFjKS8pMIRQKcPPmTSaSL1++nGvfVCyJqpDq6elj47UgLDrxlB3P2NaVMaWDB/TEHz/LT3HD+3DJw8VxKYCLY+3pzNefx2PohhuoWcEU89paoXePbsjJyWHvCWwqQ7fTt1CIdGCgI0KmRAZxyC1kXPwbFd08IGr6OSKEr3zipDe3w2f3cuaKUXjk+FPAplqxVLcrDvigzNtX2+F9uOQgC+7MAw9x8VksDEUKZN7YjqwnF8lTGUL35iw4WaBrCKPYR+hql45zJ48gRs8RIq/+EJhYw9ZQiJGtqzC3td13wpAjlaOqnQkmtnNH99oOEAkFzBf5l19+wd27d3Ptm6rtDR06FAExqZi62xf3w5JRzd4ESwfVRY0K+QMItRneh0seLo5LAVwca0dnTkzPYX7GKVkS/De6AcZ92g8BAQHsPYFlJeh2mQYF5QhV8fgUcu7sgWOdFkiq0Q8S3Vf5i2X+l3Dk56Hw9PQsUs5R1PkEaDgKHxs+KPP21XZ4Hy5ZZDIZdl59huWXwxGbmg0reSIiDy8DkiMBA3NWclroVAeKzGR0MI9HDdMcrPlrPbKdm0BUuzsTzzXsDDC5Y3V4hyRh641gZOTImIva+NaV0cfLEbpiIctq0bt3b7Y/TajISI1atbHm/HOsOufPsmNO6VAFY1u5QSz6sMWhSgreh0ue0i6Oy0ZP5mg9FCBC+YmjUrIwp3ctbFg2Xy2MYeYA3U5Tcwlj2d09TBibNeyD2DrDkCPUgzzaX51Af9mnTQsXxgRZjGl7lq7Kv5R7lMPhcEo55CfcvooFTk9pyQLkEkQWMOgzB1ZtRrAZMem5VZBe3sCCkc9mu+FffxH+WPc3BntaQ3pgJmRPz+FRZBrGbfVGUEwKdoxugsnt3BGfls3G4DaLz2Pz9SDUqO2J0NBQ7N69O9f+u3btiob166Gzkwz7JzSHi5URFp/0w8C/riMwThn8x+FoO1wcc0oFZL04/TgavevYQnjmN+zctA6QZADGVkpXipcuDwIoIL3yD2SPTkHs0gCZ1btDnhoD2a0dEFi7QJEej8/ccjCgX5/X75BcKchinBCo/EtJ+TkcDkdLMDXQwa99amHv+GbwsDNBaqXmsPlsOXScakEeeBOSgz9DHnwXqeYeGHcwDPKK9XDq8D400wuD5NAsyEPv4dTTOPRbc4WVkD4+pRWmd6mKbKmcFQJpsfA8/rr4HHUbNmGlp2fMmKHeN1n8KKXmgu8nYHFHK4xu6Yp7oUnouuISNl2jCnzKNHIcjrYiUJDJrozD3SpK9zTI06gU9Fp9FXamelhb+S76TJiF9Bw5YGgB3T7zARNlNLZIIUPWuT+gCH/Apg91es0ikzMcXxxGhGt3KHT0UT/xIvasW/rmUtCFVKv62PDpPN6+2g7vwx++fSlA7p8rgVj2MmDPMiUAUSfWAFmpEDjXh7jRUAgMTGGZEYotX3dHVNAz5lf8PFWkDNqzcoaRDvBt5xoY0MAJ/90Jw7pLL9hMHhX/GNncFSOauUAgzWR+x3kLiVC10o5Dx+OPu6kIS8xkmTAoL3IFcwNoI7wPlzyl3a2Ci2POR+3MFD3da/UVNh23Y1RDzBzUCA9eRDJhrNNjNgQWTuxzIlk2sk4tgzAhCF26dsXJ7CoQOlRHe4NgXIvXR6ahHayfHcL5Tb/DxOSV77G2wQdl3r7aDu/DH699KWDv54MPcd4vFgYiBXM/S7t/BtAzgqjRYIhcGwM5GRhW0wA/D++E7du3Y9Hi35FiURUir74QGFnA0VQHP/eugzZVbbDPOxx/XnjOsgIZ64kxrIkzRrVwxWPvGxgyZEi+Y2vbqSsMmw3D6RcZMNEXY3bPmuhXz7FU55YvCN6HS57SLo65WwXno/Lr0cfwj0nDt52q4sjm1XgQkgjoGEDc9adXwjgnDVlH58MgPZJFTF+K0WXC2EMQhWdpukwYi/zOYM+SH7RaGHM4HM77UNHSEP+MaIg1n9aDiaE+cuoOgsOwhYC+CWSXN0ByfjUU0hxs9RegxYwdaNq+G25cv4ZxXbyAo7Mh9TmAsPhUjN1yF0PWXmVZg8592xrLPvGEvZk+1l58jhYLz+F8sjUu3vbF8OHDc+3//KnjODr7U9SIuwAx5Ph2jy/bVlxaNj+xHK2Ci2POR+P4g0hsvxnCpuDccp5j/fr1ymjr7j9DaO3GPiPKiEPm4bmwEufgv//+w7bjV5BVpSN0MuNha6KHUJED5GG+2PB1b7i6uvKzyeFwyjVkpe1W2wFnvm2Nz5rmDthTRDxm/saygKuIEVmjy4orWHH0LvMnvnjuDLo5CyDZPwOyZ5fgHZrM3N2m7vJBY1crnJrSCmuH1YOHnTH+vRaEbmtuQ9DoU6z8ZwecnJSGDBX3jm1D9KbJsEoPwqnH0ei87BJOPIz6aG3C4bwt3K2C81GmQcKTMtF1+SWW+mfTkKro3KoJi64Wt/8aQodqys6ZFIrsU8vg6mDNpv8OHz+FxfdFEBhZoq1NJs7FGUOREoVpdYFJ40aXiTPJp/N4+2o7vA+Xrvb1CUnEj/uUFfbMhDlIOrsWktAHEFSoCXHTz9h46iBMwZbJXeFub45bt25h9uzZ8A2Khaj+AAgda0FHoMDo1u6Y0NYdRroiXHgWiz/OBeBOcCKEAqBrDVvIHx3Hwc1r8+1f6NIAes0/h0ykz1wsZvWsyfyYSzO8D5c8pd2tgotjzgfvzFQeesj6G6zs84bh9TCya2NIZQqI24yDsGJd9hlFlB8k51ahXp2a2Lx5M4uW7v7LdggqN0dVUQz8skygkMvRET5Yv3S+1vm0FQYflHn7aju8D5e+9qWAvY1XA7HstD8rnmSREoBoCtiTSSCqNwCiqq0hkEkwtqk9pvdpzIqKHDhwAPPmzUOUwFIpki2cYKYnwPRuNfFJg4qsYAiVk/7jfAAu+8ex/TSwF+P54TWIfnI79wEYmDEhTvmXLfUFWPFpQ7T0KKA4UymB9+GSp7SLY+5WwfngrDoXwIQxRUDvWv4zpDIZRM2/eCWMU6KRc3YFOrVrjT179kBfXx8j/reCCWOjrBg8TxNBIdKDa+R5/LFgdpkRxhwOh1MS6IiEGNOqMk590wptq9og0dQdZkOWwLh6K8huboPk1BLIM5Kx9lY8Wv2yDy/i0tGvXz9cuXIF04f3gM65JZBe34ykpCTM2P8QHX8/x6r0NXGzwpZRjXFgYnNWhvpOlBSJDcfA+bNFENh6qPfvam8JweW1bBvxqZkY/vctTPr3MgvI5nBKI1wccz4oN1/Es6pKNRxM4Z7+EEePHoWo8TCI3MhaoUR6bROGD/kEGzZsgIGBAabN/g2xLh0hyEmHTCGA1MAKRs9OYNeK2dDT0+NnkMPhcN4pYG8g7D9dyKrpSQ7PhuzJGYRm6TLxu+iQN3T19PH111/j+tUrGNyworKIyP0jeBGbihEbb2PoX1fxJDIFdSuaY8PnDXD865bo6VkBUQIL6HSZDqNeMyFwqIEXgYFo2rwFetaygeL4PMij/HD4aQpa/3YC3hSEzeGUMrhbBeeDTYMkZSjLQydlSPBHr4oY1rsjRPUHQlSzk/ozMr+LqJX1AEeOHGG5io8cPYoJe/wgtK8Kvax4ZOtbAS+uY/+P/eHl5VXmzh6fzuPtq+3wPqwd7ZuSJcHvJ/2w5UYws5IZBF9BwuXtEFg5Q9xsBASmdqhoJMeG0a1R1V45/fz48WOWH/ny3YcQefWBqHIzVphpQH0nTOtcDbamyiqmL2LTWAq4/T7hkMoVkMcFQXb/KGxlMZj09Tc4cvggbiUYQFSvL4QiMca3cWclqKlsdWmA9+GSh7tVcDgvy0N/v/c+IpOz8FNndyaMhXV6MmFMeTXFAkCRkQT47sfixYuZMCY/4ynrjjNhLJBkMGEsj32ORYPqlUlhzOFwOB8KU30dzOldi5WArmJvirRKLWA9fDkg0oHk8C+QPTqJkFSgy7ILWHz0AfNbrlGjBnbu3Iktf62Ea/RlSI78ClnUM+y5G46WC89ixZlnyMiRws3GGIsHeuLI5Bb4pGFF6Nq6QKfdRMRV7YMZM/8HB3sHDPK0gvTIXMjiQ7DmwnP0XHmJWaE5nNJA6XhM45QtqPrc7b+BY9OUf3PSse1mCE4+ikb32vb4flArCKt3hLhuL1Q0UsDVxgRSBSC9uR1jvvgMtWrVgkwmw8jv5kBSVWlVVugYQpGeiGHOGRg0oN/H/oUcDodTJiCXiMNfNceMbtUhERlAp+NU2Pb4holj6YkFkCVH44/LIeiw8CQeRSSzGI/27dvj7NmzmPfdOBjf/pvlT85KiMKyM/5o+dsZ/Hc3DPKsNLgG7sQv4k042z4CHauYQ1TJCzrdfsT+U5dw6dIl9OvQFMIzv0Pmexh+0anoueoyszjLePlpzkeGu1Vwin8ahATx/V2AWB+QZsHP+VP0umALGxM9SA7PQbjIjkUum8jT8L9BzTD9v/uQB9+F/fMjuHDhAgwNDbF4+Sqs9jeG4GXpaIVMgrpx57Bvw7I3l4bWYvh0Hm9fbYf3Ye1t37BEqrD3COeexkBfpEDalS2Qv7gBUe3uENbqAqFAgHGt3fBNp+pqFwiK/l+5ciXWb/gbMtdmEHn2gkDfGFWNszDL7CgaGccBNPNXYyD+zm7LMmbQfSHnwl9QRDxE9erVERISggx9G+i2+hIKEzvUd7bAkoGecLE2wseA9+GSh7tVcMofsU+VwtjSFVlCI0y6ZsD8zlwizyNcbg5Rk2EQpMdh/7fdsPDoIyhyMiC9tQMLFiyAoVgB7y0/Y9XlMLUwJqyeH8eWFXPLtDDmcDicj4mThSH+/rwB/vy0HkwN9ZkRw3bQr5AH3oL02DzIEsPx56UgdFx8Gr6hSew7lA5r5syZuHjhPLp5GEGy/ydmdfZLFmFoeH+Mju6DIFSAQXIAvmrrgc2jGsHE2Bg6HSYzwf3kyRNkZ2dDER+E7EO/QP7kDO4GJ7D4lC3Xg5hLHofzoeFuFZzix6YaswwgIRBzoxrhWYYRGhrE4Py1OxC1GAlkJOPYtK7YeC0Q8ZkyyO7+h14dW6Ndu3ZIvfYvRv59B0KPVurNifzP47/fv39tTkIOh8Mpq25pHxJym+j6ssLe502dkaxrBf0+v0BYqT6kJxexEtNBCVnovfoK5h99jCyJMh2bi4sL1q1bhwN7dqC2LAA5/01jlfjOJ9mh0/MBmBXeEIkZOSy/8eFJLVDVzhTiev1h2mUKpCopIpNAensXJCd/B7JT8b+Dj/DZP7cQmZz5QduAw+HimFP8eA4G6nyCE4Lm2JpcB26G2bi0dQXErccC2elY3scd6XIxtt0MZSl9jKLvY86cOeyrU1btRaLXCPWm5KE+WDexG9zclOWkORwOp0zju1Pplhbrp/xLrz9SwN4vGgF7ojrdYTlkERRxgZAe/RXy+GCsuxzI0r6RpVdFo0aNWLah1Qt/ge2jbZAcnAlp7AtseWGEFgvOYu3F57Az1ce+Cc3QpaY9smxrwu7TRRAYW6u3oYh+huRd38Ms7hErMNJp2SXs9wnjVmTOB4OLY07xo2uECI+h+D7QC0Zi4OmmGSxlj0Ckg/6OaejaqgG+2+WttBLc2IL/zZzB/Of27N2H06Y9IdBT+pkpEsPwjdNztG/Xjp8lDodT7tzS2F8SyaUgYG9m9+qQ6RhDp+M3ENbqBun5PyC9+x9CEtLRf801zDn8iGWqIMgXut+gobji/RTTx42AzqXVkF78CxmJMVhw/CnaLDqLs09jWL7l7zpVQRKMYDl4ARzqtn61Y0kWYo8th8GdTRAqZPhmly/Gb/VGfFr2x2sMTrmBi2NOsUORxlN23kNypgRJJ1cxi4DIuT4sMkKx+NuRLBo5KDEbUt8jaFi1EoYOHYrg4GB8v/k8hHZV2DYU2elol3YKU+as4meIw+GUS7c09tem6sc+IohFQnzZ0g2np7ZC+2q2rGiTyaAFQE4GJEfmsBSb/1wNQoffz+H683j196iIEysicu0aBjevAumhWZDe2YPI+GRM3uGD3qsvo2llK6wf3gAKgQhJnsPRYuT/clU9TX58BUnbv0NN02yceBSFzssv4fTj6I/UEpzyAhfHnGJn9bkA3ApKgIciHIrw+xDVHwDIZVj/VXc8j8tg78sTwyB6dpblNKa0bSOm/Qp59S7qbTiHn8Off2+DQM+YnyEOh1Pu3NKYSKa/9LoUBexRJby1w+rB3MSQBeyJm34O2Y2tkN7eifCENAxZfwMz9z9AWrbSikzQzODvv/+O0yePo5llBiT7Z0D25CwehCWh/5/X8d/dUKwa4gVnK0PcklZCh5mb4eLmrv6+JD0JPqu/QlM8ZfmWR2++g+/2+LJCJh/bR5tTNuHimFOs3ApMwIqzz1DZUg+Pts2FoFI9CG3d0cHVAPXcnTBttw9kcjlk1zZh4oTxqFKlChYsWYEXdq8C8Az8jmP3ytnQ11dWW+JwOJxyg64R0HAU0G2R8i+9LkWQVbdLLQecmdoaI5q5QGTnDr1esyDQN4X02HwWR7L1ZgjaLTqLS89ic32Xiojs2rULm9evgWviHeQcnAV5iA9OPIrGl5tuo4GLBUvjdjE4C1aD5mH4mEm5rMgXNy+B44NNaOJixnIpd11+GddO7ysVPtqcsgUXx5xig8pDT9npAz2xCLIrG8g5gkUj60CGBcNbYevNYPiGpzKLgYupgE23Xbt2Det9UiEwslJuJOgWts78Avb29vzMcDgcTinFRF8Hs3vVxIEJzVHNwQyi2t0gbjsRsofHIb2xFdFJaSzTxHe7lS52KkjsdujQgRURmf/j1zDx3QHJycWQxAVj950w+EUmw0hXhAcRKTgrboiFf+/JFZB979oF3FnyBSY0smDZL4ZetsTsuHbINKtcKny0OWUDLo45xQLlovxh7wNEJGfh89oGeHbnEoRVWkNgaotJHaoiRypnaX8UaXGQ3TuIhQsXIjMzE2Pmb4DQtRHbhjwuEPP71kL9+vX5WeFwOBwtwLOiOQ69DNgztHSATocpENh5QHpqCeQRj/CfdzjaLjyDs09y+wnr6OhgxIgRuH79Osb37wDBqUWQXvkbqYlxSM9RpoeLS8vGL5eTMHrBJowfP1793Yz0dCz/agCm1shCA6sc/JtYG9296+Nemlmp8NHmaD9cHHOKhe23QliwRPfaDlg5ZQigYwBRnR6wMdLB6FbumLH/PrKkCmZR+KR/HzRv3hwTps9CWq3+6gp4g51SMPSTgfyMcDgcjhahGbDXobotRK6NIe4wBfJgb0ivbUJ8ShpGbbqDr7beRmJ6Tq7vqouIXLyAbjWsITnwP0h99rNsFYREpsCsI0+RVasP9h08xPyXVfw8dQIqhZ/H9zWTESYxQ/+wT7AkrjEzxnA47wMXx5z35ll0KuYcfgxHcwMYPjrA1olqd4VA3wTTulbHmSfROO8XB9mLG7DIisTPP/+Mrdt34KpuPfU2asecxfyZ0/nZ4HA4HC2FAvbWf0YBe/XhYGMFcdPhELo3h/TSOshDfXHkYQxaLTiNYw8i8n1XXURk727UEUUgh4L2nl1Uv7/tZgi+PZuCI2cuYty4cer1O3f/h1+//w7bR9Vn7h2rLgaj75qr8ItK/WC/m1P24OKY815QdaRJ231YeeifOzph278bACNLCKt3QDV7E5b25+cD94HsNMhu78Ls2bMRHx+PmbtuqZO+m/sdxraVcyEWi/nZ4HA4HC1GGbBnz6zILGDP1h3idpOgSI6A9PpmpKSlYcI2H3yx4Qpzm8iLuojI7/NhG3wWkkOzmHsGEZ6UiRa/X0GHTyewz2jSq1UDzGiki0nt3PE0KhU9V13BXxefs9SiHM7bwsUx572Yd/QJ/KJT8XU7d3zRsw1bJ6rbhxX8+LFbdZbwPTFTBumtnWjRsC569eqFEdPnQVCjo7IDBlzCnkXfwszMjJ8JDofDKWsBexObo5aTBUS1urJFdns35MF3cT4gGS3nn8S+uyH5Kt+xIiL9+uHKlSuYPnY4dK+tg+T0MvX7Q9ffxJzrmbj7yA/NWrSE6tsD+vVF+Im/sHd8MzhZGuC3408xeN11BMfz9G6ct4OLY847c/JRFLbcCEYTN0uk3d7H1gksK0Lk1gQtPawhEgiw524Y5OEPIY64h0WLFmH2/EUIqzKAfVaRGoe1YzvCw8ODnwUOh8MpowF7BycqA/YMLGwhbjESCrkM0pvbWWDd1D0P8Mnqc4hOUfoYa6IqIkJZjYa0rQvJ1rGQ+SldLe6FJqHpwkuYOGcl/lz3N1T6eufOnejWuAa2DKuJL5q74HZQIrquuIytN4J5+WlOkeHimPNORCRlYvp/92FuqIPPPORYvWolWy+qN4BNq03pUAU/7L0HSHNYEN7Ub75BYGAgtiW/EsKTq2WgU8cO/AxwOBxOOQjYO/tdW7SvZgORayOIvPpA9vQsZIG3cCs8i1mRt14NKFDAqouInDqFRqJAZO/7Sf3eF//eQbpVNdy5r3S9UNHQyxM1s59g+5eNYWGoi5kHHmLExtuISs4vwjmcvHBxzHm38tC7lLkrZ3V2w5SxX7D1ggo1IaxQA/3qOeHUoyiEJWWzqOOqTtYYNGgQPlu8R13xrlX6RUz96lVQBYfD4XDKNhS0/feIRvhreH1YmRlD7NUXAiNLSL33IzsjDTMP+6HH4mPMt7ggqIjI7l278NfyhXC4ugjy+GC2fsbBx/ht7w2Eh4dj+PDh6s9PnjwZ86eMwPbPa2FgfSdcfBaLTssu4uC98JK1IvOqfVoPF8ect2bNheesEt5nTZ1xYM2vSEtLoygM6DcZAj2xkAVjrL/8guUtlj89y9wpRnzzP4jqdGfftwk4inWL5+SqfMThcDic8kHnmva49ENHDG/sBJFtZYjq9mTZLMiK/CgBaDX/JNacvF+ggBWJhOjVrQvOnT6Fn5voQxx4la0/EiRH/a//wk8//YQ9e/aoP+/t7Y2m9euijWEYy6ShKxbi6533MHG7NxLypJUrNqhKH6/ap9Vwccx5K+6GpmDFWX9UtTeBQ9R1nD59WtmRKjeDzNiORScvO+0HuUwG2fXN+Gz4cNz18cVT5z7sc/Kw+zi06n+8NDSHw+GUY4z1xPi1rycOftUCzmY6EFVtDaG1K2SPz0CamYJF50PRds5+vIhJKfD7Bvp6+HLUSHiv/wGt9JQW5ATDivD69TTcParg7t27uQwwo0ePxn8rfsbOz2ujS017HHsQhU7LLuUrTlIsxD5VVuuzdOVV+7QULo45RSZLIsXSC6HQFQkx1lMfC+bPVb4h1oV+w4GwNNKFjkiIx5GpkD08ARudHPTt2xcLXjiot7FvSkc4OLx6zeFwOJzySx0nc5z7oQu+besMsZEFRDU6QJEYAXmID4Iy9dB+8TnM23MV8kJSslERkc2/TMDSrhXYa7nYAI2X3kJqthTPnz9nGZJUHD58GN07tEZ380gs+8QT2VIZK07y/X/3kZr1qsT1e2NTDZBmAQmByr+8ap/WwcUxp8jcC03G8/gs9Kpti99nTIFEohxMhNU7QqpjhD51HbHu0nMoUqIhu38EM2bMwIAF/6m/P7t2Kho0aMBbnMPhcDi5AvYmda6FSz92RA0zKYROtSGwrwpZ0B3IM5Ox/m4SGs/YhYfBMYW2Wr/WXtg+xF39utPa+7j5NBh//vkny6+vIjU1FWPHjsGJtb9i5+e1WGalXXdC0WX5ZVx/Hl88Z8VzMFDnE6VIpr/0mqNVcHHMKTL/XgsGTVJl+h5DUFCQcqW+CfQ8u8PFyhAPw5OQI1OwRO8d27fFmn3nIKqqzH3cJOMGRnzKBwgOh8PhFIyjhSGO/dgb87pUgq5ADpFLAygkWZBHPkWM3Ag9Vl/D1PXHCy3s0cyzKo59WVP9+vOdAdh89h7GjBnDUrwZGhrmsiJ/0rMzBlhH49c+tZj/8ZD1N1i1Vypu9V7oGgENRwHdFin/0muOVsHFMadIhCZksDLQDRwNcGTHP+r1Is9ekAl14G5rjFtBiZA9uwSD1DBUqlIbgc492Gdkfhewdeks3tIcDofDeSOftqmNO7/2QX3jZAjMKkBg5w5F7HMoslKx77kcXtO24sqD5wV+t4a7C85N9FK/nnUmAj9uu4IWLVrizJkzqFq1qvq9hIQEjB8/Dpf/mYftn9VCfWcL/HM1ED1XX8WTaF44pDzDxTGnSGy+HgR6WDeN8kZ29suSn6b2EFdpxazGt17EA1kpkN39D0OHDsXmpFf5jB9snMFLQ3M4HA6nyJgZ6mHvzKFY0cMReplxENp6AAIh5IlhSBaZY9jmhxixaCeycvL7CrtVrIBr3zSCbqbSTWLHg2QMXHUG1g5OzGLcpUuXXJ+nUtTD+nTG8Apx+L5LVWYM+nLnUyw74w+JTM7PWjmEi2POG0nPlmLn7VC4WRngwq616vW6DQZAIRBCIlMgJVsGyY1tqFnFDX8fvQqBSMw+s7lvBZibm/NW5nA4HM5b06dVPfgsHIpG4hBARx9CCycgPQEKaRYuJJjA87ttOHDxTr7vVbCzwfXZvWCaFsJe343MQbuFpxGfLcCGDRvw7bff5vp8YmIiJowfh7tbF2Dj0OrsfrfqXAD6rrmKZ9Gp/MyVM7g45ryRfd5hSM2Swl0RjnTKaUwFP+gp3skT+jpClrCdIosFYfdg71gROh2nss/UEoShVeNX01scDofD4bwtRoYG2D13PFZ2sYFOzGMITKwBkS5zs8gysMLXRyPQa8Y6JCbnFrFWFua4uuAz2KQGsNcxGXK0/f0Crj6PZ+L4n3/+gZFRbn/gY8eOYcygHvjEKgTjW7vhcUQKeqy6gvWXXhTq68wpe3BxzHktlD5n47UgmOqLcXvPavV63