ML_solution.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "a429ea62-9d9c-4e14-b2a1-5c62a7665d2a",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "[notice] A new release of pip is available: 23.2.1 -> 23.3.1\n",
      "[notice] To update, run: python.exe -m pip install --upgrade pip\n"
     ]
    }
   ],
   "source": [
    "!pip install pandas numpy tqdm scikit-learn catboost shap >> None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "8ef781b4-ef34-4c46-ad27-8561d3e25aea",
   "metadata": {},
   "outputs": [],
   "source": [
    "import shap\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "from tqdm.notebook import tqdm\n",
    "\n",
    "from sklearn.model_selection import KFold, StratifiedKFold, TimeSeriesSplit, train_test_split\n",
    "from catboost import CatBoostClassifier, CatBoostRegressor, Pool\n",
    "\n",
    "from sklearn.metrics import roc_auc_score, f1_score\n",
    "\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3fef580f-90e2-4ac9-bf25-f0e66f69b8cb",
   "metadata": {},
   "source": [
    "# Load data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "616c98e1-eb8c-487a-a65e-2b28f2b7179b",
   "metadata": {},
   "source": [
    "загружаем данные и убираем лишние пробелы до и после слова"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "17154996-f14a-4d29-ac98-cb1f0d42b21e",
   "metadata": {},
   "outputs": [],
   "source": [
    "events_df = pd.read_csv(\"Данные/Происшествия.csv\")\n",
    "events_df.replace('^\\s+', '', regex=True, inplace=True) \n",
    "events_df.replace('\\s+$', '', regex=True, inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4752cb6d-c252-4fa0-8528-b4529e2d26a3",
   "metadata": {},
   "source": [
    "приводим названия МО в таблице Происшествия к единому формату "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "3c8a13f8-777c-4269-bb03-7854ec188f2e",
   "metadata": {},
   "outputs": [],
   "source": [
    "events_df.replace(\"й\", \"й\", inplace=True)\n",
    "events_df.rename(columns={'Дата и время  пожара ': 'Местное время'}, inplace=True)\n",
    "events_df = events_df[events_df[\"Район\"] != \"Пермский край\"]\n",
    "\n",
    "administrative_territories_to_replace = {\n",
    "    \"Верещагино ГО\": \"Верещагинский ГО\",\n",
    "    \"Березники ГО\": \"ГО город Березники\",\n",
    "    \"Гремячинск ГО\": \"Губахинский МО\",\n",
    "    \"Губахинский ГО\": \"Губахинский МО\",\n",
    "    \"Кизеловский ГО\": \"ГО город Кизел\",\n",
    "}\n",
    "\n",
    "events_df.replace({\"Район\": administrative_territories_to_replace}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "da5d8d85-5bb7-4ea5-8f7e-e33e93548604",
   "metadata": {},
   "outputs": [],
   "source": [
    "weather_stats_df = pd.read_csv(\"Данные/Данные по метеостанциям.csv\", low_memory=False)\n",
    "weather_stats_df.replace('^\\s+', '', regex=True, inplace=True) \n",
    "weather_stats_df.replace('\\s+$', '', regex=True, inplace=True)\n",
    "weather_stats_df.meteostation = weather_stats_df.meteostation.str.replace(\"й\", \"й\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "69e6cb34-02b4-42d3-9a28-16f0567ddcd3",
   "metadata": {},
   "outputs": [],
   "source": [
    "respective_weather_df = pd.read_csv('Данные/Данные_по_метеостанциям_Соответствие_МО.csv')\n",
    "respective_weather_df.replace('^\\s+', '', regex=True, inplace=True) \n",
    "respective_weather_df.replace('\\s+$', '', regex=True, inplace=True)\n",
    "respective_weather_df.replace(\"й\", \"й\", inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "951f6fd0-4430-4127-b87b-e5d702dfd3c3",
   "metadata": {},
   "outputs": [],
   "source": [
    "predict_df = pd.read_csv('Данные/ОЯ и НЯ.csv')\n",
    "predict_df.rename(columns={'Дата, время начала': 'Местное время'}, inplace=True)\n",
    "predict_df.replace('^\\s+', '', regex=True, inplace=True) \n",
    "predict_df.replace('\\s+$', '', regex=True, inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6f6878cd-54ca-44e2-8513-304c827733d4",
   "metadata": {},
   "source": [
    "# Data preprocessing"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "23c86819-325d-4770-a5df-a7f8bdf18ce9",
   "metadata": {},
   "source": [
    "переводим МО в название метеостанции которая отвечает за этот округ"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "3f294297-20c1-4c98-a9a0-6e603858317e",
   "metadata": {},
   "outputs": [],
   "source": [
    "MO_to_weather_station = dict(respective_weather_df.values)\n",
    "MO_to_weather_station = {x: y.split()[1] for x, y in MO_to_weather_station.items()}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "6d3bba49-5c3c-4c84-933a-cdd914bbef17",
   "metadata": {},
   "outputs": [],
   "source": [
    "events_df['meteostation'] = events_df['Район'].map(MO_to_weather_station)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4e293371-cc7f-4fc0-bc03-a067abb16f7d",
   "metadata": {},
   "source": [
    "приводим названия метеостанций к одному формату"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "72bc5c66-58b5-4076-8d5a-aba4c203bc58",
   "metadata": {},
   "outputs": [],
   "source": [
    "meteostations_to_replace = {\n",
    "    \"Большая\": \"Большая Соснова\",\n",
    "    \"Кочёво\": \"Кочево\",\n",
    "}\n",
    "\n",
    "events_df.replace({\"meteostation\": meteostations_to_replace}, inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "08a7f0b7-613b-4936-83bc-5a111cb43d50",
   "metadata": {},
   "source": [
    "переводим время из строки в таймстампы"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "0a21bc4e-be0f-4cc0-be38-0e3ff0e40b63",
   "metadata": {},
   "outputs": [],
   "source": [
    "weather_stats_df[\"Местное время\"] = pd.to_datetime(weather_stats_df[\"Местное время\"], format=\"%d.%m.%Y %H:%M\")\n",
    "events_df[\"Местное время\"] = pd.to_datetime(events_df[\"Местное время\"], format=\"%Y-%m-%d %H:%M:%S\")\n",
    "predict_df[\"Местное время\"] = pd.to_datetime(predict_df[\"Местное время\"], format=\"%Y-%m-%d %H:%M:%S\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "4c67043d-85bd-4009-b6b5-437aa72552f3",
   "metadata": {},
   "outputs": [],
   "source": [
    "events_df[\"timestamp\"] = pd.to_numeric(events_df[\"Местное время\"])\n",
    "weather_stats_df[\"timestamp\"] = pd.to_numeric(weather_stats_df[\"Местное время\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7080af2a-8a5f-4659-9fca-514d11359b1f",
   "metadata": {},
   "source": [
    "изменяем дату в events_df на максимально приближенную к дате из weather_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "da9f338a-79e0-4069-a883-88528e5b6084",
   "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>Район</th>\n",
       "      <th>Тип</th>\n",
       "      <th>Местное время</th>\n",
       "      <th>meteostation</th>\n",
       "      <th>timestamp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ГО город Березники</td>\n",
       "      <td>Пожар с гибелью 2-х и более человек</td>\n",
       "      <td>2012-12-18 20:39:00</td>\n",
       "      <td>Березники</td>\n",
       "      <td>1357005600000000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Пермский МО</td>\n",
       "      <td>Особо опасные болезни сельскохозяйственных жив...</td>\n",
       "      <td>2013-01-01 00:00:00</td>\n",
       "      <td>Пермь</td>\n",
       "      <td>1357005600000000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Юрлинский МО</td>\n",
       "      <td>Особо опасные болезни сельскохозяйственных жив...</td>\n",
       "      <td>2013-01-01 00:00:00</td>\n",
       "      <td>Кочево</td>\n",
       "      <td>1357005600000000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Карагайский МО</td>\n",
       "      <td>Особо опасные болезни сельскохозяйственных жив...</td>\n",
       "      <td>2013-01-01 00:00:00</td>\n",
       "      <td>Верещагино</td>\n",
       "      <td>1357005600000000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Добрянский ГО</td>\n",
       "      <td>Пожар с гибелью 2-х и более человек</td>\n",
       "      <td>2013-01-01 07:50:00</td>\n",
       "      <td>Добрянка</td>\n",
       "      <td>1357016400000000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3649</th>\n",
       "      <td>Пермский ГО</td>\n",
       "      <td>Аварийные ситуации на ЖД транспорте</td>\n",
       "      <td>2022-12-08 11:50:00</td>\n",
       "      <td>Пермь</td>\n",
       "      <td>1670497200000000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3650</th>\n",
       "      <td>Пермский ГО</td>\n",
       "      <td>Пожары на объектах торговли</td>\n",
       "      <td>2022-12-16 18:01:00</td>\n",
       "      <td>Пермь</td>\n",
       "      <td>1671210000000000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3651</th>\n",
       "      <td>Пермский МО</td>\n",
       "      <td>Крупные ДТП (повреждение 10 и более а/м, ДТП с...</td>\n",
       "      <td>2022-12-25 21:00:00</td>\n",
       "      <td>Пермь</td>\n",
       "      <td>1671998400000000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3652</th>\n",
       "      <td>Пермский ГО</td>\n",
       "      <td>Аварии на магистральных газо-, нефте-, продукт...</td>\n",
       "      <td>2022-12-28 13:03:00</td>\n",
       "      <td>Пермь</td>\n",
       "      <td>1672225200000000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3653</th>\n",
       "      <td>Куединский МО</td>\n",
       "      <td>Пожары на объектах торговли</td>\n",
       "      <td>2022-12-28 17:20:00</td>\n",
       "      <td>Чернушка</td>\n",
       "      <td>1672246800000000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3653 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                   Район                                                Тип  \\\n",
       "0     ГО город Березники                Пожар с гибелью 2-х и более человек   \n",
       "1            Пермский МО  Особо опасные болезни сельскохозяйственных жив...   \n",
       "2           Юрлинский МО  Особо опасные болезни сельскохозяйственных жив...   \n",
       "3         Карагайский МО  Особо опасные болезни сельскохозяйственных жив...   \n",
       "4          Добрянский ГО                Пожар с гибелью 2-х и более человек   \n",
       "...                  ...                                                ...   \n",
       "3649         Пермский ГО                Аварийные ситуации на ЖД транспорте   \n",
       "3650         Пермский ГО                        Пожары на объектах торговли   \n",
       "3651         Пермский МО  Крупные ДТП (повреждение 10 и более а/м, ДТП с...   \n",
       "3652         Пермский ГО  Аварии на магистральных газо-, нефте-, продукт...   \n",
       "3653       Куединский МО                        Пожары на объектах торговли   \n",
       "\n",
       "           Местное время meteostation            timestamp  \n",
       "0    2012-12-18 20:39:00    Березники  1357005600000000000  \n",
       "1    2013-01-01 00:00:00        Пермь  1357005600000000000  \n",
       "2    2013-01-01 00:00:00       Кочево  1357005600000000000  \n",
       "3    2013-01-01 00:00:00   Верещагино  1357005600000000000  \n",
       "4    2013-01-01 07:50:00     Добрянка  1357016400000000000  \n",
       "...                  ...          ...                  ...  \n",
       "3649 2022-12-08 11:50:00        Пермь  1670497200000000000  \n",
       "3650 2022-12-16 18:01:00        Пермь  1671210000000000000  \n",
       "3651 2022-12-25 21:00:00        Пермь  1671998400000000000  \n",
       "3652 2022-12-28 13:03:00        Пермь  1672225200000000000  \n",
       "3653 2022-12-28 17:20:00     Чернушка  1672246800000000000  \n",
       "\n",
       "[3653 rows x 5 columns]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "events_timestamps = list(events_df[\"timestamp\"])\n",
    "weather_stats_timestamps = list(sorted(set(weather_stats_df[\"timestamp\"])))\n",
    "new_events_timestamps = events_timestamps\n",
    "weather_stats_iterator = 0\n",
    "events_iterator = 0\n",
    "while events_iterator != len(events_timestamps):\n",
    "    if weather_stats_iterator == len(weather_stats_timestamps) - 1:\n",
    "        new_events_timestamps[events_iterator] = weather_stats_timestamps[weather_stats_iterator]\n",
    "    elif events_timestamps[events_iterator] > weather_stats_timestamps[weather_stats_iterator + 1]:\n",
    "        weather_stats_iterator += 1\n",
    "    else:\n",
    "        new_events_timestamps[events_iterator] = weather_stats_timestamps[weather_stats_iterator]\n",
    "        events_iterator += 1\n",
    "events_df[\"timestamp\"] = new_events_timestamps\n",
    "events_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "e874d742-2cdb-4f81-a28f-02d223b0bc72",
   "metadata": {},
   "outputs": [],
   "source": [
    "events_df[\"Местное время\"] = pd.to_datetime(events_df[\"timestamp\"])\n",
    "events_df.drop(\"timestamp\", axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1e8f7ac2-7ecb-4237-9f28-6272e89b52cd",
   "metadata": {},
   "source": [
    "соединяем данные по погоде и события, которые произошли в определенное время "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "26a02ed5-9119-4daf-945c-a8eeb668898e",
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_df = weather_stats_df.merge(events_df, on=['meteostation', 'Местное время'], how='left')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cb7c4631-55e5-4863-adf7-f2fa9e91a448",
   "metadata": {},
   "source": [
    "добавляем фичи дня"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "6288c063-83b3-4386-80f2-491f9bf60f1f",
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_df[\"day_of_week\"] = merged_df[\"Местное время\"].map(lambda x: x.day_of_week)\n",
    "merged_df[\"is_weekday\"] = merged_df[\"day_of_week\"].isin([6, 7])\n",
    "merged_df[\"hour\"] = merged_df[\"Местное время\"].map(lambda x: x.hour)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e2bf7046-969c-49b4-ae18-7f742a5b0dc7",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "cb5ba5e7-fc94-419a-b0f2-1797ac40be4f",
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_df['Тип'] = merged_df['Тип'].fillna('Нет события')\n",
    "merged_df['Район'] = merged_df['Район'].fillna('')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dfff2335-0268-4cbf-9bf3-2420f100eb52",
   "metadata": {},
   "source": [
    "уменьшаем количество уникальных названий классов до базового набора из пяти категорий"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "9eba188d-6f86-4b26-9a66-3acbdbc2c585",
   "metadata": {},
   "outputs": [],
   "source": [
    "type_convertation = {\"Абразия\": \"Опасные природные явления\",\n",
    " \"Аварии (катастрофы) на автодорогах (на транспорте, перевозящем опасные грузы)\": \"ДТП\",\n",
    " \"Аварии на внутрипромысловых нефтепроводах\": \"Аварии с выбросом опасных/токсичных веществ\",\n",
    " \"Аварии на коммунальных системах жизнеобеспечения Водозаборы, очистные сооружени\": \"ЖКХ\",\n",
    " \"Аварии на коммунальных системах жизнеобеспечения ГВС\": \"ЖКХ\",\n",
    " \"Аварии на коммунальных системах жизнеобеспечения Газоснабжение\": \"ЖКХ\",\n",
    " \"Аварии на коммунальных системах жизнеобеспечения Канализации\": \"ЖКХ\",\n",
    " \"Аварии на коммунальных системах жизнеобеспечения Котельные\": \"ЖКХ\",\n",
    " \"Аварии на коммунальных системах жизнеобеспечения Теплоснабжение\": \"ЖКХ\",\n",
    " \"Аварии на коммунальных системах жизнеобеспечения ХВС\": \"ЖКХ\",\n",
    " \"Аварии на магистральных газо-, нефте-, продукто-проводах\": \"Аварии с выбросом опасных/токсичных веществ\",\n",
    " \"Аварии на магистральных газо-, нефте-, продукто-проводах Канализации\": \"Аварии с выбросом опасных/токсичных веществ\",\n",
    " \"Аварии на очистных сооружениях\": \"ЖКХ\",\n",
    " \"Аварии на электроэнергетических системах\": \"ЖКХ\",\n",
    " \"Аварии на электроэнергетических системах ГВС\": \"ЖКХ\",\n",
    " \"Аварии на электроэнергетических системах ХВС\": \"ЖКХ\",\n",
    " \"Аварии с выбросом или сбросом (угрозой выброса) АХОВ\": \"Аварии с выбросом опасных/токсичных веществ\",\n",
    " \"Аварийные ситуации на ЖД транспорте\": \"ДТП\",\n",
    " \"Авиационные катастрофы, аварии в аэропортах, в населённых пунктах\": \"ДТП\",\n",
    " \"Взрыв газа\": \"Взрывы/пожары/разрушения\",\n",
    " \"Внезапное обрушение зданий, сооружений, пород\": \"Взрывы/пожары/разрушения\",\n",
    " \"Гидродинамические аварии\": \"Аварии с выбросом опасных/токсичных веществ\",\n",
    " \"Загрязнения окружающей среды\": \"Аварии с выбросом опасных/токсичных веществ\",\n",
    " \"Зажор, затор, дождевой паводок\": \"Опасные природные явления\",\n",
    " \"Заморозки\": \"Опасные природные явления\",\n",
    " \"Засуха\": \"Опасные природные явления\",\n",
    " \"Инфекционные заболевания людей невыясненной этиологии\": \"Прочие опасности\",\n",
    " \"Карстовая просадка (провал) земной поверхности, просадка лессовых пород\": \"Взрывы/пожары/разрушения\",\n",
    " \"Кораблекрушения, аварии, повреждения грузовых, пассажирских судов, повреждения судами береговых, гидротехнических и др. объектов\": \"ДТП\",\n",
    " \"Крупные ДТП (повреждение 10 и более а/м, ДТП с тяжкими последствиями)\": \"ДТП\",\n",
    " \"Крушение поезда\": \"ДТП\",\n",
    " \"Крушения и аварии грузовых и пассажирских поездов\": \"ДТП\",\n",
    " \"Лесные пожары\": \"Взрывы/пожары/разрушения\",\n",
    " \"Массовая гибель рыб\": \"Прочие опасности\",\n",
    " \"Нет события\": \"Нет события\",\n",
    " \"Обнаружение неразорвавшихся боеприпасов, взрывчатых веществ\": \"Прочие опасности\",\n",
    " \"Опасные кишечные инфекции людей\": \"Прочие опасности\",\n",
    " \"Оползни, обвалы, осыпи\": \"Опасные природные явления\",\n",
    " \"Особо опасные болезни сельскохозяйственных животных\": \"Прочие опасности\",\n",
    " \"Отравления людей\": \"Прочие опасности\",\n",
    " \"Отрыв прибрежных льдин\": \"Опасные природные явления\",\n",
    " \"Перерыв в движении на главных магистралей\": \"Прочие опасности\",\n",
    " \"Переувлажнение почвы\": \"Опасные природные явления\",\n",
    " \"Пожар с гибелью 2-х и более человек\": \"Взрывы/пожары/разрушения\",\n",
    " \"Пожары на магистральных газо-, нефте-, продуктопроводах\": \"Взрывы/пожары/разрушения\",\n",
    " \"Пожары на объектах сельскохозяйственного назначения\": \"Взрывы/пожары/разрушения\",\n",
    " \"Пожары на объектах торговли\": \"Взрывы/пожары/разрушения\",\n",
    " \"Пожары на социально-значимых объектах (образования, культуры, здравоохранения и др.)\": \"Взрывы/пожары/разрушения\",\n",
    " \"Пожары на транспортных средствах (ж/д, водный, автомобильный, воздушный)\": \"Взрывы/пожары/разрушения\",\n",
    " \"Пожары повышенного номера\": \"Взрывы/пожары/разрушения\",\n",
    " \"Половодье\": \"Опасные природные явления\",\n",
    " \"Происшествия на водных объектах\": \"Прочие опасности\",\n",
    " \"Пропажа людей\": \"Прочие опасности\",\n",
    " \"Сильный ветер\": \"Опасные природные явления\",\n",
    " \"Сильный дождь\": \"Опасные природные явления\",\n",
    " \"Сильный снег\": \"Опасные природные явления\",\n",
    " \"Сильный туман\": \"Опасные природные явления\",\n",
    " \"Транспортные катастрофы и аварии на мостах, переправах, на ж/д переездах\": \"ДТП\",\n",
    " \"Угроза террористического акта\": \"Прочие опасности\",\n",
    " \"Ухудшение качества питьевой воды\": \"Прочие опасности\"}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "b71c1cc2-55f9-4a93-b4a3-c71f4e9f45ec",
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_df['Тип'] = merged_df['Тип'].map(lambda x: type_convertation[x] if x in type_convertation else 'Нет события')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ed2806f8-d1af-4d51-b1cd-f32a2a54ae11",
   "metadata": {},
   "source": [
    "список новых названий классов"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "5d07b454-4f3d-4cb0-ba8e-822e6a39954b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Тип\n",
       "Нет события                                    566496\n",
       "ЖКХ                                              1791\n",
       "Взрывы/пожары/разрушения                          538\n",
       "Прочие опасности                                  285\n",
       "Аварии с выбросом опасных/токсичных веществ       221\n",
       "Опасные природные явления                         179\n",
       "ДТП                                               139\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "merged_df['Тип'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "0fe51040-dfea-43e3-a413-7407506baf9c",
   "metadata": {},
   "outputs": [],
   "source": [
    "target_values = list(set(merged_df['Тип']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "6fc98aab-78b6-4dc4-ae86-ae721034b302",
   "metadata": {},
   "outputs": [],
   "source": [
    "for target_type in target_values:\n",
    "    merged_df[target_type] = merged_df['Тип'] == target_type"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "d885589c-ed50-493c-b6fa-8e043d9f2ea7",
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_selected_df = merged_df.copy()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6262a3f3-e40d-4734-894b-37961823675f",
   "metadata": {},
   "source": [
    "# Data shifts"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "503ca905-6a28-46df-aac1-65ffe3f1dc91",
   "metadata": {},
   "source": [
    "для каждого из сэмплов формируем окно значений погоды за последние 10 сообщений"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "110247c4-b818-4fa5-ac62-bf008e3c1ae0",
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_selected_df['RRR'] = merged_selected_df['RRR'].map(lambda x: x if x not in ['Осадков нет', 'Следы осадков'] else 0)\n",
    "merged_selected_df['sss'] = merged_selected_df['sss'].map(lambda x: x if x not in ['Снежный покров не постоянный.'] else 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "56818a41-ffd5-41a0-b5f5-25217a1dda3b",
   "metadata": {},
   "outputs": [],
   "source": [
    "selected_columns = ['T', 'Td', 'P', 'Po', 'Pa', 'U', 'DD', 'Ff', 'ff10', \n",
    "                       'ff3', 'N', 'WW', 'W1', 'W2', 'Tn', 'Tx', 'Cl', 'Nh', 'H', 'Cm',\n",
    "                   'Ch', 'RRR', 'tR', 'E', 'Tg', \"E'\"]\n",
    "cat_features = ['DD', 'N', 'W1', 'W2', 'WW', 'Cl', 'Nh', 'H', 'Cm', 'Ch', 'E', \"E'\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "e60b5a72-dbf7-428d-ae1b-e35259b5db84",
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_selected_df = merged_selected_df[['Местное время', 'Тип', 'Район', 'meteostation', \n",
    "                                         \"day_of_week\", \"is_weekday\", \"hour\", *selected_columns, *target_values]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "c586bc04-5a19-44f8-8fa8-ecc268c56d50",
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_selected_df = merged_selected_df.sort_values(by=['Местное время'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "12426333-2355-47a4-869d-5cfe14cfaacf",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "dfs = []\n",
    "for meteo in set(merged_selected_df['meteostation'].tolist()):\n",
    "    class_df = merged_selected_df[merged_selected_df['meteostation'] == meteo]\n",
    "    for column in selected_columns:\n",
    "        for i in range(10):\n",
    "            class_df[f\"{column}_lag{i}\"] = class_df[column].shift(i)\n",
    "    dfs.append(class_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "d235bde4-39b8-4d87-8790-8f26fbc9dbc3",
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_selected_df = pd.concat(dfs, ignore_index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "003b1323-7293-4541-9f37-cc4302411561",
   "metadata": {},
   "source": [
    "# Metric decision"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3d36ff32-57bc-43de-abd8-e60b747771bb",
   "metadata": {},
   "source": [
    "Так как на одно время может произойти два инцидента, то это уже не задача мултиклассификации, а задача мультиклассификации. Также у нас присутствует дисбаланс классов, соотвественно в качестве метрики валидации мы будем использовать F1_score(mode='macro'). Это среднее из f1_score по каждому из возможных классов"
   ]
  },
  {
   "attachments": {
    "f7daf581-9f4e-440f-8507-4a09d370cfdc.png": {
     "image/png": "iVBORw0KGgoAAAANSUhEUgAABQMAAADZCAIAAAAbuLHjAAAgAElEQVR4AeydB5wdVdn/n6m3bk/ZlE2HVAkBqSIdBOkoiCIgRZq8IqD4WkAEQZoiCiL+wVdqII2O9C4kEGpIIb1s2va9fer5f57nzMydLXezSTYNznwmd+fOnTtzzm/O3JzveZ7zPMDEIhQQCggFhAJCAaGAUEAosIMVcBnbjNVhjliFAru4AsxhYhUK7EgFYAf/7IvLCwWEAkIBoYBQQCggFBAKbA4GM+bu4ggkMF4o4AgMFgrscAUECYv/e4UCQgGhgFBAKCAUEArscAU2wyAsSFgMBHwpFNiRxsAdzmCiADuDAoKEd/j/fKIAQgGhgFBAKCAUEAoIBQQJCzPpV00BQcJCgR2sgCBh8V+vUEAoIBQQCggFhAJCgR2ugCDhnjiw1HzSHWcaLdWD76kWXUvbu3rxa23embtea1vs6V35wyUvpZvYLxTYAQoIEt7h//OJAggFhAJCAaGAUEAoIBQQJBzmpc7bm09cnc/Q1xxYqte+edftXb34tTbvzH1d3+6v3rvyh79bSjexXyiwAxQQJCz+6xUKCAWEAkIBoYBQYCdVgE8k20kL1wfF4vTLTyRIOMxLnbdJI4cxi1auGL7dPrzX3VVK9do7l7y77/Jj8AyllvAUSpsxm4iT4Wvvz7+tj+xt+TuWuZRuuJ/XNKhvWASxLRTYFgoIEi71EyT2CwWEAkIBoYBQQCiwvRVwXZdfMtzp2d6F2Jzr9QAzJU7jhjr8Af2yAmN5SqNEyGMyZjLcx1eTvoKowEHQ8tBoW6POTnR+EtNhzPiKkDBBrweH3u33JegIljvwHmHxSi3h57djgfFbpVZBwqWUEfu3kQKChEs9wmK/UEAoIBQQCggFhALbVYFCoRCQsM2Y4SL49dDb3q6FC13MdW3LcQyHGY7Xpw992M0mr5Rt2bZtWCxrslbGUg5Lp9JN8z795K47/nzSaT+onHzwEgttYlhjt52xHMuuZ4WNLUs/vPnXPx3cryIajSdqhp569qWfrGytTzkmHrcDKWh7XLojMXE2NBCGXd4s+MgA9pC3lxQlEa5LN72nInXTRBiNgnQ6C413mET/Jv8ODRPRUT2df1ur0X35S+/tVK2e3rq9V1gcKRToAwUECZd+cMUnQgGhgFBAKCAUEApsLwVcfwkuuHMxcNGRGbcCT86gtD1sOI5jW7brumvXrXr3gxd/8/vLjj52/3gS+KKABFL0wAuuno/4y9Dm6Tay1Eq2YcnvLj27Voc4gOQdq4LUr3L4fi/8d9FXjIS5ur5NmLMUdoO5jXy7YWHve949Fan7psLxPoyJ5BhA6I8eArh86Um4Qx17r7Y4UiiwhQoIEu7+50jsFQoIBYQCQgGhgFBghyiAFlTHxpWbhB1kSMdxLIfPCnW6XdDiug1Wx7KZaaN5mvyUXRsLYrmW4TKbr9w82aNSruuapnnxxRdrOhGtChFdBoCIptJ7CbTkdU++vJixJjyPydhGtmH+YRMmVigRGUBR8ChNkTSlDCAJUHfEMeeRH2lPuLWtDYPb4fy+TTgQ1ydhzyZMowYEwz0Xhp+n52N692nve9s93ZqgPp03Amd5zsMu+oLnySaMzgL+WEzvitpTAbbyDJ2Lvan3Ybrvuh14RNPIF3m/00HBYBOfI931i2KPUGDrFRAkvKnHV3wuFBAKCAWEAkIBocD2VMCxrXzmxSenvfjUY/954vHnnpj25KyZM2fOnD5z1vSZs2bMmDWju2X6NltmTps+6/GZsx57atbjTzw5fdrMWY/PeurJWU/9Z+ZT78x86i3eFetZnmw2yxhrbW3dsHH1yjUf//n2/62Oq3FJiusakXACKoa/tKphLWMtboaxvLV89v6DyxMQHTf+67ffdfvd//rz+PF1iaisgApQoSq7ScoI9B5HCHS3kmp25q/7JBzu7pK/vMeExf0912K7kbBNzgJUrJ5AtEtr8SsSkDDeVYu5iMEZGhpBSvRr3XNlt8OnncrvumFW7fQhvvWr1/0GJ2GaC+0Q+Of5cbaLY0L8057PID4VCmyxAoKEu3lixS6hgFBAKCAUEAoIBXaUAmgTdo0zjzs4AVBF8KcDoAkVfYTxr6pH4/E4vg8thJR98FJWVsbPoihKMpmUZbyiLskKgAagFEsSBxi0+4RDNjanNu3FXSQcZpsF12i94VfnV2i8UgmAwXsfc34rAQMzNqRXfXTgnrsDwK1/+1eesSzNGJ722D9UAB0gAtUAI0Ae1dheMOwCv0fbAX52yCW6YSiPBpnrGIVc1jYLpoWrYXe/5ml/z8eU+i7fz7/rvZrFxbDsYLVdVjCwFZi2azO+2jazS4nWpV5BHDS6n67FjAxvDiua7NXtNHvcr3ipc27P/aHyo3XedQzbzJEHB3pv22Yhl8sYZr411bpm7erFyxYvWLRwxepVjc1NmRy6/5u2a1ho5DZpWnyugH/yzDGZxawsEr+VXf7FklX1Zt6LDNc9Qm8x/IgvCgUCBQQJ76j/6MV1hQJCAaGAUEAoIBQooYCTN9Z8tletUg1QIyOC+hNlZUlRVT2qkMdwCIT9z7eOhWVZ1jRNUZRIJFJeXs5PpsqIwUihEugyqEjDCsjlINeNn3j0xqbMZpEwVtjK3n7Nj8sBYgARvRa0CZdd/1Aez5JhxtpTjvy6EtGfemNOK8PIWiaRwXuvPaUSiqtQDtLoQeO+3ZJjXxkS9kGRO0UjbZlmIZVJtQVIVooDbQqmxQ8rdUzP+4NL0EaxuQY9aYeILpXOWw45zG8GCfN68YnAeYRBO08oaDIj9cl7L40fNwbk/rf96zmnWPGeTM09V6TPPyX05dZgg7kF34nbevP1F88796yy8pgWI7d+/5GUVSUaj03aY/JVv7j6vTnvZ/MGZ2DbZe0FHNMxMBZajuVaXpz5cCISTdbssaYJ239YarEtFOhbBQQJF3/UxJZQQCggFBAKCAWEAjuFAk6WWWvnTrtjIECUCJBIF+2ykoJzayVJisVifh+7L/9KkqRpmiThBWOxGLcJqxyDAVQJFDQSywBlAMPGjz9y80gYe/U2y7de/J1jKnHKr4wvAw975bOUZTQxY+lll/1QjmsvvjOHG5rzFEqX2cYLj9yX5MMBUhSiux953rUpl9loLPTMzX3OOTvDCUMQ5EXGQv1cds/dd2gqxCLct7wv7/7mnEuiG8Jf5fKKmtnvf1AwrZ5twrxGNOE9wGBu+E8zlnUd9KJ//73X99trnAQQQyeEwX9+8A2PhLdfiOySvE2uznwSP07bJ/rFfY6dzqQbbr3lukG1FQqNFmmaJGueBwVXVZIkRVFklc+Vh0g0+fV9D5w6fVY6a6QMI8tYo9HIWFuhZcnJh32jTK1M9Nv73XntgoT7FvzE2TopIEh4p/gfXxRCKCAUEAoIBYQCQoGiAk6eGRuYufoPF55YTTBM3tEYPYr3qjVZH1Y35rGZzz828/npM2fNmjFt1rTH+Tpz+uPTZkzvtPZ+EvETtNx///1/+MMf9ttvP87bFYmoQtf2/LOxEDpA7ejdDl3XkuGd9WLhu24F3tFo9Wq1mpZMHlhRgzZhCaB8t0POW4+zQBvvvPknNUNq7p32RJryCPPTGGbWaF938xUXDZLIMC5FIDH2opsezCJUOTZDL9weHHF3BqDd4jKESNhPmEQk/Nc7b0UneRk0RdEUiPirpoCsKqDruGo6Qhe9KioyGGhR0KOgKTJCWm9X/G6wKsUFLy3riqIrchxN+7H+L789h7sIEwx3uCkdW4TDSdjx4mFlGUsz1s5Y++z339h3nz0qEsiKOACkKVAx9sb7XnZ45x1J2NpiMbf+i3Q7mOvatsfAWUz3ZTUyZs+a+XC//uhDEVEhqoEXCS6Y0CAXnTrCYw2qHtdiyURF9RnnnDvjlZfnr16ydNmH53//BHrYY5Cc9Mw7a4yOjaATxoi3QoGtVECQcMdfJ/FOKCAUEAoIBYQCQoEdrwD3OM64y949amxFZcj3WSdIiCCIVt31wJvtjGVMlm5vdfN5Zlo8YhFhk4sWK5dZGHR6CxfDMNpb22774821FYkIQEzyDNR+b77fpCnfXp7KpMlPFDXzJ3OW1i/H2Mq3n7ijH8745Yt+yfV3NjP27LOPVibhkp//opXAyJsBzCzbrWf2yu8d+vVBCs0r1iKgDHvygw2FrwAJB/BGNNQht3R7W8P8+fPefvP1y39y8aCqWJxczT1ro6yCrIIko0VSUkN2yTKABO0MWXPDlt1g27/B4b+SjOzNF1WGiCzpkhyTYjG5WpHqpOjubyzY0OIbSTs3AO7eTI2aAqHzutA0WWaztjVvvzXrwEMmFucAEOdDtBzUutsfeg2Pw+9iuw402Y4b3PxLRmx0TXdMjOSVY3YTczeylk+uvegElIXzqwSVUUiQtT5ZQWwciw4aNnTQ0CEDBw0eWNufC6hrEs5EiMRA0UGlCOpKHOSECnEVJ0MA3il15D0z5hidpRTvhQJ9qYAg4b5UU5xLKCAUEAoIBYQCQoG+UsBoa2JOy6LXpicB+kV5FxoJOIrYoylQ1W/EER+uYFmXYQAennQFDQQYsshwHb5uDQljRRyXuaypftkBew2q0LG3HwNVR8pSAKpHjj9kSSbDZ/PiwZskYQwNveyOX34Xz6PqICcgVvnaRx9+sWHNhMlfO/a4E0xCKTNAa0wwvNZqnTtpYLwMQI/IkKyAuq9/lsJIWtuRhUq6y263MpAkiIOYdprfWy9kMbPaN744/a66cmwYOFwgdSJhNRaPAKggDwBlYO2IMXXDh/a8Dhs+NFgnThw/bPjQsrKEb4WWVBlDeEdliClKBDEuIcFAUMfMnLN6I8U+7qb9ByRMCYFxonMh5bA0c9rXfvLfHxy9vyr5LgdeMwecBRCtBrXutkdewWTCaAz3DOPbTXP/QhaZr2k+s1d+F43Ydj3LLb7giNH9vDIjClckeWIwGD5y5B13/qWhqdHGpwIXbrtrbVk/593XTjn52AD76dsKKDGQyhSoUDBPGADEQR993V1PCxLupjmJXX2ngCDhvtNSnEkoIBQQCggFhAJCgb5RwGFGppBq5ie75fIfJIkUIpouSZIKUgKkKFpVq/c74kyaYckNZohLro1GYMvFrL/BumVGYcZYLpMlxk7n2z457fgp1VIkAYk46FE5BlBZPXiPhe2tjczJMbxCr0g4teCU/YdSV78fQL/hEw9Nt6376WXnjBi/fzvv9eeyFH/IJKuvzdjGRa/eX4EkR5G6kv0PuPCXqxhLfeVIuOgF6dqWYxmmVbCMvJUrUFixdc89/tcE+qyTEZjswchaaBNWNYyvFplwyOmLW1nKZDkj1/NaKOSC1W/OmEd6zaqVM6c/ftQRh+sSVMeiFbpHfWRirrj32c/Xd0fCfIjGZMzEqd3UmswcczLz5756/CGTB0g4uFMdBWRtLLZnd96ZSJh7ZVNML+Y1cjvX6m786PxvTyjHyOpQVl0JIMUqK0HVIFFzyz8fbErls3ls0DbzvDPytp0znWzBtcg2vmbtF7/85SUxHcqjGvp8YFg6tLWrnrdEBLQx5151J44CiEUosM0UECS8zaQVJxYKCAWEAkIBoYBQYAsVcJx8qvhVc+1xX+uf9EJnaQpoMZBiICWjNaD1/9P9zzRZzKAet2PyBLuW4xihdUv8o23LNgzDRmuzy6w0Yw2F1gUHTxqfpEhXSaUMoEquGPlZS+MGZqTxWCMgYerqF4tf3HJy7pK39xlASZkgCVUTz778D49PfbQsqj33xqftFh1o55GsGaaQQexwVk+94cIkJ2FQQOv3+2feWPwVJmHXtV27g6c02uLdJie14pvjx6DpFq23RJVIwjJ3NFYjVVff+kgzkapv6ixp6CZc5eBdvHW0kzmWwRx7/ZqVJx9zSJnnjc0vGb/90dnrSpOwRRjMSXjep58cfvABA6tj5SoaQCWAfgrEOb2jXzeeMCDh2x96DWkQ3Q12lE2Yk7CBlmFu3DazzGq+9Zc/Kudu5ipAWRXo5fiAxgY+8trcdt9LnEYBsNwuDVcEBl4Lpz4bjFmrl3xy2H6Tvarj9yWNRi9w4Esdc/yZ12DsdBetyvw1fD/EtlBg6xUQJLz1GoozCAWEAkIBoYBQQCiw7RSwmZte8+GroyIwUEYDmgIRHSQdIKHoAFK0dq/P6nG2sGEwDN3rYFAf18m7DgbjdZ08IfHmwTBN5nTQEZd64ZyCmNH02RsPj++HYaPjUAkwEKD2s5bmDcxodzCtLWXRcW0X/3rMEGZiBJnsu/++ZTeAGjT/RUAu++N9M6K1e5596e/Wt1MepZCKyMUOY22LzjuoDpEDg1ZXwIBJLzVkV9JcYo5HvXzdJP7t5AcEdOp6TtFc4sDt2GJu9o5fXVrNvYw5CXPA4hNYofzPD761mSQcXIL797o0ymJjnDI3Z+WWX3nZaaEIamW3PPxOfXck7EEsDqbkP/t4zuGHHkz2Tw3U8rqRu33nhKPHDo7HsEVxm7CK5cXzKhCpAnnobQ++uhOQsGNTal+qi8HY+qUfPTYwgmWOSUSukUpIDIOKiddPfX0DY81mh+xHvCH74b/xHMXFMYy2NZddcCo3hcs0k5uGM2SQRu976PkNzWZAwt7zyGNXe97WRU+BoIWU2CheU2wJBQIFBAkHUogNoYBQQCggFBAKCAV2CgU62uNc5hjMbLnvt+eNUaAcDX4RCTRJUuK6FomWA9QefOLljVkMaOtSTCHXMZiTZU47c9oJiY3NDZvF+9xez5tLYjNWaGX2oktPnVyBXrhJgGGgDZ/Xlmpgdoa5JidguwsJW45rmNhvR1fRwnVnnzwYiRaX2hG7f/Oks/uPO+6jZWjldJnFaYG/YmVsl614Z59qKEP33gjow4fse8pCl632iCuMapvY3slBd5PF64g3xcrypkI+tMYjt/+Kp90ikvTNwuRzDFB5+4Pv9h0JZxjbmG79ZJ8JA+M8nLladdtD76y10NDZYeGo7lrMaX33rad4KKmzf3TBq29+2JLFiG6MWW1rF5x98J44EV4iAzZCoQyytjORMPp1o0Ebhc8z9vml3xs/IIKxrSKchPVyiNUNOOxHc3NsnUUtN3TDuCD+Q10cHDIyPO5bO7M2XH3FGcP7QQ06XHhPB8DoUWOPb2g2cd4BZmryRqZwKATBvPcM7F+5w40Rb4QCqIAgYdEOhAJCAaGAUEAoIBTYuRQId3K9krmtrPmj8/erG4IRs2SMNaUm9EQikqhC26w08r5H3mjJYe/YMDGiEmFwI3OamZ3HaaWbuXRDwliOHMssXPzaI6OjUQlUUEdB+Zh57W2dSZhhyC7qqDtkoXbyqTSvRXrDxq/XDikjDK5MkGkZqu998O0cBfzizq84obQYMSvz/jN3DCJDmaTHQB/zw6v/0UApd8iTugiEm7QMbxI1d/IDOpJPUHHEM9MbF8i/M/XuIYRSiJLcIEwzUImJK29/YHYjHbnJmvrXCq7i24R54irEsBxjzYy1PPfI75I82FW06rYH399QwMJ0WfiNLcz7/IN77r6jqbmBMdbeargWK+TdXL6NOa3tH74xirtxawomrZaVnYyEEUY5fTLHyC5+aTAFrtM1jB8mK2QW1uPX/Ovp1XRYwT+4OwblWYhxtMc2c75Wppn94jc/PbmaMLgCb5gEsFuifN+mJo+ELdfL1YweH94jwp3kw78WPWz7lxJ/hQIhBQQJh8QQm0IBoYBQQCggFBAK7DQKFLvRVp7lm5ix0pz30pQKCXOzKBHQEkokSna0KoDaxMA95n7RknZZFkHHYnaaOa3Iw7bh2pvJwUXrExIxn6NoY+ifDHM3sOzKC444MIYZhYdC1dgFaYNbGol+0VpF6WS5J61D0YJ8I1i6fen7c+q0Moy1pXFv0MqDvnU+nwSK7EwdfJMogm5CnrGGv/zuB3VJSiyjRAGG/eO5BZhAiVhtk/QbPmCT+LeTH+DTKW8UAaMinhk+CS96+fFB24eEXX6XGs21L+01kgdzq7ntwdkbjG5I2GEYv8108pZr2MwsFHz8IzMrkrDdxNLLJ8dAx0nNEig7LQmT7E72X7+7rJbmYsfiiiJhvG5V0kFJvvHF2vWMbTTROyNstO34i8JnCOMk4UIeR4jyFsOHi61jxpJzjtozAuj3Qc/18HjZlPpWlqGBCNMxLHycEIMFCXeUVLzbcgUECW+5duKbQgGhgFBAKCAUEAr0qQJhk05wYhc7wE6e2e3M3Ljo+X/Ukk1VisRlWdVBV7Hn3A/0QQec8KMGF4MqGy5zbctfnS0gYY7ORcuwg4lckYrdNmbXv3P/74ZjWOIqqBj75AfL2hkjI5iLTptEwngkxmNC85eDRUffTsay9/zi4nIypmHKVFkGffS/X1rBGEvlGbIVdfMxYyxylsFYI8vNP3BcvIpTs1IOg/b7IscaDZbK58LgGibeUtvh43fF7VIkHMw+Zcyc++z/DQSII5vhKvP8wL539G0PvdvQO5twWJ+Oehbbp2XkKWzZsou/u1dClUEbdNMDr7R6TB40Xdzgrci2DZvZNrP5W3QZ4DiPENjC2t4/cWIkioGuJSz3TmYT5vWxnSxW2Vh3xkFTqih8XTSmahiULK7A6MpB+6+1nFYcLsLFprEgnLTvDwT59eUjP0S3ga0dh3bSLLUi88krx4yrxfauJUGuVip2f2Fhaz2dMETCBQquvlkG4aAQvCriVSjgKSBIWDQFoYBQQCggFBAKCAV2EgWKpOGDAuGCS9YgmmzJGt+/+ocHRAEimqpKSMIRiGII3tgAkKtveeSVNRbLuQiuiJQ2TS/cfJtwVxK2Hercsxxz1rXNeXCcDrISh8rxT76/ipMwhsMNSNirBydhq81xstk2lll18RGTywjSkM4UffRhP5qXYhhJybeT0XUpELadY2xVw6fTxg9CtEMbmVqz34n/s9JA6iabcDH0cQetPOAIrKbeRpjudsXtnkkYFXHtD56/f4AMMVIYSZhbFj0SLr/toXf6joSZVbCIx1Y+cufVFZKuR4Zc9/+eaCG36U7PUnE8BWe4eu9oxi3dRrfA3AaWfu+0KTrOIsZYX5IkKRhEeWeKmEX8mmNsY67x0ylDqnFARwNdk3RQFSiLyJMqaw/aQAMBeWZn7DzW1Jsc4OvhkT+9DTOw10ILLNfAnOZ3Hr2nPEEjAmockkP/Pad+Od37vF0wPJuwIGFfUvF3qxUQJLzVEooTCAWEAkIBoYBQQCjQNwpskoTTzFjRuuTVw6cM0dGqqstSPAZ6nLLtgpKEYfu/sCjDzcIeBodAZHOdpItfJQzGzj0rMKeBtc4+5ev9Ebgqxz09J0TCROye+QmrwkkYvToxMtLSD3bXKEQwjxKsxv86/b1WxlpsVuA+viZlJXYsMhhmGVv+zH1X99PJNRqnTVbfcPdTXoZlZAMkfb4KEsbW15GENT7igIMIlPVXKr/l4be2koRDgjM08boF5qxeM+/lah2DKP/4N3/KlLYJF9uSv0VPjGNkU8zcwHLvn3VAeZICo0u47HQkbKBt12T2unXzXolRyOh4BDMJJ5S4DEkdBibLx89tyPCZ2OlcG9ayBxLmPxfBcA22+AIGnrPzrHXdyUdMwXumxiEy+I/PfLiGVM1ZeYO5BtrVC/gYogt2+Odik9t98wu1nc4SHjXo/SUDPflAQ8j6zudT9NIyzqXs/WV36SMFCe/St08UXiggFBAKCAWEAl8mBcLd2VC3LrAJu3lmrGepJW8/ee8AHVRZViGaAJw5TB7EKkD1Xqf+tJGSDKE/s2353EE9881E4fB3adtGJ2anhZmf33j5yWiojY16Zs6Kduqpo024MwlzU2+em8reffyf1Rh0GkBWQYmDWr00w9qoO2/wmY9hEmYF5i67/ZenkD0YKMlOzVOvL26h2MSFQisFyS6ahUOQFtItZB8ODthFNzpiT1BH5CGehrl7EsZ2wUk4ecvDb21gLIsn6ka3UjvDowzBMYgYGOzbcDMrG1d9ctl5Pz7//J/eP/OlNj+PbviZ7NKKvB10jJNubWKF9Sz/wfmHVO/EJMzyNk1rNtctfmtmlAzCUQ2iqhRT4irEdIhr6qCpsxct4YTq0HPXexJGcuPTA1DaZx+4dWAMcDhAH/yrB19dR4+zIOFwo+p+O3gsOpEwvfWif3f/zQ57BQl3kEO8EQoIBYQCQgGhgFBAKLBdFChBwh7R8TjBOcSN/Orf/8+ZKkBC0pI0bZhwRwGpDJSBv3/g1fWMpQoOzrZ1vJmZpYCk1/s5BtvMSTFj4Ssz7qC5vrVPvbOUR8wyDMOnJpLK64xi8iUM9mxu+MUPT67xiB0gUXfmz//YyliGMYvmQfMoQ1QYdOlGeDYXfWOciuSMNFzTf/TB61tYDn2jLTpncAH0uQ0LF2yHb1lAcbvoxhaQMN6gEAnf/MjWkjC/vw6aIqk4roONwcng5F/HC9xFwaLCrOvdhHAzK94Xfg8LG1lu7kVHDEDvaBzR4fSugl4J0pBbH3h5J8gnzHImjTk4Gz98/t9xCV2jIwqSsA6qCpqOZU6eeuUtXzjYpG3LD4bO22JXQitKEGz5BkuXsdzSoyf3p9sX+85v/7GcsUaHFTAXGaPzGjTgIGzCgXT+RledvSReeMAmSJh/l87Eb5p/0i/5X2ET/pLfYFE9oYBQQCggFBAK7DoKBBAXBBQKd+6wL2cww3HaGWvJrPzokEmjMViWLJWDQjZhSmgjVUC/PV9fYXBTIY/k7HEIRb0KM0mvt8kV1jUpJG6OsYb1y98c1q8/QOUDz85Za/lmSY/YSW+vZ4n9eiThzPKjvjY6AhCLxUCtAKnu0deXpLFHjxCF3y8iFHlh220NC5/lcYnQz1cZcvQZP1/XQLRgpZ3cxg/nzr7hphs5PfA68tewguH7vosCcFDszSJhjWzv246EPWhDzDBpxZvo5ZOmHFq8XYX1D7e04n5OwkYjy36wk5Mw2YQNxhq+ePeJCnSKRhjWVDCWV4sAACAASURBVDS40yoBJPSRB37SgiSczRgdPJfDD3Gx8p23vKEdVHXNv2++JIIDGZFDLvzdQgwphuHkto6EnZCHRIcxq87lKP3eh/XSR4Q/CbFl+IchfEg321206tqWuvlW4AgdfJ3v8aAWC87HEUtWIVTabUfCwa9T91XYEXsFCe8I1cU1hQJCAaGAUEAoIBToRoGgp1SShDN23mCULpjlV7791PgERo6Ootk0WGOgDtzjuHOXp1iejEc2v5DrpXZx3WII3zCfhLcxoFGHlVLIugXkVcc0rSWM1R+y9967DZ8y/ZVPN3hd+KA3Gq6YS3bfpg0fPT0shgGNcVGqyr928pICIrKLIbgoiUyYhI0cc1PPP3hdjQwqj/sUH3vHQ29nsT5oLrZSa7550D7fPuH4ppYW6uOiWViQcBAxqw9J2KeCgDCK5ne8374DMEULxzaD2W5xxaVDOwg1r+L+XYeETQwGXWCsoX7RazgCBBCPgKQgBtMgFLljxEadcuHvGih4dFEm/igHD3Sx8h22Ql7uDkt9sfajp6o0jCA97qQL59NTzP2nyV8Cp9/T0FH456IX28W75d82Mu93gUPc4d/3IjMXMyp3KHiJN9wYG9Tau3Tng4tX4Z/wNOT+1fmnxZDjQak7n8YvZKABb1dBGaiaPFF5l8r65+pCwsHJgq903eN/uVd/uYCBaTo4bYcRiuIISq/OufUHCRLeeg3FGYQCQgGhgFBAKCAU6BMFwn2tInv4HT3ck8MUQww75W4bSy2bfusvkgAV1BknGFYAIrjK1dfeM62JEt0YiMDUNyTr61aTMKa9sZ11djrFyOBLU0959UPdSU8PmxIvLXvtX9cN0CAi6WhC0/qffPlf+exiirpEJEwuzpSoCSuYb1xx1TmHVwInYRkqJ7+1KO9V3Elf8ZMLhg0f+uHHH/GLBBjcyU3aKwL9CYyru+hGRzoI2oZnTSdlirGj+5aEqfkFnvkF3qHnXIFz0f1oajxyOHnjfwlJ2EYXC4ex5hWLXh8CMAhwcjBIYRJOQHwoQPU/H5zZnkOffztMbsGT3WF8oNhCiySMVzHc5sX7TBoJij5g7yPnpfEu+ySMseJ8J4rgpCU2wrvRld3qvOJsb7wexzN+T3mZwtt4hJ/ru8hvxbLz3ydeW9rmY2IB8+GlDVotGjuhb/o6BFfnLYwnqg6yMdvMtVzLcrubdx0UIHgaOtSXasb3uJYfnJ6cWkLPEp4DY/LzKH34LlyeoAbhE3s76beGk35wGJUofCzt8PE+qGC4jlQALiC/ePjm0te38Ysg4W0ssDi9UEAoIBQQCggFhAJbq0DQ18OEQ9gZdU1MXJpdzeo/PvfQ8QO5YUpSMd6sFAVJk6JVUDFmxvv1TYy1udj/w94W9bLQbOendQ0Z6kpuesZhvKzvYIkmKYM6jZ7PoWd27uAEiUmNMQ6Q08Kcxb85bT+c8QuKHOkPkLxr1ntN3NIV9Bs9iCZYtwss3Xjg7tX9yOYGIFdOOWU9Y3bzemY33vDb31ZW1c568ulMLuf3XHtjE+YV2FVfQ733oqGNz5nmmBSOmNWHJEwt16WJ2W3MXZnPLH34kQfvf/iJpU0WbwQ5A/Mp4VAL5dCiV68t9a7Vu6ywkaXf//Fh/eMqnxNOTv6wc80T5lnETJa1C2uP233oIO7d4CeswudPUhU9GS3vX149+NmX33WpeXPcI3FchGPTRZ38e9mjPvaf//jHMWN2n3zAYU0GRtOi0NUc5ILpBJ2mCvv3ChuIN5vbY+jgSoUMy6f5OzvTVr988br61c3tmTaTNVusgbGNFJGOMZYm9LcwBh55Ztss7+JagoQdzIfsZL26MZbNFQ/Mt7fggJeZoeDbeWaiHtn2lvqVSzZuXM/bsE2x3Jro6hmbZWi4z3BZwbENhiTsjXa5xUTNqIVNDBkIGmzw/Yy5ecNIpRizsoVs2mSrGvK80XLOx4zYaZ75K49AXLAaGxtXrF5luCxPv2ipAiqPSdL9wQK+UXBc07QxU50X6Ax/DXFYsGDaTpqxbFt7Q8HIpFKp+vp1a+obWtMYHT9YDcZSmI676xKGZX+ooOtRfbpHkHCfyilOJhQQCggFhAJCAaFA3yvgYSJjmEGF+pgOc3Msv5Gxlvz8lyZWYMdc16J6+UBQKYSWEgVt0JhDfrAsg07IxMFmiIRLcm/XDzwSDmC8ux5asdtbrLtLvUST2S2scc4xo3SKfVUOck3FoMkrWzGIsRm4L9LJvXoxg9m5pmXzx1TI/WSojGLV6g48YU2eMTP9l99cWZbo//D0F1J57CyTAaeIwT3ahL1eetBb3rU2fHripQ5uBr+l/AZvE5uwb64ycbKqu2jpp0/iTO/ooNfmrUfHdFo6kjD3kcZ25H/e899dg4QdItkc4kz6+rNOHgFQQxN5Zc872idjABVzSmn3PvpCmrEMtWkv1bZlMwS6YjssqYvD7Lxl5QgHvUaOF+a2RP8x4cbDcCvm56M9/MniEyJsxjKum0rlWhqem/n4xT/6YV2/yqSKUdyTulRVUQaKDrF+kSETTzz/yqfe/ChVcAoOOoJnGMvkCraBI2i2jT8jOcK5boqNjgFpWnFQxBsX4z8KZo7l27IbVn34xitXXPCjfSZN1JWIApLvVY66yYlkonb4d869bMaL7y3dkG6jenPcN1yn4BoFFzOxIQx3JeGwAME2Nj2XmQZivGFk2hsfm/7Y6K/tVztmTx7I3iAxc2YGAZ61PfS3m+pq+lWXVUU0/JUaMGTod868YMHybDtORME1sOJymkUAtvmj59iY2gpB2GFs9erV9/7zL4cevu/IUUNAAl3nYdtBiUQBpAMPP/qfD01dvq7BRRK2MiZlj+ugpv+b6V2wl09Qh1Ns7htBwpurmDheKCAUEAoIBYQCQoHtrECRhH3rq0NmljbGmlmh/rn/d1MVQFwCxGCZJg5DDOK1ADW/+cv0DRS+x0SLDRqD0VWaezJ2pd7Sezxzb7EgHRTgnd4OJmGv624yp23V6w8P5sCgV0Pl6KNPu8R0EQq8b9GZeCeW+pMdSDjhI8bQsRMmjdtTk8qv+PUtnbKp9s47ukggQYd5F9rYISRMFw2oO8XaZ7809UYcc6kZN/2tRU1041zX7kTCQSPq0ERKvtlFSJggEA15jj3/hal1ZL2OqpLip/ny2ylEIjFVj+rJqt/fdncbYymLMk6ZFj56OI8aic5v7SVFCX8Q+OsGJEwG324ar4NPVd7G+AB06/I2a21jbmH95++fdeoJMkBMkWIKTlHAmwhQU5GUAeLJckjWAsTwGZW1eHnVD350wbufLC4wljeZkcfyWy6eNE9uHuGy4bY3nmVQpHrvQxtjvLcya2V+xdxfn/+9ChkiMo8SoIMclyQFM1FRWmZVAk3jIipyvBIA9v7mUY8++bzBc7MRiOZdO++6hsuQyskbv8M07LASHFXpl8XIZBbM/fgn554XpdhxEKsZscc3sjylG8pjMZb55O2nJo2tAwx3Fo+pUU2RcNq3JIEUG7LbgW9/XM+PD2CYz60IKo2zT+z1jDUs+OiVI76xF+bD7tgeZBVzxnmLAqBGJa3s4MNPnj13hUE/5eHfQNIuMAv7du3Ocvfxe0HCfSyoOJ1QQCggFBAKCAWEAn2tQFfjgIumCDvD3DQzm1lm7QUnHhzzOlwaKGW44lIGUPX6J2spdJaNc/0cz5aBns6bs/SKhDtUmwxS2O3PPPaXa8p52bRyiA2544GnHM+jlzv64tf8HixN6LNzLLtx/7ED+smgASQU7LqrelTRKq/5w19Tnq2ZEz3yrSBhVNDte5twQMLkFp9iuY9+f8kx0WgcoHL6W4taCYLaWlr56Ar5Rdsc9njL6tAcSr7ZJUgYMzDb3OHZYfbGzw8YrkcAqqJ6J/IBgIoK8tCgBn/K93+yvL6AU/tNxN+goaLBE1OFdb8YmQJN4MVPEZ/9p6M3JIz+2+hvYSC/ulZh/aoLTjpygI7B6ngwAf5aU1UxeuSI0SNH9K+pxhsqxUFO+IfgLAYAOP2Hl7VnmGEifFKQLrPAOL51LDYnYcsuTow2WKG5Id3y+VUXH9sfoDo4b6QfQFm8ctCkSXt898RvTRgxICZBRKYsaXQMuhvgox4BSa0dMeaV/87hkcEKLqZi4yRs+Gbh0MiWg5e2O0QAuOsvtw8fMjghcQiGSEwHvXrk5IMoXj3H4Oxzj/29XwzrGqus5r9P3mskolfWQmLEkadekiK3bR4KgQYgeLgyPiBZYOZ6ll5yxF6Ddb+OyURlPF5sAFoENPIdCJ1fo5/lfj+5/LbmFBp/uY974B+DzQTXEoOOHbXf+neChLdeQ3EGoYBQQCggFBAKCAW2vwIOwrCTZ3aWWe3OuoX7jKwhs4uMnVq5TAJNgghI/UdNPrLRYCnM+EowjPMVWZhYekPEW1Q93pkzL/7BSROHDxg9evSAEeP6Tzpg9uJ1/tn4PDt851vJ0D/Qzbcwln5p+r01MtQmqylPTWzwgNHPv/KKi9OePV/E4Fthm1B4279K8fzhT3etbV8hXurATstdNKmi24qEmW0WbPT/bGXWoiMm0JhGsm7q6/NbGdIHs30LZ4n4wOG70N32LkHCyD84U5dCqrN8/UtT/1wjQQU1zRDkgOQvZFrEXNhVFcMem/X2xhTO8A8vloMo3HGf/3nx9nrPBWcjf7SIf6vb9msRQVm2tZG5jTMfvrs2IVcpNJJEpRw/cfKMWU97+a4IKdNZo7217c83XDsgwU3Ffm1kCSA2csy3//t+K03izRfcPBGxX8jwX5vhHGLDcfCXJcvc1bPuvXZADcTjUKFAQgJJjkYq6669/b65C9dymLYLOD2CObls89prf/nTIVUYgCypS8jgCkA0UlYzQImVH33cd5rbcHpHzixK5W1xF2gsBk1lZnZ7Y/09f71tn8njuBVWouCBaHxW+Pzz5NBJ32jBMTiXsZZXn/wXRv9WAXRdjiU0OaqCguMa6O8ug5KA6FAoH72w0UhR3EGKbW8WMk00Zz7NcitYdsk9117QDyDONYuXQaTy+FPOuP76m56Y9dSDD/77ttuvP+uc70SI7mtUrFb/aIWGxvgIfamium6POZ+vzbmsLe/xcKfHu+sIaFj1PtkWJNwnMoqTCAWEAkIBoYBQQCiw/RXwo9NiaNb0+888OjyBwWxB0kCOKaAo2PuKg1R91fV/58YNm5nMMnGWm+MF+O0NBvd6zmdPChiGlbFZK9lY/F5tMC8OkYp28j0uKzQx1ppeu/jpaU8/+8QbCz6pRxdNPMZNO1mDEEyQ8CYjZsnYrycukJM3P/LWBk98tHD2ZkWbJIFLwWxidv28V/9dxw1cZXUz3llI2YLwjvv3zo8e3FMr6PrZLkDCZLfLo2M0J2E3ZTYv+t5heyXJx9hnxyIGIw7L6CgbRbBSQR1w/Jn/syFl4ynI3sfbv69bF00CEiYSCmzCvSPhHMtvcLP1Z51+ZEyChKrKRGqDh+w+Y+azGGjARlfnYKX7ZzM39dfrruAA6cWhx1rpAMOqhx7y7ueruxSxyw7Dbly70nYabGPZD741YoACsQgoUX5KOOCIk1ryLEUTIgxyTPF/AagJG20Ny+efcsSBnpIqwrCERnf8+oDBo977YJ4/K4Sls8GQgsNcy8i1rF+16O93/nH/vcdjiSVc+aKRw3eIhMuHTTygLZstZBuenXp3WRBvX9WkSLwDCSsKqAmQB4A6ZNZ7Cxt4hHy8FxY+QPl1jK1nbMWlJ+3ZH6A/VbFsYN0djz292kTI9hesYr7Qaljt/3ly6rjKaA2gh0vEsx5rIMVASiZqJyyqN9Lks412b99EHITF88+2rf4KEt5WyorzCgWEAkIBoYBQQCiwHRTADhcaRQrMbP3LFT8coWP/WwJNAUWS+Nw8Ta6Y/OZHLSmMfeMyK4dxXG3eIfWyvwZ+m3yjKx5vTUWC3mHetk3G1qbT7eThTL1h6lyiPyc/ipt3MLoqToHGV5Z1WUvBxXCy/MMuReFE0fU1fGDXT3etPT5v8lIHqLQJmzD19ulFKdsCEuZNyyxkGNvI2KoLTzvE83KP1Mx4Z2GKytSBhPloRlj3TW/vKiSMwdIdLww7Y3Z7dunc8ZWyPyUB4cu3B0sKBnDHx5BsjERmSplUNermux9xGMZDppaP0gQbHXQKmiY9EwEJB7vpW/iO7ym2DTePpvvGRcftvwdNsI8A9APoN3n/05ozrC3L76cXAorHTy6YlplJsXyTsXb+XsPK4wARKm8ijvZSvXoIKLUjJp+ytqVDOOgOpcU3rsMytlvP2uadf9KUChWSEcoypUYhNvKs/72ryQdZDANmu8EUDe88rsHcLCs0Xn35eeUU7w9/t7AYHGqjg4eMefHVty0Kzux/ha7JDLPQ+IurLihP4qFo+yXOjChYC46dmFPOcw0vHzlhn/TGFS2rPh45vJ8akaVoFKJxUDW6kKKDpKL5GnCPWgbKQNDqZr63dD0PHsZ/n5w2Vqg31n144cmoMAf9kXsc+unabBsdlmdZh6XJm9urM2WGy22Y+8yeVdCfYNgjdSUKcgXoQ0fvfcJnaxCGMeo0WZwDf+kuOvf9DkHCfa+pOKNQQCggFBAKCAWEAttBAd4P5rk38XJGO1v14bET+nNTlWeOwUl/EdBH7b7P91ZliCwdmxmpgISZY3fCYLu7GcRbXB2HsbacxcPgZrNZfp527LBzDHCo+5clUxlneupBmu1me72d3WCzgkmmEi9UtMOMdsvm4bD9MoUJIbztf45/w/t3xe2ONdgECSvU4dbIz5TMwgBq2c2PvLG5NmHOabaZY2z9n278cZJ8V2NaAqTy6W8vaPfvmU90XNew6r3Z3iVImEcyRvL0Hjc7z9zUmg9fLSfLqUYAJkmKpKggI1npRRL2wAcgAVDdf/CoBx55zGEsl8eg7Di4EwwUoY7ksswnB/uT6HtBwjytWp7Z7WzDwmP3HJ4ASMpxgFqAEfsfddH6NBab8xW3bHIM9vmUblNu4x1XnTOAAu/JCqBPrwygyzipVZp4850vepGsO9zSoPQFxpodc+G5J+1dI4NO5KmXJQEih5/9qxUOa7JZnocGML3sQTZjzXmnLYczHSyDA2CGFTZeds63+5G3sSSDpkYqKgdwG2rtkJGLV6zLUJQDVAzbGk4Pca22Z556eOrUe9esnNfWuOLNl6b96PuHR/z2722gx7UCUD5qwt7Z9UtPP+4bNG6hnXvZ5V8sW9meNR566JEY3R6cNeyRcAWoQ0Ab/lmTNyCHYQ/yjYy1s/Ylpx4yrhwggXOQo1A2+oOlmXZOyzjIYFOA/4KNQb7w1hqYgZuxwrJ3pt04gGZN+7/MGkA5DVXUnHjuNXki4RyRME+a1UHpbfZGkPA2k1acWCggFBAKCAWEAkKB7aIA7yvjpfKNn78wfURSkwGqImSywIDSUVArQR9y1jX3rnDQ7IDZPm0evhndpG0eXYf6yn1uE+Z4ZIe6+zwzqhczCwvdiaAoL4mZZXaBOUGUaNKRs0ER/T1x+fe7vnof+1/tesAutMdXiRe5iFA8lw5WMTRPmM+15K+eE6gev+H/XlhLTpib9IvmumGQY1refvP1k044iDvLEl1HASqmv73AH8vgR/FRjNBtDnZvYmNXIWGMGsbV95MnF5jVNvfZB6YMwDm1OoAWSYBeBoigOsGwpJKJkW6BxH1jJa0MIHbE4d/7aO5abN0Eprm09zS4tmObFKKK+2EjK+NFw/OE/eaPYMs/ooMKrHU5a1l2+qH7l5OtkpI5jYCqb85djAmEA29bTpHBCfnNIVazPn36/4bR6AmCooZn0SVQIQkweuzk76yjtMbMdTGlklcils9zb44Mc5bdfvX3KuJkXsUw0Sok6hK7HThvTRs/2m8FoSL7wwoYt8At4IoZ1z4+dmJFTRlE4qokR0FOghIHJa4n+o/d+5BPNxje+AveCQdLgkLZLsu5mPWpwMwNbvvHJx42PE4jEZyEFZBA0UAuK+s/4jdX/gxrN2TUP16fg1Duh/ta/e6LUwaoOoAqQbyyP0QHAgy86IZ/NdFAnYHDCBlWWM2aF95785WV6PIOWiIBUPXQqwv8YTw+XBIEuwo/Cw6RcuN5x+8fo4LR6IgCkIhH6gAGQXLyrP+uzvje0aHfRl+2bfZXkPA2k1acWCggFBAKCAWEAkKB7aIAulwGtqV88z9vuFqlOXIJBdBIhSSsg1apjf7mo7PXrXVYC49Ca5EBysH4zpTqBV+3BQljh7O32WWpc+paaLLGbm6Q8YnmoDo0TZO/hoQtBbShQzyGKXXkzr9/s0gY0UvCtZjtVtf3Pu7sM352w1kX/+SCi87veT3r7DOvvPJnZ575/d12Hx2Px2PoXYqLRquC806rv1IkTEGDvdaESOkvOFKTqZ/38mOT68rLydKnRRKY0xszmmkq+hejn28Qt5nuSkSCalUZHNGGXXrhdevXWG3NDHHSXxzTxcDR1CL5U1OahK0iZFppZm/81x+u4O7r8YikJ2pAGXv/88tbqfRhLAsGnzi0Ubh2unz9hwcN9PIkg0be0ZIcgaQEdYmKPf+7dJ2HjuFhLcac9Abmbvxk1p95tHo0h2sJUKtgwF5X3jWzoyW5E4Z74Z5djESGK7NTLLts9VuPDiSncxlzECUxM5wWpzTp5Uf+6FfraWIF6ePQdGfUysbpz9xRpJWx9Y/87bIkASd3i0CrshwBuVyKD0ITt5z8w+PPrqPfTJ+ETeY0tM1/88LTj4+hR7sE0QEnX3zNygJ6LFNW4hyzmxirX/T6Q0OSEMNcwTpAbO+TLm4kfKXyWMyPie/fTO8v3UGDscKbM//fIDk0VRhicaU/WobVMd844ZI2GmrpqFinM/X9W0HCfa+pOKNQQCggFBAKCAWEAttZAc5ydNE8yyw+7Rtj4gAx7P/FQY2qsizJOkQHDpx85GLykcapihQOlREJE3Ry9CQ7i9/XD/5uZXU2k4T9SGBeOGL/rSBhj+g3YRNGbFVkymTqe2LqEkhJKBsEWlTWpJ5XhF5N0TQEYG5SVmWI6t7cSw2NhP14FiXuPu23jcBX1t/Rq7+7gE04TMJ8WCf0XLhuIb1u8UdH7TOuRsOcvDE9oqkxVVK7I2H0W1fk/pT0G29SNNr/H/dMy6DVNrS4XsZv7iUdBleOvtyEyzB7MBpnkQ2t1vWfvzm6HG2VMYw1hSc/+pxrGhjGqbJLGBn5bwCPN42XbPno16fvg0bLWJmsAFp2lShI1SoMUpShL3602AN2P9NtAeNFp5hTb6x/7/A9qsoCe7Kug5xUJhy9iGKO09ABxiPwbcB8g7+iT0ORhN0cyzex3OrrL/1WJY3l4c+XlgQ9judWyiA+9k8PvEqBx/yURjRqYGEMQG5hzzG2+LNX7+VJ1Sn2FrViKQZqOUg1AHUX/fwvWcZaMGgCX2w0R2eb8Acxny6kUxvWN9Q3tKcoGxUNDhKlO22Mrbn6zG9i6OeooiUqIDrgjqmvNHMSRsZ3eiRh9Ihn+SX7jYpFsDJkboeoJvcnv2x8pt6Y32RSSDMqVsdnyytq3/8RJNz3moozCgWEAkIBoYBQQCiw4xQoMNa28dPXRldG0EdT1kHWIhIkNQ0NI1B+3q//ksd0pp6HI8Kmi9mViqvbTVjprayOIOGtNDtvlk2YOEhHjEE3XZopjCly0CzWSxLGmaIRTVExZJEiQ1RDJOY2YQ0nNw4QJOzSY4MYisBaYIWWa396bj8aeahQUfcwCROa8oEFdKlV9KQSiRJraqoy4OCDT168JEgthoHRvZUsuUHLCQiSIAnBy5tQmmtg+XWXfv8YGSCmybFEBYCiJkdMe33BWo6HpamKn9yzGKe++PS5fyYBynQ9IoGKM2ajIFcCDCmrnjRn6QZvji/3tKZXJ9vK2MoH77iUj5hURvyURcl+J1512/w8TpzlrIvzZjHHUlCJbkgYf4KQ2ttzi56cVI7tTJZ1xGAkYQA5BnLtnkf8gHsRoznXn/1sMdNCEiYn5MKC5bMfrqTA3f44EOZlQvOyPGDQnqcvbWV5k1mFIBkb4idjrpXPMZPKGzaguwXm0DRmp2XeC/cNp3iEdDdjicHjP1+ba6XMUVhLPmzX3Q8ll980Wpi55HeXnlQdicpoV8YMTiBX02xhXYsNvPSGB9YxrAktpe9Zd5fg+zCh2WYugoQ3UzBxuFBAKCAUEAoIBYQCO60CxLf51gyzjcf/9rtqsqvIAFEZKiNaVJLUSBkoAx56/iMMxUxLa0s7mmvcAuUKyTG3gP3RLjC8fWvMvSiDXrpvAg33UEMFClCh00boEM+W2umAXejtVpEwd5aOVlDPW+WO0z29IkhhHiC+8L9l8Y4k/FWaJ9zJJhxuV8E8Xs4gH7/xwt6DI0mAChXikSha1xWMweQvlNUME5v5+a1kBaQEQMWgIeMeeOhJl7FUFhEvZ3BjL16KTwX3Paa9lsyvSw+xy1jzglemDo5RXAAAJVYN8sAJh5y+ugGHuHqz8OQ9eGR707nHHTCAz4PFTEiSoicBBp920c1toRMVrf/5FGuae8xelTHKYJQAiKMlWYKKgQ+/vaCJflP4EJvjGN2RcJGH+ehcDqdBZ5m94JaLDq7gqkkqqBi/C6NBKxHQyma8tqCFTOEuulfjtBAbzeOGYxnMzTBnybpPZvSTMUAZxrTCxSfh6JC/PT0fLduIzDhIQJOuDW7WDdWP5mjgewuDWlvNOMvXajjrmMk1/F5ihuayPQ49JctYHmeNb3qh5pFlbM1jf/vfBFrbYzSbW8MIDnIyArIkJaec+LOVDjMIxjd9Rn/KiWliK+AMbFu9KUuHcwsS7iCHeCMUEAoIBYQCQgGhwC6oALePUKfMYQa33TQuuOTEvYA6XDFFwaQh1C8HtXrM3t9a3eDmTHRMxMU1mZujlEUptG65OysJl7gxpYA2fHipY3aV/ZtPwh28o+W49r2Lr/zn5+ATawAAIABJREFUtP88OvOJaTMe73l97LFH/37PXVdf/fPvnXHaqJF1ABCPQ3ky8I4mm7AgYa95IeNwnsMddpoV6h+/87oxg/rhoyfLuqopFDeLkAzj1+HKJ3B7c4hVkKKS3g/k6nN/fFVjm5t3WTMla+YmX07C4YbKr8wjz+HFU4t/dtK+STII41W0Gojs9oOr7kQM5l8ryUc8LxlOb/X4y7CctZ8dMb66nKO6AvGqmoNPvXxeA+PGT87A3CrMizF35j9HapQySia8kwBiSaiu+2BVSwudlgA4wHk+JaMby7BHwi5zHItZS9+bdlMtTuoFGb2J0Tcah3DQTB079LSfNRNjYwE8EjbCJLz+s1kDtSIJU0IlHdRybeC4V5dgCjccXchzhGaON9TAaxO8csks5mRwhrC7ftnsJ+vK0dSPWA0axOt+ftt9/nBi8K2eN1xmLJv/xlSaBM0dLHSKB4bu7JKUrNrjpAV8vIFnjOv5ZD4Jo1eCiQMDhYLpOszx4txt6sv+54KEfSXEX6GAUEAoIBQQCggFdkkFqHeKIamw9+Yw1t7agn+dZvOLd/YfEq2iPjjZbHB2WiRRA5C4+rd3klGEGaaLJOxkGGtirAVJ2Ol77+jN1JXXpetr96cJQ0J4O3x0eP+uuO0zDS97oEyoFx+KHY2Gqy4Rs665+/FVpg88YWm6bKcz7UgjuLiZVEt7W8Mvf3lJIgHxCBrZFO4d3SGLktf2upypNzt2vXnCvFaeIyp/5jDYnGG1b/QiDWea161YdtXProhHolXlFR3MwpyB8VXGlaIsE1wlAWoAYmP3OHh9K95WjqaBKbgrzPokbBYWvDiUnnEtGkFneKkC4ns//OIyuoUGTl4tJjDudEf4CJr3u1EoFDJNa5ExM/WvTXv0yp/85Lc3XT9/XdM6Cp1sUagqXh5vajFB2M9POL4MMIoa53qE1eSgmknfaDTQ55iCO5shw3bPJIzRsLGpF1az1MJvTxmJdl2kTwzpTA7nAHIiMnL/jxrY+nCYMWY63LPaSTN76YZ5TwyOIkXTV0HX0DgPkarBe+yzkfK3FQjM/Sjc9Gyhb3OxWkWZ0DW6heU/v++3Z+NJFKLxSBL0oY+8+hkVtlDI27mCWyiYhULBMKxgpT0m35/KtpMI6ZaF73JLtQKaBBFQVVBlFRtDOZTt8Z8P02SvLl5/k1s4z4V1B8C9eygFCW9SYXGAUEAoIBQQCggFhAI7swLU5aHgzLyfir3ofDN2yp3mtx790zCaZKeoZdT1pi4zJGR90Buzl67LsIyL84SJhFvQLOyYmNTUCUICeRvbt/4B6XXa6L4Upcg2fHSpY3aV/ZtFwpxJPIsjv+GR5B3T30mRJa1TdLROby3XCvInobsp4rCVza974ME/Vcc58CRBGjTjnYUpKhOJ7HW6PTgM677p7V2chLlnqmswJ8/sjNm2AScOF7I035UtXbzke989LaZHMJEPX2RCRoksw0C7KayToiZoInelFBk0do+jPlmcacqzNNkrufOu/2hzQV0+VGEiMqee+NPPKsjWTAHOJMTSigM/a0GrJ5o+bYuiAnR7J4LnCz9NpVuIdvOs0MyPzpOjiJ/syZuWSyWhgNNWm7Nm6aRYsooCNXvjL7IKkUFTTji3jUCcu0b3koT9BuW0Na9krOXeay4cQqJh8HsiYSRbOQHQ/6aHX99gsVTRJksJkh2bEQlv/PzpwVFI0ExczP6FJCyD3n/CQcesNnONzMlRsTwS9n4CKHw+DiZ6Yw78j2ORBqmPf/LtCWgMjmGeqFi/oRAbBlqtommUSlqVpagsy7ICCi46X2UkXL7KmGSavLVryPOcTMtIwpKiSgp+FY3D0pD7Hp/T3hoUodtbFt5pvfD09MGDB0djFeMnTvl80bKiHsGN7TqCEj4BY4KEO+oh3gkFhAJCAaGAUEAosIspwLty4UJzU4+J/djMgmt/+l2eUwRNEOTGyIMnjZ5yyCdt7jrspbrMyXFrMGFwR+zipw93rXqzHS5On20XL7yrEGxflbPjLQl06N4mjBGKaaX8PTS1USu/6eHX11FW007oy9/SzHAM8EsY7JeaO8+i030bY/UXf2ffQTIktEpQh3AS9oL7+PzwpSZhXxPP4TioK99vkWOFyWzDdQzXtsLryy/+Z8K4sWiVjOloVETrqQwQUQH5iBL3klUYWSkKUAnK4EFjDv3wC5eDTc5l6NBLNEl/TcZyZiFFkajqmbP4wm9PGRvDM2qxcnIGiA476rIFDmYAQlshDwlPoY27ppIm6vJoyQ/HbDDXICTG/YENl4fYI2Oqi3mLWIrll3466x6evthLtIXeyzrodd//xZ+z9F06ZyBd0G5LbXD7JsPxOKtl/rN/HwYQpTEEXUbnfBVHd+Kg7XbYD3+7xvSiRVMVfBK208xa2jjv6SExKJPRBQb5U9YxPrNUV/e1o+od1sRcDBnoTRKmb3ve3kE5iz9YjmXj4Wvf/1oZYDYx+odW98hAiKIDvCKBgr+rOvfCkEILUbz/IqF7N8XI8u61ApoCOGcFaRnT3ckAiSt+cXMh72bTnYKJF8vDvcHpfWHB29P6yxCPqphiSk4OGDFpeR4fVKyZkw/SXwe1Cm8EZxQkHEghNoQCQgGhgFBAKCAU2BUV4D2cbktumpkVLas/OGj0wEEK6BiqNMGtR3K0DLSK713zp7XcD5Ocooud/HCniW+X6ruW2t9tcbZ2Z/FiXQv45d6zWSSskV8oWat4vGIJ1PKbHn5zS0iYDGcOyzK2fs0Hj46OYQIYgNqpr89vZiyPogc3BZvP5i+7ik24a/vidfX280EEx7H4GiZhblr/29/u1HVVkjGGFsZ0h7gGGh+aihAP04iFCpSkCvTRIyactHgtjjGkyaE34xtMKdBxzjLT5Im9qH3dS6PK8Du08AS68WHHX/E5Q37GmO29ImG8iWFODhNycE8DCfBTt41l5z93++U8gzEVngBf1SE66vu/+Atl4uVNIvhe0FRKbPCJzQipLrPaMp8+MSmGgKnGkIF1b+Z7FPSxw/Y7Yx3Zq6lsLnG7iTWws8xc3vT5M2ESVtECXwEwvG7ScesYw3ZLX/Bswrwsnry8qLzGOBDgmC6z8item7ZbhChYA1mKox+7PBjUQaDEFEXXpLgq6ZIiSyq3CZNhWCFTL5p7cZEUWY14tmJVUhVQNDmqSXFNSWhqRNMkTVMiWsVZZ15E1w4XI5Df/8QbuMhdf+mx1fy2K3GUPjbwV/9vJoUgt5lLXtaUoDhQP7wRnFSQcCCF2BAKCAWEAkIBoYBQ4EupgPHx0w/UYfgZCcNmqWQ3KtOhaiCUT3hyzlq07LnkMOiW7jqV6Lt65sCun24TIYuXCffqvgrbW0rCaKTD+73FJFw0l7Uye95RU6oUDOE75PE3F7b5c1kDavqKkXCHdtcDCQf3bsXSBcccOCkBkMAEV5h7GKN00x1SJcxWRUsUIKklRoEy8tjv/mxpG2siCyxFRfID4+FTRxmD2PyFb91TrZOJGdSIFNeVCKhVu53yyyUdSBhjb4VBd5PbdE87P8O8wrTXZU4bKyy6+ZzDqrxiU9nR4K31FQmzxveP2Q0pNl4uyyrmnMKgy2oUpDqldp/FOTJ6Y2lCeZGtDDOWNs9/fmgcysnqKkkSkXAZJ+EGPk/Ys1cTqIdImMvit2f8AG3CZmbazVcP5vZgFSjy4CDQ9/t4KUsbFLTZQGA2bde0bSe02JYdrPgRWejRg9t0bcPGqeUWft3LY+e6uMHTQ1s8I1R3z1MxWFnhzCMnVPImAxBLJEFKfP9/79pAVaOpMWSPDxpfx43g1goSDqQQG0IBoYBQQCggFBAKfAkVcDJ5tm7eXVecVQOyKunoyqeRG2YkkdztiFc/bkYSdgqua9sMe2PYPey6FiE03FEsvb1NhCwWomsBv9x7Ot6SQIfuvaMDmzA30iEJK1tkE/YYw6HxjiwzP/vjr07Dvrc6eMZbC9oFCYeek02RMCsUcmiJLDT+35+ujUuY+JfcpL1EPxLN6kbsU3SQkxhKWhoI0P+2h/6zKhc8SwWyZQZv2xib986MawAgmVATih4DvSySAK16wum/Wb61JBxcpdsNImHji4sPGdUNCevDuU3YS3dcVClotyU2OtqEWXbehUcMr4qBFAWJZ2fm+kAFlI17a2UzhX8OkzBjRMItC/7TiYQljJ+FNuGNGP8KbcJ83jWyJi+L94DheIG/Cz/AecJm5sbzv1MNEMFhJZlIeAgoB324CFMo4UKe1raLluzwEo610OGjIDhXx+PRn9nGENDhk3Te9kvL3MKvLzq1vwq614IUUCtv/NcLaY+E80TY+O1ufxiD0woSDqQQG0IBoYBQQCggFBAKfNkUoGSTjGVWsaWz9x29G0JRXIJyGaIKJKunvjgvS66IzDVtHju1235Th85huKNYensbChl0Brvv5JWqQZ/s77lafXKJbk/SsUPLFcB+NM9Og6UKxY7uOxLmMYwtShNkMLbm+Rm3YJzbyJBHX/2ouUjCWGQ+j7Rnfbr7dNf1ju58ozgMd33teO8clm+c++ZDe49NRAEikq5EcNYCrt6kYRlhGN9E0I2j/5gPVjTbPGgVTowNYZLTyNyPp/3pbDQmyxAHJQpyVbIclMq9z7lx2XYg4bZPDquFSj8WGA6RcJuwWnfsj69LUVnJrBkIFbTbUhskFfdatttZdt4fzv1mZZSM5lwcDSQ0olfK8ZGPzV6w3qM+mzzGKfWQmWaFJc3zn6+LQrkEKtmrFZyZmwQYNnTit9c66B0dmFw3RcJIumbjmguP3K8fj0QtK5STqT/Ej3jlPXwsTMO7vbyS3bXwvt7nJccy5735TLUGsSjgjznExx10/Nq0jRMW6AAsmz8XOrgBwUZQJkHCgRRiQyggFBAKCAWEAkKBL5sCbW0UQiW75tpzTuY5RSARwTWSPP/XN7aQ46VN4aM5U+3cNmF+dwQJByCxHUg4j8CLKXQMxjbO/3AWhnyK1T4/d1mOsZzLDMNDAUHCXRmY7+lIwnz4YtWG5S8fsf/YyoSKAMlJmL/iJGLf5xUUiPe74Oe/R8ueyW86z8SLgyCO28jY+/ddd7yGx2NUqChAHKcel084/TdfbC8SrvAImMrsk/A3v//z1q0n4fZPnv/b5dVxUGI0MkB+5JKiKhDXY3VTZy+u70zCjBlZVljWAwmvsTBZ3OaQsMPa1p53yF4DAxJG830VxPab9fJGTLvkwyVv/905NPf+/xTu/d6L4/FiJjPa57338tFHH1A7aujRZ13y3vLmAmP5bI7bqE0aqOKjCn4Zi0OHwTUECQdSiA2hgFBAKCAUEAoIBb5kCpiMNbLCqll331KDBqioHOHZhWOHnfHTeht7hKlCivrrmEwz3LErdpp4NyqAr15u9IGQ/MJdT8RLgPu79vC25R7uOcnL033Z/KvjjL/wcf7+LS9wx7oG92Cbk7CDeW0NTMNj5hlram2YN2vGY4/MeG51jq230MuUFqyfIOFNkTA2WMrK4zIM8dtqFBadcOzECEBlmYIUGyMzME3r5tO7ORD3Hzlho8mKLtL41BbILbeVOe/84X/24TmCOEfjV5TKwcf+z8KtImHvvpb+4zI7xcgmzEm4Q8SsyMiJR5/baGOZt8ombCya+ocfeVGpEYMxAZEsyzrosdigR2YvXeWZyE3fJuyT8OcvFG3CmEuJMkuRTTggYf4I9cImXGD5pjMPmDgoTMJSDOLD7532hhc7nXDTZpbJDG9ubmnhunxCPxTerwW343Y5pOsOHJmyWSFPAuNTyB008DnkFmOX3Dg6/mqEf4WCUwoSDqQQG0IBoYBQQCggFBAK7KIKcAzx+lNYB8cmJ7kMY2tenfXnfjrGTVL0JOg1ABXDxx/2RSPLUF1dlmGOjVFXaZ5bYOLoDG0BfPVyY0uFpDpwwxevQtcT8RJQLUt39cLdvj7a5iTMT8bHDUIWIIo1RrlcGAXB8QrfR5fuRP3BPdjWJOxQRlsLc9LaPLlsmjHHcjFYUTtlJzYskyde+hKTMI0Rmcy1HN/jNDQroHiHe0HCXmO2XMt22l13ZT4z74QjJqsAMk+wo6MnsEyJcHl2JSRbvfxPj73QwGcxYLPKM5Y38DlJM+utq88ZG8UJxzr3qMbj5Yp+B1+wyCPhTWRRKhE6q+tDF5pQyz+0U6zp48NqIUG8jiQs0bR0JQqRkaMOOH2dhbhPxB5IFLTbUhv0PAfe0bnPv3j+r+jJ4kUV4yRMNuH4kKmzF/k24W1Kwllmtp7+9d0Hk1kaY35jeDId4rW3/Otxz1XdI2FM/bSFJMzVRbU8P/hubwD/FcCPuKK4YTI3xVihjR7GFOX/xnzw/iGlNoLzCxIOpBAbQgGhgFBAKCAUEArsmgq4FiYEdrEXRWYKByNgpVuY1WSsfGvf3YiDVW5sqgZ9/IxXVmcDiMNgpth14lFMS/ahSvVdS+0vJWTQhwsOcJiTNhiGVmWZfIq6eWj1ol50wPb8Mvw7xe1S/by+3e+XlJ+VT501uNI8QQ72Y0l3njmGAgXZpKj/1S5/N7eEHXu2gejbnIS962LjCCqO5l+yFDMDTU/uV4CEDWZhKCKTsY2WbRAH+7fUKQXAwX7/3vnf8P66pGJLU8MX+31t1EAFTcKahhZQPx20ooKma1FQ49+65KZGhmozy8W8326hgLa/DDPeuuqMkZhxV4roIGFOW00HqICqA+Y08dDKOYIlbG4loLfb/Z2KykENn8piq7ZybOOH36zG2cxI3zICPIFiOSjDAQbPS7MNVEO/+r0JNuDNccXLOylWWLjmjftqK/0EzLKEECpHJL0OEmNe+mJFI2XOZWiJzaPzgsNYIc1yS5vm/WdoBGNkoSc/fZsM7nWDJxy72mQNNJbAH6FN2oQdO82M5uPHD63tQMJxiAw74cJfUagsl8fcKs7Y7ypeT3vwJ87/Md4ECQcxsr3zofnXpXjT3kn88/R0vU6fCRLuJIh4KxQQCggFhAJCAaHArqYAggr2j3mnCu3BzGKZDaxxwcUnTMYcoxGARDlabdTBdzw8ey03B3tHh2CnBz4L4KuXG6Uk5Jfo+ikZsTMuzka10es2i1UIQJh3Nr1v8RLgmx7K24cfhQob9EWp30uaO3Y2k2rLZ+yNLeaaAqIpzULckSTsoxRPJkwjIFseO7p7IW3sgHtr4BpNBsCQWr3a3GTErBfR8ob3nIvfLblt252ITrn1395v4qiJk395+91ZzymXtz87IN5SG3477SyHYWbb0xsdll4+//3xca0cIBHRJdWLnKVhrOKYpMigJCeccMVq7hJtOszOOw5OzkYXa+Pt3184WUcE1XSKr4wkrFRAcso/X1ycxkcph+4hBEpbR8I2GR2RhL1qWDnW9OF3JsaweREG+ySclKEWoPaJeWvrKUqzX/3NJ+Hc5/Vv3h+TyC9ZlnBBEo5BYjz033N+W1sTkjkO0Gw7EnadLDOaz9h3/ACyTJPhGyfKg9x/+D7HNpITMhek1A9b57ve+T3p6f2kbWKecPjXxztN8Gvc+bS9fS9IuLdKieOEAkIBoYBQQCggFNg5FQh6SFg8l+WzZCuwVt94yVFVNPcwiqYRHZKTTvrp3zYYfvzZ7hkn1HENHxB0uXq5UUqpDh1G6gXaDktzuEDWwVjWaHHkU+D4WYJLht/idriA2267RFUo+ZTTylh64YIPTzv1DDlS8/LHy9B2hzfB4oUu8d3NLnnHuhYFKcrpx44mX1mPpri5Drvv25KEeRm4R3up+pbeH5DwwIQ3JRTDLiF36JUgDbn1gZ2BhNMsvfC4vUcAxKvH7PdZJoezdl3GsnmcWeBuYvXvXWcNDDNLp8mxbONff/7jWsIsINsjuUZH6NkF0MrUscctsRD4sFm5BZrgz5jbzNw5d//m6KgEkqR4JKyqoFdDZLfvXn5rGz1FPH8P3aOS4wV+CYNnqFNRvTjERYMwusJnWH7uFSeNxiDXalIFCQ2wOAdDkqQkKANvePTV9VtJwoWFn866FUOKKSBJksIvgSS8+6C9jq933DS6E6O3QhcSfi5sE6Y5zDGAIVtgE8an2U7/5NgDqj0SDmz2MYjXPj9/Q1swLNJJs8166/9Y9GjUxYN83x1/PKL4S7BZ1yseLEi4qIXYEgoIBYQCQgGhgFBgV1TAph6nF77I6yOZM+/4RRX1qpM6ggVAZOi+P3hvneeJV3I+cNAZ7rQRdLl6uVFKR24Zog5svtCOR2Val8x++/D99xwxesT/3nqXn5vHthz0wqUluGT4LW53KuM2esuv6r8GF7EcO71q1fyTTz5c10ChnDcP/uedVs8mbBaxwf9m+G9wll5udKxrURD+dTyza895/v4BMoZeCtmEuVkYdmoSzjey9NwfH1aShNH+iTXegTbhdpae9/sLT4jKCZD6//7RJ9tpJr7T3rY5JOw3Z68duLZtmBZFwnIz9sr3hvPJ/IriZVMKRjO0Miib8nmmCwk7jYx9/H+3npHAVLeYilhFZNQgUgHywN2+eWozziTGcRm/jW0pCbs484LPkfbqgGfMMPfTGy/+hg5QplaqEFHQ5YTzogxK1fk3/nvDVpIwW/Hi36+sIPdrnnwZmVaOgj7s66ddtsa0vJkgAQm7zv9n7zvgtCjO/2d33369cAccx9FRsFfsxthNLEks0ShFBY2aon8j0aiJxhSN+ZmoSTSJsRdAIcZC7BqNioqogNTjOK7C1bdvnf/neZ7dffd9732P9+AOjrjL8t68887OzjwzO/t852lc7uWJdR1f5ELCp6B2tO3pDUeCHiaTSkAimm3Wp8aN2A3nwpYiuvkmJCyB1jrz3frk25tx8y6tInN88/yDtzcMS88ZtinJtXWf60li3MeHgtX+PuXzynCRcF5kcgu5FHAp4FLApYBLAZcCw5YCNhJGDy4a11ob//vcgVW+QsaKRFYUBCDsK538n1W9Wyy4bBiWVMGEGRbjR3zVzn/2SyxEiQroPyudH7/4dCnKLuHDU3rUt2ZHkRdMOFuYozaLxbd5/SFJcM4TCSCcrstJpUvWOjmPt7ZsnPW97/pFocAPEncPK2RS5eL3VrZj40Gw3e8x0Jan4Ewap272F26FSHjErkLCudrfb6dz/JgkJDyqyHRATEDelgm/6kDCcNuBaPkOTmFuRHhszZ9/enEIJLalo2ecvT4V1ldNHxqbMBmdpXw0LIW5QQ7VrGfQiHLjy/OPH0nPAfm+AljLxECBnyES/rQ7hYQN069SB0+8/8krd5ZKECPcwwSP4INtENHPhBCTyl7+tBlDqPGEShtfOamR0QWksD21NB5u4cnuiMw70RQXOgYy4TjnX76/5DdjgqIfYjj5IWwvgXjwnlU05dS5G3VqqE2T7S4ruAoR2tbDPPbFC3+4hiYFOdOGtOBngbHX/XUJxis2qaoDbRDyy708CUh4tJ8VCKadMNYQZGzU6GmnNICd8ACRMI//Zu55FegpCwdIEpknIMGiOvmbP1zFOfg2gAPNBWztcTOz/z+GHoPdwGRS4arK9ciyt1557vWP2/vImUHvgCuJOEmgk6b9Nd1V5729imxNyEgMIgrbZ/+355y7SHi7JHILuBRwKeBSwKWASwGXAsOaAiSuSSIbCX5E1775tSkFZYwVmnIMgbGCxW+t7tJNFlDRQS0zdWyXQd2BAqnaM1Mo2jN0FH4Y0YYrzzyyHBQsGSsMgTpo2cTnP9jQidyqBhxdf4fN8A1pwm6BrsuK3tXc9uVFF51ZVOgB+0xQOpf8LBAUy5lnxL+WrbXCmYCf4X5aP9AGO5jbNJlV6i57MBLu5JHll3+tDxL2ljNWe+cjwwMJx9c/f98NJSDz9LHivf/wzAfbMBY3ihCzDqY9a+wEybTJv3i6X3QjyvVPLj97NKE+CwnD9PL4PYCEx56woieFhDE+kcZ5F5c/bvny6ZoQKAIEBF9xQQUqBAjgvUoq+Nr35rdy3oGzENWIsyPh9KkFfUEkDB7S8FKFR5tbVy6bNe/6hm7a4CHfxAqPru1a+9rUEm8phDL2AhImmTCot5ewCcf9pxFdF6Rw2XbXEastALbDPL7y/mvPJEksIWGgiOBjVdMWrtkG8Yp1EnirGpfRpAJlwsl121YiErY8ZiF8zY2EcXysIYS+p28Kgkx48d0316QjYRwpD/PsteDTWBd5z+Mp51X2kG8vYW6FaOSFrbPhhIP3Z8G6v7+yqo8LaV1T4pwbhtrFudzZseWNN177zW/vPvm0bxcW14ZKajY1bo3J5u6b1Zf+1h+7YS4StknhJlwKuBRwKeBSwKWAS4E9kgLEYmPTk4nmtReePKPajwFKPQGfWOL1j/zhLfc0y+SECmUXGYKL7TKoWQsQqbL+ZMm6+lLTRgOIhBM8vumY8b5KkoV5JCYEmW/kHY+/1wSOcbkMcqfcdVlss5PzG4q0YRgkE25oqJ87b47Pzzyobh7wsLLiQIGXBSXJD9sOZc+9v2qrKa/ZdUjYJLKFhH3oYddUUyXCAqIq/NVj/2lBqlpxdTRnwmnsmg9dc9G574hvPyeJSPj4mkLcWbCUuz3MiYTJuTliqt0jE5br17z6UFUQvSOz4r0OPWdd3AziiuSiKK5OqvTttz33FUsmbJUxolx9+6dzp/i9UL+FhGHwQOnZV1xx2IWrYiD0RHBDDqISgIS1FTz6wdH7lIUYC4p+v1QskiUEYFEvq57xwiolynlY0XIh4VxjrYKvcHKInuRbPj9ur1omVt/7+GsItlD+CM6r47yr4fSD6krQNTOYCotMEhn8E4qZVPvLx19FRGeTxYn/sy4cVnMICSe/nHfiONM3NW7qoUzYV3ny+Q2mTjIFLdJzIWFLn5lhA3PIhGEQwBADz75IWOdGoumd56f4wHE1HhK68KI6xx59/g31EQL8AFStEc3nL1IgFgZ/1+AdsPf2qy71MTbiwNNWxtOk1oqivPvuu3Mvm3PWGSfrHoHpAAAgAElEQVRPn1pVVS7g8oO7cLDslB8049TOiEI4WIGIeNQXnCzba4iLhLdHIfd3lwIuBVwKuBRwKeBSYPhQANkn4nWoURTHFaU3Kufhv//y+gpgReEQvAWMFZ534TUx8DMLmooqCHVkbiiDoB1t0yQHT2v/7kwQGkBxk4JmjNsuOe3ACpEVVYQwRIqHhap+t+Ddds57QOamDAckLMtya2vrnDlzxo8fP2fOrPfe+8/mTfWvLn2+qhh8cheCkTDziUVMLF/8wZfkMYukRLtAJpyij6F8+NI/RoigoEoCNAcMGN5IWO3l4RXzjqspJqBBZs4Cen4SLJnw7kXCPMK1hq41/64phA0mgFWBsdff/XCruetBUkR4KEk7A8NoOWc9pTOQsFUAHp8ol9/8zXWHFAQAwAbQU5bXcqDFvIWTT567TjbRkQ7OoEEKCo+P3sCjqy4989ByC5whFMXHXxBY2b6HfPsnvRhYGGFSxrLhREwZaRX1czGYmR697coLyxg7+NhzNkY4mTWbTU/Eudz7yG//X7UIUYW92Ab09SWBkNhf97WZ15OhMgJEU9bqaES2hYPyQL4a5vKGUyZ6kOA0MzCWrxia9fsHWux5D8+YAwkr3Ty5xpYJOx6BIGPVo6ef1KDw9gztaISNVv/7ImFwjMbbV+5b7LQT9onBAuYtZMJIVjjp0dc/6kbwaY1onn/R7ldJAKm13hVvLi2DPRB2zS//1mVaWJj1KIry3nvvzZk1+5Lvfuvw6ZXVFMEZIldh+Gnf+CtvuKczZnpVkFWIR2ef222Ki4S3SyK3gEsBlwIuBVwKuBRwKTA8KICSJ90AgAisLXKNCtdlrspoJvfucw9OLGAlJFuDkCNs8vSD2jqjPTIEKaKrKPgNCA4GdGTjWqkB8Emcl7NMWuUme0t5VBYMW9UeznufeuQPAmOixfhPP/CQhs5ojPOEARIkm6WzdRaty6Ey569Dl5ZVpTvc++lnK7D92Fv4SPz8anBWVOIFNt0rFDOxcvEHa00rRB1EM0Zu3dmBtja9r05CW7Q1kivfWliGMNjvh9g2NgyAeeAt/s0T77aiJC2HTHXwiZk2Bfr5osd4cvWsGZWFJhIGMShMCH8FE8fe9ejrKIK0pW2ku9v3c/Dbb48Rkm2zEf5k3xHMtmkvGXno5+vTlFgNw1ANTsiTwLBdg/W02bMYu0NDB4WSXP3s5ksPBv0CvxBgDO0EAFuCQ2Zf6OY/L9qsAwrVDDjJ2lg1VB0QcezTF/7q8GwMj7019IXMV3PT35e2mMS3WkFLh65yOk09aGdjZT3ZgQoE0WfuvL06APU98sJbURBDW6olEA45xrVw7+p39ypltUUFAYTuEBWZCYjlK7yjD/igESxYrRtDrx1xd23nT84E57qShOjNiZa3F073srIA6SVLQBXfKO+ovT/pjphh4OgBg8XQ0o5Wu7m8Ztuqf40MADi36EDy25xIuM9j6ny+OCycxtr5lxwBVuKMBcQgeM3yCMzvZ1IxC44uPfSM1zfHEpyHZfThlTnVrScU8p01484IipPr//ufKaWjGKsuG31EYwes7RbF8Bpb1GxEOd/48t9vpAcFpeVFrOLwx//TFsU3gnMI+0/bbXSRsE0KN+FSwKWASwGXAi4FXAoMbwoYpg0f6ToSW6VxVecxbvRuWfnOAaN9xWgeDKwSONoZ+foHHyU5j6pqwgBkSaeTzcq3w04Wrm+a2C5nflq99ENaFnCFekLXIvFE189/cSMTmT8YOOjww1atWysbumwhdSc/R7VQDtXl/HXo0kkIcgz+cBB7IyevG1xLPHTHVdWMFeG+g5eVMrHyuQ/XtpGccNchYRpSzo3kqrefrbA0OPcUJJyMy1wJ88Tqy48ZPYyRcILzFh5bfeZBY6oQhXj9BYxVff2UuVsjsMGEB7j/JSTcjXMgY0Kapew/9sMCSh1Jbqw/+5DiAM4lGwmLTAqWVLJA2ZL3Pu8wQCasGbC9AngYo+jqCueyyps+PGWfEot6UIXHBI8+JpUKdYc//tZalXNDVuIR9JjnvDVsKajpO06gD8z5Ns7b3330j+O8AmOhr33j4q0qYOBOdBBldlmJcaWLd274xdwLQoyVMVaFfqp8oOIdwHWo4rdPvLVJA/Rrae/aSDjDE7K1OGkKl0Ghm/PoS3/+1SQJ5KRSKMQkD4jiC6Zccv3dsCMAZNS4phJkVLiskMcsRMLtq7Mi4RGjp389UyYMJs8kB3YOl00g2uQLc77utUdvLgFJNxyFBaVMYILPy3w+WGZLxp4w+9pNSVggskFhqo0G3q7ZvJ3S09a0csVeZVVl0gjGRv3k14sgRlYcFpwshxHmiRUfL7ojKDIJfYwDkUcd/3YL2JIQfnZ2o5+0XbmLhG1SuAmXAi4FXAq4FHAp4FJgWFOAdC9RGqw6VPwS3NiW2PDG0XUieckqACmlxKQx/+8v/5IRBkN8XgNEP4OMhIHVsjjJDMkwENJm+5y8YAaFgXWkWDIq6vclFVUmjzzpYlBbmkbsHdXSD6s3iD9BC8HQEjz0oMI2RW9JLvnTzTUiiO9g22H3IGEN9ShJaKr9559/LsPGUOgkCxSjruxwlQkrCYUnunnX8gsOKC4g/AY4brjJhMGvL0+033jxmdVoxxwqCjFvCWPlV9z2AMgnIaCPynWNHkuCan1nYNrUJyG3+dREo2v/PbWM+VC5GdxfASkkxooYG3HQmXPak4Dz8Pk1uK7AjQwQg+pw2xiPfHHfLTNLGCsNlKOSMmopwEwQwb+UUBysmf7ye6uiOihVJxBlkbd50ym3bgeptRJ6hKsb/v3Iz2tLWShUwkL7LXnblCtHUCZsImE1wRMdnPc2ffTKfhVSNWOVPl9A8HkEjwAmwz7GCmsOPm1DnHeht2vagLMccWVHwkYyxtUITzbxrrXnn3A4SeB9BSj+LChnI2es6SEYDCJpOJGM6Uh49cCRMAmqbTfg9sKl6fDod8C+QPunR44M1IA1hB+oKvmYPwBgWGQimu1ecvPv29K1mq3hdq5+lDanRrR3W+v6zw6cNK7YV8JYSdHEU7YkeG9PAoea5odVB/01wjz++aI7rygNMMmPC49QfMBZP2o2dT0GoBZh1+siYZsUbsKlgEsBlwIuBVwKuBQY1hRwmAQDF2iKdrVerjRff8FRFPHSwj/Bs6+7q53zTkRtIBGyYLAtOhhYV23m0JnQdWDENYjbAcwx/gStMrk4Z1FKO+9JhaA4uk6FnzSDZ0QBccIJqoJyqCLnr0OXDkcj1u1oBFAipQMSrkbJV9ATACQsVS5alpIJk2XrYLUqfVvAJmwaEl7w55uDGFqWMQ8onFuoGDFRwfDVjtYivOndE0cD7MM2W0g4UMmEvtrRphCvj473AGDAQAcFRl/ReDz6wE0/GGkTFtxbeVnZpEXvrOScd3VsNTDWDWjFg/5F6nTejiaS+QlFFFCN1nuef+jnY4rAg3pxURAdQQsgAhWqWWjvR/7bQLHPQCAMgWeTXE9qBjz/iIQTXN4Uafxg2ijaBgFnWbaJcWFxKWC2QDHzlv/m/ie6EwAfQZCMqs9QLUBxCwCTQj+oHEfu/8XlZYz5fYwJJVf89G9bZZ5EHeI4mmaYKw8g4R4IMcV77ps/swZvLDGfIIkiyHIFRPQVv3tiaTt6rSc1aQsJ2x4DbBRqaTfwJO9YtXXFv8vQ6B1nrwfq9pbe8H9PteNeFBIQLzSRsGrKhPUwT6xsX/2v6iALCeBuzDp8jFWSTDgNr5oyYbsNNFap50vhiXCiFWTUsZZHbrpqJNljezwgDfYIzCd4MIRVMORlwVHn/uSuNHV5a5jtBRETtPnBe8Odqz//pKbEC8rWrJAVTnnqjbWtca6BC2inazHHlAF78vXXnrl/ETgD8MMoe8quv/fZMG6LOqfZdtN2pS4StknhJlwKuBRwKeBSwKWAS4FhTgHdlIQgMIqSaCay+bn/uxEc1QjM62GS38d8JROOOGltewRdZAFC6CtfcLJKefXZZg6dCU3l8W3LXn52wXMvfbhqM2Fsm5+12L60Cxx4HPPTW+ZsVd80VUT51Oa+ZYY0B+2Wrf7pyUX3/rQSkTA4kxZKmVi96MP1pnY0KrAOvZ2whuIjWVchnMzDd88vYizoBc1okZwcWziAseDsW/4+DO2EFcXg0S7e8sFxlelIGDzjVjCx7s5HXiWRN4w4zpYhHeKslcOtZc7j2juP/3kE4j2gK3qI8hSVSCWjnn7j03bwpGwdWpLrGtn02p8q1kG2+ogkySuVxo3e3oZPZkwrK5KYF6W5YBHr9QMS9k/a77z56zjvIW1gGwnzmMHjGke1Y5UbyYgRbrnv1h+O8kIQ3YyDNkQKi8GUeMrBX3/5v2t6cPXoiBjRmKbEEkYigTJtIG64s/M/b74xaVxtVVlQAihbMW7ycV+sa4shco5jF7AtOBaawtUE12Jc7eRdX5y6XzUpDwseJnrQ+xPzME/ZhBmnre4AhZQujYcVrmiGAVrNCpzkoDr1KfN4F0+0cm3L/POPtzoigsW4MPr4c6+OQQP0BDbD3HFD/RGFAxJW4SZhnlzZsX5puY8FaaRMgvgYqyqqOWhtjLcibrQD8FJPrHXJuVgZOvhfiKk8Curicpg3fT49xEZg1wo8uNvklZhX8nqY6BVYoJT5yw858rjly1fEYrF43J4O9hpnbiAYXOnobDv/gm8zxkYERY8vwMQx59z45zCOizWHMv7S1I/y5mXTJAYicikIGw0V495Y2x62CJJ19mbNtGt3kbBNCjfhUsClgEsBlwIuBVwKDHMKqKgHRyIiDmwoj3684C+TA+hZJ1DAPCgoqJ6yvDkcsxgwmxNy9s3ONLkz529Z02ksogVytQRPtlz/vTMZC1x10+9I5kPxUrGOjGsIQ9Lvps+vjFs5W9U3vduRMMohM5FwAXgH8klCOZOqFy5b38K5CvaCll+jdGFu307lmZNejZOwYCFs6DFuRK6dfXalCPAlQzsaswJfn3Vr1NSi7Otuyo4iM5iS1YzBzfoV1Hw5N1YvnVEESFgCiSvKhAUvROIJTH7wuQ9TSBgna54UG8RigMAVzpPGqqVPjcH2AYnJVzKkBFYw9rE3V21TLT0NJabruqqDBb9KZr04HVCdFzA1gknSdw5z3nvDvLOKJIZ+qaC60pDP5w0wVinWHvNGvULhaskGFYzVAe8lDB4F6SyJRekZjmw6eVpZKWMlttQaKgNqmmbDzMu85WJJnX/E5Mt/+LPXXn93y+bGZCS2tbl59eefvf/OO7ffektdzWiaPAx8NgeLqw57/T+bVUDcIPkNk+oHbUmYHuFUcNin9/KOlU0rXphUBZbOHkTCkmhJpgMj5/7sHvJdDxs2GjfkOEJo2cTDNirWEzy5lXdvWLno/hrGQgLWECxj4oiaQ89d2cUjekLnESIj7btxRMLoMhC1X7ROrq2JNL5ZjMrZoLZsImEvE0eJZVPXxTm4Aktb9cDyOhcSVnhC5XEQ3atxvnX14nt/Vo0ULmOsQEK9aK8keiTYevIGkGhCKFR48MEHL1y4cNu2bbFYjPZvNFXTNLm1dcuHy947+5xvFBUHSacaRkgqPObKX3yBew2cg35NWuvgsaHnnXM9uvGfD0xFpXnmKWTeyvJ9j1oXNSDEcPoCsd3Jbz+MLhK2SeEmXAq4FHAp4FLApYBLgWFDARvsIF9EnI0GXKcM/C/oOic4b4uvfuWompIK4F4l8N1SOIaNmv78R2st+ZmJhvtyV05Waft9zrABtiGpEuPda06cXM6EwpnX/ppYfOJTsU67D5Qg99VJFJ4lTTiQfu++GpPUTlJetG/rbPygpgki9oMGdYtpBn+/JBPeVUiYKEXdTRFWV7VEvIfzSM+21QeOra5OU4lGJAR+biGKUsFep33WBU6P+ugVb7fX/RCkv5/Sxzb3t0TX2iX3Tkb+HpEwOD/2BwohIzjllj8+m0LCJBgbINO/8zMEmo7oJLrmzSlFACwBB4MOBispgMg3AOHLan5x7yPbVLLmNTGwEwmTz2ew2Ac8qcNfuZ3HNj1z/w0VAgt5bMxGoyax0ukL322McN4c0XrxaSGPWfBowwGi1ZQGvhbhybbWdxYcUszGMDa2mPCfhHidKsQwv5gUGCsvLikOhjwQrlgwoSLeXzR1ChjzB71VYx/+17KtSQ6Bp3XwyRQxH9q+2tQJrm/jeuMHr/9tXDULihhwG8eSiUFWOJp5R/3+yX9vtaaAlkxwTYaTMDB9wrIW43pbfN1/TptYXkehpIICKxoVHDfjv00gcu/QYGMBiIdK2jb8IyQMMmG1gxtr3n/pnhAMDmysWIeXSSOZr/apd5Zv5SBSh8NcHftBwhCwSuVJVGKPc97Nk41XnnTYGMbKGbjoLxSYn6A/wGAR1MLxgPVYFIPBYEVFRd242r322qtiRDWYTjMWLPR5g5I/6AP8zBgrrht7xBkrE+SKDKIAOFxtW/QygbrBtfj9V3x7EmPlHg/sEwVrTpp3QyfnPdC+/p7Evo+AXbWLhG1SuAmXAi4FXAq4FHAp4FJgeFAghTyBh8EIosANKcA+6ygJkrnazbd+Muvo8eWMlXkLBA96yxqx7/d/9bcI570ajyVjmpaw4wZnyBmcvNH2+0zGhNAQLGs3T0nwNW9MDUGQnqtu+wtw6qR5aNZoQVcI4YSefsxOkEDY5ENTZVHTl5RIbemZbcBHps72nZ3t38k0NsCkM6LEfnjK3YuE7Y7ajDz4K+Zc1Y1tTz12T6nACjHuTqHtSBhxgCAyIVjCvLV3PvY6mRSSsaitqbu9XvdDkP5+MkcW/9jz0JnJOU8mk1yP3TLr9PEMwt7YSNjjCzFPKZNqvnHxdSkknI3dJ5/eQx1Sy5ysnSsOGEFI2E9ImDFWAvGDmKeojBWN2vfY0z9dt7kzkpSNFBg2FaRRFdiJhJV42/dnfaNcgI732cEI/vrBF0i/VgUEBsHAEyYeB7EqPFI6jL5hGAkDH5feFt67ae3zf5rqYRUII63Bh8ohGhPC4vKQx5YSS0zwewN+r88rAUqjMlDMD7ar9z76dJvOuw1Awjq4p6dtFJ5SaabBh6dc5XqYq62cb1729qMhAVx/AZngpl4m4g5dce1DS97oxQXEkJM5kHCiefV/j5xYPUYEyXYRQsWysfu8t6Z7G+dbFN6hgrdsui1ZO1sWGarOE9HedgjzFF/5yx+dTarRImMBCCiGomFPOQuMnvmTO7bgPgTMQ9NuxOlCIbXHhEAZHCFoXAHcjeGOYCia1xw3pqwCfWWXMha0kDDohGODfT7aHMHbmh8CE7weX4jAMBOBLoXFpUKgrGqf49/e0NUGLrmSUVD8hg5mrNW4rOL46z2XzJhaw1iFT2JiIRNrb/z7i504N4gO9gKx3YT9GLpI2CaFm3Ap4FLApYBLAZcCLgWGAQVstGfG0wXBEsFg2RZj8F7eU3/73O9UMVYGbJWfSbUsOO34b17V2I3Oq6AfqDTbl60acBd14HQ1PEEz00KwwH7r7zz2+wJg6yp/89i/UNgFenoIXWylPgLyBvC+igy/OdAtSIfUGLTVsqpLYgiXBPL99ifVSeydrciIiZRfIqePor7pvqyhxTviL2SpiJsOxGj3LY85uxEJU4uIdqkhlGVwALxoweMlhR5TCIY6pRITJCZAbFn0bStIHlCn9JXf+ciSrRwUROOcb0sCjKYjR38JdOzgpwlZOGoK6wDarLvBX00B70Kaknz07w+OKgK5NSnDmriFSej3uJCJBW+9+4GiwbXJhGmjarc5qaj4yw62MM9ey6gY3CuHeXTlZSdNAmXjYBAwUACAliSAANDCPZ5AYemxx5/w+ptvd3b3EtpSNANPLamDC3KD8/aejl/f+cuiYn/QL4BzM+qzz888HkH0Bb1Fjz68iHMes7prtxMuxkBKAIPNjSZD0TRdV0HfWO7lyZZPXnx0cqlUjGDYj58jCwCZgqDSw3xe8CbgEaHNoohybcaCQXTbRO1gTPIGn1zwbG9CUTiPKhi0CWWw5DM9CxIG2bQKD64W5lr443ffGDdmFAynj4SyXuYDzMg8RZfMvbqlbStNDC0Zk2Nh0JSGw4hFul9Z+lJZYanIRB9awjLGjjjm6yu+bOxVQJ85jisHPPVWUGWKNRRNREE7ONnJE+1caXn1H7eN8ZKVstWf1F+vWLH34y8sbzdd4GFDNK73hnFYMmCwSV9rGtMeXYLrMa7EjhhfW4oz1vTT7QUC+/wi7DoJAmBigMX4+MGWgA9PD8Qd98J6CbsNBWVHHn9KXEebb0SzKhKRAmXR4NK4c67KsSbOm7rXvVnJIKJ1AdQdZBWHv76eb+o11RDsSZJPgp4gzrmLhG1SuAmXAi4FXAq4FHAp4FJgGFCAkDABPgirK+sQMAUgJMiEDU3RwlzveOaB35SiuyaI6sGKCssPmHboBV0x4IrwMJC5iqeA6473jJCwDGDY0HWNg5cjcqIsJ+75yWxgogtH/OaJl2MovCCoa7XBBD/gI4dr4C2HlKPjAMWMhLy1afOrS5fcfdfvvn7yrMl7n9QWifeCAx8Qf9mnLcAkr7ODi4QdbC4pCZtsZA5ucrcjYWoXb2tra25uXrhw4fz582tra722OC/F9PdJgTtpDMpaNvaiH9300D9f+XzL1o/Xb+7FIbL2BQYTUlq0NdsM4l/rSMSikd7ujz58/5tnnB4KgAytAD19SQIACcASEAGIhJcAPK+/4adbGlutq82/hIEtGeFgtjxj9AkJJ3iMJ1bdMvNIv8RYgR/seoN+EqR6ydVVOslDhcWHzThy3pVXLVj0LJ1PL3p27jXXTN1vP8Hn9QT9Xn/Q4wsIog98UDMPkwKMeafutf/KFRsNheNGQXqPCZpRno3aUCys4wEOnniSx9o2r/jvwRNHljBWzCDK14igidRF8KjHRC9qRUuiIHmKCsGs2OcN+X0F2Aw2fuLUz75YFY2bg4WRw0A/WuOojw06zNZJJDdbhSrT8HyARLOzveXSy2bCAAqmhJl5QXWEST6/3z975sXLl72/rbUJdsfUREvjpgfu/+N+++yF9PNjsVLBU3bzrb/riUDgMlkHYEiDQgb4tCEEPsMgG1aTtg0rFjx498mHTPIyNnmEr8DL/F6Pz+eRfBK4s/KgebfkBSApjbzokmv/+8G6RBRMTLSwTWSbps5ExlxAAw5d7d5Q/70TTx0hgmMGsk2hwTdnbwoJmzC4sLgSRfIBj6+CtB9u/9Xv2jt7wAYbJfxgPu3QOncgYR3GVGnm+trXn/pVCQyoCCrZUmjiUbPWxqDzJNTu29B+cuw+u0jYJoWbcCngUsClgEsBlwIuBYYBBWymD5T3bCQMYmHk+wARf/TGkhEeVuBlQZ9fZH6Pp5KFxi19b4Mj+Ibp5AcvSpPF7UgPwZJOBnVP5K1iEXJO083bV33j0FrQDQ2N+u0Tr6ENKhSxDpuhRD85XF/xyfLnFi656/Y7L/zGmYdOnVrqBzfHPsZKSsoYqzzum3M1Dk5xHKrR4H2ZZMq2LNmJ2frKfnPl5GIKicGmjQZHy022u89VuwsJkwE1AgLO/3jvXV5UwBQB2TLJlioSM97vJ0iJGRMCAV9ZJQNfRCW1M05d063ZJpd9upyLFNvPt8CLWaWmyRdfctH48XX7Tt977JiR/TYz7UfJGwSQJnglb7C2bpx9rvgcwhfRqA1is/tWRdMDAIm6+rFfX0KNE0ZOvevvz3XH5La2tpdffnnu3LnBYBAxvFNEDGW9/gL7tDsmeQoY8wliwOMt9PqLGPN6/CW33XGXrED8Jc55bzjp8G+Mj5T9PDkToCOtGZoKm02AhDm4X+YJubv1ph9fWVtRUOpD3054Yw+4pwb9AFFCMIyKw6FCwsuFklRx002/bmnbinbBIMG3MJqJhC3kaVEoDbphJuWQKYSaWL7s/SMPO7isKFQc8hX4PQGPgKHOoSlBlBaLjFUWB4v8YoEfROvWIZ12xnkfLV+DztR4OA7LFz0AssHtE9YgPcmTzVd/59iJJWAXgEGYGegeezySJEkeQfJgJCXy3YUzv1gSQ4yV+goKAxWMhQKlU0prDj393CtwFjnJaqetzqbPdyOu8qS66K/3T6kIVjsF6qIkCCgQtmXCYIstMq/kCRSgLXzopFO/V7+lO2nwuCqTJTmpu6eR09qfQqvoOOdtXP78+pnHoKjfBw++p/T7tzwEWgd4ZG9lxhPo+Gpd58qEbUq4CZcCLgVcCrgUcCngUmCYUIDAcIoDTGo8jiE5wYFre8OnR0wbD/gTJANB5h3NWPD+x18Cjy7ICSNXRyF+s1nk7kAfoT0q6Q6bXoZBd3Nt62dLxvjBMw0LjP3tU29bLq3pBjYraeJwTUnec/fvx4+bOKa6qiYkjhCBcw1K6HIG9LvH3vSHpyKg5miGnwE9bNB9jnCjl1v64XYwUuL8cuHevvlZOUVisHX02kp+calYbjHj7kbCSNo//fmevadNHDNm9MSJEydNmlBZWT5l6qSq6koHSKwdP75u/Pi6ugl40pfx48ePrxs7qnyvybUTJo6tHju2cuK0kkkHHHrmJQ3RAfuezUrPjEwH4w1TMhYLX3zJRdOn7z15Yt3kiXX7TJsyvq5mfF1NzeiqyRPrRlaV19WOqqutqcVjzNg6OGvH146dkHbWjTN/Glv34UefkNZ07vFKhy/pDcpobT9fgeqqCkjYaPj3o3eECkUmFcy79b5mUheGRwKOZDL5zjvvXHLJJT6fj0xtTWiHMJ7APGOCx1uI+xfkKjvg9ZUUFFb8+q4/dHSDmQDUI+tyUrcl3pQJn85Hyk47kTAYmeKOmQIavJwntzZt/NlPrxtdXUba0R6R+fogYcaEESPGXXfdLzY3dkMgpQg8x3E1kTTkASNhcMemgdAYIjDHOFf0WPeXn35w3VWXFXpAVxii+hbCLo7IWEhkhT4hQC3Dxczv9c2ff+OadWmYvP8AACAASURBVPUUKr0rHOuNybLBk1oKAGci4UTLjZefddikiqljK2pGV40ZP75m8vTRk6bVTZw0YcK4cRPH1U2oGzu+trZuTM24MXW1oyZUFe9fO3LfCeP32Wu/MeP3nbDPccV1h5w180fYYpumzkTWeWGNSTLCe1of/dPv9psyDizGS0oEoQ8SBu0G6GSgqHjO5dcs+3Q99MjgCQO2OcGPGtIsi0zYHEt0rsa38fjyr+1bhLb0QeYtY55qUPNOgG/rrE3sP9PqQD9I2EkEZ9q+dM9M5KLLntkbt9UuBb5iFHCuRZT+ihHA7a5Lga8IBZwva2SHkpz3GryN8yYe+fiCEyaUBcECTQgWMKmU+adddssj28CDKGjKoXDPlKDgNxOI7jTpwGE1+ujhaiJuRBo5X/v3O2ZVAFPrYYFJiz5sM5GweUPnggVQARWk6Tedb/3ilrnnhNBDEoqCvKxyr5dXU2BMdL+LH9j+7k9eevi1FxY9teCZZ55f+szzLy1evHjJ4oWLFz/77OLFi557Ns/z2cWL+56LFz+76OnHFi7555NvfhpHh0CkdWkimywks93M7hbf0RRNGpl2aJslKM7GCac7iIYSliw9JWMnn0Mkgbd+zVbXjgPI9ImcqtthKW5ajWeUzEJ6O8tZ1B6pjMzB/QoGqko37Mjo4Y0fLROEYlY07qPN4RhartoNcybWrNvw9388fPo3ziwpK0c8jC6kABL7EQ+Cj629ph1wy89/1dDYpupcVgw4VU3RDBVMEEwDeLNO58PUJ42upOGD/GiltJcNBZwK47lu7arHHnvk/AvOnTBh3PjxdYEC/5ixNWd/6zsP/ePpjz7+si/qRgKa9gLOuZRzLtitMm2GVW7IcIITKPOsX/nxgof/csl5Z+09vmZybdX4URUTa0dOnzzux1fNXbLgiZ6Otmi4B5xiW313DiIBYOe8QU/1BjiyMuIgHDYUw9AQhYP2Nulyk1I3fVImeRtD19uIQg0z3hlOJNpnIz9adn+yJjR0c51AhWZ0Xca53N782uIFV18x78QTT5w0aVJtbe2IESOmT59+yazZf7j3z1+sWt/RBXrYBjqjRqcLcGuyebZH0E7Argeaw6BMWOGJTfUfP1FD246siAXqQqMPbei0PDKQR8KcY5N68GyS2nM1t3Z01o4P1tvEvv8uTUD3bRJkJHZpQ9ybuRRwKbBjFOi7Ltn19POTXWb7CZup2H5Rt4RLAZcCQ0QBekHb1rDI3sTQZU8r11b9fN6xdQFWAJIUDwuNYELVqEO/+1EbBeFQ4+iWSgORQxJhpMPB1c41l+QWJHLiRpTrLbzzvVP2LUA2v4AFpi34sAME0rZqd3YBFjFSMc4bn73nh4SEAScI/oIDT3i/B5B8LBYmr1rIncd4dPXsY8YWQlwd8jrT1y+rpVO5Q39RIiUd/J2r2qxwqf2C4d2LhDmOKcr9gZDkwieDoTO/OtFLxoyyp5ap7A71QJHsFQ2Evc6oIfelTkRj09u+ejsz1S63yxJAarkL/FFpRndT77q1Xesa4zFUJej/rUkt5JxHYvLmLa2rVq/b0tS2dVt3V0+UfH2pOk8qalxOJhVZ0WwlZAVM6tE016RF31e8I8eGTxk+yfBaaIIZcglGGQ6Na0klATGH8NANDggcDHLRxpcyYfB2CAnDzCQHWpY5MaiToJsAuh/oXCeNZESJ9mjJGAqQaR+Gftaw79B9bICJ4qy6TNd2tof5hLXM4Q0ARVNn6TMDCTsrsUviTkESn6aBIGHLlzaEdSLKaQkUg6fNCAgjTKAWOyfzmMx7Zd6r8jjNe0DCFBXLOYqgmQ71mOstqOU3PnbPlSGUnDNPJQvsffTZPzItufu4VcvnuTCHoj+PWY4ZlraY25dmT9h3h5/TiJG9/ODm2k+cpUGRqp42DkECbzfRSsBETxV0Uy4FtkcBx3RxJLd31VD87pjxGbpRQ3G3YVBnrnXJZj2tpxqeabtwWsOdNMsYPlu8QBc460qrwv3iUsClwJBSgMR3JK9DJ8w6VyOcd3Jj1cJHbxhZAD5rfYVeJgSYp0YYcfBnTWonoDg1xiMJHlV53ELCxGsNwi4+NklVuCyD764IhBJJ1L/84E216MgU4okEDljwQY8Z4ATYWnsBykyEudybaOD653/80SmFqCQJAFYKnnTt7Q3k+iYaTrmfNnp54osfnDCGxGrg7Ufyi+h5xtKpzBv+kmPijE+G4V58hfued906FKqTWzJLB7vvOO8WJJzBp9Hi7ASQGW9AKJAOY+CdkH46tQZIKkxXZRTbqa+Z90xVRo13fjpb15fsaTnOokOdBpRC99AQcuHDFItBT7p7yZ43rW15fpG1pKzFddiD0HQQEiZUntBAC9h5Aho0wUTmY+R4xaMraZiXKGDMRRAQlup6UpHjcspvWVLBXRVsNOBOR+uxnoxZRPMwxx0yWuioipK6rmuqplua5LBpZgqxManLSjKqJKOaHrOIoOioOEwz2eG83gTDFI8KXe+Z3vVkhLOELZ0Y2JIJk/t78xNjI9GOIdVBV1PvMjqT8RUajI6ubHUK8IHOk6oF3qHHyWSS/Lqruh6X40kloeiJpBFLGjGwMIZTUXGzIxcSJvsWxLlRrjfNPnPqmBAud/7RzL//HX97B9ZbCCRsHjkGJvXUOQtYF/XRjk7tpmT02v5qX+pI2LY4MIcwwpe9teYoNQRJOwKW+ZqjPqY9HnjX1HrnIuEhGIavXJXWPMNXr7kiOdfPXUQQYhbT9sIgoh1u+aWt53m1Bx5cWFMy+Im8rt1FhexVKCPhXNsy0lQyrX0ZJTIGjvgSusDcPtsNe3ppDXa/uBT4ylGgDxJWud7FefPW+n+XOQWigVLmrXv6tbVhZKjjPBHnETmFhMnDlBkxcyeJiE2COMYK8Jy9nLcrrctPmFZZBV5QBcZGsODBCz4Ig4zJ9OuVsU6lvqLwpZPHP77w0IJiBi5dQUVaKrr7n290kJBNlVNIWAvzjk/PO6A8yFhpWTHzBJjHJwkQloScDJuBdzLwbdavogCudNJPQQR7RcbY3qdd2sIB4tvgLGNxtAi4G5Gws0XUTMfbuA+72y8SpgttuZqLhPuQzwLxhITRoBPADsHFWNwci56OdNN4a5bYfwmd0leNG4ouoyTWwrdQn2wBPycGpjRIhs2GpB6gdCYf88nTUj9ImPSN7QlEMmGE2Vx1QFO72cgJZYHBlhJBX0YCfbnZjcSKqJCtgJCxIaehIDoFuzBSUCIZUbSoRRBCwqZLqb5I2ApqBIok5Gc+HQmDgrTztGogJExSd0DCBnixH5BMGLqn2HHd9YTc2caTCWAgNZ6MmMbeBIaJpN2RTlmLylo0aSSSRgIxsKKaux6w35EuD4ZvBCRx+FWutcvbVhw0DjyBw9oWqmG+A5d+lkSPYlb0OTQ5zmBhs42TOdXtsU7XjqYrOFeTqqHpqTNNmp7KJ5flgPVxDLZEeSyJOvq4fWRVZt9rCBI0v1KOw+Apg0VaT6nY45pNmgb2c2XODM45xuCChtlK+U4h/hC02K1yD6eAiRehF7iZp3C1B05d6fsYD3WOhtHtTIKqCeS/MPQ57JBlLLnbJ7smYzhzKKjhAw2KZ7RztP2Lh6iE/VLJSGSsbRBSAPurqrCznArY6CAC7tAZmqqoSYMrkVgn7Zz29oJ7DJIeg56USUbIwxeezHUZlZrsF+gQddWt1qWAS4EUBegRt8Im6QjQ2js3vHrUtPJRiDsBuolB5qu848ElTch3Kcm4ashJDieKbRWDg9Uc2Nxlcj6pGw00BdwNHEme2PTMH+cjhISor0wqZwWHfNyOkmCQchnAhZiYOCNhRLguG13dX/xzEoaA8kFslSLmGbl8XZPKuaKZZnGWxq7Co9vUzsaulsYtmxvXbmxYu7Fh08b1DfUbGhrqN2/e1NjYuCW/oynbsWXLlpamxtb2ba29cnOvasGOnLgIVcSBNUeuN7Hwj/PLMUqNj3kloZxJ1QuXrW/huCRTsBfcY81Ys3fsa/pgUR3peZnf8rmPE/gP9TqfR3ucL7vM7uzO77ZHZHBlbr006dVpfubbupxEQPknoT57UJz7HfmMeL6NyKecs6H5lM8o47w8azq9vLMI9NopDbb2dDKfyoyVhb46K3Kks6N6m+bpO2CO6zLvmfrJ2hRAiEjq3zoFQerLgYJcwb4XriEAxCgHb5310YMtChTZJzhve2/BnSUCC0mgEOQPVfsnn7VZh6h23IjG5O6wYsdHBrqm+aBOJ3Tfb2lIOB4BQJ+IJmPhWAoGEyQmMOyExyDhh0PFWM/doCrEe1IaB6YSct9bDmaOhjuzMW5EYa3V1YSuyrDLoeLGlQI00lTYglL0hKLH8AQ7BIrxbbckngTfe5Y6gYN7tku4CZcCQAEdkZIRiUQczhVoJ273EKgnriZ0iE+Jt4edIPSdkOy7DvXTPljYSAST1HgszI1YPNbxyrtvvP7he/a2bT+XD+FPTp6All/KSS3FxOLqXJb1aAweezmuxSJNDRvffev15xY8tWjBk4sWPPnmG69sWLcWtjCT8UQihhsIZHuixuNRs/2kX63EXn3+uScWPr+xqRMXZpSxg3lP1mV6CLvuVu1S4KtMAXrELQuxBOedeueKbx0/aYQPIoWAh9KiUcxTddq532+O8R6FJ5NxDm97WYYgKnCSod3gImHHRqi2fOmTRYSDSaIqlbHAQR80W0hY7wcJQ/xOXe9e+vef1yESFqUQYyP3PfysbR1RDtepwI6YEADXdlidDENOqqBsqCoQLwb+67qKHdxppgUj0Og6RGV1zrrMhdbcUdA1ZGp3NxJ2tjRXOkcPcjL3ueoZrPw82uN86w3WbQejHhMJZ62K2pz1p8xMokBmLn238J5TA3lPRcJ5jHQGETKvcFLDSmdO3H6hb0bhXEg4lW+B4cyWmM99Rn39fM0yH5xIGO5IkmrqVx5IOMbVdQ/ffBHET5JYQBB8oVFHXXhrI5HQiHAel/Uecxbh1tuOI2Ed1lZdV7Xmxi3r165bv3aN41yHOWmf69atW7Nuw5fr6z/f0FjfrbYicoc2QBNARWjImUe6Qa/CtzV3rP1kS/3q+vr69fUt6+vbNta31G9sqq9vrK9v2NiwYWPDuo0NayBR37RxY5t51rc0NXd2dsUAOVvuyxwIJ2Oaul+/8hQARejI1vUr2rfUt7a2bmhoqm9obKpf11S/bvOm+oZdftQ3NG7c0ra5PbxqU7s1Nvkg4WxvWtPONsJ59N2XFh508H7MV3r4t2alDGisG+zSv86W0uIMfJoj8jrsTqN7Rjna1rj+pSXPfO3Yw4g79XpYCEMThAJicZFfEpjPL1ZVV86ZM+f111+XZdWS7dBCZSoZ9TbX19WMZL6iBxYtBT7P3NAb8pVsl1LVvZlLgT2BAqSaonGVa9080nj7D2dKjFUETFVexooPPPZ7W1AjLK5EY0on50nVAMVLCwmDzGFwkXCcQpVqbe1fvF1bXekRfBT+0yOKzDeCFR76aTvyPeDF1Lbd6iu2UUDpRt921bcOG8VYMQBpP/PVXfvzB4B7At1TDfC0yeRavqqtIdMUODQ85QEeeGnmhwZBmrMcudlhFwlnIVd+WbmJauMK51svv0r/Z0pZeM+JhPtSbNd113nvgd7VeW2udHqdmaUsatgzYycTKcRr1ZyZM2DEm6tF2ZFw5u3sZlj3TacH5MLeI3KhMd69/NLjxhabkcOLWHDcnU+92wLKOehCL9q05OE//PtfS6iKeCy540iYVuEP//t+0O/1SoJ1Mq9kn5QpeSU4JUkSPRJYrfhCrHDEmd/7fpjzsMmqysBHDvVh3sKIfvnewSN9BRKTvH7mK2K+Iskb9Hr8cHq9ktcr+SQ4vX7JUyJJFdZZVlAwqqmpOw47tHgiHh7qVrv176kUMGSebD+wrgwjsVM0OOBiUrHQCYTtwk/JV8ikosNO/14n6uHBDhQ468cYImlUdr5d+3JmmhaLcLn3rYUPnX30PtB8wcNKJhx05pztGACl3WIIvtitttFvhj9GjiEKjMTtP58/ckRRKMgk0Tw9EvNa4wQdQrmNKDGfzycwbyhYdu53Lvr30tcjETKLS6D6ZeTH82ZC94OVV/36PvB3g1qKu1kwPgR0dat0KTD8KYB6cTqE9VHa31/w1xLGCnCpRWdRlX7/1Hc+2hrlPCaDDrXB42Ftm8zBGeyQIWECt+H3n39gtJ9Jko+xoA9UtAUPEwAJF834vDMfJJwEFTqt6ZBasdzslJ8VTHn8heWAoAkJm1uT6HZLNyI9vck42uAN2bDZtjzOO2Ty5ibH6iJhJ5EGlM5BUSeisN96u4CFHlDbh76wDY0stdus5Br6dlh3cN7eysv3r/PaXOn0ujJLWdRwTo6dSecEotaNMhtgAdSB5+dAwihbyNIM60bp9IBcBxJueW9GJSsRGPN4mFDFCia/sbGnA/YUNc6TDZ++d/CEMW+8/DKJYknUYYPhzHr7fE/TjqZf33z9tamTxxcFvH6J+QQW8gsei9MP+SS/T/R5JALJkkdIgWHm91VOeGN1R8S0n8QYU33uN9gZ5AC9d8l9N1WZ4clEJviYCC8mOkRRFCTSXqIsCdWRigVPhSiWVldPbm0D5XLyAUADPtiNdOv7X6GAIfPOTafP2LvQhlgCpCTGaGMInYAKgiCIoujBkyZhMBg0pyMTvZIEAoT8DlEU4RHDT0Gw5nTqWjFYUMIClQeec1UrehwFSQLJSGHBJGkneYsx0W88HkZFFEPRomgMnOQ8oSc73l/65CETyir90BcGeLGCeUcdfsbFKZd8u2cM7SXY8lXAdUMhfWaDg4MC7cXnF4wfW8UY8/nNuJyeAPRAoEDuAqTpCAR9dtrnBRXLwoLSYDB4ycwLl/zz6SXPPz5r1rfMgQ2Nuvhn/wc7kTo46+i7r7B7iOHe1aXAV4sCKudxJbq5+dM3ppZKQYiYRM6lmCc4+S9/fTOJdmEoYOXoJSuMbmY0JxLWQGVaVQ0VjaRk8tZDOSr4kd3OQfQGdTkZVoKN65fNvvh0gbEiH/MJoo95QnCKsDQLpaX7XfBlLzJ12WXCxB5qEH5JbWxYtrCYsQC6gGYs5B112JYEl1EL2rS6NBc/kAmTtRou74b56y6ZCfb6m55wkfAOUz+dkFlxzVcYCVuIKA8qmaTb4YEYkgvzbzeVTG9ErqtTpXKVyC8/CwS1MTAlBk7//O6cdZ7nkUnqMSiPibV9+PS+QXCXxSQvY1VTT5nVSuLi2GbesebAvabfdMMvUpQaYCoNCUfDPbFIL4BwsjbkWsPGVU88+uAF537T5tw9IsiHTTwM0lY8PH7gNT2l377u7lZUjuZaYpe4mUlwvo1HvjjnuMnACkMjRCaiJ0SRUAnxwALyyCDrKSodfettd/73w5Ut7ZF4AiyKwfG1iq6BrFkwQBq6xb8yFKDQc1ovVxOxWOy9Dz++8Wc3j6+pKvSBxFEQBL8/GAiEvF6vjYQlSfL5TFengUBIFE2oJeZx0NyFWRsMer1e+ipJFrDD735/kHnLp5z94/UY/QLdQmng+UoHR3zWjCahJsWyo7QCkhYe0YzuN95eMnlieZHESiVWVQQ2GEzwMm8Fk6qPPu2i7Gpzu27AaZ0lSyHw7RnubkVBbVxPdnG1e/5184guPp8H2h0MSf4AyX8lf2BU7diZs2fPvmz2zDkzv/u9744cVWUjYVhOJZ/YZ0sCnLhKQeapPvg7V0UxkB1F6nPVo3fdmLt3cilgUkCGMBytnx8yoayA/CSLzCPAShss3/ec82+4+vLrfnTlVVdc/cPLf3DN5dfMvfyaufOuvuKKq66Ye9W8uVfNu+IqSF9x1RXzvj9v3vfnzb1yLp30lT6vzOO4/PLLr/n+Vad+/aTigISMDqy8ggCK0UHcVqegHowVl+53wdooLrrbQcI9PL5y8b0/LDL37hhjRUec9YMmxYGEYU+T1m8XCVuU2JHnYqCc+o7cYyDX5NEeFwmnBny75BoI7Ye+7Habm1EgvUUZP9pfU6XsrB1K7IlIWNVAy5FzHnv38V9WEsoL+FlxzUU33dXFeUJL9jZ8+s2j9j9yxombtqQ8SKdIll8qDQlriikBMkC7khxChxPRVq731q//bFQVOaqA1wBoHkoMmXkPsfSoIhpgtUe81aiCqa0WAxVNEE7n15Dtl6KhzyjXy/UV/15wY1CCF6REcfmAlzURMVgUmYI0f2Hh6Nmzf9jZA67fk7ohG4B+QRpsv3Es3JBxD/erSwGLAjrXwSpBUzWaPIA5Y1t/f+v/KykIAHskeQqKSuiJ8Iim+AL2aHy+UCj0re+cd8EFF82eOfPyS2dfdtlll27vuPLKKy+99NJTTjllypQpFVUjSPBAeDj9s2jkGdd+DKE2cUrjTpblRoWeGtMDq8HjitaTkDsVrcfg4XeXLT3qa6ALHSpm3hATEVOW+kFM4ZMKmFB9zKkXDxskTCEuElzrBYkKT8qxllnfOwPWImgtY6JfDIxkrIQJZeMmH/rgQ8+F4zxJLi5h8AxZIUVvfeGiZw459ADJizLkNDpamiMF5cw7du9TZneY/vhcmbA1/d2/LgV2JQX0GG9ZfsusU83H1AN7XKihJkAEX89ISSyksLqgZSMWwClJ8PanAEK0G048gFOfxhleKG0FyPnF3n/0I/T1eoLEkvkFkFEUUuWCAwmDshqpUtMnBei1ZcJdPLzsipNrweaN7ikU3/GPt9tJxGEDIQDDls4yOStFDwlDPwJONjt1N0euKxNOkWWAKQcVc+E9ewIMGvM8wDbuzuJ50CeNbruzrX3vvXOtz3V16j65SuSXv+chYQ6ODwkJJ176009gP5QxMeRlAe8VN9/EOV+/dt1hBx1ZPWL8qrVbozuhwZiGhB2oVUfNSYrFF0eb5NhbLz4xIgQiI2IY8VNERhRyygqLmbeM+SbO++UjCMzR2yr4sxmUp9kO7AnvF/CsCB78dW708tiHF54x1hJZ46vPev+BxhKl4WUTPPSwU1vbAcYkNPiMKWn7B86ZlJp3bsqlQBYKwA4KxhciSWGSR1q/e8bxJBnw+AKgHi0IkgDBHkFLweNjYujp518xa8IJnOm0JNt3551pur759lvf+dY5XlSoo7mOIuZQ1ck/XqbxDg4iBQgZggLhdJkwgeFkuHsL5/zNt1+cceR+0DZT0myq6BUXsBIvCzHmF3yMVR11yiXDBgmjRyuD3HSHOe++4drLbL7V7/cyVsb8Y5k44oab7+mN8ajM46jYbNNQB7fyCYynAivSRx+/N+Pw/b2mIgkTRVR1RFkPC5UxVlw05YR2zre6qtE2Bd2ES4EhowC9fx3V68D/qN0L77y+Gp5KfNYJCeMuN0bt8ZPFWBD8TYFYlYkloDg3BEjY5/P5/X7TOxatO74CQQDDsRCekCeVFB5w4co42FMAr56GhJOoX0LBLnVu9PCeDw8pQ00/uNLHxPLXV4e7qJANhBxImHS4nbFSHbQa9GR2bsiR6yLhHaa5g4ppiM6Rb0+AQeGdd7ilu+dCBx1y0Sctf/e0Mtddd671ua5O3S1Xifzy92gkrL799O/LSIkGFBdh1ayrrQ55/eMnHrRidZvBeRw1fFPEGkgqHQmnrgSVYQTAEFIbxKac820rTtm3klxW0LuA5DGokcw8TPDDwl5TOvao1R0GBetEa+dUpTuWwjHWIXgdnPiOkcNcRn/Zurx+2csjQ/CiDPol9FsRYGKACSDWAnhsImGJSaX/eOpV0LUyuGyAJaUGSBpfWPDG+iouOTs2HO5V6RTALX89/NGz949EOOn1B5koOZEw8xaW73/SygioI8MBOsugutyPgRoJAhw3MjBgLlnFy40rXz12/1E4u0UQ5Xqryw+Z+Umcd6Oagw5R3dIArCUOhXu/8+9Fp37tYDRrg4eYtpC8aBoMMmcRXOJ5GAMtbnHEoafudjthogE+qbCtlpB7N3J1y39efSoo4u4gdMLDWDErmMLEsQ88/iJ42TE4GgoaMjzmsPoj0W3+AjQXcSCUfz78f2Nx9fCSzjlplYh+Jo5inrqPO5St5v1ptXAMiJt0KeBSYJAoQBEj06zx1QSPNb2/4IFSBhrIEK1XlJgIO4weeM1LuIXlQ6thMDjBAl6AwaIEG5C4lJlvf+IBnHJgZ9piZbL+pT1N+1NigoSsGGOiIHlokfcIJKZmzFc+/uybVqEiCW6R0vY9yYRJnABR53VwR88//tcfR6J4GVF8Se2hJ4Pna/J6SO+IPiyJZnDyXJ3UDTw1RdM0dRBOCJ6kqaqcQIYo7cY5OGwXCe/w1M9BUSe6s99UaUOxw3fcsy7Mgz5OWg2zzu1c63NdnepkrhL55e+RSBj3RMEpVsMX71WScSGxrQKTPMI+e0/77IsviUBJmURTKXLln+ofCSM/TTTmOo+tvm2WKfiC1wZs04JmIr1CBDAuDPpZJWPVt//tpUbOu4hvz78tOUpanYN2mGk1Bm8NFZy+3nHDD3zI0KN9owcxcCG+YgqheaQp7StgrOSdTxp6VND5lg3QH3WRcA56u9kDogC+rPSwvurFAwsoHoa5AZNSR/CVhg486+MYtwLXmkEyBoSEdVD1JeWPJOctya2fHjZtnA8koSwkVlcfdPHnUd6NQR77IuFkEpQgVq1adcpJJwYYKwlBMKFzz7/grrvumnvxRaNRngKPMfCawErCkw3pqgNOvWQ3R1Gyh4JMno0Y583R1g+PO3xiALYFRfQ4E2KsnLEx5175ayquGwndiKk8KXPF8gPhZC4c2wS9LWvfe3F0EFaxQvQIDn62BImJFcxT88qGDhMJa0ncN7Nb4yZcCrgUGDQKULSkhOn9nnMNTCFal710ULUXYTBoPAPAFcyEB9zDgz40nRIY7dro15mwtsJtBWknBqZ0VgRsZdoY2E5Yv6T+0k/wPVBZ962ffoFWKhbnpOKmm4a2Zrh6q1xTdEyojgAAIABJREFUktzo/cOtl1bi4gVIXxh91lW3dSMMtbgdc7POIrHKjdi2lg1bmjZtbmla39i4vrFxw+aG+s2DE7pv86aGxo0b6uvrG9q6Cbhvj7FOQ8LP/HF+OcrofcwrCeVMql64bH0L56qB/8k9hdW77dWchnL6FrYIkv/fvnX0n5N/zTtWsv+746/Ol9WO3WQPvioP+qTNkeHV1Z1rfa6rU33MVSK//D0OCYP4U8dQKMT7/uP+uwuCfpA2MVZWVjZ//vympiabOhBi3RFh087PJ9EPEiZns/Yuocr55uf/8GOwhiQYbCJh0wMQqoL6YYNTKh857eTVnSABo+HJpx3bKWO5/knz46q0dK56fXptia08hQ3z4iZrCN8y6DpLYmKw3F86fksn70hkRcLbubn7s0uB7VBAD/PWD44BNT6ajICC+yLhMD4R8IFWA5YFWBbhcF+ZMHKJtEVsKHIXN7b964k7RyAETCFhnXS2M2XCnPPm5uaxY8eWlpZffc2Pl3++hjb/41Fo0cqlj07wg+AFEaADCUtV+582bJCwKUuPcN7w0G/mlhQwX0GQCSHGCiWEsaHRh6/rQr0RoG7C0GMaT6qIhK0lLI2/oDjBmmzIvS1bV7903D5BH0Zs95FOu1jEvKMee3cthESBhRBdHmxnErg/uxRwKbAjFEhHwioYPXXWf+vYAwpxo51gcManDYNFkhTbUNWRGHSZcDbv/YjCyat/qJI8F5L3aFznnUgYtiNBGSXWGWtbc9IhdRXIniDLNPFvL67qNnXeHAS0Vyw9ltz4zkXHTyiAvUqBBQtZsNDnF/0+5vN5fDt9+K1wc0eee2UjulIkBbzc7LWFhHXOtcQzf5xfhvyWx0XCjtHLkcxNVBvg2ePuyoRtmuRM5CDzbsrOY3TTepLezFxXp0rlKpFf/p6FhGkRAoc8Cm2SQkyARGvzltVrvty8pbk3YomWLPIA2zzYSNihoEOqhTzJ5XWvPXijKQImnh9eNeA3C51UkPUiWOQyVvmPxcvbErxXB6H2IBykZ6SB65oY2dJoGudbnrnnR1ZDvMDJo5AapdWgNwU/ERzxVoybdnx7gueQCQ9CA90qvtIU0MM8tmrO0TXW0zHoSNjxPIK/N43zKJdXn3PEyErgpCpNmXAOJGwYhqZqS5cu7eyO9MQALoK5rca5muDJFq6um/X1SQXDHwmDZUSEd350xgHFaDYIJg9MKPaC/77A2bNu7ua8V+GxhIqO6wEJa7CE0jvCyVyAWYRq8KRiK5ls3vrl86MLYMkoNAU1PiZV/uGFTzo41MI1qucrPcfdzrsUGCIK2NrRwC0YMo9t+eXV30WdLrC6ysDA9HVYIWHTl2Gwcr8z5zUjiwKEomh24MDfQJ5FgaWIc97b2LH23WLGik1nWQHm2fvzdmCWVFivHOjHXrS0Xt714eXHVJWQWqAUYJKP2S5CTR5ox/+IjJUGBCb69zr9ig0Yg4A8xORmr10kvMOPQm6i2hDJHnfHXNjh++1pF+ZBH5tQ5hM1jLq4c63PdXWqg7lK5Je/ZyFhu0/meBMVZDzA31MiIcugsyjLMkiD6RhaJAxgOMFjq5f++XqL10deFNdeCQ4fms/QOu1jUs2Bh5/fKgMMJqF2aiR3LJWOhOPwvjB4+8fH7uUPUUN8ZRU1E04+dkaB+TqAqDbQGtCAEpl3VN30E9qSvEsHO2FVtx0yAvF2rEXuVS4FUhTQw7xr+awjR4HGBISy9aIiH8YaZoz5ywoPOnt5nGPwS8S06FPdfHRz/0nVD0uBpaNBWmdGkutr/vngtaME5mFlxdPO/SLGe8y5TXGSHFdbSeA4dZ7QeRys5TkHK4NtvPfDO648qRBtgwXRJ0gQSAhil3uqDzh95nDRjqYu6JGmZU9PLcRdLsnPPGVMLPExr4cF73zw5W1oaCcrBvj3NuLpSJhWVZvFQOkx7AWQIEaOda9ftvTBsX6QL1taLsW3Pf52G+4aoJ8ti4juX5cCLgUGlQJpq5ueePTum0d5WVBgQZ/fK/qynh7BY58UcT2rwHbH0eGAr5SYr3TmdbfHTB1vFTx+aTJ5DEX2jGsQr0Lj21a/8/idBaYjE8aEwskzZjag+05Nj+iqHI/GlIQlQKYVS+vlsRVzjqq0wjV5TWvlATcy5wXARkmBqafN2zZwJPz0vfNLXJlwvk+Ezd/nTtivqa8ic5qbLGkA2C6WL913TTm7WXkm0luV66JUqVwl8svf45AwddzsHP2hLCfXbJsZQubQImGw9IgTEra0o/1gF2wdgiBZkTlFcBxRNIaximffam5KghR3EI507Wio0NBee+a+MX5W5GVSsJixkt8/+NjpxxxSajUphYSZj3lq66af2CLzLgQA4B/CcQxC89wqvuIU2BVIOJnyqAKco8KNDfUfPTXCkgmvUXgYjCp0hG0pc7O0kTG4pgEzlgCfUpwbSc7befzje3767ZAAXrKGPRIOP/H7H44gR9fgF6eACUUe5hdY0T/f2hjmPEy+CRAJG2kyYVpFHSyGoYJEGJEwbPTpMc4bH//dVSMYs3bT/FffvbgZLfzSaOh+2RkKfBV5u52h11fnWnxC4alM3Hj1rOsu+94P5s659JKZc2ZeOvfSuX3Py+dcbp+X4eGMSTcHj1mXzrLPmXNmZj1n93tQPfRJ9dO97Bh45o3mXDbnimvOm3XF40uW4tY/mrZpMVROATk3bbgpcYMnElzd9JvLvg6YFtCnj7HK86++pxn5JEOPgfYOBrPUVc3QMAglMD8Rvm3Fk3dec8MPgCLfvWzedy+bN3v2xXNmXzR7NkRM38ljzqyZ111z5Zy5V97whydJpk0NNjlQS6/G8TVNJvzkffOL4TXETO1oMd1OGD2yOK7NDmjyLDDw5yHPiu1iA7/DwK6wb5Q74XhNDazu/4XSucmSfeIMrz7vXOtzXZ3qY64S+eXvSUjY7hGamcjcUEmwiqwtudox5cA2EgZkN+hI2PQ0Q7u1oCsT49FVL//p/xUT/vUUjh478ZAZB4AvRdKrNP1PBBgLFYoBxqQjz/xpGzHcqXHckZTZc2DdzRiuIPNJ9px6+P7FJu4VhFH7rG2Pn3nsgdX0fgHfjqbKErhy8NTVTTsZkbAuG9xFwgMcBntOUmKAV38Viuth3rN89lEjwZDf43HIhFFjoa9MGEKBpeji2JZJS1IJJLqq87AOUFcBbQhw657kRgs3Wg6fWBdgFaHJZ6/T4GcN6s09TJYaoqGiOxOQM2/l2qp/3HM1PLHggRXshE2ZsHdYyYQhtCbXo7O/PQMcxQsAf50esxa/vSmFhGGtSGD8I4zllnqDWiyGoXE5AacO/hiQzirnUb1r9RkH1lULDBc54Ywf3NfMTUl+jq2F1CC6qewUwHluzkiamzhBsxd2c7+iFCDomAA5qkGgkKvhLp7ItAQbzuSJqBBkA5wvGDJgYC0CXBO68VdNky7OY728+5PTx0PMdgZOr8HV38Mvre6gh0JPcp2/+PyLiSjKhKG39OioXI0YiS7qPq1ig04KwzDi+G5xvJrSbkJNwc+cSNjLylkfJKzhIuu4PLUiDzQzrUF5fRn6O+TVDLtQHu2xXlNOJsG+/n89kQd90qbP8KLHzrU+19WpPuYqkV/+noSEyUGyhXtj3JAV5OqsHEsdOo1lHgKZcA4k/JNCEwmX773PIQ//9e6QqUqIzh+Agw4wFihgLOQpYqEDX1wWIbPe5E5EFcYx1nXQ61SRz1c5j3/65pKqYDAAQqFSJlaef83tcc7POfrAkaQUnYGEpfEuEk49TANL0SNGe8T2TjFWMUQv5IE1b3iU7ouEmVdikge3/bNoRw8QCWtcVSFYWhTEwobGFY0rMtfD3Ahfdv65IguxgmmbFB4znEg4G3YzEB7CnhLXzN+3cb7m2YduzIqE9ztjZmJ4EBgxfpJrvd8+8QARXRH4TTM5dF8vlP7yb6+2kdAcnNSoYM2Be2cWK0nT2GIxCAkrMTSXtn3cQ6STfz1w28SA6WPg0AtuaeKwv0AM6XChxJ7VDmSrTc1+GBegdrapuWf1ym3toFIAA6RxPYZIGGZIOKFwOcbliJqIKwml75mMy6kzmZRlWVEUCilEO904zVKcnwbaMFnODD4q42uenaQNxp64qqsE42VUjc6UCUNtSlL+ZNE0Lzi/Zp4AY2MYm7ByqwGGM2qEa7ENX64dVzvh5Rde0RUCpNAFzV6tsUFg14KRiSkERsZnRhfy+Zqtm9ZS6fjNwW3b6jQq13ufvO8noOwNe6k+PytnnppFH2wyfUfr4M1VM9Kwi6OeAec7mpNncqB3y7PaHS6WR3ts2ufak9jhm+8ZF+ZBotTEGXZd2onW57o01cdcJfLLT62HhO/6fqbzS/nVmhqLQS8PVcOBO6TkLxm+pnucTVvido1M+KU/zQ+AX6oixqr3nX54vGHZkVMCXsLGFKgAXVShIIz5C8ZedMXN8FbjvAcc6lu9or4N8JNoDEhYS3Jt8/e/fYQfgqKWs+B+/hFHfbo+rCv8vCMOrrGQsKUljR68xHFj9z6pOck7dS2pm656XQvhHCPgXIYpTVoJiun2AwSJ7pFOAa2X93yckgkDkvKIYDGMhxVFyfQdbSpwpNeQ+xupZZBwE23QkuiTBd0/63zV8uVPP/HMsy++3torW2XgWem7xmGOuWhxZE3gnnon5ysX3vf9AtBtM2XC0Gyvj/lGHvCNWbsNCdvT0KRMkvMurrcdOXFUOWNBjPQkEYXBUZ7vxHm3beE8Sm4JAHHJeJLvVrsuZ4LQmRVbGGkBoUDDX5w0Dff3pKLRx81Zq4G3+WhcybXG5x637L/kqidXfvZa9qBcJDn6flQ57wYRO0r3v5LijizDlmvcc+VnqWKQs3LdOVf+QG+fox7aJYGNEmufxMxBH1LOB7f/9ECbM9DyWe9uVwK/WrwYxHXXLTthwK6gHa3pq156uJJEBqBAtNcp357fA12UOW+JNiw/dJ99b/zpbTKZCQOLp5luEG3vftl4Vfv+uzJhQPzjMOdbFz90UwHq6UjMV+QdJXrGL/pvexP5s9Bh15b2XXMMvPlKyvPXXdnBIblX1vnjZv7PUGBIJs0wrnRIB87uNy2q9BXuaJmN0N1BW9A6d7Q9uaIoceKkYZUiHITa0S/96UY/YN0Kxmr2mX600bnmpUdv8zLYCYRsK0BfkY/8agVKRk59d2PXVs6jiJ5osbN7N6CEtTdgcC0c+eLlaaXM6/EzNpKxfU4+7xe9Mo9u7frujANrLO1oFwkPiLyOwn2nkouEHeTJmiQkfHSVqR2dQsKgJsF2DgmnbyFp4Dqax3TgnICBAj/IXFcNVdaSqk4Bt+E5y4qENa7TfgZcZMBWPdd7uP75kj9egZ5aHUjY52P+4YGEzX3xOCpyNx8ysqSEsZA/4A0Wi8wnofYH84jSlOM+7QHn8HBA7wkM94OEnfPcHtQI583/+O088oZQcej5KxPAwtIQZOXV7CvzTGStpJ/MPKsdpsUsGltIuBP2GsA+xYY7w7Thu6xZ/Qx91p+GvmFZb9tP5kBb1E9V2X6yphBsneRzDrQ5Ay2fTxvsMmbcDYqbx+MRlSfkB2/7cbnAvOCi0M/YxN/+5RVkjmS9s/6bRx58xtdP3doWhmdE4yjrAM1iS7k4ra1OYqX9sEu+xJK4R6o0c63h1h+dFUSZsIcFBFbo9U5Y8n73VtTJ4aDKFMdodgNDvM7eOdO7pHNDeRN7briJ/0kKDOXcGY51D/UgUp/T7kJbjY7XgQ2DdbQ6TCvsKNZvfk4kTKpEwCyruKQpUR5Z+dJ91wdB4FvEWN3kfU9u7djU0/XZ9LFBP7kyRFmwxASfCUcBM5937X3bHANovRocWQNIGtzo4Vrb/dfPqTCRbhErOvDVj7thCzUZv/DIA8a4SHgA9MxatO98IekZYaiUB+OsF39FM7VeHv5kzjHV4MncS+6HPSJE9xkcJGwJ0CgaRxKV40zGwiK4oRpgRY8bRvB3oEi4CB5h0NhD59eMIRLe75u7TyZMHaPJCGnZ4B1cbz56/IhyxnweiYFnbq9I+ueiwEITbnnghRYVVMSBCpoBauSwjdd3PvfNsajIE9xoaPry5XHggUBgVce822h7/LbLpLF0qdz8Uk6uLp90frUOy1I2mXEYNCBbBNQZUE5mzdVh2fJd2Kh85oCzzNA3zXm3fNIDbVE+dTrK2LMoz8RAmzPQ8nk2g4qZ7A5oskG4i6TB48nbrryg3ORefMw76VcPPqdzo62zYb+9jxw36pDmLb3UIjlp+vUcaAN3QXmD5NRqgvMOzpuP3Xc0hbzyiiACkTwjb7n3FfS6bxiwOZl0ZcKpQRnQ/HEL73EUSI30VyO1xw1QjgYPGAkXwiIeZKxuwr7f6AKes/vumy8fRT5mRIgyj7w/rfRexirZ6OM/3wq+aehA5acdRlMK11u41nDo6MISxrxFlaxi0kGnXt5BVSu9Fx29by1hcvNNQ3+CbOe0o3dej3rna9iFT1XWmUJD5oQWebXIlOTnVXZPLjSkSNgcEA28sICurkL6pRRwHIIh4RGX48Q/kvbcDiBh7/BCwjh3qO/QwaTBu7jWes5he9Wklhi/SM6tmMRYKSud+s7arTEUd6O8Gy5z0iE3KiYSAt7mfJuWXH/+sftWiAFWdMh7m7MYAzj4dPvCLAlTQp/+i/PafNLpV++p36yZqRq6qcswHHoyHJblfOaAs8zQ0815t3zSA21RPnU6ymR9HfWTOdDmDLR8P7fu+1M6EoZb6fwvt/4wBKsVC7BixkqLa/c/9bzvFFaPrB1/3LKPemALE2XEmmqAWHi4vkFRn0jjvOPh+24tRX/7AmxQoqWcWDzj9LmbYjwOXgGiiIRT1iVEb8cAp20sbjd/oMM17Mr3nSRuzv8SBYbdhBviBv2vjN3AkPDS+65Dd81+xmrH73dmK+fb9I5w/Tv7lbMSPwhpCAmbXrTAh0I580+//YFXIR4wCgVQ1ZCkiwMcIYPzRA/nTa8+dGsVYwVSAAC5v/pvL3/YhsFQuNF14VHTxqBUKw0IsxBL95hFBtZ53j4cDisY1cAuT0Gc6dPOpAS5886aOThcl8GTyVSEZllWNTTABuY/R2DkcJg8/mQ0qp+v9tTO8hImUljhVdVYrCdXRRo3VNLVzVUiPT8WAwhDBzk+0cBretphd1YGk1geSwABFHDJkXqbpl2wy75kQcK+QZUJ63LPVp7o5DypxHtlJEySgwctFUSfpicYJ+rLmu5HO7oIQhOZMmFAmrtZJkwBxEkxmUZRM3iYGx3XfPv4IGMFEPRJYswvQvAOCU8vk8rG7H9GfTdXgE3jPd3RpEwTI4uEPOvUMMAzA7g1ePR34EKMeacufKvJ2YisV+XKVBQlEjFjSMv/n73vgI+juP6fLXd7VXdqVrctWwbbGFMdCMUYQi8BQiC00AKY0BJaKKEEQgKBOPAnEH4JHWKIwVTTAqGDKTbuvapbXbrT1W3z/7w3u3t70km6cwMZ7ed8nl3t7c68mZ1933nvfV8ymUhYfLDwC0WjbPQqGlQ4ISsJWUnKelLWVUWXZRjY2A7mJAmT1kA3GuS4pmnW9KVpGstBb0F0xi1k7bLr9Nkd5OKZ/sSeRb03EgqHwxDmaD3Fmi6rFvs/NKqnx5g9VEVNJBLWmblWgJ2v6zqTNhNUf3EBNW4sxu5iFdi7oP/JmZq2Q47Z565syjukEmkXzaYW9nPSfpzFjv23WZSt11GWhSxqsE2nZFkNdlo6ElbjGk0oX7/2VCkh+UbaIb/griKkyFE17e1PNyQwxZ2RFERFB2kz6Nh0C9qmum/jj9lMEg6HW7a0ffX1t//vbw8ccfCeXkIChPgI78Qp2SEIwATmyiPu4A133fPWhx9sblq3btOKcKQ3EpPN5B+p93UWIyDt5G1swnf/85zGz8jJw04C3/0I27k1GHYdNECFc0XCN5hIuKx66vEb46pCNdqx5Jbzj/aa8Je5LaKjtBNmSEdZ6aTD6yI0igYXRFeMgji3DpMjESp30fia0/cvEgkJ5BUSsaBk/6PX9NIucLnrpbTjjAPGl6aDYNzbeiSs67qSSKqyzFSlUCjEtK5YFJTaWDRRX1+/fPny2bNnz5s376uvvlq1alV7e8oZPJlEA8jAMDU3EWAXJhKyBYaTSaW9vXP16rVfffXN3LlzP/jggxUrVjQ1NamKatcmc9Tz7CPFqGAiIcdjyU8//VxTaSwWw6NJ9PgyTojFIhs3blywYMGnn366atUqqICtbfay7XDfIlNS2Xc8Hm9vb6+vr3/99XkfffTJ8mUr6+sao5F4LJpgeDjUG4nGk2AXVcEt2Hqh9r3oztlXu2l4wUWHFtu8o21I2BH07HPywihgLBBFjoxZ0AJIs9lLk63vzH702pkX//H+J7op7aG0PR7fdZEwI3+2xo6mA/Dv/OcfLishBFj6wPvEzUP2XzcmGOZd/hJCSmaccll7TInJSiyhh3riSKyaYaUg47jQgX06odJw69pvbpg584Ir7nnn6yY8c2tQaMZbyKre0taxdPnKTz77/KWXX/nks0/rGpoSssKGMUW3bmYUsn6+LWhNlmU7Ao/H4xs2bFi+fPnLuC1fvnz9+vXmE03ZahSj4bXunmMB0lmbK2XwU4hgl+NUVVcuWdHZDdAeE7WkFrkYQO3o6Fi4cOELz7+waNEi+6JYNndniJrB4Gg0qipqQ0PD2rVr38Bt7dq1LS0t9vnQfk0jGBNWGdgUYv/jzihbE1eWhR1fpywrYp2Wa42sH2ZXsL+OsinnWp1cz8+mDvZzjOvDITmi0FiCtqy+4tSDGBJ2EpfLN+bos3/XpELaNyCFBhyc1DT43+LegrVeds1ca7sDzr/99tsn7j65ZsKk3Wt2nzxu7NSxoycUlY8rrhxdVlFRUVZZVVo2urxoXFVg3NiC6gllNRMqx5eUVgZfn/cGq4uMFpHs+j71WrfO3wEN2rmXtI+NkfKuJ4GdO5q++7vtKj2YGxJ+++Eb0TvaS0jhuL2OaAWDpEZpW/23b5YVQq54sKMImI8Ui6ihEjF/1B3//nA9Mt6m5rbc+hAjbbTOZXP+VsKsvkBhXXDToy9hqIoOlhy54ecHVm9fJGzUUdFUWZbjgH7r6ureefu/Z591ntef7/YGnA4XBzXB1joEl9uZ5/PyhMyYMWPOnDlJRc7CLmoNpUElYjurtzc6f/5XZ555jiS5kTYs5S3KakIIOeKII5577rktW7ZYlpZBr27/IwuwZPfTEonE2vUbf3fTrcVFFR53fl0tAwYsP2IsnuxsbWu48qrL8gIep1N0Oo3lkPyCwFPPPhOVQWJZSADuripqT0/PggULLrroouKSUq8/YOWDZo3iOM7lgjE2cffJ1157/dr1GymlvVFwu7dekxZssrdnZ5QzI2GnwNLxmEjYSI65FUgYvHZjNNF00U/2LyDk4OMvaKa0C3ImbTebcF4mm/CUn35XccIKzhZRtnSAPcgylXSv+/iFySZJACJhP7rmuXnidEl5kDycBCcefEp9CBAthVC7VOqUdDs5DJs+G6YnYT4XMHR7ddrB1rKYcbbP2dntxuPxZDIZi8X+85//HHzwwQ7JLTpdotMjOl0Ol8Q7jNmjoGjUGWee9ea77ydUGjP9Pjq7QxYMzryexR7TQWvS0tIyd+7co48+urq6mj1KgmCQmluFysrKww8/fO7cuR0dEGtimbIHvXD/PzJLPls10DRNkbVkJNn73L+fmjRuHCHk2Tnvg2dPIoQs9CB+TVFfmD17/PjxLhcjKYMKSm7X2Wef2RsxAibN57v/7YwjzO7d3d390MP/2H3SHqyNYJvCTRAEdvHx48ffeeedtbWbwNU+kQCnFfxACm7wQ2UZnoxBYZ9S+o+TAasy1B/YZe1n9bnRkLv23257mT0R7Dpm3YasQp8Tcq1Fn58PtWu9+LIs5FqdXM/PshrWacb1cZ+NJKWNyg0fv/7EM48/NHfeW2u3hFUMoI8oSkyOaHocPnYkzHyk2QVzre12Ol/XIeOGqoAexBzQYEa0shogIR48RnAopgHchyx2vZSGtGRcD2t6NKnISXQPUbYtqdJ2atB3dxlrYIwUdkkJfHcj67u5867Sibki4RvMOOH8cXvP6EjqSKjVSaP1Z510GNM8xBQuM3QRQsio6ResQWdOWNoEDXXQXstggFFpdAvtWnfF8dMKCXE7COH54J6HLo/SNko79YhOw71dK844eNy2ImGrX1MVlKkeV+KxSCh87tm/JETkwCET2Hr6hySDdynHOHxAGJxI7ntgVjRBWSKslFtz6uJMHFb8beoPtpIORLjso2s3Xn9TIJDvcnmcDhf75nmR53newZlJe8DJVTBrcfTRRy9eukxFYJEudUNhNW/EWo400TpYSijVn3risUmTdg8U5CEzuMcXKGvvjctUj8baKNUTydYvvng3v9Dtchu8RWwA8AJxOKDrH392DgtshVuk0vbgDjsEaRggiSul6q233BQMBiVJ4gRRkDyc5OMkH7JPwVXZNVOLDqIkOKSJk6f858VXFExAEUflOmVpMluFa+mDizd16taXBkbCAHQcQfc+py2IgE0TRj4s8CMLfPb3A71iCw2tOqjU4ydk72MurkebcEJLwTzMOZnBDdgO/wbxjs4D8k+eR8Ysyzv6u0PCGiZRx0xRKSkpVA/RzqUHjGXU9ACaGAw2PA3xmSOOIHFUFdYctGDxCmZitXJvMlGg1yE47qc/DsZtEnqSUi0pRxMyOKHHdJpMZDzRqhZjVrdnY2IjHB7tno6WZYsWHPOTw/IDfsic4mNcOcDRzQmmFw0+Ng4HJzgFGPCi6667/9zRBUENUQ0sRcCtMNCcifoo1A8KGtWSoILCBzLRL1644Gc/P4MIAvi648bzPASw8MYc7RA4r4P4JR4YyHDzeDzHHnvs4qXLUhzj0NCBpGUJgRXYagWmq6YfL4mNAAAgAElEQVT0408+PPW001we4FP3EKfbEXjo6XcSlMp6L1VaKFXrV6844pCDIB8Wz4lOFxMILhMAKj7+pBNZHWAJAHwisI14H+wPHCHwXOjffvvtoYce6pDcTng4nERwYaZW1iCYDC1UzA4dfsRhK1et6w7LSXRTAiSsy1RLmGm3YGxYvAhZtryPIAbaZVezX9N+JJvyQFfeiuN4O/uMwZ6IbGphPyfXO9t/m0XZeiNnWci1Ormen2U1rNOM67N9SuNRGm6k8UaqdlCkT+lI6l3A8ABzVULu+X4iYbuQ0uKtWGrj1ASpazSq0miSqjFKu1TZDC0Bq3YsAWEr25hb2F6TYVm2BsZIYZeUwLAclNtQ6V2lE3NEwo9c5wPVXiCksHrvIxt743HQGShVezd/+kqNh+RLBDN9osrB2UChc59/zKs1abP0RBzW+63XYFo/sLVGBtOYlBkTbHRz06J5FX5U4nhCPJ6rHn68FZJy0LgWQVzQ/IvpNRnyCXNeIo4bu8cxWxTaDSrVoBtTa5nxDub3ONUa1PjGG6+53ONwIz2PiHYnx4QJ+9x3/8MbN9TVbq6vr69taWl8/MmHp+45zlBvBWYOBzkcedwlaxrAqNfN7px6bTAZMG7utKNwokaTEayHLtNEJ+1pfuJv95QFvBDC6fAIgouRlE3cfcrZZ513+eWXX/WbS44/4Uc+FylF/3UXOI+aGW1FzwmX3LIyBA7q4YQRvKsC/pQhfFtNykpClWNUBzYmPdm94LMPLjv/Ar/TtNIYmaJdxFuxKhpqpeGY1khp/bNP/QVIw20bj4gZtWyeE4r9ZfusCsFSBfQ0UqXpsCTAYqGQBjm+mUZWP/PgbRUBUMR5iE518qKTcM7CqglnX3bdLff89S+z/nTppWfuNXWCW4JzfG6fg3dJTi8vCvARfEQMPj7vQxYMHUbPq/QOVqkWw/dyJpicfurW7w2OhJ0F4j6/nB9FJMzyZeaMhEM0trz1i2fG4PCaeuI1jbjiroE05e0SJ/w9Q8LW9GDvE43qvZS2P/2X35irbSxvm/XNBqAIE5Qj4HL7H3/iOVUH1mkGJBEDKzJVZECNTO+3Xx/KLBMV+FTjr4xBk2Ftjv1Qp1oEM1HFKERrgyFaBbMJDO+WjSvPOfU4hkG9Eq6dcQWco7S6ep8TTjz98iuvmHnFr6YdNAnoFAwcis+SE1NxSYF/Pv9eBzMWQQhACgf2rbG9IaEmqjRS2tG7af7M04+E/HaBQiI4iSQRQXA6nQcddNCll176+OOP33bbbccee2zN6DKRkHL2oPMCL8LTBw+gw3foyefNX9Ns3SseNmYv64jRfoxejsViukZD3V2xnvb61cv+8ee7Sj0ur1vyeAJO4nETn5u4CJ9393PvIhlAktItPSu/OGHvCQE2e4gcER2EF5FqAtjL8bDw77e+CWHrtFiEKrjoqlpUaDqlHW3rv7jonJ953ZKxRsZLnCsfwbDP4faJHHERYCVi7XOKgkXnmF864Y6/PJlAJKwA7XoPVXphltBkqkFkh4p/ku2y7dPyrHet1xwrWCnp7COwzzmD7GZ92wFPtJbG2JIQTsdJFWYS9lAANmZSz+57wBsN8IecLp6eNzIbrWuAu263w9nUwX6OcWOm1mCOXTWBIy1BNZUhQ4juAekr+GGu0d9f7+i+kuw7N8IikoaKlvGug+EEJ+XY8ZnP73v3YbdvHxsj5V1PAsNuQG5jhXeVHtw6JEwACe91dH0kbjh8Aqtt2+XH7w04zdRSQZ/hABiLYgERpxx06q1taOhRtV40A6ZmurS+YK+MPrOmKlPa8uCN54CGxHuJq5Dkl3zQ3NIKF9TRctNFafOZ03evxDsaqhRTs7JDwgYSRT0l3oF4ORqhaufyhXP33SOPJ0SC5FCC5BxVWDzuzbc/sYhgwB6jqxpwooJm9eyTs/I8xO8TIckL4hZBmli+2wn1lHaydqYptQwGgxmqrxDYvhyl8a7Q5uUHTxztIaQkD0zyPO8sLa185OF/btpUHw5B1C66FjeoakMyuuwvt59f5SI+0K4F6ADeAfXwj6368cnretAIiyol1Brssaj86/HeUPvaVYuuuPTc0nwh6ALd0c07RCNFDVNN84h33LJ4dzPEZbd/9t4TPAIQyxWeQVmGReAHXBEhFXMXtmyhiLgTwOwEeW0UnYLXtEyVloYlbx42OZCP3WS4EnDiGWdfsHTNBo3SrtSiRYLqgNVv+t0NPo/LIQgizwsih0jYQxx5JFh5yq9v3NgLyyKx/qIErBK2RzWniXq77AyMhKFxjiJ+nwvnM5uwqR3kdlstROmGdx68ogg8DYITjr92M4WoY1Xr3XWRcCYJ6XHatZ52rNl/Yjl7uNO/jaUYnzfI807R6SFEOvjQE1etb2VDCX36NRnI+6yMU33vwiYeu+2i7xlp+yoMLbWH6gmm7UXBkKzTWOfLTz9a5hfyeOLliNdBBIe07yFHz5n3OfxaB1UR1F4aVWhnQ9s3Bx5c7kHWVWiO6OR5Ea3cZedcO2tDBOYNwNZpofdplTDmjmiIRgEJv/ivP1T7SCEH5lEiBomUR7zBvz7yaDwOS5FWVDBeQq5f8ulJB06EBTZCBMEAw5DtlC8geVUPPfWfzl5wFNfV1NOYdm/cicVi9XWNs+69Z8/xVT5CSnmYQPCJFp3EJxGfk7iIGLx99of1cH6ya+OHM2ryyggpEMG1BnAsL+DHAUlWeQcaiqWLfv/IFtSzlXAXlTG+GB4f9u4Nf/vZy/mWcwDhRMl/7kUz//XsnNUNnbWtkdXrat9987XrZ/5in7GBMXmS9VLC4BVCiNddsNuRJ1zUGmagNwnrgDsFCfcxODNh9nndDbLbX/i5HrF7RqCTCDwR+FAgHEPXglxgS+73T735B2mo+adcNa1cq5Pr+VtZH1RrDIc4Hcaw7aOrmvUx6LLAF9kklEDSLFBz+mLOXKu+885nPF+2Ko8gYVP4uY6fkfOHlwTMfv6h/D+8emfg2m4vJJykycYvXnqoAOlrACMZkIhzEsHNB4k4mrgnPPv5ig4wnaArK7rLstdd30FjB8PgTKNSNdy99J19ShCZOYuJf/fpF9/YBQYTXQEtMQYZVmjLmdN3L+fRKG0pPoAcwCY8esoQNmFj+Z+9sBSqRVUa2dK18v3RBcTvAkjrdngJ8XgLJ7z90cIIM4TLluck/EynCVXtkuN1//fg7QGsgGmC8BNSdtiZv2uUaTSBr0CjwQquDDBnwnQkjCeA7pvo/OSlx8u8vN+IJ+QEwXX3XfeFQwnrvagpzAzVS2lY1VqU2KZLTzzAb2iihLjziTOPk3x8Xmnlwad9vYWGdRpJKKDrY0esXPPFFVef4wZDFGwlQeIXDedvQ4qGTbiAeGuWxzvbaNuH8x6tdBNP0EE44vG7nOgczaFli9H4woU4iNh8/N0lEHeIbogQsklVDHeUKd207O2HqgpJSb4fAyUFMF9VTPjgy8UKpTEZTMjpGxvCdPOmlZMmVaW0dgdnRloKY3983PyGSDvKFH6bkmgMHfO/E5switBZIO114YIw2OShVluh0Gi9tPWbC6aP9RPCu0dNOPG6WrQW6lpU32Vtwun9b+xh6iil88t5T+WbpkM2bvHbQMJOhwfTinOEiA7Bm1dQPuvRZ5gXPvK8seWnjNc3lHQ7Es48Rxm/1gFBAQyGEcumBRqq+/2vTi6Eh4MEEKoFK3d7+f35uOYEPzONbhr4CcOgaFu1+p0p1byXhQyLIlaemTMrZt70zx6dhrR+D0Sf6id1Gu+mXfU3XHTGKBfxOyA7dKG7kiflex55+squhH3QpYccw3Px2atPlUlEEggvOgUBPLRBunklxFN0zq9v6ErQrlCcaipahu1Xok1NTbNmzaqpqWG9IBKSjyRm+eYMzPFO9iHOgmuf/KSO0rZNK2dM8DA7rYeDNTskPYOVQ4DEgoPwDifhCCcddt61jJaAJqO4fKGjZThG9c65T9wP/OEuCfjD80pvuOOe5q5ITEabPDwSeihqsBXSaOfTd19hGJ8Fl0eCG+EmEj5v+onntuHyGUYLJ9CUBROHicOMQh9hb9MuzGQA6PvcIsvdbbo1/pjdyHL/ZtbvJK62oFk4y4pYp+VaI+uH2RUG1l0McNjnhFyrk+v5fW435K5xfRMJ2wCwBYYtGAwr6oiBYU2iDxLOtZ7f2fl2iViVgIOWM0J6oe+zNviwsK44XAt28YyUdz0JDNdxubX13lV6MDckbGZR6mcT1hQI/epZdejkwiDaIkENBGWId6JhlJOKSWDMIZfcADRHlKpRSC5iTXkZOoGBYVmnqgzOh2rbv++9toyQPKeXuMcS317PL26rjUdU0Hwg1lSnPVRvO3P6pAxImPizRMJWHC/ohopCm5cfNylQhERgpvdi8bFnXd2t0W5Ij0dVRY/HgAuHmYUZ9kpGm2mk+eQD9i0AMzITgoMQPyHjX/5wM9POTF0yiaitn45r/plS/a1n/l7lQtsuCDJP9Je//Nq7sYgqJ63AYUt4UYCQei/Vutd++HQRIUbonzvIuRkwJ0SsOvrSu9Z30x6ZdvdiCh6aeOjhP4LWbHqyOwnxCBB/iTZe1M0Nf/hRxFOzPtIapm1Tq/PHgOFYIvnB/KJ8rwsilO0f/LWLK6x+/pMl4LcMkZoUfLCZS7bSvuztB0sxptrhCqIDY35exdTPVjZSSkMKuLOyLRmJmbQ2GlqlmD9jz/HH/tjFo+IOajPYvngA3j7n+H03a2AZhg1sR2xjDonWrnl4O/4/gE0YHNQhmjXomXrOkhDEfKIscr+xHl381pNl6JdPpMKan15bi+o71RI/MCTMRCdT2n3PdeeD4RGHmvllIGGPJ0AIJzm9TofH7fXhQ+iomrTPp4tXMDdp9P/s2wvWdMSiMfrEs/U9O7UPSqvJhq/TaMdFJx5U5QZTMGyia8LeByza3ANBHAr4jshJKmPQKw6FqAwW3zil3XdcdWIxMwsj1ZPfl09IgHAVhIx/+On/daXlFTXvZtZBiccAK/Y03Tzzl4QQvxEVLBFS/rMzbmyIwz1Ccbm7uzuZTGqaFo/HGRiGxSmwUyWp3NG64nM/R1w8Uh6CUZqDAA8Iui487ORzuwDssy31HGmatnr16qIi8FSQ3B6nKHgR1mJ8Bhv6HExDDgGcnx2Be1/8slGhN11xBbiBoHwmjSspEiCjTBCXUMEJhWdIWCDEffLVd7RSyDUFN1ZVsFzKYZpse+wPV+ZD9lIvcRaR4NgX//dNLwoxtfaFc3FcQQd3IMjaPPvBmwIi+KFDn7D4FlaUimbe8/RGBZ/NTNHY9muaEtiq/0FsOMT0JFVDVIU3oDkaU69C+yDMWN6qe6f9iF3WigNnEflRRMKYjjDjbQc5mHbxLHYGuVSmP+WqaWVRg206ZSvrkzUSNn3WR5BwptGwTV33vfhxruNn5PzhJYHvxSDbiZUYXr0zcG23DglzhBRX73VsfW/S8I4Gj9fWZHjtcw9cUyYRJ4er+5xDIJKLONxE4kUXKC7V+723qbsLlQs2zfUnglEwyAR6UqM0waw3bTS+7rDdi/JAickj3OiDf/nnFUlYEY6qXXo/JAz6JIscRL2H8HnEMWHMnie0KJB4RgG3YNjsg4XFsZpRYaiz9Lb87dpzixmSNb5dRKj86+yPkuCYqif0pAIZczHrAdAqQhZQDaA7pGJ5++G7S1ApBJWSlwhXTMjog4+9rAusk5oCdYegQuYlbq8JNBsUMp1q3S88eGeJi3g44nZKzmA5cY+b+1VDzMrCw6yLbMkAOjiODDvAs0O7vz3jEHQfdbuIw4XKpUAENyS1clb8++O1HYCjGD2v1tbWUlu7qaOtWY53UqVlzrN3ezmSJ4Aii9TH6CLNuQjJd/p3W7Zq4VFH78cYd66580/tihyJR9qaaw/58X5BXypoGMlv/CSvYkkrhAm3d8F3IhGDkN3ehk9f/T8XARwLYNtdBEmn+fLZb81vUyGStq97s/U+BDDACCm7Ke088aApJgriCHFznI/wLuIqrNr/xOXNSkLTLYfVdPHumL0BkLCB1Fx5xfv/Yl0vECBpalTWkrKqWjmTrULKL04DymNriIbD4f/OeyUfU1A6AQAUjD/ZQsIQ1sjy1mDqGktYmQsaxEGCEwU8gmyoaj1UW/baQ5d9/+KEM/cUglh0ptAADDMGKnzQ2bITW5CBpw4C6XkHjH/48JxIRMk57ZDpyzc2wbSCs5DFCW3JS9VgAsraJgyPcVyjXTEtGQ3TztXnzZgIAcEC53K53PklhZMOWtkJfuzAvKWDTzb7YGi3rNKYSiMacKy2tix6abSTeA0CfiISQYSVtHxCxlbUTN8USRiJiWF9B+YNVmF0MEiAf0Cy/ZYrz4Nb805BBHYuQtyTD7hgc5i2KzRi8dX1E6qB9HSFqr0Niz+q9OD6HS5s4fPpIUKQCAXX3Hovrv2lOVbIstzT0/P11193hXrYGl/j158fUA59AnZlQoDZ3iESSQQmMMH11OxXHnnkaUIkhxA48MgT1jU3wuOuttx19bkBA50KgkMSJI/Aewjx/v2V/3WZFU7EMA2fWv/eP24uhlBgQlylxD3mb8++takH5g3DPd043zC6wkDXaUKJUhq55qyjAxgzLOK34Y0uFRD3hH++s9x4kZm3Q8Gy16YhodRftqIEV8LYSZQzpe1fvfHMm2+89tLrb895+bWXXn4l+8/cua/k+Jnbf3vp5Veef/W1/7zy2isvvTb3xXdnv/plZxIcd1Qtim9I62nIppCrOLK5pu2cgXWXYWUTRikN1BabodhILGCmEbZ2c5Xy9+/8zI239XRW60Hfv3blWKPMYrDNNyMnDGsJ5Dgchv3pw7qzbJXPGQkjdzRPyKgxU49OIWFwgI0mE5tp+5cHVQPk4jiBAw1I8sIH475EBwlWn3Pj/W2Utsr43jUXxe2jQaZqArQG1NZVncZbKd3w3nO3lTlIoegF1OSa/Ni8NVsM1QeiTvvYhDMgYbFm9JTjWhRQSS0kbCEN6+6Ih4EvltLuZMfqKSUCM24QgPZA0kWE6svu/XcYkLCaAIoRsMgBDrG9t+BqvT298+ftyYOhAww8AsY2Ex/x1yxuDaNBB/miqIwUzdb9oc2awtKoJr569fFxhlosEtFHhIJf3vIIc+SzbmeoAuxlYmQBoUAbSzc9cPNpYI7EiptheBJxBUlgzCnX/7/VCcCUGg1heDOrAAPGnVTfcN8dFwZEZuMVUefkUXUszvONv+m66yH8Typ45f0FXUAOCasGkCcxET7/3DMEQlwmPCWuUadeek0XENFE41oMqIgAq9PI+gXjcAwBVbHXA0YnUnz5dfd3yuAnCrxD+FJIqZ8wWJl82bIIAH6qh1qXfVYT4N1OgB1QSQ5IpwEMk+KfnPWbDkp7FMOOxsRjl/L2Lw+OhP2BoqlHrWyDMY8mXFllw4bh3fRve93i8fgHH3xwyimnuB2iD6ECtNFVNPqU32ykNIrZJ3+ASJg518GySLT+potPRtTHvnD1BkYukGYZSFh0ECe6qQiQnsfjLyRi8PTzrtrShuTMcSWWQN4aBoDt37bwjdRotHcP5v6FXE1s0yN3XXxipVEbkXhLiL/qjUVNUfRjT+gwWSAY1hRdwehuGV1aYsgP3U7jS46fIjEKOoFwLsK54Vty8IXEWfLsh18bLAMGEpYZJw0+LhEarX3jyfsKXcSDi1Gir5jwJYQf8+LnzS0UoiGsOpp1Tf1vaaLospt4+q+3OpkN3Vj+cxLeA8HGjoLH//N6LG39EHIRpS5klJTFb700KgAJnyUWJAP9gD7PRLj5+t+VFlQQ4rvkqrs7mIOEHMKA9+SL/3dfsQBu0uhEwQvugmDlxJUdyW7DJgzcfjTaIK966+BiUgxmZi8hZcdfeNumHpg3EjixsyrYTfrMG6QHlj305vkv1+DqJDjNW4NG9BBpTM3Bp9d2KYzoDC/CViqtVaOB+r9f6wc8oKEPTxLoqdVuGlt34U/2gPeU6AOSMGj0jvtYTbUXePDW4lwCOAEEpaIfrYe1SqrBu4MxGVjjYsjCgG0e4A9DXjD9BJuykhn69jlhgLtut8N9bjfkrv3GA508goRR3Uvv+MzA2C7OYVkeaAyMHN81JDAsB+U2VHrX6DWdbh8kzGyqlHbR5NI59830AvoDg7CTSG6wCWPCEA6chN3le35Q391qBm6mbMJo27QyY4DHMCj6CtU7aOyrI6aQSi/4yxES3G2/M3qSoPr0oIkQzzO8o39x6MQyztRy0PsOQly5PCKOHwIJI6Cj4LeYgFbQxs9fexBd6Tj03UVrNPERsWbaeb9vh8qrSXDMjqUhYQzsicRQO1z95QwJDDqAhJ0mnCTSS19taEajHKiwjDvFplwyGKxpoYbNS/YbP6qQ2ZOJhwRrxkw7rrZL6ZXR5zfj4GMSg/CvXhpd+u0n//K4CScAhbTATPS8IPoKiaey8NDzN6D6CYYxPYraD3sSdKDilld9/dFjEOPr4NEm7ES9kamOkuQYzUtjX3z3q26kwjZ8JoFwFVIKvv/OmxNrqr2SGMgPXv/n+7ugj+IxtF8lAZb2Rjo3XnDSUUFCfMwiA5Jxecf/aHUnroygKAy/eOzcBLwgkxhLyaKpUUuDMwAVv/bkg4VI0S0a5MEOInggLNxRdM3fX9mCCyRsRYUpd9vwuA/10wGRsJOIiE6CJcW77zO6Zo/x48ePr6mAz/hqKPfbSmybYG4OQfAwM5vDSdxFVaeaSBjMu+jiqoNleEh1YtewCZtIWIFllkjzrJsvLpMYgrI89IGtDpEwDmEHjhKBB7cIMQihCmKhwxl47F/PAHMVpp8yTMHpSHioXoe/J3qjMAWpnYvefb7MD34OgkOClSOp6sr7/t2NIA2WeHT4ABimiqIzxm9YEFGpDC6peieNL/zLzOnoJSGJMIiB9NhHCKz1EHL1rKe3YO/i2hk85Mj+pWuwDtjTtPTNSjwN7swT4vMS58TpZ9/ZQWmDCsT6g2yW9gnnqEmqhM895kDkXzbAGRp4fUQIEHHU56ub+oBCGQizqQzc1jgB6b3JrtoZe4wFxmYzoy8iMPC1FryFor/85+ddE2Kgk90bf6lG2zYs++hnx+wf8EJLjjjj7CVNYA+GZoKfehTmG3nDWdPHjiKkiCNOwesomfbqF1sQUMO7gs1/SVx3YL6+sjlxwN/inTSyduaRexh8CBargSiJ/koijnripbd78V2CV4LVVZhI4b6DLCPguVl9aVSLw9X0KFXb6ZYvfnUIpgITPLhQapGf74iCHQBbZQjDJsQtwiu1kIh7rWyGeRa3ESQ8aI9mfPkOctB+sUFOs/5kP3+XKlstTCtY80+WhWEvkrTW44wzcmRXksCwH6A5NmBX6buckTCyjAqElIyecpRlE1YxqS94H+sbE7UfsaAvJ+jukhvUT1j3FgCR+QkpOO/ORzdj9iNc5wc8aGzo+BhHCAQHwfs5TOONi9+8t5SlDCIBQiru/cdHaDwFqyYk2wT2kUGQMM+RIOHHVe1xbIsCboiZbcIGEmY0Il2ULp/z0OV5bkIEH1ihAdI6UTXdvfLQ89sACctJ8Gzsi4RVjYbjCVChGpcf4SCFhAC1shOIYEHnIeTSe/9db0fCmfiTlEjjrVdfyBHi5RxEzCNSKXFNeO6DFUY8caaRxzA1SIwhYb02Efpm4m6YjgWXBhDOcqIjj4gl+T/+5YIe2q1TTUsifTTiTSqrciwRbaHJ1Q1r55W4yCgnROxh7QGRg7ciFyRk7G9vfS6ETZApjQPFCrNO6YlIjxxDF041mUhAhkQMg4bgayMAO7nlf68/6WacapZKFhx10W2zIBrQHAXMjJNALRYN10y5NfKqQhvxnamF22h0y8+mjS9HoloMOxTBGxY2Nxk9Y1G3wWDEoipTw8y80fb8fxAkTBzEFyAuF/H5EZCx+FGr/akCS/QqCILD4XA6nZIEENrpdLqBzYxzEifHu4jTRTxFFT+7yrAJ//CQcN9e02JUaf3f7Ic95oINTjUQ+S7AMhSGv4ORmAMRwubDXmBuCeLEyXsvXrImAwxGSNz3Xhn2dT0cBtKspqVHTB7lhUTnhPO4iSuQN3n64nbw2o3gNJEJCRu5oGHa00M0tnLZGw95cHS4sJYMCUPMAU8uue3BthQShikAIywABlO95fpzj/QBcxTx5IkQ4i/wZNTBLy6Ot+MkOsjIZ/MGm/XYaVqsvWHxvGofkD+jDNkSA4xEIhYfecavTesrrFOGQqnUSnI0SWU1KfdSNXTOYfuCA7NB8mddQSJS2W4//mlDL8zbsCmqJtNYRI1FcfVQbqLKFqonVR0mavDMiVh1T9D4+o+e/zOQL+BEzBH31J9c0KTAfG5t7E3ATMQshTuDxHiCSsOb3nn8PjYIYCBwhPhcLNOy4C6Yccr5G7sVJPGHsBy4P2BXdOOwbrDVBWALiiISTlK1k8aXnndQGZimpXxedFnJq7IpCAJQmuXyMZfTbP9DsJIjjwh5LuJxkFFS8LAlGxjit97JWQITq4OyF032V8YzM73v4C0w0Cf7imzdmQPdd6Dj9rsMdI79uP38Xapsb2SqnONoGP4SSTV94DE8cs7wlcDwH6G5tWD49lR6zbcOCTsIKRs95Zi6sMyi11RAVIk4OOL1UK3twd+dX0RIHsc5wSAMMaagiHKcyzsKMgFVH7uqHXQd5HFlYsc1fYSFjMSS2Uu1RBdVWu741QluBGSE85dOPnN9B9pF2StYV3RQVjIhYVR7kQoqSLjxlXsc36bCTQdb4Td8liNUWf76/13JlGV0Sg2CkYcvImTS5GkXNvaABSRJ4wqQ4EC0sOWuDCo1IL8IbVx4XBAANNBHYyZRT1AinHTG7U+vRz3LqAjrDJSBRmlLAvDgus/f2M0JyUW8xC+C+s7tcfSl9ehDyUwfGUeqzewZo1oT1Tb+6Y5zmaOjgOojmliMSNsAACAASURBVCYdRKycfu7dTegziV6ahjoOFkXg6IhQ2qD1LJpSKAbwV2iyAFG6nW7RUeKvOmpzGPT7JOriKli2Ux8IlVYVvA7TqGDZwahtIkI71x+zz5g8wKk8uFhD3uAiUjDx8+UN7HXIzkR2YFBwzRYZ+D6963RIeRJp/uz5e8ZJoLXjAAEOa4cb4499E39+9d8gh3MYyM1MdTnTazf9eRhQwbJOyyj9gZEw5w4AHS7HYVYZtBmi567AEZ7nbdppqsjCsHH8EkmS8A9OjrhhXUbyEB8g4fXMO9pw6MzULrZgkP69C9iEQfxWX7AFLF2marRn44pj9t2dcS+5jKBTADsOA9GJAuR3Q2stjBcIVMcxI0iOwO133BczwBmQE7MtiQtzlmTTu50dRnf9WA9N9D5z761A0gWBAxC0TiTxjvsfUgHLAU1cUjM/ECoM3tHwwTwpLBc01RKwvNbTfEDNbqzyQSLkwbiBDRIjz37DrIAxSeH8l6Bq3bI5DwYsd1/MRkQET8WRFy9D6rgMCbbNC5nuEpDkmz0gGtA4dVN15XUzj8TpAuzBbDUPXXjBfjjn/dVm7C6apdnVTKIFoDhQ4n+49NQSi7EPm4A+z17i3v35TzZjSmEktENSMnAWAe9QFYSgYWQMunMDvV4Shi9sukJ715+w7yhWKy+4UTvueeL1Jkw3y05h40LJFBEtJyI4eyRqly/0EvBNMjbmJw2GWMlZse+iBhlNwMyWHGf1siYw6y5bU2BIWI3j4xil8cVnHzXGKxCHk3VdTq7R28VuzBysILkUT9xOYf+NSGWpgLP6VoDbXERif3izKedy7eF67ogchmvPjdR7RAIjEhjOEjDn3pyRMJo2nYSUV+1xtIWEdfBhTcRpLA72v3B49f/2DoAOyrhJIA8vvnJ5KZ9wZUTa98//7322Wo8OnaoKEbOoqyCShFcxaFQapZFoy6oJAYCUbge4/V155wth9q4G3UsDJ0GwCccZd3SadzQqOwYSJtXlk49jhEyDIWEGwmSVRtau+uzJ6gDJF4ngzCM801dchEw599x7FMhWFEvSJIPBCuM3wlBWINphSLj5m2PzIU6Y40zaKYkj3sKDLr6v1kDCrAdSY0ijtCMGHt/XnXlsNYcaNfG5eC/hAtfNmtOGNtLU2f1KNuiY0NU2qtYne5dOrQEW3TwA1RBzCKYm19gPloQTaE7H/L4syy98Q+A0ImG9e9Fe+WIBwgWD45UQl5RHSMGdj3zYJLNoXhWdPROZkTDotQr6MFs4NvbeUw+MQmOXk+OBbBoi9Cbk7370ptZUrCVrFsMZ6U3sIy68bLSZRpdPn+B0gz4HiVcIb1qFhVHeqh+vqo9HDIalgT2HzSchDV8NcjC9WsbeAEgYutHpAfjqDxJPHudySS7BI4FXokMAJNxng4SuPKIGU1dnZmECfu4ShI46PcRbVPGz3/xwkXDmrkH82lk/+8E/VbhIkQBIt9jlZMDXiQz2DohddYiwICGinwTL2Qt4WBD9++8/Y936ls5OgISKDhwADCiyoZgODtgx9sBFqdxBW9fuWxWExE08cQRwAU10Ll2zDsjrYhpLXmxxZTGSPfZtPoDg2R4Jd+mxyBdvvD0K+pjkc6JPYPZLcb99D27vsHycDZ8JZDuO0/CKe35+CBJkGfgOchHzwVNveaTeiIAwAX2mcYsTni7jHIo5g3vUZB3VV3z5v3/m4+If54BrM4CPHPaBH598NSJhVYk0Wt4egJ4QF0PUgqz84dKTSjmc881hjGPaP+HwS+qwVr3RKMw2BjFcWm5VuBTz+1CxWtixeiL61Wv/Kiak0Adhx27IVV704fLaGM5jbCFSU2lSBoEnk8BiyJYYIHoFXklGB4Y7WieOGYsZosBTBzaDZ81BCvd45t3lYcjTZIRjMEGnd30mIWZ1jHlHIxLW47R36SlHVgPLt68QX5LfCRKWcJmI8ETKd+1Tu5Y1o19m+6xal8tJmZ/f9OUt+zm5XHu4nmtv70Dl4dq2kXqPSGBEAiMS+L5KwJxvtxsSVsBhOJGkaFeMbr72jBkSWmYABnNoLYFEPS4ijCKkeNL+p29qA4UHlA45grGglsaGiib47QE2vOf6XwYI8YsOQpxS/m5fru9hGqqhewG9EgQXMyR8xiG7l6LeZipg4OCHxFVjy/YYFAnDPVPmk2ioiyqdZ83YrZBdjfdCgk2hwOncc/6X7XGV9sTDCqT1UEFbM5l+DYUMkHCYbln0kwKAoMBWBUGLqHW4gpNOu7kxExJGfStJaahnw5cVLjR5gMzckrPUX7zHx6uA3sc0bA45pGSMwo5R2lq74oO9yzxjvMwC5veV7/fOF5siOlwKExulYLANCTfRnsV7FzhMJGzI0iH4iVjx30VhM/2mamiZIDdDl2V5EWGXLQlAq1hlOmnXxvOPOqSkDxJ27XbEmTd2ZcgePGQbUUFWuqla/9jdlzOHV9C2gZyMKbh5xFH++wdfYlow2joGiH8zn4QdgYQNW53ozK85YHWrvKGhoa5+bWPd2vpNazZvXL8h01ZfX798+fKPPvro9ddfv+WWW/bee2+/H9AIZgzigIMX44QNJAyrRTl8dgWbcMb+gulCpfEIpXrbxrWX/uJUmDQIKRSJB0UHxGpgDARSNR5opRm3FgsgZt+SIAbeffeLzm41LhuWYfutBhiRUUprH/3zr/yYOtglchBGwJdN3O/Y9iigwoSKeXBNymgbHsZ1PMoeQJhJAAokYzQcffuZ54Ow8AGbIDhPO/Wc9nbI02ZuVvQAUPNFvnnjQC+0lCfI8c4TAWgF/Q++9e1aHXw34BFkwNL8vf1/+9DB8yK62kL1DVSvP/GgqeCOAhR0sKGMOBCqNP7TNV24UMAIpfB6QyNhSK78m/tfrcXJR1Yxr7g5b1gTCBT6C12Rla6mmaccZjrpEJ4UjBlzYF1IZfxORovQrpxMAI0/HmEXAtdvVY2rcowq8WRP1xknHA/+44SYGdpxxhAEIo2+5Ja/w0okIGGIxWC1g/ps48YqAu86RnYQh9ggvY1SGXINfOcb+LlbawU7vjb2/s2mvONrNHKHEQmMSGBEAiMS+CFKwHwH5YyEIfaUkwhXUTn5KLtNWAH7rKwAEk7QWMOKD14oQS5Q9INF1QMUKokIBRh5O+qJFz4CLzSddke70PMXPeFAb8EIL1jTj0c3frNfuQQwTiogpPL0C29jDDSg3jHda2gkLPIQ6DumdNKxbSqwjGa2CYOyqFMdUDVQP4GDcyRWu/Cqs48SWOReXmFR2YQvv25BfhvIccRgMLPq2Nh/wZZN5R7asmJ6EdNQmaZFBBdAtEmn3WzahJmKCmOPqSJg9JXXPfGHCwIgLR7JmyXRPb605tB13WoHLhRkN1LVWLRDlSGHKLB/ddXWL13w4ovzVmzubo6DLgYLEDLDqhmRcAPtWbp3gYQe4UwNhm+e5HHOMetCoF4jw7NB1GTXYlNI2OodEGcXDa/Y9PW8CX4PEiDzKZuwMPbae1/oiG+FtqlDTHKyhyotmxe+wxy5gTeLGXlA0QUD+JRjLqlFcrVkrMNU9+zKP5bNJwEqkc0nYx8MYBM2kLDLN2qvk1Z3wWjX1F5di0OENhBcDb2pitrY2PiL006BOG3oBg65o6/ZiMxAQCk3goRZrxlzgkaTcRoLbVz05bEH7MU8jV3oBs3wLiPqxZFieZkyq6focucT4nn40ec6u6FrFBOUGUtvmfuqV2v77JhpEB/ME5InuQVSRMjul137fyxOHuazAWAweLSkI2FAa4kkTarJ1tav/vfhpx993NHVE4nqiQQ425gbQ8IxCDdV2+fNuqUcBwZ6ngByFdwFxD/2w43hRlwPM6ZK+G/IDWM+qKZqzVRueOr+m4sYHHeAGxCzWkJkg3fszbOeNbypGYA3XHhg6rRswmV8H5swIOE/PvWZwfuFqezsUweWWYAGeDwbAJ5VWelWtqwel4/jH54ogRDv9MNOW9/cXtfRubmhub4OPnW1TZs21a9ft3njxs11dXX1sNXW19c2Nde1tjW1tTY112+uX7fmhst/LeFjme8yQ4ahbQIhBYecfGk3hI3IKixToON1Jh6HIeWY4QQAvGz50SDA1+VwPBbGGSfD6TvpkDXdgT/AzqqLdVOrwF6B1m6fwk6SxchtRiQwIoERCYxIYBeVQJ/XirUEbR7fKiRM3ISrqpx0TF0IAsvgRWZodWAjpRB3FqFy8+U/nwFUP2DNZYiARe250UjjO+iIcyCnjmGIjRnwSpUBCWtxqoap1vXmo3/yEXAiJSRApD3e/LwBiJTAhoz6v+HxCnmPBrYJMyRcXToJ4oSRgWmwnkbvbIQoIKkYpZG21qYtre3NbZ3Nrb2WLqMB6Ss0lm2xmGW00SCvb6KTdqybURlADm10wDOynPr6I2FmsEG+qF7a/dXPp4IBlxBXXl4ewrny4865EeNyTZyWsfpmdxpYDtR4yNsCTdDjzPmR6bksR5NlnjB1cbt3dD0NLdun0M0yfAJ8x38CGcU5qld20HbWVbBuMOAGGjIanfG/dkqb/vvkHwvBICy4wC7HQUQlJNsa/Y+5C4BZF4J5c900qsZpoofG247ZbwxgbGYNBtcDJm4/KZr2ySYF2XgwGWkKNNrkZSvuACQMtjriCQT2/vm6BIgEBQL3yW3TE9edd3wxPjmEeGtOuaHO8JbPDQnT4Z9PeKA+Yisc8FdYhZFpsjfZ2fzRvLlTq8v9PJg0SwPegBNMnIxSy4R2prsKL+BU5Ral/Kt/exujcLL6KD0c1ToMVtnNXz9Z6SD5SEkcEH0eoZJw+/3f7JWAeHD1KkskDMuIkGJHhsorMhgy8bm12LzYnIewmrnv9tDQpguO2G+cg7mQoFc9PFdeV8key9rVdhwhMJexEW6vdYYyi2JIohk7RmlP64r3q5EdEZeXrC8HkUpr9juqOQkZ6dCZQwGBs+ddQUClqHde9tNygYiCFVpBIIKfBB54aTFmEDZvn+HRw2xDumb0Jpyo0GTjh8/8mfl9oFXfS4jL6UFHCT9j9WdeN+wbOpbnRXDGQX91jgecK/LEJXIupOtzo7OShJH7wOWIDPaEeKqm/mRDdwwSRLMFUzaWcn5WzdbZ/sfmaKbXEywX4OwIbjW2s77D4k7DweZotMakXbwZxsN3KJORW49IYEQCIxIYkcDwlwB7s7B17Yxvma3IooQqX18kDHlANDQzAkBSQQdTWt55+s+lzBzjQu86nmkqTD0SCV/16ptLTAAEmBO9hhVIE6JGgPOpa8XPD5gIQMzlJ1L5ASf/tk1DeBeLGvouU5eApGVbkbARIAxatOnyZ7yhmXHTMFGwHQXMPCkkrOEGdhGICIxTmojULnvpvjtqnOCZydAZ6IEQGTsEEu5cOHu/IGjqgeAo1D05Io2+8s7HAcAzw7GNgio1PK1uNgQCf2EQF6OodZVCICL7qJDRMuWo1xcJQy/U0fDyfYs8eagvmkhYdJAKzjF+ZaeeQsLawGDYrAkqf920d+kfLjwCUsIQRx8k/NRbKxibTao5uZaUyO/OP6oA/a6xtjYk7J/6xLurEH9+t0g46Nr37JVJELthuRyqjX3WGKga7lr65p5+ID4mxDHx5BvrDSScM9erlSMVUv+Art9DtWWvPXRZHiACHoE7Yiqnk0ilU356QcoYOVSdd8bfM85iFtiz/sqcpVWVJqNUDv39r3dWFfuAkBnJ0AVCgn4BPKTRKoyLMszoySEY9hC+4Nbb748lAa7EVTUBM8wAmx6e9+hviwnxOYnfl+8nXq9QSRwHz/lvZ1KHaWIQGGy3CSuQv0hWgC4BlwJhNRAoEMCb1rwO0ktRzCCloENAO21btdcoj0UPBoMfGhPwV0/bEKLdJqIzJq+MLbAkBlMqDA2kXsBFwLYF+5VjMC2ARdwAY3Kct9hbOnFtL21FGzDMJea0CG8BgMTbiISTVDOc/rHKMRpd8+TvTmNzqQNC7P2Sy4dZxElBAKgYcG61EmhBpwI7g21jmeQkAZoDMJjjvLzgFgSng3M4OFHiOclDiMdVOml9KNmUVCIoCwPh26FaRhkOdZDNhWjvhlM1XEmEBOsgrYGH1lCX3U5/hzGONeqfGno73aHPZawhxzzh8eHVFFVXcbnJ+qtV6PPzkd0RCYxIYEQCIxIYkUD2EsBFbWCUMoEJFKxXDBa2j03YjoQBd4E1tYu2fHHSIUXgpuohxMmBwy9TOCHJLyEk/4yzbulF+KvDy7gXX8kMGsdodNn8Z28f6zSRpK98zle1DdYCutUeaOFgSNgBxoACQoawCWMeY6iEDIje1H2YSQdi/RIqMKkYKFkBfTGBKmmqK+JxgAxffvbhWSdML+EgQngUWvCGRMJMjWdo7Y1/3cy4ufwSM8cIxF1273PvQDSZZY82a5e6N+tRZplhkkGQY0O8ugrMMWAl1hHDmwg5zTsa7PkMCfeu2K/Y2w8JVwES7pLTkDDyhFnU2fYCG2pokorSnuVHVUOotwO4i2w2YbF67qd1vf1blGpbFiVdefepO8tRxwVXdvgw7+h8wo+94q7HUdlkZp70sc8ehUzH+jwkabsZazSIdzTYhAvE/c9eJhtIGOxu+BhmvJL9oIWHqdpDe7+965cHeglxuosnnnh9LT4waPbv83xbz0bmwrBBwvZ+YULpf8R+3Bz5zLeWBXlCtANVlWRb/eZlF5xzMgNHkoOIImRYgiRL+IHFJ8NSzIlAFZdPiPT447PDcWCJj0O8hrVZgxULeuSyY6eA8wbP8XyeG8K4Czn/jM/XwOIcPJFAFt3/o/VhzFLAvUFOMnonlp3bRMIstBTi7vGCKSScqNs0/zUfIUUi8PMbrITQpKJRexzRphnzKabZwnU0qwX2AjN7GliZIWFmplRpbOnlp05C/xRjvoa5WCC8K0Bco95aXt+CcxJOmfiCsXIxqVommzDwQz3w0rdD24RVEwkzSWudtP2LmT8uYPHKEvBpwxvFHSyoGVtRXuQvLy0tLa20PqOKy0tLyyvKq8rL4S/l5fCpqCyprCgZW1k2tqxkXEXF+MrKmqqq8aOrqsdUjBlbMWZcVeX48b7C6jFTDnl/ydoYrjEx92/jfWmX2FaVcaUVUq5TXBzpxcAVTJ1gDaetuu62/kjDinQii1mWc9K23tI+lwKbHOSGV1uat9gWLn7gReMdxt5kw/D7B9592735w308jNR/RALfvQQg7WQfTdmuT+4Im7BKVaTfjNDY2jeeuR25rtADjbnEIoEW4kOP4Bz7/pcbEOKhGRmgqCKDoYrS1s+uPnJsAWCnIHGWTzrqrNURABKoTaSj+e2EhBmdMpNV+tue+Q3CG9w0HYNNGHA7OB6D3tbV0fbII4+OKh/N8+B+WSqSCgLJhBlJKUyNjKAnk02Y3UsFpBu6ZeYpfrZKwKZTh4c4ix948cMUEs5oQmA92gcJIwuZzQuaEVwBGFZNg7bdIMzKgyLhMZw4YWVXIlskjN2Imk6c1n15UBF4qGJ6YhsSdta8+U3r1iFh1miQnq5s+WZONWYVdghoCwLHSAeB7NOjjj7nerTEMl/t9LHPlLJMx+zqWt9y+uAw9gZBwryBhJcoBhJmkYLZqMApJAyW29WLXvpTEUA2zx4nXGt6R++6NmGrX8DBAens2BPYX3DszHQkDLgMedzQAYItN0XmvvjY6EpYY5JcsEIGhPamqzTPAbEf+G0Qj4MvJMTtCZb9b+GKdtOSb3Y7A3xsaSpGta79yvggPK0iIT4JQ1j5/OlrujCWflAYbGZRAsOghYSRRR8swzqkIzPcN2DsGkhYM72ju2nvyk+fv59R4bksJAweycVV+x4XRjjHMtJBEEl/obH2pCFhFtRsIuHOBf/566VmLiV8kxmIWCSe0lmvfMEi7w3DMJt8WLiKqv3h0pP6xQkTQrwPzF0wNBLWYEY1QoWh37pp0yfTC2AuZcRd4OrOFVx+893xRC+z3ZpdA//j2of9QJ+yThMJPR6H72RCVhJJNZbQ48zsz0JcDL4s9js2tPpcI/fdFBKGBOyw3GgmERyoY3K/R66/gDszJNyN076RkC3Xy+R8vvVc61SOgyFaleUN69azN97I9zCEvn3U3JE+3L4S6CPekd0RCYxIIGcJGEjYUnh2tE0YrIK4ofoaijYsO2CMp1gkkguswCJnhAwzGipOyjv71/czLz4F4BkQbsWpqsVaowtfPaqYlBMu6KshZPd7XljYxWy19oAm412eyqLUnzs6S5swUwrQHm29qI1XPFOwWerMqEbDGu0xuEnDVO9oWzX/+ovPLsnz8EAuHSC+4j323uft/zxBW5celAeMWbCxRB2gdvfxjsYbwpem0y5K28468scF+AunxPs8HiLmEUfpvc/+F7BcOtA14Bmrox0JmB3Mqm3ZaXAXAtXMD1yv/wf153oaXTltlC9gMPEgPuAcTm4s79xtVbfSYbPdD6YGoY6HjUuGFr+9pwuAKQblcak4YWnCvK9bwKE8I8If7Oo2AehxWv/RvgHwHHaKDkPaIEYPIVVj9j91S4wZ9ayeTS+k7/XFvf3/mrFWAyBh5mxMpAJpvzOWmUg44wWGOhilNKw0fj3VL/qIa9rPrmvDXsA1of7dONiRYWcT1qgWBS/nJHQNrlRZzUsTGuspdshehiPsl+AF2hPa/Iszf8IwFWTJDiKMxYWTPMw3him4RXTkzw8eeNKiCO1NZeHBsAOqRhNRqoWoXt+68p0iF+QpBijKeZ0wwN0keMDCBoNVzp4zKVVm+YR1hWX6YUg4CQ7SkE+OUbKzyRBpCKHRyMUOSWlNJNxO6ZoX7/sVxqoQhoTBEwKRcPHkwztkYAdkpADMvxpkkHFLyQpKTLbgtaDXf/nyLIY/cSkPSKVwnUkkYtGv7nquAycQ8PBlv7HmmkxIGCnKXLNe+qoBpxic+NIXNFk1oIks+5pZEaq0zn9tT48RaQIO7UAmWHnKlX/C1mCrjN9mbF6/g9boMSYoDf1awE0ZqRphuSqtwweSW78LD3KALaHCJIdVtURlyGGQX+64PxntYq7tTII77ma2K1udZQpWV7XNGzbie2/ka0QCIxIYkcCIBEYksJ0l0N3RwzI1DGQZ3j7e0WDXwDhhw4oF2oYOOo0SeuLWmaMIKfRDulcLCRuRZyJHfPsuqoU18gSlSV1LAstMgtLwvRedVAlGPSchFYVTz1ofB63O2HTzFWqoYNsHCZt6GdNS2HdKWbCUPVYHTUm+9+ITh+5WzNAdgHzJf+bFV3+7phEtkAna9u2JpVkgYWZvhi5qoXrjSdP2CqLhlAPLlITqePmfnnqbIWFITmQCXaZRmQIx627pM1iwVL7sC1kgYW0rkHDz/Fd3E8BjvC8SdtbMfn9dgvHTmN2b5f9WW8Ey3/zxEaNJgCOQTBVXi9DpwMmQcGNq8Ni1MLNs/m+IdMjdjPUbAAkjUw9PnAWufc9Yvk1IOElpJ6Wdp+27p5uQH53861bT1plpQWOwDh92SFgFi5USgQAMY6BYzcvYFXCQdWLfPyuaFpLltlC49oFZtwK1FEtqLULwtROzLvmxAI8gLxGxkBTtdf1j78iUhkKWgzQgJlmLU72HKiubF8/xMDZqHrI9O9EmTALTvmkCJJyCvprhDg1HTBis2JAwA8OYmA0WBNmHUfHbAhmsoOMk1dtp6POHf3sUJDHGOsOMCqnaOEIK/OMOrA9rYRMJs3DlvsLIsG/NqxgnrDUsf++xPM5w5eAhrgEiyeE2fMmv7nyhnSFhDYMq0m3Cd1xyYikLwzZfZ1kjYQTj8LJiXuHQl5u/eGO8ZCBhELBDJI7qQ8+/DUNt4R2QenIzNKrfIfsD3u/hYWezAdbvl1t/wFrGhEuYYt7ud8mtfpmfkdyusTVnp8ufCaSpvsEhCBk/TtHxw/oIknM4fxwj26ASyHkwD+fBMKxH8kjldyUJdHf0gFoBLqTpb33zfbT9kDD676UhYSCCDtGmbxjzisQ5JVSnRM6WYEOoPu+6f4YQgSUAPsep3Na65ou9CxwB0KIKiHvKjfe/1CYz/YEpEZaXJNMldigSNtUW9samNN6x5ZVn/lVV6A2YOTiIt+DKm+7aEkIITGk0iSi6fdHhARMJM9fojDZhptxD57RQrfnH1RUAF1FtRGsPT/hRtz3+OgBgyE+F7rVmz1kaVfbqiAUhBioMioRHo014a5Dw+veeq2KZhEGZFkQiGNzRwtgHn/8itcaRfUtM7Re7JU7bvjhjH3ceUgaBro7pYhEggE14l0DCSJKk99z4yzM5QqaddEEHLh6hGAbqzMzHhw0SZmlsMDiW2TYR8ADwsZvUrEamjZ3MWr6G/AVAzpeI1L3x71lFInHzDgRrTg6BJOBJsDoKhHcTPp84xwul01ox5pbNomgvZDGkIZpYVjv/KQ9SE2PksVNgSDhoImEb6LUDYKts2YQHKDBIzPKWw7cZcYxIOLHgrvP290CIsmBw8xtIOCCWTFnRGutBHiSwJzOjZ5qAhtyJUbW2fuHLRTzxolMIT0BSyDjmtJAwPn0mtLN6QtMHQMLOWS/NH8omzB5sg0EaiLNUed1Hr1Yx1nSDd5An0pgpx13cgz7GlqE1h/nQmkJt78QhJbKNJ3yX5t9trPr2/bklfHy1AlEWpW0trYzLvf+3GcvPIvp/CN+MwWC4fnMj26ASyH08D9eRgEQcI5UfkcD3QgI9nSFcNf+ukLDSQ+XaWb87r1gkQdHlJqKTcAIPRDXIn0UIKfWMnt6FlB1IKNJLkw3/uus3QVBMHYRUuasOW90s91gmGdB3LMMoU752KBLG9Mh6DJIk64lHZ90zuijPzxO3Q8TQPO/lt9yzuiUSoTQMFFqglYUjUarGaPuiE8sRCTPXaAFNvZm8o3GhQqW0nWotexYEfYRIzIACGjpPSOHv//kqJBVRgLILIqVztYHY9JiUsmpTAe0H+yBhg/saEKboIFUYJwzZcBHqCAAAIABJREFUWYZMRgX3REUHaqsll7/xWCkyWjlRmRagXcAJTvgxdz769nZAwp1fXnJ4GSJhHuhz05FwUwI8RVETtdpqU8dsRcRaTBcf+Nsmz1Rxh9uENV0OUy2itW5prW1opbSZAvsZ5j1lrLRW04YoDB8kbHYBQl9j2IN7MHRlfzCc6gvrATEOsQ5mP5JlNUxpMtxZR5W2j155opBjAaho70RMjCNThMHJ5aGduPze597qNAJSGcEyQ8JhGlm07uN/+nicyiDs2CkCzXIeQSQcBY49ZfDPAADYgr6DImF96T+uPdZHiEtwGEgYTNwcrDgFxn1ZH27D6E8gm986JKzXda17r1wA9jCM8BcdAIYx0p8zbMIgYEbkZQ+/2X5IOJHUqKKsfO/FElzWgpsD2idEKhl30CkdlPawfrEe4bRBMOgO+4l5Ck4O5s7I/ztUAlZnwZjRGWlWT1f30kWLM36WLVnyw/osXrZsOH+WjGyDSiDnwTycB8OwHskjld+VJBDqDjMMsPNswlTTMdpMx9yYYRpa377mk0kFoK55QGm0iLMg2wUojq4x9zz2fpeBVRSaqD28ptBNiMCXEWHyxbc+hVqvFUia9iLFH+0QJByN9qhqXI930dBGmmxY88kbkyoKAM+KLuLMI76K0T864c2Fdd0pHcrQrYDxKxGiHUuOCKYjYTBoeCeeetNmw5pngEVEwhql7bHOtVPy8/yEuBEJo8nZSfJ3O+u6+yIJGo8a6knqhjtEX4lR2kBjq39U6g/2zaKUQsIxVIBZTPiAtbAh4TXvPlOBSJiZlQwkzHmIWH3Wbx7oiAOwyXVLqbJ6nMaX/u3XhyPdtoAGPWA/IpxEuNF542ZskcH9HuWWCSXaB1Q25YwV3eFIGCPFNRnIlXRwfA1jF8DqiLUuBALP1MD0g8MJCadEDctMukaTurESZqQSZwG06I2SOtdCwoZAWAQoazfjjYMcRXo0RuOhtR88X2QBYMiwg07GnINwbgyTlwg/qnrGaa2pZ9bKlB2mvd9+OPs2H4/BucA+LTpEiTiLSN7+cz5rygUJMwp3CwD3KaT8pU2bsALe0ZEvbz9rrwAHSFhCbn6A4WDhdhOu8N9frGtGJAzDng2KNAENuROjtJH2LtunUCoGFMpx4JQsiMThlQoJKb7wjtmb2UOrYpdkQsL2REZsSsvVJpwEqkD9m1efKmItY0gYRO13lU7anICAgdQa6ZBt+v6dEAr1pvHtZ1FD+/nbq6wquiyrsT5JtLOozNacYs2xmq6rGiBhnYJlmLG99/+2zh8pDAsJbM2Y+CH9Zlh04kglRySwK0kAVSCAWvpOtAmnIWE1AmbGcN2t558QQP9YQMIc42nlMJOJi5CCvN1OaJZpFAKAe5a9+aSHkCBxE76KFB/y0vzmXlVTtV5U6Pp0DlPxskLCZZNPaFEGN2ayizNrE/iSy0oMFODeTQ9df26lA7wEDUVZKj/9N3dvTkBAswnhZDAaw55uIOHObJEwC+Omemeye9OewbyAiYQlIIeRiGvslMPP7sTUpkmVRmNoKNseLw47ZrJdLw0JW3w5BLqqkgg1yzuUNvS6BAUY2dFsv00v2pBw3cdzqk1cLTKjLYjSTRzjp59+fWsEAF2fzerpPsetXXYC7BpI+CcecG11pCFhfsykQ89u14GpFRQuYCDuZ0G17pRlwaqBvbAzkLBFZ6RFIW81WiZZL1o1Twe9JgaydzWQrxuEQAxAfl/zCVuVRposIAxLIGdWGHl+WVQBY1RmQ9HeGynbPkhJRhgLRASQnZdtGuYs6gnRZOOch39fEAAEyXOM6x2D2cFU7AJXaSHIjZn2cV3MTLcKSFiDy/bQ2NJv3/xrHkecAgEUikiYk4qJe+od/3w/dyRspTTLAgnTdhr76uHfHuVFm7BEgAobYDCsnzmIUHjXCx/V90HC/R6xNIn13YlRuoVGlu1fIJWBIFJIWBKChJRc88Bb9cgFgV4JfQaa4R297UgYvKMpzYSEPYT4P1rXhjmN+1b9e76vQ8pmPZlU1P6z3ndd9UEyZ2+3qlmTFYJeXdXYZwQJp2YtS0TDsbDdBsoueqHh2KcjdR6RwLCWAGqTzBr8HdmEVaZAxta//9QYApxGhrGOd0NiSMOBPEhI5QMvfBuBOL4tP9t/QilgZoEQ36E/vwVoghLxhMxsxn16g2nL2SDhsWWTT9iSHBwJM1sSIOFEMhKLhVUtGu9tn7FvjZHKUpLAGkwC193/9KYIjag0ITM/SQraNrhPR6iegEXuRIh2LT08H9iMCaRnQae+wWzCkCxEDdXuGcxjlljQZ0GRdBBSRlw1S5p6IMtlgsYjyE+zNe8IJrrULy2kYS/08Y7eJiSMtwJVVo13fvv2JCdYyDG7EcQJ45oCIOHSqSc19VgLCkb17BamgRR4I0UO/CJOE0v/9msTCQvgG20MM6H6pxf9oRssqCy7ir2tZrnPmBpyNyVCW2nHI2HGl4Y+vlqSRlUaBd2R2UazAMAWWBkWSNjyf8aVpiilDbffcCYh/tETZ9Q1dycwKVBcVRkSZtjW1hmmWzUMHQVhcAxXb8CpH07GnK4qpd2hKFUUmgxdcfpRpaJhEWZhDLhu4wJ3XMlHSPmcjzei7dEA4JDriHZTZeXmzx8L8sQlsCBageMEAkBxt5+c94f+SJjRZtn9pU3vaBMD49qELbcZg8Qpm7BKk/hRIJgi/vW8By4B8Mvx6UiYEGfRyb9/ZDM6DkC/M2kO9CClCc7aQSQcXXZgUKpEKzPhHBj35eCB+a70L//5dstASFjtg4RhBRHDIaSsbMLg48AYpDUGzBgShimDRy52pKEjhNz+2NyWlK3eqnl2hSykYU4Q2V1wyLNsEwubjFiu6SF/txNOYJhclXeKfd2SQ3/zb8Yj1vkjhWEhgZ0wXof1LYZFJ45UckQCu5IErHd5/4LZzO3PmJVuE5YhwLWjmSbXX3jEmADL4clJoGhyQJwEmIj3Ev/Umh+d3tHb+9Un84oR/0kk6Ars8fz7a4ApSAZoalNvzbobACArJFy+x4nNiSGRMAPDMlp3e1etWTRlcg3kwRR5QFagcPqOOPuqRpW2MjZnyPeJIbyQrIoh4dhWI2G9t37PfFcAdXAgswUaWA5dNEe9NH9V17b6ATKhpd4hum7q36atEQ0wUUpTWZS2HQnD/bRofPX7e/khixIarmAJBMnDJSJUkMK9NnZrfUKFWWYPxvE2kMoKYZgYNQpIOL581q+PAhMe7wSXAwwWBtIjcbfLb3+ym9IQlREQ9H8OMDeXNaCyKaREaCvtFCQM6XBgaGoAg2mU6uganfJ/zdS6fiB5GCFh9hwiL9Km5x6+TnTkEXHMv+b8twtHakJWwBoMQkCR2AeK1Y8GEk4gYoKnlTmUMkkplMYAais9i9/dw0v8htMHA24YeysQ4vIT14R/vfxtPyQcpsqmjV/PCfLE5wA/Bx4yGLHcXaUV+57UhfVOQEphTdOSmpbUVYNM2gLDiITxMWQY2PqG3ESQUQlzKcVVyDLGPhYS7qTaii+ev5MRR7sJZnIybMKAhPc+6+Y6tAnDWsBWIuEW2rtq/3xvOcY1gCMPz+Ec6CfcqKfeWdXDJkul34O1XZAwrtyquNK4+OWnSlgCJ4GH9wUTM3Eddt71XchZwOz+BoeCYVmzPZsZi/2mQ5xf03LqWo9TxgvkfDC1vKdooaZnHv3r+Rf88rIrr7pk5mWXXmp8Lpt5+ZAf6+RtLFwy87JLZl522SWXXXbxVb++6vaOELzKkpEYyG+HbtazmRH39j9onT9SGBYS2KGDZxe4+LDoxJFKjkhgV5KA9S7vXzCbucORcDyG+TAS9Zs+fp4hK7CcIAIGoAcbWF5cecUffPDfmTPPckv4x8DEiTPOb4zQhEaTcaoZxMxmrdMciQAJA9ih7b84dGIZB6Qq1uYADpl8QsZU7HlSUxz8mYd40cMdklRtbGv4asKEYowaFATgTOWI6OYq9v2qUekEAxNmKAV1TaE609iNQETwdo6n24QZd3QGmzBo9Mw+APabWMOZ03cPENDIJdPuAQ0R/Mdd+sd2ROfb8CIARSxpXESnigIxWirc3frgxeN2JCxZckxnzLLHCfcfWtYRo7Z6iIa+mVFNvLDG4XQQhwT+ANAwICXyVz3yxmeNmJTU6hsNPVlZHpWB+kuHhFtJVJXjNLLsksPHgmWMc4KbqkEE7CXilP+8V9eusxUQq172yFozyjHjyMp4MGMf7AQkjIsuZhtwlYDp1hkrOdBBU9/HkQdsR/R76R3N8q+aSDhKae2aT2bDbOEsGHXgqfVo4U0mk2gNVqkmwydjkwEFsgeTXcy28IHPAs4r3TSy8Ipja2yjHdegOAIphMRiQqrveeJ/JhI2UsOqmOyto3FZvpPkOWxk+DD9iEQMPPPp6mY0ScN40aJU7YWEwQriYsDGsDFrtvUAWgU0C8uIfmMJGorrXVG1iwI2Zkg4idNd0+ZvXy/hwdUij9HsWUiYLwjsdeaiDqgizI0MCdvGT/oQziQ4PUGVjo6ln42VxCAiYQFidDnicJO8isJx+zaHQaywGAVJkbGd7BtmRO32i08oMezAjIzbsAn/9cXPmc82DDwGpPvcnI1vXVOpAp0q68CZv+DTCbg8ykRLOJJfWAxLhMXTFnfQThVYADop7WDU+uADz1IpDzRzpLfe2ItT2kZpSIt260nDNIpiA4a27bBBM1kqd0qVEI0sPv8nVTCegSkj160/v/JWHQH6fh5t9XkkcNCaFmylbHtA+nTNDto1Z7TMz+8OuunIZUckMCKBEQmMSOAHJYFB3zU7GgmroM2A+3Ab7V57/MGTJVQyjS/UlnhCCoJ5Ik+mTK7x+RHGerwkf8+HXl3US2kyoYHOZcTjZuy3rJBw5dSTG2NZIuE41etvv+ZUUSCSB7Q/DoKWHcThP+y839WCWyRNwpUQBhvfLLoY4oTB0yxnJKyDZ3Wi4Y6LjykgxMcRlwhan8AByTYRJc+Eo9YnaEil3b2xrVXLwLEzIRsrCmAokqGqhjktxS67vZEw9FiIRhded/a0AoG4iMMBHwFQKyyEOElexYV3/aONYQZTd2XpRBkSHkAT1XQa12gckXCUdn37i2klHtDVRRuPrp8Ef7SglnbjZdNNV7aBZCtmpY1l7ICdhYQz3Dyn+pu4cLggYez9KNUatyx5u8hLiNNDPOOfmF9XF8KAfBi+CINZVtv+ojBQIHs8ceXAdg6mJUNPY7ry1fsuKLW5QMDY5IiT9xNSQrgJf5/7jQGP4IJgfVQBAsq94dbqUQ4vIW4ncbhgvYzjgBufCNJxV97dQGm7zGL7FapGqQJIGFagzG1gJKwiDI7owI7cFlWaElSOw5IbhChjlHKc0q5o67L9yqVSRMKQBgr56HCJqZB4p725qBfrnDRd6KHq7JM+imwSsR4ALUHjbQ0fv1UiuZwcZBKGRgHSlkjxmIN/elYorisJBeCqoqamQGNGNJCwuSIJYI95R2eLhGHNTolTDWzCMqUrFh4ahOeaLZ9CAyEsWyK+CX+e/UEYVwVaKTW9ZpJojI8OueZpE4KGP2mikTU/O+XEjz6ez/6EngYsVNl27tYVQcaIhIG+rZuGv7jkiEKwbW8FEAZZbo9PimM/SIJHrtyCDWOuM5lGhDU0tnPBGpQ786Yj9xqRwIgERiQwIoEflAQGfdfsYCSsaUBVBD6CHVRv/PS1h/y4Ag5KJi5Is1d62qq4VyKCx7Pb4ZsTtEdlZivU4OBNnbHfskHCoyunntwQzQIJg1rXqzYuHu0mkgvhGidwRIJ8KkLh3f9+M2IYH1IWENMgYmjblk34sAKwgqJLIepwA9iEUQFBJCxveetftxaaeXcNJAymGAdx1/zxsffaULEdABkOrKAxCyiIDuw8lCZ+e90dZ593tawAcQtDwhYepnqM0jrau2K/Yi8YmlIdk5ZFKQebMAy+Xhpf/b/n76n0gls8ukY7/n97VwIfRXX/38zsvblJCDdYxANvrUe1Wttq1WJtta3aeiAQIATwtlr929qq1Xq0HtVWq9Zab/A+0ap4U2/wAASCQDhzb/aa8/0/v9+bmcwmu8kmJEDCb13J29mZN+9935u3v+/7XX6bCftZaOjo7/zUkWKhF2K6utsMOformHACWUEbX/f+D8dHwPYeI5ODtgNCZ5Xuf/ysDWgwYIOcdf5knVNdHMyKdP8z4ay3hYNdNLXzVwOHCYvHCadDglub1A2fThxbJAUVxvwTJ03dgIo/oMFGEgJBW6jbdzHydjwDIGctxM2RdiaceH/lyzcPQbqGRMvmGtFgKWNDmH/3R99c7TBhuAjpKFgvcyv+sx8eEGUsEmTRoig+7IIJ+1l019drwfPS3r3SOPgkG2AjbRNg94/HLgPUwtAZzeJJCwKEb+F8TUvbF7scfsiNjzyJTgSotYOnpI1r6y885bCxqLNlSsBWmfp8TCphbJdLbpyXEluIotPOk9XpgfKC5ZTNONfWfT7vzkIf5CwSUeiAiMLH6C0PPA1VahY3dchWLHTO7StiByYMikdkfaEeMWEVTdkBibplZx68J25TuDncxcIU2f3YU2oNiIfXnIbxwTnjula4s6Hrgth2a+bq8r/dcCFj8oHHnynG2kSrk64vzvdbwBWZMKxrbTz+8eRjhtszrX2NzbPUFzSYycCm7WxzBazku19twpWkQyx6ZzpkPEN0kBAgBAgBQoAQGEQI9IYJS3KESaNH7XlcbbPahrIASETo5IZue94sSiBy6imeNlpFjJlDd/EVQTDSgMiwC0KSBIFPwWUYjFp9zBdhoaEXXvdAq4hRnCFrZAU+LyY8er+frY13y4SFZV3zLZdVlTMWgCSjYtM+ANGs5fInP/l6IzLhlCVCFbl8TcjtoBMGATHZwpsWu0zYdlsFnUaHLEquUGoCBTUaVix8aCj6CQuiqEDoWgZJWsKj9zjiF6tbwcrXDdKVAUyuD64BLUiJcW5teeDe2xkrG7PnD9P9zoRRAAcpOcW1dS2r3t61zCf2FmRQC4NOWJYijBUrww54f2Vb2iOpO3wFZGxHT9yhh5YJOuEEGj63plYuPGQYKw8AEwZxEsyjC5lUeclND28RKjw9lZMyZp1TXRzs0BDxkZhwVlh6e9DhVkIBm+BWI0/UHbHvqACwgCjzD7vrufdinIOruwlWEu6DZN/QO3wZbXBmVgcmzJesff329n0fkQNcZoXhcsZKWXDPFz+tR3YEPhGtsUbBhGEXxopfdcFZpTILuvREsAvYjImecdGfN6O2UVVty110FXaMorMyYcuCxDJmm8VjqBDewPm6Y76/OysoO/mS6yCTmniiRYdT6+ddVz1KrFJKUHAbeLBYmJXsceiJVW24DQTxxfB8p/MZiGR/LowYN2vfu/ePYViUQ0yJgFGMiJVQOPKrTUn7qexzJmw31dRAgY6cXzd5bO2ck78fATrtQwKHkbJ9kDacRUoefndZA+dxiO0NjiZ2CLkOXcz5Ufx2JDhv/PiV+4YWS2zE3jc++T5mrHMNU3Je3MMv3C2DFE9++asf7QIrlbv5kicLtq9xZ9tWFEQYCqggwEoPWL4Jgy64sRm8TxCVCQFCgBAgBAiBwYtAfkwYzJvj3Fi14I6Ly3xMVgoYG1Y8+vDaZohkBZKobtvtgfObAWFpMLCtBo5kBuQVToCHWxNXl953XXUJ0BTI9+H+jAsabEel8pVLQ/dZvhGCM6ugwgJPLUfSzToOBobDiXHecPpRe46QwYhPECIMvuwDEstGjdr3p9/EgEnmYFbYDaDrjVZ85aRD9yxsl1JkbCww4Y8a0qvRDC9pahhwSgOLRfATdimxaeuE2778YQXDvD4o4whCLUd2/9mltUJLCV1y/odCCsT9tuXfHgrng/5YwIMYyeGhzD9yzh//1sLBRhqi5mImVVe0dfDpJJtZPNWImxVqjOuNG7/+YNyICla8+y0PLUw6kYa83sKoE16XqRO29wL8LHs+YW8bPGXH1ReEUxP3NOLnnHDYcFup45fBwFFC69MI84879w//SCFZENbbnhHv1CPPAXAiBL1ZwztP3FGGNqKOQSZjwbHh8gMXr08125b1mEosq2oj65zq4qCnAe3Fwc6E8xyRdkC2uuRo0FKg/TXjPF0/9eSjwAA+FGasMDT+4GXNYGsiXs7fbu/qjiuciS6/BmxtaR9+/cKNsOkkHCdF3lrBUqRhkXHf/6IRNtFcGoqcBnR8nCfeeu5fmMi6M5UJsaIxj7+/zE4EnrIdPHI+qnaLdNNKaGaTBda+DZw3/fPm80M+xir3XLAiDout45SbTia40awtfe7QoULjageSQv1rAHeChny+CfSlSWyo59nsAiUHH7OFxz75/WkHh+GJkoEJQ4LlAGOho0+eui4Oqzm8DAzD7RJQewPDjh2NsQDstVhYR9/42NtrcGtAF+4Ybngk57b2Awq6fvATBqwsk2vrX3zo5nLwrXBeigR7lGAfEBh56KRVca5qXPj32kw4rwmhq8kGzpNtjV9vWfvJgbtCYunIfscu1mFNQWMTwQu7gKuHX9m7GDpPfVM9+QejxpTstteE0WNHjR47Jv/32LG79MV77Nhxo8fsMnLk+MrKXccO2/N7G1rBDR1mdIexoI+EACFACBAChMCgRiAvJgwiZ6qF8zUL7rg4zJAJ+3bZZd8T18Ygzy9IRBpXLXgLD7gMJqzDTywqM+PcrONbFu9bzMowWWc2JhxmoXFzrn2wvo0n0yLWDaTSQPEx1zhkY8LOXrsf1Ah5MWGTq9xs5Nb6RN1HE4cXtyuIUOCCAC1S6bOL19ahW1rcSqNtZBYmbEBQmzjf+MHhhciEbXkaZTiHCQPf8xJ8gFDlZjNvW/7nWT8LMVYRCaL2A/MNyawggrqpom898BrEsI2l0ibn8TT4sLnvroQyI8VTzen1tfuNG1FSUj7m4J98soVDZuIOZpkgBmE+4bavDqoodKyjOzLhzUJX75gAuA3wFnBrAOMbC+EYGhd/e96dxbYo64cwz0wKAEKFTB572I+mpDnfnEwnOU91yRVEN3VuxHlCg07EOK+/46pZo4MQN0i0NVg+jEljT5p8tdCJIUjC1zzbFMp2rCtxMCvQxISzwtL7gyZSERV3jVKGkeJG2+9nnyp8K0DtKZedfck17qDmfR93sOGKdiaceH/FSzcVgbsvPvcOE/ZLBYyVfOek84WNvcsoRewjwYStxmX7jQkVw6ZX55dcNvHI5a2Q8yqdgiXDcdPv3F7x9AgVZZvFm0y+SeNr33/viWIfxIE7+6Kb1hn2MysyrxrpFDdaefNH19X8yDEahlDXMvBVDEkoFf7fbc/WabB3JjrrPqGdb+8ccfCxWvjmN089oABqltGAg4WYXMR8Q2/+9wsb0fUYLnGZsFs1btH9fvqJwyDamHjBQ+llwnHMyGz/Rggy7NxWNAPXDOGxIRq+6evPXxhTwMplJ9OVJNvqYUlhpbtN+80tSVw07MUmtyWJ002RXitlaPWwsanXnXj8oZC1Llj6uwdfWodrCno/57ESeWrsviiYMLD0uGltxrltpLVUWlPzf2ua0RdvTdPTqhGP85Y2+LnCuBfOD21Xq587UlQgBAgBQoAQIAQGBQJ5MWE9aYBO2Kp95JqppaB7CDBWJJfss6bViOHvv2DCEEUKk2Y62U00bhqGiRFGQZmBElm87tY5p+yhAAfKxoSjbNhhi9Yi9QUXY44hWUAq6nKzGuihqxNG3TIIYajMzZcJpyyogadWbPj4OUeGA+tkGbxlMZ61FL3p8bdAiQBaKuGhiqpgEUHaVgubRkrnZpp/9dK+PlRiZDDh0O4nX1IrLCW9TBhkI51bzdxcu2XRU2Ns/usDnGXICBRVhKBdxir3WrJ6Q9q0VFR6uPJnR5EN3CaFWsRCz9CUvnLJKQce4GPMHx1z8/yFLdiLLEzYTHO+judmwp83GUKC8966cxmZcCtsGojvQLeT4vqaY/evBLYK/sIgH0tgJF/OpHEsOP7F9xa3ofNni+r0Rjxg2QRMi/NGtQW71qjWf3Hst8dWItr2BkhhBas84pUvVEAJJG47rHf2KdTTxzhbe3g2JqyAXbsMOxr+8tCBp36ud0GEslaa38EetR91iq7erHPsaGGACjAGAyw0bO+TzhEJrtxhzK9NW3+WlwmradgN0x+67YoyxsoK/EowwqQQKxn14BuL1yNl6fJ+3mnkggVXtDPh1P8e+ePpJYwF/QGwa4bnBOhk2F/MWPmN/35bZOtxCbBbgAlrrP/DuaeVZmXC4OhR+YOTp29Af1+4o60Y7tBega4wx03gVluDytcu+vSpyhFBHysYMeyQL1cCiXbiX6sWT4IdjhnjfH3togcnoM8p84X84UoZwu3hjpAU/NYR5yyNwSpa3wrpotznNPP22fCxWta+fd+3fKxEwfxMsJoqzDd8r+/8Ym2K17u9EOuMC6rtouvVCYu9KWTCUlDohOOIg+0RjSbibnvcFpp2aiyOQaQ3m3z52ZP2LEWKj/RaZlIQh0phcjmTRt3z5McJXHRyddK9BRYcq3ve0Nr01aknHxqKMHnI2OKJRy1NgWd2HKAWGunM67bik901THymw2+IbsF6IFbprah3qy41VJ5MwbTCJVI43nT9S+uONRUIAUKAECAECIFBgUB2JmyCnAUxlYBHCMNSM8VTX9w0+zg0UZPAFjW068oWSF8hdMJpC/aVVcsEMgwCtwVmw2ArDYbToDwQP/pGa+zLF/YuZAUo1gjqIvyEIbuSXHzsmb+t0zDgKlQCZnIYw7Pr32edY5DV0763eyWmvRSioKCSCtxqxJh9J62OQ3oNh2Z1lCDa9DjQ6cSXK1//DxrZAuNHxSyGeYIPwYv+Mj+GXEIFUJwXZDTV8a1yU7WSsMPe+tHT43DDAK2AhW6EMclmwiB6iEsgEbGwrFbBBJS38uYVZx1zQMi+dYhl3a1PAAAgAElEQVTJPknxhUDJIzOlnCkVhZVjX3vznZQOMpRHcHQaA38tSEdqv9Pcamxe/clJ+++G+tjyH/zi/LVooAiG4ll0wsiE47ZOGHVNkEwG0huzsZJvwuJmQ3jeCiLhbUBmGQKPwRuPxpKGBeFgN7zy2O0OFgz9/JgilUnySMYqfjnzojjnTYZj8uo+Xd6eecqYJpXzVO03nzxfjnlG0Rgeq48OP7HmL5vRQh8kajBfd9WHbr1OwfmbrybE0wa3y9xs5bEPp323IgIhfEX4aqDBGM1XZr7ywv1P/VIDy4i+f/Wo/ThncIMJCBIwGqOZG0ueubW6CPddkLojhsECFhyz909mxPE0cXKuZ6fvOwU8xHXU1NPg6qq/Pf+2YehFLyvMH44wFomOP+TTjWC2rNmLS4eGeKERX3U8ogMGTZwv/fVhpQW4d+ZDcwUM3a4wVhQo3GPxRjtDj0uA3QKq9WIbV7wz3LF2sKe3rEiQMzuA7S3d+5hfrYx7l4wO7YR5hP0V2eAaOY+/8/4zQ0f5pSALRcbc9Y+XOefNrUhXYCqrqMPTuA6e/1yrnf3rQ0OM+f3+oqLh7dRV9jNlwq3zvwTmiYPYPl0z7i8wcfERJjZN/7y2upSxYQVCnYvdCu1yz5Mfm5w3iYhSsNLgjpsLqocJDxVeKrh2iB1PxsLXP/5+LTYdHkrPS9xbNM+ZaSaSYWDCKvzIfPPwrTXjw7AO4gvCzgfh+ZKDwXL4MSrY6/HXlsOinKOTGT2GocAdB23LGb86Dpb2iMIClQ+9tmQLrhQpZ72ACd9HXNVuF6YET8KGI0+orbqdkb691W7zcxUyO7I1n2CTNM2TqhOVUhcbHO5oUoEQIAQIAUKAENgJEMjJhFOQt8OAhCGcazELipveuuzX+0dRnednISUw5oE31q0QOStTKWHIiMTOTrMLEVBNMI02xFEoC2/Q9RdWTQoxVghxZuDl90N0Usb84eKx7y1eFXMCwYo8FoadOCT7aBhp9CbmTTxVe/D4cAnQNhnDqfp9EEFYQqXuUKlswpeJRL3QG+SUH0ye/mrVuw9WRFAlJBoHtYl2Bsp3P3H5ZvR7FTTU1bMYJm9pgVrbtgCd1jad/+tjSsFOUQFRWAoxSS4sAkXo0KOrPtdxEx74apInG8AkG/YQhL8xNxvWvv/8rWjgCAFhgtEhihIKMDkIfQnIoGaWgtHSu+55oKkV1EQCFA39tCGKVDyN2mCDt24BMb1pzer3Xjx64oggY+FIqTz2u6+sArlSBMB1A0e7lJhbac43cG3FoSPKCjGoFTB5iDIaZmyIL7TnV3FgBSAJo3JDpAHOBickbcK3/aVp6pYW41rjtFO+X4BKdtxukBgLQ5IqFmXFY1/8GGTZVErsV7jDna16mJa6pm3i+td3XH4Wg+sVDOnjZ9LQkglHL2vkjWLuQiIWFcRpCHLm1ukpeIpZv28/2C6v2iXRQ9zXSPBv/nvMSIwWDtMG5oyCTAjNbUtGHzL18xbQNTmvviOV+bdfnInDh5bryM2MGE98cH3V4SUYGUmQeNiTClSywL57n3hRiz077S4DGu606yOS4GDi/SuYobNwAJNLtXz27HgZEPaJB9IXYtHyw48/eWXdJqTK3NCNVCqFG3Giqy7l74yRgVYRMHRtzeu+efvhAwuhZnhMMFYy5jyPstDE8//0aAxok6l5SJbLUgww/1A5T1513nHljJXbEdtApa74imVZhEQoYb5hu048/H//+7ItZttHG7ph2MwDeq1bekKLW1Yb0PJk3S3XX1lYEPKHC0JFFT8/fUob3sO1bcCWiK6puNvYZK57/aBRrDzEQhK439sTEFKVjRxz6Kmr2yCRmHfjzAs0LBemveCq4IlscWtzuv6Lo/afUIS8E