modelling_1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "import os\n",
    "import time\n",
    "import random\n",
    "import pandas as pd\n",
    "import pickle\n",
    "import numpy as np\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.preprocessing import MaxAbsScaler, StandardScaler, LabelEncoder\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.metrics import mean_squared_error, r2_score\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from darts.datasets import TrafficDataset, AirPassengersDataset, AustralianTourismDataset\n",
    "from ucimlrepo import fetch_ucirepo "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import Data\n",
    "- We try multiple datasets to see which provides us with the best outcome"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>Month</th>\n",
       "      <th>#Passengers</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1949-01-01</td>\n",
       "      <td>112.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1949-02-01</td>\n",
       "      <td>118.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1949-03-01</td>\n",
       "      <td>132.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1949-04-01</td>\n",
       "      <td>129.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1949-05-01</td>\n",
       "      <td>121.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>139</th>\n",
       "      <td>1960-08-01</td>\n",
       "      <td>606.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>140</th>\n",
       "      <td>1960-09-01</td>\n",
       "      <td>508.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>141</th>\n",
       "      <td>1960-10-01</td>\n",
       "      <td>461.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>142</th>\n",
       "      <td>1960-11-01</td>\n",
       "      <td>390.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>143</th>\n",
       "      <td>1960-12-01</td>\n",
       "      <td>432.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>144 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         Month  #Passengers\n",
       "0   1949-01-01        112.0\n",
       "1   1949-02-01        118.0\n",
       "2   1949-03-01        132.0\n",
       "3   1949-04-01        129.0\n",
       "4   1949-05-01        121.0\n",
       "..         ...          ...\n",
       "139 1960-08-01        606.0\n",
       "140 1960-09-01        508.0\n",
       "141 1960-10-01        461.0\n",
       "142 1960-11-01        390.0\n",
       "143 1960-12-01        432.0\n",
       "\n",
       "[144 rows x 2 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "room_occupancy_estimation = fetch_ucirepo(id=864) \n",
    "room_df = room_occupancy_estimation['data']['features']\n",
    "room_df = pd.concat([room_df, room_occupancy_estimation['data']['targets']], axis = 1)\n",
    "room_df['Capacity %'] = room_df['Room_Occupancy_Count']/max(room_df['Room_Occupancy_Count'])\n",
    "\n",
    "airpass_raw = AirPassengersDataset().load().pd_dataframe().reset_index()\n",
    "airpass_raw.columns.name = None\n",
    "\n",
    "airpass_raw"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Airpassenger Model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data preprocessing\n",
    "- Scale the data\n",
    "    - Make sure each data entry is $0\\leq x\\leq 1$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>Month</th>\n",
       "      <th>LMin</th>\n",
       "      <th>LMax</th>\n",
       "      <th># Passengers</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1949-01-01</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.180064</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1949-02-01</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.189711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1949-03-01</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.212219</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1949-04-01</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.207395</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1949-05-01</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.194534</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>139</th>\n",
       "      <td>1960-08-01</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.974277</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>140</th>\n",
       "      <td>1960-09-01</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.816720</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>141</th>\n",
       "      <td>1960-10-01</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.741158</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>142</th>\n",
       "      <td>1960-11-01</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.627010</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>143</th>\n",
       "      <td>1960-12-01</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.694534</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>144 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         Month  LMin  LMax  # Passengers\n",
       "0   1949-01-01     0     0      0.180064\n",
       "1   1949-02-01     0     0      0.189711\n",
       "2   1949-03-01     0     1      0.212219\n",
       "3   1949-04-01     0     0      0.207395\n",
       "4   1949-05-01     1     0      0.194534\n",
       "..         ...   ...   ...           ...\n",
       "139 1960-08-01     0     0      0.974277\n",
       "140 1960-09-01     0     0      0.816720\n",
       "141 1960-10-01     0     0      0.741158\n",
       "142 1960-11-01     1     0      0.627010\n",
       "143 1960-12-01     0     0      0.694534\n",
       "\n",
       "[144 rows x 4 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## Find local minimums and plot the month\n",
    "airpass_data = airpass_raw.copy()\n",
    "airpass_data['LMin'] = 0\n",
    "airpass_data['LMax'] = 0\n",
    "passengers = []\n",
    "\n",
    "for row, index in airpass_data.iterrows():\n",
    "    passengers.append(index['#Passengers'])\n",
    "    \n",
    "for i in range(len(passengers)):\n",
    "    if i > 0 and i < len(passengers)-1:\n",
    "        if passengers[i] < passengers[i-1] and passengers[i] < passengers[i+1]:\n",
    "            airpass_data.loc[i, \"LMin\"] = 1\n",
    "        if passengers[i] > passengers[i-1] and passengers[i] > passengers[i+1]:\n",
    "            airpass_data.loc[i, \"LMax\"] = 1\n",
    "\n",
    "## Keep out these values to prevent any errors\n",
    "to_conc = airpass_data[['Month', \"LMin\", \"LMax\"]]\n",
    "\n",
    "## Scale train and test\n",
    "scaler = MaxAbsScaler()\n",
    "airpass_scaled_data = pd.DataFrame(scaler.fit_transform(np.array(airpass_data.drop(columns = ['Month', \"LMin\", \"LMax\"])))).rename(columns = {0:\"# Passengers\"})\n",
    "\n",
    "## Add month to both\n",
    "airpass_scaled_data = pd.concat([to_conc, airpass_scaled_data], axis = 1)\n",
    "airpass_scaled_data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plot scaled data\n",
    "- Looking at the data we see some trends\n",
    "    - Cyclical in nature\n",
    "    - Upward trend from 1950 onwards"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA+UAAAIiCAYAAACuWWkyAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy80BEi2AAAACXBIWXMAAA9hAAAPYQGoP6dpAADitklEQVR4nOzdd3hUVf7H8fdMeiMhhJCQQEJvsoA0V0BQFFRQNCLW/Ymg4rIq6rqWVRH72lbsurpiW0VlYy8oKi5dLCC9kwYhAUJLSEIm9/fHcIcMSchMMiWZfF7Pw5Pkzpk7ZyYX5XPPOd9jMQzDQERERERERER8zurvDoiIiIiIiIg0VwrlIiIiIiIiIn6iUC4iIiIiIiLiJwrlIiIiIiIiIn6iUC4iIiIiIiLiJwrlIiIiIiIiIn6iUC4iIiIiIiLiJwrlIiIiIiIiIn6iUC4iIiIiIiLiJwrlIiLikokTJ5Kenu7Rc6anpzNx4kSPnlPEl0aMGMFJJ53k7240Km+88QYWi4Wff/7Z310REWkSFMpFRJqgVatWMX78eNLS0ggPDyclJYWzzjqL5557zt9d84l+/fpx1113AbB3716sVivz58936bnp6elYLBbHn8TERIYNG8ZHH33kxR5LU7Zjxw5mzJjBihUr/N2VRuXFF1/kjTfe8Hc3RESaPIVyEZEmZvHixQwYMICVK1dy7bXX8vzzz3PNNddgtVp55pln/N09ryspKWH16tWccsopACxduhSr1crAgQNdPkffvn15++23efvtt7ntttvYsWMHGRkZvPzyy97qtjRhO3bs4P7771coP45CuYiIZwT7uwMiIuKehx9+mNjYWJYvX05cXJzTYwUFBf7plA/9+uuvVFRUOEL5kiVL6NWrF1FRUS6fIyUlhSuvvNLx8//93//RuXNnnn76aa6//nqP9zlQFRcXu/W5N2aB9F5ERKRp0Ui5iEgTs2XLFnr16lUtkAMkJiZWO/bOO+8waNAgIiMjadmyJaeddhrffPON4/FPPvmEMWPG0LZtW8LCwujUqRMPPvggNputzr5UVlYyc+ZMevXqRXh4OG3atGHKlCkUFRU5tTMMg4ceeojU1FQiIyM5/fTTWbNmjcvvuaSkhN27d7N7927+97//kZqaSlBQELt372bhwoX07t3b8Xh9JCUl0aNHD7Zt2wbA77//zsSJE+nYsSPh4eEkJSUxadIk9uzZ4/S8gwcPcvPNN5Oenk5YWBiJiYmcddZZ/Prrr442mzZt4qKLLiIpKYnw8HBSU1O59NJL2b9/v9O53nnnHfr3709ERATx8fFceuml5OTkOLUx1y+vXbuW008/ncjISFJSUnj88cervaesrCzOP/98oqKiSExM5JZbbmHu3LlYLJZqU/2XLVvG2WefTWxsLJGRkQwfPpxFixY5tZkxYwYWi4W1a9dy+eWX07JlS4YOHQpAfn4+V199NampqYSFhZGcnMy4cePYvn17nZ/9999/z7Bhw4iKiiIuLo5x48axbt06x+Nz5szBYrHw448/VnvuK6+8gsViYfXq1Y5j69evZ/z48cTHxxMeHs6AAQP49NNPnZ5nrnn+8ccfmTp1KomJiaSmptbYv/nz5ztmYVx99dWOZQ/HjxC78jspKyvjvvvuo3PnzoSFhdGuXTtuv/12ysrK6vyczN/977//zvDhw4mMjKRz587MmTMHgB9//JHBgwcTERFBt27dmDdvXrVz/Pbbb5xzzjm0aNGC6OhoRo4cydKlS2v8bBYtWsStt95K69atiYqK4sILL6SwsNDRLj09nTVr1vDjjz86PpMRI0ZUe78nOgfAzz//zOjRo0lISCAiIoIOHTowadKkOj8PEZFAopFyEZEmJi0tjSVLlrB69eo6C0zdf//9zJgxg1NPPZUHHniA0NBQli1bxvfff8+oUaMA+z/Co6OjufXWW4mOjub7779n+vTpHDhwgCeeeOKE558yZQpvvPEGV199NTfddBPbtm3j+eef57fffmPRokWEhIQAMH36dB566CHOPfdczj33XH799VdGjRpFeXm5S+/58ccf5/7773c61rp1a6ef//Of/wD2GwDuOnLkCDk5ObRq1QqAb7/9lq1bt3L11VeTlJTEmjVr+Ne//sWaNWtYunQpFosFgOuvv545c+Zwww030LNnT/bs2cPChQtZt24dJ598MuXl5YwePZqysjJuvPFGkpKSyMvL4/PPP2ffvn3ExsYC9tkP9957LxMmTOCaa66hsLCQ5557jtNOO43ffvvN6QZMUVERZ599NhkZGUyYMIE5c+Zwxx130Lt3b8455xzAPup7xhlnsHPnTqZNm0ZSUhLvvvsuP/zwQ7X3/v3333POOefQv39/7rvvPqxWK7NmzeKMM85gwYIFDBo0yKn9xRdfTJcuXXjkkUccn/VFF13EmjVruPHGG0lPT6egoIBvv/2W7OzsExYHnDdvHueccw4dO3ZkxowZHD58mOeee44hQ4bw66+/kp6ezpgxY4iOjuaDDz5g+PDhTs9///336dWrl+PvwZo1axgyZAgpKSnceeedREVF8cEHH3DBBRfw3//+lwsvvNDp+VOnTqV169ZMnz6d4uLiGvvYo0cPHnjgAaZPn851113HsGHDADj11FPd+p1UVlZy/vnns3DhQq677jp69OjBqlWrePrpp9m4cSMff/xxrZ9T1dcZO3Ysl156KRdffDEvvfQSl156Kf/5z3+4+eabuf7667n88st54oknGD9+PDk5OcTExDg+m2HDhtGiRQtuv/12QkJCeOWVVxgxYoQj0Fd144030rJlS+677z62b9/OzJkzueGGG3j//fcBmDlzJjfeeCPR0dHcfffdALRp08atcxQUFDBq1Chat27NnXfeSVxcHNu3byczM7POz0JEJKAYIiLSpHzzzTdGUFCQERQUZPzxj380br/9dmPu3LlGeXm5U7tNmzYZVqvVuPDCCw2bzeb0WGVlpeP7kpKSaq8xZcoUIzIy0igtLXUcu+qqq4y0tDTHzwsWLDAA4z//+Y/Tc7/++mun4wUFBUZoaKgxZswYp9f9+9//bgDGVVddVed73rJli/Htt98a33zzjREREWH87W9/M7799lvj6aefNgDjgw8+ML799lvj22+/rfNcaWlpxqhRo4zCwkKjsLDQWLlypXHppZcagHHjjTfW+pm89957BmD873//cxyLjY01/vKXv9T6Wr/99psBGB9++GGtbbZv324EBQUZDz/8sNPxVatWGcHBwU7Hhw8fbgDGW2+95ThWVlZmJCUlGRdddJHj2FNPPWUAxscff+w4dvjwYaN79+4GYPzwww+GYdivgy5duhijR4+udk106NDBOOussxzH7rvvPgMwLrvsMqd+FhUVGYDxxBNP1Poea9O3b18jMTHR2LNnj+PYypUrDavVavzf//2f49hll11mJCYmGhUVFY5jO3fuNKxWq/HAAw84jo0cOdLo3bu303VbWVlpnHrqqUaXLl0cx2bNmmUAxtChQ53OWZvly5cbgDFr1qxqj7n6O3n77bcNq9VqLFiwwOn5L7/8sgEYixYtOmEfzNd59913HcfWr19vAIbVajWWLl3qOD537txq/b3ggguM0NBQY8uWLY5jO3bsMGJiYozTTjvNccz8bM4880yna+KWW24xgoKCjH379jmO9erVyxg+fHi1vrp6jo8++sgAjOXLl5/wvYuIBDpNXxcRaWLOOusslixZwvnnn8/KlSt5/PHHGT16NCkpKU7TdD/++GMqKyuZPn06Vqvzf+7NkV6AiIgIx/cHDx5k9+7dDBs2jJKSEtavX19rPz788ENiY2M566yzHFPHd+/eTf/+/YmOjnaMys6bN4/y8nJuvPFGp9e9+eabXX7PHTt25Mwzz6RNmzYcPnyYa665hjPPPJOKigrS09O5+OKLOfPMMznzzDNdOt8333xD69atad26NX369OHDDz/kT3/6E4899li1z6S0tJTdu3c71rBXnZoeFxfHsmXL2LFjR42vY46Ez507l5KSkhrbZGZmUllZyYQJE5w+x6SkJLp06VJtdDs6OtppPXxoaCiDBg1i69atjmNff/01KSkpnH/++Y5j4eHhXHvttU7nWrFiBZs2beLyyy9nz549jtcuLi5m5MiR/O9//6OystLpOcevuY+IiCA0NJT58+dXW7ZwIjt37mTFihVMnDiR+Ph4x/E//OEPnHXWWXz55ZeOY5dccgkFBQVO0+7nzJlDZWUll1xyCWCvwv/9998zYcIEx3W8e/du9uzZw+jRo9m0aRN5eXlOfbj22msJCgpyuc+1ceV38uGHH9KjRw+6d+/u9Hs+44wzAGqcxVDT61x66aWOn7t160ZcXBw9evRwGuk2vzdf32az8c0333DBBRfQsWNHR7vk5GQuv/xyFi5cyIEDB5xe67rrrnP6+zps2DBsNhtZWVkufSaunMOcAfL5559z5MgRl88rIhJoFMpFRJqggQMHkpmZSVFRET/99BN33XUXBw8eZPz48axduxawrz23Wq307NnzhOdas2YNF154IbGxsbRo0YLWrVs7Asbx656r2rRpE/v37ycxMdERcM0/hw4dchSdM/8B3qVLF6fnt27dmpYtW9b5XsvKyhwB5ssvv6RNmzbEx8eze/du5s2bx+DBgx2PHx8gazN48GC+/fZb5s2bx+LFi9m9ezdvvfWWI4zv3buXadOm0aZNGyIiImjdujUdOnSo9pk8/vjjrF69mnbt2jFo0CBmzJjhFMQ6dOjArbfeymuvvUZCQgKjR4/mhRdecDrHpk2bMAyDLl26VPsc161bV614X2pqqlPQAWjZsqVTIM7KyqJTp07V2nXu3Nnp502bNgFw1VVXVXvt1157jbKysmrXgPk5mMLCwnjsscf46quvaNOmDaeddhqPP/44+fn5J/gNHLsuunXrVu2xHj16OG4OAI717ua0Z7BPXe/bty9du3YFYPPmzRiGwb333lvtvdx3331A9UKIx7+X+nLld7Jp0ybWrFlTrW9m/10p0ljT68TGxtKuXbtqxwDH6xcWFlJSUlLrZ11ZWVmtfkH79u2rvZ+q53RFXecYPnw4F110Effffz8JCQmMGzeOWbNmubTGXkQkkGhNuYhIExYaGsrAgQMZOHAgXbt25eqrr+bDDz90hJC67Nu3j+HDh9OiRQseeOABOnXqRHh4OL/++it33HHHCUNuZWUliYmJjrXcxzt+zXd9vffee1x99dUnPLcZ1rZt23bCNcymhISEE46qT5gwgcWLF/O3v/2Nvn37Eh0dTWVlJWeffbbTZzJhwgTHHufffPMNTzzxBI899hiZmZmOtcRPPfUUEydO5JNPPuGbb77hpptu4tFHH2Xp0qWkpqZSWVmJxWLhq6++qnHUNjo62unn2kZ2jXqspTffyxNPPEHfvn1rbHP861edRWC6+eabOe+88/j444+ZO3cu9957L48++ijff/89/fr1c7tfxwsLC+OCCy7go48+4sUXX2TXrl0sWrSIRx55pNp7ue222xg9enSN5zn+pkRN76U+XPmdVFZW0rt3b/75z3/W2Pb4YO3O63jymvDkOes6h8ViYc6cOSxdupTPPvuMuXPnMmnSJJ566imWLl1a7doTEQlUCuUiIgFiwIABgH1aMECnTp2orKxk7dq1tQau+fPns2fPHjIzMznttNMcx80q5CfSqVMn5s2bx5AhQ04YbtLS0gD7SGHVqbOFhYUujbqNHj2ab7/9FoCMjAyuu+46zj77bDZs2MANN9zA7NmzHQXakpKS6jxfXYqKivjuu++4//77mT59uuO4Oap8vOTkZKZOncrUqVMpKCjg5JNP5uGHH3aEcoDevXvTu3dv7rnnHhYvXsyQIUN4+eWXeeihh+jUqROGYdChQwfHqGlDpaWlsXbtWgzDcBpZ3bx5s1O7Tp06AdCiRQuXp/7XplOnTvz1r3/lr3/9K5s2baJv37489dRTvPPOO7X2EWDDhg3VHlu/fj0JCQlOW5RdcsklvPnmm3z33XesW7cOwzAcU9cBx7UVEhLS4PdyvONHp+ujU6dOrFy5kpEjR3rkfO5o3bo1kZGRtX7WVqvVpZsCx/PU+zjllFM45ZRTePjhh3n33Xe54oormD17Ntdcc41Hzi8i0thp+rqISBPzww8/1DhaZa7BNaeoXnDBBVitVh544IFqI97m882RrKrnKy8v58UXX6yzHxMmTMBms/Hggw9We6yiooJ9+/YBcOaZZxISEsJzzz3n9DozZ86s8zXAHnrPPPNMOnXqxMGDB7nyyis588wzCQoKIikpiUsuucSxnjw8PNylc55ITZ9JTf212WzVpnYnJibStm1bx/TbAwcOUFFR4dSmd+/eWK1WR5uMjAyCgoK4//77q72mYRjVtmFzxejRo8nLy3OqMVBaWsqrr77q1K5///506tSJJ598kkOHDlU7z/HbV9WkpKSE0tJSp2OdOnUiJibmhNOQk5OT6du3L2+++abjWgFYvXo133zzDeeee65T+zPPPJP4+Hjef/993n//fQYNGuQ0/TwxMZERI0bwyiuvOG5MufteamPeHKjaT3dNmDCBvLy8ar8DgMOHD9da/d0TgoKCGDVqFJ988onTNnW7du3i3XffZejQobRo0cLt80ZFRTXoMykqKqp2zZs3EDWFXUSaE42Ui4g0MTfeeCMlJSVceOGFdO/enfLychYvXsz7779Penq6Y6p3586dufvuu3nwwQcZNmwYGRkZhIWFsXz5ctq2bcujjz7KqaeeSsuWLbnqqqu46aabsFgsvP322y5NUR0+fDhTpkzh0UcfZcWKFYwaNYqQkBA2bdrEhx9+yDPPPMP48eNp3bo1t912G48++ihjx47l3HPP5bfffuOrr74iISHB5fe9aNEiYmJi6N27NwCLFy/mj3/8Y/0+xBNo0aKFY130kSNHSElJ4Ztvvqk2e+DgwYOkpqYyfvx4+vTpQ3R0NPPmzWP58uU89dRTgH27sRtuuIGLL76Yrl27UlFRwdtvv01QUBAXXXQRYA+wDz30EHfddRfbt2/nggsuICYmhm3btvHRRx9x3XXXcdttt7n1HqZMmcLzzz/PZZddxrRp00hOTuY///mP46aFOcJptVp57bXXOOecc+jVqxdXX301KSkp5OXl8cMPP9CiRQs+++yzE77Wxo0bGTlyJBMmTKBnz54EBwfz0UcfsWvXLqeiZDV54oknOOecc/jjH//I5MmTHVuixcbGMmPGDKe2ISEhZGRkMHv2bIqLi3nyySerne+FF15g6NCh9O7dm2uvvZaOHTuya9culixZQm5uLitXrnTjUzymU6dOxMXF8fLLLxMTE0NUVBSDBw92a036n/70Jz744AOuv/56fvjhB4YMGYLNZmP9+vV88MEHzJ071zHbxRseeughvv32W4YOHcrUqVMJDg7mlVdeoaysrMY91V3Rv39/XnrpJR566CE6d+5MYmKio3CdK958801efPFFLrzwQsdNt1dffZUWLVpUuykjIhLQfFztXUREGuirr74yJk2aZHTv3t2Ijo42QkNDjc6dOxs33nijsWvXrmrtX3/9daNfv35GWFiY0bJlS2P48OFOW4ctWrTIOOWUU4yIiAijbdu2ji3WqLJ1lmFU3xLN9K9//cvo37+/ERERYcTExBi9e/c2br/9dmPHjh2ONjabzbj//vuN5ORkIyIiwhgxYoSxevVqIy0tzaUt0QzDMK6//npj5MiRjp87d+5cr2240tLSjDFjxpywTW5urnHhhRcacXFxRmxsrHHxxRcbO3bsMADjvvvuMwzDvu3V3/72N6NPnz5GTEyMERUVZfTp08d48cUXHefZunWrMWnSJKNTp05GeHi4ER8fb5x++unGvHnzqr3mf//7X2Po0KFGVFSUERUVZXTv3t34y1/+YmzYsMHRZvjw4UavXr2qPbem383WrVuNMWPGGBEREUbr1q2Nv/71r8Z///tfA3DaPssw7Fu3ZWRkGK1atTLCwsKMtLQ0Y8KECcZ3333naGNuiVZYWOj03N27dxt/+ctfjO7duxtRUVFGbGysMXjwYOODDz444WdsmjdvnjFkyBAjIiLCaNGihXHeeecZa9eurbHtt99+awCGxWIxcnJyamyzZcsW4//+7/+MpKQkIyQkxEhJSTHGjh1rzJkzx9HG3LLLna24PvnkE6Nnz55GcHCw03Zj7vxOysvLjccee8zo1auX4+9j//79jfvvv9/Yv3//CV+/ttep7XoGqm3X9+uvvxqjR482oqOjjcjISOP00083Fi9e7NSmts/mhx9+qPbfhPz8fGPMmDFGTEyMATi2R3P1HL/++qtx2WWXGe3btzfCwsKMxMREY+zYscbPP/98ws9CRCTQWAyjAVVAREREpMmYOXMmt9xyC7m5uaSkpPi7OyIiIgIolIuIiASgw4cPV9tvvV+/fthsNjZu3OjHnomIiEhVWlMuIiISgDIyMmjfvj19+/Zl//79vPPOO6xfv77WLexERETEPxTKRUREAtDo0aN57bXX+M9//oPNZqNnz57Mnj3baRsxERER8T9NXxcRERERERHxE+1TLiIiIiIiIuInCuUiIiIiIiIifqJQLiIiIiIiIuInCuXSJFRWVrJt2zYqKyv93RVpQnTdSH3oupH60HUj9aHrRupD103gUSgXERERERER8ROFchERERERERE/USgXERERERER8ROFchERERERERE/USgXERERERER8ROFchERERERERE/USgXERERERER8ROFchERERERERE/USgXERERERER8ROFchERERERERE/USgXERERERER8ROFchERERERERE/USgXERERERER8ROF8gBQWFhIaGgoxcXFHDlyhKioKLKzs0/4nBkzZmCxWLBYLAQHB5Oens4tt9zCoUOHfNRrERERERERCfZ3B6ThlixZQp8+fYiKimLZsmXEx8fTvn37Op/Xq1cv5s2bR0VFBYsWLWLSpEmUlJTwyiuv+KDXjY/NZsNisWC16l6ViIiIiIj4htJHAFi8eDFDhgwBYOHChY7v6xIcHExSUhKpqalccsklXHHFFXz66acAvP322wwYMICYmBiSkpK4/PLLKSgocDy3qKiIK664gtatWxMREUGXLl2YNWsWAOXl5dxwww0kJycTHh5OWloajz76qOO5+/bt45prrqF169a0aNGCM844g5UrVzoenzFjBn379uXtt98mPT2d2NhYLrvsMqdR/IMHD3LFFVcQFRVFcnIyTz/9NCNGjODmm292tCkrK+O2224jJSWFqKgoBg8ezPz58x2Pv/HGG8TFxfHpp5/Ss2dPwsLCyM7OZv78+QwaNIioqCji4uIYMmQIWVlZrv9CREREREQCnM0G8+fDe+/Zv9ps/u5R0+XXkfI5c+bw0UcfsXnzZiZNmsSUKVNqbFdZWcnTTz/NZ599RmhoKFdddRVXXHGF1/rVp08fpwDqC4mJiU7BtC7Z2dn84Q9/AKCkpISgoCDeeOMNDh8+jMViIS4ujssvv5wXX3zR5XNGRERQXl4OwJEjR3jwwQfp1q0bBQUF3HrrrUycOJEvv/wSgHvvvZe1a9fy1VdfkZCQwObNmzl8+DAAzz77LJ9++ikffPAB7du3Jycnh5ycHMfrXHzxxURERPDVV18RGxvLK6+8wsiRI9m4cSPx8fEAbNmyhY8//pjPP/+coqIiJkyYwMsvv8xzzz0HwK233sqiRYv49NNPadOmDdOnT+fXX3+lb9++jte54YYbWLt2LbNnz6Zt27Z89NFHnH322axatYouXbo4PrvHHnuM1157jVatWhEfH0/fvn259tpree+99ygvL+enn37CYrG4/DmKiIiIiASyzEyYNg1yc48dS02FZ56BjAz/9aup8msoT0hI4LrrruPrr78+Ybv//ve//PLLL2RmZnLo0CGmTJlCly5dGDRokI962vi0bduWFStWcODAAQYMGMCyZcuIioqib9++fPHFF7Rv357o6GiXz/fLL7/w7rvvcsYZZwAwadIkx2MdO3bk2WefZeDAgRw6dIjo6Giys7Pp168fAwYMACA9Pd3RPjs7my5dujB06FAsFgtpaWmOxxYuXMhPP/1EQUEBYWFhADz55JN8/PHHzJkzh+uuuw6w34h54403iImJAeDKK69k3rx5gH2U/M033+Tdd99l5MiRAMyaNYu2bds69WHWrFlkZ2c7jt922218/fXXzJo1i0ceeQSw33x48cUX6dOnDwB79+5l//79jB07lk6dOgHQo0cPlz9HEREREZFAlpkJ48eDYTgfz8uzH58zR8HcXX4N5SNGjABg0aJFJ2z35ZdfcuWVVxIfH098fDwXXHABX3zxhddCuTsj1v5iFmf74IMPGDhwIH/4wx9YtGgRbdq04bTTTnPpHKtWrSI6OhqbzUZ5eTljxozh+eefB+whfcaMGaxcuZKioiIqKysBe9jt2bMnf/7zn7nooov49ddfGTVqFBdccAGnnnoqABMnTuSss86iW7dunH322YwdO5ZRo0YB9s/20KFDtGrVyqkvhw8fZsuWLY6f09PTHYEcIDk5mT179gCwdetWjhw54vT7j42NpVu3bk7vzWaz0bVrV6fXKSsrc3rt0NBQx4wDgPj4eCZOnMjo0aM566yzOPPMM5kwYQLJyckufaYiIiIiIoHKZrOPkB8L5IeAICACwwCLBW6+GcaNg6Agv3WzyWkShd62bt3qmG4M0LlzZxYuXFhr+/Lycsc0bFNwcDChoaFe66Ov9e7dm6ysLI4cOUJlZSXR0dFUVFRQUVFBdHQ0aWlprFq1qtbnG4ZBt27d+PjjjwkODqZt27aOz+fgwYOMHj2aUaNG8fbbb9O6dWuys7M555xzKC0tpbKyktGjR7Nt2za+/PJL5s2bx8iRI5k6dSpPPPEEffv2ZcuWLXz11Vd89913TJgwgZEjR/Lhhx9y8OBBkpOT+f7776v1KS4ujsrKSgzDICQkxHEjwOxvZWWl4w/g9P3x7Q4cOEBQUBDLly8n6Lj/IkRHRzueGxERgWEYGFVu9f373//mhhtuYO7cubz//vvcc889zJ07l1NOOcX9X5T4VdVrRcRVum6kPnTdSH3oupH68Od1s3Ah7NsH0dH2f3eXlPTCYkkjIuIHx3LPoiJ7u2HDfN69RsmVItJNIpQfPnyYqKgox89RUVGUlJTU2n7WrFm8+uqrTscuvvhiJkyY4LU++torr7zCkSNHuPLKK7nzzjs56aSTuOmmmxg/fjynnXYaISEhJyxOtm/fPgBCQkIA2Llzp+OxVatWsWfPHv7yl784pn6vW7fO0a5ly5aOtiNGjGDEiBH07NmTf/zjH9xwww2Ox0455RROOeUUhg0bxsSJE1m5ciVt27YlPz+fXbt2kZqa6tSn4uJiiouL2bdvH+Xl5U79N/ubk5NDcHAwISEhfPnll5xzzjkAHDhwgA0bNtCnTx+ysrJITEzEZrPx+++/V5tRUVZWRlZWFnv27KGysrLGzyk+Pp7LLruMyy67jIsuuoh//etfGi1vwqrWNBBxla4bqQ9dN1Ifum6kPvxx3bRvD7//bv9+//799OuXjWFk8/LL71UrNq06yXYdOnSos02TCOUREREUFxc7fi4uLiYyMrLW9ldffXW1QnCBNlKelpZGfn4+e/bsYfLkyVgsFjZv3sw111zjUniMi4sjNDTUab23KTIyktDQUD7++GOmTJnC6tWrefnllwH7NPK0tDTuu+8+Tj75ZHr16kVZWRlLliyhZ8+epKWl8fTTT5OUlES/fv2wWq3873//Iykpid69e/OHP/yBV199lRtvvJF//OMfdO3alR07dvDll19ywQUXMGDAgBr7FhcXB0C7du2wWq383//9H08++STdunUjMTGRBx54gKCgIGJjY0lLSyMtLY3LL7+cO++8kyeeeIJ+/fpRWFjI999/T+/evRkzZgytWrXCarU6vc62bdt49dVXOe+882jbti0bNmwgOzubyZMn1/hZSeNWWVlJTk6O47oRcYWuG6kPXTdSH7pupD78ed0sWABjx5r92Og4PnHim0REXO74+fPPNVLujiYRyjt27MjmzZsdU9i3bNlCx44da20fGhoaUAG8Nv/73/8YOHAgkZGRLFiwgNTUVFJSUlx6rjm9pKa/yG3atOGNN97g73//O8899xwnn3wyTz75JOeffz5WqxWr1UpYWBh3330327dvJyIigmHDhjF79mysVistWrTgySefZNOmTQQFBTFw4EC+/PJLgoPtl9uXX37J3XffzeTJkyksLCQpKYnTTjuN5ORkrFZrjX0zvzdf/+mnn+b666/n/PPPp0WLFtx+++3k5uYSERHhaPvGG2/w0EMP8be//Y28vDwSEhI45ZRTOO+88xznOf51oqOj2bBhA2+99RZ79uwhOTmZv/zlL/z5z3/W/yybsKq/bxFX6bqR+tB1I/Wh60bqwx/XzdChEBdnL+pmGHscx222eRw6tAKL5WRSU+3tdEm7zmIYx9fN852KigpsNhtPPPEE8fHxTJ48meDg4GprgD/44AM+/vhjXnjhBUf19RkzZjTr6uvNjTnNPC0trcb/+BQXF5OSksJTTz3F5MmT/dBDaYzqum5EaqLrRupD143Uh64bqQ9/XzfHqq9/DFwI9ALWAJdgscxW9fV68Ovf/n//+98MGTKEjz/+mNdff50hQ4bw5Zdf8ttvvzGsynyH8ePH079/fy688EImTZrE5ZdfrkDezP3222+89957bNmyhV9//dWxXGHcuHF+7pmIiIiISODKyLBvexYXV3D0yHVAJ+BDnn9+iwJ5Pfh1+vqUKVOYMmVKjY8tWLDA8b3VauWvf/0rf/3rX33VNWkCnnzySTZs2EBoaCj9+/dnwYIFJCQk+LtbIiIiIiIBLSMD1qwpZPp0uOmmJIKCbuPpp//M6tVPAS/6u3tNjubJSJPUr18/fvnlFw4dOsTevXv59ttv6d27t7+7JSIiIiLSLOzZUwjABRe05uGHryIxMZFZs2ZRUFBQxzPleArlIiIiIiIi4pbCQnsob926NREREdx0002Ulpby3HPP+blnTY9CuYiIiIiIiLilaigHmDp1KtHR0Y7i3OI6hXIRERERERFxixnKW7VqBUDLli257rrrKCoq4tVXX/Vn15ochXIRERERERFxS2FhIa1atSI4+Fjt8FtuuYWQkBD++c9/Ul5e7sfeNS0K5SIiIiIiIuIywzAoKChwTF03paamcsUVV5Cbm8t7773np941PQrlIiIiIiIi4rIDBw5w5MiRaqEc4G9/+xsAr7zyiq+71WQplEu9zJgxg759+3rsfPPnz8disbBv3z6PnVNERERERDzv+CJvVfXs2ZOEhASysrJ83a0mS6G8iZo4cSIXXHCBv7tRq+3bt2OxWAgKCiIvL8/psZ07dxIcHIzFYmH79u0AnHrqqezcuZPY2Fg/9FZERERERFx1olAOkJCQwJ49ezAMw5fdarIUyj3FZoP58+G99+xfbTZ/96hRSElJ4a233nI69uabb5KSkuJ0LDQ0lKSkJCwWiy+7JyIiIiIibqorlLdq1YqysjKKi4t92a0mS6HcEzIzIT0dTj8dLr/c/jU93X7cT3788UcGDRpEWFgYycnJ3HnnnVRUVDger6ys5PHHH6dz586EhYXRvn17Hn74Ycfjd9xxB127diUyMpKOHTty7733cuTIEbf7cdVVVzFr1iynY7NmzeKqq65yOnb89PU33niDuLg45s6dS48ePWjRogUTJ05k586dbvdBREREREQ8xwzliYmJNT6ekJAAwJ49e3zWp6ZMobyhMjNh/HjIzXU+npdnP+6HYJ6Xl8e5557LwIEDWblyJS+99BL//ve/eeihhxxt7rrrLv7xj39w7733snbtWt59913atGnjeDwmJoY33niDtWvX8swzz/Dqq6/y9NNPu92X888/n6KiIhYuXAjAwoULKSoq4rzzzqvzuSUlJTz55JO8/fbbzJ8/nx07djgKR4iIiIiIiH8UFBQAJx4pB4VyVwXX3URqZbPBtGlQ01oJwwCLBW6+GcaNg6Agn3XrxRdfpF27djz//PNYLBa6d+/Ojh07uOOOO5g+fTrFxcU888wzPP/8844R606dOjF06FDHOe655x7H9+np6dx2223Mnj2b22+/3a2+hISEcOWVV/L6668zdOhQXn/9da688kpCQkLqfO6RI0d4+eWX6dSpE5WVlfzpT3/ixRdfdOv1RURERETEs1yZvg6we/dun/WpKVMob4gFC6qPkFdlGJCTY283YoTPurVu3Tr++Mc/Oq3PHjJkCIcOHSI3N5f8/HzKysoYOXJkred4//33efbZZ9myZQuHDh2ioqKCFi1a1Ks/kyZN4tRTT+WRRx7hww8/ZMmSJU5T6WsTGRlJp06dHD8nJiY67sqJiIiIiIh/uFLoDTRS7ipNX28IV9c3N7J10BERESd8fMmSJVxxxRWce+65fP755/z222/cfffdlJeX1+v1evfuTffu3bnsssvo0aMHJ510kkvPO3403WKxqIKjiIiIiIifaaTcsxTKGyI52bPtPKRHjx4sWbLEKcAuWrSImJgYUlNT6dKlCxEREXz33Xc1Pn/x4sWkpaVx9913M2DAALp06dLgfQYnTZrE/PnzmTRpUoPOIyIiIiIi/mWGcnNE/HgaKXePpq83xLBhkJpqL+pW0wiuxWJ/fNgwr7z8/v37WbFihdOxVq1aMXXqVGbOnMmNN97IDTfcwIYNG7jvvvu49dZbsVqthIeHc8cdd3D77bcTGhrKkCFDKCwsZM2aNUyePJkuXbqQnZ3N7NmzGThwIF988QUfffRRg/p67bXXcvHFFxMXF9eg84iIiIiIiH8VFhbSsmXLWutEqdCbexTKGyIoCJ55xl5l3WJxDubmeu6ZM71W5G3+/Pn069fP6djkyZN57bXX+PLLL/nb3/5Gnz59iI+PZ/LkyU7F2+69916Cg4OZPn06O3bsIDk5meuvvx6wV0y/5ZZbuOGGGygrK2PMmDHce++9zJgxo959DQ4OrvVOmoiIiIiINA2GYVBQUEBaWlqtbTR93T0WQ4t0Gy4z016FvWrRt3bt7IE8I8Nv3QoklZWVZGVlkZaWhtWqVRfiGl03Uh+6bqQ+dN1Ifei6kfrw93Vz8OBBWrRowZAhQxzbHh+vsLCQxMREzjrrLL755hsf97Dp0Ui5J2Rk2Lc9W7DAXtQtOdk+Zd2H26CJiIiIiIh4W11F3gBatmwJaKTcVQrlnhIU5NNtz0RERERERHzNlVAeHBxMy5YttabcRZonIyIiIiIiIi5xJZSDfV25QrlrFMpFRERERETEJe6E8uLiYkpLS33RrSZNoVxERERERERcYobyxMTEE7bTXuWuUygXERERERERlxQUFACujZSDir25QqFcREREREREXOLq9HWNlLtOoVxERERERERc4s6aclAod4VCuYiIiIiIiLjEDOXmSHhtNH3ddQrlUi8zZsygb9++/u6GiIiIiIj4UGFhIS1atCAsLOyE7TR93XUK5R5is8H8+fDee/avNpt3X2/ixIlccMEF3n2RBti+fTsWi4WgoCDy8vKcHtu5cyfBwcFYLBa2b9/unw6KiIiIiIjbCgsL66y8Dhopd4dCuQdkZkJ6Opx+Olx+uf1rerr9eHOXkpLCW2+95XTszTffJCUlxU89EhERERGR+iguLqakpKTO9eSgkXJ3KJQ3UGYmjB8PubnOx/Py7Mf9Fcx//PFHBg0aRFhYGMnJydx5551UVFQ4Hq+srOTxxx+nc+fOhIWF0b59ex5++GHH43fccQddu3YlMjKSjh07cu+993LkyBG3+3HVVVcxa9Ysp2OzZs3iqquucjpms9mYPHkyHTp0ICIigm7duvHMM884Hi8tLWX06NFMmTLFcWzLli3ExMTw+uuvu90vERERERFxj6tF3kCF3tyhUN4ANhtMmwaGUf0x89jNN3t/Kvvx8vLyOPfccxk4cCArV67kpZde4t///jcPPfSQo81dd93FP/7xD+69917Wrl3Lu+++S5s2bRyPx8TE8MYbb7B27VqeeeYZXn31VZ5++mm3+3L++edTVFTEwoULAVi4cCFFRUWcd955Tu0qKytJTU3lww8/ZO3atUyfPp2///3vfPDBBwCEh4fz9NNP89Zbb/HJJ59gs9m48sorOeuss5g0aVJ9PiYREREREXFDfUK5pq/XLdjfHWjKFiyoPkJelWFATo693YgRPusWL774Iu3ateP555/HYrHQvXt3duzYwR133MH06dMpLi7mmWee4fnnn3eMWHfq1ImhQ4c6znHPPfc4vk9PT+e2225j9uzZ3H777W71JSQkhCuvvJLXX3+doUOH8vrrr3PllVcSEhJSrd3999/v+LlDhw4sWbKEDz74gAkTJgDQs2dPHnzwQa655houvfRSsrKy+Pzzz93+fERERERExH3uhPLQ0FBiYmI0Uu4ChfIG2LnTs+08Zd26dfzxj3/EYrE4jg0ZMoRDhw6Rm5tLfn4+ZWVljBw5stZzvP/++zz77LNs2bKFQ4cOUVFRQYsWLerVn0mTJnHqqafyyCOP8OGHH7JkyRKnqfSmF154gddff53s7GwOHz5MeXl5tQrvt956K5988gnPP/88X331leMOnIiIiIiIeJc7oRzso+UaKa+bpq83QHKyZ9v5SkRExAkfX7JkCVdccQXnnnsun3/+Ob/99ht333035eXl9Xq93r170717dy677DJ69OjBSSedVK3N7Nmzue2225g8eTLffPMNK1as4Oqrr672mgUFBWzcuJGgoCA2bdpUr/6IiIiIiIj7zFDuSvV1sBd7O3DgQL1qUzUnCuUNMGwYpKZClQFpJxYLtGtnb+dLPXr0YMmSJRhVFrsvWrSImJgYUlNT6dKlCxEREXz33Xc1Pn/x4sWkpaVx9913M2DAALp06UJWVlaD+jRp0iTmz59f6/rvRYsWceqppzJ16lT69etH586d2bJlS7V2kydPpnfv3rz55pvccccdrFu3rkH9EhERERER19RnpBxg7969XutTIND09QYICoJnnrFXWbdYnAu+mUF95kx7O2/Yv38/K1ascDrWqlUrpk6dysyZM7nxxhu54YYb2LBhA/fddx+33norVquV8PBw7rjjDm6//XZCQ0MZMmQIhYWFrFmzhsmTJ9OlSxeys7OZPXs2AwcO5IsvvuCjjz5qUF+vvfZaLr74YuLi4mp8vEuXLrz11lvMnTuXDh068Pbbb7N8+XI6dOjgaPPWW2+xdOlSfv/9d9q1a8cXX3zBFVdcwdKlSwkNDW1Q/0RERERE5MQKCgoA90P57t27nYpKizONlDdQRgbMmQPHb7udmmo/npHhvdeeP38+/fr1c/pz//33k5KSwpdffslPP/1Enz59uP7665k8ebJT8bZ7772Xv/71r0yfPp0ePXpwySWXOP6SnX/++dxyyy3ccMMN9O3bl8WLF3Pvvfc2qK/BwcEkJCQQHFzzfaApU6aQkZHBJZdcwuDBg9mzZw9Tp051PL5+/Xr+8Y9/8Pzzz9OuXTvAXtBu9+7dDe6biIiIiIjUzd2Rcu1V7hqLYdS0oZe4y2azV1nfudO+hnzYMO+NkDdHlZWVZGVlkZaWhtWqe0niGl03Uh+6bqQ+dN1Ifei6kfrw53UzePBgfvrpJw4fPkx4eHid7R944AHuu+8+/vvf/5LhzdHKJk7T1z0kKMi3256JiIiIiIh/NNcBucLCQqKjo10K5KCRclcplIuIiIiIiLgoMxOmTYPc3GPHUlPttaYCfTC4sLDQ5crrcGxNuUL5iWmejIiIiIiIiAsyM+1FnqsGcoC8PPvxzEz/9MsXDh8+zKFDh1xeTw7Ohd6kdgrlIiIiIiIidbDZ7CPkxypylQP2H8xjN99sbxeI3C3yBpq+7iqFchERERERkTosWFB1hDwLaA9MczxuGJCTY28XiOoTyjVS7hqFchERERERkTrs3Gl+VwlMBnYBy07QLrA0JJRrpPzEFMpFRERERETqkJxsfvcK8N3R76uPAB9rF1jqE8ojIyOJiIhQKK+DQrmIiIiIiEgdhg2DpKStwN+ASCAOOBY2LRZo187eLhDVJ5SDfbRc09dPTKFcRERERESkDhZLJfHxVwPFwONAJ2A/cASLxd5m5szA3a/cDOXubIkG9mJvRUVF2AK1Ap4HKJSLiIiIiIjU4dlnn2Xt2v/Ru/cZpKT8GUg4+sgeUlNhzpzA3qe8oKAAqN9IuWEYFBUVeaNbAUGhXERERERE5AQ2btzIXXfdRXR0NJ9++m+ysqyceaY9lL/++m62bQvsQA71n76ubdHqplAuIiIiIiJSC5vNxsSJEyktLeWf//wn6enpBAVBr172sNmhw+6AnbJeVUPWlINC+YkolIuIiIiIiNTiqaeeYsmSJZx99tlcc801juPNLWwWFhYSGRlJZGSkW8/TXuV1UygXERERERGpwYEDB5g+fTqxsbG8+uqrWMyKbhyblt1cwmZhYaHbo+Sg6euuUCgXERERERGpQVZWFmVlZYwdO5bU1FSnx5pTKC8rK+PAgQNuV14HjZS7QqFcRERERESkBjt37gQgOTm52mPNKZSb71Ej5d6hUC4iIiIiIlKD/Px8AJKSkqo91pxCeX23Q4Pmt/a+PhTKRUREREREauBKKG8OYbO+lddB09ddoVAuIiIiIiJSgxOF8uYUNhsSypvTzYv6UigXERERERGpwYnWlIeGhhITE6NQXofo6GhCQkKaxedUXwrlIiIiIiIiNTjRSDnYR4GbQ9g0Q3l9qq9bLBYSEhI0Un4CCuUiIiIiIiI1yM/PJyQkhJYtW9b4eEJCAgcPHqS8vNzHPfOthoyUg32q/549ezAMw5PdChgK5SIiIiIiIjXIz88nKSkJi8VS4+PNZb10Q6qvgz2U22w29u/f78luBQyFchERERERkeOUlpayb9++GteTm5pLsbeGjpQ3l5sX9aVQLiIiIiIicpy61pND89mrPD8/n8jISKKiour1/OZy86K+FMpFRERERESOo1BuV1lZSW5uLu3bt691Gn9dNFJ+YgrlIiIiIiIix1Eot9u1axfl5eW0b9++3ucwR8oVymumUC4iIiIiInIcd0J5IIfN7OxsAI+E8kC+edEQCuUiIiIiIiLHMUP5iQq9NYeRcjOUp6Wl1fsczeHmRUMolIuIiIiIiBxn586dwIlHypvDCLBGyr1PoVxEREREROQ4WlNul5WVBTQslGuk/MQUykVERERERI5jhvI2bdrU2qY5jAB7cqRcobxmCuUiIiIiIiLHyc/PJzY2loiIiFrbhISEEBsbG9BhMzs7G4vFQkpKSr3PERsbi9VqDeibFw2hUC4iIiIiIlKFYRjk5+efsMibKSEhIaDDZnZ2NklJSYSFhdX7HFarlVatWgX0zYuGUCgXERERERGpoqioiPLy8hOuJze1atWKQ4cOUVpa6oOe+VZxcTF79uxpUOV1U6tWrdi9ezeGYXigZ4FFoVxERERERKQKV4q8mQK5iJkn1pObEhISKC8vp7i4uMHnCjQK5SIiIiIiIlXUJ5QH4hR2T4ZyFXurnUK5iIiIiIhIFWYod3VNOQRm2PRGKA/EmxcNpVAuIiIiIiJShUbK7Tw9fR0C8+ZFQymUi4iIiIiIVLFz505AodwM5Z4q9AaB+Tk