LOOK-ALIKE-X5-FINAL.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "881987d3",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:39:17.110637Z",
     "start_time": "2022-03-27T18:39:11.161224Z"
    }
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "import sklearn\n",
    "from sklearn.preprocessing import PowerTransformer\n",
    "from sklearn.model_selection import train_test_split, StratifiedKFold\n",
    "from sklearn.metrics import confusion_matrix, f1_score, roc_auc_score\n",
    "\n",
    "import lightgbm as lgb\n",
    "from catboost import CatBoostClassifier, Pool\n",
    "from catboost.utils import eval_metric, get_roc_curve\n",
    "from catboost.utils import get_roc_curve\n",
    "from catboost.utils import get_fpr_curve\n",
    "from catboost.utils import get_fnr_curve\n",
    "from catboost.utils import select_threshold\n",
    "\n",
    "from hyperopt import hp, fmin, tpe, Trials\n",
    "from hyperopt.pyll import scope\n",
    "import optuna\n",
    "from optuna.samplers import TPESampler\n",
    "\n",
    "import shap"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "42715cb2",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:39:17.117065Z",
     "start_time": "2022-03-27T18:39:17.112785Z"
    }
   },
   "outputs": [],
   "source": [
    "pd.set_option('display.max_rows', 500)\n",
    "pd.set_option('display.max_columns', 500)\n",
    "pd.set_option('display.width', 1000)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8f960b21",
   "metadata": {},
   "source": [
    "## EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "2c931836",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:40:14.084191Z",
     "start_time": "2022-03-27T18:40:10.717785Z"
    }
   },
   "outputs": [],
   "source": [
    "df = pd.read_csv('../first_part/2. Cup_IT_2022_Датасет_Data_Science.csv')\n",
    "df_2 = pd.read_csv('Cup_IT_2022_Data_Science_additional_data.csv')\n",
    "\n",
    "df = pd.concat([df, df_2], axis=1)\n",
    "df.drop(['Unnamed: 0', 'index'], axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "8756f00e",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:40:14.268365Z",
     "start_time": "2022-03-27T18:40:14.086779Z"
    }
   },
   "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>client_id</th>\n",
       "      <th>is_in_club</th>\n",
       "      <th>cnt_checks_10_Мясная гастрономия</th>\n",
       "      <th>rto_std_11</th>\n",
       "      <th>rto_6</th>\n",
       "      <th>rto_12_Сыры</th>\n",
       "      <th>rto_12_Птица и изделия из птицы</th>\n",
       "      <th>rto_8</th>\n",
       "      <th>cnt_checks_9</th>\n",
       "      <th>rto_9_Мясная гастрономия</th>\n",
       "      <th>rto_10_Птица и изделия из птицы</th>\n",
       "      <th>cnt_checks_9_Рыба и рыбные изделия</th>\n",
       "      <th>rto_stddev_6_Рыба и рыбные изделия</th>\n",
       "      <th>rto_12_Мясная гастрономия</th>\n",
       "      <th>rto_stddev_8_Крупы и зерновые</th>\n",
       "      <th>rto_12</th>\n",
       "      <th>rto_6_Птица и изделия из птицы</th>\n",
       "      <th>rto_6_Рыба и рыбные изделия</th>\n",
       "      <th>cnt_checks_11</th>\n",
       "      <th>rto_stddev_7_Птица и изделия из птицы</th>\n",
       "      <th>rto_stddev_9_Сыры</th>\n",
       "      <th>cnt_checks_10_Овощи - Фрукты</th>\n",
       "      <th>cnt_checks_12_Птица и изделия из птицы</th>\n",
       "      <th>rto_6_Мясная гастрономия</th>\n",
       "      <th>rto_stddev_6_Овощи - Фрукты</th>\n",
       "      <th>cnt_checks_6_Мясная гастрономия</th>\n",
       "      <th>rto_stddev_6_Мясная гастрономия</th>\n",
       "      <th>cnt_checks_11_Сыры</th>\n",
       "      <th>cnt_checks_12_Овощи - Фрукты</th>\n",
       "      <th>rto_stddev_10_Птица и изделия из птицы</th>\n",
       "      <th>cnt_checks_6_Сыры</th>\n",
       "      <th>rto_stddev_10_Овощи - Фрукты</th>\n",
       "      <th>rto_12_Крупы и зерновые</th>\n",
       "      <th>rto_7_Птица и изделия из птицы</th>\n",
       "      <th>cnt_checks_12_Рыба и рыбные изделия</th>\n",
       "      <th>rto_stddev_10_Крупы и зерновые</th>\n",
       "      <th>rto_7_Крупы и зерновые</th>\n",
       "      <th>rto_7_Овощи - Фрукты</th>\n",
       "      <th>cnt_checks_10_Птица и изделия из птицы</th>\n",
       "      <th>rto_9_Рыба и рыбные изделия</th>\n",
       "      <th>cnt_checks_11_Крупы и зерновые</th>\n",
       "      <th>rto_stddev_11_Крупы и зерновые</th>\n",
       "      <th>rto_stddev_6_Крупы и зерновые</th>\n",
       "      <th>rto_stddev_12_Овощи - Фрукты</th>\n",
       "      <th>cnt_checks_11_Рыба и рыбные изделия</th>\n",
       "      <th>rto_stddev_9_Крупы и зерновые</th>\n",
       "      <th>rto_11_Крупы и зерновые</th>\n",
       "      <th>rto_stddev_10_Сыры</th>\n",
       "      <th>cnt_checks_12_Мясная гастрономия</th>\n",
       "      <th>rto_stddev_10_Мясная гастрономия</th>\n",
       "      <th>rto_stddev_8_Рыба и рыбные изделия</th>\n",
       "      <th>cnt_checks_9_Крупы и зерновые</th>\n",
       "      <th>rto_10_Сыры</th>\n",
       "      <th>rto_stddev_6_Сыры</th>\n",
       "      <th>cnt_checks_7_Сыры</th>\n",
       "      <th>cnt_checks_6_Птица и изделия из птицы</th>\n",
       "      <th>cnt_checks_6_Рыба и рыбные изделия</th>\n",
       "      <th>rto_8_Овощи - Фрукты</th>\n",
       "      <th>cnt_checks_10_Крупы и зерновые</th>\n",
       "      <th>rto_12_Овощи - Фрукты</th>\n",
       "      <th>rto_11_Мясная гастрономия</th>\n",
       "      <th>cnt_checks_12_Крупы и зерновые</th>\n",
       "      <th>rto_stddev_11_Мясная гастрономия</th>\n",
       "      <th>rto_stddev_12_Птица и изделия из птицы</th>\n",
       "      <th>rto_stddev_7_Крупы и зерновые</th>\n",
       "      <th>cnt_checks_6</th>\n",
       "      <th>cnt_checks_9_Овощи - Фрукты</th>\n",
       "      <th>cnt_checks_7_Крупы и зерновые</th>\n",
       "      <th>cnt_checks_8_Птица и изделия из птицы</th>\n",
       "      <th>rto_11_Птица и изделия из птицы</th>\n",
       "      <th>rto_stddev_11_Сыры</th>\n",
       "      <th>cnt_checks_7_Птица и изделия из птицы</th>\n",
       "      <th>rto_10_Овощи - Фрукты</th>\n",
       "      <th>rto_stddev_7_Мясная гастрономия</th>\n",
       "      <th>rto_9_Сыры</th>\n",
       "      <th>rto_10_Мясная гастрономия</th>\n",
       "      <th>rto_stddev_12_Рыба и рыбные изделия</th>\n",
       "      <th>cnt_checks_7</th>\n",
       "      <th>rto_7_Мясная гастрономия</th>\n",
       "      <th>cnt_checks_9_Мясная гастрономия</th>\n",
       "      <th>cnt_checks_7_Мясная гастрономия</th>\n",
       "      <th>rto_9_Овощи - Фрукты</th>\n",
       "      <th>rto_stddev_9_Птица и изделия из птицы</th>\n",
       "      <th>rto_std_9</th>\n",
       "      <th>rto_stddev_10_Рыба и рыбные изделия</th>\n",
       "      <th>rto_stddev_9_Рыба и рыбные изделия</th>\n",
       "      <th>cnt_checks_9_Сыры</th>\n",
       "      <th>rto_stddev_6_Птица и изделия из птицы</th>\n",
       "      <th>rto_10_Рыба и рыбные изделия</th>\n",
       "      <th>rto_12_Рыба и рыбные изделия</th>\n",
       "      <th>rto_10_Крупы и зерновые</th>\n",
       "      <th>cnt_checks_12</th>\n",
       "      <th>cnt_checks_10_Рыба и рыбные изделия</th>\n",
       "      <th>rto_stddev_9_Овощи - Фрукты</th>\n",
       "      <th>rto_11_Рыба и рыбные изделия</th>\n",
       "      <th>rto_6_Овощи - Фрукты</th>\n",
       "      <th>cnt_checks_10</th>\n",
       "      <th>rto_stddev_7_Овощи - Фрукты</th>\n",
       "      <th>cnt_checks_12_Сыры</th>\n",
       "      <th>rto_std_7</th>\n",
       "      <th>cnt_checks_8_Овощи - Фрукты</th>\n",
       "      <th>rto_stddev_8_Мясная гастрономия</th>\n",
       "      <th>rto_stddev_11_Овощи - Фрукты</th>\n",
       "      <th>cnt_checks_7_Овощи - Фрукты</th>\n",
       "      <th>rto_9_Птица и изделия из птицы</th>\n",
       "      <th>cnt_checks_6_Крупы и зерновые</th>\n",
       "      <th>cnt_checks_8_Крупы и зерновые</th>\n",
       "      <th>cnt_checks_7_Рыба и рыбные изделия</th>\n",
       "      <th>cnt_checks_11_Овощи - Фрукты</th>\n",
       "      <th>cnt_checks_8_Рыба и рыбные изделия</th>\n",
       "      <th>cnt_checks_11_Мясная гастрономия</th>\n",
       "      <th>rto_8_Мясная гастрономия</th>\n",
       "      <th>rto_10</th>\n",
       "      <th>rto_std_8</th>\n",
       "      <th>rto_stddev_11_Птица и изделия из птицы</th>\n",
       "      <th>rto_stddev_8_Птица и изделия из птицы</th>\n",
       "      <th>rto_stddev_12_Мясная гастрономия</th>\n",
       "      <th>rto_6_Сыры</th>\n",
       "      <th>rto_7</th>\n",
       "      <th>rto_std_6</th>\n",
       "      <th>cnt_checks_9_Птица и изделия из птицы</th>\n",
       "      <th>rto_stddev_11_Рыба и рыбные изделия</th>\n",
       "      <th>rto_9</th>\n",
       "      <th>rto_std_10</th>\n",
       "      <th>rto_stddev_8_Сыры</th>\n",
       "      <th>rto_8_Птица и изделия из птицы</th>\n",
       "      <th>rto_11_Овощи - Фрукты</th>\n",
       "      <th>cnt_checks_11_Птица и изделия из птицы</th>\n",
       "      <th>rto_7_Рыба и рыбные изделия</th>\n",
       "      <th>cnt_checks_6_Овощи - Фрукты</th>\n",
       "      <th>rto_9_Крупы и зерновые</th>\n",
       "      <th>rto_7_Сыры</th>\n",
       "      <th>rto_8_Крупы и зерновые</th>\n",
       "      <th>cnt_checks_8_Мясная гастрономия</th>\n",
       "      <th>rto_stddev_7_Сыры</th>\n",
       "      <th>rto_8_Сыры</th>\n",
       "      <th>rto_11</th>\n",
       "      <th>rto_std_12</th>\n",
       "      <th>cnt_checks_10_Сыры</th>\n",
       "      <th>rto_stddev_12_Сыры</th>\n",
       "      <th>rto_6_Крупы и зерновые</th>\n",
       "      <th>rto_stddev_7_Рыба и рыбные изделия</th>\n",
       "      <th>rto_stddev_8_Овощи - Фрукты</th>\n",
       "      <th>rto_8_Рыба и рыбные изделия</th>\n",
       "      <th>rto_11_Сыры</th>\n",
       "      <th>cnt_checks_8_Сыры</th>\n",
       "      <th>cnt_checks_8</th>\n",
       "      <th>rto_stddev_9_Мясная гастрономия</th>\n",
       "      <th>rto_stddev_12_Крупы и зерновые</th>\n",
       "      <th>cnt_unique_plu_6</th>\n",
       "      <th>cnt_unique_plu_7</th>\n",
       "      <th>cnt_unique_plu_8</th>\n",
       "      <th>cnt_unique_plu_9</th>\n",
       "      <th>cnt_unique_plu_10</th>\n",
       "      <th>cnt_unique_plu_11</th>\n",
       "      <th>cnt_unique_plu_12</th>\n",
       "      <th>cnt_unique_plu_10_Мясная гастрономия</th>\n",
       "      <th>cnt_unique_plu_10_Овощи - Фрукты</th>\n",
       "      <th>cnt_unique_plu_10_Птица и изделия из птицы</th>\n",
       "      <th>cnt_unique_plu_10_Рыба и рыбные изделия</th>\n",
       "      <th>cnt_unique_plu_11_Мясная гастрономия</th>\n",
       "      <th>cnt_unique_plu_11_Овощи - Фрукты</th>\n",
       "      <th>cnt_unique_plu_11_Птица и изделия из птицы</th>\n",
       "      <th>cnt_unique_plu_11_Рыба и рыбные изделия</th>\n",
       "      <th>cnt_unique_plu_12_Мясная гастрономия</th>\n",
       "      <th>cnt_unique_plu_12_Овощи - Фрукты</th>\n",
       "      <th>cnt_unique_plu_12_Птица и изделия из птицы</th>\n",
       "      <th>cnt_unique_plu_12_Рыба и рыбные изделия</th>\n",
       "      <th>cnt_unique_plu_6_Мясная гастрономия</th>\n",
       "      <th>cnt_unique_plu_6_Овощи - Фрукты</th>\n",
       "      <th>cnt_unique_plu_6_Птица и изделия из птицы</th>\n",
       "      <th>cnt_unique_plu_6_Рыба и рыбные изделия</th>\n",
       "      <th>cnt_unique_plu_7_Мясная гастрономия</th>\n",
       "      <th>cnt_unique_plu_7_Овощи - Фрукты</th>\n",
       "      <th>cnt_unique_plu_7_Птица и изделия из птицы</th>\n",
       "      <th>cnt_unique_plu_7_Рыба и рыбные изделия</th>\n",
       "      <th>cnt_unique_plu_8_Мясная гастрономия</th>\n",
       "      <th>cnt_unique_plu_8_Овощи - Фрукты</th>\n",
       "      <th>cnt_unique_plu_8_Птица и изделия из птицы</th>\n",
       "      <th>cnt_unique_plu_8_Рыба и рыбные изделия</th>\n",
       "      <th>cnt_unique_plu_9_Мясная гастрономия</th>\n",
       "      <th>cnt_unique_plu_9_Овощи - Фрукты</th>\n",
       "      <th>cnt_unique_plu_9_Птица и изделия из птицы</th>\n",
       "      <th>cnt_unique_plu_9_Рыба и рыбные изделия</th>\n",
       "      <th>cnt_unique_plu_10_Крупы и зерновые</th>\n",
       "      <th>cnt_unique_plu_10_Сыры</th>\n",
       "      <th>cnt_unique_plu_11_Крупы и зерновые</th>\n",
       "      <th>cnt_unique_plu_11_Сыры</th>\n",
       "      <th>cnt_unique_plu_12_Крупы и зерновые</th>\n",
       "      <th>cnt_unique_plu_12_Сыры</th>\n",
       "      <th>cnt_unique_plu_6_Крупы и зерновые</th>\n",
       "      <th>cnt_unique_plu_6_Сыры</th>\n",
       "      <th>cnt_unique_plu_7_Крупы и зерновые</th>\n",
       "      <th>cnt_unique_plu_7_Сыры</th>\n",
       "      <th>cnt_unique_plu_8_Крупы и зерновые</th>\n",
       "      <th>cnt_unique_plu_8_Сыры</th>\n",
       "      <th>cnt_unique_plu_9_Крупы и зерновые</th>\n",
       "      <th>cnt_unique_plu_9_Сыры</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>300.021342</td>\n",
       "      <td>2418.116714</td>\n",
       "      <td>853.016746</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5092.170392</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5738.687177</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>41.017712</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>73.644488</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1012.977676</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>470.600157</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>124.564463</td>\n",
       "      <td>5.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>81.154671</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1583.623070</td>\n",
       "      <td>839.257740</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>440.974542</td>\n",
       "      <td>357.220317</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1658.894353</td>\n",
       "      <td>184.518332</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>482.557307</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1720.373691</td>\n",
       "      <td>675.763593</td>\n",
       "      <td>NaN</td>\n",
       "      <td>73.846425</td>\n",
       "      <td>148.316558</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>40.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>519.031384</td>\n",
       "      <td>19899.211736</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>32392.636661</td>\n",
       "      <td>7.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>176.866878</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>10.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>211.090724</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>96.228636</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11849.260237</td>\n",
       "      <td>1.0</td>\n",
       "      <td>975.070527</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>160.83624</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>564.287939</td>\n",
       "      <td>238.798235</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>14559.692214</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>12.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>228.800923</td>\n",
       "      <td>220.173709</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3522.302011</td>\n",
       "      <td>NaN</td>\n",
       "      <td>108.200176</td>\n",
       "      <td>148.900417</td>\n",
       "      <td>NaN</td>\n",
       "      <td>16.0</td>\n",
       "      <td>276.814607</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1156.885192</td>\n",
       "      <td>NaN</td>\n",
       "      <td>524.139352</td>\n",
       "      <td>370.327074</td>\n",
       "      <td>192.14527</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4228.863632</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5.0</td>\n",
       "      <td>102.446072</td>\n",
       "      <td>374.702932</td>\n",
       "      <td>7025.919801</td>\n",
       "      <td>11.0</td>\n",
       "      <td>203.249441</td>\n",
       "      <td>NaN</td>\n",
       "      <td>829.838894</td>\n",
       "      <td>19.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>98.340478</td>\n",
       "      <td>15.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19329.981907</td>\n",
       "      <td>685.952065</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1597.936059</td>\n",
       "      <td>28890.382271</td>\n",
       "      <td>758.343961</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6337.074013</td>\n",
       "      <td>765.052043</td>\n",
       "      <td>74.362466</td>\n",
       "      <td>NaN</td>\n",
       "      <td>572.071018</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2001.937586</td>\n",
       "      <td>9.0</td>\n",
       "      <td>160.913398</td>\n",
       "      <td>1556.301144</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>158.711615</td>\n",
       "      <td>1926.327565</td>\n",
       "      <td>4327.469741</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>225.219998</td>\n",
       "      <td>301.404157</td>\n",
       "      <td>525.868862</td>\n",
       "      <td>1153.293331</td>\n",
       "      <td>7.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>60.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>76.0</td>\n",
       "      <td>32.0</td>\n",
       "      <td>70.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>22.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>20.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>26.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>23.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>131.165948</td>\n",
       "      <td>411.347689</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1416.656061</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>210.972647</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>42.392075</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>117.746642</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>35.434240</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>316.437420</td>\n",
       "      <td>294.356180</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>378.216233</td>\n",
       "      <td>34.539829</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>421.193294</td>\n",
       "      <td>216.621431</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1706.023073</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6.651813</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>35043.643030</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>8460.544337</td>\n",
       "      <td>7.0</td>\n",
       "      <td>1425.707450</td>\n",
       "      <td>237.934884</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>442.722464</td>\n",
       "      <td>1293.152471</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4333.598862</td>\n",
       "      <td>58.131341</td>\n",
       "      <td>9.0</td>\n",
       "      <td>137.65366</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>30.453683</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>407.527435</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.234178</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>190.736930</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>108.424011</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>240.583003</td>\n",
       "      <td>155.338766</td>\n",
       "      <td>NaN</td>\n",
       "      <td>648.079128</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7.0</td>\n",
       "      <td>651.794322</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>832.122662</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1663.238330</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>51.475337</td>\n",
       "      <td>NaN</td>\n",
       "      <td>991.632283</td>\n",
       "      <td>7.0</td>\n",
       "      <td>71.715674</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1679.613800</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>12.854454</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9538.245656</td>\n",
       "      <td>1210.802212</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>16998.291564</td>\n",
       "      <td>1854.095011</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>17864.856294</td>\n",
       "      <td>736.949861</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>69.597897</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1636.070094</td>\n",
       "      <td>6.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>279.048253</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>743.245025</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4.0</td>\n",
       "      <td>104.784828</td>\n",
       "      <td>NaN</td>\n",
       "      <td>113.0</td>\n",
       "      <td>65.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>42.0</td>\n",
       "      <td>14.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>10.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>856.579037</td>\n",
       "      <td>8992.835847</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4914.387768</td>\n",
       "      <td>14.0</td>\n",
       "      <td>663.142041</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1665.378471</td>\n",
       "      <td>NaN</td>\n",
       "      <td>15990.279906</td>\n",
       "      <td>600.864642</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>20.936036</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>61.722697</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>13.195965</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>289.356804</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>82.225959</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>67.321531</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>206.805669</td>\n",
       "      <td>NaN</td>\n",
       "      <td>985.379711</td>\n",
       "      <td>301.601374</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-0.08862</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>11.0</td>\n",
       "      <td>5.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>341.232519</td>\n",
       "      <td>29.690702</td>\n",
       "      <td>NaN</td>\n",
       "      <td>332.733042</td>\n",
       "      <td>NaN</td>\n",
       "      <td>389.682415</td>\n",
       "      <td>456.246166</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>465.404972</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>228.235798</td>\n",
       "      <td>NaN</td>\n",
       "      <td>639.404908</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>195.927226</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>20.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>23.493134</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1441.066796</td>\n",
       "      <td>11.0</td>\n",
       "      <td>89.619418</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1592.704570</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>68.188554</td>\n",
       "      <td>2.0</td>\n",
       "      <td>210.38563</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>7.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9021.667138</td>\n",
       "      <td>549.407033</td>\n",
       "      <td>28.907638</td>\n",
       "      <td>NaN</td>\n",
       "      <td>80.552285</td>\n",
       "      <td>243.054746</td>\n",
       "      <td>3486.691576</td>\n",
       "      <td>496.890829</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>10976.344743</td>\n",
       "      <td>505.726186</td>\n",
       "      <td>60.994215</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2026.054832</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>134.608140</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>404.605895</td>\n",
       "      <td>17881.204613</td>\n",
       "      <td>570.521973</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>30.906907</td>\n",
       "      <td>NaN</td>\n",
       "      <td>533.725164</td>\n",
       "      <td>2.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>143.967128</td>\n",
       "      <td>NaN</td>\n",
       "      <td>64.0</td>\n",
       "      <td>21.0</td>\n",
       "      <td>33.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>40.0</td>\n",
       "      <td>90.0</td>\n",
       "      <td>62.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>15.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>12.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>4.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   client_id  is_in_club  cnt_checks_10_Мясная гастрономия  rto_std_11         rto_6  rto_12_Сыры  rto_12_Птица и изделия из птицы         rto_8  cnt_checks_9  rto_9_Мясная гастрономия  rto_10_Птица и изделия из птицы  cnt_checks_9_Рыба и рыбные изделия  rto_stddev_6_Рыба и рыбные изделия  rto_12_Мясная гастрономия  rto_stddev_8_Крупы и зерновые        rto_12  rto_6_Птица и изделия из птицы  rto_6_Рыба и рыбные изделия  cnt_checks_11  rto_stddev_7_Птица и изделия из птицы  rto_stddev_9_Сыры  cnt_checks_10_Овощи - Фрукты  cnt_checks_12_Птица и изделия из птицы  rto_6_Мясная гастрономия  rto_stddev_6_Овощи - Фрукты  cnt_checks_6_Мясная гастрономия  rto_stddev_6_Мясная гастрономия  cnt_checks_11_Сыры  cnt_checks_12_Овощи - Фрукты  rto_stddev_10_Птица и изделия из птицы  cnt_checks_6_Сыры  rto_stddev_10_Овощи - Фрукты  rto_12_Крупы и зерновые  rto_7_Птица и изделия из птицы  cnt_checks_12_Рыба и рыбные изделия  rto_stddev_10_Крупы и зерновые  rto_7_Крупы и зерновые  rto_7_Овощи - Фрукты  \\\n",
       "0          0           0                               NaN  300.021342   2418.116714   853.016746                              NaN   5092.170392           3.0                       NaN                              NaN                                 NaN                                 NaN                        NaN                            NaN   5738.687177                             NaN                          NaN            3.0                                    NaN                NaN                           NaN                                     NaN                       NaN                          NaN                              NaN                              NaN                 NaN                           2.0                                     NaN                NaN                           NaN                      NaN                             NaN                                  NaN                             NaN                     NaN                   NaN   \n",
       "1          1           0                               1.0  519.031384  19899.211736          NaN                              NaN  32392.636661           7.0                       NaN                       176.866878                                 2.0                                 NaN                        NaN                            NaN           NaN                             NaN                          NaN            2.0                                    NaN                NaN                          10.0                                     NaN                       NaN                   211.090724                              NaN                              NaN                 2.0                           NaN                                     NaN                3.0                     96.228636                      NaN                             NaN                                  NaN                             NaN                     NaN          11849.260237   \n",
       "2          2           0                               NaN  131.165948    411.347689          NaN                              NaN   1416.656061           NaN                       NaN                              NaN                                 NaN                                 NaN                        NaN                            NaN    210.972647                             NaN                          NaN            6.0                                    NaN                NaN                           NaN                                     NaN                       NaN                          NaN                              NaN                              NaN                 NaN                           NaN                                     NaN                NaN                           NaN                      NaN                             NaN                                  NaN                             NaN                     NaN                   NaN   \n",
       "3          3           0                               2.0         NaN  35043.643030          NaN                              NaN   8460.544337           7.0               1425.707450                       237.934884                                 NaN                                 NaN                        NaN                            NaN           NaN                      442.722464                  1293.152471            1.0                                    NaN                NaN                           2.0                                     NaN               4333.598862                    58.131341                              9.0                        137.65366                 NaN                           NaN                                     NaN                NaN                     30.453683                      NaN                             NaN                                  NaN                             NaN                     NaN            407.527435   \n",
       "4          4           0                               1.0  856.579037   8992.835847          NaN                              NaN   4914.387768          14.0                663.142041                              NaN                                 NaN                                 NaN                1665.378471                            NaN  15990.279906                      600.864642                          NaN           18.0                                    NaN          20.936036                           2.0                                     NaN                       NaN                    61.722697                              NaN                              NaN                 2.0                           4.0                                     NaN                2.0                     13.195965                      NaN                             NaN                                  NaN                             NaN                     NaN            289.356804   \n",
       "\n",
       "   cnt_checks_10_Птица и изделия из птицы  rto_9_Рыба и рыбные изделия  cnt_checks_11_Крупы и зерновые  rto_stddev_11_Крупы и зерновые  rto_stddev_6_Крупы и зерновые  rto_stddev_12_Овощи - Фрукты  cnt_checks_11_Рыба и рыбные изделия  rto_stddev_9_Крупы и зерновые  rto_11_Крупы и зерновые  rto_stddev_10_Сыры  cnt_checks_12_Мясная гастрономия  rto_stddev_10_Мясная гастрономия  rto_stddev_8_Рыба и рыбные изделия  cnt_checks_9_Крупы и зерновые  rto_10_Сыры  rto_stddev_6_Сыры  cnt_checks_7_Сыры  cnt_checks_6_Птица и изделия из птицы  cnt_checks_6_Рыба и рыбные изделия  rto_8_Овощи - Фрукты  cnt_checks_10_Крупы и зерновые  rto_12_Овощи - Фрукты  rto_11_Мясная гастрономия  cnt_checks_12_Крупы и зерновые  rto_stddev_11_Мясная гастрономия  rto_stddev_12_Птица и изделия из птицы  rto_stddev_7_Крупы и зерновые  cnt_checks_6  cnt_checks_9_Овощи - Фрукты  cnt_checks_7_Крупы и зерновые  cnt_checks_8_Птица и изделия из птицы  rto_11_Птица и изделия из птицы  rto_stddev_11_Сыры  \\\n",
       "0                                     NaN                          NaN                             NaN                             NaN                            NaN                     41.017712                                  NaN                            NaN                      NaN                 NaN                               NaN                               NaN                                 NaN                            NaN          NaN                NaN                NaN                                    NaN                                 NaN             73.644488                             NaN            1012.977676                        NaN                             NaN                               NaN                                     NaN                            NaN           5.0                          NaN                            NaN                                    NaN                              NaN                 NaN   \n",
       "1                                     1.0                   975.070527                             NaN                             NaN                            NaN                           NaN                                  1.0                            NaN                      NaN           160.83624                               NaN                               NaN                                 NaN                            1.0   564.287939         238.798235                3.0                                    NaN                                 NaN          14559.692214                             NaN                    NaN                        NaN                             NaN                               NaN                                     NaN                            NaN          12.0                          5.0                            NaN                                    NaN                       228.800923          220.173709   \n",
       "2                                     NaN                          NaN                             NaN                             NaN                            NaN                           NaN                                  NaN                            NaN                      NaN                 NaN                               NaN                               NaN                                 NaN                            NaN          NaN                NaN                NaN                                    NaN                                 NaN             42.392075                             NaN                    NaN                        NaN                             NaN                               NaN                                     NaN                            NaN           2.0                          NaN                            NaN                                    1.0                              NaN                 NaN   \n",
       "3                                     1.0                          NaN                             NaN                             NaN                            NaN                           NaN                                  NaN                            NaN                      NaN                 NaN                               NaN                          3.234178                                 NaN                            NaN          NaN                NaN                NaN                                    1.0                                 1.0            190.736930                             NaN                    NaN                 108.424011                             NaN                               NaN                                     NaN                            NaN          11.0                          1.0                            NaN                                    NaN                              NaN                 NaN   \n",
       "4                                     NaN                          NaN                             NaN                             NaN                            NaN                     82.225959                                  NaN                            NaN                      NaN                 NaN                               4.0                               NaN                                 NaN                            NaN          NaN          67.321531                1.0                                    2.0                                 NaN            206.805669                             NaN             985.379711                 301.601374                             NaN                          -0.08862                                     NaN                            NaN          11.0                          5.0                            NaN                                    NaN                       341.232519           29.690702   \n",
       "\n",
       "   cnt_checks_7_Птица и изделия из птицы  rto_10_Овощи - Фрукты  rto_stddev_7_Мясная гастрономия  rto_9_Сыры  rto_10_Мясная гастрономия  rto_stddev_12_Рыба и рыбные изделия  cnt_checks_7  rto_7_Мясная гастрономия  cnt_checks_9_Мясная гастрономия  cnt_checks_7_Мясная гастрономия  rto_9_Овощи - Фрукты  rto_stddev_9_Птица и изделия из птицы    rto_std_9  rto_stddev_10_Рыба и рыбные изделия  rto_stddev_9_Рыба и рыбные изделия  cnt_checks_9_Сыры  rto_stddev_6_Птица и изделия из птицы  rto_10_Рыба и рыбные изделия  rto_12_Рыба и рыбные изделия  rto_10_Крупы и зерновые  cnt_checks_12  cnt_checks_10_Рыба и рыбные изделия  rto_stddev_9_Овощи - Фрукты  rto_11_Рыба и рыбные изделия  rto_6_Овощи - Фрукты  cnt_checks_10  rto_stddev_7_Овощи - Фрукты  cnt_checks_12_Сыры    rto_std_7  cnt_checks_8_Овощи - Фрукты  rto_stddev_8_Мясная гастрономия  rto_stddev_11_Овощи - Фрукты  cnt_checks_7_Овощи - Фрукты  rto_9_Птица и изделия из птицы  cnt_checks_6_Крупы и зерновые  cnt_checks_8_Крупы и зерновые  \\\n",
       "0                                    NaN                    NaN                              NaN         NaN                        NaN                                  NaN           1.0                       NaN                              NaN                              NaN                   NaN                                    NaN   470.600157                                  NaN                                 NaN                NaN                                    NaN                           NaN                           NaN                      NaN           11.0                                  NaN                          NaN                           NaN            124.564463            5.0                          NaN                 3.0          NaN                          1.0                              NaN                     81.154671                          NaN                             NaN                            1.0                            NaN   \n",
       "1                                    NaN            3522.302011                              NaN  108.200176                 148.900417                                  NaN          16.0                276.814607                              NaN                              1.0           1156.885192                                    NaN   524.139352                           370.327074                           192.14527                1.0                                    NaN                   4228.863632                           NaN                      NaN            NaN                                  5.0                   102.446072                    374.702932           7025.919801           11.0                   203.249441                 NaN   829.838894                         19.0                              NaN                     98.340478                         15.0                             NaN                            NaN                            NaN   \n",
