06_descriptives.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "44b56386-9bd6-40db-8e61-0cebc1fd983b",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Modify the file 00_setup.py to define input/output file paths on your system\n",
    "# The information in 00_setup.py will be used across notebooks\n",
    "from importlib.machinery import SourceFileLoader\n",
    "setup = SourceFileLoader(\"setup\", \"./00_setup.py\").load_module()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5975abcf-8c2a-4ad4-9531-79799842931f",
   "metadata": {},
   "source": [
    "# 06: Base Model Descriptives\n",
    " * Compare the results for Friedman's H, ALE, and Shapley.  \n",
    " * Get some descriptive plots relating to term and interest rate"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f2f60612-7830-4112-b1d3-1617d123de4d",
   "metadata": {},
   "source": [
    "*This script takes about 1 minutes on my MacBook Air*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "61ff97aa-210f-4817-9e6b-44934d50ad6a",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import pickle\n",
    "from pathlib import Path\n",
    "import sklearn as sk\n",
    "import matplotlib as mpl\n",
    "import scipy\n",
    "import random\n",
    "from matplotlib.ticker import FormatStrFormatter, PercentFormatter\n",
    "from sklearn.inspection import partial_dependence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "d0ced6de-78b1-49fb-9108-79df75db9f51",
   "metadata": {},
   "outputs": [],
   "source": [
    "from interactions_package.interactions_package import ip_plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "6830d40c-3717-46e9-a7f6-e53ac0fb1dbd",
   "metadata": {},
   "outputs": [],
   "source": [
    "ip_plot.plot_defaults()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2d3dbffa-7110-4c13-bd9b-5148bbb48798",
   "metadata": {},
   "source": [
    "## Import Data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "edd1ff00-2636-4b05-b251-3c5dff94e39e",
   "metadata": {},
   "source": [
    "##### Predictors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "6fa544ef-5b04-45a2-a4e6-626769a8d8cc",
   "metadata": {},
   "outputs": [],
   "source": [
    "predictor_features = pd.read_csv(Path(setup.temp_path) \\\n",
    "                                 .joinpath('01_REPORT_predictor_features.csv'))['feature'].tolist()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fd64c3e4-2d2d-4787-b72c-92ee456fa038",
   "metadata": {},
   "source": [
    "##### Dataset - train, test, and sampled train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5ab17675-38f7-4f82-9507-54b494679a5c",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_df = pd.read_parquet(Path(setup.temp_path).joinpath('01_DATA_combined_base_test.parquet'))\n",
    "train_df = pd.read_parquet(Path(setup.temp_path).joinpath('01_DATA_combined_base_train.parquet'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "a4d5b29c-9191-4060-874f-67015a559a39",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_samp_df = pd.read_parquet(Path(setup.temp_path) \\\n",
    "                                .joinpath('03_DATA_combined_base_train_samp.parquet')) "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "700b4c25-1325-4904-baf1-aa364ba650c2",
   "metadata": {},
   "source": [
    "##### Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "a5d41d78-cb6f-4e02-82bf-c8419a3f360c",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(Path(setup.temp_path).joinpath('02_MODEL_base_gc.pkl'), 'rb') as pi:\n",
    "    base_model = pickle.load(pi)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4d31c639-2181-4955-929b-befa334441f3",
   "metadata": {},
   "source": [
    "##### Permutation importances"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "d2800ff4-c1a2-4e22-9095-77a5a46065b1",
   "metadata": {},
   "outputs": [],
   "source": [
    "perm_importance_df = pd.read_csv(Path(setup.temp_path).joinpath('02_DATA_importance_perturb.csv'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "248436b0-1d8a-4809-adaa-47054d3579b9",
   "metadata": {},
   "source": [
    "##### Friedman's H Scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "f30c9b37-056c-4be3-962c-8de07fbaa4fd",
   "metadata": {},
   "outputs": [],
   "source": [
    "h_comb = pd.read_csv(Path(setup.temp_path).joinpath('03_DATA_h_scores.csv'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dbc26aa7-7cd1-4cfa-9024-4a9fd6051474",
   "metadata": {},
   "source": [
    "##### 2-Way ALE Mean Abs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "4153d13a-5392-4000-884b-1a5b309b28e6",
   "metadata": {},
   "outputs": [],
   "source": [
    "ale_mean_abs = pd.read_csv(Path(setup.temp_path).joinpath('04_DATA_ale_magnitudes_2way.csv'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d827ef6e-9d6f-4504-b52d-22d34d0a9879",
   "metadata": {},
   "source": [
    "##### Shapley Interaction Mean Abs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "0699bb7b-e148-4998-814f-3597b3788a4c",
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_mean_abs = pd.read_csv(Path(setup.temp_path).joinpath('05_DATA_shap_mean_abs.csv'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8b436add-a13b-4c51-a853-022bec783792",
   "metadata": {},
   "source": [
    "## Compare Ranks for the Three Methods"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9b106a9f-bda0-47f8-823d-0581fa3377bc",
   "metadata": {},
   "source": [
    "##### Align ranks and save table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "929c8900-fb89-4c0d-bc57-fac3ed45f00e",
   "metadata": {},
   "outputs": [],
   "source": [
    "h_comb['rank'] = h_comb['h_score'].rank(method='first', ascending=False) \n",
    "h_comb = h_comb.rename(columns={'h_score':'score'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "21666ec0-1374-40b1-9ae8-d16dc83032cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "ale_mean_abs['rank'] = ale_mean_abs['abs_mean'].rank(method='first', ascending=False)\n",
    "ale_mean_abs = ale_mean_abs.rename(columns={'abs_mean':'score'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "33e9bce8-24df-4117-a3ed-dd72e9c0de56",
   "metadata": {},
   "outputs": [],
   "source": [
    "shap_mean_abs_int = shap_mean_abs[shap_mean_abs['feat_1'] != shap_mean_abs['feat_2']] \\\n",
    "    .copy() \\\n",
    "    .rename(columns={'shap_mean_abs':'score'})\n",
    "shap_mean_abs_int['rank'] = shap_mean_abs_int['score'].rank(method='first', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "bf6bb44b-26eb-4935-8b48-6f94bf03c50a",
   "metadata": {},
   "outputs": [],
   "source": [
    "rank_comb = h_comb[['rank', 'score', 'label']] \\\n",
    "    .merge(ale_mean_abs[['rank', 'score', 'label']], how='outer', on='rank',\n",
    "           suffixes=('_h', '_ale')) \\\n",
    "    .merge(shap_mean_abs_int[['rank', 'score', 'label']], how='outer', on='rank') \\\n",
    "    .rename(columns={'score':'score_shap', 'label':'label_shap'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "1fe84f1b-06a8-4b0a-9347-b89ba182fe59",
   "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>rank</th>\n",
       "      <th>score_h</th>\n",
       "      <th>label_h</th>\n",
       "      <th>score_ale</th>\n",
       "      <th>label_ale</th>\n",
       "      <th>score_shap</th>\n",
       "      <th>label_shap</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.159343</td>\n",
       "      <td>loan_amnt:verification_status</td>\n",
       "      <td>0.004654</td>\n",
       "      <td>annual_inc:sec_app_fico_range_low</td>\n",
       "      <td>0.029502</td>\n",
       "      <td>int_rate:term</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2.0</td>\n",
       "      <td>0.146755</td>\n",
       "      <td>loan_amnt:revol_bal</td>\n",
       "      <td>0.004026</td>\n",
       "      <td>int_rate:home_ownership__RENT</td>\n",
       "      <td>0.021683</td>\n",
       "      <td>int_rate:sec_app_fico_range_low</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3.0</td>\n",
       "      <td>0.136924</td>\n",
       "      <td>term:home_ownership__RENT</td>\n",
       "      <td>0.003251</td>\n",
       "      <td>int_rate:term</td>\n",
       "      <td>0.014491</td>\n",
       "      <td>int_rate:loan_amnt</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4.0</td>\n",
       "      <td>0.134237</td>\n",
       "      <td>loan_amnt:annual_inc</td>\n",
       "      <td>0.003209</td>\n",
       "      <td>annual_inc:fico_range_low</td>\n",
       "      <td>0.011650</td>\n",
       "      <td>int_rate:mort_acc</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5.0</td>\n",
       "      <td>0.127820</td>\n",
       "      <td>term:sec_app_fico_range_low</td>\n",
       "      <td>0.002340</td>\n",
       "      <td>int_rate:annual_inc</td>\n",
       "      <td>0.010035</td>\n",
       "      <td>int_rate:annual_inc</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   rank   score_h                        label_h  score_ale  \\\n",
       "0   1.0  0.159343  loan_amnt:verification_status   0.004654   \n",
       "1   2.0  0.146755            loan_amnt:revol_bal   0.004026   \n",
       "2   3.0  0.136924      term:home_ownership__RENT   0.003251   \n",
       "3   4.0  0.134237           loan_amnt:annual_inc   0.003209   \n",
       "4   5.0  0.127820    term:sec_app_fico_range_low   0.002340   \n",
       "\n",
       "                           label_ale  score_shap  \\\n",
       "0  annual_inc:sec_app_fico_range_low    0.029502   \n",
       "1      int_rate:home_ownership__RENT    0.021683   \n",
       "2                      int_rate:term    0.014491   \n",
       "3          annual_inc:fico_range_low    0.011650   \n",
       "4                int_rate:annual_inc    0.010035   \n",
       "\n",
       "                        label_shap  \n",
       "0                    int_rate:term  \n",
       "1  int_rate:sec_app_fico_range_low  \n",
       "2               int_rate:loan_amnt  \n",
       "3                int_rate:mort_acc  \n",
       "4              int_rate:annual_inc  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rank_comb.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "39ea7740-0c13-4ab8-9a55-dedb52faf83d",
   "metadata": {},
   "outputs": [],
   "source": [
    "rank_comb.to_csv(Path(setup.temp_path).joinpath('06_REPORT_top_interactions_all.csv'),\n",
    "                   index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9e60ed6e-5ae0-42a1-a8b2-00ac2e40248f",
   "metadata": {},
   "source": [
    "##### Correlations among ranks and with perturbation importances"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "6b1c6cfa-a561-4209-8b5f-57e6aeec6882",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Merge ranked data by feature"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "0d868cdb-540e-4f44-9135-5d6fd8944cc3",
   "metadata": {},
   "outputs": [],
   "source": [
    "comb_score_by_feature = h_comb[['rank', 'score', 'label', 'feat_1', 'feat_2']] \\\n",
    "    .merge(ale_mean_abs[['rank', 'score', 'feat_1', 'feat_2']], how='outer', \n",
    "           on=['feat_1', 'feat_2'],suffixes=('_h', '_ale')) \\\n",
    "    .merge(shap_mean_abs_int[['rank', 'score', 'feat_1', 'feat_2']], how='outer', \n",
    "           on=['feat_1', 'feat_2']) \\\n",
    "    .rename(columns={'score':'score_shap', 'rank':'rank_shap'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "2c079200-fcc7-4ad0-b39b-f6c347fd16bf",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Append permutation feature importances\n",
    "# We have 1 permutation importance per feature, so to compare the interaction measures, look at both component features.\n",
    "comb_score_by_feature = comb_score_by_feature.merge(perm_importance_df[['features', 'abs_importance']] \\\n",
    "                                      .rename(columns={'abs_importance':'pi_1_a'}),\n",
    "                                  how='left', left_on='feat_1', right_on='features') \\\n",
    "    .drop(columns=['features']) \\\n",
    "    .merge(perm_importance_df[['features', 'abs_importance']] \\\n",
    "                                      .rename(columns={'abs_importance':'pi_2_a'}),\n",
    "                                  how='left', left_on='feat_2', right_on='features') \\\n",
    "    .drop(columns=['features']) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "3276971f-5af8-465e-8388-efe4b8fbdc2e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Reorder permutation importances by strength\n",
    "comb_score_by_feature['pi_1'] = np.where(comb_score_by_feature['pi_1_a'] >=  comb_score_by_feature['pi_2_a'],\n",
    "                                   comb_score_by_feature['pi_1_a'], comb_score_by_feature['pi_2_a'])\n",
    "comb_score_by_feature['pi_2'] = np.where(comb_score_by_feature['pi_1_a'] >=  comb_score_by_feature['pi_2_a'],\n",
    "                                   comb_score_by_feature['pi_2_a'], comb_score_by_feature['pi_1_a'])\n",
    "comb_score_by_feature.drop(columns=['pi_1_a', 'pi_2_a'], inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "b964c714-ce05-4bb1-986c-546401012c37",
   "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>rank_h</th>\n",
       "      <th>score_h</th>\n",
       "      <th>label</th>\n",
       "      <th>feat_1</th>\n",
       "      <th>feat_2</th>\n",
       "      <th>rank_ale</th>\n",
       "      <th>score_ale</th>\n",
       "      <th>rank_shap</th>\n",
       "      <th>score_shap</th>\n",
       "      <th>pi_1</th>\n",
       "      <th>pi_2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>16.0</td>\n",
       "      <td>0.066456</td>\n",
       "      <td>annual_inc:sec_app_fico_range_low</td>\n",
       "      <td>annual_inc</td>\n",
       "      <td>sec_app_fico_range_low</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.004654</td>\n",
       "      <td>38.0</td>\n",
       "      <td>0.002256</td>\n",
       "      <td>0.000243</td>\n",
       "      <td>0.000137</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>29.0</td>\n",
       "      <td>0.033647</td>\n",
       "      <td>int_rate:home_ownership__RENT</td>\n",
       "      <td>int_rate</td>\n",
       "      <td>home_ownership__RENT</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.004026</td>\n",
       "      <td>12.0</td>\n",
       "      <td>0.006824</td>\n",
       "      <td>0.001057</td>\n",
       "      <td>0.000478</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>7.0</td>\n",
       "      <td>0.101831</td>\n",
       "      <td>int_rate:term</td>\n",
       "      <td>int_rate</td>\n",
       "      <td>term</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.003251</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.029502</td>\n",
       "      <td>0.001057</td>\n",
       "      <td>0.000776</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>34.0</td>\n",
       "      <td>0.025367</td>\n",
       "      <td>annual_inc:fico_range_low</td>\n",
       "      <td>annual_inc</td>\n",
       "      <td>fico_range_low</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.003209</td>\n",
       "      <td>73.0</td>\n",
       "      <td>0.000890</td>\n",
       "      <td>0.000243</td>\n",
       "      <td>0.000150</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>24.0</td>\n",
       "      <td>0.043082</td>\n",
       "      <td>int_rate:annual_inc</td>\n",
       "      <td>int_rate</td>\n",
       "      <td>annual_inc</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0.002340</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0.010035</td>\n",
       "      <td>0.001057</td>\n",
       "      <td>0.000243</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    rank_h   score_h                              label      feat_1  \\\n",
       "15    16.0  0.066456  annual_inc:sec_app_fico_range_low  annual_inc   \n",
       "28    29.0  0.033647      int_rate:home_ownership__RENT    int_rate   \n",
       "6      7.0  0.101831                      int_rate:term    int_rate   \n",
       "33    34.0  0.025367          annual_inc:fico_range_low  annual_inc   \n",
       "23    24.0  0.043082                int_rate:annual_inc    int_rate   \n",
       "\n",
       "                    feat_2  rank_ale  score_ale  rank_shap  score_shap  \\\n",
       "15  sec_app_fico_range_low       1.0   0.004654       38.0    0.002256   \n",
       "28    home_ownership__RENT       2.0   0.004026       12.0    0.006824   \n",
       "6                     term       3.0   0.003251        1.0    0.029502   \n",
       "33          fico_range_low       4.0   0.003209       73.0    0.000890   \n",
       "23              annual_inc       5.0   0.002340        5.0    0.010035   \n",
       "\n",
       "        pi_1      pi_2  \n",
       "15  0.000243  0.000137  \n",
       "28  0.001057  0.000478  \n",
       "6   0.001057  0.000776  \n",
       "33  0.000243  0.000150  \n",
       "23  0.001057  0.000243  "
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "comb_score_by_feature.sort_values('rank_ale').head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "55a2d1d7-dc60-47d8-9c5e-abb932b977f7",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Correlations\n",
    "comb_score_corr  = comb_score_by_feature[['score_h', 'score_ale', 'score_shap', 'pi_1', 'pi_2']].corr(method='spearman')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "6e11eb0b-1235-4aa4-9125-114ebd2c5500",
   "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>score_h</th>\n",
       "      <th>score_ale</th>\n",
       "      <th>score_shap</th>\n",
       "      <th>pi_1</th>\n",
       "      <th>pi_2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>score_h</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.195784</td>\n",
       "      <td>0.687088</td>\n",
       "      <td>0.179698</td>\n",
       "      <td>0.261687</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>score_ale</th>\n",
       "      <td>0.195784</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.295652</td>\n",
       "      <td>0.274342</td>\n",
       "      <td>0.408549</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>score_shap</th>\n",
       "      <td>0.687088</td>\n",
       "      <td>0.295652</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.514559</td>\n",
       "      <td>0.663469</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>pi_1</th>\n",
       "      <td>0.179698</td>\n",
       "      <td>0.274342</td>\n",
       "      <td>0.514559</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.467320</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>pi_2</th>\n",
       "      <td>0.261687</td>\n",
       "      <td>0.408549</td>\n",
       "      <td>0.663469</td>\n",
       "      <td>0.467320</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             score_h  score_ale  score_shap      pi_1      pi_2\n",
       "score_h     1.000000   0.195784    0.687088  0.179698  0.261687\n",
       "score_ale   0.195784   1.000000    0.295652  0.274342  0.408549\n",
       "score_shap  0.687088   0.295652    1.000000  0.514559  0.663469\n",
       "pi_1        0.179698   0.274342    0.514559  1.000000  0.467320\n",
       "pi_2        0.261687   0.408549    0.663469  0.467320  1.000000"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "comb_score_corr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "d4f08c0c-145b-4360-af11-da5b216484aa",
   "metadata": {},
   "outputs": [],
   "source": [
    "comb_score_corr.to_csv(Path(setup.temp_path).joinpath('06_REPORT_interactions_corr.csv'),\n",
    "                   index=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6b9c9844-4641-4ff2-9f4d-f67f498087c7",
   "metadata": {},
   "source": [
    "## Term and Interest Rate\n",
    "This is an interaction which shows opposite yet similar effects with SHAP vs. ALE.  Get more information about these features and how the model responds to them."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "80203de4-90e1-4fad-a847-a5e342207057",
   "metadata": {},
   "source": [
    "### Correlations between term and other features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "fd49834d-35db-4d28-bb4b-a4ff72901e9a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    100000.000000\n",
       "mean      14398.332000\n",
       "std        8736.021155\n",
       "min        1000.000000\n",
       "25%        7875.000000\n",
       "50%       12000.000000\n",
       "75%       20000.000000\n",
       "max       40000.000000\n",
       "Name: loan_amnt, dtype: float64"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_samp_df['loan_amnt'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "291fa1b7-b513-4aa6-87b2-bd5414e753ef",
   "metadata": {},
   "outputs": [],
   "source": [
    "term_corr = pd.DataFrame({'feature': predictor_features,\n",
    "                          'corr': [train_samp_df[['term', f]] \\\n",
    "                                   .corr(method='pearson').iloc[0,1] for f in predictor_features],\n",
    "                         'corr_spearman':[train_samp_df[['term', f]] \\\n",
    "                                   .corr(method='spearman').iloc[0,1] for f in predictor_features]})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "6e8a8af4-8ee9-497d-9530-a41deb848547",
   "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>feature</th>\n",
       "      <th>corr</th>\n",
       "      <th>corr_spearman</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>term</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>int_rate</td>\n",
       "      <td>0.413934</td>\n",
       "      <td>0.398523</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>loan_amnt</td>\n",
       "      <td>0.383568</td>\n",
       "      <td>0.413985</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>verification_status</td>\n",
       "      <td>0.125375</td>\n",
       "      <td>0.125375</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>home_ownership__MORTGAGE</td>\n",
       "      <td>0.102143</td>\n",
       "      <td>0.102143</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     feature      corr  corr_spearman\n",
       "1                       term  1.000000       1.000000\n",
       "2                   int_rate  0.413934       0.398523\n",
       "0                  loan_amnt  0.383568       0.413985\n",
       "5        verification_status  0.125375       0.125375\n",
       "14  home_ownership__MORTGAGE  0.102143       0.102143"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "term_corr.sort_values('corr', ascending=False).head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b8521c7b-c804-45bf-b770-3ffcd22f21c8",
   "metadata": {},
   "source": [
    "### Correlations between interest rate and other features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "bb0219e1-0e7d-45aa-97ee-45fb6aaa80e8",
   "metadata": {},
   "outputs": [],
   "source": [
    "int_rate_corr = pd.DataFrame({'feature': predictor_features,\n",
    "                          'corr': [train_samp_df[['int_rate', f]] \\\n",
    "                                   .corr().iloc[0,1] for f in predictor_features],\n",
    "                         'corr_spearman':[train_samp_df[['int_rate', f]] \\\n",
    "                                   .corr(method='spearman').iloc[0,1] for f in predictor_features]})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "00f96328-4dd0-4b61-8f82-de7bb169cef3",
   "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>feature</th>\n",
       "      <th>corr</th>\n",
       "      <th>corr_spearman</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>int_rate</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>term</td>\n",
       "      <td>0.413934</td>\n",
       "      <td>0.398523</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>verification_status</td>\n",
       "      <td>0.225272</td>\n",
       "      <td>0.229827</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>loan_amnt</td>\n",
       "      <td>0.141949</td>\n",
       "      <td>0.114624</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>home_ownership__RENT</td>\n",
       "      <td>0.062024</td>\n",
       "      <td>0.068737</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 feature      corr  corr_spearman\n",
       "2               int_rate  1.000000       1.000000\n",
       "1                   term  0.413934       0.398523\n",
       "5    verification_status  0.225272       0.229827\n",
       "0              loan_amnt  0.141949       0.114624\n",
       "15  home_ownership__RENT  0.062024       0.068737"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "int_rate_corr.sort_values('corr', ascending=False).head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "c700c488-121c-4366-a8cb-20d841330314",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    100000.000000\n",
       "mean         13.252114\n",
       "std           4.776319\n",
       "min           5.310000\n",
       "25%           9.750000\n",
       "50%          12.740000\n",
       "75%          15.990000\n",
       "max          30.990000\n",
       "Name: int_rate, dtype: float64"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_samp_df['int_rate'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "74d84ba0-0971-43cd-bb28-41018d61f26b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "36    0.75937\n",
       "60    0.24063\n",
       "Name: term, dtype: float64"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_samp_df['term'].value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d55be4e-d37e-41ef-a3ce-ae6610239056",
   "metadata": {},
   "source": [
    "### Counts by interest rate / term / income bucket\n",
    "This section is informed by later analyses, in which I see that the reason for differences in SHAP and ALE have to do with nodes for lower-income, higher-interest rate, higher-term loans"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f24a523a-1cca-4a53-a759-cc69916721ce",
   "metadata": {},
   "source": [
    "##### Functions to get counts, rates by bucket"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "9b58eca0-d264-4c01-a0d2-a5e16661b236",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_samp_df['high_int'] = np.where(train_samp_df['int_rate'] > 30, 1, 0)\n",
    "train_samp_df['high_income'] = np.where(train_samp_df['annual_inc'] > 40000, 1, 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "9decea1d-7222-4894-9155-e9b118e422fe",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    0.81761\n",
       "0    0.18239\n",
       "Name: high_income, dtype: float64"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_samp_df['high_income'].value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "a99e100b-6441-42e9-83a7-128c361ed687",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.99648\n",
       "1    0.00352\n",
       "Name: high_int, dtype: float64"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_samp_df['high_int'].value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "ac6226b3-8654-40b4-ab05-ab2fe89bb8ac",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "36    0.75937\n",
       "60    0.24063\n",
       "Name: term, dtype: float64"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_samp_df['term'].value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "be0a8f4f-c74a-4286-81d7-bc33f71b38c0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.98081\n",
       "1    0.01919\n",
       "dtype: float64"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.Series(np.where(train_samp_df['int_rate'] >= 25, 1, 0)).value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "86ab0f25-9aca-42e0-90a6-84570a86d8fa",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.91743\n",
       "1    0.08257\n",
       "dtype: float64"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.Series(np.where(train_samp_df['int_rate'] >= 20, 1, 0)).value_counts(normalize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "aa8e4b09-082b-471a-b010-0a206b44d222",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['index', 'id', 'grade', 'sub_grade', 'hardship_flag',\n",
       "       'debt_settlement_flag', 'hardship_amount', 'settlement_amount',\n",
       "       'addr_state', 'purpose', 'total_rec_prncp', 'total_rec_int',\n",
       "       'tot_coll_amt', 'tot_cur_bal', 'application_type', 'loan_amnt', 'term',\n",
       "       'int_rate', 'emp_length', 'home_ownership', 'annual_inc',\n",
       "       'verification_status', 'fico_range_low', 'sec_app_fico_range_low',\n",
       "       'open_acc', 'initial_list_status', 'num_actv_bc_tl', 'mort_acc',\n",
       "       'pub_rec', 'revol_bal', 'target', 'home_ownership__MORTGAGE',\n",
       "       'home_ownership__RENT', 'home_ownership__OWN',\n",
       "       'home_ownership__ZZ_OTHER', 'high_int', 'high_income'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_samp_df.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "847eb776-0adb-4b0e-a012-35bf19e61d39",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Add binary high/low income and interest rate indicators\n",
    "def get_level_indicators(data):\n",
    "    data['high_int'] = np.where(data['int_rate'] > 25, 1, 0)\n",
    "    data['high_income'] = np.where(data['annual_inc'] > 40000, 1, 0)\n",
    "    return data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "155def04-4e13-4e7a-82f0-8a3f61fe5243",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to aggregate within categories.  Assumes get_level_indicators() has been done\n",
