Diagram_feature_importance.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index(['Unnamed: 0.4', 'Unnamed: 0.3', 'Unnamed: 0.2', 'Unnamed: 0.1',\n",
      "       'Unnamed: 0', 'CLINIC', 'YEAR', 'GENDERSP', 'Susceptible', 'MSM',\n",
      "       'MSMW', 'MSW', 'Oth/Unk/Missing', 'REGION', 'Midwest', 'Northeast',\n",
      "       'Southeast', 'Southwest', 'West', 'PREV_REGION', 'PREV_CLINIC',\n",
      "       'DELTA_REGION', 'DELTA_CLINIC'],\n",
      "      dtype='object')\n"
     ]
    }
   ],
   "source": [
    "#%reset\n",
    "import seaborn as sb\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import sklearn as sk\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "from sklearn.inspection import permutation_importance\n",
    "from imblearn.over_sampling import RandomOverSampler\n",
    "\n",
    "hfont = {'fontname':'Helvetica'}\n",
    "\n",
    "## read data \n",
    "CIP_data = pd.read_csv(\"CIP_data_encode_prev.csv\")\n",
    "CIP_data_no_drop = pd.read_csv(\"CIP_data_encode_prev_not_dropped.csv\")\n",
    "print(CIP_data_no_drop.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "#nn_data \n",
    "best_features_by_year_nn = {2005: ['PREV_REGION', 'West', 'Midwest', 'DELTA_REGION', 'DELTA_CLINIC', 'Southeast', 'Southwest', 'MSW'], 2006: ['PREV_CLINIC', 'PREV_REGION', 'DELTA_CLINIC', 'MSM', 'DELTA_REGION', 'Midwest'], 2007: ['PREV_CLINIC', 'PREV_REGION', 'MSW', 'DELTA_REGION', 'Midwest', 'MSMW', 'Oth/Unk/Missing'], 2008: ['PREV_REGION', 'PREV_CLINIC', 'DELTA_CLINIC', 'MSW', 'West', 'DELTA_REGION', 'MSM', 'Midwest', 'Northeast', 'MSMW'], 2009: ['PREV_CLINIC', 'DELTA_REGION', 'MSW', 'DELTA_CLINIC'], 2010: ['PREV_CLINIC', 'MSW', 'MSM', 'DELTA_CLINIC', 'MSMW', 'PREV_REGION', 'West', 'Southeast', 'Southwest', 'Oth/Unk/Missing']}\n",
    "best_hyperparameters_by_year_nn = {2005: {'solver': 'sgd', 'learning_rate': 'constant', 'hidden_layer_sizes': 12, 'alpha': 0.12618568830660204, 'activation': 'tanh'}, 2006: {'solver': 'sgd', 'learning_rate': 'constant', 'hidden_layer_sizes': 12, 'alpha': 0.12618568830660204, 'activation': 'tanh'}, 2007: {'solver': 'sgd', 'learning_rate': 'constant', 'hidden_layer_sizes': 12, 'alpha': 0.12618568830660204, 'activation': 'tanh'}, 2008: {'solver': 'sgd', 'learning_rate': 'constant', 'hidden_layer_sizes': 12, 'alpha': 0.12618568830660204, 'activation': 'tanh'}, 2009: {'solver': 'sgd', 'learning_rate': 'constant', 'hidden_layer_sizes': 12, 'alpha': 0.12618568830660204, 'activation': 'tanh'}, 2010: {'solver': 'sgd', 'learning_rate': 'constant', 'hidden_layer_sizes': 12, 'alpha': 0.12618568830660204, 'activation': 'tanh'}}\n",
    "ROC_by_year_nn = {2005: 0.707710113960114, 2006: 0.7363098705025091, 2007: 0.6839585775693631, 2008: 0.6843168957154405, 2009: 0.6645239980510848, 2010: 0.6755020080321286}##lr_data \n",
    "\n",
    "best_features_by_year_nn = {2005: ['DELTA_CLINIC', 'MSW', 'MSM', 'West', 'PREV_CLINIC', 'PREV_REGION', 'DELTA_REGION', 'Northeast', 'MSMW', 'Oth/Unk/Missing'], 2006: ['DELTA_CLINIC', 'PREV_REGION', 'MSM', 'MSW', 'Southwest'], 2007: ['PREV_REGION', 'MSM', 'MSW', 'PREV_CLINIC', 'Midwest', 'Southwest', 'MSMW'], 2008: ['PREV_REGION', 'West', 'DELTA_CLINIC', 'MSM', 'MSW', 'PREV_CLINIC', 'MSMW', 'DELTA_REGION', 'Oth/Unk/Missing', 'Southeast'], 2009: ['PREV_REGION', 'PREV_CLINIC', 'MSW', 'DELTA_CLINIC', 'MSM', 'MSMW', 'Northeast'], 2010: ['DELTA_CLINIC', 'MSM', 'PREV_CLINIC', 'PREV_REGION', 'Southwest']}\n",
    "best_hyperparameters_by_year_nn = {2005: {'solver': 'sgd', 'learning_rate': 'constant', 'hidden_layer_sizes': (12,), 'alpha': 0.12618568830660204, 'activation': 'tanh'}, 2006: {'solver': 'sgd', 'learning_rate': 'constant', 'hidden_layer_sizes': (12,), 'alpha': 0.12618568830660204, 'activation': 'tanh'}, 2007: {'solver': 'sgd', 'learning_rate': 'constant', 'hidden_layer_sizes': (12,), 'alpha': 0.12618568830660204, 'activation': 'tanh'}, 2008: {'solver': 'sgd', 'learning_rate': 'constant', 'hidden_layer_sizes': (12,), 'alpha': 0.12618568830660204, 'activation': 'tanh'}, 2009: {'solver': 'sgd', 'learning_rate': 'constant', 'hidden_layer_sizes': (12,), 'alpha': 0.12618568830660204, 'activation': 'tanh'}, 2010: {'solver': 'sgd', 'learning_rate': 'constant', 'hidden_layer_sizes': (12,), 'alpha': 0.12618568830660204, 'activation': 'tanh'}}\n",
    "\n",
    "#lr_data \n",
    "best_hyperparameters_by_year_lr = {2005: {'solver': 'liblinear', 'penalty': 'l1', 'C': 8.65}, 2006: {'solver': 'liblinear', 'penalty': 'l1', 'C': 35.730000000000004}, 2007: {'solver': 'liblinear', 'penalty': 'l1', 'C': 8.65}, 2008: {'solver': 'liblinear', 'penalty': 'l1', 'C': 54.85}, 2009: {'solver': 'liblinear', 'penalty': 'l2', 'C': 83.42}, 2010: {'solver': 'liblinear', 'penalty': 'l1', 'C': 8.65}}\n",
    "best_features_by_year_lr = {2005: ['PREV_CLINIC', 'MSM', 'DELTA_CLINIC', 'PREV_REGION', 'MSW', 'Southwest', 'Northeast', 'Oth/Unk/Missing'], 2006: ['DELTA_REGION', 'PREV_CLINIC', 'DELTA_CLINIC', 'MSW', 'PREV_REGION', 'Oth/Unk/Missing', 'MSM', 'Southwest', 'Southeast'], 2007: ['MSM', 'PREV_CLINIC', 'MSW', 'Oth/Unk/Missing', 'MSMW'], 2008: ['PREV_CLINIC', 'DELTA_CLINIC', 'West', 'MSW', 'MSM', 'PREV_REGION', 'MSMW', 'Oth/Unk/Missing', 'Northeast'], 2009: ['PREV_CLINIC', 'Oth/Unk/Missing', 'DELTA_CLINIC', 'Northeast'], 2010: ['MSW', 'MSM', 'DELTA_CLINIC', 'PREV_CLINIC', 'Oth/Unk/Missing', 'West', 'Southwest', 'MSMW', 'DELTA_REGION', 'Southeast', 'Northeast']}\n",
    "ROC_by_year_lr = {2005: 0.7317511805463615, 2006: 0.7423710317796873, 2007: 0.7048918256421187, 2008: 0.6971529715492738, 2009: 0.6166128405256475, 2010: 0.6792793175522966}\n",
    "##rf_paper \n",
    "best_hyperparameters_by_year_rf = {2005: {'n_estimators': 163, 'min_samples_split': 5, 'min_samples_leaf': 14, 'max_depth': 81}, 2006: {'n_estimators': 163, 'min_samples_split': 5, 'min_samples_leaf': 14, 'max_depth': 81}, 2007: {'n_estimators': 163, 'min_samples_split': 5, 'min_samples_leaf': 14, 'max_depth': 81}, 2008: {'n_estimators': 163, 'min_samples_split': 5, 'min_samples_leaf': 14, 'max_depth': 81}, 2009: {'n_estimators': 163, 'min_samples_split': 5, 'min_samples_leaf': 14, 'max_depth': 81}, 2010: {'n_estimators': 163, 'min_samples_split': 5, 'min_samples_leaf': 14, 'max_depth': 81}}\n",
    "best_features_by_year_rf = {2005: ['MSW', 'MSM', 'MSMW'], 2006: ['MSW', 'DELTA_CLINIC', 'MSM', 'MSMW', 'Oth/Unk/Missing'], 2007: ['MSW', 'PREV_CLINIC', 'MSM'], 2008: ['PREV_CLINIC', 'MSW', 'MSM', 'DELTA_CLINIC', 'Oth/Unk/Missing', 'Southeast'], 2009: ['MSW', 'MSM', 'PREV_CLINIC', 'DELTA_CLINIC', 'PREV_REGION', 'Midwest'], 2010: ['DELTA_CLINIC', 'PREV_CLINIC', 'MSW', 'PREV_REGION', 'West', 'DELTA_REGION', 'Midwest', 'Southeast', 'MSMW']}\n",
    "ROC_by_year_rf = {2005: 0.7418091168091168, 2006: 0.6955865853135346, 2007: 0.6980265728800312, 2008: 0.7012934518997574, 2009: 0.6612742833389195, 2010: 0.6738955823293172}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "########### Try functionalised \n",
    "\n",
    "\n",
    "def effective_unnecessary_threshold(threshold_seq, y_predict_proba, y_test, cipro_R_prevalence):\n",
    "\n",
    "    get_effective_threshold = []\n",
    "    incorrectly_get_X_threshold = [] #no bootstrapping, no 95% CI \n",
    "    sensitivity_threshold = []\n",
    "    specificity_threshold = []\n",
    "    for threshold in threshold_seq:\n",
    "\n",
    "        y_predict_test = np.where(y_predict_proba[:, 1] > threshold, 1, 0)\n",
    "\n",
    "        tn_test , fp_test , fn_test , tp_test  = confusion_matrix(y_true=y_test, y_pred=y_predict_test).ravel()\n",
    "\n",
    "        sensitivity_test  = tp_test  / (tp_test   + fn_test )\n",
    "        specificity_test   = tn_test / (tn_test + fp_test )\n",
    "\n",
    "        sensitivity_threshold.append(sensitivity_test*100)\n",
    "        specificity_threshold.append(specificity_test*100)\n",
    "        get_effective_threshold.append(sensitivity_test * cipro_R_prevalence*100 + (100 -  cipro_R_prevalence*100)) #q_p\n",
    "        incorrectly_get_X_threshold.append((100 - cipro_R_prevalence*100) * (1 - specificity_test)) #c_p\"\n",
    "    return(sensitivity_threshold, specificity_threshold, get_effective_threshold, incorrectly_get_X_threshold)\n",
    "\n",
    "def get_best_hyperparameters(model, cv, space, X_train, y_train):\n",
    "        search = RandomizedSearchCV(model, space, scoring='roc_auc', n_iter=1,  n_jobs=-1, cv=cv, random_state=1)\n",
    "        result = search.fit(X_train, y_train)\n",
    "        return(result.best_params_)\n",
    "\n",
    "def get_best_features(feature_names, model_fit, X_test, y_test):\n",
    "    PI = permutation_importance(model_fit, X_test, y_test, n_repeats = 10, random_state = 42)\n",
    "    important_features = []\n",
    "    for q in PI.importances_mean.argsort()[::-1]:\n",
    "        if PI.importances_mean[q] - 2 * PI.importances_std[q] > 0:\n",
    "          important_features.append(feature_names[q]) #works cos they are in same order as the x columns\n",
    "    return(important_features)\n",
    "\n",
    "\n",
    "def get_feature_effects(feature_names, model_fit, X_test, y_test):\n",
    "    PI = permutation_importance(model_fit, X_test, y_test, n_repeats = 10, random_state = 42)\n",
    "    \n",
    "    return(PI.importances_mean)\n",
    "\n",
    "oversample = RandomOverSampler(sampling_strategy = 0.5, random_state=42)\n",
    "\n",
    "def get_test_train_data_overfit(CIP_data_no_drop, year, feature_names, oversample_size):\n",
    "    years_train = np.array(range(year - 5, year))\n",
    "\n",
    "    # first do for all clinics \n",
    "    train_data = CIP_data_no_drop.loc[CIP_data_no_drop['YEAR'].isin(years_train)]\n",
    "    X_train = train_data[feature_names] #need to consider all columns BEFORE feature engineering\n",
    "    y_train = 1 - train_data['Susceptible']\n",
    "    X_train, y_train = oversample.fit_resample(X_train,y_train)\n",
    "    #test\n",
    "    test_data = CIP_data_no_drop.loc[CIP_data_no_drop['YEAR'].isin([year])]\n",
    "    X_test = test_data[feature_names]\n",
    "    y_test = 1 - test_data['Susceptible']\n",
    "    cipro_R = y_test.sum()/len(y_test) \n",
    "    X_test, y_test = oversample.fit_resample(X_test,y_test)\n",
    "    return(test_data, train_data, X_train, y_train, X_test, y_test, cipro_R)\n",
    "\n",
    "\n",
    "def get_test_train_data(CIP_data_no_drop, year, feature_names):\n",
    "    years_train = np.array(range(year - 5, year))\n",
    "\n",
    "    CIP_data_training_years = CIP_data_no_drop.loc[CIP_data_no_drop['YEAR'].isin(years_train)]\n",
    "    CIP_data_testing_years = CIP_data_no_drop.loc[CIP_data_no_drop['YEAR'].isin([year])]\n",
    "    # first do for all clinics \n",
    "    train_data = CIP_data_no_drop.loc[CIP_data_no_drop['YEAR'].isin(years_train)]\n",
    "    X_train = train_data[feature_names] #need to consider all columns BEFORE feature engineering\n",
    "    y_train = 1 - train_data['Susceptible']\n",
    "    #test\n",
    "    test_data = CIP_data_no_drop.loc[CIP_data_no_drop['YEAR'].isin([year])]\n",
    "    X_test = test_data[feature_names]\n",
    "    y_test = 1 - test_data['Susceptible']\n",
    "    return(test_data, train_data, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "########### Now do proportion receiving effective and unnecessary treatment for years surrounding 2007 decision \n",
    "### now do a loop over different years to get the proportion receiving effective and unnecessary treatment \n",
    "### has LOOCV \n",
    "\n",
    "## NEEDED FOR FEATURE ENGINEERING \n",
    "feature_names = ['MSM','MSMW', 'MSW', 'Oth/Unk/Missing','Northeast', 'Southeast', 'Southwest', 'West', 'Midwest','PREV_REGION', 'PREV_CLINIC', 'DELTA_REGION', 'DELTA_CLINIC']\n",
    "\n",
    "\n",
    "#### Loop set up \n",
    "test_years = [2005, 2006, 2007, 2008, 2009, 2010]\n",
    "\n",
    "imporances_all_models = pd.DataFrame(0, index = feature_names, columns=np.arange(len(test_years)*3) + 1)\n",
    "\n",
    "### effect features by year\n",
    "best_effect_by_year_lr = {}\n",
    "best_effect_by_year_nn = {}\n",
    "best_effect_by_year_rf = {}\n",
    "i = 1\n",
    "for year in test_years: \n",
    "    years_train = np.array(range(year - 5, year))\n",
    "\n",
    "    CIP_data_training_years = CIP_data_no_drop.loc[CIP_data_no_drop['YEAR'].isin(years_train)]\n",
    "    CIP_data_testing_years = CIP_data_no_drop.loc[CIP_data_no_drop['YEAR'].isin([year])]\n",
    "    \n",
    "    ## LOGISTIC REGRESSION - no overfit \n",
    "    test_data, train_data, X_train, y_train, X_test, y_test =  get_test_train_data(CIP_data_no_drop = CIP_data_no_drop, year = year, feature_names = feature_names)\n",
    "    cipro_R_prev = y_test.sum()/len(y_test) #get prevalence for all clinics for \"general\" model for year \n",
    "    model_lr = LogisticRegression(class_weight = 'balanced', max_iter=4000, solver = best_hyperparameters_by_year_lr[year]['solver'], C = best_hyperparameters_by_year_lr[year]['C'], penalty = best_hyperparameters_by_year_lr[year]['penalty'])\n",
    "    model_fit_lr = model_lr.fit(X_train, y_train)\n",
    "    imporances_all_models[i + 6] = get_feature_effects(feature_names, model_fit_lr, X_test, y_test) #want it to be the \"second\" block\n",
    "   \n",
    "    ## NEURAL NETWORK \n",
    "    test_data, train_data, X_train, y_train, X_test, y_test, cipro_R_prev =  get_test_train_data_overfit(CIP_data_no_drop = CIP_data_no_drop, year = year, feature_names = feature_names, oversample_size = 0.5)\n",
    "    model_nn = MLPClassifier(solver = best_hyperparameters_by_year_nn[year]['solver'], activation = best_hyperparameters_by_year_nn[year]['activation'], max_iter = 3000 ,hidden_layer_sizes= best_hyperparameters_by_year_nn[year]['hidden_layer_sizes'], alpha =  best_hyperparameters_by_year_nn[year]['alpha'], random_state=10, learning_rate = 'adaptive' )\n",
    "    model_fit_nn = model_nn.fit(X_train, y_train)\n",
    "    imporances_all_models[i] =  get_feature_effects(feature_names, model_fit_nn, X_test, y_test) #want it to be the \"first\" block\n",
    "    \n",
    "    ## RANDOM FOREST \n",
    "\n",
    "    model_rf = RandomForestClassifier(n_estimators = best_hyperparameters_by_year_rf[year]['n_estimators'], min_samples_split = best_hyperparameters_by_year_rf[year]['min_samples_split'], min_samples_leaf=best_hyperparameters_by_year_rf[year]['min_samples_leaf'], max_features = 'sqrt', max_depth = best_hyperparameters_by_year_rf[year]['max_depth'], random_state = 10)\n",
    "    model_fit_rf = model_rf.fit(X_train, y_train)\n",
    "    imporances_all_models[i + 12] = get_feature_effects(feature_names, model_fit_rf, X_test, y_test) #want it to be the \"third\" block\n",
    "    i += 1\n",
    "\n",
    "\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                       1         2         3         4         5         6   \\\n",
      "MSM              0.029095  0.010817  0.036173  0.016006  0.004407  0.007885   \n",
      "MSMW             0.001555  0.000267  0.001290  0.002506  0.001954 -0.002811   \n",
      "MSW              0.030188  0.003254  0.029530  0.013959  0.004945 -0.003507   \n",
      "Oth/Unk/Missing  0.000332 -0.000953 -0.001641  0.001011 -0.000026  0.000228   \n",
      "Northeast        0.006054  0.000038 -0.005992 -0.000094  0.000839 -0.001191   \n",
      "Southeast       -0.000012 -0.004068 -0.020008  0.000862  0.000131 -0.001044   \n",
      "Southwest       -0.000404  0.000445  0.000964  0.000000 -0.000813  0.001526   \n",
      "West             0.027695 -0.017910 -0.005393  0.041323 -0.009864 -0.001272   \n",
      "Midwest          0.000036  0.000076  0.001850 -0.000862 -0.002164 -0.001539   \n",
      "PREV_REGION      0.008511  0.025779  0.119773  0.078766  0.012854  0.002436   \n",
