{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "with open(\"/app/suno/data/dpo/30b_t6_v20/30b_t6_bt16_cached_loss.json\", \"r\") as fp:\n",
    "    data = json.load(fp)\n",
    "\n",
    "len(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "82972"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(data[\"train\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>z_loss</th>\n",
       "      <th>semantic_0</th>\n",
       "      <th>coarse_0</th>\n",
       "      <th>orig_idx</th>\n",
       "      <th>coarse_1</th>\n",
       "      <th>coarse_2</th>\n",
       "      <th>coarse_3</th>\n",
       "      <th>coarse_4</th>\n",
       "      <th>coarse_5</th>\n",
       "      <th>coarse_6</th>\n",
       "      <th>coarse_7</th>\n",
       "      <th>coarse_8</th>\n",
       "      <th>coarse_9</th>\n",
       "      <th>coarse_10</th>\n",
       "      <th>coarse_11</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>380.5000</td>\n",
       "      <td>0.964844</td>\n",
       "      <td>2.187500</td>\n",
       "      <td>0</td>\n",
       "      <td>2.968750</td>\n",
       "      <td>3.250000</td>\n",
       "      <td>3.296875</td>\n",
       "      <td>3.406250</td>\n",
       "      <td>3.328125</td>\n",
       "      <td>3.328125</td>\n",
       "      <td>3.359375</td>\n",
       "      <td>3.312500</td>\n",
       "      <td>3.500000</td>\n",
       "      <td>3.437500</td>\n",
       "      <td>3.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>520.6875</td>\n",
       "      <td>1.734375</td>\n",
       "      <td>3.218750</td>\n",
       "      <td>1</td>\n",
       "      <td>4.062500</td>\n",
       "      <td>4.531250</td>\n",
       "      <td>4.875000</td>\n",
       "      <td>4.843750</td>\n",
       "      <td>4.750000</td>\n",
       "      <td>4.812500</td>\n",
       "      <td>4.781250</td>\n",
       "      <td>4.687500</td>\n",
       "      <td>4.687500</td>\n",
       "      <td>4.687500</td>\n",
       "      <td>4.687500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>425.1250</td>\n",
       "      <td>0.683594</td>\n",
       "      <td>1.851562</td>\n",
       "      <td>2</td>\n",
       "      <td>2.656250</td>\n",
       "      <td>3.234375</td>\n",
       "      <td>3.375000</td>\n",
       "      <td>3.390625</td>\n",
       "      <td>3.281250</td>\n",
       "      <td>3.328125</td>\n",
       "      <td>3.203125</td>\n",
       "      <td>3.203125</td>\n",
       "      <td>3.250000</td>\n",
       "      <td>3.343750</td>\n",
       "      <td>3.359375</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>433.5000</td>\n",
       "      <td>0.878906</td>\n",
       "      <td>1.937500</td>\n",
       "      <td>3</td>\n",
       "      <td>2.890625</td>\n",
       "      <td>3.406250</td>\n",
       "      <td>3.609375</td>\n",
       "      <td>3.640625</td>\n",
       "      <td>3.609375</td>\n",
       "      <td>3.609375</td>\n",
       "      <td>3.531250</td>\n",
       "      <td>3.484375</td>\n",
       "      <td>3.515625</td>\n",
       "      <td>3.640625</td>\n",
       "      <td>3.609375</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>640.3750</td>\n",
       "      <td>3.734375</td>\n",
       "      <td>4.062500</td>\n",
       "      <td>4</td>\n",
       "      <td>5.812500</td>\n",
       "      <td>6.187500</td>\n",
       "      <td>6.625000</td>\n",
       "      <td>6.718750</td>\n",
       "      <td>6.687500</td>\n",
       "      <td>6.656250</td>\n",
       "      <td>6.625000</td>\n",
       "      <td>6.593750</td>\n",
       "      <td>6.593750</td>\n",
       "      <td>6.593750</td>\n",
       "      <td>6.625000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     z_loss  semantic_0  coarse_0  orig_idx  coarse_1  coarse_2  coarse_3  \\\n",
       "0  380.5000    0.964844  2.187500         0  2.968750  3.250000  3.296875   \n",
       "1  520.6875    1.734375  3.218750         1  4.062500  4.531250  4.875000   \n",
       "2  425.1250    0.683594  1.851562         2  2.656250  3.234375  3.375000   \n",
       "3  433.5000    0.878906  1.937500         3  2.890625  3.406250  3.609375   \n",
       "4  640.3750    3.734375  4.062500         4  5.812500  6.187500  6.625000   \n",
       "\n",
       "   coarse_4  coarse_5  coarse_6  coarse_7  coarse_8  coarse_9  coarse_10  \\\n",
       "0  3.406250  3.328125  3.328125  3.359375  3.312500  3.500000   3.437500   \n",
       "1  4.843750  4.750000  4.812500  4.781250  4.687500  4.687500   4.687500   \n",
       "2  3.390625  3.281250  3.328125  3.203125  3.203125  3.250000   3.343750   \n",
       "3  3.640625  3.609375  3.609375  3.531250  3.484375  3.515625   3.640625   \n",
       "4  6.718750  6.687500  6.656250  6.625000  6.593750  6.593750   6.593750   \n",
       "\n",
       "   coarse_11  \n",
       "0   3.500000  \n",
       "1   4.687500  \n",
       "2   3.359375  \n",
       "3   3.609375  \n",
       "4   6.625000  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "df = pd.DataFrame.from_dict(data[\"train\"], orient=\"index\")\n",
    "\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss_discount_facs = np.concatenate(\n",
    "    [\n",
    "        np.linspace(1.0, 1, 1) * 4,\n",
    "        np.linspace(1.0, 0.5, 12),\n",
    "    ],\n",
    "    axis=0,\n",
    ")\n",
    "loss_discount_facs = loss_discount_facs / loss_discount_facs.sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss_discount_map = {}\n",
    "for n in range(1):\n",
    "    loss_discount_map[f\"semantic_{n}\"] = loss_discount_facs[n]\n",
    "for n in range(12):\n",
    "    n2 = n + 1\n",
    "    loss_discount_map[f\"coarse_{n}\"] = loss_discount_facs[n2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>z_loss</th>\n",
       "      <th>semantic_0</th>\n",
       "      <th>coarse_0</th>\n",
       "      <th>orig_idx</th>\n",
       "      <th>coarse_1</th>\n",
       "      <th>coarse_2</th>\n",
       "      <th>coarse_3</th>\n",
       "      <th>coarse_4</th>\n",
       "      <th>coarse_5</th>\n",
       "      <th>coarse_6</th>\n",
       "      <th>coarse_7</th>\n",
       "      <th>coarse_8</th>\n",
       "      <th>coarse_9</th>\n",
       "      <th>coarse_10</th>\n",
       "      <th>coarse_11</th>\n",
       "      <th>loss</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>380.5000</td>\n",
       "      <td>0.964844</td>\n",
       "      <td>2.187500</td>\n",
       "      <td>0</td>\n",
       "      <td>2.968750</td>\n",
       "      <td>3.250000</td>\n",
       "      <td>3.296875</td>\n",
       "      <td>3.406250</td>\n",
       "      <td>3.328125</td>\n",
       "      <td>3.328125</td>\n",
       "      <td>3.359375</td>\n",
       "      <td>3.312500</td>\n",
       "      <td>3.500000</td>\n",
       "      <td>3.437500</td>\n",
       "      <td>3.500000</td>\n",
       "      <td>2.504097</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>520.6875</td>\n",
       "      <td>1.734375</td>\n",
       "      <td>3.218750</td>\n",
       "      <td>1</td>\n",
       "      <td>4.062500</td>\n",
       "      <td>4.531250</td>\n",
       "      <td>4.875000</td>\n",
       "      <td>4.843750</td>\n",
       "      <td>4.750000</td>\n",
       "      <td>4.812500</td>\n",
       "      <td>4.781250</td>\n",
       "      <td>4.687500</td>\n",
       "      <td>4.687500</td>\n",
       "      <td>4.687500</td>\n",
       "      <td>4.687500</td>\n",
       "      <td>3.646962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>425.1250</td>\n",
       "      <td>0.683594</td>\n",
       "      <td>1.851562</td>\n",
       "      <td>2</td>\n",
       "      <td>2.656250</td>\n",
       "      <td>3.234375</td>\n",
       "      <td>3.375000</td>\n",
