restaurant_classification.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3d6e56f8",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "7dbcd0d0",
   "metadata": {},
   "source": [
    "## Table of Contents\n",
    "* [1. Exploratory Data Analysis](#eda)\n",
    "    * [1.1. Downloading data](#download)\n",
    "    * [1.2. Data Evaluation](#evaluation)\n",
    "    * [1.3. Selecting the country](#countries)\n",
    "        * [1.3.1. Dutch dataset](#select_dutch)\n",
    "        * [1.3.2. Italian dataset](#select_italian)\n",
    "        * [1.3.3. The UK dataset](#select_uk)\n",
    "        * [1.3.4. German dataset](#select_german)\n",
    "    \n",
    "<!--     * [1.2. Deleting outliers and mistakes in data](#1_2)\n",
    "    * [1.3. Adding new columns](#1_3)\n",
    "    * [1.4. Adding Geospatial data](#1_4) -->\n",
    "<!--     * [2.1. Comparing of cities](#2_1)\n",
    "    * [2.2. Bivariate Analysis of Warsaw](#2_2) -->\n",
    "* [2. Model Fitting & Evaluation](#models)\n",
    "\n",
    "    * [2.1. Classification of the Italian dataset by the \"Serves Alcohol\" feature](#italian_serves_alcohol)\n",
    "<!--        * [2.1.1. Logistic Regression](#logreg)\n",
    "        * [2.1.2. SVM](#svm)\n",
    "        * [2.1.3. Naive Bayes](#nb)\n",
    "        * [2.1.4. k-Nearest Neighbors](#knn)\n",
    "        * [2.1.5. Decision Tree](#tree)\n",
    "        * [2.1.6. Random Forest](#forest)\n",
    "        * [2.1.7. Voting Classifier](#voting)\n",
    "        * [2.1.8. Bagging Classifier](#bagging)\n",
    "        * [2.1.9. AdaBoost Classifier](#adaboost)\n",
    "        * [2.1.10. Gradient Boosting Classifier](#gradient)\n",
    "        * [2.1.11. XGBoost](#xgboost)\n",
    "        * [2.1.12. Stacked Ensembles](#stacked) -->\n",
    "    * [2.2. Classification of the UK dataset by the \"Vegetarian Friendly\" feature](#uk_vegetarian)\n",
    "    * [2.3. Classification of German dataset by the \"price_level\" feature](#german_price_level)\n",
    "      \n",
    "\n",
    "* [3. Conclusion](#conclusion)\n",
    "\n",
    "    * [3.1. For Classification of Italian dataset by the \"Serves Alcohol\" feature](#conclusion_italian_serves_alcohol)\n",
    "    * [3.2. For Classification of the UK dataset by the \"Vegetarian Friendly\" feature](#conclusion_UK_vegetarian)\n",
    "    * [3.3. For Classification of German dataset by the \"price_level\" feature](#conclusion_german_price_level)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "135e64fd",
   "metadata": {},
   "source": [
    "<a id=\"eda\"></a>\n",
    "\n",
    "# 1. EDA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 236,
   "id": "879f6098",
   "metadata": {},
   "outputs": [],
   "source": [
    "import opendatasets as od\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import os\n",
    "import copy\n",
    "import time\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "71d447e5",
   "metadata": {},
   "source": [
    "<a id=\"base\"></a>\n",
    "\n",
    "## 1.1. Opening the base dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 238,
   "id": "31caac15",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/zh/mnvg84rj3_7413ycjzpn5y8r0000gp/T/ipykernel_3296/2763966263.py:1: DtypeWarning: Columns (14,16,18,19,20,21,22,23,24,25,26,27,28) have mixed types. Specify dtype option on import or set low_memory=False.\n",
      "  base_rest_df = pd.read_csv('restaurant_base.csv', sep=';')\n"
     ]
    }
   ],
   "source": [
    "base_rest_df = pd.read_csv('restaurant_base.csv', sep=';')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 239,
   "id": "e52dd791",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>restaurant_link</th>\n",
       "      <th>country</th>\n",
       "      <th>region</th>\n",
       "      <th>province</th>\n",
       "      <th>city</th>\n",
       "      <th>latitude</th>\n",
       "      <th>longitude</th>\n",
       "      <th>price_level</th>\n",
       "      <th>cuisines</th>\n",
       "      <th>features</th>\n",
       "      <th>...</th>\n",
       "      <th>meals_Lunch</th>\n",
       "      <th>meals_Breakfast</th>\n",
       "      <th>meals_Drinks</th>\n",
       "      <th>meals_Brunch</th>\n",
       "      <th>meals_After-hours</th>\n",
       "      <th>special_diets_Vegetarian Friendly</th>\n",
       "      <th>special_diets_Gluten Free Options</th>\n",
       "      <th>special_diets_Vegan Options</th>\n",
       "      <th>special_diets_Halal</th>\n",
       "      <th>special_diets_Kosher</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>g10001637-d10002227</td>\n",
       "      <td>France</td>\n",
       "      <td>Nouvelle-Aquitaine</td>\n",
       "      <td>Haute-Vienne</td>\n",
       "      <td>Saint-Jouvent</td>\n",
       "      <td>45.961674</td>\n",
       "      <td>1.169131</td>\n",
       "      <td>€</td>\n",
       "      <td>French</td>\n",
       "      <td>Reservations, Seating, Wheelchair Accessible, ...</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>g10002858-d4586832</td>\n",
       "      <td>France</td>\n",
       "      <td>Centre-Val de Loire</td>\n",
       "      <td>Berry</td>\n",
       "      <td>Rivarennes</td>\n",
       "      <td>46.635895</td>\n",
       "      <td>1.386133</td>\n",
       "      <td>€</td>\n",
       "      <td>French, European</td>\n",
       "      <td>Reservations, Seating, Table Service, Wheelcha...</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>g10002986-d3510044</td>\n",
       "      <td>France</td>\n",
       "      <td>Nouvelle-Aquitaine</td>\n",
       "      <td>Correze</td>\n",
       "      <td>Lacelle</td>\n",
       "      <td>45.642610</td>\n",
       "      <td>1.824460</td>\n",
       "      <td>€</td>\n",
       "      <td>French</td>\n",
       "      <td>Reservations, Seating, Serves Alcohol, Table S...</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>g10022428-d9767191</td>\n",
       "      <td>France</td>\n",
       "      <td>Occitanie</td>\n",
       "      <td>Aveyron</td>\n",
       "      <td>Saint-Laurent-de-Levezou</td>\n",
       "      <td>44.208860</td>\n",
       "      <td>2.960470</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>French</td>\n",
       "      <td>Reservations, Seating, Wheelchair Accessible, ...</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>g10055971-d10040330</td>\n",
       "      <td>France</td>\n",
       "      <td>Auvergne-Rhone-Alpes</td>\n",
       "      <td>Ain</td>\n",
       "      <td>Journans</td>\n",
       "      <td>46.144340</td>\n",
       "      <td>5.331200</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>French</td>\n",
       "      <td>Reservations, Seating, Table Service, Serves A...</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 40 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       restaurant_link country                region      province  \\\n",
       "0  g10001637-d10002227  France    Nouvelle-Aquitaine  Haute-Vienne   \n",
       "1   g10002858-d4586832  France   Centre-Val de Loire         Berry   \n",
       "2   g10002986-d3510044  France    Nouvelle-Aquitaine       Correze   \n",
       "3   g10022428-d9767191  France             Occitanie       Aveyron   \n",
       "4  g10055971-d10040330  France  Auvergne-Rhone-Alpes           Ain   \n",
       "\n",
       "                       city   latitude  longitude price_level  \\\n",
       "0             Saint-Jouvent  45.961674   1.169131           €   \n",
       "1                Rivarennes  46.635895   1.386133           €   \n",
       "2                   Lacelle  45.642610   1.824460           €   \n",
       "3  Saint-Laurent-de-Levezou  44.208860   2.960470      €€-€€€   \n",
       "4                  Journans  46.144340   5.331200      €€-€€€   \n",
       "\n",
       "           cuisines                                           features  ...  \\\n",
       "0            French  Reservations, Seating, Wheelchair Accessible, ...  ...   \n",
       "1  French, European  Reservations, Seating, Table Service, Wheelcha...  ...   \n",
       "2            French  Reservations, Seating, Serves Alcohol, Table S...  ...   \n",
       "3            French  Reservations, Seating, Wheelchair Accessible, ...  ...   \n",
       "4            French  Reservations, Seating, Table Service, Serves A...  ...   \n",
       "\n",
       "   meals_Lunch  meals_Breakfast  meals_Drinks  meals_Brunch meals_After-hours  \\\n",
       "0            1                0             0             0                 0   \n",
       "1            1                0             1             0                 0   \n",
       "2            1                0             0             0                 0   \n",
       "3            1                0             0             0                 0   \n",
       "4            1                0             0             0                 0   \n",
       "\n",
       "   special_diets_Vegetarian Friendly special_diets_Gluten Free Options  \\\n",
       "0                                  0                                 0   \n",
       "1                                  0                                 0   \n",
       "2                                  0                                 0   \n",
       "3                                  0                                 0   \n",
       "4                                  0                                 0   \n",
       "\n",
       "   special_diets_Vegan Options special_diets_Halal special_diets_Kosher  \n",
       "0                            0                   0                    0  \n",
       "1                            0                   0                    0  \n",
       "2                            0                   0                    0  \n",
       "3                            0                   0                    0  \n",
       "4                            0                   0                    0  \n",
       "\n",
       "[5 rows x 40 columns]"
      ]
     },
     "execution_count": 239,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "base_rest_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 240,
   "id": "8b5de25e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['restaurant_link', 'country', 'region', 'province', 'city', 'latitude',\n",
       "       'longitude', 'price_level', 'cuisines', 'features', 'avg_rating',\n",
       "       'total_reviews_count', 'price_min', 'price_max', 'min_open_dttm',\n",
       "       'min_open_sec', 'max_close_dttm', 'max_close_sec', 'break_flg',\n",
       "       'max_break_hours', 'max_no_weekend_hours', 'max_weekend_hours',\n",
       "       'open_mon_hours', 'open_tue_hours', 'open_wed_hours', 'open_thu_hours',\n",
       "       'open_fri_hours', 'open_sat_hours', 'open_sun_hours', 'meals_Dinner',\n",
       "       'meals_Lunch', 'meals_Breakfast', 'meals_Drinks', 'meals_Brunch',\n",
       "       'meals_After-hours', 'special_diets_Vegetarian Friendly',\n",
       "       'special_diets_Gluten Free Options', 'special_diets_Vegan Options',\n",
       "       'special_diets_Halal', 'special_diets_Kosher'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 240,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "base_rest_df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "000e062b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "bbcef1d9",
   "metadata": {},
   "source": [
    "<a id=\"evaluation\"></a>\n",
    "\n",
    "## 1.2. Data Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 241,
   "id": "afb5dda8",
   "metadata": {},
   "outputs": [],
   "source": [
    "drop_columns = [\n",
    "    \n",
    "    'restaurant_link', \n",
    "#     'country', 'region', 'province', \n",
    "    'city', 'latitude', 'longitude', \n",
    "#     'price_level', 'cuisines', 'features', \n",
    "    'avg_rating',\n",
    "       'total_reviews_count', \n",
    "    'price_min', 'price_max', \n",
    "    'min_open_dttm',\n",
    "       'min_open_sec', 'max_close_dttm', 'max_close_sec', 'break_flg',\n",
    "       'max_break_hours', 'max_no_weekend_hours', 'max_weekend_hours',\n",
    "       'open_mon_hours', 'open_tue_hours', 'open_wed_hours', 'open_thu_hours',\n",
    "       'open_fri_hours', 'open_sat_hours', 'open_sun_hours', \n",
    "#     'meals_Dinner',\n",
    "#        'meals_Lunch', 'meals_Breakfast', 'meals_Drinks', 'meals_Brunch',\n",
    "#        'meals_After-hours', \n",
    "#     'special_diets_Vegetarian Friendly',\n",
    "#        'special_diets_Gluten Free Options', 'special_diets_Vegan Options',\n",
    "#        'special_diets_Halal', 'special_diets_Kosher'\n",
    "    \n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 242,
   "id": "6fc257f6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['country', 'region', 'province', 'price_level', 'cuisines', 'features',\n",
       "       'meals_Dinner', 'meals_Lunch', 'meals_Breakfast', 'meals_Drinks',\n",
       "       'meals_Brunch', 'meals_After-hours',\n",
       "       'special_diets_Vegetarian Friendly',\n",
       "       'special_diets_Gluten Free Options', 'special_diets_Vegan Options',\n",
       "       'special_diets_Halal', 'special_diets_Kosher'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 242,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rest_df_clf2 = base_rest_df.drop(drop_columns, axis=1)\n",
    "rest_df_clf2.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5b55e212",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "06c7a1e9",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "989eaf01",
   "metadata": {},
   "source": [
    "<a id=\"countries\"></a>\n",
    "\n",
    "## 1.3. Selecting the country"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6087231b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# dummies without not popular options of feature columns\n",
    "\n",
    "def features_to_flags(df, feature_col_list, threshold):\n",
    "    for col in feature_col_list:\n",
    "        top_tags = [f[0] for f in rating_of_column_values(df, col, separator=', ') if f[1] >= threshold * len(df)]\n",
    "        for t in top_tags:\n",
    "            df[col + '_' + t] = df[col].apply(lambda x: 1 if t in str(x) else 0)\n",
    "        df = df.drop(col, axis=1)\n",
    "    return df\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "6083f9a7",
   "metadata": {},
   "outputs": [],
   "source": [
    "def features_percentage(df):\n",
    "    features_percent_df = (df.select_dtypes(include=['int64']).sum() / df.shape[0] * 100.).sort_values(ascending=False)\n",
    "    return features_percent_df\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0a49f8e2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "id": "61233683",
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_country_dataset(rest_df_clf2, country_name, sub_country, feature_col_list, threshold_1, threshold_2):\n",
    "    \n",
    "    dutch_df = rest_df_clf2[rest_df_clf2['country'] == country_name].drop(['country', sub_country], axis=1)\n",
    "    dutch_df = features_to_flags(dutch_df, feature_col_list, threshold_1) # 0.02\n",
    "\n",
    "#     print(dutch_df.head())\n",
    "    dutch_features_percent_df = features_percentage(dutch_df)\n",
    "    print(dutch_features_percent_df)\n",
    "    \n",
    "    dutch_df2 = dutch_df[\n",
    "        list(dutch_df.select_dtypes(include=['object'])) + \n",
    "        list(dutch_features_percent_df[dutch_features_percent_df.values >= threshold_2].index)\n",
    "    ]\n",
    "    dutch_df2\n",
    "\n",
    "    return dutch_df2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9e93be52",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "c9008144",
   "metadata": {},
   "source": [
    "<a id=\"select_dutch\"></a>\n",
    "\n",
    "### 1.3.1. Dutch dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "id": "88fda0e3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "country          province              \n",
       "The Netherlands  Drenthe Province            78\n",
       "                 Flevoland Province          47\n",
       "                 Friesland Province         100\n",
       "                 Gelderland Province        280\n",
       "                 Groningen Province          76\n",
       "                 Limburg Province           248\n",
       "                 North Brabant Province     419\n",
       "                 North Holland Province    1047\n",
       "                 Overijssel Province        159\n",
       "                 South Holland Province     678\n",
       "                 Utrecht Province           213\n",
       "                 Zeeland Province           154\n",
       "dtype: int64"
      ]
     },
     "execution_count": 243,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rest_df_clf2[['country','province']][rest_df_clf2['country'] == 'The Netherlands']\\\n",
    "    .groupby(['country','province']).value_counts()\n",
    "# 'region',"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 244,
   "id": "d9def44d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "meals_Dinner                         83.309517\n",
      "features_Seating                     80.594456\n",
      "features_Reservations                75.650186\n",
      "features_Table Service               68.648185\n",
      "cuisines_European                    61.103172\n",
      "meals_Lunch                          59.731352\n",
      "features_Wheelchair Accessible       52.757931\n",
      "cuisines_Dutch                       51.100314\n",
      "features_Serves Alcohol              41.354673\n",
      "features_Outdoor Seating             25.521578\n",
      "meals_Brunch                         23.006573\n",
      "features_Takeout                     21.863390\n",
      "meals_Breakfast                      17.833667\n",
      "meals_Drinks                         15.718777\n",
      "features_Free Wifi                   15.518720\n",
      "features_Highchairs Available        15.347242\n",
      "cuisines_Seafood                     15.318662\n",
      "features_Accepts Credit Cards        13.575307\n",
      "cuisines_Bar                         13.232352\n",
      "features_Full Bar                    12.889397\n",
      "cuisines_Cafe                        10.374393\n",
      "cuisines_Asian                        9.859960\n",
      "special_diets_Vegetarian Friendly     9.174050\n",
      "cuisines_Chinese                      8.373821\n",
      "features_Wine and Beer                7.802229\n",
      "cuisines_International                7.573592\n",
      "cuisines_Pub                          7.459274\n",
      "cuisines_Italian                      7.202058\n",
      "features_Parking Available            6.887682\n",
      "meals_After-hours                     6.773364\n",
      "features_Digital Payments             6.659045\n",
      "special_diets_Vegan Options           6.173192\n",
      "cuisines_Fast food                    5.887396\n",
      "special_diets_Gluten Free Options     5.801658\n",
      "features_Street Parking               5.230066\n",
      "features_Accepts Mastercard           4.829951\n",
      "cuisines_French                       4.801372\n",
      "features_Accepts Visa                 4.687053\n",
      "cuisines_Steakhouse                   4.486996\n",
      "features_Delivery                     4.344098\n",
      "features_Free off-street parking      4.172621\n",
      "cuisines_Mediterranean                4.029723\n",
      "special_diets_Halal                   0.228637\n",
      "special_diets_Kosher                  0.000000\n",
      "dtype: float64\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>province</th>\n",
       "      <th>price_level</th>\n",
       "      <th>meals_Dinner</th>\n",
       "      <th>features_Seating</th>\n",
       "      <th>features_Reservations</th>\n",
       "      <th>features_Table Service</th>\n",
       "      <th>cuisines_European</th>\n",
       "      <th>meals_Lunch</th>\n",
       "      <th>features_Wheelchair Accessible</th>\n",
       "      <th>cuisines_Dutch</th>\n",
       "      <th>features_Serves Alcohol</th>\n",
       "      <th>features_Outdoor Seating</th>\n",
       "      <th>meals_Brunch</th>\n",
       "      <th>features_Takeout</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>44523</th>\n",
       "      <td>North Brabant Province</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44524</th>\n",
       "      <td>South Holland Province</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44525</th>\n",
       "      <td>South Holland Province</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44526</th>\n",
       "      <td>South Holland Province</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44527</th>\n",
       "      <td>Friesland Province</td>\n",
       "      <td>€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48017</th>\n",
       "      <td>Gelderland Province</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48018</th>\n",
       "      <td>Gelderland Province</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48019</th>\n",
       "      <td>Gelderland Province</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48020</th>\n",
       "      <td>Gelderland Province</td>\n",
       "      <td>€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48021</th>\n",
       "      <td>North Holland Province</td>\n",
       "      <td>€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3499 rows × 14 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     province price_level  meals_Dinner  features_Seating  \\\n",
       "44523  North Brabant Province      €€-€€€             1                 1   \n",
       "44524  South Holland Province      €€-€€€             1                 1   \n",
       "44525  South Holland Province      €€-€€€             1                 1   \n",
       "44526  South Holland Province      €€-€€€             1                 0   \n",
       "44527      Friesland Province           €             1                 1   \n",
       "...                       ...         ...           ...               ...   \n",
       "48017     Gelderland Province      €€-€€€             0                 1   \n",
       "48018     Gelderland Province      €€-€€€             1                 1   \n",
       "48019     Gelderland Province      €€-€€€             1                 1   \n",
       "48020     Gelderland Province           €             1                 1   \n",
       "48021  North Holland Province           €             1                 1   \n",
       "\n",
       "       features_Reservations  features_Table Service  cuisines_European  \\\n",
       "44523                      1                       0                  0   \n",
       "44524                      1                       1                  0   \n",
       "44525                      1                       1                  0   \n",
       "44526                      1                       1                  1   \n",
       "44527                      0                       0                  0   \n",
       "...                      ...                     ...                ...   \n",
       "48017                      1                       1                  1   \n",
       "48018                      1                       1                  0   \n",
       "48019                      1                       1                  0   \n",
       "48020                      1                       1                  1   \n",
       "48021                      0                       0                  1   \n",
       "\n",
       "       meals_Lunch  features_Wheelchair Accessible  cuisines_Dutch  \\\n",
       "44523            0                               0               0   \n",
       "44524            0                               0               0   \n",
       "44525            1                               0               1   \n",
       "44526            0                               0               1   \n",
       "44527            1                               1               0   \n",
       "...            ...                             ...             ...   \n",
       "48017            1                               0               1   \n",
       "48018            1                               1               0   \n",
       "48019            1                               1               0   \n",
       "48020            1                               0               1   \n",
       "48021            1                               1               1   \n",
       "\n",
       "       features_Serves Alcohol  features_Outdoor Seating  meals_Brunch  \\\n",
       "44523                        0                         0             0   \n",
       "44524                        0                         0             0   \n",
       "44525                        0                         0             0   \n",
       "44526                        0                         0             0   \n",
       "44527                        0                         1             0   \n",
       "...                        ...                       ...           ...   \n",
       "48017                        0                         0             0   \n",
       "48018                        1                         1             0   \n",
       "48019                        1                         0             0   \n",
       "48020                        0                         0             0   \n",
       "48021                        0                         1             0   \n",
       "\n",
       "       features_Takeout  \n",
       "44523                 1  \n",
       "44524                 1  \n",
       "44525                 0  \n",
       "44526                 0  \n",
       "44527                 0  \n",
       "...                 ...  \n",
       "48017                 0  \n",
       "48018                 1  \n",
       "48019                 1  \n",
       "48020                 0  \n",
       "48021                 1  \n",
       "\n",
       "[3499 rows x 14 columns]"
      ]
     },
     "execution_count": 244,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_col_list = ['cuisines', 'features'] # 'meals', 'special_diets',\n",
    "\n",
    "dutch_df2 = create_country_dataset(rest_df_clf2, 'The Netherlands', 'region', feature_col_list, 0.04, 20)\n",
    "dutch_df2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "cb2b4b20",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "price_level\n",
       "€          777\n",
       "€€-€€€    2656\n",
       "€€€€        66\n",
       "dtype: int64"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# dutch_df.groupby('price_level').size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "id": "a9e180b5",
   "metadata": {},
   "outputs": [],
   "source": [
    "dutch_df2.to_csv('df_for_classifications/dutch_df_for_classification.csv', sep='\\t', encoding='utf-8', index=False)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "095f2d6f",
   "metadata": {},
   "source": [
    "<a id=\"select_italian\"></a>\n",
    "\n",
    "### 1.3.2. Italian dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "id": "371a8518",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "country  region               \n",
       "Italy    Abruzzo                  1165\n",
       "         Basilicata                272\n",
       "         Calabria                  914\n",
       "         Campania                 2934\n",
       "         Emilia-Romagna           3402\n",
       "         Friuli Venezia Giulia     948\n",
       "         Lazio                    3919\n",
       "         Liguria                  1766\n",
       "         Lombardy                 5544\n",
       "         Marche                   1209\n",
       "         Molise                    206\n",
       "         Piedmont                 2836\n",
       "         Puglia                   2088\n",
       "         Sardinia                 1096\n",
       "         Sicily                   2390\n",
       "         Trentino-Alto Adige       973\n",
       "         Tuscany                  3044\n",
       "         Umbria                    484\n",
       "         Valle d'Aosta             279\n",
       "         Veneto                   3269\n",
       "dtype: int64"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rest_df_clf2[['country','region']][rest_df_clf2['country'] == 'Italy']\\\n",
    "    .groupby(['country','region']).value_counts()\n",
    "# 'region', 'province'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 245,
   "id": "81b783a1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "features_Seating                     87.536786\n",
      "cuisines_Italian                     86.558418\n",
      "features_Reservations                79.281326\n",
      "features_Table Service               79.175487\n",
      "meals_Dinner                         74.170066\n",
      "meals_Lunch                          66.250194\n",
      "features_Wheelchair Accessible       62.269606\n",
      "features_Serves Alcohol              55.772110\n",
      "features_Takeout                     32.838556\n",
      "features_Outdoor Seating             26.366875\n",
      "meals_Breakfast                      25.636326\n",
      "cuisines_Mediterranean               24.066808\n",
      "cuisines_Pizza                       22.001652\n",
      "cuisines_Seafood                     21.805462\n",
      "features_Highchairs Available        20.793536\n",
      "features_Accepts Credit Cards        18.903919\n",
      "cuisines_Bar                         18.387630\n",
      "features_Full Bar                    16.544478\n",
      "meals_Drinks                         15.276989\n",
      "meals_After-hours                    15.163405\n",
      "cuisines_Cafe                        12.705870\n",
      "meals_Brunch                         11.479684\n",
      "features_Parking Available           10.904022\n",
      "features_Free Wifi                   10.292220\n",
      "features_Wine and Beer               10.016005\n",
      "cuisines_Pub                          8.645258\n",
      "cuisines_Barbecue                     7.142857\n",
      "features_Television                   6.969900\n",
      "features_Free off-street parking      6.076721\n",
      "features_Street Parking               5.885694\n",
      "features_Delivery                     5.565595\n",
      "features_Accepts Visa                 5.144819\n",
      "features_Accepts Mastercard           4.933141\n",
      "cuisines_Southern-Italian             4.736951\n",
      "cuisines_Northern-Italian             4.535598\n",
      "cuisines_European                     4.249058\n",
      "special_diets_Vegetarian Friendly     3.077082\n",
      "special_diets_Gluten Free Options     2.372348\n",
      "special_diets_Vegan Options           2.003201\n",
      "special_diets_Kosher                  0.012907\n",
      "special_diets_Halal                   0.007744\n",
      "dtype: float64\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>region</th>\n",
       "      <th>price_level</th>\n",
       "      <th>features_Seating</th>\n",
       "      <th>cuisines_Italian</th>\n",
       "      <th>features_Reservations</th>\n",
       "      <th>features_Table Service</th>\n",
       "      <th>meals_Dinner</th>\n",
       "      <th>meals_Lunch</th>\n",
       "      <th>features_Wheelchair Accessible</th>\n",
       "      <th>features_Serves Alcohol</th>\n",
       "      <th>features_Takeout</th>\n",
       "      <th>features_Outdoor Seating</th>\n",
       "      <th>meals_Breakfast</th>\n",
       "      <th>cuisines_Mediterranean</th>\n",
       "      <th>cuisines_Pizza</th>\n",
       "      <th>cuisines_Seafood</th>\n",
       "      <th>features_Highchairs Available</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>126442</th>\n",
       "      <td>Emilia-Romagna</td>\n",
       "      <td>€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126443</th>\n",
       "      <td>Trentino-Alto Adige</td>\n",
       "      <td>€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126444</th>\n",
       "      <td>Sicily</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126445</th>\n",
       "      <td>Sicily</td>\n",
       "      <td>€</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126446</th>\n",
       "      <td>Sicily</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>165175</th>\n",
       "      <td>Lombardy</td>\n",
       "      <td>€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>165176</th>\n",
       "      <td>Emilia-Romagna</td>\n",
       "      <td>€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>165177</th>\n",
       "      <td>Calabria</td>\n",
       "      <td>€</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>165178</th>\n",
       "      <td>Campania</td>\n",
       "      <td>€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>165179</th>\n",
       "      <td>Campania</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>38738 rows × 17 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     region price_level  features_Seating  cuisines_Italian  \\\n",
       "126442       Emilia-Romagna           €                 1                 1   \n",
       "126443  Trentino-Alto Adige           €                 1                 1   \n",
       "126444               Sicily      €€-€€€                 1                 1   \n",
       "126445               Sicily           €                 1                 0   \n",
       "126446               Sicily      €€-€€€                 1                 1   \n",
       "...                     ...         ...               ...               ...   \n",
       "165175             Lombardy           €                 1                 1   \n",
       "165176       Emilia-Romagna           €                 1                 1   \n",
       "165177             Calabria           €                 0                 1   \n",
       "165178             Campania           €                 1                 1   \n",
       "165179             Campania      €€-€€€                 1                 1   \n",
       "\n",
