{ "cells": [ { "cell_type": "markdown", "id": "243c8d6c", "metadata": {}, "source": [ "# Прогнозирование `Exam_Score` по данным `StudentPerformanceFactors`\n", "\n", "Цель ноутбука — решить задачу **регрессии**: по учебным, семейным и поведенческим факторам предсказать итоговый балл `Exam_Score` с помощью **нейронных сетей**.\n", "\n", "Что делается в ноутбуке:\n", "- выполняется первичный анализ данных;\n", "- выделяются числовые и категориальные признаки;\n", "- оценивается относительная важность факторов;\n", "- подбираются архитектуры `MLPRegressor`;\n", "- отдельно сравниваются модели на **всех признаках** и на **ключевых признаках**;\n", "- финальная модель дообучается и проверяется на тестовой выборке.\n", "\n", "> В ноутбуке **нет текстового заключения**: здесь только код, комментарии, таблицы и графики." ] }, { "cell_type": "code", "execution_count": 1, "id": "37095b12", "metadata": {}, "outputs": [], "source": [ "import warnings\n", "warnings.filterwarnings(\"ignore\")\n", "\n", "from pathlib import Path\n", "import ast\n", "import random\n", "\n", "import numpy as np\n", "import pandas as pd\n", "import matplotlib.pyplot as plt\n", "\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.compose import ColumnTransformer\n", "from sklearn.pipeline import Pipeline\n", "from sklearn.impute import SimpleImputer\n", "from sklearn.preprocessing import OneHotEncoder, StandardScaler\n", "from sklearn.neural_network import MLPRegressor\n", "from sklearn.metrics import r2_score, mean_absolute_error, mean_squared_error, make_scorer\n", "from sklearn.inspection import permutation_importance\n", "\n", "SEED = 42\n", "np.random.seed(SEED)\n", "random.seed(SEED)\n", "\n", "plt.rcParams[\"figure.figsize\"] = (10, 5)\n", "plt.rcParams[\"axes.grid\"] = True" ] }, { "cell_type": "code", "execution_count": 2, "id": "5d66f4a9", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Используем файл: /home/konnilol/Downloads/StudentPerformanceFactors.csv\n" ] } ], "source": [ "# Поиск CSV в нескольких типичных местах.\n", "candidate_paths = [\n", " Path(\"StudentPerformanceFactors.csv\"),\n", " Path(\"/mnt/data/StudentPerformanceFactors.csv\"),\n", "]\n", "\n", "data_path = None\n", "for candidate in candidate_paths:\n", " if candidate.exists():\n", " data_path = candidate\n", " break\n", "\n", "if data_path is None:\n", " raise FileNotFoundError(\"Файл StudentPerformanceFactors.csv не найден.\")\n", "\n", "data = pd.read_csv(data_path)\n", "print(f\"Используем файл: {data_path.resolve()}\")" ] }, { "cell_type": "markdown", "id": "616fda5c", "metadata": {}, "source": [ "## 1. Первичный осмотр данных" ] }, { "cell_type": "code", "execution_count": 3, "id": "2e9fc5ea", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Размер датасета: (6607, 20)\n" ] }, { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	Hours_Studied	Attendance	Parental_Involvement	Access_to_Resources	Extracurricular_Activities	Sleep_Hours	Previous_Scores	Motivation_Level	Internet_Access	Tutoring_Sessions	Family_Income	Teacher_Quality	School_Type	Peer_Influence	Physical_Activity	Learning_Disabilities	Parental_Education_Level	Distance_from_Home	Gender	Exam_Score
0	23	84	Low	High	No	7	73	Low	Yes	0	Low	Medium	Public	Positive	3	No	High School	Near	Male	67
1	19	64	Low	Medium	No	8	59	Low	Yes	2	Medium	Medium	Public	Negative	4	No	College	Moderate	Female	61
2	24	98	Medium	Medium	Yes	7	91	Medium	Yes	2	Medium	Medium	Public	Neutral	4	No	Postgraduate	Near	Male	74
3	29	89	Low	Medium	Yes	8	98	Medium	Yes	1	Medium	Medium	Public	Negative	4	No	High School	Moderate	Male	71
4	19	92	Medium	Medium	Yes	6	65	Medium	Yes	3	Medium	High	Public	Neutral	4	No	College	Near	Female	70

\n", "

" ], "text/plain": [ " Hours_Studied Attendance Parental_Involvement Access_to_Resources \\\n", "0 23 84 Low High \n", "1 19 64 Low Medium \n", "2 24 98 Medium Medium \n", "3 29 89 Low Medium \n", "4 19 92 Medium Medium \n", "\n", " Extracurricular_Activities Sleep_Hours Previous_Scores Motivation_Level \\\n", "0 No 7 73 Low \n", "1 No 8 59 Low \n", "2 Yes 7 91 Medium \n", "3 Yes 8 98 Medium \n", "4 Yes 6 65 Medium \n", "\n", " Internet_Access Tutoring_Sessions Family_Income Teacher_Quality \\\n", "0 Yes 0 Low Medium \n", "1 Yes 2 Medium Medium \n", "2 Yes 2 Medium Medium \n", "3 Yes 1 Medium Medium \n", "4 Yes 3 Medium High \n", "\n", " School_Type Peer_Influence Physical_Activity Learning_Disabilities \\\n", "0 Public Positive 3 No \n", "1 Public Negative 4 No \n", "2 Public Neutral 4 No \n", "3 Public Negative 4 No \n", "4 Public Neutral 4 No \n", "\n", " Parental_Education_Level Distance_from_Home Gender Exam_Score \n", "0 High School Near Male 67 \n", "1 College Moderate Female 61 \n", "2 Postgraduate Near Male 74 \n", "3 High School Moderate Male 71 \n", "4 College Near Female 70 " ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	dtype	nunique	missing	missing_pct
Hours_Studied	int64	41	0	0.00
Attendance	int64	41	0	0.00
Parental_Involvement	object	3	0	0.00
Access_to_Resources	object	3	0	0.00
Extracurricular_Activities	object	2	0	0.00
Sleep_Hours	int64	7	0	0.00
Previous_Scores	int64	51	0	0.00
Motivation_Level	object	3	0	0.00
Internet_Access	object	2	0	0.00
Tutoring_Sessions	int64	9	0	0.00
Family_Income	object	3	0	0.00
Teacher_Quality	object	4	78	1.18
School_Type	object	2	0	0.00
Peer_Influence	object	3	0	0.00
Physical_Activity	int64	7	0	0.00
Learning_Disabilities	object	2	0	0.00
Parental_Education_Level	object	4	90	1.36
Distance_from_Home	object	4	67	1.01
Gender	object	2	0	0.00
Exam_Score	int64	45	0	0.00

\n", "

" ], "text/plain": [ " dtype nunique missing missing_pct\n", "Hours_Studied int64 41 0 0.00\n", "Attendance int64 41 0 0.00\n", "Parental_Involvement object 3 0 0.00\n", "Access_to_Resources object 3 0 0.00\n", "Extracurricular_Activities object 2 0 0.00\n", "Sleep_Hours int64 7 0 0.00\n", "Previous_Scores int64 51 0 0.00\n", "Motivation_Level object 3 0 0.00\n", "Internet_Access object 2 0 0.00\n", "Tutoring_Sessions int64 9 0 0.00\n", "Family_Income object 3 0 0.00\n", "Teacher_Quality object 4 78 1.18\n", "School_Type object 2 0 0.00\n", "Peer_Influence object 3 0 0.00\n", "Physical_Activity int64 7 0 0.00\n", "Learning_Disabilities object 2 0 0.00\n", "Parental_Education_Level object 4 90 1.36\n", "Distance_from_Home object 4 67 1.01\n", "Gender object 2 0 0.00\n", "Exam_Score int64 45 0 0.00" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Число полных дубликатов: 0\n" ] } ], "source": [ "print(\"Размер датасета:\", data.shape)\n", "display(data.head())\n", "\n", "summary = pd.DataFrame({\n", " \"dtype\": data.dtypes.astype(str),\n", " \"nunique\": data.nunique(dropna=False),\n", " \"missing\": data.isna().sum(),\n", " \"missing_pct\": (data.isna().mean() * 100).round(2),\n", "})\n", "display(summary)\n", "\n", "print(\"Число полных дубликатов:\", data.duplicated().sum())" ] }, { "cell_type": "markdown", "id": "bdb9d3f7", "metadata": {}, "source": [ "## 2. Постановка задачи и типы признаков\n", "\n", "Целевая переменная — `Exam_Score`. Это **числовой** показатель, поэтому задача машинного обучения здесь — **регрессия**.\n", "\n", "Далее:\n", "- отделяем целевую переменную;\n", "- делим признаки на числовые и категориальные;\n", "- пропуски будем обрабатывать внутри `Pipeline`, чтобы исключить утечку данных." ] }, { "cell_type": "code", "execution_count": 4, "id": "58f1452f", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Целевая переменная: Exam_Score\n", "Числовые признаки: ['Hours_Studied', 'Attendance', 'Sleep_Hours', 'Previous_Scores', 'Tutoring_Sessions', 'Physical_Activity']\n", "Категориальные признаки: ['Parental_Involvement', 'Access_to_Resources', 'Extracurricular_Activities', 'Motivation_Level', 'Internet_Access', 'Family_Income', 'Teacher_Quality', 'School_Type', 'Peer_Influence', 'Learning_Disabilities', 'Parental_Education_Level', 'Distance_from_Home', 'Gender']\n" ] } ], "source": [ "TARGET = \"Exam_Score\"\n", "\n", "X = data.drop(columns=[TARGET]).copy()\n", "y = data[TARGET].copy()\n", "\n", "numeric_features = X.select_dtypes(include=[np.number]).columns.tolist()\n", "categorical_features = X.select_dtypes(exclude=[np.number]).columns.tolist()\n", "\n", "print(\"Целевая переменная:\", TARGET)\n", "print(\"Числовые признаки:\", numeric_features)\n", "print(\"Категориальные признаки:\", categorical_features)" ] }, { "cell_type": "code", "execution_count": 5, "id": "576ab373", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "count 6607.000000\n", "mean 67.235659\n", "std 3.890456\n", "min 55.000000\n", "25% 65.000000\n", "50% 67.000000\n", "75% 69.000000\n", "max 101.000000\n", "Name: Exam_Score, dtype: float64\n" ] }, { "data": { "image/png": "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", "text/plain": [ "

" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "print(y.describe())\n", "\n", "fig, ax = plt.subplots(1, 2, figsize=(14, 4))\n", "\n", "ax[0].hist(y, bins=25, edgecolor=\"black\")\n", "ax[0].set_title(\"Распределение целевой переменной Exam_Score\")\n", "ax[0].set_xlabel(\"Exam_Score\")\n", "ax[0].set_ylabel(\"Частота\")\n", "\n", "ax[1].boxplot(y, vert=False)\n", "ax[1].set_title(\"Boxplot целевой переменной\")\n", "ax[1].set_xlabel(\"Exam_Score\")\n", "\n", "plt.tight_layout()\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "afaea877", "metadata": {}, "source": [ "## 3. Быстрый однофакторный анализ" ] }, { "cell_type": "code", "execution_count": 6, "id": "dab5b418", "metadata": {}, "outputs": [ { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	spearman_corr
Attendance	0.672366
Hours_Studied	0.480956
Previous_Scores	0.191941
Tutoring_Sessions	0.163940
Physical_Activity	0.029150
Sleep_Hours	-0.007629

\n", "

