notebooks/Post-processing.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## This notebook demonstrates how to use bias mitigation algorithm during \"Post-processing stage\" from AI 360 fairness toolkit"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Post-processing algorithm\n",
    "A bias mitigation algorithm that is applied to predicted labels."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Insert your credentials as credentials in the below cell\n",
    "Click on dropdown from Pipeline_LabelEncoder-0.1.zip under Data tab and select 'Credentials'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from ibm_botocore.client import Config\n",
    "import ibm_boto3\n",
    "\n",
    "cos = ibm_boto3.client(service_name='s3',\n",
    "    ibm_api_key_id=credentials['IBM_API_KEY_ID'],\n",
    "    ibm_service_instance_id=credentials['IAM_SERVICE_ID'],\n",
    "    ibm_auth_endpoint=credentials['IBM_AUTH_ENDPOINT'],\n",
    "    config=Config(signature_version='oauth'),\n",
    "    endpoint_url=credentials['ENDPOINT'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/home/wsuser/work'"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import os\n",
    "os.getcwd()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "cos.download_file(Bucket=credentials['BUCKET'],Key='Pipeline_LabelEncoder-0.1.zip',Filename='/home/wsuser/work/Pipeline_LabelEncoder-0.1.zip')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pipeline_LabelEncoder-0.1.zip\r\n"
     ]
    }
   ],
   "source": [
    "!ls"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing ./Pipeline_LabelEncoder-0.1.zip\n",
      "Building wheels for collected packages: Pipeline-LabelEncoder\n",
      "  Building wheel for Pipeline-LabelEncoder (setup.py) ... \u001b[?25ldone\n",
      "\u001b[?25h  Created wheel for Pipeline-LabelEncoder: filename=Pipeline_LabelEncoder-0.1-py3-none-any.whl size=2062 sha256=ca70e060b2091f8204a7377e68a004e429d9e435e77d79d6577cd377e78125ab\n",
      "  Stored in directory: /tmp/wsuser/.cache/pip/wheels/6b/4b/1e/43f3c8b97ffade4539a329b9eaa5755e4df16a248960947534\n",
      "Successfully built Pipeline-LabelEncoder\n",
      "Installing collected packages: Pipeline-LabelEncoder\n",
      "  Attempting uninstall: Pipeline-LabelEncoder\n",
      "    Found existing installation: Pipeline-LabelEncoder 0.1\n",
      "    Uninstalling Pipeline-LabelEncoder-0.1:\n",
      "      Successfully uninstalled Pipeline-LabelEncoder-0.1\n",
      "Successfully installed Pipeline-LabelEncoder-0.1\n",
      "Requirement already satisfied: aif360 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (0.4.0)\n",
      "Requirement already satisfied: scikit-learn>=0.22.1 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from aif360) (0.23.2)\n",
      "Requirement already satisfied: matplotlib in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from aif360) (3.3.4)\n",
      "Requirement already satisfied: numpy>=1.16 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from aif360) (1.19.2)\n",
      "Requirement already satisfied: scipy<1.6.0,>=1.2.0 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from aif360) (1.4.1)\n",
      "Requirement already satisfied: tempeh in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from aif360) (0.1.12)\n",
      "Requirement already satisfied: pandas>=0.24.0 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from aif360) (1.2.4)\n",
      "Requirement already satisfied: python-dateutil>=2.7.3 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from pandas>=0.24.0->aif360) (2.8.1)\n",
      "Requirement already satisfied: pytz>=2017.3 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from pandas>=0.24.0->aif360) (2021.1)\n",
      "Requirement already satisfied: six>=1.5 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from python-dateutil>=2.7.3->pandas>=0.24.0->aif360) (1.15.0)\n",
      "Requirement already satisfied: joblib>=0.11 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from scikit-learn>=0.22.1->aif360) (0.17.0)\n",
      "Requirement already satisfied: threadpoolctl>=2.0.0 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from scikit-learn>=0.22.1->aif360) (2.1.0)\n",
      "Requirement already satisfied: kiwisolver>=1.0.1 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from matplotlib->aif360) (1.3.1)\n",
      "Requirement already satisfied: pillow>=6.2.0 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from matplotlib->aif360) (8.4.0)\n",
      "Requirement already satisfied: cycler>=0.10 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from matplotlib->aif360) (0.10.0)\n",
      "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.3 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from matplotlib->aif360) (2.4.7)\n",
      "Requirement already satisfied: shap in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from tempeh->aif360) (0.40.0)\n",
      "Requirement already satisfied: memory-profiler in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from tempeh->aif360) (0.59.0)\n",
      "Requirement already satisfied: pytest in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from tempeh->aif360) (6.2.3)\n",
      "Requirement already satisfied: requests in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from tempeh->aif360) (2.25.1)\n",
      "Requirement already satisfied: psutil in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from memory-profiler->tempeh->aif360) (5.8.0)\n",
      "Requirement already satisfied: attrs>=19.2.0 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from pytest->tempeh->aif360) (21.2.0)\n",
      "Requirement already satisfied: iniconfig in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from pytest->tempeh->aif360) (1.1.1)\n",
      "Requirement already satisfied: packaging in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from pytest->tempeh->aif360) (21.3)\n",
      "Requirement already satisfied: pluggy<1.0.0a1,>=0.12 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from pytest->tempeh->aif360) (0.13.1)\n",
      "Requirement already satisfied: py>=1.8.2 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from pytest->tempeh->aif360) (1.10.0)\n",
      "Requirement already satisfied: toml in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from pytest->tempeh->aif360) (0.10.2)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from requests->tempeh->aif360) (2021.10.8)\n",
      "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from requests->tempeh->aif360) (1.26.6)\n",
      "Requirement already satisfied: chardet<5,>=3.0.2 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from requests->tempeh->aif360) (3.0.4)\n",
      "Requirement already satisfied: idna<3,>=2.5 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from requests->tempeh->aif360) (2.8)\n",
      "Requirement already satisfied: cloudpickle in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from shap->tempeh->aif360) (1.6.0)\n",
      "Requirement already satisfied: numba in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from shap->tempeh->aif360) (0.54.1)\n",
      "Requirement already satisfied: slicer==0.0.7 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from shap->tempeh->aif360) (0.0.7)\n",
      "Requirement already satisfied: tqdm>4.25.0 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from shap->tempeh->aif360) (4.59.0)\n",
      "Requirement already satisfied: llvmlite<0.38,>=0.37.0rc1 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from numba->shap->tempeh->aif360) (0.37.0)\n",
      "Requirement already satisfied: setuptools in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from numba->shap->tempeh->aif360) (52.0.0.post20211006)\n",
      "Requirement already satisfied: fairlearn in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (0.7.0)\n",
      "Requirement already satisfied: pandas>=0.25.1 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from fairlearn) (1.2.4)\n",
      "Requirement already satisfied: scipy>=1.4.1 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from fairlearn) (1.4.1)\n",
      "Requirement already satisfied: numpy>=1.17.2 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from fairlearn) (1.19.2)\n",
      "Requirement already satisfied: scikit-learn>=0.22.1 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from fairlearn) (0.23.2)\n",
      "Requirement already satisfied: python-dateutil>=2.7.3 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from pandas>=0.25.1->fairlearn) (2.8.1)\n",
      "Requirement already satisfied: pytz>=2017.3 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from pandas>=0.25.1->fairlearn) (2021.1)\n",
      "Requirement already satisfied: six>=1.5 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from python-dateutil>=2.7.3->pandas>=0.25.1->fairlearn) (1.15.0)\n",
      "Requirement already satisfied: joblib>=0.11 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from scikit-learn>=0.22.1->fairlearn) (0.17.0)\n",
      "Requirement already satisfied: threadpoolctl>=2.0.0 in /opt/conda/envs/Python-3.8-main/lib/python3.8/site-packages (from scikit-learn>=0.22.1->fairlearn) (2.1.0)\n"
     ]
    }
   ],
   "source": [
    "!pip install Pipeline_LabelEncoder-0.1.zip\n",
    "!pip install aif360\n",
    "!pip install fairlearn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'2.4.4'"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "tf.compat.v1.disable_eager_execution()\n",
    "tf.__version__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "# Load all necessary packages\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from aif360.datasets import BinaryLabelDataset\n",
    "from aif360.metrics import BinaryLabelDatasetMetric\n",
    "from aif360.metrics import ClassificationMetric\n",
    "from aif360.metrics.utils import compute_boolean_conditioning_vector\n",
    "\n",
    "from sklearn.preprocessing import scale\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "from IPython.display import Markdown, display\n",
    "import matplotlib.pyplot as plt\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Insert the data as Pandas Dataframe and change the name from df_data_ to df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Gender</th>\n",
       "      <th>Married</th>\n",
       "      <th>Education</th>\n",
       "      <th>Fraud_risk</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Male</td>\n",
       "      <td>No</td>\n",
       "      <td>1</td>\n",
       "      <td>Risk</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Male</td>\n",
       "      <td>Yes</td>\n",
       "      <td>1</td>\n",
       "      <td>Safe</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Male</td>\n",
       "      <td>Yes</td>\n",
       "      <td>1</td>\n",
       "      <td>Safe</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Male</td>\n",
       "      <td>Yes</td>\n",
       "      <td>0</td>\n",
       "      <td>Safe</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Male</td>\n",
       "      <td>No</td>\n",
       "      <td>1</td>\n",
       "      <td>Risk</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Gender Married  Education Fraud_risk\n",
       "0   Male      No          1       Risk\n",
       "1   Male     Yes          1       Safe\n",
       "2   Male     Yes          1       Safe\n",
       "3   Male     Yes          0       Safe\n",
       "4   Male      No          1       Risk"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Gender</th>\n",
       "      <th>Married</th>\n",
       "      <th>Education</th>\n",
       "      <th>Fraud_risk</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>921</td>\n",
       "      <td>921</td>\n",
       "      <td>921.000000</td>\n",
       "      <td>921</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>top</th>\n",
       "      <td>Male</td>\n",
       "      <td>No</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Safe</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>freq</th>\n",
       "      <td>703</td>\n",
       "      <td>501</td>\n",
       "      <td>NaN</td>\n",
       "      <td>562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.730727</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.443823</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Gender Married   Education Fraud_risk\n",
       "count     921     921  921.000000        921\n",
       "unique      2       2         NaN          2\n",
       "top      Male      No         NaN       Safe\n",
       "freq      703     501         NaN        562\n",
       "mean      NaN     NaN    0.730727        NaN\n",
       "std       NaN     NaN    0.443823        NaN\n",
       "min       NaN     NaN    0.000000        NaN\n",
       "25%       NaN     NaN    0.000000        NaN\n",
       "50%       NaN     NaN    1.000000        NaN\n",
       "75%       NaN     NaN    1.000000        NaN\n",
       "max       NaN     NaN    1.000000        NaN"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe(include = 'all')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "privileged_groups = [{'Gender': 1}]\n",
    "unprivileged_groups = [{'Gender': 0}]\n",
    "cost_constraint='weighted'\n",
    "favorable_label = 1 \n",
    "unfavorable_label = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature Gender\n",
      "mapping {'Female': 0, 'Male': 1}\n",
      "Feature Married\n",
      "mapping {'No': 0, 'Yes': 1}\n",
      "Feature Fraud_risk\n",
      "mapping {'Risk': 0, 'Safe': 1}\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Gender</th>\n",
       "      <th>Married</th>\n",
       "      <th>Education</th>\n",
       "      <th>Fraud_risk</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Gender  Married  Education  Fraud_risk\n",
       "0       1        0          1           0\n",
       "1       1        1          1           1\n",
       "2       1        1          1           1\n",
       "3       1        1          0           1\n",
       "4       1        0          1           0"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn import preprocessing\n",
    "categorical_column = ['Gender', 'Married', 'Fraud_risk']\n",
    "\n",
    "data_encoded = df.copy(deep=True)\n",
    "#Use Scikit-learn label encoding to encode character data\n",
    "lab_enc = preprocessing.LabelEncoder()\n",
    "for col in categorical_column:\n",
    "        data_encoded[col] = lab_enc.fit_transform(df[col])\n",
    "        le_name_mapping = dict(zip(lab_enc.classes_, lab_enc.transform(lab_enc.classes_)))\n",
    "        print('Feature', col)\n",
