{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Saving and Reloading\n",
    "\n",
    "In this example, we show how to save and reload results with `alabi`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-02-21T08:26:59.243640Z",
     "iopub.status.busy": "2026-02-21T08:26:59.243533Z",
     "iopub.status.idle": "2026-02-21T08:27:00.227749Z",
     "shell.execute_reply": "2026-02-21T08:27:00.227186Z"
    }
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import alabi\n",
    "from alabi.core import SurrogateModel\n",
    "import alabi.utility as ut\n",
    "\n",
    "np.random.seed(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-02-21T08:27:00.230124Z",
     "iopub.status.busy": "2026-02-21T08:27:00.229817Z",
     "iopub.status.idle": "2026-02-21T08:27:00.232954Z",
     "shell.execute_reply": "2026-02-21T08:27:00.232301Z"
    }
   },
   "outputs": [],
   "source": [
    "def test1d_fn(theta):\n",
    "    theta = np.asarray(theta)\n",
    "    return -np.sin(3*theta) - theta**2 + 0.7*theta\n",
    "\n",
    "# domain of the function\n",
    "bounds = [(-2,3)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If `cache` is set to `True` when initializing the surrogate model (the default behavior), `alabi` will automatically save your results at intermediate points:\n",
    "- after initial training\n",
    "- after each hyperparameter optimization\n",
    "- after running MCMC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-02-21T08:27:00.234266Z",
     "iopub.status.busy": "2026-02-21T08:27:00.234116Z",
     "iopub.status.idle": "2026-02-21T08:27:02.100628Z",
     "shell.execute_reply": "2026-02-21T08:27:02.099976Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Initialized GP with squared exponential kernel.\n",
      "Successfully initialized GP on attempt 1\n",
      "\n",
      "Optimizing GP hyperparameters using 5-fold cross-validation...\n",
      "Evaluating 100 hyperparameter candidates using 5-fold CV...\n",
      "Running 20 active learning iterations using bape...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 45%|████▌     | 9/20 [00:00<00:00, 88.26it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Optimizing GP hyperparameters using 5-fold cross-validation...\n",
      "Evaluating 100 hyperparameter candidates using 5-fold CV...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 90%|█████████ | 18/20 [00:00<00:00, 25.78it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train MSE: 1.3201468829865077e-09\n",
      "Test MSE: 0.0005274731839308958\n",
      "\n",
      "Optimizing GP hyperparameters using 5-fold cross-validation...\n",
      "Evaluating 100 hyperparameter candidates using 5-fold CV...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 20/20 [00:01<00:00, 18.10it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train MSE: 7.161131001054863e-13\n",
      "Test MSE: 2.2569896946485492e-06\n",
      "Caching model to results/test1d/surrogate_model...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "sm = SurrogateModel(lnlike_fn=test1d_fn, \n",
    "                    bounds=bounds, \n",
    "                    savedir=f\"results/test1d\",\n",
    "                    cache=True,\n",
    "                    random_state=7)                 # enable caching of results\n",
    "\n",
    "sm.init_samples(ntrain=10, ntest=100, sampler=\"lhs\")\n",
    "sm.init_gp(kernel=\"ExpSquaredKernel\", fit_amp=True, fit_mean=True, white_noise=-12)\n",
    "sm.active_train(niter=20, algorithm=\"bape\", gp_opt_freq=10, obj_opt_method=\"nelder-mead\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "At this point you should find several files saved in your specified save directory (`savedir`)\n",
    "\n",
    "```\n",
    "results/test1d/\n",
    "    initial_training_sample.npz  \n",
    "    surrogate_model.txt\n",
    "    surrogate_model.pkl\n",
    "```\n",
    "\n",
    "- `initial_training_sample.npz` contains the initial training samples saved to a numpy file (which is useful if you run the initial samples in parallel, but switch to sequential sampling when running the active learning chain)\n",
    "\n",
    "- `surrogate_model.txt` contains a summary of `alabi`'s configuration in a text file\n",
    "\n",
    "- `surrogate_model.pkl` contains the saved `SurrogateModel` object which can be reloaded and rerun"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Results file after GP training\n",
    "\n",
    "**surrogate_model.txt**\n",
    "\n",
    "```\n",
    "==================================================================\n",
    "GP summary \n",
    "==================================================================\n",
    "\n",
    "Configuration: \n",
    "-------------- \n",
    "Kernel: ExpSquaredKernel \n",
    "Function bounds: [[-2  3]] \n",
    "fit mean: True \n",
    "fit amplitude: True \n",
    "fit white_noise: True \n",
    "GP white noise: -12 \n",
    "Hyperparameter bounds: [[ -3.63198075   1.15217628]\n",
    " [-15.          -9.        ]\n",
    " [  0.1         10.        ]\n",
    " [ -1.           1.        ]] \n",
    "Active learning algorithm : bape \n",
    "\n",
    "Number of total training samples: 30 \n",
    "Number of initial training samples: 10 \n",
    "Number of active training samples: 20 \n",
    "Number of test samples: 100 \n",
    "\n",
    "Results: \n",
    "-------- \n",
    "GP final hyperparameters: \n",
    "   [mean:value] \t0.22212174710423627 \n",
    "   [white_noise:value] \t-13.414108468805813 \n",
    "   [kernel:k1:log_constant] \t1.539832622117638 \n",
    "   [kernel:k2:metric:log_M_0_0] \t-0.779896563686618 \n",
    "\n",
    "Active learning train runtime (s): 1.0 \n",
    "\n",
    "Final test error (MSE): 4.4642838595549054e-06 \n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Reload the surrogate model function from pickle file\n",
    "\n",
    "Load using the save directory you saved the results to:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-02-21T08:27:02.102074Z",
     "iopub.status.busy": "2026-02-21T08:27:02.101917Z",
     "iopub.status.idle": "2026-02-21T08:27:02.105452Z",
     "shell.execute_reply": "2026-02-21T08:27:02.104779Z"
    }
   },
   "outputs": [],
   "source": [
    "sm = alabi.load_model_cache(\"results/test1d\")  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-02-21T08:27:02.106902Z",
     "iopub.status.busy": "2026-02-21T08:27:02.106775Z",
     "iopub.status.idle": "2026-02-21T08:27:17.009343Z",
     "shell.execute_reply": "2026-02-21T08:27:17.008878Z"
    }
   },
   "outputs": [],
   "source": [
    "sm.run_dynesty(like_fn=sm.surrogate_log_likelihood)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After running an MCMC sampler (with `emcee` or `dynesty`), `alabi` will update the `surrogate_model.txt` file \n",
    "\n",
    "```\n",
    "results/test1d/\n",
    "    initial_training_sample.npz  \n",
    "    surrogate_model.txt\n",
    "    surrogate_model.pkl\n",
    "    dynesty_samples_final_surrogate_iter_20.npz\n",
    "```\n",
    "\n",
    "- `dynesty_samples_final_surrogate_iter_20.npz` contains the final posterior samples computed from dynesty after n iterations"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**surrogate_model.txt** \n",
    "\n",
    "```\n",
    "==================================================================\n",
    "GP summary \n",
    "==================================================================\n",
    "\n",
    "Configuration: \n",
    "-------------- \n",
    "Kernel: ExpSquaredKernel \n",
    "Function bounds: [[-2  3]] \n",
    "fit mean: True \n",
    "fit amplitude: True \n",
    "fit white_noise: True \n",
    "GP white noise: -12 \n",
    "Hyperparameter bounds: [[ -3.63198075   1.15217628]\n",
    " [-15.          -9.        ]\n",
    " [  0.1         10.        ]\n",
    " [ -1.           1.        ]] \n",
    "Active learning algorithm : bape \n",
    "\n",
    "Number of total training samples: 30 \n",
    "Number of initial training samples: 10 \n",
    "Number of active training samples: 20 \n",
    "Number of test samples: 100 \n",
    "\n",
    "Results: \n",