       "2                                    NaN                    NaN                              NaN         NaN                        NaN                                  NaN           3.0                       NaN                              NaN                              NaN                   NaN                                    NaN          NaN                                  NaN                                 NaN                NaN                                    NaN                           NaN                           NaN                      NaN            1.0                                  NaN                          NaN                           NaN                   NaN            1.0                          NaN                 NaN   117.746642                          1.0                              NaN                     35.434240                          NaN                             NaN                            NaN                            NaN   \n",
       "3                                    NaN             240.583003                       155.338766         NaN                 648.079128                                  NaN           7.0                651.794322                              2.0                              2.0            832.122662                                    NaN  1663.238330                                  NaN                                 NaN                NaN                                    NaN                           NaN                           NaN                      NaN            NaN                                  NaN                    51.475337                           NaN            991.632283            7.0                    71.715674                 NaN  1679.613800                          1.0                              NaN                     12.854454                          2.0                             NaN                            NaN                            NaN   \n",
       "4                                    NaN             332.733042                              NaN  389.682415                 456.246166                                  NaN           2.0                465.404972                              2.0                              1.0            228.235798                                    NaN   639.404908                                  NaN                                 NaN                2.0                             195.927226                           NaN                           NaN                      NaN           20.0                                  NaN                    23.493134                           NaN           1441.066796           11.0                    89.619418                 NaN  1592.704570                          1.0                              NaN                     68.188554                          2.0                       210.38563                            NaN                            NaN   \n",
       "\n",
       "   cnt_checks_7_Рыба и рыбные изделия  cnt_checks_11_Овощи - Фрукты  cnt_checks_8_Рыба и рыбные изделия  cnt_checks_11_Мясная гастрономия  rto_8_Мясная гастрономия        rto_10    rto_std_8  rto_stddev_11_Птица и изделия из птицы  rto_stddev_8_Птица и изделия из птицы  rto_stddev_12_Мясная гастрономия   rto_6_Сыры         rto_7    rto_std_6  cnt_checks_9_Птица и изделия из птицы  rto_stddev_11_Рыба и рыбные изделия         rto_9  rto_std_10  rto_stddev_8_Сыры  rto_8_Птица и изделия из птицы  rto_11_Овощи - Фрукты  cnt_checks_11_Птица и изделия из птицы  rto_7_Рыба и рыбные изделия  cnt_checks_6_Овощи - Фрукты  rto_9_Крупы и зерновые   rto_7_Сыры  rto_8_Крупы и зерновые  cnt_checks_8_Мясная гастрономия  rto_stddev_7_Сыры   rto_8_Сыры        rto_11  rto_std_12  cnt_checks_10_Сыры  rto_stddev_12_Сыры  rto_6_Крупы и зерновые  rto_stddev_7_Рыба и рыбные изделия  rto_stddev_8_Овощи - Фрукты  rto_8_Рыба и рыбные изделия  rto_11_Сыры  cnt_checks_8_Сыры  cnt_checks_8  \\\n",
       "0                                 NaN                           1.0                                 NaN                               NaN                       NaN   1583.623070   839.257740                                     NaN                                    NaN                               NaN          NaN    440.974542   357.220317                                    NaN                                  NaN   1658.894353  184.518332                NaN                             NaN             482.557307                                     NaN                          NaN                          1.0                     NaN          NaN                     NaN                              NaN                NaN          NaN   1720.373691  675.763593                 NaN           73.846425              148.316558                                 NaN                          NaN                          NaN          NaN                NaN           5.0   \n",
       "1                                 2.0                           2.0                                 1.0                               NaN                       NaN  19329.981907   685.952065                                     NaN                                    NaN                               NaN  1597.936059  28890.382271   758.343961                                    NaN                                  NaN   6337.074013  765.052043          74.362466                             NaN             572.071018                                     1.0                  2001.937586                          9.0              160.913398  1556.301144                     NaN                              NaN         158.711615  1926.327565   4327.469741         NaN                 2.0                 NaN                     NaN                          225.219998                   301.404157                   525.868862  1153.293331                7.0          21.0   \n",
       "2                                 NaN                           5.0                                 NaN                               NaN                       NaN    316.437420   294.356180                                     NaN                                    NaN                               NaN          NaN    378.216233    34.539829                                    NaN                                  NaN           NaN         NaN                NaN                      421.193294             216.621431                                     NaN                          NaN                          NaN                     NaN          NaN                     NaN                              NaN                NaN          NaN   1706.023073         NaN                 NaN                 NaN                     NaN                                 NaN                     6.651813                          NaN          NaN                NaN           5.0   \n",
       "3                                 1.0                           1.0                                 1.0                               1.0                       NaN   9538.245656  1210.802212                                     NaN                                    NaN                               NaN          NaN  16998.291564  1854.095011                                    NaN                                  NaN  17864.856294  736.949861                NaN                             NaN              69.597897                                     NaN                  1636.070094                          6.0                     NaN          NaN                     NaN                              NaN                NaN          NaN    279.048253         NaN                 NaN                 NaN                     NaN                                 NaN                          NaN                   743.245025          NaN                NaN           4.0   \n",
       "4                                 NaN                           7.0                                 NaN                               2.0                       NaN   9021.667138   549.407033                               28.907638                                    NaN                         80.552285   243.054746   3486.691576   496.890829                                    1.0                                  NaN  10976.344743  505.726186          60.994215                             NaN            2026.054832                                     1.0                          NaN                          4.0                     NaN   134.608140                     NaN                              NaN                NaN   404.605895  17881.204613  570.521973                 NaN                 NaN                     NaN                                 NaN                    30.906907                          NaN   533.725164                2.0           6.0   \n",
       "\n",
       "   rto_stddev_9_Мясная гастрономия  rto_stddev_12_Крупы и зерновые  cnt_unique_plu_6  cnt_unique_plu_7  cnt_unique_plu_8  cnt_unique_plu_9  cnt_unique_plu_10  cnt_unique_plu_11  cnt_unique_plu_12  cnt_unique_plu_10_Мясная гастрономия  cnt_unique_plu_10_Овощи - Фрукты  cnt_unique_plu_10_Птица и изделия из птицы  cnt_unique_plu_10_Рыба и рыбные изделия  cnt_unique_plu_11_Мясная гастрономия  cnt_unique_plu_11_Овощи - Фрукты  cnt_unique_plu_11_Птица и изделия из птицы  cnt_unique_plu_11_Рыба и рыбные изделия  cnt_unique_plu_12_Мясная гастрономия  cnt_unique_plu_12_Овощи - Фрукты  cnt_unique_plu_12_Птица и изделия из птицы  cnt_unique_plu_12_Рыба и рыбные изделия  cnt_unique_plu_6_Мясная гастрономия  cnt_unique_plu_6_Овощи - Фрукты  cnt_unique_plu_6_Птица и изделия из птицы  cnt_unique_plu_6_Рыба и рыбные изделия  cnt_unique_plu_7_Мясная гастрономия  cnt_unique_plu_7_Овощи - Фрукты  cnt_unique_plu_7_Птица и изделия из птицы  cnt_unique_plu_7_Рыба и рыбные изделия  \\\n",
       "0                              NaN                             NaN              18.0               8.0              40.0               8.0               15.0               15.0               38.0                                   NaN                               NaN                                         NaN                                      NaN                                   NaN                               3.0                                         NaN                                      NaN                                   NaN                               6.0                                         NaN                                      NaN                                  NaN                              1.0                                        NaN                                     NaN                                  NaN                              NaN                                        NaN                                     NaN   \n",
       "1                              NaN                             NaN              60.0              60.0              76.0              32.0               70.0               20.0                NaN                                   1.0                              22.0                                         1.0                                      5.0                                   NaN                               3.0                                         1.0                                      1.0                                   NaN                               NaN                                         NaN                                      NaN                                  NaN                             20.0                                        NaN                                     NaN                                  1.0                             26.0                                        NaN                                     2.0   \n",
       "2                              NaN                             NaN               4.0               4.0              17.0               NaN                1.0               10.0                2.0                                   NaN                               NaN                                         NaN                                      NaN                                   NaN                               3.0                                         NaN                                      NaN                                   NaN                               NaN                                         NaN                                      NaN                                  NaN                              NaN                                        NaN                                     NaN                                  NaN                              NaN                                        NaN                                     NaN   \n",
       "3                       104.784828                             NaN             113.0              65.0              37.0              42.0               14.0                9.0                NaN                                   2.0                               2.0                                         1.0                                      NaN                                   1.0                               2.0                                         NaN                                      NaN                                   NaN                               NaN                                         NaN                                      NaN                                 10.0                             10.0                                        1.0                                     1.0                                  3.0                              4.0                                        NaN                                     1.0   \n",
       "4                       143.967128                             NaN              64.0              21.0              33.0              60.0               40.0               90.0               62.0                                   1.0                               6.0                                         NaN                                      NaN                                   2.0                              15.0                                         2.0                                      NaN                                   3.0                               7.0                                         NaN                                      NaN                                  NaN                             12.0                                        2.0                                     NaN                                  1.0                              3.0                                        NaN                                     NaN   \n",
       "\n",
       "   cnt_unique_plu_8_Мясная гастрономия  cnt_unique_plu_8_Овощи - Фрукты  cnt_unique_plu_8_Птица и изделия из птицы  cnt_unique_plu_8_Рыба и рыбные изделия  cnt_unique_plu_9_Мясная гастрономия  cnt_unique_plu_9_Овощи - Фрукты  cnt_unique_plu_9_Птица и изделия из птицы  cnt_unique_plu_9_Рыба и рыбные изделия  cnt_unique_plu_10_Крупы и зерновые  cnt_unique_plu_10_Сыры  cnt_unique_plu_11_Крупы и зерновые  cnt_unique_plu_11_Сыры  cnt_unique_plu_12_Крупы и зерновые  cnt_unique_plu_12_Сыры  cnt_unique_plu_6_Крупы и зерновые  cnt_unique_plu_6_Сыры  cnt_unique_plu_7_Крупы и зерновые  cnt_unique_plu_7_Сыры  cnt_unique_plu_8_Крупы и зерновые  cnt_unique_plu_8_Сыры  cnt_unique_plu_9_Крупы и зерновые  cnt_unique_plu_9_Сыры  \n",
       "0                                  NaN                              1.0                                        NaN                                     NaN                                  NaN                              NaN                                        NaN                                     NaN                                 NaN                     NaN                                 NaN                     NaN                                 NaN                     3.0                                1.0                    NaN                                NaN                    NaN                                NaN                    NaN                                NaN                    NaN  \n",
       "1                                  NaN                             23.0                                        NaN                                     1.0                                  NaN                              8.0                                        NaN                                     1.0                                 NaN                     2.0                                 NaN                     3.0                                 NaN                     NaN                                NaN                    2.0                                NaN                    3.0                                NaN                    4.0                                1.0                    1.0  \n",
       "2                                  NaN                              2.0                                        1.0                                     NaN                                  NaN                              NaN                                        NaN                                     NaN                                 NaN                     NaN                                 NaN                     NaN                                 NaN                     NaN                                NaN                    NaN                                NaN                    NaN                                NaN                    NaN                                NaN                    NaN  \n",
       "3                                  NaN                              1.0                                        NaN                                     1.0                                  4.0                              4.0                                        NaN                                     NaN                                 NaN                     NaN                                 NaN                     NaN                                 NaN                     NaN                                NaN                    NaN                                NaN                    NaN                                NaN                    NaN                                NaN                    NaN  \n",
       "4                                  NaN                              4.0                                        NaN                                     NaN                                  2.0                              4.0                                        1.0                                     NaN                                 NaN                     NaN                                 NaN                     4.0                                 NaN                     NaN                                NaN                    2.0                                NaN                    1.0                                NaN                    2.0                                NaN                    2.0  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "2699d134",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:08.677157Z",
     "start_time": "2022-03-27T18:40:54.854403Z"
    }
   },
   "outputs": [],
   "source": [
    "PT = PowerTransformer()\n",
    "\n",
    "x = PT.fit_transform(df.drop(['client_id', 'is_in_club'], axis=1))\n",
    "x = pd.DataFrame(data=x, columns=df.columns[2:])\n",
    "x['is_in_club'] = df['is_in_club']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "8ce5db87",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:08.861458Z",
     "start_time": "2022-03-27T18:41:08.679752Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.boxplot(data=x, x=\"is_in_club\", y='cnt_checks_12_Овощи - Фрукты', palette=\"tab10\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "715aaeb8",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:09.071654Z",
     "start_time": "2022-03-27T18:41:08.863986Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x[x.is_in_club==1]['cnt_checks_12_Овощи - Фрукты'].hist();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "95700fd8",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:09.440758Z",
     "start_time": "2022-03-27T18:41:09.074968Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x[x.is_in_club==0]['cnt_checks_12_Овощи - Фрукты'].hist();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "9e88bb77",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:09.846437Z",
     "start_time": "2022-03-27T18:41:09.445773Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x[x.is_in_club==0]['cnt_checks_12_Рыба и рыбные изделия'].hist();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "1ea21978",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:10.063666Z",
     "start_time": "2022-03-27T18:41:09.848969Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x[x.is_in_club==1]['cnt_checks_12_Рыба и рыбные изделия'].hist();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "bd4643a1",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:10.843430Z",
     "start_time": "2022-03-27T18:41:10.065673Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 7 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cotinuous_column_list = [ 'rto_6', 'rto_7', 'rto_8', 'rto_9',\n",
    "                         'rto_10', 'rto_11', 'rto_12']\n",
    "fig = plt.figure(figsize=(10,10))\n",
    "# fig, ax = plt.subplots(3,3, figsize=(15,20))\n",
    "row, col = 3, 3\n",
    "for i in range(1,8):\n",
    "    ax = fig.add_subplot(row, col, i)\n",
    "    sns.boxplot(data=x, x=\"is_in_club\", y=cotinuous_column_list[i-1], ax = ax, palette=\"tab10\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fa386924",
   "metadata": {},
   "source": [
    "- Инфляция"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "2625a857",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:11.062883Z",
     "start_time": "2022-03-27T18:41:10.846638Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import re\n",
    "\n",
    "infliation = {6: 7.9,\n",
    "              7: 7.43,\n",
    "              8: 7.7,\n",
    "              9: 9.21,\n",
    "              10: 10.89,\n",
    "              11: 10.81,\n",
    "              12: 10.62\n",
    "              }\n",
    "\n",
    "plt.plot(list(infliation.keys()), list(infliation.values()), color = 'green')\n",
    "plt.xlabel('month', fontsize=12)\n",
    "plt.ylabel('infliation, %', fontsize=12)\n",
    "plt.title('Infliation', fontsize=17)\n",
    "\n",
    "rto_columns = [ 'rto_6', 'rto_7', 'rto_8', 'rto_9',\n",
    "                         'rto_10', 'rto_11', 'rto_12']\n",
    "                         \n",
    "for cat in (rto_columns):\n",
    "    month = int(re.findall(r'\\d+', cat)[0])\n",
    "    inf_coef = 1+infliation[month]/100\n",
    "    df[cat] = df[cat]*inf_coef"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "92fc298f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:11.084814Z",
     "start_time": "2022-03-27T18:41:11.065314Z"
    }
   },
   "outputs": [],
   "source": [
    "rto_columns = [ 'rto_6', 'rto_7', 'rto_8', 'rto_9',\n",
    "                         'rto_10', 'rto_11', 'rto_12']\n",
    "for cat in (rto_columns):\n",
    "    month = int(re.findall(r'\\d+', cat)[0])\n",
    "    inf_coef = 1+infliation[month]/100\n",
    "    df[cat+'_inf_diff'] = df[cat]*inf_coef - df[cat]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d1e5ca6a",
   "metadata": {},
   "source": [
    "- quantile"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "27ef7c8d",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:11.196871Z",
     "start_time": "2022-03-27T18:41:11.086806Z"
    }
   },
   "outputs": [],
   "source": [
    "categories = ['rto_6', 'rto_7', 'rto_8', 'rto_9', 'rto_10', 'rto_11', 'rto_12']       \n",
    "for i, cat in enumerate(categories):\n",
    "    df[f'cat_rto_{i}'] = (df[cat] > df[cat].quantile(0.8)).astype('int')\n",
    "    \n",
    "categories = ['rto_6_Овощи - Фрукты', 'rto_7_Овощи - Фрукты', 'rto_8_Овощи - Фрукты', \n",
    "              'rto_9_Овощи - Фрукты', 'rto_10_Овощи - Фрукты', 'rto_11_Овощи - Фрукты', 'rto_12_Овощи - Фрукты'] \n",
    "for i, cat in enumerate(categories):\n",
    "    df[f'cat_fruit_rto_{i}'] = (df[cat] > df[cat].quantile(0.8)).astype('int')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cea78ab9",
   "metadata": {},
   "source": [
    "- Рост зарплаты"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "fbc395fa",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:11.462226Z",
     "start_time": "2022-03-27T18:41:11.199983Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "salary_growth = {\n",
    "    6: 4.9,\n",
    "    7: 2.2,\n",
    "    8: 1.5,\n",
    "    9: 2,\n",
    "    10: 0.6,\n",
    "    11: 3.4,\n",
    "    12: 3.6\n",
    "}\n",
    "\n",
    "plt.plot(list(salary_growth.keys()), list(salary_growth.values()), color = 'green')\n",
    "plt.xlabel('month', fontsize=12)\n",
    "plt.ylabel('salary growth, %', fontsize=12)\n",
    "plt.title('Salary growth', fontsize=17)\n",
    "\n",
    "rto_columns = [ 'rto_6', 'rto_7', 'rto_8', 'rto_9',\n",
    "                         'rto_10', 'rto_11', 'rto_12']\n",
    "for cat in (rto_columns):\n",
    "    month = int(re.findall(r'\\d+', cat)[0])\n",
    "    sal_coef = 1+salary_growth[month]/100\n",
    "    df[cat+ '_sal_diff'] = df[cat]*sal_coef - df[cat]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "10705e38",
   "metadata": {},
   "source": [
    "## Data Preprocess"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "35ceb4b2",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:11.591833Z",
     "start_time": "2022-03-27T18:41:11.464746Z"
    }
   },
   "outputs": [],
   "source": [
    "def data_preprocess(df):\n",
    "    '''\n",
    "    Fill nans with zeros, add features:\n",
    "    cnt_rto_n_cat - сумма покупок клиента по всем категориям в разрезе месяца\n",
    "    rto_n_category_part - доля данной категории от всех покупок в разрезе месяца\n",
    "    '''\n",
    "    df.fillna(0, inplace=True)\n",
    "\n",
    "    for col in df.columns:\n",
    "        if df[df[col] < 0].shape[0] > 0 and col.find('std') == -1:\n",
    "            df.loc[df[col] < 0, col] = 0\n",
    "\n",
    "    rto_n = df.columns[df.columns.str.contains(r\"^rto_[0-9][0-9]*$\")].sort_values(ascending = True)\n",
    "    rto_n_category = df.columns[df.columns.str.contains(r\"^rto_[0-9][0-9]*_[^0-9]+$\")].sort_values(ascending = True)\n",
    "    df[rto_n_category] = df[rto_n_category].fillna(0)\n",
    "\n",
    "    for i in range(7):\n",
    "        for rto_n_cat in rto_n_category[i*6:i*6+6]:\n",
    "            df[\"cnt_rto_\" + str(i) + \"_cat\" ] = (df[rto_n_category[0:6]]>0).sum(axis = 1)\n",
    "\n",
    "    for i,r_n in enumerate(rto_n):\n",
    "        for rto_n_cat in rto_n_category[i*6:i*6+6]:\n",
    "            df[rto_n_cat + '_part'] = df[rto_n_cat]/df[r_n]\n",
    "\n",
    "    for col in df.columns:\n",
    "        if col.find('_part')!= -1:\n",
    "            df.loc[df[col] == np.inf, col] = 0\n",
    "    \n",
    "    df = df.fillna(0)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "a3e23164",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:13.763383Z",
     "start_time": "2022-03-27T18:41:11.594759Z"
    }
   },
   "outputs": [],
   "source": [
    "df = data_preprocess(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "ba25d451",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:13.887661Z",
     "start_time": "2022-03-27T18:41:13.765748Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0, (135061, 275))"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sum(df.isna().sum() != 0), df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "642c66fc",
   "metadata": {},
   "source": [
    "## Split data on train and test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "887b3a97",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:20.012384Z",
     "start_time": "2022-03-27T18:41:19.451851Z"
    }
   },
   "outputs": [],
   "source": [
    "TEST_SIZE = .3\n",
    "RANDOM_STATE = 42\n",
    "x = df.drop(['client_id', 'is_in_club'], axis=1)\n",
    "y = df.is_in_club\n",
    "\n",
    "X_train_val, X_test, y_train_val, y_test = train_test_split(x, \n",
    "                                                            y, \n",
    "                                                            test_size=TEST_SIZE, \n",
    "                                                            shuffle=True, \n",
    "                                                            stratify=y,\n",
    "                                                            random_state=RANDOM_STATE)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1fb41049",
   "metadata": {},
   "source": [
    "## CatBoostClassifier "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c06e6b80",
   "metadata": {},
   "source": [
    " - Hyperparams opt (подбираем гиперпараметры на кросс валидации с искользованием библиотеки optuna)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "6f3babf9",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:21.647448Z",
     "start_time": "2022-03-27T18:41:21.644255Z"
    }
   },
   "outputs": [],
   "source": [
    "N_SPLITS = 3\n",
    "RANDOM_STATE = 42 \n",
    "FOLD = StratifiedKFold(n_splits=N_SPLITS, shuffle=True, random_state=RANDOM_STATE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "165b1e1c",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:30.347068Z",
     "start_time": "2022-03-27T18:41:30.342374Z"
    }
   },
   "outputs": [],
   "source": [
    "def hyper_opt_loss(train_loss, val_loss, beta = 1.5):\n",
    "    return -val_loss * np.exp(- beta * np.abs(val_loss - train_loss))\n",
    "\n",
    "def opt_loss(train_loss, val_loss):\n",
    "    return hyper_opt_loss(train_loss, val_loss)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "9c61c6ce",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:37.789278Z",
     "start_time": "2022-03-27T18:41:37.781784Z"
    }
   },
   "outputs": [],
   "source": [
    "def train_cb(params, \n",
    "             train_index, \n",
    "             val_index, \n",
    "             X_train_val,\n",
    "             y_train_val):\n",
    "    \n",
    "    train_pool = Pool(X_train_val.iloc[train_index, :], label=y_train_val.iloc[train_index])\n",
    "    val_pool = Pool(X_train_val.iloc[val_index, :], label=y_train_val.iloc[val_index])\n",
    "    \n",
    "    class_wei =  y_train_val.iloc[train_index].value_counts()[0] / y_train_val.iloc[train_index].value_counts()[1]\n",
    "    model = CatBoostClassifier(**params, \n",
    "                               class_weights = (1, class_wei),\n",
    "                               custom_metric='AUC:hints=skip_train~false',\n",
    "                               eval_metric='AUC',\n",
    "                               #task_type=\"GPU\",\n",
    "                               random_seed=42, \n",
    "                               early_stopping_rounds=20,\n",
    "                               thread_count=16,\n",
    "    )\n",
    "    model.fit(train_pool, verbose=0, eval_set=val_pool)\n",
    "\n",
    "    train_loss = model.best_score_['learn']['AUC']\n",
    "    val_loss = model.best_score_['validation']['AUC']\n",
    "    \n",
    "    output = {\n",
    "        'model': model,\n",
    "        'train_loss': train_loss,\n",
    "        'val_loss': val_loss\n",
    "    }\n",
    "    \n",
    "    return output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "2ac89c24",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:43.027451Z",
     "start_time": "2022-03-27T18:41:43.021677Z"
    }
   },
   "outputs": [],
   "source": [
    "def stat_score(params):\n",
    "    train_losses, val_losses = [], []\n",
    "    for train, validation in FOLD.split(X_train_val, y_train_val):\n",
    "        result = train_cb(params, train, validation, X_train_val, y_train_val)    \n",
    "        train_loss, val_loss = result['train_loss'], result['val_loss']\n",
    "        train_losses += [train_loss]\n",
    "        val_losses += [val_loss]\n",
    "        \n",
    "    train_losses = np.array(train_losses)\n",
    "    val_losses = np.array(val_losses)\n",
    "    \n",
    "    output = {\n",
    "        'train': train_losses,\n",
    "        'val': val_losses\n",
    "    }\n",
    "    return output\n",
    "\n",
    "def score(params):\n",
    "    stat = stat_score(params)\n",
    "    \n",
    "    train_loss = stat['train'].mean()\n",
    "    val_loss = stat['val'].mean()\n",
    "    \n",
    "    return opt_loss(train_loss, val_loss)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "cc225923",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:45.272377Z",
     "start_time": "2022-03-27T18:41:45.264827Z"
    }
   },
   "outputs": [],
   "source": [
    "def objective(trial):\n",
    "    param = {\n",
    "        'n_estimators': trial.suggest_int('n_estimators', 200, 500),\n",
    "        \"learning_rate\": trial.suggest_loguniform(\"learning_rate\", 1e-5, 1e0),\n",
    "        \"l2_leaf_reg\": trial.suggest_loguniform(\"l2_leaf_reg\", 1e-2, 1e0),\n",
    "        \"colsample_bylevel\": trial.suggest_float(\"colsample_bylevel\", 0.01, 0.1),\n",
    "        \"depth\": trial.suggest_int(\"depth\", 1, 10),\n",
    "        \"boosting_type\": trial.suggest_categorical(\"boosting_type\", [\"Ordered\", \"Plain\"]),\n",
    "        \"bootstrap_type\": trial.suggest_categorical(\"bootstrap_type\", [\"Bayesian\", \"Bernoulli\", \"MVS\"]),\n",
    "        \"min_data_in_leaf\": trial.suggest_int(\"min_data_in_leaf\", 2, 20),\n",
    "        \"one_hot_max_size\": trial.suggest_int(\"one_hot_max_size\", 2, 20), \n",
    "        'max_ctr_complexity':  trial.suggest_int('max_ctr_complexity', 0, 8), \n",
    "    } \n",
    "\n",
    "    if param[\"bootstrap_type\"] == \"Bayesian\":\n",
    "        param[\"bagging_temperature\"] = trial.suggest_float(\"bagging_temperature\", 0, 10)\n",
    "    elif param[\"bootstrap_type\"] == \"Bernoulli\":\n",
    "        param[\"subsample\"] = trial.suggest_float(\"subsample\", 0.1, 1)\n",
    "    return score(param)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "28822d0d",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T15:37:19.822693Z",
     "start_time": "2022-03-22T15:34:22.822112Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[32m[I 2022-03-22 18:34:22,826]\u001b[0m A new study created in memory with name: no-name-e4626435-775e-42ca-8d45-cd9338b00ac8\u001b[0m\n",
      "\u001b[32m[I 2022-03-22 18:34:28,403]\u001b[0m Trial 0 finished with value: -0.7015587779483291 and parameters: {'n_estimators': 452, 'learning_rate': 0.45886988140164003, 'l2_leaf_reg': 0.38026771974147267, 'colsample_bylevel': 0.05470951384808257, 'depth': 1, 'boosting_type': 'Plain', 'bootstrap_type': 'Bayesian', 'min_data_in_leaf': 11, 'one_hot_max_size': 16, 'max_ctr_complexity': 5, 'bagging_temperature': 0.0330344331790966}. Best is trial 0 with value: -0.7015587779483291.\u001b[0m\n",
      "\u001b[32m[I 2022-03-22 18:34:33,896]\u001b[0m Trial 1 finished with value: -0.7075607972474863 and parameters: {'n_estimators': 430, 'learning_rate': 0.004323515841480409, 'l2_leaf_reg': 0.126766630509968, 'colsample_bylevel': 0.019611151935028673, 'depth': 5, 'boosting_type': 'Plain', 'bootstrap_type': 'MVS', 'min_data_in_leaf': 16, 'one_hot_max_size': 11, 'max_ctr_complexity': 5}. Best is trial 1 with value: -0.7075607972474863.\u001b[0m\n",
      "\u001b[32m[I 2022-03-22 18:34:40,556]\u001b[0m Trial 2 finished with value: -0.7056077427309009 and parameters: {'n_estimators': 374, 'learning_rate': 4.434410742738529e-05, 'l2_leaf_reg': 0.08640433640669903, 'colsample_bylevel': 0.09773506799498918, 'depth': 4, 'boosting_type': 'Ordered', 'bootstrap_type': 'Bayesian', 'min_data_in_leaf': 2, 'one_hot_max_size': 8, 'max_ctr_complexity': 0, 'bagging_temperature': 8.149461553272097}. Best is trial 1 with value: -0.7075607972474863.\u001b[0m\n",
      "\u001b[32m[I 2022-03-22 18:35:34,022]\u001b[0m Trial 3 finished with value: -0.6962453527207604 and parameters: {'n_estimators': 245, 'learning_rate': 4.420481868081677e-05, 'l2_leaf_reg': 0.013937198077345453, 'colsample_bylevel': 0.05221651920635816, 'depth': 9, 'boosting_type': 'Ordered', 'bootstrap_type': 'Bernoulli', 'min_data_in_leaf': 5, 'one_hot_max_size': 12, 'max_ctr_complexity': 0, 'subsample': 0.96072309879189}. Best is trial 1 with value: -0.7075607972474863.\u001b[0m\n",
      "\u001b[32m[I 2022-03-22 18:35:59,656]\u001b[0m Trial 4 finished with value: -0.6587612895078604 and parameters: {'n_estimators': 246, 'learning_rate': 0.22885637711573112, 'l2_leaf_reg': 0.6219862658992136, 'colsample_bylevel': 0.09021096126621228, 'depth': 9, 'boosting_type': 'Ordered', 'bootstrap_type': 'Bernoulli', 'min_data_in_leaf': 15, 'one_hot_max_size': 4, 'max_ctr_complexity': 0, 'subsample': 0.39511469374909247}. Best is trial 1 with value: -0.7075607972474863.\u001b[0m\n",
      "\u001b[32m[I 2022-03-22 18:36:05,371]\u001b[0m Trial 5 finished with value: -0.7062478497603437 and parameters: {'n_estimators': 424, 'learning_rate': 2.2991348932685423e-05, 'l2_leaf_reg': 0.3247265224171686, 'colsample_bylevel': 0.018981142302126386, 'depth': 4, 'boosting_type': 'Ordered', 'bootstrap_type': 'Bayesian', 'min_data_in_leaf': 4, 'one_hot_max_size': 15, 'max_ctr_complexity': 7, 'bagging_temperature': 3.1507570041086206}. Best is trial 1 with value: -0.7075607972474863.\u001b[0m\n",
      "\u001b[32m[I 2022-03-22 18:36:07,873]\u001b[0m Trial 6 finished with value: -0.708373054624765 and parameters: {'n_estimators': 455, 'learning_rate': 6.178662590477509e-05, 'l2_leaf_reg': 0.024838087441953816, 'colsample_bylevel': 0.049304397029700825, 'depth': 2, 'boosting_type': 'Plain', 'bootstrap_type': 'Bernoulli', 'min_data_in_leaf': 7, 'one_hot_max_size': 16, 'max_ctr_complexity': 8, 'subsample': 0.6315772792144269}. Best is trial 6 with value: -0.708373054624765.\u001b[0m\n",
      "\u001b[32m[I 2022-03-22 18:36:10,692]\u001b[0m Trial 7 finished with value: -0.6688640732744414 and parameters: {'n_estimators': 317, 'learning_rate': 0.9644362691128544, 'l2_leaf_reg': 0.014706637716578274, 'colsample_bylevel': 0.013245158611782125, 'depth': 4, 'boosting_type': 'Plain', 'bootstrap_type': 'Bernoulli', 'min_data_in_leaf': 4, 'one_hot_max_size': 5, 'max_ctr_complexity': 6, 'subsample': 0.32789013165661146}. Best is trial 6 with value: -0.708373054624765.\u001b[0m\n",