" ], "text/plain": [ " spearman_corr\n", "Attendance 0.672366\n", "Hours_Studied 0.480956\n", "Previous_Scores 0.191941\n", "Tutoring_Sessions 0.163940\n", "Physical_Activity 0.029150\n", "Sleep_Hours -0.007629" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAA04AAAHICAYAAAB51rptAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjgsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvwVt1zgAAAAlwSFlzAAAPYQAAD2EBqD+naQAAkwFJREFUeJzs3Xd8DPn/B/DXpieipyBCdKeGaHEX5ZSoJ8rph3A45A45LRzRW5xeD9HbRXcIEeWUaIdTTjv1tEiQQkjb9+8Pv51v1oaNCJuN1/PxyIP97GdmPzvvmdl5z3zmMyoRERAREREREdFbmRi6AURERERERJkdEyciIiIiIiI9mDgRERERERHpwcSJiIiIiIhIDyZOREREREREejBxIiIiIiIi0oOJExERERERkR5MnIiIiIiIiPRg4kRERERERKQHEyciIiIiIiI9mDgRvWH58uVQqVQ4ffq0znuLFy+GSqWCl5cXkpOTDdA6Sovw8HAMGjQIpUuXho2NDbJlywY3NzeMHz8eUVFRhm4eEX0mRo8eDZVK9da/R48eGbqJH+TIkSNo3LgxnJycYGVlhUKFCqF58+ZYu3atoZtG9FGYGboBRMZiy5Yt6NOnDzw8PLB+/XqYmpoaukmUilOnTqFJkyZ4/vw5OnfuDDc3NwDA6dOnMXnyZPz555/Yu3evgVtJRJ+TBQsWwNbWVqc8V65cn74xGSQoKAjt2rWDq6sr+vfvj9y5c+PWrVv4888/sXjxYnTs2NHQTSTKcEyciNLg4MGD6NChA8qUKYMdO3bAysrK0E2iVERFRaFly5YwNTXF2bNnUbp0aa33J0yYgMWLFxuodUT0uWrTpg3s7OwM3YwMNXr0aJQpUwbHjx+HhYWF1nuPHz/+ZO0QEbx69QrW1taf7DPp88WuekR6nDt3Di1atED+/PmxZ88e5MyZU6dOUFAQ3NzcYG1tDTs7O3Tu3Bn379/XqtOtWzfY2tri5s2b8PT0RLZs2VCgQAGMHTsWIqLUu337NlQqFaZNm4YZM2agcOHCsLa2Ru3atXHx4kWdz75y5QratGmDPHnywMrKClWqVMH27dtT/S516tRJtbvI8uXLteotWLAA5cqVg42NjVa9jRs3as2rXLlyOp8xbdo0qFQq3L59WynTdH9MWaZWq1GhQoVUP3/jxo2oUqUKsmfPrvX506ZNS/V7aSxatAj379/H9OnTdZImAHB0dMQvv/yivHZxcXlnN5qUVCoVRo8erVUWEBAAlUqFOnXqKGUHDx6ESqXChg0bMHz4cOTLlw/ZsmXDN998g//++09rek08vLy8dNrau3dvqFQqnWWsVqsxc+ZMlC1bFlZWVnB0dETv3r3x7NkzrXouLi5o1qyZznx9fHx0vtuyZcvw9ddfw8HBAZaWlihTpgwWLFigM21qXFxc0K1bN60yzTI4ePCg1ndNuZyA11cHU1vWALB69WpUq1YNNjY2yJ07N2rVqqVzpVDzOW/+ubi4aNV7/PgxevToAUdHR1hZWaFixYpYsWKFVh3Ndqf5Mzc3h4uLCwYPHoyEhASl3ru68r7tu3bt2hVWVla4fPmyVj1PT0/kzp0bDx48eOu83mzXm38fsu6lJR5PnjxB48aNUbBgQVhaWiJ//vzo1KkT7ty5o9PG1LbPcuXKaX1OQkICRo0aBTc3N+TMmRPZsmWDh4cHDhw4kOr3TrlviI2NhZubG4oUKYKHDx8q5S9evMDPP/8MZ2dnWFpaolSpUpg2bZrWfhWA1nIzNTWFk5MTevXqlebuu2lZJ9MrrevI06dPMWjQIJQvXx62trbIkSMHGjdujL///ltrOs268Pvvv2PMmDFwcnJC9uzZ0aZNG0RHRyM+Ph4DBgyAg4MDbG1t4e3tjfj4+Pdq840bN1C1alWdpAkAHBwctF6r1WrMmjUL5cuXh5WVFezt7dGoUSOt7SgpKQnjxo1DsWLFYGlpCRcXFwwfPlynXZp92549e1ClShVYW1tj0aJFAF6fPBswYICyLhQvXhxTpkyBWq1+r+9G9Da84kT0Djdu3ECjRo1gaWmJPXv2IH/+/Dp1li9fDm9vb1StWhWTJk1CeHg4Zs2ahaNHj+Ls2bNaXTGSk5PRqFEj1KhRA1OnTkVwcDD8/f2RlJSEsWPHas135cqViI2NRb9+/fDq1SvMmjULX3/9NS5cuABHR0cAwKVLl/Dll1/CyckJw4YNQ7Zs2fD777/Dy8sLmzZtQsuWLXXaW7p0aYwYMQIAEBkZiYEDB2q9v2HDBvTt2xd16tTBjz/+iGzZsuHy5cuYOHHihy5OLatWrcKFCxd0ysPCwtC2bVtUrFgRkydPRs6cOVNtZ2q2b98Oa2trtGnTJs3tcHV1xc8//6xVtnLlSoSEhLxzuqioKEyaNOmt70+YMAEqlQpDhw7F48ePMXPmTNSvXx/nzp3TOjNqZWWFnTt34vHjx8rBxsuXL7Fhw4ZUr2z27t1bWed++ukn3Lp1C3PnzsXZs2dx9OhRmJubp/m7ayxYsABly5bFN998AzMzM+zYsQN9+/aFWq1Gv3793nt+aTV06NBUy8eMGYPRo0ejZs2aGDt2LCwsLHDixAns378fDRs21Kk/fPhwfPHFFwCA3377DXfv3lXee/nyJerUqYN///0XPj4+KFKkCIKCgtCtWzdERUWhf//+WvPq1asXPDw8EB8fjz179mDatGmwsrLCuHHj0v09Z82ahf3796Nr164ICwuDqakpFi1ahL1792LVqlUoUKCA3nl06NABTZo00Srz8/NLtW5a1703pRaPhIQEZM+eHf3790fevHlx48YNzJkzB+fPn091+9UnJiYGS5YsQYcOHdCzZ0/ExsZi6dKl8PT0xMmTJ+Hq6prqdImJiWjdujXu3r2Lo0ePKvtiEcE333yDAwcOoEePHnB1dcWePXswePBg3L9/HzNmzNCaT8uWLdGqVSskJSUhLCwMv/32G16+fIlVq1a9s93vu06+6enTpzplZmZmyu9DWteRmzdvYuvWrfj2229RpEgRhIeHY9GiRahduzb++ecfnXVp0qRJsLa2xrBhw/Dvv/9izpw5MDc3h4mJCZ49e4bRo0fj+PHjWL58OYoUKYJRo0bp/S4ahQsXRmhoKO7du4eCBQu+s26PHj2wfPlyNG7cGN9//z2SkpJw+PBhHD9+HFWqVAEAfP/991ixYgXatGmDn3/+GSdOnMCkSZNw+fJlbNmyRWt+V69eRYcOHdC7d2/07NkTpUqVQlxcHGrXro379++jd+/eKFSoEI4dOwY/Pz88fPgQM2fOTPN3I3orISIty5YtEwDyxx9/SLFixQSANGzYMNW6CQkJ4uDgIOXKlZOXL18q5X/88YcAkFGjRillXbt2FQDy448/KmVqtVqaNm0qFhYWEhERISIit27dEgBibW0t9+7dU+qeOHFCAMjAgQOVsnr16kn58uXl1atXWvOsWbOmlChRQqe9X375pdStW1d5rfmsZcuWKWUdOnSQXLlyaX2fAwcOCAAJCgpSymrXri1ly5bV+YyAgAABILdu3VLKNMtUU/bq1SspVKiQNG7cWOfz/fz8BIA8fPhQp50BAQE6n5dS7ty5pWLFiu+sk1LhwoWladOmOuX9+vWTN3ePAMTf3195PWTIEHFwcBA3NzepXbu2Uq5ZVk5OThITE6OU//777wJAZs2apZRplmGFChVk2rRpSvmqVaukYMGC4uHhobWMDx8+LABkzZo1Wm0LDg7WKX+f7xYXF6dTz9PTU4oWLapT/qYiRYpIly5dtMo0y+DAgQNa3zXlctq1a5cAkEaNGmm15/r162JiYiItW7aU5ORkrfmq1Wqt1yEhIQJADh06pJR17dpVChcurLyeOXOmAJDVq1crZQkJCeLu7i62trZKjFLbFkREChQoIE2aNFFea9blU6dOvXWZvPldRUT27NkjAGT8+PFy8+ZNsbW1FS8vr7fOQ+Nd637ZsmU/aN1LSzxSM3XqVAEgkZGR793GpKQkiY+P16rz7NkzcXR0lO7du+t872XLlolarZZOnTqJjY2NnDhxQmvarVu3Kss1pTZt2ohKpZJ///1XKXtzGxYRqVmzppQpU+ad3/d91sk3+fv7C4BU/0qVKqVVNy3ryKtXr3TacOvWLbG0tJSxY8cqZZp1oVy5cpKQkKCUd+jQQVQqlTRu3FhrHu7u7lrbTVosXbpUAIiFhYXUrVtXRo4cKYcPH9Zp3/79+wWA/PTTTzrz0Cy/c+fOCQD5/vvvtd4fNGiQAJD9+/crZYULFxYAEhwcrFV33Lhxki1bNrl27ZpW+bBhw8TU1FTu3r37Xt+PKDXsqkf0Ft26dcN///2Hjh07Yu/evQgKCtKpc/r0aTx+/Bh9+/bVujrQtGlTlC5dGjt37tSZxsfHR/m/SqWCj48PEhISsG/fPq16Xl5ecHJyUl5Xq1YN1atXx65duwC8PoO5f/9+tG3bFrGxsYiMjERkZCSePHkCT09PXL9+Xae7YEJCAiwtLd/5vWNjY2FjY/NR7+OaN28enjx5An9//1Q/38TEJF03TcfExCB79uwZ0MJ3u3//PubMmYORI0emesM3AHTp0kWrLW3atEH+/PmV+KXk7e2NZcuWKa+XLVuGrl27wsREexcdFBSEnDlzokGDBkq8IyMj4ebmBltbW53uTomJiVr1IiMj8erVK53PT3kVIjo6GpGRkahduzZu3ryJ6Ojody4LBwcH3Lt375113iQi8PPzQ+vWrVG9enWt97Zu3Qq1Wo1Ro0bpfP83u/RputC9a53etWsX8uXLhw4dOihl5ubm+Omnn/D8+XMcOnRIq/7z588RGRmJ+/fv47fffsOjR49Qr149nflqllNsbGyavnPDhg3Ru3dvjB07Fq1atYKVlZXSvSijvc+6B7w7HhqxsbF4/PgxwsLCsG7dOpQtWxZ58uTRqhMXF6ezvr05+qipqanStUutVuPp06dISkpClSpVcObMmVQ/e/DgwVizZg1+//13VKtWTeu9Xbt2wdTUFD/99JNW+c8//wwRwe7du1Nt46NHj7Bp0yb8/fffqcY3pfdZJ99m06ZNCAkJ0fpLuc0DaVtHLC0tlTYkJyfjyZMnsLW1RalSpVJdfl26dNG6Cl29enWICLp3765Vr3r16vjvv/+QlJSUpu8DAN27d0dwcDDq1KmDI0eOYNy4cfDw8ECJEiVw7Ngxre+uUqlS3d9rlp9m3fT19dV6X9Mb4M3f0iJFisDT01OrLCgoCB4eHsidO7fWOli/fn0kJyfjzz//TPN3I3obJk5Eb/H06VOsXr0aK1asUEYNevMgUtPPv1SpUjrTly5dWus+AAAwMTFB0aJFtcpKliwJAFr3/wBAiRIldOZZsmRJpd6///4LEcHIkSNhb2+v9af5gXrzBt2oqKi3HuhruLu748GDBxg9ejTu3r2LyMhIvQfP7yM6OhoTJ06Er6+v0uXwzc9Xq9Xo378/bty4gcjISJ37d94mR44caT6Q/RD+/v4oUKAAevfu/dY6b8ZPpVKhePHiOnEGgE6dOuHatWs4efIkbt++jYMHD+rcNwQA169fR3R0NBwcHHRi/vz5c5147927V6fe0qVLdeZ79OhR1K9fH9myZUOuXLlgb2+P4cOHA4De2NesWROHDh3C+vXr8fjx4zStL2vWrMGlS5dS7f5548YNmJiYoEyZMu+cBwDl3pR3rdN37txBiRIldA54NV373txGf/zxR9jb26NgwYLo3bs3unbtmmo30fr168Pe3h45cuRA7ty50bdvX7x48eKd7Z02bRry5MmDc+fOYfbs2Tr3gWSU91n3gHfHQ6Nnz55wdHREzZo1YWZmhn379ukkDf7+/jrr25UrV3TmtWLFClSoUAFWVlbImzcv7O3tsXPnzlTXm0WLFuHXX38FgFT3A3fu3EGBAgV0Tpi8Lb4BAQGwt7dH/vz50aZNG3h4eGDKlClv/d7A+62Tb1OrVi3Ur19f68/d3V2nnr51RK1WY8aMGShRogQsLS1hZ2cHe3t7nD9/PtXlV6hQIa3Xmnt0nZ2ddcrVavV77+s9PT2xZ88eREVF4c8//0S/fv1w584dNGvWTNkf3bhxAwUKFNBJtFO6c+cOTExMULx4ca3yfPnyIVeuXDpxLFKkiM48rl+/juDgYJ11sH79+gA+7YAVlHXxHieitwgICMC3334L4PV9EzVq1ICfnx/mz59v4Ja9prnZddCgQTpn3jTe/BF69OjRW+tqDBw4EFevXsW4ceMwZsyYjGlsClOmTIGJiQkGDx6MJ0+e6Lzfvn17nDlzBnPmzMFvv/32XvMuXbo0zp07h4SEhFRvWM4Ily9fxvLly7F69ep03U+UGnt7ezRv3hzLli2Do6MjvvzyS53YAa9j7uDggDVr1rx1PilVr14d48eP1yqbO3cutm3bpry+ceMG6tWrh9KlS2P69OlwdnaGhYUFdu3ahRkzZui9qXr48OE4evSo1hWdd0lISMDIkSPRo0cP5aRBemmegZMvX74Pmk9KgwcPRsOGDZGcnIxLly4pg7e8eXVg3rx5KFmyJOLj43Hw4EFlYIR37R/Onj2rHLxduHAhzcvsY0prPH755Rd4e3vjxo0bmDp1Ktq3b499+/bBzOx/hxG9evVS9pkaPXv21Hq9evVqdOvWDV5eXhg8eDAcHBxgamqKSZMm4caNGzqfe/z4cUyYMAGnTp3CwIED0ahRow8ane67775Dly5doFarcfPmTYwbNw7NmjVLNRE0BH3ryMSJEzFy5Eh0794d48aNQ548eWBiYoIBAwakuq2+7bEZbyuXNwbUSCsbGxt4eHjAw8MDdnZ2GDNmDHbv3o2uXbu+13zSGoPU7tVTq9Vo0KABhgwZkuo0H7q/IQKYOBG9Va1atZT/V61aFf369cO8efPQpUsX1KhRA8Drm2OB1zeqfv3111rTX716VXlfQ/NjnXIHfu3aNQDQGQns+vXrOm26du2aUk9z5crc3Fw5o/Yu9+7dQ2xsrHIm9m2sra2xePFinD17Fjlz5oS/vz/+/vtvDBo0SO9n6PPgwQPMmjULkyZNQvbs2VNNnExMTDBt2jRcuHABt27dwvz58xEeHo7OnTvrnX/z5s0RFhaGTZs2fbSDUj8/P7i6uqJdu3bvrPdm/EQE//77LypUqJBq/e7du6NTp07ImTOnzuh9GsWKFcO+ffvw5ZdfpmnoXTs7O511Y+vWrVqvd+zYgfj4eGzfvl3r7PSb3f7e9RlhYWH4559/lETmXevL/Pnz8fjx43d+R7VajX/++eetAwVo/PPPP7C3t0fevHnfWqdw4cI4f/481Gq11lUnzZWQN7fRMmXKKMvM09MT8fHxGD58OCZMmKB14321atWUm9qbNm2Kv//+G8HBwW9tx4sXL+Dt7Y0yZcqgZs2amDp1Klq2bImqVau+8zumx/use/rioVGuXDllhMfy5cujVq1aCAkJQePGjZU6JUqU0FnfsmXLpvV648aNKFq0KDZv3qx1kJxaNy7g9XYxfPhwPHjwAGXKlMHAgQO1BnIoXLgw9u3bh9jYWK2rTm+Lb9GiRbXamDNnTnTs2BHHjx9P9QoQ8H7r5IdIyzqyceNG1K1bV+fKcVRUVKYZ7lyzXWhGPixWrBj27NmDp0+fvvWqU+HChaFWq3H9+nWt36jw8HBERUXpxDE1xYoVw/Pnz9P0e0iUXuyqR5RGEyZMQP78+dGrVy+lH3iVKlXg4OCAhQsXag2Zunv3bly+fBlNmzbVmc/cuXOV/4sI5s6dC3Nzc51+9lu3btW6R+nkyZM4ceKEcqDi4OCAOnXqYNGiRVpD82pERERovV6/fj0A6CR4qfHz88Pdu3exevVq1K9fX3mI7IcaM2YMHB0d8cMPP7yz3pw5c7B//36sWbMG9evXx5dffpmm+f/www/Inz8/fv75ZyUhTenx48c6V2DeR1hYGLZt24bJkyfrPTOqGRVRY+PGjXj48KHWgWZKjRo1QrZs2fD06VO0bds21Tpt27ZFcnJyqiO8JSUlpXlY5ZQ0Z55TnmmOjo7WucLyLiYmJihXrpzSBelt60tsbCwmTJiAgQMHvvUqkZeXF0xMTDB27FidM+gp2xgbG4tdu3bpXZ+bNGmCR48eYcOGDUpZUlIS5syZA1tbW9SuXfud0798+RIAtIYkT41arX7nQ7GHDh2Ku3fvYsWKFZg+fTpcXFzQtWvX9x4COi3Suu6lJR6piYyMBIB0tT219e3EiRMICwtLtb6HhwcAoECBApgyZQpWr16tNQR4kyZNkJycrLVfBYAZM2ZApVK9dXvT0MT3Xd8lrevkh0rLOmJqaqrzmUFBQTr3s34KoaGhqZZr7lfSdGFv3bo1RCTVHgya76IZMfLNke+mT58OAKn+lr6pbdu2CAsLw549e3Tei4qKeq/7t4jehleciNIoe/bsmDNnDlq1aoVff/0VQ4cOhbm5OaZMmQJvb2/Url0bHTp0UIYjd3Fx0bk3wsrKCsHBwejatSuqV6+O3bt3Y+fOnRg+fLhON6vixYvjq6++Qp8+fRAfH4+ZM2cib968Wt0Q5s2bh6+++grly5dHz549UbRoUYSHhyMsLAz37t3D33//jfDwcPj7+2PJkiVo3759qs83Smnfvn2YMWMGVq1apfcs3/Pnz3XOsl+9ehUAcOjQIZibm2sNcLF3716sWbPmnd3oLl26hCFDhmD06NHvfTY+d+7c2LJlC5o0aQJXV1d07txZOYg/c+YM1q1b99azymmxd+9eNGjQIE1nNPPkyYOvvvoK3t7eCA8Px8yZM1G8eHGdrksapqamuHz5MkRE5yy9Ru3atdG7d29MmjQJ586dQ8OGDWFubo7r168jKCgIs2bNeq+h2IHXN6RbWFigefPm6N27N54/f47FixfDwcEh1YT8Q5w5cwZ2dnZv7UoDvF7vR4wYodxo3qpVK1haWuLUqVMoUKAAJk2apDyb5tmzZxg2bNg7P7NXr15YtGgRunXrhr/++gsuLi7YuHEjjh49ipkzZ+rcGxMWFgYzMzOlq96cOXNQqVIlnSvCYWFhiIyMVLrqhYaGvvUq2/79+zF//nz4+/ujcuXKAF4PAFKnTh2MHDkSU6dOTcPSS7u0rntpicfixYvx559/onLlysiRIwf++ecfLF68GPnz59c7qEJqmjVrhs2bN6Nly5Zo2rQpbt26hYULF6JMmTJ4/vz5O6ft1asX1q5dix9++AEXL16EjY0Nmjdvjrp162LEiBG4ffs2KlasiL1792Lbtm0YMGAAihUrpjWP8+fPY/Xq1RAR3LhxA7Nnz0bBggWVqySpScs6qc/GjRtTvRevQYMGcHR0TPM60qxZM4wdOxbe3t6oWbMmLly4gDVr1ujcO/sptGjRAkWKFEHz5s1RrFgxvHjxAvv27cOOHTtQtWpVNG/eHABQt25dfPfdd5g9ezauX7+ORo0aQa1W4/Dhw6hbty58fHxQsWJFdO3aFb/99huioqJQu3ZtnDx5EitWrICXlxfq1q2rtz2DBw/G9u3b0axZM3Tr1g1ubm548eIFLly4gI0bN+L27duZ5qocGbFPPIofUaanb7jhFi1aiI2Njdy8eVMp27Bhg1SqVEksLS0lT5480qlTJ62hxEVeD5OcLVs2uXHjhjRs2FBsbGzE0dFR/P39tYZvTTm076+//irOzs5iaWkpHh4e8vfff+u058aNG9KlSxfJly+fmJubi5OTkzRr1kw2btwoIiJHjx6V4sWLy+jRo3WGAX5zCObIyEgpUKCAdOjQQave24Yjx1uG2dX8aearWaaurq5aw/e++fmvXr2SChUqyFdffSVJSUmpLpO0ePDggQwcOFBKliwpVlZWYmNjI25ubjJhwgSJjo5W6r3vcOQqlUr++usvrfI3h3XWLKt169aJn5+fODg4iLW1tTRt2lTu3LmjM21qQ7rre/+3334TNzc3sba2luzZs0v58uVlyJAh8uDBg3R9t+3bt0uFChXEyspKXFxcZMqUKRIYGKgzrHxavW04cgAyY8YMrbqa4ZrfFBgYqGxTuXPnltq1a0tISIiIiLRs2VIaN26sMzS1iO5w5CIi4eHh4u3tLXZ2dmJhYSHly5fXGXZcs45p/kxMTKRgwYLStWtXrW1Zsy5r/iwsLKR48eIyatQoZftKuU7ExMRI4cKFpXLlypKYmKj1mQMHDhQTExMJCwt767JMz3DkaV330hKPQ4cOiYeHh+TKlUssLS3FxcVFevbsqbVevE8b1Wq1TJw4UQoXLiyWlpZSqVIl+eOPP3Ti9rbh4a9evSpWVlZaj2WIjY2VgQMHSoECBcTc3FxKlCghAQEBOkOFp4ybSqWSfPnySatWreTy5cs67U7Nu9bJt3nXcOSabeR91pFXr17Jzz//LPnz5xdra2v58ssvJSws7K37oZT7bJG3/75p2ql5LEZarFu3Ttq3by/FihUTa2trsbKykjJlysiIESO0hsMXeT0MfUBAgJQuXVosLCzE3t5eGjdurLU/TUxMlDFjxkiRIkXE3NxcnJ2dxc/PT+txGyJv37eJvF4X/Pz8pHjx4mJhYSF2dnZSs2ZNmTZtmtaw7ETppRLJwOvMRPRW3bp1w8aNG/WeVb19+zaKFCmCgICADLmvyFBcXFwwevToVEeHy8oOHjyIunXrIigo6L2v/hB9CK57REQfF+9xIiIiIiIi0oP3OBHRR1G7dm2t+5uIiCjzi46OVgbNeJuMfAQAkTFh4kREH8WKFSsM3QQiInpP/fv317v/5l0e9LniPU5EREREBOD189EePHjwzjp8VhJ9rpg4ERERERER6ZEpuurNmzcPAQEBePToESpWrIg5c+agWrVqqdatU6cODh06pFPepEkT7Ny5U+9nqdVqPHjwANmzZ9f7AEsiIiIiIsq6RASxsbEoUKAATEzePW6ewROnDRs2wNfXFwsXLkT16tUxc+ZMeHp64urVq3BwcNCpv3nzZq0nuD958gQVK1bEt99+m6bPe/DgAZydnTOs/UREREREZNz+++8/FCxY8J11DN5Vr3r16qhatSrmzp0L4PUVIWdnZ/z44496nwgPADNnzsSoUaPw8OFDZMuWTW/96Oho5MqVC//99x9y5Mjxwe3/3CUmJmLv3r1o2LAhzM3NDd0c+kgY56yPMf48MM5ZH2Oc9THGGSsmJgbOzs6IiopCzpw531nXoFecEhIS8Ndff8HPz08pMzExQf369REWFpameSxduhTt27d/a9IUHx+P+Ph45XVsbCwAwNraGtbW1h/QegIAMzMz2NjYwNramhtvFsY4Z32M8eeBcc76GOOsjzHOWImJiQCQplt4DJo4RUZGIjk5GY6Ojlrljo6OuHLlit7pT548iYsXL2Lp0qVvrTNp0iSMGTNGp3zv3r2wsbF5/0ZTqkJCQgzdBPoEGOesjzH+PDDOWR9jnPUxxhkjLi4uzXUNfo/Th1i6dCnKly//1oEkAMDPzw++vr7Ka83luIYNG7KrXgZITExESEgIGjRowLMeWRjjnPUxxp8HxjnrY4yzPsY4Y8XExKS5rkETJzs7O5iamiI8PFyrPDw8XO9TqV+8eIH169dj7Nix76xnaWkJS0tLnXJzc3OubBmIy/PzwDhnfYzx54FxzvoY46yPMc4Y77MM3z3m3kdmYWEBNzc3hIaGKmVqtRqhoaFwd3d/57RBQUGIj49H586dP3YziYiIiIjoM2fwrnq+vr7o2rUrqlSpgmrVqmHmzJl48eIFvL29AQBdunSBk5MTJk2apDXd0qVL4eXlhbx58xqi2URERERE9BkxeOLUrl07REREYNSoUXj06BFcXV0RHBysDBhx9+5dnYdRXb16FUeOHMHevXsN0WQiIiIiIvrMGDxxAgAfHx/4+Pik+t7Bgwd1ykqVKgUDP36KiIiIiIg+Iwa9x4mIiIiIiMgYMHEiIiIiIiLSg4kTERERERGRHkyciIiIiIiI9MgUg0N87lyG7TR0E9LN0lQwtRpQbvQexCerDN2cdLs9uamhm0BEREREmRivOBEREREREenBxImIiIiIiEgPJk5ERERERER6MHEiIiIiIiLSg4kTERERERGRHkyciIiIiIiI9GDiREREREREpAcTJyIiIiIiIj2YOBEREREREenBxImIiIiIiEgPJk5ERERERER6MHEiIiIiIiLSg4kTERERERGRHkyciIiIiIiI9GDiREREREREpAcTJyIiIiIiIj2YOBEREREREenBxImIiIiIiEgPJk5ERERERER6MHEiIiIiIiLSg4kTERERERGRHkyciIiIiIiI9GDiREREREREpAcTJyIiIiIiIj2YOBEREREREelh8MRp3rx5cHFxgZWVFapXr46TJ0++s35UVBT69euH/Pnzw9LSEiVLlsSuXbs+UWuJiIiIiOhzZGbID9+wYQN8fX2xcOFCVK9eHTNnzoSnpyeuXr0KBwcHnfoJCQlo0KABHBwcsHHjRjg5OeHOnTvIlSvXp288ERERERF9NgyaOE2fPh09e/aEt7c3AGDhwoXYuXMnAgMDMWzYMJ36gYGBePr0KY4dOwZzc3MAgIuLy6dsMhERERERfYYMljglJCTgr7/+gp+fn1JmYmKC+vXrIywsLNVptm/fDnd3d/Tr1w/btm2Dvb09OnbsiKFDh8LU1DTVaeLj4xEfH6+8jomJAQAkJiYiMTExA79R+lmaiqGbkG6WJqL1r7HKLOtCZqVZPlxOWRdj/HlgnLM+xjjrY4wz1vssR4MlTpGRkUhOToajo6NWuaOjI65cuZLqNDdv3sT+/fvRqVMn7Nq1C//++y/69u2LxMRE+Pv7pzrNpEmTMGbMGJ3yvXv3wsbG5sO/SAaYWs3QLfhw46qoDd2ED8L75NImJCTE0E2gj4wx/jwwzlkfY5z1McYZIy4uLs11DdpV732p1Wo4ODjgt99+g6mpKdzc3HD//n0EBAS8NXHy8/ODr6+v8jomJgbOzs5o2LAhcuTI8ama/k7lRu8xdBPSzdJEMK6KGiNPmyBerTJ0c9Lt4mhPQzchU0tMTERISAgaNGigdJOlrIUx/jwwzlkfY5z1McYZS9MbLS0MljjZ2dnB1NQU4eHhWuXh4eHIly9fqtPkz58f5ubmWt3yvvjiCzx69AgJCQmwsLDQmcbS0hKWlpY65ebm5plmZYtPNt6EQyNerTLq75FZ1oXMLjNtN/RxMMafB8Y562OMsz7GOGO8zzI02HDkFhYWcHNzQ2hoqFKmVqsRGhoKd3f3VKf58ssv8e+//0Kt/l+3sGvXriF//vypJk1EREREREQZwaDPcfL19cXixYuxYsUKXL58GX369MGLFy+UUfa6dOmiNXhEnz598PTpU/Tv3x/Xrl3Dzp07MXHiRPTr189QX4GIiIiIiD4DBr3HqV27doiIiMCoUaPw6NEjuLq6Ijg4WBkw4u7duzAx+V9u5+zsjD179mDgwIGoUKECnJyc0L9/fwwdOtRQX4GIiIiIiD4DBh8cwsfHBz4+Pqm+d/DgQZ0yd3d3HD9+/CO3ioiIiIiI6H8M2lWPiIiIiIjIGDBxIiIiIiIi0oOJExERERERkR5MnIiIiIiIiPRg4kRERERERKQHEyciIiIiIiI9mDgRERERERHpwcSJiIiIiIhIDyZOREREREREejBxIiIiIiIi0oOJExERERERkR5MnIiIiIiIiPRg4kRERERERKQHEyciIiIiIiI9mDgRERERERHpwcSJiIiIiIhIDyZOREREREREejBxIiIiIiIi0oOJExERERERkR5MnIiIiIiIiPRg4kRERERERKQHEyciIiIiIiI9mDgRERERERHpwcSJiIiIiIhIDyZOREREREREejBxIiIiIiIi0oOJExERERERkR5MnIiIiIiIiPRg4kRERERERKQHEyciIiIiIiI9mDgRERERERHpkSkSp3nz5sHFxQVWVlaoXr06Tp48+da6y5cvh0ql0vqzsrL6hK0lIiIiIqLPjcETpw0bNsDX1xf+/v44c+YMKlasCE9PTzx+/Pit0+TIkQMPHz5U/u7cufMJW0xERERERJ8bM0M3YPr06ejZsye8vb0BAAsXLsTOnTsRGBiIYcOGpTqNSqVCvnz50jT/+Ph4xMfHK69jYmIAAImJiUhMTPzA1mcMS1MxdBPSzdJEtP41VpllXcisNMuHyynrYow/D4xz1scYZ32MccZ6n+WoEhGDHfEmJCTAxsYGGzduhJeXl1LetWtXREVFYdu2bTrTLF++HN9//z2cnJygVqtRuXJlTJw4EWXLlk31M0aPHo0xY8bolK9duxY2NjYZ9l2IiIiIiMi4xMXFoWPHjoiOjkaOHDneWdegV5wiIyORnJwMR0dHrXJHR0dcuXIl1WlKlSqFwMBAVKhQAdHR0Zg2bRpq1qyJS5cuoWDBgjr1/fz84Ovrq7yOiYmBs7MzGjZsqHfhfCrlRu8xdBPSzdJEMK6KGiNPmyBerTJ0c9Lt4mhPQzchU0tMTERISAgaNGgAc3NzQzeHPgLG+PPAOGd9jHHWxxhnLE1vtLQweFe99+Xu7g53d3fldc2aNfHFF19g0aJFGDdunE59S0tLWFpa6pSbm5tnmpUtPtl4Ew6NeLXKqL9HZlkXMrvMtN3Qx8EYfx4Y56yPMc76GOOM8T7L0KCDQ9jZ2cHU1BTh4eFa5eHh4Wm+h8nc3ByVKlXCv//++zGaSEREREREZNjEycLCAm5ubggNDVXK1Go1QkNDta4qvUtycjIuXLiA/Pnzf6xmEhERERHRZ87gXfV8fX3RtWtXVKlSBdWqVcPMmTPx4sULZZS9Ll26wMnJCZMmTQIAjB07FjVq1EDx4sURFRWFgIAA3LlzB99//70hvwYREREREWVhBk+c2rVrh4iICIwaNQqPHj2Cq6srgoODlQEj7t69CxOT/10Ye/bsGXr27IlHjx4hd+7ccHNzw7Fjx1CmTBlDfQUiIiIiIsriDJ44AYCPjw98fHxSfe/gwYNar2fMmIEZM2Z8glYRERERERG9likSJ6KszmXYTkM34YNYmgqmVns9dL6xjp54e3JTQzeBiIiIjJhBB4cgIiIiIiIyBkyciIiIiIiI9GDiREREREREpAcTJyIiIiIiIj2YOBEREREREenBxImIiIiIiEgPJk5ERERERER6MHEiIiIiIiLSg4kTERERERGRHkyciIiIiIiI9GDiREREREREpAcTJyIiIiIiIj2YOBEREREREenBxImIiIiIiEgPJk5ERERERER6MHEiIiIiIiLSg4kTERERERGRHkyciIiIiIiI9GDiREREREREpAcTJyIiIiIiIj2YOBEREREREenBxImIiIiIiEgPJk5ERERERER6MHEiIiIiIiLSg4kTERERERGRHkyciIiIiIiI9GDiREREREREpAcTJyIiIiIiIj2YOBEREREREemRKRKnefPmwcXFBVZWVqhevTpOnjyZpunWr18PlUoFLy+vj9tAIiIiIiL6rBk8cdqwYQN8fX3h7++PM2fOoGLFivD09MTjx4/fOd3t27cxaNAgeHh4fKKWEhERERHR58rM0A2YPn06evbsCW9vbwDAwoULsXPnTgQGBmLYsGGpTpOcnIxOnTphzJgxOHz4MKKiot46//j4eMTHxyuvY2JiAACJiYlITEzMuC/yASxNxdBNSDdLE9H611h97HXBmGMMZI04Z5btPbPSLB8up6yNcc76GOOsjzHOWO+zHFUiYrAjoYSEBNjY2GDjxo1a3e26du2KqKgobNu2LdXp/P39cf78eWzZsgXdunVDVFQUtm7dmmrd0aNHY8yYMTrla9euhY2NTUZ8DSIiIiIiMkJxcXHo2LEjoqOjkSNHjnfWNegVp8jISCQnJ8PR0VGr3NHREVeuXEl1miNHjmDp0qU4d+5cmj7Dz88Pvr6+yuuYmBg4OzujYcOGehfOp1Ju9B5DNyHdLE0E46qoMfK0CeLVKkM3J90ujvb8qPM35hgDWSPOHzvGxi4xMREhISFo0KABzM3NDd0c+kgY56yPMc76GOOMpemNlhYG76r3PmJjY/Hdd99h8eLFsLOzS9M0lpaWsLS01Ck3NzfPNCtbfLJxHoimFK9WGfX3+NjrgjEvm5SMOc6ZZXvP7DLTvpE+HsY562OMsz7GOGO8zzI0aOJkZ2cHU1NThIeHa5WHh4cjX758OvVv3LiB27dvo3nz5kqZWq0GAJiZmeHq1asoVqzYx200ERERERF9dgw6qp6FhQXc3NwQGhqqlKnVaoSGhsLd3V2nfunSpXHhwgWcO3dO+fvmm29Qt25dnDt3Ds7Ozp+y+URERERE9JkweFc9X19fdO3aFVWqVEG1atUwc+ZMvHjxQhllr0uXLnBycsKkSZNgZWWFcuXKaU2fK1cuANApJyIiIiIiyigGT5zatWuHiIgIjBo1Co8ePYKrqyuCg4OVASPu3r0LExODP26KiIiIiIg+YwZPnADAx8cHPj4+qb538ODBd067fPnyjG8QERERERFRCryUQ0REREREpAcTJyIiIiIiIj2YOBEREREREenBxImIiIiIiEgPJk5ERERERER6MHEiIiIiIiLSg4kTERERERGRHkyciIiIiIiI9GDiREREREREpMd7J05JSUlYuXIlwsPDP0Z7iIiIiIiIMp33TpzMzMzwww8/4NWrVx+jPURERERERJlOurrqVatWDefOncvgphAREREREWVOZumZqG/fvvD19cV///0HNzc3ZMuWTev9ChUqZEjjiIiIiIiIMoN0JU7t27cHAPz0009KmUqlgohApVIhOTk5Y1pHRERERESUCaQrcbp161ZGt4OIiIiIiCjTeu/EKTExEV9//TX++OMPfPHFFx+jTURERERERJnKew8OYW5uzhH1iIiIiIjos5KuUfX69euHKVOmICkpKaPbQ0RERERElOmk6x6nU6dOITQ0FHv37kX58uV1RtXbvHlzhjSOiIiIiIgoM0hX4pQrVy60bt06o9tCRERERESUKaUrcVq2bFlGt4OIiIiIiCjTSlfipBEREYGrV68CAEqVKgV7e/sMaRQREREREVFmkq7BIV68eIHu3bsjf/78qFWrFmrVqoUCBQqgR48eiIuLy+g2EhERERERGVS6EidfX18cOnQIO3bsQFRUFKKiorBt2zYcOnQIP//8c0a3kYiIiIiIyKDS1VVv06ZN2LhxI+rUqaOUNWnSBNbW1mjbti0WLFiQUe0jIiIiIiIyuHRdcYqLi4Ojo6NOuYODA7vqERERERFRlpOuxMnd3R3+/v549eqVUvby5UuMGTMG7u7uGdY4IiIiIiKizCBdXfVmzpyJRo0aoWDBgqhYsSIA4O+//4alpSX27t2boQ0kIiIiIiIytHQlTuXLl8f169exZs0aXLlyBQDQoUMHdOrUCdbW1hnaQCIiIiIiIkNLV+I0adIkODo6omfPnlrlgYGBiIiIwNChQzOkcURERERERJlBuu5xWrRoEUqXLq1TXrZsWSxcuPC95zdv3jy4uLjAysoK1atXx8mTJ99ad/PmzahSpQpy5cqFbNmywdXVFatWrXrvzyQiIiIiIkqrdCVOjx49Qv78+XXK7e3t8fDhw/ea14YNG+Dr6wt/f3+cOXMGFStWhKenJx4/fpxq/Tx58mDEiBEICwvD+fPn4e3tDW9vb+zZsyc9X4WIiIiIiEivdCVOzs7OOHr0qE750aNHUaBAgfea1/Tp09GzZ094e3ujTJkyWLhwIWxsbBAYGJhq/Tp16qBly5b44osvUKxYMfTv3x8VKlTAkSNH0vNViIiIiIiI9ErXPU49e/bEgAEDkJiYiK+//hoAEBoaiiFDhuDnn39O83wSEhLw119/wc/PTykzMTFB/fr1ERYWpnd6EcH+/ftx9epVTJkyJdU68fHxiI+PV17HxMQAABITE5GYmJjmtn5MlqZi6Cakm6WJaP1rrD72umDMMQayRpwzy/aeWWmWD5dT1sY4Z32McdbHGGes91mOKhF57yMhEcGwYcMwe/ZsJCQkAACsrKwwdOhQjBo1Ks3zefDgAZycnHDs2DGt5z8NGTIEhw4dwokTJ1KdLjo6Gk5OToiPj4epqSnmz5+P7t27p1p39OjRGDNmjE752rVrYWNjk+a2EhERERFR1hIXF4eOHTsiOjoaOXLkeGfddCVOGs+fP8fly5dhbW2NEiVKwNLS8r2mT2/ipFarcfPmTTx//hyhoaEYN24ctm7dijp16ujUTe2Kk7OzMyIjI/UunE+l3GjjvT/L0kQwrooaI0+bIF6tMnRz0u3iaM+POn9jjjGQNeL8sWNs7BITExESEoIGDRrA3Nzc0M2hj4RxzvoY46yPMc5YMTExsLOzS1PilK6uehq2traoWrVquqe3s7ODqakpwsPDtcrDw8ORL1++t05nYmKC4sWLAwBcXV1x+fJlTJo0KdXEydLSMtWEztzcPNOsbPHJxnkgmlK8WmXU3+NjrwvGvGxSMuY4Z5btPbPLTPtG+ngY56yPMc76GOOM8T7LMF2DQ2QUCwsLuLm5ITQ0VClTq9UIDQ3VugKlj1qt1rqqRERERERElJE+6IpTRvD19UXXrl1RpUoVVKtWDTNnzsSLFy/g7e0NAOjSpQucnJwwadIkAK8fvlulShUUK1YM8fHx2LVrF1atWoUFCxYY8msQEREREVEWZvDEqV27doiIiMCoUaPw6NEjuLq6Ijg4GI6OjgCAu3fvwsTkfxfGXrx4gb59++LevXuwtrZG6dKlsXr1arRr185QX4GIiIiIiLI4gydOAODj4wMfH59U3zt48KDW6/Hjx2P8+PGfoFVERERERESvGfQeJyIiIiIiImPAxImIiIiIiEgPJk5ERERERER6MHEiIiIiIiLSg4kTERERERGRHkyciIiIiIiI9GDiREREREREpAcTJyIiIiIiIj2YOBEREREREenBxImIiIiIiEgPJk5ERERERER6MHEiIiIiIiLSg4kTERERERGRHkyciIiIiIiI9GDiREREREREpAcTJyIiIiIiIj2YOBEREREREenBxImIiIiIiEgPJk5ERERERER6MHEiIiIiIiLSg4kTERERERGRHkyciIiIiIiI9GDiREREREREpAcTJyIiIiIiIj2YOBEREREREenBxImIiIiIiEgPJk5ERERERER6MHEiIiIiIiLSg4kTERERERGRHkyciIiIiIiI9MgUidO8efPg4uICKysrVK9eHSdPnnxr3cWLF8PDwwO5c+dG7ty5Ub9+/XfWJyIiIiIi+lAGT5w2bNgAX19f+Pv748yZM6hYsSI8PT3x+PHjVOsfPHgQHTp0wIEDBxAWFgZnZ2c0bNgQ9+/f/8QtJyIiIiKiz4XBE6fp06ejZ8+e8Pb2RpkyZbBw4ULY2NggMDAw1fpr1qxB37594erqitKlS2PJkiVQq9UIDQ39xC0nIiIiIqLPhZkhPzwhIQF//fUX/Pz8lDITExPUr18fYWFhaZpHXFwcEhMTkSdPnlTfj4+PR3x8vPI6JiYGAJCYmIjExMQPaH3GsTQVQzch3SxNROtfY/Wx1wVjjjGQNeKcWbb3zEqzfLicsjbGOetjjLM+xjhjvc9yVImIwY6EHjx4ACcnJxw7dgzu7u5K+ZAhQ3Do0CGcOHFC7zz69u2LPXv24NKlS7CystJ5f/To0RgzZoxO+dq1a2FjY/NhX4CIiIiIiIxWXFwcOnbsiOjoaOTIkeOddQ16xelDTZ48GevXr8fBgwdTTZoAwM/PD76+vsrrmJgY5b4ofQvnUyk3eo+hm5BuliaCcVXUGHnaBPFqlaGbk24XR3t+1Pkbc4yBrBHnjx1jY5eYmIiQkBA0aNAA5ubmhm4OfSSMc9bHGGd9jHHG0vRGSwuDJk52dnYwNTVFeHi4Vnl4eDjy5cv3zmmnTZuGyZMnY9++fahQocJb61laWsLS0lKn3NzcPNOsbPHJxnkgmlK8WmXU3+NjrwvGvGxSMuY4Z5btPbPLTPtG+ngY56yPMc76GOOM8T7L0KCJk4WFBdzc3BAaGgovLy8AUAZ68PHxeet0U6dOxYQJE7Bnzx5UqVLlE7WWiOjtXIbtNHQTPoilqWBqtddXR401OQaA25ObGroJRESURRm8q56vry+6du2KKlWqoFq1apg5cyZevHgBb29vAECXLl3g5OSESZMmAQCmTJmCUaNGYe3atXBxccGjR48AALa2trC1tTXY9yAiIiIioqzL4IlTu3btEBERgVGjRuHRo0dwdXVFcHAwHB0dAQB3796Ficn/Rk1fsGABEhIS0KZNG635+Pv7Y/To0Z+y6URERERE9JkweOIEAD4+Pm/tmnfw4EGt17dv3/74DSIiIiIiIkrB4A/AJSIiIiIiyuyYOBEREREREenBxImIiIiIiEgPJk5ERERERER6MHEiIiIiIiLSg4kTERERERGRHkyciIiIiIiI9GDiREREREREpEemeAAuERGRMXAZttPQTfgglqaCqdWAcqP3ID5ZZejmpMvtyU0N3QQi+kzxihMREREREZEeTJyIiIiIiIj0YOJERERERESkBxMnIiIiIiIiPZg4ERERERER6cHEiYiIiIiISA8mTkRERERERHowcSIiIiIiItKDiRMREREREZEeTJyIiIiIiIj0YOJERERERESkBxMnIiIiIiIiPZg4ERERERER6cHEiYiIiIiISA8mTkRERERERHowcSIiIiIiItKDiRMREREREZEeTJyIiIiIiIj0YOJERERERESkBxMnIiIiIiIiPZg4ERERERER6WHwxGnevHlwcXGBlZUVqlevjpMnT7617qVLl9C6dWu4uLhApVJh5syZn66hRERERET02TJo4rRhwwb4+vrC398fZ86cQcWKFeHp6YnHjx+nWj8uLg5FixbF5MmTkS9fvk/cWiIiIiIi+lwZNHGaPn06evbsCW9vb5QpUwYLFy6EjY0NAgMDU61ftWpVBAQEoH379rC0tPzErSUiIiIios+VmaE+OCEhAX/99Rf8/PyUMhMTE9SvXx9hYWEZ9jnx8fGIj49XXsfExAAAEhMTkZiYmGGf8yEsTcXQTUg3SxPR+tdYfex1wZhjDGSNODPG75YVYgwwzvpkhThnlt/uzEqzfLicsi7GOGO9z3JUiYhB9p4PHjyAk5MTjh07Bnd3d6V8yJAhOHToEE6cOPHO6V1cXDBgwAAMGDDgnfVGjx6NMWPG6JSvXbsWNjY26Wo7EREREREZv7i4OHTs2BHR0dHIkSPHO+sa7IrTp+Ln5wdfX1/ldUxMDJydndGwYUO9C+dTKTd6j6GbkG6WJoJxVdQYedoE8WqVoZuTbhdHe37U+RtzjIGsEWfG+N2yQowBxlmfrBDnjx1jY5eYmIiQkBA0aNAA5ubmhm4OfQSMccbS9EZLC4MlTnZ2djA1NUV4eLhWeXh4eIYO/GBpaZnq/VDm5uaZZmWLTzbOH6+U4tUqo/4eH3tdMOZlk5Ixx5kxThtjjjHAOKeVMcc5s/x2Z3aZ6TiHPg7GOGO8zzI02OAQFhYWcHNzQ2hoqFKmVqsRGhqq1XWPiIiIiIjI0AzaVc/X1xddu3ZFlSpVUK1aNcycORMvXryAt7c3AKBLly5wcnLCpEmTALweUOKff/5R/n///n2cO3cOtra2KF68uMG+BxERERERZW0GTZzatWuHiIgIjBo1Co8ePYKrqyuCg4Ph6OgIALh79y5MTP53UezBgweoVKmS8nratGmYNm0aateujYMHD37q5hMRERER0WfC4IND+Pj4wMfHJ9X33kyGXFxcYKBBAImIiIiI6DNm0AfgEhERERERGQMmTkRERERERHowcSIiIiIiItKDiRMREREREZEeTJyIiIiIiIj0YOJERERERESkBxMnIiIiIiIiPZg4ERERERER6cHEiYiIiIiISA8mTkRERERERHowcSIiIiIiItKDiRMREREREZEeTJyIiIiIiIj0YOJERERERESkBxMnIiIiIiIiPZg4ERERERER6cHEiYiIiIiISA8mTkRERERERHowcSIiIiIiItKDiRMREREREZEeTJyIiIiIiIj0YOJERERERESkh5mhG0BERESUWbgM22noJnwQS1PB1GpAudF7EJ+sMnRz0u325KaGbgKRDl5xIiIiIiIi0oOJExERERERkR5MnIiIiIiIiPRg4kRERERERKQHEyciIiIiIiI9mDgRERERERHpwcSJiIiIiIhIj0yROM2bNw8uLi6wsrJC9erVcfLkyXfWDwoKQunSpWFlZYXy5ctj165dn6ilRERERET0OTJ44rRhwwb4+vrC398fZ86cQcWKFeHp6YnHjx+nWv/YsWPo0KEDevTogbNnz8LLywteXl64ePHiJ245ERERERF9LgyeOE2fPh09e/aEt7c3ypQpg4ULF8LGxgaBgYGp1p81axYaNWqEwYMH44svvsC4ceNQuXJlzJ079xO3nIiIiIiIPhdmhvzwhIQE/PXXX/Dz81PKTExMUL9+fYSFhaU6TVhYGHx9fbXKPD09sXXr1lTrx8fHIz4+XnkdExMDAEhMTERiYuIHfoOMYWkqhm5CulmaiNa/xupjrwvGHGMga8SZMX63rBBjgHHWJyvEmTF+t6wQY+Djx9mYaZYNl1HGeJ/lqBIRg21ZDx48gJOTE44dOwZ3d3elfMiQITh06BBOnDihM42FhQVWrFiBDh06KGXz58/HmDFjEB4erlN/9OjRGDNmjE752rVrYWNjk0HfhIiIiIiIjE1cXBw6duyI6Oho5MiR4511DXrF6VPw8/PTukIVExMDZ2dnNGzYUO/CIf0SExMREhKCBg0awNzc3NDNoY+Ecc76GOPPA+Oc9THGaVNu9B5DNyHdLE0E46qoMfK0CeLVKkM3J90ujvY0dBMA/K83WloYNHGys7ODqampzpWi8PBw5MuXL9Vp8uXL9171LS0tYWlpqVNubm7OHUoG4vL8PDDOWR9j/HlgnLM+xvjd4pONN+HQiFerjPp7ZJb1833aYdDEycLCAm5ubggNDYWXlxcAQK1WIzQ0FD4+PqlO4+7ujtDQUAwYMEApCwkJ0erqR0RERET0NrcnNzV0E9ItMTERu3btwsXRnpkm+fhcGLyrnq+vL7p27YoqVaqgWrVqmDlzJl68eAFvb28AQJcuXeDk5IRJkyYBAPr374/atWvj119/RdOmTbF+/XqcPn0av/32myG/BhERERERZWEGT5zatWuHiIgIjBo1Co8ePYKrqyuCg4Ph6OgIALh79y5MTP43anrNmjWxdu1a/PLLLxg+fDhKlCiBrVu3oly5cob6CkRERERElMUZPHECAB8fn7d2zTt48KBO2bfffotvv/32I7eKiIiIiIjoNYM/AJeIiIiIiCizY+JERERERESkBxMnIiIiIiIiPZg4ERERERER6cHEiYiIiIiISA8mTkRERERERHowcSIiIiIiItKDiRMREREREZEeTJyIiIiIiIj0MDN0Az41EQEAxMTEGLglWUNiYiLi4uIQExMDc3NzQzeHPhLGOetjjD8PjHPWxxhnfYxxxtLkBJoc4V0+u8QpNjYWAODs7GzglhARERERUWYQGxuLnDlzvrOOStKSXmUharUaDx48QPbs2aFSqQzdHKMXExMDZ2dn/Pfff8iRI4ehm0MfCeOc9THGnwfGOetjjLM+xjhjiQhiY2NRoEABmJi8+y6mz+6Kk4mJCQoWLGjoZmQ5OXLk4Mb7GWCcsz7G+PPAOGd9jHHWxxhnHH1XmjQ4OAQREREREZEeTJyIiIiIiIj0YOJEH8TS0hL+/v6wtLQ0dFPoI2Kcsz7G+PPAOGd9jHHWxxgbzmc3OAQREREREdH74hUnIiIiIiIiPZg4ERERERER6cHEiYiIiIiISA8mTkRERERERHowcSIiIiIiItKDiRMREREREZEeTJyIPlOXLl1CcnKyoZtBRB9ArVZrveYTRoiIPh4mTpQmPMDOWkaOHAk3NzccOXJE58CLsg4eRGdtarUaJiavf8Y3btyIGzduQKVSGbhV9LFwX531McaZHxMn0qI50Lpy5Qr+/PNPBAcHAwBMTU0N2SzKYOPGjUONGjXg7e2Nw4cPMzHOgtRqtXIQHRERYeDWUEYTESVp8vPzw4ABA7B9+3a8fPnSwC2jj0UT71OnThm4JfQxpDwRsnbtWpw8edLALaLUqISnJOn/iQhUKhU2b96MYcOGwdzcHGZmZnj16hW2bduG0qVLG7qJlAFevXoFKysrAECdOnXw6NEjLFiwAB4eHjAzMzNw6ygjpPwBnjp1Kh48eIAuXbqgcuXKBm4ZZbTx48dj1qxZ2L17N8qUKQMbGxtDN4k+ohMnTsDd3R1Hjx6Fu7u7oZtDGURz/AUAw4YNw4YNG9C9e3f0798f2bNn55XkTIRXnEihUqlw5MgRdOvWDUOGDMHFixexePFiXL9+HaGhoUo95trGS61WK0nTzp070aVLF1y7dg2+vr44cuQIrzxlEZqkaciQIQgICEDNmjWRL18+rTrcjo1fVFQUjhw5gl9//RVVqlTB06dPceDAAXTs2BELFy7Ev//+a+gmUgYrUKAAateujTNnzgBgN/qsQpMYTZo0CUuWLEFQUBD8/PyQI0cOJk2ZDBMn0nL+/Hm0b98e33//PW7fvo22bdvihx9+QL9+/ZQ6KpWKB11GSnNAPWLECHTr1g3x8fH45ZdfICLstpfFBAUFYd26dQgNDUXbtm1RoEABxMbG4uLFiwC4HWcFJiYmuHz5Ms6ePYuDBw9i4MCBGD58OO7du4dffvkFmzZtAsAk2Vildr+Ls7MzatSogQkTJiAuLo7d6I1cyhi/evUKBw4cwIwZM1ClShXcv38fwcHBaNWqFcaOHYs7d+4YsKWkwcSJtFy4cAFRUVGIiIhA7dq14enpiXnz5gEA1qxZgwkTJgAAz4AYsVu3bmH16tVYsGAB+vTpg7Fjx+LkyZNwcnJC9+7dcfjwYSQlJRm6mfSBnj17hkKFCqFChQq4du0aAgIC4OrqiubNm6NTp04AuB0bk9QOonPkyIEJEyZg1apV8PLyQrFixTBhwgT8+eefaN26Nf7++28AjLOx0pzounXrFmJiYpTyoUOHokCBAliwYAFEhImxkUp5n+LatWthamqKmJgYbN26Ffv378ePP/6IcePGQaVSYeLEicqxGBkWE6fPmGZne+/ePTx8+BAA0KJFCzx79gxlypSBp6cnFi1apOyYT5w4gdu3byMuLs6QzaYMkJSUBAcHBwBAQkICLCwssGPHDiQlJWHEiBHYt28fR/cxIqkdOFlbW+PZs2fw8vJC06ZNcf78efTs2RNjxoxBSEgITp8+bYCWUnqkPMBaunQphgwZglmzZuHmzZvo3Lkzzp07h9OnT2Py5Mn4+uuvoVarcfPmTRQqVMjALacPFRISgtKlS6N79+7Ytm0bkpOTkStXLtSoUQN79+6FSqXi1WMjlPKepoCAAHTv3h2XL1+Gn58frl69Ci8vL5QrVw4TJ07Epk2bMHToUNy8eZMnNTMB3gn+mdJstFu3bsXEiRPRq1cvtGvXDmXLloWJiQlsbW3RtGlTAK/70f/666/YsGEDDh06xJuPjUjKnbNGkSJFkC1bNqxYsQK1atWChYUFkpOTYWlpiWLFiuHQoUNYsmQJGjVqZKBW0/tIORDE/fv3ERsbi2LFiuG7777Dq1evcOzYMYwcORJ16tRBoUKFcOHCBRQqVAi2trYGbjmlRcr4Dhs2DIGBgShdujSioqKwevVqLFy4EG5ubgCA2NhYnDlzBtOmTUN4eDjGjx9vyKZTOly/fh358+eHra0tJkyYAB8fHyxbtgynTp1C27Zt0axZMzRq1Ag//vgjqlatinXr1qFDhw68qmhkNPEKCwvDnTt3sH37dlSoUAHlypVDgwYNEBERgcKFCyv1jxw5ggoVKnAAp8xA6LO1c+dOsbKykhkzZsh///2nlF+7dk1q1Kgh5cuXl0KFCkndunWlYMGCcubMGQO2lt5XcnKy8v8bN27I3bt35caNGyIisnbtWilcuLD4+fkpdRITE6Vbt25y7do1rWkp81Kr1cr/f/nlF6lRo4bY2tpKs2bNZOzYsVp1k5OTJTo6Wpo3by5ff/01Y2xkrl+/Lr169VL2wwcPHhQvLy8pWbKknD17VkRE9u/fL9988400atRIEhISREQkKSnJUE2m93TixAlxcnKSFStWiI+Pj6hUKrl+/bry/smTJ2XYsGFStGhRqVChgtjZ2UmnTp0kISGB27MR2rZtm5QvX16KFCkily5dEhFRtlsRkefPn0toaKg0btxYypcvL4mJiYZqKqXAxOkzpFarJTY2Vho2bCgjR47Uek+z0d6/f1/27dsnY8aMkS1btsjt27cN0VRKp5QH1KNGjZJq1aqJi4uLVKtWTRYsWCBqtVoCAgKkQIECUq9ePfH19ZWaNWvKF198ofwA84DLeIwfP17s7Oxk7969cufOHWnVqpXkzp1bOch+8eKFzJ8/Xzw9PaVSpUrKds6DLeOwfv16KVKkiNSoUUMiIiKU8rCwMPHy8pLSpUvLhQsXRETk3LlzSlx5oGUczp07p/y/b9++kjdvXrGxsZGjR4+KyOt9sSamycnJEh8fL+PGjZMmTZqIlZWV1vSUeb25vz1x4oS0a9dOrKys5Ndff9Wpt3//funSpYs0bdpU2WdzmzY83uP0GVKpVLCwsMC9e/dQsGBBAP+78djc3BwAYGVlhXr16mHUqFHw8vLSumRMmZcmjppuAGPHjsXcuXMxceJE7N69G8WKFUPfvn1x//599OzZE+vXr4eVlRVu3bqFYsWK4e+//4aJiQnUajVHa8rE5P/vZxARPHnyBPv27cOiRYvQoEEDXLt2DXv27EFAQAAqVaqExMRE2NjYICkpCZUqVcLJkydhbm6OpKQkpQsYZW4iAhcXF1y5ckXrHtMaNWpg6NChKFOmDDw8PHD9+nVUrFhR2YbZrSfz69WrF1atWoXExEQAr2P66tUr5MmTB9euXcOzZ8+09sUiAgsLC/zyyy9Ys2YNmjRpgtmzZyMxMZH3OWVymv3tmjVrEB0djWrVqmHkyJFo2bIl1qxZgxUrVmjVq1q1KoYNG4bt27cr+2xu05mAQdM2Mojk5GSJi4sTFxcXGTp0qFa5yOsuIQsWLJDHjx8bqomUAaKiosTT01O2bdsmIiI7duyQXLlyyYIFC0Tk7VeUeEYrc0t51jImJkZiY2PFzc1N/vnnH9m2bZvY2toqMX716pX89ttvytUIDV5NzLzedhVw586dUqVKFXF3d5ebN29qvXfo0CEZNmwY42qEwsLClKsJjx8/lujoaImIiJC+fftKsWLFZP78+fLs2TOd6TSxHjdunNSvX/9TNpk+wLVr16REiRLy1VdfSUxMjIiInD17Vjp16iRffvmlrFy5MtXp2Dsg8+Dpxs+A/P9ZqOfPnwN4fTXC2toa/fv3x9KlSxEYGAjgf2c5Fi5ciPXr1ytXnyjz++abb+Dv769VFh8fj7Nnz6JgwYLYs2cPOnTogEmTJuGHH35AfHw8pk2bhr/++ktrGhHhGa1MTrOd/vTTT/Dz88PTp0+hUqkwevRoeHt7Y+rUqfjhhx8AAHfu3MHmzZtx7do1rXnwamLmlHIgiCNHjuDAgQPKw8ebNGmCsWPHwsrKCt26dcPt27eV6WrVqoVJkybB1NSUz2EzIiKCGjVqwNzcHMuWLUO7du1w5swZ2NnZYd68efj666/x66+/IigoSBmOvHv37nj8+LGyDcfFxeG///5DbGwsrzhlQm/GpEiRIpg6dSrUajWaN2+OmJgYuLq64ueff0aRIkWwZMkSLFy4UGc+7B2QiRg0baNP5o8//pBmzZpJnTp1JDAwUMLDw+Xp06fy448/Sp48eeTHH3+U8ePHS48ePSRnzpzsM21Enj9/Ljt27JD4+Hit8pcvX0qnTp2kR48ekiNHDlm0aJHy3o0bN6R58+ayadOmT91cSqeU961dvXpVSpQoIYcPHxYRke3bt4ulpaW0atVKqRMTEyNNmjSRunXr8kqEkRk0aJA4OTlJkSJFxMrKSlq0aKFcNfzjjz+kXr16UrduXa2BA8i4pNyeRUT27t0rVapUkXbt2kloaKhS3qtXLylRooR4e3vL119/LXnz5lV6Bdy5c0fat2/PgZuMTFJSkmzdulWqVasmtWvX1rry1KRJE/nhhx8M3EJ6FyZOn4GjR4+KlZWVDB48WBo1aiQVKlSQ3r17y8OHDyUuLk4CAwPF1dVVvvzyS2nVqpVOtx7KvO7cuSMi/7uMP2PGDGndurXy/qRJk0SlUsl3330nL1++FBGRZ8+e8YDaiE2cOFH69u0r/fr1E7VarfzNmDFDVCqVNGvWTJo0aSK1a9eW8uXLc3Q1I7Nw4UKxs7OTkydPyq1bt+Ts2bNSrFgxqVevnjJIz5YtW8TV1VX69u1r4NbSh5oxY4YsW7ZMRER2794tNWrUkDZt2mglT/7+/tKlSxfp2LGjzsAuz58//+RtpvezdOlS+eabb7T2wUlJSbJlyxYpW7asNGrUSGJjY0Xk9UkxTWzfTK4pc2DilMXdvn1bRo8erTViy9y5c6VGjRry/fffKz/EmqsVr169Mkg76f1pkqKLFy+KyOsrTL/99pvY2dmJt7e3Um/AgAGSJ08eady4sbRp00a++uorqVixIkdWM0Lx8fHSv39/UalUUrt2bZ33Dxw4ID4+PuLj4yMzZsxQzkzzvrXMaefOnRIXF6dV1rdvX+ncubOI/G/bvHPnjjg6OkqvXr2UeocPH+a2mwV06NBBihUrpsRyz5494u7uLq1bt9ZKnlL+NicmJvKg2kgkJibKjBkzpEKFCtKlSxedE1gDBw4UlUolrq6uWkkwt+3Mi4lTFnb16lWpXr26FCpUSObNm6f13ty5c6V69erSq1cvuXz5slLOnbHxuHPnjrRo0ULs7e3l/PnzIvJ6QIhVq1ZJ/vz55bvvvlPqBgYGytChQ6VHjx4yffp0HlAbidS2x4iICBkzZoyYmJjI8uXLlXpvO0vJK02Z0+jRo6VFixZa8UpKSpJvvvlGvLy8lDLNAfPy5culUKFCcv/+fa35ML7GSbO9Xr58WapWrSqLFy9W3gsJCZGaNWtK27ZtZffu3VrT8Tc6c0st4YmNjZXffvtNXF1dpXPnzlrPavrtt9+kRYsW4uvry23ZSDBxyuKGDx8uDg4O0qpVK63nf4iILFiwQEqXLi0//vgjD6CNVHh4uLRr107s7e2VBDg6OlpWrlypkzy9iTvpzC3lD/Djx4+1RlJLSkqSQYMGiYmJiWzYsEFE/ndAxTOVxkOz3z137pxyn8OGDRvEyspKfv/9d626y5cvl8qVKytdeihrePHihbRu3VqaNWumVb5v3z4pXry4jBgxwkAto/eVct974sQJOXjwoHL/WWJioixevFgqVaok7du3l5iYGHn27Jm0a9dOAgIClOn4u5z5MXH6DPj7+0v58uXll19+kfDwcK33Fi9eLLdu3TJMwyhdUu6cV6xYIVOmTBGVSiXOzs5Ktz1N8lSgQAHp3r27oZpK6fTmA4wrVaokuXLlkq+++krmz5+vdO/6+eefxdTUVOcgmzK3lAdHmzdvFgcHB1myZIk8f/5cYmJipG/fvlK0aFFZvXq1vHr1SsLDw6VJkybyzTff8IqDkZk8ebLcuHFDeb1y5UoZNGiQVne7S5cuia2trc5Q1KdOneKBtBEaPHiwODk5SeHChcXU1FS6du0qZ8+eFbVaLStWrJBy5cqJtbW1lClTRsqUKaOcQOG2bRyYOGURKXfAO3bskIMHD8p///2nvO/n5yeVKlWSESNG8PlMWcSQIUOkYMGCMmvWLBkwYIBUqlRJ8ubNq3Tbi46OllWrVolKpZIJEyYYuLWUHuPHj5e8efPKqlWrZO/evdKuXTupUaOGjBgxQl69eiXx8fEybNgwUalUWvdDUOaV2hXBtm3bStmyZWX58uWSlJQkt27dEl9fXzE3N5fChQtLyZIlpVKlSrwv0cj8+eef0qBBAyX5efnypfj4+EipUqWkSJEiMnnyZOWKRJ8+faRHjx4SHR2t0wOEyZPxWLBggdjb28uRI0fk/v37EhwcLK6urtK6dWu5cuWKJCcny71792TBggWyfPlyJdaMsfFg4pQFaJKmTZs2Sb58+aRixYpSoEABadeunVb/aD8/P6lWrZoMGDBAp9seGZebN29KkSJFtIYTv3LlijRq1Ejs7e3l0qVLIvJ6BL3du3dzp2xk1Gq1REZGiru7u9a9D/Hx8fLLL7+Iq6urBAcHi4jI06dPZf78+exuawRSJjwbN27USnY7d+4sJUuWlGXLlikJ0rlz52TVqlWyfft2ZRtmnI2L5vd5+/btyiioarVahgwZIo0aNZLcuXPL3Llz5eeff5YCBQpwaHEjpdk+vb29pVu3biLyv9gfPnxYChcuLMOGDXvntGQcmDgZsZQ/wvv27ZO8efMqg0CsX79ecuTIIR4eHrJ161al3k8//SS1a9fmVScj8+Yl/AsXLoiNjY0cO3ZMKUtOTpYzZ86Ig4ODFC9eXOdZXNw5Z25vxjguLk4qVKigjIiZMn6VK1fWGjlRgwfVmVfK+A4ZMkSKFy8uAQEBWt2nO3ToICVKlJBly5ZJdHS0zjy4DRsHtVqtNQDArVu3xMrKSry9vZXu1CIid+/elUWLFkm5cuWkUaNGolKp5McffzREkykdQkNDZfbs2VonNdq0aSMdO3ZUXmvemzlzpjg6OsrTp095xdjI8VHERmjlypWIj4+HiYkJ1Go1Xr58iY0bN8Lb2xt9+/bF3bt3MXz4cHh4eMDExAQTJ05EcHAwAGDWrFkICgqCvb29gb8FvQ+VSgUASEhIAACUK1cOFStWxPLlyxEfHw/g9ZPFv/jiC5QpUwYPHz7E0KFDteahedI8ZT5qtVqJ8b179wC8jleePHlw4MAB5XVycjIAoGbNmnj58qXOfMzMzD5Ri+l9aeI7adIkBAYGYvXq1fD19YWDg4NSZ+3atahRowYCAgKwcuVKvHr1Smse3IaNQ0xMDMzNzQEAO3bsQOHChbF+/XocOHAAs2fPxoULFwAAzs7O6NWrF7Zu3Yo+ffqgR48emD59uiGbTmmUmJiI+fPnIzAwEEuXLkVycjLMzMxQr149rFu3DidOnICZmRlMTF4fZufIkQNFixaFtbW1UkZGytCZG72fR48eSZ48eaRatWrKs5dEXj9x+syZMxIVFSWVKlVSBgQICgoSa2trKV++vNaVJzI+EydOlF69esmzZ8+U1zVq1JCpU6cqdWJjY6Vly5Z8xosRSRmnMWPGSO3atSUsLExERP7++2/Jnj27/PDDD/Ly5UtJSkqSxMREcXd3lwEDBhiqyZROERERUq9ePVm1apWIvH7O3u7du6V9+/bi5+en1GvcuLF06NCBN4sboUOHDom9vb08ffpUBg8eLIULF1aGkN+6dasULFhQevbsqXSnTk3Kq1WUeT19+lQ6deokNWvWlHnz5ilXlzp06CA5c+aUffv2SWRkpERFRYmnp6fO4wfIOPH0pJFxcHDAnj170LlzZ9SuXRuHDh2ChYUFypQpAwsLC2zZsgXm5uYYO3YsACBXrlxwdXVF+fLlUalSJQO3nj6Ei4sLRowYAVtbWwQEBGDAgAF4+PAh1q9fj507d6JOnTrYs2cPkpKS4O7urlyR5NmtzE0TnxEjRmDp0qWYM2cOChQoAACoUKEC1q1bhw4dOuD06dPIlSsXXr58iejoaAQEBBiy2ZQGb25/dnZ2iI+Px/bt2+Hg4IB58+YhMjISjo6OmDlzJmJiYjB37lzs2rVLuQopIsrVKsr8smfPjmrVqqFkyZJITk7GhQsXUKBAAajVarRo0QIA4OPjA5VKhQEDBuCLL77QmYfmahVlXsnJycidOzcCAwPRq1cvLF26FGZmZujZsydmzJgBGxsbNG7cGIULF4aZmRksLS1x6tQpbtNZAI+ojIjmh7RKlSpYs2YNIiMj4eXlhfj4eFhYWAAA4uPj8eDBA9y9excAcODAAdSoUQNTpkxBoUKFDNl8eg9qtVrrtYigQ4cO2Lx5M2bNmoWBAwfC2toaU6dOxdChQ+Ho6Ihjx46hWLFiOHbsGExNTZk0GZFLly5h48aNWLp0Kb799lutbbVp06a4dOkSGjZsiLJly6J+/fr4+++/YWZmhqSkJAO2mt4l5fa3c+dOHDlyBADQvXt3/Pfff/Dy8kLZsmUxadIkbN68Gb6+vnj69KnSHVdz4oMHWMZBRAAAlSpVQrly5fDkyRNYWVkp8UtOToaIoEWLFpg7dy727t0Lf39/3L5924Ctpvfx6NEjnTILCws8efIEV65cQWBgIBYvXgx7e3ssWbIEu3btwtixYzFmzBj89ddfMDc3R1JSErdpY2fQ6130XjSXeHfu3Cne3t5SrVo1UalUUrduXaXb3tGjR8XDw0PKlSsnX331ldja2irDU5PxOXHihPL/lKMnmpqayoABA+TVq1fK+yn/z0ECMjdNlw5NTA8cOCD58uVTuvSo1WrlvZRxTW0elPm8ORBEqVKllOdvvXz5UmJiYuT69eta09SpU0cGDhz4qZtKGSBld9vIyEg5evSo7Nq1S1q0aCHOzs5y9epVEXk9HLnGli1bxMvLi12qjcSpU6fExMREtm/frrV9t2rVSlxdXeXcuXPSuXNnqVGjxltHOeU+O2tg4mRkQkJCxNzcXObOnSt79uyRuXPnirOzs9SsWVPpF717926ZOHGiDBw4UC5fvmzgFlNaqdVqrR3rmTNnRKVSybRp07TqiIisWrVKTE1NZcyYMTpDy7MPdeYWHBwsfn5+EhkZqZRduHBB7O3tZfv27UqZZl14c9hqMh7jx49XnumS2kFTTEyMHD58WDw9PaVChQo84WGEUiY+kyZNkkGDBsmpU6dEROT06dPSpEkTKVSokPz7779KvaVLl2rdo8zkyTh89913kitXLtm/f7+IiLRu3VrKlSunxPbp06fSsWNH+eqrryQgIIBxzaKYOBkZPz8/8fLyUl4nJyfL0aNHxcnJSerUqcObSo2YZtAHEZGDBw/K8+fPZcaMGWJhYSHTp0/Xqnv79m3Jly+fqFQqmT179iduKaVXYGCgFCpUSDp27Cj79u1Tyh8+fChfffWVtG3bVk6ePKmUJyUlSf369cXHx8cQzaUPEB4eLjVr1pQ1a9aIyOuhp/ft2yc9e/aUsWPHisjrJLpTp07SpEkTZd/Ns9LGaciQIWJnZydr166Ve/fuKeVnzpyRpk2biqOjowQFBUn9+vWlSpUqPKg2Ut9//71kz55datSoIa6ursqzuTQnPZ4+fSqNGjWS3r178yRmFsXEych0795dKlasqFWmVqtl6tSpolKppGrVqlpnssg4HDx4UCpXriyRkZEycOBAKVasmPKsrZkzZ4qJiYlMnz5d2RE/evRIhg4dKnv37uVZaiOxdu1ayZ49u6xbt07i4uJ03g8ODpYyZcpIgwYNZMyYMbJkyRKpXbu2lC9fnjE2QvHx8VKrVi3p2bOn7Nq1S1q1aiU1atQQT09PsbKykhEjRoharZZz584pB9GMs3Hat2+fuLi4yPHjx5WylAfN//zzj3Ts2FGKFSsmjRo1UpJkHlgbp4EDB4pKpZLAwECtcs1Jj9jYWGWbZoyzHo6qZ2Q6dOiAI0eOYOXKlejSpQuA188HKV68OGrXro2XL1/i/v37KFKkiIFbSu8jNjYWDg4OqFy5MmJjY3H27FnY29tDRNC/f39lBKZ///0Xrq6u2LZtG169eoXJkycDAJKSkvgMn0zsyZMnWLhwIcaNG4f27dsr5c+fP8c///yDnDlzwtPTE+vXr8eCBQuwZMkSFC5cGE5OTggJCYGZmRmSk5P5HJ9MKrWBWMzNzdG8eXNs2rQJK1aswIABA/Djjz+iTp066N+/Px48eACVSoWKFSsq8+A2bJyePHkCa2trFC1aVGfENBHBF198gTVr1uDu3btwdnaGSqXiPtuITZs2DXFxcfjpp5/g6OiIJk2aAIAyKJOtrS2A1PcLZPy41WZSmp3v1atXERERgUKFCqFQoULKsOJr166FWq1Gt27dkJycjBMnTuCLL77A9OnTYWVlZejmUxo9e/YMuXPnRrNmzbBr1y7s2bMHFSpUUA6QNevBTz/9hHz58mHkyJE4evQo7OzssHv3bqUOf4AztxcvXuDGjRvKQTIALFq0CLt378b27dtha2uLFi1aYPr06Zg/fz5+/fVXiAhsbGwAMDHOzFIeHK1cuRLnz5+HiMDDwwODBg3C999/j8ePH6NkyZLKNOfOnUP16tW15sMDLOOSMkF69OgRIiMjlQfLJyYmwtzcHCKC0NBQ2NraokaNGspomUySjZuJiQkWLlyI5ORktG/fHhs2bEDjxo2V91LWo6yHUc2kVCoVNm/ejOrVq6NLly4oXrw45syZA3t7e0yYMAF58uTB+PHjUaJECdSrVw/z5s1Dnz59mDQZkR07dqB3797YtGkTAKB+/fpYvHgxChUqhG+//RZXrlyBiYkJEhMTAQBt27bFiRMnsH//foSEhHBoUyNib2+PggULYubMmTh//jxatWqFefPmIX/+/Dhx4gRmzZqFQ4cOYe/evQAAa2trJWliYpy5aQ6OhgwZgqFDhyrP2erWrRuGDh2KXLlyoWTJknjx4gVOnz4NT09PREVFYeLEiQZuOb2v69evIzY2FsDr32jNYyPatGmDbNmyKb1ANM9hev78OX799VecOXNGaz48oM4aFi9ejI4dO6Jp06Y4fvy4oZtDn4qBugjSW2j6w96+fVsqVaokCxYskJs3b8q4ceMkW7ZsMmbMGElMTJQnT55IWFiY+Pr6ypQpUzh6npFZunSpODg4yC+//CI7d+7Ueu+PP/4QT09PqVGjhjKMrcjr4WtTjsTGm4uNR3JysqxevVrKly8v+fPnl4oVK8q+ffu0RkQsW7asDB061ICtpPTas2ePuLi4SFhYmIiIrFu3TqytrWXp0qVKnU2bNkmbNm207nHhQBDGY+fOnaJSqaRfv34yYMAASUxMVOIYFxcnCxYskNKlS0vz5s3l9OnTsn37dmnSpIlUrFiR965lcZMnT2aMPyNMnDKhffv2yfTp06V3795ao+RNmzZNsmfPLmPHjtU6gCbjsnHjRsmVK5f8/vvvbz1w2rNnj3h6ekqlSpVk165d4unpKdWrV+eNpkYsMTFRoqOj5cKFCzrvPXz4UGrWrCkrVqwwQMsovTQnLxYvXix16tQRkdcJUvbs2WXhwoUi8vpG8ZMnT0pycrKcOHGCA0EYqSVLlohKpZLNmzdLrVq1xMPDQwYOHKg8J/Hly5eyY8cOqVixouTJk0fKlCkjjRs3ZpJsJFL7bX3f31uOavx5YP+PTOiPP/7ArFmzUKpUKTx79gwODg4AgJ9//hkqlQoTJkzAy5cvMWDAAOU9Mg7x8fHYunUrfH198e233yrlt27dwoULFxAeHo4GDRqgYcOGsLa2xvTp09GrVy+UKlUKhw8fhkql0rn5mIyDmZkZcuTIgXLlyillIoLo6Gj07NkTKpUKnTp1MmALKS2OHTuGq1evomPHjrC0tAQA5MqVC87Ozvj999/Ro0cPTJs2Db179wYA/Pnnn9izZw+KFi2KatWqAeA9LsaoR48e+OOPPxAZGYlDhw5h7ty5uHXrFqpWrYp+/fqhbt26aNasGZo1a4a7d+9CpVKhYMGCHAjCCKS8TzEiIgIJCQlwcnJSfmffNshDyvK4uDilezVlcYbO3Ch1o0ePFpVKJfPnz5cXL15ovTdu3DgpVKiQzoNPKfN78eKFlC5dWoYPH66UBQQESKNGjcTW1lZsbGykbNmySpefly9fyo0bN3iWOguKioqS6dOni6enp1SuXJlnpo3AihUrpESJEvL999/LsWPHlPKTJ09K9uzZRaVSybx585TyuLg48fT0lB49evBqsRHTPJzc399fWrVqpfVetmzZxM3NTezs7KRx48Yyc+ZMefnypfI+u1QbjxEjRki5cuUkZ86c4uXlJStXrlTeezOOKbfn5cuXy8CBA7XiTlkX71A0MBEBANy7dw///vsvLl26BADw9/fHTz/9hAEDBmD9+vV4+fKlMs0vv/yCs2fPws7OziBtpvSzsrJCkyZNsG/fPkybNg2enp5YvHgx3NzccPz4cTx79gxxcXFYuHChUr9o0aIwMTHhWeos5tChQ7h27RpKly6NEydOKIN9cMjxzGnlypXo168fhg8fjgkTJsDd3V15r2rVqliwYAEA4OHDhwgODsahQ4fQokULPHz4EAsXLlSuFpPxUalUMDU1Re/evXHgwAFs3boVAFChQgVUrVoVGzduxKFDh/D8+XOcOnUKFhYWyrQcCCLz0gzuAQDz58/HkiVLMHDgQCxevBhxcXFYtGgRpk6dCkA7jpKi18eiRYvQu3dvfP311xyc6zOhEu7JDUaz8W3ZsgXjx4/Hs2fP4OjoiPz582Pz5s0AgEGDBmHOnDlYtGgR2rZtqzXSFrtrGaejR49i6dKl+Ouvv2Bvb49p06ahWLFiyJ49O4DXXUJevnyJtWvXGrillFbpeV5HYmIioqOjlRMgfE5T5nXjxg20bNkSgwYNUkZOA17vh+/cuQM7OzvY2tpixYoVyr68ePHisLe3x+bNm2Fubs74GjlN/CZNmoRLly7h+PHjyJcvH7Zs2aIMQx4fHw8LCwt2qTYyhw8fxrFjx5AvXz507doVwOtnc40ZMwYnT55EQEAAPDw8AOgmTUOHDsXSpUvRunVrg7WfPjFDXeqi10JCQsTa2loWLFggd+/elcDAQFGpVFo3iQ8aNEhUKpWsWrXKgC2ljJSQkKDTBVPk9Y3ktWvXljFjxhigVZQeKbtw7N69W+lm+SHzoczl+PHjUqhQIa3RS9esWSOdOnWS7NmzS+HChZUR9O7fvy/Xr1+Xe/fuKd152MU269izZ4+Ym5tLrVq1tLrVptx+uS0bj6tXr4pKpRKVSiVTpkwRkf91w4uNjZXixYvLkCFDdKZbuHCh5MyZUzZu3PhJ20uGx2vIBiQiCAkJQf/+/fHDDz/AxMQEY8aMQd++fbXOagYEBMDPzw9ubm4GbC1lJDMzM9jY2ChddxITE/Hw4UO0bdsWL168wPDhww3cQkoLEVGuNA0bNgw//fQTrl69iidPnuidTiM0NBRPnz5ll55MzNzcHDly5MDevXsRFRWF7t27Y8aMGUhMTMT8+fPRpEkT+Pj44PLlyyhQoACKFy+u3FzOLrZZS8OGDdGnTx+oVCqtLvR88KlxkDc6WZUsWRJ//PEHcuXKhWPHjuHZs2fKe7a2tqhVqxbu3r2rNd3y5cvRv39/BAYG8krTZ4hb9yfw5oaqoVKpcOHCBVhaWiIiIgI1atSAp6cn5syZAwBYs2YNVqxYAQCYMGECvvjii0/WZvq4NJf6VSoVoqOjERAQgM6dOyMqKgrHjh2DmZkZkpOTDdxK0kcTx4kTJ2L58uVYunQpOnXqhLx582rVSxlLSdHVY+HChejUqRNu3Ljx6RpN761kyZKoUqUKZs+ejSJFiuDIkSPw8/PD3Llz0blzZ8ycORM2NjY4duyYzrQ8iM78Xr16laZ6mt/yr7/+GtHR0Th79uzHbBZlMLVarex7X7x4gaSkJCQkJKBJkyZYunQpdu/ejREjRuDhw4cQEbx69Uq5n1zT/fL58+f466+/EBQUhFatWhn4G5Eh8DTYJ6BSqRAREQETExPkzZsXW7duRWJiIr799lvUrVsXZ86cQeXKldG4cWMsWrQIIoKXL1/izz//RP78+ZGQkKB1syllLTdu3EB8fDzq16+PIUOGwNTUlMPXGgkRQWRkJHbs2IHJkyfDw8NDGehl06ZNsLOzg7+/P0xNTZWDrtT6x1etWtWQX4PeQURga2uLWbNm4fr164iIiECDBg207le6ceMGChYsiCJFihiwpZQe3333HSIiIhAUFKTcZ/o2mm23RYsW6Nu3L4KCgpR7XyhzS3kf6rRp03Ds2DHcv38f1apVQ58+fdCyZUts2LABHTp0wJEjR1C6dGkkJCQgOTkZv/76qzIfW1tbTJ06FdbW1ob6KmRgHBziIxMRxMTEoFixYujbty9cXFzw/fffY82aNejQoQOOHz+Oli1bIkeOHNi+fTtKlSqF+Ph4jB07FitXrsT+/ftRokQJQ38NSqOUO+f3SX5evHiBbNmyvfd09Om9ORBEXFwcmjZtisqVK+PLL7/EunXr8ODBA5ibm+PKlSto27YtZs+erTXdokWLMGTIEHb1MBLvGvzjxYsX6NChA16+fIng4GAOAGFkDh48iNatW6NRo0ZYuHCh3uRJM0jEuXPnUK5cOe6rjYyfnx+WLFmCqVOnIiEhAfPnz0dCQgJOnDiBHDlyYOfOnejUqRMKFSqE2bNnw8PDA6ampkhMTIS5ubmhm0+Zwae+qepztWnTJjE3NxcTExPlOR+aGxD37t0rOXLkkDp16oiHh4e0bNlS7O3t5cyZM4ZsMr2nlDcEr1ixQjZt2iRRUVF6p+PzXYzT1q1b5dGjRyIiMnz4cPnqq6/EzMxMhg0bJocOHRIRkV69eomPj4/WdHPnzpVcuXLxpmIjFxUVJZs2bZJGjRpJxYoV+RwuI3bs2DHJlSuXtG/fXmJiYt5r2qdPn36kVlFGu3Tpkri6usrhw4dFRGTnzp2SPXt2WbRokYj87zd8+/btYm5uLv3791ee4UWkwc7XH0HKZwPEx8cDeP28B+B/XXsiIyOVPrMNGjRASEgIWrRogSJFiqBOnTo4evQoKlWqZJD20/uTFIMEDB06FEOGDEFsbKwS/3dNp+n+sX37duzZs+ejt5U+3IMHD9CyZUv069cPL1++xIQJE7B69WpcvHgRkyZNQq1atQAA165d0+rSceLECQQEBGDRokW80mTERATz58/HqlWrkC9fPpw+fZrP4TJi7u7u2LlzJ4KDg9GrVy/Exsa+tW7K3/fly5dj1KhRWoNEUOb1/PlzREZGwt3dHdu2bUO7du0wdepU9OrVC3FxcVi1ahWePXuG5s2bIygoCEuWLEHPnj15vzFpM2jaloXdvXtXbt++LSKvz14EBgbKxYsXJSgoSFQqlfj5+UlkZKSBW0kZbfr06ZIvXz45ffq0VnlcXJyIaF9dSvn/+fPnS65cuZQrFZS5pHZVMCwsTHLlyiVt27bV2pZjYmLk7Nmz4unpKRUqVNAaivrOnTty4cKFT9JmSrv0DB8dGxsr//zzj7Ju8Ky08XjbVf6jR49Kzpw533rlKeV0CxcuFEtLS9m+fftHayelX8pYabbvf/75R+rVqyezZs2SHDlyyMKFC5U6x48fl44dO2r19AkKChJHR0elZwGRiAgTp48gNjZWvLy8pFq1ajJv3jxRqVTy+++/K++vWLFCVCqV/PLLLxIRESEiIlOnTpVNmzYZqsmUTm8ecHXt2lUGDRokIiI3b96UjRs3Sv369aVjx45y8OBBpd6bP8C5cuWSoKCgT9NoSjdNvDXxO378uNja2kqHDh2UH9fff/9dmjRpIo0bN1a6b/E5PsbhzRMeb5PagTe73BqHlPvs//77T27duqX1/tuSp5TTaZ7hw9/szCllrJKSkuT58+fK6wYNGohKpZLx48crZXFxcdK4cWP55ptvdH7TY2NjP36DyahwcIiPJCQkBAMHDsTVq1cxZcoU+Pr6IiEhAebm5lCpVFi5ciV69OiBtm3bQqVSISgoCCdOnICrq6uhm07psHLlSpQqVQrTp09HQkICatasiZCQEJibm8Pa2hqJiYl48eIFtmzZAhsbG6U7z2+//YbBgwdzkAAjMHnyZNy9exezZs2Cubm50s3y+PHjqF+/Ptq1a4dp06Yhd+7cOHbsGGrUqAETExMO9pGJpRz04ezZs3Bzc8OyZcvQtWvXd04nKbrY/vPPPyhSpAhH2TICKeM9duxYBAUF4fnz58iWLRuWLl2KihUrwsrKCkePHkXz5s3RpEkTzJ8/Hzly5FDmkXI0TO6zM5+UMf7111/x559/4saNG2jQoAF+/vln5M6dGx4eHkhKSkLbtm1hbW2NXbt2ITw8HGfPnoW5ufk7B4Mh4hWnDJBaN487d+5IiRIlpESJEvL111/LzZs3RUQkISFBOTP5+++/S4sWLaR58+by999/f9I204dJGfMpU6ZIzpw55cGDBxIcHCz16tUTFxcXmThxopw8eVJERKZNmybNmjXT6Z5na2vLs5ZGYtWqVaJSqWTo0KHKlSTNejBu3DhRqVTSokULrTOU6ekCRp9Gym1x5syZMmLECDExMRGVSqXcLK5vurlz50q1atWU/TsZh1GjRkn+/Pll/fr18uDBA3Fzc5OyZcvKtm3b5NWrVyLy+sqTSqUSf39/Zbrly5eLpaUlB3YxAn5+fpIvXz6ZPn267Ny5U0xNTZX9c3R0tHTu3Fm+/PJLqV+/vvTt21fpFcDeAaQPE6cMcvnyZRk+fLjcvn1b1Gq1qNVquXnzpuzevVvq1q0rtWrV0kqeNJKSkpQdNRmff/75R/z9/bWSn2fPnildMEVeH2g1adJEvvvuO63pqlevzu55mdTbEp7ff/9dzMzMZNCgQRIfH6+Uz5kzR9q3by8NGjRgsmRkRowYIQ4ODvL777/LkiVLpHv37qJSqWT+/Pk6dd/sYpsjRw7ZsGHDp2wufaDjx49LtWrVJCQkREREgoODJWfOnFKuXDnJnTu3bNu2Tbkn9fz588qB9IsXL2TgwIGyY8cOg7WdUvdmN9nz589L6dKlle7xJ06cEAsLC1m6dKlWvbi4OK39OJMmSgsmThkgISFBqlatKiqVSkqUKCG+vr5aB8R79uyRWrVqSZ06dZTkadq0abJw4UIeZBmxgwcPikqlEhsbm1TPQEZFRcmOHTukSZMmUq5cOa2EWUR4ljqTSrlNnj59Wnbv3i1Xr15Vhh1et26dmJubi6+vr1y/fl25p3H9+vWpzoMyr8jISKlSpYrWFaYnT56Iv7+/qFQqrQOtlIM/aJImXi3O/N7cFi9evCgLFiwQEZHQ0FBxcHCQ3377TUREXF1dpUyZMrJ+/Xqt/bXm/y9fvvxEraa06tu3r4SGhmrF+dSpU1KpUiURef0oGFtbWyXmMTEx8scff+jMh/coUloxccogU6dOlenTp8vevXvF399fcufOLR06dJAlS5aIWq2WXbt2yTfffCMFCxaULl26iEqlYvc8I5PawfCUKVNEpVLJqFGjdM5WXb58WVq0aCHt2rXT6gbAg2rjMGTIEHFxcZG8efNKiRIlpF69enL16lURef0MJxsbG3FxcREXFxed0fPIOISHh0uuXLlk7ty5OuUeHh6iUqlk2bJlIvK/A6sFCxZIzpw52V3LCKTc16b8vX38+LGo1Wpp2bKlDBw4UNRqtcTHx0uLFi0kZ86c4unpaYjmUjoULVpUihcvLkeOHFHiffnyZSlcuLCMGzdOcubMqXX1+MiRI1KvXj05f/68oZpMRo6JUwY5cOCA5MiRQ06dOiUiIg8ePJDRo0eLpaWl1KpVSxYvXixLliyRiRMnStu2beXixYsGbjGl1+rVq5UH6ImIjB49WkxNTSUwMFCn7r1795SdOQ+sM7eUZxwXLFggefLkkQMHDsj9+/dl/fr10rhxYylbtqxcv35dRF53twwMDJSlS5cqseWQ1JnX205Y9OjRQ+rVqyfXrl3TKu/Tp4989dVXYmpqKtu2bROR12evc+fOzS62RiDl9jxy5EgpX768EkeR16OlVa5cWaZPny4ir9ePzp07y+3bt3lyy8jUqlVLihYtKocPH5aEhARJTk6WXr16iZWVlQwYMECp9+rVK2nevLl4eXkxxpRuHFUvAw0ePBgPHz7EkiVLYGVlhfbt2+Pvv/9GlSpV8PDhQ/z5559YtGgROnXqBAsLC0M3l9Lh+fPnKFKkCEqXLo1p06ahevXqAIBRo0Zh4sSJWLJkCbp166YzHUfpybx27dqFJk2aKK+Tk5PRq1cvZMuWDbNnz1bKw8LCMGLECJQuXRozZsyApaWl1nySk5P58NNMKuX2d/XqVcTHxysPJd+yZQsCAgLg6uqKwYMHo0iRIoiLi0OnTp3Qvn17/PHHH4iOjsbmzZtx8uRJxMfHo27duob8OvQeRo8ejfnz52P16tUoV64cChQooLz3zTff4MKFC/juu++wf/9+REdH49y5czA1NeU+O5Pbu3cvTp06hbZt26JEiRKoUaMGIiMjsWLFCnz55Zc4fvw4xowZgzt37sDb21uZ5tGjRzhz5gxHz6P0M3TmlpUEBQWJu7u7JCcnS48ePcTR0VG5snT58mWZMWMGrzQZmdT6Pd+7d0+++OILqVOnjoSFhSnlI0eOFEtLS51uP5R5zZs3T9lmU56B/O6776RBgwY69f38/KR8+fIc0MVIDR48WJydnSVHjhxSu3ZtOXv2rIiILFmyRGrWrClFixaV1q1bS8WKFaVixYoiIjJo0CBxd3fnPRBG6N69e+Lm5iZr167VKtdcIU5ISJCWLVtKvXr1pHXr1jqjZVLmFBgYKE5OTtKnTx85evSoUl69enUpUqSIHDt2TEReDwQydOhQcXZ2lkaNGkmvXr04eh59MCZOGaxWrVpiYmIiBQoUkHPnzhm6OZRBoqKiROR/idT9+/elZMmS4uHhIcePH1fqDRgwQDw8PHiQZSTCw8OV7nUp74GYPn26lCtXTkJCQrRGXfr999+latWqEhkZ+cnbSu8vZdfJjRs3SokSJWTr1q1y8OBBqVixolSoUEHpdnvq1CmZNm2adOrUSfz8/JTk2NvbW7777jsODJDJffXVVxIaGqpVduHCBcmePbvShT7lfjllPFM+IJUH1JnbunXrxMbGRjZs2CDR0dEior2df/XVV+Ls7KwkTyKi1NNgjOlDMHHKIJod8s6dO6VkyZKyZcsWrXIyXtOmTZPatWvLjRs3tMofPHggzs7OUqtWLa2dtCbmjL3xCAkJEZVKJUuWLBGR1z+sNWvWlEqVKsmmTZvk0aNH8uTJE6lfv760aNGCsc3kNCc6NDZt2iSTJ0+W2bNnK2UvX76U6tWrS4UKFeTQoUPKVQZNbB8/fiyDBw+WXLlysadAJvf8+XMJCAjQuRIcEREhpUqVkoCAAKVMc9C8bds2Wbx4sVZ9bteZ2+PHj6VOnTo6vTpiY2PlyJEjcuXKFRERady4sRQqVEiOHDmikyQxxvSh2Lkzg2ieIu/m5ga1Wo2//vpLq5yMV8OGDXHy5EkMGjQIN2/eBPD6non8+fNj+vTpOHLkCAYMGICLFy8CeB1zEWHsjURMTAxq1qyJAQMGYPDgwViyZAnMzMxw8OBB5M2bF/7+/ihdujQaNGiAiIgIBAUFKTGmzKd169bYvn07gNfbaVxcHL777jv4+fnh9u3bSj0rKyscPHgQNjY28PX1RXBwMNRqNVQqFSIjIzFp0iTs378fBw4cQNmyZQ30bSgtsmXLhkGDBsHS0hITJkzAhg0bAAA2NjZwdXXF1q1blXXCzMwMSUlJWLhwIQ4cOKC1HXOfnfk9fvwYTk5OyusFCxbA29sbHh4e8PDwgJeXF3bt2oVSpUrB09MTly9f1pqeMaYPZuDELUtatWqVZMuWTU6cOGHoptB7elvf9osXL0r27Nnlm2++0bry9Pvvv0v37t2lffv2HFHNCK1bt06aNm0q8fHxcu/ePRkyZIhkz55dea5LUlKSHDlyRJYvXy6bN29WYsyuHpnXjBkzlO6VsbGxIvL6eU1ffPGFlC1bVs6dO6fTZato0aLSrVs3rfn8999/8ujRo0/XcEqXlPvshIQE6dGjh6hUKtm8ebOIvI5j7dq1pWrVqtK+fXvx9/eXmjVrStmyZZXtmFchjMPjx4+lYMGC8v3330toaKi0bt1aypcvL3369JG9e/dKUFCQODs7y7x580RE5Pvvv+fvMmU4jqr3Edy/fx+dO3fGqlWrULBgQUM3h9Io5Qg7+/btw/379+Hg4ICSJUuiWLFiuHDhAr788kvUq1cPffr0QcWKFdGrVy94enqib9++ADiymrGZM2cOZs2ahWPHjsHBwQF37tzB/PnzsWDBAsyYMQM9evTQmYYxzpzeHCFr5syZeP78Obp3744CBQogIiIClStXhouLCxYuXKh1FSkhIQGmpqYwNTXl1WIjkjLmERERsLe3x8uXLzFy5EjMnDkT69atw7fffotHjx5h+fLlCA0Nhbm5OVxcXDB79mzl6pOZmZmBvwmlVWhoKFq3bo28efMie/bsmD59OipWrIi8efPi2bNn+Prrr9G4cWNMnDhRmYb7bMpITJw+klevXsHKysrQzaB0GDJkCNatWwdLS0uYmZkhLi4OK1euRJ06dXDx4kW0a9cOMTExEBE4Ojri+PHjMDc3N3SzSY+UB8QpD7gqVKiAChUqYPXq1QCAe/fuYe7cuVi0aBEmTJigJMWUuaWMaVJSEsaOHYslS5ZgyJAhaNeuHfLnz4/Hjx+jcuXKKFq0KBYsWKDTBY8HWMYjZbwnTJiAu3fvolevXnBzc0N0dDTGjBmD2bNnY/369WjTpo0yXcoYM2kyThEREcqjQVJ69uwZWrRogc6dO6NXr148CUIfBfcYHwmTJuOi+RFetWoVAgMD8ccff6Bs2bK4fv06Zs+ejcaNG2Pfvn348ssvsW/fPvzzzz94/vw5mjVrBlNTU/4AG4GUP6Apr0z4+flh7ty5OH36NKpUqYKCBQvixx9/RHR0NLZs2YI+ffrwxzeTExElpj4+Pvjvv/+wbds2JCYmYubMmVCr1ejQoQPy58+PM2fOoFq1amjdujV2796tdfDFpMl4aOI9dOhQLF++HDNnzlTufcmZMyfGjRuH5ORkdOjQAWZmZvDy8gLwvxiLCPfZRsre3h729vZaZREREfD29kZCQoLSU4D7bfooDNVHkCgzCA4OlqdPn4rI677yI0aMkDZt2mjVefjwobRt21bq1Kmj1E2JfaiNx4wZM6Ru3boSGhoqz549ExGRGzduSLFixcTf31+r7uPHjzlCohFIGZtTp07Jl19+KYcOHVLKhgwZIoULF5Zff/1VHjx4ICKvt2kvLy9uu0Zuy5YtUqBAAa1HCURGRsr58+eVEfb69+8vKpVKa52grCMiIkImTZokTZs2lapVqyrP4uK2TR8LR9Wjz1ZsbCx8fX1RuXJlREVFwcTEBCKCc+fOIS4uTqmXL18+NGrUCDdv3kR8fLzOfHiWOvOSFD2RExMTUaFCBbx69QrDhg1D48aNsX//fhQtWhTTp0/HokWLcObMGaW+vb09R0g0AprYbNiwAePGjYOLiwtq1aqFV69eAQCmTJmCdu3aYc6cOdiwYQPu3buHfPnyYcuWLTA1NUVycrIhm0/vQa1Wa71+/vw5ihYtimLFiuHKlSsYP348qlSpgm+//RatWrXCy5cvMWHCBMyaNQs1a9Y0UKvpY7p37x6OHj2K4sWL49ixYzA3N0dSUhJ/l+mjYeJEn63s2bMjKCgI9vb2cHd3x7Nnz9C4cWNYWloiMDAQ0dHRSt0SJUoge/bsePHihQFbTO9DM7Q08Pq+huTkZHz99dc4duwYpkyZgtKlS6NNmzbw8vLCoUOH4OrqiqNHj+rMh0lT5vfixQsEBwfj1KlT+PfffwG87i6tOdExZcoUtG/fHsOGDcORI0cA/C+p5gGW8dB0z1u9ejUeP36MnDlz4ubNm+jSpQsaNGiAa9euwdfXF8OHD8fly5dx4cIFZMuWDT/++KMyEARlLa6urli1ahVmzJgBMzMzJCcnswsmfVRcu+izpLmnqWTJkggKCsK3336Lb775Bn/88QcaNWqEJUuW4NmzZ2jbti0sLCwwfvz4/2vv3oOirNs/jr8BZfFETh6JZDB11FDHsRxDGVKxYSgS4uCYlA5Rk6aE2OgkkiSjIKCBeCJCxVAgNTAFSklNPIxW0HgoMDDDw3hoMhVRQA6/Pxz2J9rzUD3Z7sbn9R/svTtfZmeX+3Nf1/e66d27N0888YSply5/wL0bx5cvX05RURGnT59mwoQJzJgxg3HjxjFu3Dh2795NUVER69ev59KlSzzyyCOEhoaaePXSmvurgJ06dSImJoZu3bqRkZFBVFQUixYtwmAwUFdXh62tLbGxsfTp04fAwEBAgdhS/fjjjyxevBiDwUBgYCDnz5+noqICf39/xo0bh4ODA5WVldjb2z/wHuuE+t+pa9euwN3vBV0IkYdNU/WkTfn111/p1q0bgPGECsDLy4tdu3bh6upKQUEBy5cvp6CggJKSEoYOHYqtra2xDeD+scdiviIiItiwYQPz5s1j4MCBTJw4ER8fH1atWoWDg4PxuLNnz7Jp0ybmzZunkyszd+/nr7KyEjs7O6ytrenRowdXrlxhyZIlHDlyBF9fX+bPnw9AbW0tBoPB+BqanmfZvL29uX37Nnv27AHutuG2b9+ehoYGqqurmTJlClVVVezbt0/f1SLyt9I3irQZBw4cICAggKKiIgBjaAoMDOTChQsUFhZSXV2Np6cnc+bMoaioiPz8fFJSUowjx+vr6/WP2EIcO3aM3NxcsrOzCQ8Pp3v37tjY2ODt7Y2Dg4OxVauhoQEnJyciIiLUzmPmmu6ZnhcVFYW3tzeurq6MHj2arKwsevbsSWRkJKNGjWLHjh3ExcUBtAhNoPY8S3H/nqbm/WixsbH89NNPbNmyBbhbSaqpqSE+Pp5JkyZx8eJFvvzyS6ytrR94DRGR/4XOAKXN6NmzJ01NTcTFxVFcXAxAQEAApaWl5Ofn4+HhQVZWFrW1tYwdO5bbt2/j5eWFq6urcRO5qhHm6/dOsuzs7Hj22WfJycnBw8ODFStWEBwcTFVVFbt27frd1g69x+arufUqJiaG1atXEx0dTVxcHL6+vgQFBREfH0+PHj2YP38+o0aNIjU1lYyMDBOvWv6q5pBcUFBATU2N8WJHr169GDJkSIs9iXZ2djz++OOMGDGCo0eP6kKXiDwUatWTNqW8vJy3334bGxsbrl+/TnV1NTk5OTg7OxuPKSsr47nnnsPd3Z3NmzebbrHyl8ydO5dhw4bh5ubG+PHjCQkJYdmyZcTGxjJjxgwADh06RFRUFImJiQwdOtTEK5Y/o7q6mhdeeAFfX19mz55t/P3KlSsJCwujsLAQDw8PLly4wNatWwkNDVWFyYKVlZUxbNgwhg8fzlNPPcW8efPo27cvhYWFTJw4kaKiIkaOHPnA89SOKSIPgy7FSJsyYMAAkpOTqa2t5cSJE8yfP98YmporFoMGDeLgwYN8/PHHJlyp/FH3jpPOz88nMzMTR0dHHnvsMTw9PVmyZAnBwcHG0FRbW0tcXBxdunTBxcXFVMuWv6ChoYE7d+5QUVFBx44dAaivr6exsZHQ0FB8fHxIS0ujrq4OR0dHZs+erZHjFub+yvGgQYO4cuUKfn5+VFZWMnz4cMLDw/nll194+eWX2b59O01NTQ+02Co0icjDoIqTtEmnT59m5syZWFtbExERgZubG8ADgx901dJy7Ny5k7y8PPr378/cuXOBu/vaYmNj+fnnnwkODgZg9+7dXLp0iZKSEg37MHPFxcUMHjyYjh07Eh8fj5ubG6NHj2bq1KmUlpayc+dOevfuTX19Pe3atSMkJITq6mqys7NNvXT5C5qHPMDdz67BYKBDhw4MHTrU+F28fv16jhw5Qk5ODlevXqVfv34cP36cDh066J5rIvLQKThJm9XctgcQGRnJmDFjTLwi+TMOHjzI0aNHAejcuTN5eXkcOHCA8PBwoqKiWhyXn5/P5s2bcXFxwcnJidWrVxsHQWhPk3n64YcfeOWVV3B3d6e+vp41a9Zw8uRJnnzySQoKCoiPj8fR0ZHk5GS6detGfX09np6euLi4kJycbOrly5/w6quv4ufnx0svvQTcbbdNT0/Hzs4Og8HAokWLCAoKMh5fW1vLmTNnSElJITc3l2nTphEdHW2q5YtIG6LgJG1aeXk54eHhXL58mXXr1jFs2DBTL0n+gLS0NCIiIujTpw8VFRU4OzszcOBArl27RmVlJVlZWYwYMaLFc27evEnnzp2NPys0mbeGhgaSkpJISEjg5s2bFBYW4urqanw8LS2NjRs3Ul5ezqhRozh//jw1NTUcO3aMdu3aqfpgIS5evEhoaCh79+4lMzOTIUOGMH78eDZv3syNGzcoLCwkPj6eDRs2MG3aNOD/K1O3bt0iOjqakydPsmPHDlWOReSh07eMtGkDBgwgISEBd3d3hgwZYurlyB+QlpbGzJkzWbVqFQcPHiQ3N5fu3btz9epVfHx86NWrF9HR0Rw/fhy4O8K6sbGRTp06GV+jqalJoclMNTY2Gqcd9uvXDysrK/r27cvWrVu5fv268bjXX3+dtWvXEh4ejrOzM35+fsbQVF9fr9BkIRwcHEhKSiIoKIgpU6aQnJxMQEAAI0eOxMPDg7lz5xIREUFwcLBx32nze9yxY0f8/Pz45ptvOH36tIn/EhFpC1RxErmH9ruYt6+++orx48fz/vvvs3DhQmNVYenSpaxZs4bS0lK++OIL1q5dS5cuXYiOjtbUPAt15swZ+vbty9mzZ8nOziYnJ4enn36amJgY7O3t/+PztC/RchQVFVFSUoK1tTX29vacOXOGhIQEfH19yczMNB539epVPvjgA5YuXcrKlSuNg14A42e/uLiYHj16mOLPEJE2RGeIIvdQaDJvjo6OuLm5UVJSQlFRkbGqYG1tjbW1NTU1Nfj7+/PGG29QXV3NzJkzdSXaAhUUFODi4kJOTg5OTk6EhYXh7e1NcXExCxcu5ObNmwDMmjWLb7/9tsVzFZosQ1paGgEBAWzcuJH33nuPuLg4bG1tCQsLY8uWLRQUFBiPffTRR5kzZw5vvfUWmzZtMlaRGxsbuXLlCp999plCk4j8I1RxEhGL0jzUo7GxkVWrVnHu3Dmef/55MjMz8fPzMx6Xnp7Od999R2JiogKxmbt165ZxvDjAiRMnSEpKYu/evSxbtgx/f3/q6upISEggLy+P9u3bYzAYOH78OBcuXFDbpYVpbrfNyMjgxRdf5MiRIyxatAiAJUuWsG7dOnJycsjKysLLy8v4vKqqKjp37oyVlZUqiyJiEgpOImJxysvLCQsL4/Lly5w4cYINGzYQFBRkvF/P/SdUasE0X+np6Zw7d44FCxa0eI9KS0tJTEzk888/JykpCX9/f+7cuUNmZiZff/01tbW1pKSk0K5dO51EW5D/1G4bFxdHcnIyJ0+epKqqisWLF7Nt2zaysrLw9PRs8Roa/CEipqLLdCJicQYMGMCKFSuYPn06AwcOpH///sDdwNTU1PTAiZVCk3lKTU1l+vTpHDp0iKysLBwdHRk7diwAgwcPZvbs2QCEhYVhMBjw9vZm6tSpxulqoOmIlub+dlt3d3cArKyssLW15c6dOzg5OfHuu+9ibW2Nl5cXhw8f5plnnjG+hkKTiJiKKk4iYrEqKioIDQ0FdC8uS5ORkUFISAi5ubmMGTOGgIAALl26RGpqKqNHjzYed+zYMYKCgvjtt99ITExk0qRJxsdUebBMf7Td9tSpU2zfvp133nlH4VhEzIKCk4hYNN2Ly/Kkp6fz2muvMWHCBHbv3g3Anj17SE1N5dSpU6xevbpFCA4MDDTe/PbTTz811bLlb/Tf2m2trKweqBKrsigi5kD9KyJi0XQvLsvy0UcfERISQkhICN9//z2zZs0CwMPDg+nTp9O/f39CQ0M5evQoANXV1XTo0IGYmBi2bdtmyqXL36i53bZr164PtNv+XhVRoUlEzIEqTiLyr6JBEOYrKSmJOXPmkJ+fj5eXFx9++CGRkZFMnjyZlStXArBv3z5SUlLYtWsXPj4+lJWVAXD48GFsbGz0/v7LqN1WRCyJgpOIiPwj9u/fz8WLF5k8eTIA169f55NPPmHBggUtwlNZWRl5eXns378fJycnkpKSaN++vULTv5TabUXEUig4iYjIP+reoQ43btwgOzv7gfAEUFdXh62tLaA9Lv92paWlpKWlkZCQoHAsImZLwUlEREyqOTxFRkYSFBREYmJii8c1Pa9tUWVRRMyVLt+JiIhJ2dvbM3nyZKysrHjzzTdxdnYmLCzM+LhCU9ui0CQi5koVJxERMQvXrl1j//79eHt7Y2NjY+rliIiItKDgJCIiZkd7mkRExNwoOImIiIiIiLRCjcQiIiIiIiKtUHASERERERFphYKTiIiIiIhIKxScREREREREWqHgJCIiIiIi0goFJxERERERkVYoOImIiIiIiLRCwUlERERERKQVCk4iIiIiIiKtUHASERERERFphYKTiIiIiIhIK/4PQ+I6ub7Z7OwAAAAASUVORK5CYII=", "text/plain": [ "

" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	feature	n_levels	min_group_mean	max_group_mean	mean_range
1	Access_to_Resources	3	66.203	68.092	1.889
0	Parental_Involvement	3	66.358	68.093	1.735
11	Distance_from_Home	4	66.433	67.512	1.079
9	Learning_Disabilities	2	66.271	67.349	1.079
10	Parental_Education_Level	4	66.894	67.971	1.077
8	Peer_Influence	3	66.564	67.623	1.059
6	Teacher_Quality	4	66.641	67.677	1.036
5	Family_Income	3	66.848	67.842	0.994
3	Motivation_Level	3	66.752	67.704	0.952
4	Internet_Access	2	66.535	67.293	0.758
2	Extracurricular_Activities	2	66.931	67.442	0.510
7	School_Type	2	67.213	67.288	0.075
12	Gender	2	67.229	67.245	0.016

\n", "

" ], "text/plain": [ " feature n_levels min_group_mean max_group_mean \\\n", "1 Access_to_Resources 3 66.203 68.092 \n", "0 Parental_Involvement 3 66.358 68.093 \n", "11 Distance_from_Home 4 66.433 67.512 \n", "9 Learning_Disabilities 2 66.271 67.349 \n", "10 Parental_Education_Level 4 66.894 67.971 \n", "8 Peer_Influence 3 66.564 67.623 \n", "6 Teacher_Quality 4 66.641 67.677 \n", "5 Family_Income 3 66.848 67.842 \n", "3 Motivation_Level 3 66.752 67.704 \n", "4 Internet_Access 2 66.535 67.293 \n", "2 Extracurricular_Activities 2 66.931 67.442 \n", "7 School_Type 2 67.213 67.288 \n", "12 Gender 2 67.229 67.245 \n", "\n", " mean_range \n", "1 1.889 \n", "0 1.735 \n", "11 1.079 \n", "9 1.079 \n", "10 1.077 \n", "8 1.059 \n", "6 1.036 \n", "5 0.994 \n", "3 0.952 \n", "4 0.758 \n", "2 0.510 \n", "7 0.075 \n", "12 0.016 " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# Для числовых признаков используем корреляцию Спирмена с целевой переменной.\n", "numeric_corr = (\n", " data[numeric_features + [TARGET]]\n", " .corr(method=\"spearman\", numeric_only=True)[TARGET]\n", " .drop(TARGET)\n", " .sort_values(ascending=False)\n", ")\n", "\n", "display(numeric_corr.to_frame(\"spearman_corr\"))\n", "\n", "plt.figure(figsize=(10, 4))\n", "plt.bar(numeric_corr.index, numeric_corr.values)\n", "plt.xticks(rotation=45, ha=\"right\")\n", "plt.title(\"Корреляция Спирмена числовых признаков с Exam_Score\")\n", "plt.ylabel(\"corr\")\n", "plt.show()\n", "\n", "# Для категориальных признаков оцениваем диапазон средних значений целевой переменной по категориям.\n", "categorical_effect_rows = []\n", "for col in categorical_features:\n", " group_means = data.groupby(col, dropna=False)[TARGET].mean().sort_values(ascending=False)\n", " categorical_effect_rows.append({\n", " \"feature\": col,\n", " \"n_levels\": int(data[col].nunique(dropna=False)),\n", " \"min_group_mean\": round(group_means.min(), 3),\n", " \"max_group_mean\": round(group_means.max(), 3),\n", " \"mean_range\": round(group_means.max() - group_means.min(), 3),\n", " })\n", "\n", "categorical_effect = pd.DataFrame(categorical_effect_rows).sort_values(\n", " \"mean_range\", ascending=False\n", ")\n", "display(categorical_effect)" ] }, { "cell_type": "markdown", "id": "9fecbaa5", "metadata": {}, "source": [ "## 4. Разделение данных и вспомогательные функции\n", "\n", "Будем использовать схему:\n", "- **train** — обучение моделей;\n", "- **valid** — выбор лучшей архитектуры;\n", "- **test** — финальная контрольная оценка.\n", "\n", "Для категориальных переменных используется `OneHotEncoder`, для числовых — медианная импутация и стандартизация." ] }, { "cell_type": "code", "execution_count": 7, "id": "ed553461", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Train: (4624, 19) (4624,)\n", "Valid: (991, 19) (991,)\n", "Test: (992, 19) (992,)\n" ] } ], "source": [ "X_train_full, X_test, y_train_full, y_test = train_test_split(\n", " X, y, test_size=0.15, random_state=SEED\n", ")\n", "\n", "# Второй split подобран так, чтобы получить примерно 70/15/15.\n", "X_train, X_valid, y_train, y_valid = train_test_split(\n", " X_train_full,\n", " y_train_full,\n", " test_size=0.1764705882,\n", " random_state=SEED\n", ")\n", "\n", "print(\"Train:\", X_train.shape, y_train.shape)\n", "print(\"Valid:\", X_valid.shape, y_valid.shape)\n", "print(\"Test: \", X_test.shape, y_test.shape)\n", "\n", "\n", "def make_ohe():\n", " # Совместимость с разными версиями scikit-learn.\n", " try:\n", " return OneHotEncoder(handle_unknown=\"ignore\", sparse_output=False)\n", " except TypeError:\n", " return OneHotEncoder(handle_unknown=\"ignore\", sparse=False)\n", "\n", "\n", "def build_preprocessor(num_cols, cat_cols):\n", " return ColumnTransformer(\n", " transformers=[\n", " (\n", " \"num\",\n", " Pipeline(\n", " steps=[\n", " (\"imputer\", SimpleImputer(strategy=\"median\")),\n", " (\"scaler\", StandardScaler()),\n", " ]\n", " ),\n", " num_cols,\n", " ),\n", " (\n", " \"cat\",\n", " Pipeline(\n", " steps=[\n", " (\"imputer\", SimpleImputer(strategy=\"most_frequent\")),\n", " (\"ohe\", make_ohe()),\n", " ]\n", " ),\n", " cat_cols,\n", " ),\n", " ],\n", " remainder=\"drop\",\n", " )\n", "\n", "\n", "def build_mlp_pipeline(num_cols, cat_cols, hidden_layers, alpha, learning_rate, random_state):\n", " preprocessor = build_preprocessor(num_cols, cat_cols)\n", " model = MLPRegressor(\n", " hidden_layer_sizes=hidden_layers,\n", " activation=\"relu\",\n", " solver=\"adam\",\n", " alpha=alpha,\n", " learning_rate_init=learning_rate,\n", " max_iter=400,\n", " early_stopping=True,\n", " validation_fraction=0.15,\n", " n_iter_no_change=25,\n", " random_state=random_state,\n", " )\n", " return Pipeline(\n", " steps=[\n", " (\"preprocessor\", preprocessor),\n", " (\"model\", model),\n", " ]\n", " )\n", "\n", "\n", "def regression_metrics(y_true, y_pred):\n", " rmse = float(np.sqrt(mean_squared_error(y_true, y_pred)))\n", " mae = float(mean_absolute_error(y_true, y_pred))\n", " r2 = float(r2_score(y_true, y_pred))\n", " return {\"rmse\": rmse, \"mae\": mae, \"r2\": r2}\n", "\n", "\n", "def search_mlp_configs(\n", " X_train_part,\n", " y_train_part,\n", " X_valid_part,\n", " y_valid_part,\n", " num_cols,\n", " cat_cols,\n", " configs,\n", " seed=SEED,\n", "):\n", " rows = []\n", " fitted_models = {}\n", "\n", " for idx, cfg in enumerate(configs, start=1):\n", " model = build_mlp_pipeline(\n", " num_cols=num_cols,\n", " cat_cols=cat_cols,\n", " hidden_layers=cfg[\"hidden_layers\"],\n", " alpha=cfg[\"alpha\"],\n", " learning_rate=cfg[\"learning_rate\"],\n", " random_state=seed,\n", " )\n", " model.fit(X_train_part, y_train_part)\n", "\n", " valid_pred = model.predict(X_valid_part)\n", " metrics = regression_metrics(y_valid_part, valid_pred)\n", "\n", " row = {\n", " \"config_id\": idx,\n", " \"hidden_layers\": str(cfg[\"hidden_layers\"]),\n", " \"alpha\": cfg[\"alpha\"],\n", " \"learning_rate\": cfg[\"learning_rate\"],\n", " \"rmse_valid\": round(metrics[\"rmse\"], 6),\n", " \"mae_valid\": round(metrics[\"mae\"], 6),\n", " \"r2_valid\": round(metrics[\"r2\"], 6),\n", " \"n_iter\": getattr(model.named_steps[\"model\"], \"n_iter_\", np.nan),\n", " }\n", " rows.append(row)\n", " fitted_models[idx] = model\n", "\n", " results = pd.DataFrame(rows).sort_values([\"rmse_valid\", \"mae_valid\", \"r2_valid\"], ascending=[True, True, False]).reset_index(drop=True)\n", " return results, fitted_models\n", "\n", "\n", "def rmse_loss(y_true, y_pred):\n", " return np.sqrt(mean_squared_error(y_true, y_pred))\n", "\n", "\n", "rmse_scorer = make_scorer(rmse_loss, greater_is_better=False)" ] }, { "cell_type": "markdown", "id": "da9ea958", "metadata": {}, "source": [ "## 5. Подбор архитектуры MLP на полном наборе признаков\n", "\n", "Подбираем несколько разумных конфигураций:\n", "- разные размеры скрытых слоёв;\n", "- небольшая L2-регуляризация `alpha`;\n", "- разные стартовые скорости обучения.\n", "\n", "Критерий выбора — **минимум RMSE на validation**." ] }, { "cell_type": "code", "execution_count": 8, "id": "eba3e23d", "metadata": {}, "outputs": [ { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	config_id	hidden_layers	alpha	learning_rate	rmse_valid	mae_valid	r2_valid	n_iter
0	6	(128, 32)	0.00050	0.0007	1.842312	0.466257	0.757518	96
1	1	(64, 32)	0.00010	0.0010	1.847151	0.486388	0.756243	99
2	2	(128, 64)	0.00010	0.0010	1.848015	0.512239	0.756015	100
3	3	(128, 64, 32)	0.00010	0.0007	1.876093	0.588299	0.748544	141
4	5	(96, 48, 24)	0.00005	0.0010	1.878440	0.555151	0.747915	91
5	4	(256, 128)	0.00010	0.0005	1.882249	0.586328	0.746891	110

\n", "

" ], "text/plain": [ " config_id hidden_layers alpha learning_rate rmse_valid mae_valid \\\n", "0 6 (128, 32) 0.00050 0.0007 1.842312 0.466257 \n", "1 1 (64, 32) 0.00010 0.0010 1.847151 0.486388 \n", "2 2 (128, 64) 0.00010 0.0010 1.848015 0.512239 \n", "3 3 (128, 64, 32) 0.00010 0.0007 1.876093 0.588299 \n", "4 5 (96, 48, 24) 0.00005 0.0010 1.878440 0.555151 \n", "5 4 (256, 128) 0.00010 0.0005 1.882249 0.586328 \n", "\n", " r2_valid n_iter \n", "0 0.757518 96 \n", "1 0.756243 99 \n", "2 0.756015 100 \n", "3 0.748544 141 \n", "4 0.747915 91 \n", "5 0.746891 110 " ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Лучшая конфигурация на полном наборе признаков:\n", "{'config_id': 6, 'hidden_layers': '(128, 32)', 'alpha': 0.0005, 'learning_rate': 0.0007, 'rmse_valid': 1.842312, 'mae_valid': 0.466257, 'r2_valid': 0.757518, 'n_iter': 96}\n" ] } ], "source": [ "mlp_configs = [\n", " {\"hidden_layers\": (64, 32), \"alpha\": 1e-4, \"learning_rate\": 0.0010},\n", " {\"hidden_layers\": (128, 64), \"alpha\": 1e-4, \"learning_rate\": 0.0010},\n", " {\"hidden_layers\": (128, 64, 32), \"alpha\": 1e-4, \"learning_rate\": 0.0007},\n", " {\"hidden_layers\": (256, 128), \"alpha\": 1e-4, \"learning_rate\": 0.0005},\n", " {\"hidden_layers\": (96, 48, 24), \"alpha\": 5e-5, \"learning_rate\": 0.0010},\n", " {\"hidden_layers\": (128, 32), \"alpha\": 5e-4, \"learning_rate\": 0.0007},\n", "]\n", "\n", "full_results, full_models = search_mlp_configs(\n", " X_train_part=X_train,\n", " y_train_part=y_train,\n", " X_valid_part=X_valid,\n", " y_valid_part=y_valid,\n", " num_cols=numeric_features,\n", " cat_cols=categorical_features,\n", " configs=mlp_configs,\n", ")\n", "\n", "display(full_results)\n", "best_full_config_id = int(full_results.iloc[0][\"config_id\"])\n", "best_full_model = full_models[best_full_config_id]\n", "best_full_row = full_results.iloc[0].to_dict()\n", "print(\"Лучшая конфигурация на полном наборе признаков:\")\n", "print(best_full_row)" ] }, { "cell_type": "markdown", "id": "406e2d2d", "metadata": {}, "source": [ "## 6. Ключевые переменные через permutation importance\n", "\n", "Ключевые переменные будем определять не только по однофакторным связям, а по тому, **насколько ухудшается качество лучшей нейросети при перемешивании конкретного признака**.\n", "\n", "Это даёт более прикладную оценку важности именно для итоговой модели." ] }, { "cell_type": "code", "execution_count": 9, "id": "626c2282", "metadata": {}, "outputs": [ { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	feature	importance_mean	importance_std
0	Attendance	1.891556	0.052341
1	Hours_Studied	1.249790	0.054106
2	Access_to_Resources	0.258330	0.019634
3	Previous_Scores	0.255576	0.023212
4	Parental_Involvement	0.252574	0.012115
5	Tutoring_Sessions	0.191517	0.011300
6	Peer_Influence	0.094267	0.008479
7	Family_Income	0.091834	0.013912
8	Parental_Education_Level	0.077539	0.008235
9	Motivation_Level	0.071868	0.009680
10	Distance_from_Home	0.052813	0.007905
11	Internet_Access	0.041834	0.008134
12	Extracurricular_Activities	0.039941	0.009552
13	Teacher_Quality	0.039414	0.009580
14	Learning_Disabilities	0.034346	0.005191
15	Physical_Activity	0.013012	0.006122
16	Sleep_Hours	0.000898	0.002343
17	School_Type	-0.000170	0.001520
18	Gender	-0.002435	0.002779

\n", "

" ], "text/plain": [ " feature importance_mean importance_std\n", "0 Attendance 1.891556 0.052341\n", "1 Hours_Studied 1.249790 0.054106\n", "2 Access_to_Resources 0.258330 0.019634\n", "3 Previous_Scores 0.255576 0.023212\n", "4 Parental_Involvement 0.252574 0.012115\n", "5 Tutoring_Sessions 0.191517 0.011300\n", "6 Peer_Influence 0.094267 0.008479\n", "7 Family_Income 0.091834 0.013912\n", "8 Parental_Education_Level 0.077539 0.008235\n", "9 Motivation_Level 0.071868 0.009680\n", "10 Distance_from_Home 0.052813 0.007905\n", "11 Internet_Access 0.041834 0.008134\n", "12 Extracurricular_Activities 0.039941 0.009552\n", "13 Teacher_Quality 0.039414 0.009580\n", "14 Learning_Disabilities 0.034346 0.005191\n", "15 Physical_Activity 0.013012 0.006122\n", "16 Sleep_Hours 0.000898 0.002343\n", "17 School_Type -0.000170 0.001520\n", "18 Gender -0.002435 0.002779" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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", "text/plain": [ "

" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Ключевые переменные:\n", "['Attendance', 'Hours_Studied', 'Access_to_Resources', 'Previous_Scores', 'Parental_Involvement', 'Tutoring_Sessions', 'Peer_Influence', 'Family_Income', 'Parental_Education_Level', 'Motivation_Level']\n" ] } ], "source": [ "perm = permutation_importance(\n", " estimator=best_full_model,\n", " X=X_valid,\n", " y=y_valid,\n", " n_repeats=10,\n", " random_state=SEED,\n", " scoring=rmse_scorer,\n", ")\n", "\n", "importance_df = pd.DataFrame({\n", " \"feature\": X.columns,\n", " \"importance_mean\": perm.importances_mean,\n", " \"importance_std\": perm.importances_std,\n", "}).sort_values(\"importance_mean\", ascending=False).reset_index(drop=True)\n", "\n", "display(importance_df)\n", "\n", "plt.figure(figsize=(10, 6))\n", "top_to_plot = importance_df.head(12).sort_values(\"importance_mean\")\n", "plt.barh(top_to_plot[\"feature\"], top_to_plot[\"importance_mean\"])\n", "plt.title(\"Permutation importance (по validation)\")\n", "plt.xlabel(\"Ухудшение качества при перемешивании признака\")\n", "plt.show()\n", "\n", "# В качестве ключевых переменных берём top-10 по permutation importance.\n", "key_features = importance_df.head(10)[\"feature\"].tolist()\n", "key_numeric_features = [col for col in key_features if col in numeric_features]\n", "key_categorical_features = [col for col in key_features if col in categorical_features]\n", "\n", "print(\"Ключевые переменные:\")\n", "print(key_features)" ] }, { "cell_type": "markdown", "id": "5fcfe71d", "metadata": {}, "source": [ "## 7. Повторный подбор моделей только на ключевых переменных\n", "\n", "Это проверка гипотезы: можно ли слегка упростить модель, не потеряв качество." ] }, { "cell_type": "code", "execution_count": 10, "id": "eaca10e7", "metadata": {}, "outputs": [ { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	config_id	hidden_layers	alpha	learning_rate	rmse_valid	mae_valid	r2_valid	n_iter
0	1	(64, 32)	0.00010	0.0010	1.960040	0.791034	0.725538	132
1	6	(128, 32)	0.00050	0.0007	1.970644	0.796516	0.722560	177
2	4	(256, 128)	0.00010	0.0005	1.972069	0.843776	0.722158	136
3	2	(128, 64)	0.00010	0.0010	1.974057	0.805843	0.721598	142
4	5	(96, 48, 24)	0.00005	0.0010	1.974669	0.822600	0.721425	127
5	3	(128, 64, 32)	0.00010	0.0007	2.000959	0.847101	0.713958	130

\n", "

" ], "text/plain": [ " config_id hidden_layers alpha learning_rate rmse_valid mae_valid \\\n", "0 1 (64, 32) 0.00010 0.0010 1.960040 0.791034 \n", "1 6 (128, 32) 0.00050 0.0007 1.970644 0.796516 \n", "2 4 (256, 128) 0.00010 0.0005 1.972069 0.843776 \n", "3 2 (128, 64) 0.00010 0.0010 1.974057 0.805843 \n", "4 5 (96, 48, 24) 0.00005 0.0010 1.974669 0.822600 \n", "5 3 (128, 64, 32) 0.00010 0.0007 2.000959 0.847101 \n", "\n", " r2_valid n_iter \n", "0 0.725538 132 \n", "1 0.722560 177 \n", "2 0.722158 136 \n", "3 0.721598 142 \n", "4 0.721425 127 \n", "5 0.713958 130 " ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	feature_set	hidden_layers	alpha	learning_rate	rmse_valid	mae_valid	r2_valid
0	all_features	(128, 32)	0.0005	0.0007	1.842312	0.466257	0.757518
1	key_features	(64, 32)	0.0001	0.0010	1.960040	0.791034	0.725538

\n", "

" ], "text/plain": [ " feature_set hidden_layers alpha learning_rate rmse_valid mae_valid \\\n", "0 all_features (128, 32) 0.0005 0.0007 1.842312 0.466257 \n", "1 key_features (64, 32) 0.0001 0.0010 1.960040 0.791034 \n", "\n", " r2_valid \n", "0 0.757518 \n", "1 0.725538 " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "key_results, key_models = search_mlp_configs(\n", " X_train_part=X_train[key_features],\n", " y_train_part=y_train,\n", " X_valid_part=X_valid[key_features],\n", " y_valid_part=y_valid,\n", " num_cols=key_numeric_features,\n", " cat_cols=key_categorical_features,\n", " configs=mlp_configs,\n", ")\n", "\n", "display(key_results)\n", "\n", "best_key_config_id = int(key_results.iloc[0][\"config_id\"])\n", "best_key_model = key_models[best_key_config_id]\n", "best_key_row = key_results.iloc[0].to_dict()\n", "\n", "comparison = pd.DataFrame([\n", " {\n", " \"feature_set\": \"all_features\",\n", " \"hidden_layers\": best_full_row[\"hidden_layers\"],\n", " \"alpha\": best_full_row[\"alpha\"],\n", " \"learning_rate\": best_full_row[\"learning_rate\"],\n", " \"rmse_valid\": best_full_row[\"rmse_valid\"],\n", " \"mae_valid\": best_full_row[\"mae_valid\"],\n", " \"r2_valid\": best_full_row[\"r2_valid\"],\n", " },\n", " {\n", " \"feature_set\": \"key_features\",\n", " \"hidden_layers\": best_key_row[\"hidden_layers\"],\n", " \"alpha\": best_key_row[\"alpha\"],\n", " \"learning_rate\": best_key_row[\"learning_rate\"],\n", " \"rmse_valid\": best_key_row[\"rmse_valid\"],\n", " \"mae_valid\": best_key_row[\"mae_valid\"],\n", " \"r2_valid\": best_key_row[\"r2_valid\"],\n", " },\n", "]).sort_values(\"rmse_valid\")\n", "\n", "display(comparison)" ] }, { "cell_type": "markdown", "id": "15fe2194", "metadata": {}, "source": [ "## 8. Финальная нейросеть и контроль на тестовой выборке\n", "\n", "Дальше автоматически выбирается лучший режим:\n", "- либо модель на всех признаках;\n", "- либо модель на ключевых признаках.\n", "\n", "Чтобы снизить разброс прогноза, финальная оценка строится как **ансамбль из нескольких MLP одной лучшей архитектуры**, обученных с разными `random_state`." ] }, { "cell_type": "code", "execution_count": 11, "id": "7fa6cf8a", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Финальный режим: all_features\n", "Используемые признаки: ['Hours_Studied', 'Attendance', 'Parental_Involvement', 'Access_to_Resources', 'Extracurricular_Activities', 'Sleep_Hours', 'Previous_Scores', 'Motivation_Level', 'Internet_Access', 'Tutoring_Sessions', 'Family_Income', 'Teacher_Quality', 'School_Type', 'Peer_Influence', 'Physical_Activity', 'Learning_Disabilities', 'Parental_Education_Level', 'Distance_from_Home', 'Gender']\n", "Архитектура: (128, 32)\n", "alpha: 0.0005\n", "learning_rate: 0.0007\n" ] }, { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	feature_set	hidden_layers	alpha	learning_rate	rmse_test	mae_test	r2_test
0	all_features	(128, 32)	0.0005	0.0007	1.711333	0.462977	0.796033

\n", "

" ], "text/plain": [ " feature_set hidden_layers alpha learning_rate rmse_test mae_test \\\n", "0 all_features (128, 32) 0.0005 0.0007 1.711333 0.462977 \n", "\n", " r2_test \n", "0 0.796033 " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "best_mode = comparison.iloc[0][\"feature_set\"]\n", "\n", "if best_mode == \"all_features\":\n", " final_features = X.columns.tolist()\n", " final_numeric_features = numeric_features\n", " final_categorical_features = categorical_features\n", " final_hidden_layers = ast.literal_eval(best_full_row[\"hidden_layers\"])\n", " final_alpha = float(best_full_row[\"alpha\"])\n", " final_learning_rate = float(best_full_row[\"learning_rate\"])\n", "else:\n", " final_features = key_features\n", " final_numeric_features = key_numeric_features\n", " final_categorical_features = key_categorical_features\n", " final_hidden_layers = ast.literal_eval(best_key_row[\"hidden_layers\"])\n", " final_alpha = float(best_key_row[\"alpha\"])\n", " final_learning_rate = float(best_key_row[\"learning_rate\"])\n", "\n", "print(\"Финальный режим:\", best_mode)\n", "print(\"Используемые признаки:\", final_features)\n", "print(\"Архитектура:\", final_hidden_layers)\n", "print(\"alpha:\", final_alpha)\n", "print(\"learning_rate:\", final_learning_rate)\n", "\n", "\n", "def fit_ensemble_and_predict(\n", " X_train_part,\n", " y_train_part,\n", " X_test_part,\n", " num_cols,\n", " cat_cols,\n", " hidden_layers,\n", " alpha,\n", " learning_rate,\n", " seeds=(7, 11, 21, 42, 99),\n", "):\n", " models = []\n", " predictions = []\n", "\n", " for seed in seeds:\n", " model = build_mlp_pipeline(\n", " num_cols=num_cols,\n", " cat_cols=cat_cols,\n", " hidden_layers=hidden_layers,\n", " alpha=alpha,\n", " learning_rate=learning_rate,\n", " random_state=seed,\n", " )\n", " model.fit(X_train_part, y_train_part)\n", " pred = model.predict(X_test_part)\n", " models.append(model)\n", " predictions.append(pred)\n", "\n", " ensemble_pred = np.mean(np.column_stack(predictions), axis=1)\n", " return models, ensemble_pred\n", "\n", "\n", "ensemble_models, test_pred = fit_ensemble_and_predict(\n", " X_train_part=X_train_full[final_features],\n", " y_train_part=y_train_full,\n", " X_test_part=X_test[final_features],\n", " num_cols=final_numeric_features,\n", " cat_cols=final_categorical_features,\n", " hidden_layers=final_hidden_layers,\n", " alpha=final_alpha,\n", " learning_rate=final_learning_rate,\n", ")\n", "\n", "final_metrics = regression_metrics(y_test, test_pred)\n", "final_metrics_df = pd.DataFrame([{\n", " \"feature_set\": best_mode,\n", " \"hidden_layers\": str(final_hidden_layers),\n", " \"alpha\": final_alpha,\n", " \"learning_rate\": final_learning_rate,\n", " \"rmse_test\": round(final_metrics[\"rmse\"], 6),\n", " \"mae_test\": round(final_metrics[\"mae\"], 6),\n", " \"r2_test\": round(final_metrics[\"r2\"], 6),\n", "}])\n", "\n", "display(final_metrics_df)" ] }, { "cell_type": "code", "execution_count": 12, "id": "9be22188", "metadata": {}, "outputs": [ { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	y_true	y_pred	abs_error
28	89	61.165762	27.834238
704	98	70.908403	27.091597
721	97	72.902072	24.097928
480	88	67.932695	20.067305
945	80	64.922223	15.077777
340	67	68.649078	1.649078
448	65	66.508582	1.508582
515	65	66.380505	1.380505
512	58	59.354995	1.354995
920	60	61.305158	1.305158
572	71	72.249651	1.249651
540	63	64.210179	1.210179
718	68	69.182157	1.182157
814	57	58.175722	1.175722
761	60	61.145952	1.145952

\n", "