      "PREV_CLINIC      0.014613  0.001424  0.006943  0.009634  0.007765  0.005127   \n",
      "DELTA_REGION     0.008143 -0.014707 -0.003074  0.003180 -0.007358 -0.001593   \n",
      "DELTA_CLINIC     0.047958  0.065120 -0.007308  0.021490  0.003843  0.038340   \n",
      "\n",
      "                       7         8         9         10        11        12  \\\n",
      "MSM              0.005067  0.031997  0.015577  0.023554 -0.020306  0.035490   \n",
      "MSMW             0.000000  0.001232  0.001631  0.002848 -0.008321  0.001916   \n",
      "MSW              0.033000  0.010348  0.022683  0.016914 -0.016625  0.041185   \n",
      "Oth/Unk/Missing  0.004567  0.000000  0.004809  0.000384  0.008499  0.009263   \n",
      "Northeast        0.002695  0.042280  0.000000 -0.000856  0.000284  0.001143   \n",
      "Southeast       -0.009327 -0.000066 -0.005991 -0.000122 -0.001085  0.002496   \n",
      "Southwest        0.000000  0.000000  0.000000 -0.001083  0.000036  0.006187   \n",
      "West            -0.009214 -0.000329 -0.001331  0.038878 -0.005850  0.005818   \n",
      "Midwest         -0.010731 -0.000181 -0.012964 -0.001730 -0.003254 -0.003182   \n",
      "PREV_REGION      0.018267  0.023686 -0.023665  0.005032 -0.003841  0.000123   \n",
      "PREV_CLINIC      0.044715  0.111991  0.028224  0.069789  0.009157  0.026648   \n",
      "DELTA_REGION    -0.021155  0.114520 -0.031403 -0.001311 -0.009602  0.002021   \n",
      "DELTA_CLINIC     0.019300  0.075657 -0.044799  0.048681  0.004499  0.031324   \n",
      "\n",
      "                       13        14        15        16        17        18  \n",
      "MSM              0.013675  0.008097  0.007763  0.009580  0.015110 -0.000228  \n",
      "MSMW             0.000321  0.000725 -0.001980  0.000243  0.000013  0.002410  \n",
      "MSW              0.021842  0.055282  0.049407  0.030989  0.032857  0.013722  \n",
      "Oth/Unk/Missing  0.000036  0.000559  0.000078  0.001078  0.000341 -0.000495  \n",
      "Northeast        0.000071 -0.000051 -0.000234 -0.000013  0.000000  0.000161  \n",
      "Southeast       -0.000012  0.000000 -0.007034  0.000822 -0.000131  0.002691  \n",
      "Southwest       -0.001745 -0.001373 -0.001706 -0.000674  0.000380 -0.001593  \n",
      "West            -0.019124 -0.016512 -0.005210  0.000054 -0.000708  0.007470  \n",
      "Midwest         -0.004179 -0.001169 -0.006148 -0.000485  0.002243  0.004003  \n",
      "PREV_REGION     -0.007479 -0.004398 -0.006630 -0.002978  0.006925  0.010348  \n",
      "PREV_CLINIC      0.001496 -0.000585  0.013560  0.034640  0.015058  0.017510  \n",
      "DELTA_REGION    -0.020715 -0.004983 -0.002644 -0.000040 -0.003332  0.006586  \n",
      "DELTA_CLINIC    -0.008832  0.010334 -0.002788  0.002883  0.009405  0.022784  \n"
     ]
    }
   ],
   "source": [
    "print(imporances_all_models)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABw0AAAKSCAYAAAAgf0RAAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/P9b71AAAACXBIWXMAAA9hAAAPYQGoP6dpAAEAAElEQVR4nOzdeVzU1f7H8fcMiwqoiIq540qKaaiB4houYHptuaWVcc3ymtY1TVu0W7fNe83KNFssy36mJddyq1DTXBHFXTQX3FFBcUWURUCY3x/cmSRgnAFkAl7Px2Mel77fc8738z0zzAz34+ccg8lkMgkAAAAAAAAAAABAhWV0dAAAAAAAAAAAAAAAHIukIQAAAAAAAAAAAFDBkTQEAAAAAAAAAAAAKjiShgAAAAAAAAAAAEAFR9IQAAAAAAAAAAAAqOBIGgIAAAAAAAAAAAAVHElDAAAAAAAAAAAAoIIjaQgAAAAAAAAAAABUcM6ODgAAAADlX6tWrRQbG6s777xTBw8edHQ4AACUCXx+AgCAP5PvO3Ut9WsO2hJV6tesyEgaAsVwbvtWR4dQZtS5J1CSdP3COQdHUnZUrl1HkhS/fq2DIylbGvQMliRdOxnn2EDKkKqNfSRJWSnXHBtIGeLiUVWSlJl02cGRlB3XruW+vq5dvaqsa1cdHE3Z4VK1miTp7Gb+SLJV3aDcP2JTTp90cCRlh0fDxpKkS/t+c3AkZUfNNndJkuJO8d3WHj6Ncr/fJmxY5+BIyg7z5+fVK1eUdibBwdGUDW716kvie5o9XGt4SWLO7GWet+sXLzg4krKjcq3akvieZg/z97SstDQHR1K2uLi5OToEoMwiaQgAAAAAAAAAAADrjAZHR4DbjD0NAQAAAAAAAAAAgAqOpCEAAAAAAAAAAABQwbE8KQAAAAAAAAAAAKwyGKhDK+94hgEAAAAAAAAAAIAKjkpDAAAAAAAAAAAAWGUwGhwdAm4zKg0BAAAAAAAAAACACo5KQwAAAAAAAAAAAFhFpWH5R6UhAAAAAAAAAAAAUMFRaQgAAAAAAAAAAADrDNShlXc8w7il+Ph4+fr6Wh6TJk2yqd/s2bMtfbp3715gm3Xr1mn8+PHq1auX2rVrJ39/f/Xu3Vvjx4/XqlWrCh17woQJlrFbt26ty5cv3zKerKwsBQQEWPr98MMPNt0HAAAAAAAAAABAeUfSEHZbuXKlTCbTLdstX7680HPXr1/XiBEjNHLkSEVERCgjI0MtWrRQ06ZNde3aNUVERGj06NF6/PHHlZycbPU62dnZVhOMZps2bbrlWAAAAAAAAAAAABURSUPYxdnZWefPn9fOnTuttjt9+rT27dtX6Pk33nhDGzZsUNOmTbVw4UJFRUVp4cKFWrRokbZs2aKvvvpKtWrV0s6dO/Xcc89ZjUeSfvnll1vGbi2JCQAAAAAAAAAACmcwGkr9gdJF0hB26dSpk6RbJ+nMCbrWrVvnO5eQkKCffvpJkvTJJ5/orrvuynPeYDCoW7dumjZtmiRp+/btio6OLvA6AQEBMhqN2rZtm9UlSjMzM7VmzRp5eXnpjjvusBo7AAAAAAAAAABARUPSEHYJDQ2VJK1atcrqEqXLly+X0WhUv3798p07cOCAcnJy5OHhoWbNmhU6RkBAgHx8fCRJe/bsKbBNzZo11bFjR2VnZ+vXX38tdKzIyEilpKQoJCRETk5OhbYDAAAAAAAAAAD5GQyGUn+gdJE0hF06duyo2rVr69y5c9q1a1eBbY4fP67Y2FgFBASoVq1a+c67uLhIklJSUrRjxw6r15s1a5ZWrlypsLCwQtuYE5PWqh/NlY8DBgywej0AAAAAAAAAAICKiKQh7GI0GhUSEiKp8CSdOUHXv3//As+3b99ebm5ukqQRI0Zo2rRpOnr0aIFtGzduLB8fH7m7uxcak7l6sLAlStPT07Vu3Trdcccd6tChQ+E3BwAAAAAAAAAACmQwGkv9gdLFjMNu5sq+wpYoXbFihVxcXNS3b98C+1erVk0TJ06UJKWmpurzzz9X//791b17d40fP17h4eGKi4uzOZ6aNWvqnnvu0Y0bN7R69ep859etW6e0tDT169ePcmYAAAAAAAAAAIACkDSE3Tp06CBvb28lJiZq9+7dec4dOnRIR48eVVBQkDw9PQsdY9CgQZo5c6bq169vOXbu3DlFRETozTffVEhIiO677z4tXrzY6t6JZtaWKF2xYoUk6b777rPl9gAAAAAAAAAAwB8ZDaX/QKkiaQi7GQwGhYaGSsqfpDMvTWpLgi44OFi//vqrvv76a4WFhal58+Z5zh87dkwTJ07UqFGjlJmZaXWsvn37ytnZWVu3blVSUpLleEpKijZs2KBGjRqpbdu2Nt0fAAAAAAAAAABARUPSEEViruxbuXJlnkrAFStWqFKlSurdu7dN4zg5OalLly567bXXtGzZMkVHR2v69Ol64IEH5OrqKil3edGPP/7Y6jheXl4KDAzMt0Tp2rVrlZGRQZUhAAAAAAAAAACAFSQNUST+/v6qW7euEhMTFRMTI0nav3+/Tp48qZ49e8rDw6NI43p5ealfv36aMmWKVq1apZYtW0qS5s+ff8tlSgtaotRc+di/f/8ixQMAAAAAAAAAAHJXISztB0oXSUMUicFgUEhIiKTfk3S2LE2anZ2tRx55RPfee6+2bt1q9Rp169bVK6+8Iil3mdELFy5Ybd+nTx85Oztry5YtSkpK0tWrVxUVFaUWLVpYko8AAAAAAAAAAADIj6Qhisxc2bdq1SqZTCatWLFC7u7u6tmzZ6F9nJycdPHiRZ05c0br16+/5TVq164tSTIajfL09LTa1tPTU506dbIsUbp69WplZWWxNCkAAAAAAAAAAMVkMBpL/YHSxYyjyO6++27Vq1dPZ86c0XfffaeEhAQFBwercuXKVvsNHDhQkvTf//5Xhw4dstp22bJlkqSgoCDLHofW3JzINFdAsjQpAAAAAAAAAACAdSQNUSyhoaGSpA8//FCSbQm6YcOGycfHR2lpaQoLC9O8efOUnJycp83ly5c1depUzZo1S5UrV9a4ceNsiqdPnz5ycXFRdHS0Nm/eLD8/PzVu3NjOuwIAAAAAAAAAADdjT8Pyj6QhisVc2Zeamqrq1aura9eut+zj6emp//u//9Pdd9+t5ORkTZo0SUFBQQoJCdEjjzyikJAQBQUFadasWfLy8tLHH38sPz8/m+KpXr26OnfurKysLGVlZVFlCAAAAAAAAAAAYANnRweAsq1t27Zq0KCB4uPjLVV+tqhXr57++9//as2aNVqzZo1iYmJ0+fJlxcfHq3r16vL391dwcLAGDRqk6tWr2xXTfffdp8jISBkMBvYzBAAAAAAAAACgJBip/CvvSBrilho0aGB178E1a9YUeu6hhx7SQw89VOA5g8Gg3r17q3fv3nbH9O677+rdd98t8NyDDz6oBx98sNC+a9eutft6AAAAAAAAAAAA5RnLkwIAAAAAAAAAAAAVHJWGAAAAAAAAAAAAsMpgoA6tvOMZBgAAAAAAAAAAACo4Kg0BAAAAAAAAAABglcFocHQIuM2oNAQAAAAAAAAAAAAqOCoNAQAAAAAAAAAAYJXBSB1aecczDAAAAAAAAAAAAFRwVBoCAAAAAAAAAADAOkPZ3dNw//79mj17trZt26YrV66oRo0aat++vYYOHar27duXyDVu3LihQYMGaf/+/dqwYYPuuOMOq+1PnjypWbNmafPmzbpw4YKqVasmPz8/PfHEE+rRo0eJxGQvkoYAAAAAAAAAAAAol3755ReNHz9eN27csBw7f/68fvnlF61atUrjx4/X8OHDi32dqVOnav/+/Ta13blzp55++mmlp6dbjl26dEmRkZGKjIxUWFiYXnvttWLHZC+WJwUAAAAAAAAAAEC5s2/fPr300ku6ceOGOnbsqPnz5ys6Olrh4eEKDAxUTk6OPvjgA61fv75Y15kxY4a+/vprm9qeOXNGzz77rNLT09WiRQvNnj1b0dHRWrJkiUJDQyVJ8+bN03fffVesmIqCpCEAAAAAAAAAAACsMhgNpf4oro8++kiZmZlq0aKFvv76a3Xo0EFeXl5q3769Zs+erXvuuUcmk0nvvfeecnJy7B4/JSVFzz//vD799FOb+8yaNUtXrlxRrVq1NHfuXHXt2lVeXl5q3bq1PvroIw0YMECS9PHHHys1NdXumIqDpCEAAAAAAAAAAADKlWPHjikyMlKSNGrUKFWqVCnPeRcXF7344ouWtrt27bJ57JycHP3www8KDQ3VypUrJUl+fn637Hf16lUtXrxYkjR06FB5eXnla/Pyyy/LyclJSUlJWr16tc0xlQT2NASKoc49gY4OocypXLuOo0Mocxr0DHZ0CGVS1cY+jg6hzHHxqOroEMoc1xr5v9jhFgwGuVSt5ugoypy6QV0dHUKZ49GwsaNDKHNqtrnL0SGUOT6N+G5bFPV73OvoEMocg5OT3OrVd3QYZQrf0+zHnBVN5Vq1HR1CmcP3NPu5uLk5OgRAkmQwlK06tI0bN0qSnJyc1L179wLbtGvXTjVr1tSlS5e0Zs0adezY0aaxDx06ZNlzsEaNGnrjjTeUnp6uiRMnWu23detWZWRkSJLuvbfg78V16tRRq1attG/fPq1Zs0b333+/TTGVhLL1DAMAAAAAAAAAAAC3cPDgQUlSw4YNVbVqwf9Y3mAw6M4775Qk7d+/367xXV1dFRYWpuXLl6tfv352xeTq6qrmzZsX2q5169ZFiqm4qDQEiiErrXTXEy7LXNzcJTFn9mDOisYyb6kpDo6k7HBx95DEnNnDPGeZaWkOjqQMMpl4X7MDnwX2Y87sZ5mzFD4HbOXikfs5kHKVObOHRzW+cxQZn58243PAfuY547utfVz/V/nFa812/H7aj/+Po2jMf7PjNiiBPQZLU0JCgiSpfn3rKzbUrVtXkhQfH2/z2A0aNND69etVs2bNIsVUr149GQyFz6c5prNnzyonJ0dGY+nUAJI0BAAAAAAAAAAAQLmSlJQkSapWzfo2KeYqxKtXr9o8dmGViyUdU3Z2tlJSUm7ZvqSQNAQAAAAAAAAAAIBV1irjbpdevXpZPb9mzZpCz5n3DqxcubLVMSpVqpSn/e1ka0w3ny+NuMzY0xAAAAAAAAAAAADlipOTk6NDyOfPGNPNqDQEAAAAAAAAAADAn461SsJbqVKliqRbV+rZWv1XEmyN6fr165afSyMuM5KGAAAAAAAAAAAAsMpgLFuLV5r3Bbx27ZrVdua9DGvUqHHbYzLvTWhrTC4uLvLw8LjtcZmVrWcYAAAAAAAAAAAAuIUmTZpIks6ePWu1XWJioiSpbt26tz0mHx8fSbbHdMcdd5TqXpIkDQEAAAAAAAAAAGCdwVD6j2Jo0aKFJOnkyZNKS0srsI3JZNLBgwclSa1bty7W9eyJKT09XXFxcYW2279/f6nFdDOShgAAAAAAAAAAAChXevToIUnKysrSxo0bC2wTExOjy5cvS5K6det222MKDAy07FG4du3aAtskJiZaEpmlEdPNSBoCAAAAAAAAAADAKoPRWOqP4mjUqJH8/f0lSTNmzFBqamqe81lZWZo6daokqWXLlurcuXOxrmcLd3d39e7dW5I0e/ZsnT9/Pl+b999/Xzk5OfLy8tJf/vKX2x7TzUgaAgAAAAAAAAAAoNyZOHGiDAaDjh49qrCwMG3ZskVJSUmKiYnR8OHDtX37dhkMBr3wwgt59g7cu3evQkNDFRoaqm+//bZEYxo3bpyqVKmiixcvasiQIVq7dq0uX76s2NhYjR07VhEREZKkf/zjH5aqxNJC0rCciI+Pl6+vr+UxadIkm/rNnj3b0qd79+4Ftlm3bp3Gjx+vXr16qV27dvL391fv3r01fvx4rVq1qtCxJ0yYYBm7devWlhJfa7KyshQQEGDp98MPP0iS1qxZYzlWUObdbNasWZZ2zz//fKHtTCaTAgMD5evrq3nz5t0yLgAAAAAAAAAAKjKDwVDqj+Jq166d3nrrLRmNRu3fv19Dhw5Vp06dNHjwYG3ZskVSbi4jODg4T7/09HSdOHFCJ06cUFJSUrHjuFn9+vU1ffp0VapUSadOndKoUaPUuXNn3X///VqxYoUk6W9/+5uGDBlSote1BUnDcmrlypUymUy3bLd8+fJCz12/fl0jRozQyJEjFRERoYyMDLVo0UJNmzbVtWvXFBERodGjR+vxxx9XcnKy1etkZ2dbTTCabdq0qcCxAgMD5eTkJEnavXt3of1vXpc4Ojpa2dnZBbY7dOiQrly5Iknq0qXLLeMCAAAAAAAAAABlz+DBg7Vw4UINGDBA3t7ecnFxkaenp+69917NmTNHTz75ZKnH1LNnT0VEROiRRx5R/fr15eLioqpVq6pTp06aMWOG/vnPf5Z6TJLk7JCr4rZydnbW+fPntXPnTnXs2LHQdqdPn9a+ffsKPf/GG29ow4YNatq0qd577z3dddddlnMmk0lRUVGaMGGCdu7cqeeee67QEl1nZ2fduHFDv/zyix599FGrsReWxPTw8NBdd92lmJgY7dq1SyEhIfnapKSkaPfu3XJzc1P16tV19uxZ7d69u8A52LFjh6TcjH7Tpk2txgQAAAAAAAAAAMouPz8/y/6FtggMDNShQ4fsusZDDz2khx56yOb2jRo1snnVyNJCpWE51KlTJ0nSL7/8YrWdOUHXunXrfOcSEhL0008/SZI++eSTPAlDKbcMuVu3bpo2bZokafv27YqOji7wOgEBATIajdq2bZvVJUozMzO1Zs0aeXl56Y477ij0vnbt2lVg/y1btigrK0vt27e3VA/eXHl4s+3bt0uSgoKCCo0HAAAAAAAAAADkMhgNpf5A6SJpWA6FhoZKklatWmV1idLly5fLaDSqX79++c4dOHBAOTk58vDwULNmzQodIyAgQD4+PpKkPXv2FNimZs2a6tixo7Kzs/Xrr78WOlZkZKRSUlIUEhJiWYr0ZuYE38GDB3X9+vV8580JwqCgIHXr1i3PsT8yVxqyNCkAAAAAAAAAAABJw3KpY8eOql27ts6dO1doVd7x48cVGxurgIAA1apVK995FxcXSblLfpoTbIWZNWuWVq5cqbCwsELbmBOT1qofzZWPAwYMKPC8v7+/qlSpoqysLO3duzff+aioKElS165d1blzZxmNRh04cCBfdePx48d18eJFGY1Gde7c2eq9AQAAAAAAAAAASQZj6T9QqpjxcshoNFr2/CssSWdO0PXv37/A8+3bt5ebm5skacSIEZo2bZqOHj1aYNvGjRvLx8dH7u7uhcZkrh4sbInS9PR0rVu3TnfccYc6dOhQ4Biurq6Wc39Mhh4/flzx8fHy9vaWr6+vqlevrrvuuksmkylftaE5Cern5ydPT89CYwYAAAAAAAAAAKgoSBqWU+bKvsKWKF2xYoVcXFzUt2/fAvtXq1ZNEydOlCSlpqbq888/V//+/dW9e3eNHz9e4eHhiouLszmemjVr6p577tGNGze0evXqfOfXrVuntLQ09evXTwZD4esUF7av4c1VhmaF7Wto3s+QpUkBAAAAAAAAALANexqWfyQNy6kOHTrI29tbiYmJ2r17d55zhw4d0tGjRxUUFGS10m7QoEGaOXOm6tevbzl27tw5RURE6M0331RISIjuu+8+LV682OreiWbWlihdsWKFJOm+++6zOoZ5X8OYmJg81zQnBgtKGm7atClPW3Ol4c1tAQAAAAAAAAAAKjKShuWUwWBQaGiopPxJOvPSpLdK0ElScHCwfv31V3399dcKCwtT8+bN85w/duyYJk6cqFGjRikzM9PqWH379pWzs7O2bt2qpKQky/GUlBRt2LBBjRo1Utu2ba2O0bp1a3l6eio5OdmyXGpmZqa2b98uo9FoSSpK0t133y0PDw9dvnxZsbGxkqT4+HidOXNGbm5uuvvuu295/wAAAAAAAAAAQDIYjKX+QOlixssxc2XfypUr81TarVixQpUqVVLv3r1tGsfJyUldunTRa6+9pmXLlik6OlrTp0/XAw88IFdXV0m5y4t+/PHHVsfx8vJSYGBgviVK165dq4yMDJuSmAaDQYGBgZJ+X6J027ZtSk9Pl5+fn2rUqGFp6+zsbFnOdOvWrZJ+rzIMDAyUi4uLTfcPAAAAAAAAAABQ3pE0LMf8/f1Vt25dJSYmKiYmRpK0f/9+nTx5Uj179pSHh0eRxvXy8lK/fv00ZcoUrVq1Si1btpQkzZ8//5bLlBa0RKm58rF///42Xd9cTWhOGpr3Myxoj0LzsS1btkhiP0MAAAAAAAAAAICCkDQsxwwGg0JCQiT9nqSzZWnS7OxsPfLII7r33nstFXqFqVu3rl555RVJucuMXrhwwWr7Pn36yNnZWVu2bFFSUpKuXr2qqKgotWjRwpJ8vJXOnTtLkvbs2SMpd89CSerWrVu+tuZ9C3fu3Cnp90pDkoYAAAAAAAAAANjBaCj9B0oVScNyzlzZt2rVKplMJq1YsULu7u7q2bNnoX2cnJx08eJFnTlzRuvXr7/lNWrXri1JMhqN8vT0tNrW09NTnTp1sixRunr1amVlZdm0NKlZ48aNVb9+fcXFxen06dM6cuSIPDw8CtyjsFGjRmrYsKGuXr2qvXv3Ki4uTvXq1VPTpk1tvh4AAAAAAAAAAEB5R9KwnLv77rtVr149nTlzRt99950SEhIUHBysypUrW+03cOBASdJ///tfHTp0yGrbZcuWScpdNtS8x6E1NycyzRWQti5NatapUyeZTCbNnTtXJpNJnTp1krOzc4FtzdWGCxYssMQJAAAAAAAAAABsZzAYSv2B0kXSsAIIDQ2VJH344YeSbEvQDRs2TD4+PkpLS1NYWJjmzZun5OTkPG0uX76sqVOnatasWapcubLGjRtnUzx9+vSRi4uLoqOjtXnzZvn5+alx48Z23ZM58bdo0SJJvycGC2I+Z16alaVJAQAAAAAAAAAA8iq4NAvlSr9+/fT1118rNTVV1atXt5pgM/P09NT//d//6YUXXlBMTIwmTZqkd999Vw0aNFC1atV09epVnTx5UiaTSTVr1tS7774rPz8/m+KpXr26OnfurMjISEn2VxlKuZWGBoNBqampkqwnDc1ViGlpaTIajVQaAgAAAAAAAABgJ4OROrTyjme4Amjbtq0aNGgg6fcqP1vUq1dP//3vf/Xpp5/qoYceUqNGjXTlyhUdOHBA165dk7+/v1588UWtWLFC3bt3tysm8x6GBoPBrv0MzWrVqqUWLVpIyt3jsGHDhoW29fDwULt27SRJfn5+t9x3EQAAAAAAAAAAoKIxmEwmk6ODAMqqrLRUR4dQZri4uUtizuzBnBWNZd5SUxwcSdnh4u4hiTmzh3nOMtPSHBxJ2dG0ZUslJCSofr16OnHksKPDKTP4LLAfc2Y/y5yl8DlgKxeP3M+BlKvMmT08qvGdw15NfO/k89NOfA7YzzxnfLe1j6ubmyRea/bg99N+/H8cRWP+mx0lb8Nzo0r9mj0+nVnq16zIqDQEAAAAAAAAAAAAKjiShgAAAAAAAAAAAEAF5+zoAAAAAAAAAAAAAPDnZjBSh1be8QwDAAAAAAAAAAAAFRyVhgAAAAAAAAAAALDOYHB0BLjNqDQEAAAAAAAAAAAAKjgqDQEAAAAAAAAAAGAVexqWfzzDAAAAAAAAAAAAQAVHpSEAAAAAAAAAAACsMrCnYblHpSEAAAAAAAAAAABQwZE0BAAAAAAAAAAAACo4licFAAAAAAAAAACAdUaWJy3vqDQEAAAAAAAAAAAAKjgqDYFicHFzd3QIZQ5zZj/mrGhc3D0cHUKZw5zZz9XNzdEhlD0GA+9rRcCc2Y85s5+LB58D9vKoxpwVBd85ioDPT7sxX/bju23R8FqzH3NmPz478WdhMFCHVt7xDAMAAAAAAAAAAAAVHJWGQDEkbo12dAhlxh2BnSVJWSkpDo6k7DD/a/9z27c6OJKypc49gZKktDMJDo6k7HCrV1+SlHXtqoMjKTtcqlaTJGWmpTk4krInOzNTZ6IiHR1GmVGva3dJ0vUL5xwcSdlRuXYdSdLlA/scHEnZ4dW6jSTpWtxxB0dSdlT1aSpJOnv2koMjKVvq1q0pSYpbHuHgSMqenBs3lBR7wNFhlAk17mwtSTo4Z7aDIyk7Wj35tCTpyPf/dXAkZUuLQY9KkhK3bHZwJGXHHZ2CJIm/B+xg/nuA/4/DPub/nwMlz8CehuUelYYAAAAAAAAAAABABUelIQAAAAAAAAAAAKxjT8Nyj2cYAAAAAAAAAAAAqOBIGgIAAAAAAAAAAAAVHMuTAgAAAAAAAAAAwCqD0eDoEHCbUWkIAAAAAAAAAAAAVHBUGgIAAAAAAAAAAMAqg4E6tPKOZxgAAAAAAAAAAACo4Kg0BAAAAAAAAAAAgHXsaVjuUWmIAsXHx8vX19fymDRpkk39Zs+ebenTvXv3AtusW7dO48ePV69evdSuXTv5+/urd+/eGj9+vFatWpWvfXJyslq1aiVfX19FREQUeu2YmBjLtTt37iyTyVRo2+eee06+vr4aPXq0TfcFAAAAAAAAAABQnpE0hE1WrlxpNQlntnz58kLPXb9+XSNGjNDIkSMVERGhjIwMtWjRQk2bNtW1a9cUERGh0aNH6/HHH1dycrKlX/Xq1dW6dWtJ0u7duwsdf+PGjZafL1++rH379hXYzmQyaceOHZKkLl263PKeAAAAAAAAAACo6AwGY6k/ULqYcdySs7Ozzp8/r507d1ptd/r06UITdZL0xhtvaMOGDWratKkWLlyoqKgoLVy4UIsWLdKWLVv01VdfqVatWtq5c6eee+65PH2DgoIkSbt27Sp0fHPSsHnz5nn++4+OHDmiK1euSCJpCAAAAAAAAAAAIJE0hA06deokSfrll1+stjNXGZqrAm+WkJCgn376SZL0ySef6K677spz3mAwqFu3bpo2bZokafv27YqOjrac79y5syTp0KFDSk1NzTf+lStX9Ntvv8nNzU1PPfWUpMKThtu3b5ckNW7cWA0bNrR6TwAAAAAAAAAAABUBSUPcUmhoqCRp1apVVpcoXb58uYxGo/r165fv3IEDB5STkyMPDw81a9as0DECAgLk4+MjSdqzZ4/leIcOHVSpUiVlZ2fnOW62efNm5eTkKCAgQD179pTBYNDevXt17dq1fG3NSUOqDAEAAAAAAAAAsJHRUPoPlCqShriljh07qnbt2jp37lyhy4MeP35csbGxCggIUK1atfKdd3FxkSSlpKRY9hMszKxZs7Ry5UqFhYVZjlWqVEn+/v6SCl6i1FxV2LVrV9WsWVN33nmnbty4oU2bNuVry36GAAAAAAAAAAAAeZE0xC0ZjUaFhIRIKnyJUvPSpP379y/wfPv27eXm5iZJGjFihKZNm6ajR48W2LZx48by8fGRu7t7nuPW9jW8OWl48//+cYnSuLg4XbhwQc7OzpZlVwEAAAAAAAAAgHUGg7HUHyhdzDhsYl5ytLAlSlesWCEXFxf17du3wP7VqlXTxIkTJUmpqan6/PPP1b9/f3Xv3l3jx49XeHi44uLirMZg3tcwJiZG2dnZluOxsbG6cOGC6tevryZNmkj6vYowKioqzxjmpUnbtm0rDw+PW902AAAAAAAAAABAhUDSEDbp0KGDvL29lZiYqN27d+c5d+jQIR09elRBQUHy9PQsdIxBgwZp5syZql+/vuXYuXPnFBERoTfffFMhISG67777tHjx4gITk23atFG1atWUmpqqw4cPW47/scrQHG+VKlWUmJiYpy37GQIAAAAAAAAAYD+D0VDqD5QukoawicFgUGhoqKT8S5Salya97777bjlOcHCwfv31V3399dcKCwtT8+bN85w/duyYJk6cqFGjRikzMzPPOaPRqMDAQEl5lyg1Jw27detmOebq6qp77rlHkrR582bLcfYzBAAAAAAAAAAAyI+kIWxmXqJ05cqVeSoBV6xYoUqVKql37942jePk5KQuXbrotdde07JlyxQdHa3p06frgQcekKurqyRp3bp1+vjjj/P1NS9Rak4apqamateuXXJ2dracMzNXHm7dulWSdPbsWSUkJKhatWpq27atPbcOAAAAAAAAAEDFZjCU/gOliqQhbObv76+6desqMTFRMTExkqT9+/fr5MmT6tmzZ5H3CPTy8lK/fv00ZcoUrVq1Si1btpQkzZ8/P98ypX9MGm7dulVZWVkF7lFoThru2LFDOTk52rZtmySpU6dOcnJyKlKsAAAAAAAAAAAA5RFJQ9jMYDAoJCRE0u9LlNqyNGl2drYeeeQR3XvvvZaqv8LUrVtXr7zyiiQpJSVFFy5cyHO+adOmuuOOO3TmzBlduHDBsvTozfsZmjVr1kx169bV1atXdfjwYe3cuVMSS5MCAAAAAAAAAAD8EUlD2MW8ROmqVatkMpm0YsUKubu7q2fPnoX2cXJy0sWLF3XmzBmtX7/+lteoXbu2pNw9DD09PfOdN1cb7t27V9u3b5eUdz/DmwUFBUmSDh48SNIQAAAAAAAAAIAiMhiNpf5A6WLGYZe7775b9erV05kzZ/Tdd98pISFBwcHBqly5stV+AwcOlCT997//1aFDh6y2XbZsmaTchJ95j8ObderUSZIUHR2tw4cPy9PTU23atClwLHMF4vbt23Xs2DE1atRIDRs2tH6TAAAAAAAAAAAAFQxJQ9gtNDRUkvThhx9Kkvr373/LPsOGDZOPj4/S0tIUFhamefPmKTk5OU+by5cva+rUqZo1a5YqV66scePGFTiWudJw8eLFysnJUVBQkIyF/IsD87kVK1bIZDJRZQgAAAAAAAAAQBEYDIZSf6B0kTSE3cxLlKampqp69eoF7if4R56envq///s/3X333UpOTtakSZMUFBSkkJAQPfLIIwoJCVFQUJBmzZolLy8vffzxx/Lz8ytwrDp16qhZs2ZKTU2VVPB+hjdf18/PT2lpabdsCwAAAAAAAAAAUFE5OzoAlD1t27ZVgwYNFB8frz59+sjFxcWmfvXq1dN///tfrVmzRmvWrFFMTIwuX76s+Ph4Va9eXf7+/goODtagQYNUvXp1q2MFBQXp2LFjkm6dCOzatat+++03OTs7W5Y2BQAAAAAAAAAAdmCPwXKPpCEK1KBBA6t7D65Zs6bQcw899JAeeuihAs8ZDAb17t1bvXv3LlZ8r732ml577TWb2o4dO1Zjx44t1vUAAAAAAAAAAADKM5KGAAAAAAAAAAAAsIo9Bss/akkBAAAAAAAAAACACo6kIQAAAAAAAAAAAFDBsTwpAAAAAAAAAAAArDNQh1be8QwDAAAAAAAAAAAAFRyVhgAAAAAAAAAAALDKYDQ4OgTcZlQaAgAAAAAAAAAAABUclYYAAAAAAAAAAACwysCehuUezzAAAAAAAAAAAABQwVFpCAAAAAAAAAAAAOvY07Dco9IQAAAAAAAAAAAAqOBIGgIAAAAAAAAAAAAVHMuTAgAAAAAAAAAAwCqDgTq08s5gMplMjg4CAAAA5VuDBg2UkJCg+vXrKz4+3tHhAABQJvD5CQAA/kz2TP+w1K/Zbuy4Ur9mRUalIQAAAAAAAAAAAKwyGA2ODgG3GUlDoBiy0lIdHUKZ4eLmLknKSktzcCRlh4ubmyTmzF6/zxu/n7bi99N+vM6KwWRSVso1R0dRZrh4VJXEa80elve0a1cdHEnZ4VK1miReZ/Ywv84yLl90cCRlSyWvWpKYtyIxmfgdtdHv322ZL1sxZ0XDvNmPObMff68XjflvdgD2I2kIAAAAAAAAAAAA69jTsNzjGQYAAAAAAAAAAAAqOCoNAQAAAAAAAAAAYBV7GpZ/VBoCAAAAAAAAAAAAFRxJQwAAAAAAAAAAAKCCY3lSAAAAAAAAAAAAWGegDq284xkGAAAAAAAAAAAAKjgqDQEAAAAAAAAAAGCVwWhwdAi4zag0BAAAAAAAAAAAACq4YiUNMzMz9f3332vkyJHq2bOn2rZtq7vvvlv33XefXn/9de3ateuWYxw5ciTfseDgYPn6+uqHH34ocmxfffWVfH19tXr1aknSxx9/LF9fXwUHB9vUPywsTL6+vpowYUKRY7hZce7pxo0b6tixox544AFJUnx8vHx9fS2PSZMm2TTO7NmzLX26d++e59yECRPk6+urF1980e74issc0+bNm0v92gAAAAAAAAAA4NYMBmOpP1C6ijzjUVFR6tu3r15//XWtW7dO169fV/PmzeXt7a1Tp07p+++/12OPPabnn39eKSkp+fqfP39e48eP1/Dhw4t1A4XZsGGDXFxcFBQUdFvGL027du3StWvX1KNHjwLPr1y5UiaT6ZbjLF++vKRDAwAAAAAAAAAAQDlQpD0Nly5dqldffVXZ2dnq2LGjXnjhBXXo0EEGQ+56tikpKfrhhx/06aefauXKlTp69Ki++eYb1a5d2zJGVFSUIiIiVKdOnZK5k5ukpKRo9+7duueee+Tm5lbi45e2DRs2SFK+6kBJcnZ21vnz57Vz50517Nix0DFOnz6tffv2FXp+3Lhx+vvf/66qVasWP2A7mZOZ9erVK/VrAwAAAAAAAAAAG7CnYblnd6Xh/v379frrrys7O1uPPvqo5s2bp44dO1oShpLk4eGhYcOGKTw8XN7e3jp27JheffXVEg3cmk2bNikrK6vQyryyJjIyUp6enrr77rvznevUqZMk6ZdffrE6hjkx17p16wLPe3t7q1mzZvL29i5esEXQrFkzNWvWTFWqVCn1awMAAAAAAAAAAKAIScP33ntPmZmZuuuuu/Svf/1LRmPhQ7Ro0UJvv/22pNzE19KlS4scqD0iIyMlFVyZV9YkJibq8OHD6tKli5ycnPKdDw0NlSStWrXK6hKly5cvl9FoVL9+/W5brAAAAAAAAAAAoHwyGI2l/kDpsmvGjxw5oi1btkiSnn766QKTWH907733yt/fX5L07bffSpJ8fX01ceJESdK5c+fk6+srX1/fAvvv3LlTI0eOVGBgoNq2bauQkBBNnTq1wH0SzSIjI9WgQQM1bdrUntuzSVhYmHx9fRUZGanY2FiNGTNGQUFBatOmjXr16qX//Oc/unz5ss3jXb58WQMGDJCvr6/+8pe/6NKlS3nOW1uaVJI6duyo2rVr69y5c9q1a1eBbY4fP67Y2FgFBASoVq1aBbaZMGGCfH199eKLL+Y5np2drfnz5+uxxx5Thw4d1KZNG3Xt2lXPPfec1q1bV+BYx44d08SJExUcHKw2bdqoQ4cOeuCBBzRt2rR89yfJ8vxv3rzZcmzx4sXy9fXVCy+8oLS0NE2fPl0hISG66667FBgYqJEjR2rHjh0FXl+SNm/erBEjRqhr165q27at7r//fn333XfKycmx+noDAAAAAAAAAACoiOxKGpqTOkajUd26dbO5X58+fSRJ+/bt06VLl9S+fXv5+PhIklxcXNS+fXu1b98+X78lS5ZoyJAh2rx5s+rWrSsvLy/FxcVp1qxZevTRR3X9+vV8fQ4ePKjz58/f9qVJIyMj9fDDD2v16tWqUaOG6tatq/j4eH3zzTd69NFHrSY1zZKSkvTkk0/qyJEjat26tebOnauaNWvmabNhwwYZDIZC59toNCokJERS4UuUmpcm7d+/vz23KJPJpBdeeEFvvfWWdu3apZo1a8rX11fZ2dlavXq1Ro4cqY8++ihPn927d+vhhx/W4sWLde3aNbVo0UJ16tTR4cOH9fnnn+vBBx/U2bNnbY7h6tWrGjx4sGbOnKm0tDQ1b95caWlpWrdunf72t79p/fr1+fp8+umnGjZsmDZs2CCTyaQWLVooPj5eb7/9tsaMGWPXHAAAAAAAAAAAAFQEdiUNY2NjJUn169eXh4eHzf3M++iZTCYdPnxY4eHheuaZZyRJXl5eCg8PV3h4eL5+O3fu1IABAyxLm65fv17vv/++pNyqx8WLF+frY67Mu91Jw3nz5qlLly5at26dli1bpl9//VWfffaZnJycdPLkSS1cuNBq/+TkZD311FM6dOiQ7rrrLn3zzTeqUaNGnjaZmZmKjo5WmzZt8iUTb2ZecrSwJUpXrFghFxcX9e3b16573Lhxo1auXCkvLy/99NNPWrVqlRYtWqSoqCiNGzdOkvTFF18oMTHR0mfy5MlKS0tTWFiYNm3apCVLlmj58uVauXKlfHx8dO7cOc2cOdPmGKKiopSUlKTZs2dr48aNWrJkidasWWNJXk6bNi1f+xkzZshoNOq1117Txo0btWjRIm3atElPPPGEVq1aZdccAAAAAAAAAAAASQZD6T9QquxKGiYlJUmSPD097brIzQkve5bubNasmaZMmZLnegMHDlSXLl0k5SYV/ygyMlKVKlVSYGCgXTHaq2bNmpoxY4a8vb0tx3r16mVZRrSwpUIl6dq1axo2bJgOHDggf39/zZkzR9WqVcvXbufOnUpLS7tlArRDhw7y9vZWYmKidu/enefcoUOHdPToUQUFBdn9vJmTxP7+/nmW83RyctIzzzyj0NBQDRgwQMnJyfn6/PWvf5Wrq6vleMOGDfXKK6/o3nvvVf369e2K41//+pe6du1q+W9vb2/94x//sFwvNTXVcs5c+fjkk08qLCzMsudm5cqV9frrr5eLfS4BAAAAAAAAAABKml1Jw4yMDEm5S4ra4+a9DwuqhCtMr169Ctw30ZzA+mMC8urVq4qJiVFAQIAqV65sV4z26ty5sypVqpTveLNmzSTlJgYLkpKSoqeeekr79+9Xu3btNHv27EKrNm+1n6GZwWBQaGiopPxLlJqXJr3vvvusjlEQ8xKyGzZs0BdffJFvWdGPPvpI7733Xp6EYuPGjSVJb7zxhqKjo5WVlWU5FxwcrM8//9xSZWoLJyenAu/fPM+SLEvBnjt3Tr/99psk6fHHHy9wvKFDh9p8bQAAAAAAAAAAkMtgNJb6A6XLrhk3V6rdXFlmC3OFopS7HKmt6tSpU+BxNzc3Scq3p2FUVJSys7MLrMwz2vniMic3C+tXWGzmZOWNGzcKPD99+nTt3btXknTp0iUZrJTXbtiwQV5eXrrrrrtuGa95idKVK1fmScyuWLFClSpVUu/evW85xh8FBwcrICBAN27c0IcffqiePXuqX79+euedd7R+/XpLEvlmL730kpydnbVnzx49+eSTCggI0IgRIzRnzhzFxcXZHUP16tULTADfnLA1z/WRI0dkMpnk5uamhg0bFjhemzZt7I4BAAAAAAAAAACgvLMrk3bnnXdKkk6dOqX09HSb+x08eNDyc8uWLW3uV1AlnzXW9jOsUqWKJBWY6CpIWlpann5/ZG+1pdn169cVEBCg2rVrKz4+3rJH4x+dPn1ax48fV9euXW1KePr7+6tu3bpKTExUTEyMJGn//v06efKkevbsadcelGbOzs6aPXu2JkyYYHnejh8/rm+//VbPPPOMgoKCNHPmzDxJyu7du2vhwoXq37+/3N3dlZaWpg0bNmjy5MkKCQnR448/rqNHj9ocgy3zbL6+OTnt7u5eaNuizAMAAAAAAAAAABWdwWAo9QdKl11Jw549e0qSsrKytH79epv7rV69WpLk5+enWrVq2XNJm5lMJm3cuFE+Pj5q1KhRvvO1a9eWJF25ckXZ2dm3HO/8+fN5+pWULl266Msvv9Rrr70mSQoPD9eWLVvytYuMjJRUcAK0IAaDQSEhIZJ+X6K0OEuTmrm6umrYsGH6+eeftWHDBk2ZMkUPPfSQatWqpZSUFE2fPl1z5szJ06dVq1b68MMPtXXrVn333Xd6/vnn1bFjRxmNRu3cuVNPPvmkJSlbkswJXvNypQW5ef9DAAAAAAAAAAAA5LIraejr66vAwEBJ0ueff67MzMxb9tm+fbslKfbEE08UIUTb7Nu3T5cuXSo0yWaukrxx44YOHTpkdazExERduHAhT7+S0q9fP1WuXFmhoaHq3bu3TCaT/vnPf+ZLokVGRspoNKpr1652jS1Jq1atkslk0ooVK+Tu7m5J9torOTlZMTExlr0M77jjDj3wwAOaPHmy1q9fr3vvvVeS9OOPP0qSsrOzdfLkSW3fvl1SbpVgx44d9dxzz+m7777Td999J4PBoAsXLmjz5s1Fiska896K6enpOnXqVIFtYmNjS/y6AAAAAAAAAACUe0Zj6T9Qquye8XfeeUdubm6KjY3VG2+8YbVq79SpU3rppZck5VbYPfjgg79f+H9P9s1LWxaHeWnS7t27F3i+RYsW8vHxkSR9+eWXVseaP3++pNw9HM1J0tvhX//6l6pWrZpvmdKMjAxt3bpV7dq1s+wjaYu7775b9erV05kzZ/Tdd98pISFBwcHBBe4JaItXX31VgwcPLnC+XFxcFBAQIEmW18CRI0fUt29fDR061JJ0vZm/v79l6dCcnJwixWRNw4YNLUnehQsXFthmwYIFJX5dAAAAAAAAAACAss7upGHjxo3173//W66urlq8eLGGDh2q3bt352mTlpamBQsWaNCgQTp79qwaNWqkKVOm5Fl/1s3NTVJuNZu15SRtFRkZKTc3N0siqyD//Oc/JeUu2/nPf/5T586dy3P+2rVr+uSTT/TVV19Jkl566aVC9zQsCXXq1NHLL78sKe8ypVu3blV6enqhCVBrQkNDJUkffvihJKl///5Fju/++++XlJtoW7p0aZ4E75EjRzRv3jxJvy+heuedd6ply5bKzs7WuHHjlJiYaGmfmZmpadOmKSUlRW5uburYsWOR47Jm9OjRkqTZs2fr+++/t8SclZWljz/+WMuWLbst1wUAAAAAAAAAAH9O+/fv17hx49S1a1e1adNG3bp105gxY7Rr165ijbtlyxaNHDlSnTp1Ups2bdSzZ09NnDhRhw8fttpvxowZ8vX1veXj22+/LVZ89nIuSqf77rtP9evX19ixY7V9+3Y9+uijqlmzpurWrauMjAydPHnSsnRpv379NGnSJHl4eOQZw9fXV0ajURkZGQoNDZW3t7dmz56tGjVq2B3P5cuX9dtvv6lHjx5ydXUttF337t31zjvv6J133tHChQu1ZMkSNWjQQJ6enrp69apOnTql7OxsOTk56dVXX9XDDz9sdyz2euSRR/Tzzz9r27ZtevXVV/Xzzz/bvZ/hzfr166evv/5aqampql69ul3Lm/5R3759NWjQIH3//fd65ZVXNGXKFNWtW1cpKSk6deqUTCaT2rZtq5EjR1r6TJs2TY8++qi2bdum3r17q0GDBqpSpYri4+N19epVOTk56e2335aXl1eR47Kmd+/eGj58uL766iu9/vrrmjFjhurWrauTJ08qOTlZ7dq10549e+Tk5HRbrg8AAAAAAAAAQHl0c2FYWfLLL79o/PjxunHjhuXY+fPn9csvv2jVqlUaP368hg8fbve4c+bM0eTJk/McO3v2rBYvXqyIiAj95z//0V/+8pcC+x44cMDu65WGIiUNJaldu3ZauXKlfvzxR/366686cOCADh06JCcnJzVq1EgdO3bUgw8+qLvvvrvA/o0bN9bkyZM1c+ZMJSQkKDs7WwkJCUVKGkZFRSknJ8emJNugQYPUqVMnLV26VJGRkUpMTNSZM2dUvXp13XnnnercubMGDx6sRo0a2R1HURgMBk2aNEkDBw5UQkKC3n//fW3evFm1a9dW69at7R6vbdu2atCggeLj49WnTx+5uLgUK7633npL/v7+Wrp0qQ4dOqRDhw7J3d1dHTp00H333adBgwbluUbz5s21ZMkSzZ49W9HR0Tpz5oxMJpO8vb3Vp08fDRs2TC1atChWTLfy0ksvyd/fX99++63279+v2NhYNW/eXC+++KKaNWumxx9/vMhLtgIAAAAAAAAAgLJh3759eumll3Tjxg117NhR48aNU5MmTRQXF6fp06dr69at+uCDD9S8eXP17NnT5nHXrl2rd999V5LUp08fjRo1SnXr1lVsbKymTJmi2NhYTZw4Uc2aNSsw17N//35JufmMxx57rNDrWCuUux0MppLaVBAoA9avX69nnnlGPj4+WrlyZbHHy0pLLYGoKgYXt9z9LLPS0hwcSdnh8r9lnJkz+/w+b/x+2orfT/vxOrNfk5a+SkhIUP169XTiUKyjwykzXDyqSuK1Zg/Le9q1qw6OpOxwqVpNEq8ze5hfZxmXLzo4krKlklctScybPZq1vfv3z88j1pe4Qq7fv9vynmYr5qxomDf7MWf24+/1ojH/zY6Sd3j+vFK/ZsvHw4rV/+9//7siIyPVokULLVq0SJUqVbKcy8rK0rBhw7R9+3Y1a9ZMERERMhpvvaufyWTSX/7yFx05ckRdunTRV199ladfSkqKHn74YZ04cUJdu3bV7Nmz8/S/dOmSgoKCJOVuXde+ffti3WNJsntPQ+DPbMCAARo8eLAlS/9HGzZskKQiVXECAAAAAAAAAICy4dixY5bt4EaNGpUnYShJLi4uevHFFy1tbd3fMCoqSkeOHJEkPf/88/kSjR4eHho9erQkadOmTTp79mye8/v27ZMkOTs7/+lyFSQNUa74+PgoJiZG7777rs6fP285fuPGDS1YsEALFiyQwWCwWu4LAAAAAAAAAAD+wGAs/UcxbNy4UZLk5OSk7t27F9imXbt2qlmzpiRpzZo1No1rLk7y8vJSu3btCmzTo0cPOTk5yWQyae3atXnOmfczbN68+Z9uK7Ui72kI/BmNHz9eO3fu1LZt2xQcHKxGjRqpcuXKSkhI0JUrV2Q0GvXyyy8rICDA0aECAAAAAAAAAIDb5ODBg5Kkhg0bqmrVqgW2MRgMuvPOO7Vp06ZCVzD8o9jY3G1XWrVqJYPBUGAbDw8PNWzYUHFxcZbKQjNz0rB169ZaunSpFi9erAMHDigjI0P16tVTcHCw/v73v8vLy8umeEoSSUOUK02aNNHy5csVHh6u1atXKyEhQenp6apdu7Z69uypxx9/vNDMPwAAAAAAAAAAKJjBWHCC7M8qISFBklS/fn2r7erWrStJio+PL/Fx4+Li8o1rTk7+/PPPWrx4cZ5zcXFx+vrrr7V48WJ99tln6tChg00xlRSShih3atSooWeffVbPPvuso0MBAAAAAAAAAABF1KtXL6vnrS0pmpSUJEmqVq2a1THMVYhXr161KabijJucnGxJOmZlZemBBx7QkCFD1LBhQ126dEk///yzZs+erStXrmjkyJFatGiRGjVqZFNcJYE9DQEAAAAAAAAAAFCuZGRkSNIt9w2sVKlSnva3c9wzZ86oXr16MhqNmjhxoqZMmaK2bduqRo0aat68uV544QV9+OGHknKTje+//75NMZUUKg0BAAAAAAAAAABgnaH069CsVRLeipOTUwlGknfcnJycIvVt1aqV1q1bp8zMTLm6uhbYpm/fvrr33nu1bt06rV69WsnJyapevXpxQrYZlYYAAAAAAAAAAAAoV6pUqSLp1hWEtlYOluS4hSUMzczLsubk5Gjfvn02xVUSqDQEAAAAAAAAAACAVQajwdEh2MW8p+C1a9estjPvOVijRg2bx7169WqJj3uzunXrWn6+fPmy3f2LikpDAAAAAAAAAAAAlCtNmjSRJJ09e9Zqu8TEREl5E3W3e1yTyWS1b1ZWluVnc2VjaSBpCAAAAAAAAAAAAKsMBmOpP4qjRYsWkqSTJ08qLS2twDYmk0kHDx6UJLVu3dqucc39CnLt2jWdPn0637j//ve/1alTJ3Xu3NnqvohHjx61/Ozj42NTXCWBpCEAAAAAAAAAAADKlR49ekjKrdrbuHFjgW1iYmIsy39269bNrnHPnz9f6H6DGzZsUHZ2tiSpa9euluOenp5KSkpSUlKSYmJiCuxrMpm0bNkySVL9+vXVrFkzm+IqCSQNAQAAAAAAAAAAYJ3RUPqPYmjUqJH8/f0lSTNmzFBqamqe81lZWZo6daokqWXLlurcubNN4wYEBFiWHH3vvfd048aNPOdTUlL0ySefSJJ69uyZp1Kwf//+MhpzU3OTJ0+2JBZv9uWXX1qqGIcOHSqDofT2kjSYbrVwKgAAAFBMDRo0UEJCgurXr6/4+HhHhwMAQJnA5ycAAPgzObZkYalfs9mDDxer/549ezR48GCZTCb5+fnp5Zdflq+vr06ePKlp06Zpy5YtMhgM+uyzzxQcHGzpt3fvXr388suSpCeeeEJPPPFEnnFXrFihsWPHSpK6dOmiMWPGqFGjRoqNjdWUKVN08OBBubq6Kjw8XG3atMnT99///rfmzp0rSfL399eYMWPUsmVLXbx4Ud99950WLFggSerYsaPmzp0rJyenYs2BPZxL7UoAAAAAAAAAAABAKWnXrp3eeustvfnmm9q/f7+GDh2ar82ECRPyJAwlKT09XSdOnJAkJSUl5evTr18/HTp0SDNnztSmTZu0adOmPOednZ31/vvv50sYStIrr7yiK1eu6KefftLu3bv15JNP5msTEBCgmTNnlmrCUCJpCBRLVvIVR4dQZrhU95QkZaWlWm8ICxc3d0lSegpzZo8qHrnzllXI5sbIz8XNTRK/n/Yw/37yOisCk0lZqSmOjqLMcHH3kCTFLfvZwZGUHT79/yKJ9zR78J5mPz47i+b31xrzZjeTiXmzEe9p9vv9PY05swfzZj8+P+3HZ2fRmOcNJc9gKJs73g0ePFht2rTR119/rW3btikpKUnu7u7y9/fX0KFDbV6W9I/Gjh2rwMBAzZ07V3v27FFycrJq1Kihe+65R8OHD5efn1+B/cwJxf79+2vBggXau3evkpOTVa1aNd15550aOHCg7r///lJdltQSW6lfEQAAAAAAAAAAACglfn5+lv0LbREYGKhDhw7dsl3nzp2LnHTs2bOnevbsWaS+twtJQwAAAAAAAAAAAFhlMJZ+5RtKV9msJQUAAAAAAAAAAABQYqg0BAAAAAAAAAAAgHVldE9D2I5nGAAAAAAAAAAAAKjgqDQEAAAAAAAAAACAVQYjdWjlHc8wAAAAAAAAAAAAUMGRNAQAAAAAAAAAAAAqOJYnBQAAAAAAAAAAgHUGg6MjwG1GpSEAAAAAAAAAAABQwZE0vI3i4+Pl6+srX19fvfLKK7dsv3XrVkv70mIymXTs2LE8x26O++TJk6UWS2k6cuSIo0MAAAAAAAAAAKDMMBiNpf5A6WLGS8nSpUu1bt06R4eRx969ezVo0CDNnDnT0aGUmvPnz2v8+PEaPny4o0MBAAAAAAAAAAD40yBpWIpef/11JScnOzoMi/nz52vv3r2ODqNURUVFKSIiQiaTydGhAAAAAAAAAABQZhiMhlJ/oHSRNCwlBoNBFy5c0KRJkxwdCgAAAAAAAAAAAJAHScNSMmTIEEnSTz/9pDVr1jg4GgAAAAAAAAAAADsYjKX/QKlydnQAFUVYWJhiY2O1Y8cOvfHGG+rQoYM8PT1t7n/u3DnNmTNHGzZsUEJCgoxGoxo1aqQ+ffrob3/7m6pVq5an/ccff6xPPvlEf//733X33Xfr/fffV0JCgry9vTV+/HiNGzfO0vbnn3/Wzz//rICAAM2bNy/POCaTSUuWLNGCBQt06NAhGQwGNWnSRA8//LAeffRRGQz5y4MzMzMVHh6u5cuX6+jRo8rKylLdunXVs2dPPf300/L29i7wHg8ePKj58+drx44dOnfunDIzM1WtWjX5+fnpr3/9q0JDQwucly+//FIbN25UQkKCXFxcVLduXQUFBenJJ59UgwYNLG19fX3z9DP/96FDh2x4BgAAAAAAAAAAAMovkoalxGAw6D//+Y/uv/9+XbhwQe+8846mTp1qU9/o6GiNHj1a165dk4uLi5o3b64bN27o8OHDio2N1cKFC/XFF1/kSYqZbd++XV9//bWqV6+uZs2a6dixY2rQoIHat2+vkydP6tKlS/Ly8pKPj49atmyZr/9rr72m7du3q3r16mrSpIkSEhK0b98+7du3T3v37tXkyZPztD9//rxGjBihgwcPymAwqF69evL09NTRo0c1Z84cLV26VJ999pk6dOiQp9/8+fP1zjvvKCcnR9WrV1fjxo11/fp1xcfHKzIyUpGRkRo5cqReeOEFS59Tp07p0Ucf1aVLl+Tm5qYmTZpIkuLi4jRv3jwtWbJE8+bNU+vWrSVJ7du31+XLlxUXFycXFxfdddddNs0/AAAAAAAAAABAeUdtZylq3LixpcIvIiJCq1evvmWfhIQEPfvss7p27ZqCg4O1fv16LV26VBEREVq1apX8/f119uxZjRw5UteuXcvXPyYmRsHBwdqwYYN+/PFHRUZGql27dgoPD1f37t0lSV26dFF4eLhef/31fP137typCRMmaNOmTVq8eLGioqL0+OOPS5IWL16sY8eOWdqaTCY9//zzOnjwoDp06KDly5dr7dq1Wrx4sTZt2qS//vWvunLlip577jlduHDB0i8uLk7/+c9/lJOTo7Fjx2rTpk1asmSJVqxYoY0bN6pfv36SpNmzZys5OdnSb9q0abp06ZJCQkIUFRVlqZhct26d/P39lZKSkicxGx4ermeeeUaS5OXlpfDwcIWHh9/yOQAAAAAAAAAAoKIzGA2l/kDpImlYysLCwnTPPfdIkt544w0lJSVZbf/FF18oLS1NLVu21EcffaRatWpZzjVs2FBffPGFateurTNnzuRbWtTslVdekaurq6TcZJk9Hn74YQ0bNkwuLi6SJBcXF73yyivy8PCQJO3YscPSds2aNdq9e7e8vb311VdfqWnTppZzVatW1b///W+1a9dOSUlJmjNnjuXcpk2b5OTkJD8/P40aNcpyLUny9PTUK6+8IknKysrSiRMnLOdiY2MlSQMHDpS7u7vleK1atfTPf/5T3bp1U/Pmze26XwAAAAAAAAAAgIqIpGEpMy9TWqVKFV28eFHvvPOO1fbr16+XJD322GOWxN/Nqlevrr/+9a+SVGDlYu3atdWwYcMix9u3b998xypXrqzGjRtLki5fvmw5br5+79695ebmlq+fwWDQwIEDJUnr1q2zHB8yZIj27Nmj+fPnFxhD5cqVLT+np6dbfjbH8MEHH2j16tW6fv265dxdd92lr776ShMnTrz1TQIAAAAAAAAAAKsMBmOpP1C62NPQARo1aqTx48dr0qRJWrZsmUJDQwtMzqWkpOjcuXOSpDZt2hQ6np+fnyTlqcIz8/b2LlasderUKfC4ubLv5kTd4cOHJeUmBM1VgH909epVSblLkppMJhkMv5cXu7i4aO/evTp8+LBOnz6tU6dO6fDhwzp+/Liljclksvw8ZswYbd26VSdOnNBzzz0nV1dX+fv7q0uXLurRo4fuvPPOIt41AAAAAAAAAABAxULS0EGeeOIJrVy5Utu3b9ebb76pjh075muTmppq+dm8HGhBzOfS0tLyJeIqVapUrDjt6Z+SkiJJOnv2rM6ePWu1bXZ2tlJTUy2xL1myRFOnTs2z16EkNWjQQA8//LC+//77fGO0atVKP/30k7744gv9+uuvunLlirZu3aqtW7fqww8/VMuWLfXGG28UOLcAAAAAAAAAAMAORir/yjuShg5iXqZ04MCBunTpkt5++2099thjedrcvE+fOSFXkOTkZEmSm5tbnoRhaatSpYok6fXXX9cTTzxhc78lS5ZowoQJkqRu3bqpT58+atGihZo1a6bq1asrKyurwKShlLuv46RJk/T2229r37592rZtm6Kjo7V161YdPnxYw4cP14oVK1S3bt3i3yAAAAAAAAAAAEA5RVrYgczLlErSihUrtHLlyjznPTw8VLt2bUnSvn37Ch3HfM7Hx+f2BGqjJk2aSJKOHDlSaJuzZ88qJibGsuyqJH3xxReSpAceeEBfffWVBg8erPbt26t69eqSpMTExHzjmEwmxcfHa/PmzZIko9Gotm3bavjw4Zo9e7Z+/vlneXh4KD09XatWrSqxewQAAAAAAAAAoCIyGAyl/kDpImnoYE888YQCAgIkSfPnz893Pjg4WJIUHh6uzMzMfOeTk5O1dOlSSVL37t3turb5F+7mfQKL495775UkLV++XJcuXSqwzauvvqrBgwdbkqWSFB8fL+n3vRn/aOHChZafb9y4IUm6cuWKQkJCNGzYMP3222/5+jRp0kT16tWTJOXk5FiOG/9XPl1S9wwAAAAAAAAAAFAekDR0MPMypW5ubgUmsv7+97/L3d1dhw8f1pgxY/Ik406fPq1nnnlGFy9eVJ06dTR06FC7rm1e/vTMmTPFu4n/ue+++9SyZUtdvXpVTz/9dJ6Kw5SUFL355pvavHmzDAaDRowYYTnXtGlTSdKCBQvyVCCmpKTo448/1qxZsyzHrl+/LkmqUaOGunXrJik3EXns2DFLm5ycHH333Xc6fPiwjEajpZ2Uu4SrlJtstbbkKwAAAAAAAAAAQEXCnoZ/Ag0bNtT48eP1zjvvFHhuxowZGjNmjNauXasePXqoefPmys7O1tGjR5WTk6N69erpk08+kZeXl13XbdWqlSRp165dCg0NVfPmzfXJJ58U+T5cXFz02Wefafjw4Tp48KAGDBigJk2aqEqVKoqLi1NaWpokaeLEiXmqIl944QU9++yzOnr0qHr16mVZ5vTkyZPKyMhQw4YNZTAYdOrUqTxLlb799tsaPHiwDh8+rAEDBqhBgwaqWrWqzpw5o6SkJMvYzZs3t/Tx9fWV0WhURkaGQkND5e3trdmzZ6tGjRpFvm8AAAAAAAAAAMo7g5E6tPKOZ/hPYsiQIZZlSv+oa9euWrZsmZ588kk1aNBAJ06c0NmzZ9WqVSuNHz9eP/74Y6FLe1rzwAMPaPjw4apdu7bi4+N18ODBPEt5FkXDhg21ZMkSvfzyy2rXrp0uXLigw4cPy93dXSEhIfr222/zVUTee++9WrhwoXr37q3atWvr+PHjOnv2rFq2bGm5v7/85S+SpHXr1ln6eXt7a+HChXr66afVvHlzy7UqVaqk/v37Kzw8PE9FoyQ1btxYkydPlo+Pj65cuaKzZ88qISGhWPcMAAAAAAAAAABQ1hlMbO4GFFlW8hVHh1BmuFT3lCRlpaU6NpAyxMUtdwnh9BTmzB5VPHLnLet/1c24NZf/Ld3M76ftzL+fvM5s16RlSyUkJKh+vXo6cfiQo8MpM1zcPSRJcct+dnAkZYdP/9x/bMZ7mu14T7Mfn51F8/trjXmzVZOWvr9/fh457OhwygTe0+z3+3sac2YP5s1+fH7aj8/OojHPG0remajIUr9mva7db90IJYZKQwAAAAAAAAAAAKCCY09DAAAAAAAAAAAAWMWehuUfzzAAAAAAAAAAAABQwVFpCAAAAAAAAAAAAKuoNCz/eIYBAAAAAAAAAACACo6kIQAAAAAAAAAAAFDBsTwpAAAAAAAAAAAArDMYHB0BbjMqDQEAAAAAAAAAAIAKjkpDAAAAAAAAAAAAWGUwUodW3vEMAwAAAAAAAAAAABUclYYAAAAAAAAAAACwysCehuUelYYAAAAAAAAAAABABUelIVAMLtU9HR1CmePi5u7oEMqcKh7MWVG4uLk5OoQyh99P+/E6KwKDQS7uHo6Ooszx6f8XR4dQ5vCeZj/e0+zH66xomLciMBiYNzvxnmY/5qxomDf78X5mP+YMfxbsaVj+8QwDAAAAAAAAAAAAFRyVhkAxZKWlOTqEMsP8L+8yUpkzW1Vyz52zuFPnHBxJ2eLTqI4kKSst1cGRlB3mf7HInNnu9znjPc1eJpNJ15KvOTqMMqNq9aqSeK3Zw/ydIz2F9zRbmVc14HVmO/PrjDmzj2XeUlMcHEkZZDLxerPR77+ffA7YyvzdNu0ac2YPt6r8HWUv/o6yH5+dRcPqNkDRkTQEAAAAAAAAAACAdSxPWu7xDAMAAAAAAAAAAAAVHJWGAAAAAAAAAAAAsMpgMDg6BNxmVBoCAAAAAAAAAAAAFRyVhgAAAAAAAAAAALDKwJ6G5R7PMAAAAAAAAAAAAFDBUWkIAAAAAAAAAAAA6wzUoZV3PMMAAAAAAAAAAABABUfSEAAAAAAAAAAAAKjgWJ4UAAAAAAAAAAAAVhmMBkeHgNuMpOFtEBsbq4ULFyo6Olrnzp1TRkaGvLy81KJFC/Xs2VMPP/ywKleu7OgwJUlHjhxRixYt8hwLDg5WQkKCJk2apEceecRBkd0+iYmJ8vDwkIeHh6NDAQAAAAAAAAAA+FNgedISNmPGDD344IOaN2+eEhMT1bBhQ7Vq1UpGo1EbN27UO++8o9DQUO3fv9+hcZ4/f17jx4/X8OHDHRpHacrMzNRHH32kkJAQXbp0ydHhAAAAAAAAAABQZhiMxlJ/oHRRaViCFi1apE8//VRubm6aPHmy+vTpIycnJ8v5Y8eO6dVXX1VMTIyefvppLV++XF5eXg6JNSoqShEREapTp45Dru8I58+f12effeboMAAAAAAAAAAAAP50SNOWoM8//1yS9PLLLys0NDRPwlCSmjVrppkzZ6pmzZpKSkrS3LlzHREmAAAAAAAAAACAXQwGQ6k/ULpIGpaQ5ORknTp1SpLUrl27Qtt5eXmpd+/ekqS9e/eWSmwAAAAAAAAAAACANSxPWkKcnX+fynXr1ql169aFth09erT+9re/qWbNmvnO/fbbb5o7d662b9+uixcvys3NTb6+vrr//vv14IMP5qteDAsL07Zt2zRy5Ei98MIL+cb7+OOP9cknnyggIEDz5s2TJPn6+lrOnzt3zvLfhw4dytd/586d+vLLL7V7926lp6erbt266tu3r5555hl5eHgUeH+rV6/W999/r99++03Xrl1TjRo1FBAQoKeeekp+fn4F9jl//ry+++47bdq0SadOnVJqaqrc3d3VtGlT9e3bV48//rgqV66cp8/169c1d+5c/fLLLzpx4oRu3LihWrVqqX379nr88cfVoUOHfPNk1rdvX0nS3LlzFRgYWGBMAAAAAAAAAADgf9hjsNwjaVhC3N3d1b59e+3atUsff/yxTp8+rb/+9a9q3759vkRf7dq1Vbt27XxjfPnll/rwww+Vk5MjDw8P+fr6KikpSdu2bdO2bdv0448/6rPPPlPVqlWLFWv79u11+fJlxcXFycXFRXfddVeB7ZYsWaJdu3bJ1dVVTZs21ZUrVxQXF6dZs2Zp3bp1WrhwYZ5E3o0bNzRhwgT9/PPPkqSaNWvK19dX8fHxioiI0IoVK/Tqq6/qiSeeyHOdmJgY/f3vf9fVq1dVqVIlNWrUSM7OzoqPj9fu3bu1e/durVmzRnPnzrXMZWZmpp588knt3r1bTk5Oaty4sapUqaLTp08rIiJCy5Yt0zvvvKNHHnlEktSyZUulpaVp3759kiQ/Pz9VqlSp2HMJAAAAAAAAAABQHpAWLkGvv/663NzcZDKZtGTJEj3xxBMKCAjQiBEjNGvWLO3Zs0c5OTkF9l25cqU++OAD5eTk6Nlnn1V0dLQWLVqktWvX6ptvvlGtWrW0bds2vfzyy8WOMzw8XM8884yk3OVSw8PDFR4enq/dzp07NWDAAEVGRmrp0qVav3693n//fUnSkSNHtHjx4jztP/roI/3888+644479NVXX2nz5s1atGiRNm/erNdee00Gg0GTJk3Spk2bLH2ys7P10ksv6erVq+rdu7c2btyoiIgILV26VNHR0Ro/frwkaceOHdq4caOl36JFi7R79275+Pho9erVWrFihRYvXqyoqCgNGTJEJpNJU6ZMUUZGhqTc5+ajjz6y9J82bZrCw8OtVoQCAAAAAAAAAABUFCQNS1Dr1q31ww8/5FkWMyUlRRs2bNDUqVM1aNAgde3aVdOmTVN6enqevtOmTZMkDR48WGPGjJGrq6vlXKdOnfTJJ59IktauXasdO3aUwt1IzZo105QpU+Tp6Wk5NnDgQHXp0kVSblLR7OLFi5ozZ44k6bPPPlO3bt0s55ycnBQWFqYnn3xSJpNJ06dPt5yLjY3VlStX5OrqqkmTJql69eqWcy4uLhoxYoQaNmwoSTp8+HCefpLUvXt31atXz3K8UqVKmjBhgrp27ao+ffroypUrxZ4HAAAAAAAAAAAqOoPBWOoPlC5mvIQ1b95c8+fP19KlS/WPf/xD/v7+cnFxsZy/dOmSPv/8cw0cOFCJiYmSpLi4OJ04cUKSNHTo0ALH9ff3l7+/vyRpzZo1t/kucvXq1Svf0qrS73siXr582XIsMjJSmZmZat68eaH7Ft5///2SpL179+rSpUuScpcJ3b59u7Zv364aNWrk65OZmWlJJN6caPXx8ZEkLVy4UPPnz88Ti6urq2bPnq3JkyerTp069twyAAAAAAAAAABAhcSehrdJq1at1KpVK40ePVrp6enatWuXoqKi9OOPP+rSpUs6deqUxowZowULFuj48eOSpCpVqqhZs2aFjtmmTRvt3r3bkmC83QpLuLm5uUmSrl+/bjl25MgRSVJiYqIee+yxAvuZTCbLz8ePH1fNmjUt/125cmUdP35cBw4c0KlTp3T69GkdPXpUhw4dsiwxevPSro888ogWLlyoo0eP6q233tLbb7+tVq1aqXPnzurWrZvuueceOTvz8gYAAAAAAAAAoCQYjAZHh4DbjKxKKahSpYq6dOmiLl26aMyYMXr11Ve1bNkyxcTEaP/+/UpJSZEkeXh4WB3H3d1dkpSamnrbY5Zyl/q01bVr1yTlLse6a9euW7a/evWq5ec9e/bozTff1IEDB/K0qVGjhnr06KEDBw4oPj4+zzkPDw8tWLBAX3/9tSIiInTy5EkdOHBABw4c0OzZs1WzZk2NHTtWgwYNsvkeAAAAAAAAAAAAKiqShiXkX//6l7Zs2aIHH3xQo0aNKrRd5cqV9fbbb2vVqlXKysrSiRMnLMlAc/KwMOZEm7m9LdLS0mxuWxxVqlSRJIWEhGjGjBk29zt27Jj+9re/6fr162revLn++te/6s4771SzZs0slY6PPvpovqShlJs4fP755/X888/r5MmT2rp1q7Zu3arIyEhdunRJr7/+ujw9PdW3b9+SuUkAAAAAAAAAACoog5Ed78o7nuESkpGRoZMnT2r16tW3bOvh4WFJ/Hl5ealp06aScvfsO3bsWKH99u3bJ0lq3Lix5Zh5z8HMzMwC+5w/f962GyimJk2aSPp9mdKCpKena9u2bTp9+rSys7MlSd98842uX7+upk2bauHChXrqqacUFBSUZ2nUc+fO5Rvr0qVL2rFjh2Uvw8aNG2vQoEGaOnWq1q9frzZt2kiSfvzxxxK7RwAAAAAAAAAAgPKKpGEJGThwoKTcxN7ixYutto2KitKVK1fk6empdu3aqUmTJpak2zfffFNgn127dmnv3r2SpO7du1uO16hRQ5Is+yLeLDU1VdHR0QWOZ/zfvwi4eZ/B4ujRo4ecnJx0/Phxbdq0qcA2c+bMUVhYmO6//36lp6dLkhISEiRJzZo1s1Qr3mzTpk06c+aMJFkSjZL09NNPa8iQIVqyZEm+Pu7u7rr77rvz9THe9K8gSuq+AQAAAAAAAACoEAzG0n+gVDHjJaRLly4KCQmRJL322mv697//nW9JzYyMDC1atEhjx46VJI0dO9ZScThmzBhJ0oIFCzRjxow8lYNbt27V888/L0nq1q2bgoKCLOc6dOggSdq4caNWrVplOX7+/Hk9//zzunTpUoHxurm5SZKSk5NvuSyqLerXr69HHnlEkjRu3DitXbvWci4nJ0c//PCDPvnkE0nSkCFDLPs3mpOlmzZt0o4dOyx9bty4oYiICL3wwguWY9evX7f8fP/990uSPvnkE0VGRuaJZceOHZYKwx49eliOm+9ZkiURCQAAAAAAAAAAAPY0LFEffPCB3NzctHTpUs2dO1dz585VvXr1VLNmTWVkZCguLk6ZmZlycXHR+PHj9dhjj1n69uvXT6dOndK0adP06aef6ptvvlGTJk10+fJlSzVeQECA3n//fRkMBku/Bx98UN9++61OnDih0aNHq1GjRnJzc9OxY8fk5OSkkSNH6vPPP88Xq6+vr4xGozIyMhQaGipvb2/Nnj3bUrlYFK+++qrOnTundevWadSoUfL29ladOnWUkJBgWUY0JCTEkjSVpKeeekoRERFKSkrSkCFD5OPjI3d3d8XHxys5OVlubm7y9/fX7t27lZiYaOn3t7/9TZs3b1ZkZKT+/ve/y9vbW97e3kpKSrLMV3BwsCWRKUmenp6qX7++EhIS9Nxzz6lp06YaM2ZMnspNAAAAAAAAAACAiohKwxLk6uqqd999Vz/88IOeeuop+fn5KTMzU7GxsUpMTFSTJk309NNP66efftKIESPy9X/mmWf0/fffa8CAAfLw8FBsbKyuX7+uzp07a8qUKfrmm2/yJfXc3d21YMECDR8+XI0bN9bZs2d18eJFhYSEaMmSJQoMDCww1saNG2vy5Mny8fHRlStXdPbsWUuyragqVaqkmTNnatq0aerWrZuysrJ08OBBZWdnKzAwUFOmTNH06dMt+zBKUr169fTTTz/psccek4+Pj86ePasTJ06oVq1aCgsL008//WRJMm7dulVpaWmScvdy/PTTT/Xqq6/K399f169fV2xsrNLT09W1a1e9//77+uyzz+TsnDcv/tFHH8nf3185OTmKi4vTqVOninXPAAAAAAAAAABUBAajodQfKF0GE5u7AUWW9b8kJm7N5X/Lw2akMme2quSeO2dxp845OJKyxadRHUlSVlqqgyMpO1zccpfKZs5s9/uc8Z5mqyYtWyohIUH16tVT7IFYR4dTZlStXlUSrzV7mL9zpKfwnmarKh68p9nL/DpjzuxjmbfU4m+RUVE08b1TCQkJql+vnk4cOeLocMqE338/+Rywlfm7bdo15sweblX5O8pe/B1lPz47i8bF3cPRIZRbKadPlvo1PRo2LvVrVmQsTwoAAAAAAAAAAACrDAYWryzveIYBAAAAAAAAAACACo5KQwAAAAAAAAAAAFhlMFKHVt7xDAMAAAAAAAAAAAAVHJWGAAAAAAAAAAAAsM5ocHQEuM2oNAQAAAAAAAAAAAAqOJKGAAAAAAAAAAAAQAXH8qQAAAAAAAAAAACwymCgDq284xkGAAAAAAAAAAAAKjgqDQEAAAAAAAAAAGCVwWhwdAi4zag0BAAAAAAAAAAAACo4Kg0BAAAAAAAAAABgHXsalns8wwAAAAAAAAAAAEAFR6UhUAwubm6ODqHMqeTOnNnLp1EdR4dQJrm4uTs6hDKHObMfnwP2MxgMqlq9qqPDKHN4rdmvigfvafbidWY/5qxoXNw9HB1C2WMw8HqzE99t7edWlTkrCl5r9uP9zH58duLPgj0Nyz8qDQEAAAAAAAAAAIAKjkpDoBiy0tIcHUKZYf5XZFlpqQ6OpOww/2tF5sw+zJv9fp8z3tNsxXtaMZhMzJsdeE+zH3NmPz4H7MfnQNFYXmvJVxwbSFlkMvE7aqPffz+ZL1sxZ0XDvNmPz0/78T2taKhmBYqOpCEAAAAAAAAAAACsMhhYvLK84xkGAAAAAAAAAAAAKjgqDQEAAAAAAAAAAGCdweDoCHCbUWkIAAAAAAAAAAAAVHBUGgIAAAAAAAAAAMAqA5WG5R6VhgAAAAAAAAAAAEAFR6UhAAAAAAAAAAAArDNSh1be8QwDAAAAAAAAAAAAFRxJQwAAAAAAAAAAAKCCY3nSEhAbG6uFCxcqOjpa586dU0ZGhry8vNSiRQv17NlTDz/8sCpXruzoMCVJR44cUYsWLfIcCw4OVkJCgiZNmqRHHnnEQZGVnitXrigrK0u1a9d2dCgAAAAAAAAAAJQNBoOjI8BtRqVhMc2YMUMPPvig5s2bp8TERDVs2FCtWrWS0WjUxo0b9c477yg0NFT79+93aJznz5/X+PHjNXz4cIfG4Whz5sxR3759deTIEUeHAgAAAAAAAAAA8KdBpWExLFq0SJ9++qnc3Nw0efJk9enTR05OTpbzx44d06uvvqqYmBg9/fTTWr58uby8vBwSa1RUlCIiIlSnTh2HXP/PYvLkyY4OAQAAAAAAAACAMqjsVhru379fs2fP1rZt23TlyhXVqFFD7du319ChQ9W+ffsij7tlyxbNmTNHMTExSklJUa1atdS5c2cNGzZMLVu2tNr35MmTmjVrljZv3qwLFy6oWrVq8vPz0xNPPKEePXoUOabioNKwGD7//HNJ0ssvv6zQ0NA8CUNJatasmWbOnKmaNWsqKSlJc+fOdUSYAAAAAAAAAAAAFdIvv/yiQYMGadmyZbpw4YKysrJ0/vx5/fLLLxoyZIi++uqrIo07Z84cDR06VOvWrVNSUpKysrJ09uxZLV68WH/961/1888/F9p3586duv/++7Vw4UKdOXNGWVlZunTpkiIjIzVixAhNmjSpqLdbLCQNiyg5OVmnTp2SJLVr167Qdl5eXurdu7ckae/evaUSGwAAAAAAAAAAQIkyOOBRTPv27dNLL72kGzduqGPHjpo/f76io6MVHh6uwMBA5eTk6IMPPtD69evtGnft2rV69913JUl9+vTR4sWLFR0drf/7v//TnXfeqczMTE2cOFEHDhzI1/fMmTN69tlnlZ6erhYtWmj27NmKjo7WkiVLFBoaKkmaN2+evvvuu2Lfv71IGhaRs/PvK7uuW7fOatvRo0dr2bJlmjp1ar5zv/32m1566SX17NlTbdq0UUBAgMLCwrRw4UJlZ2fnax8WFiZfX19NmzatwGt9/PHH8vX1VVhYmOWYr6+vJk6cKEk6d+6cfH195evrW2D/nTt3auTIkQoMDFTbtm0VEhKiqVOnKiUlxdJm06ZN8vX1lb+/vzIzM/ONMWzYMPn6+uqhhx7Kdy4nJ0edOnWSr6+vfvvttzznVq9erREjRqhz585q06aNunXrpvHjxxe6H2R2drbmz5+vxx57TB06dFCbNm3UtWtXPffcc/mekwkTJuS5Z3OMixcvLnBsAAAAAAAAAABQtn300UfKzMxUixYt9PXXX6tDhw7y8vJS+/btNXv2bN1zzz0ymUx67733lJOTY9OYJpNJH374oUwmk7p06aIZM2bIz89PXl5eCgoK0nfffacmTZooKyurwLzQrFmzdOXKFdWqVUtz585V165d5eXlpdatW+ujjz7SgAEDJOXme1JTU0t0Pm6FpGERubu7W9a5/fjjjzVhwgRt3769wERf7dq11bx5c9WoUSPP8S+//FKDBg3STz/9pGvXrsnX11ceHh7atm2b/vnPf+rJJ5/UtWvXih1r+/bt5ePjI0lycXFR+/btC1yjd8mSJRoyZIg2b96sunXrysvLS3FxcZo1a5YeffRRXb9+XZIUEBAgDw8PpaWladeuXXnGyMjI0M6dOyVJBw8e1NWrV/Ocj4mJUVJSku644w7dddddkqQbN27oxRdf1HPPPacNGzbIYDDI19dXmZmZioiI0COPPKJvv/02zzgmk0kvvPCC3nrrLe3atUs1a9aUr6+vsrOztXr1ao0cOVIfffSRpb2Pj0+ee27ZsqXat2+vmjVrFnFWAQAAAAAAAACoQAyG0n8Uw7FjxxQZGSlJGjVqlCpVqpTnvIuLi1588UVL2z/mOwoTFRWlI0eOSJKef/55GY15U20eHh4aPXq0pNwirLNnz1rOXb161VLMNHToUHl5eeUb/+WXX5aTk5OSkpK0evVqm2IqKSQNi+H111+Xm5ubTCaTlixZoieeeEIBAQEaMWKEZs2apT179hSamV65cqU++OAD5eTk6Nlnn1V0dLQWLVqktWvX6ptvvlGtWrW0bds2vfzyy8WOMzw8XM8884yk3OVSw8PDFR4enq/dzp07NWDAAEVGRmrp0qVav3693n//fUnSkSNHLC9kFxcXdevWTVLuC/5m27dvV0ZGhqTcqsJt27blOb9hwwZJUq9evSzHPvroI/3888+644479NVXX2nz5s1atGiRNm/erNdee00Gg0GTJk3Kc62NGzdq5cqV8vLy0k8//aRVq1Zp0aJFioqK0rhx4yRJX3zxhRITEyVJI0eOzHPPEydOVHh4uMM2EwUAAAAAAAAAALfPxo0bJUlOTk7q3r17gW3atWtnKS5as2aNTeOa8xxeXl6Fbl/Xo0cPOTk5yWQyae3atZbjW7duteRQ7r333gL71qlTR61atbIrppJC0rAYWrdurR9++EEdOnSwHEtJSdGGDRs0depUDRo0SF27dtW0adOUnp6ep695edHBgwdrzJgxcnV1tZzr1KmTPvnkE0m56+Lu2LGjFO5GatasmaZMmSJPT0/LsYEDB6pLly6SZKkglKTg4GBJuRn1m5kTe/fcc48kacuWLXnOm5cNNe/zePHiRc2ZM0eS9Nlnn1mSkVLuL3JYWJiefPJJmUwmTZ8+3XIuNjZWkuTv759n2VEnJyc988wzCg0N1YABA5ScnGzfJAAAAAAAAAAAgDLv4MGDkqSGDRuqatWqBbYxGAy68847JanQrdL+yJyfaNWqlQyFVEN6eHioYcOGknL3VfxjTK6urmrevHmh12jdurVdMZUUkobF1Lx5c82fP19Lly7VP/7xD/n7+8vFxcVy/tKlS/r88881cOBAS9VbXFycTpw4ISm3/LQg/v7+8vf3l1R6meRevXrJyckp33FzUu7y5cuWYz169JCzs7MOHjyY5/jmzZtlMBg0atQoSblZc7OzZ8/q0KFDqlatmiWpGBkZqczMTDVv3lx+fn4FxnX//fdLkvbu3atLly5JkmW51Q0bNuiLL77IU94r5VYvvvfee4Xu3QgAAAAAAAAAAMqvhIQESVL9+vWttqtbt64kKT4+/raPa+5br169QhOON/c9e/aszXstlgTnUrtSOdeqVSu1atVKo0ePVnp6unbt2qWoqCj9+OOPunTpkk6dOqUxY8ZowYIFOn78uCSpSpUqatasWaFjtmnTRrt377YkGG+3OnXqFHjczc1Nkix7GkpS9erV1b59e23btk1RUVEaOHCgLl68qEOHDunOO+9Up06dVLVqVR05ckSXLl1SzZo1tX79eklS9+7dLYlV87q/iYmJeuyxxwq8vslksvx8/Phx1axZU8HBwQoICNC2bdv04Ycf6sMPP1TTpk0VFBSkbt26qXPnzvnWJwYAAAAAAAAAAGXHzVudFcRa0VVSUpIkqVq1albHMFchXr161aaYijOuvX2zs7OVkpJyy/YlhUrD26BKlSrq0qWLXnnlFa1du1b9+/eXJMXExGj//v1KSUmRlFueao27u7skKTU19fYG/D/2JtnMv6zmJUk3b94sk8mkzp07y8nJSffcc49MJpOl2tCcNDQvTSpJ165dk5S7rOuuXbsKfOzevdvS3vzL5ezsrNmzZ2vChAlq2bKlpNyE4rfffqtnnnlGQUFBmjlzZp6EIwAAAAAAAAAAKCqDAx5FZ947sHLlylbbmXMj5va3c1xb+9583ta4SgKVhkX0r3/9S1u2bNGDDz5oWYqzIJUrV9bbb7+tVatWKSsrSydOnLAkA83Jw8KYE2Tm9rZIS0uzuW1x9erVS5MnT86TNJRy92SUpKCgIK1du1ZbtmxRr169tGXLFrm6uubZt7BKlSqSpJCQEM2YMcOu67u6umrYsGEaNmyYEhMTtWXLFm3dulWRkZG6ePGipk+frsqVK2vYsGElcbsAAAAAAAAAAKAUFWf7toK2YysJTk5ORV4y9HbFVFKoNCyijIwMnTx5UqtXr75lWw8PD0viz8vLS02bNpUkpaen69ixY4X2M2+O2bhxY8sx8wsqMzOzwD7nz5+37QZKQMOGDdWiRQtduHBBsbGx2rp1q1xcXCz7FQYFBUnK3dcwOjpa169fV6dOnfJUWDZp0kTS78uUFiQ9PV3btm3T6dOnlZ2dLUlKTk5WTEyMZS/DO+64Qw888IAmT56s9evX695775Uk/fjjjyV/4wAAAAAAAAAA4E/NXLR0q0o9W6v/SmJcW/vevF2crXGVBJKGRTRw4EBJuYm9xYsXW20bFRWlK1euyNPTU+3atVOTJk0sybJvvvmmwD67du3S3r17JeXuAWhWo0YNSbLsi3iz1NRURUdHFzie0Zj7VJf0cp3BwcGSpAULFujMmTNq166dZQ/EZs2aydvbW3FxcQoPD5eUd2lSSerRo4ecnJx0/PhxS8XiH82ZM0dhYWG6//77lZ6eLkl69dVXNXjwYH355Zf52ru4uCggIECSLElGM/PGoixbCgAAAAAAAABA+WXeF9C8TVphzKs+mvMvt3Nc896EtvZ1cXG55VZ3JYmkYRF16dJFISEhkqTXXntN//73vxUfH5+nTUZGhhYtWqSxY8dKksaOHWupOBwzZoyk3GTbjBkz8lQObt26Vc8//7wkqVu3bpaKPUnq0KGDJGnjxo1atWqV5fj58+f1/PPP69KlSwXGa07kJScn33JZVHuYk4Y//PCDJKlz5855zpv/e/369TIYDJb2ZvXr19cjjzwiSRo3bpzWrl1rOZeTk6MffvhBn3zyiSRpyJAhll+O+++/X1Lu/C1dujRPEvDIkSOaN2+epNyk5M3M83DmzJmi3jIAAAAAAAAAABWOyQGP4jAXb5lXLCxMYmKiJKlu3bq3fVwfHx+7+t5xxx2WYqjSwJ6GxfDBBx/Izc1NS5cu1dy5czV37lzVq1dPNWvWVEZGhuLi4pSZmSkXFxeNHz9ejz32mKVvv379dOrUKU2bNk2ffvqpvvnmGzVp0kSXL19WQkKCJCkgIEDvv/9+nhfEgw8+qG+//VYnTpzQ6NGj1ahRI7m5uenYsWNycnLSyJEj9fnnn+eL1dfXV0ajURkZGQoNDZW3t7dmz55tc+a8MO3atVOtWrV08eJFSb/vZ2gWFBRkWSK0Xbt2ql27dr4xXn31VZ07d07r1q3TqFGj5O3trTp16ighIUGXL1+WlLvnoTn5Kkl9+/bVoEGD9P333+uVV17RlClTVLduXaWkpOjUqVMymUxq27atRo4cmedarVu31vbt2/X2228rPDxcjz/+uB5++OFizQEAAAAAAAAAAPhzadGihSTp5MmTSktLsxQV3cxkMungwYOScvMHto4bFRVl6VeQa9eu6fTp0/nGNceUnp6uuLg4SxLxj/bv329XTCWFSsNicHV11bvvvqsffvhBTz31lPz8/JSZmanY2FglJiaqSZMmevrpp/XTTz9pxIgR+fo/88wz+v777zVgwAB5eHgoNjZW169fV+fOnTVlyhR98803+ZJ67u7uWrBggYYPH67GjRvr7NmzunjxokJCQrRkyRIFBgYWGGvjxo01efJk+fj46MqVKzp79qwlOVkcBoPBsn+gm5ub2rVrl+f8zZWHvXr1KnCMSpUqaebMmZo2bZq6deumrKwsHTx4UNnZ2QoMDNSUKVM0ffr0fBuEvvXWW5o8ebICAwOVk5OjQ4cO6cqVK+rQoYP+9a9/af78+fnKdv/zn/+oS5cucnZ21okTJxQXF1fsOQAAAAAAAAAAAH8u5pUIs7KytHHjxgLbxMTEWIqXunXrZte458+f1759+wpss2HDBsv2aV27drUcDwwMtOxRePPKizdLTEy0JCRtjamkGExs7gYUWVZamqNDKDNc/vevOLLSUh0cSdnh4pa7nDFzZh/mzX6/zxnvabbiPc1+TVr6KiEhQfXr1dOJI4cdHU6ZwXua/Zgz+/E5YD8+B4rG8lpLvuLYQMqQJn5tbvr8POLocMqE338/eU+zFXNWNMyb/fj8tB/f04rGpYBqMpSMTAe8Fl2L+Xw++uij2r17t5o3b67vv//esoWclJtMHDZsmLZv366WLVvqp59+smkp0OzsbPXq1Utnz55VYGCgvv76azk7/76wZ0pKih5++GGdOHFCPXv21BdffJGn//jx4xUREaFatWppyZIl8vb2LvC8l5eX1q1bZ0kylgYqDQEAAAAAAAAAAFDuTJw4UQaDQUePHlVYWJi2bNmipKQkxcTEaPjw4dq+fbsMBoNeeOGFPAnDvXv3KjQ0VKGhofr222/zjOnk5KRXXnlFkrR161aNGDFCe/bsUVJSkqKjo/XEE0/oxIkTcnV11ejRo/PFNG7cOFWpUkUXL17UkCFDtHbtWl2+fFmxsbEaO3asIiIiJEn/+Mc/SjVhKLGnIQAAAAAAAAAAAG6hLK5b2a5dO7311lt68803tX//fg0dOjRfmwkTJig4ODjPsfT0dJ04cUKSlJSUlK9Pv379dOjQIc2cOVObNm3Spk2b8px3dnbW+++/rzZt2uTrW79+fU2fPl3PP/+8Tp06pVGjRuVr87e//U1Dhgyx615LAklDAAAAAAAAAAAAlEuDBw9WmzZt9PXXX2vbtm1KSkqSu7u7/P39NXToUHXu3LlI444dO1aBgYGaO3eu9uzZo+TkZNWoUUP33HOPhg8fLj8/v0L79uzZUxEREZo1a5Y2b96s8+fPq3LlyvLz89Pjjz+ukJCQot5usZA0BAAAAAAAAAAAgFUmlcFSw//x8/PT1KlTbW4fGBioQ4cO3bJd586di5x0bNSokSZNmlSkvrcLexoCAAAAAAAAAAAAFRyVhgAAAAAAAAAAALCqLO5pCPtQaQgAAAAAAAAAAABUcCQNAQAAAAAAAAAAgAqO5UkBAAAAAAAAAABgFcuTln9UGgIAAAAAAAAAAAAVHJWGAAAAAAAAAAAAsCqHUsNyj0pDAAAAAAAAAAAAoIKj0hAAAAAAAAAAAABWmag0LPdIGgLF4OLm5ugQyhwXN3dHh1DmMGdFw7zZj/c0+/E6KwKDgXkrAubMfsyZ/fgcsB+vs6Jxqe7p6BDKHoOB31E7MV/2Y86KhnmzH5+f9uN1BqC0kDQEAAAAAAAAAACAVRQaln8kDYFiuLBrh6NDKDNqt+8oSTodf8HBkZQdDRvUliRdT011cCRlS2X33H+xGLc8wsGRlB0+9w2QJF29cs3BkZQd1TyrSpKy0tIcHEnZYzKZdC2Z15qtqlbntWYv87/CZs5sZ56zU7+udHAkZUejPiGSpPQUvqfZo4pH7ve0rOQrjg2kDMrOztGp0+cdHUaZ0KihtyTp7NlLDo6k7Khbt6Yk/va0l/lvz6w05s1W5gpDvqfZzvw9LfNKkoMjKVtcPWs4OgSgzDI6OgAAAAAAAAAAAAAAjkWlIQAAAAAAAAAAAKzKYX3Sco9KQwAAAAAAAAAAAKCCo9IQAAAAAAAAAAAAVpmoNCz3qDQEAAAAAAAAAAAAKjgqDQEAAAAAAAAAAGBVTg6VhuUdlYYAAAAAAAAAAABABUelIQAAAAAAAAAAAKxiT8Pyj0pDAAAAAP/P3n3HVVn+fxx/A4IIiOLAgSiIHtwrd+6RMzMtzZzf0sqRZWpp37Jhy9Tc1tfSbGg5cuQ2R87U3BPFAQoqqIAKqKzz+4PfOUkMAYHDgdfz8TiP7/Hc13Xfn/vqhgu+Hz7XBQAAAAAA8jmShnikzz//XL6+vurQoUOa7V5++WX5+vrK19dXP//8c6rt/Pz8zO3Onj2b1eE+kr+/f45fEwAAAAAAAAAAIDcjaYhHatKkiSQpICBAYWFhKba5f/++/v77b/O/d+3aler5Dhw4IEkqWbKkfH19szDStIWGhmr06NEaPHhwjl0TAAAAAAAAAIC8IMFozPEXchZJQzxSgwYNZG9vL0k6cuRIim0OHDigBw8eqFKlSuZ/x8TEpNj24MGDkqQnn3wyG6JN3e7du7V27VrWXQYAAAAAAAAAAPgXkoZ4JGdnZ9WsWVOSdPjw4RTb7N69W5L0zDPPqHz58oqOjjYnB//NUklDAAAAAAAAAACQOQkJxhx/IWeRNES6NG3aVFLqSUPTcqRNmjRR8+bNk3z2sAsXLujWrVuysbEhaQgAAAAAAAAAAJBLFLB0ALAOTZo00ezZs3Xy5EnFxMTIwcHBfCw4OFgXL16Um5ubqlevrpCQEC1atEi7du3SO++8k+Q8pn0PfX19Vbx4cfPnkZGR+uGHH/THH38oMDBQRqNRnp6eat++vQYNGiRXV9dkMYWEhOjbb7/Vrl27FBwcLHt7e5UpU0ZNmzbVoEGDVK5cOXPbh/dODAkJMf/77NmzWTNAAAAAAAAAAADkYez8lfdRaYh0qV27tpycnBQTE6OTJ08mOWaqKHzyySdla2urxo0by97eXv7+/rp+/XqStqak4cNVhhcuXFC3bt00c+ZMnTt3Tu7u7qpQoYIuXryoOXPmqHv37rpw4UKS81y+fFnPPvusfvrpJ4WGhsrb21vlypXTlStX9NNPP+mZZ57R6dOnze3r1asnLy8vSZK9vb3q1aunevXqZdn4AAAAAAAAAAAAWDOShkgXe3t71a9fX1LyJUpN+xk2a9ZMkuTi4qLatWtLSr5EqWk/Q1Pb6OhoDR06VMHBwWrbtq22b9+uTZs2afXq1frzzz/VqlUrBQcHa9iwYbp//775PNOmTdOtW7fUoUMH7d69W2vWrNGaNWu0fft21a1bV5GRkZo6daq5/S+//KJXX31VklSsWDH98ssv+uWXX7JsfAAAAAAAAAAAyMsSjMYcfyFnkTREuqW0r2FcXJz++usvSUmrB03vH04aXrlyRdevX5ejo6M5Abls2TIFBgaqevXqmjVrlkqVKmVuX7JkSc2YMUMeHh4KCAjQihUrzMf8/PwkSd26dZOzs7P58xIlSui///2vmjdvrkqVKmXZvQMAAAAAAAAAAORlJA2Rbk2aNJEkHTlyxPzZkSNHFBkZKV9fX7m7u5s/N1US7tu3TwkJCZL+WZq0fv365j0Rt2zZIknq3Lmz7Ozskl3T0dFRHTp0kCRt377d/HmFChUkSVOmTNGWLVuSVCHWrFlT3333ncaPH/+YdwwAAAAAAAAAAJA/FLB0ALAevr6+Kl68uG7duqVLly7J29vbXEloShKa1KhRQ0WLFlVERIT8/PxUrVq1FPczPHfunKTEisOtW7emeN2bN29Kki5evGj+7I033tD+/ft16dIlDR8+XA4ODqpbt66efPJJtWzZUlWqVMm6GwcAAAAAAAAAIJ9LSGC50LyOSkOkm42NjRo3bizpnyVK/72foYmtra25MnH//v2S/tnP8OGkYWRkpCQpICBAhw8fTvF1+fJlSdLdu3fN/apWrarff/9dzz//vIoWLaqYmBjt379fX331lZ555hk9/fTT5usBAAAAAAAAAAAgbSQNkSGmROCxY8cUHh6u06dPq1ChQuY9Ch9mSg4ePHhQoaGhunz5skqWLClfX19zm0KFCkmSvvnmG509ezbN14EDB5Kc39PTU5988on++usvLVu2TGPHjlWzZs1kb2+vc+fOafDgwbp27Vp2DQUAAAAAAAAAAPmG0WjM8RdyFklDZIgpaXjixAkdOnRIRqNRDRo0MO9R+DBT9eGZM2d06NAhSUmrDCXJ29tbkuTv75/qNQMCAnTixAmFhYVJSvzGFBQUpL1790pKrGqsVauWBg8erPnz52vNmjVycXHRvXv3tHnz5se8YwAAAAAAAAAAgLyPpCEypFy5cvL09JS/v7/27NkjKfnSpCZlypSRj4+PgoODzXsfNm3aNEmb1q1bS5KWL1+u+/fvJztHXFychg0bpueee06TJk2SJEVERKhDhw76z3/+oxMnTiTr4+3trbJly0qSEhISzJ/b2iY+7vx1AgAAAAAAAAAAGZNgNOb4CzmLpCEyrGnTpoqNjdXq1aslpZ40lP6pLNywYYNsbGySVRr27dtXJUuWVGBgoIYOHaqrV6+aj4WFhenNN9/UhQsXZG9vr5deekmS5ObmpubNm0uS3n33XV24cMHcJyEhQYsWLdK5c+dka2trbidJTk5OkqTbt2+b91IEAAAAAAAAAACAVMDSAcD6NG7cWEuWLFFUVJTKli0rHx+fVNs2a9ZMP/74o6Kjo1WlShWVKFEiyfEiRYro66+/1tChQ7V37161bdtWlSpVko2NjS5duqSYmBgVKFBAX331VZK9ED/++GP17t1b586dU9euXVWuXDkVLlxYV69eVXh4uCRp1KhRqlSpkrmPr6+vbG1t9eDBA3Xs2FHu7u6aP3++3NzcsniEAAAAAAAAAADIW6j8y/uoNESGNWnSRDY2NpKS71H4bw0bNjTvd5ha25o1a2rNmjUaPny4fH19FRQUpIsXL6pEiRLq3r27fvvtNz311FNJ+ri7u2v58uV6+eWXValSJd24cUPnzp1TwYIF1aVLF/3yyy965ZVXkvSpUKGCPv/8c3l5eSkiIkLXrl1TcHBwZocBAAAAAAAAAAAgz6DSEBnm5uYmPz+/dLUtVKhQivsOpnTOkSNHauTIkemOo3jx4nr77bf19ttvp7tP9+7d1b1793S3BwAAAAAAAAAAyA9IGgIAAAAAAAAAACBNxgSWJ83rWJ4UAAAAAAAAAAAAyOeoNAQAAAAAAAAAAECaEoxUGuZ1VBoCAAAAAAAAAAAA+RyVhgAAAAAAAAAAAEgThYZ5H5WGAAAAAAAAAAAAQD5HpSEAAAAAAAAAAADSlJBAqWFeR6UhAAAAAAAAAAAAkM9RaQgAAAAAAAAAAIA0JbCpYZ5HpSEAAAAAAAAAAACQz5E0BAAAAAAAAAAAAPI5licFAAAAAAAAAABAmowsT5rnUWkIAAAAAAAAAAAA5HM2RlLDAAAAyGblypVTcHCwPDw8FBQUZOlwAACwCsyfAAAgNznhdznHr1mzSvkcv2Z+RqUhAAAAAAAAAAAAkM+xpyHwGGKjoywdgtWwd3KWxJhlBGOWOf+MW7SFI7Ee9k5OkqQYxizdHP5/zHjOMs6YYFRE+B1Lh2E1irq5SuJZywjT97RbNyMsG4gVKV6iqCSes4xg7swcB8Yt84xGPYhi3NKjoDM/p2WU6XvavUh+98yIQi787plRpmctNirSwpFYD3tnF0lSbORdC0diXexdCls6hDwrIYGFK/M6Kg0BAAAAAAAAAACAfI5KQwAAAAAAAAAAAKTJSKFhnkelIQAAAAAAAAAAAJDPkTQEAAAAAAAAAAAA8jmWJwUAAAAAAAAAAECaElifNM+j0hAAAAAAAAAAAADI56g0BAAAAAAAAAAAQJoSEqg0zOuoNAQAAAAAAAAAAADyOSoNAQAAAAAAAAAAkCYjexrmeVQaAgAAAAAAAAAAAPkcScNcLigoSL6+vubXJ598kq5+8+fPN/dp0aKF+fNx48bJ19dXY8aMyVAc/fv3l6+vr6ZNm5ahfrnV9evXFRkZaekwAAAAAAAAAACwCgnGnH8hZ5E0tDKbNm1KVwnw+vXrcyAa6xMTE6MZM2aoQ4cOunXrlqXDAQAAAAAAAAAAyBXY09CKFChQQKGhoTp06JDq16+farsrV67o5MmTKR576623NGTIEBUuXDi7wszVQkNDNXfuXEuHAQAAAAAAAAAAkKtQaWhFGjduLEnauHFjmu1MVYbVqlVLdszd3V0+Pj5yd3fP+gABAAAAAAAAAECeZDQac/yFnEXS0Ip07NhRkrR58+Y0v1jWr18vW1tbderUKadCAwAAAAAAAAAAgBUjaWhF6tevr5IlSyokJESHDx9Osc3Fixfl5+enhg0bqkSJEsmOjxs3Tr6+vhozZkyyY3fu3NHs2bPVpUsX1alTR82aNdN7772nGzduJGu7dOlS+fr66qmnnkoxjqeeekq+vr4aOXJksmO3b99WtWrVVLVq1ST7CsbHx2vlypUaMGCAGjZsqBo1aqhNmzZ6//33FRAQkOJ17t+/r3nz5qlHjx6qW7euatasqdatW2v06NE6dOhQkrb9+/dX27Ztk8W4f//+FM8NAAAAAAAAAAASJRiNOf5CziJpaEVsbW3VoUMHSakvUWpamrRLly4ZOvfVq1fVu3dvzZo1S5cuXVKFChXk4uKiZcuWqUePHgoNDU3SvnXr1rKxsVFgYKCCgoKSHAsKClJgYKAkaf/+/cmqInfu3Kn4+HjVqVNHxYsXlyRFRUVp8ODBGjdunPbv3y9HR0cZDAZFRERo6dKleuaZZ7R58+Yk54mJidGgQYM0depU+fn5qXTp0qpcubIiIyO1du1a9e3bV8uWLTO3NxgMqlGjhvnf1atXV7169fLt/o4AAAAAAAAAAAAmJA2tjGnJ0dSWKN2wYYPs7e1TrQBMzXvvvaeLFy/K19dXmzZt0urVq7Vx40YtW7ZMNjY2ySr9SpYsqZo1a0qSdu/eneTYnj17zO8jIiLk5+eX5Piff/4pSWrXrp35s/fff1979+5V5cqVtWzZMu3cuVMrVqzQX3/9pddee03379/XmDFjdO7cOXOf3377TUeOHJGXl5e2bNmiDRs2aMWKFdq9e7f69u0ro9GoSZMm6cGDB+ZrzJgxw9x/2rRp+uWXX1Lc+xEAAAAAAAAAAPwjIcGY4y/kLJKGVuaJJ56Qu7u7rl+/riNHjiQ5dvbsWZ0/f15NmzZV0aJF033O48ePa8+ePbKzs9Ps2bPl6elpPlarVi1NnTo1xX5t2rSRlHrSsEGDBpKUZPnP+Ph4c3tT0tDPz0/r1q1ToUKFNH/+fNWqVcvcvmDBgho1apQ6deqkBw8eaO7cueZjpmRkixYtVLZs2SR9xo0bp2bNmql9+/aKiIhI91gAAAAAAAAAAADkRyQNrYyNjY06duwoKfkSpaalSTt37pyhc27fvl1SYkKyfPnyyY43aNBAlSpVSva5KWm4b98+xcXFSZISEhK0f/9+ubq6ql+/fubjJkeOHFFERIQqVaqkChUqSJL++OMPSVLDhg1VqlSpFGN85plnJP2ztKkkeXl5SZKWL1+uxYsXKywszNzewcFB8+fP1+eff57qOQEAAAAAAAAAQPoYjTn/Qs4qYOkAkHGdOnXSjz/+qE2bNmn8+PGysbGRlLg0acGCBZMs+5kely5dkpS4519qqlSpovPnzyf5zNfXV+XKlVNQUJCOHTumJ554QidPnlRERITatWunhg0bysbGRn///bfi4+NlZ2eX4tKk/v7+kqSTJ0+qT58+KV7ftMRoVFSUQkJCVLZsWT3//PNavny5zp8/r48++kgff/yxqlatqiZNmqh58+Zq0KCBChTgEQcAAAAAAAAAAHgUKg2tUN26dVWmTBldv35dR48elSSdOnVKgYGBatWqlVxcXDJ0vjt37kiSnJycUm1TpEiRFD83VRualiQ1/W+TJk1UrFgxGQwGRUZG6uTJk5L+2c+wbdu25nPcvXtXknTr1i0dPnw4xdepU6eSxevi4qIlS5Zo+PDhqlChgoxGo06fPq358+dr0KBBatGihZYuXZqhsQAAAAAAAAAAAMiPSBpaIRsbG3Xo0EHSP0uUZnZpUknm/Q8jIyNTbXP//v0UP/930nDv3r2SEpOGktS0aVNJifsaBgcHy9/fX6VKlVLNmjXN5yhUqJAk6aWXXtLZs2cf+apSpYq5r4uLi0aOHKnNmzdr8+bNmjhxorp27SpXV1fdunVL77//vjZv3pzhMQEAAAAAAAAAAP9IMBpz/IWcRdLQSnXq1EmStHnzZhmNRm3YsEHOzs5q1apVhs/l7e0tSTpz5kyqbf69NKlJgwYN5OrqqhMnTig0NFRHjx6Vu7u7fHx8JP2TPNy3b5+5yrBNmzbmJVUfvr5pmdKUhIeH69ChQ7p69aqM//+N4tatWzp48KB5L8MKFSqoV69emjp1qv7880/VqFFDkrR69epHjgEAAAAAAAAAAEB+RtLQStWpU0dly5bV1atXtWjRIgUHB6tNmzZydHTM8LmeeuopSdLRo0d1+vTpZMf9/Px0/PjxFPsWKFBALVq0UHx8vL7++mvFxMSYE4VSYlLR3t5ehw8f1qZNmyQp2Z6LrVu3liT99ddfunDhQorXmTp1ql588UX179/fnDR8+eWX1bdvX61cuTJZe2dnZ9WpU0eSFB8fb/7c1vafR97IXykAAAAAAAAAAJAuRqMxx1/IWSQNrVjHjh0lSV999ZUkqUuXLpk6j6+vr7p27Sqj0agRI0YkqTj09/fXyJEj0/ziNC1RumzZMklKkjR0cnJS7dq1de/ePe3fv18uLi5q2LBhkv7169dXs2bNFBcXpyFDhujw4cPmYzExMZo7d6753EOGDDEn/p555hlJ0uzZs7Vz584k5zx48KC5wrBly5ZJ4jG5evXqo4YGAAAAAAAAAAAgXyhg6QCQeZ06ddKCBQsUFRWlIkWKqFmzZpk+1wcffKCrV6/q8OHD6t69uypXriwbGxv5+/vL1dVVDRs21IEDB1Ls26JFC9nb2ys2NlZS0qSh6d8HDx6UlJjAc3BwSHaOKVOm6NVXX9WxY8fUp08flStXTkWKFNGVK1d0584dSdKgQYP0wgsvmPsMGDBAe/fu1c6dOzVkyBC5u7vL3d1d4eHhCg4OlpSY0Hz++efNfYoWLSoPDw8FBwdr+PDhqlixot544w21aNEi02MHAAAAAAAAAEBel5BA5V9eR6WhFatVq5bKlSsnSWrfvr3s7e0zfS5XV1f98MMPevfdd1W1alUFBwcrNDRUHTp00LJly1S+fPlU+xYuXFgNGjSQJHl5eal06dJJjjdt2tT8/t9Lk5q4ublp0aJF+vjjj9WwYUPdvXtXZ8+eVYECBdSyZUvNnTtX48ePT9LHzs5Oc+bM0bvvvqu6devq/v378vPz071799SsWTNNnjxZc+fOVYECSXPjM2bMUN26dZWQkKCAgABdvnw5Q2MFAAAAAAAAAACQ19gYWRQWyLTY6ChLh2A17J2cJTFmGcGYZc4/4xZt4Uish/3/L90cw5ilm8P/jxnPWfp5GwwKDg5W2TJlderUmUd3gCSpqJurJJ61jDB9T7t1M8KygViR4iWKSuI5ywjmzsxxYNwyrOL/z58eZcvqwjl/S4djFQo683NaRpm+p92L5HfPjCjkwu+eGWV61mKjIi0cifWwd3aRJMVG3rVwJNbF3qWwpUPIs7bs9cvxa7ZrWiXHr5mfUWkIAAAAAAAAAAAA5HMkDQEAAAAAAAAAAIB8rsCjmwAAAAAAAAAAACA/Y7e7vI9KQwAAAAAAAAAAACCfo9IQAAAAAAAAAAAAaUqg0jDPo9IQAAAAAAAAAAAAyOeoNAQAAAAAAAAAAECaKDTM+6g0BAAAAAAAAAAAAPI5Kg0BAAAAAAAAAACQpoQESg3zOioNAQAAAAAAAAAAgHyOpCEAAAAAAAAAAACQz7E8KQAAAAAAAAAAANKUYGR50ryOpCHwGOydnC0dgtVhzDKOMcsceycnS4dgdRwYswzjOcs4G1sbFXVztXQYVodnLeOKlyhq6RCsDs9ZxjF3Zg7jlgk2NirozLhlBN/TMq6QC797ZgbPWsbZO7tYOgSrY+9S2NIhAMgnSBoCAAAAAAAAAAAgTRQa5n0kDYHH8CDspqVDsBoFi5WQJN08dsTCkViPErXrSpKWNm5m4UisS699uyVJ+94dZ+FIrEfjz76QJEVfv2rhSKyHU+mykqTYqEgLR2J9jPFxunPxvKXDsBquFStJkm6dPGHhSKxH8Ro1JUnnLvA9Lb0MPv//Pe3uHQtHYj3sCydWTD+IirZwJNbFVCkXGx1l4Uisj9FoVPRdxi09nAonVsvx9Zl+pq9NnrGMMT1rMdE8a+llqjSPiQi3cCTWw6GomyTp/o0QC0diXRxLlrJ0CMhjTp06pfnz5+vAgQOKiIiQm5ub6tWrp4EDB6pevXqPde59+/Zp4cKFOnr0qCIjI1WiRAk1adJE//nPf2QwGFLtN3PmTM2ZM+eR53///ffVr1+/dMdjm+6WAAAAAAAAAAAAyJcSjMYcf1naxo0b1atXL61bt043btxQbGysQkNDtXHjRvXt21ffffddps+9cOFCDRw4UNu3b1d4eLhiY2N17do1rVixQj179tSaNWtS7Xv69OlMXzctVBoCAAAAAAAAAAAADzl58qTGjh2ruLg41a9fX2+99Za8vb0VEBCg6dOna//+/ZoyZYoqVaqkVq1aZejc27Zt0xdfJK7+1b59ew0dOlRlypSRn5+fJk2aJD8/P40fP14+Pj6qVq1asv6nTp2SJI0dO1Z9+vRJ9ToODg4ZiotKQwAAAAAAAAAAAKTJaDTm+MuSZsyYoZiYGFWuXFkLFizQE088oWLFiqlevXqaP3++GjRoIKPRqC+//FIJCQnpPq/RaNRXX30lo9GoJ598UjNnzlT16tVVrFgxNW3aVIsWLZK3t7diY2M1derUZP1v3bql0NBQSVK9evXk7Oyc6sve3j5D90zSEAAAAAAAAAAAAPh/Fy5c0M6dOyVJQ4cOVcGCBZMct7e315gxY8xtDx8+nO5z7969W/7+/pKkkSNHytY2aarOxcVFr7/+uiRpz549unbtWpLjJ0+elCQVKFAgxSrEx0HSEAAAAAAAAAAAAPh/u3btkiTZ2dmpRYsWKbapXbu2ihcvLknaunVrus+9Y8cOSVKxYsVUu3btFNu0bNlSdnZ2MhqN2rZtW5Jjpv0MK1WqJEdHx3RfNz1IGgIAAAAAAAAAACBNCcacf1nKmTNnJEmenp4qXLhwim1sbGxUpUoVSf/sMZgefn5+kqSqVavKxsYmxTYuLi7y9PSU9E9loYkpaVitWjWtWrVKAwYMUP369VWzZk116NBBkyZNUlhYWLrjeRhJQwAAAAAAAAAAAOD/BQcHS5I8PDzSbFemTBlJUlBQUI6d25SgXLNmjd555x3t379fd+/eVUxMjAICArRgwQJ16tRJhw4dSndMJgUy3AMAAAAAAAAAAAD5SoIFSv/atm2b5vGMLAuaEeHh4ZIkV1fXNNuZqhDv3LmTI+e+ffu2OekYGxur7t27q2/fvvL09NStW7e0Zs0azZ8/XxEREXrttdf022+/qXz58umOjaQhAAAAAAAAAAAA8pRZs2Zp9uzZGeozYsQIvf7663rw4IEkPXLPwIIFC0qSuX16PM65r169qrJly+r69et65513NGjQIPMxNzc3jRo1StWrV9frr7+uO3fuaPLkyZo1a1a6YyNpCAAAAAAAAAAAgDQZLbDHYHZVEj6KnZ1dtp47ISEhU32rVq2q7du3KyYmRg4ODim2eeqpp9S6dWtt375dW7Zs0e3bt1WkSJF0nT/PJQ2DgoLSLFe1t7eXi4uLvLy81KpVK/Xr108uLi45GGH2GjdunFauXJnqcRsbGzk6Osrd3V01a9bUwIEDVatWrRyMMLn9+/drwIABkhLX4i1QIM89lgAAAAAAAAAAIAf17dtXnTt3zlAfNzc3SVKhQoUkPbqCML1Vgw8rVKiQYmNjH+vcqSUMTdq2bavt27crISFBJ0+e1JNPPpmu2PJ0dsZgMCRLCMbGxiosLExHjhzRkSNHtGTJEi1cuFAVKlSwUJTZw8XFRQaDIdnnRqNRN2/eVGBgoAIDA7V+/Xp9/PHHev755y0QJQAAAAAAAAAAsAYJlig1fAzFihVTsWLFMtXXtJ/g3bt302xn2m/QlGxM77nv3LmTLec2KVOmjPl9WFhYuvvl6aThe++9p0aNGqV4bP/+/Ro2bJiuXr2qd955R7/++msOR5e9qlWrpp9++inV4xcuXNBbb70lPz8/ffTRR2rcuLE8PT1zMMJ/1KpVS+vXr5ckqgwBAAAAAAAAAIBFeXt768CBA7p27Vqa7a5fvy4paZIuPecODg5+rHMbjUbZ2Nik2jc2Ntb83lQ1mR626W6ZxzRq1EhvvfWWJOnIkSM6efKkhSPKWT4+PpoxY4bs7OwUGxur5cuXWyyWQoUKycfHRz4+PhaLAQAAAAAAAAAAQJIqV64sSQoMDFR0dHSKbYxGo86cOSMpsZAro+c29U3J3bt3deXKlWTn/vTTT9W4cWM1adIkzX0Rz58/b37v5eWV7tjybdJQktq3b29+f+zYMQtGYhleXl7y9vaWJJ07d87C0QAAAAAAAAAAgNzKaDTm+MtSWrZsKSmxYm/Xrl0ptjl69Kh56c/mzZtn+NyhoaGpFrTt2LFD8fHxkqRmzZqZPy9atKjCw8MVHh6uo0ePptjXaDRq3bp1kiQPD48MFWzl66ShaU1aSYqKijK/79+/v3x9fbVjxw7NmTNHTZs2Ve3atdW1a1dduHDB3O7mzZv68ssv1blzZ9WuXVt169ZVz549tWDBgmQbWI4fP16+vr4aOnRoqvH8/vvv8vX1VevWrdPMEGclW9vERyC1L74tW7bolVdeUZMmTVSjRg01b95co0eP1qlTp1I9Z0REhGbNmqUuXbqoTp06atKkiUaPHq1Lly5p1qxZ8vX11axZs8zt9+/fL19fX/n6+iouLi7Z+U6cOKGxY8eqVatWqlGjhho2bKj+/ftr+fLl5i+ah5n+++3cuVN+fn5644031LRpU9WoUUNt27bVZ599lqE1fAEAAAAAAAAAQP5Rvnx51a1bV5I0c+bMJDkkKTGZOHXqVEmSwWBQkyZN0n3uhg0bmpcc/fLLL5PlRSIjIzV79mxJUqtWrZJUCnbp0sWc1/n8889TzJF8++235irGgQMHprmM6b/l66RhYGCg+X3p0qWTHf/mm280c+ZMOTs7q3Tp0oqMjDT/xzl06JC6dOmi+fPn6/Lly/L09FTZsmV16tQpTZo0Sb169dKNGzfM5+rRo4ckadeuXYqIiEgxntWrV0uSunfvbv6Pnp38/f3l7+8vSeaH3yQuLk5jxozR8OHDtWPHDtnY2MjX11cxMTFau3atnn/+ef3888/JznnlyhX16tVLs2fP1qVLl1ShQgUVKVJEa9euVY8ePVLNfKfm22+/Va9evfT777/r7t278vX1lYuLiw4cOKD//ve/GjRoUKqbhe7cuVPPPfectmzZIjc3N5UpU0ZBQUH64Ycf9MILLygyMjJDsQAAAAAAAAAAkF8lGHP+ZUnjx4+XjY2Nzp8/r/79+2vfvn3mCr/Bgwfr77//lo2NjUaNGpUsMXf8+HF17NhRHTt2TJZLsbOz0zvvvCMpsajqlVde0bFjxxQeHq6//vpL/fr106VLl+Tg4KDXX389SV8vLy/169fPfI2+ffvqr7/+0q1bt3T27FlNmDDBnMysX7++uW16FchQ6zzm+++/lyTZ29uradOmyY4fPnxYY8aM0ZAhQyRJYWFhsrOzU0hIiIYNG6aIiAj16tVLY8eOlaurqyTp8uXLGjNmjI4dO6Y333xTixYtkpT4H6d8+fK6fPmyNmzYoD59+iS51o0bN/TXX3/JxsZGzz77bHbetqTEfRzHjRsno9GoEiVKqFevXkmOz5gxQ2vWrFHp0qX1ySefmEtr4+PjtXjxYn3xxRf65JNP5O3trSeffFJSYrXi22+/rcDAQNWoUUPTp0+Xp6enpMTlX0eMGKHdu3enO8ZNmzZpypQpkqRhw4Zp6NChcnBwkCTt27dPo0eP1oEDB/T222/r66+/Ttb/p59+UqtWrTRx4kS5u7tLkrZu3arXX39dgYGBWr58uQYNGpSxgQMAAAAAAAAAAHle7dq19dFHH+nDDz/UqVOnNHDgwGRtxo0bpzZt2iT7/N69e7p06ZIkKTw8PNnxTp066ezZs/r666+1Z88e7dmzJ8nxAgUKaPLkyapRo0ayvu+8844iIiL0+++/68iRIynmORo2bKivv/5adnZ26b3dxOtmqHUecP/+fV28eFFLlizRqlWrJEmDBg1SiRIlkrX18PDQ4MGDzf8uVqyYJGn+/PmKiIhQmzZtNHHixCR9ypcvr7lz56pDhw46ePCgduzYoZYtW5qTgTNmzNDvv/+eLGm4du1axcfHq0GDBipfvvxj3+fp06eTXUNKvP+QkBDdunXLfI+zZ8+Wm5ubuc3Nmze1cOFCSdLcuXNVvXp18zE7Ozv1799f169f13fffafp06ebk4Y7d+7U4cOH5eTkpG+++UYlS5Y096tdu7Zmz56dLDmZlmnTpkmSevfurTfeeCPJscaNG2v27Nl64YUXtG3bNh08eFD169dP0qZ48eKaOXOmChYsaP6sbdu2atGihbZv367Dhw+TNAQAAAAAAAAAIB0SLLjHoKX07t1bNWrU0IIFC3TgwAGFh4fL2dlZdevW1cCBAzO0LOm/vfnmm2rUqJF+/PFHHTt2TLdv35abm5saNGigwYMHJ8nNPMyUUOzSpYuWLFmi48eP6/bt23J1dVWVKlXUrVs3PfPMMxlaltR87kzfjRUYMGDAI9s8//zzyRJSJnXr1k1xULds2SJJ6tatW4r9SpQooSeffFKbNm3S9u3bzZtaPvvss5o1a5aOHDmioKAglStXztzHlMA0LWP6uCIjI3X48OEUjzk5OalTp05q2bKlunTpYq7eM9m5c6diYmJUqVKlVB/KZ555Rt99952OHz+uW7duqXjx4uZxad++fZKEoYlp38cjR448Mv6AgABzFj6l7L2U+N/HdL6tW7cmSxo2adIkScLQxMfHR9u3b091WVMAAAAAAAAAAABJql69unnJz/Rq1KiRzp49+8h2TZo0yXTisVWrVmrVqlWm+qYmTycNDQaDXFxczP+2sbFRwYIFVbRoUfn6+qpdu3aqVKlSqv1TSnxFRUUpODhYUmIV3o8//phiX1Obixcvmj8rU6aMmjZtqt27d2vNmjUaOnSoJOns2bPy8/OTk5OTOnTokPEbTUHDhg31008/SUpcNjQqKkqbNm3Sl19+qYiICMXExKhNmzbJEoaSzPscXr9+PcVqRdM5TS5evKjixYub+1WpUiXVuGrUqJGupKFp3AoVKiQfH59Hns+UYHxYqVKlUuzj6OgoSck2FwUAAAAAAAAAACkzJlg6AmS3PJ00fO+999SoUaNM90+pSi0yMtL8/ty5c488x7+r2Xr06JEsabh69WpJUseOHeXs7JzpeFNjY2MjFxcX9ezZUzVq1FCfPn20detWDRkyRAsXLpSTk1OKMadVrfiwO3fuSPpnXd5/n+9h6b0/0zg/nPRN63xRUVHJjtnb26frWgAAAAAAAAAAAPldnk4aZodChQqZ369Zs0YGgyFD/du1aydXV1dduHBBp0+fVpUqVbR27VpJicuXZjdfX199+OGHGjt2rI4dO6aPPvpIkyZNStLGdI8dOnTQzJkz031uU7+HE6v/llJyLyWmZGBa55L+SVhmR7IVAAAAAAAAAAAgv7C1dADWxtXVVSVKlJAknT9/PtV2Z8+e1ZkzZ3T79u0knxcsWFBdunSRJG3atEn79+9XSEiIPD091aBBg+wL/CHdunUzL4O6atUqbdq0Kclxb29vSf8sU5qSe/fu6cCBA7py5Yri4+MlyZxATWud3vSs4StJFStWNF/nwoULqbY7efKkJKlChQrpOi8AAAAAAAAAAMi4BKMxx1/IWSQNM8G0seTPP/+shITki/jevXtXAwYMUPfu3fXDDz8kO96zZ09J0h9//KGNGzdKSqwytLGxyb6g/+WDDz5Q0aJFJUkTJ040V+xJUsuWLWVnZ6eLFy9qz549KfZfuHCh+vfvr2eeeUb37t2TJD311FOSpG3btiksLCxZnwsXLujgwYPpis/b29ucvExpDCXp8OHDOn78uCSpRYsW6TovAAAAAAAAAAAAkiNpmAmvvPKKnJycdOjQIY0dOzZJgiw4OFivvPKKIiIiVLhwYfXt2zdZ/5o1a8pgMOjChQtas2aNbGxski1Naqqwu3Dhgjkpl5WKFy+usWPHSpJu3LiRZIlSDw8PPf/885Kkt956S9u2bTMfS0hI0LJlyzR79mxJUt++fc37DrZu3VrVq1dXZGSkhg8frmvXrpn7nTt3TsOGDUsxyZqaN954Q5K0ZMkSzZw5UzExMeZj+/fv18iRIyVJzZs3V9OmTTN0/wAAAAAAAAAAIP2Mxpx/IWexp2EmVKhQQdOnT9eoUaO0du1abdq0SZUqVVJsbKwCAgIUFxcnJycnzZs3T8WLF0/xHD169NAXX3yhqKgoNWnSRGXLlk1y/Pjx4xowYIAk6ccff1SjRo2y/D6ee+45rV69WgcOHNDy5cv19NNPq3HjxpKkd999VyEhIdq+fbuGDh0qd3d3lSpVSsHBweYkaYcOHfTmm2+az2dnZ6dp06apX79+Onz4sNq1a6fKlSsrNjZWFy5ckKurq7y8vBQQECA7O7tHxtepUyddvnxZ06ZN05w5c/TDDz/I29tbYWFhCg4OliQ1bNhQkydPztEqTQAAAAAAAAAAgLyGSsNMatmypdatW6dBgwapfPnyunTpkgIDA+Xh4aEXX3xRv//+u+rVq5dq/27dusne3l6SklUZ5qSJEyeqYMGCkqQJEybo/v37khL3Xvz66681bdo0NW/eXLGxsTpz5ozi4+PVqFEjTZo0SdOnT0+W/KtQoYJWr16tgQMHqkyZMrpw4YLCwsL09NNP67ffflO5cuUkSYUKFUpXfK+++qqWLl2qrl27ysXFRX5+frp//76aNGmiSZMm6YcffpCbm1sWjggAAAAAAAAAAPg39jTM+2yMRkYdOadnz546efKkpkyZoqefftrS4Ty2B2E3LR2C1ShYrIQk6eaxIxaOxHqUqF1XkrS0cTMLR2Jdeu3bLUna9+44C0diPRp/9oUkKfr6VQtHYj2cSieuEBAbFWnhSKyHt28VBQcHq2zpUjqzZ7elw7EarhUrSZJunTxh4UisR/EaNSVJ5y7wPS29DD7//z3t7p1HtISJfWFXSdKDqGgLR2JdCjo7SZJio6MsHIn18Db4Js6fZcvK3++cpcOxCk6FnSXx9ZkRpq/N6Lt8bWaE6VmLieZZSy8Hp8RnLSYi3MKRWA+HoonFEvdvhFg4EuviWLKUpUPIsxasOpjj13ype/0cv2Z+RqUhsszy5cvVoUMHffLJJykeDwkJ0dmzZyVJ1apVy8nQAAAAAAAAAADAY0gw5vwLOYs9DZFlatSooYCAAF25ckW1a9dW165dzXsNBgUFacyYMYqNjVXjxo3l4+Nj4WgBAAAAAAAAAABgQtIQWaZKlSoaMGCAfvzxR40ZM0aff/65ypQpo8jISF2+fFkJCQmqVKmSJk2aZOlQAQAAAAAAAAAA8BCShshS//3vf9WmTRv9/PPPOnv2rPz9/eXk5KSaNWuqU6dOeuGFF1SoUCFLhwkAAAAAAAAAADLAaGS90LyOpCGyXJMmTdSkSRNLhwEAAAAAAAAAAIB0ImkIAAAAAAAAAACANCVQaZjn2Vo6AAAAAAAAAAAAAACWRaUhAAAAAAAAAAAA0kShYd5HpSEAAAAAAAAAAACQz1FpCAAAAAAAAAAAgDQlJFBqmNdRaQgAAAAAAAAAAADkcyQNAQAAAAAAAAAAgHyO5UkBAAAAAAAAAACQJlYnzftsjEYj/5kBAACQrcqVK6fg4GB5eHgoKCjI0uEAAGAVmD8BAEBuMnvJ/hy/5ojejXL8mvkZlYYAAAAAAAAAAABIEyVoeR9JQ+AxRAVdsXQIVsO5nKckKTY6ysKRWA97J2dJUkxEuIUjsS4ORd0kSfdv3rBwJNbDsURJSVJMeJiFI7EeDm7FJEmxd+9YOBLrE3//ni79vtLSYVgN727PSpJWte9g4UisR/c/NkmSQg8esHAk1sO9fkNJ0uFJn1k4EutR7513JUkPwm5aOBLrUrBYCUnSjaOHLRyJFTIa+V0qnUy/R8VGR1s4Euth7+QkSYq+GmzhSKyLU1kPSVLwju0WjsR6eLRsLUmKvBJo4Uish4tnBUnS3YCLFo7EuhT2qmjpEACrRdIQAAAAAAAAAAAAaUqg1DDPs7V0AAAAAAAAAAAAAAAsi0pDAAAAAAAAAAAApIlCw7yPSkMAAAAAAAAAAAAgnyNpCAAAAAAAAAAAAORzLE8KAAAAAAAAAACANCWwPmmeR6UhAAAAAAAAAAAAkM9RaQgAAAAAAAAAAIA0JVBomOdRaQgAAAAAAAAAAADkc1QapiIoKEht27ZN9bi9vb1cXFzk5eWlVq1aqV+/fnJxccnBCHPW7t27tXbtWh05ckTXr19XfHy8SpQooVq1aqlbt25q165div3atGmj4OBgffLJJ3r++eclSfv379eAAQMkSadOnVKBAo/3GPbv318HDhzQa6+9plGjRj3WuQAAAAAAAAAAQHJG9jTM80gapoPBYEiWEIyNjVVYWJiOHDmiI0eOaMmSJVq4cKEqVKhgoSizR0hIiMaOHav9+/dLkgoWLKiyZcvK3t5eQUFB2rRpkzZt2qSGDRtq+vTpKl68uIUjBgAAAAAAAAAAQEaRNEyH9957T40aNUrx2P79+zVs2DBdvXpV77zzjn799dccji77XLhwQf369VNYWJi8vLw0cuRIdezYUXZ2dpKkuLg4rV69WlOnTtWBAwc0cOBA/frrr4+suKxVq5bWr18vSY9dZShJkyZN0r179+Tm5vbY5wIAAAAAAAAAAMmxp2Hex56Gj6lRo0Z66623JElHjhzRyZMnLRxR1oiJidFbb72lsLAwVatWTUuWLFGXLl3MCUMpMeHXs2dPLVy4UI6OjvL399f06dMfee5ChQrJx8dHPj4+WRJr2bJl5ePjo2LFimXJ+QAAAAAAAAAAAPIbkoZZoH379ub3x44ds2AkWWfhwoXy8/OTra2tJk+erKJFi6ba1mAwqH///pKkZcuWKTIyMoeiBAAAAAAAAAAAQFYgaZgFChcubH4fFRVlft+/f3/5+vpqx44dmjNnjpo2baratWura9euunDhgrndzZs39eWXX6pz586qXbu26tatq549e2rBggV68OBBkmuNHz9evr6+Gjp0aKrx/P777/L19VXr1q2VkJCQqXtaunSpJKl169aqVKnSI9v3799fkydP1oYNGx65POn+/fvl6+srX19fxcXFmT8fN26cfH199csvvygoKEjjx49XixYtVKNGDbVo0UL//e9/FRQUlOK1fX19NW3atGTHYmJitHjxYvXp00eNGjVSjRo19NRTT+mTTz7RjRs3HnlfAAAAAAAAAABAMhqNOf5CzmJPwywQGBhofl+6dOlkx7/55hsdPnxY5cuXV+HChRUZGSkvLy9J0qFDhzRs2DBFRETI3t5eXl5eMhqNOnXqlE6ePKnVq1fru+++U8mSJSVJPXr00IoVK7Rr1y5FRESkWAG4evVqSVL37t1la5vxvPCVK1d05coVSdKTTz6Zrj6lSpVSt27dMnytlJw+fVpTpkxRdHS0ypcvrwoVKuj8+fNavny5tm3bphUrVqhMmTKPPE9oaKiGDh1qXjK2QoUKKl26tC5evKiffvpJGzdu1JIlS+Th4ZElcQMAAAAAAAAAAFgrKg2zwPfffy9Jsre3V9OmTZMdP3z4sMaMGaM//vhDmzZt0ooVK2RnZ6eQkBBzwrBXr17au3ev1q5dq3Xr1mnz5s2qXbu2/Pz89Oabb5rPVb9+fZUvX16xsbHasGFDsmvduHFDf/31l2xsbPTss89m6n4uXrxofu/r65upczyOpUuXqlKlSlq/fr02bdqkdevW6ddff5Wzs7PCwsK0YMGCdJ3n3Xff1cmTJ1W+fHmtWLFCmzdv1urVq7V161bVrVtXN27c0OjRo7P5bgAAAAAAAAAAsH4JCTn/Qs4iaZhJ9+/f1+nTp/XBBx9o1apVkqRBgwapRIkSydp6eHho8ODB5n8XK1ZMkjR//nxFRESoTZs2mjhxolxdXc1typcvr7lz58rFxUUHDx7Ujh07JClJMvD3339Pdq21a9cqPj7enFzMjNu3byeLNSfZ29tr9uzZ8vb2Nn9Wt25d9ejRQ1JiEvZRjh07pl27dsnW1lZz585V9erVzcfc3d311Vdfyc7OTkeOHEmSJAUAAAAAAAAAAMiPWJ40HQYMGPDINs8//7zeeOONFI/VrVtXNjY2yT7fsmWLJKW6rGeJEiX05JNPatOmTdq+fbtatmwpSXr22Wc1a9YsHTlyREFBQSpXrpy5jymBaUqwZYaTk5P5/cN7DuaUGjVqmJdjfVjFihUlSXfv3n3kObZt2yZJqlevnipXrpzseNmyZbVy5Uq5u7vLzc3tMSMGAAAAAAAAACBvS2CLwTyPpGE6GAwGubi4mP9tY2OjggULqmjRovL19VW7du1UqVKlVPunlACLiopScHCwJGnu3Ln68ccfU+xravNwNVyZMmXUtGlT7d69W2vWrNHQoUMlSWfPnpWfn5+cnJzUoUOHjN9oCvGGh4dn+jyZVapUqRQ/d3R0lJS+RObly5clSVWqVEm1jSWWXgUAAAAAAAAAAMiNSBqmw3vvvadGjRplun/BggWTfRYZGWl+f+7cuUee49/VdT169EiWNFy9erUkqWPHjnJ2ds50vOXLl5etra0SEhLk7++f7ns/c+aMKleurAIFHu+xsre3f6z+khQRESEpadUkAAAAAAAAAADIHKORUsO8jqShhRQqVMj8fs2aNTIYDBnq365dO7m6uurChQs6ffq0qlSporVr10qSec/DzHJzc9MTTzyhv//+W3v27FG/fv0e2SckJETPPvusChUqpKlTp6pNmzaPFcPjMo1vVFSUReMAAAAAAAAAAACwBraWDiC/cnV1VYkSJSRJ58+fT7Xd2bNndebMGd2+fTvJ5wULFlSXLl0kSZs2bdL+/fsVEhIiT09PNWjQ4LHjM517x44d8vf3f2T7xYsXy2g0KjY2VrVq1Xrs6z8uLy8vSUoz9vfff1+vvvqqduzYkUNRAQAAAAAAAAAA5E4kDS2oVatWkqSff/5ZCQkJyY7fvXtXAwYMUPfu3fXDDz8kO96zZ09J0h9//KGNGzdKSqwytLGxeezYnn/+eRkMBsXHx2v8+PHm5T5Tcvz4cS1YsECS9MILL5iToZbUsmVLSdKhQ4d06dKlZMdv3bql1atX688//zTvlQgAAAAAAAAAAFKWYMz5F3IWSUMLeuWVV+Tk5KRDhw5p7NixCgsLMx8LDg7WK6+8ooiICBUuXFh9+/ZN1r9mzZoyGAy6cOGC1qxZIxsbm2RLk967d08XLlzQhQsXdO/evXTHVqBAAX366acqXLiwTpw4od69e2vjxo2Kj483t3nw4IEWL16sQYMGKSYmRgaDQW+99VYmRiLrNWrUSA0aNFB8fLxGjBihCxcumI+FhITojTfe0IMHD1SrVq3H2q8SAAAAAAAAAAAgL2BPQwuqUKGCpk+frlGjRmnt2rXatGmTKlWqpNjYWAUEBCguLk5OTk6aN2+eihcvnuI5evTooS+++EJRUVFq0qSJypYtm+T48ePHNWDAAEnSjz/+mKEEWa1atbRo0SK99tprCggI0BtvvCEnJyd5enrK1tZWFy9e1IMHDyRJTZs21VdffSUnJ6dMjkbWmzp1qgYPHqxz586pS5cu8vHxka2trS5duqTY2Fh5eHjoq6++snSYAAAAAAAAAADkeglGSv/yOpKGFtayZUutW7dOCxcu1K5du3Tp0iXFx8fLw8NDTz75pF566SV5enqm2r9bt26aOnWqYmNjk1UZZgVfX19t2LBBK1eu1LZt2+Tn56eLFy/KxsZGJUuWVK1atdS9e3fzUqu5SalSpbR06VItXrxY69ev18WLFxUXFydPT0899dRTGjx4sAoXLmzpMAEAAAAAAAAAACzOxmgkNQxkVlTQFUuHYDWcyyUmv2OjoywcifWwd3KWJMVEhFs4EuviUNRNknT/5g0LR2I9HEuUlCTFhIc9oiVMHNyKSZJi796xcCTWw7tqNQUHB6t08WLau+A7S4djNby7Jf5R2Kr2HSwcifXo/scmSVLowQMWjsR6uNdvKEk6POkzC0diPeq9864k6UHYTQtHYl0KFishSbpx9LCFI7Eedbt2U3BwsDzKltUl/3OWDscqmH6Pio2OtnAk1sP+/1eNir4abOFIrItTWQ9JUvCO7RaOxHp4tGwtSYq8EmjhSKyHi2cFSdLdgIsWjsS6FPaqaOkQ8qwJ83bl+DU/fqV5jl8zP2NPQwAAAAAAAAAAACCfY3lSAAAAAAAAAAAApCmBdSvzPCoNAQAAAAAAAAAAgHyOpCEAAAAAAAAAAACQz7E8KQAAAAAAAAAAANKUYGR90ryOSkMAAAAAAAAAAAAgn6PSEAAAAAAAAAAAAGmi0DDvo9IQAAAAAAAAAAAAyOeoNAQAAAAAAAAAAECaEqg0zPOoNAQAAAAAAAAAAADyOSoNAQAAAAAAAAAAkCYjmxrmeVQaAgAAAAAAAAAAAPkcSUMAAAAAAAAAAAAgn7MxUk8KAACAbFauXDkFBwfLw8NDQUFBlg4HAACrwPwJAABykzGzd+T4NaeMaJnj18zPqDQEAAAAAAAAAAAA8rkClg4AsGYx4WGWDsFqOLgVkySFhjJm6eXunjhmMdHRFo7Eujg4OUmSYiMjLRyJ9bB3cZEkRd+NsnAk1sOpsLMkKZavzwyLj0/QxcDrlg7DalSsUFqS9FurthaOxHr0/HOrJOng8UsWjsR61K/lLYnvaRlh//8/b9yJuGvhSKyLa9HCkqTYKH5Oyyjmz/QzzZ33o/jZNr0cnRN/tuV3z4wx/+7JuKWbaf7kWUs/03MWExFu4Uisi0NRN0uHkGclsG5lnkelIQAAAAAAAAAAAJDPUWkIAAAAAAAAAACANCUYKTXM66g0BAAAAAAAAAAAAPI5Kg0BAAAAAAAAAACQJgoN8z4qDQEAAAAAAAAAAIB8jqQhAAAAAAAAAAAAkM+xPCkAAAAAAAAAAADSlJBg6QiQ3ag0BAAAAAAAAAAAAPI5Kg0BAAAAAAAAAACQpgSjpSNAdsuRpGFkZKRWrVqlbdu26ezZs4qIiJCDg4M8PT3VpEkTvfDCC/L29k7Wb//+/RowYIAk6dSpUypQgBzno8yaNUuzZ89O9biNjY0cHBxUokQJVa1aVX369FGzZs1yMMLkgoKC1LZtW0nS5s2bVaFCBYvGAwAAAAAAAAAAkN9kexZu+/btGj9+vMLDwyVJRYsWlcFg0O3bt3X+/HmdPXtWixYt0ogRI/Taa69ldzj5hoODg2rUqJHsc6PRqIiICAUGBio4OFhbtmzR0KFD9eabb+Z8kAAAAAAAAAAAwCoYjZQa5nXZmjRcsGCBJk2aJEnq1KmThg8frsqVK5uPh4aG6uuvv9bixYs1bdo03b9/n+RVFilZsqR++eWXVI9fv35d77zzjvbt26evv/5azZo1U/369XMwwn+UKlVK69evlySVLVvWIjEAAAAAAAAAAADkZ7bZdeKDBw9qypQpkqThw4dr+vTpSRKGkuTu7q4PPvhAw4YNkyT973//08mTJ7MrJDykdOnSmj59uooUKSJJaSYYs5u9vb18fHzk4+Mje3t7i8UBAAAAAAAAAABSlmDM+RdyVrYkDY1GoyZMmKD4+HjVqVNHI0eOTLP90KFDVaZMGSUkJOj777/PjpCQAjc3N9WpU0eS5O/vb9lgAAAAAAAAAAAAYDHZsjzpoUOHdOHCBUnSkCFDHtnewcFBn332mSSZk1j/FhcXp0WLFmnVqlW6ePGiHBwc5OvrqxdffFGdO3dOsc+ZM2e0ePFiHTx4UCEhIYqJiZGrq6uqV6+unj17qmPHjknar1ixQuPHj1fnzp316aefat68edqwYYOuXr0qJycn1a1bV4MHD051Gc+9e/dq4cKFOn36tO7cuSNvb2/16tVLffr0UdWqVSVJZ8+eTdbv77//1k8//aTDhw8rIiJCrq6uqlOnjvr3768mTZo8cvweh61tYt44tbWIMxPbvXv39Ouvv2r16tUKDAyUvb296tWrp2HDhun8+fMaP368nn32WX3xxReSpKCgILVt21aStHnzZlWoUCHJ+S5duqSFCxdq7969unbtmgoWLCgfHx917txZL7zwghwdHZO0HzdunFauXKkPP/xQzZs315w5c7Rnzx6FhYWpWLFiat68uYYOHapy5co91tgBAAAAAAAAAADkFdmSNNy7d68kyc7OTo0bN05Xn6ZNm6Z5fPDgwfr7779VokQJVaxYUYGBgfr777/1999/6+LFixoxYkSS9osXL9bEiROVkJCgIkWKqEKFCrp//76CgoK0c+dO7dy5U6+99ppGjRqV7Fp37txR7969de7cObm7u6tSpUo6f/68tm/frp07d2ru3Llq1apVkj5z5szRzJkzJUklSpRQ5cqVFRAQoI8//lj79u1L9b6mTJmib7/9VpJUpEgRGQwGhYaGauvWrdq6dauGDBmiMWPGpGcIMywsLEz79++XJNWtWzdLYgsPD9crr7yi48ePy8bGRj4+PrK1tdWff/6pXbt2qUWLFhmK8ffff9d///tfxcTEyNHRUQaDQVFRUTp27JiOHTum3377Td9++61Kly6drO/p06c1ZcoURUdHq3z58qpQoYLOnz+v5cuXa9u2bVqxYoXKlCmToXgAAAAAAAAAAMiPWC4078uW5UkvXrwoSfLw8JCLi0uWnPPEiROaNGmSdu3apZUrV2rXrl1q166dJOmbb75RZGSkuW1AQIA+++wzJSQk6M0339SePXu0cuVKbdiwQbt27VKnTp0kSfPnz9ft27eTXWv37t0KDw/X/PnzzdfbunWrfH19FR8fr2nTpiVrP3PmTNna2uq9997Trl279Ntvv2nPnj3q16+fNm/enOI9/frrr/r222/l6uqqyZMn68CBA1qxYoV27dqladOmycnJSd9++62WLVuWJWP4MH9/f7322muKjo5WoUKF9J///CdLYvvkk090/PhxeXp6atWqVVq3bp3WrFmjdevWydvbW9u2bUt3jMeOHdP48eMVExOjXr16ac+ePVqxYoU2bdqkVatWycvLS+fOndOwYcMUFxeXrP/SpUtVqVIlrV+/Xps2bdK6dev066+/ytnZWWFhYVqwYEHmBg8AAAAAAAAAACCPyZakoSkRV6xYsSw759ChQ9W9e3fzcprOzs6aMGGCJCk2NlZHjx41t92zZ4/s7OxUvXp1DR06VPb29uZjRYsW1TvvvGPud+nSpRSvN2HCBDVr1sz8b3d3d3M1o5+fn6KioszHZsyYIUkaNGiQ+vfvb47R0dFR77//forVdTExMZo1a5Yk6bPPPlO3bt3Mx2xsbNS5c2eNHTtWkjRr1qwUk2JpuXHjhvr06ZPs1aNHD7Vo0UJdu3bVsWPH5Obmpjlz5sjb2/uxY7tw4YLWrl0rGxsbzZkzR1WqVDH38/Hx0f/+979kS4mmZebMmYqLi1OzZs00ceLEJAnoqlWr6rvvvpOjo6NOnTqldevWJetvb2+v2bNnJ7m3unXrqkePHpKkw4cPpzsWAAAAAAAAAADyM6Mx51/IWdmSNCxUqJCkxKRcVnnqqaeSfVaqVCkVLVpUUuJSmyZ9+/bVsWPHtHjx4hTP9XDi6t69e8mO29nZpZjo8/HxMb83VTaGhIToxIkTkqQXX3wxxesNHDgw2WdHjhzRzZs35ezsbN7P79+6desmW1tbhYSE6PTp0ym2SU1MTIwOHz6c7HXq1CndunVLbdu21YQJE7RlyxY9+eSTWRLbH3/8IUl64okn5Ovrm6yPh4eHuTr0UaKjo81Lpw4YMCDFNp6enubzbd26NdnxGjVqqGTJksk+r1ixoiTp7t276YoFAAAAAAAAAAAgr8uWPQ1NiZqIiIgsO2epUqVS/NzZ2VkRERF68OBBsmP29vY6fvy4zp07pytXrujy5cs6d+6ceflUSTKmkKouUqRIihVxBQsWNL83Vdf5+/vLaDTKyclJnp6eKcZYo0aNZJ/5+/tLSkys9u3bN8V+UmICMyEhQRcvXlStWrVSbfdvHh4eSZYCvX//vnbs2KFJkyYpODhYERERatOmTYrLx2Y2NlO/hysM/61GjRpau3btI+O/cuWKOemc0vj9+3wpVYym9syY/ttmtHoTAAAAAAAAAID8KoHSvzwvW5KGpuUgr1+/rrt376pw4cKP7BMWFqbo6GiVK1cuxeMPJ+zSY+XKlZo6dapu3LiR5PNy5crpueee09KlS1Pt+/BypqkxJRvDw8MlJSYvU5NSYs5U5WaqCHyUO3fuPLJNWhwdHdWhQwfVqVNHzz//vA4dOqQBAwZo8eLFyarxMhubaSycnJxSbZvePS4f3qMyrefHdL6Hl4s1Sc9/RwAAAAAAAAAAAGRT0rBt27b6/PPPFR8fr3379ql9+/aP7LNs2TJ99dVX8vLy0po1a+Tg4JDp669cuVLjxo2TJDVv3lzt27dX5cqV5ePjoyJFiig2NjbNpGFGmJZifTjJ9W8pJbRM/apXr64VK1ZkSSzpUapUKU2dOlUDBgzQ5cuXNXr0aC1cuNC8D+PjxJbZsUjJw0nYu3fvqnjx4im2M+2fmVbSFgAAAAAAAAAAPJ4ECg3zvGzZ09DT01O1a9eWJM2fPz/FJUAfFhMTY07iVaxY8bEShpL0v//9T5LUvXt3fffdd+rdu7fq1aunIkWKSEqsgMwqpr377t27p8uXL6fYxs/PL9lnpmrMgICAVJfJNBqN2rdvnwICAhQTE5NFEUsNGjTQoEGDJEn79+/XwoULsyQ2g8EgSTp79myq105pLFJSvnx5FSiQmNM+efJkqu1MxypUqJCu8wIAAAAAAAAAACC5bEkaStK7774rGxsbHTlyRF9//XWabadMmaKgoCDZ2tpq2LBhj33toKAgSYmVcilZvny5+f3j7mvn6elp3sPv4fM+bMmSJck+a9CggQoXLqyoqKhUq/nWrFmjgQMHqlOnTlma6JSkN99805xomzlzpq5cufLYsZkqSo8cOaILFy4k63Pr1i398ccf6YrPyclJjRs3liT9+OOPKba5cuWKed/GFi1apOu8AAAAAAAAAAAASC7bkoZ16tTRq6++KkmaMWOGRo8eLX9//yRtgoKCNGbMGP3www+SpOHDh6tmzZqPfe2KFStKSkzWhYSEmD+PjIzUrFmzNG/ePPNn9+/ff+zrvf7665ISqyqXLl1qrqyMjY3VrFmztG7dumR9nJyc9Morr0iSPv30U/32229KSEgwH9+yZYs++OADSVKnTp1Uvnz5x47zYQULFtRHH30kKbFKcsKECY8dW7Vq1dSuXTslJCRoxIgROn/+vLnP1atXNWzYsDSXLv23ESNGqECBAtq9e7fef//9JH39/Pw0ZMgQPXjwQFWqVFH37t0zPggAAAAAAAAAACBdEow5/0LOypY9DU1GjRqlokWLavLkyVq7dq3Wrl2rkiVLqnTp0rpz544CAwMlSfb29nrjjTc0ZMiQLLvusGHDdP78ebVt29a83GZgYKAePHggT09P2djY6PLly1lSwdeuXTsNHjxY3333nd5//33NnDlTZcqUUWBgoG7fvq3atWvr2LFjsrOzS9JvyJAhunLlipYuXap3331XkydPVrly5RQSEqLQ0FBJ0hNPPKFPP/30sWNMSZMmTfTss89q5cqV2rt3r3777Tf17NnzsWKbOHGiAgMD5e/vr65du6pSpUqys7OTv7+/ChQoIF9fX509ezbZWKSkbt26+vTTT/Xee+9p6dKl+v333+Xj46Po6GhdunRJUuKSqLNnz37sJW0BAAAAAAAAAADys2yrNDT5z3/+o/Xr1+vll19WzZo19eDBA50+fVo3b95U1apV9dJLL2n9+vVZljCUpNatW2v58uVq166dSpYsqYsXL+ratWsyGAwaPXq0Vq9eraefflqStH379iy55tixYzVnzhw1adJEDx48kJ+fnzw8PDRx4kS98847kiRHR8ckfWxsbDRx4kTNnz9f7du3V4ECBXTmzBlFRUWpTp06eu+997Rw4UIVKlQoS2JMyTvvvKNixYpJkr788kvdvHnzsWIrVqyYli5dquHDh6tixYq6fPmyrl69qtatW2vp0qWqU6eOJKX7nrp3767Vq1erV69eKlGihPz9/RUeHq569eppwoQJWr58uTw9PbNuQAAAAAAAAAAAQDJUGuZ9NkbTWprINn/++adeffVVeXl5adOmTZYOx6JGjhypTZs2adSoUXrttdcsHc5jiwkPs3QIVsPBLTE5HRrKmKWXu3vimMVER1s4Euvi4OQkSYrNwHLI+Z29i4skKfpulIUjsR5OhZ0lSbF8faabt8Gg4OBglS5dRnv2HbZ0OFajYoXSkqTfWrW1cCTWo+efWyVJB49fsnAk1qN+rcSVWfieln72///zxp2IuxaOxLq4Fi0sSYqN4ue09PL2rcL8mUGmufN+FD/bppejc+LPtvzumTHm3z0Zt3QzzZ88a+lnes5iIsItHIl1cSjqZukQ8qz+n23L8Wv+9G6bHL9mfpbtlYb5QdeuXdW7d2+dOnUqxeM7duyQlLjnX162Z88etW3bVm+88UaKx+/du6cDBw5IyvtjAQAAAAAAAABAXmI05vwLOYukYRbw8vLS0aNH9cUXX5j3+5OkuLg4LVmyREuWLJGNjY369OljwSizX9WqVRUSEqKNGzdq/vz5iouLMx+7deuWxowZo/DwcFWsWFFNmjSxYKQAAAAAAAAAAAB4WAFLB5AXjB49WocOHdKBAwfUpk0blS9fXo6OjgoODlZERIRsbW319ttvq2HDhpYONVsVK1ZMY8aM0eeff64vv/xS8+bNk4eHh+7du6crV64oNjZWpUuX1vTp02Vvb2/pcAEAAAAAAAAAQDrl1z0GT506pfnz5+vAgQOKiIiQm5ub6tWrp4EDB6pevXpZdp24uDj16tVLp06d0o4dO1S6dOk02wcGBmrevHnau3evbty4IVdXV1WvXl39+vVTy5YtMxUDScMs4O3trfXr1+uXX37Rli1bFBwcrHv37qlkyZJq1aqVXnzxRdWuXdvSYeaIQYMG6YknntAPP/ygEydO6OLFi7K3t1elSpXUrl079e3bV25urCkNAAAAAAAAAAByt40bN2r06NFJVlYMDQ3Vxo0btXnzZo0ePVqDBw/OkmtNnTo11W3w/u3QoUN6+eWXde/ePfNnt27d0s6dO7Vz5071799f7733XoZjIGmYRdzc3DRs2DANGzbM0qFYXM2aNTVlyhRLhwEAAAAAAAAAAJApJ0+e1NixYxUXF6f69evrrbfekre3twICAjR9+nTt379fU6ZMUaVKldSqVavHutbMmTO1YMGCdLW9evWqhg0bpnv37qly5coaN26cqlWrpuvXr+t///ufNm7cqJ9++kne3t7q27dvhuJgT0MAAAAAAAAAAACkKcFozPGXJc2YMUMxMTGqXLmyFixYoCeeeELFihVTvXr1NH/+fDVo0EBGo1FffvmlEhISMnWNyMhIjRw5UnPmzEl3n3nz5ikiIkIlSpTQjz/+qGbNmqlYsWKqVq2aZsyYoa5du0qSZs2apaioqAzFQ9IQAAAAAAAAAAAA+H8XLlzQzp07JUlDhw5VwYIFkxy3t7fXmDFjzG0PHz6cofMnJCRo2bJl6tixozZt2iRJql69+iP73blzRytWrJAkDRw4UMWKFUvW5u2335adnZ3Cw8O1ZcuWDMVF0hAAAAAAAAAAAABpSjDm/MtSdu3aJUmys7NTixYtUmxTu3ZtFS9eXJK0devWDJ3/7Nmzeu+993Tjxg25ublp+vTp6tev3yP77d+/Xw8ePJAktW7dOsU2pUqVUtWqVTMVF0lDAAAAAAAAAAAA4P+dOXNGkuTp6anChQun2MbGxkZVqlSRJJ06dSrD13BwcFD//v21fv16derUKUNxOTg4qFKlSqm2q1atWqbiKpCh1gAAAAAAAAAAAMh3LFn5l9OCg4MlSR4eHmm2K1OmjCQpKCgoQ+cvV66c/vzzT3OlYkbjKlu2rGxsbB4Z17Vr15SQkCBb2/TVEJI0BAAAAAAAAAAAQK7Ttm3bNI9ndPnN9AoPD5ckubq6ptnOVIV4586dDJ0/terFrI4rPj5ekZGRj2xvQtIQAAAAAAAAAAAAabK2SsNZs2Zp9uzZGeozYsQIvf766+Z9Ax0dHdNsX7BgQUkyt89u6Y3r4eMZiY2kIQAAAAAAAAAAAHKd7KokfBQ7OzuLXPdRsjsukobAY3BwK2bpEKyOuztjllEOTk6WDsEq2bu4WDoEq+NU2NnSIVgde74+M8zOzlYVK5S2dBhWp+eflvklyZrVr+Vt6RCsDt/TMs61aOaWFMrv7J35OS2jmD8zztGZn20zit89M4f5M+N41jLOoaibpUMArFLfvn3VuXPnDPVxc0v8eitUqJCkR1fppbfyL6ukN6779++b32ckNpKGAAAAAAAAAAAASJPRypYnLVasmIoVy1wRi2lPwLt376bZzrSXoSnZmN1MexOmNy57e3u5ZKC4gqQh8BjOL19i6RCsRqXnekuSjk3/ysKRWI/ab74lSYq+G2XhSKyLqVou4uwZC0diPYr6VpUkxUam/cMG/mHvkviD45Wtmy0cifUxxsXp9nl/S4dhNYpUqixJOvDhBAtHYj0afvixJGn9jlMWjsR6dG5ZXZLkv/RXC0diPSr3ekESc2dGmebP6/v2WjgS62M0GnX3Ns9behQukvic3b8RYuFIrIdjyVKSpJjoaAtHYl1M1XKxkZEWjsR6mFYEun/zhoUjsR6OJUpKksJOn7RwJNalWLUalg4BeYC3t7cOHDiga9eupdnu+vXrkqQyZcrkRFjy8vKSpHTHVbp0adnY2KT7/LaZjgwAAAAAAAAAAAD5QoIx51+WUrly4h/xBgYGKjqVP6wxGo06cyaxcKFatWo5Gte9e/cUEBCQartTpxL/kDajcZE0BAAAAAAAAAAAAP5fy5YtJUmxsbHatWtXim2OHj2qsLAwSVLz5s1zJK5GjRqZ9yjctm1bim2uX79uTmZmNC6ShgAAAAAAAAAAAEhTfqo0LF++vOrWrStJmjlzpqKikm4hFRsbq6lTp0qSDAaDmjRpkiNxOTs7q127dpKk+fPnKzQ0NFmbyZMnKyEhQcWKFdPTTz+dofOTNAQAAAAAAAAAAAAeMn78eNnY2Oj8+fPq37+/9u3bp/DwcB09elSDBw/W33//LRsbG40aNSrZvoHHjx9Xx44d1bFjR/38889ZGtdbb72lQoUK6ebNm+rbt6+2bdumsLAw+fn56c0339TatWslSSNGjDBXJaZXgSyNFAAAAAAAAAAAAHmOJSv/LKF27dr66KOP9OGHH+rUqVMaOHBgsjbjxo1TmzZtkn1+7949Xbp0SZIUHh6epXF5eHho+vTpGjlypC5fvqyhQ4cmazNgwAD17ds3w+cmaQgAAAAAAAAAAAD8S+/evVWjRg0tWLBABw4cUHh4uJydnVW3bl0NHDgwx5Yl/bdWrVpp7dq1mjdvnvbu3avQ0FA5OjqqevXqevHFF9WhQ4dMnZekIQAAAAAAAAAAAJCC6tWrm/cvTK9GjRrp7NmzGerTo0cP9ejRI93ty5cvr08++SRD13gUkoYAAAAAAAAAAABIU0KCpSNAdrO1dAAAAAAAAAAAAAAALItKQwAAAAAAAAAAAKQpwWjpCJDdrKrSMDIyUj///LNeeuklPfnkk6pevbrq1q2rbt266fPPP9elS5dS7Ld//375+vrK19dXcXFxORx13nDkyBF98MEH6tq1q+rWrasaNWqoWbNmevXVV7Vq1SrFx8en2K9///7y9fXVtGnTzJ8FBQWZ/3sEBgY+dmzjxo2Tr6+vxowZ89jnAgAAAAAAAAAAyI+sptJw+/btGj9+vMLDwyVJRYsWlcFg0O3bt3X+/HmdPXtWixYt0ogRI/Taa69ZONq84+7du5owYYLWr18vSbK3t1eZMmXk7OysK1eu6M8//9Sff/6pBQsWaNasWapQoYKFIwYAAAAAAAAAAFmNSsO8zyqShgsWLNCkSZMkSZ06ddLw4cNVuXJl8/HQ0FB9/fXXWrx4saZNm6b79+/rzTfftFC0ecetW7fUu3dvXblyRSVLltTw4cPVs2dPOTg4SJKMRqO2bNmiL7/8UmfPnlXfvn21fPlylS5dOs3zlipVypyELFu27GPH+dZbb2nIkCEqXLjwY58LAAAAAAAAAAAgP8r1y5MePHhQU6ZMkSQNHz5c06dPT5IwlCR3d3d98MEHGjZsmCTpf//7n06ePJnjseYlRqNR77zzjq5cuSIPDw/98ssv6tOnjzlhKEk2NjZq3769Fi1apOLFi+vGjRv66KOPHnlue3t7+fj4yMfHR/b29o8dq7u7u3x8fOTu7v7Y5wIAAAAAAAAAAMklGHP+hZyVq5OGRqNREyZMUHx8vOrUqaORI0em2X7o0KEqU6aMEhIS9P333+dQlHnThg0btGvXLknShx9+KE9Pz1Tburu76/XXX5ckbdu2TQEBATkRIgAAAAAAAAAAALJIrl6e9NChQ7pw4YIkaciQIY9s7+DgoM8++0ySVKdOnRTbxMXFadGiRVq1apUuXrwoBwcH+fr66sUXX1Tnzp1T7HPmzBktXrxYBw8eVEhIiGJiYuTq6qrq1aurZ8+e6tixY5L2K1as0Pjx49W5c2d9+umnmjdvnjZs2KCrV6/KyclJdevW1eDBg1W/fv0Ur7d3714tXLhQp0+f1p07d+Tt7a1evXqpT58+qlq1qiTp7Nmzyfr9/fff+umnn3T48GFFRETI1dVVderUUf/+/dWkSZNHjt/DlixZIkmqUqWKWrRo8cj23bp1k729vZ544gl5eXml2TYoKEht27aVJG3evNm8D+KsWbM0e/ZsDRkyRC+99JLmzp2rbdu2KTQ0VK6urmrUqJFee+01+fr6JjnfuHHjtHLlSj399NPmqlQTo9Go33//XatWrdK5c+d0+/ZtlSxZUo0bN9arr776yFgBAAAAAAAAAADyg1ydNNy7d68kyc7OTo0bN05Xn6ZNm6Z5fPDgwfr7779VokQJVaxYUYGBgfr777/1999/6+LFixoxYkSS9osXL9bEiROVkJCgIkWKqEKFCrp//76CgoK0c+dO7dy5U6+99ppGjRqV7Fp37txR7969de7cObm7u6tSpUo6f/68tm/frp07d2ru3Llq1apVkj5z5szRzJkzJUklSpRQ5cqVFRAQoI8//lj79u1L9b6mTJmib7/9VpJUpEgRGQwGhYaGauvWrdq6dauGDBmiMWPGpGcIFRMTo0OHDkl69HiaODs767nnnktX20e5evWqunfvrtDQUJUtW1Y+Pj46d+6c1q9fr+3bt2vRokWqXr36I88TFRWlkSNHavfu3ZIS9080GAy6dOmSVqxYoY0bN+rnn39O17kAAAAAAAAAAMjP4lkuNM/L1cuTXrx4UZLk4eEhFxeXLDnniRMnNGnSJO3atUsrV67Url271K5dO0nSN998o8jISHPbgIAAffbZZ0pISNCbb76pPXv2aOXKlealOzt16iRJmj9/vm7fvp3sWrt371Z4eLjmz59vvt7WrVvl6+ur+Ph4TZs2LVn7mTNnytbWVu+995527dql3377TXv27FG/fv20efPmFO/p119/1bfffitXV1dNnjxZBw4c0IoVK7Rr1y5NmzZNTk5O+vbbb7Vs2bJ0jdGVK1cUGxsrKbHSMKetW7dOTk5OWrZsmbZt26bVq1dr3bp1Kl26tO7du6c5c+ak6zyTJk3S7t275ebmpu+//17bt2/XihUrtHPnTrVr107R0dEaPny44uPjs/mOAAAAAAAAAAAAcrdcnTQ0JeKKFSuWZeccOnSounfvLlvbxFt3dnbWhAkTJEmxsbE6evSoue2ePXtkZ2en6tWra+jQobK3tzcfK1q0qN555x1zv0uXLqV4vQkTJqhZs2bmf7u7u5urGf38/BQVFWU+NmPGDEnSoEGD1L9/f3OMjo6Oev/991NcJjQmJkazZs2SJH322Wfq1q2b+ZiNjY06d+6ssWPHSkpc/jMuLu6RYxQREWF+n5VjnxFTp05VzZo1zf+uWLGiBg0aJEk6fPjwI/uHhoZq6dKlkhKThw9XTBYuXFiTJ0+Wq6urrl27Zq5oBQAAAAAAAAAAKUsw5vwLOStXJw0LFSokSeaqt6zw1FNPJfusVKlSKlq0qCQpLCzM/Hnfvn117NgxLV68OMVzOTo6mt/fu3cv2XE7O7sUE30+Pj7m96bKxpCQEJ04cUKS9OKLL6Z4vYEDByb77MiRI7p586acnZ3N+wT+W7du3WRra6uQkBCdPn06xTYPc3JyMr9PT5Ixq7m7u6e4ZGjFihUlSXfv3n3kOXbs2CGj0aiyZcuqZcuWyY47OTnp119/1e7du9W8efPHDxoAAAAAAAAAAMCK5eo9DUuWLCkpaeXb4ypVqlSKnzs7OysiIkIPHjxIdsze3l7Hjx/XuXPndOXKFV2+fFnnzp0zL58qSUZj8pR3kSJFkiQWTQoWLGh+b0rK+fv7y2g0ysnJSZ6eninGWKNGjWSf+fv7S0pMrPbt2zfFflJiAjMhIUEXL15UrVq1Um0n/TPukhQeHp5m2+yQ2n8j01imJ5EZGBgoSfL19U21zcPJWwAAAAAAAAAAkDoq//K+XJ009Pb2liRdv35dd+/eVeHChR/ZJywsTNHR0SpXrlyKxx9O2KXHypUrNXXqVN24cSPJ5+XKldNzzz1nXgIzJQ8vZ5oaU7LRlJxzdnZOtW1K+zqaqu5iYmLStWznnTt3HtmmWLFiKlq0qCIiIsxJyfQ4d+6cKlSokOEx/rf0jNujmBLND1dNAgAAAAAAAAAAIGW5OmnYtm1bff7554qPj9e+ffvUvn37R/ZZtmyZvvrqK3l5eWnNmjVycHDI9PVXrlypcePGSZKaN2+u9u3bq3LlyvLx8VGRIkUUGxubZtIwI0xLsZqWK03Jw/sf/rtf9erVtWLFiiyJxdbWVq1atdKqVau0Z8+edPWJiYnRCy+8oNjYWL377rvq06dPlsSSWaZxSWnMAAAAAAAAAABAxlBpmPfl6j0NPT09Vbt2bUnS/PnzU1wC9GExMTHmJF7FihUfK2EoSf/73/8kSd27d9d3332n3r17q169eipSpIikxArIrGJaRvPevXu6fPlyim38/PySfWaqxgwICEh12U6j0ah9+/YpICBAMTEx6YqnS5cukqSzZ89q165dj2y/evVqRUVFKSYmxvzfzJK8vLwkKc1KyZkzZ+rll1/WypUrcygqAAAAAAAAAACA3ClXJw0l6d1335WNjY2OHDmir7/+Os22U6ZMUVBQkGxtbTVs2LDHvnZQUJCkxCq+lCxfvtz8Pj377KXF09NTVapUSXbehy1ZsiTZZw0aNFDhwoUVFRWVaqXhmjVrNHDgQHXq1Cndic4WLVqoefPmkqSPPvpIwcHBqba9cuWKpkyZIklq3bq1qlWrlq5rZKcWLVrI1tZWwcHB+uuvv5Idv3//vpYvX67du3c/MhkNAAAAAAAAAEB+F2/M+RdyVq5PGtapU0evvvqqJGnGjBkaPXp0suqxoKAgjRkzRj/88IMkafjw4apZs+ZjX7tixYqSEpN1ISEh5s8jIyM1a9YszZs3z/zZ/fv3H/t6r7/+uqTEqsqlS5eak1mxsbGaNWuW1q1bl6yPk5OTXnnlFUnSp59+qt9++00JCQnm41u2bNEHH3wgSerUqZPKly+f7ng++OADlSpVSleuXFHv3r21fPlyPXjwwHw8Li5O69at0wsvvKCIiAi5u7tr4sSJGb/xbODp6amnn35akvT222/ryJEj5mO3b9/W22+/rZCQEHl4eKhz586WChMAAAAAAAAAACBXyNV7GpqMGjVKRYsW1eTJk7V27VqtXbtWJUuWVOnSpXXnzh0FBgZKkuzt7fXGG29oyJAhWXbdYcOG6fz582rbtq15KdDAwEA9ePBAnp6esrGx0eXLl7NkqdJ27dpp8ODB+u677/T+++9r5syZKlOmjAIDA3X79m3Vrl1bx44dk52dXZJ+Q4YM0ZUrV7R06VK9++67mjx5ssqVK6eQkBCFhoZKkp544gl9+umnGYrH09NTS5Ys0auvvqqzZ8/qv//9rz7++GOVK1dOjo6OCggIMO8ZWL16dc2YMUMlS5Z87HHIKhMmTNC1a9d04MABvfDCC6pQoYKcnJx06dIl3b9/X0WLFtXMmTPl6Oho6VABAAAAAAAAAAAsyiqShpL0n//8R61bt9bSpUt14MABBQYG6vTp03J0dFTVqlXVpEkT9enTJ0OVdI/SunVrLV++XHPnztXp06d18eJFFSpUSAaDQU899ZT69u2r+fPna86cOdq+fbsGDBjw2NccO3as6tatq59//lmnTp2Sn5+fKlWqpDFjxsjHx0cvvvhisiSXjY2NJk6cqA4dOujXX3/V0aNHdebMGRUsWFB16tRR165d1bt370zt8VimTBmtWLFC69at0x9//KHTp08rKChICQkJcnNzU8OGDdW1a1d16tQpWTLT0lxcXPT9999rxYoVWrVqlc6dO6erV6+qVKlSat26tYYMGaJSpUpZOkwAAAAAAAAAAHK9BJYLzfNsjGzoZjX+/PNPvfrqq/Ly8tKmTZssHQ4knV+efJ9JpKzSc70lScemf2XhSKxH7TffkiRF342ycCTWxamwsyQp4uwZC0diPYr6VpUkxUbetXAk1sPepbAk6crWzRaOxHo0GfiSgoODVbZUKZ3evcvS4ViNIpUqS5IOfDjBwpFYj4YffixJWr/jlIUjsR6dWybu4e6/9FcLR2I9Kvd6QRJzZ0aZ5s/r+/ZaOBLrUf+5XonzZ9my8