UafsL9ZEjkE0ixIb25nmSguZDmrP7N39gjeuhwOh8MpFnp3bIW7y8eiQeZdIDsNAn0TQC5jRajuyxzRYMYe/LXnRK7vmBgb4fKScXDK8Ffe0xTA8L9vscxKnTt3Zm4Wzs6v7lsqK/LMH3+A357FWP9JdVbZb96xJxiy7gZC4jP42SwHcHHMeS2XA+LwIjYdXvKniH7xCJBmQ+jcAArLV4OJ5PYemOkqWGUjii4mhjqlQE9Pj7cuh8PhcIoNymtMxTwWtzODKOAiIBBBIFBKGSpENf+2BK0mLUVgaLj6O/r6ejj3+wS4ZysFMjH/2FNM2XUPzm4eOH78OFq1apVrPyKRCCdOnMCkoT0wziWRVYS9FZSALisuYfvNkJItP8356HBxzHltHfiNVwMhhAIBh1YBcimgkEO3QX/1+4oYf8j9L0NgYAZJs9HKlVkpmD3m1Wc4HA6HwylOBvXthSurvkHtyGOQx71MLUouf0IhQoyqou2Ck5i9ejNkMmVaNl1dXZz8fRLqyF9luTh4LwID1l5DOvSwdetWjB07Vv2era0tXFxckJSUhG8mT0TogSVY1a8KzAx08NP+B/ji39uITsniJ7WMwsUxp9A68M9j03DBLxaeomAEBocBAjFENTtDbmTN3hfIZZBc20wSGanVe6tb8rOqAjYQcTgcDodTUlC11UOb/sD/mugD3ntYTmQ1JnbYGGqBBmMW4LaPr9oafHDhJDTVeSWmH0WkoNeqK7gTkoxZs2Zh5cqV7P4VGRnJ3h8wYAB0dHRw8uRJTP+8Fya6JaGflyO7N1L56cO+EfwEl0G4OOYUWgd+0zXlAJJ87xRAU0g6eixDhQqJ72EgJQqwqQxRxTpsnU5yGGaO6sdblcPhcDglL2KEQoz+chROrf4Jbk+3QR7srX6P3C0SbepiwAYfjB0/Fun7vobgzj/YPvNzdDKJgIL8lQHEp+dg2IabLGVp//79sW/fPtjY2LBiV6dOncL333+P2rVrIzk5GdO/mYSE48uxqGdliIUCTNrhg6+2eyMxPUe9X+5yof1wccwpsA58SpYE/90Ng4uZCE99bgE6BhDVGwCFqtJPcgTkj5SBD+KandWtOKyuBbcaczgcDueDUqVKFZw6sAsT6+pCdmENFOkJ6vcEJjY4adYL9bYJcGD7egju78K6GaPR3z4BCpkyT7JUrsDPBx/h+733Ub1WbezYsQP169dHSkoK5s2bh27dujGRTFZkEsw/j+qDr9yT0bmmHY7cj0Sn5Zdw8lEUkjMl6uW9K+1xPhpcHHMKrAO/+3YoMnJk0A26RkMLYF4R4trd1a0lufqvMkrYxA7CSsp0bTopYfjxiz5v7dfM4XA4HM77QsKVBOy+VbNh77MB0senoZC/KuKR7dERX+eMQ68FR5nbxNKpI/CFWyYUUmVhK6FCjt13wjB0/S0o9E2xZ88eVtSKLMELFy6En58fDh48CE9PTyaaZ3w7GZlnVmFWZxdk5sgwdstd/HL4EasNQN8hccwFsnbCxTEnXx14mdgQm64HwUxfiIdH/2UtJPLqC4VQxP4ve3gMopinqOzmAlGNDuoW/MzLqnCrcSF+zRwOh8PhFCcNGzbE2RNH0ddVAMnReZDHvnj1pr4x7nuMQoupf+LanXuYPX4IJtUCFDkZkCnkMJKmwCc0CV/seIpHUelYvHgx5s+fz6q8HjhwANOnT8fatWvx448/MjF+5vRpLBzfH2OcE9HY1RL7vMMx6K8bCIxTGoCyJbzCnjbCxTEnH2efRCM0IRN2qf4sQ4XAyhkit8bsPUF6HGS3tqFvLSMEhYRDVLUNW6+THIofRrwKyiuqXzOHw+FwOMWNsbExFv/+O1bNnwmjG39BFnyXrbfLiYBhchBkzo3x6dbH2HruHr4b0Q8zmhgBOZlIExrCMiMUqdlSDFl/E7vvhGLEiBHYtWsXLC0t8fDhQ/To0QONGjVi7hV169ZlVuTfZkwFLv6BCc0cEJWchck7fZCSKYECPOWbNsLFMScfG68GQSQEHh5YA0gymNVYRfbFdXCxNIS1qSFQta16/YgGNuwp+m38mjkcDofDKSn0dUTo27M7Tp85g5biF5BHPUWMUWXoZ8ZA33cP5EIdzDwVjtH/XMXA7h3wWwdbCLJSEW/gCMPo+7A21sX3ex/g54MP0aBRY5b3uGbNmoiPj8fAgQNx8+ZNHDp0CDNmzICOrg4unT+HDdOGoKNdOjMwTd97n5wSOVoIF8ecXDyNSsH1F/GoqIhjmSgEDjUgrFCTvUf5jOWh9zCntwv2eMdCVH8gW6+bEo7pn/V6a79mDofD4XBKEqFQACcHe2zbvBG/tHcAEoKRYN8IaTkKSA7OgjzUF6efJaH9kvOwcK6OVX3cIEhPQKJVbWT7X0ODSqbYfD2YZbMwMLdhYrhXr16QSqX44YcfmDAeM2YMTp86Bc+6dZGakoqDC7+GTao/brxIwMqzrwqPcLQHLo45ufj3qjJ9m/+Rv1jBD3GzL9TvSa9vRo/2zZGiY4eEikp3CmJkwzdYjQvwa2avORwOh8P5AAgEAnwxfCgOTu0EvawEiBsOgtCxJqTnV0N6eQOSUlIxfps3jkfqYUm/ahCkRCHetj4CbpzGJ152uBmYgF6rr+J5Qg7+/PNP/PTTT2ybW7ZsYVZkcrk4cvgQZs6cAT09XYQfXAKd5BD8fTWIZX7iaBdcHHPUJKTnYL9POOzEGVDEB0FYrT0E5hXYe9Kb22ChL8DUOcux8b4coqafsfU6aVH4bnhP3oocDofDKfV4VnPH2Z/7wlQkgajJZxBUqgd54E1k7ZsB0+TnOPYgCnOvJOPTJs4QJYchqUJjHD1yGD91qITY1Gz0//MaDvlG4KuvvsLmzZtZOetbt26hS5cuePToESZOnIjTp0+jomMFpJ9cDh1JGn7cdx/eIYkf+6dz3gIujjlqdtwKQbZUjvBz2wCxntptgpA/v4aZv8xDbGwsvKMlEAjFbP3ohtYsipfD4XA4HG3AydIYh6d2hKWhGDqtxkBAroNZqYg7uAAu4WeYn/DWx9moVc0Dwox4pDk2wtLNB7CspzPMDXXw9c57mH/sCdq0bYejR4+icuXKLDVcnz59WAERDw8PZlE20QEyTi6FXCplad4oUI+jHXBxzGFIZHJsuR4MY5EU8uC7ENXsBIGBGXtPHuIDJ3tbDBj8KXbu3Alx40/ZenFmAqZ++ir3MYfD4XA42oCzlRF2jGsOM0M96HeYBIFNZbb+2dldEJyYh05VzOEbmQmRsTVbn1WhHqb8eRALO9qhXiVzrLv0AiM23oKVQ0UmkDt06IDs7GxmUf7111+ZYF6/fj2EyRHIvrSBWZ3HbLnD8x5rCVwccxhU2ScqJQtJdw6zPJDCGq+q3sn8L2PIkCFIS0vDgSv3ITA0Z+vHNTDjVmMOh8PhaCXV7E3x78hG0NPVhVnPHyCwcGLrI4P8cejHfljTKA4W4lfWXolDHYz84wTGe+qhfz1HXPaPQ+8/riIyQ4CNGzdi8uTJ7HPkkzxs2DBWLIRyJMuD70B2/wjuhyWzCny8vHTph4tjjjp9m1iggPzZJYg8e0Ggo8fWUwlOQdRjDB48mOV51GkyTNlxJBmY+mkP3nocDofD0Vq8Kllg/WcNIFUIYffJXOhYKONsIMvBlyPHYp3R3xhg/FD9ebl9DYz5+wra2aRjRrfqCE/MRN81V3H6SQzLXkEFQgwMDHDx4kV07doVjRs3ZtksZPcOAWG+OHgvAn9d0ihKwimVcHHMwf2wJNwNTkS2/1VA3wRC9xbqVpEHXEX7du1gZ2eHFX9vg9CqEls/qZ4hhELefTgcDoej3TR3t8bKIV5IypTAcfhCNG7bBVDIQM7H3RZeQK2QnVhf3Qe2xsoKsApbD0zc7gOLtEBsHdUYhroijNvqjaWn/NCjR0+W7s3JyQlBQUHo3r07E8gdO3ZAzqX1EKZGYeHxpzj3NPpj/2zOa+DqhsOsxoT8yVmI6/WHQCiErYkuS+UmC7iCTz/9FJcuXUJ6tVeW4q+HdOUtx+FwOJwyQZda9lg0wBMRSVlI9voMg/r1BCtup5Dhp92P8e/Oo1jZz4MVBmFYueLbgwEIe3IHh75qgdqOZlh5LoD5FVeqXAXHjx9Hs2bNkJ6ejlGjRrEgveoebsg6vRICaSYm7/BBQEzqx/7ZnELg4ricE5OahSP3IyCP8gN0DSGs6IlGrpaIS82GPOIR7Ix10K5dOyxcvhpCx1rsO1NqU2J13nU4HA6HU3YYUN8Js3rWwIu4dPhVG4mJn/cHKDOTjgHO3QvCyAHdMcw1G3pi5f2PfJS/PRaGkydPYs+4pujr5YgzT2LQ54+rSJbrYceOHfjiC2WtgDVr1kBfXx/GyET2uTVIy5Lgy013kJwh+ci/mlMQXOGUc7bdCIFEpoDs6VmWuk0kUMDN2ghyCNSBeBEREXgoflXuedLgLh/1mDkcDofDKQm+aO6Kr9u7wy82C/eqT8TIr6YDOoas4EdSUiIWTRuDoW6vBK3A1B6zL6dg3dY9WDrIEzO7V0dgXDp6r76KK88TMW/ePCxZsoQVyvLx8WGB7Yqop5De2omg+Ax8tcMbUpmcn8xSBhfH5ZhsqQzbbgZDkRYHhTQHQmsXDG7kjItPo6DITAbCHjBx/PuyFRC5N2PfGV05gw0SHA6Hw+GURSa3c8dgL1vcDkpEuEsXDBo8JNf7m+ZOwdA6ylSnDCMLLLknx/w//sWXLd2weWRjiEQCjNx0G3+cD2AB7ZT/2NbWVv0Vud95yPwvsYwXC44//ZA/j1MEuDguL+SkA7f/Bo5NU/7NSccR30jEpeVA9vQ8RM712cfcbY0RmSphRT9atWzOLub9/jnqzYzs8SpYj8PhcDicsgYZgCa3cmLp2s77xULaYCi6dX+V0z8rKwsnl05Bp6qW7LWFMBvQNcK650b4ftGfaO5uhUMTW6CqnQkWn/TDV9t9UL22J06cOIG6deuqtyO7uR2KGH9suBLIS0yXMrg4Li/47gTu7wJi/dhfxb2d2HgtEApJNhPCwkpeqO5gituB8ezj5FJBgXgbNvwNUS2lG8VXtYXcaszhcDicMo9QIMBvfWuhS017HLkfBevOE9GqdWv1+5EREYg8uBh1HE2RpDBAdaN0QCTGzmhbfDVnBSpaGmDv+GboVtseRx9Eot+aa5DommL//v0YOPBl9Vm5DJILf7KUqT/s9eUlpksRXByXF2KfAmJ9wNKV/b37/P/t3Qd4FFXXB/D/bN/03jsloYUWeu9NBBtgAxHs+lrxtXwWbFheC4ooKhZsCBYERXpXeg+BkIT03nu2zXzPvZNsEggImpDs5vx8xt2dmWw2k2Fz9s6552QjNrMM4rm/IHiGQtA6YVCEJzaeyoWYfQZeWgkTJkzAol8OWJ/i0VnjW/VHIIQQQq4WlVKBxTf3wtCOXlh5KBPdb3kGMTEx1u2HD+xDcOof8HPR4ZzZHWN8qnmVp9/KQ3HH029Br1bgw1v6YMGESMTnlmPqkj04lF6O9957DwsXLpSfpKYc5u1LYDIacddXB6nFdBtBwXF74R0FmGuAomR++0VBFF9tObMNitqUihqzBRYJEBN2Y8aMGdixYwdUMfIn3NsjDDRqTAghpF3RqpRYdntf9A5xw2d/pmH0A4vQtWtX6/ZfvvsS17pnQSEAJ02+uK2jCJiN2CFGYsbji2CxWPDAqI5YPicGFouE2Z8fwOd/pmD+/Pm8sRYjFaXD8ucXKKxkFSwOUIvpNoCC4/ai5ywgeiYPkrM63owNWXpeqg1l+dBE9EeUnzP2xOdCMlRATDvCUyrmvrjU+uUv3XVdq758QgghpDU4alX48g757+R721Nw01OLER4ebt2+5OX/4u4eOuSXG3DUFIDH+mqBmgocVEdj6sOLYDQaMTrKF2seHIJQTwe8/FscHl99HP0GDsZff/0FrVZb22L6d8RmlVOL6TaAguP2QuMI9JsHTH4TK6oG8RFiy+ktEPyjYFFq4euiQ1oJyz/ei8ED+/MJB6ph8/mXjnHOpVFjQggh7Zargxor5vXnwe3rW1Nw18sfw9/f37r9vSfm4M5+PojLLsNpBOPlkR5AVRFOO/fBhIdeR3V1NTp4O2HNA0MwJsoHPx/JxIxle6Fx80VsbCw8PT1hOfYrxLSjvMX00u0JrfrztncUHLcz1UYLvj+QCqk0B1LmKShC5fypsmpjo4l4Y29/2Po1nz0jFzEnhBBC2isfZx1vF+3jrMUrW9Lx2P+Ww9HRkW8zm8344+2HMamrNzacykG+UwQWTw0FynKQ7N4Xox96C2VlZXDRqfHp7Bg8NLojTmSU4tolexCXb8Dx48cRFhYK857lEIsz8ebGeKw9mNTaP3K7RcFxO7PmWCZKq8081xiCAo6dByHQTY/YzFKIeYlwE6p5qRn1WDk4jjScpVFjQgghBECwhwMPkJ11Kry0NRsvffSd9bikpqSgbMtHiA5yxQfbEiH4dsKns7pBKMlAtldfjHpkMYqLi6FQCHh8fCQ+urUPqowW3PLpPnx/MJ2nWER1DId5+4eAsQoPfX8EG/ceo+PeHoPjXbt2YerUqQgICOBB2Jo1axptv+OOO/j6hsvEidSh7WK1iy9FkiR8vicZMFXz8m0spaJGUsIsijCJ8kQ8VmLmurset37NmtfubZlfPCGEEGKDOvk646s7+0OtVOC13UV4+4vV1m3bt2xEr/L9vILFgh9PwCssEt/OHwBFUQoKffpgxOMfITcvj+87qYc/fr5/MPxd9Xj2l1g8/fNJrN+wCZGBHjDvXAaotLj76yP44Zd1rfjTtk+tHhxXVlaiZ8+e+PDDDy+6DwuGs7OzrQvrV04urF3MH1/CX0mFSMirgCVhD2A2wD16DF/PertLxmo+IWDUqFEoiZnH13sUnIBer6dDTQghhDQQHeTG0yMMFhFvHajCW8tWWLd99sHbmB1exStY3LXiMMI6RuLnh0ZBVZCIMp+eGPnUF0hNy+D7Rvm5YO2DQ3i5uO8PpOHW5QewYtUahDsYYTm0CoKLLxb8HIfXFr0OUaQ20+0mOJ40aRJeeeUVXHfdxashsJmcfn5+1sXd3f2qvkZbqV3Mg+RL+PzPZDZ8LKdUKJQw+8nlaGrMIsTkfYjpFY23PqsPsH989pYW/xEIIYQQWzSogydPjSirNmFJrIDX3vvYuu2Vx+/Bo4M8UVBhwPyvDqFzZGf8/tRUaPLPoNqnO8Yt/AHxief4vm4OGnw5tx/uGhaOw6nFuOnTQ3jlo6/hXRLHW0wrArvjo7+yMXfuXJ63TFqeCjaA1dtlbYxZUDx69GgeTLOZnRdjMBj4UqfuZGKfuuzqk5dXFJB9AihK4SPB8IpkP2STu6YWVmLb6TyIGceBigIIAd1RIyqt21lKxfA7bsKSgi78sSLjKMJCJzU6Xuw+S82wq2PYxtAxpuNr6+gcpuPbns7hUZHeeOvGaDy2+ji+SHXGMy8twmvPP823LZw/HQ99vB7L/srEoz8cw4c398am52dg0sLveIA85fV1WPXQGPTq0ZWPMj89KQpd/J3x9M+xuPuHM3jgxY+wZMEdMLj6Q9ltPLbu+RxTpkzBF198gYiICNgysZXiicv9foLEXl0bwfKJWWvF6dOnW9etXLkSDg4OvKZgUlISnnnmGTg5OWHv3r1QKuuDu4ZefPHF+u4zDZw9exbOzs6wG6YaIGU3UJoBuAYBYcMAta7JXd/bmY6VR/Ng3PAWxJwz8Jz8CCp9evBtYkEKNDveRbch43HcX+4f/0rvaowdMfSCk6q0tBSurq5QKFr9ooNdomNMx9fW0TlMx7c9nsM/Hc/HW9vT0MlLj15FO/D1559Yt0159UdsTSjFHf39cO/gQOTl5+Pmt39FpV9vCIXJ+ODGLojp2c26/5ncSvz3tyTklpswzMeMrR8+DfWkpwC9C4wb3oRjdR5ef/11DB3a+G+0LRFbKZ4oLy9H586d+fd2cXGx3eD4fOfOnUOHDh2wZcsWjBkj58xezshxcHAwnyV6qYNhryoMZgx8bQsq8tJhXreQp1Q43/4hDJL84cK892uMDFZht+91ENQ6Hjyf++zhC05YdjLn5+fD29ubguMWQse4ZdHxbXl0jOn4ttdz+MMdSXh701n0DXGD8+GvsOmP3/h6J1cPRD3wMS/d9u6MnpjWK4DHI+MWLEWRTx9ezWLZzT0wdthA63OxdIwHvjuKgynFCNMbkLD2Q6hGPQAYq2H6/VUINaV4+umncd9999lkRSmxleIJFg+yLIS/C45tIq2iIXYpwcvLC4mJiRcNjlmOMlvOx34B7XHE86fDmagySRBZrjH7EOLfxRoYS2YDxJQDyHAfBCFIHnV+qJ8rVKqmTw32j7C9HserhY4xHV9bR+cwHd/2eA4/OKojKmrMWLbrHIYNnoegU7HISEtBRWkR9IdWwK/zTPz355MI83JE7xBP7HjvYUx47F1ke8Xg7h9O470aI6ZPGMmfy8dFj2/nD8Qrv8dhxd5UuI27H2UntkAZPQUukx9H2ZqX8Nprr+HUqVN49913odM1fdW4LRNaIZ643O9lcxFORkYGCgsLG3WmIRdntohYvjtBbgt9bj9f59FzrHW7mHwQbo46JAeO44+lkizcfdtNdEgJIYSQKwz2npoUhZv7B2N3QgH6PvgBW8m37d70G0Yqz1grWGSVVPORy20fPInw4sOAix8eXpeKb9ZssD6fRqXAS9O6Y9H1PWBQ6KHuPgFiWS5qHP0QNfsVKJUq/Prrr3weFmlerR4cV1RU4NixY3xhkpOT+f20tDS+bcGCBdi3bx9SUlKwdetWTJs2DR07dsSECRNa+6XbhMziamSUmiCmHgEsRp5SYfCOajQRr8SiheDszR+PdczgOUCEEEIIufIA+ZXpPTAl2h9/xObg1vd+t25bsfhV3NpBtFawqDKaebnUjR88hS7lRwAHDzy7rRBLv23c7+GaaH8snxMDNyc9FC6+fN05iwemLFjMClDxCXqswx6xo+D40KFD6N27N1+Yxx57jN9//vnn+YS7EydO4Nprr+UJ1PPmzUPfvn2xe/fuJtMmyIW2nEzlt1LOaX4r+HeFsTalQizOgFRwDsouo637PzqXRo0JIYSQf0qpEPDujF4Y0dkbqw9nYv6S+gD5o2fuws3dnRGXXcYrWIiiBI1Gg3WLn0aM6QSgc8IbBw1449P6fg4WUULPYDesvGcgOnvVp09syFSj33V38qoPTz75JG9hTewkOB45ciT/xZ6/fPnll/Inqo0bkZeXB6PRyEePP/nkE/j6yp+cyN/77cBZfivmyDWQ3aNHNxo1ZpRR8rrQtA3o3r07HVZCCCHkX2ApER/f1hf9wtyx4kAW7l78s3XbV0/OwpiOrth4KhfvbJb/RrN5PqvffhojFfGASoOlp1X4v8Vf8G16jTyg5eOsw+JbY8AeShYTJFMN4pxj4BTQESdPnuQjyMROgmPSctiHjLhCCx8hRk05oFDB6CPXMYbFBPHcPjh6B1v3/88No+jXQQghhDQDFtR+Nqcfuvq74MvDhZjzyufWbX8svBVdfR2wZHsi1hzNtE4W+3LRAlzjnAoICnyd7oqHF30MjVIBF50aOrUSoZ6OuHt4RwhKNaS8BJgkBaShdwMaB7zxxhvIysqi310zoODYjiXmlsOg0EHKOWOtUlGXUmFJPYywAB8Y+t3OH4v5SZg69ZpWfb2EEEKIPXHVq7FiXn9EeDni+7MWTH3oZXmDxYRjS+6Hj7MGT/50AkfTiq05y0uefxg3++YBFjPWFPlj/otL+Lw+Fhw7aFR4cHRHBLnroQvuBjF2A0w6d6iG3YWq6mr83//9X+v+wHaCgmM79vNfp/gtq1vMuPVonFIRFtERCq9w/vjWDhae90QIIYSQ5uPlpMXX8wcgwFWHLeX+6H3NHHlDdSkKf3mVB2J1FSzqAuTXn7wPd3eoBIyV2FIdhllPv2vt7saC5P+bIs8fCu/aC5aE3bzFtLLPDdiwYQNPRyX/DgXHdmzbyXRIrEVjzlmeUlHj05Wvl0pzoC5JRZLZ3brvQ3fMbMVXSgghhNivQDc9D5Dd9Gok+AyHe9QAvr466yzc4tc2qmBR59kH5+LxXgqgqhj7EYlpj//POuluQjdfDO3ohTRVILpoSyHmJUDZbQIUEYN4J2FW7Yv8cxQc2yk2AzapQgmpKA0wVfMqFZbaX7clcTeumToVuWHj+WMpcTcCAgJa+RUTQggh9quDtxO+urM/tEo5T1jwDOPr0/78FWHlsY0qWNT5z9xZeHGYC6SyXJzUdsOER97mBQrY6PILU7vyyhjaAbPgn/Q7pMpCKAfdjhyTDv/73/9a8Se1fRQc26mT6YUwK+vzjV17yF13GDFpL/w69bI+fuyGEa3yGgkhhJD2pHugKz6f2w8s/vW4/jkIrvLA1NlfFiPQktOogkWduTOn4+3JgZCKM5Dk1B2jHn4XVVVV6OTrjDmDwnAyuwJznnoD2gNfAZII1cj78ek3qxEbG9tKP6Xto+DYTq3aebw+31ihQpVPD/lxyiF0CvbD6p1Hrfved5M8gkwIIYSQltUvzIOXeasxS/C+aSHg5MXXJ698CZ5CRaMKFnVuvGYCls3sAhSmINO1O4Y/ugSlZWV4eGwneDpq8PnhInz4xkJI+76G4OAG5cj78PiTT8FisdCv8x+g4NhO7T6TDUk0Q0o/AsGng3U9S9y/5ZZbUOwkT8RTFyZBo5IrWBBCCCGk5Y2M9MG7M3uh1AR43/QSoHflFSyyV70IR4WJV7A4UlvBos7E0cOxYm4MhPwEFLh3w/AnPoWxqhxPTozkOcsHqzzx5oMzYTnxO