Beginning Machine and Deep Learning with Zeek logs/16 - Beginning Deep Learning, - Classification, Tensorflow.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Author: Nik Alleyne <br>\n",
    "Author Blog: **https://www.securitynik.com** <br>\n",
    "Author GitHub: **github.com/securitynik** <br>\n",
    "\n",
    "Author Books: [  <br>\n",
    "\n",
    "                \"https://www.amazon.ca/Learning-Practicing-Leveraging-Practical-Detection/dp/1731254458/\", \n",
    "                \n",
    "                \"https://www.amazon.ca/Learning-Practicing-Mastering-Network-Forensics/dp/1775383024/\" \n",
    "            ] "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Why this series?\n",
    "When teaching the SANS SEC595: Applied Data Science and Machine Learning for Cybersecurity Professionals \n",
    "**https://www.sans.org/cyber-security-courses/applied-data-science-machine-learning/** I am always asked,\n",
    "\"Will you be sharing your demo notebooks?\" or \"Can we get a copy of your demo notebooks?\" or ... well you get the point.\n",
    "My answer is always no. Not that I do not want to share, (sharing is caring :-D) , but the demo notebooks \n",
    "by themselves, would not make sense or add real value. Hence, this series! \n",
    "\n",
    "This is my supplemental work, similar to what I would do in the demos but with a lot more details and references.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 16. Beginning Deep Learning - Classification, Tensorflow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This post is part of my beginning machine learning series. \n",
    "The series includes the following:aq\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The series includes the following: <br>\n",
    "01 - Beginning Numpy <br>\n",
    "02 - Beginning Tensorflow  <br>\n",
    "03 - Beginning PyTorch <br>\n",
    "04 - Beginning Pandas <br>\n",
    "05 - Beginning Matplotlib <br>\n",
    "06 - Beginning Data Scaling <br>\n",
    "07 - Beginning Principal Component Analysis (PCA) <br>\n",
    "08 - Beginning Machine Learning Anomaly Detection - Isolation Forest and Local Outlier Factor <br>\n",
    "09 - Beginning Unsupervised Machine Learning - Clustering - K-means and DBSCAN <br>\n",
    "10 - Beginning Supervise Learning - Machine Learning - Logistic Regression, Decision Trees and Metrics <br>\n",
    "11 - Beginning Linear Regression - Machine Learning <br>\n",
    "12 - Beginning Deep Learning - Anomaly Detection with AutoEncoders, Tensorflow <br>\n",
    "13 - Beginning Deep Learning - Anomaly Detection with AutoEncoders, PyTroch <br>\n",
    "14 - Beginning Deep Learning - Linear Regression, Tensorflow <br>\n",
    "15 - Beginning Deep Learning - Linear Regression, PyTorch <br>\n",
    "16 - Beginning Deep Learning - Classification, Tensorflow <br>\n",
    "17 - Beginning Deep Learning - Classification, Pytorch <br>\n",
    "18 - Beginning Deep Learning - Classification - regression - MIMO - Functional API Tensorflow <br> \n",
    "19 - Beginning Deep Learning - Convolution Networks - Tensorflow <br>\n",
    "20 - Beginning Deep Learning - Convolution Networks - PyTorch <br>\n",
    "21 - Beginning Regularization - Early Stopping, Dropout, L2 (Ridge), L1 (Lasso) <br>\n",
    "22 - Beginning Model TFServing <br>\n",
    "\n",
    "But conn.log is not the only log file within Zeek. Let's build some models for DNS and HTTP logs. <br>\n",
    "I choose unsupervised, because there are no labels coming with these data. <br>\n",
    "\n",
    "23 - Continuing Anomaly Learning - Zeek DNS Log - Machine Learning <br>\n",
    "24 - Continuing Unsupervised Learning - Zeek HTTP Log - Machine Learning <br>\n",
    "\n",
    "This was a specific ask by someone in one of my class. <br>\n",
    "25 - Beginning - Reading Executables and Building a Neural Network to make predictions on suspicious vs suspicious  <br><br>\n",
    "\n",
    "With 25 notebooks in this series, it is quite possible there are things I could have or should have done differently.  <br>\n",
    "If you find any thing, you think fits those criteria, drop me a line. <br>\n",
    "\n",
    "If you find this series beneficial, I would greatly appreciate your feedback.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import some libraries\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>id.resp_p</th>\n",
       "      <th>duration</th>\n",
       "      <th>orig_bytes</th>\n",
       "      <th>resp_bytes</th>\n",
       "      <th>orig_pkts</th>\n",
       "      <th>orig_ip_bytes</th>\n",
       "      <th>resp_pkts</th>\n",
       "      <th>resp_ip_bytes</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>48798</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>48804</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>27761</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>8888</td>\n",
       "      <td>0.053276</td>\n",
       "      <td>208</td>\n",
       "      <td>976</td>\n",
       "      <td>6</td>\n",
       "      <td>528</td>\n",
       "      <td>6</td>\n",
       "      <td>1296</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>8888</td>\n",
       "      <td>0.383646</td>\n",
       "      <td>227</td>\n",
       "      <td>692369</td>\n",
       "      <td>202</td>\n",
       "      <td>10739</td>\n",
       "      <td>117</td>\n",
       "      <td>240893</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430183</th>\n",
       "      <td>9200</td>\n",
       "      <td>0.000051</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>52</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430184</th>\n",
       "      <td>9200</td>\n",
       "      <td>0.000023</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430185</th>\n",
       "      <td>9200</td>\n",
       "      <td>0.000005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430186</th>\n",
       "      <td>9200</td>\n",
       "      <td>0.000005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430187</th>\n",
       "      <td>9200</td>\n",
       "      <td>0.000005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>4236125 rows × 9 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         id.resp_p  duration  orig_bytes  resp_bytes  orig_pkts   \n",
       "1            48798  0.000000           0           0          0  \\\n",
       "2            48804  0.000000           0           0          0   \n",
       "4            27761  0.000000           0           0          0   \n",
       "5             8888  0.053276         208         976          6   \n",
       "6             8888  0.383646         227      692369        202   \n",
       "...            ...       ...         ...         ...        ...   \n",
       "4430183       9200  0.000051           0           0          1   \n",
       "4430184       9200  0.000023           0           0          0   \n",
       "4430185       9200  0.000005           0           0          0   \n",
       "4430186       9200  0.000005           0           0          0   \n",
       "4430187       9200  0.000005           0           0          0   \n",
       "\n",
       "         orig_ip_bytes  resp_pkts  resp_ip_bytes  label  \n",
       "1                    0          0              0      0  \n",
       "2                    0          0              0      0  \n",
       "4                    0          0              0      0  \n",
       "5                  528          6           1296      0  \n",
       "6                10739        117         240893      0  \n",
       "...                ...        ...            ...    ...  \n",
       "4430183             52          1             40      0  \n",
       "4430184              0          1             40      0  \n",
       "4430185              0          1             40      0  \n",
       "4430186              0          1             40      0  \n",
       "4430187              0          1             40      0  \n",
       "\n",
       "[4236125 rows x 9 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# In the notebooks on Pandas, Matplotlib and Scaling\n",
    "# we loaded our dataset such as\n",
    "#   04 - Beginning Pandas \n",
    "#   05 - Beginning Matplotlib\n",
    "#   06 - Beginning Data Scaling\n",
    "df_conn = pd.read_csv(r'df_conn_with_labels.csv', index_col=0)\n",
    "df_conn"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This file represents Zeek (formerly Bro) connection log - conn.log`. <br>\n",
    "Zeek is a framework used for Network Security Monitoring.  <br>\n",
    "This entire series is based on using Zeek's data.  <br>\n",
    "The majority of the notebooks use the conn.log <br>\n",
    "You can learn more about Zeek here:\n",
    "   \n",
    "    https://zeek.org/\n",
    "\n",
    "Alternatively, come hang out with us in the:\n",
    "SANS SEC595: Applied Data Science and Machine Learning for Cybersecurity Professionals\n",
    "\n",
    "        https://www.sans.org/cyber-security-courses/applied-data-science-machine-learning/ OR\n",
    "\n",
    "SEC503 SEC503: Network Monitoring and Threat Detection In-Depth\n",
    "\n",
    "        https://www.sans.org/cyber-security-courses/network-monitoring-threat-detection/\n",
    "\n",
    "\n",
    "Here are also some blog posts on using Zeek for security monitoring\n",
    "Installing Zeek: \n",
    "\n",
    "        https://www.securitynik.com/2020/06/installing-zeek-314-on-ubuntu-2004.html\n",
    "\n",
    "Detecting PowerShell Empire Usage: \n",
    "\n",
    "        https://www.securitynik.com/2022/02/powershell-empire-detection-with-zeek.html\n",
    "\n",
    "Detecting Log4J Vulnerability Exploitation: \n",
    "\n",
    "        https://www.securitynik.com/2021/12/continuing-log4shell-zeek-detection.html\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>duration</th>\n",
       "      <th>orig_bytes</th>\n",
       "      <th>resp_bytes</th>\n",
       "      <th>orig_pkts</th>\n",
       "      <th>orig_ip_bytes</th>\n",
       "      <th>resp_pkts</th>\n",
       "      <th>resp_ip_bytes</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.053276</td>\n",
       "      <td>208</td>\n",
       "      <td>976</td>\n",
       "      <td>6</td>\n",
       "      <td>528</td>\n",
       "      <td>6</td>\n",
       "      <td>1296</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.383646</td>\n",
       "      <td>227</td>\n",
       "      <td>692369</td>\n",
       "      <td>202</td>\n",