       "        features_Reservations  features_Table Service  meals_Dinner  \\\n",
       "126442                      1                       1             1   \n",
       "126443                      1                       1             1   \n",
       "126444                      1                       1             1   \n",
       "126445                      1                       1             0   \n",
       "126446                      1                       1             0   \n",
       "...                       ...                     ...           ...   \n",
       "165175                      1                       1             1   \n",
       "165176                      1                       0             1   \n",
       "165177                      1                       1             1   \n",
       "165178                      1                       1             1   \n",
       "165179                      1                       1             1   \n",
       "\n",
       "        meals_Lunch  features_Wheelchair Accessible  features_Serves Alcohol  \\\n",
       "126442            1                               1                        1   \n",
       "126443            1                               0                        0   \n",
       "126444            0                               1                        1   \n",
       "126445            0                               1                        1   \n",
       "126446            1                               1                        1   \n",
       "...             ...                             ...                      ...   \n",
       "165175            1                               1                        1   \n",
       "165176            1                               1                        1   \n",
       "165177            1                               1                        0   \n",
       "165178            1                               1                        1   \n",
       "165179            1                               1                        1   \n",
       "\n",
       "        features_Takeout  features_Outdoor Seating  meals_Breakfast  \\\n",
       "126442                 1                         1                0   \n",
       "126443                 0                         0                0   \n",
       "126444                 1                         1                0   \n",
       "126445                 1                         1                1   \n",
       "126446                 0                         0                1   \n",
       "...                  ...                       ...              ...   \n",
       "165175                 0                         1                0   \n",
       "165176                 0                         1                0   \n",
       "165177                 0                         0                0   \n",
       "165178                 0                         0                0   \n",
       "165179                 0                         1                0   \n",
       "\n",
       "        cuisines_Mediterranean  cuisines_Pizza  cuisines_Seafood  \\\n",
       "126442                       0               0                 0   \n",
       "126443                       0               0                 0   \n",
       "126444                       0               0                 0   \n",
       "126445                       0               0                 0   \n",
       "126446                       0               0                 1   \n",
       "...                        ...             ...               ...   \n",
       "165175                       0               0                 1   \n",
       "165176                       0               0                 0   \n",
       "165177                       0               0                 0   \n",
       "165178                       1               1                 0   \n",
       "165179                       1               0                 0   \n",
       "\n",
       "        features_Highchairs Available  \n",
       "126442                              0  \n",
       "126443                              0  \n",
       "126444                              0  \n",
       "126445                              0  \n",
       "126446                              0  \n",
       "...                               ...  \n",
       "165175                              1  \n",
       "165176                              0  \n",
       "165177                              0  \n",
       "165178                              1  \n",
       "165179                              1  \n",
       "\n",
       "[38738 rows x 17 columns]"
      ]
     },
     "execution_count": 245,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_col_list = ['cuisines', 'features'] # 'meals', 'special_diets',\n",
    "\n",
    "italian_df2 = create_country_dataset(rest_df_clf2, 'Italy', 'province', feature_col_list, 0.04, 20)\n",
    "italian_df2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "id": "553d82a8",
   "metadata": {},
   "outputs": [],
   "source": [
    "italian_df2.to_csv('df_for_classifications/italian_df_for_classification.csv', sep='\\t', encoding='utf-8', index=False)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0fa51858",
   "metadata": {},
   "source": [
    "<a id=\"select_uk\"></a>\n",
    "\n",
    "### 1.3.3. The UK dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "id": "adbfa5d5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "country         region         \n",
       "United Kingdom  Aberdeenshire       261\n",
       "                Angus                63\n",
       "                Argyll and Bute     118\n",
       "                Ayrshire            165\n",
       "                Bedfordshire        229\n",
       "                                   ... \n",
       "                West Midlands       869\n",
       "                West Sussex         507\n",
       "                Wiltshire           352\n",
       "                Worcestershire      322\n",
       "                Yorkshire          2753\n",
       "Length: 86, dtype: int64"
      ]
     },
     "execution_count": 207,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rest_df_clf2[['country','region']][rest_df_clf2['country'] == 'United Kingdom']\\\n",
    "    .groupby(['country','region']).value_counts()\n",
    "# 'region', 'province'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 246,
   "id": "feebf284",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "features_Seating                     80.751372\n",
      "meals_Lunch                          79.093851\n",
      "meals_Dinner                         74.295765\n",
      "features_Wheelchair Accessible       58.494442\n",
      "features_Reservations                52.339947\n",
      "features_Serves Alcohol              49.576474\n",
      "cuisines_British                     49.356972\n",
      "features_Table Service               48.957366\n",
      "features_Takeout                     48.740678\n",
      "features_Highchairs Available        44.879696\n",
      "special_diets_Vegetarian Friendly    40.211060\n",
      "features_Full Bar                    36.760940\n",
      "meals_Breakfast                      34.056564\n",
      "features_Accepts Credit Cards        33.611932\n",
      "special_diets_Gluten Free Options    32.776136\n",
      "features_Free Wifi                   31.360630\n",
      "features_Outdoor Seating             31.090474\n",
      "meals_Brunch                         31.042634\n",
      "special_diets_Vegan Options          30.054875\n",
      "cuisines_Bar                         27.544674\n",
      "cuisines_Pub                         24.457577\n",
      "cuisines_Cafe                        22.217532\n",
      "meals_Drinks                         21.764458\n",
      "features_Parking Available           20.312368\n",
      "cuisines_Fast food                   19.729844\n",
      "meals_After-hours                    17.731814\n",
      "features_Delivery                    14.231040\n",
      "features_Accepts Visa                13.766709\n",
      "features_Accepts Mastercard          13.423385\n",
      "features_Wine and Beer               12.199240\n",
      "cuisines_Asian                       10.215281\n",
      "features_Television                   9.356972\n",
      "cuisines_European                     9.117771\n",
      "features_Free off-street parking      8.068102\n",
      "cuisines_Seafood                      7.212607\n",
      "features_Digital Payments             6.669481\n",
      "features_Accepts American Express     6.635711\n",
      "cuisines_Pizza                        6.334600\n",
      "cuisines_Chinese                      6.283945\n",
      "features_Street Parking               5.999719\n",
      "features_Gift Cards Available         5.960321\n",
      "cuisines_Indian                       5.808358\n",
      "cuisines_Italian                      5.611369\n",
      "cuisines_Mediterranean                5.040101\n",
      "features_Live Music                   4.654566\n",
      "cuisines_Scottish                     4.074856\n",
      "special_diets_Halal                   2.479246\n",
      "special_diets_Kosher                  0.019699\n",
      "dtype: float64\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>region</th>\n",
       "      <th>price_level</th>\n",
       "      <th>features_Seating</th>\n",
       "      <th>meals_Lunch</th>\n",
       "      <th>meals_Dinner</th>\n",
       "      <th>features_Wheelchair Accessible</th>\n",
       "      <th>features_Reservations</th>\n",
       "      <th>features_Serves Alcohol</th>\n",
       "      <th>cuisines_British</th>\n",
       "      <th>features_Table Service</th>\n",
       "      <th>...</th>\n",
       "      <th>special_diets_Gluten Free Options</th>\n",
       "      <th>features_Free Wifi</th>\n",
       "      <th>features_Outdoor Seating</th>\n",
       "      <th>meals_Brunch</th>\n",
       "      <th>special_diets_Vegan Options</th>\n",
       "      <th>cuisines_Bar</th>\n",
       "      <th>cuisines_Pub</th>\n",
       "      <th>cuisines_Cafe</th>\n",
       "      <th>meals_Drinks</th>\n",
       "      <th>features_Parking Available</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>90907</th>\n",
       "      <td>Cornwall</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90908</th>\n",
       "      <td>Surrey</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90909</th>\n",
       "      <td>Surrey</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90910</th>\n",
       "      <td>NaN</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90911</th>\n",
       "      <td>East Sussex</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126437</th>\n",
       "      <td>Greater Manchester</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126438</th>\n",
       "      <td>Greater Manchester</td>\n",
       "      <td>€</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126439</th>\n",
       "      <td>Kent</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126440</th>\n",
       "      <td>Cumbria</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126441</th>\n",
       "      <td>Rutland</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>35535 rows × 26 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                    region price_level  features_Seating  meals_Lunch  \\\n",
       "90907             Cornwall      €€-€€€                 1            1   \n",
       "90908               Surrey      €€-€€€                 0            1   \n",
       "90909               Surrey      €€-€€€                 1            1   \n",
       "90910                  NaN      €€-€€€                 1            1   \n",
       "90911          East Sussex      €€-€€€                 1            1   \n",
       "...                    ...         ...               ...          ...   \n",
       "126437  Greater Manchester      €€-€€€                 1            0   \n",
       "126438  Greater Manchester           €                 0            0   \n",
       "126439                Kent      €€-€€€                 1            0   \n",
       "126440             Cumbria      €€-€€€                 1            1   \n",
       "126441             Rutland      €€-€€€                 1            1   \n",
       "\n",
       "        meals_Dinner  features_Wheelchair Accessible  features_Reservations  \\\n",
       "90907              0                               1                      1   \n",
       "90908              1                               0                      0   \n",
       "90909              1                               0                      0   \n",
       "90910              1                               1                      1   \n",
       "90911              0                               1                      1   \n",
       "...              ...                             ...                    ...   \n",
       "126437             0                               1                      1   \n",
       "126438             0                               1                      0   \n",
       "126439             1                               1                      1   \n",
       "126440             1                               0                      1   \n",
       "126441             0                               0                      0   \n",
       "\n",
       "        features_Serves Alcohol  cuisines_British  features_Table Service  \\\n",
       "90907                         1                 1                       0   \n",
       "90908                         0                 0                       0   \n",
       "90909                         1                 1                       1   \n",
       "90910                         1                 1                       1   \n",
       "90911                         1                 1                       1   \n",
       "...                         ...               ...                     ...   \n",
       "126437                        1                 0                       0   \n",
       "126438                        0                 1                       0   \n",
       "126439                        1                 1                       1   \n",
       "126440                        1                 1                       1   \n",
       "126441                        0                 0                       1   \n",
       "\n",
       "        ...  special_diets_Gluten Free Options  features_Free Wifi  \\\n",
       "90907   ...                                  1                   1   \n",
       "90908   ...                                  0                   0   \n",
       "90909   ...                                  0                   0   \n",
       "90910   ...                                  0                   0   \n",
       "90911   ...                                  1                   1   \n",
       "...     ...                                ...                 ...   \n",
       "126437  ...                                  0                   1   \n",
       "126438  ...                                  0                   0   \n",
       "126439  ...                                  0                   0   \n",
       "126440  ...                                  0                   0   \n",
       "126441  ...                                  0                   1   \n",
       "\n",
       "        features_Outdoor Seating  meals_Brunch  special_diets_Vegan Options  \\\n",
       "90907                          1             1                            1   \n",
       "90908                          0             0                            0   \n",
       "90909                          1             0                            0   \n",
       "90910                          0             0                            0   \n",
       "90911                          1             1                            1   \n",
       "...                          ...           ...                          ...   \n",
       "126437                         0             0                            0   \n",
       "126438                         0             0                            0   \n",
       "126439                         0             0                            0   \n",
       "126440                         1             0                            0   \n",
       "126441                         1             1                            0   \n",
       "\n",
       "        cuisines_Bar  cuisines_Pub  cuisines_Cafe  meals_Drinks  \\\n",
       "90907              0             0              1             1   \n",
       "90908              0             0              0             0   \n",
       "90909              1             1              0             1   \n",
       "90910              0             0              0             0   \n",
       "90911              0             0              0             0   \n",
       "...              ...           ...            ...           ...   \n",
       "126437             0             0              0             1   \n",
       "126438             0             0              0             0   \n",
       "126439             0             0              0             0   \n",
       "126440             1             1              0             0   \n",
       "126441             0             0              1             1   \n",
       "\n",
       "        features_Parking Available  \n",
       "90907                            0  \n",
       "90908                            0  \n",
       "90909                            1  \n",
       "90910                            0  \n",
       "90911                            0  \n",
       "...                            ...  \n",
       "126437                           0  \n",
       "126438                           0  \n",
       "126439                           0  \n",
       "126440                           0  \n",
       "126441                           0  \n",
       "\n",
       "[35535 rows x 26 columns]"
      ]
     },
     "execution_count": 246,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_col_list = ['cuisines', 'features'] # 'meals', 'special_diets',\n",
    "\n",
    "uk_df2 = create_country_dataset(rest_df_clf2, 'United Kingdom', 'province', feature_col_list, 0.04, 20)\n",
    "uk_df2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4f0a7ca5",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "id": "1584bdf8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "features_Highchairs Available  special_diets_Vegetarian Friendly\n",
       "0                              0                                    16770\n",
       "                               1                                     2817\n",
       "1                              0                                     4476\n",
       "                               1                                    11472\n",
       "dtype: int64"
      ]
     },
     "execution_count": 228,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "uk_df3 = uk_df2[['features_Highchairs Available', 'special_diets_Vegetarian Friendly']]\n",
    "uk_df3.groupby(['features_Highchairs Available', 'special_diets_Vegetarian Friendly']).size()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "792f85a0",
   "metadata": {},
   "source": [
    "<a id=\"select_german\"></a>\n",
    "\n",
    "### 1.3.4. German dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "id": "50409f3e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "country  region                    \n",
       "Germany  Baden-Wurttemberg             1412\n",
       "         Bavaria                       1949\n",
       "         Berlin                         748\n",
       "         Brandenburg                    210\n",
       "         Hamburg                        279\n",
       "         Hesse                          909\n",
       "         Lower Saxony                   823\n",
       "         Mecklenburg-West Pomerania     274\n",
       "         North Rhine-Westphalia        1777\n",
       "         Rhineland-Palatinate           640\n",
       "         Saarland                       103\n",
       "         Saxony                         355\n",
       "         Saxony-Anhalt                  125\n",
       "         Schleswig-Holstein             395\n",
       "         State of Bremen                 96\n",
       "         Thuringia                      164\n",
       "dtype: int64"
      ]
     },
     "execution_count": 217,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rest_df_clf2[['country','region']][rest_df_clf2['country'] == 'Germany']\\\n",
    "    .groupby(['country','region']).value_counts()\n",
    "# 'region', 'province'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "id": "dae709d8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "meals_Dinner                         84.735354\n",
      "features_Seating                     80.924067\n",
      "features_Reservations                78.711375\n",
      "features_Table Service               69.071060\n",
      "meals_Lunch                          63.524710\n",
      "cuisines_German                      44.955649\n",
      "features_Wheelchair Accessible       28.131397\n",
      "cuisines_European                    22.965201\n",
      "features_Serves Alcohol              21.005946\n",
      "meals_Breakfast                      20.372356\n",
      "features_Takeout                     18.861487\n",
      "features_Outdoor Seating             15.937226\n",
      "cuisines_Italian                     14.523833\n",
      "cuisines_Cafe                        13.919485\n",
      "meals_Drinks                         11.940735\n",
      "features_Highchairs Available        10.546837\n",
      "cuisines_Central European             9.620821\n",
      "cuisines_Pizza                        8.646067\n",
      "meals_Brunch                          8.607077\n",
      "cuisines_Bar                          8.012477\n",
      "cuisines_Mediterranean                7.661565\n",
      "features_Wine and Beer                7.408129\n",
      "cuisines_Greek                        6.833025\n",
      "features_Full Bar                     6.764792\n",
      "features_Accepts Credit Cards         6.413880\n",
      "meals_After-hours                     6.160444\n",
      "features_Free Wifi                    5.838776\n",
      "features_Parking Available            5.712058\n",
      "cuisines_Fast food                    5.634077\n",
      "cuisines_Asian                        4.639828\n",
      "cuisines_Pub                          4.591091\n",
      "special_diets_Vegetarian Friendly     4.113461\n",
      "cuisines_Seafood                      4.025734\n",
      "special_diets_Vegan Options           2.232186\n",
      "special_diets_Gluten Free Options     1.666829\n",
      "special_diets_Halal                   0.107223\n",
      "special_diets_Kosher                  0.009748\n",
      "dtype: float64\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>region</th>\n",
       "      <th>price_level</th>\n",
       "      <th>meals_Dinner</th>\n",
       "      <th>features_Seating</th>\n",
       "      <th>features_Reservations</th>\n",
       "      <th>features_Table Service</th>\n",
       "      <th>meals_Lunch</th>\n",
       "      <th>cuisines_German</th>\n",
       "      <th>features_Wheelchair Accessible</th>\n",
       "      <th>cuisines_European</th>\n",
       "      <th>features_Serves Alcohol</th>\n",
       "      <th>meals_Breakfast</th>\n",
       "      <th>features_Takeout</th>\n",
       "      <th>features_Outdoor Seating</th>\n",
       "      <th>cuisines_Italian</th>\n",
       "      <th>cuisines_Cafe</th>\n",
       "      <th>meals_Drinks</th>\n",
       "      <th>features_Highchairs Available</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>48022</th>\n",
       "      <td>Baden-Wurttemberg</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48023</th>\n",
       "      <td>Saxony</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48024</th>\n",
       "      <td>Rhineland-Palatinate</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48025</th>\n",
       "      <td>Baden-Wurttemberg</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48026</th>\n",
       "      <td>North Rhine-Westphalia</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58276</th>\n",
       "      <td>Hesse</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58277</th>\n",
       "      <td>Baden-Wurttemberg</td>\n",
       "      <td>€</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58278</th>\n",
       "      <td>Baden-Wurttemberg</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58279</th>\n",
       "      <td>Baden-Wurttemberg</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58280</th>\n",
       "      <td>Schleswig-Holstein</td>\n",
       "      <td>€€-€€€</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>10259 rows × 18 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                       region price_level  meals_Dinner  features_Seating  \\\n",
       "48022       Baden-Wurttemberg      €€-€€€             1                 1   \n",
       "48023                  Saxony      €€-€€€             1                 0   \n",
       "48024    Rhineland-Palatinate      €€-€€€             1                 0   \n",
       "48025       Baden-Wurttemberg      €€-€€€             1                 1   \n",
       "48026  North Rhine-Westphalia      €€-€€€             1                 1   \n",
       "...                       ...         ...           ...               ...   \n",
       "58276                   Hesse      €€-€€€             1                 1   \n",
       "58277       Baden-Wurttemberg           €             1                 0   \n",
       "58278       Baden-Wurttemberg      €€-€€€             1                 1   \n",
       "58279       Baden-Wurttemberg      €€-€€€             1                 1   \n",
       "58280      Schleswig-Holstein      €€-€€€             1                 1   \n",
       "\n",
       "       features_Reservations  features_Table Service  meals_Lunch  \\\n",
       "48022                      1                       1            1   \n",
       "48023                      1                       0            1   \n",
       "48024                      1                       1            1   \n",
       "48025                      1                       1            1   \n",
       "48026                      1                       1            0   \n",
       "...                      ...                     ...          ...   \n",
       "58276                      1                       1            0   \n",
       "58277                      1                       0            1   \n",
       "58278                      1                       1            1   \n",
       "58279                      1                       1            1   \n",
       "58280                      1                       1            1   \n",
       "\n",
       "       cuisines_German  features_Wheelchair Accessible  cuisines_European  \\\n",
       "48022                1                               0                  1   \n",
       "48023                0                               0                  0   \n",
       "48024                1                               0                  1   \n",
       "48025                0                               0                  0   \n",
       "48026                0                               1                  0   \n",
       "...                ...                             ...                ...   \n",
       "58276                1                               0                  0   \n",
       "58277                0                               0                  0   \n",
       "58278                1                               1                  1   \n",
       "58279                1                               0                  0   \n",
       "58280                0                               0                  1   \n",
       "\n",
       "       features_Serves Alcohol  meals_Breakfast  features_Takeout  \\\n",
       "48022                        0                0                 0   \n",
       "48023                        0                0                 0   \n",
       "48024                        0                0                 0   \n",
       "48025                        0                0                 0   \n",
       "48026                        0                0                 0   \n",
       "...                        ...              ...               ...   \n",
       "58276                        0                0                 0   \n",
       "58277                        0                0                 0   \n",
       "58278                        1                0                 0   \n",
       "58279                        0                0                 0   \n",
       "58280                        0                0                 0   \n",
       "\n",
       "       features_Outdoor Seating  cuisines_Italian  cuisines_Cafe  \\\n",
       "48022                         0                 0              0   \n",
       "48023                         0                 0              0   \n",
       "48024                         0                 0              0   \n",
       "48025                         0                 0              0   \n",
       "48026                         0                 0              0   \n",
       "...                         ...               ...            ...   \n",
       "58276                         0                 0              0   \n",
       "58277                         0                 1              0   \n",
       "58278                         1                 0              0   \n",
       "58279                         0                 0              0   \n",
       "58280                         0                 0              0   \n",
       "\n",
       "       meals_Drinks  features_Highchairs Available  \n",
       "48022             0                              0  \n",
       "48023             0                              0  \n",
       "48024             0                              0  \n",
       "48025             0                              0  \n",
       "48026             0                              0  \n",
       "...             ...                            ...  \n",
       "58276             0                              0  \n",
       "58277             0                              0  \n",
       "58278             1                              1  \n",
       "58279             0                              0  \n",
       "58280             0                              0  \n",
       "\n",
       "[10259 rows x 18 columns]"
      ]
     },
     "execution_count": 247,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_col_list = ['cuisines', 'features'] # 'meals', 'special_diets',\n",
    "\n",
    "german_df2 = create_country_dataset(rest_df_clf2, 'Germany', 'province', feature_col_list, 0.04, 10)\n",
    "german_df2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cdb54cac",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 233,
   "id": "8dbf46cb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "cuisines_Italian  price_level\n",
       "0                 €              2561\n",
       "                  €€-€€€         6001\n",
       "                  €€€€            207\n",
       "1                 €               386\n",
       "                  €€-€€€         1088\n",
       "                  €€€€             16\n",
       "dtype: int64"
      ]
     },
     "execution_count": 233,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "german_df2.groupby(['cuisines_Italian', 'price_level']).size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c86fa95c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "id": "a8ca698c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "country         price_level\n",
       "Austria         €                488\n",
       "                €€-€€€          1446\n",
       "                €€€€              61\n",
       "Belgium         €               1043\n",
       "                €€-€€€          4531\n",
       "                €€€€             250\n",
       "Bulgaria        €                 91\n",
       "                €€-€€€           426\n",
       "                €€€€              19\n",
       "Croatia         €                285\n",
       "                €€-€€€          1333\n",
       "                €€€€              61\n",
       "Czech Republic  €                419\n",
       "                €€-€€€          1548\n",
       "                €€€€              47\n",
       "Denmark         €                182\n",
       "                €€-€€€          1012\n",
       "                €€€€              70\n",
       "Finland         €                116\n",
       "                €€-€€€           510\n",
       "                €€€€              22\n",
       "France          €              12586\n",
       "                €€-€€€         30068\n",
       "                €€€€            1869\n",
       "Germany         €               2947\n",
       "                €€-€€€          7089\n",
       "                €€€€             223\n",
       "Greece          €               1800\n",
       "                €€-€€€          3144\n",
       "                €€€€             129\n",
       "Hungary         €                204\n",
       "                €€-€€€           642\n",
       "                €€€€              39\n",
       "Ireland         €                472\n",
       "                €€-€€€          1676\n",
       "                €€€€              50\n",
       "Italy           €              15938\n",
       "                €€-€€€         22092\n",
       "                €€€€             708\n",