1lEK5iIiIiIhIFe6MlAdy2NRIuW8olIuIiIiIiFSh6et2WVlZREVF0bJlywafK5BvXjSUQrmIiIiIiEgV+fn5BAUFOQL3iQRqKLfZbOTm5tK+fXssFkuDz6fq67VTKBcREREREakiPz+fNm3aYLXWHZcCdVr2rl27OHLkiEemrkPgfk6eEOzvDoiIiIiISONjs8GCBbBzJyQnw7BhEBTk7155X3l5Obt37+bkk092qX18fDwQeCPlnlxPDpq+fiIK5SIiIiIi4iQzE6ZNg9zcY8dSU+GZZyAjw3/98oWCggLAtfXkAMHBwcTFxQVc2PRk5XXAsS5dI+XVafq6iIiIiIg4ZGbC+PHOgRwgL89+PDPTP/3yFXeKvJkSEhICLpRnZWUBnhspDw4OJjExkZycHI+cL5AolIuIiIiICGCfsj5tGhiGeWQrYA+p5rGbb7a3C1RmKE9OTnb5OQkJCZSUlHD48GFvdcvnPD19HaBbt27s3r2bvXv3euycgUChXEREREREAPsa8mMj5DuAPsAljscNA3Jy7O0CVX1HyiGwpmZ7K5QDbNiwwWPnDAQK5SIiIiIiAtiLuh3zIHAI2FJHu8Cy8+ibq08oD6Qp7NnZ2VgsFlJSUjx2ToXymimUi4iIiIgIYK+ybrcJePXo99VHf92Y2d3kNGSkPNBCedu2bQkNDfXYORXKa+b3UF5UVMS0adMYOnQoGRkZ/PTTTzW227FjBzfccAMjRozgnHPO4bXXXvNxT0VEREREAtuwYfYq6zAdsGGPC6VACQAWC7RrZ28XqOqzpjzQtvs6dOgQe/fu9ejUdVAor43fQ/ljjz1Gq1atmDdvHtOmTeOuu+5i//791do98cQTJCUlMW/ePF577TXmzJnDkiVL/NBjEREREZHAFBQE06b9BswGugKjjj6yB4vF/t3MmYG9X7kZytu0aePycwJtpNwb68kBOnToQHBwsEL5cfwayktKSpg/fz5TpkwhPDyc4cOH06lTJ3788cdqbXfs2MGZZ55JcHAwKSkp9O3bl61bt/qh1yIiIiIigev77+8GID7+QcAMpntITYU5cwJ/n/L8/Hyio6OJjo52+TmBVujNW6E8JCSEzp07s2nTJioqKjx67qYs2J8vnp2dTWRkpNNdqM6dO9cYti+++GK+/fZbTj75ZPLz81m1ahVXX311jectLy+nvLzc6VhwcLBH10OIb1VWVjp9FXGFrhupD103Uh+6bqQ+GuN18+OPP/LVV19x8skns3hxBn/60xI+/BCeeqqQG2+sJCgIGlF3Pc4wDHbu3ElKSopbv5f4+HgACgsLvf779MV1s337dgDatWvn8dfp2rUr69evZ+vWrXTu3Nmj526MrNa6x8H9GsoPHz5MVFSU07GoqKgap6/369ePzMxMhg0bhs1m4/rrr3esSTjerFmzePXVV52OXXzxxUyYMMFznRe/yMnJ8XcXpAnSdSP1oetG6kPXjdRHY7luDMPgtttuA2DatGns2JFD+/b2eephYRvJzQ38AHXw4EEOHz5MXFwcWVlZLj/PHBDMyspy63kN4c3rZvXq1QCEh4d7/P2YBfQWLFhASEiIR8/dGHXo0KHONn4N5RERERQXFzsdKy4uJjIy0umYzWbjpptu4oorrmDChAns2rWLadOm0aVLF4YPH17tvFdffTVXXHGF0zGNlDdtlZWV5OTk0K5dO5fuNomArhupH103Uh+6bqQ+Gtt18+mnn/Lrr79y+umnc8UVV2CxWOjYsSMAFouFtLQ0P/fQ+zZu3AhAenq6W+/XzC+HDx/2+ufki+vGHCQdMGCAx9/PwIED+de//sW+ffuaxTXlCr+G8vbt21NSUkJBQQGJiYkAbNmyhTFjxji1O3DgALt27WL8+PGONeVDhw5l+fLlNYby0NBQBfAAZbVaG8X/tKRp0XUj9aHrRupD143UR2O4bmw2G/fccw8Ajz76KEFHK7m1bt0asO+Y5O8++kJBQQFgr7zuzvtt1aoVFouFPXv2+Oxz8uZ1Y64pT09P9/hr9OjRA7DfAGkO15Qr/PopREZGMnz4cF555RVKS0tZsGABmzdvrha0W7ZsSVJSEh999BGVlZXk5+ezcOHCZrEGQURERETE2/7zn/+wZs0aLrjgAgYPHuw4bm71FSgFzOpSnz3KwT4rt2XLlgHzOWVnZxMdHU1cXJzHz61t0arz+62JO++8k8LCQkaOHMnTTz/NI488QmxsLF999ZXTGvDHHnuMuXPncvrpp3PVVVcxZMgQzj//fD/2XERERESk6TMMgxkzZmC1WnnooYecHjMLmAVK2KzLzp07AfdDOdgrsAfClmg2m43c3Fzat2+PxdwHz4MSEhKIj49n/fr1Hj93U+XX6etgHwV/9tlnqx0/55xzOOeccxw/9+rVi9dff92XXRMRERERCXgFBQVs27aNESNG0KtXL6fHzJHyvXv3+qNrPlffkXKwh82NGzdSUlJSrUZWU5Kfn09FRYXHt0Orqlu3bixZsoT9+/cTGxvrtddpKvw+Ui4iIiIiIv6zbds2ADp16lTtseY6fT05Odnt55p7lTf10XJv7VFeVffu3QFNYTcplIuIiIiINGNmKK9p66bIyEjCw8ObXSivz0i5eQMjUEK5Nyuja125M4VyEREREZFm7EShHOxhszmFcovF4qg67w5zpLypf1bmvuTenr4OCuUmhXIRERERkWbMlVBeVFREZWWlL7vlFzt37qR169YEB7tfekvT112nUO5MoVxEREREpBnbvn07UHsoj4+Pp7Kykn379vmuU35gs9koLCys19R1UCh3R6dOnQgKClIF9qMUykVEREREmrFt27YRERFBmzZtany8uVRgLywspLKysl5F3iCwQrnVaiUlJcVrrxEaGkqHDh3YtGkTNpvNa6/TVCiUi4iIiIg0UzabjezsbNLT02vdk7q5VGBvSJE3CKxCb23btiUkJMSrr9OtWzfKysocI/PNmUK5iIiIiEgzlZeXx5EjR0hPT6+1TXMJ5Tt37gTqH8oDodDbwYMHKSoq8urUdZO2RTtGoVxEREREpJmqq8gbNJ9Q3tCR8kCYvu6L9eQmFXs7RqFcRERERKSZciWUx8fHA80nlNd3TXlcXBxWq1Wh3EUK5ccolIuIiIiINFMaKT+moSPlQUFBxMfHK5S7yAzlqsCuUC4iIiIi0mzVtR0aNJ/q6w0N5WCfwr57924Mw/BUt3zKl6E8MTGR2NhYjZSjUC4iIiIi0mxppPyYhhZ6A/tnVVZWRklJiae65VNmKE9LS/P6a1ksFrp168aOHTs4ePCg11+vMVMoFxEREZFmy2aD+fPhvffsX5vblsnbtm0jNjaWli1b1tqmuYTy/Px8wsPDadGiRb3P0dSLvWVlZQG+GSmHY1PYN27c6JPXa6wUykVERESkWcrMhPR0OP10uPxy+9f0dPvx5qCsrIy8vLwTjpIDjsAeyKHcMAzy8vJo27Ztrfu1u6Iph/J169axbNky2rRpQ2xsrE9eU9ui2SmUi4iIiEizk5kJ48dDbq7z8bw8+/HmEMyzs7MxDOOEe5QDBAcHExsbG9ChfOfOnZSUlNC5c+cGnaephnKbzcbkyZMpLy/n0UcfbdCNCXeoArudQrmIiIiINCs2G0ybBjXV4jKP3Xxz4E9ld2U9ualVq1YBXeht06ZNAM02lD///PMsWbKEUaNGMXHiRJ+9riqw2ymUi4iIiEizsmBB1RHyrcBrwP8cjxsG5OTY2wUyVyqvm1q1asWhQ4coLy/3cq/8Y/PmzQB06dKlQedpiuvvt27dyt///neio6P517/+5bNRcrDfBLFYLM1+pDzY3x0QEREREfGVoqIiPvnke+Dbo3+2Hn0kEdjl1PZoMe6A5e5IOdjDZnJyslf75Q/NdaTcMAyuvfZaSkpKeP75531Sdb2q8PBw0tPT2bhxI5WVlVitzXPMuHm+axERERFpdiZPnkxCQgIzZ44HXgFygdOBFKAAcN6WKQCzp5P6hvJA5KmR8qYWyl977TW+//57hg0bxp///Ge/9KFbt24cPnyY3OMLPDQjCuUiIiIiEvDKysp48803adGiBTfffCsJCV8BRcD3wLCjrXIAsFigXTsYNqyWkwUIM5TXVegNID4+HgjcUL5p0yasVqtLNyhOpCmF8tzcXG677TbCw8N57bXX/DZKrWJvCuUiIiIi0gxs2LABm83G6NGjefrpp3jllbOxWCKxL58192TOxlxOO3MmBAX5p6++sm3bNhITE4mKiqqzbSCPlBuGwebNm0lLSyM0NLRB5zJDeUFBgSe65jWGYXD99ddz4MABHnjgAbp27eq3vmhbNIVyEREREWkG1q5dC0DPnj0ByMiAOXMgJQXAXEebRWqq/XhGhl+66TOHDh2isLDQ5ZFhM5QHYgV2czu0hk5dB4iLiyMqKors7GwP9Mx7Zs+ezRdffMGAAQO45ZZb/NoXVWBXKBcRERGRZmDNmjUA9OrVy3EsIwO2b4dHHrGPlF9xRTbbtgV+IIdjldddmboOgT1S7qkibwAWi4UOHTqQnZ1NRUVFg8/nLa+//joA//rXvwgO9m/tb/NmyNatW+toGbgUykVEREQk4NUUysE+RX3MGHsot1iyA37Kusmd7dAgsEO5p4q8mTp06IDNZmvUhct+//134uPj6du3r7+7QlJSElarlby8PH93xW8UykVEREQk4K1Zs4bQ0NAaR0Pbt7eH8sY+5diT3Km8DoEdyj05Ug7HPlPzM25sdu3aRUFBAb179/bpnuS1CQ4Opk2bNgrlIiIiIiKBqrS0lM2bN9OtW7cap+rGxsYSExOjUH4CgVx93Rsj5dB4Q/mqVasA6N27t597ckxKSgp79uyhtLTU313xC4VyEREREQloGzdupLKy0lHk7XgWi4X27duTm5uLzWbzce/8QyPlx3hqOzRTUwnlf/jDH/zck2Patm0L2IvuNUcK5SIiIiIS0GpbT15VWloaFRUVzSYUbNu2zXEzwhUxMTEEBwcHXPV1T26HZjKL55nr9hubxjpSDjTbKewK5SIiIiIS0FwJ5c1pXblhGGzbto3U1FSXg6jFYqFVq1YBN1Luye3QTE1lpPxEfx98TaFcRERERCSAKZQ7Kyoq4sCBA25P1zZDuWEYXuqZ73m6yBtAixYtiI+Pb5Sh3GazsXr1ajp06EBMTIy/u+OgUC4iIiIiEsDMyuudOnWqtU1zCuXu7lFuio+Pp6KigoMHD3q+U37i6SJvpg4dOrBjx45GV7hsy5YtlJaWNqr15KBQrlAuIiIiIgGrtLSULVu21Fp53dScQrm7Rd5MgVjszRsj5XDss83KyvLoeRuqMa4nh2OF3nbs2OHnnviHQrmIiIiIBKwNGzZQWVlZ5/rZtLQ0oPGFKG9QKD/GmyPl0PjWlTfWUK6RchERERGRAOXKenKwj9RZrVaNlJ+AGcoDqQK7p7dDM5lLAxpbKP/999+BxrUdGkBsbCyRkZEK5SIiIiIigcbVUB4cHExKSopC+QkE2ki5N7ZDMzXmkfKwsDCPT9dvKIvFQkpKCnl5eQFVSNBVCuUiIiIiErDWrl0LuLb9U/v27dm3bx8HDhzwdrf8atu2bYSEhDjW8boq0EK5N7ZDM5mhvDHtVV5cXMyWLVvo2bPnCesr+Evbtm0pKysLqJkYrlIoFxEREZGAtWbNGsLCwujYsWOdbc1ibzk5Od7ult8YhsH27dtJS0sjKCjIrefGx8cDgRPKvVXkDRrn9PW1a9diGEajW09uMteVN8dibwrlIiIiIhKQXK28bjJDeSAXe9u1axelpaVub4cGgTdS7q0ibwDh4eEkJyc3qlDeWNeTm5pzsTeFchEREREJSOvXr3ep8rrJrMAeyOvK67ueHAKv0Js5Uu6NUA72z3jPnj2NZl/3xlp53aRQLiIiIiISYFwt8mZqDnuVeyKUB9pIubeKnjW2Ym8K5Y2XQrmIiIiIBCR3iryBQnldAnFNuTe2QzM1pnXlhmHw+++/k5CQQFJSkr+7UyOz8KDWlIuIiIiIBAiNlFfXkFAeFhZGVFRUQIRyb26HZmpMI+W7du1i9+7d9O7dG4vF4u/u1Egj5SIiIiIiAcadyusAsbGxtGjRIqALvTUklIN9CnsghHJvbodmakzbojX2qesAycnJgEK5iIiIiEhAOHz4MFu2bKF79+5ubf3Vvn178vLyqKio8GLv/CcrK4uIiAhat25dr+e3atWK/fv3N/nPx5vboZka00h5UwjloaGhJCYmKpSLiIiIiASC9evXYxiGy1PXTWlpadhsNnbu3OmlnvmPYRjs2LGDlJSUek9hNou9FRUVebJrPufN7dBM7dq1IygoqFGE8sa+HZqpbdu2FBQUUF5eXmubw4cPs3btWo4cOeLDnnmXQrmIiIiIBBx3i7yZAnld+f79+zl8+LCjoFZ9BEoFdm9vhwYQHBxMu3bt2LZtG4ZheO11XLFq1SosFovbfx98zVxXnp+fX2ub5cuX06tXL2666SZfdcvrFMpFREREJOC4W+TNFMih3JwWbAaf+giUUO7t7dBMHTp04NChQ379vGw2G2vXrqVjx45ERUX5rR+ucKXY2+rVqwH3/243ZgrlIiIiIhJwzFDes2dPt55nhvJALPZmbjXVkJHyQNkWzdvboZkaw7ZomzdvprS0tFGvJze5E8pPOukkn/TJFxTKRURERCTgrFmzhvDwcJcrr5s0Un5igTBS7ovt0EyNodhbU1lPDq6F8vrOgmnMFMpFREREJKCUlJSwdetWtyuvQ2CHck+MlAdCKPfFdmimxrAtWlOovG4yr83aQrlhGKxevZo2bdrUeweBxkihXEREREQCyoYNG+pVeR3soSAoKEihvBZmKN+7d69H+uQPvtgOzdQYRsqbUig3R8rNa/V4+fn57N27N6CmrgMEu/uEvLw8du3aRceOHYmNjeU///kPv/76K127duWaa64hONjtU4qIiIiIeExDprcGBweTkpISkKFc09ftfLEdmqmxhPLw8HCf3IRoqLqmrwdikTeoRyh/+umn+d///sf777/PkiVLeOaZZwBYuHAhR44c4cYbb/R4J0VEREREXNXQNaft27dn4cKF7N+/n9jYWE92za/M0cfk5OR6nyMQCr35Yjs0U1JSEmFhYX4L5YcOHWLLli3079/f7aUc/hAfH09YWFidoTzQRsrdnr6+ceNGWrZsSYcOHVi4cCHBwcFkZGRgsVj4/vvvvdFHERERERGXmXuUu1t53RSo68rz8vKIj48nIiKi3ucIpJFyX4wcW61W0tPT2b59O5WVlQ0+n80GCxbYv1+wwP7ziZg3qJrC1HUAi8VC27ZtycvLq3Fvd/P9NPtQvmfPHsei+i1bttCjRw/uuusuOnToQGFhocc7KCIiIiLijm3btjnCUH0EYii32Wzk5+c3aD05QFxcHFartUmHcl9th2ZKT0+nvLycnTt3Nug8mZmQng5jx9p/HjvW/nNmZu3PaUrryU0pKSmUlJRw4MCBao8F6vR1t0N5eHg4u3fvZvfu3eTk5Dgu5srKSq9vKSAiIiIiUpfc3Fzatm1b71pHgRjKCwoKsNlsDQ7lVquVli1bNtlQXlpaysaNG0lPT/dZdvHEuvLMTBg/HnJznY/n5dmP1xbMzVDeFLZDM9W2rryyspI1a9bQvn17WrRo4Y+ueY3bobxr167s3buXc889lyNHjtCnTx8qKyvZtWsXSUlJ3uijiIiIiLjAZoP58+G99+xf65raGoiKi4spKioiNTW13udIS0sDAiuUm+vJG1LkzdSqVasmW339+++/p7S0lJEjR/rsNRsaym02mDYNapjN7Th28801/31fvnw5EBihPDs7m0OHDgXc1HWoRyifOnUqMTExGIZB7969Ofvss/nll18oKSmhT58+3uijiIiIiNTBnNp6+ulw+eX2r3VNbQ1EuUeHEhsSygNxpNwT26GZWrVqRWlpKSUlJQ0+l6999tlnAJx33nk+e82G7lW+YEHVEfLFFBd3ZvHixY7HDQNyco6tNTcVFxezfPlyunfvTmJiYr1e2x9qC+UNLeDYmLk9p6d3797MmzePAwcOOKpRDhw4kKVLlzaJin4iIiIigcac2nr8SJo5tXXOHMjI8E/ffM0M5e3atav3OcxQnpWV5ZE+NQae2A7NVLUCe2RkZIPP5yuGYfD5558THh7epEbKjy1FLwMmYxjbeP/994HLamlnt2TJEioqKhg+fHi9XtdfzBtHx4fyQK28Dm6OlFdUVDBo0CBGjx5dbR6/ArmIiIiI7zVkamsgysnJARo2Ut6iRQtiY2M1Ul6LplqBfeXKleTm5nLmmWf69GZCQ0P5sR3sngTWA7BgwQIMw1ZLO7v58+cDMGLEiHq9rr+YN47Ma9akUH5UcHAwCQkJxMXFYbFYvNUnEREREXGR89RWgDeBvwP2f7DXNrU1UHlipBzso+V5eXlUVFR4olt+Z446NudQbk5dH2uWL/eR+Ph4YmJi6h3Khw2DpKQtwENALEFBZ7Bv3z4qK+3rxS0WaNfO3q6qH3/8EaDJjZTXNn199erVWCwWevTo4Y9ueZXba8ovvfRSsrKyWLp0qTf6IyIiIiJuqL7L0l3Ao8AUwDhBu8DkiTXlYC/2VllZWW20rqnydKE3oMkVe/v8888B34dyi8VCeno6OTk5HDlyxO3nW60GbdpMBUqBfxAcfAUAFRVfYY6TzpwJVScul5SUsGzZMrp27Ury8UPojVxN09dtNhvr1q2jc+fORERE+KtrXuP2mvJFixYRFBTETTfdRFpaGvHx8Y5Rc4vFwksvveTxToqIiIhIzZz/vV0ImOn730Ar4LEa2gUuT0xfB+dib+b3TdmOHTuwWq0eKfjVmEbKbTb7LJCdO+3X+LBhzuHUlJ+fz08//cTJJ5/skRsT7urQoQOrVq0iJyeHjh07uvXcDz/8kJUrv6Fr11MoLr6OoqJ8AGy2r0lNfZCZM6vXjFiyZAlHjhxpcqPkYN+COz4+3imUb9myhbKysoAs8gb1GCn/9ddfOXLkCIZhsH37dn799Vd++eUXxx8RERER8Z1hwyA1laMjZiuPHr0ASAUeBx6rcWproMrNzcVqtTZ4dDDQir3l5eWRlJRU773bq6pa6M2f3Nlx4IsvvgB8P0puqu+68v379zNt2jSCgoL44IOXycqy8uWXSfTq1YvKyp9ZtqygxiKO5tT1prae3JSSksKuXbscy0cCeT051GOkfMyYMVpPLiIiItJIBAXBM8/Yq6wfC+VjsU9hHwbcybnntiQo6Dp/ddGncnNzSU5ObnD4DKRt0crKytizZw8DBgzwyPkaw0i5uzsOmFPXfbkVWlX13RbtnnvuIT8/n1tvvdWx/fSwYfZ14mvWrGHevLn86U9/qvY8s8hbUxwpB3soX7VqFbt27SIlJUWh/HgzZszwQjdEREREpL4yMuwh5E9/Wol96+g+QHfatPmaAwdO51//up4zzohjwoQJfu6pd5WUlLB3714GDx7c4HMFUij3ZOV18H8or3nHgUrAimHYZ43cfDOMG2e/aVVaWso333xDcnIyJ598sl/6XJ+R8uXLl/PCCy+QmprK/fff7/TY8OHDefHFF/nqq6+qhfLDhw+zbNkyOnfu7Jep+p5Qtdhbcwjlbk9fN/3888/MmjWLOXPmcOjQIfLz8ykvL/dk30RERETERRkZ0KnTCqxWK2+80YsffoC8vP58+eWnhIaGcuWVVzJ37lx/d9OrPFV5HQIzlHsqoPk7lFffceBaoBewG6i+48APP/xASUkJY8aMwWqtd/xpkM6dOwOwceNGl9pXVFQwZcoUDMPgueeeIzo62unxfv36ERcXx9y5c7Edt9/h0qVLKS8vb7Kj5FC9AvuaNWsICQmhS5cu/uyW17h9VZaWljJ16lSmTp3KSy+9xJdffslPP/3E+eefz7vvvuuNPoqIiIhIHcrKyli3bh3dunXjqqsiGDHCPko4YsQI3n//fSorK7nssssCZouvmniqyBvYR5WDgoICKpR7eqTcX9XXq+8k8C32/bsnUtOOA+ZWaP6aug72UB4cHMyaNWtcav/xxx/z22+/cd555zFu3LhqjwcHB3PWWWexd+9eli9f7vRYU19PDs4V2MvKyti4cSNdu3YlNDTUzz3zDrdD+UsvvcTy5csxDAPj6JyRoUOHEhISwqJFizzeQRERERGp27p166ioqHCsO61q3LhxnHvuuRQVFQVM4bKaeHKkPCgoiNTUVLZv3+74N29TZY42emqkPDIyksjISPLz8z1yPnc51/CrBMxt674AnnZqZxgGn3/+OeHh4Zx55pm+6+RxQkND6dq1Kxs3bnRpW7Sff/4ZgGuuuabWel5nn302AF9++aXT8aa+nhyOXas7duxg48aNVFRUBOzUdahHKJ83bx5hYWFOo+KhoaEkJSUFxJ1EERERkaZo5Up7kbe+ffvW+Hi3bt0A16fPNkWe2qPclJ6ezqFDhygqKvLI+fzF0yPlAF26dCErK4vS0lKPndNVzjsO7AGOAD2BOOAOYJljx4Hff/+dnJwczjjjDCIjI33e16p69uxJRUUFmzdvrrOtOaJ+oi3AzFD+1VdfOY6VlpaydOlSOnbs6JGbU/5Sdfp6oK8nh3qE8r1799K+fftq8/mDg4M5ePCgxzomIiIiIq4zQ3lNI+UAXbt2BQI7lHty+jrYQzm4XzG7sTFHyj0Zynv06EFlZaVfridzxwE7cy/rAcC/gQrgEh56YDdBC+bz2cMPA3DemDE+7mV1PXv2BGDt2rV1tl27di3h4eGOa7AmSUlJ9OvXj59//pmCggIAli1bRllZWZMeJQeF8jolJCSQnZ3tuBMJsGHDBrZv307r1q092jkRERERcY2roXzTpk0+65OveXL6OhwL5e7uLd3YeLrQG0D37t0B+7IJfzB3HGjVygzlbYEMoqNvALL4ZGp7jNNP5/MPPwRg7IMP1ryBuQ+ZobyudeUlJSVs27aNHj16EBQUdMK255xzDoCjiGMgrCcHe+YMCQkhLy/P8XkplFcxfPhwysrKuOSSS7BYLGzYsIGrrroKwzA47bTTvNFHERERETkBwzBYsWIFrVu3JikpqcY2zWGkPDc3F6vVWutn4K767i3d2OTl5REWFkbLli09ds4ePXoA/gvlYA/mDz1kD+VXXZXCDz9A/stD6AdkHj7MfcBPQF8gddcu+wbmfgzm5lT0ukbK161bh2EYJ5y6bjJDuTmFPRDWkwNYrVaSk5MdI+URERGOv4+ByO1Qfv3119OlSxfKy8sxDIPy8nJsNhudO3dmypQp3uijiIiIiJxAXl4ee/fupU+fPrUWhUpKSiI6OjqgQ3lOTg5JSUmEhIR45HyBMH3dMAx27NhBSkpKrddGfZihfP369R47Z33s3GkP5ePGpTBimI2oO//GB0AM8CD2WuznwbFNzW++2b7RuR906dKFoKCgOkO5K+vJTaeccopja7SSkhKWLFlCeno6aWlpHumzP6WkpHDw4EG2bNni0qyBpizY3SdER0fz1ltv8fXXXztdMKNHj/bYfwBFRERExHV1FXkDsFgsdOnShRUrVnD48GEiIiJ81DvfKCkpYe/evQwePNhj5wyEUH7w4EGKi4s9up4c7DMvrFarX0fK4bjK8kc3MO8M/Au47Ggbx0ZoVTcw98P07rCwMDp37syGDRuoqKggOLjmKGaGdldCeXBwMKNGjeKDDz7ghVtvpbS0lOHdutlvPDTxEFt1uUUgT12HeoyUg/2XP3bsWO644w7uuOMOxo4dq0AuIiIi4id1rSc3de3aFcMw2LJliy+65VOerrwO9lAQFBTUpNeUe6PIG9gDZseOHdmwYQM2P408w3Hr5atsYH4pcC9wMdD/+CdV3+jcZ3r27El5efkJ/w6aA5/mGvS6nHO0rtejr7wCwIi5cyE93e9r6Bsqpco1e1JEhN9mOPiC2yPl999/f62PhYeH07VrV84991zCwsIa1DERERERcY07oRzsxd4CbeTJ00XewD4Q1b59e8de5Z6c/u0r3ijyZurRowebN29m+/btdOrUyePnd0VeXh5Wq5U2bdocv4E5D9T2pOPa+VLPnj356KOPWLt2rWObwuOtWbPG9TXUmZmc/cILAJgb9w0HyMuzr6GfM8e++L6pycyk7ZtvOn486ZVX4Isv7GX3m+L7qYPbofzzzz+v8z9I7733Hq+99hotWrSod8dERERExDUrVqwgNDTUURG7NoFc7M0bI+Vgn8K+bds29uzZQ0JCgkfP7QveGikHeyj/7LPPWLdunV9DeZs2bexTwc0NzPPyjq0hr8pisT8+bJjvO3pU1WJvF154YbXHi4uL2bZtGyeffDJWax2Tmm02mDaNJKAf8BvQHkgH+/u3WOxr6MeNa1pT2TMzYfx4Uqr8Dk+Cpn+j4QTcnr5uFs8wDIOYmBhiYmIwDIOQkBBiY2MxDIPt27fz2muveaO/IiIiIlJFcXExmzZtomfPnnUuJ+zSpQsQmKHc03uUm5r6tmjeHikH/1VgLy0tZc+ePcfeW9UNzI8fRDR/njnTrwG1rm3RzM/SlfXkLF4MR29GnXP00HDA8c6rrqFvKo7eaMAwMK/YFkAqNIpifd7idii/9dZbsVqtvPjii3z33Xd89913vPDCC1gsFu68805effVVgoKCWNCUfvkiIiIiTdTq1asxDOOERd5MgRzKvTF9HZr+tmhmKPfWSDn4L5TXeMPB3MD8+JsQqamNYoTVLJBXWwV2d4q8sWuX49urgT8A19TUzo9r6N12tFgf4AjlPWniNxpc4Pb09eeee47k5GQGDhzoODZo0CDatm3LCy+8QGZmJv369WPFihWe7KeIiIiI1MDV9eQA8fHxJCQkBHQo99ZIeVMN5U7VyT3MXC7RqEI52IP3uHH24LZzp30N+bBhjWIKd0REBB07dmT9+vXYbLZq23y5sx0abdo4vu0MrKytnR/X0Lutyg2EzsCdHF0jf4J2gcDtUL5r1y4qKiqYM2cOZ555JmDfpD4rK8tR1j80NFTV2EVERER8wJ1QDvaRusWLF7N//35iY2O92TWfysnJwWq1kuzhANLUQ7kZXD39uQDExsaSnJzM+vXr/VII74Q3HIKC/LLtmSt69erF5s2b2bZtG507d3Z6zK3K66ee2ujX0LutynVqAR51oV0gcHv6er9+/aisrOTxxx9n1KhRjBo1ikceeYTKykr69+9PZWUlGzdu9MrdOBERERFxVp9QDvYK7IEkNzfXUfvIk5r6mvK8vDxiY2OJioryyvl79OjBvn372FVlKrWveHMWgDedaF35mjVriIyMdFx3J9QE1tC7zSzWV9sNHosF2rVrWjcaXOB2KL/nnnvo3r07hmE4/enRowd33303O3fuZOTIkVx11VXe6K+IiIiIHFVZWcnKlStJTU0lPj7epecE4rryw4cPs