    "-------- \n",
    "GP final hyperparameters: \n",
    "   [mean:value] \t-0.1412005223011758 \n",
    "   [white_noise:value] \t-12.0 \n",
    "   [kernel:k1:log_constant] \t1.7443253209346983 \n",
    "   [kernel:k2:metric:log_M_0_0] \t-0.08685227966604564 \n",
    "\n",
    "Active learning train runtime (s): 1.0 \n",
    "\n",
    "Final test error (MSE): 4.4642838595549054e-06 \n",
    "\n",
    "==================================================================\n",
    "dynesty summary \n",
    "==================================================================\n",
    "\n",
    "Configuration: \n",
    "-------------- \n",
    "Results: \n",
    "-------- \n",
    "Total weighted samples: 10362 \n",
    "\n",
    "Dynesty runtime (s): 24.0 \n",
    "\n",
    "Summary statistics: \n",
    "theta_{i} = 0.2781219235609284 +/- 0.8305605738406258 \n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Saving Plots\n",
    "\n",
    "Any plots generated from `sm.plot()` will also be saved to the same savedir\n",
    "\n",
    "```\n",
    "results/test1d/\n",
    "    initial_training_sample.npz  \n",
    "    surrogate_model.txt\n",
    "    surrogate_model.pkl\n",
    "    dynesty_samples_final_surrogate_iter_20.npz\n",
    "    gp_mse_vs_iteration.png\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-02-21T08:27:17.010951Z",
     "iopub.status.busy": "2026-02-21T08:27:17.010816Z",
     "iopub.status.idle": "2026-02-21T08:27:17.551053Z",
     "shell.execute_reply": "2026-02-21T08:27:17.550550Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plotting the gp mean squared error with 100 test samples...\n",
      "Saving to results/test1d/gp_mse_vs_iteration.png\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnYAAAHWCAYAAAD6oMSKAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjgsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvwVt1zgAAAAlwSFlzAAAPYQAAD2EBqD+naQAAoM5JREFUeJzs3XdYU2f7B/BvQkjYAVRQRBli3bPiKKg4aq22WnG0bq2tHa4qtdXX1+1bR6ttbanV1tGlFi22+tO27jrqApx1KwiCooiEJSPk/v0Rc0rIJATCuD/XlYtwzjPuEwK5ec45zyMiIgJjjDHGGKvyxLYOgDHGGGOMWQcndowxxhhj1QQndowxxhhj1QQndowxxhhj1QQndowxxhhj1QQndowxxhhj1QQndowxxhhj1QQndowxxhhj1QQndowxxhhj1QQndowxVkJCQgJEIhFEIhHGjRtn63BqnMOHDwuv/4IFC2wdDmNVCid2rFxp/jiX9pGRkWHr0EuloKAAUVFRGD58OJ555hnI5XJIJBK4uroiKCgIL7zwAj788EP89ttvyMzMtHW4zArGjRsnvF83bdpksvyxY8fg7u4u1HnttddQWFhY/oEyVk1kZGRgwYIFWLBgAX799Vdbh1NpSWwdAGNV3enTpzFmzBhcu3ZNZ192djays7Nx69Yt7N27FwDg5eWF1NTUig6T2dCePXswZMgQPHnyBADw9ttvIzIyEmIx/2/NmLkyMjKwcOFCAMDYsWPxyiuv2DagSooTO1ZhduzYYXZZZ2fncozEemJjY9GrVy9kZ2cDAOrVq4fBgwejdevW8PDwwJMnT3D37l3ExsbiwIEDyMjIQFFRkY2jZhVp69atGDNmjDA6N3v2bHz00Uc2jooxVl1xYscqTHX872rixIlCUjd27Fh8/fXXcHBw0FtWqVRi//79iIqKqsgQmQ19/fXXmDRpElQqFQDg448/xvvvv2/jqBhj1RkndoxZ6PLly4iLiwMANGjQAN988w3s7e0NlpdIJOjbty/69u1bUSEyG1q6dCn+85//AADs7Oywbt06vP766zaOijFW3fEFHqzSe/z4Mfz8/CASiSAWi4Vr1QyZOHGicIG6vg9Szb6wsDCh/f/9739o3749PD094ezsjObNm2PmzJm4f/++wX6uXr0qPO/SpYvRpK60/vzzTwwaNAj16tWDg4MDGjZsiPDwcOHYTd01WNq7Ov39/SESieDv72+wzN27d/HVV1/htddeQ/PmzeHq6gp7e3vUrl0bnTp1wuzZs5GUlFTqvvLz8xEZGYmwsDDUq1cPdnZ2BuM4cuQIJk6ciGbNmsHd3R0ODg5o0KABBg8ejF9++QVEZLJ/ADh16hRGjhyJBg0awMHBAfXr10ffvn2tNpo6c+ZMIamTSqX4+eefzUrqFAoFVq5cid69e8PHxwcymQyenp549tlnMXv2bCQnJxutX/yGjoSEBADAr7/+ivDwcPj5+UEmk2ntW7BggVD+8OHDANSXF4wfPx6BgYFwcHBArVq10KNHD2zatEkYeTSlqKgIP/30E4YOHQp/f384OzvDxcUFTZo0wZtvvomYmBiz2rGW3bt3Y/jw4QgKCoKzszNkMhnq1auHVq1aYeDAgfjkk09w9+5dnXqbNm0y+wYZc37nSvveDwsLE9oEAJVKhe+//x59+/aFr68v7O3thX3FFRUV4bvvvsOAAQPg6+sLBwcHeHh4oHXr1pgxYwauX79u1usGALt27cKAAQO0/hYNGTIEBw4cAGDeHcxEhOPHj2PevHl4/vnnhZgcHR3h6+uLAQMGYMOGDSgoKNBbX/PaBgQECNu+++47vTfead7HJeXl5WHt2rV46aWXhN97uVyOli1bYurUqaV6TSo9YqwcARAeZXH06FGys7MjAOTt7U3379/XWy4qKkror0mTJpSdnW0wpu7du9PFixfJz89PK87iD3d3d/rjjz/09vXzzz9rtWUNRUVF9MYbbxiMBwBNmzaNDh06JHw/f/58nXbi4+OF/WPHjjXZr+Y18PPz07v/0KFDJBKJjMYFgKRSKX377bdm9xUfH08tW7bUaadkHI8fP6aXXnrJZP/dunWjhw8fGu1//vz5JBaLDbYxZMgQun79usnXb+zYsUKZjRs3EhGRUqmkCRMmCNudnZ1p3759RuPRiIqKIk9PT6PH5+DgQJs2bTLYRvGYrl27RoMHD9bbTnx8vPBaaLYdOnSIli1bJvye6Xu88sorVFhYaPQ4Ll68SE2bNjX5s5o8eTIplUq9bZh6f5srNzeXXn75ZZOxAKBJkybp1N+4caPOz9gQc37nSvve7969u7A9PT2dunXrpjf24m7evEktWrQweqwSiYQWL15s9HgKCwtp5MiRRtuJiIgw62c1fvx4s34GTZs2pevXrxt9bU09Dh06pFP/8OHDVL9+faP17Ozs6KOPPjL6mlQVfCqWVQmhoaGYP38+5s2bh9TUVIwdOxa///671n+rCQkJePPNNwEAMpkMW7duNXoThkKhwMCBA3Hnzh1069YNQ4YMgbe3NxITE/HTTz/h3LlzyMjIwCuvvIIjR44gODhYq35QUJDw/O+//8bp06fRsWPHMh3n9OnT8e233wJQn74bOXIkwsLCIJPJcO7cOaxfvx6ff/65WSNj1pSXlwciQpMmTdCjRw80b94ctWvXhkQiwf3793HkyBH8+uuvKCgowJtvvglvb2+89NJLRtvMz89HeHg4Ll26hM6dO2PIkCHw9fVFeno6/vnnH6FcZmYmQkJCcPnyZQBA48aNMXToUDRr1gxSqRS3b9/Gli1bcOHCBRw5cgS9e/fGyZMn9V7r+Omnnwp31QHAoEGD8OKLL8LV1RVXrlzBhg0bsH37drNHpoorKCjAyJEjsX37dgCAp6cn9uzZg06dOpms+8033+Ctt94CEUEqlWLgwIHo1q0bvL29kZ2djWPHjmHz5s3Iy8vDuHHjIJVKMXz4cKNtvvfee/j999/h5+eHMWPGoGnTpsjLy8Pp06chk8n0xrB582bUqVMH48aNQ+vWrSEWi/H333/j22+/RX5+Pn799VesWLFCGI0s6ezZs+jevTuysrIAAF27dkX//v3h5+cHlUqFCxcuYNOmTUhNTcWXX36JgoICrF271uTrY6k5c+Zg165dAIA6derg1VdfRYsWLVCrVi3k5eUhPj4ep0+fxqFDh8otBn3Mfe8XN3LkSBw5cgQtWrTA8OHD0ahRI2RlZeGvv/4SyqSkpCAkJES4697Pzw/jxo1D06ZNkZ2djT///BO//PILlEol5s6di/z8fCxevFhvf2+//TZ++uknAOrLSEaNGoXu3btDJpPhwoULWL9+PVauXGnW36Lc3FxIpVKEhoaiU6dOCAoKgpubG/Lz83Hz5k1ER0fjwoULuHr1Kl588UXExcXBzc1NqO/l5YUdO3bgwYMHeOuttwAAPXr0wNSpU3X6atmypdb3v//+OwYOHIjCwkKIxWL07dsXvXv3Rv369ZGXl4eYmBh8//33UCgUwvt69uzZJo+pUrN1ZsmqNxj4r9ISRUVFFBYWJrS3YsUKYV9hYSF17txZ2Ld69WqzYgJAy5cv1ymjVCpp8uTJQpnmzZtTUVGRVhmVSkXNmjUTyri4uNDMmTPpxIkTVFBQUOrjO3bsmDAq5uzsTEePHtUpk5KSojMaUhEjdgkJCXTu3DmjbZw9e5a8vLwIADVu3JhUKpXRvjSPZcuWGW33tddeE8ouWLBA7yhPUVERRURECOXmzJmjU+bWrVvk4OAg/He+bds2nTIKhYK6du2qFZ85I3Zffvkl9enTR/jex8eHLl26ZPS4NM6fP09SqVR43a5cuaK33OXLl8nHx4cAkKurKz169MhoTHg6wvbkyRODfRcfsQPUI88ZGRk65Q4fPiyM5NWuXZvy8/N1yuTk5FBgYCABICcnJ9q5c6fePjMyMqhHjx5Cn/pGNK0xYqdUKkkulxMAatSoEaWnpxssq1AoKC4uTmd7eY3YmfveLz5ih6ejioZGOYmI+vXrJ5Tt168f5eTk6JTZs2cPyWQyAkBisZhOnDihU+bgwYNCO25ubnTy5EmdMg8fPqQ2bdqY/FtERPTXX38Zff1VKhUtXbpUaMfQaGJp/66lpKQIo+BeXl56j5WI6O7du8LoqZ2dncHfwaqCEztWrkomUeY8jP3C3r17l2rVqkUAyN7enk6fPk1ERLNmzRLqv/zyy2bHFB4ebrBcUVERdejQQSj722+/6ZQ5ceIEOTk56RyDVCqlDh060FtvvUXfffcd3b171+RrNWjQIK1EwZDz589rnS6riMTOXN9++63Q77Fjx4z2BYAGDhxotL3z588LZSdMmGCy/5CQEAJAcrmc8vLytPZNnz5daOv999832EZKSgq5urqWKrFzcXERnjdq1Ihu375tMlYNzc/dwcGBbty4YbTs3r17jSYFxWOqX78+ZWVlGW2veGLn6elJaWlpBsu++uqrRn+2n3/+ubD/hx9+MNpvWloaubm5EQDq27evzn5rJHb37t0T2pg5c6ZFbZRnYmfqvU+kndi1b99e55/L4i5cuCCUrVevHikUCoNlly1bppX8l1T89PXatWsNtnPlyhWSSCRl/llphIaGEgAKCgrSu7+0f9eK/84fOXLEaNkrV64If1fffvttS8KvNPjmCVal1K9fHxs3bgQAFBYWYvjw4YiOjsby5ct19pvjgw8+MLhPLBYjIiJC+F5ziq24zp074/Tp0+jRo4fW9oKCAsTExGDt2rUYO3YsGjZsiOeffx7Hjx/X21d+fj52794NAJDL5XjjjTcMxtW6dWv06dPH6HHZSmhoqPD85MmTJsvrO5VS3HfffSc8N/az0hgzZgwA9Wn2U6dOae2Ljo4GoP65Tp8+3WAb9erVw6hRo0z2VZxmyhsAGDVqlNZF3sZkZGTgt99+A6A+LVz89L4+zz//POrVqwdAfYONMa+//jpcXFzMigNQv3a1atUy2rfGpUuXdPZrflb169fHiBEjjPZVq1Yt9O/fH4D64vv8/Hyz4zSXk5OT8Fxz93plYuq9X9KkSZOMTmiteX8D6tOoxU9lljR58mS4uroCUE+enZeXJ+zLy8vDH3/8AQBwd3c3evNV06ZN8eKLL5p7CCZp/n7cvHkTjx49KlNbRITvv/8egPrmtq5duxot37RpU+FSGlO/W5UdX2PHKoy5ExQ3bNjQ6P6XX34ZU6dOxerVq3Hr1i0MHjwYgPoD+8cffzT64VScm5ubyWvievfuLTw/ffq03jItWrTAwYMH8c8//+CXX37B0aNHcebMGSgUCqGMSqXC/v37ceDAASxatAj//e9/tdo4f/68cEdYSEiI3mugiuvVqxd+//13o2XKw7lz5/Djjz/ixIkTuHHjBjIzMw1+KOu7y7A4Ozs7PPfcc0bLHDlyBADg4OCAy5cvC9fZGVL8rtHLly+jW7duAIAHDx7gzp07ANR/wH18fIy206tXL6xZs8ZomeJatWqFixcvAgAWLlwIuVxuNHnUOH78uHA9n0wmM2uZJFdXV9y7d8/ka2Hqg6ykLl26GN3v6+srPH/8+LHWvszMTJw7dw6AOjHeuXOnyf407xvNtW5NmzYtVbymuLm5oXPnzjh58iQOHDiAAQMGYPLkyQgLC4NUKrVqX6Vlznu/JFM/z+L/yJj6x8/Z2RmhoaH4/fffUVBQgLNnzwo///PnzwuTaYeEhJh8rXr06CFcx2iMUqlEdHQ0fv31V5w7dw4pKSnIysoyeD3r3bt3zf5brs/ly5eF5NDDw8Os3y07OzsAQHx8PPLy8gzOSVrZcWLHKow1JyhesWIFjhw5InyYAOoLpTVTmJijUaNGeqcKKK527dpwd3dHRkYGUlJSjJZt0aIFWrRoAUD93+Lt27dx8uRJ7NmzB9u3b0dBQQGICHPnzkVgYKDWqEbxtk2N2phbxpqUSiUmTZqEb775xuwpRUytiVurVi2Tfzg103Lk5eVh0KBBZvWrkZ6eLjwv79d3xowZuHHjhrCixIwZM6BSqbRGfPXRHB+gnlrDnDVnNYofnz7FEzFz1K5d2+j+4v9sFB/hAYCkpCThAzomJqZMPytrioyMRM+ePaFQKLBr1y7s2rULjo6OCA4OxnPPPYeePXuiR48ekEgq9qPQnPd+SaZ+nvfu3ROeP/PMMybbe+aZZ4R/DovXLf670qhRI5PtBAYGmixz7do1hIeHm/xnpLiyrqld/Hdrz5492LNnT6nqp6enm/wHsLLixI5VSTKZDP7+/lqJ3dChQ0vVhrnLljk7OyMjI0PrdJspIpEIjRo1QqNGjTBy5EgsWbIEffv2FeZKmj9/vlZiV7zt4qeQyhq7tUybNg3r1q0DANjb26Nv377o2LEjfH194ezsLMzhV/yuNVNLpzk6OprsNyMjw+KYi8+JVRGv7//+9z+IRCL873//AwC8//77UKlUmDlzpsE6ZTk+zaiKIea8vsWVZd3ashwHAIPzl5VV+/btcf78eSxcuBBRUVHIycnBkydPcOTIERw5cgTLli2Dt7c3Zs2ahalTp1bY2r2l/dmYU0dzJzJg3vu3+Gn64nVzcnKE59b4XVEoFOjZs6eQMPr4+KB///5o1qwZvL294eDgILzuW7duxc8//wzA9N8PUyrre7IicGLHqqSvv/5aZ2h91KhRBqdy0Kf4HzBzypXmeqWSAgICsGnTJuH0y82bN5GQkCBMRlq87dzcXLNjshZjf0STkpLw9ddfA1BfP3Xo0CE0btxYb1lDUzVYysXFBRkZGfD09CzTNTcV9fouWbIEYrFYmELigw8+gEqlwocffmgyrtWrV2PKlCkW9WtrxY8jPDwcv/zyiw2j0ebn54cNGzZgzZo1OHXqFE6cOIFjx47h8OHDyM7ORmpqKqZPn47z58+X6vrckmy9BrTmmjlA/f419Xew+D87xesWT9Ss8bvy5ZdfCkndyJEjsWHDBoOndw1dg2yJ4u/JGTNmYOXKlVZru7LjmydYlXPp0iXMmDEDgPqX94UXXgAAXLhwoVTrcN66dcvkacVHjx4J//mVdVi+c+fOWn9sip/+qF+/vvD85s2bJtsyVab4H3VT/3kSkdFTYfv37xdOs82aNctgUgeor02xJs3pp9KOmJZU/GdnjdfXmEWLFmHevHnC97NmzcKyZcv0li1+eq2i5ya0puLv38p6HDKZDN26dcOHH36IXbt24eHDh1i7dq0w2rxp0ybExsbq1NEw9XuUlpZm/aBLQXNTDQDcuHHDZPniKy0U//0o/vzWrVsm27l9+7bR/ZrVciQSCb744guj1+xZ8+9HdfndsgQndqxKefLkCV577TU8efIEAPDVV1/h559/Fka+vvzyS7Mu5AXU13AYuiFCY//+/cJzcyaaNUYkEmldy1M8yWvdurXwIXL8+HGTdwlqlvMxxN3dXXhuahmqc+fOGf3PvPiyaqauPbP2DR3du3cHoL75pCx3qnl5eQnvkatXr5q8XtLU62vKwoULtZZXmj17tnD9XXFdu3YVrvPU3IlYFdWuXVu4vjQuLk6YILcyc3BwwMSJE/Huu+8K244ePapVxsPDQ3hu6vfo77//tm6ApVT875OpZRdzc3Nx7NgxAOol79q1ayfsa9OmjZDs/v333yYTWlOTO2v+ftSqVUvr9SwpLy/P4HJgGsVPlZv6p7xt27aQy+VCjOVx53VlxYkdq1KmT58unO4bPXo0Ro8eDblcji1btghJ0/jx401+cGt88sknBvepVCqsWrVK+H7IkCFa+zMyMkp1HcZff/0ljP45OjpqXZgsk8nQr18/AOprUjZs2GCwnUuXLpn8w+3o6Chc1Hz69GmjFyIXP0Z9ip+aMTaSdfv2ba3pSaxBM30JoB4JK3nRfmloLuhXqVT4/PPPDZZLTU0VZtwvi/nz52PRokXC93PmzMGSJUu0ynh5eQnTRVy8eBFbtmwpc7+2MnbsWADqU5LFRywru+JT0yiVSq19mmQV0P4nr6S8vLxS3UVdHjSzAwDAmjVrjP7OR0ZGCtfV9e/fX2tk0sHBQTgL8vjxY6M39Fy9etXkP3Oavx8PHjwwGtPnn39u8nKL4v8MmzoFrFm5B1CPppr6O1et2HAOPVYDoNikvWW1fft2oa2goCCdyVc/+ugjYX+PHj0MTuZZPCYAtHLlSp0yRUVFNG3aNKFMixYtdNr7/fffycfHh1asWEEpKSlGYz937hw1bNhQaG/UqFE6ZYqvPOHq6kp///23Tpn79+9T8+bNzZrtfdKkSUKZd999V2+ZTz/9VKstfRMU//3331r79U1ie+fOHZ24zFkv0xxDhw4V2uzTpw89ePDAYFmVSkXHjh2jiIgInX3FV56QSCQUHR2tUyYzM1Nntv/SrBWrz5IlS7TaW7Rokdb+s2fPCitPODk50ebNmw22RUT06NEjWrlypd4VG4rHpFkP1piSa8UaY2rS4JycHPL39xfKfPDBB0ZXYMnPz6eff/5Z72Tc1pigOC4ujhYsWGD0dzM7O5vat28v9KXvNS2+7mpUVJTO/ry8PBo2bFi5vPeLvxfN0b9/f6F8//799a488ccffwi/B2KxWO+qEgcOHBDaKevKExMnThT2T548WW/cmzdvJnt7e622DL0fi68mYmh1G427d++Su7u7cKyffvqp0Umes7Oz6ZtvvjH5O1jZ8c0TrMKYM4+QRqdOnbSuGUlMTBTWgZVKpdi6davOzQwffvghDhw4gAMHDuDQoUNYunQp5syZY7CPtm3bIjMzExEREdi5cyeGDBkCLy8vJCUl4aeffsLZs2cBqEfTNm7cqPeOuZSUFHzwwQeYNWsWOnXqhC5duuCZZ56Bp6cnlEolkpKS8Ndff+HPP/8ULq729fXFihUrdNoKCQnBlClTsHr1amRlZaFbt25a6zOeO3cO3377LdLT0xEeHq41Iak+06ZNw/r165GXl4evvvoK169fx9ChQ+Hh4YGkpCRs374dJ06cQPfu3XHz5k2Dp5q6dOmCTp064dSpU7hz5w6aNm2KiRMnolmzZigqKsLJkyfxww8/ICcnB+PGjSvVlB3mWL9+Pa5fv47z589j79698Pf3x+DBg9G5c2fUqVMHhYWFSE1NxYULF7B//37cvXsXjRo10hmNDQwMxEcffYQZM2ZAqVQiPDwc4eHhOmvFJiUlmfX6mmvOnDkQi8XCOpTz5s2DSqXC/PnzAajfh2vXrsWECROQm5uLESNGYMWKFXj55ZfRuHFjODo6QqFQ4ObNmzh9+jSOHDkCpVKJH374wSrxWYuTkxN27tyJbt26ISMjAytWrMCPP/6IIUOGoE2bNnBzc0Nubi6SkpIQFxeH/fv3IzMzExMmTCiXeBQKBRYsWIBFixbhueeew3PPPYcmTZrAzc0NGRkZuHr1KrZs2SKM7nfu3Bk9e/bUaefDDz8URo5HjBiB33//HT179oREIsGVK1fw/fffIyEhASNHjrTKSG9ZrFu3Du3bt0dqaip2796NFi1aYPz48WjSpAmys7Oxd+9ebNu2TTiNOWfOHL2XmPTs2RMTJkzA+vXrkZmZia5du2LUqFHo1q2b1lqxDx8+xLBhwxAVFQVA/53VkydPxoYNG6BUKvHll18iLi4OQ4YMQf369ZGamorffvsNBw4cgIuLCwYMGGDyxptevXohOjoat27dwrBhwxAeHg53d3fhkoaOHTvC09MTgPraz6ioKLz88svIz8/H9OnT8dVXX2HQoEFo3rw5XFxckJWVhfj4eMTExODgwYPIy8szuH5ulWHrzJJVbygxOmbuY8eOHUIbSqVSWCoKAH3yyScG+0tJSaE6deoIozL6Rr007XTv3p0uXryoNcpQ8iGXy+mPP/7Q21dsbKywdqe5j549e9KdO3cMxl9UVEQTJkww2sa0adPMHtH44YcftJYfK/no1q0bPXr0yORIQnx8PAUEBBiNa8qUKXT79m2rj1oQEWVlZdHo0aOFEU1Tj+7duxtsa968eUbbGTp0KF2/ft1qI3YaxZdx0vdz++OPP8x+P8lkMvr999+NxlTRI3YaN2/epE6dOpl1HCKRiObNm2dxX8YcPnzY7N/Lbt26GR0Jfvvtt43Wnzx5crm890s7YkdEdOPGDZ3R85IPiUSiM3JcUmFhIY0YMcJoOxEREbRv3z7h+1WrVulta/369VpLj5V81KpVi/7880+z3o/nz5/Xu4yjsXpnzpyhJk2amPVesLOzo2+++cbs17sy4sSOlStz/7CWfBRP7P773/8K2/v27Wty+H337t3CB7e/v7/OouYlP/wfP35Mixcvpnbt2pG7uzs5OjpSkyZNKCIiwuQpVpVKRadOnaLly5fToEGDqHnz5iSXy8nOzo4cHBzI29ubQkND6b333jO4dqo+v//+Ow0YMIC8vLxIKpWSr68vDRo0SEgyS/PBd+7cORo1ahQ1aNCApFIp1a5dm7p160bffvutsKC4OR84GRkZtGDBAmrdujU5OTmRk5MTBQYG0qhRo4Q/pqVZL9OSdWn/+ecfmjlzJnXs2JHq1KlDEomEnJycyM/Pj/r06UMLFiygU6dOmWznxIkTNHz4cKpfvz5JpVKqV68evfDCC7R161azj6O0iR0R0YoVK7Te53PnztXan5eXRxs2bKAhQ4ZQQEAAubi4kEQiIQ8PD2rXrh2NGzeOvv/+e3r8+LHJmGyV2Gn8+eef9MYbb1Dz5s3J3d2d7OzsyNXVlZo2bUrh4eH0+eefG1xT1xqJHRHRxYsXadWqVTR06FBq3rw5ubm5kZ2dHTk7O9MzzzxDI0aMoJ07d5rV1q+//kp9+/al2rVrk1Qqpfr161N4eDjt37+fiMrnvW9JYkekTso2btxIL730Evn4+JBUKiW5XE4tW7ak9957j65du2Z2Wzt37qSXXnpJ62/R4MGDhePeunWrEOOmTZsMthMbG0sjR44kX19fsre3J09PT2rbti3NnTuXkpOTicj89+O1a9forbfeombNmpGzs7PWP2qG6hUVFdG2bdto1KhR1LhxY+G9oHldhg8fTmvXrqV79+6Z/dpUViIiM6eRZ6ya0AzZd+/e3eRdWJXV4cOHhfVp58+fr3UHJmOMVZSIiAjhxoSzZ8+ibdu2tg2I8V2xjDHGGCs9hUIhXOtZp04dtGzZ0sYRMYATO8YYY4yVEB8fjzt37hjcn5GRgaFDh+Lhw4cAgDfeeKPC19xl+vFPgTHGGGNaYmNj8dprryEkJATdu3dHUFAQnJ2d8fjxY8TGxmLr1q3CvJxBQUFGZyBgFYsTO8YYY4zpKCoqwpEjR3DkyBGDZdq1a4fffvtNayJzZluc2DHGGGNMywsvvID169dj3759+Oeff5CWlob09HTY2dnBy8sLwcHBGDx4MIYOHap3/jpmO3xXLGOMMcZYNcEjdjWYSqVCSkoKXF1dhSlAGGOMMVY+iAhZWVnw8fEpt5FOTuxqsJSUFDRo0MDWYTDGGGM1SlJSEnx9fculbU7sajBXV1cA6jeYm5ubjaNhjDHGqrfMzEw0aNBA+PwtD5zY1WCa069ubm6c2DHGGGMVpDwvf+JbWRhjjDHGqglO7BhjjDHGqglO7BhjjDHGqokql9ht27YNYWFh8PDwgLOzM9q0aYMVK1agsLDQovZiY2MxdOhQeHt7w8HBAQEBAZgyZQoePHhgtF5qaiomT56MgIAAyGQyeHt7Y+jQoYiLizNY58iRI/joo48wePBg+Pv7QyQSQSQS4dixYybjVKlUWLt2LTp16gRXV1e4urqiU6dOWLduHXgqQsYYY4wBVWyC4vfeew+ff/45JBIJevbsCRcXFxw8eBAZGRkIDQ3F3r174ejoaHZ727dvx/Dhw6FUKhEcHIyAgADExMTg9u3b8Pb2xrFjxxAUFKRT7/r16+jatSsePHiAwMBAdOjQAfHx8Thz5gwkEgmioqIwaNAgnXpt27bF+fPndbYfPXoUoaGhBuMsKirCsGHDEB0dDScnJ/Tq1QsAsH//fjx58gRDhw7F1q1bSz0nTmZmJuRyORQKBd88wRhjjJWzCvncpSpix44dBIBcXFwoNjZW2P7w4UNq1aoVAaCIiAiz20tOTiYnJycCQGvXrhW2K5VKGjVqFAGg4OBgUqlUWvVUKhW1a9eOANDo0aNJqVQK+9auXSvEeO/ePZ0+33//fVqwYAHt3LmT7t69S35+fgSAjh49ajTWTz/9lABQ/fr16fbt28L227dvk4+PDwGgL774wuxj11AoFASAFApFqesyxhhjrHQq4nO3yozYdezYEWfOnMGSJUswZ84crX3Hjh1D165dIZPJkJqaCrlcbrK9Dz74AB9//DF69+6Nffv2ae3Lzs6Gr68vFAoF/vjjD7zwwgvCvj179qB///5wd3dHUlISXFxctOr27t0bBw4cwKxZs7B06VKjMfj7++POnTtGR+xUKhXq16+P+/fv48cff8TIkSO19v/4448YPXo0fHx8kJSUVKpROx6xY4xVRYWFhSgqKrJ1GIzBzs4O9vb2ZpeviM/dKjGPXXJyMs6cOQMAGDFihM7+0NBQNGjQAElJSdizZw+GDx9uss0dO3YYbM/FxQUDBgzADz/8gOjoaK3ETlNvwIABOkmdpr0DBw4gOjraZGJnjhMnTuD+/fuQyWQYPHiwzv7BgwdjwoQJSElJwalTp9ClS5cy98kYY5VRZmYm0tLSkJ+fb+tQGBPIZDLUrl270gyQVInE7uzZswAAT09PBAQE6C3ToUMHJCUl4ezZsyYTu6ysLNy8eVOoZ6i9H374Qei7ZCzG6gHAjRs3kJOTA2dnZ6OxmKLpr0WLFnBwcNDZ7+joiBYtWuDs2bM4e/YsJ3aMsWopMzMTycnJcHFxQe3atWFvb89rXDObIiIUFhZCoVAgOTkZACpFclclErv4+HgAQMOGDQ2W0ax5qilrTEJCgvDcUJuG2jMVi6YeESEhIQEtWrQwGY8x5h772bNnzTp2xhiritLS0uDi4gJfX19O6Fil4ejoCFdXV9y9exdpaWmc2JkrKysLAIyOfmlOi2ZmZprdnrE2DbVnKpbip2fNicUUax57fn6+1ikMa8THGGPlrbCwEPn5+ahduzYndazSEYlEkMvlSE5ORmFhYamuuSsPVW4eO2a5pUuXQi6XCw/N6CJjjFVmmhslbP2ByZghmvdmZbipp0okdq6urgCAnJwcg2Wys7MBmHd+W9OesTYNtWcqFk09c2MxxZrHPnv2bCgUCuGRlJRU5vgMunMCOPsjcPMA8OAK8CQDqBo3YDPGKikerWOVVWV6b1aJU7H+/v4AYDQR0ezTlDXGz89PeJ6YmIhWrVqZ3Z6/vz/S09ORmJhoNA6RSKTVj6U0/Rvqz1isJclkMshksjLHZJaLUUDMBu1t9k6Aa13A1Qdwq1fiuc/T7+sBEmnFxMgYY4xVM1UisWvXrh0A4NGjR4iPj9d7Z2xMTAwAoH379ibbc3NzQ1BQEG7evImYmBi9iZ2h9tq3b4+4uDhhv6F6jRs31jsdSmlp+v/nn3+Ql5enc2fskydP8M8//+iN1aZqNwEa9QQy7wFZKUCeAijMBdJvqx/GONV+muw9fbj56CaBTp5AJfoPiTHGGKsMqsSpWF9fXwQHBwMANm/erLP/2LFjSEpKgkwmQ79+/cxqU7Pkl772srOzsWvXLgBAeHi43no7d+7Ue3pU017Jepbq0qUL6tati/z8fPzyyy86+3/55RcUFBTAx8cHnTp1skqfVtH5bWD0DmDSSWBWIvCfe8CUOGDcbmDweuD5xUDnd4HmrwANOgPuDQG7pyN1uWnA/YvAjb1A3HfA4aXArmnA5qHA16HAx4HAEi/gq+eAR7dsepiMMVbdLFiwACKRCIcPH7Z1KMwCVSKxA4D//Oc/AIBly5YhLi5O2P7o0SO8++67AIDJkydrrTqxY8cONG3aVFhbtbj33nsPTk5O2L9/P7755hthe1FREd59911kZGQgODgYffr00ar34osvol27dsjIyMC7776rdaHkunXrcODAAbi4uGDatGlWOW6xWIwPP/wQAPDhhx9qTWkSHx+PWbNmAVBfP1fatWIrlNQJqNUI8A8FWg0BQqYCfZcCw74DJvwJvHcR+O8DYOZt4O1jwIhtwMufA2GzgfZjgcZ9AO9W6tE8ACgqAB78A1yIsu1xMcZYOTt8+DBEIhEWLFhg61BYVVBui5WVg6lTpxIAsre3p759+9LgwYPJ3d2dAFBISAjl5uZqld+4cSMBID8/P73tRUVFkZ2dHQGgTp060auvvkqBgYEEgLy9venGjRt66129epXq1KlDACgwMJBeffVV6tixIwEgiURC0dHReut988031KlTJ+EhlUoJADVv3lzY9sorr+jUUyqVNGjQIAJATk5ONGDAABowYICw1u2QIUOoqKiodC8mVeG1YgvziI59RjTfjei7gbaOhjFWzp48eUKXL1+mJ0+e2DoUmzh06BABoPnz51dIfw8fPqQrV65QTk5OhfRXHZj7Hq2Iz91KPMSj6/PPP8fPP/+MLl264O+//8aePXvg6+uLZcuW4eDBg3B0dCxVe0OHDsWpU6cQHh6O27dvY8eOHSgqKsKkSZNw/vx5BAUF6a3XpEkTXLhwAZMmTUJRURF27NiB+Ph4hIeH49SpU8Lp2pLu3r2LU6dOCY+CggIAwOXLl4VtJVe6ANRr0W3fvh1ff/01mjdvjgMHDuDAgQNo0aIFvv76a0RFRVXu0Tprk8iARk9HYe/GACrb317OGGPVRe3atdG0aVM4OTnZOhRmiXJLGVmlV2VH7IiIipREH/mqR+3uXbB1NIyxclSTR+zmz59PAPQ+4uPjiYho7NixBIBu3bpFn3zyCTVr1oykUimNHTuWiIiSk5Np3rx51KlTJ6pTpw5JpVLy8/Ojd955h1JTUw32eejQIWFbfHw8AaCxY8fSjRs36JVXXiF3d3dycnKiXr160blz50p1XPn5+bRy5Upq164dOTk5kYuLC4WGhtJvv/2mU9bU8fn5+ZGfnx89fvyYJk2aRL6+vmRnZ0cbN24U2ti5cyeFhYWRm5sbOTg4UOvWrWnlypVUWFio1Vfx47x8+TK98sor5OnpqfV661OZRuyqxF2xjOkQ2wG+HYBbB4HEk0Bd3TubGWOsqgsLC0NCQgK+++47dO/eHWFhYcI+d3d3rbJTpkzByZMn0b9/f7z88svw8vICABw5cgQrV65Er1690KlTJ9jb2+Ps2bNYs2YN/vzzT8TFxWldn25MQkICOnfujBYtWuD111/HrVu38Ntvv6FHjx64cuUKvL29TbaRn5+Pvn374vDhw2jbti0mTJiAwsJC7N69GwMHDsQXX3yByZMn69QzdHyaNnv27Ins7GwMGDAAEolEiGXVqlWIiIiAp6cnRowYAWdnZ+zcuRMRERE4evQooqOjdeahu3nzJjp37oxWrVph3LhxePToEaTSKjIVV7mljKzSq9IjdkREh5aqR+y2v2HrSBhj5cjYaIhKpaKc/MJK/1CpVBYfv6lr7DQjWr6+vnTnzh2d/ampqZSVlaWz/bvvviMAtGTJEq3txkbsANCyZcu0yv/3v/8lALR06VKzjuc///kPAaC5c+dqvS6ZmZnUoUMHkkqllJycbPbx+fn5EQB64YUXdK61v3nzJkkkEvLy8qLExERhe15eHoWGhhIA+v777/Ue57x588w6HiIesWPMOhp0VH9NOmnbOBhjNvOksAjN5/1p6zBMurzoBThJy/cjd+bMmWjYsKHO9uIjW8WNHj0aU6ZMwf79+zFnzhyz+ggICMDMmTO1tk2YMAFLlizBmTNnTNZXqVRYs2YNGjVqhIULF2qNlLm6umLevHkYMGAAoqOjdUbtDB2fxooVK3Sutd+8eTOUSiUiIiK0ltGUyWRYvnw5QkJCsGnTJowePVqrXt26dc1+TSobTuxY1VW/AyASAxmJ6omQ3erZOiLGGLOZjh07GtwXHR2NtWvXIi4uDo8fP9aaqislJcXsPtq2batzs56vry8AICMjw2T9a9eu4fHjx/Dx8cHChQt19j98+BAAcPXqVZ19xo7PwcFB72IDmhsSi5/C1ujSpQscHBxw7tw5nX1t2rSpOqdeS+DEjlVdDm6AVwsg9SKQdApo8YqtI2KMVTBHeztcXvSCrcMwydHertz7MHR928qVK/H++++jTp066NOnD3x9fYWRrc8++wz5+flm96FvTXKJRJ1KFE8WDUlPTwegXk1Js2qSPvoWADB2/Z6Xl5fe9VozMzMN1hWJRPD29kZycnKp+qrsOLFjVVvDTk8Tu9Oc2DFWA4lEonI/xVlV6EtslEolFi9ejHr16uHcuXNap2WJCCtWrKjIEIXEcPDgwdi+fXup6uo7PlP7NP2lpqbqrN9OREhNTdWbrBrrq7KrQZOfsWqpwdNl1Pg6O8ZYNWVnpx7tM2dErKS0tDQoFAp06dJF51q7mJgYPHnyxCoxmqtZs2Zwc3NDTEwMCgsLy70/zVrz+pZHO3XqFPLy8tC2bdtyj6MicWLHqjZNYnfvPFBYsX+gGGOsInh6egIAkpKSSl3Xy8sLjo6OiIuLQ25urrD98ePHmDJlitViNJdEIsE777yDO3fu4P3339eb3F26dAkPHjywSn8jRoyARCLBqlWrtK4lLCgoEJbrHDdunFX6qix4/LoGioyMRGRkpEX//VU67g0Bl7pA9n0g5Szg95ytI2KMMatq2rQpfHx8sHXrVshkMvj6+kIkEmHKlCkm558Ti8V49913sXLlSrRp0wYvv/wyMjMz8fvvv8PPzw8+Pj4VdBT/WrhwIeLi4rB69Wrs3r0b3bp1g5eXF5KTk3Hx4kWcP38eJ06cMHg3b2k0atQIy5cvR0REBFq3bo1hw4bB2dkZu3btwrVr1zBw4ECMGjXKCkdVeXBiVwNNmjQJkyZNQmZmptmTUlZaIpH6OrvLv6knKubEjjFWzdjZ2SE6OhoffvghtmzZgqysLADAqFGjzPobvnTpUnh6emLTpk346quv4O3tjeHDh2PBggVo2bJleYevQyaT4ffff8f69evx/fff45dffkF+fj68vb3RvHlzvP3223rvcLXUjBkzEBQUhFWrVuHHH39EQUEBnnnmGaxcuRJTp06t0tfT6SMiIrJ1EMw2NImdQqHQe/FolXEiEvjzP8AzLwIjtto6GsaYleXl5SE+Ph4BAQFwcHCwdTiM6TD3PVoRn7t8jR2r+hp0Vn9NOgXw/ymMMcZqME7sWNVXtxUgcQCepAOPbto6GsYYY8xmOLFjVZ9ECtR/Vv08kac9YYwxVnNxYseqB2Hd2FO2jYMxxhizIU7sWPUgTFTMiR1jjLGaixM7Vj34Ph2xS7sO5KbbNhbGGGPMRjixY9WDcy2gVmP186TTto2FMcYYsxFO7Fj10ZBPxzLGGKvZOLFj1QdfZ8cYY6yG48SOVR+aiYqTY4Ei3YWlGWOMseqOEztWfdQKAhw9AGUecO+CraNhjDHGKhwndqz6EIv5dCxjjLEajRM7Vr0IExXzChSMMcZqHk7saqDIyEg0b94cwcHBtg7F+jTX2SWdBohsGwtjjDFWwTixq4EmTZqEy5cv48yZM7YOxfp82gFiCZB1D1Ak2Toaxhgrs8OHD0MkEmHBggUV2u+CBQsgEolw+PDhCu2XlQ0ndqx6kToB9dqonyfydXaMMcZqFk7sWPXDN1AwxhiroTixY9UP30DBGKsmFixYgB49egAAFi5cCJFIJDwSEhKEcgUFBVi1ahXat28PZ2dnuLq6omvXrti5c6dOmwqFAvPmzUPz5s3h4uICNzc3BAUFYezYsbhz5w4AICwsDAsXLgQA9OjRQ+jT39/frLhLE8+4ceMgEolw+/ZtrFy5Es2bN4dMJsO4ceMAAP7+/vD390dGRgYmT56MBg0aQCKRYNOmTUIbu3btQo8ePSCXy+Ho6Ig2bdpg1apVUCqVWn0lJCRAJBJh3LhxuHLlCgYNGoRatWrpvJ5VmcTWATBmdZobKFL/AfKzAJmrbeNhjDELhYWFISEhAd999x26d++OsLAwYZ+7uzsAID8/H3379sXhw4fRtm1bTJgwAYWFhdi9ezcGDhyIL774ApMnTwYAEBFeeOEFnDp1CiEhIejbty/EYjHu3LmDnTt3YvTo0fDz8xOSqr/++gtjx44VEjpNn8aUJp7ipkyZgpMnT6J///54+eWX4eXlpdVmz549kZ2djQEDBkAikcDb2xsAsGrVKkRERMDT0xMjRoyAs7Mzdu7ciYiICBw9ehTR0dEQiURafd28eROdO3dGq1atMG7cODx69AhSqdTMn0olR6zGUigUBIAUCoWtQ7G+VS2J5rsR3Txo60gYY2X05MkTunz5Mj158kR3p0pFlJ9d+R8qlcXHf+jQIQJA8+fP17v/P//5DwGguXPnkqpYP5mZmdShQweSSqWUnJxMREQXLlwgAPTKK6/otJOXl0dZWVnC9/PnzycAdOjQoVLFW5p4iIjGjh1LAMjX15fu3Lmj056fnx8BoBdeeIFyc3O19t28eZMkEgl5eXlRYmKi1rGEhoYSAPr++++F7fHx8QSAANC8efNKdVzGGH2PFlMRn7s8Yseqp4adgIuJ6mlPGvWwdTSMsfJSmAt85GPrKEz7TwogdbZ6syqVCmvWrEGjRo2EU7Uarq6umDdvHgYMGIDo6GitUTJHR0edtmQyGWQymU3iAYCZM2eiYcOGBttesWKFTtybN2+GUqlEREQEGjRooHUsy5cvR0hICDZt2oTRo0dr1atbty7mzJlTlkOttDixY9VTg07AxW18nR1jrFq7du0aHj9+DB8fH+GauOIePnwIALh69SoAoFmzZmjdujW2bNmCu3fv4pVXXkFYWBjatm0Lsbjsl92XNp7iOnbsaLBdBwcHtGrVSmf72bNnAUDrFLVGly5d4ODggHPnzunsa9OmTfU59VqCxYndkSNHAACtW7c265w7YxVKc2fs3RhAVQSI7WwbD2OsfNg7qUfDKjt7p3JpNj09HQDwzz//4J9//jFYLicnBwAgkUhw8OBBLFiwAL/88gsiIiIAAHXq1MHkyZMxZ84c2NlZ/veytPEUp7lmTh8vLy+d6+QAIDMz02BdkUgEb29vJCcnl6qvqs7ixC4sLAx2dnZ48OCBNeNhzDq8mgNSFyA/E3hwBajb0tYRMcbKg0hULqc4qwo3NzcAwODBg7F9+3az6tSqVQtffPEFVq9ejatXr+LgwYP44osvMH/+fNjb22P27NkVGo+GvsTN1D5Nf6mpqfDz89PaR0RITU0VypjbV1Vn8birXC6HXC6Hh4eHNeNhzDrsJIBvB/Vzns+OMVaFaUbQioqKdPY1a9YMbm5uiImJQWFhYanaFYlEaNasGSZNmoR9+/YBgNZ0JMb6NaQs8ViiXbt2AKB3dYxTp04hLy8Pbdu2Lfc4KhOLE7ugoCBkZWUhPz/fmvEwZj08UTFjrBrw9PQEACQl6S6TKJFI8M477+DOnTt4//339SZTly5dEs6uJSQk6J2vLTU1FYD6WjZz+jWktPGU1YgRIyCRSLBq1SqkpPx7Sr6goAAffvghAAhTt9QUFp+Kfe211xAbG4uoqCidu00YqxQ4sWOMVQNNmzaFj48Ptm7dCplMBl9fX4hEIkyZMgVyuRwLFy5EXFwcVq9ejd27d6Nbt27w8vJCcnIyLl68iPPnz+PEiRPw8vLCuXPnEB4ejo4dO6J58+aoW7cukpOT8euvv0IsFmP69OlCv5qJif/zn//gn3/+gVwuh7u7u9456IorTTxl1ahRIyxfvhwRERFo3bo1hg0bBmdnZ+zatQvXrl3DwIEDMWrUqDL3U6VYOk9KYWEhPffcc+Tm5ka7d++24gwsrKJU63nsiIieZBDNl6vns8u8b+toGGMWMneOsOrs5MmT1L17d3J1dRXmYYuPjxf2K5VKWrt2LYWEhJCbmxvJZDJq2LAh9e3bl9asWUPZ2dlERJSUlESzZs2izp07k5eXF0mlUmrYsCGFh4fTiRMndPrdtGkTtWrVimQyGQEgPz8/s+I1Nx6if+exK348xfn5+Zns97fffhNeH5lMRq1ataKVK1dSYWGhVjnNPHZjx4416zjMVZnmsRMREVmSEC5atAhPnjxBZGQkcnJy0KJFC4SEhMDLy8voHTXz5s2zpDtWDjIzMyGXy6FQKPReXFotrAkBUi8Bw34Amg+wdTSMMQvk5eUhPj4eAQEBWqcKGasszH2PVsTnrsWnYhcsWACRSARNXnjp0iWjtzZrcGLHKlSDjurELukUJ3aMMcaqPYsTu27dulXr24Wrs8jISERGRpbqTqcqq0EnIGYDX2fHGGOsRrA4sdN3azGrGiZNmoRJkyYJQ8LVmuYGipRzQGEeYM+ncRhjjFVfZV8/hLHKzMMfcPYCVIVAyllbR8MYY4yVK07sWPUmEgENedoTxhhjNYPFp2KLKygowL59+xATEyNMOujl5YXg4GD07t272i60y6qIBp2AK7s4sWOMMVbtlTmxW7duHebOnYu0tDS9+2vXro0lS5bgzTffLGtXjFmmQWf116RTAJF6FI8xVuVYODsXY+WuMr03y3Qq9sMPP8Q777yDhw8fgojg4+ODjh07omPHjvDx8QER4eHDh3j77bcxa9Ysa8XMWOnUaw3YyYDcR8CjW7aOhjFWSpq5USti7VHGLKF5bxqbx7eiWJzY/fXXX/j4449BRBg8eDAuX76MpKQknDhxAidOnEBSUhKuXLmCIUOGgIjw8ccf4+jRo9aMnTHzSGRA/fbq53w6lrEqx97eHjKZDAqFolKNjDAGqEfrFAoFZDIZ7O3tbR2O5adiIyMjAQATJkzAN998o7dMkyZNEBUVhTfffBPr16/Hl19+ia5du1raJQBg27ZtiIyMxPnz51FQUICgoCCMHDkS06dPt+gFjY2NxbJly3DkyBEoFArUq1cPL730EubOnWt0HbvU1FQsXrwYu3fvRkpKCtzd3dGtWzfMnj0b7du3N1ivoKAAn376KTZv3oybN29CKpWiTZs2mDx5MoYMGWKwXmJiIlasWIF9+/YhMTERRIR69eqhW7dumDFjBtq0aVPqY69RGnQEEk8ASSeBdiNtHQ1jrJRq166N5ORk3L17F3K5HPb29jyXKrMpIkJhYSEUCgWys7NRv359W4cEALB4STFfX1/cv38fKSkpJhfyTU1NhY+PD+rVq4e7d+9aFCgAvPfee/j8888hkUjQs2dPuLi44ODBg8jIyEBoaCj27t0LR0dHs9vbvn07hg8fDqVSieDgYAQEBCAmJga3b9+Gt7c3jh07hqCgIJ16169fR9euXfHgwQMEBgaiQ4cOiI+Px5kzZyCRSBAVFYVBgwbp1MvNzcXzzz+Pv//+G+7u7ujZsyeys7Nx8OBBKJVKRERE4JNPPtGpd+rUKTz//PPIyspC/fr18eyzz8LOzg7nzp1DfHw8JBIJNm/ejKFDh5bq9awRS4ppXN0DbB0O1GkKTOJRO8aqoszMTKSlpSE/P9/WoTAmkMlkqF27tlmfoxXxuWtxYufg4ABnZ2c8evTIrPK1atVCTk4O8vLyLOkOv/76KwYNGgQXFxf89ddfwqhYWloaevbsiYsXLxpMjPRJSUlB48aNkZubi7Vr12LixIkAgKKiIowbNw4//vgjgoODcerUKa3/CokIzz77LM6ePYvRo0dj48aNwjn1devW4a233oKLiwtu3LiBunXravWpSUxbtWqFgwcPonbt2gDUo4ZhYWHIzs7Grl278NJLL2nVa9OmDS5cuICJEyfiyy+/FEYmVSoV5s+fjyVLlsDd3R337t0r1TqKNSqxy0kDPm6kfv5BPODkadt4GGMWKywsrBkr57BKz87OrlRnCyvkc5csVLt2bbK3t6cnT56YLJubm0v29vZUq1YtS7uj4OBgAkBLlizR2Xf06FECQDKZjDIyMsxqb+bMmQSAevfurbMvKyuL5HI5AaA//vhDa9/u3bsJALm7u1NWVpZO3V69ehEAmjVrltb29PR0kkqlBICOHTumU2/x4sUEgDp37qy1PS0tjQAQAHrw4IFOPaVSSY6OjgSA4uLizDp2DYVCQQBIoVCUql6Vtbo90Xw3omt/2joSxhhjNVBFfO5afPNE69atUVRUhA0bNpgsu2HDBiiVSouvA0tOTsaZM2cAACNGjNDZHxoaigYNGiA/Px979uwxq80dO3YYbM/FxQUDBqgXjI+OjtZbb8CAAXBxcdGpq2mvZL09e/agoKAADRs2REhIiMF6J0+eREpKirBdJpOZdTwAhBFAZoBmebGkk7aNgzHGGCsnFid2I0eOBBEhIiIC69evN1ju22+/RUREBEQiEUaPHm1RX2fPqpeC8vT0REBAgN4yHTp00CprTFZWFm7evKlVz9z2NN+bqnfjxg3k5OSYXS8wMBCenurTg+fOnRO2u7i4CDec/Pe//9W63V+lUmHBggV48uQJXnzxRTRo0EBv2+ypBh3VX5NO2zYOxhhjrJxYfFfsuHHj8MMPP+Cvv/7CxIkTsWjRIvTo0UO4K+Tu3bs4dOgQkpOTQUQICwvD2LFjLeorPj4eANCwYUODZTRJjaasMQkJCcJzQ20aas9ULJp6RISEhAS0aNHCrHqA+oaU9PR0nT6/+eYb9OvXD+vWrcPu3bvRoUMH2NnZ4ezZs0hOTsbo0aPx5ZdfGmyXPaWZqPhuDFBUCNjZ/rZ0xhhjzJosTuzEYjF+++03vP7664iOjkZSUhJ++OEHrTL09L6MwYMHY/369Rbfmp6VlQUAcHZ2NlhGc1o0MzPT7PaMtWmoPVOxFD89W7xuWY6hSZMmOHHiBEaPHo29e/ciOTlZ2Ne8eXOEhYWZdRFmfn6+1t1k5rxW1UrtZwAHOZCnAO5f/HduO8YYY6yaKNPKE25ubti+fTtOnjyJ6dOnIzQ0FM888wyeeeYZhIaGYvr06Th16hS2bdtW/e+6LEfHjx9Hq1atcOnSJWzevBn3799Heno6du3ahcLCQkyYMAETJkww2c7SpUshl8uFR407dSsWF7vOjqc8YYwxVv1YPGKXmJgIAPDy8hKWESsvrq6uAKB1zVpJ2dnZAGBWAqlpT9OmXC43uz1XV1ekp6cbjEVTr2RdS48hIyMDgwYNQlpaGk6cOIFOnToJ+1566SU0b94crVq1woYNGzBq1Cj06NHDYPuzZ8/GjBkzhO8zMzNrXnLXoCNwY686sev8jq2jYYwxxqzK4hE7f39/BAYGIj093ZrxGOwLAJKSkgyW0ezTlDXGz89PeK5JUM1tT/O9qXoikUirH1P1AAiTNxfvc/fu3Xj48CECAwO1kjqN4tv3799vsG1AfYetm5ub1qPG0VxnxzdQMMYYq4YsTuxcXFwgl8vh4+NjzXj0ateuHQDg0aNHBm+OiImJAQCjy3lpuLm5CStKaOqZ257me1P1GjdurHW9nal6t2/fFpJkzfEC/yaCxpIwzYhjRSTZVV799oDIDshMBjIM/6PAGGOMVUVlGrHLzc2tkNm/fX19ERwcDADYvHmzzv5jx44hKSkJMpkM/fr1M6tNzZJf+trTrAABAOHh4Xrr7dy5U+9pVU17Jev169cPUqkUiYmJOH78uMF6nTt31kqWNXcZX716FQqFQqdeYWEh4uLiAMDgVDCsGKkzUK+1+jlfZ8cYY6y6sXRm47lz55JYLKadO3daa7Jko3bs2EEAyMXFhWJjY4XtaWlp1KpVKwJAERERWnWio6OpSZMm1LNnT532kpOTycnJiQDQunXrhO1KpZJGjx5NACg4OJhUKpVWPZVKRe3atSMANGbMGFIqlcK+tWvXCjHeu3dPp89p06YRAGrdujWlpaUJ22NjY8nFxYUA0K5du7TqPHjwgJydnQkADR06VGu1i/z8fJo0aRIBIHt7e7p165apl1FLjVt5QmPPB+oVKHa/b+tIGGOM1SAV8blrcWKXnZ1NzzzzDPn4+ND58+etGZNBU6dOFZKYvn370uDBg8nd3Z0AUEhICOXm5mqV37hxIwEgPz8/ve1FRUWRnZ0dAaBOnTrRq6++SoGBgQSAvL296caNG3rrXb16lerUqUMAKDAwkF599VXq2LEjASCJRELR0dF66+Xk5FCXLl0IAHl4eNDgwYOpb9++ZG9vTwBoxowZeuv98MMPJJFICADVqVOH+vXrRwMHDqT69esTABKLxbRmzRrzX8inamxid/EXdWL3dVdbR8IYY6wGqYjPXRHR08nmSun7779HWloaFixYgLy8PPTt2xchISHw8vKCnZ2dwXpjxoyxpDtBVFQUIiMjce7cORQWFqJRo0YYNWoUpk+fDqlUqlV206ZNGD9+PPz8/LQmJS4uNjYWH330EY4ePQqFQoF69erhpZdewty5c+Ht7W0wjvv372PJkiX4v//7P9y7dw9yuRxdu3bFnDlzjF7nV1BQgFWrVuGnn37CrVu3IJVK0aZNG0yePBlDhw41WO/8+fP47LPPcOTIEWHS53r16iE0NBRTp0616K7kClmMuDJSJAOfNldfazcrEZDpLg3HGGOMWVtFfO5anNiJxWJhwmEiMmvyYZFIBKVSaUl3rBzU2MQOAD5tCSiSgDE7gcDuto6GMcZYDVARn7sWz2PXsGFDi1eSYMzmGnRUJ3ZJpzixY4wxVm1YnNgZOrXJWJXQoBNw6Re+M5Yxxli1YnFid+TIEQBA69at4e7ubq14GKsYwtJiZwCVSr3cGGOMMVbFWZzYhYWFwc7ODg8ePLBmPIxVDO+WgL0zkK8AHl4FvJvbOiLGGGOszCweptAsJO/h4WHNeBirGHYSwPdZ9fOkk7aNhTHGGLMSixO7oKAgZGVlIT8/35rxMFZxhNOxvG4sY4yx6sHixO61115DYWEhoqKirBkPYxWnQWf110QesWOMMVY9WJzYTZs2DV26dMHkyZOxZ88ea8bEWMXw7QBABDyOB7L5WlHGGGNVn8U3T3z00Ufo1q0bLl68iJdffhktWrQwa+WJefPmWdols5LIyEhERkaiqKjI1qHYlqM74NUMeHBZfTq22Uu2jogxxhgrkzKvPFG8ujkTFtf4ZKISqdErT2jsmgbEbgKemwL0WWLraBhjjFVjlXrliW7duvHKE6zqa9BZndjxDRSMMcaqAYsTu8OHD1sxDMZspEFH9deUs0BhHmDvYNt4GGOMsTLg6fZZzeYZCDjXAYoKgHvnbR0NY4wxViac2LGaTSQqNp8dT3vCGGOsajM7sUtMTERycrLFHW3btg3ff/+9xfUZKzea07F8nR1jjLEqzuzEzt/fHx07dtS7b/DgwejVq5fR+lOnTsXrr79euugYqwjFJyq27CZxxhhjrFIo1c0ThmZG+fvvv/HggekJXi2cWYWx8lWvDWAnBXLTgPTbQK1Gto6IMcYYswhfY8eYvQPg0079POmUbWNhjDHGyoATO8aAYtfZcWLHGGOs6uLEjjGg2HV2nNgxxhirujixYwz4d8Tu4RXgSYZNQ2GMMcYsxYkdYwDg4qWerBgA7p6xbSyMMcaYhTixY0xDmKiYT8cyxhirmko13UlRURGSkpJ0pi0pKioCAL37SpZhrNJq0Ak4v0U9nx1jjDFWBZUqsUtLS4O/v7/B/cb2ERFEIlFpumPlJDIyEpGRkZxsl6QZsUuOBYqUgF2pfj0YY4wxmxORmbMGi8VlP2srEok4mahEMjMzIZfLoVAo4ObmZutwbE+lApb7A/kKYOJfgE9bW0fEGGOsGqmIz12zhyQ2btxYLgEwVmmIxUCDYODmfvV1dpzYMcYYq2LMTuzGjh1bnnEwVjk06PRvYtfpLVtHwxhjjJUK3xXLWHGa6+x4omLGGGNVECd2jBVX/1lAZAdk3gUUd20dDWOMMVYqnNgxVpzMBajbUv2c57NjjDFWxXBix1hJwkTFp20bB2OMMVZKnNgxVpJwnR1PVMwYY6xq4RlYGStJk9jdvwgkPV03tqgAKMoHigqfPi/Qfq7Us62o8GkdPduV+UCtIKD3fEDmartjZYwxVq1wYsdYSe4NALf6QGYysL53+fVz+xCgSAJe2wyI7cqvH8YYYzUGJ3aM6fPcFODYZ+qEy84esJM+fdgDdjI9254+l0j1bNfUKbZNmQfs/S9w/Q9g/wKgz2JbHzFjjLFqgBM7xvTp/I76UZ4cPYBfJgB/rwbqNAHajSrf/hhjjFV7fPMEY7bSagjQfZb6+a73gITjNg2HMcZY1ceJHWO21P1DoMUgQFUI/DwKSI+3dUSMMcaqMLNOxQYGBlqlM5FIhFu3blmlLcaqBbEYGPgV8DgBSDkLbHkNmLAXcJDbOjLGGGNVkFmJXUJCgtH9IpEIRGRyn0gkKl10rFxERkYiMjISRUVFtg6FAYDUCXhtC/BND+DhVWD768DwnwE7vgSWMcZY6YjIUEZWzHfffad3++PHj7Fo0SJkZGSgS5cu6NmzJ3x9fQEAycnJOHjwIP7++294eHhg3rx5cHd3x9ixY617BMximZmZkMvlUCgUcHNzs3U4LOUssOFFQPkE6Pwu0HeprSNijDFmRRXxuWtWYqdPTk4OgoODkZqaii1btqBPnz56y+3fvx+vvvoq6tWrh1OnTsHZ2blMATPr4cSuEvrnV2Db039+XvoM6DDeltEwxhizoor43LX45omlS5fi2rVrWLNmjcGkDgB69+6NNWvW4PLly1i2bJml3TFWM7R4BejxX/XzPe8D8UdsGg5jjLGqxeIRu6ZNm+LOnTvIzs6GnZ3xWfOLiorg4uICf39/XLlyxaJAmfXxiF0lRQT88gZwaTvg4A68eRCo1cjWUTHGGCujSj1il5iYCEdHR5NJHQDY2dnBwcEBiYmJlnbHWM0hEgEDvwTqdwDyMoDNrwJPMmwdFWOMsSrA4sTO2dkZCoUCN27cMFn2+vXrUCgUcHJysrQ7xmoWe0f1GrJuvsCjG8C2cUCR0tZRMcYYq+QsTuxCQkJARHjnnXeQn59vsFxBQQHeffddiEQihISEWNodYzWPqzcwfAtg7wTcPgT8OdvWETHGGKvkLE7sZs2aBbFYjEOHDqFt27bYuHEjEhISUFhYiMLCQiQkJGDjxo1o164dDh48CJFIhNmz+YOJsVKp1xoI/0b9/PQ64PQ3to2HMcZYpWZxYte5c2esW7cOdnZ2uHbtGt544w00atQIDg4OcHBwQKNGjfDGG2/gypUrsLOzw5o1a9CpU6cyB7xt2zaEhYXBw8MDzs7OaNOmDVasWIHCwkKL2ouNjcXQoUPh7e0NBwcHBAQEYMqUKXjw4IHReqmpqZg8eTICAgIgk8ng7e2NoUOHIi4uzmi9goICLF++HG3atIGzszM8PDwQFhaG7du3m4y1oKAAq1evRmhoKDw9PeHg4ABfX1+8+OKL+Pnnn0t13KwKafYS0Gu++vnvHwK3Dto2HsYYY5UXlVFsbCy9+OKLZGdnRyKRSOthZ2dHL774IsXExJS1GyIimjZtGgEgiURCffr0ofDwcHJ3dycAFBoaSrm5uaVqb9u2bSSRSAgABQcH07BhwygwMJAAkLe3N924cUNvvWvXrpGXlxcBoMDAQBo2bBgFBwcLsUVHR+utl5OTQ8899xwBIHd3dwoPD6c+ffoIMURERBiMNSkpiZo3b04AqHbt2vTSSy/Rq6++Ss899xw5OTnR4MGDS3XsREQKhYIAkEKhKHVdVsFUKqLot4jmuxF91IDowTVbR8QYY6yUKuJzt8yJnUZGRgYdPHiQtm7dSlu3bqWDBw9SRkaGtZqnHTt2EABycXGh2NhYYfvDhw+pVatWJhOjkpKTk8nJyYkA0Nq1a4XtSqWSRo0aJSR7KpVKq55KpaJ27doRABo9ejQplUph39q1a4UY7927p9OnJjFt1aoVPXz4UNgeExNDLi4uBIB27dqlUy83N5eaNm1KAGjBggVUUFCgtT8nJ4fOnj1r9rFrcGJXxRTmEX3bR53cfd6WKOeRrSNijDFWClUqsStvmhGxJUuW6Ow7evQoASCZTGZ2Mjlz5kwCQL1799bZl5WVRXK5nADQH3/8obVv9+7dwohbVlaWTt1evXoRAJo1a5bW9vT0dJJKpQSAjh07plNv8eLFBIA6d+6ss2/u3LkEgCZOnGjWsZmLE7sqKOsB0aqW6uRuY38iZYHpOowxxiqFivjctfgau4qUnJyMM2fOAABGjBihsz80NBQNGjRAfn4+9uzZY1abO3bsMNiei4sLBgwYAACIjo7WW2/AgAFwcXHRqatpr2S9PXv2oKCgAA0bNtR7d7Cm3smTJ5GSkiJsLywsxJo1awAAM2fONOvYWDXmUgcYsRWQugAJR9WrU1g2xzhjjLFqqMyJXW5uLlavXo3+/fujZcuWaNRIe4Z8hUKBzZs3Y8uWLRb3cfbsWQCAp6cnAgIC9Jbp0KGDVlljsrKycPPmTa165ran+d5UvRs3biAnJ8fseoGBgfD09AQAnDt3TtgeFxeHtLQ0+Pj4ICgoCBcvXsTChQvx1ltvYdasWdi9ezdUKpXR42XVjHcLYPB6ACIgdhNwaq2tI2KMMVZJSMpS+dy5cxg4cCDu3r0LejpqIBKJtMq4ublhyZIluHbtGry9vdGzZ89S9xMfHw8AaNiwocEyDRo00CprTEJCgvDcUJuG2jMVi6YeESEhIQEtWrQwqx4A+Pr6Ij09XavPCxcuCPtmzZqFFStWCK81ACxfvhzt2rXDr7/+arRtVs006Qv0WQzs/a96frtaQUDj3raOijHGmI1ZPGL36NEj9O/fH0lJSWjfvj0++eQTveueiUQiTJgwAUSEnTt3WtRXVlYWAPVqF4ZoTotmZmaa3Z6xNg21ZyqW4qdni9e19BgePXoEQD3it3z5crz77ru4du0aFAoF9u3bh2eeeQZnz55F//79TU75kp+fj8zMTK0Hq8K6TAbajQJIBWwfDzy4auuIGGOM2ZjFid2nn36Ke/fuoVevXjh16hRmzJgBR0dHvWX79+8PADhx4oSl3dVYmtG5wsJCDB8+HF9++SWeeeYZuLm5oXfv3ti3bx8cHBxw6dIlbN261WhbS5cuhVwuFx6a0UVWRYlEQP9PAb8QID8T2PIqkPPI1lExxhizIYsTu127dkEkEmHFihUQi40306RJE9jb2+PWrVsW9eXq6goAWteslZSdnQ0AekcNDbVnrE1D7ZmKRVOvZF1Lj6F4rG+99ZZOnYYNGwqJ8/79+w22DQCzZ8+GQqEQHklJSUbLsypAIgWG/QC4+wGPE4Co0YCywNZRMcYYsxGLE7vbt29DKpWibdu2JsuKRCK4ublZfOrP398fAIwmIpp9mrLG+Pn5Cc8TExNL1Z7me1P1RCKRVj+m6gHA3bt3dfoMDAzU+7w4zfZ79+4ZbBsAZDIZ3NzctB6sGnCuBYyIAmRuwJ3jwO7pfKcsY4zVUBYndiqVChKJROdmCX2ICNnZ2UavLzOmXbt2ANTXmxm6OSImJgYA0L59e5Ptubm5ISgoSKueue1pvjdVr3HjxlrX25mqd/v2baSnpwP493g19TSvcVpamt66mu36pl9hNYRXU2DIBkAkBs7+CJyItHVEjDHGbMDixK5+/frIzc01uaYqAJw5cwb5+fkGpyoxxdfXF8HBwQCAzZs36+w/duwYkpKSIJPJ0K9fP7PaHDRokMH2srOzsWvXLgBAeHi43no7d+7Ue1pV017Jev369YNUKkViYiKOHz9usF7nzp3h4+MjbK9bty5CQ0MB6D/VWlhYiL/++gsA0LFjR32HymqKxs8DL3ykfr73v8ClaCD7IZCfDaiKbBsbY4yximHpzMZvvvkmicViWrZsmbCtbt26JBaLdcr269ePxGIxffDBB5Z2Z3BJsbS0NINLikVHR1OTJk2oZ8+eOu0VX1Js3bp1wnalUkmjR482a0mxMWPGWLSkWOvWrSktLU3YHhsba3RJsf379xMA8vDwoBMnTgjbCwsLacqUKQSAXF1d6f79+8ZeQh288kQ1pFIR7ZyqXpmi5GNRHaJlfkQrmxGtbk+0JoTo2+eJvhtItHk40bbXiX6dRLTnA6J984kOLyc6vpro9LdEZzcTXdpBdO1PottHiO7G8JJmjDFWShXxuSsisuxinH/++Qdt2rSBs7MzfvnlF/Tu3Rv16tXDgwcPUFSkHh1ITU3FjBkzsGXLFshkMly7dq1Mc61NmzYNq1evhr29PXr16gVnZ2ccOHAAGRkZCAkJwb59+7TuzN20aRPGjx8PPz8/rbnrNLZt24bhw4ejqKgInTp1gr+/P86cOYPbt2/D29sbx44dE07ZFnft2jV07doVDx8+RGBgIIKDgxEfH4/Tp09DIpEgKipKGNkrLjc3F71798aJEyfg4eGBnj17IicnBwcOHEBhYSFmzJiBlStX6j32JUuWYO7cuZBIJOjYsSPq1q2LuLg4JCQkwNHREdu2bRNuojBXZmYm5HI5FAoFX29XnRQVAr9NAi7vBJRPyq8fiQMQ/AYQ8p56RQzGGGNGVcTnrsWJHQCsWLECs2bNgkgkQrt27XD58mXk5+fj1VdfxZ07dxAbG4vCwkIQEb7++mtMnDixzAFHRUUhMjIS586dQ2FhIRo1aoRRo0Zh+vTpkEqlWmVNJXYAEBsbi48++ghHjx6FQqFAvXr18NJLL2Hu3Lnw9vY2GMf9+/exZMkS/N///R/u3bsHuVyOrl27Ys6cOUav8ysoKMCqVavw008/4datW5BKpWjTpg0mT56MoUOHGj32vXv34rPPPsOpU6eQlZWFunXrolevXvjwww/RtGlTo3X14cSuBiACCp88feQW+5pb7PsnQEFOiTJPgEI92zTl8jOBrKc369g7A53eAp6bAjh52vZ4GWOsEqv0iR0ArF+/Hu+//z4UCsW/jYpEwvxr7u7u+OyzzzBmzJiyRcqsjhM7ZjEi4NYB4OD/gJQ49TaZG9BlEtD5XcCB30+MMVZSlUjsAPXNBr/88guOHz+OlJQUFBUVoW7duggJCcHQoUMhl8utESuzMk7sWJkRAdd+Bw79D0i9pN7m6AGETAM6TgSklt0Jzxhj1VGlTuyOHDkCAGjdujXc3d2tGROrIJzYMatRqYDLvwKHlwJp19XbnOsAoTOADq8D9g42DY8xxiqDSp3YicVi2NnZ4cGDB/Dw8LB2XKwCcGLHrE5VBFzcpk7wHieot7n6AN0igHZj1CtlMMZYDVURn7sWz2OnWW+UkzrGmEBsB7R5DZgcA7y8GnDzBbJSgN0RwJfPqidPLlLaOkrGGKu2LE7sgoKCkJWVhfz8fGvGwxirDuzsgWfHAlPjgBc/Bly8gYxE9TQskR2BC9t40mTGGCsHFid2r732GgoLCxEVFWXNeBhj1YlEBnSaCEw9B/RZAjjVAtJvAdFvAGtC1HPt8bq2jDFmNRZfY6dUKtG9e3dcunQJW7ZsMXspL1Z58DV2rMLlZwGnvgb+/gLIezpFUt3WQM//Ao37AGasPc0YY1VVpb55YtGiRXjy5AkiIyORk5ODFi1aICQkBF5eXrCzszNYb968eRYHy6wjMjISkZGRKCoqwvXr1zmxYxXvSQZwIhI4+RVQkK3eVr+DOsELDOMEjzFWLVXqxE4sFmtNRAyoJyY2RbPcGLM9HrFjNpfzCPj7c+DUun+XP/MLBXrOAfyes21sjDFmZZU6sQsLCzMrkSvp0KFDlnTHygEndqzSyEoFjq0CYjYARQXqbXWaAbWDAM9A7YerDyC2+PJgxhizmUqd2LGqjxM7Vuko7gJHPgHO/gCoDEyLYicDPAOKJXvFnssbqKdcYYyxSogTO1auOLFjlVbWfeD+RSD9NvDolvpr+m0g447hhA8AxPaAh5/uKJ9nIODeUD0NC2OM2UhFfO5KyqVVxhgrC9e66kdJRUpAkfRvopce/+/zx/Hq07iPbqofJYns1MmdZpRP5vp0qhUq8RUGtpfiKwDYO6n7kLkBDm7qr8WfO7ip90td+dQyY8xqOLFjjFUddpKnp14DAPTS3qcqAjJTiiV9JZI/5RN18vc4Hrhlk+gNEP2bAMpcdRM/4bn83/2OHoBvMI9AMsZ0lDmxy8/Pxy+//IJjx47h7t27yMnJgaGzuyKRCAcOHChrl4wxpktsB7g3UD8Cu2vvI1Kf3i2e8BU+eTqtiujf6VW0vjfnq7jYc2i3VfgEyMsE8jPV8/flKdTP855+n5/59EYReloms3TH69MOGBUNOHla+IIxxqqjMiV2f//9N1599VWkpKSAiIS7ZDWJXfG7ZovvZ4yxCiUSAW711A//EFtH86/CvH+TPE3il5/1b0IoJIYlEsKH14CUs8Cm/sDoXwFXb1sfCWOskrA4sUtKSkL//v2hUCjQunVr9O3bFytWrICLiwvee+893L9/HwcPHsTt27dRu3ZtvP3220YnLmaMsRrH3kH9cKlTunoPrgI/vAI8uAxs7AuM+U19/SBjrMaz+K7Y6dOn4/PPP8eLL76I//u//4NIJIJYLEbdunWRkpIilFuzZg2mTp2Kfv364bfffrNa4Kzs+K5Yxqqw9Hjg+wFARiLg5guM3QnUamTrqBhjRlTE567Ft2Lt3bsXIpEICxcuNHqK9Z133sHChQvxf//3f/jmm28s7Y4xxlhxngHA+D+AWo2BzLvAhr5A6j+2jooxZmMWJ3aJiYkQi8Vo37691vaCggKdspMmTYJIJMKmTZss7Y4xxlhJ8vrA+N8B71ZAzgP1NXfJsbaOijFmQxYndkQEDw8PiIvNv+Ts7IzMzEydu2LlcjnkcjmuXr1qeaSMMcZ0udQBxu0C6ncAnjwGvhsIJBy3dVSMMRuxOLGrX78+srKytLb5+vqiqKgIV65c0dqem5uLjIwM5ObmWtodY4wxQxw9gDG/Av5dgYIs4MfBwM39to6KMWYDFid2gYGBKCgowK1b/8702alTJwDA119/rVV21apVICL4+/tb2h1jjDFjZK7AyG1A4xfUkzFvfg24vNPWUTHGKpjFiV1YWBiICPv27RO2vfHGGyAiREZGol+/fpgzZw5eeuklzJ8/HyKRCMOHD7dK0IwxxvSwdwRe/RFo/gqgKgS2jQPO/2zrqBhjFcji6U4SEhIwfvx4tGvXDqtWrRK2f/jhh/j444/VjYtEwvV23bp1w969eyGVSq0QNiuLyMhIREZGoqioCNevX+fpThirblRFwM6pwLkfAYiA/iuB4Am2joqxGq8ipjuxOLEzZv/+/di6dSuSkpIgl8vRt29fjBkzBhIJL01bmfA8doxVYyoV8Mcs4PRa9ffPLwJCptk2JsZquCqb2LGqgRM7xqo5IuDAIuDY07Mq3T8Ewmb/u54tY6xCVeoJihljjFVyIhHQez7Qa576+7+WA3/OUSd8jLFqiRM7xhir7rpGAC+qr33GyUhg11T1dXiMsWrH4oveFi1aZFG9efPmWdolY4wxS3WaCEidgJ1TgLjvgYJcYNDXgJ29rSNjjFmRxdfYicVio2vElkREEIlEKCri/xIrC77GjrEa6J8dwC9vACol0KQ/MGQDYO9g66gYqxEq4nPX4hG7bt26GU3sFAoFrly5gvz8fHh4eKB169aWdsUYY8xaWgwC7J2An0cD13YDW14FXtsMSJ1tHRljzArK9a7Y7OxsfPzxx/jf//6HFStWYMaMGeXVFbMAj9gxVoPFH1GvTlGYAzToDIyMAhzkto6KsWqtyt8V6+LigoULF+KDDz7ABx98gMOHD5dnd4wxxswV0A0Y85s6mUs6CXz3MpDzyNZRMcbKqELuio2IiAARCStSMMYYqwQaBANj/w9wqg3cOw9s6gdk3rN1VIyxMqiQxK5WrVpwd3fH6dOnK6I7xhhj5qrXGhj/O+DqAzy8Cmx8EXh8x9ZRMcYsVCFrfGVlZSEjIwMymawiumOMMVYadZ4BXv8d+H4g8Dhendy1GQ5IZOrpUOw0X6XFtkkNbDdSVmzHq14wVs4qJLFbuXIliAgBAQEV0R1jjLHS8vBXj9x9/wqQdg04+kk5dCJSJ3sSGeDiDbj5AG71Add6/z5381E/nGpxEsiYBSxO7I4cOWJ0f15eHpKSkvDLL7/gzz//hEgkwvDhwy3tjjHGWHlz81EndzHrgZyHQFEBUFQIKPOfPi/Qs60QKMovsb3YNi0EKPPUjzwFkHbdcCx2MsCtnvoUsSbZc6uv3qZJAF281aOAjDFBuU9QrGm+R48e2LNnD5+OrUR4uhPGWLkiUk+EXDIJLHwCZN0HMlOAzGQg696/zzPvATkPzGtfJAZc6hZL/J4+XOqq5+Wzdyz2cAIkDuqvmm2cFLIKVqknKAb+Tdr0sbOzg4eHB9q0aYPhw4dj3LhxEIt5adrKIDIyEpGRkbwKCGOsfIlET6+x07NsWZ0mhuspC/5N9rJSniZ9JR5Z9wAqUu/PSgGSLYjPTgpIHEskgMWTwKfP7R1KJIYO6hFFiVT7q51Ud5tEVux6Q83XCrrekEi9JjAVPf2qUj+HSH0sdvZ8ursaKtcJilnlxiN2jLEqS1WkPl2sGeUTRvxSgOxU9ahg4RNA+eTf55rvKwXRv0leyWRQLPk3CdMkZqQCVCrdJE14rtJN4mDq4/1pgieR6fmqb5uZX+2dAUePYg93XpP4qUo/YscYY4zZhNgOcK2rftQvRT2V6t/TwYW5QGHe06/Fk8Bi25V5+rdpTi9rfc1XjzYW5Re73rDYNlIVC6TY9Yb51n5xzEXqY66IZFfq+m+SVzLp0/q+xMPesfxjq2Y4sWOMMVZziMWA+OkpV3hWbN9FyqeJXrGbUTRJX/EEUaV8eqpWDIjsSjx/+lUkfrpd81xfWc3zEmVB6n6U+f8mqXq/5j2NyVgZPV/zs4C8DODJY/VNMgBQkKV+KBJL95pJHHSTvcAwoOObVv7hVB+c2DHGGGMVwU6ifkidbR1JxY2EqYrUyd2Tx8UeGSW+L/HQJIUqpTpZzLqnfmhc/T/1knjGrtOswSxO7F5//XWrBCASibB+/XqrtMUYY4yxSkRsBzh5qh+lQQQUZOsmfafWAYl/A2e+BfrxMqX6lPt0JxoluxGJRCAiiEQivjvTRvjmCcYYY1XKrYPAD4PU1+xFXAFkrraOqFQq9c0TY8aMgUgkwm+//YaMjAw4ODjg2Wefha+vLwAgOTkZsbGxePLkCTw8PDBgwACrBc0YY4yxGiggDKgVBDy6CVyIAoIn2DqiSsfixG7Tpk0YMWIEFAoFZs+ejQ8//FAn+8zKysLy5cuxdOlSFBQU4KeffipzwIwxxhirocRioMME4M/Z6tOxHV7nufhKsHjG4G+++QY///wzFixYgP/97396hxRdXV2xZMkSLFiwAFu3bsW3335bpmABYNu2bQgLC4OHhwecnZ3Rpk0brFixAoWFhRa1Fxsbi6FDh8Lb2xsODg4ICAjAlClT8OCB8ZnPU1NTMXnyZAQEBEAmk8Hb2xtDhw5FXFyc0XoFBQVYvnw52rRpA2dnZ3h4eCAsLAzbt28vVdwffPABRCIRRCIRlixZUqq6jDHGWJXVdoR6ougHl4E7f9s6msqHLNSpUyeSSCSUmZlpsmxmZiZJJBLq1KmTpd0REdG0adMIAEkkEurTpw+Fh4eTu7s7AaDQ0FDKzc0tVXvbtm0jiURCACg4OJiGDRtGgYGBBIC8vb3pxo0beutdu3aNvLy8CAAFBgbSsGHDKDg4WIgtOjpab72cnBx67rnnCAC5u7tTeHg49enTR4ghIiLCrLiPHz9OYrGYRCIRAaDFixeX6rg1FAoFASCFQmFRfcYYY8wmfptCNN+NKGqcrSMplYr43LU4sZPL5VSrVi2zy3t6epJcLre0O9qxYwcBIBcXF4qNjRW2P3z4kFq1alWqxIiIKDk5mZycnAgArV27VtiuVCpp1KhRQrKnUqm06qlUKmrXrh0BoNGjR5NSqRT2rV27Vojx3r17On1qEtNWrVrRw4cPhe0xMTHk4uJCAGjXrl1G487JyaHGjRtT/fr16ZVXXuHEjjHGWM2Tcl6d2C30JMrU/bytrCp1Yufq6kp2dnb06NEjk2UfPXpEdnZ25Orqaml3wojYkiVLdPYdPXqUAJBMJqOMjAyz2ps5cyYBoN69e+vsy8rKIrlcTgDojz/+0Nq3e/duYcQtKytLp26vXr0IAM2aNUtre3p6OkmlUgJAx44d06m3ePFiAkCdO3c2GvfUqVMJAO3evZvGjh3LiR1jjLGa6dvn1cndoWW2jsRsFfG5a/E1dq1btwYRYdGiRSbLLl68GCqVCq1atbKor+TkZJw5cwYAMGLECJ39oaGhaNCgAfLz87Fnzx6z2tyxY4fB9lxcXIS7eKOjo/XWGzBgAFxcXHTqatorWW/Pnj0oKChAw4YNERISYrDeyZMnkZKSojfmw4cP44svvsCYMWPQr18/o8fHGGOMVWvBT1efiN2oXsKNASjDzRPvvPMOiAhffPEFxo8fj9u3b+uUiY+Px+uvv47Vq1dDJBLh3Xfftaivs2fPAgA8PT0REBCgt0yHDh20yhqTlZWFmzdvatUztz3N96bq3bhxAzk5OWbXCwwMhKenegLHc+fO6ezPzs7G66+/Dm9vb3z22Wd622CMMcZqjOYDAOc66lUprpk3qFMTWJzYjRw5EuPGjQMR4fvvv0fjxo3h7++PkJAQhISEwN/fH0FBQfjuu+9ARBg9ejRGjhxpUV/x8fEAgIYNGxos06BBA62yxiQkJAjPDbVpqD1TsWjqEZFWP+Ycg2YOQH3H8P777yM+Ph5r1qyBh4eHwTYYY4yxGkEiA9qPUT8//Y1tY6lELE7sAGDDhg347LPP4OHhASJCYmIiTpw4gRMnTiAxMRFEBHd3d6xatQobN260uJ+srCwAgLOz4fX1NKdFMzMzzW7PWJuG2jMVS/HTs8XrluUY9u7di7Vr1+K1117DK6+8YrC+Kfn5+cjMzNR6MMYYY1XWs+MBkRhIOAo8uGrraCoFiyco1pg6dSreeust7N27FzExMcL8b15eXujQoQOef/55ODg4lDnQmkqhUGDChAmoU6cOvvjiizK1tXTpUixcuNBKkTHGGGM25t4AaNIPuPp/6gmL+39i64hsrsyJHQDIZDK8/PLLePnll63RnA5XV/VacMWvWSspOzsbAMxae03TnqZNuVxudnuurq5IT083GIumXsm6lh7De++9h7t37+Lnn39G7dq1DdY1x+zZszFjxgzh+8zMTOHUMWOMMVYlBb+hTuzObwV6z69y68dam1USu/Lm7+8PAEhKSjJYRrNPU9YYPz8/4XliYqLeu3UNtefv74/09HQkJiYajUMkEmn1o2nHUD0AuHv3rk6fO3bsgEQiwVdffYWvvvpKq/zVq+ph5/Xr12P//v2oW7cutm7darB9mUwGmUxmcD9jjDFW5QR0L7Z+7M/qRK8Gs3pi9+DBA2zduhXXrl2DTCZD+/btMXjwYDg6OlrcZrt27QAAjx49Qnx8vN47Y2NiYgAA7du3N9mem5sbgoKCcPPmTcTExOhN7Ay11759e8TFxQn7DdVr3Lix1vV2mnYM1bt9+zbS09MB/Hu8GkqlEn/99ZfB40lISEBCQoJWIskYY4zVCGKxOpn7YxZw+lv1WrI1eP1Ys2+eSEtLw6JFi7B48WLk5+frLbN7924EBQVh+vTp+Prrr/H5559j7NixaNmyJa5du2ZxkL6+vggODgYAbN68WWf/sWPHkJSUBJlMZvb8boMGDTLYXnZ2Nnbt2gUACA8P11tv586dek+ratorWa9fv36QSqVITEzE8ePHDdbr3LkzfHx8hO0ZGRkg9UTSOo+xY8cCUM8TWPIuXMYYY6zGaDNcvX7swyu8fqy5Mxlv2bKFRCIRdevWTe/+27dvk7Ozs7CGaclH48aNqaCgwOKZlA0tKZaWlmZwSbHo6Ghq0qQJ9ezZU6e94kuKrVu3TtiuVCpp9OjRZi0pNmbMGIuWFGvdujWlpaUJ22NjY81eUqw4XnmCMcYYe2rn1Kfrx461dSQGVaolxd59910Si8UUGRmpd/8bb7xBIpGIxGIxzZ8/n1JSUigzM5NWr15NEomExGIxbdy4sUzBapbTsre3p759+9LgwYPJ3d2dAFBISAjl5uZqld+4cSMBID8/P73tRUVFkZ2dHQGgTp060auvvkqBgYEEgLy9venGjRt66129epXq1KlDACgwMJBeffVV6tixIwEgiURC0dHReuvl5ORQly5dCAB5eHjQ4MGDqW/fvmRvb08AaMaMGaV6PTixY4wxxp66d6HSrx9bqRK7zp07k1gspuvXr+vsUyqV5O7uTmKxmMaPH6+z/7333iORSESvvPJK2aIlop9//pm6detGbm5u5OjoSC1btqRly5ZRfn6+TllTiR0RUUxMDIWHh1OdOnVIKpWSn58fTZo0ie7fv280jnv37tGkSZPIz8+PpFIp1alTh8LDw7VGE/XJz8+npUuXUsuWLcnR0ZHkcjl169aNoqKizDr+4jixY4wxxor5ts/T9WOX2joSvSric1dERGTOKduAgADcu3cPeXl5Ovvi4uLQoUMHiEQiHD9+HJ07d9baf/36dTRt2hQBAQG4detWKU4Us/KUmZkJuVwOhUJh1jQxjDHGWKV2cTvwywTAtR7w3kXAzt7WEWmpiM9ds2+eSE1N1Zr/rbgzZ84AABwdHdGxY0ed/UFBQbCzsxMmL2aMMcYYs7pmL/+7fuzV3baOxibMTuzEYjEyMjKgUql09mmm8GjZsiXEYt0mxWIx5HK5wbtpGWOMMcbKTCID2qtnjMCZb20bi42YndjVq1cPKpUK//zzj86+48ePQyQS6R2t08jMzNSa140xxhhjzOo6FF8/9oqto6lwZid2mqSt5HqlMTExwgoIPXr00Fv3+vXrUCqVPIEuY4wxxsqX3Fe9fiwAnFlv21hswOzEbty4cSAirF+/HpMmTcKff/6J9evXCxPxenp6Gpwc+NChQwCA1q1bWyFkxhhjjDEjNMuKnd8K5GfZNpYKZnZi9/zzz2PIkCEgInz99dfo168fJk6ciLt370IkEmHBggUG1yHdsmULRCIRunbtarXAGWOMMcb0CgwDajUGCrLUyV0NYnZiBwA//vgjJk+eDKlUKixr5eTkhCVLlmDSpEl668TFxeHIkSMQiUTo37+/VYJmjDHGGDNIJPp31O7Mt4B5M7tVC2bPY1dcdna2cF1dy5Yt4eDgYLDso0ePcP/+fUgkEjRp0sTySJnVREZGIjIyEkVFRbh+/TrPY8cYY6z6eZIBrGoGFOYC43YD/qG2jqhC5rGzKLFj1QNPUMwYY6xa2zUNiN0ENH8FGPadraOpXBMUM8YYY4xVKcFvqr9e/T8g855tY6kgnNgxxhhjrHqq2xJo2AVQKYE424/YVQRO7BhjjDFWfWluoojZCBQV2jaWCsCJHWOMMcaqr2YDAGcvIPu++pRsNceJHWOMMcaqL4kUePbp+rGnq//6sZzYMcYYY6x6e3Y8ILID7hyr9uvHcmLHGGOMsepNXh9o8qL6+ZnqPWrHiR1jjDHGqr+OT6c+Ob8VyMu0bSzliBM7xhhjjFV/Ad2frh+bDVz42dbRlBtO7BhjjDFW/dWQ9WPLlNgRETZt2oQXXngB9erVg0wmg52dncGHRCKxVtyMMcYYY6XTdjhg7ww8vAokHLN1NOXC4kwrPz8f/fv3x6FDh8DLzTLGGGOs0nOQA62HAbEbgTPfAAFdbR2R1Vmc2C1fvhwHDx4EAISHh2PgwIHw8fHhUTnGGGOMVV7Bb6gTuytP1491q2friKzK4ixs69atEIlEmDdvHubPn2/NmBhjjDHGyodm/djEE0DsJqDHbFtHZFUWX2MXHx8PkUiEiIgIa8bDGGOMMVa+NDdRxG6qduvHWpzYubq6Qi6Xw8XFxZrxsAoQGRmJ5s2bIzg42NahMMYYYxWv+PqxV3bZOhqrsjixCw4OhkKhQHp6ujXjYRVg0qRJuHz5Ms6cOWPrUBhjjLGKJ5ECz45TPz+z3qahWJvFid2MGTNARPj000+tGQ9jjDHGWPl7dty/68emXrZ1NFZjcWLXq1cvLF++HMuWLcPixYuRm5trzbgYY4wxxsqPvD7QtJ/6eTVaP1ZEFk5C17NnTwDAuXPnoFAoIJPJ0KJFC7i6uhruTCTCgQMHLIuUWV1mZibkcjkUCgXc3NxsHQ5jjDFWsW7/BXw/AJC6ADOuAA7l+1lYEZ+7Fk93cvjwYa3v8/LyEBsba7SOSCSytDvGGGOMMesK6AbUfgZIu65eP7bjm7aOqMwsTux47jrGGGOMVWma9WN//wA4/Y36eRUfhLL4VCyr+vhULGOMsRovTwGsbAYU5gBjd6lH8cpJRXzuWnzzBGOMMcZYlecgB9q8qn5++hvbxmIFnNgxxhhjrGbTrERxdTeQmWLbWMqIEzvGGGOM1WzeLYCGzwFUpF5mrAorc2J3/vx5TJw4Ec2bN4ebmxvs7OwMPiQSi+/VYIwxxhgrPx2fjtrFfQ+oVLaNpQzKlGl9+eWXmDFjBoqKisD3YDDGGGOsymr6EmDvBGTdA1IvAfVa2zoii1g8Ynfq1ClMmzYNRUVFePfdd7Fnzx4AgKenJ/bv348ff/wR48aNg1QqRe3atbF582YcPHjQaoEzxhhjjFmNRAb4d1U/v1V1F1OwOLFbvXo1iAjTpk3DF198gb59+wIApFIpevbsiREjRmDDhg04efIkRCIR5s6di/bt21stcMYYY4wxqwrqrf56swYmdsePH4dIJMK0adO0tpc8Jdu2bVt88cUXuHXrFj7++GNLu2OMMcYYK19BvdRfE08C+dm2jcVCFid2qampkMlk8PPz+7cxsRh5eXk6ZQcNGgR7e3tER0db2h2zosjISDRv3hzBwcG2DoUxxhirPDwDAXc/QFUIJBy1dTQWsTixc3JygpOTk9Y2V1dXZGZmIj8/X2u7vb09nJyccOfOHUu7Y1Y0adIkXL58GWfOnLF1KIwxxljlIRJV+dOxFid29evXR2ZmJpRKpbCtUaNGAKCTMKSkpEChUPCds4wxxhir3DSnY2/ut20cFrI4sWvWrBmKiopw8eJFYVtYWBiICIsWLRJOyRYUFGDq1KkAgFatWpUxXMYYY4yxcuTfFRBLgMfxQPptW0dTahYndn369AERYdeuXcK2SZMmQSaT4cCBA/D19UVISAjq16+PHTt2QCQSYfLkyVYJmjHGGGOsXDi4AQ06qZ9XwdOxFid2gwcPxvz58+Hj4yNsCwgIwObNm+Hq6or09HScOHECjx49gkgkwgcffICRI0daJWjGGGOMsXIjnI6teomdiMrhwrf09HTs2bMHSUlJkMvl6NOnD4KCgqzdDSujzMxMyOVyKBQKuLm52TocxhhjrHJIOQes6w5IXYAP4gGJ1CrNVsTnbrks3urp6YlRo0aVR9OMMcYYY+WrbmvAqTaQmwYknQICuto6IrNZfCqWMcYYY6xaEov/PR1bxZYXs0pid+HCBXz88ceYPHkyJkyYoLWvsLAQKSkpuHfvnjW6wrZt2xAWFgYPDw84OzujTZs2WLFiBQoLCy1qLzY2FkOHDoW3tzccHBwQEBCAKVOm4MGDB0brpaamYvLkyQgICIBMJoO3tzeGDh2KuLg4o/UKCgqwfPlytGnTBs7OzvDw8EBYWBi2b9+ut3xhYSEOHDiAmTNnIjg4GO7u7rC3t0fdunUxYMAA7N6926LjZowxxpgRjarotCdUBhkZGRQeHk5isZjEYjGJRCISi8VaZXJycqh27dpkZ2dHly5dKkt3NG3aNAJAEomE+vTpQ+Hh4eTu7k4AKDQ0lHJzc0vV3rZt20gikRAACg4OpmHDhlFgYCABIG9vb7px44beeteuXSMvLy8CQIGBgTRs2DAKDg4WYouOjtZbLycnh5577jkCQO7u7hQeHk59+vQRYoiIiNCps2/fPgJAAKhu3brUv39/GjZsGLVs2VLYPnHiRFKpVKU6diIihUJBAEihUJS6LmOMMVatZT0gmu+mfmSlWqXJivjctTixKygooC5dupBYLCYXFxd6+eWXydHRUSexIyKaOXMmiUQiWrBggcWB7tixgwCQi4sLxcbGCtsfPnxIrVq1MpgYGZKcnExOTk4EgNauXStsVyqVNGrUKCHZK5kwqVQqateuHQGg0aNHk1KpFPatXbtWiPHevXs6fWoS01atWtHDhw+F7TExMeTi4kIAaNeuXVp1Dhw4QIMHD6YjR47otLd161ays7MjAPTdd9+ZfewanNgxxhhjRqwJVSd257ZYpblKnditWbOGRCIRBQUFUXJyMhER1a1bV29id/LkSRKJRNSzZ0+LA9WMiC1ZskRn39GjRwkAyWQyysjIMKu9mTNnEgDq3bu3zr6srCySy+UEgP744w+tfbt37xZG3LKysnTq9urViwDQrFmztLanp6eTVColAHTs2DGdeosXLyYA1LlzZ7Pi15gwYQIBoF69epWqHhEndowxxphR+xaoE7vtb1iluYr43LX4GrstW7ZAJBLh008/1ZrLTp927dpBLBbj6tWrFvWVnJwsLFM2YsQInf2hoaFo0KAB8vPzsWfPHrPa3LFjh8H2XFxcMGDAAABAdHS03noDBgyAi4uLTl1NeyXr7dmzBwUFBWjYsCFCQkIM1jt58iRSUlLMOgZA/doCQFJSktl1GGOMMWaG4jdQqFS2jcVMFid2Fy9ehEgkQp8+fUyWlUqlkMvlePTokUV9nT17FoB6GpWAgAC9ZTp06KBV1pisrCzcvHlTq5657Wm+N1Xvxo0byMnJMbteYGAgPD09AQDnzp0zeQwaN27cAADUq1fP7DqMMcYYM4NvR/VcdrmPgPvnbR2NWSxO7HJzc+Hq6gqp1LxJ+woLCyGRWDZtXnx8PACgYcOGBss0aNBAq6wxCQkJwnNDbRpqz1QsmnpEpNWPOcfg6+urt09D7t+/j02bNgFQrwTCGGOMMSuSSIGA7urnVWQVCosTu9q1ayMzMxPZ2dkmy8bHxyM7O9vkKVtDsrKyAADOzs4Gy2hOi2ZmZprdnrE2DbVnKpbip2eL17X2MSiVSowaNQoKhQKtWrXCW2+9ZbJOfn4+MjMztR6MMcYYMyKop/prdU/sOnVSL5BrzjxqX3zxBQCga9eqM3NzZff222/jwIEDqFWrFrZv327WyOnSpUshl8uFh2Z0kTHGGGMGaOazu3sayKv8AyIWJ3avv/46iAhz5841erH/2rVr8fnnn0MkEmHixIkW9eXq6goAWteslaQZOTRn7TVNe8baNNSeqViKj2AWr2vNY5g2bRrWr18PDw8P7Nu3D88884zR8hqzZ8+GQqEQHnzDBWOMMWaCZwDgGQiolED8EVtHY5LFiV3//v0xePBg3Lx5Ex06dMD777+PJ0+eAADWrVuHOXPmoE2bNnj33XdBRHjjjTeEUb7S8vf3B2D8zk/NPk1ZY/z8/ITniYmJpWpP872peiKRSKsfU/UA4O7du3r7LC4iIgKrV6+Gu7s79u7dK9wVaw6ZTAY3NzetB2OMMcZMCOqt/loFlhcr05JiP/zwA0aOHIn79+/j008/Fa4je+edd7Bs2TJcvHgRRITXX38dkZGRFvejSV4ePXpk8MaCmJgYAED79u1Ntufm5oagoCCteua2p/neVL3GjRtrXW9nqt7t27eRnp4OAAaTtQ8++ACrVq2CXC7H3r17Dd5hyxhjjDErKr68GJFtYzGhTImdg4MDfvjhBxw5cgSjR49Go0aN4OjoCKlUioYNG2LEiBE4fPgwvv32W4vviAXUd4sGBwcDADZv3qyz/9ixY0hKSoJMJkO/fv3ManPQoEEG28vOzsauXbsAAOHh4Xrr7dy5U+9pVU17Jev169cPUqkUiYmJOH78uMF6nTt31nuTyaxZs/Dxxx9DLpdj3759wuvBGGOMsXLmHwqI7YGMRODRLVtHY1y5TX1sZYaWFEtLSzO4pFh0dDQ1adJE74oXxZcUW7dunbBdqVTS6NGjzVpSbMyYMRYtKda6dWtKS0sTtsfGxhpcUoyIaM6cOcJqF6dPnzbj1TIPrzzBGGOMmWnTS+pVKE5+bXETFfG5KyKq5GOKxUybNg2rV6+Gvb09evXqBWdnZxw4cAAZGRkICQnBvn374OjoKJTftGkTxo8fDz8/P6055TS2bduG4cOHo6ioCJ06dYK/vz/OnDmD27dvw9vbG8eOHRNO2RZ37do1dO3aFQ8fPkRgYCCCg4MRHx+P06dPQyKRICoqShjZKy43Nxe9e/fGiRMn4OHhgZ49eyInJwcHDhxAYWEhZsyYgZUrV2rV2blzJwYOHAhAPblxixYt9L42tWvXxieffFKalxOZmZmQy+VQKBR8vR1jjDFmzLHPgP3zgcYvACOjLGqiQj53yy1lLCc///wzdevWjdzc3MjR0ZFatmxJy5Yto/z8fJ2yGzduJADk5+dnsL2YmBgKDw+nOnXqkFQqJT8/P5o0aRLdv3/faBz37t2jSZMmkZ+fH0mlUqpTpw6Fh4drjSbqk5+fT0uXLqWWLVuSo6MjyeVy6tatG0VFRektrzkGUw9jx2gIj9gxxhhjZrp3QT1it6QuUWGeRU1UqhE7Y3dzloaxlRdYxeIRO8YYY8xMRMDKJkB2KjDmNyAwrNRNVMTnrtl3NPj7+0MkEpWpM5FIBKVSWaY2GGOMMcYqnEikvjv2/Gb1KhQWJHYVoVR3xRJRmR+MMcYYY1VSkGbak8o7n12p5iARiUTw9/fHuHHj0K1bt/KKiTHGGGOs8gnsAUAEPPgHyLwHuNWzdUQ6zE7sXnzxRezduxfx8fFYsGABAgMDMX78eIwdOxb169cvzxgZY4wxxmzPuRbg0w5IiQNuHQTajbR1RDrMPhW7e/duJCYm4qOPPkLjxo1x69YtzJ07F/7+/njxxRexbds2FBQUlGesjDHGGGO2pTkdW0mXFyvVNXb16tXDrFmzcPXqVRw9ehTjxo2Do6Mj/vzzT7z22mvw8fHB1KlTERcXV17xMsYYY4zZjmZ5sVsHAVWRbWPRw+IlxUJCQrB+/Xrcv38f69evR0hICNLT0/Hll18iODgYbdq0werVq/Ho0SNrxssYY4wxZju+HQCZG/DkMZByztbR6CjTWrEA4OTkhPHjx+PIkSO4ceMGZs+ejfr16+PixYuYPn06li9fbo04mRVFRkaiefPmvN4sY4wxVlp29kBgd/XzSng6tsyJXXGNGjXC66+/jhEjRkAqlVqzaWZFkyZNwuXLl3HmzBlbh8IYY4xVPZrTsTf32zYOPUo13Ykhubm5iIqKwoYNG3D8+HEA6jnvWrVqhV69elmjC8YYY4yxykFzA8XdGOBJBuDobstotJQpsTt27Bg2bNiA7du3IycnB0QEDw8PDB8+HOPHj8ezzz5rrTgZY4wxxioH94ZA7WeAtOtA/F9A84G2jkhQ6sQuJSUFmzZtwqZNm3Dr1i0QEcRiMZ5//nmMHz8egwYN4tOwjDHGGKveGvVSJ3Y3D1TNxC4qKgobN27E/v37oVKpQERo1KgRxo0bh7Fjx8LX17c842SMMcYYqzyCegGn1qgTOyL1WrKVgNmJ3WuvvQaRSAQnJycMHToU48ePR9euXcszNsYYY4yxyskvBLCTAZl31SN3dZrYOiIAFpyKdXJywuHDh3H48OFSdyYSiXDr1q1S12OMMcYYq1SkToDfc8DtQ+pRu6qY2BERHj58iIcPH1rUmaiSDFMyxhhjjJVZUC91YnfrANDlXVtHA6AUid38+fPLMw7GGGOMsaqlUS8A/wUSjgGFTwB7R1tHxIkdY4wxxphFvJoBrj5AVgpw5+9/57ezIauuPMEYY4wxVmOIREBQT/XzWwdtG8tTnNgxxhhjjFmqki0vxokdY4wxxpilAsMAkRh4eBVQ3LV1NJzYMcYYY4xZzMkTqP90CdVKcDqWEzvGGGOMsbIQTscesG0c4MSuRoqMjETz5s0RHBxs61AYY4yxqk9zN+ztQ0CR0qahiIiIbBoBs5nMzEzI5XIoFAq4ubnZOhzGGGOsaipSAh83AvIygAn7gAYd9RariM9dHrFjjDHGGCsLO4n6JgrA5qdjObFjjDHGGCsrzenYW5zYMcYYY4xVbZobKJJjgdx0m4XBiR1jjDHGWFnJ6wN1mgGkAm4ftlkYnNgxxhhjjFlDJTgdy4kdY4wxxpg1NHq6buzNA4CNJh3hxI4xxhhjzBr8QgCJI5B1D3hwxSYhcGLHGGOMMWYN9g6Af4j6uY1Ox3JixxhjjDFmLTZeXowTO8YYY4wxa9HcQHHnb6Agt8K758SOMcYYY8xaaj8DyBsARfnAneMV3j0ndowxxhhj1iISad8dW8E4sWOMMcYYsyYbzmfHiR1jjDHGmDUFdAdEdkDadSAjsUK75sSOMcYYY8yaHN0B32D18wo+HcuJXQ0UGRmJ5s2bIzg42NahMMYYY9WTjU7HiohstOYFs7nMzEzI5XIoFAq4ubnZOhzGGGOs+rgbC3zbE5C5AR/cBuzsK+Rzl0fsGGOMMcaszact4OgJ5GcCd2MqrFtO7BhjjDHGrE1sBzTqoX5egadjObFjjDHGGCsPNlhejBM7xhhjjLHyoLmBIuUskPOoQrrkxI4xxhhjrDy41gW8WwIg4PahCumSEzvGGGOMsfJSwcuLVbnEbtu2bQgLC4OHhwecnZ3Rpk0brFixAoWFhRa1Fxsbi6FDh8Lb2xsODg4ICAjAlClT8ODBA6P1UlNTMXnyZAQEBEAmk8Hb2xtDhw5FXFyc0XoFBQVYvnw52rRpA2dnZ3h4eCAsLAzbt283Gau1j50xxhhj5az4fHYVMcMcVSHTpk0jACSRSKhPnz4UHh5O7u7uBIBCQ0MpNze3VO1t27aNJBIJAaDg4GAaNmwYBQYGEgDy9vamGzdu6K137do18vLyIgAUGBhIw4YNo+DgYCG26OhovfVycnLoueeeIwDk7u5O4eHh1KdPHyGGiIiICjt2IiKFQkEASKFQlLouY4wxxsxQmEe0pC7RfDdSXP+73D93q0xit2PHDgJALi4uFBsbK2x/+PAhtWrVymRiVFJycjI5OTkRAFq7dq2wXalU0qhRo4RkT6VSadVTqVTUrl07AkCjR48mpVIp7Fu7dq0Q471793T61CRnrVq1oocPHwrbY2JiyMXFhQDQrl27yv3YNTixY4wxxirAj0PVid0fSzmx09CMiC1ZskRn39GjRwkAyWQyysjIMKu9mTNnEgDq3bu3zr6srCySy+UEgP744