      "\u001b[32m[I 2022-03-22 18:37:15,262]\u001b[0m Trial 8 finished with value: -0.6703742860635638 and parameters: {'n_estimators': 357, 'learning_rate': 0.058799115665915544, 'l2_leaf_reg': 0.7125543616844636, 'colsample_bylevel': 0.03449921561966607, 'depth': 9, 'boosting_type': 'Ordered', 'bootstrap_type': 'Bayesian', 'min_data_in_leaf': 6, 'one_hot_max_size': 17, 'max_ctr_complexity': 2, 'bagging_temperature': 2.851437223493236}. Best is trial 6 with value: -0.708373054624765.\u001b[0m\n",
      "\u001b[32m[I 2022-03-22 18:37:19,814]\u001b[0m Trial 9 finished with value: -0.7090247608101233 and parameters: {'n_estimators': 394, 'learning_rate': 0.00229646867747707, 'l2_leaf_reg': 0.027694654484639977, 'colsample_bylevel': 0.03264359223551126, 'depth': 2, 'boosting_type': 'Plain', 'bootstrap_type': 'MVS', 'min_data_in_leaf': 16, 'one_hot_max_size': 16, 'max_ctr_complexity': 2}. Best is trial 9 with value: -0.7090247608101233.\u001b[0m\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of completed trials: 10\n",
      "Best trial:\n",
      "\tBest Score: -0.7090247608101233\n",
      "\tBest Params: \n",
      "    n_estimators: 394\n",
      "    learning_rate: 0.00229646867747707\n",
      "    l2_leaf_reg: 0.027694654484639977\n",
      "    colsample_bylevel: 0.03264359223551126\n",
      "    depth: 2\n",
      "    boosting_type: Plain\n",
      "    bootstrap_type: MVS\n",
      "    min_data_in_leaf: 16\n",
      "    one_hot_max_size: 16\n",
      "    max_ctr_complexity: 2\n"
     ]
    }
   ],
   "source": [
    "study = optuna.create_study(sampler=TPESampler(), direction=\"minimize\")\n",
    "study.optimize(objective, n_trials=10) \n",
    "print(\"Number of completed trials: {}\".format(len(study.trials)))\n",
    "print(\"Best trial:\")\n",
    "trial = study.best_trial\n",
    "\n",
    "print(\"\\tBest Score: {}\".format(trial.value))\n",
    "print(\"\\tBest Params: \")\n",
    "for key, value in trial.params.items():\n",
    "    print(\"    {}: {}\".format(key, value))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "a2a941b5",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:57.372716Z",
     "start_time": "2022-03-27T18:41:57.364662Z"
    }
   },
   "outputs": [],
   "source": [
    "#best_params = trail.params\n",
    "# best_params = {\n",
    "#     'boosting_type': 'Ordered',\n",
    "#     'bootstrap_type': 'Bernoulli',\n",
    "#     'colsample_bylevel': 0.08512900052697547,\n",
    "#     'depth': 4,\n",
    "#     'l2_leaf_reg': 0.04335660768413505,\n",
    "#     'learning_rate': 0.005965363852834718,\n",
    "#     'max_ctr_complexity': 2,\n",
    "#     'min_data_in_leaf': 16,\n",
    "#     'n_estimators': 361,\n",
    "#     'one_hot_max_size': 9,\n",
    "#     'subsample': 0.9989517974264015\n",
    "# } #optuna best\n",
    "best_params = {\n",
    "    'boosting_type': 'Ordered',\n",
    "     'depth': 3,\n",
    "     'l2_leaf_reg': 4.757519417940943,\n",
    "     'learning_rate': 0.05561681993507123,\n",
    "     'max_ctr_complexity': 6,\n",
    "     'n_estimators': 306\n",
    "}#hyperopt best "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4f3e8b7f",
   "metadata": {},
   "source": [
    "- Train CatBoostClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "1e63e791",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:58.617519Z",
     "start_time": "2022-03-27T18:41:58.290012Z"
    }
   },
   "outputs": [],
   "source": [
    "X_train, X_val, y_train, y_val = train_test_split(X_train_val, \n",
    "                                                  y_train_val, \n",
    "                                                  test_size=TEST_SIZE, \n",
    "                                                  shuffle=True, \n",
    "                                                  stratify=y_train_val,\n",
    "                                                  random_state=RANDOM_STATE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "8258bcc7",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:41:58.765541Z",
     "start_time": "2022-03-27T18:41:58.688371Z"
    }
   },
   "outputs": [],
   "source": [
    "train_pool = Pool(X_train, y_train)\n",
    "val_pool = Pool(X_val,y_val)\n",
    "\n",
    "class_wei = y_train.value_counts()[0] / y_train.value_counts()[1]\n",
    "model = CatBoostClassifier(\n",
    "    **best_params,\n",
    "    class_weights = (1, class_wei),\n",
    "    custom_metric='F1',\n",
    "    early_stopping_rounds=10,\n",
    "    use_best_model=True,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "bed3d4ae",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:42:14.032806Z",
     "start_time": "2022-03-27T18:41:59.710462Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e6ad1d44f9be4042aaac728858435c84",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "MetricVisualizer(layout=Layout(align_self='stretch', height='500px'))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.fit(\n",
    "    train_pool,\n",
    "    eval_set=val_pool,\n",
    "    verbose=False,\n",
    "    plot=True\n",
    ");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "dcaa149a",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:42:34.787925Z",
     "start_time": "2022-03-27T18:42:34.614053Z"
    }
   },
   "outputs": [],
   "source": [
    "y_pred = model.predict(X_test)\n",
    "y_pred_proba_test_cb = model.predict_proba(X_test)\n",
    "y_pred_proba_val_cb = model.predict_proba(X_val)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6282cb0f",
   "metadata": {},
   "source": [
    "- Evaluating Quality"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "c55f9d16",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:42:52.343901Z",
     "start_time": "2022-03-27T18:42:50.549957Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    threshold        F1       AUC\n",
      "51       0.51  0.287837  0.677859\n",
      "50       0.50  0.283502  0.677120\n",
      "47       0.47  0.272937  0.676333\n",
      "49       0.49  0.279221  0.676217\n",
      "52       0.52  0.290115  0.676053\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>threshold</th>\n",
       "      <th>F1</th>\n",
       "      <th>AUC</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>61</th>\n",
       "      <td>0.61</td>\n",
       "      <td>0.300009</td>\n",
       "      <td>0.647045</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59</th>\n",
       "      <td>0.59</td>\n",
       "      <td>0.298464</td>\n",
       "      <td>0.654257</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60</th>\n",
       "      <td>0.60</td>\n",
       "      <td>0.298156</td>\n",
       "      <td>0.649552</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62</th>\n",
       "      <td>0.62</td>\n",
       "      <td>0.297564</td>\n",
       "      <td>0.640989</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58</th>\n",
       "      <td>0.58</td>\n",
       "      <td>0.296924</td>\n",
       "      <td>0.657110</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63</th>\n",
       "      <td>0.63</td>\n",
       "      <td>0.296845</td>\n",
       "      <td>0.636360</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57</th>\n",
       "      <td>0.57</td>\n",
       "      <td>0.295482</td>\n",
       "      <td>0.660173</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>0.56</td>\n",
       "      <td>0.293453</td>\n",
       "      <td>0.662458</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>55</th>\n",
       "      <td>0.55</td>\n",
       "      <td>0.293313</td>\n",
       "      <td>0.666505</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54</th>\n",
       "      <td>0.54</td>\n",
       "      <td>0.292079</td>\n",
       "      <td>0.669462</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    threshold        F1       AUC\n",
       "61       0.61  0.300009  0.647045\n",
       "59       0.59  0.298464  0.654257\n",
       "60       0.60  0.298156  0.649552\n",
       "62       0.62  0.297564  0.640989\n",
       "58       0.58  0.296924  0.657110\n",
       "63       0.63  0.296845  0.636360\n",
       "57       0.57  0.295482  0.660173\n",
       "56       0.56  0.293453  0.662458\n",
       "55       0.55  0.293313  0.666505\n",
       "54       0.54  0.292079  0.669462"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "qwe = {\n",
    "    'threshold': [],\n",
    "    'F1': [],\n",
    "    'AUC': []\n",
    "}\n",
    "\n",
    "for thr in np.arange(0, 1, 0.01):\n",
    "    qwe['threshold'].append(thr)\n",
    "    qwe['F1'].append(f1_score(y_test, (y_pred_proba_test_cb[:,1] > thr).astype(int)))\n",
    "    qwe['AUC'].append(roc_auc_score(y_test, (y_pred_proba_test_cb[:,1] > thr).astype(int)))\n",
    "\n",
    "qwe = pd.DataFrame(qwe, columns=['threshold', 'F1', 'AUC'])\n",
    "print(qwe.sort_values('AUC', ascending=False).head(5))\n",
    "qwe.sort_values('F1', ascending=False).head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "483ff734",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:43:03.522945Z",
     "start_time": "2022-03-27T18:43:03.511358Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[25776, 11018],\n",
       "       [ 1290,  2435]])"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(y_test, y_pred)\n",
    "# array([[28907,  8711],\n",
    "#        [ 1242,  1659]])\n",
    "# array([[26989, 10629],\n",
    "#        [ 1104,  1797]]). lgb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "97af29b9",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:43:11.150903Z",
     "start_time": "2022-03-27T18:43:10.920091Z"
    }
   },
   "outputs": [],
   "source": [
    "test_pool = Pool(X_test, y_test)\n",
    "curve_train = get_roc_curve(model, train_pool) \n",
    "curve_val = get_roc_curve(model, val_pool) \n",
    "curve_test = get_roc_curve(model, test_pool)\n",
    "\n",
    "(fpr_train, tpr_train, thresholds_train) = curve_train\n",
    "(fpr_val, tpr_val, thresholds_val) = curve_val\n",
    "(fpr_test, tpr_test, thresholds_test) = curve_test\n",
    "\n",
    "roc_auc_train = sklearn.metrics.auc(fpr_train, tpr_train)\n",
    "roc_auc_val = sklearn.metrics.auc(fpr_val, tpr_val)\n",
    "roc_auc_test = sklearn.metrics.auc(fpr_test, tpr_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "86c5f3d0",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:43:12.796843Z",
     "start_time": "2022-03-27T18:43:12.504770Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set_style(\"whitegrid\", {\"grid.color\": \".6\", \"grid.linestyle\": \":\"})\n",
    "plt.figure(figsize=(12, 8))\n",
    "lw = 2\n",
    "#roc_auc = sklearn.metrics.auc(fpr, tpr)\n",
    "plt.plot(fpr_train, tpr_train, color='darkorange',\n",
    "         lw=lw, label='ROC curve train (area = %0.2f)' % roc_auc_train)\n",
    "plt.plot(fpr_val, tpr_val, color='darkblue',\n",
    "         lw=lw, label='ROC curve val (area = %0.2f)' % roc_auc_val)\n",
    "plt.plot(fpr_test, tpr_test, color='darkgreen',\n",
    "         lw=lw, label='ROC curve test (area = %0.2f)' % roc_auc_test)\n",
    "\n",
    "plt.plot([0, 1], [0, 1], lw=lw, linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xticks(fontsize=16)\n",
    "plt.yticks(fontsize=16)\n",
    "plt.grid(True)\n",
    "plt.xlabel('False Positive Rate',size=16)\n",
    "plt.ylabel('True Positive Rate', size=16)\n",
    "plt.title('ROC-AUC', size=16)\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5b8d335c",
   "metadata": {},
   "source": [
    "## Feature importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "e4b676c5",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:43:16.563928Z",
     "start_time": "2022-03-27T18:43:16.545289Z"
    }
   },
   "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>Feature Id</th>\n",
       "      <th>Importances</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>cnt_checks_12_Овощи - Фрукты</td>\n",
       "      <td>14.623158</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>cnt_unique_plu_12</td>\n",
       "      <td>8.912469</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>cnt_unique_plu_11</td>\n",
       "      <td>4.184685</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>cnt_checks_11_Овощи - Фрукты</td>\n",
       "      <td>3.338058</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>rto_stddev_10_Сыры</td>\n",
       "      <td>2.269363</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>cnt_unique_plu_11_Овощи - Фрукты</td>\n",
       "      <td>2.157602</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>cnt_checks_12_Птица и изделия из птицы</td>\n",
       "      <td>1.929615</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>cnt_checks_12</td>\n",
       "      <td>1.714084</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>rto_9_Мясная гастрономия</td>\n",
       "      <td>1.534002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>cnt_unique_plu_7_Сыры</td>\n",
       "      <td>1.505712</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>rto_stddev_10_Птица и изделия из птицы</td>\n",
       "      <td>1.378511</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>cnt_checks_10_Овощи - Фрукты</td>\n",
       "      <td>1.370638</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>cnt_unique_plu_10_Овощи - Фрукты</td>\n",
       "      <td>1.367915</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>rto_11_Сыры</td>\n",
       "      <td>1.335201</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>rto_stddev_11_Птица и изделия из птицы</td>\n",
       "      <td>1.307900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>rto_stddev_11_Сыры</td>\n",
       "      <td>1.256216</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>rto_12_sal_diff</td>\n",
       "      <td>1.250946</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>cnt_unique_plu_12_Овощи - Фрукты</td>\n",
       "      <td>1.242309</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>rto_6_Мясная гастрономия</td>\n",
       "      <td>1.215187</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>cnt_unique_plu_10_Сыры</td>\n",
       "      <td>1.143953</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                Feature Id  Importances\n",
       "0             cnt_checks_12_Овощи - Фрукты    14.623158\n",
       "1                        cnt_unique_plu_12     8.912469\n",
       "2                        cnt_unique_plu_11     4.184685\n",
       "3             cnt_checks_11_Овощи - Фрукты     3.338058\n",
       "4                       rto_stddev_10_Сыры     2.269363\n",
       "5         cnt_unique_plu_11_Овощи - Фрукты     2.157602\n",
       "6   cnt_checks_12_Птица и изделия из птицы     1.929615\n",
       "7                            cnt_checks_12     1.714084\n",
       "8                 rto_9_Мясная гастрономия     1.534002\n",
       "9                    cnt_unique_plu_7_Сыры     1.505712\n",
       "10  rto_stddev_10_Птица и изделия из птицы     1.378511\n",
       "11            cnt_checks_10_Овощи - Фрукты     1.370638\n",
       "12        cnt_unique_plu_10_Овощи - Фрукты     1.367915\n",
       "13                             rto_11_Сыры     1.335201\n",
       "14  rto_stddev_11_Птица и изделия из птицы     1.307900\n",
       "15                      rto_stddev_11_Сыры     1.256216\n",
       "16                         rto_12_sal_diff     1.250946\n",
       "17        cnt_unique_plu_12_Овощи - Фрукты     1.242309\n",
       "18                rto_6_Мясная гастрономия     1.215187\n",
       "19                  cnt_unique_plu_10_Сыры     1.143953"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.get_feature_importance(prettified=True)[:20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "e7db29b1",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:43:39.271807Z",
     "start_time": "2022-03-27T18:43:22.694188Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsYAAAI4CAYAAABp8C/cAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAEAAElEQVR4nOydd3wVVfbAv1NeS6+0UAUVXVRU1koTFEQFFJV1rbgWlp/YAQuu4loRUERdBNa1IC7WVeNSpAWjESMIaFaKdEISEtLz8tqU3x8v7728EBIQQl7C/foZydy5c++Z8+7cOXPm3Hsl0zRNBAKBQCAQCASCExy5uQUQCAQCgUAgEAgiAWEYCwQCgUAgEAgECMNYIBAIBAKBQCAAhGEsEAgEAoFAIBAAwjAWCAQCgUAgEAgAYRgLBAKBQCAQCASAMIwFAoFAIBAIBMeZ3NxcBg0adFD6qaeeyooVK3j11VcbPP+WW27hhx9+OOZyqce8RIFAIBAIBAKB4HcyePBgBg8e3Cx1C4+xQCAQCAQCgSBi+Oyzz3j00UcB+OGHHxg+fDhXX301U6ZM4ZZbbgnm+/jjjxk1ahSDBw9m5cqVx6Ru4TEWCAQCgUAgEBx3CgsLGTly5CGP+3w+Jk2axJw5c+jZsyfPPvts2PG4uDg+++wzVq1axeuvv15vaMaRIjzGAoGgVbB169bmFiFiELrwk56ezocfftjcYkQEok2EELo4DkijwrdD0KZNG7744ouwrTZbt24lOTmZnj17AnDdddeFHb/00ksB6NGjB6WlpcdEdGEYCwSCVkFUVFRzixAxCF2EUBSluUWICESbCCF00XJQFAXDMBo8DiBJ0jGrUxjGAoFAIBAIBIKI46STTqKiooItW7YA/q9ATY2IMRYIBK2C6urq5hYhYhC6CKHrenOLEBGINhFC6OJ4cGw8uFarlZdeeolHHnkEWZbp1q0bdrv9mJR9KIRhLBAIWgUpKSnNLULEIHQRwmKxNLcIEYFoEyGELiKDjh071juTRMA7PGrUKAzDYOXKlXzwwQdERUXx9ttvs3//fgDmz5/faFm/B2EYCwSCVsGePXtISkpqbjEiAqGLEB6Pp7lFiAhEmwghdHE8ODYeY1mWSUhI4LrrrsNisZCWlsZzzz13TMo+FJJpmmaT1iAQCATHAZfLhcPhaG4xIgKhCz/p6enous7VV1/d3KI0O6JNhBC6OA5I4bNHYH7SPHL8DsTgO4FA0CpYv359c4sQMQhdhHA6nc0tQkQg2kQIoYvjgVRnazkIj7FAIBAIWiWBEezDhw9vZkkEghMM6frwffPj5pHjdyA8xgKBoFVwrAZetAaELkIcq0n/WzqiTYQQujgeCI+xQCAQCAQRhfAYCwTNhDQ6fN/8qHnk+B0Ij7FAIGgVCC9QCKGLEMJj7Ee0iRBCF4KGEB5jgUAgELRKhMdYIGgmhMdYIBAImpfMzMzmFiHIsxnVpL1YRdJL1UjPumj/UtVxrT+SdNHclJeXN7cIEYFoEyGELo4HIsZYIBAImhWfzxcRq5zZnqwkzumjJNaGEWUFWQLTBK9GtNXEqcngNVB1g9Q4iW9us9Aj1XpMZYgUXTQ36enpGIbByJEjm1uUZke0iRBCF8cB6U/h++aHzSPH70B4jAUCQavg559/bm4RqHb7UH0GB9rFYsTY/I4SjwaaARYVpxNwGaCDJsvke1VO/qfB+lzvMZUjEnQRKYh5jP2INhFC6OJ40HI9xmJJaIFA0Cro0aNHs9VtfbYanw7ogM3iN4YVGVx6KJNqgFHrJAPQTTBN+ryjoT9x7LzGzamLSEOscOZHtIkQQheChhAeY4FA0CrIy8trlnpPfa0an4bfKAa/wes1w41iAK1O1JpVBocKDhVDlpGeriC/VDsmMjWXLiIRr/fYeuNbKqJNhBC6OB60XI+xMIwFAkGrICEhoVnq3VpMuCf4UEiAUvOAUCSwKv7YY4/uP1+10nF6Fet3+o5apubSRSSiKEpzixARiDYRQuhC0BDCMD6G7N+/H007Nh4fgUBwZHg8nuNWV7lbR3qqGunJ6kMbxfX1rlEqxFog2uIflOczwKaAXQWrghFt575/lR21fLV1oRsmOXtcFJV6WbmxjF0F1UddfktCjC/3czzvj0hH6OJ40HI9xiLG+BhRXFzMtddey9dff42q/n61Dh8+nEmTJtGvX79jJltTlAlgGAaPPPIIffr04U9/Co1A/c9//sN7771HSUkJXbp04cEHH+Tss88+rDJzc3N58803+eGHH/B4PLRr144RI0bw5z//ucV7fqqrq+nfvz9ffvklHTp0aG5xWh26rjee6XcgPVnHkLRK/vjhgzICtW2wugazCUh1HhCyFEqr+fdAxVEIW0NAF1M/L2Xeirov6z7AybbXUo++ohbAsTSMDd3knls34dVkZJ8P1fTHiL/87z/gcET247Sp7o+WiNCFoCEi+05uQXg8Htxud3OLcdzIz8/nxRdf5LvvvqNPnz7B9LVr1/LGG2/wxhtvcPLJJ7No0SIefPBBPv/880Y/X+3atYs77riDESNG8MEHH5CSksLmzZt54YUX2LBhA9OnT2/iq2oafv31V+bOncuWLVsAuP322+nevTt33HEH5557bjNL13qIjY393ef6NIPEqW6cHvyeXrsCmOCtY91K+OOHY2Uw6zxcG7O/ourpbmvbyaYJmMRHhWcpqdT4JqeaKo+Xn3aYLF0Pjfu7YoGiBnP0uLfh40CLNp5N08Q0/aEUe3Y4+eWnUnRZ4qe1Lgq2eTA0A11WwKKCYYCmIckysmlimCamNTQYUtZ0ZMDq82JHwrRZ8TkceGuM7vtu347Vp5FYUYkmgReJqmgH0W4vsT4fFmDg7e3pd1NXANxVHqxRFnSvgcXubxf/Gv8DBetdmKqMvawKw25FTYviwc8uwlXhxjTBGmXB59JxxB35QM3fc38YhoEsyxT9dIB9y/fg3e+kco+Til/KsR5w4VElLGUGhiQhmQaxko8ec/oj2U2iuiQS1T0WTBktvwp7j3jUKBt7pHvpiD/G10kM1URRShIeHOjY6cJvOKhGQcZEoRoZH1YkDCR0FKCCRJwkkEwxDlw4sQOgUo4FjWriMFAAGaviRo6PQh3dh8TZN/xuXQiOlJblJa6NmMe4HtatW8err77Kzp07ad++PQ899BAXXHABU6ZMITo6mi1btrBlyxa6du3K5MmT6dmzJ4MHD6a8vByHw8G8efPo2bNng3V88sknzJ8/n7KyMk4//XQmT55Mx44dGT58OJdccgnZ2dns27ePc845h2eeeYa4uDh0Xedf//oXX375JW63m759+/Lwww8TExPTaJkBj/G//vUvvvjiC+bMmUN0dDRTpkxh/fr1REVFcf755zNp0iRsNluDsvt8Pi6//HKuvvpq1q9fz9ChQ4Me4//+978cOHCA2267LZh/8ODBTJs2jXPOOafBcu+77z7S0tJ45JFHwtLLysoYNWoUTz75JAMHDuTuu+/ml19+CXqQzz33XJ5//nmio6PZtGkTr7zyClu2bCElJYUxY8YEV72qrYe8vDxGjBjBN998Q1RUFOnp6Xz00UfMnz+ftWvX8sgjj7BixQoAfvzxR8aNG8crr7xyxF73tWvXcv/99zNx4kT69u3L5Zdfzscff8zOnTt55plnmDx5MpdeeilTpkzBZrPxv//9j927d9OrVy+efPJJEhMTGTJkCLNmzaJ3794AfPPNN7z22mt8/PHH9OnTh4ULF9KpUyf++te/kpycTHl5OZs3b8br9SJJEhaLhfPPP5+HHnoo7Jprc/fddzN48OAwz3+k8mSGjzfW6ZQ5DYzazlBF8ntgtVrGbMCxaxLy4Eq10gAsNZk0w3/scOKFbYo/NvhwkYDYWsaMafrDKFQpzIss6Qbd9lVE5COlQwxc29fB/VfGHHVZf/3LNnxN9TW7xot7kHcesPg04txunDYrPllBt/iN0thqF8mVVRiyTH5CHL7AVz/JPwe1xeulbUUVFsMINpHdbVIwA3WYJm3LKvBYFMqiowGweb0kVzqx6n6juvaD9iDJaj+GTRPJBFOR/em6geLTiCmvwuLTkHQDZ7Qdw2ZB8fiIdrrrbS8NtiHdQNIM/z0jSSheA80mgyRhq/KRVOyhOkrFa5FwxYXm4rZWa8RVaigGOGNUoqp82N1+nSjoREvVlJtRmCgYdoMOngok06QMO07VgaJBHzJJpMR/qUAVcehYMYjFSxQJ7MFOJV5sOEkhihJsuKgill30REYhDhcqPqxUUYWdtuzBihsfVmR09tOZLfTCQMaGh3P4Fi8JAKi4iKYUCQMdhXx6YMVDFAcAKwZ2FDQkdHTkoEFeQRomMjaKsKFhYkHBi44NAxMfKjJeJKIwMXGwDwUVGQMNGxIG1cRh7RQPe0sx8WLFjYmKbLfDmW2Rft5FpTsWAwcyPuxDuqG4qzD3lCIrJsrT16HEqbjuXYiyuwSjTRz2rMege7uGfu3mQbopfN9c0Dxy/A5EjHEdSkpKeOihh7juuuvIyMhg/PjxTJw4kcrKSgAWLVrEpEmTWLZsGR07duT1118HYP78+QAsXbq0UaM4KyuLf/zjHzz33HOsXLmS0047jSeffDJ4/Oeff2b27Nl89dVX5OXl8fHHHwOwYMECVq1axbx58/j8889xu91MmzbtsMoEWLhwYdAobteuHe+//z6yLLN06VI++OADNm3axOLFixvVkaIofPjhh9x7770HhY1ceeWVYUbxhg0bcDqdnHTSSQ2W6fF4yM7OZsiQIQcdS0hIoF+/fmRkZATTHnjgAb799luWLFlCfn4+X331FaWlpYwbN45BgwaxYsUKnn76aWbOnElWVlaj13QodF1nxowZv3sy+Lfffpvhw4dz9dVXB41Rm83GoEGDuO2225g7d24w71dffcUDDzzA8uXL6dixI4899hh2u50BAwawfPnyYL4lS5Zw+eWXh9Xz1Vdf0bFjR6ZPn868efPIzMxk2LBh3HDDDWRmZrZYb3td7lvs5ZkfocQrhRvFquz3yDrUmkFthIxhh0qYmRA4BqFzHKrf2D3ceGHNANshuk9LnfSgIV7L+NHNkGEtScHN5tUi0igGyKuC15a4eOjdo1tJ7q9jmtAoBr8uZTlMr0gSUV4v7csriHW5aVtW4Tc8DQOrTyO5soqKKAe5SYn4LJbQeTXlyRA0isHfHKQ6PiWvUmMU15zXoawCe41RDI1EW9aWVZb9sgXSVQXdYaO8bRK6omDKMq6keDzRUcQ43X5Z6tkaRJExbSqmqmAqMppDBVlGMkxSD3iw6hBXpWHYFFBkYsu9dNztJLnYS3S1jt2tk3TAE1aXjoLLtCOhIAMJbg+qaaIASbhJ00pJwImN0JdVCYilggQOkMguVLxUk4oHBzs4l3y6s4NzKCGVbPpTQAeqsKPiJomdJLGHTmzFhv/lwIKPErqxl+4YNZr3YGMfp+MjHo0YYihBxv9bqui0ZwcKVjy0x0V7PCTiIgkbTuIoIoZivCSgAhYMdJJx0RY3SVTTBis6DnR8ODBIQCMKnWhMYrHhwoIHBxXYqSKRPDx7nbiJJoYSbFRjowLJ7UXK3kuVOx6NBAzsaMRQ9XUefLMJ+64dWLfvRLp5Ot4Rr6HsrgJsyIVuvD0mwZZ9jf3izUDLjTEWhnEdvv32W9LS0hgxYgSKotC/f39mz54dNIz69+/PKaecgt1uZ8iQIezdu/eI61i6dClXXnklvXr1QlEU7rrrLiZMmBA8fv3115OYmEh8fDznnXdecGqZL774grvuuot27doRHR3Nfffdx+LFi/F4PI2W+d///pdXXnmF1157jXbt/G+XVquVzZs3s3TpUnw+H++//z5XX311o/LLskxKSkqj+Xbs2MEjjzzCX//610bDKCoqKtA0jeTk5HqPp6SkUFxcfFC6pml4vV6Sk5NZvXo1bdu25YYbbkBVVXr16sU111xDenp6o7Ieik8++YTu3bsf1vXWR35+Pt27d6/3WI8ePcLaz7Bhw+jTpw82m417772XnJwc9u3bx7Bhw1i+fDmGYeByucjMzAwzjF0uFwsWLODPf/7z75KxJbFgU40n0FfnQ5elliGj17JuA8ZxXWM1gCrX/3ddrEp4367XE2YBIUM7uuaFMcbi9xQHvMUBYypgyNfxalrrXlcEkvG/o5s1w3f0k278LlLLK1ENv7mkABZdB1nGomsUx8VQEhuDUV/sOOCx2SiOiQ6+T5mAXLudSRKlcSFPut3rPfamgCRR3iYR3apiKjK26vo9xUeD3W2g1FyWbEJMlYbNpZFU7EUxwFprykEJ0OWwWCD8QSN+PLUiNf0eZYjGSz5d6o06kjBR8SJhUEUSRs35Jgr76YiBSjv205tfSaIUcPjD9uuUkUAJSp03XDVsP7x2F8kYWJAJvWnL+FBrDHgJiKkVlhR+xTIGak2eKsywaz64f5AAO1U4KAuWIwEybkys+IiuVYOEhIQSJpdZczT8bJZtPKguwe9HGMZ1KCkpoW3btmFpvXr1wm73xzAlJiYG01VVxTAO57trw3U4HA5OP/304H5cXFzwb4vFEhwoUFBQwFNPPcXAgQMZOHAgf/7zn1FVlYKCgkbL/OWXX+jUqRNff/11MG3MmDFcffXVzJ8/n2HDhvHXv/6VPXv2HPH11MeaNWu44447uP766xkzZkyj+ePj41EUhf3799d7PD8/P8xonjVrFgMHDmTo0KG4XC7OOeccSktLad++fdh57dq1o7Cw8HddQ1lZGe+99x7333//IfOMHj2afv360a9fP55//vmDjqelpbF9+/Z6z922bRtpaWnB/Y4dOwb/jouLw+FwUFxczPnnn49pmqxfv57Vq1dz8sknh533l7/8hTPOOCPs926IK664gksuuYRRo0bx7rvvHtY5x4r169eTn5/Prl27yMnJobCwkJycHAoKCsjOzsYwDJYtWwbAypUrqaysJCMjA8MwyMzM5I9tdL9xqdQxCfRaD7u6x2Qp3FiuTe2wi0PlCRBd56tBoMqAy84qh4xrRQa77K+7NrUN4doD7wJJLSCyrV1MNeXl5cEvOIE+JSMjg/LycrKzs6mqqmLt2rUUFhayfft2Nm3aRH5+Pps2baop5fhfZ7hB4/fwArisVtyH8UWoIjqKSpuN0ugo8hPj0SwHx4wrhoFkmuiyXO8VHourlnT/PaBZLcdciz5VCivTZ5GR9fBaAneJ265Q3MZORbwFVIN4nNS+Qp36dVBFYj2poGHBRRSx5COFfdaB+JrQh5SaEAw/apgh6pfNioaVduSSSCl23LSjkBhcgN+QraIdJlLNBlLNBORqmCfbCJPdH9lMTRkhZHQUNH97qgmlqF1Gffiw4yU6rBz/dWioYaMHTEzM4AtCoG6znlG+P1ur8fl8ZGZmUllZycqVKwG/Ew78X5SLiorYsGEDJSUlbNy4kdzcXHJzc+uV8VgQ0rEUNOdbCmLwXR1SU1MPMqTeeustLr300iaro6qqinnz5jF+/PgGz0tJSeGJJ57gj3/8I+D3lubm5tKxY8dGy3z44YdJSkrinnvuYfDgwXTr1o3t27dz5ZVXcscdd1BUVMSMGTOYNm0ar7322lFd35dffsmMGTN47LHHDvrkfyisVisXXXQRX331VfD6AhQXF/P999/zt7/9LZh233338ac//Qmfz8fkyZN57bXXOO+881iyZEnYuXl5eSQlJf2u63jzzTe59tpradOmzSHzfPTRRw2W8Ze//IV7772X0047jb59+wL+BQdWrVrFO++8w2OPPRbMe+DAgeDfZWVluFwu2rRpg6IoXHbZZaxcuZLCwkKGDRsWVsc//vEPXnnlFT7++GOuv/76Rq9r0aJFREVFsXfvXu68885DerSbgvpmJwnoN/Al47LLLgNg0KBBAAwcOBCAfv36saQf9H/Hzbf5YCqEFtXwGmBqfmPTa4S+3qky+PRDf8mr1sChhMqo+8wJ4NbBwcHHHQpYauYjNuqcWNcDXTfmte6+aYJhHOQFiyQuOsXCe/f6B+QFfpdA+FNg/7zzzgMIDsqtff+0b9+etxbAHTdtazohAy8XNfHBgb+LYmNIrfCHxBVHRwfjgw1ZPmQEjd3twW0PjbnwyRLlUQ5/qEY92Hw+2pVX4lUUqq1W7F5v0PvklSQqHHaSq10He6Rqy1w7zaf5BwfW4Kiooiw1CcXjQ7eolCfFEVXpRNF05DrNr8E2ZJqhdlzTpiXDRDIMKmMUVB28NoWKBCuSYeK2ydg9BroM+9s5QALF6w9HKE2xI1l12hX50HDjwlpTt04lNmR0ogiFCOmYuIhCxYeCjoKBFwvruYBEnLRBw46beMqQcBFNOXEcwEdcMDwC/GajRCjO30sULlIoIwEVJ53YgYmKiomEjI4VEx03MYCJg1K8qGiAlXJ0LMSyFx0LEiaVtCWKUmQ0VJzo6IBEHAf8XwxqDFkTK06icBFNAgWY2AEDmUp0VMBEx4KMgYt4LLKBbui4icNKNQYWTFSkKB9x1fuoIA0NGzJerClWMGLwVinICpi3XIKlSwzepxYjazqmJGFd/CBnDvX3q4ExMIG+c+jQoQBcdNFFgN/2AH73M/FEQRjGdbj44ouZMWMGixYtYujQoXz33XcsWLCAa6+9tsHzrDUjmJ1O50EDm+oybNgwHn30Ua644gp69OjB22+/TU5OTqNxrFdddRXz5s2jW7duJCQk8I9//IMVK1bw6aefNlqmxWKhd+/eXHHFFTzzzDP885//5D//+Q/79u3j+eefJyEhAZvNdlRTzQFkZ2czdepUXn/99cOeoi3AQw89xJgxY5g5cyY333wzCQkJbN68malTp9K7d28uueSSg86RJIm4uDgqKyuDv93ChQu57rrr2Lx5M59//jlPPPHEEV+H0+kkOzubhQsXHvG5tTn77LN56623ePPNN3nzzTcB/0C3bt268eKLL3LBBRcE8y5atIirrrqKbt268eqrr9KnT5+gsThs2DAee+wxnE4nkydPDqsjMTGRJ598krFjx9KvX7/gOY0hSRKyLLe4qYu+GWM/JuXohsGs7938r0DHq8H8X2sOqIQMbkXyh2G4dP8m1UrH9BvFUBMfWqeCgGEWMHYM45AGVSB/WZyd5Eov2c/F83FWFad3svDeajerNh36tKakqWaleGtB0y7JW1ni5dXndpBbXIHVZmLoSTgNFWdSAlafRrTbg1WSMFQFm64R53RRGBeDbrGgaBqqYWLKEqZpoHp9mLKMbBpU2u2h37P2b2uaSIaJo9KJDvh0HV3yzx5ix980omNN+g6Npmibya6f3bTpLuPzyiS1s3PZuG7886pv0WLsoCigm6i6hq4oPL7iIqYO/RbZVU3H0+1U7Kuk55VpXPTIWb9bP5rH/xIpSRK5i/ew56f9HPjpAPGSinlSLI4KL+X7qlB+q8IXZaEszYZR6iFa82Gt9mAYEtFeDbNER1ZlrB6DAhz4AyZ0fOhU2VXaahKuGAdOj47eJRp72yhSV//APnqgYSeKShLZRxlJpLKfKLz4sGOlAgsOXERjpRITaMtOKknFCygYuLBjYK2ZhSKaIrpgx42dciQ0LHjR8OFBJxY3JhZ8bdphOz0ZU0mmKrE3ZloS5t792AaeipG5Cd0rYZt5HZau4V8eZaD9wWoMYgMaMjUD/ub65sKoOwnp4ZistsevO4xcgt+LMIzrkJCQwMyZM3n55ZeZOnUqaWlpTJ8+vdEY2eTkZC6++GKuueYaZs6cGTaFWV3++Mc/ct999/H4449TXFzMWWedxXPPPdeobLfffjs+n48xY8ZQWVlJz549eeWVV1BV9bDLvPfee7nuuuv46KOPuOeee3juuecYOXIkmqZxzjnn/C4jsjbvvvsuPp+P++67Lyx96tSpwbfWQ9GpUyfmz5/Pm2++yU033UR1dTXt27fnqquu4sYbb0Sq5U2ZOXMmr7/+OoZh0L17d5588kni4uJ47bXXmDFjBrNnzyYhIYHx48cH354BHn30UWRZDs5tGnij1nUdTdP417/+xZlnnommadx///3BF56j4dRTT+WVV14JzmP8zjvv1DuP8VlnncULL7zArl276NOnT9jv94c//AFVVTnjjDPqbYs9e/bkiiuu4OWXX+all15qUJ6hQ4diGAYxMTEMGzaM/v37s2BByxkxfChWrlwZ9ls3hiLLPHhx6CX2vXry/LDXywXv6n7j2GeEvGyG6Z9RorZxpBsgyeFevzAPIPV7BsOQsMiQFGdl7OX+R2S/Pxz51FIBXdzxxgFWb67/g/vWV5NZs6mMRT/pTBiVhCpLxDhaR3RdbJKVJ2b0JD09nbKyMm655ZZjUu6eHU6m37cZV1QURuAlR5K458FUzjmv/hCBw+WR9Yf+uvbI0r5HVTaE3x+qLfTo7zK8K12Gdz3q8g8X07wa72/70dbvRbnoJLTvttHpT+eF9e8Adb/T2Wu2AI5af8fVbIdLvX3FvZcdQQmC1oyYrk0gqGHOnDkAjB079rjXPWXKFBISEnjggQcOmef//u//GDlyZNCYFzQfhVUabWdpfkM3MA+uS/fHFlskfyCmIoWHU/hqjvv8g76CsdK1B3z5NKb30nh4+NFPiRZg/XYP97xVQXllaP7j32alHGSItEYCA28DUzYKBILjgyndGrYvmfW5HiKT1uEeEAiOATabrdE5nJuDgoICVqxYwbZt24JxnIKDqT2dX1PTJkbFfNyOOdkBbs3vQXaofmNX88cKY9aJXFVk/zd1m+r3NnsN/5RtPt1/vlcHt8ZDV0UftXy1dXF2dxtZz6fyv9dS2VaznQhGcYCysrLmFiEiOJ73R6QjdCFoCBFK0QQ88sgjh5w7t3379o0O2Gpuhg4dSnV1db3Hhg0bxuOPP37EZW7evJm77rrrkMcff/zxgwaVHW8OZ/aM5uCDDz4gPT2dyZMnR6ThHin079+/Weo9NdlkS4kUPitFlc8fWqGGx6FS6/N7cAo5E7+BbBqYTx8bT3Fz6SISiY+Pb24RIgLRJkIIXTQ9dWeiaEmv4iKUQiAQtAqys7ODMyIcb+S/V2Na/Ish4NZD08CpEtjV0JRxtado02pill0aih20xxsetHskNKcuIon09HQqKiq46aabGs/cyhFtIoTQRdNjSLeF7cvm8Z0a9GgQHmOBQNAqONx5nJsC40m/UbuhQOPsf/indQL8YRVePWQce3W/Vzkwv7FPR1Jg3V+OfpBnbZpTF5FGdPTRh6a0BkSbCCF0cTxoST7icESMsUAgaBXs2rWruUWgdzvVP/gugETNEtUmuDSQauZK1nQUtw8ssPI2C2e1O7Y+ikjQRaTgdrsbz3QCINpECKELQUMIj7FAIGgVNLQQy/GkcpKdzq+5Ka2oSfDq4DMwnz52oRKNESm6iAQamx/+REG0iRBCF8eDlusxFoaxQCBoFVRWVkbEAy/GJlMy4fgZwfURKbqIBFraAjZNhWgTIYQuBA0hDGOBQNAqUJS6a0iduAhdhDiRpqZrCNEmQghdND11Z6VoSQjDWCAQtArEVHYh6tPF7jKNU2e48Xh8gFIT/yyBqlD1iIVoe+sMORCGsR9xf4QQuhA0hBh8JxAIWgViIYcQdXVR4jLp9rIHj88AWfEvKKKb4DFBVYh52UB6zoP0bOsbqCZCKfyI+yOE0MXxQKqztRyEx1ggELQKOnTo0NwiRAy1dXHqM6Xs8FgwtZqlqPVaU9cbJkgmWGtW7fOZLN2hMfSk1vNosFqP7VR4LRVxf4QQuhA0hPAYCwSCVsG2bduaW4SIIaCL2EdK+c1tQfPVLDri8vmN4QCyBJaaeEtJAkXi3mVGPSW2XFwuV3OLEBGI+yOE0EXTY9bZWhKtxy0gEAhOaM4888zmFiFiCOjCaUhIPh0zygLlNWESKmBR/U+rqIMfAdtLW1fogVjgw4+4P0IIXQgaQniMBQJBq2DNmjXNLULEsGbNGmwTSsAwMaKsoMq003R6ejU6VXqRqr0QY/F7jAPuHMP0b2ZL8+80TGVlZXOLEBGI+yOE0MXxQMQYCwTHhP3795OcnIyqiqYpODL69evX3CI0O+VOL4lPVGCTTqeHy0snn4avykWVqgRDi6NNE49Ho1gz6F5WTX6cjUqbFQxAljD0lvUQa4z4+PjmFuGYsSPlSXzFsDc5keqLTuaU8n2UrCvDKht4vTJV1iRQTWwuJxfu/z+scaHZF8T9EULoQtAQwvoQRAzFxcVce+21fP31101qGI8ePZoHHniAiy66qMnq+L3cfffdDB48mD/96U/HpLxVq1bx9ttv89577wXTtm/fzrRp09i8eTPR0dFcffXV3HnnnS1+WquVK1cyaNCg5hbjuBL1wAFcNguSLGFFwqPrgIzbhF/tVryyxKkeH5gm5YrijyMGog2TP+wpQTJNtidEgVZjNUv+8XmtidwdVTw4aj2aLiOZJv2ujuf6v3RrbrEOyfZ1pWy78j3OOLCFalMl3iikmkTcUjKdzH1Uk0Db4mI2rTNwVrmxo1AqJ6ApVhIs5SS5qzgQHcfyk95Bw4HD46aNu4QD8XHk2xNxVPiwWZx08ezF0GKolmIpbpfIVTtvBqBi2wFWXfwFVbpC2/5JDHxvGGpM65rK70TsK443Yh5jgeAY4PF4cLubfrqojz76qMnraG40TWPBggXMmTOH7t27B9MNw+Chhx7immuu4fXXX6egoIB77rmH1NRUrr766uYT+BjQWh50Ps2guMLDb8U+3l/toXsbheeWe3EhoysqDt2gnaFjVWUMVSXNp1Mhy1RaVCQzfKDLtigbKtDF6yNJ88cOlysyFkVmr0XFY1PQpVqWsAmpjoMfaG63zpo1xZxyahQ2m8JXi4pIX+JFBiyAIoPP8NctATrgAyTDRJJlJElCliQk0wDT5KYRUYwYHT4zQNH+atZm7Kfb6Yn0PCMBV7WPSTfmoEsqpmGQ1lEiN9fEIstgmv7rNA0UJHwQTPfJMgYmFk1HkjsDEhbDxKLIIEl8n17Fjyt+5qUFhx9nahgGVWUeYhPtGAYU7iwjuUMMFofKivd3E5OocCC3Gnu0TI9zUuh4Sjyq5eA3jIoiJxWFXlI7x2BxKOiGwUvDvufP339KkqsYGZk4NAbjRUUP/CQUWhRSfEXs4XR8+L3AnfbvJ1ouJNFXSrVs59u0i9gZ3YFdhonFrZO8141qutGR2NDlZFwxVkygpI0FyTCQ8xx0LyrBjo+43UWsjp9FQVQMMRVwimc/Ft1H8aJK/tP+fWyah/YjO/LHhVcBsH3hFtaO/YY4U8N2XVcG/WvYYesyEmgtfYWgaRCGsaDJWLduHa+++io7d+6kffv2PPTQQyxZsoTo6Gi2bNnCli1b6Nq1K5MnT6Znz57ccsstAAwdOpR58+bRs2fPQ5adnp7ORx99xPz58wGorq6mf//+fPnllwDceOONjBkzhn//+98YhsHll1/Oww8/DMDw4cOZNGkS/fr1Y/PmzTz//PPs2LGD3r17k5SURFpaGmPHjj3Ie/vhhx+yYsUK5s6dC8Ann3zCggULqKio4Oyzz+bRRx8lJSWlUb306dOH8ePH88EHH6DrOqNGjWLcuHHIdVx1jdXfEC+++CJ79uzhpptuCounKy4upkuXLtx6663IskzHjh0ZOHAgGzdubPGG8fHwAr2S6eVfP3oo2+kiXjexSPD535Lo0s7Cec+UsMOpUGq3YgAWnw9JluhZ4cKhGVh1gzJFwSVJtPf62BLroNhuxabruEyI8uq09WmUKDLVkkSirqPKMpUKnOKDRJ+XSkXDYpp4ZRmLZhAj6ZSpCvti/FOSRekGztoCW1Q2O6CzVwsOKInRDYot/q7fqGcCir4bCrh2jISEiWya2HSdwDphLiqQFAUdCUlVMSUJ3TDAMJAlCRNQTAMb4FRVNFkB0yTK50M3Tdw1X4IWfuliycebkGr0JAMuWcZhmkhmJZq0F82iIlnsWHQdFIV9RSoWVcOmG5iS5PdHSf7ybYaBzzTxWlSivT4k0y+7VBM/4lHVoLccScJbDU9csc6/bxo1x2qOmyayYSBL/ugSycRfX026KQGy7I88qclvShK6JCMB3ywoqzHYqYnXlvxFm/63FtkwsPg0TFnCME0u/2Uxaa79/MwFuIghljKs7ELHgZUqoikg2ufBS0zQKAbYGdWFC6u28VObM/m22/kYkgyaiWKapJUUYzV1JEDBJKXYya54B5UJMRiKX2dpv4RmZlBNk5Mq9nJ+xT7suKkiiUK6UyZFEetSgCj2LSkjSX4Bi+llW2IavZ2lOHQfnvmlbHt3PdXYKCSBMz49i7Zzl0C5E565Eao9MOY18Onw1PWwrwS++RWGnQN///PBDXDddnjwX7C7CGIdcOW58MLNh/6U8c3/4PEFEGOHN+6G7u3qz1cL4TE+HrRcj3Er+2gmiBRKSkp46KGHuO6668jIyGD8+PFMnDiRyspKFi1axKRJk1i2bBkdO3bk9ddfBwgauUuXLm3QKD4cqqqqyMvLIz09nZdffplPPvmEn3/+OSyPx+PhoYceYuDAgWRkZDBixAiWLFlyWOUvX76ct99+m+nTp7No0SLS0tJ4/PHHD1u+b7/9lo8++oh3332XpUuX8tlnnx3R9TXG2LFjmTt3Lp07dw5LT01NZdasWUEj3Ofz8f3333PKKacc0/qbg6Z+0GXu1Hjovx527fHSRjexA4oJN08t4e//rab0gEGxw4YhSSBJ+FQVr1sn2mcQrRlYTEjRdDp5NYqsFgqj7eiKTLXVQoxpkqTpuGSJMlnGK0nsV1X2WVQqLCrrYh2UWhSSdAO76ffSAthqjMIADsMkyqqQaJq013WSdAOQCLN/az2vrJoRNtguxutDlxUMWUaXZEz83pOA2SjLCros+wft1Riauuw3CBXTb5SpgCbJfqMYv1HpVRSs+A1IU5ZxWSx4FAVdUfBareiK4jf5JAldlvFZLCBJmLKErirIgImJXdMhYBTXYEoSLquVSocdm6Yj41/tzgwYUubB00WZkr8u/ybXbP79mlnrgtdsynKNLP5/FRO/kRzQS408cuDjcU2ZZqD8gK5k2W8jSxKKYWDR/AZ/j/K95NMVFzEAVJJAER0wsOAmEQ9xRFGCgg9NCV15aWwM61LOYNVJffGqNjTFgmaxYHV5kep8yJYlk6r4GIzAC4IsU01ofmcVNx3YhYrmbweU0IEcVG9ohhJblY5igoGV7mXFOHQfMjrxmos4w0U7o4xEs4qfb/gZFv8EWVtg5Atw7UtQXAkV1fDwu/ByOqzdDs98DJ9+z0FcMxUyN8GeA/C/vfDS5zB/9cH5AHwajHgBvtsMSzfAbbPqz1cHYRQLGkIYxoIm4dtvvyUtLY0RI0agKAr9+/dn9uzZWK1W+vfvzymnnILdbmfIkCHs3bu3SWS47bbbsFqtnHHGGXTt2pU9e/aEHd+4cSMej4cxY8agqipDhgyhT58+h1X2F198wY033kj37t2x2WyMHz+enJwcdu/efVjnjx8/noSEBDp27MgNN9zA119/fcTX1xCpqamN5vH5fDz++OOoqso111xzTOs/FOvXryc/P59du3aRk5NDYWEhOTk5FBQUkJ2djWEYLFu2DPB7dSorK8nIyMAwDDIzM6mqqiInJ4ddu3axffv24L+bNm1i1apVZGVl4fP5yMzMpLKykpUrVwL+ly2ArKwsioqK2LBhAyUlJWzcuJHc3Fxyc3PZuHEjJSUlbNiwgaKiIrKyssLOXbHmF8DvXattdLi9Jj/vKsen1E4laGipdWZ5kE0Tlxye10BCBQ41UVrAQ+rf6pQnSaS5PERpOhbTJM3pIU3TSTZM2jvdWAydnzvEo0vgliW2RNuQa66jIsoSMpQlE02p9Uio7UUNynEIAQOnEAinMOukm+EpkoRRowOjxhiWwk8I1RkSr14flAkhIzTsQChuWgO0Go+2T25klHwjk3JIdf6tV+ha3umDs0l+efEb2Ao+VHxhWeSwVxkZH3EYRFOVrLKnXTJburRjT1oSq/4wEKPWbyZhYq/2AGbQiDaAaK8XuU47LIuKogKVNmynCxtR0JFrtUA7LrqzJVR2raZRt6wAVjSk2seq3Jherd68AOwvY+XKlfh8PtatW0deXh5GYdlB2XLX5VBVVUV2djbl5eVkZGQAsGLRUiivDuZz795PVVUVa9eupbCwMNg35Ofns2nTJvLy8li3bh2ZmZnBviHQzwT2V69ejcvlYsOGDezdu5etW7eSm5vL1q1b2bp1KyUlJfzwww+4XC6ysrLq7WcyMzMpKSlh3bp1lJeXN2m/V15eflT9XlNh1ryemQe9pkU+IpRC0CSUlJTQtm3bsLRevXphs9lISEgIpqmqilHf99xjQGJiYlg9Zp3OvLS0lJSUlLAQho4dOx5W2QUFBcyePZt58+YF0yRJIj8/ny5dujR6fqdOnYJ/t2nThgMHDhxWvceKsrIyJk6ciKZp/OMf/8Butx+Xes8+++yD0tq0aQNAu3b+T6CXXXYZEPLqDBw4EAiNJO/Vq1e9ZXft2hWHwxGWN1DG0KFDAYIDLgMvDklJScHzA799IC2QJ3DuI9f1ZllZNd9vV6jQdOJMv9Fx1+VRXDsgmgufKMZiGPgC7cmrYZfArciohv+ztg/wKTKd3V72eW1oVhXJMPBhstci09WnU26aeCQJh2HgUv1lWQwDu25gAhYTbIaBp8Z7Ga9p5KTG8ofiKtq7vBRY1KDRKQFStI1CSWb5yW1AM0grcmIz/cZKistHSZQVs8Yeclv9cbpIkj9EQJbQdP+DwgQkXQ95WmteEBTDQDbxG7s112gzTeyaD6+soJgmFl2nWrUEDUWbpmGpmRrOVhNK4ZNlrLruj0U2jGBeRTfQJAmjZguELhiAIUsYNSEMkmnitFmJ8XiQDb/3GvwGsa4qeC2hAWRx1a7QdULY3yaEwiRqHZMMA9k0MSBk2AayBPRT+5ya0AkC3mnT9Bv3uo5UEyNtAHviO9KzdBdO4qgggQQOEEc1YEOlGhvlVOO/R84o3MXehBTcupVdNn+aocjIuoGsG8SWVUGN7FZdx6x54bJ7vJz2Wx6/nNYZU5aQdIOi5Giiql2YhhY0zDUsWPEG9aRb/C8ShiyhOjTMmm5qZ0ISSe5qor1uTExshoYPhVIpBp+jGkzV7819eARSUTm8V+Px7d3N7znesR96dYY/92NQot9Tfu655/rzPHcTTHg3KAMnt6fj5FsgJobzzjsPCPUJg0deCY+Wwoufgapgf/FWiIkJOjgCfQtA+/btAf+qdy6XK9hXBPqIwL8DBgzwi9q7N4fi/PPPB0L9Sd1+JtD/BPqSpuz3asvxe/o9wcEIw1jQJKSmplJYWBiW9tZbb+Hz+Q5xxpEhy3JYWeXl5UdcRocOHSgqKkLTtOAsGEVFRSQnJwOgKMoh60hJSeHmm29m5MiRwbSdO3cetmFdu56CgoJg51ibhuo/GvLy8vi///s/Tj/9dJ566ilsNlvjJ7UAtmzZ0uDD7GhxWCS+GRvF7lIHlZUevsr2cmM/O93a+/WXMy2ZVzPcrMo3GXuezGnJDjJ363SOslFUJVHl0nBLCk6Pxi+/eDj9twoqFQVNMpCirdhsUOnz0cbn97HomJiGQZWqYBomP9uttPH66KSbGEhokt/bajdNulS7+S01huTcUpI1jWKLii5JHHBY8ao1EcIuHSwydj30ImoxTFRVxmeVwa2BbrLDYnCK28AwTayyiRtQIOjDlDQNCQkNE1tNuluS/TKbJrqi4q6JyVV0HUOScCqq3xFrGFh0HdPlwqrpaKbJmefFUFrmYccOAxeg+nxomoRksQRjm3XNhyxBpaqimCaSYdSEMEjIpl9PNsPApypUWSzYvD7UgKGuKFg0vcawllF0DY+mEQWg1dxfpukfQYgEmo4O6LWMf/xH0KhxKAd0WOP9lQ0DJDMs7jioZf8oQSRNA0VBq4lPViSTtr4DfNHzCk79/g1OZQMmsLj7QDBtDN2xCrnGSFVx4iMOgE5lfus0+6SeVKqqP8RDN0jNK8aU/V5xZ4wd+UAlYFKQGEuMr5rO+wtxuF1sPLUTicXV2F0+8trEs9sygCsKlqOhUqimECO56Vq9mzJLPD90OBfT7Q8hsWga+1NUohNlDjij2J6SgmwYGJrGaUU7qZLjiTknmmHfjIXiMf7Y4k41Yy4euxaqvXB2N/BqkFsMnVP8C83U5eGRcFN/0HTwaNAxGWwNzIrxws1wz+Vgt0JK3KHz1aKp+woBtOQYY2EYC5qEiy++mBkzZrBo0SKGDh3Kd999x4IFCzjrrLMOeY7V6o95czqdREVFNVh+ly5d2LNnDzt37qRDhw688847Rzzd2Omnn07btm2ZO3cud911F+vWrSMrKysY39y5c2e+//57rr/+eoqKili8eHHwjfvKK6/kvffe45xzziEtLY2PPvqI2bNnk56ejsXS+NRGc+bM4dlnn+XAgQMsXLiQv/71rwflaaj+34vb7ebee+/lggsu4NFHHz2qsiKNuvHUTYEiS5yULEGyg7O6OsKOOWwKjw6NprZWT2tbu4ut9QIyNOaI695fodHt6TL2aqbfuDUMnHFRaJJER59G+woXP7eJJc6tkeD0sNuqUh5T8yUgcG94dcoVGUeNYee0KvjUmvhamwK6ydYe7dg46dh9Qfhtm4tnX8ijym0iS/CPmR1JSTk+XyjS09P5+t1OeA3Vb5CbOqZh8MZX5xyX+g+H4t3V5BeeRtR5j8IfTyJp/AjS/jKLH+POpq23iFR3Hi7JRoEZTxolmCh4VCunbdvNpi5d8KoK8cUVWA2N8ugoJAOqbCrrenWkfUExVo+TXW2S2ZNkoGJw3rb/YVYZpJpOqmw2qkwHO4xT+aNnIu0Uma8S5rA9+WRK4hOweHyc/lJf/nDzSWEyn9HYRSXH+rcAPWs5DGyWxgfItUts+HhdOjY+6Lk2x6OvELRchGEsaBISEhKYOXMmL7/8MlOnTiUtLY3p06cHZ42oj+TkZC6++GKuueYaZs6c2WC8b69evRg9enTQoLzllluIizs8b0EARVF4+eWXee655xgyZAinn3562FKhY8aM4emnn2bIkCF07NiRK6+8kuzsbMBvGFdUVHDfffdRUlJCly5dePXVVw9bhg4dOjB69Gg0TePGG2/kqquuOihPQ/X/XjIyMti9ezf79+/nv//9bzB94MCBPPPMM0dVdnNz4MCBVv2JsG2cSvWMhg2Azo8doFiT8NktnFztZX2UFd0WGmSFAQcUhWqbjCpBZWpUyGhWZZBNEm3HNrTp5B4O3n2re+MZm4i+1285ZvOCNwXJXaKgSxSY/rnGLwC49WU61coTAxQ/uJS9M9chIeFWDdwxbs7b+BMlUW1RNAPdYlLtsOK2WVE1nbS4ai7bOhpTlvju6i9IWP0rmCbJs0fS9a+H7luvKhvblJcbEbT2viISaGlxxbWRzLqBlwLBCcykSZPo3r07Y8c23cOhT58+LFy4kB49ejRZHSciubm5hx3K0tp58sNc3v3eRoXNQlmsIxQHoWmgheJ3ibb4DWIIfvlM8Xkpejz6uMvcFKSnp+PxeLjuuuuaW5RmR9wfIYQumh6PNC5s32bObiZJjhwxK4VAIBC0Mu6+GHbPTKXMYgGb7A8SVvBPGaYq/nhaRQZ3zai7WtOvtbUIX4lAIDhxEaEUgojkkUceCU6XVZf27dtH7Op1GRkZ/O1vfzvk8VmzDm+ezYZ45ZVXGpz3ODMz86jraIlUV1c3nukEIaALyadhaqrfK6ybSDWLY4Th0f3xxaYJuklllFJPiS0XXT/UJHgnFuL+CCF0IWgIEUohEAhaBSUlJSJusIaALgzDQH2sEovin8VCQkZX6nwolCS/99g0QZW4qZfE+yOt9RfcwkhPT8fn8zFq1KjmFqXZEfdHCKGLpsct/V/Yvt38RzNJcuSIUAqBQNAqqLuAy4lMQBeyLGNMjWflnSqGoqLWnjPcNP3fDBX/unJIgGHy1pWt60Oix+NpbhEiAnF/hBC6EDSEMIwFAkGr4NRTT21uESKGurq4+KQo4qw6XkXG4dOx6ToWRQJF8S8pJwMKeB+3YlNb12MhsJDDiY64P0IIXRwPpDpby6F1uQYEAsEJy/r164MrPJ3o1KeL8idiD5G7deN0OptbhIhA3B8hhC4EDSFijAUCgUDQKklPTwdg+PDhzSyJQHBi4ZLGh+07zNebSZIjp3V9MxMIBCcsK1eubG4RIoZI10VFtY+P11Sw/Nemnx2gtLS0yetoCUR6mzieCF0IGkJ4jAUCgUDQpEz7ooQ1213kGRa27FcpS4wCE2SnF0kC33NHtmrl4SI8xgJB8+CS7g3bd5ivNZMkR47wGAsEglaB8AKFiBRd+HSDi8bmsu6TUr4rimajJ47S1BhMRaZ7ZTXddA3JhKK97iaTQXiM/URKm4gEhC4EDSEG3wkEglbBoEGDmluEiCESdOH2aKQ9XMQZmsnKtBQOJEUFFxdpX1HNBbnFAHSLtqP5kptMjsTExCYruyURCW0iUhC6aHrMFjYTRW2Ex1ggELQKTtQV/+qjuXVhmiZdJ5XQ2WuQ1TGZA0kOehZVck5eKVFejVKrhff/0IX3/9CFnQnRtD8pqslkKS8vb7KyWxLN3SYiCaELQUMIw1ggELQKLrjgguYWIWJobl30m5yPV5fo5PXRprSKrnvKSCpz0aXYyYgteXil0KNnW0IMZUXeJpMlNvbEnKauLs3dJiIJoQtBQwjDWCAQtAp+/vnn5hYhYmhuXRwo1Djd7WZDTBROWSHJMHGrKnuiHZgGRPs0/8p7NWO/VUvTfXYV8xj7ae42EUkIXQgaQsQYCyKO/fv3k5ycjKqK5ik4fHr06NHcIkQMzamLvn/ZRVcDknWDX6Md2OtMfFRiVfFIEFdYBYAr3g6mtcnkaWkr3622zCRGM5ABJxJ9zQePSbni/gghdNH0iBhjgeAYUVxczLXXXovX23SfVgFGjx5NVlZWk9bxe7n77rv58MMPj1l5q1at4tZbbz0oPTc3l0suuYTq6qafS/Z4kJeX19wiRAzNoYuud+fSfVw+RTY7Bxx29tkteC0qJYqCryaPbJpkxzhQyz2Y+FejdpS7wTTqLVPXDAr3VDL/lfWMv3YtGWuLcVd7jkiupu5Lfi8bpb+zUXqZHOllflJn8I30d7KkGXTQnERj4gBSMMiWZuCtdLHO9hLb1KdZY3mVpdFv+guprILCYtD1RusT90cIoQtBQwiXnCCi8Hg8uN1NN3VTgI8++qjJ62huNE1jwYIFzJkzh+7du4cdy8jIYOrUqVRWVjaTdMeehISE5hYhYjiWuvBqJj8XGHw6v5DftrlwabDLopJiGERrBvkOG4WKgjPgmTVNKk0Tn92K02EFReJ/0VZidR3NJlOly6DKYJpEV7ixajov/mk9uiQh18xaoSkymqpiAGV2G1GaikOT+fLve/kv4MPk1+RY2pdV0bnag+J0EeP2IMkmeHRkJJAAxQJmPNOnLkFRZDBMZE1D1g1QZHyqgqEqSJKE7PWhGAYmoCMhmQaSomD6fES7NCRNQzZBVyVMJCRJwsTEonnRLDbQDDAB2QQkUBQwdUDipP15xFcbuBUbMWYFVk3DRwKxeLDgw6GbGCQhSVCOlWjTV+Nvk4jFx464lzmDvQB0oJD/aT3ZKj1MD3YiARp+L1dDnq6e+MULR6rZjEP792TAooCnlvEdZYXzT4VJI+Df30KbBJj5pb+sOwfDZb1h2Ua4cQBkbYbLzoRzavqgAxWQsxssKpzVFWKOv0df9BXHg5brMRaGsaBJWbduHa+++io7d+6kffv2PPTQQyxZsoTo6Gi2bNnCli1b6Nq1K5MnT6Znz57ccsstAAwdOpR58+bRs2fPQ5adnp7ORx99xPz58wGorq6mf//+fPnllwDceOONjBkzhn//+98YhsHll1/Oww8/DPgn/J80aRL9+vVj8+bNPP/88+zYsYPevXuTlJREWloaY8eO5e6772bw4MH86U9/AuDDDz9kxYoVzJ07F4BPPvmEBQsWUFFRwdlnn82jjz5KSkpKo3rp06cP48eP54MPPkDXdUaNGsW4ceOQ5fBHW2P1N8SLL77Inj17uOmmm1izZk0wffHixbz55pvceeedvPDCC42W01LweI7Mk9iaOVa6cPtMBr7lIenHQhRdp71bQwYSfQaVFoVyhx2rJOGxWEInSRLFsoxXVfwGMJDq1biosAzVhN/iHKxPiQNJQrOp3PjbLqJ1Ha+qovp8yECFxYECSLJMsk8DYENKPGcXlWLKMioSvQ9UYtc0JFnGiI3GwMTm8SJLYKk5x6NY0VQZWZaxeHxgmmgWFdPif2yr+I1KZBnTZkXXdWRJwuLTkEwZ3aIiWVRMrRql5vJkwJAA2W8ca4oCkgSyjKSbmIFbWJbBlECSKEpogypVsCclEZ/clraFVZxUuZ825OPDQb6UTLzpxGr68GDFhQ0bfkNUQSeeCv/vIVupUKNRfBodzYKgIWzBb/TWNUXCDWGpJkVGwvAb95jBXPWdD4BBuFEMUO2FVb/4t7o1vrnMvwG8+bX/30eBKX+CEX+EgX+DCpc/vUsq/DAV2ibUV3OTIfoKQUOIUApBk1FSUsJDDz3EddddR0ZGBuPHj2fixIlUVlayaNEiJk2axLJly+jYsSOvv+5fRz1g5C5durRBo/hwqKqqIi8vj/T0dF5++WU++eSTgwZdeDweHnroIQYOHEhGRgYjRoxgyZIlh1X+8uXLefvtt5k+fTqLFi0iLS2Nxx9//LDl+/bbb/noo4949913Wbp0KZ999tkRXV9jjB07lrlz59K5c+ew9PPPP59PP/2UCy+88JjW19zoh/E5+UThWOli9S6DH3INJCTaeHzBB0aMYVBus6IALkxSNQ3JDBlYmkRwYB1Ap2o3S7qm8mmPdhiyjFXTQTfRZRm31W9UWzS/0W0CPosFxTD8BmcNChL/S4rze0ZNM+jrpOZft92GCVh8oWu3eHwoJlg9PhRdRzEMVE0POy/4EJQkkPyebMOioltDxr6sh0I9pMD/ArLV/leWkM1aBmbNscpoO7lJ8bitFnRVJr9dLO3Iw0kCXhzIQBxO7HiJpwpZ8hFFJfEcwIeDChKoVBysTP4j3yeexY6UtvjkcL9WfUZtSEcyEgGdGTV/HWw2Nykzv4K5y0JGMcDuIvjk+6au+SBEX9H0mDUtzgy2vJaDMIwFTca3335LWloaI0aMQFEU+vfvz+zZs7FarfTv359TTjkFu93OkCFD2Lt3b5PIcNttt2G1WjnjjDPo2rUre/bsCTu+ceNGPB4PY8aMQVVVhgwZQp8+fQ6r7C+++IIbb7yR7t27Y7PZGD9+PDk5Oezevfuwzh8/fjwJCQl07NiRG264ga+//vqIr68hUlNT601PSkpqtoGN69evJz8/n127dpGTk0NhYSE5OTkUFBSQnZ2NYRgsW+b3Nq1cuZLKykoyMjIwDIPMzEyqqqrIyclh165dbN++Pfjvpk2bkCSJrKwsfD4fmZmZVFZWBle4Wrp0KQBZWVkUFRWxYcMGSkpK2LhxI7m5ueTm5rJx40ZKSkrYsGEDRUVFwRj0wLkBeTIzM/H5fGRlZVFeXs6mTZvCZAlcW1VVFZmZmRiGQUZGRpg8y5YtwzAMsrOzKSgoCNPFrl27yM/PZ/369ZSXl7Nu3TpKSkqCc6/WlScrKwuXy8UPP/xASUkJW7dupbS0lNzcXLZu3crevXvZsGEDLpeL1atXh50bkGflypX4fD7WrVtHXl4emzZtIj8/H8+BnciSiWyaeJXwx4Vi+I3FJE1HkyTOcHvp5vGhA