       "      <td>10739</td>\n",
       "      <td>117</td>\n",
       "      <td>240893</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430183</th>\n",
       "      <td>0.000051</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>52</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430184</th>\n",
       "      <td>0.000023</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430185</th>\n",
       "      <td>0.000005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430186</th>\n",
       "      <td>0.000005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430187</th>\n",
       "      <td>0.000005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>4236125 rows × 8 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         duration  orig_bytes  resp_bytes  orig_pkts  orig_ip_bytes   \n",
       "1        0.000000           0           0          0              0  \\\n",
       "2        0.000000           0           0          0              0   \n",
       "4        0.000000           0           0          0              0   \n",
       "5        0.053276         208         976          6            528   \n",
       "6        0.383646         227      692369        202          10739   \n",
       "...           ...         ...         ...        ...            ...   \n",
       "4430183  0.000051           0           0          1             52   \n",
       "4430184  0.000023           0           0          0              0   \n",
       "4430185  0.000005           0           0          0              0   \n",
       "4430186  0.000005           0           0          0              0   \n",
       "4430187  0.000005           0           0          0              0   \n",
       "\n",
       "         resp_pkts  resp_ip_bytes  label  \n",
       "1                0              0      0  \n",
       "2                0              0      0  \n",
       "4                0              0      0  \n",
       "5                6           1296      0  \n",
       "6              117         240893      0  \n",
       "...            ...            ...    ...  \n",
       "4430183          1             40      0  \n",
       "4430184          1             40      0  \n",
       "4430185          1             40      0  \n",
       "4430186          1             40      0  \n",
       "4430187          1             40      0  \n",
       "\n",
       "[4236125 rows x 8 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Drop the port column\n",
    "df_conn = df_conn.drop(columns=['id.resp_p'], inplace=False)\n",
    "df_conn"
   ]
  },
  {
   "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>duration</th>\n",
       "      <th>orig_bytes</th>\n",
       "      <th>resp_bytes</th>\n",
       "      <th>orig_pkts</th>\n",
       "      <th>orig_ip_bytes</th>\n",
       "      <th>resp_pkts</th>\n",
       "      <th>resp_ip_bytes</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.053276</td>\n",
       "      <td>208</td>\n",
       "      <td>976</td>\n",
       "      <td>6</td>\n",
       "      <td>528</td>\n",
       "      <td>6</td>\n",
       "      <td>1296</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.383646</td>\n",
       "      <td>227</td>\n",
       "      <td>692369</td>\n",
       "      <td>202</td>\n",
       "      <td>10739</td>\n",
       "      <td>117</td>\n",
       "      <td>240893</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.054294</td>\n",
       "      <td>208</td>\n",
       "      <td>977</td>\n",
       "      <td>6</td>\n",
       "      <td>528</td>\n",
       "      <td>6</td>\n",
       "      <td>1297</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.424980</td>\n",
       "      <td>227</td>\n",
       "      <td>791633</td>\n",
       "      <td>176</td>\n",
       "      <td>9387</td>\n",
       "      <td>174</td>\n",
       "      <td>367737</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.272567</td>\n",
       "      <td>227</td>\n",
       "      <td>406607</td>\n",
       "      <td>171</td>\n",
       "      <td>9127</td>\n",
       "      <td>71</td>\n",
       "      <td>169939</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430183</th>\n",
       "      <td>0.000051</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>52</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430184</th>\n",
       "      <td>0.000023</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430185</th>\n",
       "      <td>0.000005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430186</th>\n",
       "      <td>0.000005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430187</th>\n",
       "      <td>0.000005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>207608 rows × 8 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         duration  orig_bytes  resp_bytes  orig_pkts  orig_ip_bytes   \n",
       "5        0.053276         208         976          6            528  \\\n",
       "6        0.383646         227      692369        202          10739   \n",
       "7        0.054294         208         977          6            528   \n",
       "8        0.424980         227      791633        176           9387   \n",
       "9        0.272567         227      406607        171           9127   \n",
       "...           ...         ...         ...        ...            ...   \n",
       "4430183  0.000051           0           0          1             52   \n",
       "4430184  0.000023           0           0          0              0   \n",
       "4430185  0.000005           0           0          0              0   \n",
       "4430186  0.000005           0           0          0              0   \n",
       "4430187  0.000005           0           0          0              0   \n",
       "\n",
       "         resp_pkts  resp_ip_bytes  label  \n",
       "5                6           1296      0  \n",
       "6              117         240893      0  \n",
       "7                6           1297      0  \n",
       "8              174         367737      0  \n",
       "9               71         169939      0  \n",
       "...            ...            ...    ...  \n",
       "4430183          1             40      0  \n",
       "4430184          1             40      0  \n",
       "4430185          1             40      0  \n",
       "4430186          1             40      0  \n",
       "4430187          1             40      0  \n",
       "\n",
       "[207608 rows x 8 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Looking at above, we see a number of records with 0s. \n",
    "# These will add no value to our learning process\n",
    "# Let's find all those records and drop them\n",
    "# Reference: https://stackoverflow.com/questions/13851535/how-to-delete-rows-from-a-pandas-dataframe-based-on-a-conditional-expression\n",
    "df_conn = df_conn.drop(df_conn[(df_conn.duration == 0 ) & (df_conn.orig_bytes == 0 ) \\\n",
    "                               & (df_conn.resp_bytes == 0 ) & (df_conn.orig_pkts == 0 )  \\\n",
    "                                & (df_conn.orig_ip_bytes == 0 ) & (df_conn.resp_pkts == 0 ) \\\n",
    "                                    & (df_conn.resp_ip_bytes == 0 )].index)\n",
    "df_conn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# The graph below shows this dataset is highly imbalanced.\n",
    "# As a result, using measures like accuracy is more than likely not the best approach, to \n",
    "# understand how well our eventual model has \"learned\" via the training data\n",
    "plt.title('Bar graph showing highly imbalanced dataset')\n",
    "plt.bar(x=['normal', 'suspicious'], height=[ df_conn[df_conn.label == 0].shape[0], \\\n",
    "                                            df_conn[df_conn.label == 1].shape[0] ])\n",
    "plt.ylabel(ylabel='Number of Records')\n",
    "plt.xlabel(xlabel='Normal vs Suspicious')\n",
    "plt.xticks(rotation=45)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.032272359446649454"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Getting the percentage of samples that are considered suspicious in this dataset\n",
    "# This is going to be quite a challenge for this learning algorithm\n",
    "(df_conn[df_conn.label == 1].shape[0] / df_conn.shape[0]) * 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>duration</th>\n",
       "      <th>orig_bytes</th>\n",
       "      <th>resp_bytes</th>\n",
       "      <th>orig_pkts</th>\n",
       "      <th>orig_ip_bytes</th>\n",
       "      <th>resp_pkts</th>\n",
       "      <th>resp_ip_bytes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.053276</td>\n",
       "      <td>208</td>\n",
       "      <td>976</td>\n",
       "      <td>6</td>\n",
       "      <td>528</td>\n",
       "      <td>6</td>\n",
       "      <td>1296</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.383646</td>\n",
       "      <td>227</td>\n",
       "      <td>692369</td>\n",
       "      <td>202</td>\n",
       "      <td>10739</td>\n",
       "      <td>117</td>\n",
       "      <td>240893</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.054294</td>\n",
       "      <td>208</td>\n",
       "      <td>977</td>\n",
       "      <td>6</td>\n",
       "      <td>528</td>\n",
       "      <td>6</td>\n",
       "      <td>1297</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.424980</td>\n",
       "      <td>227</td>\n",
       "      <td>791633</td>\n",
       "      <td>176</td>\n",
       "      <td>9387</td>\n",
       "      <td>174</td>\n",
       "      <td>367737</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.272567</td>\n",
       "      <td>227</td>\n",
       "      <td>406607</td>\n",
       "      <td>171</td>\n",
       "      <td>9127</td>\n",
       "      <td>71</td>\n",
       "      <td>169939</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430183</th>\n",
       "      <td>0.000051</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>52</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430184</th>\n",
       "      <td>0.000023</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430185</th>\n",
       "      <td>0.000005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430186</th>\n",
       "      <td>0.000005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4430187</th>\n",
       "      <td>0.000005</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>207608 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         duration  orig_bytes  resp_bytes  orig_pkts  orig_ip_bytes   \n",