2fPHo9PXQf7WuyQkJAmOVJeWVnJzp07vRpa/bmu3JuV5b3JDOXHrysvLi5m+/bt9OjRo+7K66ZGvobebYF4o8EFbk9fT0pK4u2332b58uWOTe87d+7MgAEDHG3++te/eq6HIiIiIlKjbdu2cejQIUa4MU03ELdF81aRN4CgoKAmu1f57t27qaio8GporRrKTz/9dK+9Tk2a+kj58aHcrcrrVTXiNfT1Yt5omDbNUfQNsN9omDmz6d1ocIHbodw0cOBAp2JvIiIiIuJb7k5dh8AO5d4YKQf7FPYtW7ZQUFBAmyrFtRo7X4TWxjBS3tRCeffu3bFYLNWmr7tV5O14jXgNfb0E2o2GOrgdyisrK/n0009Zvnw5e/fuxahSVMBisfDSSy95tIMiIiIiUrP6hHKzsFQgrSn31h7lpqrbojWlUO7N7dBM/gzlO3bsICoqihYtWvj8tRsiMjKSDh06sG7dOqcK7G4VeWsOAu1Gwwm4Hcqffvpp3n//fQCnQA40qek8IiIiIk1dfUJ5ZGQk7dq1Y+PGjU1uOnZtvDl9HZwrsA8ePNgrr+ENvlhznZSURGxsrM9DuWEY5OXlkZaW1iSv4Z49e7J161aysrLo2LEj0MCRcmnS3C70NnfuXAzDICEhgT59+nDyySc7/vTr188bfRQRERGRGqxYsYLIyEg6derk1vO6dOnCgQMHKCgo8FLPfMsX09eh6W2LVus+3h5ksVjo0aMHO3bsYP/+/V57nePt3buXsrKyJjd13VTTuvK1a9cSGRlJWlqav7olfuL2SLnNZiMxMZGPPvqI0NBQb/RJREREROqwb98+srKyGDx4sGP6q6u6du3K999/z8aNG5vUdOzamNPXvT1S3tS2RfNVdfIePXqwdOlSNmzYwKBBg7z6WqamWnndZI6Gr1mzhrFjx3Lo0CG2b9/OgAEDXK+8LgHD7d/4qFGjKCsro6KiosEvXlRUxLRp0xg6dCgZGRn89NNPtbb97LPPuPDCCxk2bBjjx4933BEVERERaY5+//13APr27ev2cwOt2Ftubi4Wi4Xk5GSvnL/qmvKmxBcj5eCfdeVNtcib6fiR8npXXpeA4PZIeWRkJMXFxVx++eWcdtppxMTEOD1+7bXXunyuxx57jFatWjFv3jyWLVvGXXfdRWZmJrGxsU7tFi5cyLvvvstTTz1Fhw4dyM3NbXIFHUREREQ8qT7ryU1mKA+UYm+5ubkkJSUREhLilfMnJSURGhraJEO5xWLx+mwIhXL3de/eHTgWylXkrXlzO5S/9dZbWCwW8vLymD17drXHXQ3lJSUlzJ8/n08++YTw8HCGDx9Op06d+PHHHzn//POd2r766qvccsstjiIIdU1NKi8vp7y83OlYcHCwpts3YZWVlU5fRVyh60bqQ9eN1Ic/rhszlPfu3dvt1zUrsG/YsKHJX+uHDx9m9+7dDBo0yKvvJS0tje3bt1NRUeGx6cXevm7y8vJITEwkKCjIq5+NeZNn7dq1PruezFCenJzcJK9hc+34unXrqKiocITyHj161Pl+9P+ppsWV/164HcqTkpI8UuEwOzubyMhIpzt3nTt3ZuvWrU7tbDYb69evZ8uWLdx///0EBwdz3nnnMXny5Fr7MWvWLF599VWnYxdffDETJkxocL/Fv8w1YyLu0HUj9aHrRurDl9fN6tWrAYiKiiIrK8ut51osFoKCgli7dq3bz21szNHrli1bevW9JCUlsWnTJn755RcSExM9em5vXDfl5eUUFhZy0kknef13bLVaCQ0NZdWqVT67njZs2ABAUFBQk72GO3ToQFZWFkuWLOHnn38GIC4uzuX3o/9PNQ3m8pcTcTuUf/bZZ/XqzPEOHz5MVFSU07GoqKhqVRv37t2LzWZj6dKlzJ49m4MHD3LDDTeQnJzMmDFjajz31VdfzRVXXOF0TCPlTVtlZSU5OTm0a9dOxS/EZbpupD503Uh9+OO62bFjB7GxsfzhD3+o14BJx44dycrKIjU11e1CcY2JOaDTtWtXr1at7t69OwsWLKCiosJjr+PN68YMdmlpaT6p5t2tWzfWrFlDUlISYWFhXn89MzP079/fawX+vK1///7Mnz+f/fv3s23bNqKiovjjH/9Y57Wg/08FHrdDuenw4cNs2rQJq9XKSSed5PbzIyIiKC4udjpWXFxMZGSk0zHzL/X//d//ERMTQ0xMDBkZGSxatKjWUB4aGqoAHqCsVqv+4yNu03Uj9aHrRurDV9dNWVkZubm59O3bt96BumvXrmzatIm8vDxHdfGmyCxm1r59e69+9uZoV3Z2NkOGDPHoub1x3eTn5wP2beJ8cU326NGDVatWsWXLlnplA3eZ6+Xbtm3bZP9bbRZ1+/nnn8nKymLgwIEEB7sez/T/qcBRr9/ia6+9xqhRo7jmmmv45z//ybfffsu4ceP4+uuvXT5H+/btKSkpcdofc8uWLY5146YWLVrQunVrj0yZFxERkabPZoP58+G99+xfbTZ/98j3srKyMAyj2r+b3BEoxd68vUe5qalti2berPDVlmG+LvZmrpf3VnE/XzCLus2ZM8fpZ2l+3A7lc+bM4ZVXXqG0tBTDMAAYOHAg+fn5zJ071+XzREZGMnz4cMe5FixYwObNmxk+fHi1tmPHjuWtt96iuLiYXbt28dFHHzF06FB3uy4iIiJNXGYmpKfD6afD5Zfbv6an2483J+aUbU+E8qa+LZq39yg3NbVt0Xy9j7cZytevX+/11yorK6OwsLDJVl43mSHcLPKm7dCaL7dD+fvvv4/VauWvf/2r41hcXByJiYlu32m98847KSwsZOTIkTz99NM88sgjxMbG8tVXXzkVZbvuuutISEjg3HPP5eqrr+bss8/m3HPPdbfrIiIi0oRlZsL48XB0YNQhL89+vDkFc0+E8i5dugBNP5T7eqTcH6G8PrNDfLVHucmXI+Xm1PymHspjYmKcbib1stma59QfcX9NeW5uLh07duTSSy/lqaeechxv0aKF29N5WrZsybPPPlvt+DnnnMM555zj+DkkJIR77rmHe+65x93uioiISACw2WDaNDg6Sc+JYYDFAjffDOPGQROuWeYyjZQfk5ub61hb7E1t2rQhPDzc59PXMzPt137Vm1GpqfDMM5CRUfvzfD1S3rVrV6xWq0dCuc0GCxbAzp2QnAzDhjn/vW7qe5Q7ZGbSs6AAs4Z6r7vughdeqPuXKwHH7ZHy6OhoCgsLKSsrcxw7ePAg2dnZREdHe7RzIiIiImD/B7rzCPl44GLHT4YBOTn2ds2BJ0J5SkoKERERTT6U5+TkkJSU5PW1xRaLhfT0dLKysny2P3RDZof4eqQ8PDycDh06sGHDhgZ9Pq4sUQmIUH70l9vzaKaKBtpD85z6I+6H8pNPPpkDBw4wceJEwH538qqrrqKsrIz+/ft7un8iIiIi7NxZ9adSIBOYAyw9QbvAtXXrVqxWK+3bt6/3OaxWK126dGH79u2Ul5d7sHe+U1payu7du70+dd2Unp7OkSNH2OmDC6367JBdwHLg2LGbb655trPNZmPjxo2EhobSqlUrr/fV1KNHDw4fPlzvfcNdvQnR5EN5lV+uuYq8B2CBun+5EpDcDuV//vOfiYyMZPPmzVgsFvbt20dOTg5RUVFMmTLFG30UERGRZi45uepPWwAzqcw8QbvAZBgGW7dupV27dg3eArZr165UVlY6Rt79qT7rpn21ntzky3Xl1WeHTAUGAz8DJ54dkpmZSW5uLmPHjvXpDkbdu3cH6reuvK4lKnAsp/p6ar7HVfnl9jt66OSqjze3qT/ifihPT0/n7bffZuzYsaSnp5Oens7YsWN54403mvQelyIiItJ4DRtmX0drzxdVp1vPAXKwWKBdO3u7QLdnzx4OHjzYoKnrpsZS7K2+VfXNUO7tyusmX26LVn0w/jfsN6P+fsJ2hmHw8MMPA/D3vzu39baGFHtzvgmxH/gT8JPj8ao51ddT8z2uyi/tZOBL4ME62klgc7vQG9j/w3ffffd5ui8iIiIiNQoKstc+Gj8ewNztpT/wC/Ai8CgzZ6rIm7saQ7E3c8ry8SOk5pTlOXNqr3nl65FyX26L5jzroxQwX/Nb4DtgZA3t4KuvvmLlypWcffbZPl9a2pBQ7pw/3zn6ZzuwoFq7Jj99/bhf2jm1NGsWU38EqMdI+dq1a/n888/Jz8+noqKCxx9/nMsuu4zp06dz6NAhb/RRREREhIwMe0CLjDQD5CNANFbrK/znPyXNplhxIIVyd6Ys18RXe5SbfDl93Xl2yGbso+Qdjj56F2BUmx1SdZT87rvv9nofj2dOX9+wYYPbz3XOn/89+nUhsKpau7y8PCIiIoiLi6tPN/3P+ZdbXXOa+iNAPUL5yy+/zAMPPIBhGHz88cd8+OGHbN68ma+//prnn3/eG30UERERAezB/OST7QHyX/8aQEbGJCorizhw4G0/98x3vBHKN23aVEdL76i+bno78DpQAdS9tDaQ15Sbs0PszJD7J+As7AXfPqo2O+THH39k8eLFDBs2jKFDh3q9j8dr2bIlrVu3rlcoN3MqFAI/AuYbewk4llOHDjXIy8sjJSXFp+vlParqL/f492D+3Fym/ghQj1C+ZcsWEhISSE5O5qeffiIsLIy//OUvBAUFsWjRIm/0UURERMRh8+ZNtGrVimuvjeexx27EYrEwc+ZMn21T5W+eDOWtWrWiZcuWfhspr75k9n5gMnA9x4r51b601vwsfBXKW7duTWRkpM/2Kjdnh7RoYYbcbthniEBKyt2cf36FU3t/jpKbunXrRmFhIUVFRW4971hO/RSoBG4B4oG3gYOAPacePLiPw4cPN92p6ybzl3v8+0hNPfGaDQlIbofyoqIix9YK27Zto2fPnkycOJGOHTuyd+9ej3dQRERExHTgwAHy8/MdBco6d+7M2LFjWb9+Pd9++62fe+cbngzlFouFLl26sGPHDr8sQ6y+ZHbz0a//Bm7HDOY1La19/PHH+frrr0lKSvJZQDP3Ks/Ozsbmo+2qMjJg3Dh7KH/ooW788MMAxo+/mLy89bz11luOdj/99BPz5s2jf//+jBo1yid9q0m3bt2A+i2JyMiAfv3MqetXAJOAQ7Rs+Y4jpzb5yutVZWTA9u3www/w7rv2r9u2KZA3Q26H8ujoaHbs2MHvv/9Obm6u438IpaWlREREeLyDIiIiIqbNm+2hzZx2DXDzzTcDMHPmTD/0yPe2bt1KTEyMx/af9ucU9upLa7OAKKAd8CTwGO1SDYbZ5jvtlfbggw9yxx130KpVK7766itCQkJ81uf09HQqKiocFcB9YeNGeyi/6aaujBgBDz30IEFBQcyYMYPSb76B997jkVtvBewV1/05rdsM5fWZwr5v3z5Wr55Hx44d+f77Pvzzn/btllNSXuTCC+03aJp85fXjBQXBiBFw2WX2r5qy3iy5Hcp79erFgQMHuOaaa7DZbPTv358jR46wa9euwPnLISIiIo2SOfpmjpQDnH766fTu3Zuvv/66XlWfm5Ly8nJycnLo2LGjx4KXP4u9Oa+bPgLkAR2xVxhPAO7i3KJhBJ1p3yvNOP107m3ZkunTp5OYmMj8+fPp27evT/vsy23RwF68bcOGDbRt25aYmBjAHnwnnX46OTk5vDR6NKsuv5xPFi2iZ3AwF/h5GYd5PdUnlH/xxRccOXKEjIwMTj/dwi23dGb06NGsXr2ahQsXAgFQeV2kBm6H8ptvvpmuXbsSERHB6NGjGTlyJCtXriQ2NpZTTz3VG30UERERAY6N5lYdKbdYLI7R8meffdYf3fKZ7OxsKisrPTJ13eTvYm/m0tqkpDzsa4nTgW60if2YCIL5V/Ei3sc+kf1O4KGDB0kGfrz7bk466SSf99eX26IBFBYWsm/fPscINACZmUyfN49w4GHgnqOH76qowDphQt0bvHtRQ0bK//tf+9T1iy66yHFs6tSpALz0kr3gm0K5BCK3Q3laWhr/+c9/+PHHH3nwwQexWq0MGDCAL774gilTpnijjyIiIiJAzSPlAJdffjkJCQm8+eabAV3jxlxPbgZDT/D3tmhgD+bvvLMdgFGj0vhhno286Mv5igpCsdccvwh4HEjFXpu7+5NP1r5Xmhf5sgI7HAu3jlB+dB+5VOBGYA/20mgdgEvNJ51oHzkv69ixI0FBQW6H8uLiYr7++mvatm3LoEGDHMfHjBlD+/btmTNnDrt27VIol4Dkdig/dOgQ+fn5lJaWAvDdd9/x5JNP8sknn3i8cyIiIiJVmaO5x4fy8PBw/vznP3P48GFee+01f3TNJzxZ5M3UuXNnwL+hHCA3NwuAs85KY0TQAoLyshkOfIh9/Pwj7GPo/wO6wIn3SvMiv4fyKvvI3QnEHm13BxAMde8j52WhoaF07NiRTZs2uVUM76uvvuLw4cNkZGRgtR6LKEFBQVx33XUcOXKEf//73wrlEpDcDuUPP/ww559/Ptu2beN///sfd955Jx988AEPP/wwb775pjf6KCIiIuJYW5ucnEx0dHS1x//85z8TEhLCK6+84ofe+Ya5jtmToTwmJobk5GQ2bNiAYRh1P8FLsrLsoTwtLc1pD7TzgA+A8dhHyJ3mCNS2V5oXmbMUzBsk3lYtlFd5z/HYd/H+EzDx+Cf64bMxdevWjbKyMnJyclx+TubRKfdVp66bJk+e7Pi7nb12LQDJGzf6bTaAiKe5HcrXr19PTEwMPXr04Pvvv8disXDKKadgGAZffPGFN/ooIiIiwp49e9i3b5/TevKqkpOT+eMf/8jWrVvJz8/3ce98wxsj5WCfwr5v3z727Nnj0fO6wymUH7cHWgb2EfP2xz+ppr3SvCw+Pp6EhARWr17doJsYNtuxwewFC2rPl9VC+XHv+TLgLSDs+Cf64bMxuVvsraysjM8//5yEhASGDh1a7fGkpCQuGjSI7OxsVm7ZQmsgdNQoSE/36/p5EU9xO5QXFhbSpk0bwL4tSbdu3Xj22WdJS0sL2P8BioiIiP/VNnW9KnMt6k8//eSTPvna1q1bsVgs9uDqQf4u9gbHQnl6enpNe6U5s1igXTt7Ox+zWCz84Q9/YM+ePfX+t29mpj1Pjh1r/3ns2Nrz5YYNGwgLCzv2O2/En43J3WJv3377LQcPHuSCCy4gODi4eoPMTP68aJHjR8fE9bw8GD9ewVyaPLdDeUhICIcOHXJsyWFO4QkJCfHrnogiIiIS2Mw1z7WNlAMMHjwYCOxQnpKSQnh4uEfP2xiKvW3fvp2IiAhat27tvFfa8f++NH+eOdNvezr37t0bgFWrVrn93MxMe448uizcoaZ8eeTIEbZu3UqXLl0IMt9rI/9swP1QfqKp62Zhu2FAr6OHHKHcnKngx8J2Ip7gdijv0KED+fn5jBo1isOHDzu2oigoKHCMoIuIiIh4Wk3boR3PHClftmyZT/rkS0VFRezbt8/jU9fB/6G8srKSnJwc2rdvf2yQx9wr7fiCXqmp9uMZGb7v6FH1DeVH8yU1zXqvKV9u3bqViooK5+3QoFF/NuBeKD9y5AiffPIJsbGxnHHGGdUbHC1sZwGmHj2UWvVxPxe2E/EEt0P5pEmTCA4Opri4mJSUFM4991xWrVrFgQMH/LJXpIiIiDQPtW2HVlW7du1o06YNy5cvp7Ky0ldd8wlvrSeHY5+pv0J5fn4+5eXl1aflZ2TA9u3www/w7rv2r9u2+T10/uEPfwDg999/d+t5VQqnA0WUl7/omLYP1fNltfXkVTXSzwagTZs2xMTEuHQ9/fjjj+zdu5fzzjuP0NDQ6g2qFKybBNwN3FTTifxY2E6koWpYtHFiQ4cO5csvvyQ/P5+OHTs6tj346KOPiI2NrfsEIiIiIvWwadMmLBYLnTp1qrWNxWJh8ODBfPrpp2zatKnmMNNEeTOUd+zYEavV6rdQ7lTk7XhBQTBihG87VIdevXphsVjcHim358ZS4DngEcrL93HPPUOwb/R2fLs6Qjk0ys8G7H8Pu3Xrxs8//0xxcTFRUVG1tv3vf/8L1DJ1HZwK1oUDD9V2Ij8WthNpKLdHygHi4uLo3r27425WVFQUqampxMTEeLRzIiIiImDfDm3Tpk2kpaURFlatzrSTQJ3C7s1QHhYWRnp6Ops3b/bLDAOnIm9NQGRkJJ06dWLt2rVUVFS49BybzcbKlW8AXYHbgTIgniVLlmAYhU5tzXxZZyhvxMw+n6h4oM1m46OPPiIyMpJRo0bV3KgJFLYTaah6hfI5c+Zw3XXXceGFFzJu3DinPyIiIiKetnPnToqLi0+4ntwUqBXYvRnKwb6uvKSkhB07dnjl/Ceyfft2oJaR8kaqd+/elJWVsXnz5jrbfv311/Tp04fHHrsa2AFcB2wmJOQGKisrqaiwjxYfny8DIZSfaF35smXL2LVrF+eeey6RkZE1N2oChe1EGsrtUP7ee+/x+OOPs2LFCnJzc9mxY4fjz06t5RAREREvcGU9uWngwIGAQrm7/Fns7YTT1xspV9eVr1u3jjFjxrBmzRoyMjJ45pnVWCyvYLG0JSTkYgAqKubUmC/Xr19PYmIicXFxXnoX3uNKKJ87dy4AY8294WrTyAvbiTSU26H8448/BqBv374AtGzZki5dutCiRQvGjBnjyb6JiIiIAK5th2aKi4ujW7durFixgtLSUm93zWe2bt1KZGQkiYmJXjm/P4u9NcVQ7moF9nnz5lFZWckjjzzCf//7X266qbsjX1qtPenatSs2248kJ+c75cu9e/eye/fuJjlKDsf+rp4olM+bNw+AM888s+4TNuLCdiIN5XYoz83NpWXLlrzyyisApKam8sYbb2C1Wr1251ZERESaN3Ndqisj5WDfr/zIkSOsXLnSm93ymYqKCrKysujYseOxLcM8zN8j5cHBwbRt29bnr11frobyRYsWAXDWWWc5jpn58vPPOTqoVcmdd/7XKV825anrUPdNngMHDrBs2TJ69OhByvEj4LUxC9tddpn9q6asS4Co15ryxMRELBYLQUFBHDhwgNDQUFq0aMHs2bM93T8RERERt0bKIfCKveXk5GCz2bw6AGJ+ticqzOUNhmGwfft22rVrR1ATClmdOnUiIiLihNPXDcNg4cKFREVFOWaZmoKC7GvHzZmmH374vtPjTT2UR0VF0a5dOzZs2IBRw8bsP/74IzabzbVRcpEA53Yoj4uLY9++fQC0bt2a7OxsbrrpJrKzszl06JCn+yciIiLCpk2bCAkJcXl6c6AVe/P2enKw7/EeFhbm85HyPXv2UFJS0qSmrgMEBQXRq1cvtm3bxsGDB2tsk52dTV5eHoMHDyY4uOadiDt27EifPn1YuHAheXl5juNNPZSDve8HDx4kPz+/2mNuTV0XCXBuh/JOnTqxa9cu8vPzGTJkCIZhsHTpUuDY/wBFREREPMVms7FlyxY6duxYa7A5Xp8+fQgNDVUod0NQUBCdO3dm69atHDlyxGuvc7ymuJ7cZE5hX7NmTY2Pm1PXhwwZcsLzXHzxxRiG4dizGwInlEPN68rnzZtHUFAQw4cP93W3RBodt0P5PffcwzvvvEN0dDQ333wz48aNo2PHjpx99tncdddd3uijiIiINGPZ2dmUl5e7PHUdIDQ0lH79+rFp0yb27t3rxd75hi9COdjXAVdUVDi2KPOFQAjlta0rN0P50KFDT3ieiy+2V2F///1jU9g3bNhAcHAwHTp08ERX/aK2Ym87duxg7dq1DB48mNjYWH90TaRRce12cxX79+/HYrEQHR0N2EO6iIiIiLe4sx1aVYMGDWLZsmUsX76c0aNHe6NrPuOrUF612Ju7n3d9maE8PT3dJ6/nSXVti7Zw4UKsViunnHLKCc/TuXNnTj75ZBYvXkxOTg5t27Zl8+bNdOrUiZCQEI/321dqGynX1HURZy6PlO/bt48rr7ySK664giuuuIIrr7yS/fv3e7NvIiIiIm4XeTMNHjwYCIx15WYo93Zw9UexN3NUPtBGyvfv38+qVavo3bs3LVq0qPNcl1xyCQAffvgh27dvp7y8vElPXYdjofz4OgUK5SLOXA7lb775pqN6omEYbNy4kTfffNObfRMRERFxezs0UyBVYN+6dStt27YlIiLCq6/jj23RmvL09cTERBITE1m1alW1CuNLly7FMIw6p66bzCnsH3zwQUCsJwdo37494eHhTiPlhmEwb948oqKiHDfORJo7l0P5jz/+iMViYeTIkYwcORLDMJg/f74XuyYiIiJS/5Hyzp0707JlS3766acat2RqKvbt28fevXu9PnUd/BfKLRYL7dq189lrelLv3r3Zu3cvO3bscDq+cOFCoO4ib6YOHTo4llzM/fe/AehmGGCzebbDPmS1WunSpQvbtm2jvLwcgHXr1rFz506GDx9OaGion3so0ji4HMp37dpFcnIy//jHP/jHP/5B27ZtKSgo8GbfRERERNi0aRMRERG0bdvWredZLBYGDRpEYWGhYzTW32w2mD8f3nvP/tWVvLVt2zYAnxT8SkxMJCYmxuehvG3btk02oJnryo+fwu5qkbeqJvToAcCrmZkAdHvySUhPh6M/N0Vdu3Z17KAAx6aun3XWWf7slkij4nIoLy8vp1WrVo6f4+PjHXe8RERERLyhvLyc7du306VLF6xWtzeNaVRT2DMz7fnq9NPh8svtX13JW74q8gb2Gxldu3YlJyeHkpISr7/egQMHKCoqapJT1001rSs/cuQIy5Yto127dq7PAMjM5OKjS0MPHz3UDSAvD8aPb7LB/Phib1pPLlKdW9XXd+3axauvvgrgGCU3fzZde+21HuqaiIiINHdbt26lsrLS7anrJjOU//TTT45CWv6QmWnPVcfPojfz1pw5kJFR83N9GcrBPrL5yy+/sGXLFkfgrC+bDRYsgJ07ITkZhg2DoKBjjzfl9eSmmkL5ihUrKCkp4fzzz3ftJDYbTJtGe+CPwBKgJZAA9ovGYoGbb4Zx45w/wCagarG3I0eOMH/+fNq0aUOvXr383DORxsOtUF5YWFgthCuUi4iIiLfUdzs0U9VQ7i9H81a1QA5V8tY0g3GxPxJU4JxeDcPg119/BXwbysH+2TcklGdm2t93bu6xY6mp8Mwzx25ABEIo79mzJ1ar1WlbNLenri9e7PigJmAP5d0Ai/m4YUBOjv0Ox4gRHuq5b1QdKf/pp584ePAg559/PhaLpY5nijQfbs0DMyuv1/ZHRERExJPMyuv1HSlPTEwkPT2dX375hSNHjniyay5bsMA5mMKPwKnAK8ARe97KtbDgzBlO89oLZs3iggsuYPbs2cTFxXHSSSf5pL+eKPZmzgxwft/VZ2IHQiiPjIykc+fOrFu3znGNmaHc1SJv7Nrl+PYSIB6ocXL3zp0N6qs/mNfThg0bNHVdpBYuj5R/+umn3uyHiIiISDUNHSkH+2j5Bx98wJo1a+jbt6+Heua66jnqDexjoUuAp4BHgIvYSbKjxae5uVwzaRKFQP/+/XnnnXeIi4vzSX/Nz7q+obz6zIAd2GNmeLWZ2GYo9/b+697Wu3dvNm7cyKZNm+jRowcLFy4kJibG9ZkGbdo4vk0G8qnlH+nJyTUdbdRatmxJ69at2bBhg2N0XKFcxJnLI+XJycku/RERERHxFE+EcnMvZH8Ve6v+zyNz7fEEYBNwMXAK+azlIHAtMA7YA9zbogVLFi6ke/fuvupug0O588yAH4B04GzAntKrzsTevn070LRHyuHYuvLff/+dbdu2kZ+fzx//+EeCXF3/feqp9rn9R0NrCFWmroP9eLt29mUNTVC3rl3ZvXs3ixctonv79qQqM4g4cb+MqYiIiIiPrF27llatWtG6det6n8Pf68qHDauat2zAWqAj8D7wC/aJyj9xK7/THngN6AQsBB44cICQpUt92t+4uDgSExMdSwfcdWxmwBZgPHAE+5T9z6q1M0fK27dvX8/eNg5Vt0Vze+o62Iu3PfOM/fvj11qbP8+c2eSKvAGQmUm3334DoNIwODM7u8lv8ybiaQrlIiIi0ijt3r2bgoICevbs2aCiUCeffDJBQUF+GymvmrdgG/YNr+zrwy30xcJcpjOE/sA+4DpgBfYq3IBf1hF37dqVwsJCioqK3H6ufRD0AHA+sBf409FH7sZ+U+JYu6ysLBISEoiKimpwn/2pagX2hQsXAm6GcrBXv5szB1JSnI+npp64PH9jdrS4QLcq2+udCU1+mzcRT1MoFxERkUZp3bp1gL26dUNERkbSq1cv1q1b55O9t2ti5q1Wrcyp6/YQl0oucxjP/SziJ2An9vJv0VWf7IepvmZxrvqMlp96qo3w8Cuwzwi4EngTuBBYDbzrmIk9cGApu3btavLrycFeGT8yMtIxUh4UFORYNuGWjAzYvh1++AHefdf+ddu2phnIqxQXMMs0WoERcKzgwM0329uJNHMK5SIiItIorV27FsAj+xn369ePyspKVq9e3eBz1VdGBtx4o/31b7jhJH6YZ2NbyjAyLB8D9n+UJVV9gh/XETdkXfn06XdTWvo5MBh4Ffvq6Iewv8PpGEYZM2dCXl420PTXkwNYrVZO6tWL7du32wsKdupEdERE/U4WFGTf9uyyy+xfm+KUdXAqLtDj6KFBQKz5eNXiAiLNnEK5iIiINEpr1qwBGj5SDjiqrv92dG2rv6xZYx8pv/763owYGUTQs0/bH2hk64jruy3aO++8w2OPPUZKSgqvvfYRqanhRx/pCVwFbGfy5H8dHRDeDgRGKCczk95Hr1eAIRs3at10lWUXXYEXgRfqaCfSXLkdyq+//noef/xxb/RFREREjrLZYP58eO89+9fmOMPTHCn3RCjv168fACtWrGjwuRpi9erVhISEHKsm30jXEddn+vqyZcu45pprCA8P5+OPP2by5GSnmdizZ88gNDSUTz99kIMHDwbEHuWAY9107ypLI4aC1k0ft+ziz8DJLrQTaY5c3qf8448/5owzzuCXX36hvLzccXzQoEGcdNJJvP76617poIiISHOTmWlfinlsWyl7Rnvmmaa5tLS+1q5dS1xcHElJSXU3rkOfPn0A/46Ul5WVsXHjRnr06EFoaOixBzIy7Jt2L1hgHzVMTrZPWffjtOVOnTphsVjYsGGDS+0PHz5MRkYGZWVlzJ49mwEDBgDHZmLbtWfp0qnMnDmTmf/8J2WbNwOQduCA/a5TU5ymXWXddNUdyYcA1TZlb4rvryHMbQfy8qpuWn+MxWJ/vIlu8ybiSS6PlD/88MOMHj0agF27dvH555+zbds2jJr+komIiEi9HB10cwrk0PwG3YqKiti5cye9evVqUOV1U1xcHOnp6fz+++/Y/DTtYP369dhsNkelbieNbB1xREQEnTt3Zs2aNZSVldXZ/pdffmHHjh1cdtllXHLJJbW2+/vf/050eDhP3n8/v7zzDgDp997bdKd6V1k3/Yejh9KBtubjzXnddCBv8ybiYS6