       "Poland          €                397\n",
       "dtype: int64"
      ]
     },
     "execution_count": 216,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rest_df_clf2.groupby(['country', 'price_level']).size()[:40]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "id": "893ebab3",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/zh/mnvg84rj3_7413ycjzpn5y8r0000gp/T/ipykernel_3296/1764051508.py:2: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  rest_df_clf3['target'] = rest_df_clf3['special_diets'].apply(lambda x: 1 if 'Vegetarian Friendly' in str(x) else 0 )\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "country          target\n",
       "Austria          0          1831\n",
       "                 1           164\n",
       "Belgium          0          5695\n",
       "                 1           129\n",
       "Bulgaria         0           442\n",
       "                 1            94\n",
       "Croatia          0          1189\n",
       "                 1           490\n",
       "Czech Republic   0          1817\n",
       "                 1           197\n",
       "Denmark          0          1164\n",
       "                 1           100\n",
       "Finland          0           580\n",
       "                 1            68\n",
       "France           0         44017\n",
       "                 1           506\n",
       "Germany          0          9837\n",
       "                 1           422\n",
       "Greece           0          3550\n",
       "                 1          1523\n",
       "Hungary          0           724\n",
       "                 1           161\n",
       "Ireland          0          1316\n",
       "                 1           882\n",
       "Italy            0         37546\n",
       "                 1          1192\n",
       "Poland           0          1620\n",
       "                 1           237\n",
       "Portugal         0          7578\n",
       "                 1           740\n",
       "Romania          0           949\n",
       "                 1           174\n",
       "Slovakia         0           486\n",
       "                 1            31\n",
       "Spain            0         28409\n",
       "                 1          2222\n",
       "Sweden           0          1215\n",
       "                 1           115\n",
       "The Netherlands  0          3178\n",
       "                 1           321\n",
       "United Kingdom   0         21246\n",
       "                 1         14289\n",
       "dtype: int64"
      ]
     },
     "execution_count": 206,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# searching countries with a high percentage of vegetarian friendly feature\n",
    "\n",
    "rest_df_clf3 = rest_df_clf2[['country', 'special_diets']]\n",
    "rest_df_clf3['target'] = rest_df_clf3['special_diets'].apply(lambda x: 1 if 'Vegetarian Friendly' in str(x) else 0 )\n",
    "# rest_df_clf3.head()\n",
    "rest_df_clf3.groupby(['country', 'target']).size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1ec3a74b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "1ce7843d",
   "metadata": {},
   "source": [
    "<a id=\"models\"></a>\n",
    "\n",
    "# 2. Model Fitting & Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "id": "9acb495a",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split, GridSearchCV, KFold\n",
    "# from sklearn.model_selection import cross_val_score, cross_val_predict\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "# from sklearn.preprocessing import StandardScaler, PowerTransformer, MinMaxScaler\n",
    "\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.svm import LinearSVC\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "from sklearn.ensemble import VotingClassifier\n",
    "from sklearn.ensemble import BaggingClassifier\n",
    "from sklearn.ensemble import AdaBoostClassifier\n",
    "from sklearn.ensemble import GradientBoostingClassifier\n",
    "from xgboost import XGBClassifier\n",
    "from sklearn.ensemble import StackingClassifier\n",
    "\n",
    "from sklearn.metrics import classification_report, confusion_matrix, roc_auc_score, accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "id": "c3c5667b",
   "metadata": {},
   "outputs": [],
   "source": [
    "def splitting_country_dataset(df, target_col):\n",
    "    '''\n",
    "    This function: \n",
    "    - encodes categorial features;\n",
    "    - defines the target variable;\n",
    "    - split the dataset into train and test.\n",
    "    '''\n",
    "    df2 = df[list(df)]\n",
    "    \n",
    "#     to labelencode all categorial variables\n",
    "    le_dict = {}\n",
    "    for cat_col in df2.select_dtypes(include=['object']):\n",
    "        le_dict[cat_col] = LabelEncoder()\n",
    "        df2[cat_col] = le_dict[cat_col].fit_transform(df2[cat_col])\n",
    "        print(cat_col)\n",
    "        print(dict(zip(le_dict[cat_col].classes_, range(len(le_dict[cat_col].classes_)))))\n",
    "        print()\n",
    "    \n",
    "    X = df2.drop(target_col, axis=1) # Features\n",
    "    y = df2[target_col] # Target variable\n",
    "\n",
    "    # Normalize the numeric features\n",
    "    # scaler = StandardScaler()\n",
    "    # X[num_names_1] = scaler.fit_transform(X[num_names_1])\n",
    "\n",
    "    # Split the data into a training set and a test set\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)\n",
    "        \n",
    "    return X_train, X_test, y_train, y_test, X.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "id": "197a66cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "def namestr(obj, namespace):\n",
    "    return [name for name in namespace if namespace[name] is obj]\n",
    "\n",
    "def clf_model_implementation(model, grid, X_train, y_train, X_test, y_test):\n",
    "    '''\n",
    "    This function:\n",
    "    - fits and returns classification model with GridSearchCV;\n",
    "    - prints confusion matrix and classification report.\n",
    "    '''\n",
    "    start_time = time.time()\n",
    "    mod = model()\n",
    "    mod_cv = GridSearchCV(mod, grid, cv=10)\n",
    "    mod_cv.fit(X_train, y_train)\n",
    "    \n",
    "    print(\"Model: {0}\".format(namestr(model, globals())[0]))\n",
    "    if grid != {}:\n",
    "        print(\"Tuned hyperparameters:\") # , mod_cv.best_params_\n",
    "        for k, v in mod_cv.best_params_.items():\n",
    "            print(\"\\t{0}: {1}\".format(k, v))\n",
    "    print()\n",
    "        \n",
    "    mod = model(**mod_cv.best_params_)\n",
    "    mod.fit(X_train, y_train)\n",
    "    y_pred_mod = mod.predict(X_test)\n",
    "\n",
    "    print(\"Confusion matrix:\\n\", confusion_matrix(y_test, y_pred_mod))\n",
    "    print()\n",
    "    print(\"Classification report:\\n\", classification_report(y_test, y_pred_mod))   \n",
    "#     print()\n",
    "    print(\"Time using clf_model_implementation(): {} sec\".format(time.time() - start_time))\n",
    "    \n",
    "    return mod, accuracy_score(y_test, y_pred_mod)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c0449e6c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# columns - predicted\n",
    "# rows - actual\n",
    "\n",
    "# for row in confusion_matrix(y_test, y_pred_knn).T:\n",
    "#     print(row)\n",
    "#     for i in row:\n",
    "#         print(i / sum(row))\n",
    "        \n",
    "# precision == column[i] / sum(column)\n",
    "# recall == row[i] / sum(row)\n",
    "# support == sum(row)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "33059791",
   "metadata": {},
   "outputs": [],
   "source": [
    "def show_feature_importances(model, col_names): # list(X_train) !!!!!!! \n",
    "    '''\n",
    "    This function plots the feature importances for given model.\n",
    "    '''\n",
    "    resultdict = {}\n",
    "    importance = model.feature_importances_\n",
    "\n",
    "    print(\"score:\\t  feature:\")\n",
    "    for i in range(len(col_names)):\n",
    "        resultdict[col_names[i]] = importance[i]\n",
    "    \n",
    "    resultdict = dict(sorted(resultdict.items(), key=lambda item: -item[1]))\n",
    "    for k, v in resultdict.items():\n",
    "        print(\"{1:.3f}\\t- {0}\".format(k, v))\n",
    "\n",
    "    plt.bar(resultdict.keys(),resultdict.values())\n",
    "    plt.xticks(rotation='vertical')\n",
    "    plt.title('Feature Importance')\n",
    "    plt.show()\n",
    "    \n",
    "    return 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7cef31ee",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ff42b210",
   "metadata": {},
   "outputs": [],
   "source": [
    "# X_train, X_test, y_train, y_test, X_columns = splitting_country_dataset(dutch_df2, 'price_level')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "696c15c6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# X_train, X_test, y_train, y_test, X_columns = splitting_country_dataset(italian_df2, 'price_level')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "458fb856",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "id": "794a7c8e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def run_classification_models(df, target_column, drop_columns, clf_models_dict):\n",
    "    \n",
    "    X_train, X_test, y_train, y_test, X_columns = splitting_country_dataset(\n",
    "        df.drop(drop_columns,axis=1), \n",
    "        target_column\n",
    "    )\n",
    "    \n",
    "    full_clf_dict = {}\n",
    "    \n",
    "    for model_name, attr in clf_models_dict.items():\n",
    "        if attr['status'] == True:\n",
    "            model, score = clf_model_implementation(attr['model'], attr['grid'], X_train, y_train, X_test, y_test)\n",
    "            full_clf_dict[model_name] = {\n",
    "                'model': model, \n",
    "                'accuracy_score': score\n",
    "            }\n",
    "            print()\n",
    "            print()\n",
    "    \n",
    "    return full_clf_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "id": "3fb9fcec",
   "metadata": {},
   "outputs": [],
   "source": [
    "def find_the_best_model(full_clf_dict):\n",
    "    \n",
    "    max_score = 0\n",
    "    best_model = 0\n",
    "    for k, v in full_clf_dict.items():\n",
    "        if v['accuracy_score'] > max_score:\n",
    "            max_score = v['accuracy_score']\n",
    "            best_model = v['model']\n",
    "            \n",
    "    return max_score, best_model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c51dffb0",
   "metadata": {},
   "source": [
    "<a id=\"italian_serves_alcohol\"></a>\n",
    "\n",
    "## 2.1. Classification of the Italian dataset by the \"Serves Alcohol\" feature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "id": "9e6752c7",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Logistic Regression\n",
    "logreg_grid = {\n",
    "    \"C\": np.logspace(-2, 1, 4), # np.logspace(-3, 3, 7) - from 0.001 to 1000 by 7 steps\n",
    "    \"penalty\": [\"l1\", \"l2\"], # , \"elasticnet\", None\n",
    "    \"multi_class\": [\"multinomial\"], # \"auto\", \"ovr\", \n",
    "    \"max_iter\": [1000],\n",
    "    \"solver\": ['lbfgs', 'newton-cg'] # For multiclass problems, only ‘newton-cg’, ‘sag’, ‘saga’ and ‘lbfgs’ handle multinomial loss;\n",
    "}\n",
    "\n",
    "# SVM\n",
    "svm_grid = {\n",
    "    \"C\": np.logspace(-2, 1, 4), # from 0.01 to 1 by 4 steps\n",
    "    \"penalty\": [\"l1\", \"l2\"],\n",
    "    \"loss\": [\"hinge\", \"squared_hinge\"],\n",
    "    \"max_iter\": [1000]\n",
    "}\n",
    "\n",
    "# Naive Bayes\n",
    "# no grid\n",
    "\n",
    "# k-Nearest Neighbors\n",
    "knn_grid = {\n",
    "    \"n_neighbors\": [6, 8],\n",
    "    \"leaf_size\": [12],\n",
    "    \"algorithm\": ['auto']\n",
    "    ,\"n_jobs\": [-1]\n",
    "}\n",
    "\n",
    "# Decision Tree\n",
    "tree_grid = {\n",
    "    \"min_samples_split\": range(4, 10, 2),\n",
    "    \"max_depth\": range(4, 10, 2),\n",
    "#     \"criterion\": ['squared_error', 'friedman_mse', 'absolute_error', 'poisson'],\n",
    "    \"splitter\": ['best', 'random']\n",
    "}\n",
    "\n",
    "# Random Forest\n",
    "forest_grid = {\n",
    "    \"n_estimators\": [200, 300], # range(100, 400, 100)\n",
    "    \"max_depth\": [8, 10], # range(4, 10, 2)\n",
    "    \"min_samples_split\": [20, 30], # range(20, 40, 10)\n",
    "    \"n_jobs\": [-1],\n",
    "    \"random_state\": [42]\n",
    "}\n",
    "\n",
    "# Gradient Boosting\n",
    "gb_grid = {\n",
    "    \"n_estimators\": [300], # range(100, 400, 100)\n",
    "#     \"max_depth\": [6], # range(4, 10, 2),\n",
    "    \"learning_rate\": [0.2], # [0.1, , 0.3],\n",
    "#     \"n_jobs\": [-1],\n",
    "#     \"objective\": ['reg:squarederror'],\n",
    "    \"random_state\": [42]\n",
    "}\n",
    "\n",
    "# XGB\n",
    "xgb_grid = {\n",
    "    \"n_estimators\": [300], # range(100, 400, 100)\n",
    "    \"max_depth\": [6], # range(4, 10, 2),\n",
    "    \"learning_rate\": [0.2], # [0.1, , 0.3],\n",
    "    \"n_jobs\": [-1],\n",
    "    \"objective\": ['reg:squarederror'],\n",
    "    \"random_state\": [42]\n",
    "}\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "id": "18140779",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "region\n",
      "{'Abruzzo': 0, 'Basilicata': 1, 'Calabria': 2, 'Campania': 3, 'Emilia-Romagna': 4, 'Friuli Venezia Giulia': 5, 'Lazio': 6, 'Liguria': 7, 'Lombardy': 8, 'Marche': 9, 'Molise': 10, 'Piedmont': 11, 'Puglia': 12, 'Sardinia': 13, 'Sicily': 14, 'Trentino-Alto Adige': 15, 'Tuscany': 16, 'Umbria': 17, \"Valle d'Aosta\": 18, 'Veneto': 19}\n",
      "\n",
      "price_level\n",
      "{'€': 0, '€€-€€€': 1, '€€€€': 2}\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py:378: FitFailedWarning: \n",
      "80 fits failed out of a total of 160.\n",
      "The score on these train-test partitions for these parameters will be set to nan.\n",
      "If these failures are not expected, you can try to debug them by setting error_score='raise'.\n",
      "\n",
      "Below are more details about the failures:\n",
      "--------------------------------------------------------------------------------\n",
      "40 fits failed with the following error:\n",
      "Traceback (most recent call last):\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py\", line 686, in _fit_and_score\n",
      "    estimator.fit(X_train, y_train, **fit_params)\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/linear_model/_logistic.py\", line 1162, in fit\n",
      "    solver = _check_solver(self.solver, self.penalty, self.dual)\n",
      "             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/linear_model/_logistic.py\", line 54, in _check_solver\n",
      "    raise ValueError(\n",
      "ValueError: Solver lbfgs supports only 'l2' or 'none' penalties, got l1 penalty.\n",
      "\n",
      "--------------------------------------------------------------------------------\n",
      "40 fits failed with the following error:\n",
      "Traceback (most recent call last):\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py\", line 686, in _fit_and_score\n",
      "    estimator.fit(X_train, y_train, **fit_params)\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/linear_model/_logistic.py\", line 1162, in fit\n",
      "    solver = _check_solver(self.solver, self.penalty, self.dual)\n",
      "             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/linear_model/_logistic.py\", line 54, in _check_solver\n",
      "    raise ValueError(\n",
      "ValueError: Solver newton-cg supports only 'l2' or 'none' penalties, got l1 penalty.\n",
      "\n",
      "  warnings.warn(some_fits_failed_message, FitFailedWarning)\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_search.py:953: UserWarning: One or more of the test scores are non-finite: [       nan        nan 0.76058385 0.76058385        nan        nan\n",
      " 0.76091566 0.76095254        nan        nan 0.76091573 0.76087885\n",
      "        nan        nan 0.76098948 0.76098948]\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: LogisticRegression\n",
      "Tuned hyperparameters:\n",
      "\tC: 10.0\n",
      "\tmax_iter: 1000\n",
      "\tmulti_class: multinomial\n",
      "\tpenalty: l2\n",
      "\tsolver: lbfgs\n",
      "\n",
      "Confusion matrix:\n",
      " [[3604 1585]\n",
      " [1249 5184]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.74      0.69      0.72      5189\n",
      "           1       0.77      0.81      0.79      6433\n",
      "\n",
      "    accuracy                           0.76     11622\n",
      "   macro avg       0.75      0.75      0.75     11622\n",
      "weighted avg       0.76      0.76      0.76     11622\n",
      "\n",
      "Time using clf_model_implementation(): 16.83566689491272 sec\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py:378: FitFailedWarning: \n",
      "80 fits failed out of a total of 160.\n",
      "The score on these train-test partitions for these parameters will be set to nan.\n",
      "If these failures are not expected, you can try to debug them by setting error_score='raise'.\n",
      "\n",
      "Below are more details about the failures:\n",
      "--------------------------------------------------------------------------------\n",
      "40 fits failed with the following error:\n",
      "Traceback (most recent call last):\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py\", line 686, in _fit_and_score\n",
      "    estimator.fit(X_train, y_train, **fit_params)\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_classes.py\", line 274, in fit\n",
      "    self.coef_, self.intercept_, n_iter_ = _fit_liblinear(\n",
      "                                           ^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py\", line 1223, in _fit_liblinear\n",
      "    solver_type = _get_liblinear_solver_type(multi_class, penalty, loss, dual)\n",
      "                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py\", line 1062, in _get_liblinear_solver_type\n",
      "    raise ValueError(\n",
      "ValueError: Unsupported set of arguments: The combination of penalty='l1' and loss='hinge' is not supported, Parameters: penalty='l1', loss='hinge', dual=True\n",
      "\n",
      "--------------------------------------------------------------------------------\n",
      "40 fits failed with the following error:\n",
      "Traceback (most recent call last):\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py\", line 686, in _fit_and_score\n",
      "    estimator.fit(X_train, y_train, **fit_params)\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_classes.py\", line 274, in fit\n",
      "    self.coef_, self.intercept_, n_iter_ = _fit_liblinear(\n",
      "                                           ^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py\", line 1223, in _fit_liblinear\n",
      "    solver_type = _get_liblinear_solver_type(multi_class, penalty, loss, dual)\n",
      "                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py\", line 1062, in _get_liblinear_solver_type\n",
      "    raise ValueError(\n",
      "ValueError: Unsupported set of arguments: The combination of penalty='l1' and loss='squared_hinge' are not supported when dual=True, Parameters: penalty='l1', loss='squared_hinge', dual=True\n",
      "\n",
      "  warnings.warn(some_fits_failed_message, FitFailedWarning)\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_search.py:953: UserWarning: One or more of the test scores are non-finite: [       nan 0.75597383        nan 0.76032568        nan 0.75785477\n",
      "        nan 0.76091568        nan 0.75840811        nan 0.75818706\n",
      "        nan 0.72562192        nan 0.70703813]\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: LinearSVC\n",
      "Tuned hyperparameters:\n",
      "\tC: 0.1\n",
      "\tloss: squared_hinge\n",
      "\tmax_iter: 1000\n",
      "\tpenalty: l2\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix:\n",
      " [[3578 1611]\n",
      " [1212 5221]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.75      0.69      0.72      5189\n",
      "           1       0.76      0.81      0.79      6433\n",
      "\n",
      "    accuracy                           0.76     11622\n",
      "   macro avg       0.76      0.75      0.75     11622\n",
      "weighted avg       0.76      0.76      0.76     11622\n",
      "\n",
      "Time using clf_model_implementation(): 46.355055809020996 sec\n",
      "Model: GaussianNB\n",
      "\n",
      "Confusion matrix:\n",
      " [[3287 1902]\n",
      " [1182 5251]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.74      0.63      0.68      5189\n",
      "           1       0.73      0.82      0.77      6433\n",
      "\n",
      "    accuracy                           0.73     11622\n",
      "   macro avg       0.73      0.72      0.73     11622\n",
      "weighted avg       0.73      0.73      0.73     11622\n",
      "\n",
      "Time using clf_model_implementation(): 0.10889315605163574 sec\n",
      "Model: KNeighborsClassifier\n",
      "Tuned hyperparameters:\n",
      "\talgorithm: auto\n",
      "\tleaf_size: 12\n",
      "\tn_jobs: -1\n",
      "\tn_neighbors: 8\n",
      "\n",
      "Confusion matrix:\n",
      " [[3944 1245]\n",
      " [1802 4631]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.69      0.76      0.72      5189\n",
      "           1       0.79      0.72      0.75      6433\n",
      "\n",
      "    accuracy                           0.74     11622\n",
      "   macro avg       0.74      0.74      0.74     11622\n",
      "weighted avg       0.74      0.74      0.74     11622\n",
      "\n",
      "Time using clf_model_implementation(): 30.723657846450806 sec\n",
      "Model: DecisionTreeClassifier\n",
      "Tuned hyperparameters:\n",
      "\tmax_depth: 8\n",
      "\tmin_samples_split: 4\n",
      "\tsplitter: random\n",
      "\n",
      "Confusion matrix:\n",
      " [[3752 1437]\n",
      " [1417 5016]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.73      0.72      0.72      5189\n",
      "           1       0.78      0.78      0.78      6433\n",
      "\n",
      "    accuracy                           0.75     11622\n",
      "   macro avg       0.75      0.75      0.75     11622\n",
      "weighted avg       0.75      0.75      0.75     11622\n",
      "\n",
      "Time using clf_model_implementation(): 2.6478428840637207 sec\n",
      "Model: RandomForestClassifier\n",
      "Tuned hyperparameters:\n",
      "\tmax_depth: 10\n",
      "\tmin_samples_split: 30\n",
      "\tn_estimators: 300\n",
      "\tn_jobs: -1\n",
      "\trandom_state: 42\n",
      "\n",
      "Confusion matrix:\n",
      " [[3584 1605]\n",
      " [1118 5315]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.76      0.69      0.72      5189\n",
      "           1       0.77      0.83      0.80      6433\n",
      "\n",
      "    accuracy                           0.77     11622\n",
      "   macro avg       0.77      0.76      0.76     11622\n",
      "weighted avg       0.77      0.77      0.76     11622\n",
      "\n",
      "Time using clf_model_implementation(): 40.27792477607727 sec\n",
      "Model: GradientBoostingClassifier\n",
      "Tuned hyperparameters:\n",
      "\tlearning_rate: 0.2\n",
      "\tn_estimators: 300\n",
      "\trandom_state: 42\n",
      "\n",
      "Confusion matrix:\n",
      " [[3727 1462]\n",
      " [1202 5231]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.76      0.72      0.74      5189\n",
      "           1       0.78      0.81      0.80      6433\n",
      "\n",
      "    accuracy                           0.77     11622\n",
      "   macro avg       0.77      0.77      0.77     11622\n",
      "weighted avg       0.77      0.77      0.77     11622\n",
      "\n",
      "Time using clf_model_implementation(): 33.75854825973511 sec\n",
      "Model: XGBClassifier\n",
      "Tuned hyperparameters:\n",
      "\tlearning_rate: 0.2\n",
      "\tmax_depth: 6\n",
      "\tn_estimators: 300\n",
      "\tn_jobs: -1\n",
      "\tobjective: reg:squarederror\n",
      "\trandom_state: 42\n",
      "\n",
      "Confusion matrix:\n",
      " [[3741 1448]\n",
      " [1222 5211]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.75      0.72      0.74      5189\n",
      "           1       0.78      0.81      0.80      6433\n",
      "\n",
      "    accuracy                           0.77     11622\n",
      "   macro avg       0.77      0.77      0.77     11622\n",
      "weighted avg       0.77      0.77      0.77     11622\n",
      "\n",
      "Time using clf_model_implementation(): 24.89090394973755 sec\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'logreg': {'model': LogisticRegression(C=10.0, max_iter=1000, multi_class='multinomial'),\n",
       "  'accuracy_score': 0.756152125279642},\n",
       " 'svm': {'model': LinearSVC(C=0.1), 'accuracy_score': 0.7570986060918947},\n",
       " 'nb': {'model': GaussianNB(), 'accuracy_score': 0.734641197728446},\n",
       " 'knn': {'model': KNeighborsClassifier(leaf_size=12, n_jobs=-1, n_neighbors=8),\n",
       "  'accuracy_score': 0.737824815006023},\n",
       " 'tree': {'model': DecisionTreeClassifier(max_depth=8, min_samples_split=4, splitter='random'),\n",
       "  'accuracy_score': 0.7544312510755464},\n",
       " 'forest': {'model': RandomForestClassifier(max_depth=10, min_samples_split=30, n_estimators=300,\n",
       "                         n_jobs=-1, random_state=42),\n",
       "  'accuracy_score': 0.7657029771123731},\n",
       " 'gb': {'model': GradientBoostingClassifier(learning_rate=0.2, n_estimators=300, random_state=42),\n",
       "  'accuracy_score': 0.7707795560144554},\n",
       " 'xgb': {'model': XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
       "                colsample_bylevel=None, colsample_bynode=None,\n",
       "                colsample_bytree=None, early_stopping_rounds=None,\n",
       "                enable_categorical=False, eval_metric=None, feature_types=None,\n",
       "                gamma=None, gpu_id=None, grow_policy=None, importance_type=None,\n",
       "                interaction_constraints=None, learning_rate=0.2, max_bin=None,\n",
       "                max_cat_threshold=None, max_cat_to_onehot=None,\n",
       "                max_delta_step=None, max_depth=6, max_leaves=None,\n",
       "                min_child_weight=None, missing=nan, monotone_constraints=None,\n",
       "                n_estimators=300, n_jobs=-1, num_parallel_tree=None,\n",
       "                objective='reg:squarederror', predictor=None, ...),\n",
       "  'accuracy_score': 0.7702632937532267}}"
      ]
     },
     "execution_count": 185,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# PIVOT DICTIONARY OF CLASSIFICATION MODELS\n",
    "clf_models_dict = {\n",
    "#     simple models\n",
    "    'logreg': {'status': True, 'model': LogisticRegression, 'grid': logreg_grid},\n",
    "    'svm': {'status': True, 'model': LinearSVC, 'grid': svm_grid},\n",
    "    'nb': {'status': True, 'model': GaussianNB, 'grid': {}},\n",
    "    'knn': {'status': True, 'model': KNeighborsClassifier, 'grid': knn_grid},\n",
    "    'tree': {'status': True, 'model': DecisionTreeClassifier, 'grid': tree_grid},\n",
    "#     ensemble models (not based on other models)\n",
    "    'forest': {'status': True, 'model': RandomForestClassifier, 'grid': forest_grid},\n",
    "    'gb': {'status': True, 'model': GradientBoostingClassifier, 'grid': gb_grid},\n",
    "    'xgb': {'status': True, 'model': XGBClassifier, 'grid': xgb_grid}\n",
    "}\n",
    "\n",
    "italian_clf_dict = run_classification_models(italian_df2, 'features_Serves Alcohol', [], clf_models_dict)\n",
    "# drop_columns = ['special_diets_Gluten Free Options', 'special_diets_Vegan Options']\n",
    "italian_clf_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "id": "2303bbf3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.7707795560144554,\n",
       " GradientBoostingClassifier(learning_rate=0.2, n_estimators=300, random_state=42))"
      ]
     },
     "execution_count": 188,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "find_the_best_model(italian_clf_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "id": "c8843f75",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "score:\t  feature:\n",
      "0.334\t- features_Highchairs Available\n",
      "0.186\t- features_Seating\n",
      "0.174\t- features_Outdoor Seating\n",
      "0.095\t- features_Table Service\n",
      "0.065\t- meals_Breakfast\n",
      "0.025\t- meals_Lunch\n",
      "0.021\t- region\n",
      "0.020\t- price_level\n",
      "0.014\t- features_Takeout\n",
      "0.012\t- features_Reservations\n",
      "0.012\t- cuisines_Mediterranean\n",
      "0.011\t- cuisines_Seafood\n",
      "0.011\t- meals_Dinner\n",
      "0.007\t- features_Wheelchair Accessible\n",
      "0.007\t- cuisines_Italian\n",
      "0.005\t- cuisines_Pizza\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiwAAAKACAYAAAC/qSN8AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8o6BhiAAAACXBIWXMAAA9hAAAPYQGoP6dpAAC2zklEQVR4nOzdd3xN9/8H8NdNZO9IZCGD1I4gNWoTYrRm7RqxWpsYNWOPKhpKzarQWkXR0qCx996bRCgJoqiEhOT9+yO/3G+uhIqcc3Pxej4e5yE59+TzPjdyz33fz/l83h+NiAiIiIiIDJhRbp8AERER0X9hwkJEREQGjwkLERERGTwmLERERGTwmLAQERGRwWPCQkRERAaPCQsREREZPCYsREREZPCYsBAREZHBY8JCREREBo8JC5EeLFmyBBqNJstt6NChqsTcv38/xowZg4cPH6rSfk6k/z6OHj2a26fy1n744QcsWbIkt0+D6IORJ7dPgOhDMm7cOHh7e+vsK1mypCqx9u/fj7Fjx6JTp06wt7dXJcaH7IcffoCTkxM6deqU26dC9EFgwkKkR/Xr10dAQEBun0aOJCQkwMrKKrdPI9ckJibC0tIyt0+D6IPDW0JEBuTPP/9E1apVYWVlBRsbGzRs2BDnzp3TOeb06dPo1KkTfHx8YG5uDldXV3Tu3Bnx8fHaY8aMGYPBgwcDALy9vbW3n6KjoxEdHQ2NRpPl7QyNRoMxY8botKPRaHD+/Hm0bdsWDg4OqFKlivbxn3/+GeXKlYOFhQUcHR3RunVr3Lx5862ee6dOnWBtbY2YmBh8+umnsLa2hoeHB+bMmQMAOHPmDGrVqgUrKyt4enpi+fLlOj+ffptp9+7d+PLLL5E3b17Y2tqiQ4cO+OeffzLF++GHH1CiRAmYmZnB3d0dvXr1ynT7rEaNGihZsiSOHTuGatWqwdLSEsOHD4eXlxfOnTuHXbt2aX+3NWrUAAA8ePAAgwYNQqlSpWBtbQ1bW1vUr18fp06d0ml7586d0Gg0WL16NSZOnIj8+fPD3NwctWvXxtWrVzOd76FDh9CgQQM4ODjAysoKfn5+mDlzps4xFy9exOeffw5HR0eYm5sjICAAGzduzO5/BZFBYg8LkR49evQI9+/f19nn5OQEAFi2bBk6duyIoKAgfPPNN0hMTMTcuXNRpUoVnDhxAl5eXgCAbdu24fr16wgODoarqyvOnTuHBQsW4Ny5czh48CA0Gg2aNWuGy5cvY8WKFfjuu++0MZydnXHv3r1sn3eLFi3g6+uLSZMmQUQAABMnTsSoUaPQsmVLdO3aFffu3cP333+PatWq4cSJE291GyolJQX169dHtWrVMHXqVPzyyy/o3bs3rKysMGLECLRr1w7NmjXDvHnz0KFDB1SqVCnTLbbevXvD3t4eY8aMwaVLlzB37lzcuHFDmyAAaYnY2LFjERgYiB49emiPO3LkCPbt2wcTExNte/Hx8ahfvz5at26NL774Ai4uLqhRowb69OkDa2trjBgxAgDg4uICALh+/TrWr1+PFi1awNvbG3FxcZg/fz6qV6+O8+fPw93dXed8p0yZAiMjIwwaNAiPHj3C1KlT0a5dOxw6dEh7zLZt2/Dpp5/Czc0N/fr1g6urKy5cuIA//vgD/fr1AwCcO3cOlStXhoeHB4YOHQorKyusXr0aTZo0wdq1a9G0adNs/38QGRQhItX99NNPAiDLTUTk33//FXt7e+nWrZvOz8XGxoqdnZ3O/sTExEztr1ixQgDI7t27tfu+/fZbASBRUVE6x0ZFRQkA+emnnzK1A0BGjx6t/X706NECQNq0aaNzXHR0tBgbG8vEiRN19p85c0by5MmTaf+rfh9HjhzR7uvYsaMAkEmTJmn3/fPPP2JhYSEajUZWrlyp3X/x4sVM55reZrly5SQ5OVm7f+rUqQJANmzYICIid+/eFVNTU6lbt66kpKRoj5s9e7YAkMWLF2v3Va9eXQDIvHnzMj2HEiVKSPXq1TPtf/bsmU67Imm/czMzMxk3bpx2344dOwSAFCtWTJKSkrT7Z86cKQDkzJkzIiLy4sUL8fb2Fk9PT/nnn3902k1NTdV+Xbt2bSlVqpQ8e/ZM5/FPPvlEfH19M50n0buGt4SI9GjOnDnYtm2bzgakfYJ++PAh2rRpg/v372s3Y2NjVKhQATt27NC2YWFhof362bNnuH//PipWrAgAOH78uCrn/dVXX+l8v27dOqSmpqJly5Y65+vq6gpfX1+d882url27ar+2t7dHkSJFYGVlhZYtW2r3FylSBPb29rh+/Xqmn+/evbtOD0mPHj2QJ08ebN68GQDw119/ITk5Gf3794eR0f8ugd26dYOtrS02bdqk056ZmRmCg4Pf+PzNzMy07aakpCA+Ph7W1tYoUqRIlv8/wcHBMDU11X5ftWpVANA+txMnTiAqKgr9+/fP1GuV3mP04MEDbN++HS1btsS///6r/f+Ij49HUFAQrly5gr///vuNnwORIeItISI9Kl++fJaDbq9cuQIAqFWrVpY/Z2trq/36wYMHGDt2LFauXIm7d+/qHPfo0SMFz/Z/Xr7tcuXKFYgIfH19szw+Y8KQHebm5nB2dtbZZ2dnh/z582vfnDPuz2psysvnZG1tDTc3N0RHRwMAbty4ASAt6cnI1NQUPj4+2sfTeXh46CQU/yU1NRUzZ87EDz/8gKioKKSkpGgfy5s3b6bjCxYsqPO9g4MDAGif27Vr1wC8fjbZ1atXISIYNWoURo0aleUxd+/ehYeHxxs/DyJDw4SFyACkpqYCSBvH4urqmunxPHn+91Jt2bIl9u/fj8GDB8Pf3x/W1tZITU1FvXr1tO28zstv/OkyvrG+LGOvTvr5ajQa/PnnnzA2Ns50vLW19X+eR1ayaut1++X/x9Oo6eXn/l8mTZqEUaNGoXPnzhg/fjwcHR1hZGSE/v37Z/n/o8RzS2930KBBCAoKyvKYwoULv3F7RIaICQuRAShUqBAAIF++fAgMDHzlcf/88w8iIyMxduxYhIaGaven99Bk9KrEJP0T/MszYl7uWfiv8xUReHt746OPPnrjn9OHK1euoGbNmtrvnzx5gjt37qBBgwYAAE9PTwDApUuX4OPjoz0uOTkZUVFRr/39Z/Sq3++aNWtQs2ZN/Pjjjzr7Hz58qB38nB3pfxtnz5595bmlPw8TE5M3Pn+idw3HsBAZgKCgINja2mLSpEl4/vx5psfTZ/akfxp/+dN3WFhYpp9Jr5XycmJia2sLJycn7N69W2f/Dz/88Mbn26xZMxgbG2Ps2LGZzkVEdKZY69uCBQt0fodz587FixcvUL9+fQBAYGAgTE1NMWvWLJ1z//HHH/Ho0SM0bNjwjeJYWVllWUXY2Ng40+/k119/fesxJGXLloW3tzfCwsIyxUuPky9fPtSoUQPz58/HnTt3MrXxNjPDiAwNe1iIDICtrS3mzp2L9u3bo2zZsmjdujWcnZ0RExODTZs2oXLlypg9ezZsbW21U36fP38ODw8PbN26FVFRUZnaLFeuHABgxIgRaN26NUxMTPDZZ5/BysoKXbt2xZQpU9C1a1cEBARg9+7duHz58hufb6FChTBhwgQMGzYM0dHRaNKkCWxsbBAVFYXffvsN3bt3x6BBgxT7/WRHcnIyateujZYtW+LSpUv44YcfUKVKFTRq1AhA2tTuYcOGYezYsahXrx4aNWqkPe7jjz/GF1988UZxypUrh7lz52LChAkoXLgw8uXLh1q1auHTTz/FuHHjEBwcjE8++QRnzpzBL7/8otObkx1GRkaYO3cuPvvsM/j7+yM4OBhubm64ePEizp07hy1btgBIG9BdpUoVlCpVCt26dYOPjw/i4uJw4MAB3Lp1K1MdGKJ3Ti7NTiL6oGQ1jTcrO3bskKCgILGzsxNzc3MpVKiQdOrUSY4ePao95tatW9K0aVOxt7cXOzs7adGihdy+fTvTNF8RkfHjx4uHh4cYGRnpTHFOTEyULl26iJ2dndjY2EjLli3l7t27r5zWfO/evSzPd+3atVKlShWxsrISKysrKVq0qPTq1UsuXbqU7d9Hx44dxcrKKtOx1atXlxIlSmTa7+npKQ0bNszU5q5du6R79+7i4OAg1tbW0q5dO4mPj8/087Nnz5aiRYuKiYmJuLi4SI8ePTJNG35VbJG0KecNGzYUGxsbAaCd4vzs2TMZOHCguLm5iYWFhVSuXFkOHDgg1atX15kGnT6t+ddff9Vp91XTzvfu3St16tQRGxsbsbKyEj8/P/n+++91jrl27Zp06NBBXF1dxcTERDw8POTTTz+VNWvWZPkciN4lGhE9jFojIlLZkiVLEBwcjCNHjrzzyx8QUWYcw0JEREQGjwkLERERGTwmLERERGTwOIaFiIiIDN5b9bDMmTMHXl5eMDc3R4UKFXD48OFXHrtu3ToEBATA3t4eVlZW8Pf3x7Jly3SO6dSpk3aJ9vStXr16b3NqRERE9B7Kdh2WVatWISQkBPPmzUOFChUQFhaGoKAgXLp0Cfny5ct0vKOjI0aMGIGiRYvC1NQUf/zxB4KDg5EvXz6dEtL16tXDTz/9pP3ezMzsjc8pNTUVt2/fho2NzSurTxIREZFhERH8+++/cHd311mM9FUHZ0v58uWlV69e2u9TUlLE3d1dJk+e/MZtlClTRkaOHKn9vmPHjtK4cePsnorWzZs3BQA3bty4cePG7R3cbt68+Z/v9dnqYUlOTsaxY8cwbNgw7T4jIyMEBgbiwIED//nzIoLt27fj0qVL+Oabb3Qe27lzJ/LlywcHBwfUqlULEyZMyHJlUwBISkpCUlKSTrsAcPPmTZ1VbYmIiMhwPX78GAUKFICNjc1/HputhOX+/ftISUmBi4uLzn4XFxdcvHjxlT/36NEjeHh4ICkpCcbGxvjhhx9Qp04d7eP16tVDs2bN4O3tjWvXrmH48OGoX78+Dhw4kOVKppMnT8bYsWMz7be1tWXCQkRE9I55k+EcellLyMbGBidPnsSTJ08QGRmJkJAQ+Pj4oEaNGgCA1q1ba48tVaoU/Pz8UKhQIezcuRO1a9fO1N6wYcMQEhKi/T49QyMiIqL3U7YSFicnJxgbGyMuLk5nf1xcHFxdXV/5c0ZGRihcuDAAwN/fHxcuXMDkyZO1CcvLfHx84OTkhKtXr2aZsJiZmWVrUC4RERG927I1rdnU1BTlypVDZGSkdl9qaioiIyNRqVKlN24nNTVVZwzKy27duoX4+Hi4ubll5/SIiIjoPZXtW0IhISHo2LEjAgICUL58eYSFhSEhIQHBwcEAgA4dOsDDwwOTJ08GkDbeJCAgAIUKFUJSUhI2b96