jvtZ+lwrELhIonP2f0bIY9oCRPHkol/UB0THW3hSKyLg5OTJCk2MtLCkVgPexcXSdL9mzcsHIn1cCyRuKpb2OmTFo7EuhSrVsPSIeRZDUZtzfFr/j2tbY5fMz/L9Xsa5iemisBTp1L+Pzh27NghSapWrVpOhgUAAAAAAAAAAPK5eGPOv5CzSBrmIl5eXjp69Ki++OIL816EkhQXF6clS5ZoyZIlsrGxUZ8+fSwYJQAAAAAAAAAAAPIaq9nTMD8YPXq0Dh06pAMHDqhNmzYqX768HB0dFRwcrIiICNna2urtt99Ww4YNLR0qAAAAAAAAAADIR9jTMO8jaZiLeHt7a/369frll1+0ZcsWBQcH6969eypZsqRatWqlF198UbVr17Z0mAAAAAAAAAAAAMhjSBrmMm5ubho2bJiGDRtm6VAAAAAAAAAAAAAkscdgfsCehgAAAAAAAAAAAEA+R9IQAAAAAAAAAAAAyOdYnhQAAAAAAAAAAABpYnnSvI9KQwAAAAAAAAAAACCfo9IQAAAAAAAAAAAAaYpPsLF0CMhmVBoCAAAAAAAAAAAA+RyVhgAAAAAAAAAAAEgTexrmfTZGo5H/zAAAAMhW5cqVU3BwsDw8PBQUFGTpcAAAsArMnwAAIDepNGxbjl/z/Nw2OX7N/IykIQAAAAAAAAAAAJDPsachAAAAAAAAAAAAkM+RNAQAAAAAAAAAAADyOZKGAAAAAAAAAAAAQD5H0hAAAAAAAAAAAADI50gaAgAAAAAAAAAAAPkcSUMAAAAAAAAAAAAgnyNpCAAAAAAAAAAAAORzJA0BAAAAAAAAAACAfI6kIQAAAAAAAAAAAJDPkTQEAAAAAAAAAAAA8jmShgAAAAAAAAAAAEA+R9IQAAAAAAAAAAAAyOdIGgIAAAAAAAAAAAD5XAFLBwDkN9evX9eCBQu0a9cuXb16VZJUrlw5tWrVSi+99JKKFy+erM+9e/f03Xffad26dQoKCpKzs7Nq1KihAQMGqGXLlqleKzw8XF9//bW2bt2qkJAQubq6ql69eho8eLDq1KmTYp+goCC1bds2zXuoUqWKVq9enf6bzgK5fdwkKSEhQb/99ptWrVolf39/RUdHy8PDQ23atNFrr72mIkWKPPY4ZERuHbP0PGMPGzFihF5//fV0t38cuXXMHubv76958+Zp3759Cg8Pl4uLi2rXrq1BgwapSZMmjz0GGWUNY3bw4EF9//33Onz4sKKiouTh4aFWrVpp0KBBKlWq1GOPQWbk5Lj929SpUzVv3jx9//33atq0aartMjve2SW3P2u5cf7M7WMmMXfmhblTyr3j9jDmT+ufP5k7Mye3P2vMn8yf+XX+zK1j9rDcNndK1jFuzJ//sOb5E8jPbIxGo9HSQQD5xcGDBzV06FDduXNHdnZ2Kl++vBISEhQUFKT4+HiVLFlS3333napUqWLuEx0drUGDBunYsWOyt7dX5cqVFRERYZ7oX3/9dY0YMSLZtW7evKk+ffro8uXLKlSokCpWrKiQkBDdvHlTdnZ2+vjjj/Xcc88l67dlyxYNHz5cRYsWVcWKFVO8Dy8vL33++edZNCqPZg3jFh0draFDh2rfvn2SEsdIki5fvqyEhAR5eHho8eLFKl26dDaMUHK5ecxu3LihkSNHphl/SEiIgoODJUmTJ09Wt27dsmJY0pSbx8xkx44dev311/XgwQMVKlRIFSpU0PXr1xURESFJeuutt/Tqq69mzwClwBrGbN68efrqq69kNBrl5OSkihUr6vr167p586ZcXV01e/ZsNWrUKPsGKQU5OW7/tn37do0YMUJxcXFp/uKW2fHOLtbwrOW2+dMaxoy50/rnTil3j5sJ86f1z5/MnZljDc8a8yfzZ36cP3PzmJnktrlTso5xY/78hzXPn0C+ZwSQI27fvm1s3Lix0WAwGF9++WVjSEiI+djly5eNL7zwgtFgMBjbtGljvH//vvnYO++8YzQYDMZnnnnGePXqVfPnK1euNFarVs1oMBiMe/bsSXa9fv36GQ0Gg/E///mPMSIiwmg0Go3x8fHG//3vf0aDwWCsXr268fz588n6zZo1y2gwGIwTJkzIytvPNGsZt7FjxxoNBoOxWbNmxmPHjpk/9/PzMz711FNGg8FgHDJkSJaMyaNYy5il5u7du8Z27doZDQaD8YMPPsjECGScNYzZ7du3jQ0bNjQaDAbj66+/brxz547RaDQa4+LijDNmzDAaDAajwWAw/v3331k6NqmxhjHbunWreVzee+89Y1RUlLnfd999ZzQYDMZatWoZL1++nKVjk5acHreHrVu3zlijRg3zmKTVPqu/rh+HNTxrRmPumj+tZcyYO6177jQarWPcmD+tf/5k7swca3jWjEbmT+bP/Dd/WsOY5ba50xRTbh835s9/WPP8CcBoJGkI5JDvv//e/IP93bt3kx2/deuWsUGDBkaDwWBctWqV0Wg0GgMDA41Vq1Y1VqlSxejv75+sz7Rp04wGg8H4wgsvJPl83759RoPBYKxTp47x1q1byfqNGTPGaDAYjKNHj052bMSIEUaDwWD8+eefM3urWcoaxu3YsWNGg8FgrFq1qtHPzy9Zv7/++stoMBiMvr6+xuvXr2fo/jPDGsYsLW+99Zb5B9QHDx6ku9/jsIYxW7VqldFgMBjr169vjI6OTtavb9++RoPBYBw3blyG7j2zrGHMunfvbjQYDMaXXnopxXsYNWqU0WAwGIcNG5bu+35cOTluJrdv3zZ+8MEH5l/YHvWLW3Z8XT8Oa3jWjMbcNX9aw5gxd1r/3Gk0Wse4MX9a//zJ3Jk51vCsGY3Mn8yf+W/+tIYxy21zp9FoHePG/Jk35k8ARqOtpSsdgfxi//79kqTWrVvLxcUl2fFixYqpbt26kqQTJ05IklavXq34+HjVqVNHlSpVStanT58+kqTDhw+blwmQpJUrV0qS2rZtq2LFiiXr98ILL0iStm7dqvv37yc55ufnJ0kyGAwZu8FsYg3jZurXvXt3+fr6JuvXqFEjvfnmm3rvvfdka5v933atYcxSs337dq1du1YFChTQ559/LgcHh0f2yQrWMGbXr1+XJJUvX16FChVK1q9mzZqSpGvXrj3qdrNEbh+zGzdu6PTp05KkwYMHp3gPAwcOlJT43JmW2cluOTluknTkyBG1b99ev/zyixwdHfXZZ589Msas/rp+XLn9WTPJTfOnNYwZc6f1z52SdYwb86f1z5/MnZmT2581E+ZP5s9/y+vzpzWMWW6bO6XcP27Mn3ln/gQgkTQEcsjQoUM1adIk9ezZM9U2xv/fYjQhIUGSdPToUUnSE088kWL7UqVKycPDQ5J04MAB8+dHjhxJs1+tWrVUoEABRUdH6+TJk+bPo6KidOXKFUm545c2yTrGbe/evZKkp556KsV+NjY2Gjp0qPr166eSJUumeh9ZxRrGLCWxsbH64osvJEn9+/dX1apV02yflaxhzMqUKSNJCgwMVHR0dLJ+Z8+elSTzNbNbbh+zh3+BqV69eor9vL29JUnx8fGPfD6zSk6OmyRdunRJERERatGihX7//fc0r2uSlV/XWSG3P2tS7ps/rWHMmDutf+6UrGPcmD+tf/5k7syc3P6sScyfafVj/sy786c1jFlumzul3D9uzJ95Z/4EIBWwdABAflGrVi3VqlUr1eNhYWHmCdf01zyBgYGSEv+6KzUeHh4KDg5WQECAJJk3NE6rn729vUqVKqXg4GBdunRJ9evXl5T4g5/RaJS7u7vCwsK0YMECnT59WvHx8fLy8lKXLl1SncizS24ft3v37uny5cvm60dGRur333/Xvn37dOfOHZUtW1adOnVS8+bNM3bjjyG3j1lqFi1apICAABUtWlTDhw9P/QazgTWMWbt27eTu7q7Q0FC9++67+uSTT+Ti4iKj0ajvv/9ee/bskb29vfr27Zuxm8+k3D5mNjY25uMFCqT8405cXJz5fXBwcKoxZaWcGjeTypUra/Hixen+3p3VX9dZIbc/a1Lumz9z+5gxd+aNuVOyjnFj/rT++ZO5M3Ny+7MmMX+m1Y/5M+/On9YwZrlt7pRy/7gxf+ad+RMASUMg1/j000917949FSpUSB06dJAk3bp1S5JSLNE3KVq0qCQpPDxcknT79m3zDyKP6hccHGzuJ/2zNMydO3fUpUsXxcfHm4/t2bNHixYtUs+ePfXRRx/J3t4+E3eZ9Sw9bteuXTP/Rdb169fVv3//ZEs0/Pbbb+rcubMmTZqUo0uGpcbSY5aS2NhYzZ8/X5I0YMAAFS5cOP03lANyw5g5OTlp4cKFGjt2rDZs2KAdO3aoQoUKCg0N1a1bt+Tl5aUPP/xQ1apVe6x7zSqWHrNy5cqZj505cybFX1z8/f3N72/fvp2e28p2WTVuJqalg9IrK7+uc4qlnzXJ+uZPS48Zc2f+mDul3DFuzJ95f/5k7swcSz9rEvNnevoxf+a/+TM3jJm1zZ2S5ceN+TP/zJ9AfsDypEAuMHfuXK1du1aSNGzYMBUvXlySzOt1p/XDfsGCBZO0fXiN74z0k/75pe3Bgwd6/vnntW7dOp04cULbt2/Xm2++KXt7e/3222/69NNPM3yP2SE3jFtUVJT52Ouvvy4bGxvNmTNHR48e1b59+zRhwgQ5Ojpq/fr1uWLccsOYpWTdunUKDQ2Vi4uL+vfvn55byTG5acwcHR1Vp04d2dnZKTo6WmfOnEnyQ/3Df91oSblhzB7en+Hrr782L7vysG+++cb8PjY29hF3lf2yctwyKyu/rnNCbnjWJOuaP3PDmDF3KsP9UpKb504pd40b82fenT+ZOzMnNzxrEvNnRvsxfyrD/VKSm+fP3DRm1jJ3Srlj3Jg/M84a508gvyBpCFjY7NmzNWPGDEmJG/8OGTLEfMzOzi7d5zH9wJbRzc4f/kGvfv366tWrlyZMmKCPPvpIlSpVkoODg8qWLauhQ4eaNzH+9ddfk/yFlCXklnF78OCB+bOYmBj99NNPateunQoVKiQ3Nzf17dtXEyZMkCQtXbpUFy9ezNB1slJuGbOU/Pjjj5KkF198Ua6urhk6b3bKTWPm5+en5557TosWLVKHDh20du1anThxQlu3btWQIUN09OhRvfTSS+Yf/C0lN43ZqFGjZGtrq127dmn48OE6d+6cYmNjFRAQoLffflv79u1TkSJFJMnif72e1eOWWVn5dZ3dctOzZi3zZ24ZM+bO9LPGuVPKXePG/JmxfpL1zJ/MnZmTm5415s+M9WP+TD9rnD9z05hZy9wp5a5xY/7MGGubP4H8hKQhYCFxcXGaMGGCZs2aJUlq1qyZpk2blmQSdHJykpT4C0FqTL84ODo6SpKcnZ3NxzLST5K6deumiRMn6sUXX0yxT7du3eTl5SWj0aitW7emeX/ZJbeN28Pj16NHjxQ3Ajd9npCQoO3bt6d9g9kgt43Zv12+fFmnTp2SJD377LOPvJ+ckBvH7OOPP1ZYWJhatmypadOmqXLlynJwcFC5cuU0ZswYvf3224qPj9eHH36oO3fuZPSWH1tuHLNGjRrps88+k729vbZu3aqnn35aNWrUUIcOHbRx40ZNnDhRXl5eya6Tk7Jr3DIrq76us1NufNZy+/yZ28aMuVMZ7vdvuXHulHLnuDF/5r35k7kzc3Ljs8b8yfyZmvw0f+bGMcvtc6eUO8eN+TNjrGX+BPIjkoaABURGRuqVV17RkiVLJEmdO3fW119/bS65N3Fzc5OUfJ3wh5mOmZYScHJyMpf1R0REpLtfelWtWlWSzJsV56TcOG4P/2WiaWz+zcbGxryp9JUrV1I9d3bIjWP2b3/88YckqXr16qpYseKjbinb5cYxu3Hjhg4dOiRJGjFiRIp9BgwYoKJFi+ru3bvasWNHmveY1XLjmJk8++yz2rBhg1555RU1a9ZMLVu21GuvvaZ169bpueee082bNyVJpUqVSu/tZpnsHLfMyu455HHl5mftUSw1f+bGMWPuzHtzp5Q7x435M+/Nn8ydmZObn7VHYf5k/pTy7vyZG8cst8+dUu4cNxPmz/SzhvkTyK9IGgI57Pr16+rTp4/27NkjSXr55Zf11Vdfpbh+t+mH2ODg4FTPZzpm+mslW1tbeXt7S0r9F6vY2FiFhoYm6ffwsYc3oP8308brBQoUSLVNdsit4+bh4WH+ISutv4wyLfGQk5vR59Yx+zfTXw137dr1EXeU/XLrmF29ejXZdf/Nzs7ukefODrl1zB7m6emp0aNHa/78+Zo3b55GjRolT09PhYeHm6/n6+ubjrvNOtk9bpmVVV/X2SG3P2u5cf7MrWPG3Jm35k4p944b82femj+ZOzMntz9rzJ/MnynJD/Nnbh2z3Dx3Srl33B7G/Jk+uX3+BPIzkoZADgoJCVH//v117tw52dnZ6cMPP9Tbb7+d6rrctWvXliQdPXo0xePXr183/0Bn2nA5Pf2OHz+uuLg4FSxYUNWqVZMk3b59Ww0bNlSNGjXSXPrlzJkzkmT+68WckJvHzc7OTjVr1pQkHTt2LNV7MO0nUb58+VTbZKXcPGYPi4mJ0fHjxyVJTZs2Tde9ZZfcPGYuLi7m46YfmFNi2pj+4fbZKTePmSStX79e8+bNU2BgYIr9tm3bJkkqV66cPD09U7/RLJZT45ZZj/t1nR1y87OWW+fP3DxmzJ15Z+6Ucve4MX/mnfmTuTNzcvOzxvzJ/Jmf58/cPGa5de6Ucve4ScyfmZFb508gvyNpCOSQmJgYDR06VJcvX5a9vb1mzJihPn36pNmnY8eOkqQDBw6kuIn5L7/8Iklq2LChypUrZ/68U6dOkqRNmzalWOJv6te5c2fzmuBFihRRiRIlJEkrVqxIMZ6NGzea42/fvn2asWeV3D5ukvT0009LShyfa9euJeu3Y8cOBQQEyNbWNkfGzRrGzOTs2bOKjY2Vo6Njjiai/y23j1nFihXl7u4uSVq6dGmK8Rw4cECXL1+WJDVu3DjN2LNCbh8zSVqyZImmTp2q3377LcX458+fL0nq379/mnFnpZwct8x63K/rrJbbn7XcOH/m9jGTmDsl6587pdw/bsyfeWP+ZO7MnNz+rDF/Mn/m1/kzt49Zbpw7pdw/bhLzZ2bkxvkTgCQjgBwxe/Zso8FgMBoMBuPSpUvT3e+tt94yGgwGY+fOnY0BAQHmz1etWmWsVq2a0WAwGPfs2ZOkT0JCgvGFF14wGgwG44svvmi8ceOG0Wg0GuPj443z5s0zGgwGY/Xq1Y3nz59P0m/lypXmGKdMmWJ88OCB+diGDRuMderUMRoMBuO0adMyMQKZ83/t3XdUFFfDBvAHqQIGo7FFjViyoNiwoBhUiu2NibFhx9hiO5ZXjYo9gtijAvauINherFgTuzQVVOyVImoEAVFBWGW+P/hmssvuUhYQ0ed3jueEmXtn7tydzJ07t5WEfEtPTxc6d+4syGQyoXPnzsL9+/elfZGRkUKbNm0EmUwmzJ49W5ssyLeSkGeinTt3CjKZTOjatasWV1p4SkKeiXllaWkprF+/Xun/z5CQEKFVq1aCTCYTJkyYoE0W5FtJyLP9+/cLMplMaNCggdIx4+PjheHDhwsymUxo166dUl4WtY+Zb5qI59cUvqD/Xxe2knCvfWrlZ0nIM5adJb/sFISSkW8sP0t++cmyUzsl4V5j+cny80ssP0tCnn1qZacglIx8Y/mpqiSWn0QkCDqCIAjF3XBJ9LnLyMiAnZ0dXr16BT09PTRs2DDH8OJCyUDWgr8DBw6UphKQyWRISUmR5hCfMGGCFFZRbGwsBgwYgOfPn8PAwADff/89Xrx4gfj4eOjo6GDx4sXo0qWLSrxFixZh8+bNAAATExOYm5sjISEB//zzDwDA2dkZbm5uKFWq6Acql6R8i42NxdChQxEdHQ0dHR3UqlULOjo6ePDgAQDA1tYWq1evhrGxcUGzJUclKc8AwMvLC6tWrYKtrS22bt1asIvXUknKsyVLlmDjxo0AgDJlyqBGjRpK6yK0bNkSa9as4X32/wRBwH//+18cO3YMQNYUTcbGxnj48CHkcjmqVq2K7du3F0oPybwojnxTR1w/Y8uWLRqnZirI/9eFqaTca8CnU36WpDxj2Vlyy06gZOUby8+SW36y7NROSbnXAJafLD+/rPKzJOXZp1J2AiUn31h+qipp5ScRZWGjIdFHcOPGDfTo0SPP4bt164aFCxdKf6empmLTpk04evQoYmNjoaenh/r168PFxQUdOnTQeJzExESsXbsWp06dwvPnz1G6dGk0atQIw4YNy3EKiaCgIPj6+uLq1at49eoVzMzM0LBhQ/Tp0wf29vZ5vo6CKmn5lpaWhm3btuHYsWOIjo6Grq4uatWqha5du6JXr17Q09PL87Voq6Tl2R9//AF/f3907NgRXl5eeU53YSppeRYaGoodO3YgPDwcSUlJMDExgYWFBbp27YquXbtCV1c3z9eirZKUZ5mZmdizZw/27t2L+/fv4/379/juu+/Qvn17DBkyBGZmZnm/8AIqrnzLLi8VN0D7e7QwlaR7Dfg0ys+SlmcsO0tm2QmUvHxj+Vkyy0+WndopSfcawPKT5eeXU36WtDz7FMpOoGTlG8tPZSWt/CSiLGw0JCIiIiIiIiIiIiIiIvrCFf38gkRERERERERERERERET0SWOjIREREREREREREREREdEXjo2GRERERERERERERERERF84NhoSERERERERERERERERfeHYaEhERERERERERERERET0hWOjIRERfZE+fPhQ3EkgIqLPEMsXIiIi+hJkZmYWdxKIiKgIsNGQiEiBo6MjLCwsUL9+fdy/fz9PcVxcXGBhYQEXF5ciTt2nxdXVFRYWFnB0dCzupORLfHw8JkyYgCtXrhR3UgosICAAFhYWsLCwwJMnT4o7OUT0hRKfQ66ursWdlHwR0+3t7V1ox7xx4wacnZ1VtoeGhkrnCw0NLbTz0aejKO4nIvq8iXXPnP5ZWVnBxsYGv/zyC+bMmYNHjx4Vd7K1Jl7v5/acPHjwIJydndG0aVM0atQI7dq1Q1BQUHEnq0hFR0dj8ODBePr0aXEnhYiIigAbDYmI1JDL5XB1deVogc9MYmIifvzxRxw5cgSCIBR3coiI6DNy7tw59OrVCzdv3izupBAR0Wfi/fv3ePXqFe7cuYOdO3eiS5cuCAgIKO5k0f/bu3cvJk+ejOvXr+PNmzd49+4dYmNjUbFixeJOWpG5c+cOfv7558++YZSI6EumV9wJICL6VN24cQMbN27EiBEjijspVEhSU1ORkpJS3MkgIqJPwHfffQcAMDMzK5TjJSQkaOxsZGRkJJ3PyMioUM5Hn5bCvp+I6MvRtGlTbNiwQe2+jIwMxMXF4ciRI9i2bRvkcjlmz56NunXrom7duh85pZTd//73PwDAt99+i+XLl8Pc3BypqamoXLlyMaes6Lx69Qrp6enFnQwiIipCbDQkIsrBypUr4eTkhDp16hR3UoiIiKgQnTx58qOdq1GjRh/1fPTx8fclIm3p6urCxMRE7T4TExN8/fXXqF+/PmQyGaZOnQq5XI7NmzdjyZIlHzmllF1CQgIAoFOnTmjcuDEAoGzZssWXICIiokLA6UmJiNSoXbs2dHR0kJGRgWnTpnGaUiIiIiIiIio2Xbt2lUawXb58uZhTQwCk7wTGxsbFnBIiIqLCw5GGRERqNGzYEK1atYKPjw+uX7+OTZs2Yfjw4VofLyIiAjt27MDly5fx8uVLlC5dGjKZDJ07d0bPnj2hr6+vEsfR0RFxcXHo1q0bFi5cqPa4rq6u2LdvH6pWrYpTp05J2wMCAjBt2jRUqlQJx48fx4IFC3D06FHI5XJUr14d06dPh62tLQAgMzMTx44dw/HjxxEZGYnExER8+PABZcuWhZWVFX755Rd06tQJOjo6Wl+/Im9vb6xcuRI2Njbw8fHB+fPn4evrK60DUalSJdjb2+O3335DpUqVNB7nwYMH2LZtG4KDg/HixQsYGBjA3NwcHTt2xIABA1C6dGml8BYWFkp/Dxw4EADQrVs3yOVyHD58GNWrV8dff/2lcq6IiAj06dMHADB8+HBMmjRJJYyHhwe2b98OKysrlXVGLl++DH9/f4SHhyM+Ph7GxsaoWbMm2rVrh/79+6utZLq4uCAsLAwjR45Ehw4d4Obmhlu3bsHY2Bj169fHmjVrNOZN9jSJ1ztjxoxc4xARFYfY2Fhs374dFy9exNOnT6Gjo4Nvv/0WP/zwAwYPHowqVapojJucnAxfX1+cPHkSsbGx0NfXR6NGjTB06FC0aNECDRo0QEZGBrZv344WLVpI8cRyYcyYMRg7dqzSMR8+fIjt27cjJCQET58+hZ6eHipWrIjmzZujf//+SlPChYaGSmVK9mMvWLAA3bt3VwqTPR2iBw8eYOfOnQgKCpLyoGbNmujUqRNcXFxUyrWciGVtkyZNsGrVKsydOxfnzp2Djo4OzM3NsWDBAqVyUZv3FNGNGzewefNmXL9+Hf/88w8qVKgAR0dHjB49GsHBwZg4caLKe8rHSl9+fkdF165dk8734sULGBoaokqVKmjZsiUGDhwoTUWqKKf7CdDuHhff58T8u379OjZv3ozLly8jOTkZ5cuXR6tWrTBs2DDUrl1b429ERJ+PSpUq4fnz54iPj1e7PyMjAwcOHMDff/+NW7duITk5GaVKlcLXX3+Nhg0bwtnZGXZ2dirxxHqlWPcMDAzErl27cPfuXaSlpaFq1apo3749hg4dqnEaZrFe6e/vj4cPHyItLQ21atVCnz590LNnz1yvLSEhAT4+Pjh79iyio6Px4cMHVK5cWXr2qpv9R7F8vXv3Lm7fvo0NGzYgLCwMKSkpqFy5Mjp16oSRI0fC2NgYHz58gJ+fHwICAvD48WPo6uqifv36GDlypFQ/zovsdcuVK1di5cqVAP4t+wtyXfmpzwP5rxuLnj17hu3bt+P8+fOIjY0FAJQvXx5NmjRBr169YGNjI4V98uQJnJyclOKLf2sq+4iIqGRioyERkQaTJk3C2bNnERMTA29vbzg5OeX7g0xmZiYWL16MLVu2KG3PyMjApUuXcOnSJezevRtr167NsYFMW4IgYMKECTh9+rS07cGDB6hZsyYAIDExESNHjsS1a9dU4r548QIvXrzA6dOnc2y4LIjly5dj7dq1SttiY2Ph4+ODgwcPwtfXFzKZTCXeli1bsGTJEqURoOnp6YiMjERkZCT8/f2xfv36PE8r27ZtWxw+iLjD5AAAI2RJREFUfBixsbF48uQJqlWrprQ/ODhY+u9Lly6pPca5c+cAAA4ODtK2zMxM/PHHH9i1a5dS2FevXuHq1au4evUq/Pz8sGbNGlhaWqo9bmxsLH799Ve8fv0aQNa9AwAGBgY5XtPy5culBsMBAwawwZCIPll79uyBu7u7yvo4Dx48kBrSPDw88PPPP6vEffjwIYYMGYLnz58rbT979izOnTuHKVOm5Ds9Z86cwdixY6XnLZD17I2KikJUVBT27t2LmTNnYsCAAfk+tibbt2/HokWL8P79e6XtN2/exM2bN7Fv3z5s3bo13+8KGRkZGDZsGG7evClti4mJQY0aNQAU/D3Fx8cHHh4eEARB2hYXFwcfHx8cPXoUvXv3Lrb0afs77tq1C3PmzFG6Jrlcjvv37+P+/fvYuXMnVqxYgXbt2uV4bYoKco+Ldu7cCTc3N6V3n+fPnyMgIACHDx/G+vXr8/XBm4hKnoyMDERHRwMAKlasqLI/NjYWw4YNQ1RUlMq+tLQ0PH36FMeOHcuxgUcQBEyZMgUHDhxQ2v7o0SOsW7cOhw8fhr+/v8ozNy0tTaXeCWR1LJk5cyaCg4ORmZmp8drOnDmDyZMnq6w/Hx0djejoaOzduxeTJk3C0KFDNR7jyJEjmDJlCuRyuVL8devWISwsDJs3b8aoUaMQEhKiFC8kJARhYWHw9vbO17M9Lwp6XbnV5wHt68aRkZEYMmSIStri4uIQFxeHQ4cOYejQoVq9SxERUcnG6UmJiDQoXbo05s+fX6BpSr28vKQPXR06dICfnx9CQ0Px119/wdXVFaamprh16xZGjBhRJIuJi41+/fv3x+nTp3Hs2DHMnz9fmtZm2rRpuHbtGnR1dTFmzBgcOnQIISEhOHr0KDw8PKRe7/v27dPYWKatyMhIrF27FtbW1ti0aRNCQkJw7NgxuLi4AMhqWJs3b55KvD179mDhwoX48OEDbGxssHnzZgQHB+PMmTOYN28eKlSogLi4OAwdOhSJiYlSvPDwcAQGBkp/r1+/HuHh4XBzc0Pr1q1RqlRWkajYQChS3Hbjxg2kpqYq7Y+NjZUq54qNhvPnz5caDG1tbaXenydPnsSkSZNgbGyMp0+fYtCgQSofvEWBgYEQBAGenp4ICgrC1q1bMXr06BzzduPGjVJjbL9+/TBr1qwcwxMRFZfjx49j1qxZSE9Px3fffYc///wT58+fx4ULF7B06VJUq1YN6enpmDx5Ms6ePasUNzU1FUOHDsXz589hbGyM6dOnS42Fc+fOhZmZGRYtWqTUaJSb1NRUTJ06FRkZGWjYsCE2b96Mc+fO4fz581i9ejXMzc0hCAIWLlyIJ0+eAACaNWuG8PBwzJ07VzpOeHg4wsPD0aVLl1zPefjwYXh4eOD9+/eoW7cu1qxZg6CgIJw4cQLjxo2Dnp4eHj16hAkTJuT5OkQ3btzAzZs3MX78eJw/fx4HDhyAu7s7jIyMABTsPeXvv//GvHnzIAgCrKyssGXLFoSEhODgwYNwdnZGQkICVq9eXSzp0+Z3BLI+lLq7u0MQBLRu3Ro7duzAxYsXcfr0aSxZsgQVKlSAXC7HjBkz8ObNmzz9BgW5x0Xx8fFwc3NDzZo14e3tjaCgIPz9998YO3YsdHV1kZGRgdmzZys1dBLR52fbtm1ITk4GAJURXx8+fMCYMWMQFRUFY2NjTJs2DUePHkVISAgOHz6MadOmSWvtrV69WhpVlt3x48dx4MAB2Nvbw8/PDyEhIThw4AA6d+4MIOs5uWLFCpV4c+fOlRq2evXqJdUrt27disaNGyMwMBDPnj1Te86IiAiMHTsWKSkp+Oabb+Dm5obTp08jODhY6lz54cMHLF68GP7+/hrzx9XVFZUqVYKnpycuXLiA/fv3SyPlIiIi0L17d4SFhWHYsGE4evQogoODsWzZMpiZmSEzMzNfa0SK5fy3334LABgxYoRK2V8Y15VbfV7burHYOJySkgJzc3OsWrUKp0+fluqbDRs2BABs2rQJ4eHhAICqVasiPDwc69evl44TGBiI8PBwjBgxIs95R0REJYBAREQSBwcHQSaTCVOnTpW2ubu7CzKZTJDJZML69etV4gwYMECQyWTCgAEDlLY/fvxYsLS0FGQymeDu7q72fNeuXRPq1q0ryGQyYfPmzbmmJbupU6cKMplMcHBwUNr+v//9T0pz37591ca9f/++FGb16tVqw9y+fVsK8+eff+bp3Lnx8vKSjtmrVy8hIyNDJcz48eMFmUwmWFhYCC9fvpS2p6SkCNbW1oJMJhNGjx4tfPjwQSVubGys0KRJE7X5HhsbK507JCREaV/v3r0FmUwmTJgwQWl7WlqaUL9+fcHCwkJo2rSpIJPJhAsXLiiF8fX1FWQymWBnZydkZmYKgiAId+7cESwsLASZTCaMHz9ebVrDw8MFKysrQSaTCePGjVPaJ95XMplM2L17t7qsVPqdY2NjBUEQhB07dkjbZs2aJaWHiKioiM+cnMorddLT04XWrVsLMplM6NChg5CUlKQS5uXLl4KTk5Mgk8mENm3aCOnp6dI+T09PQSaTCXXr1hXCwsJU4t6+fVto0KCBxue+uN3Ly0vadvr0aWn7s2fPVI4ZExMjle3Zy23FZ3J2ISEhatPx7t07wdbWVpDJZEL37t2F1NRUlbhbtmyR4l66dEllvzqKZe2kSZPUhinIe4pcLhfatWsnyGQy4ZdffhHevn2rEvfPP/+U0pD9XaGo06ft7yiW540bN1a610ShoaHScY8fP660T939VNB7XPGecnJyEl6/fq0Sf/HixVKYyMhItflERJ8msb7Xt29f4c2bNyr/Xr9+LTx9+lS4cOGC4OrqKtUt7OzshISEBKVjnTlzRnoWHDhwQO35Tp48KYXZuXOn0j6xbieTyYSxY8eqjd+zZ09BJpMJTZo0UapjREZGSmlbtGiRSrx3794JvXr1UvucFARB6N69uyCTyYQWLVpIdRpFqampgrOzsyCTyYRGjRop1Q8Vy1dra2vh+fPnSnGfP38u1KtXTwqzbt06leMr1p/i4uLUXrsm4m+Y/ZoKel15qc8XpG587969HN8tXr16JTRv3lxtGayY5+qui4iISj6ONCQiysWkSZNQvXp1AFlr8Dx8+DBP8Xbt2oXMzEyULl1a4+iAhg0b4scffwQA7N69u3ASnE2nTp3Ubv/w4QOGDBmCjh07om/fvmrDWFpa4quvvgIApZ6JhWXYsGFq1yGyt7cHkNUDMi4uTtp+8OBBvH37FkBWT1JxdKCiatWqSVONBQQEqEz1pol4zpCQEKWe+leuXEFGRgYsLS3RrFkzAEBYWJhSXHFqUnt7e2ntxz179kAQBBgYGGD27Nlq02ptbY1+/foBAE6cOIGXL1+qhNHR0UHHjh3zdA0HDhyAm5sbAMDZ2Rlz584ttLUoiYgK29mzZ/HPP/8AyHqmiyMgFJUrVw6urq4AsqZiVJyea//+/QCAn376Cc2bN1eJa2lpif79++crTYqjEtWtF1W9enWsX78ee/bsydP6TLkJDg6Wnv1Tp05Vu+ZQ3759IZPJYGdnJ5WB+fGf//xH7faCvKdcuXIFMTExUrrVrc07duxYVKhQoVjSp+3vKMZ7//49kpKSVOLZ2Nhg1apV2Ldvn9o1wbIr6D2uaMCAATA1NVXZrjjDgeKoSSIqOa5cuYImTZqo/GvatCns7e0xZMgQBAQEQBAE1KtXDz4+PihfvrzSMUxMTDBw4EB07txZei5mp7iebk51u5EjR6rdLtaX3rx5o/SMFGdGMTU1xbhx41TiGRoaYtq0aWqPeevWLdy4cQNA1rp42ZeJALJmABLrOGlpaVL5n12XLl1Upk2tVKmSVJc3NDRUWYMYyKqTicRndkEV5nVpqs8XpG6sWE4mJCSoxPvqq6+watUq7Ny5M9dZboiI6PPDRkMiolwoTlOanp6OGTNm5Lgeg0hsWKpVqxYA4O3bt2r/NWrUCEDWOhHqPlAVVL169dRut7CwwNSpU+Hl5aX2I9bbt29x8eJFqfKR36lZ80Kc9iQ7xUrwu3fvpP8ODQ0FAHz99dcoV66cxjwVj/v27VvcuXMnT2kRK8EvX77E3bt3pe3i1KQ2NjbScRUbDTMyMqR0KX64E6dztbGxQbly5TSeV6zUZ2Zm4sqVKyr7q1WrJjXc5uSvv/7C9OnTIQgCrK2t4e7uzgZDIvqkic/S0qVLo02bNhrDOTg4SI1p4rP10aNHUqeS7FO0KcprpwtR48aNpc4sgwcPxtKlS3HlyhWlMrB169Zo2LAhypQpk69jqyOWMaampmobPoGsj5yHDh3Cpk2b0LZt23yfQ9N7QEHeUy5cuCClu2XLlmqPr6+vD0dHx2JJn7a/o9g5KCMjAz169MCqVatw69Ytpc5E7dq1Q7169dQ2lGq6Bm3u8ew0vTMpvmMovjMR0eejatWq6Nu3LzZs2ICAgACYm5urhGnWrBlmzJiBZcuWQU9PT2X/q1evlOowmup2hoaGGtdaz62OZmNjI00vnV3jxo3VdiRRTFNOZbalpaW0jl/2DpwisTzQlO5atWqpTZ9ih4z8TGmek8K8Lk3lZEHqxnXq1JG+AUyZMgVubm4ICgpSuv7mzZvD2to6x7osERF9nlTfJIiISIWNjQ369+8PX19fREREYMuWLTkuwg7829v75s2baNKkSZ7O8/z5c3z99dcFTq+ivBzv5s2biIiIQFRUlLQ+X0xMjFLjqFAE6+RoqoAYGBhI/62YBjFPk5KS8pWn9evXzzWcpaUlqlSpgmfPniE4OFiqLIsfdFu0aCFVKCMjI/Hu3TsYGRkhLCwMaWlpMDIyQqtWrZTOCwC1a9fO8byK+58+faqyP6+VtIULF0q/0Y0bN3D37l2NFX4iok+B+Jw0NzeHrq6uxnC6urowNzfH7du3peek4ih0dR9PRWKDU15VrFgRkyZNwsKFC/H69Wts2LABGzZswFdffQVbW1u0bdsWTk5OajvbaEMc0VCjRo0i6+ihqRwpyHuKmP+5pTsv+V8U6dP2d2zQoAFcXFzg4+OD+Ph4eHl5wcvLC+XLl8cPP/wABwcH2Nvb56nBUEwPoN09nl1+35mIqOSwsbGBj4+P9HdGRgbu3bsHT09PnDt3Di9evEDZsmVhZ2eXa1mRmZmJiIgIREZGSnW7x48f4+nTp0r1OU11OzMzM7Uj1gDNzxtxrcIaNWrkmLaaNWuqjP4Wn5NfffVVrqPTa9eujcePH2tcG1FTvVe8Hk2dfYqi/P0Y11WQurGhoSHmzJmD33//Henp6dixYwd27NgBY2NjNG/eHG3btkX79u1RsWLFPB2XiIg+LxxpSESUR4rTlHp6euLRo0c5hn/z5k2+z6FNnNwYGhpq3Hfnzh10794d3bt3h7u7O3x8fHDmzBlERUWhfPny6Nq1K8zMzAo9TSJ1U5PmpKjzVBwFIDYUvnr1Crdu3UKpUqWknpaGhoaQy+WIiIgA8O/UpLa2tko9V8Xz5vZhUXEqutTUVJX9Of1+igRBgKOjI8qUKQO5XI6ZM2cWyehQIqLCktfnJPDvs1J8TiYnJ6vsUyevjTuKBg8ejG3btqFNmzZSOZWSkoLjx49j+vTpsLOzw6JFiwrlGfvq1SsA0DgyozBoKkcKUqaK+Z9buk1MTHI9ZlGkD9D+d5w5cya8vLzQvHlz6UPzy5cvcfDgQUyYMAF2dnbYtGlTvtKjzT2enbqRQ0T0eTIwMED9+vWxbt06/PTTT5DL5VizZg3mzp2bY7yQkBB06tQJ/fr1w4IFC+Dv748LFy4gLi4OVatWRZ8+fXI9d37rZwDw+vVrALmXCeoa7QrzOZnT+8DHVpjXVVTl5I8//og9e/agU6dO0m+XmpqKs2fPws3NDfb29pg2bZrGdBER0eeLNQ8iojwyNjaGh4cHfv31V6Snp2P69Onw8/PTGN7IyAhv3rzBjz/+iOXLlxdJmgoyDdWTJ08wYMAAvH79Gvr6+mjXrh0aN26MOnXq4Pvvv5fWg2jTpo30UbO4iZWZRo0aFckakPb29ti1axcuXboEuVyOsLAwZGZmwsrKSpoitHHjxggNDUVoaChsbW1x/vx5AMpTkwJZ90tKSkqulSzF9am0+bgt+s9//oOlS5di165dcHNzQ2RkJLZv347BgwdrfUwioqIkPvPy8jFKfFaKH9Zy63AhSktL0yptLVu2RMuWLfHmzRsEBQUhODgYFy5cQExMDORyOTZv3gxBEKS16LQlXkdxTCtZkPcUxY+LOSnIh8bCeI/S9nfs2LEjOnbsiMTERFy8eBFBQUG4ePEi/vnnH7x9+xaLFy+GgYEBXFxccjx/Qe5xIqJSpUph/vz5uHv3Lu7fv4+dO3eidu3aatflu379OoYNGwa5XA4TExO0b98eDRo0QJ06dSCTyVCuXDm8f/8eO3fuLPR0mpmZISEhIdcyV93Un5/rc/JjXFdh1I2trKzg6emJd+/eITQ0FMHBwbh48SLu3buHDx8+ICAgAG/evIG3t7dWxyciopKJIw2JiPKhRYsW6Nu3LwAgIiICW7du1Rj222+/BaA8hZo6mqaGEXu3y+VyjXEVR1rk17p16/D69Wvo6urCz88PK1aswKBBg2BnZyc1GAqC8Mk0GAIFz9Pc2NrawtDQEKmpqbh27Zq0TkSLFi2kMOLaTWFhYYiNjZVGnIprImZP68OHD3M854MHD1TiaOP333+Hnp4e+vbtiwYNGgDIGhEbGxur9TGJiIqS+MyLiorKcdTe+/fvERUVBSBrXScA+O6776T94j51ctqXF6ampujQoQPmzJmDkydPYs+ePVIa/Pz88P79+wIdv0qVKgCQ67N6165d2LBhgzQSvjAUpEwVp6CLjo7OscyNjo4ulvRlp+3vWK5cOfz8889YsGABzp49i02bNkkzMChOJZjbNWhzjxMRAVmjzJYuXSqNAFyyZAnu3bunEm7FihWQy+UoU6YMDhw4gEWLFmHAgAFo2bKlNL2xuO5rYROfdbnNxKOurBPjpqSkqExdmp1YrypInelj+RjXVZjlpJGREdq2bQtXV1ccOnQIR44cgZWVFQDgxIkT0nTqRET0ZWCjIRFRPv3++++oVq0agKxGGU0v6c2aNQOQtRaPuKaBOrNnz0aLFi3Qo0cPpelCxJ6GiYmJauO9f/8ekZGRWl0DAGl6zbp160qLo2cXHh4ujX74FNbJEfM0ISEB165d0xhu3bp1aNasGbp06YKYmBhpe27rVZQuXVpqIAwKCkJ4eDgA5UZD8b+vX7+OEydOAMjqoSk2tGZPa1hYmMbfEACOHTsmpa1x48Y5pi8vSpUqBTc3N+jq6iItLQ1z5swp8DGJiIqC+JxMS0uTpnpW58yZM0hPTwcAWFtbAwDq1KmD8uXLS/tzipsf4lRwYgeh7Bo2bCiN8EhPT1fqWKPNmkjiGkQpKSm4evWq2jCCIMDT0xNLly7F0aNH830OTQrynmJjYwMgawRFWFiYxnSfPXu2WNKn7e84b948dOzYERMnTlSJo6OjAzs7O3Tp0gUA8vQBtSD3OBGRyNLSEsOGDQOQNVpv9uzZKnUzsW7XqlUraUmN7BQ7nhRm3U5c4iEsLAwpKSlqw8TExKjtSCI+JwHg+PHjGs9x+/ZtKX5e1+8rTh/jugpSN967dy+6du0KR0dHtY2KtWvXxqhRo6S/Fcu8olqDmYiIPh1sNCQiyicTExN4eHhAR0cH796909ho2KtXLwBZjXtz585V28P82rVr2LdvH5KTk1G2bFmYmppK+8zNzQEAly9fVts7cdOmTQVaA1FXVxdAVs9EddOivXr1Cm5ubtLfOY14/Fi6du0KAwMDAIC7u7vaKXBiYmKwZcsWvH79GhkZGUqVZsX1gDRdjzhi8OTJk7hz5w50dXWVKn0NGzaEsbEx5HI5Nm7cCEB1alIAcHZ2BpBVsXdzc1NbMb9+/Tr8/f0BAG3bti20hebr1auHAQMGAAAuXryIgICAQjkuEVFhcnBwQIUKFQAACxcuVDuyPSkpCYsWLQKQNerL0dERQFYHiZ49ewIADh06hOvXr6vEjY2NxbZt2/KVJj09Pdy/fx8RERFSx5Hsbt++DSBr9Jo4egP4t1wF1E/Bpo6TkxPKli0LAFi6dKnassnHxwcvX74EAHTu3DlPx82LgryntG7dWholuXTpUqnBS9HWrVvx5MmTYkmftr9jZmYmoqKicPr0abUftwVBwN27dwEoj3bVpCD3OBGRotGjR0v1w4iICKkOIRLLoMePH6utdzx79gxLly6V/i7Mul2XLl2gr6+Pd+/eYcGCBSr7MzMzsWDBArWNU/Xr15dGtK1cuVJt3frdu3fSeo6GhoZS541P2ce4roLUjU1NTXH79m3ExcUhMDBQ7fHFcrJUqVJSp2lA+X3nU/hGQEREhY+NhkREWmjZsmWuC8nXrVsX/fr1AwCcOnUKAwcOxIULF5CYmIiYmBj4+Pjgt99+g1wuh6GhISZPnqwUv3379gCyPjwOGzYMwcHBSExMxO3btzF79mwsW7ZMmiJLG3Z2dgCyPlaNGjUKERERSExMRFRUFPz8/NCtWzfcuXNHCq+49l5x+eabbzB+/HgAQGRkJHr16oUTJ04gISEBT58+xb59++Di4oLk5GTo6OhgxowZSj0hFfPr2LFjSE5OVvmA17ZtWwCQ1nGwsrJSaszV19eXeoGKIwjVNRpaWlpKax0dPXoUQ4cORWhoKJKSkhAbG4tNmzZh0KBBkMvlMDMzwx9//FEIOfSvcePGSaMfFy1aJH1wJiIqKlFRUdizZ0+u/+7fvw8AMDAwwOzZs6W4zs7OCAwMRHx8POLj43HkyBH06tVL6hU/f/58pfV+hg8fjkqVKuH9+/cYMmQIduzYgefPnyM+Ph779+9Hv3798Pr1ayl8XnrG9+jRA2XLloUgCBg9ejR8fX3x6NEjJCYm4s6dO3B3d8f+/fsBAH379lU6ptj4B2Q1ZKakpORadhoZGWHKlCkAgEuXLuHXX39FcHAwkpKS8ODBAyxfvhyLFy8GkFXWKI58L6iCvKfo6elh2rRpALI6wLi4uEjpfvjwIRYsWCA1hBVH+rT9HV1cXGBgYIDU1FQMHjwY+/btQ0xMDBITE3Ht2jVMmjRJGlkppi0nBb3HiYhEBgYGSvWFZcuWKY3+Eut29+7dw++//47bt29Lz+SNGzeia9euSuELs25Xo0YN/PbbbwCAgIAAjBkzBpGRkUhOTsbVq1cxfPhwnDp1SqmxSdGcOXOgr6+PpKQk9O7dG7t378azZ8+QmJiIM2fOoF+/ftJIyqlTp6Jy5cqFlvaiVNTXVZC6sZOTk9QIPXPmTKxevRr379+X3j+8vb2xbt06AECnTp2UOkkpvu8EBgYiJSWlQJ2ZiYjo06OXexAiIlJn8uTJOHfuXI5rCEyfPh0ZGRnYu3cvLl++jKFDh6qEMTExwbJly2Bpaam0vXPnzjh8+DDOnj2LO3fuYNCgQUr7ra2t8csvv2jd2DR8+HCcPn0aDx8+RFBQEIKCglTCWFtbo0yZMjh37lyB1iUqTEOHDsXbt2+xZs0a3Lt3D2PHjlUJo6+vjzlz5qB169ZK242MjNC4cWNcvXpV+nhtY2OjtC5RtWrVUKdOHWmtQXUfaFu2bIkLFy4AACpWrCj1Is3O1dUVGRkZ2LVrl8Y8rl69Ojw9PaURG4XF1NQUM2fOxNixY5GcnAx3d3esWLGiUM9BRKQoIiJC+viVk2nTpuH7778HAHTo0AHz5s3D3LlzER0drXZaSGNjY7i5ual00DA1NcX69esxaNAgJCUlwc3NTWmEfKlSpdC7d2/s2rULADR+rFRUtmxZrFixAqNHj0ZSUhLc3d3VhnNwcMC4ceOUttWvXx/GxsZITU3F9OnTMX36dIwZM0ZtOaWoR48eSEhIwPLly3HlyhWV8h4AmjZtqjRCpLAU5D2lY8eO+O9//wtPT09cu3ZNJd1Vq1ZF7dq1ce7cOaWR/h8jfdr+jjVr1oSHhwemT5+OuLg4uLq6qo3Xp08fjVOfZleQe5yISJGtrS26dOmCgwcP4s2bN5g3bx68vb0BZC2hIc5QExgYqHb0mIODAxISEhAZGVnodTuxzuHn54eTJ0/i5MmTSvvt7e0hl8tx8eJFlbiNGjXCypUrMXHiRMTHx2PWrFkqYfT09DBx4kT079+/UNNdlD7GdWlbN9bX14eXlxeGDBmChIQEeHp6wtPTU+01KL5bAVmNxFWqVMGzZ8/g7e0Nb29vdOvWDQsXLtTqGoiI6NPDkYZERFpSnKZUE319fXh4eMDX1xc//fQTqlatCgMDAxgZGeH777/HkCFDcOTIEWlKTEW6urpYs2YN3NzcYG1tDRMTExgbG8PKygrTpk2Dr68vjI2NtU6/mZkZdu/ejZEjR6J27dowMDCAvr4+KlSoADs7OyxatEhKN5DVQ/7evXtan6+w6OjoYPz48di/fz+cnZ1Ro0YNGBkZwcDAAObm5ujTpw8OHDggTQ+a3fLly+Ho6IgyZcrA0NAQqampKmHE0YbAv+s2KVJsSLS3t9d4D+jp6cHNzU3KxypVqkBfXx/ly5dH06ZNMXv2bBw4cEBjo2NBdejQQfoAefToUZw6dapIzkNEVBDOzs44evQoXFxcUKtWLZQuXRqmpqawtLTEqFGjcOTIEfz8889q41paWuLIkSMYMmQIzM3NYWhoCDMzMzg4OMDf3x9du3aVwhoaGuYpPba2tggMDMSgQYMgk8lgbGwMfX19VKxYEY6OjvDy8sLatWulKcFE5cqVw9q1a9GoUSMYGRnB1NRU7XSU6owYMQL79u1D9+7dpXcFExMTNGnSBHPnzoWPj4/SqPfCUpD3FAAYNWoUfHx80L59e3zzzTfQ19dH1apVMWTIEOzbt08aNZHXvC/M9Gn7O3bp0gX79u1D7969UbNmTRgZGUFfXx9VqlRB586dsXXrVmk6ubwqyD1ORKTI1dVVmj3lxIkT+PvvvwFkdXwUR5Z999130NfXh4GBAapUqQInJyesWrUKa9eulaZAvnz5co7rrudXqVKlMGfOHGzevBn29vaoUKECDA0NIZPJMHXqVKxevTrHerO9vT1OnDiB4cOHw8LCAsbGxjA2NkadOnUwcOBAHDx4UG3HkU9dUV9XQerGFhYWOHz4MEaPHg0rKyuYmJhAT08P5cuXxw8//AAPDw/4+/ujTJkySvH09PSwdu1atGjRQrqeT2FWIiIiKjw6grpJxYmIiIiIiEq4c+fOSVOmnTlzptBHdVPOJkyYgCNHjqBly5b5Xl+SiIiIiIiIPj5OT0pERERERCXOrFmzYGpqig4dOsDa2lptmBs3bgDImv5RXOeVCu7OnTvYsGEDatasiUGDBqkdBSkIAm7dugUAqFWr1sdOIhEREREREWmBjYZERERERFTiREVFISwsDJGRkfD19VXZn5iYCH9/fwBAq1atUKoUV2YoLKampjh8+DAAoHLlyujZs6dKmEOHDiEqKgoA8MMPP3zM5BEREREREZGWWHMmIiIiIqISR1wD7tKlSxg/fjwiIiKQmJiI2NhYBAYGon///njx4gUMDQ0xceLEYk7t56VatWrS6M758+dj48aNePjwIZKSknD37l14enpixowZALLWBnZycirO5BIREREREVEecU1DIiIiIiIqcTIzMzF16lQcPHhQY5gyZcpg8eLFcHR0/Igp+zJER0dj0KBBePr0qcYw1tbW8Pb2RoUKFT5iyoiIiIiIiEhbbDQkIiIiIqIS69SpU9i7dy+uX7+O5ORkmJiYoEqVKnBwcEDv3r1RuXLl4k7iZ+vNmzfw8/PDX3/9hYcPHyI9PR3ly5dH7dq10aVLF3Tu3Bn6+vrFnUwiIiIiIiLKIzYaEhEREREREREREREREX3huKYhERERERERERERERER0ReOjYZEREREREREREREREREXzg2GhIRERERERERERERERF94dhoSERERERERERERERERPSFY6MhERERERERERERERER0ReOjYZEREREREREREREREREXzg2GhIRERERERERERERERF94dhoSERERERERERERERERPSF+z8mTgWnagwd8gAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 2000x700 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(20,7))\n",
    "#fig, ax = plt.subplots(figsize=(11.69,8.27)) \"A4\"\n",
    "sb.set(font_scale=1.8)\n",
    "x_axis_labels = [2005,2006,2007,2008,2009,2010]*3\n",
    "y_axis_labels =['MSM','MSMW', 'MSW', 'Oth/Unk/Missing','Northeast', 'Southeast', 'Southwest', 'West', 'Midwest','Prev. Region', 'Prev. Clinic','Change Region', 'Change Clinic']\n",
    "\n",
    "ax.tick_params(axis='both', which='major', labelsize=18)\n",
    "\n",
    "sns.heatmap(imporances_all_models,fmt=\"\",cmap='vlag',linewidths=0.30,ax=ax, xticklabels = x_axis_labels, yticklabels = y_axis_labels, vmin =-abs(imporances_all_models.max()).max(), vmax = abs(imporances_all_models.max()).max())\n",
    "ax.text((ax.get_xlim()[1])/18 * 1.5, (ax.get_ylim()[0]) + (ax.get_ylim()[0])*0.10, \"Neural network\", fontsize=20)\n",
    "ax.text((ax.get_xlim()[1])/18 * 7.5, (ax.get_ylim()[0]) + (ax.get_ylim()[0])*0.10, \"Logistic regression\", fontsize=20)\n",
    "ax.text((ax.get_xlim()[1])/18 * 13.5 , (ax.get_ylim()[0]) + (ax.get_ylim()[0])*0.10, \"Random forest\", fontsize=20)\n",
    "ax.axvline((ax.get_xlim()[1])/18 * 6, color = \"black\", linewidth = 2)\n",
    "ax.axvline((ax.get_xlim()[1])/18 * 12, color = \"black\", linewidth = 2)\n",
    "##c_bar = ax.collections[0].colorbar\n",
    "#c_bar.set_ticklabels(size = 14)\n",
    "#print(abs(imporances_all_models.max()).max())\n",
    "plt.tight_layout()\n",
    "#plt.savefig(\"Graphs/Magnitude_of_feature_importance_2005_2010_all_models.png\", dpi = 600)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{2005: array([ 5.59116809e-03, -1.38888889e-03, -1.34140551e-03, -9.82905983e-03,\n",
      "        2.17711301e-02,  7.12250712e-05, -3.52564103e-03,  1.02552232e-01]), 2006: array([ 0.02365578, -0.00087708,  0.02249905,  0.07006483, -0.00677514,\n",
      "        0.        ]), 2007: array([ 0.0301029 , -0.00455907,  0.13523512, -0.00678651, -0.00358213,\n",
      "        0.0024098 ,  0.00192784]), 2008: array([-9.43142010e-05,  6.29749394e-02,  5.99568849e-03,  2.81460523e-02,\n",
      "        4.04203719e-04, -6.19779035e-04,  7.36998114e-03,  4.98517920e-04,\n",
      "        9.97035839e-04, -1.10482350e-03]), 2009: array([ 0.03197796, -0.00208552,  0.02736097,  0.00893232]), 2010: array([ 0.01817938,  0.01899598, -0.00694779,  0.0214324 ,  0.0001071 ,\n",
      "        0.00230254, -0.00161981, -0.00115127, -0.00149933,  0.00037483])}\n",
      "{2005: array([ 0.17540361,  0.02877493,  0.01003086,  0.02688746,  0.05232669,\n",
      "       -0.00205366,  0.01862536,  0.09057455]), 2006: array([0.03256642, 0.11785941, 0.05896784, 0.06630228, 0.14319309,\n",
      "       0.0168298 ]), 2007: array([ 0.02293865,  0.04827407,  0.15304155,  0.02141462, -0.00704702,\n",
      "       -0.00076853,  0.00569233]), 2008: array([0.02918351, 0.1100512 , 0.05278901, 0.04353274, 0.07592293,\n",
      "       0.00774724, 0.0404069 , 0.0402048 , 0.01783886, 0.0069523 ]), 2009: array([ 0.03283054, -0.0043809 ,  0.05473505,  0.0201469 ]), 2010: array([ 0.04423025,  0.05951807,  0.00095047,  0.02097724,  0.00637216,\n",
      "       -0.0003079 ,  0.05447122,  0.02563588,  0.00507363,  0.00506024])}\n",
      "{2005: array([ 0.00506697,  0.        ,  0.03299984,  0.00456673,  0.00269485,\n",
      "       -0.00932709,  0.        , -0.00921414, -0.010731  ,  0.0182669 ,\n",
      "        0.04471518, -0.0211554 ,  0.01929966]), 2006: array([ 3.67115637e-02,  2.80880420e-03,  1.15801577e-02,  0.00000000e+00,\n",
      "        4.23455979e-02,  0.00000000e+00,  6.35676741e-03, -3.12089356e-04,\n",
      "       -1.64257556e-05,  2.36859396e-02,  1.11990802e-01,  1.14520368e-01,\n",
      "        7.56570302e-02]), 2007: array([ 0.01557664,  0.00163089,  0.02268264,  0.00480945,  0.        ,\n",
      "       -0.00552505,  0.        , -0.01808953, -0.01296389, -0.0236645 ,\n",
      "        0.02822433, -0.0314029 , -0.04479947]), 2008: array([ 0.02054866,  0.0019046 ,  0.02295999,  0.00122313,  0.02378123,\n",
      "       -0.002621  , -0.002953  ,  0.05869299, -0.00679713,  0.00503233,\n",
      "        0.06978857, -0.0013105 ,  0.04868076]), 2009: array([-2.03058321e-02, -8.32147937e-03, -1.66251778e-02,  8.49928876e-03,\n",
      "        2.84495021e-04, -1.08463727e-03,  3.55618777e-05, -5.84992888e-03,\n",
      "       -3.25391181e-03, -3.84068279e-03,  9.15718350e-03, -9.60170697e-03,\n",
      "        4.49857752e-03]), 2010: array([ 0.0183512 ,  0.00059764,  0.02564598,  0.00950958,  0.        ,\n",
      "        0.00249604,  0.00606433,  0.01300756, -0.0026894 ,  0.00012304,\n",
      "        0.02664792,  0.00202144,  0.03132361])}\n"
     ]
    }
   ],
   "source": [
    "print(best_effect_by_year_rf)\n",
    "print(best_effect_by_year_nn)\n",
    "print(best_effect_by_year_lr)\n",
    "\n",
    "best_effect_by_year_rf={2005: array([ 5.59116809e-03, -1.38888889e-03, -1.34140551e-03, -9.82905983e-03,\n",
    "        2.17711301e-02,  7.12250712e-05, -3.52564103e-03,  1.02552232e-01]), 2006: array([ 0.02365578, -0.00087708,  0.02249905,  0.07006483, -0.00677514,\n",
    "        0.        ]), 2007: array([ 0.0301029 , -0.00455907,  0.13523512, -0.00678651, -0.00358213,\n",
    "        0.0024098 ,  0.00192784]), 2008: array([-9.43142010e-05,  6.29749394e-02,  5.99568849e-03,  2.81460523e-02,\n",
    "        4.04203719e-04, -6.19779035e-04,  7.36998114e-03,  4.98517920e-04,\n",
    "        9.97035839e-04, -1.10482350e-03]), 2009: array([ 0.03197796, -0.00208552,  0.02736097,  0.00893232]), 2010: array([ 0.01817938,  0.01899598, -0.00694779,  0.0214324 ,  0.0001071 ,\n",
    "        0.00230254, -0.00161981, -0.00115127, -0.00149933,  0.00037483])}\n",
    "best_effect_by_year_nn={2005: array([ 0.17540361,  0.02877493,  0.01003086,  0.02688746,  0.05232669,\n",
    "       -0.00205366,  0.01862536,  0.09057455]), 2006: array([0.03256642, 0.11785941, 0.05896784, 0.06630228, 0.14319309,\n",
    "       0.0168298 ]), 2007: array([ 0.02293865,  0.04827407,  0.15304155,  0.02141462, -0.00704702,\n",
    "       -0.00076853,  0.00569233]), 2008: array([0.02918351, 0.1100512 , 0.05278901, 0.04353274, 0.07592293,\n",
    "       0.00774724, 0.0404069 , 0.0402048 , 0.01783886, 0.0069523 ]), 2009: array([ 0.03283054, -0.0043809 ,  0.05473505,  0.0201469 ]), 2010: array([ 0.04423025,  0.05951807,  0.00095047,  0.02097724,  0.00637216,\n",
    "       -0.0003079 ,  0.05447122,  0.02563588,  0.00507363,  0.00506024])}\n",
    "\n",
    "best_effect_by_year_lr={2005: array([ 0.00506697,  0.        ,  0.03299984,  0.00456673,  0.00269485,\n",
    "       -0.00932709,  0.        , -0.00921414, -0.010731  ,  0.0182669 ,\n",
    "        0.04471518, -0.0211554 ,  0.01929966]), 2006: array([ 3.67115637e-02,  2.80880420e-03,  1.15801577e-02,  0.00000000e+00,\n",
    "        4.23455979e-02,  0.00000000e+00,  6.35676741e-03, -3.12089356e-04,\n",
    "       -1.64257556e-05,  2.36859396e-02,  1.11990802e-01,  1.14520368e-01,\n",
    "        7.56570302e-02]), 2007: array([ 0.01557664,  0.00163089,  0.02268264,  0.00480945,  0.        ,\n",
    "       -0.00552505,  0.        , -0.01808953, -0.01296389, -0.0236645 ,\n",
    "        0.02822433, -0.0314029 , -0.04479947]), 2008: array([ 0.02054866,  0.0019046 ,  0.02295999,  0.00122313,  0.02378123,\n",
    "       -0.002621  , -0.002953  ,  0.05869299, -0.00679713,  0.00503233,\n",
    "        0.06978857, -0.0013105 ,  0.04868076]), 2009: array([-2.03058321e-02, -8.32147937e-03, -1.66251778e-02,  8.49928876e-03,\n",
    "        2.84495021e-04, -1.08463727e-03,  3.55618777e-05, -5.84992888e-03,\n",
    "       -3.25391181e-03, -3.84068279e-03,  9.15718350e-03, -9.60170697e-03,\n",
    "        4.49857752e-03]), 2010: array([ 0.0183512 ,  0.00059764,  0.02564598,  0.00950958,  0.        ,\n",
    "        0.00249604,  0.00606433,  0.01300756, -0.0026894 ,  0.00012304,\n",
    "        0.02664792,  0.00202144,  0.03132361])}"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "GISP_init",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.9"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "37b4cc77837642ee25c4eb6578aebe03e17eb3bb59efdde49edbbc888dbc418f"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}