7MUwzpWdent63E3p73hGUzYVndyzfrj7J8V4xgVREtcwx8mFf/hkQ9W4nQSkcOEAZGOec3drTsvExY4o1F5M08uO2K05Mcw3WJhxsTNmPNZjoDltlx0xx33pYSddObMxJz20GlD02A14AZP169a9Oph++4WiYYYBPJTps29VMBigc0O7FCKGJD2458BWW8/IELO5LH/Orsq7YuLc0L2I+5anWehy7baJujidwx2vTs/KHSEECAECAFCgBAY1AjkZMIqT6e4lhS0J82BCetLj9tLjqDlYATE0fKLbn11g8MtU2D/assQmAnWMpAJg0whQkqBs1xrWt3I+ZZFb80rldgQP4sG7AyzwB18w46adE4CxQUkUiJpp5tCM9c4YHBq3shTq741RDiSQaoPxgQTBl0q85dKQ8Z+2BTbICyTs0kHtuxk1jYvf7kCw1UJecuhwRCSmPl3/f2f5sHVLgcW0gMcMiGhC0+m65ZUn/bDILKgABhZFgAZlpH3RQpKDjltqQb78DzWwlPN3EoIHYmDIueqyo3as0/aD+8ekuWAAsE95RADcQ8zpYh2+Sfuc/CTzyxAMRrSfYCSIY5KR13n8VZuJFtWfDr1hCNGSeBfh3sCZVc98EY9akPA2NQxQMyMmAVM2Gj9cq8h0TDcR4KWw+VBxoaECiauTPPNeTHhbBBDW1ONy/+3eyn49Ga+/FLpuGNOnR4zhH5PKLEEd81alWnxuJVcHV++YI8CVgjCvxTwlTB5KFPGPv76yphopKhG00H87UsmDHC7TbSZ8KZ3j64AKoUUGLZOBBPGTZSoVHH0l4083nf8tx2UXI9F7uMYRhxoOdKyODcXXz/jSNQJK3Y2Zth1Gc58++876ZJmnDAuD+woK7e3o49L6IWrg92CrWlM8BX/PWQIPtjCnASU7j4mhyb94vSYZiUsnjQs1QJdoiPRiyaJTZmOcKBVtQiD1VLz0yNHYWA9hbFiyednsiwXhSsmDt3350vQox6z+9r+DF6iotkYqg0rnz1kJHBQDNcXxEWjWIIIBcW4AhQzFi6MVp5+2uSVK74BysGBt6fTaQi2b79Uzlvf/+/8g3cfYdtiyOXRoXt+tqyWc94Yj4vcuuKOGMoJzRyA4KZ58qvn77u80gcuJ3a0e4jzjMugVH7fC+80OOuLaLxzR/wrPHXbDyU533Tvny8t8fvQvN/nhwDv/nF7HL4xDWYediRqAaeXJgKjFk+Fdm3NqUOcx9vdR2RS4TXzFwuDFOi9pefJhAEt8K3YuOBf15fLrCSACxLOAUg1xyCFM2Rxln1+X/gHP/jRZx9/ZqjgTNvaKqJYQN9cMswtTW+su2LGGWI/LlI4hEVGjvv2j75auU5VbUsfoOUYBwFMDeyetgO0NaVO+MN0xbcNnnd2ZS3nSYC7P03o5NEsyM77IMKYuV2mAiFACBAChAAhsBMgkJ0Jo8CT1jlEMHK0GIl17z48Dv17/aA+8jFWtv9xc5clIN0HiBr4Frvp7UzYskBBjEzY1CwtHTOMBotv4ryx5qTDyhgrCYGVW1iEfCnf57Z570I2DGCasF3tSAle6bbDmKBYYqQ537R60ZOFdpINDBiD0qhfEFl/gEWG3Pns/7Y4EmEHacbCxuPBet7y+aR9KgocSY5JqOhBRQsDt8+yW+6YL1L+oJCkc7BwM3hbfbzu65uvPH9ksRJFzVKxD2g5Rjx1mDDoLvd4+M017t1N3Vi/rnn+fLB+bH8lm2o/fGN8aVGAsaGFUfRRljFEDMRuRcWV3x8uAlV6uKiobNiPf/Lz2/9+z7//88h77y168fnnn3r00T9c9ptvVZaH8GwQ+ECjs+v3zrxqVb3AlifVJMQ2Eygb7WbSGDFrQ2L9B6OiEirXZSAbEKVVUVhUCY9+aUnz5gwMXTm+vfldlgytZdOHrz9d4YcdAjDptHFW/OECpkTvmgdpkwVT6FII1Tlv5Ob6M47eVygJpWAEEt6wkXc/8d4GFWwfoWW26h7yuCITzqYW7jChuvoorMmFNbst0zp/EhvefnSvMIy4Pwh7CCKHqpjXMAuUvR9/q9mOuNslQD3+sqsGCxQcIMWZ+FiJpxX9b2O87aPpx++KMDpMWFZYoIL5dt/vxLkt248JA6fF/QYDoqjpvOHjn+4VQccBmDSyD1SWSrSUMfbdE35WlwTny6RjJi3sXx1z9I4YmTB7eJLrPLVy3o3nlaERcxgmObwKwsVwoHDC4y8ugTrBHjju3ftyKYrjMKtbqRUP3vab4TgFFT9EFWQsIklFsgxxuIqKhCNvUAaHf3nE8DFnn33Offfd/9p/33j55Veefvrp+Y8/eNapx4yA6MXg5lsckP1KlCmjb3ngdTHb4rouAhmLj1gW2lNYLNOJTZzHrrv4rFLxsGMv7H8Uf3D0fh+vS4jER/Z09U4yoOL2Io2HjTVvv7hPATyXCvMFwMBbKhx54Ce1SRtMUYU9l+w/cCEcx6TlPHbRKZMqxMIL/4LnNbzkYNU18zZw+L3QzJRu6aplirdjSQS7pg62GfwQl5w2nt447fjDKhmL4jihQ7ekMKDB4o3mSnJRUclBBx38z7vvXbp0eWtrm/DNTiXVpUuXv/T0kycddUABnie2LQLRIaxy/wVfNumcNzQ0QT8866HYs/Ci1edlp7/bnAl3co0R5Bl/1DJXjI5PD31LCBAChAAhQAgMHgS6YMKawSHpLDqwcZ5uuGr2LyDLLkhILACWuiWsaJ8XPtrcbPL6pApRf4E2gzQjmDCEyrIw0w36DYNYa0BKEcNo4OamtW89/C0JdGg+JHmQwWK3730RA0PhzjTYo+fp8LMsJBOTJ1f/69pznTQbwrUVpC/bkBIFsZNn/qktg8W1SzUWymeoEUjy5Oo7LjgFkxkxFrZddIXJocJYSSDK2NATf3Xu50u/RooPVuEvPzvvpOO/X+CH2zHGiqOB6y4747Rj90Qm7EMxGzmRT2L+4Yf+eEpjE0iVrW3xBa8vGjv+ez85ZU4cFIs8ndCNNhM2Dgzz3WfnDQsCVxSJlyGtsC+K5NEHttJg0WerW6RAOBQuDIXsoD1hX6AwFEY7SdFqxpTgsIN/8XkLWjUDyzQSsQbLVDG6GShmLMsQb/BVtuqaVr5VLDgc6NygT7bVp1R05e0vNKEWux27npXE8KmP/v3GAkym5Q5Z0O9T/EFWOPzF/y11CIzgsllvoHJe/9BtV44IsnJhYx8oYHLF7N/d2cR5s8lbVNjBQW0SyvkmWDvCzOos6HWYUF19FFKraI+JMx0nu5XiRvMjN1w8lLGikF/2hVyLUEjMAzNCYsFdLr7+UTcWWNYu9fJgVw3G7mYYM2OzIUUNzjKwyW9oW/XqQWOR8aBvKzjKgnV0KQuMPfDEaS4TFg+xvbng3rSXje7+MrELJpiwBgpRg6eXTz5qJOAptqV8PiUYCRaWweMgRyYc/IM3Pl0J1s7ofZBGP3UnLoDbXLtgcr5ZtxI89cL91w5nrDzsKykoZywYCpYAZQsVMjb0ur8/o1mwAmo44QURzcZb9AQk40lzbfNlUyaNs7fQxBacyDzkB5MKKSiLdNrwQEuy7EMDbFDdKooSDLCo39kQlJk/GGWsYPbv7m7hPOawbYyqBwbt7k6MO5nbDNxubFp+2ZRJUTvoPKwO8ApKrHjUqL2ObHYMEnCbyYu/0JnDjNDTLXrDhlO/c3AAFNmyzAKyXBQoHvfogsUtuOUFTvwCAgGk/TRhQ+C4xXkLtxp+cehBYGeuKLIsyxB0EF+Kss9xc0QU7uZ4zKXBEFrCeZaQhIo7dGDCZrytgSfqN36y8MBKf4UPpqvYbHJ1wognk3G9CoejsuwL+EP4UaSHhjZgXnhYVAsDPr8PDXeUkic+XA+e2cIvmfO2toRXoepFqj/K2WaUjXHWr7xt649y5wWSjhAChAAhQAgQAoMYgVxMGH6Mk3pKjW/hSVDhLnj8zpIgw6waIFL4QJlXyFjpmH1/tHSz1uDEt0lzM20aLSnBZVB6ElakQk6C7Ltp02zlRgOPfXX6d0ZjSCHBWAtumf/2WguMNg10LsZY0+3CAOTvhZeZjMf0VBKVB9zUhcY63fTFwn2GgKADYrIQlMGmFz6DdiYK9sVK+cELFm0WBnBtbW1ePz0DY0/b2m/e2vzZUxNQGwLXCwUzVuXYjpYxX6UkOwo/xoLhgOIHLYocLmNFu4PRsvn5byfvDZeDsWy0MFAKWTpAJCxgrCgSLRtaMYIFi5mvkrGyn5z+W9EqUF+IaDgQfKz533+7Zjga/qL2iwXDEdCvKKigdVMwKxgiTBbSHlAwTF8ifOgg4SZjpeMP/NHaNKhJbYYJ4cv0dLxJ1VuBfmstFk9aPGlaCcjkpK6+5coqiTFf0IdOwkHQt9jCrP/QH9esxMg9tnyOgjHnPKmioWkeoiKkfoK7Nr388D/GRsF+tABjZYf9cjSMxsWFFfNf/wgCT8eS3voM3bAsS1XFKJm3/u5iYYEJ7WQBFh5+/rW3oZc29DTJeRqC01jtaZ8zxEsPNfIUu3/OOW9s8/Y0ydV6nqzjG784ZGwRpHGSbPJrAybmH0yD4OjxByxZ1RLTeTIN4LmeoiYGqfX2tGflbtsvqoOYUybXMbA59BP4MTimausf+sflZaD5E6GiYJJKkgRbLYFyefhui1Z+U69BvADbctRpnAUWIzAWKfthd77oo7+4+WVCS01uM2Gr9obZk0KoCWZMCUSjxeUVvgDspUiKr3DIUCZHpl/yx5WtfJPJW3G2A301TV1NqRq87aaZXFdTDS3ra86dUoxuHgrowtGRwQ8aZlZQMufK22CYca6KEFJZdcIAoqmCdz3YV8R5YtWZJx2MD73IewxFgazn3/apIbkvmSlBcH5wzi+q+f0/GhBxEepNbOG4q6FjZ8PFxoaI8MS1TTy58qJzTkC/BibJEebDZUsKMBY88KDvfPXlSuGoDCEN8ZXW9DjsTlpca+bxLem1K08+8vDyoE8otRmL+sr2fOWDtSnXKNptARbizbCAuC8tHudtdesWvVgZAtIfkcRGJDB/mF2KJFUc/Npn8RYLNheEuYZ7resn7B7pVECFs1G/eOHj40OQh0+8FAmmK8zYji93P6q9IDNWHPCJz7CBUrbrPS9/Bu46GNc9K7Hs1Iw+PpCJaMYilfWrrI3sw4Pdr4HdLjh0AiFACBAChAAhMHAQyM6E01oqprWhDi3FW+r+/Jsp5WjGqgSE/aDkY0GZRSWlkjG5dNdDX/jf0iY7EJQtZcfTKc1UNV3IiAaoC4QxLpJhbrbwlmVv/OfGUaj2ZCxYMHKfzzYmG1F+zcqEUQDx6DN0Hfgw3u2VZx/btzI0QrI1qCARgeAlmKcjH4GZ4siiEd9++LFnhCyTTtuURtM0kcdTGGYDGTXX/uU3vywWCj0PE3bqCiGhdSVaUE4ES4czqXC3Qyd9uEYQzhXvPHJZu4m1IOVwPZiWh0PF4PYLYndEGX74kwtXtho8ZYCxqnCrRsV6I7caXnjqX2HHgDjgh6hQ0C90j7Mpuuim3RaRAQbFegVvLpWedPZl37QCOQRyjSOK0paua62cp01w/ROmsgkoG/X333LJsCAD+RKUSz6FBYHLA99mTAkzNmzyBTeuaQbExEgbOldBxShYQ/diosj/BOepzUvefGm3IcB+i7GPwMUExCWjT5524doGT4QpT8VffPHFicceK/wv0SjarxSPfuSFt8CKFU15YTsGonibmMraAH91kL29gqXnAfUUu5cCIUwx8Nh0vIWbKQ6O582fvvifQ3cpLrLVv2IIcNND9MXRCzOmDP/WAQ89/WYibaXTaUGA3YKnfz0sdtt+UR9EijW4puppkX0Kj5r1L79w7+gK1EaKSYXjDA0HA+kICxcdfcrPl27YJHLSdmiZqQlG3eFw33z0MmGs0eKpZU/99YIATBWFgdYUoo8HAqAkDgbch9HPikd+96dnvPXZcpiTaO/ATdvkQZD2tatq//TH35WWKkiiBf9XmFLG/KOZVM4iRb+/4y4RYkr0RMSwzsWEIZaWqWF4OI1bG1ItS86de4Y/gF66OAFknyI5alExI9x/XSIMz1cYzSP8QXnI6Cv//uS6NG/DPR/XHd3bALFepfChFup9XBnj3NjIW5c9cOvv7DXDLzM/IBMNRWU5EAmX3n/fw2qaJ9ocKmz3sJUnNn/95gujMPe6vW5KhaMmHPLOl43r445vsPcBwrKehCXUUuHxB66upjd/+MZxuw8FVbizIYdh3dGmBELIDR2z76SFn61uMWCjyrOaQ1mQ4eyzBya5CBmW4Eb956/MP2hYqALNLdCEISsTFstJOw3GfskFRWBE4I+URoZN+O+y1jrEENqfy1o4e4P67GgnUDPWqc7f9iHpzVpV92tgtwsOnUAIEAKEACFACAwcBDKYsCCHW7ZseebZef9+8NbZs047YMKEcl+wWGJFfhYIMH8Y9/kxy67MwhF/OfNXMKWSlX6rdMz4U88666777pn3xPxlK1elTSORTqhaAuJlQTolxyfVLgAvMxtXHrPXWGRsJVMuvhElTsNhLxBzOpO9CMHJSicTr7704ryHH77q/3578knHF4YhXYlwEi1EJ1qU41E5KnxdgZVE8RSwxQ5GiwOBwAknnHDdddc99dRTzzxjE2MdRSJHOEtxa33VMXuMEzlJBP139cN4AwRCOO4WMVYil+459/f/XG+A/yFEkdYatnyxcM8AuK4WBhztMTjcMjsalxxlwUo2bO/5H9dtQbEqqacENmluJtE7O5kGq8sPFi3YfTfwIJQg2ja4LQMT9r6Vdlc8/CLAlALmL2MlY8+96tYNad6Iptcqml0LRojAGqg5S1o8ZljNL78y//QzThxWqbj22DKYYisQo8uHTBjcBiNMBn9SubTy/67+60eLalsbhG4eOuBA1yMZUd+y/NNZJx9dKcEICgVqNByUfQEWLmUFI449fcZzC15fsmRJXV3dkiVLbr755sMPPzwUAs5QEohKMKxDJh5z1qerNqA/JxAzHQhfGg37IVKWSAaGyXK8UqXnAfUU85ACVa41ffHGUy8+eOe//nr1lJ8fP7oUbKHBDN4m8WCP71B6UFfZI2UztTDzl5YPG/3zX/7q6muve/7FF5955pnNmzf3CLKOJ3fdfux0ojUGpAWZcKqtnnN1c92655+Zd8ghYyWJ+UXLhXkGthP+ATdKBbIKB/yVEyZcf+ttHyz+PKVbmmYZKrydfRAI+9SxSX3xuRMTNnjiq9o3H4CGBUpwR4kFg8EhxQVBbL9346ugIBKOwA7LmLGjpkyZPAVek6dMmXzooYdWVlbiRGORIh8MjT1wEuxthSeykYc9v3xdHectjv5YcDwD4yt7J5AoQ0dNy04NC/tBKXz6G++/58ZQBKLUhV0rGnHXzH89TNjPispYpNy3x3cf+2CNiOwNCKfQfAB5pnc1FPnTW9ACXBMjAU3BgNhaI29bt+Ttp/afMLQopPggoTUAVFhQKoPSWQ6HCy+84DcvvfjqurWbV39d+/Zzz9908bn7lwdHyGDkA0udFGQFw+defdeKet6YwFD5nXsujmgGpODRjGVfLrvltjsPmDhxpAwrHsQwwJUIZ5PYlVSg2lAlpt8bftLZs59b8HrtN9941w1BhrPOHcvQuZ6CDS5hq663LnvvhaN3Ky5HdxEBYya0Yga3W8oIGoz20UHmr/jRlMuXJvhag2/CtRF1+7CUdX5nbU8fHswFba7jnVvYt0fyWANxD7TrZYe+JQQIAUKAECAEBggCGUw4lUpZlnX33Xej4gUiLRX6gkVSRNAuWUJFB8hVisyCPhYOgq1vGM2kC1iw0BcCuzxJkX9+2uloNol2dxlMWEiUoB82dIjv++w9f8NMnCXvf7UlCYrGJAhW2ayjE4kY6OKSiXfffqs4EgkpCjQvrBSHQZ6NIsmMuEwYuIgdNws9kYtQPJOYz+cLBoQpXUEBcPDS0tLaWojOKiLUunROTa7lTV/O/sn+IYfWoNQsREW0IIV/RL7QosO/94v3ljQkOW/mfIswzU0neXP9v37329HiCmGqjXoS8CAEkTRUvMt+79Y2xiCJMJiemtyyMPdykqsJnlpXv1blKQ0SqyS09OYrL51ejGmCOjJh7KVjvezohOXooUf/5PM1zZvSvAXzPKeFVt72DgWX7aQOClSLx2bM+jWwCxQaITAX5jUqLfIFMDVRiPlRMexjgagSLoLtEJgcElLi0oh/KGPy4d89wsUtDxkRSJQICI4na7zlmy/ffGbSD48UYW/DfjBfBMk1UM58YAUZiUQCAXvUgkG0oGZyUah4l/EHPrjgi02ct6CttsH1tJVCbRMwYXy3M2GbWdkaF8/T6Sl2LwWaKo/VHTgqUoSbK4XuFolosK2v9zBhd/fE5skSiu7MF4iEIlFhUX///ffnAVruU7puv8tYOL//nnsVDMgkRltszxRGHQYhuAvMJBnxR+7gl1jAB29wv3QmMU6Xq6+6WlCwrbXuztEze0ahEa0BocZSXF+9/qNnRwwpZVIBU4quvPUf4tLmDes/e//d6VPOjKJWNSAxXBKAyOd6KYoiBTMCxA+pHH/h9Q/VWnypBptZuqPtBvxy0GC7hWLhAPNxwwTbC0y6pG1ZsuSdHx53RK4GiOMZTNgXrbrq5uWYPLfB4NDjpK25FQ7uHZhwm30ncPUWeX1xqDEHEEzyVh5f8+Bddw4pEGHawd0BbwpruSyDylaWA0WBaIXsK2SsgrGhGOmeyeyEU894e+n6uLAZNriexFxpWZmZZjz35NMib1MwGAY/FAZ2yyGHBuOsF0zYD5r76BDwPweXCGhVMBwKhkOK37/wrTfFUOZaRiC6n5YEMgwvMG/gPM7rl/3n5isgrRy+OkHtLGowb913YNzeh738UW0956t1e+FNxG3le1ZKKRrWf/9mxbWLg1kb2YcHu18Du15w6FtCgBAgBAgBQmBAIcC8v3wocfI3Xlv4m4svnjt75kXnzbno3AvPnTV3bnXN3FnVM6dPmzVr1qzq2bOq59bMPA/fc2fMmF1Vff6MORfNnH3utOrqmTWzzr3g/Fz45BgAACAASURBVH//5yHNCeKKHNP5ZfdCg3L05jWrLz3//LvveaBFGOzZdowiw4MjjaIYAqkg8bV65aqZVdNnTp82c/q06qopzntazTR4V1dNmz59+tSZ06fOnDl15syqGdOnT585s2rW9OmzqmbMnDa9avr06bNmzZoxrWrOrJqa6lmzZlZz04JAyhnCDvoGWnFuxZ97+uF99h4P0VYCvmBBKBS1kyBH/MHCSNGvf3Xme+9+AFHFwCsVBDTh14f1WTwRn3k2ZLsNR0HlqfikINK8cLjwmutv3tQUa4inhWmyuEoINDokYAYKBxmFwLI3jUzZ2LKh9i9/ubZyRJkkZFoZ3HjdNwqCMpP8Z02ZuXRVXUsCOtSmQowbFZkn6LVERRA8SlPNGH5K33P/HdVzplbPmVozd/rsudPmzpkyd86U2XOmz54za27N7Lk1c6vnzp5x7uzqOXNnzJ07/Vx4Tzv3vOlzL6iZfcHsmvNqauZcc801Bo5mBoTdfYBoahBiR7hAwuBu3rjpd/935bAKEa1MwiTK4MDqOgH6/X5ZlsvLy88+++yvl64QjsRtGuwjuLsnEPoLzaEhQq0n/HjGPLT5ME5L75zsvmyaidZL5s6cW3XO+bNmTpt8RtXkybNmVs+YVjV9+szp02dW4TSbUTV15rRzqqqqpk2vmjbjnGkzzqmaPnXatGlTp1RNq6oW76lTqmbNmnXmmWcuWrSoO6i6/L77NgsvT/7mf18/b3bN3Fkz8F09d1b13FlVc2qq5tTMqKmprp5dXT27pqZmTk3NHHg0amqqZ9fMOHf2jLkw7lWzZ1fNqplRXYOP/+xZ1bMfefgxEVnXzT3TZSt7/KX9PGLv0Bc3znlTfP0XBw0tkxg77FdzVuKWE7Ajsd0B2WKTr7+24OSfTqooLynEJ85lR7IsBwKBYDCoiPjQ9sYEbNuNHz/+5pv+urkx1oyEKIFOBOLuYtkSK6SzhHmnDpThJXxeYbkST5iwp9ZUre3d996onjW1pDTitzXvYHXufQUCgWHDRlzzp+vT6KyfsNCRAR9YvBPsjokn19uAjKVGNKDd2NjC4HCgPkX/Ef7Ga68ff/yPy4cMc9TC7vYMevBia0Iy22PC2CuvuLQ51pwyIeqYyNIOHcw9wUzNWvTuoprq2dOmVNVUz55bPefc6TNmT5laPWXKjCnTZkyZNn3q9BnTZlRVzaiaBtO+alr1lGmzJk+tmTK9pmr6rJqaOVVVM2bMqF67pq7r+QELo2FaYMGM7RH+0UaKm6mWpsa77757woQJAlW/3x8MBsVwh0IhP4ZvkCSlpKTs/PMuXL1mvYDOXXXbR7DrFuw83+Ye7i5mAn1FCBAChAAhQAgMUAQymLCWQrZpclPD7Ef482/p8NHUrHRSFXkaNc0QFpLwMaXpad3ENLDJtKbqGhi5oVNwu/Bgy4iOHGNLM1xPo+zJQcoxNC2dTIG443m114DngCTEeWtzi421EErt/XCPxOYhPXYNzh3bs3cCB3NeqPTBiFyuFOAVO0HAXbO69p6776qeUTX1nMkXnHfuDdf/6X/vvtPS1Ajfmbpt+y3a49QKMqRlGbqxYsWK22677bzz5k4+56yrrrrq1VdfrduwMZkGqFVVFWwNU0bZtnnoLwfG4fjWMBiZp7XciiWa//vGK1f98fczZs2w39XVDzzwwPr1G8X+QyJtJdKWZgjVqwmpO/Ht1Amus3B3QzjiWpqZwrwmqm6Jt32+m/BTN+18J6oFLrhp4XVsmbppishPvWDCLk7uoLh2tmvWrFm4cOHtt99+0UUXTZ06de7cueeff/7ll1+e1ZbYM19E0XYKtWC6iPxb3tHsVHbHPL+CoWmmjjpoDhNYbNDYYroJvAWldt0ydMuADRYnQDcApWkafoZxES/TNNva2rOeupj0oJBHs+1dJM3QEil7kgkYvNcCu3BuK47jJ3EixCvG4Lp2B5329xMNFpyy/TG30kh7E2bbpuP2mBBl7Pf/enqzzYTFrhE0Dp5EeAbhtWLFinvvvffCCy/cbY89ho8aGQ6LMFLM7/dXVA4dP2HXM8+e/PCjj8eTabFeiavEfTtMEfFVh4PuR6eRbgVuwW6JxdOGmVi7rnb+E49ffsVl02dMq6mpvuCCC37zm98sWLAgFotpmtHRdwRdSQTUIterezu34N6m6wKssUCNoTFrvlk/7/Gnrr3mzzOq5+C7ZkZ19T333fvO++8lUrAUCx94Z2w9FXvnSWbZ+0sBF1h2GGjxkyG+hV8Q0SP8KTFUw06XhCmLYEn3JP713DWzmHlfG3acsSKK3vr16z/66KP77rvvggsuOPvss6dMmTJ79uybbrrpmWee2bKlQdPsZzazUvrUCYFcONNxQoAQIAQIAUJgMCLgYcKm2G4H3uvIk87PJPYcNMauIOYUXEkI4plAcF9XmgahFPJUmuhS2AE70wI6gV8JKc3Uja6ZMDctVy3sNKuP/trmj17zQ6d73EwnE5ah6yoY5qUSdgyneAx8gcUrGY/lYsLOKVxwPMuyUqmUkNtM0xS8qDsmDHQO67EEj3X1nG7lopBGau09qJtc1Q3NMFxCm1kDWChjpBgx8J6B89bSqeyOpPuNZhhbw4TdekQhmcwIGS0OCr7tUuUUvjTHTCCzhvax82yJ5D7o9ie/QjKesHkvKCE1e052mOT2jAKlt/sCC08Mf+19RrwBzDN7kfenfJqNlaUSSXfvKe/a20/sQJDcVKtuof3UvijZt4NZCYpWy0xZEBFcAw1w2tqU4ltMSHEEWzpA03UNcp/jtpQTl1ug7bbFC7t7EMKe44OTSEGKq168vVW5ZdgCM1Qn3D0c1vQ0hA/El1gHYrGYdx+hMxneWiYMmur2ZdY0uKFbIlmu207wq3esYYQVj8eWx3NWrjmG1NdQDR1GIvPlXpJ5GBhsh5nU4YRcH90KOxRwwxEQ1rTWVjuWdTqddkOau+uJu4nTTzM2V8MH2PEO8NJHQoAQIAQIAUJgUCPgYcJuP/GnG0isU/DqhA3V8L6FTtgCi14rmda85EQok4HoujU7Bcsw9bQqODC3eLw1BueAJOsSh0xSLbQNwAqRV5ug8OvF29QNl3iL3qFkJhwBszPhlqZGQ0tjuiaQ4ExdBVYM+Y5AbLYMHb4SYqwtSNsVZ/3TQRy3QIkIekvUCYOYiml/XIWwCBsGqIFnr5qMJUDUcwmtZhhpTU1rqqprSMpA9aObsCXheXdkwuJyx8oQNMYYnCxfJty5X5qBKqfOX/TwiKqqnZmhyHflJE8CLToSni6k6R7SGWdOdp6l2Y9w3tbSmk6mtFTanrQ97KY43bKsRCLhium9qgMv6rb9omo0goBHRtUsFQh8Pm9TN0C3rVtpTU+mtTRONVXVccIZglEIgtH79ue4EgfYHUqdWxAA3NTR3N8ZfDfQunBuaG5uxKcSDDGEBh7+NQxV18TWSRpfqq55N25A1418uBFzArm3zLOQtfmwz2VTX9FW++HycmN4kCECGegq4fHuEFIQU+sK0+jO1tEOAFlvnnEQDBMcE4aMLzwfXKf9nNXmmGPtwcPdxSOHThhGCt8mRFzrqBP2tCV3MUcbEEYjmUy6Di7eNdbQwTADrW/0znuFuW+2E3+TC2c6TggQAoQAIUAIDEYEcjLhjrKA2/kcEmKWHXf7zOw6YeC03jdqMDKOdNCziQY4TLjjmd6quil7RE6bu+boUv7fOsQ4f32Hy/hdHS8WxC2BG3d4C/pqQCRX+ys8Yrn6XiHOZvsXbJjF262kQ8GtEwv5dxvOFC+34BzYjn97OJruxO5FoYe99MrosAnivNzJ4D0h37p72uxuno7Mp9IELSKoEy149xPpzdFT4W0LCl+x6yQeLtF8cOaFnTbkV2iRLgzRwUvcTjxu1+omqoUC9kLMD/emwjpaPDg9nDrt89+tTRSEWwQOa8f2dDjT/ehdP71l94TtWcg1x7K2SZycDUqXD6OtkGdUc9Wfz/GsbaCDvUYgH8zpHEKAECAECAFCYLAgkI0Jd+5brp9V75mdz/F+27mcj0Te+SpxJJ9rc57jYcJeRUzPCGA2Qc+uoTMQWY645CcbC81Jg3dAJtwFEFm/yoJFHx/KetvcB3PNsXyO52q591rnHEF9xScx+g4RzjCEcE7P+6/3XvmUcz4XHTmw2GwSrEzAl3eb+uREFZMSQVDm9g0mC/eBQAMJMY9AZYqPBFpWgAv21jBhwZO9bDn3pGlfLHJ11XnAiQm3M15iwrlmyw50PJ81hM4hBAgBQoAQIAQGCwL5MeH+6G0+Enmu++Zzbc5zdggm7Io+mUy4IwcW33pVuO75nXTC7bpfVwncoeCtx1t268RCu5SfDxPo6Tlux/ut0MMW5Zpj+RzP1QfvtZ5zXJXvgGPCnk5sm6KJXLcN/3X9NDDQOIZZN7iuQbJdDFGGhsRbqRN2aXAPZ09ONIgJd15HiAnnnC47zhfetYvKhAAhQAgQAoTAYEeAmLBDjDsLbr080jOhJpOFEhPuGXrZzu4hl9maJzzb7eGYt07POcSEPWB0W/TqhPFkEWwJQ2Rp3FQh564GkGK+XWLC3QLa+xO889lbzlqjOCHbU0hMOCtgO9ZB7/hSmRAgBAgBQoAQGOwIbD8mvL2Q7SW/zSbZZa+qp4JN90DkYstevW6GPyRkFHXfGfG3Mi+xv8qsP3uv8u9/12f2FJ2eny/w7LoVnm+7hz+T2XrP72HjXCYMZDnHq4dV5m6bt53eck5bia6so7eT52o293McuvZ/IH4avBFOO4GWF0PPgwARvty39xzhKtxTu2hRlbceKtsIeOebt+yi36HgPaenZQK9bxHoKf50PiFACBAChAAhMJARICbcQSjb+o89FUy6nz6ZTLVdb5xJa13q26FATLjLMe0e/txss4dDPTiYsGDFPex6P52OVBlHkJhwP0Hcm2pzPVO5HsRc5+dzvDfto2tyI5AP5nQOIUAIEAKEACEwWBAgJpxLOuv18dxCRvZvup9K+THh7hlvJnNuPz+z/na9Wa8h6OLC7Bj05VGBZxdNyPyqe/j7jAl7e5lDJeymo/Ge22W5p+3fap3wDsaEbb2xiJXVdcQs78B7MfUe72nZWw+VbQRyzclc4OY6P5/jBHrfIpAP5nQOIUAIEAKEACEwWBAgJpxLOuv18Z4KJt1PpUymmksn3M5sczHeXMcz6ycmnJv6dhirng6153xiwh4wel1sf0iJCfcaxL6/sMNj4n5sHy6PnbqZ9+Pm1uMt9H3rd+4avdhSmRAgBAgBQoAQGOwIEBPOJZ31+nh/CFLZp2Emg21nyFt3nJhw3qJ5fwx1T+vMPjVyd2EQ6oQ7P6rtIHb+ThxpPyOzlOv8XMczr6ZPXSLQ07maz/ld3pC+7DEC+WBO5xAChAAhQAgQAoMFAWLCuUTcXh/vseyRxwXZp9vWMd5czJmYcG4a2WEc8hi5fj+lQ5O6/UhMGPWRucalp499rnroeBYEup2cvTghy23o0FYg0IshoEsIAUKAECAECIEBiwAx4Z6Kvt2e31MppPdzh5hwNqwFnt0Ok3NC7+HPdvNtf6yn7ScmTEx4289ScceeztV8zt9efRms980HczqHECAECAFCgBAYLAjsfEwYYhI5LKhfCj0VkXo/lYgJZ8Na4Jn3EPce/mw33/bHetr+nYwJ53racw1U3vPGXjty1UPHsyDQ07maz/lZbkOHtgKBfDCncwgBQoAQIAQIgcGCwE7JhLfZ4G2FQEKXbiMEttlkoBttFwS20TTaYW6TD8j93dh82kDnEAKEACFACHSNQE/X6q5r6/xtf9ff+Y50hBDY8RAgJpy3U2gvBq+nqwydv+0R6MWw0iUDCIFtP6O27x3zGZr+bmE+baBzCAFCgBAgBLpGoKdrdde1df62v+vvfEc6QgjseAgQE+5PJpzPePd0JaLz+xaBfMaIzhm4CPTtbNnxaxu4I0UtJwQIAUKAEPAi0NNfHO+1+ZT7u/582kDnEALbGwFiwsSEe7oWDq7zt/cTyKkB/YrA4Jqt3femX8GkygkBQoAQIAS2GQKdVnzLsgzd6HTYOdDThjnX5fu3p/XT+YTAQECAmDAx4XzXwMF53kB4Sokt9x6BwTlrc/eK5jMhQAgQAoTA4ECAc9M029raXPZrGGoqlbAsw31nhIC1uJ5WTS9VNi3LMLkdKRPPFcjk/g3p6pvBgSr1ghDIRICYMDHhrta9wf9d5vPQe8ZF9eyYCAz+GZzZwx1zFKhVhAAhQAgQAj1EwNCNVCqlqqpY5dPppGnqhqG6NNiyDJcJQ9mE5CjctPS0amia4MCWYTpkmJjw9hb4ezgBSCLdNggQE97eD0amHEuftjUCtDANbgS29Xza3vcb3KNJvSMECAFCYCdBQNBa/EnR9HQiEcOiyTmYSLtvlwmDLGtxr0LYMp3fIwvosWVyOCLQg8tM8XZOyuPvToI8dXMnQ4CYsLMubK+Bz2PxoVP6EYHtNe50322DQD9OnR2y6m2DKt2FECAECAFCoKcIcBQ4Rdp6+wcE6ampIy11dLaQrl7nXLe49s03tffdd/+UaTOvvuEmVA0jl7U03v5OcSvFucq5+dXnS/50zdXHHHPMtw85ZI+Je196xe++XL4qnjJ1C+rSOTdQZwxqY2LCPR07On/wIkBMmJjwDinQb7NGDd5ne9tYlezod9lmE2kHuRHNZ0KAECAECIEdEAHkn/CPoXJLs5DsogDKuRbT6tfxtoSZ0D58f9F99/xj0qQjh42MMPsV8oeHP/7GRw12DZxz1VDb9HQMDiTWc2PLR2+9fMA+u8uMyfYlshIpYKGCYPmwg394Ul2ct+EvFDBhYUANV5quzjjfn68dEFVqEiGw1QgQEyYmnO8aODjP2+pHaEengjt5BwfnrM3dq518uKn7hAAhQAjssAgIZS9ogPVEosXk+qYN6z99763rLq055XsHVwZDxUpQMFklwETRF2JMKi8cuv9H38SagDxbnEPsaF1NcCvBrVajbvF5vzq22M9CPqa0M2HGmML8IeYPs2D56ENOrY2BTjiZ1lwmbBAT3mHnCTVs2yJATJiYcG6pemf4Zts+b0SbtzUCO8Mc9vaR5jMhQAgQAoTADogA2DWr3NI15LLxLXX7jBvuVyRJZmHGCvEdZSwkqLDEmI8FwkyO+Jky6qiTL1A5xwu1lJkEO2dgww1LXnlkRIAVKXCN4pMUxvwOGfaFwsznZ0xigXLGRl/+lydbRPYl1Akb3CQmvK2lkR1wTlKTEAFiwsSEvXL0zlemhWBwI7CzzejBPZrUO0KAECAEBigC4Jurc0sXnDa+pe7HRxw6dvRwX9AnMRZhrIKxobZtMxBYxpjf1hCPuvxvL+FPmW7pLYlkA+c8lYwtfPrf5Y4xNGNytKjQz1gAybAiM8nHmCKUxBEW3G3PI89KCHtsYsIDdP5Qs/sNAWLCxIR3Nq6Q2d9+e7Rou3GHQCBztAf/J5rPhAAhQAgQAjsaAm6sLEvnPAVvywRbabBZVlPJ2Gdvv/6rI/YdLQEllhVktAHGgoz5Qqxo72c+a22LQQwtuFDbxJNr33r6XxUR0AMzX7lv1H6zLr36w48/XfzB6z8+av8QY2EfU4TTsCIxXyGTdqnca1Jdi5HUTPG7rHFTI+voHW2SUHu2EwLEhIkJD3520FUPt9ODt0OwxJ2h712N/WD8bmcYU+ojIUAIEAIDCwHxawNtNjlPIBkWh1TOVU3TrLYWbdn/vsVYgZcJy4wVllcefcYKA2JF83SS80bOt7z/0r+HR1F97C/89cV//qIJAmKZEACrNbZu8R4jC6MyKIcViUnAh6MsukfFniesbcTg04gbMWGSwQgBFwFiwsSEByMfyL9PA+vXlFrbUwTynwmD48ye4kPnEwKEACFACPQ3Avbvi8iZZHLQDIM2WGQ/SgsP3o1fHFYCTBjoqw/VwqEQi1ZMvu6+dUB0Odd0Hlv58TuPjRxVwYKFLDL8H8+8aXLepmN6JCvF29Zwa/PJ30O1MFagSIIJTxx70M8b0tgI7CkxYZcFUYEQICZMTHhwMIDe9qK/f/+o/u2LQG/nxUC9bvuiTXcnBAgBQoAQ6IyAsI7mggkLzbCgwbr9jZps++TFA6PAhG1drqIwX4E0euITS2rrOdeSQJeb1n588AFjQB1cOvrB/35ap0PoLMvQQZBNNHJzPdfW/vpHh5cyiMLls3XChYztcuAx0znnQLmxbcSEif4RAi4CxIS3NxPuvGLSEUKAECAECAFCgBAgBAiBQYUAaoMt5MNCLQzG0iCQp+Mtbz14wzDGopD+iMmgFC5gytAh+373q1QM/IlbVmgblxx/7BFMYtHisqeef7U5xds0rgsFs6ZztZXzzTy5clxULkKFMITdUmSmFDE2ZtYV/0xYPKVbQvoXTNjABMUQySuPd292hwfV2BFZGLQIEBMetEPr7nZQgRAgBAgBQoAQIAQIAUJguyIgOHDmv9xMGRo32/5Q89PRMjBhP5DhAGNlLDLhuKkXbuJJlW/h6a+vrPklY0wqrHz6+QWCl6aSummgqbVmcrWN8/UtX7w6MqgERC4lcBL2MaWEsZGPvLw8ZmwVEwbK3tMXMWFCYCAgQEyYmDAhQAgQAoQAIUAIEAKEACHQnwh4da+2ZtjkXDW0er1t9ff2rBgpQz5hDPysMFbMwhPvX/B5ArTGjY//7ZrKAKQN/uv9z7UmMymphSpdLcn1tf978CZMrSSBZhnyDCOjLjnoqy0QVQtMsfEF2Zy41SOdMDHh7bqH0p/TciCQ1X4Fn5gwTS9CgBAgBAgBQoAQIAQIAUKgPxHIzoQT3KprWPnG2CgrZywEmYCDMiiGCwvHH1vbxjU9uXLx2yOjBQoLXfKHOxtMHtd4Kgmstq0t1U6otSRPrXni6rkRcCPGChRJAX/hsvEH/HJVKxcBs4gJ9yunosoHIgLEhPtz1dvpN1oG4iNBbSYECAFCgBAgBAgBQqCPEfAyYVEGItvKzeUP3jSrhLFiySdhdqRIeAhjhcedeen6GOextd/ZtZKx8Pj9f7ayEZisJeis+Nfk3MDcTFqSp9ee/6P9fUCEFaWskPlkxoISG3LJNY/WxtoVwhxVyKQT7uPBJYF/wCJATJiYMCFACBAChAAhQAgQAoQAIdCfCGRlwkYzT3x2xa8PKsEwWYXBQvQULmas+Lb75sfT+uWzf1XEWLBi//dX841ahmrXZsTod8z1JK///KihEH3aLwcgXFYYzKkZ8z360ooNKUjZ5L6wIRCtq0dv9/J8CwOWGhFJ3qkQICbcn6serQKEACFACBAChAAhQAgQAjs5Ai7pFAzURsPkRjNf89/DSlkhuPWySDQgsSBjpUMqdmvatHrha4+VlsIXtz/8xpokb8nUB9uMVFRlpFYufHQkYxWMRZgMDDgcZIyFyyfU1vNmvaNO2HSqctvVbSFfAuyet5OPOHV/gCBATJiYMCFACBAChAAhQAgQAoQAIdBvCKAZMzdE1iMndpWV5Eb90iduHI75k1CFy1AnXH7IQT/gsbUH7DmEMXbCtDktnLeCHXRH42g4YFica9xofuzmi0czcDYGHgzRtWQmFR7989l1MR7TgAm7ZtUierVgrN0SYPcEl+HmWxggRGin0n9SZzsjQEy431Y9WgIIAUKAECAECAFCgBAgBAgBlwkDJ0VKC4UtnK+766JTKpAJF0YjoYIyKVDGfCV/u/XvV8w+R2ZM2WXv19e3NXOu4lVZYjhbaW7Ucb7h9IOGj2IsEkYaHJGYEmRFe19596v1aZ62wBCamHBnFkRHCAFiwsSECQFCgBAgBAgBQoAQIAQIgX5DQDBhU4e0SVxPm4ZppTiv59qyM47etQhMmpks+xR/CTBhJt1w3Q0F/iBTQlc/8Nw6zuOg+Y2DY69Q0dqkVvxRefpzY83Lhw1jlch/wUm4OMCCRazk4Ec+aGqCi1JiL8LRA0OoLqecr7dwvqpg9zza/iAEBgICxIT7bdUbCMNPW0GEACFACBAChAAhQAgQAv2LgMOETa6iqTI3jQTnjRuXvzYsDKl/hUlzIFzKWGBE5YihlaOZUv6TaZdsRmKpQqisODdT3NSBudqvdib80Qs3DJFYqc8xjQ5FmL+k8JBffByHTMKgUfZIu2Qd3b9jTfL/gEKAmLBnbRhQI0ePMSFACBAChAAhQAgQAoTAwEAAGKyuOUyYc51bm956/p8hjJXFmA89hCGPUjgcZYHS4O7ff311cnNSQ1fgNOdJbrZxM9GRCVtt3Fp+7UUnFAVZ2A+1yEyWfBB9+ti5f1jOeYqYMIn3hEBuBIgJExMmBAgBQoAQIAQIAUKAECAE+hMB8A82Va6qbiDnhmUXTf4J+vX6GQvZZFgKM18Rk4dU3/rEJlv3q3LT4JbGrTbQDAPvFy8s6M08/skPDy6QZSYpwIR9LKgEh7HI6LteXbTa4hqcnxZXCTrgXA5/3YBY3Ra8V+VVzs09BsbOBbV/50CAmHB/rno7xxyiFY0QIAQIAUKAECAECAFCICcCGCZLBftmYMKYSsngW5adsP9YEeEKFcJQlAsqmTJ01CE/+TTFmzhXLW6lE3A+vJL4bmfCpm5wozWxceH4SjSwloR1dJApw6XKvT6oq28DrmtAcGlomftqL3dLgN0T3IvzLZAMTAgMBASICRMTJgQIAUKAECAECAFCgBAgBPoLARNddRNg3OzohHWj7p3ndw1CfCvGAsiHGfMXM7mCybtc+e/X05y3cJ7SeVtMNWzlrWECp21/pRLAjV976vpiH2RNsr2EpRDzjdrr2DOb7BPBvtq5xgSrbJuvwzGX6HZbcGrI++9AYEE5dy6o8TsNAsSEnQVhpxlyeuwJAUKAECAECAFCgBAgBLYdAqiWTdlMGOJXca49fesfRwr2KkHALIgdHS5jhePC+5/0pQGRrtoMODPWpgv+qqFvsYeJGjzRyHnyilmTKkJMJy8LNgAACppJREFUFkwYMi9FWMFuk6+4M2bz33YNMNJgFYmxfbBbAuye4LlvfkWSqwmBgYAAMWFiwoQAIUAIEAKEACFACBAChEB/IgAWzqbKdZW3JbXVPLHqol9OGiZ0waDPZT4pwFhILtvlD/95YS3nCVTYCtKZg44mubGGNy/eb7i/RGbApv2MKQrzFbPivR97ex1c6waatpmvihWnPVri/GhtL84aCCxo222FEBo7KgLEhPtz1dtRR52efEKAECAECAFCgBAgBAiBbYSA4LLc1Lhu8FbOa9ObP5pQBCGewaTZJykSk5if+UoCIyYuj/MY6pBd+mlf7X4WBSvOk1+0fPrECIkNCShQj19hSpD5S5VxR37VjIbUhiPlmliAdEoJDKDl2kt3qLTvPpIMTAgMBASICTtrxEAYrW20XhMUhAAhQAgQAoQAIUAIEAJ9iADoYYEJc97KtyxZ++bj5T4W8oMuV5LRy9enMLns7D/c0YQ02NXm5uSmVpyrXz31lzmljJUqkuRjLBBg/gIWqPz+mZfZzsGiFstxCIa7pzoF0Mp5h636og+ho6oIgX5DgJgwMWFCgBAgBAgBQoAQIAQIAUKg3xBA9onZkHRutfJ03QOXVZWGmORjigSpjySFsYDCinaZt7S5Pp0fA7XinK+68Of7V2IKJkUJsEAB85ey4PDb572dwDralcki6BYcFLQY/+Z3n16e1W/UhdRChEAfIkBMuN9WPVoCCAFCgBAgBAgBQoAQIAR2egRMVMWmuGlyneutfPNX53x7F3AO9ttBoyGEdMBf+eOpi9F8OT/ymeBbFh1QxkoZUxgL+iLMX8akoaxyv4VfbUqhGli1uOqkYBJ1gpV02rADSOd3m16etdMPeh+yNaqq/xAgJkxMmBAgBAgBQoAQIAQIAUKAEOgvBEzktykIYKVzrZUvf/eQIhZmTPGzAj/ohOHlK5h26+Mr0Ys3P/KZWPbq3XtGWBFmYQr6CphcxuRR3zry9JVNEHU6xXnKMtr0OPSqZdPj9/5zcyypYtgu+M4bUjq/+/XsLGLChMBAQICYcH+tev23e0E1EwKEACFACBAChAAhQAgMFARQJwyBo7mlcq31v/f9qZCxCGMhhZVFkAZLQalyr1dXqZuQwarc6p6oWqk7f3vaSMYKGAszOShFmDSE+XY99aJbUpguKcHVBE+pECUrfdefLi8PsMOPPb5J1AuEuGfEtsdnDwQWNFDmD7Wz/xAgJkxMmBAgBAgBQoAQIAQIAUKAEOgvBERCYIOr3IpzvbHml0cVIP8NMUchHBk58tsn1SZ5I4dMwinLyE1UTcyBlOK8ccaP9yljLMpYVAoGWYixISww8bf/WKCi6jnFE2mIUx1747UnAni7W/71UAxTK2kxi5hw/5ErqnkAIUBMuL9WvQE0CaiphAAhQAgQAoQAIUAIEAL9hYCtUY1xXs/VNUfsNjzIwDVYZpgHGPx8h9fc8Ogmgye4oRmGZeXMcmSaOudxzr9JbVg4oYAJJhxmvhDz+5QRrOTwN+t4TOWWFgM7a31T29oP9hpX7i8ce9ikqiY00jZMbiRJJ0zyPyEACBATpnlACBAChAAhQAgQAoQAIUAI9BsCwIQNyJ/E16/56JmxxX4FmTCkE5ZlFihi/vH3vL26gXPNauMGBpnOYY6sWzrnLZx/vvTN20oYOAmHmRRkUikriEZ2qTxsxhcqjwGP1ri2Pr5lyUlHToxIbNwBx328AW6vWhzUzSLPcI5b9M1hso4mBAYCAsSE+23VGwjD3197n9R3QoAQIAQIAUKAECAECAGBAOdtRgwYrLnuwT+fD1GjMW60BNbRCguVBw6b9IHJmzg3zAQ3DBDPcrzQajrGzcWP3nC2HSsLsigphSyiBMYcOe22b9DT2FCbeXzp5JMPY4yVjJjwyerNcagcObDJM+NJ57jTVh6moScEBgICxISJCRMChAAhQAgQAoQAIUAIEAL9hgDnDYl6zmNcW3v5GccgEQbT6PJgCeiGA5Xfv+iPqzgwYdNKcb1bJ95Wri2+dsbhYQlqCjIpBPbVjLHoqZf8YwPnMYs3bqmb/LOjojJjvrLHX//MNrbWMZSWaVl5BOTaSiJMuhZCYEAgQEy431a9gbARMiDmKDWSECAECAFCgBAgBAiBAYwA5ypPa+mNvP7zXf2YRViCWFkBFpAZ2DjfvvD9WjCetoCqmt068bby9CfXVX83hPQ3yCS/zYSZPGT0DXfe/8jD8/befa+iSEWgcPRf/vP62pRgtWmugdG0ZXIjt855awmwez2JwYTAQECAmDAxYUKAECAECAFCgBAgBAgBQqDfEMBgzoa++bOX7h3ngyjPMlpHy1AsCA7b6/0N9a0Qzso0OfzXRWDntGlAxKzkh9fOPDKCoacFEw5Hi8MlZaHCIsaYogQYC4XLdq3+7V+/SfM0sFODG3H4xwSVc/daZ5fQ9rowEFjQAN5bIXj7CAFiwv226vXRCNFTSggQAoQAIUAIEAKEACEwgBFAPmnpDb8795ThPlYInr2oGYakwmXHnnNhDJW0Gjc1IMNdEGGuqipYWRsfP3/fBcWYlNiPOuFiVqAwVugvsisuLb/sltubQQeMteka17UUpmhKYtys3Fmaes19My8kMZgQGAgIEBMmJkwIEAKEACFACBAChAAhQAj0JwKcm2bzDw7ftUSBDMAhhgbNcoiFh/3loWc1nVtq2gQmDHGdc+ZQ4thCrYnzNan69wqQCfuY5GNKlPnKWaHMWCRcGiwf+rd5j69WE40cIk1z0+CawXUridGriQkP4C2VgcAtBxa8xIT7c9Wj+UoIEAKEACFACBAChAAhQAhYWkvz5ldffnr+Y/9+et6jT897fP78+Y898fQjTz+/OZbUNDtedPeqWkBSA7WwFfvw3YUVJcWMsYAC+YkL/QURf/iXvzxNtXja4klIKGwaHONjGRY3gRWrHK2jUVHc/b0ytbw9+0QjTggMBASICRMTJgQIAUKAECAECAFCgBAgBPoLAcswDS3Jua6ChXLGC62duaF3pQbOuMCCANPiBepeU6+rq3vppZcWLnyrtvYbcaZpAAM28WWfKv5g/iSDWwbmUSImPLC0l9Ta/kCAmHB/rXr9MVpUJyFACBAChAAhQAgQAoTAAEMAGKplGECDTdO0rAzea+iGpmkZdDe/D6IeTdNEVqRkMsk5T6fTggNnYcIeWpzfHbbirIGgDxxgs4gg7QcEiAkTEyYECAFCgBAgBAgBQoAQIAT6EwHOhUI4lUppelqzX4ZlWalUSmiGe8E7DUPlnHuJtKEbxISJ4hICeSJATLg/V71+2LrIc1zpNEKAECAECAFCgBAgBAiBHQUBJzmSbafsZaum2QPr6Dzoskf1m7OYRzVbdwrJwITAQECAmDAxYUKAECAECAFCgBAgBAgBQqA/EQBeCZ65WZmwl7BuHQGFq7215Spv/V26qWEgsKAdZZeEsNp+CBAT7s9Vb/uNKz3bhAAhQAgQAoQAIUAIEAI7CgIOcSQmvKOMCEnphIDFiQkTEyYECAFCgBAgBAgBQoAQIAT6EwFiwsS7CIEdDwFiwv256u144037cIQAIUAIEAKEACFACBAC2xoBhwl3Yb3sOWXgF0kGJgQGAgL/D7qYyvUIcdo/AAAAAElFTkSuQmCC"
    }
   },
   "cell_type": "markdown",
   "id": "84e95c8e-2d49-4da7-9a3a-303b418fa6c2",
   "metadata": {},
   "source": [
    "![image.png](attachment:f7daf581-9f4e-440f-8507-4a09d370cfdc.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4efaa258-a1d3-4a04-b217-403eaedcc354",
   "metadata": {},
   "source": [
    "тренировать мы будем 5 бинарных моделей, чтобы работать с каждым классом"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c8536fa3-6a0f-45df-92d0-404ee157d3df",
   "metadata": {},
   "source": [
    "# Model training (CatBoost)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "fadf109e-c49e-40c7-9a2d-a4e6a803e302",
   "metadata": {},
   "outputs": [],
   "source": [
    "def downsample(df, value, fraction=1):\n",
    "        \"\"\"\n",
    "        function to downsample majority class.\n",
    "        \"\"\"\n",
    "        # Separate the majority and minority classes\n",
    "        if value == 'Нет события':\n",
    "            majority_class = df[df[value] != False]\n",
    "            minority_class = df[df[value] == False] \n",
    "        else:\n",
    "            majority_class = df[df[value] == False]\n",
    "            minority_class = df[df[value] != False]  \n",
    "        \n",
    "        # Calculate the number of samples needed from the majority class\n",
    "        num_samples = len(minority_class)\n",
    "        \n",
    "        downsampled_majority = majority_class.sample(frac=(len(minority_class) / len(majority_class)) * 2, replace=True) \n",
    "        \n",
    "        downsampled_df = pd.concat([downsampled_majority, minority_class], axis=0)\n",
    "        \n",
    "        return downsampled_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "af2bc5d6-8863-4b2a-b45e-efcb594e5753",
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_model(X, y):\n",
    "    \"\"\"\n",
    "    функция для тренировки модели\n",
    "    \"\"\"\n",
    "    clfs = []\n",
    "    scores = []\n",
    "    kf = StratifiedKFold(n_splits=n_splits, shuffle=True, random_state=0)\n",
    "    \n",
    "    for train_index, test_index in kf.split(X, y):\n",
    "    \n",
    "        X_train, X_test = X.iloc[train_index], X.iloc[test_index]\n",
    "        y_train, y_test = y.iloc[train_index], y.iloc[test_index]\n",
    "        \n",
    "        clf = CatBoostClassifier(\n",
    "            depth=7,\n",
    "            iterations=500,\n",
    "            learning_rate=0.06,\n",
    "            loss_function=\"MultiClass\", \n",
    "            eval_metric=\"TotalF1:average=Macro\", \n",
    "            custom_metric=\"TotalF1:average=Macro\",\n",
    "            boosting_type=\"Ordered\",\n",
    "            \n",
    "            # Главная фишка катбуста - работа с категориальными признаками\n",
    "            cat_features=cat_features,\n",
    "            # ignored_features = ignored_features,\n",
    "            \n",
    "            # Регуляризация и ускорение\n",
    "            colsample_bylevel=0.098,\n",
    "            subsample=0.95,\n",
    "            l2_leaf_reg=9,\n",
    "            min_data_in_leaf=243,\n",
    "            max_bin=187,\n",
    "            random_strength=1,\n",
    "            \n",
    "            # Параметры ускорения\n",
    "            task_type=\"CPU\",    \n",
    "            thread_count=-1,\n",
    "            bootstrap_type=\"Bernoulli\", \n",
    "            \n",
    "            # Важное!\n",
    "            random_seed=7575,\n",
    "            # auto_class_weights=\"SqrtBalanced\",\n",
    "            early_stopping_rounds=50)\n",
    "    \n",
    "        clfs.append(clf)\n",
    "    \n",
    "        clf.fit(\n",
    "            Pool(X_train, y_train, cat_features=cat_features),\n",
    "            eval_set=Pool(X_test, y_test, cat_features=cat_features),\n",
    "            verbose=False,\n",
    "            use_best_model=True,\n",
    "            plot=False)\n",
    "    \n",
    "        y_pred = clf.predict(X_test)\n",
    "        score = f1_score(list(map(str, y_test)), y_pred, average=\"macro\")  # np.mean(np.array(y_pred == y_test))\n",
    "        scores.append(score)\n",
    "        print(f\"fold: f1 score: {score}\")\n",
    "    \n",
    "    assert len(clfs) == n_splits\n",
    "    print(\n",
    "        \"mean f1 score ====>>> \" + str(np.mean(scores, dtype=\"float32\")) + \" <<<====\", np.std(scores).round(4), '\\n'\n",
    "    )\n",
    "    return clfs, np.mean(scores, dtype=\"float32\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6d4f340c-590b-47c5-9428-0e724612bb29",
   "metadata": {},
   "source": [
    "начинаем тренировку моделей! CatBoostClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "144b87fa-0626-44f5-a738-737be56b208a",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Прочие опасности\n",
      "Тип\n",
      "False    570\n",
      "True     285\n",
      "Name: count, dtype: int64\n",
      "fold: f1 score: 0.7890420899854862\n",
      "fold: f1 score: 0.8466128466128466\n",
      "fold: f1 score: 0.7654320987654321\n",
      "fold: f1 score: 0.8525862068965517\n",
      "fold: f1 score: 0.8065610859728507\n",
      "mean f1 score ====>>> 0.8120469 <<<==== 0.0334 \n",
      "\n",
      "Аварии с выбросом опасных/токсичных веществ\n",
      "Тип\n",
      "False    442\n",
      "True     221\n",
      "Name: count, dtype: int64\n",
      "Cannot setitem on a Categorical with a new category (), set the categories first\n",
      "fold: f1 score: 0.7451203066221875\n",
      "fold: f1 score: 0.8446368446368446\n",
      "fold: f1 score: 0.8589607635206786\n",
      "fold: f1 score: 0.7858201058201058\n",
      "fold: f1 score: 0.7200136845706466\n",
      "mean f1 score ====>>> 0.79091036 <<<==== 0.0542 \n",
      "\n",
      "ЖКХ\n",
      "Тип\n",
      "False    3582\n",
      "True     1791\n",
      "Name: count, dtype: int64\n",
      "Cannot setitem on a Categorical with a new category (), set the categories first\n",
      "fold: f1 score: 0.753943175192045\n",
      "fold: f1 score: 0.7851921276502944\n",
      "fold: f1 score: 0.7803980705210394\n",
      "fold: f1 score: 0.7741958090966431\n",
      "fold: f1 score: 0.7422418746045483\n",
      "mean f1 score ====>>> 0.76719415 <<<==== 0.0164 \n",
      "\n",
      "Взрывы/пожары/разрушения\n",
      "Тип\n",
      "False    1076\n",
      "True      538\n",
      "Name: count, dtype: int64\n",
      "Cannot setitem on a Categorical with a new category (), set the categories first\n",
      "fold: f1 score: 0.742874116319135\n",
      "fold: f1 score: 0.7628599848427435\n",
      "fold: f1 score: 0.7796207415772632\n",
      "fold: f1 score: 0.7436023436023436\n",
      "fold: f1 score: 0.7611022787493376\n",
      "mean f1 score ====>>> 0.75801194 <<<==== 0.0137 \n",
      "\n",
      "ДТП\n",
      "Тип\n",
      "False    278\n",
      "True     139\n",
      "Name: count, dtype: int64\n",
      "Cannot setitem on a Categorical with a new category (), set the categories first\n",
      "fold: f1 score: 0.6612252736700596\n",
      "fold: f1 score: 0.7606837606837606\n",
      "fold: f1 score: 0.7194591984548527\n",
      "fold: f1 score: 0.7844155844155845\n",
      "fold: f1 score: 0.7688779688779688\n",
      "mean f1 score ====>>> 0.73893243 <<<==== 0.0444 \n",
      "\n",
      "Опасные природные явления\n",
      "Тип\n",
      "False    358\n",
      "True     179\n",
      "Name: count, dtype: int64\n",