       "5        0.053276         208         976          6            528  \\\n",
       "6        0.383646         227      692369        202          10739   \n",
       "7        0.054294         208         977          6            528   \n",
       "8        0.424980         227      791633        176           9387   \n",
       "9        0.272567         227      406607        171           9127   \n",
       "...           ...         ...         ...        ...            ...   \n",
       "4430183  0.000051           0           0          1             52   \n",
       "4430184  0.000023           0           0          0              0   \n",
       "4430185  0.000005           0           0          0              0   \n",
       "4430186  0.000005           0           0          0              0   \n",
       "4430187  0.000005           0           0          0              0   \n",
       "\n",
       "         resp_pkts  resp_ip_bytes  \n",
       "5                6           1296  \n",
       "6              117         240893  \n",
       "7                6           1297  \n",
       "8              174         367737  \n",
       "9               71         169939  \n",
       "...            ...            ...  \n",
       "4430183          1             40  \n",
       "4430184          1             40  \n",
       "4430185          1             40  \n",
       "4430186          1             40  \n",
       "4430187          1             40  \n",
       "\n",
       "[207608 rows x 7 columns]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Extract the X_data\n",
    "X_data = df_conn.drop(columns=['label'], inplace=False)\n",
    "X_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5          0\n",
       "6          0\n",
       "7          0\n",
       "8          0\n",
       "9          0\n",
       "          ..\n",
       "4430183    0\n",
       "4430184    0\n",
       "4430185    0\n",
       "4430186    0\n",
       "4430187    0\n",
       "Name: label, Length: 207608, dtype: int64"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Extract the labels\n",
    "y_label = df_conn.label\n",
    "y_label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# prepare to split the data into training and testing sets\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((166086, 7), (166086,), (41522, 7), (41522,))"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Split the data into training and testing sets\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_data.values, y_label, test_size=0.2, \\\n",
    "                                                    train_size=0.8, stratify=y_label, random_state=10)\n",
    "X_train.shape, y_train.shape, X_test.shape, y_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# With a statistical understanding of the normal and suspicious datasets, time to build the model\n",
    "# Scaling was covered in \n",
    "#   06 - Beginning Data Scaling\n",
    "# Scaling the data first\n",
    "# import the scaler library\n",
    "# https://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.MinMaxScaler.html\n",
    "from sklearn.preprocessing import MinMaxScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[2.63072563e-07, 9.18795041e-08, 7.10178687e-08, ...,\n",
       "        3.59579942e-07, 1.75013957e-06, 3.05765612e-07],\n",
       "       [3.61672231e-07, 5.89415687e-08, 2.38314996e-08, ...,\n",
       "        2.75234030e-07, 1.75013957e-06, 1.34744168e-07],\n",
       "       [2.79225614e-09, 0.00000000e+00, 0.00000000e+00, ...,\n",
       "        2.30841444e-07, 1.75013957e-06, 6.90995734e-08],\n",
       "       ...,\n",
       "       [1.68135854e-09, 0.00000000e+00, 0.00000000e+00, ...,\n",
       "        2.30841444e-07, 1.75013957e-06, 6.90995734e-08],\n",
       "       [1.65133428e-09, 0.00000000e+00, 0.00000000e+00, ...,\n",
       "        2.30841444e-07, 1.75013957e-06, 6.90995734e-08],\n",
       "       [3.49932746e-07, 7.80108997e-08, 8.38868784e-08, ...,\n",
       "        3.24065874e-07, 1.75013957e-06, 3.52407824e-07]])"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Setup the scaler\n",
    "min_max_scaler = MinMaxScaler(feature_range=(0,1))\n",
    "\n",
    "# Fit on the training data\n",
    "min_max_scaler.fit(X_train)\n",
    "\n",
    "# Transform the train data\n",
    "X_train = min_max_scaler.transform(X_train)\n",
    "X_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1.73484676e-05, 3.90054499e-07, 8.33234064e-04, ...,\n",
       "        7.60578166e-05, 6.94805411e-04, 1.16707452e-03],\n",
       "       [5.08070521e-07, 6.41422953e-08, 5.29059290e-08, ...,\n",
       "        2.88551805e-07, 1.75013957e-06, 2.40121018e-07],\n",
       "       [4.43809598e-06, 2.05948775e-06, 1.84884774e-05, ...,\n",
       "        9.69534066e-06, 5.60044664e-05, 6.65187043e-05],\n",
       "       ...,\n",
       "       [5.47522398e-07, 6.41422953e-08, 8.96064383e-08, ...,\n",
       "        2.88551805e-07, 1.75013957e-06, 3.73137696e-07],\n",
       "       [1.83147984e-09, 0.00000000e+00, 0.00000000e+00, ...,\n",
       "        2.30841444e-07, 1.75013957e-06, 6.90995734e-08],\n",
       "       [0.00000000e+00, 0.00000000e+00, 0.00000000e+00, ...,\n",
       "        3.95094010e-07, 0.00000000e+00, 0.00000000e+00]])"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Scale the test data\n",
    "X_test = min_max_scaler.transform(X_test)\n",
    "X_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import PCA to leverage dimensionality reduction\n",
    "# PCA was covered in notebook\n",
    "#   07 - Beginning Principal Component Analysis (PCA)\n",
    "from sklearn.decomposition import PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-1 {color: black;background-color: white;}#sk-container-id-1 pre{padding: 0;}#sk-container-id-1 div.sk-toggleable {background-color: white;}#sk-container-id-1 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-1 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-1 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-1 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-1 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-1 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-1 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-1 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-1 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-1 div.sk-item {position: relative;z-index: 1;}#sk-container-id-1 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-1 div.sk-item::before, #sk-container-id-1 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-1 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-1 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-1 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-1 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-1 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-1 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-1 div.sk-label-container {text-align: center;}#sk-container-id-1 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-1 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>PCA(n_components=3, random_state=10)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" checked><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">PCA</label><div class=\"sk-toggleable__content\"><pre>PCA(n_components=3, random_state=10)</pre></div></div></div></div></div>"
      ],
      "text/plain": [
       "PCA(n_components=3, random_state=10)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Setup PCA to use 3 principal Components\n",
    "pca = PCA(n_components=3, random_state=10)\n",
    "pca"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-0.00025828, -0.00023497, -0.00047665],\n",
       "       [-0.00025839, -0.00023502, -0.00047653],\n",
       "       [-0.00025846, -0.00023515, -0.00047686],\n",
       "       ...,\n",
       "       [-0.00025846, -0.00023515, -0.00047686],\n",
       "       [-0.00025846, -0.00023515, -0.00047686],\n",
       "       [-0.00025826, -0.00023494, -0.00047657]])"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Fit on the X_train\n",
    "pca.fit(X_train)\n",
    "\n",
    "# transform the training data\n",
    "X_train = pca.transform(X_train)\n",
    "X_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.00134937,  0.000691  , -0.0008094 ],\n",
       "       [-0.00025832, -0.00023494, -0.0004764 ],\n",
       "       [-0.00018612, -0.00015918, -0.0004979 ],\n",
       "       ...,\n",
       "       [-0.00025824, -0.00023488, -0.00047639],\n",
       "       [-0.00025846, -0.00023515, -0.00047686],\n",
       "       [-0.00025896, -0.0002359 , -0.00047662]])"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Use the opportunity to PCA transform the X_test\n",
    "X_test = pca.transform(X_test)\n",
    "X_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# In this notebook, I'm using  the Sequential API. In the autonencoder anomaly detection notebook, \n",
    "# we used model subclassing. In another I will use the  Functional API\n",
    "#   12. Beginning Deep Learning - Anomaly Detection with AutoEncoders, Tensorflow\n",
    "#   14 - Beginning Deep Learning - Linear Regression, Tensorflow <br>\n",
    "#   18 - Beginning Deep Learning - Classification - regression - MIMO - Functional API Tensorflow <br> "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import tensorflow\n",
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"classification_model\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " hidden_1 (Dense)            (None, 8)                 32        \n",
      "                                                                 \n",
      " output (Dense)              (None, 1)                 9         \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 41\n",
      "Trainable params: 41\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# Set the random seed\n",
    "# This is needed for reproducibility \n",
    "tf.keras.utils.set_random_seed(seed=10)\n",
    "\n",
    "# Setup the model using the Sequential Class\n",
    "tf_clf_model = tf.keras.Sequential(name='classification_model')\n",
    "\n",
    "# Build a model that takes input based on the index position 1, returned from X_train.shape. \n",
    "# # This gives the number fo features\n",
    "tf_clf_model.add(tf.keras.layers.InputLayer(input_shape=(X_train.shape[1],), name='input_layer'))\n",