w+tfbt37xZG3LKysnTq9urViwDQrFmztLanp6eTVColAHTs2DGdeosXLyYA1LlzZ5191j52DU7sGGOMsQpw8mt1YrfmhXL/3K0S19glJyfjzJkzAIARI0bo7A8NDUWDBg2Qn5+PPXv2mNXmjh07DLbn4uKCAQMGAACio6P11hswYABcXFx06mraK1lvz549KCgoQMOGDRESEmKw3smTJ5GSkiJsL49jZ4wxxlgFCuqt/pp4uty7qhKJ3dmzZwEAnp6eCAgI0Fumw/+3d+dBUd/3/8CfK8eyXAsoAnIsXsG0IIERYwIxHrGlhNgiOmlSFadmEhM7Rk0TtdWilkRjM81IptNO2xRqokUbNYWvpoIEU9Gp1SDxSGwlcpmAFbawHBU5Xr8/+O3WlV1YkN11d5+PGWfWz/t6vfeTd/bl55wxw6juYNra2lBVVWXUztL+9H8fqt3Vq1fR0dFhcbtJkyYhKCgIAFBZWTmg3WjNnYiIiGwsaBIQoAGkx+pDOURiV11dDQCIiooyWycyMtKo7mBqamoMn831aa6/oWLRtxMRo3EsmUNERMSAMUd77kRERGRjCsX/7o61MnebjHKP2traAAA+Pj5m6+hPi+p0Oov7G6xPc/0NFcudp2fvbDvSOYzm3Lu6utDV1WUyvtG270wdTn/ZZLX+iYiIHElc+0Q8Y4NxHCKxo9GxY8cObNu2zSZjXfyqBf93ocEmYxEREd3vyjABcaIBcMmq4zhEYufn5wcARtes3a29vR0A4O/vb3F/+j7VarXF/fn5+UGr1ZqNRd/u7rYjncNozn3Tpk1Yv3694e86nc5wGne0pU+fgAdC/Iau6OIU9g6AiIgMrP344MqOvcCueKuO4RCJXXR0NACgvr7ebB19mb7uYDQajeFzXV0d4uLiLO4vOjoaWq0WdXV1g8ahUCiMxtH3Y64dAFy/fn3AmKM5d6VSCaVSOWid0ZI8ZRySp4yzyVhERESOQKfTYbWVx3CImycSEhIAAM3NzWZvEDh37hwAIDExccj+/P39MWXKFKN2lvan//tQ7aZOnWp0vd1Q7a5duwatVgvgf/O98/NozZ2IiIicl0MkdhEREUhKSgIA7Nu3b0B5eXk56uvroVQqkZaWZlGfGRkZZvtrb29HUVERAGDRokUm2xUWFpo8Parv7+52aWlp8PT0RF1dHU6dOmW23axZszBhwgTDdmvMnYiIiJyU1R59PMrMvVarqanJ7Gu1Dh06JDExMTJv3rwB/d35SrHf/va3hu09PT2ybNkyi14ptnz58hG9Umz69OnS1NRk2P7pp5+O6JVig83dEnzzBBERke3Y4nfXYRI7EZE1a9YIAPHw8JDU1FTJzMyUgIAAASDJycnS2dlpVD8vL08AiEajMdnfgQMHxM3NTQDIww8/LE8//bRMmjRJAEhISIhcvXrVZLsrV65IcHCwAJBJkybJ008/LTNnzhQA4u7uLocOHTLZrqOjQx555BEBIIGBgZKZmSmpqani4eEhAGT9+vWjNndLMLEjIiKyHSZ2Juzfv19mz54t/v7+olKpJDY2Vnbu3CldXV0D6g6V2ImInDt3ThYtWiTBwcHi6ekpGo1GVq9eLY2NjYPG0dDQIKtXrxaNRiOenp4SHBwsixYtMjqiZkpXV5fs2LFDYmNjRaVSiVqtltmzZ8uBAwdGde6WYGJHRERkO7b43VWIiLXv7qX7lE6ng1qtRmtrq0WPiSEiIqKRs8XvrkPcPEFEREREQ2NiR0REROQkmNgREREROQkmdkREREROgokdERERkZNgYkdERETkJJjYERERETkJd3sHQPajf4ShTqezcyRERETOT/97a81HCDOxc2HNzc0AgMjISDtHQkRE5Dqam5uhVqut0jcTOxcWFBQEAKirq7PKf2BJSUk4e/bsqPfL/oem0+kQGRmJ+vp6qz3d3JG/H0fv39H3ryN/99bu39H3LfsfXGtrK6Kiogy/v9bAxM6FjRnTf4mlWq22yv9A3NzcrPqqMvY/NH9/f6uN4ejfj6P3Dzju/nX07577lv3fK/3vr1X6tlrP5PJWr17N/u3Yv7U5+vfj6P1bmzXjd/TvnvuW/d/PFGLNK/jovmaLlxGTfXDfOjfuX+fFfevcbLF/ecTOhSmVSmRnZ0OpVNo7FBpl3LfOjfvXeXHfOjdb7F8esSMiIiJyEjxiR0REROQkmNgREREROQkmdi7oz3/+M+bMmYPAwED4+PggPj4eu3btQnd3t71Do3uwYsUKKBSKQf/cunXL3mGSGf/85z/xzjvvYMWKFYiLi4O7uzsUCgVycnKGbHv8+HGkpaVh3LhxUKlUmDZtGn7605+ivb3dBpHTUEayb7du3Trker5y5YoNZ0GmdHd3o7S0FK+++iqSkpIQEBAADw8PhIaGYuHChThy5Mig7a2xdvkcOxezdu1a7N69G+7u7pg3bx58fX3x8ccfY8OGDSgqKkJxcTFUKpW9w6R7kJycjClTppgsc3Nzs3E0ZKlf//rX2L1797Dbvf3221i/fj0UCgUee+wxhISE4OTJk3jjjTdw8OBBlJeXY9y4cVaImCw10n0LAPHx8XjooYdMllnrzQVkuU8++QQLFiwAAISGhiIlJQU+Pj74/PPPUVRUhKKiIjz//PP4zW9+A4VCYdTWamtXyGUcPnxYAIivr698+umnhu03b96UuLg4ASCvvPKKHSOke5GVlSUAJC8vz96h0Aj87ne/kx//+Meyd+9e+eKLL2TZsmUCQH7+85+bbVNRUSEKhULc3Nzk6NGjhu0dHR0yf/58ASCZmZm2CJ8GMZJ9m52dLQAkOzvbdoHSsJWWlkpmZqb87W9/G1BWUFAgbm5uAkD++Mc/GpVZc+0ysXMhSUlJAkBycnIGlJ08eVIAiFKplJaWFjtER/eKiZ1z0e/PwX78lyxZIgDkueeeG1BWU1MjY8aMEQDyxRdfWDNUGiZL9i0TO+ewcuVKASDz58832m7Ntctr7FzEV199ZXj33bPPPjugPCUlBZGRkejq6sLRo0dtHR4RDdPt27cN1++YWtMajQbJyckAgMOHD9s0NiLql5CQAACor683bLP22uU1di7i/PnzAICgoCBMnDjRZJ0ZM2agvr4e58+fxzPPPGPL8GgUlZWV4eLFi2hra8PYsWMxc+ZMpKWl8YGnTuZf//oXOjs7AfSvXVNmzJiBkydPGtY/OZ6Kigps3LgRWq0WarUaCQkJeOqpp+Dn52fv0MgCV69eBQCEhYUZtll77TKxcxHV1dUAgKioKLN1IiMjjeqSY9qzZ8+AbWFhYfjDH/6A1NRUO0RE1qBfpwEBAWZ/5LmmHZ/+Avw7qdVq5ObmYvny5XaKiizR2NiI/Px8AEBmZqZhu7XXLk/Fuoi2tjYAgI+Pj9k6vr6+APrfZUeOJz4+Hrt378alS5eg0+lw48YNFBcX49FHH0VDQwMWLlyIEydO2DtMGiVc085t8uTJeOONN3D+/HlotVpotVqUl5cjPT0dra2tyMrKwt69e+0dJpnR09ODpUuXorW1FXFxcXjhhRcMZdZeuzxiR+Qk1q1bZ/R3Pz8/LFiwAE888QQyMjLwl7/8BWvXrkVlZaV9AiQiiy1btmzAtuTkZBQVFWHNmjV45513sG7dOixZsgSenp52iJAGs2rVKpSWlmLs2LH44IMPbLqPeMTORegP93Z0dJito38gor+/v01iIttQKBTYtm0bAOCzzz4zuoiXHBfXtOvaunUr3NzccPPmTZw5c8be4dBdXn75Zbz77rsIDAxESUkJHnjgAaNya69dJnYuIjo6GgAG/VHXl+nrkvN48MEHDZ+vX79ux0hotOjXaUtLi+HUzt24pp1TUFAQxo8fD4Dr+X7zyiuvIDc3FwEBASguLjbcFXsna69dJnYuQv8fV3Nzs9mLMc+dOwcASExMtFlcZBvNzc2Gz7ybzjnExMTA29sbwP/W7t24pp1Tb28vWltbAXA9309ee+01/PKXv4RarUZxcbHZO16tvXaZ2LmIiIgIJCUlAQD27ds3oLy8vBz19fVQKpVIS0uzdXhkZQUFBQD6D+vHxMTYORoaDZ6ennjyyScBmF7TtbW1OH36NAAgIyPDprGRdRUWFqKzsxMKhcJs8kC2tXHjRvziF7+AWq1GSUmJ4ffWFGuvXSZ2LuQnP/kJAGDnzp2oqKgwbG9ubsZLL70EAPjRj37E9w86oMrKShQWFqKnp8doe19fH959913Dvl+zZg08PDzsESJZwcaNG6FQKJCXl4e//vWvhu2dnZ1YuXIlent7kZmZiWnTptkxShquuro6vP/++7h169aAsg8//BDPPfccAOAHP/gBQkNDbR0e3WXz5s148803ERAQMGRSp2fNtasQERl2K3JYL7/8MnJzc+Hh4YH58+fDx8cHpaWlaGlpQXJyMkpKSqBSqewdJg3Thx9+iIyMDAQGBiIxMREhISFoaWnBpUuXUFdXBwB45plnsGfPHri782b4+1FFRYXhH1gA8OWXX6KpqQkREREIDw83bD98+LDRw07vfJH4448/jvHjx+PkyZNoaGhATEzMyF8kTqNmuPu2srISCQkJ8PX1RUJCAsLDw/Hf//4Xn3/+ueGBt3PnzkVhYaHhsRhkH4WFhfjud78LoP+hwt/85jdN1hs3bhzeeusto21WW7sjfP0ZObD9+/fL7Nmzxd/fX1QqlcTGxsrOnTulq6vL3qHRCF27dk3Wrl0rKSkpEh4eLl5eXqJUKiUqKkoWL14sR44csXeINISysjIBMOSf6urqAW1LSkokNTVVgoKCRKlUytSpU2XTpk2i0+lsPxEaYLj7tqmpSTZs2CDz5s2TqKgo8fHxEQ8PDwkLC5P09HTZt2+f9Pb22ndSJCIieXl5Fu1bjUZjsr011i6P2BERERE5CV5jR0REROQkmNgREREROQkmdkREREROgokdERERkZNgYkdERETkJJjYERERETkJJnZEREREToKJHREREZGTYGJHRERE5CSY2BGRy1ixYgUUCgVWrFhh71DsqqamBgqFAgqFAjU1NfYOh4hGEd8GTkQuLz8/HzU1NZgzZw7mzJlj73DuydatWwH0J7HR0dF2jYWIbI+JHRG5jLCwMMTExCAsLMxoe35+Pj755BMAcPjEbtu2bQD652EusfPw8EBMTIzhMxE5DyZ2ROQyduzYgR07dtg7DLsLDw/HlStX7B0GEVkBr7EjIiIichJM7IjIZdx980R+fj4UCoXhNOy2bdsMNxUMdnPBqVOnsHTpUmg0Gnh5eUGtVmPmzJl488030d7ePuTYIoLf//73SElJwdixY6FQKJCfn2+o+/e//x0bNmzAY489ZhgjICAAs2bNMjuGvn+9uXPnGs3jztOyltw80draiu3btyMxMRH+/v5QqVSYOnUqXnzxRVy7ds3sd6zv98SJE2hra8PmzZsxbdo0qFQqjB07Funp6Thz5ozZ9kR0b3gqlohclkqlQkhICLRaLbq7u+Hj4wNfX1+jOm5ubobPfX19WLduHXJzcw3bfH190dHRgbNnz+Ls2bPIy8vDsWPHoNFoTI4pIliyZAkOHjyIMWPGQK1WY8wY439jP/LII4bP3t7e8Pb2xn/+8x+cOXMGZ86cwZ49e1BWVobx48cb6qnVaoSEhODGjRsAgMDAQHh6ehrKg4ODLf5eLl++jNTUVFy/fh0A4OXlBQ8PD1RVVaGqqgp5eXnYu3cvMjMzzfbR0NCAxMREVFVVwcvLC2PGjIFWq8WRI0dQUlKCoqIifOtb37I4JiKykBARuYisrCwBIFlZWUbbH3/8cQEg2dnZg7bfvHmzAJDx48fLr371K2lubhYRkdu3b0tZWZkkJCQIAElMTJTe3l6TY/v6+oq7u7u89dZb0traKiIibW1t8vXXXxvqPvXUU7J//35paGgwbOvs7JRDhw5JTEyMAJCMjAyTMQIQAFJWVmZ2HtXV1YZ61dXVRmU6nU4mTpwoACQ8PFyOHDlimEtlZaXMmjVLAIhSqZTKykqz4wcGBso3vvEN+fjjj6W3t1f6+vrkH//4hyF+jUYz4DsionvHxI6IXMa9JHbV1dXi5uYmKpXKZEIj0p8URURECAA5fPiwybEBSG5u7ojncP36dVEqlaJQKKS2tnZA+b0mdjt37hQA4uHhIRcvXhzQVqfTSXR0tACQJ5980uz4wcHBcuPGjQHlFy5cMNQpLy8fesJENCy8xo6IyAL5+fno7e1Famoq4uPjTdbx8/PD9773PQDAsWPHTNYJDAzECy+8MOI4wsPDER8fDxHB6dOnR9yPOfv37wcALF68GLGxsQPK/fz88NprrwEAPvroI7S2tprs5/nnnzc6VawXFxeHiRMnAgAuXLgwWmET0f/Ha+yIiCxw6tQpAEBxcTFCQ0PN1tPf2FBbW2uyPCkpyejaN1P6+vpQUFCAgoICVFZW4ubNm7h169aAevpr4EbL7du3DcnWE088YbbeggULDHFWVFRg7ty5A+o8/PDDZttPmDAB1dXV0Gq19xgxEd2NiR0RkQW+/vprAEBHRwc6OjqGrN/Z2Wlyu6mjWHe3S09PR1lZmWGbp6cngoKCDA8T1t/sYUkcw6HVatHb2wug/8igOREREYbP//73v03W8fPzM9ve3b3/p6e7u3skYRLRIHgqlojIAvqEZ8OGDZD+65MH/XPixAmT/dx5l60pr7/+OsrKyqBSqfD222+jtrYWt27dQnNzMxobG9HY2Gg4GiYiozpHInJ8TOyIiCygP/1q7hTraCkoKAAA/OxnP8PatWsRFRVl9Hw6AGhsbLTK2EFBQYbEc7DTvHeWDXUEkohsi4kdEbk8/XPkBjsClpycDAA4fvy4yevdRkt9fT0AICEhwWR5TU0NqqqqzLbXJ4EjOZrn6emJ6dOnAwBKS0vN1jt+/DiA/u8tMTFx2OMQkfUwsSMil+fv7w8AaGlpMVvnhz/8Idzd3dHU1ITs7OxB+7t9+7bZN1AMRa1WAwA+++wzk+UbN24ctL0lcxnM97//fQDABx98gEuXLg0ob29vx65duwAAaWlphniJ6P7AxI6IXJ7+sR5Hjx7FV199ZbLO5MmTsWXLFgDArl27sHz5cqPEp6enB5WVldi+fTumTJmCysrKEcWSmpoKAMjJycGhQ4fQ09MDAKiursazzz6LAwcOIDAwcMi57N271+wNHIN58cUXMXHiRHR3d+M73/kOPvroI/T19QEALl68iG9/+9uorq6GUqlETk7OsPsnIutiYkdELi8rKwteXl6oqqpCVFQUQkNDER0djejoaKPrybZs2YItW7ZAoVDgvffeQ1xcHLy9vTFu3Dh4eXkhISEB2dnZqK+vH3BdnKVycnIQEhKCtrY2ZGZmQqVSISAgAJMmTcKf/vQnvP7664bTpaasWrUKAHDw4EEEBAQgIiIC0dHRSElJsWh8Pz8/FBYWIjw8HNevX0daWhp8fHygVqsxffp0nD59GkqlEu+//77Z5/kRkf0wsSMilzd16lSUlZVh4cKFCA4ORnNzM2pra1FbW2s4Ygb0X7+2fft2XLhwAS+99BIefPBBuLm5obW1FYGBgXj00Ufx6quv4vTp04Zr8oZLo9Hg3LlzWLlyJSZMmACg/12t6enpOHbsGDZt2jRo+6VLl+K9995DSkoKvL290dDQgNra2mE98y42NhaXL1/G1q1b8dBDD8Hd3R1dXV2YPHkyVq1ahcuXL2Px4sUjmh8RWZdCeL88ERERkVPgETsiIiIiJ8HEjoiIiMhJMLEjIiIichJM7IiIiIicBBM7IiIiIifBxI6IiIjISTCxIyIiInISTOyIiIiInAQTOyIiIiInwcSOiIiIyEkwsSMiIiJyEkzsiIiIiJwEEzsiIiIiJ8HEjoiIiMhJ/D+eXfh+f46zpQAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sm.plot(plots=[\"test_mse\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "py314",
   "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": "3.14.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}