      "Cannot setitem on a Categorical with a new category (), set the categories first\n",
      "fold: f1 score: 0.7511520737327189\n",
      "fold: f1 score: 0.6974789915966387\n",
      "fold: f1 score: 0.7915584415584416\n",
      "fold: f1 score: 0.7967370655939211\n",
      "fold: f1 score: 0.7683982683982684\n",
      "mean f1 score ====>>> 0.761065 <<<==== 0.0358 \n",
      "\n",
      "Нет события\n",
      "Тип\n",
      "True     6306\n",
      "False    3153\n",
      "Name: count, dtype: int64\n",
      "Cannot setitem on a Categorical with a new category (), set the categories first\n",
      "fold: f1 score: 0.7384264048010931\n",
      "fold: f1 score: 0.7364317821955335\n",
      "fold: f1 score: 0.7326677891122775\n",
      "fold: f1 score: 0.7496194063850831\n",
      "fold: f1 score: 0.7322258861693051\n",
      "mean f1 score ====>>> 0.73787427 <<<==== 0.0063 \n",
      "\n",
      "[0.8120469, 0.79091036, 0.76719415, 0.75801194, 0.73893243, 0.761065, 0.73787427]\n",
      "['Прочие опасности', 'Аварии с выбросом опасных/токсичных веществ', 'ЖКХ', 'Взрывы/пожары/разрушения', 'ДТП', 'Опасные природные явления', 'Нет события']\n"
     ]
    }
   ],
   "source": [
    "scores = []\n",
    "total_clfs = []\n",
    "for column in target_values:\n",
    "    print(column)\n",
    "    train_df = merged_selected_df.copy()\n",
    "\n",
    "    train_df = downsample(train_df, column, fraction=0.01)\n",
    "    train_df['Тип'] = train_df['Тип'] == column\n",
    "\n",
    "    print(train_df['Тип'].value_counts())\n",
    "    \n",
    "    n_splits = 5\n",
    "    \n",
    "    target_column = 'Тип'\n",
    "    cat_features = [column for column in train_df.columns if any(map(lambda x: column.startswith(x), cat_features))] + ['meteostation']\n",
    "    for cat_feature in cat_features:\n",
    "        try:\n",
    "            train_df[cat_feature] = train_df[cat_feature].fillna('')\n",
    "        except Exception as e:\n",
    "            print(e)\n",
    "        train_df[cat_feature] = train_df[cat_feature].astype('category')\n",
    "    text_features = ['WW']\n",
    "    \n",
    "    X = train_df.drop(columns=[target_column, 'Район', 'Местное время', *target_values]) # 'meteostation'\n",
    "    y = train_df[target_column]\n",
    "\n",
    "    clfs, f1_value = train_model(X, y)\n",
    "    total_clfs.append(clfs)\n",
    "    scores.append(f1_value)\n",
    "print(scores)\n",
    "print(target_values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "4c653e19-979e-416b-b9a0-c577dd494ce1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.76657635"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cea2272e-f980-410a-9339-394034740f9f",
   "metadata": {},
   "source": [
    "**итоговый f1-score CatBoost модели равен среднему из каждого скора ==== 0.76657635** "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "faaa1c73-67a5-44b8-bf97-27f9cc0942a3",
   "metadata": {},
   "source": [
    "# Explain reasons to select such predicts"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6b503c85-ca8c-484c-a86a-1389d517a0d8",
   "metadata": {},
   "source": [
    "смотрим на фичи, которые оказались важными для модели класса ЖКХ"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "76b8af75-5aaf-4b72-abe6-7c90b3b7d948",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cannot setitem on a Categorical with a new category (), set the categories first\n"
     ]
    }
   ],
   "source": [
    "column = 'ЖКХ'\n",
    "\n",
    "train_df = merged_selected_df.copy()\n",
    "\n",
    "train_df = downsample(train_df, column, fraction=0.01)\n",
    "train_df['Тип'] = train_df['Тип'] == column\n",
    "\n",
    "n_splits = 5\n",
    "\n",
    "target_column = 'Тип'\n",
    "cat_features = [column for column in train_df.columns if any(map(lambda x: column.startswith(x), cat_features))] + ['meteostation']\n",
    "for cat_feature in cat_features:\n",
    "    try:\n",
    "        train_df[cat_feature] = train_df[cat_feature].fillna('')\n",
    "    except Exception as e:\n",
    "        print(e)\n",
    "    train_df[cat_feature] = train_df[cat_feature].astype('category')\n",
    "text_features = ['WW']\n",
    "\n",
    "X = train_df.drop(columns=[target_column, 'Район', 'Местное время', *target_values]) # 'meteostation'\n",
    "y = train_df[target_column]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "ca5436c9-8b68-49f4-b8ac-615f859a5463",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, stratify=y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "424fa8b4-e0d8-43dc-ad50-096aab3340c9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxYAAAU8CAYAAACpWQPNAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy81sbWrAAAACXBIWXMAAA9hAAAPYQGoP6dpAADzxElEQVR4nOzdfVyVVb7///dmg9yJoJgYktiIhQGKipnpaHmTJiqpaNmURg5K3p3mjD5Gj55qJuc3TaPlbQnqoMlY3qQQeDM6mR5tDhVM2DiU5piaSEcoQGsDyvb6/eGD/W0LWnpt2Qmv5+NxHoe91tprfS6cP3i3rnVdFsMwDAEAAACACR7uLgAAAADArY9gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYIEfLS0tTRcvXnR3GQAAAPgJIlgAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0i2EYhruLwK3BsrDG3SUAwA0zZie6uwQAuHFGprsr+EHsWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANILFT1xcXJxeeOEFd5cBAAAAXNNPPlhkZ2drw4YN7i6jjjNnzig1NVVHjhwxNc/58+eVmpqqvLw8F1UGAAAANDxPdxfwQ7Kzs1VcXKzHH3/c3aU4OXPmjFatWqXQ0FDdfffdNzzP+fPntWrVKkmXdyeu9P7778tqtd7w/AAAAEBD+MkHi6bO29vb3SUAAAAAP8hiGIZxPV/Izs7Wb3/7W7322ms6dOiQsrKyVFZWpoiICM2aNUsxMTHKz8/Xa6+9piNHjsjf319jx47VL3/5S6d5CgsL9ec//1kff/yxbDabbr/9dsXHx2vixIny9Lycd0aMGKHi4uI6NaxcudLxX/dPnTqlVatW6cMPP1RFRYVuu+02DRo0SJMnT5avr6/T9z7//HOlpqbq448/VmVlpdq1a6fhw4friSeecNoV+Oqrr5SamqqPPvpIX3/9tZo3b6477rhDo0eP1vDhwx2/gyt1795daWlpunTpktLT05Wbm6tTp06poqJCwcHB6tu3r5555hkFBQVJkvLy8pSSklJnnttvv13Z2dmSLu9iDB8+vM45i8zMTG3evFknTpyQp6enoqOjlZycrNjYWKdxtd8fPXq0li9frsLCQnl7e+uBBx7Qr3/9a/n5+dXzr1w/y8KaHz0WAH5qjNmJ7i4BAG6ckenuCn7QDe9YLF++XHa7XY899phqamqUkZGh6dOn67e//a1efPFFjRo1Sg8//LD27NmjlStXKjQ0VMOGDZMkHTx4ULNnz9Ydd9yhJ554Qi1atNA///lPpaam6ujRo/rjH/8oSfr1r3+t5cuXq7y8XP/5n//pWPvOO++UJH366adKSUlRQECARo8erTZt2ujo0aN66623dOjQIaWlpTlCSmFhoSZPnixPT0+NHTtWwcHBOnDggJYtW6bPP/9cCxYskCTV1NRo2rRpKikpUWJiotq3b69vv/1Wx44d08cff6zhw4erW7duSkpKUnp6ukaNGqVu3bpJklq1aiVJunjxotavX68BAwaof//+8vHxUWFhobKyslRQUKCMjAx5eXnpzjvv1H/+53/qlVde0YMPPqgHH3xQkn7wj/2lS5fqjTfeUFRUlKZOnSqbzaZt27ZpypQpWrRokfr27es0/ujRo/rVr36lESNGaMiQIcrPz1dWVpY8PDw0b968G/2fAAAAAOBww8HCbrdr7dq18vLyknT5j/1f//rX+s1vfqP09HTdc889kqSEhAQNHz5cmzdv1rBhw1RdXa0XX3xR0dHRev311x1/+I8ZM0adOnXSq6++qry8PMXFxemBBx7Qhg0bVF1d7Qgl3/e73/1OrVu31htvvCF/f39H+7333qvZs2dr586dGjFihCRp4cKFunjxotLT09WpUydJ0qOPPqq5c+dq165dGjlypO6991598cUXOnnypGbMmKGJEyfWe+1hYWHq1auX0tPT1aVLlzq1NWvWTLt27ZKPj49Te5cuXbRgwQLt27dPgwcPVnBwsB544AG98sorioiIqPcar3TixAmtX79eXbt21cqVKx2//0ceeURjx47VH//4R/Xu3dtpB+bzzz9Xenq6oqOjHb/r7777Tu+8845+9atfXdeuBQAAAFCfG34qVGJiouOPWkmO/2ofHR3tCBWS5OXlpaioKJ06dUqS9MEHH+jrr7/WiBEj9O2336q8vNzxf3369HGM+SHHjh3T559/rqFDh+rixYtO88TGxsrX11e5ubmSpG+++UaffPKJ+vXr5wgVkmSxWPT0009Lkt577z1JUvPmzSVJ+fn5+uabb27od2OxWByhwm636/z58yovL1fPnj0lSYcPH76heSVp//79MgxDEyZMcPr933bbbY5bx658UlVMTIwjVNTq2bOn7Ha7zpw5c8O1AAAAALVueMeiXbt2Tp9btGghSQoNDa0ztkWLFqqoqJAkffHFF5Iu7zZczddff/2D69fOk5qaqtTU1HrH1AaD2j+ef/azn9UZc+edd8rDw0NFRUWSLp9vePrpp7V27VoNHTpUd911l3r27KlBgwYpKirqB+uqtWfPHmVkZOjIkSOqqXE+m3Du3LkfPc+Vaq+lY8eOdfpq24qKipzC3ZX/VpIUGBgoSY5/FwAAAMCMGw4WHh71b3b80KNRa8+K/8d//IfuuuuuesfcdtttP7h+7TxPPPGEevfuXe+Y2rBzvaZOnaqRI0fq4MGDKigoUFZWltavX68JEyZo5syZP/j9vXv3au7cuYqKitKsWbMUEhKiZs2a6dKlS5oxY4au87y8adf6N2noWgAAANA4NfjjZtu3by9J8vX1Va9evX5wvMViueY8Hh4ePzhP7S7K8ePH6/SdOHFCly5dqvNf9cPCwvTYY4/pscceU3V1tWbMmKE33nhDTzzxhFq1anXVuiRpx44d8vb2VmpqqtM5ixMnTvzo67ua2jr//e9/KywszKmv9vrq26EAAAAAbqYGf/N279691apVK61du7be23Cqqqr03XffOT77+fnp3Llzdf7L+t13362OHTvq7bff1unTp+vMU1NT45i/VatW6tKli/7nf/5Hx44dc4wxDEPp6emS5Hgi07ffflvn1iVvb2916NBB0v+7jan2wHN911C7m3Pp0iWntdasWVNnbO0jcX/s7VH9+vWTxWLR+vXrneosLS1Vdna2br/9dlMv7AMAAABuRIPvWPj6+uq3v/2tZs2apTFjxmjkyJG64447dP78eZ04cULvvfee/vSnPzneUxEdHa0DBw7o5ZdfVpcuXeTh4aGePXuqVatW+t3vfqdnnnlG48eP18iRI/Wzn/1MVVVVOn36tPbu3avp06c7ngo1a9YsTZ48WcnJyY7HzR48eFD/+7//q6FDh+ree++VdPndEr///e81YMAAhYeHy8/PT59++qmysrIUHR3tCBh33nmn/P39tWXLFvn4+CggIECtWrVSz549NXDgQO3du1cpKSmKj49XTU2N9u/fr6qqqjq/j6CgIN1xxx3avXu3wsLC1KpVK/n6+qpfv371/v46dOigJ598Um+88YaSk5M1ePBgx+NmbTabXnzxRd7UDQAAgAbnljdv9+7dW+vWrdO6deu0c+dOlZWVqUWLFgoLC9MvfvELpyc3/eIXv1BRUZHeffddvf3227p06ZJWrlypVq1a6e6779Zf/vIXpaen63/+53/09ttvy9/fX7fffrtGjBjheAqTJN1zzz3685//rNTUVG3ZssXxgrwZM2boiSeecIzr1KmTHnzwQeXn52vXrl2y2+1q27atkpKSnMb5+Pjo97//vV5//XW98sorunDhgrp3766ePXtqyJAhstls2rBhg5YsWaKAgAD169dP06dP18CBA+v8Pl588UW98sorWrFihaqqqnT77bdfNVhI0syZM3XHHXdo8+bNWr58uePJWwsWLHA8nQsAAABoSNf95m00Xbx5G8CtjDdvA7il3QJv3m7wMxYAAAAAGh+CBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANN43Cx+tLS0NCUlJcnLy8vdpQAAAOAnhh0LAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAawQIAAACAaRbDMAx3F4Fbg2VhjbtLqMOYnejuEuoyMt1dAQAAQINjxwIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsGtiIESM0efJkd5cBAAAAuJSnuwtwl7y8PKWkpDi1+fr6Kjw8XPHx8Ro3bpysVqubqnONkydPaufOncrNzdXp06d14cIFhYWFaeDAgXr88cfl6+vr7hIBAADQSDTZYFFryJAh6tOnjwzDUElJiXJycrRo0SIdP35c8+bNc3d5przzzjvavHmz+vXrp6FDh8rT01P5+fl6/fXX9be//U3p6eny8fFxd5kAAABoBJp8sIiMjNSwYcMcnxMTEzV27FhlZmYqJSVFwcHBbqzOnIEDByopKUnNmzd3tCUmJuqOO+7Qn//8Z2VlZenRRx91Y4UAAABoLJp8sLhS8+bNFRMTo7179+rLL79UZmamcnNzderUKVVUVCg4OFh9+/bVM888o6CgIJesmZubq6ysLBUWFqq0tFReXl6KiorS008/rR49etQZ/+6772r16tU6efKkWrZsqYSEBHXt2lXTpk3T888/rxEjRkiS7rnnnnrXe+ihh/TnP/9Z//73v11SPwAAAECwuIJhGDp9+rSkyyFj/fr1GjBggPr37y8fHx8VFhYqKytLBQUFysjIkJeXl+k1s7OzVVFRoWHDhikkJERnz55VVlaWpk6dqpUrV6pbt26Osbt379a8efMUFham5ORkWa1W5eTk6MCBAz96vf/7v/+TpFt6NwYAAAA/LU0+WFRVVam8vFyGYai0tFQbN27U0aNHFRMTo44dO2rXrl11ziF06dJFCxYs0L59+zR48GDTNcyfP7/OQeoxY8Zo3LhxSk9PdwSLmpoavfrqq2rZsqXWrVunFi1aSLp8e9P48eN/1Fp2u11r1qyR1WrVkCFDTNcOAAAASAQLpaamKjU11fHZw8ND/fr107x582SxWByhwm63y2azyW63q2fPnpKkw4cPuyRYfD9U2Gw2XbhwQVarVdHR0Tp8+LCj77PPPlNJSYkmTJjgCBWS5Ofnp9GjR2vZsmU/uNaiRYv0ySefaNq0aerQoYPp2gEAAACJYKFRo0Zp0KBBslgs8vX1Vfv27RUYGOjo37NnjzIyMnTkyBHV1NQ4fffcuXMuqeH06dNasWKFcnNzdf78eac+i8Xi+LmoqEiSFB4eXmeO+tqu9Prrr2vTpk0aNWqUkpKSTFYNAAAA/D9NPli0b99evXr1qrdv7969mjt3rqKiojRr1iyFhISoWbNmunTpkmbMmCHDMEyvb7PZlJycrMrKSo0fP14RERHy9/eXxWLR2rVr9dFHH5leQ7q8M7NmzRqNGDFC//Vf/+WSOQEAAIBaTT5YXMuOHTvk7e2t1NRUp3MWJ06ccNkaH374oUpKSvTcc89p5MiRTn2vv/660+fQ0FBJl198d6X62mqlpqZq1apVGj58uP77v//baRcEAAAAcAUPdxfwU+bhcfnXc+nSJUebYRhas2aNy9aofbv3lbsfubm5TucrJKlz585q3bq1cnJynG7Dstls2rp1a73zr1q1SqtWrdKwYcP03HPPOa4JAAAAcCV2LK5h4MCB2rt3r1JSUhQfH6+amhrt379fVVVVLlsjNjZWwcHBWrx4sYqLi9WmTRsdPXpUO3bsUEREhI4dO+YY6+npqWeffVbz58/XxIkTlZCQIKvVquzsbAUGBqqoqMhpN2LTpk1KTU1V27Ztde+992rXrl1Oa7dq1Ur33Xefy64FAAAATRfB4hqGDBkim82mDRs2aMmSJQoICFC/fv00ffp0DRw40CVrBAQEaPny5Vq6dKk2btwou92uyMhILVmyRFlZWU7BQpKGDh0qT09PrV69WqmpqWrVqpUSEhLUqVMnzZ49W97e3o6xhYWFkqSvvvpKL7zwQp21u3fvTrAAAACAS1gMV5xAhttlZGRo8eLFSk9PV0xMzE1Zw7Kw5ocHNTBjdqK7S6jLyHR3BQAAAA2OG+5vMRcvXpTdbndqs9ls2rx5swIDAxUZGemmygAAANCUcSuUC5SVldX5Y/9Kfn5+8vPzM71WUVGRZs6cqYceekihoaEqLS3V9u3bVVRUpDlz5sjLy8v0GgAAAMD1Ili4wIQJE1RcXHzNMcnJyZoyZYrptYKCghQdHa2dO3eqrKxMVqtVERERmj59ukveAg4AAADcCM5YuEBBQYGqq6uvOaZdu3YKCwtroIpuDs5Y/EicsQAAAE0QOxYuEBsb6+4SAAAAALfi8DYAAAAA07gVCj9aWlqakpKSOCAOAACAOtixAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmGYxDMNwdxG4NVgW1rhlXWN2olvWlZHpnnUBAABuQexYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0gkUDGzFihCZPnuzuMgAAAACX8nR3Ae6Sl5enlJQUpzZfX1+Fh4crPj5e48aNk9VqdVN1rmOz2fTWW2/pr3/9q4qLi+Xl5aXw8HCNGjVKw4cPl8VicXeJAAAAaASabLCoNWTIEPXp00eGYaikpEQ5OTlatGiRjh8/rnnz5rm7PFMuXbqkmTNn6pNPPlF8fLweffRRVVVV6a9//at++9vf6osvvtDMmTPdXSYAAAAagSYfLCIjIzVs2DDH58TERI0dO1aZmZlKSUlRcHCwG6sz5/DhwyooKND48eP161//2tE+duxYJSYmauvWrQQLAAAAuESTDxZXat68uWJiYrR37159+eWXyszMVG5urk6dOqWKigoFBwerb9++euaZZxQUFOSSNXNzc5WVlaXCwkKVlpbKy8tLUVFRevrpp9WjR4864999912tXr1aJ0+eVMuWLZWQkKCuXbtq2rRpev755zVixAhJ0nfffSdJuu2225y+7+XlpcDAQF24cMEl9QMAAAAEiysYhqHTp09Luhwy1q9frwEDBqh///7y8fFRYWGhsrKyVFBQoIyMDHl5eZleMzs7WxUVFRo2bJhCQkJ09uxZZWVlaerUqVq5cqW6devmGLt7927NmzdPYWFhSk5OltVqVU5Ojg4cOFBn3qioKAUEBOiNN95QaGiooqOjVVVVpZycHH322WeaO3eu6doBAAAAiWChqqoqlZeXyzAMlZaWauPGjTp69KhiYmLUsWNH7dq1Sz4+Pk7f6dKlixYsWKB9+/Zp8ODBpmuYP3++fH19ndrGjBmjcePGKT093REsampq9Oqrr6ply5Zat26dWrRoIeny7Vvjx4+vM2+LFi30yiuv6MUXX9ScOXMc7f7+/nr55Zf1wAMPmK4dAAAAkAgWSk1NVWpqquOzh4eH+vXrp3nz5slisThChd1ul81mk91uV8+ePSVdPsPgimDx/VBhs9l04cIFWa1WRUdH6/Dhw46+zz77TCUlJZowYYIjVEiSn5+fRo8erWXLltU7d8eOHdW/f3916dJFFRUV2rx5s+bNm6dFixbpvvvuM10/AAAA0OSDxahRozRo0CBZLBb5+vqqffv2CgwMdPTv2bNHGRkZOnLkiGpqapy+e+7cOZfUcPr0aa1YsUK5ubk6f/68U9/3HwdbVFQkSQoPD68zR31tx44d06RJk/SrX/1KiYmJjvahQ4fq0Ucf1e9//3tlZmY2isfqAgAAwL2afLBo3769evXqVW/f3r17NXfuXEVFRWnWrFkKCQlRs2bNdOnSJc2YMUOGYZhe32azKTk5WZWVlRo/frwiIiLk7+8vi8WitWvX6qOPPrrhuf/yl7+ourpagwYNcmr38fFRnz59tGnTJhUXFyssLMzsZQAAAKCJa/LB4lp27Nghb29vpaamOp2zOHHihMvW+PDDD1VSUqLnnntOI0eOdOp7/fXXnT6HhoZKkk6ePFlnnvraSkpKJF1+n8WV7Ha7JNXZhQEAAABuhIe7C/gp8/C4/Ov5/h/mhmFozZo1Lluj9jakK3c/cnNznc5XSFLnzp3VunVr5eTkON2GZbPZtHXr1jpz33nnnZIuP3Xq+86fP6/9+/erRYsWuuOOO1xyHQAAAGja2LG4hoEDB2rv3r1KSUlRfHy8ampqtH//flVVVblsjdjYWAUHB2vx4sUqLi5WmzZtdPToUe3YsUMRERE6duyYY6ynp6eeffZZzZ8/XxMnTlRCQoKsVquys7MVGBiooqIipzMZjz/+uHbs2KHly5fr2LFj6tq1q86dO6fMzEyVlpbqN7/5DecrAAAA4BIEi2sYMmSIbDabNmzYoCVLliggIED9+vXT9OnTNXDgQJesERAQoOXLl2vp0qXauHGj7Ha7IiMjtWTJEmVlZTkFC+nywWtPT0+tXr1aqampatWqlRISEtSpUyfNnj1b3t7ejrG333671q1bp1WrVumjjz7S7t275ePjo7vuukvPPvusBgwY4JJrAAAAACyGK04gw+0yMjK0ePFipaenKyYm5qasYVnonvMYxuzEHx50UxbOdM+6AAAAtyDOWNxiLl686Dh4Xctms2nz5s0KDAxUZGSkmyoDAABAU8atUC5QVlZW54/9K/n5+cnPz8/0WkVFRZo5c6YeeughhYaGqrS0VNu3b1dRUZHmzJkjLy8v02sAAAAA14tg4QITJkxQcXHxNcckJydrypQpptcKCgpSdHS0du7cqbKyMlmtVkVERGj69OkueQs4AAAAcCM4Y+ECBQUFqq6uvuaYdu3a3fIvouOMBQAAAK6GHQsXiI2NdXcJAAAAgFtxeBsAAACAadwKhR8tLS1NSUlJHBAHAABAHexYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATLMYhmG4uwjcGiwLaxp0PWN2YgMultlwawEAADRC7FgAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCRQPJzs5WXFyc8vLy3F0KAAAA4HIECwAAAACmESwAAAAAmEawaOKqqqpUU9Owb9QGAABA40OwaGCGYWj9+vVKSEhQ7969NXr0aOXk5NQZl5mZqV/84hfq06eP+vfvr2nTpqmgoMBpzJkzZxQXF6fU1NQ6309NTVVcXJzOnDnjaHvhhRcUFxensrIy/fa3v9VDDz2kn//85zp79qzLrxMAAABNi6e7C2hqVqxYoerqao0ePVrNmjXTli1b9MILLygsLEyxsbGSpKVLl+qNN95QVFSUpk6dKpvNpm3btmnKlClatGiR+vbta6qGadOmKTg4WJMmTVJlZaX8/PxccGUAAABoyggWDezChQt644035OXlJUkaOHCgEhIStGnTJsXGxurEiRNav369unbtqpUrVzrGPfLIIxo7dqz++Mc/qnfv3rJarTdcQ8eOHfXiiy+65HoAAAAAiVuhGtzYsWMdYUGS2rRpo/bt2+vLL7+UJO3fv1+GYWjChAlO42677TaNGDFCxcXFOnLkiKkannjiCVPfBwAAAK5EsGhg7dq1q9MWGBioiooKSXKciejYsWOdcbVtRUVFpmoIDw839X0AAADgSgSLBubhUf+v3DCM657LYrFctc9ut1+1z8fH57rXAgAAAK6FYPETU7uj8e9//7tO3/Hjx53GtGjRQpJ07ty5OmPN7moAAAAA14Ng8RPTr18/WSwWrV+/3un9EqWlpcrOztbtt9+uu+++W5Lk7++v4OBgffTRR047HqdPn9a+ffsaunQAAAA0YTwV6iemQ4cOevLJJ/XGG28oOTlZgwcPdjxu1maz6cUXX3R6ItS4ceP0+uuva+bMmerfv79KS0v19ttvq2PHjiosLHTjlQAAAKApIVj8BM2cOVN33HGHNm/erOXLl8vLy0tRUVFasGCBunXr5jR24sSJ+vbbb7Vjxw7l5+frzjvv1H//93/r008/JVgAAACgwViMGzk1jCbJsrDmhwe5kDE7sQEXy2y4tQAAABohzlgAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDQeN4sfLS0tTUlJSfLy8nJ3KQAAAPiJYccCAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgmsUwDMPdReDWYFlY0yDrGLMTG2CRzJu/BgAAQBPCjgUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYNLARI0Zo8uTJ7i4DAAAAcClPdxfgLnl5eUpJSXFq8/X1VXh4uOLj4zVu3DhZrVY3Vec6X3/9tVJTU/X+++/r66+/VnBwsB588EFNmTJFAQEB7i4PAAAAjUSTDRa1hgwZoj59+sgwDJWUlCgnJ0eLFi3S8ePHNW/ePHeXZ8o333yjp556SiUlJRo9erQ6duyof//739qyZYv+8Y9/6M9//rN8fHzcXSYAAAAagSYfLCIjIzVs2DDH58TERI0dO1aZmZlKSUlRcHCwG6szJz09XcXFxVqwYIGGDh3qaO/SpYvmz5+vjIwM/fKXv3RjhQAAAGgsOGNxhebNmysmJkaGYejLL7/UmjVrlJycrCFDhui+++5TfHy8/vCHP6i8vNxla+bm5mru3LlKSEhQnz599MADD2jatGnKz8+vd/y7776r8ePH6/7771d8fLzS0tL0wQcfKC4uTtnZ2Y5xeXl58vb21pAhQ5y+/9BDD8nb29tpLAAAAGBGk9+xuJJhGDp9+rSkyyFj/fr1GjBggPr37y8fHx8VFhYqKytLBQUFysjIkJeXl+k1s7OzVVFRoWHDhikkJERnz55VVlaWpk6dqpUrV6pbt26Osbt379a8efMUFham5ORkWa1W5eTk6MCBA3XmvXDhgry9vWWxWJzaPTw85O3traKiIpWXlysoKMj0NQAAAKBpa/LBoqqqSuXl5TIMQ6Wlpdq4caOOHj2qmJgYdezYUbt27apzDqFLly5asGCB9u3bp8GDB5uuYf78+fL19XVqGzNmjMaNG6f09HRHsKipqdGrr76qli1bat26dWrRooWky7dvjR8/vs68P/vZz/Tee+/pyJEjuvvuux3tR44c0blz5yRJX331FcECAAAApjX5W6FSU1M1aNAgDR48WOPHj9c777yjfv36aeHChbJYLI5QYbfbdf78eZWXl6tnz56SpMOHD7ukhu+HCpvNpvLyclmtVkVHR+tf//qXo++zzz5TSUmJhg8f7ggVkuTn56fRo0fXmffxxx+Xh4eH5s6dq4MHD+qrr77S+++/r7lz58rT83KmrKqqcsk1AAAAoGlr8jsWo0aN0qBBg2SxWOTr66v27dsrMDDQ0b9nzx5lZGToyJEjqqmpcfpu7X/1N+v06dNasWKFcnNzdf78eae+79/GVFRUJEkKDw+vM0d9bd26ddP/9//9f/rTn/6kZ599VpJktVqVkJCgsrIyvffee/L393fJNQAAAKBpa/LBon379urVq1e9fXv37tXcuXMVFRWlWbNmKSQkRM2aNdOlS5c0Y8YMGYZhen2bzabk5GRVVlZq/PjxioiIkL+/vywWi9auXauPPvrI1PyDBg3Sgw8+qGPHjslmsyk8PFytWrXShAkTZLVadccdd5i+BgAAAKDJB4tr2bFjh7y9vZWamup0zuLEiRMuW+PDDz9USUmJnnvuOY0cOdKp7/XXX3f6HBoaKkk6efJknXnqa6tltVqdzliUlpbqyJEj6tGjB++xAAAAgEs0+TMW1+LhcfnXc+nSJUebYRhas2aNy9aofbv3lbsfubm5dc5wdO7cWa1bt1ZOTo7TbVg2m01bt279UetdunRJCxcu1KVLl/T000+brB4AAAC4jB2Laxg4cKD27t2rlJQUxcfHq6amRvv373fpgefY2FgFBwdr8eLFKi4uVps2bXT06FHt2LFDEREROnbsmGOsp6ennn32Wc2fP18TJ05UQkKCrFarsrOzFRgYqKKiIqczGTabTRMnTtQDDzygdu3a6dtvv9Vf//pXffrpp5o6dari4uJcdh0AAABo2ggW1zBkyBDZbDZt2LBBS5YsUUBAgPr166fp06dr4MCBLlkjICBAy5cv19KlS7Vx40bZ7XZFRkZqyZIlysrKcgoWkjR06FB5enpq9erVSk1NVatWrZSQkKBOnTpp9uzZ8vb2doz18vJSp06d9Ne//lWlpaXy8fHRPffco2XLlql3794uqR8AAACQJIvhihPIcLuMjAwtXrxY6enpiomJuSlrWBbW/PAgFzBmJzbAIpk3fw0AAIAmhDMWt5iLFy/Kbrc7tdlsNm3evFmBgYGKjIx0U2UAAABoyrgVygXKysrq/LF/JT8/P/n5+Zleq6ioSDNnztRDDz2k0NBQlZaWavv27SoqKtKcOXPk5eVleg0AAADgehEsXGDChAkqLi6+5pjk5GRNmTLF9FpBQUGKjo7Wzp07VVZWJqvVqoiICE2fPl2DBw82PT8AAABwIzhj4QIFBQWqrq6+5ph27dopLCysgSq6OThjAQAAgKthx8IFYmNj3V0CAAAA4FYc3gYAAABgGrdC4UdLS0tTUlISB8QBAABQBzsWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0yyGYRjuLgK3BsvCmps6vzE78SZOnnnz5gYAAAA7FgAAAADMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDRPdxeAG5OXl6eUlBSnNl9fX4WHhys+Pl7jxo2T1Wp1U3UAAABoaggWt7ghQ4aoT58+MgxDJSUlysnJ0aJFi3T8+HHNmzfP3eUBAACgiSBY3OIiIyM1bNgwx+fExESNHTtWmZmZSklJUXBwsBurAwAAQFNBsGhkmjdvrpiYGO3du1dffvmlMjMzlZubq1OnTqmiokLBwcHq27evnnnmGQUFBbm7XAAAADQSBItGxjAMnT59WtLlkLF+/XoNGDBA/fv3l4+PjwoLC5WVlaWCggJlZGTIy8vLzRUDAACgMSBY3OKqqqpUXl4uwzBUWlqqjRs36ujRo4qJiVHHjh21a9cu+fj4OH2nS5cuWrBggfbt26fBgwe7qXIAAAA0JgSLW1xqaqpSU1Mdnz08PNSvXz/NmzdPFovFESrsdrtsNpvsdrt69uwpSTp8+DDBAgAAAC5BsLjFjRo1SoMGDZLFYpGvr6/at2+vwMBAR/+ePXuUkZGhI0eOqKamxum7586da+hyAQAA0EgRLG5x7du3V69evert27t3r+bOnauoqCjNmjVLISEhatasmS5duqQZM2bIMIwGrhYAAACNFcGiEduxY4e8vb2VmprqdM7ixIkT7isKAAAAjZKHuwvAzePhcfmf99KlS442wzC0Zs0ad5UEAACARoodi0Zs4MCB2rt3r1JSUhQfH6+amhrt379fVVVV7i4NAAAAjQzBohEbMmSIbDabNmzYoCVLliggIED9+vXT9OnTNXDgQHeXBwAAgEbEYnCCFz+SZWHNDw8ywZideBMnz7x5cwMAAIAzFgAAAADMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjcfN4kdLS0tTUlKSvLy83F0KAAAAfmLYsQAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhmMQzDcHcRuDVYFtbctLmN2Yk3bW4ZmTdvbgAAAEhixwIAAACACxAsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawaGAjRozQ5MmT3V0GAAAA4FKe7i7AXfLy8pSSkuLU5uvrq/DwcMXHx2vcuHGyWq1uqs514uLi6m339fXVgQMHGrgaAAAANFZNNljUGjJkiPr06SPDMFRSUqKcnBwtWrRIx48f17x589xdnkt069ZNo0aNcmrz9Gzy//QAAABwoSb/12VkZKSGDRvm+JyYmKixY8cqMzNTKSkpCg4OdmN1rtGuXTunawQAAABcrckHiys1b95cMTEx2rt3r7788ktlZmYqNzdXp06dUkVFhYKDg9W3b18988wzCgoKcsmaubm5ysrKUmFhoUpLS+Xl5aWoqCg9/fTT6tGjR53x7777rlavXq2TJ0+qZcuWSkhIUNeuXTVt2jQ9//zzGjFiRJ3vXLx4URcvXpSfn59LagYAAAC+j2BxBcMwdPr0aUmXQ8b69es1YMAA9e/fXz4+PiosLFRWVpYKCgqUkZEhLy8v02tmZ2eroqJCw4YNU0hIiM6ePausrCxNnTpVK1euVLdu3Rxjd+/erXnz5iksLEzJycmyWq3Kycm55nmJd999Vzt37pTdblfLli01ePBgTZ06Vc2bNzddOwAAACARLFRVVaXy8nIZhqHS0lJt3LhRR48eVUxMjDp27Khdu3bJx8fH6TtdunTRggULtG/fPg0ePNh0DfPnz5evr69T25gxYzRu3Dilp6c7gkVNTY1effVVtWzZUuvWrVOLFi0kXb59a/z48fXOHRUVpUGDBumOO+7Qt99+q/fff1+bNm3SP/7xD/35z39mBwMAAAAu0eSDRWpqqlJTUx2fPTw81K9fP82bN08Wi8URKux2u2w2m+x2u3r27ClJOnz4sEuCxfdDhc1m04ULF2S1WhUdHa3Dhw87+j777DOVlJRowoQJjlAhSX5+fho9erSWLVtWZ+5169Y5fR4+fLg6deqk1157TW+++aYmTZpkun4AAACgyQeLUaNGadCgQbJYLPL19VX79u0VGBjo6N+zZ48yMjJ05MgR1dTUOH333LlzLqnh9OnTWrFihXJzc3X+/HmnPovF4vi5qKhIkhQeHl5njvrarmbChAlatWqV3n//fYIFAAAAXKLJB4v27durV69e9fbt3btXc+fOVVRUlGbNmqWQkBA1a9ZMly5d0owZM2QYhun1bTabkpOTVVlZqfHjxysiIkL+/v6yWCxau3atPvroI9NrXMnT01O33XabysvLXT43AAAAmqYmHyyuZceOHfL29lZqaqrTOYsTJ064bI0PP/xQJSUleu655zRy5Einvtdff93pc2hoqCTp5MmTdeapr+1qqqur9X//93+KiYm5gYoBAACAujzcXcBPmYfH5V/PpUuXHG2GYWjNmjUuW6P27d5X7n7k5uY6na+QpM6dO6t169bKyclxug3LZrNp69atdea+2o7EypUrZbfb9fOf/9xk9QAAAMBl7Fhcw8CBA7V3716lpKQoPj5eNTU12r9/v6qqqly2RmxsrIKDg7V48WIVFxerTZs2Onr0qHbs2KGIiAgdO3bMMdbT01PPPvus5s+fr4kTJyohIUFWq1XZ2dkKDAxUUVGR05mMNWvW6PDhw4qLi1NISIgqKyv1/vvvKy8vT9HR0Xr00Udddh0AAABo2ggW1zBkyBDZbDZt2LBBS5YsUUBAgPr166fp06dr4MCBLlkjICBAy5cv19KlS7Vx40bZ7XZFRkZqyZIlysrKcgoWkjR06FB5enpq9erVSk1NVatWrZSQkKBOnTpp9uzZ8vb2dozt0aOHvvjiC+Xk5KiiokIeHh5q3769pk6dql/84hdOYwEAAAAzLIYrTiDD7TIyMrR48WKlp6fftLMTloU1PzzoBhmzE2/a3DIyb97cAAAAkMQZi1vOxYsXZbfbndpsNps2b96swMBARUZGuqkyAAAANGXcCuUCZWVldf7Yv5Kfn59L3nJdVFSkmTNn6qGHHlJoaKhKS0u1fft2FRUVac6cOfLy8jK9BgAAAHC9CBYuMGHCBBUXF19zTHJysqZMmWJ6raCgIEVHR2vnzp0qKyuT1WpVRESEpk+f7pK3gAMAAAA3gjMWLlBQUKDq6uprjmnXrp3CwsIaqKKbgzMWAAAAuBp2LFwgNjbW3SUAAAAAbsXhbQAAAACmESwAAAAAmMYZC/xoaWlpSkpK4slTAAAAqIMdCwAAAACmESwAAAAAmEawAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgmsUwDMPdReDWYFlYc1PmNWYn3pR5ZWTenHkBAABQBzsWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYNHAXnjhBcXFxbm7DAAAAMClPN1dQEO6nj/o33nnHYWGht7EahrOwYMH9ec//1lHjx5Vs2bN1LNnT82cOVPt2rVzd2kAAABoJJpUsPjd737n9Pnjjz/Wtm3bNGrUKHXr1s2pr2XLlg1Z2k2zd+9e/eY3v1GnTp30H//xH/r222/15ptvatKkSVq/fr1uu+02d5cIAACARqBJBYthw4Y5fbbb7dq2bZu6dOlSp68xqKmp0Z/+9CeFhIRo9erV8vPzkyTdf//9evLJJ5WWlqZ58+a5uUoAAAA0BpyxuIa8vDzFxcUpOztb77zzjsaNG6fevXtr+PDhWrduncvWOXHihF566SWNGzdO/fr1U58+ffTEE08oMzOz3vGff/65pk2bpr59+2rgwIF6/vnnVV5erri4OL3wwguOcfn5+SopKdEjjzziCBWSdPfdd6tHjx7avXu3ampqXHYdAAAAaLqa1I7FjXr77bf1zTffaOTIkQoICNDOnTu1bNkyhYSEaOjQoabnz8vL0z/+8Q/17dtXoaGhqqqq0t/+9jctWLBAZWVlSkpKcow9deqUfvnLX8owDD322GO67bbb9P7772vGjBl15i0sLJQkxcTE1OmLjo7WRx99pJMnT6pjx46mrwEAAABNG8HiR/jqq6+0ZcsWNW/eXJKUkJCg4cOHa+PGjS4JFvHx8UpMTHRqe/zxx5WSkqK1a9fqySeflKfn5X+q1157Td99951Wr16t2NhYSdKjjz6quXPn6tNPP3Wao6SkRJLUpk2bOmvWtpWUlBAsAAAAYBq3Qv0II0aMcIQKSfLx8VFMTIxOnTrlkvl9fX0dP1dXV6u8vFznzp3Tfffdp++++04nTpyQdPlMyPvvv6+oqChHqKj1i1/8os68VVVVkiQvL686fc2aNXMaAwAAAJjBjsWPUN9jWQMDA1VRUeGS+W02m9LS0rRnzx793//9X53+c+fOSZLKyspUWVmp8PDwOmM6dOhQp83Hx0eSdPHixTp9Fy5ccBoDAAAAmEGw+BGsVutNnX/evHk6ePCgRo0ape7duyswMFAeHh56//33tWHDBl26dOmG5q19lOzZs2d15513OvWdPXvWaQwAAABgBsHCzc6fP6+DBw9q2LBh+q//+i+nvg8//NDpc8uWLeXr66uTJ0/Wmaf2dqnvu+eeeyRJ//znP9WrVy+nvsOHD8vf37/e3Q8AAADgenHGws08PC7/ExiG4dReWlpa53GzVqtV999/v/71r3+poKDAqe8vf/lLnbl79Oih1q1bKzMzUzabzdF+9OhR5efna9CgQY5D4QAAAIAZ/FXpZv7+/rrvvvu0c+dOeXt7KyoqSsXFxdq6davatWtX5xzHM888o9zcXM2cOVPjxo1TmzZtdPDgQZWXl0uSLBaLY6ynp6dmzZqluXPn6pe//KVGjRql7777Ths2bFDLli01ZcqUhrxUAAAANGLsWPwEvPjiixo5cqQOHDigl19+Wfv27dPUqVM1duzYOmM7dOigtLQ0RUdH680339TKlSsVFBSkP/zhD5Ikb29vp/GDBg3SK6+8ombNmmnx4sVat26dYmNjtWbNmnofQwsAAADcCItx5T04uCV9+umnevLJJzV9+nQ99dRTN2UNy8Kb85ZuY3biDw+6oYkzb868AAAAqIMdi1vQle+eMAxDb7zxhiTVOaQNAAAANATOWLjAt99++4MvmvPy8lJgYKBL1nv88cfVs2dPRUREqLKyUgcOHNDHH3+swYMHq3Pnzi5ZAwAAALgeBAsXWLhwoXJycq45pnv37kpLS3PJev3799eBAwe0Y8cO2e12hYaGKiUl5abdAgUAAAD8EM5YuMDx48dVUlJyzTEtWrS45XcTOGMBAACAq2HHwgV+9rOf6Wc/+5m7ywAAAADchsPbAAAAAEwjWAAAAAAwjTMW+NHS0tKUlJQkLy8vd5cCAACAnxh2LAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAawQIAAACAaRbDMAx3F4Fbg2VhjcvnNGYnunxOGZmunxMAAADXxI4FAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWDSwESNGaPLkye4uAwAAAHApT3cX4C55eXlKSUlxavP19VV4eLji4+M1btw4Wa1WN1XnGidOnNDq1av12WefqaSkRDU1NWrbtq369OmjCRMmqHXr1u4uEQAAAI1Ekw0WtYYMGaI+ffrIMAyVlJQoJydHixYt0vHjxzVv3jx3l2fK2bNnVVpaqgcffFBt2rSR1WrVsWPHtG3bNu3evVsbNmxQq1at3F0mAAAAGoEmHywiIyM1bNgwx+fExESNHTtWmZmZSklJUXBwsBurM+fee+/VvffeW6e9e/fumjNnjrKzszVx4kQ3VAYAAIDGpskHiys1b95cMTEx2rt3r7788ktlZmYqNzdXp06dUkVFhYKDg9W3b18988wzCgoKcsmaubm5ysrKUmFhoUpLS+Xl5aWoqCg9/fTT6tGjR53x7777rlavXq2TJ0+qZcuWSkhIUNeuXTVt2jQ9//zzGjFixDXXa9u2rSTp/PnzLqkfAAAAIFhcwTAMnT59WtLlkLF+/XoNGDBA/fv3l4+PjwoLC5WVlaWCggJlZGTIy8vL9JrZ2dmqqKjQsGHDFBISorNnzyorK0tTp07VypUr1a1bN8fY3bt3a968eQoLC1NycrKsVqtycnJ04MCBq85fXV2tyspKVVdX64svvtDSpUslSX369DFdOwAAACARLFRVVaXy8nIZhqHS0lJt3LhRR48eVUxMjDp27Khdu3bJx8fH6TtdunTRggULtG/fPg0ePNh0DfPnz5evr69T25gxYzRu3Dilp6c7gkVNTY1effVVtWzZUuvWrVOLFi0kXb59a/z48VedPzMzU3/6058cn0NDQ/Xiiy86BRYAAADAjCYfLFJTU5Wamur47OHhoX79+mnevHmyWCyOUGG322Wz2WS329WzZ09J0uHDh10SLL4fKmw2my5cuCCr1aro6GgdPnzY0Vf7dKcJEyY4QoUk+fn5afTo0Vq2bFm98z/wwAPq0KGDKisrdeTIEe3fv1/l5eWm6wYAAABqNflgMWrUKA0aNEgWi0W+vr5q3769AgMDHf179uxRRkaGjhw5opqaGqfvnjt3ziU1nD59WitWrFBubm6dcw8Wi8Xxc1FRkSQpPDy8zhz1tdUKCQlRSEiIpMshY8CAAZowYYKqqqqUlJTkiksAAABAE9fkg0X79u3Vq1evevv27t2ruXPnKioqSrNmzVJISIiaNWumS5cuacaMGTIMw/T6NptNycnJqqys1Pjx4xURESF/f39ZLBatXbtWH330kek1rtSpUyfdfffd2rJlC8ECAAAALtHkg8W17NixQ97e3kpNTXU6Z3HixAmXrfHhhx+qpKREzz33nEaOHOnU9/rrrzt9Dg0NlSSdPHmyzjz1tV1LdXW1KioqrrNaAAAAoH4e7i7gp8zD4/Kv59KlS442wzC0Zs0al61R+3bvK3c/cnNznc5XSFLnzp3VunVr5eTkON2GZbPZtHXr1jpzl5aW1rtmXl6e/v3vfysmJsZs+QAAAIAkdiyuaeDAgdq7d69SUlIUHx+vmpoa7d+/X1VVVS5bIzY2VsHBwVq8eLGKi4vVpk0bHT16VDt27FBERISOHTvmGOvp6alnn31W8+fP18SJE5WQkCCr1ars7GwFBgaqqKjI6UzGSy+9pNLSUvXs2VNt27bVhQsX9Omnn2r37t3y8/PTs88+67LrAAAAQNNGsLiGIUOGyGazacOGDVqyZIkCAgLUr18/TZ8+XQMHDnTJGgEBAVq+fLmWLl2qjRs3ym63KzIyUkuWLFFWVpZTsJCkoUOHytPTU6tXr1ZqaqpatWqlhIQEderUSbNnz5a3t7dT/du3b9eOHTtUVlYmi8Witm3bavTo0ZowYYLjRXkAAACAWRbDFSeQ4XYZGRlavHix0tPTb9otTpaFNT886DoZsxNdPqeMTNfPCQAAgGvijMUt5uLFi7Lb7U5tNptNmzdvVmBgoCIjI91UGQAAAJoyboVygbKysjp/7F/Jz89Pfn5+ptcqKirSzJkz9dBDDyk0NFSlpaXavn27ioqKNGfOHHl5eZleAwAAALheBAsXmDBhgoqLi685Jjk5WVOmTDG9VlBQkKKjo7Vz506VlZXJarUqIiJC06dPd8lbwAEAAIAbwRkLFygoKFB1dfU1x7Rr105hYWENVNHNwRkLAAAAXA07Fi4QGxvr7hIAAAAAt+LwNgAAAADTCBYAAAAATOOMBX60tLQ0JSUl8eQpAAAA1MGOBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwzWIYhuHuInBrsCyscck8xuxEF0ySaX4OAAAAuAw7FgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWDRwEaMGKHJkye7uwwAAADApTzdXYC75OXlKSUlxanN19dX4eHhio+P17hx42S1Wt1UnWukpqZq1apVV+23Wq364IMPGrAiAAAANFZNNljUGjJkiPr06SPDMFRSUqKcnBwtWrRIx48f17x589xdnikDBgzQHXfcUaf9888/1/r169WvXz83VAUAAIDGqMkHi8jISA0bNszxOTExUWPHjlVmZqZSUlIUHBzsxurM6dSpkzp16lSn/fe//70kKSEhoaFLAgAAQCPV5IPFlZo3b66YmBjt3btXX375pTIzM5Wbm6tTp06poqJCwcHB6tu3r5555hkFBQW5ZM3c3FxlZWWpsLBQpaWl8vLyUlRUlJ5++mn16NGjzvh3331Xq1ev1smTJ9WyZUslJCSoa9eumjZtmp5//nmNGDHiqmtVVlZq9+7dCgkJUe/evV1SPwAAAECwuIJhGDp9+rSkyyFj/fr1GjBggPr37y8fHx8VFhYqKytLBQUFysjIkJeXl+k1s7OzVVFRoWHDhikkJERnz55VVlaWpk6dqpUrV6pbt26Osbt379a8efMUFham5ORkWa1W5eTk6MCBAz9qrb/97W/67rvv9Nhjj93yZ0gAAADw09Hkg0VVVZXKy8tlGIZKS0u1ceNGHT16VDExMerYsaN27dolHx8fp+906dJFCxYs0L59+zR48GDTNcyfP1++vr5ObWPGjNG4ceOUnp7uCBY1NTV69dVX1bJlS61bt04tWrSQdPn2rfHjx/+otbKysmSxWDRy5EjTdQMAAAC1mnywSE1NVWpqquOzh4eH+vXrp3nz5slisThChd1ul81mk91uV8+ePSVJhw8fdkmw+H6osNlsunDhgqxWq6Kjo3X48GFH32effaaSkhJNmDDBESokyc/PT6NHj9ayZcuuuc6JEydUUFCge++9V+3atTNdNwAAAFCryQeLUaNGadCgQbJYLPL19VX79u0VGBjo6N+zZ48yMjJ05MgR1dTUOH333LlzLqnh9OnTWrFihXJzc3X+/HmnPovF4vi5qKhIkhQeHl5njvrarpSVlSWJQ9sAAABwvSYfLNq3b69evXrV27d3717NnTtXUVFRmjVrlkJCQtSsWTNdunRJM2bMkGEYpte32WxKTk5WZWWlxo8fr4iICPn7+8tisWjt2rX66KOPTK8hXb6NaseOHQoMDNSDDz7okjkBAACAWk0+WFzLjh075O3trdTUVKdzFidOnHDZGh9++KFKSkr03HPP1Tn38Prrrzt9Dg0NlSSdPHmyzjz1tX3fgQMH9PXXX2v8+PFq1qyZyaoBAAAAZx7uLuCnzMPj8q/n0qVLjjbDMLRmzRqXrVH7ZKYrdz9yc3OdzldIUufOndW6dWvl5OQ43YZls9m0devWa67DbVAAAAC4mdixuIaBAwdq7969SklJUXx8vGpqarR//35VVVW5bI3Y2FgFBwdr8eLFKi4uVps2bXT06FHt2LFDEREROnbsmGOsp6ennn32Wc2fP18TJ05UQkKCrFarsrOzFRgYqKKiIqczGbVKSkr0v//7v4qKilJERITLagcAAABqESyuYciQIbLZbNqwYYOWLFmigIAA9evXT9OnT9fAgQNdskZAQICWL1+upUuXauPGjbLb7YqMjNSSJUuUlZXlFCwkaejQofL09NTq1auVmpqqVq1aKSEhQZ06ddLs2bPl7e1dZ43s7GzZ7XY98sgjLqkZAAAAuJLFcMUJZLhdRkaGFi9erPT0dMXExNyUNSwLa3540I9gzE50wSSZ5ucAAACAy3DG4hZz8eJF2e12pzabzabNmzcrMDBQkZGRbqoMAAAATRm3QrlAWVlZnT/2r+Tn5yc/Pz/TaxUVFWnmzJl66KGHFBoaqtLSUm3fvl1FRUWaM2eOvLy8TK8BAAAAXC+ChQtMmDBBxcXF1xyTnJysKVOmmF4rKChI0dHR2rlzp8rKymS1WhUREaHp06e75C3gAAAAwI3gjIULFBQUqLq6+ppj2rVrp7CwsAaq6ObgjAUAAACuhh0LF4iNjXV3CQAAAIBbcXgbAAAAgGkECwAAAACmccYCP1paWpqSkpJ48hQAAADqYMcCAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhmMQzDcHcRuDVYFta4ZB5jdqILJsk0PwcAAABchh0LAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgGsGigb3wwguKi4tzdxkAAACAS3m6u4CGdD1/0L/zzjsKDQ29idU0DMMw9Ne//lWbNm3SyZMndfHiRbVt21aDBw/W+PHj1bx5c3eXCAAAgEagSQWL3/3ud06fP/74Y23btk2jRo1St27dnPpatmzZkKXdNK+99prS09PVs2dPJScny9PTU/n5+UpNTdX777+v9PR0WSwWd5cJAACAW1yTChbDhg1z+my327Vt2zZ16dKlTl9jUFNTozfffFORkZFasWKFPDwu3/mWmJgoT09P7dy5U0ePHtXdd9/t5koBAABwq2tSweJ65eXlKSUlRc8//7wMw1BGRoa+/PJLBQcHa+zYsZo4caJL1jlx4oTeeust/eMf/9BXX30lu92uO++8U4mJiXrkkUfqjP/888+1ePFiHTp0SN7e3urbt69+9atfadCgQRo+fLheeOEFSZeDRXV1tYKDgx2holbr1q0lSb6+vi65BgAAADRtBIsf4e2339Y333yjkSNHKiAgQDt37tSyZcsUEhKioUOHmp4/Ly9P//jHP9S3b1+FhoaqqqpKf/vb37RgwQKVlZUpKSnJMfbUqVP65S9/KcMw9Nhjj+m2227T+++/rxkzZtSZ18fHR926ddP//u//au3atRo4cKCsVqvy8/O1ZcsWPfzww2rfvr3p+gEAAACCxY/w1VdfacuWLY6DzgkJCRo+fLg2btzokmARHx+vxMREp7bHH39cKSkpWrt2rZ588kl5el7+p3rttdf03XffafXq1YqNjZUkPfroo5o7d64+/fTTOnMvWLBAL7zwgpYvX67ly5dLkiwWi55++mmlpKSYrh0AAACQCBY/yogRI5yenuTj46OYmBh98sknLpn/+7cjVVdXq7KyUpJ033336R//+IdOnDihiIgI2e12vf/++4qKinKEilq/+MUvtGfPnjpzN2vWTO3atVN8fLzuv/9+SdLevXu1Zs0aNWvWTJMmTXLJNQAAAKBpI1j8CO3atavTFhgYqIqKCpfMb7PZlJaWpj179uj//u//6vSfO3dOklRWVqbKykqFh4fXGdOhQ4c6bVVVVXr66ad199136w9/+IOjfciQIZo7d65SU1M1cODAer8LAAAAXA+CxY9gtVpv6vzz5s3TwYMHNWrUKHXv3l2BgYHy8PDQ+++/rw0bNujSpUs3NO/f/vY3nTp1StOmTavTN2jQIO3Zs0cFBQUECwAAAJhGsHCz8+fP6+DBgxo2bJj+67/+y6nvww8/dPrcsmVL+fr66uTJk3XmOXHiRJ22kpISSao3mNjtdqf/DwAAAJjh8cNDcDPVPgbWMAyn9tLSUmVmZjq1Wa1W3X///frXv/6lgoICp76//OUvdea+8847JUk5OTl1+mrboqKibrR0AAAAwIEdCzfz9/fXfffdp507d8rb21tRUVEqLi7W1q1b1a5duzrnOJ555hnl5uZq5syZGjdunNq0aaODBw+qvLxckpzeov3zn/9cUVFRev/995WcnKwHH3xQkvTee+/p448/1qBBgxQZGdlg1woAAIDGix2Ln4AXX3xRI0eO1IEDB/Tyyy9r3759mjp1qsaOHVtnbIcOHZSWlqbo6Gi9+eabWrlypYKCghyHs729vR1jrVarXnvtNSUlJembb77RsmXLtHz5cp07d04zZszQggULGuwaAQAA0LhZjCvvwcEt6dNPP9WTTz6p6dOn66mnnropa1gW1rhkHmN24g8P+sFJMs3PAQAAAJdhx+IWVFVV5fTZMAy98cYbkqRevXq5oyQAAAA0cZyxcIFvv/22zh/7V/Ly8lJgYKBL1nv88cfVs2dPRUREqLKyUgcOHNDHH3+swYMHq3Pnzi5ZAwAAALgeBAsXWLhwYb1PXvq+7t27Ky0tzSXr9e/fXwcOHNCOHTtkt9sVGhqqlJSUm3YLFAAAAPBDOGPhAsePH3e8M+JqWrRoccvvJnDGAgAAAFfDjoUL/OxnP9PPfvYzd5cBAAAAuA2HtwEAAACYxq1Q+NHS0tKUlJQkLy8vd5cCAACAnxh2LAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAawQIAAACAaRbDMAx3F4Fbg2VhjanvG7MTzRVgZJr7PgAAAG4adiwAAAAAmEawAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBIsGNmLECE2ePNndZQAAAAAuRbC4iuzsbMXFxSk7O7ve/jNnziguLk4vvPBCwxZmQlVVlRISEhQXF6c//vGP7i4HAAAAjQjBoglZuXKlysrK3F0GAAAAGiGCRRPx2Wef6c033+Q2LAAAANwUBIufgNzcXM2dO1cJCQnq06ePHnjgAU2bNk35+fn1jn/33Xc1fvx43X///YqPj1daWpo++OCDq966ZbfbtWDBAvXu3VsDBgy42ZcDAACAJsjT3QXg8nmOiooKDRs2TCEhITp79qyysrI0depUrVy5Ut26dXOM3b17t+bNm6ewsDAlJyfLarUqJydHBw4cuOr8GzZs0IkTJ/Tyyy83xOUAAACgCSJY/ATMnz9fvr6+Tm1jxozRuHHjlJ6e7ggWNTU1evXVV9WyZUutW7dOLVq0kCQlJiZq/Pjx9c5dVFSk1NRU/fKXv1RoaKjOnDlzcy8GAAAATRK3Qv0EfD9U2Gw2lZeXy2q1Kjo6Wv/6178cfZ999plKSko0fPhwR6iQJD8/P40ePbreuf/whz+oXbt2euKJJ27eBQAAAKDJY8fCJIvFYnqO06dPa8WKFcrNzdX58+evOn9RUZEkKTw8vM4c9bXt2LFDH3zwgVatWiVPT/6pAQAAcPPw1+ZVeHt7S7r87of6VFZWOo27UTabTcnJyaqsrNT48eMVEREhf39/WSwWrV27Vh999NENzXvhwgW9+uqr6tOnj4KDg/Xll19Kks6ePStJ+vbbb/Xll18qKChIAQEBpq4BAAAAIFhcRWhoqCTpiy++qLe/tr123I368MMPVVJSoueee04jR4506nv99dfrrenkyZN15rmyrbq6WmVlZTp48KAOHjxYZ/zOnTu1c+dO/cd//IeefPJJU9cAAAAAECyuIjIyUiEhIdq9e7eSkpJ02223OfouXryoTZs2yWKxqF+/fqbWsVqtkiTDMJzac3NzdfjwYae2zp07q3Xr1srJydFTTz3lOGdhs9m0detWp7G+vr566aWX6qxXXl6ul156Sffff79GjhypTp06maofAAAAkAgWV+Xp6am5c+dq1qxZeuyxx5SQkKCwsDB988032r17t44fP66kpCR16NDB1DqxsbEKDg7W4sWLVVxcrDZt2ujo0aPasWOHIiIidOzYMaeann32Wc2fP18TJ05UQkKCrFarsrOzFRgYqKKiIseZDE9PTw0aNKjOerVPhWrXrl29/QAAAMCNIFhcQ9++fbVmzRq98cYb2r59u8rLy+Xr66u7775bf/jDHzR48GDTawQEBGj58uVaunSpNm7cKLvdrsjISC1ZskRZWVlOwUKShg4dKk9PT61evVqpqalq1aqVEhIS1KlTJ82ePdv0mQ8AAADgRliMK+/BwS0pIyNDixcvVnp6umJiYm7KGpaFNaa+b8xONFeAkWnu+wAAALhpeI/FLebixYuy2+1ObTabTZs3b1ZgYKAiIyPdVBkAAACaMm6FcoGysrI6f+xfyc/PT35+fqbXKioq0syZM/XQQw8pNDRUpaWl2r59u4qKijRnzhx5eXmZXgMAAAC4XgQLF5gwYYKKi4uvOSY5OVlTpkwxvVZQUJCio6O1c+dOlZWVyWq1KiIiQtOnT3fJmQ8AAADgRnDGwgUKCgpUXV19zTHt2rVTWFhYA1V0c3DGAgAAAFfDjoULxMbGursEAAAAwK04vA0AAADANG6Fwo+WlpampKQkDogDAACgDnYsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpFsMwDHcXgVuDZWGNqe8bsxNNfDnT1NoAAAC4udixAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsGtiIESM0efJkd5cBAAAAuJSnuwtwl7y8PKWkpDi1+fr6Kjw8XPHx8Ro3bpysVqubqnOdS5cu6c0339TWrVtVXFysli1batCgQUpJSZGvr6+7ywMAAEAj0WSDRa0hQ4aoT58+MgxDJSUlysnJ0aJFi3T8+HHNmzfP3eWZ9sorr+itt97Sgw8+qCeeeEJffPGF3nrrLR05ckSvvfaaPDzYtAIAAIB5TT5YREZGatiwYY7PiYmJGjt2rDIzM5WSkqLg4GA3VmfOv//9b23cuFEPPvig/vSnPznaQ0NDtXDhQu3evVtDhw51Y4UAAABoLPjP1Vdo3ry5YmJiZBiGvvzyS61Zs0bJyckaMmSI7rvvPsXHx+sPf/iDysvLXbZmbm6u5s6dq4SEBPXp00cPPPCApk2bpvz8/HrHv/vuuxo/frzuv/9+xcfHKy0tTR988IHi4uKUnZ3tGPfXv/5VhmHo8ccfd/r+qFGj5OPjox07drjsGgAAANC0NfkdiysZhqHTp09Luhwy1q9frwEDBqh///7y8fFRYWGhsrKyVFBQoIyMDHl5eZleMzs7WxUVFRo2bJhCQkJ09uxZZWVlaerUqVq5cqW6devmGLt7927NmzdPYWFhSk5OltVqVU5Ojg4cOFBn3sLCQnl4eCgqKsqp3dvbW3fddZcKCwtN1w4AAABIBAtVVVWpvLxchmGotLRUGzdu1NGjRxUTE6OOHTtq165d8vHxcfpOly5dtGDBAu3bt0+DBw82XcP8+fPrHKQeM2aMxo0bp/T0dEewqKmp0auvvqqWLVtq3bp1atGihaTLt2+NHz++zrwlJSUKCgpSs2bN6vS1adNGn3zyiS5evOiScAQAAICmrckHi9TUVKWmpjo+e3h4qF+/fpo3b54sFosjVNjtdtlsNtntdvXs2VOSdPjwYZcEi++HCpvNpgsXLshqtSo6OlqHDx929H322WcqKSnRhAkTHKFCkvz8/DR69GgtW7bMad6qqqqrhobasHGtMQAAAMCP1eSDxahRozRo0CBZLBb5+vqqffv2CgwMdPTv2bNHGRkZOnLkiGpqapy+e+7cOZfUcPr0aa1YsUK5ubk6f/68U5/FYnH8XFRUJEkKDw+vM0d9bT4+PiorK6t3zQsXLjjGAAAAAGY1+WDRvn179erVq96+vXv3au7cuYqKitKsWbMUEhKiZs2a6dKlS5oxY4YMwzC9vs1mU3JysiorKzV+/HhFRETI399fFotFa9eu1UcffXTDc99222364osvdOHChTq3Q509e1ZBQUHsVgAAAMAlmnywuJYdO3bI29tbqampTv9l/8SJEy5b48MPP1RJSYmee+45jRw50qnv9ddfd/ocGhoqSTp58mSdeepru+eee5Sbm6t//etfTgfAq6urdfToUXXv3t0VlwAAAADwuNlrqX153KVLlxxthmFozZo1Lluj9u3eV+5+5ObmOp2vkKTOnTurdevWysnJcboNy2azaevWrXXmfuihh2SxWLRhwwan9m3btqmqqop3WAAAAMBl2LG4hoEDB2rv3r1KSUlRfHy8ampqtH//flVVVblsjdjYWAUHB2vx4sUqLi5WmzZtdPToUe3YsUMRERE6duyYY6ynp6eeffZZzZ8/XxMnTlRCQoKsVquys7MVGBiooqIipzMZERERGjt2rDZt2qTZs2erT58+jjdvd+/enWABAAAAlyFYXMOQIUNks9m0YcMGLVmyRAEBAerXr5+mT5+ugQMHumSNgIAALV++XEuXLtXGjRtlt9sVGRmpJUuWKCsryylYSNLQoUPl6emp1atXKzU1Va1atVJCQoI6deqk2bNny9vb22n8r3/9a4WGhmrr1q06ePCggoKC9OijjyolJcWxIwMAAACYZTFccQIZbpeRkaHFixcrPT1dMTExN2UNy8KaHx50DcbsRBNfzjS1NgAAAG4u/pP1LebixYuy2+1ObTabTZs3b1ZgYKAiIyPdVBkAAACaMm6FcoGysrI6f+xfyc/PT35+fqbXKioq0syZM/XQQw8pNDRUpaWl2r59u4qKijRnzhweHwsAAAC3IFi4wIQJE1RcXHzNMcnJyZoyZYrptYKCghQdHa2dO3eqrKxMVqtVERERmj59ukveAg4AAADcCM5YuEBBQYGqq6uvOaZdu3YKCwtroIpuDs5YAAAA4GrYsXCB2NhYd5cAAAAAuBWHtwEAAACYxq1Q+NHS0tKUlJTEAXEAAADUwY4FAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADDNYhiG4e4icGuwLKwx9X1jdqKJL2eaWhsAAAA3FzsWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIJFA4uLi9MLL7zg7jIAAAAAl2qUweLvf/+74uLi9Nprr9Xp++c//6m4uDj17t1bVVVVdfpnzJihnj17qry8XKWlpVqxYoVmzJihQYMG3ZKh4MKFC1q5cqVGjhyp3r17KyEhQatXr1ZNjbmX3QEAAADf1yiDRWxsrKxWq/Lz8+v05eXlyWq16uLFizp06JBTX01NjQoKCtSxY0cFBQXpxIkTSk9P1/Hjx3XPPfc0VPkuNXfuXK1evVpxcXH6zW9+ox49emjlypVasGCBu0sDAABAI+Lp7gJuBj8/P0VFRamwsFBVVVXy8fFx9OXn56tXr146cuSI4+dahYWFqqysVI8ePSRJnTt31p49e9SyZUuVl5dr0KBBDX4tZhw8eFD79+/XL37xC/3qV7+SJD3yyCMKCAjQX/7yF40aNUpdu3Z1c5UAAABoDBrljoUk9ejRQxcvXlRBQYGjraamRocOHVL37t3VvXt35eXlOX2ndocjLi5OkuTv76+WLVve9Fp3796tX/3qV4qPj1fv3r01cOBA/frXv9bnn39e7/gtW7Zo9OjR6t27t0aNGqWNGzcqOztbcXFxTtf017/+VZI0fvx4p+/Xft65c+dNuiIAAAA0NY02WNSGg+/fDvX9HYkePXo4PtfKz8+XxWJR9+7dG7TWTZs2ycPDQ6NGjdJvfvMbjRo1SgUFBZo0aZJOnTrlNHbt2rV66aWX5Ofnp2nTpik+Pl7r16/Xpk2b6sz7r3/9S23atFHbtm2d2tu2bavbbrtNhYWFN/W6AAAA0HQ0yluhJKlr167y8vJyChb5+fny8/NT586d1bx5c8cOxn333ef4uVOnTgoMDGzQWpctWyZfX1+ntvj4eD3++OPasGGD5syZI0mqqKjQqlWrFBERoTVr1sjb21vS5dubxowZU2fe0tJS3XnnnfWuedttt+ns2bMuvhIAAAA0VY12x8LHx0fR0dFOuxL5+fnq0qWLPD09deedd6pVq1aO4HHl+YqGVBsqDMPQt99+q/LycrVs2VLh4eE6fPiwY9wHH3yg6upqJSYmOkKFJLVu3VoPP/xwnXmrqqrUrFmzetf09vau96lYAAAAwI1otDsW0uXboT7++GMVFBSoZ8+eOnTokJKSkhz93bp1c5xJqA0Y7ggWn332mVauXKn8/HynW7MkqV27do6fz5w5I0kKDw+vM0d9bT4+Prpw4UK9a1ZXVzsdagcAAADMaNTBokePHlq1apXy8/Pl7++vyspKp/MT3bt31yuvvCKbzab8/Hx5eHg0+PmKr776SpMnT5a/v78mTZqkDh06yMfHRxaLRYsWLaoTNK5H69atVVJSUm9fSUmJ2rRpc8NzAwAAAN/XqINFly5d5O3trby8PPn7+8vb21tRUVGO/h49eshutys/P1+HDh3SXXfdpRYtWjRoje+9955sNpteeeUVx4HzWhUVFU63Mt1+++2SpJMnT6pnz55OY0+ePFln7qioKO3cuVNfffWV0wHur776SiUlJerXr58rLwUAAABNWKM9YyFJzZo1U0xMjD799FMdOHBAXbp0kZeXl6O/Y8eOCgwM1Pr16912vsLD4/I/gWEYTu3btm3T119/7dTWq1cvNWvWTFu2bFF1dbWjvbS0tN5Hxw4ZMkSS9Oabbzq1136u71wGAAAAcCMa9Y6FJMe7HT755BNNmTLFqc9isahbt27at2+fY+yVVq9eLUmOP+Q///xzR1vt+zDM6NOnj5YtW6bnnntO48aNU0BAgA4dOqS///3vCgsLk91ud4wNCgpScnKyVqxYoUmTJunhhx9WVVWVtm3bpvDwcBUWFspisTjG9+3bVz//+c/1l7/8Rd9++61iYmL0z3/+U1lZWXr44YcVGxtrqnYAAACgVpMIFrXqCwHdu3fXvn37ZLVa1a1btzr9K1eudPp85MgRHTlyRJKUnJxsOliEhYVp6dKlWrFihdLT0+Xh4aGuXbsqNTVVL7/8soqLi53GJyUlyd/fX2+99ZaWL1+utm3b6sknn5RhGCosLHR6WpQkvfTSS1qzZo127typHTt2qE2bNkpJSdFTTz1lqm4AAADg+yzGlffg4Jb08ssva9OmTdq1a5dat259U9awLKwx9X1jdqKJL2eaWhsAAAA3V6M+Y9EYff9sRa3S0lJt375dHTt2vGmhAgAAALiWRn8rVEMoLS39wTHNmzd3yXsj8vPztWTJEg0YMEBt2rTRmTNnlJmZqcrKSs2YMcP0/AAAAMCNIFi4wNChQ39wzPPPP68RI0aYXuuOO+5QWFiYtm3b5ngc7T333KOnnnpKvXr1Mj0/AAAAcCM4Y+ECH3zwwQ+OaQy3KXHGAgAAAFfDjoULsFMAAACApo7D2wAAAABM41Yo/GhpaWlKSkpyens5AAAAILFjAQAAAMAFCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwzWIYhuHuInBrsCysue7vGLMTr/MLmde9BgAAANyPHQsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGCap7sLwI8XFxf3o8e+8847Cg0NvYnVAAAAAP8PweIW8rvf/c7p88cff6xt27Zp1KhR6tatm1Nfy5YtG7I0AAAANHEEi1vIsGHDnD7b7XZt27ZNXbp0qdMHAAAANCTOWDRieXl5iouLU3Z2tt555x2NGzdOvXv31vDhw7Vu3Tp3lwcAAIBGhB2LJuDtt9/WN998o5EjRyogIEA7d+7UsmXLFBISoqFDh7q7PAAAADQCBIsm4KuvvtKWLVvUvHlzSVJCQoKGDx+ujRs3EiwAAADgEtwK1QSMGDHCESokycfHRzExMTp16pQbqwIAAEBjQrBoAtq1a1enLTAwUBUVFW6oBgAAAI0RwaIJsFqt7i4BAAAAjRzBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgmsUwDMPdReDWYFlYc93fMWYnXucXMq97DQAAALgfOxYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA03mOBHy0tLU1JSUny8vJydykAAAD4iWHHAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYZjEMw3B3Ebg1WBbWXNd4Y3bi9S1gZF7feAAAAPxksGMBAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1g0sBEjRmjy5MnuLgMAAABwKYLFVWRnZysuLk7Z2dn19p85c0ZxcXF64YUXGraw63DixAnNnz9fiYmJ6t+/v/r06aMxY8bolVdeUWlpqbvLAwAAQCPi6e4CcPOcPXtWpaWlevDBB9WmTRtZrVYdO3ZM27Zt0+7du7Vhwwa1atXK3WUCAACgESBYNGL33nuv7r333jrt3bt315w5c5Sdna2JEye6oTIAAAA0NgSLn4Dc3FxlZWWpsLBQpaWl8vLyUlRUlJ5++mn16NGjzvh3331Xq1ev1smTJ9WyZUslJCSoa9eumjZtmp5//nmNGDHimuu1bdtWknT+/Pmbcj0AAABoeggWPwHZ2dmqqKjQsGHDFBISorNnzyorK0tTp07VypUr1a1bN8fY3bt3a968eQoLC1NycrKsVqtycnJ04MCBq85fXV2tyspKVVdX64svvtDSpUslSX369Lnp1wYAAICmgWDxEzB//nz5+vo6tY0ZM0bjxo1Tenq6I1jU1NTo1VdfVcuWLbVu3Tq1aNFCkpSYmKjx48dfdf7MzEz96U9/cnwODQ3Viy++6BRYAAAAADMIFj8B3w8VNptNFy5ckNVqVXR0tA4fPuzo++yzz1RSUqIJEyY4QoUk+fn5afTo0Vq2bFm98z/wwAPq0KGDKisrdeTIEe3fv1/l5eU37XoAAADQ9BAsTLJYLKbnOH36tFasWKHc3Nw65x6+P39RUZEkKTw8vM4c9bXVCgkJUUhIiKTLIWPAgAGaMGGCqqqqlJSUZLp+AAAAgGBxFd7e3pKkqqqqevsrKyudxt0om82m5ORkVVZWavz48YqIiJC/v78sFovWrl2rjz76yNT89enUqZPuvvtubdmyhWABAAAAlyBYXEVoaKgk6Ysvvqi3v7a9dtyN+vDDD1VSUqLnnntOI0eOdOp7/fXX663p5MmTdeapr+1aqqurVVFRcZ3VAgAAAPXjzdtXERkZqZCQEO3evVslJSVOfRcvXtSmTZtksVjUr18/U+tYrVZJkmEYTu25ublO5yskqXPnzmrdurVycnJ07tw5R7vNZtPWrVvrzH21t2vn5eXp3//+t2JiYkzVDgAAANRix+IqPD09NXfuXM2aNUuPPfaYEhISFBYWpm+++Ua7d+/W8ePHlZSUpA4dOphaJzY2VsHBwVq8eLGKi4vVpk0bHT16VDt27FBERISOHTvmVNOzzz6r+fPna+LEiUpISJDValV2drYCAwNVVFTkdCbjpZdeUmlpqXr27Km2bdvqwoUL+vTTT7V79275+fnp2WefNVU7AAAAUItgcQ19+/bVmjVr9MYbb2j79u0qLy+Xr6+v7r77bv3hD3/Q4MGDTa8REBCg5cuXa+nSpdq4caPsdrsiIyO1ZMkSZWVlOQULSRo6dKg8PT21evVqpaamqlWrVkpISFCnTp00e/ZspzMfQ4YM0fbt27Vjxw6VlZXJYrGobdu2Gj16tCZMmOB4UR4AAABglsW48h4c3JIyMjK0ePFipaen37RbnCwLa65rvDE78foWMDKvbzwAAAB+MjhjcYu5ePGi7Ha7U5vNZtPmzZsVGBioyMhIN1UGAACApoxboVygrKyszh/7V/Lz85Ofn5/ptYqKijRz5kw99NBDCg0NVWlpqbZv366ioiLNmTNHXl5eptcAAAAArhfBwgUmTJig4uLia45JTk7WlClTTK8VFBSk6Oho7dy5U2VlZbJarYqIiND06dNdcuYDAAAAuBGcsXCBgoICVVdXX3NMu3btFBYW1kAV3RycsQAAAMDVsGPhArGxse4uAQAAAHArDm8DAAAAMI1bofCjpaWlKSkpiQPiAAAAqIMdCwAAAACmESwAAAAAmEawAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgmsUwDMPdReDWYFlYc13jjdmJ1zE48/qKAQAAwE8KOxYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0gkUDi4uL0wsvvODuMgAAAACXapTB4u9//7vi4uL02muv1en75z//qbi4OPXu3VtVVVV1+mfMmKGePXuqvLxcpaWlWrFihWbMmKFBgwbdcqHAZrMpLS1Nv/rVrzRs2DDFxcVp8uTJ7i4LAAAAjVCjDBaxsbGyWq3Kz8+v05eXlyer1aqLFy/q0KFDTn01NTUqKChQx44dFRQUpBMnTig9PV3Hjx/XPffc01Dlu0x5ebnS0tJUWFioTp06yWq1urskAAAANFKe7i7gZvDz81NUVJQKCwtVVVUlHx8fR19+fr569eqlI0eOOH6uVVhYqMrKSvXo0UOS1LlzZ+3Zs0ctW7ZUeXm5Bg0a1ODXYkbr1q21fft2hYSESJJ+/vOfu7kiAAAANFaNMlhIUo8ePfTJJ5+ooKBA9913n6TLOxKHDh3SpEmT5O/vr7y8PKfv1O5wxMXFSZL8/f3l7+9/02vdvXu3du7cqaNHj+qbb76Rn5+fYmNjlZKSok6dOtUZv2XLFm3YsEHFxcVq27atHnvsMfn5+em3v/2tVq5c6ai/WbNmjlABAAAA3EyN8lYo6f+Fg+/fDvX9HYkePXo4PtfKz8+XxWJR9+7dG7TWTZs2ycPDQ6NGjdJvfvMbjRo1SgUFBZo0aZJOnTrlNHbt2rV66aWX5Ofnp2nTpik+Pl7r16/Xpk2bGrRmAAAA4Psa7Y5F165d5eXl5RQs8vPz5efnp86dO6t58+aOHYz77rvP8XOnTp0UGBjYoLUuW7ZMvr6+Tm3x8fF6/PHHtWHDBs2ZM0eSVFFRoVWrVikiIkJr1qyRt7e3JOmRRx7RmDFjGrRmAAAA4Psa7Y6Fj4+PoqOjnXYl8vPz1aVLF3l6eurOO+9Uq1atHMHjyvMVDak2VBiGoW+//Vbl5eVq2bKlwsPDdfjwYce4Dz74QNXV1UpMTHSECunyWYqHH364wesGAAAAajXaHQvp8u1QH3/8sQoKCtSzZ08dOnRISUlJjv5u3bo5zlnUBgx3BIvPPvtMK1euVH5+vtOtWZLUrl07x89nzpyRJIWHh9eZo742AAAAoKE06mDRo0cPrVq1Svn5+fL391dlZaXT+Ynu3bvrlVdekc1mU35+vjw8PBr8fMVXX32lyZMny9/fX5MmTVKHDh3k4+Mji8WiRYsW1QkaAAAAwE9Row4WXbp0kbe3t/Ly8uTv7y9vb29FRUU5+nv06CG73a78/HwdOnRId911l1q0aNGgNb733nuy2Wx65ZVXHAfOa1VUVKhZs2aOz7fffrsk6eTJk+rZs6fT2JMnT978YgEAAICraLRnLKTLj1uNiYnRp59+qgMHDqhLly7y8vJy9Hfs2FGBgYFav369285XeHhc/icwDMOpfdu2bfr666+d2nr16qVmzZppy5Ytqq6udrSXlpZq586dN79YAAAA4Coa9Y6FdPmcRV5enj755BNNmTLFqc9isahbt27at2+fY+yVVq9eLUmOP+Q///xzR1v37t1N3zrVp08fLVu2TM8995zGjRungIAAHTp0SH//+98VFhYmu93uGBsUFKTk5GStWLFCkyZN0sMPP6yqqipt27ZN4eHhKiwslMVicZp/48aNOn/+vKTL7/H46quvHPXfdddd6tevn6n6AQAAAKmJBIta9YWA7t27a9++fbJarerWrVud/pUrVzp9PnLkiI4cOSJJSk5ONh0swsLCtHTpUq1YsULp6eny8PBQ165dlZqaqpdfflnFxcVO45OSkuTv76+33npLy5cvV9u2bfXkk0/KMAwVFhY6PS1KkjIyMpzmOHPmjOOahg8fTrAAAACAS1iMK+/BwS3p5Zdf1qZNm7Rr1y61bt36pqxhWVhzXeON2YnXMTjz+ooBAADAT0qjPmPRGH3/bEWt0tJSbd++XR07drxpoQIAAAC4lkZ/K1RDKC0t/cExzZs3l4+Pj+m18vPztWTJEg0YMEBt2rTRmTNnlJmZqcrKSs2YMcP0/AAAAMCNIFi4wNChQ39wzPPPP68RI0aYXuuOO+5QWFiYtm3b5ngc7T333KOnnnpKvXr1Mj0/AAAAcCM4Y+ECH3zwwQ+OaQy3KXHGAgAAAFfDjoULsFMAAACApo7D2wAAAABM41Yo/GhpaWlKSkpyens5AAAAILFjAQAAAMAFCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwzWIYhuHuInBrsCysua7xxuzE6xiceX3FAAAA4CeFHQsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAawaKBxcXF6YUXXnB3GQAAAIBLNcpg8fe//11xcXF67bXX6vT985//VFxcnHr37q2qqqo6/TNmzFDPnj1VXl6u0tJSrVixQjNmzNCgQYNu+VBQWlqqBx98UHFxcVq/fr27ywEAAEAj0iiDRWxsrKxWq/Lz8+v05eXlyWq16uLFizp06JBTX01NjQoKCtSxY0cFBQXpxIkTSk9P1/Hjx3XPPfc0VPk3zcsvvyy73e7uMgAAANAINcpg4efnp6ioKBUWFtbZlcjPz1evXr0UHBxcJ3gUFhaqsrJSPXr0kCR17txZe/bs0fbt2/W73/2uweq/Gfbv3699+/bpl7/8pbtLAQAAQCPk6e4CbpYePXrok08+UUFBge677z5Jl3ckDh06pEmTJsnf3195eXlO36kNGnFxcZIkf39/+fv73/Rad+/erZ07d+ro0aP65ptv5Ofnp9jYWKWkpKhTp051xm/ZskUbNmxQcXGx2rZtq8cee0x+fn767W9/q5UrVzrqr/Xdd9/p5Zdf1pgxYxrFzgsAAAB+ehrljoX0/8LB93clvr8j0aNHD8fnWvn5+bJYLOrevXuD1rpp0yZ5eHho1KhR+s1vfqNRo0apoKBAkyZN0qlTp5zGrl27Vi+99JL8/Pw0bdo0xcfHa/369dq0adNV51++fLnsdrumTZt2sy8FAAAATVSj3bHo2rWrvLy8nIJFfn6+/Pz81LlzZzVv3tyxg3Hfffc5fu7UqZMCAwMbtNZly5bJ19fXqS0+Pl6PP/64NmzYoDlz5kiSKioqtGrVKkVERGjNmjXy9vaWJD3yyCMaM2ZMvXP/85//1Ntvv60FCxaoefPmN/dCAAAA0GQ12h0LHx8fRUdHO+1K5Ofnq0uXLvL09NSdd96pVq1aOYLHlecrGlJtqDAMQ99++63Ky8vVsmVLhYeH6/Dhw45xH3zwgaqrq5WYmOgIFZLUunVrPfzww3Xmramp0YIFC9SrVy899NBDN/9CAAAA0GQ12h0L6fLtUB9//LEKCgrUs2dPHTp0SElJSY7+bt26Oc5Z1AYMdwSLzz77TCtXrlR+fr7TrVmS1K5dO8fPZ86ckSSFh4fXmaO+trVr1+r06dNatGiRiysGAAAAnDXqYNGjRw+tWrVK+fn58vf3V2VlpdP5ie7du+uVV16RzWZTfn6+PDw8Gvx8xVdffaXJkyfL399fkyZNUocOHeTj4yOLxaJFixbVCRo/VmlpqdLT0xUfHy/DMPTll19KkkpKSiRdvq3qyy+/VOvWrevchgUAAABcr0YdLLp06SJvb2/l5eXJ399f3t7eioqKcvT36NFDdrtd+fn5OnTokO666y61aNGiQWt87733ZLPZ9Morr9R5mlNFRYWaNWvm+Hz77bdLkk6ePKmePXs6jT158qTT56+//lrV1dXaunWrtm7dWmfdtWvXOg6CDxo0yFWXAwAAgCaqUQeLZs2aKSYmRh9//LE8PDzUpUsXeXl5Ofo7duyowMBArV+/3m3nKzw8Lh9zMQzDqX3btm36+uuvHWFCknr16qVmzZppy5YtGjFihOOcRWlpqXbu3On0/Xbt2umll16qs97x48eVlpam+Ph4/fznP1eXLl1cfUkAAABoghp1sJAun7PIy8vTJ598oilTpjj1WSwWdevWTfv27XOMvdLq1aslSdXV1ZKkzz//3NHWvXt307dO9enTR8uWLdNzzz2ncePGKSAgQIcOHdLf//53hYWFOb0pOygoSMnJyVqxYoUmTZqkhx9+WFVVVdq2bZvCw8NVWFgoi8UiSWrevHm9OxG1Z0oiIiLYqQAAAIDLNIlgUau+ENC9e3ft27dPVqtV3bp1q9O/cuVKp89HjhzRkSNHJEnJycmmg0VYWJiWLl2qFStWKD09XR4eHuratatSU1P18ssvq7i42Gl8UlKS/P399dZbb2n58uVq27atnnzySRmGocLCQqenRQEAAAANxWJceQ8Obkkvv/yyNm3apF27dql169Y3ZQ3LwprrGm/MTryOwZnXVwwAAAB+Uhrteywaq9pbsr6vtLRU27dvV8eOHW9aqAAAAACupdHfCtUQSktLf3BM8+bN5ePjY3qt/Px8LVmyRAMGDFCbNm105swZZWZmqrKyUjNmzDA9PwAAAHAjCBYuMHTo0B8c8/zzz2vEiBGm17rjjjsUFhambdu2OR5He8899+ipp55Sr169TM8PAAAA3AjOWLjABx988INjGsNtSpyxAAAAwNWwY+EC7BQAAACgqePwNgAAAADTuBUKP1paWpqSkpKc3l4OAAAASOxYAAAAAPj/27vv8CjKvY3j96aTTocQKRJqCCWEIiBNBKRKL0o3gIigR1AUVOweOlKkGpByRJBy6EXAIyAtUqSrFCWihBICpJFk3j+4si9LEgiZhYTk+7kuL9lnnp35zezsZu+deWbsgGABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0yyGYRhZXQQeD5ZxiRnuawzvmPEZGysfvBgAAABkKxyxAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAApuW6YBEREaE33nhDTZo0UUhIiEaPHq24uDiNHTtWLVu2VM2aNdW6deuHsuy//vpLISEhmjlz5kOZPwAAAJBVnLK6gEftgw8+0K+//qq+ffsqf/788vf31/z587VkyRL16NFDAQEB8vDwkCQtXLhQ//vf/3Tu3DlFR0fL29tbJUuWVNeuXdWoUaMsXpOMuXHjhqZPn65t27bp2rVr8vf3V+fOndWhQwdZLJasLg8AAAA5RK4KFgkJCTpw4IA6d+6sHj16WNsnTZqkgIAADR061Kb/0aNH5efnp7p168rX11fR0dHasmWLhg8froEDB+qll1561KvwQG7duqVBgwbp5MmT6tKli0qVKqVdu3bp888/1+XLlzVgwICsLhEAAAA5RK4KFleuXJFhGPL29rZpv3z5sgoXLpyq/2effZaqrVu3burRo4e+/vpr9enTR46Ojg+tXrNWrlypY8eOadiwYerataskqV27dho+fLjCwsLUpk0bFS1aNIurBAAAQE6Qa4LF6NGjtWbNGknS7NmzNXv2bJvpERERCgkJkSSFhoam+2u+k5OTChYsqN9++02JiYl2CRZLly7V9u3bdfr0aV29elU+Pj6qWbOmXn75Zfn5+dn0TUpKUlhYmFauXKkrV66oePHi6tu3r86cOaPZs2frv//9r/U5GzZskJubm9q1a2czj+7du2vbtm3atGmTevXqZbp+AAAAINcEi/bt26ts2bKaMGGCGjVqpEaNGikiIkJFixbVpEmT5Ovrq759+0qSypQpY/Pca9euKTk5WVFRUdqyZYt++uknhYSEyNXV1S61LVy4UJUqVVKXLl3k4+Oj33//XStXrtS+ffv0zTffyNfX19p3zJgx+u677xQSEqIXX3xRUVFR+ve//50qgCQnJ+vEiRMqX758qjoDAwNlsVh07Ngxu9QPAAAA5JpgUblyZRUoUEATJkxQQECAWrRoYZ02c+ZM5cuXz6btTu3bt9e1a9ckSY6OjmrcuLFGjBhht9q++eYb5cmTx6atfv36GjRokFatWmU9qvD777/ru+++01NPPaXJkyfLweH2Rb2aNGmi7t272zw/Ojpa8fHxKlSoUKrlubi4yNfXV5GRkXZbBwAAAORuuSZYmDF27FglJCTo4sWL2rJli+Lj43Xz5k3lzZvXLvNPCRXJycmKiYlRYmKiypYtK09PTx05csTa78cff5Qkde3a1RoqJCkgIEC1a9fWrl27rG1xcXGSJGdn5zSX6eLiYu0DAAAAmEWwyIDg4GDrv9u0aaN33nlH/fr109KlS1MNBM+Mffv2afbs2Tp69Kji4+Ntpl2/ft3677/++kuSVKJEiVTzKFGihE2wcHNzk3T7ylBpSUhIsPYBAAAAzCJYZEKrVq20adMmbd26Vc8//7ypeR09elSDBw+Wv7+/Bg8eLD8/P7m6uspiseidd95RcnJypubr7e0tV1dXXbx4MdW0hIQERUVF2QQmAAAAwAyCRSakHFWIjo42Pa8NGzYoKSlJX3zxhYoVK2Ztj42NtTlaIck6QPvcuXPy9/e3mXbu3Dmbxw4ODipfvrxOnjyphIQEubi4WKcdPXpUhmGoQoUKpusHAAAAJMnh/l1yp9jYWMXExKRqT0pK0tKlSyVJQUFBppeTcrlawzBs2r/66qtURyuefvppSbcHe9857bffftPu3btTzbtZs2aKi4vT8uXLbdoXL14sR0dHNW3a1HT9AAAAgMQRi3T98ccf6t+/v5555hmVKFFCPj4+unjxojZu3Khz586pVatWqlatmunlNGzYUIsXL9bQoUPVrl07OTs7a8+ePfrtt99sLjMrSaVLl1a7du20YsUKDRo0SA0bNlRUVJSWLl2qcuXK6fjx47JYLNb+7dq10+rVqzVx4kRduHBBpUqV0s6dO7Vt2zb169cv1SVqAQAAgMwiWKSjcOHCatGihQ4ePKjt27fr5s2b8vT0VLly5fTSSy+pefPmdllO1apVNWbMGM2ZM0czZsyQq6uratasqVmzZik0NDRV/xEjRqhgwYJatWqVJk+erBIlSmjEiBE6evSojh8/bnPPCmdnZ02fPl3Tp0/Xxo0bde3aNfn7+2v48OHq3LmzXeoHAAAAJMli3H0ODh5Lr7/+uvbt26cffvjBLncDT4tlXGKG+xrDO2Z8xsbKBy8GAAAA2QpjLB4zad174tdff9WuXbtUo0aNhxYqAAAAgHvhVCiTkpKSdPXq1fv28/HxSfdmdQ9izZo1WrdunerWrau8efPq7NmzWrFihZycnDRgwADT8wcAAAAyg2Bh0j///KM2bdrct9+MGTMUEhJiennly5fX9u3btWTJEl27dk0eHh4KCQlR//79Vb58edPzBwAAADKDMRYmxcfH6+DBg/ftV6FCBbvcpTsrMcYCAAAA6eGIhUmurq6qVatWVpcBAAAAZCkGbwMAAAAwjVOhkGGzZs1Snz597DIIHQAAADkLRywAAAAAmEawAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkWwzCMrC4CjwfLuMQM9zWGd8z4jI2VD14MAAAAshWOWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFo9Y69at1b9//6wuAwAAALArp6wuIDvZv3+/Bg4cmO50R0dH7dmzR5IUEhKi0NBQDRgw4FGVlynJycn6z3/+o+XLl+vChQvKmzevmjRpooEDBypPnjxZXR4AAAByCIJFGpo1a6a6deumandwePwO8EyYMEHffPONGjVqpBdffFFnzpzRN998o5MnT2r69OmP5ToBAAAg+yFYpKF8+fJq0aJFVpdh2u+//64lS5aoUaNGGjt2rLXdz89P48aN06ZNm9S8efMsrBAAAAA5BT9XZwO7d+/W22+/rbZt26pu3bpq2LChXnnlFYWHh6fZ//vvv1e3bt1Up04dtWzZUrNmzdKePXsUEhKi1atXW/tt3LhRhmGoe/fuNs9v166d3NzctG7duoe6XgAAAMg9OGKRhri4OEVFRaVqd3Jykqenp92Xt3r1al27dk0tWrRQ4cKFdfHiRa1atUqDBg3SjBkzVK1aNWvfTZs2aeTIkfL391doaKgcHR21Zs0a/fjjj6nme+zYMTk4OCgwMNCm3dXVVWXLltWxY8fsvi4AAADInQgWaZg5c6ZmzpyZqr1evXqaNGmSpNsDve1l1KhRqQZSd+jQQZ07d1ZYWJg1WCQmJmrixInKmzev5s+fL29vb0lSx44d1a1bt1TzjYyMlK+vr1xcXFJNK1SokA4fPqxbt27J2dnZbusCAACA3IlgkYZ27dqpSZMmqdrz5s37UJZ3Z6iIiYlRQkKCHB0dValSJR05csQ67cSJE4qMjFTPnj2toUKS3N3d1b59e02ZMsVmvnFxcemGhpSwca8+AAAAQEYRLNJQvHhx1apV65Et7/z585o2bZp2796t69ev20yzWCzWf0dEREiSSpQokWoeabW5ubnp6tWraS4zISHB2gcAAAAwi2CRxWJiYhQaGqrY2Fh169ZNAQEB8vDwkMVi0bx587Rv375Mz7tgwYI6c+aMEhISUp0OdfHiRfn6+nK0AgAAAHbBVaGy2N69exUZGal//etfGjBggJ555hnVrl1btWrVUmxsrE1fPz8/SdK5c+dSzSettooVKyo5OVlHjx61aY+Pj9epU6dUsWJFO64JAAAAcjOCRRZzdHSUJBmGYdO+e/dum/EVklShQgUVKFBAa9asUXR0tLU9JiZGy5cvTzXvpk2bymKxaPHixTbtK1asUFxcHPewAAAAgN1wKlQaTpw4ke49Hho2bCh3d3e7Latq1arKnz+/Jk2apAsXLqhQoUI6deqU1q1bp4CAAP3222/Wvk5OTnrttdc0atQo9erVS23btpWjo6NWr14tHx8fRURE2IzJCAgIUKdOnfTtt99q+PDhqlu3rvXO28HBwQQLAAAA2A3BIg0bN27Uxo0b05y2YsUKuwYLLy8vTZ06VV988YWWLFmipKQklS9fXpMnT9aqVatsgoUkNW/eXE5OTpozZ45mzpypfPnyqW3btipTpoyGDx8uV1dXm/5vvPGG/Pz8tHz5cu3YsUO+vr7q0qWLBg4cKAcHDlgBAADAPizG3efg4LG0cOFCTZo0SWFhYQoKCnooy7CMS8xwX2N4x4zP2Fj54MUAAAAgW+En68fMrVu3lJSUZNMWExOjpUuXysfHR+XLl8+iygAAAJCbcSqUHVy9ejXVl/27ubu72+UUqoiICA0ZMkRNmzaVn5+fLl26pLVr1yoiIkIjRozg8rEAAADIEgQLO+jZs6cuXLhwzz6hoaEaMGCA6WX5+vqqUqVKWr9+va5evSpHR0cFBARo8ODBevbZZ03PHwAAAMgMxljYwcGDBxUfH3/PPsWKFZO/v/8jqujhYIwFAAAA0sMRCzuoWrVqVpcAAAAAZCkGbwMAAAAwjVOhkGGzZs1Snz59GCAOAACAVDhiAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMsxiGYWR1EXg8WMYlZrivMbxjBjuuzFwxAAAAyFY4YgEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAExzyuoCYD/79+/XwIED053u6OioPXv2SJJCQkIUGhqqAQMGPKryAAAAkIMRLHKgZs2aqW7duqnaHRw4QAUAAICHg2CRA5UvX14tWrTI6jIAAACQi/ATNgAAAADTOGKRA8XFxSkqKipVu5OTkzw9PR99QQAAAMjxCBY50MyZMzVz5sxU7fXq1dOkSZMk3R7oDQAAANgLwSIHateunZo0aZKqPW/evFlQDQAAAHIDgkUOVLx4cdWqVSurywAAAEAuwuBtAAAAAKYRLAAAAACYRrAAAAAAYBpjLHKgEydOaN26dWlOa9iwodzd3R9xRQAAAMjpCBY50MaNG7Vx48Y0p61YsYJgAQAAALsjWOQgISEh3J8CAAAAWYIxFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSLYRhGVheBx8OsWbPUp08fOTs7Z3UpAAAAyGY4YgEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATLMYhmFkdRF4PFjGJWaonzG8YwY6rTRXDAAAALIVjlgAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBaPWOvWrdW/f/+sLgMAAACwK6esLiCr7N+/XwMHDrRpy5Mnj0qUKKGWLVuqc+fOcnR0zKLqHo64uDh16dJFERER6tSpk956662sLgkAAAA5RK4NFimaNWumunXryjAMRUZGas2aNRo/frxOnz6tkSNHZnV5djVjxgxdvXo1q8sAAABADpTrT4UqX768WrRooZYtW6p3796aN2+eChYsqJUrV+ry5ctZXZ7dnDhxQv/5z384DQsAAAAPRa4/YnE3T09PBQUFaevWrfrzzz+1cuVK7d69W3/88YeuXbum/Pnzq169enr55Zfl6+trl2Xu3r1bq1at0rFjx3Tp0iU5OzsrMDBQffv2VfXq1VP1//777zVnzhydO3dOefPmVdu2bVWlShW98sorev/999W6dWub/klJSfr444/11FNPqXHjxpo0aZJd6gYAAABSECzuYhiGzp8/L+l2yFiwYIEaN26sBg0ayM3NTceOHdOqVat08OBBLVy4UM7OzqaXuXr1al27dk0tWrRQ4cKFdfHiRa1atUqDBg3SjBkzVK1aNWvfTZs2aeTIkfL391doaKgcHR21Zs0a/fjjj+nOf/HixTp79qzGjBljulYAAAAgLbk+WMTFxSkqKkqGYejSpUtasmSJTp06paCgIJUuXVobNmyQm5ubzXMqV66sjz/+WNu3b9ezzz5ruoZRo0YpT548Nm0dOnRQ586dFRYWZg0WiYmJmjhxovLmzav58+fL29tbktSxY0d169YtzXlHRERo5syZeumll+Tn56e//vrLdL0AAADA3XJ9sJg5c6Zmzpxpfezg4KD69etr5MiRslgs1lCRlJSkmJgYJSUlqUaNGpKkI0eO2CVY3BkqYmJilJCQIEdHR1WqVElHjhyxTjtx4oQiIyPVs2dPa6iQJHd3d7Vv315TpkxJNe/PPvtMxYoV04svvmi6TgAAACA9uT5YtGvXTk2aNJHFYlGePHlUvHhx+fj4WKdv3rxZCxcu1MmTJ5WYmGjz3OjoaLvUcP78eU2bNk27d+/W9evXbaZZLBbrvyMiIiRJJUqUSDWPtNrWrVunPXv2aPbs2XJyyvUvNQAAAB6iXP9ts3jx4qpVq1aa07Zu3aq3335bgYGBGjZsmAoXLiwXFxclJyfr1VdflWEYppcfExOj0NBQxcbGqlu3bgoICJCHh4csFovmzZunffv2ZWq+CQkJmjhxourWrav8+fPrzz//lCRdvHhRknTjxg39+eef8vX1lZeXl+n1AAAAQO6W64PFvaxbt06urq6aOXOmzTiLs2fP2m0Ze/fuVWRkpN577z21adPGZtqXX35p89jPz0+SdO7cuVTzubstPj5eV69e1Y4dO7Rjx45U/devX6/169dr6NCh6tGjh9nVAAAAQC5HsLgHB4fbt/lITk62thmGoblz59ptGSl397776Mfu3bttxldIUoUKFVSgQAGtWbNGvXv3to6ziImJ0fLly2365smTR59//nmq5UVFRenzzz9XnTp11KZNG5UpU8Zu6wIAAIDci2BxD88884y2bt2qgQMHqmXLlkpMTNQPP/yguLg4uy2jatWqyp8/vyZNmqQLFy6oUKFCOnXqlNatW6eAgAD99ttv1r5OTk567bXXNGrUKPXq1Utt27aVo6OjVq9eLR8fH0VERFjHZDg5OalJkyaplpdyVahixYqlOR0AAADIDILFPTRr1kwxMTFavHixJk+eLC8vL9WvX1+DBw/WM888Y5dleHl5aerUqfriiy+0ZMkSJSUlqXz58po8ebJWrVplEywkqXnz5nJyctKcOXM0c+ZM5cuXT23btlWZMmU0fPhwubq62qUuAAAA4EFYDHuMQEaWW7hwoSZNmqSwsDAFBQU9lGVYxiXev5MkY3jHDHRaaa4YAAAAZCsOWV0AHsytW7eUlJRk0xYTE6OlS5fKx8dH5cuXz6LKAAAAkJtxKpQdXL16NdWX/bu5u7vL3d3d9LIiIiI0ZMgQNW3aVH5+frp06ZLWrl2riIgIjRgxQs7OzqaXAQAAADwogoUd9OzZUxcuXLhnn9DQUA0YMMD0snx9fVWpUiWtX79eV69elaOjowICAjR48GC73AUcAAAAyAzGWNjBwYMHFR8ff88+xYoVk7+//yOq6OFgjAUAAADSwxELO6hatWpWlwAAAABkKQZvAwAAADCNU6GQYbNmzVKfPn0YIA4AAIBUOGIBAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEyzGIZhZHUReDxYxiVmqJ8xvGMGOq00VwwAAACyFY5YAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWj1jr1q3Vv3//rC4DAAAAsCunrC4gq+zfv18DBw60acuTJ49KlCihli1bqnPnznJ0dMyi6uwnLCxMJ06c0IkTJxQREaGiRYtq9erVWV0WAAAAcphcGyxSNGvWTHXr1pVhGIqMjNSaNWs0fvx4nT59WiNHjszq8kybNm2afHx8VK5cOV2/fj2rywEAAEAOleuDRfny5dWiRQvr444dO6pTp05auXKlBg4cqPz582dhdeatXLlS/v7+kqTOnTsrNjY2iysCAABATpTrg8XdPD09FRQUpK1bt+rPP//UypUrtXv3bv3xxx+6du2a8ufPr3r16unll1+Wr6+vXZa5e/durVq1SseOHdOlS5fk7OyswMBA9e3bV9WrV0/V//vvv9ecOXN07tw55c2bV23btlWVKlX0yiuv6P3331fr1q2tfVNCBQAAAPAwESzuYhiGzp8/L+l2yFiwYIEaN26sBg0ayM3NTceOHdOqVat08OBBLVy4UM7OzqaXuXr1al27dk0tWrRQ4cKFdfHiRa1atUqDBg3SjBkzVK1aNWvfTZs2aeTIkfL391doaKgcHR21Zs0a/fjjj6brAAAAADIr1weLuLg4RUVFyTAMXbp0SUuWLNGpU6cUFBSk0qVLa8OGDXJzc7N5TuXKlfXxxx9r+/btevbZZ03XMGrUKOXJk8emrUOHDurcubPCwsKswSIxMVETJ05U3rx5NX/+fHl7e0u6ffpWt27dTNcBAAAAZFauDxYzZ87UzJkzrY8dHBxUv359jRw5UhaLxRoqkpKSFBMTo6SkJNWoUUOSdOTIEbsEiztDRUxMjBISEuTo6KhKlSrpyJEj1mknTpxQZGSkevbsaQ0VkuTu7q727dtrypQppmsBAAAAMiPXB4t27dqpSZMmslgsypMnj4oXLy4fHx/r9M2bN2vhwoU6efKkEhMTbZ4bHR1tlxrOnz+vadOmaffu3amu3GSxWKz/joiIkCSVKFEi1TzSagMAAAAelVwfLIoXL65atWqlOW3r1q16++23FRgYqGHDhqlw4cJycXFRcnKyXn31VRmGYXr5MTExCg0NVWxsrLp166aAgAB5eHjIYrFo3rx52rdvn+llAAAAAA9brg8W97Ju3Tq5urpq5syZNuMszp49a7dl7N27V5GRkXrvvffUpk0bm2lffvmlzWM/Pz9J0rlz51LNJ602AAAA4FFxyOoCsjMHh9ubJzk52dpmGIbmzp1rt2Wk3N377qMfu3fvthlfIUkVKlRQgQIFtGbNGpvTsGJiYrR8+XK71QQAAAA8KI5Y3MMzzzyjrVu3auDAgWrZsqUSExP1ww8/KC4uzm7LqFq1qvLnz69JkybpwoULKlSokE6dOqV169YpICBAv/32m7Wvk5OTXnvtNY0aNUq9evVS27Zt5ejoqNWrV8vHx0cRERE2YzIkae3atbpw4YIkKSoqSrdu3dKcOXMkSUWLFlXLli3tti4AAADIvQgW99CsWTPFxMRo8eLFmjx5sry8vFS/fn0NHjxYzzzzjF2W4eXlpalTp+qLL77QkiVLlJSUpPLly2vy5MlatWqVTbCQpObNm8vJyUlz5szRzJkzlS9fPrVt21ZlypTR8OHD5erqatN/1apV+vnnn23aZsyYIUkKDg4mWAAAAMAuLIY9RiAjyy1cuFCTJk1SWFiYgoKCHsoyLOMS799JkjG8YwY6rTRXDAAAALIVxlg8Zm7duqWkpCSbtpiYGC1dulQ+Pj4qX758FlUGAACA3IxToezg6tWrqb7s383d3V3u7u6mlxUREaEhQ4aoadOm8vPz06VLl7R27VpFRERoxIgRcnZ2Nr0MAAAA4EERLOygZ8+e1gHS6QkNDdWAAQNML8vX11eVKlXS+vXrdfXqVTk6OiogIECDBw+2y13AAQAAgMxgjIUdHDx4UPHx8ffsU6xYMfn7+z+iih4OxlgAAAAgPRyxsIOqVatmdQkAAABAlmLwNgAAAADTOBUKGTZr1iz16dOHAeIAAABIhSMWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0i2EYRlYXgceDZVziffsYwzvef0bGSvPFAAAAIFvhiAUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADCNYPGIhYSEaPTo0VldBgAAAGBXTlldQHYVEhKS4b7//e9/5efn9xCryZxffvlFCxYs0KlTp3TlyhVJUpEiRdSkSRN1795dnp6eWVwhAAAAcgqCRTo+/PBDm8cHDhzQihUr1K5dO1WrVs1mWt68eR9laRl27tw5xcXF6bnnnlOBAgVkGIaOHj2qr776St9//73mz58vNze3rC4TAAAAOQDBIh0tWrSweZyUlKQVK1aocuXKqaZlV61atVKrVq1s2jp27KhSpUrpiy++0I8//qhnn302i6oDAABATkKwyAY2bdqk9evXW09Zcnd3V9WqVTVw4ECVKVMmVf9ly5Zp8eLFunDhgooUKaKuXbvK3d1dH3zwgWbMmHHf07iKFi0qSYqOjn4o6wMAAIDch2CRDXz77bfy8fFRu3btVKBAAZ0/f14rVqxQv379tHDhQhUvXtzad968eZo6darKly+vV155RXFxcVqwYME9T8eKi4uz/nf8+HFNmTJFzs7OqlWr1qNYPQAAAOQCBItsYMqUKcqTJ49NW8uWLdW9e3ctXrxYI0aMkCRdu3ZNs2fPVkBAgObOnStXV1dJ0vPPP68OHTqkO/8ZM2Zo4cKF1sdPPvmkJk6cKH9//4ewNgAAAMiNCBbZQEqoMAxDN2/eVGJiovLmzasSJUroyJEj1n579uxRfHy8OnbsaA0VklSgQAE999xzWrZsWZrzb9++vZ566ildv35dv/zyi8LDwxUVFfVQ1wkAAAC5C8EiGzhx4oRmzJih8PBwxcbG2kwrVqyY9d9//fWXJKlEiRKp5pFWW4rixYtbT6dq0qSJfvrpJ7366quSpObNm5uuHwAAACBYZLG///5b/fv3l4eHh/r166eSJUvKzc1NFotF48ePTxU07OGpp55S/vz5tWzZMoIFAAAA7IJgkcW2bdummJgYTZgwIdXVnK5duyYXFxfr45SrOZ07d041atSw6Xvu3LkHWm58fDxXhQIAAIDdOGR1Abmdg8Ptl8AwDJv2FStW6PLlyzZttWrVkouLi5YtW6b4+Hhr+6VLl7R+/fpU87506VKay1yzZo1u3LihSpUqmS0fAAAAkMQRiyxXt25dTZkyRe+99546d+4sLy8vHTp0SLt27ZK/v7+SkpKsfX19fRUaGqpp06apX79+eu655xQXF6cVK1aoRIkSOnbsmCwWi7X/0KFD5ePjo8qVK6tIkSK6ceOGDh48qB9++EGFCxdW//79s2KVAQAAkAMRLLKYv7+/vvjiC02bNk1hYWFycHBQlSpVNHPmTI0ZM0YXLlyw6d+nTx95eHjom2++0dSpU1WkSBH16NFDhmHo2LFjNleLateunbZu3aqVK1cqKipKTk5O8vf3V69evfTiiy/K19f3Ea8tAAAAciqLcfc5OHgsjRkzRt9++602bNigAgUKPJRlWMYl3rePMbzj/WdkrDRfDAAAALIVxlg8Zu4cW5Hi0qVLWrt2rUqXLv3QQgUAAABwL5wKZQfpDZK+k6enp9zc3EwvKzw8XJMnT1bjxo1VqFAh/fXXX1q5cqViY2Ot96YAAAAAHjWChR1k5F4Q77//vlq3bm16WU888YT8/f21YsUK6+VoK1asqN69e6tWrVqm5w8AAABkBmMs7GDPnj337ZMTTlNijAUAAADSwxELO+BIAQAAAHI7Bm8DAAAAMI1ToZBhs2bNUp8+feTs7JzVpQAAACCb4YgFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAAAAADDNYhiGkdVF4PFgGZd43z7G8I73n5Gx0nwxAAAAyFY4YgEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAExzyuoCIEVERGjChAk6dOiQoqKi1KpVK40YMUJTpkzR9u3bFRkZqcKFC2v16tVZXSoAAACQJoJFNvDBBx/o119/Vd++fZU/f375+/tr/vz5WrJkiXr06KGAgAB5eHhIko4cOaL169fr+PHj+vXXXxUbG6v3339frVu3TnPeCQkJ+uqrr7Ru3TpFRkaqUKFCat26tXr37i0nJ15+AAAA2AffLLNYQkKCDhw4oM6dO6tHjx7W9kmTJikgIEBDhw616b9z504tXbpUJUuWVJkyZXT48OF7zv/tt9/WDz/8oDZt2qhy5co6fPiwZsyYofPnz2v06NEPY5UAAACQCxEsstiVK1dkGIa8vb1t2i9fvqzChQun6t+xY0f17NlTefLk0ZYtW+4ZLHbs2KEffvhBL7zwgl5//XVJ0vPPPy8vLy8tWrRI7dq1U5UqVey7QgAAAMiVGLydhUaPHq1WrVpJkmbPnq2QkBDrfxEREfr555+tj2fOnClJyp8/v/LkyZOh+W/cuFGS1K1bN5v2lMfr16+316oAAAAgl+OIRRZq3769ypYtqwkTJqhRo0Zq1KiRIiIiVLRoUU2aNEm+vr7q27evJKlMmTIPPP+jR4+qUKFCKlKkiE17kSJFVLBgQR07dswu6wEAAAAQLLJQ5cqVVaBAAU2YMEEBAQFq0aKFddrMmTOVL18+m7YHdenSJZUqVSrNaQULFtTFixczPW8AAADgTpwKlYPFxcXJxcUlzWmurq6Ki4t7xBUBAAAgpyJY5GBubm5KSEhIc1p8fLzc3NwecUUAAADIqQgWOViBAgUUGRmZ5rSUe1oAAAAA9kCwyMECAwN18eJF/f333zbtf//9tyIjI1WhQoUsqgwAAAA5DcEiB2vWrJkk6T//+Y9Ne8rj55577pHXBAAAgJyJq0I9Zi5cuKC1a9dKkk6fPi1J+t///qd//vlHktSyZUsVLVpUklSvXj09/fTTWrRokW7cuKGgoCD98ssvWrVqlZ577jlVrVo1S9YBAAAAOQ/B4jETERGhGTNm2LRt27ZN27ZtkyRVrVrVGiwk6fPPP9fcuXO1fv16rVu3ToUKFdLAgQPVu3fvR1k2AAAAcjiLYRhGVheBx4NlXOJ9+xjDO95/RsZK88UAAAAgW2GMBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABM43KzyLBZs2apT58+cnZ2zupSAAAAkM1wxAIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmGYxDMPI6iLweLCMS7xvH2N4x/vPyFhpvhgAAABkKxyxAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsMiGdu3apZCQEE2fPj3VtF9++UUhISF66qmnFBcXl2r6q6++qho1aigqKkrh4eH697//rS5duqhBgwZq0qSJ+vbtqw0bNoj7IgIAAMCenLK6AKRWtWpVOTo6Kjw8PNW0/fv3y9HRUbdu3dKhQ4dUq1Yt67TExEQdPHhQpUuXlq+vr6ZMmaKLFy+qYcOGCggIUGxsrDZv3qxRo0Zp//79GjVq1KNcLQAAAORgBItsyN3dXYGBgTp27Jji4uLk5uZmnRYeHq5atWrp5MmT1n+nOHbsmGJjY1W9enVJt49epISUFN26ddPAgQO1cuVKde3aVQEBAY9uxQAAAJBjcSpUNlW9enXdunVLBw8etLYlJibq0KFDCg4OVnBwsPbv32/znJQjHCEhIdZ53BkqJMnBwUGNGzeWJP3+++8PcQ0AAACQmxAssqmUcHDn6VB3HpGoXr269XGK8PBwWSwWBQcH33PeFy9elCTlz5//IVQOAACA3IhgkU1VqVJFzs7ONsEiPDxc7u7uqlChgqpXr249giH9/9GMMmXKyMfHJ935RkZGasWKFSpWrJiqVq36sFcDAAAAuQTBIptyc3NTpUqVbI5KhIeHq3LlynJyclKpUqWUL18+a/C4e3xFWuLi4jRs2DDFxMRo9OjRcnJiiA0AAADsg2CRjYWEhFiv9HTn+IoU1apVs46zSAkY6QWL+Ph4vfHGGzp+/Ljef/99VatW7eGvAAAAAHINgkU2lhISwsPDrUck7gwWwcHBOnbsmGJiYhQeHi4HB4c0x1fEx8dr2LBh2rt3r0aNGqUWLVo8snUAAABA7sC5MNlY5cqV5erqqv3798vDw0Ourq4KDAy0Tq9evbqSkpIUHh6uQ4cOqWzZsvL29raZR0qo2L17t0aOHKk2bdo86tUAAABALkCwyMZcXFwUFBSkAwcOyMHBQZUrV5azs7N1eunSpeXj46MFCxakOb4iISFBw4cP1+7du/X222/r+eeff8RrAAAAgNyCYJHNhYSEaP/+/Tp8+LAGDBhgM81isahatWravn27te+dRo0apV27dqlmzZpyc3PTunXrbKaXKVNGZcqUeaj1AwAAIHcgWGRzd4aFtMZPBAcHa/v27XJ0dEw1IPv48eOSpL1792rv3r2pnhsaGkqwAAAAgF1YDMMwsroIPB4s4xLv28cY3vH+MzJWmi8GAAAA2QpXhQIAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAawQIAAACAaVxuFhk2a9Ys9enTx+bu3wAAAIDEEQsAAAAAdkCwAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkWwzCMrC4CjwfLuMT79jGGd7xPh5X2KQYAAADZCkcsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAawQIAAACAaQQLAAAAAKYRLLKhXbt2KSQkRNOnT0817ZdfflFISIieeuopxcXFpZr+6quvqkaNGoqKitKlS5c0bdo0vfrqq2rSpIlCQkI0evToR7AGAAAAyG0IFtlQ1apV5ejoqPDw8FTT9u/fL0dHR926dUuHDh2ymZaYmKiDBw+qdOnS8vX11dmzZxUWFqbTp0+rYsWKj6p8AAAA5EJOWV0AUnN3d1dgYKCOHTumuLg4ubm5WaeFh4erVq1aOnnypPXfKY4dO6bY2FhVr15dklShQgVt3rxZefPmVVRUlJo0afLI1wUAAAC5A0cssqnq1avr1q1bOnjwoLUtMTFRhw4dUnBwsIKDg7V//36b56Qc4QgJCZEkeXh4KG/evI+sZgAAAOReBItsKiUc3Hk61J1HJKpXr259nCI8PFwWi0XBwcGPvF4AAADkbgSLbKpKlSpydna2CRbh4eFyd3dXhQoVVL16desRDOn/j2aUKVNGPj4+WVU2AAAAcimCRTbl5uamSpUq2RyVCA8PV+XKleXk5KRSpUopX7581uBx9/gKAAAA4FEiWGRjISEh1is93Tm+IkW1atWs4yxSAgbBAgAAAFmBYJGNpYSE8PBw6xGJO4NFcHCwjh07ppiYGIWHh8vBwYHxFQAAAMgSXG42G6tcubJcXV21f/9+eXh4yNXVVYGBgdbp1atXV1JSksLDw3Xo0CGVLVtW3t7eWVgxAAAAciuCRTbm4uKioKAgHThwQA4ODqpcubKcnZ2t00uXLi0fHx8tWLCA8RUAAADIUgSLbC4kJET79+/X4cOHNWDAAJtpFotF1apV0/bt26197zZnzhxJUnx8vCTp119/tbal3A8DAAAAMItgkc3dGRbSCgHBwcHavn27HB0dVa1atVTTZ8yYYfP45MmTOnnypCQpNDSUYAEAAAC7sBiGYWR1EXg8WMYl3rePMbzjfTqstE8xAAAAyFa4KhQAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATONys8iwWbNmqU+fPjZ3/wYAAAAkjlgAAAAAsAOCBQAAAADTCBYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEyzGIZhZHUReDxYxiXet48xvON9Oqy0TzEAAADIVjhiAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATHPK6gLwYCIiIjRhwgQdOnRIUVFRatWqlUaMGKEpU6Zo+/btioyMVOHChbV69eqsLhUAAAC5CMHiMfPBBx/o119/Vd++fZU/f375+/tr/vz5WrJkiXr06KGAgAB5eHhIkhYuXKj//e9/OnfunKKjo+Xt7a2SJUuqa9euatSoURavCQAAAHISgsVjJCEhQQcOHFDnzp3Vo0cPa/ukSZMUEBCgoUOH2vQ/evSo/Pz8VLduXfn6+io6OlpbtmzR8OHDNXDgQL300kuPehUAAACQQxEsHiNXrlyRYRjy9va2ab98+bIKFy6cqv9nn32Wqq1bt27q0aOHvv76a/Xp00eOjo4PrV4AAADkHgSLx8To0aO1Zs0aSdLs2bM1e/Zsm+kREREKCQmRJIWGhmrAgAFpzsfJyUkFCxbUb7/9psTERIIFAAAA7IJg8Zho3769ypYtqwkTJqhRo0Zq1KiRIiIiVLRoUU2aNEm+vr7q27evJKlMmTI2z7127ZqSk5MVFRWlLVu26KefflJISIhcXV2zYlUAAACQAxEsHhOVK1dWgQIFNGHCBAUEBKhFixbWaTNnzlS+fPls2u7Uvn17Xbt2TZLk6Oioxo0ba8SIEY+kbgAAAOQOBItcYOzYsUpISNDFixe1ZcsWxcfH6+bNm8qbN29WlwYAAIAcghvk5QLBwcGqXbu22rRpoy+++ELu7u7q16+foqOjs7o0AAAA5BAEi1yoVatWunz5srZu3ZrVpQAAACCHIFjkQvHx8ZLEEQsAAADYDcEih4qNjVVMTEyq9qSkJC1dulSSFBQU9KjLAgAAQA7F4O0c6o8//lD//v31zDPPqESJEvLx8dHFixe1ceNGnTt3Tq1atVK1atWyukwAAADkEASLHKpw4cJq0aKFDh48qO3bt+vmzZvy9PRUuXLl9NJLL6l58+ZZXSIAAAByEIthGEZWF4HHg2Vc4n37GMM73qfDSvsUAwAAgGyFMRYAAAAATCNYAAAAADCNYAEAAADANIIFAAAAANMIFgAAAABMI1gAAAAAMI3LzSLDZs2apT59+sjZ2TmrSwEAAEA2wxELAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGCaxTAMI6uLwOPBMi7xntON4R3vPQNjpf2KAQAAQLbCEQsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAaweIRa926tfr375/VZQAAAAB2lSODxa5duxQSEqLp06enmvbLL78oJCRETz31lOLi4lJNf/XVV1WjRg1FRUUpPDxc//73v9WlSxc1aNBATZo0Ud++fbVhwwY9LvcVTE5O1qJFi9ShQwfVqVNHLVu21MSJExUbG5vVpQEAACAHccrqAh6GqlWrytHRUeHh4amm7d+/X46Ojrp165YOHTqkWrVqWaclJibq4MGDKl26tHx9fTVlyhRdvHhRDRs2VEBAgGJjY7V582aNGjVK+/fv16hRox7lamXKhAkT9M0336hRo0Z68cUXdebMGX3zzTc6efKkpk+fLgeHHJktAQAA8IjlyGDh7u6uwMBAHTt2THFxcXJzc7NOCw8PV61atXTy5Enrv1McO3ZMsbGxql69uqTbRy9SQkqKbt26aeDAgVq5cqW6du2qgICAR7diD+j333/XkiVL1KhRI40dO9ba7ufnp3HjxmnTpk1q3rx5FlYIAACAnCLH/lxdvXp13bp1SwcPHrS2JSYm6tChQwoODlZwcLD2799v85yUIxwhISHWedwZKiTJwcFBjRs3lnT7i7s97N69W2+//bbatm2runXrqmHDhnrllVfSPOIiSd9//726detmPbVp1qxZ2rNnj0JCQrR69Wprv40bN8owDHXv3t3m+e3atZObm5vWrVtnl/oBAACAHBssUsLBnV/O7zwiUb16devjFOHh4bJYLAoODr7nvC9evChJyp8/v11qXb16ta5du6YWLVpo+PDh6t69u86ePatBgwbpwIEDNn03bdqkESNGKC4uTqGhoerSpYu2bNmiqVOnpprvsWPH5ODgoMDAQJt2V1dXlS1bVseOHbNL/QAAAECOPBVKkqpUqSJnZ2ebYBEeHi53d3dVqFBBnp6e1iMYtWvXtv67TJky8vHxSXe+kZGRWrFihYoVK6aqVavapdZRo0YpT548Nm0dOnRQ586dFRYWpmrVqkm6fcRl4sSJyps3r+bPny9vb29JUseOHdWtW7c0a/X19ZWLi0uqaYUKFdLhw4d169YtOTs722U9AAAAkHvl2