4YkgWES5fKCabIlMYZqi4ouS2xMjcVbK2Ti01P8L25GLSNYNgwk08Qj1+STJDYlxmAxTHRJQqoTE+A3iLWDyjFlyR8kUMtIl0wzaBL6/aVS2DHAH7NrhIxhs3aZNEBN2XXzyIYRXoYEBgqBR7CDcA+m025lvyOO3aThQ6WMRHLsPfHJlprrkvlf9BkNSRJxOGOtfg9xHTaW5dXb9gL3VWFhIWvXrqWqqors7GzKy8vJyMgACPaZGRkZlJeXk52dTVVVFWvXrqWwsDDYNwTKy8vLY926dTgcjrC2X/teWL16NS6Xiw0bNrB37162bt0avJe2bt1KSUkJP/zwAy6Xi6ysrHr7mczMTEpKSli3bh3l5eVN2u+Vl5cfVb8nOBgRSiFoMkpKSmjbtm1YWq9evbDZbGExXqqqYhj1D745WmqveqWqKmadEfKlpaWkpKSEhTB07NjxsMouKChg9uzZzJs3L5gmSRL5+fl06dKl0fM7deoU/LtNmzYcOHDgsOptyZx99tkHpbVp0waAdu3aAXDZZZcBodWpBg4cCEC/fv0Afxuqj5ycHC666KKwvIEyhg4dChA8HnhpSEpKCp4f+N0DaYE8gXMDZQXKDpQVHx9/yOsN5A1cQ6CMwDWed955Ydce0AVA+/btATj33HPDyqorT0CO888/Pyh/Tk5OWDsOtLUBAwaEnVv330BdHTp0AGBEe1gQrfHGv+0Y+zQUyaBaUdhpt6LUmH+6qpLo9hBlQAI6aR4vO+wW9sY6iPFoXHSggu87hvQM+D2pNbeiJst4VRWvoqAaOqpuEF1djcdqJcrlJrt9Kr/Fx5Di8tCzpJxfU+LpWeakwqJiUyRiPT5sPg2bx4Mpy3itFlRdB0nCa7OALKPrErJRE00rSeiyhGLUCGAaSLrpN7Z1HUORsXr8wwV9NiuyYWBgBL1IRvB/fl9YvKeSSmu0P0ft7sUI7aTtLyNW87KlQyq6LNO5qAQ3jpqCZOymho6EgolHVsmLTcKrWOjj3kSl6V8V0NTqeIi1wzTWa6Q2/UEggBK4gFovBf4Sjtivp0ig11O74n8xwqr6q+qUTPTqZ/whE8WV8J8fwG6BOy7lrAeHAwe3vUD7h9B9EbhfAvfTkCFDwvYDxwPOjfrupw4dOpCTk3PIeyBwj/Tu3fuQlx241wL3Xt1+JnCvBvqSpuz3asvxe/q9pqKleYlrIwxjQZORmppKYWFhWNpbb72Fz+c7xBlHhizLYWX9nhWuOnToQFFREZqmBb2oRUVFJCf7H0aKohyyjpSUFG6++WZGjhwZTNu5c+dhG9a16ykoKAh2kLVpqH5BOF27dm1uESKGY6mLG85UueHMNkCbeo/vzK1i+OR8LDYHhiRxVmkF0bqNTe0TuXhPEd92SsZpqXnUmCb4DGRNx1D8UbuGYTLpk3NwxFmYdmc2pbv849c0rxdDUbgwdz99cwuosFnZF23njIISJF1nzvLGQ4HS09PRdZ0BfxzM2yNWY/gARcZweTAAw2qiemR8eo3VbFVQHRL3ZA8/FqoLsuvxRRRO3UyvbTuJJ5/yDu0pLk0iFhcqBmVSFIWOGBTVoNgRjyYrSKZBG6kE2TQpIYUoj0FiJVTZbaiaQXvXOgwCQ+dAWfUUnNIOM20cgXhiacDpcM8QOLs77pPvxXZqO6T/veYfGOhXPuzZBx3agrXppswLY9pt/q0ZEX2FoCGEYSxoMi6++GJmzJjBokWLGDp0KN999x0LFizgrLPOOuQ51prO2el0EhXV8DKxXbp0Yc+ePezcuZMOHTrwzjvvIElH9pZ6+umn07ZtW+bOnctdd93FunXryMrKCsY3d+7cme+//57rr7+eoqIiFi9eHHzrvvLKK3nvvfc455xzSEtL46OPPmL27Nmkp6djqT0Y6RDMmTOHZ599lgMHDrBw4UL++te/HpSnofoF4fz6669Bj9GJzvHURbeOMeS8ezK6YXLplD18HJVCmseHZOg4LQpOa63HjOE3jE2fhmz4Qx9MVUKx+v2xE/957GV2Op0kpsXy0LqrjnnZh0vX56+g6/NXhKXtfC6D8id+ACx4TR+XOseyZfJqvLN/w6conFq1G6dhR/vLH+nz1lWUbi4i5/av6Ta4C10mn4/qCE3BqNQqVzLrH6vw8w+PHtwmZBm6dqo3f2tG9BWChhCGsaDJSEhIYObMmbz88stMnTqVtLQ0pk+fHpw1oj6Sk5O5+OKLueaaa5g5c2aD8b69evVi9OjRQYPylltuIS4u7ohkVBSFl19+meeee44hQ4Zw+umnc+aZZwaPjxkzhqeffpohQ4bQsWNHrrzySrKzswG/YVxRUcF9991HSUkJXbp04dVXXz1sGTp06MDo0aPRNI0bb7yRq646+MHdUP2CcA43NvxEoDl0ocgSq/7ehf97bR+rfzFxVHn5X0qc3xiWQy+sCbKJ0wCfrebx49PRfCZWe9PIFalLQnebPBAmDwxLO/W5AZz63ICwtEBAQWLPVPp9f9Pvrk/cHyGELgQNIZl1gy4FghOcSZMm0b17d8aOHdtkdfTp04eFCxeKFZiOIRkZGcG4vBOd5tbFaWP3kuuw40qIRtdM/yd7SQKLDDFWKPeEZnMwTHxPOVAtx34seHp6OmVlZcFpIE9kmrtNRBJCF01PhfRw2H6cOaOZJDlyxKwUAoGgVSAedCGaWxdvjIvh/OIKYkqqkHy6PwhWN/2bWeNBVmsW+pDAU91002eJxRz8NHebiCSELgQNIUIpBBHLI488cshlm9u3bx+xq9dlZGTwt7/97ZDHZ82addR1vPLKKw3OexyY2utEYuXKlcHR2Cc6za2LQb0T+Vs7F3FOnfKwadhMTM3AtKuh8AoJnBU60fGNx+X/HsrKypqk3JZGc7eJSELo4njQcmelEKEUAoFAIGgS0u7ajhEdT4HdhmyadKl2s7NDQs1iGjWZTJOKuyE26djPipCeng74V7oUCATHjwppQth+nDm9mSQ5ckQohUAgaBUEJv0XRI4u9s3rjul2I2FwsstNlAyyFj5neZRbaxKjOIDwGPuJlDYRCQhdND0teeU74TEWCAStAsMwwhZqOZGJZF2MmrGfr0qjsCj+RTecXhfG1KaZgjA9PR3TNBkxYkSTlN+SiOQ2cbwRumh6yqWJYfvx5rRmkuTIES1DIBC0CtauXdvcIkQMkayLzx5ui/fZWJxPx1D5TEyTGcUBKisrm7T8lkIkt4njjdBF09OSPcbCMBYIBK2C008/vblFiBiELkJER0c3twgRgWgTIYQuBA0hDGOBQNAq2LVrV3OLEDEIXYRwu93NLUIYn/2qIU0P344Hok2EELo4Hkh1tpaDMIwFAkGroE2bNs0tQsQgdBHicJZnP178kq9x7aKD04+HcSzaRAihC0FDCMNYIBC0CkQsaQihixC63nSLhxwpZy5ovrpFmwghdNH0mHW2loQwjAUCQatAUZTmFiFiELoIIUkt4DPucZgcSrSJEEIXgoYQK98JBIJWgc1ma24RIgahixARZRjrJijNI49oEyGELpqeljYTRW2Ex1ggELQKxEIOIYQuQkRSKEVzItpECKELQUMIj3EEs2/fPtLS0ppbDIGgRdChQ4fmFiFiaI26aH/zVhJlldQqNxarQmqVi39/1bvR86zWpltV74hRJH/YRDN4sVtjm/i9CF0cD4THWHCM+fDDD5k1a9ZRl+NyuXjhhRe47LLLGDp0KLNmzULTDm8EdJ8+fbj44otxOp1h6ZqmMXjwYIYPH37U8gkEx4pt27Y1twgRQ2vSxS1X/MRtQ38kNj4eX1wM33Zqy4q2Kazp0oFbRv3c6Pkul+s4SHkENFNoR2tqE0eL0IWgIYRhHKGUlZVxLFbrfvXVV/n11195//33+eSTT9i2bRtvvPHGYZ9vt9tZvXp1WFpWVtZhG9cCwfHizDPPbG4RIobWoIt/X/gZz/X6L6UJceQlxpNqmOTJMoYkgSSxS1XYmpxEh/8roMetO1j1TVG95TT3Ah87SzWkF1xIz7mggbAOabrG97ubrl9tDW3iWCF00fS05JXvJPNYWF+CoyIvL48///nPXHLJJWRkZDBhwgSeffZZTNPk1FNP5b333mPTpk288sorbNmyhZSUFMaMGXNYHtshQ4YwZcoULrroIgBycnJ44IEHWLZsWaODUvr06cOoUaPYv38/r776ajD9sccew2azsW7dOtLT03G73Tz99NNkZWVhGAamaeJ2u/nmm2+IiopixYoVzJkzh/z8fLp168Zjjz3Gaaedxt13383gwYP505/+BPi95CtWrGDu3Lm43W5mzpzJmjVrOHDgAKmpqdx///0MHDiw0fpqM2XKFJYsWRKcy3T69OmUlZUxf/588vLyALj00kt57LHHkCSJgoICpk6dyk8//URUVBQ33XQT1157LUOGDAH8iwVYrVZkWWbs2LHEx8ezePFiYmJiyMrKokOHDkyYMIHzzjsPgDVr1vDGG2+wZ88e0tLS+L//+z/69u0b1O/ChQvp0aMHAJMmTaJ79+6MHTuW4uJiXn75ZdasWYPdbmfIkCGMGzcOq9XKlClT+Oqrr3jmmWcYNmwYAJmZmTz44IPcfPPNPPDAA0yZMoWEhAQeeOABAJ588km+++47VqxYQUlJCVOmTGHDhg2YpolhGNjtdlasWNFoe4pkMjMz6devX9MU/u4qePEz6JQCn02CGEfj53y3Cd5ZBad0gIeGw+GOhJ//Daz+FewquD24nVaqluzBEmMSl+SEXYVQEfKC1r6Lj6Qz17DjIw4JDSulSICXBEwUVJxIGMh40bBhYEPHggUPKk5M4ONT+/P5Kb0ZvTmLAXt+JcZjoKNSTQI6Vix4cGMhlnKs+A3C/7Vpx0e9BnBS6QE2piazPiUNl2yji6uUG3/aiOSOwqHp7O3YhsKYaFZ060h5VBRuIAX4yWrBXdNvWQ2Ti6tdxJgmTknCXu1iSZskDIcFZNnvmT2oi/MndIiF63sp/G2gleSo8EzXfODm883G4StSlcCm+EMkDAkMA1TZX5UshX4YCb9MhwqlME3w6mAAkgmmBBbJX4ZHB71WfSb+fIHCzVp1SRKKZGJVwapIVLpNrJKJHS9l3vA5nRPskBolM6KnzNODLDyzWuPDbSa7y01MHVRMkqMlSnQZSTM4yWEyd4SFfl39bfnNHzXW5hlce7rCsJND7TuvwuTFb32YJiiKhNeAs9pKTM/S2FliogdklwnpwgRJhngbzB+pctWpTRfp2aR9hQCAYunxsP1k8/lmkuTIETHGEYLT6aR9+/Z8/fXXGIbBvn372L59Oy+99BKlpaWMGzeOv/71r/zjH/9g8+bN3H///SQnJwcN3kOh6zp2uz24L8syZWVlVFRUEB8f36hcQ4YM4f7776esrIyEhAScTifr169n4sSJrFu3DoCvvvqKHTt28NVXXxEbG0teXh4jRowA/J+snnzySaZNm8YFF1zAJ598woQJE/jyyy8brPf9999n586dvP/++zgcDt59912mTZvGwIEDG6yvPm644YaggZiXl8eECROYPXs2vXr1YseOHdx+++1ceumlnHfeeTz66KOcfPLJLF26lKKiIu6880569OhBZmYmAMOHD2fSpEnBTjU9PZ3s7Gzuv/9+nn/+eRYvXszEiRP5/PPPKSkp4eGHH+aZZ56hf//+rFmzhkcffZR33nknaAwfiokTJ9KhQwe+/PJLnE4nEyZMYM6cOdx7770ApKWlsXr16qBhvHTp0kP+nr/88gsrVqwItoP3338fXdf5+uuvsdvtrF27lkceeaRBeVoCTfagy9kDY17z/715Hwx6CrJfavicvQdgyN+h2uPfd3vhb6Mbr+vzH+HW0BcdDYkCzgSiaVO2BSm3vMHTD9cvY6Dgoi2Bj4b+82Q0/N5VHTtR5CIDVtyAO2h0S8D61JOZPHAUWfOfoG11Rc05EuDASSog4QYS2YUDDyb+WQBOPVCM22ohylPMP84J3bPuAjvW4m3IgIFKqSOWH7p3pDTa/6JrBUpNE9Oi+B9aPp0umkZMjV8n2jSxKAr9SitYLSdAtHKwNiSCbw55lfDq9zo5hR6Wjwn1j698561lFNc64SACZZtgVUCp9fHVlA42fGvvHsohIUlgU0E3wGf4DV21xtg0anmatRqZgka3FJruraZs3QSXD1w1hrrbADfWg66nzA1lboMZWSYZOw3WlUp+o94wQTPRgP3lJtgBWWFzlcGgd7y4n7Tzzgadcf/1AfDOBp21d9vo3d6vh2Hve/h5vxl6OahdbX1GcY2OTNMv0/APNX4eK3NG26b5qC2M4qanpXmJayNCKSKIYcOGYbVawwxZgNWrV9O2bVtuuOEGVFWlV69eXHPNNaSnpzdaZv/+/Zk3bx7FxcVUVFTw1ltvAeD1eg9LpoSEBM4991xWrlwJwKpVq+jbt2/YgJbo6Gi8Xu9BscgAK1as4MILL+Siiy5ClmWuu+46XnjhhUbDRK6//nqmTp2Kw+Fg//79REVFUVhY2Gh9jZGamsqHH35Ir169gi8IsbGxFBYWsm/fPnJycrj//vux2+106tSJ2bNnc+qppzZYZufOnbnllltQVZXhw4eTlpZGZmYmX3/9Needdx6DBg1CVVX69u1L//79Wbx4cYPl5ebm8vPPPzNhwgSio6Np06YN48aN46uvvgrmueCCC/jpp5/w+Xy43W5++umnoJe6NqZpMm3aNG688cZgWnR0NG63u1mWyl2/fj35+fns2rWLnJwcCgsLycnJoaCggOzsbAzDYNmyZQCsXLmSyspKMjIyMAyDzMxMqqqqyMnJYdeuXWzfvj3476ZNm1iyZAlZWVn4fD4yMzOprKwMttulS5cC/jCgoqIiNmzYQElJCRs3biQ3N5fc3Fw2btxISUkJGzZsoKioiKysLAB+efs/Ydfg25KLz+cjKyuL8vJyNm3aFCbLrl272LksK2QUA4Ur1obJs2zZMgzDIDs7m4KCgqAu9i/7MawuA5VAN22l+pj9DiYKtbt/A0tNXQGUmi1E7YVdd8W3oV11edAoBtBlGR8OaluBGvaauvxYDYNe+/ewObltWNmdSirCHkYxVa6DHqw7Yh147FY0uxXZYcVd1/aUJdp6fTXVN/BQrnVo/T5fWNub/0P94RgNFlbXuXy0McSK7Dcka0/rVre7lOr8XZ9nPPgmUzdz/Ww5oIWM2LrXZIaMcc2AT5d+y9q9ofatm/DDzkq2bt3K5i1bySk0GqsuXK56HgfLNuSyfft2CgsLWbt2LVVVVWRnZ1NeXk5GRgYAX3/9NQAZGRmUl5eTnZ1NVVUVa9eupbCwMNg35Ofns2nTJvLy8li3bh3Lly8P3ouBfiawv3r1alwuFxs2bGDv3r1s3bqV3Nxctm7dytatWykpKeGHH37A5XKRlZVVbz+TmZlJSUkJ69ato7y8vEn7vfLy8qPq9wQHIzzGEURycnK96aWlpbRv3z4srV27dqxfv77RMh9++GFmzJjBn//8Z+Li4rjppptYvXo1sbGxhy3X0KFD+fLLLxk1ahSLFy/mL3/5C9XVoYf0kCFD2LNnDzfeeCNOpzPMsC8uLg5bflOW5bD4rlmzZjF79mwAfD4ff/jDHwCoqqpi6tSp5OTk0LFjR9LS0oLGdEP1NYaiKPznP//hyy+/xOFw0LNnTzRNwzRNSkpKiIqKIiYmJpj/pJNOarTMjh07hu23adOG4uJiSkpKaNeuXdix9u3bBw18gDvuuCMY0uJ2u+nevTslJSU4HA4SEhLCzisuLg7GdjscDv7whz/w448/4nQ66zWKwe/RTk1N5fzzz+ezzz4D4KabbqKkpIQrr7wSwzBQVfW4jdw/++yzD0oLtI+Ari677DIABg0aBMDAgQOBkJenV69e9ZZ92mmnBf8O5A2UMXToUIDgF5bU1FQAkpKSgucEfsdAWiDPGY/fDv/IArffO2a5oT9YLMGy6vXUj24L05b5PcyKTJv7R0FsbFCewDUGfrdgO/nrVfDeD1Dlf2lR8KDiRsOOk2TiKaj32o8UGS8KLnQcgImFSkDGgxWQUKgm8O0+EAXgjxUEGZOBe37mbwNG8VPbrpyzfxcAqqEjU4GMDwMLfnO7GhlP0OjeHx3LspN78/A3n/D6uQMpt0ehGjrrO7el0m4h1u3DkCUK2ybS2emi2GbBpSgogKdWKIpXVdhrsxBnmCRrOi5Z4nSXiyJVCYVQHOrdu1b6mHNttGsXG9T/G9e25aK3Ds9pECxMqikwECJRX6jEkcxEYdTMdew1/P9Kkj98Iugpxu8lrn09da/VNGuMXNNfXiDtEEjATWdZmPOz4Q8LUSS/tRs4GPBOawbtYmD0sH502K3z9s9ePLo/NGXkGXG0i/XfCzed6WX+Rj3UePA7orXaBncD+oq2wM19O9Em2p8e6CMC90ugTwiEuAX2A8f79OkTdh4QfIbWnZEicE8G/h0wYAAAvXv3PqS+zj//fCDUn9TtZwL9T6Avacp+r7Ycv6ffExyMMIwjiEPF/LZr144lS5aEpeXl5R1W4z5w4AAPPPAATz31FOB/c+zcufMRGZOXXHIJU6dOZdOmTeTm5nLOOefw7bffBo8rikJaWhoWi4VPPvkERVGCoQ1t2rRhy5YtwbymaTJr1ixuvfVWAO67776DYowBnn/+eU466SRefvllVFXlp59+Cr5RN1RfY3z99dcsW7aMBQsWkJKSAsDIkSMBf6dRXV1NVVVV0DhesmQJcXFxDYasFBWFe5ny8/MZMmQIhmHwyy+/hB3Ly8sL66zfeuutsBhj8P/eLpcrGL4C/qn74uPjUdXQLTtw4EBWr15NSUkJo0aNCnoHAjidTubNm8fs2bMpKAgZVHa7ncTERNq1a8e8efPYsWNHqwilWLlyZfCBcExJjoOdb8Ks/0KvznBj/8bPiXHAmhdhVQ50bwdndDm8us7oDL9Mg/W7oF08Sn4p7dun4J6xCsuZp4Ndx/wtH2PResgvRqlj69QXGVvLNsEH+JBwYGKnAAMbEjoyGgYyVrz4PcUGPqIAEx0LCjoGFr+NhAeHV2fJ/Gm8f8ZF7HMk0X/vFmRdRceOjQNINV7iSmJxE4uKDxUPstfNbWsy2BZ3MhOzs8hObsu+qARi8RLt2E9xbBsORMcTVe0kTfNS5PXikVSK2ycTpelUW2rav2GCquDSfFRIMl1cbiokidVx0URpXjwe0ANhCMF+1e+TbhsNY/+oMKCbyqCTwr3iF3ZRWf0XuOEjL/urTCwyeOoNN66leLdOMM7XwG/AWiS/cSnXGLa64TdmA9O1KfV8rDUMvzFqBLytErg0sMrhVeqAUlMfkl8XEkHD165Cn/bQPk4i2Q7Z+yTaR0ukerazrrorRU6o8oLDAn/9o4RHVxndS+bcDgq39tZZtcfg498k9ldInJYAA06SWL5XIlY2OK+NzGP9/HHKfbso/DzOxv+KTC7uJNMmJvT8eudqCzeeoWBVwKpKlLlNLkiDV7NNftyn81OuQZGr5qJqDPjuCSArEoO6yjxziYXU6Kb7FN9kfYWgVSAM4wjFarUGQwUuvvhiZsyYwcKFC7nuuuvYvHkzn3/+OU888USj5bz33nv4fD6efPJJioqKeP3117n22muPSJaoqCguvvhinnrqKYYMGXKQAX/gwAGmTZvGxIkT6dSpU3BQG/gHtr377rtkZ2fTp08fPv30U5YvX8748eMbrNPpdGKz2VAUhYKCAt58803AP1VcWVnZIetrjKqqqqCH1Ov18u9//5t9+/ahaRrt2rXj7LPP5vXXX+fBBx9k//79zJw5k6effrrBMn/77Te++uorLr/8cr766isOHDhA3759KS8v5+2332blypUMGDCANWvW8M033zB37twGy2vTpg3nnXce06dP5/HHH6eqqoo5c+Zw+eWXh+UbMGAAb775JoqicN555x1kGC9atIibbrqJjh07hhnGW7duDRrMSUlJ7Nix47D1F8k06YOuXSI8f/ORnRMfDVeff+R1dW3j32pQgOhPQjHpEnWDHEI0Fhtnq9lql32459alI/DoEZ4TCxx6BtneB6U8fM5y+m/dz28dkim95Ey2JsSEeRl7llWyNy6ai6+PZeLw1IPOD4SbHcnUkv27quRNappHozTNe8iBdxZZwjupKb/cnN5ojos6K1zUWWFy3/D0pw6R/5QUmVNSDk6XZYnLTz64lT5zCUSC2SGM4qZHxBgLjjn9+vVj+/btjBo1iri4OF577TWWL1/O4MGDmTx5MuPHjz+sm/u+++7D6XRy2WWXcccddzBo0KCweNPDZdiwYezYsSM42Ks2r776Kt26dePKK6886FjXrl15/vnnefnll7nkkktYsmQJr7zySqNr1T/00ENkZmYyYMAAxo4dy8UXX4zD4WDnzp0N1tcYV111Fd27d2f48OFcddVVbNq0iUsuuYSdO3cCfk/1gQMHuOKKKxg3bhx33nln8LPZoejatSvffvstl156KZ999hmvvvoqcXFxdOrUienTp/PWW29xySWX8Oqrr/Lss88Gw0Ua4plnnkHXdUaMGMGNN97IWWedxX333ReWJyEhgU6dOtGvX7969ZmQkMBf/vKXg9KnTZvGoEGDOOeccxqVoyURiAkWtB5dzPjpUtZdcRKabGKrqsZW4xm16zqD8g9gcXv4+a0u9RrFASoqKg557LhjHtpY8E6wHPLYsaC1tIljgdCFoCHEdG0CwVGQnp7ORx99xPz585tblBMel8uFw3EY06idALRGXWiazln37MRV5uNsU2P6y13p1rHhsRLp6enous7VV199fIRsBGma75AeY3Ni0xrGrbFN/F6ELpqeQulvYfttzGeaSZIjR3iMBQJBq6B2LPuJTmvUhaoq/G9OD3Z8eBqffnRGo0ZxgIha+a6ZVr2D1tkmfi9CF4KGaP5gH8FRMXTo0LAZImozbNgwHn/88XqPLVy4sMEV8D7++OODZlQQCCKZzp07N7cIEYPQRQibzdZ4phMA0SZCCF0cD1pujLEwjFs4dQdcHS433HADN9xwwzGW5sRj+PDhRzSwR9B0HDhwQExDVIPQRQifz9fcIkQEok2EELoQNIQwjAUCQaug7nLgJzJCFyEaG+gbERyHkT6iTYQQumh6WvLgNRFjLBAIBALBcWD3HfUkBhbjEAgEEYEwjAUCQavgULH2JyJCFyF0XW9uEYJ0TlQpuwdqz1Z8dw8Jc0LTf7wVbSKE0EXT418pM7S1JEQohUAgaBUEVjIUCF3UxmJp2mnQjpR4h4pnwvGvV7SJEEIXgoYQHmOBQNAq2LNnT3OLEDEIXYTweDzNLUJEINpECKGLpkd4jAUCgaCZOfXUU5tbhIgh0nVx5XyNRQXhafoEBbkJ5vkVCzn4ifQ2cTwRuhA0hPAYCwSCVsH69eubW4SIIZJ1sSGvxiiWwjdlRtPEAjudziYpt6URyW3ieCN0cTyoc4O3IIRhLBAIWgUXXXRRc4sQMUSyLs5eQP3PySaa3ykuLq5pCm5hRHKbON4IXQgaQhjGAoGgVbBy5crmFiFiaJG6aCKnUmlpadMU3MJokW2iiRC6aHpacoyxMIwFAkGrYNCgQc0tQsQgdBEiMTGxuUWICESbCCF0IWgIYRgLBIJWgfAChRC6CCE8xn5EmwghdNH0mHW2loQwjAVHxL59+5q0fLfbTXFxcZPWIWidCC9QCKGLEMJj7Ee0iRBCF4KGEIax4LD58MMPmTVrVpPWcdddd/Hrr7/We6y6upo+ffqQl5dX7/FJkyYxZ86cphQPAMMwmDhxIh9++GFY+pIlSxg5ciR9+/blgQceOCIDPycnhwcffJBLL72UgQMHcscdd7BmzZpjLXqrJjMzs7lFiBhapC6ayK301Bft6HBPHrYH97M378SdoaJFtokmQuhC0BBiHmPBYVNWVoZpNu1HkbKysiYt/2jJz8/nxRdf5LvvvqNPnz7B9N9++40XXniBN954gx49ejBt2jSefvrpw3qRyMrK4vHHH2fChAm8+OKLKIrC119/zcSJE5kxYwbnnXdeU15Sq+GCCy5obhEihkjThdOjsXAj3PlNw/kWb9cY1v3oHks+3aSi0seyn0qZsOgcPCkxlDisGLrJGa94oLqCPrEu7BhceIaNs6INMpZVU7E/1LedfaHEJZe3oarQS1HeAZa/XYGOjATYDANZkdEVBcPjI23XXgxJBUNHUVUcmoEiy7hsKpYqN0l5VbgNsCgSNt3AIpnopgl4Sb26DUZJFfzqRD7gQUbHwIeJgoqEDRcKOhYUzJr/LJhYqMJOGQ4qUAAZFQMZBQAvMlbkmpIMdGTgHEDjlfCHfizw0HVw7klwVlfo2AZkGQzDf1yu5TszTX+6ooQrXNf9+QJzUPs0/7+WyDUvIu3+aI20tAF3tREeY8EhycvLY8CAAUyZMoX+/fvzz3/+k9WrV3PrrbcCsGnTJu6++24GDBjAtddeS3p6+mGVu2XLFsaMGcPAgQMZNWoU77//PgATJkygoKCARx99lIULFwKwYMECLr/8cgYPHswHH3wQVs7mzZu59dZbgx7aysrK4DFd15k3bx7Dhw/nsssu4+mnn6aqqgrDMLjyyiv59ttvg3nXrl3L5Zdfjq43PI+qz+fj5ptvpkePHpx55plhxxYvXsyAAQPo1asXdrude++9l++//75Rr7Fpmrz00kuMGzeOq666CpvNhqqqXHHFFYwdO5bdu3cDMGXKFGbOnHnQ+WvXrmXw4MHBv0ePHs2LL75I//79GTlyJEuXLg3m7dOnD9u2bTuojEOV3dL4+eefm1uEiCFSdGEYBjEzNWJeq2UUH+p5KcEV/4FrPtd+V11FVQbnzHJy2sMHuHxSIXd9ZcGbGkd1jJUkXUexyThtFpIVlY0VUTj2W8hcCTM/N3EXgkWWsCgyqizzyxp4d9JO/j2riOUfG6hR0diio7DbbditFmyKQpRp4rBZOHBqdzxd22N2SCUWiRjdxOHTSap045BkKk5KoSMSqTrYLFbspoVorFitcWjp5Vi+8WI9oBCFhbYYpCHhIBaJBCSSaINMHBLRWIlGwYEFmRQMOmHDgUR7JNqgkAKkAG2w4sWChgOdKCAKiKYeT1gl8PQnMOIl6PJ/oFwHf5oOMTf6t4U1/eTinyDhFoj6M7wZ6lOYvQQcf/YfW/wTXP0CWEf7t2un+o3pCCRS7g9BZCIMY0GDOJ1O2rdvz/Lly7n55psZMGAA7733HqWlpYwbN45BgwaxYsUKnn76aWbOnElWVlajZU6bNo1LL72UjIwMpk2bxrx589i9ezfTp0+nXbt2vPjii9xwww18++23vP3227z22mv897//DRqJAF6vl4cffpjBgweTkZHB1Vdfzbp164LHFyxYwKpVq5g3bx6ff/45brebadOmIcsyl19+eZjBuHjxYoYNG4ZS1xNSB0VR+PDDD7n33ntR1fBHzK5duzjppJOC+wkJCcTFxYXJXB979+4lNze33pi3m2++meuvv77B8+uyY8cOLBYLy5cv57HHHuPpp59m+/btR1RGS6VHjx7NLULEECm6mLHWxHmEdu7n2+DXA0duUM3K8rG+wOSMcif/i4ums6ZhShIX7y7h3Pxy+u8pwWqYSIbJybrJAUUhUdM43emqWYPAb7FLEpiyhNdqBUlCQgoeUw0jZNdLEqphYMoSkm5gdXuxeWpdrCRh8enYXR68soyKidVnBJc7sHoNKqw2FExUTOz4sKDjxIFWY8J6sOHEgYoBmMgoGKiAhIadalJq/Ni1UQm8fUj8jof8R1ng8vq3cTWhafe9BRXV4NXg3n+C1wcen/9vn+Y/dvds+OLHUDmf/QArfznS2o8LkXJ/tG7EAh+CVsywYcOwWq3Y7fZg2urVq2nbti033HADqqrSq1cvrrnmmsPyGlutVjIzM8nMzCQtLY1Vq1bRpUuXg/J9/fXXXHHFFZx88slBL2yAjRs34vF4uOWWW1BVlYEDB4aFNnzxxRfcddddtGvXjujoaO677z4WL16Mx+Phyiuv5JtvvsHj8eDz+Vi5ciVXXHFFo3LLskxKSkq9x9xud5h+AOx2O263u8EyA6Ejx2qAUFRUFOPHj8dqtXLBBRdw4YUXsnz58mNS9rFg/fr15Ofns2vXLnJycigsLCQnJ4eCggKys7MxDINly5YB/pHjlZWVZGRkYBgGmZmZVFVVkZOTw65du9i+fXvw302bNrF9+3aysrLw+XxkZmZSWVkZHH0eeBHKysqiqKiIDRs2UFJSwsaNG8nNzSU3N5eNGzdSUlLChg0bKCoqCr7kBc4NyJOZmYnP5yMrK4vy8vJg3bVlysnJoaqqiszMTAzDICMjI0yeZcuWYRgG2dnZFBQUhOli165d5Ofns379esrLy1m3bh0lJSXBuMi68mRlZeFyufjhhx8oKSlh69atrF+/ntzcXLZu3crevXvZsGEDLpeL1atXh50bkGflypX4fD7WrVtHXl4emzZtIj8/P3hthYWFrF27lqqqKrKzsykvLycjIwPw36cAGRkZlJeXk52dTVVVFWvXrsVwV/2udrJhXTYul4sNGzawd+9etm7dGryerVu3UlJSwg8//IDL5SIrK4vKykr27dkBgIGEDFQqMp1Lq4MPOathcmqZE6thIJsmMmbDI+YDB2o9083aS1bXeENlMzRTq67I9R6XDTOsyIb+lupIFNivT06pCcb81y7Np0BJSQlO3RtKtCisWr2aSmcVuhq63tp/B9hfXtpsba+wsDDYNwTKy8vLY926dezduzes/tryrF69+ojbXt1+JjMzk5KSEtatW0d5eXmT9nvl5eVH1e8JDkYymzpoVNBiycvLY8SIEaxevZro6GjmzJnD9u3beemll3j77bfZuHFj2Cf4Tz75hMWLF/PWW281WG5ZWRlvvPEGWVlZlJaWMmTIECZNmkRUVBTDhw9n0qRJ9OvXj3vuuYcLL7yQm2++OXju+eefz3/+8x9+/vln3n777bABcM8++yypqamMHTuWiy++GEVRkGvFyHm9Xv7973/TpUsXbrnlFsaMGYOiKMyZM4d///vfR6Sbu+++m8GDB/OnP/0JgAcffJCzzjqLMWPGBPMMHjyYGTNm0Lt370OWs2/fPkaOHMl///tf2rZtG3bM6XSiqio2m40pU6bw9ddfY7VasdlsnHrqqTz44IMUFxfzyCOPsGLFCtauXcuLL77IJ598Eixj6tSpaJrG5MmT6dOnD9HR0UiSRHx8PIMGDeKee+7h2WefJSEhgQceeOCIdBBp5Ofn0759++YWIyKIJF10mq2R66ReQ7M+HjgHXhl05PGpVR6TGxe6WP+zi/aVLn6Kj+Vcp4tEnz9EygQKVAWvIlNlGJzh8VKuKsg+jbMqqzGlwNRSEpJpElddjWZRMSUJi0/DlGQkw8CqayimiWSaGJKE7NWw6Ro2lxfF6yW+3IWiG0iagST5vWXt8qvQkXDbFGxef/yuzyoT7fP4DXUMrPiIxYUFjVJi8WIhCjdJlGMg10QK+w16MLHiIoFcfCSiYASVK+HBRjESDZvMh/wZHh0FH2f5Dfu542DwmfDDVhjzGlR7YOZf4JqaGN3P1sCD/4JoO7x7L8xfDXP8xirjhsLMO474dzweRNL90VrZKz0Xtt/JnNxMkhw5kRsdL4gYJOngLrRdu3YsWbIkLC0vL4+kpKQGyzJNk23btvHwww8zefJkfvvtNx5//HE++uijMKMSIDU1lfz8/OB+SUlJMA44JSWFoqIiNE0LhjUUFhaSmpoaPP7EE0/wxz/+EQBN08jNzaVjx44AXHHFFSxfvhxZlg/LW9wY3bp1CwubKCsro6Kigq5duzZ4XlpaGp07d2bVqlXccMMNYcfmzJnD5s2bmTt3LgCjR4/mgQcewOfzMXfuXJ566inuu+++sHMCOgqEhRQUFPCHP/whePytt96iR48eFBUVMX78eDp06HA0lx1ReDye5hYhYogkXewdF3rMXPZPjeWlHHJJ6OoHZByW3/chM8Ym8eVtgYhaP23HVqHIMlbTxC1LeCXIM012TU0iMdZ6UBnbNuXTpUcqlnoGjlWWV2GY8PdRG0FRkADV62Xg9SpbP9mPp8zA61PBKiHrEhZVxefWuGL6mXS76TSKfy0ka+RXVG0zoK3EwPUjSWx/8JciTdOI+l8++z7fhK/AiTr+j+BQUBWF0g9+xPf4KiTcGH3boiyYhvbo+1T9+wfUmiALGRkPCgo6biDG+BRJkti1a1ej/VGQF24O3z//FNj02sH5Rl3g3wL88WSYdefh1dGMRNL9IYg8hGEsOGysVitOp3+6o4svvpgZM2awcOFCrrvuOjZv3sznn3/OE0880WAZkiTx0ksvMXjwYO644w5SU1ODHkwAi8USrOOKK65g4sSJDBs2jJNPPpnXXgt1zL179yY2Npa5c+dy1113kZ2dzQ8//BA0Aq+66irmzZtHt27dSEhI4B//+AcrVqzg008/BeDyyy/nn//8J6ZpHhNP6dChQ7n77rsZOXIkp512Gq+//joXXXQRCQkJjZ774IMP8re//Y3Y2Fguu+wyTNPkq6++4tNPP+WVV145KL8kSSiKUu9gwYqKCt555x1uu+021qxZw48//shDDz10UD5Zlg9ZRkulNV3L0RKpulh2p4o07RBBxxK/2yg+FP+86ieqqlxMW9uHAWfYeWVMw17CHqcd+nhsfAwAM1ZefNCxy++5sFFZkk9vw/Df/tJoPlVVST2rE6lndTroWNJjV8JjV4bn/2Ai0R8clBUAS62/I7VNNAdCF01PSw5FEIax4LDp168fH374IaNGjeKzzz7jtddeY8aMGcyePZuEhATGjx9/WBOnP//887z44ot88MEHWCwWhg0bxogRIwC/Qfvss8+Sm5vLnXfeyQMPPMAjjzyC0+lk9OjRWK1+L4+qqrz66qs888wzfPDBB5x22mn07ds3WMftt9+Oz+djzJgxVFZW0rNnT1555ZWgdzkxMZEzzzwTr9cb9DIfDaeeeipPPPEEf//73ykuLqZ379489dRTh3Vuv379eOGFF3j77beZMWMGhmFwyimnMHPmzKDHG2DhwoV8+umnmKbJSSedxOTJk6murg4rKzY2lsLCQoYOHUpSUhIvvvginTqFHrC33XYb4I9/7tu3L9dccw0vvPDCUV9/JBAbG9vcIkQMQhchrFaFn2ac1HjGVo5oEyGELgQNIWKMBYJWwtq1a4PxxiciOTk59OrVq7nFiAgiWRfSNO2QoRTmxGPrq0lPT8fpdB4UpnQiEslt4ngjdNH07JaeD9vvYj7eTJIcOWJWCoFA0Co47PjJEwChixB1Z4s5URFtIoTQhaAhRCiF4JgzdOjQgz7xBxg2bBiPPx6Zb45NIXdGRgZ/+9vfDnl81qxZnH322UdcruBgfv31V7FKYA1CFyECYxZOdESbCCF00fS05JXvRCiFQCBoFRiGETY934lMJOtCmn7oFT/MCcc+lMI0zeAYhhOZSG4Txxuhi6ZnlxQ+dqWr+VgzSXLkiJYhEAhaBd98803jmU4QIlkXx9r4bYzy8vLjWl+kEslt4ngjdHE8ECvfCQQCQbMycODA5hYhYoh0Xcztd3BaUxnMhzNl4olApLeJ44nQhaAhhGEsEAhaBYFlUAWRr4u7zlcxJ4RvTUVg2fUTnUhvE8cToYumx6yztSREjLFAIBAIWiXp6ekADB8+vJklEQhOLHZKL4btdzMfbSZJjhzhMRYIBK2CjIyM5hYhYmgpupAfK8MyuZKeTxU1WR3CY+ynpbSJ44HQRdNjIoVtLQnhMRYIBK0CMdI8REvQhfJ3F4ZdAUlC0gw6eyrYNeXoV6GsjZiVIkRLaBPHC6GLpmeHNDVs/yTzkWaS5MgRLUMgELQK1q5d29wiRAyRrov//lKJRTLpuq+Ck/aUkVjlZbdmbZK6Kisrm6Tclkakt4njidBF09OSPcbCMBYIBK2C008/vblFiBgiXRcf/6KTVOZGMfyPzMQKN1ajaT5eRkdHN0m5LY1IbxPHE6ELQUMIw1ggELQKdu3a1dwiRAyRrouTEiVqB/GZAFrTGMZut7tJym1pRHqbOJ4IXTQ9LdljLJaEFggErYI2bdo0twgRQ6TqIuHRIm45C0xkdEUGn98oroqy4LNISE9WQowNTJOKB63EWo/+gWqxWI5e8FZApLaJ5kDoQtAQwmN8jNm/fz+aduglTwUCQdMgYklDRJou5n9fgfUpJ6pm4f31FubkKMS5/f2kBNg9GqbNCnYrVOvgNoh72cuxGBuu6/pRl/F7KCmq4u+9F/HSmYuapf66RFqbaE6ELpqeljyPsTCMjyHFxcVce+21eL3eoypn+PDhZGZmHiOpmq5M8I/unThxIh9++OFBxxYsWMCkSZOOqLzc3FyeeOIJLrvsMvr378/o0aN5//33m+3hdiyprq6mT58+5OXlNbcorRJFUZpbhIghUnTx8+5ybA8UcftnXtqXuUjw+Eh2+Uh0G2hyyBusqTIoErj10FPJa7C73DhqGSTp93mddY/niM/Z8v12fs3awfJ//cyUO7ax+/ST2XLGKYwdvSmY58GBmTx52bc8PmTNce0LIqVNRAJCF4KGEKEUxxCPx3NCxbPl5+fz4osv8t1339GnT59gusvlYu7cubz//vtccsklh13erl27uOOOOxgxYgQffPABKSkpbN68mRdeeIENGzYwffr0priMJufXX39l7ty5bNmyBYDbb7+d7t27c8cdd3Duuec2s3StB5vN1twiRAyRoAvpviI6mZBslTG8Bj5ZxmKYmIDNZ5DfNpqkcg+mBMWJUeDTIFoFpcYy9mhUuA3g6IyYIzWMv5Nv52zTi4RCNTZ2RyezU27HtRUP8E3qc0iaRlFUIqphp5tvD6Yks0ntxObuPShPTEAxTSwuF3pCAtTUrVksTBz4HSWlJVhO6UGF1R/eMffGnUzJ6HBU13e4REKbiBSELo4HLSuuuDbCY3wI1q1bx6233kq/fv0YPXo0a9asAWDKlClMmzaNO++8k379+nHLLbewefNmAG655RYAhg4dGkxriE8++YSRI0cyYMAAxo0bR25ubvDYjz/+yA033EC/fv24//77qaioAPyfBefNm8fw4cO57LLLePrpp6mqqjqsMgP861//YuTIkRQUFFBZWcnDDz/MoEGDuOqqq3jmmWfwHIanxOfzcfPNN9OjRw/OPPPMsGMTJkxg7969jBo1qtFyavPyyy8zZMgQ7r//flJTU5EkidNOO41Zs2bx008/BSdlv/vuu7nwwgvp27cvffv25f7778fpdAKwadMm7r77bgYMGMC1114bXPkKwr3meXl59OnTh+rqasA/32ng91u7di2DBw8Onvfjjz/Sp0+f3+VxX7t2LXfddRcDBw5k/vz5ALz55ptcd911TJgwgeXLlwP+dvXCCy9w8803069fP8aNG0d+fj5ut5v+/fuzYcOGYJnffPMN119/PQB9+vRh27ZteDwebr/9diZMmMBdd91Fv379OP/887ngggvo168fEyZMOOiaa3P33XfX6/VvSYiFHEI0py463lvESfcW0VaC/e1jyW8fz4EEB1Ga3/trACVRVrwek4I4O/tTotFkwJRCRjGAReHzX44+LK3216bKvU6WPfM/Vs7ahqsyVPbnnxQze1Y+e3e76WWaWPAhSVYqHW1I8/o4Ta9gjzKB3xJ6sinlTHITuxPvqCD9zGEs7nUp7RzluGNjSXa5SXB7sKgqUdW1nCSGiaLKdIhPwGcNxTzvbZ/KhDH/444/b+WOG7Yc9bU2hLg/QghdCBpCGMb1UFJSwkMP/T97Zx4XVdU+8O+djVVARUVxX9LMysys3FMTzaXUMstc2jR7zSzXzEory33vNTWzXEpt0zB3DRyl5JXE4hdKLpgICDIwMDAMM3Pv749xBka2QUEYvN/PZ5S5955znvPcM/c+9znPfc7bPP3004SFhTFx4kSmTZvmiEvas2cP06dP5+DBgzRs2JDVq1cDOAyf/fv306ZNmxLbiIiI4L///S/z5s3jyJEj3H333bz//vuO/X/++Sdr1qxh9+7dJCYm8t133wG28IRff/2V9evXs3PnTnJzc1m0aJFLdQJs27aNXbt2sXbtWoKCgtiyZQsKhYL9+/fzzTffEBsby969e0vVkVKpZPv27bzxxhuoVM4TD3PmzGHx4sXUqlWr1HrsmEwmIiMj6du3b6F9AQEBdOvWzWm1osmTJ3Ps2DH27dtHUlISu3fvJj09nQkTJtCrVy8OHz7M3LlzWb58ORERES7LcSNWq5UlS5bc9As8GzduZNCgQTz11FN4e3sDNm9Fr169GDNmDOvWrXMcu3v3biZPnsyhQ4do2LAh77zzDp6envTo0cNhQAPs27ePfv36ObWze/duGjZsyOLFi1m/fj1arZb+/fszYsQItFqt23rby0KDBrfH8+YOVJYuGryRige2G4voqSYgM5eGiZn45Fmxm6c6Lw05vhrwVtk8qiKAAF43TGAK8EG0ioXhtxaaptHY8iNnXzXy9XO/ExmeyW8/p7LhlWisFpHPliUR+lM6J09k8+HsBIz4AAr2N+jM4QYP8kujR7EoBE42vB/BIqGUJLyNRo606saV2jW5XDuQE40fJCAr/4FTYxVtLxfaUQh4WCVyvD1v6KNAep4GFApQKnn5ubhb6mtJyL+PfGRdVDzunJVCNoyL4NixYwQHBzN48GCUSiXdu3dnzZo1DuOoe/fu3HXXXXh6etK3b18uX75c5jb279/PgAEDaNeuHUqlkldffZWpU6c69j/zzDPUrFkTf39/OnXq5IhF27VrF6+++ipBQUH4+PgwadIk9u7di8lkKrXOX375hWXLlrFq1SqCgoIA203jzJkz7N+/H7PZzJYtW3jqqadKlV+hUBAYGFjkvjp1yr56VWZmJhaLhdq1axe5PzAwkLS0tELbLRYLeXl51K5dm/DwcOrVq8eIESNQqVS0a9eOIUOGOHmNy8r3339PixYtiu1raSQlJdGiRYsi97Vs2dJp7PTv35+OHTvi4eHBG2+8QUxMDFeuXKF///4cOnQIURQxGo1otVonw9hoNLJ161aee+65m5KxunDu3LnKFqHKUFm68CR/AtU710KNbDMeZit1DHkk1fTiYkN/Mv004KkClQI8lIAECruBfAOCwLd/3dr7BUajEYDU0zqyhXzjW59mJjMlj7gzRsc2UQQFAhmaGlzzDABAEhQk1KhDmsYHyW7sCgqSagaQq9Fg9NDwb2Bd/NPz61GbLFgLxFAjSVgVAhrRuZOCKDrCLez9rSjk30c+si5kSkI2jItAp9NRr149p23t2rXD09P2tF+zZk3HdpVKhSiW/QWRG9vw8vJySjru5+fn+FutVjumA5OTk/nggw/o2bMnPXv25LnnnkOlUpGcnFxqnX/99ReNGjXiwIEDjm1jx47lqaeeYvPmzfTv35/XXnuNf//9t8z9uVX8/f1RKpVcvXq1yP1JSUlORvPKlSvp2bMnISEhGI1GOnToQHp6OvXr13cqFxQUREpKyk3JlJGRwaZNm3jzzTeLPWb48OF069aNbt268cknnxTaHxwczPnz54sse+7cOYKDgx3fGzZs6Pjbz88PLy8v0tLSePjhh5EkiVOnThEeHk6rVq2cyr300kvce++9Lietf+KJJ3jssccYOnQoX3/9tUtlyotTp06RlJREfHw8MTExpKSkEBMTQ3JyMpGRkYiiyMGDBwE4cuQIWVlZhIWFIYoiWq0Wg8FATEwM8fHxnD9/3vF/bGwsTZo0ISIiArPZjFarJSsriyNHjgC2B1GwzaqkpqYSHR2NTqfj9OnTJCQkkJCQwOnTp9HpdERHR5OamuqYabCXtcuj1Woxm81ERESg1+uJjY11ksXeN4PBgFarRRRFwsLCnOQ5ePAgoigSGRlJcnKyky7i4+NJSkri1KlT6PV6oqKi0Ol0jlCeG+WJiIjAaDRy4sQJdDodcXFxeHl5kZCQQFxcHJcvXyY6Ohqj0Uh4eLhTWbs8R44cwWw2ExUVRWJiIrGxsSQlJTn6lpKSwsmTJzEYDERGRqLX6x0zOPbrSVhYGLnkv4GuKLBgh1UhYKrhgahUgLpAzLAk5VvSRd2NJIluQQbi4uIc/YmLi0On03HixAmMRiMRERFFnmutVotKpUKSJPR6PcnSFTwsZkfVXr4C/179h3r1CxqjEr7k4Wm1oBLzj/XPy0AhXZcXkCQJa4ElhUVJoOG5VBqdS6X+JR0t/krCXHCW6bphbPJwXt1PIL9OOxU19uzhAxU99gqeq9s59vR6PZGRkRgMBk6ePElKSorj2mCvLzExkaioKO6++26n9gvKEx4ejtFoJDo6msuXL9/02NPpdERFRaHX6yv0uqfX62/puldRuLPHWJDKIx9ONeOXX37hm2++YevWrY5tGzZsoE+fPmzcuJGAgAAmT54M2H4ACxcuJDQ0lMTERAYPHszRo0cd0+bF8eGHH+Lr68vbb78NgMFgYP369UycOJGhQ4cyffp0unXrBsDy5cvJyMhgzpw5PPnkk8yePZuHHnoIsHlMExISaNSoEfPmzSuxzrfeeotatWrxn//8hy1bttCsWTPi4uLw8/MjKCiI1NRUlixZQnZ2NqtWrXJZX+PGjaN37948++yzTtvXrl3L+fPnWbhwoUv1vPXWW/j5+TF37lyn7WlpaTz99NO899579OrVy6k9s9nMu+++i4+PD506deKrr75yipVduXIlly9fZtGiRQwaNMih1xvPVWhoKDt27GDz5s2cPHmSGTNm8Pjjj1O3bl1eeuklp7Jl4dSpU7zxxhtMmzaNrl27EhISwvfff8/FixeZO3cu77zzDiEhIcyZMwdfX1+Hhz8jI4M+ffqwe/dugoKCWLx4MZIkkZKSQqdOnZxijNesWcOyZcsYMmSIYzvYQloKjtUb+3z58mVeeeUV3nvvPTZt2lTkOXQntFptmc9PdaUyddHwejiFSQGe130GVoVAfEP/fI+oKIEo4ak3oRIlsr3VSJ62cAIHksSrLaysG+pZqA1XCQ0NRa/X88ILLwBw7e8Mjn0Rj7KGhp5vtKRGoAZJkti0IYWryRYGD61J/Xtm4Eku6Z5Kzvi1wsdiQGkVkURfrngEcq2GH4IocalhQy43CUQQJRrHp+CTncvdZ5KQBIGoB5txuUkgYgHj2S8jE7NSidHX26EHTV4eglLCJNiM6DkLfGjUOLhwR8oB+feRj6yLiuessNTpe2vp7UqSpOzIHuMi6NKlC0lJSezZswer1crRo0fZunUr/v7+JZZzxLJdfxGsJPr3788vv/zCmTNnsFgsbNy4kZiYmFJjWQcOHMj69eu5du0aFouF//73v7z55ptIklRqnWq1mvbt2/PEE0/w0UcfIYoiP/30E5988gkGg4GAgAA8PDxK7WdF8fbbb3Ps2DGWL1/u6F9MTAyTJ0+mffv2RWa4EAQBPz8/cnJy6NKlC2lpaWzbts1RdufOnfTv37/MsmRnZxMZGem4od4sDzzwABs2bCAsLMxR17hx49i2bRvz588nJCTEceyePXs4c+YMJpOJFStW0LFjR0fIS//+/dFqtfzxxx88/vjjTm3UrFmT999/n9WrV5OcnOyybIIgoFAoqkUqPEC+0RWgMnWRsKoO51fVIVWEHKVAngB6lUDtdCOCKKG0iNRJzcI/w0iwIZd6OSaC03PAUngcrhh069kDCl7PAtsG8NTS9gya25YagbbrtSAIjHmlHtNnB9OmrTcp6DDhjSbXgxaZiWhyzfzt0YhGx56hpjWFppnxeBp11DEmUy8xnTopGQRmJ1MrI53Ytg34676G1NOnorhhJtFqscCVZPyzcxBEEbXZTI1sI//ddg8bvr2LDd/eVWFGMci/j4LIuqh43DmPsZyurQgCAgJYvnw5S5cuZcGCBQQHB7N48WICAgJKLFe7dm26dOnCkCFDWL58uVMKsxt56KGHmDRpErNmzSItLY3777+fefPmlSrbiy++iNlsZuzYsWRlZdGmTRuWLVuGSqVyuc433niDp59+mh07dvCf//yHefPm8eSTT2KxWOjQoQOzZ88uVY6KoFGjRmzevJnPP/+ckSNHkpOTQ/369Rk4cCDPP/+8U9ql5cuXs3r1akRRpEWLFrz//vv4+fmxatUqlixZwpo1awgICGDixIn06tXLUW7mzJkoFArHwgF2w9RqtWKxWPjyyy+57777sFgsvPnmm46HnVuhdevWLFu2jJycHLp3785XX31V5Msf999/P59++inx8fF07NjR6dzdc889qFQq7r333iLHYZs2bXjiiSdYunRpqR76kJAQRFHE19eX/v370717d6fZEXflyJEjTuf6TqYq6MK0qg4t303lcraAmCfSJM9EQJYt402eABpl/u3SwyqiNomYNZJTnK0uWyQ44NbStaWnp5fp+FbSVsy5uZyrO4O7M5fQCLj/+r66uhnFlLLN4H3U4SckUcCjYy3qXs0gsVZtEASUZjMvvetH++6dmNldS1BuLqIoEdT61l4sLAtVYUxUFWRdyJSEHEohI3OdtWvXAjB+/Pjb3vaNYQ9F8frrr/Pkk086eZllZNyBmT/o2XLUjOf1PMbpGgU+VgkPq+32IwJX/DXk+V8PM5AkkCTEaeqbXqADcLx4O2jQoHLoRdk4vvYvQrfoyfP2YPjrPjzypGvvAMjIVAdihWVO3++W3qokScqOHEohI3MdDw+PKpn4PTk5mcOHD3Pu3Dl69uxZ2eJUWewvnMhUPV3MH+bP6Xn+JCvhEtC1hYjeR+OYYs1TK8gTBDDm2VJDiBLh/Sy3ZBTbKavHuLzoMv5e5mu7snT/Q1XCKK5qY6IykXUhUxJyKEUFMWPGjGLz59avX58dO3bcZonKRkhISJELQYAt3nXWrFllqu/MmTO8+uqrxe6fNWvWTcUClydjx46t1PaL45tvviE0NJR33323ShruVQV5ajSfqqiL2r5qDMttqRzHbkjFasmfrBRFCQQl0vs+BUrcehgTOGcRupOpimOispB1UfG4WyaKgsihFDIyMtWCiIgIOnfuXNliVAmqui6Gf5lKZDSoCtx9LtfwxPRJjXJtJzQ0lMzMTEaOHFmu9bojVX1M3E5kXVQ8fwvLnb63lSZXihw3gxxKISMjUy144IEHKluEKkNV18VjrQrPfAgV9O66j49P6QfdAVT1MXE7kXVR8bhzVgrZMJaRkakWnD17trJFqDJUdV1M6OZHund+jLFJpcB0PWNFeWNf+e5Op6qPiduJrAuZkpANYxkZmWpB48aNK1uEKoM76EL3SQ2SvFUk+6hJsIpIq8q+lLwryHH5NtxhTNwuZF1UPO688p388p2MjEy14Nq1a9SqVauyxagSuIMulAoFOQsq/sU4s9lc+kF3AO4wJm4Xsi5kSkI2jGVkZKoFpS3Dfich6yIfpfLWFgipLshjIh9ZFxWPu3mJCyIbxjIyMjIyVYoR78fzY4YPVoUST9GCbmEgHp5y5J+MjEzFI19pZGRkqgXF5d2+E3F3XezS+9DDkMMwXQbdDDn0e+XCTddltVrLUTL3xd3HRHki66LikbNSyMjIyFQygYGBlS1ClcHdddEuJ5eGeWZ8RJH6eRYEDzUhz91cJgG1Wl3O0rkn7j4myhNZFzIlIYdSyMjIVAv+/fdf+YWa67iTLoSPTDYXjQBYQXrPA2/R2ceU4uFBM9PNvURnMlVMGjh3w53GREUj66LikWOMZWRkZCqZ1q1bV7YIVQZ30YWw2AIeAgjXb6JKiSYrshE91NS2WPC3ilxRqzjrqSEg9+ZuV15eXuUosfviLmPidiDrQqYk5FAKGRmZasGpU6cqW4QqgzvoQlhsAUkCBBCvb1QquJacR6JGzU+1/NlSOwBtDR9q5pkJyLPcVDvZ2dnlJrM74w5j4nYh66LikfMYuwlXr16ldu3aqFR3VLdlZO4IOnfuXNkiVBncQheSZDOI865bxQLgoSDHzxuF1QRqJRZJwmIyMyAxlauam0u75ufnV+T2BV33oc6RsAgSbxzthZeP5831wwVShQnUIJMcfMmkJiDhjYE61tUIittjNLjFmLhNyLqQKYk7xmOclpbGsGHDyMvLq9B2hg8fTkRERIW2cbOMGzeO7du3l1t9v/76K6NHj3balpiYyIQJE+jevTtDhw5Fq9W6XF9OTg4rVqxg8ODBdO3alYEDB7Jo0SL0en25yVyZjBo1itDQ0MoWo9py5MiRyhahylDVdbHouMUWPmGVQCWAhwLUgu31dVFCVF83ggWBGoCPKJLuobmpttLT0wttW/DAL2DVkOfni+jry4rex26+M9e5du5akRkwEoVXyKUmF3kQJRJNiaMJ/2BFzWXlxFtu11Wq+pi4nci6qHjcOSvFHeM6NZlM5ObmVng7O3bsqPA2KhuLxcLWrVtZu3YtLVq0cNo3c+ZMOnXqxKpVqzhx4gSzZs1i+/btBAUFlVhnbm4u48aNo1atWixbtoxmzZqRkpLCmjVrGD16NJs3by7W81OVSU1NZe3atURFRZGYmMjixYv57rvvGDJkCE899RSC4F5TTFWZXr16VbYIVYaqrovopOt/2F+6EyVQCKAQUEoSVklyxB03MebyV00/GmQZbqqtmjULr66nEkGVZwaTCatajUWlZH7PX5gZNqDM9X+r2gR+SiSNAAgoRAllhoWAF+ry+Jf9UKDBQE1qoEeDlevBI9QmiQu05R/hLVpJy26qb2Whqo+J24msC5mScEuPcVRUFKNHj6Zbt24MHz6c33//HYA5c+awaNEiXnnlFbp168aoUaM4c+YMYPPWAYSEhDi2FUdoaKjjeLB5Mjt27EhiYiKJiYn07NmTr776ipCQEB5//HGWLFniOHbQoEEOL+mZM2cYPXo0Xbt2ZeLEibz//vusXbsWKOy93b59O+PGjXN8//777xkyZAi9e/dm6tSpXLt2zSXddOzYka+++oq+ffvSu3dvPvvsM0RRLHRcae2XxPz58zl+/DgjR4502n7x4kXOnTvHK6+8gkqlokuXLnTo0IH9+/eXWue2bduwWq0sW7aMFi1aoFAoCAoKYu7cudSrV4/169cDsHbtWh5++GG6detG586dee655zh37hxgmxV499136d27NwMGDGDFihWOGYI5c+awfPlyR3sFz1NiYiIdO3Z05Lbs2LGjo86MjAx69erlVNZVdDodL7zwAmq1mg0bNnDXXXfx1ltvMXPmTLZt28aiRYsA23h7/fXXmT59Ol27dmX48OFERkYC8Nprr7FhwwZHnXq9ni5dunDt2jWnc7hw4ULH+bbrplOnTnTr1o3evXsX6nNB1q5dy/Tp08vcv6qG7AXKp6rr4u461/9QCqBUgEph8x5LElYRh1EsSBK+IvhJIHh50XvUeQImplD/9WRCRtl+o4T/H7R6HXyegw5TYc9JqPciaJ6h/9ObeGLo19D7A6g7BpNiOKsf+gVPQG2xoLKKeOaa8M42ojB5sOT+X1j1wF5WPbCXL+//HovmWQgcA//7B9Ky4Kn5cM+b8Pl+vvPfxDc1tyIFeaBQKfAySqitEqKHCnMdD+psjeJo7VVoMFGTdLwxYaQOqdyFGQ9UWKlDGvUwcF6YSqrwNleEd/lHmM8/wnwuCu8jCUNBGArKYdBiArSZCCOXQU7ZM21U9TFxO5F1UfG4c4yx2xnGOp2Ot99+m6effpqwsDAmTpzItGnTyMrKAmDPnj1Mnz6dgwcP0rBhQ1avXg3A5s2bAdi/fz9t2rS5JRkMBgOJiYmEhoaydOlSvv/+e/7880+nY0wmE2+//TY9e/YkLCyMwYMHs2/fPpfqP3ToEBs3bmTx4sXs2bOH4OBgZs2a5bJ8x44dY8eOHXz99dfs37+fH3/8sUz9K43x48ezbt06Gjdu7LQ9Pj6e+vXr4+mZH6vXpEkT4uPjXZK5d+/eRS7fOnDgQMLCwhzfe/TogVar5ejRo7Rp04Z169YBMG3aNARB4Oeff2bjxo1ERUU5HkRuljVr1tz0TMPOnTvx9/dn+vTpjtRASqWStm3b8t577/H999+TlpYGQGRkJPfeey9hYWGMGjWKadOmkZ6eTr9+/Th06JCjzkOHDnH//fc75eGMj48nIiKC/fv385///AetVss777xD69at0Wq1HD58+BY04D7IXqB8qrouErOu/1FwxkRpewnPx5iflk0SBATRigIQBYEjtQPQq1Uke2g47+vD1zuuwbCFcC7ZZiyeugBDF0GKHsxWVHlWlBYRjvyFlJqFtmkXTGovx23a/r9CkvAwW9EIAkpRRCmKGAUv/gpsYzOIn5wP72yBXZHw92V4fR0oVeCpQGkW8dOZ8cq2UiPdgjrXCgqBXI0H3tZ0/MhGSb5zQkRNAm2w4omANxI+1CSHPPwx4I8FDyx4YMabf7jveiEJLlyFs4nwjRYW/lRmnVf1MXE7kXUhUxJuZxgfO3aM4OBgBg8ejFKppHv37qxZs8aRxL179+7cddddeHp60rdvXy5fvlwhcowZMwaNRsO9995L06ZN+ffff532nz59GpPJxNixY1GpVPTt25eOHTu6VPeuXbt4/vnnadGiBR4eHkycOJGYmBguXbrkUvmJEycSEBBAw4YNGTFiBAcOHChz/0qiTp06RW43Go1ORjGAp6enS4alTqejdu3aRe4LDAx0GJA3kpOTQ+3atUlISODPP/9k6tSp+Pj4ULduXSZMmMDu3btLbbs44uLiiIyMpEePHjdVPikpiebNmxcZLtGiRQtEUeTKlSsANG7cmFGjRqFSqRg0aBDBwcFotVr69OnDpUuXHA8X+/fvp3///o56JEliw4YNDBs2zC1eKj116hRJSUnEx8cTExNDSkoKMTExJCcnExkZiSiKHDx4ELB5dbKysggLC0MURbRaLQaDgZiYGOLj4zl//rzj/9jYWA4fPkxERARmsxmtVktWVpbDM2SftYiIiCA1NZXo6Gh0Oh2nT58mISGBhIQETp8+jU6nIzo6mtTUVMe7Avaydnm0Wi1ms5mIiAj0ej2xsbFOstj7ZjAY0Gq1iKJIWFiYkzwHDx5EFEUiIyNJTk520kV8fDxJSUmcOnUKvV5PVFQUOp3O4e2/UZ6IiAiMRiMnTpxAp9MRFxfH7t27SUhIIC4ujsuXLxMdHY3RaCQ8PNyprF2eI0eOYDabHSE/sbGxJCUlOfqWkpLCyZMnMRgMREZGotfrHQ+r9utLWFgYer2eyMhIDAYDJ0+eJCUlxXF+7PUlJiYSdzneNiCkAtGHokS9LCMd0jId22vmmWlgyEGQJBQWs5MhbREELicYETOcs06I5uKzV+SqPLAqlUXGPJpVKoSC8kiQq7JdzyxZ2VhSMwrsux76IYDK4uwPU+WJIElICgVq0YwCM5YCRxjxxIwXudgelq14osGE5IgrsVO8py3p//4p89jbtm0bUPFjLy4urkqPvaioKMLDw53aLyhPeHg4RqOR6OhoLl++7NSfuLg4dDodJ06cwGg0EhERUeR1RqvVotPpiIqKQq/XV+h1T6/X39J1r+IQbvi4D1X/TnoDOp2OevXqOW1r166d4++C8WQqlarIMILy4MZ2JMn5Upuenk5gYCAKRf6zR8OGDV2qOzk5mTVr1jjCBwAEQSApKYkmTZqUWr5Ro0aOv+vWretyGMat4unpWSiZfm5uLt7e3qWWrVWrFlevXi1yX1JSkpPRfPToUXr27InJZMJisbBmzRp0Oh1eXl4EBAQ4jqtfvz5paWlYLDeX5mnx4sW8+eabTt7qgnzyySfs3bvX0daN8eXBwcHs2bMHSZIKGcfnz59HEAQaNGjApUuXCo2NunXrkpaWhq+vL127duXAgQM89dRT/P333yxdutRx3JIlS2jatKnLMwqzZs1CqVTi4+PDI488wpQpU1wqV1488MADhbbVrVsXwBGH/vjjjwP5Xp2ePXsC0K1bN8D5916Qli1bOh6Q7cfa6wgJCQHy30a3P9wVTPJvPwf2bfZj7GXtddnrttfl7+9fbH/tx9r7YK/D3sdOnTo59d2uC7CNKYAHH3zQqa4b5bHL8fDDDzvkb9asmdOKb/Zrgv0hz172xv/tbTVo0MBJhoKy2WW296lv375O3+377Y6AovrkXdMCdnvWev0abZbwN5qpZbHSKcOASaEgwGwhycOD+qY8klQK7s008JefL15WkZY5ucx+uykwGqZ8ZatDrUQxIQRW7bUZp1y/Jft7I+hz6HHxGH/Xa0NODR9UpjwUkoQggVmtwqJWkpsLXtcv5YIK7k/+CwDVxy9A97Zw7Cxcy4RnuyD9YgGFErNaQBJAkGwvGVlVAqpsM/6mHP7xb06zrEtcpDG+5GBBSR4aGpKIEhEzIgqspOOLJ1koMJJJLQQk1OTgR5xDdygVNl01DqT+glfB379MY2/YsGFAxY+9G6lqY69BgwaYzWbH7+NGOexytm/fvlBf7Nj7a+//jdcZu77s+qjI615BOW7muidTGLfzGNepU4eUlBSnbRs2bHDZm+oKCoUCszl/Ou9msiI0aNCA1NRUJ6MsNTXV8bdSqSy2jcDAQKZNm0ZYWJjjs2XLFseFozQKtpOcnFzki28ltX+zNGvWjKSkJKfMH5cuXaJZs2allu3ZsycHDhwolDVEkiR++eUXunfv7tjWvXt3wsLCiIiIYPLkyXz44YcEBQVhNBrJyMhwHHflyhX8/f1vypN66NAh1Gq14+JUFLNmzUKr1aLVaot86fKpp57CYDCwaNEix5vxVquVv//+m48//pghQ4Y4QiIKnjOwPQzYHwD79+/PkSNHOHLkCF27dsXX19dx3KRJk7j33nuZO3euS/365JNPCAsL45tvviEpKcnp4cvduTGc6U6mquvC126zC9dzGAsCaBRc9vUkTxDwtYrUNltQAv/U8EZCIlmjJgeJphkZPOmVyaGvmtrqeHsw6DZB/OeQvhlWvAK6ryFmOXu2jeSX+Y9DxhbI3EqNLeNpc+U3yMvD5OOFsYYPOX4+mD01kJlN1yl3o2npiTXQk8EbOuF9egGkfQ1vDoQHmkPCerj6JWybwvCsUdTp6IOYayGnhgqzWsCiFiDPgsZq4awlkBZplzDgiYRAFr4Y8cKbHALIQACUGDFgJuC3iZgECdCjIAGJdKThDah9cAacXgpnVkH2N5D8JZz7LzQv+WXmoqjqY+J2Iuui4pFjjG8jXbp0ISkpiT179mC1Wjl69Chbt24t8ckZQKOxpfpxJdl7kyZN+Pfff7l48SImk4mvvvqqzNkD2rZtS7169Vi3bh1ms5nff//dKY1b48aN+e233zCZTCQkJDg8jwADBgxg69atXL58GVEU2bZtG2PHjsVoNLrU9tq1a8nOzubSpUts27aNJ554otAxJbV/szRr1oxmzZqxZs0a8vLyOH78OFFRUfTp06fUssOHD8fX15cpU6Zw4cIFR5jBe++9R2pqKq+++mqhMoIg4O/vT05ODnXr1qVTp04sXryYnJwcUlJSWLt2Lf369bupvnz77bdMnTr1psraCQgIYPPmzVgsFl588UXi4uJYtmwZn376KUOGDGHmzJmOY//55x92796NxWJh586dXLt2ja5duwLQtWtXUlJS2LVrV6H+aDQapk2bxl9//eUUi1waSqUSpVJZZHopd6Vly5aVLUKVoarrom3dAl/UCltYglLA6KMmvIY3JgBJQmm1orZYMCBhleDchiZc3Nicbxc0d66wpi80qQv2XMQBvnBPY0RvDXl3XTcia3jBs90YfOEdrAoBRZ7Z5oG1WlEbc5l5egCPjGrOqz/0ZtLh3jRsXwfubgS1auS346GGugGOr4/veZLnM0aRp7GQ5QVZCng6azRPZY5liOU12lnm4EEurTiNFwb8yKARtulrCchBoqG0khqPtKKhuIgG0mfcJS2njfQJzbePhz7t4b6m0DoYPDRQLwDUNzfRW9XHxO1E1oVMSbhdKEVAQADLly9n6dKlLFiwgODgYBYvXuw0hV4UtWvXpkuXLgwZMoTly5eXGO/brl07hg8fzmuvvQbYMlqUNVWYUqlk6dKlzJs3j759+9K2bVvuu+8+x/6xY8cyd+5c+vbtS8OGDRkwYIAjE8GAAQPIzMxk0qRJ6HQ6mjRpwooVK1yWoUGDBgwfPhyLxcLzzz/PwIEDCx1TUvu3wqJFi5g3bx6PP/44tWvXZt68eaWmagObgbdu3To2bNjAW2+9RVpaGv7+/nTv3p3Nmzc7nd/w8HC6deuGKIoEBgY6Mip89NFHLF68mMGDBwM2T+vEifl5Qrdt28YPP/wA2EI8Zs6ciUKhcITBzJw5k5UrVwLw5JNPuuTpLo3atWs7whxGjRrF8OHDGTRoUKHjmjZtyrFjx1i8eDGNGzd2Ot9qtZrevXtz+PBhunTpUqisl5cXb731FsuXL6dr166F4rwLYjfG1Wo1HTp04KWXXirX3NaVSWJiYqkPyHcKVV0XbzwI753AFqvr5HQQyAn0xffiVYJMZgSgrslEnihR11j2TAxAkbnrZ/zWhwWPHEaBBcFiway+uXArO88ljy52X4DpS3I9XuI+jmNFRQ5BmFFhxkqtc3Nuqd2yUNXHxO1E1kXF4265iwsiSDcGx8pUGNOnT6dFixaMHz++wtro2LEj27Ztk5+Iy8jJkydZt26dI8PF7SQ0NJQdO3Y4MqcUxYYNG7h69WqZspPcaSQlJTnFJd7JuIMuHEtCFzSMJYnARB09ddn4WvPfD/nD1weVxULUF42KqKl4QkNDMZlMPP300+Ul9k2jG/cFmn5tyRr5LaKPF0GJn6DU3D7flDuMiduFrIuK56Swxul7R2lCJUlSdtzOYywjUxGoVCqXXhK83aSnp3PlyhV+/PFHFi5cWNniVGlufPHzTsYddJE9CXxW4GwcSxI5Nby4ZjDjY8xFAK54eqBTK/C9yci/quL7qbXuFQB8jZWzHLE7jInbhayLisfd4ooLckcaxjNmzCh22eaisgtUFcLCwnjvvfeK3W8PA7gVli1bVmLe47Is8VyQkJAQxwIaN9K/f/9K94S2b9/+phbxqGhOnTrF+++/zzPPPMM999xT2eJUaapTvPSt4g668NaosE6VUM7LA8315aDzJLI/9KPbawYSvL1QSBJWhQKFBL6Wm8swVFUM48rGHcbE7ULWhUxJyKEUMjIy1YKUlBSn9Ex3Mu6ui67jE/G4/rdFgAg/bx65mob26xYllruR0NBQ8vLyHKnK7mTcfUyUJ7IuKp5I4XOn752k1ypJkrLjdlkpZGRkZIrixjSOdzLuroscpUSmAJkKgT89PWiSa2bPZ8E3VVfBtJR3Mu4+JsoTWRcyJSEbxjIyMtWCpk2bVrYIVQZ310WaVULvo0GpgLpYGdo2lxq+xWdbKYmSsrTcSbj7mChPZF1UPCKC08eduCNjjGVkZKoff//9t2Plqzsdd9fFpbWurRLqCq7krr8TcPcxUZ7IupApCdkwlpGRqRaUlJv8TkPWRT41atQo/aA7AHlM5CProuJx56wUciiFjIxMteDo0aOVLUKVQdZFPje73H2thTkIH2fy7A5TtchsIY+JfGRdyJSEnJVCRkZGRqZaEhoaClDkapPFcS7JRKtNgm2ZajuSxOlRAvfVV5e3iDIy1ZIIwXmxrM7SuEqSpOzIHmMZGZlqwZEjRypbhCqDrIt8MjIyXD5WFCVafSnajGKhwEeh4P6N7p37Vh4T+ci6kCkJ2TCWkZGpFvTq1auyRagyyLrIJyAgwKXjJElCOTcbNCrnZartKN03ZhLkMVEQWRcVj4Tg9HEnZMNYRkamWhAWFlbZIlQZZF3k46rHeMdxA3ipizaKrZLt48bIYyIfWRcyJSFnpZCRkakWdO/evbJFqDLIusjH39/fpeMuZ5JvFIsSilwzklKBpFHaXEgW9zaM5TGRj6yLisfdvMQFkT3GMjIy1YKTJ09WtghVBlkX+WRlZbl0nKdaBAEQRVQpBlQ6I+rUbBTZZpvB7OaGsTwm8pF1IVMSsmFchbly5UpliyAj4za0bdu2skWoMsi6yMfHx8el464ZRBAEBIuEYBWxeGsw+3ggmCwgSYCVpAxTxQpbgchjIh9ZFxWPdMPHnZAN4yrK9u3bWblyZbnU9f333zNw4EB69OjB66+/TlJSkkvlOnbsSJcuXQqtHGWxWOjdu3eZUiDJyFQ08fHxlS1ClUHWRT65ubkuHRfga4sslFQCFm8PRA81kihhNYuQngueGhp/auR/F9xzJT15TOQj60KmJOQY4ypKRkZGuSSVP3r0KBs2bGD16tU0adKEJUuW8PHHH/PZZ5+5VN7T05Pw8HCeeOIJx7aIiAgsFsstyyYjU57UrVu34iq3WuEbrS1jwfAu+bGo+mykbRHQqj5CnRqQY4KH78ovZzLDV0fg0jVbWe3fSNHxtn0NayK0CAI/L1Ao4f3h0LQuRJyB7ceQAv3gwbswBdXDnA2apr4YDl/C/HcK6sY1yDx5Dp+/4lA0r4tP79Z4nv4HqVYNcgUlZv9ADD+exXd/FFLCNSSlgtxLGQiigIQKBUZUiI4oQAkBMyrUmAGbxyT/6qNAQoEZDQqsWPElkxroqIkFD9TkosRKLr7oPAKwSEoUFispdXyxqJTkqZUolJAe4E2mbw1yfLxRSCK5SiVmpRJJoUASBJI9NYgKJWaFAg+LmeYpiSQFBJDm44dVqUQQJSTRCioVSqsZi1JNnqBAKYAoAYKAKACShCBJIEHrK/W4K/Uq0776nWsBPhgkFcnePlg0KkwKAQsCVqWAQSGgsUgQpABvDyQPFeRZbR8AkxWyzFg8NHT53Ixk0WHxVqKQBJv8FtGmMUEAlc3fVEMp0jwQmngqaN9Awdz+3qUOs9QMK5dTrbRppMLbs3i/1bUckYn7LKiUMPIeBffUUaJSwrl06FAPfDWF4ztd/X1EX7FiESU6Nqq+5kGFXitkAPeOMZYX+KgCJCYm8txzz/HYY48RFhbG1KlT+fjjj5EkidatW7Np0yZiY2NZtmwZZ8+eJTAwkLFjx7rksX3jjTfo3r07zzzzDAA5OTkkJibSsmXLUst27NiRoUOHcvXqVVasWOHY/s477+Dh4UFUVBShoaHk5uYyd+5cIiIiEEURSZLIzc3l6NGjeHt7c/jwYdauXUtSUhLNmjXjnXfe4e6772bcuHH07t2bZ599FrB5yQ8fPsy6devIzc1l+fLl/P7771y7do06derw5ptv0rNnz1LbK8icOXPYt28farUtMf/ixYvJyMhg8+bNJCYmAtCnTx/eeecdBEEgOTmZBQsW8Mcff+Dt7c3IkSMZNmwYffv2BWzeJ41Gg0KhYPz48fj7+7N37158fX2JiIigQYMGTJ06lU6dOgHw+++/89lnn/Hvv/8SHBzM66+/TteuXR363bZtm+NcTJ8+nRYtWjB+/HjS0tJYunQpv//+O56envTt25cJEyag0WiYM2cOu3fv5qOPPqJ///4AaLVa3nrrLV544QUmT57MnDlzCAgIYPLkyQC8//77HD9+nMOHD6PT6ZgzZw7R0dFIkoQoinh6enL48OFSx0RV5vz587Ro0aL8K5YkuHcy/N9l2/dHW0PEp3A1A6nVfxCyjEiQfxt4uTd88R/IM0PT1yBJj83UFJGuG6P2453KAUzsD6v3Or5aUXGarljRICkEclBQEyN/tGpMjqcHKquZJy/soV5OilN9ZkGBWhId9YiokVCgxBYKUKhdV9RQ4O8/6Uw2foCEL7mIKAHQazzJVHtSx5iFQpC4WK8mBi8NaqsVQbS1mFqnJpcaB2Hw9ABBwCsvD+88MxaFQGztmkiCAiQJf2M2em9fm/4FAQkQwWZ8ShJWQchPIHG9QwX7JQKeubls3PQV13x9WRAygEu1a5KnUNjqE0Ws1/MUmySJfzzUmPw8oIaHLQOFwQxGc36nvVXg7wlWCcFoRvK5nsHCaLEZ0NflwleDw1pXCAiiRP2rWXjVUHBuTkCx+v2/eDMTVqSTY5JoUk/JV9NqUcO7sHGcZBAJXma2RXdIEigFVGoBtYcCowXa1ILfnlcS4Ol8hl35fXywP5cPD9rGyKSuGlY85VXi8e5KhV0rZBwcFTY4fe8uvVxJkpQdOZSiipCdnU39+vU5cOAAffr04cUXX6RHjx5s2rSJ9PR0JkyYQK9evTh8+DBz585l+fLlRERElFrvmTNnsFgsjB49mj59+vDBBx+4nNcToG/fvpw8edKR8ig7O5tTp07RrVs3xzG7d+/mwoUL7N69G61Wy44dOxz7zp07x/vvv8/kyZMJDw9n4MCBTJ06Fau15GT5W7Zs4eLFi2zZsoXw8HAGDRrEokWLSm2vKEaMGIFWq0Wr1dKoUSM+/vhjZs6cyZEjR/jiiy/Yv38///vf/wCYOXMmgYGB7N+/n3Xr1rF582ZOnz7tKB8UFMT8+fPRarW88MILAERGRnLvvfcSFhbGqFGjmDZtGunp6Zw/f54pU6bw4osvcvjwYV5//XVmzpzJuXPnStX7tGnTEASBn3/+mY0bNxIVFcXatWsd+4ODgwkPD3d8379/f7Fv3//1119ORu+WLVuwWq0cOHAArVbr9NDjziiVyoqp+FJqvlEM8NtZm6d45wmELCNwg5G54TAYTfDHBUhKJ/8ym+9DufF/p7IFUGLBF9uSxoIo4SeayPD1JsfTAwCLUk1srbsK1VfQKAYQ8URBfnzszfpyBMCC+rpRbK87X+/eljwQBHJVajytFpql6GyeXin/VlP7WgY5HhqH192o0dj6olDajGIAQbAZxdf/huuGud0SFgTnPghO/zn+zvH0JMnfn7oGA/G1aiLZF+wArNcNZICrKqXNqFYK+V7fGmpQX5dHKYCnymbsKgUkrwK5jr1U+avkCQKYr1/brm+TFAIGXw3GdAuWElK+hf5mJMdk23/pqpUTZ/KKPG71SRGHO0sQwCphQcB4fRLvjA4OXSrcjiu/j9XH89v8LCIPUayefrMKu1bIOJDzGMuUC/3790ej0eDp6em0PTw8nHr16jFixAhUKhXt2rVjyJAhjuVOSyIzM5Mff/yRjz76iNDQUDw9PXn//fddlikgIIAHH3zQsVLQr7/+SteuXdFcv5mB7eWWvLy8QrHIAIcPH+bRRx+lc+fOKBQKnn76aT799NNSw0SeeeYZFixYgJeXF1evXsXb25uUlJRS2yuNOnXqsH37dtq1a0dGRgaZmZnUqFGDlJQUrly5QkxMDG+++Saenp40atSINWvW0Lp16xLrbNy4MaNGjUKlUjFo0CCCg4PRarUcOHCATp060atXL1QqFV27dqV79+7s3bu3xPoSEhL4888/mTp1Kj4+PtStW5cJEyawe/duxzGPPPIIf/zxB2azmdzcXP744w+Hl7ogkiSxaNEinn/+ecc2Hx8fcnNzXY69LE9OnTpFUlIS8fHxxMTEkJKSQkxMDMnJyURGRiKKIgcPHgRsq1NlZWURFhaGKIpotVoMBgMxMTHEx8dz/vx5x/+xsbGYzWYiIiIwm81otVqysrIc43b//v2ALQwoNTWV6OhodDodp0+fJiEhgYSEBE6fPo1OpyM6OprU1FTHg+eh6EhQ5d9IrRolZg8V/5eXXmQfLfUDiDkXR3ZNr+te1uLH+o178vy9Cu034eX42ywo8cozQ4Hfj78ps9j67AhYoYABeyvmjgozXhgKbMk3wvMUtul3jWgzDpWShFKUMKvzp+VzPTUoxfwygijajFJJcuoXJV0jbtxX3KES1MyxPbz45+aiKHCcUKAOu3RSnphft1IBgd5Q1xvqeIM9vOJ6mIaTLAW/243kAvWrzSIWAdJ114ode7V9Czy4CBLBgUrHuLX/FrRaLQ/WE537Lzi3pUAiSGNEq9UiiiJhYWFkZWXxf//3fwAcPHgQURSJjIwkOTnZ6XfYqEa+h7xpAJw69Qc6nQ6tVgtQSJ6IiAiMRiMnTpxAp9MRFxdHXFwcCQkJxMXFcfnyZaKjozEajY4HeXtZ+2/zyJEjmM1moqKiSExMJDY2lqSkJGJjYzl//jwpKSmcPHkSg8FAZGQker3ekYf4wIEDgC0vsV6vJzIyEoPBwMmTJ0lJSXFcG+z1JSYmEhUVhVKpdGq/oDzh4eEYjUaio6O5fPmyU3/i4uLQ6XScOHECo9FIREREkdcZrVaLTqcjKioKvV5fodc9vV5/S9c9mcJU3yAiN6R27dpFbk9PT6d+/fpO24KCgjh16lSpdarVaoYPH06TJk0AeP311xk8eDDZ2dkuv60dEhLCzz//zNChQ9m7dy8vvfQSOTk5jv19+/bl33//5fnnnyc7O9vJsE9LS3OK51IoFNx3332O7ytXrmTNmjUAmM1m7rnnHgAMBgMLFiwgJiaGhg0bEhwc7DCmS2qvNJRKJT/99BM///wzXl5etGnTBovFgiRJ6HQ6vL298fX1dRzfvHnzUuts2LCh0/e6deuSlpaGTqcjKCjIaV/9+vUdBj7Ayy+/jHDd85Sbm0uLFi3Q6XR4eXk5efbr169PWlqaI7bby8uLe+65h//9739kZ2cXaRQDhIaGUqdOHR5++GF+/PFHAEaOHIlOp2PAgAGIoohKpXJ60KlIHnjggULb7OPDrqvHH38cyF+dqmfPngCOWYp27doVWXdsbCydO3d2OtZeR0hICIBjf506dQCoVauWo7z9PNq32Y/p89RACA2GCWtBqUD59SSUnh60e2MEUrYHfLYPatWANvVtzsb3h9OubSNbpV+/AVO+Bn0uSCBZLNgsGQEKxPiiEOCJDmi2ToZec+B0PJJSidiiEQH1GpDrXwtlbQ+u/RKPKiOHe68mkuEFdYwptEz7B6OXD555NgMwtUYtLtZsQJurF/HLyboeXpCDBU8UKLGiRETCg7I/GEkIiCgJ5iLJNMabPHzIIQN/zIKGPKUGL3MuPpZcRARimzVAlGzhBLpafkiCQErdWtTUZ5FZwwdREPA05ZGrUqLz9EAtilgUCrzy8uhxNoY/mjQn2S/A1rg9VAHyHzikfG+y/eMIpZCg/aV4YuvW40ydINr+m8zfjYPAy3a9yJMkzColIFBTFMlUCoiShEUoYNhaJJsxrBDyPcNGC+SYodb1hxij+XqQs2T7X1RAngVMVhRqBV4mC+o8C5+/UoM6dTyKHXvt7pVAmc3Zyxb6PujJ3Y3V3N3YNm7t49g+rl9on8fWP20PF80CBZ5prUCjEfjrGoxoo6BrMx9oZjvW/vuxX8vsvy/7NcP+u6tbty6hDUVm783FLMLcvh60qvOgU7v235FdHvvv6eGHHy7UJzuNGtl+Cz169HAqe+P/Dz5oa6tBgwYATvc8+zXCLrO9T/YQN/t3+/6OHTs6lStYX4MGDYiNjS1WDruc7du3L9QXO/b+2vt/43XGri+7PiryuldQjpu57lUU7jzXIBvGVQi7gXQjQUFB7Nu3z2lbYmKiS4O7SZMm5OXlT4/ZQxjKElr+2GOPsWDBAmJjY0lISKBDhw4cO3bMsV+pVBIcHIxareb7779HqVQyePBgwPbjP