    "\n",
    "# Setup the first hidden layer with 8 units\n",
    "# Use Rectified Linear Unit (ReLU) as the activation funtion\n",
    "tf_clf_model.add(tf.keras.layers.Dense(units=8, activation='relu', name='hidden_1'))\n",
    "\n",
    "# Setup the output layer with 1 unit and Sigmoid activation function\n",
    "# We learned about Sigmoid in notebook\n",
    "#   10. Beginning Supervise Learning - Machine Learning - Logistic Regression, Decision Trees and Metrics\n",
    "tf_clf_model.add(tf.keras.layers.Dense(units=1, activation='sigmoid', name='output'))\n",
    "tf_clf_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-0.5, -0.4, -0.3, -0.2, -0.1, -0. ,  0.1,  0.2,  0.3,  0.4,  0.5])"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Before compiling the model, let's see what ReLU does\n",
    "# Setup some samples between -10 and 10, space them by 0.1\n",
    "sample_numbers = np.arange(-10, 10, 0.1)\n",
    "np.round(sample_numbers[95:106],2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Range of Values')"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the numbers that were created above\n",
    "# we see there are 200 numbers between -10 and 10\n",
    "plt.title('Range of values between -10 and +10')\n",
    "plt.plot(sample_numbers)\n",
    "plt.xlabel('Number of records')\n",
    "plt.ylabel('Range of Values')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "# With ReLU, anything less than 0 will be made 0 and anything above 0 will be kept the same\n",
    "# Setup a function to take care of this \n",
    "def my_relu(x: np.array) -> list:\n",
    "    ''' Computes  RELU from the x '''\n",
    "    return [ 0 if i <=0 else i for i in x ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "([0], [0], [10])"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Testing our function with 2 values\n",
    "# One less than 0 and another greater than 0\n",
    "# We see below when x is less than 0, the value returned is 0\n",
    "my_relu(np.array([-10])), my_relu(np.array([0])),   my_relu(np.array([10])), "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0,\n",
       " 0,\n",
       " 0,\n",
       " 0,\n",
       " 0,\n",
       " 0,\n",
       " 0.09999999999996412,\n",
       " 0.19999999999996376,\n",
       " 0.2999999999999634,\n",
       " 0.39999999999996305,\n",
       " 0.4999999999999627]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Running ReLU against our samples\n",
    "my_relu(sample_numbers)[95:106]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x227be3e7b50>"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Making this more visual by plotting the original numbers vs the numbers which ReLU has been applied to\n",
    "# As we can see, anything below 0 has now become 0\n",
    "plt.plot(sample_numbers, lw=10, label='original values')\n",
    "plt.plot(my_relu(sample_numbers), lw=2, linestyle='--', label='values after ReLU activation')\n",
    "plt.xlabel('Number of records')\n",
    "plt.ylabel('Range of Values')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "# With that understanding of RELU activation function\n",
    "# Let's compile the model\n",
    "tf_clf_model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=0.001), loss='binary_crossentropy', \\\n",
    "              metrics=[tf.keras.metrics.Recall()])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "2596/2596 [==============================] - 9s 3ms/step - loss: 0.5178 - recall_1: 0.0000e+00 - val_loss: 0.1513 - val_recall_1: 0.0000e+00\n",
      "Epoch 2/10\n",
      "2596/2596 [==============================] - 7s 3ms/step - loss: 0.4423 - recall_1: 0.0185 - val_loss: 0.1295 - val_recall_1: 0.0769\n",
      "Epoch 3/10\n",
      "2596/2596 [==============================] - 6s 2ms/step - loss: 0.4205 - recall_1: 0.0185 - val_loss: 0.1260 - val_recall_1: 0.0769\n",
      "Epoch 4/10\n",
      "2596/2596 [==============================] - 6s 2ms/step - loss: 0.3949 - recall_1: 0.0185 - val_loss: 0.1214 - val_recall_1: 0.1538\n",
      "Epoch 5/10\n",
      "2596/2596 [==============================] - 7s 3ms/step - loss: 0.3706 - recall_1: 0.1296 - val_loss: 0.1163 - val_recall_1: 0.1538\n",
      "Epoch 6/10\n",
      "2596/2596 [==============================] - 7s 3ms/step - loss: 0.3426 - recall_1: 0.2222 - val_loss: 0.1395 - val_recall_1: 0.3077\n",
      "Epoch 7/10\n",
      "2596/2596 [==============================] - 7s 3ms/step - loss: 0.3208 - recall_1: 0.2963 - val_loss: 0.1109 - val_recall_1: 0.3077\n",
      "Epoch 8/10\n",
      "2596/2596 [==============================] - 6s 2ms/step - loss: 0.3003 - recall_1: 0.3333 - val_loss: 0.0950 - val_recall_1: 0.3077\n",
      "Epoch 9/10\n",
      "2596/2596 [==============================] - 7s 3ms/step - loss: 0.2809 - recall_1: 0.4259 - val_loss: 0.0694 - val_recall_1: 0.3077\n",
      "Epoch 10/10\n",
      "2596/2596 [==============================] - 7s 3ms/step - loss: 0.2646 - recall_1: 0.4074 - val_loss: 0.1276 - val_recall_1: 0.3077\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x227be774700>"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Train the model\n",
    "# Notice I introduced the class_weight parameter here to try to reduce the imbalancing of the classes\n",
    "# https://stackoverflow.com/questions/43459317/keras-class-weight-vs-sample-weights-in-the-fit-generator\n",
    "tf_clf_model.fit(x=X_train, y=y_train, epochs=10, batch_size=64, validation_data=(X_test, y_test), \\\n",
    "          shuffle=True, class_weight={0:1, 1:500})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'loss': [0.5178429484367371, 0.4422512948513031, 0.42047640681266785, 0.39487582445144653, 0.37055543065071106, 0.3426285982131958, 0.3208451569080353, 0.3002537786960602, 0.2808709144592285, 0.2645675539970398], 'recall_1': [0.0, 0.018518518656492233, 0.018518518656492233, 0.018518518656492233, 0.12962962687015533, 0.2222222238779068, 0.29629629850387573, 0.3333333432674408, 0.42592594027519226, 0.40740740299224854], 'val_loss': [0.1513395458459854, 0.1294926255941391, 0.12596602737903595, 0.12136714905500412, 0.11625183373689651, 0.13949763774871826, 0.11093323677778244, 0.0950491726398468, 0.06944186985492706, 0.1275871843099594], 'val_recall_1': [0.0, 0.07692307978868484, 0.07692307978868484, 0.1538461595773697, 0.1538461595773697, 0.3076923191547394, 0.3076923191547394, 0.3076923191547394, 0.3076923191547394, 0.3076923191547394]}\n"
     ]
    }
   ],
   "source": [
    "# How did our model perform?\n",
    "# Getting the history\n",
    "print(tf_clf_model.history.history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "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>loss</th>\n",
       "      <th>recall_1</th>\n",
       "      <th>val_loss</th>\n",
       "      <th>val_recall_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.517843</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.151340</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.442251</td>\n",
       "      <td>0.018519</td>\n",
       "      <td>0.129493</td>\n",
       "      <td>0.076923</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.420476</td>\n",
       "      <td>0.018519</td>\n",
       "      <td>0.125966</td>\n",
       "      <td>0.076923</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.394876</td>\n",
       "      <td>0.018519</td>\n",
       "      <td>0.121367</td>\n",
       "      <td>0.153846</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.370555</td>\n",
       "      <td>0.129630</td>\n",
       "      <td>0.116252</td>\n",
       "      <td>0.153846</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.342629</td>\n",
       "      <td>0.222222</td>\n",
       "      <td>0.139498</td>\n",
       "      <td>0.307692</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.320845</td>\n",
       "      <td>0.296296</td>\n",
       "      <td>0.110933</td>\n",
       "      <td>0.307692</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.300254</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.095049</td>\n",
       "      <td>0.307692</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.280871</td>\n",
       "      <td>0.425926</td>\n",
       "      <td>0.069442</td>\n",
       "      <td>0.307692</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.264568</td>\n",
       "      <td>0.407407</td>\n",
       "      <td>0.127587</td>\n",
       "      <td>0.307692</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       loss  recall_1  val_loss  val_recall_1\n",
       "0  0.517843  0.000000  0.151340      0.000000\n",
       "1  0.442251  0.018519  0.129493      0.076923\n",
       "2  0.420476  0.018519  0.125966      0.076923\n",
       "3  0.394876  0.018519  0.121367      0.153846\n",
       "4  0.370555  0.129630  0.116252      0.153846\n",
       "5  0.342629  0.222222  0.139498      0.307692\n",
       "6  0.320845  0.296296  0.110933      0.307692\n",
       "7  0.300254  0.333333  0.095049      0.307692\n",
       "8  0.280871  0.425926  0.069442      0.307692\n",
       "9  0.264568  0.407407  0.127587      0.307692"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# A cleaner view for easier comparison\n",
    "pd.DataFrame(tf_clf_model.history.history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>loss</th>\n",
       "      <td>0.517843</td>\n",
       "      <td>0.442251</td>\n",
       "      <td>0.420476</td>\n",
       "      <td>0.394876</td>\n",
       "      <td>0.370555</td>\n",
       "      <td>0.342629</td>\n",
       "      <td>0.320845</td>\n",
       "      <td>0.300254</td>\n",
       "      <td>0.280871</td>\n",
       "      <td>0.264568</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>recall_1</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.018519</td>\n",
       "      <td>0.018519</td>\n",
       "      <td>0.018519</td>\n",
       "      <td>0.129630</td>\n",
       "      <td>0.222222</td>\n",
       "      <td>0.296296</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.425926</td>\n",
       "      <td>0.407407</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>val_loss</th>\n",
       "      <td>0.151340</td>\n",
       "      <td>0.129493</td>\n",
       "      <td>0.125966</td>\n",
       "      <td>0.121367</td>\n",
       "      <td>0.116252</td>\n",
       "      <td>0.139498</td>\n",
       "      <td>0.110933</td>\n",
       "      <td>0.095049</td>\n",
       "      <td>0.069442</td>\n",