    "def cat_agg(data, agg_col = 'target', agg_func = 'count'):\n",
    "    bucket_overall = data \\\n",
    "        .groupby('high_income')[agg_col] \\\n",
    "        .agg(agg_func) \\\n",
    "        .rename('all_loans') \\\n",
    "        .reset_index()\n",
    "    bucket_int_rate = data \\\n",
    "        .groupby(['high_income', 'high_int'])[agg_col] \\\n",
    "        .agg(agg_func) \\\n",
    "        .rename('value') \\\n",
    "        .reset_index() \\\n",
    "        .pivot(index='high_income', values='value', columns='high_int') \\\n",
    "        .rename(columns={0:'int_low', 1:'int_high'})\n",
    "    bucket_term = data \\\n",
    "        .groupby(['high_income', 'term'])[agg_col] \\\n",
    "        .agg(agg_func) \\\n",
    "        .rename('value') \\\n",
    "        .reset_index() \\\n",
    "        .pivot(index='high_income', values='value', columns='term') \\\n",
    "        .rename(columns={36:'term_36', 60:'term_60'})\n",
    "    bucket_term_int = data \\\n",
    "        .groupby(['high_income', 'term', 'high_int'])[agg_col] \\\n",
    "        .agg(agg_func) \\\n",
    "        .rename('value') \\\n",
    "        .reset_index() \\\n",
    "        .pivot(index=['high_income', 'high_int'], values='value', columns='term') \\\n",
    "        .reset_index() \\\n",
    "        .pivot(index=['high_income'], columns='high_int') \\\n",
    "        .rename(columns={0:'int_low', 1:'int_high'}, level=1)\n",
    "    bucket_term_int.columns = [str(c[1]) + '_' + str(c[0]) \\\n",
    "                                   for c in bucket_term_int.columns.to_flat_index()]\n",
    "    return pd.concat([bucket_overall, bucket_int_rate, bucket_term, bucket_term_int], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "89b56b02-9dc0-4f7a-9451-681f8093294b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get counts, rates for one data set\n",
    "def get_data_stats(data):\n",
    "    data = get_level_indicators(data)\n",
    "    count_agg = cat_agg(data)\n",
    "    rate_agg = cat_agg(data, agg_func='mean')\n",
    "    \n",
    "    return pd.concat([count_agg, rate_agg], axis=0, keys=['count', 'rate'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "00f03f8e-a1f8-41c5-a834-b0a67c59379a",
   "metadata": {},
   "source": [
    "##### Get stats for each data set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "40d5151a-17d8-44cd-bf83-15e254b56a8b",
   "metadata": {},
   "outputs": [],
   "source": [
    "stats_all = pd.concat([get_data_stats(test_df), get_data_stats(train_df), get_data_stats(train_samp_df)],\n",
    "                      axis=0, keys=['test', 'train', 'train_samp'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "8f3288b8-f1b0-4a3d-a828-10f385b5b8da",
   "metadata": {},
   "outputs": [],
   "source": [
    "stats_all.to_csv(Path(setup.temp_path).joinpath('06_REPORT_dataset_stats_by_term_income_interest_rate.csv'), index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "14b7897a-6936-4ed6-a6e1-c9f5c188295c",
   "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></th>\n",
       "      <th></th>\n",
       "      <th>high_income</th>\n",
       "      <th>all_loans</th>\n",
       "      <th>int_low</th>\n",
       "      <th>int_high</th>\n",
       "      <th>term_36</th>\n",
       "      <th>term_60</th>\n",
       "      <th>int_low_36</th>\n",
       "      <th>int_high_36</th>\n",
       "      <th>int_low_60</th>\n",
       "      <th>int_high_60</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"4\" valign=\"top\">test</th>\n",
       "      <th rowspan=\"2\" valign=\"top\">count</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>68749.000000</td>\n",
       "      <td>67358.000000</td>\n",
       "      <td>1391.000000</td>\n",
       "      <td>60086.000000</td>\n",
       "      <td>8663.000000</td>\n",
       "      <td>59478.000000</td>\n",
       "      <td>608.000000</td>\n",
       "      <td>7880.000000</td>\n",
       "      <td>783.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>307938.000000</td>\n",
       "      <td>301838.000000</td>\n",
       "      <td>6100.000000</td>\n",
       "      <td>225587.000000</td>\n",
       "      <td>82351.000000</td>\n",
       "      <td>224390.000000</td>\n",
       "      <td>1197.000000</td>\n",
       "      <td>77448.000000</td>\n",
       "      <td>4903.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">rate</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.235931</td>\n",
       "      <td>0.229876</td>\n",
       "      <td>0.529116</td>\n",
       "      <td>0.210332</td>\n",
       "      <td>0.413483</td>\n",
       "      <td>0.207371</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.399746</td>\n",
       "      <td>0.551724</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.191522</td>\n",
       "      <td>0.185851</td>\n",
       "      <td>0.472131</td>\n",
       "      <td>0.146790</td>\n",
       "      <td>0.314058</td>\n",
       "      <td>0.145194</td>\n",
       "      <td>0.446115</td>\n",
       "      <td>0.303649</td>\n",
       "      <td>0.478483</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"4\" valign=\"top\">train</th>\n",
       "      <th rowspan=\"2\" valign=\"top\">count</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>159932.000000</td>\n",
       "      <td>156825.000000</td>\n",
       "      <td>3107.000000</td>\n",
       "      <td>140050.000000</td>\n",
       "      <td>19882.000000</td>\n",
       "      <td>138619.000000</td>\n",
       "      <td>1431.000000</td>\n",
       "      <td>18206.000000</td>\n",
       "      <td>1676.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>714519.000000</td>\n",
       "      <td>700645.000000</td>\n",
       "      <td>13874.000000</td>\n",
       "      <td>523701.000000</td>\n",
       "      <td>190818.000000</td>\n",
       "      <td>520989.000000</td>\n",
       "      <td>2712.000000</td>\n",
       "      <td>179656.000000</td>\n",
       "      <td>11162.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">rate</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.236438</td>\n",
       "      <td>0.231092</td>\n",
       "      <td>0.506276</td>\n",
       "      <td>0.209561</td>\n",
       "      <td>0.425762</td>\n",
       "      <td>0.207129</td>\n",
       "      <td>0.445143</td>\n",
       "      <td>0.413545</td>\n",
       "      <td>0.558473</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.191386</td>\n",
       "      <td>0.185836</td>\n",
       "      <td>0.471674</td>\n",
       "      <td>0.146372</td>\n",
       "      <td>0.314928</td>\n",
       "      <td>0.144940</td>\n",
       "      <td>0.421460</td>\n",
       "      <td>0.304432</td>\n",
       "      <td>0.483874</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"4\" valign=\"top\">train_samp</th>\n",
       "      <th rowspan=\"2\" valign=\"top\">count</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>18239.000000</td>\n",
       "      <td>17899.000000</td>\n",
       "      <td>340.000000</td>\n",
       "      <td>16007.000000</td>\n",
       "      <td>2232.000000</td>\n",
       "      <td>15847.000000</td>\n",
       "      <td>160.000000</td>\n",
       "      <td>2052.000000</td>\n",
       "      <td>180.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>81761.000000</td>\n",
       "      <td>80182.000000</td>\n",
       "      <td>1579.000000</td>\n",
       "      <td>59930.000000</td>\n",
       "      <td>21831.000000</td>\n",
       "      <td>59603.000000</td>\n",
       "      <td>327.000000</td>\n",
       "      <td>20579.000000</td>\n",
       "      <td>1252.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">rate</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.237074</td>\n",
       "      <td>0.232862</td>\n",
       "      <td>0.458824</td>\n",
       "      <td>0.212345</td>\n",
       "      <td>0.414427</td>\n",
       "      <td>0.210261</td>\n",
       "      <td>0.418750</td>\n",
       "      <td>0.407407</td>\n",
       "      <td>0.494444</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.192268</td>\n",
       "      <td>0.186700</td>\n",
       "      <td>0.474984</td>\n",
       "      <td>0.148440</td>\n",
       "      <td>0.312583</td>\n",
       "      <td>0.146587</td>\n",
       "      <td>0.486239</td>\n",
       "      <td>0.302882</td>\n",
       "      <td>0.472045</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    high_income      all_loans        int_low      int_high  \\\n",
       "test       count 0            0   68749.000000   67358.000000   1391.000000   \n",
       "                 1            1  307938.000000  301838.000000   6100.000000   \n",
       "           rate  0            0       0.235931       0.229876      0.529116   \n",
       "                 1            1       0.191522       0.185851      0.472131   \n",
       "train      count 0            0  159932.000000  156825.000000   3107.000000   \n",
       "                 1            1  714519.000000  700645.000000  13874.000000   \n",
       "           rate  0            0       0.236438       0.231092      0.506276   \n",
       "                 1            1       0.191386       0.185836      0.471674   \n",
       "train_samp count 0            0   18239.000000   17899.000000    340.000000   \n",
       "                 1            1   81761.000000   80182.000000   1579.000000   \n",
       "           rate  0            0       0.237074       0.232862      0.458824   \n",
       "                 1            1       0.192268       0.186700      0.474984   \n",
       "\n",
       "                          term_36        term_60     int_low_36  int_high_36  \\\n",
       "test       count 0   60086.000000    8663.000000   59478.000000   608.000000   \n",
       "                 1  225587.000000   82351.000000  224390.000000  1197.000000   \n",
       "           rate  0       0.210332       0.413483       0.207371     0.500000   \n",
       "                 1       0.146790       0.314058       0.145194     0.446115   \n",
       "train      count 0  140050.000000   19882.000000  138619.000000  1431.000000   \n",
       "                 1  523701.000000  190818.000000  520989.000000  2712.000000   \n",
       "           rate  0       0.209561       0.425762       0.207129     0.445143   \n",
       "                 1       0.146372       0.314928       0.144940     0.421460   \n",
       "train_samp count 0   16007.000000    2232.000000   15847.000000   160.000000   \n",
       "                 1   59930.000000   21831.000000   59603.000000   327.000000   \n",
       "           rate  0       0.212345       0.414427       0.210261     0.418750   \n",
       "                 1       0.148440       0.312583       0.146587     0.486239   \n",
       "\n",
       "                       int_low_60   int_high_60  \n",
       "test       count 0    7880.000000    783.000000  \n",
       "                 1   77448.000000   4903.000000  \n",
       "           rate  0       0.399746      0.551724  \n",
       "                 1       0.303649      0.478483  \n",
       "train      count 0   18206.000000   1676.000000  \n",
       "                 1  179656.000000  11162.000000  \n",
       "           rate  0       0.413545      0.558473  \n",
       "                 1       0.304432      0.483874  \n",
       "train_samp count 0    2052.000000    180.000000  \n",
       "                 1   20579.000000   1252.000000  \n",
       "           rate  0       0.407407      0.494444  \n",
       "                 1       0.302882      0.472045  "
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats_all"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cc5ccb8e-2003-4efd-ae99-dedabbcb1a59",
   "metadata": {},
   "source": [
    "### Purposes by bucket"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "b693b257-f54e-4ddb-8e45-3052857e13ee",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_purpose_stats(data, cat_col='purpose'):\n",
    "    purpose_dummies = pd.get_dummies(data[cat_col])\n",
    "    purpose_comb_df = pd.concat([get_level_indicators(data)[['id', 'high_int', 'high_income', 'term']], \n",
    "                                 purpose_dummies], axis=1)\n",
    "    \n",
    "    return pd.concat([cat_agg(purpose_comb_df, agg_col = c, agg_func='mean') for c in purpose_dummies.columns], axis=0,\n",
    "          keys = [c for c in purpose_dummies.columns])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "bc1b1323-4f31-4da1-a244-eee63fcb4b85",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_purpose_stats = get_purpose_stats(test_df)\n",
    "test_purpose_stats.to_csv(Path(setup.temp_path).joinpath('06_REPORT_purpose_by_term_income_interest_rate_test.csv'), index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "cf996fa2-06c2-4fa7-b8bf-e013da11f5ed",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_purpose_stats = get_purpose_stats(train_df)\n",
    "train_purpose_stats.to_csv(Path(setup.temp_path).joinpath('06_REPORT_purpose_by_term_income_interest_rate_train.csv'), index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "916fe7ab-8cca-4c4e-ae52-9bdb8d99348d",
   "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></th>\n",
       "      <th>high_income</th>\n",
       "      <th>all_loans</th>\n",
       "      <th>int_low</th>\n",
       "      <th>int_high</th>\n",
       "      <th>term_36</th>\n",
       "      <th>term_60</th>\n",
       "      <th>int_low_36</th>\n",
       "      <th>int_high_36</th>\n",
       "      <th>int_low_60</th>\n",
       "      <th>int_high_60</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">car</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.015901</td>\n",
       "      <td>0.015909</td>\n",
       "      <td>0.015449</td>\n",
       "      <td>0.016237</td>\n",
       "      <td>0.013530</td>\n",
       "      <td>0.016210</td>\n",
       "      <td>0.018868</td>\n",
       "      <td>0.013622</td>\n",
       "      <td>0.012530</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.009686</td>\n",
       "      <td>0.009775</td>\n",
       "      <td>0.005190</td>\n",
       "      <td>0.010676</td>\n",
       "      <td>0.006970</td>\n",
       "      <td>0.010699</td>\n",
       "      <td>0.006268</td>\n",
       "      <td>0.007097</td>\n",
       "      <td>0.004927</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">credit_card</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.222895</td>\n",
       "      <td>0.225583</td>\n",
       "      <td>0.087222</td>\n",