59sH+fYC19++RX9Ov8BCo7tkNkiIsOoh5SbwGe6ukQNtm5TFSRi6LBhUPhF8sezR8qT9AghhBBy9VwTHYDXruuBUosaHtc/z2sVswoWJWtfh2gx4+4Vh5FZW8GizvDBA7D6wVFQ5p5GqUcURjz9FQYFaBAd5IrP9yQjZvQ1uGdwIMT0Y1B2GIQXVu5CTk4O/VqvEAXHdujP0+kQVTpIWbGASofqEDk4Nh9dg8mTJ+G9L1db971n6tBWfKWEEEJI+3Vz/xA8PSkKFUoXuEx9ClBp+UT66u3L+GjwXedVsGBievfE2ievgTr7BKo8OmPcwlWYH+MFsyjhpd/i8PTTT2GkLhViSRYv73bDg8+12s9nqyg4tkOrd53gt2LmCQiB9e2gxcQ/ceutt+KPoynWdT4u9X3aCSGEEHJ13TOiAx4Y1QE1jv7Qj3+EzxOS0o7AcmxNkxUsmG5dorDhhVnQZR2BwaMDHv9iG4aEOWPX2Xxsjy/AsiWLIez+GDBWIyNkPD769hf6tV4BCo7t0IGUEkhmI6SidDj0vZ6vE4syEObrhsGDB0PZYzJf18FUHyQTQgghpHU8MT4Stw8MhcWrI9Qj7uHto1nusJRykFeweHtz/AVf0yEiHFsX3QmnzAMwuYfirzMZfD0bPRZUGhzYth7mnR/z0ejX9xQhI6+oFX4y20TBsZ2pMZmRJzlDyk8EVHqYvOXcYvYJlE3E27v/AASW1wTgv7eMa+VXSwghhBBWt3jhtd0wrVcAhOBeUA5mXfQEmP78AihKxYfbky6oYMEEBgZi29sPwD1zLySdK1+XVlSF5StXw9fdCb9+8iYsh1ZBcPXD4Mc/5XOSyN+j4NjO/PHnMUCphphxHIJvR+t6Zf5ZzJgxA//3zifWdWN6d2qlV0kIIYSQhhQKAf+7qSfGRPlA2WEwlP1m8AoWxq0fANUlWLD6+AUVLBgfHx9s/+AJ+J773brurVPOOLfre3SL7o1b+gVaW0xPeHIpHfTLQMGxnVmzeTu/lTJPwKmT3J5SqinH+DGj+D+gRLOnvK6igHfWIYQQQkjboFYq8OGtfdA/zB3KLmOh7DmVV7Awbf0AJpMB877Yf0EFC8bNzQ3rb1Ai5PQK67p7twMGswUvvPQqLPu/g5iXiCR1OF74Yv1V/qlsDwXHduZYiR6SsRpScSYMYXIlCjHlIKZNm4aamhooI+VOeddGUG1jQgghpK3RqZVYfkc/dPVzhLLntVB0GcMrWJh2L0dxtQV3fLYXlYbGFSwYfXAPrB+fhw7xcm3jsxZ/LFu9AWwq39HDB2HesRRSZRG+jDPix+2HWuEnsx0UHNuRsqoalDoEQ8qOg+BfX79Yyj+HmJgYfPXNd9Z1z9xxbSu9SkIIIYRcirNOjW/uGoxgVzVU/WZB0WEwr2DBSrImFFTjgRV7IR5YDqxfABxcznsaaPvcAlXvm/HLbQFQikb+PEsPleL3X9fA19cXLz71OMzbP+Qtph//JR4nE9Pol3ARFBzbkdXbjwBKFW8Z7dhlhHW9J8rg7++P1z6rr2/s7+HSSq+SEEIIIX/Hw1GD1fcPh6cOUA6aAyGkN8STv8OSvB87ksrw2uYUID8eOPEDcHwllDon6AbfDafr3sVbM/vy5xAcPfDERz/j9OnTmDdvHsLdVLD89SVvMX3j4s0oLCmjX0QTKDi2I+sPJ/JbVsJNDO0v368uQ0wXuVW01E0u4eZUIu9HCCGEkLbLz1WHnx4cCUeVCNWwuyD4d4Hlr68g5iXgs/KB+KFmAG/2xYPkBhP7ru0VBE9Hjfw4ZgY+Wv4VlEolPv74Y4gph3iZOIOjHya/8A3M5gtTNNo7Co7tSFyBGZKhEtC7wAw5p1gqOIe+ffogOTkZCu8Ivm7B5PrGIIQQQghpu8K8HPHjgyOhFiSoRt4PwT0Y5i3v8Yn1TyV1w4ESF6C2bGsdlVKBx8Z35vcFpRprkyWUlJSge/fuuOeee2A59itvMZ2rD8UtLy1vpZ+s7aLg2E5kF5SgSucNKecMFCHy5RRGzE9Cnz598M4H9eVbbpk0vJVeJSGEEEKuVNcAV6yYNxAKhRKqMf+B4OQN85Z3IYoibku/FunBUy/4mhv6BMHLSR49RqfheH/FT/zuE088gYAAf5j3LOctpvfXBOD5j3+gX0oDFBzbiW837YegUEDMPwdtRH1wLBSmIjo6Gr8cy7GuU6vVrfQqCSGEEPJPDI70xzvXRUJQaaEa9ygcHR0h7f0KRpUjpr67HRU1pguqXswfJl8xZr6JM/AUCvZ1r732GmCqgXnbEsBUja/OKrBqw076xdSi4NhObDqRKt8RADNU/K4kWtDJU83zjJQxN/J1Xc2Ub0wIIYTYousGd8Fjg9wArTMMg+5GBzcFT5EoEZww5eVVEEVWuK3erQNC4KyVYwKjWyje+mErvz9+/HhMnDgRqMiHedcy3mL6v+vO4WjsmVb5udoaCo7txJkiEVJVCT/B60jFGejXuyd27doFQSH/43h59thWfJWEEEII+Tf+c/0IXB9UBcHJC+eCJ2N4gMArWKRa3HDzovqSrXUl4WYPDrU+/uJoKaqM8gS8V155RR59zj7NW0xLzj6Y9f5G5BcUtvtfEAXHts5Yibhf34PCIxhidhyCutSnVLDJeCzf+PX361tG9+lRX/+YEEIIIbbn3f/MRG+cg+AWgL/UvXFtkJGnVe4vd8Nj769stO/cIeHQKOWOuCwFY9Gaw/x+QEAAnnzySX5fPLMNQvI+GDw6YvLTy2A0ynWS2ysKjm3d8ZX4dvMBfldKPwZR79ao+QcLjs9q5RmrUmkWBIFaRhNCCCG2btXL98K/8Cgk92BsrgzBdd55kCoK8VOaFu+++iwfPGO8nLSY1T/E+nXfHslDelEVvz937lz06NGD3+9QcghCYTLyPXtixn/fhiQ1TtFoT64oOE5PT2+5V0L+mfwz2FUpn/RS3lkUWBysmxyq8/jkO2XHwfzxzM40EY8QQgixB+zv+7pF90Kfvp/nE+8scceE7M95B7z3crti5dJXrPveNSwCCgGQJBEiFFj460m+XqVS4a233oJCoUDcyeO4ObiCt5g+quqCp976GO3VFQXHUVFReP7551FVJX/iIK1P9OyMNG0EpPI8QKGCyGbksUC5phy9Ogbit99+t+774A2jW/GVEkIIIaQ5eXl5YfUzMyGk7EeJWyTO+oxGv5yfeQe8p+KCsW3Xn3y/YA8HXNszAIKg4PWRt8QXYHdCPt/GKlqxEWRmx4a1mB9pYVE0Vma4YMXqX9vlL+yKguPNmzdj48aN6NSpE7788suWe1Xksu0RoyG4BULMOoXArjGNUipi+vbFG1/Vn9ghIfWXVQghhBBi+7p364bFtw2AmHIQKa59IXh3QFjKGsAjFPM+3YWjx47x/e4b2ZHfSjUVEEQTFq6Lg8ki8nUs99jPzw8pKSnQVOZipD6dB9jPb8rA3gOH0N5cUXA8ePBg7N+/H4sWLcJzzz2Hvn37Yvfu3S336sjfWrVHLrsi5SXBNXKQdb1YcA7dunWD0OcG/ti1uL61JCGEEELsx8SJk3BnNxXEzFgc1A1Ar5494V6WCCkwGjNe/hpnz55FpJ8zRnTygMIzBObUo0jMq8CKvXIZWGdnZ169glm6dCmevnU8QipOA55huP39P5CWlob25B9NyJs9ezbi4+MxZcoUTJo0CTfeeCNvT0yuvv2pJfxWzImHxcm30cgxm22qcA/ijx8e3YF+PYQQQogdEiUJ/3n0CQxVJkAszcaaAj8smT8GusocmDqNxs1PvcMn2P1nbCQgyKGfzlKF9zafRUGFgT9m8dy4ceNgMpnw9NNP47dF98CxJAnmoD6Y/t8PUFZWhvbiX1WrYEWk58+fj19++QVdu3blw/IVFRXN9+rIJdXU1CBf8IBUmg1Ul6BY0vP1kigiyMGCjdt2WfedNnYoHU1CCCHEDrFueHqNGv97639wL0sC1HoseOVdfHP/aKCqGAUdJmLNnuPoG+qBHr46KEL7oOrobyg3mPHWBvnKMqtm9eqrr0Kv1/MsgfW//45fn7oeiopcFAYPx4yHX+Qd9tqDKwqOP/74Y8ybN48nb7u6umLMmDE8reLee+/F4sWLcejQIR4ks1vS8tbvPAA4evB8Y0gWFNfIZVekkkz0je6G3+LLrJ3yPD096VdCCCGE2Cm9Rgl/L1e889gd/HG66IYvlvwP0aV7AVHEs3+kIbOkGo9N7sEbg0kOHghUVWDV4XQcT5evQgcFBWHBggX8/sKFC+HmpMPSm3tCMFfjtPsgPPx/r7WLEm9XFByzTxSlpaU8rWL79u0oKSnB4cOH8eGHH+Luu+/Gtm3beKB8xx3yL4Y0E1ar8OByYP0C+ba2duGavXH8Vkw/Clf/kEYpFT4+PlD1uZ4/jhapZTQhhBBi79jo7+g+neHtoIAiuBe/sp8Zuw/mP5ejSlRi/leH0C/MA4GOgKLTMBTt/hYKQcCL605ZW0/Pnz+fD3SyeO+ll17CxCF98XCMM+/A+1tpID785HPYuyuuc/zjjz/iiSeewNChQ/nQ+/nYyPLp06eb8zWS4yuBEz8A+fHyLXsM4Fi2XFJPyo6DT0Q363GS8pNQUlpa/zsZLRf4JoQQQoj9B8jX9gmF4OgBwSMY2dnZkNKOQjz2K05nl+GxH47hcTZ6rNaiwsEfg71MOJpWgl+OZjaqfSwIAo/59uzZg0dvmYQxHqUQXP3w5p4CbNq8Gfas2TvksRFLNoJMmlH+GUClAzzC5dv8eBQVFaFU5wexMBWoLoXW09+6u7IkDX+dybA+Hju8vooFIYQQQuzbuK7yBP2eU2Zb15lP/IYYLwmb4nIRn1sJV6URiqgxyN35LTwcNXh9wxmU15j4vr1797ZmAfz3v/+FwWDAZ0/eijAxC4rA7rj3ow2Ii5OvXv/dVW5b1OzBMfukMWLEiOZ+2vbNOwow1wBFyfKtdyRWbdwNQe8KKfcsT7w3ukXwXSVDJTr5uyHbpav8tUWpcHJyat3XTwghhJCrJibUHW4Oakj+PTBgwADreodTv6BnsBuW7TqHcD9PCFpHxFXoMLOLHvnlBizZVp+GyYJiX19fXo1syZIlvIvebwtnw6kmD2Ln0bhxwVtISMlAWY0J1UYLxGPfN3mV2xY1e3BMWkDPWUD0TDlIZrc9Z2HDoQS+SeTBsQOyVIHWfGPRYoEyvD9/PC5I7phHCCGEkPZBpVRgdJQPTueUY+H/lljXb9n4B59g5++qw7HMcr5O2XU80nf/hO6BLvj8z2Qk5ctVx1xcXHjOMfP+++8jMTERjjoNvvnPBCgMZajsOg1zH3kGVVXVMJgtMGWfhnjeVW5bRcGxLdA4Av3mAZPf5LeS2gFnM/LkbWV5EBzcYbTIifRifhIv8VbnllG9WutVE0IIIaSVjO/qx28P5xjx+++/W9c/NH82Pp0dA71ayR8Lju5YeyIbDw0NhMki4aV1cdaKFNdccw1Gjx7Nax+zkWQ2Qhzm54X/3dANbOgtPWQCvv95Ld/X4tUZorG60VVuW0XBsQ1iyfXlZvmklqpL4BvZ27pNKjiHgqJi6+OhA+tbShNCCCGkfRje2QtalQKb43J5DnGdffv24fRfm/DmjfWT9YXu1+Do1l9xTbQ/dp7Nx7bYNJ43LPzxJF6bGQ2dVou9e/fixx9X8/1H94nCYF0GbzG9MV4uA2fsNgPGrjc2usptqyg4tkGsjSM7ISWzATBWwSOsvlKFm6UElTpv62O1Wt1Kr5IQQgghrcVBo8LQjl7Yn1yE0ioT3n77bes2VsvYpSIDD4zqYB09/nrrETwwIhwOGiVeWnMMhmM/8tSIkLwteGLGEL7f66++guKiokYj01llJutVbjGm/io3v+ptoyg4tkGZmZmAgxtQJX9aU3nJNY6lymLE9OwORQe5OoWzqX4EmRBCCCHtr2qFRZSwLT6Xp0fUMRqNeOj+e3BtJwf0DHTh6yq6Tse+3Tswe1AoUitV+LS0H0R3OX/4riG+6NKlC0pKivH2m4v4/j17dINUVYIyUQuLxQKFQrCmatg6Co5tNDjmI8e1wXGm6MpvxcyT6N6dBcfyJ7zxngV2UVKFEEIIIVduTBdfCAJ4agWrPMFihDr5eXl4/713ML2PPKFfUOvw0hEFtp7OQ6ijCUsLeyMzJxeSqQZq/6548803+X4//7gaJ48cRKcOHSBUFkBy8kJuZjqctSooFfZRBICCYxuUlpkFQevE843ZWV9uEqz5xjl5+RAU8ie3cfo4uyipQgghhJAr5+2sRZ8Qd+yMz+cVJcaMGcPXa7VaHjTHxZ5EpK88csxIChXSCivx+KQeqJI0eKNsHKToGTx/uG/fvrj99tv5fs8+8zQ0SsBdZebl4I7FnualfO0FBcc2KCWnNl2iqgQaD/kTX10Zt/jEc9bHoQ5GuyipQgghhJB/nlpRabTgr6RCjB07lq9jTT2Y5HNJ6BvsCietyrq/wSIhOtwPIzp747cCPxz2ud6aP/zMM8/A29ubl3X78MMPEeyu4+v3nUqyq18PBcc2KLNYTpFgaRX+Xfpa13fwccKRU3L9YyZQKLCLkiqEEEII+Xfd8lhqRa9eveDu7t4o9zgrK4PXOG4ou6AEL17bDWqlgBd+PcXzlhlXV1csXLiQ31+8eDG8tCK/H5eWb1e/HgqObQyrPZhXbpQfVJXAKVjuhCdVl6Fvnz4QdM7WfV16X2cXJVUIIYQQcoVq2zl3OLgQEU5mbD6VA0FQYNSoUY12i4+PR49Aee5SnY9/3Y1wL0fMGxqBuOwyrDyYZt02bdo03gmZBda71v/M16UVVVtrI9sDCo5tTHl5OQwKPb/Pco5zdHKlCjHpT/Tp0wfQ1wfHDRuH2HJJFUIIIYRcITbXqLad8zj1MeRXGHE8o8Sad9wwOO5eGxz7a838dmeRM0RRxIOjO8LHWYv/bYxHSZU8MMdyixctWsTzlqvyUvm6apUT8vJqm5PZAQqObbJShas1raLcIucJiTnxvMyKoHfjj1WCfKmDEEIIIe1Q/hl5zpFHOMa7ZFhTK0aOHAmFQtHkyHGvDv7W9Ws27eS5yE9PjkJxlQnvbj5r3RYWFoZHH30UMNXwK9eCiw9iY2NhLyg4tqFLI6wsW8bWT+Uax4yhvjybtiKb5xEJHvJIcoRbfXI9IYQQQtoZllbJ5hwVJaO3OhVeWgsPjlmswK80NwiOwzwdeSBcWGWxrn/792Mwm80YFqxDD39HfL0vFWdyyqzb77vvPnTu3BlSeR4EZwqOSSteGsk8sgmCzgVSTTmc/MOsu/Tq2hlFRUUQvML54+jzkusJIYQQ0o6wuUZszpF3FBQ9Z2Js9yAk5FUguaDSWrWCYZUnRNGCrgEuPL/Y31nurJuhDcWyL75BdHQ0yncsB8sofnHtKWtuMevAy2of8+BY54zlX38Pe0Ejx7Yg5yRQVQRU5CIzrxiCgztPqfCLrK9Uwfqmnzt3DgqPYP54QKf6SyOEEEIIaWfYXKMGc4/GdZdLv26Oy2mUd2wymZCamspTKyoMZmuvBEGlxVd75Zxic34yZvULxr5zRVh/Msf6tf3794eLIJeFK6yRn8seUHBsC2rKgPJsoDwHGfklgKMHUF0CbYBcnk3MS+TFuY8dO2b9ki5Bnq34ggkhhBDSlgzp6AUHjZKnVnTt2pV3zGsq77i4Sg5wJUMlcl27ACotKioq8MT4SDjrVHj19zhUG+vTLx696zZ+y/KO33rrLdgDCo5tgc4FcPbnS5rRFYJKI/cz13jxzVJRGiIiInDkyBHrlwS5yxUtCCGEEEJ0aiWGd/LG4dRiFFUaG40eN6xYUW2SA1+HwjO8+52i03BeKcvTSYvHxnVGVmkNPtpZ3/SjX1RtiqezD5YsWYKUlBSbP9gUHNsCvx6Agwfg5INMsXYyXlUJcg21k+4MlXBzc8PJkyetX+Kql3OGCCGEEELqGoKwfh5bz+Q1yjtmwXGElyMcWU/oWmOjQ/lAnLLbOFQbTHxy3m0DQ9HZ1wkf70xCelEV349N5mPYpDzm6aeftvmaxxQc21BSvdmjM/IFOV1CMtc2AmH3jdW8aw1vlF7LnnqcE0IIIeTfGx3lA6VCwKZTuRg6dKh1/YYNG6BQCHxSXp3O0TGQzmzl85wUEQNRVVUFtVKBF6Z2g9Es4tXfT/P9XB3U0Alma3C8c+dOrF27FraMgmMbSqrP7fMoBGc5R8jZUz4JGaVohE6nA7QNGoAQQgghhDTg7qhBTKg79iTmQ6nR83rFDOt2x0aG61Ir6ibYTe/hxXOPld0moKS0zJq7PKm7HzacysGehAK+zs9JxXOO6zz//PMoK6sv+2ZrKDi2sQYgdTWO3XwCrOsdNQo+UizUVqoghBBCCGnK+G5+qDGJ2J2Qjzlz5ljXs1zhhm2kM4qrcPfc2RDjd0Bw9cOqv+Kt2/47MRJalQIvrI3lnfMifFwg6F355D0WcOfn5+O1116z2V8ABcc21x1PDo61LvXVKFiOEGvzKHjKnwA9da32EgkhhBDSho3vKl+BZlUrxo0bZ11/+PDh84LjanTv3h2WM1t5KucX+zJ5LnGV0Qw3Bw3uGByGpPxKfLs/DZ0Ca2MSJx/Mnz+f312xYgV/TltEwbHNBcfukEQzoKtPoXDRq/mnNIWXHBz3DKo/uQkhhBBC6gR7OCDKz5lPygsNkxuHMcuXL0eEtxMv98ZjjuJqeWJdTTnExD2o1Hpi++kcnm/M3DEkDF5OGny3Pw0hHrWT8ly8oXdwxMyZM/njBQsW2GTtYwqObUhGRgagdwOqS6FQ1w8PuzlqeaWKuu54A6kBCCGEEEIuUbWClXNjZd3qxMbG8sl6Xf3lSXnlBjPKqs28j4Ll1CZIogWv/3oYKqUcOurVSmhUCjhqlQj2kMvHskl5p+NO4bnnnuNtqs+cOYPvv7e9znkUHNuQAwcO8LQKVlpFo60Pjt2ddLwBCM/3YSPHwbXl3gghhBBCmgiO67rl1aVBMGykuHuD1Ir04ip4eHgAlYUQkw/gbKmAxLwKOGhUYIPKOaU1fCQ63MuJ788m5cXFxfGvmT59Ol/Huu/ZGgqObciZ+LNyQ5CqEogNagh6Ojs2agDS2ZeqVhBCCCGkaSy32M9Fx/OOG07KO378eKPgOLOkGk5OcuArxm7gt2+sPcJHjIurjLxmckcfJ954jBcHcPbhI9AsyC4qKuL7N+zEZysoOLYlOhcICgUfObawM5J9yrOY4enm0ig4dnOgBiCEEEIIaRqrcMVGj1MKq2Bx9Lau/+yzzy6YlOfsLA+4SaVZENOPYW96NRLzypFaWN8EhD0fy1dmI8eshBtLA83NzeXb/fz8bO7XQMGxjWAnW12lCqWxgn9a40zVcHNzbVRPkBqAEEIIIeSyUitO51nX/fzzz+jgLU+uqyvnVjdyzHOPT/7B77+/+QxSCyv5/VAPB34b5uXEiwZAqcGpU6eQk5PD19PIMWkxWVlZ1hrHGkuVdeQYdd3xGlSvIIQQQgi5lIERnnDWqrApLhfDhw+3ri8vK7VWvUovqk+riIyMRFdfPcScePx2Mhd7zxXy9aFecjAd5ikHyYKzN0+tqBs5puCYXJUax6xahTWtgo8cu0Fw8qKjTwghhJDLwvKGR0b54Hh6CfoOHW1dv2PHDvSqndi/P7nQmlZRUVGBO++8E5bYPyBCwMZTudAoFTx3mQn1rB1xdvbB3r17UV1dzR9ScExaNjhmZdyY6tL6CXkmeeRYcPXnDxWon6hHCCGEEPJ3qRXlznKfBGbLli3WSXnlNWbryDELjlkFCufKDIhFaXydk07Fy781Gjl28ebBMePi4gK9Xi7zZkso59imWke78/tCTSnMDdIqFAqFtTteZy9ta75MQgghhNiIkZHeUCsFnK2qLw+7efNmdPOvT9VU6OTguLy8nAe6t916K8TYjXwdq5Vcp27kmFWsqGOLk/EYCo5tKq3CFZKpBjDVoNJgtqZVFBQUWBuA9A1xBQ4uB9YvkG+NcsI8IYQQQkhDLjo1zz0+klEJrZOrdYS4PDPBuo9B6WBdz8yePRtSYbJ1e0VtPMK65TlqlNB7hwCmKsBQBl91pU3GIRQc29rIcVVxbW9zi7zBWI3k5GRr6+hoJAAnfgDy4+Xb4ytb94UTQgghpE2nVhgtIoJixlnX7di+zXq/QtI2Co5DQkLQd/h46/aVB9KslbLCvByhcPLgKZ8QzfCTcm0yDqHg2KZyjl15jWPe67yWwmLA6dOnrY99jGmASgd4hMu3LEgmhBBCCGnC2C61TToCoxvlHXfykdMpEkphTauo02f4BOv9T3YlwWC2WGse16hcAJUGUKjg6+Zok3EIBcc2wGKxIDuvAILW8YLgWKcEbx1dx9MnADDXAEXJ8q13ZCu9akIIIYS0dQFuet74I1/tCwhKvo7VKR4WLucd/5laaR05ros/NB4B/NZydhfyyo1YczSTPw5lk/IEgVes4CPHzgqbjEMoOLYBeXl5sCjlZHnJWNkoONargHPnzlkfe/a6BoieCXhHybc9Z7XKayaEEEKI7aRW1IgKCL6drOt0hXLecU65POnObDbDaJTvpxVVQ4AEy9GfoTDX4KMdSTCZRYS5qfh2wcUfrHiWW+QQm4xDKDi2lXxjUU54FxTqRsExK+DdkKe7G9BvHjD5TflWU9/phhBCCCHkYiXdFMG9rOtO7918wX51qRUphZXwd9XBQSnBGLeVt6FecywTfnl75FiFl5eVUFJpssk4hIJjG1BVVQVYasulKM8LjvXqRvvq1Ep5ZujlVqy4kn0JIYQQYnei/JwR7K6HMrS3dd2eXTut9x0b1DpmMUhaURUivJ1x7fTrYYnbAoVkxjf7UhFqkPOLBdcAnl5RkFl/ZduWUHBsAxwdHQGLPHIMpapRcOzqoGm0L9/GZoZebsWKK9mXEEIIIXaHVZoY19UPcPCA4B7M11VW1g+W6X3CrCPH+RUGXjErxNMBt86eA5gNsLCUC5MJXv5h0MIEwcWHBSTIM9bXT7YlFBzbAAcHB0ASIYlmCCpNox54Hk76C4Pj/DOXX7HiSvYlhBBCiF2nVggNUiscIF+1FkL6WIPjtMIqa0e8YH9fKIN7QlDrMCLCBZbomfDXmeW0CrUeZ0sap37ailYPjnft2oWpU6ciICCAf3JZs2bNBcHe888/D39/f96ZZezYsUhIqC9O3R7UtW6E2cTTKiDU/9o8XeTi3HVMrHMem4x3uRUrrmRfQgghhNilfmHu0CtEKELqg2NF1kl+a/DuYh1NZvnFTIiHIzZt3ABFWD8+gDd3ZFe4uLjB09UJcPIGtM6IOxMPURRha1o9OGYHumfPnvjwww+b3P7mm2/i/fffx8cff4z9+/fzFIMJEyagpqYG7S44tph43UAJch9zxoudhA0YzKI8M/RyK1Zcyb6EEEIIsUsqpQK9/TRQeITAKzQSKpUKxXG7+bYavVeDkWM53SLMywFr1m+EENQDIToDfF3lK9lOUjUEhQJw9OJzplJSUmBrWn28e9KkSXxpChs1fu+99/B///d/mDZtGl+3YsUK+Pr68hHmWbNmtZ+cY0Y08cLabHC47lONt7vc7rFORY0ZLqzoNqtUcTk0V7AvIYQQQuzW1D4h+CvrHEocgtCzpxMOx6c22s6C4xSDPHLsIhhxINMIRYgaMweGWPdRG0pYSzIILt6QynMRGxuLiIgI2JJWD44vhbVFzsnJ4akUdVxdXTFgwADs3bv3osGxwWDgS52ysjJ+y4b2bXF4n316U6vV1rSKhiPHLi5yke46+eU18HORWz02N3bs2AcWWzyGtoKOMR1fW0fnMB1fW9eez+EpfSLw9K/xQFA0wlzScPhofZMx6N14cJyKSvg4a7Fj6yYgNIaXmr11eDfr8ZLK8uTg2NkXEmJ5cHzNNde0iWN8ud+vTQfHLDBm2EhxQ+xx3bamLFq0CAsXLrxgfX5+vs2mY2g0GhhYWoVSDYskQd3gF+3s7Iy6jwJJGXnwVdd/MGhO7HuVlpbyE1rBLpkQOsY2hs5hOsa2js5hOsYtzbEiAxW+kXDQVkFkg3J1PIJ57JWsqkC4hx4r16yHEDYTIepS1JQXo6a2u3RJRgIQ2B1uQZ1QELcZhw4d4s3M2sJ53LAFts0Gx//U008/jccee6zRyHFwcDC8vb3h4uICW+Tm5oZci4m3kFYo639tIaFhCAoKQqLZAEGlhVmlh4+PT4u8BnYys0mT7DhScNwy6Bi3LDq+LY+OMR1fW9fez+FIJwOOKJS8nXRoaChyauMLpVc476JXVmNBqIcevxYrIUQoMDu8DD6H366duzQD1UXZkPxMcAvuhCKFgnfxPT8uaa1jrNPpbD849vPz47e5ubm8WkUd9rhXr/rZlOfTarV8OR/7BdjqiS7XOm6iWoWHO/x9vZFYXcp7mRec+ROKnte2WEcadjLb8nG0BXSM6fjaOjqH6fjauvZ6DpssIvqHOONwhogThcCI4SOxKv0klGExEHw7I7+6BNAChsJMSCExvHX0bdI6KArUQM5xsKzPvNxcSMH5MAUG8ePIRo0LCgouCJBb4xhf7vdq07/18PBwHiBv3bq10Sgwq1oxaNAgtCe8YkUTwTFLqfBXlkCqYgnwQG7qWWrkQQghhJArYhElVBrM6BHZAVJhKgrVPhg2YiSkUjmNVeEXicLarM34s/FQeEegszkJOo3a2itBzD0jp1CU56GgWkREx058/1OnTtnUb6PVg2PWivDYsWN8qZuEx+6npaXxTxWPPPIIXnnlFaxduxYnT57E7NmzeU3k6dOnoz252MhxZn4xvNXVQI086TDX4kSNPAghhBByRdhUfwEC3PzDILgFQKwoQHhEBKTKQus+eZJcISu5VJ7YdnsPXaNeCUXaQJjNZkjlebCIQHj3vjYZHLd6WgVL1B41apT1cV2u8Jw5c/Dll1/iySef5LWQ7777bpSUlGDo0KHYsGHDZeeN2FNwLFWaoFAoeQvpOmfScuDsHQzpjDxynG/Syo08jJXyCDLrgMfygFj94hZKtSCEEEKIbVMoBDjrVFh9qoznGJsO/YhFL+8AGgTHhY61Jds8QqEwVWHGzXcApxzlQTnvSOSqo621jqtZAYWIbvwxq1hhS1o9OB45cqTc8vgi2OjxSy+9xJf2jKdVlMmzRtlJW+dcViH8u44Gju/jjwtVvkDPyXJgfOIHuSV09nF5Z6pnTAghhJCLiMsuw5pjWXCqyEBRVix2ZbH8TW/rdotK7sorOLihq7YIGgeXRrFFTm0arKdWRAabAOcdbJPBcaunVZArTKtgGgTHaQVlcHb3gmRkn9GAEslRHiFmI8YsMK7NA+Kf6gghhBBCmsAGKl/5PQ6CAAx2qC+X2zUsoMnjNWdU9wvWsYIJTKCLXHDWrPOwpsyyLACOXdk+9EX9wh63MRQc22JwrK4PjrNLquTydLUFBksNtaPwLJWiQR4QT7UghBBCCGnC5rhc7DtXhJkxwYj0dbKuX/DYI/ByaJxooKwpxfXDel7wHHU9KIK9nKFWCsitEnlvChZ4nzlzRt6JXdk+uQoozZBv2eM2hoLjlsA+BR1cDqxfIN82w6eiRsGxoq4FCFCQnQbntK2QDHJwbE1QYTnG0TNr6w7OlB8TQgghhJwftphFLPrjDBw0Sjw2vjN27Nhh3TZ+/HgEuDWe59XduQrKI19cEOfUjRz7+/oi2MMBKYWV6N5dHmFmRRW4uivbzr5t9sp2q+cc26UWyPdlOcdSXXBcGwgzpRYNXDLWAwZz4y9gqRWUY0wIIYSQv/Ht/lQkF1Ti8XGdobHU4PDhw9ZtBoMBIZ6OOJFVYV03J7QQOLH1gjinbuSYleENMzlid0I+xnbtxkvyWitWsEG77BNAeW6bvbJNwXFLaJjvy9IamuFTkVzn2Cg/sJjhgiqUwQGVKje4sG2mi7fTJoQQQghpSmmVCYu3JsDPRYf5wyKwZPE7jbYnJibCz1WeiFdnomcuzCU6qM6Lc+pGjl3dveBv0MFkkeAeGsWvap+oGzlmV7LZClYPuVOfNnllm4LjlsA/FR1v1nxfOa2idnRYqYaf1ogygwOMGlc4q8yAJNccJIQQQgi5XEu2J6CkyoR3ZnSFsboCn332WaPtJ0+dhpdf78Zf5BMJ5J+AWJTMu+TVxTk5OblgBcg8vH0QXKnn6yokLQ+GU1NTUW20QM+ubMfMlYNj1jWvDXYhpOC4JdR9Cqqt+9ccn4oaV6vQwNOFNfsAJAd3CD1nQFJ89K+/ByGEEELaj9TCSnz5Vwp6BLpieq9AvPPO2ygvL0eXLl3QoUMHrPvtN5xNTESn4PrgONqxHMWdboA7AH1JIuAj91JgzT/y8/P5PqxVtFO2HLN88/Nv/Hb0mHEwmC3QqtpeMHw+Co5bQgvk+/Lg2FybVqFUwVlhtP76znpPBCwfNOv3I4QQQoh9e2PDGZ768OyULqioKLeOGo8dOxZLly7lI769e/fBkh2nrTUcbo52hUXpAClmHgRtfRhZmJsLURShUCph0TjhvS0HoIQF+bF/wcvbG/995v94Bz5WKu4S7S3ahLYfvpMLJuQJSjXMxawytyw+Lac+5YIlz5stdNQIIYQQclEHU4qw/mQOxnf1xcAIT3z66acoKytDREQENm/eDIvFgmumToVPZAzOltWHi5KxBp5OGjg2CIwZNhmPBb5evn5Y8ONJFFQYYdi1HFJZLha+8hrc3NzhoFXy5m5tHQXHNqJxzrEGZZmJ1m3JOUX1KRcAymvOq1xBCCGEEFJLFFnDj9NQKQQ8NSmKB8UsOGbUajWvSezh4YFXX30Fsz6VO/AylqS9fC7V+YFxw8l4ir434URmKZwz98OSchDTpl+HayZPgoteBbXSNsJO23iVpDY4rkurUCMr/nh9l7z8MkCsD4jLqusDZUIIIYSQhtadyMLx9BLcPigUEd5OWL58OQ+Q66pTMK+++irmfl/buKOWlJeIqrpOd+dhI8eKTsNR7NENQapyFG5bDh9vbyx67RVoVAqbGDGuQ8GxjWg0IU+pQnpqsnVbTll1o7QKGjkmhBBCSFNqTBa8uSEeLjoVHh7TiQfFy5Yts263WCyYOHEiXCIH4lSWHDDXkcrzUFFRX++4oWMZZVD2vxlqYylSv3+BJxa//vrrcHdnU/dsCwXHNsLBwaF+dFip4Sevu6WEPyyoNMPX18e6LwXHhBBCCGnK538mI7OkGv8Z0wluDhp8/vnn1lFjxtXVFS+98hrmfnnogq+VyvObDI5zy2qwvjQAEC2o2vAuLNXlmDZtGiZNmmSTvwQKjm0EuxzhoJFzfNQ6uRi3Q41cMiXP4oDw8HDrvjmlVa30KgkhhBDSVhVUGLB0exJCPR0we1AYL9vWcNSYefnllzFm6VHrY0mU+ygoIAJVRY0C6boiAPd+cxg1ghaWv76EVJIJT09PnpZhqyg4tiF6pVz7ROfhz2+VVYX8tlrtij59+kDMS+CP/zhyrhVfJSGEEELaonc3n0WFwYynJ0XxPOAvvvgCpaWl1u1jxoyBMbgfDGY5IO5iToJQ26RDLGPNFSRERUU1es4X18bhaFoJLLEbIKbKo80snYJN6LNVFBzbEA+9AKk0B2aXQMBUBak027qN5fSImbH8/t6Uxp/qCCGEENK+nc0tx/cH0tA/zAMTuvnx9IiPP/64dqsEZ7WIZ8e445lf5Fji9eu64UyZCpLJwB+LZbkYN24c5syZY33O7/an8ecc0sETlqM/83UdO3bElClTYMsoOLYhvQJ0EDNPwOTgBWidYMqs7VNem2ohphzk96tEZSu+SkIIIYS0Na+tPw1RAm/4wVI1v/zyS5SUyHOXYKrGixO8MfHMBP6QBcrIT4LkEQqpMIWvc1WZ8fZ770MQ5NDxcGoRXlgbi2APPTrm7bJ29mCVL2wdBcc2ZHCYDlLWKX5fEdgdFTn1FSuKqi1AuZyDzJgt8iURQgghhLRvu87mY0d8Pqb3CkDPYDdUVlbiww8/tG4f3tUP6Z1mWB/v77cLn2+TB+DEAjnWuHnaRKi0epTVmJBeVIV7vzkCpULAk4Pc8emH71m/lo0c2zoKjm3IkKEjIGbJlzuEgG4oqqitewyguKZxL8ZPNtYn0xNCCCGkfbKIEh811qoUWDBRzhdmo8Z1ucY6rRaP3TgESwv78Mff+P8Ik0sEzlQ5Qqoph+Dkzdf37hjEb41mEQ9+dwT55QYsmt4NixcusH6vwMBAm6pnfDEUHNuQwHH3IczLic8EVQRGA2q9dZvS2QtK1s/89Bb+eNnezFZ8pYQQQghpC1YfSseZnHLMHxaOQDc9HzVuWEni+dtHYUbm9fx+L1UKhgYIePDHZCjcAiDonKEM64tRkd7oHy5PsHtjQzyOZ5TinuERSN75I+Li4qzP5efnB3tAwbEt0Thi8PjpEHMTIDh5AU7eEPPkTjYJJRJiYmJgiZOD41IT5R0TQggh7RmrTPG/TWfh5aTBfSPldIevvvrKun3AgAHY7zzM+vjnjhtxMNOEHSn1JWG7BbjgzRujodco8euxLPx0JANDO3phaqiExYsX83369evHbyk4Jq1iyJAhkHLi+X2FX5T1fkIpMGrUKKBSLu/GZJdU02+JEEIIaaeW7UzitY0fGxcJJ60KVVVVeOWVV6zb731mEf4olJuIrQ9fjcz8Esx+fxuU3eXmHe46AV/c0Q/ezjqczi7HK7/H8Ql479zUHY8/9ihMJhPvpjdw4EC+v6+vL+wBjRzbmMGDB0PMlQNiwbez9X6JqJWD4wa+2XW6VV4jIYQQQlpXVkk1Ptl1Dp19nTAjRs4XXrRokXX7c8+/gHt+SuL3rw+pRrCnG+b8kI/yoEHWfb6eNxA+LjrkldXgvm8O8wl4y26LwfdffoaTJ0/ybnq33347vv32W75/aGgo7AEFx7bEWAnftN/Q0aEKYnE6FH6RkIrrc4s7RkbBy8sLluPr+ONP9ma14oslhBBCSGv538Z43szj2SldoVIq+KhxwzJra2q6WO+/efd0PLC2CGcKLVANnsvXdTMnoHuwB5+Ad9+3R5BXbsCbN/aEoiwL77zzDt/npptuwn333YeioiJER0dj5syZsAcUHNuS4yuBEz9gCCvplnESgqMHK3DcqFTLiBEjIGac4I9NksDbOhJCCCGk/TiRUYKfj2ZieGdvjOgsV5to2Jhj0Ve/IT6nnN//86nRfILelm07oB7zkHWfRyd257cvrjuFw6nFuHt4BCZ388EjjzzC0ylYB7yVK1fydtJ9+/bFqlWr4OLiAntAwbEtyT8DqHQY3DsKYnZtvWM2elwjn+Artsfy1AqpKN36JfvPFbXayyWEEELI1SVJEl75/TQUAvDsZHl0ODk5GfHxchpmh8iueGGr3GH34TGdsHP9L1i2bBmUg++A4BXB16vi/uCDbaz7HeuCxybgPTkhknfUO3FCHoBjo8Wsy96gQYPw/fff201gzFBwbEu8owBzDQb7mSDV1Tv2i4JDWSq//1eGAQOHDAMkCyRDJV+37rDc2YYQQggh9m9TXC4OJBdhZr8QRPo582CZTeavkzP4Mev9AY4FeOqpp6DsORXKiAHW9SMinHAyuwIv/HoKQe56fHBzb5xLSsRbb73V6HsNHz4c33zzDZycnGBPKDi2JV2uBVyC4CmUoYu3BmKRnHfs76TimyVBgeO5RvTs2RNiwm6+bvXxPP4PgxBCCCH2jeUHL1p/Go4aJR4b15mv++6776zbh935LEwWOSZYe0ck5s+fDzEkBsqe16KTjxMgiRDzz2Hi2NG49+vDUCiAT26PgbNWgUcflatT1Bk3bhwvC6fX1/dcsBcUHNuS02uBsgzA0au2lXQsBL0rdA4O1l2+2n6Sp1bU1T9mkgvkUWRCCCGE2K9v9qUipbAK94/qCG9nLXJzc7FgQW0HO50z9pvD+N2FUzrhgbvmolTjDfXgO+DnouXNPyAooM6JxXepDnwC3hs3RKNrgAs++eQTHD1a33l38uTJ+Oyzz6DVamGPKDi2wZxjeIRjSGcfayvpwtL6Yt2HsgwYPHwkpHy5PAuz7Uxeq7xcQgghhFwdJVVGLN6agABXHeYNDedXjR97rD6FQjNDrjDhqlfhjw+fR1JeOTSjH4Req8YXc/tj64k0SKIIr6j+OJpeiruGhWNar0AkJiY2qo08ffp0nnusVqvt9ldLwbEN5hyjKBmDwvSQcs9CkkTkFJUgQlvBdzELKhRpfAFDBaTKYr5u8yk58Z4QQggh9umDbYkorTbhyYlR0KmVWLduHbZv3863KfveZN1vYspS7Nq1A5ox/4GgdcCHt/XlAXWa2RmCQoFctR+GdPTEfydGwWKx8LziOqxU2wcffACVSk7ntFcUHNuSnrOA6Jk8SHYdeBtC3LWQClIAvy4YL/5p3e3rHad4yRYxW24Csj+lhLeQJIQQQoj9SSmoxIq9KYgOcsW1PQNQWFiIxx9/XN7o4gdlt/H87i2qPfj61y1QjXsCcPHFC9Gl6B/mgQ/W7gMUcsAb6KbDkpv78NrIw4bVt5YeO3Ys3n77bSiVStg7Co5ticYR6DcPmPwmv53cK8Cad5x4Nh4KSQ6Aj+SaMHTEKEj59XnHexLyW/GFE0IIIaSlvP7HGT7R7v+mdIVCIeDZZ59FZWUlzyHWTH+Z79PBVYGvl30I5cgHoPDvgnkeJ3Cz0xGk5RXjq8P16ZeLZ/WGu6OGV7FISZErXrFqFGzynYLN0GsH2sdPaacGDxoIMfMkv78jzxH9nOU0CougguTf9d/lHRsrgYPLgfUL5Fv2mBBCCCFtCivbtuFUDiZ280P/cA+sX78ea9eu5dtUI++z7pfx9ROQet8IZecRGOuUjGc8dmBNri+mvLsNJqVcceLO7hp08XfBe++9hxUrVli/ltVIFgQB7YV9J43YuQGzFgCv9eEJ9EafHujmpcV+OfUY645mQGcshdlYBUHjgG0nUyHqv4HCJ0pOz2Cj0JfRjY9PAMw+DpgNgFIL5OUBPj5Ar8t4DkIIIYS0GFFkDT/ioFYKeGpSFIqLi/mILyP4dIIiuBe/73Xkc+S4d4aqzw3o5mrEe90y8MD+Ydhg7Axo5FBwXKgaj980Eh8ufocHx3UOHjzYrgJjhkaObZizpx969eoNqSgVQkBXHPrjBzgIcg3Co/kW3HLbbEj55/jjAoMSp9IL5YCXBb5XUBmD3575HTi5CijNkG8v5zkIIYQQ0mLWHs/CiYxSzB4UhjAvRzz//PMoKCgAVBqoJz7J9/EoS0J2XgFUQ+6Ar7MGr9w0AKO2hmGDpTeglAPj/qGuWHrXGLz71huNAuM333wTgYGB7e43SMGxjRs8eDCknHgIWiccTszCQD/5050oqKCL6Nuo3vE2S0850M2XW0hebmUMfsuwr3X2vfznIIQQQkizqzFZkF9ew3ONXfQq3DsiAps3b8ZPP/3Et6snP2vdN2fbl1CPegB6jRo3dVbjhsWbkK8PhlQlp2IGuunx0e39sPDFF7B06VLr17HJeLfeemu7/O1RcGzjhg4dCjFHDlQFv0g4F5yybtufZeSd8iSLPFFva76LHOh6R15RZQx+22WK/LXluZf/HIQQQghp1rk+LDDOLavBUz+dRE5ZDe4Z3gHVlRV48kl5pFgR1h+CWwC/b1q/COrRD0LQOiJIWYIP9hfDoneHU8ZeeHp6QadWYNltffDGS8/h888/t34PR0dH/O9//2t36RR1KOfYxvXr1w+q4hRIogWCf1ds/ulreE/vgHzRCSdK1OjVLQon4zZB2WMyTtT4IK/TLPj0nHn5lTHqsH+kitqc40595OCZEEIIIc3j/Lk+/I98g7/DtfadK8STP57gHezGRPlgRkwQXnzmv7wbHrROUA2/i+9nObUJyj7XQ3DxBWsYnWBwBQzFGKY+h8pBE3A0vQzvzojGp2+9gNWrVzf6Hs899xyCg4Pb7W+WRo5tnIODA3p26wIpOw6KgO4okXSI0pZYt4cNngLLqY3Wx19WDf5nE+nY18TMrV9oMh4hhBDSfM6f63Ne+iIbMV647hTu+OIgKg1mLLquBz68tTeOH/gLv/wkB7fqm96u/wKdMxR+9Vd5hdx4vDLMEWHDr+OB8Z2DQ7Hxk1d5YNywRNuQIUNw2223tevfLAXHdmDEiOGwHF/HO9soe16L8li5Iw5zpsoBMFbx7czSHUnIKa3NISaEEEJI23D+XJ8G6YuxmaWY+sEefPFnCmJC3fHHw8Nx84AQFOblYvbtciCr6D6RxwGM5exOKDsMsn69V+4BbHnuOmjC++KbfWkYGO6Bc2ve5SXfWBtoNn+pbsDt7bffbjf1jC+mff/0doJ9wtOUZ0FMPw5lWD8cPJWEQHMW33bO5Ibg8I6NRo/ZJ09CCCGEtCHnz/XpOQsWUcLSHYm4bumfSC6oxIIJkfjhnkEI9tDzwJalVnLO3rxMGyPmn+O1jBnJbMREXSL2fvJ/KFO54bk1p3iraOnPT7Fpwx/QaDR48cUXceTIEb4/ax4SEhKC9o6CYzvg6+uL2bNnw3L8V/5Y2WsaPMsSrNv9+ozhdYrNB3/gj/+IzcHh1KJWe72EEEIIuXQX3PQKATd/sg9vbohHqKcj1jwwBA+M6ojsrEzMmTMH9957r/x1rAveda9Zn0bhHSGvri7GuyN1+HiKE4p/fxH3fr4bgiDB/dRq7N6yATqdjne9Y01DqqqqMGjQIP68hIJju/HAAw9AW5UPMfUwFCG9kRB3AhAtfFuRUwS/RCLG77Duv+DHE7x4OCGEEELaDkmSsPpQOiYt3o0DKUW4Y3AYfntoKKJ8HbFs2TKMGDECW7Zs4fvBVIXO18y/4DlcqnOw9YnRmOKVjuqjq3H/8XDk1ijhd3oVjm5by6tRfPvtt0hNTcWff/4JvV5P6RQN0MixnfDx8eGf+CzH1kKSRFSGj4RH7iG+LV3ywIxbbwdEM8x75XaQ5/Irsfpweiu/akIIIYTUKao04r5vjvABLAeNEivu7I8Xr+2Gs6dPYcqUKVi4cCGqqqvB4mKYquEdHI4U99rUilpd9OXY8eqt8PD0gsQm4RWNwqFqf3imbETCjh/h7OyMlStXIigoCC+//DL/mmeeeQZhYWH0i6hFwbEduf/++6E1FEFMPghFYA9YSnOs2zxipvJ/EGLiX5Cqy/i6Z3+JRVmN3FGPEEIIIa1nR3weJry3CxtO5WBKD39sfGQ4+gY68JzgyZMn4+TJk3w/pVIJycEdygG3onT8K42eY1SoFt89cZ11Qt3qql74tqQ79AVxyN74KVycnfHNdyvRp08fPP7446isrMSAAQMwd+7cVvmZ2yoKju2It7c3P8FZZQpJFFHsGgmhVJ6Y99nhIsx/8FFAssByaBVfZxYlvL+lPjeZEEIIIVdXtdGC59bE8hJtNUYL3pnRE0tu6Y1Df+3EyJEj8cknn0AURb6v4BECYdBc6G56E6oekxs9z2TfYrwwYyBv3JFdWo2nfz6J5457QFldhJL178DdwwNf/fAL9E5OuPvuu7Fnzx6ed/zOO++0++oU56MmIHbmvvvuw5dffgnjub+g7DgUUsoBwJV1yhGQ4t6Xz0JNSzkAscdkKNwC8NmeZMzqH4KOPk6t/dIJIYSQduVERgke+eEYT3UcEO6Bt2f0hNpYzifbrVsnl2BlhIBuUHabAIV/lyafZ5BHBV68fRJMSgXe35rAy7UZLSK0RUko27Ec3t6+eOf9Jfjxl7X44ftvYTaZeBD90ksvITw8/Cr+xLaBRo7tjJeXlzx6fOJ3SKIZYC0kzUa+7ffYPNzwwP8BkgjLoZXWr3lp7Qk5sZ8QQgghLc5sEfHB1gRcv/QvpBdV4elJUfhmXn9sW/cjn3DHA2OFEoqIQVBNfQHqsY9cNDDu6KnF0geuw6bECkz94E98/mcKAlzUcDv+LcrWvQ43lRnDho/AfXfdiW9WfMkD41GjRvFJfe292cfF0MixneYef/bZZ7Ak7IEyciTUlmrUZRb/mqFDv8hgHDy6B2L0VCh8OmFXYjG2ns7D2K6+rfzKCSGEEPuWWliJR384hiNpJYj0dca7M3tBWZ6NG66/DocOHQLUeii6jYcyaiwER/cmn4PNHRL0LvDQCZg3ojNmLNuLhLwKeDtrsWBEAL54/l7kZqazi8YoKSnGzz+tZnfRo2sXPDcjBsOCRKDiL8AYQh1vm0Ajx3bIw8OD5xNZTq6HZDHBpNRbt2WUGhA59iZAEGA5JNc9Zl7+PQ4Gs1z6jRBCCCHNi12hXXkgjZdoY4Hx/KHhWH1XDNZ9/RHGjx+PQ3FJUPa9Eeob3oCq701NBsaSaJGbegks1AUEpZrnFqcXV+E/ozviixtCsfzZecjKTOfBcN0SFBiIJUuWYMOrszBMeVRuTX3iB+B4/VVkUo+CYzvOPUZVMcT4nRds+6k8ChOGRENKPwoxK5avSy2swud7UlrhlRJCCCH2raDCgLtWHMZTP5+Eq16N7+YPwEi3IkwcPxbvf/UjxP63Q339azyvWNDUD2ipJSMvz8pIlUUwb3wDio5DIeic+bqiKiNu6BOE7U+MxOQQCeNHj0BOTn2lKhcXFzz//PN88t31118PReFZQKUDPMLlWxYkkwtQcGyn3N3d8Z///AeW2D8gmQ18nVD7D8wCBYoGLwAUalgO/2T9miXbEpBXVtNqr5kQQgixN1tP52Lie7uw5XQupvUKwPdzeuCHJa9g5n+eR0aHaVBf+yKUHQZBUNRnujqoBWgzDsFYWghBUEDMjoPp95egnvQMBK0j32egawnWPTiUT+I7vncnRo8e3ej7skl9+/bt47darVZeyVpTm2uAomT51jvy6h4MG0HBsR3jrSVryiCe2c4fBzrJl2GY42UOmPTou5CKMyBmnODrKo0WvL7hTKu9XkIIIcReVBrMPOVh3leHYDSLeH9WLwxWJmHc7IfxS2VnqMc/DsG/K6SyPChYoMo6SCsF9NIXoWzH56jx7gK4+vIJ9g75p6GZ8Z71uT/1X4vvx9Qg2Al8RPjOO++0bmNB8v79+/mIsZubW+MX1XMWED1TDpLZLXtMLkAT8uwY+0fxyCOP4L2PPoMiciSy8irg46JHnkX+1HnAFATonGE5tgaKoGi+7ucjmbhtYCj6hDQ9CYAQQgixGcZKOa82/4wcELJgUCP/DWxJR9OK+aS7lMIqDO7giXti3PHckk+QqusAIeY2CCYDxIyT0Lh4wOwSCEkAJnRyxbrXH8DhzsOgGjIXkqECQvxW+MeMQ55RY33uk4N2QOXZC8sOVOCl6Y0rWKxevRpDhgy5+AtjP3u/eS35o9sFGjm2c/fccw9UlhqIcZshOnpByDkDwSKXdis3WDDw3jchFaXDkiK3mmYWrj0FUaTSboQQQmwcC4zZxLOrNAHNZBHx7uazuPHjvcgqrcFdQ8Ngyk3C7BXHkebVH1CqYTm9FcqcWKiCuvPAeFgHd3RK+hHrXr0LqsFzoOw+CZKhEt7OOiCqPjAOcNNh94IRWFPVD8Me/RwvvfZGo++dkJBw6cCYXDYKju2cq6urnHsct5n/Y8vRBcM/d591+5EiFbQh0bAc+5V31WOOZ5TixyMZrfiqCSGEkGbARoyv0gS0c/kVuPGjv7B4awKUgoBQFyU+3ZmIgxVuQE05zAdXQUz6E7ouIyEG90VXf2dM1p3F1hduxKmEZKiveQ4KPzkHmOUVFxjrL+7PiAnCwz0EzJkxHU8+9giyMjOt24YNG4akpCQ4Orb8iHh7QcFxO8DKurno1bDEbYLg7IO01BQI2XKVCsZ3ysNART4sR360rnv19zgUVcgT+QghhBCbdBUmoLESbd/sS8Xk93fzwSWGdadLKDJBLEyBaefHsMRvh0PMND4qHODuiNGaJBx542as+eR/UHSfBPWEBRD0rhc894IJkVhxfRBSf3odj9x1G86cjqur4saNGzcOK1asgF5fX+GC/HuUc9wO8qpYKZe77roLby9eAqnLWCh7TIFhy3tQTwzl5WByKkV4D7oO+ZuXweLoDmWXcSitNuOurw/jm3kDoNcoW/unI4QQQq5c3YQzNmLMAuNmnoCWV16DJ1afwK6z+dZ1rPSalHZMHpBy9ITDwJth0rrCxVGNDlWnsf/jxUgw1AAaB6iGzociqMcFz/v+zb3R2xN45+238NqqVTwAbxATc1OmTMHSpUuhVqub9WciFBzbX14Vu2yUfVxe1yDpngXHn376KcqP/gr1oNt4XpPl4A9QDZvPt1d2GAcc+h2WPz/n/5gVIX1wOLUYd399CJ/OjoFOTQEyIYQQG9OCE9B+PZaJh1cea7TOEr8DlrgtEJw8oOx/MxSeodBqFIisTsDJL95HVk0V30/w6QT1xCcveM7v7xqIbt5q3rDj4U8/hcFgaLJU69y5c/mEe5WKxjhbAh1Ve8yrYpePzsurYqPH//d//4cnn3wSFmcvKLtPBEzVsJzbD2XEAJihQMjUh5D29VMwb/ofVNNfhcIrHLsTCvDAt0fw0W19oaIkHEIIIe1cWmEVhr8ll0itYzm+jqdOCI4eUA28BQr/rlArBIQYzuHs6iU4Ul0m7+jgDs2Nb17wnL89NBSRPg48RWLOG2+goqLivD0kRId64s6R4Zh2zRRo+90OUGDcYig4thcslYKNGF8ir+q2225DZmYmFi9eDGgdoOw0nJeSkSoL+WhxjkMnOPeahPLY7TBv+wDqKc/y9VvP5OHhlUexeGbPVvnRCCGEkNaWXFCJe78+jPjccus6VoPYcnI94OAKZb9ZUIb35+kPAYY0pPy2FKcrC/l+gmcYbw1dN+GuDmvi0T3QBevWrcOUhx6CyWRqtJ2lTFx77bWY298dvQ17IagF4MzPgEZNJdlaEAXH7Syvio0c5+fn47vvv4GgdoAiLAZSSZZ1u2LwfCjTT8NSVQYTy0ue9DQEjQP+iM2BRqXAk8P9rtZPRAghhLQ6lmK4dHsiHyiqI+bEw7z5HUDrxINeZefhgEIJT0MWcjZ+guSSTEBQoPOYmUh3jITFLbjRc350ax9M6uGPvXv3IrD/DRd8Tx8fH546ccstt8Db2xtYvwDI11/06jBpXhQct7O8KkEQ8Prrr6OwsBAbNy+HSqOHIqCbdXu1BYi5920c+eABmEpzYN62BKrxT0BQKPDrsSxIZiPemeUDBaVYEEIIsVOiJGFTXC4+3Z3Mg+OGjL8+B1SVQBl9DRRdx0FQ6+BsyEfR9i+QnZfAB5S6Xv8Qcl2ikCLWN+9gOvs64Y+Hh+PkieMICOh7wfft2bMnHnzwQUyYMKFxPvFlXB0mzYeC43aI/YP76KOPcPPNN2P/jo+gGvcoFN4drNuPFgD9HvkYhz/8Dwx5CbDs+RSq4ffwbWtjC+D222ksnNaNB9qEEEKIvagxWfDT4XQs25mItOLGk+F4CkXsH1B0HMIDY1btSWcsQfn25ShMPwaFiy/cx96DKr+eSFSoAbl1gNXyOTFwLE1GcFDgBd930qRJWLBgAaKiolql6gZpjILjdkqn0+Grr77Cddddh9Nb34dqwgIo3IOs2w/lWhB1zwdI+vK/qEw5BLODB1QxN/FtK/al8vJuT02KogCZEEKIzSuuNOLrfan46q8UFFbKXWTrSOV5MP/5BZ9sp576AgRnbyiN5TD89RXKkv6C4NMRqlEPQBHcC+dPo2MivBww068At4/pfcG22bNn45lnnuGT5i+J2j5fVRQct2PsH+N3332HqddMQcaGRVBPeR6Ci691+5lCM4JufQP5Py5EedwmWJw8oYwazbct23WOl3d7dFznVvwJCCGEkH9XeWL5nnNYdSgD1SYLFALg5aRBcaUBFkmA5exuiFmxclk2jxDeadZ8+CcYz+7kwbD6ulchOHk1+dzjIj3gW3AEy995FQvP28bKsD3xxBNQUI5im0TBcTvn6+uL75+5AVMfeBUlvy+Eevoia5ceX0s+Msu94TbteSg2voPSgysBvRuUoX34dtYikwXI942sT8kghBBC2rrj6SX4ZNc5/BGbDVGqX8/uF1QYIVWVwZK4m6ccqkfeB8ligiV2AyyJe6CMGgPNze9f9LkfHOCJzN2r8MOzK/njhgmIrKTqfffdR1dd2zgKjgk6aIvw7b19MWXxYZh/Wwj1Te/wo5Kr9IZvdTqKlaFQjnkU7nuXo3DXpxAmPmHNUX5jwxloVQrcOTT8sjv2EUIIIVcL6y4n3wI7zuZh2c5z2J9cdOF+FhOk9OMQ88/xJh2q6GsgiSIsCXsgFaZANfA2uUfARdwUbkH8+s/xzooDF2xjQTFrxkXd7GwDBceEB6+9A4/ju7uiccvHh2Fa9yLUU1/kRyZXFwyH6iJY9B4w9psLN40zyrYtgXrSU9YUjJd+i+MjyDNigvinbnZZSvU3HfsIIYSQlmKyiDCYRZgtIoxmEetPZvP5MufyKy/YV8xNgHhuL8S8BD4qrOx7AwSFElJpDqB3gbLTUIAtTdAoBfTQFCDpt6X4NjPtgu0PP/ww7r//fjg7O7fIz0laBgXHxDrrdaR/PJZ49MKD7/0i1zge+whfX6V2A0xmXrPREn0j3J29ULz1fTlA1sn/4J/95SREScQ10QH8sUN2HFQqHRSsJmNBIhD7M40iE0IIaXEsID6bW4595wqxN6kQW07X1yeuI5XlwpK0F2LyfsBQAWW3CVBP/j8Iaq11H8G1cV3/EcFqeHh5Y+3xbFhECf7GDGSu/wh7yy58flafmOUU+/lRbwBbRMExaTQL9vrJQIFbL7z44osw71pmLeHGAuM6leEj4aZ3Q+mOpVCNfQSCSgt20er5X09BrVRgQjc/mD07Q5FzAkJRMoTSDDnpSqH8+1FkSscghBByhSoMZuxLKsTuhHzsSijg3ezOJ1nMEBN2wZy0DyhMBgQVlFGjoOo346LPG+Qg4aM7BmJXUglvBFKZngVdeSZqdn2J1MKUC/YfN24crz4RGUl1iG0ZBcfkgqD07t5RyJs/B0s/+ghmQQHVsLv4LmyWrqCV84ar/HrBeYAzyvd+DdXQOyEICp5S8ezPJ+D61yKMcUiCxckXkiVhaBQAAC7kSURBVKMn1KJFnpJwOZ19KB2DEELI32Ajt7GZpdZg+EhqMcwNZ9Y13DflEMRz+3jVCUkSeAtn1eRnofAKu+jz39nbFfPG9cLuxCLMW3EYeRVmCBUFMB1aDWPakSabd7C84iFDhtDvzg5QcEyaDEqf7R6Igl6uWHVoA8waPVQDboNUVQwx9yyUIXKtxhr3DnDo7YSqAyuh6jWNB85mScC92dfgfs16POC2F8KAe6EK6AHhxKrL6+zDJvCx10AtMgkh5PK0kytuWSXV1mD4z8QClFSZLrqvmB0H8exOiKwOsVsAhLAB0I5+8JLPH+2rxd2ju2JQhDt2nkzGzR/tQlq5xAeGWAMQMX47IJobfU1ISAiefvppTJ06lcqy2REKjkmTQalQEI+3ZvVAkcUBW47+AovOBcqe18JSUwHzsbVQ9ZgIKDUwOflC3/taVG1bCvWIu61l4JYaJ+O3gv54LjEDo2c+DCUbOb6czj7UIpMQQq6MnV5xqzSYsT+5ELvOFvCgOKl2Mh1rzurpqOHZelITecRS6gHeqEMRPgiqEfdd8nsoJBHX9grEtb0DebC9OTYTz646gFKLGpJohuXMDlhO/AawK6cNarK5ubnh0UcfxZw5c6DRNG4RTWwfBcfkIkFpFNRlGVh2cwRmFBbg8KEfIGkcoeoyBpJnKCzxO6EO7AbRNQBmjRM0ExfAfOxXPoJcJ030wl2nvTD5p7N4Yeqt8O2n+/ujTS0yCSHkytjJFTdRlHAqqwy7EvJ5MHw4tRgmixz++rnoMLmHH4rKq5GcXYjccolHyVJNOcS0o5Aq8iFoHKDsOARC7+l/+70cVRJiInwQ4uGA2KxSzPvqEC/1xkjlJbBkxkI8vQXSeZPtunTpgsmTJ/OybH/b1Y7YLAqOSdNBaZdrgdNroc+Px9fvTsSM19Yg9uBKGDNOQD14DpRdx8GcEw9V6UGIIf3kk6nXNJiPrYEyeiovg1OHldDZfTYfT0yIxG0DQ6Fktd7a4SVBQghpEW31ittlvLfnlNbwQHh3QgH2JBagqLZ1s06twJCOXhjcwRMVFZXYdiwR64+VA0o1JIsEqSwLqCnjxYuVnYc3+e2l0ixIxZmAiy/vbteQGUrsPJvP7wvmGliyTsOSdQpiVhxQngdRFHmahFanw6AhQzFuzBhMGD8OgYGBLXa4SNshSHXVse1YWVkZXF1dUVpaSp/0/iGj0Yj3338f77zzDqDWQdnnRigjR0Ay1UDKjYciqKd1X1XKXhidA6HwbPxmxPQIdMVr1/VAjyA5/YI7uLz+kiB7Y4+eaReXBP8J9oacl5cHHx8fyl+j42uT6BxuheNrDUIbpK61hQGGJt7bq3vewVMlWDDMguKzuRXW3bsFuGBYJ28M6+SF6ooyfLrxMPZlmyHoGo/Qsr87grrpK5FixkmImScAlZZXohAcPS/YRykA+uo8lMazSXpxkAqSAUmsfXIJge46DOwWiinD+2LIbU/B2d27uY9Muye20t+6y40H2/zIMSsptnBh467krETKmTNnWu01tUcsp+qxxx5D//79sWjRIhzf/w3EpD1QDZ3HA2NWLL2uJqQ5bBDUkhnmyiIIjh6NnudkZimmfbgHsweF4bHxneGiU9vNJUFCCGntcpxtSv4ZiEodTmuisbtQwO6tWhxcswlGixyI+jhrcUOfIAzv7MVHibNz8/HGD9vx0a81fBId4AChiRj4/MBYTDkEC0utKEqHasDNUA249YKvCXLTQleSgtPbf4Ex5wyqTdWNtoeFheG2227DWJ9idCjZiXx9J/jUxEKRuLZtHlvSotp8cMx069YNW7ZssT5WqWziZdulzp07Y+3atVi+fDleemkhTL88A2XfG6HoNpG33lQoVZAgwCKoLgiM67BqO1/+lcLTLV6Y2g2TvaIgtMVLgoQQQq5YXhlLlSjA7uQB2JPVHwUWB75eq5AwoIMHhnfyxvDO3ujs64SUzFw889EqPJRpgsIvCoAHBLdLPz+rHiFmnICUcQLQOUMZfQ3UYTEX7NfTTwuXkiRs/W4pzlXIKRQNKZVKvPvWIoz3LYFLdSrgrQVy8iGq9YCzL2DOoMGadsomokwWDF9JlxmDwcCXhsPodcP4bCH/DDt2LAtHEATcfffdSEtLw5dffgnLkV/4pSzV4Dsg1baUvhiputRa0SKv3IAHvjuCER27YGGHWxBSFScHxj1msG924RezS4esJFxd/lr0jLZx6bAFjjGdp3R8bRWdw+3v+NaYLDiYUoTdCYXYnViA+Jzy2i16RLmacJ1TOoaGOaLfyKnQObogLSMTDy9chKMVrlAEdgfQAYq/+RMvVZdBzIwFLEYeRCs7DALYch5/vYTQwoPY9dNyHKxLlTjPyy+/jNmzZ/PgGIe+AE6ug8grfZwAXIIgmg2QWN6x2QB4RTb994jY5Hl8ud/PJoLjhIQEBAQEQMcS4wcN4pf1WW3Bi2Hbz0/FYPLz81FTU9PCr9Z+sZOK5enUndDr16+vP9FyzsLy6wt8Uh5rwylcJIeIBcaW9BNQeIdbW0/vTCzGhGQf3DGgJ24N8IWmhJXrubC7ERI2Ayn7AYUaKGItPzVAp3Gw12N8NfOw2gs6vnSMbV1bOIfZ904qrMH+1DIcSC3DscxyGGqrSrg7qDAxygMDQl3QL8QFXo5qvj4+Ph7jb3sQad4DofAKB1yHQOH6N9+nugySoQKCiw8EvQuUHQc3uV8vXQFydnyD5NMncQ5A8nlzvplZs2bxQR0PD/mKZmFhobwhLw/QdZRHistzAa0/xE4jUVpaAil4ABT+w+R9iF2cx+XldR/cbHxC3h9//IGKigqeZ5ydnc2D3szMTMTGxsLZWQ6uLmfkODg4GMXFxTQh71+ezOwDhre3Nz+Z2f233noL33//PT/B6wieYVAOmQsFzxm7xCgAayhy3qWwDt6OeGVaNwyIuHASBf74L1BwFnAPB4qT5U/0k16HPR9jQsfX1tA5bJ/HN7/cgD+TCrCndnSYPWY0KgX6hbljWEcvPpEu0tcZCoXA/ybs2LEDD7/2Icq7Tm9yYtw/NTRQBeOxddjzx8+X3K9fv3645557MHbs2IunY/KR41X1kwZ7zIDYZw69D9vpecziQXd397+dkNfmg+PzlZSUIDQ0lFdNmDfv8pLkqVpFy84uPXXqFF544QX89ddf9TsrVFBGT4Gi+6RGZd3OZ4ldD0XEYAgOjZPM2CSNZyZHwdNJ266qWtBMfzq+to7OYfs4vixVgtUZ5jWHzxYgLltOT2RYAMwC4WGdvdE/zAN6jfweX11dje9//AXPL/sJykG3Q1DKo8bnkyoLAa0TBFWD9/dLGBLuCueMvVj7+Xu8msSlJo7ffPPNGDBgAF/8/f3//smbqPQhqvRUNaiFUbWKZsa60rBJYYmJic391ORfTJhcvXo1H+V/6aWXeC4ya7FpOfYrxLQjUA6eC4VHML885ioYUKapH0FQdp8MsSgNUmURFN4R1vU/HcnAltO5eHpSFGbEBPORCGoQQgghLYONkyXkVWDXWbnmMCu3VmOS0+Y8HDW4tmeAHBB38oafa321iJS0dHzw7Vqs2h3LJ8YBrlANvfPC568shuDozu//3ShyN39HdEEGflzyKoyVpdh+iX1vv/12PjLMRolZfNAslT4ox7jds4mc44ZYikVSUhL/B0HaDjZJj3UNGjNmDD799FO89tprfD0rrWP+/VUouk/kb5ylggOk5AMIDwtFqiBP3qsrzs5qV7KW1HX5yqXVJjz180ks23QEM71TcX1XF/gMaCP1OwkhxAY70ImSxFsus7Tc4koj/jpXX3M4t0xOlVArBcSEemBYZy9eWaKrv4s8QMGu9lksWL9zPz5ZuxtHsqqgCO3LCqFBGR12ye9dFxg3xUVhxFDPapzYuBLnYg/hKMCXpowfPx7z589Hnz594OAgV8EgpLm1+bSKJ554AlOnTuWpFFlZWfzy/bFjxxAXF8dzVS4HpVVc/csgbL/XX38dK1eutK5jdSuVg+/gkzFEVhf59GYIAy/8kCOxmcJF6VD4dLxg22i/Gtw0ZjDGdPHluW72hi5J0/G1dXZzDrfRzp3/5PhaRAlVRjPPEz6ZUYrjGSXYe64Qp7PrJyd19HHiI8MsGB4Q4QEHTf3YWVZ+ERZ//wd+3ncWppD+//pnUElmhBuSUHZ2P9KP77lkqgRr1/zUU09h5MiRUKubTtNo7t+12GMG8koqbf8cbsNEagLy72RkZPAcIjazlAXDQ4cOxb59+y47MCZXUYM3GB/vKLzz+suYO3cunnvuORw4cABSSRbMf7wORZdxUPaexgNjsSQT6uI0WMLrS/KwPDTBpyPEghRImSegiBxlrWyxLUeHbd8egYtOhev7BOHGvkHoHvg3U54JIeRKsfeyujkOrA47Y0NzHAxmCw9+j6YV41BKMY6llyCzpL7xhatejQndfDEowhNDOnmhk4/8HsvGy9IzMrF41Wb8FFcGeNcNUrgB/yIw7qrOh1N5Ovb8/DmMZgNOXWQ/R0dHPPPMM7j11lt5DnGr/K5ZrB4y5ep8b9Imtfm0ioYjj8T2/pj06DcPv/zyC3777Tc8++yzKCgogBi3EZbUw1B2GQ1lh8GwuAVCMhmgqciGyb3+0pzCKwzwCuN5y5bE3VB0GFp7CQ8oqzHzRiJs6eLvwoPk6b0CGk/gI4SQf4qNIrKykRYDUJIOxP7cZkaPz8cC2oziahxNL8GxtBIcTS/Gqcwyayc6JtzLEdN6BaB7gCt6BLqis58zFBCRdC4Fv2/YhN8OnkOiY9cGz+oDePv849fUxakamsIkHFn/LSzlBTjW+BUDrEOdaOaTt++9/0HMufMufoW4VTTVpZWC43atzQfHxIZcpA00y0dmqTHjxo3DJ598wtMthMoCiIdWwXJ0DRRhMVBGjoSJ1b5kQbHZALHBLGZFSB++MOa9KwC9G5SdhlgndZzOLsPLv8XxZXxXX9wUE4yRkd5QKxU2dbmUELtli//m2Os8uwmoLpIfl2XIP0MbGD2uNFqwN6kQxzJKcTSthI8KF1QYGo0KD+rgid4hbugd4o5eQW5w1AjYf/w0Vmzfh3dSKlDtFNjgGXVAo8C4Mdb9FKLlgrbNDYXoDFAXJiLz8FaUpcbi+CVSJSbFdMSsjlUYEuULB6UZiA4HWiswrvtdU5dW0gAFx+SqvcGwJi7/+c9/MHPmTN6oZdWqVRAsRkhJf8Gc9BcEjxAoIkdCCu/PJ4s0RTVoNr8Vc87AEruRd0pShMqBM7MpLpcvXk4aTO8VyAPlSD9nu7pcSojNscV/cyyAZ6PFLDB0DwYEdau0EmaT6JLyK3gQzEaEj6SVICG3HGJt7KlUCOji74xJ3f3QK9gNPYJcYTGb8NPuk/huSyLew/lpZ66A08VT0cS8RMBsgODsDcFZHjnmJdnOK8vmpjBAV3IOBbF7UJVyHImsm9wl8oanTZvGq0qw+8IfT8rH8ryBlFb9XTN15dxYl1bejIq0VxQck5Z7g6l7fB5fX1+89957uOOOO/D888/j0KFDfL1UlAbL3hWwHP4RiohBUEaOgODadJ1KHhT7RfH7lvgdfEKHIqAb76TEFFQY8dmeZL6wS4g3xQTxUkRuDpqmR7htcWSLEBu/qtSmsX//3a+Xg3qltskP/C2hsMLAR4LZwgLi4+klKDeYrdv9XLQY3sENHfzc4KbXQLSYsPZgIr7eV4av96We92wXD4JZrWExaS8PhOHgDmXUaL6+qYnQjJNggKIkHUWn/oQl8xRyq4ohXGQUQ6/XY8KECTwYHjVqFG+60KZHas8v58ZLuVFw3J5RcEyaT1P1Ii+hV69e+PXXX7F27Vre655VI+GMVRDPbOWL4BvJR5MVIb0gKJo+XVlKBiNVlUBMOwpoHCB4d4CglPc/mVnKl1d+O41xXX1xo0s0hptOQNnwjdkWR7YIsRVtLRhq5g/8/5TRLPK0MDZpjucLp5cgtbCq0T56tRJ+LjqeKlFtMCCtxICcMgN2JJY02OvijZYYy7l9kHLP8nryrOGSovMIXjVIrkvctCAUIP34nzDlJUEqTIPB0CBYFFi6XOP92YgwK+XJAmJWZu2iHemuwnElxO5LuTUHKuXW9kuvsM5KH3/8MV+a7H2ud4Wi41AoOw+H4Ojx96+1OFMOjlmgXFvpoiFfnQXXeWfixig9Og67CdjyYuPLfOyP+eQ3cbXZTRmsNoqO71U6xplp8MneCkXBGfnflNkIJGyWd4iaAvSdY9tXZv7BlSb2pzartEYOhFmKRFoxYrPKeIDcXNjEZjHtMKSCZEis2o+xCspOw6DoOu6SnUrdDNkoP7ULNVnxkEqzWDHjS34fVl+YNeNiCwuEWVAcGNgwh9m20ftEy6NSboRcBnYZ7tFHH8X999+PnTt38tHkjRs3orKydrSiuhTiyd8hxv4BIagHlJ1HQhHY/aLPp3APbFQ3+fw2pbk1SnycHoKP04HQIwcw2i0GA6oqEGPOhhdsaGSLkLYoZTdwdhXA/t0lbJJLY7mHyqPGbJ0tB8bMZVxpqjSYcSKjlI8GH0krxua43GZ9CfxKWfZpSIWpkApTIFWXQgjoBmX4AF4F6FJU6YdRfXYP/1rUlCGvQeUIqPW1LUJkrAXz8OHDrcFwZGQkAgIC6MM7sWuUVkHaFK1WyzsgsaWmpgY7duzggfKmTZtQVVUFSCKk9OMwpx8HnL2hZJcHOwxucnS4zvmBMSNZzNa0C3YZ84tCJ3wB+RJjoKocQw1AjJiIfh38EerpwCtuEHJZKH8dKM2ozTEOAwoS5ONiS/nGV5hDLebF41xeOXaeLcDPRzJwKqusWb+dVFEAMf8cpNJsnnYmmY1QBHSFssMggC2XIFSXwHRyA8TUQ3yQgTE23gNQy53mvLy8eH3h/v3780CYBcH03kfaIwqOSZvFqltMnDiRLyxQ3rZtGw+UN2/ezNMwUJ7PJ+/VlYPjeXS1k0nY5URBc/HWoiwwlioK+WgLu+Sr8Jcn9zGZZmf8kAb8kMb+iMfDw0GNgR08eTvV/uEeiPJzhupiZeIIofx1wDWIXZ6Rg2GVRh45trV844tgaRDHhR74OdMDP5/uAgPGyxt272qW55eqiiGx/F42mmusASxGCD6doAz/+wYckmiGmHwQlrO7IOUnXrTqD+Pi4oLZs2dj1qxZCA8PpyCYkAYoOCY2EyhPnjyZLyww3rp1K9atW8cDZRY4i+f28UVwD4KCpVxcxh8SwcmTL4yYfQZSwTlApYbg7AtFULR1v6IqE9afzOELo1crEBPmgX5hHogJc0fvYHfoNZeeEEPaEVuszNDcwoYBWiNQEA9ETpSD4+KUNjv5ipVLs9ROv2Efe1mZtJyyaiQXVCExrxx/xObgYEpxg69g7xvye0dzkViFBJYC5uDOl799zVmnICYf4KkVqGr42mTCeWlrLBBm5dS6devWcm2YCbETFBwTm8Pe6K+55hq+sFSLLVu28ECZ5SibizNg2f8NLAe+43WTBd/OcsUL346XHEnmI8cNRo/ZHx6ef6fR88L3dfU+mWqTiN0JBXxhVAqBt7DuF+ZuDZo9HK9S29PmRikB7bcyQ3NizSJi5gJtdFIpC4YLK43IKa1BamElUgorEZdVhsNpxcgtu3i93pYksGOl0V/4WoszIGWcgMjyilmOcGURn8x0sQm7w4YNw6BBg9ChQweeH9yxY0fKDybkClFwTGwamzV97bXX8oVN3mOBMku9YLcm/sckBVLcJj6OIrgHQ+EXyQNmhV/nSwfLAd0aPZaM1YBSJRfDP49ZlKx1ST/dnczXdfB25EFy3RLsobeNy5a2nBLQVgJ7KlPVqiyihPxyA7JLq5FdWoMzOeX83yarF1xabUJbxd5jpMJkHgCLrNJEURpQIX8Abyolgg0ODBw4kAfBERERF9YSJoT8YxQcE7vh6OjILxuyhZWDqwuUWa6yyWQCitMgFqcBpzfDIggQ3ILkQJmNLrMGIuoLJ+7VEZoY0bmUpPxKvqw8mM4f+7po5VHlUHl0uYu/C+9s1ea0VkpAcwS2bSWwv8J63+TymSwi8soNyCmtRkpBFY5nyB9KWVUIW8LmRLDawfzDexELhlOBivwL9uvatSsmTZqEAQMG8CCYNVBiqNwjIS2LgmNil5ydnXHdddfxhdW5ZrnJLFBm1S94oCxJkIrT+cKajfCRZbeA+mA5tA8EofkuCbNLtb+fyOZLHZam3D/UDf0ifBDorkegmx7+rlooLc1X9/SKg8/WSglojsCWcn1tmsFsQW6pAZkl1TiVJZdAY0tGcTXaIqXFCEfBAEm0oMZohlFSyGXq1Ho5RaKWZKqpL7dWmMbTI9hkYjkRW8aq87AgmFWJCAoKumROMEupaFVt5QoNIS2IgmNif85783bpOQs33HADX1igzHKT2ajysWPHkJ7ORnYlwFQFKfc0pPwEiPHb5Oksrn68ixTLWVZ2HNL8L9MC7DlXwpemuOhU6B/uiQhvRwS46hDgprcG0axb1hWlaVxu8NlaKQHNEdhSrm+bVW208DSHxLwK62jvsbQSVLJ/BFcZT5HigWltmze+KFgJm0ZBLSOIZiglM/8wLVrMsJhNvB4wmzwnqbUoYxPn2Jeo5AlwPBDOT4TIg+FUORAuy+Pfz8/PjwfAEyc+jr59+/KUMJvUVq7QENKCKDgm9ucSb94sV++mm27iC1NQUIBjPyzCsZ3rcDS9CkdTS1DCSiixup+lORBLc4Ckv2DZuZTXVVYEdIcQ2IMX2m9pZTVmbDmdC5xuerubgxrRQW48WA50qw2e3fT81s9VB3XDcnOXG3y2VkpAcwS2bSXX185H1lint4YfzCoMZmQUV1kbXrDOb6ezy1v3NbKGFqYawGSoDXxrg2CV9rJSpCRJBEQLJEEBSamTV7IAuC7zqqZcrj2cm8BvpZIcOV+4LBe9evfGNZMnYdKk//Bc4BY9z06uuvrnGV2hIe0ABcfE/lzBmzcrej82xIKxw1xYWz1IFl+kOvbEMc9rcfToUb4cOnRIDpZrKiGe2w+c2w/L7s/kYNk3EgIrxh/Wr9lePmtQwmb5/11aR0mVCbvOXpinWMfdQY0IbyceLAdUxyCwWEJAuQGBCicEdIyCy3lBTqtqjsC2OQP7fxPg2sHIGgt40workVZUhcziaiTkVfBR36T8ChRXtd1JbXUEVmlG6yQv/4TZyANdlgcsluXKAXB1OZ8gx+53jgjFoEGDMeDaQejXfyA8PeWW9uwD6VWbS3BiFXCyFc4zukJD2gEKjon9udI375oyoDybl25jl1HD/HsibPp0TJ8+nW9mOcrx8fE8UD58+DBWrVolf115PkSWO5i4B5ZdnwAKJeDoBcEzFMpOQ6Hw7/KPXn5d575/iwUxh1OL+QKwIGFU/cY0AOvW87v+rjqEeDjwlI0AJyVcimPhXJMNZw9fOHceDmdnJ57i4axTw0mrgoNG2fxBdVubxPZvAtxWGFljo7k1JpEHtWwprzHxMmVZJdXIKq3hebzZ7H5JDXLKamBPWCoDryTDAuIrGVmuLJKD3nI54K0LfNkS7GjGEH8zhkb6oFdPLYJH3A71oHvQprTWCG5buUJDSAui4JjYnyt989a5AM7+cnArWgCda6PNbHJM9+7d+XL77bfjvffe42XjTp48iV27duG7777js8f515bnQirPhSn5QP0TsEu5rDlJYDcogntC4RGCtoSVu2IL5Cp0LPmkdgGw/8Tffr1aKcij065ySgerzMECaWceUKvgYr2vhlPtOieNCoq2WK2jOQKPy/xwxkqOVRrNqDTIS4XBgooaObhljwsqDNbglt2y31FRZePGv/ZKYmkLlUVy+kJNOU+NEFinPb5o+S1vC88WncsFucL8OVjwW54HiX2ArS6tDYTz+a2XoxqDBw5Alz5dEBExlFeCCAsLq88Dtl45aMMBIDvPclph8mxb+yBLSAsQJDbkYOfYJCxXV1eUlpbynFPyz7BZ0nZZQujg8vpRQvZHJnrmP3rzz8/P55P82MJGmVngbKmdWd5kGCgoAWev2sC5OxSB0RD0rXd+8nQOnm+ta7Kec0tSCgL83XTwdtbCWauCs17NA2gWTLPRataBkI1Ys4YritpRa/bGJbKqI5J8X6q7zyZPnb8O8jp2DldUVELv6ChPoGLrarfXPQ+ffJV1DFJuHCSFGiLLPWWBiF90o+9Xd98sijzQNVkkVBnNqDEYUVGUg8oaAyokPSqg5xPP2MhueyZVFsoBb3VtwFtTBokt1WX8g6Uc/GoBrQMEdgXGSV7AOlmybU2VQ2MjvnzkVw562eJuKUYnZTa8HZVw1yvg2bEvwofP5Pm/9lAP2Po+7OYIBc85bhDA21Fue2uy2791bYjYSsf4cuNBCo7JZbPbN4ymRoma4Y8MC8xSU1OtwTJbTp44AUNlCZ/xzjvwqfUXhM68+5WjOwRX//rA+bymJIS0NhbUStWltUEuC3bL5XU1ZRCrSiEYK/h6gQW/TG2wKzg3DHxr72sv/PcmWUzWVAetuRL+zmp08HNDJ393hHk7I9DbHR4eHjzYZQsb9eXpPi3077mtsNv34TaEjnHLo+C4DaCR4+ZBbxj/nmXfpyj46xvkVgnILSpHrltv5Dp24X/scnJykJubi8TERN4Wu0lsVNfFzxo4CyxwdgtohldG2jueu1sX6FY3CHhZTj4f4S0HzAbAYuSlzHjTHHa1RaODwG/18mPWOpqlEmmdILBRXxYMs5JnF3xDERpLFdw1EoLc9YgM8ECnAE9EBnkiwscVPs7atp1600rofZiOsT0Q2/jIMeUcE3IVKYvOwtfTFb6danNZvT2AyY9f8Ibh6emJkpISHiw3XNg2Njlw37698hccXt0gRcMbgmtt4Fx3ywLpS5SukqpK+Ogfu2WBEM/pZAE4+xrWzMDBvclRPdKGWxBXFgKVxZDYCKrFWH9+sOon7I9Q3X1e7qxaDnjZfSX73eshsOCWnQPugfyxggXB7AqHStdkbu+lOGkEBLnpEeLlzCd9hns7Itjdgd9n+ekaFY18EkLaHgqOCbmaLnOyllKphLe3N1/YRMCLYcF0UVERD5rrAuisrCwkJCQgPn4vzpw5A+jdLgya2S0LfB3c+ALP0PpL2WajvLARQjapqTQbYOtZ+gcLjtjERRZg1QZaPGDij2sXHoQ1vK+4okoC9orlyILlybImFPyWPa4GampTE1gwa6jko7IsUOWl/KzHtjaYZZM+2XFlE9K0zgBbGo3c1nZoY90e64Jctv8VUliMUAsi9CrASaeGq4MW7s56uDnq+OKi18CR5YbrVHCszQ131CihVSmhVgKGkjyEF+2Ga9kZSN6REHrOgtaB5nsQQmwD/cUi5Gr6p2WQLlJ3l12OYrWa2dK1a9fz9nUCvG/l+0pqBxQXF1tHn1kKR0pWMs5klyC3GigVdagQNag2WWAwi3wiGtRauSKAztlaJeCfBFoXw9ru8qCbB+ImORhv9NgEqe4+qzvLbk3VEFiDBh44CvJoJ5s8yBs9NFzqup4JcmDOXj+fHMmm3rF15+/X1GP5PstjZc0guNrX1XDhr6tu4fOb67quof4+f1B3nxfi5UEun4Dp5MmDWEXtaD0Lcv/JhwnBbIBSMkEjiNApAQcN4KyzwNVBxQNbT1cneLk4wvX84FYrT3rki04FB7XyX6Uz8Ksfqb/DK+lHKNj5k38SUCmpwgEhxGZQcEzI1fRPyyBdSd3dJvYV+s3jk5fY0qVLff1lo1lEjdkCkZVsqKVSKqBTKfhtHVbrubq6GuUVlSgqq0RxeSVKK6tRWlGN/KJSZOUVILewmN8vLC1HcVklD7TlclssgG1YgkvTYF3dYy2gceD7X+ml+6uhpTJfFaIJGskAnaUSesEIJ6EcTo6ucPYO4qO1bJSWBbYezg58BLdhEOukVcJJq4ajVslHb9tUfm5pRm0pvLCrW4OXEEKaAQXHhNhCG+Erqbt7BfuynE+21JU8u1iAxWo9s4VNYAhsgfl/dcE3m4hYWlGJkvIqlFRUIb+kAkXl1aioNqC8ygAp9xQ0RfEQVCqoJTOUft2hComBUsmCeSXvTqZQKHlgz9YpFcraWwVUKiUvA1deXg5PDw+oNSp5vZLtz/aT9617zPdXyF+rVsv7suPDRpLZYRLYf9ZB4tp1bKSaBb21I8f16yQYzRIs7DizAWxB4q9VyUamzZXQxa2GqjARCh92ntxo+9UVXIOA3JqrX4OXEEKaAQXHhLSU5mwjfCVd//5Be1cewLXiwGPD4NvvvG0scGc1hFldYfaBQ3nyBygKz0Lpc+UfOFpzpr+j9mJb3ICBd8GuhA0DtEagoA030SCEkIug4JgQW2jveiW5ypezb3OOarcwFrhrVLWRu9rF/gJJe8QmA8bMlScSEkKIjaHgmJCW8g9GcJslV/ly9m3OUe3mYkMBe5tFx5AQQv41Co4JaWuVKdryqHZLBl9tMWC3NXQMCSHkX6PgmJC2VpmiLY9qt2Tw1ZxpKO0VHUNCCPnXKDgmpD36p6PaLRl8NWcaSntFx5AQQv41Co4JaY/+6ah2SwZfbTkNxVbQMSSEkH+NgmNCSNsIvv5tGgpNRmvbqTyEEGIjKDgmhNhH8PVv86EpuCaEEELBMSHEbvzbfOjmmGxIATYhhNg8qtBOCLEPLB+a5UH/03zohsE1u/0nkw3rAmz2teyWPSaEEGJTKK2CEGIf/m0+dHNMNqRSaoQQYvMoOCaE2Id/mw/dHJMNqZQaIYTYPAqOCSGkuSYbUik1QgixeRQcE0JIe6jmQQgh5LLQhDxCCCGEEEJqUXBMCCGEEEJILQqOCSGEEEIIqUXBMSGEEEIIIbUoOCaEEEIIIaQWBceEEEIIIYTUouCYEEIIIYSQWhQcE0IIIYQQUouCY0IIIYQQQmpRcEwIIYQQQkgtCo4JIYQQQgipRcExIYQQQgghtSg4JoQQQgghpBYFx4QQQgghhNRSoR2QJInflpWVtfZLsWmiKKK8vBw6nQ4KBX2uomNse+gcpmNs6+gcpmNsD8RWiifq4sC6uLBdB8fsF8AEBwe39kshhBBCCCGtHBe6urpedLsg/V34bCefULKysuDs7AxBEFr75dgs9omLfcBIT0+Hi4tLa78cu0THmI6vraNzmI6vraNz2H6PMQt5WWAcEBBwyRHrdjFyzA5AUFBQa78Mu8FOZAqO6RjbMjqH6RjbOjqH6RjbA5dWiCcuNWJchxJHCSGEEEIIqUXBMSGEEEIIIbUoOCaXTavV4oUXXuC3pGXQMW5ZdHxbHh1jOr62js7hltfWj3G7mJBHCCGEEELI5aCRY0IIIYQQQmpRcEwIIYQQQkgtCo4JIYQQQgipRcExIYQQQgghtSg4JhfYtWsXpk6dyjvIsI6Ca9asabSdzeF8/vnn4e/vD71ej7FjxyIhIYGOZDMd3zvuuIOvb7hMnDiRju9lWrRoEfr168c7Yvr4+GD69OmIj49vtE9NTQ0eeOABeHp6wsnJCTfccANyc3PpGDfjMR45cuQF5/G9995Lx/gyffTRR4iOjrY2SRg0aBD++OMPOoev0vGl87f5vf766/x94JFHHmnz78UUHJMLVFZWomfPnvjwww+bPDpvvvkm3n//fXz88cfYv38/HB0dMWHCBH6Sk39/fBkWDGdnZ1uX77//ng7tZdq5cyd/s923bx82b94Mk8mE8ePH8+Ne59FHH8W6deuwevVqvj9rL3/99dfTMW7GY8zcddddjc5j9t5BLg/r6sqCicOHD+PQoUMYPXo0pk2bhlOnTtE5fBWOL52/zevgwYNYtmwZ/0DSUJt9L2al3Ai5GHaK/PLLL9bHoihKfn5+0ltvvWVdV1JSImm1Wun777+nA/kvjy8zZ84cadq0aXQsm0leXh4/zjt37rSer2q1Wlq9erV1n9OnT/N99u7dS8e9GY4xM2LECOnhhx+m49mM3N3dpc8++4zO4RY+vgydv82nvLxc6tSpk7R58+ZGx7UtvxfTyDG5IsnJycjJyeGpFA37lA8YMAB79+6lo9lMduzYwS9XR0ZG4r777kNhYSEd23+otLSU33p4ePBbNlLERjobnsNRUVEICQmhc7iZjnGdb7/9Fl5eXujevTuefvppVFVV/dNv0a5ZLBasXLmSj8yzy/90Drfs8a1D52/zYFeZpkyZ0ug9l2nL57GqVb87sTksMGZ8fX0brWeP67aRf4elVLDLSuHh4UhKSsIzzzyDSZMm8TcLpVJJh/cKiKLI89uGDBnCA7S6c1ij0cDNzY3O4RY6xswtt9yC0NBQnlt/4sQJ/Pe//+V5yT///DOdw5fp5MmTPFhjKWssH/OXX35B165dcezYMTqHW/D40vnbfNiHjiNHjvC0ivO15fdiCo4JaWNmzZplvd+jRw+eo9WhQwc+mjxmzJhWfW22OGIRGxuLPXv2tPZLaXfH+O677250HrMJvOz8ZR/42PlM/h67csQCYTYy/+OPP2LOnDk8L5O07PFlATKdv/9eeno6Hn74YT4vQafTwZZQWgW5In5+fvz2/Nmk7HHdNtK8IiIi+KXpxMREOrRX4MEHH8Rvv/2G7du388k3ddh5ajQaUVJS0mh/Ooeb7xg3haVeMXQeXz42qtaxY0f07duXVwhhE3kXL15M53ALH9+m0Pl75VjaRF5eHvr06QOVSsUX9uGDTehn99kIcVt9L6bgmFwRdqmfnbRbt261risrK+NVKxrmapHmk5GRwXOO2cgb+XtsniML2tgl0m3btvFztiH2h1CtVjc6h9nl/rS0NDqHm+kYN4WN0DF0Hv+7FBaDwUDncAsf36bQ+Xvl2JUilrrCjl3dEhMTg1tvvdV6v62+F1NaBblARUVFo9EdNgmPnchssg1LlGf5ha+88go6derE/yg+99xzPK+Q1Tol/+74smXhwoW81iP7EMIuQT/55JN8dIOVyyOXd5n/u+++w6+//srr8NblrrGJo6wuN7udN28eHnvsMX68WY3Thx56iL8ZDxw4kA5xMxxjdt6y7ZMnT+b1S1nOMSvZNHz48AtKOZGmsQmMbK4Be88tLy/nx5OlVm3cuJHO4RY+vnT+Ng/23tBwHgLDSr+y94S69W32vbhVa2WQNmn79u28lMr5CysxVlfO7bnnnpN8fX15CbcxY8ZI8fHxrf2y7eL4VlVVSePHj5e8vb15iZvQ0FDprrvuknJyclr7ZduMpo4tW7744gvrPtXV1dL999/PSzc5ODhI1113nZSdnd2qr9uejnFaWpo0fPhwycPDg79HdOzYUVqwYIFUWlra2i/dZtx55538379Go+HvB+x9dtOmTdbtdA633PGl87flnF8ir62exwL7X+uG54QQQgghhLQNlHNMCCGEEEJILQqOCSGEEEIIqUXBMSGEEEIIIbUoOCaEEEIIIaQWBceEEEIIIYTUouCYEEIIIYSQWhQcE0IIIYQQUouCY0IIIYQQQmpRcEwIIYQQQkgtCo4JIYQQQgipRcExIYQQQgghtSg4JoQQO/X9999Dr9cjOzvbum7u3LmIjo5GaWlpq742QghpqwRJkqTWfhGEEEKaH3t779WrF4YPH44PPvgAL7zwAj7//HPs27cPgYGBdMgJIaQJqqZWEkIIsX2CIODVV1/FjTfeCD8/Px4g7969mwJjQgi5BBo5JoQQO9enTx+cOnUKmzZtwogRI1r75RBCSJtGOceEEGLHNmzYgDNnzsBiscDX17e1Xw4hhLR5NHJMCCF26siRIxg5ciSWLVuGL7/8Ei4uLli9enVrvyxCCGnTKOeYEELsUEpKCqZMmYJnnnkGN998MyIiIjBo0CAeMLM0C0IIIU2jkWNCCLEzRUVFGDx4MB81/vjjj63rWbDM0itYqgUhhJCmUXBMCCGEEEJILZqQRwghhBBCSC0KjgkhhBBCCKlFwTEhhBBCCCG1KDgmhBBCCCGkFgXHhBBCCCGE1KLgmBBCCCGEkFoUHBNCCCGEEFKLgmNCCCGEEEJqUXBMCCGEEEJILQqOCSGEEEIIqUXBMSGEEEIIIZD9Pyg7rhXhEWK0AAAAAElFTkSuQmCC", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "from matplotlib.patches import Ellipse, Rectangle\n", + "from matplotlib.collections import PatchCollection\n", + "\n", + "# Align lengths in case filtering returns an extra prior state.