    "        print('mapping', le_name_mapping)\n",
    "        \n",
    "\n",
    "data_encoded.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing ./Pipeline_LabelEncoder-0.1.zip\n",
      "Building wheels for collected packages: Pipeline-LabelEncoder\n",
      "  Building wheel for Pipeline-LabelEncoder (setup.py) ... \u001b[?25ldone\n",
      "\u001b[?25h  Created wheel for Pipeline-LabelEncoder: filename=Pipeline_LabelEncoder-0.1-py3-none-any.whl size=2062 sha256=7c148783d4f7fad1b0f928863dd159e80e0f76c02298408db5b5bd44b7d0e78a\n",
      "  Stored in directory: /tmp/wsuser/.cache/pip/wheels/6b/4b/1e/43f3c8b97ffade4539a329b9eaa5755e4df16a248960947534\n",
      "Successfully built Pipeline-LabelEncoder\n",
      "Installing collected packages: Pipeline-LabelEncoder\n",
      "  Attempting uninstall: Pipeline-LabelEncoder\n",
      "    Found existing installation: Pipeline-LabelEncoder 0.1\n",
      "    Uninstalling Pipeline-LabelEncoder-0.1:\n",
      "      Successfully uninstalled Pipeline-LabelEncoder-0.1\n",
      "Successfully installed Pipeline-LabelEncoder-0.1\n"
     ]
    }
   ],
   "source": [
    "!pip install Pipeline_LabelEncoder-0.1.zip"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Inside fit transform\n",
      "Feature Gender\n",
      "mapping {0: 0, 1: 1}\n",
      "Feature Married\n",
      "mapping {0: 0, 1: 1}\n",
      "Feature Fraud_risk\n",
      "mapping {0: 0, 1: 1}\n",
      "-------------------------\n"
     ]
    }
   ],
   "source": [
    "from Pipeline_LabelEncoder.sklearn_label_encoder import PipelineLabelEncoder\n",
    "preprocessed_data = PipelineLabelEncoder(columns = ['Gender','Married', 'Fraud_risk']).fit_transform(data_encoded)\n",
    "print('-------------------------')\n",
    "#print('validation data encoding')\n",
    "#validation_enc_data = PipelineLabelEncoder(columns = ['Gender','Married', 'Fraud_risk']).transform(validation_input_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Create binary label dataset that can be used by bias mitigation algorithms\n",
    "fraud_dataset = BinaryLabelDataset(favorable_label=favorable_label,\n",
    "                                unfavorable_label=unfavorable_label,\n",
    "                                df=preprocessed_data,\n",
    "                                label_names=['Fraud_risk'],\n",
    "                                protected_attribute_names=['Gender', 'Married'],\n",
    "                                unprivileged_protected_attributes=unprivileged_groups)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_orig_train, dataset_orig_vt = fraud_dataset.split([0.6], shuffle=True)\n",
    "dataset_orig_valid, dataset_orig_test = dataset_orig_vt.split([0.5], shuffle=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training data characteristics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "#### Dataset shape"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(552, 3)\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "#### Favorable and unfavorable labels"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.0 0.0\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "#### Protected attribute names"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Gender', 'Married']\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "#### Privileged and unprivileged protected attribute values"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[array([1.]), array([1.])] [array([0.]), array([0.])]\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "#### Dataset feature names"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Gender', 'Married', 'Education']\n"
     ]
    }
   ],
   "source": [
    "# print out some labels, names, etc.\n",
    "display(Markdown(\"#### Dataset shape\"))\n",
    "print(dataset_orig_train.features.shape)\n",
    "display(Markdown(\"#### Favorable and unfavorable labels\"))\n",
    "print(dataset_orig_train.favorable_label, dataset_orig_train.unfavorable_label)\n",
    "display(Markdown(\"#### Protected attribute names\"))\n",
    "print(dataset_orig_train.protected_attribute_names)\n",
    "display(Markdown(\"#### Privileged and unprivileged protected attribute values\"))\n",
    "print(dataset_orig_train.privileged_protected_attributes, dataset_orig_train.unprivileged_protected_attributes)\n",
    "display(Markdown(\"#### Dataset feature names\"))\n",
    "print(dataset_orig_train.feature_names)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Metric for the original datasets (without any classifiers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "#### Original training dataset"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Difference in mean outcomes between unprivileged and privileged groups = -0.335727\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "#### Original validation dataset"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Difference in mean outcomes between unprivileged and privileged groups = -0.396668\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "#### Original test dataset"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Difference in mean outcomes between unprivileged and privileged groups = -0.460317\n"
     ]
    }
   ],
   "source": [
    "metric_orig_train = BinaryLabelDatasetMetric(dataset_orig_train, \n",
    "                             unprivileged_groups=unprivileged_groups,\n",
    "                             privileged_groups=privileged_groups)\n",
    "display(Markdown(\"#### Original training dataset\"))\n",
    "print(\"Difference in mean outcomes between unprivileged and privileged groups = %f\" % metric_orig_train.mean_difference())\n",
    "\n",
    "metric_orig_valid = BinaryLabelDatasetMetric(dataset_orig_valid, \n",
    "                             unprivileged_groups=unprivileged_groups,\n",
    "                             privileged_groups=privileged_groups)\n",
    "display(Markdown(\"#### Original validation dataset\"))\n",
    "print(\"Difference in mean outcomes between unprivileged and privileged groups = %f\" % metric_orig_valid.mean_difference())\n",
    "\n",
    "metric_orig_test = BinaryLabelDatasetMetric(dataset_orig_test, \n",
    "                             unprivileged_groups=unprivileged_groups,\n",
    "                             privileged_groups=privileged_groups)\n",
    "display(Markdown(\"#### Original test dataset\"))\n",
    "print(\"Difference in mean outcomes between unprivileged and privileged groups = %f\" % metric_orig_test.mean_difference())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train classifier (logistic regression on original training data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.metrics import roc_curve\n",
    "\n",
    "# Placeholder for predicted and transformed datasets\n",
    "dataset_orig_train_pred = dataset_orig_train.copy(deepcopy=True)\n",
    "dataset_orig_valid_pred = dataset_orig_valid.copy(deepcopy=True)\n",
    "dataset_orig_test_pred = dataset_orig_test.copy(deepcopy=True)\n",
    "\n",
    "dataset_new_valid_pred = dataset_orig_valid.copy(deepcopy=True)\n",
    "dataset_new_test_pred = dataset_orig_test.copy(deepcopy=True)\n",
    "\n",
    "# Logistic regression classifier and predictions for training data\n",
    "scale_orig = StandardScaler()\n",
    "X_train = scale_orig.fit_transform(dataset_orig_train.features)\n",
    "y_train = dataset_orig_train.labels.ravel()\n",
    "lmod = LogisticRegression()\n",
    "lmod.fit(X_train, y_train)\n",
    "\n",
    "fav_idx = np.where(lmod.classes_ == dataset_orig_train.favorable_label)[0][0]\n",
    "y_train_pred_prob = lmod.predict_proba(X_train)[:,fav_idx]\n",
    "\n",
    "# Prediction probs for validation and testing data\n",
    "X_valid = scale_orig.transform(dataset_orig_valid.features)\n",
    "y_valid_pred_prob = lmod.predict_proba(X_valid)[:,fav_idx]\n",
    "\n",
    "X_test = scale_orig.transform(dataset_orig_test.features)\n",
    "y_test_pred_prob = lmod.predict_proba(X_test)[:,fav_idx]\n",
    "\n",
    "class_thresh = 0.5\n",
    "dataset_orig_train_pred.scores = y_train_pred_prob.reshape(-1,1)\n",
    "dataset_orig_valid_pred.scores = y_valid_pred_prob.reshape(-1,1)\n",
    "dataset_orig_test_pred.scores = y_test_pred_prob.reshape(-1,1)\n",
    "\n",
    "y_train_pred = np.zeros_like(dataset_orig_train_pred.labels)\n",
    "y_train_pred[y_train_pred_prob >= class_thresh] = dataset_orig_train_pred.favorable_label\n",
    "y_train_pred[~(y_train_pred_prob >= class_thresh)] = dataset_orig_train_pred.unfavorable_label\n",
    "dataset_orig_train_pred.labels = y_train_pred\n",
    "\n",
    "y_valid_pred = np.zeros_like(dataset_orig_valid_pred.labels)\n",
    "y_valid_pred[y_valid_pred_prob >= class_thresh] = dataset_orig_valid_pred.favorable_label\n",
    "y_valid_pred[~(y_valid_pred_prob >= class_thresh)] = dataset_orig_valid_pred.unfavorable_label\n",
    "dataset_orig_valid_pred.labels = y_valid_pred\n",
    "    \n",
    "y_test_pred = np.zeros_like(dataset_orig_test_pred.labels)\n",
    "y_test_pred[y_test_pred_prob >= class_thresh] = dataset_orig_test_pred.favorable_label\n",
    "y_test_pred[~(y_test_pred_prob >= class_thresh)] = dataset_orig_test_pred.unfavorable_label\n",
    "dataset_orig_test_pred.labels = y_test_pred"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Results before post-processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "#### Original-Predicted training dataset"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Difference in GFPR between unprivileged and privileged groups\n",
      "-0.18984395152026312\n",
      "Difference in GFNR between unprivileged and privileged groups\n",
      "0.20564974255683074\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "#### Original-Predicted validation dataset"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Difference in GFPR between unprivileged and privileged groups\n",
      "-0.21128922832410973\n",
      "Difference in GFNR between unprivileged and privileged groups\n",
      "0.19262371319267885\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "#### Original-Predicted testing dataset"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Difference in GFPR between unprivileged and privileged groups\n",
      "-0.21800999378346622\n",
      "Difference in GFNR between unprivileged and privileged groups\n",
      "0.170629864716808\n"
     ]
    }
   ],
   "source": [
    "cm_pred_train = ClassificationMetric(dataset_orig_train, dataset_orig_train_pred,\n",
    "                             unprivileged_groups=unprivileged_groups,\n",
    "                             privileged_groups=privileged_groups)\n",
    "display(Markdown(\"#### Original-Predicted training dataset\"))\n",
    "print(\"Difference in GFPR between unprivileged and privileged groups\")\n",
    "print(cm_pred_train.difference(cm_pred_train.generalized_false_positive_rate))\n",
    "print(\"Difference in GFNR between unprivileged and privileged groups\")\n",
    "print(cm_pred_train.difference(cm_pred_train.generalized_false_negative_rate))\n",
    "\n",
    "cm_pred_valid = ClassificationMetric(dataset_orig_valid, dataset_orig_valid_pred,\n",
    "                             unprivileged_groups=unprivileged_groups,\n",
    "                             privileged_groups=privileged_groups)\n",
    "display(Markdown(\"#### Original-Predicted validation dataset\"))\n",
    "print(\"Difference in GFPR between unprivileged and privileged groups\")\n",
    "print(cm_pred_valid.difference(cm_pred_valid.generalized_false_positive_rate))\n",
    "print(\"Difference in GFNR between unprivileged and privileged groups\")\n",
    "print(cm_pred_valid.difference(cm_pred_valid.generalized_false_negative_rate))\n",
    "\n",
    "cm_pred_test = ClassificationMetric(dataset_orig_test, dataset_orig_test_pred,\n",
    "                             unprivileged_groups=unprivileged_groups,\n",
    "                             privileged_groups=privileged_groups)\n",
    "display(Markdown(\"#### Original-Predicted testing dataset\"))\n",
    "print(\"Difference in GFPR between unprivileged and privileged groups\")\n",
    "print(cm_pred_test.difference(cm_pred_test.generalized_false_positive_rate))\n",
    "print(\"Difference in GFNR between unprivileged and privileged groups\")\n",
    "print(cm_pred_test.difference(cm_pred_test.generalized_false_negative_rate))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Perform odds equalizing post processing on scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Odds equalizing post-processing algorithm\n",
    "from aif360.algorithms.postprocessing.calibrated_eq_odds_postprocessing import CalibratedEqOddsPostprocessing\n",
    "from tqdm import tqdm\n",
    "\n",
    "# Learn parameters to equalize odds and apply to create a new dataset\n",
    "cpp = CalibratedEqOddsPostprocessing(privileged_groups = privileged_groups,\n",
    "                                     unprivileged_groups = unprivileged_groups,\n",
    "                                     cost_constraint=cost_constraint,\n",
    "                                     seed=None)\n",
    "cpp = cpp.fit(dataset_orig_valid, dataset_orig_valid_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Transform validation and test data using the post processing algorithm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_transf_valid_pred = cpp.predict(dataset_orig_valid_pred)\n",
    "dataset_transf_test_pred = cpp.predict(dataset_orig_test_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Results after post-processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "#### Original-Transformed validation dataset"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Difference in GFPR between unprivileged and privileged groups\n",
      "-0.22387251471792324\n",
      "Difference in GFNR between unprivileged and privileged groups\n",
      "0.17486203209269477\n"
     ]
    },
    {
     "data": {
      "text/markdown": [
       "#### Original-Transformed testing dataset"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Difference in GFPR between unprivileged and privileged groups\n",
      "-0.21800999378346622\n",
      "Difference in GFNR between unprivileged and privileged groups\n",
      "0.16020525050725246\n"
     ]
    }
   ],
   "source": [
    "cm_transf_valid = ClassificationMetric(dataset_orig_valid, dataset_transf_valid_pred,\n",
    "                             unprivileged_groups=unprivileged_groups,\n",
    "                             privileged_groups=privileged_groups)\n",
    "display(Markdown(\"#### Original-Transformed validation dataset\"))\n",
    "print(\"Difference in GFPR between unprivileged and privileged groups\")\n",