MZcuWYe7cuQCAJ0+eYOzYsWjevDlcXV1x7do1DBkyBIULF9aZ9kxEREQfrmwnLK1atcK9e/cQGhqK2NhY+Pv7IyIiQjsQNyYmRmcudUJCAnr27Ilbt27BwsICRYsWxc8//4xWrVoBAIyNjXH69GmEh4fj4cOHcHd3R926dTF+/Hje9iEiIiIA70lp/sePH8POzg6PHj3iLCEiIqJ3RHbev7n4IRERERk8JixERERk8JiwEBERkcFjwkJEREQGjwkLERERGTwmLERERGTwmLAQERGRwWPCQkRERAaPCQsREREZvGyX5v8QeQ3dpFrb0VMaqtY2ERHR+4I9LERERGTwmLAQERGRwWPCQkRERAaPCQsREREZPCYsREREZPCYsBAREZHBY8JCREREBo8JCxERERk8JixERERk8JiwEBERkcFjwkJEREQGjwkLERERGTwmLERERGTwmLAQERGRwWPCQkRERAaPCQsREREZPCYsREREZPCYsBAREZHBY8JCREREBo8JCxERERk8JixERERk8JiwEBERkcFjwkJEREQGjwkLERERGTwmLERERGTwmLAQERGRwWPCQkRERAaPCQsREREZPCYsREREZPCYsBAREZHBY8JCREREBo8JCxERERm8t0pY5syZAy8vL5ibm6NChQo4fPjwK49dt24dAgICYG9vDysrK/j7+2PZsmU6x4gIQkND4ebmBgsLCwQGBuLKlStvc2pERET0Hsp2wrJq1SqEhIRg9OjROH78OEqXLo2goCDcvXs3y+MdHR0xYsQIHDhwAKdPn0ZwcDCCg4OxZcsW7TFTp07FrFmzMG/ePBw6dAhWVlYICgrCs2fP3v6ZERER0XtDIyKSnR+oUKECPv74Y8yePRsAkJqaigIFCqBPnz4YOnToG7VRtmxZNGzYEOPHj4eIwN3dHQMHDsSgQYMAAI8ePYKLiwuWLFmC1q1bZ/r5pKQkJCUlab9//PgxChQogEePHsHW1jY7T+eNeA3dpHib6aKnNFStbSIiIkP2+PFj2NnZvdH7d7Z6WJKTk3Hs2DEEBgb+rwEjIwQGBuLAgQP/+fMigsjISFy6dAnVqlUDAERFRSE2NlanTTs7O1SoUOGVbU6ePBl2dnbarUCBAtl5GkRERPSOyVbCcv/+faSkpMDFxUVnv4uLC2JjY1/5c48ePYK1tTVMTU3RsGFDfP/996hTpw4AaH8uO20OGzYMjx490m43b97MztMgIiKid0wefQSxsbHByZMn8eTJE0RGRiIkJAQ+Pj6oUaPGW7VnZmYGMzMzZU+SiIiIDFa2EhYnJycYGxsjLi5OZ39cXBxcXV1f+XNGRkYoXLgwAMDf3x8XLlzA5MmTUaNGDe3PxcXFwc3NTadNf3//7JweERERvaeydUvI1NQU5cqVQ2RkpHZfamoqIiMjUalSpTduJzU1VTto1tvbG66urjptPn78GIcOHcpWm0RERPT+yvYtoZCQEHTs2BEBAQEoX748wsLCkJCQgODgYABAhw4d4OHhgcmTJwNIGyAbEBCAQoUKISkpCZs3b8ayZcswd+5cAIBGo0H//v0xYcIE+Pr6wtvbG6NGjYK7uzuaNGmi3DMlIiKid1a2E5ZWrVrh3r17CA0NRWxsLPz9/REREaEdNBsTEwMjo/913CQkJKBnz564desWLCwsULRoUfz8889o1aqV9pghQ4YgISEB3bt3x8OHD1GlShVERETA3NxcgadIRERE77ps12ExRNmZx/02WIeFiIhIearVYSEiIiLKDUxYiIiIyOAxYSEiIiKDx4SFiIiIDB4TFiIiIjJ4TFiIiIjI4DFhISIiIoPHhIWIiIgMHhMWIiIiMnhMWIiIiMjgMWEhIiIig8eEhYiIiAweExYiIiIyeExYiIiIyOAxYSEiIiKDx4SFiIiIDB4TFiIiIjJ4TFiIiIjI4DFhISIiIoPHhIWIiIgMHhMWIiIiMnhMWIiIiMjgMWEhIiIig8eEhYiIiAweExYiIiIyeExYiIiIyOAxYSEiIiKDx4SFiIiIDB4TFiIiIjJ4TFiIiIjI4DFhISIiIoPHhIWIiIgMHhMWIiIiMnhMWIiIiMjgMWEhIiIig8eEhYiIiAweExYiIiIyeExYiIiIyOAxYSEiIiKDx4SFiIiIDB4TFiIiIjJ4b5WwzJkzB15eXjA3N0eFChVw+PDhVx67cOFCVK1aFQ4ODnBwcEBgYGCm4zt16gSNRqOz1atX721OjYiIiN5D2U5YVq1ahZCQEIwePRrHjx9H6dKlERQUhLt372Z5/M6dO9GmTRvs2LEDBw4cQIECBVC3bl38/fffOsfVq1cPd+7c0W4rVqx4u2dERERE751sJywzZsxAt27dEBwcjOLFi2PevHmwtLTE4sWLszz+l19+Qc+ePeHv74+iRYti0aJFSE1NRWRkpM5xZmZmcHV11W4ODg5v94yIiIjovZOthCU5ORnHjh1DYGDg/xowMkJgYCAOHDjwRm0kJibi+fPncHR01Nm/c+dO5MuXD0WKFEGPHj0QHx//yjaSkpLw+PFjnY2IiIjeX9lKWO7fv4+UlBS4uLjo7HdxcUFsbOwbtfH111/D3d1dJ+mpV68eli5disjISHzzzTfYtWsX6tevj5SUlCzbmDx5Muzs7LRbgQIFsvM0iIiI6B2TR5/BpkyZgpUrV2Lnzp0wNzfX7m/durX261KlSsHPzw+FChXCzp07Ubt27UztDBs2DCEhIdrvHz9+zKSFiIjoPZatHhYnJycYGxsjLi5OZ39cXBxcXV1f+7PTpk3DlClTsHXrVvj5+b32WB8fHzg5OeHq1atZPm5mZgZbW1udjYiIiN5f2UpYTE1NUa5cOZ0Bs+kDaCtVqvTKn5s6dSrGjx+PiIgIBAQE/GecW7duIT4+Hm5ubtk5PSIiInpPZXuWUEhICBYuXIjw8HBcuHABPXr0QEJCAoKDgwEAHTp0wLBhw7THf/PNNxg1ahQWL14MLy8vxMbGIjY2Fk+ePAEAPHnyBIMHD8bBgwcRHR2NyMhING7cGIULF0ZQUJBCT5OIiIjeZdkew9KqVSvcu3cPoaGhiI2Nhb+/PyIiIrQDcWNiYmBk9L88aO7cuUhOTsbnn3+u087o0aMxZswYGBsb4/Tp0wgPD8fDhw/h7u6OunXrYvz48TAzM8vh0yMiIqL3gUZEJLdPIqceP34MOzs7PHr0SJXxLF5DNyneZrroKQ1Va5uIiMiQZef9m2sJERERkcFjwkJEREQGjwkLERERGTwmLERERGTwmLAQERGRwWPCQkRERAaPCQsREREZPCYsREREZPCYsBAREZHBY8JCREREBo8JCxERERk8JixERERk8JiwEBERkcFjwkJEREQGjwkLERERGTwmLERERGTwmLAQERGRwWPCQkRERAaPCQsREREZPCYsREREZPCYsBAREZHBY8JCREREBo8JCxERERk8JixERERk8JiwEBERkcFjwkJEREQGjwkLERERGTwmLERERGTwmLAQERGRwWPCQkRERAaPCQsREREZPCYsREREZPCYsBAREZHBY8JCREREBo8JCxERERk8JixERERk8JiwEBERkcFjwkJEREQGjwkLERERGTwmLERERGTwmLAQERGRwWPCQkRERAbvrRKWOXPmwMvLC+bm5qhQoQIOHz78ymMXLlyIqlWrwsHBAQ4ODggMDMx0vIggNDQUbm5usLCwQGBgIK5cufI2p0ZERETvoWwnLKtWrUJISAhGjx6N48ePo3Tp0ggKCsLdu3ezPH7nzp1o06YNduzYgQMHDqBAgQKoW7cu/v77b+0xU6dOxaxZszBv3jwcOnQIVlZWCAoKwrNnz97+mREREdF7QyMikp0fqFChAj7++GPMnj0bAJCamooCBQqgT58+GDp06H/+fEpKChwcHDB79mx06NABIgJ3d3cMHDgQgwYNAgA8evQILi4uWLJkCVq3bv2fbT5+/Bh2dnZ49OgRbG1ts/N03ojX0E2Kt5kuekpD1domIiIyZNl5/85WD0tycjKOHTuGwMDA/zVgZITAwEAcOHDgjdpITEzE8+fP4ejoCACIiopCbGysTpt2dnaoUKHCK9tMSkrC48ePdTYiIiJ6f2UrYbl//z5SUlLg4uKis9/FxQWxsbFv1MbXX38Nd3d3bYKS/nPZaXPy5Mmws7PTbgUKFMjO0yAiIqJ3jF5nCU2ZMgUrV67Eb7/9BnNz87duZ9iwYXj06JF2u3nzpoJnSURERIYmT3YOdnJygrGxMeLi4nT2x8XFwdXV9bU/O23aNEyZMgV//fUX/Pz8tPvTfy4uLg5ubm46bfr7+2fZlpmZGczMzLJz6kRERPQOy1YPi6mpKcqVK4fIyEjtvtTUVERGRqJSpUqv/LmpU6di/PjxiIiIQEBAgM5j3t7ecHV11Wnz8ePHOHTo0GvbJCIiog9HtnpYACAkJAQdO3ZEQEAAypcvj7CwMCQkJCA4OBgA0KFDB3h4eGDy5MkAgG+++QahoaFYvnw5vLy8tONSrK2tYW1tDY1Gg/79+2PChAnw9fWFt7c3Ro0aBXd3dzRp0kS5Z0pERETvrGwnLK1atcK9e/cQGhqK2NhY+Pv7IyIiQjtoNiYmBkZG/+u4mTt3LpKTk/H555/rtDN69GiMGTMGADBkyBAkJCSge/fuePjwIapUqYKIiIgcjXMhIiKi90e267AYItZhISIieveoVoeFiIiIKDcwYSEiIiKDx4SFiIiIDB4TFiIiIjJ4TFiIiIjI4DFhISIiIoPHhIWIiIgMXrYLx5F+qFX7hXVfiIjoXcQeFiIiIjJ4TFiIiIjI4DFhISIiIoPHhIWIiIgMHhMWIiIiMnhMWIiIiMjgMWEhIiIig8eEhYiIiAweExYiIiIyeExYiIiIyOAxYSEiIiKDx4SFiIiIDB4XPyQA6i22CHDBRSIiyjn2sBAREZHBY8JCREREBo8JCxERERk8JixERERk8JiwEBERkcFjwkJEREQGjwkLERERGTwmLERERGTwmLAQERGRwWPCQkRERAaPCQsREREZPCYsREREZPCYsBAREZHBY8JCREREBo8JCxERERk8JixERERk8JiwEBERkcFjwkJEREQGjwkLERERGTwmLERERGTwmLAQERGRwXurhGXOnDnw8vKCubk5KlSogMOHD7/y2HPnzqF58+bw8vKCRqNBWFhYpmPGjBkDjUajsxUtWvRtTo2IiIjeQ9lOWFatWoWQkBCMHj0ax48fR+nSpREUFIS7d+9meXxiYiJ8fHwwZcoUuLq6vrLdEiVK4M6dO9pt79692T01IiIiek9lO2GZMWMGunXrhuDgYBQvXhzz5s2DpaUlFi9enOXxH3/8Mb799lu0bt0aZmZmr2w3T548cHV11W5OTk6vPDYpKQmPHz/W2YiIiOj9la2EJTk5GceOHUNgYOD/GjAyQmBgIA4cOJCjE7ly5Qrc3d3h4+ODdu3aISYm5pXHTp48GXZ2dtqtQIECOYpNREREhi1bCcv9+/eRkpICFxcXnf0uLi6IjY1965OoUKEClixZgoiICMydOxdRUVGoWrUq/v333yyPHzZsGB49eqTdbt68+daxiYiIyPDlye0TAID69etrv/bz80OFChXg6emJ1atXo0uXLpmONzMze+3tJSIiInq/ZKuHxcnJCcbGxoiLi9PZHxcX99oBtdllb2+Pjz76CFevXlWsTSIiInp3ZSthMTU1Rbly5RAZGandl5qaisjISFSqVEmxk3ry5AmuXbsGNzc3xdokIiKid1e2bwmFhISgY8eOCAgIQPny5REWFoaEhAQEBwcDADp06AAPDw9MnjwZQNpA3fPnz2u//vvvv3Hy5ElYW1ujcOHCAIBBgwbhs88+g6enJ27fvo3Ro0fD2NgYbdq0Uep5EhER0Tss2wlLq1atcO/ePYSGhiI2Nhb+/v6IiIjQDsSNiYmBkdH/Om5u376NMmXKaL+fNm0apk2bhurVq2Pnzp0AgFu3bqFNmzaIj4+Hs7MzqlSpgoMHD8LZ2TmHT4+IiIjeB2816LZ3797o3bt3lo+lJyHpvLy8ICKvbW/lypVvcxpERET0geBaQkRERGTwmLAQERGRwWPCQkRERAaPCQsREREZPCYsREREZPCYsBAREZHBY8JCREREBo8JCxERERk8JixERERk8JiwEBERkcFjwkJEREQGjwkLERERGTwmLERERGTwmLAQERGRwWPCQkRERAaPCQsREREZPCYsREREZPCYsBAREZHBY8JCREREBo8JCxERERk8JixERERk8JiwEBERkcFjwkJEREQGjwkLERERGTwmLERERGTwmLAQERGRwWPCQkRERAaPCQsREREZPCYsREREZPCYsBAREZHBY8JCREREBo8JCxERERk8JixERERk8JiwEBERkcFjwkJEREQGjwkLERERGTwmLERERGTwmLAQERGRwWPCQkRERAaPCQsREREZPCYsREREZPDeKmGZM2cOvLy8YG5ujgoVKuDw4cOvPPbcuXNo3rw5vLy8oNFoEBYWluM2iYiI6MOS7YRl1apVCAkJwejRo3H8+HGULl0aQUFBuHv3bpbHJyYmwsfHB1OmTIGrq6sibRIREdGHJdsJy4wZM9CtWzcEBwejePHimDdvHiwtLbF48eIsj//444/x7bffonXr1jAzM1OkTSIiIvqwZCthSU5OxrFjxxAYGPi/BoyMEBgYiAMHDrzVCbxNm0lJSXj8+LHORkRERO+vbCUs9+/fR0pKClxcXHT2u7i4IDY29q1O4G3anDx5Muzs7LRbgQIF3io2ERERvRveyVlCw4YNw6NHj7TbzZs3c/uUiIiISEV5snOwk5MTjI2NERcXp7M/Li7ulQNq1WjTzMzsleNhiIiI6P2TrR4WU1NTlCtXDpGRkdp9qampiIyMRKVKld7qBNRok4iIiN4v2ephAYCQkBB07NgRAQEBKF++PMLCwpCQkIDg4GAAQIcOHeDh4YHJkycDSBtUe/78ee3Xf//9N06ePAlra2sULlz4jdokIiKiD1u2E5ZWrVrh3r17CA0NRWxsLPz9/REREaEdNBsTEwMjo/913Ny+fRtlypTRfj9t2jRMmzYN1atXx86dO9+oTXo/eQ3dpEq70VMaqtIuERHlnmwnLADQu3dv9O7dO8vH0pOQdF5eXhCRHLVJREREH7Z3cpYQERERfViYsBAREZHBY8JCREREBo8JCxERERk8JixERERk8JiwEBERkcFjwkJEREQGjwkLERERGTwmLERERGTwmLAQERGRwWPCQkRERAaPCQsREREZPCYsREREZPCYsBAREZHBY8JCREREBo8JCxERERk8JixERERk8JiwEBERkcFjwkJEREQGjwkLERERGTwmLERERGTwmLAQERGRwWPCQkRERAaPCQsREREZPCYsREREZPCYsBAREZHBY8JCREREBo8JCxERERk8JixERERk8JiwEBERkcHLk9snQKQvXkM3qdZ29JSGqrVNRETsYSEiIqJ3ABMWIiIiMnhMWIiIiMjgMWEhIiIig8eEhYiIiAweExYiIiIyeExYiIiIyOAxYSEiIiKDx4SFiIiIDB4TFiIiIjJ4TFiIiIjI4L1VwjJnzhx4eXnB3NwcFSpUwOHDh197/K+//oqiRYvC3NwcpUqVwubNm3Ue79SpEzQajc5Wr169tzk1IiIieg9lO2FZtWoVQkJCMHr0aBw/fhylS5dGUFAQ7t69m+Xx+/fvR5s2bdClSxecOHECTZo0QZMmTXD27Fmd4+rVq4c7d+5otxUrVrzdMyIiIqL3TrYTlhkzZqBbt24IDg5G8eLFMW/ePFhaWmLx4sVZHj9z5kzUq1cPgwcPRrFixTB+/HiULVsWs2fP1jnOzMwMrq6u2s3BweGV55CUlITHjx/rbERERPT+ylbCkpycjGPHjiEwMPB/DRgZITAwEAcOHMjyZw4cOKBzPAAEBQVlOn7nzp3Ily8fihQpgh49eiA+Pv6V5zF58mTY2dlptwIFCmTnaRAREdE7JlsJy/3795GSkgIXFxed/S4uLoiNjc3yZ2JjY//z+Hr16mHp0qWIjIzEN998g127dqF+/fpISUnJss1hw4bh0aNH2u3mzZvZeRpERET0jsmT2ycAAK1bt9Z+XapUKfj5+aFQoULYuXMnateunel4MzMzmJmZ6fMUiYiIKBdlq4fFyckJxsbGiIuL09kfFxcHV1fXLH/G1dU1W8cDgI+PD5ycnHD16tXsnB4RERG9p7KVsJiamqJcuXKIjIzU7ktNTUVkZCQqVaqU5c9UqlRJ53gA2LZt2yuPB4Bbt24hPj4ebm5u2Tk9IiIiek9le5ZQSEgIFi5ciPDwcFy4cAE9evRAQkICgoODAQAdOnTAsGHDtMf369cPERERmD59Oi5evIgxY8bg6NGj6N27NwDgyZMnGDx4MA4ePIjo6GhERkaicePGKFy4MIKCghR6mkRERPQuy/YYllatWuHevXsIDQ1FbGws/P39ERERoR1YGxMTAyOj/+VBn3zyCZYvX46RI0di+PDh8PX1xfr161GyZEkAgLGxMU6fPo3w8HA8fPgQ7u7uqFu3LsaPH89xKkRERATgLQfd9u7dW9tD8rKdO3dm2teiRQu0aNEiy+MtLCywZcuWtzkNIiIi+kBwLSEiIiIyeExYiIiIyOAxYSEiIiKDx4SFiIiIDB4TFiIiIjJ4TFiIiIjI4DFhISIiIoPHhIWIiIgMHhMWIiIiMnhMWIiIiMjgMWEhIiIig8eEhYiIiAweExYiIiIyeExYiIiIyOAxYSEiIiKDx4SFiIiIDB4TFiIiIjJ4TFiIiIjI4DFhISIiIoOXJ7dPgOh95jV0k2ptR09pqFrbRESGhj0sREREZPCYsBAREZHB4y0hovcIb0ER0fuKPSxERERk8NjDQkQ5wl4dItIH9rAQERGRwWPCQkRERAaPCQsREREZPCYsREREZPCYsBAREZHB4ywhInqncFYS0YeJPSxERERk8JiwEBERkcHjLSEiov+g79tQuXHb60N4jvRuYw8LERERGTwmLERERGTwmLAQERGRweMYFiIi+iBw3My7jQkLERGRStRKkj7EBIkJCxER0Xvife5F4hgWIiIiMnhvlbDMmTMHXl5eMDc3R4UKFXD48OHXHv/rr7+iaNGiMDc3R6lSpbB582adx0UEoaGhcHNzg4WFBQIDA3HlypW3OTUiIiJ6D2U7YVm1ahVCQkIwevRoHD9+HKVLl0ZQUBDu3r2b5fH79+9HmzZt0KVLF5w4cQJNmjRBkyZNcPbsWe0xU6dOxaxZszBv3jwcOnQIVlZWCAoKwrNnz97+mREREdF7I9sJy4wZM9CtWzcEBwejePHimDdvHiwtLbF48eIsj585cybq1auHwYMHo1ixYhg/fjzKli2L2bNnA0jrXQkLC8PIkSPRuHFj+Pn5YenSpbh9+zbWr1+foydHRERE74dsDbpNTk7GsWPHMGzYMO0+IyMjBAYG4sCBA1n+zIEDBxASEqKzLygoSJuMREVFITY2FoGBgdrH7ezsUKFCBRw4cACtW7fO1GZSUhKSkpK03z969AgA8Pjx4+w8nTeWmpSoSrvAq89ZrZj6jpcbMT+E5/iqmHyOysfLjZh8jsrHe99iGtJzVKJNEfnvgyUb/v77bwEg+/fv19k/ePBgKV++fJY/Y2JiIsuXL9fZN2fOHMmXL5+IiOzbt08AyO3bt3WOadGihbRs2TLLNkePHi0AuHHjxo0bN27vwXbz5s3/zEHeyWnNw4YN0+m1SU1NxYMHD5A3b15oNJpcO6/Hjx+jQIECuHnzJmxtbRnzHY2XGzH5HN+PmHyO70fMD+E55lbMl4kI/v33X7i7u//nsdlKWJycnGBsbIy4uDid/XFxcXB1dc3yZ1xdXV97fPq/cXFxcHNz0znG398/yzbNzMxgZmams8/e3j47T0VVtra2ev/P/xBi8jm+HzH5HN+PmHyOjKkUOzu7NzouW4NuTU1NUa5cOURGRmr3paamIjIyEpUqVcryZypVqqRzPABs27ZNe7y3tzdcXV11jnn8+DEOHTr0yjaJiIjow5LtW0IhISHo2LEjAgICUL58eYSFhSEhIQHBwcEAgA4dOsDDwwOTJ08GAPTr1w/Vq1fH9OnT0bBhQ6xcuRJHjx7FggULAAAajQb9+/fHhAkT4OvrC29vb4waNQru7u5o0qSJcs+UiIiI3lnZTlhatWqFe/fuITQ0FLGxsfD390dERARcXFwAADExMTAy+l/HzSeffILly5dj5MiRGD58OHx9fbF+/XqULFlSe8yQIUOQkJCA7t274+HDh6hSpQoiIiJgbm6uwFPUHzMzM4wePTrT7SrGfLfi5UZMPsf3Iyaf4/sR80N4jrkVMyc0Im8yl4iIiIgo93AtISIiIjJ4TFiIiIjI4DFhISIiIoPHhIWIiIgMHhMWBXF1aSIiInW8k6X5DUlqaiomTpyIefPmIS4uDpcvX4aPjw9GjRoFLy8vdOnSJbdPMcdOnz6d5X6NRgNzc3MULFhQ8WlxDx8+xJo1a3Dt2jUMHjwYjo6OOH78OFxcXODh4aFoLCIigNcdQ8eEJYcmTJiA8PBwTJ06Fd26ddPuL1myJMLCwhRPWF61WqZGo4GZmRlMTU0VjQcA/v7+r12jycTEBK1atcL8+fMVqZ1z+vRpBAYGws7ODtHR0ejWrRscHR2xbt06xMTEYOnSpTmOQe8nHx8fHDlyBHnz5tXZ//DhQ5QtWxbXr19X/RweP36M7du3o0iRIihWrJhqcVJTU3H16lXcvXsXqampOo9Vq1Ytx+1nZ2Xe3CzrrhRedwwf67DkUOHChTF//nzUrl0bNjY2OHXqFHx8fHDx4kVUqlQJ//zzj6LxjIyMXps85M+fH506dcLo0aN1CvjlxIYNG/D1119j8ODBKF++PADg8OHDmD59OkaPHo0XL15g6NChaNWqFaZNm5bjeIGBgShbtiymTp2q8zvdv38/2rZti+jo6BzHyIohvNm9LzZu3PjGxzZq1EixuEZGRoiNjUW+fPl09sfFxaFgwYJISkpSLFa6li1bolq1aujduzeePn2K0qVLIzo6GiKClStXonnz5orHPHjwINq2bYsbN27g5Uu4RqNBSkpKjmP817UmIyXiZfT8+XNYWFjg5MmTOkVG1ZQb152UlBQsWbIEkZGRWSae27dvVzzm0aNHsXr1asTExCA5OVnnsXXr1ikeT0nsYcmhv//+G4ULF860PzU1Fc+fP1c83pIlSzBixAh06tRJJ3kIDw/HyJEjce/ePUybNg1mZmYYPny4IjEnTpyImTNnIigoSLuvVKlSyJ8/P0aNGoXDhw/DysoKAwcOVCRhOXLkCObPn59pv4eHB2JjY3Pc/qtER0dneeFNSkrC33//rXi8hIQETJky5ZUXKzUSJH3FfNNlNZR6c82YIG3ZskVnMbWUlBRERkbCy8srx3Gysnv3bowYMQIA8Ntvv0FE8PDhQ4SHh2PChAmqJCxfffUVAgICsGnTJri5uamySv2OHTu0X0dHR2Po0KHo1KmTdo23AwcOIDw8XLsMi5JMTExQsGBBxROh18mN606/fv2wZMkSNGzYECVLllTl/zGjlStXokOHDggKCsLWrVtRt25dXL58GXFxcWjatKmqsZXAhCWHihcvjj179sDT01Nn/5o1a1CmTBnF44WHh2P69Olo2bKldt9nn32GUqVKYf78+YiMjETBggUxceJExRKWM2fOZHp+AODp6YkzZ84ASLttdOfOHUXimZmZZdkdffnyZTg7OysSI6PcerPr2rUrdu3ahfbt26v2ppNbMV9OhNSWniBpNBp07NhR5zETExN4eXlh+vTpqsR+9OgRHB0dAQARERFo3rw5LC0t0bBhQwwePFiVmFeuXMGaNWuy/LCklOrVq2u/HjduHGbMmIE2bdpo9zVq1AilSpXCggULMv3OlTBixAgMHz4cy5Yt0/5+1aTv6w6QlkCsXr0aDRo0UKX9l02aNAnfffcdevXqBRsbG8ycORPe3t748ssv4ebmppdzyBGhHFm/fr3Y2dnJlClTxNLSUr799lvp2rWrmJqaytatWxWPZ25uLpcvX860//Lly2JhYSEiItevX9d+rQR/f3/p2LGjJCUlafclJydLx44dxd/fX0RE9u7dK15eXorE69KlizRp0kSSk5PF2tparl+/Ljdu3JAyZcpIv379FImRkUajEY1GI0ZGRtqv0zdTU1P56KOP5Pfff1c8rp2dnezdu1fxdg0tZkZPnz5VtX0vLy+5d++eqjFe5uvrK6tWrZInT56Is7OzREZGiojIyZMnJW/evKrErFmzpvz555+qtJ0VCwuLLK87ly5dUvRak5G/v79YW1uLmZmZfPTRR1KmTBmdTWn6vu6IiLi5ucmlS5dUaTsrlpaWEhUVJSIijo6Ocvr0aREROX/+vLi6uurtPN4We1hyqHHjxvj9998xbtw4WFlZITQ0FGXLlsXvv/+OOnXqKB6vQIEC+PHHHzFlyhSd/T/++CMKFCgAAIiPj4eDg4NiMefMmYNGjRohf/788PPzA5DW65KSkoI//vgDQNqthJ49eyoSb/r06fj888+RL18+PH36FNWrV0dsbCwqVaqEiRMnKhIjo/TeAG9vbxw5cgROTk6Kx8iKg4ODXj455nbMlJQUTJo0SS8z6aKiohRr6031798f7dq1g7W1NTw9PVGjRg0AabeKSpUqpUrMPn36YODAgYiNjUWpUqVgYmKi83j661QpBQoUwMKFCzF16lSd/YsWLdJed5T2prcVlaLv6w4ADBw4EDNnzsTs2bP10sPq4OCAf//9F0Dara6zZ8+iVKlSePjwIRITE1WPn1McdPuO2bhxI1q0aIGiRYvi448/BpA2iOrixYtYs2YNPv30U8ydOxdXrlzBjBkzFIv777//4pdffsHly5cBAEWKFEHbtm1hY2OjWIyX7du3D6dOncKTJ09QtmxZBAYGqhbrVR4+fAh7e3tV2v7555+xYcMGhIeHw9LSUpUYhhBz3LhxCA8Px7hx49CtWzecPXsWPj4+WLVqFcLCwnDgwAFFY71OaGioYrEyOnr0KG7evIk6derA2toaALBp0ybY29ujcuXKisfLakC9RqOBiCg2LiijzZs3o3nz5ihcuDAqVKgAIG3s3JUrV7B27Vq93dLQh7179+L06dN6ue40bdoUO3bsgKOjI0qUKJEp8VR6EGzbtm0REBCAkJAQjB8/Ht9//z0aN26Mbdu2oWzZsgY/6JYJyzsoKioK8+fP10kevvzyS9UGFX4ovvnmG3h5eaFVq1YAgBYtWmDt2rVwc3PD5s2bUbp06RzHKFOmjM4nqatXr0JE4OXllelidfz48RzHyyr+tWvX9BpTnzPpXh439vz5c0RFRSFPnjwoVKiQKs8vN9y4ceO1j2c15iynbt26hR9++AEXL14EABQrVgxfffWVaj0swPtfFyU4OPi1j//000+Kxnvw4AGePXsGd3d3pKamYurUqdi/fz98fX0xcuRIRXvm1cCE5S04ODi8cffdgwcPVD4b/bhy5Qp27NiR5cwSpT+19u3bF4ULF0bfvn119s+ePRtXr15FWFiYovHSeXt745dffsEnn3yCbdu2oWXLlli1apV2CuDWrVtzHGPs2LFvfOzo0aNzHC+78dWIaWFhgYsXL8LT01MnYTl//jzKly+PJ0+eKB4zo8ePH6NTp05o2rQp2rdvr3j7uTE19UPwcl2US5cuwcfHByNHjlSsLsqsWbPe+NiXr0ekf0xY3kJ4ePgbH6vG6PmHDx/i8OHDWV4cO3TooHi8hQsXokePHnBycoKrq6tOsqbRaBT/1Orh4YGNGzeiXLlyOvuPHz+ORo0a4datW4rGS2dhYYHLly+jQIEC6NevH549e6btyapQoYLiNXU+FOXKlcOAAQPwxRdf6CQs48aNw7Zt27Bnzx7Vz+HMmTP47LPPVKml0bt3b+3U1KxmXn333XeKx0x3/vz5LOtpKFnbJt3Dhw/x448/4sKFCwCAEiVKoHPnzjqz6pSkj7oo3t7eb3ScRqN5L+ow1apVC9WrV8/0weSff/5B8+bNDT+5zqXBvvSWNm7cKDY2NqLRaMTOzk7s7e21m4ODgyoxCxYsKFOmTFGl7ayYmZnJlStXMu2/cuWKmJmZqRbXzc1N9u3bJyIiH330kaxevVpERC5evCg2NjaKxzt8+LAcPHgw0/6DBw/KkSNHFI+X0dGjR2XZsmWybNkyOX78uKqx9D2TLit79uwRe3t7VdrOmzevbNq0SZW2X+XatWvi5+eXaXabkZGRGBkZKR7vyJEj4ujoKB4eHtK0aVNp2rSp5M+fX/LmzSvHjh1TPJ6IiK2trVy9elVERKytreXatWsiIhIdHa3qdUDffv31V2nRooVUqFBB9ZlQGo1GnJycpHHjxvLkyRPt/tjYWFX+bpTGhEUBL168kF9//VXGjRsn48aNkzVr1sjz589VieXr6yv9+vWThIQEVdrPio2NjfZioQ8lSpSQ77//PtP+WbNmSbFixVSL26tXL/H09JTAwEDJmzev/PvvvyIismLFClUuHh9//LH8+uuvmfavXbtWypcvr3g8EZG4uDipWbOmaDQacXBwEAcHB9FoNFKrVi25e/euKjFFRHbv3i2BgYHi7OwsFhYWUrlyZdmyZYvicWbOnKmzhYWFyddffy3u7u7Spk0bxeOJ6H9qqojIp59+Ko0bN5Z79+6JtbW1nD9/Xvbs2SPly5eX3bt3Kx6vSpUq0qlTJ53r2vPnz6Vjx45StWpVxeOJiDg7O2uT6YwJy9atWyV//vyqxEyXmpoqqampqsYQSft7tba2lt69e4upqal8+eWXEhgYKHZ2djJ8+HDF42k0Gjl58qRUqFBBSpYsqZ3izITlA3H27Fnx8fERS0tLbVZsZWUlXl5ecubMGcXjWVpa6jV5EBHp3LmzzJ07V2/xfvzxR7GwsJDQ0FDZuXOn7Ny5U0aNGiWWlpayYMEC1eImJyfLt99+K3379tXpdZgxY4YsXLhQ8XhWVlZZ/l9ev35drK2tFY8nItKyZUsJCAiQ8+fPa/edO3dOAgICpHXr1qrE1CcvLy+dzcfHRypUqCDDhg2Tx48fqxJz2rRp0rNnT728waXLmzevnDp1SkTSeiIuXrwoIiKRkZHa2khKMjc3lwsXLmTaf+7cOdXqsORGXZRFixZJiRIlxNTUVExNTaVEiRKqvPbTFSlSRJYvXy4iuknZqFGjpFevXorH02g0EhcXJ8+ePZM2bdqIk5OT7NixgwnLh6JixYry2WefyYMHD7T7Hjx4II0aNZJKlSopHq9p06ayatUqxdt9nUmTJomTk5N07NhRpk2blulTrBp++OEH8fDw0HZ1e3t7S3h4uCqxcoujo6Ps378/0/59+/apdvvC1tZWDh8+nGn/oUOHxM7OTpWYXbp0kR07dqjStiFo0qSJ2NnZibe3t3z66afaWybpmxrs7e3l+vXrIiLi4+Mj27dvFxGRq1evqpJA5MuXL8sesYiICMmXL5/i8UREHj58KIGBgWJvby/GxsZSoEABMTExkWrVqunczlDKqFGjxMrKSoYOHSobNmyQDRs2yNChQ8Xa2lpGjRqleDyRtIJ80dHRIpLWo3Ty5EkRSSsE6ujoqHg8IyMjiYuL034/fvx4MTMzk9DQ0HciYWHhuBw6efIkjh49qjMdzMHBARMnTtTWSVFSernv8+fPZ1kwSo3BdgsWLIC1tTV27dqFXbt26Tym0WhUGT3fo0cP9OjRA/fu3YOFhYW2toU+6GsgY926dTFs2DBs2LBBO3Dx4cOHGD58uCpFB4G0Inkv/80AaeXr1Sqnf+/ePdSrVw/Ozs5o3bo12rVrB39/f1ViZZQ+ODt//vyqxrG3t9f7OiwlS5bEqVOn4O3tjQoVKmDq1KkwNTXFggUL4OPjo3i8Vq1aoUuXLpg2bRo++eQTAGl1kgYPHqxTrl9JdnZ22LZtm97qosydOxcLFy7MtPyAn58f+vTp8581ft6Gq6srHjx4AE9PTxQsWBAHDx5E6dKlERUVlWlRSyW83ObIkSNRrFgxVSaHqCK3M6Z3nZ+fn7YUd0aRkZFSsmRJxeO9XDo+4/YuZMiGTN8DGW/duiU+Pj5iZ2cnNWrUkBo1aoi9vb0UKVJEYmJiFI8nItKoUSOpVq2a/P333zrnUb16dWnSpIkqMUXSeh3nz58v1atXFyMjIylevLhMnDhRew9dKSkpKTJ27FixtbXV/r/Z2dnJuHHjJCUlRdFYuSkiIkLWrl0rImmD0YsUKaIdUJnV9SinkpKSpG/fvmJqaqr9vZqZmUn//v3l2bNnisfLDXZ2dq9cfkDN3scxY8aIiMjs2bPFwsJC26vUuXNnxeNFR0dneevy7NmzsmTJEsXjKY3Tmt9CxgWy9u7diyFDhmDMmDGoWLEigLSl38eNG4cpU6a8VxUg1VS2bFlERkbCwcEhU3G1l6lV/Ouzzz6DsbExFi1aBG9vbxw+fBjx8fHaVairVq2qeMyEhAT88ssvOHXqFCwsLODn54c2bdpk2QuihJs3b6JRo0Y4d+6ctuDXzZs3UbJkSWzcuFH13gggredjxYoVWLx4Ma5cuYIXL14o1vawYcPw448/YuzYsdoKs3v37sWYMWPQrVs31UqsA2k9SZcuXQKQVsxRrQXzXuXBgwfZqhH1NhITE3Ht2jUAQKFChVSvlhwZGfnK+jaLFy9WNFafPn1gYmKSqUL4oEGD8PTpU8yZM0fReEBaj2dqairy5Em72bFy5UptIbcvv/wSpqamisZbunQpAgICULx4cZ39z549w+rVq1Upi6EkJixvwcjISOeikP4rTN+X8Xt9Lo+upPTSzVZWVggJCXntsUosATB27FgMHjwYlpaWGDNmzGsvumoUNwMAJycnbN++HX5+frCzs8Phw4dRpEgRbN++HQMHDsSJEydUiatvIoK//vpLp2KpvpY9eP78OTZt2oSff/4ZmzZtgqOjI/7++2/F2nd3d8e8efMy3b7bsGEDevbsqWisdAkJCejTpw+WLl2qfVM1NjZGhw4d8P3336v6pn716lVcu3YN1apVg4WFhbY0v5r0datt7NixGDduHAICArKsb/Pbb7/lOEbGa9uLFy+wZMkSFCxYUPvh89ChQ4iJidH+X77rjIyMYGVlhSVLlqB58+ba/XFxcXB3dzf49yuOYXkLO3bs0Gu8WbNmoXv37jA3N//PyoxKjSc5ceIEnj9/rv1abRmTkDFjxqgeLyspKSnatZGcnJxw+/ZtFClSBJ6entpPzkrTZwXhdBqNBnXq1FFtnExWduzYgeXLl2Pt2rVITU1Fs2bN8Mcff6BWrVqKxnnw4AGKFi2aaX/RokVVqzodEhKCXbt24ffff9fp1enbty8GDhyIuXPnKh4zPj4eLVu2xI4dO6DRaHDlyhX4+PigS5cucHBwwPTp0xWNl5qaigkTJmD69OnaysQ2NjYYOHAgRowYkeXaRjk1b948LFmyRJXqxOlevralF6tM70VycnKCk5MTzp07p1jM06dPo2TJkjAyMsLp06dfe6zSi1gCaYlg+/btcebMGZ1r7bvQd8EelneAt7c3jh49irx58762MuP7Uo2xa9eu+OKLL7Sr3upL1apVMXDgQDRp0gRt27bFP//8g5EjR2LBggU4duwYzp49q2g8fVUQzo2ENyMPDw88ePAA9erVQ7t27fDZZ5/BzMxM8TgAUKFCBVSoUCHT8+zTpw+OHDmCgwcPKh7TyckJa9asyfT3umPHDrRs2RL37t1TPGaHDh1w9+5dLFq0CMWKFdNWgd2yZQtCQkIUfYMFcudWW968eXH48GEUKlRI8bZzk5GREWJjY5EvXz5tb31Wb8Nq9NCnx75+/TqaNm2KypUrY9myZXj8+PE70cPCQbcKSUhIkAsXLsipU6d0tvdBcHBwljUsnjx5IsHBwYrHa9SokZiZmUn+/Pll0KBB2ql+atP3QEZ9VRD28vKS+/fva79+1ebt7a1K/AULFsg///yjStsv27lzp1hZWUmxYsWkc+fO0rlzZylWrJhYW1urUlBNJG1qasa6NunOnj0rlpaWqsR0cXHRvi4y1u+4du2aWFlZKR7Pzc1NNmzYkGn/+vXrxd3dXfF4IiJDhgyRcePGqdJ2bso48DU6Ovq1m9IyTmu+ceOGlC5dWvz9/eXgwYPvxKQN9rDk0L179xAcHIw///wzy8eVzljHjRuHQYMGZbov/vTpU3z77beq3EYwNjbGnTt3kC9fPp399+/fh6urq6KDJtP9888/+PXXX7F8+XLs2bMHRYsWRbt27dC2bVtFV6XO2D2bFTUHMtra2uLkyZOqTEM1RPoab3H79m3MmTNHZ4xOz5494e7urngsAKhduzby5s2LpUuXwtzcHEDa67Fjx4548OAB/vrrL8Vj2tjY4Pjx4/D19dVZZ+fo0aMICgpCfHy8ovHMzc1x+vRpfPTRRzr7L126BH9/fzx9+lTReADQr18/LF26FH5+fvDz88s0EF2JsXPNmjXDkiVLYGtri2bNmr322HXr1uU4Xm7L2LsDpA2ibteuHSIjI5GQkGDwPSxMWHKoXbt2uHHjBsLCwlCjRg389ttviIuL097vbdiwoaLxXpU8xMfHI1++fIr+wT1+/BgiAgcHB1y5ckVn1kNKSgp+//13DB06FLdv31YsZlbUnFWS8ffp4+ODI0eOIG/evIq1/zpdunTBxx9/jK+++kov8XLLq8ZbdO7cWZXxFvp29uxZBAUFISkpCaVLlwYAnDp1Cubm5tiyZQtKlCiheMwGDRqgXLlyGD9+PGxsbHD69Gl4enqidevWSE1NxZo1axSNlxu32mrWrPnKxzQajSIL9QUHB2PWrFmwsbFBcHDwa4/96aefchwPADZu3PjGxypd+ynj5IaMRo8ejd27d+t9fGZ2MWHJITc3N2zYsAHly5eHra0tjh49io8++ggbN27E1KlTsXfvXkXjGRkZIS4uLtOUye3bt6NVq1aK3i9/eTbUyzQaDcaOHYsRI0YoFvNlas8qyZs3LzZv3owKFSq88nerlsmTJ2PGjBlo2LBhlkUA1RhP8qoZXxqNBubm5ihcuDAaN24MR0dHxWLqe7zFnj17MH/+fFy/fh2//vorPDw8sGzZMnh7e6NKlSqKxkqXmJiIX375RadXp127drCwsFAl3tmzZ1G7dm2ULVsW27dv105Vf/DgAfbt26f4uI9du3ahYcOGKFiwICpVqgQAOHDgAG7evInNmzerMuX/ffWmA5Tf5VmmamHCkkO2trY4ffo0vLy84OnpieXLl6Ny5cqIiopCiRIlkJiYqEic9NsSjx49gq2trU4ikZKSgidPnuCrr75StFbArl27ICKoVasW1q5dq/MmZmpqCk9PT9W62bOaVdKuXTvUqlVL0dsI3bt3x9KlS+Hm5oaYmBjkz58fxsbGWR6r9IDm3BhAXbNmTRw/fhwpKSkoUqQIAODy5cswNjZG0aJFcenSJWg0GuzduzdTrYa35erqii1btqB06dI6ty+uX78OPz8/7awTJaxduxbt27dHu3btsGzZMpw/fx4+Pj6YPXs2Nm/ejM2bNysWK7c9evQIs2fPxqlTp7RVYHv16gU3NzdV4un7Vpu+PX36FCKi7X24ceMGfvvtNxQvXhx169bN5bN7exs3bkT9+vVhYmLy2t4djUaDzz77TI9nln1MWHLo448/xoQJExAUFIRGjRrB3t4ekydPxqxZs7BmzRrt9LicCg8Ph4igc+fOCAsL05ZyB9KSBy8vL+0nH6XduHEDBQoUUGXqYlb0OasEACIiInD16lX07dsX48aN005tflm/fv1UOwd9CQsLw549e/DTTz/B1tYWQNobX9euXVGlShV069YNbdu2xdOnT7FlyxZFYupzvEWZMmUwYMAAdOjQQSfWiRMnUL9+fcTGxioS503fBAB1lsvQh4zjO5YuXYpWrVqp+jp8WUJCAqZMmfLKwnFKJ/R169ZFs2bN8NVXX+Hhw4coUqQITE1Ncf/+fcyYMQM9evRQNJ6+vDwr6VXehR4dJiw59PPPP+PFixfo1KkTjh07hnr16uHBgwcwNTXFkiVL0KpVK0Xj7dq1C5988olqlVBfJzExMcs1dpSuFbBw4UK0aNEC9vb2irb7XzLez35feXh4YNu2bZl6T86dO4e6devi77//xvHjx1G3bl3cv39fkZj6HG9haWmJ8+fPw8vLK1NvTvHixfHs2TNF4hjCm8DDhw9x+PDhLN/MlahYampqihs3bsDNze2VY+fU1KZNG+zatQvt27fPsnCc0h8gnJycsGvXLpQoUQKLFi3C999/jxMnTmDt2rUIDQ3FhQsXFI2XLiEhAbt27cry2qrGbeF3GQvH5dAXX3yh/bpcuXK4ceMGLl68iIIFC8LJyUnxeNWrV9d+/ezZs0x/4OmfmpWkz5lQz58/R48ePVCpUiW9Jyzpn8qzMmfOHPTq1UvReJ07d37t40qXHgfSelPu3r2bKWG5d++edskJe3v7TH9XOTF16lTUrl0bR48eRXJyMoYMGaIz3kJJrq6uuHr1aqaZZHv37lV0NlbGBEGtRSNf5/fff0e7du3w5MmTTLeINRqNIglL0aJFMWzYMNSsWRMigtWrV7/y+qJGSfc///wTmzZt0tZ9UVtiYqL29b9161Y0a9YMRkZGqFixIm7cuKFKzBMnTqBBgwZITExEQkICHB0dcf/+fVhaWiJfvnyKJizR0dHYtm0bnj9/jurVq6syGFx1+p5HTTmTkJAgvXr1EmdnZ+0iZBk3NbRt21YqV64sR44cESsrK9m6dassW7ZMihQpIn/88Yfi8by9vfVWeyUjBwcHOXr0aKb9YWFhYmNjo3i8Jk2a6GwNGzYUT09PsbOzk6ZNmyoeTyTt/9Lb21vWrVsnN2/elJs3b8q6devEx8dHvvjiCxERWbFihZQrV07RuA8fPpQJEyZIixYtpH79+jJixAi5ffu2ojFERCZNmiTFixeXgwcPio2NjezZs0d+/vlncXZ2llmzZikeT0QkPDw8ywUAk5KSJDw8XJWYvr6+0q9fP0lISFClfRGRffv2SYUKFcTJyUm7iKS9vX2mzcHBQZX4Xl5eWda3UUupUqVk5syZEhMTI7a2trJ//34RETl69Ki4uLioErN69erSrVs3SUlJ0dbTiYmJkWrVqmlrQilh+/btYmlpqV3Q1cTERJYtW6ZY+/rChOUtDBgw4I03pfXs2VOKFSsma9asEQsLC1m8eLGMHz9e8ufPLz///LPi8UREXF1d5dChQyIiYmNjI5cuXRIRkQ0bNkjlypUVj7do0SJp0KCBxMfHK9726yxcuFCcnZ3lwoUL2n3Tpk0TW1tb1YqOvSwlJUW6d+8u33zzjSrt//vvv9K1a1edVXdNTU2lW7du8uTJExEROXHihJw4cUKV+GpLTU2VCRMmiJWVlfbibG5uLiNHjlQtZsZiXBndv39ftQ8RlpaW2mJx+qDRaLJ8jmpatmyZfP7556omZRn9+uuvYmJiIkZGRlKnTh3t/kmTJkm9evVUiWlnZycXL17Ufp2eoB08eFCKFCmiWJzKlStL48aN5fbt2/LgwQPp2bOnuLm5Kda+vnAMy1t4XX2AjJSqFZBRwYIFsXTpUtSoUQO2trY4fvw4ChcujGXLlmHFihWqzILQ10yodGXKlMHVq1fx/PlzeHp6wsrKSudxtVZrBtJuX8yaNQt79+7FqlWrMGnSJGzevFlv3dJAWjGuGjVq4M6dO6rFePLkiXbQoo+PD6ytrRVt/7/WSMlIyTFQz58/h4mJCZKTk3H16lU8efIExYsXh7W1Ne7fv6/KbdpXTYc/deoUatasqcoaRs2aNUPr1q3RsmVLxdvOyo0bN1CwYEHVF1bMqEyZMrh27RpEBF5eXpnG7alxHYiNjcWdO3dQunRp7dikw4cPw9bWNss1qnLK2dlZuzrzRx99hO+//x5BQUG4ePEiypUrh4SEBEXi2NvbY//+/dpbwYmJibC1tUVcXJze6k4pgWNY3kJuFtd58OCB9l68ra2t9mJYpUoV1UaxFylSBJcuXYKXlxdKly6N+fPnw8vLC/PmzVNlCmWTJk0Ub/NNDRkyBPHx8QgICEBKSgq2bNmiXblVX65du6ZK9eCM0i/MalWd9ff3f+UaKRkpPSi1devWWLNmDUxNTXXG6cTFxaF27dqKrgdVpkwZaDQaaDQa1K5dG3ny/O9ympKSgqioKNSrV0+xeBk1bNgQgwcPxvnz57Os4aP0zCRPT08A+ht4D+TOdcDV1RWurq4A0gpnbt++HUWKFFElWQHS/oaOHDkCX19fVK9eHaGhobh//z6WLVuGkiVLKhbn8ePHOsm6paUlLCws8OjRIyYspB4fHx9ERUWhYMGCKFq0KFavXo3y5cvj999/V22Qar9+/bSf9kePHo169erhl19+0c6EUlrGlZvVltVigB4eHrC0tES1atVw+PBhHD58GIDyI/ZfLuImIrhz5w42bdqEjh07Khornb5W+Y2KilKkneyKiYlB165d8eOPP2r33blzB7Vq1VJ8kGH6G+rJkycRFBSk00uVXmqgefPmisZM161bNwBpS3W8TI2ZSfpeggTQ73UAAFq2bIlq1aqhd+/eePr0KQICAhAdHQ0RwcqVK1X5v5w0aRL+/fdfAMDEiRPRoUMH9OjRA76+vjp/w0rYsmWLTjmM1NRUREZG6iTxhj4Fn7eE3kJurj/x3XffwdjYGH379sVff/2Fzz77DCKC58+fY8aMGXqpFZKYmKjqTCggbcpmeh2bwYMHw9HREcePH4eLiws8PDwUi/O64m0ZqVHI7eVbi0ZGRnB2dkatWrXQuXNnnU/sStF31dnsaNiwIRYtWpSjXrt79+6hWrVqqF+/PmbMmIHbt2+jZs2aKF26NFauXKlKLaHw8HC0atVKu47Q+0jfS5DkhowFDpcvX47Ro0fj1KlTCA8Px4IFC3DixIncPsW39iZ/9+9CHRb2sLwFOzs7bfd5xoxVHwYMGKD9OjAwEBcvXsSxY8dQuHBhVbplM0pOTkZUVBQKFSqEsmXLqhbn9OnTCAwMhJ2dHaKjo9GtWzc4Ojpi3bp1iImJwdKlSxWLlVs9AcCrby0mJibi8OHD+OSTTxSPuXXrVmzZsgX58+fX2e/r66va1M03tXv37hwvoufs7IytW7dqS/D/8ccfKFu2LH755RfVCh+q1RtmSLZv344NGzYgICAARkZG8PT0RJ06dWBra4vJkycrlrA4Ojri8uXLcHJy+s9FR5UeG/To0SNtNe+IiAg0b94clpaW2ttvaqhVqxbWrVuXqXf88ePHaNKkiWJjIHNj6r0amLC8hYyLYCm1INbbePbsGTw9PbX3l9WSmJiIPn36IDw8HEBaKXcfHx/06dMHHh4eGDp0qKLxQkJC0KlTJ0ydOlWnLkqDBg3Qtm1bRWMZoitXrqBq1aqqfNpJSEjItPAZkHbx12cVUzUVKFAA27ZtQ9WqVVGnTh0sW7ZM1cGiKSkp+O6777B69eosx3eoMegW0G/BsYSEBG3ROAcHB9y7dw8fffQRSpUqpejg1++++077mg8LC1Os3TdRoEABHDhwAI6OjoiIiMDKlSsBpK0cr1bv2c6dO7OsefTs2TPs2bNHlZhvQoneTjUwYXnHpKSkYNKkSZg3bx7i4uK0ycOoUaPg5eWFLl26KB5z2LBhOHXqFHbu3KkziDAwMBBjxoxRPGE5cuQI5s+fn2m/h4eHYqXVX+XWrVvYuHFjlm8CSixnn9uqVq2KpUuXYvz48QDSuoFTU1MxderUN579Zmhe9Uk8MTERv//+u86gQjWSh7Fjx2LRokUYOHAgRo4ciREjRiA6Ohrr169HaGio4vEA/RYcA/Q38D5jb5W+e6769++Pdu3awdraGgULFkSNGjUApPX8lSpVStFYGWfRnT9/Xue6lpKSgoiICEVvfWeXEr2damDCooA1a9a88tOV0lPvJk6ciPDwcEydOlU78A4ASpYsibCwMFUSlvXr12PVqlWoWLGizhtDiRIlFFsrKSMzMzNt1dWMLl++rOpKypGRkWjUqBF8fHxw8eJFlCxZUjvoTs1bYPr07bffolatWnqpOqsv+v4k/rJffvkFCxcuRMOGDTFmzBi0adMGhQoVgp+fHw4ePKhKefUBAwbgs88+w7x582BnZ4eDBw/CxMQEX3zxhSrj2PQ98B5Iu0Wzbds2REdHQ6PRwMfHB7Vr11almjcA9OzZE+XLl8fNmzdRp04d7S1EHx8fTJgwQdFY6bPoNBoNatWqlelxCwsLfP/994rGfC/kUv2X98bMmTPF2tpaevfuLaampvLll19KYGCg2NnZyfDhwxWPV6hQIfnrr79ERLSVEUVELly4IPb29orHExGxsLDQxskY8+TJk2Jra6t4vC5dukiTJk0kOTlZrK2t5fr163Ljxg0pU6aM9OvXT/F46T7++GMJDQ0Vkf89z3///VcaNWokP/zwg2pxX3by5ElVCo4lJydLrVq15NChQ3qpOptdGf+23iWWlpZy48YNEUkrsnjs2DEREbl27Zoqrw8R/RUce5WEhAQ5duyY3Lt3T5X2ly1bJnZ2dtrif+mbvb29rFy5UpWY6ZKSkuTixYvy/Plz1WJER0dLVFSUaDQaOXLkiERHR2u327dvy4sXL1SL/SYM9bXIHpYc+uGHH7BgwQK0adMGS5YswZAhQ+Dj44PQ0FBVup///vtvFC5cONP+1NRUPH/+XPF4ABAQEIBNmzahT58+AKDtZVm0aJEqK0RPnz4dn3/+OfLly4enT5+ievXqiI2NRaVKlTBx4kTF46W7cOECVqxYAQDIkycPnj59Cmtra4wbNw6NGzdWrM7Nf63uq9ZAYBMTE5w+fRoODg4YMWKEKjEMib7W2sqfPz/u3LmDggULolChQti6dSvKli2LI0eOqDYuyMTERNsDkC9fPsTExKBYsWKws7PDzZs3VYkJ6Gfg/fHjxxEcHIx27dphwIABKFq0KEQE58+fR1hYGNq3b4+iRYuidOnSisbV51i99HGH78tgWH1hwpJDMTEx2tkcFhYW2jn17du3R8WKFTF79mxF4xUvXhx79uzJNNB2zZo1KFOmjKKx0k2aNAn169fH+fPn8eLFC8ycORPnz5/H/v37sWvXLsXj2dnZYdu2bdi3bx9OnTqFJ0+eoGzZsggMDFQ8VkZWVlbaNzg3Nzdcu3ZNW7tDqZWLgTcriKXWINEvvvgCP/74I6ZMmaJK+zkxfPhw7SyNt5WQkICvv/4aq1evRnx8fKbH1RjI3LRpU0RGRqJChQro06eP9nccExOjM6tPSfoqOJZOn2/m33//PZo0aZLpVlPZsmWxdOlSJCYmYubMmYovDqrPsXr/9aElnaHXRdG73O7iedd5e3vL8ePHRUSkXLlyMm/ePBER2bJliyqLgq1fv17s7OxkypQpYmlpKd9++612bZitW7cqHi/d1atXpWvXrvLxxx9LsWLFpF27dnL69GnV4uWGxo0by4IFC0REZODAgVK4cGGZMGGClC1bVmrXrp3LZ6eM3r17i62trZQrV066d++u+tpX6ZYuXSqffPKJuLm5SXR0tIiIfPfdd7J+/XpF4+TGWlsvO3DggEyfPl02btyoWowjR47I9u3bRUQkLi5OgoKCxMbGRsqWLavKwqF9+/aVcuXKyZ49e8TKykp7u2D9+vXi7++vaCxfX1/Ztm3bKx/ftm2b+Pr6KhpTRKRgwYJy4MABEdG9JXLlyhXFFz99+VZXVpta61C9CUO9JcSEJYe6dOkiY8aMERGR2bNni4WFhQQGBoq9vb107txZlZi7d++WwMBAcXZ2FgsLC6lcubJs2bJFlVj6tH//fvn999919oWHh4uXl5c4OztLt27dslwVVynXrl2TU6dOiYjIkydP5Msvv5RSpUpJs2bNtG+yuaFBgwaKjTGpUaPGK7eaNWsqEuNlP/zwgzg5OcmECRN0xkP99NNPUqNGDUVjFShQQHbs2CEiaQt1XrlyRUTSEqb69esrGivd06dPVWn3VVJTU+XGjRt6javPN3MrKyvtmKCs3LhxQywtLRWNKaL/sXqGbNKkSfLPP//k9mlkwoQlh1JSUnQGZ61YsUL69Okjs2bNkqSkpFw8s5x7/vx5pgQhNjZWxowZI4MHD5Y9e/YoGq9evXoyZcoU7fenT5+WPHnySNeuXWX69Oni6uoqo0ePVjTmu8BQP+28qWLFislvv/0mIrrP5cyZM5I3b15FY2V8s/Pw8NCuMn79+nWxsrJSNFY6Gxsb6dChg2zdulVSUlJUiZFRSkqKmJiYyOXLl1WPlU6fb+b/tTJ0bGysKr0PVatWlVmzZomIaAf7i6T1SgYFBSkeLzuU+tCyZMkS+eOPP7TfDx48WOzs7KRSpUq5+qHsTXEMSw4ZGRnpVNBs3bo1WrdurZfYz549w6pVq5CYmIjAwED4+voq2n63bt1gamqqrYny77//4uOPP8azZ8/g5uaG7777Dhs2bECDBg0UiXfy5EltfRAAWLlyJSpUqICFCxcCSCvsNHr0aIwZM0aReFnR15IAH5KoqKgsx1eZmZkpthptutxYays8PBzLly9H48aNYWdnh1atWuGLL75AQECAKvGMjIzg6+uL+Ph4xV/zr6Lvgfcvr3uT0cOHDxWPB+h/rF52KFUXZdKkSZg7dy4A4MCBA5gzZw6+++47/PHHHxgwYIDiS8koLrczpnddoUKFZPTo0XLp0iVV4wwYMEB69+6t/T4pKUlKly4tJiYmYmdnJ1ZWVrJ//35FY/r6+urcapo9e7a4u7vLw4cPRURkyJAhinbpm5mZSUxMjPb7ypUry4QJE7TfR0VFibW1tWLxXnbq1ClxdnaWwoULS548ebSfIkeMGCHt27dXLe5/eR96WNLHqmR8LrNmzZIyZcooEsPIyEji4uJkxowZMnPmTBFJG+tgbm4uZmZmYmRkJGFhYYrEepXHjx/L4sWLpU6dOmJsbCy+vr4yduxYVWJt3LhRqlSpImfOnFGl/Zft2bNHrK2t5auvvhJzc3Pp16+f1KlTR6ysrOTo0aOKxsrN8R2GOlZPqWuAhYWFtgdyyJAh2uva2bNnxcnJKcftq40JSw7NmDFDAgICxMjISAICAiQsLEzu3LmjeJwSJUrIhg0btN8vXrxYHBwcJDo6WlJTU6VTp07SoEEDRWNaWlpqu0VFRJo2bSp9+vTRfn/u3DlxdnZWLF7BggVl165dIpKWkFlYWGhrzoik3SJSYyBzutq1a8vgwYNFRPcCsW/fPvH09FQt7n951xOWhQsXioeHh6xcuVKsrKxkxYoVMmHCBO3XSnjVbYTo6GhZu3atdmySvpw7d078/f1Ve2O1t7cXU1NTMTIyEnNzc3FwcNDZ1GCob+YfAqWuAc7OztpJIv7+/rJ06VIRSfu/VeuWqZJ4SyiHBgwYgAEDBuDy5cv45ZdfMGfOHAwaNAg1a9bEF198gQ4dOigSJyYmBsWLF9d+v3XrVnz++efa6c39+vVT7NZMOnNzc51uyIMHD+Lbb7/VefzJkyeKxWvQoAGGDh2Kb775BuvXr4elpSWqVq2qffz06dMoVKiQYvFelptLArzPunbtCgsLC4wcORKJiYlo27Yt3N3dMXPmTNVvn+pjra10z549w8aNG7F8+XJERETAxcVFtUXzcqO6b6FChbS3Zw2JUuvebN68GcbGxggKCtLZv2XLFqSmpqJ+/fo5at8Q1KlTB127dkWZMmVw+fJl7XvGuXPn4OXllbsn9waYsCjko48+wtixYzF27FgcPHgQPXr0QHBwsGIJi5GREURE+/3BgwcxatQo7ff29vb4559/FImVzt/fH8uWLcPkyZOxZ88exMXF6ZSRvnbtGtzd3RWLN378eDRr1gzVq1eHtbU1wsPDYWpqqn188eLFqFu3rmLxXpZbSwJ8CNq1a4d27dohMTERT5480S6kp6RFixbB2tr6tceoUSZ/y5YtWL58OdavX488efLg888/x9atW1GtWjXFYwHA8+fPsWvXLowaNQre3t6qxEj34sULpKSk6BTAi4uLw7x585CQkIBGjRppV8bOLUqN7xg6dGiW9YlEBEOHDn0vEpY5c+Zg5MiRuHnzJtauXatdZ+vYsWNo06ZNLp/dG8jtLp73yaFDh6Rfv37i6uoqlpaW0qpVK8XarlixokyfPl1E0u43GhkZ6dyu2blzp+K3LXbu3CkWFhbi4+MjFhYWmaZp9+jRQzp06KBoTBGRhw8fZlmaOj4+XtWZV7m1JMB/MdQphm/q+vXrWc5ouXz5skRFRSkSQ6PRSIECBcTLy+uVm7e3tyKxXmZhYSEtW7aU9evXS3JysioxXmZra6vz+ldLp06dpHv37trvHz9+LAUKFBBnZ2fx8/OTPHnyyKZNm1Q/j9dR6naJubl5ln+PUVFRqkyjzo53/bawUpiw5NClS5ckNDRUfH19JU+ePFK3bl0JDw+Xf//9V9E469atE1NTU6lVq5a4uLjIp59+qvP4kCFDpEWLForGFBE5f/68hIWFycqVKzNN2Zw/f76cOHFC8Zi55eHDh9oaOsbGxlKgQAExMTGRatWqyZMnTxSP965PMXxT1apVkyVLlmTav2zZMqlevboiMf5rKqxanj9/LrNmzVJl3NrrdOjQQWbMmKF6HH0PvH8bSr2Zu7i4SGRkZKb927ZtU3Ss3ttQ8kPL7t27pV27dlKpUiW5deuWiKTVKVK6TIUamLDkkEajkfLly0tYWJjExsaqGuuvv/6S/v37y5QpUyQhIUHnsTFjxmgLZlHO7N27V+bMmSPffPPNaytu5tRHH32kvUDu379fLC0tZf78+fLZZ59J06ZNVYurbxkLuGV05coVsbOzUyRG+iyh3GBhYaH3BHP8+PFib28vzZs3l0mTJsnMmTN1NqXoe+D921AqYenevbuUKlVKrl69qt135coV8fPzky5duuS4/azo+0NLegXorl27ipmZmfb39v3336tWWFFJTFhyKKuu7tTUVNm8ebM0b948F84oTY8ePVRbSfV9lJycLMbGxnqbJiry7k8xfFO2trbamQkZHT16VLFp6rnVwyIiUr16dW1hPH3R160vR0dHOXfunPZ7Nzc3nSUOrl27JhYWForFextKJSwPHz6UihUrSp48ebS/yzx58kjNmjVVuyWr7w8t/v7+Eh4eLiK6v7fjx4+Li4uL4vGUxkG3OZSxcFNUVBQWL16MJUuW4N69e6ov1vc6P//8MwYNGgQnJ6dcO4d3iYmJCQoWLKjK4nivYm1tjfj4eBQsWBBbt25FSEgIgMyzs9511apVw+TJk7FixQoYGxsDSFuEcPLkyYoN2Bw9evR/DrjNqGfPnhg3bpwir4+ePXti4MCBuHXrFsqVKwcrKyudx/38/HIc42Vqrej9Mn0PvM9NdnZ22L9/P7Zt24ZTp07BwsICfn5+qg2eBoCbN2+icOHCAID169ejefPm6N69OypXrowaNWooHu/SpUtZPh87OzvVCvIpKrczpnfds2fP5Oeff5aaNWuKiYmJGBkZyYwZM+TRo0e5el4cpJV9ixYtkgYNGkh8fLxe4rVt21bKli0rXbp0EUtLS7l//76IiGzYsEFKlCihl3PQh3PnzknevHmlUKFC0qlTJ+nUqZMUKlRInJ2d9dqjlZGNjY1ir49XFTbTxwJ2SUlJcvHiRZ3lQZSUWwPvs+NdHpSu77oo3t7e2tvcGd8jwsPDpVixYorHU5rRf6c0lJVjx46hZ8+ecHV1RVhYGJo0aYKbN2/CyMgIQUFBsLW1ze1TVERKSgp2796t9+x72bJlqFy5Mtzd3XHjxg0AabUnNmzYoFrM2bNnY/fu3XB3d0eRIkVQtmxZnU1pc+bMQaVKlXDv3r13c4rhGypevDhOnz6Nli1b4u7du/j333/RoUMHXLx4ESVLlsyVc5IMJQJyKioqKtN2/fp17b9qSExMRJcuXWBpaYkSJUogJiYGANCnT58sp+a+rerVq+PYsWPo27cvfvrpp0x1WPz9/TFgwADF4mUUHh6OTZs2ab8fMmQI7O3t8cknn2ivCQAwbNgwRZZd6Nu3L2bNmpVp/+zZs9G/f/8ct5+V9LooXbt21UtdlG7duqFfv344dOgQNBoNbt++jV9++QWDBg1Cjx49FI+nuNzOmN5VxsbG0r9/f7l48aLO/jx58ujc880tSvawmJmZ6WUKZTp9ru6b0ZgxY1670fvjXe+B7Nu3r5QrV0727NkjVlZW2ueyfv168ff3z7XzUnJlcX2P73B3d89ymYFjx46Jh4eH4vFERP755x/p1auXNGrUSP7880/t/tDQUJ1lSZSSmpqqrTKd3htobm4uI0eOVDyWGjQiCn7U+IAEBQXhwIED+Oyzz9C+fXsEBQVBo9HAxMQEp06d0qlKmxtsbGxw6tQp+Pj45LitgIAAfPPNN6hdu7YCZ/bfihcvjkmTJqFJkyY6z+Ps2bOoUaMG7t+/r5fzUMPp06ff+Fg1xj7oy+nTp1GyZEkYGRn953POjeep5OsDSOsRnDdvHqKionDgwAF4enoiLCwM3t7eaNy4sSIxMvL09MSqVatQsWJFnedy9epVlC1bNssCiPqg5O/V0tISFy9eRMGCBfH111/jzp07WLp0Kc6dO4caNWrg3r17Cpzx/5ibm+Ps2bPaMSXprl69ipIlS+LZs2eKxstNycnJuHr1Kp48eYLixYtna/xXbuKg27e0ZcsW3Lx5Ez/99BN69OiBp0+folWrVgD+t5Lp+2LChAkYNGgQxo8fn+WgQqVvf+lzdd/XuX79Op4+fYpixYrprMidE/7+/tBoNK+8JZH+mEaj0esAYKX5+/sjNjYW+fLle+1zftefJwDMnTsXoaGh6N+/PyZOnKh9Pvb29ggLC1MlYbl3716W1YITEhLem+uPvgelFy5cGBEREejdu7fO/j///FOxxDYre/bswfz583H9+nX8+uuv8PDwwLJly+Dt7a1aFWFTU9Nc/1D9Npiw5ECBAgUQGhqK0NBQbNu2DT/99BPy5MmDxo0b4/PPP8fnn3+uytiHN/HFF18olkik31dt1KiRzsVQrTdWb29vnDx5MtMaMBEREShWrJiisYC0UucTJkzA8ePHUbFiRQwdOhRffPEFVq9eDQAoUqQINm/erMg9ZX3N7shtUVFR2uUM3vfn/P3332PhwoVo0qSJzviRgIAADBo0SJWYAQEB2LRpE/r06QPgfx+SFi1ahEqVKqkSU9/0ve5NSEgIevfujXv37mlnQkVGRmL69Omqrd20du1atG/fHu3atcPx48eRlJQEAHj06BEmTZqEzZs3KxovISEBU6ZMQWRkJO7evYvU1FSdx9Uac6WY3Lwf9T568OCBzJo1S7WVWv/880+dioSzZ8+W0qVLS5s2beTBgweKxxNJmynwuk1p+ljdN6OQkBBxdnaWrl27io+PjzRq1EiKFCkiK1eulNWrV0upUqWkbdu2isf9ECQnJ0twcLBex0C9ia+++kqxOkXm5ubaIl8Zx8ZcvnxZzM3NFYnxsj179oi1tbV89dVXYm5uLv369ZM6deqIlZVVluMw9EXJsUH6Ht8hkjZ+zsPDQzu+w9vbW1u3RA36rovSunVrcXNzkyFDhsh3330nYWFhOpuhY8KiomPHjineZsmSJbVrd5w+fVrMzMxk2LBhUrFiRenUqZPi8XLLzz//LIULF9ZeODw8PGTRokWqxCpYsKD2d3rp0iXRaDSyefNm7eM7d+5UbdCdSNq03z///FM2bNigs70v9LXujUjuJPTFihWT9evXi4jum86sWbOkTJkyqsQUSZv62rVrV/n444+lWLFi0q5dOzl9+rRq8d7Euz6YOd3du3d1lldRq9SBhYWFdv2ijL+7a9euiZmZmeLx7OzsZO/evYq3qy+8JaQiNW4HRUVFae89rl27Fp9++ikmTZqE48ePa7tM1fDw4UP8+OOPuHDhAgCgRIkS6Ny5M+zs7FSJp4/VfdPdvn0bpUuXBpC26raZmZnOwLuPPvoIsbGxise9fv06mjZtijNnzuiM8Ujv3n/Xx3aka9KkCdavX6/a9NeMBg8ejG+++QYAcObMGQwcOBAhISHYsWMHQkJC8NNPPykeMyQkBL169cKzZ88gIjh8+DBWrFiByZMnY9GiRYrHS1eoUKFM04zfdYYyKD39dubWrVuxaNEi/P7776qMm3F1dcXVq1cz3eLau3evKuNmHBwc4OjoqHi7epPbGRNlj4ODg3badOXKlWX+/PkikraiqFolso8cOSKOjo7i4eEhTZs2laZNm0r+/Pklb968qvQi6dvLZd1f/pQYGxuryu29Tz/9VBo3biz37t0Ta2trOX/+vOzZs0fKly8vu3fvVjxebtHXujciIlZWVtpPrKNHj9Yuj3Hs2DFVS4/rs0dQ5NVrJ92/f1/1YnWvk9MibhmL7uVWQb7o6GgJDQ0VT09PsbW1lVatWsnq1atViTVp0iQpXry4HDx4UGxsbGTPnj3y888/i7Ozs8yaNUvxeMuWLZPPP/8801p07wpOa37HNGrUCMnJyahcuTLGjx+PqKgoeHh4YOvWrejduzcuX76seMyqVauicOHCWLhwIfLkSeuUe/HiBbp27Yrr169j9+7dOY5RpkyZN57dcPz48RzHy8jIyAjh4eHa3qI2bdogLCwMLi4uANJ6l4KDgxXv8XBycsL27dvh5+cHOzs7HD58GEWKFMH27dsxcOBAnDhxQtF4ucXb2/uVj2k0GkUH+jk6OmLv3r0oXrw4qlSpgg4dOqB79+6Ijo5G8eLFkZiYqFisrOijRxBI+5tNn4WV0e3bt1GoUCHFewPCw8Ph5OSEhg0bAkgr4rZgwQIUL14cK1asyDRA/m1lLAj3X5SKCaRN8123bh0WLVqEffv2ITAwEH/++SdOnDiBUqVKKRbnZSKCSZMmYfLkydq/TTMzM+2sTKWVKVMG165dg4jAy8sLJiYmOo8rfW1VGm8JvWNmz56Nnj17Ys2aNZg7dy48PDwApE29q1evnioxjx49qpOsAECePHkwZMgQBAQEKBKjSZMmirTztjp27Kjz/ZdffqnzvRpTRVNSUmBjYwMgLXm5ffs2ihQpAk9PT1y6dEnxeLkl4ywheem2l9KqVKmCkJAQVK5cGYcPH8aqVasAAJcvX0b+/PlVifn06VOICCwtLWFpaYl79+4hLCwMxYsXR926dRWNlV6JVaPRYNGiRTr1M9KrUhctWlTRmAAwadIkzJ07FwBw4MABzJkzB9999x3++OMPDBgwAOvWrVMkjpJJyJvq06cPVqxYAV9fX3zxxRdYtWoV8ubNCxMTE+3aV2rRaDQYMWIEBg8erJe6KLl9nc2x3OzeeR8kJibqdK9FR0fLd999J1u2bMnFs1JWvnz5snw+ERERki9fvlw4o/dDlSpVtKv8tmnTRurVqyd79+6VDh06vFdrCYmkrdNUokQJMTU1FVNTUylRooQsXLhQ8Tg3btyQhg0bip+fn84tmf79+0ufPn0UjyciUqdOHZk7d66IpM1syZcvn+TPn1/Mzc3lhx9+UDRW+irCGo1GChQooLNK80cffSR169aVgwcPKhpTJHdXFld7ULqxsbEMHz5cHj9+rLPfUKqW0/8wYcmhly9WLi4uql2s0l29elVGjBghrVu31t7H3rx5s5w9e1aVeH369JH8+fPLypUrJSYmRmJiYmTFihWSP39+6devnyoxRdLGzixdulSWLl2aq1M1X6ZU+fGIiAhZu3atiIhcuXJFihQpIhqNRpycnLQlyd8Ho0aNEisrKxk6dKj2zWbo0KFibW0to0aNyu3Ty7G8efNqX3sLFy4UPz8/SUlJkdWrV0vRokVViVmjRg3VZj1lRd+L9ImkzZTx8/PLNK7FyMhI0TEsy5cvl8DAQLGyspKWLVvK77//Li9evNBLwvLkyRMZOXKkVKpUSQoVKiTe3t46G+liwpJD+r5Ypa+eGhgYKKamptrBoZMnT9YOMFRaUlKS9O3bV0xNTbUXCzMzM+nfv788e/ZM8Xg3b96UKlWqiEajEQcHB3FwcBCNRiOVK1eWmzdvKh4vu9ScuhkfHy+pqamqtJ1bnJycZPny5Zn2L1++XPLmzat4PH0n9Bl7H1q0aKFddyomJka1gfD6lhsri+t7UPr169clNDRUChYsKE5OTmJkZCS//vqr4nEy0kddFAcHB23NIXt7e+01NavN0HHQbQ5lXO+iZcuWKFGiBEaPHo2bN2+iSJEiig/yq1SpElq0aIGQkBCddTsOHz6MZs2a4datW4rGyygxMRHXrl0DkDal0tLSUpU49erVw8OHDxEeHo4iRYoAAC5duoTg4GDY2toiIiJClbhvSul1aK5evYpr166hWrVqsLCw0FYQfl/Y29vjyJEj8PX11dl/+fJllC9fXtGVwHft2oX69eujcuXK2L17Ny5cuAAfHx9MmTIFR48exZo1axSLlc7Pzw9du3ZF06ZNUbJkSURERKBSpUo4duwYGjZsqNiU+JCQEIwfPx5WVlbaMvWvMmPGDEVipnv48CFGjhyJmzdvokePHtrxcqNHj4apqSlGjBihaDwg9waliwi2bt2KH3/8ERs3boSTkxOaNWuW5UrOOWVvb49NmzahcuXKiredLjw8HK1bt4aZmRmWLFny2mvLy2P5DE7u5kvvvlKlSsnMmTMlJiZGbG1tZf/+/SIicvToUVWmUVpZWWmLcGX8pB8VFaVKoaGX3bx5U/VeDnNzc233c0ZHjx41iE+sSvWw3L9/X2rVqqXt5k5vMzg4WEJCQnLcvqHo3bu3DBgwINP+gQMHSs+ePRWNVbFiRZk+fbqI6P4/HTp0SLXif7/++quYmJiIkZGRBAYGavdPmjRJ6tWrp1icGjVqaKcM16hR45VbzZo1FYuZm+zt7bXXOh8fH9m+fbuIpPWg6es6EB8fL9999534+fmp0r6Xl5ecP39elbbfR0xYcijjxapOnTra/UpfrNJ5eHjIvn37RET3grxu3Trx8fFRPJ6ISEpKiowdO1ZsbW21t4Ts7Oxk3LhxkpKSong8X19fOXToUKb9hw4dkkKFCikeL7uUSljat28vQUFBcvPmTZ02IyIipHjx4jlu31D07t1bbG1tpUSJEtKlSxfp0qWLlCxZUmxtbbXJTPqWU7mV0N+5c0eOHz+u83o4dOiQXLhwQbWY+rZ7925p166dVKpUSW7duiUiIkuXLtWpLKykD2FQur7rohw7dkynGvL69eulcePGMmzYMElKStLLOeQEExYF6PNiNXDgQKlSpYrcuXNHbGxs5MqVK7J3717x8fHR3jtX2tChQ8XZ2Vl++OEHOXXqlJw6dUrmzJkjzs7OMnz4cMXjrV+/XsqXLy9HjhzR7jty5IhUrFhRewHLTUolLC4uLnLy5MlMbV67dk21gYy54XW9AUr3DORGQp/uypUrEhERIYmJiSIi79VYpDVr1oiFhYV07dpVzMzMtL/X77//XurXr69KzA9hULq/v7/Y2NiItbW1lCxZUsqUKaOzKS0gIEDWrFkjIv8r/9+mTRspXLiwqhMolMIxLDnw/PlzWFhY4OTJkyhZsqReYiYnJ6NXr15YsmQJUlJSkCdPHqSkpKBt27ZYsmSJKnUD3N3dMW/ePDRq1Ehn/4YNG9CzZ0/8/fffOY7h4OCgc281ISEBL1680ClUlydPHlhZWeHBgwc5jpcTSo1hsbGxwfHjx+Hr66vT5tGjRxEUFIT4+HiFzvjDMWjQIBw6dAi//vorPvroIxw/fhxxcXHo0KEDOnTogNGjRyseMz4+Hi1btsSOHTug0Whw5coV+Pj4oHPnznBwcMD06dMVidOsWbM3PlapuijpypQpgwEDBqBDhw46f6snTpxA/fr1VVm6IisPHjzIdK14l40dO/a1jyv992pnZ4fjx4+jUKFC+Oabb7B9+3Zs2bIF+/btQ+vWrXHz5k1F4ymNheNywMTEBAULFtTbmi8igtjYWMyaNQuhoaE4c+YMnjx5gjJlymQa0KikBw8eZFmMqmjRooolD2ot366G4cOHK7IeR9WqVbF06VJtRUuNRoPU1FRMnToVNWvWzHH7H6JJkyahV69eKFCgAFJSUlC8eHFtQj9y5EhVYg4YMAAmJiaIiYlBsWLFtPtbtWqFkJAQxRKWjOt2iQh+++032NnZaYs3Hjt2DA8fPsxWYvOmLl26hGrVqmV5TkoOms5KxkHpjo6OeJ8+Y6uRQL+OiCA1NRUA8Ndff+HTTz8FABQoUAD379/X67m8ldzs3nkfLFq0SBo0aKDaap4ZpaSkiImJiVy+fFn1WBmVL18+y6JbvXv3lgoVKuj1XNS0ZMkS+eOPP7TfDx48WOzs7KRSpUoSHR2teLwzZ85Ivnz5pF69emJqaiqff/65FCtWTFxcXOTq1auKx3vfpaamyo0bNyQxMVFiYmJk06ZNsmrVKtVfL7lxa2/IkCHStWtXefHihXbfixcvpHv37jJo0CDF43l7e8u2bdtERPc5hoeHS7FixRSPJ6LfQenPnz+XsWPHGkTZBDXVrFlTOnToIEuXLhUTExO5cuWKiKSVy/D09Mzdk3sDTFhyyN/fX6ytrcXMzEw++ugj1e9BFi9eXA4cOKB4u6+zc+dOsbKykmLFiknnzp2lc+fOUqxYMbG2tlZ9kb6nT5/Ko0ePdDa1fPTRR9p74/v37xdLS0uZP3++fPbZZ9K0aVNVYv7zzz8yYcIEadGihdSvX19GjBihSFG6D1FuJfTW1tbamBnfzNMXDVWDk5OTXLx4MdP+ixcvqhJT34v0ieh/ULq1tbV24Uw15WZdlFOnTmkHvGcc89i7d29p06aN4vGUxltCOaTvtRmmTJmCwYMHY+7cuXobN1O9enVcvnwZc+bMwcWLFwGk3U/v2bMn3N3dFY+XkJCAr7/+GqtXr85yHIdat+Bu3ryJwoULAwDWr1+P5s2bo3v37qhcuTJq1KihSkxzc3PUqVMHpUuX1nbVHjlyBAAyjRmi1zMyMoKvry/i4+NVvUX6sty4tffixQtcvHhRW6co3cWLF7V/R0oaOnQoUlNTUbt2bSQmJqJatWraRfr69OmjeDwA2Lp1K7Zs2ZJpDShfX99sLZL4pmrVqoVdu3bBy8tL8bYz+u6777RriH333Xd6HY/j5+eHM2fOZNr/7bffqr5ukiJyO2Oi7LG3t9dWnDU3N3/nKhW+iZ49e0qxYsW0MxMWL14s48ePl/z588vPP/+sWlx9lx//888/tRU108uOZyw/Ttm3ceNGqVKlipw5c0ZvMXPj1t6AAQMkb968Mn36dNmzZ4/s2bNHpk2bJk5OTopMD3+VpKQkOXfunBw6dEj+/fdf1eKI6L/nau7cueLq6ioDBw6U5cuXq7Z2Eb09zhJ6x4SHh7/2caUqFZ4+ffqNj/Xz81MkZrqCBQti6dKlqFGjBmxtbXH8+HEULlwYy5Ytw4oVK7B582ZF46Vr164dLl68iDJlymDFihWIiYlB3rx5sXHjRgwfPhxnz55VNJ6vry/q1q2L0NBQuLi4KNr2h8rBwQGJiYl48eIFTE1NYWFhofO4WjPMHj16hNmzZ+PUqVN48uQJypYti169esHNzU2VeKmpqZg2bRpmzpyJO3fuAADc3NzQr18/DBw48N34tPwfGjRogHLlymH8+PGwsbHB6dOn4enpidatWyM1NVXxqsVGRkavfEyj0ajSs3v8+HGYmJigVKlSANJmXv70008oXrw4xowZA1NT0xzHcHR0xOXLl+Hk5PSfM6xyewbmf2HC8hbepz+AVzEyMoJGo8lUJj79zyXjPqVfyNbW1jh//jwKFiyI/PnzY926dShfvjyioqJQqlQpPHnyRNF46fRdftzW1hYnTpxAoUKFFG33Q6avhP5NPHv2DLNnz8agQYNUjfP48WMAaX9PaklISMCUKVMQGRmJu3fvZrrtdP36dcVjnj17FrVr10bZsmWxfft2NGrUCOfOncODBw+wb9++9+J18/HHH2Po0KFo3rw5rl+/juLFi6NZs2Y4cuQIGjZsqMjsyYyl+Q3p9fE2mLC8hdz8A4iJiXnt4wULFlQkTsZ7xCdOnMCgQYMwePBgVKpUCQBw4MABTJ8+HVOnTlV8HI+fnx++//57VK9eHYGBgfD398e0adMwa9YsTJ06VdX1kvSpc+fOqFy5Mrp06ZLbp0Jv6d69ezh06BBMTU1Ru3ZtGBsb4/nz5/jhhx8wefJkvHjxQrXpoi9evMDOnTtx7do1tG3bFjY2Nrh9+zZsbW1hbW2taKw2bdpg165daN++Pdzc3DJ9SOvXr5+i8dI9fPgQc+bM0VvPlb6963VR9I0JyzsmvefjVdTotixfvjzGjBmDBg0a6OzfvHkzRo0ahWPHjikSx9jYGHfu3MEvv/wCY2Nj9O3bF3/99Rc+++wziAieP3+OGTNmKHpxzM1bX4mJiWjRogWcnZ1RqlQpmJiY6Dzet29fReN9CPSV0APA3r178emnn+Lx48fQaDQICAjATz/9hCZNmiBPnjzo27cvOnbsmOm2lBJu3LiBevXqISYmBklJSbh8+TJ8fHzQr18/JCUlYd68eYrG08cifVl59uwZTp8+nWWvjhKD0mfNmoXu3bvD3Nz8Pxc3VOP1aGtri2PHjsHX1xd16tTBp59+in79+iEmJgZFihTB06dPFY2nz9eHGpiwKOjZs2dITk7W2ad0N+2pU6d0vn/+/DlOnDiBGTNmYOLEiaoUjbKwsMDx48d1imIBwIULF1C2bFnFXlRGRkaIjY1Fvnz5dPbfuHEDx44dQ+HChRVPGjLe+spKxttiSieDP/74I7766iuYm5sjb968OomoRqNRpZv9fafPhL5GjRpwd3fH8OHDER4ejunTp8PX1xcTJ07E559/rlicrDRp0gQ2Njb48ccfkTdvXm3l2Z07d6Jbt264cuWKovG8vb2xefPmTNcANUVERKB9+/Z48OBBptenUq9Hb29vHD16FHnz5oW3t/crj1Pr9VirVi0UKFAAgYGB6NKlC86fP4/ChQtj165d6NixI6KjoxWNlxsfeBWl/3G+75cnT55Ir169xNnZWbswYMZNX/744w+pXr26Km2XKVNG2rdvr7M4VlJSkrRv317RWjMajUbi4uIUa+9NREdHv/GmNBcXF5k4caIqC0h+qE6ePKmzHTlyRBYsWCBFixbVrkujFEdHRzl37pyIiCQmJoqRkZGsX79e0Rivi51eh+XlRR7VWMlY34v0iYgULlxYevbsKbGxsXqLqW/6rouiz9eHGtjDkkO9evXCjh07MH78eLRv3x5z5szB33//jfnz52PKlClo166dXs7j6tWrKF26NBISEhRv+/Dhw9rbMuk9HKdPn4ZGo8Hvv/+O8uXLKxLHyMgIEyZM+M/77+/LrRJHR0ccOXLkvRg8aOg2bdqEb7/9Fjt37lSszZd7BG1sbHDy5Em9/H86ODhg3759KF68uM7aPnv37kXz5s0RFxenaLwyZcrg2rVrEBF4eXllun15/PhxReMBH/ag9GfPnsHY2DjT71ktarw+1MDCcTn0+++/a6fgBgcHo2rVqihcuDA8PT3xyy+/KJ6wpM8ISCciuHPnDsaMGaNasazy5cvj+vXr+OWXX7SF41q1aoW2bdvCyspK0Vjz5s177ZRMjUajesJy/vx5xMTEZLq9p3Qht44dO2LVqlUYPny4ou1SZkWKFNEW5FPS+fPntQv/iQguXbqU6UOD0rcxAaBu3boICwvDggULAKS9Lp48eYLRo0dnGmumBH0XyASAzz//HDt37tRrwnLr1i1s3Lgxy9f/jBkz9HYe5ubmeosFqPf6UBp7WHJI31Nws7oHKSIoUKAAVq5cqZ3Fo6Tdu3fjk08+0a6cnO7FixfYv39/louivY1XjWHRl+vXr6Np06Y4c+aMzriW9N+30vd3+/bti6VLl6J06dLw8/PL9GlKnxfI98XrEvqLFy/i5MmTisV63fgnNcc+AWlvrEFBQRARXLlyBQEBAbhy5QqcnJywe/fuXHsNKUnfg9IjIyPRqFEj+Pj44OLFiyhZsiSio6MhItqp1UrIzbIY+nx9qIE9LDnk4+ODqKgoFCxYEEWLFsXq1atRvnx5/P7777C3t1c83o4dO3S+NzIygrOzMwoXLpwpoVBKzZo1cefOnUwXwUePHqFmzZqKXZBze8n4fv36wdvbG5GRkfD29sbhw4cRHx+PgQMHYtq0aYrHO3PmDMqUKQMAmYrS5fbv4l1lb2//2oReSVFRUYq2lx358+fHqVOnsHLlSpw+fRpPnjxBly5d0K5dO1VmJeWGFStWYOvWrTA3N8fOnTszDUpXOmEZNmwYBg0ahLFjx8LGxgZr165Fvnz50K5dO21NJiVkLM2v71XqX/f6WLFihV7P5W2whyWHvvvuO71NwQX019uRkZGREeLi4uDs7Kyz//LlywgICMiUteckTm72sDg5OWH79u3w8/ODnZ0dDh8+jCJFimD79u0YOHAgTpw4kSvnRW9u165dOt/rI6F/Uz179sS4cePg5OSUq+fxpnK7QKarqyv69u2LoUOHvrYKrVIyjkFycHDA3r17UaJECZw6dQqNGzdWfMZObjDk18ebMPwzNHADBgzQfh0YGIiLFy+qNgUX0F9vBwDtFGmNRoNOnTrBzMxM+1hKSgpOnz6NTz75RLF4o0ePzlbBK6XfAFJSUrSffJycnHD79m0UKVIEnp6euHTpkiIxSF0ajeaVCf3u3btVSejf1M8//4xBgwbl6O919+7db3ScEs8zNxfpA4Dk5GS0atVKL8kKAFhZWWnHrbi5ueHatWsoUaIEAKhW/E/fdVH2798PFxcXdO7cWWf/4sWLce/ePXz99deKxlMaExaFeXp6wtPTU7X25aVS+eni4+MVHwBrZ2enjWljY6PT1WxqaoqKFSuiW7duisUbPXp0to5X4g0go5IlS+LUqVPw9vZGhQoVMHXqVJiammLBggXw8fFRJAapS58JfXYp0Zldo0YN7ev/Ve0pNW4mY5XuTp065bi9t4mvz0HpFStWxN69e1GsWDE0aNAAAwcOxJkzZ7Bu3TpUrFhRlZheXl56rYsyf/58LF++PNP+EiVKoHXr1kxYPgSRkZGvXGNj8eLFisTQd28HAPz0008A0l5UgwYNUjwhyiml72aOHDlSO8Nj3Lhx+PTTT1G1alXkzZsXq1atUjQWqUOfCX1ucHBwgI2NDTp16oT27dvr7faSPhbpe1lKSgqmTp2KLVu26GVQ+owZM7STJMaOHYsnT55g1apV8PX1VW0A/Mu3mV8uBKq02NjYLJc1cHZ21i6iaciYsOTQ2LFjMW7cOAQEBGS5xoZS9N3bkVF2ez7eVUFBQdqvCxcujIsXL+LBgwf/ef+ecl9uJPS54c6dO/jtt9+wePFiTJ06FQ0aNECXLl1Qr149Vf9Gv/zySwwdOhSlSpXC9evX0apVKzRr1gy//vorEhMTVRk8qu9B6Rl7Ua2srBRf3iArpUuXzrQvICAA7u7u+PbbbxWvXF6gQAHs27cvU1Xfffv2wd3dXdFYamDCkkPz5s3DkiVL0L59e1Xj5GZvh7e392svEO9bCfmrV6/i2rVrqFatGhwdHRXvySHl5WZCr0+mpqZo1aoVWrVqhZiYGCxZsgS9e/dGUlISOnbsiLFjx6oyePLy5cvw9/cHAPz666+oXr06li9frl2kT42E5eUZkfqSnJycZW+5PtfZUasuSrdu3dC/f388f/4ctWrVApB2h2DIkCEYOHCg4vGUxoQlh5KTk/X6yS03ejv69++v8316t2VERAQGDx6s9/NRS3x8PFq2bIkdO3ZAo9HgypUr8PHxQZcuXeDg4IDp06fn9inSKxj67Us1FCxYEKGhoWjfvj26dOmCKVOmYODAgXB0dFQ8loho38D/+usvfPrppwDSPrGrNSBV3y5fvowuXbpg//79OvvVrKej70KggwcPRnx8PHr27KkdYGxubo6vv/4aw4YNUzye0piw5FDXrl2xfPlyjBo1Sm8x16xZg9WrV2dZjVGNEtmvmpo9Z84cHD16VPF4uWXAgAEwMTFBTEyMziJvrVq1QkhICBOWd4Ah37784osvFFsMNSkpCWvXrsXixYtx4MABNGzYEJs2bVIlWQHSblNMmDABgYGB2LVrF+bOnQsgrRaNi4uLKjH1LTg4GHny5MEff/yh6u39jPRdF0Wj0eCbb77BqFGjcOHCBVhYWMDX11fnFqohYx2WtxASEqL9OjU1FeHh4fDz89PLwLBZs2ZhxIgR6NSpExYsWIDg4GBcu3YNR44cQa9evVQZqPUq169fh7+/v2J1WLKrR48eGD9+vGIDD11dXbFlyxaULl1aZ32W69evw8/PT/GqxaQOfSf0ERERsLa2RpUqVQCkJfILFy5E8eLFMWfOHDg4OCgW6/Dhw/jpp5+wcuVKeHl5ITg4GF988YVqiUq606dPo127doiJiUFISIg2MezTpw/i4+OznHnyrrGyssKxY8dQtGhRvcV81+ui6Bt/I2/h5ZHd6fd29TEw7IcffsCCBQvQpk0bLFmyBEOGDIGPjw9CQ0NVKd70OmvWrFHlQvmmbwDpn/KUkpCQAEtLy0z7Hzx48M58AvnQZUzoN2zYkCmhV8PgwYPxzTffAEgbKDpw4ECEhIRgx44dCAkJ0d6uUkLFihVRsGBB9O3bF+XKlQMA7N27N9NxSq975efnhzNnzmTa/+2337527a93SfHixfV+e+tdr4uid+ouBk1Ks7CwkOjoaBERcXZ2lpMnT4qIyOXLl8XR0VGVmP7+/lKmTBnt5u/vL66urmJsbCzz589XPF7JkiVl06ZNIiJy+vRpMTMzk2HDhknFihWlU6dOisdLV79+fRk5cqSIiFhbW8v169clJSVFWrRoIc2bN1ctLimnSJEisnz5chFJ+z+8du2aiIiMGjVKevXqpUpMKysriYqKEhGR0aNHa/9Wjh07Ji4uLorG0mg0/7kZGRkpGvN99ujRI+0WGRkplSr9X3v3H9Xkdf8B/J2ggSi/lIJYJEYGykBLc2YrCgWBdtWuO2fDddMWaFmlHrHCiEE504lVpFZXaylUS2tB0ckUeqqzamtX+V0QSyo//IGB0DjFVmGiGVM0ud8//JIZQUXyPHn48Xmdk3OW+3Dy+WyTPB/uc+/nzmTHjh1jV65cMbvW0dHBS/yJEyey8vLyHuOVlZVMLpfzEnMwoxkWC3V0dMBgMPSYaWhvb8eIESM4e2bdzd3dHe3t7Zg4cSJkMhkqKysREBAArVbL226We09q7Z62nD17Ni/Tp1qtFn5+fgCAwsJCvPjii0hPT0dNTQ0vJ9F227hxIyIiInDixAl0dXVh+fLlaGhoQHt7O8rLy3mLS7ij0+lMi+ClUimuX78OAIiOjkZgYCAyMzM5jymRSNDZ2QngzoLUmJgYAHda23P9uPTenSt8Ero1vzXcu4aEMYaIiAizn2E8Lrod7H1RrI0KFgvNnz8fv/71rxEfH282vnfvXhw4cACHDh3iNF54eDgOHDgAhUKB2NhYJCUloaCgACdOnOB8z343ay9ktOYN4G5Tp07F2bNnkZWVBQcHB+j1ekRGRmLJkiW9fqmQgUeIgj44OBhKpRJBQUE4fvy4qclgY2MjJkyYwEvMvvrVr36FTz75pF//foU8pM9ahNo63W2w90WxNlp0a6GxY8eivLzcbFcJAJw5cwZBQUFoa2vjNJ7RaITRaDQtyMrPz0dFRQV8fHywaNEiXjpOXrhwAYWFhWhsbARwp0dAZGQkPDw8OI8F3Hn+3tXVhaCgIKxbtw5arRYeHh746quv8Oabb5ry4MONGzdQW1vbax8GrtcFEO4tXLgQnp6eSE1NRVZWFpKTkxEUFGQq6Ldv3855TJ1Oh/j4eJw/fx4JCQl4/fXXAdzZdWYwGJCRkcF5zL66e/E4GXg2btyIjRs3YtOmTb32RRkMW42tiQoWC40ePRqVlZWmltXd6urqMGPGDNNMwWD14YcfQqlUoqury/R469q1a5BIJNi8eXOPmSUuCHUDOHLkCKKjo9He3t7jr3G+poQJt4Qo6AcyrgoWax/SJ4ScnBzY29vjpZdeMhvv7uZ799lKXGGMISUlBRkZGT36oqxevZrzeIOeUItnhorZs2ezN998s8d4fHw8Cw4O5iVmSUkJe+WVV1hgYCD717/+xRhjbOfOnay0tJTTOAcPHmQ2NjZs2bJl7OLFi6bxixcvsqSkJDZixAjT4tihwNvbm8XHx7NLly4JnQoZZDQaDVu5ciWbP38++/HHHxljjB06dIjV19cLmtfdC48t0b2Y936vocDHx4d98803PcaLiorY5MmTeY19/fp1dvz4cVZXV8du3LjBa6zBjAoWC5WVlTE7Ozv2zDPPsDVr1rA1a9awZ555htnZ2bGSkhLO4xUUFDCpVMoWLlzIbG1tTV9GH3zwAZs7dy6nsUJDQ9nKlSvve33lypUsNDSU05jdhLgBODg4MI1Gw9vnE+uwVkHfraioiEmlUvbss88yiURi+p18++23Bd9dxlXB8v3335u9qqurWXZ2NvP19WWFhYUcZCo8W1tb026vu2m1WmZnZ2f9hEgPVLBwQK1Ws5dffpn5+fmxX/ziFyw2NpY1NjbyEuvJJ59kO3bsYIyZfxnV1NRwvoXSwcGBnTlz5r7Xz5w5wxwcHDiNyZhwN4DY2Fj2ySef8Pb5hH/WLOi7BQYGsnfffZcxZv47WVVVxTw8PHiJ2VdcFSz3c/DgQd7+aLE2T09Ptn///h7jn3/+ueD/P5I7aJcQB5588kns3r3bKrHOnj2LkJCQHuNOTk64evUqp7EMBkOPzr13GzlyJC/rOlJSUpCWlgalUmnapQDc2SHFx7bUbpmZmXjppZdQWlqKadOm9fjvnpCQwFtswo20tDRs27YNMTExyM/PN40HBQUhLS2Nl5h1dXW9dnp1c3MbMufs3A9fh/QJYcGCBUhISICDg4PpO7a4uBiJiYmYP3++wNkRgLY1c8JoNEKj0fS6s6S34sIS7u7u0Gg0kMvlZuNlZWWc7wTw9/fH/v37kZSU1Ov1zz//HP7+/pzGBIS7AezZswdfffUV7OzsUFRUZNafQSQSUcEyCFizoO/m7OyM1tbWHltT1Wo1bzvp+urPf/4zJ92orX1InxDWrVuHlpYWREREmBZtG41GxMTEID09XeDsCABadGupb7/9lk2aNImJxWKrdJxMT09nfn5+rLKykjk4OLDS0lK2a9cu5urqyjIyMjiNlZuby6RSKcvKymK3bt0yjd+6dYtlZmYyqVTKcnJyOI3JGGMeHh6m7o93T2l/9tlnzMvLi/N43caNG8fWr1/PDAYDbzEIvyZNmsSOHj3KGDP/t7Njxw7285//nJeYy5YtY8HBway1tZU5ODiwc+fOsbKyMubl5cXWrFnDS8zc3Fx28OBB0/vk5GTm5OTEZs6caeqEzaXeFt2KRCImk8l67dQ6mJ09e5bt3buX/eMf/+Dlf0vSf1SwWCggIIC99NJL7NSpU+zf//43u3r1qtmLCydPnjS7iaalpbHRo0ebCiM7OztTS3muLVu2jIlEIubo6Ghqy+/o6MjEYjH705/+xFtMa98AGGNszJgxtOh2kLNmQd/t5s2bbOHChWzEiBFMJBKxkSNHMrFYzKKiotjt27d5iTl58mT2z3/+kzHGWEVFBRs1ahT76KOP2K9//Wv229/+lvN4RUVFZq+SkhJ2+vRpsz9kCOEb9WGx0OjRo3Hy5El4e3vzFsPGxgatra1wc3ODl5cXqqur4eDgAI1GA71eDz8/P9jb2/MWv7KyEnv27MG5c+cAAJMnT8b8+fMRGBjIS7yuri4sWbIEubm5MBgMGDFiBAwGA15++WXk5ubydthaUlISXF1d8ec//5mXzyf8qK2txdSpUyEWiwEA69evx9tvv23qgWRrawuVSoV169ZxHpsxhvPnz8PV1RVXrlxBXV0d9Ho9FAoFr49KRo0ahTNnzkAmk2HFihVobW3Fzp070dDQgNmzZ+Py5cucxnv77beH5CF9SqWyzz+7efNmHjMhfUEFi4XCw8OxfPlyzJkzh7cYLi4uOHToEGbMmAGxWIwff/wRrq6uvMXrr/j4eKxduxaPPfZYvz9DqBsAcGdR7c6dOxEQEIAnnniix6Jb+sIamIQs6I1GI+zs7NDQ0GDVtRxubm748ssvoVAooFAooFQqER0djaamJgQEBECv13MaTy6X429/+5vpnKZuVVVVmD9/PrRaLafxrCUsLKxPPycSifDNN9/wnA15GFp02w+1tbWm/7x06VIsW7YMly5d6nVnyRNPPGFxvHnz5iE0NBTjx4+HSCTC9OnT7zvL0NzcbHG8/tq1axdUKpXFBYu3t7fpBuDp6clhhg9WV1cHhUIBAKivrze79qCD34iwnJ2dodVq4ebmhpaWFhiNRkgkEtMBmnwSi8Xw8fFBW1ubVQuW5557DgsXLoRCoUBjY6PpUNCGhoYeC/K5MFQP6RP6LCHyaKhg6Ycnn3wSIpHIrH373VOl3de4aueenZ2NyMhIaDQaJCQkIC4uzmy770DBxWSdUDcAgL68BiuhC/oNGzYgOTkZW7duxdSpUzn//N5kZWVh1apVOH/+PAoLC+Hi4gIA+O6777BgwQLO4w2nQ/o0Gg2ampoQEhICqVRq+i4nwqNHQv3www8/9PlnJ06cyGns2NhYZGRkDMiChatzS/7xj39g48aNVr0BkMHtyJEjpoJ+7dq19/39SExM5Dz2mDFj0NnZidu3b0MikUAqlZpdb29v5zymtQ2HQ/ra2trw+9//HseOHYNIJMK5c+fg5eWFP/7xjxgzZgzeffddoVMc9qhgIZzhqmAZDjcAwg8hCvodO3Y88Dofh+YBQGlpKT766CM0Nzdj37598PDwQF5eHiZNmoTg4GBOY7FhcEhfTEwMfvrpJ3zyySf4+c9/bvou+/LLL6FUKtHQ0CB0isMeFSwWOnDgQK/jIpEIdnZ28Pb27jGNOlRxVbAIdQMgZLAoLCxEdHQ0XnnlFeTl5eHUqVPw8vJCZmYmDh06hEOHDvESV6/X4/Tp05BKpfDx8YGtrS0vcYTg7u6OL7/8EgEBAWbfZc3NzXjiiSc4X8hMHh2tYbHQb37zmx7rWQDzdSzBwcH4/PPPMWbMGIGyHFyoICGDiU6ne+B1mUzGeUwhjiAAAHt7ezz11FO8fb6Q/vOf/2DUqFE9xtvb24dUYTaYiYVOYLA7evQonnrqKRw9ehQdHR3o6OjA0aNHMWPGDBw8eBAlJSVoa2uDSqUSOlXeRUVFwdHR0eLP0el0D3wRMpDI5XJMmjTpvi8+CHEEwVD3zDPPYOfOnab3IpEIRqMRGzdu7PP2Z8IvmmGxUGJiIrKzs836E0RERMDOzg5vvPEGGhoasGXLlh4NlwaTI0eOwN7e3vRcPCsrCx9//DH8/PyQlZVlmjnaunUrJ/HkcvkDV+XzceAiIf2lVqvN3t+6dQtqtRqbN2/G+vXreYlpzTPFhouNGzciIiICJ06cQFdXF5YvX46Ghga0t7ejvLxc6PQIqGCxWFNTU6+zCo6OjqYtlD4+PoP61Nbk5GS88847AO70Klm2bBmUSiWOHTsGpVKJnJwcTuMJcQMgpL8CAgJ6jE2fPh2PP/44Nm3ahMjISM5jxsXFITExEZ9++ilEIhEuXryIb7/9FiqVCn/5y184jzccTJ06FY2Njfjggw/g4OAAvV6PyMhILFmypNceNEQA1j0JYOgJCgpic+bMYT/99JNp7KeffmJz5sxhzzzzDGOMsaNHj7LJkycLlaLFRo8ezbRaLWOMsdTUVDZv3jzGGGPfffcdGzdunNXyOHjwIAsNDbVaPEIsce7cOTZq1ChePttoNFr1TDFCBgJaw2Kh7du3Q6vVYsKECfD29oa3tzcmTJiAlpYWfPLJJwDurKxftWqVwJn2n0QiMZ3L8vXXX+OXv/wlAGDs2LE9jp3n05QpU1BdXW21eIT0xbVr18xeHR0dOHPmDFatWsVb80ORSISVK1eivb0d9fX1qKysxOXLl3k5L2k4KS0tRVRUFGbNmoULFy4AAPLy8lBWViZwZgSgR0IWmzJlCk6dOoWvvvoKjY2NprHnnnvOdBjbb37zGwEztFxwcDCUSiWCgoJw/Phx/P3vfwcANDY2YsKECZzHu7cIYoyhtbUVa9assXr3W0IextnZuceaK8YYPD09zXbw8MFaRxAMB3dvFa+pqcHNmzcBAB0dHUhPT+dtqzjpO+rDQh5Kp9MhPj4e58+fR0JCAl5//XUAd043NhgMyMjI4DSeWCx+4A1g5syZnMYjxBLFxcVm78ViMVxdXeHt7Y0RI/j5m/A///kPNmzYgH/+85/46aefYDQaza4LeabYYKVQKJCUlISYmBizPixqtRpz587FpUuXhE5x2KMZln7IyMjAG2+8ATs7u4ferBMSEqyUFX9kMhkOHjzYY/y9997jJd69Z/pY4wZASH+JRCLMmjWrx7/N27dvo6SkpNftx5ZauHAhiouLER0dbTpDiViGtooPfPTt3w/vvfceXnnlFdjZ2T3wpi0SiYZEwQLc2Q2Vk5ODpqYmvP/++3Bzc8Phw4chk8ng7+/PaSwhbgCE9FdYWBhaW1vh5uZmNt7R0YGwsDBetuEfPnwYX3zxBYKCgjj/7OGKtooPfLToth+0Wq3pdFStVnvf11CZli0uLsa0adNQVVWFzz77zNSi+uTJk0hNTeU8XlhYWK/nBXXfAAgZSNh9TvNta2vD6NGjeYk5ZswYjB07lpfPHq66t4pXVVWZtorv3r0bKpUKixcvFjo9ApphIX2QkpKCtLQ0KJVKs0PlwsPDkZmZyXk8IW4AhDyq7v4qIpEIr732mln7doPBgNraWrOGklxat24dVq9ejR07dvTaTp48upSUFBiNRkRERKCzsxMhISGwtbWFSqXC0qVLhU6PgAqWflMqlX36uc2bN/OcCf/q6urwt7/9rce4m5sbpw3xhLwBEPKonJycANwpsB0cHMxOFZdIJAgMDERcXBwvsd999100NTVh3LhxkMvlGDlypNn1mpoaXuIOZd1bxZOTk6HRaKDX6+Hn5wd7e3uhUyP/jwqWfrq3G2tZWRl+8YtfmH1pDZWFcM7Ozmhtbe1xLoparYaHhwdncYS8ARDyqLo7PMvlcqhUKqvO/g32VgkDGW0VH7hoWzNH7t4GN9SoVCpUVVVh3759mDx5MmpqavDjjz8iJiYGMTExnK9jeeutt6x+AyCEDG+0VXzgo4KFI0O5YOnq6sKSJUuQm5sLg8GAESNGwGAw4OWXX0Zubi5sbGyETpEQQRUUFGDv3r3Q6XTo6uoyu0aPZwaHBQsWPHCreGJiokCZkW5UsHBkqBYsjDGcP38erq6uuHLlCurq6qDX66FQKHjtOks3ADJYZGRkYOXKlXjttdeQnZ2N2NhYNDU1obq6GkuWLOHswM6xY8eisbERjz32GMaMGfPAR8697bIjD+bs7ExbxQc4WsNCHogxBm9vbzQ0NMDHxweenp68x7z7BrB///4eNwBCBpIPP/wQ2dnZWLBgAXJzc7F8+XJ4eXlh9erVnBYO7733nmmX3nvvvTdk1sgNFLRVfOCjGZZ+qq2tNXs/a9Ys7N27t8fZOk888YQ10+KFv78/tm/fjsDAQKvE8/X1RWpqKhYsWGA2c9V9A+BjKzUh/TVq1CicPn0aEydOhJubG44ePYqAgACcO3cOgYGBaGtrEzpF0ge7du3C/v37aav4AEYzLP305JNPQiQS4e5678UXXwQA07hIJOKly6W1bdiwAcnJydi6dSumTp3KezydTmfaviyVSnH9+nUAQHR0NAIDA6lgIQOKu7s72tvbMXHiRMhkMlRWViIgIABarRZ8/T1YU1ODkSNHYtq0aQCA/fv3IycnB35+flizZg0kEgkvcYcahUJhNlOl0Whoq/gARgVLP2m1WqFTsJqYmBh0dnYiICAAEonEbLsxwP3zciFuAIT0V3h4OA4cOACFQoHY2FgkJSWhoKAAJ06cMPUW4tqiRYuQkpKCadOmobm5GX/4wx8QGRmJffv2obOzE1u2bOEl7lBD28MHF3okZCXx8fFYu3YtHnvsMaFTeWQ7dux44PVXX32V03gLFy6Ep6cnUlNTkZWVheTkZAQFBZluANu3b+c0HiGWMBqNMBqNprOv8vPzUVFRAR8fHyxatIiX2Q4nJyfU1NTgZz/7Gd555x188803+PLLL1FeXo758+fj/PnznMckRGhUsFiJo6Mjvv/++yG3i4gPQtwACBlMHB0d8d1338HHxwfPPfccXnzxRSQmJkKn02HKlCn473//K3SKg8arr76KiIgIzJ49GzKZTOh0yANQwWIlg3nbs06ne+B1+iUnw11paSk++ugjNDU1oaCgAB4eHsjLy8OkSZMQHBzMebzw8HB4enri2Wefxeuvv45Tp07B29sbxcXFePXVV9HS0sJ5zKFq9uzZqKqqQldXF+RyOcLCwhAWFobw8HCMHz9e6PTIXei0ZvJQcrkckyZNuu+LD6WlpYiKisLMmTNx4cIFAEBeXh7Kysp4iUdIfxUWFuL555+HVCqFWq3GzZs3Adw5XTw9PZ2XmFu2bEFNTQ3efPNNrFy5Et7e3gDu9C+i87YeTVFREa5evYqvv/4aUVFROHfuHF5//XVMmDABvr6+WLx4Mfbt2yd0mgQ0w2I1g3mG5eTJk2bvb926BbVajc2bN2P9+vWcLywsLCxEdHQ0XnnlFeTl5eHUqVPw8vJCZmYmDh06hEOHDnEajxBLKBQKJCUlISYmxuz3XK1WY+7cubh06ZLVcrlx4wZsbGx67HAhj+bGjRuoqKjA4cOHkZ2dDb1ePyR2fA52tEuIPFRAQECPsenTp+Pxxx/Hpk2bOC9Y0tLSsG3bNsTExCA/P980HhQUhLS0NE5jEWKps2fPIiQkpMe4k5MTrl69atVc7OzsrBpvqOnq6sK3336LoqIiHDt2DFVVVXj88ccxb948oVMjoIKFWGDKlCmorq7m/HMH0g2AkIdxd3eHRqOBXC43Gy8rK+N0RpVa8/OjpKTErECRyWQIDQ3FG2+8gV27dvVoBkqEQwWLlURFRcHR0VHoNPrl2rVrZu8ZY2htbcWaNWt4OU/IWjcAQrgQFxeHxMREfPrppxCJRLh48SK+/fZbqFQq/OUvf+Eszt2t+anPCne6dwetWLEC+fn5GDdunNApkfthxCKHDx9mpaWlpveZmZksICCALViwgLW3twuYGXdEIhETi8VmL5FIxGQyGauoqOA8Xnp6OvPz82OVlZXMwcGBlZaWsl27djFXV1eWkZHBeTxCHtXJkyeZwWAwvU9LS2OjR49mIpGIiUQiZmdnx1atWiVghqSvVqxYwWbMmMEkEgmbNm0ae/PNN1lBQQG7fPmy0KmRe9CiWwtNmzYN77zzDl544QXU1dXhqaeeglKpxLFjx+Dr64ucnByhU7RYcXGx2XuxWAxXV1d4e3ubeqVwiTGG9PR0vP322+js7AQA2NraQqVSYd26dZzHI+RR2djYoLW1FW5ubvDy8kJ1dTUcHByg0Wig1+vh5+cHe3t73uJTqwHu6fV6lJaWoqioCEVFRVCr1Zg8eTJCQ0MRFhaG3/3ud0KnOOxRwWIhe3t71NfXQy6XY82aNaivr0dBQQFqamrwwgsvWHWHAF9KSkowa9asHsXJ7du3UVFR0et6k0dVW1uLqVOnQiz+3077rq4uq90ACHkULi4uOHToEGbMmAGxWIwff/wRrq6uVosvFosfuIaFdrRYrr29HZs3b8YHH3xAu4QGCFrDYiGJRGKaBfj6668RExMD4M4CuXvXfgxWYWFhpr8m79bR0YGwsDBOfpEVCkWPv1hdXFzg5+dn8WcTwrV58+YhNDQU48ePh0gkwvTp02FjY9PrzzY3N3MeX61Wm72/t9UAeXRGoxHV1dWmGZby8nLo9XrIZDLezoQij4YKFgsFBwdDqVQiKCgIx48fx9///ncAQGNj45BZXc7+/+Tpe7W1tWH06NGcxHB2doZWq4WbmxtaWlpgNBo5+VxC+JCdnY3IyEhoNBokJCQgLi7OtCDWGqzdamAo27hxo6lAuX79Ojw8PDB79mxs2bIFYWFhvDXHJI+OChYLZWZmIj4+HgUFBdi6dSs8PDwAAIcPH8acOXMEzs4y3V96IpEIr732GmxtbU3XDAYDamtrOeuqKfRfrIQ8qu7f7++++w6JiYlWLVjuh69WA0PZli1bMHv2bPz1r39FWFiYqWswGXhoDQu5r9jYWAB3Tmv+/e9/D6lUaromkUggl8sRFxfH2QnUR44cMf3Funbt2vveABITEzmJR8hg9aBWA2fOnMH3338vTGLDQHx8PNauXcvZ9x7pOypYONDU1IScnBw0NTXh/fffh5ubGw4fPgyZTAZ/f3+h07PYW2+9BZVKxdnjn4eJjY1FRkbGgPiLlZCBqLdFt4wxeHp6Ys+ePXSeEI8cHR3x/fffU08oAVDBYqHi4mLMnTsXQUFBKCkpwenTp+Hl5YUNGzbgxIkTKCgoEDpFQsgQY+1WA+R/BvO5cIMd/cu2UEpKCtLS0qBUKs1mBMLDw5GZmSlgZtwqKCjA3r17odPp0NXVZXatpqZGoKwIGZ4qKiowbtw4/PGPfzQb//TTT3H58mWsWLFCoMwI4Y/44T9CHqSurg6//e1ve4y7ubnhypUrAmTEvYyMDMTGxmLcuHFQq9V4+umn4eLigubmZsydO1fo9AgZdj766CP4+vr2GPf398e2bdsEyIgQ/lHBYiFnZ2e0trb2GFer1aYdQ4Pdhx9+iOzsbHzwwQeQSCRYvnw5jh49ioSEBHR0dAidHiHDzqVLlzB+/Pge466urr1+HxEyFFDBYqH58+djxYoVuHTpEkQiEYxGI8rLy6FSqUxN5AY7nU5nWsQnlUpx/fp1AEB0dDT27NkjZGqEDEuenp4oLy/vMV5eXo7HH39cgIwI4R8VLBZKT0+Hr68vPD09TS3kQ0JCMGvWLKxatUro9Djh7u5uOq5eJpOhsrISAKDVakFrtgmxvri4OPzpT39CTk4OfvjhB/zwww/49NNPkZSUhLi4OKHTG9KioqLg6OgodBrDEu0SsgBjDOfPn4erqyuuXLmCuro66PV6KBQK+Pj4CJ0eZxYuXAhPT0+kpqYiKysLycnJCAoKwokTJxAZGYnt27cLnSIhwwpjDCkpKcjIyDAtgrezs8OKFSuwevVqgbMbnI4cOQJ7e3sEBwcDALKysvDxxx/Dz88PWVlZGDNmjMAZEipYLGA0GmFnZ4eGhoYhVaDcy2g0wmg0mrZL5ufno6KiAj4+Pli0aBEkEonAGRIyPOn1epw+fRpSqRQ+Pj5m3ajJo5k2bRreeecdvPDCC6irq8NTTz0FpVKJY8eOwdfXFzk5OUKnOOxRwWIhf39/bN++HYGBgUKnQgghpJ/s7e1RX18PuVyONWvWoL6+HgUFBaipqcELL7yAS5cuCZ3isEdrWCy0YcMGJCcno76+XuhUeFVaWoqoqCjMnDkTFy5cAADk5eWhrKxM4MwIIcRyEokEnZ2dAICvv/4av/zlLwEAY8eO7XEUAhEGFSwWiomJwfHjxxEQEACpVIqxY8eavYaCwsJCPP/885BKpVCr1bh58yYAoKOjA+np6QJnRwghlgsODoZSqcS6detw/Phx/OpXvwIANDY2YsKECQJnRwDqdGuxLVu2CJ0C79LS0rBt2zbExMQgPz/fNB4UFIS0tDQBMyOEEG5kZmYiPj4eBQUF2Lp1q6mP1uHDh00ncxNh0RoW8lCjRo3CqVOnIJfLzc7RaG5uhp+fH27cuCF0ioQQQoY4mmGxkE6ne+B1mUxmpUz44+7uDo1GA7lcbjZeVlZGB4ARQoaMpqYm5OTkoKmpCe+//z7c3Nxw+PBhyGQy+Pv7C53esEdrWCwkl8sxadKk+76Ggri4OCQmJqKqqgoikQgXL17E7t27oVKpsHjxYqHTI4QQixUXF2PatGmoqqrCZ599Br1eDwA4efIkUlNTBc6OADTDYjG1Wm32/tatW1Cr1di8eTPWr18vUFbcSklJgdFoREREBDo7OxESEgJbW1uoVCosXbpU6PQIIcRiKSkpSEtLg1KphIODg2k8PDwcmZmZAmZGutEaFp588cUX2LRpE4qKioROpV9qa2sxdepUiMX/m4Tr6uqCRqMxHUFgb28vYIaEEMIde3t71NXVYdKkSWZr9VpaWuDr60tr9QYAeiTEkylTpqC6ulroNPpNoVDgypUrAAAvLy+0tbVBIpHAz88PTz/9NBUrhJAhxdnZudeTrtVqtWnHEBEWFSwWunbtmtmro6MDZ86cwapVqwZ1u35nZ2dotVoAQEtLC4xGo8AZEUIIf+bPn48VK1bg0qVLEIlEMBqNKC8vh0qlQkxMjNDpEdAjIYuJxWKIRCKzMcYYPD09kZ+fj5kzZwqUmWXeeOMN7Ny5E+PHj4dOp8OECRNgY2PT6882NzdbOTtCCOFWV1cXlixZgtzcXBgMBowYMQIGgwEvv/wycnNz7/v9R6yHChYLFRcXm70Xi8VwdXWFt7e36bDAwerIkSPQaDRISEjA2rVrzRai3S0xMdHKmRFCCHcYYzh//jxcXV1x5coV1NXVQa/XQ6FQDOqZ8qGGChYLlZSUYNasWT2Kk9u3b6OiogIhISECZcad2NhYZGRk3LdgIYSQwcxoNMLOzg4NDQ1UoAxgVLBYyMbGBq2trXBzczMbb2trg5ubGwwGg0CZEUII6St/f39s374dgYGBQqdC7oMW3VqIMdZjDQtwp2AZPXq0ABkRQgh5VBs2bEBycjLq6+uFToXcB82w9FNkZCQAYP/+/ZgzZw5sbW1N1wwGA2prazFlyhQcOXJEqBQJIYT00ZgxY9DZ2Ynbt29DIpFAKpWaXW9vbxcoM9JtcK8KFZCTkxOAOzMsDg4OZv+4JRIJAgMDERcXJ1R6hBBCHsGWLVuEToE8BM2wWOitt96CSqWixz+EEEIIj6hgIYQQMuzpdLoHXpfJZFbKhNwPFSwcKCgowN69e6HT6dDV1WV2raamRqCsCCGE9FVvTUDvRjs+hUe7hCyUkZGB2NhYjBs3Dmq1Gk8//TRcXFzQ3NyMuXPnCp0eIYSQPlCr1aipqTG9qqqqsG3bNkyePBn79u0TOj0CmmGxmK+vL1JTU7FgwQKzEz5Xr16N9vZ2OpacEEIGsS+++AKbNm1CUVGR0KkMezTDYiGdTodZs2YBAKRSKa5fvw4AiI6Oxp49e4RMjRBCiIWmTJmC6upqodMgoILFYu7u7qb9+TKZDJWVlQAArVYLmrwihJDB4dq1a2avjo4OnDlzBqtWraJ2/QME9WGxUHh4OA4cOACFQoHY2FgkJSWhoKAAJ06cMDWXI4QQMrA5Ozv3WHTLGIOnpyfy8/MFyorcjdawWMhoNMJoNJoOP8zPz0dFRQV8fHywaNEiSCQSgTMkhBDyMMXFxWbvxWIxXF1d4e3t3eNwWyIMKlgIIYQMeyUlJZg1a1aP4uT27duoqKhASEiIQJmRbrSGhQOlpaWIiorCzJkzceHCBQBAXl4eysrKBM6MEEJIX4SFhfV6XlBHRwfCwsIEyIjciwoWCxUWFuL555+HVCqFWq3GzZs3Adz5R56eni5wdoQQQvqCMdZr47i2tjY6emWAoAdzFkpLS8O2bdsQExNjtjArKCgIaWlpAmZGCCHkYbo3R4hEIrz22muwtbU1XTMYDKitrTW1riDCooLFQmfPnu312aaTkxOuXr1q/YQIIYT0mZOTE4A7MywODg6QSqWmaxKJBIGBgYiLixMqPXIXKlgs5O7uDo1GA7lcbjZeVlYGLy8vYZIihBDSJzk5OQAAuVwOlUpFj38GMFrDYqG4uDgkJiaiqqoKIpEIFy9exO7du6FSqbB48WKh0yOEENIHqampVKwMcDTDYqGUlBQYjUZERESgs7MTISEhsLW1hUqlwtKlS4VOjxBCSB8VFBRg79690Ol06OrqMrtWU1MjUFakG82w9ENtbS2MRiOAOwu1Vq5cifb2dtTX16OyshKXL1/GunXrBM6SEEJIX2VkZCA2Nhbjxo2DWq3G008/DRcXFzQ3N2Pu3LlCp0dAjeP6xcbGBq2trXBzc4OXlxeqq6vh4uIidFqEEEL6ydfXF6mpqViwYAEcHBxw8uRJeHl5YfXq1Whvb0dmZqbQKQ57NMPSD87OztBqtQCAlpYW02wLIYSQwUmn05m2L0ulUly/fh0AEB0djT179giZGvl/tIalH+bNm4fQ0FCMHz8eIpEI06dPh42NTa8/29zcbOXsCCGEPCp3d3e0t7dj4sSJkMlkqKysREBAALRaLehBxMBABUs/ZGdnIzIyEhqNBgkJCYiLi4ODg4PQaRFCCOmn8PBwHDhwAAqFArGxsUhKSkJBQQFOnDhhai5HhEVrWCwUGxuLjIwMKlgIIWQQMxqNMBqNpsMP8/PzUVFRAR8fHyxatAgSiUTgDAkVLIQQQggZ8GjRLSGEEAKgtLQUUVFRmDlzJi5cuAAAyMvLQ1lZmcCZEYAKFkIIIQSFhYV4/vnnIZVKoVarcfPmTQBAR0cH0tPTBc6OAFSwEEIIIUhLS8O2bdvw8ccfY+TIkabxoKAg6nI7QFDBQgghZNg7e/YsQkJCeow7OTnh6tWr1k+I9EAFCyGEkGHP3d0dGo2mx3hZWRm8vLwEyIjciwoWQgghw15cXBwSExNRVVUFkUiEixcvYvfu3VCpVFi8eLHQ6RFQ4zhCCCEEKSkpMBqNiIiIQGdnJ0JCQmBrawuVSoWlS5cKnR4B9WEhhBAyTNXW1mLq1KkQi//3sKGrqwsajQZ6vR5+fn6wt7cXMENyNypYCCGEDEs2NjZobW2Fm5sbvLy8UF1dDRcXF6HTIvdBa1gIIYQMS87OztBqtQCAlpYWGI1GgTMiD0JrWAghhAxL8+bNQ2hoKMaPHw+RSITp06fDxsam159tbm62cnbkXlSwEEIIGZays7MRGRkJjUaDhIQExMXF0UG2AxitYSGEEDLsxcbGIiMjgwqWAYwKFkIIIYQMeLTolhBCCCEDHhUshBBCCBnwqGAhhBBCyIBHBQshhBBCBjwqWAghhBAy4FHBQgghhJABjwoWQgghhAx4/welEGd9/XYa0QAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 189,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "show_feature_importances(find_the_best_model(italian_clf_dict)[1], X_columns)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "id": "8e95d002",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "score:\t  feature:\n",
      "0.349\t- features_Highchairs Available\n",
      "0.289\t- features_Seating\n",
      "0.101\t- features_Outdoor Seating\n",
      "0.055\t- meals_Breakfast\n",
      "0.035\t- features_Table Service\n",
      "0.023\t- meals_Lunch\n",
      "0.018\t- price_level\n",
      "0.017\t- features_Takeout\n",
      "0.017\t- cuisines_Seafood\n",
      "0.017\t- cuisines_Mediterranean\n",
      "0.015\t- meals_Dinner\n",
      "0.014\t- features_Reservations\n",
      "0.013\t- cuisines_Italian\n",
      "0.013\t- features_Wheelchair Accessible\n",
      "0.012\t- cuisines_Pizza\n",
      "0.011\t- region\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 190,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "show_feature_importances(italian_clf_dict['xgb']['model'], X_columns)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b923d91f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "535a307d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# ENSEMBLE MODELS\n",
    "\n",
    "# Voting Classifier\n",
    "\n",
    "\n",
    "# Bagging Classifier - based on forest\n",
    "bag_on_forest_grid = {\n",
    "    \"estimator\": [forest],\n",
    "    \"n_estimators\": [40],\n",
    "    \"max_features\": [7],\n",
    "    \"max_samples\": [0.5],\n",
    "    \"bootstrap\": [True],\n",
    "    \"random_state\": [42]\n",
    "}\n",
    "\n",
    "# AdaBoost - based on forest\n",
    "ada_on_forest_grid = {\n",
    "    \"estimator\": [forest],\n",
    "    \"n_estimators\": [60],\n",
    "    \"learning_rate\": [0.01],\n",
    "    \"random_state\": [42]\n",
    "}\n",
    "\n",
    "# AdaBoost - based on logreg\n",
    "ada_on_logreg_grid = {\n",
    "    \"estimator\": [logreg],\n",
    "    \"n_estimators\": [60],\n",
    "    \"learning_rate\": [0.01],\n",
    "    \"random_state\": [42]\n",
    "}\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "78266681",
   "metadata": {},
   "outputs": [],
   "source": [
    "# PIVOT DICTIONARY OF CLASSIFICATION MODELS\n",
    "clf_ensemble_models_dict = { \n",
    "#     ensemble models (based on other models)\n",
    "    'bag_on_forest': {'status': True, 'model': BaggingClassifier, 'grid': bag_on_forest_grid},\n",
    "    'ada_on_forest': {'status': True, 'model': AdaBoostClassifier, 'grid': ada_on_forest_grid},\n",
    "    'ada_on_logreg': {'status': True, 'model': AdaBoostClassifier, 'grid': ada_on_logreg_grid}\n",
    "}\n",
    "\n",
    "italian_ensemble_clf_dict = run_classification_models(italian_df2, 'features_Serves Alcohol', [], clf_ensemble_models_dict)\n",
    "# drop_columns = ['special_diets_Gluten Free Options', 'special_diets_Vegan Options']\n",
    "italian_ensemble_clf_dict\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7434d0e9",
   "metadata": {},
   "outputs": [],
   "source": [
    "find_the_best_model(italian_ensemble_clf_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4ade7813",
   "metadata": {},
   "outputs": [],
   "source": [
    "show_feature_importances(find_the_best_model(italian_ensemble_clf_dict)[1], X_columns)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "eae3996f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8e1d90d1",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "802d37a5",
   "metadata": {},
   "source": [
    "<a id=\"voting\"></a>\n",
    "\n",
    "### Voting Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "id": "105e7fb0",
   "metadata": {},
   "outputs": [],
   "source": [
    "def clf_voting_by_weights(weights_list): \n",
    "    clf_voting = VotingClassifier(\n",
    "        estimators=[\n",
    "            ('logreg', logreg),\n",
    "            ('svm', svm),\n",
    "#             ('nb', nb),\n",
    "            ('tree', tree)\n",
    "        ],\n",
    "        weights=weights_list\n",
    "    )\n",
    "\n",
    "    clf_voting.fit(X_train, y_train)\n",
    "    y_pred_clf_voting = clf_voting.predict(X_test)\n",
    "\n",
    "    return accuracy_score(y_test, y_pred_clf_voting)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "id": "ed38a4e3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1, 1, 1]\n",
      "[1, 1, 2]\n",
      "[1, 1, 3]\n",
      "[1, 2, 1]\n",
      "[1, 2, 2]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1, 2, 3]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1, 3, 1]\n",
      "[1, 3, 2]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1, 3, 3]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[2, 1, 1]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[2, 1, 2]\n",
      "[2, 1, 3]\n",
      "[2, 2, 1]\n",
      "[2, 2, 2]\n",
      "[2, 2, 3]\n",
      "[2, 3, 1]\n",
      "[2, 3, 2]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[2, 3, 3]\n",
      "[3, 1, 1]\n",
      "[3, 1, 2]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[3, 1, 3]\n",
      "[3, 2, 1]\n",
      "[3, 2, 2]\n",
      "[3, 2, 3]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[3, 3, 1]\n",
      "[3, 3, 2]\n",
      "[3, 3, 3]\n",
      "\n",
      "0.7601101359490621\n",
      "[1, 1, 3]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "from itertools import product\n",
    "\n",
    "rrange = [1,2,3]\n",
    "max_score = 0\n",
    "weights_for_max = []\n",
    "\n",
    "for p in product(rrange, repeat=3):\n",
    "    weights = list(p)\n",
    "    print(weights)\n",
    "    score = clf_voting_by_weights(weights)\n",
    "    if score > max_score:\n",
    "        max_score = score\n",
    "        weights_for_max = weights\n",
    "print()\n",
    "print(max_score)\n",
    "print(weights_for_max)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "17f2f752",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "fb597065",
   "metadata": {},
   "source": [
    "<a id=\"stacked\"></a>\n",
    "\n",
    "### Stacked Ensembles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "id": "bce0ed6a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix:\n",
      " [[3658 1531]\n",
      " [1169 5264]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.76      0.70      0.73      5189\n",
      "           1       0.77      0.82      0.80      6433\n",
      "\n",
      "    accuracy                           0.77     11622\n",
      "   macro avg       0.77      0.76      0.76     11622\n",
      "weighted avg       0.77      0.77      0.77     11622\n",
      "\n",
      "Time using clf_model_implementation(): 16.233088970184326 sec\n"
     ]
    }
   ],
   "source": [
    "start_time = time.time()\n",
    "\n",
    "# Prepare the list of tuples with the first-layer classifiers\n",
    "classifiers = [\n",
    "#     ('logreg', logreg),\n",
    "#     ('svm', svm),\n",
    "#     ('nb', nb),\n",
    "    ('knn', knn),\n",
    "    ('tree', tree)\n",
    "    ,('forest', forest)\n",
    "]\n",
    "# Instantiate the second-layer meta estimator\n",
    "clf_meta = LogisticRegression(max_iter=500)\n",
    "\n",
    "clf_stack = StackingClassifier(\n",
    "    estimators=classifiers,\n",
    "    final_estimator=clf_meta,\n",
    "#     stack_method='predict_proba',\n",
    "    passthrough = False,\n",
    "    cv=5, \n",
    "    n_jobs=-1\n",
    ")\n",
    "\n",
    "clf_stack.fit(X_train, y_train)\n",
    "y_pred_clf_stack = clf_stack.predict(X_test)\n",
    "\n",
    "print(\"Confusion matrix:\\n\", confusion_matrix(y_test, y_pred_clf_stack))\n",
    "print()\n",
    "print(\"Classification report:\\n\", classification_report(y_test, y_pred_clf_stack))\n",
    "print(\"Time using clf_model_implementation(): {} sec\".format(time.time() - start_time))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 380,
   "id": "d5a4268f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix:\n",
      " [[23613  5674]\n",
      " [ 4662 17656]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.84      0.81      0.82     29287\n",
      "           1       0.76      0.79      0.77     22318\n",
      "\n",
      "    accuracy                           0.80     51605\n",
      "   macro avg       0.80      0.80      0.80     51605\n",
      "weighted avg       0.80      0.80      0.80     51605\n",
      "\n",
      "Time using clf_model_implementation(): 64.43616819381714 sec\n"
     ]
    }
   ],
   "source": [
    "# from sklearn.ensemble\n",
    "\n",
    "from sklearn.ensemble import StackingClassifier\n",
    "\n",
    "start_time = time.time()\n",
    "\n",
    "# Prepare the list of tuples with the first-layer classifiers\n",
    "classifiers = [\n",
    "    ('clf_ada', clf_ada),\n",
    "    ('clf_xgb', clf_xgb)\n",
    "    ,('forest', forest)\n",
    "]\n",
    "\n",
    "# grid = {\n",
    "#     \"n_estimators\": [200, 300], # range(100, 400, 100)\n",
    "#     \"max_depth\": [8, 10], # range(4, 10, 2)\n",
    "#     \"min_samples_split\": [20, 30], # range(20, 40, 10)\n",
    "#     \"n_jobs\": [-1],\n",
    "#     \"random_state\": [42]\n",
    "# }\n",
    "# mod = RandomForestClassifier()\n",
    "# mod_cv = GridSearchCV(mod, grid, cv=10)\n",
    "# mod_cv.fit(X_train, y_train)\n",
    "    \n",
    "# print(\"Model: {0}\".format(namestr(model, globals())[0]))\n",
    "# if grid != {}:\n",
    "#     print(\"Tuned hyperparameters:\") # , mod_cv.best_params_\n",
    "#     for k, v in mod_cv.best_params_.items():\n",
    "#         print(\"\\t{0}: {1}\".format(k, v))\n",
    "# print()\n",
    "\n",
    "# clf_meta = RandomForestClassifier(mod_cv.best_params_) # max_iter=500 # GaussianNB # DecisionTreeClassifier\n",
    "\n",
    "# Instantiate the second-layer meta estimator\n",
    "clf_meta = RandomForestClassifier(max_depth=8,min_samples_split=20,n_estimators=300,n_jobs=-1) # max_iter=500 # GaussianNB # DecisionTreeClassifier\n",
    "\n",
    "clf_stack = StackingClassifier(\n",
    "    estimators=classifiers,\n",
    "    final_estimator=clf_meta,\n",
    "#     stack_method='predict_proba',\n",
    "    passthrough = False,\n",
    "    cv=5, \n",
    "    n_jobs=-1\n",
    ")\n",
    "\n",
    "clf_stack.fit(X_train, y_train)\n",
    "y_pred_clf_stack = clf_stack.predict(X_test)\n",
    "\n",
    "print(\"Confusion matrix:\\n\", confusion_matrix(y_test, y_pred_clf_stack))\n",
    "print()\n",
    "print(\"Classification report:\\n\", classification_report(y_test, y_pred_clf_stack))\n",
    "print(\"Time using clf_model_implementation(): {} sec\".format(time.time() - start_time))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fa7c8aa2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "bd1d041c",
   "metadata": {},
   "source": [
    "<a id=\"uk_vegetarian\"></a>\n",
    "\n",
    "## 2.2. Classification of the UK dataset by the \"Vegetarian Friendly\" feature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "id": "714293b7",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Logistic Regression\n",
    "logreg_grid = {\n",
    "    \"C\": np.logspace(-2, 1, 4), # np.logspace(-3, 3, 7) - from 0.001 to 1000 by 7 steps\n",
    "    \"penalty\": [\"l1\", \"l2\"], # , \"elasticnet\", None\n",
    "    \"multi_class\": [\"multinomial\"], # \"auto\", \"ovr\", \n",
    "    \"max_iter\": [1000],\n",
    "    \"solver\": ['lbfgs', 'newton-cg'] # For multiclass problems, only ‘newton-cg’, ‘sag’, ‘saga’ and ‘lbfgs’ handle multinomial loss;\n",
    "}\n",
    "\n",
    "# SVM\n",
    "svm_grid = {\n",
    "    \"C\": np.logspace(-2, 1, 4), # from 0.01 to 1 by 4 steps\n",
    "    \"penalty\": [\"l1\", \"l2\"],\n",
    "    \"loss\": [\"hinge\", \"squared_hinge\"],\n",
    "    \"max_iter\": [1000]\n",
    "}\n",
    "\n",
    "# Naive Bayes\n",
    "# no grid\n",
    "\n",