\n", + "n = min(obs_values.shape[0], states.shape[0], filtered_mean.shape[0])\n", + "fm = filtered_mean[:n]\n", + "st = states[:n]\n", + "obs = obs_values[:n]\n", + "\n", + "# Use previously computed std if available, otherwise derive from covariance.\n", + "if \"filtered_std\" in globals():\n", + " fs = filtered_std[:n]\n", + "else:\n", + " fs = jnp.sqrt(jnp.maximum(jnp.diagonal(filtered_cov[:n], axis1=-2, axis2=-1), 1e-10))\n", + "\n", + "fig, ax = plt.subplots(figsize=(7, 6), constrained_layout=True)\n", + "\n", + "ellipses = [\n", + " Ellipse(xy=(fm[i, 0], fm[i, 1]), width=2 * fs[i, 0], height=2 * fs[i, 1])\n", + " for i in range(len(fm))\n", + "]\n", + "ec = PatchCollection(ellipses, facecolor=\"C0\", edgecolor=\"none\", alpha=0.07)\n", + "ax.add_collection(ec)\n", + "\n", + "ax.plot(st[:, 0], st[:, 1], \"k-\", label=\"True state\", linewidth=1.5, alpha=0.9)\n", + "ax.scatter(obs[:, 0], obs[:, 1], s=6, c=\"C1\", alpha=0.6, label=\"Noisy observations\")\n", + "ax.plot(fm[:, 0], fm[:, 1], \"C0-\", label=\"Filtered mean\", linewidth=1.2)\n", + "ax.legend(\n", + " handles=[\n", + " ax.lines[-1],\n", + " ax.collections[-2],\n", + " ax.lines[0],\n", + " Rectangle((0, 0), 1, 1, facecolor=\"C0\", alpha=0.35),\n", + " ],\n", + " labels=[\"Filtered mean\", \"Noisy observations\", \"True state\", \"Filtered 1σ ellipses\"],\n", + " loc=\"upper right\", fontsize=9,\n", + ")\n", + "\n", + "ax.set_xlabel(\"$x$\")\n", + "ax.set_ylabel(\"$y$\")\n", + "ax.set_title(\"Phase space: true state, noisy data, filtering distribution (mean ± 1σ)\")\n", + "ax.grid(True, alpha=0.3)\n", + "ax.set_aspect(\"equal\")\n", + "ax.autoscale_view()\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "id": "ddc91ed5", + "metadata": {}, + "source": [ + "## Summary\n", + "\n", + "This notebook demonstrated **Universal ODEs (UODEs) for Bayesian system identification** on a\n", + "Lotka-Volterra predator-prey system using the dynestyx library.\n", + "\n", + "**Key take-aways:**\n", + "\n", + "| Aspect | Details |\n", + "|--------|--------|\n", + "| **Known physics** | Linear self-regulation: $[\\alpha x,\\; -\\delta y]$ |\n", + "| **Unknown interaction** | Polynomial library $\\phi(x, y) = [1, x, y, x^2, xy, y^2]^\\top$; Laplace(0, 0.005) prior promotes SINDy-like sparsity |\n", + "| **Both species observed** | $H = I_2$ at $\\Delta t = 0.2$ (150 points); denser data breaks library-feature collinearity |\n", + "| **SVI (MAP)** | Fast point estimate using `AutoDelta` with ADAM optimizer |\n", + "| **Library changes** | **None** — arbitrary drift callables compose with existing `Filter` / `SDESimulator` handlers |\n", + "\n", + "The recovered $\\Theta$ concentrates mass on the $xy$ column, correctly identifying the predator--prey\n", + "coupling.\n", + "\n", + "The same pattern extends to any hybrid model: replace the polynomial library with a neural network\n", + "to get a neural ODE / physics-informed neural network, or add additional known ODE terms for\n", + "structured mechanistic learning." + ] + }, + { + "cell_type": "markdown", + "id": "a576b354", + "metadata": {}, + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": ".venv", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.12.10" + } + }, + "nbformat": 4, + "nbformat_minor": 5 +} diff --git a/scripts/lv_uode_debug.py b/scripts/lv_uode_debug.py new file mode 100644 index 00000000..67e3693c --- /dev/null +++ b/scripts/lv_uode_debug.py @@ -0,0 +1,510 @@ +#!/usr/bin/env python3 +""" +LV-UODE debug script — standalone CLI version of lv_uode.ipynb. + +Usage: + python scripts/lv_uode_debug.py [--num-steps 100] [--no-plots] [--no-nuts] +""" +import argparse + +import jax.numpy as jnp +import jax.random as jr +import numpyro +import numpyro.distributions as dist +from numpyro.infer import SVI, Trace_ELBO, Predictive +from numpyro.infer.autoguide import AutoDelta +from numpyro.infer.initialization import init_to_median, init_to_value + +import dynestyx as dsx +from dynestyx import ( + ContinuousTimeStateEvolution, + DynamicalModel, + Filter, + LinearGaussianObservation, + SDESimulator, +) +from dynestyx.inference.filter_configs import ContinuousTimeEKFConfig + +# --------------------------------------------------------------------------- +# True system +# --------------------------------------------------------------------------- +_alpha = 1.0 # prey self-growth +_beta = 0.1 # prey-predator interaction (unknown to model) +_gamma = 0.075 # predator-prey interaction (unknown to model) +_delta = 1.5 # predator self-decay + +state_dim = 2 # (x=prey, y=predator) + + +def lv_true_drift(x): + prey, pred = x[0], x[1] + return jnp.array([ + _alpha * prey - _beta * prey * pred, + _gamma * prey * pred - _delta * pred, + ]) + + +def lv_known_drift(x): + prey, pred = x[0], x[1] + return jnp.array([ + _alpha * prey, + -_delta * pred, + ]) + + +# --------------------------------------------------------------------------- +# Config +# --------------------------------------------------------------------------- +sigma_x_true = 0.05 +sigma_y_true = 0.5 + +obs_times = jnp.arange(0.0, 30.0, 0.5) # 60 observations + +H_obs = jnp.array([[1.0, 0.0]]) +observation_dim = 1 + +# Overridden by --observe-both +OBSERVE_BOTH = False + +x0_mean = jnp.array([10.0, 5.0]) +x0_cov = 0.5**2 * jnp.eye(state_dim) + +initial_condition_kwargs = dict( + initial_condition=dist.MultivariateNormal( + loc=x0_mean, + covariance_matrix=x0_cov, + ) +) + +key = jr.PRNGKey(42) +key, k_data, k_svi, k_filter = jr.split(key, 4) + + +DIFFUSION_FLOOR = 0.0 # overridden by --diffusion-floor +LAPLACE_SCALE = 0.1 # overridden by --laplace-scale +OBS_DT = 0.5 # overridden by --obs-dt + +def get_obs_config(): + if OBSERVE_BOTH: + return jnp.eye(state_dim), state_dim + else: + return jnp.array([[1.0, 0.0]]), 1 + +def make_state_evolution(drift_fn, diffusion_coeff): + safe_coeff = jnp.maximum(diffusion_coeff, DIFFUSION_FLOOR) if DIFFUSION_FLOOR > 0 else diffusion_coeff + return ContinuousTimeStateEvolution( + drift=lambda x, u, t: drift_fn(x), + diffusion_coefficient=lambda x, u, t: safe_coeff * jnp.eye(state_dim), + ) + + +# --------------------------------------------------------------------------- +# Polynomial library +# --------------------------------------------------------------------------- +TERM_NAMES = ["1", "x", "y", "x²", "xy", "y²"] +N_TERMS = len(TERM_NAMES) + + +def interaction_library(x): + prey, pred = x[0], x[1] + return jnp.array([1.0, prey, pred, prey**2, prey * pred, pred**2]) + + +true_Theta = jnp.zeros((state_dim, N_TERMS)) +true_Theta = true_Theta.at[0, 4].set(-_beta) +true_Theta = true_Theta.at[1, 4].set(_gamma) + + +# --------------------------------------------------------------------------- +# UODE model +# --------------------------------------------------------------------------- +COV_RESCALING = 1.0 # overridden by --cov-rescaling +FILTER_STATE_ORDER = "first" # overridden by --filter-state-order + +def uode_lv_model(obs_times=None, obs_values=None): + Theta = numpyro.sample( + "Theta", + dist.Laplace(0.0, LAPLACE_SCALE).expand([state_dim, N_TERMS]).to_event(2), + ) + sigma_x = numpyro.sample("sigma_x", dist.LogNormal(jnp.log(0.05), 1.0)) + sigma_y = numpyro.sample("sigma_y", dist.HalfNormal(0.5)) + + def drift(x): + known = lv_known_drift(x) + phi = interaction_library(x) + unknown = Theta @ phi + return known + unknown + + H, obs_dim = get_obs_config() + return dsx.sample("f", DynamicalModel( + state_evolution=make_state_evolution(drift, sigma_x), + observation_model=LinearGaussianObservation( + H=H, + R=sigma_y**2 * jnp.eye(obs_dim), + ), + **initial_condition_kwargs, + ), obs_times=obs_times, obs_values=obs_values) + + +def data_conditioned_model(obs_times=None, obs_values=None): + with Filter(filter_config=ContinuousTimeEKFConfig( + record_filtered_states_mean=True, + record_filtered_states_cov=True, + cov_rescaling=COV_RESCALING, + filter_state_order=FILTER_STATE_ORDER, + )): + return uode_lv_model(obs_times=obs_times, obs_values=obs_values) + + +# --------------------------------------------------------------------------- +# Main +# --------------------------------------------------------------------------- +def main(): + parser = argparse.ArgumentParser(description="LV-UODE debug script") + parser.add_argument("--num-steps", type=int, default=100) + parser.add_argument("--lr", type=float, default=1e-3) + parser.add_argument("--clip-norm", type=float, default=0.0, + help="Gradient clip norm (0=no clip, use Adam)") + parser.add_argument("--lr-decay", type=float, default=0.0, + help="Exponential decay rate (0=no decay)") + parser.add_argument("--cov-rescaling", type=float, default=1.0) + parser.add_argument("--filter-state-order", type=str, default="first", + choices=["zeroth", "first", "second"]) + parser.add_argument("--diffusion-floor", type=float, default=0.0) + parser.add_argument("--fwd-mode", action="store_true", + help="Use forward_mode_differentiation=True") + parser.add_argument("--no-plots", action="store_true") + parser.add_argument("--no-nuts", action="store_true") + parser.add_argument("--two-phase", action="store_true", + help="Phase 1: cov_rescaling=2 fast convergence, " + "Phase 2: cov_rescaling=1 to refine sigmas") + parser.add_argument("--init", type=str, default="prior", + choices=["prior", "median", "uninformed", "moderate"], + help="Init strategy: prior=prior mode/median (default, principled), " + "median=numpyro init_to_median, " + "uninformed=sigma_x=0.1, sigma_y=1.0, " + "moderate=sigma_x=0.08, sigma_y=0.8") + parser.add_argument("--laplace-scale", type=float, default=0.1, + help="Scale for Laplace prior on Theta (default 0.1)") + parser.add_argument("--observe-both", action="store_true", + help="Observe both prey and predator (H=eye(2))") + parser.add_argument("--obs-dt", type=float, default=0.5, + help="Observation interval (default 0.5)") + args = parser.parse_args() + + global COV_RESCALING, FILTER_STATE_ORDER, DIFFUSION_FLOOR, OBSERVE_BOTH, LAPLACE_SCALE, OBS_DT + COV_RESCALING = args.cov_rescaling + FILTER_STATE_ORDER = args.filter_state_order + DIFFUSION_FLOOR = args.diffusion_floor + OBSERVE_BOTH = args.observe_both + LAPLACE_SCALE = args.laplace_scale + OBS_DT = args.obs_dt + obs_times = jnp.arange(0.0, 30.0, OBS_DT) + + # --- Data generation --- + print("=" * 60) + print("Generating synthetic data …") + print("=" * 60) + + def model_with_true_drift(obs_times=None, obs_values=None): + H, obs_dim = get_obs_config() + return dsx.sample("f", DynamicalModel( + state_evolution=make_state_evolution(lv_true_drift, sigma_x_true), + observation_model=LinearGaussianObservation( + H=H, + R=sigma_y_true**2 * jnp.eye(obs_dim), + ), + **initial_condition_kwargs, + ), obs_times=obs_times, obs_values=obs_values) + + predictive = Predictive(model_with_true_drift, num_samples=1, exclude_deterministic=False) + with