       "      <td>0.127587</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>val_recall_1</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.076923</td>\n",
       "      <td>0.076923</td>\n",
       "      <td>0.153846</td>\n",
       "      <td>0.153846</td>\n",
       "      <td>0.307692</td>\n",
       "      <td>0.307692</td>\n",
       "      <td>0.307692</td>\n",
       "      <td>0.307692</td>\n",
       "      <td>0.307692</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     0         1         2         3         4         5   \n",
       "loss          0.517843  0.442251  0.420476  0.394876  0.370555  0.342629  \\\n",
       "recall_1      0.000000  0.018519  0.018519  0.018519  0.129630  0.222222   \n",
       "val_loss      0.151340  0.129493  0.125966  0.121367  0.116252  0.139498   \n",
       "val_recall_1  0.000000  0.076923  0.076923  0.153846  0.153846  0.307692   \n",
       "\n",
       "                     6         7         8         9  \n",
       "loss          0.320845  0.300254  0.280871  0.264568  \n",
       "recall_1      0.296296  0.333333  0.425926  0.407407  \n",
       "val_loss      0.110933  0.095049  0.069442  0.127587  \n",
       "val_recall_1  0.307692  0.307692  0.307692  0.307692  "
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# How about a much better view\n",
    "# Transposing the dataframe, now we can look across all the metrics\n",
    "pd.DataFrame(tf_clf_model.history.history).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA04AAAHyCAYAAADcN56mAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAAChoklEQVR4nOzdd1xV9R/H8dcFBEQEtzjIgZqae+a2xNDK0lyZZZKaWe5tlquBW3OUpqmllaZm2XAgjpy5Nc29zYkDVBQU7u+P7w+SXIDAYbyfj8d5eO7h3HM+93rPvfdzv9/v52uz2+12RERERERE5IEcrA5AREREREQkpVPiJCIiIiIi8ghKnERERERERB5BiZOIiIiIiMgjKHESERERERF5BCVOIiIiIiIij6DESURERERE5BGUOImIiIiIiDyCEicREREREZFHUOIkIiJpms1mY8iQIVaH8djq1q1L3bp1Y24fP34cm83GrFmzLItJRCQ9UeIkIpKGzJo1C5vNxtatW60OJVWJTkKiFwcHB7Jly0bDhg3ZuHGj1eGJiEgK4GR1ACIiIilFq1ateP7554mMjOTgwYN8/vnnPPPMM2zZsoXSpUtbHZ6IiFhIiZOIiMj/VahQgddffz3mdq1atWjYsCFffPEFn3/+uYWRiYiI1dRVT0QkHdqxYwcNGzbEw8MDd3d36tWrx6ZNm2Ltc/v2bYYOHUrRokVxdXUle/bs1KxZk8DAwJh9zp07h7+/P/nz58fFxYU8efLw8ssvc/z48Yeef/fu3bRt25bChQvj6uqKl5cXb731FpcuXYq135AhQ7DZbBw+fJi2bduSJUsWPD098ff3JywsLNa+4eHh9OjRg5w5c5I5c2ZeeuklTp8+/VjPU61atQA4cuRIrO1Xr16le/fueHt74+LiQpEiRRgxYgRRUVGx9ouKiuKzzz6jdOnSuLq6kjNnTho0aBCrK+XMmTN59tlnyZUrFy4uLpQsWZIvvvjiseIWEZHEpxYnEZF0Zu/evdSqVQsPDw/69u1LhgwZmDp1KnXr1mXNmjVUrVoVMElLQEAA7du3p0qVKoSGhrJ161a2b99O/fr1AWjatCl79+6lS5cuFCxYkAsXLhAYGMjJkycpWLDgA2MIDAzk6NGj+Pv74+Xlxd69e/nyyy/Zu3cvmzZtwmazxdq/RYsWFCpUiICAALZv38706dPJlSsXI0aMiNmnffv2zJkzh9dee43q1auzcuVKXnjhhcd6rqITwKxZs8ZsCwsLo06dOvzzzz907NiRJ554gg0bNjBgwADOnj3L+PHjY/Zt164ds2bNomHDhrRv3547d+6wdu1aNm3aRKVKlQD44osveOqpp3jppZdwcnLil19+4d133yUqKor33nvvseIXEZFEZBcRkTRj5syZdsC+ZcuWB+7TuHFju7Ozs/3IkSMx286cOWPPnDmzvXbt2jHbypYta3/hhRceeJwrV67YAfuoUaPiHWdYWNg9277//ns7YP/jjz9itg0ePNgO2N96661Y+zZp0sSePXv2mNs7d+60A/Z333031n6vvfaaHbAPHjz4ofEcO3bMDtiHDh1qv3jxov3cuXP2tWvX2itXrmwH7PPnz4/Z96OPPrJnypTJfvDgwVjH6N+/v93R0dF+8uRJu91ut69cudIO2Lt27XrP+aKioh76XPj5+dkLFy4ca1udOnXsderUuSfmmTNnPvSxiYhI4lBXPRGRdCQyMpLly5fTuHFjChcuHLM9T548vPbaa6xbt47Q0FAAsmTJwt69ezl06NB9j5UxY0acnZ1ZvXo1V65ciVccGTNmjFm/desWwcHBPP300wBs3779nv3feeedWLdr1arFpUuXYmL9/fffAejatWus/bp37x6vuAYPHkzOnDnx8vKiVq1a7Nu3jzFjxtCsWbOYfebPn0+tWrXImjUrwcHBMYuvry+RkZH88ccfACxcuBCbzcbgwYPvOc/dLWp3PxchISEEBwdTp04djh49SkhISLziFxGRpKPESUQkHbl48SJhYWE8+eST9/ytRIkSREVFcerUKQCGDRvG1atXKVasGKVLl6ZPnz7s3r07Zn8XFxdGjBjBkiVLyJ07N7Vr12bkyJGcO3fukXFcvnyZbt26kTt3bjJmzEjOnDkpVKgQwH2ThSeeeCLW7eiuc9EJ24kTJ3BwcMDHxyfWfvd7nA/z9ttvExgYyC+//EKPHj24efMmkZGRsfY5dOgQS5cuJWfOnLEWX19fAC5cuACYcVF58+YlW7ZsDz3n+vXr8fX1JVOmTGTJkoWcOXPy/vvvA/d/LkRExBoa4yQiIvdVu3Ztjhw5ws8//8zy5cuZPn0648aNY8qUKbRv3x4wLTqNGjXip59+YtmyZXz44YcEBASwcuVKypcv/8Bjt2jRgg0bNtCnTx/KlSuHu7s7UVFRNGjQ4J4CCwCOjo73PY7dbk+cB/t/RYsWjUmAXnzxRRwdHenfvz/PPPNMzJikqKgo6tevT9++fe97jGLFisX5fEeOHKFevXoUL16csWPH4u3tjbOzM7///jvjxo2773MhIiLWUOIkIpKO5MyZEzc3Nw4cOHDP3/bv34+DgwPe3t4x27Jly4a/vz/+/v5cv36d2rVrM2TIkJjECcDHx4devXrRq1cvDh06RLly5RgzZgxz5sy5bwxXrlwhKCiIoUOHMmjQoJjtD+oSGBcFChQgKiqKI0eOxGplut/jjI+BAwcybdo0PvjgA5YuXQqYx3v9+vWYBOtBfHx8WLZsGZcvX35gq9Mvv/xCeHg4ixcvjtWqtmrVqseKW0REEp+66omIpCOOjo4899xz/Pzzz7FKhp8/f57vvvuOmjVr4uHhAXBPaXB3d3eKFClCeHg4YKrL3bp1K9Y+Pj4+ZM6cOWafB8UA97YW3V2NLr4aNmwIwIQJExLtmGDGeXXs2JFly5axc+dOwLSWbdy4kWXLlt2z/9WrV7lz5w5gKg7a7XaGDh16z37Rj/1+z0VISAgzZ858rLhFRCTxqcVJRCQNmjFjRkwLyd26devGxx9/TGBgIDVr1uTdd9/FycmJqVOnEh4ezsiRI2P2LVmyJHXr1qVixYpky5aNrVu3smDBAjp37gzAwYMHqVevHi1atKBkyZI4OTmxaNEizp8/z6uvvvrA2Dw8PGLGQ92+fZt8+fKxfPlyjh07luDHW65cOVq1asXnn39OSEgI1atXJygoiMOHDyf4mNG6devG+PHjGT58OHPnzqVPnz4sXryYF198kbZt21KxYkVu3LjBX3/9xYIFCzh+/Dg5cuTgmWee4Y033mDChAkcOnQophvi2rVreeaZZ+jcuTPPPfcczs7ONGrUiI4dO3L9+nWmTZtGrly5OHv27GPHLiIiiUeJk4hIGvSgCVTbtm3LU089xdq1axkwYAABAQFERUVRtWpV5syZEzOHE5gKdYsXL2b58uWEh4dToEABPv74Y/r06QOAt7c3rVq1IigoiNmzZ+Pk5ETx4sX54YcfaNq06UPj++677+jSpQuTJ0/Gbrfz3HPPsWTJEvLmzZvgxzxjxgxy5szJt99+y08//cSzzz7Lb7/9FqvrYULkzZuX1157jdmzZ3PkyBF8fHxYs2YNn376KfPnz+ebb77Bw8ODYsWKMXToUDw9PWPuO3PmTMqUKcNXX31Fnz598PT0pFKlSlSvXh0wxSsWLFjABx98QO/evfHy8qJTp07kzJmTt95667HiFhGRxGWzJ/bIWhERERERkTRGY5xEREREREQeQYmTiIiIiIjIIyhxEhEREREReQQlTiIiIiIiIo+gxElEREREROQRlDiJiIiIiIg8ghInERERERGRR1DiJCIiIiIi8ghKnERERERERB5BiZOIiIiIiMgjKHESERERERF5BCVOIiIiIiIij6DESURERERE5BGUOImIiIiI5Ww2G0OGDIm5PWvWLGw2G8ePH7csJpG7KXESSWbRHwRbt261OhQREUlHoj9/ohcnJyfy5ctH27Zt+eeff6wOTyTFc7I6ABERERFJPsOGDaNQoULcunWLTZs2MWvWLNatW8eePXtwdXW1OjyRFEuJk4iIiEg60rBhQypVqgRA+/btyZEjByNGjGDx4sW0aNHC4uhEUi511RNJgXbs2EHDhg3x8PDA3d2devXqsWnTplj73L59m6FDh1K0aFFcXV3Jnj07NWvWJDAwMGafc+fO4e/vT/78+XFxcSFPnjy8/PLL6i8uIiIxatWqBcCRI0ditu3fv59mzZqRLVs2XF1dqVSpEosXL77nvlevXqVHjx4ULFgQFxcX8ufPT5s2bQgODgYgIiKCQYMGUbFiRTw9PcmUKRO1atVi1apVyfPgRBKRWpxEUpi9e/dSq1YtPDw86Nu3LxkyZGDq1KnUrVuXNWvWULVqVQCGDBlCQEAA7du3p0qVKoSGhrJ161a2b99O/fr1AWjatCl79+6lS5cuFCxYkAsXLhAYGMjJkycpWLCghY9SRERSiugf07JmzQqYz6EaNWqQL18++vfvT6ZMmfjhhx9o3LgxCxcupEmTJgBcv36dWrVqsW/fPt566y0qVKhAcHAwixcv5vTp0+TIkYPQ0FCmT59Oq1at6NChA9euXeOrr77Cz8+PzZs3U65cOYsetUgC2EUkWc2cOdMO2Lds2XLfvzdu3Nju7OxsP3LkSMy2M2fO2DNnzmyvXbt2zLayZcvaX3jhhQee58qVK3bAPmrUqMQLXkREUq3oz58VK1bYL168aD916pR9wYIF9pw5c9pdXFzsp06dstvtdnu9evXspUuXtt+6dSvmvlFRUfbq1avbixYtGrNt0KBBdsD+448/3nOuqKgou91ut9+5c8ceHh4e629Xrlyx586d2/7WW2/F2g7YBw8efE+8x44de9yHLpIo1FVPJAWJjIxk+fLlNG7cmMKFC8dsz5MnD6+99hrr1q0jNDQUgCxZsrB3714OHTp032NlzJgRZ2dnVq9ezZUrV5IlfhERSfl8fX3JmTMn3t7eNGvWjEyZMrF48WLy58/P5cuXWblyJS1atODatWsEBwcTHBzMpUuX8PPz49ChQzEV+BYuXEjZsmVjWqDuZrPZAHB0dMTZ2RmAqKgoLl++zJ07d6hUqRLbt29PvgctkgiUOImkIBcvXiQsLIwnn3zynr+VKFGCqKgoTp06BZiqSFevXqVYsWKULl2aPn36sHv37pj9XVxcGDFiBEuWLCF37tzUrl2bkSNHcu7cuWR7PCIikvJMnjyZwMBAFixYwPPPP09wcDAuLi4AHD58GLvdzocffkjOnDljLYMHDwbgwoULgBkTVapUqUee7+uvv6ZMmTIx43Fz5szJb7/9RkhISNI9SJEkoMRJJJWqXbs2R44cYcaMGZQqVYrp06dToUIFpk+fHrNP9+7dOXjwIAEBAbi6uvLhhx9SokQJduzYYWHkIiJipSpVquDr60vTpk1ZvHgxpUqV4rXXXuP69etERUUB0Lt3bwIDA++7FClSJM7nmjNnDm3btsXHx4evvvqKpUuXEhgYyLPPPhtzLpHUQsUhRFKQnDlz4ubmxoEDB+752/79+3FwcMDb2ztmW7Zs2fD398ff35/r169Tu3ZthgwZQvv27WP28fHxoVevXvTq1YtDhw5Rrlw5xowZw5w5c5LlMYmISMrl6OhIQEAAzzzzDJMmTeKtt94CIEOGDPj6+j70vj4+PuzZs+eh+yxYsIDChQvz448/xnTfA2Jar0RSE7U4iaQgjo6OPPfcc/z888+xSoafP3+e7777jpo1a+Lh4QHApUuXYt3X3d2dIkWKEB4eDkBYWBi3bt2KtY+Pjw+ZM2eO2UdERKRu3bpUqVKF8ePH4+HhQd26dZk6dSpnz569Z9+LFy/GrDdt2pRdu3axaNGie/az2+2A+Vy7+zbAn3/+ycaNGxP7YYgkObU4iVhkxowZLF269J7tQ4YMITAwkJo1a/Luu+/i5OTE1KlTCQ8PZ+TIkTH7lSxZkrp161KxYkWyZcvG1q1bWbBgAZ07dwbg4MGD1KtXjxYtWlCyZEmcnJxYtGgR58+f59VXX022xykiIilfnz59aN68ObNmzWLy5MnUrFmT0qVL06FDBwoXLsz58+fZuHEjp0+fZteuXTH3WbBgAc2bN+ett96iYsWKXL58mcWLFzNlyhTKli3Liy++yI8//kiTJk144YUXOHbsGFOmTKFkyZJcv37d4kctEj9KnEQs8sUXX9x3e9u2bVm7di0DBgwgICCAqKgoqlatypw5c2LmcALo2rUrixcvZvny5YSHh1OgQAE+/vhj+vTpA4C3tzetWrUiKCiI2bNn4+TkRPHixfnhhx9o2rRpsjxGERFJHV555RV8fHwYPXo0HTp0YOvWrQwdOpRZs2Zx6dIlcuXKRfny5Rk0aFDMfdzd3Vm7di2DBw9m0aJFfP311+TKlYt69eqRP39+wHymnTt3jqlTp7Js2TJKlizJnDlzmD9/PqtXr7bo0YokjM1+d9upiIiIiIiI3ENjnERERERERB5BiZOIiIiIiMgjKHESERERERF5BCVOIiIiIiIij6DESURERERE5BGUOImIiIiIiDxCupvHKSoqijNnzpA5c2ZsNpvV4YiIpCt2u51r166RN29eHBz02100fTaJiFgjPp9L6S5xOnPmDN7e3laHISKSrp06dSpmgkzRZ5OIiNXi8rmU7hKnzJkzA+bJ8fDwsDgaEZH0JTQ0FG9v75j3YjH02SQiYo34fC6lu8QpuguEh4eHPpxERCyi7mix6bNJRMRacflcUgdzERERERGRR1DiJCIiIiIi8ghKnERERERERB4h3Y1xEpF72e127ty5Q2RkpNWhSCrn6OiIk5OTxjAlAV2nkhLoGpf0TImTSDoXERHB2bNnCQsLszoUSSPc3NzIkycPzs7OVoeSZug6lZRE17ikV0qcRNKxqKgojh07hqOjI3nz5sXZ2Vm/IkqC2e12IiIiuHjxIseOHaNo0aKa5DYR6DqVlELXuKR3SpxE0rGIiAiioqLw9vbGzc3N6nAkDciYMSMZMmTgxIkTRERE4OrqanVIqZ6uU0lJdI1LeqafCUREvxhKotLrKWnoeZWUQq9FSa/0yhcREREREXkEJU4iIiIiIiKPoMRJRAQoWLAg48ePt/wYIvJgaeUaO378ODabjZ07dwKwevVqbDYbV69etTQuEXk4JU4ikqrYbLaHLkOGDEnQcbds2cLbb7+duMGKpFMp+TqtW7duTByurq4UK1aMgIAA7Hb7Yx1XRNI+VdVLiGvXIHNmq6MQSZfOnj0bsz5v3jwGDRrEgQMHYra5u7vHrNvtdiIjI3FyevRbXc6cORM3UJF0LKVfpx06dGDYsGGEh4ezcuVK3n77bbJkyUKnTp0S5fgikjapxSk+bt+G99+HAgXg5EmroxFJfHY73LhhzRLHX3u9vLxiFk9PT2w2W8zt/fv3kzlzZpYsWULFihVxcXFh3bp1HDlyhJdffpncuXPj7u5O5cqVWbFiRazj/rcLkM1mY/r06TRp0gQ3NzeKFi3K4sWL4/V0njx5kpdffhl3d3c8PDxo0aIF58+fj/n7rl27eOaZZ8icOTMeHh5UrFiRrVu3AnDixAkaNWpE1qxZyZQpE0899RS///57vM4vaZhV12oauU7d3Nzw8vKiQIEC+Pv7U6ZMGQIDA2P+Hh4eTu/evcmXLx+ZMmWiatWqrF69OtYx1q9fT926dXFzcyNr1qz4+flx5coVAJYuXUrNmjXJkiUL2bNn58UXX+TIkSNxeu5EUrwzZ+DFF6F7dzh82OpokpUSp/hwcoK1a+HKFejZ0+poRBJfWBi4u1uzhIUl2sPo378/w4cPZ9++fZQpU4br16/z/PPPExQUxI4dO2jQoAGNGjXi5CN+ABk6dCgtWrRg9+7dPP/887Ru3ZrLly/HKYaoqChefvllLl++zJo1awgMDOTo0aO0bNkyZp/WrVuTP39+tmzZwrZt2+jfvz8ZMmQA4L333iM8PJw//viDv/76ixEjRsT6lV7SOauu1TR2ndrtdtauXcv+/ftxdnaO2d65c2c2btzI3Llz2b17N82bN6dBgwYcOnQIgJ07d1KvXj1KlizJxo0bWbduHY0aNSIyMhKAGzdu0LNnT7Zu3UpQUBAODg40adKEqKioBD5bIinIkCHw22/w2WdQtCi88AIsXQrp4fVtT2dCQkLsgD0kJCRhB9i92253dLTbwW5fujRxgxNJZjdv3rT//fff9ps3b5oN16+b17YVy/Xr8Y5/5syZdk9Pz5jbq1atsgP2n3766ZH3feqpp+wTJ06MuV2gQAH7uHHjYm4D9g8++CDm9vXr1+2AfcmSJQ885t3HWL58ud3R0dF+8uTJmL/v3bvXDtg3b95st9vt9syZM9tnzZp132OVLl3aPmTIkEc+jpTontfVXR77PTiNetjzct/n06prNQ1cp3Xq1LFnyJDBnilTJnuGDBnsgN3V1dW+fv16u91ut584ccLu6Oho/+eff2Ldr169evYBAwbY7Xa7vVWrVvYaNWo8Mv5oFy9etAP2v/76y2632+3Hjh2zA/YdO3bY7fZ/n5MrV67E+ZhWetg1LmncmTN2u7OzeT+oXdtut9n+fX8oVsxunzDBbk9l7+/x+VzSGKf4Kl0aunaFceOgc2fYswdcXKyOSiRxuLnB9evWnTuRVKpUKdbt69evM2TIEH777TfOnj3LnTt3uHnz5iN/yS5TpkzMeqZMmfDw8ODChQtximHfvn14e3vj7e0ds61kyZJkyZKFffv2UblyZXr27En79u2ZPXs2vr6+NG/eHB8fHwC6du1Kp06dWL58Ob6+vjRt2jRWPJLOWXWtppHrtHXr1gwcOJArV64wePBgqlevTvXq1QH466+/iIyMpFixYrHuEx4eTvbs2QHT4tS8efMHHv/QoUMMGjSIP//8k+Dg4JiWppMnT1KqVKmHxiaSoo0fDxERUKMGrFljuupNngwzZsDBg+Y78sCB0Lat+Z78n+sotVPilBBDhsDcuebFMnq0eYGIpAU2G2TKZHUUjy3Tfx5D7969CQwMZPTo0RQpUoSMGTPSrFkzIiIiHnqc6G5z0Ww2W6J2tRkyZAivvfYav/32G0uWLGHw4MHMnTuXJk2a0L59e/z8