H8s8++4y//OUvABQUFHD//fdzySWXYLFYyM7O5vXXX69XMRARERGxqzLoVk1zK1bsyanrpn79+nH48GG//Xtl1Sr7evKTTjrJL6/vrsGDB1NeXs7vv/9eZ9uff/4ZgNNOO+2E7VovWMBtpaUcMAy+PnosDZruXacq66ETsK+bfqmOds1KI12eIdLYuBzKw8PDufLKKwH7XoyZmZnMmDEDwzAoKyvj3//+t+NxERERcV+VQbejZgH3ApVA8xp080YoN4u9+WtduVn5vcaR8kZo0KBBAC7t7758+XIABg4cWHujo3edbsUeYAFaAHHQdO861bBu+mwX2jUrgbTNm4iXuBzKR48ezdSpUwEoLS0lISGBc889F7Bvm/Hdd9/x5JNPeqeXIiIizYDzYJoB3Il9K6kZJ2gXmLw1Ug7+C+VNbaTcDOU//fRTnW2XL19OaGjoiW84HL3rFAPcffSQU4m3pnjXyVw3XdsSCz9ua9eoNLLlGSKNjcuh/MYbb6S0tBSAzZs3M2LECCZMmADAvn372LFjR51TlkRERKR2zoNpO4GCo98/CHxQS7vA5Mk9yk3+3hZt9erVREdHN5lq43369CEkJKTOUL5v3z42bdpEnz59nAvYHa/K3aQ/A2OBq+to1+hp3bSIeIDLofzKK690VFhv164df/7zn2nXrh0Aubm5XHrppYwbN847vRQREWkGnAfdVhw9ejoQBkwEfm02g25r1qyhRYsWtG3btu7GLkpNTSU+Pp4VK1b4vFDt/v37yc7OplevXlitbu9I6xfh4eH07duXDRs2sG/fvlrb/fLLL0AdU9fB6W5SGPAZcEsd7ZoErZsWkQaq1/8VYmNjmThxIk899RRgn77+8MMPO+5Ai4iIiPuqDrodC+VXA68Ch4Fx3HdffsAPuu3fv5+8vDx69uzpkcrrJovFQr9+/SgsLGTHjh0eO68r1qxZAzSd9eQmcwq7uWa8JmaRN3MbtFoF8lRvrZsWkQZwO5S//PLL3H777U7HwsLCOOuss7j//vs91jEREZHmyBx0i4gwp1j3A/5ETMzfgFxee+1Cx3KyQLVu3TrAs+vJTf4q9tbU1pObXFlX7lKRNwj8qd5aNy0i9eR2KO/fvz/du3d3/Lx8+XLHtHYRERFpuIwMSElZQUhIGG+91Y0ffoDdux9lzJgxLF26lClTpvh8+rUveWM9uclfxd6aWuV10+DBg4G6Q3lkZKTTvw9rpaneIiLVBPu7AyIiIuLswIEDbN68mQEDBvCnP4UcPRrEu+++yymnnMJbb71F7969ue222/zaT28xp3p7c6Tc18XemupIeZcuXYiNjWXZsmUYhlFtOUFBQQHZ2dkMHTqU4GAX/1mZkQHjxtmrrO/caV9DPmyYRpZFpNlqGpVGREREmpHff/8doFqtlhYtWvDpp5/SsmVL7r77bkpKSvzQO+/zxnZopm7duhEeHu7TkfL/b+++w6Oqtj6OfyeTkJCEJh1CL1KlChgIoRephg7ea0O4YgEblityFQsoKmC5giL6SjcEAQURuARC7xCaAtK7lACpZGbePyYTEtKTKZnw+zxPnknO2efMOuQkZJ2999oWi4X9+/dTpkwZypQp47T3tQcPDw8efPBBLl68yOnTp9Pst80nz3Lo+t001FtEJJmSchERkXzG1oubXgHVmjVrEhISQkJCQnLV64Lm4MGD+Pv7J6/yYk+enp40bNiQY8eOERUVZffzp+fChQtcuXLF7XrJbTIbwm6bT55lkTcREcmQknIREZF8xtaLa5v/fLdWrVoBsGXLFmeF5DQ3b97k1KlTdq+8npLtYYdtRIKj2eaTu2tSbiv2tnXr1jT7ct1TLiIiyZSUi4iI5DN79uzBYDDwwAMPpLu/ICflhw8fBhwzdN3G9rDDXvPKTSbr9GiwvppMqffb5pO7W5E3m4wqsFssFrZv306xYsWoUaOGK0ITESkQlJSLiIjkI7dv32b//v3UqlULf3//dNvUrVsXf39/Nm/eXOCqsDuyyJuNPZdFCwuDqlWhZ0/r1z17Wr8OC7vTxt17ysuVK0flypXZsWMHiYmJydvPnj3LxYsXad68OR4e+pNSRCS39BtUREQkHzl06BAJCQnpzie3MRqNtGjRgvPnz3PmzBnnBecEjizyZtOwYUMMBkOek/KwMOjfH+7+Fpw9a91uS8xtSbkjlnhzlhYtWhATE5P8/YEcrE8uIiKZUlIuIiKSj9iGVGc0n9ymoA5hd+Qa5Tb+/v7Url2b/fv3k5CQkKtzmEwwejTYBipYLNHJoxZs28aMgdu3zRw4cICqVatSpEgRO0TvGukNYVeRNxER+1BSLiIiko/Yem8z6ymHgp2U+/r6UrlyZYe+T+PGjbl9+zaHDh3K1fERESl7yGcQHV2K4cOHY7FYl6mzWOD0aVi48DgxMTFuO5/cJr0K7CryJiJiH0rKRURE8pHs9pTbkqSClJRHR0dz/Phx6tat6/A5yraHHrkt9nb+PEAC8C9gJJDA2rVriY3tAdxIbrdzyXIAGhQpkrYCnBtp2rQpHh4eyUm5xWJhx44dlC5d2iFL14mI3EuUlIuIiOQTFouFPXv2UK5cOcqWLZtp2zJlylC9enV27tyZ6yHY+Y0zKq/b2B565HZeubf3RaAjMB2oQeHCmwkKCsJsXp+0/W8Aon76FICGc+emrQDnRvz9/alfvz6RkZFER0fz119/ce3aNR588EGHLV0nInKvUFIuIiKST5w4cYKoqKgsh67btGrVivj4ePbu3evYwJzEGfPJbfJSgX3Hjh2MHt0c2AB0AbZjNLZgxowZGI2PADuAYCqwnZucAKABpK0A52ZatmyJ2Wxm165dmk8uImJHSspFRETyCVuCmNXQdZuCNq/cGZXXbcqWLUv58uXZs2dPjpaVmz17NkFBQZw5c4a+fccCyzEYSgDg7e2Nj8984B/AQUy0ZwvgCdwPqSvAueFQ9pTF3lR5XUTEfpSUi4iI5BPZLfJmU9CScmesUZ5S48aNiYqK4sSJE9lq/8cff/DPf/4Tg8HA3LlzWbx4EosWGalY8U4bg8GTAN6lO9W5SDSnsSbkhWwNbBXgIiLsezFOYEvKt27dmlzkTT3lIiJ5p6RcRETyHZMJwsNh3jzrqxt2KuZKdou82TRq1Ahvb2+2bt3qyLCc5uDBg/j4+FC1alWnvF9Oi71t2bIFi8XCW2+9xZAhQwAICYETJ+CXX6xtfnk1nBPU4Ff+4s2k49JNW62V4txK/fr18fX1Zcu6dezato2A0qUpV7q0q8MSEXF7SspFRCRfCQuz1sNq3x6GDrW+unF9rBzZs2cP/v7+1KhRI1vtCxUqRLNmzTh27BiXL192cHSOFRsby19//UXdunUxGo1Oec+cFnvbv38/kHYkg9EIQUHWz4OCjRgxYwDexzrr/JP0Tla+fC4idi3PpUtplpjI6UuXuBUXx4OXL987P5wiIg6kpFxERPKNsDBrHaw76z9buXl9rGy5cuUKp0+fplGjRjlaDsw2hN3de8v/+OMPLBaL04auQ857yiMjIwFo0KBBxo0CAyEgAJIqkrcGSqbcbzBApUp3snh3kfTD2SJFpf/mcG/8cIqIOJiSchERyRdMJhg9+k4tLDgFrAPcvj5WtuR0PrlNQVmv3NnzyQFq1KhBkSJF2LVrV7baR0ZGUqxYsczX5TYaYepU6+d3LxVm+3rKFGs7d5Hih7NFis0Pwr3xwyki4mBKykVEJF+IiEjZQ34UaAF0AM4Bbl0fK1tyOp/cpqAUe3Nm5XUbDw8PmjZtyrlz5zifxRzvq1evcu7cORo0aJD1utwhIRAaSqoKcGDtQQ8Nte53Jyl+OFum2Jw8V76g/3CKiDiYp6sDEBERgZR1r84CnYGLSV9vAULSaVew5LanvFKlSpQvX55t27ZhMpmcNh87L27fvs2JEyc4duwYR48e5ejRoyxZsgRwzhrlKTVv3px169axc+dOevbsmWE729D1hg0bZu/EISHQp481UT1/3jqHPCjIvXrIbVL80FUG6gBFgBKZtBMRkexTUi4iIvmCte7VFaALcAJrP9wOYBspk3I3rI+VLXv27MFoNOY4KTUYDLRq1YrFixdz6NChzOc7u0hCQgKbN29m1apVrFq1ip07d2JKZ6hznTp1qF69ulNja9asGYD9k3KwJuDt2uUlvPwhxQ+dAdhMBkMtC+oPp4iIgykpFxGRfKFJk1sUKtSDhISDwD+AD4BKWJNy63TcgAD3q4+VHbGxsRw+fJh69erh4+OT4+NtSfmWLVvyTVJ+9epV/u///o/ff/+ddevWERMTk7yvbt263H///dSsWZMaNWokv1auXNnpPf22pNy27nZGcpWUFxRBQdYfvrNnwWKh+N37C/IPp4iIEygpFxERl4uPj6d//0dISNgK9AZmAl5ARWA7YAKMblcfK7v279+PyWTK8dB1m5TzyocPH27HyHJv7NixzJw5E4AKFSowYMAAOnfuTMeOHSlXrpyLo7ujZs2aFC1alJ07d2baLluV1wsqW/G6/v2tCfidaozuW7xORCQfUaE3ERFxqcTERIYOHcrq1asJDg5m3rwFBAR4Je1tAdyiXLnDblkfK7tyW+TNplmzZhiNxnxV7G3Lli14enoSGRnJmTNn+P777xk2bFi+SsjhTrG38+fPc+7cuXTbWCwW9u/fT0BAACVKpJlJfW8oaMXrRETyESXlIiLiUm+++SZhYWE0bdqUpUuXMniwDydOwNq1MHiwdQGm997bVqD/5rcl5bntKffz8+OBBx7g4MGDREVF2TGy3ImOjubQoUPUr18/e9XKXax5c2sd8Yx6y0+dOsXNmzfvzaHrKYWEkPzDOXeu9fX4cSXkIiJ5pKRcRERcJi4ujunTp1OiRAl+++03ihYtCtypjzVihHUBpu3bt7owSsfbtGlTco9tbrVq1QqLxcL27dvtGFnu7Nu3D7PZnKfrcaaUxd7Sc08PXb+b7YdzyBDrq4asi4jkmZJyERFxmZUrV3Ljxg369+9P6dKl0+xv1qwZBoOBbdu2uSA657h69SqRkZE0adKEYsWK5fo8tnnlW7e6/gGGLbm1Jbv5XVbF3u7pIm8iIuJwSspFRMRl5s+fD8DgwYPT3V+0aFHq1q3Lvn37iI2NdWZoTrN+/XosFgvBwcF5Ok/KYm+utmvXLgC36SmvUaMGxYoVy7KnXEm5iIg4gpJyERFxiejoaJYuXUrZsmUzTUhbtmyJyWRKTvQKmnXr1gHkOSmvVasWJUqUYMuWLVhSVsd2gZ07d+Lh4UGjRo1cGkd22aYOXLhwId1ib5GRkRiNRurWreuC6EREpKBTUi4iIi7x66+/EhMTw4ABAzJdm7pFC2uxt4I6hD08PByDwUBQHtd4NhgMPPTQQ/z999/8+eefdoou5+Li4jhw4AB169bF19fXZXHklK3Y291D2BMSEjh8+DC1a9fG29vbFaGJiEgBp6RcRERcwjZ0fdCgQZm2K8hJ+bVr19i7dy+NGjWyy1JbtsQ+IiIiz+fKrX379mEymdxmPrlNRsXe/vjjDxITEzV0XUREHEZJuYiION2NGzdYvnw5AQEBBAYGZtq2YcOG+Pj45IsCZva2YcMGu8wnt7El5evXr7fL+XLD3eaT22SUlGs+uYiIOJqSchERcbolS5YQHx/PoEGD8PDI/L8iLy8vmjZtyvHjx7l8+bKTInSO8PBwANq1a2eX8zVv3hwfHx+X9pS7W+V1G1uxtx07dqSak6+kXEREHE1JuYiIOF12h67b2Iaw54c1uO3JVuQtr/PJbby9vWnZsiUnTpzg9OnTdjlnTu3atQuDweA2Rd5sDAYDzZo14+LFi6mKve3fvx/QGuUiIuI4Lk3Kr127xujRo2nTpg0hISFZzhc8d+4crVu3ZsKECU6KUERE7O3q1av8/vvvVK9ePbm4VlZatmwJ5I81uO0lKiqK3bt388ADD1CyZEm7ndeV88rj4+OJjIykdu3aFClSxOnvn1fpFXuLjIzEz8+PatWquSosEREp4FyalE+aNImSJUuyevVqRo8ezRtvvEFUVFSG7T/99FPuv/9+J0YoIiL2FhYWRmJiIoMGDcJgMGTrmIJY7G3Dhg2YzWa7zSe3adu2LeCapPzAgQPcvn3b7Yau29w9r/zGjRucPHmS+vXrZznNQkREJLc8XfXGMTExhIeHs2TJEnx8fAgODqZGjRqsW7eO3r17p2m/efNmLBYLLVu25NKlS5meOyEhgYSEhFTbPD09KVSokF2vQZzHbDanehXJDt03+ZNt6PrAgQOz/b2pUqUKpUqVYtu2bZhMpmwn87nhrPvGNp88KCjIru/VsmVLjEYjERERTr/3bdMLmjRp4pY/d02aNAGsPeVms5l9+/YB1qHrWV2Pft9Ibui+kdzQfeNesvNQ12VJ+alTp/D19aVs2bLJ22rWrMlff/2Vpu3t27eZOnUqkydP5tdff83y3LNmzeKbb75JtW3AgAEMHDgw74GLS7lqjqS4N903+cfly5dZu3YtNWvWpGjRopw8eTLbxzZo0IDw8HAiIiKoUqWKA6O0cvR98/vvvwNQrVq1HP07ZEe9evWIjIxkz549dllqLbtsVd8rVqxo92tyBqPRSNGiRdm2bRsnTpxInvOfk+vR7xvJDd03khu6b9xDdqY/uSwpj42Nxc/PL9U2Pz+/dIevz5kzh9atWxMQEJCtcz/xxBMMGzYs1Tb1lLs3s9nM6dOnqVSpkoYQSrbl9/vGZIJNm+DiRShbFgIDwWh0dVSO9euvv2I2mxk2bBhVq1bN0bHBwcGEh4dz5syZ5CHajuCM++bmzZscOHCA+vXrO2TpsE6dOhEZGcnJkydp3LixXc6Znfv1yJEjAHTr1o1ixYrZ5X2d7cEHH2TNmjV4enomF3xr27Ztlg+C8vvvG8mfdN9Ibui+KXhclpQXLlyY6OjoVNuio6Px9fVNte3SpUssXbqU2bNnZ/vchQoVUgJeQHl4eOiXj+RYfrxvwsJg9Gg4c+bOtoAAmDoVQkJcF5ejLVy4EIDBgwfn+HtiK/a2Y8cOHn30UbvHdjdH3jebN2/GZDIRHBzskPdo27Ytn332GRs3buSRRx7J8/myc7/evn2bffv2UbNmTaf2zttb8+bNWbNmDbt3706uvN6oUaNsf5/y4+8byf9030hu6L4pOFyWlFeuXJmYmBguXbpEmTJlADh27Bg9evRI1e7gwYNcvHgx+Y+KmJgYzGYz58+f56uvvnJ63CIieRUWBv37Q4qlkAE4e9a6PTS0YCbmZ86cISIigsaNG1OnTp0cH1+Qir3Ze33yu7Vp0wa4M5w8L7J7vx48eJD4+HiH9Pw7k63Y244dO4iMjKRs2bKULl3axVGJiEhB5rJHK76+vgQHBzN9+nTi4uKIiIjg6NGjaarQBgYGsmTJEubMmcOcOXPo168f7du354MPPnBR5CIiuWcyWXscUyc4JuDOtjFjrO0Kmnnz5gHZX5v8biVLlqRGjRrs2rUrTTFPd2Obq+yoYfilSpWiXr167Nq1i1u3buX6PGnv1zs35t33665duwDctvK6jS3+ZcuWce3aNa1PLiIiDufS8Q6vv/46ly9fpmPHjnz22Wd88MEHFCtWjBUrViQXZStUqBClSpVK/ihcuDDe3t4UL17claGLiORKRETqIcDwOlAKuABYE53Tp63tCpLo6Gg++eQTvLy8GDp0aK7P07Jly+S1sN3VrVu32LFjB3Xq1ElV7NTegoKCMJlMbNmyJdfnSH2/XgBqAINJ+SDJdr/alhFz957yatWqUaJECfbu3QtAwxIlCuZTMhERyTdcmpSXKFGCadOmsXHjRsLCwpLnC3bv3j153uHdRo4cybhx45wZpoiI3Zw/n/KrbcBHwHVgTSbtXMdkgvBwmDfP+prb3GTatGlcvHiRf/3rX1SuXDnX8RSEIeybNm0iMTHRYUPXbYKCgoC8DWFPfR++BZwEFgBvpGln6yl396TcsHgxzWJikr9uGBoKVatax/GLiIg4gCoDiIg4Ufnyts8SgZGAbVzw5gzauU5YmDUXad8ehg61vuYmN7l69SqTJk3Cz8+Pf//733mKyZaUb926NU/nsYfcPrCwDV2/e7qWvdmGxkfkYdjFnftwN/AdUBkIAD4GfkhuV6ZMInv27KFq1arcd999uX4/l0uaQN88Pj55U0O4M4FeibmIiDiAknIREScKCrJWrYYvgD1AJ8AAbALAYIBKlaztXMlW3Cv1UPvc5SaTJk0iKiqKl156Kc/DtZs0aYKnp6fLe8rz8sDCWUl5pUqVqFKlClu2bMn1HPygIKhY0QK8iPUB0mRgCVAYGAFsplKAhdInfiQ2NpZmlSu771DvFBPobbPiDUB9KPgFH0RExKWUlIuIOJHRCG+/fQYYB/gBM7H+2b8PsBbkmjLFteuVpy3uZQH+BP7GkrQxu7nJ2bNnmTZtGiVLluTll1/Oc2w+Pj40atSIw4cPc/369TyfLzfy8sAiJiaGbdu2Ubt2bco7YThEUFAQcXFx7NixI1fHG40wZEgYsA4IAvoDTbH2kicAj/DWraHsGf4kAE3Xr3ffod4pJtDbkvIaQPJCrQW14IOIiLicknIRESf77bfRwC2KFXsH63DgQMBE6dLb88VyaGmL0c0F7gdKA8WxWJpw+nQ/hg17le+++47Y2NgMzzVhwgTi4uJ48803KVasmF3iCwoKwmKxJC8r5kxpH1hEAcsBU7Y6Uzdv3szt27cd3ktuk9ch7HFxcSxa9CoGg4EyZT7D2ncMMICihd8ALvL19fnYzt4M3Heod4oJ9FWBl7DOos+snYiIiD0oKRcRcaJffvmFsLAwGjVqxMWLo1m7FkaOfAiA55/f5PKEHNLLOTYkvTYBigN7gTAWLJjMU089RXBwMOfOnbM2STHR+sjs2Xz77bcEBAQwatQou8XXpUsXAFauXGm3c2ZX2gcWk4EewCjAkmVn6v/+9z/AceuT381W7C23SfnUqVM5fvw4jz/+OOfONWPtWpg7F9auNvF3idn0wzrb/Nuk9k3BfYd6pxi5YAA+AR7Lop2IiIg9eLo6ABGRe0V0dDTPPfccBoOBr7/+Gm9vT9q1gwoVApk+HbZu3ZzlOZwhbc5xIOl1FVASiAdOMnHiMTZu/C/Lli2jRYsWLB0zhqZTpyZnreOwLpz1n1698PHxsVt8wcHBeHt7s3LlSiwWCwaDIeuD7CTtAwvb0mwzsP7bfJBBO+sDmcmTJ+Pl5UX79u0dF2QK999/P6VLl2bDhg2YTCaMOZgXceHCBd5//338/f15//33MRoh+VlCeAScO80PwDGs1REqYR1LAaQe6u2kBxB5Ziv4cPZsyqEQdxgM1v2uLvggIiIFjnrKRUSc5N133+XkyZOMGDGCVq1aJW+vVasWJUuWZPPmzclztl3JlptYc10L1qS8LNakEwwGbypVqs0rr3Tn559/5rXXXuPs2bO0efVVQpMS8l1YF866H3jsv/+161BmX19fgoKCOH78OEePHrXbebMj7QOLo4ARqAh8iLUqOZS/uCdVWfbFixcTEhKC2Wxm3rx5TplPDmAwGAgKCiIqKor9+/fn6Nhx48Zx8+ZN3nzzzbTxJj118MNa9q06MCC9k7jTUG+jEaZOtX5+94Me29euLvggIiIFkpJyEREniIyM5NNPP6VMmTJ8+OGHqfYZDAYeeughrl69yp9//umiCO9ImZvAReAqSTWo0+QmHh4eTHz/fX4oUQIT1sRsAvBm0tHvA54Gg92HMnft2hVw/hD21A8szFj7iasBvwP3AWMpYfiYoBebJZdlX1CmDAP698dgMBAWFka/fv2cHHPOh7Dv3r2bmTNnUrVqVV588cW0DVIk6ZWxPpr4JL0TudtQ75AQCA2FihVTbw8IIF8UfBARkQJJSbmIiINZLBZGjRpFYmIin376KSVKlEjTJjAwEIBNmzY5O7x02XKTUqVsQ9etSXm6uUlEBP+8do3/YR2+/DawEmgOhIBDqlbb5pX//vvvdjtndqR+YHEWiANqAfWwFnzzJ8oylp8xA/AjMPTqVTzNZpaMHUuvXr2cGi/cKfa2fv36bLW3WCy8+OKLWCwWPv744/SnHqR+OkGaCQT5ZW2/3AgJgRMnuDOBfi0cP66EXEREHEZJuYiIg82ePZsNGzbQrl07hg4dmm6bhx6yFnvLL0k5WHOQf//bmpQ/9VT9jHOTpCHKrYFtQEOsSdok7krW7DiUuWHDhpQvX561a9fmeg3u3LrzwOJI0pZaAFQyBvAfGuMJDAXGYC0U5g38CnT74QeXFD5r1KgRRYsWJTw8HLPZnGX7yMhI1q1bR2BgYMa9+gV9qLdtAv2QIdZXd70OERFxC0rKRUQcKCoqildffRVPT0+++OKLDIuSPfjggxiNRjZvzh/F3mwOHbIm5Y8/Xj/j3CTFEOWqWBPzP4EOmbTLK4PBQJcuXbh165ZLHmSEhMCECdb57I89VpO1n+3huKky49nAAiARmIp1jesVQEdw2RrXRqORTp06cfnyZXbt2pVle9uUgKFDh2ZeRE9DvUVEROxCSbmIiAONHz+eixcvMnr0aOrXr59hOz8/Pxo3bsyBAwe4fv268wLMwoED1qQ8s9jvHsrsA9RMud9BQ5lduTQawLFj1p7ywYNr0a7sIYxJQ9b7Yh22/iDWmeapViR3UeGz7t27A7B8+fIs29r+PW3z9jOlod4iIiJ5pqRcRMRB9u3bxxdffEH58uUZP358lu1tQ9i3bt3q6NCyxWKxcODAAcqXL5/uPPhkLhrK3LlzZwwGg9PnldvYKr/XrFkzzSiAoVhHDATefZCLCp9169YNgBUrVmTaLiYmhoiICKpXr269ruzQUG8REZE8UVIuIuIAFouFZ599FpPJxOTJkylSpEiWx+S3Ym/nz5/n+vXr1KtXL+vGLhjKXLp0aZo2bcquXbu4dOmS3c+flSNHjuDp6UnVqlXTjBZIw8WFzwICAmjYsCFbt27lypUrGbZbt24dCQkJ2eslFxEREbtQUi4i4gBz5sxhw4YNBAcHM2TIkGwdk9+KvWVr6HpKLhjKbEseV61a5bD3SI/ZbObYsWNUrVoVT09Ptyh81r17dywWS6YjC2xD121TA0RERMTxlJSLiNiZrbib0WjMtLjb3apUqUL58uXZunUrJhdU6b5bjpNycPpQZlfNKz979ixxcXHUqlXrzsZ8XvjMNq88syHsK1euxNPTkw4d0pTpExEREQdRUi4iYmf/+c9/uHDhAi+88AINGjTI9nEGg4GHHnqImzdvJifErnTw4EEgh0m5kz300EP4+/vz+++/Z2u5L3s5csRa5C3NvOt8XPisdevWFClShN9++y3df6tTp05x+PBhHnroIYoWLeqCCEVERO5NSspFROxo//79fP7555QrV47//Oc/OT7eNq88PyyNlquecicrVKgQHTp04OLFi0RGRjrtfW1F3lL1lNvk08JnXl5edO7cOcOl0WzD2jWfXERExLmUlIuI2NEPP/yAyWTi/fffz1VvY34p9marvF6hQgWKFy/u0liyYksinTmEPcOe8nwus6XRNJ9cRETENZSUi4jYkW05M1vyk1NNmzalUKFCLu8pP3fuHFFRUfm6l9zGFfPKM+0pz8cyWhotMTGR1atXU7JkSZo2beqK0ERERO5ZSspFROwkMTGRnTt3UqlSJcrncj1qb29vmjVrxpEjR7h8+bKdI8w+dxi6blOzZk2qV6/Ohg0biI6Odsp7ploOzY1ktDTa9u3buX79Op07d8aYT4bbi4iI3CuUlIuI2MnBgweJiYmhZcuWeTqPbWk0V/aWu1NSDtYh7AkJCaxbt87h75VmOTQ3k97SaJpPLiIi4jpKykVE7GTbtm0AtGjRIk/nyQ/F3twtKXfmEPZ0l0NzI+ktjab55CIiIq7jfo/4RUTyKdt88rwm5baeclcWe7Ml5fXq1XNZDDnRoUMHPD09nZKU2+aTu1uRN5u7l0aLiopi69atNGjQgAoVKrg6PBERkXuOespFROxk27ZteHh40KxZszydp0KFClSpUoXt27dz+/ZtO0WXfRaLhYMHDxIQEECxYsWc/v65UbRoUVq0aMEff/zBhQsXHPpetsrr7tpTfvfSaGvWrMFsNmvouoiIiIsoKRcRsYPo6Gj2799P/fr18ff3z/P5AgMDiY2NdUlv+ZkzZ7hx44bb9JLbBAUFARAREZHrc5hMYDs8IsL69d3cvaccUiyN9uWX/P7VVwB07dTJlSGJiIjcs5SUi4jYwc6dOzGbzXkeum4zcOBAAGbOnGmX8+WEu80nt2nbti2Q+6Q8LAyqVoWePa1f9+xp/TosLHU7d+8pB+hmsQCw4vvvWbl2LT5A0PDhaS9WREREHE5JuYiIHdiKvOW18rpNz549KV++PAsXLuTq1at2OWd2uWtSHhgYiMFgYP369Tk+NiwM+veHM2dSbz971ro9Za569OhRjEYjVapUyWPELhIWRsDIkTQEtgCngGDA59y5tBcrIiIiDqekXETEDuxVed3G09OTp556ivj4eH788Ue7nDO73DUpL168OA888AD79u3j+vXr2T7OZILRoyGp8xiz+a/k423bxoyxtjObzRw9epRq1arh5eVl1/idIsXFdk+xuSukvVgRERFxCiXlIiJ2sHXrVnx9fe2ayA4fPhyDwcCMGTOw2BImJ3C3yusptW3bFovFkqO5+BERcOZMLPAj0IaYmFoMGzYMi8UMWHPV06et7c6dO0dcXJz7zie3XixA2qQcUl+siIiIOIWSchGRPLpw4QKnTp2iWbNmeHrab6XJKlWq0K1bNw4ePOi0gm+2yuuVKlWiaNGiTnlPe8ppsbfDhw/z0UcvARWBfwIbgSIcOnQIk+nXVG3Pny8A88nPn0/+tDVQCqgB1M2knYiIiDiWknIRkTzavn07YL+h6ymNGDECgOnTp9v93Ok5deoUt27dcruh6za2pDyreeUWi4WhQ4dSt25dVqz4DEgEngH2ULjwcgASEj5OdUz58neScrftKS9fPvlTL2ATsBowZNJOREREHEtJuYjYTUxMDNu3b0/34+zZs64Oz2G2bt0KOCYp79Gjh1MLvrnrfHKbcuXKUatWLbZv305sbGyG7bZu3cq8efOoXr06X389gwoVzmEwfAU0wmgMpHnz5pjNG4GNGAxQqRIEBd1ZDs1te8qDgiAgAAzWNLwWUDXl/pQXKyIiIk6hpFxE7GbgwIG0aNEi3Y/atWtz+vRpV4foEPauvJ6Sl5cXTz75JPHx8cyePdvu57+buyflYO0tv337dvL3JT0LFiwAYMKECYwc+TSff25dWz4pV2XkyJFJLScBMGUKGI0FYPi60QhTp1o/N9zVP2772naxIiIi4hRKykXELi5cuMCKFSuoUKECI0aMSPXRrVs3YmJimDhxoqvDzJDJBOHhMG+e9TW7xafNZjPbt2+nTJkyVK5c2SGxObPgW0FIym3rlWc0hN1sNrNw4UIKFy5M7969AQgJgdBQqFjR2qZ9+/Z4eNQHlvHppwcICbFud/vl0CDtxdoEBFi32y5WREREnMJ+FYlE5J4WGhqK2Wxm5MiRvP3226n2xcbGUq1aNb799lveeOMNAgICXBRl+sLCrKtEpVyjOiDA2qGYVX5y9OhRrl+/Ts+ePTHc3fNoJ1WrVqVr16789ttvbNq0idatW+f6XMeOHWPVqlWsXr2ac+fOpdnvzpXXbbIq9rZhwwbOnTtH//798ff3T94eEgJ9+sCGDeDh4cHYsS8zceKT7NnzMfC9+y+HlpLtYiMirEXdype3DllXD7mIiIjTKSkXEbuwDQceNGhQmn2FCxdm7NixvPzyy3z00UdMmzbN2eFlKCwM+ve/s0Szzdmz1u1ZdRza5pM7Yuh6SiNGjOC3335jxowZOUrKb968ycqVK1m1ahWrVq3i+PHjWR7TsWPHVMmqu6lWrRoVKlRg06ZNJCYmpqmIP3/+fAAGDx6c5lij0ZqbnjwJ48cPYfbst5kzZw4TJkzAYDC493JodzMaoV07V0chIiJyz9PwdRHJs9OnT7NhwwaaNGnC/fffn26bf/3rX5QpU4YZM2ak20PrCiaTtYfcmpCbgf8Bt4E7SfqYMZkPZbfNW3ZEkbeUevbsSbly5Vi4cCHXrl3L1jFms5m2bdsyYMCA5H/3jh07MnHiRHbs2EFMTAyxsbFpPlatWuXQa3E0g8FA27ZtiY6OZvfu3an2JSYmEhoair+/Pw8//HCm5ylUqBAvvfQSiYmJTJkyxf3nk4uIiEi+pKRcRPJs4cKFQPq95Da+vr68+uqrxMfH89FHHzkrtExFRKQcsj4N6Ai8mLzfYoHTp63tMmJLyh988EFHhQlYC7499dRTxMXFZbvg27Zt29izZw8tWrRg5cqVXLt2jdWrV/Paa6/RrFkzChcujI+PT5oPRw3Dd6aMhrCvXbuWy5cv06dPHwoXLpzleZ5++mlKlCjBjBkz2J7Uw17TYsl+0QERERGRLCgpF5E8y2zoekrPPPMMpUqVYvr06Zw/f94ZoWXqTghRwHtJn38FbMugXWrx8fHs2bOH2rVrU6JECccEmcJTTz2FwWBg+vTp2Sr4FhoaCsDo0aPp0qVLtpLQgiKjpDyzoevp8ff359mOHbl16xYfzpgBQK0vvoCqVa1zH0RERETySEm5iOTJsWPH2L59O61ataJq1aqZtvXz8+OVV14hLi6OyZMnOyfATJQvb/tsMnAFCAQswL+AxHTapbZ3714SEhIcPnTdplq1anTp0oUDBw4kz2XPiMViYdGiRRQqVIiePXs6Jb78pH79+pQoUYKIiAjMZjMACQkJhIWFUbx4cbp06ZK9E4WF8XxoKD7A9aRNNeFO0QEl5iIiIpJHSspFJE+yM3Q9pWeffZaSJUvy3//+l4sXLzoytCwFBUH58heAT4FiwDKgD7Ab+AKDASpVsrZLj7Pmk6c0fPhwAL799ttM2+3atYsTJ07QpUsXihYt6ozQ8hUPDw/atGnDlStXOHz4MAC///47169fJyQkhEKFCmV9kqSiA2WAJ5I2GYGqkP2iAyIiIiJZUFIuInkyf/58DAYDAwYMyFZ7f39/XnrpJWJjY13eW240QqNG7wExwGvAfVjnlvsC47BYzjBlSsarRDmr8npKvXv3plSpUsyfP5+bN29m2M42dL1fv37OCi3fuXsIu23oenYfILFpU3LRgVew/odZE0heDC07RQdEREREsqCkXERy7dChQ+zbt4+2bdtSsWLFbB/33HPPUaJECb766isuXbrkwAgzd+zYMVavnk6JEuWpUGF00tbKwDvALVq1GkNIHxOEh8O8edbXFL2i27Ztw8vLi0aNGjkt5kKFCvHYY48RHR2dnGTezWKxEBoaiqenJ71793ZabPlN27ZtAVi/fj2xsbEsWbKEUqVK0aFDh+ydIMVIjurAr8AP6bXLB/URRERExH0pKReRXMtugbe7FS1alJdeeomYmBg++eQTR4SWLePGjSMxMZEPPnibU6d8WbsW5s6FVatG88ADD7BlyyJ+LVsW2reHoUOtr0kFvq5du8aff/5J48aN8fb2dmrcWQ1hj4yM5OjRo3Ts2JH77rvPmaHlK02bNsXX15eI1atZ/uab3Lp1iwH9+6dZtzxDZcum+rIbkO6YiIyKDoiIiIhkg5JyEckVi8XC/PnzMRqNuRoi/fzzz1O8eHG+/PJLoqKiHBBh5nbv3s28efOoVasWTz31FEYjtGsHQ4ZAp05efD14MAbg2StXiEl5YFKBr22ffgo4d+i6TZ06dWjTpg3btm1j3759afYvWrQIuLeHrgN4LVtGK5OJ05cu8dGUKQAMWrQo+8XZAgMhIAAyWiIuq6IDIiIiItmgpFzkHmHKeBR2ruzdu5c//viDjh07UqZMmRwfX6xYMZ588kmio6P55Zdf8hZMLrz55psAvPfee3h5eaXeaTLx0FdfMQI4CbyLtSb7HmCSxUIHi4Ve71mXUHNmkbeUbL3lM2fOTLMvNDQUDw8P+vbt6+So8pGwMOjfn7bx8YB1kbsKQJvLl7NfNd1ohKlTrZ/fnZjbvs6s6ICIiIhINigpF7kHhIVZR12nMwo713I7dD0lW3E4W8+us4SHh/Pbb7/RrFkz+vfvn7ZBRAScOcOHQGngE6wJXRPgdWAt1pnno/v1IyQkxHmBp9C/f3+KFi3Kjz/+SFxcXPL2Q4cOcfDgQdq1a0fp0qVdEpvLJVVNx2IhZR/2AKzV04HsV00PCYHQULi7ZkJAgHW7i77/IiIiUnAoKRcp4JI6DG1FpJPlZZll29B1Ly8vHnnkkVzH1qJFCwICAlixYgW3bt3K9Xly4vbt27z66qsATJw4EQ+PdH4NJhXuKgFMxbpi+S2gN/AlcDTpY0q/fvj5+Tkj7DT8/PwYOnQo165dY/HixcnbNXSd5IcqAK0A2wzywbb9Oa2aHhICJ06QXHRg7Vo4flwJuYiIiNiFknKRAixFh2ESC2Ctdp6XZZa3bdvGiRMn6Nq1KyVKlMh1fB4eHoSEhBAXF8fy5ctzfZ6cGDNmDDt27KB79+506tQp/UYpCncNwTqE/QqwBBgF1EinnSukV/AtNDQUg8GQp4clbi9FNXRf4FGgA+kUactJ1fSURQfatdOQdREREbEbJeUiBViKDsMkc4CygLU3NTfLLF+6dImXXnoJgMGDB2fROmu24ePOGML+3//+l6+++orq1avz448/ZtwwKChVga/KQKGU+/NJga+mTZvSuHFj/ve//3Hs2DGOHj3K3r17ad26NeXv5Yrgd137LGANkKZc2738byQiIiL5hpJykQIsbUfg0qTXcYApk3bp27lzJ82bN2fTpk107tw5/fnYORQYGEi5cuX49ddfiYmJyfqAXFq7di3PP/88RYoUYenSpZQsWTLjxm5S4MtgMCT3ln83fjyLxo0DoP+9Pqz6rocqaeSThyoiIiIioKRcpEBL3RFoAdYnfX4ICM2gXfrmzJlDmzZtOH36NC+//DLLly+3y/rcRqORRx55hOjoaFauXJnn86Xn2LFj9O/fH7PZzLx586hfv37WB7lJga+hRYrgA8yaM4cF8+cDEPLRR3mr4ufu3OShioiIiAgoKRcp0FJ3GB4FLgI1k/ZOAMxZdhgmJibyyiuv8OijjwLw448/MnnyZDw9PTM+KIccOYT95s2b9O3bl6tXrzJx4kR69OiR/YPze4GvsDBKPP44/YHzwG6s86YrXbyY+yp+BYWbPFQRERERsd9f1SKS79g6DK05r23i+BNYV21eAixiypQBGXYYXrt2jUGDBrFq1SoCAgL4+eefadasmd3jbNu2LaVKlWLp0qXEx8fbpQcewGQyMWbMGA4ePMg//vGP5KrrOWIr8JXfpKjiNxyYnbS5P1iLBRgM1ip+ffrcuz3CISHW64+IsM7RKF/e+gTqXv33EBERkXxJPeUiBZytw9DX1zZ0vS3wNgCVK79L377mdI9LSEigV69erFq1ijZt2rBjxw6HJOQAnp6e9O3bl5s3b7Jq1Sq7nNNsNjN27FjWrl1Lq1atmDFjBoaM5hi7oxRV/NoCtZI2Jy+ElpsqfgWRqqaLiIhIPqekXCSH1qxZQ9++fendu3eaj5EjRxIbG+vqENMICYFy5SLw8vLmhx8eZO3apvTs2YtTp/anWuM6pTFjxrBx40a6dOnCmjVrKFu2rENjtOcQ9uPHj9OxY0emTJlCuXLlCA0NxcfHJ8/nzVdSVOczAGHAL0C1TNqJiIiISP6j4esiOfTyyy+zd+/eDPeXLVuWd99914kRZe3s2bP89ddfBAUF8c9/WoeG+/u/zS+/LOPdd9/lkUcewcPjzjO6mTNn8t///pfq1aszb948ChUqlNGp7aZDhw6UKFGCn3/+menTp+fqPS0WC9OnT+eVV14hOjqa4OBg3n333YK5PNhd19Qg6SOrdiIiIiKSv6inXCQHDh8+zN69ewkMDOTUqVOpPg4ePEjJkiWZOHEihw8fdnWoqUQkDWFu27Zt8rbmzZvz8MMPs2/fPpZ+8AHMmwfh4WzbvJlRo0bh6+vLzz//zH333eeUGL28vOjduzfXr19n7dq1OT7+1KlTdOnShWeeeQaz2cy0adNYvXo1lSpVckC0+YCW/RIREREpEJSUi+TAggULABg2bBiVKlVK9VG3bl0+/vhjbt++zTPPPIPFYnFxtHfYkvKguxK08UlJ+rvjxmEZOpQL7dsTEhREQkICs2bNomHDhk6NM7dD2OfMmUODBg1YvXo1rVu3Zt++fTz//POpev8LHC37JSIiIlIgFOC/WEXsy2KxMH/+fDw8POjXr1+6bR5//HGCgoIIDw9n9uzZ6bZxhYiICDw8PHjooYfubAwLo8Ubb9AN61JaYcAA4KzJxFhgoB2XPMuuzp07U6RIERYvXkxiYmK2jjlz5gz//Oc/uX37Np988gnr1q2jZs2aWR9YEGjZLxERERG3p6RcJJsiIyM5fPgwHTp0yLDomcFg4Ouvv8bT05OXX36Zq1evOjnKtK5evUpkZCRNmjShaNGi1o0pltMan9RuKLAB6AJ8ANbltEwmp8bq7e1Nr169+Pvvv1m/fn3WBwBz587FbDbzzjvv8NJLL2G813qG8/ta6iIiIiKSKSXlItk0f/58AAYPHpxpu3r16vHqq69y+fJlXn/9dWeElqmNGzcCdw1dT7GcViusiXgC1srd8wAjuGw5rZwOYZ89ezYGg4GhQ4c6Mqz8Tct+iYiIiLgtJeUi2WCxWFiwYAGenp488sgjWbZ/6623qFatGt98801yUuwq6c4nv2uZrE+Ah4GlQKqybi5YTqtbt274+fnx008/ERcXl2nbffv2ERkZSfv27QkICHBShCIiIiIi9qOkXCQbduzYwV9//UXXrl2zVY3c19eXL7/8EoB//etf3L5929EhZsg2DDxVUp7Oclq/ks6SWi5YTqtw4cIMGzaMy5cv8+OPP2ba1jZv/9FHH3VGaCIiIiIidqekXCQbbEPXBw0alO1junfvzoABA9i/fz+fffaZo0LLVHR0NDt37qROnTqULl36zo58vpzWK6+8gsFg4OOPP8aUwbx2k8nE3Llz8fHxIUTzp0VERETETSkpF8mC2Wxm4cKFeHt706dPnxwdO2XKFIoUKcJ//vMfLl686KAIM7Z161YSExPTLIWW35fTqlWrFiEhIRw5coSff/453Tbr1q3j7Nmz9O7dm2LFijk3QBERERERO1FSLpKFTZs2cebMGXr06HGnenk2VahQgTFjxhAbG8uyZcscFGHGbEPX2yatR55KPl9O67XXXgNg0qRJ6a75rqHrIiIiIlIQKCkXyUJuhq6n1Lt3bwBWrFhht5iyK90ibynl4+W0HnzwQdq3b8/27dtZt25dqn2xsbGEhoZSsmRJunbt6qIIRURERETyTkm5SCYSExP56aef8PPzo0ePHrk6R9OmTSlTpgyrVq0iISHBzhFm7Pbt22zZsoVKlSpRpUqVjBvm4+W0UvaWp7Rs2TJu3rzJoEGDKFSokCtCExERERGxCyXlIplYt24dly5dolevXvj5+eXqHB4eHnTr1o2bN2+yadMmO0eYsV27dhETE5P+0HU30aVLFxo3bsxvv/3G3r17k7dr6LqIiIiIFBRKykUysWDBAgAGDx6cp/M8/PDDACxfvjzPMWVXukuhuRmDwcDYsWMB+OijjwD4+++/WbFiBdWrV6dVq1auDE9EREREJM+UlItkICEhgUWLFlGsWDG6deuWp3N17twZDw8Ppybl4eHhgHsn5QADBgygatWqLFiwgBPz57Nw7FgSExN5dOhQDBkt6SYiIiIi4iaUlItkYPXq1Vy9epW+ffvi7e2dp3Pdd999PPTQQxw4cIBTp07ZKcL0XblyhaFDh7J8+XIqVqxI3bp1Hfp+jubp6cnLHTtiMpn4dMgQZs+aBcCwb76BsDAXRyciIiIikjdKyiVdYWFh1KlTh1q1aqX5ePDBBzl58qSrQ3S4RYsWATBw4EC7nM82hN2RVdiXLl1K/fr1mTdvHvXq1WPZsmXu35scFsaTM2dSCpgBbAZaALUvXYL+/ZWYi4iIiIhbU1Iuady6dYtRo0bxxx9/cOHChVQf586dY8eOHYwaNSrdtaMLCpPJxLJlyyhSpAgdO3a0yzm7d+8OOCYpv3btGo899hh9+vTh8uXLvP766+zcuZMmTZrY/b2cymSC0aPxBZ4H4pM2Pwpgu//GjLG2ExERERFxQ0rKJY3Jkydz8eJFXnjhBW7evJnq48aNGzRr1ozly5ezePFiV4eaLpMJwsNh3jzra27ytS1btnD58mW6deuW56HrNo0bN6Z8+fKsXr2a+Pj4rA/IphUrVtCgQQP+7//+j/vvv5+NGzfy4Ycf4uPjY7f3cJmICDhzBoBnAV/ACCSvGG+xwOnT1nYiIiIiIm5ISbmkcu7cOT7++GOKFSvGuHHj0uw3Go1Mnz4dDw+P5KQ9PwkLg6pVoX17GDrU+lq1as5HOP/8888A9O3b126xGQwGunXrRnR0NBF2SCKjoqIYPnw4Dz/8MOfPn+ell15i9+7dBasi+fnzyZ+WBMKAhUCZTNqJiIiIiLgTJeWSyvjx44mJieHNN9+kVKlS6bZp1qwZzz33HGfPnuXtt992coQZCwuzTjFO6lhNdvZszqYeWywWlixZgqenZ/I8cHux17zy1atX07BhQ2bOnEnNmjVZv349n3zyCYULF7ZHmPlH+fKpvuwKhGSjnYiIiIiIu1BSLskiIyP57rvvqFy5Mi+88EKmbSdMmECFChWYNm0au3fvdlKEGUuaesydae5XAesXOZ16fPjwYY4cOUJwcDDFixe3a5ydO3fGaDTmemm0W7du8cwzz9C5c2dOnz7NCy+8wJ49e2jTpo1d48w3goIgIAAyKlZnMEClStZ2IiIiIiJuSEm5JBs7dixms5kPPvggy/nIRYsWZerUqZjNZkaOHInJxYW2Ukw9BpYCpYCBwC0gZ1OPlyxZAkCfPn3sHmexYsVo3bo1hw8f5vjx4zk6duPGjTRs2JCvv/6aatWqsXbtWqZOnYqfn5/d48w3jEaYOtX6+d2Jue3rKVOs7URERERE3JCScgFg1apV/PbbbzRr1owhQ4Zk65h+/frRvXt3tm/fzvTp0x0cYebuTCm2ABOSXkOBQOB4Ou0y5sikHHI3hP327dv07t2bEydO8Mwzz7Bv3z7atWvnkPjynZAQCA2FihVTbw8IsG4PSXdAu4iIiIiIW1BSLphMJl599VXAWnndwyN7t4XBYODLL7/Ex8eHN954gwsXLjgyzEzdmVK8AdgBtAUGA5FAc2DNXe3Sd/78ebZs2UKTJk2oXLmyQ2K1LY2WkyHsW7Zs4erVqwwdOpSvvvoKf39/h8SWb4WEwIkTsHYtzJ1rfT1+XAm5iIiIiLg9JeVuyh7LftnMnj2bvXv30qtXrxz3vlarVo23336bGzdu8OKLL+Y+iDyyTT2GT5K2vAbMBSYB14CuFC8+hTZtMl9bfdmyZYDjeskBGjZsSMWKFfnf//5HXFxcto5ZuXIlcCehvycZjdCuHQwZYn3VkHURERERKQCUlLshey37BRATE8O///1vjEYjkyZNylU8L7/8MvXq1WP+/Pls2LAhV+fIK6MRXnvtCNb55HWAboABGAssB4pw/fqLDB/+RKaJsKOHroN1hMHDDz9MbGws69aty9YxtqS8c+fODotLREREREScT0m5m7HXsl8206ZN4+zZszz99NPUrVs3VzEVKlQoOaGfaivK5QKHDk0BLBQv/hIpb+1Klbrx+efbqFevHj/88AOjRo1K9/hbt26xZs0aqlSpQqNGjRwaq63HOzvzyv/++2927txJ48aNKVu2rEPjEhERERER51JS7kbSLvt1R06X/QK4du0akyZNwtfXl/Hjx+cptu7du1O9enUWL17MmbufGDjBlStXmDVrFqVLl+b06UfTTD1+7rlabN68mdq1azNr1iwWLFiQ5hwrV64kPj6ePn36YMhoCS476dixI15eXvz6669Y0vuGprB69WosFgtdu3Z1aEwiIiIiIuJ8SsrdSOplv04DDwHDsRY2y9myXwCTJk3i+vXrvPjii5QrVy5PsRmNRkaNGoXJZHJJJfbp06cTGxvLqFGj8PcvnO7U46JFizJv3jy8vLwYOXIkJ0+eTHWOn3/+GXDs0HWbokWL0q5dO44ePcquXbsybWsbuq6kXERERESk4FFS7kZSL+f1HrAFmAk8CDQDZgC3srXs17lz55g2bRr33XdfcuX1vHryyScpXLgwM2bMID4+3i7nzI74+Hg+//xzvL29MxyabtO0aVM+/PBDoqKiGDZsGImJiYB1ybFff/2V4sWLExQU5IyweeKJJwD49ttvM2xjsVj4/fff8fX1JTAw0ClxiYiIiIiI8ygpdyN3lvM6C3wPlAPmAO2AXcBIoALz5z/DuXPnMj3XhAkTiI2N5Y033qBYsWJ2ia9EiRI8+uijXLp0iZ9++sku58yO+fPnc+HCBf7xj39QpkyZLNu/+OKLdO7cmY0bN/LBe+9BeDgb3nmHa9eu0bNHD7y8vJwQNTzyyCOUKFGCuXPnEh0dnW6b/fv3c+7cOdq3b4+3t7dT4hIREREREedRUu5GUi/7lQC8AgwF1gKHgBfx8PBi6dKv6dChA1euXEn3PEeOHOGbb76hYsWKPPvss3aN0Xa+L774wq7nzYjFYuGTT6zLoI0ZMyZbx3h4ePDDDz9QqmhR3nnnHTa2b8+S998HoM+KFbkrY58LPj4+/OMf/+DGjRuEhoam20ZD10VERERECjaXJuXXrl1j9OjRtGnThpCQELZt25Zuu88++4w+ffrQtm1bBg8eTER2J00XMEYjTJjwNzAduA9rz7iVwVAHg+FT5sw5yxNPPMEff/xBz549iYmJSXOet99+G5PJxPjx4ylcuLBdY2zUqBFBQUFs3bqV7du32/Xc6VmzZg2RkZF069aN+vXrZ/u48ps3M+vGDczAMCAMKAR0vXo1d2Xsc+mpp54CMh7C/vvvvwNKykVERERECiqXJuWTJk2iZMmSrF69mtGjR/PGG28QFRWVpp2vry/Tpk0jPDycV155hXHjxnH27FkXROx6x45NBWIoWnQ04J+8PSAAQkNh8GAfZsyYQc+ePdmyZQuDBg0iMT4ewsNh3jz2fPst8+fPp3bt2slzmu3t+eefB+DLL790yPltLBYL7yf1cL/88svZPzCpjH1P4DngJNayeR2BIrY2OSljnwcPPPAALVq0YMOGDRw+fDjVvpiYGNavX0+VKlWoVauWw2MRERERERHn83TVG8fExBAeHs6SJUvw8fEhODiYGjVqsG7dOnr37p2q7ciRd3qEmzdvTvXq1Tl8+DAVK1ZM99wJCQkkJCSk2ubp6UmhQoXsfyFOFBUVxeeff46/vz9HjjzLoUNmLl6EsmUhMNDak242W4dnz5s3j86dO/PLL78wskwZZphMGAwG3oiNBeDd3r3x8PDAbDbbPc7evXtToUIF5s+fz6RJkyhdunSez2mLM2W806dPJzw8nNatW9O+ffvsX8uGDXD9Ovj7M9FiITw2lv1mM728vTHb5pNfu2Zt54Sib08++STbtm3j22+/5aOPPkreHh4eTnx8PF26dMFisWS5dJqkld59I5IV3TeSG7pvJDd030hu6L5xLx4eWfeDuywpP3XqFL6+vpQtWzZ5W82aNfnrr78yPe7GjRscO3aM6tWrZ9hm1qxZfPPNN6m2DRgwgIEDB+YtaBf76quviIqKYsSIEURH36By5RtUrmzdl97S4F988QUDBw7ku7/+wu/55wkMDOS3IUNo0KABzf/1rzRLgtnToEGD+Oyzz5g8eXKWFdFz4vTp08mvr7zyCj4+Prz77rucOnUq+yepXBn27Uv+8otTp1i0aBHtnnmGkz4+qds68N/IJjAwEF9fX77//nuefvrp5IdHtnnmTZo0cej36l5gu29EckL3jeSG7hvJDd03khu6b9xDtWrVsmxjsLio+2337t28/fbbLFu2LHnbl19+SVRUFG+++Wa6x5jNZl599VWKFi3K+PHjMzx3Qewpj4mJoXr16ty8eZO//vor1cOMdJlM0KABJ86coU1sLOctFsoZDFywWFjh40MXLy+oUAEiI+8s5G1HFy5coGrVqpQrV46jR4/i6Zn585+lS+G11yBl0fgKFWDSJOjd2/q9P336NJUqVQKgU6dOrFu3jilTpiQPl8+2iAjo2TPrdr/84pSecoDhw4cza9YsFi5cSL9+/QBo0KABf/75J5cuXaJ48eJOiaOgSXnfZOcppQjovpHc0X0juaH7RnJD9417ydc95YULF06zDFR0dDS+vr4ZHjNx4kRu3brFhx9+mOm5CxUq5NYJeHq+++47Ll++zHPPPUf5O2ujZWz9evjzT6oDK4C2wAWLhfZA17g4DHFx8McfsHEjtGtn93grVKjAwIEDmTNnDr/++iuPPPJIhm3Dwqy11e5+PHTkCISEWOfK9+1r3ebh4cEXX3zBunXraNeuHc8//3zOfxm1aQPFi8PZs2nfFMBgsE7Sb9MGnPSLbsSIEcyaNYuZM2cyYMAATp8+zaFDhwgMDOS+++5zSgwFmYeHh/7TkhzTfSO5oftGckP3jeSG7puCw2XfxcqVKxMTE8OlS5eSt2U2LH3q1KkcPnyYTz/9tMAl3FlJSEjg448/xtPTk1dffTV7B50/n/xpI2AZ0A34HDBk0M7ennvuOQA+//zzDNsk1VxLNze2bRsz2oIp3Fpx/8+5c3n99dfx9/fnu+++y90vIqMRpk61fm4wpN5n+3rKFIeMIMhIy5YtqV+/Pr///jsnT55U1XURERERkXuEy5JyX19fgoODmT59OnFxcURERHD06FGCg4PTtP3222/ZsGED06ZNw8/PzwXRutasWbM4c+YM//jHP6hsm0Selbt609ti7TFPs2hYdnrdc6lly5Y0a9aMtWvXsnv37nTbRETcPR8+HPgGuA5YE/PTZwxs6vMhJpOJJx97jNjYWCYPG5at+RkZsnXB310s0FbGPiQk9+fOBYPBwPDhw7FYLMwaP56VSTURunbq5NQ4RERERETEuVw63uH111/n8uXLdOzYkc8++4wPPviAYsWKsWLFilRF2b7++mvOnDlDr169CAoKIigoiBUrVrgwcuewWCx88cUXPPvss3h6evL6669n/+CgIGuCeXdPsI3BAJUqOXTOtMFgYOzYsQC8++676bZJ3VH/N9ALGAFUAJ4AtgIWLlKOmTNnstlspjMwYvr0vK8lHhICJ07A2rUwd6719fhxpyfkNo8WL04h4LsffmD11q2UAJoPHOi0NdNFRERERMT5XFboTTIXHx/PqFGj+O677yhatChz5syhZ3aKk6Vkm6wNqceH2xJ1J/QIm81mGjZsyMGDB9mzZw+NGjVKtT88HNq3t301FvgYaAccAi4mbW/E04Vi+IGT+CQksB+oZJv3ffy4U4eZO0zS92qwxcKCpE0DgIVO/F4VRGazmZMnT1KlShXNuZJs030juaH7RnJD943khu6bgkffxXzo3LlzBAcH891333H//fezdevWnCfkkC+GaHt4eDBu3DgAJkyYkGa/rUMfLgBfAPcBPwOngJ+AjsBevkk4QkJCAp96e1MJksa1n7aOf3d3KSbWD0+xuSukmFg/xtpOREREREQKFCXl+cyWLVto3rw5W7dupUePHmzdupU6derk/oT5YIj2gAEDqFOnDosWLSIyMjLVvjs11yYCscCrQDGgEAZCMPA7n9OZsV5ejB49msfvXlrNgYXqnCbFxPoOgG2mfBfb/oL0AEJERERERFJRUp6PfPfddwQHB3P+/Hn+/e9/s3TpUooVK5b3ExuN1mXPhgyxvjp5uLfRaMy0t7xFizN4ev4XD48ywJ01xwM4Qyj9eY5VfOjtzejRozHcPUfegYXqnCbFgwUPIAzrWIFKmbQTEREREZGCwWXrlEtqJpOJH374AS8vL+bMmUN/21zwAmLQoEG88847hIaGcuDAAerXv1MH/v333ycxMYHJk1+nWTM/zp+H8mVMBD0WhPHcaUiv6oFtTrkDC9U5zV0PFhonfWTVTkRERERE3J96yvMJo9HITz/9xObNmwtcQg7W63vrrbewWCy89957yduPHz/Ot99+S4UKFRg16l93OvQ7GjFO+8zaKJ+sJe4w+aBSvoiIiIiIuIaS8nykTJkyNGzY0NVhOMyQIUOoWbMmCxYs4NAPP8C8eUwYNYrExET+/e9/U7hw4dQH5INCdU5xZ2J9wX8AISIiIiIiqSgpF6fx9PTkrW7dsFgsvP/44/w5dCg//PYbVYxGnipZMv2DbIXqfvnF+vUvv7h0LXGHuVceQIiIiIiISCqaUy7OExbGsC++YAIwDzgNmIFxJhPeQ4aAl1f6yafRaB26ffKk9bWgrscYEgJ9+lirrJ8/b51DHhSkHnIRERERkQJMSbk4R9Ja3J7Av4EngfVATeCftjZjxliT0ns5CbVVyhcRERERkXtCAe1ylHwnxVrcj3JnLe7xgBdoLW4REREREbknqadcnCPFGttewBxgLTAkk3YiIiIiIiIFnZJycY671th+KOkjq3YiIiIiIiIFmYavi3NoLW4REREREZE0lJSLc2gtbhERERERkTSUlIvzaC1uERERERGRVDSnXJxLa3GLiIiIiIgkU1Iuzqe1uEVERERERAANXxcRERERERFxGSXlIiIiIiIiIi6ipFxERERERETERZSUi4iIiIiIiLiIknIRERERERERF1FSLiIiIiIiIuIiSspFREREREREXERJuYiIiIiIiIiLKCkXERERERERcREl5SIiIiIiIiIuoqRcRERERERExEWUlIuIiIiIiIi4iJJyERERERERERdRUi4iIiIiIiLiIgaLxWJxdRAiIiIiIiIi9yL1lIuIiIiIiIi4iJJyERERERERERdRUi4iIiIiIiLiIkrKRURERERERFxESbmIiIiIiIiIiygpFxEREREREXERJeUiIiIiIiIiLqKkXERERERERMRFlJSLiIiIiIiIuIiSchEREREREREXUVIuLhEaGsqwYcNo2bIl06dPT95usViYPn06PXr0oF27drz33nvcvn07ef+IESMIDAwkKCiIoKAgXnjhhVTn/f777+nUqRMdOnRg6tSpWCwWp12TOJYj7plly5YxbNgwgoOD6dmzJ99//70zL0mcwFG/awASExMZNGgQffv2dcaliBM56r7Zv38/jz/+OEFBQfTo0YM1a9Y47ZrE8Rxx3yQkJPDee+/RqVMnOnbsyLhx44iOjnbqdYlj5fa+Afjhhx/o0aMHbdu2ZejQoanuDf1N7F6UlItLlCpVihEjRtChQ4dU25ctW8aaNWuYNWsWv/76K1euXOGbb75J1eatt94iIiKCiIgIpk2blrx9w4YN/PTTT3z//fcsXLiQTZs2sWTJEqdcjzieI+6Z+Ph4xo4dy5o1a5gxYwbLli3jt99+c8r1iHM44r6xWbhwIf7+/g6NX1zDEffN33//zWuvvcaIESNYu3Ytc+bM4f7773fK9YhzOOK+WbhwIYcPHyY0NJSlS5dy9epVZs2a5ZTrEefI7X2zcOFCNm/ezMyZM1m3bh3vvPMOXl5egP4mdkdKysUl2rVrR3BwMEWKFEm1fcOGDYSEhFCmTBn8/Px47LHH+OWXX7J1zuXLl/PII48QEBBAqVKlePTRR1m+fLkjwhcXcMQ9079/fxo1aoSnpycVKlSgQ4cO7Nu3zxHhi4s44r4BuHLlCosXL+aJJ56wd8iSDzjivpk7dy49e/YkMDAQT09PihcvTkBAgCPCFxdxxH1z7tw5HnroIYoXL46fnx/t2rXjr7/+ckT44iK5uW9MJhPfffcdb731FuXKlcNgMFCrVi0KFSoE6G9id6SkXPKdu4fXXLp0iVu3biV//emnn9KpUydGjRrFkSNHkrcfP36cWrVqJX9ds2ZNjh075viAxeVye8/cbdeuXVSvXt1hcUr+kpf75vPPP+eJJ57Ax8fHKbFK/pHb++bAgQMADBo0iK5duzJ+/PhUx0nBltv7pmfPnuzdu5erV69y69Yt1q5dS6tWrZwWt7hWRvfNpUuXiIuLY/Xq1XTp0oWQkBAWL16c3E5/E7sfJeWSrwQGBrJo0SLOnz/PzZs3k4doxcbGAvDCCy+wdOlSfvnlF1q2bMkLL7yQPH8mJiYGPz+/5HP5+fklHycFV17umZRmz57NjRs36Nmzp1PjF9fIy32zb98+Tp06Rffu3V0Wv7hGXu6by5cvs3z5cj766CMWL15MTEwMn376qcuuRZwnL/dNpUqVKF68OF27dqV9+/YYjUb69evnsmsR58nsvrEl56dOnWLp0qVMmjSJL7/8kt27dwP6m9gdKSmXfKV379507tyZESNGMGjQIFq0aIGnpyf33XcfAA0aNMDX1xcfHx8ee+wxfH19iYyMBMDX1zdVshUdHU3hwoVdch3iPHm5Z2xWrFjBvHnzmDJlino+7xG5vW/MZjOTJ0/m5ZdfxmAwuPgqxNny8vvG29ubXr16UaVKFXx9fXnyySfZuHGjKy9HnCQv983EiRPx9PQkPDyctWvX4ufnx9SpU115OeIkmd033t7eADz99NP4+PhQq1YtunTpkvw7RX8Tux9PVwcgkpKHhwcjR45k5MiRAGzZsoU6depgNBozbG9TrVo1jh49SnBwMADHjh2jRo0ajg9aXCov9wxAeHg4U6ZM4auvvqJixYoOj1fyh9zeN9HR0Rw+fJiXXnoJgNu3bxMdHU3Xrl1ZtGiRCr8VcHn5fVOjRo1UD3L0UOfekZf75siRI7z44ovJvZ49evTgq6++cnzQ4nKZ3TdVqlTBy8srw98p+pvY/ainXFwiMTGR+Ph4zGYzJpOJ+Ph4TCYT169f58yZM1gsFo4dO8Znn33G008/DcDNmzfZsmULCQkJ3L59mzlz5nDjxg0aNGgAwMMPP0xYWBhnzpzhypUrzJkzh4cfftiVlyl25Ih7Ztu2bUyYMIFPPvlE/1kVUPa+b/z9/Vm+fDlz5sxhzpw5vPXWW5QtW5Y5c+akGioo7s0Rv2969uzJsmXLOHPmDHFxcXz//fe0adPGlZcpduaI+6Zu3bosX76cuLg4YmNjWb58OTVr1nTlZYqd5ea+KVy4MB07dmTmzJkkJCRw/PhxVq1aRevWrQH9TeyODBYtWicuMH369DTLgYwfP54GDRrw4osvcvnyZUqXLs1TTz1Fr169ALh27RovvPACJ0+exNPTk9q1azNmzBjq1KmTfI5Zs2Yxe/ZszGYzffv25YUXXlBvRAHhiHtm5MiR7NmzJ7laKUD37t158803nXdh4lCO+l1js2PHDt577z1+/vlnZ1yOOImj7pv58+cza9YsEhMTCQwMZOzYsWkqLov7csR9c/XqVSZOnMjOnTsBaNKkCa+//jqlSpVy7sWJw+TmvgHrA513332XrVu3Urx4cR5//HFCQkKS9+tvYveipFxERERERETERTR8XURERERERMRFlJSLiIiIiIiIuIiSchEREREREREXUVIuIiIiIiIi4iJKykVERERERERcREm5iIiIiIiIiIsoKRcRERERERFxESXlIiIiIiIiIi6ipFxEREQcpnnz5jRv3pxly5a5OhQREZF8SUm5iIiImxsxYkRy8jtkyJBU+65fv07r1q2T93/++ed2f/9ly5Yln19ERERyRkm5iIhIAXLkyBF27dqV/PXPP/9MfHy8CyMSERGRzCgpFxERKSA8PT0BWLBgAQAmk4nQ0NDk7SlFRUUxadIkevToQcuWLenSpQvjxo3jwoULyW2mT59O8+bN6dWrF6tXr6Zfv360adOGp59+mhMnTgDwn//8h3feeSf5GFuP+fTp01O9361bt3jnnXcIDg6me/fufPvtt/a+fBEREbekpFxERKSAqF27NhUrViQ8PJyLFy+yfv16Lly4QMeOHVO1i4+PZ8SIEfz000/8/fffVKlShejoaFasWMETTzzBtWvXUrW/dOkSb731FgaDgfj4eHbv3s27774LQEBAABUrVkxu26BBAxo0aEDZsmVTneOLL75g8+bNeHl5cfnyZb7++mu2bNnioH8JERER96GkXEREpIDw8PBgwIAByT3kth7zQYMGpWq3cuVKjh07BsCkSZNYuHAhM2fOxMPDg8uXL7Nw4cJU7U0mEx999BGhoaHJc9b37dtHXFwcw4cPZ/jw4cltv//+e77//nv69u2b6hz3338/y5YtS9Vzv337drtev4iIiDtSUi4iIlKA9OnTh8KFC7Nw4UJ27NhB3bp1eeCBB1K1OXjwIAA+Pj60a9cOgDp16lClSpVU+238/f1p27YtANWrV0/efnePemY6deqEl5cXxYsX57777gPg6tWrObs4ERGRAkhJuYiISAFSpEgRunfvTnR0NJC2lzy357QxGo3Jn1ssljydIyfHi4iIFFRKykVERAqYgQMHAlCiRAm6dOmSZn+9evUAiIuLIzw8HIDDhw9z8uTJVPuzy8fHJ/nz2NjY3IQsIiJyz0pbjlVERETcWs2aNVmzZg1Go5FChQql2d+1a1fmzJnDsWPHeO2116hSpQpnz57FbDZTunTp5KQ+u6pWrZr8+YABAyhVqhRjxoyhcePGebwSERGRgk895SIiIgVQsWLF8Pf3T3eft7c3M2bMSE6gT548ia+vL927d2fWrFmUKFEiR+9Vq1Ythg8fTsmSJblw4QL79+/n5s2b9rgMERGRAs9g0YQuEREREREREZdQT7mIiIiIiIiIiygpFxEREREREXERJeUiIiIiIiIiLqKkXERERERERMRFlJSLiIiIiIiIuIiSchEREREREREXUVIuIiIiIiIi4iJKykVERERERERcREm5iIiIiIiIiIsoKRcRERERERFxESXlIiIiIiIiIi7y/02vt4BaEllIAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 1200x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(nrows = 1, ncols = 1, figsize = (12, 6))\n",
    "lmins = airpass_scaled_data[airpass_scaled_data['LMin'] == 1]\n",
    "lmaxs = airpass_scaled_data[airpass_scaled_data['LMax'] == 1]\n",
    "\n",
    "axes.plot(airpass_scaled_data['Month'], airpass_scaled_data['# Passengers'], label = '# Passengers')\n",
    "axes.set_xlabel(\"Month\")\n",
    "axes.set_ylabel(\"# Passengers\")\n",
    "axes.set_title(\"Scaled # Passengers over the months\")\n",
    "axes.scatter(lmins['Month'], lmins['# Passengers'], color='red', label = 'Local Min')\n",
    "axes.scatter(lmaxs['Month'], lmaxs['# Passengers'], color='blue', label = 'Local Max')\n",
    "axes.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Fit a model on the data\n",
    "- Invert the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# ## Format to TimeSeries\n",
    "# past = 24\n",
    "# future = 24\n",
    "# airpass_ts = TimeSeries.from_dataframe(airpass_scaled_data, \"Month\", \"# Passengers\")\n",
    "# airpass_ts = airpass_ts.astype('float32')\n",
    "\n",
    "# model = NHiTSModel(input_chunk_length = past, output_chunk_length = future)\n",
    "# model.fit(airpass_ts, verbose = 0)\n",
    "\n",
    "# ## Predict on self\n",
    "# pred = model.predict(n = len(airpass_scaled_data['Month']), verbose = 0)\n",
    "\n",
    "# ## Plot predict v train\n",
    "# results_df = pd.DataFrame(scaler.inverse_transform(pred.values())).rename(columns = {0:\"Predicted\"})\n",
    "# results_df = pd.concat([airpass_scaled_data['Month'], results_df, \\\n",
    "#                         pd.DataFrame(scaler.inverse_transform(airpass_ts.values()))], axis = 1)\\\n",
    "#                         .rename(columns = {0:\"Actual\"})\n",
    "\n",
    "# plt.plot(results_df['Month'], results_df['Predicted'], label = 'Predicted')\n",
    "# plt.plot(results_df['Month'], results_df['Actual'], label = 'Actual')\n",
    "# plt.legend()\n",
    "# plt.show()\n",
    "\n",
    "# mse = round(sum((results_df['Predicted'] - results_df['Actual'])**2)/len(results_df['Predicted']), 3)\n",
    "# mape = round(sum(abs((results_df['Predicted'] - results_df['Actual'])))/len(results_df['Predicted']), 3)\n",
    "\n",
    "# print(\"MSE =\", mse, \"\\nMAPE =\", mape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Air dataset\n",
    "- https://www.opendatanetwork.com/dataset/datahub.transportation.gov/xgub-n9bw"
   ]
  },
  {
   "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>Year</th>\n",
       "      <th>Month</th>\n",
       "      <th>usg_apt</th>\n",
       "      <th>fg_apt</th>\n",
       "      <th>carrier</th>\n",
       "      <th>carriergroup</th>\n",
       "      <th>Total</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2023</td>\n",
       "      <td>1</td>\n",
       "      <td>MIA</td>\n",
       "      <td>CLO</td>\n",
       "      <td>AA</td>\n",
       "      <td>1</td>\n",
       "      <td>18315</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2023</td>\n",
       "      <td>1</td>\n",
       "      <td>IAD</td>\n",
       "      <td>DND</td>\n",
       "      <td>AOQ</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2023</td>\n",
       "      <td>1</td>\n",
       "      <td>IAD</td>\n",
       "      <td>DEL</td>\n",
       "      <td>AI</td>\n",
       "      <td>0</td>\n",
       "      <td>5385</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2023</td>\n",
       "      <td>1</td>\n",
       "      <td>IAD</td>\n",
       "      <td>CPT</td>\n",
       "      <td>UA</td>\n",
       "      <td>1</td>\n",
       "      <td>5662</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2023</td>\n",
       "      <td>1</td>\n",
       "      <td>IAD</td>\n",
       "      <td>CPH</td>\n",
       "      <td>SK</td>\n",
       "      <td>0</td>\n",
       "      <td>5190</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32448</th>\n",
       "      <td>2023</td>\n",
       "      <td>12</td>\n",
       "      <td>TEB</td>\n",
       "      <td>LBG</td>\n",
       "      <td>VJT</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32449</th>\n",
       "      <td>2023</td>\n",
       "      <td>12</td>\n",
       "      <td>TEB</td>\n",
       "      <td>LBG</td>\n",
       "      <td>1HQ</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32450</th>\n",
       "      <td>2023</td>\n",
       "      <td>12</td>\n",
       "      <td>TEB</td>\n",
       "      <td>KEF</td>\n",
       "      <td>AOQ</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32451</th>\n",
       "      <td>2023</td>\n",
       "      <td>12</td>\n",
       "      <td>TEB</td>\n",
       "      <td>LTN</td>\n",
       "      <td>HUQ</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32452</th>\n",
       "      <td>2023</td>\n",
       "      <td>12</td>\n",
       "      <td>FXE</td>\n",
       "      <td>MHH</td>\n",
       "      <td>IWQ</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>32453 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       Year  Month usg_apt fg_apt carrier  carriergroup  Total\n",
       "0      2023      1     MIA    CLO      AA             1  18315\n",
       "1      2023      1     IAD    DND     AOQ             0      3\n",
       "2      2023      1     IAD    DEL      AI             0   5385\n",
       "3      2023      1     IAD    CPT      UA             1   5662\n",
       "4      2023      1     IAD    CPH      SK             0   5190\n",
       "...     ...    ...     ...    ...     ...           ...    ...\n",
       "32448  2023     12     TEB    LBG     VJT             0      7\n",
       "32449  2023     12     TEB    LBG     1HQ             0      7\n",
       "32450  2023     12     TEB    KEF     AOQ             0      7\n",
       "32451  2023     12     TEB    LTN     HUQ             0      2\n",
       "32452  2023     12     FXE    MHH     IWQ             0      1\n",
       "\n",
       "[32453 rows x 7 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "carrier_raw = pd.read_csv(\"carrier_passengers.csv\")\n",
    "carrier_df = carrier_raw[carrier_raw['Year'] == 2023].sort_values(by = 'data_dte').reset_index(drop = True)\n",
    "carrier_df = carrier_df.drop(columns=[\"data_dte\", \"usg_apt_id\", \"usg_wac\", \"fg_apt_id\", \"fg_wac\", \"airlineid\", \"type\", \"Scheduled\", \"Charter\"])\n",
    "\n",
    "carrier_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Numeric encoding\n",
    "- Tried one-hot and PCA, didn't work well\n",
    "    - Create the encoding for the flight and carrier, and use pca to determine which columns are most useful in determining the total\n",
    "- Scale the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>Year</th>\n",
       "      <th>Month</th>\n",
       "      <th>carriergroup</th>\n",
       "      <th>Total</th>\n",
       "      <th>usg_apt_encoded</th>\n",
       "      <th>fg_apt_encoded</th>\n",
       "      <th>carrier_encoded</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2023</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>18315</td>\n",
       "      <td>214</td>\n",
       "      <td>127</td>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2023</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>157</td>\n",
       "      <td>164</td>\n",
       "      <td>56</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2023</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>5385</td>\n",
       "      <td>157</td>\n",
       "      <td>160</td>\n",
       "      <td>52</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2023</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>5662</td>\n",
       "      <td>157</td>\n",
       "      <td>140</td>\n",
       "      <td>184</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2023</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>5190</td>\n",
       "      <td>157</td>\n",
       "      <td>139</td>\n",
       "      <td>170</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Year  Month  carriergroup  Total  usg_apt_encoded  fg_apt_encoded  \\\n",
       "0  2023      1             1  18315              214             127   \n",
       "1  2023      1             0      3              157             164   \n",
       "2  2023      1             0   5385              157             160   \n",
       "3  2023      1             1   5662              157             140   \n",
       "4  2023      1             0   5190              157             139   \n",
       "\n",
       "   carrier_encoded  \n",
       "0               48  \n",
       "1               56  \n",
       "2               52  \n",
       "3              184  \n",
       "4              170  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "columns_to_encode = ['usg_apt', 'fg_apt', 'carrier']\n",
    "temp_df = carrier_df.copy()\n",
    "temp_df = temp_df[columns_to_encode]\n",
    "\n",
    "model_df = carrier_df.copy()\n",
    "\n",
    "## Construct encoders\n",
    "le_usg_apt = LabelEncoder()\n",
    "le_fg_apt = LabelEncoder()\n",
    "le_carrier = LabelEncoder()\n",
    "\n",
    "## Encode the data\n",
    "model_df['usg_apt_encoded'] = le_usg_apt.fit_transform(model_df['usg_apt'])\n",