       "      <td>0.227255</td>\n",
       "      <td>0.192184</td>\n",
       "      <td>0.228951</td>\n",
       "      <td>0.062893</td>\n",
       "      <td>0.199934</td>\n",
       "      <td>0.107995</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.219336</td>\n",
       "      <td>0.221878</td>\n",
       "      <td>0.090962</td>\n",
       "      <td>0.229261</td>\n",
       "      <td>0.192099</td>\n",
       "      <td>0.230087</td>\n",
       "      <td>0.070428</td>\n",
       "      <td>0.198073</td>\n",
       "      <td>0.095951</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">debt_consolidation</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.567454</td>\n",
       "      <td>0.566581</td>\n",
       "      <td>0.611522</td>\n",
       "      <td>0.555901</td>\n",
       "      <td>0.648828</td>\n",
       "      <td>0.556186</td>\n",
       "      <td>0.528302</td>\n",
       "      <td>0.645721</td>\n",
       "      <td>0.682578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.582201</td>\n",
       "      <td>0.580856</td>\n",
       "      <td>0.650137</td>\n",
       "      <td>0.558546</td>\n",
       "      <td>0.647124</td>\n",
       "      <td>0.558588</td>\n",
       "      <td>0.550516</td>\n",
       "      <td>0.645433</td>\n",
       "      <td>0.674342</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">educational</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.000694</td>\n",
       "      <td>0.000708</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000750</td>\n",
       "      <td>0.000302</td>\n",
       "      <td>0.000757</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000330</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.000143</td>\n",
       "      <td>0.000146</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000191</td>\n",
       "      <td>0.000010</td>\n",
       "      <td>0.000192</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000011</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">home_improvement</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.041912</td>\n",
       "      <td>0.042040</td>\n",
       "      <td>0.035404</td>\n",
       "      <td>0.041885</td>\n",
       "      <td>0.042098</td>\n",
       "      <td>0.041913</td>\n",
       "      <td>0.039133</td>\n",
       "      <td>0.043008</td>\n",
       "      <td>0.032220</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.070272</td>\n",
       "      <td>0.070305</td>\n",
       "      <td>0.068618</td>\n",
       "      <td>0.070733</td>\n",
       "      <td>0.069008</td>\n",
       "      <td>0.070754</td>\n",
       "      <td>0.066740</td>\n",
       "      <td>0.069004</td>\n",
       "      <td>0.069074</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">house</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.005115</td>\n",
       "      <td>0.004725</td>\n",
       "      <td>0.024783</td>\n",
       "      <td>0.005012</td>\n",
       "      <td>0.005834</td>\n",
       "      <td>0.004689</td>\n",
       "      <td>0.036338</td>\n",
       "      <td>0.004998</td>\n",
       "      <td>0.014916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.005499</td>\n",
       "      <td>0.005249</td>\n",
       "      <td>0.018091</td>\n",
       "      <td>0.005320</td>\n",
       "      <td>0.005990</td>\n",
       "      <td>0.005173</td>\n",
       "      <td>0.033555</td>\n",
       "      <td>0.005472</td>\n",
       "      <td>0.014334</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      high_income  all_loans   int_low  int_high   term_36  \\\n",
       "car                0            0   0.015901  0.015909  0.015449  0.016237   \n",
       "                   1            1   0.009686  0.009775  0.005190  0.010676   \n",
       "credit_card        0            0   0.222895  0.225583  0.087222  0.227255   \n",
       "                   1            1   0.219336  0.221878  0.090962  0.229261   \n",
       "debt_consolidation 0            0   0.567454  0.566581  0.611522  0.555901   \n",
       "                   1            1   0.582201  0.580856  0.650137  0.558546   \n",
       "educational        0            0   0.000694  0.000708  0.000000  0.000750   \n",
       "                   1            1   0.000143  0.000146  0.000000  0.000191   \n",
       "home_improvement   0            0   0.041912  0.042040  0.035404  0.041885   \n",
       "                   1            1   0.070272  0.070305  0.068618  0.070733   \n",
       "house              0            0   0.005115  0.004725  0.024783  0.005012   \n",
       "                   1            1   0.005499  0.005249  0.018091  0.005320   \n",
       "\n",
       "                       term_60  int_low_36  int_high_36  int_low_60  \\\n",
       "car                0  0.013530    0.016210     0.018868    0.013622   \n",
       "                   1  0.006970    0.010699     0.006268    0.007097   \n",
       "credit_card        0  0.192184    0.228951     0.062893    0.199934   \n",
       "                   1  0.192099    0.230087     0.070428    0.198073   \n",
       "debt_consolidation 0  0.648828    0.556186     0.528302    0.645721   \n",
       "                   1  0.647124    0.558588     0.550516    0.645433   \n",
       "educational        0  0.000302    0.000757     0.000000    0.000330   \n",
       "                   1  0.000010    0.000192     0.000000    0.000011   \n",
       "home_improvement   0  0.042098    0.041913     0.039133    0.043008   \n",
       "                   1  0.069008    0.070754     0.066740    0.069004   \n",
       "house              0  0.005834    0.004689     0.036338    0.004998   \n",
       "                   1  0.005990    0.005173     0.033555    0.005472   \n",
       "\n",
       "                      int_high_60  \n",
       "car                0     0.012530  \n",
       "                   1     0.004927  \n",
       "credit_card        0     0.107995  \n",
       "                   1     0.095951  \n",
       "debt_consolidation 0     0.682578  \n",
       "                   1     0.674342  \n",
       "educational        0     0.000000  \n",
       "                   1     0.000000  \n",
       "home_improvement   0     0.032220  \n",
       "                   1     0.069074  \n",
       "house              0     0.014916  \n",
       "                   1     0.014334  "
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_purpose_stats.head(12)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0a778d50-28fb-4a74-a255-8698111bf3b4",
   "metadata": {},
   "source": [
    "##### Grades by bucket"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "482867ec-278e-4ec4-98e4-e097309f5bdc",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_grade_stats = get_purpose_stats(test_df, cat_col='grade')\n",
    "test_grade_stats.to_csv(Path(setup.temp_path).joinpath('06_REPORT_grade_by_term_income_interest_rate_test.csv'), index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "304c5fb9-3005-413e-bfa3-0f26809796fb",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_grade_stats = get_purpose_stats(train_df, cat_col='grade')\n",
    "train_grade_stats.to_csv(Path(setup.temp_path).joinpath('06_REPORT_grade_by_term_income_interest_rate_train.csv'), index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "b8da0211-08e5-401f-a64c-17eed3a1420f",
   "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></th>\n",
       "      <th>high_income</th>\n",
       "      <th>all_loans</th>\n",
       "      <th>int_low</th>\n",
       "      <th>int_high</th>\n",
       "      <th>term_36</th>\n",
       "      <th>term_60</th>\n",
       "      <th>int_low_36</th>\n",
       "      <th>int_high_36</th>\n",
       "      <th>int_low_60</th>\n",
       "      <th>int_high_60</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">A</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.108421</td>\n",
       "      <td>0.110569</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.122663</td>\n",
       "      <td>0.008098</td>\n",
       "      <td>0.123930</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.008843</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.189740</td>\n",
       "      <td>0.193497</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.251392</td>\n",
       "      <td>0.020538</td>\n",
       "      <td>0.252700</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.021814</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">B</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.278012</td>\n",
       "      <td>0.283520</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.304248</td>\n",
       "      <td>0.093200</td>\n",
       "      <td>0.307389</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.101780</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.294984</td>\n",
       "      <td>0.300826</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.346925</td>\n",
       "      <td>0.152433</td>\n",
       "      <td>0.348731</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.161904</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">C</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.320042</td>\n",
       "      <td>0.326383</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.325355</td>\n",
       "      <td>0.282617</td>\n",
       "      <td>0.328714</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.308635</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.275177</td>\n",
       "      <td>0.280626</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.255537</td>\n",
       "      <td>0.329078</td>\n",
       "      <td>0.256867</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.349524</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">D</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.185829</td>\n",
       "      <td>0.189511</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.173688</td>\n",
       "      <td>0.271351</td>\n",
       "      <td>0.175481</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.296331</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.141537</td>\n",
       "      <td>0.144340</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.110300</td>\n",
       "      <td>0.227269</td>\n",
       "      <td>0.110874</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.241389</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">E</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.078127</td>\n",
       "      <td>0.073470</td>\n",
       "      <td>0.313164</td>\n",
       "      <td>0.058115</td>\n",
       "      <td>0.219093</td>\n",
       "      <td>0.054928</td>\n",
       "      <td>0.366876</td>\n",
       "      <td>0.214655</td>\n",
       "      <td>0.267303</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.067777</td>\n",
       "      <td>0.063650</td>\n",
       "      <td>0.276200</td>\n",
       "      <td>0.029171</td>\n",
       "      <td>0.173731</td>\n",
       "      <td>0.027292</td>\n",
       "      <td>0.390118</td>\n",
       "      <td>0.169084</td>\n",
       "      <td>0.248522</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">F</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.023723</td>\n",
       "      <td>0.016305</td>\n",
       "      <td>0.398133</td>\n",
       "      <td>0.013431</td>\n",
       "      <td>0.096218</td>\n",
       "      <td>0.009508</td>\n",
       "      <td>0.393431</td>\n",
       "      <td>0.068054</td>\n",
       "      <td>0.402148</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.023876</td>\n",
       "      <td>0.016288</td>\n",
       "      <td>0.407092</td>\n",
       "      <td>0.005459</td>\n",
       "      <td>0.074422</td>\n",
       "      <td>0.003445</td>\n",
       "      <td>0.392330</td>\n",
       "      <td>0.053530</td>\n",
       "      <td>0.410679</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">G</th>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0.005846</td>\n",
       "      <td>0.000242</td>\n",
       "      <td>0.288703</td>\n",
       "      <td>0.002499</td>\n",
       "      <td>0.029424</td>\n",
       "      <td>0.000050</td>\n",
       "      <td>0.239693</td>\n",
       "      <td>0.001703</td>\n",
       "      <td>0.330549</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.006908</td>\n",
       "      <td>0.000774</td>\n",
       "      <td>0.316708</td>\n",
       "      <td>0.001216</td>\n",
       "      <td>0.022529</td>\n",
       "      <td>0.000090</td>\n",
       "      <td>0.217552</td>\n",
       "      <td>0.002755</td>\n",
       "      <td>0.340799</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     high_income  all_loans   int_low  int_high   term_36   term_60  \\\n",
       "A 0            0   0.108421  0.110569  0.000000  0.122663  0.008098   \n",
       "  1            1   0.189740  0.193497  0.000000  0.251392  0.020538   \n",
       "B 0            0   0.278012  0.283520  0.000000  0.304248  0.093200   \n",
       "  1            1   0.294984  0.300826  0.000000  0.346925  0.152433   \n",
       "C 0            0   0.320042  0.326383  0.000000  0.325355  0.282617   \n",
       "  1            1   0.275177  0.280626  0.000000  0.255537  0.329078   \n",
       "D 0            0   0.185829  0.189511  0.000000  0.173688  0.271351   \n",
       "  1            1   0.141537  0.144340  0.000000  0.110300  0.227269   \n",
       "E 0            0   0.078127  0.073470  0.313164  0.058115  0.219093   \n",
       "  1            1   0.067777  0.063650  0.276200  0.029171  0.173731   \n",
       "F 0            0   0.023723  0.016305  0.398133  0.013431  0.096218   \n",
       "  1            1   0.023876  0.016288  0.407092  0.005459  0.074422   \n",
       "G 0            0   0.005846  0.000242  0.288703  0.002499  0.029424   \n",
       "  1            1   0.006908  0.000774  0.316708  0.001216  0.022529   \n",
       "\n",
       "     int_low_36  int_high_36  int_low_60  int_high_60  \n",
       "A 0    0.123930     0.000000    0.008843     0.000000  \n",
       "  1    0.252700     0.000000    0.021814     0.000000  \n",
       "B 0    0.307389     0.000000    0.101780     0.000000  \n",
       "  1    0.348731     0.000000    0.161904     0.000000  \n",
       "C 0    0.328714     0.000000    0.308635     0.000000  \n",
       "  1    0.256867     0.000000    0.349524     0.000000  \n",
       "D 0    0.175481     0.000000    0.296331     0.000000  \n",
       "  1    0.110874     0.000000    0.241389     0.000000  \n",
       "E 0    0.054928     0.366876    0.214655     0.267303  \n",
       "  1    0.027292     0.390118    0.169084     0.248522  \n",