       "      <td>3.390625</td>\n",
       "      <td>3.281250</td>\n",
       "      <td>3.328125</td>\n",
       "      <td>3.203125</td>\n",
       "      <td>3.203125</td>\n",
       "      <td>3.250000</td>\n",
       "      <td>3.343750</td>\n",
       "      <td>3.359375</td>\n",
       "      <td>2.334845</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>433.5000</td>\n",
       "      <td>0.878906</td>\n",
       "      <td>1.937500</td>\n",
       "      <td>3</td>\n",
       "      <td>2.890625</td>\n",
       "      <td>3.406250</td>\n",
       "      <td>3.609375</td>\n",
       "      <td>3.640625</td>\n",
       "      <td>3.609375</td>\n",
       "      <td>3.609375</td>\n",
       "      <td>3.531250</td>\n",
       "      <td>3.484375</td>\n",
       "      <td>3.515625</td>\n",
       "      <td>3.640625</td>\n",
       "      <td>3.609375</td>\n",
       "      <td>2.562445</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>640.3750</td>\n",
       "      <td>3.734375</td>\n",
       "      <td>4.062500</td>\n",
       "      <td>4</td>\n",
       "      <td>5.812500</td>\n",
       "      <td>6.187500</td>\n",
       "      <td>6.625000</td>\n",
       "      <td>6.718750</td>\n",
       "      <td>6.687500</td>\n",
       "      <td>6.656250</td>\n",
       "      <td>6.625000</td>\n",
       "      <td>6.593750</td>\n",
       "      <td>6.593750</td>\n",
       "      <td>6.593750</td>\n",
       "      <td>6.625000</td>\n",
       "      <td>5.455310</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     z_loss  semantic_0  coarse_0  orig_idx  coarse_1  coarse_2  coarse_3  \\\n",
       "0  380.5000    0.964844  2.187500         0  2.968750  3.250000  3.296875   \n",
       "1  520.6875    1.734375  3.218750         1  4.062500  4.531250  4.875000   \n",
       "2  425.1250    0.683594  1.851562         2  2.656250  3.234375  3.375000   \n",
       "3  433.5000    0.878906  1.937500         3  2.890625  3.406250  3.609375   \n",
       "4  640.3750    3.734375  4.062500         4  5.812500  6.187500  6.625000   \n",
       "\n",
       "   coarse_4  coarse_5  coarse_6  coarse_7  coarse_8  coarse_9  coarse_10  \\\n",
       "0  3.406250  3.328125  3.328125  3.359375  3.312500  3.500000   3.437500   \n",
       "1  4.843750  4.750000  4.812500  4.781250  4.687500  4.687500   4.687500   \n",
       "2  3.390625  3.281250  3.328125  3.203125  3.203125  3.250000   3.343750   \n",
       "3  3.640625  3.609375  3.609375  3.531250  3.484375  3.515625   3.640625   \n",
       "4  6.718750  6.687500  6.656250  6.625000  6.593750  6.593750   6.593750   \n",
       "\n",
       "   coarse_11      loss  \n",
       "0   3.500000  2.504097  \n",
       "1   4.687500  3.646962  \n",
       "2   3.359375  2.334845  \n",
       "3   3.609375  2.562445  \n",
       "4   6.625000  5.455310  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[\"loss\"] = df.apply(\n",
    "    lambda row: sum([v * row[k] for k, v in loss_discount_map.items()]),\n",
    "    axis=1,\n",
    ")\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkQAAAHHCAYAAABeLEexAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAACRjklEQVR4nOzdeVxUZdvA8d/MwAzDLrIrIrnvJpZRZhamppWmLablkksLZmqpWbmXmrtmPVqZS2nZYpuluWuammFqKqEiisrmBsg6w8x5//BlcgQROMgi1/f5zOfpbNe5zoAzF/e57/toFEVREEIIIYSowrTlnYAQQgghRHmTgkgIIYQQVZ4UREIIIYSo8qQgEkIIIUSVJwWREEIIIao8KYiEEEIIUeVJQSSEEEKIKk8KIiGEEEJUeVIQCSGEEKLKk4JIiDJSu3Zt+vfvX95p3PZmzpzJHXfcgU6no2XLljfcr3///tSuXbvM8lJj27ZtaDQatm3bZltXUP7p6ekMGjQIf39/NBoNw4cPByApKYknn3yS6tWro9FomDdvXpnlLkRlIQWRECWwbNkyNBoNf/31V4Hb27dvT9OmTVWf59dff2XixImq41QVGzZsYPTo0dx3330sXbqUqVOnlndKZWrq1KksW7aMl19+mc8//5znn38egBEjRvDbb78xduxYPv/8czp37lzOmQpR8TiUdwJCVBXR0dFotcX7G+TXX3/lww8/lKKoiLZs2YJWq2XJkiXo9fryTueW+uSTT7BarXbrtmzZwj333MOECRPyre/WrRtvvPFGWaYoRKUiLURClBGDwYCjo2N5p1EsGRkZ5Z1CsSQnJ2M0Gm/7YgjA0dERg8Fgty45ORlPT898+95ofUnl5uZiMplKLZ4QFYEUREKUkev7EJnNZiZNmkS9evVwcnKievXqtG3blo0bNwJX+4h8+OGHAGg0GtsrT0ZGBq+//jpBQUEYDAYaNGjArFmzUBTF7rxZWVkMGzYMb29v3NzcePzxxzl37hwajcau5WnixIloNBqOHj1K7969qVatGm3btgXg0KFD9O/fnzvuuAMnJyf8/f154YUXuHjxot258mIcO3aM5557Dg8PD3x8fBg3bhyKonDmzBm6deuGu7s7/v7+zJ49u0jvXW5uLlOmTKFOnToYDAZq167NW2+9RU5Ojm0fjUbD0qVLycjIsL1Xy5YtK1L84r6nGzdupG3btnh6euLq6kqDBg1466237Pb54IMPaNKkCc7OzlSrVo3WrVuzatWqm+Zw9uxZunfvjouLC76+vowYMcLuOvNc24cor49RbGwsv/zyi931azQaFEXhww8/zPc7lJKSwvDhw23XW7duXd5//327lqdTp06h0WiYNWsW8+bNs/0Mjh49CsC///7Lk08+iZeXF05OTrRu3ZqffvrJLte8PHbt2sXIkSPx8fHBxcWFJ554gvPnz+e7tnXr1vHAAw/g5uaGu7s7d911V773bu/evXTu3BkPDw+cnZ154IEH2LVr103fXyFuRG6ZCaFCamoqFy5cyLfebDbf9NiJEycybdo0Bg0axN13301aWhp//fUX+/fv5+GHH+bFF18kPj6ejRs38vnnn9sdqygKjz/+OFu3bmXgwIG0bNmS3377jVGjRnHu3Dnmzp1r27d///58/fXXPP/889xzzz1s376drl273jCvp556inr16jF16lRbIbBx40ZOnjzJgAED8Pf358iRI3z88cccOXKEPXv22H3JAjzzzDM0atSI6dOn88svv/Duu+/i5eXF4sWLeeihh3j//fdZuXIlb7zxBnfddRft2rUr9L0aNGgQy5cv58knn+T1119n7969TJs2jaioKL7//nsAPv/8cz7++GP+/PNPPv30UwDuvffem/4civueHjlyhEcffZTmzZszefJkDAYDJ06csPsy/uSTTxg2bBhPPvkkr732GtnZ2Rw6dIi9e/fSu3fvG+aQlZVFeHg4cXFxDBs2jMDAQD7//HO2bNlSaO6NGjXi888/Z8SIEdSsWZPXX38dgDvvvNPWl+jhhx+mb9++tmMyMzN54IEHOHfuHC+++CK1atXijz/+YOzYsSQkJOTreL106VKys7MZMmQIBoMBLy8vjhw5wn333UeNGjV48803cXFx4euvv6Z79+589913PPHEE3YxXn31VapVq8aECRM4deoU8+bNY+jQoaxevdq2z7Jly3jhhRdo0qQJY8eOxdPTk7///pv169fb3rstW7bwyCOPEBoayoQJE9BqtSxdupSHHnqI33//nbvvvrvQ90uIAilCiGJbunSpAhT6atKkid0xwcHBSr9+/WzLLVq0ULp27VroeSIiIpSC/pn+8MMPCqC8++67duuffPJJRaPRKCdOnFAURVEiIyMVQBk+fLjdfv3791cAZcKECbZ1EyZMUADl2WefzXe+zMzMfOu+/PJLBVB27NiRL8aQIUNs63Jzc5WaNWsqGo1GmT59um395cuXFaPRaPeeFOTAgQMKoAwaNMhu/RtvvKEAypYtW2zr+vXrp7i4uBQa79p9g4ODbctFfU/nzp2rAMr58+dvGLtbt275fv5FMW/ePAVQvv76a9u6jIwMpW7dugqgbN269Yb5K8rV37GCfqcAJSIiwm7dlClTFBcXF+XYsWN26998801Fp9MpcXFxiqIoSmxsrAIo7u7uSnJyst2+4eHhSrNmzZTs7GzbOqvVqtx7771KvXr1bOvy/r106NBBsVqttvUjRoxQdDqdkpKSoiiKoqSkpChubm5KmzZtlKysLLtz5R1ntVqVevXqKZ06dbKLlZmZqYSEhCgPP/xwvusXoijklpkQKnz44Yds3Lgx36t58+Y3PdbT05MjR45w/PjxYp/3119/RafTMWzYMLv1r7/+OoqisG7dOgDWr18PwCuvvGK336uvvnrD2C+99FK+dUaj0fbf2dnZXLhwgXvuuQeA/fv359t/0KBBtv/W6XS0bt0aRVEYOHCgbb2npycNGjTg5MmTN8wFrl4rwMiRI+3W57WC/PLLL4UeX1RFfU/z+uL8+OOP+To15/H09OTs2bPs27ev2DkEBATw5JNP2tY5OzszZMiQYsUpim+++Yb777+fatWqceHCBdurQ4cOWCwWduzYYbd/z5498fHxsS1funSJLVu28PTTT3PlyhXb8RcvXqRTp04cP36cc+fO2cUYMmSIXWvi/fffj8Vi4fTp08DVlsgrV67w5ptv4uTkZHds3nEHDhzg+PHj9O7dm4sXL9rOm5GRQXh4ODt27Ljhz0WIwsgtMyFUuPvuu2ndunW+9XlfMoWZPHky3bp1o379+jRt2pTOnTvz/PPPF6mYOn36NIGBgbi5udmtb9SokW173v9rtVpCQkLs9qtbt+4NY1+/L1z98ps0aRJfffUVycnJdttSU1Pz7V+rVi27ZQ8PD5ycnPD29s63/vp+SNfLu4brc/b398fT09N2rWoV9T195pln+PTTTxk0aBBvvvkm4eHh9OjRgyeffNI2inDMmDFs2rSJu+++m7p169KxY0d69+7Nfffdd9Mc6tatm+8WZIMGDUrlGq91/PhxDh06ZFfkXOv6n/P1vxcnTpxAURTGjRvHuHHjbhijRo0atuXrfy+qVasGwOXLlwGIiYkBKHTKirw/IPr163fDfVJTU22xhSgqKYiEKCft2rUjJiaGH3/8kQ0bNvDpp58yd+5cFi1aZNfCUtaubQ3K8/TTT/PHH38watQoWrZsiaurK1arlc6dOxf417hOpyvSOiBfh+Ubub5IKC9Go5EdO3awdetWfvnlF9avX8/q1at56KGH2LBhAzqdjkaNGhEdHc3atWtZv3493333HR999BHjx49n0qRJ5X0JAFitVh5++GFGjx5d4Pb69evbLV//e5H3c3/jjTfo1KlTgTGuL2LV/g5ce96ZM2fecOJNV1fXIscTIo8UREKUIy8vLwYMGMCAAQNIT0+nXbt2TJw40VYQ3agICA4OZtOmTVy5csWuRePff/+1bc/7f6vVSmxsLPXq1bPtd+LEiSLnePnyZTZv3sykSZMYP368bX1JbvWVRN41HD9+3NZaA1dnX05JSbFda2mcpyjvKYBWqyU8PJzw8HDmzJnD1KlTefvtt9m6dSsdOnQAwMXFhWeeeYZnnnkGk8lEjx49eO+99xg7dmy+20HX5nD48GEURbH72UdHR5fKNV6rTp06pKen2/ItrjvuuAO4Ovy/pDEKygng8OHDN2zFzNvH3d291M4rBMiweyHKzfW3ilxdXalbt67dEGsXFxfg6vDoa3Xp0gWLxcLChQvt1s+dOxeNRsMjjzwCYPvL/aOPPrLb74MPPihynnl/1V//V3xZPf6hS5cuBZ5vzpw5AIWOmCvueYrynl66dCnfsXktFXk/u+t/tnq9nsaNG6MoSqEjELt06UJ8fDzffvutbV1mZiYff/xxia6pME8//TS7d+/mt99+y7ctJSWF3NzcQo/39fWlffv2LF68mISEhHzbCxpOfzMdO3bEzc2NadOmkZ2dbbct7/cvNDSUOnXqMGvWLNLT00vlvEKAtBAJUW4aN25M+/btCQ0NxcvLi7/++otvv/2WoUOH2vYJDQ0FYNiwYXTq1AmdTkevXr147LHHePDBB3n77bc5deoULVq0YMOGDfz4448MHz7c9ld0aGgoPXv2ZN68eVy8eNE27P7YsWNA0W5Dubu7065dO2bMmIHZbKZGjRps2LCB2NjYW/Cu5NeiRQv69evHxx9/TEpKCg888AB//vkny5cvp3v37jz44IOlcp6ivqeTJ09mx44ddO3aleDgYJKTk/noo4+oWbOmbd6mjh074u/vz3333Yefnx9RUVEsXLiQrl275uujdK3BgwezcOFC+vbtS2RkJAEBAXz++ec4OzuXyjVea9SoUfz00088+uij9O/fn9DQUDIyMvjnn3/49ttvOXXqVL4+X9f78MMPadu2Lc2aNWPw4MHccccdJCUlsXv3bs6ePcvBgweLlZO7uztz585l0KBB3HXXXbb5sA4ePEhmZibLly9Hq9Xy6aef8sgjj9CkSRMGDBhAjRo1OHfuHFu3bsXd3Z2ff/5ZzVsjqqryGt4mRGWWN4x43759BW5/4IEHbjrs/t1331XuvvtuxdPTUzEajUrDhg2V9957TzGZTLZ9cnNzlVdffVXx8fFRNBqN3RD8K1euKCNGjFACAwMVR0dHpV69esrMmTPthiIrytVh2xEREYqXl5fi6uqqdO/eXYmOjlYAu2HweUPmCxpOfvbsWeWJJ55QPD09FQ8PD+Wpp55S4uPjbzh0//oYNxoOX9D7VBCz2axMmjRJCQkJURwdHZWgoCBl7NixdsO9CztPQQoatl6U93Tz5s1Kt27dlMDAQEWv1yuBgYHKs88+azd8ffHixUq7du2U6tWrKwaDQalTp44yatQoJTU19aZ5nT59Wnn88ccVZ2dnxdvbW3nttdeU9evXl/qw+7zrHTt2rFK3bl1Fr9cr3t7eyr333qvMmjXL9nuYN+x+5syZBeYbExOj9O3bV/H391ccHR2VGjVqKI8++qjy7bff2va50b+XrVu35rsuRVGUn376Sbn33nsVo9GouLu7K3fffbfy5Zdf2u3z999/Kz169LC9x8HBwcrTTz+tbN68ucA8hbgZjaIUozebEOK2cODAAe68806++OIL+vTpU97pCCFEuZM+RELc5rKysvKtmzdvHlqt9qYzRAshRFUhfYiEuM3NmDGDyMhIHnzwQRwcHFi3bh3r1q1jyJAhBAUFlXd6QghRIcgtMyFucxs3bmTSpEkcPXqU9PR0atWqxfPPP8/bb7+Ng4P8TSSEECAFkRBCCCGE9CESQgghhJCCSAghhBBVnnQgKAKr1Up8fDxubm4V5nlKQgghhCicoihcuXKFwMBA28OXb0QKoiKIj4+X0ThCCCFEJXXmzBlq1qxZ6D5SEBVB3lT7Z86cwd3dvZyzEUJc70DiAR5Y+gDbB2ynpX9LFYEOcOCBV3mA39m+HW7wMPUiuXLlAAcOPEDLlttxc1MRSAhRYmlpaQQFBRX6yJw8UhAVQd5tMnd3dymIhKiAGmgbMPvx2TQIbIC7q4p/ow0a0GDiy8xGQ4MGbqj55+7k1IDmzWfj49MAvV4+N4QoT0Xp7iLD7osgLS0NDw8PUlNTpSASQgghKonifH/LKDMhRKV3Oesy3xz5hstZl1UGuszlpT/wzdJ0LqsMZTZfJjn5G8xmlYGEEGVCCiIhRKUXmxLL098+TWxKrMpAscS+MJmnX3AlVmWo7OxYjh59muxslYGEEGVC+hAJISo8i8WC2Wy+4XbFrBDsEoxiVsjOzi75iRQFJdifYLJRFFATymRS0GqDMZlU5iSEKJRer7/pkPqikD5ERSB9iIQoH4qikJiYSEpKSqH7mSwmEq4kEOAWgF6nL/kJTSZMCRdJIICAANCrCGW1mjCZEtDrA9BqVQQSQhRKq9USEhKCvoB/sMX5/pYWIiFEhZVXDPn6+uLs7HzDkSKZpkzMKWaCPINw1juX/ISZmWSadZgJISgInFWEslgyyc424+QUhE6nIpAQ4obyJk5OSEigVq1aqiZPloJICFEhWSwWWzFUvXr1QvdVdArOOc4YjUacHJ1KflJFQXFywhk9RqMWJxWhLBYFcMbJyYhOpyKQEKJQPj4+xMfHk5ubi6OjY4njSEEkhKiQ8voMORehmcboaKSxT2P1JzUaMTatSylEQqcz4uJSGpGEEIXJu1VmsVhUFUQyykwIUaHJ8wOFEIUprc8IKYiEEJVepimTyPhIMk2ZKgNlkhkZRWSkQqbKUBZLJleuRGKxqAwkhCgTUhAJISo95Zr/qQukoCgKiqKhpONv27dvz/DhwwHlmlf5+S+fq2rXrs28efNsy4mJiTz88MO4uLjg6el5w3VC3O6kD5EQotKZu/GY3bLZYuJ8Zjo+zrE4qhl2bzZhPp/KeY7hkwyOjjDi4foqs61Y9u3bh4uLi2157ty5JCQkcODAATw8PG64TojbnRREQghRhfj4+Ngtx8TEEBoaSr169QpdV1wmk6nAeWGEqKjklpkQQtxCly9fpm/fvlSrVg1nZ2ceeeQRjh8/btt++vRpHnvsMapVq4aLiwtNmjTh119/tR3bp08ffHx8MBqN1KtXj6VLl97wXBkZGfTt2xdXV1cCAgKYPXt2vn2uvWVWu3ZtvvvuO1asWIFGo6F///4FrgNISUlh0KBB+Pj44O7uzkMPPcTBgwdtcSdOnEjLli359NNPCQkJwen/5ywo6nGff/45tWvXxsPDg169enHlyhXbPlarlRkzZlC3bl0MBgO1atXivffes20/c+YMTz/9NJ6ennh5edGtWzdOnTpV9B+SEEgLkRDiNuCgdcTXxQedRuVHmoMjDj6e+KKgc1A3ckWrNeLs3ITnnnuG48dP8NNPP+Hu7s6YMWPo0qULR48exdHRkYiICEwmEzt27MDFxYWjR4/i6uoKwLhx4zh69Cjr1q3D29ubEydOkJWVdcNzjho1iu3bt/Pjjz/i6+vLW2+9xf79+2nZsmWB++/bt4++ffvi7u7O/PnzMRqNmEymfOsAnnrqKYxGI+vWrcPDw4PFixcTHh7OsWPH8PLyAuDEiRN89913rFmzBp1OV+TjYmJi+OGHH1i7di2XL1/m6aefZvr06baiZ+zYsXzyySfMnTuXtm3bkpCQwL///gtcnZ6hU6dOhIWF8fvvv+Pg4MC7775L586dOXTokLRSiSKTgkgIUelpNBocNCWff+SaQGRaNYAFF5XFlUaj5eTJs/z008/s2rWLe++9F4CVK1cSFBTEDz/8wFNPPUVcXBw9e/akWbNmANxxxx22GHFxcdx55520bt0auNqicyPp6eksWbKEL774gvDwcACWL19OzZo1b3iMj48PBoMBo9GIv7+/bf3163bu3Mmff/5JcnIyBoMBgFmzZvHDDz/w7bffMmTIEODqbbIVK1bYbssV9Tir1cqyZctwc3MD4Pnnn2fz5s289957XLlyhfnz57Nw4UL69esHQJ06dWjbti0Aq1evxmq18umnn9qGXy9duhRPT0+2bdtGx44dC/kpCfEfKYiEEJWexZrLFVM6bnpXdFoVH2uWXBzS07mCG04OoNNBUtrVB7P6uRdvtmmrNYcDB7bh4OBAmzZtbOurV69OgwYNiIqKAmDYsGG8/PLLbNiwgQ4dOtCzZ0+aN28OwMsvv0zPnj3Zv38/HTt2pHv37rbC6noxMTGYTCa7c3l5edGgQYNi5V2QgwcPkp6enm/G8KysLGJiYmzLwcHBdn2Uinpc7dq1bcUQQEBAAMnJyQBERUWRk5NjK/IKyu3EiRN2xwNkZ2fbnUOIm5GCSAhR6VkVK5nmTFwcndGpCmRFm2PGhA6r9WpBVFKKkovVmnbT/QYNGkSnTp345Zdf2LBhA9OmTWP27Nm8+uqrPPLII5w+fZpff/2VjRs3Eh4eTkREBLNmzSp5YiWQnp5OQEAA27Zty7ft2mH5145eK85x188urNFosFqtALZbdoXlFhoaysqVK/Ntu74DuRCFkU7VQogqLyMnl4ycXLJMllKNW79+CLm5uezdu9e27uLFi0RHR9O48X+P9QgKCuKll15izZo1vP7663zyySe2bT4+PvTr148vvviCefPm8fHHHxd4rjp16uDo6Gh3rsuXL3Ps2LEC9y+OVq1akZiYiIODA3Xr1rV7eXt7l/px16pXrx5Go5HNmzff8BzHjx/H19c33zlkygBRHFIQCSHELVK3bi0ef/xRBg8ezM6dOzl48CDPPfccNWrUoFu3bgAMHz6c3377jdjYWPbv38/WrVtp1KgRAOPHj+fHH3/kxIkTHDlyhLVr19q2Xc/V1ZWBAwcyatQotmzZwuHDh+nfvz9arfqP+Q4dOhAWFkb37t3ZsGEDp06d4o8//uDtt9/mr7/+KvXjruXk5MSYMWMYPXo0K1asICYmhj179rBkyRIA+vTpg7e3N926deP3338nNjaWbdu2MWzYMM6ePav62kXVIbfMhBDiFlqy5H+MHPkWjz76KCaTiXbt2vHrr7/abhNZLBYiIiI4e/Ys7u7udO7cmblz5wJXH1o5duxYTp06hdFo5P777+err7664blmzpxJeno6jz32GG5ubrz++uukpqaqvgaNRsOvv/7K22+/zYABAzh//jz+/v60a9cOPz+/Uj/ueuPGjcPBwYHx48cTHx9PQEAAL730EnD14b87duxgzJgx9OjRgytXrlCjRg3Cw8Nxd3dXfe2i6tAoSkknqK860tLS8PDwIDU1Vf6BCVFGsrOziY2NtZvT5kZMFhPJGcn4uviiL8FM1Xkdp7VmM46XUknW+FK7lh69Xk2nahNmczKOjr5otTL0W4hbpbDPiuJ8f0sLkRCi0tPr9NR0v/Hw8qKyOjqS4+eNB1bUTl+j1eoxGNTnJIQoG1IQCSEqPYvVQqY5E2dHZ3Takg8N01gtaDOzycSFBGs213a/yWspgqK1FimKBYslE53OGY1G1dg3IUQZkE7VQohKLzs3m+iL0WTnZt9850LoTCaM8UkkxzuRa1I3U7XVmk1WVjRWq7qchBBlQwoiIYQQQlR5UhAJIYQQosqTgkgIIYQQVZ50qhZCVHoajQZHraPt4Z4lpWg0KDodOqygLhSgQaNxpBQCCSHKgBREQohKz9nRmRb+LVTHsRicyKgTTCAm1bF0OmdcXdXnJIQoG3LLTAghhBBVnhREQohKL9OcycHEg2SaM1XF0eVk4xJzmvgYPaYcdbe6LJZM0tMPYrGoy0kIUTbKtSDasWMHjz32GIGBgWg0Gn744Ycb7vvSSy+h0WiYN2+e3fpLly7Rp08f3N3d8fT0ZODAgaSnp9vtc+jQIe6//36cnJwICgpixowZt+BqhBDlRVEUzFYzap9EpFEUNBYLFosWShiqf//+aDQaXnnlVRTFzLWBIiIi0Gg09O/fX1WeZSk7O5uIiAiqV6+Oq6srPXv2JCkpqdBjFEVh/PjxBAQEYDQa6dChA8ePH7fb59ixY3Tr1g1vb2/c3d1p27YtW7duvZWXIkShyrUgysjIoEWLFnz44YeF7vf999+zZ88eAgMD823r06cPR44cYePGjaxdu5YdO3YwZMgQ2/a0tDQ6duxIcHAwkZGRzJw5k4kTJ/Lxxx+X+vUIIQRAUFAQq1d/R1bWf5MyZmdns2rVKmrVqlWOmRXfiBEj+Pnnn/nmm2/Yvn078fHx9OjRo9BjZsyYwYIFC1i0aBF79+7FxcWFTp06kZ393/vx6KOPkpuby5YtW4iMjKRFixY8+uijJCYm3upLEqJgSgUBKN9//32+9WfPnlVq1KihHD58WAkODlbmzp1r23b06FEFUPbt22dbt27dOkWj0Sjnzp1TFEVRPvroI6VatWpKTk6ObZ8xY8YoDRo0KHJuqampCqCkpqYW/8KEECWSlZWlHD16VMnKyrrpvuk56cq+c/uU9Jz0Ep0rMTVLSUzNUi4kXVLS9x1R9u1TlLikbNv661+F6devn9KtWzeladPGyiefTFZyc6/mtHLlSqV58+ZKt27dlH79+tn2t1gsytSpU5XatWsrTk5OSvPmzZVvvvnGtj03N1d54YUXbNvr16+vzJs3r8Bzzpw5U/H391e8vLyUV155RTGZTCV6P/KkpKQojo6OdvlERUUpgLJ79+4Cj7FarYq/v78yc+ZMuzgGg0H58ssvFUVRlPPnzyuAsmPHDts+aWlpCqBs3LhRVc6i6inss6I4398Vug+R1Wrl+eefZ9SoUTRp0iTf9t27d+Pp6Unr1q1t6zp06IBWq2Xv3r22fdq1a4f+mic1durUiejoaC5fvnzrL0IIUSX179+XlSt/ti1/9tlnDBgwIN9+06ZNY8WKFSxatIgjR44wYsQInnvuObZv3w5c/RysWbMm33zzDUePHmX8+PG89dZbfP3113Zxtm7dSkxMDFu3bmX58uUsW7aMZcuW2bZPnDiR2rVrF+saIiMjMZvNdOjQwbauYcOG1KpVi927dxd4TGxsLImJiXbHeHh40KZNG9sx1atXp0GDBqxYsYKMjAxyc3NZvHgxvr6+hIaGFitHIUpLhR52//777+Pg4MCwYcMK3J6YmIivr6/dOgcHB7y8vGzNromJiYSEhNjt4+fnZ9tWrVq1fHFzcnLIycmxLaelpam6DiFEKUtIuPr6f05WCw1dnXDyd4LsbDh6NP8xrVpd/f/oaMjIsNuk8fJH8fLCmpaK6WIiNUjFIQO0GrD6+2P1Dyh2is8/34+3357AmTNJaDQ6du3axVdffcW2bdts++Tk5DB16lQ2bdpEWFgYAHfccQc7d+5k8eLFPPDAAzg6OjJp0iTbMSEhIezevZuvv/6ap59+2ra+WrVqLFy4EJ1OR8OGDenatSubN29m8ODBAHh7e1OnTp1iXUNiYiJ6vR5PT0+79X5+fje8tZW3Pu9ztqBjNBoNmzZtonv37ri5uaHVavH19WX9+vUFfiYLURYqbEEUGRnJ/Pnz2b9/v+rJ1opr2rRpdh9AQogKZvFiuObfqA5w7dMHvvgCzsZCQa0MeR2u+/eHPXvsNhk+/ozsZ57F8OMPuL8xwm5b+ptvkzH2nWKn6OvrT9euXVm+/HMURaFr1654e3vb7XPixAkyMzN5+OGH7dabTCbuvPNO2/KHH37IZ599RlxcHFlZWZhMJlq2bGl3TJMmTdDpdLblgIAA/vnnH9vy0KFDGTp06A3znTp1KlOnTrUtHy2oqCwliqIQERGBr68vv//+O0ajkU8//ZTHHnuMffv2ERBQ/AJUCLUqbEH0+++/k5ycbNcB0WKx8PrrrzNv3jxOnTqFv78/ycnJdsfl5uZy6dIl/P39AfD39883IiJvOW+f640dO5aRI0faltPS0ggKCiqV6xJClIIXX4THH7ctmi1mLhgsVLeY0NesCZGRNz522bJ8LUQ5Xlc/C0yPduNKnQZc0nhhdFPQOShYb/A5cTNWq4nnn3+cESMmAJoCB4/kjYj95ZdfqFGjht02g8EAwFdffcUbb7zB7NmzCQsLw83NjZkzZ9q6BeRxdHS0W9ZoNFit1iLn+9JLL9m1OAUGBuLv74/JZCIlJcWulSgpKemGn59565OSkuwKm6SkJFsRt2XLFtauXcvly5dxd3cH4KOPPmLjxo0sX76cN998s8h5C1FaKmxB9Pzzz9vdg4arfX+ef/552334sLAwUlJSiIyMtN133rJlC1arlTZt2tj2efvttzGbzbYPjI0bN9KgQYMbNs0aDAbbh5EQogIKCLj6+n8mUwbnLkThbjGjd3L57/ZYQRo0yLdKSbs6+klTzRNtzWDO0xi/WjnonUo+jF9RzLRv3xCTyYRGo6VTp0759mncuDEGg4G4uDgeeOCBAuPs2rWLe++9l1deecW2LiYmpsR53YiXlxdeXl5260JDQ3F0dGTz5s307NkTgOjoaOLi4my3+K4XEhKCv78/mzdvthVAaWlp7N27l5dffhmAzMyrczNptfbdWLVabbGKOCFKU7kWROnp6Zw4ccK2HBsby4EDB/Dy8qJWrVpUr17dbn9HR0f8/f1p8P8faI0aNaJz584MHjyYRYsWYTabGTp0KL169bIN0e/duzeTJk1i4MCBjBkzhsOHDzN//nzmzp1bdhcqhKiSdDodhw9HotO52N3OyuPm5sYbb7zBiBEjsFqttG3bltTUVHbt2oW7uzv9+vWjXr16rFixgt9++42QkBA+//xz9u3bl69v5M0sXLiQ77//ns2bNxf5GA8PDwYOHMjIkSPx8vLC3d2dV199lbCwMO655x7bfg0bNmTatGk88cQTaDQahg8fzrvvvku9evUICQlh3LhxBAYG0r17d+DqH6rVqlWjX79+jB8/HqPRyCeffEJsbCxdu3Yt1nUJUVrKtSD666+/ePDBB23Lebep+vXrZzc6ojArV65k6NChhIeHo9Vq6dmzJwsWLLBt9/DwYMOGDURERBAaGoq3tzfjx4+3m6tICCFuFXd3d3Q6lxtunzJlCj4+PkybNo2TJ0/i6elJq1ateOuttwB48cUX+fvvv3nmmWfQaDQ8++yzvPLKK6xbt65YeVy4cKFELUtz5861fbbm5OTQqVMnPvroI7t9oqOjSU1NtS2PHj2ajIwMhgwZQkpKCm3btmX9+vU4OTkBVzt4r1+/nrfffpuHHnoIs9lMkyZN+PHHH2nRQp7/JsqHRlFUTu1aBaSlpeHh4UFqaqrtfrcQ4tbKzs4mNjaWkJAQ2xfpjWSYMoi6EEUj70a46G9cfNxI0v/fMnPIzsIpLoGom9wy83MvPB8AiyWDzMwonJ0bFVoQCSHUKeyzojjf3xW2D5EQQhSVg9YBb2dvHLTqPtKsOh25bi64YkKb/w5XsWg0Djg4eKPRyMesEJWB/EsVQlR6BgcDtT1rq45jddSTHeBHNdR37NVqDRiN6nMSQpSNCj1TtRBCFIXVaiXLnKV+hJLVijY7B3M2qA2lKFYsliwURUZNCVEZSEEkhKj0snKzOHL+CFm5WariOJhycI47R2KcE7kmdRPCWq1ZZGYewWpVl5MQomxIQSSEEEKIKk8KIiGEEEJUeVIQCSGEEKLKk1FmQohKT3PN/9QH06Ch8OnZ8uYtKnw+Is01LyFERScFkRCi0nPWOxMaWMAT7osp18lIer0QapKjOpZO54ybm/qchBBlQ26ZCSFEJdG+fXuGDx9e3mmUu2XLluHp6Wlbnjhxou1BskKUlBREQohKL8ucxdHzR8kyF2+Ie1Jatu32F4AuJwfnU2dIOuWAOadkt7qmTZvGXXfdhZubGz4+1enW7TGio6Pt9mnfvj0ajcbu9dJLL9m2b9u2DY1GQ0pKSolyuJ3Url2befPm2a175plnOHbsWLnkExcXR9euXXF2dsbX15dRo0aRm5tbpGNzcnJo2bIlGo2GAwcO2G07dOgQ999/P05OTgQFBTFjxoxbkL0ojNwyE0JUelbFSqY5E6vKSRA1ihWtyYwJBxTFUqIY27dvJyIiglatmpCefpx3311Bx44dOXr0KC4u/z3TbPDgwUyePNm27OzsrCr3sqQoChaLBQeH8vkKMRqNGI3GMj+vxWKha9eu+Pv788cff5CQkEDfvn1xdHRk6tSpNz1+9OjRBAYGcvDgQbv1aWlpdOzYkQ4dOrBo0SL++ecfXnjhBTw9PeVB5GVIWoiEEKIUrV+/nv79+9OkSWOaNavPZ58tIi4ujsjISLv9nJ2d8ff3t73yHjx56tQpHnzwQQCqVauGRqOhf//+tuOsViujR4/Gy8sLf39/Jk6cWGg+/fv3p3v37kyaNAkfHx/c3d156aWXMJlMdjGnTZtGSEgIRqORFi1a8O2339q257VYrVu3jtDQUAwGAzt37sRqtTJjxgzq1q2LwWCgVq1avPfee7bjzpw5w9NPP42npydeXl5069aNU6dO5ctt1qxZBAQEUL16dSIiIjCbzcDVlrTTp08zYsQIW0sa5L9lVpBPP/2URo0a4eTkRMOGDfnoo48K3b8oNmzYwNGjR/niiy9o2bIljzzyCFOmTOHDDz+0ez8Lsm7dOjZs2MCsWbPybVu5ciUmk4nPPvuMJk2a0KtXL4YNG8acOXNU5yyKTgoiIYS4hVJT0wDw8vKyW79y5Uq8vb1p2rQpY8eOJTMzE4CgoCC+++47AKKjo0lISGD+/Pm245YvX46Liwt79+5lxowZTJ48mY0bNxaaw+bNm4mKimLbtm18+eWXrFmzhkmTJtm2T5s2jRUrVrBo0SKOHDnCiBEjeO6559i+fbtdnDfffJPp06cTFRVF8+bNGTt2LNOnT2fcuHEcPXqUVatW4efnB4DZbKZTp064ubnx+++/s2vXLlxdXencubNd8bB161ZiYmLYunUry5cvZ9myZSxbtgyANWvWULNmTSZPnkxCQgIJCQlFes9XrlzJ+PHjee+994iKimLq1KmMGzeO5cuX2/Zp3769XaFZFLt376ZZs2a2awTo1KkTaWlpHDly5IbHJSUlMXjwYD7//PMCWwJ3795Nu3bt0Ov1dnGjo6O5fPlysXIUJSe3zIQQlU7ClQQS0v/7cswyZ3Eh8wKNvBuRnZvN0fNH8x3TKqAVANEXoskwZwBwMf3qaLIgj2CqOXlxIesCF1KPEks2yedNOOoV/Fz88XMNKFGeVquVkSPHcN9999G0aVPb+t69exMcHExgYCCHDh1izJgxREdHs2bNGnQ6na148vX1zdcS0rx5cyZMmABAvXr1WLhwIZs3b+bhhx++YR56vZ7PPvsMZ2dnmjRpwuTJkxk1ahRTpkzBbDYzdepUNm3aRFhYGAB33HEHO3fuZPHixTzwwAO2OJMnT7ad58qVK8yfP5+FCxfSr18/AOrUqUPbtm0BWL16NVarlU8//dTWsrN06VI8PT3Ztm0bHTt2BK62gi1cuBCdTkfDhg3p2rUrmzdvZvDgwXh5eaHT6XBzc8Pf37/I7/uECROYPXs2PXr0ACAkJISjR4+yePFiW661atUiIKB4P9fExES7YgiwLScmJhZ4jKIo9O/fn5deeonWrVvbtZBdGzckJOSGcatVq1asPEXJSEEkhKh0FkcuZtL2SXbrnmz8JF3rd+VUyilCP84/3F2ZcHVuof4/9mfP2T122xZ2/ownGz3L97E/M3bn63bbXr/nbUaFvVPsHDUaA6NH/48jR6LYuXOn3bZr+4U0a9aMgIAAwsPDiYmJoU6dOoXGbd68ud1yQEAAycnJhR7TokULu5aJsLAw0tPTOXPmDOnp6WRmZuYrqEwmE3feeafdutatW9v+OyoqipycHMLDwws858GDBzlx4gRubm5267Ozs4mJibEtN2nSBJ1OZ3c9//zzT6HXU5iMjAxiYmIYOHAggwcPtq3Pzc3Fw8PDtrxixYpC4zzyyCP8/vvvAAQHBxfaAlSYDz74gCtXrjB27NgSHS/KjhREQohK58XQF3m8weN266o5VcNB60BN95pEDom8wZGwrNsy4i5fslsX5BEMwGMNnyK0RpjdNj+XordMXGvYsOH8+usmduzYQc2aNQvdt02bNgCcOHHipgWRo6Oj3bJGo8FqLXln8vT0dAB++eUXatSoYbfNYDDYLV/bKfxmnZrT09MJDQ1l5cqV+bb5+PjY/vtWXc8nn3xie1/zXFt43cynn35KVlaWXY7+/v78+eefdvslJSXZthVky5Yt7N69O9972bp1a/r06cPy5cvx9/e3xSlqXFH6pCASQlQ6AW4BBLj9d7vDbDFzMesiZosZJwcn2+2xgjTwboCnPrvAbb76agTqHLik8cLooaArwSekoii8+uqrfP/992zY8A3BwYUXQ4BtCHbeLZy8viQWS8lGul3v4MGDZGVl2YqYPXv24OrqSlBQEF5eXhgMBuLi4uxuj91MvXr1MBqNbN68mUGDBuXb3qpVK1avXo2vr6+tw3hJ6PX6Yr0Pfn5+BAYGcvLkSfr06VPi815fHMLVlrX33nuP5ORkfH19Adi4cSPu7u40bty4wDgLFizg3XfftS3Hx8fTqVMnVq9ebSvYwsLCePvttzGbzbbia+PGjTRo0EBul5Uh6VQthKj0TBYTZ9POYrIUPtLnZrS5ZvSXLnP5ogFLbsnmIYqIiOCLL77g88+XYDCkkZBwhsTERFtrQ0xMDFOmTCEyMpJTp07x008/0bdvX9q1a2e7HRYcHIxGo2Ht2rWcP3/e1upRUiaTiYEDB3L06FF+/fVXJkyYwNChQ9Fqtbi5ufHGG28wYsQIli9fTkxMDPv37+eDDz6w64R8PScnJ8aMGcPo0aNZsWIFMTEx7NmzhyVLlgDQp08fvL296datG7///juxsbFs27aNYcOGcfbs2SLnXrt2bXbs2MG5c+e4cOFCkY6ZNGkS06ZNY8GCBRw7dox//vmHpUuX2o3a6tu3b7FvY3Xs2JHGjRvz/PPPc/DgQX777TfeeecdIiIibC1Af/75Jw0bNuTcuXPA1b5KTZs2tb3q168PXO1vlddy2Lt3b/R6PQMHDuTIkSOsXr2a+fPnM3LkyGLlJ9SRFiIhhChF//vf/wAID3/Ebv3SpUvp378/er2eTZs2MW/ePDIyMggKCqJnz568885//ZRq1KjBpEmTePPNNxkwYAB9+/a1jbwqifDwcOrVq0e7du3Iycnh2WeftRuuP2XKFHx8fJg2bRonT57E09OTVq1a8dZbbxUad9y4cTg4ODB+/Hji4+MJCAiwTTDp7OzMjh07GDNmDD169ODKlSvUqFGD8PDwYrUYTZ48mRdffJE6deqQk5ODohT+nDmAQYMG4ezszMyZMxk1ahQuLi40a9bMbpbvuLg4tNritQnodDrWrl3Lyy+/TFhYGC4uLvTr189uPqnMzEyio6NtUwcUhYeHBxs2bCAiIoLQ0FC8vb0ZP368zEFUxjRKUX67qri0tDQ8PDxITU1V1fQrhCi67OxsYmNjCQkJwcmpsIeoQoYpg6gLUTTyboSL3qXQfQG72amv5ZCdhVNcAlE0xq9WDnqnwj8eC3u4q8WSQWZmFM7OjdDpbp7TrdK/f39SUlL44Ycfyi0HIW6lwj4rivP9LbfMhBBCCFHlyS0zIUSlp9Pq8DB4oNMWfRRRQRStDouzE0bMFPNuSj4ajQ6dzgONRl1OQoiyIQWREKLSc3Jwol71eqrjWPR6smoG4o360V1arRPOzupzUktN3yMhqhK5ZSaEqPSsihWzxaz64a4oChpzLtZcBbW9KxXFitVqRlGbkxCiTEhBJISo9LLMWRxMOkiWOUtVHIecbFxi4zh30og5p2TD7vNYrVlkZBzEalWXkxCibEhBJIQQJZSUln3DEWtCiMpFCiIhhBBCVHlSEAkhhBCiypOCSAghhBBVnhREQohKz9nRmTv978TZ0VlVnFyDE+l1gqlRJxtHg7phZlqtM66ud6LVqsvpWu3bt7d7/ERVtWzZMjw9PW3LEydOpGXLlqrjXrx4EV9fX06dOqU6VnFt27YNjUZDSkoKkP8aS8OFCxfw9fUt1rPkqhIpiIQQlZ5Go0Gn1aHRqBsZhkYDOh1a3dX/LImJEyei0WjQarVotQ5otVoaNmxot092djYRERFUr14dV1dXevbsSVJSkm379V+OVVnt2rWZN2+e3bpnnnmGY8eOlfq53nvvPbp160bt2rVLPXZx3Ypr9Pb2pm/fvkyYMKFU4imKwvjx4wkICMBoNNKhQweOHz9+0+M+/PBDateujZOTE23atOHPP/+0296+fXs0Go3dK+8ZebeSFERCiEovOzebYxePkZ2rbsSXzpSD8Uw8F87oMJtKXlw1adKEs2djiYnZydmzsezcudNu+4gRI/j555/55ptv2L59O/Hx8fTo0UNV7mVJURRyc3PL7fxGoxFfX99SjZmZmcmSJUsYOHBgqcYtqVtxjQADBgxg5cqVXLp0SXWsGTNmsGDBAhYtWsTevXtxcXGhU6dOZGff+N/h6tWrGTlyJBMmTGD//v20aNGCTp06kZycbLff4MGDSUhIsL1mzJihOt+bkYJICFHpWawW0nLSsFjVzTCtsVrRZWWTleWImvkUHRwc8Pf3wcfHgL+/D97e3rZtqampLFmyhDlz5vDQQw8RGhrK0qVL+eOPP9izZw+nTp3iwQcfBKBatWpoNBr69+9vO95qtTJ69Gi8vLzw9/e3e2p9Qfr370/37t2ZNGkSPj4+uLu789JLL2EymexiTps2jZCQEIxGIy1atODbb7+1bc9rsVq3bh2hoaEYDAZ27tyJ1WplxowZ1K1bF4PBQK1atXjvvfdsx505c4ann34aT09PvLy86Natm93tqLzcZs2aRUBAANWrVyciIsL2pPj27dtz+vRpRowYYWspgKLdTvr0009p1KgRTk5ONGzYkI8++qjQ/X/99VcMBgP33HNPvuv+7bffuPPOOzEajTz00EMkJyezbt06GjVqhLu7O7179yYzM7PI72fe+erXr4/RaOTBBx/Md5vu+muMiYmhW7du+Pn54erqyl133cWmTZvsjqlduzZTp07lhRdewM3NjVq1avHxxx/b7dOkSRMCAwP5/vvvC30/bkZRFObNm8c777xDt27daN68OStWrCA+Pr7QBwnPmTOHwYMHM2DAABo3bsyiRYtwdnbms88+s9vP2dkZf39/26ssHqwuBZEQQpSy48ePExRUl+bNu/H88y8QFxdn2xYZGYnZbKZDhw62dQ0bNqRWrVrs3r2boKAgvvvuOwCio6NJSEhg/vz5tn2XL1+Oi4sLe/fuZcaMGUyePJmNGzcWms/mzZuJiopi27ZtfPnll6xZs4ZJkybZtk+bNo0VK1awaNEijhw5wogRI3juuefYvn27XZw333yT6dOnExUVRfPmzRk7dizTp09n3LhxHD16lFWrVuHn5weA2WymU6dOuLm58fvvv7Nr1y5cXV3p3LmzXTG2detWYmJi2Lp1K8uXL2fZsmW2x42sWbOGmjVrMnnyZFtLQVGsXLmS8ePH89577xEVFcXUqVMZN24cy5cvv+Exv//+O6GhoQVumzhxIgsXLuSPP/6wFXnz5s1j1apV/PLLL2zYsIEPPvigyO/nmTNn6NGjB4899hgHDhxg0KBBvPnmm4VeU3p6Ol26dGHz5s38/fffdO7cmccee8zudwtg9uzZtG7dmr///ptXXnmFl19+mejoaLt97r77bn7//Xfb8rJly4p9uzk2NpbExES732MPDw/atGnD7t27CzzGZDIRGRlpd4xWq6VDhw75jlm5ciXe3t40bdqUsWPH2hWct4wibio1NVUBlNTU1PJORYgqIysrSzl69KiSlZWVb1t8vKJERv732rknU/lh90ElPSddycqy35b3yvPvv4qyYXu23SsqNktJTM1S/j1yWdn5eYzy+eeK8svGHGXD9mzlYPTVbYW9rvXrr78qX3/9tbJ//x5lzZoFyj33tFFq1aqlpKWlKYqiKCtXrlT0en2+a7rrrruU0aNHK4qiKFu3blUA5fLly3b7PPDAA0rbtm3zHTdmzJgbvo/9+vVTvLy8lIyMDNu6//3vf4qrq6tisViU7OxsxdnZWfnjjz/sjhs4cKDy7LPP2uXzww8/2LanpaUpBoNB+eSTTwo87+eff640aNBAsVqttnU5OTmK0WhUfvvtN1tuwcHBSm5urm2fp556SnnmmWdsy8HBwcrcuXPtYi9dulTx8PCwLU+YMEFp0aKFbblOnTrKqlWr7I6ZMmWKEhYWVmCuiqIo3bp1U1544QW7dXnXvWnTJtu6adOmKYASExNjW/fiiy8qnTp1UhRFKdL7OXbsWKVx48Z228eMGWP3M7/+GgvSpEkT5YMPPrAtBwcHK88995xt2Wq1Kr6+vsr//vc/u+NGjBihtG/f3ra8Zs0apUGDBoWe63q7du1SACU+Pt5