3v2rONYSZJYuXIlo0ePBmDSpEmFYowBPvnkE5o3b87SpUtRqVT88ccfjifqktorjQMHDnDw4EG2bt1KYGAgAE8++SRgu2jk5ORgMBgcxvG+ffvw8/NzXFiKIjU11el7UlISffv2RRRF/vrrL6d9iYmJThfrDRs2OMUYg+18G41GMjIyHMbxlStX8Pf3R6XK/8n27NmT8PBwdDodQ4cOdXgH7GRnZ7N+/XrWrFlDcnKyY7unpyc1a9YkKCiI9evXc+HCBWbMmFG68qo49htqhdDvAbj4eaHNwswhMHNI8eVGP2b72I93pa2TixzHKoFGBXaV/pgGdQCNXo/f9dAauxlunx4syySyZf1eLP/qsJ75F9PFTHzffhxF/weoW7MmBV9fqg8k7PqHxM1RZB/I4LLoiaaGB6LBSI63F4JSILWGJ5JGgyIrC0mlRhIEsjVqnptcg87dmqBS3SiZK70tmdDQKHItgbw1pPT3MYavucZ3WV6OmGYEAdQCqATIMoPSCmYRDEauvetJ7QD19ZLqEut1FaVC4NUnfEs/ENj8pIbNT5atfld+H40CFHz9XOkvCbo7FXqtkHF75FCKKopGo3GECnTp0oW0tDS2bduGxWIhJiaGnTt3Ol68KolBgwaxbds2Ll26RG5uLp9//jmPPvqok1e0NLy9venSpQsffPABffv2LWTAX7t2jUWLFvHGG284vAN2+vTpw++//+6YLvr+++85dOgQfn5+lER2djYeHh4olUqSk5P5/HObUWKxWEpsrzQMBoPDQ5qXl8fXX3/NlStXsFgsBAUF8cADD7B69WpMJhP//vsvy5cvLzUe7Z9//mH37t1YLBZ27tzJtWvX6Nq1qyM++8iRI1itVo4fP87Ro0cdXo7iqFu3Lp06dWLx4sXk5OSQkpLC2rVr6devn9NxPXr04OjRo8TGxhbpMd6zZw/9+/cv5NGOi4tj/fr1vPvuu7fFc3C7cCV2+06hvHSherU/nh+NxOe7d6h18lM0z/dCVbNmkcc2fLIV/b4fwbDM13gxbzwj08YyOmkk/zk/lNfjhvDx0R7MO/QoHx9/jHnhXfnv3o5s2nU/3R9rXoRRXH4YjUaXjmsWCA6vb44ZMk1gstgMZA8F5Fogz4y0qBa1A9zPeJR/H/nIuqh45BhjmXKnW7dunD9/nqFDh+Ln58eqVas4dOgQvXv35t1332XixImO6ZKSePbZZ3n22WeZNGkSISEhZGdn88EHH5RZnv79+3PhwoUijfEVK1bQrFkzBgwYUGhf06ZN+eSTT1i6dCmPPfYY+/btY9myZaUam2+//TZarZYePXowfvx4unTpgpeXFxcvXiyxvdIYOHAgLVq0YNCgQQwcOJDY2Fgee+wxLl68CNg81deuXeOJJ55gwoQJvPLKK45ps+Jo2rQpx44do0+fPvz444+sWLECPz8/GjVqxOLFi9mwYQOPPfYYK1as4OOPP3aEi5TERx99hNVqZfDgwTz//PPcf//9TJo0yemYgIAAGjVqRLdu3YrUZ0BAAC+99FKh7YsWLaJXr1506NChVDnciYIhOnc6si7ycTVkzFuhtBnF2RabUZxtBl2uzSC2SOAhIM0r+YG+KiOPiXxkXciUhJyuTUbmFggNDWXHjh1s3ry5skW549FqtU7ZUu5kZF3YCA0NRa/XOzLIlMRLG66xMcUPsvNwpHgA8FSCRgkKkN5zP0+xHXlM5CProuI5Imx0+t5LerGSJCk7ssdYRkamWiDf6PKRdZFPcWkMb+TBxhqQrI4FOhxoVNeDtN37dimPiXxkXciUhPzynZsTEhLilCGiIP3792fWrFlF7tu2bVuJq9999913hTIqyMhUZY4cOeJSeNGdgKyLfNLTi06rdyMv9/Bh4gkTeClB6QF5oi2PsZcK9Cbwde/ct/KYyEfWRcXjbnHFBZFDKWRkZGRkqiX2XO+urBIK0GlNDv9LF0Cjzs8NbLTYYo+9VUjTb09aQxkZd+ew8JXT997S2EqR42Zw77khGRkZmevYk9rLyLooiKseY4DICd5gup6XWJJsXmOlwrYcdJ659AqqMPKYyEfWRcXjzlkpZI+xjIyMjEy1pKweYzvC3CxQqEBz3XdksSK9674v3snI3G4OCV87fe8jjakkScqO7DGWkZGpFtiXcJaRdVGQzMzM0g+6AemDGkjveSHN8LB9qoFRLI+JfGRdVDziDR93QjaMZWRkqgVFLbt6pyLrIh9X8xhXd+QxkY+sC5mSkA1jGRmZakHBpcfvdKqDLrovSsX3w2w0n+QSMDvjputxdeW76k51GBPlhayLikdSCE4fd0I2jGVkZKoFjRs3rmwRqgzurovZ+4xEWGpg9NRg1qjI8vPB/z39TdXl4eFRztK5J+4+JsoTWRcyJSEbxjIyMtWCa9euVbYIVQZ318Vnx/PwskqI1z1NokJAUN1c2n2z2b2zSZQX7j4myhNZFxWPJDh/3AnZMJaRkakWeHu7/wtS5YW768KsVuFrsjht87ZY8fo4h8DFuWWqS6l074U5ygt3HxPliawLmZKQV76TkZGRkalSNMw0kuzrRbcLqahECaNaSYKfJ3kaNWKu7AGWkanquFtccUFkw1hGRqZaUNzS6Hci7q6LBvpcNCL45VkB8LaI+JrMGNVK9F5lW33OarVWhIhuh7uPifJE1oVMSVRYKMXVq1exWCylHygjIyNTDgQGBla2CFUGd9eFRSkgCc4eJ4UEviYLCklEWGhwuS61Wl3e4rkl7j4myhNZFxWPpHD+uBMVIm5aWhrDhg0jLy/vluoZNGgQWq22nKSquDoBRFFk2rRpbN++3Wn7Tz/9xJAhQ+jRowejR4/m1KlTLtU3btw4Hn30Ubp16+b0efjhh1m+fHm5yy8j4+78+++/lS1ClcHddXGlhgcXanljVOXfotQSNMswkqdUAq57jU0mUwVI6H64+5goT2RdyJREhYRSmEwmcnPL9oKEO5OUlMT8+fM5fvw4HTt2dGw/efIkn332GZ999hmtWrViz549vPXWW+zcuZOAgIBS6508eTLPPvus07Y5c+aUs/QyMtWD1q1bV7YIVQZ310WASeSiWkVY8zr0PJ+Cl1UCwMdsRSGBKLgev+jl5VVRYhZiQ8sfnb7f/1ELOj53f6nlRIuF9NAT5A1di89f71HjnpYkKJ5HwpP6F+ahbtbglmVz9zFRnsi6qHgkZTWOMY6KimLFihVcvHiR+vXr8/bbb/PII48wZ84cfHx8OHv2LGfPnqVp06a8++67tGnThlGjRgEQEhLC+vXradOmTYltfP/992zevJmMjAzatm3Lu+++S8OGDQH43//+x2effcaVK1fo0KEDH330EX5+flitVr788kt+/vlncnNz6dq1K1OmTMHX17fUOu18+eWX7Nq1i7Vr1+Lj48OcOXM4deoU3t7ePPzww0yfPr3UHJhms5kXXniBp556iqysLKd9V69eZdSoUY4f4cCBA1m2bBkXLlygQ4cOpam+RCIiIpgxYwaSJJGbm+u4+H/++ed89913BAQEMHnyZADef/99jh8/zuHDhwkNDWXHjh1s3rwZgMTERAYPHszRo0fx9vbmzJkzrFy5kvPnz2MwGGjfvj0ffvghtWvXLiRDx44d2bZtGy1btiQjI4OhQ4cyePBgR7t2bmwjPj6eESNG8NZbb/Hss89y7NgxFi1ahF6vp3PnzsyZMweNRsOJEyeYNGkSGo3NO5SXl8dLL73E+PHjAds53rp1K5mZmTzwwAPMnDmTwMBAR3sFb4hGo5Ft27bh7e1doiy7du1i3bp1ZGRkoFAoMJlMTJkypdADysmTJ5kxYwaHDx8GbON0woQJLFu2jG7durFr1y4+//xz8vLy6NevH1OnTkW44WYeGhrKRx995Bhjoiji6enJ7t276du3LwC5ubloNBoUCgXjx4/H39/f6fzt2rWLjz76qMi+gW2GZPr06XTr1o05c+Y4jYvi+uKunDp1is6dO1e2GFUCd9TFulNmxh+QQCHgW7cG91zN5EzdGpxsVAuzUkG9rFzqGEy2FG6S5HK92dnZ5SKfk9FbUvv237kkcXr2OU7PPue03d+YRi2TEZUo0irrLDUt6aiw4Ikf/piw3PshIkYaIgImaD4RE15c5S4kVICIH4loMJFJPSwoqcUVPMlCREUmAdSulwdjesJ7z8AnP8DhP9GcjgelAqwSmAq8wOihBF9v6NwaziaCpxo2TYL7m9n2J+pgznYwGOHH36FgtpB6/tA6GNZNICvsKrlh/+IV0hzfsffdorYrFnf8fcjcPkoMpdDpdLz99ts8/fTThIWFMXHiRKZNm+YwAPfs2cP06dM5ePAgDRs2ZPXq1QCOm/b+/ftLNYojIiL473//y7x58zhy5Ah3330377//vmP/n3/+yZo1a9i9ezeJiYl89913AGzdupVff/2V9evXs3PnTnJzc1m0aJFLdQJs27bNYRQHBQWxZcsWFAoF+/fv55tvviE2Npa9e/eWqkClUsn27dt54403UN2QZ3PAgAGMGTPG8T06Oprs7GyaN29ear2l0blzZ7RaLTt27ABsutZqtdxzzz1Ox/31119lMnhmzpxJ9+7d2bdvH7/88gsGg8HRRkmsWbPG5VmCpUuXolDYhl5ubi6zZ89mypQp7Nu3j8uXL7Nr1y7ANvMwePBgtFotWq2WHj16OOo4dOgQGzduZPHixezZs4fg4GBmzZrl1I5dJyWFzhSUJScnh08++YQPPviA48ePo9Vqad++fan9sVqtLFmyxBHLmJyczMKFC1m6dCk//fQTv/76K8ePHy+ybOvWrR0yrlixArB5uOzbgoKCmD9/PlqtlhdeeMGprMFgYP369aXKd6cg3+jycTdd/H1NYsJeq81wEwQMnmou1vahZVo2mZ5qjGol8bV8+KteDQA8zFae2+naOyx+fn63LN+NnmAEofhPccdcR+9Vm4sBDfmnVmO09TrjQR5mfPAhAxVmPMlGiYgAjo8HRjzJvv5NSSaNkFBQl/NIeHCN5oCAmjxqcg2uZsDCndB3Lnz6I0SeQ2myQE6es1EMYLJCWhaEnoS4RPjzEnR9FyzXX1ocsgDWH4RvjzkbxQBX9XD0b8SH30H32l5ytv1N2ou7Me6/cMs6r0jc7ffhjogKwenjTpRoGB87dozg4GAGDx6MUqmke/furFmzxmEAdO/enbvuugtPT0/69u3L5cuXyyzA/v37GTBgAO3atUOpVPLqq68ydepUx/5nnnmGmjVr4u/vT6dOnUhMTARsnrJXX32VoKAgfHx8mDRpEnv37sVkMpVa5y+//MKyZctYtWoVQUFBAGg0Gs6cOcP+/fsxm81s2bKFp556qnQFKhQuBfJfuHCBGTNm8Nprr7kURlEeSJLEokWLeP75510us3r1aoYPH05ubi4pKSkEBASQkpJSYpm4uDgiIyOdDNfi0Gq1GAwG2rVrB4Cnpyd79uyhe/fuiKKIxWJBuu6NycjIKHYadNeuXTz//PO0aNECDw8PJk6cSExMDJcuXXK5rzfKolQq8fT0RK/Xl+lN9u+//54WLVo4xkFQUBAHDx7k7rvvRhTFCnsr/osvvqB///4VUndFcerUKZKSkoiPjycmJoaUlBRiYmJITk4mMjISURQ5ePAgAEeOHCErK4uwsDBEUXScr5iYGOLj4zl//rzj/9jYWPbt20dERARmsxmtVktWVhZHjhwBbNcZsD00p6amEh0djU6n4/Tp0yQkJJCQkMDp06fR6XRER0eTmppKRESEU1m7PFqtFrPZTEREBHq9ntjYWCdZ7H0zGAxotVpEUSQsLMxJnoMHDyKKIpGRkSQnJzvpIj4+nqSkJE6dOoVerycqKgqdTud4wLtRnoiICIxGIydOnECn0xEXF8fOnTtJSEggLi6Oy5cvEx0djdFoJDw83KmsXZ4jR45gNpuJiooiMTGR2NhYkpKSHH1LSUnh5MmTGAwGIiMj0ev1hIWFAXDgwAEAwsLC0Ov1REZGYjAYOHnyJCkpKY7zY68vMTGRqKgozGazo/0fj55GVDnfjrLVSiw33FBzPWz3njyVgp3nTBiNRiIiIoo811qtFpVKRWpqKnq9/pbGXkVhUimuG7+ueMCdjzFRAyUWlFgAJVZsuhHIl1e6WPK1u1gMuWReTiYqKgrp79Lv6wq9wand1N/OVemxd+jQIaf2C8oTHh6O0WgkOjqay5cvExcX5/gtxcXFodPpOHHiRKljT6fTERUVdctjr7Trnl6vv6XrnkxhSgyl0Ol01KtXz2mb3YgAqFmzZn5FKtVNXUB0Oh2tWrVyfPfy8qJt27aO7wWf9tVqtWNaLDk5mQ8++IC5c+c6yZCcnFxqnX/99ReNGjXiwIEDvPLKKwCMHTsWsHm7P/zwQ9q3b8/s2bPLZenI33//nXfeeYeRI0c62rkdhIaGUqdOHR5++GF+/PHH0gsAMTExTJo0iZycHFq2bElmZqbTeS6KxYsX8+abbzouVsVhNptZvnw58+bNY+nSpY7t3t7eHDlyhHfeeQcPDw+6du0KwJUrVwqNPzvJycmsWbPGyWMqCAJJSUkunbOiZPHw8GDhwoWsXLmSWbNm4eXlhclkonfv3sXWk5GRwaZNm9i4cSMvv/yyU5+2bt3KypUrCQwM5P77S48zLAvx8fH8+uuvbNu2jY0bN7pcbseOHezcuRMPDw9at27NW2+9Va5ylcYDDzxQaFvdunUBHA+pjz/+OAC9evUCoGfPngB069YNcL4GFeTuu+92/G0/1l5HSEgIkO8pqlOnDgC1atVylLGHWtm32Y+xl7XXZa/bXpe/v3+x/bUfa++DvQ57Hzt16uTUd7suAOrXrw/Agw8+6FTXjfLY5Xj44Ycd8t91111OcjRq1AjA8fBqL3vj//a2GjRo4CRDQdnsMtv7ZA/7sX+377e/b1FUn+z129t9Y8D9fPppFjk+Xg7vag2TBY3ZikIUEa/P6iBJINjMyMinPfDyUjn6f+O57tatm+Ma6O/vf0tjD2+gAjJ8eVmsSIAHBsx4oMaEFSVK8h+mJSAXP4wEFCgp4YUeIzWwoEFDFmpsM3Z5eONBDigVCC/1hgU/gbWM9+ZHW+PXLJgHmwXD+L6w5OcSDxd7tUc45YOUnouirjf1x3ZC1dQmb1Ude3ZulMMuZ0mzhfbfWkljD/KvJRV53Ssox81c9yqKqpCJIjMz86ZmjEoUvU6dOoW8hRs2bCiTV640bmzDYDCwbNmyUpfxDAwMZMmSJYSFhREWFsahQ4f45ptvaNiwYal1TpkyhdmzZ7Nx40YuXrwIwPnz5xkwYADbt2/nl19+oVatWo7QjFvh559/ZsaMGcyYMcNhhN8OsrOzWb9+fZkMn6tXr/LBBx/w4Ycfsm/fPlavXk2LFi1KLHPo0CHUanX+DaQEvv32Wzp06OAUXiOKImlpafTq1Yvjx4/Ts2dPli1bBsDp06edjJ2CBAYGMm3aNMf5DwsLY8uWLY6LK1Aoprc0WQBatGhBeno6U6ZM4ejRo6UatJ9//jnDhg1zugBbLBbS09MZOXIk4eHhNG7cuNxDHpYuXcqECRPK/GLR8OHDCQsLY/fu3bRu3ZoPPvigXOWqTOxeEhn304W/h0DSOzUIMGSD1YpfVg5ZaiV/1/fPN4rBZjRfn1G6t6lr746np6ffsnwv/zkUdeCNt0sRRNEmT8GPHfv3Gx1GOXoa6pNplJFE2+S/yEKFBCgxYUKFEW/S8UdHEBkEYsAHPTVQko6KDFQY8CCFbPxIoSmeXEXNVbKoRTp1Ydwj8OlIOLsa5o2Ev5bDnndJGNAWpj4JUwfDI62gZRD0vAdmDLHFIv+5FD57Fb6eBMfm5cu7eCxEfAraeTBzCLRrBHX84KEWsOoV2DUTxcHZ1P+/cdTdN4IGf73qMIqrKu72+5ApGxcuXGDAgAEMGDCAq1ev0r9/f86fP+9y+RKvLF26dGHJkiXs2bOHkJAQjh8/ztatWxk2bFiJldpflsrOzi516cX+/fszc+ZMnnjiCVq2bMnGjRuJiYkpNffkwIEDWb9+Pc2aNSMgIID//ve/HD58mB9++KHUOtVqNe3bt+eJJ57go48+4osvvuCnn37iypUrfPLJJwQEBODh4VEoZrisREZGsmDBAlavXl3kE2NFsmfPHkaOHEnDhg1JTk52qYzRaARs4Q2SJBEREcHhw4fp3r17sWW+/fZbvvrqK5fq37lzJ9u2bXPadvbsWV577TXWrVtHy5YtHbMCycnJJCYmct99Rb/EMWDAADZt2kSHDh0IDg5mx44drFmzhtDQUAwGA4IglDiGipIFYOHChTRp0oQRI0aU2p/s7GwiIyML1RMeHs6yZctYt24dtWvXRq1WO8JDyoMLFy5w9913069fv5uuQxAElEpltVr8wO4lkXFPXfh5CKTPDQCgxfRsDN4gShIt07IJMOaRXMOTxBqeqK1WTGrXr82lzXi5yujfn3LpuNif/0Hlq6ZVr6ZlbsPj+udGapS5pgLc3RDubkjD/g+Wfuy9TYve/uj1LA5d74ZPRxV5iKq+L6r6vjcn423GHX8f7kZlrnz38ccfM2vWLBYtWkS9evV44YUXeP/999m6datL5Uu8ugQEBLB8+XKWLl3KggULCA4OZvHixaXGyNauXZsuXbowZMgQli9f7pTC7EYeeughJk2axKxZs0hLS+P+++9n3rx5xR5v58UXX8RsNjN27FiysrJo06YNy5YtQ6VSuVznG2+8wdNPP82OHTv4z3/+w7x583jyySexWCx06NCB2bNnlypHSXz99deYzWYmTZrktH3BggUVHvwfEBDASy+9VOS+s2fPOqZc7AbbmDFj+O6773jllVd47bXXEEWRpk2bMnToUP73v/8V286TTz5Js2bNXJJp7NixhcbO3Xffzfjx45k8eTI5OTm0atWK9957j6FDh2K1Wnnsscccx+bl5XH06FE6derEgAEDyMzMZNKkSeh0Opo0acKKFSs4evQo8+bNo2/fviU+2BQly4kTJwgPD3f5x2OxWHjzzTcdD4J2evXqxd9//82YMWMwm83cd999vPjiiy7V6Qpms9kpZv5G7NNnYHu5cebMmfzwww+A7aXTH374AUmSaN68Oe+++261WQVKq9U6xvWdjrvrwipAm6uZqIBGmbYQgXrZedybnMmhFoFgtQKuLdyh1+srTtAiuHtwq9IPqgTcfUyUJ7IuqjcZGRl06dLFMes/cuRIl5II2BGk8nRlybgliYmJjB8/ntDQ0MoWxUHHjh05efJkoe3jxo1j3LhxJT5syTgzaNAg1q5dWyi2rrphNpvlVc6u4+666PNKAha1EgBfs/OshrZJTfTv+bhUT2hoKKIo8uSTT5a7jO6Gu4+J8kTWRcWzs9Y3Tt+f0rmeBOBWGTp0KN9++y0jRozgp59+IjU1lZdeesllG6cKhEfLVDYKhQIfH9duNLcLez7qG/H29r7lEJc7DR8fH0dKuurMn3/+WdkiVBncXRdmlUBsHV/+CA7AfH1K1gqIgEGlLFNd5ZXH2N1x9zFRnsi6qN48//zzvPzyy6SlpbFkyRKeffZZnnvuOZfLV7jHeMaMGY60RzdSv379Mrm3K4OQkJBip5r79+9fKHeuK5w5c4ZXX3212P2zZs1yuzRcMjKVjV6vLzFDxJ2Eu+uizZRr/FO3BqJSicZixSfPitIiIipA5+OBNM01b19oaCgWi4UhQ4ZUsMRVH3cfE+WJrIuK56fAb52+D7nmumFaHkRGRhIeHo4oinTp0sWR7coVKtz1tmDBgopuokKx5wEsT9q0aVPiohMyMjJlJzExUb7ZXcfddaGWJCTBNsuRp1KSp1Lik2smWymQPK5sdeXl5VWAhO6Hu4+J8kTWRfWnU6dOjvR9ZUWek5aRkakW3K6Fc9wBd9eFRrSilCQs2MIoBFFCKUpIs8uWnhBsC/fIuP+YKE9kXVQ8YiUudvfAAw8Uma71jz/+cKm8bBjLyMhUC0wmU2WLUGVwd12cWFIXn3k5qDUaBCR88ixkGY3YVtooG/L75TbcfUyUJ7Iuqje7d+92/G02mzlw4ECZHpCr/xs5MjIydwTVKSfzreLuulApFZje9+Uuaxqq9CyeucuK+dPAm6pLNoxtuPuYKE9kXVQ8kkJw+txOgoODHZ+mTZsybtw49u3b53J52WMsIyNTLahR45aWQahWVBddxLxbv/SDSkEOpbBRXcZEeSDr4s7i/PnzpKWluXy8bBjLyMhUC1JSUpyW5r6TkXWRj9lsrmwRqgTymMhH1kXFI1WRGGNJkjCbzUybNs3l8rJhLCMjUy1o2rRpZYtQZaguuhAWFTZqXU3VZsfT07O8xHFrqsuYKA9kXVRvCsYYC4KAn59fsWsjFIVsGMvIyFQL/v7775tOz1PdqFa6KPh2uSQhLMxDmq4p/vgbkBf4sFGtxsQtIuuienLgwIES9/ft29elemTDWEZGplogLxOeT7XVxXUj2WyVUCtdm6uV40ltVNsxcRPIuqh4pCLSpVU0mzdvLnafIAiyYSwjI3NncfToUXr27FnZYlQJqrsurhgkmvq7duPV6/UVLI17UN3HRFmQdVE9KckwLguyYSwjI1MtkG90+VRLXUiSw2Pc7HML0gzXwinkxRxsVMsxcZPIuqh4KnOBj/j4eLZs2UJOTg6SJCGKIpcuXWLbtm0ulZfzGMvIyFQLjhw5UtkiVBmqpS7sU7OCAAoBYYFrSz1nZGRUnExuRLUcEzeJrIvqzZQpUzCbzZw6dYrg4GDOnTvHXXfd5XL5cjGMr1y5Uh7VFEtubm6ZctDJyMjcefTq1auyRagyVBtdWCSbp/hGrhvHrlAVPMYnfD4kS3iJNGECJ5TzKkWGajMmygFZFxVPZS7wkZ2dzdy5c+natSvdu3dn48aN/N///Z/L5W/ZMN6+fTsrV6681WpK5NVXX+Xvv/8ucl9OTg4dO3YkMTGxyP3Tp09n7dq1FSkeAKIoMm3aNLZv3+60fd++fTz55JN07dqVyZMnu2zgDxo0iC5dutCtWzenT6dOnQq1ISMjA2FhYZUtQpVB1kU+le0xPuL7PrUU1xA9RHLVAr4aI4e93r/tcshjIh9ZF9Ub+8NwkyZN+Oeff/Dz80MURZfL33KMcUZGRoUvuVnZF7bSSEpKYv78+Rw/ftzpbdd//vmHTz/9lM8++4yWLVuyaNEi5s6d6/KDxPz58+nWrZvTtnHjxpWr7DIy1YXu3btXtghVhuqgC2GRGcph0Tp/f3/+OJfKgytB0KhQCAJWswUkkZrecHlOID4epfuIzBYJsxW8PQQyjBL+njgWESiJeoKB5oZE/va9B72hAZJZSS0y0PqvIkMDXnlGDHV8aapP5s+ARnjmqTBLZkb+m3+tt5osGNNzUKSb8L67DqJVwmK0oPG15XTOybbi5a0oUZ7qMCbKC1kXFU9lLvDRpEkT5s2bx5AhQ3j33XfJyckhL8+10Csoo8c4MTGRHj16MGfOHHr27Mnu3bvZuHEj4eHhjB49GoDY2FjGjRtHjx49GDZsGKGhoS7VffbsWcaOHUvPnj0ZOnQoW7ZsAWDq1KkkJyczc+ZMR+D01q1b6devH7179+abb75xqufMmTOMHj3a4aHNyspy7LNaraxfv55Bgwbx+OOPM3fuXAwGA6IoMmDAAI4dO+Y49uTJk/Tr16/UNdXNZjMvvPACLVu25L777nPat3fvXnr06EG7du3w9PTkjTfe4LfffiuXsJAff/yRbt260aVLFzp27OjwKOt0OsaNG+fkVR43bhyjRo0CYO3atUyfPt2pn71793Z8j4yM5KWXXqJPnz706NGDGTNmkJubW6j9xMREOnbsSE5ODmALdn/kkUeK9Gbf2Mb//vc/OnbsiFarBWDXrl3079+f3r17s2jRIseD1s6dO3n00UedvOX28VTcubS3V1AnXbt2dchamixffvklISEhDm/9Qw895NhXkNDQUIdO7X3o2LEj586dA2DDhg306dOHxx9/nK+++qrwCbx+Lh5++GGHnJ07d2bUqFFcvHjRsa1jx44OWQ4fPlzo/H3++efF9g1wkunGcVFcX9yVkydPVrYIVYZqoQtJAoXCOY/xDQiLLaVWM+LwI3Raq0bp44lUwwurjwd4qEGpIt3qQcAHWaw6YiyxjrC/THSZnsojU1N5+KN0an6QRdsl2SRnleyFuuwziYbGNL7s8AJWozfSdUvfQAD3ZJ5m4LUwemSdpMf58zS4ZqVT+gUGJh3iMX00O5t+AcC/P/wfRv+x+NYfy++DN7L+4d1s6nWADZ33s//daD6ec5n/jL/IB+9exmAo/n5VLcZEOSHronozZ84cOnbsSNu2bXnmmWf4/fff+fDDD10uX+ZQiuzsbOrXr8+BAwfo06cPL774Ij169GDTpk2kp6czYcIEevXqxeHDh5k7dy7Lly8nIiKi1HoXLVpEnz59CAsLY9GiRaxfv55Lly6xePFigoKCmD9/PiNGjODYsWNs3LiRVatW8csvv3Dp0iVHHXl5eUyZMoXevXsTFhbGU089RVRUlGP/1q1b+fXXX1m/fj07d+4kNzeXRYsWoVAo6NevH/v373ccu3fvXvr3749SWbLLQqlUsn37dt544w1UKmcHfHx8PM2bN3d8DwgIwM/Pz0nmm2Xo0KFotVpWrFiBv78/Wq0WrVZLrVq1nI47cOAAMTExLtVpNBqZNm0aY8aM4dChQ+zYsYO///6bffv2lVp26dKlKBSlDyer1cqSJUtQq22ejuTkZBYuXMjSpUv56aef+PXXXzl+/Dhgiy1/9dVXHX1r3bq1o57izqWdgjrZsWOHS7JcvHiRdevWsXbtWo4fP45WqyUoKKjUPhkMBtavX+/4Hh0dzTfffMOmTZvYtGkT69ev5/z580WW7dGjh0POd955B4BmzZo5tgF8/fXXaLXaQkZvYmKiHFZTgLZt21a2CFWGO0IXLuZJVSpUWC0iVg9VfjlPNUi2vy0irN6TU2IdS3caMObZbPXsq2YESeJMisjq4yV7oRrmXCHZpzY5Hr54iKZ80bHijREBCbVkoQYJKDHTJu0s3mYjDTKTaWW0XTPMs77B12STr9e531EaczCm21YE/O24gfPnbPVe/jeP8F8zi5XljhgTLiLrouKRBMHpczv573//y9133w3A888/z2effcaDDz7ocvmbijHu378/Go2m0FKb4eHh1KtXjxEjRqBSqWjXrh1DhgxxyWus0WgcxkBwcDC//vorTZo0KXTcgQMHeOKJJ2jVqpXDC2vn9OnTmEwmRo0ahUqlomfPnk6hDbt27eLVV18lKCgIHx8fJk2axN69ezGZTAwYMICjR49iMpkwm80cOXKEJ554olS5FQoFgYGBRe7Lzc0tpCNPT88iPbAVQW5uLp999hnPPvusS8drNBq2bt1Kjx49MBgMpKamEhAQQGpqaonltFotBoOBdu3aldrG999/T4sWLRw6CwoK4uDBg9x9992Ioujkodfr9Xh7exdZT0nn0lVulMXLy8vRblnCg7744gv69+/v+N6+fXtCQ0Np0KABVqsVSZIqJNxo+fLlPP300+Veb0Vy6tQpkpKSiI+PJyYmhpSUFGJiYkhOTiYyMhJRFDl48CBge3M8KyuLsLAwRFF0jLOYmBji4+M5f/684//Y2FhiY2OJiIjAbDaj1WrJyspyvH1uf+iNiIggNTWV6OhodDodp0+fJiEhgYSEBE6fPo1OpyM6OprU1FTHA729rF0erVaL2WwmIiICvV5PbGyskyz2vhkMBrRaLaIoEhYW5iTPwYMHEUWRyMhIkpOTnXQRHx9PUlISp06dQq/XExUVhU6nczwo3ShPREQERqOREydOoNPpiIuLIzIykoSEBOLi4rh8+TLR0dEYjUbCw8OdytrlOXLkCGazmaioKBITE4mNjSUpKcnRt5SUFE6ePInBYCAyMhK9Xu+I07SvNhUWFoZerycyMhKDwcDJkydJSUlxnB97fYmJiURFRTmus0XJY5ezdEQiIiKKPNdardbmrBAAARQFf4Pi9b+v/ycI1hLHnmg2OIoW/CXnZeuIjY1Fr9cXOfYkINE7mKdPHKOmlEddkqlJKm04hSf5XmoBCQHna4TKamX//v1Yvb3yxRYEzMr8JbGVkrPHOicnvdixZ+9bRY+9uLi4Kj/2zp8/X+LYMxqNREdHc/nyZaf+xMXFodPpOHHiBEajscSxp9PpiIqKQq/XV+h1r7ixV1Cekq571ZVRo0YxevRodu/eXaYwCrjJGOPatWsXuT09PZ369es7bQsKCuLUqVOl1vnJJ5/w2WefMX/+fNLT0+nbty/Tp08vZBilpaU5pd2oW7euw6ublpZG7dq1nTyXDRo0cPydnJzMBx98wNy5cx3bVCoVycnJNG/enMaNG3Ps2DGUSiVBQUG0atWqVLlLwsPDo5ARnJubW6yxV958/fXXdO/enebNm7s0daRUKjl69KgjPKVVq1YYjcYSg9bNZjPLly9n3rx5LF26tMT6MzIy2LRpExs3buTll192bPf29mbr1q2sXLmSwMBA7r//fgASEhKKTbFS0rl0haJkCQoKYu7cucydO5eEhAS8vLwcoSLFER8fz6+//sq2bdvYuHGjU5+WLFnCtm3baNOmTZEPebfCyZMnSUxMZNKkSU7tlsbKlStZs2YNnp6etG/fnqlTp5arXKXxwAMPFNpWt25dAId3/vHHHwfy3xy35xy1x9sX9wCWkpLiqMt+rL2OkJAQADp37gxAnTp1AJxmWBo2bOi0zX6Mvay9Lnvd9rr8/f2L7a/9WHsf7HXY+2hfltbed7v8gONaavd02Ou6UR67HA8//LBD/oCAAKe6GjVqBNhmKAqWvfF/e1v262bB67m9PrvM9j7ZV5Oyf7fvtzsliuqTvf7i5OjRowf8XsrNTJJAUDj6f+O57tatG6GhoTRTXuIfoRkehlxyvTSYJZCMeUgKBQigFgT+OzYAhUJR7NhbOj6YD77JJDtXIqCBBmMqPNJYyYeD6+OptvXFLkfBsXfR+2cQFahFKyBQExPXvH35tdGjPB7/KzXyDCCBgboIWEn1qk2t3HQyPWqQYK1LSEgIuvr3c/GpZXin6jjWrBNB3iKKdoFkJuTQ+bmGnDb7EvNXDq1bezFkWF1UKsFJDnsfunTpAlT82LuRqjj2UlJSHPeZIsceNgdHcdj7W9LYK6iPirzuFZTjZq57FUVl5jGeMmUKb731Flqtlh9//JFFixYREhLCrFmzXCp/U4ZxcQH+QUFBhabdExMTSz0JkiRx7tw5pkyZwrvvvss///zDrFmz2LFjB2PHjnU6tk6dOiQlJTm+63Q6h5cxMDCQ1NRULBaLI6whJSXFMRgCAwOZPXs2Dz30EAAWi4WEhATHDfGJJ57g0KFDKBQKl7zFpdGsWTOnsImMjAwyMzNp2rTpLdddGsnJyRw8eJBt27bx66+/ulTm9OnTrF+/nq+//prGjRsD8Nprr5VY5ttvv6VDhw60adOm1Po///xzhg0b5nSxslgsZGVlMXLkSIYNG8Zbb73F+vXrefvtt/nzzz/5z3/+U2RdJZ3L1NTUUl+KKUoWgJYtW6LT6Vi4cCE9e/Zk0KBBJdazdOlSJkyY4PA2gy2kx2g0MmXKFF599VVeeeUVfvjhB0aMGFFiXa4iiiJLlixh5syZLoWvFGTSpEk8++yzGI1GPv30UxYvXuy4Ybo7WVlZhc7nnUp10MXX3RSMOS6VGDIhTS39Fjb7oX8YPvwBWs5MISHLigJ49l7Y8orrxsFdwSq+nVZ2Y6JZ9nIOtfyGVtcy0YgWJEBhFkjyq0lU3XuJ15kZcvQpfuu3Ex+FguS8ukht6/DIml70fqgZALXuC6LWhQUADCuijftdlKU6jInyQtZF9UehUNCuXTsuXrzIxYsXyxRXfsvp2jQaDdnZ2YDtiTQtLY1t27ZhsViIiYlh586dTtPMRSEIAgsXLuTrr7/GYrFQp04dBEFweGLUarWjjSeeeILdu3cTExODyWRi1apVjnrat29PjRo1WLduHWazmePHj3PixAnH/oEDB7J+/XquXbuGxWLhv//9L2+++aZjmrtfv35ERkby22+/0a9fv1tVDSEhIRw5coTo6GhMJhOrV6+mc+fOtyWv5o4dO3jllVfw9fV1uUx2djYKhQIPDw+sViu7d+/m1KlTWCzFv+Cyc+fOYo3XG+uOjIzkhRdecNoeHh7OqFGjSExMRBAE1Go1kiTx559/EhAQUGyMb0nn0mAw4OHhUWZZJEli7ty5PPbYYy6tjHThwgVycnIKjZUdO3bwn//8h/T0dBQKBSqVqkypYkrjt99+o2XLlg6Px80gCAIqlarUl0vdidLeB7iTqA66GN25fBZmtT8kn5tfl9wVdTCvqMOWV+qUS92uUC/xHNH165Oh8kWv9OWfurV57Hkvev87hZcNM6nVoQ0DUmbSM3k6I3Qzee7kyzS7bhSXJ9VhTJQXsi4qnsqMMT5w4ACvvfYaAwYMID4+nk8//ZQff/zR5fK3fOXp1q0b27dvZ+jQofz444+sWrWKJUuWsGbNGgICApg4caJLybQ/+eQT5s+fzzfffINaraZ///4MHjwYsBlBH3/8MQkJCbzyyitMnjyZGTNmkJ2dzfDhw9FobEuDqlQqVqxYwUcffcQ333zD3XffTdeuXR1tvPjii5jNZsaOHUtWVhZt2rRh2bJlDu9yzZo1ue+++8jLy3N4mW+F1q1bM3v2bD788EPS0tJo3749H3zwwS3X6wrNmzd36O9GwsPDHVMuoihiMpmYNGkSy5cvp0+fPowYMQKFQkHbtm0ZOHAg8fHxxbYzduxYlwx9i8XCm2++6ThXdnr16sXff//NmDFjMJvN3HfffYwcOZIhQ4YgSZJTujqTycSnn37KQw89VOy5/Oqrr9iwYQPPP/98mWXZtWsXly5dYtmyZaX2B2xhJEWFIgwfPpzz588zbJjNv/Poo48ydOhQl+p0BaVSyaRJk4rcp9frC6X4GzNmjOOFxuXLl7Nq1SrH+Z09e7ZLoU7uQEkPQ3casi7ycSWlWkUyOOd9vglcQnyQGgEQdUncO/nlUsuVN/KYyEfWRfXmyy+/ZPjw4SxbtsxpNtdVBKmikxDLVHlOnjzJunXrWLduXWWLAtjCb8aPH1/kS5uDBg1i7dq1TrHjMiXTsWPHOyI9UWxsrONN5Dud6qILYZG56FAKSQIJpOnqwvsKEBoaSk5OjssvIFdnqsuYKA9kXVQ8Xzf73un7mIvu86J4uSwJLePeqFSq2/ZCoCsoFAp8fHyK3Ofj41PmuNo7nbKE07gz8sNSPtVaF9d9OaUZxXZunBm6U6nWY6KMyLqQKYnb5jEOCQkp9g3//v37u/y24O2mIuQOCwvjvffeK3b/ypUri3yLVUZGpniioqLKlKuyOlNddFGkx1iSyHxTRQ1N6SESoaGhZGVllRhadadQXcZEeSDrouLZ2PwHp+8vXijq1dGqiRxKISMjUy0wm82OxVrudKqLLoozjKVprvUtNDQUURR58sknK0A696K6jInyQNZFxePOhrE8Jy0jI1Mt+P333ytbhCpDtdWFJOUvzuEiWVlZFSSMe1Ftx8RNIOui4pEE58/tJDU1lXHjxhESEsK1a9d4+eWXSUlJcbm8bBjLyMhUC27MxnEnU610IUmOz6L2AtKMssUMl7QAy51EtRoTt4isi+rN3Llz6dOnDx4eHvj7+9OmTRtmz57tcnnZMJaRkakW2JdBlak+upCmqZ0+Ux8ve4bR9PT0CpDM/aguY6I8kHVR8VRmHuMrV64wfPhwFAoFarWaadOmOS0MVxqyYSwjI1MtcCVf+p2CrIt8atasWdkiVAnkMZGPrIvqjSAITotqGQyGMi2yJRvGMjIy1QLZC5RPddHFjv/LQ/VBFuoPcxCmJt9UHbLH2EZ1GRPlgayLiqcyPcZ9+/Zl6tSpZGVlsW3bNsaMGVPqCswFkbNSyMjIyMhUOSbvz2XFCRFyLGCVwFsF2VlIi4teJr4o7IsEDRo0qKLElJGRKYIvWv3k9P2Vf4bc1vZ37txJWFgYoijStWtXnnnmGZdXwZQ9xjIyMtWCiIiIyhahylAddLHiNwvkWsEs2jJRGMzgV6PM9WRmZlaAdO5HdRgT5YWsi4qnMrNSTJ8+naeeeorly5ezcuVKhg8fXqal4WXDWEZGplogL4qTT3XQhcpogRvnM61ln+AsbhXNO43qMCbKC1kX1ZszZ85wK8EQZX/FV0ZGRqYKcvbsWdq3b1/ZYlQJ3F0XQW9fA7UKlAW8PDfpdTIajeUjlJvj7mOiPJF1UfFIitvsJi5AnTp1GDBgAPfff7/Tg7GrKdtkw1hGRqZa0Lhx48oWocrgzrqoN+Ua13w8ECWFzRiu5QUKAYxmCruQS8fDw6PcZXRH3HlMlDeyLqo3DzzwwC3NCtxxhvHVq1epXbs2KtUd13UZmWrNtWvXqFWrVmWLUSVwR10ETr6KSq0m29cTyXp9o0KwBfypFaDSQHZemes1m83lKqe74o5joqKQdVHx3O5MFAWZOHHiLZW/o2KM09LSGDZsGHl5Zb+4loXhw4dX2eD+cePGsX379nKr79dff2X06NGFtickJPDYY4+Rk5Pjcl0Wi4Uvv/ySYcOG0bVrV/r168fcuXPLlJi7KjN9+nTWrl1b2WJUW7y9vStbhCqDO+rCV1Bi9dLQSmdEY72ec9QqQbYFLCIIAijLfstSKpXlLKl74o5joqKQdVG9GTRoUJEfV7mj3KYmk4nc3NwKb2fHjh0V3kZlY7FY2Lp1K2vXrqVFixZO+8LCwliwYAFZWVku1yeKIlOmTEGv1zN37lzatGlDVlYWmzZt4oUXXuDrr7+mYcOG5d2NCicrK4sNGzZw7NgxEhISOH78OOHh4Tz++OOMGjVKnrmQkQEC30zB4OuFh0pBdJOakGOGvOvGsZfK5jnOs9oM5HLi3M//x//eisKiB1OgHx7ZeQg5Zqw+HlhEEbNgRaVQk+PriaAy8cbpYbfUnrghFGKvolj8Sjn1QEam6lKZMcbvvfee42+z2cyhQ4eoW7euy+Xd9q4cFRXFihUruHjxIvXr1+ftt9/mkUceYc6cOfj4+HD27FnOnj1L06ZNeffdd2nTpg2jRo0CICQkhPXr19OmTZti6w8NDWXHjh1s3rwZgJycHLp3787PP/8MwPPPP8/YsWP59ttvEUWRfv36MWXKFMD2tDJ9+nS6devGmTNn+OSTT7hw4QLt27enVq1aBAcHM378eMaNG0fv3r159tlnAdi+fTuHDx9m3bp1AHz//fds3bqVzMxMHnjgAWbOnElgYGCpuunYsSMTJ07km2++wWq1MnToUCZMmIBC4extKa39kpg/fz7//vsvI0eO5Pfff3ds37t3L59//jmvvPIKn376aan12Dl06BBnzpzhp59+cjzN16xZkzfffBOdTseyZctYsmQJoaGhfPTRR3h4eGC1Wqlbty6zZ8+mY8eO5OTksGrVKkfy9q5du/LWW2/h6+vL2rVrOX/+PAsXLiyy7x07dmTbtm20bNnS6fyZzWaGDx9Oq1atHGVdxWQy8fLLL9OsWTNWr17N0qVLadGiBSEhIXzyySf8+eefLFu2jJMnT7Jw4UI6dOjAnj17qFmzJq+//johISHMnTsXpVLpeGnAarXSv39/li5dyvfff09AQACTJ09m06ZNfPHFF7z88st88cUXiKKIyWTCy8sLgF27djFz5kynPtu5cay7K2WZnajuuJMu6ryWgOTrgwkJk7fGtlGjtBnGXirwuO7xVStsOY1vYNWc/+P8X8XV3gyAY9/+n2NLUHwyTS+nklnLE0ntT14DJao8kYYpuYhKBZkaCclTTWqjIMweGiQk/HUGPm+7l1xvDUHnT9M5IwUVKsx4ICHggQkVFjRkI6AgDwVGfMjDG5CoyWUEbP1QLnmVGqQVkFGg+NhpAdvErljMMfayquvHmRF8NLbtPdqCyQzaWAACm9aGLvdA+2aQmwfPdYNGBe4nkgRbj0JyOozqCfUCCjd3IBr+uACDOsI97hun606/D5my06lTJ6fvnTt3ZsSIEUyYMMGl8m4ZSqHT6Xj77bd5+umnCQsLY+LEiUybNs3hodyzZw/Tp0/n4MGDNGzYkNWrVwM4bvz79+8v0Sh2BYPBQGJiIqGhoQ4j5c8//3Q6xmQy8fbbb9OzZ0/CwsIYPHgw+/btc6n+Q4cOsXHjRhYvXsyePXsIDg5m1qxZLst37NgxduzYwddff83+/fv58ccfy9S/0hg/fjzr1q0r9BLDww8/zA8//MCjjz5apvq0Wi1dunQpcopr0KBBHD9+HIvFdlNs3bo1Wq2WY8eO0b9/f1asWAHAvHnziI+PZ9u2bXz33XekpaUxb968m+yhja1bt3L16tWbKvvrr7+SlpbG3LlzCQrKX5SgadOmfPLJJxw7doz/+z/bDfvChQuo1WoOHTrEO++8w9y5czl//jz9+/fn119/dfQ9MjISHx8f2rVr56gvMzOTrVu38ssvvzBmzBi0Wi0rVqzA398frVaLVqu9I+LpXHlovFNwF110fS8FLy8v6uSZaZqdR9uretqkZqJUgJcGBKesFIItU0UB/okxlGAUg81wdPZc3XUxhcwAD5QWEZXFipcxD0+TBUkhENu+AfGtg0gNDiD40jXUZgsCAvpAP1Ia1SKzti+X2j6ILxJe5OFPFjUw4IkJNVZEPFFjJhd/cvFHRI2IBhUSNUmkJol4kI2E4JBMQCrw940fCbA6ekKR+wE8ANH2PdsE2bmw5w84/BfkWSDPgmfcVdh4BN7cADM2Q+d3ILOAgTj7Gxi1AqZtgq7v2ozngnwfASEfwjtb4JGZcOHmViKsCrjL78OtEQTnTyWSnp5OSkqKy8e7pWF87NgxgoODGTx4MEqlku7du7NmzRrUajUA3bt356677sLT05O+ffty+fLlCpFjzJgxaDQa7r33Xpo2bcq///7rtP/06dOYTCbGjh2LSqWib9++dOzY0aW6d+3axfPPP0+LFi3w8PBg4sSJxMTEcOnSJZfKT5w4kYCAABo2bMiIESM4cOBAmftXEnXq1Clye61atW4qPECn01G7du0i9wUGBmKxWMjIyHDaLggC2dnZ1KpVi9zcXA4fPsykSZOoWbMmfn5+TJ48mUOHDt10+My1a9f47rvvePLJJ2+qfGJiIsHBwXh6ehbaFxgYiL+/v2Nsent7M3HiRDQaDY888giPPvoohw4domPHjmg0GiIjIwHbQ12/fv2c6tq8eTO9evWiRo2yL35wuzl16hRJSUnEx8cTExNDSkoKMTExJCcnExkZiSiKHDx4ELAt25qVleVYvUir1WIwGIiJiSE+Pp7z5887/o+NjSUuLo6IiAjMZjNarZasrCzH7MH+/fsBW2L/1NRUoqOj0el0nD59moSEBBISEjh9+jQ6nY7o6GhSU1Md7wnYy9rl0Wq1mM1mIiIi0Ov1xMbGOsli75vBYECr1SKKImFhYU7yHDx4EFEUiYyMJDk52UkX8fHxJCUlcerUKfR6PVFRUeh0OrRabZHyREREYDQaOXHiBDqdjri4OP744w8SEhKIi4vj8uXLREdHYzQaCQ8Pdyprl+fIkSOYzWaioqJITEwkNjaWpKQkR99SUlI4efIkBoOByMhI9Ho9YWFhAI5rS1hYGHq9nsjISAwGAydPniQlJcVxfuz1JSYmEhUVhdlsJjkTMlVK7jIYeUCfzf0pmTyQrKf7xWsYBSWSKNk8mWALo5AgLi4OnU7HiRMn0O69VuYxKNyQ31SQwOjrQWqQH9m+3uR5aDD6eJNZ05sa+hxu9NTmeWscxiyAEtHJaLWiIg+NUxkPDAX+zkVRhuwaQqF/izvKWuzeIklI41rEn46xl7P7RP6+c0lk/d9F57F3pMATiCEXY/hfRY69uLi4Kj/2Ll686NR+QXnCw8MxGo1ER0dz+fJlp/4UHHtGo5GIiIgirzNarRadTkdUVBR6vb5Cr3t6vf6WrnvVkRtjix9//PFC982ScMtQCp1OR7169Zy2FfSg1axZ0/G3SqVCFMsvLq0gN7ZzY0Lp9PR0AgMDnUIYXI2TTU5OZs2aNaxfv96xTRAEkpKSaNKkSanlGzVq5Pi7bt26XLtW9hvI7aRWrVrFemaTkpJQKpUEBAQAthtjz549MZvNmEwm5syZg8FgwGKxUL9+fUe5+vXrI0kSqampNyXTqlWrGDlyJFlZWaSlpRXa/+WXX7Jx40bHd7vhYic4OJgrV65gMpkKpYy6du0aer2ehg0bkpubS926dZ2OsZ8zhUJBSEgIBw8e5MEHH+TXX391CnnYunUrNWvW5JtvvnGpTytXrmTNmjV4enrSvn17pk6d6lK58qKoFDr22C+7V/3xxx8HoFevXgD07NkTgG7dugHOv/WCGI1GR+iI/Vh7HSEhIYBtSg3yH+wKetLtv037Nvsx9rL2uux12+vy9/cvtr/2Y+19sNdh76N9ys/e94JxcPax/OCDDzrVdaM8djkefvhhh/yNGjVy6ALyrwc9evRwKnvj//a2GjRo4CRDQdnsMtv71LdvX6fv9v12J0BRfbLXX8cfDNki2Wol9XLzs0fUMubZvEwiYBJtxrEggAR33XWXo7+1fAz8+Xs2ZeFKUG08zdlI2MxJs4cKSYAsX42TZ0tUKLAqldwVk8jFVvXI87I5XlR5zuEc9nrsCIj4YCCT/PtDLj54YPPO5qFBTZ7LaZltdxWxyLacj1IChUNNiqVRIIGd7yPQ7/os3aBH4M/vbX+3qk+Nds150MPW527dukGqGtbYDC1qeOHV8z46N7PdhwuOvULNVMGxV/BacaMcdjlLynNs76/9t3fjdcb+W7XroyKvewXluJnrXkVRVWKMBUGgVq1ahd6FKgm39BjXqVOnkFt8w4YNLntTXUGhUDil+dHr9WWuo0GDBqSmpjqmwQEnI02pVBbbRmBgINOmTSMsLMzx2bJli+PiURoF20lOTnaaynel/dvNY489xvHjxwt5hQF2795N586dHZ7ou+66i7CwMI4fP87ChQuZP38+vr6+aDQaEhMTHeUSExNRKBRODzCuEhMTQ2xsLMOHDy/2mJdeeskRqnCjUQy2i1PdunV5//33SU7On3aMj49n1qxZPPLII46LnU6nw2rN9/gkJyc7Hv769etHeHg4v/32G02aNHEKXxkxYgTDhg1j+vTpTuOsOCZNmkRYWBg//fQTGo2GxYsXl64MN+HUqVOVLUKVwV108duHdfE2GjlRPwBdgTCJZE+Nw1MsSBIqqwTmwi/ftWrnS4fuJbUgcaPH1yM7G48cC2aN7cU6i1KJR44JVV4eiuvZMARRpP6/6dx/8l8C0rLxMhgRrCIqk5k2f0aSCVhRkIeKLLzJQ4UVAQsiZlR4ko0/iajIAczoqIueQAwEIF33F0vX/c7534v+FOwJjjI46rBhAhS2776eUMMTBneEvu3BQwUeKrJa1oFX+8DqV2DxGPjtU/ArELr20XOwdTIsGQvH5sF1o9jB0Efg4AewcDT8Ph+aOTun3Al3+X3I3Bw7d+6kU6dOdOrUiYceeogWLVrwxhtvuFzeLT3GXbp0YcmSJezZs4eQkBCOHz/O1q1bGTas5LeGNRrb9FZ2dnap6VqaNGnCv//+y8WLF2nQoAFfffVVmdbaBmjbti316tVj3bp1vPrqq0RFRREREeGIb27cuDG//fYbzzzzDKmpqezdu9fxRDdgwAA2bdpEhw4dCA4OZseOHaxZs4bQ0FBHyEhJrF27lo8//phr166xbds2XnvttULHlNT+7eaxxx5j9+7dTJo0iRkzZtC6dWsyMjLYtGkTERERfP3110WW8/f3x2QyOV5KW7VqFZ9++ikKhYIVK1bQpUsXfH19yyzPli1bWLp06S1ljVCr1axfv54vv/ySiRMnOrJShIWF0atXL8aMGeM4NjMzk6+++ooxY8bw+++/87///Y+3334bgDZt2hAYGMi6desKpZxRKpW8+uqrHDt2jG+//dbxgmlpCIKASqWqVquC2b0iMu6liwv/DabGO+kcbFCLpmkGrAL86+9jy0KhEFBeNwFb6I2c9y8cljT6zXsY/WbRdYeGhgLc8Lu5B4DU2CS0/9GSdlFFnrcHXplGal9KRvTwRGEycdVfQWpgIKJCQOOZyXuHHr9evutN9VM8Ho31zDXUL/e5qfK3SqmBVoIAz5f4lAF97rd93Bx3+n24K5WRx/iDDz7g6tWrjhAgOxaLhQsXLrhcj1saxgEBASxfvpylS5eyYMECgoODoExdxAAAtJdJREFUWbx4sWOqvThq165Nly5dGDJkCMuXLy8x3rddu3YMHz7cYVCOGjUKPz+/MsmpVCpZunQp8+bNo2/fvrRt25b77rvPsX/s2LHMnTuXvn370rBhQwYMGOCIJR0wYACZmZlMmjQJnU5HkyZNWLFihcsyNGjQgOHDh2OxWHj++ecZOHBgoWNKav92IwgCCxcuZOvWrY7cxd7e3jzyyCNs3brVMb0GtuU8u3XrhiiK+Pn5MWXKFHx8fHj77bdZuXIlzz77LHl5efTo0cORKQRssWP2qSaTycRff/3leDETYObMmXz/vW0qsXPnzo7pslvB19eXSZMmMWnSJKZPn06LFi0YP358oeNq1KhBSkoKISEh1KpVi/nz5zuFw/Tr14+1a9c6pg4LolAomDFjBhMnTqR///4lvliyfPlyVq1ahUKhoG3btsyePbvaeE+OHDnimEK803E3XWR9WpPGb6VywcfTlrsYridbUGDxst2mLnhrCCj4spiLpKenF7m9zt31GXqk+Bmh8kbRpT2KLretuUK425ioSGRdVE+efvpp/vnnH86ePesIIwGbLVaWlfAE6cbAWJkKpSTjqLwomHpMxjUSExMZP368w8N0Ozl58iQzZszg8OHDxR6zd+9e9uzZw6pVq26jZDIyt4+gt67iIypIqOFJnofKZiB7qZwW9fDJyMbwcfEx3TdStMdYRkamovms/V6n7/+J7n/b2i4ufNRV3DLGWEamvFEoFPj4+FS2GIUwGAycPXuWzZs333R2jDsF+5vYMu6pi+Rl9Ti3PBBFbh5KlQJ8Nc4r3YkSuTexil1xHuM7DXccExWFrIvqTVJSEhMmTGDMmDGMHj2aF154wfEyoyu4ZShFeTBjxoxil22uX79+lV29LiwszOmNyxtZuXLlLbexbNmyEvMeF/WiWWmMHj2aixcvFrnvgQceKBe5b4WgoCC2bdtWqTIUxaVLlxg/fjw9evSgd+/elS1OlUaeGs3HXXUhCALGxbVRT0sDP28ElcL2Dp4ggCjhkVeGrAvXuZmXb6sj7jomKgJZFxVPZWalmD17Nk8++ST79+9nxIgRHD58uMgwxOKQQylkZGSqBVqt1hFDfqdTHXSh/thIC10OFwJ9MSsgKMuEJjePS0tcX5whNDQUvV7PCy+8UIGSugfVYUyUF7IuKp7VHZwXM5v4h+t5hG+VgQMHsnv3bhYsWECfPn1o164dzz33nMsLncmhFDIyMtWCRx55pLJFqDJUB11YFAJ+uXl0SEjn4X8zuOuagas+pWfkuRF3WPjmdlAdxkR5Ieui4pEEwelzO7GHRTZu3Jh//vkHDw8Pp3SopSEbxjIyMtWCG5dkv5OpDroIfULkjyB/spUKQOL/An0xWUxlric7u2wLgFRXqsOYKC9kXVRv7rvvPiZPnswjjzzCl19+yfz581GW4f2EOzbGWEZGpnohZ2HJpzroYmB7b0745dLlK08EpZJccy7SvLqlF7yBgisA3slUhzFRXsi6uA1UXogxs2bN4vTp0zRr1oxZs2YRERFRpsWsZMNYRkamWpCYmFji8sx3EtVFFw829yT3Q/uiHjeXNSYvL6/8BHJjqsuYKA9kXVRvBEFAoVCwbds2hg4dir+/P82bN3e5vBxKISMjUy0obYGfOwlZF/mUZQq1OiOPiXxkXVQ8lRlj/MMPP/DOO+/wxRdfkJWVxeuvv16mTGOyYSwjI1MtMJnKHn9aXZF1kU9JiZf0ORaEeUaED3MQPjIizDchfJp7G6W7fchjIh9ZF//P3rnHRVGuD/w7O7vLzQsqKl5KS7tbmceuhqaWdz1lN09Z2c0uP7M6qZlWR09l5iXtYqZWdrPUOh2NjncJ3OQkQUJxQimVBAFBFhYW9j7z+2PZhRWERQWW5f1+PqPszLzvPPPsMzPPPvO8zxvcfP7552zYsIE2bdrQqVMnvvnmGz755BO/2wvHWCAQBAUNGXUc7AhdVFGXYxy5zAFIEK6HUC1oZZA1DPkg+NIvhE1UIXTR+KgayWdpSjQaDW3atPF+7tatW4PeHAnHWCAQBAWiLFcVQhdV1PlAVIGQyqE2nte9Gg17jCDNNze6bE2JsIkqhC6Cm8jISDIyMpAqr+lvv/22QTnlYvCdQCAICgoKCujSpeFVC4IRoYsqHA7HqTdKEqhqlVNcfX2Int/y7VwarW9cAZsIYRNVCF00Pk2dV1ydOXPm8PTTT3P06FFuvPFGQkJCeO+99/xuLxxjgUAQFPTu3bu5RQgYhC6qCA0NPfVGWarpFEPlFNQubvnEybHng8MxFjZRhdBFcNOnTx82b95MVlYWLpeL8847D53O/8mBRCqFQCAICn777bfmFiFgELqoos4JPlTcEePakDUUWJ2NIlNzIGyiCqGLxqc5qlK89NJL3r9NJhN9+vThwgsvbJBTDMIxDmiOHTvW3CIIBC2GgQMHNrcIAYPQRRV15pO6Kp1iVQVntQFZlZ9D5WacpeAsI2yiCqGL4CQ9Pd3798MPP3za/QjHOEDZsGEDb7/99hn3c9dddxETE+NdbrjhBgYOHEhhYWG9bQcOHMigQYNqRFycTifDhw9n/PjxZyyfQHC22LNnT3OLEDAIXVRhMplqXa//exE4FXfahCSBQwWzA6wuMDtB1WCWQ2m3KDgG4QmbqELoovFpjohx9Qo0dVWjqQ+RYxyglJSUnNEX66F6UWtFUXjyySe5/PLL6dy5s1/tQ0NDSUhIYMyYMd51iYmJOJ3B84pREBzcdNNNzS1Cs+Aos6AcL8J+ohT1aBHWpTu55JejGO/KQJ56Ey5DBpafTyB/tROJULRocKCnLRYkJBQUXFhRkbCiYh83knNjp55dGa129r//P/6IzUCpULnsnn70/78rkM5SGacfvjzAnpXHMKoyJRHhSKEhVOh0mMMuw6XVsnXbH7S3WikK0ZPSoS3Fsg5tZBiOcC2U2twRYlkDTtyRYwnQSUhAhVNLn2XlHHr29GbeCxRa6/VRG0IXwY90Bs64cIwDgNzcXP72t78xdOhQ4uPjmTFjBmvXrkVVVe6//34+/fRTMjIyWLZsGQcPHiQqKoopU6Y0OGK7fv16zGYzjz/+uN9thg8fzvbt230c461btzJ06FBSUlIAsFqtzJ8/n8TERBRFQVVVrFYre/bsITw8nN27d7Nq1Sry8vI477zzeOGFF7jkkkuYOnUqw4cP5+677wbcUfLdu3ezevVqrFYry5cv58cff+TEiRN07tyZp59+mptuuqne41Vn3rx5bNu2zZtjtGTJEkpKSvjss8/Izc0F4Oabb+aFF15AkiTy8/N54403+PnnnwkPD+fee+/l9ttvZ8SIEd5z1ev1aDQaHnvsMdq3b8/WrVtp06YNiYmJdO/enRkzZnDNNdcA8OOPP7JixQqOHj1Kjx49ePLJJ7nxxhsBd0R+/fr19O3bF4BZs2bRp08fHnvsMYqKinjzzTf58ccfCQ0NZcSIETzxxBPo9XrmzZvHd999xyuvvMLo0aMBMBgMPPvss0yePJlnnnmGefPmERkZyTPPPAPAyy+/zN69e9m9ezdGo5F58+aRmpqKqqooikJoaCi7d+/235iaGOOqX8h7eg8um4JTgouHOwj/6H6Oj/mEowdC0XePoOi17lyX5sL26c9oLBZCu8noTPlIJ0qgdxcI1yHZnfD2wyiD+5Hz0C7K9+bRdlxv1INZRCV8D+3DcayYxoFFR3CWObhg8UA6//Vcrxx/rPiNwud+QOdw0iMqn+iw40j9zqW8zxXYN2Wg/Ut32sy4BteUFdgPFXFIvpAo5wkUtT2h2JBRKaEdTnSoaAingE7k4kSLmZ4o6JGx0U7KwqoLo1jbgZ4Vx7ATRhHRhONEwkJbitDg4gj9CEEiBAsdOIqeClS60IYKKj79meOfnqAbebTDhokeRFCEhINCenKCUMIopwN5gJbfuRINLnp+t5cKyUA+3cmjFy5JwqoLpbBzOC5Zg+xS6F2ezVWl+zkmn8MBfT93qq6kEl7hxCVrKOochjVcR2iFA1lRCLUqAIRoQHZB8lvZfLFdRULlnD9zCbPYcGk1OPRaNE4VVecusSa5XO4aqJIGXYUNV4gWW3gIkgouWcallVEBRQJJ1kNYGF2dTsJLSlFkmSOdOuCSnXSpsFAcFopJr+WYLgSbBB1MVsJMVox6T7k2t4xE6CBMS6jFQfcT5ehVKNdruHJaORU6mSORYbj0WqIiVG7pJrH3qML4izS8PUaHppHrtX6SZOOfOyx0jZC4MhwO5jgZfJmel+9qU++x4+LiGDZsWKPK11IQumh8mqMqhaIomEwmVFXF5XJ5//bg74yHwjEOEMrLy+nWrRs7duxAURSOHTvGoUOHWLRoEcXFxTzxxBM8/vjjvPfeexw4cICnn36aTp06ccMNN/jVf2lpKatXr+bdd99tUKHrESNG8PTTT1NSUkJkZCTl5eXs37+fmTNneh3j7777jsOHD/Pdd9/Rtm1bcnNzmTBhAgB//PEHL7/8MosXL+a6667j66+/ZsaMGXz77bd1Hvfzzz/nyJEjfP7554SFhfHJJ5+wePFibrrppjqPVxuTJk3yOoi5ubnMmDGDlStX0q9fPw4fPsyDDz7IzTffzDXXXMPs2bO54IIL2L59O4WFhTzyyCP07dsXg8EAwPjx45k1axYxMTEAxMbGkpSUxNNPP82CBQvYunUrM2fOZNOmTRiNRp577jleeeUVBg8ezI8//sjs2bP5+OOPvc7wqZg5cybdu3fn22+/pby8nBkzZrBq1SqeeuopAHr06EFCQoLXMd6+ffsp6zT++uuv7N692zs6//PPP8flcrFjxw5CQ0NJTk7m+eefr1Oe5sSRV07+k3FIivuGpagyh3apXHTrcn5P7wVosB11ce7ruVh/ywTABTjLignhhLuTQ7lVHd61hKIXn6XkS/e+xvd+pSe/EYYZisz8cn8CZrs7Opj+twSGFN+DJkSm/KiZnBmJRNrsRFBCtwJ3Ppv9zwqsuNON7EdLsPz3F8Lzs9EClyi/8is3cB7u8QJm2mIjHB1OnNjowWEAyumEE/ePOgUtLrUtHewn6GAvAUCPHRvh2AmnG1Xn0hYHEhpARwnn0okj6LCjwcWfXEIUhYTinuUrkjI02CkjEguRbnkJRY+NduQRhpl8egPQmwNkcREggQqm9iG4tO77hkvWkBN6DudZsrjAlkmRpivHwrrSocR9HFlR6GC0kh+uQ5UlQiwKnkekrKioSPx2eW8U2Z3Jd+ycaC7I/BOtU8ElK6g6rbdShCrLyC4XTh0oei16qwNHiA5Vo3HLI7mjumgkVKBbicmbH2jX6ehTVExRmwhKwkJBoyErNAxbpQNZLEt0c1TmFVd/iDsVqHBgLXdwOERHV4eLDnYFWVFpb3PStdxOrqThRJnElyfc7VckuRjcW8Nd/RrvkXq8TOHhDeW4FCguUChW3A/8DT9YGdhHx7ir66i+AcIRrIbQRXCSmZnJdddd53WGr732Wu82SZLIyMjwqx+RYxxAjB49Gr1eX6O8UEJCAl27dmXSpElotVr69evHbbfdRmxsrN99f/XVV1x++eX069evQTJFRkbyl7/8hbi4OAC+//57brzxRvT6qhJGERER2O32Wkd/7969m+uvv54bbrgBjUbDHXfcweuvv15vmsidd97JG2+8QVhYGMePHyc8PJyCgoJ6j1cfnTt3ZsOGDfTr14+SkhJKS0tp27YtBQUFHDt2jPT0dJ5++mlCQ0M555xzWLlyJRdddFGdfZ577rncd999aLVaxo8fT48ePTAYDOzYsYNrrrmGYcOGodVqufHGGxk8eDBbt26ts7+cnBx++eUXZsyYQUREBF26dOGJJ57gu+++8+5z3XXX8fPPP+NwOLBarfz888/eKHV1VFVl8eLF3HPPPd51ERERWK1WrNamn/p2//795OXlkZWVRXp6OgUFBaSnp5Ofn09SUhKKorBz507AHdUpKyvjv/E/gFLVhwQoyFiLK6h+C6s44ZtLqlJ7xEK12Dn2+9GT9q3qx+WqaqdYXezctgOAxIRENJV2q8FVa1sAV0VV3VwZF1KlJJL3L08fyin7OPkzgISCVF0RJ58XMgoSKm4H1omWmleZigvfEdou988N7zEdhKCggWr684n+SG4n1CHpKs/R6T2nKllBo6ju9Sd9DaoESrXopiL58RhS3e2qevdFUkGj+mpNo7p15ZJllEr5tSfddxyeBtXXa4Dyyu9QkijSySjVjqrxnNNJyk397Q8cDgcGg4GysjLvPXP79u2AOwWtsLCQ1NRUjEYjaWlp5OTkkJOTQ1paGkajkdTUVAoLC0lMTPRpGxcXx4kSMy6lSsTqZOed4NChQ97rymw2YzAYUBSF+Ph4ysrK+PLLLwHYuXMniqKQlJREfn6+z3WYlZVFXl4e+/fvx2QykZKSgtFo9AYGqstTVlZGYmIiFouFffv2YTQayczMJDMzk5ycHDIzM8nOziY1NRWLxUJCQoJPW49+4uLicDgcpKSkkJubS0ZGBnl5eWRkZHDo0CEKCgpITk7GbDaTlJSEyWQiPj4egB073NdmfHw8JpOJpKQkzGYzycnJFBQUcOjQIZ/+cnNzSUlJIS4uzuf41eVJSEjAYrGQmppKdna2z/lkZmZiNBrZt28fFouFxMTEWr9rg8GA0WgkJSUFk8l0Wve9+Ph4FEXBYDBgNpu934/ne/acm8lkIjEx8bRtr7FojhzjAwcOkJGRwYEDB2os/jrFICLGAUWnTp1qXV9cXEy3bt181kVHR7N//36/+46NjfVGTRvKyJEj+fbbb5k4cSJbt27loYceoqKiwrt9xIgRHD16lHvuuYfy8nIfx76oqMinkLpGo+GKK67wfn777bdZuXIl4C7Ef9lllwFgNpt54403SE9Pp2fPnvTo0cPrTNd1vPqQZZl///vffPvtt4SFhXHxxRfjdDpRVRWj0Uh4eLjPVJLnn39+vX327NnT53OXLl0oKirCaDQSHR3ts61bt25eBx/cI2c9uVBWq5U+ffpgNBoJCwvzee3TrVs3ioqKvLndYWFhXHbZZfz000+Ul5fX6hSD+3vv3Lkz1157Ld988w0A9957L0ajkbFjx6IoClqt1ueHTmNy1VVX1VjnsQ+Prm655RagKqoz+G8jyd2rp2RFGgpgR+aCnrm0+/wxeozayLGy7mj1KueuvwN50U4c2zKRcCHrQEWP5LCjhoeB4kRyuJDeeph+dwzm8I//xvrLCcJv7E5FlpmOOfkoGi19nr+IjLfycFlc9H3jL/T6q/vH5C33jeDnPT/g+PAXytSOGHXRdHTko+8djq53bxzxWWgujCJ8ztUoj61AstnJ0PWjves4ZiWCcCqIoBwbelxoUAmjiC50ogAdpWix4iQULVbCKMKq0VGi7US0PR87IViJQIeEmfa0wYQLDWWE0RYboNKW4+iwYSIaGTtdyaaQ7oRTThgVVBBCW0poSwHFROMiBB1W2pGPEw0F9ETCRXcOEYKFHhzmGOejwUmHYivH9TKKRkJ2qUTZi+hpzSZX2518ORqtU6UiTCbc4sIlS5S30WEL0SI7XFhDZEKt7h8SiiShdapcePAYv1/YA0lV6JZ73J0OodWABBqnC6UyOq1RFK9Tq7M5cOpkFFlCUkHrdOLUat0P3cpJOpwaDVpFQQWcGpnctm0oCg+jjdMJksRV5goK9FqMWi2hqkqRVobIEHAo7goVsgThWrA6vT/GNIBZL9PGqlChk8lvEwKyhEZWuayDxK/HVYb01vDSbReh00net0ke+x05ciSA9+2eZ3xHx44dvdeA5x7iWefZx9PW09fs4RUs3G1FjdBwTihkF7r4Sx8dU0ZFEab3dT48cnjyaT0pa57ry3PP8Fx31e/TnufNX/7yF5++TpbHc06eyFz1c/JwzjnnADBkyBCftif/7zlW9+7dfWSoLptHZs85eVLcPJ892z1VJ2o7p+7du6MoChqNplY5PHL279+/xrl48Jyv5/xP/q49+vLo43Tue55z8vRVV1DLI8fp2J6gJsIxDiBOlSweHR3Ntm3bfNbl5ub6bdxHjhzBaDQyaNCg05Jr6NChvPHGG2RkZJCTk8OAAQP44YcfvNtlWaZHjx7odDq+/vprZFn2pjZ06dKFgwcPevdVVZW3336b+++/H4Dp06fXyDEGWLBgAeeffz5vvvkmWq2Wn3/+2fuLuq7j1ceOHTvYuXMn69atIyoqCoC//vWvgPumUVFRgdls9jrH27Zto127dnWmrJxc4SMvL48RI0agKAq//vqrz7bc3Fyfm/WHH37ok2MM7u/bYrF401fAXbqvffv2aLVVl+xNN91EQkICRqORiRMneqMDHsrLy1mzZg0rV64kPz/fuz40NJQOHToQHR3NmjVrOHz4cECnUgB0f3co0YtuxFlqR6cHqaM71eE800ucW2hG0ymCn1KSuWbrg6hmG+g07tfsOhk1vwSpWwdwOMGlQKgeLXBh2j24zHbkNu4fBc6CcrQdQ+mqlek8X0F1KMhhvrfIAWtuxLHsWlwmO6E9IsBsQYoIpZ0kuY8boXdfx/fegFJh57x8M2Tmobn8HGRVxbrndzqlHkN3VXfMC3ehDB9CPiHoVBvObZlYCSO8vUpeig67qkcrSWTRCQWwYCMUO1Y0nKAzElbCOIgdPVp0lBKCmU64qCCfKMIx0pEySgmhhHC0WHDRFg06OpOBFtBiQ0LFhZa+/IgTDfl04BAd0F3THU1KMY6IEGx2Kx3yLEh6C/bwNpS3C+MreSRalxaXxYnDpqKoEka9hKwHtcxGSIUFuyxToddg1UhoHSqSDKASmX2ci/OOo4bIEKrDqpGRo1UcZRpUu4LsdGKVJJBAB2hcLsrbhSB1kInoDOW5KrYKBZfLDhLIWhnVJWGXwBQaiqSROBrZnoL2bZFUlTY2G1qnC7teR2erHa1OoUyvQ+0UDnoZQk4yuMhQKLODJGGPkMktd5AboqF7pMTiIRKj+mi4uLM7Ym62qbQJaZqI2OvjwnlpRBghWpA1EuVWhYhQ/178Jicnn/IHdGtD6KLxac6Z784U4RgHKHq93psqMGjQIJYuXcr69eu54447OHDgAJs2beLFF1/0q69ff/2Viy++uMFFrj2Eh4czaNAg/vGPfzBixIgaDvyJEydYvHgxM2fO5JxzzvEOagP3wLZPPvmEpKQkBg4cyL/+9S927drFtGnT6jxmeXk5ISEhyLJMfn4+77//PuAuFVdSUnLK49WH2Wz2Rkjtdjtffvklx44dw+l0Eh0dzVVXXcW7777Ls88+y/Hjx1m+fDnz58+vs8/ff/+d7777jlGjRvHdd99x4sQJbrzxRkwmE2vXriUuLo4hQ4bw448/smfPHlavXl1nf126dOGaa65hyZIlzJkzB7PZzKpVqxg1apTPfkOGDOH9999HlmWuueaaGo7xli1buPfee+nZs6ePY5yZmel1mDt27Mjhw4f91l9zognXoQ8/yYYlCbmLu07tpZde6l7VxtfLkbp1cP+h03JSFoHXKQbQdqmqOqDRakBbu8Oha6ND16ayozZhVcepflytjKZdGGHtwuDCqgowbXpfC/dX/n3fdbWf6Fmg+o87f9ACnvcu0XXt2MJY91ks5RYXd08aSft2Vd9VSamVwTOL+N0ZilVfy5gLnQwdKjVidaC+1u6Ux2gqp9hDeLXIsL9OMVRdHwKhC0HdiBzjACUmJoZDhw4xceJE2rVrxzvvvMOuXbsYPnw4c+fOZdq0aX4PIMjLyztlmoa/jB49msOHD3sHe1Xnrbfe4rzzzmPs2LE1tvXu3ZsFCxbw5ptvMnToULZt28ayZcvqHQD497//HYPBwJAhQ3jssccYNGgQYWFhHDlypM7j1ce4cePo06cP48ePZ9y4cWRkZDB06FCOHDkCuCPVJ06cYMyYMTzxxBM88sgjPgn8tdG7d29++OEHbr75Zr755hveeust2rVrxznnnMOSJUv48MMPGTp0KG+99RavvvqqN12kLl555RVcLhcTJkzgnnvu4corr2T69Ok++0RGRnLOOecQExNTqz4jIyN56KGHaqxfvHgxw4YNY8CAAfXK0ZLIyspqbhECBqELN+0ioV07q49TDBDZLpRfVvXAqteeeuY7SUK2WlFfCq99ewtD2EQVQheNj6qRfJaWhKSejWK5AkErJTY2lo0bN/LZZ581tyitnoKCAp80ldaM0IWb2NhY7HY7t99+e63bpdes3rQbTn71q6rgsqHO9j/yHsgIm6hC6KLxWTTEdxKVWQmDm0mShiMixgKBICgoKytrbhECBqGLKlwu16k3OlxgU8BW+z5ttS0r0lUXwiaqELpofJqjKsXZQuQYt3BGjhzpUyGiOqNHj2bOnDm1blu/fj0rVqw4Zb9fffVVjYoKAkEg05D63MGO0EUVdc+AJdVW/c3L+zefPCqv5SJsogqhC0FdCMe4hXPygCt/mTRpEpMmTTrL0rQ+xo8f3+AZCAWNQ0hI8DgxZ4rQRRV1O8Yq6DU1ahKjqqCq3NP/9AYsByLCJqoQumh8WlqUuDoilUIgEAQFJSUlzS1CwCB0UUWdqRQaqbJ+cbVHoaqC6kKd1TS1vZsKYRNVCF0I6kI4xgKBICjwTAwgELqoTl2T17w3AkAFRfFGiXE6UWf5P2lQS0HYRBVCF41PS84xFo6xQCAICv7444/mFiFgELqowmKxnHLbE9eFo74QijpLjzpT515eCDvl/i0ZYRNVCF0I6kLkGAsEgqCg+lTjrR2hiyoiIiLq36kVIGyiCqGLxqelRYmrIyLGAoEgKPjxxx+bW4SAIRh1If2jGGl+GTM2NKzUlijN5SYYbeJ0EboQ1IVwjAUCQVAQExPT3CIEDMGmC2meCdq0gfBQlv4ZgvRktt9t27dv34iStRyCzSbOBKGLxkfkGAsEAkEzExcX19wiBAxBp4uI8KrZ6TQaOLez302Li4sbUbCWQ9DZxBkgdCGoC5FjLBAIgoJhw4Y1twgBQ9DrogERqA4dOjSiIC2HoLeJBiB00fi0tChxdUTEWCAQBAUiClRFsOmindVR9UFV0bgUv9uKiLGbYLOJM0HoQlAXwjEWCARBgYgCVRFsugh3OH0+t7U6T7FnTUTE2E2w2cSZIHTR+KiS79KSEI6xQCAIChITE5tbhIAh2HRR0KbahBuShKltKLoXS/1qW1rq337BTrDZxJkgdCGoC+EYC1oNx44da24RBI3IVVdd1dwiBAzBpguFk0JOkoQzIsSvtqKOsZtgs4kzQeii8RFVKQSCAGfDhg28/fbbZ60/RVGYOXMmGzZsqLFt3bp1zJo1q0H9ZWVlMWfOHEaMGMGQIUO499572b59+9kSt1Vw8ODB5hYhYAg6XVSfstmz+PmwrWvmu/pYfs5XvNV942m3DySCzibOAKELQV2IqhSCVkFJSQmqqp6VvvLy8li4cCF79+5l4MCB3vUWi4XVq1fz+eefM3ToUL/7y8zMZOrUqTzyyCPMnTuXsLAw9u3bx9y5c3E4HIwbN+6syB3snHvuuc0tQsAQTLpY9L0NkEAFNNWcYcXJwxvL+fCuuiPCISH+RZY9bHnpJ3I+zkIN16ONDMMWqufD3t+gUyA7ui337o8nwulAQcGmC8EZ2oY2ZUdx0BYXoWhxoMeCBgU9ObRpyMHVbxoka0MIJps4U4QuGp+WFiWujogYC4KS3NxchgwZwrx58xg8eDAffPABCQkJ3H///QBkZGQwdepUhgwZwu23305sbKxf/TocDiZPnkzfvn1rTCs6Y8YMsrOzmThxYoNkXbZsGbfeeiuTJ08mIiICjUbD9ddfz4wZM7zpH6tWrao1Cp2bm8vAgQOpqKggNzeXmJgYPvjgA4YOHcqoUaP48ssvvfuOHz8eg8FQo49T9d3SOHHiRHOLEDAEiy7GrKlgXoLD7RCf9MNWQuKj/6nErKqosw+Hw1Hn9urEr87ixEd/otFpaGu20bXQTNsyGxlX9OF/V5xH12IbHZw2QlAIQUJy6OlcdgwNoYSgIqNFJQyVUDpiREcDB/5JDbt3NIRgsYmzgdCFoC5ExFgQtJSXl9OtWzd27drFypUrOXbsGIsWLaK4uJgnnniCxx9/nPfee48DBw7w9NNP06lTJ2644YY6+5RlmQ0bNhAVFcXUqVN9ts2bN4/OnTuzatUqv0tE2e12kpOTefzxx2tsGzNmjP8nW4nFYuH3339ny5YtZGVl8eSTT9KrV696zysYCA8Pb24RAoZg0UVSPljQgAS4VNBUplCoKqpDBY2Gn07UHZmSZdnv4+1fmUm0xYFTr0NXWRIu88Jzceh1OIBDF3RH+sO9rwYVLU60KESQRx6XePuxEwaAHhsOQNeQk24kgsUmzgZCF42PIiLGAkFgMnr0aPR6PaGhVaPaExIS6Nq1K5MmTUKr1dKvXz9uu+02v6LGGo2GqKioWrd17uz/bFweSktLUVX1rJaUeu655wgLC+OSSy5h3LhxAZWrvH//fvLy8sjKyiI9PZ2CggLS09PJz88nKSkJRVHYuXMn4K41WlZWRnx8PIqiYDAYMJvNpKenk5WVxaFDh7z/Z2RkUF5eTmJiIg6HA4PBQFlZmbdeqUcHiYmJFBYWkpqaitFoJC0tjZycHHJyckhLS8NoNJKamkphYaF35LqnrUceg8GAw+EgMTERk8lERkaGjyyeczObzRgMBhRFIT4+3keenTt3oigKSUlJ5Ofn++giKyuLvLw89u/fj8lkIiUlBaPR6I32nyxPYmIiFouFffv2YTQayczM9J5TZmYm2dnZpKamYrFYSEhI8GnrkScuLg6Hw0FKSgq5ublkZGSQl5fnPbeCggKSk5Mxm80kJSVhMpmIj48HYMeOHQDEx8djMplISkrCbDaTnJxMQUGB9/vx9Jebm0tKSgoOh8Pn+NXlSUhIwGKxEC653MPuFBVJlqryiiWJMI0Kikqo6qr1uzYYDGi1WiwWCyaTyS/bs7dXkGQJRa56NCqaqr+d1ZxsFXBWxpZcaNFSFZmWcbnbommQU6xAo9leampqk9iex/4C2facTme9tpeamkp2drbP+WRmZmI0Gtm3bx8Wi4XExMRT2p7RaCQlJcVv2zvd+57JZDqj+56gJpJ6thIvBYIAIjc3lwkTJpCQkEBERASrVq3i0KFDLFq0iLVr15KWlsby5cu9+3/99dds3bqVDz/80O9jTJ06leHDh3P33Xf7rK9+rPpwOBzceOONrFq1iv79+/tss1qtuFwur/xr164lNDQUrVbL+eefz//93//RuXNnJkyYwJ49eygpKeHOO+9k79693j4+++wz9u3bx7vvvsv48eMpKSlBlmUiIiK47rrreO655/jss8/8ljeQyczM5MILL2xuMQKCYNHFiTIn58zJJbxTR4zVS7apKpFmGyWKgnN+W2RN7dGp2NhYKioqalyjp0Jxqay6YBOSrCXMYSPE5iQ/uiNHzu+BRlHpnZXH3b/Eo1ccONEBTpxoCaGUEKxY6ISCTDjF6HBiQaIDhf6fcOzDMG6s//s3gGCxibOB0EXj8/LoZJ/P/9w68BR7Bh4ilUIQ1Ei1vM6Jjo5m27ZtPutyc3Pp2LFjU4nlRafTcfXVVxMXF1fDMd60aRNffPEFmzdvBmDw4MEsWrQIp9PJpk2bmDlzJmvXrvVpY7PZKC0tpV27doB7oGDXrl292xcsWEBMTAwmk4kXXniBNWvW+ETTWzKniuS3RoJFF1FttVjeOZfwBVaf9XqHC4tWg/pS/UPbdDr/Y7YaWeKJw7cBsOScL3Cd2xWtQ6Fb8q9odTqe/uN24JZTtj9ZmlqvLGkiRADmxhtoVxvBYhNnA6ELQV2IVApBq0Cv11NeXg7AoEGDKCoqYv369TidTtLT09m0aROjR49uFtmmTZvG5s2b+eKLL6ioqPC+5lu5ciVTp06t4dxrNBpkWcblctXa37vvvovdbic9PZ0tW7YwdmzNCJQsy3X20RI5evRoc4sQMASbLhwnRYSdsga73r+4js1mO61jzsi+h+f3Dee5n29h9p+TmPHH7afVTw3Ub5rcKYbgs4kzQeii8WnJdYxFxFjQKoiJiWHDhg1MnDiRb775hnfeeYelS5eycuVKIiMjmTZtWrNNE3rxxRfz3nvvsWrVKj766CMcDge9evXipZde4uabb/bul5CQQExMDIqi0KNHD1555ZVaI+Lh4eGMGzeO0NBQZsyYwYABA7zbZs+eDbijaAMGDOChhx6qtRZzS+Siiy5qbhEChmDThVOjqapdrKooGgkc/k0LHRYW1sjStQyCzSbOBKELQV2IHGOBIEjw5FXv2bOnVY66TkxMbBXVN/wh2HQhLbCCJ0KsquB0Qc5x1BW96mwXGxtLaWkp9957bxNIGdgEm02cCUIXjc+LY3/2+fzqfwacYs/AQ6RSCASCoEA86KoIOl3I1d6MSBI41XqdYg+efPvWTtDZxBkgdCGoC5FKIRBUY+TIkVRU1D5hwOjRo5kzZ06D+jtw4ACPPvroKbfPmTOn2XKbg424uLhmS4cJNIJOF5qTYjg6/2M6/tYUD3aCzibOAKGLxqel5RVXR6RSCAQCgSCgkRbZfZ1jhwP1hfpzhz21ycePH99YogkEglqYO26/z+fXvruqmSRpOCKVQiAQBAWeovaC4NOF1uGqmhJaVcHm38A7EBFjD8FmE2eC0EXjo0q+S0tCOMYCgSAoEK9Gqwg2XTjmhhFqsaN3OMBiQ53X1u+2Z3NWyZZMsNnEmSB0IagL4RgLBIKgwDNtrSA4dWF5OQLbC2GoL9c/qUd1TCZTI0nUsghGmzhdhC4aH0WSfJaWhHCMBQJBUHDdddc1twgBg9BFFW3b+h9dDmaETVQhdCGoC+EYCwSCoOCXX35pbhEChtagC4vVycWPHiX6mROcM72Avnen17qfZ8bL1k5rsAl/EbpofMTMdwKBQNDM9O3bt7lFCBhagy4ue7aQgnZtcUgSEqDt2aPW/cTMd25ag034i9CFoC5ExFggEAQFubm5zS1CwNAadGGXZcqRsKtgU8GCxNA7f665n93eDNIFHq3BJvxF6KLxackRY+EYCwSCoCAyMrK5RQgYWpIuer1iIWSehTKbo0HtXCd9llSVyy0qaam+5dlkWT5DCYODlmQTjY3QhaAuhGMsEAiCApvN1twiBAwtQReKoiDNs3BUJ2MP0dBuifuzv1xgttDGWVXP2CVr2NKrB/P/me2zn5jDyk1LsImmQuii8RFVKQQCgaCZcblOjiG2XlqCLuR/2kArgdUFdgVQQef/AzRCcXF+uRVkiXYSRLoUSrUyFfj2IRxjNy3BJpoKoQtBXYjBd4JWxfHjx+nUqRNarTD9YEOU5aqixejCqYJc6cg61AaFakw6LbYQLd0cLjo53JHjcknCoio++4lUCjctxiaaAKGLxqelzXZXHRExFrQaioqKuP322894MM748ePPeoH4xugT3K+rZ86cyYYNG3zW//vf/+a2225jyJAh3H///ezfv/8UPbQcCgoKmluEgKHF6EKvgTCtewmVoQHBXUlRKQrR09ZZFf2LUFVCZd8fvQ5Hw3KXg5UWYxNNgNCFoC6EYyxoNdhsNqxWa3OL0WTk5eXx7LPP8v333/usT05OZsWKFSxcuJDvv/+eu+66i2effZaSkpLmEfQs0bt37+YWIWBoEbqQcEeL7UplKoV73Rfp9V+jD9z8X9qg0qfCilVTFZpyAkplxPiToRsoesKCbaaGj7pvOEVPjY/zUDYFI1/D5XSh+POjXFHq3+c0aBE20UQIXTQ+KpLP0pIQ75MFLZaUlBTeeustjhw5Qrdu3fj73//Otm3biIiI4ODBgxw8eJDevXszd+5cLr74Yu677z4ARo4cyZo1a7j44ovr7P/rr7/ms88+o6SkhEsvvZS5c+fSs2dPAH766SdWrFjBsWPHGDBgAK+88grt2rXD5XLx0Ucf8e2332K1Wrnxxht57rnnaNOmTb19evjoo4/YvHkzq1atIiIignnz5rF//37Cw8O59tprmTVrFiEhIXXK7nA4mDx5MrfeeitlZWU+244fP859993HRRddBMC4ceNYtmwZhw8fZsCAAf5/AQHGb7/9xjXXXNPcYgQEgaqLlUk2ntypgKRxO8ZWF0iS20FWJJA13PsvhXu/trijyRoJHAqySwEVhv6Zx6XHjXTXypRZbJi0OspVMGpl2roU2tocDMvI5sNeB9EpGnQuBadOgxSi4cNzNxLqUuiaV46EhEujQVYUtNVSLxwaGYdG445cSxIRLgvhqsv7WJf8CGn3CjuIQwnFYVOI4gQhlNMRLeW6/0MlDFDQ4EKDEyMdkJHpSC56TN7+7URgpy1hnEBCwkk4KqCnDCc6AMrpRHtycStSwoEWBxHoKMbn7tC9A5itSISSc+ENaAtO0L7wd3RdQ6i4/gY0/05E67RAt/aEXtcDbrgIvk+Hy86B+ZNAp4XMXHj5S/ffr90D53Y+Y1toTgL1+hAEBiJiLGiRGI1G/v73v3PHHXcQHx/PtGnTmDlzJmVlZWzZsoVZs2axc+dOevbsybvvvgvAZ599BsD27dvrdYoTExN57733eO2114iLi+OSSy7h5Zdf9m7/5ZdfWLlyJd999x25ubl89dVXAKxbt47vv/+eNWvWsGnTJqxWK4sXL/arT4D169d7neLo6Gg+//xzNBoN27dv54svviAjI4OtW7fWqx9ZltmwYQNPPfVUjXzqsWPH8sADD3g/p6amUl5ezvnnn19vv4HMwIEDm1uEgCEQdXHUpPLk9xJIlSkTHn9UI1WlUEiS28/TaUAvg1YDoTIuJC4pLObmrFy6W6y0t9ooB1IjwijSaWlnd9CjvIKbMo/Sp7iMtnYXoU4FRavBoZdRZA3oZELL7MiAXavDJWnQqQoa8C4higuNCmg0IEmUy+7JQaTKpeqvUy9/Wi4k19abdpQTSmmlE+xCoS0qOlRCcBGOg3a0p4xSOmJFj6YyriYBIZQThhEtTmQchGBCQkZFiw4bWmyEU4KTUDSoaFDQY6cNxehPVnxuMZRaiCgtpkOygS5HkwizGNFm5aH9Mo421jxCnSWEZP+J66skeHYtfPsTvP4NLPzG3cfY12DDXvg8Ae5cclbtojkIxOsj2BBVKQSCJuaHH36gR48eTJgwAVmWGTx4MCtXrkSv1zN48GAuvPBCQkNDGTFiBNnZ2fV3eBLbt29n7Nix9OvXD1mWefTRR5kxY4Z3+5133kmHDh1o374911xzjbdg/ObNm3n00UeJjo4mIiKC6dOns3XrVmw2W719/uc//2HZsmW88847REdHA6DX6zlw4ADbt2/H4XDw+eefc+utt9Yrv0ajISoqqt79Dh8+zPPPP8/jjz/eJLU99+/fT15eHllZWaSnp1NQUEB6ejr5+fkkJSWhKAo7d+4EIC4ujrKyMuLj41EUBYPBgNlsJj09naysLA4dOuT9PyMjg127dpGYmIjD4cBgMFBWVkZcXBzg/j7B/eOksLCQ1NRUjEYjaWlp5OTkkJOTQ1paGkajkdTUVAoLC0lMTPRp65HHYDDgcDhITEzEZDKRkZHhI4vn3MxmMwaDAUVRiI+P95Fn586dKIpCUlIS+fn5PrrIysoiLy+P/fv3YzKZSElJwWg0enPQT5YnMTERi8XCvn37MBqNZGZmEhsbS05ODpmZmWRnZ5OamorFYiEhIcGnrUeeuLg4HA4HKSkp5ObmkpGRQV5envfcCgoKSE5Oxmw2k5SUhMlkIj4+HoAdO3YAEB8fj8lkIikpCbPZTHJyMgUFBd7vJy3rRJXj63GEtZVR4WrpECj4fq5s0628qpSbBHSx2rBpNdxYVMzDR3IYnVdI+3Bft1Cp/oSTJFStxvta91QveH2iwpKE0uDXwDX7V5C962vb10nNN0ASTp/PGpyoaJEAFzIydjTVqjn7I6UeK5pq56epdgwJdyy7OieSfsN44gTqkfyqlYfy67S9zMzMgLM9T3+5ubmkpKQQHx/vc/zq8iQkJGCxWEhNTSU7O9vnfDIzMzEajezbtw+LxUJiYmKt9xmDwYDRaCQlJQWTydSo9z2TyXRG9z1BTSRV1LIRtEA+/vhj0tLSWLZsmc/6efPmERkZyTPPPAO4b1CLFi0iNjaW3NxcJkyYwJ49ewgPD6+z/6eeeoprr72WyZMn19g2fvx4Zs2aRUxMDADLly+npKSEefPmMWjQIGRZRqOpesDY7Xa+/PJLlixZUmefiqIQFhbGqFGjeOSRRwB3WaGPP/6YHTt2cOTIEfr378+LL77Iueee67eupk6dyvDhw7n77rt91v/444+88MIL3Hvvvd7jCQSNhd2lErLA5k6bUAGH4o4Kn4yjMr0iTHb/71LB4qR7aTmPpR4kRFFwAYUREbx7WV+eOHCYSEeVg3fe//4kovKzSyNREVrZj6oSXlRBd5MNi1aPCoQ5Hcio7swJwClJWGWde39Ao7iIcrlTL/xDRcKKDpXO/EkHTOgwoyJRQVectAHv0RRKicBGOOfyCyFYqvUCFXQinKLK3xESdjqioxgnYWgpp5RoIihGi837W0NFOqXDrwIFnEdbThBOGSoSpXSnLXloUHCiQ8aB1KktFJVBRChsewluvASe/QiWf+fu6J+T4KW7/NSHoLXy9B2/+Xx+6+tLm0mShiNyjAUtks6dO9cYWfzhhx+etRHoJ/dvNptZs2YN06ZNq7NdVFQUL774IldffTUATqeTnJwcevbsWW+fzz33HB07duT//u//GD58OOeddx6HDh1i7NixPPzwwxQWFrJ06VIWL17MO++8c0bn9+2337J06VJeeOEFRo0adUZ9BQpxcXEMGzasucUICAJRF3pZwjYnhLYLrdi9tZw8TmI11Mp/LE7QSITJKhYUzKFafurcgfPKLHQtKSXhgs44ZQ0nQkO8jrFNo+F4mIwrIpTuJWa0qOitTlyShMulcsmrA/h9/n7anbCilSTsqMi4g9SuysNqXA5kAFUhDBsqCgoab6qDCwkZpRbZ3e7p5TtG4dh/jLI/L6L0vS2EosOBDgd6tJgBiRDKUQA9TkIpopx22NAjY6vMKZbQUYEVLaDDQSigYqEzGoxoCUFLMS6gnAgk7Ci0R0Wi46QLYXOS+wfG/TfBA0Mh/U9+PZpN3/HjUCvs2JL+R0j/brS5si/W9ftRjxkJnXAxUlu9O7f458NwXlfo3tF9assegvuGuHOML+91dg2jGQjE60MQOAjHWNAiGTRoEEuXLmXLli2MHDmSvXv3sm7dOq688spTttHr3a9Zy8vL640Yjx49mtmzZzNmzBj69u3L2rVrSU9PR6fT1dlu3LhxrFmzhvPOO4/IyEjee+89du/ezb/+9a96+9TpdPTv358xY8bwyiuv8MEHH/Dvf/+bY8eOsWDBAiIjIwkJCTnjGsxJSUm88cYbvPvuu1x11VVn1FcgIR50VQSqLvSyhG1uZd7uPEuVbykBigouBXVeWB09RHr/+vqRbM6tsLKpV3cG5xXS02rDHBLC9oGXkLrOPbA0NjYWcL+R8dB/2mVn96RqQXvLZYQBrBhf537+VtOtSyOh/nQw+DKuqP75lvMAkIHwGTfV3H/QJTXXDejjz5FaBIF6fQQTagvLK66OyDEWtEgiIyNZvnw5GzduZNiwYbz//vssWbKE9u3bn7JNp06dGDRoELfddhvJycl19n/11Vczffp05syZw/Dhw/n999957bXX6pXrwQcfpH///kyZMoXhw4fzv//9j2XLlqHVav3u86mnniInJ4eNGzfyf//3f0RERPDXv/6Vm2++mdLSUp599tn6FVQHn3zyCQ6Hg+nTpxMTE+NdPDm1LRVP3qGgheii+nOzsgpEQzih1TDseBFDikyEaLVIWi1pbcLpWVrhs19LL0N4tmgRNtFECF0I6kLkGAsEgqBAURSf3O7WTEvQhTTP4p4SurqcTgX1Zb9ioEQ/foyi8HD6lVXgkCR+Dw/h3PJy/nI0l/XbrwPcEWNVVZkwYUJjnEKLoiXYRFMhdNH4TLsrw+fzuxtreQsRoAjLEAgEQUF9bwFaEy1BF4O6454S2hObUVV3OoWfdLW7cEoSqe0i+F/bcDo4XYTJWiTZN93p5DrerZWWYBNNhdCFoC5EjrGgVfL888+fMnWgW7dubNy4sYklahgjR46koqKi1m2jR49mzpw5TSxR83PppS1n1HNj0xJ08cPUMHfU2KF4S7hZ5/r/SOpmtZLpisAquytbdLPZ6Wa1c1mMb+ZuRETE2RS7xdISbKKpELpofFpyjrFwjAWtkjfeeKO5RTgjPPUpBVVkZWXRr1+/5hYjIGgpuqh7oF3daCUNN5ZWUKjVEOFy0cWlgMXCiy/4Dq5rTdPA10VLsYmmQOhCUBfCMRYIBEFBly5dmluEgKE16MKslWkjQU9FBUmDTVKZ+kTHGvvVV0mmtdAabMJfhC4EdSFyjAUCQVAgckmraA266FhejsVqxeVy4XI6oaSUW4d3q7Gfy+WqpXXrozXYhL8IXTQ+CpLP0pIQjrFAIAgKZLmWWdRaKa1BF9983Y+vV5xDibOcsCgHW/59ea37SS041/Fs0hpswl+ELgR1IVIpBAJBUBASEtLcIgQMrUUXHSJD+e9ndQ+kEo6xm9ZiE/4gdNH4tOTBdyJiLBAIggIxkUMVQhdViFQKN8ImqhC6ENSFiBgLBIKgoHv37s0tQsAQjLqQXimjg03FrtMS4nBSIoNrfrt623mmgm/tBKNNnC5CF42P0nIDxiJiLBAIgoM//vijuUUIGIJRF1qNluL2YZRH6DFGhhPiZ56oxWJpZMlaBsFoE6eL0IWgLkTEWCAQBAVXXHFFc4sQMASjLpx6LVTLW7To/SvDJib4cBOMNnG6CF00PorIMRYIBILm5ccff2xuEQKGVqELP59eojSXm1ZhE34idCGoCxExFggEQUFMTExzixAwtApd+BmRat++fSML0jJoFTbhJ0IXjY+oSiEQCATNTFxcXHOLEDAIXVRRXFzc3CIEBMImqhC6ENSFcIwFDebYsWON2r/VaqWoqKhRjyEIPoYNG9bcIgQMQhdVdOjQoblFCAiETVQhdNH4KJLv0pIQjrGgQWzYsIG33367UY/x6KOP8ttvv9W6raKigoEDB5Kbm1vr9lmzZrFq1arGFA8ARVGYOXMmGzZsqLFt3bp1zJo1q0H9ZWVlMWfOHEaMGMGQIUO499572b59+9kSt1UgokBVtApdKCrt5xvr3c3fiPEH7VewqcuHfNVmxZlKFpC0CpvwE6ELQV2IHGNBgygpKUFV1UY/RiCTl5fHwoUL2bt3LwMHDvSut1gsrF69ms8//5yhQ4f63V9mZiZTp07lkUceYe7cuYSFhbFv3z7mzp2Lw+Fg3LhxjXEaQYeIAlURbLpwKQqoQPXIkwSlLj1peXY+zdAw6WIYEC0ja3zDU/5EjD845wuulBz0KjyOAiR0XcVRm8xtB+7EVuBCHvgyOgeACn3bIc+/Fd3IS1H1GpwvxKL+Jw2ySkArQdyThMX0A0A1loOqIGWfwPZzNry+HcrNkFcBaGBgFJLeBVnFkFv9LZkOsJ1CWgmJU9+DPWevAprzOiAf/hCotAlFgbxi6NIedKd4/B8vgYgQsDlB1kBk8FX1CLbrIxBRaWFh4mqIiLGgTnJzcxkyZAjz5s1j8ODBfPDBByQkJHD//fcDkJGRwdSpUxkyZAi33347sbGxfvV78OBBpkyZwk033cTEiRP5/PPPAZgxYwb5+fnMnj2b9evXA+4I7KhRoxg+fDhffPGFTz8HDhzg/vvv58Ybb+SZZ57xGYHucrlYs2YN48eP55ZbbmH+/PmYzWYURWHs2LH88MMP3n2Tk5MZNWpUvbNkORwOJk+eTN++fWuU/JkxYwbZ2dlMnDjRLx14WLZsGbfeeiuTJ08mIiICjUbD9ddfz4wZM7xpK6tWrao1Cp2bm8vAgQOpqKggNzeXmJgYPvjgA4YOHcqoUaP48ssvvfuOHz8eg8FQo49T9d3SSExMbG4RAoZg0sVvBU60S1xup05VCXW4aGd1uAff6WX6v2flzWSVa76Anu+7KLH6Oo2lpaV19v/hxZuQ2obSy+SOPmuA7mXFhHZpyxc3/YDuypfRO9zuqISM9IcZ5d5PcUQ9j6PdTFixB7JMgITkBAa/h2XIUpzLd2PvNAOiHsRx1cvw8BfwhxHybLhjUhpILkCbmIk+txA9CnoUtIAOG3rwLjrvokWPFi1aZGR0lT2d/L/nb82RYlyS+3704+54uOEF6PkoXPwU5JyoqYxpayD6Iej0AEQ9AJ2nwNrdDf7OAp1guj4EZx/hGAvqpby8nG7durFr1y4mT57MkCFD+PTTTykuLuaJJ55g2LBh7N69m/nz57N8+XK/bjqLFy/m5ptvJj4+nsWLF7NmzRr+/PNPlixZQnR0NAsXLmTSpEn88MMPrF27lnfeeYf//Oc//Pnnn94+7HY7zz33HMOHDyc+Pp5bb72VlJQU7/Z169bx/fffs2bNGjZt2oTVamXx4sVoNBpGjRrlk6qwdetWRo8ejVzPpAGyLLNhwwaeeuoptFrfiMu8efNYsmQJHTt29Fe12O12kpOTa40wjxkzhscee8zvvsAdtf7999/ZsmULy5YtY/Xq1a3mIXDVVVc1twgBQzDpot+nuMOgGonuZVZG/57PiEMFXJdtBL3sjm5aHADkV8C6DF/HuK46xi6bi9L2bYjOK/KJweqcTlIv6EN0dgmSpKkUwBMBc/+t1lhT9Td7DuGavQkN5UgoKEScvEcl2hpxNQ1KjXVS5b+Sz19StW3UGp+TqHrIDzhsg32/uz8cPg4rT0rV+rMAVmx1/213uv93umD257X03LIJpusjUFEkyWdpSQjHWOAXo0ePRq/XExoa6l2XkJBA165dmTRpElqtln79+nHbbbf5FTXW6/UYDAYMBgM9evTg+++/p1evXjX227FjB2PGjOGCCy4gNDSUp556yrstLS0Nm83Gfffdh1ar5aabbvJJbdi8eTOPPvoo0dHRREREMH36dLZu3YrNZmPs2LHs2bMHm82Gw+EgLi6OMWPG1Cu3RqMhKiqq1m2dO3eut/3JlJaWoqrqWR0g9NxzzxEWFsYll1zCuHHjAipXef/+/eTl5ZGVlUV6ejoFBQWkp6eTn59PUlISiqKwc+dOwJ0HWFZWRnx8PIqiYDAYMJvNpKenk5WVxaFDh7z/Z2RkkJaWRmJiIg6HA4PBQFlZmTeX0KODxMRECgsLSU1NxWg0kpaWRk5ODjk5OaSlpWE0GklNTaWwsND7g8LT1iOPwWDA4XCQmJiIyWQiIyPDRxbPuZnNZgwGA4qiEB8f7yPPzp07URSFpKQk8vPzfXSRlZVFXl4e+/fvx2QykZKSgtFo9Eb7T5YnMTERi8XCvn37MBqNZGZmsnfvXnJycsjMzCQ7O5vU1FQsFgsJCQk+bT3yxMXF4XA4SElJITc3l4yMDPLy8rznVlBQQHJyMmazmaSkJEwmE/Hx8YD7GgWIj4/HZDKRlJSE2WwmOTmZgoIC7/fj6S83N5eUlBTvdVebPAkJCVgsFlJTUwnB7fQiSVx4woxc6cH2LLPSxu5CVlWolj6Rk/kLAAaDAa1WS0lJCSaTqVbb++3Ab0iqiqZcg1PSoOJOQcjq3AVF0uDQyadIWqg/ncwe4Y4Kq37uXz8q6mn0o+Ku5Zx8+IDP+qPlxT62l5GdhaqvGRhwRYY3yPYyMzMD3vZ+++03v2wvOzvb53wyMzMxGo3s27cPi8VCYmJirfcZg8GA0WgkJSXllLZ3tu57JpPpjO57gppIamMnjApaNLm5uUyYMIGEhAQiIiJYtWoVhw4dYtGiRaxdu5a0tDSWL1/u3f/rr79m69atfPjhh3X2W1JSwooVK0hMTKS4uJgRI0Ywa9YswsPDGT9+PLNmzSImJob/+7//4/rrr2fy5Mnettdeey3//ve/+eWXX1i7dq3PALhXX32Vzp0789hjjzFo0CBkWUajqfr9Z7fb+fLLL+nVqxf33XcfU6ZMQZZlVq1a5ZN24A9Tp05l+PDh3H333T7rq+uoPhwOBzfeeCOrVq2if//+PtusVisul8ur97Vr1xIaGopWq+X888/n//7v/+jcuTMTJkxgz549lJSUcOedd7J3715vH5999hn79u3j3XffZfz48ZSUlCDLMhEREVx33XU899xzfPbZZ37LG8gYjcYGReuDmWDThbTQCrLMNTnFnFvqnuLZKUkkRrWhIFQPbfSg0fDQZfDBKBmpMkIVGxuLw+GoM71p47CtaH4upJupgrAIK8ejOrD3osvo/UcuYU47o47GE+aWorKFioSKFK5FtTihlkeotGs6+nA9zme/Qv7jDzCacartKvuoHmu2oKPMJ9qroEfCiYRSo18V2RspVnH6lcWptNEil21028SybbApCa67EFY8CifPHvjNj/DKVxCiBasD2oS697vyPD+O1HIItusjEJl8/xGfz59/2nJsSAy+E/iFVMurkOjoaLZt2+azLjc3t94bjqqq/PHHHzz33HPMnTuX33//nTlz5rBx40amTJnis2/nzp3Jy8vzfjYajd484KioKAoLC3E6nd60hoKCAm/kNioqihdffJGrr74aAKfTSU5ODj179gTcqQq7du1Co9H4FS1uDHQ6HVdffTVxcXE1HONNmzbxxRdfsHnzZgAGDx7MokWLcDqdbNq0iZkzZ7J27VqfNjabjdLSUtq1awe4Bwp27drVu33BggXExMRgMpl44YUXWLNmjc9bgJbMiRMnxMOukmDThTo7FGmxg9Ru7VEkCHMqHOzUBpfNjvrPNnW2dTgcdW6/K2406zq/T4k+DEt5KNpylf6m/2Hqp+Few33AHactt/7H571/152kVUVd+0mn+Nuf/k6cOEHHV+6BV+459c4Tr3MvQU6wXR+Cs4tIpRA0CL1eT3l5OQCDBg2iqKiI9evX43Q6SU9PZ9OmTYwePbrOPiRJYtGiRXzyySc4nU46d+6MJEneGap0Op33GGPGjOG7774jPT0dm83GO++84+2nf//+tG3bltWrV+NwONi7dy/79u3zbh83bhxr1qzhxIkTOJ1O3nvvPZ5++mlvVY1Ro0aRlJTEf//7X0aNGnVW9dQQpk2bxubNm/niiy+oqKjA6XQSFxfHypUrmTp1ao0fJRqNBlmWTzlQ8N1338Vut5Oens6WLVsYO3ZsjX1kWa6zj5ZIeHh4c4sQMASrLuxameSeHTH0jqKgTQhps9vW26a+cQMA9xY+zljbA9ysTuEWdQp3FD/Aw4b7zobIAUOw2sTpIHTR+LTkOsYiYixoEDExMWzYsIGJEyfyzTff8M4777B06VJWrlxJZGQk06ZN86sUzoIFC1i4cCFffPEFOp2O0aNHM2HCBMDt0L766qvk5OTwyCOP8Mwzz/D8889TXl7OXXfdhV6vB0Cr1fLWW2/xyiuv8MUXX3DJJZdw4403eo/x4IMP4nA4mDJlCmVlZVx88cUsW7bMG13u0KEDV1xxBXa7/bTyg88WF198Me+99x6rVq3io48+wuFw0KtXL1566SVuvvlm734JCQnExMSgKAo9evTglVdeqTWSHx4ezrhx4wgNDWXGjBkMGDDAu2327NmA+8fHgAEDeOihh2qtxSwQtAR6ROrq30kgEAgagMgxFgiCBE8++J49e1plRCQzM5MLL7ywucUICIJRF/JCG4q2Mvpb+dhSZ9btGMfGxlJRUVFjHEBrJBht4nQRumh8Jj2Q5fN5/Se9m0WO00GkUggEgqDgVNVCWiPBqAutq1oMR5KQ/Izp6HQiqgzBaROni9CFoC5EKoWgURg5ciQVFRW1bhs9ejRz5sxpYon842zLfeDAAR599NFTbp8zZ069OdkC/zh69KgYUFNJMOpC53RhD6l6ZPk7s5bNdqoZ5FoXwWgTp4vQReOjtrDaxdURqRQCgSAosFgshIWFNbcYAUEw6kK70IbOoWAN0zUolcLlcnHrrbc2gYSBTTDaxOkidNH43D3lT5/PGz6uOU9BoCJSKQQCQVCwf//+5hYhYAhGXbhkjdspBvd00H7GdDwVblo7wWgTp4vQRePTkqtSCMdYIBAEBTfccENzixAwBKUuKipAUdwOsUuB8hK/mnlqerd2gtImThOhC0FdCMdYIBAEBZ5pUAXBqQv1H+1RZ+lRZ+pQn9ejzu/iV7vi4uJGlqxlEIw2cboIXTQ+iiT5LC0J4RgLBIKgwJ/62a0FoYsqOnTo0NwiBATCJqoQuhDUhXCMBQJBUCCiQFUEoy7W/eJEWmBFet2KtNCKtMhGfqmz3nYiYuwmGG3idBG6aHwUJJ+lJSEcY4FAEBSIKFAVwaaLpXsdTN7iAp0MOi1otSBp6LZK5a71dTvHImLsJths4kwQuhDUhXCMBQJBUGAwGJpbhIAh2HQxw6C4K1FUz1Ws/PzVn0qdbU0mUyNL1zIINps4E4QuGh+X5Lu0JIRjLBAIgoLrrruuuUUIGIJOF6oKmlM8XU+1vpK2bds2gkAtj6CziTNA6EJQF8IxFggEQcEvv/zS3CIEDEGnizMY1S7qGLsJOps4A4QuGh9RlUIgEAiamb59+za3CAFDUOriNB+uYoYzN0FpE6eJ0IWgLs6aY3zs2LGz1VWtWK1WioqKGvUYAoGg5ZKbm9vcIgQMQacLl+r3THcnY7fbz7IwLZOgs4kzQOii8Wn1M99t2LCBt99++2x0dUoeffRRfvvtt1q3VVRUMHDgwFMa+6xZs1i1alVjigeAoijMnDmTDRs21Ni2bt06Zs2a5XdfU6dO5frrrycmJsZnufbaa1m+fPlZlFogCA4iIyObW4SAITh1cXpPV1mWz7IcLZPgtInTQ+hCUBfas9FJSUkJ6mn+mm/IMQKZvLw8Fi5cyN69exk4cKB3vcViYfXq1Xz++ecMHTq0QX0+88wz3H333T7r5s2bdzbEFQiCDpvN1twiBAxBpwsNgMrpOMeN/WxqKQSdTZwBQheNT0urXVydBjvGubm5/O1vf2Po0KHEx8czY8YM1q5di6qq3H///Xz66adkZGSwbNkyDh48SFRUFFOmTGH8+PH19n3w4EFef/11srKy6NixIxMnTmTy5MnMmDGD/Px8Zs+ezVNPPcWkSZNYt24dn332GQ6Hg7/97W8+/Rw4cIAFCxZw+PBhBg4c6HMRuFwuPvroI7799lusVis33ngjzz33HOHh4YwfP54XXniBG2+8EYDk5GRefPFF/vOf/9QZdXA4HEyePJlbb72VsrIyn20zZswgLCyMiRMnntVC84mJiTz//POoqorVavXm0b3//vt89dVXREZG8swzzwDw8ssvs3fvXnbv3k1sbCwbN27ks88+A9zf54QJE9izZw/h4eEcOHCAt99+m0OHDmE2m+nfvz///Oc/6dSpUw0ZBg4cyPr16+nbty8lJSVMnDiRCRMmeI/r4eRjZGVlMWnSJJ599lnuvvtufvjhBxYvXozJZOKGG25g3rx56PV69u3bx/Tp09Hr9YD7lehDDz3EY489BsDXX3/NunXrKC0t5aqrrmL27NlERUV5j1c9t9BisbB+/XrCw8PrlGXz5s2sXr2akpISNBoNNpuN5557rsYPlOTkZJ5//nl2794NwE8//cQTTzzBsmXLiImJYfPmzbz//vvY7XZGjRrFjBkzkE7KkYyNjeWVV14hJCQEcL9xCA0N5bvvvmPEiBGAO4VIr9ej0Wh47LHHaN++vc/3t3nzZl555ZVazw1g/PjxzJo1i5iYGObNm+djF6c6l5aKy+VqbhEChkDVhaqqfJyuklWqct+lGvp2kCg64eCH+FIiO2i56Nq2rPkV7A6FV5OoSp/Q1n7/DS+10q2onHOmSbhkmQ4OJ20dDob9mU2oAk61I5KsJfXDn1BkGQlQrFYkrczMy4wYEwr4rTiEYkUmxO7A1DYCRZYIsTrROl04tBqUEBmNoqBVFfpl/Ukbl40yfThhTicWp5YwXGhRABtdKUQF9NjQoCJjQ0JFwYEOKxKgIiEhQeUWCbWaC+HeozHo1Si9NgMaCTq3gxsuhtIK6BIJvTtDhzZgtUObMHhyFIToTtlFoF4fgsDgtFIpysvL6datGzt27ODmm2/mwQcfZMiQIXz66acUFxfzxBNPMGzYMHbv3s38+fNZvnw5iYmJ9fa7ePFibr75ZuLj41m8eDFr1qzhzz//ZMmSJURHR7Nw4UImTZrEDz/8wNq1a3nnnXf4z3/+w59//untw26389xzzzF8+HDi4+O59dZbSUlJ8W5ft24d33//PWvWrGHTpk1YrVYWL16MRqNh1KhRbN++3bvv1q1bGT16dL2v4mRZZsOGDTz11FNotb6/NebNm8eSJUvo2LGjv+r1ixtuuAGDwcDGjRsB2L59OwaDgcsuu8xnv19//bVBDs/s2bMZPHgw27Zt4z//+Q9ms9l7jLpYuXIlVqvVr2O8+eabaDRu07Narbz44os899xzbNu2jezsbDZv3gy4f9VPmDABg8GAwWBgyJAh3j527drF2rVrWbJkCVu2bKFHjx7MmTPH5zgendRVs7K6LBUVFSxYsIB//OMf7N27F4PBQP/+/es9H5fLxdKlS9Hp3Dfi/Px8Fi1axJtvvsm///1vvv/+e/bu3Vtr24suusgr41tvvQW4Bwt51nns3mAwMHnyZJ+2ZrOZNWvW1Ctfa0GU5aoiUHWxYJ/KQ9sV/vlflUFfujhucvHGP3OI/cbIpx8WcPUHDl7eW+kUQ1Xt4pNrGAMal0L7EisVsoxLo6G7zU5bQNVqievVE1QFdHoUrRaNRuN2imUNhIWhU2F5egeO/VnOcW0IOqeKNTQEVQMaVcWh16DIGioiI7BGhFHRNgIUaKM4CcXFOVYjUc5SemIkBBcSGiRCkYEwrEAICmE4aI8GBV2lMywhoUGDhA4JLRp0J8XVGi+63XLjdyehqHDcBP/eB7t/hS8N8Po3MOtTeHk9/H0tTHmnzi4C9foIJlyS5LO0JE47x3j06NHo9XpCQ0N91ickJNC1a1cmTZqEVqulX79+3HbbbcTGxtbbp16v9zoEPXr04Pvvv6dXr5q/c3fs2MGYMWO44IILCA0N5amnnvJuS0tLw2azcd9996HVarnpppt8Uhs2b97Mo48+SnR0NBEREUyfPp2tW7dis9kYO3Yse/bswWaz4XA4iIuLY8yYMfXKrdFoiIqKqnVb586d623fWKiqyuLFi7nnnnv8bvPuu+9y1113YbVaKSgoIDIykoKCgjrbZGZmkpSU5OO4ngqDwYDZbKZfv34AhIaGsmXLFgYPHoyiKDidTu+rz5KSklOOKN+8eTP33HMPffr0ISQkhGnTppGenu7zI6mhssiyTGhoKCaTqUERha+//po+ffp4bSA6OpqdO3dyySWXoChKo0UnPvjgA0aPHt0ofTcW+/fvJy8vj6ysLNLT0ykoKCA9PZ38/HySkpJQFIWdO3cC7mlby8rKiI+PR1EU7/eVnp5OVlYWhw4d8v6fkZHBn3/+SWJiIg6HA4PBQFlZmXfqV88P3sTERAoLC0lNTcVoNJKWlkZOTg45OTmkpaVhNBpJTU2lsLDQ+2Pe09Yjj8FgwOFwkJiYiMlkIiMjw0cWz7mZzWYMBgOKohAfH+8jz86dO1EUhaSkJPLz8310kZWVRV5eHvv378dkMpGSkoLRaPT+wDtZnsTERCwWC/v27cNoNJKZmUl6ejo5OTlkZmaSnZ1NamoqFouFhIQEn7YeeeLi4nA4HKSkpJCbm0tGRgZ5eXnecysoKCA5ORmz2UxSUhImk4n4+HjAfT8GiI+Px2QykZSUhNlsJjk5mYKCAu/3k5eXx7aDVW/UCirg8+9SMZ5wz1xnkzXkOP3PBw6t8B1UJynu+4YkSYQpICGhaqoecZIn+qyRcErglGWKwtoAILsUbCHVnFRJwqGXfWokS7KEKkvolaqZ9iRAxuX9ZCcEFZkqN1TChR4JFwpytVbVe9DUsl5wJtjj0mrYXkZGBrm5uaSkpJCXl+dj+9WvhYSEBCwWC6mpqWRnZ5OZmem9ljIzMzEajezbtw+LxUJiYmKt9xmDwYDRaCQlJQWTydSo9z2TyXRG9z1BTU47x7i2V+vgnpe+W7duPuuio6PZv39/vX0uWLCAFStWsHDhQoqLixkxYgSzZs3yvhb2UFRUxIUXXuj93KVLF29Ut6ioiE6dOnmjgADdu3f3/p2fn88//vEP5s+f712n1WrJz8/n/PPP59xzz+WHH35AlmWio6O54IIL6pU7UImNjaVz585ce+21fPPNN361SU9PZ/r06VRUVNC3b19KS0vrnVJ1yZIlPP30094H5alwOBwsX76c1157jTfffNO7Pjw8nLi4OF544QVCQkK8qSzHjh2ja9eutfaVn5/PypUrfSKmkiSRl5fHueeeW+951iZLSEgIixYt4u2332bOnDmEhYVhs9kYPnz4KfspKSnh008/Ze3atTz88MM+57Ru3TrefvttoqKiuPLKK+uVqSFkZWXx/fffs379etauXet3u40bN7Jp0yZCQkK46KKLePbZZ8+qXPVx1VVX1VjXpUsXwH2fALjllluAqmlbb7rpJgBiYmIAvD9kTsZsNtOmTRuffT19jBw5EnC/aYGqH6zV3+T07NnTZ51nH09bT1+evj19tW/f/pTn69nXcw6ePjzneM011/icu0cXgPc++pe//MWnr5Pl8chx7bXXeuXv3r27VxcA55xzDoD3x6un7cn/e47luWdWv5d7ZPPI7DknT9qP57NnuycgUf2c7r5c4Yc490x157WHh269imWp2eQesxPiUrgoxMlBm3+PJWu4jjYmG7KiEuJSCEXFgdvJteKuYiEpitc5ViujVpJLQaeqyDYbXctNFIW1xanVEGpxYAnTut1TVUVrd/q0lx0utE4XChokXEiACwmH9zGqEIIVieo/hFVk7Cjo0FDhXVd9Oyi1rBecCfqJN9ClS5dar6fu3btjNpu99+STrwHPNVLX20LPtea59k6+z3iuVc+9pDHve9XlOJ37XmPR0ipRVOe0HeOT8yU9REdHs23bNp91ubm59X4Rqqryxx9/8NxzzzF37lx+//135syZw8aNG5kyZYrPvp07dyYvL8/72Wg0eqNyUVFRFBYW4nQ6vWkNBQUFXoOIiorixRdf5OqrrwbA6XSSk5PjfSiOGTOGXbt2odFo/IoWByrl5eWsWbOGlStXkp+f71eb48eP849//IMPP/zQexHOnz+/zsEru3btQqfTcdNNN9XrGH/55ZcMGDCAiy++2LtOURSKi4sZNmwYe/fu5Z///CfLli1j8eLFpKWlMXXq1Fr7ioqKYvLkyfz1r3/1rjty5Ag9e/aksLAQOLWNnkoWgD59+lBcXMxzzz3HpEmTTnl8D++//z633367zw3Y6XRSVlbGvffey+23386zzz7LmjVr+Pvf/15nXw3hzTff5Iknnmhwjda77rqLZ555BofDwerVq/nHP/7B9OnTz5pczclvv/3mdcpaO4Gqi2kDNPTtAEdMMPECiQ4RErP+0ZPkH8uI7KDlrctD+PKAil6jMmWb6k6HUKk1lUKRZQqj2/CXrCLC7U6cGg0dKyyYgWHZue4Hs8OBRpbdrqesQVZcqDY7Nq3M3PEaiiLPoSLLSSEaQkst2HUROHQyIXYXqNDGVIpeAhQFyalSpG9DmGJHiwuHRovZqqUDNnTYUbHhAOyEocWBDhcSTpzokbAgoVa6vgrgwB3TdjvFTeFDKLTwiQtCtKAooNXChd1g5FWQXww9o6B7B+jUFsptEBEKd91QZ1eBen0IAoOzUpVCr9d7ZxcaNGgQS5cuZf369dxxxx0cOHCATZs28eKLL9bZhyRJLFq0iOHDh/Pwww/TuXNnJEnyRmN0Op33GGPGjGHmzJmMHj2aCy64gHfeqcon6t+/P23btmX16tU8+uijJCUlsW/fPm/u7bhx41izZg3nnXcekZGRvPfee+zevZt//etfAIwaNYoPPvgAVVVrDFJqSWzZsoV7772Xnj17+u0YWywWwJ3eoKoqiYmJ7N69m8GDB5+yzZdffsnHH3/sV/+bNm1i/fr1PusOHjzI448/zurVq+nbt6/3e87Pzyc3N5crrrii1r7Gjh3Lp59+yoABA+jRowcbN25k5cqVxMbGYjabkSTJm/PrrywAixYtolevXkyaNKne8ykvLycpKalGPwkJCSxbtozVq1fTqVMndDrdWR0Zf/jwYS655BJGjRp12n1IkoQsy0E1CKV6ylRrJ5B1Meo8X/esTRuZm26O9H5+9Aq3m/hAtQCZtNAO2pruo6qVSevVEfsLoSdt6Q2435qpqsqECRNqlaXts549WwGKApoW7RqfNQL5+ggWXC04NeisOMYxMTFs2LCBiRMn8s033/DOO++wdOlSVq5cSWRkJNOmTfOG9utiwYIFLFy4kC+++AKdTsfo0aO9N7Rx48bx6quvkpOTwyOPPMIzzzzD888/T3l5OXfddZe3coFWq+Wtt97ilVde4YsvvuCSSy7xvpoHePDBB3E4HEyZMoWysjIuvvhili1b5o0ud+jQgSuuuAK73d6s+cFnSmRkJA899FCt2w4ePOh95eJx2B544AG++uorHnnkER5//HEURaF3795MnDiRn3766ZTH+etf/8p5553nl0xTpkypUT/ykksu4bHHHuOZZ56hoqKCCy64gJdeeomJEyficrl8StzZ7Xb27NnDNddcw9ixYyktLWX69OkYjUZ69erFW2+9xZ49e3jttdcYMWJEjYGQ9cmyb98+EhISWLdunV/n43Q6efrpp72252HYsGH89ttvPPDAAzgcDq644goefPBBv/r0B4fDwYwZM0653fP6DNyDG2fPnu394bd+/Xr+9a9/oaoq559/PnPnzqWiouJUXbUo9uzZ43392NoJOl0oCqg1o8YAjlNUrPBgMpkaS6oWRdDZxBkgdCGoC0kVRR5bPbm5uTz22GN+DZBsKgYOHEhycnKN9VOnTmXq1KniF38DGD9+PKtWrfLJtRcIWhLSQps72qmpJQqlqqgza3875Lmn+VMuVCAQnD1iHs/z+Wx4v9sp9gw8xHsVARqNhoiIiOYWw4fqA4eqEx4eXmckWFCTiIgIn8GowYpnJLYgCHUhcdqJuIE+OVRTEXQ2cQYIXQjqokkjxiNHjjzla9vRo0fXqEMbKJxtuQ8cOMCjjz56yu1z5sxpcWW4BAKBoLGQXre6J/mobUCtiBgLBAHHoCd8xzbtXRndTJI0nCYNvVWfPKMlcbblvvjii+ucdEIgEDSc+Ph4kTdYSdDp4hT5xe5tdcd2RMTYTdDZxBkgdCGoC/FOWiAQBAV1VU9pbQSdLvyf96MGddWZbk0EnU2cAUIXjU9Lm+2uOsGfeCgQCFoFtQ3WbK0Emy7Sp8iVlSkqo8Oq6l1mX1X3A7isrKzO7a2FYLOJM0HoQlAXwjEWCARBwaWXXtrcIgQMwaaLy7pqOfSoBuwOcLnA6eL5yyTUmTpev+XU9cqBgBtY3FwEm02cCUIXjY/zpKUlIRxjgUAQFGRlZTW3CAFDMOri/A5a1LnhqM+HoL4QysIxfk4dbbU2smQtg2C0idNF6EJQFyLHWCAQBAXVp+Vu7QhdVFHXDJitCWETVQhdND4tOcdYOMYCgSAoKCsrEw+8SoQuqjjVtOd3j/kvxZFR6FxO1KIytuy6pokla1qETVQhdCGoC5FKIRAIggJZPoPSBUGG0EUVUi2Rq7vHJvNn93PI69ieo507UdCnJ/u+/7UZpGs6hE1UIXTR+Dgl36UlIRxjgUAQFISEhDS3CAGD0EUVtTnGHctsWPRaQhSFUEXBLmt4b/qRZpCu6RA2UYXQhaAuhGMsEAiCAjGRQxVCF1XUlkqh1zjo4HTR3qXQzqUQ5XRhDQtuZ0nYRBVCF42PE8lnaUkIx1ggEAQF3bt3b24RAgahiyr0en2NdSfaRvgMDpIkiba28qYUq8kRNlGF0IWgLoRjLBAIgoI//vijuUUIGIJBF9LLFUjzLe7l5YrT7sdisdRYZ5N0VI8j6+1WRhQeZ9P0Tad9nEAnGGzibCF00fg4JN+lJSEcY4EgADl27Fhzi9DiuOKKK5pbhIChqXUhLXT4LGcFWQKNxr3IGpKO1HRw/aG2CT7aO1R+DdXxe6ieY1oN9+7bzm15++n70S9nKnXAIq6PKoQuBHUhHGOBIMDYsGEDb7/99lnrT1EUZs6cyYYNG2psW7duHbNmzTprx2pOfvzxx+YWIWBoSl1ICx3uJ4lWAhnOajqhU3EvqFy7Vj2tLmqbErpcligPCSE7LISDbSOYfstEAM61FJyJtAGNuD6qELpofByS5LO0JIRjLBAEGCUlJajq6TkBJ5OXl8ezzz7L999/77PeYrHw1ltvsXz58rNynEAgJiamuUUIGJpcF5rKB58kgcTZiRorNf/uM6/hKRXt27evse6G348SUsugPGNomwb331IQ10cVQheCuhCOsUDQzOTm5jJkyBDmzZvH4MGD+eCDD0hISOD+++8HICMjg6lTpzJkyBBuv/12YmNj/erX4XAwefJk+vbtW+PV4YwZM8jOzmbixIln/Xyai7i4uOYWIWDwVxcXrKpKf5j5vfPsHFwCNCAtciC9UbWM//oMnWVJ4rAqIf3DgvSyhYoK/1IriouLa6xL7doJu6KC3YnscvLwL0mkR57H/k59zkzGAEZcH1UIXTQ+jpOWloRwjAWCAKC8vJxu3bqxa9cuJk+ezJAhQ/j0008pLi7miSeeYNiwYezevZv58+ezfPlyEhMT6+1TlmU2bNjAU089hVbrO8nlvHnzWLJkCR07dmysU2pyhg0b1twiBAz+6CIuy8kf1XzGJfvOzluKU/FdQ8sEy1LtfwNI0GGJf9106NDB5/PY29L4+C+XoFS+3lVdsPGya8ns0B1jx04NFLLlIK6PKoQuBHUhHGOBIEAYPXo0er2e0NBQ77qEhAS6du3KpEmT0Gq19OvXj9tuu82vqLFGoyEqKqrWbZ07dz5rcjeE/fv3k5eXR1ZWFunp6RQUFJCenk5+fj5JSUkoisLOnTsBd1SnrKyM+Ph4FEXBYDBgNptJT08nKyuLQ4cOef/PyMhg27ZtJCYm4nA4MBgMlJWVeSND27dvByAxMZHCwkJSU1MxGo2kpaWRk5NDTk4OaWlpGI1GUlNTKSws9P748LT1yGMwGHA4HCQmJmIymcjIyPCRxXNuZrMZg8GAoijEx8f7yLNz504URSEpKYn8/HwfXWRlZZGXl8f+/fsxmUykpKRgNBoxGAy1ypOYmIjFYmHfvn0YjUYyMzPZtGkTOTk5ZGZmkp2dTWpqKhaLhYSEBG/bojIL4OsMp6SkkJubS0ZGBnl5ed5zKygoIDk5GbPZTFJSEiaTifj4+MpWKigqqJ7l1N9/RkYGubm5pKSk4HA4vPrwnIvns0dONJzyKeVSXFgsFhITE2v9rg0GA1qtlsLCQkwmk9f2dIri048CGGUNud3OpWdO4WnZnslkCnjb+/LLL5vE9jIzM+u1ver6iYuLw+FwnJbt7dixA4D4+HhMJhNJSUmYzWaSk5MpKCjwfj+e/jy2t2vXrjptz2KxkJqaSnZ2ts/5ZGZmYjQa2bdvX722ZzQaSUlJ8bG9xrjvnantNRYVkuSztCQk9WwlMwoEgtMiNzeXCRMmkJCQQEREBKtWreLQoUMsWrSItWvXkpaW5pML/PXXX7N161Y+/PBDv48xdepUhg8fzt133+2zvvqxBK2PLm87KKxM273nUlg3QdfgPqSFjpoD7jyfqz1dLo+CXx72v3/pn1bfFa5qnalwcDpcGBVWZx+eH5Djx4/3WX/dPb+zL7ryR6OsAUnil2UfYNdq+IvlGb9lFAgEtRP5dJHP55K3Ws7bGBExFggChNqmro2OjiYvL89nXW5ublClQJwt/EkvaS34q4uC6TosMyTU2brTcopPiQLqLB0xbWHvraA+r2uQU1wrlRHpI0+B+s+wep1iD6WlpTXWnWO307/CCloZJInuFVbCnU5CNcE7yYe4PqoQumh8LJLv0pIQjrFAEGDo9XrKy90P6EGDBlFUVMT69etxOp2kp6ezadMmRo8e3cxSBh5XXXVVc4sQMDREF6En5Z+fDdTZbid4z5M6brjoNB3iWt5lqv8Mp7efDrGH2uoYdy4v56qycsYVGhl5opgrKqx05SCyPngdY3F9VCF0IagL4RgLBAFGTEwMhw4dYuLEibRr14533nmHXbt2MXz4cObOncu0adPE4JFaOHjwYHOLEDA0pS48TnDVirPU8UlRpg/Gnl43tc18hyxj0slYZJkKnRZFAnNYCL9GX3h6B2kBiOujCqGLxseO5LO0JESOsUAgCAqMRqNIMakkGHQhvVzhTnUAcLlQ54c3uI/Y2FgcDkeNsoS3TTrA0Q7t+DkqEoCeZgv/iNuOM/8Yjx+feaaiByTBYBNnC6GLxkd6xujzWV3ecvR99t+hCQQCQTNw4sQJ8bCrJBh0of6z4Y5wbTgcNauolmok/mhXlWKR0yaMnSEd2XD8nrNyzEAkGGzibCF00QS0rCCxD8IxFghaKCNHjqSiovaZwEaPHs2cOXOaWKLmJTz87DhSwYDQRRWyLNdYp3E6qD7vnc6loHEqNfYLJoRNVCF0IagL4RgLBC0UT41KgUDQMDog8ZesfBL7dEenqDyc+CsmfXNLJRAEES2sdnF1hGMsEAiCglNFz1sjQhdVuFyuGutskoYLjWXM37QBraqytf8FXDCpWzNI13QIm6hC6EJQF8IxFggEQcGpZvlrjQhdVKHT1SwXt3nDJYy97w+W3XkTuBQsKOz4+8VNL1wTImyiCqELQV0Ix1ggEAQFR48eFQNqKhG6qMJms9W6/j+f9W1iSZoXYRNVCF0I6kI4xgKBICi46KKLmluEgEHoooqwsIZNCBKsCJuoQuiiCWjBOcZigg+BQBAU7N+/v7lFCBiCURdxv5SjmVOENLeYG1b6P0OdZxbJ1k4w2sTpInQhqAsxwYdAIBAIApqXvjrBW7+G0s5sw6mTKewQhuKwo77Uts52sbGxAIwfP74pxBQIBJVIM0p8PqtLIptFjtNBRIwFAkFQEBcX19wiBAzBpos3ftXhkiWOdWtHYcdw2pXZ0PqZCVhcXNzI0rUMgs0mzgShC0FdCMdYIBAEBcOGDWtuEQKGYNOFotNRERECkoQiayhtG4JWqVmGrTY6dOjQyNK1DILNJs4EoYumQDppaTkIx1ggEAQFIgpURbDp4uSMP63DhTXUvxk5RMTYTbDZxJkgdCGoC+EYCwSCoEBEgaoINl3IDidUc47toTrQ+BeFEhFjN8FmE2eC0EUT0HIDxsIxFggEwYHBYGhuEQKGYNOFI1TnW/5Jknwc5bowmUyNJFXLIths4kwQuhDUhXCMBRw/fhyn09ncYggEZ8R1113X3CIEDMGkC3OFnVC74usIqyr4WU+pbdu6K1e0FoLJJs4UoYsmQESMBS2VoqIibr/9dux2e6Me56677iIxMbFRj3G6TJ06lQ0bNpxRHx999BExMTE+y8CBA/noo4/8al9YWMhrr73GmDFjGDJkCHfeeecZy9Ta+OWXX5pbhIAhmHTRYUEFkkaqMWFAiNXhV/tT1TFe2GsDH1zwb9b1/Jr1Pb5i6WWxZyxrIBNMNnGmCF0I6kLMfNfKsdlsWK3WRj/Oxo0bG/0YzclDDz3EQw895P28efNmPv/8c+6666562xYUFHDfffcxbtw4vvjiC9q3b8///vc/Zs+ejclkYurUqY0petDQt2/rmuK3LlqELjbvg29/gkvOgfe3wZECLnv4DU6060h5hA6rVkNIUQWatiFYIkJ820oSEnD5s1kMH92Dy6NlHr7CHedxKSo7Usvoui6B7pY/KW+rIfHNDfxSFII5uh0au0rbUgvn2LWEmuw4QrQYu7YjVJJI071ON6f5JEE9DrkKKHR5sAesrRq85aq2R4MiTbIE4SHgdIKl8o3d3Nvhv5kw/i/QtzuEaMGpQHEZxKXDlhSw2KFbR/gjDxyVlTnuvAHifoGik2Wv4qqGyKZ+05C9a+JwwtafoX0EDLnszPpqBFrE9dHiaWFh4mqIiHGQkZKSwv33309MTAx33XUXP/74I/PmzWPx4sU88sgjxMTEcN9993HgwAEA7rvvPgBGjhzpXXcqYmNjvfsDVFRUMHDgQHJzc8nNzeWmm27i448/ZuTIkdxyyy0sXbrUu+/48eO9eV0HDhzg/vvv58Ybb2TatGm8/PLLrFq1CqgZvd2wYYOPY/j1119z2223MXz4cGbMmMGJEyf80svAgQP5+OOPGTFiBMOHD2fFihUoilJjv/qO7w/Hjx/nzTffZP78+bRp06be/d9//32uvPJKnnrqKSIjI5EkiX79+vHSSy95z+9k3Z98bn/88Ue953mqyHhdfbckcnNzm1uEgCHgdTH7M7j1DfgoDmZ+AoeO8+6AkVhD2tPW5qJLsRVZ0qC2CSHSUjnwrvoCWMP1ZIW145stpUzd6uTe79xO4pQNZs4f9QID3lzLgJVxhC6yklbelqI+XbFFhGHpEE5Br05kXdKN3PM6cejy7hR3aUtJ5zbEDbwGFQ34LJ53wRpAS8HaHJ9Tkavt2SBcKpRZq5xigNf+BXG/wrMfw/gFMOKfMOZVuPct+HA35JVASQVk5FQ5xQBfJdbpFNNQ+aSJDdm7Jn9d6F5uegn+GXhBkYC/PgTNinCMgwij0cjf//537rjjDuLj45k2bRozZ86krKyMLVu2MGvWLHbu3EnPnj159913Afjss88A2L59OxdffPEZHd9sNpObm0tsbCxvvvkmX3/9dY1XVjabjb///e/cdNNNxMfHM2HCBLZt2+ZX/7t27WLt2rUsWbKELVu20KNHD+bMmeO3fD/88AMbN27kk08+Yfv27XzzzRlGRU7BihUrGDVqFJdeeqlf+//3v/+tdZT0tdde26Dz89BU5xloREZGNrcIAUPA6+LT+JqrrhhWFXlVoY3FgU6FCHulAyhJ3kXjVECScOpDUCQItTn57rCK1amS9/3vXFSYB4ATHaV0wnFyxBlAI1HaMdx3lepPnEu8aK2TojJ3tNjD5wnNJ8spCPjrIxgQOcaCQOCHH36gR48eTJgwAVmWGTx4MCtXrkSv1zN48GAuvPBCQkNDGTFiBNnZ2Y0iwwMPPIBer+fyyy+nd+/eHD161Gd7WloaNpuNKVOmoNVqGTFiBAMHDvSr782bN3PPPffQp08fQkJCmDZtGunp6fz5559+tZ82bRqRkZH07NmTSZMmsWPHjgafX33k5eWxe/duHnzwQb/blJSUnNUbdVOc5+myf/9+8vLyyMrKIj09nYKCAtLT08nPzycpKQlFUdi5cyfgrjVaVlZGfHw8iqJgMBgwm82kp6eTlZXFoUOHvP9nZGRgNBpJTEzE4XBgMBgoKyvz1ivdvn07AImJiRQWFpKamorRaCQtLY2cnBxycnJIS0vDaDSSmppKYWGhNyfe09Yjj8FgwOFwkJiYiMlkIiMjw0cWz7mZzWYMBgOKohAfH+8jz86dO1EUhaSkJPLz8310kZWVRV5eHvv378dkMpGSkoLRaPS+cTlZnsTERCwWC/v27cNoNJKZmcmRI0fIyckhMzOT7OxsUlNTsVgsJCQk+LT1yBMXF4fD4SAlJYXc3FwyMjLIy8vznltBQQHJycmYzWaSkpIwmUzEx8cDeO0rPj4ek8lEUlISZrOZ5ORkCgoKvN+Pp7/c3FxMvWuWUOt//LDPeDqrXkZrcyKpEGLxHQMh4Y4cSw4nWkXFoZOJks2EaiXMPSIp17kdYS0OdNjQWfzLR+5Y5E8FCz9H/bVgVDht28ssOIbznI7evop7dwgo20tJSaG8vNzn+NXlSUhIwGKxkJqaSnZ2NpmZmd5rKTMzE6PRyL59+7BYLCQmJtZ6nzEYDBiNRlJSUjCZTI163zOZTGd03xPURFJPrpwuaLF8/PHHpKWlsWzZMp/18+bNIzIykmeeeQZwX7SLFi0iNjaW3NxcJkyYwJ49ewgPD6+l1ypiY2PZuHGjN8pcUVHB4MGD+fbbbwFq9HPfffdx1113MX78eMaPH8+sWbOoqKjgo48+8nmlv2DBAjp16sRjjz3G1KlTGT58OHfffTfgTmXYvXs3q1ev5s477yQvLw+ttipi43A4WLp0ab2jjAcOHMj27dvp1KkT4I4+v/fee3zzzTc+x6zr+P7wwQcfkJGR4ZNGUh/jxo3jySefZMyYMT7rXS4XZrOZ9u3bExsby6uvvkpYWBiyLNOjRw8effRR7yC/9evX07dv33rP83//+x86nY7Q0FD69+/PjBkz+O9//+vzvbZUDh06RJ8+fZpbjIAg4HVhtsBjK+H7dIiOhP1ZAFz2xHKKw9tQHqbDHK6nzYky0Gtx6rRUtAv1Ntc6XGhcCudnH4LL+tAjWs83EyTahWrIKVP5auWvXLtxK92tRkrDQ8lMOY+ivlEUdW2HVlFxarVoFAWXTq4a1KeqXPnbQa7OP1RPgOsIXbB4P1V/gJ5xYKxHByguh77d4LoLQacFRYGsAvgx073Nc6DqBw7Tu3OP60BtqHxnkmd85Dgs/Rbah8PsidA27PT7agQC/voIAqTny3w+q2+0nOow4p1QENG5c2cKCgp81n344Yc4HP5FS+pDo9H49HU69UG7d+9OYWEhTqfT6+AWFhZ6HTlZlk95jKioKCZPnsxf//pX77ojR47Qs2dPv45d/Tj5+flER0fX2Keu4/uDwWDgb3/7W4PaXH/99Xz//fc1HOMffviBF1980ZtqcuGFF/LZZ5+hKAqJiYnMmDGDXbt21eivrvOcPn06d999NxaLhddff50lS5YwaNCgBskbqIiyXFUEvC7ahMG6v9dY/b8aa0KQXy5BCdW6c4srnVhFA6oTMlb1q9GiZ1uJZ2ddAbOuIDY2Frvdzh233+6zz39e3kPqfywQ2dabsxxSZmFcnv9vejy0lLfEhQUFdOnSpWkOdl5XePfRpjnWaRDw14egWRGpFEHEoEGDyMvLY8uWLbhcLvbs2cO6deuoqKg4ZRu93j2t6qlKGlWnV69eHD16lCNHjmCz2fj444+RpIY9Fi699FK6du3K6tWrcTgc/Pjjjz5l3M4991z++9//YrPZyMnJYevWrd5tY8eOZd26dWRnZ6MoCuvXr2fKlClYLJbaDlWDVatWUV5ezp9//sn69etrOKL1Hb8+7HY7Bw8e5PLLL/e7DcAjjzzCzz//zIoVKzCZTLhcLn766Sdef/117r//fiIiInz2lyQJWZZRVRWXy1WjP3/OU5IktFptre1bKif/KGzNBJMulBIH4RU2n3JtKhKKzr/HV22BgbH/HMzclJHM3X2Dd5mRNPysyRyIBJNNnClCF01AC84xFhHjICIyMpLly5fz5ptv8sYbb9CjRw+WLFniTXWojU6dOjFo0CBuu+02li9fXme+b79+/bjrrrt4/PHHAXeqRLt27RokoyzLvPnmm7z22muMGDGCSy+9lCuuuMK7fcqUKcyfP58RI0bQs2dPxo4dS1JSEuB2jEtLS5k+fTpGo5FevXrx1ltv+S1D9+7dueuuu3A6ndxzzz2MGzeuxj51Hb8+PJHwqKgov/b30LVrV9auXct7773HnXfeidVqJTo6mkceeYQ77rjDu9/BgweJiYlBURQ6d+7Myy+/TPv27Rt0nsuXL+edd95Bo9Fw6aWX8uKLL7J///4GyRuo9O7du7lFCBiCSRfq250JnVtCeLkNWVGRVLCEanFo/XOMQ0ND69+pFRBMNnGmCF0I6kLkGAuanVmzZtGnTx8ee+yxRjtG9TzcYKa1nGdtJCUlcc011zS3GAFBsOlC+qcZrU6HUycDEGZxYAmRUWfp62wXGxtLaWkp9957b1OIGdAEm02cCUIXjY/0wkk5xq+3nPQVkUohEAiCAn+rm7QGgk0XssvldYoBv6PFIPJJPQSbTZwJQheCuhCpFAIvzz///Cmnbe7WrVvAzl4XHx/PSy+9dMrtb7/99hkfY9myZXXWA/aUMzqZkpISxo8ff8p2Dz74oM+MeYLTZ8+ePdx0003NLUZAEGy60LkkcLpwaauc47AKB1B3xBhOb5BwMBJsNnEmCF00BS0ssbgaIpVCIBAIBAGN9Eo5hPo6we1LKih5rWaOfXViY2MB6vxxKhAIzj7SC74zMaqv1z8LbKAgUikEAkFQ4ClqLwhCXRgt3rJqSBKy00V5qH8vPEtKShpPrhZE0NnEGSB00QS04KoUImIsEAgEgoBHeqEIqV0EkqKiulwoL9cfgRIRY4GgeZDmnBQxXiAixgKBQNCkeKaJFQSnLtTXO6G8EIprbphfTrEHETF2E4w2cboIXTQBkuS7tCCEYywQCIKCwYMHN7cIAYPQRRW11fpujQibqELoQlAXwjEWCARBQXJycnOLEDC0Jl1IL5qRXi5Herm81lnuysrKamnlxlxq4W+jf+L+MclkHTzemGI2O63JJupD6EJQF8IxFggEQcGll17a3CIEDK1FF9KcUlABlwou0L9ir7HPyVOqe6got/D0A7/TXdHQzQEL5p7gRH5RI0vcfLQWm/AHoQtBXYg6xgKBICjIysqiX79+zS1GQNCUurBaHYS9jXvkeaWTqr5Qf33hs4KsAb3W/b+qgqVmxNhqtdZYpygKC4fvZszxbJyuCOyuCE5ERjBwfic0ah59Tpxg59eXN8UZNBni+qhC6KIJaFlpxT4Ix1ggEAQFXbp0aW4RAoam1EXYWypoPINsmuywbjQat1MM7uPraz7SdDpdjXVvXPUvbvs1i3LaYwqXWHXrVfTNPYHN7sSqk9nfLbqxJW9yxPVRhdCFoC5EKoVAIAgK6solbW00qS4kfEedN6VzrCgnfa5ZfdTlctVYNyDnOFa1PbIKHcst3L17Px1NFXz90VY2rf4PMRlHG0viZkNcH1UIXTQFLbeQsYgYCwSCoECW5fp3aiU0qS6aoRK+9EKpOzqsSmB1gk7jzjO213SCpVpKRTmlMKBq386lpZx73AiAVlV56MffgL80lvjNgrg+qhC6ENSFiBgLgopjx441twiCZiIkJKS5RQgYmlQXEu78XkWtmp2uMQ83ywQ6rdsRlnA7w+UOt4Ms13SCT3aMnRYLpeESquSRVaW4eygaqcpRdoTUTL9o6Yjrowqhiyag5QaMhWMsCB42bNjA22+/fdb6UxSFmTNnsmHDBp/1//73v7ntttsYMmQI999/P/v37z9rxzyZ3NxcBg4cSEVFRYPbLl++nHnz5gHw0Ucf8dJLLwGQn5/PpEmTiImJ4ZNPPuH1118nJiaGe+6552yK3uSIiRyqaApdSK/bkRbaQaEqaqzi/tyYyJLbKa5+LEVxL7U45ienUvzScTkxx37nMvUgfTjMJfzOVScOkNa/F0WRERREtWNLzGWNfBJNj7g+qhC6ENSFSKUQBA0lJSWcrRnO8/LyWLhwIXv37mXgwIHe9cnJyaxYsYIVK1ZwwQUXsGXLFp599lk2bdpEZGTkWTl2Y/DQQw95/05JSaGiooL4+HhkWeaaa65hxYoVXH311c0o4ZnTvXv35hYhYGhsXUiv2wHJ7ZxKNa85RVX5PNXJ/VfVjLxa7Q5Sj6lc/6XvenXOWahk4azplev1eo5Jf6cNTuxoUPTn8of+PAZYDxCGu7zbnx260r7AggqUh+kJqSUvuaUjro8qhC6agBYWJa6OcIwFLZbc3Fz+9re/MXToUOLi4rBYLGg0Gu6//34+/fRTMjIyWLZsGQcPHiQqKoopU6Ywfvz4evt1OBxMnjyZW2+9tcYgjePHj3Pfffdx0UUXATBu3DiWLVvG4cOHGTBgQJ39btu2jVWrVlFSUsI555zDk08+yXXXXYeiKKxatYpdu3ZRWFhI27Zteeihh7j99tsbrI9XXnmF9PR0LrjgArp27ep9Zbhq1SoOHTrE4MGDWbBgAU6nk8GDB2O321FVlWeeeYapU6fywAMPNOiYgcQff/zBX/4SXHmhp8uZ6qLzOw5OVFD1TlGp9rdLBSSQcVejqOGLqshLnKDAAzsclZHdU/xglbxNkBbYazjHnV81c6JCrWrv+eErKe7GamU6hU4DTrdc0iwT6qKq2e4umLiXruSjaBzMHTyJ8JI2DMks5A/dOXRznuBox46kR5zP+TnuHOO2ZRZ65JxoiLpaBOL6qELoQlAXwjEWtGjKy8vp1q0bu3btYuXKlRw7doxFixZRXFzME088weOPP857773HgQMHePrpp+nUqRM33HBDnX3KssyGDRuIiopi6tSpPtvGjh3r8zk1NZXy8nLOP//8Ovu0Wq3Mnz+fjz76iEsuuYRvv/2WV199ldjYWLZu3UpcXByrVq2iU6dObNu2jVdffZXRo0c3SBezZ8+mX79+vPXWW2RkZPDUU08xbNgwn33GjRuHqqps3LiRzz77DICBAwfyySef0Ldv3wYdL9C44oormluEgOFMdXHCitsR9uTnalT332qlUyzhdorVylJt1R1ftfKzx2n2OK9Qc6BeHS94Jm+wcMKCOwqskar212sr+5RAI1eulNxPM5dSo88uWHEh87+IKxhwtII1V17EtYeN5NGZI+Hd2BZzOT0KSgCjt43GZfFbVy0FcX1UIXTRFLTckLHIMRa0eEaPHo1eryc0NNS7LiEhga5duzJp0iS0Wi39+vXjtttuIzY2tt7+NBoNUVFR9e53+PBhnn/+eR5//HG/0ij0ej3ffPMNv/zyC2PGjCE2NhZJkhgyZAgrV66kU6dOFBQUoNfrsdlslJaW1tunh5ycHH777TeefPJJ9Ho9V155JSNGjPC7fVOxf/9+8vLyyMrKIj09nYKCAtLT08nPzycpKQlFUdi5cycAcXFxlJWVER8fj6IoGAwGzGYz6enpZGVlcejQIe//GRkZ7Nmzh8TERBwOBwaDgbKyMuLi4gDYvn07AImJiRQWFpKamorRaCQtLY2cnBxycnJIS0vDaDSSmppKYWEhiYmJPm098hgMBhwOB4mJiZhMJjIyMnxk8Zyb2WzGYDCgKArx8fE+8uzcuRNFUUhKSiI/P99HF1lZWeTl5bF//35MJhMpKSkYjUYMBkOt8iQmJmKxWNi3bx9Go5HMzEy2b99OTk4OmZmZZGdnk5qaisViISEhwaetR564uDgcDgcpKSls+zm39i/v5KivorodX9dJ6z2pFbVUg6gb1StPQkICKcecNfvw1CxWcdcwdh/Id7vLHcI2GAxotVoKwjpg5Hyiy+yMOnyEp1N+4qnbb2THX85h0y39MbWP4H99u/Nnj044tBqOdWxDVpfQBtmeyWQKeNv717/+BTS+7WVmZp627eXm5pKRkUFeXp733AoKCkhOTsZsNpOUlITJZCI+Ph6AHTt2ABAfH4/JZCIpKQmz2UxycjIFBQXe78fTX25uLikpKSQmJvocv7o8CQkJWCwWUlNTyc7O9jmfzMxMjEYj+/btw2KxkJiYWOt3bTAYMBqNpKSkYDKZGvW+d6a2J6iJpJ6tpEyBoInJzc1lwoQJJCQkEBER4U0XWLRoEWvXriUtLY3ly5d79//666/ZunUrH374od/HmDp1KsOHD+fuu+/2Wf/jjz/ywgsvcO+99/LII4/41deBAwf48MMPSUpKQqvVct999zFlyhRKSkpYuHAhSUlJREdHc+GFF/Ldd9/x7bffAjBhwgT27NlDeHj4KftOS0vjqaeeYs+ePd51H374IdnZ2cybN89HN7GxsTUixuvXr2/xEWPB2UNa7PAGYqsixZV4BttV91k9EWRwO8qayn08aRSe/et52lRPpfjtuJ3L3nZURoHVqgi1Vnb3KWtq1k9WVXC6UBe0BSA2Npau9/xCT3NVzvDBjh15fsT1bPzPKr68+jaSe1UNtFOcTkySi9v++xtPZf7VP2UJBIIaSP/wfeuizg9rJkkajogYC1o8tdUpjY6OJi8vz2ddbm4uHTt2POPjffvttzz//PM8//zzfjvFZrOZ8vJyFi9ezO7du3nllVdYvXo1v/76KytWrEBRFLZu3coXX3zB448/3mCZOnfuTEVFhc9o68LCwgb305LxREkEZ64LdaaOc9vhLvXrVOnTHl6+Dh65vDJNoq5gsMbdxlvCDXc6ckOcYoBLu+o5NlPH6Itld58upaoknFZTrTxctY6dCvJJA/AqIo0UtIkA3D59pO0Eeze8Tu+yE/Qp8p3IIxSJ7qqWvQODryqFuD6qELoQ1IXIMRYEDXq9nvLycgAGDRrE0qVLWb9+PXfccQcHDhxg06ZNvPjii2d0jKSkJN544w3effddrrrqKr/bWa1Wpk+fzpIlS7j++uuJiopCkiTatWuH2WwmJCQEWZYpKSnxRrmdTidarX+XaPfu3enfvz/Lly9n9uzZHDp0iK1btzJ06NDTOc0Wycn51K2Zs6GLPx+vvZbvmlHu/6U3HL4bPA6zCuoL/lWYkBa4q0KcqiJF90g9W6b4bpNeKndHi3WVEWS7e6CfO2oNzmoD7wDK3ruJyKUHyf9FokwbQu/C42grRwyO+y2ezC69+T3qXCQkHNrK+ra64JsAQlwfVQhdCOpCRIwFQUNMTAyHDh1i4sSJtGvXjnfeeYddu3YxfPhw5s6dy7Rp0874hvjJJ5/gcDiYPn06MTEx3sWTE3gqoqKimD9/PkuXLiUmJoYZM2Ywa9YsevXqxeOPP052djbDhg3j3nvv5ZxzzqFnz54cOXKkQbItXLgQo9HILbfcwquvvtqqnGIQUaDqNIUu1Od1qM/r3M7padYxVufoT6NMm+qOHttdYHO6o9OKWlUt4ySKi4vpHz+DAcaXGFIwCxulHKcrKhDusPHizlV88uVclm5a5G1tDsKZ0cT1UYXQhaAuRI6xQCAQCE4baaG9Kr8YQFVRnz8LNYlPQV6hme7vye78YrvLt3axLKG+1sb70TPY9uQyjQe1M+kgl9DJbkSu9OoNva9g3i2PAuAA9qzu0WjnIBAEOyLHWCAQCJqZ+qL2rYkm10X1iLHauGWaunVug/qPMDaOd4Cq+DrlmpqPtNqquxR1bE8XexEmTQcAFCQWDHvQ7WxLEnKDK2oEPuL6qELooglowVNCixxjQatj5MiRp5xiefTo0cyZM6fBfR44cIBHH330lNvnzJnT4LrETdl/MNCQnO9gp8l14TMGrmleQt55VVv4l9k9EM9TW7mWiUQiIiJqrHNWZkp0VIzY0bP1vL9gk7Xe5/f/t3fn4TGd7QPHvzNZiRJFKrW2qKUtQTZLhETFFmKtV6lQaqloK1VpX4rXUmopVfVW6qdqqa21xF5CxZJEbH0VVSmtNInIKpFttt8faSZGViSZZNyf68rFzDnnOfe55wn3PPPMc5Qm+EGq/H7kkVyIokhhLJ46uWs7lqYWLVro1/osC2Xdvin47bffcHBwMHYYFUK55qLgqb3lQ6kwHCUuII6MjPw369A2q8N5WtI4JYrLdRox32Moje+lklA9Z5m3zMyssorYaOT3I4/kohxU4k9dpDAWQpiEhg0bGjuECqM8c6H7yBLFp9k5E/O0JV+RolSoNDmFce5ay1pNvl1yb4v+oK4nxvPvzgfQWmqoprPA99cI6sddYWsDT7Tm5nx/xLk8oi9X8vuRR3IhiiJzjIUQJiE+Pt7YIVQY5Z0L3UeW6KZblm9RDPz0OpCpBrXmn1Uq8hfGKpUq/4HA/JO9IMWCK1VtOGtbkz327mw63tEki2KQ348HSS5EUWTEWAhhEoq6M+DT5mnJRXeH6ugcYNXP93jHvXqB+5gVsfTap+e8yiiyiudp6RMlIbkQRZHCWAghRKVWWFEshDCSyjvFWKZSCCFMQ2ErjTyNTDUXcfGPfl0aTf7pFU8jU+0Tj0NyIYoiI8ZCCJNQu3ZtY4dQYZhaLhSz0//5gh1gloFuZslvFmBhUfCtrZ82ptYnnoTkojxU3iFjGTEWQpiEv/76y9ghVBgmlwuFIueWz1pApUMxK7XEh2Zlmd7Sa4/D5PrEE5BciKLIiLEQwiQ0b97c2CFUGCaXi3w33Cj5aFSVKpXnVrRlyeT6xBOQXJSDyjtgLCPGQgjTcOHCBWOHUGGYWi4sHiiMzXQ6lAXc4a4w9+/fL4uQKh1T6xNPQnIhiiIjxkIIk9CxY0djh1BhmFou6t7LRKMDlVJBNZWGGJuSzxuuXl1WrADT6xNPQnIhiiIjxkIIkxAcHGzsECoMU8tFtkJBbDVr7lazJqGKJWqLkn9Om5SUVIaRVR6m1ieehORCFEVGjEvRnTt3qFWrFubmklYhypuHh4exQ6gwTC0XSdYWaJU5xfA9awss1dklPrZmzZpFbl/TaBk1dVVJTVMxJtGPb17+gWv2DbDOVqPMyuI/Yd2eKPaKwtT6xJOQXJQDmWMsEhISGDRoENnZJf8HuyDe3t6EhISUUlRl1yaAVqtl2rRpbN261eD5gwcP0r9/fzp37sx7771HQkJCidu8du0a77//Pt26dcPd3Z2RI0eye/fu0g7dKG7cuIGjo6OxwzBZMgqUx9RyobYwA0sl2JiDtRnqkk8xLnLE+CvzJfSPvYpLwq9437/EHvvlpJjVoNW1WOrGpXG33nNM73a2FK7A+EytTzwJyYUoigxtlpKsrCwyMzONHUa5iYmJYeHChZw6dcqg2Pv999/59NNPWbVqFU2bNmXx4sXMmTOHL774otg2z507x9SpUxk9ejSzZs3imWee4cKFC8yfP58bN27g7+9flpdUZkJDQ/n222+JjIwEoE+fPrz00ktMmjSJZs2aGTk60yGjQHkqcy4GfBnPnltQXatAA2RbW4CNJVj989+VGZhVsSxxe4WNGB+xfJ9OVW04VtOd+xZVMdOqaZXwO1X+97d+sKvZ77Gc6NISar4OyaqCT2AG5N5DpEkNODAPXvLL2677scSxlqXK3CdKm+RCFEVGjAtw7tw53nzzTdzc3Bg6dCihoaEAzJ49m8WLFzN27Fjc3NwYOXIk165dA2DkyJEAeHl56Z8ryo4dO+jfvz/u7u5MnDiRqKgo/bazZ88ybNgw3NzcePfdd7l37x6QcwenwMBAvL29ee2115gzZw5paWklajPX//3f/9G/f39iY2NJTU3F398fDw8P+vbty9y5c0u05qdKpWLEiBE0bdqU1q1bG2w7cOAA7u7uvPLKK1hbW+Pn58eZM2dKNGq8cOFCxo4di6+vL7a2tpiZmeHo6MiKFSvYtm0bv/32G5AzAt6pUyc6d+6Mu7s7n3zyCWq1GsgpQkeOHIm7uzvDhw/n5MmT+vYdHR25ceMGABEREXh6euq3ff3113z44YcABAUF6V9PgN27dxsc+yj27dtHQEAAr7/+OitWrABg+/btuLu7M3r0aH755RcA3n77bVasWMHAgQPp0qUL06ZNIyUlhZiYGJydnblz546+zS1btjBlyhSio6NxdHQkPT1d/4nF0qVL8fb2xs3NDWdnZzp06ICbmxtLly7Nd80PKqtPFcpTZY+/NFXWXNhNjWfXHWuslRYk21ijs7SgXmoWlmrDu9eZ6+D5DxNL1GZKSkq+584rJvGC2paYqvbct6gKgEZpTpKilsEnwGZaHS9djy68KIa8ohggMsWwKAZQDCxRnGWtsvaJsiC5EEWRwvghiYmJTJ06lcGDB3P8+HEmT57MtGnTSE3NWVB+//79fPjhh/z000/Ur1+fL7/8EoANGzYAcOjQIVq0aFHkOU6fPs1XX33F/PnzCQ4OpmXLlnzyySf67b/88gurV69m7969REdHs337dgA2bdrEsWPHCAwMZNeuXWRmZrJ48eIStQk5BdXu3bv5+uuvqVu3Lhs3bkSpVHLo0CE2b97M1atXOXDgQLE5MjMzY+vWrfj5+eWbT33r1i1efPFF/WNbW1uqV6/On3/+WWSbUVFR3Lx5k9deey3ftoYNG9KmTRuOHz+uf27hwoWcPHmSH374gTNnznDmzBkiIyPx9/dn9OjRHD16lEmTJhEQEPBYBW2utLQ0AgMDH/v4tWvXMnr0aLp164alZc4oV9WqVfHx8aFPnz6sXbtWv+/+/ftZtGgRBw4cQKVSsXDhQuzt7WndujVHjhzR73fo0CF69eplcJ7vv/+e7t274+/vT1BQECEhITg4OPDee+8REhJSaUfbH4Wrq6uxQ6gwKmsu0qpYYanWkmFlDgoFNtlqtAoFygenTuh0KLPUZGlLNonxmWeeyfdcA5Rk6Wpgn5yEQqfVP2+rTUGJYRGusjB7rGsxDOJfT97GE6qsfaIsSC7KgUJh+FOJSGH8kJMnT1KvXj369euHmZkZXbp0YfXq1frbinbp0oWXXnoJa2trevTowe3btx/5HIcOHaJPnz688sormJmZMW7cOD744AP99iFDhlCzZk1q1KiBs7Mz0dHRQM7I5bhx46hbty42NjZMmTKFAwcOkJWVVWyb+/bt4/PPP2flypXUrVsXAEtLS65du8ahQ4dQqVRs3LgRHx+fYuNXKpWF3lIzMzMTa2trg+esra2LnWaSO6JcWLu1atUqcNRZrVajUql49tlnOXz4MM7Oznh4eGBubk7nzp3p0qVLiYr9wnzzzTf5itBHER0dTZMmTQrc1qRJE4P+M3ToUJo1a4aNjQ0TJ07k+PHjZGdn07NnT3766ScA/v77b27cuEHXrl31xyUkJLB3714GDRr02HGWlwsXLhATE8OtW7e4fPkycXFxXL58mdjYWMLDw9FqtfprDQ4OJjU1lePHj6PVagkJCSEtLY3Lly9z69YtIiMj9X9evXqV8PBwTp8+jUqlIiQkhNTUVP1cwkOHDgE5byDv3r3LxYsXSUxM5NKlS0RFRREVFcWlS5dITEzk4sWL3L17l9OnTxscmxtPSEgIKpWK06dPk5KSwtWrVw1iyb22tLQ0QkJC0Gq1HD9+3CCen376Ca1WS3h4OLGxsQa5uHXrFjExMVy4cIGUlBTOnTtHYmKifpTr4XhOnz5NRkYGYWFhJCYmcv36dYKDg4mKiuL69evcvn2bixcvkpGRwc8//2xwbG48wcHBqFQqzp07R3R0NFevXiUmJkZ/bXFxcURERJCWlkZ4eDgpKSn6N6qHDx8G4Pjx46SkpBAeHk5aWhoRERHExcXpX5/c9qKjozl37hwqlcrg/KmpqSh1OtRmShT/FMJZ5kqUOh0ZVR5Ynk2hIMNSiU4BGRkZnD59usDXOiQkBHNzc5KSkkhJSTHoe6AGdJhlgWfsCZqmRuIafxa7zCTsuEc1MjBHTWY1c863f+GJ+rwOoGNzo/e93O9qlHXfu379eoXvexcuXMjX93If//zzz2RkZHDx4kVu375tcD3Xr18nMTGRsLCwYvteYmIi586dy9f3SvvfvZSUlCf6d0/kp9Dp8t1S6Kn27bffcunSJT7//PN822bPno2trS3vvfcekNP5P/vsM4KCgoiOjqZfv36cOHGCqlWrFnkOPz8/XFxcGDFiRL5t3t7efPjhh7i5uQGwfPlykpOTmT17Np06dcLMzAylMu/9THZ2Nt9//z1Lliwpsk2tVkuVKlXo2bMnY8eOBXKmZnz77bccPnyYmzdv4uDgwIwZM2jYsGGJ8/X222/j6enJ66+/DsD7779PmzZt8PX11e/j6enJ0qVLcXBwKLSdqKgofHx82LVrF/Xr18+3ffTo0bi6ujJ+/Hi8vb1JTk5GqVSSnp7OCy+8wHfffcfSpUsxNzdn+vTp+uNWrlxJXFwcc+fOxdHRkS1bttC0aVMiIiKYPn06R48eBXKmUkRGRupfz23btjF37lzeffddtmzZgpubm/7YB+W+TrkxjhkzxmD7oEGD6NevH6NGjeLGjRsMGzaMiIgIAD799FNiYmL44osvePvttxk8eDA9evQAIDk5me7du3Pw4EEsLCzo2bMnP/zwAwcPHuSPP/5g3rx5+j6nUCgYO3Ys48ePL/K1iYiIYOLEidjY2KBUKrGzs2P48OH069cvX7+rjFJSUqhRo4axw6gQKmsuRgQms/0vC6wzskitYoUOqK7WcK9mFVA+MOqUmo1vYxXrRhd9jUFBQajVagYMGGDw/J1DkWT3/Yq/qU999W3siCadZ7hrY8Of1m14PiGFDEsLdni4El+jGmu2fvRkX7KvAPOMK2ufKAuSi7KnWGC4EIHu45J/L8DY5Mt3D6lTpw5xcXEGz61du5bu3buX2TlyP66fPHlykcfVrl2bGTNm4OTkBOSMlkZFRVG/fv1i2/T39+fZZ5/lnXfewdPTkxdeeIHIyEj69OnDW2+9xd27d1m6dCmLFy9m5cqVj31tL7zwgsG0ieTkZO7du0fjxo2LPK5+/fo0bdqUffv25Svwct8ZP1jwLliwADc3N9LT05k8eTLfffcddevW5X//+5/BsdHR0djZ2T3WtSxbtoyJEycWeUvZ4uaqjRs3jkWLFtGoUSOee+45ANLT0zl8+DB79+5l1apV+n3v3r2r/3tMTAxWVlbUqFEDCwsLOnToQHBwMMHBwUyYMMHgHJs3b+b999/n5ZdfpnPnzkXGU716dY4ePYpOp+PatWuMHTuWV155pchjKovo6Gj5z+4flTUXG8fZshFQ+KVTS5GNTqfjnoU5igwVuqoWOR/JqrSg0RVbFOcqaKWg57yagGop5+uvpv7fWu5Qj+f5i5vW9TjW/mXUSiWp1WzIsjDHMiMbRQUobJ9UZe0TZUFyIYoiUyke0qlTJ2JiYti/fz8ajYYTJ06wadOmYn+JcuePluT2o7169WLfvn1cu3YNtVrNunXruHz5sn66RmH69u1LYGAg8fHxqNVqvvrqK9599110Ol2xbVpYWODg4EDv3r2ZO3cuWq2WnTt3smDBAtLS0rC1tdUXYk/Cy8uL4OBgLl68SFZWFl9++SUdO3bE1ta22GM/+ugjNm/ezLfffktycjIqlYrw8HD8/f0ZMGBAgXO3zc3NqVatGhkZGfTo0YOIiAiCg4PRaDScOnWKEydO6EdhH8Uff/xBeno6PXv2fORjH9SzZ08WL17Mtm3bmDJlCgCDBw8mODiYb775xmAUfdu2bURFRZGWlsbq1avp0aOH/vXr3bs3e/fuJS4uLt/8uHr16vHBBx8wb968Et/+VqFQ6D950Gg0xexdOZSkjz0tKnsudCvrEL+4JglLnkX1aXVqpWRAWjakqSBdhZlWXeK2zMwKnyPcP2oiv9ZrBlU0RNRuxwHbFjyTeh+1hTlKbc7c4+pJJfuSX0VX2ftEaZJciKLIiPFDbG1tWb58OcuWLWPRokXUq1ePJUuWFPuLVKtWLTp16sSAAQNYvnx5kevVOjk5MWXKFD7++GMSEhJo06YN8+fPLza20aNHo1Kp8PX1JTU1lRYtWvD5559jbm5e4jb9/PwYPHgw27Zt45133mH+/Pn0798ftVpNu3btmDFjRrFxFKV58+bMmDGD//znPyQkJODg4MCsWbNKdGybNm0IDAwkMDCQjRs3kp2dTaNGjfD19c039zkgIEBf2LVs2ZJ//etf2NnZsWTJElauXMns2bOpW7cu8+bN4+WXX9YfN2rUKJRKJVqtlqysLP3UAZVKhU6n089HVqlUBnO0n4Sjo6N+VYthw4axf//+Avdr3bo1U6dOJS4ujm7dujFt2jT9Njc3N+bOnUvv3r0LvIFM165d2blzJ4GBgfqpPgVJSUnBzc0NjUZDzZo1mTBhgsksGVeSFVWeFqaWi6RnrKmu0qAy04FKQ4Z1yb8QV9xsQa+odwBoALgAnzbfwrMKUKiy+TjcC5DfD1MjuRBFkTnGQvxj9uzZtG/fHm9v73I/98PzgQsycOBA5syZw6uvvlqOkVUekZGRhX7R8Wljarkwm5+JNncdY50OReJ9tJ/aFntcUFAQGRkZDB06tGwDrARMrU88CclF2VN8+tAc449kjrEQlY61tXWx01mMISoqilOnTmFubi5FcREKWpbraWVqudBaPjBCrFCgsyr5f11FTaV4mphan3gSkgtRFCmMy8D06dP1y+08zN7enm3btpVzRI/Gy8uL9PT0Arf16tWLjz/++JHbPH78ODNnzix0+xdffEHbtm0fud3SFBAQYNTzF2bFihX88ssvLFq0yNihVGhxcXGP/UVLU2PyuXiEzzlVqiJuzvEUMfk+8QgkF+Whcq1d/CCZSiGEMAlpaWlUq1bN2GFUCKaWC8XCbDD/57viOh0kp6ObX/wXhYOCgtBoNCVan93UmVqfeBKSi7Kn+NTwDanuo4r3aWxhZFUKIYRJuHLlirFDqDBMLRdm97NBo80pijPVkJl/CbbClHSlFlNnan3iSUguyoHioZ9KRKZSCCFMQlErwTxtTC0XmuQMrNVadApQK0G9tE6Jj5X5pDlMrU88CcmFKIqMGAshTMKJEyeMHUKFYWq50K2sQ8an1clcUB31vOqPdGxKSkoZRVW5mFqfeBKSC1EUmWMshBDCJAUFBQEYZQlGIZ5mioUPzTEOkDnGQghRroKDg40dQoUhuciTnJxs7BAqBOkTeSQX5aASzzGWEWMhhBAmSUaMhTAOxaKHRoyny4ixEEKUq+PHjxs7hApDcpFHRoxzSJ/II7kQRZERYyGESdBqtSiV8l4fJBe5goKC0Ol09OvXz9ihGJ30iTySi7InI8ZCCGFkERERxg6hwpBc5ElNTTV2CBWC9Ik8kotyUInnGEthLIQwCa1atTJ2CBWG5CKPjY2NsUOoEKRP5JFciKJIYSyEMAm3bt0ydggVhuQiT2ZmprFDqBCkT+SRXIiiSGEshDAJdnZ2xg6hwpBc5LGwqDxzG8uS9Ik8kgtRFCmMhRAmQeaS5pFc5NFoNMYOoUKQPpFHclEOFArDn4eEhYUxcuRIIwRWPCmMhRAmwczMzNghVBiSizyKAv5TfhpJn8gjuRBFkcJYCGESrKysjB1ChSG5yCOFcQ7pE3kkF+XgMVel+O9//0vv3r3x9vZm4cKFaDQaJkyYwM8//wzA559/ztixYwGIi4ujb9++pR66eam3KIQQhfj9999Rq9Vl0nZSUpLczOEfkoscTZs2JSkpiatXrxo7FKOTPpFHcpHH3NycZs2alXq7ug8evbz8+eefCQ4O5scff8Tc3Bw/Pz+2bNmCu7s7oaGhuLu7c/bsWWJjY9FoNISEhNClS5dSj10KYyFEuSmLf4BzjRw5kg0bNpRZ+5WJ5CKP5CKH5CGP5KJiCg0NpU+fPlhbWwMwaNAgdu3axb///W8mTpxIWloaAM2bN+fXX3/lxIkTjBgxotTjkMJYCCGEEEIYlVarzfecWq3G3t4erVbL4cOHadeuHbVr1yY0NJRff/2Vdu3alXocMsdYCCGEEEIYlaurK/v27SMzMxO1Ws0PP/yAq6srAF26dGH16tU4Ozvj6urKhg0baNOmTZl8kVJGjIUQJmHAgAHGDqHCkFzkkVzkkDzkkVxUDBEREbRt21b/2Nvbm65duzJo0CDUajVubm76qRJdu3Zl3bp1tG/fnqpVq6JSqejatWuZxKXQ6XS6MmlZCCGEEEKISkSmUgghhBBCCIEUxkIIIYQQQgBSGAshKqnY2FjGjRvHoEGDmDp1Kunp6fn2iY+Px8/Pj+HDh/PGG29w9uxZI0Ra9kqSi1yhoaFMnDixHKMrewcPHmTIkCH4+Piwbdu2fNt/++033nzzTQYOHMjcuXPLbC3tiqC4XOSaNWsWQUFB5RhZ+SsuF8ePH2f48OH861//wt/fn3v37hkhSlHh6IQQohJ69913dQcPHtTpdDpdYGCgbsWKFfn2mTFjhm7Lli06nU6nu3nzpq5Hjx46tVpdrnGWh5LkQqPR6DZs2KDz8PDQjRs3rrxDLDN37tzReXt765KTk3Xp6em6YcOG6SIjIw32GTJkiO6XX37R6XQ63Zw5c3Tbt283RqhlriS5iIuL07333nu6jh076vbs2WOkSMtecblITU3VeXl56e7cuaPT6XS61atX6xYvXmyscEUFIiPGQohKR61Wc+HCBTw9PQHo27cvR48ezbdft27d6NWrFwANGjQgKyuLjIyMco21rJU0Fzdv3uTmzZv8+9//Lu8Qy1R4eDiOjo7UqFGDKlWq4OnpaXD9MTExZGVl8eqrrwI533w/cuSIscItU8XlAuDAgQO4u7vz2muvGSnK8lFcLtRqNQEBAdjZ2QE5d0mMjY01VriiApHCWAhR6SQnJ2NjY4O5ec6Kk7Vr1+bOnTv59vPw8KB69eoAbNiwgebNm1OtWrVyjbWslTQXTZo0YebMmfp8mIq7d+9Su3Zt/ePatWsTFxdX4u2mpCTX+uabb+Lj41POkZW/4nJha2urX+4rMzOT9evXl9nyX6JykXWMhRAV2pEjR1i2bJnBcw0bNsy3n1JZ+Pv8zZs3s3PnTr7++utSj688lUYuTI2ugBVHFQpFibebkqfpWotT0lykpaXh7+9Ps2bN6Nu3b3mEJio4KYyFEBVa9+7d6d69u8FzarUaT09PNBoNZmZmxMfHU6dOnQKPX7FiBadOnWLNmjU899xz5RFymXnSXJgiOzs7Lly4oH/88PXb2dmRkJBQ6HZTUlwuniYlyUV8fDyTJ0/GycmJqVOnlneIooJ6eoYVhBAmw9zcHAcHB3766ScA9u3bR8eOHfPtt3nzZs6dO8fatWsrfVFcmJLmwlQ5Oztz9uxZkpKSyMzMJDg4mA4dOui329vbY2lpycWLFwHTzk9xuXiaFJcLjUbDe++9R/fu3fH3939qR9ZFfjJiLISolAICApg1axZr166lbt26zJ8/H4AdO3YQHx/P+PHjCQwMxMbGhvHjx+uPW7FihcmNohWXiwkTJhg5wrJjZ2fHpEmTGD9+PGq1mv79+/PKK68wZcoUJkyYQKtWrZg3bx7z5s0jPT2d5s2bM2zYMGOHXSZKkounRXG5uHPnDr/99htarZbg4GAAWrZsycyZM40cuTA2uSW0EEIIIYQQyFQKIYQQQgghACmMhRBCCCGEAKQwFkIIIYQQApDCWAghhBBCCEAKYyGEEOKJ3b5929ghVFqllbvY2FjUanWptCWeXlIYCyGEkd28eZOJEyfi5ORE27Zt6devH9u3b9dv//HHHxk4cGC+444dO4aHh0e+54cPH46LiwtZWVkGz69cuZJWrVrRtm1b/Y+HhwerVq0q1esZOXIkGzduLNU2AVQqFW+88QZ3795lz549vPHGG6V+jsdx9OhR3n//fWOH8dhcXFwICwsrdr+AgAAWLVpUqucurdzFx8fTs2dPfZ9fvXo1P/744xO3K54+UhgLIYQRabVaxo4dyyuvvEJISAjnzp1jxowZLF68mEOHDj1ye5GRkcTGxtKqVSuCgoLybe/evTsXLlzQ/wQGBrJp0ya2bNlSGpdTpv7v//6PDh06UKdOHfr168emTZuMHRIAKSkpaLVaY4dRKZVW7jIzM8nIyNA/HjNmDGvXriUxMfGJ2xZPFymMhRDCiJKSkoiKiqJfv35YW1ujVCpxdnZm2rRpqFSqR25v69ateHp6MnDgwBIVjk2aNMHR0ZHr168bPJ+Wlkbr1q35/fff9c/98MMPDBkyBIAzZ84wbNgwXF1dadeuHVOmTDEoTHI9PHq8ceNGRo4cqX+8efNmevTogYuLC++88w53794tMM709HS+/fZbhg4dChiOoq9cuZKPP/6Y8ePH07ZtW3x8fLh06RJjx46lbdu2DBkyhJiYGCDvZigDBw6kbdu2jBo1ir///hvIeZOyfPlyevbsSdu2bXF3dzd4wxAeHs6gQYNo27Ytffr04eTJk/zyyy/MmjWLq1ev0qlTpwJj37t3L71796Z9+/YMGzaMS5cuARAVFYWjoyNr1qyhU6dOdOjQgQULFhTYRlhYGAMGDOCzzz7DycmJLl26EBwczPz583F0dMTDw4MzZ87o91+/fj2enp44OTkxZswY/vjjD/22oKAgPD09adeuHYsXLzY4T3JyMtOmTaNDhw54eHiwZs0aSnK7g8uXLzNixAjat29Pz549DUZrPTw8OHbsmP7xokWLCAgIKDB3Hh4efPnll3Tt2pX27dszc+ZM/SjwwyPWD35iMmjQIAA6d+7MlStXsLKyolu3bnz33XfFxi7Eg6QwFkIII6pVqxbOzs6MHj2aL774gtDQUNLT0xkyZAh9+/bV73ft2jUcHR0NfqZOnWrQVnZ2Nrt372bQoEF4eXkRExPDuXPnCj23RqPh/PnzhIWF4eLiYrCtWrVqeHp6sm/fPv1zQUFB+Pj4kJ6ezuTJkxk3bhyhoaHs37+fy5cvs3fv3ke69gMHDrBmzRpWrVrFiRMnaNCgQaEfqx85coQXX3wROzu7Arfv2bOHcePGER4ezjPPPMOoUaOYNGkSZ86cwdra2qBA2rVrF9OnTyc0NJSGDRvqz7lnzx4OHz7Mhg0bOH/+PP7+/ixYsID79++TkJDAhAkTGD58OBEREfj7++Pn50fjxo2ZM2cOLVu25NSpU/niCgkJ4ZNPPmHOnDmEhYUxePBg3nrrLf0bgNTUVKKiojh27BirV69m8+bNXLhwocBrvHLlCrVq1dIXyZMnT6ZevXqcOXOGPn36sGTJEiDnzdHatWtZtWoVp06dol27dowbN47MzEyuXbvGjBkzWLBgAaGhoSgUCpKTk/Xn+PDDD1EoFBw9epTvvvuOPXv2FDslITExEV9fX7y8vAgNDWXRokUsWrSIEydOFHlc69atC8zdvn372LRpE4cOHeLKlSusXLmyyHYg500bwMmTJ/V3+PPy8mLHjh3FHivEg6QwFkIII/vmm28YMWIEYWFhjB07FmdnZ6ZOnUpSUpJ+nxYtWhAREWHws2zZMoN2Dh06RKNGjWjRogWWlpYFjhoHBwfrC2tnZ2dmzpzJ22+/jZeXV764fHx8OHDgAAB3797l/Pnz9OrVCysrK3bu3ImnpyepqanExcVha2vLnTt3Hum6d+zYga+vL82aNcPKyoqpU6dy6dIlbt68mW/fiIgIWrduXWhbbdu2xdHREQsLC9q3b4+DgwPt2rXD2toaR0dHoqOj9ft6e3vj4uKClZUVH3zwAZcuXeL27dt0796d9evXU7t2be7cuYOVlRVZWVmkpKRw/PhxGjZsyKBBgzAzM8PDw4P169djaWlZ5DXu2bOHAQMG4OTkhLm5OYMHD6ZJkyYcOXJEv8+4ceOwtLTEwcGBF198kT///LPAtiwsLBg1ahRKpRJXV1eUSiVvvvkmFhYWdOzYUX+Nu3fvxtfXV98PJk2aRHZ2NuHh4Rw6dAg3NzdcXFywtLRkypQpVK1aFch5jU+cOMFHH31E1apVqV+/Pm+99ZbBfPeCHD16FHt7e0aOHImFhQVt2rRh6NCh7Ny5s8jjCjNx4kTq1atH7dq1mThxosGbs0fRokULkpOTuXXr1mMdL55O5sYOQAghnnZWVlb4+vri6+tLVlYW586dY8mSJXz88cesXr26xO1s27aN69ev6z+Wzs7OJiMjg7i4OP1Iq4eHB1988UWJ2uvcuTNpaWlcvnyZiIgIOnXqxLPPPgvkFNjr168HoHnz5mRkZJToI/cHxcTEsHz5cr788kv9cwqFgujoaF544QWDfWNjY3F1dS20LVtbW/3fzczMqF69uv6xUqk0iK1hw4b6v9eoUYOqVasSHx9PtWrVmDdvHmfOnMHe3p6WLVsCOVMsEhISqFu3rsE5iyrUcyUmJtKiRQuD555//nliY2P1j3NzCmBubl7onFsbGxvMzc3112RjY4NSqdQ/zj0uISGB559/3uD67e3tuXPnDvHx8Tz33HP6bZaWltSpUwfIeT10Oh2vvfaafrtWqzXIbWHX+OD5cq8xIiKiyOMK06hRI/3fn3vuuUKn1xTHwsICW1tbYmNjady48WO1IZ4+UhgLIYQR7d+/n6VLl3LkyBEUCgVWVlZ07NgRPz8//vOf/5S4nZs3b3Lx4kX27t2rHwEE8PPzY+vWrfj5+T1ybGZmZvTp04eDBw8SERGBr68vAOfPn2fVqlVs375dX3C8+eabBbahVCoN5ko/+LF9nTp1GDNmDIMHD9Y/FxkZSYMGDQpsR6PRFBqrQqEo8XXFxcXp/56UlER6ejp169Zl2bJl6HQ6QkJCsLKyIjo6Wj/qaWdnl29EfPXq1fTs2bPIc9nb2+vnMOeKioqiXbt2JY73UT3//PMG59RqtURHR1OrVi3s7Oz49ddf9dvUajUJCQlAzuthbm7O6dOn9SPhKSkp3L9/v8jz2dvbG4zIQ8411q5dGyi6DxTkwdcnOjoae3v7x2oHcqYL5b55EKIkpLcIIYQRdejQgfv37zN//nwSEhLQ6XT8+eefbNiwgW7dupW4nW3bttG5c2caNWpEnTp19D8DBgxg69atj/VFPsiZTrFnzx7++OMP/Red0tLSUCqVWFtbo9Fo2LVrFxEREQWuIdu4cWNCQkLIysri9u3b7NmzR79twIABrFu3jj///BOtVsuGDRsYOnRogV/iq1u37mOPHD5sz549XLlyhaysLD777DNcXFywt7cnLS0NS0tLzMzMSEpK0n/RS61W4+7uzt9//83u3bvRaDQEBwezbt06bG1tsbS05P79+wWOmPv4+LBr1y7Onj2LWq1mx44d3Lhxg+7du5fKtRTEx8eH9evX89tvv5Gdnc1XX30FgKurK7179+b06dMcO3YMlUrFqlWrSEtLA3IK3Pbt27N48WIyMzNJTk5mypQpfP7550Wez93dnbt377JhwwZUKhWXLl1i+/bteHt7Azl94NixY2g0Gq5cuUJwcLD+2IJyt2bNGhISEoiLi+O///0v/fv317cTFhZGamoqCQkJBl+MzC3kc68Fcj4xSUlJ0RfWQpSEFMZCCGFENWvWZPPmzcTFxdG3b18cHBwYPXo0r776KgEBASVqIzs7m507dxp8WS9Xr169uHfv3mMt/QbQqlUratSogZeXl7746Ny5Mz179sTb25uOHTsSFBTEgAEDiIyMzHf822+/jVqtpmPHjkyZMgUfHx/9tv79+zNkyBDGjRuHo6Mju3fv5uuvv6ZGjRr52unQoYN+NYcn1a5dO2bNmkWHDh1ISUnRz9WeMmUKf/31F05OTvj4+NCoUSMaNmxIZGQkNWvW5Ouvv2bTpk04OzuzYsUKVq1aRc2aNXFycgLAyckp39rRjo6OzJkzh1mzZuHk5MSWLVsIDAws02Ktf//+jBkzhnfeeQcXFxfCw8NZt24dVatWpUmTJixbtoyFCxfi7OxMXFycwdSFZcuWkZCQgIeHB15eXtjZ2TFr1qwiz1ejRg2++eYbDhw4gIuLC/7+/vj7+9OjRw8A/P39uXHjBk5OTixYsMBgTe6CcteiRQtef/11vL29cXV1ZcKECQAMGzaMBg0a4OHhwfDhw+nVq5e+nTp16uDu7q7/AiDA//73P+zt7Qv8BEKIwih0jzopTAghhChnGRkZeHp6snPnToM5so8qICCAmjVrMn369FKMTpQWDw8PZs6c+UiflhRm0aJFWFtb8+6775ZCZOJpISPGQgghKrwqVarg6+vL999/b+xQRCWQkZHBkSNHGDVqlLFDEZWMFMZCCCEqBV9fX8LCwkptrrEwXWvXrmX8+PHFrqghxMNkKoUQQgghhBDIiLEQQgghhBCAFMZCCCGEEEIAUhgLIYQQQggBSGEshBBCCCEEIIWxEEIIIYQQgBTGQgghhBBCAPD/beRdRNzpTskAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 576x684 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "shap_values = model.get_feature_importance(train_pool, fstr_type='ShapValues')\n",
    "shap.summary_plot(shap_values[:,:-1], X_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f742ed51",
   "metadata": {},
   "source": [
    "## LightGBM\n",
    "- Hyperparam opt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "578b2273",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:46:07.332541Z",
     "start_time": "2022-03-27T18:46:07.329030Z"
    }
   },
   "outputs": [],
   "source": [
    "RANDOM_STATE = 42\n",
    "N_SPLITS = 3\n",
    "EARLY_STOPPING_ROUNDS = 10\n",
    "NUM_BOOST_ROUND = 2000\n",
    "FOLD = StratifiedKFold(n_splits=N_SPLITS, shuffle=True, random_state=RANDOM_STATE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "3ec11124",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:46:08.913207Z",
     "start_time": "2022-03-27T18:46:08.907228Z"
    }
   },
   "outputs": [],
   "source": [
    "best = {\n",
    "    'bagging_fraction': 0.9,\n",
    "     'feature_fraction': 0.1,\n",
    "     'feature_fraction_bynode': 0.15000000000000002,\n",
    "     'lambda_l1': 7.0,\n",
    "     'lambda_l2': 13.0,\n",
    "     'learning_rate': 0.025,\n",
    "     'max_bin': 322,\n",
    "     'metric': 'auc',\n",
    "     'min_data_in_leaf': 3700,\n",
    "     'neg_bagging_fraction': 0.6000000000000001,\n",
    "     'nthread': 16,\n",
    "     'num_leaves': 106,\n",
    "     'objective': 'binary',\n",
    "     'pos_bagging_fraction': 0.65\n",
    " }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "10d29c72",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:46:13.341952Z",
     "start_time": "2022-03-27T18:46:10.817465Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "'early_stopping_rounds' argument is deprecated and will be removed in a future release of LightGBM. Pass 'early_stopping()' callback via 'callbacks' argument instead.\n",
      "'verbose_eval' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LightGBM] [Info] Number of positive: 6085, number of negative: 60094\n",
      "[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.038258 seconds.\n",
      "You can set `force_row_wise=true` to remove the overhead.\n",
      "And if memory is not enough, you can set `force_col_wise=true`.\n",
      "[LightGBM] [Info] Total Bins 49828\n",
      "[LightGBM] [Info] Number of data points in the train set: 66179, number of used features: 264\n",
      "[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.091948 -> initscore=-2.290083\n",
      "[LightGBM] [Info] Start training from score -2.290083\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n"
     ]
    }
   ],
   "source": [
    "train_data = lgb.Dataset(X_train.values, label=y_train)\n",
    "val_data = lgb.Dataset(X_val.values, label=y_val)\n",
    "\n",
    "model_lgb = lgb.train(best, train_data, valid_sets=[train_data, val_data], valid_names=['train', 'val'],\n",
    "                  num_boost_round=NUM_BOOST_ROUND, verbose_eval=False, early_stopping_rounds=EARLY_STOPPING_ROUNDS)\n",
    "\n",
    "y_pred_proba_test_lgb = model_lgb.predict(X_test.values)\n",
    "y_pred_proba_val_lgb = model_lgb.predict(X_val.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "7e8213e5",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:46:30.816111Z",
     "start_time": "2022-03-27T18:46:29.167093Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    threshold        F1       AUC\n",
      "9        0.09  0.260053  0.666291\n",
      "10       0.10  0.276645  0.659932\n",
      "8        0.08  0.232237  0.646528\n",
      "11       0.11  0.286124  0.645501\n",
      "12       0.12  0.280688  0.623025\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>threshold</th>\n",
       "      <th>F1</th>\n",
       "      <th>AUC</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.11</td>\n",
       "      <td>0.286124</td>\n",
       "      <td>0.645501</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.12</td>\n",
       "      <td>0.280688</td>\n",
       "      <td>0.623025</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.10</td>\n",
       "      <td>0.276645</td>\n",
       "      <td>0.659932</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.13</td>\n",
       "      <td>0.266163</td>\n",
       "      <td>0.600576</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.09</td>\n",
       "      <td>0.260053</td>\n",
       "      <td>0.666291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.14</td>\n",
       "      <td>0.237444</td>\n",
       "      <td>0.577271</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.08</td>\n",
       "      <td>0.232237</td>\n",
       "      <td>0.646528</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.07</td>\n",
       "      <td>0.198808</td>\n",
       "      <td>0.589707</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.15</td>\n",
       "      <td>0.190323</td>\n",
       "      <td>0.553283</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.01</td>\n",
       "      <td>0.168384</td>\n",
       "      <td>0.500000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    threshold        F1       AUC\n",
       "11       0.11  0.286124  0.645501\n",
       "12       0.12  0.280688  0.623025\n",
       "10       0.10  0.276645  0.659932\n",
       "13       0.13  0.266163  0.600576\n",
       "9        0.09  0.260053  0.666291\n",
       "14       0.14  0.237444  0.577271\n",
       "8        0.08  0.232237  0.646528\n",
       "7        0.07  0.198808  0.589707\n",
       "15       0.15  0.190323  0.553283\n",
       "1        0.01  0.168384  0.500000"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "qwe = {\n",
    "    'threshold': [],\n",
    "    'F1': [],\n",
    "    'AUC': []\n",
    "}\n",
    "\n",
    "for thr in np.arange(0, 1, 0.01):\n",
    "    qwe['threshold'].append(thr)\n",
    "    qwe['F1'].append(f1_score(y_test, (y_pred_proba_test_lgb > thr).astype(int)))\n",
    "    qwe['AUC'].append(roc_auc_score(y_test, (y_pred_proba_test_lgb > thr).astype(int)))\n",
    "\n",
    "qwe = pd.DataFrame(qwe, columns=['threshold', 'F1', 'AUC'])\n",
    "print(qwe.sort_values('AUC', ascending=False).head(5))\n",
    "qwe.sort_values('F1', ascending=False).head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "49df503d",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-27T18:46:58.717053Z",
     "start_time": "2022-03-27T18:46:58.706592Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[29761,  7033],\n",
       "       [ 1929,  1796]])"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "thr = .11\n",
    "confusion_matrix(y_test, (y_pred_proba_test_lgb > thr).astype(int))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 296,
   "id": "1475c35d",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-23T06:04:47.683585Z",
     "start_time": "2022-03-23T06:04:47.632184Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Threshold .5, f1_score:  0.0010732492621411322\n",
      "THreshold .5, ROC-AUC:  0.5002684563758389\n",
      "Threshold 0.1, f1_score: 0.2879162254382869\n",
      "THreshold 0.1, ROC-AUC: 0.6813399981686539\n"
     ]
    }
   ],
   "source": [
    "print('Threshold .5, f1_score: ', f1_score(y_test, (y_pred_proba_test_lgb > .5).astype(int)))\n",
    "print('THreshold .5, ROC-AUC: ', roc_auc_score(y_test, (y_pred_proba_test_lgb > .5).astype(int)))\n",
    "print('Threshold {}, f1_score: {}'.format(thr, f1_score(y_test,\n",
    "                                                      (y_pred_proba_test_lgb > thr).astype(int))))\n",
    "print('THreshold {}, ROC-AUC: {}'.format(thr, \n",
    "                                         roc_auc_score(y_test, (y_pred_proba_test_lgb > thr).astype(int))))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "32c5bd12",
   "metadata": {},
   "source": [
    "## Feature importance"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "94669752",
   "metadata": {},
   "source": [
    "## Merge CatBoost & LightGBM predictions"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b6e9fcd6",
   "metadata": {},
   "source": [
    "- Choose gamma & threshold on validation dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 297,
   "id": "7536ae81",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-23T06:05:21.748184Z",
     "start_time": "2022-03-23T06:05:06.635056Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "      gamma  threshold  F1_weighted  Precision_weighted        F1  Precision\n",
      "50      0.0        0.5     0.000000            0.000000  0.000000   0.000000\n",
      "150     0.1        0.5     0.005352            0.913043  0.005341   0.538462\n",
      "250     0.2        0.5     0.018963            0.868726  0.018748   0.423729\n",
      "350     0.3        0.5     0.059191            0.841121  0.056657   0.370370\n",
      "450     0.4        0.5     0.141036            0.829436  0.126375   0.350785\n",
      "550     0.5        0.5     0.259526            0.809779  0.208646   0.321116\n",
      "650     0.6        0.5     0.397282            0.780217  0.274432   0.282865\n",
      "750     0.7        0.5     0.510785            0.745706  0.301040   0.245754\n",
      "850     0.8        0.5     0.592550            0.718405  0.307171   0.220860\n",
      "950     0.9        0.5     0.640181            0.691111  0.298522   0.199104\n",
      "1050    1.0        0.5     0.675514            0.671252  0.290751   0.184919\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>gamma</th>\n",
       "      <th>threshold</th>\n",
       "      <th>F1_weighted</th>\n",
       "      <th>Precision_weighted</th>\n",
       "      <th>F1</th>\n",
       "      <th>Precision</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1041</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.41</td>\n",
       "      <td>0.701952</td>\n",
       "      <td>0.617219</td>\n",
       "      <td>0.256064</td>\n",
       "      <td>0.151940</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>419</th>\n",
       "      <td>0.4</td>\n",
       "      <td>0.19</td>\n",
       "      <td>0.701571</td>\n",
       "      <td>0.598376</td>\n",
       "      <td>0.243301</td>\n",
       "      <td>0.142031</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>213</th>\n",
       "      <td>0.2</td>\n",
       "      <td>0.13</td>\n",
       "      <td>0.701227</td>\n",
       "      <td>0.605140</td>\n",
       "      <td>0.247763</td>\n",
       "      <td>0.145506</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>731</th>\n",
       "      <td>0.7</td>\n",
       "      <td>0.31</td>\n",
       "      <td>0.701225</td>\n",
       "      <td>0.618538</td>\n",
       "      <td>0.256875</td>\n",
       "      <td>0.152661</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>524</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0.701161</td>\n",
       "      <td>0.617101</td>\n",
       "      <td>0.255877</td>\n",
       "      <td>0.151876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>625</th>\n",
       "      <td>0.6</td>\n",
       "      <td>0.25</td>\n",
       "      <td>0.701139</td>\n",
       "      <td>0.594365</td>\n",
       "      <td>0.240609</td>\n",
       "      <td>0.140013</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>522</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.22</td>\n",
       "      <td>0.701028</td>\n",
       "      <td>0.595884</td>\n",
       "      <td>0.241593</td>\n",
       "      <td>0.140774</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>316</th>\n",
       "      <td>0.3</td>\n",
       "      <td>0.16</td>\n",
       "      <td>0.701012</td>\n",
       "      <td>0.601041</td>\n",
       "      <td>0.245001</td>\n",
       "      <td>0.143390</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>730</th>\n",
       "      <td>0.7</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.700973</td>\n",
       "      <td>0.609698</td>\n",
       "      <td>0.250802</td>\n",
       "      <td>0.147898</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>729</th>\n",
       "      <td>0.7</td>\n",
       "      <td>0.29</td>\n",
       "      <td>0.700871</td>\n",
       "      <td>0.601226</td>\n",
       "      <td>0.245106</td>\n",
       "      <td>0.143484</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>937</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.37</td>\n",
       "      <td>0.700830</td>\n",
       "      <td>0.612023</td>\n",
       "      <td>0.252361</td>\n",
       "      <td>0.149135</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>214</th>\n",
       "      <td>0.2</td>\n",
       "      <td>0.14</td>\n",
       "      <td>0.700700</td>\n",
       "      <td>0.620884</td>\n",
       "      <td>0.258426</td>\n",
       "      <td>0.153953</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>832</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.32</td>\n",
       "      <td>0.700670</td>\n",
       "      <td>0.598984</td>\n",
       "      <td>0.243595</td>\n",
       "      <td>0.142340</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1040</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.40</td>\n",
       "      <td>0.700659</td>\n",
       "      <td>0.609434</td>\n",
       "      <td>0.250583</td>\n",
       "      <td>0.147759</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>627</th>\n",
       "      <td>0.6</td>\n",
       "      <td>0.27</td>\n",
       "      <td>0.700659</td>\n",
       "      <td>0.612833</td>\n",
       "      <td>0.252890</td>\n",
       "      <td>0.149569</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      gamma  threshold  F1_weighted  Precision_weighted        F1  Precision\n",
       "1041    1.0       0.41     0.701952            0.617219  0.256064   0.151940\n",
       "419     0.4       0.19     0.701571            0.598376  0.243301   0.142031\n",
       "213     0.2       0.13     0.701227            0.605140  0.247763   0.145506\n",
       "731     0.7       0.31     0.701225            0.618538  0.256875   0.152661\n",
       "524     0.5       0.24     0.701161            0.617101  0.255877   0.151876\n",
       "625     0.6       0.25     0.701139            0.594365  0.240609   0.140013\n",
       "522     0.5       0.22     0.701028            0.595884  0.241593   0.140774\n",
       "316     0.3       0.16     0.701012            0.601041  0.245001   0.143390\n",
       "730     0.7       0.30     0.700973            0.609698  0.250802   0.147898\n",
       "729     0.7       0.29     0.700871            0.601226  0.245106   0.143484\n",
       "937     0.9       0.37     0.700830            0.612023  0.252361   0.149135\n",
       "214     0.2       0.14     0.700700            0.620884  0.258426   0.153953\n",
       "832     0.8       0.32     0.700670            0.598984  0.243595   0.142340\n",
       "1040    1.0       0.40     0.700659            0.609434  0.250583   0.147759\n",
       "627     0.6       0.27     0.700659            0.612833  0.252890   0.149569"
      ]
     },
     "execution_count": 297,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "jkl = {\n",
    "    'gamma': [],\n",
    "    'threshold': [], \n",
    "    'F1_weighted': [], \n",
    "    'Precision_weighted': [], \n",
    "    'F1': [], \n",
    "    'Precision': [],\n",
    "}\n",
    "class_wei = int(y_val.value_counts()[0] / y_val.value_counts()[1])\n",
    "d = {0: 1, 1: class_wei}\n",
    "\n",
    "#y_test = pd.Series(y_test)\n",
    "weights_for_val_set = y_val.copy()\n",
    "weights_for_val_set = weights_for_val_set.map(d).values\n",
    "#gamma = .5\n",
    "for gamma in np.arange(0, 1.1, 0.1):\n",
    "    for thr in np.arange(0, 1, 0.01):\n",
    "      #model.set_probability_threshold(x)\n",
    "        jkl['gamma'].append(gamma)\n",
    "        jkl['threshold'].append(thr)\n",
    "        y_pred = ((gamma*y_pred_proba_val_cb[:,1]  + (1-gamma)*y_pred_proba_val_lgb) > thr).astype(int)\n",
    "        jkl['F1_weighted'].append(eval_metric(y_val.values, y_pred, 'F1', weight=weights_for_val_set)[0])\n",
    "        jkl['F1'].append(eval_metric(y_val.values, y_pred, 'F1')[0])\n",
    "        jkl['Precision_weighted'].append(eval_metric(y_val.values, y_pred, 'Precision', weight=weights_for_val_set)[0])\n",
    "        jkl['Precision'].append(eval_metric(y_val.values, y_pred, 'Precision')[0])\n",
    "\n",
    "jkl = pd.DataFrame(jkl)\n",
    "print(jkl[jkl.threshold == .5])\n",
    "jkl.sort_values('F1_weighted', ascending=False).head(15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 298,
   "id": "dae579dc",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-23T06:05:58.635647Z",
     "start_time": "2022-03-23T06:05:39.385038Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "      gamma  threshold  F1_weighted  Precision_weighted        F1  Precision\n",
      "50      0.0        0.5     0.001073            1.000000  0.001073   1.000000\n",
      "150     0.1        0.5     0.003216            0.915254  0.003212   0.545455\n",
      "250     0.2        0.5     0.016486            0.869159  0.016324   0.424658\n",
      "350     0.3        0.5     0.062651            0.878935  0.060561   0.446494\n",
      "450     0.4        0.5     0.157980            0.859923  0.143236   0.405507\n",
      "550     0.5        0.5     0.284458            0.826043  0.229473   0.345386\n",
      "650     0.6        0.5     0.404499            0.780240  0.277869   0.282893\n",
      "750     0.7        0.5     0.511525            0.739921  0.297538   0.240185\n",
      "850     0.8        0.5     0.589968            0.713120  0.302649   0.216422\n",
      "950     0.9        0.5     0.637377            0.688545  0.296009   0.197198\n",
      "1050    1.0        0.5     0.666905            0.666696  0.285780   0.181838\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>gamma</th>\n",
       "      <th>threshold</th>\n",
       "      <th>F1_weighted</th>\n",
       "      <th>Precision_weighted</th>\n",
       "      <th>F1</th>\n",
       "      <th>Precision</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>315</th>\n",
       "      <td>0.3</td>\n",
       "      <td>0.15</td>\n",
       "      <td>0.696475</td>\n",
       "      <td>0.586262</td>\n",
       "      <td>0.234814</td>\n",
       "      <td>0.136027</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>212</th>\n",
       "      <td>0.2</td>\n",
       "      <td>0.12</td>\n",
       "      <td>0.696386</td>\n",
       "      <td>0.588159</td>\n",
       "      <td>0.236023</td>\n",
       "      <td>0.136949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>418</th>\n",
       "      <td>0.4</td>\n",
       "      <td>0.18</td>\n",
       "      <td>0.696187</td>\n",
       "      <td>0.584857</td>\n",
       "      <td>0.233881</td>\n",
       "      <td>0.135348</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>521</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.21</td>\n",
       "      <td>0.695796</td>\n",
       "      <td>0.583810</td>\n",
       "      <td>0.233169</td>\n",
       "      <td>0.134844</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>625</th>\n",
       "      <td>0.6</td>\n",
       "      <td>0.25</td>\n",
       "      <td>0.695505</td>\n",
       "      <td>0.592359</td>\n",
       "      <td>0.238637</td>\n",
       "      <td>0.139014</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>728</th>\n",
       "      <td>0.7</td>\n",
       "      <td>0.28</td>\n",
       "      <td>0.695501</td>\n",
       "      <td>0.590637</td>\n",
       "      <td>0.237521</td>\n",
       "      <td>0.138164</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>831</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.31</td>\n",
       "      <td>0.695177</td>\n",
       "      <td>0.588999</td>\n",
       "      <td>0.236426</td>\n",
       "      <td>0.137360</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>624</th>\n",
       "      <td>0.6</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0.695123</td>\n",
       "      <td>0.582495</td>\n",
       "      <td>0.232255</td>\n",
       "      <td>0.134214</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1038</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.38</td>\n",
       "      <td>0.695074</td>\n",
       "      <td>0.593873</td>\n",
       "      <td>0.239570</td>\n",
       "      <td>0.139767</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>727</th>\n",
       "      <td>0.7</td>\n",
       "      <td>0.27</td>\n",
       "      <td>0.694952</td>\n",
       "      <td>0.581766</td>\n",
       "      <td>0.231773</td>\n",
       "      <td>0.133866</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>937</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.37</td>\n",
       "      <td>0.694845</td>\n",
       "      <td>0.609155</td>\n",
       "      <td>0.249648</td>\n",
       "      <td>0.147611</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>316</th>\n",
       "      <td>0.3</td>\n",
       "      <td>0.16</td>\n",
       "      <td>0.694830</td>\n",
       "      <td>0.598744</td>\n",
       "      <td>0.242727</td>\n",
       "      <td>0.142218</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>213</th>\n",
       "      <td>0.2</td>\n",
       "      <td>0.13</td>\n",
       "      <td>0.694791</td>\n",
       "      <td>0.602537</td>\n",
       "      <td>0.245225</td>\n",
       "      <td>0.144158</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.06</td>\n",
       "      <td>0.694713</td>\n",
       "      <td>0.593346</td>\n",
       "      <td>0.239185</td>\n",
       "      <td>0.139505</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>934</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.34</td>\n",
       "      <td>0.694706</td>\n",
       "      <td>0.587679</td>\n",
       "      <td>0.235522</td>\n",
       "      <td>0.136715</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>419</th>\n",
       "      <td>0.4</td>\n",
       "      <td>0.19</td>\n",
       "      <td>0.694486</td>\n",
       "      <td>0.595458</td>\n",
       "      <td>0.240533</td>\n",
       "      <td>0.140560</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>834</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.34</td>\n",
       "      <td>0.694407</td>\n",
       "      <td>0.611869</td>\n",
       "      <td>0.251419</td>\n",
       "      <td>0.149053</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>522</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.22</td>\n",
       "      <td>0.694390</td>\n",
       "      <td>0.593413</td>\n",
       "      <td>0.239190</td>\n",
       "      <td>0.139538</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>110</th>\n",
       "      <td>0.1</td>\n",
       "      <td>0.10</td>\n",
       "      <td>0.694327</td>\n",
       "      <td>0.607904</td>\n",
       "      <td>0.248742</td>\n",
       "      <td>0.146951</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1040</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.40</td>\n",
       "      <td>0.694267</td>\n",
       "      <td>0.606307</td>\n",
       "      <td>0.247667</td>\n",
       "      <td>0.146114</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1037</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.37</td>\n",
       "      <td>0.694263</td>\n",
       "      <td>0.586405</td>\n",
       "      <td>0.234654</td>\n",
       "      <td>0.136096</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>730</th>\n",
       "      <td>0.7</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.694260</td>\n",
       "      <td>0.606746</td>\n",
       "      <td>0.247959</td>\n",
       "      <td>0.146344</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>935</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.35</td>\n",
       "      <td>0.694225</td>\n",
       "      <td>0.594802</td>\n",
       "      <td>0.240074</td>\n",
       "      <td>0.140231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>109</th>\n",
       "      <td>0.1</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.694192</td>\n",
       "      <td>0.589405</td>\n",
       "      <td>0.236573</td>\n",
       "      <td>0.137559</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1039</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.39</td>\n",
       "      <td>0.694136</td>\n",
       "      <td>0.599974</td>\n",
       "      <td>0.243451</td>\n",
       "      <td>0.142844</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>729</th>\n",
       "      <td>0.7</td>\n",
       "      <td>0.29</td>\n",
       "      <td>0.694119</td>\n",
       "      <td>0.598110</td>\n",
       "      <td>0.242224</td>\n",
       "      <td>0.141896</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>830</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.694103</td>\n",
       "      <td>0.580820</td>\n",
       "      <td>0.231079</td>\n",
       "      <td>0.133416</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>833</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.33</td>\n",
       "      <td>0.694065</td>\n",
       "      <td>0.603921</td>\n",
       "      <td>0.246053</td>\n",
       "      <td>0.144873</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>933</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.33</td>\n",
       "      <td>0.693976</td>\n",
       "      <td>0.580278</td>\n",
       "      <td>0.230722</td>\n",
       "      <td>0.133159</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>211</th>\n",
       "      <td>0.2</td>\n",
       "      <td>0.11</td>\n",
       "      <td>0.693930</td>\n",
       "      <td>0.571456</td>\n",
       "      <td>0.225188</td>\n",
       "      <td>0.129045</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>832</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.32</td>\n",
       "      <td>0.693793</td>\n",
       "      <td>0.596092</td>\n",
       "      <td>0.240863</td>\n",
       "      <td>0.140878</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1036</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.36</td>\n",
       "      <td>0.693768</td>\n",
       "      <td>0.579744</td>\n",
       "      <td>0.230362</td>\n",
       "      <td>0.132906</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>523</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.23</td>\n",
       "      <td>0.693703</td>\n",
       "      <td>0.603081</td>\n",
       "      <td>0.245450</td>\n",
       "      <td>0.144438</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>524</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0.693681</td>\n",
       "      <td>0.613571</td>\n",
       "      <td>0.252474</td>\n",
       "      <td>0.149965</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>936</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.36</td>\n",
       "      <td>0.693680</td>\n",
       "      <td>0.601445</td>\n",
       "      <td>0.244365</td>\n",
       "      <td>0.143596</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>626</th>\n",
       "      <td>0.6</td>\n",
       "      <td>0.26</td>\n",
       "      <td>0.693667</td>\n",
       "      <td>0.600131</td>\n",
       "      <td>0.243497</td>\n",
       "      <td>0.142924</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>627</th>\n",
       "      <td>0.6</td>\n",
       "      <td>0.27</td>\n",
       "      <td>0.693545</td>\n",
       "      <td>0.609449</td>\n",
       "      <td>0.249677</td>\n",
       "      <td>0.147766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>932</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.32</td>\n",
       "      <td>0.693531</td>\n",
       "      <td>0.573181</td>\n",
       "      <td>0.226219</td>\n",
       "      <td>0.129839</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>829</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.29</td>\n",
       "      <td>0.693417</td>\n",
       "      <td>0.573025</td>\n",
       "      <td>0.226110</td>\n",
       "      <td>0.129767</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1041</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.41</td>\n",
       "      <td>0.693357</td>\n",
       "      <td>0.612905</td>\n",
       "      <td>0.251981</td>\n",
       "      <td>0.149607</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1035</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.35</td>\n",
       "      <td>0.693320</td>\n",
       "      <td>0.573236</td>\n",
       "      <td>0.226231</td>\n",
       "      <td>0.129864</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>731</th>\n",
       "      <td>0.7</td>\n",
       "      <td>0.31</td>\n",
       "      <td>0.693254</td>\n",
       "      <td>0.614817</td>\n",
       "      <td>0.253262</td>\n",
       "      <td>0.150636</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>417</th>\n",
       "      <td>0.4</td>\n",
       "      <td>0.17</td>\n",
       "      <td>0.693211</td>\n",
       "      <td>0.571947</td>\n",
       "      <td>0.225417</td>\n",
       "      <td>0.129270</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>726</th>\n",
       "      <td>0.7</td>\n",
       "      <td>0.26</td>\n",
       "      <td>0.693042</td>\n",
       "      <td>0.572627</td>\n",
       "      <td>0.225822</td>\n",
       "      <td>0.129583</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>108</th>\n",
       "      <td>0.1</td>\n",
       "      <td>0.08</td>\n",
       "      <td>0.693004</td>\n",
       "      <td>0.569866</td>\n",
       "      <td>0.224105</td>\n",
       "      <td>0.128317</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>938</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.38</td>\n",
       "      <td>0.692967</td>\n",
       "      <td>0.615656</td>\n",
       "      <td>0.253794</td>\n",
       "      <td>0.151090</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>420</th>\n",
       "      <td>0.4</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.692883</td>\n",
       "      <td>0.606145</td>\n",
       "      <td>0.247382</td>\n",
       "      <td>0.146029</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>623</th>\n",
       "      <td>0.6</td>\n",
       "      <td>0.23</td>\n",
       "      <td>0.692723</td>\n",
       "      <td>0.572193</td>\n",
       "      <td>0.225518</td>\n",
       "      <td>0.129383</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>314</th>\n",
       "      <td>0.3</td>\n",
       "      <td>0.14</td>\n",
       "      <td>0.692560</td>\n",
       "      <td>0.571174</td>\n",
       "      <td>0.224869</td>\n",
       "      <td>0.128915</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>317</th>\n",
       "      <td>0.3</td>\n",
       "      <td>0.17</td>\n",
       "      <td>0.692504</td>\n",
       "      <td>0.610475</td>\n",
       "      <td>0.250231</td>\n",
       "      <td>0.148310</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      gamma  threshold  F1_weighted  Precision_weighted        F1  Precision\n",
       "315     0.3       0.15     0.696475            0.586262  0.234814   0.136027\n",
       "212     0.2       0.12     0.696386            0.588159  0.236023   0.136949\n",
       "418     0.4       0.18     0.696187            0.584857  0.233881   0.135348\n",
       "521     0.5       0.21     0.695796            0.583810  0.233169   0.134844\n",
       "625     0.6       0.25     0.695505            0.592359  0.238637   0.139014\n",
       "728     0.7       0.28     0.695501            0.590637  0.237521   0.138164\n",
       "831     0.8       0.31     0.695177            0.588999  0.236426   0.137360\n",
       "624     0.6       0.24     0.695123            0.582495  0.232255   0.134214\n",
       "1038    1.0       0.38     0.695074            0.593873  0.239570   0.139767\n",
       "727     0.7       0.27     0.694952            0.581766  0.231773   0.133866\n",
       "937     0.9       0.37     0.694845            0.609155  0.249648   0.147611\n",
       "316     0.3       0.16     0.694830            0.598744  0.242727   0.142218\n",