       "F 0    0.009508     0.393431    0.068054     0.402148  \n",
       "  1    0.003445     0.392330    0.053530     0.410679  \n",
       "G 0    0.000050     0.239693    0.001703     0.330549  \n",
       "  1    0.000090     0.217552    0.002755     0.340799  "
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_grade_stats"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8be62bbf-a390-4d4c-8436-2ce9610ca75d",
   "metadata": {},
   "source": [
    "### 2-Way plots of actual and predicted rates.\n",
    "What does the raw data / response show?  Bucket the interest rates and plot the mean term rates / model response for both term values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "372557f4-d804-44e6-a2ce-aca13de0eb61",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get model predictions\n",
    "model_predict = pd.Series(base_model.predict_proba(train_samp_df[predictor_features])[:,1])\n",
    "model_predict.name = 'predict'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "7c3d8203-3d56-4a63-9565-90ef2d6d4ee7",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Append to data\n",
    "train_pred_df = pd.concat([train_samp_df[['int_rate', 'term', 'target']].reset_index(), \n",
    "                           model_predict], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "d3f494cc-2dc9-40a3-accd-ab5a9ba5de2d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Bucket the interest rates\n",
    "train_pred_df['int_rate_grp'] = pd.qcut(train_pred_df['int_rate'], 25, duplicates='drop')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "b5b20dd5-0cf5-4ec3-a180-7d8047d3e7d3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Aggregate\n",
    "train_samp_agg = train_pred_df.groupby(['int_rate_grp', 'term']) \\\n",
    "    [['target', 'predict']] \\\n",
    "    .agg(['mean', 'std']) \n",
    "train_samp_agg.columns = [c[0]+'_' + c[1] for c in train_samp_agg.columns]\n",
    "train_samp_agg.reset_index(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "8d69fa6c-d3da-462e-9dae-c1faae587d08",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Add interest rate midpoints for plotting\n",
    "train_samp_agg['plot_x'] = train_samp_agg['int_rate_grp'].apply(lambda x: float(x.mid))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "701c3bb0-7de0-4048-bae5-cb25f521a3b3",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_samp_agg_plot = train_samp_agg[['plot_x', 'term', 'target_mean', 'predict_mean']] \\\n",
    "    .pivot(index='plot_x', columns='term', values=['target_mean', 'predict_mean'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "3b9925d9-24ad-4748-99c8-3aebf268e86b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Colors array for plot\n",
    "cmap = mpl.cm.coolwarm\n",
    "n = len(train_samp_agg['term'].value_counts())\n",
    "colors = cmap(np.linspace(0,1,n))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "50110c04-4ce4-4761-addc-a2e7fe719d61",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 800x200 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1,2, sharex=True, sharey=True)\n",
    "figsize=(8,2)\n",
    "    \n",
    "train_samp_agg_plot[[c for c in train_samp_agg_plot.columns if c[0] == 'target_mean']] \\\n",
    "    .droplevel(0, axis=1) \\\n",
    "    .plot(color=colors, legend=None, figsize=figsize, ax=ax[0])\n",
    "train_samp_agg_plot[[c for c in train_samp_agg_plot.columns if c[0] == 'predict_mean']] \\\n",
    "    .droplevel(0, axis=1) \\\n",
    "    .plot(color=colors, legend=None, figsize=figsize, ax=ax[1])     \n",
    "\n",
    "ax[0].legend(frameon=True)\n",
    "ax[1].xaxis.set_major_formatter(FormatStrFormatter('%.0f'))\n",
    "ax[1].yaxis.set_major_formatter(PercentFormatter(1, decimals=0))\n",
    "ax[1].set_xlabel('int_rate')\n",
    "ax[0].set_xlabel('int_rate')\n",
    "ax[0].set_ylabel('mean default')\n",
    "ax[0].title.set_text('Actual')\n",
    "ax[1].title.set_text('Model Prediction')\n",
    "fig.savefig(Path(setup.temp_path).joinpath('06_PLOT_int_term_actual_pred.png'),\n",
    "            bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a9205050-2bb8-4cee-9f71-8253c869056d",
   "metadata": {},
   "source": [
    "## PDP / ICE Plots - Term and Interest Rate"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8f360bf1-9f8c-437d-89b5-264031caa7fa",
   "metadata": {},
   "source": [
    "##### 2-Way PDP Average"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "97beb97f-db4b-4373-a044-1b25e537c2cc",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get 2-Way PDP\n",
    "avg_pdp = partial_dependence(base_model, train_samp_df[predictor_features], \n",
    "                                  ['int_rate', 'term'], kind='average',\n",
    "                             percentiles=(0,1),\n",
    "                            method='brute')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "89ec695d-1254-4d1f-a765-942fcabeed29",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Convert to data frame\n",
    "avg_pdp_df = pd.concat([pd.DataFrame(avg_pdp['values'][0], columns=['values']),\n",
    "                         pd.DataFrame(avg_pdp['average'][0], columns=avg_pdp['values'][1])],\n",
    "                        axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "a2a6647c-96ce-447d-a6d7-159850f6b676",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot\n",
    "fig, ax = plt.subplots()\n",
    "avg_pdp_df.set_index('values').plot(color = colors, ax=ax)\n",
    "ax.set_ylabel('2-way partial dependence')\n",
    "ax.set_xlabel('int_rate')\n",
    "ax.legend(title='term')\n",
    "fig.savefig(Path(setup.temp_path).joinpath('06_PLOT_int_term_pdp_avg.png'),\n",
    "            bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7a7e0e32-fea2-4e52-90b4-4f72196e2dbf",
   "metadata": {},
   "source": [
    "##### Overlay 1-Way Averages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "b9e4ae7b-fc20-486a-a7f1-1e5b1fb6e1fa",
   "metadata": {},
   "outputs": [],
   "source": [
    "term_value_counts = train_samp_df['term'].value_counts()\n",
    "color_categories = term_value_counts.index\n",
    "cnorm  = mpl.colors.Normalize(vmin=0, vmax= len(color_categories) - 1)\n",
    "color_scalar_map = mpl.cm.ScalarMappable(norm=cnorm, cmap=cmap)\n",
    "color_dict = {color_categories[i]: color_scalar_map.to_rgba(i) \n",
    "              for i in range(0, len(color_categories))}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "26b5d728-a488-402c-a1c3-c77b76f0b052",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAk0AAAG4CAYAAABYTdNvAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy89olMNAAAACXBIWXMAAA9hAAAPYQGoP6dpAACL7ElEQVR4nOzde3yO9f/A8de9e+cjsxNzHtswcl5iziGlyCmSkOSUQklJpVSUQ/UtPwkhcqgQUjLnw5hDwhg2hhk7sPN59339/li7s7Zx797hvre9n4+Hx6Nd1+e+rvd9d2/3+/4c3h+VoigKQgghhBDigcyMHYAQQgghREUgSZMQQgghhB4kaRJCCCGE0IMkTUIIIYQQepCkSQghhBBCD5I0CSGEEELoQZImIYQQQgg9SNIkhBBCCKEHc2MHUJm0bduWrKwsXF1djR2KEEIIIfQUGxuLpaUlJ0+efGA7SZpKUWZmJhqNxthhCCGEEKIYcnJy0GeDFEmaSpGbmxsAe/bsMXIkQgghhNBXjx499Gonc5qEEEIIIfQgSZMQQgghhB4kaRJCCCGE0IMkTUIIIYQQepCkSQghhBBCD5I0CSGEEELoQZImIYQQQgg9SNIkhBBCCKEHSZqEEEIIIfQgSZMQQgghhB4kaRJCCCGE0IMkTUIIIYQQepCkSQghhBBCD5I0CSGEEOUs8a8LpIbfMHYYopgkaRJCCCHKUdLZUI50GMShNs8QH/SXscMRxSBJkxBCCFGOQmctRNFo0KSmEfzUWBJPnTd2SEJPkjQJIYQQ5SRu/zFi/ziIytwcp7bNyUlK4Xjfl0g6d8nYoQk9SNJUxSmKYuwQhBCiSlAUhdC3FwBQ9+WhPPrnKqq1f4Tsewkc7zOalEtXjRyheBhzYwcgjGfHjh3s2bOHxYsXGzsUIYSo9O5s3kXiyXOo7WxpPGsi5g72tN/xHccef5Gkvy9y7PGRuPYKMF6ACqAouV+m8/4ZmaVLdWzqeWJT3xPberWx9aqDub2d0eKRpKmKOnHiBNOnT6d169bGDkUIISo9bXY2l2bnfkFtOG0MVu4uAFhUd8L/j5UE9XiBlAthRK7ebMwwTZ7azpbHDq7HsYWvUe4vSVMVpdVqjR2CEEJUGTdX/kzqlQgsXZ1pOHV0vnOWLs502LeOqPU70KSlGSnCf6hUoFKhMjMDVEYNRVG0ZMXcJf36LdIibpF+/RYoCioLC6PFJEmTEEIIUYZyUlK58tHXADSeNQlzB/sCbSydq1F/0ojyDq3CURQFlcp4yZwkTVXQzJkz2bJlCwCnT5/Gx8eH9u3b88MPPwAQERHB0qVLOXr0KPfu3cPZ2ZmAgAAmTpyIp6dnvmt1796dpKQkfvzxR2bOnMnly5dxdXXl888/58aNG7z99tvMnj2bpk2b8vXXX3PmzBnUajVt27bl7bffpm7dugQGBvLtt99y+fJlatSoQY8ePZg6dSq2trbl/toIIaqmlMvXODfhPXKSUkp+MUVBydGgzclGm5WNJiWNrNh72DasQ92Xh5T8+lWYMRMmkKSpSmrVqhXR0dEcPXoUZ2dnOnbsiJeXFwBBQUFMnDiRtLQ0vL29admyJdeuXePnn38mMDCQFStW4Ofnl+962dnZjBs3DnNzc7p06cLFixfx9fXlxo3carcHDhzgk08+oU6dOjz22GOcO3eOvXv3cuHCBUaPHs2nn35K8+bN6dixI0FBQaxZs4bbt2/z9ddfl/trI4Somm58t5F7B4PL9B6+82ZgZmlZpvcQZUuSpipo6NCh1K9fn6NHj1K/fn0WLMhdAhsfH8/rr79OZmYmX3zxBU888YTuMRs3buS9997j9ddfZ+fOnVje94ufkZGBs7MzGzZswNLSEq1Wi5nZv9UsDh48yJgxY5gxYwYqlYqUlBSeeuopbt++zaeffspnn33GM888A8CVK1d45plnCAwM1PVyCSFEWbt3+CQAjd6ZgHPHNiW+nsrcHJWFOWYWFphZmGNRoxq29WuX+LrCuCRpEjo//fQTCQkJvPDCC/kSJshNtPbt28e+ffvYvXs3Tz75ZL7zQ4YM0SVS9ydMANWqVWPq1Km6blV7e3u6du3K+vXradWqlS5hAmjcuDENGjQgLCyMGzduSNIkhChzOSmpJP11AYC6Lw3Bpm4tI0ckTJUUtxQ6x48fB8Df37/Q8wEBAfna3c/Xt+jln02bNs3XMwXokqHCHufg4ABAZmamHlELIUTJJASfRdFosKlbSxIm8UDS0yR0bt++DcDkyZMf2O7OnTsFjjk5ORXZvrBzeb1ODzonhBDlIW9orvpjUrdOPJgkTUJHo9EA0K1bN+ztCy6JzdOoUaMCx/47JHc/c3N5mwkhTNe9I6cAcO7U1siRCFMnn2ZCx83NjYiICEaOHMljjz1m7HCEEKLMabOzSTh2BqBUJoCLyk3mNFVRhQ2BtWvXDsgtEVCYzz77jP79+7Np06YyjU0IIcpL0l8X0aSlY1HdCfumBXvRhbifJE1VlJWVFQApKf8Wchs6dCi2trasXbuW3377LV/7vXv3smbNGkJDQ2nevHm5xiqEEGUlb2iuesfW/2wdIkTRZHiuiqpduzZqtZrLly/z4osv4uPjwzvvvMP8+fOZNm0a06ZN45tvvqFhw4bcvn2b8+fPA/DOO+/QpEkTI0cvhBCl496R3EngMjQn9GHyaXVwcDBjxoyhQ4cOtGrViueee46dO3eW6Jrbt2/Hx8eHN954o8g2u3fvZtiwYbRr1462bdsyevRogoKCSnRfU1KjRg0+/vhjateuzalTp9i3bx8AvXr14pdffuHpp58mOTmZ/fv3ExcXR7du3VizZg0vvviikSMXQojSoSgK8XmTwDvKJHDxcCpFURRjB1GUbdu2MWPGDMzNzfH390etVhMUFERWVhaTJk1iypQpxb7m7du3efrpp0lKSqJfv366atj3W7p0KYsXL8bW1hZ/f38yMjIIDg5Gq9Uyd+5cBg0aVOi1e/ToAcCePXuKHZcQQojylRIazoHmfTGztqL33ZOyxUkVpu/nt8kOz8XFxTF79mxsbGxYu3YtzZo1AyA8PJyRI0eyZMkSevTooTuuD0VReOutt0hKSiqyTWhoKIsXL8bNzY3169dTu3Zu2fuTJ0/y8ssv89FHHxEQEIC7u3vJnqAQQgijunc4t5epWvtHJGESejHZ4bl169aRkZHBiBEj8iVGXl5eTJs2DUVRWL16dbGu+f3333P8+HHdKrHCrFy5EoBJkybpEiZAN0SXkZHBhg0bivlshBBCmBpdfaaOUtRS6Mdkk6a8Ze89e/YscK5nz56oVCr279+v9/UuXbrE4sWL6datG88++6xB93388ccBinVfIYQQpkk3CVyKWgo9mWTSpCgKYWFhQO4Grv/l5OSEi4sLiYmJREdHP/R6WVlZvPHGG9jZ2TF37twi28XGxpKQkED16tVxcXEpcN7LywuVSkVYWJiuerYQQoiKJ+NWNOnXIsHMjGqPtjJ2OKKCMMmkKTExkczMTOzs7LC1tS20jZubG5A79+lhFi1axOXLl5kzZ06hyVCemJgYAFxdXQs9b2lpiZOTE1lZWSQnJz/0vkIIIUxT3tCcYwtfLByL3jZKiPuZZNKUnp4OgI2NTZFt8oozpqWlPfBaQUFBrFq1iqeffprevXuX232FEEKYrrxNeqU+kygOk1w9l7f5qz673Wu12iLPJSUl8fbbb+Pu7s7s2bPL7b5CCCHKR9LfodzasB0lO6dYj4veFgiAcydJmoT+TDJpsrOzAyAjI6PINpmZmQBFDt8BzJkzhzt37rBy5UocHR0fet+8a5X0vkIIIcqWoijcXPETIa99iDYr27CLmJlRXSaBi2Iw2aTJzs6O5ORkMjIysLa2LtAmb/5R3tym/zp37hw7duygWrVqbN68mc2bN+vORUZGAvDXX3/xxhtv4OXlxYQJE3S1l4qaJ5WVlUViYiIWFhZUq1atJE9RCCGEgTTpGZx/dQ6Rq3P/rrv07IhTa/1r9uWp1v4RrD0Kn8MqRGFMMmlSqVQ0btyYM2fOEB4eXqCAZUJCAnFxcTg5ORVZZDJvzlFCQgLbt28vtE1kZCSRkZG0b9+eCRMmUL16dVxdXYmNjSU+Pp7q1avnax8WFoaiKDRu3Fg3lCeEEKJ0KBoN2pwHr0zOuBHF6WGvk/T3RTAzw+fD1/F682XZbFeUC5NMmgACAgI4c+YMgYGBBZKmwMBAFEWhc+fORT7e39+fS5cuFXpu8+bNvP3224VuoxIQEMDmzZvZs2dPge1Sdu/eDfDA+wohhCi+lEtXORrwHNnxiXq1t3R1ptXaRbh071DGkQnxL5NNzQcNGoSNjQ2rVq3i9OnTuuNXr17liy++AGDs2LG64zExMYSHh+uG7Qw1fPhwzMzM+OKLL7h69aru+OnTp1m1ahVWVla88MILJbqHEEKI/KJ/26d3wuQc0I5OwVskYRLlzmR7mjw8PJg1axazZ89mxIgR+Pv7Y2lpSVBQEJmZmUyfPh1fX19d+0WLFrFlyxYGDBjAvHnzDL5v8+bNGT9+PEuWLKF///506NCBrKwsjh8/jlar5fPPP39grSchhBDFl/x3KACNZk2k4bSXimynUoG5g9RVEsZhskkTwODBg/Hw8GDZsmWcOXMGtVpN06ZNGTNmDL169Sqz+7722ms0bNiQNWvWcPz4caytrfH392f8+PH4+/uX2X2FEKKqSvonaarWroUUmxQmS6UoimLsICqLHj16ALBnzx4jR1K5KIrCTz/9xIYNGwgLC8PCwgIfHx+GDBlC//79C7TPyclhw4YNbN68mWvXrqEoCr6+vowaNYo+ffqU/xMQQjyQJiOTXdVaoWg0dL92AJvaHsYOSVQx+n5+m3RPkxAAH330EevWrcPGxoZ27dqhUqk4deoUb731FsePH+fTTz/Vtc3KymLChAkcPnwYBwcHHn30UZKSkjh9+jSvvfYa7733Hs8//7wRn40Q4r9SLoShaDRYOFfD2rPwFdFCmAJJmoRJO3DgAOvWraNmzZqsX7+emjVrAnD79m2GDRvG5s2b6dOnD126dAFg2bJlHD58mFatWrF06VJdPa3jx4/z0ksv8emnn/LEE0/g7OxsrKckhPiPpL8vAuD4iK9eOzIIYSwmu3pOCIBt27YBMGXKFF3CBFCzZk1dj9GhQ4eA3EruK1euxM7Ojq+++ipfAVJ/f38GDRqEm5sb586dK78nIIR4qKSzufOZHFv4PqSlEMYlPU3CpM2bN48JEyZQq1atAufyCpiq1WoADh8+TGpqKkOGDCm0UvwHH3xQprEKIQyTNwnc8ZEmRo5EiAeTpEmYNAsLCxo1alTg+F9//cWPP/6IWq2mX79+AISEhADQokULcnJyCAwM5OTJk2RlZeHn58fTTz9d6JY8QgjjURRFeppEhSFJkwlTFIWMTK2xwygWayuzMp2TMH36dMLDw7l48SLVqlVj4cKF+Pn5AXDjxg0gdxueIUOG6JIogI0bN7Js2TKWLVtGw4YNyyw+IUTxpF+/RU5iMioLC+ybyO+mMG0lTpqys7O5ePEit2/fxsXFhTZt2hAVFVXocIrQn6IoTHzrDOcuJhk7lGJp3sSRJfNblkniFB8fz44dO3Q/q1QqLl++TK9evVCr1SQnJwMwf/58HBwc+Pbbb2nbti137txh0aJF7Nmzh1deeYXt27dLj5MQJiJvaM6hiRdmlpZGjkaIBzN4InhOTg5ffvklHTt2ZOjQobz++uts3LgRgBkzZjBw4EDdN38hSoOdnR1Hjx7l1KlTLFu2DEdHR5YsWcL7778P5JYbAEhPT2f58uV07doVe3t7GjVqxNdff02TJk24ceMGW7duNeKzEELcTzc0J/OZRAVgUE9TTk4Or7zyCkePHsXc3Bw/P798K5KSkpK4fPkyw4cPZ+vWrbLtiAFUKhVL5reU4bn7WFpaUqNGDQC6dOlCgwYNePrpp/nll1945ZVXdL1HHTt2LDAEZ2ZmxpAhQ5gzZw7Hjh3jueeeK5MYhRDFk1duwEHmM4kKwKCepnXr1nHkyBE6duzI3r17+emnn/Kd37RpE8888wxxcXGsXLmyVAKtilQqFTbW6gr1rzxrrNStW5dWrVqh1Wq5ePGirvZS7dq1C23v6ekJ5A7zCSFMg66nqaUkTcL0GZQ0bdmyhWrVqvHll1/i6upa4Ly1tTVz587Fzc2NgwcPljhIUXUtWrSI1157TVde4L8s/5kDkZOTg7e3NwDR0dGFto2LiwPQ9VYJIYwrOyGJ9GuRgKycExWDQUlTREQEbdu2xc7Orsg2FhYWtGjRglu3bhkcnBAHDhzgjz/+IDAwsMC5pKQkzpw5A0CzZs10VcGPHDlSaG9SXgLftm3bsgtYCKG3pHOXALCuUxNL52rGDUYIPRiUNJmbm5OU9PBVXQkJCZibS1UDYbi8uUefffYZERERuuOJiYm8+eabJCQk0LNnT+rVq0eDBg3o3r07aWlpzJgxg9TUVF37LVu28Mcff1CjRg2eeuqp8n4aQohC/FvUUnqZRMVgUEbTpEkTzp07x+3bt/NtbXG/mzdvcu7cOVq0aFGiAEXVNnToUI4fP87vv/9Ov379aNOmDebm5pw9e5bExESaNWvGJ598omv/0UcfcePGDQ4ePEj37t1p06YNt2/f5sKFC1hbW/Ppp5/i6OhoxGckhMij23NOhuZEBWFQT9OwYcNIT09n4sSJhIaGFjh/6dIlXn31VbKyshg0aFCJgxRVl5mZGYsXL2bu3Ln4+vry119/ceLECWrWrMn06dNZv349Tk5OuvYuLi5s2rSJyZMn4+zszKFDh4iOjqZPnz5s3LhRN4QnhDC+ZCk3ICoYlaIoiiEPfO+999i0aRMqlQoHBweSk5OpVq0aVlZWREdHoygKTz75JAsXLiztmE1Wjx49ANizZ4+RIxFCCNOmzclhV7VWaDOz6HrxT+wa1TN2SKIK0/fz2+AJRx9++CEtW7bku+++49q1a8C/S7lr1arFiy++yMiRIw29vBBCiEos9dI1tJlZqO1tsW1Yx9jhCKGXEs3SfvbZZ3n22WeJi4sjKioKrVaLq6urrh6OEEIIURjdfKbmPqjMDN6cQohyVaJ36r179zh+/DguLi60aNGCli1bkpaWxpo1a4qslSOEEELokiaZzyQqEIOTpvXr19O5c2fmz5+f7/jZs2f55JNP6NOnD9u3by9xgEIIISoHRVHIuBNL3N4g4vYeAyRpEhWLQcNzhw4dYs6cOdjY2NCxY8d859q0acPYsWNZt24dM2bMwMXFhQ4dOpRKsEIIIUqPJjOL1NCrJJ0LJfncJZIvhKFJKbz6fklps7JJDbtO9r2EfMedWjUtk/sJURYMSpqWL1+Oubk5P/zwA35+fvnO1a9fnzfeeIPevXvz3HPPsXz5ckmahBCiFCiKwumhU4jbc7RUrqdJy0DJySmVa+lNpcLWqy4OTbyo0fVRHFs3K9/7C1ECBiVNV65cwd/fv0DCdL/mzZvTtm1b3TYXQgghSib2z0Pc2fJnqV7T3MkBx+Y+ODT3wcHPG8sa1Ur1+jpmZtg2qIO9TwPUNtZlcw8hyphBSVNGRgY2NjYPbefk5EROeX+LEUKISura4u8BqDt2KA2njSnx9cxsrLH2dEelUpX4WkJUBQYlTfXr1+fEiROkpKRgb29faJuMjAxOnjxJ3bp1SxSgEEKI3H3a4vYcBTMzvN4ah2392sYOSYgqx6DVc08//TSJiYlMmTKFuLi4Aufv3bvH66+/Tnx8vGyOKoQQpeDal6sAqDmwtyRMQhiJQT1Nzz//PL///jtHjx6la9eu+Pn56TbuvXPnDufPnyc7O5tmzZoxatSo0oxXCCGqnIyoaG5t2AFAw9dLPiwnhDCMQUmThYUF33//PYsWLWLz5s2cOXMm34RvS0tLhg4dyptvvomVlVVpxSqEEFVSxJJ1KNnZVO/YhmrtWxg7HFFCaekaZnx4DksLM14f14i6tW2NHZLQk8HbqNja2vLuu+8yY8YMzp8/T2xsLBqNBldXV5o1a4atrbwJhBCipHJS07ixbAMADadKL1Nl8MuOW5w5nwjAqNdOMfb5+gx9pjZqtUzIN3Ul2nsOcnuVWrduXRqxCCGE+I/IVZvJjk/E1qsu7k91M3Y4ooRS03JYv+UmAPVq23I9Mo0l319l/9FY3nnNh/p17IwaX1a2luSUHJJTsklOySE7RzFqPHY2apyrW1LN0QILC+PvUViipCklJYWrV6+Snp6OVqstsp0UtxQldefOHZYsWcKhQ4eIjY3FycmJDh06MGXKlAIrNOPi4nRto6OjcXV1pU+fPkyYMKHI1Z5CmCJFo+HaV6sAaDBlFCq12rgBiRL7afstkpJzqOtpw+qv2/LH3jv8b3k4Fy4lM3LySWysjff/OCdHITOr6M9yY7O3M6d+HVs+fKspbi7GmfpjUNKk1Wr55JNP2LBhAxqN5oFtVSoVFy5cMCg4gODgYJYuXcrFixfJyMjAx8eHkSNH0rdvX72vcePGDb755huOHDlCQkICrq6uBAQEMGHCBN0E9vutWrWKTz/9tMjrvfzyy7zxxhsGPR9RfBcuXGD06NEkJCTg5eVF165dCQ0NZfv27Rw+fJhffvkFT09PAGJiYnjuuee4desW3t7edO3alXPnzrF8+XIOHjzI+vXrJXES5UrRaon5/QA3lm0g805ssR6rycwi7epNLKo7UfvFAWUUoSgvKak5bNgSCcDoYfUxV6t46vGatG/lzGdfX+bYqXukpj34M7U8qFS5CYqDvTmWRu7dSU7NISEhC4029/U7H5rEnZiMipU0rV69mrVr1wJQu3Zt3NzcUJfBN6Bt27YxY8YMzM3N8ff3R61WExQUxNSpUwkLC2PKlCkPvUZERARDhgwhMTGRhg0b0rJlS65du8bGjRvZuXMna9euxdfXN99jQkJCAOjatSsODg4Frtm0qeyVVF6ysrKYPn06CQkJTJ8+nXHjxgGg0WiYO3cuP/74Ix9//DFLliwB4MMPP+TWrVuMHz+eqVOn6q4xY8YMfv/9d7744gveffddoz0fUXVoMrOIWr+dq4tXknIhrETXqv/qSMztZJ5oRbdpWyQpqTnUr2NL906uuuNuLlYs+KA50bEZZGUbr6fHzEyFg7059rbmmJmZzvwqrVYhOSWHewlZWJibUbvWw4trlxWVoijFHrB86qmnuHbtGt988w1du3Ytg7Byh1h69OiBmZkZa9eupVmz3P2JwsPDGTlyJHfv3uWXX37RHS/K4MGDOXv2LK+++iqTJk3SVb79+uuv+d///keLFi346aefCjy/sLAwTp48WaxeiR49egCwZ8+e4jxV8QDbtm3jzTffpHfv3nz11Vf5zqWnp9O3b18cHR3ZvHkzkZGR9O7dGw8PDwIDAzE3//c7QXJyMl27diUnJ4ejR49iZ2fceQOiYsi4E8vFGfPJTkgq9mOT/r5IZlQMAOYOdtR9eSjOXfyLXX3bzMYK505tMTMv8RRUYURJKdkMGXuclFQNc2Y0oUeAm7FDEvfR9/PboN/CGzdu0K5duzJLmADWrVtHRkYG48aNy5cYeXl5MW3aNN555x1Wr17NZ599VuQ1IiIiOHv2LJ6envkSJoCJEyeyYsUKzp49S0JCAtWqVQNyK5lfvXqV+vXryzCOCdi1axdAofW+bGxs2Ldvn+7ngwcPoigKXbp0yZcwATg4OODv78+ePXs4duyY7hdEiAeJXPULUeu3G/x4q1puNJjyInXHDsXCqWCvtag6Nm2NJCVVQ8N6dnTr6PrwBwiTZFDS5OjoWOb1lw4cOABAz549C5zr2bMns2bNYv/+/Q+8Rv369QkKCiIpKanAt7vs7Gyys7MB8g0tXrp0CY1G89AeLFE+QkJCMDMzw8/Pj5iYGHbs2MG1a9ewt7enW7dutG/fXtf2ypUrAHh7exd6rUaNGrFnzx4uX74sSZPQS9K5SwDUeu4pXHo8VqzHWlR3wu2JzphZWpZFaKICSUrOZtO2WwCMGVbPpIa+RPEYlDQFBASwd+9eEhMTcXJyKu2YUBSFsLDcOQCNGzcucN7JyQkXFxdiY2OJjo7G3d29yGs5Ozvj7Oyc71h6ejofffQR2dnZ9OzZM9+8pbz5TE5OTrz33nscOXKEmJgYPD096devH2PHji23gp2KoqBJSy+Xe5UWta1NqW3+mZWVxe3bt6levToHDhzgrbfeIjU1VXd+5cqVDBgwgLlz52Jubk5MTO5QiKtr4d/i8o7fvXu3VOITlV/y+csAeA5/Grcnuhg5GlER3Y3PYsE3l0lL19CogR2dO7gYOyRRAgYlTdOmTePo0aO89tprzJo1q9DEpiQSExPJzMzEzs6uyCKZbm5uxMbGEhcX98Ck6X6BgYGsX7+es2fPkpSURLdu3Zg3b16+Nnkr/datW4ezszOtWrXC3d2dkJAQvvrqKw4ePMj3339f5sU7FUUhqMsw4oP+KtP7lLbqj7Wmw/4fSyVxSklJASAtLY1p06bRrVs3pkyZgoeHBydOnOD9999ny5YtuLm5MW3aNNLTcxNMa2vrQq+Xd/z+xEuIomgys0i9HAGAg1/hvZdCFCUnR8vPO26x8sfrpKVrUKlg/IsNpZepgjMoaZozZw4eHh4cP36cp59+GmtraxwdHQv9oFSpVPnmnegj78PPxqboGfJ5vT1paWl6XzcoKIjDhw/nu09ERATNmzfXHctLmgYOHMgHH3yA5T9d6zdv3mTy5MmcOXOGhQsXMnv2bP2fkKFKqcemosrKygIgMzOT9u3b55sI3q1bN1xcXBg8eDCrVq1i7NixumHWhyVsBqx9EFVQ6qWrKDk5mDs5YF3bw9jhiHJw5WoKCYlZJb5OUkoO36+/TsTN3M+nJo0deP2VRjTzcSzxtYVxGZQ0BQYG5vs5PT1dl+j8lyE9DmZmZno/9kFFNf9r/PjxvPXWW0RHR7N27VpWrVrFyJEj+fnnn/Hy8gJye5giIyNp2LBhvrlOderUYd68eQwYMICffvqJGTNmlOkwnUqlosP+H6v08Nz9PUbPP/98gfPNmzenefPmnD17lr/++kvX+5eZmVno9TIyMgBkix+hl7yhOYdmjUvtPS1M13drr7F6441SvWY1JwvGj2xA354e0sNUSRiUNJX1kvq85eB5H3KFyftgLM4HYN6cljp16vD222+Tnp7Oxo0b+e6773TDdDY2NkUONzZp0gQPDw9u377NlStX8PPz0/vehlCpVFW6NouDgwMWFhZkZ2dTu3btQtt4enpy9uxZ4uPjdcO0sbGFFxDMO17UnCch7pcckruwQIbmKr+gk3d1CZNXfbsSd/KrVCpat6jGqKH1cLCXUhGViUH/N/OqL5cVOzs77OzsSE5OJiMjo9A5KnmTft3cDK918cwzz7Bx40bd5G99uLi4cPv27QcmdKJ0qNVqvLy8CA0NJTo6utAkNS4uDoAaNWrokt3w8PBCr/ew1XVC3E/X0yRJU6V2JyaDjxaFAvDsk7WYNr505+iKyqXE9dFjYmLYuXMnK1eu1A3bnTt3Trec3xAqleqBH4AJCQnExcXh5OT0wEngJ0+eZPbs2QWKV+bJm6+Uk5MDQHR0NLNmzeLNN98s8po3b+ZutKjv5HNRMl265K5Y2rlzZ4Fzd+/eJSQkBEtLSx555BECAgJ0c+j+u71PcnIyx48fx8bGJl+ZAiGKIklT5ZedreW9zy6QlJyDbyMHJr/kZeyQhIkzOGlKTk5mxowZdO/enenTp/P555/z559/AvDxxx/To0cP/v77b4MDCwgIAArOn8o7pigKnTt3fuA1EhIS2LRpEytWrCh07tPBgwcBdDWZ7O3t2bZtG9u2bePixYsF2u/fv5+EhATq169PnTp1iv2cRPE999xz2NrasmPHjnzJb1paGrNmzSItLY0BAwbg6OiIp6cn3bp1IzIyks8//1w34TsrK4v33nuP1NRUhg4dWujWOELcLzsphfQbUQA4StJUaX3z/VUuXErG3s6cD99qYvR91oTpM2h4Lj09nRdffJELFy7g4uJC27Zt+eOPP3TnLSwsiImJYfTo0fz6668GJRiDBg1i+fLlrFq1ioCAAFq3bg3A1atX+eKLLwAYO3asrn1MTAzJyck4ODjohuw6d+6Mp6cn165dY8GCBbzxxhu6SeZ79+5l6dKlqNVqRo8eDeQOC/br149ffvmFd999l++++05X4ykiIoIPP/wQgEmTJhX7+QjD1KpVi/nz5zNt2jTeffdd1qxZQ+3atTl37hyxsbH4+vrm6xl87733CAkJ4fvvv+fAgQM0btyYc+fOERUVRbNmzfTar1CIvF4ma093LKqXfi06UXoys7Scv5hIdk7xVsVev5nGz9tzC06+O82HWh7G289MVBwGJU0rVqzgwoULDBgwgA8++AArK6t8m97+8MMPLF68mG+//ZbvvvtOl2wUh4eHB7NmzWL27NmMGDECf39/LC0tCQoKIjMzk+nTp+e756JFi9iyZQsDBgzQTeq2tLRk4cKFvPTSS6xYsYLdu3fj4+PDzZs3CQ0NxdzcnDlz5uSr/v3WW29x/vx5zp8/T69evWjdujUajYbg4GCysrIYNWoUTz/9tCEvmzBQr169+OWXX1i6dCnHjx8nIiKCWrVqMWTIEMaOHZtvMUDNmjX56aef+N///sf+/fvZt28ftWrVYvz48bz88suy55zQiwzNVRxfrwhny84ogx8/fGAdOrWXgpNCPwYlTTt37sTd3Z0PP/wQCwuLQtu8/vrr7Ny5k+PHjxsc3ODBg/Hw8GDZsmWcOXMGtVpN06ZNGTNmDL169dLrGq1ateLXX39lyZIlHD58mH379uHk5ESfPn0YO3ZsvhpNkFsJfMOGDaxcuZKdO3cSFBSElZUVLVu2ZOTIkTz++OMGPx9hOB8fHxYvXqxXW3d3d+bOnVvGEYnKTLdyrplMCjZ1Zy8kAlDH0wYba/VDWuf3SDMnxo2oXwZRicrKoKQpMjKSrl27FpkwQe5kbl9fX928IUMFBATo5jc9yLx58wpU985Tp04dPv30U73vaWtry+TJk5k8ebLejxFCVB7/9jT5GDkS8SA5GoUbkbkFJBd+0FyG2ESZM2jWm42NjW6p94PExMQUuaWFEEKYIkVRZHiugoi6k052joK1lRkebvJZI8qeQUlTixYtOHfunK7uTWEuXrxISEgILVq0MDg4IYQob5l3Ysm+lwBmZtj7NjR2OOIBrt3I7WWqV8dWKm6LcmFQ0jR69Giys7N55ZVXCAwMJDk5WXcuOzubffv2MWnSJDQaTaHbXwghhKnK62Wya1wPtY30XpiyazdyN99uUFcWeIjyYdCcpscee4xp06axePFiXn31VSB3DtOuXbv47bff0Gq1KIrCSy+9RNeuXUszXiGEKFMyNFdxRPzT01S/TtXdbkqUL4M3xRk3bhwtWrRg+fLlnDp1ivT0dDIzMzE3N6dt27a8+OKL9OzZszRjFUKIMpd8Pm/lnCRNpk56mkR5K9FOgo8++iiPPvooWq2WhIQENBoN1apVe+CqOiGEMGXJ5y8B0tNk6nI0Cjdv/dPTVFd6mkT5KJXtl83MzHSVs4UQoqJSNBqSL4QBsn2KqYu6k05WtoKVpRk1ZeWcKCd6JU1bt24t0U369+9foscLIUR5SLt6E21GJmY21tg2lP0lTVmErJwTRqBX0jRz5kxUKsPflJI0CSEqgqS8SeBNvFCpi1ddWpQvmc8kjEGvpOmpp54qkDSdPXuW69evU6NGDTp16kTt2rVRq9XExMRw+PBhIiMjadmyJZ06dSqTwIUQorRJJfCKI+KmrJwT5U+vpGnBggX5fj558iS///47AwcO5P3338fS0jLfea1Wy7x581i7di3jx48vvWiFEKIMJYdIuYGKQtfTVE+SJlF+DJoI/uWXX+Lh4cGHH36IupAubDMzM95++20OHjzIN998I7WahBDlKuHEWa4uXok2K/uhbVXmaszMzVGZm3N3f+4G4w5+slGvKdPct+dcgzoyPCfKj0FJ0/nz5+nSpUuhCVOevA179+/fb2hsQghhkAtvfEr80dMGPValVuPYwreUIxKlKSr635VzsuecKE8GJU329vZEREQ8sI2iKISGhlK9enVDbiGEEAZJv3k7N2FSqWi6aBZqK8si2yqKgqLRoGTnoM3JQcnR4PiIL1buLuUYsSgu3cq52rao1bJyTpQfg5KmRx99lB07drBixQpeeumlAucVRWH+/Plcv36dYcOGlThIIYTQ1+2ffwfAuVNbGkx+wcjRiLLw78o5mc8kypdBSdPkyZM5cOAACxYsYOfOnQQEBODu7g5AVFQUe/bs4dq1a3h4eDBx4sRSDVgIIR4kauNvANQc/ISRIxFlRbdyTsoNiHJmUNJUr149Vq5cyaxZswgJCSEkJERXkkBRFADatGnDp59+iouLdHMLIcpHavgNEk+dBzMzag7sY+xwRBm59s/wnPQ0ifJm8DYqfn5+bN26ldOnT3PixAliYmJQqVS4u7vToUMHmjdvXppxCiHEQ93+aScALt07YOVWw8jRiLKg0Shcj8yr0SQ9TaJ8lWjvOZVKRZs2bWjTpk1pxSOEEAaL2pSbNMnQXOV1OzqDrCwtlpZm1HSXlXOifJV4w9579+6RlpamG5YrTJ06soeTEKJsJV8II/ncJVQWFnj0f9zY4YgyEnEzdxK4rJwTxmBw0rRq1SqWL1/O3bt3H9hOpVJx4cIFQ28jhBB6yRuac328I5bO1YwbjCgzMp9JGJNBSdPmzZuZN28eAJaWllSvXv2BhS6FEKIsKYpC1D9JU80hfY0cjShLEf+UG5A954QxGJQ0rV27FpVKxbvvvsuQIUOwsLAo7biEEEJvSX+HknrpGmbWVrj362HscEQZ+renSSaBi/JnUNIUHh5Oy5Ytef7550s7HiGEKLbbm3JrM7k90QULR3sjRyPKikajEJG3ck6G54QRGJQ02djYyPYoQgiTcP/QXC0ZmquQsrO1/LE3ml//uE1Kak6R7TRaJXflnIWKWu425RihELkMSprat2/P6dOnycjIwNpalnwKIYwn6e9Q0iNuobazxa1vV2OHI4ohM1PDjt13WPfLTWLiMvV+XItmTrJyThiFQUnT1KlTGTx4MLNmzeL999/H0dGxtOMSQgi9xAedBsC5Y2vUttL7UNbi7mYSFZ1R7MdptQqJSdncS8jmXkIWd+9lcST4LnfjswCo4WzJ8AF1aOrj8MDrqFTQuIEMwQrjMChp2rBhAy1atGDnzp3s2rWL+vXr4+joqNtK5X4qlYq1a9eWOFAhhChMQvBZAJzatTByJJWboihs+vUW/7f6Kjk5RdflKy53VytGDKpL354eWFmaldp1hSgLBiVNq1ev1v13Tk4OYWFhRbYtLJESQojSknDibwCqtX/EyJFUXvGJWXz6xSWOnrwH5CY6FhbFS3BUKnBysKB6NQucq1niXM2SenVs6dLBpdjXEsJYDEqa1qxZU9pxCCFEsWXHJ5J66RoA1aSnqUycPpfAhwsuEncvC0sLFZNf8mJA31ryhVhUSQZPBBdCCGNLOHkOANuGdbBydTZyNJXP73vv8MkXl1AUqOtpw5y3msp8IlGllXjvuezsbC5evMjt27dxcXGhTZs2REVFUatWrdKITwghipQQLENzZSVHo/DdDxEoCvTu5s4bExtjYy07P4iqzeCkKScnh2+++YZ169aRnJwMQL9+/WjTpg0zZswgPT2dxYsXU7du3RIFGBwczNKlS7l48SIZGRn4+PgwcuRI+vbVvx7LjRs3+Oabbzhy5AgJCQm4uroSEBDAhAkTqFmzZqGP2b17NytXriQsLAxFUWjevDnjxo2jQ4cOJXo+QojSkzcJXIbmSt+xk3eJicukmqMFb73qjaXMOxICg34LcnJyeOWVV1i6dCnp6en4+fmhKP+upkhKSiIkJIThw4cTFxdncHDbtm1j5MiRBAcH07RpU9q1a0dISAhTp07lq6++0usaERERDBo0iK1bt+Lg4EDXrl2xtbVl48aN9OvXj9DQ0AKPWbp0KZMnTyY0NJQ2bdrg5+fH8ePHGT16ND///LPBz0cIUXoURZGepjK09ffbAPTt6S4JkxD/MOg3Yd26dRw5coSOHTuyd+9efvrpp3znN23axDPPPENcXBwrV640KLC4uDhmz56NjY0NGzduZMWKFSxbtoytW7fi4uLCkiVLCAkJeeh13nzzTRITE3n11VfZuXMnX3/9Nb/99huvvvoqycnJzJ49O1/70NBQFi9ejJubG9u3b2fp0qWsWrWKNWvWYGNjw0cffUR0dLRBz0kIUXrSr0WSFRePysICx5ZNjB1OpRJ1J53jp3NXyj3dW6ZaCJHHoKRpy5YtVKtWjS+//BJXV9cC562trZk7dy5ubm4cPHjQoMDWrVtHRkYGI0aMoFmzZrrjXl5eTJs2DUVR8pU+KExERARnz57F09OTSZMm5VvtMXHiRGxtbTl79iwJCQm643lJ3qRJk6hdu7bueNu2bRk9ejQZGRls2LDBoOckhCg98f/0Mjk+4ova2srI0VQu2/+8jaJAu5bVqV1LCoYKkcegpCkiIoK2bdtiZ1f0LtMWFha0aNGCW7duGRTYgQMHAOjZs2eBcz179kSlUrF///4HXqN+/foEBQWxcuXKAstjs7Ozyc7OBkCt/ndy44Pu+/jjjwM89L5CiLInQ3NlIztby47ddwDo/0Thcz6FqKoMSprMzc1JSkp6aLuEhATMzYs/11xRFF3BzMaNGxc47+TkhIuLC4mJiQ8dKnN2dqZ+/fr5jqWnpzNnzhyys7Pp2bMnDg65ZftjY2NJSEigevXquLi4FLiWl5cXKpWKsLAwNBpNsZ+XEKL06CaBt5dJ4KXp4LE44hOyqeFsScf2NYwdjhAmxaCkqUmTJpw7d47bt28X2ebmzZucO3cOX1/fYl8/MTGRzMxM7OzssLW1LbSNm5sbQLEmmgcGBvLSSy/RuXNnfvnlF7p168a8efN052NiYgAKHXIEsLS0xMnJiaysLN2KQSFE+dNmZZF05gIA1aWnqVT9+kfu3/V+j3tgbi4TwIW4n0G/EcOGDSM9PZ2JEycWuvrs0qVLvPrqq2RlZTFo0KBiXz89PR0AG5uix9KtrHLnMKSlpel93aCgIA4fPqzrJUtPTyciIqLM7yuEKF1Jf4eizczCwrkato3qGTucSuP6zTROn03AzAye6iVDc0L8l0FJU9++fRkyZAgXL15kwIABtG/fHpVKxaFDh+jatSv9+/cnNDSUvn378swzzxQ/KLPcsPQp06/VavW+7vjx4zl37hyBgYGMGjWKY8eOMXLkSMLDw8v0vkKI0vVvfabmsp1HKdq2KwqADm1r4OFmbeRohDA9Bve9fvjhh3zyySfUr1+fpKQkFEUhPj6eO3fuULNmTd5++20WLFhg0LXzJphnZGQU2SYzMxOgyOG7wri6umJpaUmdOnV4++23GTp0KGlpaXz33Xf5rlXa9xVClC6ZBF76MjM17NyTO0dUJoALUbgSbaPy7LPP8uyzzxIXF0dUVBRarRZXV1c8PT1LFJSdnR12dnYkJyeTkZGBtXXBbzx584/y5jYZ4plnnmHjxo26ek/u7u5A0fOksrKySExMxMLCgmrVqhl8XyFEySSckEngpW31phskp+Tg4WZF+1ayj58QhSmVWX4uLi60aNGCli1bljhhgtzhsbxVc3lDZ/dLSEggLi4OJycnXaJTmJMnTzJ79uwCxTfzWFpaArkVzgGqV6+Oq6srcXFxxMfHF2ift6VK48aNdUN5QojylXUvgdQrEYBsn1JarlxLYd0vNwF49SUv1GoZ8hSiMCX65A8JCeHdd9+lZ8+etGjRglatWtG3b1/mzZvH9evXSxRYQEAAkLvi7b8CAwNRFIXOnTs/8BoJCQls2rSJFStWFDoHKa/w5v3FM/Puu2fPngLtd+/eDfDQ+wohyk5eL5Nto3pY1qhu5GgqvhyNwryvLqHRKHTp4EKXxwpfPSyEKEHS9PXXXzN48GB+/vlnIiMjycrKIj09natXr7Jq1SqefvrpEu3TNmjQIGxsbFi1ahWnT5/WHb969SpffPEFAGPHjtUdj4mJITw8XDdsB7nJjaenJ9euXWPBggX5Eqe9e/eydOlS1Go1o0eP1h0fPnw4ZmZmfPHFF1y9elV3/PTp06xatQorKyteeOEFg5+XEKJkZJPe0rXp10guhaVgb2fO1PGNjB2OECbNoDlNu3bt4uuvv8be3p4JEybw+OOP4+HhAcCtW7fYvXs33377Le+//z4NGzakdevWxb6Hh4cHs2bNYvbs2YwYMQJ/f38sLS0JCgoiMzOT6dOn56sBtWjRIrZs2cKAAQN0tZcsLS1ZuHAhL730EitWrGD37t34+Phw8+ZNQkNDMTc3Z86cOfl6mpo3b8748eNZsmQJ/fv3p0OHDmRlZXH8+HG0Wi2ff/55oYUvhRBlR9FquXfkFFE/bidq02+ATAIvDZFR6SxfFwHA5Jca4uIs29EI8SAGJU3ff/895ubmfP/99zRv3jzfuQYNGjBu3DjatWvHiBEjWLZsGUuXLjUouMGDB+Ph4cGyZcs4c+YMarWapk2bMmbMGHr16qXXNVq1asWvv/7KkiVLOHz4MPv27cPJyYk+ffowduzYAvEDvPbaazRs2JA1a9Zw/PhxrK2t8ff3Z/z48fj7+xv0XIQQxZd6JYKbqzcTtX476TeidMdt6nvi8XQPI0ZW8SmKwvyvL5GVpaXNI9V4sqeHsUMSwuSpFEVRivugNm3a0LJlS1asWPHAdqNHj+bChQscP37c4AArkh49cv+IFzYfSgihH016Brc37+Lmyp+5dzBYd9zc0R6PAb3wfP5panRuj+q+PSOrkj/2RnPiTMGFKsWVnJzN0ZP3sLI0Y803bfH0kI15RdWl7+e3QT1NFhYW+Ta5LYqtrS0G5GRCiEpGURQuz/mKu/sf8gVKUUgOuUJO4j/bFJmZ4do7gNov9Mf9qe6obap2wcXUtBw+/TIUTSnW1n35hfqSMAmhJ4OSpq5du7Jz507Cw8Px8vIqtE1cXBzHjh2ja9euJYlPCFEJRK7ZQtjHS/Rub1PPkzqjB1J75LPY1JFCi3kuXEpCo4Ua1S15bkDtEl+vupMlvboaXutOiKrGoKRp5syZhISE8OKLL/LOO+/w+OOPY2FhoTt/8uRJ5syZg6OjI9OnTy+w3F9qHAlRdaSGXSfk9Y8AqD/5BZwD2j2wvZVbDao/1hqV/J0o4OzF3H0z2zxSjWED6hg5GiGqHoOSphdeeIGMjAzi4uKYPn06FhYW1KxZEzMzM2JjY0lNTdW1zRsnzKNSqbhw4ULJohZCVAja7GzOjHwDTUoazp3b03TB21V2LlJpOP9P0tS8iZORIxGiajIoabpy5YruvxVFISsrS+9iljLHSYiq48rcb0g4cRbzao60XPWZJEwloNEohFzKTZr8mjgaORohqiaDkqbQ0NDSjkMIUcncO3ySsHnfAtB8yRyZm1RC126kkpauwcZGTcO6dsYOR4gqqUQb9gohqp7spBRi/ziINiur6EYKXP7gS9BqqT3yWWoN7lt+AVZS50Nze5ma+TjI3nBCGEmJk6azZ89y/Phx7ty5g6+vL4MHD2bfvn088sgjODvLTtlCVDYhUz7k1rpf9Wpr27AOzb6YVcYRVQ3n8uYz+cp8JiGMxeCkKSoqihkzZnDq1CndsX79+jF48GC++eYbwsLCWLBgAT179iyVQIUQxpcdn8jtn38HoEa3RzGzKPpPiJmVJd7vv4a5g315hVepnQ9NBGQ+kxDGZFDSFB8fz4gRI4iKisLHx4dOnTrlqw5er149zp8/z+uvv87PP/+cb484IUTFdWvDDrSZWTi28MV/1ypUKhkmKg/34rO4dTsDlQqaekvSJISxGFQI5dtvvyUqKooJEybw66+/8uabb+Y7v3DhQt5//31ycnL47rvvSiVQIYTxRa76BYDaowZKwlSO8uYzNahrh4O9TEUVwlgMSpoCAwOpV68er732WpFthg0bRuPGjfn7778NDk4IYTqS/g4l8XQIKgsLPIc9ZexwqhTd0Jyv9DIJYUwGJU3R0dF6Dbk1aNCAmJgYQ24hhDAxN1fn9jK5P90dSxdZ5FGedJPAZT6TEEZlUNLk4OBAVFTUQ9tFRkbi4OBgyC2EECZEk5nFrR+3AVBn1EAjR1O1ZGVruRSWu4GxTAIXwrgMSpratm1LSEgIJ0+eLLJNUFAQFy5coE2bNgYHJ4QwDTE79pJ9NwFrT3dcH+9k7HCqlMvhyWRlK1RzsqB2TRtjhyNElWZQ0vTKK69gZmbG+PHjWb16ta5CuEaj4ebNm6xdu5YpU6ZgZmbG6NGjSzVgIUT5u7l6MwCeI/rLVijl7N/6TI4y+V4IIzNoGUazZs34+OOPmT17NvPmzQNyN+LduXMnO3fuBMDMzIxZs2bRqlWr0otWCFHuMm5FE7vrEAB1XnzWyNFUPXmb9MrQnBDGZ/Da1WeeeYYWLVqwatUqgoODuX37NhqNBldXV9q3b88LL7xAs2bNSjNWIYQRRK7dClotzp3aYte4vrHDqVIURdGVG/CTSuBCGF2JCn40aNCAOXPmlFYsQggjSwkNJ3bXIRStojt247uNQG5tJlG+bkdncDc+C3NzFb6NpLK6EMYmVdKEEOQkp3D5o2+I+N8alJycAufVdrbUHNjbCJFVDoqi8PP2W9yOzijW4/Lae3vZY2Ulc8mEMDa9kqavv/7a4BuoVComTZpk8OOFEGVHURRub9rJhRnzyIzKralWo6s/VjXddG1UKhUeA3tjbm9nrDArvH1H4vjyu3CDH9+ymQzNCWEK9E6a7l+1oShKgZ/vl3cur50kTUIYl6Io3Fz5M9HbAvP9vmbeiSPprxAAbL3q0mzxu7g90cVYYVZKiqKw9ucbAHRo64xX/eIln9bWagY8UassQhNCFJNeSdMrr7xSYKlrYGAgYWFh+Pj40KNHD2rXro1arSYmJob9+/dz6tQpWrVqxZAhQ8okcCGEfhStlgtvfErE/9YUet7M2opGb4+n4bSXUFtblXN0ld/JM/FcDk/B2sqMWa/7Us3JwtghCSEMpFfSNHXq1Hw/BwYGsnTpUiZNmsSrr75aoP3LL7/MqlWrmD9/viRNQhiRJjOLv8e8xe1NuaVAvGaMw867ge68ysyMGl3aY1NXejLKytqfbwLQr3dNSZiEqOAMmgj+f//3fzRu3LjQhCnPqFGj2LZtGytXrmTAgAEGByiEMEx2UgqnBk3i7r5jqCwseGTFp3gO62fssKqUC5eTOHU2AbVaxdBnahs7HCFECRlUETwsLAwvL6+HtqtTpw43btww5BZCiBLIjk/kWPcR3N13DLW9Le22fSsJkxGs+6eXqVcXNzzcrI0cjRCipAzqaapRowahoaEFJoTfLysri7///ht3d/cSBSiEKL7r364n6e+LWLrVoP22ZTi18TN2SFXO9ZtpHDwWB8DwgXWMHI0QojQY1NPUpUsXrl+/zscff0xOITVd0tPTmTlzJtHR0TzxxBMlDlIIUTwxvx8AwPv9KZIwGcmPW26iKBDgX4MGdaVcgxCVgUE9TRMnTmTfvn2sW7eOP/74gw4dOuh6lKKiojhy5AiJiYl4e3vz8ssvl2rAQogHy45PJP7YGQBcewcYN5gqKiYuk137ogF4flBdI0cjhCgtBiVNrq6urFu3jo8++oj9+/ezffv2fOfNzMzo168f77zzDvb2UvpfiPIUt+coaLXYN/HCtp6nscOp0G5EpvHR4lDS0jTFelxKWg45OQqtmjvh5ysb7QpRWRi8jYqnpydLly4lOjqa4OBgYmJiUKlUuLu78+ijj1KjRo3SjFMIoaeYPw4C4Nq7s5Ejqfg2bI3k4uVkgx//4tB6pRiNEMLYSrz3nLu7O/36yaocIUyBoijE/nkIkKG5ksrJ0XLgaCwA08Y3omG94s1LquZkQf06MpdJiMrE5DfsDQ4OZunSpVy8eJGMjAx8fHwYOXIkffv21fsa165dY9myZQQFBREXF4etrS3Nmzdn1KhRBAQU/GCZO3cuP/zwQ5HX++CDDxg2bJhBz0eIspR89hKZt2NR29rgHNDO2OFUaKfOJpCYnEM1Jwue7lMLc3XhK4WFEFWHSSdN27ZtY8aMGZibm+Pv749arSYoKIipU6cSFhbGlClTHnqNU6dOMXbsWNLS0qhfvz5dunQhOjqaw4cPc/jwYWbMmMFLL72U7zEXLlwA4IknnsDcvOBLVL9+/VJ5fkKUtphduUNzNbo9itrK0sjRVGx7DuX2MnV9zFUSJiEEYMJJU1xcHLNnz8bGxoa1a9fSrFkzAMLDwxk5ciRLliyhR48euuOFycnJYcaMGaSlpTF9+nRefvllXV2pI0eO8Morr7BgwQICAgLw9vYGcoc3QkNDqV69Ol988UWZP08hSlPsrn+G5nrJ0FxJZGdrORiUW2OpR4CrkaMRQpgKg+o0lYd169aRkZHBiBEj8iVGXl5eTJs2DUVRWL169QOvERwcTGRkJM2bN2fcuHH5CnF27NiRoUOHotVq2blzp+54REQEqampD0zGhDBF2UkpxB89DYBbH5kEXhLBf8WTkppDDWdLWjR1MnY4QggTYbJJ04EDucX5evbsWeBcz549UalU7N+//4HXSE1NpXnz5nTuXPgHSN4wW0xMjO7YxYsXAfDzk4KAomK5uzcIJScHO+/62DaUCtQlsfdQ7t+Ebh1dUcvQnBDiHyY5PKcoCmFhYQA0bty4wHknJydcXFyIjY0lOjq6yK1aHn/8cR5//PEi73P27FkAPDw8dMdCQkIAMDc3Z/r06Zw8eZL4+HgaNGjAoEGDeP755zEzM9lcU1RhulIDMjRXIpmZGg4dvwvI0JwQIj+9Pv21Wm2J/hVXYmIimZmZ2NnZYWtrW2gbNzc3IHfukyEuXbrEb7/9hkqlolevXrrjeZPAv/76a06dOkXz5s3x8fEhPDycuXPnMnnyZDSa4hW6E6KsKYpC7D+TwF1laK5Ejp26R1