    "model_df['fg_apt_encoded'] = le_fg_apt.fit_transform(model_df['fg_apt'])\n",
    "model_df['carrier_encoded'] = le_carrier.fit_transform(model_df['carrier'])\n",
    "model_df = model_df.drop(columns=['usg_apt', 'fg_apt', 'carrier']).rename(columns = {0:\"Month\", 1:\"carriergroup\", 2:\"Total\"})\n",
    "\n",
    "model_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Fit the model on average passenger per month"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE = 0.5287902181638201\n",
      "R^2 = 0.7203809051742596\n"
     ]
    },
    {
     "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>Predicted</th>\n",
       "      <th>Predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.133549</td>\n",
       "      <td>0.559985</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.152497</td>\n",
       "      <td>0.407262</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-1.286046</td>\n",
       "      <td>-0.967246</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Predicted  Predicted\n",
       "0   0.133549   0.559985\n",
       "1   1.152497   0.407262\n",
       "2  -1.286046  -0.967246"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## , 'carriergroup', 'usg_apt_encoded', 'fg_apt_encoded', 'carrier_encoded'\n",
    "scaler = StandardScaler()\n",
    "X = model_df[['Month']].drop_duplicates().reset_index(drop = True)\n",
    "X['Year'] = 2023\n",
    "model_df['Avg'] = model_df.groupby(by = 'Month')['Total'].transform(\"mean\")\n",
    "y = model_df[['Avg']].drop_duplicates().reset_index(drop = True)\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n",
    "\n",
    "X_train = scaler.fit_transform(X_train)\n",
    "X_test = scaler.fit_transform(X_test)\n",
    "y_train = scaler.fit_transform(y_train)\n",
    "y_test = pd.DataFrame(scaler.fit_transform(y_test))\n",
    "\n",
    "model = LinearRegression()\n",
    "model.fit(X_train, y_train)\n",
    "\n",
    "y_pred = model.predict(X_test).flatten()\n",
    "results_df = pd.DataFrame(y_pred).reset_index(drop = True)\n",
    "results_df = pd.concat([y_test, results_df], axis = 1).rename(columns = {0:\"Predicted\", \"Total\": \"Actual\"})\n",
    "\n",
    "rmse = np.sqrt(mean_squared_error(y_test, y_pred))\n",
    "r2 = r2_score(y_test, y_pred)\n",
    "\n",
    "print(f\"RMSE = {rmse}\", f\"R^2 = {r2}\", sep = '\\n')\n",
    "results_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Forecast future data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>Month</th>\n",
       "      <th>Year</th>\n",
       "      <th>Predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>2024</td>\n",
       "      <td>0.686782</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>2024</td>\n",
       "      <td>1.373565</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>2024</td>\n",
       "      <td>2.060347</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>2024</td>\n",
       "      <td>2.747130</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>2024</td>\n",
       "      <td>3.433912</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>6</td>\n",
       "      <td>2024</td>\n",
       "      <td>4.120695</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>7</td>\n",
       "      <td>2024</td>\n",
       "      <td>4.807477</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>8</td>\n",
       "      <td>2024</td>\n",
       "      <td>5.494260</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>9</td>\n",
       "      <td>2024</td>\n",
       "      <td>6.181042</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>10</td>\n",
       "      <td>2024</td>\n",
       "      <td>6.867825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>11</td>\n",
       "      <td>2024</td>\n",
       "      <td>7.554607</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>12</td>\n",
       "      <td>2024</td>\n",
       "      <td>8.241390</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Month  Year  Predicted\n",
       "0       1  2024   0.686782\n",
       "1       2  2024   1.373565\n",
       "2       3  2024   2.060347\n",
       "3       4  2024   2.747130\n",
       "4       5  2024   3.433912\n",
       "5       6  2024   4.120695\n",
       "6       7  2024   4.807477\n",
       "7       8  2024   5.494260\n",
       "8       9  2024   6.181042\n",
       "9      10  2024   6.867825\n",
       "10     11  2024   7.554607\n",
       "11     12  2024   8.241390"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "future_data = X.copy()\n",
    "future_data['Year'] = 2024\n",
    "\n",
    "forecast_x = pd.DataFrame(model.predict(future_data).flatten()).rename(columns = {0:\"Predicted\"})\n",
    "forecast_x = pd.concat([future_data, forecast_x], axis = 1)\n",
    "forecast_x"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Use the API for modelling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>count</th>\n",
       "      <th>node_id</th>\n",
       "      <th>noise</th>\n",
       "      <th>timestamp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>Wed, 26 Jun 2024 17:57:14 GMT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>Wed, 26 Jun 2024 17:57:58 GMT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>Wed, 26 Jun 2024 17:58:41 GMT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>Wed, 26 Jun 2024 17:59:25 GMT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>Wed, 26 Jun 2024 18:00:09 GMT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>164</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>Wed, 26 Jun 2024 22:29:56 GMT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>165</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>Wed, 26 Jun 2024 22:30:20 GMT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>166</th>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>Wed, 26 Jun 2024 22:36:27 GMT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>167</th>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>Wed, 26 Jun 2024 22:42:25 GMT</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>168</th>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>False</td>\n",
       "      <td>Wed, 26 Jun 2024 22:42:57 GMT</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>169 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     count  node_id  noise                      timestamp\n",
       "0        1        1  False  Wed, 26 Jun 2024 17:57:14 GMT\n",
       "1        2        1  False  Wed, 26 Jun 2024 17:57:58 GMT\n",
       "2        0        1  False  Wed, 26 Jun 2024 17:58:41 GMT\n",
       "3        0        1  False  Wed, 26 Jun 2024 17:59:25 GMT\n",
       "4        1        1  False  Wed, 26 Jun 2024 18:00:09 GMT\n",
       "..     ...      ...    ...                            ...\n",
       "164      0        1  False  Wed, 26 Jun 2024 22:29:56 GMT\n",
       "165      1        1  False  Wed, 26 Jun 2024 22:30:20 GMT\n",
       "166      3        1  False  Wed, 26 Jun 2024 22:36:27 GMT\n",
       "167      6        1  False  Wed, 26 Jun 2024 22:42:25 GMT\n",
       "168      6        1  False  Wed, 26 Jun 2024 22:42:57 GMT\n",
       "\n",
       "[169 rows x 4 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from datetime import datetime, timedelta\n",
    "from darts import TimeSeries\n",
    "from src.forecasting_engine import get_data, write_forecast\n",
    "import pandas as pd\n",
    "\n",
    "## Timestamp ('%Y-%m-%dT%H:%M:%S'), node_id, noise (binary = loud or not), count\n",
    "## Data to be collected every 15 mins\n",
    "## Want to forecast for 6 hours = 24 future steps\n",
    "\n",
    "end = datetime.now()\n",
    "start = (end - timedelta(hours=24)).strftime('%Y-%m-%dT%H:%M:%S')\n",
    "end = end.strftime('%Y-%m-%dT%H:%M:%S')\n",
    "\n",
    "data_raw = get_data(start = start, end = end) ## only use the last 24 hours as input\n",
    "n_future = 24\n",
    "\n",
    "## Get forecast dates and times\n",
    "end_time = datetime.now() + timedelta(hours = int(n_future/4))\n",
    "future_times = []\n",
    "current_time = datetime.now()\n",
    "while current_time <= end_time:\n",
    "    future_times.append(current_time.strftime('%Y-%m-%dT%H:%M:%S'))\n",
    "    current_time += timedelta(minutes=15)\n",
    "data_raw"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Preprocess data and run model on new data\n",
    "- We want to remove the node_id and, convert to timeseries\n",
    "- Train the model on the new data\n",
    "- Forecast future data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "https://occupi-rp-hack-24.uc.r.appspot.com\n",
      "[{\"Forecast\":4.1206949265,\"node_id\":1,\"Timestamp\":1719432000000},{\"Forecast\":4.8074774143,\"node_id\":1,\"Timestamp\":1719432900000},{\"Forecast\":5.494259902,\"node_id\":1,\"Timestamp\":1719433800000},{\"Forecast\":6.1810423898,\"node_id\":1,\"Timestamp\":1719434700000},{\"Forecast\":6.8678248775,\"node_id\":1,\"Timestamp\":1719435600000},{\"Forecast\":7.5546073653,\"node_id\":1,\"Timestamp\":1719436500000},{\"Forecast\":8.241389853,\"node_id\":1,\"Timestamp\":1719437400000},{\"Forecast\":0.6867824878,\"node_id\":1,\"Timestamp\":1719438300000},{\"Forecast\":1.3735649755,\"node_id\":1,\"Timestamp\":1719439200000},{\"Forecast\":2.0603474633,\"node_id\":1,\"Timestamp\":1719440100000},{\"Forecast\":2.747129951,\"node_id\":1,\"Timestamp\":1719441000000},{\"Forecast\":3.4339124388,\"node_id\":1,\"Timestamp\":1719441900000}]\n"
     ]
    },
    {
     "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>Forecast</th>\n",
       "      <th>node_id</th>\n",
       "      <th>Timestamp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>4.120695</td>\n",
       "      <td>1</td>\n",
       "      <td>2024-06-26 20:00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4.807477</td>\n",
       "      <td>1</td>\n",
       "      <td>2024-06-26 20:15:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5.494260</td>\n",
       "      <td>1</td>\n",
       "      <td>2024-06-26 20:30:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>6.181042</td>\n",
       "      <td>1</td>\n",
       "      <td>2024-06-26 20:45:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6.867825</td>\n",
       "      <td>1</td>\n",
       "      <td>2024-06-26 21:00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>7.554607</td>\n",
       "      <td>1</td>\n",
       "      <td>2024-06-26 21:15:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>8.241390</td>\n",
       "      <td>1</td>\n",
       "      <td>2024-06-26 21:30:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.686782</td>\n",
       "      <td>1</td>\n",
       "      <td>2024-06-26 21:45:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1.373565</td>\n",
       "      <td>1</td>\n",
       "      <td>2024-06-26 22:00:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>2.060347</td>\n",
       "      <td>1</td>\n",
       "      <td>2024-06-26 22:15:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>2.747130</td>\n",
       "      <td>1</td>\n",
       "      <td>2024-06-26 22:30:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>3.433912</td>\n",
       "      <td>1</td>\n",
       "      <td>2024-06-26 22:45:00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Forecast  node_id           Timestamp\n",
       "0   4.120695        1 2024-06-26 20:00:00\n",
       "1   4.807477        1 2024-06-26 20:15:00\n",
       "2   5.494260        1 2024-06-26 20:30:00\n",
       "3   6.181042        1 2024-06-26 20:45:00\n",
       "4   6.867825        1 2024-06-26 21:00:00\n",
       "5   7.554607        1 2024-06-26 21:15:00\n",
       "6   8.241390        1 2024-06-26 21:30:00\n",
       "7   0.686782        1 2024-06-26 21:45:00\n",
       "8   1.373565        1 2024-06-26 22:00:00\n",
       "9   2.060347        1 2024-06-26 22:15:00\n",
       "10  2.747130        1 2024-06-26 22:30:00\n",
       "11  3.433912        1 2024-06-26 22:45:00"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## Remove node_id\n",
    "processed_data = data_raw.copy()\n",
    "processed_data['Month'] = pd.to_datetime(processed_data['timestamp']).dt.month\n",
    "processed_data['Year'] = pd.to_datetime(processed_data['timestamp']).dt.year\n",
    "processed_data = processed_data[['Month', 'Year']].drop_duplicates()\n",
    "new_data = pd.DataFrame({'Month': [7, 8, 9, 10, 11, 12, 1, 2, 3, 4, 5], 'Year': [2024, 2024, 2024, 2024, 2024, 2024, 2025, 2025, 2025, 2025, 2025, ]})\n",
    "\n",
    "processed_data = pd.concat([processed_data, new_data], axis = 0).reset_index(drop = True)\n",
    "\n",
    "## Predict model to our data\n",
    "preds = list(model.predict(processed_data).flatten())\n",
    "forecast_df = pd.DataFrame(preds).rename(columns = {0:\"Forecast\"})\n",
    "start_time = datetime(2024, 6, 26, 20, 0, 0) \n",
    "end_time = datetime(2024, 6, 26, 23, 0, 0)  \n",
    "\n",
    "# Generate timestamps in 15-minute intervals\n",
    "timestamps = []\n",
    "current_time = start_time\n",
    "while current_time < end_time:\n",
    "    timestamps.append(current_time)\n",
    "    current_time += timedelta(minutes=15)\n",
    "\n",
    "forecast_df['node_id'] = 1\n",
    "forecast_df['Timestamp'] = timestamps\n",
    "\n",
    "write_forecast(forecast_df, node_id = 1)\n",
    "forecast_df"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.10.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}