/Pjtt99Yvnw5AQEBjBkzhi5duiTa+SUVSwPXqpXXqaenJ0WKFAHghx9+oEiRIjz99NP4+vpy/fp1HB0d2bZtG46OjrHuF91dNmPGjA89fqNGjShQoADTpk0jb968REVFUapUqUc+FpEU7epV+OILs96/v/m3SBHTmPDRR/DNNzBxIuzfb/6dOBEaNDDJlJ8fOKT+EUKp/xFYwcMDxowx6598AsePWxqOiDzc+vXradu2LU2aNKF06dJ4eXlxPImv2xIlSnDq1ClOnToVs+3vv//m6tWrlCxZMmZbsWLF6NGjB8uXL+eVV15h5syZMX/z9vbmnXfe4ccff6RXr15MmzYtSWMWsZIV1ymYZKhbt2707t0bu91O+fLliYyM5MKFCxQpUiTW4uXlBZhWrqCgoPse79KlSxw4cIAPPviAevXqUaJEiZiiESKp2hdfmMrSpUrB88/H/pu7O7z7Lvz9NyxfDo0amR94li41+xYvDhMmQGioNbEnEiVOCfXqq1C3Lty8aaqKiEiKVbRoUX788Ud27tzJrl27eO2115J8kLavry+lS5emdevWbN++nc2bN9OmTRvq1KlDpUqVuHnzJp07d2b16tWcOHGC9evXs2XLFkqUKAFA9+7dWbZsGceOHWP79u2sWrUq5m8iaZEV12m0jh07cvDgQRYuXEixYsVo3bo1bdq04ccff+TYsWNs3ryZgIAAfvvtNwAGDBjAli1bePfdd9m9ezf79+/niy++IDg4mKxZs5I9e3a+/PJLDh8+zMqVK+mpglKS2t28aYpBAPTt++DWI5sN6teHxYvh0CFTTM3T06x36wb58kGXLnDX9ASpiRKnhLLZTJ9OJyf4+WdTXUREUqSxY8eSNWtWqlevTqNGjfDz86NChQpJek6bzcbPP/9M1qxZqV27Nr6+vhQuXJh58+YB4OjoyKVLl2jTpg3FihWjRYsWNGzYkKFDhwIQGRnJe++9R4kSJWjQoAHFihXj888/T9KYRaxkxXUaLVu2bLRp04YhQ4YQFRXFzJkzadOmDb169eLJJ5+kcePGbNmyhSeeeAIwLcXLly9n165dVKlShWrVqvHzzz/j5OSEg4MDc+fOZdu2bZQqVYoePXowatSoZHkcIknm66/h/Hl44gnTeBAXPj6mh9bp0/D551CihBmbOWmSaYHy8zPfn1NRNT6b3Z6+psoODQ3F09OTkJAQPDw8Hv+AffvCqFFQqBDs3QuP6PcskpLcunWLY8eOUahQIVxdXa0OR9KIh72uEv09OI142POi61RSGr0m05k7d+DJJ+HoUdPdLqFjbe12WLnSHOOXX/6dF87HxxSS8Pc3rVPJLD6fS2pxelwffmiaHY8dg5EjrY5GRERERCTxLFxokqYcOaBdu4Qfx2aDevVMT63Dh6FXL5MoHTkCPXqY79PvvQf79iVe7IlMidPjypwZxo416wEB5j9fRERERCS1s9th+HCz3qVL4k1JULiwqUz9zz8wZQqULAk3bpgufSVLwnPPmVap/08qnVKkiMRp8uTJFCxYEFdXV6pWrcrmzZsfuO+sWbOw2WyxFsubiZs3B19fCA83JRfTV+9HEREREUmLli+HnTvN9AfvvZf4x8+UCTp2NPOiBgXByy+blqnAQHjpJTMP1NixphR6CmB54jRv3jx69uzJ4MGD2b59O2XLlsXPz++hk9d5eHhw9uzZmOXEiRPJGPF92GxmoFuGDPD776aSiIiIiIhIahbd2vT22/D/CaCThM0Gzz4LP/1kem/17g1Zspgugr16mW58nTqZcucWsjxxGjt2LB06dMDf35+SJUsyZcoU3NzcmDFjxgPvY7PZ8PLyilly586djBE/wJNPmv9kMOUWw8KsjUdEREREJKH+/BNWrzYVpHv0SL7zFipkCq+dPg1Tp5p5o8LCTJe+p54yvbx+/tmSbnyWJk4RERFs27YNX1/fmG0ODg74+vqycePGB97v+vXrFChQAG9vb15++WX27t37wH3Dw8MJDQ2NtSSZgQPB2xtOnDDjnUREREREUqMRI8y/r79uvt8mt0yZTEvX7t2mGl+TJmb+qKAgaNwYihQx5c6TcYJpSxOn4OBgIiMj72kxyp07N+fOnbvvfZ588klmzJjBzz//zJw5c4iKiqJ69eqcPn36vvsHBATg6ekZs3gn5X98pkwwfrxZHznSTPYlIiIiIpKa7NsHixaZ9b59rY3FZoNnnoEffzRd9/r2haxZ4fhx09srf34zTmr//iQPxfKuevFVrVo12rRpQ7ly5ahTpw4//vgjOXPmZOrUqffdf8CAAYSEhMQsp06dStoAmzSBBg0gIsLUpFehCBERERFJTaInbW7c2Excm1IUKGBawk6fhmnToHRp043vyy9N18IkZmnilCNHDhwdHTl//nys7efPn8fLyytOx8iQIQPly5fn8OHD9/27i4sLHh4esZYkZbOZib2cnU0lkuhsXURSlLp169K9e/eY2wULFmR8dIvxA9hsNn766afHPndiHedhhgwZQrly5ZL0HCJJLa1fpylFQp5nScNOn4Y5c8x6v37WxvIgbm7Qvj3s2mXGYbVpAy1bJvlpLU2cnJ2dqVixIkFBQTHboqKiCAoKolq1anE6RmRkJH/99Rd58uRJqjDjr2jRf19o3bubuvQikigaNWpEgwYN7vu3tWvXYrPZ2L17d7yPu2XLFt5+++3HDS+WByUvZ8+epWHDhol6Lklc8Zkm425z587FZrPRuHHjpA0whdN1Gjd3T7Hi4OBAnjx5aNmyJSdPnkzS84o81LhxcPs21KkDTz9tdTQPZ7OZOL/+GpJheiLLu+r17NmTadOm8fXXX7Nv3z46derEjRs38Pf3B6BNmzYMGDAgZv9hw4axfPlyjh49yvbt23n99dc5ceIE7du3t+oh3F///lCwIJw6BR9/bHU0ImlGu3btCAwMvO+4xpkzZ1KpUiXKlCkT7+PmzJkTt8Sa2O8RvLy8cHFxSZZzSfwlZJoMgOPHj9O7d29q1aqVTJGmXLpO4y56ipV//vmHhQsXcuDAAZo3b57k5xW5r8uXTSU7MN9lJRbLE6eWLVsyevRoBg0aRLly5di5cydLly6NKRhx8uRJzp49G7P/lStX6NChAyVKlOD5558nNDSUDRs2ULJkSasewv25ucFnn5n1MWOSZcCaSHrw4osvkjNnTmbNmhVr+/Xr15k/fz7t2rXj0qVLtGrVinz58uHm5kbp0qX5/vvvH3rc/3ZNOXToELVr18bV1ZWSJUsSGBh4z3369etHsWLFcHNzo3Dhwnz44Yfcvn0bML8kDx06lF27dsX8ohwd83+7AP311188++yzZMyYkezZs/P2229z/fr1mL+3bduWxo0bM3r0aPLkyUP27Nl57733Ys4VF1FRUQwbNoz8+fPj4uJCuXLlWLp0aczfIyIi6Ny5M3ny5MHV1ZUCBQoQ8P/qoHa7nSFDhvDEE0/g4uJC3rx56dq1a5zPndokZJqMyMhIWrduzdChQylcuHAyRpsy6TqN+3UaPcVKnjx5qF69Ou3atWPz5s2xqgD//PPPVKhQAVdXVwoXLszQoUO5c+dOzN+vXr1Kx44dyZ07N66urpQqVYpff/0VIEHPs6Rjn39uekqVLQt+flZHk+I4WR0AQOfOnencufN9/7Z69epYt8eNG8e4ceOSIapE8NJL8OKL8OuvplBEYKBpUhRJoex2O2G3rZmDzC2DG7Y4XB9OTk60adOGWbNmMXDgwJj7zJ8/n8jISFq1asX169epWLEi/fr1w8PDg99++4033ngDHx8fqlSp8shzREVF8corr5A7d27+/PNPQkJCYvX/j5Y5c2ZmzZpF3rx5+euvv+jQoQOZM2emb9++tGzZkj179rB06VJWrFgBgKen5z3HuHHjBn5+flSrVo0tW7Zw4cIF2rdvT+fOnWN96Vy1ahV58uRh1apVHD58mJYtW1KuXDk6dOjwyMcD8NlnnzFmzBimTp1K+fLlmTFjBi+99BJ79+6laNGiTJgwgcWLF/PDDz/wxBNPcOrUqZhiOgsXLmTcuHHMnTuXp556inPnzrFr1644nTe1iZ4m4+6eDnGZJmPYsGHkypWLdu3asXbt2iSP06prVddp0l6nFy5cYNGiRTg6OuLo6AiYro1t2rRhwoQJ1KpViyNHjsR0Vxw8eDBRUVE0bNiQa9euMWfOHHx8fPj7779j7n/r1q3Hep4lHQkL+/dH/3799J31PlJE4pSmffYZrFhhas7Pnw8tWlgdkcgDhd0Owz3A3ZJzXx9wnUzOmeK071tvvcWoUaNYs2YNdevWBUz3n6ZNm8ZMPdA7ekJqoEuXLixbtowffvghTl8UVqxYwf79+1m2bBl58+YF4NNPP71nvMMHH3wQs16wYEF69+7N3Llz6du3LxkzZsTd3R0nJ6eHFrv57rvvuHXrFt988w2ZMpnHP2nSJBo1asSIESNiWt+zZs3KpEmTcHR0pHjx4rzwwgsEBQXF+QvZ6NGj6devH6+++ioAI0aMYNWqVYwfP57Jkydz8uRJihYtSs2aNbHZbBQoUCDmvidPnsTLywtfX18yZMjAE088kWa/cD1smoz9D+g5sG7dOr766it27twZ5/OEh4cTHh4eczu+cwxada3qOk386zQkJAR3d3eTDIeZZLhr164x5xk6dCj9+/fnzTffBKBw4cJ89NFH9O3bl8GDB7NixQo2b97Mvn37KFasWMw+0fLly/dYz7OkIzNmQHCwmYBW3UXvy/Kuemle4cIQ/ctljx5w7Zq18YikAcWLF6d69eoxXacOHz7M2rVradeuHWC6TX300UeULl2abNmy4e7uzrJly+I84Hrfvn14e3vHfBkD7luwZt68edSoUQMvLy/c3d354IMP4j2oe9++fZQtWzbmSxJAjRo1iIqK4sCBAzHbnnrqqZhfkAHy5MnzyDE30UJDQzlz5gw1atSItb1GjRrs27cPMN2Mdu7cyZNPPknXrl1Zvnx5zH7Nmzfn5s2bFC5cmA4dOrBo0aJY3YTSs2vXrvHGG28wbdo0cuTIEef7JescgxbRdRq36zRz5szs3LmTrVu3MmbMGCpUqMAnn3wS8/ddu3YxbNgw3N3dY5YOHTpw9uxZwsLC2LlzJ/nz549Jmv7rcZ9nSSdu34bRo816nz7gpLaV+9Gzkhz69jXVPo4ehWHD/q2NL5LCuGVw4/qA64/eMYnOHR/t2rWjS5cuTJ48mZkzZ+Lj40OdOnUAGDVqFJ999hnjx4+ndOnSZMqUie7duxMREZFo8W7cuDFmTIufnx+enp7MnTuXMWPGJNo57pYhQ4ZYt202G1FRUYl2/AoVKnDs2DGWLFnCihUraNGiBb6+vixYsABvb28OHDjAihUrCAwM5N13341pSfhvXKldfKfJOHLkCMePH6dRo0Yx26L/X5ycnDhw4AA+Pj733G/AgAH07Nkz5nZoaGi8kierrlVdpw+XkOvUwcGBIkWKAFCiRAmOHDlCp06dmD17NmDGhQ0dOpRXXnnlnvu6urqSMWPGhx4/OZ5nSQN++AFOnIBcuaBtW6ujSbGUOCUHV1eYOBFeeAHGjzcvyKeesjoqkXvYbLY4d8OxWosWLejWrRvfffcd33zzDZ06dYoZR7F+/XpefvllXn/9dcB8kT148GCci8iUKFGCU6dOcfbs2ZipDjZt2hRrnw0bNlCgQAEGDhwYs+3EiROx9nF2diYyMvKR55o1axY3btyI+TV7/fr1ODg48OSTT8Yp3kfx8PAgb968rF+/PuZLa/R57u6q4+HhQcuWLWnZsiXNmjWjQYMGXL58mWzZspExY0YaNWpEo0aNeO+99yhevDh//fUXFSpUSJQYU4q7p8mILikePU3G/cbiRj8Pd/vggw+4du0an3322QOTIRcXl8eq2JZarlVdp/HXv39/fHx86NGjBxUqVKBChQocOHAgJrn6rzJlynD69GkOHjx431anx32eJR2w22H4cLPerRs8IhlPz9RVL7k8/7yZffnOHXjvPfMiFZEEc3d3p2XLlgwYMICzZ8/S9q5fyIoWLUpgYCAbNmxg3759dOzY8Z4WhIfx9fWlWLFivPnmm+zatYu1a9fG+uIVfY6TJ08yd+5cjhw5woQJE1j0nwmvCxYsyLFjx9i5cyfBwcGxxrREa926Na6urrz55pvs2bOHVatW0aVLF9544417xtk8jj59+jBixAjmzZvHgQMH6N+/Pzt37qRbt26AqST3/fffs3//fg4ePMj8+fPx8vIiS5YszJo1i6+++oo9e/Zw9OhR5syZQ8aMGWONg0pL4jNNRnQFs7uXLFmykDlzZkqVKoWzs7OVD8Vyuk7jz9vbmyZNmjBo0CAABg0axDfffMPQoUPZu3cv+/btY+7cuTFjt+rUqUPt2rVp2rQpgYGBMS3H0VUzH/d5lnTg999hzx7InBnefdfqaFI0JU7Jadw4k8WvWQMqBSry2Nq1a8eVK1fw8/OLNc7hgw8+oEKFCvj5+VG3bl28vLziNSGpg4MDixYt4ubNm1SpUoX27dvHGnMA8NJLL9GjRw86d+5MuXLl2LBhAx9++GGsfZo2bUqDBg145plnyJkz531LALu5ubFs2TIuX75M5cqVadasGfXq1WPSpEnxezIeoWvXrvTs2ZNevXpRunRpli5dyuLFiylatChgxlmMHDmSSpUqUblyZY4fP87vv/+Og4MDWbJkYdq0adSoUYMyZcqwYsUKfvnlF7Jnz56oMaYU8Z0mQx5O12n89ejRg99++43Nmzfj5+fHr7/+yvLly6lcuTJPP/0048aNi/XDxcKFC6lcuTKtWrWiZMmS9O3bN6YV7XGfZ0kHolubOnaELFksDSWls9nt6avpIzQ0FE9PT0JCQvDw8Ej+AD75BD74ALy8zNxO9yl7KpJcbt26xbFjxyhUqBCuyTDjtqQPD3tdWf4enEI97HnRdSopjV6Tacj69VCzJjg7w7FjcNePG+lFfD6X1OKU3Hr3hqJF4dw5GDLE6mhEREREJL0aMcL826ZNukya4kuJU3JzcYHopv2JE2H3bmvjEREREZH0Z88e+OUXM9Ftnz5WR5MqKHGywnPPQbNmEBmpQhEiIiIikvyip8d55RV4wDxgEpsSJ6uMHQtubrBuHfx/rgYRERERkSR34gR8951Z79fP2lhSESVOVvH2hv+XGqVPH7h61dJwRERERCSdGDvWTJFTrx5Urmx1NKmGEicr9egBxYvDhQvwn/KoIskpnRXXlCSm11PS0PMqKYVei6lccDBMm2bW1doUL0qcrOTsDJMnm/XPP4cdO6yNR9KdDBkyABAWFmZxJJKWRL+eol9f8nh0nUpKo2s8lZs0CW7ehAoVwNfX6mhSFSerA0j3nn0WXn0V5s41szWvXw8OymcleTg6OpIlSxYuXLgAmEkebTabxVFJamW32wkLC+PChQtkyZIFR0dHq0NKE3SdSkqhazwNuH7dVHUG6N/fVNSTOFPilBKMHg2//gqbNsGsWfDWW1ZHJOmIl5cXQMyXMpHHlSVLlpjXlSQOXaeSkugaT8WmT4fLl6FIEVNNT+JFiVNKkC+fmQy3d2/T17RxY8iWzeqoJJ2w2WzkyZOHXLlycfv2bavDkVQuQ4YM+hU6Ceg6lZRC13gqFhFhikKAKUym/8d4U+KUUnTtCjNnwt69MHAgfPGF1RFJOuPo6KgPQ5EUTtepiCTY99/DqVPg5QVt2lgdTaqkwTQpRYYM/xaKmDoVtm61Nh4RERERSRuiomDECLPeowe4ulobTyqlxCklqVMHXn8d7HZTKCIy0uqIRERERCS1++UX2LcPPD3hnXesjibVUuKU0owaBR4esGULfPWV1dGIiIiISGpmt8Pw4Wa9UyfzPVMSRIlTSuPlBcOGmfUBA8wkZSIiIiIiCbF2ranc7OIC3bpZHU2qpsQpJXrvPShTxpSLHDDA6mhEREREJLWKbm3y9zc/0EuCKXFKiZyc/i0UMX26+ZVARERERCQ+du2CJUvAwcFMeyOPRYlTSlWzJrz5pll/7z0VihARERGR+Bk50vzbvDn4+FgbSxqgxCklGznSVD/Zvt2UKBcRERERiYujR2HuXLPer5+1saQRSpxSsly54JNPzPrAgXDhgrXxiIiIiEjqMGaMmb/Jzw/Kl7c6mjRBiVNK98475sV+9ap+LRARERGRR7twAWbMMOv6/pholDildI6O8PnnZn3WLFi/3tJwRERERCSFmzABbt2CKlWgbl2ro0kzlDilBk8/De3amfV334U7d6yNR0RERERSptDQf6sz9+8PNpu18aQhSpxSi+HDIWtW2L373xYoEREREZG7ffmlGeLx5JPw8stWR5OmKHFKLXLkgIAAs/7hh3DunLXxiIiIiEjKEh4O48aZ9b59zfxNkmj0bKYm7dtD5cqmCbZPH6ujEREREZGUZM4cOHMG8uWD1q2tjibNUeKUmjg6mj6rNpu5MP74w+qIRERERCQliIz8d8Lbnj3BxcXaeNIgJU6pTeXK8PbbZv3dd+H2bWvjERERERHr/fQTHDwIWbJAhw5WR5MmKXFKjT79FLJnh717YeJEq6MRERERESvZ7TBihFnv3BkyZ7Y2njRKiVNqlC3bvxfH4MGmL6uIiIiIpE+rVsGWLZAxI3TtanU0aZYSp9TK39/M73T9Orz+OuzbZ3VEIiIiImKF4cPNv2+9BTlzWhtLGqbEKbVycDDzOWXIYH5leOopaNoUtm2zOjIRERERSS7bt0NgoCki1quX1dGkaUqcUrPy5eHPP+GVV0zf1h9/hEqVoEEDVdwTERERSQ+ih2+8+ioUKmRtLGmcEqfUrnx5WLjQFIp44w3za8OyZVCnDtSqBUuWmKRKRERERNKWQ4dgwQKz3q+ftbGkA0qc0oqSJeGbb8wF9M474OwM69bB889DhQowf76p7y8iIiIiacPo0RAVZb7vlS5tdTRpnhKntKZQIfjiCzh+HHr3hkyZYOdOaNHCjIOaNUtzP4mIiIikdmfPmu91AP37WxpKeqHEKa3KkwdGjYITJ0zJ8qxZ4cABU42vSBGYNAlu3rQ6ShERERFJiM8+g4gIqF4data0Opp0QYlTWpc9OwwZYhKoUaPAywtOnoQuXaBgQTOgMDTU6ihFREREJK5CQkwPIzCtTTabtfGkE0qc0ovMmU3XvWPHTBnzggXhwgVzsRUoAB9+CMHBVkcpIiIiIo8yZYr54btkSXjhBaujSTeUOKU3rq7QqRMcPAhffw3Fi8PVq/DxxyaB6tkT/vnH6ihFRERE5H5u3YJx48x6v35mbk9JFnqm06sMGaBNG1PGfOFCU3kvLMxciIULQ8eOcOSI1VGKiIiIyN2+/hrOnwdvb2jVyupo0hUlTumdg4OZQHfrVli6FGrXNgMNv/wSihWD1q1hzx6roxQRERGRyEgzZh2gVy/zQ7gkGyVOYths4OcHa9bA2rXQsKGZF+C778y8AI0bw+bNVkcpIiIikn4tXGh6BGXLBu3bWx1NuqPESe5Vsyb8/jts3w7Nm5uk6uefoWpV8PWFlSvBbrc6ShEREZH0w26H4cPNeteuZq5OSVZKnOTBypeHH36AffugbVtwcoKgIKhXz8wZ8MsvSqBEREREkkNgIOzYAW5u0Lmz1dGkS0qc5NGefBJmzoTDh+G998DFBTZtgpdegrJl4fvvTZ9bEREREUka0a1NHTqYeTol2SlxkrgrUAAmTYLjx035y8yZ4a+/4LXXTFnz6dMhPNzqKEVERETSls2bYdUq0/unZ0+ro0m3lDhJ/Hl5mV89TpyAYcPMAMXDh80vID4+8NVXaoESERERSQyRkTB0qFlv3RqeeMLaeNIxJU6ScFmzwocfmgRq7FjIm9dMntu+vZkXasUKqyMUERERSb1u3DDTxvz+u5lCpm9fqyNK15Q4yeNzd4cePeDoUZNAZckCu3dD/frw4oumuISIiIiIxN2ZM2Z+zcWLzfjyb7+FkiWtjipdU+IkicfFxSRQhw9Dt26mH+5vv5l5oN57Dy5etDpCERERkZRv1y4zDcz27ZAjh5kK5tVXrY4q3VPiJIkve3YYPx727oWXXzZ9cz//HIoUMbNd37pldYQiIiIiKdPvv5s5NU+fNsW3/vzTTAMjllPiJEmnWDH46SfzK0m5chAaavrmliwJ8+drDigRERGRu02aBI0awfXr8OyzsGEDFC5sdVTyf0qcJOk98wxs3WrmgsqbF44dgxYtzK8pf/5pdXQiIiIi1oqMNMMcunSBqCh46y1YssQU4pIUQ4mTJA9HR2jbFg4ehCFDzKzXGzbA00+beaBOnLA6QhEREZHkd/06NG4MEyaY2wEBZm5MZ2dLw5J7KXGS5JUpEwwebBKotm3BZoPvv4cnn4QBA0x3PhEREZH04PRpqFULfv0VXF3hhx+gf3/z/UhSHCVOYo18+UzXvW3bTFe+8HAzqW6RIjB1Kty5Y3WEIiIiIklnxw5TOW/nTsiVC1atgubNrY5KHkKJk1irfHkICoKffzbFJC5ehHfegbJlYelSq6MTERERSXy//GLGep85Y4pm/fmnGb4gKZoSJ7GezQYvvQR79pj+vdmywd9/Q8OG0KCB2S4iIiKS2tnt8NlnZrqWsDDw9YX166FgQasjkzhQ4iQpR4YMpprM4cPQs6e5vWyZaX3q2BHOn7c6QhEREZGEuXPHfM/p3t0kUB06mDmbsmSxOjKJoxSROE2ePJmCBQvi6upK1apV2bx5c5zuN3fuXGw2G40bN07aACV5Zc0KY8bAvn3QtKkpy/nll2b8U0AA3LxpdYQiIiIicRcaanrXTJ5setqMHGnGdGfIYHVkEg+WJ07z5s2jZ8+eDB48mO3bt1O2bFn8/Py4cOHCQ+93/PhxevfuTa1atZIpUkl2Pj6wYAH88QdUqmTKdb7/vplF+/vvNYGuiIiIpHynTpnxTEuWQMaM5rtNnz6qnJcKWZ44jR07lg4dOuDv70/JkiWZMmUKbm5uzJgx44H3iYyMpHXr1gwdOpTCmk057atVywyanDMH8ueHkyfN3E9PP236BYuIiIikRNu2mcp5f/0FuXPDmjXwyitWRyUJZGniFBERwbZt2/D19Y3Z5uDggK+vLxs3bnzg/YYNG0auXLlo167dI88RHh5OaGhorEVSIQcHaN0aDhyAjz8Gd3fYvNn8gtOiBRw9anWEIiIiIv/66SeoXRvOnoVSpcyPwJUrWx2VPAZLE6fg4GAiIyPJnTt3rO25c+fm3Llz973PunXr+Oqrr5g2bVqczhEQEICnp2fM4u3t/dhxi4Xc3GDgQDh0yAyqdHCA+fOhRAnT7H31qtURioiISHpmt5ux2q+8Yirn+fmZHjIFClgdmTwmy7vqxce1a9d44403mDZtGjly5IjTfQYMGEBISEjMcurUqSSOUpKFl5cpGLFjhynlGREBo0ebAhKTJ8Pt21ZHKCIiIunNnTvQqRP07m0SqHfegV9/BQ8PqyOTROBk5clz5MiBo6Mj5/9TZvr8+fN4eXnds/+RI0c4fvw4jRo1itkWFRUFgJOTEwcOHMDHxyfWfVxcXHBxcUmC6CVFKFMGli83Ay579zaV+Dp3hokTTSL1wgsafCkiIiJJLzTUDB9Ytsx89xgzxpQe1/eQNMPSFidnZ2cqVqxIUFBQzLaoqCiCgoKoVq3aPfsXL16cv/76i507d8YsL730Es888ww7d+5UN7z0ymaD55+H3bvh888hRw4zFqpRI9MatWWL1RGKiIhIWnbiBNSoYZImNzdYtAh69FDSlMZY2uIE0LNnT958800qVapElSpVGD9+PDdu3MDf3x+ANm3akC9fPgICAnB1daVUqVKx7p/l/5OG/Xe7pENOTqZ5/LXX4NNPYfx4WLkSqlSBZs3gk0+gWDGroxQREZH/sNvtjFw/km1nt1kdSvxdvgzr1kGJcCjvCjWrQ8S3MP9bqyNLV/pU70PlfElbfMPyxKlly5ZcvHiRQYMGce7cOcqVK8fSpUtjCkacPHkSB4dUNRRLrObpCSNGmCRq8GCYPdvMmbBoEbRrB4MGQb58VkcpIiIi/7f7/G76B/W3OoyEKxK9cgvOroCzVgaTPrUu3TrJz2Gz29PXLKKhoaF4enoSEhKChwbqpQ9//WUq8f3yi7nt6grdukG/fpA1q7WxiaQzeg++Pz0vkt4NXzecAUEDqJS3Em+WfdPqcB7NboegIPj5Z3O7ZEnw9zffMcQSzxd9nsJZ4z+/a3zefy1vcRJJcqVLw+LFphm9f39TEnTECJg61dzu0sX0RxYR+b/JkyczatQozp07R9myZZk4cSJVqlS5774//vgjn376KYcPH+b27dsULVqUXr168cYbbyRz1CKp19LDSwFoW7Yt71V5z+JoHuH2bXj3XZj+/6Spc2cYN84MGZA0TX3gJP2oWRPWrjUtT6VKmTmf+veHokVNaXOVMBcRYN68efTs2ZPBgwezfft2ypYti5+fHxcuXLjv/tmyZWPgwIFs3LiR3bt34+/vj7+/P8uWLUvmyEVSp9DwUNafWg9AgyINLI7mEa5eNQWppk83hR8++8xU8lXSlC4ocZL0xWaDF1+EnTvh66/NZHRnzkDHjiaZmj/fNL+LSLo1duxYOnTogL+/PyVLlmTKlCm4ubkxY8aM++5ft25dmjRpQokSJfDx8aFbt26UKVOGdevWJXPkIqnTymMruRN1hyLZiuCTzefRd7DKsWOmct6KFZApk+mm17Wr1VFJMlLiJOmToyO0aWPKlo8fb0qYHzxo5l+oUsX0WxaRdCciIoJt27bh6+sbs83BwQFfX182btz4yPvb7XaCgoI4cOAAtWvXTspQRdKM6G56DXxScGvTpk3w9NPw99+QN6/pwXLXvKKSPihxkvTNxcUUijh61FTgc3eHrVvN/E/168O2VFgWVUQSLDg4mMjIyJjKrtFy587NuXPnHni/kJAQ3N3dcXZ25oUXXmDixInUr1//gfuHh4cTGhoaaxFJj+x2+7+JU0rtpjd/PjzzDFy4AOXKwebNUL681VGJBZQ4iQBkzgxDhsCRI6bZPUMG0xRfqZJphTp40OoIRSQFy5w5Mzt37mTLli188skn9OzZk9WrVz9w/4CAADw9PWMWTeAu6dWBSwc4EXICZ0dn6hasa3U4sdntEBBgvgfcumVamNau1ZQm6ZgSJ5G75cplBnoeOABvvGHGRM2fb8qMduxoxkOJSJqVI0cOHB0dOX/+fKzt58+fx8vL64H3c3BwoEiRIpQrV45evXrRrFkzAgICHrj/gAEDCAkJiVlOnTqVaI9BJDWJbm2qU6AOmZwzWRzNXSIioH17eP99c7tbNzMfpLu7tXGJpZQ4idxPoULwzTemiMQLL0BkpKm8V6QIDBgAV65YHaGIJAFnZ2cqVqxI0F3jHKOioggKCqJatWpxPk5UVBTh4eEP/LuLiwseHh6xFpH0KMV20xs4EGbMAAcHmDTJjId2dLQ6KrGYEieRhylTBn79Ff74A6pXh5s3Yfhw8PGBkSPNbRFJU3r27Mm0adP4+uuv2bdvH506deLGjRv4+/sD0KZNGwYMGBCzf0BAAIGBgRw9epR9+/YxZswYZs+ezeuvv27VQxBJFW7evsmaE2uAFJY47d9vEiWA77+H91L4vFKSbFR0XiQuatUyE+j++qtpcdq7F/r1M936hgwxs4VrDgeRNKFly5ZcvHiRQYMGce7cOcqVK8fSpUtjCkacPHkSB4d/f3e8ceMG7777LqdPnyZjxowUL16cOXPm0LJlS6segkiqsObEGm7duYW3hzclcpSwOhzDbjfd8u7cMWOaWrSwOiJJQWx2e/qatCY0NBRPT09CQkLUNUISJjISvv0WBg2CEyfMtmLF4JNPoGlTMy5KRO5L78H3p+dF0qPuS7vz2Z+f0aFCB75s9KXV4RiLF8PLL4Ozs/mRtEgRqyOSJBaf91911ROJrwfNAdW8ueaAEhERiaMUN77p1i3o0cOs9+yppEnuocRJJKGi54A6csS0PmXK9O8cUM89pzmgREREHuDYlWMcuHQAR5sj9QrVszocY9w4M69j3rymOITIfyhxEnlcHh4wdKh5s+3SxcwBFRho5oBq2RIOHbI6QhERkRRl2ZFlAFT3ro6nq6fF0QD//GO63AOMGKGy43JfSpxEEkuuXDBhgunC9/rrZqzTDz9AiRLwzjuaA0pEROT/Ulw3vX794MYNqFYNWre2OhpJoZQ4iSS2QoVg9uzYc0BNnWr6Sg8davpQi4iIpFMRkREEHTPjgVNE4rR+vSn6ZLPBxIkq8iQPpMRJJKncbw6oIUOgdGnTlU9ERCQd2nBqA9cjrpMrUy7KeZWzNpjISNPNHqBdO6hY0dp4JEVT4iSS1KLngJo3D/LkgcOHTfGIli3VfU9ERNKd6G56fj5+ONgs/io6Ywbs2AGenv+OcRJ5ACVOIsnBZjOT6O3fbyrxOTiY8U/Fi5tJdO/csTpCERGRZLHk8BIgBXTTu3IF3n/frA8dasYqizyEEieR5OThYeZ+2roVqlaFa9ege3eoXBk2bbI6OhERkSR15toZdp/fjQ0b9QvXtzaYIUMgOBhKloR337U2FkkVlDiJWKF8ediwwRSNyJrVFJKoXh06doTLl62OTkREJEksO2zKkFfKW4mcmXJaF8jevTB5slkfP95MJSLyCEqcRKzi4ABvv2267735Jtjt8OWXpvve11+b2yIiImnI0iMpoAy53W66zUdGQpMmUN/ili9JNZQ4iVgtVy6YNQvWrDHdBS5ehLZtoU4d84uYiIhIGnAn6g6BR0xVWUsTp0WLICgIXFxgzBjr4pBUR4mTSEpRu7bpsjdiBLi5wdq1UK7cv5PyiYiIpGJb/tnClVtXyOqalSr5qlgTxM2b0KuXWe/Tx8y9KBJHSpxEUpIMGaBvX/j7b3j5ZVNtb+RI0xL1889WRyciIpJg0WXI6/vUx8nByZogRo+G48chf37o39+aGCTVUuIkkhIVKAA//QSLF5v1kyehcWN46SXzhi8iIpLKxIxv8rGom97JkxAQYNZHjYJMmayJQ1ItJU4iKVmjRmacU//+4OQEv/xiWp+GD4eICKujExERiZPgsGC2/LMFAL8iftYE0bev6apXq5aZhF4knpQ4iaR0mTKZX8h27TIFI27ehAEDzPin1autjk5EROSRAo8EYsdOmdxlyJs5b/IHsGYNzJtnKtpOmGAmpheJJyVOIqlFyZKwahV88w3kzAn79sEzz0CbNnDhgtXRiYiIPJCl3fTu3IGuXc3622+bHx5FEkCJk0hqYrPBG2/AgQPwzjvm9uzZ8OSTMGWKmZNCREQkBYmyR8VMfGtJGfJp02D3bjPh/EcfJf/5Jc1Q4iSSGmXNCl98AZs2QYUKcPUqdOoE1avD9u1WRyciIhJj17ldnL9xnkwZMlHjiRrJe/LLl+GDD8z6Rx9BjhzJe35JU5Q4iaRmVarA5s2mv7aHh1mvXNl0SQgJsTo6ERGRmDLk9QrXw9nROXlPPmiQSZ5Kl4aOHZP33JLmKHESSe0cHaFLF9i/H159FaKiYOJEKF4c5s4Fu93qCEVEJB2zbHzT7t2mdwbAZ5+Z6rQij0GJk0hakScPfP89LF8ORYvCuXPQqhU89xwcPGh1dCIikg6F3Aphw6kNQDKXIbfbTe+LqCho3twUUxJ5TEqcRNKa+vXNr2zDhoGLC6xYYbooDBpkSpmLiIgkk5XHVnIn6g7FshejcNbCyXfiBQtMCXJXVzPZrUgiUOIkkha5usKHH5rJcxs0MJPlfvSRSaCWLrU6OhERSSeixzclaze9sDDo1cus9+8PBQok37klTVPiJJKW+fjA77/D/PmQLx8cOQING5puC//8Y3V0IiKShtnt9n/HNyVnGfIRI+DUKZMw9e2bfOeVNE+Jk0haZ7NBs2ZmwtyePU0xiQULoFgx84vcmTNWRygiImnQvuB9nAw5iYujC3UK1kmekx4/DiNHmvXRoyFjxuQ5r6QLSpxE0ovMmWHMGNi2zcz3FBYGY8dCoUJmMt2jR62OUERE0pDobnp1CtbBLYNb8py0d2+4dcsUg2jaNHnOKemGEieR9KZsWVi3DpYsgVq1zPinqVNNC9Qbb5hxUSIiIo8p2cc3rVwJCxeanhWffWZ6XIgkIiVOIumRzWaKRvzxh1kaNIDISJgzB0qVgiZNYMsWq6MUEZFU6kbEDdacWANAw6INk/6Ed+6Y8uMAnTqZYkgiiUyJk0h6V6uWaX3ats2MhbLZ4KefoEoVMwfU6tWaRFdEROJlzYk1RERGUMCzAE9mfzLpT/jFF6bHRPbsMHRo0p9P0iUlTiJiVKhgqu/t3Qtvvmm6OgQGmn7iNWrAr78qgRIRkTiJ6aZXpAG2pO4yd/GimasQ4JNPIFu2pD2fpFtKnEQkthIlYNYsOHwY3n3XTKK7cSM0agTlysG8eaZbn4iIyAPcnTgluQ8/hKtXzWdU+/ZJfz5Jt5Q4icj9FSwIkyeb0q59+4K7O+zeDa++apKrGTNMYQkREZG7HLl8hEOXD+Hk4MSzhZ5N2pPt2AFffmnWJ0wwvSVEkogSJxF5OC8vM5ngyZOm33i2bHDoELRrZybYnTDBlDYXEREBlh1ZBkAN7xp4uHgk3YnsdujSxfzbqpUZsyuShJQ4iUjcZM1q+pCfOGEmFcyTB06fhm7dTOtUQACEhFgdpYiIWCzZuunNnQvr14Ob27+T3ookISVOIhI/7u7Qq5eZMHfKFDOB7sWL8P77UKAAfPCBuS0iIulO+J1wVh5bCSRx4nT9OvTpY9bffx/y50+6c4n8nxInEUkYV1fo2BEOHoTZs6FkSdPi9MknJoHq3t20SImISLqx/tR6bty+gZe7F2Vzl026EwUEwD//mB/vevVKuvOI3EWJk4g8HicneP11+OsvWLQIKlWCmzfNrO2FC0OHDqZCn4iIpHnR3fT8fPySrgz5kSOmyzjA2LHmhzyRZKDESUQSh4MDNG4MmzfD8uVQpw7cvg3Tp8OTT8Jrr5nkSkRE0qxkGd/Uq5ep6lq/Prz8ctKdR+Q/lDiJSOKy2cyH2erVsG4dvPACREXB999DmTLw0kuwaZPVUYqISCL7J/Qf/rrwFzZs1C9cP2lOsnw5/Pyz6e3w2WfmM0ckmShxEpGkU6MG/PqrmWejRQvzAffLL1CtGtSrB0FBpoysiIiketFlyKvkq0J2t+yJf4Lbt00lV4DOnc2cgiLJSImTiCS9cuVg3jzYvx/eesv8UrhyJfj6wtNPw+LFplVKRERSrSTvpjdpkvkcyZkTBg9OmnOIPIQSJxFJPsWKwVdfmYG9XbqYAb2bN5s+6qVLw4wZEB5udZQiIhJPd6LuEHg0EEiixOnCBRgyxKwHBECWLIl/DpFHUOIkIsnviSdgwgQ4fhz694fMmeHvv6FdO1PK/JNP4NIlq6MUEZE42vzPZq7eukpW16xUzls58U/w/vsQGgoVK4K/f+IfXyQOlDiJiHVy5za/HJ46BaNGmQkMz583k+g+8YRplTpyxOooRUTkEZYcWgLAcz7P4ejgmLgH37rV9EgA86Obg76+ijX0yhMR63l6Qu/ecPQozJljxkSFhZn+7EWLQrNmsHGj1VGKiMgDLD1ixjc1LNIwcQ8cFQVdu5pCQm+8AdWrJ+7xReJBiZOIpBwZMkDr1rB9u6m417Ch+bBcuNB8WNaoYSbZjYy0OlIREfm/CzcusPXMVsC0OCWqb781P5xlygTDhyfusUXiycnqAERE7mGzwbPPmmXvXjMz/Jw5sGEDvPIKFCkCPXpA27bg5mZ1tGKh3bt3x3nfMmXKJGEkIulX4BFTFKKcVznyZM6TeAe+dg369jXrH34IefMm3rFFEkCJk4ikbE89ZSrxffKJ6br3+edw+DC89x4MGgSdOpn5PHLntjpSsUC5cuWw2WzYHzAfWPTfbDYbkWqpFEkS0d30GvgkcjW9Tz6Bc+fMj2XduyfusUUSwGZ/0KdNGhUaGoqnpychISF4eHhYHY6IxNeNGzBzJowbZ8ZEAbi4mL7vPXtqQsQULrHfg0+cOBHnfQsUKPDY50sq+myS1CrKHoXXaC8uhl1k9ZurqVOwTuIc+NAh88PZ7dtm4vQXX0yc44r8R3zef9XiJCKpS6ZMpoWpUyf46ScYPRo2bYLp083ywgum0ESdOqbLn6RpKTkZEkkPdpzdwcWwi2R2zkw172qJd+AePUzS1LCheV8XSQGUOIlI6uToCE2bmmXDBpNA/fQT/PabWSpUMAlUs2am6ISkSYsXL47zvi+99FISRiKSPi09bLrp1StcD2dH58Q56O+/m/fxDBlM7wL9CCYphBInEUn9qleHH380XTvGjzdd+bZvh9deMxPsdusG7duDukClOY0bN47TfhrjJJI0En18U0TEv+OZunWDJ59MnOOKJIIElSM/deoUp0+fjrm9efNmunfvzpdffplogYmIxFvRojB5Mpw8CR99BLlymfVevcDb21Rnuuu9S1K/qKioOC1KmkQS39VbV9l4ysyx51fEL3EO+tln5kew3LlNJT2RFCRBidNrr73GqlWrADh37hz169dn8+bNDBw4kGHDhiVqgCIi8ZYjB3zwAZw4AdOmQfHiEBoKo0ZBoUKmkMTOnVZHKSKSqgUdDSLSHknxHMUpmKXg4x/w7FmI/h45YoR6CUiKk6DEac+ePVSpUgWAH374gVKlSrFhwwa+/fZbZs2aFe/jTZ48mYIFC+Lq6krVqlXZvHnzA/f98ccfqVSpElmyZCFTpkyUK1eO2bNnJ+RhiEha5+pquujt3Qu//gp168KdO2ZOqPLlwdcXli41k+xKmnDjxg1+//13pkyZwoQJE2ItIpK4osc3JVo3vQED4Pp1qFLF/MAlksIkaIzT7du3cXFxAWDFihUxA26LFy/O2bNn43WsefPm0bNnT6ZMmULVqlUZP348fn5+HDhwgFy5ct2zf7Zs2Rg4cCDFixfH2dmZX3/9FX9/f3LlyoWfXyI1E4tI2uLgYKoyvfACbNsGY8bADz9AUJBZSpUypcxfe82UNpdUaceOHTz//POEhYVx48YNsmXLRnBwMG5ubuTKlYuuXbtaHaJImmG32/8d31QkERKnPXvg66/N+sSJ5n1bJIVJ0KvyqaeeYsqUKaxdu5bAwEAaNDAXzJkzZ8iePXu8jjV27Fg6dOiAv78/JUuWZMqUKbi5uTFjxoz77l+3bl2aNGlCiRIl8PHxoVu3bpQpU4Z169Yl5KGISHpTsSJ89x0cOWKSJXd384H91ltQsCB8+ilcvmx1lJIAPXr0oFGjRly5coWMGTOyadMmTpw4QcWKFRk9erTV4YmkKX9f/JvToadxdXKldoHaj3/AESPMv82amRYnkRQoQYnTiBEjmDp1KnXr1qVVq1aULVsWMGVhq8TjxR4REcG2bdvw9fX9NyAHB3x9fdm4ceMj72+32wkKCuLAgQPUrn3/izY8PJzQ0NBYi4gIBQqYlqdTp2DkSMiXz8xQP3CgKSTxzjumi5+kGjt37qRXr144ODjg6OhIeHg43t7ejBw5kvfff9/q8ETSlOhuenUL1iVjhoyPd7ATJ+D77816//6PGZlI0klQ4lS3bl2Cg4MJDg6O1TL09ttvM2XKlDgfJzg4mMjISHLnzh1re+7cuTl37twD7xcSEoK7uzvOzs688MILTJw4kfr1699334CAADw9PWMWb2/vOMcnIulAlizQpw8cPQqzZ0O5chAWBlOnmi58zz4LixaZsVGSomXIkAGH/3fvyZUrFydPngTA09OTU6dOWRmaSJqTqGXIx4yByEgz7rRixcc/nkgSSVDidPPmTcLDw8maNSsAJ06cYPz48Q8cl5TYMmfOzM6dO9myZQuffPIJPXv2ZPXq1ffdd8CAAYSEhMQs+vAUkftydobXXzfzP61aBa+8YvrYR6/7+JiuJJcuWR2pPED58uXZsmULAHXq1GHQoEF8++23dO/enVKlSlkcnUjacSPiBn+c+ANIhPFNFy/C9OlmXa1NksIlKHF6+eWX+eabbwC4evUqVatWZcyYMTRu3JgvvvgizsfJkSMHjo6OnD9/Ptb28+fP4+Xl9eCgHRwoUqQI5cqVo1evXjRr1oyAgID77uvi4oKHh0esRUTkgWw2U31v4UI4dsxUecqe3cwH1b8/5M8P7drBjh1WRyr/8emnn5InTx4APvnkE7JmzUqnTp24ePEiU6dOtTg6kbRj9fHVRERGUDBLQYplL/Z4B5s4EW7eNC1Nzz6bOAGKJJEEJU7bt2+nVq1aACxYsIDcuXNz4sQJvvnmm3iVfHV2dqZixYoEBQXFbIuKiiIoKIhq1arF+ThRUVGEh4fH/QGIiMTFE0+YYhGnT8PMmVChAty6BTNmmPWaNWHePLh92+pIBahUqRLPPPMMYLrqLV26lNDQULZt20a5cuWsDU4kDYke39SwSENsNlvCD3T9OkyaZNb79zc/XImkYAlKnMLCwsicOTMAy5cv55VXXsHBwYGnn36aEydOxOtYPXv2ZNq0aXz99dfs27ePTp06cePGDfz9/QFo06YNAwYMiNk/ICCAwMBAjh49yr59+xgzZgyzZ8/m9ddfT8hDERF5NFdXaNsWtm6F9euhVStwcjLrr75qCk0MGwb/aT2X5HXs2DEOHTp0z/ZDhw5x/Pjx5A9IJI1KtDLk06bBlStQtCg0aZIIkYkkrQQlTkWKFOGnn37i1KlTLFu2jOeeew6ACxcuxLsrXMuWLRk9ejSDBg2iXLly7Ny5k6VLl8YUjDh58mSsuaFu3LjBu+++y1NPPUWNGjVYuHAhc+bMoX379gl5KCIicWezQfXqppz5yZMweDB4eZnZ7gcPNtX4Xn8d/vzT6kjTpbZt27Jhw4Z7tv/555+0bds2+QMSSYMOXz7M4cuHyeCQgWcKPpPwA0VEmKIQAH37gqNj4gQokoRsdrvdHt87LViwgNdee43IyEieffZZAgMDAdMa9Mcff7BkyZJEDzSxhIaG4unpSUhIiMY7icjji4iABQtMP/1Nm/7dXrkydO4MLVtqUt27JOV7sIeHB9u3b6dIkSKxth8+fJhKlSpx9erVRD1fYtJnk6QWkzZPosuSLjxT8BlWvrky