       "213     0.2       0.13     0.694791            0.602537  0.245225   0.144158\n",
       "6       0.0       0.06     0.694713            0.593346  0.239185   0.139505\n",
       "934     0.9       0.34     0.694706            0.587679  0.235522   0.136715\n",
       "419     0.4       0.19     0.694486            0.595458  0.240533   0.140560\n",
       "834     0.8       0.34     0.694407            0.611869  0.251419   0.149053\n",
       "522     0.5       0.22     0.694390            0.593413  0.239190   0.139538\n",
       "110     0.1       0.10     0.694327            0.607904  0.248742   0.146951\n",
       "1040    1.0       0.40     0.694267            0.606307  0.247667   0.146114\n",
       "1037    1.0       0.37     0.694263            0.586405  0.234654   0.136096\n",
       "730     0.7       0.30     0.694260            0.606746  0.247959   0.146344\n",
       "935     0.9       0.35     0.694225            0.594802  0.240074   0.140231\n",
       "109     0.1       0.09     0.694192            0.589405  0.236573   0.137559\n",
       "1039    1.0       0.39     0.694136            0.599974  0.243451   0.142844\n",
       "729     0.7       0.29     0.694119            0.598110  0.242224   0.141896\n",
       "830     0.8       0.30     0.694103            0.580820  0.231079   0.133416\n",
       "833     0.8       0.33     0.694065            0.603921  0.246053   0.144873\n",
       "933     0.9       0.33     0.693976            0.580278  0.230722   0.133159\n",
       "211     0.2       0.11     0.693930            0.571456  0.225188   0.129045\n",
       "832     0.8       0.32     0.693793            0.596092  0.240863   0.140878\n",
       "1036    1.0       0.36     0.693768            0.579744  0.230362   0.132906\n",
       "523     0.5       0.23     0.693703            0.603081  0.245450   0.144438\n",
       "524     0.5       0.24     0.693681            0.613571  0.252474   0.149965\n",
       "936     0.9       0.36     0.693680            0.601445  0.244365   0.143596\n",
       "626     0.6       0.26     0.693667            0.600131  0.243497   0.142924\n",
       "627     0.6       0.27     0.693545            0.609449  0.249677   0.147766\n",
       "932     0.9       0.32     0.693531            0.573181  0.226219   0.129839\n",
       "829     0.8       0.29     0.693417            0.573025  0.226110   0.129767\n",
       "1041    1.0       0.41     0.693357            0.612905  0.251981   0.149607\n",
       "1035    1.0       0.35     0.693320            0.573236  0.226231   0.129864\n",
       "731     0.7       0.31     0.693254            0.614817  0.253262   0.150636\n",
       "417     0.4       0.17     0.693211            0.571947  0.225417   0.129270\n",
       "726     0.7       0.26     0.693042            0.572627  0.225822   0.129583\n",
       "108     0.1       0.08     0.693004            0.569866  0.224105   0.128317\n",
       "938     0.9       0.38     0.692967            0.615656  0.253794   0.151090\n",
       "420     0.4       0.20     0.692883            0.606145  0.247382   0.146029\n",
       "623     0.6       0.23     0.692723            0.572193  0.225518   0.129383\n",
       "314     0.3       0.14     0.692560            0.571174  0.224869   0.128915\n",
       "317     0.3       0.17     0.692504            0.610475  0.250231   0.148310"
      ]
     },
     "execution_count": 298,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "jkl = {\n",
    "    'gamma': [],\n",
    "    'threshold': [], \n",
    "    'F1_weighted': [], \n",
    "    'Precision_weighted': [], \n",
    "    'F1': [], \n",
    "    'Precision': [],\n",
    "}\n",
    "class_wei = int(y_test.value_counts()[0] / y_test.value_counts()[1])\n",
    "d = {0: 1, 1: class_wei}\n",
    "\n",
    "#y_test = pd.Series(y_test)\n",
    "weights_for_test_set = y_test.copy()\n",
    "weights_for_test_set = weights_for_test_set.map(d).values\n",
    "#gamma = .5\n",
    "for gamma in np.arange(0, 1.1, 0.1):\n",
    "    for thr in np.arange(0, 1, 0.01):\n",
    "        jkl['gamma'].append(gamma)\n",
    "        jkl['threshold'].append(thr)\n",
    "        y_pred = ((gamma*y_pred_proba_test_cb[:,1]  + (1-gamma)*y_pred_proba_test_lgb) > thr).astype(int)\n",
    "        jkl['F1_weighted'].append(eval_metric(y_test.values, y_pred, 'F1', weight=weights_for_test_set)[0])\n",
    "        jkl['F1'].append(eval_metric(y_test.values, y_pred, 'F1')[0])\n",
    "        jkl['Precision_weighted'].append(eval_metric(y_test.values, y_pred, 'Precision', weight=weights_for_test_set)[0])\n",
    "        jkl['Precision'].append(eval_metric(y_test.values, y_pred, 'Precision')[0])\n",
    "\n",
    "jkl = pd.DataFrame(jkl)\n",
    "print(jkl[jkl.threshold == .5])\n",
    "jkl.sort_values('F1_weighted', ascending=False).head(50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 321,
   "id": "655cbc56",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-23T06:08:35.439694Z",
     "start_time": "2022-03-23T06:08:35.435493Z"
    }
   },
   "outputs": [],
   "source": [
    "gamma, thr = 0.8, 0.35\n",
    "y_pred = ((gamma*y_pred_proba_test_cb[:,1]  + (1-gamma)*y_pred_proba_test_lgb) > thr).astype(int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 322,
   "id": "b225c010",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-23T06:08:35.643659Z",
     "start_time": "2022-03-23T06:08:35.631398Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[20551, 16243],\n",
       "       [  804,  2921]])"
      ]
     },
     "execution_count": 322,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 334,
   "id": "79fe4f36",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-23T06:13:31.380959Z",
     "start_time": "2022-03-23T06:13:31.234118Z"
    }
   },
   "outputs": [],
   "source": [
    "pd.DataFrame(data={'client_id':X_test.index.values,\n",
    "                   'score':(gamma*y_pred_proba_test_cb[:,1]  \n",
    "                            + (1-gamma)*y_pred_proba_test_lgb)}).to_csv('fin_fin.csv', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "07a3dd4c",
   "metadata": {},
   "source": [
    "## Power Transformer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "id": "02e513a8",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T18:43:13.391341Z",
     "start_time": "2022-03-22T18:42:52.439992Z"
    }
   },
   "outputs": [],
   "source": [
    "PT = PowerTransformer()\n",
    "\n",
    "x = PT.fit_transform(df.drop(['client_id', 'is_in_club'], axis=1))\n",
    "x = pd.DataFrame(data=x, columns=df.columns[2:])\n",
    "y = df.is_in_club"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "id": "554cf3ce",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T18:43:13.594055Z",
     "start_time": "2022-03-22T18:43:13.394069Z"
    }
   },
   "outputs": [],
   "source": [
    "TEST_SIZE = .3\n",
    "RANDOM_STATE = 42\n",
    "\n",
    "X_train_val, X_test, y_train_val, y_test = train_test_split(x, y, test_size=TEST_SIZE, \n",
    "                                                            shuffle=False, random_state=RANDOM_STATE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "id": "347a36ab",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T18:43:20.665985Z",
     "start_time": "2022-03-22T18:43:20.451194Z"
    }
   },
   "outputs": [],
   "source": [
    "X_train, X_val, y_train, y_val = train_test_split(X_train_val, \n",
    "                                                  y_train_val, \n",
    "                                                  test_size=TEST_SIZE, \n",
    "                                                  shuffle=True, \n",
    "                                                  stratify=y_train_val,\n",
    "                                                  random_state=RANDOM_STATE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "id": "047d3fca",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T18:43:21.763089Z",
     "start_time": "2022-03-22T18:43:21.758191Z"
    }
   },
   "outputs": [],
   "source": [
    "best_patams = {\n",
    "    'boosting_type': 'Ordered',\n",
    "    'bootstrap_type': 'Bernoulli',\n",
    "    'colsample_bylevel': 0.08512900052697547,\n",
    "    'depth': 4,\n",
    "    'l2_leaf_reg': 0.04335660768413505,\n",
    "    'learning_rate': 0.005965363852834718,\n",
    "    'max_ctr_complexity': 2,\n",
    "    'min_data_in_leaf': 16,\n",
    "    'n_estimators': 361,\n",
    "    'one_hot_max_size': 9,\n",
    "    'subsample': 0.9989517974264015\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "id": "8a653641",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T18:43:23.059497Z",
     "start_time": "2022-03-22T18:43:22.959844Z"
    }
   },
   "outputs": [],
   "source": [
    "train_pool = Pool(X_train, y_train)\n",
    "val_pool = Pool(X_val,y_val)\n",
    "\n",
    "class_wei = y_train.value_counts()[0] / y_train.value_counts()[1]\n",
    "model = CatBoostClassifier(\n",
    "    **best_params,\n",
    "    class_weights = (1, class_wei),\n",
    "    custom_metric='F1',\n",
    "    early_stopping_rounds=10,\n",
    "    use_best_model=True,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "id": "ab8c39c2",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T18:43:37.473994Z",
     "start_time": "2022-03-22T18:43:24.381761Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "be48a50705bd4a6c8b0f2c3cf0cfe9ff",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "MetricVisualizer(layout=Layout(align_self='stretch', height='500px'))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.fit(\n",
    "    train_pool,\n",
    "    eval_set=val_pool,\n",
    "    verbose=False,\n",
    "    plot=True\n",
    ");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "id": "6ce12052",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T18:43:37.632603Z",
     "start_time": "2022-03-22T18:43:37.476635Z"
    }
   },
   "outputs": [],
   "source": [
    "y_pred_test_pt_cb = model.predict(X_test)\n",
    "y_pred_proba_test_pt_cb = model.predict_proba(X_test)\n",
    "y_pred_proba_val_pt_cb = model.predict_proba(X_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "id": "e49dcfa0",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T18:43:37.647500Z",
     "start_time": "2022-03-22T18:43:37.636731Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[28825,  8793],\n",
       "       [ 1275,  1626]])"
      ]
     },
     "execution_count": 200,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(y_test, y_pred_test_pt_cb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 201,
   "id": "b26e82ba",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T18:46:20.195609Z",
     "start_time": "2022-03-22T18:46:20.176710Z"
    }
   },
   "outputs": [],
   "source": [
    "ANDOM_STATE = 42\n",
    "TEST_SIZE = 0.25\n",
    "N_SPLITS = 3\n",
    "EVALS = 50# \n",
    "EARLY_STOPPING_ROUNDS = 10\n",
    "NTHREAD = 16\n",
    "NUM_BOOST_ROUND = 2000\n",
    "FOLD = StratifiedKFold(n_splits=N_SPLITS, shuffle=True, random_state=RANDOM_STATE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "id": "b913bf52",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T18:46:20.873619Z",
     "start_time": "2022-03-22T18:46:20.862132Z"
    }
   },
   "outputs": [],
   "source": [
    "best = {\n",
    "    'bagging_fraction': 0.9,\n",
    "     'feature_fraction': 0.1,\n",
    "     'feature_fraction_bynode': 0.15000000000000002,\n",
    "     'lambda_l1': 7.0,\n",
    "     'lambda_l2': 13.0,\n",
    "     'learning_rate': 0.025,\n",
    "     'max_bin': 322,\n",
    "     'metric': 'auc',\n",
    "     'min_data_in_leaf': 3700,\n",
    "     'neg_bagging_fraction': 0.6000000000000001,\n",
    "     'nthread': 16,\n",
    "     'num_leaves': 106,\n",
    "     'objective': 'binary',\n",
    "     'pos_bagging_fraction': 0.65\n",
    " }"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "id": "1e6a4301",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T18:46:48.326550Z",
     "start_time": "2022-03-22T18:46:26.946937Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "'early_stopping_rounds' argument is deprecated and will be removed in a future release of LightGBM. Pass 'early_stopping()' callback via 'callbacks' argument instead.\n",
      "'verbose_eval' argument is deprecated and will be removed in a future release of LightGBM. Pass 'log_evaluation()' callback via 'callbacks' argument instead.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LightGBM] [Warning] Found whitespace in feature_names, replace with underlines\n",
      "[LightGBM] [Info] Number of positive: 6662, number of negative: 59517\n",
      "[LightGBM] [Warning] Auto-choosing row-wise multi-threading, the overhead of testing was 0.011120 seconds.\n",
      "You can set `force_row_wise=true` to remove the overhead.\n",
      "And if memory is not enough, you can set `force_col_wise=true`.\n",
      "[LightGBM] [Info] Total Bins 44371\n",
      "[LightGBM] [Info] Number of data points in the train set: 66179, number of used features: 189\n",
      "[LightGBM] [Warning] Found whitespace in feature_names, replace with underlines\n",
      "[LightGBM] [Info] [binary:BoostFromScore]: pavg=0.100666 -> initscore=-2.189842\n",
      "[LightGBM] [Info] Start training from score -2.189842\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n",
      "[LightGBM] [Warning] No further splits with positive gain, best gain: -inf\n"
     ]
    }
   ],
   "source": [
    "train_data = lgb.Dataset(X_train, label=y_train)\n",
    "val_data = lgb.Dataset(X_val, label=y_val)\n",
    "\n",
    "model_lgb = lgb.train(best, train_data, valid_sets=[train_data, val_data], valid_names=['train', 'val'],\n",
    "                  num_boost_round=NUM_BOOST_ROUND, verbose_eval=False, early_stopping_rounds=EARLY_STOPPING_ROUNDS)\n",
    "\n",
    "y_pred_proba_test_lgb = model_lgb.predict(X_test)\n",
    "y_pred_proba_val_lgb = model_lgb.predict(X_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "id": "8d4c51ad",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T19:07:53.963721Z",
     "start_time": "2022-03-22T19:07:52.642649Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    threshold                    F1                   AUC\n",
      "8        0.08  [0.6728781975266814]  [0.6802456776753719]\n",
      "9        0.09  [0.6560315787911614]   [0.678301337403495]\n",
      "7        0.07  [0.6831088973322772]  [0.6762397743575455]\n",
      "10       0.10  [0.6396179904187531]  [0.6758427792851263]\n",
      "6        0.06  [0.6930640446452299]  [0.6696564224087682]\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>threshold</th>\n",
       "      <th>F1</th>\n",
       "      <th>AUC</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.05</td>\n",
       "      <td>[0.6992470521380878]</td>\n",
       "      <td>[0.6566857465115539]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.04</td>\n",
       "      <td>[0.6953470265053899]</td>\n",
       "      <td>[0.6274188095869453]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.06</td>\n",
       "      <td>[0.6930640446452299]</td>\n",
       "      <td>[0.6696564224087682]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.07</td>\n",
       "      <td>[0.6831088973322772]</td>\n",
       "      <td>[0.6762397743575455]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.03</td>\n",
       "      <td>[0.679543830335484]</td>\n",
       "      <td>[0.5757068659273307]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.08</td>\n",
       "      <td>[0.6728781975266814]</td>\n",
       "      <td>[0.6802456776753719]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.09</td>\n",
       "      <td>[0.6560315787911614]</td>\n",
       "      <td>[0.678301337403495]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.02</td>\n",
       "      <td>[0.6526258462520774]</td>\n",
       "      <td>[0.5082163291063126]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.00</td>\n",
       "      <td>[0.6492232520840716]</td>\n",
       "      <td>[0.5]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.01</td>\n",
       "      <td>[0.6492232520840716]</td>\n",
       "      <td>[0.5]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   threshold                    F1                   AUC\n",
       "5       0.05  [0.6992470521380878]  [0.6566857465115539]\n",
       "4       0.04  [0.6953470265053899]  [0.6274188095869453]\n",
       "6       0.06  [0.6930640446452299]  [0.6696564224087682]\n",
       "7       0.07  [0.6831088973322772]  [0.6762397743575455]\n",
       "3       0.03   [0.679543830335484]  [0.5757068659273307]\n",
       "8       0.08  [0.6728781975266814]  [0.6802456776753719]\n",
       "9       0.09  [0.6560315787911614]   [0.678301337403495]\n",
       "2       0.02  [0.6526258462520774]  [0.5082163291063126]\n",
       "0       0.00  [0.6492232520840716]                 [0.5]\n",
       "1       0.01  [0.6492232520840716]                 [0.5]"
      ]
     },
     "execution_count": 229,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "qwe = {\n",
    "    'threshold': [],\n",
    "    'F1': [],\n",
    "    'AUC': []\n",
    "}\n",
    "\n",
    "for thr in np.arange(0, 1, 0.01):\n",
    "    qwe['threshold'].append(thr)\n",
    "    qwe['F1'].append(eval_metric(y_test.values, \n",
    "                                 (y_pred_proba_test_lgb > thr).astype(int),\n",
    "                                'F1', weight=weights_for_test_set))\n",
    "    qwe['AUC'].append(eval_metric(y_test.values,\n",
    "                                    (y_pred_proba_test_lgb > thr).astype(int),\n",
    "                                   'AUC'))\n",
    "\n",
    "qwe = pd.DataFrame(qwe, columns=['threshold', 'F1', 'AUC'])\n",
    "print(qwe.sort_values('AUC', ascending=False).head(5))\n",
    "qwe.sort_values('F1', ascending=False).head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 205,
   "id": "a7c235be",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T18:48:30.947654Z",
     "start_time": "2022-03-22T18:48:10.159595Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "      gamma  threshold  F1_weighted  Precision_weighted        F1  Precision\n",
      "50      0.0        0.5     0.000689            0.923077  0.000689   0.500000\n",
      "150     0.1        0.5     0.000689            0.750000  0.000688   0.200000\n",
      "250     0.2        0.5     0.002753            0.774194  0.002741   0.222222\n",
      "350     0.3        0.5     0.015712            0.862500  0.015499   0.343284\n",
      "450     0.4        0.5     0.052787            0.857143  0.050351   0.333333\n",
      "550     0.5        0.5     0.142741            0.850957  0.125486   0.322397\n",
      "650     0.6        0.5     0.269236            0.813206  0.200901   0.266212\n",
      "750     0.7        0.5     0.396987            0.778188  0.244698   0.226222\n",
      "850     0.8        0.5     0.503221            0.747875  0.260324   0.198198\n",
      "950     0.9        0.5     0.575587            0.717540  0.256250   0.174709\n",
      "1050    1.0        0.5     0.618280            0.689348  0.244144   0.156061\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>gamma</th>\n",
       "      <th>threshold</th>\n",
       "      <th>F1_weighted</th>\n",
       "      <th>Precision_weighted</th>\n",
       "      <th>F1</th>\n",
       "      <th>Precision</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>313</th>\n",
       "      <td>0.3</td>\n",
       "      <td>0.13</td>\n",
       "      <td>0.700497</td>\n",
       "      <td>0.573100</td>\n",
       "      <td>0.181012</td>\n",
       "      <td>0.100616</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>416</th>\n",
       "      <td>0.4</td>\n",
       "      <td>0.16</td>\n",
       "      <td>0.699824</td>\n",
       "      <td>0.574164</td>\n",
       "      <td>0.181551</td>\n",
       "      <td>0.101010</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>210</th>\n",
       "      <td>0.2</td>\n",
       "      <td>0.10</td>\n",
       "      <td>0.699464</td>\n",
       "      <td>0.570475</td>\n",
       "      <td>0.179509</td>\n",
       "      <td>0.099650</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.699247</td>\n",
       "      <td>0.594732</td>\n",
       "      <td>0.193126</td>\n",
       "      <td>0.108966</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>108</th>\n",
       "      <td>0.1</td>\n",
       "      <td>0.08</td>\n",
       "      <td>0.699235</td>\n",
       "      <td>0.592865</td>\n",
       "      <td>0.192045</td>\n",
       "      <td>0.108217</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>622</th>\n",
       "      <td>0.6</td>\n",
       "      <td>0.22</td>\n",
       "      <td>0.698518</td>\n",
       "      <td>0.575837</td>\n",
       "      <td>0.182385</td>\n",
       "      <td>0.101634</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>519</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.19</td>\n",
       "      <td>0.698506</td>\n",
       "      <td>0.574810</td>\n",
       "      <td>0.181818</td>\n",
       "      <td>0.101251</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>211</th>\n",
       "      <td>0.2</td>\n",
       "      <td>0.11</td>\n",
       "      <td>0.698369</td>\n",
       "      <td>0.590793</td>\n",
       "      <td>0.190788</td>\n",
       "      <td>0.107392</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>623</th>\n",
       "      <td>0.6</td>\n",
       "      <td>0.23</td>\n",
       "      <td>0.698146</td>\n",
       "      <td>0.588836</td>\n",
       "      <td>0.189652</td>\n",
       "      <td>0.106619</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>520</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.697968</td>\n",
       "      <td>0.588909</td>\n",
       "      <td>0.189680</td>\n",
       "      <td>0.106648</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>829</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.29</td>\n",
       "      <td>0.697929</td>\n",
       "      <td>0.588365</td>\n",
       "      <td>0.189367</td>\n",
       "      <td>0.106434</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>725</th>\n",
       "      <td>0.7</td>\n",
       "      <td>0.25</td>\n",
       "      <td>0.697928</td>\n",
       "      <td>0.576491</td>\n",
       "      <td>0.182705</td>\n",
       "      <td>0.101879</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>417</th>\n",
       "      <td>0.4</td>\n",
       "      <td>0.17</td>\n",
       "      <td>0.697833</td>\n",
       "      <td>0.589369</td>\n",
       "      <td>0.189933</td>\n",
       "      <td>0.106829</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>107</th>\n",
       "      <td>0.1</td>\n",
       "      <td>0.07</td>\n",
       "      <td>0.697772</td>\n",
       "      <td>0.567277</td>\n",
       "      <td>0.177665</td>\n",
       "      <td>0.098487</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>726</th>\n",
       "      <td>0.7</td>\n",
       "      <td>0.26</td>\n",
       "      <td>0.697702</td>\n",
       "      <td>0.588368</td>\n",
       "      <td>0.189352</td>\n",
       "      <td>0.106435</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>828</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.28</td>\n",
       "      <td>0.697506</td>\n",
       "      <td>0.577092</td>\n",
       "      <td>0.183009</td>\n",
       "      <td>0.102104</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>314</th>\n",
       "      <td>0.3</td>\n",
       "      <td>0.14</td>\n",
       "      <td>0.697506</td>\n",
       "      <td>0.589723</td>\n",
       "      <td>0.190111</td>\n",
       "      <td>0.106969</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>212</th>\n",
       "      <td>0.2</td>\n",
       "      <td>0.12</td>\n",
       "      <td>0.697396</td>\n",
       "      <td>0.610481</td>\n",
       "      <td>0.202298</td>\n",
       "      <td>0.115518</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>315</th>\n",
       "      <td>0.3</td>\n",
       "      <td>0.15</td>\n",
       "      <td>0.697282</td>\n",
       "      <td>0.607236</td>\n",
       "      <td>0.200337</td>\n",
       "      <td>0.114133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>724</th>\n",
       "      <td>0.7</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0.697221</td>\n",
       "      <td>0.564152</td>\n",
       "      <td>0.175951</td>\n",
       "      <td>0.097363</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>932</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.32</td>\n",
       "      <td>0.697177</td>\n",
       "      <td>0.587783</td>\n",
       "      <td>0.188981</td>\n",
       "      <td>0.106206</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>518</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.18</td>\n",
       "      <td>0.696964</td>\n",
       "      <td>0.559557</td>\n",
       "      <td>0.173491</td>\n",
       "      <td>0.095735</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>418</th>\n",
       "      <td>0.4</td>\n",
       "      <td>0.18</td>\n",
       "      <td>0.696963</td>\n",
       "      <td>0.604315</td>\n",
       "      <td>0.198570</td>\n",
       "      <td>0.112902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>931</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.31</td>\n",
       "      <td>0.696861</td>\n",
       "      <td>0.577394</td>\n",
       "      <td>0.183131</td>\n",
       "      <td>0.102218</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1035</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.35</td>\n",
       "      <td>0.696807</td>\n",
       "      <td>0.587419</td>\n",
       "      <td>0.188747</td>\n",
       "      <td>0.106063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1034</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.34</td>\n",
       "      <td>0.696577</td>\n",
       "      <td>0.577900</td>\n",
       "      <td>0.183392</td>\n",
       "      <td>0.102408</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>621</th>\n",
       "      <td>0.6</td>\n",
       "      <td>0.21</td>\n",
       "      <td>0.696548</td>\n",
       "      <td>0.561836</td>\n",
       "      <td>0.174673</td>\n",
       "      <td>0.096539</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>827</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.27</td>\n",
       "      <td>0.696458</td>\n",
       "      <td>0.565409</td>\n",
       "      <td>0.176575</td>\n",
       "      <td>0.097813</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1033</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.33</td>\n",
       "      <td>0.696407</td>\n",
       "      <td>0.568193</td>\n",
       "      <td>0.178071</td>\n",
       "      <td>0.098818</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>521</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.21</td>\n",
       "      <td>0.696250</td>\n",
       "      <td>0.601771</td>\n",
       "      <td>0.197007</td>\n",
       "      <td>0.111842</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>109</th>\n",
       "      <td>0.1</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.696224</td>\n",
       "      <td>0.614411</td>\n",
       "      <td>0.204583</td>\n",
       "      <td>0.117221</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>930</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.695953</td>\n",
       "      <td>0.566632</td>\n",
       "      <td>0.177200</td>\n",
       "      <td>0.098254</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.04</td>\n",
       "      <td>0.695347</td>\n",
       "      <td>0.562357</td>\n",
       "      <td>0.174874</td>\n",
       "      <td>0.096723</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>624</th>\n",
       "      <td>0.6</td>\n",
       "      <td>0.24</td>\n",
       "      <td>0.695333</td>\n",
       "      <td>0.599675</td>\n",
       "      <td>0.195703</td>\n",
       "      <td>0.110978</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>727</th>\n",
       "      <td>0.7</td>\n",
       "      <td>0.27</td>\n",
       "      <td>0.695137</td>\n",
       "      <td>0.598124</td>\n",
       "      <td>0.194780</td>\n",
       "      <td>0.110342</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>625</th>\n",
       "      <td>0.6</td>\n",
       "      <td>0.25</td>\n",
       "      <td>0.695104</td>\n",
       "      <td>0.611670</td>\n",
       "      <td>0.202823</td>\n",
       "      <td>0.116031</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>522</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.22</td>\n",
       "      <td>0.695075</td>\n",
       "      <td>0.614441</td>\n",
       "      <td>0.204502</td>\n",
       "      <td>0.117234</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1032</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.32</td>\n",
       "      <td>0.694925</td>\n",
       "      <td>0.557814</td>\n",
       "      <td>0.172446</td>\n",
       "      <td>0.095125</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>929</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.29</td>\n",
       "      <td>0.694922</td>\n",
       "      <td>0.555808</td>\n",
       "      <td>0.171394</td>\n",
       "      <td>0.094427</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>415</th>\n",
       "      <td>0.4</td>\n",
       "      <td>0.15</td>\n",
       "      <td>0.694717</td>\n",
       "      <td>0.555176</td>\n",
       "      <td>0.171051</td>\n",
       "      <td>0.094208</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>935</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.35</td>\n",
       "      <td>0.694561</td>\n",
       "      <td>0.614860</td>\n",
       "      <td>0.204713</td>\n",
       "      <td>0.117417</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>832</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.32</td>\n",
       "      <td>0.694472</td>\n",
       "      <td>0.617406</td>\n",
       "      <td>0.206263</td>\n",
       "      <td>0.118538</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>830</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.694403</td>\n",
       "      <td>0.596681</td>\n",
       "      <td>0.193883</td>\n",
       "      <td>0.109754</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1036</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.36</td>\n",
       "      <td>0.694354</td>\n",
       "      <td>0.595017</td>\n",
       "      <td>0.192915</td>\n",
       "      <td>0.109081</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>826</th>\n",
       "      <td>0.8</td>\n",
       "      <td>0.26</td>\n",
       "      <td>0.694290</td>\n",
       "      <td>0.553163</td>\n",
       "      <td>0.169977</td>\n",
       "      <td>0.093515</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>729</th>\n",
       "      <td>0.7</td>\n",
       "      <td>0.29</td>\n",
       "      <td>0.694059</td>\n",
       "      <td>0.620128</td>\n",
       "      <td>0.207903</td>\n",
       "      <td>0.119749</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1038</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.38</td>\n",
       "      <td>0.694023</td>\n",
       "      <td>0.612598</td>\n",
       "      <td>0.203292</td>\n",
       "      <td>0.116432</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>933</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.33</td>\n",
       "      <td>0.693771</td>\n",
       "      <td>0.595394</td>\n",
       "      <td>0.193089</td>\n",
       "      <td>0.109233</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>419</th>\n",
       "      <td>0.4</td>\n",
       "      <td>0.19</td>\n",
       "      <td>0.693533</td>\n",
       "      <td>0.617177</td>\n",
       "      <td>0.206039</td>\n",
       "      <td>0.118436</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>312</th>\n",
       "      <td>0.3</td>\n",
       "      <td>0.12</td>\n",
       "      <td>0.693498</td>\n",
       "      <td>0.550717</td>\n",
       "      <td>0.168664</td>\n",
       "      <td>0.092680</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      gamma  threshold  F1_weighted  Precision_weighted        F1  Precision\n",
       "313     0.3       0.13     0.700497            0.573100  0.181012   0.100616\n",
       "416     0.4       0.16     0.699824            0.574164  0.181551   0.101010\n",
       "210     0.2       0.10     0.699464            0.570475  0.179509   0.099650\n",
       "5       0.0       0.05     0.699247            0.594732  0.193126   0.108966\n",
       "108     0.1       0.08     0.699235            0.592865  0.192045   0.108217\n",
       "622     0.6       0.22     0.698518            0.575837  0.182385   0.101634\n",
       "519     0.5       0.19     0.698506            0.574810  0.181818   0.101251\n",
       "211     0.2       0.11     0.698369            0.590793  0.190788   0.107392\n",
       "623     0.6       0.23     0.698146            0.588836  0.189652   0.106619\n",
       "520     0.5       0.20     0.697968            0.588909  0.189680   0.106648\n",
       "829     0.8       0.29     0.697929            0.588365  0.189367   0.106434\n",
       "725     0.7       0.25     0.697928            0.576491  0.182705   0.101879\n",
       "417     0.4       0.17     0.697833            0.589369  0.189933   0.106829\n",
       "107     0.1       0.07     0.697772            0.567277  0.177665   0.098487\n",
       "726     0.7       0.26     0.697702            0.588368  0.189352   0.106435\n",
       "828     0.8       0.28     0.697506            0.577092  0.183009   0.102104\n",
       "314     0.3       0.14     0.697506            0.589723  0.190111   0.106969\n",
       "212     0.2       0.12     0.697396            0.610481  0.202298   0.115518\n",
       "315     0.3       0.15     0.697282            0.607236  0.200337   0.114133\n",
       "724     0.7       0.24     0.697221            0.564152  0.175951   0.097363\n",
       "932     0.9       0.32     0.697177            0.587783  0.188981   0.106206\n",
       "518     0.5       0.18     0.696964            0.559557  0.173491   0.095735\n",
       "418     0.4       0.18     0.696963            0.604315  0.198570   0.112902\n",
       "931     0.9       0.31     0.696861            0.577394  0.183131   0.102218\n",
       "1035    1.0       0.35     0.696807            0.587419  0.188747   0.106063\n",
       "1034    1.0       0.34     0.696577            0.577900  0.183392   0.102408\n",
       "621     0.6       0.21     0.696548            0.561836  0.174673   0.096539\n",
       "827     0.8       0.27     0.696458            0.565409  0.176575   0.097813\n",
       "1033    1.0       0.33     0.696407            0.568193  0.178071   0.098818\n",
       "521     0.5       0.21     0.696250            0.601771  0.197007   0.111842\n",
       "109     0.1       0.09     0.696224            0.614411  0.204583   0.117221\n",
       "930     0.9       0.30     0.695953            0.566632  0.177200   0.098254\n",
       "4       0.0       0.04     0.695347            0.562357  0.174874   0.096723\n",
       "624     0.6       0.24     0.695333            0.599675  0.195703   0.110978\n",
       "727     0.7       0.27     0.695137            0.598124  0.194780   0.110342\n",
       "625     0.6       0.25     0.695104            0.611670  0.202823   0.116031\n",
       "522     0.5       0.22     0.695075            0.614441  0.204502   0.117234\n",
       "1032    1.0       0.32     0.694925            0.557814  0.172446   0.095125\n",
       "929     0.9       0.29     0.694922            0.555808  0.171394   0.094427\n",
       "415     0.4       0.15     0.694717            0.555176  0.171051   0.094208\n",
       "935     0.9       0.35     0.694561            0.614860  0.204713   0.117417\n",
       "832     0.8       0.32     0.694472            0.617406  0.206263   0.118538\n",
       "830     0.8       0.30     0.694403            0.596681  0.193883   0.109754\n",
       "1036    1.0       0.36     0.694354            0.595017  0.192915   0.109081\n",
       "826     0.8       0.26     0.694290            0.553163  0.169977   0.093515\n",
       "729     0.7       0.29     0.694059            0.620128  0.207903   0.119749\n",
       "1038    1.0       0.38     0.694023            0.612598  0.203292   0.116432\n",
       "933     0.9       0.33     0.693771            0.595394  0.193089   0.109233\n",
       "419     0.4       0.19     0.693533            0.617177  0.206039   0.118436\n",
       "312     0.3       0.12     0.693498            0.550717  0.168664   0.092680"
      ]
     },
     "execution_count": 205,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "jkl = {\n",
    "    'gamma': [],\n",
    "    'threshold': [], \n",
    "    'F1_weighted': [], \n",
    "    'Precision_weighted': [], \n",
    "    'F1': [], \n",
    "    'Precision': [],\n",
    "}\n",
    "class_wei = int(y_test.value_counts()[0] / y_test.value_counts()[1])\n",
    "d = {0: 1, 1: class_wei}\n",
    "\n",
    "#y_test = pd.Series(y_test)\n",
    "weights_for_test_set = y_test.copy()\n",
    "weights_for_test_set = weights_for_test_set.map(d).values\n",
    "#gamma = .5\n",
    "for gamma in np.arange(0, 1.1, 0.1):\n",
    "    for thr in np.arange(0, 1, 0.01):\n",
    "        jkl['gamma'].append(gamma)\n",
    "        jkl['threshold'].append(thr)\n",
    "        y_pred = ((gamma*y_pred_proba_test_pt_cb[:,1]  + (1-gamma)*y_pred_proba_test_lgb) > thr).astype(int)\n",
    "        jkl['F1_weighted'].append(eval_metric(y_test.values, y_pred, 'F1', weight=weights_for_test_set)[0])\n",
    "        jkl['F1'].append(eval_metric(y_test.values, y_pred, 'F1')[0])\n",
    "        jkl['Precision_weighted'].append(eval_metric(y_test.values, y_pred, 'Precision', weight=weights_for_test_set)[0])\n",
    "        jkl['Precision'].append(eval_metric(y_test.values, y_pred, 'Precision')[0])\n",
    "\n",
    "jkl = pd.DataFrame(jkl)\n",
    "print(jkl[jkl.threshold == .5])\n",
    "jkl.sort_values('F1_weighted', ascending=False).head(50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "id": "945af3ad",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T19:02:58.641765Z",
     "start_time": "2022-03-22T19:02:58.637676Z"
    }
   },
   "outputs": [],
   "source": [
    "gamma, thr = 0.7, 0.32\n",
    "y_pred = ((gamma*y_pred_proba_test_cb[:,1]  + (1-gamma)*y_pred_proba_test_lgb) > thr).astype(int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "id": "77444112",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-03-22T19:02:58.929295Z",
     "start_time": "2022-03-22T19:02:58.917310Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[23968, 13650],\n",
       "       [  827,  2074]])"
      ]
     },
     "execution_count": 225,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(y_test, y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e74d29f7",
   "metadata": {},
   "outputs": [],
   "source": [
    "array([[23711, 13907],\n",
    "       [  807,  2094]])"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}