CMWbm5uqlSpks1RifDwcFWuXFlOTk4qVaqU8uXLZw0ed4+vSEtcXJyGDRummJgYjR49Wk5O9slld4aKmJgYRUVFydHRUZUqVdLRo0et006cOKHIyEi1atXKGiqk22NK2rdvn2a96YWGlLCR1pWxAAAAgAeVY49YSLdPhzpw4IAOHjyoGjVq6NChQ+rTp491erVq1azjLFICRnrBIj4+Xm+88YaOHz+u0aNHW48i2MP58+c1bdo07d69W9evX7eZZrFYrP+OiIiQJJUoUSLVPNJqc3Nz09WrV9NcZkJCgrUPAAAAYFaODhbVq1fX7NmzFR4eLg8PD8XGxtqMnwgODtaECRMUExOj8PBwOTg4pDm+Ij4+XsOGDdPevXv17rvvqkWLFnarMSYmRqGhoYqNjVW3bt0UEBAgDw8PWSwWzZs3T/v27cv0vAsWLKgzZ84oISEh1elQFy9elK+vL6dBAQAAwC5ydLCoXLmyXF1dtX//fnl4eMjV1dVmIHP16tWVlJSk8PBwHTp0SGXLlrU5xUj6/1Cxe/dujRw5Um3atLFrjXv37lVkZKTee++9VPP+8ssvbR77+flJks6dO5dqPmm1VaxYUbt379bRo0dtjrDEx8fr1KlT9x2kDgAAAGRUjh1jId0eRxAUFKTjx4/rxx9/VOXKlW1+oS9durR8fHy0YMGCNMdXJCQkaPjw4dbLwT7//PN2rzHlcrZ338l79+7dOnLkiE1bhQoVVKBAAa1Zs0bR0dHW9piYGC1fvjzVvJs2bSqLxaLFixfbtK9YsUJxcXHcwwIAAAB2k6OPWEi3x1ns379fhw8f1oABA2ymWSwWVatWTdu3b7f2vdOoUaO0a9cu1axZM837PpQpU0ZlypQxVV/VqlWVP39+TZo0SRcuXFChQoV06tQprVu3TgEBAfrtt9+sfZ2cnPTaa69p1KhR6tWrl9q2bStHR0etXr1aPj4+ioiIsBmTERAQoE6dOunbb7/V8OHDVbduXeudt4ODgwkWAAAAsJtcESxSpHXqT3BwsLZv3y5HR8dUA7KPHz8u6fbpSnv37k313NDQUNPBwsvLS1OnTtUXX3yhJUuWKCkpSeXLl9fkyZO1atUqm2AhSc2bN5eTk5PmzJmjmTNnKl++fGrbtq3KlCmj4cOHy9XV1ab/G2+8IT8/Py1fvlw7duyQr6+vunTpooEDB8rBIUcfsAIAAMAjZDHuPgcHj6WFCxdq0qRJCgsLU1BQ0ENZhmVc4j2nG8M73nsGxkr7FQMAAIBshZ+sHzO3bt1SUlKSTVtMTIyWLl0qHx8flS9fPosqAwAAQG6W40+FehSuXr2a6sv+3dzd3eXu7m56WRERERoyZIiaNm0qPz8/Xbp0SWvXrlVERIRGjBjB5WMBAACQJQgWdtCzZ09duHDhnn1CQ0NTDR7PDF9fX1WqVEnr16/X1atX5ejoqICAAA0ePFjPPvus6fkDAAAAmcEYCzs4ePCg4uPj79mnWLFi8vf3f0QVPRyMsQAAAEB6OGJhB1WrVs3qEgAAAIAsxeBtAAAAAKZxKhQybNasWerTpw8DxAEAAJAKRywAAAAAmEawAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAgA0qWLKnevXtndRnZFsECAAAAudrvv/+uAQMG6Mknn5Sbm5u8vb1Vt25dTZ48WbGxsVld3n3Fx8frrbfekp+fn/LkyaNatWpp8+bNj7wOp0e+RAAAACCbWLt2rTp16iRXV1f17NlTlSpVUkJCgnbs2KHhw4fr6NGjmjVrVlaXeU+9e/fWsmXL9Nprr6lMmTKaN2+eWrRooW3btqlevXqPrA6CBQAAADLNMi4xq0uQMSxzX2nPnDmjrl27qkSJEtq6dauKFi1qnfbKK6/ot99+09q1a+1V5kOxd+9effPNNxo7dqyGDRsmSdaA9Oabb2rXrl2PrBZOhUKGDYjuK5fJFlnGJab5nyzPp/8fAABANjNmzBjduHFDc+fOtQkVKQICAjR06NB0n3/lyhUNGzZMQUFB8vT0lLe3t5577jkdOnQoVd8pU6YoMDBQ7u7uyps3r0JCQrR48WLr9OvXr+u1115TyZIl5erqqkKFCunZZ5/Vzz//fM91WLZsmRwdHdW/f39rm5ubm/r166effvpJf/75Z0Y2hV1wxAIAAAC50urVq/Xkk0+qTp06mXr+6dOntXLlSnXq1EmlSpXSP//8o5kzZ6pBgwY6duyY/Pz8JEmzZ8/WkCFD1LFjRw0dOlRxcXE6fPiw9uzZo+7du0uSBg4cqGXLlmnw4MGqWLGiLl++rB07duj48eMKDg5Ot4YDBw6obNmy8vb2tmmvWbOmJOngwYN64oknMrV+D4pgAQAAgFwnOjpaERERatu2babnERQUpFOnTsnB4f9PAurRo4fKly+vuXPn6t1335V0exxHYGCgli5dmu681q5dq9DQUI0fP97a9uabb963hgsXLqR5tCWl7a+//srw+pjFqVAAAADIdaKjoyVJXl5emZ6Hq6urNVQkJSXp8uXL8vT0VLly5WxOYfL19dX58+e1b9++dOfl6+urPXv2PHAQiI2Nlaura6p2Nzc36/RHhWABAACAXCfl1KHr169neh7JycmaOHGiypQpI1dXVxUoUEAFCxbU4cOHde3aNWu/t956S56enqpZs6bKlCmjV155RTt37rSZ15gxY3TkyBE98cQTqlmzpkaPHq3Tp0/ft4Y8efIoPj4+VXtcXJx1+qNCsAAAAECu4+3tLT8/Px05ciTT8/j000/1r3/9S/Xr19fChQu1ceNGbd68WYGBgUpOTrb2q1Chgk6ePKlvvvlG9erV03fffad69erp/ffft/bp3LmzTp8+rSlTpsjPz09jx45VYGCg1q9ff88aihYtqgsXLqRqT2lLGefxKBAsAAAAkCu1atVKv//+u3766adMPX/ZsmVq1KiR5s6dq65du6pp06Zq0qSJoqKiUvX18PBQly5dFBYWpj/++EMtW7bUJ598Yj2yIN0OCYMGDdLKlSt15swZ5c+fX5988sk9a6hatapOnTplPbUrxZ49e6zTHxWCBQAAAHKlN998Ux4eHnrppZf0zz//pJr++++/a/Lkyek+39HRUYZh2LQtXbpUERERNm2XL1+2eezi4qKKFSvKMAzdunVLSUlJNqdOSVKhQoXk5+eX5mlOd+rYsaOSkpJsbuIXHx+vsLAw1apV65FdEUriqlCZ1r9/f124cEGrV69+oOeNHj1aa9as0f79+x9SZQAAAMiI0qVLa/HixerSpYsqVKhgc+ftXbt2aenSperdu3e6z2/VqpU+/PBD9enTR3Xq1NEvv/yiRYsW6cknn7Tp17RpUxUpUkR169ZV4cKFdfz4cU2dOlUtW7aUl5eXoqKi5O/vr44dO6pKlSry9PTUli1btG/fPpurRKWlVq1a6tSpk95++21dvHhRAQEBmj9/vs6ePau5c+faYzNlWK4KFiEhIRnu+9///veRnpP2sBiGoY0bN+rbb7/VuXPndOvWLRUpUkTPPvusunXrJk9Pz6wuEQAAIMu0adNGhw8f1tixY7Vq1Sp9+eWXcnV1VeXKlTV+/HiFhoam+9x33nlHN2/e1OLFi7VkyRIFBwdr7dq1GjFihE2/AQMGaNGiRZowYYJu3Lghf39/DRkyRKNGjZIkubu7a9CgQdq0aZOWL1+u5ORkBQQEaPr06Xr55Zfvuw5ff/213n33XS1YsEBXr15V5cqVtWbNGtWvX9/cxnlAFuPu4zc52Lp162weHzhwQCtWrFC7du1UrVo1m2mNGjW65yj6x+WIxbRp0xQWFqYaNWqoQYMGcnJyUnh4uDZv3qxKlSopLCxMFoslQ/OyjEu853RjeMd7TFz5AFUDAADgcZOrjli0aNHC5nFSUpJWrFihypUrp5qWEyQmJuo///mPypcvr2nTplmvs9yxY0c5OTlp/fr1OnXqlMqVK5fFlQIAAOBxl6uCRWZER0friy++0LZt2xQfH6+KFSvq9ddft+syzp49q2+++UY///yz/v77byUlJalUqVLq2LGjnn/++VT9f/31V02aNEmHDh2Sq6ur6tWrp9dff11NmjRRq1atNHr0aEm3g0V8fLzy589vc0dISSpQoICkR3ttYwAAAORcBIt7SExM1ODBg3Xs2DG1aNHCetv2QYMGycfHx27L2b9/v37++WfVq1dPfn5+iouL05YtW/Txxx/r6tWr6tOnj7XvH3/8oZdeekmGYahr164qWLCgdu7cqVdffTXVfN3c3FStWjX99NNPmjdvnp555hk5OjoqPDxcy5Yt03PPPafixYvbbT0AAACQexEs7uG///2vjh07ptDQUA0YMMDaXqpUKU2YMEFFixa1y3Jatmypjh1txyd0795dAwcO1Lx589SjRw85Od1+qaZPn66bN29qzpw51usSd+nSRW+//baOHz+eat4ff/yxRo8eralTp2rq1KmSJIvFor59+2rgwIF2qR8AAADgPhb3sH37djk6OuqFF16wae/YsaM8PDzstpw7T0eKj49XVFSUoqOjVbt2bd28eVNnz56VdHtMyM6dOxUYGJjqZid315jCxcVFxYoVs96E5ZNPPlHjxo01d+5cffXVV3ZbBwAAAORuHLG4h4iICBUoUCDVJVlTvqxfv37dLsuJiYnRrFmztHnz5jRvzpJyJ8WrV68qNjZWJUqUSNWnZMmSqdri4uLUt29flStXTp999pm1vVmzZnr77bc1c+ZMPfPMM2k+FwAAAHgQBItsYOTIkdqxY4fatWun4OBg+fj4yMHBQTt37tTixYuVnJycqflu2bJFf/zxh1555ZVU05o0aaLNmzfr4MGDBAsAAACYRrC4h2LFimnPnj26ceOGzVGLhIQERUREyNvb2/Qyrl+/rh07dqhFixZ65513bKbt3bvX5nHevHmVJ08enTt3LtV8Uk6XulNkZKQkpRlMkpKSbP4PAAAAmMEYi3to0KCBkpKStGjRIpv2ZcuW6ebNm3ZZRsplYO++T+GlS5e0cuVKmzZHR0fVqVNHR48e1cGDB22m3V2jdHuQuSStWbMm1bSUtsDAwMyWDgAAAFhxxOIe2rRpoxUrVmj27NmKiIhQ5cqVdfLkSW3ZskX+/v52+bXfw8NDtWvX1vr16+Xq6qrAwEBduHBBy5cvV7FixXTt2jWb/i+//LJ2796tIUOGqHPnzipUqJB27NihqKgoSbK5i/bTTz+twMBA7dy5U6GhoWrUqJEkadu2bTpw4ICaNGmi8uXLm14HAAAAgCMW9+Ds7Kxp06apTZs22rlzpyZNmqRz585p2rRpKlSokN2W89FHH6lNmzb68ccfNWbMGG3fvl2DBg1Sp06dUvUtWbKkZs2apUqVKuk///mPZsyYIV9fX+vgbFdXV2tfR0dHTZ8+XX369NGVK1c0ZcoUTZ06VdHR0Xr11Vf18ccf220dAAAAkLtZjLvPwcFj6fjx4+rRo4cGDx6s3r17P5RlWMYl3nO6MbzjPSautG8xAAAAyFY4YvEYiouLs3lsGIa+/vprSVKtWrWyoiQAAIAcr2TJkg/tB9ycgDEWdnDjxo1UX/bv5uzsLB8fH7ssr3v37qpRo4YCAgIUGxurH3/8UQcOHNCzzz6rChUq2GUZAAAAucXvv/+uMWPGaPPmzfrrr7/k4uKioKAgde7cWf3797e5mXF2c+PGDY0dO1Z79uzR3r17dfXqVYWFhWVJACJY2MG4cePSvPLSnYKDgzVr1iy7LK9Bgwb68ccftW7dOiUlJcnPz08DBw4kQQMAgEfP8nxWV2DqlOu1a9eqU6dOcnV1Vc+ePVWpUiUlJCRox44dGj58uI4ePWq373APw6VLl/Thhx+qePHiqlKlirZv355ltRAs7KBnz5567rnn7tnHHve8SDF06FANHTrUbvMDAADIjc6cOaOuXbuqRIkS2rp1q4oWLWqd9sorr+i3337T2rVrs7DC+ytatKguXLigIkWKaP/+/apRo0aW1cIYCzt48sknVatWrXv+xylKAAAA2cuYMWN048YNzZ071yZUpAgICLjnj7lXrlzRsGHDFBQUJE9PT3l7e+u5557ToUOHUvWdMmWKAgMD5e7urrx58yokJESLFy+2Tr9+/bpee+01lSxZUq6uripUqJCeffZZ/fzzz/dcB1dXVxUpUuQB1vrh4YgFAAAAcqXVq1frySefVJ06dTL1/NOnT2vlypXq1KmTSpUqpX/++UczZ85UgwYNdOzYMfn5+UmSZs+erSFDhqhjx44aOnSo4uLidPjwYe3Zs0fdu3eXJA0cOFDLli3T4MGDVbFiRV2+fFk7duzQ8ePHFRwcbLd1fpgIFsiwmd5fqU+fPnJ2dk67w7CVj7QeAACAzIqOjlZERITatm2b6XkEBQXp1KlTcnD4/5OAevToofLly2vu3Ll69913Jd0exxEYGKilS5emO6+1a9cqNDRU48ePt7a9+eabma4tK3AqFAAAAHKd6OhoSZKXl1em5+Hq6moNFUlJSbp8+bI8PT1Vrlw5m1OYfH19df78ee3bty/defn6+mrPnj3666+/Ml1PViNYAAAAINdJubDO9evXMz2P5ORkTZw4UWXKlJGrq6sKFCigggUL6vDhw7p27Zq131tvvSVPT0/VrFlTZcqU0SuvvKKdO3fazGvMmDE6cuSInnjiCdWsWVOjR4/W6dOnM11bViBYAAAAINfx9vaWn5+fjhw5kul5fPrpp/rXv/6l+vXra+HChdq4caM2b96swMBAJScnW/tVqFBBJ0+e1DfffKN69erpu+++U7169fT+++9b+3Tu3FmnT5/WlClT5Ofnp7FjxyowMFDr1683tZ6PEsECAAAAuVKrVq30+++/66effsrU85ctW6ZGjRpp7ty56tq1q5o2baomTZooKioqVV8PDw916dJFYWFh+uOPP9SyZUt98sknNjdZLlq0qAYNGqSVK1fqzJkzyp8/vz755JPMrt4jR7AAAABArvTmm2/Kw8NDL730kv75559U03///XdNnjw53ec7OjrKMAybtqVLlyoiIsKm7fLlyzaPXVxcVLFiRRmGoVu3bikpKcnm1ClJKlSokPz8/BQfH/+gq5VluCoUAAAAcqXSpUtr8eLF6tKliypUqGBz5+1du3Zp6dKl6t27d7rPb9WqlT788EP16dNHderU0S+//KJFixbpySeftOnXtGlTFSlSRHXr1lXhwoV1/PhxTZ06VS1btpSXl5eioqLk7++vjh07qkqVKvL09NSWLVu0b98+m6tEpWfq1KmKioqyDvxevXq1zp8/L0l69dVX5ePjk/mN9AAsxt0xC0jHrFmz7n25WQAAkPtYns/qCiRjpamn//rrrxo7dqw2b96sv/76S66urqpcubK6du2q0NBQubq6SpJKliyphg0bat68eZKk+Ph4jRw5UosXL1ZUVJSCg4M1btw4jRgxQpK0fft2Sbe/Qy1atEhHjx7VjRs35O/vr/bt22vUqFHy9vZWQkKCRo0apU2bNun06dNKTk5WQECABgwYoJdffvm+9ZcsWVLnzp1Lc9qZM2dUsmRJU9snowgWyDCCBQAAANLDGAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmESwAAAAAmEawAAAAAGAawQIAAACAaQQLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACmOWV1AXg8GIah2NhYRUdHy9nZOavLAQAAwCPk5eUli8Vyzz4WwzCMR1QPHmOXLl1SwYIFs7oMAAAAZIFr167J29v7nn04YoEMcXV1VdWqVbV27Vp5enpmdTmPjRs3bqhly5ZstwfANsscttuDY5tlDtvtwbHNMoft9uAe5jbz8vK6bx+CBTLEYrHI0dFR3t7evLkfgIODA9vtAbHNMoft9uDYZpnDdntwbLPMYbs9uKzeZgzeBgAAAGAawQIAAACAaQQLZIiLi4tCQ0Pl4uKS1aU8VthuD45tljlstwfHNsscttuDY5tlDtvtwWX1NuOqUAAAAABM44gFAAAAANMIFgAAAABM43Kz0NmzZzVmzBgdPnxYHh4eatGihQYNGnTfO2wbhqH58+dr6dKlioqKUtmyZfWvf/1LQUFBj6jyrJXZ7bZ06VLt3LlTR44cUVRUlD7//HM1adLkEVWdtTKzzS5duqRFixZpz549On/+vDw9PVWtWjUNHjxYRYsWfYTVZ53M7mvvvvuujhw5osjISDk7OysgIED9+vVT7dq1H1HlWSez2+xOixcv1oQJE1SvXj1NmjTp4RWbjWR2u7Vu3VoXLlxI1b5z5065uro+rHKzBTP72sWLFzVt2jTt3LlTsbGxKlq0qPr166fnnnvuEVSetTKz3fbv36+BAwemOa1EiRL67rvvHla52UJm97WoqChNnz5dO3fu1LVr1+Tn56fOnTurY8eOdq+RYJHLRUdHa+DAgSpevLjGjh2rixcvauLEiYqLi9Nbb711z+fOnz9fM2fO1ODBg1WmTBktXbpUgwcP1qJFi+Tv7/+I1iBrmNlua9eulSTVrVvX+u/cILPb7Pjx49q2bZvatGmjoKAgRUVFac6cOerVq5eWLFmivHnzPsK1ePTM7Gu3bt3SCy+8oCeeeEIJCQlatWqVhg4dqhkzZqhatWqPaA0ePTPbLMWlS5c0e/Zs5cuX7yFXm32Y3W7PPPOMXnzxRZu2nD7o1sw2u3Tpkvr06aMSJUpo5MiR8vDw0OnTp5WQkPCIqs86md1u5cuXV1hYmE3bzZs3NWTIENWpU+dhl52lzOxrI0aM0NmzZ/XKK6+oSJEi2rlzpz7//HM5OjqqXbt29i3UQK721VdfGfXq1TOioqKsbd99951Rs2ZN4+LFi+k+Ly4uzqhfv74xdepUa1tCQoLRqlUr47PPPnuoNWcHmd1uhmEYSUlJhmEYRkREhFG9enVj8+bND7XW7CKz2yw6Otq4deuWTdvff/9thISEGAsWLHho9WYXZva1uyUmJhotWrQwPv74Y3uXma3YY5u9++67xnvvvWeEhoYaQ4cOfUiVZi9mtlurVq2Mzz///GGXmO2Y2WajRo0y+vTpYyQmJj7sMrMde36u/fe//zWqV69uHDlyxN5lZiuZ3WaRkZFG9erVjf/+97827aGhocbAgQPtXidjLHK5Xbt2qWbNmvLx8bG2Pfvss0pOTtbu3bvTfd7hw4d18+ZNm1N4nJ2d1ahRI+3cufOh1pwdZHa7SbfvipkbZXabeXl5ycnJ9uBq4cKFlTdvXkVGRj60erMLM/va3RwdHeXl5aVbt27Zu8xsxew2O3jwoH744Qe9+uqrD7PMbMee+1pukdltduPGDW3ZskWdOnWSo6Pjoyg1W7HnvrZhwwYVL15cgYGB9i4zW8nsNktMTJSkVHfh9vDwkPEQLgybO7/hwOrs2bMqWbKkTZuXl5cKFCigs2fP3vN5klI9t1SpUvr7778VFxdn30Kzmcxut9zMntvs3LlzunLlikqVKmW/ArMps9vNMAwlJiYqKipKCxYs0J9//qn27ds/nGKzCTPbLCkpSWPGjFGfPn1UoECBh1dkNmR2X9uwYYOeeuopPf300xoyZIh+++23h1NoNpLZbXbixAndunVLTk5O6t+/v2rVqqVmzZrpiy++sH4RzMns9ffg8uXL2r9/v5o1a2bfArOhzG6zIkWKqHbt2goLC9Pp06d18+ZNbd68WXv27FGnTp3sXidjLHK56OhoeXl5pWr38vJSdHT0PZ/n4uKSalCel5eXDMPQ9evX5ebmZvd6s4vMbrfczF7bzDAMjRs3TgULFswVf0zMbrdVq1bp448/liS5u7vr008/VeXKle1eZ3ZiZpstXbpUsbGxeuGFFx5WedmWme1Wv359VapUSUWKFFFERIS++uor9evXL8ePucvsNrt8+bIk6eOPP9bzzz+v/v3768iRI5o5c6YcHBw0ePDgh1ZzdmCvvwebN29WUlKSmjdvbs/ysiUz22zs2LF6++231blzZ0m3j14PHz5czzzzjN3rJFgAeKzMmjVLe/fu1ZQpU5QnT56sLifba9iwocqWLauoqCht2bJFb7/9tsaOHau6detmdWnZzpUrVzRz5kx98MEHD3T1KEjDhw+3/rtatWqqXbu2OnTooIULF2rEiBFZWFn2lHIKSs2aNfX6669LkkJCQhQTE6OFCxfqpZdeytE/ztnL+vXrVaFCBZUoUSKrS8m2DMPQBx98oD/++EMff/yxChQooD179mj8+PHy8vKy+w90BItcztvbWzdu3EjVfv36dXl7e9/zeQkJCYqPj7c5anH9+nVZLJY0U3VOktntlpvZY5utWLFCs2fP1rvvvquaNWvau8Rsyex28/X1la+vrySpTp06io6O1uTJk3N0sMjsNpsxY4bKlCmjatWq6fr165JunxqVlJSk69evK0+ePKnG++Qk9vxcK1CggKpWrarjx4/bq7xsKbPbLOVvZEhIiE17zZo19dVXX+n8+fMKCAiwb7HZiD32tfPnz+vo0aPWYJbTZXab7dixQ1u2bNE333xj3adCQkJ09epVTZo0ye7BgjEWuVzJkiVTnZt348YNXbp0KdW5fHc/T7p9rvudzp49qyJFiuT4X1oyu91yM7PbbNu2bfr88881cOBAtW3b9uEUmQ3Ze18rX768zp8/b5/isqnMbrOzZ8/q559/VqNGjaz/HTp0SD/99JMaNWqkvXv3PtzCsxifaw8us9vsySefvOd84+Pj7VBd9mWPfW3Dhg1ycHDIFafESpnfZqdPn5ajo6NKly5t016uXDlFRkbafUwswSKXq1Onjvbu3Wv9dU6StmzZIgcHh3veRKty5cry8PDQli1brG2JiYnatm1bjv4lNEVmt1tuZmab7d+/XyNHjtTzzz+vl1566WGXmq3Ye187dOiQihUrZs8Ss53MbrM33nhDM2bMsPmvbNmyCgoK0owZM3L8VWfsua9FRkbq4MGDqlixor3LzFYyu82KFi2qgICAVGF1z549cnV1vW/weNzZY1/buHGjqlevnmsusmBmX0tKStKvv/5q0378+HHly5fP7j8E59xjusiQDh06aMmSJXrjjTfUt29fXbx4UZMnT1b79u1VsGBBa7+XX35ZFy5c0MqVKyVJrq6u6tOnj2bNmqW8efMqICBAS5cu1bVr11LdICknyux2k6Rjx47pr7/+UlRUlCTpyJEjkqS8efOqevXqj3I1HqnMbrMzZ85o2LBheuKJJ9SiRQv98ssv1r558+bN0QNDpcxvtx07dmjt2rWqV6+eChcurOjoaG3YsEE//fSTPvnkkyxam0cjs9usXLlyqebl6ekpd3f3VKes5ESZ3W4bNmzQjh07VLduXRUsWFDnz5/XvHnz5OjomOP/Hpj5WzBo0CC98cYbGj9+vOrWratjx45pwYIF6tmzZ44fP2Zmu0m3r6p15syZXHWRhcxus7p166pIkSJ66623FBoaqgIFCmj37t1as2aN+vfvb/c6CRa5nLe3t7788kuNHTtWb7zxhjw8PPT8889r0KBBNv1SzjO+U69evWQYhhYuXKirV6+qbNmymjJlSo7/oieZ227ffvut1qxZY328cOFCSVJwcLBmzZr18IvPIpndZkeOHNGNGzd048YN9evXz6Zvq1atNHr06EdRfpbJ7Hbz9/dXQkKCpk6dqqioKPn6+qpMmTKaOXNmjg6wkrn3Z26W2e1WrFgxRUZGavz48bp+/bq8vLxUo0YNDRgwIMcfHTOzr9WvX1+ffPKJ5syZo2XLlqlAgQIaMGCAevfu/QjXIGuYfY9u3LhRLi4uD+WqRtlVZreZh4eHvvzyS02fPl1TpkzR9evX5efnp9dff916lSh7shgP4+4YAAAAAHIVxlgAAAAAMI1gAQAAAMA0ggUAAAAA0wgWAAAAAEwjWAAAAAAwjWABAAAAwDSCBQAAAADTCBYAAAAATCNYAA/BxYsX5ePjo9mzZ9u09+7dWyVLlsyaonKI0aNHy2Kx6OzZs49kefPmzUu1vNjYWPn5+emDDz544Pmlt28g81Jeo+3bt2d1KchiZj8f2Jdyr7Nnz8pisWj06NGPdLnbt2+XxWLRvHnzMvX8gwcPysHBQT/88IN9C8skggXwEIwaNUoFCxZUnz59MtT/77//1rBhw1SpUiV5eXnJ29tbZcqUUdeuXbV8+XKbvg0bNpSnp2e680r5w7p///40p1+9elV58uSRxWLRggUL0p1PyZIlZbFYrP+5uLioZMmSeumll/Tnn39maL1yqjx58mjEiBEaO3asLly48EDPfdB9A7nbwYMHNXr06EcWpJH1zp49q9GjR+vgwYOPdLnsa6lFRUVp9OjR2TpoVq1aVc8//7zeeOMNGYaR1eUQLAB7O3/+vL766iu9+uqrcnJyum//c+fOqUqVKpo2bZpq166tzz//XJ999platWqlEydOKCwszK71LVq0SPHx8SpVqpS++uqre/b19/fXggULtGDBAk2ePFm1atXSV199pVq1aunSpUt2retx069fP1ksFk2YMCHDz3nQfQMZ06NHD8XGxqp+/fpZXYrdHTx4UB988AFf9nKRs2fP6oMPPsiSYJGb97USJUooNjZWo0aNsrZFRUXpgw8+yNbBQpJee+01hYeHa926dVldivjLBtjZzJkzZbFY1K1btwz1HzdunC5evKiVK1eqbdu2qab//fffdq1v7ty5atSokdq2bavXXntNp0+f1pNPPplmXx8fH7344ovWxy+//LIKFSqkqVOnKiwsTMOHD7drbY8TDw8PtW/fXvPmzdPHH38sV1fX+z7nQfeNrJaUlKT4+Hi5u7tndSn35OjoKEdHx6wuA8BjzGKxyM3NLavLyJSnn35aJUuW1IwZM9SyZcssrYUjFshyKee0fv/99/rwww9VokQJ5cmTR7Vq1dLu3bslST/88IPq1asnDw8PFS1aVB999FGa89q/f7/atWunAgUKyNXVVeXKldMnn3yixMREm3579+5V7969VbZsWbm7u8vLy0t169bVihUrUs2zd+/eslgsunbtmvWLtZubm+rWras9e/ak6r906VKFhISoUKFCGVr/X3/9VZL0zDPPpDm9SJEiGZpPRvz88886ePCgevXqpe7du8vJyem+Ry3u1qxZM0nSb7/9lm6f9evXy2Kx6Isvvkhz+lNPPaWCBQvq1q1bkh7s9UhLymuUFovFot69e6dqX7JkierVqycvLy+5u7urVq1aWrZsWYaWl+K5557TpUuXtG3btgz1T2/fSE5O1ieffKL69eurSJEicnFxUfHixfXyyy/r8uXL1n5RUVFyc3NT+/bt05z/22+/LYvFYvNL57Vr1/TWW28pICBArq6uKliwoLp166bTp0/bPDflfbhlyxZ99NFHKl26tNzc3PTtt99KkjZt2qQuXbroySefVJ48eeTr66umTZume17vd999pypVqsjNzU3FixfXBx98oC1btqR5LnF8fLw+/fRTBQYGys3NTb6+vmrdurUOHDiQoe2a1nnx9vpcKVmypBo2bKiff/5ZjRs3lqenp/Lly6devXrp4sWLNn2vX7+uUaNGqVatWtbPoICAAI0YMUIxMTGp5m0YhmbPnq1atWrJ09NTnp6eCgoK0nvvvSfp9mmNKafMNWrUyHpaYlr7890OHz6sdu3aKX/+/HJzc1PFihU1ZswYJSUl2fR70M+3tKScfnns2DG99tprKlq0qNzd3fXMM8/o5MmTkqTly5crODhYefLkUcmSJTVr1qw05zVnzhxrPx8fHzVt2lQ7duxI1S85OVmfffaZSpUqJTc3N1WqVEmLFi1Kt8YLFy7o5ZdfVvHixeXi4iI/Pz/1798/1Wv4oDK6nRs2bJjm+Lq7z+ufN2+eGjVqJEnq06eP9TVv2LChJNvz8adMmaKyZcvKzc1NZcuW1ZQpU1LNP2X/vdvd5/Vndl9L2X8uX76s3r17q0CBAvLy8tLzzz9v/VFs1qxZqlChgtzc3FS+fHmtWrUq1XymT5+upk2bqlixYnJxcVHRokX14osvpnn0JCkpSR999JFKlCghNzc3Va5cWUuWLElzfM2D7N93vxbbt29XqVKlJEkffPCBdZukvI73GhuR3t+kVatWqVq1anJzc9MTTzyhd9991/p38G4P8rlosVjUrFkzbdiwQTdu3Ehzfo8KRyyQbYwYMUJJSUkaOnSoEhISNH78eDVt2lRff/21+vXrp/79++uFF17Qt99+q/fee0+lSpWy+TV97dq1at++vQICAvTGG28oX758+umnn/Tee+/p4MGDWrp0qbXvihUrdOLECXXu3FklSpTQ5cuXNX/+fLVv316LFi1S9+7dU9XXrFkzFSxYUO+9954uX76sCRMmqGXLljpz5oy8vLwkSf/8849OnjypIUOGZHi9S5cuLUmaPXu2XnvttXS/IN8tvVOR0voCk2Lu3Lny9PRUhw4d5OHhoVatWmn+/Pn68MMP5eCQsd8ZUoJQgQIF0u3TtGlTFSlSRF9//XWqbfHrr79q9+7dGjJkiJydnSVl7vUwY9SoUfrkk0/UvHlzffTRR3JwcNCKFSvUqVMnTZ06Va+88kqG5vPUU09Juv0Hpnnz5vfse699IyEhQWPHjlWHDh3Utm1beXh4aN++fZo7d6527Nih8PBwubi4yNfXV23atNGqVat05coV5cuXzzqP5ORkLVq0SJUrV1bVqlUl3Q4VderU0R9//KG+ffsqMDBQFy5c0PTp01WrVi3t379fJUqUsKll2LBhunXrlkJDQ+Xt7a1y5cpJuv2F58qVK+rZs6f8/f0VERGhOXPm6JlnntG2bdv09NNPW+exZMkSdevWTaVLl9b7778vJycnzZ8/X6tXr0617rdu3VLz5s21a9cu9ejRQ4MHD9a1a9c0e/Zs1a1bV//73/8UEhKSodcjLWY/V6Tbp7A988wz6tChgzp27Kiff/5ZX331lfbv3699+/ZZj+ikbJMOHTpYg/sPP/ygMWPG6MCBA9q4caPNfHv06KFFixapVq1aGjlypHx9fXXixAktW7ZMH374odq3b68LFy5o1qxZeuedd1ShQgVJ//+ZkZ79+/erQYMGcnZ21iuvvKIiRYpo9erVeuutt3To0KE0v4Bn5PPtfnr16iVPT0+98847ioyM1Pjx49WsWTN99NFHevPNN/Xyyy+rb9++mjt3rgYMGKCKFSuqXr161ue/9dZbGjNmjGrWrKlPP/1U169f16xZs9SoUSOtWrVKLVq0sPb917/+pcmTJ6t+/fp6/fXXdfHiRb3yyitpHn39448/9NRTTykhIUH9+vVT6dKl9dtvv+nLL7/Utm3btH//fvn4+GRoHc1u5/upX7++3nnnHX366afq37+/9X1VuHBhm35TpkzR33//rQEDBsjLy0v/+c9/NGTIEF25ckXvv//+Ay83s/taiubNm8vf318ffvihfvvtN33xxRdq166d2rdvr1mzZqlfv35yc3PTF198oY4dO+rUqVPWL+3S7SP3tWvX1pAhQ5QvXz4dOXJEc+bM0datW/XLL78of/781r6DBw/WjBkz1KhRIw0bNkyRkZEaNGiQzfzulpn9u0KFCpo4caJef/1167pIuucYx3tZsWKFOnTooJIlS+q9996Tk5OTwsLCtHbt2lR9M/O5+NRTT2nmzJnasWPHff8ePVQGkMXCwsIMSUa1atWM+Ph4a/uqVasMSYaTk5Oxb98+a3t8fLxRpEgRo3bt2ta22NhYo3DhwsbTTz9t3Lp1y2b+EyZMMCQZ27Zts7bduHEjVR03b940ypYta1SoUMGmvVevXoYk4+WXX7Zp//bbbw1JxowZM6xtW7duNSQZkydPTnNde/XqZZQoUcKm7ffffze8vb0NScYTTzxhdO/e3Zg4caKxf//+NOfRoEEDQ9J9/7tzm6VsI19fX6NXr17WtpUrVxqSjHXr1qVaTokSJYzy5csbkZGRRmRkpHH69Gnjq6++Mnx8fAwnJyfjl19+SbO+FMOGDTMkGUePHrVpHzVqlCHJCA8Pt7Y9yOvx/vvvG5KMM2fOWNtSXqO0SLJZ5/DwcEOS8fbbb6fq27ZtW8PLy8uIjo62tqXsn3cu705OTk5Gq1at0px2p3vtG8nJyUZMTEyq9jlz5hiSjCVLlljb1qxZY0gypk2bZtN3y5YthiRj/Pjx1rYhQ4YYbm5uxsGDB236nj171vDy8rLZLinrWbZsWePmzZupaknrNfr777+N/PnzG88995y17datW4afn59RqFAh48qVK9b269evG6VKlTIkGWFhYdb2lPfnhg0bbOZ97do144knnjAaNGiQarl3S6n9zve4PT5XDOP2+0CSMXHiRJv2lLo/++wzm3kkJCSkqi9ln9+zZ4+1bcmSJYYk48UXXzSSkpJs+t/5OK11u586deoYjo6OxqFDh6xtycnJRqdOnQxJxpYtW6ztD/L5lp6U92SrVq2M5ORka/vkyZMNSYaXl5fxxx9/WNsvXrxouLq6Gl27drW2nThxwrBYLEbdunVtXq+IiAjDx8fHKFGihJGYmGjTt3HjxtY2w7j93rZYLKner23atDEKFixo/PnnnzZ179u3z3B0dDTef/99a9uDbO8H2c4NGjRI9dlvGIZx5swZQ5JNDdu2bUv1Prl7mqenp836xMfHGzVq1DCcnJxs2kuUKJHmeyitZWRmX0vZfwYNGmTT/vrrr1v/pl27ds3afujQIUOSMWLECJv+aX2+pHym/fvf/7a2HTlyxJBkNGvWzOZ9cvjwYcPBwSHdvw0Z2b/Tei3Saktxr9fp7r9JiYmJxhNPPGHkz5/fiIyMtLZHRUUZxYsXt8vn4o8//mhIMsaNG5dq2qPEqVDINl5++WW5uLhYH6f8UlOrVi2bZO7i4qKaNWtafzmXpM2bN+uff/5Rnz59FBUVpUuXLln/S/mVa9OmTdb+Hh4e1n/HxMTo8uXLiomJUePGjXX8+HFFR0enqu/111+3edy4cWNJsqkjMjJSkmx+Sb6fJ598UocOHbL+Sr548WK9/vrrCgkJUeXKlRUeHp7qOW5ubtq8eXOa//Xo0SPN5SxfvlxRUVHq1auXta1FixYqWLBguqdDnThxQgULFlTBggX15JNPqm/fvipQoIBWrVqlSpUq3XO9Upbz9ddfW9sMw9DChQtVqVIlBQcHW9sz83pk1qJFi2SxWNSrVy+b/eTSpUtq06aNrl+/rp9++inD88uXL1+GTqe4175hsViUJ08eSbcP86fswyn72J2H7Js1a6bChQvbbFfp9nZ2cnLSCy+8IOn2tl60aJHq16+vYsWK2aynh4eHateubfOeSPHyyy+nOabiztfoxo0bunz5shwdHVWrVi2b+sLDw/XXX3+pd+/eyps3r7Xd09NTAwcOTDXfhQsXqnz58qpevbpNjQkJCXr22We1Y8cOxcbGprFFM8bM50oKb29vDRo0yKZt0KBB8vb2tjldz8XFxXoULjExUVevXtWlS5fUpEkTSbavY8qv2ePGjUt1tDCjRw/TcvHiRe3atUtt2rRR5cqVre0Wi0UjR46UpDRPMczI59v9DBkyxOaIa8q2btOmjZ544glre8GCBVWuXDmbea9atUqGYejNN9+0eb38/PzUp08fnTt3znoKSErff/3rXzZja4KDg/Xss8/a1HTt2jWtWbNGbdq0kZubm80+VrJkSQUEBKT5PrifzG5ne3nhhRfk7+9vfezi4qLXX39diYmJaR4ZfNhee+01m8cpr33Pnj3l7e1tba9cubK8vb1T7Vcpny/Jycm6du2aLl26pCpVqsjHx8fmfbNmzRpJ0tChQ23eJ0FBQdbTdNNij/3bjPDwcP3555/q06ePzdF+Hx8fu30uphzVMXt6n1mcCoVs4+5D2ClfStI6vJk3b16bc8+PHz8uSerbt2+68//nn3+s/7548aJGjRqlVatWpfkmjIqKsvkwTKu+lDfxnXWk/FE1HvCSbyVLltTUqVM1depUXbhwQTt27NCCBQu0evVqtWrVSkePHrX5Quro6Gj9snK3tM5Hlm6fBlWwYEH5+/vbjI9o2rSpli5dqkuXLqU6valkyZLW+y2knJccEBCQoXVKCQ+LFi3Sp59+KgcHB/3vf//T2bNnNWbMGJu+mXk9Muv48eMyDEPly5dPt8+d+8r9GIaRodPX7rdvfPvttxo/frwOHDiQ6pzbq1evWv+dEh4mTJigU6dOqWzZsrp586aWL1+upk2bWk+ZiIyM1OXLl7Vp0yYVLFgwzWWm9QW2bNmyafb9/fffNXLkSG3cuFFRUVFprpsknTlzRpKsp1DdKa2248ePKzY2Nt0apdun/d35xfRBmPlcuXMed37ZlSRXV1c9+eSTqcaqTJ8+XTNmzNDRo0eVnJxsM+3O1/HXX39V0aJFU53iYlbK9g8MDEw1rUKFCnJwcEhVs5Sxz7f7edBtfe7cuQzVndJ2+vRphYSEWOtP6z1csWJFm6Bw8uRJJScna+7cuZo7d26G6s6IzG5ne0k5VelOFStWlKSHutz0mH2fbd26VR9++KH27NmjuLg4m2l3vm/u9/myfv36DNWXmf3bjPvts3fLzOdiyt+WjJ5O/bAQLJBtpHdVl4xc7SXlDTV27Fjr+eV38/Pzs/Zt2rSpjh8/rqFDhyokJEQ+Pj5ydHRUWFiYFi9enOoLwb3quPOLYsqHwJUrV+5bc3qKFi2qTp06qVOnTnrhhRe0ePFirVu3LtV53w/izJkz2rZtmwzDSPeL48KFC1P96uTh4ZFugMmInj176rXXXtPWrVvVpEkTff3113J0dLRZl8y+HndK74P07kH7KcuzWCxav359uq9pWl8W0nP16tV7fvinuNe+sXz5cnXp0kU1a9bU5MmT9cQTT8jNzU1JSUlq3rx5qvXv2bOnJkyYoK+//loff/yxli9frhs3btgcjUrZL5s0aaK33norw+uT1tGKGzduqH79+rp586Zee+01BQUFycvLSw4ODvrss8+0devWDM//boZhKCgo6J6X7c3I9k2Pmc+VBzVhwgS98cYbatq0qYYMGSI/Pz+5uLgoIiJCvXv3vu9+nJUy8vmW2XnYY96ZlbKMF1980eb9caeUo4UP04N8Rj2OyzXz2u/bt09NmzZVQECAPv/8c5UqVcp6r6WuXbva5X3zMPbBe32BN7t9M/O5mPK3xcznpT0QLJAjlClTRlLGvggfPnxYhw4d0nvvvZfqzslz5swxVUfKF1J7HV6tXbu2Fi9erIiICFPzCQsLs16BxtfXN9X0UaNG6auvvkoVLMzq3r27hg8frq+//lp169bVsmXL9Oyzz6po0aLWPvZ4PVKO5tw9oDmtX+7KlCmjDRs2qHjx4mn+6vcgzp49q8TExPueFibde99YsGCB3NzctG3bNpsv9idOnEhzXlWqVFGVKlW0cOFCffTRR/r666+tA7tTFCxYUL6+voqOjjYVDiXp+++/119//aWvvvoq1Y397rzmuyTrFVNSrgZ0p7TaypQpo8jISDVu3NjUKUAP0+nTp5WQkGBz1CI+Pl6nT5+2+QVywYIFKlmypNavX2+zLhs2bEg1z7Jly2rVqlX6559/7nnU4kF/fUz5hfjo0aOppp04cULJycmZ+oX+YUup6ejRo6kGDB87dsymT8r/T5w4kW7fFAEBAbJYLEpISDD9PrjTg27nfPnypXlaa1qfURl5zVOO0t/p7u2Usty0fszI7HIfhsWLFyspKUnr16+3OcJx8+ZNm6MVku3ny937cVqfL2bda5vc+Xfnbndv3zv32bvdvc9KmftcTDkTISN/jx6m7PkpDjygZs2aqVChQvr888/TfJPHxsbq+vXrkv7/l4u7f6k4cuSI6XNiCxYsqMDAQOvlLDNi+/btaZ5DnpycbD1XNq1DpRmVnJysefPmKSgoSC+99JI6duyY6r9u3brpl19+0b59+zK9nLQULFhQzz33nJYvX65FixYpOjo61a+G9ng9Uo7CbNmyxaZ9/PjxqfqmjEF55513Ul0SUnqw06BSXucGDRrct++99g1HR0dZLBabX+YMw9DHH3+c7vx69eqlc+fOafHixdq6dau6dOlicw12BwcHvfDCC9q7d2+6l9HN6Lm46b1GmzZtSnXJxpCQEBUtWlTz5s2z+VJw48YNzZgxI9W8e/bsqb///jvdX+Ye5PV4WKKjozV9+nSbtunTpys6OlrPP/+8tS3ldbxzOyUmJurzzz9PNc+UsTBvvvlmql9k73x+yhVoMnoUtFChQqpTp45Wr16tI0eO2Mzzs88+kyS1a9cuQ/N6lNq0aSOLxaKxY8fanAp44cIFhYWFqUSJEqpWrZpN3wkTJti8h3/++edUnwH58+dXixYttHz58jTfe4ZhWMc/PYgH3c5ly5bV9evXtXfvXmtbcnKyJk6cmGreGXnNFy1apPPnz1sfJyQkaOLEiXJ0dFSrVq1slnvixAmbH6fi4+M1bdq0TC33YUjv8+XTTz9N9d5o3bq1JGny5Mk203755ZdUV12zh3ttk1KlSsnJySnVPrdr165U+1r16tXl7++vsLAwmys6RkdH2+1zcffu3XJyclLdunXvv2IPEUcskCN4eHjo66+/1vPPP69y5cqpb9++CggIUFRUlE6cOKHly5drxYoVatiwoSpUqKDAwECNGTNGMTExKleunE6dOqWZM2cqKCgozV+VHkSnTp300Ucf6cKFCza/zKdn3Lhx2rlzp1q3bq3g4GD5+Pjo77//1nfffafw8HA1atTI1A1vNm3apD///FP9+vVLt0+HDh00evRozZ07VzVq1Mj0stLSq1cv/fe//9Ubb7whHx8fmy9ikuzyenTr1k3vvPOO+vfvrxMnTihfvnzasGFDmpfkrVGjhkaPHq3Ro0eratWq6tSpk/z8/HThwgXrnUsTEhIytG7r1q1TgQIFrNedv5/09o2OHTvqu+++U+PGjdWzZ0/dunVLK1euvOelg1944QW9+eabGjRokJKTk9M8zeOTTz7Rzp071blzZ3Xu3Fm1a9eWi4uLzp07p3Xr1ql69eppXoP9bvXq1VORIkX0xhtv6OzZs/L399fBgwe1YMECBQUF6ZdffrH2dXJy0rhx4/TCCy+oZs2a6tevn5ycnDRv3jzlz59fZ86csfkVcOjQodq8ebOGDx+urVu3qnHjxvL29tYff/yh77//3nokJyuVLl1aH3zwgY4cOaLq1asrPDxcX331lcqXL29z+eCOHTvq7bff1nPPPaf27dsrOjpaixcvtg7ovlOnTp3UpUsXff311/r111/Vpk0b5c2bV6dOndLGjRutX1Zr1KghBwcHffLJJ7p69ao8PDxUqlQp1apVK916J0+erAYNGujpp5+2XgZ1zZo12rhxo7p3757uPXOyUrly5TR8+HCNGTNG9evXV5cuXayXm71x44YWLVpk/QJavnx5vfLKK5o6daoaN26sDh066OLFi5o6daqqVKmS6jr/X375perVq6f69eurZ8+eqlatmpKTk3X69GmtWrVKPXv2tN674EE8yHbu37+/xo8fr3bt2mno0KFycXHRsmXL0jxlpmLFivLy8tL06dPl7u4uX19fFSpUyDrgWLodGGrVqqWBAwfKy8tLixcv1r59+/Tuu+/anHc/ePBgffPNN2rSpIkGDhyohIQELViwIM1THjOzr9lDu3btNHHiRLVo0UL9+/eXi4uLNm/erMOHD6ca9xcYGKj+/ftr1qxZatKkidq1a6fIyEhNmzZN1apVU3h4uF2PvOTPn18BAQH65ptvVLp0aRUuXFgeHh5q3bq1PD091bt3b82ZM0fdunVTw4YN9euvvyosLEyVK1fWoUOHrPNxdHTUxIkT1blzZ9WsWVOhoaHW+0jlz59ff/zxh81yH/Rz0TAMbdiwQc2bN8/05XDt5iFfdQq4r3td4k53XSo0RXqXF/3ll1+MF154wfDz8zOcnZ2NQoUKGU899ZTx4YcfGpcvX7b2O3v2rNGxY0ejQIECRp48eYwaNWoYy5cvN30pU8O4fXlEJyenNC/5ltblZn/66SfjX//6lxESEmIUKlTIcHJyMnx8fIzatWsb48ePN+Li4mz6N2jQwPDw8EizHsP4/0s/plxKs2PHjoYk4/Dhw+k+xzAMo2zZsoaPj4/1sqclSpQwAgMD7/mcjIiPjzfy5ctnSDJeeumlNPs8yOuRVpthGMbu3buNOnXqGK6urkb+/PmN0NBQ4+rVq+nuQ2vWrDGaNm1q5M2b13BxcTH8/f2N5s2bG19++aVNv/QuN3vjxg3Dw8PDGDZsWIa3xb32jVmzZhkVKlQwXF1djSJFihihoaHG5cuX063fMAyjVatWhiSjTJky6S7z5s2bxocffmhUqlTJcHNzMzw9PY3y5csbL730krF79+5U65nepSYPHTpkNGvWzPD19TU8PT2NBg0aGP/73//SfX98++23RlBQkOHi4mI88cQTxujRo43ly5enunyuYdy+RO3kyZONkJAQw93d3XB3dzcCAgKM7t27Gxs3bkx33e5Vu70+V1Iu1xkeHm40atTIcHd3N3x9fY0XX3zR+Pvvv236JiYmGp9++qlRunRpw8XFxShevLgxfPhw49ixY2lesjIpKcmYOnWqUa1aNSNPnjyGp6enERQUZIwePdqm37x584wKFSoYzs7O99wf7nTw4EGjbdu21v27fPnyxr///W+by7Omt8732053S+89ea9LdaZ3+dVZs2YZVatWNVxdXQ0vLy+jSZMmxv/+979U/ZKSkoyPP/7YKF68uOHi4mIEBgYaCxcuTLeWyMhIY9iwYUaZMmUMV1dXw8fHx6hUqZIxZMgQm0tiP+glVzO6nQ3DMNauXWtUqVLFcHFxMYoWLWq8+eabxokTJ9LcRmvXrjWqVatmuLq6GpKslxe98xKnkydPNgICAgwXFxcjICDAmDRpUpo1zps3zyhbtqzh7OxslCxZ0vj3v/9tfP/992leKvVB97X09p97XYo1rUvgrlixwggODjbc3d2N/PnzG126dDHOnTuXZt/ExERj9OjRxhNPPGG4uLgYQUFBxpIlS4w33njDkGT8888/963PMFLv3+ntr3v27DHq1KljuLu7G5Js9tvr168b/fr1M/Lly2fkyZPHqFevnrFz5850l/vdd99Z9wF/f39j1KhRxqZNm9LcVg/yubh9+3ZDkrFmzZo01/VRshjGIxg9BeQyAwcO1KZNm3Ty5EmbXyt79+6t7du3p3k3UWRP8+bNU58+fXTmzBmbO+dOnjxZI0eOtF7dJ6PS2zdyg/Hjx2vYsGH66aefVLt27awuJ0NKliypkiVL2tzVG8gq27dvV6NGjRQWFpahO7DnJq1bt9bWrVsVHR39UC7OkJ21a9dOf/75p/bt25flV4VijAXwEHz44Ye6fPmywsLCsroUPASxsbH6/PPPNXz48AcKFVLu2DcSEhJSjV+5ceOGpk2bpvz589vcwwQAHkRaYxIPHz6s9evXq3HjxrkuVBw4cECrVq3S+PHjszxUSIyxAB6KQoUK6dq1a1ldBh6SPHny6MKFC5l6bm7YN06fPq3nnntOXbt2ValSpXThwgXNnz9fZ86c0ZdffpnqnhAAkFHz58/X119/rZYtW6pgwYI6ceKEZs2aJRcXF3344YdZXd4jlzJmKLsgWAAA7KpgwYKqXbu2Fi1apIsXL8rJyUlBQUH6/PPP1blz56wuD8BjLDg4WCtWrNAXX3yhK1euyMvLS40bN9b7779vvXIYsg5jLAAAAACYxhgLAAAAAKYRLAAAAACYRrAAAAAAYBrBAgAAAIBpBAsAAAAAphEsAAAAAJhGsAAAAABgGsECAAAAgGkECwAAAACm/R+DcSTpupai6AAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 800x1350 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "explainer = shap.TreeExplainer(total_clfs[target_values.index(column)][0])\n",
    "\n",
    "val_dataset = Pool(data=X_test, label=y_test, cat_features=cat_features)\n",
    "shap_values = explainer.shap_values(val_dataset)\n",
    "shap.summary_plot(shap_values, X_test, max_display=30)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "63ee66f5-a997-4810-9e8a-9203315d62bc",
   "metadata": {},
   "source": [
    "сохраняем модели"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "c802768b-0a04-4b3a-a7cd-d5a53ae677b5",
   "metadata": {},
   "outputs": [],
   "source": [
    "for j, column in enumerate(target_values):\n",
    "    for i, clf in enumerate(total_clfs[j]):\n",
    "        clf.save_model(f\"models/{column.replace('/', '_')}_classificator{i}.cbm\",\n",
    "               format=\"cbm\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fa219b74-7f04-4860-9a02-94417abebe3e",
   "metadata": {},
   "source": [
    "# Inference"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a2faac16-9a67-4dec-b4fc-64468e88b81b",
   "metadata": {},
   "source": [
    "код для инференса только для одной метеостанции"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "36e4a1bc-0e57-43db-a76f-48f2723df483",
   "metadata": {},
   "outputs": [],
   "source": [
    "inf_columns = X.columns.tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "659cbbef-d400-4a12-9f1c-2d31afe83689",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Прочие опасности',\n",
       " 'Аварии с выбросом опасных/токсичных веществ',\n",
       " 'ЖКХ',\n",
       " 'Взрывы/пожары/разрушения',\n",
       " 'ДТП',\n",
       " 'Опасные природные явления',\n",
       " 'Нет события']"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "target_values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "dd810484-9e4f-4647-9514-c128ac579581",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "weather_stats_df = pd.read_csv(\"Данные/Данные по метеостанциям.csv\", low_memory=False)\n",
    "weather_stats_df.replace('^\\s+', '', regex=True, inplace=True) \n",
    "weather_stats_df.replace('\\s+$', '', regex=True, inplace=True)\n",
    "weather_stats_df[\"Местное время\"] = pd.to_datetime(weather_stats_df[\"Местное время\"], format=\"%d.%m.%Y %H:%M\")\n",
    "\n",
    "predicted_matrixes = []\n",
    "\n",
    "meteostation = 'Чердынь'\n",
    "    \n",
    "meteo_df = weather_stats_df[weather_stats_df['meteostation'] == meteostation].copy()\n",
    "\n",
    "meteo_df = meteo_df.sort_values(by=['Местное время'])\n",
    "meteo_df['RRR'] = meteo_df['RRR'].map(lambda x: x if x not in ['Осадков нет', 'Следы осадков'] else 0)\n",
    "meteo_df[\"day_of_week\"] = meteo_df[\"Местное время\"].map(lambda x: x.day_of_week)\n",
    "meteo_df[\"is_weekday\"] = meteo_df[\"day_of_week\"].isin([6, 7])\n",
    "meteo_df[\"hour\"] = meteo_df[\"Местное время\"].map(lambda x: x.hour)\n",
    "\n",
    "for column in selected_columns:\n",
    "    for i in range(10):\n",
    "        meteo_df[f\"{column}_lag{i}\"] = meteo_df[column].shift(i)\n",
    "\n",
    "for cat_feature in cat_features:\n",
    "    try:\n",
    "        meteo_df[cat_feature] = meteo_df[cat_feature].fillna('')\n",
    "    except Exception as e:\n",
    "        pass\n",
    "    meteo_df[cat_feature] = meteo_df[cat_feature].astype('category')\n",
    "\n",
    "X = meteo_df.drop(columns=['Местное время'])\n",
    "X = X[inf_columns]\n",
    "\n",
    "for column in target_values:\n",
    "\n",
    "    if column == 'Нет события':\n",
    "        continue\n",
    "\n",
    "    clfs = [CatBoostClassifier().load_model(f\"models/{column.replace('/', '_')}_classificator{i}.cbm\", format=\"cbm\") for i in range(5)]\n",
    "    predict_matrix = np.array([clf.predict_proba(X) for clf in clfs]).mean(axis=0)\n",
    "    predicted_matrixes.append(predict_matrix)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "377df9fa-ddd0-400a-991f-2835ae92480f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Прочие опасности',\n",
       " 'Аварии с выбросом опасных/токсичных веществ',\n",
       " 'ЖКХ',\n",
       " 'Взрывы/пожары/разрушения',\n",
       " 'ДТП',\n",
       " 'Опасные природные явления',\n",
       " 'Нет события']"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "target_values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "e7666035-a350-491e-b74c-ecb967a732b1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6, 29194)"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predicted_matrixes = np.array(predicted_matrixes).argmax(axis=2)\n",
    "predicted_matrixes.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7b706709-c58d-4e0a-af71-4a933ebabe07",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "59c34d5d-711f-443e-9ec6-535e591c7525",
   "metadata": {},
   "outputs": [],
   "source": [
    "meteostation = 'Чердынь'\n",
    "meteo_df = weather_stats_df[weather_stats_df['meteostation'] == meteostation].copy()\n",
    "meteo_df = meteo_df.sort_values(by=['Местное время'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "0f2d0f71-13d6-438c-bdb0-4362fcb921b0",
   "metadata": {},
   "outputs": [],
   "source": [
    "submit_df = pd.DataFrame()\n",
    "submit_df['Местное время'] = meteo_df['Местное время']\n",
    "i = 0\n",
    "for column in target_values:\n",
    "    if column == 'Нет события':\n",
    "        continue\n",
    "    submit_df[column + \"_pred\"] = predicted_matrixes[i]\n",
    "    i += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "44b7f8d7-2732-4e33-aca1-8bf47247119e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Прочие опасности',\n",
       " 'Аварии с выбросом опасных/токсичных веществ',\n",
       " 'ЖКХ',\n",
       " 'Взрывы/пожары/разрушения',\n",
       " 'ДТП',\n",
       " 'Опасные природные явления',\n",
       " 'Нет события']"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "target_values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "bffb3846-3153-4ddc-bcd9-c4de11346134",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Опасные природные явления_pred\n",
       "0    26321\n",
       "1     2873\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "submit_df['Опасные природные явления_pred'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "61f09a5a-bcea-4060-a339-7375fd17e118",
   "metadata": {},
   "outputs": [],
   "source": [
    "submit_df['total'] = sum([submit_df[column + \"_pred\"] for column in target_values if column != 'Нет события'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "7a3dfe7e-3f29-452f-8423-4b0a7041ce70",
   "metadata": {},
   "outputs": [],
   "source": [
    "cur_df = merged_df[merged_df['meteostation'] == 'Чердынь']\n",
    "cur_df = cur_df.sort_values(by=['Местное время'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "bd64cea5-f531-4100-a855-03f748599ae6",
   "metadata": {},
   "outputs": [],
   "source": [
    "cur_merged = submit_df[['Местное время', 'ЖКХ_pred']].merge(cur_df[['Местное время', 'ЖКХ']], on='Местное время')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "473ea6b7-d6d2-4e34-8b87-71df6baac0aa",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}