u/VNPPaU8/fTTBR5z7tw5Bcj33owaNUq5++67bcuLFy9W1q9frxw6dEj54osvlBo1aihPPPHEDXMp7LOiON/f8nBXIUSls3gxXNPAARjp3KMGHb6Es2ehoD/0855N1r8/7NljsNu28GMTTz5j5cef9IwZe8f/r9UD8PqbuYwaW/T+Mo888ggAFksGdevqaNeuB3fc0Zivv/66VPqnNG/e3G45ICAgX/+L67Vo0QJn5/9Gu4WFhZGens6ZM2dIT08nMzOThx9+2O4Yk8nEnXfeabeudevWtv+OiooiJyeH8PDwAs958OBBTpw4gZubm9367OxsYmJibMtNmjRBp9PZXc8///xT6PUUJiMjg5iYGAYOHMjgwYNt63Nzc/Hw8LjhcVlZWTg5ORW47dr33M/PD2dnZ+644w67dXkdg0+cOHHT9zMqKoo2bdrYbQ8LCyv0utLT05k4cSK//PILCQkJ5ObmkpWVla+F6NpcNRoN/v7++X4/jEajXYvLE088wRNPPHHDc69cuZIXX3zRtrxu3Tq7n1lpGzJkiO2/mzVrRkBAAOHh4cTExFCnTp1bdt5yLYh27NjBzJkziYyMJCEhge+//57u3bsDV5tb33nnHX799VdOnjyJh4cHHTp0YPr06QQGBtpiXLp0iVdffZWff/4ZrVZLz549mT9/Pq6urrZ9Dh06REREBPv27cPHx4dXX32V0aNHl/XlCiFKyYsvwuOP/7ecazGjcVbQ6/TUrAmRkTc+dtkyiEvKsVsXVOtqtfRoDyt3Nz1PKh44uShodeDnX7Lh9xqNHoOhFi4u1ahfvz4nTpwAwN/fH5PJREpKil0fkaSkJPz9/W8a19HR8brzaLBaS97hKT09HYBffvmFGjVq2G0zGOwLRxcXF9t/G43Gm8YNDQ1l5cqV+bb5+PjY/vtWXc8nn3ySr+go7Evc29uby5cvF7jt2hw1Gk2hORfn/SyON954g40bNzJr1izq1q2L0WjkySeftLv9eH2u1+eW59KlS3Y/g5t5/PHH7d7LGjVq2G5fJiUlERAQYNuWlJR0wykQvL290el0diMq844p7Hc/79wnTpy4fQuijIwMWrRowQsvvJBvhEVmZib79+9n3LhxtGjRgsuXL/Paa6/x+OOP89dff9n269OnDwkJCWzcuBGz2cyAAQMYMmQIq1atAiAtLY2OHTvSoUMHFi1axD///MMLL7yAp6enXRUqhKg8AgKuvv7jCFztuOyog1atbnxsgwbgGVBwkVPd34Hq/m6A+ifUa7WO6PW+pKenExMTw/PPPw9AaGgojo6ObN68mZ49ewJX+wrFxcXZWgn0+qutUxaLuk7ieQ4ePEhWVpatiNmzZw+urq4EBQXh5eWFwWAgLi6OBx54oMgx69Wrh9FoZPPmzQwaNCjf9latWrF69Wp8fX1VdYjV6/XFeh/8/PwIDAzk5MmT9OnTp8jH3XnnnXzxxRclSdFO48aNb/p+NmrUiJ9++slu3Z49ewqNu2vXLvr3729ryUlPTy/xfEmHDx+mffv2Rd7fzc0tX0tfSEgI/v7+bN682VYApaWlsXfvXl5++eUC4+j1ekJDQ9m8ebOt8cNqtbJ582aGDh16w/MfOHAAwK7wuhXKtSB65JFHbM3L1/Pw8MjXUXDhwoXcfffdxMXFUatWLaKioli/fj379u2zNeV+8MEHdOnShVmzZhEYGMjKlSsxmUx89tln6PV6mjRpwoEDB5gzZ44URELcJnKtuaRmp+Lh5IGD9sYfazd7EKvGYsHhSjqpeGBwg5LcFXjjjTd47LHHCAqqwdmzx5k8eTY6nY5nn30WuPrZNnDgQEaOHImXlxfu7u68+uqrhIWF2UY4BQcHo9FoWLt2LV26dMFoNNq1eheXyWRi4MCBvPPOO5w6dYoJEyYwdOhQtFotbm5uvPHGG4wYMQKr1Urbtm1JTU1l165duLu7069fvwJjOjk5MWbMGEaPHo1er+e+++7j/PnzHDlyhIEDB9KnTx9mzpxJt27dmDx5MjVr1uT06dOsWbOG0aNHU7NmzSLlXrt2bXbs2EGvXr0wGAx2I/ZuZNKkSQwbNgwPDw86d+5MTk4Of/31F5cvX2bkyJEFHtOpUyfGjh3L5cuXqVatWpFyK0hR3s+XXnqJ2bNnM2rUKAYNGkRkZKStI/mN1KtXjzVr1vDYY4+h0WgYN25ciVrSMjMziYyMZOrUqbZ133//PWPHjuXff/8tchyNRsPw4cN59913qVevHiEhIYwbN47AwEBbsQMQHh7OE088YSt4Ro4cSb9+/WjdujV333038+bNIyMjgwEDBgBXR9OtWrWKLl26UL16dQ4dOsSIESNo165dvtvFpa1S9SFKTU1Fo9HYmpl3796Np6en3X3tDh06oNVq2bt3L0888QS7d++mXbt2tr+44Oov/vvvv3/DX/ycnBxycv5rUk9LS7t1FyWEUC0nN4fYlFgaeTfCQV/yjzWd2YQh+SIX8cPPKQedrvi3y86ePcuzzz7LxYsX8fb2pG3b+9mzZ4/dLYq5c+fabvHn5OTQqVMnu2HhNWrUYNKkSbz55psMGDCAvn373vQLszDh4eHUq1ePdu3akZOTw7PPPms3XH/KlCn4+Pgwbdo0Tp48iaenJ61ateKtt94qNO64ceNwcHBg/PjxxMfHExAQYJtAz9nZmR07djBmzBh69OjBlStXqFGjBuHh4cVqMZo8eTIvvvgiderUIScnB0W5+c9k0KBBODs7M3PmTEaNGoWLiwvNmjUrdJbvZs2a0apVK77++mu7/jIlcbP3s1atWnz33XeMGDGCDz74gLvvvts2XP5G5syZwwsvvMC9996Lt7c3Y8aMKdF3048//kitWrW4//77betSU1PzjUQritGjR5ORkcGQIUNISUmhbdu2rF+/3q4vVkxMDBcuXLAtP/PMM5w/f57x48eTmJhIy5YtWb9+vW10ol6vZ9OmTbZCKSgoiJ49e/LOO+8UO7/i0ihF+e0qAxqNxq4P0fWys7O57777aNiwoe2e9NSpU1m+fHm+H6Svry+TJk3i5ZdfpmPHjoSEhLB48WLb9qNHj9KkSROOHj1Ko0aN8p1r4sSJdkNS86SmppbJXAhCiKv/5mNjYwkJCblhZ9c8GaYMoi5E0ci7ES56lxvud7MWIofsLJziEoiiMX61ctA7Fe3j0c89f34WSwaZmVE4OzdCp7txTrda//79SUlJKXRuGHHVL7/8wqhRozh8+DBa7e05K80999zDsGHD6N27d3mnUmoK+6xIS0vDw8OjSN/fleInbjabefrpp1EUhf/973+3/Hxjx44lNTXV9jpz5swtP6cQovJKSsu+abElKr6uXbsyZMgQzp07V96p3BIXLlygR48ettu3wl6Fv2WWVwydPn2aLVu22FV4BQ0nzM3N5dKlS7Ye6/7+/gX2aM/bVhCDwaBqNIAQQojK6XZ+eK63t7eMsC5EhW4hyiuGjh8/zqZNm6hevbrd9rCwMFJSUoi8Zoztli1bsFqttmF6YWFh7NixwzYVPMDGjRtp0KCBqo5zQoiKQ6fR4eLogk6jbm4URavFatBjMOSiUf3pqEOrdQFu3XwtRbFs2TK5XSZEEZRrQZSens6BAwdsQ+piY2M5cOAAcXFxmM1mnnzySf766y9WrlyJxWIhMTGRxMRE27wLjRo1onPnzgwePJg///yTXbt2MXToUHr16mWbq6h3797o9XoGDhzIkSNHWL16NfPnz7/hSAMhROXj5OhEI59GODkW3tfoZix6A5nBNfENzsVRr657pU7nhItLI3Q6dTkJIcpGud4y++uvv2wPMQRsRUq/fv2YOHGibZ6G6yd52rp1q20OhZUrVzJ06FDCw8NtozYWLFhg29fDw4MNGzYQERFBaGgo3t7ejB8/XobcCyGEEMKmXAui9u3bFzqEsigD4Ly8vGyTMN5I8+bN7R5kJ4S4vRR1lNnNlHSUWUEqyigzIUTRVOg+REIIIYQQZUEKIiGEEEJUeVIQCSGEEKLKk4JICCEqifbt29/W8+QU1bJly2yPcIKrTxe40RPWhSgqKYiEEJWe0dFIU9+mGB2NquLk/v+w+4Dg7BIPu69duzYajQYHB1fc3e/CwcEVjUZDRESEbZ/27duj0WjsXnnPAAPYtm0bGo2GlJQUVddzO6hduzbz5s2zW/fMM89w7NixW37uS5cu0adPH9zd3fH09GTgwIGkp6cXuv+rr75KgwYNMBqN1KpVi2HDhpGamppv32XLltG8eXOcnJzw9fW1+/0Q5aPCz1QthBA3o9VocXIohfl+/n9iRjUfjPv27cNisdiWDx8+zMMPP8xTTz1lt9/gwYOZPHmybdnZ2VnFWcuWoihYLBYcHMrnK8RoNGI0qit+i6JPnz4kJCSwceNGzGYzAwYMYMiQITcc2RwfH098fDyzZs2icePGnD59mpdeeon4+Hi+/fZb235z5sxh9uzZzJw5kzZt2pCRkcGpU6du+fWIwkkLkRCi0svJzeHk5ZPk5OaoiqM1mTCeS+TSOS25Jk2JYvj4+ODv74+vbzU8PDL5+ecfqVOnDg888IDdfs7Ozvj7+9teeY8lOnXqlG1+tmrVqqHRaOjfv7/tOKvVyujRo/Hy8sLf39/uqfUF6d+/P927d2fSpEn4+Pjg7u7OSy+9ZJvgNi/mtGnTCAkJwWg00qJFC7sv8LwWq3Xr1hEaGorBYGDnzp1YrVZmzJhB3bp1MRgM1KpVi/fee8923JkzZ3j66afx9PTEy8uLbt262X3x5+U2a9YsAgICqF69OhEREbYnC7Rv357Tp08zYsQIW0sa5L9lVpBPP/2URo0a4eTkRMOGDfnoo48K3f96UVFRrF+/nk8//ZQ2bdrQtm1bPvjgA7766ivi4+MLPKZp06Z89913PPbYY9SpU4eHHnqI9957j59//pnc3FwALl++zDvvvMOKFSvo3bs3derUoXnz5jz++OPFyk+UPimIhBCVXq41l0tZl8i15qqKo7Va0GVkkpGhx2pVl5Oi5JKZmcTKlat44YUXbF/meVauXIm3tzdNmzZl7NixZGZmAhAUFMR3330HQHR0NAkJCcyfP9923PLly3FxcWHv3r3MmDGDyZMns3HjxkJz2bx5M1FRUWzbto0vv/ySNWvWMGnSJNv2adOmsWLFChYtWsSRI0cYMWIEzz33HNu3b7eL8+abbzJ9+nSioqJo3rw5Y8eOZfr06YwbN46jR4+yatUq/Pz8gKuPXurUqRNubm78/vvv7Nq1C1dXVzp37mxXjG3dupWYmBi2bt3K8uXLWbZsGcuWLQNgzZo11KxZk8mTJ5OQkEBCQkKR3vuVK1cyfvx43nvvPaKiopg6dSrjxo1j+fLltn3at29vV2heb/fu3Xh6etK6dWvbug4dOqDVatm7d2+R8gBsT1nPa03buHEjVquVc+fO0ahRI2rWrMnTTz8tDxGvAOSWmRCi0snJScBk+u/LMcuchdV8AWiExZJNZubRfMe4ubW6emz2MazWDLtten0wOgcvcnMvkGE5ioVscrJMWKwKDo7+ODoGlCjPtWu3kZKSmu+Lt3fv3gQHBxMYGMihQ4cYM2YM0dHRrFmzBp1Oh5eXFwC+vr75WkKaN2/OhAkTAKhXrx4LFy5k8+bNPPzwwzfMQ6/X89lnn+Hs7EyTJk2YPHkyo0aNYsqUKZjNZqZOncqmTZsICwsD4I477mDnzp0sXrzYrmVr8uTJtvNcuXKF+fPns3DhQvr16wdAnTp1aNu2LQCrV6/GarXy6aef2orBpUuX4unpybZt2+jYsSNwtRVs4cKF6HQ6GjZsSNeuXdm8eTODBw/Gy8sLnU6Hm5vbDR/GXZAJEyYwe/ZsevToAUBISAhHjx5l8eLFtlxr1apFQMCNf66JiYn4+vrarXNwcMDLy4vExMQi5XHhwgWmTJli92SEkydPYrVamTp1KvPnz8fDw4N33nmHhx9+mEOHDqHX64t8naJ0SUEkhKh04uMXc/r0JLt1Dm6dIaADOTlniYwMzXdM+/ZXO0knnB5MVsafdtsCgz/Do/qzpKb9QHzW1aeBn429us3b/218At8pUZ4rVvxE584dbc9WzHPtF2SzZs0ICAggPDycmJgY6tSpU2jM5s2b2y0HBASQnJxc6DEtWrSw66MUFhZGeno6Z86cIT09nczMzHwFlclk4s4777Rbd21rSVRUFDk5OYSHhxd4zoMHD3LixAnc3Nzs1mdnZxMTE2NbbtKkCTrdfw/ADQgI4J9//in0egqTkZFBTEwMAwcOZPDgwbb1ubm5eHh42JZXrFhR4nMURVpaGl27dqVx48Z2tzWtVitms5kFCxbYisIvv/wSf39/tm7dSqdOnW5pXuLGpCASQlQ6gYEv4u39X5+LLHMWp65cAMBgqEloaOQNjw0I/qTAFiIAD/fueKaGEMsdVPc34WC42kJUEqdPx7Ft2598+23hjxYCaNOmDQAnTpy4aUHk6Ohot6zRaLCquL+XN2rql19+oUaNGnbbDAaD3bKLy3+PILlZp+b09HRCQ0NZuXJlvm0+Pj62/75V1/PJJ5/Y3tc81xZeN+Pv75+v0MzNzeXSpUs3ba26cuUKnTt3xs3Nje+//97uGvNapRo3bmxb5+Pjg7e3N3FxcUXOT5Q+KYiEEJWOwRCAwfDf7Q6DxURNxws46hzR6fS222MFHutU/4bbtE7+OFZ3xkvjjbObgk7FJ+Ty5avw9fWma9fHbrrvgQMHgP++LPNum1w7Wk2NgwcPkpWVZSti9uzZg6urK0FBQXh5eWEwGIiLi8vX8bsw9erVw2g0snnzZgYNGpRve6tWrVi9ejW+vr62DuMlodfri/U++Pn5ERgYyMmTJ+nTp0+JzxsWFkZKSgqRkZGEhl5tcdyyZQtWqzVfoXWttLQ0OnXqhMFg4KeffsLJyX7043333Qdc7R9Ws2ZN4Opw/QsXLhAcHFzifIV60qlaCFHp6XV6At0C0evy979ISsu2vW7G6uBIjm913H3UFUNWq5Vly1bQr98A9Hr74fQxMTFMmTKFyMhITp06xU8//UTfvn1p166d7XZYcHAwGo2GtWvXcv78+ULnvikKk8nEwIEDOXr0KL/++isTJkxg6NChaLVa3NzceOONNxgxYgTLly8nJiaG/fv388EHH9h1Qr6ek5MTY8aMYfTo0axYsYKYmBj27NnDkiVLgKtD1r29venWrRu///47sbGxbNu2jWHDhnH27Nki5167dm127NjBuXPnuHDhQpGOmTRpEtOmTWPBggUcO3aMf/75h6VLlzJnzhzbPn379mXs2LE3jNGoUSM6d+7M4MGD+fPPP9m1axdDhw6lV69etlug586do2HDhvz559VbsGlpaXTs2JGMjAyWLFlCWloaiYmJJCYm2oq6+vXr061bN1577TX++OMPDh8+TL9+/WjYsKFtdKEoH9JCJISo9CxWC+mmdFz1rui0Rb8tcj2NxYIuM4t0XNE7Q0lDbdq0ibi4OPr2fQpFsaDR/BdIr9ezadMm5s2bR0ZGBkFBQfTs2ZN33vmvn1KNGjWYNGkSb775JgMGDKBv3762kVclER4eTr169WjXrh05OTk8++yzdv1apkyZgo+PD9OmTePkyZN4enrSqlUr3nrrrULjjhs3DgcHB8aPH098fDwBAQG2CSadnZ3ZsWMHY8aMoUePHly5coUaNWoQHh5erBajyZMn8+KLL1KnTh1ycnJQlJtPmDlo0CCcnZ2ZOXMmo0aNwsXFhWbNmtnN8h0XF4dWW3ibwMqVKxk6dCjh4eFotVp69uzJggULbNvNZjPR0dG2EYL79++3jUCrW7euXazY2Fhq164NXO2/NGLECLp27YpWq+WBBx5g/fr1+W4firKlUYry21XFpaWl4eHhYRs+KYS49bKzs4mNjSUkJCTfbYfrZZgyiLoQRSPvRrjoXey2FaVlKI9DdhZOcQlE0Ri/WjnonYr38ejn/l+eFksGmZlRODs3QqdzKeSoW6t///6kpKTwww8/lFsOQtxKhX1WFOf7W26ZCSGEEKLKk4JICCGEEFWe9CESQojbmJq+R0JUJVIQCSEqPa1Gi0FnQKtR1+itaDQoDjocsKApwaPMru2v5O2iRaMxIA3xQlQOUhAJISq0ooz7MDoaaebXTPW5LAYnMu4IJgCz6lg6nRFXV/U5CSEKV1pjw+RPFyFEhZQ3BDlvSLMQQhQk72HBxZmJvCDSQiSEqJB0Oh2enp62xyc4Ozvne2J8nixzFqdSTlHbszZGR/tHSphNOUU/pykHfcIFTlEbr4BcHPUl/8szIyOL7OxTODnVRqcr/DEXQoiSsVqtnD9/HmdnZxwc1JU0UhAJISqsvGdG3ezhpSaLiaQrSWhTtflmq07LKvrtL12uGcdLqZxHQ44lF51DyQuiVIOCyZSEXq9Fq5UnmAtxq2i1WmrVqnXDP5iKSgoiIUSFpdFoCAgIwNfXF7P5xoXNkeQjvLT+Jb57+jsa+Daw27ZsV2yRz1f99GnumPw/XmINA8afwy/YVOLcn2qZyZEjL9GkyXe4uDS4+QFCiBLR6/U3nXW8KKQgEkJUeDqdrtD+ARpHDaczTqNx1OSbqTbLWvR+BabsXDSnEzmNE2nZDrhbS/5wVb1eg9V6Gr0+f05CiIpHOlULIYQQosqTFiIhRKVXv3p9/njhD+pXr68qzuWatTn0/tuMJBqfmur+XjQa63PnnX9gNKrLSQhRNqQgEkJUeq56V8KCwlTHMRtduHxnM2oCoG5uEwcHVzw81OckhCgbcstMCFHpnU07y8jfRnI27ayqOK7nE2kwZxG/znUm5by6vxezs89y4sRIsrPV5SSEKBtSEAkhKr3kjGTm7plLckbhw/NvxjnlIv7rt7FpXU3SU9RN8mY2J3P27FzMZnU5CSHKhhREQgghhKjypCASQgghRJUnnaqFELeluRuPlXcKQohKRAoiIUSl5+3szSutX8Hb2VtVnCyPaiSFt6W95hwuHiWflBHA0dGbwMBXcHRUl5MQomxoFEVRN7a0CkhLS8PDw4PU1FTc3d3LOx0hRBGUdwvRiIdl/iEhyltxvr+lD5EQotLLNGeyP2E/meZMVXEcsrPwOPIvCUfAlK3uQZEWSyZXruzHYlGXkxCibJRrQbRjxw4ee+wxAgMD0Wg0/PDDD3bbFUVh/PjxBAQEYDQa6dChA8ePH7fb59KlS/Tp0wd3d3c8PT0ZOHAg6enpdvscOnSI+++/HycnJ4KCgpgxY8atvjQhRBn698K/hH4cyr8X/lUVx+vMSVqOmMjMEfVJPqPuCfWZmf8SGRlKZqa6nIQQZaNcC6KMjAxatGjBhx9+WOD2GTNmsGDBAhYtWsTevXtxcXGhU6dOZGdn2/bp06cPR44cYePGjaxdu5YdO3YwZMgQ2/a0tDQ6duxIcHAwkZGRzJw5k4kTJ/Lxxx/f8usTQgghROVQrp2qH3nkER555JECtymKwrx583jnnXfo1q0bACtWrMDPz48ffviBXr16ERUVxfr169m3bx+tW7cG4IMPPqBLly7MmjWLwMBAVq5ciclk4rPPPkOv19OkSRMOHDjAnDlz7AonIcTtobz7DgkhKqcK24coNjaWxMREOnToYFvn4eFBmzZt2L17NwC7d+/G09PTVgwBdOjQAa1Wy969e237tGvXDr3+v+bvTp06ER0dzeXLlws8d05ODmlpaXYvIYQQQty+KmxBlJiYCICfn5/dej8/P9u2xMREfH197bY7ODjg5eVlt09BMa49x/WmTZuGh4eH7RUUFKT+goQQt4xWo8VN74ZWo+4jTdFqyXUy4ORkRmUoQItO50YF/pgVQlxD/qUWYOzYsaSmptpeZ86cKe+UhBCFaOnfkrSxabT0b6kqzvk6jdj103Km/hRLjTo5qmJ9useZv7L/ws1NXU5CiLJRYQsif39/AJKSkuzWJyUl2bb5+/uTnGz/4MTc3FwuXbpkt09BMa49x/UMBgPu7u52LyGEEELcvipsQRQSEoK/vz+bN2+2rUtLS2Pv3r2EhYUBEBYWRkpKCpGRkbZ9tmzZgtVqpU2bNrZ9duzYgdlstu2zceNGGjRoQLVq1croaoQQt9LR80dp8lETjp4/qiqO1+kTtOn3GrP6BZB4Wt2weydO0NChKxkZ6nISQpSNci2I0tPTOXDgAAcOHACudqQ+cOAAcXFxaDQahg8fzrvvvstPP/3EP//8Q9++fQkMDKR79+4ANGrUiM6dOzN48GD+/PNPdu3axdChQ+nVqxeBgYEA9O7dG71ez8CBAzly5AirV69m/vz5jBw5spyuWghR2rJzszl6/ijZudk337kQDqYcnBKSiE9wI9ekbmJGDTkYNSewWtXlJIQoG+U67P6vv/7iwQcftC3nFSn9+vVj2bJljB49moyMDIYMGUJKSgpt27Zl/fr1ODk52Y5ZuXIlQ4cOJTw8HK1WS8+ePVmwYIFtu4eHBxs2bCAiIoLQ0FC8vb0ZP368DLkXQgghhE25FkTt27ensEepaTQaJk+ezOTJk2+4j5eXF6tWrSr0PM2bN+f3338vcZ5CCCGEuL1V2D5EQgghhBBlRQoiIUSld0e1O/ix14/cUe0OVXFSA4KIejOCF988TvUA880PKISJIGJyP8LJSV1OQoiyUa63zIQQojR4OnnyeIPH/38pudB9C5Pj6k7yQ/fTAAW48e38orDgTpoSjqOjp6o4QoiyIS1EQohKLzE9kWm/TyMxveDZ54vK+dJ57liyih2f6Um7pFMVy4Hz+GkXk5OjLichRNmQgkgIUenFX4nnrS1vEX8lXlUc14vJBK3+iR++qk3aRXUN6I4kE6ibg8mkLichRNmQgkgIIYQQVZ4UREIIIYSo8qQgEkIIIUSVJ6PMhBCVnqeTJ082fhJPJ08gt8Rxclzdudy6OaGaJIyuVlU5WXDnsrUTDg6equIIIcqGRilsqmgBXH2orIeHB6mpqfLkeyEquLkbj5V3CnZGPFy/vFMQosoqzve33DITQlR6JouJs2lnMVlMquJozSacEpJIS7CSq25eRjSYcCQRq1VdTkKIslGigujkyZOlnYcQQpTY4eTDBM0N4nDyYVVxvE8dp02/15jYryGJpwyqYjlxnKaOD5CRoS4nIUTZKFFBVLduXR588EG++OILsrOzSzsnIYQQQogyVaKCaP/+/TRv3pyRI0fi7+/Piy++yJ9//lnauQkhhBBClIkSFUQtW7Zk/vz5xMfH89lnn5GQkEDbtm1p2rQpc+bM4fz586WdpxBCCCHELaOqU7WDgwM9evTgm2++4f333+fEiRO88cYbBAUF0bdvXxISEkorTyGEEEKIW0bVsPu//vqLzz77jK+++goXFxf69evHwIEDOXv2LJMmTSItLe22uJUmw+6FqNisihWzxYyjzpH5m06oCGRFY8rFjCNavQatqj8ZrWjI5bUOjdFoZECvEOWhON/fJSqI5syZw9KlS4mOjqZLly4MGjSILl26oL3m0+Ps2bPUrl2b3NyST5JWUUhBJETFV9HmH7qezEckRNm75fMQ/e9//6N3796cPn2aH374gUcffdSuGALw9fVlyZIlJQkvhBDFcuziMT48/DzJWbGq4niejaXVq2+x+FVvks86qoplIJa6uucxoC4nIUTZKNGjO44fP37TffR6Pf369StJeCGEKJZ0UzoxaX9ismSqiqPPysQt+iTReGHKuqIqlpZM3LR/olWZkxCibJSohWjp0qV88803+dZ/8803LF++XHVSQgghhBBlqUQF0bRp0/D29s633tfXl6lTp6pOSgghhBCiLJWoIIqLiyMkJCTf+uDgYOLi4lQnJYQQQghRlkrUh8jX15dDhw5Ru3Ztu/UHDx6kevXqpZGXEEIUWS2PWjxd5108DQGq4qT5BnDypefoQwyeviWekQQAEwHE5b6LCXU5CSHKRokKomeffZZhw4bh5uZGu3btANi+fTuvvfYavXr1KtUEhRDiRq4dan+P31Oq42V7eHGmx6OEYlEdy4IXFxX1OQkhykaJbplNmTKFNm3aEB4ejtFoxGg00rFjRx566CHpQySEKHPp5kvsSfqGdPMlVXGcUi8RtGYtkWt0pKeqm0xRxyWqa75Bh7qchBBlo0QtRHq9ntWrVzNlyhQOHjyI0WikWbNmBAcHl3Z+QghxUyk5CXwd8w41XRrj6uhV4jjuyQncsegLVjKSkc1O4+qRU+JYehKo5fAOmebGZFHynIQQZaNEBVGe+vXrU7++zL4qhBBCiMqtRAWRxWJh2bJlbN68meTkZKxWq932LVu2lEpyQghxu8jr7ySP8BCiYipRQfTaa6+xbNkyunbtStOmTdFoNKWdlxBCCCFEmSlRQfTVV1/x9ddf06VLl9LORwghik2vc6aO+93odc6q4piMzlxpcAcNuITeaL35AYWw4swV691YUZeTEKJslLhTdd26dUs7FyGEKBFfYwgRTT9XHSelZgj7P5jKi1xQHSuHEE5Y1OckhCgbJRpX+vrrrzN//nwURd3EZUIIURqsipVcqwmroq5VB6sVTbaJ3GwFq8pQYEWDCVAdSAhRBkrUQrRz5062bt3KunXraNKkCY6Ojnbb16xZUyrJCSFEUcRnRDHnUA9GNl9DTdcmJY7jGxNFo4gphLKfkR+epma9kg+7NxJFQ8ce/GteQxYlz0kIUTZKVBB5enryxBNPlHYuQgghhBDlokQF0dKlS0s7DyGEEEKIclPiuelzc3PZtGkTixcv5sqVKwDEx8eTnp5easlZLBbGjRtHSEgIRqOROnXqMGXKFLu+S4qiMH78eAICAjAajXTo0IHjx4/bxbl06RJ9+vTB3d0dT09PBg4cWKp5CiGEEKJyK1FBdPr0aZo1a0a3bt2IiIjg/PnzALz//vu88cYbpZbc+++/z//+9z8WLlxIVFQU77//PjNmzOCDDz6w7TNjxgwWLFjAokWL2Lt3Ly4uLnTq1Ins7GzbPn369OHIkSNs3LiRtWvXsmPHDoYMGVJqeQohhBCictMoJRgq1r17d9zc3FiyZAnVq1fn4MGD3HHHHWzbto3Bgwfna6EpqUcffRQ/Pz+WLFliW9ezZ0+MRiNffPEFiqIQGBjI66+/bivEUlNT8fPzY9myZfTq1YuoqCgaN27Mvn37aN26NQDr16+nS5cunD17lsDAwJvmkZaWhoeHB6mpqbi7u5fKtQkh1Mub/TnXaiLdfAlXRy8ctPoSx9OaTegvXCYZH5y9tTg43vyYG9FgwoFL5OKFwn85yUzVQpSd4nx/l6iF6Pfff+edd95Br7f/4Klduzbnzp0rScgC3XvvvWzevJljx65+6B08eJCdO3fyyCOPABAbG0tiYiIdOnSwHePh4UGbNm3YvXs3ALt378bT09NWDAF06NABrVbL3r17CzxvTk4OaWlpdi8hRMXloNXjafBXVQwBWB31ZAf44R6grhgCUNBjxt+uGBJCVFwlKoisVisWiyXf+rNnz+Lm5qY6qTxvvvkmvXr1omHDhjg6OnLnnXcyfPhw+vTpA0BiYiIAfn5+dsf5+fnZtiUmJuLr62u33cHBAS8vL9s+15s2bRoeHh62V1BQUKldkxCi9F3MPsPy6GFczD6jKo5HwhmavzWVlW97cDFBXUWk5wy1dcPQoy4nIUTZKFFB1LFjR+bNm2db1mg0pKenM2HChFJ9nMfXX3/NypUrWbVqFfv372f58uXMmjWL5cuXl9o5CjJ27FhSU1NtrzNn5ANNiIosKzeNgxd/IytXXWuuIT2Nan8dInKfH1npJR5zAoCONKppf0OHtDALURmUaNj97Nmz6dSpE40bNyY7O5vevXtz/PhxvL29+fLLL0stuVGjRtlaiQCaNWvG6dOnmTZtGv369cPf3x+ApKQkAgICbMclJSXRsmVLAPz9/UlOTraLm5uby6VLl2zHX89gMGAwGErtOoQQQghRsZXoT6CaNWty8OBB3nrrLUaMGMGdd97J9OnT+fvvv/PdnlIjMzMTrdY+RZ1Oh/X/59QPCQnB39+fzZs327anpaWxd+9ewsLCAAgLCyMlJYXIyEjbPlu2bMFqtdKmTZtSy1UIIYQQlVeJWojgaj+c5557rjRzyeexxx7jvffeo1atWjRp0oS///6bOXPm8MILLwBXb9UNHz6cd999l3r16hESEsK4ceMIDAyke/fuADRq1IjOnTszePBgFi1ahNlsZujQofTq1atII8yEEEIIcfsrUUG0YsWKQrf37du3RMlc74MPPmDcuHG88sorJCcnExgYyIsvvsj48eNt+4wePZqMjAyGDBlCSkoKbdu2Zf369Tg5Odn2WblyJUOHDiU8PBytVkvPnj1ZsGBBqeQohCh/7npfutQaibteXQt1enVfzjzzON01p3Cvnn/gSHGY8SXeMhIz9jnlTRUAMgRfiIqkRPMQVatWzW7ZbDaTmZmJXq/H2dmZS5culVqCFYHMQyRExXJtUVGZSUEkxK11y+chunz5st0rPT2d6Oho2rZtW6qdqoUQoiiyctM4fGlzqYwy893yO9FbNKUyysxds1lGmQlRSaj7F3+NevXqMX36dF577bXSCimEEEVyMfsMn/37SqnMQ9Ro+ocsnl6vVOYhquPwisxDJEQlUWoFEVztaB0fH1+aIYUQQgghbrkSdar+6aef7JYVRSEhIYGFCxdy3333lUpiQghxvdul75AQouIpUUGUN6Q9j0ajwcfHh4ceeojZs2eXRl5CCCGEEGWmRAVR3sSIQghREThoDfgZ6+KgVTfDfK7eQHaAH4FcwUFf7AG4dhQMZCl1UZBZ74WoDEo07L6qkWH3QlQMt9stMxl2L8StVZzv7xK1EI0cObLI+86ZM6ckpxBCCCGEKDMlKoj+/vtv/v77b8xmMw0aNADg2LFj6HQ6WrVqZdtPo9GUTpZCCFGIcxlRLDzch6FNV1LDpVGJ4/jERFF/xHQeYAcRc+OpUSenxLGMRFHPoQ/Hc1eSRclzEkKUjRIVRI899hhubm4sX77cNmv15cuXGTBgAPfffz+vv/56qSYphBCFURQrOZYMFEVd/0aN1YpDdg7ZOKIyFGBFp8kApM+lEJVBieYhmj17NtOmTbN7hEe1atV49913ZZSZEEIIISqdEhVEaWlpnD9/Pt/68+fPc+XKFdVJCSGEEEKUpRIVRE888QQDBgxgzZo1nD17lrNnz/Ldd98xcOBAevToUdo5CiGEEELcUiUadp+Zmckbb7zBZ599htlsBq4+tmPgwIHMnDkTFxeXUk+0PMmweyEqhhsNuzdZskjOOomv8Q70OmOJ4ztkZ+ESc5p/aUj1Ohr0TiWflURDFk6cJJs7UCg4Jxl2L8StVZzvb1XzEGVkZBATEwNAnTp1brtCKI8UREJUDDIPkRCiOIrz/a3q4a4JCQkkJCRQr149XFxckDkehRDl4XJOPN+dnMTlHHUPl3ZLjqfh+wv5aYYLl5NLNAjXxpF4amon4Yg88FqIyqBEBdHFixcJDw+nfv36dOnShYSEBAAGDhwoQ+6FEGUuw3yZXYmryDBfVhXHmHoZv8072bapBhmpOlWxHLiMj24VDqjLSQhRNkpUEI0YMQJHR0fi4uJwdna2rX/mmWdYv359qSUnhBBCCFEWStQmvGHDBn777Tdq1qxpt75evXqcPn26VBITQgghhCgrJWohysjIsGsZynPp0iUMBnmysxBCFMXcjcduu47iQlRWJSqI7r//flasWGFb1mg0WK1WZsyYwYMPPlhqyQkhRFG4OlbngYD+uDpWVxUn07M6iZ3b0+GRs7h6WlTFyqU6yZb+5KIuJyFE2SjRsPvDhw8THh5Oq1at2LJlC48//jhHjhzh0qVL7Nq1izp16tyKXMuNDLsXomK4XVtTZPi9ELfGLR9237RpU44dO0bbtm3p1q0bGRkZ9OjRg7///vu2K4aEEBVfjiWDU1f+JseSoSqOY1YG1f7+h7N/K+RkaVTF0pKBs+ZvtKjLSQhRNordqdpsNtO5c2cWLVrE22+/fStyEkKIYjmfdYoF//RiZPM11HRtUuI41c6eotGY9xjAfkZ+eJqa9XJKHMvAKRo49OJf8xqyKHlOQoiyUeyCyNHRkUOHDt2KXIQQokC3660yIUTFUaJbZs899xxLliwp7VyEEEIIIcpFieYhys3N5bPPPmPTpk2Ehobme4bZnDlzSiU5IYQQQoiyUKyC6OTJk9SuXZvDhw/TqlUrAI4ds2/K1mjUdUQUQoji0mp0uDhUQ6tR97gNq06H2dUZN3LQ6tQ9m1FBh1mphoK6nIQQZaNYw+51Oh0JCQn4+voCVx/VsWDBAvz8/G5ZghWBDLsXonzd7n2IZNi9ELfGLRt2f33ttG7dOjIyZEipEEIIISq3EnWqzlOCOR2FEKLUJWYe5739D5OYeVxVnOqnjnNPn1eY3qcGiaf0qmI5cZzGDg/jhLqchBBlo1h9iDQaTb4+QtJnSAhR3nKtJi5mx5FrNamKozObMJy/RDIu5JovqIqlwYRBE4cGdTkJIcpGsQoiRVHo37+/7QGu2dnZvPTSS/lGma1Zs6b0MhRCCCGEuMWKVRD169fPbvm5554r1WSEEKIqyus0Lp2rhSg/xSqIli5deqvyEEIIIYQoNyWamFEIISoSb6dghjT6FG+nYFVxUgKDOTp+OK9wHO9Adf0jcwjmRO6n5KAuJyFE2VA1yqwsnDt3jueee47q1atjNBpp1qwZf/31l227oiiMHz+egIAAjEYjHTp04Phx+1Edly5dok+fPri7u+Pp6cnAgQNJT08v60sRQtwiTg6uNKx2P04OrqrimFxcOd/2Huq2VXBysaqKZcWVK8r9WFGXkxCibFTogujy5cvcd999ODo6sm7dOo4ePcrs2bOpVq2abZ8ZM2awYMECFi1axN69e3FxcaFTp05kZ2fb9unTpw9Hjhxh48aNrF27lh07djBkyJDyuCQhxC2QZkpmfdwHpJmSVcVxuZhM3UXL2bLYibSL6maYdiAZf+0HOKAuJyFE2ajQt8zef/99goKC7PouhYSE2P5bURTmzZvHO++8Q7du3QBYsWIFfn5+/PDDD/Tq1YuoqCjWr1/Pvn37aN26NQAffPABXbp0YdasWQQGBpbtRQkhSl2a6Twbzi6kqddDuOt9SxzH5dJ5aqxZx1reo/5Dp3GvbilxLEfOE6BbSKr1IXIpeU5CiLJRoVuIfvrpJ1q3bs1TTz2Fr68vd955J5988olte2xsLImJiXTo0MG2zsPDgzZt2rB7924Adu/ejaenp60YAujQoQNarZa9e/cWeN6cnBzS0tLsXkIIIYS4fVXogujkyZP873//o169evz222+8/PLLDBs2jOXLlwOQmJgIkO9Zan5+frZtiYmJtmev5XFwcMDLy8u2z/WmTZuGh4eH7RUUFFTalyaEEEKICqRC3zKzWq20bt2aqVOnAnDnnXdy+PBhFi1alG9OpNI0duxYRo4caVtOS0uTokiIMna7P9BVCFGxVOgWooCAABo3bmy3rlGjRsTFxQHg7+8PQFJSkt0+SUlJtm3+/v4kJ9t3aszNzeXSpUu2fa5nMBhwd3e3ewkhKi6jgzutvB/D6KDu32q2qzsXw0JpE5aA0bXk/YcALLhzyfoYFuTzQ4jKoEIXRPfddx/R0dF2644dO0Zw8NV5PUJCQvD392fz5s227Wlpaezdu5ewsDAAwsLCSElJITIy0rbPli1bsFqttGnTpgyuQghxq1V3CuK5+rOo7qSuJTctIIjDk0bxzKQrVA/IVRXLRBCnLbMwIa3LQlQGFbogGjFiBHv27GHq1KmcOHGCVatW8fHHHxMREQFcfbDs8OHDeffdd/npp5/4559/6Nu3L4GBgXTv3h242qLUuXNnBg8ezJ9//smuXbsYOnQovXr1khFmQtwmzNYczmedxmzNURVHZ8rB+dQZLp1SMJvUTcyoIQc9p9GgLichRNmo0AXRXXfdxffff8+XX35J06ZNmTJlCvPmzaNPnz62fUaPHs2rr77KkCFDuOuuu0hPT2f9+vU4OTnZ9lm5ciUNGzYkPDycLl260LZtWz7++OPyuCQhxC2QlHmCaX93JCnzhKo41U+f4K4ho3h3SAOSTutVxXLiBE0cO+KEupyEEGWjQneqBnj00Ud59NFHb7hdo9EwefJkJk+efMN9vLy8WLVq1a1ITwghhBC3gQrdQiSEEEIIURakIBJCCCFElScFkRBCCCGqPI2iKEp5J1HRpaWl4eHhQWpqqsxJJEQZqYoTM454uH55pyDEbaU439/SQiSEEBXE3I3HqmQhKERFIAWREKLSS846yfxDz5CcdVJVnGpnThL60mg+fMmX5DOOqmIZOEl93TMYUJeTEKJsVPhh90IIcTMmSxan0w9gsmSpiuOYnYXryThi8MSUnaoqlpYsXLQH0KrMSQhRNqQgEkJUKHLLSAhRHuSWmRBCCCGqPCmIhBBCCFHlyS0zIUSlV81Qg971ZlDNUENVnFS/Gpx4dQD9NDFU81M3I4mJGpzKnYEJdTkJIcqGFERCiErPxdGT1j7dVMfJcffk3GOdaIFFdSwLnlxW1OckhCgbcstMCFHppZsvsTNhJenmS6riGFMuEbz6e/Z+5Uh6ik5VLAcu4a1diQPqchJClA0piIQQlV5KTgJrYieTkpOgKo7b+QRqL1nN6s9CSDmvrgHdkQSCdJNxpPg55U3QKCPuhCg7UhAJIYQQosqTgkgIIYQQVZ4UREIIIYSo8mSUmRCiQlDTX8agc6GBZ1sMOhdVOZicXUht0oAmXMDgbFUVy4oLada2WFGXkxCibGgURVE32UYVkJaWhoeHB6mpqbi7u5d3OkLclqQDccFGPFy/vFMQotIqzve33DITQlR6VsVCdm46VkXd/EEaiwWHtHRy0hSsqqcisqAlHUphTiMhxK0nBZEQotKLz/iXt/4MJT7jX1VxfE7+y31PDmLskw2IP2lQFcvIv7RwDMWIupyEEGVDCiIhhBBCVHlSEAkhhBCiypOCSAghKgGZuVqIW0sKIiGEEEJUeTIPkRCi0gtwrs/ku3Zj1LmpinMhpD67V/2PdzX/YvBQ9/diFvX5x7ybXNTlJIQoG1IQCSEqPZ3WEVetl+o4VgdHTN7VcC6FnMCRXNTnJIQoG3LLTAhR6V3IjmNJ1EtcyI5TFccjPo4WoyezfHQ1LsQ7qoqlJ447dC+hR11OQoiyIQWREKLSy869wpHLW8nOvaIqjiHjCp4HjnLwgA/ZGeo+HnVcwUO7FR3qchJClA25ZSaEKFcyckoIURFIC5EQQgghqjwpiIQQQghR5cktMyFEpeeh9+Px2m/iofdTFSfd24/Tz/XkSU0sHt5WVbHM+HHW8iZm1OUkhCgbUhAJISo9N7037QMHqI6TWc2bU32f4l7MqmPl4s15q/qchBBlQwoiIUSll5mbyrGUP6jveS/ODh4ljmO4korXH5FsogMh94KzW8lbiXSk4qb5gyvKvVgoeU7S6VyIsiF9iIQQld6l7LOsODacS9lnVcXxSDxL/dkfs2R2XS4lqp2H6CwhDsPRoy4nIUTZqFQF0fTp09FoNAwfPty2Ljs7m4iICKpXr46rqys9e/YkKSnJ7ri4uDi6du2Ks7Mzvr6+jBo1itzc3DLOXgghDygVQlRUlaYg2rdvH4sXL6Z58+Z260eMGMHPP//MN998w/bt24mPj6dHjx627RaLha5du2Iymfjjjz9Yvnw5y5YtY/z48WV9CUIIIYSooCpFQZSenk6fPn345JNPqFatmm19amoqS5YsYc6cOTz00EOEhoaydOlS/vjjD/bs2QPAhg0bOHr0KF988QUtW7bkkUceYcqUKXz44YeYTKbyuiQhhBBCVCCVoiCKiIiga9eudOjQwW59ZGQkZrPZbn3Dhg2pVasWu3fvBmD37t00a9YMP7//hr526tSJtLQ0jhw5UuD5cnJySEtLs3sJISouR60TNVwa46h1UhUn1+BERq0aBNdKxdGgqIplxYlMpTFW1OUkhCgbFX6U2VdffcX+/fvZt29fvm2JiYno9Xo8PT3t1vv5+ZGYmGjb59piKG973raCTJs2jUmTJpVC9kKIsuDnXIfXW3yvOs6lWnW49OlsXiPp5jvfRA51iM5Vn5MQomxU6BaiM2fO8Nprr7Fy5UqcnMrur6yxY8eSmppqe505c6bMzi2EEEKIslehC6LIyEiSk5Np1aoVDg4OODg4sH37dhYsWICDgwN+fn6YTCZSUlLsjktKSsLf3x8Af3//fKPO8pbz9rmewWDA3d3d7iWEqLjOph9l1O6mnE0/qiqOz4mjtO3yHKO71OHsCYOqWEaO0sKhKUbU5SSEKBsVuiAKDw/nn3/+4cCBA7ZX69at6dOnj+2/HR0d2bx5s+2Y6Oho4uLiCAsLAyAsLIx//vmH5ORk2z4bN27E3d2dxo0bl/k1CSFuBQWLYgbU9fvRKAq63Fxyc3VqQwEKWo36nIQQZaNC9yFyc3OjadOmdutcXFyoXr26bf3AgQMZOXIkXl5euLu78+qrrxIWFsY999wDQMeOHWncuDHPP/88M2bMIDExkXfeeYeIiAgMBnV/AQohiub6uYdkLiIhREVToQuiopg7dy5arZaePXuSk5NDp06d+Oijj2zbdToda9eu5eWXXyYsLAwXFxf69evH5MmTyzFrIYQQQlQkla4g2rZtm92yk5MTH374IR9++OENjwkODubXX3+9xZkJIYQQorKqdAWREEJcz9dYh9Et1+JlCFIV52KtOuxf+B5vEo1nLXVdLLOpQ5R5LTmoy0kIUTakIBJCVHp6nRP+zvVUx7EYnLhSvw6+gNrO0ApOZKM+JyFE2ajQo8yEEKIoLmWfY/WJt7mUfU5VHLekczR+by5r3nPlUpK6vxcdOUeQ7m0cUZeTEKJsSEEkhLhlyurp9pm5KexN/pbM3BRVcYxpKfhs38vO7YFkpulUxXIgBW/ttzigLichRNmQgkgIIYQQVZ4UREIIIYSo8qQgEkIIIUSVJ6PMhBCVnqujNw/VGIKro7eqOBnVvIl/7GE6a87gWs2iKlYu3iRahpCLupxu5Nq+WSMern9LziFEVSIFkRCi0vM0+PFo8Ouq42R4+3H81YF0JEt1LDN+JFjV5ySEKBtSEAkhKr1sSzpn049Q07UJTjrXEsdxzEzH/Z9oIgnFv5kGJ+eSz0WkJR1nzREylSZYKXlO15PnwAlxa0gfIiFEpXch6zQfHenLhazTquJUO3eapuNmsmBcfS6c06uKZeA09Rz6YkBdTkKIsiEFkRBCVHJlNd+TELczuWUmhChV8sUshKiMpIVICCGEEFWetBAJISo9rcYBD70fWo26jzSLgwMmT3eqkYXWQe3DXR0wKX4o8jErRKWgURRF3b/6KiAtLQ0PDw9SU1Nxd3cv73SEqNDklln5kfmIhLBXnO9vuWUmhBBCiCpPCiIhRKUXnxHNpL/aEZ8RrSpO9dhowp4ewpSng4iPVTfs3olomji0wwl1OQkhyobc3BZCVHpWJZdUUxJWJVdVHF1uLvqUNC5jxJqrURVLQy56TRIa1OUkhCgb0kIkhBC3CZmPSIiSk4JICCGEEFWeFERCCCGEqPKkD5EQotLzNgbzSpMVeBuDVcW5XCOYw1NGMYxjeNdQ14coh2CO564gB3U5CSHKhhREQohKz0nnSl2PNqrjmJ1dudgmlNoAqJuizYor6Yr6nIQQZUNumQkhKr2UnCTWnp5NSk6SqjguF5Ko98ESNiw0knJB3d+LjiQRoJ2NI+pyEkKUDSmIhBClojxHOKWbL7Dl3Mekmy+oiuNy+QKBP29k/U9BpF/WqYrlwAX8dR/jgLqchBBlQ26ZCSFUkWHeFc+1PxN5nIcQRSMtREIIIYSo8qQgEkIIIUSVJ7fMhBCVnrODJ218n8TZwVNVnCx3T84/0Ia2xOPsblEVKxdPLlifJBd1OQkhyoZGURR1Y0urgLS0NDw8PEhNTcXd3b280xGiQpE+RBWb9CESVVlxvr/llpkQotIzWbJJzDyOyZKtKo4uJxu3YzEkH1Mw5ah9uGs2ThxHg7qchBBlQwoiIUSJVKQHiSZnxTDjwKMkZ8WoilM9LoZWQ99m+tAGJMfpVcVyIoZGjo/ihLqcSktF+nkJURFJQSSEEEKIKk86VQshikxaGIQQtyspiIQQ4jYmRawQRVOhb5lNmzaNu+66Czc3N3x9fenevTvR0dF2+2RnZxMREUH16tVxdXWlZ8+eJCXZPzsoLi6Orl274uzsjK+vL6NGjSI3N7csL0UIcUtp0GkcAXUdoRWNBouDAw4OFrWhAA1WRX1OQoiyUaELou3btxMREcGePXvYuHEjZrOZjh07kpGRYdtnxIgR/Pzzz3zzzTds376d+Ph4evToYdtusVjo2rUrJpOJP/74g+XLl7Ns2TLGjx9fHpckhLgFaro2ZmbYYWq6NlYV53zdxuz89Qtm/BpDzbo5qmJl0ZiDuYfJQl1OQoiyUanmITp//jy+vr5s376ddu3akZqaio+PD6tWreLJJ58E4N9//6VRo0bs3r2be+65h3Xr1vHoo48SHx+Pn58fAIsWLWLMmDGcP38evf7mI0lkHiIhrpLbL5WfzEskqpLbdh6i1NRUALy8vACIjIzEbDbToUMH2z4NGzakVq1a7N69G4Ddu3fTrFkzWzEE0KlTJ9LS0jhy5EgZZi+EuFWSMmOYffAJkjLVDXH3iouh9aDXmT/IjySVw+4NxNDA4QkMFWTYvRCicJWmU7XVamX48OHcd999NG3aFIDExET0ej2enp52+/r5+ZGYmGjb59piKG973raC5OTkkJPzX3N5WlpaaV2GEOIWMFuzOZdxFLNV3SSIDjnZuMSd4zQemHNSVMXSko2z5ihamZhRiEqh0hREERERHD58mJ07d97yc02bNo1Jkybd8vMIUVnIrbLbR97PUm6dCWGvUtwyGzp0KGvXrmXr1q3UrFnTtt7f3x+TyURKSord/klJSfj7+9v2uX7UWd5y3j7XGzt2LKmpqbbXmTNnSvFqhBBCCFHRVOiCSFEUhg4dyvfff8+WLVsICQmx2x4aGoqjoyObN2+2rYuOjiYuLo6wsDAAwsLC+Oeff0hOTrbts3HjRtzd3WncuODRHwaDAXd3d7uXEEIIIW5fFfqWWUREBKtWreLHH3/Ezc3N1ufHw8MDo9GIh4cHAwcOZOTIkXh5eeHu7s6rr75KWFgY99xzDwAdO3akcePGPP/888yYMYPExETeeecdIiIiMBgM5Xl5QohS4uVUk7715+HlVPPmOxci1b8mx14fwkBO4FVwA3KRmahJbO48TKjLSQhRNir0sHuNpuAJzZYuXUr//v2BqxMzvv7663z55Zfk5OTQqVMnPvroI7vbYadPn+bll19m27ZtuLi40K9fP6ZPn46DQ9HqQRl2L6oq6Tt0+5I+RKIqKM73d4UuiCoKKYhEVVVZCqIrpgtEXviZUO/HcNN7lziO8+UL+P68lVWa3jR71IpbNUuJYzlwgWran7lsfYxcSp5TWZDiSNyubtt5iIQQoiCppiR+OjWdVFPSzXcuhOuFJIK/+I5vPw8h9YK6HgWOJFFTNx1H1OUkhCgbUhAJIYQQosqr0J2qhRBlp7LcHhNCiFtBWoiEEEIIUeVJC5EQotJzcnCjSbUHcXJwUxUnx8WNlJaNacF5nFysqmJZcCPV+iAW1OVUFmT2aiFklFmRyCgzURXILTMhBZG43cgoMyFElWKxmkk3X8JiNauKo801o79wmcyLViy5arMy48AlQF1OZWnuxmNSGIsqSwoiIUSll5B5jPH7wkjIVPdl7h17jLDeL/POsw1JiFU3k72RYzRzDMOIFBhCVAZSEAkhhBCiypOCSIgqSm6PCCHEf2SUmRBVnBRFQgghLURCCCGuI62HoiqSYfdFIMPuxe3odvrCsyoWTJYs9DojWo2uxHE0Fgu6jCwycMHRRYO25KEAC1qysGIEVAWqEGRIvqiMivP9LbfMhBCVnlajw8nBVXUcRacj190VdePL8uiwoj4nIUTZkFtmQohK73zWKRYfHcj5rFOq4nieO0XLERNYMsKL8+ccVcUycIo6uoEYUJeTEKJsSAuREFXA7XR7rCA5lgyiU3aSY8lQFUefmYHHkWiO4E2nTHWxtGTgrt2JVmVOQoiyIS1EQgghbko6WovbnRREQgghikwKI3G7kltmQlRyBX05yYggIYQoHimIhLiNVZW/5D0NAfQIGY+nIUBVnCs+AZwa+AzPKLF4+lhVxTITwBnLeMyoy0kIUTakIBJCVHqujl60DeijOk6Wpxenn3mCNqXwhPpcvLhgVZ9TZZJXgEsLpaiMpA+REJWU9OX4T4Y5hb/O/0iGOUVVHENaCjV+/o2Da3VkpKn7eNSRQjXNj+hQl5MQomxIC5EQt6GqVihdzjnHquOjGdl8DS6OniWO45F0jrofLGU5rzKywWlc3HNKHEvPOWo7jOZf8xqyKHlOFV1V+10Tty9pIRJCCCFElSctREJUMvIXuajopC+RqIykhUgIIYQQVZ60EAkhKj29zkiwa0v0OqOqOGYnI+l31KIOKeid1A27t2Ikw9ry/592f/uRlkpxu9EoiqKUdxIVXVpaGh4eHqSmpuLu7l7e6YgqpKBbD/JFJCobuXUmyktxvr+lhUiICkiKHnE7ufb3WYojUVFJHyIhKoG8OYekUCrY2fQjjPyjAWfTj6iK43v8CA907MXIjvU5e9ygKpaRI9zp2AAj6nK63cjvsaiopIVIiHImXw6iKrr+drC0IonyJgWREGXk+sJHPvSFEKLikIJIiHIiLUNCyL8DUXFIQSSEEKJSkNtq4laSYfdFIMPuRUEK+3CWv3rLltmaQ0pOIp4Gfxy1Je8MrTPlYIhP5iw1cQvU4qgv+cejhhwcScSMPwrqOmhXdQX1M7p+mxAFkWH3QtxCBX0oy6MKypej1oCPMVh1HIveQGbtoP9r7+5jmrr3MIA/B+5aQBGFAUJ4HdsEI6IBy4AlyiAyt+mYGVuW3VmQoJhiJM1lwLJJlrhgdvfCxpiwNzS7ElhckJjpJuEKJONFhzAdE+40bLyKCJOXOuko5/5h1owL1wFtOZTzfJIT0l8PPc83bU+/PT0vcAYAmPZdUYQSepieiYgWhqwOuy8oKICfnx/s7OwQHh6O8+fPSx2JrMhsDhfmIcXSGLzThX/95x8YvNNl0uOs6OvCupx/oizHEYN9pn1fVKALvrb/gAKmZSK+r2hhyGYLUVlZGbRaLQoLCxEeHo68vDzExcWhvb0dbm5uUscjifEIMOv228QILt48hS2eSSY9jt3YCFzqm9AID0T9/RcAE/N+LFuMwNnmFG4YTMtE9zafRonvb5qJbBqid955BykpKUhKurtyKiwsxFdffYXPPvsMWVlZEqcjqfy/lSm/jRItXTPt/8efvUkWDZFer0dTUxOys7ONYzY2NoiNjUV9fb2EychceM0vIpqP/11PzGa9MZeTSfLIOOshi4bo5s2bMBgMcHd3nzLu7u6Otra2afOPj49jfHzceHt4eBjA3b3VLaHg31cBAJrHHpzz/8z1/0zNM5f7/pxxoeSevLjgyyTpjd++Ddy5+/eOMDbvx7n9222MwQBgBOO/jeKOTm9CqtvQ3Qf89vtt3MH8M9HiM9N6Zjbrnvmsn+61fp/LOnY2nxOW/FyZzXItscw/PrdndUC9KAM9PT0iALGurm7KeEZGhqhSqabNn5OTI+LuISacOHHixIkTJyufurq6/rJXkMUWovvvvx+2trbo7++fMt7f34/Vq1dPmz87OxtardZ4e3JyEkNDQ3BxcYEgCBbP+79GRkbg7e2Nrq6uJXcepKVcG8D6rB3rs26sz3qZqzZRFDE6OgpPT8+/nFcWDZFCoUBoaCiqqqoQHx8P4G6TU1VVhbS0tGnzK5VKKJVTT6S2cuXKBUh6bytWrFhyL/o/LOXaANZn7VifdWN91ssctTk5Oc1qPlk0RACg1WqhVqsRFhYGlUqFvLw86HQ641FnREREJF+yaYief/55DAwM4ODBg7h+/To2bNiAr7/+etqO1kRERCQ/smmIACAtLW3Gn8gWO6VSiZycnGk/4y0FS7k2gPVZO9Zn3Vif9ZKiNl7clYiIiGRPVtcyIyIiIpoJGyIiIiKSPTZEREREJHtsiIiIiEj22BBZqfHxcWzYsAGCIKClpUXqOGazY8cO+Pj4wM7ODh4eHnjppZfQ29srdSyz+Pnnn5GcnAx/f3/Y29sjICAAOTk50OtNuV7W4vHGG28gMjISDg4Oi+JEpqYqKCiAn58f7OzsEB4ejvPnz0sdyWxqa2uxfft2eHp6QhAEnDx5UupIZpObm4tNmzbB0dERbm5uiI+PR3t7u9SxzObIkSNYv3698YSFEREROHPmjNSxLObw4cMQBAHp6ekWXxYbIiv18ssvz+pU5NYmOjoaX3zxBdrb2/Hll1/i2rVrePbZZ6WOZRZtbW2YnJxEUVERWltb8e6776KwsBCvvPKK1NHMQq/XIyEhAfv27ZM6isnKysqg1WqRk5ODixcvIiQkBHFxcbhx44bU0cxCp9MhJCQEBQUFUkcxu5qaGmg0GjQ0NKCyshK///47tm7dCp1OJ3U0s/Dy8sLhw4fR1NSE7777Do899hiefvpptLa2Sh3N7C5cuICioiKsX79+YRZonsun0kI6ffq0GBgYKLa2tooAxObmZqkjWUxFRYUoCIKo1+uljmIRb775pujv7y91DLMqLi4WnZycpI5hEpVKJWo0GuNtg8Egenp6irm5uRKmsgwAYnl5udQxLObGjRsiALGmpkbqKBazatUq8ZNPPpE6hlmNjo6KDz30kFhZWSlu3rxZPHDggMWXyS1EVqa/vx8pKSn4/PPP4eDgIHUcixoaGsLx48cRGRmJ++67T+o4FjE8PAxnZ2epY9Cf6PV6NDU1ITY21jhmY2OD2NhY1NfXS5iM5mN4eBgAluT7zGAwoLS0FDqdDhEREVLHMSuNRoMnn3xyyvvQ0tgQWRFRFJGYmIjU1FSEhYVJHcdiMjMzsWzZMri4uKCzsxMVFRVSR7KIq1evIj8/H3v37pU6Cv3JzZs3YTAYpl3Wx93dHdevX5coFc3H5OQk0tPTERUVhXXr1kkdx2wuX76M5cuXQ6lUIjU1FeXl5Vi7dq3UscymtLQUFy9eRG5u7oIulw3RIpCVlQVBEO45tbW1IT8/H6Ojo8jOzpY68pzMtr4/ZGRkoLm5GWfPnoWtrS127doFcRGfUH2u9QFAT08PHn/8cSQkJCAlJUWi5H9tPrURLRYajQY//PADSktLpY5iVmvWrEFLSwsaGxuxb98+qNVq/Pjjj1LHMouuri4cOHAAx48fh52d3YIum5fuWAQGBgYwODh4z3keeOABPPfcczh16hQEQTCOGwwG2Nra4sUXX8SxY8csHXVeZlufQqGYNt7d3Q1vb2/U1dUt2k3Cc62vt7cXW7ZswSOPPIKjR4/Cxmbxfi+Zz3N39OhRpKen49atWxZOZxl6vR4ODg44ceIE4uPjjeNqtRq3bt1aclssBUFAeXn5lFqXgrS0NFRUVKC2thb+/v5Sx7Go2NhYBAQEoKioSOooJjt58iSeeeYZ2NraGscMBgMEQYCNjQ3Gx8en3GdOsrq462Ll6uoKV1fXv5zv/fffx6FDh4y3e3t7ERcXh7KyMoSHh1syoklmW99MJicnAdw9zcBiNZf6enp6EB0djdDQUBQXFy/qZggw7bmzVgqFAqGhoaiqqjI2CZOTk6iqqrLKi0PLjSiK2L9/P8rLy1FdXb3kmyHg7utzMa8j5yImJgaXL1+eMpaUlITAwEBkZmZarBkC2BBZFR8fnym3ly9fDgAICAiAl5eXFJHMqrGxERcuXMCjjz6KVatW4dq1a3jttdcQEBCwaLcOzUVPTw+2bNkCX19fvPXWWxgYGDDet3r1agmTmUdnZyeGhobQ2dkJg8FgPD/Wgw8+aHytWgutVgu1Wo2wsDCoVCrk5eVBp9MhKSlJ6mhmMTY2hqtXrxpvd3R0oKWlBc7OztPWM9ZGo9GgpKQEFRUVcHR0NO735eTkBHt7e4nTmS47Oxvbtm2Dj48PRkdHUVJSgurqanzzzTdSRzMLR0fHaft7/bFPqcX3A7P4cWxkMR0dHUvqsPtLly6J0dHRorOzs6hUKkU/Pz8xNTVV7O7uljqaWRQXF4sAZpyWArVaPWNt586dkzravOTn54s+Pj6iQqEQVSqV2NDQIHUkszl37tyMz5VarZY6msn+33usuLhY6mhmsXv3btHX11dUKBSiq6urGBMTI549e1bqWBa1UIfdcx8iIiIikr3FvQMDERER0QJgQ0RERESyx4aIiIiIZI8NEREREckeGyIiIiKSPTZEREREJHtsiIiIiEj22BARERGR7LEhIqIlJTExccldqJSILI8NEREREckeGyIiko2amhqoVCoolUp4eHggKysLExMTxvtPnDiB4OBg2Nvbw8XFBbGxsdDpdACA6upqqFQqLFu2DCtXrkRUVBR++eUXqUohIjNjQ0REstDT04MnnngCmzZtwvfff48jR47g008/xaFDhwAAfX19eOGFF7B7925cuXIF1dXV2LlzJ0RRxMTEBOLj47F582ZcunQJ9fX12LNnDwRBkLgqIjKXv0kdgIhoIXz44Yfw9vbGBx98AEEQEBgYiN7eXmRmZuLgwYPo6+vDxMQEdu7cCV9fXwBAcHAwAGBoaAjDw8N46qmnEBAQAAAICgqSrBYiMj9uISIiWbhy5QoiIiKmbNWJiorC2NgYuru7ERISgpiYGAQHByMhIQEff/wxfv31VwCAs7MzEhMTERcXh+3bt+O9995DX1+fVKUQkQWwISIiAmBra4vKykqcOXMGa9euRX5+PtasWYOOjg4AQHFxMerr6xEZGYmysjI8/PDDaGhokDg1EZkLGyIikoWgoCDU19dDFEXj2LfffgtHR0d4eXkBAARBQFRUFF5//XU0NzdDoVCgvLzcOP/GjRuRnZ2Nuro6rFu3DiUlJQteBxFZBvchIqIlZ3h4GC0tLVPG9uzZg7y8POzfvx9paWlob29HTk4OtFotbGxs0NjYiKqqKmzduhVubm5obGzEwMAAgoKC0NHRgY8++gg7duyAp6cn2tvb8dNPP2HXrl3SFEhEZseGiIiWnOrqamzcuHHKWHJyMk6fPo2MjAyEhITA2dkZycnJePXVVwEAK1asQG1tLfLy8jAyMgJfX1+8/fbb2LZtG/r7+9HW1oZjx45hcHAQHh4e0Gg02Lt3rxTlEZEFCOKftx8TERERyRD3ISIiIiLZY0NEREREsseGiIiIiGSPDRERERHJHhsiIiIikj02RERERCR7bIiIiIhI9tgQERERkeyxISIiIiLZY0NEREREsseGiIiIiGSPDRERERHJ3n8BtWsWXF3HaGgAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "negative_loss = df[\"loss\"][::2].values\n",
    "positive_loss = df[\"loss\"][1::2].values\n",
    "\n",
    "loss_difference = positive_loss - negative_loss\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.hist(positive_loss, bins=30, alpha=0.5, label=\"Positive Loss\")\n",
    "plt.hist(negative_loss, bins=30, alpha=0.5, label=\"Negative Loss\")\n",
    "plt.xlabel(\"Loss\")\n",
    "plt.ylabel(\"Frequency\")\n",
    "plt.legend(loc=\"upper right\")\n",
    "plt.title(\"Histogram of Positive and Negative Loss\")\n",
    "plt.show()\n",
    "\n",
    "\n",
    "mean_loss_diff = np.mean(loss_difference)\n",
    "std_loss_diff = np.std(loss_difference)\n",
    "percentiles_loss_diff = np.percentile(loss_difference, [25, 50, 75])\n",
    "\n",
    "plt.hist(loss_difference, bins=200, alpha=0.5, label=\"loss difference\")\n",
    "plt.axvline(mean_loss_diff, color='r', linestyle='dashed', linewidth=1, label=f'Mean: {mean_loss_diff:.2f}')\n",
    "plt.axvline(percentiles_loss_diff[0], color='g', linestyle='dashed', linewidth=1, label=f'25th percentile: {percentiles_loss_diff[0]:.2f}')\n",
    "plt.axvline(percentiles_loss_diff[1], color='b', linestyle='dashed', linewidth=1, label=f'50th percentile (median): {percentiles_loss_diff[1]:.2f}')\n",
    "plt.axvline(percentiles_loss_diff[2], color='y', linestyle='dashed', linewidth=1, label=f'75th percentile: {percentiles_loss_diff[2]:.2f}')\n",
    "plt.xlabel(\"Loss\")\n",
    "plt.ylabel(\"Frequency\")\n",
    "plt.legend(loc=\"upper right\")\n",
    "plt.title(\"Histogram of loss difference\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "suno_env_dev",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.15"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}