    "print(cm_transf_valid.difference(cm_transf_valid.generalized_false_positive_rate))\n",
    "print(\"Difference in GFNR between unprivileged and privileged groups\")\n",
    "print(cm_transf_valid.difference(cm_transf_valid.generalized_false_negative_rate))\n",
    "\n",
    "cm_transf_test = ClassificationMetric(dataset_orig_test, dataset_transf_test_pred,\n",
    "                             unprivileged_groups=unprivileged_groups,\n",
    "                             privileged_groups=privileged_groups)\n",
    "display(Markdown(\"#### Original-Transformed testing dataset\"))\n",
    "print(\"Difference in GFPR between unprivileged and privileged groups\")\n",
    "print(cm_transf_test.difference(cm_transf_test.generalized_false_positive_rate))\n",
    "print(\"Difference in GFNR between unprivileged and privileged groups\")\n",
    "print(cm_transf_test.difference(cm_transf_test.generalized_false_negative_rate))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Testing: Check if the rates for validation data has gone down after transformation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.19262371319267885\n",
      "0.17486203209269477\n"
     ]
    }
   ],
   "source": [
    "print(np.abs(cm_pred_valid.difference(cm_pred_valid.generalized_false_negative_rate)))\n",
    "print(np.abs(cm_transf_valid.difference(cm_transf_valid.generalized_false_negative_rate)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/markdown": [
       "#### Classification thresholds used for validation and parameter selection"
      ],
      "text/plain": [
       "<IPython.core.display.Markdown object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 25/25 [00:00<00:00, 115.97it/s]\n"
     ]
    }
   ],
   "source": [
    "# Thresholds\n",
    "all_thresh = np.linspace(0.01, 0.99, 25)\n",
    "display(Markdown(\"#### Classification thresholds used for validation and parameter selection\"))\n",
    "\n",
    "bef_avg_odds_diff_test = []\n",
    "bef_avg_odds_diff_valid = []\n",
    "aft_avg_odds_diff_test = []\n",
    "aft_avg_odds_diff_valid = []\n",
    "bef_bal_acc_valid = []\n",
    "bef_bal_acc_test = []\n",
    "aft_bal_acc_valid = []\n",
    "aft_bal_acc_test = []\n",
    "for thresh in tqdm(all_thresh):\n",
    "    \n",
    "    dataset_orig_valid_pred_thresh = dataset_orig_valid_pred.copy(deepcopy=True)\n",
    "    dataset_orig_test_pred_thresh = dataset_orig_test_pred.copy(deepcopy=True)\n",
    "    dataset_transf_valid_pred_thresh = dataset_transf_valid_pred.copy(deepcopy=True)\n",
    "    dataset_transf_test_pred_thresh = dataset_transf_test_pred.copy(deepcopy=True)\n",
    "    \n",
    "    # Labels for the datasets from scores\n",
    "    y_temp = np.zeros_like(dataset_orig_valid_pred_thresh.labels)\n",
    "    y_temp[dataset_orig_valid_pred_thresh.scores >= thresh] = dataset_orig_valid_pred_thresh.favorable_label\n",
    "    y_temp[~(dataset_orig_valid_pred_thresh.scores >= thresh)] = dataset_orig_valid_pred_thresh.unfavorable_label\n",
    "    dataset_orig_valid_pred_thresh.labels = y_temp\n",
    "\n",
    "    y_temp = np.zeros_like(dataset_orig_test_pred_thresh.labels)\n",
    "    y_temp[dataset_orig_test_pred_thresh.scores >= thresh] = dataset_orig_test_pred_thresh.favorable_label\n",
    "    y_temp[~(dataset_orig_test_pred_thresh.scores >= thresh)] = dataset_orig_test_pred_thresh.unfavorable_label\n",
    "    dataset_orig_test_pred_thresh.labels = y_temp\n",
    "    \n",
    "    y_temp = np.zeros_like(dataset_transf_valid_pred_thresh.labels)\n",
    "    y_temp[dataset_transf_valid_pred_thresh.scores >= thresh] = dataset_transf_valid_pred_thresh.favorable_label\n",
    "    y_temp[~(dataset_transf_valid_pred_thresh.scores >= thresh)] = dataset_transf_valid_pred_thresh.unfavorable_label\n",
    "    dataset_transf_valid_pred_thresh.labels = y_temp\n",
    "    \n",
    "    y_temp = np.zeros_like(dataset_transf_test_pred_thresh.labels)\n",
    "    y_temp[dataset_transf_test_pred_thresh.scores >= thresh] = dataset_transf_test_pred_thresh.favorable_label\n",
    "    y_temp[~(dataset_transf_test_pred_thresh.scores >= thresh)] = dataset_transf_test_pred_thresh.unfavorable_label\n",
    "    dataset_transf_test_pred_thresh.labels = y_temp\n",
    "    \n",
    "    # Metrics for original validation data\n",
    "    classified_metric_orig_valid = ClassificationMetric(dataset_orig_valid,\n",
    "                                                 dataset_orig_valid_pred_thresh,\n",
    "                                                 unprivileged_groups=unprivileged_groups,\n",
    "                                                 privileged_groups=privileged_groups)\n",
    "    bef_avg_odds_diff_valid.append(classified_metric_orig_valid.equal_opportunity_difference())\n",
    "\n",
    "    bef_bal_acc_valid.append(0.5*(classified_metric_orig_valid.true_positive_rate()+\n",
    "                              classified_metric_orig_valid.true_negative_rate()))\n",
    "\n",
    "    classified_metric_orig_test = ClassificationMetric(dataset_orig_test,\n",
    "                                                 dataset_orig_test_pred_thresh,\n",
    "                                                 unprivileged_groups=unprivileged_groups,\n",
    "                                                 privileged_groups=privileged_groups)\n",
    "    bef_avg_odds_diff_test.append(classified_metric_orig_test.equal_opportunity_difference())\n",
    "    bef_bal_acc_test.append(0.5*(classified_metric_orig_test.true_positive_rate()+\n",
    "                              classified_metric_orig_test.true_negative_rate()))\n",
    "\n",
    "    # Metrics for transf validing data\n",
    "    classified_metric_transf_valid = ClassificationMetric(\n",
    "                                     dataset_orig_valid, \n",
    "                                     dataset_transf_valid_pred_thresh,\n",
    "                                     unprivileged_groups=unprivileged_groups,\n",
    "                                     privileged_groups=privileged_groups)\n",
    "    aft_avg_odds_diff_valid.append(classified_metric_transf_valid.equal_opportunity_difference())\n",
    "    aft_bal_acc_valid.append(0.5*(classified_metric_transf_valid.true_positive_rate()+\n",
    "                              classified_metric_transf_valid.true_negative_rate()))\n",
    "\n",
    "    # Metrics for transf validation data\n",
    "    classified_metric_transf_test = ClassificationMetric(dataset_orig_test,\n",
    "                                                 dataset_transf_test_pred_thresh,\n",
    "                                                 unprivileged_groups=unprivileged_groups,\n",
    "                                                 privileged_groups=privileged_groups)\n",
    "    aft_avg_odds_diff_test.append(classified_metric_transf_test.equal_opportunity_difference())\n",
    "    aft_bal_acc_test.append(0.5*(classified_metric_transf_test.true_positive_rate()+\n",
    "                                  classified_metric_transf_test.true_negative_rate()))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fd19ca6f100>"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3IAAAHnCAYAAAD94284AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAADin0lEQVR4nOzdd3hURffA8e+kd9ILoST0XgKIiAq8KAIKIqD4Q1FsiKgIiu0FAbFg4UURC6Ii2AALFlQsoIAoNlAsJPSSQCgJpPdkfn/MbrJppLDJAjmf59lnd++dvXd2s0n27Myco7TWCCGEEEIIIYQ4ezg5ugNCCCGEEEIIIWpGAjkhhBBCCCGEOMtIICeEEEIIIYQQZxkJ5IQQQgghhBDiLCOBnBBCCCGEEEKcZSSQE0IIIYQQQoizjARyQgghhBBCCHGWcXF0B4QQQgghGpqtW7de5uLiMktrHY58sS6EKFGklErRWm8rKCh4skePHrsqa6ikILgQQgghRP3ZunXrZe7u7i9GRUXleXp65jg5OcmHMSEEAFpr8vPzXdLS0ryPHDmic3NzJ/fo0eOritrKN0BCCCGEEPXIxcVlVlRUVJ63t3e2BHFCCFtKKdzc3AqCg4NTo6KiClxdXR+orK0EckIIIYQQ9UhrHe7p6Znj6H4IIc5s3t7eWVrr6Mr2SyAnhBBCCFG/nGQkTghRFaUUgKpsvwRyQgghhBBCCHGWkUBOCCGEEELU2gsvvBCklOphvTg7O/cIDQ3tMnTo0Bbbtm1zr80xR40aFRUZGdnZ3n2tTzV9Drt373Z1dnbu4erqGpOYmHhWZpZfvHhxQK9evdr6+vp28/T07N6mTZsODz30UHhGRkalo0plWd9PO3bscKvLvp4LJJATQgghhBCnbcmSJXvXrl0bt2bNmriZM2ce+vfff70GDRrUNjk52dnRfTsbvPbaa0FFRUUUFBSoJUuWBDq6PzU1duzY5hMnTmwRFRWV+9prr+17//33dw8bNuzkwoULI/r06dPuxIkT1Yo7Ro8enbp27dq4Zs2a5dd1n892Z2W0L4QQQgghziy9evXK6tSpUy7AoEGDMps0aZJ31VVXtVm3bp33Nddck+bo/p3pVq5cGdy6devsjIwM5+XLlwdNnz79mKP7VF0vvPBC0PLly4PnzJkT/8gjjxT3e9iwYenDhw9PvfTSS9tNmDCh2Ycffri/smPk5uYqV1dX3bhx44LGjRsX1EvHz3IyIieEEEIIIeyuUaNGhQD5+fnF0+r++ecf9xEjRkRHRkZ29vDwiGnSpEnn6667rtnx48erHLWbOnVq4w4dOrT39fXtFhAQ0PX8889vs27dOm/bNp9//rmvUqrHu+++2+iGG25oFhAQ0DUgIKDrlVdeGZ2UlFTqHPn5+UyfPj28ZcuWHd3d3WMCAgK6XnTRRa3/+OMPD2ubxMREl+uuu65ZaGhoFzc3t5jo6OiO8+bNCy7bt08//dS3Q4cO7d3d3WOaNm3a6dlnny3X5lTWrl3rfeDAAfcxY8Ykjx49Ovnff//1+v333z3Ktjt8+LDL9ddf3yw8PLyLm5tbTHh4eJcRI0ZEZ2dnF7/Gmzdv9rz00ktb+vv7d/Pw8IiJiorq9PDDD4fXpD819fzzz4e3atUqp6Lgs1+/flnXXHNN0ieffBK0f/9+V4AdO3a4KaV6PPXUUyETJ05sEhoa2sXT0zMmOTnZuaKplenp6U7XXXddM39//27e3t7dL7300pbffvutt1KqxwsvvBBUl8/tTCYjckIIIYQQ4rQVFhaSn59PQUGBio2NdZ8+fXqTwMDAgiFDhqRb28THx7tGRkbmjR49Oj4oKKhg165d7vPnz4+49NJLvf7888+4Ux3/8OHDrnfeeefRZs2a5WdkZDi9++67QYMHD267adOm2N69e2fbtn3ggQeaXXLJJSlvvPHGvtjYWI/HHnusyYQJE/SqVav2W9sMGzasxdq1awNuvvnmo4MGDUrLzs522rBhg09CQoJr9+7dc06cOOHUp0+fdrm5uerBBx883LJly9w1a9Y0evDBB5vn5uY6WYOWrVu3elx99dWtO3XqlPX666/vzcnJUXPnzm2clZXl7OzsXK3spEuWLAlycnLilltuOZGWlua0cOHCiDfeeCOoZ8+eh6xtjh8/7tynT592qampLlOnTj3crVu37CNHjriuXr3aPycnR3l6eurvv//ea+jQoW2bNm2a+/jjj8c3a9Ysb8eOHe5//fWXVzV/jDW2f/9+13379nlMmjTpiJNTxWNEI0aMSHn33XdDvv76a9/bb7/9hHX7/PnzI7p06ZK5YMGCA4WFhcrT07Ooosdff/31zdesWRNw3333HT7vvPOyvv32W9/x48e3qKOndNaQQE4IIYQQwsFuvpmm//xDnX3Yro5OnchasoT42j6+W7dunWzvh4SE5K9atWpXYGBg8YfzIUOGZAwZMiTDev+SSy7JaNu2be7gwYPb/vjjj559+/YtFZDZWrly5QHr7YKCAkaPHp3aunXrTosWLQru3bt3qX737t07fdmyZdZtaTt27PBYuXJlcFFR0X4nJyc+++wz36+//jrgsccei58xY0bxKNK4ceNSrLfnzp0blpiY6Pb777//27lz51yAESNGpKempjrPmzcv4oEHHjjm6urK7NmzI7y9vQvXr1+/08/PrwhgwIABGe3bt+8cGhpa5Tqv7Oxs9fnnnwf26dMnLSoqKh+ga9eumR999FHQiy++eMjZ2QwkPvHEE2EJCQnuGzdu3G77OtkGRvfff3/TRo0aFW7ZsiXO19fX+rqnU4f27t3rBhAVFZVbWZuWLVvmgQnkbbcHBwfnf/PNN3sqCwABtm3b5r569erA//73vwmPP/74UYCrrroqLSsry2nZsmWhdnkSZymZWimEEEIIIU7bW2+9tWfDhg2x69evj3377bf3tGrVKnvEiBGtt27dWjxFMCcnRz300EPh0dHRHT08PGLc3Nx6DB48uC3Av//+W24qoa1PPvnEt3fv3m38/f27ubq69nBzc+tx4MAB9927d5d73NChQ1Nt73fu3DkrLy9PJSQkuAB89dVXfkoppkyZcryy861bt65Rly5dMtu1a5ebn5+P9XLZZZelpaSkuGzdutUTYOvWrT79+/dPtQZxAK1atcqPiYnJqOzYtt59913/9PR05+uuuy7Zum3s2LHJx48fd/3000/9rNu+//57v06dOmVWFuymp6c7/fHHHz4jR45MtgniaqWoqAjb55yfX3k8qnXVg46VtRk6dGjKqYI4gB9++MFHa831119/0nb7mDFjTlb2mIZCRuSEEEIIIRzsdEbCzhTdu3fPtiY7AbJGjBiR2rRp0y6PPPJI4y+++GIvwN133x25dOnS0KlTpyZeeOGFGY0aNSo8cOCA24033tgyJyen0k/0mzZt8rr66qtbX3TRRWkLFy7cHxkZme/i4qInTJgQlZubW+5xwcHBpZJluLu7a4Ds7GwngBMnTrg0atSowMfHp9IoJDk52eXgwYPubm5uPSraf+zYMReApKQk19DQ0HLJOUJCQgoSEhKqLL/wzjvvBHl4eBQNGTIk3bqOb/jw4anTp0/XS5cuDRo5cmQaQEpKikv79u2zKjtOUlKSc1FREU2aNMmr6pxV+fLLL32HDRvWxnab1npLRW1btGiRB7B///5Kn6t11K5p06alIsKIiIgqRywTExNdAcomQGncuHGDz2opgZwQQgghhLA7Hx8f3aRJk7zY2FhP67ZPP/00cOTIkcnPPPNMonXb6tWrq0x0snLlygAXFxe9Zs2aPdagDCAtLc3Zz8+vsKZ9CwoKKkhNTXXJyMhQlQVz/v7+BUFBQQXPP//8wYr2d+7cOQfM9EBrUGfr+PHjVX7Ojo+Pd9m0aZNfYWGhio6O7lJ2/7fffut/8uRJp4CAgKKAgICCI0eOVFpbLTg4uNDJyYlDhw6ddv21vn37Zm7YsCG2Om2jo6Pzo6Kicr755hv/oqKiQxWNsH3yySf+zs7OXHbZZaWmeSpVdXk5a7B3+PBhFz8/v+Ig9fDhw66VP6phkKmVQgghhBDC7tLT053i4+PdAwMDi0dScnJynFxcXEoFTkuWLKky62BWVpaTk5MTTk5OxY/97LPPfBMTE2sVtAwePDhNa82CBQtCKmszcODAtL1793q0bNky7+KLL84qewkICCgCiImJyVi/fn2jtLS04s/Vu3fvdt26datPVf14/fXXgwoLC9XTTz99cPXq1TttL3PmzInPyclxWrZsWQDAgAED0v7++2/vzZs3e1Z0LF9f36KYmJiMjz76KKgmBbgrEhAQUFT2+Z6q/eTJk4/u3r3b48knnyy3Zm3Dhg1e77//fvDw4cOTrWsAa+Kiiy7KUErxzjvvBNhuX7FiRUBlj2koZEROCCGEEEKctt9++83r6NGjLlprDh065PrKK6+EpqamOk+aNKk4mUi/fv1SV61aFfTUU09lt2nTJvfDDz/037JlS5UBz9ChQ1OXLFkSOnr06Oibb745KS4uzuN///tfRHWSiVRk2LBh6ZdddtnJWbNmNYmPj3e75JJL0vLy8tSGDRt8hw0blnrFFVekT58+/egnn3wS0Ldv33aTJk062r59+5yMjAyn2NhYjx9//NFn3bp1ewBmz56d2Lt374D+/fu3mTp16pHc3Fz15JNPNg4KCqqyFtqKFSuCIiMj86ZNm3a87EjW4MGD0xcuXBj+7rvvBk+ZMiV5+vTpR1etWhU4dOjQNvfee29i165ds48dO+by2Wef+S9btuxAQEBA0bx58+IHDx7ctmfPnu3vvvvuI82aNcvfvXu3+59//ulpTf4ybdq0iOeff77x9u3b/27Tps1pT8MEmDp1atLmzZu9Z86c2fSvv/7yHDVqVIqXl1fRhg0bfBYtWhTeunXr7MWLF9dq+nDXrl1zhw0bduKZZ56JLCoqUr169cpau3at79q1a/0BbF+3F198MWjKlClRn3766Y7LL7+8WmsUz2YSyAkhhBBCiNN28803F6eDDwgIKGjTpk32hx9+uGvUqFHFxcAXL14cf9ttt6knn3wyEqB///6pb7/99t7+/fu3P9WxR40alfb444/Hv/zyy2Fff/11QKtWrbIXL16878knn2xc2/6uXr1674wZM8JXrlwZ/MYbb4T6+PgUdunSJXPSpEnHAYKCggp/+eWXuIceeqjxggULwo8dO+bq6+tbGB0dnXPllVcWJ9qIiYnJ+eCDD3Y9/PDDTW+55ZYWoaGh+ffcc0/izz//7PPzzz/7Vnb+H3/80XPnzp2e06ZNO1zRdEQXFxfGjBmT/NJLL4XHxcW5tWvXLu+nn36KmzZtWuQLL7wQnpKS4hIUFFRwwQUXpHl4eGgwNdvWrl0b98gjj0Q+9NBDzfLz850iIiJyx44dW5xIpaioSBUWFlYrSUlNvP/++wcWLVqU/vrrr4fccsstLQoKClTTpk1zJ02adOSRRx45ejoJWN55550DEydOLHzppZfC8/Pz1fnnn5/+v//97+D//d//tfL39y+eWltUVITluZ3WiOTZQtn7hyiEEEIIISq3bdu2/V27dk1ydD+EOJs98sgjYU888USTHTt2/N26dWu7jCyeibZt2xbctWvXqIr2yYicEEIIIYQQ4oy1fPnyRn///bdnTExMlpOTExs2bPB59dVXw4cOHXryXA7iqiKBnBBCCCGEEOKM5efnV/jFF1/4L1y4MDw7O9spLCws/6abbjr6v//977Cj++ZIEsgJIYQQQgghzliXX355xuWXXx7n6H6caaT8gBBCCCGEEEKcZSSQE0IIIYQQQoizjARyQgghhBBCCHGWkUBOCCGEEEIIIc4yEsgJIYQQQgghxFlGAjkhhBBCCCGEOMtIICeEEEIIIYQQZxkJ5IQQQgghRK298MILQUqpHtaLs7Nzj9DQ0C5Dhw5tsW3bNvfaHHPUqFFRkZGRne3d1/pU0+ewe/duV2dn5x6urq4xiYmJZ1Wt53vvvbex7XvA19e3W+fOndsvWrQo0N7neuGFF4Kef/75IHsf92x0Vr1JhBBCCCHEmWnJkiV7mzVrlldYWMjOnTs9nn322YhBgwa13b59+79BQUGFju7fme61114LKioqoqioSC1ZsiRw+vTpxxzdp5r6+uuv45ydnUlKSnJ5/fXXg++4447onJwcNWXKlGR7neOdd94JLiwsxJ7HPFvJiJwQQgghhDhtvXr1yho4cGDmoEGDMu+6667kBQsWHDh27JjrunXrvB3dt7PBypUrg1u3bp0dERGRt3z58rNyxGnAgAGZAwcOzBwzZkzql19+uadZs2a5L7/8cpij+1WV3NxcVVRU5Ohu1JgEckIIIYQQwu4aNWpUCJCfn6+s2/755x/3ESNGREdGRnb28PCIadKkSefrrruu2fHjx52rOt7UqVMbd+jQob2vr2+3gICArueff36bskHi559/7quU6vHuu+82uuGGG5oFBAR0DQgI6HrllVdGJyUllTpHfn4+06dPD2/ZsmVHd3f3mICAgK4XXXRR6z/++MPD2iYxMdHluuuuaxYaGtrFzc0tJjo6uuO8efOCy/bt008/9e3QoUN7d3f3mKZNm3Z69tlny7U5lbVr13ofOHDAfcyYMcmjR49O/vfff71+//13j7LtDh8+7HL99dc3Cw8P7+Lm5hYTHh7eZcSIEdHZ2dnFr/HmzZs9L7300pb+/v7dPDw8YqKiojo9/PDD4TXpjz24urrSsWPHrPj4+OLptS+//HJg27ZtO1hf7xEjRkQfOHDA1fZxixYtCmzfvn0HLy+v7r6+vt3atGnTwfp6nnfeeW1/++03n61bt/pYp3Ged955baFkiu+aNWt8LrnkkpZeXl7d/f39u40bN65ZRkZG8euzY8cON6VUj6eeeipk4sSJTUJDQ7t4enrGJCcnOxcVFfHoo4+GRkVFdXJ1dY0JCQnpcsMNNzQ7ceJEqZipOu+d+iBTK4UQQgghxGkrLCwkPz+fgoICFRsb6z59+vQmgYGBBUOGDEm3tomPj3eNjIzMGz16dHxQUFDBrl273OfPnx9x6aWXev35559xpzr+4cOHXe+8886jzZo1y8/IyHB69913gwYPHtx206ZNsb179862bfvAAw80u+SSS1LeeOONfbGxsR6PPfZYkwkTJuhVq1btt7YZNmxYi7Vr1wbcfPPNRwcNGpSWnZ3ttGHDBp+EhATX7t2755w4ccKpT58+7XJzc9WDDz54uGXLlrlr1qxp9OCDDzbPzc11sk593Lp1q8fVV1/dulOnTlmvv/763pycHDV37tzGWVlZzs7Ozro6r92SJUuCnJycuOWWW06kpaU5LVy4MOKNN94I6tmz5yFrm+PHjzv36dOnXWpqqsvUqVMPd+vWLfvIkSOuq1ev9s/JyVGenp76+++/9xo6dGjbpk2b5j7++OPxzZo1y9uxY4f7X3/95VXNH6NdHTx40N3X17cQYN68ecH3339/88svv/zknDlzDh06dMj18ccfj+zXr1/bbdu2bW/UqFHR119/7TNp0qTo8ePHH5s7d258UVGR2r59u0dKSooLwCuvvHJg3Lhx0YWFherll18+AODv719q2u7NN98cPWzYsJOTJk3a8/PPP3s/99xzEVlZWU4fffTRftt28+fPj+jSpUvmggULDhQWFipPT8+iyZMnR7700kvh48aNO37llVem/PPPPx5PP/105L///uv566+/7nB2Nt8FVPXeqY/XFiSQE+KsoJTqC9wN9AVCgRxgB/Ax8KLWOv0UDy97rChgn+XuMq31+Fr057SPUcPzjQfetNy9SWu9tC7PVxNlXosNWuv+VbRfD/Qrs7kASATWAnO01vvt2snS5+8GjLDc/URr/WddnUuUUEr1B7633K3z35lK+rAfaA4c0FpHWbaNpwa/W0op64fS4mPUsA/dOMX7r6I+ninKvFaPaq1n1/IYUT/88IM/kFR2/3nn0bbstpEjOfHQQxxPT8dp4EBal91//fUkTZ5McmIiLldeScuy+2+7jWO33cbJ3btxHTuWFmX3T5nCkbFjSd22DfeuXcmt6XOy1a1bt06290NCQvJXrVq1KzAwsHjO2pAhQzKGDBmSYb1/ySWXZLRt2zZ38ODBbX/88UfPvn37lgrIbK1cufKA9XZBQQGjR49Obd26dadFixYF9+7dO962be/evdOXLVtm3Za2Y8cOj5UrVwYXFRXtd3Jy4rPPPvP9+uuvAx577LH4GTNmFK9FGzduXIr19ty5c8MSExPdfv/99387d+6cCzBixIj01NRU53nz5kU88MADx1xdXZk9e3aEt7d34fr163f6+fkVAQwYMCCjffv2nUNDQ/Oret2ys7PV559/HtinT5+0qKiofICuXbtmfvTRR0EvvvjiIWvw8MQTT4QlJCS4b9y4cbvt63T77befsN6+//77mzZq1Khwy5Ytcb6+vtbXvdqfEU5XQUGBAvSxY8dc5s+fH/Lvv/963XTTTccKCgqYO3du5HnnnZf++eef77W279ixY87gwYPbLly4MHjGjBnHfvzxR29fX9/CJUuWFP88R44cmWa93aNHjxwfH5+iwsJCBg4cmFlRHwYMGJC6ePHiBOtjlVJ63rx5kX/99Vdily5dit/jwcHB+d98880eJycz2Hb06FHn1157LWzkyJHJb7311kGAUaNGpYWEhBTceeed0StWrGh03XXXpVbnvVNfZGqlEGc4pdRsYBMwBmgCuAF+QC/gSeAPpVS5f97irOICNAVuAn5XSpX7sGVH3YBZlku3OjyPEBXpRsN+/40HZmVkZDRydEfqwltvvbVnw4YNsevXr499++2397Rq1Sp7xIgRrbdu3Vo83SwnJ0c99NBD4dHR0R09PDxi3NzcegwePLgtwL///nvKaWmffPKJb+/evdv4+/t3c3V17eHm5tbjwIED7rt37y73uKFDh6ba3u/cuXNWXl6eSkhIcAH46quv/JRSTJky5Xhl51u3bl2jLl26ZLZr1y43Pz8f6+Wyyy5LS0lJcdm6dasnwNatW3369++fag3iAFq1apUfExOTUdmxbb377rv+6enpztddd11x8o6xY8cmHz9+3PXTTz/1s277/vvv/Tp16pRZWbCbnp7u9Mcff/iMHDky2SaIq5WioiJsn3N+fpXxKABeXl4xbm5uPZo0adJ14cKFETfffPOxF198MWHbtm0eJ06ccBkzZswJ2/aXXXZZRuPGjfN++OEHX4DevXtnpqWlOV955ZXRy5cvb1R2Omx1XHvttSdt799www0ni4qK2LRpU6lpuEOHDk2xBnEA69ev98nLy1Pjxo0rlUTltttuO+Hs7KzXr1/vC9V779QXGZET4gymlLoa84EHIA3zQf9LIAz4HzAKaAmsUkrFaK1PmRVMKeVhGe1Rp2pXFXscQwAwQGu9XinVDFgJnA8EATOAmx3as9OglHIBdFXvR+F4lhG4pQ7uRrEzbRSuvv36Kzsq2+frS9Gp9kdEUHCq/a1akX+q/ac7GgfQvXv37E6dOlmPkzVixIjUpk2bdnnkkUcaf/HFF3sB7r777silS5eGTp06NfHCCy/MaNSoUeGBAwfcbrzxxpY5OTmVDjBs2rTJ6+qrr2590UUXpS1cuHB/ZGRkvouLi54wYUJUbm5uuccFBwcX2N53d3fXANnZ2U4AJ06ccGnUqFGBj49PpVMfk5OTXQ4ePOju5ubWo6L9x44dcwFISkpyDQ0NLSi7PyQkpCAhIaHK8gvvvPNOkIeHR9GQIUPSrYHL8OHDU6dPn66XLl0aZB2RSklJcWnfvn1WZcdJSkpyLioqokmTJnlVnbMqX375pe+wYcPa2G7TWm+p6nHr1q2Lc3Fx0UFBQYWtWrXKs77uSUlJLgCNGzcuFxEGBwfnp6SkOANcfvnlGUuWLNn78ssvh44bN64lwHnnnZfx3HPPxZedPluZsudo0qRJPsChQ4fcbLdHRESUapecnOxs297K1dUVf3//wpMnT7pA9d479UVG5IQ4s820uT1Da71Ka52jtT4AjAWOWPZ1Aa4EM9VPKaUtl6VKqQlKqTilVD5wbdn9tidTSt2olNqplMpRSv2hlBqslFpv0z6qonPYPH6pzfYLlFLvKKVOKqWSlVIfKqXCbdr6KKWWKaX+tuzPV0qlKKU2KqXG1PYFU0o1U0qtsDznk5bjJimlvlZKXVqm7Wyb/l6plHpJKXVUKZWqlPpGKdW6TPtGSqnXLcdNU0otxwTVp0VrfRB41mbTeTbn9FZKPaqU+lcpla2UyrL8bO61BEy2/btNKfW7UuqEUipXKXVIKfWtUupGy/71lEwNA3jT5vmPt7TZb7m/Xyl1vlJqk+WciUqpp5VSrjbnG2/z+IlKqf8ppQ4DeZgRRut75XWl1EGlVJ7lZ7xOKTW87OuglApXSj1n8x5MU0pttfbfpt1Yy/sk1fI8dyqlnlBKeZVpN1Ip9YNS6ril3RHL4+6zaRNs+bnvtTzPNKXUDqXUcqVUW5t2TkqpO5VSvymlMiw/i7+VUtMq+Dk0trzfMyzv7VcA3wp/+OVfA1el1DHLaxpbZl87m9f7Lcu2MZafcbyl/7mW57JIKVXle7PMz3C8zXY3pdSzlt+HTKXUl0qpVpUco5tSapVSarfl9cu3vNarlFI9bdqtpwbvvzLnqNb7qMzzmWB5X8RbfhY/KaXOoxosr/U3lp9zolLqCcC1krZ3KqU2KKUOW9632cr8/XlaKeVr03+NzbTq33//vYf1ApCTk+O2a9euFn/99VfHrVu3dtuyZUvMH3/80TUuLq71yZMn/So695nOx8dHN2nSJC82NtbTuu3TTz8NHDlyZPIzzzyTOHz48PR+/fplBQQEVPmlz8qVKwNcXFz0mjVr9owbNy7lP//5T+bFF1+clZaWVuMRG4CgoKCC1NRUF9skGGX5+/sXdO/ePXPDhg2xFV0uvPDCTDCBiDWos3X8+PEqB0zi4+NdNm3a5JeTk+MUHR3dJSQkpFtISEi39u3bdy4oKFDffvut/8mTJ50AAgICCo4cOeJW2bGCg4MLnZycygUstdG3b99yz7s6j7vooosyL7744qyOHTvmWoM4S98KABITE8v9HiUlJbkGBAQUB8I33XTTyd9++21HcnLyn2+//faeY8eOuQ4bNqx1YWH1vhs8fPhwqXMkJCS4AkRGRpYKcJUq/aO3lsg4dOhQqcfn5+eTkpLiHBgYWGBpV+V7p75IICfEGUopFQHYrjd4y3a/1joPWGGzaVAFh7kCeBVoSxUj8EqpcZhv5lsD7phpT6uB9jXrebEvgOsAfyAQM3r4js1+H+AGzHMMtPSvEXARsKLsh/caaIyZhtrWcm4XzCjXIOArpdSASh63FJiEWYPoB1wKrLZ+SFfmL/4nwC2W4/oC12LWKdpDub/HSilvYCMmoO8AeACemJ/N/4DPlFJOlrZXA4uBHkAAZgpuY+ASYFgt+hMCfIdZl+kJhAMPAG9U0v5x4F4gAstorVKqA7AV85o1xXwQbgT8B/hUKfWwzXNtBfwJTKHkPegLdAcG2LRbCLyLeZ/4WZ5na+C/wAallKelXW/gA+BCINjSLszyuP+z6fcyzM892vI8fYE2mJ9ta8uxnIBVwItAT8Ab87PohAnAP7a8P7Ccfx3m/e6NeW9PxPweVklrnU/J70k7pVR3m91jbW4vsVwPwPyMm1j672Z5LrdbXo/afqBbDEzD/D54AUMw78WKtAOuwswO8MX8zoVZtm1QStX2bwhQs/dRGU9j3hdNMD+LPsAapdQppzUqpUKBDZi/AR6Y9/5/gUcrecgw4GLMe9/d8pi2mN+XT6v1JIG8vDzX1NTUgLy8PI+ioiJnrbUqLCx0ycjI8NuzZ0/rlJSUan0ZcCZJT093io+Pd7d+AAbIyclxcnFxKTWSsWTJkipT7WdlZTk5OTnh5ORU/NjPPvvMNzExsVbv8cGDB6dprVmwYEFIZW0GDhyYtnfvXo+WLVvmXXzxxVllLwEBAUUAMTExGevXr2+UlpZW/Hd89+7drlu3bvWpqh+vv/56UGFhoXr66acPrl69eqftZc6cOfE5OTlOy5YtCwAYMGBA2t9//+29efNmz4qO5evrWxQTE5Px0UcfBZ1ukBEQEFBU9vmezvG6du2aExQUVPDBBx8E2G7/9ttvvQ8fPux20UUXlZuG2qhRo6L/+7//Sx0/fvzx48ePux49etQFwM3NrehUo7crVqwodY633norwMnJCWvgXZn+/ftnuLm56eXLl5cqYv76668HFhYWqv79+6dD9d479UUCOSHOXM1sbp/UWqdW0Gafze1mFewPAp7CfJANBb6p6ESWD6pP2my6BfNB6WHL42pjH+aDXRvAuhh4oCVABbP4egwQhfmg6AFcAFj/WUyt5XkPYEYnm1iO6U1JIOME3FPJ41IxAVJjwPrNY1vMWkQwgWB/y+3dmA+vEcCuWvazmDJTK++z2fSL5XoKEGO5/bXlfC0wH2rBfLi+1nL7Yst1hqXf7pikEdcAXwFYErHcZHOem7TWynJZWqZbXpjRkwDMCKH1ZzhOKdW1gqfhgwmQfIBWlvYLLI8HeALznroYSLFsm2N57gAvUDK6+bHlGL6W9t8BKKXOB+6ytFmK+YDtBdxv2dYTuMNy+0JK/sf1wQQ4TTDvheU2/ba+bqss/fPDjHDfByRY9l2DZcQbmIsJzvyA5y3brsAELWC+nGhnuf2z5ZytbJ5zdSyxuW0bvFkD0L2YQAPgPaA35nfcFfMaWke92gJDa3BewIxGAdYvUpIwP/8gKg/ktgKXURLI+FHyc/DCBJU1ff/Zqsn7qKz+lr6vt9wPpOrXZColf/c+wby2MVQ+nfx5zN+OQMzPoAmW3zlggFKqm9Z6v9ZaUfJzo2fPnlusFwB3d/e86Ojo3Z07d/4rJiZma/fu3f9o0aLFbmv7o0eP1vZvcb357bffvNatW+e9du1a72XLlvlfeumlrVJTU50nTZpUnBCiX79+qatWrQp66qmnQlatWuU3duzYZlu2bKky4Bk6dGhqVlaW0+jRo6M//fRT36effjrk1ltvja5OMpGKDBs2LP2yyy47OWvWrCYTJ05s8uGHH/q99957jW6//fYmn3/+uS/A9OnTjwYGBub37du33TPPPBOyevVq3+XLlzeaOXNm2MCBA4vXps+ePTsxIyPDuX///m3efvtt/9dffz1g0KBBbYKCgspNtyxrxYoVQZGRkXnTpk07fsUVV6TbXh5++OFjISEh+e+++26wtT9NmjTJHTp0aJvHHnss9LPPPvN9/fXXA4YPHx5tHbWbN29efGpqqnPPnj3bv/TSS4GrV6/2fe6554JvvPHGptZzTps2LcLFxaXHzp07T3vkrrpcXFx46KGHDm3evNnvyiuvjP7www/9nnvuueD/+7//a9W8efPcu+66KwlgypQpjceOHdt88eLFAWvWrPFZtGhR4OLFi0PbtWuX3bhx4wKAtm3b5uzatcvztddeC9i4caPXtm3bSk1f/f777xvdfvvtTT7++GO/Bx98MHz+/PmNr7rqqmTbRCcVCQsLK7ztttuOrly5Mvjmm29uumrVKr/HHnssdNq0ac1jYmIyxowZkwrVe+9YnnOPa665prn9X80SskZOiLNDZfOwVRVtdgD/1VoX71OW6ZFltMF8+ADYprVeYmn7P0zg06SCx1RlptZ6r+U4P2BGKMAEF4mYgC0YszasPeZDu+3zKZe9rZqSgc7AHMwH6LKFaCs77jyt9TZLf7+kZCSyObAZ8+2/1UKt9Q5L2zmYbJO18X3ZqR3ASUqC6stttj+stT5ic85PLNuHYj7MW4N6b8wauy2YgPSbSr4EqEo+cL/WOgv4TSn1OmZUAswI0LYy7d/SWltHiPdYRqasI2kngNla6wLgB2Wm407B/A8apJR6FzP6ASYQHae1tn5z+oPlAqVHFsdbLmUNAuZT+kuOhzAJg+KAX7TWn9vs24d5v/TBvG7bgb+B57XW1mQBtud92HKp6LyrKP0+mau1PgTFv0uvV/C4crTW/yilfsN8iXCtUuoBTJBqndr4ps3vdCJmxLYfJrAtO22pNr9HtqPWb2utf7M8h+mYL1/KOoL58ud5zBczZUcLavu7TE3eR5R/fd/QWm+wHOcjSr6IqeqDle3PcJbWOhlItvwOzKig/XHMWubzMSPZZT9btcWMNp+Sq6trQXZ2tmdiYmJkXl6ee1FRUakv2/Py8uq1PlVt3HzzzcWJmgICAgratGmT/eGHH+4aNWpUcdbBxYsXx992223qySefjATo379/6ttvv723f//+pxy5HTVqVNrjjz8e//LLL4d9/fXXAa1atcpevHjxvieffLJxbfu7evXqvTNmzAhfuXJl8BtvvBHq4+NT2KVLl8xJkyYdBzPV7pdffol76KGHGi9YsCD82LFjrr6+voXR0dE5V155ZXFCjZiYmJwPPvhg18MPP9z0lltuaREaGpp/zz33JP78888+P//8c6UjqT/++KPnzp07PadNm3bYNumGlYuLC2PGjEl+6aWXwuPi4tzatWuX99NPP8VNmzYt8oUXXghPSUlxCQoKKrjgggvSPDw8NEC/fv2y1q5dG/fII49EPvTQQ83y8/OdIiIicseOHVucwKOoqEgVFhZi89GgXkybNi3Jy8uraMGCBeFjx45t5eXlVdS/f//UBQsWJDRq1KgI4Pzzz8988cUXQ6dPn940NTXVJTAwsODiiy9OfeaZZw5bjzNr1qzE3bt3u0+ZMiUqKyvLqVevXhm//vpr8frPJUuW7Js3b17Ydddd19LV1VVfe+21Sa+88kp8RX0q64UXXjgUEhKS/+abb4a+/fbbIf7+/gUjR45MXrBgQYI1eyhU/d4BU46jsLCwTqdfSiAnxJnroM3tQKVUowo+kEfZ3K7oj9Q2Xb2/1LaFS4vPq7XWSqkEahfI2S6qt53OYP0w8iBmdKMytf3Q8gKWEYBKVDglhar7azv1J6GS27VVgPkwvA5TfsCamtl22obt++GAzW3rt/QvYz5IjgTGWS4AuUqp2Vrrp2rYp2RLEFfR+SsqdPtHmfuBgPW/3mHLh2+rsv23Tq0FOGgTxJVVnREJ68/pY8xrcgtmNM06olaolFqktbaO7N2GmV7ZlpKRPYADSqkRlvT4NTmvvd4nSzCBXBPMdFDriF8RluQklimCm6roX2Xv91Op6XN4HzMiZ88+WNXkfVRWVb/Tlan281cmw+wPnPo5Vuv579+/v+mJEycqnapVNrA7k0yePDl58uTJyVW3hIiIiALb9PNWZRNplK35BTB9+vRj1tptViNGjCiVwOWKK65IrygpR0V9dHV15emnnz7y9NNPHynb3iokJKTwjTfeiKfi/7G2/UgfMWLE9jKby5WYsNW3b9/sqhKILFy48NDChQuLa8lFRkYWLF++/MCpHtO3b9/s7777bndl++fPn394/vz5hyvbXxvVPeakSZNOTJo06URl+6+99trUa6+99pRfPjZr1qxgw4YNlT6/pk2b5q9bt25PZfvbtm2bV9nr7uTkxKxZs47NmjXrWEX7rarz3qlOcpjTdcb+URCiodNaJwL/2GwaZ7vfsvbF9tvxiqZNVivDE6X/2UTanENRuyAOzIiOVUXB5LU2t0cA7papR9X6MHAK1uPmYqZqumKmelWlqv7avkZNKrldUwMs08pctdZNtdbjbYI4MN/0WzWr5PYxAEsSnGswH3wvxGS9/AUz1e1JpZT151rdr2CDVOnkIbbnrOjDSdn32gnAujK9sVLKNiFB2f6fwASzAM3KnJcyba2us5mWV3zBkihGG3dipuT1xvz+rMEEBXcqpfpY2v2itW6HWVc2BDN6l4EZtXm6gvP2reS811Tw2pzO+2Q5Ja/pdTbH/0ZrbQ0oBlASwKwDIix9mVzDc5VV7eeglAqgJIg7CnTEvMZdKjl2TYcAavI+Kquq3+nK1ORnOJySQO0dINDyM5hfybEr7UdKSkogmFp9bdq0iYuJidnSrVu3sl+QCCFEMQnkhDizPWZz+3Gl1FVKKQ/LepC3MWtSwEwFq/ai+grsoOTb5hhlsgL6YtYJnU6gciq236ynAK5KqUco/W346Ry3yHJcb0pnhKyt721u36WUaqtMFs5H7HDsythOAXxCKRVmmRprm830CwCl1Cil1F2YQHwbJtGHdfqjbUBuGyh3UmUyLtpwBZ5SJlNnL+BWm31VTiXVWmdjWduGCS5nKaX8lCluP96yvQATmGQD31q2+QDLlFItlMnYeb5S6gbLPtvX43GlVF/L70NTpdQQpdR7mKAHpVQ/pdSDmGnDO4EPMVNkrZpZ2j2hlBqGCRa+w4wunbRtU+a8LyilulrO20KZzJifU7LWzvZ98pBSKlKZOo+2ayCrZBl9tybSuRmzdhNKr5+z/R3KATKVUh2Bu2tyrgrYPodxSqleloDtiQraFlASnBRgyqQEU/pvl63qvv+Amr2PTnWcGrJ9/o8qpYIsSWduraCt7c8gC8ixfElwQwVtweb5Z2RklBqpUyXF1rWzs3NhUVGR88GDB+vq768Q4hwggZwQZzCt9fuYtV5gFvivwnxLf4CSb+j3AFedTs0uy/TL/9psehfzgWwupUeF7Dmh3jbb43rMKMhkapYU4lTH9cSsd0qh9JqX2vqGkoQJrTHrrRIxIxB1ZQGlE5scwazpstY0WoNZYwhmTd9CzHNOt1wmWPYlAn9Zbv+BKQ8AJrjIVzalJWxkYD4opwC/UjLy87Z1LWE1TKEkKHoEk1BmEyWJKx6xlF4A87M/ark9GvO+zsBmfaLWejPwiqVNtOVY2Zhpn19ikoFYA4OmmEQ/f1r6kE3J71Km5bFgRrU/sxwjF5NIxJoU4GvL9UosATPmtf/Tcrw9wEeYtYzWdRBvYd4bYNbdJWCS49Qmu5k1aLM+p2RKf2HzIyW/n5djfmdtR/FrxbL+c5nlbjDm538CszaybNt0zGggmC8R4jE/xw6VHL667z9bU6j++8genqNkhG8EZoRuKxUvR1mDCaLB/L5lAT/Z9Lesn6034uLiOvz+++89YmNj2wL4+fmlAGitnWJjYzv++eef3TIyMs7KsgNCOMLkyZOTtdZbbGoZnvMkkBPiDKe1noVZI/M+cAgzXSgd+B2YDsRorSudC16D87yNySi3G/NBaxtm2pDtlKVK57XXwtOYpB6HMB+KN2A+sNcmMYetqcAiTL8zMaMp5T6A1pQl2B2B+XCdigkyVlGydsnuLGvFLsYEILGYQCMHE0hMA4bbJORYh0l6stvSt0JMALcC6GcZ2cCSfOMGTMB3qn92yZipexst5zyGGdm8pQb9347J9vcG5gN+Aea1+x4YYbtuT2u9G5P5bwEmE2iu5Xn8ic0IidZ6EnA95v2Sivl9SLC0eQDzwRpMspc3Ma9bquX1SMK8H/pbk5BgSgp8B1jr3+UA/2KSV9xvOWcRZo3dXZjpqhmW/h3AZCe8C0vAbXmdL8G8NzIxgfCbmLV4NfUdsN/m/ruWsiPW1+IkJsDfhAkgDgOzMQHs6ZqAKXFxHPP7uZaSUceyrscEuycxr/U7VJwUpSbvP9vHVPt9ZA9a62OYxChrKXnvz6eCRCeWv73DMQFqDuaLlkmY38WKvAQscnJyKvfFW/PmzeMDAwOPOzs7Fzg5ORX5+vqmtmnTZqcdnpIQ4hyl6jtjjRDizKSUCsR8i/6T1rrIsj7uRkzgojDZ/s53ZB9F3VOmEHNz4IDWOsqxvRHi3LRt27b9Xbt2PWUiDCGEANi2bVtw165doyraJ1krhRBWjTHZ13KVUscw05astX2s0x6FEEIIIcQZQKZWCiGsjmGmRx3BrOdxxUzTWwx01Vr/6sC+CSGEEEIIGzIiJ4QAiteFXFtlQ3FOk+mUQgghxNlBRuSEEEIIIYQQ4izT4EfknJyctKenZ9UNhRBCCCHsYNWqVRQWFjZ3dD+EsLeioiJ69uy5xdH9aCgafCDn6elJZmamo7shhBBCiAYiNjaW9u3bO7obdrN06VJuuummCvc1atSIlJSU+u1QBZRSzJo1i9mzZzu6K3a3f/9+oqOjefPNNxk/fjwA48ePZ/369ezfv7+4XVxcHJMmTeL3338nPT2djz/+mBEjRrBkyRLmzp3LgQMH8PLyOq2f15YtW4oq25eRkaGeeOKJsFWrVgUePHjQ3cXFRbdt2zb71ltvPT5x4sRqlzcaNWpU1M8//+x76NChv2vd0XNEgw/khBBCCCHE6fvggw9o0qRJqW0uLvJR0xEeeeQR7rnnnlLb7r33Xvbu3cv777+Pv78/bdu25fDhw0yYMIHrrruON998Ew8PjzrpT3JysnP//v3b7N271+P2228/2r9///Ts7GynVatW+U+aNCl648aNvu+9996B6hxrzpw5iSkpKUfrpKNnGfntEkIIIYQQp61bt260atXK0d0QQMuWLctti42N5eKLL2bw4MHF2/766y8KCwu58cYbufDCC+usPxMmTGi6Y8cOz2+//TauX79+WdbtY8aMSe3UqVP2zJkzm/bp0yfj7rvvTq7sGNnZ2crT01N37Ngxt846epaRZCdCCCGEEKJerFu3jpiYGDw8PGjZsiWvvvoq48ePJyoqqrjN+vXrUUqxfv36Uo9dunQpSqlS0wVXrFjBf/7zH0JCQvDx8aF79+4sW7as1v379ddfueSSS/Dx8cHb25uBAwfy66+lq++MHz+eJk2a8NNPP9GrVy88PDyIiopi4cKFFfZ348aNjBgxAh8fH4KCgrjzzjvJzs6uVf+ysrKYNGkSQUFB+Pj4MHz4cBISEsq1s31Nra/n/v37efvtt1FKoZRi/Pjx9O/fH4CBAwcWb7O3/fv3u3766adBY8aMSbIN4qymT59+rGXLljnPPfdcuHXbCy+8EKSU6rFmzRqfIUOGtPD19e3WvXv39mCmVkZGRna2Pcb27dvd+vXr18rT07N7YGBg19tuu63JvHnzgpVSPXbs2OFm9yd1hpAROSGEEEIIcdoKCwspKCgotc3JyQknJzNuEBsby9ChQ+nZsycrVqwgNzeX2bNnk5GRgbOzc63OuXfvXkaPHs1DDz2Ek5MTGzdu5NZbbyU7O5uJEyfW6Fh//fUX/fr1o0OHDsVB2FNPPUW/fv34+eef6dq1a3HbtLQ0xowZw4MPPkirVq1YsWIFkydPxtfXt1wwdP3113PNNdcwadIkfv31V+bMmUNmZiZLly6t8fO9/fbbWblyJbNmzaJXr158++23jB079pSPiYmJYfPmzQwfPpxevXrxyCOPABASEkKPHj2YPHkyL730EjExMYSEhNS4T1X56quvfAsLCxkxYkRKRfudnJwYNGhQyiuvvBJ+4MAB1+bNm+db9910000tRowYceL222/fU1BQoCp6fE5Ojrrsssva5ObmOj399NMHw8LCCt54443gL774IsDuT+YMI4GcEEIIIYSjTZkCf/7p2D506wbPP1/rh7dr167ctssvv5zPP/8cgMcffxxfX1+++eYbvL29Abjgggto2bIljRs3rtU5//vf/xbfLioqon///iQmJvLKK6/UOJCbM2cO7u7urFu3Dn9/fwAuvfRSoqKiePTRR1m1alVx2/T0dBYvXsy115ryq4MHD+bQoUPMmjWLG2+8EaVKYo6hQ4cyb948AAYNGoRSipkzZ/Lf//6XNm3aVLt/O3bs4L333uOJJ57goYceKj5eRkYGixYtqvRxfn5+nH/++bi5uRESEsL5559fvM+adKdDhw6ltttTfHy8G0CrVq0qnRIZFRWVB7B3795Sgdzll19+ctGiReWHHG28+OKLQQkJCe7fffdd7IABA7IArr766tT27dt3SExMtM+TOEPJ1EohhBBCCHHaPv74Y3777bdSl+dtAsPNmzczdOjQ4iAOoGnTpvTt27fW59y1axf/93//R2RkJK6urri6uvL666+zY8eOGh9r48aNXHHFFcVBHJggaPjw4WzYsKFUW2dnZ0aNGlVq27XXXsvBgwc5dOhQqe3XXHNNuXZFRUXlpmxW5ZdffqGoqKjC49W3oqIiCgoKii+FhYWVttVaV3m8ytqMHj36ZFWP/eWXX3wiIiLyrEEcmFG+YcOGVfnYs52MyAkhhBBCONppjISdKTp16nTKZCeJiYmEhYWV2x4WFsa+fftqfL6MjAwuvfRSvLy8eOqpp2jZsiVubm688sorLFmypMbHO3HiBBEREeW2h4eHc/Jk6ZggICAAV1fXUtusz+3QoUOlsneWfc627WrCOrpU2fHq05w5c3j00UeL7/fr16/cmkarpk2b5gHs3r3bvWvXrhWOyh04cMANIDo6Ot92e5MmTfIram/r6NGjroGBgQVlt4eFhVX52LOdBHJCCCGEEKLORUREcPRo+azxZbdZU+Dn5eWV2p6cXDqh4ebNmzlw4AA//PBDqYyLZdfpVVdgYCBHjhwpt/3IkSMEBgaW2nby5Eny8/NLBXPW5xEZGVmq7dGjR+nYsWOV7apiDTKPHj1KixYtyh2vPk2YMIErrrii+L6vr2+lbQcPHpzu5OTEJ5984j9q1Ki0svuLior45ptv/KOjo3OioqJKBV+2U1QrExYWlr979+5ydROOHj3qWlH7c4lMrRRCCCGEEHWuT58+fPnll2RmZhZvi4+P58cffyzVrnnz5gD8888/pbZ/+eWXpe5nZZmZdLbB1MmTJ/n0009r1b9+/frxxRdfkJ6eXrwtPT2d1atX069fv1JtCwsL+eijj0ptW7FiBc2aNSsXoL3//vvl2jk5OXHeeefVqH+9e/fGycmpwuPVt8aNG9OzZ8/iS9u2bSttGx0dnT9s2LATK1euDN6wYYNX2f1PPPFE6J49ezzuueee8lF0NfTu3TsjMTHR7fvvvy8+dlFREatXr5ZkJ0IIIYQQQlTlzz//JCkpqdz2nj174uLiwowZM/jggw8YNGgQ999/P3l5ecyaNavc1MCIiAj69evH3LlzCQ4OJjQ0lHfeeYc9e/aUanfBBRfg5+fHnXfeyaOPPkpmZiaPP/44wcHBpKam1rj/jzzyCJ9//jkDBw7kwQcfRCnF008/TVZWFjNnzizV1tfXlwceeICkpCRat27N8uXLWbt2bXG2S1tffvkl999/P4MGDeLXX3/l0Ucf5YYbbiiV6KRVq1Y0b96cdevWVdq/tm3bMnbsWGbOnElRUVFx1sqyAe7pGjhwIAcOHGD37t12O+brr79+sF+/fh5Dhw5tO3HixCP9+/fPyM7OVh999FHABx98EHzNNdck3XPPPZXWkDuVu+66K3nBggXh1157basZM2YcCg0NLXjjjTeC09LSnIHirKkA11xzTfNVq1YFFxQUbLHTU3MoGZETQgghhBCn7eqrr6ZPnz7lLikpKYDJkPjll1+SlZXFmDFjeOihh5gyZQoDBw4sd6x33nmH888/n8mTJzN+/HiaNWvGjBkzSrUJCQnh448/prCwkNGjR/Pwww9z6623cv3119eq/126dGH9+vX4+flx4403Mm7cOHx8fNiwYUOp0gNgkqCsWLGCZcuWceWVV/L999+zYMECbrzxxgqfy86dO7nqqqv43//+x2233cbLL79cqk1VCUOsXn31VW655RbmzZvHVVddRVxcHO+9916tnm9lKiojcbqCg4MLf/755x1333134urVqwOuvvrq1jfddFPL3bt3e7z44ov7Vq5ceaC2x/bw8NBfffXVrnbt2mXdf//9zSdOnBgVGRmZd/PNNx8HCAwMLH5hCwsLVXVe57OFqk4mmXOZt7e3th3iF0IIIYSoS7GxscVp34UpXr1+/fpShb7PZOPHj2ft2rUVFuK2tXTpUm666SZ27dp1yiQw55ItW7YU9ejR4w9H9wNgwIABrXbv3u0RHx//T9Wtz1zbtm0L7tq1a1RF+2RqpRBCCCGEEOKsNXv27DAfH5/Ctm3b5qalpTl98MEHgevXr2/09NNPH3R03+qSBHJCCCGEEEKIs5a7u3vRyy+/HJaYmOhWVFSkoqKicubPn39g6tSp5RdtnkNkaqVMrRRCiDNaSgpUMYNJiLNKYWEsbdrI1EpH8PCAamS0P31aQ2EhuDSsMZMzaWrluUKmVgohhDirHDkCn3wCH38M330Hdl53L4RDrVkD+ed8qeIzk5cXtGoFbm51fKKkJDh0CLp0ASfJLSjqhgRyQgghzgj79pnAbdUq+Okn84V2q1Zw333Qo0c9fYsuRD0ICYHoaEf3ouEpKDCj+9u3Q8uWcIoa1qcvK8ucsKCgHqJG0VBJICeEEMIhtIbYWBO4rVoFf1gm43TtCrNnw8iR0LGjBHDi3BMbC4GBju5Fw+TrC7t3w86d0LSpCarr5G9MXp65lkBO1CEJ5IQQQtQbrWHLlpLgbccOs/2CC2DePLjqKmjRwrF9FEKcuzw9oX17MwPg4EEzcNasWR3MfrTOnZV54aIOSSAnhBCiThUWwqZNJnD7+GOIjwdnZxgwAO65B668Eho3dnQvhRANhYuLmbZ9+DAkJkJ2tplqadeBM9sROSHqiARyQggh7C431yQpWbUKPv0Ujh8Hd3e47DJ47DG44goICnJ0L4UQDZVSEBlpRuj27zfTXVu2BB8fOxy8qKgkgJNATtQhCeSEEELYRWYmfPWVCd4+/xzS0syHoiuuMOvdhgyx04ckIYSwk8BAU5Jgzx4z1bt5cwgOPs2DWkfjQAI5UackH6oQQohaO3kS3n7brG0LDobRo+Hrr83155+bkbjly+HqqyWIE+JctXTpUpRSFV78/f0d3T0AlFLMnj27wn1eXmbdnK+vGZ07eNAMqtWabW2Jegjk9u/fj1KKpUuXFm8bP348UVFRpdrFxcXxn//8Bz8/P5RSfPLJJwAsWbKE1q1b4+bmVmc/r/POO6+tUqqH9RISEtLloosuav3dd9952/tc9957b+PPPvusLnOSnjFkRE4IIeqA1vDZZ3DihKN7UjfS0uDLL0tqvEVGwm23mYDuoosaXA1cIQTwwQcf0KRJk1LbXM6SPwYuLtC6tSlPcPSoSYLSsiW4utbiYNYROaUcVjDwkUce4Z577im17d5772Xv3r28//77+Pv707ZtWw4fPsyECRO47rrrePPNN/Hw8KizPrVp0yb75ZdfPgCwZ88et2effbbxkCFD2v7000/be/TokWOv8zz33HMRBQUFDB8+PN1exzxTnR2/XUIIcZb55x8YMcLRvahb1hpvI0dCz55S81aIhq5bt260atXK0d2oNaVMSQIvr9Lr5rxrOmZkDeQ8PBw2tbJly5bltsXGxnLxxRczePDg4m1//fUXhYWF3HjjjVx44YV12icfH5/CgQMHZgIMHDgws1+/fpnt2rXrvHDhwpClS5fG1+nJT1N2drby9PTUju5HWfJvVwgh6sA//5jrb781HwjOtcuhQ6YO01NPwXnnSRAnhKiedevWERMTg4eHBy1btuTVV18tNw1w/fr1KKVYv359qcdap3Du37+/eNuKFSv4z3/+Q0hICD4+PnTv3p1ly5bVun+//vorY8ZcQr9+PvTu7c2AAQP59ttfS7UZP348TZo04aeffqJXr154eHgQFRXFwoULTYP8fHB2Zunnn6PatWPjxo2MGDECHx8fgoKCuPPOO8nOzq5V/7Kyspg0aRJBQUH4+PgwfPhwEhISyrWzfU2tr+f+/ft5++23i6e9jh8/nv79+wMwcODA4m31pW3btnkBAQEF+/btcwfIzc1VkydPbhwZGdnZ1dU1JjIysvPkyZMb5+bmFlf6y8/P55577mnctGnTTu7u7jEBAQFde/To0fbrr7/2AVBK9QBYuHBhhHUa57333tsYYNSoUVFhYWFdvv32W+9OnTq1d3d3j4mMjOz8xBNPhNr264UXXghSSvVYs2aNz5AhQ1r4+vp26969e3uAEydOON1www3NQkNDu7i5ucVERUV1evTRR0OLyszFPXz4sMv111/fLDw8vIubm1tMeHh4lxEjRkRnZ2fbtWqhjMgJIUQdiIszwc2FF5ovZUXtXX45/PJL6W29esGaNeb2hRea19vWgAHwwQfmdrduZrqUrWHD4M03ze1WrSAlpfT+//s/sH4mi4goPzvqtttg7lyTnTMysnyfp0yBGTMgORnati2/f8YM0+bAAejRo/z+p56CW2+Ff/+Ffv3K73/xRbj2Wvj1Vxg6tPz+pUtNkpnvvoNrrinZ7ulpRlCnTTMjD0LYU2FhIQVlRqCcnJxwsnzTExsby9ChQ+nZsycrVqwgNzeX2bNnk5GRgbOzc63OuXfvXkaPHs1DDz2Ek5MTGzdu5NZbbyU7O5uJEyfW6Fh//fUX/fr1o0OHDixdupTCQsXjjz/FsGH9+PTTnxk0qGtx8fC0tDTGjBnDgw8+SKtWrVixYgWTJ0/G19eX8RdeaGoZWJ7T9ddfzzXXXMOkSZP49ddfmTNnDpmZmaXWtFXX7bffzsqVK5k1axa9evXi22+/ZezYsad8TExMDJs3b2b48OH06tWLRx55BICQkBB69OjB5MmTeemll4iJiSEkJKTGfaqt5ORk57S0NBc/P79CgNGjR0etWbMm8K677kq8+OKLM3788UfvBQsWROzbt8999erV+wBmzJgR/tprr4U9/PDDh2JiYrJSU1Odf/vtN++kpCRngLVr18Zdcskl7UaNGpV8xx13HAeIiooqzj6TmZnpPG7cuJaTJ09ObNOmTe7y5csDZ8yY0dTX17dw8uTJybb9u+mmm1qMGDHixO23376noKBAFRYWcumll7bevn271wMPPHC4a9eu2atXr240e/bspsePH3d98cUXDwEcP37cuU+fPu1SU1Ndpk6derhbt27ZR44ccV29erV/Tk6OXUf2JJATQog6EBcH0dESxNWE1vDHHybr5W+/mQyYSsHFF0NUFMUfoMC8tlZDhphgzVb79iW3hw0zSVls2bYfOdKsh7HVu3fJ7WuuMbXwbMXEmGsnJxNQldWli7l2d694f4cO5trbu+L9rVuba3//ivdbn39wcMX7rcuUwsNL7z9yBF59Fe6809xPSICAgFpMHRN1wzI6Uso118CkSeZNWlHUPn68uSQlmSxDZd1xB4wZYwo4jhtXfv9995lfkh07Kv7WoQbatWtXbtvll1/O559/DsDjjz+Or68v33zzDd6WN90FF1xAy5YtaVzLYpL//e9/i28XFRXRv39/EhMTeeWVV2ocyM2ZMwd3d3fWrVtXnPTjsssuJSoqiqeeepSoqFVYZyymp6ezePFirrX8gg0ePJhDhw4xa9YsbvzyS5SbW/FUhaFDhjBv3jwABg0ahFKKmTNn8t///pc2bdpUu387duzgvffe44knnuChhx4qPl5GRgaLFi2q9HF+fn6cf/75uLm5ERISwvnnn1+8r73lj2WHDh1Kba8r+ZZvxXbu3Ol+zz33NCksLOSaa645+dtvv3l8/vnngVOnTk2cP3/+YYCRI0emOTs78+yzzzb+5ZdfjvTu3Tv7119/9bnwwgvTHnnkkWPWY44dOzbVets6dbNx48Z51tu2MjMznebPn79/woQJJwFGjx6ddsEFF7g+9dRTje+6665kJ5vpJZdffvnJRYsWFX8NuHz58kZbt271WbBgwX5r0Ddy5Mi0rKwsp8WLF4dNnz79aERERMETTzwRlpCQ4L5x48btffv2LR56vf322+2+al4COSGEqANxcVDBZxpRgdhYeO01E8AdOGA++/TrZ4KvwEB48MFTP3769FPvf+yxU+9/5plT71+woPJ9rq5mdKwyPj6n3h8cfOr9kZGn3t+ixan3d+hQfn9qKjRqZG5PmWKm/44da0YZrQGqELXx8ccfl0t2YpsFcfPmzQwdOrQ4iANo2rQpffv2Zd++fbU6565du5g5cyYbN27kyJEjWKe4ubu71/hYGzdu5IorrijVZ39/P0aMGM5nn60mIwO2bzfL3pydnRk1alSpx1977bXceuutHEpIoEm7dsUjcteUCbCvvfZaZsyYwa+//lqjQO6XX36hqKiIa2yH2S3HO1UgVxeKioqwnU6olKpyVHXr1q0+bm5uxXMQAgICCp555pkD48aNS3nqqadCAG6++eZSo2K33HJL8rPPPtt47dq1vr17986OiYnJfPHFFyPuvvvuyCuuuCK1X79+mR4eHtUe4XJ2dubGG29Msd129dVXn7z33nub79u3z7Vly5bF8y9Gjx5d6ivADRs2+Do5OXHbbbeVCsjGjRt34v333w/+/vvvvceOHZv6/fff+3Xq1CnTNoirKxLICSGEnRUWmi+3Bw1ydE/OTPn5sH49tGljajbt2AEvvQSXXgozZ8Lw4Xao4yQqZQ3iAKZONaNxS5fCokUmkLv//opH+UQ9KLMmrBQvr1PvDw4+9f6mTU+9/zRH4wA6dep0ymQniYmJhIWFldseFhZWq0AuIyODSy+9FC8vL5566ilatmyJm5sbr7zyCkuWLKnx8U6cOEFERES57eHh4aSknKRtW1NvLjUV/P0DcC2T0tL63A4lJtKkc+fiEbmwMn/QitsdOlSj/iUmJpZ6fNnj1ac5c+bw6KOPFt/v169fuTWNZbVt2zZ78eLF+5VSNG7cOD86OjrfOgJ24sQJF4BmzZqVmsjetGnTfMt+Z4Ann3zyiIeHh/7ggw8CX3zxxXAvL6+iIUOGnFy4cGFCRERElZllfH19C9zd3UsFfuHh4fkABw4ccLMN5Jo0aVKqLydPnnT28/MrKDs1MjIyMh8gOTnZBSAlJcWlffv2ZeZ51A0J5IQQws4OHDBrp2RErkRWFnzzjRl1W73arEl77DGzVmzIEFNvzs/P0b1sePr2NZfnn4d33zUjo7GxZl9+PmzdapLZKLsuzxcNVUREBEePHi23vew2awr8PNvC2kBycqnBGjZv3syBAwf44YcfSmVcLLtOr7oCAwM5cuRIue1HjhwhMDAQHx8zyu3iAikpJ9m3L5+oKNfi3w/r84gMCSm1Ru7o4cN07N693PONrGiB7SlYg8yjR4/SokWLcserTxMmTOCKK64ovu/rW3XZNm9v78KLL744CyAxMTHk77//Di8oKHB1d3fP9vX1TQeIj4937dixY671MfHx8a4AgYGBBYcPHw5NSkoKGTlypPvVV19dkJube+zHH3/MnjlzZtNbb73V6YsvvthbVR/S09NdcnNzlW0wd+TIEVeA5s2bl3rDqTJ/+AICAgrT0tJccnJylO0o4KFDh1wBgoODCyztCo4cOeJW5QtiB5JnTAgh7MyaeKOhB3LWWTd5eWYw4Kqr4IsvTFmGzz4zS3PArCOTIM6xAgLgrrvgzz9NcA3mZ3X++dC1q0n8UnadoRA11adPH7788ksyM0uWLsXHx/Pjjz+Wate8eXMA/rGm/7X48ssvS93PsixutR0ZO3nyJJ9++mmt+tevXz+++OIL0tNLyo+lp6ezevVq+lmyDrm6mr9XhYWFLF/+Ebt2lVQYWLFiBc2aNiUyNNQEcpbRpvc//LDUeVasWIGTkxPnnXdejfrXu3dvnJyceP/998sdr741btyYnj17Fl/a1mBE9/jx4wGHDx9uGhYWltiuXbvt3t7eGe3atQsBeOuttwJs2y5ZsiQQoGvXrt5JSUmhkZGRhzp06PBPq1atdrVu3Trt3nvvTbrgggvSduzY4Wl9jKurq87Ozq4wxiksLGTZsmX+tts++OCDgIiIiLzo6OhTFv3r379/elFREW+++WapPr7zzjuBrq6uun///pkAAwYMSPv777+9N2/e7FnxkexHRuSEEMLOGnIgd+yYCdJWrYLMTNiwwXyeeewxM3Pr4otrWWBX1AulSn4+AweaxCivvQaTJ8MDD8DVV5ugznZ6phBWf/75J0lJSeW29+zZExcXF2bMmMEHH3zAoEGDuP/++8nLy2PWrFnlpgZGRETQr18/5s6dS3BwMKGhobzzzjvs2bOnVLsLLrgAPz8/7rzzTh599FEyMzN5/PHHCQ4OJjU1lZp65JFH+Pzzzxk4cCAPPvggSimefvppsrKymDlzZnE7pcwI1MsvP0BKShItWrTmxx+Xs3btWpa++KIZyXF1LR6R+/Lbb7n//vsZNGgQv/76K48++ig33HBDqfVxrVq1onnz5qxbt67S/rVt25axY8cyc+ZMioqKirNWlg1wT9fAgQM5cOAAu3fvtutxrY4dOxbm7++fHB4engQQHR0dn56e3mjIkCH6f//7X+OCggJ14YUXZv7444/ezz//fMTll1+eGhEREdiyZcvtw4cPj+zcubNXjx49sgIDAwu2bNkS+sMPP/iNHTu2+I3XsmXLnLVr1zb6+OOP04KCggqaNWuWHxUVlQ/g7e1dNGvWrCZJSUkubdu2zX3vvfcCN2/e7PfCCy/sd6qijs7VV1+d+uyzz2ZMmzat+fHjx106d+6c8/nnnzdauXJl8J133nnEOrVz+vTpR1etWhU4dOjQNvfee29i165ds48dO+by2Wef+S9btuxAQEBA0RdffOFz5ZVXtn3++ef333XXXcmnPPEpSCAnhBB2FhsLISEQFOTontSfTz+F+fNh0yYzEteiBYwaZW47OZmke+Ls4usLEyaYyx9/mIDu55/NdjBBeseOsp5RlLj66qsr3H78+HGCg4Np3749X375Jffffz9jxowhMjKSBx98kM2bN5dbX/XOO+9wxx13MHnyZDw8PLj55puZMWMGt912W3GbkJAQPv74Y+677z5Gjx5N48aNueeeezhx4kSp9VvV1aVLF9avX8/06dO58cYb0Vpz/vnns2HDBrp27VqqrZ+fH++/v4K77rqHf//9m8DAMObOXcCNI0eaQps2UyvfWbiQ/y1dyiuvvIKbmxu33XZbcRZLq4KCAgrLpsetwKuvvoqPjw/z5s0jLy+P//znP7z33nt2LeZdURkJeykqKlLZ2dneYWFhpeaD+vr6ps2dO9ezVatWJ5cvXx68YMGCiNDQ0PxJkyYdmTp1amF6erpHamqqX9euXX2//fbbRkuXLiUnJ4fw8PC8O+644+jcuXMTrcdasGDBgalTpza79tprW+Xl5SnbTJje3t6Fb7311t6pU6c227Vrl2dQUFD+Y489Fn/33XdXGUw5Ozvz7bff7poyZUqThQsXhqekpLhERkbmzZ49O942i2ZwcHDhTz/9FDdt2rTIF154ITwlJcUlKCio4IILLkizTsnUWqvCwkLK1p+rKaX1GVekvF41bdpUv/32247uhhDiHLJjh7m2pql2sXxlVma5B2CCnNPZ7+xc/FnhtPZrXb5WmrXvTk7l9xcWmgX/wcGmzfHj5uLvb6bpedb5hBLhSFrDX3+Z94G/v/niohpLZIRFo0aNTpkUpKGZOHEimzZtKjeV8kw1ceJE1q9fT5xl+oXWkJNjvrjyTzuKW0Yama1b8+4773DHpEn8u24dTXv1cnCv68fOnTu1u7t7cbbG4ODg49aRN4Dc3FzXv//+u0vr1q13NGrUKMO6PT4+PuLkyZNBXbp0Kfcm2Lt3b7OTJ08Ge3h4ZDVp0iQBICEhoQlAhw4d4squZavMqFGjojZt2uR39OjRv07jKda7bdu2BXft2jWqon0NfkTuxIkTxVXthRDCHq6+2mRgvPNOU/bps8/M9vBwKLsmfexYk2QCTKp6m6UjANx+u8kmqHXxcotS7rsP5s2DtLSKp7vNng2zZpkviMtkBQfgf/+De+8100Fta69ZLV5s0tL//rspwl3Whx+WHnkTDUdIiBmle/xxs36uZUtTquHyyx3dszNfbGxstZJDNBSurq4opc6a16Si/vr4mERX7mlHUU7ueHn54mH5RsvNyemseW52oDt16hRb28dWsl1prVV0dPQ+Ly+vXABXV9d927dv75Senu7t5+dXrl5cQ9HgAzkhhLCnpCRzcXc3oxW2ZYYWLDDf2tqyLWz96qsli+atbEsMLV1a/nwdO5prD4+K91tnA/n7V7y/Z09zHRFR8f4+fcx1VFTp/a6uZr2bNTiUIK7h6dTJvKefesqsiXztNfM+A9i5E/buhQsuKJ3x0vpZNien4hFg6/7s7PK/C0qZD8tgsqCWnYVW1X4np5LC55mZJcl4rJydTYZ/gIwM8+VJTfa7uJSMRNvkyijm6mp+T8Gcu6L+W0e/K5pt5ej9Tk6mTVX7i4rKvzZV7be9X5vHV2d/FSXOTpuTk/k7WZCeR1aeG/GxNu/xakyZbChcXV0LAPLz80utli4oKHB1cXGpcD6ni4tLvlJKW4M4AE9Pz1yllM7Ly3MDGmwg1+CnVnp7e+vMsl+BCyFELW3aBBddBOPGwdtvmw901g+XQjQU994Lzz1Xfrv1I8eECSbws+XjUxIAjR0Ly5eX3h8RAYcPm9vDhsHnn5fe37q1CSAB+vc3a/hsxcTAli3mdo8eprSCrX79SsqstWkDu3aV3n/FFaZ0BkDjxpCYWHr///0fvPeeue3ra4I9W7fdZka4Ab76Kpbg4NJD4GFhJrtrYaFZk1hW48bmkpdnprWW1aSJGfXPyYGKZig2b25GUTMzS0pM2IqONut609NLpofbatXKBOopKVBRDoy2bc3zTk6GikrCtW9vAunjx83IVVmdOplA98gRSEgov79LF7Ps7PDhkveBre7dTbAWH19+5gOUfGmVmWnOU2eB3bZt5Hk1YntmFEpBF48dKK0bTParLVu2FPXo0aOCd3CJf//9t52np2d2ixYtit8Jf/31V6dGjRqdbN68ebnieidPnvTbs2dP644dO/7j6emZC5Cdne3+77//dmrTpk2sn59fvdRsc5QzbmqlUmoScD8QAfwLTNFa/3CK9pcBs4FOQC7wI3C/1nqnZX9/4PsKHtpeax1nz74LIcSpWDNWnjxpvp2VIE40RE89BRdeWPEHeoDRo8vXn7bNZnrddSbYsmUdTQO4+WYTrNkKsEkIfscdJtizFRpacnvKFJNh1Zbt1OOHHipfbsGmbBczZ5afBm37Of2JJ8qPOHbuXLqvZac6W5+fUhVPg7b+LXF2rni/dTTTxaXi/dbju7lVvN862ujuXvF+62iih0fF+93dS45T0X43t5J+VLTfuhbY17fi/dbAy8+v4hkA1pFff//KM+MWFpoA3cXFBKbW52Q3RUWQn4+btxuNG8HBg6C9XVA52VU/tgEJDQ09evDgwegjR45k+vr6Zhw7diykoKDANSws7DjAgQMHIrOysrzbt2+/E8Df3z/Nw8Mja9++fVHNmjU7CHDw4MFmnp6emb6+vud0EFeVeh+RU0qNAd4BJgGbLNc3AR201gcraB8NxAIvAIsBH+AZoIXWupWlTX9MINcROGHz8ONa61OOZ8uInBDCnqZNg5deMt8eh4SUHzUQQojY2FjaV7QoVdS59HTYs8eMDkdHl0wHtovcXPj7b4iK4oRTMHv3QteAg7imn4Bu3ex4ojNXdUbkwBQEP3bsWHFB8KZNm8Zbk5/s3r07KjMz07dr165/W9vn5ua6HjhwoGlGRkYjpVSRr69vWrNmzeLd3NzqJr3mGeRMG5G7F1iqtbZOqrhbKTUYuAN4uIL2PQBX4GFrUKaUmgt8p5QK1lrbFiw5Vua+EELUq7g4My3r55/LT60SQggrrTXVzbYn7MfX10zz3LPHTBGNjDRTUu3yo7CmBnZzw9VyvCInF7PgU2s7neTcEBERcTwiIuJ4RftatWq1v+w2d3f3/DZt2uyt846dYSwDbpWOutXr8nSllBsmMPumzK5vgAsqedjvQD5wq1LKWSnlC9wI/FZB0Pa7UipRKbVOKTXAnn0XQojqiI01HxJMwVhH90YIcSZydXUlO1um2zmKu7uZ2hsYWD4B1WmxBnKursVTRQusYyZ1VJdNnNsyMzO9lFKVTFKv50AOCAacgbLLUI8C4RU9QGu9H7gUeBSzPi4V6AxcYdMsETOiNwoYCewA1imlLrZj34UQ4pRycsyaIHd3uOkms6hfCCHKCg0N5dChQ2RlZdHQk845irOzmVrZvLn54i0nxw5BnXVhpJtb8To9CeRETWmtycvLc0lKSvLfv3+/S35+/jOVtXVU+YGyf7VUBdvMDqXCgTeAt4DlgC8wB3hfKfUfrXWR1noHJniz2qyUigKmARsrOOYEYAKAm3X1rRBCnKZdu8zsmaws+Ogjs1ZOCCHK8vPzA+Dw4cPkV1SHQdS7I0dMHBYcXFJGosZOnDBz6nfuRGtTikZn5+CTmWTm3ds9u8qZJykpSW3bti3Y0f04yxUppU5qrTcWFBTM7dGjx67KGtZ3IJcEFFJ+9C2U8qN0VncCmVrrB6wblFLXA/GY6ZibKnncL8C1Fe3QWi/GJE7B29tbvgoTQtiFNWNlWpr5pteaBU4IIcry8/MrDuiE47m7w4gR8O+/8PTTcN99tVjSdtVV5hu9f/6BY8d4pu9mogdEMevjIfDhh6ULi56jOnTokK21jnJ0PxqKep1aqbXOA7ZgpkrauhT4qZKHeWGCP1vW+6fqfzfMlEshhKgXsbHmH398fEmhbiGEEGe+Fi1g82YTa91/vymBkVXTxPYJCaYYIMDQoSxNGcGRVMvwXpLk4hP2V99r5ADmA+OVUrcqpdorpRYAjYFFYDJSKqXW2bT/AohRSs1SSrVWSsUAb2JG5LZYHjNFKTXCsr+jJavlCODFenxeQogGLi4OmjUzmdA6dXJ0b4QQQtSEtzesXAlz55r1zhXVqzulhISSInh9+wKQkmxZGyeLpkUdqPdATmu9EpgCzAD+BC4EhmqtrdXdI4CWNu2/A8YCVwJ/AF9jslgO1lpbC8C5AfOAv4AfLMe8XGu9qo6fjhBCFIuLM1MqW7QwdeSEEEKcXZQyBeE3bjRL2lJSYFNli3hs5eXB0aMlgdzgwQAUJKWaFMYyIifqgEOSnWitXwZermTf+Aq2rQBWnOJ4z2CKhAshhEMUFcGOHTBhAnz/vaN7I4QQ4nRYs07OmAGLFsFzz8Fdd51i3dzhwybblXVq5Q6Tg885+Rg0DpFATtQJR0ytFEKIc058vFlP0a6do3sihBDCXp58Ei6/HCZPhptvPkWJgoQEc92kiQnopk41m3OCKAoMlqmVok5IICeEEHZgzVi5Zg3ccotj+yKEEMI+/Pzg449h1ixYuhQuvhgOHaqgYXy8uW7SBFJTAfh59Dx+5EJy/YJlRE7UCQnkhBDCDqyB3PbtkJzs2L4IIYSwHycnmD3bBHSpqZVMr7SOyDVtWnzbOaopoMn0kqmVom5IICeEEHYQFwf+/rB3r5QeEEKIc5G1zlzjxlBYCKtX2+xMSDDDd76+xYFcr3ljeJfrSHOTqZWibkggJ4QQdhAba0oPFBZKICeEEOcqF0uawGXLYPhwuP12k7CS+PiSjJWWaZb5jZsRzhFSnIMhO7sWhemEODUJ5IQQwg7i4iAw0NyWQE4IIc5tN94IDz8MixfDgAGQt690MXA++QS6diWMoxwjxGyX6ZXCziSQE0KI03TypCkfFB0NAwdC27aO7pEQQoi65OxsMlquXAl//gkn/krguLtlRC4yEq68EpdmkYRyjCP5wWa7TK8UduaQOnJCCHEusZQL4qqrYNgwx/ZFCCFE/bnmGmjXIo/QXkdIDLUEcmvWQEAAKiyUIJI5nO1vtsuInLAzCeSEEOI0xcaa61atHNsPIYQQ9a9LSCKgiTzfTK1Mm3AfPj3b43TH7awMuZtDGf6moQRyws5kaqUQQpymuDhwdYWYGFiwwNG9EUIIUa9sioFv3aJRCfGs+q0px7sP4q0eC4jLtKydk6mVws4kkBNCiNMUF2cyVubkQHi4o3sjhBCiXtkUA49plYYvGWw52oQLL4SIgBzikzzNojoZkRN2JoGcEEKcJslYKYQQDVgFxcAv/L8msHMHS5Z7csHRj9FBQRLICbuTQE4IIU5DXh7s2WOmVrq4QJs2ju6REEKIepWQYAqB+/kVj855tWnKcUvZAf/8YxQFSlFwYX+S7EQIIU7D7t2mCHhuLrRuDW5uju6REEKIemVbDPyii2DLFpq5tGVqoSeFc1wILTpGnl8wnjIiJ+xMAjkhhDgNcXHmesgQCAtzbF+EEEI4QIJNMXBvb4iJoSXwSBfIeSGEsBNHyfYOwfNorEO7Kc49EsgJIcRpsAZyDz4IPj6O7YsQQggHSEiATp3M7U8/hexs9JhrOXwYfAPCCD1xjDT3SAJlaqWwM1kjJ4QQpyEuDho3htRU0NrRvRFCCFGv8vMhMbFkauXLL8P8+YCpLfpZ4ztYyRhSnIMhORmKihzYWXGukUBOCCFOQ2ysyVjZpElJYXAhhBANRGKi+RbPOrUyPh6aNkUpM93+yyYTeJfrTeKToiJISXFod8W5RQI5IYSoJa3NiJy7u8lY2bq1o3skhBCiXtkUAy++b7kdHg5pR7Pp6LmXo4XBZr9MrxR2JIGcEELU0uHDkJFhShC0aWNKEAghhGhAbIqBk5YG6enFgVxYGAzcvpB/sltyLNPbtJPMlfVLqZYoNRal7rVcrkOpc6ZQkCQ7EUKIWrImOklKggsvdGxfhBBCOEAFxcBtA7l9WSad8ck0y9iJBHJ1T6kA4DbLpUUlbQ4CrwGvonVy/XXOviSQE0KIWrKuiTtyBDp2dGxfhBBCOEBCgklZ7OdnLkePgpcXANdfD3m+oTAf0lMtSU5kamXdUmoa8F+gEaBO0bI58BhwP0o9gdbz6qN79iaBnBBC1FJcHPj6miRlXbs6ujdCCCHqnbUYuLLEDKGhxbsuvhjwNoFcVkq+2SgjcnXtGct1FvA58AOwDUjCBHZBQFfgIuAKTMD3NCCBnBBCNCRxcdC+vfnWVQghRANkWwz8s8/gjz9g5kxQisxM2HU4lG6AU1oK2ssLJYFcXdsPzAXeQ+vMStpsAl5CKS/geuCBeuqb3UmyEyGEqKW4OAgOht9/d3RPhBBCOIRNlko++wwWLSoenfvjDzhveBjv93mOTUV9KAoMlqmVda81Wr92iiCuhNZZaL0YaFv33aobEsgJIUQtpKXBoUOwZw/cfLOjeyOEEKLeFRSULgZuG9Rhkp3k48b3XafwL53IbxQsUyvr3gco9REASs1EqZlVPkLrwrruVF2RqZVCCFELO3aY65Mn4T//cWxfhBBCOEBioinybVsMvG3J4E54uLn2P7GXtuSR7R2ChwRydW0EYMksw2xAA3Mc1Zm6JoGcEELUgrX0wLFjkrFSCCEaJNsacmBG5AYOLN7t4wOennDT9zdwCe6ke0QQEL/TAR1tUAoBJ5Rq5uiO1AeZWimEELUQFwfOzua2BHJCCNEA2daNy8yE3NxSUyuVMtMrT7iGEcZRUpxlamU9SLRc78OMxoFShZVcChzWSzuRQE4IIWohNhZCQsxtCeSEEKIBsi0G7u0N2dkwdWqpJi+9BNG9QwnlGMlOIZCebgI+UVe+wJQZqO7lrCZTK4UQohbi4qBHD5g2DVq0cHRvhBBC1Lv4eBPANWpk7isFrq6lmgwdCvwSStHHSRwrCDQbk5IgMrJ++9pw3AukAjHAJZhRuY0O7VEdkkBOCCFqKD8fdu+GK6+E/v0d3RshhBAOYc1SqRR88w2sXAnz55cEdpj/FdnHQumMJjnT3WyUQK7uaJ0NPASAUkWWbQMc2KM6JVMrhRCihvbtM8HcgQPw22+O7o0QQgiHsC0G/vPPsGSJyW5i4623YOSiy3ik7fvsz7MEb7JOru4otaq4/AAss1zOWRLICSFEDcXGmuvly82XsEIIIRqg+PiS5Cbx8SaziZtbqSZhYbCbVmxpcTVxmZagT4qC16URlgvAjcANDutJPZBATgghashaegCgUyfH9UMIIYSDVFEM3Co8HNzI5byM78hKzjYbZUSuLpni3iXlB876hCan4pBATik1SSm1TymVo5TaopS6qIr2lymlNiul0pVSSUqpT5VSbcq06Wc5Vo5Saq9SamLdPgshREMVF1eyBEIyVgohRAN05EjpYuC20yxthIWBN5nM/mEgXVI2oJWSQI6axQJKqSillK7gMriC5lJ+oC4ppcYAC4Ange7AT8AaVUnhPqVUNPAp8IOl/SWAJ/BlmTZfWo7VHZgLLFRKjaq7ZyKEaKji4sDPDzw8IDra0b0RQghR78oWA3dzg5YtyzULC4MU/ClQLgSRjPYPaPBTK2saC9gYDETYXL6roI2UH6hj9wJLtdavWe7fbYmo7wAerqB9D8AVeFhrXQiglJoLfKeUCtZaJwETgcNa67stj4lVSvUGpgEfVXBMIYSoFa3NGjl/f2jfvqQouBBCiAbEthg4wJYtFTZr1gzWfedE7shQwlKOkt8oGHcZkatpLGCVrLU+Uo1jp2DiByk/YE9KKTfMCzuvzK5vgAsqedjvQD5wq1LqdcALs3jxN0sQB9DHcgxbXwM3KqVctdb59ui/EEIcPQqpqfDoo/B//+fo3gghhHAI22Lgp+DuDgMGQHpwKKEpx8j2CWnQgVwtYwGrVUopD2AX8JzW+sNyLUz5gYctJzvnyw/U94hcMOAMHC2z/Sgmai5Ha71fKXUp8AHwEmY66B/AEJtm4cDaCo7pYjlnou0OpdQEYAJAZGQk69evr8VTEUI0RBkZMG+eKQK+fbu5CCGEaFha/vQTjT08+OHPP/Hbvp2opUvZNXky2RUEdmlpcJ63Cz2b7SbLKxCX/fv5/dz97OmilPrd5v5irfVim/s1jgWADMwsux+BAmA4sFIpdaPW+p1T9OWcX/zgqILgusx9VcE2s0OpcOAN4C1gOeALzAHeV0r9R2tddIpjVrQdyxtqMYC3t7fuLxV9hRDVtGgRTJsG110Hjz8OUVGO7pEQQoh69/LL0KwZ/QcMgP374fff6X3hhRUunO7TBzo5LWLLQcXbkS/TOHsv5/BnzwKtdc9qtKt2LGCZgfc/m02/K6WCgQeA0oGcUhdbHrQRaG7Z1rzyXuizetplfQdySZi0oOFltodSPjK3uhPI1Fo/YN2glLoeiMcMwW4CjlRyzAIg+fS7LYQQRmysWdP+7rswZ46jeyOEEMIhbLNUWqdZNm5cYdOwMPhlbw/+UZCsQkzWSq1BnfW5NmqjNrFARX4Bbqpg+3qgCBPjrKeS4NBC47hBLbuo16yVWus8YAtwaZldl2Iy1lTEC2tNiBLW+9b+b6b8cOylwO+yPk4IYU9xcSbRiZeXjMYJIUSDVbYYeGioWRBXgfBw8Di8l0k+b5FU0Ajy8818ywaolrFARbpRZumUDVXmtmSttKP5wNtKqV8xc10nAo2BRVCckfI8rfVAS/svgKlKqVnAe5iplU9iRuSsKYIWAXcppZ4HXgX6AuMBSUUghLCruDhwcjIZK50cUolTCCGEQ1WzGLhVWBh0TN7Ii9zEsqxnzMakpJKCpA1PjWIBpdSNmMSHf2BG24ZhZuw9WMGx51AyCvdoHT6HM0K9B3Ja65VKqSBgBqYGxD/AUK31AUuTCKClTfvvlFJjMfNg7weygZ+BwVrrTEubfUqpocBzmNSlh4HJWmspPSCEsJvMTDh4EHx9pRC4EEI0WEePQmFhydTKiIhTTtEIC4PfCAUgLcNSsyYpqcK6cw1BTWMBixmYNW+FwE7g5goTnWg92+a2BHJ1QWv9MvByJfvGV7BtBbCiimNuAGLs0T8hhKjIzp3m2tMTOnd2bF+EEEI4SNli4G+8ccrmY8bAFeGhMAoyMyw5+hp4UfCaxAJa62XAsnro1lnnrF7gJ4QQ9Sk21lx/9x106ODYvgghhHCQssXAqxAUBEG9wgDIS88zGxtwLbk6pdR3NWitKVnKdVaSQE4IIarJuj6uVauGmmxMCCFEqWLgu3fDsGHw4oswsOKYICMDXn83hCmAzs0xGyWQqyv9OXWmSqtKyx2cTWSpvhBCVJM1Y+X48Y7uiRBCCIeJjzdz7AMCTA25uDhwqXxsRGuY+rAH0/r/zgtMRru5NfiplXWsqkyV58xXsRLICSFENVlH5PbscXRPhBBCOIw1S6VSpUfnKuHjY+K+nb49OEEw+Y2CZUSu7kTbXM7HlCjYiClT1s5yvRE4DlzkoD7ajUytFEKIaigsNMlOXFwkY6UQQjRotsXArYlPKikGDibeCw+HrvGf40squb4huEkgVzdKMl+CKWMQDnRH62OWrTtRajsmw/0Uala77owjgZwQQlTD/v2Qm2suEsgJIUQDFh8PAwaY2wkJEBICHh6nfEhYGAzcuYQx7CTDMxxfmVpZHy63XPsBx2y2+1muL6vf7tifTK0UQohqiIsruS2BnBBCNFCFhXD4cEnGyjZtYMSIKh8WFgZHi0IJ5RhprjK1sp4UWq6/QKm7UGoESt0FfFZm/1lLRuSEEKIarIHc+edDp06O7YsQQggHKVsM/L77qvWwN94A32dDcXk6id0qSAK5+rEKuBloBSwos08DH9V7j+xMAjkhhKiG2FgIDYXNmx3dEyGEEA5Tthi41tWqRxMUBDQJBTSp+V5w8iTk54Ora511VXAv0BLoV8G+9Zb9ZzWZWimEENUQFwdt2zq6F0IIIRzKthh4Zib4+cGrr1b5sD/+gGVfmaLg6TmW4O3EibrqpQDQOg2tBwCDgWeAN4CngcFo/R+0Tndo/+ygWoGcUvSu644IIcSZLC7O/COePdvRPRFCCOEwtuUGDh0y1b69vat82L59cOcXQ7gi5jB/6S5mo0yvrB9af4PWD6H1bWj9MFp/4+gu2Ut1R+Q2K8VfSnG3UgTUaY+EEOIMk5QEycnm/7WfX9XthRBCnKPi402GysDA0qNzVQgLg0x8yGoUwb7MULNRMleK01STqZUdgeeBQ0rxrlIMqJsuCSHEmSU2tuS2ZKwUQogGzLYYuHW93CmKgVuFhYEreYw/+CiNTuw1G2VETpym6gZyzwEJgAI8gGuBtUqxSykeUorwuuqgEEI4mpQeEEIIAZQuBm4dkYuMrPJh4eFQgAvX7X2Mdnl/mY0SyInTVK1ATmvu05rmwMXAK8BxTFDXEngCOKgUHyhFt7rqqBBCOEpcHDg7m2mV1fh/LYQQ4lwVH18ylbJTJ7j99iqLgQP4+ICXtxMpbqH4kGE2ytRKcZpqlLVSazZpzZ1AL2CDzS4XYCTwi1Jcacf+CSGEw8XFmQDunnuqlWVaCCHEuahsMfArr4RFi6r98JMnwaVxKMEkUeDtJyNy4rTVKJBTikuV4iNgD2Z0DszI3B/AXsAVM0InhBDnjNhYuOACmDPH0T0RQgjhMMeOQUFBydTK9HRTR66aXF1Bh4QSyjFy/UIkkKtrSr2CUjGO7kZdqm75gfuVYhfwFTACMwKngU+AflrTA+gGpAFt6qKjQgjhCNnZJm10NdayCyGEOJeVLQYeFWWmalTTG29AXLIZkcvyDJaplXXvduA3lNqKUneg1DmXd7q6I3JPAy0wo2/pmOyVrbRmpNb8AKA1mcARwLkO+imEEA6xa5e5fvZZ+O47x/ZFCCGEA9mWG8jKMgW9w6uf7+/33+Hqk4tpw07S3IJlRK5+KKAr8CKQiFJLUeoiB/fJbmoytXIfMBVoojX3as3+Ctr8BxPwCSHEOcE2Y2W7do7rhxBCCAcrWwzceruawsLg0Ekv3D2cOOEsUyvrQWfgSczyLwV4AuOA9SgVh1L3oVSwIzt4uqobyF0FtNaaBVpbU+2UpzWHteaAfbomhBCOZ60h5+8PEREO7YoQQghHio8Hd3cICio/zbIawsKgm97K604TSMvzMFMra7DGTtSQ1v+i9Qy0bg2cB8wHDmGCujbAM8ABlHoGpc7KGYXVDeTWA02VolTUqhTBStFMKRrZvWdCCHEGiIsz/7c7dpSMlUII0aDZFgO3nWZZTeHhEM4Rrst6jawcJ8jJMVM0RX3YBxwD8jB5PjQlo3T3YWpmn3WqG8gtwbwAY8tsv9ay/Q17dkoIIc4UsbFQVCSFwIUQosGzLQbesSM8/HCNArmICMj3DwUgM8cyACTTK+uWUheh1LtAAjAXiMYEcHuAe4CnLPfHOKyPp8Glmu16W64/KrN9FfCCzX4hhDhnFBXBjh0wYACMG+fo3gghhHCo+Hi4yJIno0cPc6mB88+Hb/8Kg2aQk11kNh4/Ds2b27mjAgCl/gHaW+9ZrtdjkjauRlvmtSp1PRBZz72zi+oGciGW65Qy21PL7BdCiHPGwYNm5svo0XDhhY7ujRBCCIcpKjIJTqwjcAcOQKNGZgF1TYSYj8z5WfnmvozI1aUOlutcYDnwPFr/VUG79cBZWWSoulMr0y3Xg8pst96vNAGKEEKcrawZK52cZD26EEI0aGWLgY8YAddfX+PDXH+rByluIeQWuZoNEsjVpaPAbKAZWt9cSRAHWt+A1gPqs2P2Ut1AbitmSHKJUsxQilFKMQOzNk4DW+qqg0II4SjWQO7++x3bDyGEEA5WNktlfHyNSg9Ybd4MQ2OO8giPmQ1SFLwu3QisR+tz9kWu7tTKRcAlgB/wqM12hQnkXrFzv4QQwuHi4sDFBTp1koyVQgjRoNlmqczOhuTkGiU6sQoLg6wsRSqN0M7OKBmRq0tfAUVUFO8oVQQUoXV1Y6EzUrVG5LRmFab2gipzAfif1nxSJ70TQggHio01UyolY6UQQjRwp1kM3CosDEYkvMiL3EWub7BMrax75b+GVcVfzZ71X9FWOwrVmmlKsRIYDoRh5p1+pjW/1VXnhBDCkf79FwoLJZATQogGLz4e3NwgOBj++cdsq8WIXHg4NM/4h6F8TJZXCB4ytdK+lGoGRJXZdhGlg7Yuluv8+ulU3anRcKIlaJPATQhxzjtxwsycAQnkhBCiwbMtBt66Nbz8spl3X0Nt20JBQCjBR5JIcGtHoIzI2dtNwEyb+wqTlbIsDRysjw7VpWoHckrhAgwF2mKqoJeiNXPs2C8hhHCoHTvM9aOPQs+eju2LEEIIB7MtBt6kCdxxR60OM2UK4BIKdxeRjh8k7bFbF0Ux6+ibLnO/rAX10Jc6Va1ATilCMdFs21M0k0BOCHHOsGasvO468PV1bF+EEEI4WHw89O1rbv/7r6lL0779qR9TmbAwANILvSRrpf2tt7k9CxPM2SZq1EAy8DNan/VZ96s7Ivco0O4U+6XCkhDinBIbC87OkJgILVs6ujdCCCEcpmwx8IcfNgXBt22r8aHi4mDezMbMdG3ByUIfM4+/sND8wxGnT+sNwAYAlLoJ0Gj96Ckfcxarbh25QZhg7U3LfQ1MBnYBO4Fb7N81IYRwHGvGyhUrHN0TIYQQDnX8OOTnl0yttJ1mWUNubvBGXF+u7LiHv+lqgsSUFPv1VZTQOgqtox3djbpU3RG5SMv1Q5hFhGjNi0rxPfA3UKO0PUqpScD9QATwLzBFa/1DJW1nY4ZGKxKmtT6mlOoPfF/B/vZa67ia9E0IIcAkJSsqkkQnQgjR4FVUDPy882p1KMusSlxd4UBmsLlz/DgEBZ1mJwUASs3EjMI9Zrl9alqf1UvDqhvIFQKumDml+YCLUoQAByz7JwCPV+dASqkxmMWFk4BNlus1SqkOWuuKssfMwxQkt7UC0FrrY2W2dwRO2NyXicdCiBrLzYWDlr9GHTo4ti9CCCEczLYYeE6Oqf1WyxE5b29zeX7nUP5MswwWSeZKe5qNKQL+mOV2Vcu/GkQgl4wZlWsEHMGMwL0L5Fj2B9TgnPcCS7XWr1nu362UGgzcATxctrHWOgPIsN5XSjUFLgLGVXDsY1pr+W0QQpyW3bvNaBzIiJwQQjR4FRUDr0UNOavwcGh98E+OaA+zQQI5e1OV3C7rrM/xUd1AbgcmkGsJbASuAwZa9mlga3UOopRyA3pgRtlsfQNcUM2+3AKkAB9VsO93pZQ7sB14XGtd0XRLIYQ4JWvGysBAU/tVCCFEA2ZbDNzTE1avhm7dan24//wHMlaG4pNvGaeQzJX2NKCS2+ek6gZyrwG7AQ9MBstBQIhl33FgSjWPEww4A0fLbD8KXFLVg5VSTsDNwFta61ybXYmYEb3fADfMaN06pVR/rfXGCo4zATMdFDc3t2p2XQjRUFgDuU2bHNsPIYQQZ4CEBIiMNCUHfH3hiitO63CLF0PyllAabU01G2REzn5M1sryt89R1QrktOZ94H3rfaVojYlyC4AftSalhuctO5SpKthWkSFAU+D10v3TOzCjhlablVJRwDTMCCJl2i8GFgN4e3uf9cOqQgj7iouDZs1qXyJICCHEOcQ2S+Wff8LhwzBkCKhTzdo7NafwUELYTYGHNy4SyNU9pUIxA1KlVZyf46xRZSCnFNapigCXa02c1qQBn9bifEmYxCnhZbaHUn6UriITgJ+01v9Wo+0vwLU1654QQsBff5nr7dsl2YkQQjR48fHQp4+5/dprsHy5qf9WS4sWwfGNXehFEsFeRfjJ1Mq6oZQPMB+4HnCvoIWm+rMTz0hV1pHTmlwgCIgC9p7OybTWecAW4NIyuy4FfjrVY5VSjYHLMdM8q6MbZsqlEEJUm9awY4fJWpmc7OjeCCGEcChrMXDbGnKnkejEambGAwzhKzLcg2VqZd15DrgVMxKnKrmc1aobha4FrgK6YtahnY75wNtKqV+BH4GJQGMsJQaUUnOB87TWA8s87mYgE5spnlZKqSnAfkxNOjdM5D0CGHWafRVCNDCHDpnyAyAZK4UQosFLSoK8vNI15GpZesDKWkvOzw9OuobQWAK5unIFZtQtD/gHE0ecU0uqqhvIPQ/0A5YrxXTgTyDbtoHWVGuOqdZ6pVIqCJiBKQj+DzBUa22tSReByY5ZTCmlMNkq39VaZ1VwWDdMJsxIS7/+BS7XWn9ZnT4JIYSVbcbKwEDH9kUIIYSDlS0GnpAAvXqd1iHDwqA3P7Ms6xYyvZo3yKyVSqlJwP2Yz/3/AlO01j9U43GtMdnyldbap4rm3pbri9H6dAeizkjVDeQ2YiLYQOC9CvbXaI6p1vpl4OVK9o2vYJsGok9xvGeAZ6p7fiGEqExsrLmW0TghhBClasjl5Jig6zSnVoaHQyHOtC3Yzqa8lpDRsEbklFJjgAXAJGCT5XqNUqqDPkXyEUsZsxWYuKRfNU61HrMsa/9pdvmMVeUaORuVzS09J+aYCiEEmEDOyem0SgQJIYQ4V1gDuSZNwNXVZK0cP/60DhkeDt0vM/Mrs/OcISPDBIkNx73AUq31a1rrWK313ZSUEjuVp4G/gA+qeZ77gVRgGUp1QCnnWvf4DFXdUbRlddoLIYQ4Q+zYYWbNPPeco3sihBDC4eLjTQAXEmK+5eva9bQP6eUFiz8OAS/Iy7Us2UpKsksSlTOdZVStB2ZJlK1vgAtO8bjLMWveYqh+Dgxr1v3LLJeyJSM0Wp/VWSurW0fuprruiKMEBgayfv16R3dDCHGGuPJKswD9hypn6gshhDjXtd+yBb/gYH7ZuBHv3bvxi43l6CWXUOTpedrH7uvhRbvooxALv3/1FRmtWtmhxw7nopT63eb+Ykv9ZqtgwJnyZceOApdUdEClVAQma/1IrXW6qn79Pmud6nN25uBZHYXaw4kTJ+jfv7+juyGEOAOkpsKAAWZ93B9/mC9hhRBCNGCzZkGrVuaz4s8/w/z5tH30UfCpKs/GqV1xBaS6jWRfbBBT+JmeUVFwbnweLdBa96xGu7LZI1UF26zeAV7RWv9cw75Yc3ycs6oVyCnFkiqaaK25xQ79EUIIh9mxw1wfOSJBnBBCCMwaud69S277+592EAfg5gZ3+b2NT1osU1jQkDJXJgGFQHiZ7aGUH6Wz+g/QTyk1y3JfAU5KqQJgUpkRvxJa9z/t3p7hqjsiN57KI1prBC2BnBDirGYtPdCunWP7IYQQ4gygtQneRo829+1UDBxMwpO0NMgh2GxoILXktNZ5SqktwKWUTlpyKfBRJQ/rXOb+lcB04DzgkN07eRapydTKc3Z+qRBCQEnpgZ7VmRQihBDi3Hb8uN2LgVuFhcH9aTMYw0q0UqgGEshZzAfeVkr9CvwITAQaA4sAlFJzgfO01gMBtNb/2D5YKdUTKCq7vRylZlbZE63n1KL/Z4zqBnJla7i5AC2AR4DumCwyQghxVtuyxVx3LvvdnxBCiIbHtvSA9X6PHnY5dFgYHEfRgr3keAbg2XCmVqK1XqmUCgJmYAqC/wMM1VofsDSJAFra4VSzqXqN3LkfyGnNgQo271GKzZi5rncAG+zZMSGEqG+7dkFAgF2ySwshhDjb2RYDB9i+HQoL7XLomBg4cHEYzhuLyHQPwLNhjcihtX4ZeLmSfeOreOxSYGk1T3WqGYVnfSKU081a6YJ5EQbboS9CCOEw+fnmf/b998vUSiGEEJiplFAyIhcUZLdDn3cenHdnKGyETCc/ghtYIFdPBpS5b51ROBVoiskBclY7nayVHkBfwB1TNV0IIc5ae/dCQYEkOhFCCGGRkGBSGIeGmmxY77wDd9wBkZF2OXx+QCiuQKb2akhZK+uP1hXNFlyHUp8BCZgi4ZUlWDkrOFWz3XjgxjKXMZhoFuBLu/dMCCHqkTXRyZo1ju2HEEKIM0RCggnanJzMIuonnoD0dLscOjsbWg2KZpXPDRzToQ0ma+UZIgXIBa5ycD9O2+lmrcwFlgNT7NIbIYRwkO3bzXVwsGP7IYQQ4gwRH1860QnYbTTO0xNO+DTn4cbLmHr8v/RPWm3KHShJEm83FWet9MCUOvDCxDFntdpmrQTI1Zoj9uyMEEI4yu+/m+uYGMf2QwghxBkiIQF69Sq53agR+Pra7fDh4YDWHMoKMHP7U1NNwXFhL7OpPKGJ5hxI1Hg6WSuFEOKc8Y+lGk3Hjo7thxBCiDOAtRj4yJHmvh2LgVuFhcF7v7Rkb0EzsyEpSQI5+6tsiHM9MKke+1EnqpvsZDCmevofWrPaZvtwoBvwq9Z8VSc9FEKIOqY1HDxobnfo4Ni+CCGEOAMkJUFubknwduyY3YqBW4WFQTq+uFtn+CUlQatWdj1HA1c2ayWY6ZQH0DqxvjtTF6o7tXIm0BsYUmZ7BmbYcjNIICeEODsdOWL+X19wAfj4OLo3QgghHK5sMfBNmyAnx66nGDsW1C+heB+yrFSSzJX2tg8ArQ86uB91prpZK60JuTeX2f6r5bq9fbojhBD1Ly7OXD/6qGP7IYQQ4gxRthi4UiZDiR2NGgUhHULxt1bxksyV9rYf2FvhHqX2odSeeu1NHahuIOdluS77XbVvmf1CCHHWsWaslBpyQgghgNLFwBMT4cYbTQkCO8rNhWy/MAI5YTZIIFcXKlsj1xyIqsd+1InqBnLWeaTTy2z/r+X6sH26I4QQ9e+XX8z1Dz84th9CCCHOEAkJ4OJiioHv3g1vvQXJyXY9xZo1MPGjS3iBuylwdpOplfaglB9KNUOpZjbbmhZvM5dhlj2Fjumk/VR3jdxa4BbgDqUYBOwA2gItMek719ZN94QQou5t22auW7d2bD+EEEKcIazFwJ2dy0+ztJOwMFjDUL51GcqdHm/TSEbk7GEqJreHlcJMsSxLA4fqo0N1qbojck9hEpuACd6GWq4VkGnZL4QQZ7ZFi2DcOFOrx8b+/ea6vaz2FUIIAaWLgdtOs7Sj8HAATROvE6S4BsnUSvtRlJ5SqSq5vF3/XbOv6taR22MZiXuD0olNtgO3al3JQkIhhDhTJCTAlClmUcJ338GePeDhQUYGpKVBQAB4ezu6k0IIIc4ICQnQo0fJbT8/uxYDBzMi14Mt/J7Wiz2+XWVqpX38CSyz3L4RM/L2ls1+DSQDP6P1R/XbNfur7tRKtOZnoKNStATCgKNac9ZnexFCNBBDh0JhIQwcCOvWQdeuEBvLzp1mYkKLFg7unxBCiDODtRj4iBHmfmEhtGlj99N4eUGmVyhkQU6Rm4zI2YPWnwKfAqDUjZZtNzmwR3WqulMri2nNHq35SYI4IcRZY8UK+PtviImBtWvhootg5044/3xit2sAxo93bBeFEEKcIZKTTc0461TKl16C336rk1M9OC8EgJx8Fwnk7Ekpd8zI3DKUauno7tSVagVySvGOUhQqVWrxIEoxw7L9nbrpnhBCnKaiIpg0ydQAev99s239eujcGX77jZ6zL8fZGW67zaG9FEIIcaYoWwy8Do2/w5McN1/yCzSkpEB+fp2fs0HQOhcYA9zAOZxdv7ojcn0t12UXBb6DWSzYFyGEOBM9/jicPAlXXQXNm5ttTk6wdSu0akXbPWuY1mgxLtWeaC6EEOKcZpulMj/fTMn/+OM6OdWhQ5DmGYZzUYHZYOcSBw3cn5g4JcTB/agz1Q3kIizXR8psP2q5DrdPd4QQwo5yc+HJJ8HVFZYtK73PxQW2b2et2xCePDGRgoUvO6aPQgghziy2WSoPHzYJsk6cqJNTPfkkzMl+gO8YYDbI9Ep7egDIBRahVJijO1MXqhvI5Viu+5TZ3qfMfiGEOHO89ZYJ5h58EHx8yu0uUK6MyP+QPao17lPvhLlzHdBJIYQQZxRrMfCwsDorPWAVFgYv5d3GVwwxGyRzpT29hSn6fRlwGKUSUWqvzeWsz/dR3UDub8zQ5FKluF4peijF9cCbmDSef9dVB4UQolYyMmDmTOjbF+bMqbDJ/v2Qqb2Y3O5r8PCA//4X3nijfvsphBDizJKQAI0b12kxcKvwcGhECn5Y6pvKiJw9RQFeltsKk3U/qszlrFbdVSFLMevgIimpzQDmRdGW/UIIcea46io4cgRWrTKJTirwt+UrqNBeUbD8Z+jZ02Q9CQoqSTsthBCiYamHYuBWYWFwH//jvzxpNkggZ08HMXHKOataI3Ja8wbwEeUrogN8qDVL6qZ7QghRC3/8YcoMNG0KfcrOCC/x44/mum9fTF25detM0DdqFPz1V/30VQghxJklIaEkcPP0hO7dTUHwOhAWBkcJw5kis0GmVtqP1lFoHX3Ky1mu2nXktOZq4FrgXWCt5XqM1oypo74JIUTtXHutuX7rrVM2O3YM/P1h2DDLhosvhg8/NNMsR440I3pCCCEaDmsxcOtUyrvuMlmO60i7djB2SigAOa4+MiInaqRGBcG15n2tGac1gyzXHyiFj1LcWFcdFEKIGlm1yhT7jomB/v1P2XTPHjMQFxFhs/Gqq8xoXmKiCewOHKjT7gohhDiDnDgB2dn1UkMOzJeJF1xpArksFz8J5OxNqVCUWoRSB1Eqx3L9MkqFOrpr9lCjQM5KKZyUYqhSvIcpSVCj7ABKqUlKqX1KqRyl1Bal1EWnaDtbKaUruYTatOtnOVaOUmqvUmpibZ6bEOIsN3Fi6eLfldAatm0z/0TL6dMH3nkHdu2CLl3kH6sQQjQUZYuBX3QRzJtXp6fcEm8J5PCSqZX2pFQQ8DNwGybPh5vl+nZgM0oFOrB3dlGjQE4peinFAiARWI2pmO5FyXq5ahxDjQEWAE8C3YGfgDVKqWaVPGQepo6d7WUDsF5rfcxyzGjgS8uxugNzgYVKqVE1eX5CiLPct9+af4LXXw8tW56yaVISZGbC3r2VNLjqKpgyBdLSoEMHyMqye3eFEEKcYcoWA//xR5MFuQ7dPLsZDzR6lcPOTeWLQ/uagclMaY1TLKlBUZbt0+u/S/ZVZSCnFNFKMUMp4jBR7V2YCunWhCfZwIoanPNeYKnW+jWtdazW+m5MYHhHRY211hla6yPWC+AKXAS8ZtNsInBYa3235ZivYbJrTqtBv4QQZ7PCQpg2DaKi4LXXqmy+bZu57tjxFI2eew6uu84Ehx06QEGBXboqhBDiDFW2GLjWdT7N0jfCh7c9J7C3KEoCOfsahslauQQIROsAINByXwHDHdg3u6g0kFOK25ViE7AbeBRoTelslWBenDCtua46J1NKuQE9gG/K7PoGuKCafb4FSMFk0bTqU8ExvwZ6KqVcq3lcIcTZbOJEk2ly9mxwd6+y+Q8/mOtTJLU03nkHBg0ya+UGDTL/1IUQQpybEhJM/bjw8DqvIWcVHg6tsv8mJ1ujjx+X/zP2Y43A70NrMxpnru+zbK/bH2w9ONWI3CuYAMkavOVjpi/eAvS3NtKamow3BwPOwNEy248C4VU9WCnlBNwMvKX/v737jpOrrv4//jrbUja97KaRhFBC6KFLM1IEo4JIExUNiKAgfAEpKihNivykWQAD0sQCglKUJiWAgEgCAUJJgJCy6b1vP78/PnfMZHZ2d3Z3du7O7vv5eMxjZu7nzr1nJze7c+ZTjntVUtOQRo5ZFJ0z9Tinm9lUM5taq2/YRfLf8uVw111QWgonn5zRS6ZODfcHH5zBzk8+GZa2fOGFkCiKiEjnlFwMvJ1ryCWUl8Nv136Tvfy/WFVVGPcv2VAd3acmbInnVeS5TObIOWExkyHufMmdu4EVbTxv6lcNlmZbOl8gvPl3ZnjMdNtx98nuvpe771VUlGlNdBHpsE46Cerr4Re/gILMpv7OnBnux43LYOeCAnj0UTjlFLjyyoyTRRERyTPJxcD79oVDD81JIrfEy+ieyCs0vDJbEgVhH8Ps/zA7BrP/Ax4l5AfvxhdadmS62MmpwIdm3GbGYYR5aq2xHKijYe9bGQ171NI5HXjV3d9L2b64kWPW0vakU0Q6shkzQrmAoUPhrLMyfllNDUycmNEozMAMJk+GrbcOwy2/p4VxRUQ6neRi4F/4Qvj70rdvu57y5JNhxB7llBL1xGnlymy5k9CxMwq4EXgouk8UAp8cU1xZ01Qi9wtgPpuHVpYREqmngX+35mTuXg1MAw5PaTqcsOJko8xsGPBFtlzkJOE14LA0x5zq7jWtiVVE8sRJJ4X7u+/O+CUbN4YvXffdt4XnKioKhWH79YPf/Q5+9rMWHkBERDqs1GLgObL11jBwXBl9WBs2qEcuO9zvIeQNlnIDmIz7fTFFljWNJnLu/Nid0YT5cL8nLDCSeAN6Eg1ZNGO+Gde24Jw3ApPM7DQzG2dmtwDDgNvD8exaM3suzetOBTYA6YpD3Q6MMLObo2OeBkwilC4Qkc7qgw/g/ffh8MPhiCMyftmMGeHvdUlJK87Zr184QM+ecNVV8Otft+IgIiLS4axaFb7pS/TIHXIITJrU7qddswY+WlNGKVGZGyVy2eN+BnAgoTTZndH9AbinXS0/3zQ7tNKdl9z5LmHo4vHAY4TJg4mkbjhwUaYndPcHgHMJtR2mE97cie4+N9plKLBFASgzM8IiK3909wbFnNz9U2AicHB0zEuAc9z94dR9RaQTufhi6NUL/vjHFr1sypRwX1jYyvMOHw7TpoVM8LzzQi+diIjkt9Ri4DNnZjzvui1WrYJJ/ziW43kgbNDQyuxyfxX3S4CzcL8E99fiDilbMr463al252F3vkJIts4iDGlsMXe/1d1Hu3s3d9/T3V9Kapvk7qNT9nd339rdz2zimC+6+x7RMbd299tbE5uI5Imbb4bHHw/z4gYPbtFLX3893B9ySBvOv8MOoYbBkCFw5JEwa1YbDiYiIrFLLQa+aFG7L3QCYbGTj9mOx0uOp66gSD1y2WS2I2YPYbYGqMRsDWZ/xWzHuEPLhlZ9zeDOKnduc+cAYFvg8qxGJSLSlNpauOSS0KV2wQUtfvkHH4T7XXZpYxz77APPPRdWzNxzT3jjjTYeUEREYpNcbmDx4jAGPwfz5Xr0gCG91vPV4sdZX9xfiVy2mO0DvA4cA/QmjCTsDXwVeB2zvWOMLiva3F/szmx3rspGMCIiGfm//wvzGL77XRgwoMUvX7AgjMhs1Ry5VGPHwm23wfr1cNBB8NFHWTioiIjkXEVFGEqZXAw8Bz1yAGMHreBPG46m2os1tDJ7bgRK2VwPe3F0b9H2G+MLLTvaf+CviEg2rV4dVozs0aNVC43U18O6daFaQdYcfzzcdBNUVcH48bBwYRYPLiIiOZEoBl5UBL17h4VOxo7NyamLhoYpApXeTT1y2bMnYXHGXwH9cB8G9AN+k9Se15TIiUh+OflkqKsLK0YWFbX45fPmhdEy552X5bjOPRd+8hPYsAF23hnWrs3yCUREpF0lFwPfeedQ1mbMmJyc+vb7erKpqBfV9Zojl0Wrovuf4r4JILq/NNqe97WmlciJSP5YvhxefBF23x1++MNWHSJr8+PSufpqOO20sATZcceFhFNERPJDcjHwysrwrV+ObLstbOxdTn2d4xpamS2JOnE7pGxPdLPelcNY2kXLv84WEYnLFVeEuXF/+lOrD/G3v4X78vIsxZTqjjtg9Gi49FI49VQ49tiG+2yzDZiFeRBr1jRs33bbcL9kSRgHmqywMFSPhbCi2oYNW7YXF8OoUeHxwoXh/UpWUgIjR4bHFRXhw0qy7t03f5CZNw+qq7ds79kzDD3q3j1n31SLiLS7RDHwiRPD85NPDnOep0/PyemnT4dulNGNxbByZfgisNU1ciTyCaHX7XHM7gDmASOB04AFwFzMvvW/vfOwQLgSORHJD//6F/zmN/DVr8K4ca0+zNSp4X706OyEldYll4QE6ppr4L68+7uQuYceSp+oiojkm9WrwxdjiS+yKipaXNqmLd54A3676laO5SF+6leHkR2DBuXs/J3U7whz5AB+kqb9jqTHzuYevLzRaCJnxsEtOZA7LzW/l4hIK02aFO4vu6xNh5k3L3QqFRe3PaQmXXUV9OuXfuGTffcNK6PNnh163ZKZwX77hccff9xw9bKiItg7WjH5ww/DH/tkJSWhFALA++837PHr0SMMTQV4992w2mayXr02jzudPh02bdqyvW9f2HFH+Na34MwzlciJSOeQukrl/Plw+OE5O315ObzN7owjGv+/bJkSueywuANoT031yE1hcxbbHG/mWCIirXfrrSEhOuQQ2HXXNh1q9erNIw/bVUEBXHhhDk7UBiec0Pr2m2+G116DadM2J44iIvkquRh4bW0Yup6DGnIJ5eWwI++xH/8JGzr5gidmdiZwITAUeA84191fbmTfHYHfAjsCfYGFwF+Ay929Ot1rIqdkNegOqLnkq1NnsSKSB+rr4aKLQmL0l7+06VDz54fDbbddlmLryq66Cg47LKzU+fTTcUcjItI2qcXA6+tzVkMOQum6Q3ie/+NXYUMnTuTM7ETgFuBM4N/R/ZNmtqO7z0vzkmrgXuAtYDWwG2FYZBFwUaMncr83q4F3QE0lcqk//OeBIcArQAUwAjiAMInwH+0SnYjIj38c5i2cckqb5yu88Ua433ffLMTV1R16aBhm+cIL4QNPgRZBFpE8ligGPnRoWGzk0ktz+seivByWUrZ5Q+deufJ84B53T8xRO9vMjgS+D/w4dWd3/xj4OGnTXDObAByU0dnMdgEOAwYBy4FncX+3tcF3JI3+5XXnlMQNeJaQxJ3ozsHufN2dg4GTgIGE5E5EJLuqquDBB8Mqi7ff3ubDrV4d7hPT7aSNTjwRampaVZhdRKRDqagISVxREZSVhVEHu+2Ws9N37w6/fTApkeukPXJmVkIoxP1MStMzwP4ZHmNb4EjgxWZ2LMLsHmA68EvgR9H9dMzuxizvlwXNdF5bonDeUynbnyAMv7wQ+H22gsqlAQMGMGXKlLjDEJE0RjzwANvOmcPb/+//serVV9t0rPr6kHPccENY8GReusEb0iJFX/gCn7nnHpY/9hgf5PADj4hItu36zjsU9enDm1OmULxmDQ7U9ukTFqDKkZ7r5zAIqC0qZtFbb/FJfn4+LTKzqUnPJ7v75KTng4BCIGWlL5YQes0aZWavAnsA3QhDK9OtRJns58C3Gmn7FrCYND2A+STTRG50dH8mcH3S9rOi+1wsHdAuVq5