" ], "text/plain": [ " y_true y_pred abs_error\n", "28 89 61.165762 27.834238\n", "704 98 70.908403 27.091597\n", "721 97 72.902072 24.097928\n", "480 88 67.932695 20.067305\n", "945 80 64.922223 15.077777\n", "340 67 68.649078 1.649078\n", "448 65 66.508582 1.508582\n", "515 65 66.380505 1.380505\n", "512 58 59.354995 1.354995\n", "920 60 61.305158 1.305158\n", "572 71 72.249651 1.249651\n", "540 63 64.210179 1.210179\n", "718 68 69.182157 1.182157\n", "814 57 58.175722 1.175722\n", "761 60 61.145952 1.145952" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAABW0AAAHqCAYAAAB/bWzAAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjgsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvwVt1zgAAAAlwSFlzAAAPYQAAD2EBqD+naQAAxAxJREFUeJzs3Xd8FHX+x/H3bnoPCaRB6CVUEbAEFQQCiIgoIHaxcj8ECyh6eCACKop3FhTrcainWCji2ZCACgqIFCkBRHogkIQASUjf3ZnfH7nsGRMghJBJyOv5eHBkv/Od73zmy9w6+8l3P2MzTdMUAAAAAAAAAKBGsFsdAAAAAAAAAADgf0jaAgAAAAAAAEANQtIWAAAAAAAAAGoQkrYAAAAAAAAAUIOQtAUAAAAAAACAGoSkLQAAAAAAAADUICRtAQAAAAAAAKAGIWkLAAAAAAAAADUISVsAAAAAAAAAqEFI2gIAAAAAAABADULSFsB5484771RgYKDVYQAAAKAWePfdd2Wz2dx/fH191bp1a40ZM0ZpaWlWhwcAqOM8rQ4AAM7G0aNH9eGHH+rHH3/UihUrlJ+fr6uuukoXXnihhg8frgsvvNDqEAEAAFCDTZ06Vc2aNVNBQYF++uknvfHGG/r666+VlJQkf39/q8MDANRRrLQFUGt9/PHHatq0qR566CGtW7dOXl5estlsOnr0qP7xj3+oS5cuuvPOO1VUVGR1qAAAAKihBgwYoNtuu0333nuv3n33XT388MPau3evPv/8c6tDAwDUYSRtAdRKK1eu1G233aaoqCitXLlSe/fuVUJCgnx9fbV27VodOnRIN998s9577z2NHTu21L5///vf1b17d4WHh8vPz09du3bV/PnzyxzDZrPpqaeecr92Op26+uqrFRYWpm3btrn7nOrPlVdeedJz6NChg3r16lWm3TAMNWzYUMOGDXO3ffzxx+ratauCgoIUHBysjh076pVXXjnlHO3bt6/Csf3www+y2Wz65JNP9MQTTygqKkoBAQG69tprdeDAgTJjz5s3T127dpWfn5/q16+v2267TSkpKaX63HnnneUet2XLlu4+TZs21TXXXFNm/DFjxshms5VqczqdmjZtmlq0aCEfHx81bdpUTzzxhAoLC0v1Gzx4sJo2bSpfX19FRETo2muv1ZYtW045VwAAACV69+4tSdq7d68k6dixY3r00UfVsWNHBQYGKjg4WAMGDNCmTZvK7FtQUKCnnnpKrVu3lq+vr6KjozVkyBDt3r1b0rm9P1uzZo2uuuoqhYSEyN/fXz179tTKlSvLPccrr7yy3OP/8d63xAcffOC+7wsLC9NNN91U7vFPdW5/ZBiGXn75ZbVv316+vr6KjIzUX/7yFx0/frxUvzO5Tywv9hdeeKHc+/HCwkJNnjxZLVu2lI+Pj2JjY/XYY4+Vuac8mYrcB0vSb7/9puHDh6tBgwby8/NTmzZt9Le//U2S9NRTT532c8QPP/wgSfrxxx91ww03qHHjxu54x44dq/z8fPexTnbf/cc/+/btc8/rnXfeWSrWkSNHytfX133MEq+//rrat28vHx8fxcTEaPTo0crMzCzV58/XUv369TVw4EAlJSVVaD4BnBzlEQDUSs8995wMw3AnM/+sfv36ev/997Vt2za99dZbmjx5siIiIiRJr7zyiq699lrdeuutKioq0scff6wbbrhBX375pQYOHHjSY95777364YcflJiYqHbt2kmS/v3vf7u3//jjj3r77bf10ksvqX79+pKkyMjIk45344036qmnnlJqaqqioqLc7T/99JMOHTqkm266SZKUmJiom2++WX369NHzzz8vSdq+fbtWrlyphx566LRzdfPNN+vqq68u1TZhwoRy+z7zzDOy2Wx6/PHHlZ6erpdfflkJCQnauHGj/Pz8JBXXf7vrrrt00UUXafr06UpLS9Mrr7yilStX6tdff1VoaKh7PB8fH/3zn/8sdYygoKDTxlyee++9V++9956GDRumRx55RGvWrNH06dO1fft2ffbZZ6X6jhw5UlFRUTp06JBee+01JSQkaO/evXzFEQAAnFZJgjU8PFyStGfPHi1atEg33HCDmjVrprS0NL311lvq2bOntm3bppiYGEmSy+XSNddco2XLlummm27SQw89pBMnTigxMVFJSUlq0aKF+xhVfX/23XffacCAAeratasmT54su92uOXPmqHfv3vrxxx918cUXlxm3UaNGmj59uiQpJydHo0aNKvfYkyZN0vDhw3XvvffqyJEjevXVV9WjR48y930lRo4cqSuuuEKStHDhwjL3aX/5y1/c95MPPvig9u7dq9dee02//vqrVq5cKS8vr3Ln4UxkZma6z+2PDMPQtddeq59++kkjR45U27ZttWXLFr300kv6/ffftWjRolOOW9H74M2bN+uKK66Ql5eXRo4cqaZNm2r37t364osv9Mwzz2jIkCGlFjKMHTtWbdu21ciRI91tbdu2lVScJM7Ly9OoUaMUHh6uX375Ra+++qoOHjyoefPmSSqe04SEBPe+t99+u66//noNGTLE3dagQYNyz2ny5MmaPXu2Pvnkk1IJ7qeeekpTpkxRQkKCRo0apR07duiNN97Q2rVry/w7xcXF6W9/+5tM09Tu3bv14osv6uqrr1ZycvIp5xPAaZgAUAuFhYWZTZo0KdU2YsQIMyAgoFTbpEmTTEnmF1984W7Ly8sr1aeoqMjs0KGD2bt371LtkszJkyebpmmaEyZMMD08PMxFixadNKY5c+aYksy9e/dW6Bx27NhhSjJfffXVUu3333+/GRgY6I7zoYceMoODg02n01mhcUvs3bvXlGS+8MILZba1b9/e7Nmzp/v1999/b0oyGzZsaGZnZ7vbP/30U1OS+corr5imWTxXERERZocOHcz8/Hx3vy+//NKUZD755JPutvL+Pf6sSZMm5sCBA8u0jx492vzjf6I2btxoSjLvvffeUv0effRRU5L53XffnfQYJeewbt26U8YCAADqlpJ7t6VLl5pHjhwxDxw4YH788cdmeHi46efnZx48eNA0TdMsKCgwXS5XqX337t1r+vj4mFOnTnW3/etf/zIlmS+++GKZYxmG4d6vqu/PDMMwW7VqZfbv3999HNMsvudt1qyZ2bdv3zLH6t69u9mhQwf36yNHjpS69zVN09y3b5/p4eFhPvPMM6X23bJli+np6VmmfefOnaYk87333nO3TZ48udQ93Y8//mhKMj/88MNS+y5evLhMe0XvE03TLBP7Y489ZkZERJhdu3YtNaf//ve/Tbvdbv7444+l9n/zzTdNSebKlSvLHK/EmdwH9+jRwwwKCjL3799faow//vv8UZMmTcwRI0aUu+3Pn11M0zSnT59u2my2MuOX+PN8nOxYb731VrmfR9LT001vb2+zX79+pa791157zZRk/utf/3K39ezZs9Qcm6ZpPvHEE6YkMz09vdwYAFQM5REA1EonTpxwr5w9lZKVrtnZ2e62khUJknT8+HFlZWXpiiuu0IYNG8od47XXXtP06dM1c+ZMDR48+Cwj/5/WrVurc+fO+uSTT9xtLpdL8+fP16BBg9xxhoaGKjc3V4mJiVV27JO54447Sq2EHTZsmKKjo/X1119LktatW6f09HTdf//98vX1dfcbOHCg4uLi9NVXX53xMR0OhzIyMkr9KSgoKNWn5Pjjxo0r1f7II49IUpnj5uXlKSMjQxs3btQ777yjyMhItW7d+oxjAwAA57+EhAQ1aNBAsbGxuummmxQYGKjPPvtMDRs2lFT8zSG7vfijs8vl0tGjRxUYGKg2bdqUun9csGCB6tevrwceeKDMMf78df4zcbr7s40bN2rnzp265ZZbdPToUff9VG5urvr06aMVK1bIMIxSYxYUFJS6lyvPwoULZRiGhg8fXuo+LSoqSq1atdL3339fqn/JcyR8fHxOOua8efMUEhKivn37lhqza9euCgwMLDNmRe4T/ywlJUWvvvqqJk2apMDAwDLHb9u2reLi4kqNWVIS48/H/6OK3gcfOXJEK1as0N13363GjRuXGqMy18EfP7vk5uYqIyND3bt3l2ma+vXXX894vBKff/657r//fo0fP15jxowptW3p0qUqKirSww8/7L72Jem+++5TcHBwmXvvkn+nI0eOaPXq1frss8/UqVMn97cPAVQO5REA1EoxMTHur66dyq5duyTJfdMtSV9++aWefvppbdy4sVTtqvJuor755hutW7dOUnE9s6p244036oknnlBKSooaNmyoH374Qenp6brxxhvdfe6//359+umnGjBggBo2bKh+/fpp+PDhuuqqq6o8nlatWpV6XVKDtqQG1v79+yVJbdq0KbNvXFycfvrppzM+5pIlS076da0S+/fvl91uL/U1MkmKiopSaGioO64SU6dOdZeSaNWqlX744YdKl2UAAADnt1mzZql169by9PRUZGSk2rRpUypRZRiGXnnlFb3++uvau3evXC6Xe1tJCQWpuKxCmzZt5OlZtR+zT3d/tnPnTknSiBEjTjpGVlaW6tWr536dkZFRZtw/27lzp0zTPGm/P5cxKKl1+udE6Z/HzMrKOunii/T09FKvK3Kf+GeTJ09WTEyM/vKXv5R5bsXOnTu1ffv2k4755+P/UUXvg/fs2SOp+PkVVSE5OVlPPvmk/vOf/5Sp+5uVlVWpMTdu3KhPP/1ULper3M84JztXb29vNW/evMy996pVq0rNaatWrbRo0aKz+mUFAJK2AGqpa665RrNmzdLs2bN1zz33lNsnLS1N7733nho0aKBLL71UUnHd2WuvvVY9evTQ66+/rujoaHl5eWnOnDmaO3dumTF++eUX3XfffQoICNDTTz+tG264odwbtcq68cYbNWHCBM2bN08PP/ywPv30U4WEhJRKyEZERGjjxo369ttv9c033+ibb77RnDlzdMcdd+i9996rsliscskll+jpp58u1fbaa6+V+8Tmit743XvvverTp48OHjyol156SUOHDtWqVasUEhJSJTEDAIDzx8UXX6xu3bqddPuzzz6rSZMm6e6779a0adMUFhYmu92uhx9+uMwKViuUxPDCCy+oc+fO5fb5YyK1qKhIhw8fVt++fU87rs1m0zfffCMPD49TjilJqampklTqWQ3ljRkREaEPP/yw3O1/TqaeyX2iVPzch3fffVcffPBBubVxDcNQx44d9eKLL5a7f2xs7Eljt4LL5VLfvn117NgxPf7444qLi1NAQIBSUlJ05513Vvr627RpkwYMGKA+ffpo/Pjxuu222075AOXT6dSpk/7xj39IKl5pPHPmTF155ZXasGHDKa8HAKdG0hZArTRx4kQtWrRIo0aN0m+//aZbbrnFveohOTlZy5Yt05NPPqnjx49r7ty57q9pLViwQL6+vvr2229LfXVrzpw55R6nb9++euONN1RQUKBFixZp5MiR7if5VoVmzZrp4osv1ieffKIxY8Zo4cKFuu6668p8rczb21uDBg3SoEGDZBiG7r//fr311luaNGlSmdWnZ6NkpUYJ0zS1a9cuderUSZLUpEkTSdKOHTvcXyMrsWPHDvf2M1G/fv1SD06QVOYhEE2aNJFhGNq5c6f7oQxScWI+MzOzzHFbtmzpnpeEhAQ1btxYc+fOLfcBGwAAAKcyf/589erVS7Nnzy7VnpmZWerr3y1atNCaNWvkcDiq5GFaJU53f1bygLPg4OAy91Tl2bRpkxwOxykT1SXjmqapZs2aVajM1LZt22Sz2U65wKFFixZaunSpLrvsslJf+z+Zitwn/tGECRPUuXPnUt9a+/PxN23apD59+pzx/XxF74ObN28uSUpKSjqj8cuzZcsW/f7773rvvfd0xx13uNvPtmxax44dNW/ePPn5+WnevHkaOXKkNm/e7C778MdzLTkfqTjhv3fv3jL/JvXq1SvVduWVVyomJkZz5sw56QP2AJweNW0B1EpRUVFavXq1BgwYoH/84x/q0qWLPvjgA+Xm5qpJkya6++675efnpy+++EI333yzez8PDw/ZbLZSX2vbt2/fSW/+unfvLg8PDwUEBOjNN9/UihUr9M4771Tpudx44436+eef9a9//UsZGRllbjKPHj1a6rXdbnffpP+xvENVeP/993XixAn36/nz5+vw4cMaMGCAJKlbt26KiIjQm2++WerY33zzjbZv366BAwdWaTwlSp6u/PLLL5dqL1klcarjZmRkSKr6uQIAAHWDh4eHTNMs1TZv3jylpKSUahs6dKgyMjL02muvlRnjz/ufidPdn3Xt2lUtWrTQ3//+d+Xk5JTZ/8iRI2Vi9/Dw0DXXXHPK4w4ZMkQeHh6aMmVKmfhN0yx1j+p0OrVgwQJdfPHFpyyPMHz4cLlcLk2bNq3MNqfT6S6xUBmrV6/W559/rueee+6kCdnhw4crJSWl3Pv5/Px85ebmnnT8it4HN2jQQD169NC//vUvJScnlxrjTK+DkhXOf9zPNE298sorZzTOn3Xp0kUBAQGy2+365z//qX379mnq1Knu7QkJCfL29tbMmTNLHXv27NnKyso67T1/fn6+JO6/gbPFSlsAtVZsbKw+//xzHT58WCtXrtQLL7ygjRs36s0331Tnzp3VuXPnMjdsAwcO1IsvvqirrrpKt9xyi9LT0zVr1iy1bNlSmzdvPuXx+vfvr9tuu02PPfaYBg0apOjo6Co5j+HDh+vRRx/Vo48+qrCwsDK/ub733nt17Ngx9e7dW40aNdL+/fv16quvqnPnzqVWnVaFsLAwXX755brrrruUlpaml19+WS1bttR9990nqbh22fPPP6+77rpLPXv21M0336y0tDS98soratq0qcaOHVul8ZS44IILNGLECL399tvKzMxUz5499csvv+i9997Tddddp169ekkqfmDZP//5T3Xv3l1hYWHas2eP3nnnHQUEBOj6668/J7EBAIDz2zXXXKOpU6fqrrvuUvfu3bVlyxZ9+OGHpVYgSsUPDHv//fc1btw4/fLLL7riiiuUm5urpUuX6v7776/0A21Pd39WkngbMGCA2rdvr7vuuksNGzZUSkqKvv/+ewUHB+uLL75Qbm6uZs2apZkzZ6p169b64Ycf3McoSfZu3rxZq1evVnx8vFq0aKGnn35aEyZM0L59+3TdddcpKChIe/fu1WeffaaRI0fq0Ucf1dKlSzVp0iRt3rxZX3zxxSnPpWfPnvrLX/6i6dOna+PGjerXr5+8vLy0c+dOzZs3T6+88oqGDRtWqXlasmSJ+vbte8rVxrfffrs+/fRT/d///Z++//57XXbZZXK5XPrtt9/06aef6ttvvz3pCuQzuQ+eOXOmLr/8cnXp0kUjR45Us2bNtG/fPn311VfauHFjhc8pLi5OLVq00KOPPqqUlBQFBwdrwYIFZWrbno0OHTro8ccf13PPPaebbrpJnTp1UoMGDTRhwgRNmTJFV111la699lrt2LFDr7/+ui666CLddtttpcZIS0vTBx98IKl4wcRbb70lT0/P0/5iAMBpmABwnhgxYoQZEBBw2n6zZ882W7VqZfr4+JhxcXHmnDlzzMmTJ5t/fkuUZE6ePLlUW0ZGhtmgQQPz+uuvLzPunDlzTEnm3r17zzj2yy67zJRk3nvvvWW2zZ8/3+zXr58ZERFhent7m40bNzb/8pe/mIcPHz7lmHv37jUlmS+88EKZbe3btzd79uzpfv3999+bksyPPvrInDBhghkREWH6+fmZAwcONPfv319m/08++cS88MILTR8fHzMsLMy89dZbzYMHD5bqU5F/jyZNmpgDBw4s0z569Ogy/x4Oh8OcMmWK2axZM9PLy8uMjY01J0yYYBYUFLj7JCUlmf369TPDw8NNb29vMzY21rzpppvMzZs3nzIOAABQ95Tcu61du/aU/QoKCsxHHnnEjI6ONv38/MzLLrvMXL16tdmzZ89S91OmaZp5eXnm3/72N/f9SlRUlDls2DBz9+7dpmme2/uzX3/91RwyZIgZHh5u+vj4mE2aNDGHDx9uLlu2rNSxT/dnxIgRpcZdsGCBefnll5sBAQFmQECAGRcXZ44ePdrcsWOHaZqm+cADD5g9evQwFy9eXCam8u6xTdM03377bbNr166mn5+fGRQUZHbs2NF87LHHzEOHDrn7nMl9oiTTZrOZ69evL9Ve3r9RUVGR+fzzz5vt27c3fXx8zHr16pldu3Y1p0yZYmZlZZU53p9V5D7YNIvvS6+//nozNDTU9PX1Ndu0aWNOmjSp3DGbNGlSZt5LbNu2zUxISDADAwPN+vXrm/fdd5+5adMmU5I5Z86ccvcp73PMqY5VUFBgxsXFmRdddJHpdDrd7a+99poZFxdnenl5mZGRkeaoUaPM48ePl9q3Z8+epa6f0NBQ87LLLjO//vrrco8PoOJspnkW39MAAJwXfvjhB/Xq1Uvz5s2r9OoGAAAAVJ2qvj/bt2+fmjVrpr1796pp06bl9nnqqae0b98+vfvuu2d9PADA2aGmLQAAAAAAAADUINS0BQAAAADgPBcYGKhbb731lA8K69Spk2JiYqoxKgDAyZC0BQAAAADgPFe/fn33w6JOZsiQIdUUDQDgdKhpCwAAAAAAAAA1iKU1bVesWKFBgwYpJiZGNptNixYtKrXdNE09+eSTio6Olp+fnxISErRz585SfY4dO6Zbb71VwcHBCg0N1T333KOcnJxqPAsAAAAAAAAAqDqWJm1zc3N1wQUXaNasWeVunzFjhmbOnKk333xTa9asUUBAgPr376+CggJ3n1tvvVVbt25VYmKivvzyS61YsUIjR46srlMAAAAAAAAAgCpVY8oj2Gw2ffbZZ7ruuuskFa+yjYmJ0SOPPKJHH31UkpSVlaXIyEi9++67uummm7R9+3a1a9dOa9euVbdu3SRJixcv1tVXX62DBw9WuIC6YRg6dOiQgoKCZLPZzsn5AQAAoOJM09SJEycUExMju93SdQa1Gve5AAAANUtF73Nr7IPI9u7dq9TUVCUkJLjbQkJCdMkll2j16tW66aabtHr1aoWGhroTtpKUkJAgu92uNWvW6Prrr6/QsQ4dOqTY2NgqPwcAAACcnQMHDqhRo0ZWh1FrcZ8LAABQM53uPrfGJm1TU1MlSZGRkaXaIyMj3dtSU1MVERFRarunp6fCwsLcfcpTWFiowsJC9+uSxcZ79+5VUFBQlcSPU3M4HPr+++/Vq1cveXl5WR0OLMJ1AInrAMW4DlCi5Fro1q2bWrduzb3ZWSqZvwMHDig4ONjiaM4th8OhJUuWqF+/fryPVDPm3jrMvXWYe+sw99Zh7qtGdna2YmNjT3ufW2OTtufS9OnTNWXKlDLtq1evlr+/vwUR1U3+/v5as2aN1WHAYlwHkLgOUIzrACX8/f21bt06SeIr/WepZP6Cg4PrRNLW399fwcHBfJCsZsy9dZh76zD31mHurcPcV63T3efW2KRtVFSUJCktLU3R0dHu9rS0NHXu3NndJz09vdR+TqdTx44dc+9fngkTJmjcuHHu1yUZ7n79+p33N7M1hcPhUGJiovr27cv/0eswrgNIXAcoxnWAEiXXQq9evawOBQAAALBMjU3aNmvWTFFRUVq2bJk7SZudna01a9Zo1KhRkqT4+HhlZmZq/fr16tq1qyTpu+++k2EYuuSSS046to+Pj3x8fMq0e3l58UGxmjHnkLgOUIzrABLXAf6H6wAAAAB1maVJ25ycHO3atcv9eu/evdq4caPCwsLUuHFjPfzww3r66afVqlUrNWvWTJMmTVJMTIyuu+46SVLbtm111VVX6b777tObb74ph8OhMWPG6KabblJMTIxFZwUAAAAAAAAAlWdp0nbdunWlvvpWUrJgxIgRevfdd/XYY48pNzdXI0eOVGZmpi6//HItXrxYvr6+7n0+/PBDjRkzRn369JHdbtfQoUM1c+bMaj8XAAAAAAAAAKgKliZtr7zySpmmedLtNptNU6dO1dSpU0/aJywsTHPnzj0X4QEAAAAAAABAtbNbHQAAAAAAAAAA4H9I2gIAAAAAAABADULSFgAAAAAAAABqEJK2AAAAAAAAAFCDkLQFAAAAAAAAgBqEpC0AAAAAAAAA1CAkbQEAAAAAAACgBiFpCwAAAAAAAAA1CElbAAAAAAAAAKhBSNoCAADglAzDVKHTZXUYAAAAQJ1B0hYAAAAntedIjm58e7VmLN5hdSgAAABAneFpdQAAAACoeVyGqdk/7dE/lvyuQqeh3w6f0JheLVUvwNvq0FBHJScnKyMj46TbDcOQJG3atEl2+5mtTalfv74aN258VvEBAABUJZK2AAAAKOX3tBMaP3+zNh3IlCRd0aq+pg/pSMIWlklOTlabuLYqyM87aR8/Pz999NFH6tGjh/Lz889ofF8/f+34bTuJWwAAUGOQtAUAAIDbrO936eWlv8vhMhXk66lJ17TTDV0byWazWR0a6rCMjAwV5Ocp/JpH5BUeW24fX8/iazTyludU4DQrPLbj6AEd/fIfysjIIGkLAABqDJK2AAAAcDueWySHy1RC2wg9c31HRQb7Wh0S4OYVHiufqJblbvP2MCW55B3ZXKaLXzIAAIDajaQtAABAHVbgcCk736GI/yZnH+nXRl2b1NNVHaJYXQsAAABY5Mwq9AMAAOC8sSH5uK559SeN+nCDXEbx18n9vD00oGM0CVsAAADAQqy0BQAAqGPyi1z6x5Idmr1yr0xTqh/oo+RjeWpWP8Dq0AAAAACIpC0AAECd8vOeo3p8wWbtP5onSRrSpaGevKadQv29LY4MAAAAQAmStgAAAHVAfpFLz369Xf/+eb8kKTrEV89e31G94iIsjgwAAADAn5G0BQAAqAM8PWzakHxcknTzxY014eo4Bft6WRwVAAAAgPKQtAUAADhPZeU75Otll4+nh7w87Pr7DRfoWG6RLmtZ3+rQAAAAAJyC3eoAAAAAUPWWbktTv5eWa9b3u91tbaODSdgCAAAAtQBJWwAAgPPIsdwiPfTxr7r3/XVKyy7U4qTDKnIaVocFAAAA4AxQHgEAAOA8YJqmvt6Sqic/T9LR3CLZbdJ9PZprbEJreXvye3oAAACgNiFpCwAAUMsdOVGoSYuStHhrqiSpdWSgXhh2gS6IDbU2MAAAAACVQtIWAACglssrcmr570fkabfp/l4tNbpXC/l4elgdFgAAAIBKImkLAABQC50ocCjI10uS1CQ8QDOGdVKLBoFqFxNscWQAAAAAzhYFzgAAAGoR0zQ1d02yuj/3ndbsOepuH3RBDAlbAAAA4DzBSlsAAIBaIvlonv66cLNW7S5O1n70S7IuaR5ucVQAAAAAqhpJWwAAgBrOMEy9v3qfnl+8Q/kOl3y97BrfP053dm9qdWgAAAAAzgGStgAAADXYniM5emz+Zq3bf1ySdEmzMD0/tJOa1g+wODIAAAAA5wpJWwAAgBps44FMrdt/XAHeHvrr1W1168WNZbfbrA4LAAAAwDnEg8gAAABqmAKHy/3z9Rc21MMJrfTt2B66/dImJGxriaZNm8pms5X5M3r0aElSQUGBRo8erfDwcAUGBmro0KFKS0srNUZycrIGDhwof39/RUREaPz48XI6nVacDgAAAKoZSVsAAIAaoshp6JWlO9XnH8uVleeQJNlsNj2c0FqN6vlbHB3OxNq1a3X48GH3n8TEREnSDTfcIEkaO3asvvjiC82bN0/Lly/XoUOHNGTIEPf+LpdLAwcOVFFRkVatWqX33ntP7777rp588klLzgcAAADVi6QtAABADbDlYJaufe0nvbT0d6Vk5mvhrwetDglnoUGDBoqKinL/+fLLL9WiRQv17NlTWVlZmj17tl588UX17t1bXbt21Zw5c7Rq1Sr9/PPPkqQlS5Zo27Zt+uCDD9S5c2cNGDBA06ZN06xZs1RUVGTx2QEAAOBco6YtAACAhQocLs1ctlNvrdgjl2EqLMBbT13bXoM6RVsdGqpIUVGRPvjgA40bN042m03r16+Xw+FQQkKCu09cXJwaN26s1atX69JLL9Xq1avVsWNHRUZGuvv0799fo0aN0tatW3XhhReWe6zCwkIVFha6X2dnZ0uSHA6HHA7HOTrDc88wDPn5+cnX0yZvD7PcPj52s9TfFWXztMnPz0+GYdTqObJSybwxf9WPubcOc28d5t46zH3VqOj8kbQFAACwyPr9x/XY/E3afSRXknRNp2hNuba9wgN9LI4MVWnRokXKzMzUnXfeKUlKTU2Vt7e3QkNDS/WLjIxUamqqu88fE7Yl20u2ncz06dM1ZcqUMu1LliyRv3/tLrHx0Ucf/fcn1yn7TetmnOHITaRBHyklJUUpKSmVig3FSsqAoPox99Zh7q3D3FuHuT87eXl5FepH0hYAAMAic9cka/eRXDUI8tG0wR10VYcoq0PCOTB79mwNGDBAMTEx5/xYEyZM0Lhx49yvs7OzFRsbq379+ik4OPicH/9c2bRpk3r06KHIW56Td2Tzcvv42E1N62Zo0jq7Co2KP7CvKG2P0ub+VStWrNAFF1xQVSHXKQ6HQ4mJierbt6+8vLysDqdOYe6tw9xbh7m3DnNfNUq+CXU6JG0BAACqkdNlyNOj+LECk65pqyBfTz2c0Eqh/t4WR4ZzYf/+/Vq6dKkWLlzobouKilJRUZEyMzNLrbZNS0tTVFSUu88vv/xSaqy0tDT3tpPx8fGRj0/ZldpeXl61+sOV3W5Xfn6+CpymTNepE7KFhk2Fp+lTqr/TVH5+vux2e62eo5qgtl9ntRlzbx3m3jrMvXWY+7NT0bnjQWQAAADV4ESBQ3/7bIv+74MNMs3impuh/sX1a0nYnr/mzJmjiIgIDRw40N3WtWtXeXl5admyZe62HTt2KDk5WfHx8ZKk+Ph4bdmyRenp6e4+iYmJCg4OVrt27arvBAAAAGAJVtoCAACcYz/sSNcTC7foUFaBJGnTwSx1jg21Niicc4ZhaM6cORoxYoQ8Pf932x0SEqJ77rlH48aNU1hYmIKDg/XAAw8oPj5el156qSSpX79+ateunW6//XbNmDFDqampmjhxokaPHl3uSloAAACcX0jaAgAAnCNZeQ5N+2qb5q8/KElqHOav54Z2JGFbRyxdulTJycm6++67y2x76aWXZLfbNXToUBUWFqp///56/fXX3ds9PDz05ZdfatSoUYqPj1dAQIBGjBihqVOnVucpAAAAwCIkbQEAAM6BJVtT9bdFSTpyolA2m3RX92Z6tH9r+Xtz+1VX9OvXz10K4898fX01a9YszZo166T7N2nSRF9//fW5Cg8AAAA1GJ8aAAAAqlih06Wnv9quIycK1bxBgF4Y1kldm4RZHRYAAACAWoKkLQAAQBUoWVFps9nk4+mh54d20vLfj+jhhFby9fKwODoAAAAAtQlJWwAAgLOUnl2giYuS1L1FuO68rJkkKb5FuOJbhFscGQAAAIDaiKQtAABAJZmmqQUbUjT1i63KLnDq5z1HdUO3WAX4cIsFAAAAoPL4RAEAAFAJKZn5emLhFi3//YgkqWPDEM0Y1omELQAAAICzxqcKAACAM2AYpj5am6zpX/+mnEKnvD3tGpvQWvdd0UyeHnarwwMAAABwHiBpCwAAcAb2ZORq8udb5TRMdW1STzOGdVKLBoFWhwUAAADgPELSFgAA4Ay0jAjU2L6t5e/toTvim8rDbrM6JAAAAADnGb7DBwAAcAq70nN041urte1QtrttdK+WuuuyZiRsAQAAAJwTJG0BAADK4XQZev2HXbp65o9as/eYpn651eqQAAAAANQRlEcAAAD4k+2Hs/XY/M3akpIlSbqyTQM9e31Hi6MCAAAAUFeQtAUAAPivIqehWd/v0qzvd8lpmAr29dTkQe01pEtD2WyUQgAAAABQPUjaAgAA/NeiX1P0yrKdkqR+7SL19HUdFBHsa3FUAAAAAOoakrYAAAD/NbRrIy3ZlqbBnWN0TadoVtcCAAAAsAQPIgMAAHXW+v3HdPe7a5Vf5JIkedht+ueIbhp0QQwJWwAAAACWIWkLAADqnLwip6Z8sVXD3lyt735L1xvLd1sdEgAAAAC4UR4BAADUKat2ZejxhZt14Fi+JOmGro10z2XNLI4KAAAAAP6HpC0AAKgTsgscmv71b/rol2RJUkyIr6YP7aSerRtYHBkAAAAAlEbSFgAA1AnTvtimeesPSpJuu7SxHr8qTkG+XhZHBQAAAABlkbQFAAB1wrh+rbXtcLYmDmyn+BbhVocDAAAAACdF0hYAAJyXvt2aqg3JxzVhQFtJUnSIn7584HLZbDaLIwMAAACAUyNpCwAAzitHcwo1+T9b9eXmw5Kknq0aqHvL+pJEwhYAAABArUDSFgAAnBdM09QXmw/rqf9s1bHcInnYbfq/ns3VpUk9q0MDAAAAgDNC0hYAANR6adkFmrgoSYnb0iRJcVFB+vsNF6hDwxCLIwMAAACAM0fSFgAA1GqGYeqWd37W7iO58vKwaUyvVhp1ZQt5e9qtDg0AAAAAKoWkLQAAqNXsdpse7ddGbyzfrRnDOikuKtjqkAAAAADgrJC0BQAAtYphmPrwl2SF+XtrYKdoSdKAjtHq1z5KHnYeNAYAAACg9iNpCwAAao19Gbl6fMFmrdl7TPX8vXRp8zCFB/pIEglbAAAAAOcNkrYAAKDGcxmm5qzcq78v2aEChyE/Lw891KeV6vl7Wx0aAAAAAFQ5krYAAKBG25V+QuPnb9avyZmSpO4twvXckE5qHO5vbWAAAAAAcI6QtAUAADVWalaBBs78SYVOQ4E+nvrbwLa66aJY2WyUQgAAAABw/rJbHcDpnDhxQg8//LCaNGkiPz8/de/eXWvXrnVvN01TTz75pKKjo+Xn56eEhATt3LnTwogBAEBViQrx1bCujdSrTQMtGdtDN1/cmIQtAAAAgPNejU/a3nvvvUpMTNS///1vbdmyRf369VNCQoJSUlIkSTNmzNDMmTP15ptvas2aNQoICFD//v1VUFBgceQAAOBMFTkNvbJ0pw4cy3O3TR7UXv+68yLFhPpZGBkAAAAAVJ8anbTNz8/XggULNGPGDPXo0UMtW7bUU089pZYtW+qNN96QaZp6+eWXNXHiRA0ePFidOnXS+++/r0OHDmnRokVWhw8AAM7A/hzp+jd+1ktLf9dfF26WaZqSJG9PO6trAQAAANQpNbqmrdPplMvlkq+vb6l2Pz8//fTTT9q7d69SU1OVkJDg3hYSEqJLLrlEq1ev1k033VTuuIWFhSosLHS/zs7OliQ5HA45HI5zcCb4s5J5Zr7rNq4DSFwHkAocLr28dKf+tcVDpnIUFuClG7s2lMPhIFlbB/GeAAAAANTwpG1QUJDi4+M1bdo0tW3bVpGRkfroo4+0evVqtWzZUqmpqZKkyMjIUvtFRka6t5Vn+vTpmjJlSpn2JUuWyN+fJ1FXp8TERKtDQA3AdQCJ66Cu2p0tfbzbQ+kFNkk2da1vaEjTfJnJG/RNstXRwUrff/+91SEAAAAAlqnRSVtJ+ve//627775bDRs2lIeHh7p06aKbb75Z69evr/SYEyZM0Lhx49yvs7OzFRsbq379+ik4OLgqwsZpOBwOJSYmqm/fvvLy8rI6HFiE6wAS10Fd9v2OI3r1w19lmlKDQG8NbpivcTcmcB3UcSXvCb169bI6FAAAAMAyNT5p26JFCy1fvly5ubnKzs5WdHS0brzxRjVv3lxRUVGSpLS0NEVHR7v3SUtLU+fOnU86po+Pj3x8fMq0e3l58UGxmjHnkLgOUIzroO7pGReplg0C1aVxPT3Wr6V++j6R6wBuXAcAAACoy2r0g8j+KCAgQNHR0Tp+/Li+/fZbDR48WM2aNVNUVJSWLVvm7pedna01a9YoPj7ewmgBAMCfZRc49OqynXK6DEmSj6eHFo2+TM8P66RgPxJ0AAAAAFCixq+0/fbbb2Waptq0aaNdu3Zp/PjxiouL01133SWbzaaHH35YTz/9tFq1aqVmzZpp0qRJiomJ0XXXXWd16AAA4L+++y1NTyxMUmp2gTw97Bp1ZQtJUoBPjb8VAQAAAIBqV+NX2mZlZWn06NGKi4vTHXfcocsvv1zffvut+ytzjz32mB544AGNHDlSF110kXJycrR48WL5+vpaHDkAAMjMK9K4Tzbq7nfXKTW7QE3C/XVh41CrwwKqRUpKim677TaFh4fLz89PHTt21Lp169zbTdPUk08+qejoaPn5+SkhIUE7d+4sNcaxY8d06623Kjg4WKGhobrnnnuUk5NT3acCAACAalbjl7cMHz5cw4cPP+l2m82mqVOnaurUqdUYFQAAOJ3FSYc1cdFWZeQUymaT7rmsmR7p10Z+3h5Whwacc8ePH9dll12mXr166ZtvvlGDBg20c+dO1atXz91nxowZmjlzpt577z33N8b69++vbdu2uRcg3HrrrTp8+LASExPlcDh01113aeTIkZo7d65VpwYAAIBqUOOTtgAAoPZ5eenvenlp8YrBlhGBmjGsk7o0rneavYDzx/PPP6/Y2FjNmTPH3dasWTP3z6Zp6uWXX9bEiRM1ePBgSdL777+vyMhILVq0SDfddJO2b9+uxYsXa+3aterWrZsk6dVXX9XVV1+tv//974qJianekwIAAEC1qfHlEQAAQO0zsGO0/L09NLpXC335wOUkbFHn/Oc//1G3bt10ww03KCIiQhdeeKHeeecd9/a9e/cqNTVVCQkJ7raQkBBdcsklWr16tSRp9erVCg0NdSdsJSkhIUF2u11r1qypvpMBAABAtWOlLQAAOGtp2QVauStDQ7o0kiS1igzST4/3VliAt8WRAdbYs2eP3njjDY0bN05PPPGE1q5dqwcffFDe3t4aMWKEUlNTJUmRkZGl9ouMjHRvS01NVURERKntnp6eCgsLc/f5s8LCQhUWFrpfZ2dnS5IcDoccDkeVnV91MwxDfn5+8vW0ydvDLLePj90s9XdF2Txt8vPzk2EYtXqOrFQyb8xf9WPurcPcW4e5tw5zXzUqOn8kbQEAQKWZpql56w5q2lfblFvoVIsGgbogNlSSSNiiTjMMQ926ddOzzz4rSbrwwguVlJSkN998UyNGjDhnx50+fbqmTJlSpn3JkiXy9/c/Z8etDh999NF/f3Kdst+0bsYZjtxEGvSRUlJSlJKSUqnYUCwxMdHqEOos5t46zL11mHvrMPdnJy8vr0L9SNoCAIBKOXg8TxMWbtGPOzMkSRc0ClGADw8ZAyQpOjpa7dq1K9XWtm1bLViwQJIUFRUlSUpLS1N0dLS7T1pamjp37uzuk56eXmoMp9OpY8eOuff/swkTJmjcuHHu19nZ2YqNjVW/fv0UHBx81udllU2bNqlHjx6KvOU5eUc2L7ePj93UtG6GJq2zq9CwVXjsorQ9Spv7V61YsUIXXHBBVYVcpzgcDiUmJqpv377y8vKyOpw6hbm3DnNvHebeOsx91Sj5JtTpkLQFAABnxDBMfbhmv5775jflFrnk42nXI/1a6+7LmsnTg3L5gCRddtll2rFjR6m233//XU2aNJFU/FCyqKgoLVu2zJ2kzc7O1po1azRq1ChJUnx8vDIzM7V+/Xp17dpVkvTdd9/JMAxdcskl5R7Xx8dHPj4+Zdq9vLxq9Ycru92u/Px8FThNma5TJ2QLDZsKT9OnVH+nqfz8fNnt9lo9RzVBbb/OajPm3jrMvXWYe+sw92enonNH0hYAAFSYaZoa+e91Wrq9ePXfxU3D9NzQjmreINDiyICaZezYserevbueffZZDR8+XL/88ovefvttvf3225Ikm82mhx9+WE8//bRatWqlZs2aadKkSYqJidF1110nqXhl7lVXXaX77rtPb775phwOh8aMGaObbrpJMTExFp4dAAAAzjWStgAAoMJsNpt6xUVo1e6j+uuAON12SRPZ7RVf0QbUFRdddJE+++wzTZgwQVOnTlWzZs308ssv69Zbb3X3eeyxx5Sbm6uRI0cqMzNTl19+uRYvXixfX193nw8//FBjxoxRnz59ZLfbNXToUM2cOdOKUwIAAEA1ImkLAABO6fe0E8rOd6hb0zBJ0s0XNVafuEhFhfieZk+gbrvmmmt0zTXXnHS7zWbT1KlTNXXq1JP2CQsL09y5c89FeAAAAKjBSNoCAIByOVyG3lq+WzOX7VKDIB99O7aHAn08ZbfbSNgCAAAAwDlE0hYAAJSx9VCWxs/brG2Hi59s2iYqSAUOlwJ9uHUAAAAAgHONT14AAMCt0OnSa9/t0hs/7JbTMBXq76WnBrXX4M4xstmoXQsAAAAA1YGkLQAAkCRl5Tk07M1V2pmeI0ka0CFKUwd3UIMgH4sjAwAAAIC6haQtAACQJIX4e6llRKCO5xVp6uAOurpjtNUhAQAAAECdRNIWAIA6bO2+Y2pWP0D1A4tX0z59XQfZbDaFBXhbHBkAAAAA1F12qwMAAADVL7fQqcmfJ+mGN1frqf9sdbeHB/qQsAUAAAAAi7HSFgCAOuannRn668LNOng8X5IU6OMpp8uQpwe/ywUAAACAmoCkLQAAdUR2gUPPfrVdH689IElqGOqn54Z21BWtGlgcGQAAAADgj0jaAgBQB2w9lKV73l2n1OwCSdKI+CZ67Ko4BfhwKwAAAAAANQ2f1AAAqAMah/nLZpOahvtrxrALdHGzMKtDAgAAAACcBElbAADOU6t2Zyi+ebhsNpuCfL303t0XK7aev/y8PawODQAAAABwCjxxBACA88yRE4Ua9cF63fLOGn267oC7vXVkEAlbAAAAAKgFWGkLAMB5wjRNLdqYoilfbFNmnkOedpuO5TqsDgsAAAAAcIZI2gIAcB44nJWvv32WpO9+S5cktY8J1oxhndQ+JsTiyAAAAAAAZ4qkLQAAtdzipMMaP2+zThQ65e1h10MJrTSyR3N5eVAFCQAAAABqI5K2AADUchHBvsopcqpzbKheGNZJrSKDrA4JAAAAAHAWSNoCAFDLGIapbYez1aFhcemDLo3rae69l+riZmHysNssjg4AAAAAcLb43iQAALXIniM5uvHt1RryxirtPpLjbo9vEU7CFgAAAADOE6y0BQCgFnAZpmb/tEf/WPK7Cp2G/L09tDMtRy0aBFodGgAAAACgipG0BQCghvs97YTGz9+sTQcyJUlXtKqvZ6/vqNgwf2sDAwAAAACcEyRtAQCowd74YbdeTNwhh8tUkK+nJg1spxu6NZLNRikEAAAAADhfkbQFAKAGM0xTDpephLYRevq6jooK8bU6JAAAAADAOUbSFgCAGqTQ6VJ6dqG79MHIHs3VOjJICW0jWF0LAAAAAHWE3eoAAABAsV+Tj2vgzJ90z3trVeh0SZK8POzq2y6ShC0AAAAA1CGstAUAwGL5RS69mLhDs3/aK8OU6gf6aF9GntpEBVkdGgAAAADAAiRtAQCw0Jo9R/X4gs3adzRPkjSkS0M9eU07hfp7WxwZAAAAAMAqJG0BALBAodOlZ77arvdX75ckRYf46tnrO6pXXITFkQEAAAAArEbSFgAAC3jZ7dqZliNJuvnixppwdZyCfb0sjgoAAAAAUBOQtAUAoJpk5TvkabcpwMdTdrtNzw/tpAPH83RZy/pWhwYAAAAAqEHsVgcAAEBdkLgtTf1eWq7nF//mbmsc7k/CFgAAAABQBittAQA4h47lFmnKF1v1+cZDkqSfdmUor8gpf2/+EwwAAAAAKB+fGAEAOAdM09RXWw5r8udbdTS3SHabdF+P5hqb0Fq+Xh5WhwcAAAAAqMFI2gIAUMUycgr1t8+26NutaZKk1pGBemHYBbogNtTawAAAAAAAtQJJWwAAqphhmlqz95g87Tbd36ulRvdqIR9PVtcCAAAAACqGpC0AAFXgeG6R6gV4S5Iignz10vDOigz2VbuYYIsjAwAAAADUNnarAwAAoDYzTVMf/ZKsHjO+17dbU93tveIiSNgCAAAAACqFlbYAAFTSgWN5+uvCzVq566gkaeGGg+rfPsriqAAAAAAAtR1JWwAAzpBhmHp/9T49v3iH8h0u+XrZNb5/nO7s3tTq0AAAAAAA5wGStgAAnIE9R3L02PzNWrf/uCTp0uZhen5oJzUJD7A4MgAAAADA+YKatgAAnIH9R/O0bv9xBXh76OnrOmjuvZeSsAVQxlNPPSWbzVbqT1xcnHt7QUGBRo8erfDwcAUGBmro0KFKS0srNUZycrIGDhwof39/RUREaPz48XI6ndV9KgAAALAAK20BADiN3EKnAnyK/5PZKy5Ck65pp6s6RKlhqJ/FkQGoydq3b6+lS5e6X3t6/u/We+zYsfrqq680b948hYSEaMyYMRoyZIhWrlwpSXK5XBo4cKCioqK0atUqHT58WHfccYe8vLz07LPPVvu5AAAAoHqx0hYAgJMochp6ZelO9ZjxvVKzCtzt91zejIQtgNPy9PRUVFSU+0/9+vUlSVlZWZo9e7ZefPFF9e7dW127dtWcOXO0atUq/fzzz5KkJUuWaNu2bfrggw/UuXNnDRgwQNOmTdOsWbNUVFRk5WkBAACgGpC0BQCgHFsOZuna137SS0t/19HcIi3YcNDqkADUMjt37lRMTIyaN2+uW2+9VcnJyZKk9evXy+FwKCEhwd03Li5OjRs31urVqyVJq1evVseOHRUZGenu079/f2VnZ2vr1q3VeyIAAACodpRHAADgDwocLs1ctlNvrdgjl2Gqnr+XpgzuoEGdoq0ODUAtcskll+jdd99VmzZtdPjwYU2ZMkVXXHGFkpKSlJqaKm9vb4WGhpbaJzIyUqmpqZKk1NTUUgnbku0l206msLBQhYWF7tfZ2dmSJIfDIYfDURWnZgnDMOTn5ydfT5u8Pcxy+/jYzVJ/V5TN0yY/Pz8ZhlGr58hKJfPG/FU/5t46zL11mHvrMPdVo6LzR9IWAID/Wr//uB6bv0m7j+RKkq7pFK2nrm2v+oE+FkcGoLYZMGCA++dOnTrpkksuUZMmTfTpp5/Kz+/clVeZPn26pkyZUqZ9yZIl8vf3P2fHrQ4fffTRf39ynbLftG7GGY7cRBr0kVJSUpSSklKp2FAsMTHR6hDqLObeOsy9dZh76zD3ZycvL69C/UjaAgDwX19tPqzdR3JVP9BHT1/XQVd1iLI6JADnidDQULVu3Vq7du1S3759VVRUpMzMzFKrbdPS0hQVVfy+ExUVpV9++aXUGGlpae5tJzNhwgSNGzfO/To7O1uxsbHq16+fgoODq/CMqtemTZvUo0cPRd7ynLwjm5fbx8dualo3Q5PW2VVo2Co8dlHaHqXN/atWrFihCy64oKpCrlMcDocSExPVt29feXl5WR1OncLcW4e5tw5zbx3mvmqUfBPqdEjaAgDqtCKnIW/P4hLvj/ZvLU8Pm+6/soVC/b0tjgzA+SQnJ0e7d+/W7bffrq5du8rLy0vLli3T0KFDJUk7duxQcnKy4uPjJUnx8fF65plnlJ6eroiICEnFq1qCg4PVrl27kx7Hx8dHPj5lvx3g5eVVqz9c2e125efnq8BpynSdOiFbaNhUeJo+pfo7TeXn58tut9fqOaoJavt1Vpsx99Zh7q3D3FuHuT87FZ07krYAgDopp9Cp577Zrt9Tc/TxyEtlt9vk7+2pJ65ua3VoAM4Djz76qAYNGqQmTZro0KFDmjx5sjw8PHTzzTcrJCRE99xzj8aNG6ewsDAFBwfrgQceUHx8vC699FJJUr9+/dSuXTvdfvvtmjFjhlJTUzVx4kSNHj263KQsAAAAzi8kbQEAdc7y34/oiYVblJKZL0n6ec9RdW9Z3+KoAJxPDh48qJtvvllHjx5VgwYNdPnll+vnn39WgwYNJEkvvfSS7Ha7hg4dqsLCQvXv31+vv/66e38PDw99+eWXGjVqlOLj4xUQEKARI0Zo6tSpVp0SAAAAqhFJWwBAnZGV59DTX23TvPUHJUmxYX56fkgnErYAqtzHH398yu2+vr6aNWuWZs2addI+TZo00ddff13VoQEAAKAWIGkLAKgTlmxN1d8WJenIiULZbNKd3ZtqfP828vfmP4UAAAAAgJqFT6oAgPOeyzD18tKdOnKiUM0bBOiFYZ3UtUmY1WEBAAAAAFAukrYAgPOSaZoyTMnDbpOH3aYXbuikLzcf1kN9WsnXy8Pq8AAAAAAAOCmStgCA8056doEmLkpS2+hgje3bWpLUPiZE7WNCLI4MAAAAAIDTI2kLADhvmKapBRtSNPWLrcoucOrHnRm6s3tT1Qvwtjo0AAAAAAAqjKQtAOC8kJKZrycWbtHy349Ikjo2DNGMYZ1I2AIAAAAAah2StgCAWs0wTH20NlnTv/5NOYVOeXvaNTahte67opk8PexWhwcAAAAAwBkjaQsAqNUOZxdo2pfbVOAw1KVxqGYMu0AtIwKtDgsAAAAAgEojaQsAqHVM05TNZpMkNQz109+ubqsil6k7uzeVh91mcXQAAAAAAJwdvjcKAKhVdqXnaPhbq/XL3mPuttvjm+qey5uRsAUAAAAAnBdI2gIAagWny9AbP+zW1TN/1Np9xzX1y60yTdPqsAAAAAAAqHKURwAA1Hi/pWZr/LzN2pKSJUnq0bqBpg/p6C6RAAAAAADA+YSkLQCgxipyGpr1/S69/sMuOVymgn099eSg9hrapSEJWwAAAADAeYukLQCgxvrutzS9smynJKlvu0g9c10HRQT7WhwVAAAAAADnVo2uaetyuTRp0iQ1a9ZMfn5+atGihaZNm1aqhqFpmnryyScVHR0tPz8/JSQkaOfOnRZGDQCoKv3bR2lol0Z69eYL9fbtXUnYAgAAAADqhBqdtH3++ef1xhtv6LXXXtP27dv1/PPPa8aMGXr11VfdfWbMmKGZM2fqzTff1Jo1axQQEKD+/furoKDAwsgBAJWxfv8x3fz2z8rKc0iSbDab/jH8Ag26IIZyCAAAAACAOqNGl0dYtWqVBg8erIEDB0qSmjZtqo8++ki//PKLpOJVti+//LImTpyowYMHS5Lef/99RUZGatGiRbrpppssix0AUHGFLunpr3/T+z8nyzSll5f9rsmD2lsdFgAAAAAAlqjRK227d++uZcuW6ffff5ckbdq0ST/99JMGDBggSdq7d69SU1OVkJDg3ickJESXXHKJVq9ebUnMAIAz8/OeY3p+k4feW12csL2hayM93Ke11WEBAAAAAGCZGr3S9q9//auys7MVFxcnDw8PuVwuPfPMM7r11lslSampqZKkyMjIUvtFRka6t5WnsLBQhYWF7tfZ2dmSJIfDIYfDUdWngXKUzDPzXbdxHdRtJwqcemHJ7/po7UFJNkUF++jZ69rrilb1JXFd1DW8H6AE1wIAAABQw5O2n376qT788EPNnTtX7du318aNG/Xwww8rJiZGI0aMqPS406dP15QpU8q0L1myRP7+/mcTMs5QYmKi1SGgBuA6qJsW7rVreWrxFz4uizR0beNcndj5i77mWZJ1Gu8HKPH9999bHQIAAABgmRqdtB0/frz++te/umvTduzYUfv379f06dM1YsQIRUVFSZLS0tIUHR3t3i8tLU2dO3c+6bgTJkzQuHHj3K+zs7MVGxurfv36KTg4+NycDEpxOBxKTExU37595eXlZXU4sAjXQd3WPc+hv3z4qx7o2VTZu9ZxHdRxvB+gRMm10KtXL6tDAQAAACxT6aRtSSkBf39/ORwObdmyRa1atVJQUFCVBZeXlye7vXTZXQ8PDxmGIUlq1qyZoqKitGzZMneSNjs7W2vWrNGoUaNOOq6Pj498fHzKtHt5efFBsZox55C4DuqKb7em6ocd6Xr2+o6y2WxqEOKlhfdfJofDoa93cR2gGNcBSnAdAAAAoC6r1IPIFi9erNDQUEVGRmrJkiXq1q2bunXrpkaNGmnlypVVFtygQYP0zDPP6KuvvtK+ffv02Wef6cUXX9T1118vSbLZbHr44Yf19NNP6z//+Y+2bNmiO+64QzExMbruuuuqLA4AQOUdzSnUmLkb9Jd/r9dHvxzQt1tPXnMcAAAAAABUcqXtxIkT9eCDD6p58+a65ZZbNHz4cP3444965JFHNHHixCqrQfbqq69q0qRJuv/++5Wenq6YmBj95S9/0ZNPPunu89hjjyk3N1cjR45UZmamLr/8ci1evFi+vr5VEgMAoHJM09QXmw/rqf9s1bHcInnYbfpLj+a6sk2E1aEBwCnl5uZq+fLlSk5OVlFRUaltDz74oEVRAQAAoC6pVNJ227Ztmj9/vpo2baoxY8bo//7v/xQcHKyHH35YV1xxRZUFFxQUpJdfflkvv/zySfvYbDZNnTpVU6dOrbLjAgDOTlp2gf72WZKWbk+TJMVFBemFYReoY6MQiyMDgFP79ddfdfXVVysvL0+5ubkKCwtTRkaG/P39FRERQdIWAAAA1aJS5RG8vb3lcrkkSa1atVK9evUkyV3fFgBQd5mmqXvfW6el29Pk5WHT2ITW+s+Yy0nYAqgVxo4dq0GDBun48ePy8/PTzz//rP3796tr1676+9//bnV4AAAAqCMqlbRt06aNtm7dKklKSkpSbGyspOIVuK1ataq66AAAtY7NZtOEq+N0QaMQffHA5XoooZW8PSv1nxsAqHYbN27UI488IrvdLg8PDxUWFio2NlYzZszQE088YXV4AAAAqCMqVR5hyZIl8vb2LtPesGFDzZo166yDAgDUHoZh6sNfkuVpt+nmixtLkrq3qK9Foy+TzWazODoAODNeXl6y24t/0RQREaHk5GS1bdtWISEhOnDggMXRAQAAoK6oVNI2JKT8r7h27tz5bGIBANQy+zJy9fiCzVqz95j8vDzUs3UDxYT6SRIJWwC10oUXXqi1a9eqVatW6tmzp5588kllZGTo3//+tzp06GB1eAAAAKgjKv191eXLl2vQoEFq2bKlWrZsqWuvvVY//vhjVcYGAKihXIapf/64R1e9ssKdsH38qjaKCva1OjQAOCvPPvusoqOjJUnPPPOM6tWrp1GjRunIkSN6++23LY4OAAAAdUWlVtp+8MEHuuuuuzRkyBD3E3RXrlypPn366N1339Utt9xSpUECAGqOXeknNH7+Zv2anClJ6t4iXM8N6aTG4f7WBgYAVaBbt27unyMiIrR48WILowEAAEBdVamk7TPPPKMZM2Zo7Nix7rYHH3xQL774oqZNm0bSFgDOU8dzi3TtayuVV+RSoI+n/jawrW66KJZSCADOG71799bChQsVGhpqdSgAAACowypVHmHPnj0aNGhQmfZrr71We/fuPeugAAA1U70Ab919WTNd2aaBloztoZsvbkzCFsB55YcfflBRUZHVYQAAAKCOq9RK29jYWC1btkwtW7Ys1b506VLFxsZWSWAAAOsVOQ299v0uDegQpbbRwZKkhxNaycNuI1kL4LzF+xsAAACsVqmk7SOPPKIHH3xQGzduVPfu3SUV17R999139corr1RpgAAAa2w6kKnH5m/WjrQT+v63dC0afZk87DZ5elT6GZYAUCtcf/318vb2Lnfbd999V83RAAAAoC6qVNJ21KhRioqK0j/+8Q99+umnkqS2bdvqk08+0eDBg6s0QABA9SpwuPRS4u9658c9MkwpPMBb/9ezhewsPANQR8THxyswMNDqMAAAAFCHVSppKxWvQLj++uurMhYAgMXW7jumx+Zv1t6MXEnS4M4xmjyovcICyl9xBgDnG5vNpvHjxysiIsLqUAAAAFCHVTppCwA4v6zanaFb/7lGpilFBPnomes7qm+7SKvDAoBqZZqm1SEAAAAAlUva1qtX75QPaDh27FilAwIAWOOSZuG6MDZUrSKC9MTAtgrx87I6JACodpMnT6Y0AgAAACxXqaTtyy+/LKl4JcKoUaM0depUvkIGADWUYZjadzRXJwqcCvL1VNPwANntNmUXOPT28j0a07ulfL085GG3ae59l8rXy8PqkAHAMpMnT5YkHTlyRDt27JAktWnTRg0aNLAyLAAAANQxlUrajhgxwv3zAw88oKFDh6p58+ZVFhQAoGokpWRpwYaD2pWeo0KHIR8vu1pGBKpxmL/eXrFHh7MKVOQy9MTVbSWJhC2AOi8vL09jxozRv//9b7lcLkmSh4eH7rjjDr366qvy9/e3OEIAAADUBXarAwAAnBtJKVmauWynthzMUqift5rWD5C/l6f+s/GQpnyxTYezCtQ03F994vimBACUGDt2rJYvX67//Oc/yszMVGZmpj7//HMtX75cjzzySKXHfe6552Sz2fTwww+72woKCjR69GiFh4crMDBQQ4cOVVpaWqn9kpOTNXDgQPn7+ysiIkLjx4+X0+msdBwAAACoHarkQWSnqm8LAKh+hmFqwYaDOpZbpJYRgbLZbDpwLE/r9h9TgcOQJHWICdYnI+MV4MszKQGgxIIFCzR//nxdeeWV7rarr75afn5+Gj58uN54440zHnPt2rV666231KlTp1LtY8eO1VdffaV58+YpJCREY8aM0ZAhQ7Ry5UpJksvl0sCBAxUVFaVVq1bp8OHDuuOOO+Tl5aVnn332rM4TAAAANVulPqkPGTLE/XNBQYH+7//+TwEBAe62hQsXnn1kAIBK23c0V7vScxQd4iebzabth7O18UCmJCnY11OdGobKZpfSThSouS8P3AGAEnl5eYqMjCzTHhERoby8vDMeLycnR7feeqveeecdPf300+72rKwszZ49W3PnzlXv3r0lSXPmzFHbtm31888/69JLL9WSJUu0bds2LV26VJGRkercubOmTZumxx9/XE899ZS8vb0rf6IAAACo0SqVtA0JCXH/fNttt1VZMACAqnGiwKlChyG/kOIatY3D/LXtULZaRQaqfUzxe/i+jOKHkwEA/ic+Pl6TJ0/W+++/L19fX0lSfn6+pkyZovj4+DMeb/To0Ro4cKASEhJKJW3Xr18vh8OhhIQEd1tcXJwaN26s1atX69JLL9Xq1avVsWPHUknk/v37a9SoUdq6dasuvPDCMscrLCxUYWGh+3V2drYkyeFwyOFwnHH8NYVhGPLz85Ovp03eHma5fXzsZqm/K8rmaZOfn58Mw6jVc2Slknlj/qofc28d5t46zL11mPuqUdH5q1TSds6cOZXZDQBQTQocLh3NLVRksK8CfT0V4OOpQRfEyNuzuJR5ToFTPl52BVEaAQBKefnll3XVVVepUaNGuuCCCyRJmzZtkq+vr7799tszGuvjjz/Whg0btHbt2jLbUlNT5e3trdDQ0FLtkZGRSk1Ndff586rfktclff5s+vTpmjJlSpn2JUuW1PqHqH300Uf//cl1yn7TuhlnOHITadBHSklJUUpKSqViQ7HExESrQ6izmHvrMPfWYe6tw9yfnYp+e6tSn9Zff/113X333e7VBwCAmsE0Tc1bd1DTvtqmEwVO2W02dW1STzabzZ2wNU1Th7Py1alRqJqGB5xmRACoWzp27KidO3fqww8/1G+//SZJuvnmm3XrrbfKz8+vwuMcOHBADz30kBITE6v1nnnChAkaN26c+3V2drZiY2PVr18/BQcHV1scVW3Tpk3q0aOHIm95Tt6Rzcvt42M3Na2boUnr7Co0Kv7MjaK0PUqb+1etWLHCnajHmXE4HEpMTFTfvn3l5eVldTh1CnNvHebeOsy9dZj7qlHyTajTqVTS9oEHHtCwYcNI2gJADXLweJ4mLNyiH3dmSJJaRQSqQaCPu7atn7eH8otcOpyVr7AAbw3p0lB2Ow+SBIA/WrFihbp376777rvvrMZZv3690tPT1aVLF3eby+XSihUr9Nprr+nbb79VUVGRMjMzS622TUtLU1RUlCQpKipKv/zyS6lx09LS3NvK4+PjIx8fnzLtXl5etfrDld1uV35+vgqcpkzXqf/bVWjYVHiaPqX6O03l5+fLbrfX6jmqCWr7dVabMffWYe6tw9xbh7k/OxWdO3tlBjfNM6sTBQA4dwzD1L9X71P/l1box50Z8vG064mr4/TNQ1foiYFt1bFRiDLzi7QvI1eZ+UXq1ChUD/ZppQ4NQ04/OADUMb169dKxY8fOepw+ffpoy5Yt2rhxo/tPt27ddOutt7p/9vLy0rJly9z77NixQ8nJye7aufHx8dqyZYvS09PdfRITExUcHKx27dqddYwAAACouShmCAC13AMf/aqvthyWJF3UtJ6eH9pJzRsESpI6NAxRu+hg7Tta/NCxIF9PNQ0PYIUtAJxEVS1OCAoKUocOHUq1BQQEKDw83N1+zz33aNy4cQoLC1NwcLAeeOABxcfH69JLL5Uk9evXT+3atdPtt9+uGTNmKDU1VRMnTtTo0aPLXU0LAACA80elk7bffvutQkLKX6V17bXXVjogAKgrDMOskmTqNZ2i9f2OdD1+VZxuv7RJmTHsdps7iQsAOL3Vq1erXr165W7r0aNHlR3npZdekt1u19ChQ1VYWKj+/fvr9ddfd2/38PDQl19+qVGjRik+Pl4BAQEaMWKEpk6dWmUxAAAAoGaqdNJ2xIgR5bbbbDa5XKd+oisA1HVJKVlasP6gtqRkKc/hkr+Xhzo2DNHQro1OW7ZgZ9oJHc4qUI/WDSRJAzpGq1vTMDUIYtUVAFSF66+/vtz2s73P/eGHH0q99vX11axZszRr1qyT7tOkSRN9/fXXlT4mAAAAaqdKJW0Nw6jqOACgzkhKydLTX23TniO5MkxTMiXZpORjedqemq2JA9uVm7h1uAy9vWKPXlm6U4G+nkoc20PhgcWJWhK2AFB1UlNTFRERYXUYAAAAqMMq9SAyAEDlGIapt1fs0W+HT8hlGPL2sMvHyy5vD7tchqHfDp/QOyv2yDBK11TceihL181aqRe+3aEil6HOsaFyGTwUEgCqms1GzW8AAABYr9LlEZYvX66///3v2r59uySpXbt2Gj9+vK644ooqCw4Azjd7MnK0bv8xuUxThkvKLSySYUp2m+TtaZcpae3+Y9qTkaOWEUEqdLo067tdev2H3XIapkL9vfTUoPYa3DmGxAIAnANV9SAyAAAA4GxUaqXtBx98oISEBPn7++vBBx/Ugw8+KD8/P/Xp00dz586t6hgB4Lzxe9oJZeU5VOhwKafQqSKXIYfLUJHLUE6hU4UOl7LyHPo97YTyi1y69tWVmvndLjkNUwM6RClxbE9dd2FDErYAcI4YhkFpBAAAAFiuUittn3nmGc2YMUNjx451tz344IN68cUXNW3aNN1yyy1VFiAAnE8MUyp0GnL+qbRBycKuIpcpwzRkmJKft4e6Nq2no7mFmjq4g67uGG1BxABQt0yfPl2RkZG6++67S7X/61//0pEjR/T4449bFBkAAADqkkqttN2zZ48GDRpUpv3aa6/V3r17zzooAKhtDMPUniM52nQgU3uO5JSpSVvCz8teJmH7Z07DlJ9X8dvzhAFxShzbk4QtAFSTt956S3FxcWXa27dvrzfffNOCiAAAAFAXVWqlbWxsrJYtW6aWLVuWal+6dKliY2OrJDAAqC2SUrK0YMNB7UrPUaHDkI+XXS0jAjW0SyN1aBhSqm9OobNCY54ocEiSgny9qjxeAMDJpaamKjq67C/KGjRooMOHD1sQEQAAAOqiSiVtH3nkET344IPauHGjunfvLklauXKl3n33Xb3yyitVGiAA1GRJKVmauWynMk4UyMvTQ6ZMFTkNbUo+rpTj+XqwT6tSidtth7IrNO6WlCxddyG/BAOA6hYbG6uVK1eqWbNmpdpXrlypmJgYi6ICAABAXVOppO2oUaMUFRWlf/zjH/r0008lSW3bttUnn3yiwYMHV2mAAFBTGYapBRsOamf6CR3PLdKJQqdMU7LZpCAfTx3Pd2jhhhS1iw6W3V784LDNB7MqNPa2QyfOZegAgJO477779PDDD8vhcKh3796SpGXLlumxxx7TI488YnF0AAAAqCsqlbSVpOuvv17XX399VcYCALXKvqO5WrUrQ8nH8uQy/rDBlDLznTpR6NTKXUe072hjNW8QKEnKOFFYobEr2g8AULXGjx+vo0eP6v7771dRUZEkydfXV48//rgmTJhgcXQAAACoKyqdtJWkdevWafv27ZKkdu3aqWvXrlUSFADUBsfzirT7SE7phO0fuAxp95EcHc8rcrfV86/Y225F+wEAqpbNZtPzzz+vSZMmafv27fLz81OrVq3k4+NjdWgAAACoQyqVFTh48KBuvvlmrVy5UqGhoZKkzMxMde/eXR9//LEaNWpUlTECQI302+FsOU+SsC3hNIr7dW0SpuW/H1G3pvX0y/7Tl0i4uHlYFUUJAKiMwMBAXXTRRVaHAQAAgDqqUknbe++9Vw6HQ9u3b1ebNm0kSTt27NBdd92le++9V4sXL67SIAGgJtqw/3iF+q3alaGVu4/q6y2p6tWmfoX2CfbzPpvQAABnYd26dfr000+VnJzsLpFQYuHChRZFBQAAgLrEXpmdli9frjfeeMOdsJWkNm3a6NVXX9WKFSuqLDgAqMky84pO30nSt9vS9fWWVHnYbcoucFZon3Rq2gKAJT7++GN1795d27dv12effSaHw6GtW7fqu+++U0hIiNXhAQAAoI6oVNI2NjZWDoejTLvL5VJMTMxZBwUAVjIMU3uO5GjTgUztOZIjwzDL7dcgqGL1DZ2GqXbRwfp89GUa2DG6QvvEhPhVOF4AQNV59tln9dJLL+mLL76Qt7e3XnnlFf32228aPny4GjdubHV4AAAAqCMqVR7hhRde0AMPPKBZs2apW7dukoq/RvbQQw/p73//e5UGCADVKSklSwvWH9SWlCzlOVzy9/JQx4YhGtq1kTo0LL3CKsTfq0JjXtQ0VHPvi5eXh13enjb5etpV6DRks0kyJVOSTcX/Y5qSj6ddl7eqWBkFAEDV2r17twYOHChJ8vb2Vm5urmw2m8aOHavevXtrypQpFkcIAACAuqBSSds777xTeXl5uuSSS+TpWTyE0+mUp6en7r77bt19993uvseOHauaSAHgHEtKydLTX23T7vQcOQ1ThmHKbrdp/9FcbU/N1sSB7Uolbp2uio3bqWGovDyKv9jQskGQLmkWpp92ZcgwJZvtvwlbFSds7TbpkmbhatkgqIrPDgBQEfXq1dOJEyckSQ0bNlRSUpI6duyozMxM5eXlWRwdAAAA6opKJW1feukl2Wy203cEgFrCMEy9vWKPklKy5XA6Zep/73E2mUpKydY7K/bopRs7y24v3hYdUrHyCH/sZ7fbNP6qOB1bsFk700/IaZgy/5u89fSwqVVEkMZf1cZ9DABA9erRo4cSExPVsWNH3XDDDXrooYf03XffKTExUX369LE6PAAAANQRZ5S0zc7OliQNGTLklP2Cg4MrHxEAWGBPRo5W785QQZFThimZMt1lC2ySXIZTq3ZnaE9GjlpGFK+C9fWu2Fvon/t1aBii6UM7ad66ZK3bn6m8Qpf8fTx0UZN6GtYttkwZBgBA9XnttddUUFAgSfrb3/4mLy8vrVq1SkOHDtXEiRMtjg4AAAB1xRklbUNDQyu0wtblquB3hgGghvgtNVvH84rk/NMzx8z//jFM6XhekX5LzVbLiCC5DFOJ29IqNPbxvLIPbuzQMETtojto39FcnShwKsjXU03DA1hhCwAWKVmc4OnpqcDAQPfr+++/X/fff7+VoQEAAKAOOuPyCPPnz1dYWNi5iAUALJOeXSinceo+TqO43+9pJzR+/mZtOpBZobELHc5y2+12m5o3CDzDSAEA5wKLEwAAAFCTnHHS9rLLLlNERMS5iAUALFNkVOxD+E+7M/TcNztU5DLk52VXgcNwl1H44yLdktd2Sa2jKBkDADXd999/X+q1aZq6+uqr9c9//lMNGza0KCoAAADUVZV6EBkA1BaGYVaoBEHSgawKjXckq0BFLkN94iJ03xXN9H8frFdWfvFKWo8/DGuYxYnbYD9PtYsmaQug5qjo+2Jd07NnzzJtHh4euvTSS9W8eXMLIgIAAEBdRtIWwHkrKSVL89cf0JaUbOUXueTn7aGODYM1rGvZh31lFZStO1ueEH9PvXvXRerZuoFMU+rROkKJ29JU6HTJNOVedWu3Sd6eHurZOkLN61MCAUDNkJSSpQUbDmpXeo4KHYZ8vOxqGRGooV0a8RBEAAAAoAY5o6StzWarUK0vALBaUkqWnv5ym3YczlSe05TLMOVht2lPWpZ+O3xCE69pVypBERvqV6FxG9fz15VtikvE2GzSyB7NlX6iQLvTc1ToNOQyTXnYbPLxtKtFRKDu69GcFWwAaoSklCzNXLZTx3KLFB3iJ78QD+UXubTlYJZSjufrwT6tSNz+wYEDB5SXl6fw8HCrQwEAAEAddEZJW9M0deedd8rHx+eU/RYuXHhWQQHA2TAMU2+v2K11+4+VeriY0zBV6DS1bv8xvbNij166sbM7odqleajmrk057dhdmoeWet2hYYgmDmynBesPaktKlvIcLvl7eahjwxAN7crKNQA1g2GYWrDhoI7lFqllRKD7l/CBvp5q6ROoXek5WrghRe2ig+vsL5pmzpzp/jkjI0MfffSRevfurZAQ3scBAABQ/c4oaTtixIhzFQcAVJk9GTn6dmtqqYTtHzkNafHWw3ogo6VaRgRJkj5edaBCY3+86oCGXdi0VFuHhiFqFx1MjUgANda+o7nalZ6j6BC/Mt+astlsig7x0870E9p3NFfNG9TNki4vvfSSpOL5qF+/vgYNGqSJEydaHBUAAADqqjNK2s6ZM+dcxQEAVWbjweMqdJqn7FPoNLXx4HF30nbP0dwKjb33WF657Xa7rc4mOgDUfCcKnCp0GPIL8Sh3u5+3h9KyDZ0ocFZzZDXH3r17rQ4BAAAAcLNbHQAAVJRhmNpzJEebDmRqz5EcGUb5idkvfj19mYM/9/P3qtjvsPw8y094AEBNFuTrKR8vu/KLXOVuzy9yycfLriBfnlELAAAA1ATcmQOoFZJSsjR//QFtSclWfpFLft4e6tgwWMO6xpapG5t8ktWwf7Y3I1fHc4tUL8Bb7RoG6WBW4Wn3adcwqFLxA4CVmoYHqGVEoLYczFJLn8BSJRJM09ThrHx1ahSqpuEBFkYJAAAAoAQrbQHUeEkpWXr6y236ekuq9hzJ0aHMfO05kqOvt6Tq6S+3KSklq1R/06xYLdmUzEJN+jxJktQwxL9C+1S0HwDUJHa7TUO7NFJYgLd2pecop8Apl2Eqp8CpXek5Cgvw1pAuDanFDQAAANQQJG0B1GiGYertFbuVdChb2QUO5RW5lF/kVF6RS9kFDiUdytY7K/aUKpVguCpWk9FlStsOZSsrz6HYsIolYyvaDwBqmg4NQ/Rgn1bq2ChEmflF2peRq8z8InVqFKoH+7Qq860FAAAAANahPAKAGm1PRo5W7z6qQodLdpvk4WGXVLwSzOUyVOhyadXuDO3JyHE/VOxorqNCY3vZpK8fukK+Xh66olWEfD1/V4HTOGl/X0+7rmgVcdbnBABW6dAwRO2ig7XvaK5OFDgV5OuppuEBrLAFAAAAahiStgBqtN9STygr3yHJlCGbHA6XTFOy2SQPu02Sqax8p35LPeFO2hZV9OHnpuTrVfxgsRYRgbqkeZh+2pkhl1mSFnZ3k4dNuqR5uFpEBFbh2QFA9bPbbWregPcyAAAAoCajPAKAGu1IdoGchimHITlcpgyzOIlqmMWvHYbkNAwdyS5w73PytbKl/bGf3W7T+P5xahcTLF9PmzzsxYlbD7vk62lTu5hgje/fhtVoAAAAAADgnGOlLYAarX6Qj/5QrrZchlncr4TdVlyv9nT+nH/t0DBE04d00vx1B7Ru/3HlFTnl7+2pbk3DNKxrI+o9AgAAAACAakHSFkCN5uNVsZWtJf1M05RZgYStVP5XDaj3CAAAAAAArEbSFkCNtnbP8Qr3i4sM1V8XblZFS9o2CCr/LZB6jwAAAAAAwEokbQHUaEu3HqpQvwXrD+iDNQeU73DJZlOFVtvWD/E7y+gAAAAAAACqHg8iA2AJp9PQyp1HJEkrdx6R01n+48MOHC+q0HjH8l3Kd7h0SbMwXdkqvEL7NAr1r1iwAFDHGYapPUdytOlApvYcyZFxumLjAAAAAM4KSVsA1e7zjSka9NqPenT+ZknSo/M3a9BrP+rzjSll+rrOYNxp13XQR/ddqoR2URXqH9+i/hmMDgB1U1JKlqZ9tU2T/7NVz3y1XZP/s1XTvtqmpJQsq0Or0d544w116tRJwcHBCg4OVnx8vL755hv39oKCAo0ePVrh4eEKDAzU0KFDlZaWVmqM5ORkDRw4UP7+/oqIiND48ePldFa0CBAAAABqM5K2AKrV5xtTNPnzJO1My1G+ozglm+9waWdajiZ/nlQmcetRwXE9JN1+aRPZ7Tbd0CVWAd6n3jPA20M3dImtxBkAQN2RlJKlmct2asvBLIX6eatp/QCF+nlry8HidhK3J9eoUSM999xzWr9+vdatW6fevXtr8ODB2rp1qyRp7Nix+uKLLzRv3jwtX75chw4d0pAhQ9z7u1wuDRw4UEVFRVq1apXee+89vfvuu3ryySetOiUAAABUI5K2AKqN02no5aW/K7vAKacpOf/79VqnYcppStkFTr2ydGepUgleFcza/jFH6+3toft7tZCHrfy+Hjbp/l4t5H2axC4A1GWGYWrBhoM6lluklhGBCvT1lIfdpkBfT7WMCNSx3CIt3JBCqYSTGDRokK6++mq1atVKrVu31jPPPKPAwED9/PPPysrK0uzZs/Xiiy+qd+/e6tq1q+bMmaNVq1bp559/liQtWbJE27Zt0wcffKDOnTtrwIABmjZtmmbNmqWiooqVDgIAAEDtRdIWQLX5cdcRHTiap5N9vjdMKflorn7cdcTdVlDB+gh/7je6VyuN69daob4e7jc6u6RQPw+N69dao3u1OuP4AaAu2Xc0V7vScxQd4iebrfRvwWw2m6JD/LQz/YT2Hc21KMLaw+Vy6eOPP1Zubq7i4+O1fv16ORwOJSQkuPvExcWpcePGWr16tSRp9erV6tixoyIjI919+vfvr+zsbPdqXQAAAJy/PK0OAEDdkZSSJedpFmQ5zeJ+veKKP6RWdP1Wef1G92ql+y5rrs82piglM18NQ/10feeGrLAFgAo4UeBUocOQX0j575l+3h5KyzZ0ooAaqyezZcsWxcfHq6CgQIGBgfrss8/Url07bdy4Ud7e3goNDS3VPzIyUqmpqZKk1NTUUgnbku0l206msLBQhYWF7tfZ2dmSJIfDIYfDURWnZQnDMOTn5ydfT5u8Pcq/O/Cxm6X+riibp01+fn4yDKNWz5GVSuaN+at+zL11mHvrMPfWYe6rRkXnj6QtgGpz4GjOGfVbv//4WR/T29tDN17c+KzHAYC6JsjXUz5eduUXuRToW/aWMb/IJR8vu4LK2YZibdq00caNG5WVlaX58+drxIgRWr58+Tk95vTp0zVlypQy7UuWLJG/v/85Pfa59tFHH/33p1N/DWdaN+OU28tqIg36SCkpKUpJKftQVFRcYmKi1SHUWcy9dZh76zD31mHuz05eXl6F+nGXDaDabDqYeUb9Vu/OqPDYXhR7AYAq1TQ8QC0jArXlYJZa+gSWKpFgmqYOZ+WrU6NQNQ0PsDDKms3b21stW7aUJHXt2lVr167VK6+8ohtvvFFFRUXKzMwstdo2LS1NUVFRkqSoqCj98ssvpcZLS0tzbzuZCRMmaNy4ce7X2dnZio2NVb9+/RQcHFxVp1btNm3apB49eijylufkHdm83D4+dlPTuhmatM6uQuMkhe3LUZS2R2lz/6oVK1boggsuqKqQ6xSHw6HExET17dtXXl5eVodTpzD31mHurcPcW4e5rxol34Q6nRqftG3atKn2799fpv3+++/XrFmzVFBQoEceeUQff/yxCgsL1b9/f73++utlvk4GwHoHj+WfUb+/9GyhV5b+LkcFFsz4UPEAAKqU3W7T0C6NlHI8313b1s/bQ/lFLh3OyldYgLeGdGkou73iybG6zjAMFRYWqmvXrvLy8tKyZcs0dOhQSdKOHTuUnJys+Ph4SVJ8fLyeeeYZpaenKyIiQlLxqpbg4GC1a9fupMfw8fGRj49PmXYvL69a/eHKbrcrPz9fBU5TpuvU11yhYVPhafqU6u80lZ+fL7vdXqvnqCao7ddZbcbcW4e5tw5zbx3m/uxUdO5qfNJ27dq1crn+9xWopKQk9e3bVzfccIMkaezYsfrqq680b948hYSEaMyYMRoyZIhWrlxpVchAnWMYpvYdzdWJAqeCfD3VNDyg3A/xuRUse1jSz8vDrt6tw/Xtb0dPu89lLcLPJGQAQAV0aBiiB/u00oINB7UrPUdp2YZ8vOzq1ChUQ7o0VIeGIVaHWGNNmDBBAwYMUOPGjXXixAnNnTtXP/zwg7799luFhITonnvu0bhx4xQWFqbg4GA98MADio+P16WXXipJ6tevn9q1a6fbb79dM2bMUGpqqiZOnKjRo0eXm5QFAADA+aXGJ20bNGhQ6vVzzz2nFi1aqGfPnsrKytLs2bM1d+5c9e7dW5I0Z84ctW3bVj///LP7phfAuZOUkqV5a5O1as9R5Ra6FODjoe7Nw3XDRY0r/GHeOMXzQh7o10bf/rbqtGM80K9NRUMGAJyBDg1D1C46uEK/nMP/pKen64477tDhw4cVEhKiTp066dtvv1Xfvn0lSS+99JLsdruGDh1a6ttiJTw8PPTll19q1KhRio+PV0BAgEaMGKGpU6dadUoAAACoRjU+aftHRUVF+uCDDzRu3DjZbDatX79eDodDCQkJ7j5xcXFq3LixVq9eTdIWOMeSUrL08Me/am9Grlx/SLzuOZKrlbuP6uWbLjxl4tY0pQ932eU4Rd25dlGhahrmp32nKK3QNMxP7aJCK3MKAIAKsNttat4g0OowapXZs2efcruvr69mzZqlWbNmnbRPkyZN9PXXX1d1aAAAAKgFalXSdtGiRcrMzNSdd94pSUpNTZW3t3epBzhIUmRkpFJTU086TmFhoQoLC92vSwoAOxwOORyOKo8bZZXMM/NdexmGqWmfb9aBYznytJd9MzlwLEfTPt+s9++5xL0ay8fjf5ldlyk5DOmXI3ZJpjxskodNKnnOzR+vjZk3XaDRH6xXem5RmTgiArw186YL5HI55Tr1w6RRQ/F+AInrAP/DtQAAAADUsqTt7NmzNWDAAMXExJzVONOnT9eUKVPKtC9ZskT+/v5nNTbOTGJiotUh4Czc0rD4z8kd0+LF37hfzbi4+G/TlF7bZteubLsifE3d0tKlZkGl9/zzyqJH25/sGPna++tP2vvrmUaPmob3A0hcB/if77//3uoQAAAAAMvUmqTt/v37tXTpUi1cuNDdFhUVpaKiImVmZpZabZuWlqaoqKiTjjVhwgSNGzfO/To7O1uxsbHq16+fgoODz0n8KM3hcCgxMVF9+/bliYO11ItLduhfq/adtt/d3ZtqXL82Mk1THZ5a4l5Ja5qSt93UYxe4NHWDXYV/KpGQ9FT/MmMZhqnkY3nKKXQq0MdTjcP8qal4HuD9ABLXAf6n5Fro1auX1aEAAAAAlqk1Sds5c+YoIiJCAwcOdLd17dpVXl5eWrZsmYYOHSpJ2rFjh5KTkxUfH3/SsXx8fMp96q6XlxcfFKsZc17zGIZZoYfN/LznqApdp0+Y/rznqI7nuzTp8yQV/Skx6+NhyssuFRq2UmPZpJNeF62ivc/shFBr8H4AiesA/8N1AAAAgLqsViRtDcPQnDlzNGLECHl6/i/kkJAQ3XPPPRo3bpzCwsIUHBysBx54QPHx8TyEDKiEpJQsLdhwULvSc1ToMOTjZVfLiEAN7dKozAPFNqfkVGjMjSk56vvSCmXlV7w2ob8Xq2cBAAAAAEDdVSuStkuXLlVycrLuvvvuMtteeukl2e12DR06VIWFherfv79ef/11C6IEareklCzNXLZTx3KLFB3iJ78QD+UXubTlYJZSjufrwT6tSiVuK/rML1NSVr5DHRoGK/V4njLynKfdJ9i3Vrw1AQAAAAAAnBN2qwOoiH79+sk0TbVu3brMNl9fX82aNUvHjh1Tbm6uFi5ceMp6tgDKMgxTCzYc1LHcIrVsECBTprLyimTKVMsGATqWW6SFG1JkGGalxn/sqjZadP9ligmtWGmDBkF8JRYAAAAAANRdLGcDzmMVrU+772iudqXnyN/bQ+uTjysjp0gOlyEvD7vqB3orOsRPO9NPaN/RXDVvEHjGcdx/ZUtJUuPwIG0+lHfa/o3Dg874GAAAAAAAAOcLkrbAeepM6tOeKHDqWG6RUo7n6USBUy73glqXsvMdOnKiUA3r+etEwelLG5xK95b19eWWtAr1AwAAAAAAqKtqRXkEAGempD7tloNZCvXzVtP6AQr189aWg8XtSSlZpfoH+Hgo5XieMvP/mLAt5jKlzHynUo7nKcDH46ziGnZhrAK8Tz1GgLeHhl0Ye1bHAQAAAAAAqM1I2gLnmVL1aSMCFejrKQ+7TYG+nmoZEVhufVqnYSgr/9SraLPynXIaxlnF5u3toft7tZBnOSUaJMnTbtP9vVrI+zSJXQAAAAAAgPMZSVvgPFNSnzY6xE82W+nkqM1mK1WftsTy34/odI8YM//bT5J+S82udHyje7XS2L6tVM/fSyW5W7tNCvP30ti+rTS6V6tKjw0AAAAAAHA+oKYtcJ45UeBUocOQX0j5q1X9vD2Ulm2Uqk/7zYbkCo391fr9yi8yNOv7XWcV4+herXTfZc312YZkKW2znrqmva7v0pgVtgAAAAAAAGKlLXDeCfL1lI+XXflFLpmGocNZ+dp7JEeHs/JlGobyi1zy8bIryPd/v7PZmJpfobE3pxXo5aU75fhz4dtK8Pb20JCujSRJQ7o2ImELAAAAAADwX6y0Bc4zTcMD1DIiUD/sSNfREwXKKzJkmMUlCPy97QoP8lWvNpFqGh5QqfHvuqypujSup/Gf/KqCCpS49SMXC9RahmFq39FcnShwKsjXU03DA2Q/SU1qAAAAAEDVIWkLnGfsdpvqB3or+Wie/rgg1jCl7EJDuUV5Cg/0qnTiZfKg9pKkhev26/udx07b/9LmYZU6DgBrJaVkacGGg9qVnqNChyEfL7taRgRqaJdG6tAwxOrwAAAAAOC8RnkE4DzjdBr6cE2yTlbBwGVKc9cky+mswDLZU7jl0qZV2g9AzZGUkqWZy3Zqy8Eshfp5q2n9AIX6eWvLweL2pJQsq0MEAAAAgPMaSVvgPLN8Z7oOZRacsk9KZoGW70w/q+P0ahOpUL9TL9av5+epXm0iz+o4AKqXYZhasOGgjuUWqWVEoAJ9PeVhtynQ11MtIwJ1LLdICzekyDDOvrY1AAAAAKB8JG2B88xPvx+p0n4n4+lp15TBHeTrWf7biK+nXU8N7iDPk2wHUDPtO5qrXek5ig7xk81WuoyKzWZTdIifdqaf0L6juRZFCAAAAADnP2raAueZHeknqrTfqQzu3FCS9Mb3O7XvWJ6cLlOeHjY1DQvQqF4t3dsB1B4nCpwqdBjyCyn/KYJ+3h5KyzZ0osBZzZEBAAAAQN1B0hY4z6Rn5Vdpv9MZ3LmhBnaI1srdGTpyolANgnx0WYv6rLAFaqkgX0/5eNmVX+RSoG/Z24T8Ipd8vOwKKmcbAAAAAKBqkFUBagnDMLXnSI42HcjUniM5J60nmZxx6nq2JfZlFOj+D9fLMMwKvxGcrJ+np10920RoWLdY9WwTQcIWqMWahgeoZUSgDmflyzRLv8+YpqnDWflqFRGkpuEBFkUIAAAAAOc/lskAtUBSSpbmrUvWuv2Zyit0yd/HQ92ahOqGbo3VoWFIqb5FFRzTKenrLama3+agWtT30c6MwtPu06K+z5kHD6BWsdttGtqlkVKO57tr2/p5eyi/yKXDWfkKC/DWkC4NZbfbTj8YAAAAAKBSSNoCNVxSSpYmLNis39Oy5TAk05RsNmlXWrY27M/U9KGdyiRuK+r+K1vo2gtitDXluHZmHDht/+6tIip1HAC1S4eGIXqwTyst2HBQu9JzlJZtyMfLrk6NQjWkS8NKv+cAAAAAACqGpC1QgxmGqRcW/6akQ9n645eUTVMqdElJh7L1wuIdmnPXRZVa9fbYVXGSpCtbR+q91adP2l7ZOvKMjwGgdurQMETtooO172iuThQ4FeTrqabhAaywBQAAAIBqQNIWqMF2HTmhlbszVH71WsmUtHL3Ee06ckKtI4MrfZwrWjVQw1BfpWSevB5uw1A/XdGqQaWPAaD2sdttat4g0OowAAAAAKDOIWkLWMAwzAqtXlu+I11O49RjOY3ifmeTtPX0tOuxq+I0cVGScgqcpZLENhU/Tf6xq9rwgDEAAHBaycnJysjIqNIxt2/fXqXjAQAA1HQkbYFqlpSSpfnrD2hLSrbyi1zy8/ZQx4bBGtY1tkydyDV7jlVozDV7jum+HmcX1+DODSVJb/6wWweO5clpmPK029Q43F9/6dnCvR0AAOBkkpOT1SaurQry86wOBQAAoFYjaQtUo6SULD395TbtyciVYf5vPeuBY3n67fAJTbymXanE7f5juRUat6TfvoyK9T+ZwZ0bamCHaK3cnaEjJwrVIMhHl7WozwpbAABQIRkZGSrIz1P4NY/IKzy2ysbN37NOWT9+UGXjAQAA1HQkbYFqYhim3l6xW7+lnpCHXfL29JBNxXVpi5wu/ZZ6Qu+s2KOXbuzsLpVwNPvkNWb/KCMrX//8cY/+vmTHWcfp6WlXzzYRZz0OAACou7zCY+UT1bLKxnMcPf0DUwEAAM4nJG2BarInI0fr9h+XYZoyXFJekUOmacpms8nLozhJu3b/Me3JyFHLiCBJ0rGC0xS0/a/jhaae/urMar2xdhYAAAAAAKBmIm8DVJPf03KUledQkctQQZFTLtOQyzDkMotfF7kMZeU59HtaTqXGD/Tx1LPXd5RP2eeZlcub//cDAAAAAADUSKRtgGpimIaKXIYcDkNFhlTkNOX4799FhorbXYYMs2Kra/9sydgeuuWSxooO9a5Q/6iQivUDAAAAAABA9SJpC1STQB9PGYYp10m2u1Rc9zbQp3JVS2JC/SRJgzs3qlD/ivYDAAAAAABA9SJpC1QTfy8PucxT93GZxf3OxqierXS6Cgm2//YDAAAAAABAzUPSFqgmq/YcrdJ+J+Pr66nb4xufss/t8Y3l68tzCAEAAAAAAGoisjZANdmw71iV9juVqYM7SpI+WpMsxx9K5HrZpZsvaezeDgAAAAAAgJqHpC1QTbamZJ1Rv00HMs/qeFMHd9QT/dvqnyv36sDxPMXW89e9lzVjhS0AAAAAAEANR/YGqCZH80/2CLLy+3VqFHLWx/T19dSYPtSuBQAAAAAAqE2oaQvUUDabTQ38K/Z/0YiAs3t4GWAVwzC1LyNXkrQvI1eGcZqn9QEAAAAAUAeQtAXOkmGY2nMkR5sOZGrPkZwqTToN6hRdoX7XdIyqsmMC1SUpJUvTvtqmZ7/ZLkl69pvtmvbVNiVVsJQIANRk06dP10UXXaSgoCBFRETouuuu044dO0r1KSgo0OjRoxUeHq7AwEANHTpUaWlppfokJydr4MCB8vf3V0REhMaPHy+n01mdpwIAAAALUB4BOAtJKVmau3q3vtl2RPlFhvy87RrQroFuiW+hDg3PvrxBTHhQlfYDaoqklCzNXLZTx3KL1CjER5IU4uutLQezlHI8Xw/2aVUl/x8CAKssX75co0eP1kUXXSSn06knnnhC/fr107Zt2xQQECBJGjt2rL766ivNmzdPISEhGjNmjIYMGaKVK1dKklwulwYOHKioqCitWrVKhw8f1h133CEvLy89++yzVp4eAAAAzjGStkAlJaVk6Za3Vyu78H+1aguchuauO6wvt6Rr7sj4s046RYX4ycMuuYyT9/GwF/cDagvDMLVgw0Edyy1Sy4hAedqKV6cH+nqqpY+XdqXnaOGGFLWLDpbdbrM4WgConMWLF5d6/e677yoiIkLr169Xjx49lJWVpdmzZ2vu3Lnq3bu3JGnOnDlq27atfv75Z1166aVasmSJtm3bpqVLlyoyMlKdO3fWtGnT9Pjjj+upp56St7e3FacGAACAakB5BKASDMPUnf9aUyph+0fZhS7d+a81Z10qIS4qSGH+3jpZ3spuk8L8vRUXxUpb1B77juZqV3qOokP8ZPv/9u47PKoq/QP4906fZJJJI500Qg9EirAIAkoggMsPF1xljUsQAQuIyK4FFQuoKLvusgi2dRcs6FpBREExSjWGjpQYAiSGxBRIzyRT7/n9ETMyJMAASSYk38/z5CFz77nnvvfMzXDmnTPnSK43tyRJCDPqkV1SjdxSk4ciJCJqfpWV9VO/BAQEAAD27t0Lm82GpKQkZ5kePXogKioK6enpAID09HT06dMHISEhzjLJycmoqqrCkSNHWjF6IiIiImptHGlLdBl+/KUMZ0y2C5Y5Y7Lhx1/KcE1k4GWfJy7IgCFdgvDtTyWw2R2QAQgBSFL9Jy5qlRLXdQlCXJDhss9B1NqqzXZYbDL0xqYX0NNrlCiuklFt5pyNRNQ+yLKMefPmYejQoUhISAAAFBUVQaPRwM/Pz6VsSEgIioqKnGXOTtg27G/Y1xSLxQKLxeJ8XFVVBQCw2Wyw2S7cd2kOsixDr9dDp5KgUTbfPP92tfKi9WoVwuVfd0kqCXq9HrIst0obtUcN7cb2a31se89h23sO295z2PbNw932Y9KW6CyyLJBbakK12Q4fnQoxgd5Nfj372fVH3arv2fVH8fF916Oo0nxZ8SgUEmYNj0NJtRknT5tgc8iQhYBCkqBWKhDXyRszh8fxK+R0VfHRqaBVK1BndcCga/zfUJ3VAa1aAZ8m9hERXY1mz56Nw4cPY8eOHS1+riVLluCZZ55ptP3rr7+Gl5dXi58fAN5///1ff2v6G0mXZdB1QOp1btW7eOAF5pVqUjQw4X0UFBSgoKDgssKjeps3b/Z0CB0W295z2Paew7b3HLb9lamtrXWrHN8RE/3qcEElPtmXj+MlNbDYZGjVCsQHGzC5f2SjuWlPnK5xq87jJdX4cPcpLP7CvSRvUxIijHjipl74ZG8+DhVUotbmgJdaiT4RRkwe0Dg2orYuJtAb8cEGHMqvRLzWAJz1mYMQAoWVdegb6YeYQG/PBUlE1EzmzJmDDRs2YNu2bYiMjHRuDw0NhdVqRUVFhcto2+LiYoSGhjrL7Nq1y6W+4uJi576mLFiwAPPnz3c+rqqqQufOnTFmzBj4+vo212Wd18GDBzF8+HCE3P4CNCFxzVavKXM7yja9fMF6tQqBxQNlLNyjgEV2/wNta/FJFL/3KLZt24bExMTmCrlDsdls2Lx5M0aPHg21Wu3pcDoUtr3nsO09h23vOWz75tHwTaiLYdKWCK4r2YcZ9dAblaizOs67kn1VnXsjOCrMAg9/8uMVx5cQYUSvMF+3RgETtXUKhYTJ/SNRUF6H4yU1iDRqAW+gxmxHfqUFAd4aTOofwfubiK5qQgjcf//9WLt2LbZs2YLY2FiX/QMGDIBarUZaWhomT54MAMjKykJeXh6GDBkCABgyZAiee+45lJSUIDg4GED9yBZfX1/06tWryfNqtVpotdpG29Vqdau8uVIoFKirq4PZLiAczfc6brY53K7XIkuwXMK5LXaBuro6KBQKvgG9Qq11n1FjbHvPYdt7Dtvec9j2V8bdtmPSljq8c1eyb1gYqX4le0OTK9lfypf9tCoF/jKmG/626SfY3Mj1apqe5hMKhYS4Tpy7ltqHhAgj5o7qik/25SP3dBXgDVSaregb6YdJ/SM4gpyIrnqzZ8/Ge++9h88++ww+Pj7OOWiNRiP0ej2MRiPuuusuzJ8/HwEBAfD19cX999+PIUOG4He/+x0AYMyYMejVqxf+/Oc/Y+nSpSgqKsITTzyB2bNnN5mYJSIiIqL2g0lb6vDOXslelmVkFtXAZLHDW6tCz1CDy0r2l5M03fjA9YjrZMBXP+Zjb/7Fp1XoE8bELHUMDSPITxRX4uiuEjw2rie6hBg5wpaI2oVXX30VADBy5EiX7atWrcK0adMAAP/85z+hUCgwefJkWCwWJCcn45VXXnGWVSqV2LBhA+69914MGTIE3t7eSE1NxaJFi1rrMoiIiIjIQ5i0pQ6vYSX7oxWVOHHa5LLvUEElunTyRqC39rJXsm9I9D41oQ/+79X0i5Z/akKfyzoP0dVIoZAQE+SNowBigjjlBxG1H0KIi5bR6XRYuXIlVq5ced4y0dHR+PLLL5szNCIiIiK6CjBpSx2ej06FU+W1KKw0N7n/xGkTaq2OK17JPqGzP4IMGpypsZ63TCeDBgmd/a/oPEREREREREREdHVTeDoAIk8L9daeN2HboLDSjFDvK5s7TqGQsPrOQfA9T/LXV6fCqjsHcaQhEREREREREVEHx6QtdXgrth275HLnWSusEdU5+deECCPem/k7pFwbiQAvNfRqBQK81Ljj2ki8N/N3XHyJiIiIiIiIiIg4PQLRJ3vy3S73QFIPrPj2OBxu1m3UNU7vJkQYsfgPfXHXcBOqzXb46FSICeRcnkREREREREREVI9JW2q3ZFkgt/TiidHSWvcWGDtjsuPmld8js7DK7Ri6Bhua3K5QSM4FyoiIiIiIiIiIiM7GpC21S4cLKvHJvnwcL6mBxSZDq1YgPtiAyf0jG01BYJfdq9MhgBt7dMLpajNG9QjGB26M0L0pMeJywiciIiIiIiIiog6MSVtqdw4XVGJ5WjbOVJtRXmuHxW6HVqWCyWxDQXkd5o7qetlzx84d1RUzhsXhx/wKfLwvH44LJHyVCiAq0Psyr4KIiIiIiIiIiDoqJm2pXZFlgU/25WN3binKXaY9sCCnFPD3MuHTfV7oFeZ7WXPIalVKaFVKDO0ShO4hPvipsBpN5W0VAHqE+GJol6DLvRQiIiIiIiIiIuqgFJ4OgKg55Zaa8PnBX85J2P6mvNaO9QcLkFtqcm5Tu1n32eVUKgVu6hsGlVKChPo/pLP/VSkljO8bCpWKf2JERERERERERHRpmFGiduV0pRlnaqwXLHOmxorTlWbnY8nNAbdnl5NlgTM1VkT6e8FXr4JKpYBSIUGlUsCoVyHS3wulNTbIsricyyAiIiIiIiIiog6M0yNQu/Le7ly3yw2Or5+6QKcGrBfO8zrLNcgtNeF4SQ16hxvhrVGiqMoMs80BnVqJUF8dTFYHskuqkVtqQlwnw2VcCRERERERERERdVRM2lK7ciC/8pLLeek0qHIja+ul0zh/rzbbYbHJ0BuVkBQSwvz0LmX1GiWKq2RUm5uepoGIiIiIiIiIiOh8OD0CtSunK8wXL/RrubTMYphtDvSP9HPrmLPL+ehU0KoVqLM6mixbZ3VAq1bAR8fPRYiIiIiIiIiI6NIwaUtXBbPZjhVp2Xjk44NYkZYN83lGsNpk9+qrcwB3vbUH/9x8DL/rGujWMWeXiwn0RnywAYWVdRDCdd5aIQQKK+vQNdgHMYHe7gVERERERERERET0Kw4DpDbvyc8O4YPdp2Cx/5Ycffm7bNx2bWcsmtjHpaxaCdjcmJFAAFAqJOjUSoT5erkVx9nlFAoJk/tHoqC8DsdLahBm1EOvUaLO6kBhZR0CvDWY1D8CCoWbq5wRERERERERERH9iiNtqU178rNDeCc9Dxa7gEIClBKgkACLXeCd9Dw8+dkhl/L+3prz1ORKrZDw2eyheHB0N2SfrsbFUqsSgOzT1S7bEiKMmDuqK/pEGlFRZ0XuGRMq6qzoG+mHuaO6IiHCeAlXSkREREREREREVI8jbanNMpvt+GD3qfpRsRIACRACkCRACcAhgA92n8JjyT2h+3XuWH+NhAI36u4WoHYmVfVqFRSK+sSsvYnpFVSK+pG5enXjP5eECCN6hfkit9SEarMdPjoVYgK9OcKWiIiIiIiIiIguG0faUpv15s4c55QIDgE4ZED+9V/HrzMlWOwCb+7McR5zqsrmVt351b/NoTA0PhBalRKABG+NAjqVAhpl/b/eGgUACTqVEkPjm577VqGQENfJgMTOfojrZGDCloiIiIiIiIiIrgiTttRmnSqvbdZy5xPfyQfXRvsDqE8CKxQSNCoJCoXkTBoPjA5AfCefKzoPERERERERERGRO5i0pVYnywK5Z0wAgNwzJsiyaLJcmJ/WrfrOLqdSundL69RK5+8KhYSHxvZAz1AfqJUSbA4ZZpsMm0OGWimhZ6gPHhrbnSNoiYiIiIiIiIioVXBOW2pVhwsq8cm+fOSersLNgcDzGzMR08kXk/tHNlq4y+ildqtOo5caDlngvztyUF5nv/gBALoFe7s8TogwYsnkvvhoTx72/FyBWosDXlolro32xy0DO3NRMSIiN8my4DzfREREREREV4hJW2o1hwsqsTwtG2UmK/x19SNiVZKEQ/mVKCivw9xRXV2So/t/rnSr3q1ZZ7BufxEOnqpwO5bOgYZG2+oXFUtgsoGI6DI1fDB3vKQGFpsMrVqB+GBDkx/MERERERER0fkxaUutQpYFPtmXj/zyWtgcMn4pt+L3AcCRwip46zSotdrx6b4C9ArzdSZJrTaHW3VvzToDAcBHq8JNfUPxv935Fz1mcJcLLypGRESX5uwP5sKMeuiNStRZHef9YI6IiIiIiIjOj3PaUqvILTXhQF4FiqssOFVeh6o6GwCgqs6GU+V1KK6yYH9eOXJLTc5jYoK83KpbABjVIxib54/AjOvj4KNVXrC8j1aJ3uFMHBARNZeGD+bKTFbEBxtg0KmgVEgw6FSIDzagzGTFp/sKzjuHOREREREREbli0pZaRWWdDT+X1aKi1lq/yNevb9xtsoDZJqOi1oq8slpU/prMBYAgg8atuu8YHIk3Uwci1KhDXJABN/QIhk6lwLmTGkgAdCoFbuwRgrggjqYlImouuaUmHC+pQZhRD0lyffWVJAlhRj2yS6pdPpgjIiIiIiKi8+P0CNQqKuusqKqzwXGeQVYOUZ/YrayzOrdJiguPmG0Q28nHmSRQKCTMGt4FJVUWnDhdgzqbAw5ZQKmQ4KVWIq6TATOHx3GeWiKiZlRttsNik6E3Nv26rdcoUVwlo9rs3mKRREREREREHR2TttQqKutssF/ka7F2WbiMtA3wVrtV97nlEiKMeOL3vfDx3lM4VFCFOqsDeo0SfSOMmDyAi+EQETU3H50KWrUCdVYHDLrGXYs6qwNatQI+TewjIiIiIiKixvjuiVrFsaLqSy6XX1bn1jFNlUuIMKJXmC9yS02oNtvho1MhJtCbI2yJiFpATKA34oMNOJRfiXitwWWKBCEECivr0DfSDzGB3h6MkoiIiIiI6OrBpC21ihqLe1+JbSgnhMDXR4vdOsZiczS5XaGQENeJc9cSEbU0hULC5P6RKCivc85tq9coUWd1oLCyDgHeGkzqH8EPzoiIiIiIiNzEhcioVZTWWC9e6Kxy5y5kcyEalXtz3xIRUctJiDBi7qiu6BNpREWdFblnTKios6JvpB/mjurKqWmIiIiIiIguAUfa0hWRZeHWFAS+bs5jqD7r2DsGReORtYcuegy/bktE1DZwahoiIiIiIqLm0eZH2hYUFOCOO+5AYGAg9Ho9+vTpgz179jj3CyHw5JNPIiwsDHq9HklJScjOzvZgxB3H4YJKLP7iKJ5afwTPfZGJp9YfweIvjuJwQWWjsp18tW7VufdUBYSoX7DMx0sNrcr1Fj33bb9WpYBK1eZvYyKiDqNhaprEzn6I62RgwpaIiIiIiOgytOlsV3l5OYYOHQq1Wo2NGzfi6NGjeOmll+Dv7+8ss3TpUixfvhyvvfYaMjIy4O3tjeTkZJjNZg9G3v4dLqjE8rRsHMqvhEohwahTQaWQcCi/fvu5iVsfTdMjbX/NzzrVmO04XWMBAHQLMSDQoIFOrXAmaxuKSwB0agUCDRp0C+G8tURERERERERE1H606ekRXnzxRXTu3BmrVq1ybouNjXX+LoTAsmXL8MQTT2DixIkAgLfffhshISFYt24dpkyZ0uoxdwSyLPDJvnzkl9fC5pCRW2qCXRZQKST4e6lRa7Xj030F6BXm6xxhZRUCEn5LugoB7C+VYJVd604dGo1gHx0AIC7IgIHR/th27AwMGhVkAchChkJSQCEBdlng2ugAxAUxaUtERERERERERO1Hm07arl+/HsnJyfjjH/+IrVu3IiIiAvfddx9mzpwJAMjJyUFRURGSkpKcxxiNRgwePBjp6ennTdpaLBZYLBbn46qqKgCAzWaDzWZrwStqH3LPmHDkVBnKa8yotdqdmVgbgBKbDV4aFQ6fKsWJ4krEBNXPN9sz2AsGjQS7Q4ZDALIAVh+rX0BMgoBGAWhUCvQNM7g8BzOGRqOsug65pbWQhQB+Tf0qJAkxgd64a2gUHA47HI7WbQNqHg3PNf/uOjbeBwTwPqDf8F4gIiIiImrjSduTJ0/i1Vdfxfz58/HYY49h9+7dmDt3LjQaDVJTU1FUVAQACAkJcTkuJCTEua8pS5YswTPPPNNo+9dffw0vL6/mvYh2KiUCQMT59toA1OHorq04etbWZwf8ulcG/vajEqfNAqMjBMZEyKifltaBquw9+PKcKYn/GALA9Sn+VS1y9u9Azv4ruRJqCzZv3uzpEKgN4H1AAO8D+s13333n6RCIiIiIiDymTSdtZVnGwIED8fzzzwMA+vXrh8OHD+O1115DamrqZde7YMECzJ8/3/m4qqoKnTt3xpgxY+Dr63vFcV+tZFkgr6wWNRY7DFoVogK8mlxAZkd2Ceb+7yCEENCrlZBF/VQVkiRBIQF1NgckScLyKYkY1jUYAFBcZcaunDIs3fQTaix2qCDwlz4OrMpSYmexCgatCgvG98T4PmFXFBtdXWw2GzZv3ozRo0dDrVZ7OhzyEN4HBPA+oN803As33HCDp0MhIiIiIvKYNp20DQsLQ69evVy29ezZE5988gkAIDQ0FABQXFyMsLDfkn3FxcW45pprzluvVquFVqtttF2tVnfYN4qHCyrxyb58HC+pgcUmQ6tWID7YgMn9I5EQYXQpa7IDZgcAIcHicECGVD9JrSRBgV+nMJDqy6lUKvxv9yk8/0UmZt8YjwW/T8DrW0/gdGUtIr0dkJT1Cdi7R3TBxGvOO3QXANA1TNNyDUAe1ZH/9ug3vA8I4H1Av+F9QEREREQdWZtO2g4dOhRZWVku244dO4bo6GgA9YuShYaGIi0tzZmkraqqQkZGBu69997WDveqdbigEsvTslFmsiLMqIfeqESd1YFD+ZUoKK/D3FFdXRK3CkkBpSTBKssQAlBIApJUP9rWLgBJEtAoFCgz2XDHfzKw83gpAGBr1mmsmTEYNyWEYcexIlRm78bfbumLYd1CoaqfH4GIiIiIiIiIiKjDa9NJ2wcffBDXXXcdnn/+edx6663YtWsX3njjDbzxxhsAAEmSMG/ePDz77LPo2rUrYmNjsXDhQoSHh+Pmm2/2bPBXCVkW+GRfPspMVsQHGyBJ9VMOGHQqxGsNOF5Sg0/3FaBXmK9zOoL4YO/63x2ASgHIqB9oCwlQSYBdBmQh8NwXmbDYZejUCjyU3APTrouBQiFBoZAwtGsnfJkNDO3aiQlbIiIiIiIiIiKis7TppO21116LtWvXYsGCBVi0aBFiY2OxbNkypKSkOMs8/PDDMJlMmDVrFioqKjBs2DBs2rQJOp3Og5FfPXJLTTheUoMwox4SgGqzDTa7DLVKAR+tCmFGPbJLqpFbakJcJwMAQCFJ8NYoYbPLkCRAe9bcsjaHDACwOgQAgcGxAXhxcl/EBHl74OqIiIiIiIiIiIiuPm1+iOPvf/97HDp0CGazGZmZmZg5c6bLfkmSsGjRIhQVFcFsNuObb75Bt27dPBTt1afabIfFJsNic2BvXjn25JZjX14F9uSWY29eOcw2Byw2GdVmu/MYk8WBYB8d/LzUUCgkyALOH0khQQBQSMA9I+Lw/szfMWFLREREHc62bdswYcIEhIeHQ5IkrFu3zmW/EAJPPvkkwsLCoNfrkZSUhOzsbJcyZWVlSElJga+vL/z8/HDXXXehpqamFa+CiIiIiDylzSdtqWX56FSwOWT8mF+BMpMVGpUCPjoVNCoFykxWHMqvgM0hw0encjkmwKBBzzAjOvvr4aVRQqdWwqBVISbACz1CfTAg2h+3DuzsnFKBiIiIqCMxmUxITEzEypUrm9y/dOlSLF++HK+99hoyMjLg7e2N5ORkmM1mZ5mUlBQcOXIEmzdvxoYNG7Bt2zbMmjWrtS6BiIiIiDyoTU+PQC0vyt8LFrsDNRY7gn21kKT6PL5aKcFXp0JJlQVWu4wofy/nMTGB3ogPNuDgqQpoVUpU1NlwbUwAggxaGLRKnDhtQp8IP8QEcoQtERERdUzjxo3DuHHjmtwnhMCyZcvwxBNPYOLEiQCAt99+GyEhIVi3bh2mTJmCzMxMbNq0Cbt378bAgQMBAC+//DLGjx+Pv//97wgPD2+1ayEiIiKi1sekbQeXV14LrUoJb50aVWY7vDQqKBUSHLJArdUOg04NjUqBvPLa3+a0VUhIjPTDR3vyUWOpnzahvNaKYB8dTpw2IcBbg0n9IzjKloiIiKgJOTk5KCoqQlJSknOb0WjE4MGDkZ6ejilTpiA9PR1+fn7OhC0AJCUlQaFQICMjA3/4wx+arNtiscBisTgfV1VVAQBsNhtsNlsLXdFvZFmGXq+HTiVBoxTNVq9drbxovVqFcPnXXZJKgl6vhyzLrdJG7VFDu7H9Wh/b3nPY9p7Dtvcctn3zcLf9mLTt4KrNdqiVCvSNNOLnUhOq6uxwyAJKhYRAbw2iAr1RWWtzzmlrtjmwPC0br287CYcsoFUpEBPkDV+tChV1VvSN9MOk/hFIiDB6+MqIiIiI2qaioiIAQEhIiMv2kJAQ576ioiIEBwe77FepVAgICHCWacqSJUvwzDPPNNr+9ddfw8vLq4kjmt/777//62+O5qt00HVA6nVu1bt4oHyJlUcDE95HQUEBCgoKLis8qrd582ZPh9Bhse09h23vOWx7z2HbX5na2lq3yjFp207JskBuqQnVZjt8dCrEBHo3OfLVR6eCVq2ATqVE/yh/VJvtsDlkqJX1c9uaLA6Y1Q746FTYl1eOhz46iBOnTQCAm/qG4anf90KNxX7R8xARERFRy1uwYAHmz5/vfFxVVYXOnTtjzJgx8PX1bfHzHzx4EMOHD0fI7S9AExLXbPWaMrejbNPLF6xXqxBYPFDGwj0KWGT3+6PW4pMofu9RbNu2DYmJic0Vcodis9mwefNmjB49Gmq12tPhdChse89h23sO295z2PbNo+GbUBfDpG07dLigEp/szcehgkrU2hzwUivRJ8KIyQMiG42AbZif9lB+JeKDDfDV//ZHJ4RAYWUd+kbWz0/73q48nDhtQpBBi2dvTsDYhFAAgOsYECIiIiK6kNDQ+j5UcXExwsLCnNuLi4txzTXXOMuUlJS4HGe321FWVuY8vilarRZarbbRdrVa3SpvrhQKBerq6mC2CwhH832Qb7Y53K7XIkuwXMK5LXaBuro6KBQKvgG9Qq11n1FjbHvPYdt7Dtvec9j2V8bdtmPStp05XFCJZ784ipOnTZCFAAQACcgrq0VmURWeuKmXS+JWoZAwuX8kCsrrcLykBmFGPfQaJeqsDhRW1sFXp3LOT3v7oChUmW24fVAU/Lw0nrtIIiIioqtYbGwsQkNDkZaW5kzSVlVVISMjA/feey8AYMiQIaioqMDevXsxYMAAAMC3334LWZYxePBgT4VORERERK1E4ekAqPnIssAb207ip8JqOGQZXhoVfPVqeGlUcMgyfiqsxr+3nYQsuy7OkBBhxNxRXdEn0oiKOityz5hwxmRGjcWOI4VViAnyBlCf4L1vZDwTtkREREQXUVNTgwMHDuDAgQMA6hcfO3DgAPLy8iBJEubNm4dnn30W69evx6FDhzB16lSEh4fj5ptvBgD07NkTY8eOxcyZM7Fr1y7s3LkTc+bMwZQpUxAeHu65CyMiIiKiVsGRtu3IyTM12PNzGSQJvyZW678WplZK8PPSoMxkxe6fy3DyTA3ig31cjk2IMKJXmC9yS03Yeuw0Xt1yAiXV9SsPf/tTCf4vkW8OiIiIiNy1Z88e3HDDDc7HDfPMpqamYvXq1Xj44YdhMpkwa9YsVFRUYNiwYdi0aRN0Op3zmDVr1mDOnDkYNWoUFAoFJk+ejOXLl7f6tRARERFR62PSth05VlyNGrMdRr0aDQnb30gwaFWorLPhWHF1o6QtAFSb7Xh1ywl8tDcfANA5QI8XJvXF0Piglg+eiIiIqB0ZOXIkhBDn3S9JEhYtWoRFixadt0xAQADee++9lgiPiIiIiNo4Jm3bFXcXXGhcbvPRYjy+9hBKqi2QJCB1SAweSu4Oby1vESIiIiIiIiIiotbEjFw70i3EAINOBZPFDrVSAUn6LTkrhIDJYodBp0K3EEOjYz/Zm4+Sagvigrzx4i19cW1MQGuGTkRERERERERERL9i0rYdiQsyYGC0P7YdO4Mqsw1eGhWUCgkOWaDWaocsgGujAxAXZIAQAlaHDK1KCQBYdHNvdA0xYPYN8dCplR6+EiIiIiIiIiIioo6LSdt2RKGQMGt4F5RUWXDyjAm1Vvtv+yQJPUINmDk8DmdMFixcdxgalRIv/6kfACDYR4e/jOnuqdCJiIiIiIiIiIjoV0zatjMJEUY88fte+HjvKRwqqEKd1QG9Rom+EUZM6h+BrKJqpLyZgco6G1QKCSeTuiKuU+PpEoiIiIiIiIiIiMgzmLRthxIijOgV5ovcUhOqzXb46FTQKBV44rPD2JJ1+tcyvvjbLYlM2BIREREREREREbUxTNq2UwqFhLhOBsiywPu787Dky59QY7FDo1JgXlJXzLo+DiqlwtNhEhERERERERER0TmYtG3nam0OrPj2OGosdvSP8sPSWxIRH8zRtURERERERERERG0Vk7btkEMWUEiAJEkwaFVYMqkPTp42IfW6GCgVkqfDIyIiIiIiIiIiogvg9+PbmeMlNbj19XR8tCffuW1k92BMHxbLhC0REREREREREdFVgCNt2wm7Q8a/t+fgn98cg9Uuo7CiDjf3i4BGxbw8ERERERERERHR1YRJ23bgp6IqPPTRjzhUUAkAGN6tE5ZM6sOELRERERERERER0VWISdurmNUu45Utx7Hyu+OwOQR8dSo8OaE3JvePgCRxKgQiIiIiIiIiIqKrEZO2V7HMwir8Ky0bQgCje4XguZsTEOyr83RYREREREREREREdAWYtG1lsiyQW2pCtdkOH50KMYHeUFzCAmFCCOco2sTOfpg3qhtiO3ljQt8wjq4lIiIiIiIiIiJqB5i0bUWHCyrxyb58HC+pgcUmQ6tWID7YgMn9I5EQYbzo8Xt/LsPCdUfw8u390KWTAQDwQFLXlg6biIiIiIiIiIiIWhFXqmolhwsqsTwtG4fyK+Gn1yAmyBt+eg0O5ddvP/zrImJNqbXasejzo7jltXQcLazC3zZltWLkRERERERERERE1Jo40rYVyLLAJ/vyUWayIj7Y4JzGwKBTIV5rwPGSGny6rwC9wnwbTZXw/YkzePSTQ8grqwUA3DIgEgtv6tXq10BEREREREREREStg0nbVpBbasLxkhqEGfWN5p2VJAlhRj2yS6qRW2pC3K/THlSbbXhh409Yk5EHAAg36rBkcl+M6Nap1eMnIiIiIiIiIiKi1sOkbSuoNtthscnQG5VN7tdrlCiuklFttju3fbD7lDNhe8fvovDI2B7w0albJV4iIiIiIiIiIiLyHCZtW4GPTgWtWoE6qwMGXeMmr7M6oFUr4HPWvtTrYrAntxyp18VgSJfA1gyXiIiIiIiIiIiIPIgLkbWCmEBvxAcbUFhZByGEyz4hBAor66BRKvDU+iOw2mUAgFqpwGt/HsCELRERERERERERUQfDpG0rUCgkTO4fiQBvDY6X1KDGbIdDFqgx25FZWIX88lqk/VSC7dln8M4PP3s6XCIiIiIiIiIiIvIgTo/QShIijJg7qis+2ZeP4yU1KKp0oMpix89nTDDbZSgVEmYNj0PK4ChPh0pEREREREREREQexKRtK0qIMKJXmC/2/lyGv399DFlF1QCAHqE+WHpLX/SN9PNsgERERERERERERORxTNq2MoVCwhvbc5CRUwa1UsLsG+Jx38h4aFScqYKIiIiIiIiIiIiYtPWIx8f3RGWdDYsm9kaPUF9Ph0NERERERERERERtCJO2HhAT5I0P7x7i6TCIiIiIiIiIiIioDeJ38omIiIiIiIiIiIjaECZtiYiIiIiIiIiIiNoQJm2JiIiIiIiIiIiI2hAmbYmIiIiIiIiIiIjaECZtiYiIiIiIiIiIiNoQJm2JiIiIiIiIiIiI2hAmbYmIiIiIiIiIiIjaECZtiYiIiIiIiIiIiNoQJm2JiIiIiIiIiIiI2hAmbYmIiIiIiIiIiIjaECZtiYiIiIiIiIiIiNoQJm2JiIiIiIiIiIiI2hAmbYmIiIiI2rCVK1ciJiYGOp0OgwcPxq5duzwdEhERERG1MCZtiYiIiIjaqA8++ADz58/HU089hX379iExMRHJyckoKSnxdGhERERE1IJUng6AiIiIiIia9o9//AMzZ87EnXfeCQB47bXX8MUXX+C///0vHn30UQ9H175kZmY2e51BQUGIiopq9nqJiIg6qry8PJw5c6bZ622L/2czaQtACAEAqKqq8nAkHYfNZkNtbS2qqqqgVqs9HQ55CO8DAngfUD3eB9Sg4V6orq4G8Fs/rSOyWq3Yu3cvFixY4NymUCiQlJSE9PT0Jo+xWCywWCzOx5WVlQCAsrIy2Gy2lg0Y9f1pnU4HqTQHQrZc/AA3KaoLL1qvrAJqaztDLjwFYXe/brnwGHR6PWbMmNFM0f5Gq9PjjddfQ3BwcLPXrVAoIMtym6hXlmXU1tZi+/btUCia/jJnS8XbknVfDTGf2/YduS1aq96Guu12+0Xv+8up92psi9aO2Z3XnMut+0pcbfVeTt3utn1LxVxSUoJZd98Di7mu2evW6fXYumULIiIimr3uc7nbz5VER+4J/yo/Px+dO3f2dBhEREREdI5Tp04hMjLS02F4xC+//IKIiAh8//33GDJkiHP7ww8/jK1btyIjI6PRMU8//TSeeeaZ1gyTiIiIiC7Dxfq5HGkLIDw8HKdOnYKPjw8kSfJ0OB1CVVUVOnfujFOnTsHX19fT4ZCH8D4ggPcB1eN9QA0a7oW8vDxIkoTw8HBPh3RVWbBgAebPn+98LMsyysrKEBgY2O77uXwd8Ry2veew7T2Hbe85bHvPYds3DyEEqqurL9rPZdIW9cO2O+oIDk/z9fXlHzrxPiAAvA+oHu8DamA0Gjv8vRAUFASlUoni4mKX7cXFxQgNDW3yGK1WC61W67LNz8+vpUJsk/g64jlse89h23sO295z2Paew7a/ckaj8aJlmmfiFSIiIiIialYajQYDBgxAWlqac5ssy0hLS3OZLoGIiIiI2h+OtCUiIiIiaqPmz5+P1NRUDBw4EIMGDcKyZctgMplw5513ejo0IiIiImpBTNqSR2i1Wjz11FONvr5HHQvvAwJ4H1A93gfUgPeCq9tuuw2nT5/Gk08+iaKiIlxzzTXYtGkTQkJCPB1am8N7x3PY9p7Dtvcctr3nsO09h23fuiQhhPB0EERERERERERERERUj3PaEhEREREREREREbUhTNoSERERERERERERtSFM2hIRERERERERERG1IUzaEhEREREREREREbUhTNpSiyooKMAdd9yBwMBA6PV69OnTB3v27HHuF0LgySefRFhYGPR6PZKSkpCdne3BiKm5xcTEQJKkRj+zZ88GAJjNZsyePRuBgYEwGAyYPHkyiouLPRw1NTeHw4GFCxciNjYWer0eXbp0weLFi3H2Wph8Peg4qqurMW/ePERHR0Ov1+O6667D7t27nft5L7Q/27Ztw4QJExAeHg5JkrBu3TqX/e4852VlZUhJSYGvry/8/Pxw1113oaamphWvgtq6lStXIiYmBjqdDoMHD8auXbs8HVK79/TTTzfq4/Xo0cPTYbVLzfE6SpfnYm0/bdq0Rn8HY8eO9Uyw7ciSJUtw7bXXwsfHB8HBwbj55puRlZXlUobvJVuGO20/cuTIRvf9Pffc46GI2y8mbanFlJeXY+jQoVCr1di4cSOOHj2Kl156Cf7+/s4yS5cuxfLly/Haa68hIyMD3t7eSE5Ohtls9mDk1Jx2796NwsJC58/mzZsBAH/84x8BAA8++CA+//xzfPTRR9i6dSt++eUXTJo0yZMhUwt48cUX8eqrr2LFihXIzMzEiy++iKVLl+Lll192luHrQccxY8YMbN68Ge+88w4OHTqEMWPGICkpCQUFBQB4L7RHJpMJiYmJWLlyZZP73XnOU1JScOTIEWzevBkbNmzAtm3bMGvWrNa6BGrjPvjgA8yfPx9PPfUU9u3bh8TERCQnJ6OkpMTTobV7vXv3dunr7dixw9MhtUvN8TpKl+dibQ8AY8eOdfk7eP/991sxwvZp69atmD17Nn744Qds3rwZNpsNY8aMgclkcpbhe8mW4U7bA8DMmTNd7vulS5d6KOJ2TBC1kEceeUQMGzbsvPtlWRahoaHib3/7m3NbRUWF0Gq14v3332+NEMkDHnjgAdGlSxchy7KoqKgQarVafPTRR879mZmZAoBIT0/3YJTU3G666SYxffp0l22TJk0SKSkpQgi+HnQktbW1QqlUig0bNrhs79+/v3j88cd5L3QAAMTatWudj915zo8ePSoAiN27dzvLbNy4UUiSJAoKClotdmq7Bg0aJGbPnu187HA4RHh4uFiyZIkHo2r/nnrqKZGYmOjpMDqcy3kdpeZxbtsLIURqaqqYOHGiR+LpSEpKSgQAsXXrViGE4HvJVnRu2wshxIgRI8QDDzzguaA6CI60pRazfv16DBw4EH/84x8RHByMfv364d///rdzf05ODoqKipCUlOTcZjQaMXjwYKSnp3siZGphVqsV7777LqZPnw5JkrB3717YbDaXe6BHjx6IioriPdDOXHfddUhLS8OxY8cAAAcPHsSOHTswbtw4AHw96EjsdjscDgd0Op3Ldr1ejx07dvBe6IDcec7T09Ph5+eHgQMHOsskJSVBoVAgIyOj1WOmtsVqtWLv3r0u95BCoUBSUhJfN1pBdnY2wsPDERcXh5SUFOTl5Xk6pA6H/3d63pYtWxAcHIzu3bvj3nvvRWlpqadDancqKysBAAEBAQDA95Kt6Ny2b7BmzRoEBQUhISEBCxYsQG1trSfCa9dUng6A2q+TJ0/i1Vdfxfz58/HYY49h9+7dmDt3LjQaDVJTU1FUVAQACAkJcTkuJCTEuY/al3Xr1qGiogLTpk0DABQVFUGj0cDPz8+lHO+B9ufRRx9FVVUVevToAaVSCYfDgeeeew4pKSkAwNeDDsTHxwdDhgzB4sWL0bNnT4SEhOD9999Heno64uPjeS90QO4850VFRQgODnbZr1KpEBAQwPuCcObMGTgcjibvoZ9++slDUXUMgwcPxurVq9G9e3cUFhbimWeewfXXX4/Dhw/Dx8fH0+F1GPy/07PGjh2LSZMmITY2FidOnMBjjz2GcePGIT09HUql0tPhtQuyLGPevHkYOnQoEhISAPC9ZGtpqu0B4Pbbb0d0dDTCw8Px448/4pFHHkFWVhY+/fRTD0bb/jBpSy1GlmUMHDgQzz//PACgX79+OHz4MF577TWkpqZ6ODryhP/85z8YN24cwsPDPR0KtbIPP/wQa9aswXvvvYfevXvjwIEDmDdvHsLDw/l60AG98847mD59OiIiIqBUKtG/f3/86U9/wt69ez0dGhERXYKGb8wAQN++fTF48GBER0fjww8/xF133eXByIhaz5QpU5y/9+nTB3379kWXLl2wZcsWjBo1yoORtR+zZ8/G4cOHOWe2B5yv7c9eV6BPnz4ICwvDqFGjcOLECXTp0qW1w2y3OD0CtZiwsDD06tXLZVvPnj2dX5kKDQ0FgEarOxYXFzv3Ufvx888/45tvvsGMGTOc20JDQ2G1WlFRUeFSlvdA+/PQQw/h0UcfxZQpU9CnTx/8+c9/xoMPPoglS5YA4OtBR9OlSxds3boVNTU1OHXqFHbt2gWbzYa4uDjeCx2QO895aGhoowWl7HY7ysrKeF8QgoKCoFQq+brRBvj5+aFbt244fvy4p0PpUPh/Z9sSFxeHoKAg/h00kzlz5mDDhg347rvvEBkZ6dzO95It73xt35TBgwcDAO/7ZsakLbWYoUOHIisry2XbsWPHEB0dDQCIjY1FaGgo0tLSnPurqqqQkZGBIUOGtGqs1PJWrVqF4OBg3HTTTc5tAwYMgFqtdrkHsrKykJeXx3ugnamtrYVC4fpfjlKphCzLAPh60FF5e3sjLCwM5eXl+OqrrzBx4kTeCx2QO8/5kCFDUFFR4TIa+9tvv4Usy843CdRxaTQaDBgwwOUekmUZaWlpfN1oZTU1NThx4gTCwsI8HUqHwv8725b8/HyUlpby7+AKCSEwZ84crF27Ft9++y1iY2Nd9vO9ZMu5WNs35cCBAwDA+765eXolNGq/du3aJVQqlXjuuedEdna2WLNmjfDy8hLvvvuus8wLL7wg/Pz8xGeffSZ+/PFHMXHiRBEbGyvq6uo8GDk1N4fDIaKiosQjjzzSaN8999wjoqKixLfffiv27NkjhgwZIoYMGeKBKKklpaamioiICLFhwwaRk5MjPv30UxEUFCQefvhhZxm+HnQcmzZtEhs3bhQnT54UX3/9tUhMTBSDBw8WVqtVCMF7oT2qrq4W+/fvF/v37xcAxD/+8Q+xf/9+8fPPPwsh3HvOx44dK/r16ycyMjLEjh07RNeuXcWf/vQnT10StTH/+9//hFarFatXrxZHjx4Vs2bNEn5+fqKoqMjTobVrf/nLX8SWLVtETk6O2Llzp0hKShJBQUGipKTE06G1O83xOkqX50JtX11dLf7617+K9PR0kZOTI7755hvRv39/0bVrV2E2mz0d+lXt3nvvFUajUWzZskUUFhY6f2pra51l+F6yZVys7Y8fPy4WLVok9uzZI3JycsRnn30m4uLixPDhwz0cefvDpC21qM8//1wkJCQIrVYrevToId544w2X/bIsi4ULF4qQkBCh1WrFqFGjRFZWloeipZby1VdfCQBNPrd1dXXivvvuE/7+/sLLy0v84Q9/EIWFhR6IklpSVVWVeOCBB0RUVJTQ6XQiLi5OPP7448JisTjL8PWg4/jggw9EXFyc0Gg0IjQ0VMyePVtUVFQ49/NeaH++++47AaDRT2pqqhDCvee8tLRU/OlPfxIGg0H4+vqKO++8U1RXV3vgaqitevnll0VUVJTQaDRi0KBB4ocffvB0SO3ebbfdJsLCwoRGoxERERHitttuE8ePH/d0WO1Sc7yO0uW5UNvX1taKMWPGiE6dOgm1Wi2io6PFzJkz+YFRM2iqzQGIVatWOcvwvWTLuFjb5+XlieHDh4uAgACh1WpFfHy8eOihh0RlZaVnA2+HJCGEaIUBvURERERERERERETkBs5pS0RERERERERERNSGMGlLRERERERERERE1IYwaUtERERERERERETUhjBpS0RERERERERERNSGMGlLRERERERERERE1IYwaUtERERERERERETUhjBpS0RERERERERERNSGMGlLRERERERE1I7ZbDZPh0BERJeISVsioqtcVVUVrrnmGtTU1CA/Px/x8fGeDomIiIiIPGjLli245ZZb0KVLFxiNRkRHR0MI0eLnzcrKQmhoKKqrq1v8XHR5YmJisGzZMgCA1WpFTEwM9uzZ49mgiKhJTNoSUauaOnUqJkyY4Okw2hVfX18MGzYMfn5+iImJwb333nvRY0aOHAlJkhr93HPPPa0Q8eU7ffo07r33XkRFRUGr1SI0NBTJycnYuXOnp0MjIiIiAgCcOnUK06dPR3h4ODQaDaKjo/HAAw+gtLS0Vc6/Zs0aTJgwAQMHDsTatWuxd+9e7N+/H5Iktfi5FyxYgPvvvx8+Pj5XXFdubi4kScKBAweuPDBqkkajwV//+lc88sgjng6FiJogidb4uI2IOrQjR45g0aJF2LlzJwoKCgAABoMBw4YNw/z58zF69GgPR9g+lJWVQaVSwdfX96JlR44ciW7dumHRokUu2728vNw63lOGDx8Oq9WKJUuWIC4uDsXFxUhLS0Pv3r3xf//3fy1yTqvVCo1G0yJ1ExERUfty8uRJDBkyBN26dcOzzz6L2NhYHDlyBA899BCsVit++OEHBAQEtNj5a2pq0LlzZ3z44Yet3sfOy8tDfHw8cnJyEBERccX15ebmIjY2Fvv378c111xz3nIdsa92JdccExODefPmYd68eQCA8vJyhIaGYt++fejdu3czRklEV4ojbYmoRa1duxaJiYmwWCx49913ceutt2Ls2LHYuHEjQkNDMWbMGKxcudJZfvfu3Rg9ejSCgoJgNBoxYsQI7Nu3z6VOSZKwbt06AIAQAlOnTkXfvn1RXl6O1atXNzmCVJIkxMTEAACefvppl46f1WpFfHw8JElCRUUFAGDatGm4+eabz3teoH4Uxa233go/Pz8EBARg4sSJyM3NdTnmv//9L3r37g2tVouwsDDMmTPHretwN87Vq1fDz88PABAQEABfX18MHz7crVEJXl5eCA0NdflpSNi+/fbbMBgMyM7Odpa/77770KNHD9TW1gIA3nnnHQwcOBA+Pj4IDQ3F7bffjpKSEmf5LVu2QJIkfPXVV+jXrx/0ej1uvPFGlJSUYOPGjejZsyd8fX1x++23O+u8kIqKCmzfvh0vvvgibrjhBkRHR2PQoEFYsGCBS8K2oqICd999N0JCQqDT6ZCQkIANGzY493/yySfO5yQmJgYvvfSSy3liYmKwePFiTJ06Fb6+vpg1axYAYMeOHbj++uuh1+vRuXNnzJ07FyaT6aJxExERUccxe/ZsaDQafP311xgxYgSioqIwbtw4fPPNNygoKMDjjz/uLHv219QBYNmyZc7+aoOz+4tNjTxduHAhJEly1rN9+3bo9XqsXbsW4eHh8PLyQlJSEo4cOeI85uz+IwD8/PPP6Ny5M5544gnntov185ry4YcfIjEx0ZmwNZlM8PX1xccff+xSbt26dfD29r7oFAqxsbEAgH79+kGSJIwcORLAb/305557DuHh4ejevXujtmrg5+eH1atXOx+703+/VNOnT8fvf/97l202mw3BwcH4z3/+c9HjR44ciTlz5mDOnDkwGo0ICgrCwoULXaazuNz+aUlJCSZMmAC9Xo/Y2FisWbOm0fn9/f0xdOhQ/O9//7vcJiCiFsKkLRG1qHnz5mHkyJFYt24dRo4cCb1eD61Wi2HDhmHVqlWYNm0aHn74YWfnorq6GqmpqdixYwd++OEHdO3aFePHjz9vp27u3Ln4/vvv8fXXX8Pf3x+33XYbCgsLUVhYiGXLliEyMtL5ePfu3U3WsWLFChQXF1/SddlsNiQnJ8PHxwfbt2/Hzp07YTAYMHbsWFitVgDAq6++itmzZ2PWrFk4dOgQ1q9ff975Zs+9jsuN89NPP8X+/fsv6VqaMnXqVIwfPx4pKSmw2+344osv8Oabb2LNmjXw8vICUN8GixcvxsGDB7Fu3Trk5uZi2rRpjep6+umnsWLFCnz//ffOjvKyZcvw3nvv4YsvvsDXX3+Nl19++aIxGQwGGAwGrFu3DhaLpckysixj3Lhx2LlzJ959910cPXoUL7zwApRKJQBg7969uPXWWzFlyhQcOnQITz/9NBYuXOjSmQeAv//970hMTMT+/fuxcOFCnDhxAmPHjsXkyZPx448/4oMPPsCOHTtckvBERETUsZWVleGrr77CfffdB71e77IvNDQUKSkp+OCDD5ptbtn8/HwsW7bM5VynT59GYWEhtmzZgv/973/IyMiAj48Pxo4di7q6ukZ1FBUVISkpCRMnTsSzzz7r3O5uP+9s27dvx8CBA52Pvb29MWXKFKxatcql3KpVq3DLLbdcdAqFXbt2AQC++eYbFBYW4tNPP3XuS0tLQ1ZWFjZv3uzy4fyFuNN/vxwzZszApk2bUFhY6Ny2YcMG1NbW4rbbbnOrjrfeegsqlQq7du3Cv/71L/zjH//Am2++6VLmcvqn06ZNw6lTp/Ddd9/h448/xiuvvNJk8n3QoEHYvn37ZbYAEbUYQUTUQoqKigQA8c9//tO5LTU1VUycONH5+NNPPxUAxA8//NBkHQ6HQ/j4+IjPP//cuQ2AWLt2rXj88cdFRESEyMnJafLYVatWiejo6Ebbn3rqKZGYmCiEEKK0tFT4+/uLxYsXCwCivLxcCCHEPffcI8aMGeNyXMN5hRDinXfeEd27dxeyLDv3WywWodfrxVdffSWEECI8PFw8/vjjTcbmznW4E+eqVauE0WgUQghhtVpFfHy8s8z+/fvPe+4RI0YItVotvL29XX7effddZ5mysjIRGRkp7r33XhESEiKee+6589YnhBC7d+8WAER1dbUQQojvvvtOABDffPONs8ySJUsEAHHixAnntrvvvlskJydfsO4GH3/8sfD39xc6nU5cd911YsGCBeLgwYPO/V999ZVQKBQiKyuryeNvv/12MXr0aJdtDz30kOjVq5fzcXR0tLj55ptdytx1111i1qxZLtu2b98uFAqFqKurcyt2IiIiat9++OEHl/7iuf7xj38IAKK4uFgIUd/nOLuf/M9//rNR3/Xs+nJyclz6eFOnThV33XWXSz2rVq0SAMTOnTuddVRVVQk/Pz/x73//21nGaDSKsrIy0adPH5GSkuLSp23Kuf28piQmJopFixa5bMvIyBBKpVL88ssvQgghiouLhUqlElu2bLng+Zq63gapqakiJCREWCwWl+1Ntb3RaBSrVq0SQrjXf79cvXr1Ei+++KLz8YQJE8S0adPcOnbEiBGiZ8+eLnE98sgjomfPns7Hl9M/zcrKEgDErl27nPszMzMbvT8TQoh//etfIiYmxq14iaj1cKQtEbWYhnmWLvTV94Z9Op0OAFBcXIyZM2eia9euMBqN8PX1RU1NDfLy8lyOW7FiBZ577jl079690dfILsWiRYtwww03YNiwYS7bExIS8MMPPyAnJ6fJ4w4ePIjjx4/Dx8fHOQI0ICAAZrMZJ06cQElJCX755ReMGjXqgud39zrOF+fZVq5cCaPRiJSUlAues0FKSgoOHDjg8nP2NAP+/v74z3/+g1dffRVdunTBo48+6nL83r17MWHCBERFRcHHxwcjRowAgEbPVd++fZ2/h4SEwMvLC3FxcS7bLvZ1uwaTJ0/GL7/8gvXr12Ps2LHYsmUL+vfv7xwpe+DAAURGRqJbt25NHp+ZmYmhQ4e6bBs6dCiys7PhcDic284eJQLUP9+rV692PtcGgwHJycmQZfm89wgRERF1TKIVlo3Zt28f1q5di8WLFzfap1KpMHjwYOdjHx8fJCYm4ujRo85tdrsd48ePx6FDhzBmzJhGi5S52887W11dnbNP32DQoEHo3bs33nrrLQDAu+++i+joaAwfPvzSL/osffr0ueQ5XS/Wf29K7969nWXHjRt33rpnzJjhHFFcXFyMjRs3Yvr06W7H9rvf/c7lORgyZMgV908zMzOhUqkwYMAA5zE9evRwmRqjgV6vd2u6MiJqXSpPB0BE7Ze/vz8GDx6Mt99+Gw888AC8vb1d9tvtdrz++uuIjIxEQkICACA1NRWlpaX417/+hejoaGi1WgwZMqTRV5Z27dqFL7/8EtOmTcPrr7+Ou++++5Ljy87OxptvvokDBw4gPz/fZd/06dOxdu1axMXFNYobqF/kYcCAAU3OC9WpUycoFO59JubOdVwozgbl5eVYvHgx1q5d6/bKwEaj8bzTNTTYtm0blEolCgsLYTKZnF9jM5lMSE5ORnJyMtasWYNOnTohLy8PycnJjZ4rtVrt/F2SJJfHDdtkWXYrZqA+wT969GiMHj0aCxcuxIwZM/DUU09h2rRpjb6KeLnOfc5rampw9913Y+7cuY3KRkVFNcs5iYiI6OrWsPZAZmYm/vCHPzTan5mZCX9/f3Tq1OmKz/WXv/wFf/3rXxEWFuay/XzTbAFw6SOaTCbo9Xq8/vrrmDdvHsaMGYPQ0FDnPnf7eWcLCgpyrs1wthkzZmDlypV49NFHsWrVKtx5551u91fPp6n+uSRJjRLmNpvN+fvF+u9N+fLLL511XKifOXXqVDz66KNIT0/H999/j9jYWFx//fVuXYu7LrV/euzYMbfrLisra5b7koiaF0faElGLevPNN2E2m9GzZ08888wzyMnJQUFBAZ5//nkkJCTg8OHDWLNmjXPO0Z07d2Lu3LkYP368c7GoM2fONKp32bJlGDduHF555RU89NBDF/zU/3weeeQRzJgxo8nEpV6vxzfffIOioiLnKNSz9e/fH9nZ2QgODkZ8fLzLj9FohI+PD2JiYpCWlnbBGNy5jgvF2WDx4sW4/vrrr3jUwtm+//57vPjii/j8889hMBhc5sf66aefUFpaihdeeAHXX389evTo4fZo2ebWq1cv55zIffv2RX5+/nk7qT179sTOnTtdtu3cuRPdunVz3oNN6d+/P44ePdrouY6Pj+9wqxUTERFR0wIDAzF69Gi88sorjeaPLSoqwpo1a3DbbbddccJy/fr1OHbsGP7617822tejRw/Y7XZkZGQ4t1VXV+PgwYPo1auXc5uXlxfWr1+PWbNmYejQoS4DBy63n9evXz+X0bwN7rjjDvz8889Yvnw5jh49itTUVLeus6GPdfZo0wvp1KmTy7yy2dnZLqNHL9Z/b0p0dLSzTMMCa00JDAzEzTffjFWrVmH16tW488473Yq5wdnPFwDn2h5X0j9tuBf27t3rPCYrK8u5oPHZDh8+jH79+l1SzETU8pi0JaIWlZCQgKysLDz22GPIzs5GZmYmjh8/jvT0dEyfPh1ZWVkuicauXbvinXfeQWZmJjIyMpCSktLkp9oBAQEA6r8uP378eMyYMeOS4jp+/Di2bNmCJ5988oLlQkJCnJ2fs6WkpCAoKAgTJ07E9u3bkZOTgy1btmDu3LnO0bBPP/00XnrpJSxfvhzZ2dnYt29fowW3LnYd7sRZW1uLN954A0uXLnX7+huOKyoqcvlpGB1RXV2NP//5z5g7dy7GjRuHNWvW4IMPPnCu/hsVFQWNRoOXX34ZJ0+exPr165v8el5zKi0txY033oh3330XP/74I3JycvDRRx9h6dKlmDhxIgBgxIgRGD58OCZPnozNmzcjJycHGzduxKZNmwDUj0pJS0vD4sWLcezYMbz11ltYsWJFk296zvbII4/g+++/x5w5c3DgwAFkZ2fjs88+40JkRERE5GLFihWwWCxITk7Gtm3bcOrUKWzatAmjR49GREQEnnvuOZfydrsdZrMZZrMZdrsdQgjnY7PZ3OQ5li5dimeffda5OOzZunfvjnHjxmHGjBnYvn07Dh06hKlTp8JgMOD22293llOr1TAYDACAN954A9u3b8e7774L4PL7ecnJyUhPT2+UZPX398ekSZPw0EMPYcyYMYiMjLxoXQAQHBwMvV6PTZs2obi4GJWVlRcsf+ONN2LFihXYv38/9uzZg3vuucflG17u9N+vxIwZM/DWW28hMzPT7cR0g7y8PMyfPx9ZWVl4//338fLLL+OBBx644DEX6592794dY8eOxd13342MjAzs3bsXM2bMaPK91fbt2zFmzJhLipmIWoGH59Qlog7m3IXIzrVv3z4xcOBAodPpRNeuXcVHH33UaJEGnLPIwOnTp0VwcLB4/fXXXeq60EJkAMTf//5357aGRbMaFvhqyrnnLSwsFFOnThVBQUFCq9WKuLg4MXPmTFFZWeks89prr4nu3bsLtVotwsLCxP333+/2dbgTZ8NiE3PmzHGWOd+iDWcbMWKEANDop2FBsDvvvFP06dNHmM1m5zEvvfSSCAgIEPn5+UIIId577z0RExMjtFqtGDJkiFi/fr3LeZtq07MXTmtw9oJrF2I2m8Wjjz4q+vfvL4xGo/Dy8hLdu3cXTzzxhKitrXWWKy0tFXfeeacIDAwUOp1OJCQkiA0bNjj3f/zxx6JXr15CrVaLqKgo8be//c3lPOfebw127dolRo8eLQwGg/D29hZ9+/a96OJsRERE1PHk5uY6F8tSq9Wic+fO4v777xdnzpxxKRcdHd1kf+zcn3MXIktMTBQOh8OlnrP7LqdPnxa33367MBqNQq/Xi6SkJHHkyBHn/qb6Y2+//bYICAgQhYWFQoiL9/OaYrPZRHh4uNi0aVOjfWlpaQKA+PDDD91sxXr//ve/RefOnYVCoRAjRowQQpz//URBQYEYM2aM8Pb2Fl27dhVffvmly0JkQrjXf79csiyL6OhoMX78+Es6bsSIEeK+++4T99xzj/D19RX+/v7isccec1mY7HL7p4WFheKmm24SWq1WREVFibfffrtRXd9//73w8/Nz6U8TUdsgCdEKs6QTERERERERUbu2cuVKrF+/Hl999ZXL9nfeeQcPPvggfvnll3Y7tVRNTQ0iIiKwatUqTJo0ye3jRo4ciWuuuQbLli1rueAu4LbbbkNiYiIee+wxj5yfiM6PC5ERERERERER0RW7++67UVFRgerqavj4+KC2thaFhYV44YUXcPfdd7fLhK0syzhz5gxeeukl+Pn54f/+7/88HZLbrFYr+vTpgwcffNDToRBREzinLRERtQl5eXkwGAzn/bmcxeaIiIiIqPWoVCo8/vjj8PHxAVA//26PHj0QGhqKBQsWuJR9/vnnz9vvGzdunCfCvyx5eXkICQnBe++9h//+979QqVQu+9py/1aj0eCJJ55ocp5bIvI8To9ARERtgt1uR25u7nn3x8TEuHSCiYiIiOjqVVZWhrKysib36fV6REREtHJEzY/9WyK6EkzaEhEREREREREREbUhnB6BiIiIiIiIiIiIqA1h0paIiIiIiIiIiIioDWHSloiIiIiIiIiIiKgNYdKWiIiIiIiIiIiIqA1h0paIiIiIiIiIiIioDWHSloiIiIiIiIiIiKgNYdKWiIiIiIiIiIiIqA1h0paIiIiIiIiIiIioDfl/SnTYk86he/wAAAAASUVORK5CYII=", "text/plain": [ "

" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "predictions_df = pd.DataFrame({\n", " \"y_true\": y_test.values,\n", " \"y_pred\": test_pred,\n", " \"abs_error\": np.abs(y_test.values - test_pred),\n", "}).sort_values(\"abs_error\", ascending=False)\n", "\n", "display(predictions_df.head(15))\n", "\n", "fig, ax = plt.subplots(1, 2, figsize=(14, 5))\n", "\n", "ax[0].scatter(y_test, test_pred, alpha=0.6)\n", "min_val = min(y_test.min(), test_pred.min())\n", "max_val = max(y_test.max(), test_pred.max())\n", "ax[0].plot([min_val, max_val], [min_val, max_val], linestyle=\"--\")\n", "ax[0].set_title(\"Факт vs прогноз\")\n", "ax[0].set_xlabel(\"Фактический Exam_Score\")\n", "ax[0].set_ylabel(\"Прогноз\")\n", "\n", "residuals = y_test.values - test_pred\n", "ax[1].hist(residuals, bins=25, edgecolor=\"black\")\n", "ax[1].set_title(\"Распределение остатков\")\n", "ax[1].set_xlabel(\"Ошибка (y_true - y_pred)\")\n", "ax[1].set_ylabel(\"Частота\")\n", "\n", "plt.tight_layout()\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": 13, "id": "7ca307d9", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Файлы сохранены:\n", "- mlp_search_full_features.csv\n", "- mlp_search_key_features.csv\n", "- feature_importance_permutation.csv\n", "- test_predictions_neural_net.csv\n" ] } ], "source": [ "# Сохраняем технические результаты, чтобы потом было удобнее ссылаться на них в выводе.\n", "full_results.to_csv(\"mlp_search_full_features.csv\", index=False)\n", "key_results.to_csv(\"mlp_search_key_features.csv\", index=False)\n", "importance_df.to_csv(\"feature_importance_permutation.csv\", index=False)\n", "predictions_df.to_csv(\"test_predictions_neural_net.csv\", index=False)\n", "\n", "print(\"Файлы сохранены:\")\n", "print(\"- mlp_search_full_features.csv\")\n", "print(\"- mlp_search_key_features.csv\")\n", "print(\"- feature_importance_permutation.csv\")\n", "print(\"- test_predictions_neural_net.csv\")" ] }, { "cell_type": "markdown", "id": "dea1344a", "metadata": {}, "source": [ "Вывод.\n", "В работе решалась задача регрессии — прогнозирование Exam_Score по характеристикам учащегося. Для решения применялась полносвязная нейронная сеть (MLPRegressor) с масштабированием числовых признаков и кодированием категориальных переменных. Ключевыми факторами оказались прежде всего Attendance, Hours_Studied, Previous_Scores, Access_to_Resources, Parental_Involvement и Tutoring_Sessions.\n", "\n", "Наилучший результат показала модель на полном наборе признаков с архитектурой (128, 32): на тестовой выборке получено RMSE = 1.711, MAE = 0.463, R² = 0.796. Это означает, что модель достаточно точно прогнозирует итоговый балл и объясняет около 79.6% изменчивости целевой переменной. Полный набор признаков оказался информативнее сокращённого, поэтому именно его использование следует считать наиболее эффективным." ] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "name": "python" } }, "nbformat": 4, "nbformat_minor": 5 }