SDESimulator(dt0=1e-3): + synthetic = predictive(k_data, obs_times=obs_times) + + obs_values = synthetic["observations"][0] + states = synthetic["states"][0] + + print(f" obs_values shape: {obs_values.shape}") + print(f" states shape: {states.shape}") + + print("\nTrue Theta:") + print(f" Terms: {TERM_NAMES}") + print(f" {true_Theta}") + + # --- SVI --- + print("\n" + "=" * 60) + print(f"Running SVI ({args.num_steps} steps) …") + print("=" * 60) + + print(f" cov_rescaling: {args.cov_rescaling}") + print(f" filter_state_order: {args.filter_state_order}") + print(f" forward_mode_diff: {args.fwd_mode}") + print(f" init: {args.init}") + print(f" two_phase: {args.two_phase}") + if args.init == "prior": + # Principled: prior median for sigma_x (LogNormal median = exp(loc)), + # prior scale for sigma_y (HalfNormal scale), prior mode for Theta (Laplace mode = 0) + init_fn = init_to_value(values={ + "sigma_x": jnp.array(0.05), # = exp(log(0.05)) = prior median + "sigma_y": jnp.array(0.5), # = HalfNormal scale + "Theta": jnp.zeros((state_dim, N_TERMS)), # = Laplace mode + }) + elif args.init == "median": + init_fn = init_to_median() + elif args.init == "uninformed": + init_fn = init_to_value(values={ + "sigma_x": jnp.array(0.1), + "sigma_y": jnp.array(1.0), + "Theta": jnp.zeros((state_dim, N_TERMS)), + }) + elif args.init == "moderate": + init_fn = init_to_value(values={ + "sigma_x": jnp.array(0.08), + "sigma_y": jnp.array(0.8), + "Theta": jnp.zeros((state_dim, N_TERMS)), + }) + guide = AutoDelta( + data_conditioned_model, + init_loc_fn=init_fn, + ) + + def run_svi_phase(guide, num_steps, lr, lr_decay, clip_norm, cov_rescaling, label="SVI"): + """Run one phase of SVI, returning (svi_result, losses_array).""" + global COV_RESCALING + COV_RESCALING = cov_rescaling + + if clip_norm > 0: + optimizer = numpyro.optim.ClippedAdam(step_size=lr, clip_norm=clip_norm) + opt_str = f"ClippedAdam(lr={lr}, clip_norm={clip_norm})" + elif lr_decay > 0: + import optax + schedule = optax.exponential_decay( + init_value=lr, + transition_steps=max(1, num_steps // 5), + decay_rate=lr_decay, + ) + optimizer = numpyro.optim.optax_to_numpyro(optax.adam(schedule)) + opt_str = f"Adam(lr={lr}, exp_decay={lr_decay}, transition={num_steps // 5})" + else: + optimizer = numpyro.optim.Adam(step_size=lr) + opt_str = f"Adam(lr={lr})" + + print(f"\n [{label}] {num_steps} steps | {opt_str} | cov_rescaling={cov_rescaling}") + + svi = SVI(data_conditioned_model, guide, optimizer, loss=Trace_ELBO()) + svi_kwargs = dict( + obs_times=obs_times, obs_values=obs_values, + progress_bar=True, stable_update=True, + ) + if args.fwd_mode: + svi_kwargs["forward_mode_differentiation"] = True + result = svi.run(k_svi, num_steps, **svi_kwargs) + return result + + if args.two_phase: + # Phase 1: cov_rescaling=2.0 for fast, stable Theta convergence + phase1_steps = args.num_steps // 2 + svi_result = run_svi_phase( + guide, phase1_steps, lr=args.lr, lr_decay=args.lr_decay, + clip_norm=args.clip_norm, cov_rescaling=2.0, label="Phase 1 (rescaled)", + ) + # Warm-start phase 2 from phase 1 MAP + median_p1 = guide.median(svi_result.params) + guide_p2 = AutoDelta( + data_conditioned_model, + init_loc_fn=init_to_value(values={ + "sigma_x": median_p1["sigma_x"], + "sigma_y": median_p1["sigma_y"], + "Theta": median_p1["Theta"], + }), + ) + # Phase 2: cov_rescaling=1.0 for unbiased sigma refinement + phase2_steps = args.num_steps - phase1_steps + svi_result = run_svi_phase( + guide_p2, phase2_steps, lr=args.lr * 0.3, lr_decay=max(args.lr_decay, 0.5), + clip_norm=args.clip_norm, cov_rescaling=1.0, label="Phase 2 (refine)", + ) + guide = guide_p2 + else: + svi_result = run_svi_phase( + guide, args.num_steps, lr=args.lr, lr_decay=args.lr_decay, + clip_norm=args.clip_norm, cov_rescaling=args.cov_rescaling, label="SVI", + ) + + median = guide.median(svi_result.params) + Theta_inferred = median["Theta"] + sigma_x_inferred = float(median["sigma_x"]) + sigma_y_inferred = float(median["sigma_y"]) + + print(f"\nInferred sigma_x = {sigma_x_inferred:.4f} (true = {sigma_x_true:.4f})") + print(f"Inferred sigma_y = {sigma_y_inferred:.4f} (true = {sigma_y_true:.4f})") + + print("\nTheta recovery:") + print(f"{'':>12s} {'x-row (prey drift)':>22s} {'y-row (predator drift)':>24s}") + print(f"{'':>12s} {'true':>8s} {'inferred':>10s} {'true':>8s} {'inferred':>10s}") + print("-" * 68) + for j, name in enumerate(TERM_NAMES): + t0, t1 = float(true_Theta[0, j]), float(true_Theta[1, j]) + i0, i1 = float(Theta_inferred[0, j]), float(Theta_inferred[1, j]) + print(f"{name:>12s} {t0:>8.4f} {i0:>10.4f} {t1:>8.4f} {i1:>10.4f}") + + # Theta error summary + import numpy as np + err = np.asarray(Theta_inferred - true_Theta) + print(f"\nTheta error: Frobenius={np.linalg.norm(err):.4f} " + f"MAE={np.abs(err).mean():.4f} max|err|={np.abs(err).max():.4f}") + # Sparsity: fraction of entries < 0.01 (should be 10/12) + n_sparse = (np.abs(np.asarray(Theta_inferred)) < 0.01).sum() + print(f"Sparsity: {n_sparse}/{Theta_inferred.size} entries <0.01 " + f"(true: 10/{Theta_inferred.size} are zero)") + # xy term accuracy + xy_prey_err = float(Theta_inferred[0, 4] - true_Theta[0, 4]) + xy_pred_err = float(Theta_inferred[1, 4] - true_Theta[1, 4]) + print(f"xy term: prey err={xy_prey_err:+.4f} pred err={xy_pred_err:+.4f}") + + # Check for NaN in losses + import numpy as np + losses = np.asarray(svi_result.losses) + n_nan = np.isnan(losses).sum() + first_nan = int(np.argmax(np.isnan(losses))) if n_nan > 0 else None + print(f"\nLoss: init={losses[0]:.2f} final={losses[-1]:.2f} " + f"min={np.nanmin(losses):.2f} NaN count={n_nan}" + + (f" first NaN at step {first_nan}" if first_nan is not None else "")) + + # Save losses for external analysis + np.save("lv_uode_losses.npy", losses) + print("Saved lv_uode_losses.npy") + + # --- Tail ELBO diagnostics --- + n_tail = min(500, len(losses) // 4) + if n_tail >= 50: + tail = losses[-n_tail:] + tail_clean = tail[~np.isnan(tail)] + if len(tail_clean) > 20: + print(f"\n--- Tail ELBO analysis (last {n_tail} steps) ---") + print(f" mean={tail_clean.mean():.2f} std={tail_clean.std():.2f} " + f"min={tail_clean.min():.2f} max={tail_clean.max():.2f}") + print(f" range={tail_clean.max()-tail_clean.min():.2f} " + f"CoV={tail_clean.std()/abs(tail_clean.mean())*100:.2f}%") + # Autocorrelation for periodicity detection + tc = tail_clean - tail_clean.mean() + acf = np.correlate(tc, tc, mode="full") + acf = acf[len(acf)//2:] + acf /= acf[0] + 1e-30 + from scipy.signal import find_peaks + peaks, props = find_peaks(acf[1:], height=0.05) + if len(peaks) > 0: + top = sorted(zip(props["peak_heights"], peaks+1), reverse=True)[:5] + acf_desc = ", ".join(f"lag={l} r={h:.3f}" for h, l in top) + print(f" ACF peaks: [{acf_desc}]") + print(f" Dominant period ~ {top[0][1]} steps") + else: + print(" No clear periodicity in autocorrelation") + # Print every 50th in last 200 + print(f" Sample losses (every 50th, last 200):") + start = max(0, len(losses) - 200) + for i in range(start, len(losses), 50): + print(f" step {i}: {losses[i]:.2f}") + print(f" step {len(losses)-1}: {losses[-1]:.2f}") + + # --- NUTS --- + if not args.no_nuts: + print("\n" + "=" * 60) + print("Running NUTS …") + print("=" * 60) + + from numpyro.infer import MCMC, NUTS + + k_mcmc = jr.PRNGKey(99) + nuts_kernel = NUTS( + data_conditioned_model, + init_strategy=init_to_value(values=median), + ) + mcmc = MCMC(nuts_kernel, num_warmup=100, num_samples=200, num_chains=1, + progress_bar=True) + mcmc.run(k_mcmc, obs_times=obs_times, obs_values=obs_values) + mcmc.print_summary() + posterior = mcmc.get_samples() + + Theta_posterior = posterior["Theta"] + Theta_nuts_mean = Theta_posterior.mean(axis=0) + Theta_nuts_std = Theta_posterior.std(axis=0) + + print(f"\n{'':>8s} {'── prey (x) drift ──':>28s} {'── predator (y) drift ──':>30s}") + print(f"{'term':>8s} {'true':>6s} {'MAP':>8s} {'NUTS':>12s} {'true':>6s} {'MAP':>8s} {'NUTS':>12s}") + print("-" * 88) + for j, name in enumerate(TERM_NAMES): + t0 = float(true_Theta[0, j]); t1 = float(true_Theta[1, j]) + m0 = float(Theta_inferred[0, j]); m1 = float(Theta_inferred[1, j]) + n0 = float(Theta_nuts_mean[0, j]); n1 = float(Theta_nuts_mean[1, j]) + s0 = float(Theta_nuts_std[0, j]); s1 = float(Theta_nuts_std[1, j]) + print(f"{name:>8s} {t0:>6.3f} {m0:>8.4f} {n0:>7.4f}±{s0:.4f} {t1:>6.3f} {m1:>8.4f} {n1:>7.4f}±{s1:.4f}") + + # --- Plots --- + if not args.no_plots: + import matplotlib + matplotlib.use("Agg") + import matplotlib.pyplot as plt + + # Loss curve + fig, ax = plt.subplots(figsize=(8, 3)) + ax.plot(losses, lw=0.8, color="steelblue") + ax.set_xlabel("SVI step") + ax.set_ylabel("ELBO loss") + ax.set_title("SVI convergence") + ax.set_yscale("log") + plt.tight_layout() + fig.savefig("lv_uode_loss.png", dpi=150) + print("\nSaved lv_uode_loss.png") + + # Theta heatmap + fig, axes = plt.subplots(1, 2, figsize=(11, 3.5)) + for ax, mat, title in zip(axes, + [true_Theta, Theta_inferred], + ["True Theta", "Inferred Theta (MAP)"]): + vmax = float(jnp.abs(mat).max()) + 1e-6 + im = ax.imshow(mat, cmap="RdBu_r", vmin=-vmax, vmax=vmax, aspect="auto") + ax.set_xticks(range(len(TERM_NAMES))) + ax.set_xticklabels(TERM_NAMES, rotation=35, ha="right", fontsize=9) + ax.set_yticks([0, 1]) + ax.set_yticklabels(["prey (x)", "predator (y)"]) + ax.set_title(title) + plt.colorbar(im, ax=ax, shrink=0.85) + plt.tight_layout() + fig.savefig("lv_uode_theta.png", dpi=150) + print("Saved lv_uode_theta.png") + + # Filtered trajectories + predictive_filter = Predictive(data_conditioned_model, params=svi_result.params, num_samples=1) + pred_samples = predictive_filter(k_filter, obs_times=obs_times, obs_values=obs_values) + filtered_mean = pred_samples["f_filtered_states_mean"][0] + filtered_cov = pred_samples["f_filtered_states_cov"][0] + filtered_std = jnp.sqrt(jnp.diagonal(filtered_cov, axis1=-2, axis2=-1)) + + fig, axes = plt.subplots(2, 1, figsize=(10, 6), sharex=True) + for i, (ax, label, col_t, col_f) in enumerate(zip( + axes, ["Prey (x)", "Predator (y)"], + ["tab:blue", "tab:orange"], ["steelblue", "darkorange"], + )): + ax.plot(obs_times, states[:, i], lw=1.5, color=col_t, label="True state", zorder=3) + mu = filtered_mean[:, i] + std = filtered_std[:, i] + ax.plot(obs_times, mu, "--", lw=1.5, color=col_f, label="Filter mean", zorder=4) + ax.fill_between(obs_times, mu - std, mu + std, color=col_f, alpha=0.25, label="±1σ") + if i == 0: + ax.scatter(obs_times, obs_values[:, 0], s=12, color="black", zorder=5, label="Observations") + ax.set_ylabel(label) + ax.legend(loc="upper right", fontsize=8) + axes[-1].set_xlabel("Time") + axes[0].set_title("EKF-filtered trajectories (MAP parameters)") + plt.tight_layout() + fig.savefig("lv_uode_filtered.png", dpi=150) + print("Saved lv_uode_filtered.png") + + plt.close("all") + + print("\nDone.") + + +if __name__ == "__main__": + main()