cyYQJE+IOQ0RSzZ4NkyfD7ruz2wUXtPown3wC3/8+PPdcSOYOOwzOPz+LcXZ1p5xC+X33UT5+fBhqKSKSjzZsgJ12Cp8Jzz8ffve7sKpvDhO5B97bmX2ADd6brbp3Z6v8/Hxa6+57ZbBf6oKKlmZbqhOB3oQ5cv8PuBi4ton9vxUdcxlwJ1vWkSsDJpHniVymkxpmRffXmrHEjOlmLAGuIbxBsxp/qYhIK5xwQsi8Lr20+X3TeOutMCpm221DCboBA+AXv9C6HFk3aVIoUfCjH8UdiYhI67iHxU4Sq1RWVISFTnKYxAE8/MIA9ur5PhVFozrt0ErCHLU6wpStZGU07KXbgrvPd/f33f3PhGGSl5lZU51SiW8Xv4j7pbhPxv1S4MvR9j4tD79jyTSRuwSoJ2TLg4BdovtE9txc16aISOamTAnL2m+/fYvrlP3rXyEH3HNPeOcd2GabULd62bLNi19KFu2zD3TrBn/4Q9yRiIi0zpo1DYuB57D0QELZkAJm1I1jYf2QTpvIReUCpgGpRfoOB1oyh6KAMLKwqXlub0X3H6Vsnxnd/7cF5+uQMvpI484/CJMKXyckbokE7j/A5935Z7tFKCJdz7eiIe1/+lNGu9fWwiWXhJF9n/986HW76KJQT/vjj1Wzul0VFMAXvhA+BD38cNzRiIi0XLpi4DksPZBQXg7HVP2Z7jVr8c69auWNwCQzO83MxpnZLcAw4HYAM7vWzJ5L7GxmJ5vZ8Wa2g5mNMbMTCEMqH3L3qibO80OgEvg5Zt2jg3UHrgLWElbPzGsZF/F25zngOTN6Av2BVe5sbLfIRKRreuCB8Ef0oIOaLTS9ciWcc05Y2LKmBgoL4StfgTvugEGDchOuANddB488AldfraxZRPJPzMXAE4YMgc/we7ZhNr5sbact5uzuD5jZQMJiikOBGcBEd58b7TIU2CbpJbWEuWzbETqT5hIKhN/UzKn+TBhReBZwOmYrCKvtFwPrgYeThs867tukPUoHlnEiB2BGEbAzMNCdJ9snJBHpstzDUvb9+4eErhHLl8Ott4YVKNeuhR494Hvfg+uvD0s4S46NHRu+vZ4+PfyD9Mn7aQci0pUkFwOvrw9/YMaPz3kYoZZcOeN5i4KNG8L84x49ch5HLrj7rcCtjbRNSnn+Z0JS1lKj2TySsISQICb0Iiyc8r/TtOL4scs4kTPjeOA3hLlxDhSZ8RywNfA99wb1IEREWuavf4VXXgmrhQ0d2qD52Wfh3HNh5szwpenEiXDkkXDWWZr7FrvvfS8sTHP//XDmmXFHIyKSuYqKsLDJ0KFQXAynnx5LGBMnwqLjyih9aEPYsHx5LD2Dncg88jRBy1RGH33MOIiQCScWOEn0Q/6TkO0e1x7BiUgXsmEDnHxy+Ery1FO3aJo8GYYPh8MPh/feg3Hjwv0//wlnn60krkP40Y/CB47HH487EhGRlkkUAy8uDsMqp08P4/VzrKgIum1VRneiaV+ddMGTnHEfjfvWGd/yUKYff34c7TszZfuz0f1nshaRiHRNkyZBdXXo2Skqoq4O/va3MNLljDPC39YDDoC33w6rUe64Y9wByxYKC8O/4VNPwbvvxh2NiEjmkhc3+ctfwrDKdetyHkZ9Pfzj9bLNG5TISTMyTeT2I3RNfjll++zofnjWIhKRrmfevLDiYf/+LD/zZ3zta7D11mHdDHc47jhYvBj+/W/Ydde4g5VGHX10uD/33FjDEBFpkUTduMTjHj3CXO0cKyiAS2acxF68ETZ07pUrc8OsDLPbMZuHWWV0fytmZc2/uOPLdI5caXQ/L2V7j5R7EZGWO+EE3J0ry37LFUMKcA9/U//6VzjmmNDZI3lgzz3Dh5+XXgqTGItatJ6WiEg8KipC7ZrE4622ynkx8IQ+Q3qyuGo0VKEeubYKK2P+BxiV2ELofDoDOAKzvXFfGVd42ZDpX9kFhDchdQjlBdF9RdYiEpEGFi6EWbMabh88OHxxuHFj+t/3ZWVhFcf168NS/anKy0Mt53XrYNWqhu2JKQNr18Lq1Q3bhw0Ln9VXrw77pBoxInzDuGpV+lEqI0dC5bQZjHv9dWYzhstnnsSQIfDTn4YRlpr7lodOOims+HbzzXDBBc3uLiISqzVrwh+omGvIJYweuI4jZ/+COoxCJXJtdSlhLQ8IIwvXAH0ICd1o4BJCrbm8lWki9zQhe30kscGMDwn1HDxqF5F2sGYNfHbreQyrns1wFjRon8NoKulBP1YxhMUN2j9la6roTn9WUs6SBu0fsy21FDOIZQyi4R+Nj9iOOooYzFIGsqJB+4fsABhDWEQ/Vm/R5hgz2QGAoSykL2u2aK+jiGN5mDGU8KPRf+GJW0NtacljV10Ft90Gv/mNEjkR6fhSi4FXVMAhh8QWzuAhhZxX+0s2FvSkp4ZWttWXCXnK3cAPcV+DWV/gBuBU4Ci6SCL3c8LKlAPZvIxnoijfCkJ1dRFpBzNeXsWM6u3oRnXcobSbDef8mL/esnfcYUg2DBgAO+0EM2aEb7a1dLaIdGTJxcAB7rgj/B6LSfnWPVlHL+oopqd65Noq0bUakjggSuZ+SEjk8v4PVEaJnDsLzDgA+BVwKFAI1AHPAee6p+kmEJGsqPjLvzmAatadeRG9+6X5L7vTTtCrFyxdCp9+2rB9l12gZ8+w7OO81GmuwO67h/GVFRWwIM1/5T33DOMn580Lx0i1zz5hLsGcObAkpcevoAD2jhK0Tz5pOP6zuBgOOIDSQw9N96NLvrruOvjSl+Cxx0KRPxGRjiq5GDjAEUfEFwtwww2w7PflYb6CErm2qgaKCQlb8pCgRAJXlfOIsizjmejuzAKONKM7MABY6U5lu0UmIgAMeOT3APQ6axLsOC7eYEQy8cUvhi8I7rpLiZyIdGyJYuDDhoXlkd94Aw4+GPr2jS2kqr5lFK9ZS/3SZRkvLy9pvUNY3+MxzG4hLNo4EjiHMMIw72vlZFoQvK8ZI80Y5E6lOwvdqTRjULQ9vqtdpJMbs3EGdRRgSuIknxx9NLz5JjzwQNyRiIg0bv58GDIkjBB55RU46qgwwiQmb78Ns1aXAYYvU49cG91JmAY2CrgReCi6TxT/nhxTXFmTaaJ/F/Ap8PWU7V+Ltv8+m0GJSFBZCUN9AWsLc1/PRqRNvvWtcH/11fHGISLSlOQacqnDLGNQVQWfX/sQd3IaBSuXh2Kq0jru9wB3EJK55BvAZNzviymyrMk0kds3un84ZfvfCG/IvohI1j3/2Hp6UMmqAdvEHYpIy4wZE+pLzJiRvnaFiEhHkFoMvHv3eBc7KQ8rOi9nEFZXp9+fbeV+BnAAYWHGO6P7A3D/fqxxZUmmidzg6H51yvY1Ke0ikkWL7n4KA2rHa0VHyUNnnRW+Tf7pT+OOREQkveTVdRNJXUzFwCEkcgfyMsfz17BBC560nftruF+C++nR/Wtxh5QtmSZyiVK+n0/Znni+PjvhiEiynutDXbgRZx8TcyQirXD++VBYCH/6U9yRiIg0tHZtw2LgMZdM6d4dtu2xkP2Jcg0lctKETFetfBM4DLjLjJ2AD4BxwPmEVV+mtU94Il1bt4pPqbJu9Dzys3GHItJyRUVw5JHwz3+G0hhbb938a0REciW1GPjvfx8mqcVswA5l8Fb0REXBpQmZ9sjdHt33Aa4AHozu+0Xbb2vJSc3sTDP71MwqzWyamR3UzP5mZuea2YdmVmVmi8zsuqT2CWbmaW47tCQukY5mzPwpLO69ffhALJKPbr01DFO65564IxER2VJicZNEL9wOO8Buu8UXT+SGP5RtfqIeOWlCRomcO38jLNeZbtWXG9x5JNMTmtmJwC3ANcB44FXgSTMb2cTLbgDOBC4m9AROBF5Ks99OwNCk20eZxiXS0cz8oJ5d696isE7lGiWPjRwJhx4aErrq6rijERHZLLlHbvXq8HsqxtID/1NevvmxEjlpQsZ1Bt25gLA65dWEVV+uBvZ156IWnvN84B53v8PdP3D3s4FFQNrVY8xsLHA2cLS7P+rus939LXd/Is3uS919cdKtroWxiXQYr90+nQKc9aN3jTsUkbbZZZfwYeSGG+KORERks+Ri4LNmhQWa3o2/RvStfxnAchtEDUUaWilNalHBeHfecOen7pwe3b/RktebWQmwJ/BMStMzwP6NvOxoYDZwpJnNNrM5ZnavmZWl2XdqNOzyOTP7XEtiE+loip/+BwB9j9L8OMlzl18ePizd1qJR+CIi7Wv+/ND7VVLScJhljFauLmCwL2NZ4RD1yEmTMp54Y0ZvwpDGUUD31HZ3rszgMIOAQmBJyvYlhMVU0hkTnfNrwCTC4iq/BB43s8+4ez2be/TeAEqAk4HnzGyCuzcYgmlmpwOnA5SUlGQQtkjubTX3FQCGnnFUzJGItFGfPrDrrvD22/DRR7DddnFHJCLSsIYcxFoMPGHIkHC/pH4wQ5cvJ75iCHnI7PkW7O24H9puseRARomcGXsDTwBNVUjMJJFLSC1Tb2m2JRQA3YCT3X1WiMdOBmYCewOvu/vM6HnCa2Y2GriANHPp3H0yMBmgtLS0sfOKxGp05QfUUkjRqFFxhyLSdpdeCscfDxdfDH/7W9zRiIiE5C3xxVJFBXTrBgMHxhsToZPwEn7OQF9G/eISCuMOKL9MoPGcIllTuUfeyHRo5c3AQBoudpK86EkmlgN1wJCU7WU07KVLWATUJpK4yEdALdDUAimvA/raV/JSdTVUU8zsss/EHYpIdhx3HPTsCc89F4qEi4jErYMVA08oL4dxfEB/VlO/VEMrW6GxfKU1uUuHlunQyl0JWeuLwMPABlqRxbp7tZlNAw6HRMl6iJ4/3MjLXgGKzGwbd/8k2jYmin1uE6fbnZAEiuSdWdM3Mo45vH/AN+MORSR7Lrss9MhNmwZ77RV3NCLSla1dG26JoZS33w4rV8YbU2TECPhwQBklK6uwlUrkWsS9Ret/5LtME7nVQE/gq+6sbuM5bwT+YGb/JSRp3wOGEdWqM7NrgX1885jVZwkFye8ys3OjbTcTetymRq85F5gDvEeYI/dN4CvAsW2MVSQWM254mp2pp9euKqAsncjpp4dk7s47lciJSLwWLAj3iUSub99w6wCGDYNDvlZGt1trYP2aMExHazpIGplmrfdF9zu39YTu/gBwLnApMB04EJjo7onetaHANkn71wNfApYS5rs9DVQQyhHUR7uVEBZAeQd4OTrmF91dEzEkL/V+9u8ADNuvqdHDInmmXz/YYw+YPFkrsYlIvJJXqayrg5/8BP7zn3hjStJjVNLi7CtWxBdIZ2C2K2bHY/atBrc8l2mP3BxgDfCoGb8nLCxSk7yD+/+SvWa5+63ArY20TUqzbRFwfBPHux64PtPzi3R0o1dPx4Fuhx0cdygi2XX88fDqq2Hxk9tvjzsaEemqklepXLoUrr02PN5vv3jjitzw0CjOYCtGMT988TV0aNwh5R+zXsAjQGMlyRwyz186okx75H4H9AX6AT8krPh4d9LtrvYITqSrGlE/lw3WG4oyrhAikh9+8INwXT/wQNyRiEhXlkjkhg3rUKUHEl7pcRinWJRjqCh4a10KHEInXvSkJRMCu8TqLyJxe+etOvqwlhW9VHZAOqGiIpgwAVavhudbUu5HRCSLEsXAu3XrUMXAE4YMgVWFg8ITDUVvraMJvW5PRc+dsFbHcmAWcEVMcWVNpl/3n9KuUYjI/7x7x2vsCmzcfre4QxFpH9deC88+Cz/9KRxySNzRiEhX1EGLgScMHVTDBbWnhidK5For8Y34aYT1NcD9AsweJiy4uDimuLImo0TOnXvbOxARCXquD0Moyn7ynZgjEWkne+0FO+4IH30UFhkoVLlbEcmxigrYJlpbb/Hi0DM3aFC8MSUZPKyYcbwfnmhoZWslRgwuJtSfLsSsD/BWtP1CwnSxvNWlai2I5IP6t96mHmPgEfvEHYpI+7nyyvDh5Omn445ERLqi5GLgV18dfh91gGLgCePHw4qicqopVo9c6yUKA/YGEtnwr4GbosfDch5RlmWcyJlxshlvmrHBjLqUW217BinSlez0/oOsKhwMpaVxhyLSfr785VCO4JJL4o5ERLqadetgzZrNQynNoHfveGNKMXEiWFkZNRThSuRa6+PofiSh/rQRak2fTpgv90FMcWVNRomcGScA9wK7AT3QYici7WL1atim/iNqC7vFHYpI+yopCcOapk+HD/L+b6mI5JPUYuDnnQcPPhhfPI2oGVAGGLWLNLSylR4AniHUqP45UMnmvKUKuDi+0LIj0x65s6L7TdG9A4nqhKuBuakvEJGWe+HeeRRTy6ohO8Qdikj7+9nPwv2PfhRvHCLStSSvUllfD7/9Lbz1VtOvybFFi+APM/dhFf2pW6IeuVZxvxX3L+D+DO5vAjsB5wFnAzvj/ly8AbZdponcroTk7bDEBncGA5cRCoN/OfuhiXQ96//8GACFB+4fcyQiOXDUUdCrV5gnV18fdzQi0lWkFgOvqelQK1YC9O8PV9RcwtMcga1QIpcV7p/ifgvuv8V9dtzhZEOm5QcSk3XeJCR0mFEI3ECowfAr4NCsRyfSxZR/+CIAo876YsyRiOTIMcfAH/4Ad98N39FKrSKSA4lEbvhwePfd8LiDJXLdu4dpe8vXDaJo1TJw71CLseQFs581u4/7lTmIpN1kmsitBfoTxpSuI6z+8gVgTdS+b/ZDE+l6Bm+cSx0FlOy3Z9yhiOTGNdfA/ffDH/+oRE5EcmP+fCgrCyUHEkldByoGnnBsr6c5fd1kCmurYf36DrcgSx64nKgDqgl5nchlOrRyYXRfxuYVXh4FpkSPV6a+QERaxh2od2YNmwAFqgwiXcSIEXDyyTB1KmzYEHc0ItIVJBcDX7curBLdwXrkAPoOKqY/q8MTrVzZWukWaOw0CzVm+mnxLcIPvC9wHw3fBBUMF2mjBXNr2aFuBlVjd4s7FJHcOu208GHq7rvjjkREuoLkRO7kk8Pvn8GD440pjX2+XL75iYqCt8bnUm6HA2cAHwIbgBPiCy07Mk3kziQs3fmEO7cBPwHeBqYBlxK6LkWkDV669Bl6UElRn55xhyKSWwceGL4RV005EcmF5GLgEOaedcD5Z18/t2zzE/XItZz7iym353C/g5DUdQeOiDnCNssokXNngztL3NkYPb/OnfHu7O3ONe7UtW+YIp1fz5eeAmDoEbvGHIlIjpnBQQfB2rVhBUsRkfayfn0o2prokTv33DBXtyMaMID6xEd1JXLZtJpQR+6YmONos0YXOzFjZEsO5M68tocj0nWNXPIGDgw8eWLcoYjk3nXXwVNPwWWXwRF5/yWpiHRUqcXAH3sM9u+YJX/uvq+QHhzH13hQQytbI/2qld0JQyx7EpK5vNbUqpVzaH6llwRv5lgi0oytqj+miu5079Ur7lBEcm+33aC8HN54Ayorw9rbIiLZlloMfMGCDrnQCYRpe1/mL3yVv1PSyXrkzOxM4ELC1K33gHPd/eVG9p1AKOS9D9AX+Bi42d3vauY0l9N4LuPAiy2Nu6NpbmhlUyu9dLqVX0Tisnw5DGAlK7oPizsUkfh85zvhg9XVV8cdiYh0VsnFwJctg+rqDpvIDRkCYCxnIPXLOk8iZ2YnArcA1wDjgVeBJ82ssdGA+wPvAscBOwO3AZPN7OuZnK6R24uENUDyWlO9aFqJUiRH3nlqAYdQz8pt92V43MGIxOWSS+C3v4Xp0+OOREQ6q+Ri4O+9Fx530ESuvBxu5fsMYBU1C5bRLe6Asud84B4PC48AnG1mRwLfB36curO7p05ivM3MPgccC/ypifN8Ls22KmAu7otaHnbH02gi584puQxEpCvbNC2UZxx4kQoiSxfWsyeccQbccAMsXpz4OlpEJHvmzw9jFrt3h6oqGDcORo+OO6q0ysqgim4UUUvVkuWdIpEzsxJgT+CXKU3PEHreMtUHqGhyD/e8HzrZnC4/r23AgAFMmTIl7jCkixtS8SAAn/ZZzyxdj9KF9Rw3jn3q6lj6ta/x/uWXxx2OiHQyu7z9NiX9+jEt8bf21lvDKpYd9G/vQceup+jhOiqXzMmXz6tFZjY16flkd5+c9HwQUAgsSXndEuCwTE5gZl8CDgUOaGbHFi3ciHveLdxo7pmtZ2LGWEIRvbFAj5Rmd+fQLMeWE6Wlpb5hw4a4w5Aubk7BGIb7fIq9Ju5QROLXty9s2hQWPSnItNypiEgGdtsNRo0Kq1XmgQVX/p7hl51GVc++dNuwOu5wmmVmG929tIn2YcAC4ODkxU3M7DLgJHffoZnjHwA8CVzs7rc1E0w9LVm40T3vOrgy+gtpxp7AVOD/gCOBzybdJkQ3EWmF+noo90WsLewfdygiHcOxx0JNDfzud3FHIiKdTXIx8AsugG98I954mlE4NBQFL9m4BmprY44mK5YDdUDq2PkyGvbSbcHMDiQkcT9rNolLelkLbnkn0686fwKU0kl+aJGO5NWn1tKdSlb02zbuUEQ6hkRx3ptuijcOEelcNmyAVas2L27yxhubyxF0UNf9fSwvcWD4wL1qVdzhtJm7VwPTCLXckh1OWL0yLTM7mJDEXeHuN2d4upeApdHj+cB/onui7S8m3V7K8JgdSqaJ3P6ErsnvR88d2BV4DJgF7JH90ES6hvm/ewIDqnffO+5QRDqGIUNg++3ho49g4cK4oxGRziK1GHhFxebeuQ7Kt9ue2zgrPOk8RcFvBCaZ2WlmNs7MbgGGAbcDmNm1ZvZcYueojtyTUfsfzWxIdBvczHmuBQYA5+E+Cvf9cR8F/BDoD1yH++f+d8tDmSZyA6P7PyY2uDMDOB3YnlCkT0Raoc8bzwIw/NQjYo5EpAO54IJw/4c/xBuHiHQeqcXAKyo6bOmBhPJyWEE09aKTFAV39weAc4FLgenAgcBEd58b7TIU2CbpJZOAnsAFwKKk2xvNnOpawsIqd6ZsvwMoJtSxy2uZJnKbovvKxONo8ZPiaPtRWY5LpMsoqdlADUX0Py6jxZpEuobTTgvLgj/6aNyRiEhnkVwMfPnyDl0MPKG8HB7muPCkkyRyAO5+q7uPdvdu7r6nu7+U1DbJ3UenPLc0t9Hpjp1kXHR/dMr2xPMd2/pzxC3TRC4xvnQAMCd6/ALwWvS4PosxiXQpvdYvYU7/PaCkJO5QRDoOM/jOd+C11+DFTl8KSERyIbkYeGUlHH447NixP8uHHrloYFznGVqZK4k6c3/AbCpmjxBKI9xHmCbWsSdIZiDTRO7d6H5X4B+ERU7KgeGEN+KZ7Icm0vktW1LPjpVTWTdq57hDEel4jo6+NP3hD+ONQ0Q6h/nzYdAg6NEDRo6EZ56BQzt29azdd4eNvaMFHjtRj1yO/ILNCzOOB74c3ScWbOwyQyuvAL5O6I37OSFxS7wxzxHKEohIC71801T6so7iqnVxhyLS8Wy7LQwdCm++CRs3xh2NiOS7PJgTl2rECCgcXk4dBdQtUSLXIu53At8FFrPlivsLgVNwvye+4LIjo0TOnbfdecCdj91Z586RhGGWfd35vDvq6xVpBX/8HwCUTpwQbyAiHdXpp4M7XHFF3JGISL5LTuQuvzzMw/VM60XHZ12PchyjaoE+breY+++BEYT5cAcR5s1thfu9scaVJW2pYF4CbMhWICJd0dA5YZrp6HO0XpBIWj/6EVx1FVx/Pdx6a5hLWlMD69c33Le0NLRXV4d6Ual69YLi4sbbe/eGoiKoqkrfA9inDxQWhrk1mzY1bO/bFwoKQltlZcP2fv3C3L+NG8M5UvWPVqbbsCHEmMwsvL6x9oKCcH4I701NTePt7uF4Il3NqlWw//7h8SefhP+refB/4boPjuYmnqTvYvXItYq7Ax/GHUZ7aDKRM2MP4GtAd+ARd5434zQ212WoNOM2dy5o/1BFOp+Rmz6khiKKR+bXUA+RnOneHa68Eh54AHbYIcz8X70a3n674b477RTmvyxfDu+917B9t91CMrRkCXyY5m/6HnuEZG7RIpg1q2H73ntDz55hns3s2Q3b99sPunWDOXNg7tyG7QccEBLFTz7ZvOhCss9+NtzPmhViSFZYCAceGB5/8AEsXbple0kJfOYz4fGMGbBixZbtPXrAPvuExRKefhqOOCK8VyJdSUFB6OWHvKghlzC1/Iu8M3dXDly2JO5QOj6zuwDH/TtJ28K35e6PJW17C6jHfc8cR5hV5o10KZtxIGH+W3Ky9/+AiwgLnCS+wnDgLPdQxC/flJaW+oZ038yKtLPaWqgvLmF10SDKalT0WERyYM6ckBDW1cErr8CYMXFHJBKPbbcNX878+c9xR9Ks/feu4bypJzFxwOuUrujYCy2a2UZ3L40xgHpCIleYsq0e96Im98tDTc2Ru5BQJy55cuCFUZsBy5Men9ySk5rZmWb2qZlVmtk0Mzuomf3NzM41sw/NrMrMFpnZdSn7fDY6VqWZzTaz77UkJpFc++SdDRRRQ8UuX4g7FBHpKkaPDiv1VVeHpddTe/5EugL3vOqRO6z7yxzPw3RbqzlybdDxx9C2QlOJ3F6E3rangTOBJwlvggMnuVMGfCPaN+MiHGZ2InALYcnP8cCrwJNmNrKJl90QxXAxYZLiROB/hQPNbGvgiehY4wlDP39tZsdmGpdIri146l0KgD5f/3LcoYhIV7LjjvDEE2GI6RFHhHlDIl1JVRV84xub58t1cN1HlgNQVNvI/F3pspoaWllFGFbZ3521ZvQFVhESue7u1JhRAlQC9e6ZLZxiZq8D77j7d5O2fQQ85O4/TrP/WGAGsKu7f9DIMX8BfNXdt0vadiewk7t/pql4NLRS4vLoLpdw9IxrqJzxMd132ibucESkq3n22bCAzP33h7l/ItIhvfv8MnY5tCw8mTs31MDroDrw0Mrmt+WhppKvYsI6L2sB3FmTWNjHnZrovjrallF3pZmVAHsCv0xpegZo7GuRo4HZwJFm9k9CL+KLwIXunpjt/RkaFiV/Gvi2mRW7e8ryXSLxG/vBI+FbkXFbxx2KiHRFhx0WbgBr14aFZUpK4o1JJBdqasICQgWZllOO1y6fHUA9BRRQHxZz6sCJXIdh9rOMtuW5ZnvRzGjwQ6fblqFBQCGQuuzOEuCwRl4zBhhFWD1zEqFH8JfA42b2GXevB4YAz6Y5ZlF0zi0mAZjZ6cDpACX6oyUxGVY3n3XWhz558odERDqpqir43Odg++1D71xhXn9BLdK8O+6A886DefPCSrgd3Jr1hVDYl751q0IiJ5m4LOmxp9nWKWQyHLI93ojU8ZyWZltCAdANONndZwGY2cnATGBv4PUmjpluO+4+GZgMYWhlS4MXaau5n9QyknV8WroLfeIORkS6tm7d4MQT4eKLQy273/42L2pribRaRQXU1+dNCY6PP4aH6y7gGi4JJUSkOV3mF1hziVy234jlQB2hBy1ZGQ176RIWAbWJJC7yEVALjCQkcosbOWYtkFJMRyR+b970IqOAtdvsEXcoIiJw0UWh9tz118PAgaEIu0hnVVEBw4fnTe9zeTn8jjNCIqceueZcEXcAudRUIpf1N8Ldq81sGnA48NekpsOBhxt52StAkZlt4+6fRNvGEGJPVFx9DfhKyusOB6Zqfpx0RAXPPAVAv68eEnMkIiKR666DlSvh5z+HESPgjDPijkikfcyfH67xPFFWBiVUUYdRv2g5xXEH1JG5K5EDcG+3jPZG4A9m9l9CkvY9YBiEguJmdi2wj7sfGu3/LPAmcJeZnRttu5nQEzc1en478AMzuxn4HXAAYT7dSe30M4i0SWGPEuowRp59dNyhiIgEZnD77dC3L3xB9S2lE6uogD33jDuKjJWUwKVF12O1UDl/mRI5+Z+cr7Lg7g8A5wKXAtOBA4GJ7p7oXRsKbJO0fz3wJWApoXbc00AFcHTUhrt/Sqgtd3B0zEuAc9y9sV4+kVj1mfsuFT13oKB/37hDERHZrLAQfvnLsCpefT1Mnx53RCLZd+qpcGx+lRpeX1pGAU7NgsZmIgkAZpdh1q8F+/fDLG8XQWm0jlxXoTpykmuVlVDVoy/zy/Zk5yXPxx2OiEh6V14J11wDTz0FEybEHY1Il/bSt3/Pwfedxqqx+9L/w//EHU6jOkgduXXAQ8ADwCu4b0jZp5Qweu9rwLFAr3ytJ6d1z0Vy7JX7P6Uva/GSbnGHIiLSuLPOgjFj4KijYNq0uKMRyY6NG2Hp0tDjnEe2PygUBC9csbSZPbu894HehClWTwKrMfsIs9cwexWzj4DVUdu3o33fjynWNlMiJ5Jjq+97FADf/4CYIxERacLAgfDMM6EkwZFHwsyZcUck0nbPPx+WgXzjjbgjaZGVhSGRK1m/MuZIOrxdgTOAjwmr7xcSpmztA+wbPS6M2mYT1urYLZZIs0CJnEiODZjxEgBjzvlyzJGIiDRjxAj417/CQihHHQW1tXFHJNI28+eH+622ijeOFnrq07E8wtEUV67Lu97EnHKvx/0O3McChwC/AJ4HPgA+BF4ArgcOxX073CcTrbmRjzIpCC4iWTRy7QzqMXrtv2vcoYiING/77eHpp2HdOijSxwbJcxUV4TouL487khbpN7ofU5jAV3gUVq+GAQPiDqnjc58CTIk5inalHjmRHBtUt4Q1Bf3DN9wiIvlg/Hg4+ODw+O9/D0mdSD6qqIBhw/KmGHhCeTn0Y1V4oqLgElEiJ5JDyxfV0I0qPtr/23GHIiLScnPmwAknwNFHhyV4RfJNnhUDTxgyBM5gcniiRC4zZvtgdiZmn4ueH47ZB5htwOzhaPXKvKZETiSHPv7nTLpTRY/994g7FBGRlhs9Gu6+G154AU46SXPmJP+cdRacd17cUbRYeTksZ1B4smxZvMHkj4uAXwPbY1YM/BHYHugBfAXI2/pxCUrkRHJo+Q33AFC298h4AxERaa1vfhN+9St45BE4/XTo4vVoJc8ceywcd1zcUbTYkCFQMHQIAL5MPXIZGh/dPw/sCQwCFgPTCKtWHh1TXFmjRE4kh7b69CUcKJ+4V9yhiIi03tlnw2WXhd65Z5+NOxqRzFRWhpqIeTjHs6gIfHgYElq1QIlchhIr2swHdo4eXwd8KXo8KucRZZkSOZEcGl41m0p6QM+ecYciItI2l10GL70Ehx8edyQimZk5E/baK6zCmocq6ofhwKZ5KgqeocRwgZ6ERM6B9yCxagx1cQSVTUrkRHJk40bozyqWdxsedygiIm1nBgcdFB7/5z+hd06kI6uoCPd5VkMu4daNk1jKYGoWaY5chhZE948DiVXm3gOGRY/zvmtTiZxIjrxy2zsUUs+KYTs3v7OISD656Sb4znfgwQfjjkSkcYli4Hm4aiVA9ajtqGAr6petjDuUfPF3wly4/YC+wBu4LwH2idrfiSuwbFFlT5Ec2TTlPwAUHPq5mCMREcmyu++GhQvDQij9+sHnPx93RCINVVSE+nFDhsQdSatsN2gVRdRQtLgi7lDyxRVAH+Ag4FPg/Gj7SOA54C8xxZU1SuREcqRw7WoAdrz6m/EGIiKSbT17wuOPw4QJcMwx8NxzsN9+cUclsqWKChg6NO+KgSfs1G8Bu/Eu61eVxR1KfnCvBM5Ks/2XwC9zHk87UCInkiPdZ05ncfFWDCkbEHcoIiLZ169fWETiwANh8mQlctLx/OAH8NWvxh1Fq/XdLiRwJZvWxhxJnjHbBTiMUH5gOfAs7u/GG1R2mHfx+i+lpaW+YcOGuMOQLmCj9WBFyTC2qvok7lBERNrP4sUwaFBYL/2llxrWmRs2DLbbDurr4eWXG75+q61gzBioqYFXX23YPno0jBoVlpJ//fWG7dtsE+ZAbdgAU6c2bB87NgytW7sW3nqrYfuOO8LgwbBqFbyTZgrNLrvAgAGhKPP77zds33136Ns3vA8zZzZs33NP6NULFiyAjz9u2L7PPtCjB8ybB59+2rB9//2huDi0zZvXsP2gg6CgIBx7wYIt28zg4IPD4w8/hCVLtmwvKoIDDgiP33sPlqesBdGt2+YE/Z13wnuUrGdP2Hvv8Pitt8J7nKx3b9hjj/B46tTwb5SsXz/Ybbfw+PXXw79xskGDYKedwuNXXmlYkL68HHbYITxu6trLU+vX1FHarxjDw3vTrVvcITVgZhvdvTTuOAAwKwLuBE5O03ofcBru+b1ypbt36VvPnj1dpL29+/Iqrwd/b8ABcYciIpI73bq5h4/Tm29nnRXaqqsbtoH7xReH9hUr0rf//Oehfc6c9O233BLa3303fftdd4X2115L3/7Xv4b2Z55J3/7kk6H94YfTt//736H9nnvSt7/9dmj/9a/Tt8+eHdqvvTZ9+7Jlof0nP0nfvmlTaD/77IZtRUWb/20mTWrY3r//5vZjj23YPnLk5vYjjmjYvtNOm9v3379h+777bm7fZZeG7Ycdtrl99OiG7cccs7l94MCG7d/6VmbXXh5bX9Ar/CwLFsQdSlrABo/xc/0WN7jOob6RW53DtbHH2MabeuTUIyc58Mjx9/OVh05m6mfPZ68pN8QdjohIbrz4Yuh5SzZ8OGy/fdj+4osNXzNyZOhVq6mBf/+7YfvWW4deucpKeO21hu3bbRd65NavhzfeaNi+ww5hntSaNfDmmw3bd9oJyspg5Up4++2G7bvuCgMHhh65GTMatu+xR+iRW7Qo9Hql2nvv0CNXUQEffdSwfb/9Qo/c3Lkwe3bD9gMOgJKS0DZ3bsP2gw8Oc8BmzUrfIzdhQnj8wQeh1zBZUdHmkhIzZoSfMVm3bqFHEMJ7szJl9cTS0tCjCKHwdmqPXJ8+oUcS4L//bdgj179/6NGE8G+brkdul13C43//O1wjyYYMgXHjwuOmrr08tXEjrO47imG188L7v+uucYfUQAfrkVtIKAq+jNAzN4+w0MlpQBmwBPeh8QXYdkrklMhJDjwz/BQ+v/AeFt7zDMO+reK5IiIi0jI1NXBayT3cyynw7LNw6KFxh9RAB0vkNgDdgX1wn5a0fW/gdWATHSXWVlIdOZEcGLHsTRwY9g2VHhAREZGWKy6GGd3CHMT6pXlfyzoXEpNkU7u+E5NX/5vDWNqFEjmRHCipr2Rp4dAwbEVERESkFbbrEWrIbZqvRC4D5wHrgZ9j1h0gur8KWMvmunJ5S58qRdpZVaUzqG4Js8Z/jfK4gxEREZG8dUBx6ETaNHsReT0msL2YpU4sNUItudMxWwEMBIqBDcBDwDa5DTC7lMiJtLMZf5vFnqyhcM/d4w5FRERE8ljVgKGwDGrnzI87lI5qNOCEBI6kxyXA0KRtvSD/c2ElciLtbOH197En0K13cdyhiIiISB47+rtlcAFhVVRJZx4hUesSlMiJtLPyT17Dge3P/WLcoYiIiEgeG7BDGQBFK5Y0s2cX5T467hBySYudiLSzERtmUkMxJSOHxB2KiIiI5LH5VSGRK1yzspk9pStQj5xIO6qvh0G+lJVF5SiNExERkbZYVDySmZzAYTUvxx1Km5jZmcCFhHlr7wHnunvaH8rCSpO3A3sA44BX3H1ChicqAiYCY4EeDdrdr2x59B2HEjmRdvT2UwsZTy1LB+ygRE5ERETapHx4EW+zNX2q/w7uYNb8izoYMzsRuAU4E/h3dP+kme3o7vPSvKQQqAR+Q0jK+mV4ojJgCiGJa0xeJ3IaWinSjta88h4A1UceFXMkIiIiku/Ky2FbPqLYa2DdurjDaa3zgXvc/Q53/8DdzwYWAd9Pt7O7b3D377n7ZKCiBee5AtiBsGplulveUyIn0o4q3/0IgO0v/Eq8gYiIiEjeKyuD8UwPT5bnX1FwMysB9gSeSWl6Btg/y6f7PGEFy7uj5w6cA3wEzAK+k+Xz5VyXH1o5YMAApkyZEncY0kmNnf9nqrr15M1ln8CU1BqVIiIiIi0zdmQJzINpTz3Fuh13jDucVEVmNjXp+eSoJy1hEGGoZOqym0uAw7Icy/Do/kfAKQC4/wazF4B3gRFZPl/OdflEbuXKlUyYMCHuMKSTWj/9LSqthAmf+1zcoYiIiEgn8GH37RnKh+w+YhSFHe8zbK2775XBfqm13izNtraqA4qBFUANUITZYGBu1H468PMsnzOnNLRSpJ0sX1hNKRtY0nN03KGIiIhIJ1FTHjqa1s9eGnMkrbKckGClrgFXRsNeurZaEd33BRZHj/8I/Cl63D/L58s5JXIi7WTaL5/HgNVbj487FBEREekkZlaOAmDTrPkxR9Jy7l4NTAMOT2k6HHg1y6ebGd1vA7xE6PU7FPgioffvzSyfL+eUyIm0E3/yaQB6fvnQmCMRERGRzuJvw8+hmiIqV26MO5TWuhGYZGanmdk4M7sFGEaoFYeZXWtmzyW/wMx2NLPdCXPsepnZ7tHzptwBTAa6E1awXMbmFSuXA+dm7SeKSSxz5FpYBHA08Gmapi+4+1PRPhOAF9LsM87dP8xCyCItVj7vvziww/lfjDsUERER6SSGjunBcgZTvTD/Vq0EcPcHzGwgcCkhF5gBTHT3xNy1oYRetGRPAKOSnr8V3TdeRsD9QeDB/z032w74HFALvIL76lb/EB1EzhO5VhQBTDgSeDvp+co0++yUsn1ZG8MVabXqoh4sKB7NiEH94g5FREREOolxg5ZRQjUlc2Y2v3MH5e63Arc20jYpzbbRWTjpWuDRNh+nA4ljaGWLigAmWeHui5Nu1Wn2WZqyT13WoxfJgDuMWPchC0cfEHcoIiIi0omMGukMYgU9Vy2IOxSJWU4TuTYWAfybmS01s1fM7LhG9plqZovM7Dkz03rvEpt3HpnNcF9A7bCt4g5FREREOpHyHQdSj9G9ak3coUjMct0j11QRwNRlSBPWAxcAJwATgeeAB8zsm0n7JHr0jgW+Slil5jkzOzjdAc3sdDObamZTa2trW/uziDRq/q//DkDRgD4xRyIiIiKdya7jC6m27nSvWx93KBKzuAqCZ1wE0N2XAzckbZpqZoOAi4D7o31msnmJUYDXokVSLiAsN5p6zMmEVWwoLS3NdvFBEfq9G9buGfN/X445EhEREelsNhX1pn/NUqithaK4Ps5L3HLdI5etIoCvA9tlYR+RdjF81QzqKGDQZ3eOOxQRERHpZFb4gPBgZbq1/6SryGkil8UigLsThlO2dR+RdjGkbgGrbUDcYYiIiEgndG3JZeHBMi3Q3pXF0Rd7I/AHM/sv8ArwPVKKAAL7uPuh0fNvAzWEehH1wJeBs4CLEwc0s3OBOYSadCXAN4GvEObMieTUkk83UkYln/bejYFxByMiIiKdTnXfwbARqhcup2SnuKORuOQ8kWtlEcBLCUUA64BZwKnufn9SewnwS2A4sImQ0H3R3Z9otx9EpBGLXv6YcmDdiafFHYqIiIh0QruUfAjAulkLGZg6zk26jFhmR7akCKC73wvc28zxrgeuz1Z8Im2x4vlQt36rE5qrqCEiIiLScoN6bgKg8oNPY45E4hRHQXCRTq33I/dRSyFDP7t93KGIiIhIJ9Rth60BqJ49P+ZIJE5K5ESybNs106imG1as5YBFREQk+w4+aTgAhYu1rl9XpkROJItqq+vpx2qWloyIOxQRERHppPqNLQegaEVLqndJZ6NETiSLpt45nQKcZUN2iTsUERER6aQW15dRRQm1q9fFHYrESImcSBatvv8fANQf9NmYIxEREZHOqrSslGc5jI113eIORWKkRE4ki3rOfR+AsRceFXMkIiIi0lkNHgzLGEyfKhUE78qUyIlkkQELikfRb7dRcYciIiIinVRxMezMDAbXLow7FImREjmRLBq69G0Wlu0edxgiIiLSyfUu2EgxdbBxY9yhSEyUyIlkyfy3lrNN7YfUFxbHHYqIiIh0cht7DgwPlml4ZVelRE4kS2b+8nEMqN1q67hDERERkU6u5+hQgmDT3KUxRyJxUSInkiXF/34BgEHfPDLmSERERKSzqy0PRcHXvjs35kgkLkrkRLJkyOLpOLDdqQfHHYqIiIh0cm/X7gjA2g+14ElXpUROJEuGVM9lHb0pKCmKOxQRERHp5ObtczwAy5fWxxyJxEWJnEgW1FbX04NNzOu9U9yhiIiISBcwdFw/aimksmJ53KFITJTIiWTBnClzKKGGtceeEncoIiIi0gXsMGwtjlH+8StxhyIxUSInkgULHp0KwKBDdos5EhEREekKthrXiyJqKV2/OO5QJCZK5ESyoPcfbsWBEYeOjTsUERER6QIGDy2ihmJKa9fEHYrERImcSBYMWz+LakroOaxf3KGIiIhIF1BcDNUFPehZtz7uUCQmSuREsmCgL2NFYXncYYiIiEgXUlXci+K6TXGHITFRIifSRrNfnEcxtSzuv0PcoYiIiEgX8pFvQyG1UK8SBF2REjmRNpp902MAbNht/5gjERERka7khf7HUgD4ylVxhyIxUCIn0lZLlwIw/MyvxBuHiIiIdCm1/QYBsH6Oasl1RUrkRNqoePlCVtpAxhyj0gMiIiKSOzsVvA/A8hffizkSiYMSOZE2Gjr3Neb13RnM4g5FREREupCi8oEArJ7+acyRSByUyIm0wcaVlWxX/T7dqtfFHYqIiIh0MYP23hqAkkXzYo5E4qBETqQN3vp/z2LA8lF7xh2KiIiIdDHbHLk9AN1XLYw5EomDEjmRNqj+xzMAdJt4WMyRiIiISFfTb7dRABQsWxJzJBIHJXIibTDo0zdwYJeLJ8YdioiIiHQxhX1KWU9PVqwujDsUiYESOZE2GLrxYyrpTo/BveIORURERLqYomJjJjuwrq5n3KFIDJTIibSFOx8MOCDuKERERKSLWlkwmL5VS+MOQ2JQFHcAIvlq6YyllLGC9w7SsEoRERGJxw58QFnd4rjDkBioR06klWbe9AQApePHxhyJiIiIdFWbCntTTE3cYUgMlMiJtFLJ4w8BMHj/bWOORERERLqqun4DKcCp31gZdyiSY0rkRFpp6Mr3qaOAUYerR05ERETiUTi8HIA1b34ccySSa7EkcmZ2ppl9amaVZjbNzA5qYt/RZuZpbkem7PfZ6FiVZjbbzL7X/j+JdGWD6xax0gbGHYaIiIh0YTWDhwOw7LX8SeRakgtE++9iZi+a2SYzW2BmPzMzy1W8HVXOEzkzOxG4BbgGGA+8CjxpZiObeemRwNCk2/NJx9waeCI61njgWuDXZnZs1n8AEWDN/LV0p5LFpRpWKSIiIvGZ0293AGa/tyneQDLU0lzAzPoA/wKWAHsD5wAXAufnJOAOLI4eufOBe9z9Dnf/wN3PBhYB32/mdSvcfXHSrTqp7XvAQnc/OzrmHcC9wAXt8yNIVzfj+icwYMW2+8QdioiIiHRhRZ/ZG4Bly/Kmg6qlucA3gJ7At919hrs/DPwCOL+r98rltPyAmZUAewK/TGl6Bti/mZf/zcy6Ax8BN7n7Q0ltn4mOkexp4NtmVuzuHXYpn4Wvz+eTyc9St2rtFtutezf67jsOgLVvfkT9ug1btBeU9qDPXmFu1tqpM6nfsOW3MAW9S+mzx3YArHn9A7yyaov2wv596L3rmND+6nt4zZZvUdGg/vTaaRQAq196B7x+y/aygfQatxVe76x5+e0GP1fx0MGUbj+cuupa1r02o2H7iHJKtxlKXWU1615/v0F7yehh9BxVRu36StZP+7BBe7cxI+ix1SBq1m5kw1uzGrR3334U3Yf2p2rFOjbN+KRBe48dt6bb4L5ULVvDpvc/bdi+8zZ0G9ibTYtWUjVrXoN2n18BQNm532jQJiIiIpIrW40fBIBXVMQcSfNamQt8BnjZ3ZM/7D4NXAWMBhp+kOsicl1HbhBQSOgaTbYEOKyR16wn9Ky9AtQCRwEPmNm33f3+aJ8hwLNpjlkUnXNR20NvH7OueoAJ/7wwfeOfm3nx/c2039eaiLqIv7exHVht/Rh70h5ZCUdERESkNbbZoy8OjN7U8IvxDqg1ucAQIDVLXZLUpkQuxzzluaXZFnZ0Xw7ckLRpqpkNAi5iy1Qm3THTbcfMTgdOBxg+fDhTpkzJOPBsK/jaVkzd6gIKN2zZ4+bFxVSOGQ1At3kVFGzassetvlt3qkZvFdo/nUdB9ZY9bvU9elA1cgQA3WfPwVJ63OpKS6keMSy0f/wpVle7ZXvv3lQPGxLaZ32CpfTI1fbtS82QMnCnx6yGk2tr+vWjtnww1NXR4+PZDdsHDKB28ECoqaXH7Ib//2oGDaJ2YH+sqpruc+Y2aK8uK6Ouf18KNlXSbd78hu1Dh1DXpzcFGzfSbf6CBu1Vw4ZS37sXBevW021hwzy/aqsR1PfsQeHadZQsalhks2rUVlSPGUn9qy83aBMRERHJpW5X/RJGxfuZNlJkZlOTnk9298lp9ss4F2hi/3Tbu5RcJ3LLgTpC9pysjIaZeVNeB05Jer64kWPWAitSXxxdUJMBSktLfcKECS04dTv45onxnl9ERERE8lfcn2U3q3X3vZpob00u0NjnfJp4TZeQ08VOogVKpgGHpzQdTlixJlO7s+Vwyddo2B17ODC1I8+PExERERHpKlqZC7wGHBStlZG8/0JgTrZjzCdxrFp5IzDJzE4zs3FmdgswDLgdwMyuNbPnEjub2bfN7OvRvmPN7ALgLODXSce8HRhhZjdH+50GTKLhREoREREREYlPi3IB4E/ARuAeM9vZzL4K/Ai40d01tDKX3P0BMxsIXEqoBzcDmOjuiYlQQ4FtUl52KTCK0BU7Czg1aaET3P1TM5sI3ERYunQhcE60PKmIiIiIiHQALc0F3H2NmR0O/BaYCqwirJ9xY04D74CsiyeylJaW+oaUhUZERERERKRlzGyju5fGHUdXEcfQShEREREREWkDJXIiIiIiIiJ5RomciIiIiIhInlEiJyIiIiIikmeUyImIiIiIiOQZJXIiIiIiIiJ5RomciIiIiIhInlEiJyIiIiIikmeUyImIiIiIiOQZJXIiIiIiIiJ5RomciIiIiIhInjF3jzuGWJlZPbApR6crAmpzdC7JH7oupDG6NiQdXRfSGF0bkk4ur4se7q6Oohzp8olcLpnZVHffK+44pGPRdSGN0bUh6ei6kMbo2pB0dF10XsqYRURERERE8owSORERERERkTyjRC63JscdgHRIui6kMbo2JB1dF9IYXRuSjq6LTkpz5ERERERERPKMeuRERERERETyjBI5ERERERGRPKNELkvM7Ewz+9TMKs1smpkd1Mz+u5jZi2a2ycwWmNnPzMxyFa/kTkuuDTObYGaPmtkiM9toZu+Y2am5jFdyo6W/M5Jet52ZrTOz9e0do8SjFX9PzMzONbMPzawq+v1xXa7ildxpxbVxhJm9Fv3OWB79fdk+V/FK+zOzg83sseizpJvZpAxeo8+gnYQSuSwwsxOBW4BrgPHAq8CTZjaykf37AP8ClgB7A+cAFwLn5yRgyZmWXhvA/sC7wHHAzsBtwGQz+3oOwpUcacV1kXhdCfAX4KV2D1Ji0cpr4wbgTOBiYBwwEV0jnU4rPmtsDTwKvBztfxjQA3giJwFLrvQCZgD/B2xqbmd9Bu1ctNhJFpjZ68A77v7dpG0fAQ+5+4/T7P994BdAubtvirZdCnwfGOH6R+k0WnptNHKMB4FCdz+2ncKUHGvtdWFmNwH9gBeB37h7r/aOVXKrFX9PxhI+xO3q7h/kLlLJtVZcG8cBDwAl7l4Xbfsc8Dww2N2X5yZyyZVopMYP3P2eJvbRZ9BORD1ybRR9Q74n8ExK0zOE3pV0PgO8nPgPFHkaGAaMznaMEo9WXhvp9AFWZSsuiVdrrwsz+yLwJcK3p9IJtfLaOBqYDRxpZrPNbI6Z3WtmZe0YquRYK6+NqUANcJqZFZpZb+DbwBtK4ro0fQbtRJTItd0goJDQRZ1sCTCkkdcMaWT/RJt0Dq25NrZgZl8CDkU1YDqTFl8XZjYUuAM42d3XtW94EqPW/M4YA4wCvgZMAk4GdgAeNzP9je88WnxtuPsc4HDgCqAKWAPsQvhCSLoufQbtRPRLPntSu6Itzbbm9k+3XfJfS6+NsJPZAcCfgHPc/b/tEZjEqiXXxf3Abe7+n/YNSTqIllwbBUA3QpL/kru/TEjm9iHMf5HOJeNrw8yGAL8H7iNcCxOAdcCDSvK7PH0G7ST0H7ntlgN1NPwWo4yG33gkLG5kf5p4jeSf1lwbAJjZgcCTwM/c/bb2CU9i0prr4hDgMjOrNbNawoez0uj56e0XquRYa66NRUCtu89K2vYRUAs0uXiO5JXWXBtnARvc/SJ3f8vdXwK+CXyWlg3vl85Fn0E7ESVybeTu1cA0wvCFZIcTVpRK5zXgIDPrnrL/QmBOtmOUeLTy2sDMDiYkcVe4+83tFqDEopXXxS7A7km3nxFWJ9sd+Gv2o5Q4tPLaeAUoMrNtkraNAYqAuVkPUmLRymujJyH5S5Z4rs9/XZc+g3Yi+o+cHTcCk8zsNDMbZ2a3ECaN3g5gZtea2XNJ+/8J2AjcY2Y7m9lXgR8BN2q1oE6nRdeGmU0gJHG3A380syHRbXDuQ5d21KLrwt1nJN+ABUB99FwL4XQuLf178izwJnCXmY03s/HAXcDrhMUupPNo6bXxT2APM7vMQv3JPYC7gfmEpFA6ATPrZWa7m9nuhM/1I6PnI6N2fQbtxIriDqAzcPcHzGwgcCkwlLAU9ER3T3wbOhTYJmn/NWZ2OPBbwh/aVYQ6QDfmNHBpdy29NgiLFfQELohuCXPRalKdRiuuC+kiWvH3pD5aFOlXhNpxmwg1os539/qcBi/tqhXXxvNRDdKLCHXCNgH/AY509w05DV7a017AC0nPr4hu9xI+U+gzaCemOnIiIiIiIiJ5RkMrRURERERE8owSORERERERkTyjRE5ERERERCTPKJETERERERHJM0rkRERERERE8owSORERERERkTyjRE5EupSoMO5vzOwDM1tvZuvM7EMzu8PM9kvab46ZuZnNiTHcRCz3RLG4mY1O2l5uZn80s0VmVhe132xmo5P2v6cd4+pnZpdHt69kGneumNmEpPM3d7s8ek3i+ZRcx9uc9vx3bcm/Vcr7mtU4REQkcyoILiJdhpmdAtwGdEtpGhvdBgNfyXFYbXELcGKM5+8HXBY9vhd4JLZIREREuhglciLSJZjZIcCdhJEIDlwN/A5YCowCjgO2jy3AJrj7JGBSmqY9o/vVwBh3X5XUZu0bVfOaiDtX559C0vtgZpOAu6On90bxZZ2ZdXf3yvY4toiISIKGVopIV3Etm3/n/crdf+ruFe5e7e4fufu1wHebOoCZ7W5mfzOzj81srZnVmNniaNteKftubWb3mdk8M6s0s9VmNiMawlaWtN93zWyqma00syozW2Bm/zKzbyfts8Wwt8TQNmDbaJd+wMqofVJTQ/DMbA8z+3N0nmozW25mL5jZPlF7LzO718zeNbMV0c+42sxeMrMTk45zOfBp0qG/nXrOJoaElprZFWb2npltMrONZvaWmZ1vZkVJ+23xc5jZt6L3cJOFobHfph2Z2SFm9p/ofJ+Y2UVmlpwYXp4U3zFm9nszWw5sStpnnJn9Ien9XmpmD5nZrinnyuh6SXnNCWb2TlPvh5kdZGaPmdmypOv1L6nnb+I9GBbFuz66Hm4Dejeyb4t/BhERaQN310033XTr1DegjNALl7gNz+A1c6J95yRt+1rKcZJvG4BxSfu+18S+O0f7HN/EPg8lHeuepO2jgQlNvG5StE/i+T1JxzkGqGnsddE+Q5o4tgPfjva7vIl97kkXd7StFJjWxGufAAqifZN/jlWN7H9gC66DSenel5R9Eu3LG3mvvpm07+Up+/9vv6j9QGBjI3FvAg5q4fWS/H4sbu79AL4J1DWyXyUwobFrLNrWA/ggzWsXpnsfM/kZdNNNN910y95NPXIi0hWMTnq81t0XtPI4bwJHAEMJ8+z6AN+P2noCZwCY2UBgx2j7rwjJywBgb+CnwJqo7eDofj1hjl43wjDPE4CnGgvC3ae4uwFzo01z3d2i2z3pXmNmPYA72Dyk/mdAOTCIkFDOjravI8y7Gx39TN2B/QkJCcB5UQyXA1snneLepBgmNRY7cC6wR/T4acJ7OYbw3gJ8gZAwp+oHnBnd/yJp+8lNnKstBgLXA/2BH2RwPgOOJLxnid6uOwjJ0FzCMNhuwHhgGeF9/S206HpJVk4T74eZlQK/JvRC1xKS+D7A96L9uhGGFjflW8AO0eP/ACMIvcCrG/zwrfsZRESkDTRHTkQkc4uB7wA3ExKdHintY6P7VYQPu/0Iick6Qs/G2+7+86T9P43uS4FLCT1VHwDPuHu2P/geQEhOAKa4+1VJbQ8lPd5ISO4eAMYRhtElz7cbS9t8Menxj919MYCZXcnmxVImAn9Ked00d78t2vd+4OJo+6g2xtOYJcDP3L3OzO4FftPM+W5w96ejx++a2XZsToJGEf5tU+1iZkMI8zRX0/z1kqy59+OA6HgAT7j7I9Hj35nZ94Ddge3NbFt3/7iRcxyS9PjaxBcgZnYDYb5pskyveRERyRL1yIlIVzAn6XEfMxvWyuM8CFxESHBSkzgS29y9ntAzUgFsB1wC3E/4gP+umW0V7X8r8Fcgsf/NhF6qJWb2o1bG2JjypMfvN7HfxYSeon0JPTipi6Z0b2Mcg5Mez0t6PDfpcbr5VDOTHm/IYjyN+cTd61pwvrdSnmc6J2xgC66XZM29H429z9D8e/2/2JIeVzTyGGjRNS8iIlmiRE5EOj13Xwr8N2nThen2S15oI01bf8KwSgi9NTsBhWweRpd6zn8AIwk9WEcBVxLmK+1M6H3D3Svd/QTCELQDgVOB1wnD3q4xs+GZ/YQZWZL0eFwT+yUPa/wK0C0axrkizb7eijiWJT0e2cjjpWleV9PG87bU/87n7pmcb1PK8+Sf4V9Jw07/dyPMBXwvOkez10tj8ZH+/WjsfU59nu69Tlie9HhEI483B9Hyn0FERNpAiZyIdBWXEHq+AM6JVhwcZmbFFoqE/4Qwp6kxtWz+wFwLrCUMQbwq3c5m9mvgUML8t6eAh4GqqHlktM+xZvYDYDjwNqF37u3EIWjkA3MrvcLmZOxzZvYTMxtsZv3N7CtmlpivV5v0mtVAsZn9lC17ZxKSk7vtonlZzflH0uOrLRQ1H02Ys5fwzwyO06G5+0fArOjp4WZ2roUC6v3MbC8z+xnwl8T+mVwvLfQKYbgjwBfM7CgLK5J+lzBPD2BmE8MqAV5IevwjMxtuZtsAP0y3czv8DCIi0gQlciLSJbj7s4TFSKoJv/suAxZEz2cR6sr1b+L164DnoqfDgfmEXq4dG3nJ94F/JZ3jbcJCGBCGT0LoGfs1Yajjuuh2etS2CHinBT9ik9x9E6G8QiJRu5rQG7MS+DthwRGixwlTCB/KzyHNAhfuvp6wUiGEBVHWR0vxT2oilFvYcmGTxYS5gomaeE8S5ud1BqcTVocEuImQWK0C3gCuYMvhrplcLxlz9w3A2YQvL4qBRwnX1+Rolyo2L3zSmPuAD6PHnyEMm/yYLYdtJsvqzyAiIk1TIiciXYa73wnsRpibNoswHG4DYb7R74HrmjnENwlJxirCKnz3E1Z4TOc64N+EZKmWsIjIm4Sk6JZon+cIi3p8TEiY6ggJ3F+Az0bJV9a4+98Jc9/+QlhCvpaQyL3I5nlzvwCuIXwY3xS1HULjqw6eDLxE6KHMJIYNhNU6ryQshlFFSHamAxcAR0XzrfKeu79ISFDvIyRBNYT3+x1CAv+TpN0zuV5aev4/EkpV/IPQe1pL+PLhQWAfDwXTm3r9JuAw4G+E/yerCQXVG6u3mPWfQUREGmeZDf0XERERERGRjkI9ciIiIiIiInlGiZyIiIiIiEieUSInIiIiIiKSZ5TIiYiIiIiI5BklciIiIiIiInlGiZyIiIiIiEieUSInIiIiIiKSZ5TIiYiIiIiI5BklciIiIiIiInnm/wNLKWQExlZkBwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 936x504 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "bef_bal_acc_valid = np.array(bef_bal_acc_valid)\n",
    "bef_avg_odds_diff_valid = np.array(bef_avg_odds_diff_valid)\n",
    "\n",
    "aft_bal_acc_valid = np.array(aft_bal_acc_valid)\n",
    "aft_avg_odds_diff_valid = np.array(aft_avg_odds_diff_valid)\n",
    "\n",
    "fig, ax1 = plt.subplots(figsize=(13,7))\n",
    "ax1.plot(all_thresh, bef_bal_acc_valid, color='b')\n",
    "ax1.plot(all_thresh, aft_bal_acc_valid, color='b', linestyle='dashed')\n",
    "ax1.set_title('Original and Postprocessed validation data', fontsize=16, fontweight='bold')\n",
    "ax1.set_xlabel('Classification Thresholds', fontsize=16, fontweight='bold')\n",
    "ax1.set_ylabel('Balanced Accuracy', color='b', fontsize=16, fontweight='bold')\n",
    "ax1.xaxis.set_tick_params(labelsize=14)\n",
    "ax1.yaxis.set_tick_params(labelsize=14)\n",
    "\n",
    "ax2 = ax1.twinx()\n",
    "ax2.plot(all_thresh, np.abs(bef_avg_odds_diff_valid), color='r')\n",
    "ax2.plot(all_thresh, np.abs(aft_avg_odds_diff_valid), color='r', linestyle='dashed')\n",
    "ax2.set_ylabel('abs(Equal opportunity diff)', color='r', fontsize=16, fontweight='bold')\n",
    "ax2.yaxis.set_tick_params(labelsize=14)\n",
    "ax2.grid(True)\n",
    "fig.legend([\"Balanced Acc. - Orig.\", \"Balanced Acc. - Postproc.\",\n",
    "             \"Equal opp. diff. - Orig.\",\"Equal opp. diff. - Postproc.\",], \n",
    "           fontsize=16)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### We have observed how the predicted labels are modified to eliminate the bias"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.5        0.5        0.5        0.5        0.5        0.66444924\n",
      " 0.66444924 0.66444924 0.67811339 0.86989093 0.86989093 0.86989093\n",
      " 0.87953973 0.87953973 0.87953973 0.87953973 0.87953973 0.87953973\n",
      " 0.87953973 0.87953973 0.87953973 0.87953973 0.87953973 0.85526789\n",
      " 0.5       ]\n"
     ]
    }
   ],
   "source": [
    "print(bef_bal_acc_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.5        0.5        0.5        0.5        0.5        0.66444924\n",
      " 0.66444924 0.66444924 0.67811339 0.85754525 0.85754525 0.85754525\n",
      " 0.86719405 0.86719405 0.86719405 0.86719405 0.85526789 0.85526789\n",
      " 0.85526789 0.85526789 0.85526789 0.85526789 0.85526789 0.83099604\n",
      " 0.5       ]\n"
     ]
    }
   ],
   "source": [
    "print(aft_bal_acc_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.          0.          0.          0.          0.         -0.27272727\n",
      " -0.27272727 -0.27272727 -0.36363636 -0.30928854 -0.30928854 -0.30928854\n",
      " -0.17885375 -0.17885375 -0.17885375 -0.17885375 -0.17885375 -0.17885375\n",
      " -0.17885375 -0.17885375 -0.17885375 -0.17885375 -0.17885375 -0.63339921\n",
      "  0.        ]\n"
     ]
    }
   ],
   "source": [
    "print(bef_avg_odds_diff_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.          0.          0.          0.          0.         -0.27272727\n",
      " -0.27272727 -0.27272727 -0.36363636 -0.30928854 -0.30928854 -0.30928854\n",
      " -0.17885375 -0.17885375 -0.17885375 -0.17885375 -0.12450593 -0.12450593\n",
      " -0.12450593 -0.12450593 -0.12450593 -0.12450593 -0.12450593 -0.57905138\n",
      "  0.        ]\n"
     ]
    }
   ],
   "source": [
    "print(aft_avg_odds_diff_valid)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}