q6BjcXK5r5SGFKIcS/9OppKsn8HpVKpUtE9JWeng6AjY1NkW2srKwASEtLK3ZMd+/eZcqUKWg0Gp599ll8fX115/JinThxIpMnT0atVgO5w3YTJ05kz549rF69mjFjxhT7vkKUlZQLYWRE3sHM2ooandsbO5wKLW/VXPdOrpiZydCcEOJfevU0KYpi8D9Depryhr/un7hdlOJePzo6mpEjRxIREYGfnx/vvfdevvN79uxhx44dvPbaa7qECaBJkybMmjULgLVr1xbrnkKUNV2pgS7tUdtYGzmaiis9Q8PREzI0J4QonF49TaGhoWUdRz52drlVdDMyMopsk5mZCVDk8F1hLl++zPjx47l16xbNmzdnxYoVBXqz7O3tC51HBdClSxfUajW3bt0iPj6e6tWr631vIQx1a8MOEk+df2CbmJ37ANk6paSOnrhLRqaWWh7W+DZ2MHY4QggTY5ITwe3s7LCzsyM5OZmMjAysrQt+c85b8ZY3t+lhjhw5wpQpU0hJSaFTp0589dVXuuRMXxYWFjg5OXHv3r0HJnRClJaUy9c488J0vdtLqYFc0bEZ7NoXTY5GeWhbM5UKMzNQq1Uc+Kc2U/dOrnr1dAshqpZSSZryhuLyaLVaMjMziYuLY9++fYwaNapY11OpVDRu3JgzZ84QHh5eYE5VQkICcXFxODk5Fbly7n7bt29n5syZ5OTkMGjQIObMmVNope9Lly6xatUqHB0defvttwucT01N5d69e1hYWMiGxKJcRK3fDoCDnzduT3R5YFvHlk2xa1y/HKIyffO/vkzw6XiDH98jQL8vY0KIqsXgpGnt2rX88MMP3Lp166GryYqbNAEEBARw5swZAgMDCyRNgYGBKIpSZP2l++3du5e33noLjUbDq6++yuTJk4tsa2VlxebNm7GwsOCVV17B2dk53/mtW7cC0L59eywtZYsKUbYUReHWP0mT14xxeA6TjbH1cSMyjeDT8ahU0K9XzYdM5lbQakGrVdBqFTRa8G5oT+OG9uUWrxCi4jAoadq5cydz5859aDsXFxf69OljyC0YNGgQy5cvZ9WqVQQEBNC6dWsArl69qtveZOzYsbr2MTExJCcn4+DgkK8cwdtvv41Go2HChAkPTJggt9jlo48+yrFjx5g5cyaLFy/WDeGdPXuWL7/8EpVKxcSJEw16TkIUR+KJc6SF30Bta4P70z2MHU6FsXlnFACPtavBjMneRo5GCFGZGJQ0bdiwAZVKxbRp0xg2bBi//vorH3/8Mfv27cPGxobDhw8zd+5cNBoN48aNMygwDw8PZs2axezZsxkxYgT+/v5YWloSFBREZmYm06dPz1cqYNGiRWzZsoUBAwYwb948AL7//nsSEhIwNzfn5s2bvPHGG4Xeq3Xr1gwfPhyATz75hOeff54DBw7w+OOP88gjj5CSksKpU6fQarW8/fbbtG3b1qDnJERx3Fq/DQD3p3tgbqf/goeqLC1dw+977gDw7JO1jByNEKKyMShpunTpEg0aNODll18GcpMOrVZLcHAw/fr1o2/fvri6uvLCCy+wfPly3nnnHYOCGzx4MB4eHixbtowzZ86gVqtp2rQpY8aMyVeQsigHD+Yuw87JyWHHjh0PbJuXNHl6erJ582a+/fZb9uzZw6FDh7CzsyMgIICXXnqJ9u2lBo4oe9qcHKI25e6J6Dn8aSNHU3H8uT+a1DQNtWvZ0K6lrG4VQpQulXL/DG49+fn50b17d7766isgd/l/y5YtGT16NDNmzNC1GzBgAOnp6fzxxx+lF7EJ69Ejdwhlz549Ro5EVHSxfx4i+MmxWLpUp8eNQ5hZWBg7JJOnKAovvnqKq9dTmTLWiyHP1DZ2SEKICkLfz2+DNlGzt7fX1UmC3AnUrq6uXL16NV+7unXrcvv2bUNuIUSVdmt9bs9ozUFPSMKkpzPnE7l6PRVrKzOe6OHx8AcIIUQxGZQ0+fr6cubMmXy1iry8vDh79my+0gOxsbGyykyIYtKkpXNn658A1JIVc3rb/FvuBPBeXd1xsDfJEnRCiArOoKTpySefJDExkVGjRnHixAkAOnXqxL1791i8eDEpKSn8+uuvnD59mgYNGpRqwEJUdtE79qFJScOmvifVO7QydjgVQuzdTA4G5e4ZJxPAhRBlxaCk6dlnn6Vbt26cOXOGVatWATB06FCcnZ357rvvaNeuHTNnzkSlUhlUo0mIqixqQ25tplrP9ZOq1Hr69fcoNFp4pJkTjRpIjSUhRNkwqA9brVbzf//3f+zevZvs7Gwgd57TmjVrmDNnDn///TfOzs6MGjWKvn37lmrAQlRmWfcSiPnjEIAUs9RTYlI223blzp2UXiYhRFkq0cD/448/nu/nRo0a8cMPP5QoICGqqqx7CVyZ+w1KdjaOjzTBoWkjY4dk8g4cjWXB/10hPiEb1xqWdOngYuyQhBCVmMyWFMKIFI2GuD1HubnqF6J/DUSbldtzW/vFZ40cmWmLT8zii2/D2HModx5T/Tq2vP9GE8zNDZpxIIQQetEraXr++edRqVQsWLAADw8Pnn/+eb1voFKpWLt2rcEBClEZKYrCnS1/cnHmZ6Rfi9Qdd3ykCXXHDqHuK8OMGJ3pyMzSsnztNe4lZP97UFE4/lc8CYnZqM1g+MC6jB5WD0sLSZiEEGVLr6Tp1KlTqFQq0tPTdT/rSyayCpFf8oUwQqbO5e7eIAAsqjvhOfxpar/4LE6tmho5OtPy+547rN8SWei5hvXseOd1H3wbOZRzVEKIqkqvpGnNmjUA1KpVK9/PQohcGXdiiT96GpVajcr8n39qNSqz/L0fMTv3E/HNWpScHMysLGn4xlgazRiH2tbGSJGbtj2HYgDo3smVJt7/JkfVHC3o0dlNepeEEOVKr6Tpv/utyf5rQuR34qmXSfr7ot7t3Z/uQdPP38a2YZ0yjKpiuxufxZnziQBMHN0QDzdrI0ckhKjqDJoIPnLkSDp27Mgrr7zywHaffvop+/fvZ9euXQYFJ0RFkHwxnKS/L6IyN8epbXPQaNDmaFBycuA/WztaOFej0VvjcO0VYKRoK479R2JRFGjm4yAJkxDCJBiUNAUHB+Ph8fC9nS5dukRUVJQhtxCiwsjb8sSl52O03/6dkaOpPPYezl0Z172Tm5EjEUKIXHolTdOmTSM6OjrfsSNHjjxwFV1ycjJXrlyhdm3ZaVxUbne27AbAo38vI0dSecTezeTshdyhuW6dXI0cjRBC5NIraerevTtvvPGG7meVSsXdu3e5e/fuAx+nVquZNGlSySIUwoSlXb9F0l8hYGaGe7/uxg6n0tj3z9Bc8yaOuLlYGTscIYQA9EyannrqKTw8PNBqtSiKwosvvvjAOU0qlQorKytq166Ns7NzqQYshCmJ3prby+TcqS1WbjWMHE3lsfefopU9AmRoTghhOvSe09S2bVvdfz/11FO0b99eVtGJKu/O1ryhuccf0lLo605MBudDk1CpoOtjsi2KEMJ0GFTk5Ny5c1LlW1R5mdFx3DuSW+hVkqbSs+9Ibi/TI82ccKkhQ3NCCNNhUNJ0+/Zt6tSR+jKiaovetgcUBac2ftjUqWnscCqNf1fNyQRwIYRpMShpqlWrFjdv3iztWISoUPJKDXgMkFVzpSXqTjoXLydjZgZdH5OkSQhhWgxKmj766COioqKYOHEiR44c4e7du+Tk5KDVagv9J0Rlk52QRNy+44AMzZWmPf9MAG/lVw3n6pZGjkYIIfIzqLjlJ598go2NDfv27WPfvn0PbKtSqbhw4YJBwQlhqmJ27kfJzsa+aSPsfRoaO5wKIe5eJpfDU0jP0JCRqSUjQ0NqWg637mRw81YaN26lk5CYDUhtJiGEaTIoabp4Uf89tpT/bCMhRGUgq+aK58/90Xz29WUyMh/e8+xV305KDQghTJJBSVNoaGhpxyGEychJTUObkVnkeW1mFjF/HATA4xlJmh4kM0vL/5aHsfX32wDUrmWDi7MlNtZqrKzMsLVW4+5mTV1PW+p42lC3lg22tgb9WRJCiDInf52EuM+drbs5/dxrKBrNQ9va1PPEsVXTcoiqYoq6k8678y5wOTwFlQpeHFKX0cPqo1arjB2aEEIYpEyTpqysLA4cOMDjj8u3cVEx3Pk1UK+ESaVW0+C1UahUVS8BOHEmnhXrIsjOefBQW2RUOqlpGpwczJk9vQmPtpHdAYQQFZvBSdPevXtZt24dkZGRZGVl5Zu7pNVqyczMJCUlBa1WW6w5UEIYU/L5ywC03vgVHs/0LLqhSoXKzKDFpxXeinURnA9N0qttMx8HPnyrKe6u1mUclRBClD2DkqYjR44wadKkh07ytrW1xd/f36DAhChvikZDysUwABxb+KJSq40ckem5G59FyKXchOnDt5piY1104mhtpaZ5E0fMzatmcimEqHwMSppWr16NoigMHz6cIUOG8Oeff7J06VI2btyIubk5hw4d4v/+7/+oXr06n3/+eWnHLESZSA27jjYzCzMba2wbSsX7whwJvouigG8jB6nYLYSocgzee87T05PZs2fj6+tL165d0Wq1REVF0aRJE8aNG8f8+fO5desWK1asKO2YhSgTySFXAHBo2qjKDr09zOFjcQB0erSGkSMRQojyZ9AnQ3JyMr6+vrpJsF5eXgD5ilj26tWLBg0aEBgYWAphClH28uYzOfh5GzkS05SWruHk3/EABPhL0iSEqHoMGp6zs7PLtz2KnZ0dzs7OhIeH52vXuHFjDh06VKIAg4ODWbp0KRcvXiQjIwMfHx9GjhxJ37599b7GtWvXWLZsGUFBQcTFxWFra0vz5s0ZNWoUAQEBBdprtVo2b97Mjz/+SEREBBYWFrRp04aJEyfi5+dXoucjTJeup6mZJE2FOfHXPbKyFWq6W9Ownp2xwxFCiHJnUE9Tw4YNCQkJQXPf0uyGDRty/vz5fO2SkpJKtCR727ZtjBw5kuDgYJo2bUq7du0ICQlh6tSpfPXVV3pd49SpUzz77LNs3rwZKysrunTpQt26dTl8+DBjx44tdPjwgw8+YNasWURGRvLYY4/RoEED9uzZw3PPPcfhw4cNfj7CtCWH/NPT1FySpsIcOn4XyO1lqoqlFoQQwqCkqVevXsTExPD6669z7do1APz9/YmOjmb9+vUAnDx5khMnTlC3bl2DAouLi2P27NnY2NiwceNGVqxYwbJly9i6dSsuLi4sWbKEkJCQB14jJyeHGTNmkJaWxvTp0/njjz/45ptv+Pnnn1m5ciUWFhYsWLCAy5cv6x6zZ88eNm7ciLe3N7t27eLrr79mw4YNfPXVV2i1WmbOnEl6erpBz0mYLk16BqlXrgPg0KyxkaMxPTkahaMncpOmTo+6GDkaIYQwDoOSpuHDh+Pn58fu3buZP3++7pi1tTUffvgh/v7+vPDCC2i1WoYMGWJQYOvWrSMjI4MRI0bQrFkz3XEvLy+mTZuGoiisXr36gdcIDg4mMjKS5s2bM27cuHzfjjt27MjQoUPRarXs3LlTd3zlypUAzJgxg+rVq+uO9+7dm379+hEbG8tvv/1m0HMSpisl9CpotVjUqIaVh6wK+69zFxJJSs7B0cGcFk2djB2OEEIYhUFJk5WVFT/++CMzZ87UzQmqUaMGS5cuxdPTk8TERMzNzRkxYgTDhg0zKLADBw4A0LNnwQKDPXv2RKVSsX///gdeIzU1lebNm9O5c+dCz9evXx+AmJgYIHeC++nTp7G1taVDhw4F2udVNn/YfUXFo5sE3sxbhp4Kcfh47qq5x9rVwFy2QRFCVFEGVwS3tLRk1KhR+Y75+/sTGBhIXFwcDg4OWFlZGXRtRVEIC8stMti4ccGhEicnJ1xcXIiNjSU6Ohp3d/dCr/P4448/cAuXs2fPAuDh4QFAeHg4Wq2Whg0bYm5e8KXJWyV4/3CeqBx085lkaK4ARVHyzWcSQoiqqlT2nrt9+zYxMTGo1Wpq1qyJi0vJ5jwkJiaSmZmJnZ0dtra2hbZxc3MjNjaWuLi4IpOmB7l06RK//fYbKpWKXr16Af/2OLm6Fj48k3c8Li6u2PcTpk3KDRTt6vVUou5kYGmhol0r2T9OCFF1GZw0aTQafvjhB3788Udu3ryZ71yTJk148cUXeeaZZwy6dt5EaxsbmyLb5PVipaWlFfv6d+/eZcqUKWg0Gp599ll8fX3zXauo+1pbW+eLT1Qe/5YbkJ6m/8rrZWr7SHVsbWRrGSFE1WVQ0pSdnc348eM5evQoiqLg7u5OzZo1URSFW7duceHCBWbOnMmxY8f49NNPi319s3+qMeszt+T+elH6iI6OZsyYMURERODn58d7772nO6f+Z6+xh91Xq9WiKIrMfakkshOSyIi8A0hPE+QWsbx/X8lDuirgsmpOCFG1GZQ0rVq1iiNHjuDj48PcuXNp3rx5vvNnzpzh3XffZevWrfj7+9O/f/9iXd/OLrdwXkZGRpFtMjMzAYocvivM5cuXGT9+PLdu3aJ58+asWLEiX69S3rWKum/ecVtbW0mYKpG8oTnrOjWxcHIwcjTGExOXybz/XSL4dHyBcyoVdGwv85mEEFWbQavntm7dipOTE99//32BhAmgZcuWfP/999ja2urqNhWHnZ0ddnZ2JCcnF5nA5M0/cnNz0+uaR44cYdiwYdy6dYtOnTqxevVqnJzyL53OmxtV1Jyl2NhYoOg5T6Ji0g3NVdFeJkVR+HN/NCMnnyw0YQLo2dmNGtUtyzkyIYQwLQb1NEVGRtK5c2ecnYueFOrq6sqjjz7K0aNHi319lUpF48aNOXPmDOHh4fnqNAEkJCQQFxeHk5OTXpPAt2/fzsyZM8nJyWHQoEHMmTOnyNVxZmZmXL16Fa1WqxsmzJO3os/bu2p+uFZW/5YbqHrzmZKSs1mw5Ap7D+d+IWji7cA7r/lQyyP/vD4rS9nAWAghDEqaatSooet1eZC0tDQcHR0NuQUBAQGcOXOGwMDAAklTYGAgiqIUWX/pfnv37uWtt95Co9Hw6quvMnny5CLb2tjY0K5dO44fP87x48cL1GravXs3AF26dDHgGQlTVRV6mjQaha9XhucmR/9OVyItPYf0DC1qMxj1XD1eGFJP6jAJIUQRDPr6OHDgQP7++29+//33ItucOXOG4OBgg1fQDRo0CBsbG1atWsXp06d1x69evcoXX3wBwNixY3XHY2JiCA8P1w3bQe4w29tvv41Go2HChAkPTJjyvPDCCwB89NFH+RLDP//8kx07duDq6mrwcxKmR1GUfIUtKyOtVmH+15f5adst7t7L4m78v//SM7TU9bTh2wWtGT2sviRMQgjxAAb1NA0bNozz58/zxhtvcOzYMZ588knq1auHWq0mOjqa/fv3s2LFCtzc3GjTpg1BQUH5Hl9Yte3/8vDwYNasWcyePZsRI0bg7++PpaUlQUFBZGZmMn36dF2pAIBFixaxZcsWBgwYwLx58wD4/vvvSUhIwNzcnJs3b/LGG28Ueq/WrVszfPhwILcg5jPPPMOvv/5Knz59ePTRR4mPj+f06dNYWFiwcOFCLC1lbkdlkXk7huz4RFRqNfa+DY0dTqlTFIVFS8PYGXgHMzN461UffLzsdefNzFTUrW0ryZIQQujBoKTpscceQ6VSoSgKmzZtYtOmTQXaKIpCeno648ePz3dcpVJx4cIFve4zePBgPDw8WLZsGWfOnEGtVtO0aVPGjBmjK0j5IAcPHgRyN+7dsWPHA9vmJU0A8+bNo0WLFmzatIlDhw7h6OhI9+7defXVV2nSpIlesYuKIa+XybZxPdTWhlWwN1WKovD1inC2/h6FSgXvTvWlV9fiF4IVQgiRy6CkqV27dqUdR5ECAgJ0+9s9yLx583Q9THm2b99u0D3NzMwYMWIEI0aMMOjxouKozEUtl/0QwcZfbwHw1mRvSZiEEKKEDEqafvjhh9KOQwijSD6fmzQ5VrJJ4HsPx/LDTzcAmDq+EU/1qmnkiIQQouKTdcSiSks+fwkA+0o2CXz9ltytjYY/W5uBT3oaORohhKgcSmXDXiEqgujf9hE6ayFKVrbuWNrV3OSiMvU0hVxK4uLlZCzMVQwbUMfY4QghRKUhSZOoEhRF4eJb80m9dK3AOes6NbFtWHmSi1925M5j6tnZjerVZKWnEEKUFkmaRJVwd98xUi9dQ21vS7tfv0X1z+bMAA5NG+X7uSKLu5epq+49sJ8MywkhRGmSpElUCdeX/ghA7RH9qdG5vZGjKTu//nGbnByF5k0c8W1UdTcfFkKIsiATwUWllx55h+htewCoN374Q1pXXNnZWn79PQqAQdLLJIQQpc6gpOnkyZOlHYcQZebmik0oGg3OndtXynpMefYdieVeQjYuzpZ06eBi7HCEEKLSMShpGjFiBL179+bbb78lOjq6tGMSotRos7K4sTy3Yn298cOMHI1h4u5mcjs646HtftqeOwG8f99amJtLJ7IQQpQ2g/6yduzYkcjISBYvXkz37t0ZO3YsO3fuJCsrq7TjE6JE7vwaSOadWKw8XPHo/7ixwym2tLQcRk05xeCxx5kw4y+277pNalpOgXb3lxl4prcUshRCiLJg0ETwFStWEBMTw6+//srWrVs5fPgwR44cwdHRkSeffJIBAwbQvHnz0o5ViGK7vnQ9AHXHDsHMwsLI0RRfaFgyCUm5daXOXUzi3MUkvlgWRqdHa+BS/d9yAmcvJAFSZkAIIcqSwavn3NzcePnll3n55Zc5f/48W7Zs4bfffuPHH39k/fr1NGrUiIEDB/L000/j7OxcmjELoZfk85e5dzAYlVpNnZeGGDscg4SGpQDQrmV12rasxs7AaK5HprHnYGyh7aXMgBBClJ1SKTng5+eHn58f77zzDsHBwezYsYMtW7Ywf/58FixYQI8ePXjhhRdo27ZtadxOCL1c/za3l8n96R7Y1PYwcjSGuRSWDEDrFtV4fmBdhj9bh5BLyRw7eZfsHCVf20YN7KXMgBBClKFSq9OUlZVFYGAgf/75J4cOHUKr1WJmZkbNmjXZtWsXf/75J127duXzzz/H3t6+tG4rqhhFqyVqww6uLv6erLvxD2ybeTu3N6behIpbZiD0n6TJt3FuMqRSqfDzdcTP19GYYQkhRJVU4qTpxIkTbN26lV27dpGamoqiKNSpU4exY8fy7LPP4u7uzoULF5g7dy779+/n/fffZ+HChaURu6hgsu7Gc+1/a7BrWBfnLu2xraf/UJKiKMTuOsilWYtIOhuq9+McWzWjRtdHDQnX6JJSsrl1O3fVnE8j+aIhhBDGZlDSdP36dbZu3cq2bduIiopCURSsra3p168fAwcOxN/fP1/7pk2b8n//93906NCBffv2lUrgouIJfXsBN7//WfezTX1PanT2x8GvMagesJBTUYjesZd7B4MBMHdywOvNl3Hp8RgqleqB97TzbfjQNqbq0j/zmTxrWuNoX/EmsQshRGVjUNLUu3dvVCoViqLg5+fHwIED6dev3wOH3RwdHTEzM8PW1tbgYEXFlR2fyK0NOwBwbOFLcsgV0iNuERmxWe9rmFlZUn/SCLxmjMOyRvWyCtVkhF75Z2hO5ikJIYRJMChpcnJy4plnnmHgwIH4+Pjo9Zjs7Gy+/fZb6tWrZ8gtRQUX+cNWtOkZOPh50+nkVjQpqcQH/cXdgydIvxH10MdbebjS4NWR2NSpOjWI8iaB+0jSJIQQJsGgpOnw4cNYFLPmjaWlJR07djTkdqKCUxSF68tyV7LVe2UYKpUKcwd7XHsF4NorwMjRma7/TgIXQghhXAZVBNc3YcrKymL37t2G3EJUInf3HyP10jXU9rZ4Pv+0scOpEOITs7gTkwmAj5dMAhdCCFNg8Oq5vXv3sm7dOiIjI8nKykJR/q0Zo9VqyczMJCUlBa1Wy8WLF0slWFEx5dVL8hz+NOYOkgDoI28SeF1PG+xsS60yiBBCiBIw6K/xkSNHmDRpUr5EqTC2trYFVtKJqiXjdgzRv+4BcofmhH5kaE4IIUyPQcNzq1evRlEUhg8fztatW5k4cSJmZmb89NNPbNmyhWnTpmFjY0P16tX5/PPPSztmUYHcXPkTSk4O1Tu0wrGFr7HDqTAuyco5IYQwOQYlTefOncPT05PZs2fj6+tL165d0Wq1REVF0aRJE8aNG8f8+fO5desWK1asKO2YRQWhzcnhxvJNANSVXqZiCZWVc0IIYXIMSpqSk5Px9fXVFQ308vIC4MKFC7o2vXr1okGDBgQGBpZCmKIiitm5n4zIO1jUqEbNgX2MHU6FEXcvk9i7WZiZQeOGMgdMCCFMhUFJk52dHVqtNt/Pzs7OhIeH52vXuHFjIiMjSxahqLBu/DMBvM6ogaitrYwcTcWRNwm8Xm1bbG3URo5GCCFEHoOSpoYNGxISEoJGo8l37Pz58/naJSUlVdgtLETJZMcnEvvnYQDqvvyckaOpWHSTwGVoTgghTIpBSVOvXr2IiYnh9ddf59q1awD4+/sTHR3N+vW5vQsnT57kxIkT1K1bt/SiFRVGSuhVAKxre2DnJe+B4rgkK+eEEMIkGZQ0DR8+HD8/P3bv3s38+fN1x6ytrfnwww/x9/fnhRdeQKvVMmTIkFINWFQMKZdzk2k77wZGjqRiURRFNzwnk8CFEMK0GJQ0WVlZ8eOPPzJz5kwCAnK3wahRowZLly7F09OTxMREzM3NGTFiBMOGyaqpqij1n6TJXpKmYom7l8Xd+CzUZtC4gZ2xwxFCCHEfg0sNW1paMmrUqHzH/P39CQwMJC4uDgcHB6ysZPJvVaXraWpc37iBVDCh/9Rnql/XDisrmQQuhBCmxKCkafTo0XTv3p2uXbtSp06dAuddXFxKHJio2FKvRABg51M1e5pu3U5n8bdhxN7NLNbjEpKyAZnPJIQQpsigpCkoKIhjx47xySef4OXlpUugWrVqVeqr5YKDg1m6dCkXL14kIyMDHx8fRo4cSd++fQ26nqIojB49mqtXr3Lw4MFC2+zevZvJkycXeY2+ffuyePFig+5fFSgaDWlh14GqOTx3MyqNKe/8TezdLIOv0faR6qUYkRBCiNJgUNK0a9cuDhw4wMGDBzlx4gTLli3ju+++o1q1anTt2pWuXbvSqVMn7OxKNidj27ZtzJgxA3Nzc/z9/VGr1QQFBTF16lTCwsKYMmVKsa85f/58goKCcHd3L7JNSEgIAO3bty+0XatWrYp936ok/XoU2swszKwssalby9jhlKsbkWm8Outv7t7Lon4dW14d64VaXbwvEvZ25vh4SVFLIYQwNQYlTfXq1WPkyJGMHDmSzMxMjh07xsGDBzl48CBbtmxh69atmJub0759e7p3787zzz9f7HvExcUxe/ZsbGxsWLt2Lc2aNQMgPDyckSNHsmTJEnr06KE7/jDp6enMmTOHLVu2PLTtxYsXAXjvvfdo3LhxsWOv6lIu55YbsGtUH5W66szLuXYjlddm/c29hGwa1rPjy7ktqF7N0thhCSGEKCUGrZ67n5WVFV26dGH27Nns3r2bP/74gyFDhqAoCkePHmXu3LkGXXfdunVkZGQwYsSIfImRl5cX06ZNQ1EUVq9erde1AgMD6d+/P1u2bCl0DtZ/hYSEYGNjQ8OGDQ2KvapLvfTPJPAqNJ/p6vVUpvyTMDVqYMdXHz8iCZMQQlQyBq+eu194eDjBwcEEBwdz4sQJ7t69i6IoANSqZdjwzIEDBwDo2bNngXM9e/Zk1qxZ7N+//6HXSUpKYtKkSajVakaOHMlzzz33wPlQcXFxxMbG0qpVK9RVqJekNFW1Gk1arcKsT0OIT8jGu6E9iz9qgZOjhbHDEkIIUcoMSprCw8M5fvw4wcHBnDx5Ml+S5O7uTr9+/fD398ff35/atWsX+/qKohAWFgZQ6PCYk5MTLi4uxMbGEh0d/cD5SWZmZjz11FNMmDCBRo0aPXQvvLz5TB4eHsyfP5+9e/cSFRWFq6srvXv3Zvz48Tg5ORX7OVUlVa1G07FT97h5Kx17OzVfzG2Bo4MkTEIIURkZlDQ99dRTuv/29PTkySefpH379vj7+1OvXr0SB5WYmEhmZiZ2dnbY2toW2sbNzY3Y2Fji4uIemDTZ29uzcOFCve994cIFAH7//Xfs7e1p164dHh4enD9/npUrV7J3717Wrl2Lq6tr8Z5UFVLVepp+3n4LgKcerykJkxBCVGIGJU15vUpqtRoXFxc8PT3x9PR8YPJSHOnp6QDY2NgU2SavcGZaWlqp3DNP3iTwLl26sHDhQhwccuvl3Lt3j6lTp3Ls2DFmz57N0qVLS/W+lUVOcgqZUTEA2FeBOU0RN1MJ/iseMzMY+JSnscMRQghRhgxKmo4ePUpQUJDu37fffsuyZcswNzfHz89PNzTXunVrg6qCm5nlzk/Xp+aTVqst9vUfZMGCBbz22mvUqlUrX9Lm7OzMZ599Rp8+fdi3bx+RkZEGDT1WdqmXIwCwdKuBRTVH4wZTDn7eHgVAx/Y1qOlubeRohBBClCWDkiZnZ2eefPJJnnzySQBu3rypS6BOnDjBX3/9xbfffouFhQUtWrRg7dq1xbp+Xn2njIyMIttkZuZWWi5q+M5QlpaWeHl5FXrO3d2dpk2bcvLkSUJCQiRpKkRKFZrPlJySwx977wAwuJ/0MgkhRGVXKqvn6tSpQ506dRgyZAj37t1jw4YNrFq1iqSkJE6dOlXs69nZ2WFnZ0dycjIZGRlYWxf8Bh8TkzsE5ObmVuL4iyNvi5i8IUSRX2oV2nPut8DbZGRqaVjPjlbNqxk7HCGEEGWsxElTZmYmJ06c0PU0hYaGoigKiqJQv359unbtWuxrqlQqGjduzJkzZwgPDy9QwDIhIYG4uDicnJxKbR4V5D6XuXPncu/ePRYuXFhosnbz5k0gd3WdKEg3CbySz2fSaBR+2ZE7NDeon2epbx8khBDC9BiUNP3999+6eU1nzpwhOzsbRVGwsLDg0UcfpUuXLnTr1o26desaHFhAQABnzpwhMDCwQNIUGBiIoih07tzZ4OsXxsrKiv379xMTE8Phw4cL1IgKDQ0lNDQUBwcHWrZsWar3rizyCltW9uG5oJN3uR2dgYO9Ob26lG9vpxBCCOMwKGkaOnQoKpUKRVFwcXGhS5cudO3alY4dO5baHKNBgwaxfPlyVq1aRUBAAK1btwbg6tWrfPHFFwCMHTtW1z4mJobk5GQcHBxKNGT33HPP8dVXX/HJJ5/g4+OjqyAeFxfHO++8g0aj4aWXXiq0F6qqU7RaUq9EAJW/3MBP/5QZ6Ne7JtbWUgRVCCGqAoOSJj8/P93GvH5+fqUdE5A7/DVr1ixmz57NiBEj8Pf3x9LSkqCgIDIzM5k+fTq+vr669osWLWLLli0MGDCAefPmGXzfl19+mZMnT3L06FGeeuop2rRpg6WlJcePHyctLY3evXszbty40niKlU5GVAyatHRU5ubYNnz4djWm6Or1VL5ZGU5mZtGrMhXg75BEzMzg2b5Va0NiIYSoygxKmn7++ecCx65evcrdu3dp165diYPKM3jwYDw8PFi2bBlnzpxBrVbTtGlTxowZQ69evUrtPveztLTku+++Y+3atfz666+cOnUKMzMzGjduzODBgxk0aJDMXylC3iRw24a1MbOomEUeV/4YwfHT8Xq17dbRFQ836XEUQoiqQqXkVaosoTfffJMdO3boikNWRT169ABgz549Ro7EOCL+bx0hUz7E7alutNtS8Yp/JiRm039UEDk5Cm9MbPzA/ePM1SratKiGrW2pLEAVQghhRPp+fstffFFqKvqec7sPRJOTo+DTyJ7+T8iwmxBCiPzMjB2AqDx0NZq8Gxo5EsPs3BMNQN8eUk5CCCFEQZI0iVKjqwZeAWs0XbmawpWrKViYq3hcSggIIYQohCRNolRo0jNIv55b7LEilhvYuSd3O5RO/i44OlTMSexCCCHKVqklTV5eXrRt27a0LicqmNSw66AomDs5YOnqbOxwiiU7W8vu/bnb8vTtWXoV5oUQQlQupZY0jR8/nh9++KG0LicqmPsngVe0kgxHT9wlISkbF2dL2reqWAmfEEKI8iOr50SxKYpC5JotuurfAAknzgIVc2jut8Dcobk+3d1RqytWwieEEKL8SNIkii0u8Ahnx75d6DkHv8blHE3J3I3P4vipe4CsmhNCCPFgkjSJYgtfsBwA587tcWrZRHfc3MmBumOHGissg+zaF41GC36+jtStXTr7JgohhKic9EqaCts2pTgGDRpUoscL05F4OoS7e4NQqdU8snIetvU8jR2S3s6HJrHu5xtk5/xbBD80LBmAvj2ll0kIIcSD6ZU0zZ49u0Q3kaSp8ri6aAUANYf0rVAJk0ajMO+rS0TcTCtwztZGTfdOrkaISgghREWiV9K0evVqpk2bRlxcHLVq1cLf37+s4xImKO3aTaJ++h0Ar+ljjRxN8ew9HEvEzTTs7cyZMtaL+xf4+TRywN5ORqqFEEI8mF6fFO3bt2fNmjWMGDGCmJgYBg0aRJs2bco6NmFirn6xCrRaXHt1wvERX2OHozeNRmHVhusAPDegtgzFCSGEMIjedZoaNmzIwoUL0Wq1zJ49G41GU5ZxCROTFXePm9/nzm1r+EbF6mXacyiG65FpODqYM7hfxRlSFEIIYVqKVdyyQ4cODBkyhKtXr/LTTz+VVUzCBEX8349o0zNwbNWMGl0fNXY4esvRKHyf18vUvzZ2tjIMJ4QQwjDF/gSZMmUKlpaW5OTklEU8wgRp0tKJ+Ca32rvXG2MrVMXvwAMx3LyVjpODOYOekl4mIYQQhit20uTs7Mw777xTFrEIE3Vz