    "# k-Nearest Neighbors\n",
    "knn_grid = {\n",
    "    \"n_neighbors\": [6, 8],\n",
    "    \"leaf_size\": [12],\n",
    "    \"algorithm\": ['auto']\n",
    "    ,\"n_jobs\": [-1]\n",
    "}\n",
    "\n",
    "# Decision Tree\n",
    "tree_grid = {\n",
    "    \"min_samples_split\": range(4, 10, 2),\n",
    "    \"max_depth\": range(4, 10, 2),\n",
    "#     \"criterion\": ['squared_error', 'friedman_mse', 'absolute_error', 'poisson'],\n",
    "    \"splitter\": ['best', 'random']\n",
    "}\n",
    "\n",
    "# Random Forest\n",
    "forest_grid = {\n",
    "    \"n_estimators\": [200, 300], # range(100, 400, 100)\n",
    "    \"max_depth\": [8, 10], # range(4, 10, 2)\n",
    "    \"min_samples_split\": [20, 30], # range(20, 40, 10)\n",
    "    \"n_jobs\": [-1],\n",
    "    \"random_state\": [42]\n",
    "}\n",
    "\n",
    "# Gradient Boosting\n",
    "gb_grid = {\n",
    "    \"n_estimators\": [300], # range(100, 400, 100)\n",
    "#     \"max_depth\": [6], # range(4, 10, 2),\n",
    "    \"learning_rate\": [0.2], # [0.1, , 0.3],\n",
    "#     \"n_jobs\": [-1],\n",
    "#     \"objective\": ['reg:squarederror'],\n",
    "    \"random_state\": [42]\n",
    "}\n",
    "\n",
    "# XGB\n",
    "xgb_grid = {\n",
    "    \"n_estimators\": [300], # range(100, 400, 100)\n",
    "    \"max_depth\": [6], # range(4, 10, 2),\n",
    "    \"learning_rate\": [0.2], # [0.1, , 0.3],\n",
    "    \"n_jobs\": [-1],\n",
    "    \"objective\": ['reg:squarederror'],\n",
    "    \"random_state\": [42]\n",
    "}\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 211,
   "id": "84e6fb3c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "region\n",
      "{'Aberdeenshire': 0, 'Angus': 1, 'Argyll and Bute': 2, 'Ayrshire': 3, 'Bedfordshire': 4, 'Belfast': 5, 'Berkshire': 6, 'Borders': 7, 'Brecon Beacons National Park': 8, 'Bristol': 9, 'Buckinghamshire': 10, 'Cambridgeshire': 11, 'Carmarthenshire': 12, 'Ceredigion': 13, 'Cheshire': 14, 'Clackmannanshire': 15, 'Cornwall': 16, 'Cotswolds': 17, 'Cumbria': 18, 'Derbyshire': 19, 'Devon': 20, 'Dorset': 21, 'Dumfries and Galloway': 22, 'Dundee': 23, 'Durham': 24, 'East Anglia': 25, 'East Dunbartonshire': 26, 'East Lothian': 27, 'East Renfrewshire': 28, 'East Sussex': 29, 'Edinburgh': 30, 'Essex': 31, 'Exmoor National Park': 32, 'Falkirk District': 33, 'Fife': 34, 'Glasgow': 35, 'Gloucestershire': 36, 'Greater London': 37, 'Greater Manchester': 38, 'Hampshire': 39, 'Herefordshire': 40, 'Hertfordshire': 41, 'Inverclyde': 42, 'Isle of Arran': 43, 'Isle of Wight': 44, 'Isles of Scilly': 45, 'Kent': 46, 'Lancashire': 47, 'Leicestershire': 48, 'Lincolnshire': 49, 'London': 50, 'Merseyside': 51, 'Midlothian': 52, 'Moray': 53, 'North Lanarkshire': 54, 'North Wales': 55, 'Northamptonshire': 56, 'Northumberland': 57, 'Nottinghamshire': 58, 'Orkney Islands': 59, 'Oxfordshire': 60, 'Peak District National Park': 61, 'Pembrokeshire': 62, 'Perth and Kinross': 63, 'Powys': 64, 'Renfrewshire': 65, 'Rutland': 66, 'Scottish Highlands': 67, 'Shetland Islands': 68, 'Shropshire': 69, 'Somerset': 70, 'South Lanarkshire': 71, 'Southern Wales': 72, 'Staffordshire': 73, 'Stirling': 74, 'Stockton-on-Tees': 75, 'Surrey': 76, 'The Hebrides': 77, 'Tyne and Wear': 78, 'Warwickshire': 79, 'West Lothian': 80, 'West Midlands': 81, 'West Sussex': 82, 'Wiltshire': 83, 'Worcestershire': 84, 'Yorkshire': 85, nan: 86}\n",
      "\n",
      "price_level\n",
      "{'€': 0, '€€-€€€': 1, '€€€€': 2}\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py:378: FitFailedWarning: \n",
      "80 fits failed out of a total of 160.\n",
      "The score on these train-test partitions for these parameters will be set to nan.\n",
      "If these failures are not expected, you can try to debug them by setting error_score='raise'.\n",
      "\n",
      "Below are more details about the failures:\n",
      "--------------------------------------------------------------------------------\n",
      "40 fits failed with the following error:\n",
      "Traceback (most recent call last):\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py\", line 686, in _fit_and_score\n",
      "    estimator.fit(X_train, y_train, **fit_params)\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/linear_model/_logistic.py\", line 1162, in fit\n",
      "    solver = _check_solver(self.solver, self.penalty, self.dual)\n",
      "             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/linear_model/_logistic.py\", line 54, in _check_solver\n",
      "    raise ValueError(\n",
      "ValueError: Solver lbfgs supports only 'l2' or 'none' penalties, got l1 penalty.\n",
      "\n",
      "--------------------------------------------------------------------------------\n",
      "40 fits failed with the following error:\n",
      "Traceback (most recent call last):\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py\", line 686, in _fit_and_score\n",
      "    estimator.fit(X_train, y_train, **fit_params)\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/linear_model/_logistic.py\", line 1162, in fit\n",
      "    solver = _check_solver(self.solver, self.penalty, self.dual)\n",
      "             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/linear_model/_logistic.py\", line 54, in _check_solver\n",
      "    raise ValueError(\n",
      "ValueError: Solver newton-cg supports only 'l2' or 'none' penalties, got l1 penalty.\n",
      "\n",
      "  warnings.warn(some_fits_failed_message, FitFailedWarning)\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_search.py:953: UserWarning: One or more of the test scores are non-finite: [       nan        nan 0.94062066 0.94058047        nan        nan\n",
      " 0.94315342 0.94323382        nan        nan 0.94347503 0.94343482\n",
      "        nan        nan 0.94355544 0.94351523]\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: LogisticRegression\n",
      "Tuned hyperparameters:\n",
      "\tC: 10.0\n",
      "\tmax_iter: 1000\n",
      "\tmulti_class: multinomial\n",
      "\tpenalty: l2\n",
      "\tsolver: lbfgs\n",
      "\n",
      "Confusion matrix:\n",
      " [[6177  148]\n",
      " [ 451 3885]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.93      0.98      0.95      6325\n",
      "           1       0.96      0.90      0.93      4336\n",
      "\n",
      "    accuracy                           0.94     10661\n",
      "   macro avg       0.95      0.94      0.94     10661\n",
      "weighted avg       0.94      0.94      0.94     10661\n",
      "\n",
      "Time using clf_model_implementation(): 36.11555194854736 sec\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py:378: FitFailedWarning: \n",
      "80 fits failed out of a total of 160.\n",
      "The score on these train-test partitions for these parameters will be set to nan.\n",
      "If these failures are not expected, you can try to debug them by setting error_score='raise'.\n",
      "\n",
      "Below are more details about the failures:\n",
      "--------------------------------------------------------------------------------\n",
      "40 fits failed with the following error:\n",
      "Traceback (most recent call last):\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py\", line 686, in _fit_and_score\n",
      "    estimator.fit(X_train, y_train, **fit_params)\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_classes.py\", line 274, in fit\n",
      "    self.coef_, self.intercept_, n_iter_ = _fit_liblinear(\n",
      "                                           ^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py\", line 1223, in _fit_liblinear\n",
      "    solver_type = _get_liblinear_solver_type(multi_class, penalty, loss, dual)\n",
      "                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py\", line 1062, in _get_liblinear_solver_type\n",
      "    raise ValueError(\n",
      "ValueError: Unsupported set of arguments: The combination of penalty='l1' and loss='hinge' is not supported, Parameters: penalty='l1', loss='hinge', dual=True\n",
      "\n",
      "--------------------------------------------------------------------------------\n",
      "40 fits failed with the following error:\n",
      "Traceback (most recent call last):\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py\", line 686, in _fit_and_score\n",
      "    estimator.fit(X_train, y_train, **fit_params)\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_classes.py\", line 274, in fit\n",
      "    self.coef_, self.intercept_, n_iter_ = _fit_liblinear(\n",
      "                                           ^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py\", line 1223, in _fit_liblinear\n",
      "    solver_type = _get_liblinear_solver_type(multi_class, penalty, loss, dual)\n",
      "                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py\", line 1062, in _get_liblinear_solver_type\n",
      "    raise ValueError(\n",
      "ValueError: Unsupported set of arguments: The combination of penalty='l1' and loss='squared_hinge' are not supported when dual=True, Parameters: penalty='l1', loss='squared_hinge', dual=True\n",
      "\n",
      "  warnings.warn(some_fits_failed_message, FitFailedWarning)\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_search.py:953: UserWarning: One or more of the test scores are non-finite: [       nan 0.94251018        nan 0.94303276        nan 0.94251018\n",
      "        nan 0.94319359        nan 0.94251018        nan 0.94041915\n",
      "        nan 0.94251022        nan 0.8715022 ]\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: LinearSVC\n",
      "Tuned hyperparameters:\n",
      "\tC: 0.1\n",
      "\tloss: squared_hinge\n",
      "\tmax_iter: 1000\n",
      "\tpenalty: l2\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix:\n",
      " [[6212  113]\n",
      " [ 488 3848]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.93      0.98      0.95      6325\n",
      "           1       0.97      0.89      0.93      4336\n",
      "\n",
      "    accuracy                           0.94     10661\n",
      "   macro avg       0.95      0.93      0.94     10661\n",
      "weighted avg       0.95      0.94      0.94     10661\n",
      "\n",
      "Time using clf_model_implementation(): 60.58976197242737 sec\n",
      "Model: GaussianNB\n",
      "\n",
      "Confusion matrix:\n",
      " [[5747  578]\n",
      " [ 266 4070]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.96      0.91      0.93      6325\n",
      "           1       0.88      0.94      0.91      4336\n",
      "\n",
      "    accuracy                           0.92     10661\n",
      "   macro avg       0.92      0.92      0.92     10661\n",
      "weighted avg       0.92      0.92      0.92     10661\n",
      "\n",
      "Time using clf_model_implementation(): 0.12271618843078613 sec\n",
      "Model: KNeighborsClassifier\n",
      "Tuned hyperparameters:\n",
      "\talgorithm: auto\n",
      "\tleaf_size: 12\n",
      "\tn_jobs: -1\n",
      "\tn_neighbors: 8\n",
      "\n",
      "Confusion matrix:\n",
      " [[6097  228]\n",
      " [ 625 3711]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.91      0.96      0.93      6325\n",
      "           1       0.94      0.86      0.90      4336\n",
      "\n",
      "    accuracy                           0.92     10661\n",
      "   macro avg       0.92      0.91      0.92     10661\n",
      "weighted avg       0.92      0.92      0.92     10661\n",
      "\n",
      "Time using clf_model_implementation(): 25.549763917922974 sec\n",
      "Model: DecisionTreeClassifier\n",
      "Tuned hyperparameters:\n",
      "\tmax_depth: 4\n",
      "\tmin_samples_split: 4\n",
      "\tsplitter: random\n",
      "\n",
      "Confusion matrix:\n",
      " [[6229   96]\n",
      " [ 513 3823]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.92      0.98      0.95      6325\n",
      "           1       0.98      0.88      0.93      4336\n",
      "\n",
      "    accuracy                           0.94     10661\n",
      "   macro avg       0.95      0.93      0.94     10661\n",
      "weighted avg       0.94      0.94      0.94     10661\n",
      "\n",
      "Time using clf_model_implementation(): 3.6664459705352783 sec\n",
      "Model: RandomForestClassifier\n",
      "Tuned hyperparameters:\n",
      "\tmax_depth: 10\n",
      "\tmin_samples_split: 30\n",
      "\tn_estimators: 200\n",
      "\tn_jobs: -1\n",
      "\trandom_state: 42\n",
      "\n",
      "Confusion matrix:\n",
      " [[6181  144]\n",
      " [ 419 3917]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.94      0.98      0.96      6325\n",
      "           1       0.96      0.90      0.93      4336\n",
      "\n",
      "    accuracy                           0.95     10661\n",
      "   macro avg       0.95      0.94      0.94     10661\n",
      "weighted avg       0.95      0.95      0.95     10661\n",
      "\n",
      "Time using clf_model_implementation(): 41.67178511619568 sec\n",
      "Model: GradientBoostingClassifier\n",
      "Tuned hyperparameters:\n",
      "\tlearning_rate: 0.2\n",
      "\tn_estimators: 300\n",
      "\trandom_state: 42\n",
      "\n",
      "Confusion matrix:\n",
      " [[6125  200]\n",
      " [ 393 3943]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.94      0.97      0.95      6325\n",
      "           1       0.95      0.91      0.93      4336\n",
      "\n",
      "    accuracy                           0.94     10661\n",
      "   macro avg       0.95      0.94      0.94     10661\n",
      "weighted avg       0.94      0.94      0.94     10661\n",
      "\n",
      "Time using clf_model_implementation(): 46.451091051101685 sec\n",
      "Model: XGBClassifier\n",
      "Tuned hyperparameters:\n",
      "\tlearning_rate: 0.2\n",
      "\tmax_depth: 6\n",
      "\tn_estimators: 300\n",
      "\tn_jobs: -1\n",
      "\tobjective: reg:squarederror\n",
      "\trandom_state: 42\n",
      "\n",
      "Confusion matrix:\n",
      " [[6112  213]\n",
      " [ 405 3931]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.94      0.97      0.95      6325\n",
      "           1       0.95      0.91      0.93      4336\n",
      "\n",
      "    accuracy                           0.94     10661\n",
      "   macro avg       0.94      0.94      0.94     10661\n",
      "weighted avg       0.94      0.94      0.94     10661\n",
      "\n",
      "Time using clf_model_implementation(): 30.252930164337158 sec\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'logreg': {'model': LogisticRegression(C=10.0, max_iter=1000, multi_class='multinomial'),\n",
       "  'accuracy_score': 0.943813901134978},\n",
       " 'svm': {'model': LinearSVC(C=0.1), 'accuracy_score': 0.9436263014726574},\n",
       " 'nb': {'model': GaussianNB(), 'accuracy_score': 0.9208329425007035},\n",
       " 'knn': {'model': KNeighborsClassifier(leaf_size=12, n_jobs=-1, n_neighbors=8),\n",
       "  'accuracy_score': 0.9199887440202608},\n",
       " 'tree': {'model': DecisionTreeClassifier(max_depth=4, min_samples_split=4, splitter='random'),\n",
       "  'accuracy_score': 0.942875902823375},\n",
       " 'forest': {'model': RandomForestClassifier(max_depth=10, min_samples_split=30, n_estimators=200,\n",
       "                         n_jobs=-1, random_state=42),\n",
       "  'accuracy_score': 0.9471906950567489},\n",
       " 'gb': {'model': GradientBoostingClassifier(learning_rate=0.2, n_estimators=300, random_state=42),\n",
       "  'accuracy_score': 0.9443767001219397},\n",
       " 'xgb': {'model': XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
       "                colsample_bylevel=None, colsample_bynode=None,\n",
       "                colsample_bytree=None, early_stopping_rounds=None,\n",
       "                enable_categorical=False, eval_metric=None, feature_types=None,\n",
       "                gamma=None, gpu_id=None, grow_policy=None, importance_type=None,\n",
       "                interaction_constraints=None, learning_rate=0.2, max_bin=None,\n",
       "                max_cat_threshold=None, max_cat_to_onehot=None,\n",
       "                max_delta_step=None, max_depth=6, max_leaves=None,\n",
       "                min_child_weight=None, missing=nan, monotone_constraints=None,\n",
       "                n_estimators=300, n_jobs=-1, num_parallel_tree=None,\n",
       "                objective='reg:squarederror', predictor=None, ...),\n",
       "  'accuracy_score': 0.9420317043429322}}"
      ]
     },
     "execution_count": 211,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# PIVOT DICTIONARY OF CLASSIFICATION MODELS\n",
    "clf_models_dict = {\n",
    "#     simple models\n",
    "    'logreg': {'status': True, 'model': LogisticRegression, 'grid': logreg_grid},\n",
    "    'svm': {'status': True, 'model': LinearSVC, 'grid': svm_grid},\n",
    "    'nb': {'status': True, 'model': GaussianNB, 'grid': {}},\n",
    "    'knn': {'status': True, 'model': KNeighborsClassifier, 'grid': knn_grid},\n",
    "    'tree': {'status': True, 'model': DecisionTreeClassifier, 'grid': tree_grid},\n",
    "#     ensemble models (not based on other models)\n",
    "    'forest': {'status': True, 'model': RandomForestClassifier, 'grid': forest_grid},\n",
    "    'gb': {'status': True, 'model': GradientBoostingClassifier, 'grid': gb_grid},\n",
    "    'xgb': {'status': True, 'model': XGBClassifier, 'grid': xgb_grid}\n",
    "}\n",
    "\n",
    "uk_clf_dict = run_classification_models(uk_df2, 'special_diets_Vegetarian Friendly', [], clf_models_dict)\n",
    "# drop_columns = ['special_diets_Gluten Free Options', 'special_diets_Vegan Options']\n",
    "uk_clf_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "id": "e639a8cf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.9471906950567489,\n",
       " RandomForestClassifier(max_depth=10, min_samples_split=30, n_estimators=200,\n",
       "                        n_jobs=-1, random_state=42))"
      ]
     },
     "execution_count": 212,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "find_the_best_model(uk_clf_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "id": "a5951451",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "score:\t  feature:\n",
      "0.331\t- features_Highchairs Available\n",
      "0.086\t- cuisines_Mediterranean\n",
      "0.077\t- features_Takeout\n",
      "0.057\t- meals_Lunch\n",
      "0.045\t- meals_Breakfast\n",
      "0.039\t- cuisines_Seafood\n",
      "0.022\t- meals_Dinner\n",
      "0.010\t- price_level\n",
      "0.007\t- features_Outdoor Seating\n",
      "0.005\t- features_Reservations\n",
      "0.004\t- region\n",
      "0.004\t- features_Seating\n",
      "0.004\t- features_Wheelchair Accessible\n",
      "0.003\t- features_Table Service\n",
      "0.002\t- cuisines_Pizza\n",
      "0.001\t- cuisines_Italian\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiwAAAKACAYAAAC/qSN8AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8o6BhiAAAACXBIWXMAAA9hAAAPYQGoP6dpAAC0KUlEQVR4nOzdd3xN9/8H8NdNZO9IJAQZpHYEqVFqhli1WrtGrNYmRqkRq6iWhtpUhdYqLVoaNPYesUrtRCgJoqgEIXn//vDL/eZKjMg59x76ej4e5yE59+TzPidy733fz/l83h+diAiIiIiINMzM1CdARERE9DJMWIiIiEjzmLAQERGR5jFhISIiIs1jwkJERESax4SFiIiINI8JCxEREWkeExYiIiLSPCYsREREpHlMWIiIiEjzmLAQGcHixYuh0+my3YYNG6ZKzL1792LMmDG4c+eOKu3nRsbv4/Dhw6Y+ldc2e/ZsLF682NSnQfSfkcfUJ0D0XzJu3Dj4+voa7CtdurQqsfbu3YuxY8eic+fOcHZ2ViXGf9ns2bPh5uaGzp07m/pUiP4TmLAQGVGDBg0QFBRk6tPIleTkZNjZ2Zn6NEwmJSUFtra2pj4Nov8c3hIi0pDff/8d77//Puzs7ODg4IBGjRrh1KlTBsecOHECnTt3hp+fH6ytreHp6YkuXbogKSlJf8yYMWMwZMgQAICvr6/+9lNcXBzi4uKg0+myvZ2h0+kwZswYg3Z0Oh1Onz6Ndu3awcXFBdWqVdM//sMPP6BChQqwsbGBq6sr2rRpgytXrrzWtXfu3Bn29vaIj49H48aNYW9vDy8vL8yaNQsAcPLkSdSuXRt2dnbw9vbGsmXLDH4+4zbTzp078cknnyBv3rxwdHREx44d8c8//2SJN3v2bJQqVQpWVlYoUKAAevfuneX2Wc2aNVG6dGkcOXIE1atXh62tLT7//HP4+Pjg1KlT2LFjh/53W7NmTQDA7du3MXjwYJQpUwb29vZwdHREgwYNcPz4cYO2t2/fDp1Oh1WrVuGLL75AwYIFYW1tjTp16uDChQtZzvfAgQNo2LAhXFxcYGdnh4CAAEyfPt3gmDNnzuCjjz6Cq6srrK2tERQUhPXr1+f0v4JIk9jDQmREd+/exa1btwz2ubm5AQCWLl2KTp06ISQkBF9++SVSUlIwZ84cVKtWDUePHoWPjw8AYMuWLbh06RJCQ0Ph6emJU6dOYf78+Th16hT2798PnU6HFi1a4Ny5c1i+fDm++eYbfQx3d3fcvHkzx+fdsmVL+Pv7Y+LEiRARAMAXX3yBUaNGoVWrVujWrRtu3ryJb7/9FtWrV8fRo0df6zZUWloaGjRogOrVq2PKlCn48ccf0adPH9jZ2WHEiBFo3749WrRogblz56Jjx46oUqVKlltsffr0gbOzM8aMGYOzZ89izpw5uHz5sj5BAJ4mYmPHjkVwcDB69uypP+7QoUPYs2cPLCws9O0lJSWhQYMGaNOmDT7++GN4eHigZs2a6Nu3L+zt7TFixAgAgIeHBwDg0qVLWLt2LVq2bAlfX18kJiZi3rx5qFGjBk6fPo0CBQoYnO/kyZNhZmaGwYMH4+7du5gyZQrat2+PAwcO6I/ZsmULGjdujPz586N///7w9PTEX3/9hd9++w39+/cHAJw6dQpVq1aFl5cXhg0bBjs7O6xatQrNmjXDmjVr0Lx58xz/fxBpihCR6r7//nsBkO0mIvLvv/+Ks7OzdO/e3eDnEhISxMnJyWB/SkpKlvaXL18uAGTnzp36fV999ZUAkNjYWINjY2NjBYB8//33WdoBIOHh4frvw8PDBYC0bdvW4Li4uDgxNzeXL774wmD/yZMnJU+ePFn2P+/3cejQIf2+Tp06CQCZOHGift8///wjNjY2otPpZMWKFfr9Z86cyXKuGW1WqFBBUlNT9funTJkiAGTdunUiInLjxg2xtLSUevXqSVpamv64mTNnCgBZtGiRfl+NGjUEgMydOzfLNZQqVUpq1KiRZf/Dhw8N2hV5+ju3srKScePG6fdt27ZNAEiJEiXk0aNH+v3Tp08XAHLy5EkREXny5In4+vqKt7e3/PPPPwbtpqen67+uU6eOlClTRh4+fGjw+HvvvSf+/v5ZzpPoTcNbQkRGNGvWLGzZssVgA55+gr5z5w7atm2LW7du6Tdzc3NUqlQJ27Zt07dhY2Oj//rhw4e4desWKleuDACIiYlR5bw//fRTg+9//vlnpKeno1WrVgbn6+npCX9/f4Pzzalu3brpv3Z2dkaxYsVgZ2eHVq1a6fcXK1YMzs7OuHTpUpaf79Gjh0EPSc+ePZEnTx5s3LgRAPDHH38gNTUVAwYMgJnZ/14Cu3fvDkdHR2zYsMGgPSsrK4SGhr7y+VtZWenbTUtLQ1JSEuzt7VGsWLFs/39CQ0NhaWmp//79998HAP21HT16FLGxsRgwYECWXquMHqPbt29j69ataNWqFf7991/9/0dSUhJCQkJw/vx5/P333698DURaxFtCREZUsWLFbAfdnj9/HgBQu3btbH/O0dFR//Xt27cxduxYrFixAjdu3DA47u7duwqe7f88e9vl/PnzEBH4+/tne3zmhCEnrK2t4e7ubrDPyckJBQsW1L85Z96f3diUZ8/J3t4e+fPnR1xcHADg8uXLAJ4mPZlZWlrCz89P/3gGLy8vg4TiZdLT0zF9+nTMnj0bsbGxSEtL0z+WN2/eLMcXLlzY4HsXFxcA0F/bxYsXAbx4NtmFCxcgIhg1ahRGjRqV7TE3btyAl5fXK18HkdYwYSHSgPT0dABPx7F4enpmeTxPnv89VVu1aoW9e/diyJAhCAwMhL29PdLT01G/fn19Oy/y7Bt/hsxvrM/K3KuTcb46nQ6///47zM3Nsxxvb2//0vPITnZtvWi//P94GjU9e+0vM3HiRIwaNQpdunTB+PHj4erqCjMzMwwYMCDb/x8lri2j3cGDByMkJCTbY4oWLfrK7RFpERMWIg0oUqQIACBfvnwIDg5+7nH//PMPoqOjMXbsWIwePVq/P6OHJrPnJSYZn+CfnRHzbM/Cy85XRODr64t33nnnlX/OGM6fP49atWrpv79//z6uX7+Ohg0bAgC8vb0BAGfPnoWfn5/+uNTUVMTGxr7w95/Z836/q1evRq1atfDdd98Z7L9z545+8HNOZPxt/Pnnn889t4zrsLCweOXzJ3rTcAwLkQaEhITA0dEREydOxOPHj7M8njGzJ+PT+LOfviMiIrL8TEatlGcTE0dHR7i5uWHnzp0G+2fPnv3K59uiRQuYm5tj7NixWc5FRAymWBvb/PnzDX6Hc+bMwZMnT9CgQQMAQHBwMCwtLTFjxgyDc//uu+9w9+5dNGrU6JXi2NnZZVtF2NzcPMvv5KeffnrtMSTly5eHr68vIiIissTLiJMvXz7UrFkT8+bNw/Xr17O08Tozw4i0hj0sRBrg6OiIOXPmoEOHDihfvjzatGkDd3d3xMfHY8OGDahatSpmzpwJR0dH/ZTfx48fw8vLC5s3b0ZsbGyWNitUqAAAGDFiBNq0aQMLCwt88MEHsLOzQ7du3TB58mR069YNQUFB2LlzJ86dO/fK51ukSBFMmDABw4cPR1xcHJo1awYHBwfExsbil19+QY8ePTB48GDFfj85kZqaijp16qBVq1Y4e/YsZs+ejWrVqqFJkyYAnk7tHj58OMaOHYv69eujSZMm+uPeffddfPzxx68Up0KFCpgzZw4mTJiAokWLIl++fKhduzYaN26McePGITQ0FO+99x5OnjyJH3/80aA3JyfMzMwwZ84cfPDBBwgMDERoaCjy58+PM2fO4NSpU9i0aROApwO6q1WrhjJlyqB79+7w8/NDYmIi9u3bh6tXr2apA0P0xjHR7CSi/5TspvFmZ9u2bRISEiJOTk5ibW0tRYoUkc6dO8vhw4f1x1y9elWaN28uzs7O4uTkJC1btpRr165lmeYrIjJ+/Hjx8vISMzMzgynOKSkp0rVrV3FychIHBwdp1aqV3Lhx47nTmm/evJnt+a5Zs0aqVasmdnZ2YmdnJ8WLF5fevXvL2bNnc/z76NSpk9jZ2WU5tkaNGlKqVKks+729vaVRo0ZZ2tyxY4f06NFDXFxcxN7eXtq3by9JSUlZfn7mzJlSvHhxsbCwEA8PD+nZs2eWacPPiy3ydMp5o0aNxMHBQQDopzg/fPhQBg0aJPnz5xcbGxupWrWq7Nu3T2rUqGEwDTpjWvNPP/1k0O7zpp3v3r1b6tatKw4ODmJnZycBAQHy7bffGhxz8eJF6dixo3h6eoqFhYV4eXlJ48aNZfXq1dleA9GbRCdihFFrREQqW7x4MUJDQ3Ho0KE3fvkDIsqKY1iIiIhI85iwEBERkeYxYSEiIiLN4xgWIiIi0jz2sBAREZHmvRV1WNLT03Ht2jU4ODg8t/okERERaYuI4N9//0WBAgUMFiPNzluRsFy7dg2FChUy9WkQERHRa7hy5QoKFiz4wmPeioTFwcEBwNMLzryqLREREWnXvXv3UKhQIf37+Iu8FQlLxm0gR0dHJixERERvmFcZzsFBt0RERKR5TFiIiIhI85iwEBERkeYxYSEiIiLNY8JCREREmseEhYiIiDSPCQsRERFpHhMWIiIi0jwmLERERKR5TFiIiIhI85iwEBERkeYxYSEiIiLNY8JCREREmseEhYiIiDSPCQsRERFpXh5Tn8CbwGfYBtXajpvcSLW2iYiI3hbsYSEiIiLNY8JCREREmseEhYiIiDSPCQsRERFpHhMWIiIi0jwmLERERKR5TFiIiIhI85iwEBERkeYxYSEiIiLNY8JCREREmseEhYiIiDSPCQsRERFpHhMWIiIi0jwmLERERKR5TFiIiIhI85iwEBERkeYxYSEiIiLNY8JCREREmseEhYiIiDSPCQsRERFpHhMWIiIi0jwmLERERKR5TFiIiIhI85iwEBERkeYxYSEiIiLNY8JCREREmseEhYiIiDSPCQsRERFp3mslLLNmzYKPjw+sra1RqVIlHDx48LnH/vzzzwgKCoKzszPs7OwQGBiIpUuXGhwjIhg9ejTy588PGxsbBAcH4/z5869zakRERPQWynHCsnLlSoSFhSE8PBwxMTEoW7YsQkJCcOPGjWyPd3V1xYgRI7Bv3z6cOHECoaGhCA0NxaZNm/THTJkyBTNmzMDcuXNx4MAB2NnZISQkBA8fPnz9KyMiIqK3hk5EJCc/UKlSJbz77ruYOXMmACA9PR2FChVC3759MWzYsFdqo3z58mjUqBHGjx8PEUGBAgUwaNAgDB48GABw9+5deHh4YPHixWjTps1L27t37x6cnJxw9+5dODo65uRyXonPsA2Kt5khbnIj1domIiLSspy8f+eohyU1NRVHjhxBcHDw/xowM0NwcDD27dv30p8XEURHR+Ps2bOoXr06ACA2NhYJCQkGbTo5OaFSpUrPbfPRo0e4d++ewUZERERvrxwlLLdu3UJaWho8PDwM9nt4eCAhIeG5P3f37l3Y29vD0tISjRo1wrfffou6desCgP7nctLmpEmT4OTkpN8KFSqUk8sgIiKiN4xRZgk5ODjg2LFjOHToEL744guEhYVh+/btr93e8OHDcffuXf125coV5U6WiIiINCdPTg52c3ODubk5EhMTDfYnJibC09PzuT9nZmaGokWLAgACAwPx119/YdKkSahZs6b+5xITE5E/f36DNgMDA7Ntz8rKClZWVjk5dSIiInqD5aiHxdLSEhUqVEB0dLR+X3p6OqKjo1GlSpVXbic9PR2PHj0CAPj6+sLT09OgzXv37uHAgQM5apOIiIjeXjnqYQGAsLAwdOrUCUFBQahYsSIiIiKQnJyM0NBQAEDHjh3h5eWFSZMmAXg63iQoKAhFihTBo0ePsHHjRixduhRz5swBAOh0OgwYMAATJkyAv78/fH19MWrUKBQoUADNmjVT7kqJiIjojZXjhKV169a4efMmRo8ejYSEBAQGBiIqKko/aDY+Ph5mZv/ruElOTkavXr1w9epV2NjYoHjx4vjhhx/QunVr/TFDhw5FcnIyevTogTt37qBatWqIioqCtbW1ApdIREREb7oc12HRItZhISIievOoVoeFiIiIyBSYsBAREZHmMWEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPOYsBAREZHmMWEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPOYsBAREZHmMWEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPOYsBAREZHmMWEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPOYsBAREZHmMWEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPOYsBAREZHmvVbCMmvWLPj4+MDa2hqVKlXCwYMHn3vsggUL8P7778PFxQUuLi4IDg7Ocnznzp2h0+kMtvr167/OqREREdFbKMcJy8qVKxEWFobw8HDExMSgbNmyCAkJwY0bN7I9fvv27Wjbti22bduGffv2oVChQqhXrx7+/vtvg+Pq16+P69ev67fly5e/3hURERHRWyfHCcu0adPQvXt3hIaGomTJkpg7dy5sbW2xaNGibI//8ccf0atXLwQGBqJ48eJYuHAh0tPTER0dbXCclZUVPD099ZuLi8vrXRERERG9dXKUsKSmpuLIkSMIDg7+XwNmZggODsa+ffteqY2UlBQ8fvwYrq6uBvu3b9+OfPnyoVixYujZsyeSkpKe28ajR49w7949g42IiIjeXjlKWG7duoW0tDR4eHgY7Pfw8EBCQsIrtfHZZ5+hQIECBklP/fr1sWTJEkRHR+PLL7/Ejh070KBBA6SlpWXbxqRJk+Dk5KTfChUqlJPLICIiojdMHmMGmzx5MlasWIHt27fD2tpav79Nmzb6r8uUKYOAgAAUKVIE27dvR506dbK0M3z4cISFhem/v3fvHpMWIiKit1iOeljc3Nxgbm6OxMREg/2JiYnw9PR84c9+/fXXmDx5MjZv3oyAgIAXHuvn5wc3NzdcuHAh28etrKzg6OhosBEREdHbK0cJi6WlJSpUqGAwYDZjAG2VKlWe+3NTpkzB+PHjERUVhaCgoJfGuXr1KpKSkpA/f/6cnB4RERG9pXI8SygsLAwLFixAZGQk/vrrL/Ts2RPJyckIDQ0FAHTs2BHDhw/XH//ll19i1KhRWLRoEXx8fJCQkICEhATcv38fAHD//n0MGTIE+/fvR1xcHKKjo9G0aVMULVoUISEhCl0mERERvclyPIaldevWuHnzJkaPHo2EhAQEBgYiKipKPxA3Pj4eZmb/y4PmzJmD1NRUfPTRRwbthIeHY8yYMTA3N8eJEycQGRmJO3fuoECBAqhXrx7Gjx8PKyurXF4eERERvQ10IiKmPoncunfvHpycnHD37l1VxrP4DNugeJsZ4iY3Uq1tIiIiLcvJ+zfXEiIiIiLNY8JCREREmseEhYiIiDSPCQsRERFpHhMWIiIi0jwmLERERKR5TFiIiIhI85iwEBERkeYxYSEiIiLNY8JCREREmseEhYiIiDSPCQsRERFpHhMWIiIi0jwmLERERKR5TFiIiIhI85iwEBERkeYxYSEiIiLNY8JCREREmseEhYiIiDSPCQsRERFpHhMWIiIi0jwmLERERKR5TFiIiIhI85iwEBERkeYxYSEiIiLNY8JCREREmseEhYiIiDSPCQsRERFpHhMWIiIi0jwmLERERKR5TFiIiIhI85iwEBERkeYxYSEiIiLNY8JCREREmseEhYiIiDSPCQsRERFpHhMWIiIi0jwmLERERKR5TFiIiIhI85iwEBERkeYxYSEiIiLNY8JCREREmvdaCcusWbPg4+MDa2trVKpUCQcPHnzusQsWLMD7778PFxcXuLi4IDg4OMvxIoLRo0cjf/78sLGxQXBwMM6fP/86p0ZERERvoRwnLCtXrkRYWBjCw8MRExODsmXLIiQkBDdu3Mj2+O3bt6Nt27bYtm0b9u3bh0KFCqFevXr4+++/9cdMmTIFM2bMwNy5c3HgwAHY2dkhJCQEDx8+fP0rIyIioreGTkQkJz9QqVIlvPvuu5g5cyYAID09HYUKFULfvn0xbNiwl/58WloaXFxcMHPmTHTs2BEiggIFCmDQoEEYPHgwAODu3bvw8PDA4sWL0aZNm5e2ee/ePTg5OeHu3btwdHTMyeW8Ep9hGxRvM0Pc5EaqtU1ERKRlOXn/zlEPS2pqKo4cOYLg4OD/NWBmhuDgYOzbt++V2khJScHjx4/h6uoKAIiNjUVCQoJBm05OTqhUqdJz23z06BHu3btnsBEREdHbK0cJy61bt5CWlgYPDw+D/R4eHkhISHilNj777DMUKFBAn6Bk/FxO2pw0aRKcnJz0W6FChXJyGURERPSGMeosocmTJ2PFihX45ZdfYG1t/drtDB8+HHfv3tVvV65cUfAsiYiISGvy5ORgNzc3mJubIzEx0WB/YmIiPD09X/izX3/9NSZPnow//vgDAQEB+v0ZP5eYmIj8+fMbtBkYGJhtW1ZWVrCyssrJqRMREdEbLEc9LJaWlqhQoQKio6P1+9LT0xEdHY0qVao89+emTJmC8ePHIyoqCkFBQQaP+fr6wtPT06DNe/fu4cCBAy9sk4iIiP47ctTDAgBhYWHo1KkTgoKCULFiRURERCA5ORmhoaEAgI4dO8LLywuTJk0CAHz55ZcYPXo0li1bBh8fH/24FHt7e9jb20On02HAgAGYMGEC/P394evri1GjRqFAgQJo1qyZcldKREREb6wcJyytW7fGzZs3MXr0aCQkJCAwMBBRUVH6QbPx8fEwM/tfx82cOXOQmpqKjz76yKCd8PBwjBkzBgAwdOhQJCcno0ePHrhz5w6qVauGqKioXI1zISIiordHjuuwaBHrsBAREb15VKvDQkRERGQKTFiIiIhI85iwEBERkeYxYSEiIiLNY8JCREREmseEhYiIiDSPCQsRERFpHhMWIiIi0jwmLERERKR5TFiIiIhI85iwEBERkeYxYSEiIiLNY8JCREREmseEhYiIiDSPCQsRERFpHhMWIiIi0jwmLERERKR5TFiIiIhI85iwEBERkeYxYSEiIiLNY8JCREREmseEhYiIiDSPCQsRERFpHhMWIiIi0jwmLERERKR5TFiIiIhI85iwEBERkeYxYSEiIiLNY8JCREREmseEhYiIiDSPCQsRERFpHhMWIiIi0jwmLERERKR5TFiIiIhI85iwEBERkeYxYSEiIiLNY8JCREREmseEhYiIiDSPCQsRERFpHhMWIiIi0jwmLERERKR5TFiIiIhI85iwEBERkea9VsIya9Ys+Pj4wNraGpUqVcLBgwefe+ypU6fw4YcfwsfHBzqdDhEREVmOGTNmDHQ6ncFWvHjx1zk1IiIiegvlOGFZuXIlwsLCEB4ejpiYGJQtWxYhISG4ceNGtsenpKTAz88PkydPhqen53PbLVWqFK5fv67fdu/endNTIyIiordUjhOWadOmoXv37ggNDUXJkiUxd+5c2NraYtGiRdke/+677+Krr75CmzZtYGVl9dx28+TJA09PT/3m5ub23GMfPXqEe/fuGWxERET09spRwpKamoojR44gODj4fw2YmSE4OBj79u3L1YmcP38eBQoUgJ+fH9q3b4/4+PjnHjtp0iQ4OTnpt0KFCuUqNhEREWlbjhKWW7duIS0tDR4eHgb7PTw8kJCQ8NonUalSJSxevBhRUVGYM2cOYmNj8f777+Pff//N9vjhw4fj7t27+u3KlSuvHZuIiIi0L4+pTwAAGjRooP86ICAAlSpVgre3N1atWoWuXbtmOd7KyuqFt5eIiIjo7ZKjHhY3NzeYm5sjMTHRYH9iYuILB9TmlLOzM9555x1cuHBBsTaJiIjozZWjhMXS0hIVKlRAdHS0fl96ejqio6NRpUoVxU7q/v37uHjxIvLnz69Ym0RERPTmyvEtobCwMHTq1AlBQUGoWLEiIiIikJycjNDQUABAx44d4eXlhUmTJgF4OlD39OnT+q///vtvHDt2DPb29ihatCgAYPDgwfjggw/g7e2Na9euITw8HObm5mjbtq1S10lERERvsBwnLK1bt8bNmzcxevRoJCQkIDAwEFFRUfqBuPHx8TAz+1/HzbVr11CuXDn9919//TW+/vpr1KhRA9u3bwcAXL16FW3btkVSUhLc3d1RrVo17N+/H+7u7rm8PCIiInob6ERETH0SuXXv3j04OTnh7t27cHR0VLx9n2EbFG8zQ9zkRqq1TUREpGU5ef/mWkJERESkeUxYiIiISPOYsBAREZHmMWEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPOYsBAREZHmMWEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPOYsBAREZHmMWEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPOYsBAREZHmMWEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPOYsBAREZHmMWEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPNeK2GZNWsWfHx8YG1tjUqVKuHgwYPPPfbUqVP48MMP4ePjA51Oh4iIiFy3SURERP8tOU5YVq5cibCwMISHhyMmJgZly5ZFSEgIbty4ke3xKSkp8PPzw+TJk+Hp6alIm0RERPTfkuOEZdq0aejevTtCQ0NRsmRJzJ07F7a2tli0aFG2x7/77rv46quv0KZNG1hZWSnSJhEREf235ChhSU1NxZEjRxAcHPy/BszMEBwcjH379r3WCbxOm48ePcK9e/cMNiIiInp75ShhuXXrFtLS0uDh4WGw38PDAwkJCa91Aq/T5qRJk+Dk5KTfChUq9FqxiYiI6M3wRs4SGj58OO7evavfrly5YupTIiIiIhXlycnBbm5uMDc3R2JiosH+xMTE5w6oVaNNKyur546HISIiordPjnpYLC0tUaFCBURHR+v3paenIzo6GlWqVHmtE1CjTSIiInq75KiHBQDCwsLQqVMnBAUFoWLFioiIiEBycjJCQ0MBAB07doSXlxcmTZoE4Omg2tOnT+u//vvvv3Hs2DHY29ujaNGir9QmERER/bflOGFp3bo1bt68idGjRyMhIQGBgYGIiorSD5qNj4+Hmdn/Om6uXbuGcuXK6b//+uuv8fXXX6NGjRrYvn37K7VJRERE/206ERFTn0Ru3bt3D05OTrh79y4cHR0Vb99n2AbF28wQN7mRam0TERFpWU7ev9/IWUJERET038KEhYiIiDSPCQsRERFpHhMWIiIi0jwmLERERKR5TFiIiIhI85iwEBERkebluHAcGYdatV9Y94WIiN5E7GEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPOYsBAREZHmMWEhIiIizWOlWwKgXmVdgNV1iYgo99jDQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPOYsBAREZHmMWEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPOYsBAREZHmMWEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeXlMfQL03+UzbIMq7cZNbqRKu0REZDrsYSEiIiLNY8JCREREmvdaCcusWbPg4+MDa2trVKpUCQcPHnzh8T/99BOKFy8Oa2trlClTBhs3bjR4vHPnztDpdAZb/fr1X+fUiIiI6C2U44Rl5cqVCAsLQ3h4OGJiYlC2bFmEhITgxo0b2R6/d+9etG3bFl27dsXRo0fRrFkzNGvWDH/++afBcfXr18f169f12/Lly1/vioiIiOitk+OEZdq0aejevTtCQ0NRsmRJzJ07F7a2tli0aFG2x0+fPh3169fHkCFDUKJECYwfPx7ly5fHzJkzDY6zsrKCp6enfnNxcXm9KyIiIqK3To4SltTUVBw5cgTBwcH/a8DMDMHBwdi3b1+2P7Nv3z6D4wEgJCQky/Hbt29Hvnz5UKxYMfTs2RNJSUnPPY9Hjx7h3r17BhsRERG9vXKUsNy6dQtpaWnw8PAw2O/h4YGEhIRsfyYhIeGlx9evXx9LlixBdHQ0vvzyS+zYsQMNGjRAWlpatm1OmjQJTk5O+q1QoUI5uQwiIiJ6w2iiDkubNm30X5cpUwYBAQEoUqQItm/fjjp16mQ5fvjw4QgLC9N/f+/ePSYtREREb7Ec9bC4ubnB3NwciYmJBvsTExPh6emZ7c94enrm6HgA8PPzg5ubGy5cuJDt41ZWVnB0dDTYiIiI6O2Vo4TF0tISFSpUQHR0tH5feno6oqOjUaVKlWx/pkqVKgbHA8CWLVueezwAXL16FUlJScifP39OTo+IiIjeUjmeJRQWFoYFCxYgMjISf/31F3r27Ink5GSEhoYCADp27Ijhw4frj+/fvz+ioqIwdepUnDlzBmPGjMHhw4fRp08fAMD9+/cxZMgQ7N+/H3FxcYiOjkbTpk1RtGhRhISEKHSZRERE9CbL8RiW1q1b4+bNmxg9ejQSEhIQGBiIqKgo/cDa+Ph4mJn9Lw967733sGzZMowcORKff/45/P39sXbtWpQuXRoAYG5ujhMnTiAyMhJ37txBgQIFUK9ePYwfPx5WVlYKXSaRemsXAVy/iIhIba816LZPnz76HpJnbd++Pcu+li1bomXLltkeb2Njg02bNr3OaRAREdF/BNcSIiIiIs1jwkJERESax4SFiIiINI8JCxEREWmeJirdEr2tODOJiEgZ7GEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPOYsBAREZHmMWEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPOYsBAREZHmMWEhIiIizctj6hMgIuX4DNugWttxkxup1jYR0cuwh4WIiIg0jwkLERERaR4TFiIiItI8JixERESkeUxYiIiISPOYsBAREZHmMWEhIiIizWPCQkRERJrHhIWIiIg0jwkLERERaR4TFiIiItI8JixERESkeVz8kIhyhQsuEpExsIeFiIiINI8JCxEREWkeExYiIiLSPI5hIaI3CsfMEP03sYeFiIiINI89LEREL2HsXh32IhFl9VoJy6xZs/DVV18hISEBZcuWxbfffouKFSs+9/iffvoJo0aNQlxcHPz9/fHll1+iYcOG+sdFBOHh4ViwYAHu3LmDqlWrYs6cOfD393+d0yMiohxiUkZal+OEZeXKlQgLC8PcuXNRqVIlREREICQkBGfPnkW+fPmyHL937160bdsWkyZNQuPGjbFs2TI0a9YMMTExKF26NABgypQpmDFjBiIjI+Hr64tRo0YhJCQEp0+fhrW1de6vkoiI/vOYJL3ZcpywTJs2Dd27d0doaCgAYO7cudiwYQMWLVqEYcOGZTl++vTpqF+/PoYMGQIAGD9+PLZs2YKZM2di7ty5EBFERERg5MiRaNq0KQBgyZIl8PDwwNq1a9GmTZvcXB8REZHJqJUk/RcTpBwlLKmpqThy5AiGDx+u32dmZobg4GDs27cv25/Zt28fwsLCDPaFhIRg7dq1AIDY2FgkJCQgODhY/7iTkxMqVaqEffv2ZZuwPHr0CI8ePdJ/f/fuXQDAvXv3cnI5ryz9UYoq7QLPP2e1Yho7nili/heu8XkxeY3KxzNFTF6j8vHetpjPi1c6fJMq8QDgz7EhireZcR0i8vKDJQf+/vtvASB79+412D9kyBCpWLFitj9jYWEhy5YtM9g3a9YsyZcvn4iI7NmzRwDItWvXDI5p2bKltGrVKts2w8PDBQA3bty4cePG7S3Yrly58tIc5I2cJTR8+HCDXpv09HTcvn0befPmhU6nM9l53bt3D4UKFcKVK1fg6OjImG9oPFPE5DW+HTF5jW9HzP/CNZoq5rNEBP/++y8KFCjw0mNzlLC4ubnB3NwciYmJBvsTExPh6emZ7c94enq+8PiMfxMTE5E/f36DYwIDA7Nt08rKClZWVgb7nJ2dc3IpqnJ0dDT6f/5/ISav8e2IyWt8O2LyGhlTKU5OTq90XI4Kx1laWqJChQqIjo7W70tPT0d0dDSqVKmS7c9UqVLF4HgA2LJli/54X19feHp6Ghxz7949HDhw4LltEhER0X9Ljm8JhYWFoVOnTggKCkLFihURERGB5ORk/ayhjh07wsvLC5MmTQIA9O/fHzVq1MDUqVPRqFEjrFixAocPH8b8+fMBADqdDgMGDMCECRPg7++vn9ZcoEABNGvWTLkrJSIiojdWjhOW1q1b4+bNmxg9ejQSEhIQGBiIqKgoeHh4AADi4+NhZva/jpv33nsPy5Ytw8iRI/H555/D398fa9eu1ddgAYChQ4ciOTkZPXr0wJ07d1CtWjVERUW9cTVYrKysEB4enuV2FWO+WfFMEZPX+HbE5DW+HTH/C9doqpi5oRN5lblERERERKbDxQ+JiIhI85iwEBERkeYxYSEiIiLNY8JCREREmseERUEPHz409SkQERG9ld7I0vxakp6eji+++AJz585FYmIizp07Bz8/P4waNQo+Pj7o2rWrqU+R6D/jxIkT2e7X6XSwtrZG4cKF35gpnP91d+7cwerVq3Hx4kUMGTIErq6uiImJgYeHB7y8vN6amPTqmLDk0oQJExAZGYkpU6age/fu+v2lS5dGRESEaglLeno6Lly4gBs3biA9Pd3gserVqysay8/PD4cOHULevHkN9t+5cwfly5fHpUuXFI33NsvJiuKmLJX9pgoMDHzhemIWFhZo3bo15s2bp1qdp3v37mHr1q0oVqwYSpQooUoMY3ve361Op4OVlRUsLS0VjXfixAkEBwfDyckJcXFx6N69O1xdXfHzzz8jPj4eS5YsUTSeqWJSzrAOSy4VLVoU8+bNQ506deDg4IDjx4/Dz88PZ86cQZUqVfDPP/8oHnP//v1o164dLl++nGVJbp1Oh7S0NEXjmZmZISEhAfny5TPYn5iYiMKFC+PRo0eKxjMVYyRmZmZmr7xAp9L/j48fP4aNjQ2OHTtmULhRDevXr3/lY5s0aaJY3HXr1uGzzz7DkCFDULFiRQDAwYMHMXXqVISHh+PJkycYNmwYWrduja+//lqRmK1atUL16tXRp08fPHjwAGXLlkVcXBxEBCtWrMCHH36oSBxTetnfbcGCBdG5c2eEh4cbFA59XcHBwShfvjymTJli8Lq6d+9etGvXDnFxcbmOoYWYAHD48GGsWrUK8fHxSE1NNXjs559/VjRWWloaFi9ejOjo6Gw/7G7dulXReEpjD0su/f333yhatGiW/enp6Xj8+LEqMT/99FMEBQVhw4YNyJ8/v2orVGd+09m0aZPBAlVpaWmIjo6Gj4+PKrGTk5MxefLk5z6x1OjViYuLyzZJePToEf7++29FYmzbts0g3rBhw9C5c2f9uln79u1DZGSkfmkLJVlYWKBw4cKKJ0LZedVlNZROsL/44gtMnz4dISEh+n1lypRBwYIFMWrUKBw8eBB2dnYYNGiQYgnLzp07MWLECADAL7/8AhHBnTt3EBkZiQkTJqiSsBj7+bF48WKMGDECnTt3NkgEIyMjMXLkSNy8eRNff/01rKys8Pnnn+c63qFDhzBv3rws+728vJCQkJDr9rUSc8WKFejYsSNCQkKwefNm1KtXD+fOnUNiYiKaN2+ueLz+/ftj8eLFaNSoEUqXLq3ae4damLDkUsmSJbFr1y54e3sb7F+9ejXKlSunSszz589j9erV2SZKSsp409HpdOjUqZPBYxYWFvDx8cHUqVNVid2tWzfs2LEDHTp0UDUpA4ybmNWoUUP/9bhx4zBt2jS0bdtWv69JkyYoU6YM5s+fn+V3roQRI0bg888/x9KlS+Hq6qp4+xmefQM1lpMnT2Z5LgKAt7c3Tp48CeDpbaPr168rFvPu3bv632VUVBQ+/PBD2NraolGjRhgyZIhicTIz5vMDACIjIzF16lS0atVKv++DDz5AmTJlMG/ePERHR6Nw4cL44osvFElYrKyssr0Nde7cObi7u+e6fa3EnDhxIr755hv07t0bDg4OmD59Onx9ffHJJ58gf/78isdbsWIFVq1ahYYNGyretlEI5cratWvFyclJJk+eLLa2tvLVV19Jt27dxNLSUjZv3qxKzFq1asnvv/+uStvZ8fHxkZs3bxotnoiIk5OT7N692yixdDqd6HQ6MTMz03+dsVlaWso777wjv/76q+JxbWxs5Ny5c1n2nz17VmxsbBSPJyISGBgo9vb2YmVlJe+8846UK1fOYFPbgwcPVG0/MDBQOnXqJI8ePdLvS01NlU6dOklgYKCIiOzevVt8fHwUi+nv7y8rV66U+/fvi7u7u0RHR4uIyLFjxyRv3ryKxcnMmM8PERFra+ts/1bPnTun/1u9dOmSYn+3Xbt2lWbNmklqaqrY29vLpUuX5PLly1KuXDnp37+/IjG0ENPW1lZiY2NFRMTV1VVOnDghIiKnT58WT09PxePlz59fzp49q3i7xsIellxq2rQpfv31V4wbNw52dnYYPXo0ypcvj19//RV169ZVJWbfvn0xaNAgJCQkoEyZMrCwsDB4PCAgQNF4sbGxirb3KlxcXFTtAcgsozfA19cXhw4dgpubm1HiFipUCAsWLMCUKVMM9i9cuBCFChVSJaYpVkBPS0vDxIkTjTKTbtasWWjSpAkKFiyofx6cPHkSaWlp+O233wA8vV3Sq1cvxWIOGDAA7du3h729Pby9vVGzZk0AT28VlSlTRrE4mRnz+QE8/Vv97rvvMHnyZIP93333nf5vNSkpCS4uLorEmzp1Kj766CPky5cPDx48QI0aNZCQkIAqVargiy++UCSGFmK6uLjg33//BfD01tOff/6JMmXK4M6dO0hJSVE83qBBgzB9+nTMnDnzjbsdBHDQ7Rspu0FtOp0OIqLKoNtx48a98PHRo0crGg8AfvjhB6xbtw6RkZGwtbVVvP1XdefOHTg7O6vS9saNG/Hhhx+iaNGiqFSpEoCn4wLOnz+PNWvWvLndts8YN24cIiMjMW7cOHTv3h1//vkn/Pz8sHLlSkRERGDfvn2Kxvv333/x448/4ty5cwCAYsWKoV27dnBwcFA0TmaHDx/GlStXULduXdjb2wMANmzYAGdnZ1StWlXxeMZ+fqxfvx4tW7ZE8eLF8e677wJ4es1nzpzB6tWr0bhxY8yZMwfnz5/HtGnTFIu7e/dunDhxAvfv30f58uURHBysWNvPs2fPHhw/ftwoMdu1a4egoCCEhYVh/Pjx+Pbbb9G0aVNs2bIF5cuXV3zQbfPmzbFt2za4urqiVKlSWT7sKh1PaUxY3kCXL19+4ePZ3cPPjWfH4jx+/BixsbHIkycPihQpgpiYGMXiZM76L1y4ABGBj49PlieWUjEz+/LLL+Hj44PWrVsDAFq2bIk1a9Ygf/782LhxI8qWLat4zKtXr2L27Nk4c+YMAKBEiRL49NNPVethAYxfa8IUM+neduXKlcPFixeN+vyIjY3FvHnzDBLBTz75RLWB9/8Ft2/fxsOHD1GgQAGkp6djypQp2Lt3L/z9/TFy5EjFeqwyhIaGvvDx77//XtF4SmPC8hpcXFxeuTvt9u3bKp+Nady7dw+dO3dG8+bN0aFDB0XaHDt27CsfGx4erkjMzHx9ffHjjz/ivffew5YtW9CqVSusXLlSP+Vw8+bNisc0tmdrTZw9exZ+fn4YOXKkarUmbGxscObMGXh7exskLKdPn0bFihVx//59ReOdP38e27Zty3b2jBq9gaaYKvqy54oazw+1zZgx45WP7devn+Lx+/Xrh6JFi2Zpe+bMmbhw4QIiIiIUj0k5w4TlNURGRr7ysWrM9Mhw+vTpbOfuK1nX4kVOnjyJDz74QLX6BMZmY2ODc+fOoVChQujfvz8ePnyo/0RZqVIlVXoC7ty5g++++w5//fUXAKBUqVLo0qWLwUwlJZmi1kSFChUwcOBAfPzxxwYxx40bhy1btmDXrl2KxVqwYAF69uwJNzc3eHp6Gnyw0Ol0qvQ89OnTRz9VNLsZO998843iMU3hzp07OHjwYLZJWceOHXPdvq+v7ysdp9PpVClr4OXlhfXr16NChQoG+2NiYtCkSRNcvXpV8Zi1a9dGjRo1siSY//zzDz788EPN10UxOhMN9qVcuHjxogQEBGSZ2WJmZiZmZmZGO49du3aJs7OzKm0fPHhQ9u/fn2X//v375dChQ6rEzJ8/v+zZs0dERN555x1ZtWqViIicOXNGHBwcFI936NAhcXV1FS8vL2nevLk0b95cChYsKHnz5pUjR44oHk9ExNHRUS5cuCAiIvb29nLx4kUREYmLixMrKytVYhpzJl3hwoVl8uTJirb5Mnnz5pUNGzYYNWaGw4cPy9KlS2Xp0qUSExOjWpz169eLg4OD6HQ6cXJyEmdnZ/3m4uKiWlxjsrKykvPnz2fZf/78edWeGzqdTtzc3KRp06Zy//59/f6EhATVXst/+uknadmypVSqVMnoswRziwmLAp48eSI//fSTjBs3TsaNGyerV6+Wx48fqxavcePG0rRpU7l586bY29vL6dOnZdeuXVKxYkXZuXOn4vGmT59usEVERMhnn30mBQoUkLZt2yoeT0Tk3XfflZ9++inL/jVr1kjFihVVidm7d2/x9vaW4OBgyZs3r/z7778iIrJ8+XJVnszVqlWTzp07G/ytPH78WDp16iTvv/++4vFERNzd3fVvbJkTls2bN0vBggVViSkisnPnTgkODhZ3d3exsbGRqlWryqZNmxSP4+DgoL8mYzHFVNHExESpVauW6HQ6cXFxERcXF9HpdFK7dm25ceOG4vH8/f2lf//+kpycrHjbL5Oeni7p6emqxylVqpR8++23WfbPmDFDSpQooUpMnU4nx44dk0qVKknp0qX1U5zVSlimT58u9vb20qdPH7G0tJRPPvlEgoODxcnJST7//HPF4ymNCUsu/fnnn+Ln5ye2trb6LNXOzk58fHzk5MmTqsTMmzevHD9+XESefmI+c+aMiIhER0fra00oycfHx2Dz8/OTSpUqyfDhw+XevXuKxxMRsbOzy/aN59KlS2Jvb69KzNTUVPnqq6+kX79+Bp9Wp02bJgsWLFA8nrW1tfz1119Z9p86dUq1OiymqDVhTF26dJE5c+YYNebXX38tvXr1MsqbaoZWrVpJUFCQnD59Wr/v1KlTEhQUJG3atFE8nq2trdETwYULF0qpUqXE0tJSLC0tpVSpUqo8DzN89913YmNjI6NHj5bt27fL9u3bZdSoUWJrayvz589XJaZOp5PExER5+PChtG3bVtzc3GTbtm2qJSzFihWTZcuWiYjhB5ZRo0ZJ7969FY+nNCYsuVS5cmX54IMP5Pbt2/p9t2/fliZNmkiVKlVUiens7CyXLl0SERE/Pz/ZunWriIhcuHBBtTc6Y3N1dZW9e/dm2b9nzx7VbkMZW758+bLtZYiKipJ8+fKpEvPOnTsSHBwszs7OYm5uLoUKFRILCwupXr26QZe0krp27Srbtm1Tpe1nTZw4Udzc3KRTp07y9ddfZ+kdVEOzZs3EyclJfH19pXHjxvrbexmbGhwdHeXgwYNZ9h84cECcnJwUj9e8eXNZuXKl4u0+z6hRo8TOzk6GDRsm69atk3Xr1smwYcPE3t5eRo0apVrc2bNni5eXl/42u6+vr0RGRqoWz8zMTBITE/Xfjx8/XqysrGT06NGqJCw2NjYSFxcnIk97W48dOyYiTwsAurq6Kh5PaSwcl0vHjh3D4cOHDaafubi44IsvvtDXK1Ba6dKlcfz4cfj6+qJSpUqYMmUKLC0tMX/+fPj5+akSM0PGwLOCBQuqGqdevXoYPnw41q1bpx+AeufOHXz++eeqFeTLYKzBzK1bt0bXrl3x9ddf47333gPwtAbEkCFDDMr1K8nJyQlbtmwxan2Lmzdvon79+nB3d0ebNm3Qvn17BAYGqhJr/vz5sLe3x44dO7Bjxw6Dx3Q6nSqzS5ydnVVZ9+VF0tPTs0xlBp4umaHGsggZywycPn0622KVSj835syZgwULFmRZtiIgIAB9+/Z9aW2o19WzZ0/07NkTN2/ehI2Njb6mjlrkmTkvI0eORIkSJVSbrOHp6Ynbt2/D29sbhQsXxv79+1G2bFnExsZmORdNMnXG9KYLCAjQl+LOLDo6WkqXLq1KzKioKFmzZo2IPB0QVqxYMf3grezOJbfS0tJk7Nix4ujoqB/Y6+TkJOPGjZO0tDTF44mIXL16Vfz8/MTJyUlq1qwpNWvWFGdnZylWrJjEx8erEtPYg5kfPXok/fr1E0tLS30MKysrGTBggDx8+FDxeKZ0+/ZtmTdvntSoUUPMzMykZMmS8sUXX+jv2VPONGnSRKpXry5///23ft/Vq1elRo0a0qxZM8XjPbtkReZNjeeGk5PTc5etUKMHyVTi4uKyvZX4559/yuLFixWP17VrVxkzZoyIiMycOVNsbGz0Pa5dunRRPJ7SOK35NWReIGv37t0YOnQoxowZg8qVKwMA9u/fj3HjxmHy5MlGq1Z6+/btHNWHyYnhw4fju+++w9ixY/VVO3fv3o0xY8age/fuqpWtTk5Oxo8//ojjx4/DxsYGAQEBaNu2bbafLJXwwQcfwNzcHAsXLoSvry8OHjyIpKQk/cq+77//vipxU1JScPHiRQBAkSJFVK9cGh0d/dyaIYsWLVI1NvC0l2758uVYtGgRzp8/jydPnqge0xhu3ryJs2fPAnhaVE2tBfMA4MqVK2jSpAlOnTqlLzJ45coVlC5dGuvXr1e9B1Rtffv2hYWFRZaquYMHD8aDBw8wa9YsReKUL18e0dHRcHFxyVK48llqTIlfsmQJgoKCULJkSYP9Dx8+xKpVqxSZLp5Zeno60tPTkSfP05srK1as0Beq++STT2BpaaloPKUxYXkNZmZmBn/YGb/CjH2Zv1e6TH5mFy5cwMWLF1G9enXY2NjoS/MrrUCBApg7d26Wbt9169ahV69e+PvvvxWPaQpubm7YunUrAgIC4OTkhIMHD6JYsWLYunUrBg0ahKNHj6oW21i32saOHYtx48YhKCgo25ohv/zyi6rxHz9+jA0bNuCHH37Ahg0b4Orqmuu/n4yy5nZ2dggLC3vhsUqWjc+QnJyMvn37YsmSJfoE0NzcHB07dsS3336rWgIqIvjjjz8MqiQbo3S9WjL/3z158gSLFy9G4cKF9R8EDxw4gPj4eP3vVQljx47FkCFDYGtrizFjxrzw9VONYnxmZmaws7PD4sWL8eGHH+r3JyYmokCBAqq+f7yJOIblNWzbts2k8ZOSktCqVSts27YNOp0O58+fh5+fH7p27QoXFxdMnTpV0Xi3b99G8eLFs+wvXry4qpV8TVGxNGO9GTc3N1y7dg3FihWDt7e3/pOzktLT0zFhwgRMnTpVX+3VwcEBgwYNwogRI7JdMyq35s6di8WLFytWnfhVbdu2DcuWLcOaNWuQnp6OFi1a4LfffkPt2rVz3fbRo0fx+PFj/dfGFhYWhh07duDXX3816IHs168fBg0ahDlz5qgSV6fToW7duqqN6ZoxYwZ69OgBa2vrl1ahVWJs0LP/dxkF3DJ6H93c3ODm5oZTp07lOlaGzEnImDFjFGs3J8aOHYsOHTrg5MmTBuegVF/CiRMnULp0aZiZmeHEiRMvPFbphXOVxh6WN1DHjh1x48YNLFy4ECVKlNBXDt20aRPCwsIUfUIDQKVKlVCpUqUsL1p9+/bFoUOHsH//fkXjAaapWPr+++9j0KBBaNasGdq1a4d//vkHI0eOxPz583HkyBH8+eefisYzxa22vHnz4uDBgyhSpIjibT+Pl5cXbt++jfr166N9+/b44IMPYGVlZbT4anNzc8Pq1av1qzRn2LZtG1q1aoWbN28qEsfYCYSvry8OHz6MvHnzvrAKrVqVZ42tW7du+Pjjj7P8P6rJzMwMCQkJuHTpEpo3b46qVati6dKluHfvnmI9LBkx8uXLp787kN3bvtp3BBRhmqEzb5/k5GT566+/5Pjx4wabGjw8PPTT0TLPpb948aLY2dkpHm/79u1iZ2cnJUqUkC5dukiXLl2kRIkSYm9vr0qhOhHTVCw19mDm/Pnzy7p167LsX7t2rRQoUEDxeCIiQ4cOlXHjxqnS9vPMnz9f/vnnH6PECg0NzbY20P379yU0NFSVmDY2Ngb1UDL8+eefYmtrq1gcHx8fuXXrlv7r522+vr6KxfwvadKkiVhZWUnBggVl8ODB+tdYNWWe1nz58mUpW7asBAYGyv79+xUbzJx5YG9cXNwLN61jD0su3bx5E6Ghofj999+zfVyNjNXBwQExMTHw9/c3WJvl8OHDCAkJQVJSkuIxr127hlmzZhncL+/VqxcKFCigeCwAcHR0xLFjx1Sfpp25uzQ7ag5mtra2xokTJ/DOO+8Y7D979iwCAwPx4MEDxWP2798fS5YsQUBAAAICArIMYFZjjEcGY4y5Mjc3x/Xr15EvXz6D/bdu3YKnp6cqA3zr1KmDvHnzYsmSJbC2tgYAPHjwAJ06dcLt27fxxx9/KB7T2MaNG4fBgwdnGY/z4MEDfPXVV4rcom3RogUWL14MR0dHtGjR4oXH/vzzz7mOl51//vkHP/30E5YtW4Zdu3ahePHiaN++Pdq1a6fKqtSZez+ApwPw27dvj+joaCQnJ2u/x8PImLDkUvv27XH58mVERESgZs2a+OWXX5CYmKgfm9CoUSPFYzZs2BAVKlTA+PHj4eDggBMnTsDb2xtt2rRBeno6Vq9erXhMY+vatSveffddfPrpp6rGyfwG5+fnh0OHDiFv3ryqxsxgiltttWrVeu5jOp1OlcXWnjfmqkuXLoqNubp37x5EBC4uLjh//rzBDJ20tDT8+uuvGDZsGK5du5brWM/6888/ERISgkePHqFs2bIAgOPHj8Pa2hqbNm1CqVKlFI9pbM9LBJOSkpAvXz5F3lhDQ0MxY8YMODg4IDQ09IXHfv/997mO9zLGmM2WedBvZuHh4di5c6ci4yXXr1//yscaa+Hc18WEJZfy58+PdevWoWLFinB0dMThw4fxzjvvYP369ZgyZQp2796teMw///wTderUQfny5bF161b99Mbbt29jz549qoxP2LVrF+bNm4dLly7hp59+gpeXF5YuXQpfX19Uq1ZN8XiTJk3CtGnT0KhRo2wLVSlVACxv3rzYuHEjKlWqBDMzMyQmJqo6HTWzHTt2oFGjRihcuDCqVKkCANi3bx+uXLmCjRs3qjaN2tiMMebq2Zl7z9LpdBg7dixGjBiR61jZSUlJwY8//mjQA9m+fXvY2NioEu95s6F0Oh2sra1RtGhRNG3aFK6urorEe95zY+vWrWjdurVi43S0Qo3ZbKbyqoP334QxLExYcsnR0REnTpyAj48PvL29sWzZMlStWhWxsbEoVaoUUlJSVIl79+5dzJw5E8ePH9dXK+3duzfy58+veKw1a9agQ4cOaN++PZYuXYrTp0/Dz88PM2fOxMaNG7Fx40bFYxprkF+PHj2wZMkS5M+fH/Hx8ShYsCDMzc2zPVaNgYXGvtVmCp6enti0aRPKli1rcAvz0qVLCAgI0M+Qyo0dO3ZARFC7dm2sWbPG4I3a0tIS3t7eb9XvtFatWoiJiUFaWhqKFSsGADh37hzMzc1RvHhxnD17FjqdDrt3785S4yMnMm6H3r17F46OjgZJYVpaGu7fv49PP/1UsbooGR48eAAR0fc8XL58Gb/88gtKliyJevXqKRors+xms7Vv3x61a9dW7Pbl+vXr0aBBA1hYWLyw90On0+GDDz5QJObbgglLLr377ruYMGECQkJC0KRJEzg7O2PSpEmYMWMGVq9erZ+S9yYrV64cBg4ciI4dOxq84Rw9ehQNGjRAQkKCqU8xV6KionDhwgX069cP48aN009tflb//v1zHSvzffolS5agdevWRp0xk5ycjMmTJz+3cJwaSZkxx1xdvnwZhQoVUmVKeGav+qYDqNPNHhERgV27duH777+Ho6MjgKcfYrp164Zq1aqhe/fuaNeuHR48eIBNmza9dpzIyEiICLp06YKIiAj9MhnA00TQx8dH3zuopHr16qFFixb49NNPcefOHRQrVgyWlpa4desWpk2bhp49eyoe01iz2Z6dtfM8b0KPh7ExYcmlH374AU+ePEHnzp1x5MgR1K9fH7dv34alpSUWL16M1q1bqxL3zp07OHjwYLZvOkpXR7S1tcXp06fh4+OT5RNyyZIl8fDhQ0XjmUrme+hqsbS0xOXLl5E/f/7njgtQU9u2bbFjxw506NAh28JxSiRlzzLFmKuUlJRs14NSqs6Eqd90vLy8sGXLliy9J6dOnUK9evXw999/IyYmBvXq1cOtW7dyHW/Hjh147733VKsy/Sw3Nzfs2LEDpUqVwsKFC/Htt9/i6NGjWLNmDUaPHo2//vpL8ZgLFixAy5Yt4ezsrHjbWpKcnIwdO3Zk+/xQY60tJbFwXC59/PHH+q8rVKiAy5cv48yZMyhcuDDc3NxUifnrr7+iffv2uH//fpZuWp1Op3jC4unpiQsXLmQZJb97927VZvF06dLlhY+rUUI+owcpO7NmzULv3r1zHaN48eIYPnw4atWqBRHBqlWr9J+Qszsfpf3+++/YsGGDvu6LMUyZMgV16tTB4cOHkZqaiqFDhxqMuVKSsWbtZf6QoMZigy9z9+5d3LhxI0vCcvPmTf3SIc7OzlnekF5XjRo19F8/fPgwS7vP+xt+XSkpKfrn4ubNm9GiRQuYmZmhcuXKuHz5sqKxgKdjVnr27IkqVaoYLWGJi4vDli1b8PjxY9SoUcMog7OPHj2Khg0bIiUlBcnJyXB1dcWtW7dga2uLfPnyaT5hYR2WN5C/v7/0799fkpOTjRJv4sSJUrJkSdm/f784ODjIrl275IcffhB3d3eZMWOGKjGbNWtmsDVq1Ei8vb3FyclJmjdvrkpMFxcXOXz4cJb9ERER4uDgoEiMPXv2SKVKlcTNzU2/iKSzs3OWzcXFRZF4z/Lx8cm2Zoja7ty5IxMmTJCWLVtKgwYNZMSIEXLt2jXF47Rr106qVq0qhw4dEjs7O9m8ebMsXbpUihUrJr/99pvi8UREIiMjs12s8tGjRxIZGalKzHbt2omvr6/8/PPPcuXKFbly5Yr8/PPP4ufnJx9//LGIiCxfvlwqVKigSLzk5GTp3bu3uLu76xfqzLwprUyZMjJ9+nSJj48XR0dH2bt3r4iIHD58WDw8PBSPJyLi6+trlNorIiJbt24VW1tb/QKSFhYWsnTpUtXj1qhRQ7p37y5paWn6Gl7x8fFSvXp1fQ0qLWPC8hoGDhz4ypsabG1t9cXijCE9PV0mTJggdnZ2+ieYtbW1jBw50mjnIPJ01egePXrIl19+qUr7CxYsEHd3d/nrr7/0+77++mtxdHRUpUCeTqfTF40ylqVLl8pHH31ktGTX2Dw9PeXAgQMiIuLg4CBnz54VEZF169ZJ1apVVYmZufhXZrdu3VLlzVxE5N9//5Vu3boZrPRtaWkp3bt3l/v374uIyNGjR+Xo0aOKxOvVq5eUKFFCVq9eLTY2NrJo0SIZP368FCxYUH744QdFYmT2008/iYWFhZiZmUndunX1+ydOnCj169dXPJ6IyMKFC6Vhw4aSlJSkSvuZVa1aVZo2bSrXrl2T27dvS69evSR//vyqx3VycpIzZ87ov8748LJ//34pVqyY6vFzi2NYXsOLallkplZdixYtWqBNmzZo1aqV4m1n5/Hjx7CwsEBqaiouXLiA+/fvo2TJkrC3t8etW7dUu/WVnbNnz6JmzZq4fv26Ku1PmTIFM2bMwO7du7Fy5UpMnDgRGzduVOUWyuXLl1G4cGFVitI9T7ly5XDx4kWICHx8fLKMSVBqyYOXrVmSmZLrl5hi1t7zpvweP34ctWrVUnW9rfv37+sHSvv5+cHe3l6VOIULF8aSJUtQs2ZNODo6IiYmBkWLFsXSpUuxfPlyVWYKJiQk4Pr16yhbtqx+nNDBgwfh6OiY7dpmuVWuXDlcuHABjx8/hre3N+zs7AweV3I5EGdnZ+zdu1d/Sy8lJQWOjo5ITExUtQ6Uu7u7fnXmd955B99++y1CQkJw5swZVKhQAcnJyarFVgLHsLwGUy9+2KhRIwwZMgSnT5/OtkaJ0rMS2rRpg9WrV8PS0tLgnnliYiLq1Kmj+Bo7L3Lx4kVVCjhlGDp0KJKSkhAUFIS0tDRs2rRJv1qs0ry9vQGoP0A0s2bNmineZnYCAwOfu2ZJZkoPSi1WrBjOnj0LHx8flC1bFvPmzYOPjw/mzp2r+JT/cuXKQafTQafToU6dOsiT538vp2lpaYiNjUX9+vUVjfmsjDd1tVdsv337tn68mqOjoz4Jq1atmiozdoCnY+c8PT0BPC0MuHXrVhQrVkyVZAUw3nMDeHo9mT/o2drawsbGBnfv3lU1YSlXrhwOHToEf39/1KhRA6NHj8atW7ewdOlSlC5dWrW4SmHC8gbq3r07gKflsp+lxqyE+Ph4dOvWDd99951+3/Xr11G7dm3VBoo9WxhLRHD9+nVs2LABnTp1UixOdovIeXl5wdbWFtWrV8fBgwdx8OBBAMqPoDfFsg6ZV6dVU2xsrFHiPKt///763rfw8HDUr18fP/74o37WnpIy3uCOHTuGkJAQg96NjCm/H374oaIxMxh7xXY/Pz/ExsaicOHCKF68OFatWoWKFSvi119/VWWQaqtWrVC9enX06dMHDx48QFBQEOLi4iAiWLFihSq/V2M9NzJs2rTJYJp4eno6oqOjDT4AKv3hc+LEifj3338BAF988QU6duyInj17wt/f3+D1Xat4S+g1aGHNC2O6efMmqlevjgYNGmDatGm4du0aatWqhbJly2LFihWq1Lx49rabmZkZ3N3dUbt2bXTp0sXg02xuvKhAXWZqrEhrimUdtKxRo0ZYuHChoj0hKSkpqs/ai4yMROvWrfXrCBmDsVds/+abb2Bubo5+/frhjz/+wAcffAARwePHjzFt2jTFp8NnLja4bNkyhIeH4/jx44iMjMT8+fNx9OhRReNluHPnjr5+1pAhQ+Dq6oqYmBh4eHjAy8tLsTiv8prJOixZsYflNTg5Oem7XTNnyG8rd3d3bN68WV+C/7fffkP58uXx448/qlag63m33VJSUnDw4EG89957isQxVU8A8LSs+bp16xAUFAQzMzN4e3ujbt26cHR0xKRJkxRLWFxdXXHu3Dm4ubm9dCFHNcdbvMzOnTsVW/AxNTUVsbGxKFKkCMqXL69Im8+jZI/fq9q8eTM2bdqEggULGuz39/dXZdrvwIED9V8HBwfjzJkzOHLkCIoWLarKrcu7d+/qqxVHRUXhww8/hK2trf52uBpOnDiB4OBgODk5IS4uDt27d4erqyt+/vlnxMfHY8mSJYrFMsVUeACoXbs2fv755yy9Yvfu3UOzZs1UGXOpJCYsryHzwlvGWIQrO8Yu/lOoUCFs2bIF77//PurWrYulS5cadbBohvPnz+P9999/Kz55JCcn64vGubi44ObNm3jnnXdQpkwZRQf4ffPNN/qaFhEREYq1q0UpKSno27cvIiMjATwtV+/n54e+ffvCy8sLw4YNUzxmWloavvnmG6xatSrb56MaSWBycnKWBfMyYqldOfnhw4fw9vbWj8FSQ6FChbBv3z64uroiKioKK1asAPB0NWW1erLCwsLQuXNnTJkyxaAeU8OGDdGuXTtVYr4qpXoft2/fnm1tnocPH2LXrl25atsYmLC8gYxR/Od5n8RTUlLw66+/GgwMM+WncqVdvXoV69evz/aNZ9q0aYrGMtYA0cw9AKboDTCm4cOH4/jx49i+fbvBgNfg4GCMGTNGlYRl7NixWLhwIQYNGoSRI0dixIgRiIuLw9q1azF69GjF4wHA+++/jyVLlmD8+PEAnt4+SE9Px5QpU155FmNOpKWlYeLEiZg7dy4SExP1ieCoUaPg4+ODrl27KhpvwIABaN++Pezt7VG4cGHUrFkTwNNeuDJlyigaK8OhQ4cwb968LPu9vLxMvvxIbnsfM8/aO336tMH1pKWlISoqStFbXmphwqKA1atXP/fTlZKflDMMHDgQH3zwAebOnQsnJyfs378fFhYW+PjjjxW7l/y2fxLPTnR0NJo0aQI/Pz+cOXMGpUuX1g/0U+O2gjEHiGa4e/cutmzZgri4OOh0Ovj5+aFOnTqKVyo1lbVr12LlypWoXLmyQcJdqlQp1db1+vHHH7FgwQI0atQIY8aMQdu2bVGkSBEEBARg//79qvR4fvXVV6hdu7ZRqgcDTwdoRkZGYsqUKfpB/wBQunRpREREKJ6w9OrVCxUrVsSVK1dQt25d/a1nPz8/TJgwQdFYGaysrPRVgjM7d+6c0VZwV0vGrD2dTofatWtnedzGxgbffvutCc4sh0xU/+WtMX36dLG3t5c+ffqIpaWlfPLJJxIcHCxOTk7y+eefqxLzTS/+kxvHjh1TrRjXu+++K6NHjxYR0VeB/Pfff6VJkyYye/ZsVWJmlpycLEeOHJGbN2+q0v7SpUvFyclJX/wvY3N2dpYVK1aoEjMnMn7nuWFjY6NvI3N7x44dE0dHx1yfY3ZsbW3l8uXLIvK0cN2RI0dEROTixYuqxExNTZXatWvLgQMHjFI9WESkSJEi8scff4iI4e/1r7/+EmdnZ1ViijytFnzmzBl5/PixajEydO3aVZo1ayapqalib28vly5dksuXL0u5cuWkf//+qsd/kdw+N+Li4iQ2NlZ0Op0cOnRI4uLi9Nu1a9fkyZMnCp6tetjDkkuzZ8/G/Pnz0bZtWyxevBhDhw6Fn58fRo8erdqtEgsLC/0njnz58iE+Ph4lSpSAk5MTrly5okrMDGqvI/KylW/VHCT7119/Yfny5QCAPHny4MGDB7C3t8e4cePQtGlT1epNGGOAaExMDEJDQ9G+fXsMHDgQxYsXh4jg9OnTiIiIQIcOHVC8eHGULVtWlfjGEhQUhA0bNqBv374AoO9lWbhwoSqrCgNAwYIFcf36dRQuXBhFihTB5s2bUb58eRw6dEiV8SQWFhY4ceIEXFxcMGLECMXbz87ff/+NokWLZtmfnp6Ox48fKx7PFGORpk6dio8++gj58uXDgwcPUKNGDSQkJKBKlSr44osvFI9nTBnjjUw12FcpTFhyKT4+Xj9jxcbGRj/HvUOHDqhcuTJmzpypeExjF/9JTk7GZ599hlWrViEpKSnL40oOgH2V4k1qDfa1s7PTJ2P58+fHxYsX9XVmlFjx9lnGfFH+9ttv0axZsyy3msqXL48lS5YgJSUF06dPV2VRyVf1+eef62eGvK6JEyeiQYMGOH36NJ48eYLp06fj9OnT2Lt3L3bs2KHQmRpq3rw5oqOjUalSJfTt2xcff/wxvvvuO8THxxvMrlFSRozJkyer0v6zSpYsiV27dmUZaLt69WqUK1dO8XimGIvk5OSELVu2YM+ePTh+/Dju37+P8uXLIzg4WPFYxvayD4IZlK77ojhTd/G86Xx9fSUmJkZERCpUqCBz584VEZFNmzaptoDdoUOHZOvWrSIikpiYKCEhIeLg4CDly5dXZfEuY68jYipNmzaV+fPni4jIoEGDpGjRojJhwgQpX7681KlTR/F4/fr1kwoVKsiuXbvEzs5O3+W7du1aCQwMVDSWv7+/bNmy5bmPb9myRfz9/RWNmdmSJUvkvffek/z580tcXJyIiHzzzTeydu1axWNduHBBunXrJu+++66UKFFC2rdvLydOnFA8zvPs27dPpk6dKuvXr1ctRp8+fcTR0VEqVKggPXr0UH0Ns7Vr14qTk5NMnjxZbG1t5auvvtKvZbR582bF4xUuXFj27dsnIoa3Q86fP6/YQqRvktzeEnr2NnB2m1q32pXEhCWXunbtKmPGjBERkZkzZ4qNjY0EBweLs7OzdOnSRfF46enpcvnyZXnw4IHibT9PoUKFZNu2bSLydEG58+fPi8jTN6EGDRoY7Tyy07BhQ8Xu21+8eFGOHz8uIiL379+XTz75RMqUKSMtWrTQv8kqyZgvynZ2dvpxFtm5fPmy2NraKhozw+zZs8XNzU0mTJhgMMbk+++/l5o1a6oS05iM+VzMULNmzedutWrVUiXmzp07JTg4WNzd3cXGxkaqVq0qmzZtUiWWMcci7d27V3799VeDfZGRkeLj4yPu7u7SvXv3bFfjNqaJEyfKP//8Y9Jz0AImLLmUlpZmMCBs+fLl0rdvX5kxY4Y8evRIlXgWFhZy7tw5xdt+nsxvdl5eXvrVcC9duiR2dnZGO4/sKDFQ01SM+aL8spWhExISVPuEVaJECfnll19ExPA6T548KXnz5lUkxuPHj7O8qSQkJMiYMWNkyJAhsmvXLkXiZMfBwUE6duwomzdvlrS0NNXi/Je8//77MmPGDBER/QBYkac9SyEhIYrGql+/vkyePFn//YkTJyRPnjzSrVs3mTp1qnh6ekp4eLiiMTMsXrxYfvvtN/33Q4YMEScnJ6lSpYoqH5JelZIfBJXEMSy5ZGZmZlDttU2bNmjTpo2q8fz9/ZGUlAR/f3/V4mRm7HVETMlYpbkB4w8QfXbtkszu3LmjeLwMsbGx2Y5zsLKyUmx12O7du8PS0lJfR+Pff//Fu+++i4cPHyJ//vz45ptvsG7dOjRs2FCReJlFRkZi2bJlaNq0KZycnNC6dWt8/PHHCAoKUjyWFjx8+BArV65ESkoKgoODVXkdMuZYpGPHjunr2QDAihUrUKlSJSxYsADA0yJ24eHhGDNmjKJxgafXOWfOHADAvn37MGvWLHzzzTf47bffMHDgQJMt7aJk1WlFmTpjetMVKVJEwsPD5ezZs0aLuX79eqlWrZqcPHlS1ThmZmaSmJgo06ZNk+nTp4vI07EO1tbWYmVlJWZmZhIREaHqObyMkj0sx48fF3d3dylatKjkyZNH3+6IESOkQ4cOisTIbNeuXWJvby+ffvqpWFtbS//+/aVu3bpiZ2cnhw8fVjSWKe9hlyhRQj9WJfP/14wZM6RcuXKKxPD39ze4PTFz5kwpUKCA3LlzR0REhg4dqvrtp3v37smiRYukbt26Ym5uLv7+/jJ27FhVY6pt4MCB0qdPH/33jx49krJly4qFhYU4OTmJnZ2d7N27V5XYxhqLZGVlJfHx8frvq1atKhMmTNB/HxsbK/b29orHFXnay5rRez106FD968yff/4pbm5uqsR8FVrtuWbCkkvTpk2ToKAgMTMzk6CgIImIiJDr16+rGtPZ2VksLS3FzMxMrK2txcXFxWBTyvNuI8TFxcmaNWv04z1MScknVp06dWTIkCFZ2t2zZ494e3srEuNZph4gagwLFiwQLy8vWbFihdjZ2cny5ctlwoQJ+q+VYGtrq79tICLSvHlz6du3r/77U6dOibu7uyKxXsWpU6ckMDDwjRjI+CKlSpWSdevW6b9ftGiRuLi4SFxcnKSnp0vnzp2lYcOGJjzD3CtcuLDs2LFDRJ4mZDY2NvqaMyJPbxGpNYHC3d1dP2kjMDBQlixZIiJPXxdMebtdqwkLbwnl0sCBAzFw4ECcO3cOP/74I2bNmoXBgwejVq1a+Pjjj9GxY0fFY5q6Cq3a64iYiilKcxcpUkTf9awlSq6c3K1bN9jY2GDkyJFISUlBu3btUKBAAUyfPl2x26fW1tYGXdj79+/HV199ZfD4/fv3FYn1PA8fPsT69euxbNkyREVFwcPDQ7WF+owlPj4eJUuW1H+/efNmfPTRR/rnf//+/VW5zbZx40aYm5sjJCTEYP+mTZuQnp6OBg0aKBarYcOGGDZsGL788kusXbsWtra2eP/99/WPnzhxAkWKFFEsXmZ169ZFt27dUK5cOZw7d07/uzx16hR8fHxUifkmY8KikHfeeQdjx47F2LFjsX//fvTs2ROhoaGKJyyPHz/Gjh07MGrUKPj6+iradnYWLlwIe3v7Fx6jRulxUzBWae4nT54gLS3NoKhYYmIi5s6di+TkZDRp0kS/MrapKH0Pu3379mjfvj1SUlJw//59/aKPSgkMDMTSpUsxadIk7Nq1C4mJiQYlyC9evIgCBQooGjPDpk2bsGzZMqxduxZ58uTBRx99hM2bN6N69eqqxDMmMzMziIj++/3792PUqFH6752dnfHPP/8oHnfYsGHZ1pgREQwbNkzRhGX8+PFo0aIFatSoAXt7e0RGRsLS0lL/+KJFi1CvXj3F4mU2a9YsjBw5EleuXMGaNWv0a7QdOXIEbdu2VSXmG83UXTxvkwMHDkj//v3F09NTbG1tpXXr1qrEcXR0NOj+VotOp5NChQqJj4/PczdfX1/Vz+NFlJzuZ6zS3J07d5YePXrov793754UKlRI3N3dJSAgQPLkySMbNmxQLN7rULJL+NKlS9nOajt37pzExsYqEmP79u1iY2Mjfn5+YmNjk6WkQM+ePaVjx46KxHqWjY2NtGrVStauXSupqamqxDCVypUry9SpU0Xk6bgKMzMzg9ee7du3q3K71NraOtu/jdjYWNWm39+5cyfbEvVJSUmqzPjUMq3eEmLCkktnz56V0aNHi7+/v+TJk0fq1asnkZGR8u+//6oWs2PHjjJt2jTV2s/wsqmwajLFdL87d+7oa+iYm5tLoUKFxMLCQqpXry73799XLI4WBoi+jJIvWNWrV5fFixdn2b906VKpUaOGIjFERE6fPi0RERGyYsWKLNOL582bJ0ePHlUsVobHjx/LjBkzVB+3Zio///yzWFpaSu3atcXDw0MaN25s8PjQoUOlZcuWisf18PCQ6OjoLPu3bNli1LFIxrBz505p3769VKlSRa5evSoiT2tcqTkV/2W0WveFCUsu6XQ6qVixokREREhCQoJRYo4fP16cnZ3lww8/lIkTJ8r06dMNNqVkzBIyhXfeeUf/grV3716xtbWVefPmyQcffCDNmzdXNfbu3btl1qxZ8uWXX76wOuzr0toA0ewombBkLjaY2fnz58XJyUmRGKZkY2Nj0poZavvjjz9kwIABMnnyZElOTjZ4bMyYMfqikkrq0aOHlClTRi5cuKDfd/78eQkICJCuXbsqHs9UMqqHd+vWTaysrPTPuW+//VaVopxarfvyqpiw5FJ2Xd3p6emyceNG+fDDD1WJaaxbNKbsYTH2dL/U1FQxNzdXfaq4iIirq6ucOnVK/33+/PkNlji4ePGi2NjYqH4eL6JkwuLo6KifCZHZ4cOHVZsuakw1atTQF8b7r+vZs6ciq43fuXNHKleuLHny5NG/tuXJk0dq1aqlyU/+ryswMFAiIyNFxPA5FxMTIx4eHorHM+UHQSVw0G0uZS6aFBsbi0WLFmHx4sW4efOmaotmqblicWbh4eEvHXCbWa9evTBu3Di4ubnlOra9vT2SkpJQuHBhbN68GWFhYQCyzgZRioWFBQoXLqzoQo7PY8oBoqZQvXp1TJo0CcuXL4e5uTmApwtmTpo0yeSDi5XQq1cvDBo0CFevXkWFChVgZ2dn8HhAQICJzsz4fvjhBwwePDjXrwFOTk7Yu3cvtmzZguPHj8PGxgYBAQFvxUDmzM6ePZvtNTk5OalSzPHKlSv6VbfXrl2LDz/8ED169EDVqlVRs2ZNxeMpztQZ05vu4cOH8sMPP0itWrXEwsJCzMzMZNq0aXL37l3VYz969EjOnDljsDSAKTk4OCj2qbxdu3ZSvnx56dq1q9ja2sqtW7dERGTdunVSqlQpRWI8a+HChdKwYUNJSkpSpf0Mphwg+qqUvId96tQpyZs3rxQpUkQ6d+4snTt3liJFioi7u7tRerTU9rwifG/KgnJK0upgTa3y9fXV33bO/LuLjIyUEiVKKB5Pq3VfXpXZy1Mays6RI0fQq1cveHp6IiIiAs2aNcOVK1dgZmaGkJAQODo6qhY7JSUFXbt2ha2tLUqVKoX4+HgAQN++fY223Hx2JNP0x9yaNWsWqlSpgps3bxptut/MmTOxc+dOFChQAMWKFUP58uUNNqXUqFEDR44cQb9+/fD9999nqcMSGBiIgQMHKhYvs8jISGzYsEH//dChQ+Hs7Iz33nsPly9f1u8fPny4YssulCxZEidOnECrVq1w48YN/Pvvv+jYsSPOnDmD0qVLKxIDeNprs3PnTlWXGchObGxslu3SpUv6fynn+vXrhxkzZmTZP3PmTAwYMEC1uEuXLkXVqlVRoEAB/fMhIiIC69atUyVe9+7d0b9/fxw4cAA6nQ7Xrl3Djz/+iMGDB6Nnz56Kx8uo+9KtW7c3s+6LqTOmN5W5ubkMGDBAzpw5Y7A/T548BuMT1NCvXz+pUKGC7Nq1S+zs7PRZ+dq1ayUwMFDV2C/ypn+6GjNmzAs3U1FyIbI3/R72y1hZWRllyj9lT6nXgAIFCmS7PMWRI0fEy8sr1+1nxxSriqenp+urPmf0zllbW8vIkSNViffPP/9I7969pUmTJvL777/r948ePdpgOQKt0oko+LH4PyQkJAT79u3DBx98gA4dOiAkJAQ6nQ4WFhY4fvy4QXVIpXl7e2PlypWoXLkyHBwccPz4cfj5+eHChQsoX758tsXPjCHzubyOEydOvPKx/6VxAbn9vWZma2uLM2fOoHDhwvjss89w/fp1LFmyBKdOnULNmjVx8+ZNBc746f9l6dKlYWZm9tL/VyX/L4OCgvDll1+iTp06irX5KpYuXYq5c+ciNjYW+/btg7e3NyIiIuDr64umTZsa9VxMSam/VWtra/z555/68RYZLly4gNKlS+Phw4e5aj87JUuWxMSJE9GsWTOD6/jzzz9Rs2ZN3Lp1S/GYGVJTU3HhwgXcv38fJUuWzNHYwf8SDrp9TZs2bcKVK1fw/fffo2fPnnjw4AFat24N4H+r7qrl5s2b2VYKTU5OVj22mgIDA6HT6Z57aynjMZ1OZ5TBsZcuXcKDBw9QokQJgxW532TGGswcGBiIhIQE5MuX74X/r0r/X06YMAGDBw/G+PHjsx0Aq8at2jlz5mD06NEYMGAAvvjiC/31ODs7IyIi4j+VsCilaNGiiIqKQp8+fQz2//7774ok7tkxxqriz2Npaanqh9zMdu3ahXnz5uHSpUv46aef4OXlhaVLl8LX11fzg+CZsORCoUKFMHr0aIwePRpbtmzB999/jzx58qBp06b46KOP8NFHHyk69iFDUFAQNmzYgL59+wL4X4K0cOFCVKlSRfF4xmKs2U/Pevz4MSZMmICYmBhUrlwZw4YNw8cff4xVq1YBAIoVK4aNGze+Gfd4X8JYa5fExsbqlzMw5v9rxvU0adLEIHlXM9H99ttvsWDBAjRr1sxgDFlQUBAGDx6seDwt+/jjjxVJCsPCwtCnTx/cvHlTP4MuOjoaU6dOVW0tNV9fXxw7dizLOmlRUVEoUaKEKjGTk5MxefJkREdH48aNG0hPTzd4XOkxUGvWrEGHDh3Qvn17xMTE4NGjRwCAu3fvYuLEidi4caOi8RRnyvtRb6Pbt2/LjBkzVF2pddeuXWJvby+ffvqpWFtbS//+/aVu3bpiZ2eX7X1fY/n0008VqcFgbGFhYeLu7i7dunUTPz8/adKkiRQrVkxWrFghq1atkjJlyki7du1Mdn5Kjg0y9j3s1NRUCQ0NNdq4ku3bt79wU4O1tbW+6Fbm/6tz586JtbW1KjGN7ffffzeovDpz5kwpW7astG3bVm7fvq1KzNmzZ4uXl5d+bIevr6++ZokajLGq+LPatGkj+fPnl6FDh8o333wjERERBpvSjF33RWlMWFR05MgR1dq+cOGCdOvWTd59910pUaKEtG/fXk6cOKFKLFO8WGU4deqU/P7777Ju3TqDTUmFCxfWr91z9uxZ0el0snHjRv3j27dvV22g36t40wczG2vtK1MpUaKErF27VkQM/69mzJgh5cqVM+WpKaZ06dL658iJEyfEyspKhg8fLpUrV5bOnTurGvvGjRsGS52oWXbghx9+kKJFi+qTJC8vL1m4cKFq8ZycnGT37t2qtf8sGxsb/RpNmf9WL168KFZWVkY7j9fFW0IqUuN2UIYiRYpkmQ6rliFDhuDLL78EAJw8eRKDBg1CWFgYtm3bhrCwMHz//feKx7x06RKaN2+OkydPGox/yOjmV7Jr/9q1ayhbtiyAp6tuW1lZGQz2e+edd5CQkKBYPGMz9WDmZs2aYe3atapN1X7WnTt38N133+Gvv/4CAJQqVQpdunSBk5OTKvHCwsLQu3dvPHz4ECKCgwcPYvny5Zg0aRIWLlyoSkxji42N1Y+xWLNmDRo3boyJEyciJiZGfxtOLRm3Fjdv3oyFCxfi119/VaV4JKD+quLPcnFxgaurq6oxMvP09MSFCxey3P7dvXu3amODFGXqjIly7nlr/Ny6dUuV21B2dnb6rDw8PFy/5MCRI0dU60Zs3LixNG3aVG7evCn29vZy+vRp2bVrl1SsWFF27typaKxnlyB4tkcjISHBpAXAclvELXMhM1MUOTPW2lciIocOHRJXV1fx8vKS5s2bS/PmzaVgwYKSN29eVXs8jf3J3NhcXFz05RqqVq0q8+bNE5GnqyeruYxEXFycjB49Wry9vcXR0VFat24tq1atUi2esS1dulQ++uijLGs0qWXixIlSsmRJ2b9/vzg4OMiuXbvkhx9+EHd3d5kxY4ZRziE3OK35DWRmZqafgZHZtWvXUKRIEcU/fbi6umL37t0oWbIkqlWrho4dO6JHjx6Ii4tDyZIlkZKSomg8AHBzc8PWrVsREBAAJycnHDx4EMWKFcPWrVsxaNAgHD16VLFYZmZmiIyM1H8Cb9u2LSIiIuDh4QHg6Sf20NBQxQdsRkZGws3NDY0aNQLwtIjb/PnzUbJkSSxfvjzL4L/Xlbkg3MsoFTMzX1/f5z6m0+kUHVj4/vvvo2jRoliwYAHy5HnagfzkyRN069YNly5dws6dOxWLlR1jfTI3tiZNmiA1NRVVq1bF+PHjERsbCy8vL2zevBl9+vTBuXPnFIuVmpqKn3/+GQsXLsSePXsQHByM33//HUePHkWZMmUUiwMA5cqVe+WZlTExMYrGzoh/8eJFiAh8fHxgYWGhakwRwcSJEzFp0iT967aVlZV+Zp3W8ZbQGySj8qNOp8PChQsN5upnVPksXry44nGrVauGsLAwVK1aFQcPHsTKlSsBAOfOnUPBggUVjwc8vR4HBwcAT5OXa9euoVixYvD29sbZs2cVj9epUyeD7z/55BOD79WYLj5x4kTMmTMHALBv3z7MmjUL33zzDX777TcMHDgQP//8syJx1EhCciLzLCF55tae0g4fPmyQrABAnjx5MHToUAQFBakS88GDBxAR2NrawtbWFjdv3kRERARKliyJevXqqRLT2GbOnIlevXph9erVmDNnDry8vAA8nWZcv359xeL07dsXy5cvh7+/Pz7++GOsXLkSefPmhYWFhX4dKiU1a9ZM8Ta1HF+n02HEiBEYMmTIm1n3xZTdO2+DlJQUg+68uLg4+eabb2TTpk2Kx8pYtVSn00mhQoUMVml+5513pF69erJ//37F416+fFkaNWokAQEBBt3cAwYMkL59+yoeT0SkWrVq+hVw27ZtK/Xr15fdu3dLx44dVVtLyNiMvSJ1ZsYYzJzZwoULpVSpUmJpaSmWlpZSqlQpWbBggeJx8uXLl+1zLyoqSvLly6d4PBGRunXrypw5c0Tk6SysfPnyScGCBcXa2lpmz56tSsy3lbm5uXz++edy7949g/3GqCBO2seEJZeefbHy8PBQ/cWqZs2aqs/OMbWoqChZs2aNiIicP39eihUrJjqdTtzc3PSl5U1FqVL5pliI7OLFixIQEJBlXIuZmZlqY1hGjRoldnZ2MmzYMH1iNGzYMLG3t5dRo0YpGqtv375SsGBBWbFihcTHx0t8fLwsX75cChYsKP3791c0Voa8efPKn3/+KSJPp8YGBARIWlqarFq1SooXL65KTFO4cOGCjBgxQtq0aaMf87Vx40b9tSth2bJlEhwcLHZ2dtKqVSv59ddf5cmTJ0ZNWA4dOiRLliyRJUuWmLRMhBru378vI0eOlCpVqkiRIkXE19fXYNM6Jiy5xBcr5V6sXiYpKUnS09ONFu95lJpmbIoVqY05mDmDm5ubLFu2LMv+ZcuWSd68eRWN9ejRI+nXr59YWlrqkzArKysZMGCAPHz4UNFYGTL3lLVs2VK/7lR8fLyqA1KNKWOF8eDgYLG0tNT//U+aNEk/CF9Jly5dktGjR0vhwoXFzc1NzMzM5KefflI8TmZXrlyRatWqiU6nExcXF3FxcRGdTidVq1aVK1euKBbHxcVFX6/K2dlZHyu7TWnGrvuiNA66zaXMa7O0atUKpUqVQnh4OK5cuYJixYopNiA1LCwM48ePh52dnb6c+vNMmzZNkZgZduzYgQYNGqBq1arYuXMn/vrrL/j5+WHy5Mk4fPgwVq9erWi8zC5cuICLFy+ievXqsLGx0VcsNSWl1ku5c+cORo4ciStXrqBnz576sQDh4eGwtLTEiBEjlDhdA8YczJzB2dkZhw4dgr+/v8H+c+fOoWLFiqqsrpySkoKLFy8CeFoCwNbWVvEYGQICAtCtWzc0b94cpUuXRlRUFKpUqYIjR46gUaNGb/SU+AxVqlRBy5YtERYWZvD3f/DgQbRo0QJXr15VJa6IYPPmzfjuu++wfv16uLm5oUWLFtmu5Jxb9evXx507dxAZGYlixYoBAM6ePYvQ0FA4OjoiKipKkTiRkZFo06YNrKyssHjx4he+nj07ti63nJ2dsWHDBlStWlXRdo3GtPnSm69MmTIyffp0iY+PF0dHR9m7d6+IiBw+fFjRKb81a9bUT22tWbPmc7datWopFjND5cqVZerUqSJi2Ltw4MAB1Qqq3bp1S2rXrq2/XZERMzQ0VMLCwlSJ+are5EJuzs7O+iJufn5+snXrVhF52oOmVm9Anz59ZODAgVn2Dxo0SHr16qVKTJGnn5iV/GT8PD/99JNYWFiImZmZBAcH6/dPnDhR6tevr3p8Y7Czs9P/3WT++4+NjTVawbGkpCT55ptvJCAgQJX2ra2t9bdoMzt8+PBb01Pm4+Mjp0+fNvVpvDYmLLmU+cWqbt26+v18scqdDh06SEhIiFy5csUgZlRUlJQsWVKVmK9KyYRl586d0r59e6lSpYpcvXpVRESWLFliUFlYSaYYzNynTx9xdHSUUqVKSdeuXaVr165SunRpcXR01CczGVtupaWlydixY8XR0VF/S8jJyUnGjRsnaWlpClxN9q5fvy4xMTEGMQ4cOCB//fWXajGNycvLS/bs2SMihn//P//8s/j5+Zny1BTj7+8vBw4cyLL/wIEDUqRIEVViHjlyxKBC+dq1a6Vp06YyfPhwefTokeLxjF33RWlMWBTAFyvleXh4yLFjx7LEvHjxomoDUl+VUgnL6tWrxcbGRrp16yZWVlb6Nr/99ltp0KBBrtvPjikGM7+oR1Dp3sFhw4aJu7u7zJ49W44fPy7Hjx+XWbNmibu7u3z++ecKXM3znT9/XqKioiQlJUVERBPjrZQyaNAgqVatmly/fl0cHBzk/Pnzsnv3bvHz89OP2XnTrV27VipWrCiHDh3S7zt06JBUrlxZn+QrLSgoSFavXi0i/yuP37ZtWylatKgqg8QDAwPFwcFB7O3tpXTp0lKuXDmDTes4hiUXHj9+DBsbGxw7dgylS5dWNVaLFi1e+Vil6ndkGDx4MA4cOICffvoJ77zzDmJiYpCYmIiOHTuiY8eOCA8PVzQe8HScSExMDPz9/Q3umR8+fBghISFISkpSPGZOzk2JMSzlypXDwIED0bFjR4M2jx49igYNGhht7MPt27fh4uJi8rFBSihQoADmzp2LJk2aGOxft24devXqhb///lvxmElJSWjVqhW2bdsGnU6H8+fPw8/PD126dIGLiwumTp2qeExjS01NRe/evbF48WKkpaUhT548SEtLQ7t27bB48WJVaqQYw7N/98nJyXjy5IlB0cE8efLAzs4Ot2/fVjy+k5MTYmJiUKRIEXz55ZfYunUrNm3ahD179qBNmza4cuWKovHGjh37wsfVeC1XEgvH5YKFhQUKFy6sypL1z8q8DoqI4JdffoGTk5O+GNaRI0dw586dHCU2r2rixIno3bs3ChUqhLS0NJQsWVL/YjVy5EjF4wFPK5YuWbJEX31Rp9MhPT0dU6ZMQa1atVSJ+ao+//xzRdb/OHv2LKpXr55lv5OTkyoDUTPLPJjZ1dUVb8vnltu3b2dbPLF48eKqvOEAwMCBA2FhYYH4+HiUKFFCv79169YICwt74xMWEUFCQgJmzJiB0aNH4+TJk7h//z7KlSuXZSD1myYiIsKk8UUE6enpAIA//vgDjRs3BgAUKlQIt27dUjye1hOSlzJl987bYOHChdKwYUNVVxB91tChQ6Vbt27y5MkT/b4nT55Ijx49ZPDgwYrGSk9Pl8uXL0tKSorEx8fLhg0bZOXKlXLu3DlF4zzr5MmTki9fPqlfv75YWlrKRx99JCVKlBAPDw+5cOGCKjEXL14sv/32m/77IUOGiJOTk1SpUkXi4uIUj+fr6ytbtmwREcPbTJGRkVKiRAnF44loezCzEipWrJhtMcM+ffpIpUqVVImp5duXSkhLSxMLCwvVn/MZHj9+LGPHjjXKgGlTq1WrlnTs2FGWLFkiFhYWcv78eRF5Oo3c29vbtCenQUxYcikwMFDs7e3FyspK3nnnHaPcE3Rzc5MzZ85k2X/mzBlxdXVVNJaxX6wy++eff2TChAnSsmVLadCggYwYMUKRgm3P88477+jHcezdu1dsbW1l3rx58sEHH0jz5s0Vj2eKhci0PJhZCdu3bxc7OzspUaKEdOnSRbp06SIlSpQQe3t71erM2Nvb658fmX+nGQsxvg1Kliwp+/btM1o8e3t7/YKrpvDgwQO5e/euwaaG48eP6wegZx4L1KdPH2nbtq0iMUxZ90VpvCWUS6ZYi+LJkyc4c+aMvlZAhjNnzui7F5ViZmYGf39/JCUlGb3719raGnXr1kXZsmX113Xo0CEAyDJGQQlXrlxB0aJFAQBr167Fhx9+iB49eqBq1aqoWbOm4vGGDRuG9PR01KlTBykpKahevbp+IbK+ffsqHg8ANm/ejE2bNmVZA8rf3z9HiyRqVY0aNXDu3DnMmjULZ86cAfB0/FevXr1QoEABVWJq+falUiZPnowhQ4Zgzpw5qo/XA4DatWtjx44d8PHxUT1WhuTkZHz22WdYtWpVtmPk1Lj1HxAQgJMnT2bZ/9VXXyk2Luibb77Rr8v2zTffvNlj1UydMVHODRw4UPLmzStTp06VXbt2ya5du+Trr78WNzc3RaaGPmv9+vVSrVo1OXnypOJtP8/vv/+ur3CZUT4+cxl5NZiiVL7I0+qsp06dkgMHDsi///6rWhyR/0ZvgLGZ4valsTk7O+urB1tbW6v+yXzOnDni6ekpgwYNkmXLlhllzatevXpJiRIl9LP3Fi1aJOPHj5eCBQvKDz/8oEpMyhnOEnoDpaen4+uvv8b06dNx/fp1AED+/PnRv39/DBo0SPER+y4uLkhJScGTJ09gaWkJGxsbg8fVGMzo7++PevXqYfTo0fDw8FC8/ey0b98eZ86cQbly5bB8+XLEx8cjb968WL9+PT7//HP8+eefRjkPNTVs2BAVKlTA+PHj4eDggBMnTsDb2xtt2rRBenq6qlWL1XLixIlXPjYgIECVc7h79y5mzpyJ48eP4/79+yhfvjx69+6N/PnzqxLP2CIjI1/4uNIVWc3MzJ77mE6nU6W3o3DhwliyZAlq1qwJR0dHxMTEoGjRoli6dCmWL1+OjRs3KhLH1dUV586dg5ub20tn5yn92hoTEwMLCwuUKVMGwNPZc99//z1KliyJMWPGwNLSUtF4SmPC8hpM+Qf3rHv37gEAHB0dVYth7Bcr4On1HD16FEWKFFG87ecxdqn85ORkTJ48GdHR0bhx40aW23mXLl1SNB4A/Pnnn6hTpw7Kly+PrVu3okmTJjh16hRu376NPXv2GPX3rRQzMzPodLosyzZkvLRl3meMGX0ZHj58iJkzZ2Lw4MFGi0mvz97eHqdPn0bhwoVRsGBB/Pzzz6hYsSJiY2NRpkwZ3L9/X5E4mUvzG/u19d1338WwYcPw4Ycf4tKlSyhZsiRatGiBQ4cOoVGjRiafNfUyTFhegyn/4DI8efIE27dvx8WLF9GuXTs4ODjg2rVrcHR0hL29vSoxjalLly6oWrUqunbtaupTUU3btm2xY8cOdOjQAfnz58+S+Pbv31+VuHfu3MGsWbPemt6AzGNvjh49isGDB2PIkCGoUqUKAGDfvn2YOnUqpkyZoviYs5s3b+LAgQOwtLREnTp1YG5ujsePH2P27NmYNGkSnjx5osr0VGOLj49/4eOFCxc20pmoJyAgAN9++y1q1KiB4OBgBAYG4uuvv8aMGTMwZcoU1dZLMiZj131RGhOWN9Dly5dRv359xMfH49GjRzh37hz8/PzQv39/PHr0CHPnzlU0nilerFJSUtCyZUu4u7ujTJkysLCwMHi8X79+isQx5e0EUy1E9vDhQ5w4cSLbXh01BjMbU8WKFTFmzBg0bNjQYP/GjRsxatQoHDlyRLFYu3fvRuPGjXHv3j3odDoEBQXh+++/R7NmzZAnTx7069cPnTp1ynIL9U2U0Yv1PEr0XM2YMQM9evSAtbX1Sxc3VOr5DwDm5ua4fv06fvzxR5ibm6Nfv374448/8MEHH0BE8PjxY0ybNk2VDxDGfm11dHTEkSNH4O/vj7p166Jx48bo378/4uPjUaxYMTx48EDReEpjwqKghw8fIjU11WCfGrdqmjVrBgcHB3z33XfImzevvkLq9u3b0b17d5w/f17ReMZ4sXrWd999h08//RTW1tbImzevQXydTqfY7ZLMtxOyk/lWg9LX6evri40bNxoUG1NbVFQUOnTogNu3b2e5ZrXGBhiTjY0NYmJisvxO//rrL5QvX17RF+SaNWuiQIEC+PzzzxEZGYmpU6fC398fX3zxBT766CPF4mjB8ePHDb5//Pgxjh49imnTpuGLL75QpGClr68vDh8+jLx588LX1/e5xyn5/AeevgYkJCQgX758BvsvX76MI0eOoGjRoqqNfTL2a2vt2rVRqFAhBAcHo2vXrjh9+jSKFi2KHTt2oFOnToiLi1M0nuKMP8737XL//n3p3bu3uLu76xdby7ypwdXVVV+H5dnFCNVYVfTYsWMG26FDh2T+/PlSvHhx/bo0SvPw8JAvvvhC1QXrRETi4uJeeVOaKRYiK1q0qPTq1UsSEhKMFtOYypUrJx06dDBYOO7Ro0fSoUMHxesiubq6yqlTp0REJCUlRczMzGTt2rWKxtC63377TWrUqGHq08gVnU4niYmJJolt7NdWY9R9URN7WHKpd+/e2LZtG8aPH48OHTpg1qxZ+PvvvzFv3jxMnjwZ7du3Vzymi4sL9uzZg5IlSxqsQbN79258+OGHSExMVDxmdjZs2ICvvvoK27dvV7xtV1dXHDp06I0cBPqqypUrh4sXL0JE4OPjk+W2V0xMjOIxTTGY2ZgOHjyo78rP+FR84sQJ6HQ6/Prrr6hYsaJisZ79ZO7g4IBjx469tb/b7Fy4cAFly5ZFcnKyqU/ltZmZmWHChAkvHfun5G2ol1HztTU7Dx8+hLm5eZbXIK1h4bhc+vXXX/VT4UJDQ/H++++jaNGi8Pb2xo8//qhKwlKvXj1ERERg/vz5AJ52kd6/fx/h4eFZ7t2rqVixYvpCbkrr1KkTVq5cic8//1yV9l/k9OnTiI+Pz3J7T+nxHaYoOvjRRx9h+/btb+2basWKFXHp0iX8+OOP+sJxrVu3Rrt27WBnZ6d4vNOnT+sXqRQRnD17Nsubt1q3E4wpYzZiBhHB9evXMWbMGNUKSl69ehXr16/P9rk4bdo0RWPNnTv3heUgdDqdURMWNV9bs2NtbW20WLnBHpZcMtZUuMyuXr2KkJAQiAjOnz+PoKAgnD9/Hm5ubti5c2eWe7G59aIXqzNnzuDYsWOKxgOefppZsmQJypYti4CAgCyZv9IvWMDTacTNmzfHyZMnDca1ZNxjftPHdwDGG8xsKjt37sR7772nX203w5MnT7B3795sF5t8XS8a/6Tm2CdTyG6shYigUKFCWLFihX5GllKio6PRpEkT+Pn54cyZMyhdujTi4uIgIvop+Up53hgWYzDGa6uWynDkFntYcsnPzw+xsbEoXLgwihcvjlWrVqFixYr49ddf4ezsrErMggUL4vjx41ixYgVOnDiB+/fvo2vXrmjfvr0qMxKcnZ1f+GKlhpMnT6JcuXIAkKVgm1qlpfv37w9fX19ER0fD19cXBw8eRFJSEgYNGoSvv/5alZjGtnz5cmzevBnW1tbYvn17lsHMb3rCUqtWLVy/fj3Lm8/du3dRq1YtRZOH2NhYxdrSum3bthl8b2ZmBnd3dxQtWjRLcqiE4cOHY/DgwRg7diwcHBywZs0a5MuXD+3bt9fXR1KKKUvVv+i1dfny5YrEyFyaX+t1Vl6GPSy59M033xh9Kpyx7dixw+B7tV+sTMXNzQ1bt25FQEAAnJyccPDgQRQrVgxbt27FoEGDcPTo0VzHMPWnHU9PT/Tr1w/Dhg17YTXRN5WZmRkSExPh7u5usP/cuXMICgrK8onWmHr16oVx48bBzc3NZOfwuozZcwUYjgdycXHB7t27UapUKRw/fhxNmzZVdDaLKXtY/iuvrUrhbySXBg4cqP86ODgYZ86cUW0q3M6dO1/pOKVfPHQ63XNfrHbu3Kl4PFNJS0vTfxJxc3PDtWvXUKxYMXh7e+Ps2bOKxDD1QmSpqalo3br1W5esZEyr1el06Ny5M6ysrPSPpaWl4cSJE3jvvfdMdXoAgB9++AGDBw9+IxMWY/ZcAYCdnZ1+3Er+/Plx8eJFlCpVCgAUL8QXHh6eo2KbSiaee/fuhYeHB7p06WKwf9GiRbh58yY+++yzXMfI7E0vAMiERWHe3t7w9vZWpe2aNWvq3+Ce1zGmxj1zY79YmUrp0qVx/Phx+Pr6olKlSpgyZQosLS0xf/58+Pn5KRIjc+Xjzp07K9JmTuObajCzmpycnAA8fV44ODgY3Bq1tLRE5cqV0b17d1OdHoDnP2ffBPLMsgcZkpKSVBnMXLlyZezevRslSpRAw4YNMWjQIJw8eRI///wzKleurGis8PDwHB2vZOI5b948LFu2LMv+UqVKoU2bNoonLD4+PkavqaUkJiwKiI6Ofu56MIsWLVIsjouLCxwcHNC5c2d06NDBaJ/UjP1iZSojR47Uz/AYN24cGjdujPfffx958+bFypUrFY9nioXI0tLSMGXKFGzatMlog5mN4fvvvwfw9AV58ODBb9XfpSmZqudq2rRp+gkLY8eOxf3797Fy5Ur4+/ub/G9UycQzISEh2yUx3N3d9QvbKunZ29rPFgDUOiYsuTR27FiMGzcOQUFB2a4Ho6Tr16/jl19+waJFizBlyhQ0bNgQXbt2Rf369VWJ+yZ0syspJCRE/3XRokVx5swZ3L59+6VjTV7XJ598gmHDhqFMmTK4dOkSWrdujRYtWuCnn35CSkqKKgPkTDGY2Zhy+mmZXsxUPVeZezTt7OwUX25EKwoVKoQ9e/Zkqey7Z88eFChQQPF4ZcuWzbIvKCgIBQoUwFdffaVIxWI1MWHJpblz52Lx4sXo0KGD6rEsLS3RunVrtG7dGvHx8Vi8eDH69OmDR48eoVOnThg7dqyiA7XehG52NVy4cAEXL15E9erV4erqqlpX/rlz5xAYGAgA+Omnn1CjRg0sW7ZMvxCZGgnLs7M93ja+vr4vTLzUWAH7bWbqnqvU1NRse661PtbiVXXv3h0DBgzA48ePUbt2bQBPe+yHDh2KQYMGGe08jF335XUxYcml1NRUk/QyFC5cGKNHj0aHDh3QtWtXTJ48GYMGDYKrq6tiMUz9YmVsSUlJaNWqFbZt2wadTofz58/Dz88PXbt2hYuLC6ZOnapoPBHRvxD/8ccfaNy4MYCnn7rehhV+TWHAgAEG32d0eUdFRWHIkCGmOam3gLF7rs6dO4euXbti7969Bvvfpto2ADBkyBAkJSWhV69e+kHG1tbW+OyzzzB8+HDF45miAKCSmLDkUrdu3bBs2TKMGjXKaDEfPXqENWvWYNGiRdi3bx8aNWqEDRs2KJqsZPZf6WYfOHAgLCwsEB8fb7B4XuvWrREWFqZ4whIUFIQJEyYgODgYO3bswJw5cwA8re/h4eGhaKz/iueVEZg1axYOHz5s5LMx9PHHH6uyGKqxrF69GqtWrcq28qzSy0iEhoYiT548+O2331S/1W5KOp0OX375JUaNGoW//voLNjY28Pf3N7j9riRj1H1RE+uwvIawsDD91+np6YiMjERAQIDqgxgPHjyI77//HitWrICPjw9CQ0Px8ccfq5aoZGbMFytT8fT0xKZNm1C2bFmDNZouXbqEgIAAxasWnzhxAu3bt0d8fDzCwsL0iWHfvn2RlJSU7ewBej2XLl1CYGCgKnVYoqKiYG9vj2rVqgF4mhwtWLAAJUuWxKxZs+Di4qJ4TGObMWMGRowYgc6dO2P+/PkIDQ3FxYsXcejQIfTu3VvxAZt2dnY4cuQIihcvrmi7SujZsyfGjx//Rk5Pf9Prvmj/DDXo2ZHWGeMQ1B7EWLlyZRQuXBj9+vVDhQoVAAC7d+/OcpzSa95kfrFat25dlhert0VycjJsbW2z7L99+7Yqn3gCAgJw8uTJLPu/+uqrF65rQjm3evVq1RL7IUOG4MsvvwTwdFDzoEGDEBYWhm3btiEsLEx/a/VNNnv2bMyfPx9t27bF4sWLMXToUPj5+WH06NGqFDgsWbKk0W+LvmrimdET+iYydt0Xxam7GDQpSafTvXQzMzNTPG6xYsVk2bJlIiJib28vFy9eFBGRUaNGSe/evRWPZyoNGjSQkSNHisjT67x06ZKkpaVJy5Yt5cMPPzTx2dGrCAwMlHLlyum3wMBA8fT0FHNzc5k3b54qMe3s7CQ2NlZERMLDw/V/K0eOHBEPDw9VYhqbjY2NxMXFiYiIu7u7HDt2TEREzp07J66urorEuHv3rn6Ljo6WKlWqyLZt2+TWrVsGj929e1eReM8qXbq0bNiwQURETpw4IVZWVjJ8+HCpXLmydO7cWZWYxubt7S179uzJsn///v3i4+NjgjPKGfaw5NLdu3eRlpaW5dPb7du3kSdPHkXvWT87Ut5Y4uPj9QOLbWxs8O+//wIAOnTogMqVK2PmzJkmOS+lTZkyBXXq1MHhw4eRmpqKoUOH4tSpU7h9+zb27NmjSAxTl+Z/2z27AnZGl3fNmjVVu71gaWmJlJQUAE8HT3fs2BHA0/9rUy4FoCRPT0/cvn0b3t7eKFy4MPbv34+yZcsiNjZWsVl0z46vEBHUqVPH4BhRcdBtbGwsSpYsCQBYs2YNGjdujIkTJyImJgYNGzZUPJ4pGLvui9KYsORSmzZt8MEHH6BXr14G+1etWoX169dj48aNJjozoFGjRli4cGG2f6A5YYwXKy0oXbo0zp49i1mzZsHBwQH3799HixYt0Lt371z/DjO8TQuRaZEpBohXq1YNYWFhqFq1Kg4ePKgvMnju3DkULFjQ6Oejhtq1a2P9+vUoV64cQkNDMXDgQKxevRqHDx9WrHaHqafc/xcST2PXfVEaB93mkqurK/bs2WMwqwQAzpw5g6pVqyIpKclEZwaDgaO50a1bNxQqVAjh4eGYNWsWhgwZgqpVq+pfrL777juFztj0Hj58iBMnTmRb+0HpsUGkvL///htr1qzBuXPnADytL9GiRQt4eXmpFjM+Ph69evXClStX0K9fP3Tt2hXA01lnaWlpmDFjhmqxjSU9PR3p6en6gZkrVqzA3r174e/vj08++USVqszG1qRJE6SmpqJq1aoYP348YmNj4eXlhc2bN6NPnz76v6k32ZQpUzBlyhR89dVX2dZ9UWMqtZKYsOSSnZ0d9u/fry+vnuHkyZOoVKmSPmM3BaUSlv/CixXwdNBdhw4dcPv27Sw9R2p0Q7/pC5FpzezZsxEWFobU1FT9rdh79+7B0tIS06ZNy9ILStr1/fffw97eHi1btjTYn1EFOvOaXEr5LySeIoJhw4ZhxowZWeq+jB492sRn9wpMNXjmbVGzZk3p06dPlv29evWSatWqmeCM/ifzAFl6uaJFi0qvXr0kISHBKPEyBkk/b6P/a+/Oo5q80j+Af9+ggSibUgIuYGBAKajIqQuKgkBbl+mc0+K044JUWtGjVigYlFOtWEXcpmoRq9IqKFqpYk+11qXayiaCKFEWFwxLcRQ3GFHKKEru7w9/ZkxBR8mbvAk8n3NyTnPfNM/TcvLycHPvc1/ewYMHmYmJCZs3bx67fv26evz69essMjKSderUSb2gUheUSiVbuHAhmzhxIrt58yZjjLFDhw6xkpISncXUt6ysLDZlyhTm7e3N/vWvfzHGGNuxYwfLzs7mPZarqyv77bffWoxnZGSwvn378h6vo7l//z47ffo0Ky4uZg8ePBA6nZdGBYuWcnJymJmZGRs1ahRbsmQJW7JkCRs1ahQzMzNjWVlZgubGZ8Giz5uVUCwsLJhSqdRbvHPnzmk8CgoKWFJSEnNzc2P79u3TWx7tgZ+fH1u4cOFzry9cuJD5+fnpJHZGRgaTSCTszTffZGKxWP2ZW7FiRbvZXZaens4kEgmbPn06MzU1Vf83btiwgY0bN473eKampuqdV8+qrKxkZmZmvMd7qiMUnsaMChYeKBQKNnnyZObu7s7eeOMNFhoaysrKyoROi7eCRd83K6GEhoayb7/9Vug02MGDB3X2y7W9srCwYJcuXXru9UuXLjELCwudxPb29mZffvklY0zzM5efn8969eqlk5j6NmjQILZ9+3bGmOZ/Y2FhoU62bjs4OLD9+/e3GP/xxx919v+0IxSexo52CfFg0KBB2LVrl9Bp6ExcXBw2b96MkJAQpKWlqcd9fHwQFxcnYGb8SkxMxPvvv4/s7GwMGDCgRdfi8PBwveRhLAeRGZLm5uYWP69nde7cWWfnzxQXF7falVgqlbabM6EuX74MX1/fFuNWVla4e/cu7/EmTZqE8PBwWFhYqONmZmYiIiICEydO5D0eAMTExCAuLg5RUVHqnXzAkx1S7aV1g7GjgoUHKpUKSqWy1Z0lrX3I9eWzzz7jpbunvm9WQtm9ezd++eUXmJmZISMjQ6MnBMdxvBcsxn4QmSHx8PDA/v37ERkZ2er1H3/8ER4eHjqJbW1tjZqamhZbRRUKhU53J+mTvb09lEolZDKZxnhOTo7Wi/pbs2zZMlRVVSEwMFC92F+lUiEkJATx8fG8xwM6RuFp9ISe4jF2p06dYk5OTkwkEuml6yxjjKWkpLCDBw+qn0dHRzMrKys2fPhwdTdKPjk5ObFjx44xxjSng7dv385ef/113uMJxc7Oji1fvpw1NzfrJV5ri245jmOOjo6tdqMkz5eSksIkEgnbuHEje/TokXr80aNHLDExkUkkEpacnKyT2PPmzWMjR45kNTU1zMLCgl25coXl5OQwZ2dntmTJEp3E1Lf4+Hjm7u7O8vLymIWFBcvOzmY7d+5ktra2LCEhQWdxL1++zPbs2cN++uknndzbntWrVy/15+7Z+9wPP/zAnJ2ddRqbvBwqWLTk6enJ3n//fXbhwgX273//m929e1fjoQt9+/Zlv/76K2OMsdzcXNalSxe2ZcsW9re//Y299957vMcT6malb926ddProtuMjAyNR1ZWFrt48aLGL1zy8ubNm8c4jmOWlpbqtvyWlpZMJBKxTz/9VGdxHz58yKZPn846derEOI5jnTt3ZiKRiAUHB7PHjx/rLK6unT9/XqN4j4uLY127dlX/QWZmZqY+yqI96AiFp7GjPixa6tq1K86fPw8XFxe9xezSpQsuXboER0dHLFiwADU1NdixYwdKS0sxevRo3