9Way7yZg06A2Hs/2MnY4esvJN5epDrbSyySEEKIESm3vOVF53Vy+CYAGU0ZhZl5xEo/d+6OJvJ1ONUcLBkovkxBCiBKqOJ+AwihSw66TdDYUlVqN57CnjB0OAGlpOSxZdZV7CdkPbHf+YiIAwwfWwdZGXR6hCSGEqMQkaRIPdGfLLgBqdHsUyxrVjRxNrq+Wh7Nj9x292jpXs2BA31plHJEQQoiqQJIm8UC3f8lNmmo+29vIkeQ6euIuO3bfQaWCcS80wMH+wW/h1s2rYWMtvUxCCCFKTpImUaS0iEgST50HMzPcn+lp7HBITMpm/v8uAzDkmdq8MLiukSMSQghRlchEcFGkO1t2A+Ac0BYrtxpGjgYWLb3C3fgs6texZdwLDYwdjhBCiCpGkiZRpNub84bm+hg5ktyq3nsOxaI2g3en+mJlKW9dIYQQ5UuG50Sh0iPvkHDsLwA8+pft0JxWqzzw/L34LBb+3xUARg6ph29jhzKNRwghhCiMJE2iUHe25g7NVX+sNda13MvkHimpOcyed4ETZ+L1au/tZc+LQ2UekxBCCOOQpEkU6s7msl01l56h4c055zh3MUmv9vZ2at6d6ou5uQzLCSGEMA5JmkQBGXdiuXf4JAAeA0p/25TMLC1vf3yecxeTsLczZ8EHftSpZfvAx9jYqLG0kIRJCCGE8UjSJAqI/jUQFIVq7VpgU7d0C0Pm5GiZPS+Ek2cSsLE2Y8EHzfHzdSzVewghhBBlQb66iwLyVs15DCzdoTmNRuHDhaEcPXEPS0sz5r8nCZMQQoiKQ3qaqri0iEiid+wjJyGR7IRkshOSuHcgGICaA0o3aVq96Tp7D8dibq7i47eb0rp5tVK9vhBCCFGWTD5pCg4OZunSpVy8eJGMjAx8fHwYOXIkffv2Neh6iqIwevRorl69ysGDB8vtvqYo4cRZjj8xhpzE5ALnqrVrgW3DOqV2r9vRGaz9+SYAb73qTYe2xi+WKYQQQhSHSSdN27ZtY8aMGZibm+Pv749arSYoKIipU6cSFhbGlClTin3N+fPnExQUhLt70cvoy+K+pib++N8E9x1DTlIKDs19qP5oSyyqOWLu5IBFNUfcnuxWqvf7ZmU4WVlaWjV3ok+3silhIIQQQpQlk02a4uLimD17NjY2Nqxdu5ZmzZoBEB4ezsiRI1myZAk9evTQHX+Y9PR05syZw5YtW8r1vqYoPugvgp98iZzkVJwD2tFu27eY29uV2f1O/R3P/qNxmJnB6+MaoVKpyuxeQgghRFkx2Yng69atIyMjgxEjRuRLULy8vJg2bRqKorB69Wq9rhUYGEj//v3ZsmULdeo8eMipNO9riu4dOcXxvmNyE6Yu7Wm3fVmZJkw5GoUvvwsHYEDfWnjVty+zewkhhBBlyWR7mg4cOABAz54Ft/Do2bMns2bNYv/+/Q+9TlJSEpMmTUKtVjNy5Eiee+65B85LKq37lqbYPw9x/tU5aNIzSnyt7LsJaLOyqdHtUdptXYra1qYUIizar79HcfV6Kk4O5rw0vH6Z3ksIIYQoSyaZNCmKQlhYGACNGzcucN7JyQkXFxdiY2OJjo5+4PwkMzMznnrqKSZMmECjRo2IjIwsl/uWpuQL4aRdvVlq13N5vBNtf/66zBOmhMRsvlsbAcDLLzTA0cGiTO8nhBBClCWTTJoSExPJzMzEzs4OW9vCK0W7ubkRGxtLXFzcA5MXe3t7Fi5cWO73LU0NXnsRt75d0KZnlvhaKksL7H0blsu8ou/WXiMlNYdGDezo16tmmd9PCCGEKEsmmTSlp6cDYGNTdE+IlZUVAGlpaRX+vg+jUqmw925QbvcrDecuJrJt120Apr7SCLVaJn8LIYSo2ExyIriZWW5Y+vSGaLXaCn/fyiYzU8MnX15CUeCJ7u480qyasUMSQgghSswkkyY7u9zVXBkZRU98zszMHaoqahitIt23slm+LoKbt9Jxcbbk1Ze9jB2OEEIIUSpMNmmys7MjOTm5yAQmJiYGyJ1jVNHvW5mcD01kw9bcyfYzJnvjaC+Tv4UQQlQOJpk0qVQq3eq18PDwAucTEhKIi4vDycmpVCdjG+u+lUVmpoZPvvh3WO6xdrJVihBCiMrDJJMmgICAACC3MOV/BQYGoigKnTt3rjT3rQyWr4vgxq10asiwnBBCiErIJFfPAQwaNIjly5ezatUqAgICaN26NQBXr17liy++AGDs2LG69jExMSQnJ+Pg4FCiobPi3rcq0moVftx8k1t3/h3C1GgU/th7B4AZk2RYTgghROVjskmTh4cHs2bNYvbs2YwYMQJ/f38sLS0JCgoiMzOT6dOn4+vrq2u/aNEitmzZwoABA5g3b1653bcq2rA1kqWrrxV6rk93dzq2l2E5IYQQlY/JJk0AgwcPxsPDg2XLlnHmzBnUajVNmzZlzJgx9OrVq9LdtyIIvZLMt2tyE6Zn+tTEzcVKd87GWk2/3lLEUgghROWkUhRFMXYQlUWPHj0A2LNnj5EjKRtp6RrGvH6KyKh0unZ04aO3mpZLZXEhhBCiLOn7+W2yE8GF6fnyuzAio9Jxc7FixmRvSZiEEEJUKZI0Cb3sPRzLb7vvoFLB7Gm+MtFbCCFElSNJk3ioOzEZfPb1ZQBeGFyXVs2rGTcgIYQQwggkaRIPdCcmgzc+OEdKag5NfRwYM6yesUMSQgghjMKkV88J47oUlsyMD89zNz4L1xqWvP9GE8zNJc8WQghRNUnSJAoVdPIu782/QHqGFq/6dnz+fvN85QWEEEKIqkaSpipOURQSErNJy9CQnq4hPUPDuYtJLF19Fa0W2rWszty3m2JnK28VIYQQVZt8ElZRUXfS+X1vNH/sieZ2TEahbfr29GDGpMYyJCeEEEIgSVOFkJmp4djpeDIzNSW+Vmqahr2HY/jrXGK+4zbWZthYq7GxUWNnY06vbm4Mfaa21GISQggh/iFJUwXw4+abrPjxeqleU6WCto9U54ke7nR+1AVra3WpXl8IIYSobCRpqgA6+bsQGpZCZlbJe5rMVCpaNHOiTzd3PNysSyE6IYQQomqQpKkCaNzQnvmz/YwdhhBCCFGlyQxfIYQQQgg9SNIkhBBCCKEHSZqEEEIIIfQgSZMQQgghhB4kaRJCCCGE0IMkTUIIIYQQepCkSQghhBBCD5I0CSGEEELoQZImIYQQQgg9SNIkhBBCCKEHSZqEEEIIIfQgSZMQQgghhB4kaRJCCCGE0IO5sQOoTGJiYtBoNPTo0cPYoQghhBBCT7dv30atVj+0nfQ0lSIrKyvMzSUPFUIIISoSc3NzrKysHtpOpSiKUg7xCCGEEEJUaNLTJIQQQgihB0mahBBCCCH0IEmTEEIIIYQeZNayidi9ezeTJ08u8nzfvn1ZvHhxOUZU8QUHBzNy5Eg++ugjBg8eXOB8SkoK3333Hbt27SIqKopq1arRrVs3pkyZQo0aNYwQccXysNf36aef5tKlS0U+fufOnXh5eZVliBWKVqvlp59+YvPmzVy5coXs7Gxq1apFz549eeWVV3B0dMzX/ubNm/zvf//jxIkT3L17F09PT/r378+YMWOwsLAw0rMwXcV5fRMTE2nfvn2R13JxceHIkSPlEXaFoSgKP/30Exs2bCAsLAwLCwt8fHwYMmQI/fv3L9A+Li6OJUuWcOjQIaKjo3F1daVPnz5MmDABe3v78n8CepKkyUSEhIQA0L59e9zd3Qucb9WqVXmHVKFdvXqVadOmUdQ6h5SUFEaOHElISAh169ala9euXL58mQ0bNrBv3z42bdqEh4dHOUddcTzs9c3KyiI8PBwnJyc6d+5caBsHB4eyDLFC0Wq1TJkyhd27d2NjY0Pz5s2xtbXl7NmzLF++nN27d/Pjjz/i4uICwJUrV3j++edJTEzkkUcewc/Pj5MnT7Jo0SKCgoJYvny5rOS9T3Ff3wsXLgDQsGFDmjVrVuB6/01gBXz00UesW7cOGxsb2rVrh0ql4tSpU7z11lscP36cTz/9VNc2JiaG5557jlu3buHt7U3Xrl05d+4cy5cv5+DBg6xfv950EydFmIRx48Yp3t7eyuXLl40dSoV39OhRpUOHDoq3t7fi7e2tbNq0qUCbTz75RPH29lZmzJihZGdnK4qiKBqNRnd8/Pjx5R12haHP63v27FnF29tbefXVV40QYcWzadMmxdvbW+ndu7dy8+ZN3fHk5GTllVdeUby9vZUpU6bojg8YMEDx9vZWNm7cmK/t888/r3h7eyvff/99eYZv8or7+i5fvlzx9vZW1q5da4xwK5z9+/cr3t7eSpcuXZSoqCjd8aioKKVLly6Kt7e3sn//ft3xSZMmKd7e3sqiRYt0xzIzM5XXXntN8fb2Vj766KNyjb84ZE6TiQgJCcHGxoaGDRsaO5QK6+7du3zwwQeMGTOGxMREatWqVWi7lJQUNm3ahI2NDe+8847uG7mZmRkzZsygTp067N27lxs3bpRn+CZP39cX/v2m7ufnV17hVWibN28GYObMmdSuXVt33N7enk8++QSVSsWePXvIyMggKCiIkJAQWrVqxZAhQwptu3r16iJ7Aaui4ry+8O/7t7BeJlHQtm3bAJgyZQo1a9bUHa9ZsybPP/88AIcOHQLg+vXrBAYGUrNmTV599VVdW0tLSz766CPs7e356aefSE1NLcdnoD9JmkxAXFwcsbGx+Pr66lWRVBRu6dKlrF+/nrp167J69Wr8/f0LbRccHExaWhpt27bFyckp3zm1Wk23bt0A2L9/f1mHXKHo+/qCfOgUl5OTEw0bNuSRRx4pcM7Z2RknJyeys7OJj4/nwIEDAHTv3r1A27p16+Lj40NUVNQD55NVNcV5fSH3/atWq/H19S3vUCukefPm8dtvv9GnT58C59LS0gB0n20HDx5EURS6dOlSYAjZwcEBf39/MjIyOHbsWNkHbgBJmkxA3nwmDw8P5s+fT+/evWnevDndu3dn/vz5JCYmGjnCiqFOnTq8//777Nixg7Zt2xbZLiwsDIDGjRsXer5Ro0YAXL58ufSDrMD0fX3h36QpJiaGUaNG4e/vT6tWrRg5cqTuG6f419KlS/n999+pXr16gXM3btwgISEBCwsLnJ2duXLlCgDe3t6FXkvevwUV5/VNS0sjIiKCOnXq8Msvv/Dss8/SqlUrOnTowLRp07h69aoRnoFps7CwoFGjRtja2uY7/tdff/Hjjz+iVqvp168fQIV//0rSZALyPmB+//13Nm3aRIMGDWjdujWJiYmsXLmSIUOGEBsba+QoTd/IkSMZPnz4Q1cOxcTEAODm5lboeVdXVyC3B1D8S9/XV6PR6P7gzZw5k/j4eNq1a0etWrU4fvw4Y8eOZeXKleURcqWwaNEiALp164aVlZXu/Zv3Pv2vvON3794tnwAruP++vhcvXkSr1RIREcHHH3+MnZ0d/v7+WFhY8NtvvzFw4ECCg4ONHLVpmz59Ov379+e5554DYOHChbqh+or+/pXlFSbg4sWLAHTp0oWFCxfqVhXdu3ePqVOncuzYMWbPns3SpUuNGWalkdddbG1tXej5vON57UTxhIeHk5GRgZWVFV988UW+YaSdO3fy5ptv8vnnn9O2bVtatGhhxEhN36pVq/j999+xsbHh9ddfByA9PR0o+v2bt3+Wqc4JMSWFvb55f4/r1q3L0qVLdWUxsrOzWbhwId9//z1Tp05l9+7dBXpWBMTHx7Njxw7dzyqVisuXL9OrVy/UavVD3795x031/Ss9TSZgwYIF7Ny5ky+//DLfMmxnZ2c+++wzbG1t2bdvH5GRkUaMsvLIG1tXqVQPbCcTaQ3j7e3NkSNH+O233wrMu+nbty/PP/88Wq2W9evXGynCimHVqlV8+umnqFQqPv74Y92Ht7x/S0dRr+/w4cPZu3cv69evz1dHzMLCghkzZtCsWTPi4uLYtWuXsUI3aXZ2dhw9epRTp06xbNkyHB0dWbJkCe+//z5Q8d+/kjSZAEtLS7y8vLCxsSlwzt3dnaZNmwL/zn0SJZP37TBvpcx/5R2Xb5GGc3FxoU6dOoWey5tof/78+fIMqcJQFIXPPvuMTz/9FLVazaeffsqTTz6pO5/3vszMzCz08XnH5f1buIe9vmZmZnh6eupqNt3PzMyMLl26APL+LYqlpSU1atTA3t6eLl26sHz5cmxsbPjll1+4efPmQ9+/pv73V5KmCiDvlzevW1OUTF7x0KLmLOXNHytqzF2UTN7rKu/ngjIyMpgyZQorVqzA2tqar776igEDBuRrk/f+LWqeo7x/i6bP6/sw8ve4eOrWrUurVq3QarVcvHixwr9/JWkysszMTGbPns2kSZOK7Pm4efMmgFSoLiV5q+byVtH918NWd4gH27VrF9OnT+enn34q9Ly8nwuXkpLCqFGj+PPPP3F2dmbNmjX07NmzQLu89294eHih15H3b+H0fX2XLl3KlClTOHv2bKHXyZsmIe/ffy1atIjXXnutyHmglpaWAOTk5FT4968kTUZmZWXF/v37CQwM5PDhwwXOh4aGEhoaioODAy1btiz/ACuhtm3bYmtrS3BwMMnJyfnOaTQa9u3bh0qlIiAgwEgRVmyJiYns2LGDdevWFTovYcuWLQB06tSpvEMzWdnZ2YwbN46//vqLunXrsnHjxkJrCgG6bWkCAwMLnLtx4waXL1+mZs2a+Pj4lGnMFUlxXt8rV66wa9cutm/fXuBcRkYGf/zxByDv3/sdOHCAP/74o9D3ZFJSEmfOnAFy67YFBASgUqnYt28fGo0mX9vk5GSOHz+OjY3NA/f+MyZJmkxA3rLMTz75RPctHHKHj9555x00Gg0vvfRSkasNRPHY2NgwcOBAUlNTee+998jKygJy5zp8/vnnREZG0rNnTxo0aGDkSCumPn36UK1aNS5evMg333yTL3HatGkTu3btokaNGrr3vYCvv/6aU6dO4erqytq1a6lbt26Rbdu2bUuTJk04efIka9as0R1PSUnhnXfeQVEUxowZ89CJtlVJcV7fYcOGAbB+/fp8X2SzsrKYM2cOUVFRPPbYY7Ru3brM464o8n6XP/vsMyIiInTHExMTefPNN0lISKBnz57Uq1cPT09PunXrRmRkJJ9//rnu70NWVhbvvfceqampDB061GT3plQppjpFvQrJysrilVde4ejRo1hbW9OmTRssLS05fvw4aWlp9O7dm8WLF0u18GKaOXMmW7ZsYe7cuQwePDjfuZSUFIYNG8bly5fx9PTEz8+PK1eucPXqVTw9PdmwYUORdZxErge9vgcOHGDy5MlkZWVRv359fHx8iIiI4NKlS9ja2rJ8+XLatGljpMhNS3x8PN26dSM9PR1fX98ii65C7mvu4uLCxYsXGTFiBCkpKTRr1ozatWtz8uRJ7t69S9euXfnmm29kw95/GPL6fvnllyxZsgSAli1b4u7uzpkzZ4iOjqZhw4b88MMPhU4Ur6q0Wi3Tpk3j999/x9LSkjZt2mBubs7Zs2dJTEykWbNmfP/997odGG7fvs3QoUN1r2fjxo05d+4cUVFRNGvWjB9++AE7OzsjP6vCSdJkInJycli7di2//vorV69exczMjMaNGzN48GAGDRok3xoN8KAPdcjtNv6///s/du3aRWxsLO7u7gQEBDBx4kSTnYRoSh72+l6+fJmlS5dy/PhxEhMTcXZ2plOnTkyYMKHIlXVV0Z9//plvD66Hta1Xrx4A165d46uvviIoKIj09HTq1q1L//79eeGFF3RzSIThr+/evXtZs2YN586dIysrC09PT/r06cPLL79ssh/oxqQoCj///DObNm3SFbetX78+Tz75JC+++KKuflie6Oho/ve//7F//37dXpZ5r6+9vb0xnoJeJGkSQgghhNCDzGkSQgghhNCDJE1CCCGEEHqQpEkIIYQQQg+SNAkhhBBC6EGSJiGEEEIIPUjSJIQQQgihB0mahBBCCCH0IEmTEEIIIYQeJGkSQlQJplrH11TjEkIUJEmTEMJk/e9//8PHx4fFixcbfI1r164xduxYbty4UYqRlZxGo2HdunV8/PHHxg5FCKEnSZqEEJXa2LFjOXTokLHDKGD79u18+OGHJCQkGDsUIYSeZBtsIYTJev755+nbty/Vq1c3+BqmOvyl1WqNHYIQopgkaRJCmCxnZ2ecnZ2NHYYQQgAyPCeEMGH/ndOU93NgYCAHDhxgxIgRtGrVitatW/PSSy9x6tQp3WOPHz+Oj48Pt27dAqBXr174+PgQGRlZolh27tzJ7NmzadWqFe3atWP+/Pm6NkH/3979hTT1/3Ecf84RhhfZhVnQwkrGYo62Q2XYqotMgiIioruJdlVGaSWVjS6FguxiEuiV/aFMwouCImGYBTXR/ixWUTGaSDRXDCGLRdrqdxGOn9iXjv++fK3X4/J8zs7787l78fm8d053N9XV1WzcuBGXy4VhGOzYsYOmpiaGh4cz95WXl3PixAng5zGdw+Ggrq5uTL3nz59TXV1NSUkJLpeL0tJSTp06xeDg4KTmLyJTp50mEZl1rl+/TjAYZOnSpXi9XqLRKPfv36enp4fLly/j8XjIy8tj+/btdHZ2kkqlKC0tJScnh5ycnCnVDgQCDAwM4PV6icfjFBYWAnD+/HlOnz7NnDlzMAwDj8dDIpEgEonw6tUrXrx4wblz5wBYt24dIyMjhMNhbDYbhmFgGEamxo0bN/D7/aTTaYqKili8eDEvX77kwoULBINBLl26hM1mm9I6RGTiFJpEZNYJBoPU1dVRWVmJxWLh+/fvHD58mI6ODlpaWmhsbKSwsJCGhgY2bdpEKpXi+PHjFBQUTLl2f38/165dY+XKlcDP3qQPHz5w9uxZcnNzaWtrY/ny5Zn7Hz16REVFBcFgkEQiwaJFi6iqqmLhwoWEw2EMw6ChoSFzfywW4+TJk2RnZ9Pc3ExxcXGmTiAQoLm5mWPHjtHa2jrltYjIxOh4TkRmHafTyZ49e7BYLABkZWXh8/kAiEajM1rb7XZnAtNo7WQySVlZGfv37x8TmABWr16N3W4HMHU0ePHiRYaHhzl48GAmMI3WOXToEA6Hg8ePH/P06dPpWZCImKadJhGZddxu97hr+fn5AHz58mVGazscjnHXnE7nuHdJpdNp3r59y7Nnz/j48SMAIyMjv31+T08PAGvXrh03ZrFYWL9+Pa9fv6a3txePxzOJFYjIZCk0icisk5ubO+6a1WoFZv6v/PPnz//l9XQ6TUdHB7du3SIajRKPx/n27RtAZkfMzOsPBgYGANi5c6ep+0Tk36PQJCIyAaMB6P+lUikqKiqIRCLMnTsXl8uF1+vFbrezatUq6uvrefjwoannp9NpALZt20ZW1j93UKxYsWJyCxCRSVNoEhGZopaWFiKRCCUlJTQ2NjJv3rwx40NDQ6aflZ+fz7t376ipqZmWxnURmT5qBBcRmaJwOAyAz+cbF5jev3/PmzdvgLFHh7/asQJYs2YNAPfu3fvleG1tLbt27aKzs3PK8xaRiVFoEpE/WnZ2NgCfPn2asRqjby3v6uoa07cUj8c5cOBAprfp69ev4+b1+fPnMc8qLy/HarUSCATo7u4eM3b16lVu3rxJNBr9ZTO8iMwsHc+JyB+toKCAWCxGTU0NRUVFHD16lCVLlkxrDZ/Px+3bt2lvb+fJkyfY7XYGBwcJh8P8+PGDZcuW0dfXRzKZHDMvgLt377J3714Mw2Dfvn24XC78fj/19fVUVlbidDqx2Wz09fURjUaxWq2cOXOGvLy8aV2DiPyedppE5I/m9/spLi4mmUwSCoWIxWLTXsPtdtPa2sqGDRsYGhrizp079Pf3s3nzZtra2jhy5AjwcydqVFFREbW1tSxYsIAHDx4QCoUyYz6fjytXrlBWVkYikaCrq4tUKsXWrVtpb29ny5Yt074GEfk9y4//6ifARURERP5DtNMkIiIiYoJ6mkTkr9LU1JT5N5tZVVVVmQ/zisjfS6FJRP4qoVCI3t7eCf1m9+7dCk0iop4mERERETPU0yQiIiJigkKTiIiIiAkKTSIiIiImKDSJiIiImKDQJCIiImKCQpOIiIiICQpNIiIiIiYoNImIiIiYoNAkIiIiYsL/AFir6W+u3fjnAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "\n",
    "for key, group in train_samp_df.groupby('term'):\n",
    "    this_pdp = partial_dependence(base_model, group[predictor_features], \n",
    "                                  ['int_rate'], kind='average',\n",
    "                                 percentiles=(0,1),\n",
    "                                 method='brute')\n",
    "    this_pdp_df = pd.DataFrame({k:v[0] for k,v in this_pdp.items()})\n",
    "    \n",
    "    this_pdp_df.plot(x='values', y='average', ax=ax,\n",
    "                     color=color_dict[key], label=key)\n",
    "ax.set_ylabel('1-way partial dependence')\n",
    "ax.set_xlabel('int_rate')\n",
    "ax.legend(title='term')\n",
    "fig.savefig(Path(setup.temp_path).joinpath('06_PLOT_int_term_pdp_avg_double_1way.png'),\n",
    "            bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "efe93536-a7f2-4d59-acb5-1eab1dd3051f",
   "metadata": {},
   "source": [
    "##### 2-Way ICE Curves (Sample)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "69e1c57f-dae9-43a2-8b9f-b2774e27a84a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get stratified samp - large sample to ensure adequate grid size\n",
    "X_samp = train_samp_df[predictor_features].groupby('term').sample(100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "bac5e523-a0b0-4090-9b12-0482890b7f9f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get 2-Way PDP\n",
    "indiv_pdp = partial_dependence(base_model, X_samp, \n",
    "                                  ['int_rate', 'term'], kind='individual',\n",
    "                               percentiles=(0,1),\n",
    "                              method='brute',\n",
    "                              grid_resolution = 100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "d2334925-9e0b-41be-9ca2-a90f73c2a269",
   "metadata": {},
   "outputs": [],
   "source": [
    "# DataFrame conversion \n",
    "num_samp = indiv_pdp['individual'][0].shape[0]\n",
    "indiv_samp_df = pd.concat([pd.DataFrame(indiv_pdp['individual'][0][i,:,:],\n",
    "                                   columns = indiv_pdp['values'][1]) \\\n",
    "                           for i in range(0, num_samp)],\n",
    "                     keys = range(0, num_samp),\n",
    "                     axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "5ea877f1-0926-4e22-966f-fbe3e50d32d7",
   "metadata": {},
   "outputs": [],
   "source": [
    "indiv_samp_df.index = pd.Series(indiv_pdp['values'][0], name='values')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "32c5fcc6-fcf3-42e2-8128-56f98722403d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(100, 400)"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "indiv_samp_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "a537e0f8-7ba7-43f6-bed4-64a744bed1ce",
   "metadata": {},
   "outputs": [],
   "source": [
    "random.seed(43)\n",
    "colsamp = random.sample(list(indiv_samp_df.columns.get_level_values(0)), 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "8c67a4f2-9ddf-44b7-8fb1-ab1577cb6bfe",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get subsample of cases to make plots nicer\n",
    "colsamp = random.sample(list(indiv_samp_df.columns.get_level_values(0)), 20)\n",
    "indiv_samp_2_df = indiv_samp_df.loc[:, colsamp]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "d152808a-81a4-4a4f-9c27-15c2fe0268ab",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(100, 40)"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "indiv_samp_2_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "4614d21d-9050-4227-882f-9844e682af5c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Plot - panel, show side by side"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "fd5593ec-9cce-405a-be93-3aa2678330df",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 800x300 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(1, 2, sharex = True, sharey=True,\n",
    "                      figsize = (8,3))\n",
    "\n",
    "for i, tv in enumerate(set(indiv_samp_2_df.columns.get_level_values(1))):\n",
    "    this_df = indiv_samp_2_df.loc[:, indiv_samp_2_df.columns.get_level_values(1)==tv]\n",
    "    this_df.plot(color=color_dict[tv],  ax=ax[i], legend=None)\n",
    "    ax[i].set_xlabel('int_rate')\n",
    "    ax[i].set_title(f'term={tv}')\n",
    "\n",
    "ax[0].set_ylabel('2-way ICE')\n",
    "fig.savefig(Path(setup.temp_path).joinpath('06_PLOT_int_term_ice_samples_panel.png'),\n",
    "            bbox_inches='tight')"
   ]
  }
 ],
 "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": "3.10.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}