L59W+sYRUVF6N+/P0QiEQBg+fLlWLFihbqvjKmpKeRyOZYtW6Z1LEMRGRkJW1tbfPbZZ3qJt2LFCuM+iMwA5eXlYffu3bhy5QoAoG/fvpg4cSK8vb11Eo8xhqtXr8LW1hZ37txBcXExGhoa4OXlZfRf65mYmKCmpgZSqRTOzs4oKCiAhYUFlEolGhoa4O7uDnNzc97iRUVFvfRr165dy1vcp5qamjBnzhykpKSgubkZnTp1QnNzMyZPnoyUlBQ6kNQAUMGipYCAAMyfPx9jx47VW0ypVIqjR4/Cy8sLXl5eiIqKwtSpU1FeXg5PT080NDRoHUPfNytDEB4ejh07dsDT0xMDBw5ssYiT75ukTCbDd999pz6n6an8/HxMnDgRlZWVvMYj/zV79mwsXboUr732mlbvo1KpYGZmhtLSUqMvUP7MxsYGhw4dwrBhwyASiXDz5k3Y2trqLJ6/v/9LvY7jOPz222+8xm7PhWd7Qotu26CoqEj9z3PnzsW8efNw48aNVneWDBw4kPf4b731FqZPnw4vLy+UlZWpD+YqLS1tsSiuraytrVFZWQmpVIqqqiqoVCqIxWL14WDtUXFxMby8vAAAJSUlGtdedEhhWxn7QWTGbOfOnZDL5VoXLCKRCK6urqitrW13v9gmTJgAPz8/9OjRAxzHYfDgwc+dZaioqNA6npBnCTHG4OLioi48HRwcBMuFPB8VLG0waNAgcByn0b792Wn9p9d0daroxo0bsWjRIly9ehX79u2DjY0NAODs2bOYNGkSLzH0fbMyBPq+YRr7QWTGjM+J5ZUrVyI6OhqbNm1C//79eXtfoSUlJSEoKAhKpRLh4eEICwvT2O6rD0qlEuXl5fD19YVEIlHfV/nWngvP9oS+EmqD33///aVf26dPHx1moltHjhxR36yWLl363JtVRESEnjNrH4z9IDJjxtc5WwDQrVs3NDY24vHjxxCLxZBIJBrX6+rqtI4htNDQUCQkJOitYKmtrcUHH3yAEydOgOM4XLlyBc7Ozvjoo4/QrVs3fPnll7zH/Omnn7B69ep2V3i2J1SwGKns7Gxs2bIFFRUV2Lt3L3r16oXU1FQ4OTlh5MiRvMbS982qo2DGfhCZEeOzYNm+ffsLr+vioL72LiQkBLdu3cK3336L119/Xf2zOnr0KKKiolBaWsp7zI5QeBo7Kli0dODAgVbHOY6DmZkZXFxcWkz5a2vfvn2YOnUqpkyZgtTUVFy4cAHOzs5ITEzEoUOHcOjQIV7jEd1qaGjAxYsXIZFI4OrqClNTU6FTavf4LFgI/+zt7XH06FF4enpq/KwqKiowcOBAXjYW/BkVnoaP1rBo6d13322xngXQXMcycuRI/Pjjj+jWrRsvMTtKq/yOwtzcHEOGDBE6DdJG1dXVL7zu6Oiop0zajz/++ANdunRpMV5XV6ezgp4KEsMnEjoBY3fs2DEMGTIEx44dQ319Perr63Hs2DEMGzYMBw8eRFZWFmprayGXy3mL2VFa5ROiK8HBwbC0tOTlvWQyGZycnJ77IK9u1KhR2LFjh/o5x3FQqVRYvXr1S29/flXV1dUvfBDh0QyLliIiIpCUlKTRSyMwMBBmZmaYMWMGSktLsX79+hbNwbSh73M9CDEWR44cgbm5uXod18aNG/HNN9/A3d0dGzduVM9ybtq0ibeYCoVC4/mjR4+gUCiwdu1aLF++nLc4Hcnq1asRGBiIM2fOoKmpCfPnz0dpaSnq6upw8uRJncSUyWQv3IGkq8MzycujgkVL5eXlrf6lZmlpqd7u6+rqyuvhWWFhYYiIiMC2bdvAcRyuX7+OU6dOQS6X4/PPP+ctDiHGJjo6GqtWrQLwpK/OvHnzEBUVhRMnTiAqKgrJycm8x/T09GwxNnjwYPTs2RNr1qxBUFAQ7zHbu/79+6OsrAwbNmyAhYUFGhoaEBQUhDlz5rTau4gPVHgaAf2eBND++Pj4sLFjx7Jbt26px27dusXGjh3LRo0axRhj7NixY6xv3768xVSpVO3+XA9C2qJr166ssrKSMcZYbGwsmzBhAmOMsbNnzzI7Ozu95nLlyhXWpUsXvcYk/Dt48CDz8/MTOg3CGKM1LFraunUrKisr0bt3b7i4uMDFxQW9e/dGVVUVvv32WwBPdoEsWrSIt5gcx2HhwoWoq6tDSUkJ8vLycPv27XZ1rg8hbSEWi9XnXR0/fhxvv/02AKB79+64d++eTmLeu3dP41FfX49Lly5h0aJF1IRMC9nZ2QgODsaIESNw7do1AEBqaipycnL0mke/fv1QUFCg15ikdfSVkJb69euHCxcu4JdffkFZWZl67K233lIfHPjuu+/qJHZ7b5VPyKsaOXIkoqKi4OPjg9OnT+P7778HAJSVlaF37946iWltbd1i7QNjDA4ODhq7+MjLe7Z1Q2FhIR4+fAgAqK+vR3x8vE5aN/y5oGWMoaamBkuWLKHC00BQHxYj9Mcff2DlypX49ddfcevWLahUKo3r7aVVPiGvqrq6GrNnz8bVq1cRHh6Ojz/+GMCTk7ibm5uRkJDAe8zMzEyN5yKRCLa2tnBxcUGnTvQ3YVt4eXkhMjISISEhGn1YFAoFxo0bhxs3bvAeUyQSvbDwHD58OO8xyauhT1MbJCQkYMaMGTAzM/ufN8Dw8HDe40+fPh2ZmZmYOnWq+qwfQsiTnicHDx5sMb5u3TqdxeQ4DiNGjGhRnDx+/BhZWVmttiAgLyZE64Y/nyVGhafhoZ9CG6xbtw5TpkyBmZnZC2+EHMfppGA5fPgwfv75Z/j4+PD+3oQYu/LyciQnJ6O8vBxfffUVpFIpDh8+DEdHR3h4ePAez9/fHzU1NZBKpRrj9fX18Pf3p+2wbSBE6wYqPA0fLbptg8rKSvUJyZWVlc996OqrmW7duqF79+46eW9CjFlmZiYGDBiA/Px8/PDDD+oW7ufPn0dsbKxOYrLnnCBcW1uLrl276iRme/e0dUN+fr66dcOuXbsgl8sxa9YsncT09/dv9bygp4UnER7NsBihZcuWYfHixdi+fXur7asJ6ahiYmIQFxeHqKgojcM6AwICkJiYyGusp/1VOI7DtGnTNFrGNzc3o6ioSKOhJHl5MTExUKlUCAwMRGNjI3x9fWFqagq5XI65c+fqJCYVnoaPCpY2ioqKeqnXrV27lvfYX375JcrLy2FnZweZTIbOnTtrXC8sLOQ9JiHGoLi4GN99912LcalUymvzRuDJegrgyS86CwsLjdN9xWIxvL29ERYWxmvMjuJp64bo6GgolUo0NDTA3d0d5ubmvMeiwtN4UMHSRn/uipiTk4M33nhD46alq8WwutomTYixs7a2Rk1NTYszfBQKBXr16sVrrKddc2UyGeRyOf0VrgP6aN1AhafxoG3NPKHj6gkRnlwuR35+Pvbu3Yu+ffuisLAQN2/eREhICEJCQnS2joXwS4jWDV988QUVngaOChaeUMFCiPCampowZ84cpKSkoLm5GZ06dUJzczMmT56MlJQUmJiY6CRueno69uzZg+rqajQ1NWlco69oX92kSZNe2LohIiJCoMyIkKhg4YmuC5bu3bujrKwMr732Grp16/bCr5taW+lOSHvHGMPVq1dha2uLO3fuoLi4GA0NDfDy8tJpp9KEhAQsXLgQ06ZNQ1JSEkJDQ1FeXo6CggLMmTOHDs5rA2tra0FaN1DhadhoDYuRWLdunXrXw7p166hZHCF/whiDi4sLSktL4erqCgcHB73E/frrr5GUlIRJkyYhJSUF8+fPh7OzMxYvXkx/PLSREK0bni089+/f36LwJMKjGZY2Kioq0ng+YsQI7Nmzp8V5JQMHDtRnWoR0aB4eHti6dSu8vb31FrNLly64ePEi+vTpA6lUimPHjsHT0xNXrlyBt7c3amtr9ZZLe7Fz507s379fr60b3NzcEBsbi0mTJmnMmD8tPPneFk9eHc2wtNGgQYPAcRyerffeeecdAFCPcxynky6XhYWF6Ny5MwYMGAAA2L9/P5KTk+Hu7o4lS5ZALBbzHpMQY7By5UpER0dj06ZN6N+/v15i2tvbo66uDn369IGjoyPy8vLg6emJyspK0N+DL8/Ly0tj5lipVOq1dUN1dbV6+7JEIsH9+/cBAFOnToW3tzcVLAaACpY2qqysFCz2zJkzERMTgwEDBqCiogL/+Mc/EBQUhL1796KxsRHr168XLDdChBQSEoLGxkZ4enpCLBZrbFEFdLO+KyAgAAcOHICXlxdCQ0MRGRmJ9PR0nDlzRt3jg/xvQrdroMLT8NFXQnoye/ZsLF26FK+99prW72VlZYXCwkL85S9/wapVq/Dbb7/h6NGjOHnyJCZOnIirV6/ykDEhxmf79u0vvP7hhx/yHlOlUkGlUqnPoElLS0Nubi5cXV0xc+ZMmvE0EtOnT4eDgwNiY2OxceNGREdHw8fHR114bt26VegUOzwqWPTE0tIS586d42UXkaWlJc6ePQtXV1e89dZbeOeddxAREYHq6mr069cP//nPf3jImBBC9OvDDz9EYGAgRo8eDUdHR73GpsLT8FHBoid8bnsOCAiAg4MD3nzzTXz88ce4cOECXFxckJmZiQ8//BBVVVXaJ0yIEaqurn7hdV39EszOzsaWLVtQXl6O9PR09OrVC6mpqXBycsLIkSN1ErM9Gj16NPLz89HU1ASZTAZ/f3/4+/sjICAAPXr0EDo9IjA6rdkIrV+/HoWFhfjkk0+wcOFCuLi4AHjSQ4DOvCAdmUwmg5OT03MfurBv3z6MGTMGEokECoUCDx8+BPDklN/4+HidxGyvMjIycPfuXRw/fhzBwcG4cuUKPv74Y/Tu3Rtubm6YNWsW9u7dq7P42dnZCA4OxvDhw3Ht2jUAQGpqKnJycnQWk7w8mmHRE310wn3w4AFMTExarKgnpKM4f/68xvNHjx5BoVBg7dq1WL58uU4WwXp5eSEyMhIhISEan3OFQoFx48bhxo0bvMfsSB48eIDc3FwcPnwYSUlJaGho0Mnuy3379mHq1KmYMmUKUlNTceHCBTg7OyMxMRGHDh3CoUOHeI9JXg3tEmpHzMzMhE6BEEF5enq2GBs8eDB69uyJNWvW6KRguXz5Mnx9fVuMW1lZ4e7du7zH6yiamppw6tQpZGRk4MSJE8jPz0fPnj0xYcIEncSLi4vD5s2bERISgrS0NPW4j48P4uLidBKTvBoqWIwEteYnpO369euHgoICnby3vb09lEolZDKZxnhOTg6dLfaKsrKyNAoUR0dH+Pn5YcaMGdi5c2eLxpx8osLT8FHBoifBwcGwtLRs87//bGt+6rNCSOvu3bun8ZwxhpqaGixZskRn5wmFhYUhIiIC27ZtA8dxuH79Ok6dOgW5XI7PP/9cJzHbq6e7gxYsWIC0tDTY2dnpLTYVnkaAEa0cPnyYZWdnq58nJiYyT09PNmnSJFZXVydgZoR0PBzHMZFIpPHgOI45Ojqy3Nxc3uKcP3+eNTc3q5/HxcWxrl27Mo7jGMdxzMzMjC1atIi3eB3FggUL2LBhw5hYLGYDBgxgn3zyCUtPT2e3b9/Weez4+Hjm7u7O8vLymIWFBcvOzmY7d+5ktra2LCEhQefxyf9Gi261NGDAAKxatQrjx49HcXExhgwZgqioKJw4cQJubm5ITk7mPaZQWzcJMXSZmZkaz0UiEWxtbeHi4qLur8EHExMT1NTUQCqVwtnZGQUFBbCwsIBSqURDQwPc3d1hbm7OW7yOpqGhAdnZ2cjIyEBGRgYUCgX69u0LPz8/+Pv74+9//zvvMRljiI+Px4oVK9DY2AgAMDU1hVwux7Jly3iPR14dFSxaMjc3R0lJCWQyGZYsWYKSkhKkp6ejsLAQ48eP18kOAZFI9MI1LLpYQU+IMcjKysKIESNaFCePHz9Gbm5uq2sU2sLGxgaHDh3CsGHDIBKJcPPmTdja2vLy3qSluro6rF27Fhs2bOB1l1BRURH69+8Pkei/HT6ampqo8DRQtIZFS2KxWF2NHz9+HCEhIQCeLJL98/fpfFEoFBrP/7x1k5COyt/fXz3z8az6+nr4+/vz9otuwoQJ8PPzQ48ePcBxHAYPHgwTE5NWX1tRUcFLzI5EpVKhoKBAPcNy8uRJNDQ0wNHRkdedXl5eXi1mymxsbODu7s5bDMIfKli0NHLkSERFRcHHxwenT5/G999/DwAoKyvT2Yp2IbZuEmIM2P+fkv5ntbW16Nq1K29xkpKSEBQUBKVSifDwcISFhakXxZO2W716tbpAuX//Pnr16oXRo0dj/fr18Pf35735n7W1NSorKyGVSlFVVQWVSsXr+xN+UcGipcTERMyePRvp6enYtGkTevXqBQA4fPgwxo4dq9dcdLl1kxBD9rRI5zgO06ZNg6mpqfpac3MzioqKeO8C/fTzffbsWURERFDBwoP169dj9OjR+Oc//wl/f391F29doZky40JrWIzQi7ZuXrp0CefOnRMmMUIEEhoaCuDJac0ffPABJBKJ+ppYLIZMJkNYWBgvp6UTwzF79mwsXbpUq5/rkSNH1DNlS5cufW7hGRER0eYYhB9UsPCgvLwcycnJKC8vx1dffQWpVIrDhw/D0dERHh4evMdrbdEtYwwODg7YvXs3nSdEOqwvvvgCcrmc169/iOGytLTEuXPneOmTEhoaioSEBJopM2BUsGgpMzMT48aNg4+PD7KysnDx4kU4Oztj5cqVOHPmDNLT03US81m62rpJCCGGTB9ntBHDQb/dtBQTE4O4uDhERUVpVOYBAQFITEzUSczc3FzY2dnho48+0hjftm0bbt++jQULFugkLiHGID09HXv27EF1dTWampo0rhUWFgqUFSFEW6L//RLyIsXFxXjvvfdajEulUty5c0cnMbds2QI3N7cW4x4eHti8ebNOYhJiDBISEhAaGgo7OzsoFAoMHToUNjY2qKiowLhx44ROjxCiBSpYtGRtbY2ampoW4wqFQr1jiG83btxAjx49Wozb2tq2mgshHcXXX3+NpKQkbNiwAWKxGPPnz8exY8cQHh6O+vp6odMjhGiBChYtTZw4EQsWLMCNGzfAcRxUKhVOnjwJuVyubiLHNwcHB5w8ebLF+MmTJ9GzZ0+dxCTEGFRXV6sXnUskEty/fx8AMHXqVOzevVvI1AghWqKCRUvx8fFwc3ODg4ODupWzr68vRowYgUWLFukkZlhYGD799FMkJyfj999/x++//45t27YhMjISYWFhOolJiDGwt7dHXV0dgCdnauXl5QEAKisrQfsL2p/g4GBYWloKnQbRE9olpAXGGK5evQpbW1vcuXMHxcXFaGhogJeXl86Osn8aNyYmBgkJCepFhWZmZliwYAEWL16ss7iEGLrp06fDwcEBsbGx2LhxI6Kjo+Hj44MzZ84gKCgIW7duFTpF8hKOHDkCc3NzjBw5EgCwceNGfPPNN3B3d8fGjRvRrVs3gTMkQqCCRQsqlQpmZmYoLS3VaYHyPA0NDbh48SIkEglcXV01unsS0hGpVCqoVCr19v60tDTk5ubC1dUVM2fOhFgsFjhD8jIGDBiAVatWYfz48SguLsaQIUMQFRWFEydOwM3NDcnJyUKnSARABYuWPDw8sHXrVnh7ewudCiGEtAvm5uYoKSmBTCbDkiVLUFJSgvT0dBQWFmL8+PG4ceOG0CkSAdAaFi2tXLkS0dHRKCkpEToVQgiA7OxsBAcHY/jw4bh27RoAIDU1FTk5OQJnRl6WWCxGY2MjAOD48eN4++23AQDdu3dvcTQJ6TioYNFSSEgITp8+DU9PT0gkEnTv3l3jQQjRn3379mHMmDGQSCRQKBR4+PAhAKC+vh7x8fECZ0de1siRIxEVFYVly5bh9OnT+Otf/woAKCsrQ+/evQXOjgiFOt1qaf369UKnQAj5f3Fxcdi8eTNCQkKQlpamHvfx8UFcXJyAmZFXkZiYiNmzZyM9PR2bNm1S97Q6fPiw+pRs0vHQGhZCSLvRpUsXXLhwATKZTOOcmYqKCri7u+PBgwdCp0gIaSOaYdFSdXX1C687OjrqKRNCiL29PZRKJWQymcZ4Tk4OHZBnZMrLy5GcnIzy8nJ89dVXkEqlOHz4MBwdHeHh4SF0ekQAtIZFSzKZDE5OTs99EEL0JywsDBEREcjPzwfHcbh+/Tp27doFuVyOWbNmCZ0eeUmZmZkYMGAA8vPz8cMPP6ChoQEAcP78ecTGxgqcHREKzbBoSaFQaDx/9OgRFAoF1q5di+XLlwuUFSEdU0xMDFQqFQIDA9HY2AhfX1+YmppCLpdj7ty5QqdHXlJMTAzi4uIQFRUFCwsL9XhAQAASExMFzIwIidaw6MjPP/+MNWvWICMjQ+hUCGnXioqK0L9/f4hE/50wbmpqglKpVB+XYW5uLmCG5FWZm5ujuLgYTk5OGmuRqqqq4ObmRmuROij6SkhH+vXrh4KCAqHTIKTd8/Lywp07dwAAzs7OqK2thVgshru7O4YOHUrFihGytrZu9eR5hUKh3jFEOh4qWLR07949jUd9fT0uXbqERYsWCdKun5COxtraGpWVlQCAqqoqqFQqgTMi2po4cSIWLFiAGzdugOM4qFQqnDx5EnK5HCEhIUKnRwRCXwlpSSQSgeM4jTHGGBwcHJCWlobhw4cLlBkhHcOMGTOwY8cO9OjRA9XV1ejduzdMTExafW1FRYWesyNt0dTUhDlz5iAlJQXNzc3o1KkTmpubMXnyZKSkpDz350vaNypYtJSZmanxXCQSwdbWFi4uLuoD2AghunXkyBEolUqEh4dj6dKlGgs1nxUREaHnzMirYozh6tWrsLW1xZ07d1BcXIyGhgZ4eXnRrHUHRwWLlrKysjBixIgWxcnjx4+Rm5sLX19fgTIjpOMJDQ1FQkLCcwsWYvhUKhXMzMxQWlpKBQrRQAWLlkxMTFBTUwOpVKoxXltbC6lUiubmZoEyI4QQ4+Th4YGtW7fC29tb6FSIAaFFt1pijLVYwwI8KVi6du0qQEaEEGLcVq5ciejoaJSUlAidCjEgNMPSRkFBQQCA/fv3Y+zYsTA1NVVfa25uRlFREfr164cjR44IlSIhhBilbt26obGxEY8fP4ZYLIZEItG4XldXJ1BmREi0KrSNrKysADyZYbGwsND4QInFYnh7eyMsLEyo9AghxGitX79e6BSIAaIZFi198cUXkMvl9PUPIYQQokNUsBBCCDEo1dXVL7zu6Oiop0yIIaGChQfp6enYs2cPqqur0dTUpHGtsLBQoKwIIcQ4tdaQ81m0+7Jjol1CWkpISEBoaCjs7OygUCgwdOhQ2NjYoKKiAuPGjRM6PUIIMToKhQKFhYXqR35+PjZv3oy+ffti7969QqdHBEIzLFpyc3NDbGwsJk2apHGq6OLFi1FXV0dHoRNCCE9+/vlnrFmzBhkZGUKnQgRAMyxaqq6uxogRIwAAEokE9+/fBwBMnToVu3fvFjI1QghpV/r164eCggKh0yACoYJFS/b29uqeAI6OjsjLywMAVFZWgiavCCHk1d27d0/jUV9fj0uXLmHRokXUrr8Doz4sWgoICMCBAwfg5eWF0NBQREZGIj09HWfOnFE3lyOEEPLyrK2tWyy6ZYzBwcEBaWlpAmVFhEZrWLSkUqmgUqnUhx+mpaUhNzcXrq6umDlzJsRiscAZEkKIccnMzNR4LhKJYGtrCxcXlxYHzZKOgwoWQgghBiUrKwsjRoxoUZw8fvwYubm58PX1FSgzIiRaw8KD7OxsBAcHY/jw4bh27RoAIDU1FTk5OQJnRgghxsff37/V84Lq6+vh7+8vQEbEEFDBoqV9+/ZhzJgxkEgkUCgUePjwIYAnH6z4+HiBsyOEEOPDGGu1cVxtbS0dg9KB0ZeBWoqLi8PmzZsREhKisRjMx8cHcXFxAmZGCCHG5elGBY7jMG3aNJiamqqvNTc3o6ioSN1GgnQ8VLBo6fLly61+n2plZYW7d+/qPyFCCDFSVlZWAJ7MsFhYWEAikaivicVieHt7IywsTKj0iMCoYNGSvb09lEolZDKZxnhOTg6cnZ2FSYoQQoxQcnIyAEAmk0Eul9PXP0QDrWHRUlhYGCIiIpCfnw+O43D9+nXs2rULcrkcs2bNEjo9QggxOrGxsVSskBZohkVLMTExUKlUCAwMRGNjI3x9fWFqagq5XI65c+cKnR4hhBil9PR07NmzB9XV1WhqatK4VlhYKFBWREg0w9IGRUVFUKlUAJ4sDlu4cCHq6upQUlKCvLw83L59G8uWLRM4S0IIMU4JCQkIDQ2FnZ0dFAoFhg4dChsbG1RUVGDcuHFCp0cEQo3j2sDExAQ1NTWQSqVwdnZGQUEBbGxshE6LEELaBTc3N8TGxmLSpEmwsLDA+fPn4ezsjMWLF6Ourg6JiYlCp0gEQDMsbWBtbY3KykoAQFVVlXq2hRBCiPaqq6vV25clEgnu378PAJg6dSp2794tZGpEQLSGpQ0mTJgAPz8/9OjRAxzHYfDgwTAxMWn1tRUVFXrOjhBCjJu9vT3q6urQp08fODo6Ii8vD56enqisrAR9KdBxUcHSBklJSQgKCoJSqUR4eDjCwsJgYWEhdFqEENIuBAQE4MCBA/Dy8kJoaCgiIyORnp6OM2fOqJvLkY6H1rBoKTQ0FAkJCVSwEEIIT1QqFVQqlfrww7S0NOTm5sLV1RUzZ86EWCwWOEMiBCpYCCGEEGLwaNEtIYQQg5OdnY3g4GAMHz4c165dAwCkpqYiJydH4MyIUKhgIYQQYlD27duHMWPGQCKRQKFQ4OHDhwCA+vp6xMfHC5wdEQoVLIQQQgxKXFwcNm/ejG+++QadO3dWj/v4+FCX2w6MChZCCCEG5fLly/D19W0xbmVlhbt37+o/IWIQqGAhhBBiUOzt7aFUKluM5+TkwNnZWYCMiCGggoUQQohBCQsLQ0REBPLz88FxHK5fv45du3ZBLpdj1qxZQqdHBEKN4wghhBiUmJgYqFQqBAYGorGxEb6+vjA1NYVcLsfcuXOFTo8IhPqwEEIIEVxRURH69+8Pkei/E/9NTU1QKpVoaGiAu7s7zM3NBcyQCI0KFkIIIYIzMTFBTU0NpFIpnJ2dUVBQABsbG6HTIgaE1rAQQggRnLW1NSorKwEAVVVVUKlUAmdEDA2tYSGEECK4CRMmwM/PDz169ADHcRg8eDBMTExafW1FRYWesyOGgAoWQgghgktKSkJQUBCUSiXCw8MRFhZGh8oSDbSGhRBCiEEJDQ1FQkICFSxEAxUshBBCCDF4tOiWEEIIIQaPChZCCCGEGDwqWAghhBBi8KhgIYQQQojBo4KFEEIIIQaPChZCCCGEGDwqWAghhBBi8P4Pt/nRNFaHSk8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 226,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "show_feature_importances(find_the_best_model(uk_clf_dict)[1], X_columns)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "id": "26ac0729",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "score:\t  feature:\n",
      "0.797\t- features_Highchairs Available\n",
      "0.020\t- cuisines_Mediterranean\n",
      "0.015\t- features_Takeout\n",
      "0.007\t- meals_Breakfast\n",
      "0.004\t- features_Outdoor Seating\n",
      "0.004\t- meals_Lunch\n",
      "0.003\t- features_Reservations\n",
      "0.003\t- features_Wheelchair Accessible\n",
      "0.002\t- region\n",
      "0.002\t- meals_Dinner\n",
      "0.001\t- cuisines_Pizza\n",
      "0.001\t- price_level\n",
      "0.001\t- cuisines_Seafood\n",
      "0.001\t- cuisines_Italian\n",
      "0.000\t- features_Table Service\n",
      "0.000\t- features_Seating\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 227,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "show_feature_importances(uk_clf_dict['gb']['model'], X_columns)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "faa43a6c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "5ef725f8",
   "metadata": {},
   "source": [
    "<a id=\"german_price_level\"></a>\n",
    "\n",
    "## 2.3. Classification of German dataset by the \"price_level\" feature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "id": "5fefd9f8",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Logistic Regression\n",
    "logreg_grid = {\n",
    "    \"C\": np.logspace(-2, 1, 4), # np.logspace(-3, 3, 7) - from 0.001 to 1000 by 7 steps\n",
    "    \"penalty\": [\"l1\", \"l2\"], # , \"elasticnet\", None\n",
    "    \"multi_class\": [\"multinomial\"], # \"auto\", \"ovr\", \n",
    "    \"max_iter\": [1000],\n",
    "    \"solver\": ['lbfgs', 'newton-cg'] # For multiclass problems, only ‘newton-cg’, ‘sag’, ‘saga’ and ‘lbfgs’ handle multinomial loss;\n",
    "}\n",
    "\n",
    "# SVM\n",
    "svm_grid = {\n",
    "    \"C\": np.logspace(-4, 0, 5), # from 0.01 to 1 by 4 steps\n",
    "    \"penalty\": [\"l1\", \"l2\"],\n",
    "    \"loss\": [\"hinge\", \"squared_hinge\"],\n",
    "    \"max_iter\": [1000]\n",
    "}\n",
    "\n",
    "# Naive Bayes\n",
    "# no grid\n",
    "\n",
    "# k-Nearest Neighbors\n",
    "knn_grid = {\n",
    "    \"n_neighbors\": [6, 8, 10],\n",
    "    \"leaf_size\": [10, 12, 14],\n",
    "    \"algorithm\": ['auto']\n",
    "    ,\"n_jobs\": [-1]\n",
    "}\n",
    "\n",
    "# Decision Tree\n",
    "tree_grid = {\n",
    "    \"min_samples_split\": range(2, 6, 1),\n",
    "    \"max_depth\": range(2, 6, 1),\n",
    "#     \"criterion\": ['squared_error', 'friedman_mse', 'absolute_error', 'poisson'],\n",
    "    \"splitter\": ['best', 'random']\n",
    "}\n",
    "\n",
    "# Random Forest\n",
    "forest_grid = {\n",
    "    \"n_estimators\": [200, 300], # range(100, 400, 100)\n",
    "    \"max_depth\": range(4, 10, 2),\n",
    "    \"min_samples_split\": [15, 20, 25, 30], # range(20, 40, 10)\n",
    "    \"n_jobs\": [-1],\n",
    "    \"random_state\": [42]\n",
    "}\n",
    "\n",
    "# Gradient Boosting\n",
    "gb_grid = {\n",
    "    \"n_estimators\": range(100, 400, 100),\n",
    "    \"max_depth\": range(4, 10, 2),\n",
    "    \"learning_rate\": [0.1, 0.2, 0.3],\n",
    "    \"random_state\": [42]\n",
    "}\n",
    "\n",
    "# XGB\n",
    "xgb_grid = {\n",
    "    \"n_estimators\": range(100, 400, 100),\n",
    "    \"max_depth\": range(4, 10, 2),\n",
    "    \"learning_rate\": [0.1, 0.2, 0.3],\n",
    "    \"n_jobs\": [-1],\n",
    "    \"objective\": ['reg:squarederror'],\n",
    "    \"random_state\": [42]\n",
    "}\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "id": "a363d08e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "region\n",
      "{'Baden-Wurttemberg': 0, 'Bavaria': 1, 'Berlin': 2, 'Brandenburg': 3, 'Hamburg': 4, 'Hesse': 5, 'Lower Saxony': 6, 'Mecklenburg-West Pomerania': 7, 'North Rhine-Westphalia': 8, 'Rhineland-Palatinate': 9, 'Saarland': 10, 'Saxony': 11, 'Saxony-Anhalt': 12, 'Schleswig-Holstein': 13, 'State of Bremen': 14, 'Thuringia': 15}\n",
      "\n",
      "price_level\n",
      "{'€': 0, '€€-€€€': 1, '€€€€': 2}\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py:378: FitFailedWarning: \n",
      "80 fits failed out of a total of 160.\n",
      "The score on these train-test partitions for these parameters will be set to nan.\n",
      "If these failures are not expected, you can try to debug them by setting error_score='raise'.\n",
      "\n",
      "Below are more details about the failures:\n",
      "--------------------------------------------------------------------------------\n",
      "40 fits failed with the following error:\n",
      "Traceback (most recent call last):\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py\", line 686, in _fit_and_score\n",
      "    estimator.fit(X_train, y_train, **fit_params)\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/linear_model/_logistic.py\", line 1162, in fit\n",
      "    solver = _check_solver(self.solver, self.penalty, self.dual)\n",
      "             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/linear_model/_logistic.py\", line 54, in _check_solver\n",
      "    raise ValueError(\n",
      "ValueError: Solver lbfgs supports only 'l2' or 'none' penalties, got l1 penalty.\n",
      "\n",
      "--------------------------------------------------------------------------------\n",
      "40 fits failed with the following error:\n",
      "Traceback (most recent call last):\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py\", line 686, in _fit_and_score\n",
      "    estimator.fit(X_train, y_train, **fit_params)\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/linear_model/_logistic.py\", line 1162, in fit\n",
      "    solver = _check_solver(self.solver, self.penalty, self.dual)\n",
      "             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/linear_model/_logistic.py\", line 54, in _check_solver\n",
      "    raise ValueError(\n",
      "ValueError: Solver newton-cg supports only 'l2' or 'none' penalties, got l1 penalty.\n",
      "\n",
      "  warnings.warn(some_fits_failed_message, FitFailedWarning)\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_search.py:953: UserWarning: One or more of the test scores are non-finite: [       nan        nan 0.74544264 0.74544264        nan        nan\n",
      " 0.74586086 0.74586086        nan        nan 0.7455823  0.74572158\n",
      "        nan        nan 0.74586086 0.74586086]\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: LogisticRegression\n",
      "Tuned hyperparameters:\n",
      "\tC: 0.1\n",
      "\tmax_iter: 1000\n",
      "\tmulti_class: multinomial\n",
      "\tpenalty: l2\n",
      "\tsolver: lbfgs\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 269  633    0]\n",
      " [ 119 1994    0]\n",
      " [   1   62    0]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.69      0.30      0.42       902\n",
      "           1       0.74      0.94      0.83      2113\n",
      "           2       0.00      0.00      0.00        63\n",
      "\n",
      "    accuracy                           0.74      3078\n",
      "   macro avg       0.48      0.41      0.42      3078\n",
      "weighted avg       0.71      0.74      0.69      3078\n",
      "\n",
      "Time using clf_model_implementation(): 12.586652994155884 sec\n",
      "\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py:1244: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py:378: FitFailedWarning: \n",
      "100 fits failed out of a total of 200.\n",
      "The score on these train-test partitions for these parameters will be set to nan.\n",
      "If these failures are not expected, you can try to debug them by setting error_score='raise'.\n",
      "\n",
      "Below are more details about the failures:\n",
      "--------------------------------------------------------------------------------\n",
      "50 fits failed with the following error:\n",
      "Traceback (most recent call last):\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py\", line 686, in _fit_and_score\n",
      "    estimator.fit(X_train, y_train, **fit_params)\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_classes.py\", line 274, in fit\n",
      "    self.coef_, self.intercept_, n_iter_ = _fit_liblinear(\n",
      "                                           ^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py\", line 1223, in _fit_liblinear\n",
      "    solver_type = _get_liblinear_solver_type(multi_class, penalty, loss, dual)\n",
      "                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py\", line 1062, in _get_liblinear_solver_type\n",
      "    raise ValueError(\n",
      "ValueError: Unsupported set of arguments: The combination of penalty='l1' and loss='hinge' is not supported, Parameters: penalty='l1', loss='hinge', dual=True\n",
      "\n",
      "--------------------------------------------------------------------------------\n",
      "50 fits failed with the following error:\n",
      "Traceback (most recent call last):\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_validation.py\", line 686, in _fit_and_score\n",
      "    estimator.fit(X_train, y_train, **fit_params)\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_classes.py\", line 274, in fit\n",
      "    self.coef_, self.intercept_, n_iter_ = _fit_liblinear(\n",
      "                                           ^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py\", line 1223, in _fit_liblinear\n",
      "    solver_type = _get_liblinear_solver_type(multi_class, penalty, loss, dual)\n",
      "                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n",
      "  File \"/opt/homebrew/lib/python3.11/site-packages/sklearn/svm/_base.py\", line 1062, in _get_liblinear_solver_type\n",
      "    raise ValueError(\n",
      "ValueError: Unsupported set of arguments: The combination of penalty='l1' and loss='squared_hinge' are not supported when dual=True, Parameters: penalty='l1', loss='squared_hinge', dual=True\n",
      "\n",
      "  warnings.warn(some_fits_failed_message, FitFailedWarning)\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/model_selection/_search.py:953: UserWarning: One or more of the test scores are non-finite: [       nan 0.69293975        nan 0.6943325         nan 0.69293975\n",
      "        nan 0.74279582        nan 0.74572139        nan 0.74627868\n",
      "        nan 0.7409864         nan 0.7450254         nan 0.7409864\n",
      "        nan 0.7450254 ]\n",
      "  warnings.warn(\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: LinearSVC\n",
      "Tuned hyperparameters:\n",
      "\tC: 0.01\n",
      "\tloss: squared_hinge\n",
      "\tmax_iter: 1000\n",
      "\tpenalty: l2\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 266  636    0]\n",
      " [ 119 1994    0]\n",
      " [   1   62    0]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.69      0.29      0.41       902\n",
      "           1       0.74      0.94      0.83      2113\n",
      "           2       0.00      0.00      0.00        63\n",
      "\n",
      "    accuracy                           0.73      3078\n",
      "   macro avg       0.48      0.41      0.41      3078\n",
      "weighted avg       0.71      0.73      0.69      3078\n",
      "\n",
      "Time using clf_model_implementation(): 11.901381969451904 sec\n",
      "\n",
      "\n",
      "Model: GaussianNB\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 471  405   26]\n",
      " [ 485 1434  194]\n",
      " [   7   33   23]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.49      0.52      0.51       902\n",
      "           1       0.77      0.68      0.72      2113\n",
      "           2       0.09      0.37      0.15        63\n",
      "\n",
      "    accuracy                           0.63      3078\n",
      "   macro avg       0.45      0.52      0.46      3078\n",
      "weighted avg       0.67      0.63      0.65      3078\n",
      "\n",
      "Time using clf_model_implementation(): 0.027211904525756836 sec\n",
      "\n",
      "\n",
      "Model: KNeighborsClassifier\n",
      "Tuned hyperparameters:\n",
      "\talgorithm: auto\n",
      "\tleaf_size: 10\n",
      "\tn_jobs: -1\n",
      "\tn_neighbors: 10\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 321  581    0]\n",
      " [ 231 1880    2]\n",
      " [   1   60    2]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.58      0.36      0.44       902\n",
      "           1       0.75      0.89      0.81      2113\n",
      "           2       0.50      0.03      0.06        63\n",
      "\n",
      "    accuracy                           0.72      3078\n",
      "   macro avg       0.61      0.43      0.44      3078\n",
      "weighted avg       0.69      0.72      0.69      3078\n",
      "\n",
      "Time using clf_model_implementation(): 11.276870012283325 sec\n",
      "\n",
      "\n",
      "Model: DecisionTreeClassifier\n",
      "Tuned hyperparameters:\n",
      "\tmax_depth: 4\n",
      "\tmin_samples_split: 2\n",
      "\tsplitter: random\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 280  622    0]\n",
      " [ 143 1970    0]\n",
      " [   1   62    0]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.66      0.31      0.42       902\n",
      "           1       0.74      0.93      0.83      2113\n",
      "           2       0.00      0.00      0.00        63\n",
      "\n",
      "    accuracy                           0.73      3078\n",
      "   macro avg       0.47      0.41      0.42      3078\n",
      "weighted avg       0.70      0.73      0.69      3078\n",
      "\n",
      "Time using clf_model_implementation(): 1.0503621101379395 sec\n",
      "\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: RandomForestClassifier\n",
      "Tuned hyperparameters:\n",
      "\tmax_depth: 8\n",
      "\tmin_samples_split: 20\n",
      "\tn_estimators: 200\n",
      "\tn_jobs: -1\n",
      "\trandom_state: 42\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 249  653    0]\n",
      " [  96 2017    0]\n",
      " [   1   62    0]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.72      0.28      0.40       902\n",
      "           1       0.74      0.95      0.83      2113\n",
      "           2       0.00      0.00      0.00        63\n",
      "\n",
      "    accuracy                           0.74      3078\n",
      "   macro avg       0.49      0.41      0.41      3078\n",
      "weighted avg       0.72      0.74      0.69      3078\n",
      "\n",
      "Time using clf_model_implementation(): 39.63809680938721 sec\n",
      "\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: GradientBoostingClassifier\n",
      "Tuned hyperparameters:\n",
      "\tlearning_rate: 0.1\n",
      "\tmax_depth: 4\n",
      "\tn_estimators: 100\n",
      "\trandom_state: 42\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 291  611    0]\n",
      " [ 122 1989    2]\n",
      " [   1   60    2]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.70      0.32      0.44       902\n",
      "           1       0.75      0.94      0.83      2113\n",
      "           2       0.50      0.03      0.06        63\n",
      "\n",
      "    accuracy                           0.74      3078\n",
      "   macro avg       0.65      0.43      0.45      3078\n",
      "weighted avg       0.73      0.74      0.70      3078\n",
      "\n",
      "Time using clf_model_implementation(): 880.3389971256256 sec\n",
      "\n",
      "\n",
      "Model: XGBClassifier\n",
      "Tuned hyperparameters:\n",
      "\tlearning_rate: 0.1\n",
      "\tmax_depth: 4\n",
      "\tn_estimators: 100\n",
      "\tn_jobs: -1\n",
      "\tobjective: reg:squarederror\n",
      "\trandom_state: 42\n",
      "\n",
      "Confusion matrix:\n",
      " [[ 271  631    0]\n",
      " [ 109 2004    0]\n",
      " [   1   62    0]]\n",
      "\n",
      "Classification report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.71      0.30      0.42       902\n",
      "           1       0.74      0.95      0.83      2113\n",
      "           2       0.00      0.00      0.00        63\n",
      "\n",
      "    accuracy                           0.74      3078\n",
      "   macro avg       0.48      0.42      0.42      3078\n",
      "weighted avg       0.72      0.74      0.70      3078\n",
      "\n",
      "Time using clf_model_implementation(): 361.9502079486847 sec\n",
      "\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/opt/homebrew/lib/python3.11/site-packages/sklearn/metrics/_classification.py:1344: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'logreg': {'model': LogisticRegression(C=0.1, max_iter=1000, multi_class='multinomial'),\n",
       "  'accuracy_score': 0.735217673814165},\n",
       " 'svm': {'model': LinearSVC(C=0.01), 'accuracy_score': 0.7342430149447693},\n",
       " 'nb': {'model': GaussianNB(), 'accuracy_score': 0.6263807667316439},\n",
       " 'knn': {'model': KNeighborsClassifier(leaf_size=10, n_jobs=-1, n_neighbors=10),\n",
       "  'accuracy_score': 0.7157244964262508},\n",
       " 'tree': {'model': DecisionTreeClassifier(max_depth=4, splitter='random'),\n",
       "  'accuracy_score': 0.7309941520467836},\n",
       " 'forest': {'model': RandomForestClassifier(max_depth=8, min_samples_split=20, n_estimators=200,\n",
       "                         n_jobs=-1, random_state=42),\n",
       "  'accuracy_score': 0.7361923326835608},\n",
       " 'gb': {'model': GradientBoostingClassifier(max_depth=4, random_state=42),\n",
       "  'accuracy_score': 0.7413905133203379},\n",
       " 'xgb': {'model': XGBClassifier(base_score=None, booster=None, callbacks=None,\n",
       "                colsample_bylevel=None, colsample_bynode=None,\n",
       "                colsample_bytree=None, early_stopping_rounds=None,\n",
       "                enable_categorical=False, eval_metric=None, feature_types=None,\n",
       "                gamma=None, gpu_id=None, grow_policy=None, importance_type=None,\n",
       "                interaction_constraints=None, learning_rate=0.1, max_bin=None,\n",
       "                max_cat_threshold=None, max_cat_to_onehot=None,\n",
       "                max_delta_step=None, max_depth=4, max_leaves=None,\n",
       "                min_child_weight=None, missing=nan, monotone_constraints=None,\n",
       "                n_estimators=100, n_jobs=-1, num_parallel_tree=None,\n",
       "                objective='multi:softprob', predictor=None, ...),\n",
       "  'accuracy_score': 0.7391163092917479}}"
      ]
     },
     "execution_count": 230,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# PIVOT DICTIONARY OF CLASSIFICATION MODELS\n",
    "clf_models_dict = {\n",
    "#     simple models\n",
    "    'logreg': {'status': True, 'model': LogisticRegression, 'grid': logreg_grid},\n",
    "    'svm': {'status': True, 'model': LinearSVC, 'grid': svm_grid},\n",
    "    'nb': {'status': True, 'model': GaussianNB, 'grid': {}},\n",
    "    'knn': {'status': True, 'model': KNeighborsClassifier, 'grid': knn_grid},\n",
    "    'tree': {'status': True, 'model': DecisionTreeClassifier, 'grid': tree_grid},\n",
    "#     ensemble models (not based on other models)\n",
    "    'forest': {'status': True, 'model': RandomForestClassifier, 'grid': forest_grid},\n",
    "    'gb': {'status': True, 'model': GradientBoostingClassifier, 'grid': gb_grid},\n",
    "    'xgb': {'status': True, 'model': XGBClassifier, 'grid': xgb_grid}\n",
    "}\n",
    "\n",
    "german_clf_dict = run_classification_models(german_df2, 'price_level', [], clf_models_dict)\n",
    "# drop_columns = ['special_diets_Gluten Free Options', 'special_diets_Vegan Options']\n",
    "german_clf_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 231,
   "id": "641093ce",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.7413905133203379, GradientBoostingClassifier(max_depth=4, random_state=42))"
      ]
     },
     "execution_count": 231,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "find_the_best_model(german_clf_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 232,
   "id": "7970e63f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "score:\t  feature:\n",
      "0.316\t- cuisines_Italian\n",
      "0.145\t- features_Reservations\n",
      "0.107\t- cuisines_Seafood\n",
      "0.081\t- region\n",
      "0.053\t- meals_Breakfast\n",
      "0.049\t- features_Takeout\n",
      "0.035\t- features_Table Service\n",
      "0.035\t- features_Seating\n",
      "0.030\t- features_Wheelchair Accessible\n",
      "0.030\t- meals_Dinner\n",
      "0.019\t- features_Outdoor Seating\n",
      "0.019\t- features_Highchairs Available\n",
      "0.019\t- meals_Lunch\n",
      "0.018\t- cuisines_Mediterranean\n",
      "0.015\t- cuisines_Pizza\n",
      "0.015\t- price_level\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 232,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "show_feature_importances(find_the_best_model(german_clf_dict)[1], X_columns)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7f0c91a7",
   "metadata": {},
   "source": [
    "<a id=\"conclusion\"></a>\n",
    "\n",
    "# 3. Conclusions"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "033038af",
   "metadata": {},
   "source": [
    "<a id=\"conclusion_italian_serves_alcohol\"></a>\n",
    "\n",
    "## 3.1. For Classification of Italian dataset by the \"Serves Alcohol\" feature"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f46b0654",
   "metadata": {},
   "source": [
    "The **XGBoost model** has the best accuracy score - **77%**. The features **Highchairs Available** and **Seating** have the biggest importance for classification with XGBoost. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8910c58e",
   "metadata": {},
   "source": [
    "<a id=\"conclusion_UK_vegetarian\"></a>\n",
    "\n",
    "## 3.2. For Classification of the UK dataset by the \"Vegetarian Friendly\" feature"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "96032e67",
   "metadata": {},
   "source": [
    "The **Random Forest Classifier** has the best accuracy score - **95%**. The features **Highchairs Available** has the biggest importance for classification with Random Forest model. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "11f7b8f9",
   "metadata": {},
   "source": [
    "<a id=\"conclusion_german_price_level\"></a>\n",
    "\n",
    "## 3.3. For Classification of German dataset by the \"price_level\" feature"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e5d82fb4",
   "metadata": {},
   "source": [
    "All models show the same:\n",
    "\n",
    "- it's impossible to recognize class 2 (High)\n",
    "- there are a lot of following mistakes: restaurants predicted as class 1 (Middle), but actually are class 0 (Cheap)\n",
    "\n",
    "The **Gradient Boosting Classifier** has the best accuracy score - **74%**. "
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}