4QeaNQv8/SFPHjOeVO+TYpH4vP8mqMWpWbNmnDx5kq1bt7Js2bKY7fXq1WPcuHEJOaSISOrk7AyvvQYbN8KWLfDmm2Zb9Lq3N3zwgRknJUnKZrNx7dq1e7aHhIQQGRlpQUQiaU/0+KbH6qYXFfXvhLc9eihpklQjQS1Odzv9/y8D+fPnT5SAkpp+1RORJHfxoum7/8UX/yZMjo6mrHmXLqaoRDodBJ2U78GNGjUiY8aMfP/99zj+v9tPZGQkLVu25MaNG+oNIfKYbt25RfaR2Qm7Hcaud3ZRJneZhB3op5/MmCZPT9PtWa95sVCStzhFRUUxbNgwPD09KVCgAAUKFCBLlix89NFHREVFJShoEZE0I2dOeP990/1k/nyoXdtM7hi9Xr68mbckLMzqSNOUESNGsHLlSp588kn8/f3x9/fnySef5I8//mDUqFFWhyeS6q07uY6w22Hkcc9D6VylE3YQux2GDzfr776rpElSlQQlTgMHDmTSpEkMHz6cHTt2sGPHDj799FMmTpzIhx9+mNgxioikTk5O0KwZrFkDu3ZBhw6QMeO/6/nzm0HRqviWKEqWLMnu3btp0aIFFy5c4Nq1a7Rp04b9+/drAlyRRHB3N70ElyH/4w9TQMfFBbp1S8ToRJJegrrq5c2blylTpvDSSy/F2v7zzz/z7rvv8s8//yRagIlN3SFExFKXL5t5oCZP/jdhstmgUSPTja9evTTdjU/vwfen50VSg1Kfl2Lvxb3MazaPFk+1SNhBGjaEpUuhUyf4/PPEDVAkAZK8q97ly5cpXrz4PduLFy/O5cuXE3JIEZH0IVs26N0bDh+GxYuhfn3TdSV6/amnzJeJ69etjjRVWrt2La+//jrVq1eP+RFv9uzZrFu3Ll7HmTx5MgULFsTV1ZWqVauyefPmB+47bdo0atWqRdasWcmaNSu+vr4P3V8kNToVcoq9F/fiYHPAt7Bvwg6yc6dJmhwczPugSCqToMSpbNmyTIqe6fkukyZNokyZBA4UFBFJTxwdTSvT8uWwbx+89x64u/+7ni8fdO8O95nQVe5v4cKF+Pn5kTFjRrZv3054eDhgqup9+umncT7OvHnz6NmzJ4MHD2b79u2ULVsWPz8/Lly4cN/9V69eTatWrVi1ahUbN27E29ub5557LkX3vhCJr2VHTBXlqvmqki1jtoQdZORI82+LFlC4cCJFJpJ8EtRVb82aNbzwwgs88cQTVKtWDYCNGzdy6tQpfv/9d2rVqpXogSYWdYcQkRQrNNTMbTJpUuyEydcX3njDVKHKnNmy8BJDUr4Hly9fnh49etCmTRsyZ87Mrl27KFy4MDt27KBhw4acO3cuTsepWrUqlStXjvmBMCoqCm9vb7p06UL//v0fef/IyEiyZs3KpEmTaNOmTZzOqc8mSema/dCMhfsWMrTuUAbVGRT/Axw9CkWLmlLkO3ZAuXKJHqNIQiR5V706depw8OBBmjRpwtWrV7l69SqvvPIKe/fuZfbs2QkKWkQk3fPwgK5dYf9+053lhRfMeKcVK8ycULlzQ6tW8NtvcPu21dGmOAcOHKB27dr3bPf09Izz5LcRERFs27YNX99/uyI5ODjg6+vLxo0b43SMsLAwbt++TbZsD/5VPjw8nNDQ0FiLSEp1O/I2gUcDgceYv2n0aJM0NWigpElSrQQlTmAKRHzyyScsXLiQhQsX8vHHH3PlyhW++uqrxIxPRCT9cXAAPz/49VczFmrYMChWDG7ehLlz4cUXIW9e6NzZTLz7eNPxpRleXl4cPnz4nu3r1q2jcBy7BQUHBxMZGUnu3Lljbc+dO3ecW6z69etH3rx5YyVf/xUQEICnp2fM4u3tHadji1jhz3/+JDQ8lOwZs1MxT8X4H+D8eVMUB6Bfv8QNTiQZJThxEhGRZFC4MHz4oWmF2rzZlO/NlQuCg01lvurVTfeXwYPh4EGro7VUhw4d6NatG3/++Sc2m40zZ87w7bff0qtXLzp16pQsMQwfPpy5c+eyaNEiXF1dH7jfgAEDCAkJiVlOnTqVLPGJJER0GfLnfJ7D0cEx/geYMAHCw6FqVahTJ5GjE0k+TlYHICIicWCzQeXKZhk92nTf+/ZbWLQIjhwxrVLDhpm/t24Nr75quvalI/379ycqKop69eoRFhZG7dq1cXFxoU+fPrRv3z5Ox8iRIweOjo6cP38+1vbz58/j5eX10PuOHj2a4cOHs2LFikcWSnJxccHFxSVOMYlY7e75m+ItNNT8yAPQv3+anm5B0j61OImIpDZOTmacwOzZpgvMt9+auVEcHWHLFlONL18+s8+cOemmtLnNZmPgwIFcvnyZPXv2sGnTJi5evIinpyeFChWK0zGcnZ2pWLEiQUFBMduioqIICgqKKYZ0PyNHjuSjjz5i6dKlVKpU6bEfi0hKceHGBbad3QaAn49f/A8wdSqEhEDx4vCf+T9FUpt4tTi98sorD/17XAffiohIIsmUCV57zSwXLsC8eSZZ2rwZli0zi5sbNG5sWqLq14cMGayOOlGFh4czZMgQAgMDY1qYGjduzMyZM2nSpAmOjo706NEjzsfr2bMnb775JpUqVaJKlSqMHz+eGzdu4O/vD0CbNm3Ily8fAQEBAIwYMYJBgwbx3XffUbBgwZixUO7u7ri7uyf+AxZJRsuPLAegQp4K5HaPZyt2eDiMG2fW+/Y14zdFUrF4JU6enp6P/HtcS6+KiEgiy5ULunQxy6FDpiXq229NgYnvvjNLzpymG1/r1lClSproNjNo0CCmTp2Kr68vGzZsoHnz5vj7+7Np0ybGjBlD8+bNcXSM+7iMli1bcvHiRQYNGsS5c+coV64cS5cujSkYcfLkSRzu+gL4xRdfEBERQbNmzWIdZ/DgwQwZMiRRHqOIVWK66fkkoJve7Nlw9qxpAW/dOpEjE0l+CZrHKTXTXBkikq7Y7ab16dtvTUW+ixf//VuRIubLTOvWpsBEMkiK9+DChQszfvx4XnrpJfbs2UOZMmVo27YtX331FbZUkhjqs0lSoih7FLlH5yY4LJg/2v5BrQLxmKczMhJKlDA/4owdC/Fo9RVJTkk+j5OIiKQSNpupZDVhAvzzD/z+u+nWlzGjaYkaOtSUOq9aFSZONN39UpnTp09TsaIpkVyqVClcXFzo0aNHqkmaRFKq7We3ExwWjIeLB0/nfzp+d160yCRNWbNChw5JE6BIMlPiJCKSXmTIYIpIfPutSZBmzzbzRTk4mFaprl3N/FDPP2+69d24YXXEcRIZGYmzs3PMbScnJ40tEkkESw4tAcC3sC8ZHOMxNtJuhxEjzHrnzqDrUdIIlSMXEUmP3N3h9dfNcu6cKSrx7bemKt+SJWbJlAmaNDFd+Xx9TTW/FMhut9O2bduY8t63bt3inXfeIVOmTLH2+/HHH60ITyTVWnokgeObVq6ErVtNy3aXLkkQmYg1UuanoIiIJB8vLzOxbrducODAv0Uljh41FfrmzDGFJ1591SRalSqlqKISb775Zqzbr7/+ukWRiKQdV25eYdPpTQD4FYlnGfLhw82/7dubgjQiaYSKQ4iIyL3sdti06d+iEpcu/fu3Zs1g/vwEHVbvwfen50VSmvl759NiQQtK5izJ3nf3xv2O27aZH1ccHc04yoIFkyxGkcSg4hAiIvJ4bDaoVg0mTTLlhH/5xbQ4ZcwINWtaHZ2IJLEElyGPHtvUqpWSJklz1FVPREQeLkMGePFFs1y7ZnU0IpLE7Hb7v+ObisQjcTp0CBYsMOt9+yZBZCLWUuIkIiJxlzmz1RGISBLbc2EPZ66dIaNTxvjN3TRqlOnm++KLULp00gUoYhF11RMRERGRGNHd9J4p9AyuTq5xu9PZs/D112a9X78kikzEWkqcRERERCRGgsqQjx8PERFQo4bGQUqapcRJRERERAC4HnGdtSfWAvEY33T1KnzxhVnv3z9pAhNJAZQ4iYiIiAgAq46t4nbUbQpnLUyRbEXidqcvvjCFY556Cp5/PmkDFLGQEicRERERAWKXIbfFZaLrmzfhs8/Mer9+4KCvlpJ26dUtIiIiItjtdpYcXgJAw6IN43anr7+G8+fhiSfMXG8iaZgSJxERERHh8OXDHLt6DGdHZ+oWrPvoO9y5Y0qQA/TubeZ8E0nDlDiJiIiISEw3vVpP1MLd2f3Rd1iwAI4ehezZ4a23kjg6EespcRIRERGRf8uQx6Want0OI0aY9a5dIVOmJIxMJGVQ4iQiIiKSzt26c4tVx1YBcUycli+HnTtNwvTee0kbnEgKocRJREREJJ1be2ItN+/cJF/mfDyV86lH32H4cPNvhw6mq55IOqDESURERCSdiylDXiQOZcj//BNWrwYnJ+jZM+mDE0khlDiJiIiIpHPxGt8UPbbp9dfB2zsJoxJJWZQ4iYiIiKRjJ0NO8vfFv3G0OeJb2PfhO+/bB4sWmfW+fZM+OJEURImTiIiISDoW3U3v6fxPk8U1y8N3jp636eWXoUSJpA1MJIVR4iQiIiKSjt09vumhTp+GOXPMev/+SRyVSMqjxElEREQknbodeZsVR1cAcUicxo2D27ehTh14+ulkiE4kZVHiJCIiIpJObTy9kWsR18jhloMKeSo8eMfLl2HqVLPer1/yBCeSwihxEhEREUmnorvp+fn44WB7yNfCyZPhxg0oUwYaxKHynkgapMRJREREJJ2K0/imsDCYMMGs9+8Pj5rnSSSNUuIkIiIikg6du36OHed2APCcz3MP3nHGDAgOhkKFoHnzZIpOJOVxsjoAERERkfTiQPABzt84b3UYAKw6tgqAinkqkitTrvvvdPs2jB5t1nv3Bid9dZT0S69+ERERkWSw6tgq6n1TDzt2q0OJpWGRhg/+4w8/wIkTkDMn+PsnX1AiKZASJxEREZEkFhkVSfdl3bFjJ497HjxcPKwOCYBsGbPRvkL7+//Rbofhw8169+6QMWOyxSWSEilxEhEREUlis3bOYvf53WRxzcJfnf4iu1t2q0N6tN9/hz17wN0dOnWyOhoRy6k4hIiIiEgSuhZ+jQ9WfQDAoNqDUkfSBP+2Nr3zDmTNam0sIimAEicRERGRJDRy/UjOXT9HkWxFeK/Ke1aHEzfr18O6deDsDD16WB2NSIqgxElEREQkiZwMOcnojaYq3UjfkTg7OlscURyNGGH+bdMG8ua1NhaRFEKJk4iIiEgSeT/ofW7duUWdAnVoXLyx1eHEzZ498MsvZqLbPn2sjkYkxVDiJCIiIpIENv+zmW//+hYbNsb6jcVms1kdUtyMGmX+feUVKFbM2lhEUhAlTiIiIiKJzG6303NZTwDalG1DhTwVLI4ojk6cgO++M+v9+lkbi0gKo8RJREREJJEt3LeQ9afWk9EpI588+4nV4cTd2LFw5w7UqweVK1sdjUiKosRJREREJBGF3wmnb2BfAPrW6Es+j3wWRxRHwcEwbZpZV2uTyD2UOImIiIgkoombJ3Ls6jHyuOehT/VUVFxh0iS4eRMqVABfX6ujEUlxlDiJiIiIJJKLNy7y0R8fAfBpvU/J5JzJ4oji6Pp1mDjRrPfvbyrqiUgsSpxEREREEsnQNUMJDQ+lvFd52pRtY3U4cTd9Oly+DEWKmGp6InIPJU4iIiIiieDvi38zZesUAMY8NwYHWyr5mhURAWPGmPU+fcDR0dp4RFKoVHJFi4iIiKRsfQL7EGmP5OUnX+aZQs9YHU7cff89nD4NXl7QJhW1kokkMyVOIiIiIo9p+ZHl/H7od5wcnBhVf5TV4cRdVBSMGGHWe/QAV1dr4xFJwZQ4iYiIiDyGyKhIei3vBUDnyp0pmr2oxRHFwy+/wL594OEBHTtaHY1IiqbESUREROQxzNgxgz0X9pDVNSsf1vnQ6nDizm6H4cPN+rvvgqentfGIpHBKnEREREQS6Fr4NT5Y9QEAg+sMJlvGbBZHFA9r18KmTeDiAt26WR2NSIqnxElEREQkgYavG86FGxcomq0onSp3sjqc+ImupNe2rSkMISIPlSISp8mTJ1OwYEFcXV2pWrUqmzdvfuC+06ZNo1atWmTNmpWsWbPi6+v70P1FREREksKJqycYs9EkH6Pqj8LZ0dniiOLhzBn47TezrtYmkTixPHGaN28ePXv2ZPDgwWzfvp2yZcvi5+fHhQsX7rv/6tWradWqFatWrWLjxo14e3vz3HPP8c8//yRz5CIiIpKevb/yfcIjw6lbsC4vPfmS1eHEz9dfQ2Qk1KgBJUpYHY1IqmCz2+12KwOoWrUqlStXZtKkSQBERUXh7e1Nly5d6N+//yPvHxkZSdasWZk0aRJt4jD3QGhoKJ6enoSEhODh4fHY8YuISNzpPfj+9LykPn+e/pOnv3oaGza2vb2N8nnKWx1S3EVFQbFicOQIzJxpuuqJpFPxef+1tMUpIiKCbdu24evrG7PNwcEBX19fNm7cGKdjhIWFcfv2bbJlu/9gzPDwcEJDQ2MtIiIiIgllt9vpsawHAG3LtU1dSRPAmjUmacqcGZo3tzoakVTD0sQpODiYyMhIcufOHWt77ty5OXfuXJyO0a9fP/LmzRsr+bpbQEAAnp6eMYu3t/djxy0iIiLp1/y/57Px9EbcMrjx8bMfWx1O/H31lfm3VSvIlMnaWERSEcvHOD2O4cOHM3fuXBYtWoTrA2a6HjBgACEhITHLqVOnkjlKERERSStu3blFvxX9AOhXox95M+e1OKJ4unIFFiww6+3bWxuLSCrjZOXJc+TIgaOjI+fPn4+1/fz583g9oizm6NGjGT58OCtWrKBMmTIP3M/FxQUXF5dEiVdERETStwl/TuD41ePky5yPXtV6WR1O/H37LYSHQ5kyUKmS1dGIpCqWtjg5OztTsWJFgoKCYrZFRUURFBREtWrVHni/kSNH8tFHH7F06VIq6aIXERGRZHDhxgU+WfsJAJ/W+5RMzqmsm5vdDtOmmfX27cFmszYekVTG0hYngJ49e/Lmm29SqVIlqlSpwvjx47lx4wb+/v4AtGnThnz58hEQEADAiBEjGDRoEN999x0FCxaMGQvl7u6Ou7u7ZY9DRERE0rYhq4cQGh5KhTwVeL3M61aHE3/btsHu3eDiAq1bWx2NSKpjeeLUsmVLLl68yKBBgzh37hzlypVj6dKlMQUjTp48iYPDvw1jX3zxBRERETRr1izWcQYPHsyQIUOSM3QRERFJJ/Ze2MvUbVMBGPvcWBxsqXCYeHRRiFdegQdUIxaRB7N8HqfkprkyRESso/fg+9PzkvI9/+3zLDm8hCbFm/Bjyx+tDif+btyAvHkhNBSCguDZZ62OSCRFSDXzOImIiIikdMsOL2PJ4SVkcMjAyPojrQ4nYRYsMElT4cJQt67V0YikSkqcRERERB7gTtQdei031fO6VOlCkWxFLI4ogaZPN/+2awcO+vonkhC6ckREREQe4KvtX7H34l6yZczGB7U/sDqchDlwANatMwnTm29aHY1IqqXESUREROQ+QsND+XDVhwAMqTOErBmzWhxRAkUXhXj+eciXz9pYRFIxJU4iIiIi9xGwNoCLYRcplr0Y71R6x+pwEiYiAr7+2qy3b29tLCKpnBInERERkf84fvU44zaNA2B0/dFkcMxgcUQJ9OuvcOECeHmZFicRSTAlTiIiIiL/MSBoAOGR4Txb6FleLPai1eEkXHQ3vTffhAypNPkTSSGUOImIiIjcZeOpjczdMxcbNsY8NwabzWZ1SAlz6hQsXWrW27WzNhaRNECJk4iIiMj/2e12ei7vCYB/OX/KeZWzNqDHMWsWREVBnTpQtKjV0YikekqcRERE/mPy5MkULFgQV1dXqlatyubNmx+47969e2natCkFCxbEZrMxfvz45AtUEt28vfPYdHoTmTJk4uNnP7Y6nISLioIZM8y6ikKIJAolTiIiIneZN28ePXv2ZPDgwWzfvp2yZcvi5+fHhQsX7rt/WFgYhQsXZvjw4Xh5eSVztJKYbt6+Sb8V/QDoX7M/eTLnsTiix7ByJRw/Dp6e0LSp1dGIpAlKnERERO4yduxYOnTogL+/PyVLlmTKlCm4ubkxI/rX+/+oXLkyo0aN4tVXX8XFxSWZo5XE9Nmfn3Ey5CT5PfLTs1pPq8N5PNOnm39bt4aMGa2NRSSNUOIkIiLyfxEREWzbtg1fX9+YbQ4ODvj6+rJx48ZEO094eDihoaGxFrHW+evn+XTtpwAE1AvALYObxRE9huBgWLTIrKubnkiiUeIkIiLyf8HBwURGRpI7d+5Y23Pnzs25c+cS7TwBAQF4enrGLN7e3ol2bEmYwasHcy3iGpXyVuK10q9ZHc7jmTPHTHxboQKUL291NCJphhInERGRZDZgwABCQkJillOnTlkdUrq258Iepm2fBsDY58biYEvFX4/s9n/nblJrk0iicrI6ABERkZQiR44cODo6cv78+Vjbz58/n6iFH1xcXDQeKgXpvbw3UfYompZoSq0CtawO5/Fs3gx79oCrK7RqZXU0ImlKKv5JRUREJHE5OztTsWJFgoKCYrZFRUURFBREtWrVLIxMksrSw0tZdmQZGRwyMMJ3hNXhPL7oohDNm0OWLJaGIpLWqMVJRETkLj179uTNN9+kUqVKVKlShfHjx3Pjxg38/f0BaNOmDfny5SMgIAAwBSX+/vvvmPV//vmHnTt34u7uTpEiRSx7HPJod6Lu0Gt5LwC6Vu2KTzYfiyN6TNeuwfffm3V10xNJdEqcRERE7tKyZUsuXrzIoEGDOHfuHOXKlWPp0qUxBSNOnjyJg8O/HTbOnDlD+bsG4I8ePZrRo0dTp04dVq9endzhSzxM2zaNvy/+TfaM2fmg9gdWh/P4fvgBbtyAokWhVirvciiSAilxEhER+Y/OnTvTuXPn+/7tv8lQwYIFsdvtyRCVJKaQWyEMWj0IgKF1h5LFNYu1ASWG6KIQ7dqBzWZtLCJpkMY4iYiISLrz6dpPCQ4LpniO4rxd8W2rw3l8e/fCxo3g6Ahvvml1NCJpkhInERERSVeOXTnG+D/HAzC6/mgyOGawNqDEEN3a1KgRJGIFSBH5lxInERERSVf6B/UnIjIC38K+PF/0eavDeXzh4fDNN2ZdRSFEkowSJxEREUk3NpzawA97f8CGjTHPjcGWFsYCLV4Mly5B3rzg52d1NCJplhInERERSRei7FH0WNYDgHbl21EmdxmLI0ok0XM3+fuDk+p+iSQVJU4iIiKSLszbM4/N/2wmU4ZMfPTsR1aHkziOH4fAQLP+1luWhiKS1ilxEhERkTTv5u2b9FvRD4ABNQfg5Z5GCijMnAl2O9SrB4ULWx2NSJqmxElERETSvHGbxnEq9BTeHt70rNbT6nASR2SkSZzAzN0kIklKiZOIiIikaeeunyNgXQAAw32HkzFDRosjSiSBgXDqFGTNCk2aWB2NSJqnxCmeIqMirQ5BRERE4mHQqkFcj7hOlXxVeLXUq1aHk3iii0K88Qa4ulobi0g6oNIr8RSwLoBlR5YxsNZA/Hz80kYZUxERkUR0J+oOd6LuWB0GAH9f/JuvdpjJYcc+NxYHWxr5zfjCBfj5Z7OubnoiyUKJUzzcibrDF1u/4My1MzT8tiEV8lTg/Zrv06REk7TzRiwiIvIYlh1exstzXyY8MtzqUGJpXrI5NZ6oYXUYiWf2bLhzB6pUgTJppKy6SAqnb/vx4OTgxOb2m+n5dE/cMrix/ex2ms1vRqnPSzF712xuR962OkQRERHLRERG0HlJ5xSXNOVwy8EI3xFWh5F47PZ/u+mptUkk2djsdrvd6iCSU2hoKJ6enoSEhODh4ZHg4wSHBfPZps+YuHkiIeEhABTMUpB+NfrRtlxbXJ3U11hE5L8S6z04rUkrz8v4TePpsawHuTLlYmfHnWRyzmR1SAC4ZXDDySENdbJZvx5q1gQ3Nzh7FlLxa0bEavF5/1Xi9LjHCw/l8y2fM3bjWC6GXQQgj3seelXrRcdKHXF3dn/sc4iIpBVpJUFIbGnhebl88zJFJhThyq0rfPnil3So2MHqkNIuf3+YNcv8O2OG1dGIpGrxef9VV73H5OHiQf+a/Tne/TgTGkwgv0d+zl4/S+/A3hQYX4Bha4Zx5eYVq8MUERFJUsPWDOPKrSuUzlWat8q/ZXU4aVdoKPzwg1lv397aWETSGSVOicQtgxtdqnbhSNcjTG80nSLZinD55mUGrx5MgfEF6L+iP+evn7c6TBERkUR38NJBJm+ZDMCY58bg6OBocURp2Ny5EBYGxYtDtWpWRyOSrihxSmTOjs60q9CO/e/t5/um31M6V2muRVxjxPoRFPysIF1+78LJkJNWhykiIpJo+gb25U7UHZ4v+jz1fepbHU7aFl0Uon170JQoIslKiVMScXRw5NVSr7LznZ38/OrPVMlXhVt3bjFpyyR8JvjQ7ud2HLx00OowRUREHsuqY6v4+cDPONocGVV/lNXhpG27dsGWLZAhg5n0VkSSlRKnJOZgc+ClJ19iU7tNrHhjBc8UfIY7UXeYsXMGJSaX4NUFr7L7/G6rw5R0xG63E3Q0iC+3fcmtO7esDkdEUrHIqEh6Lu8JQMeKHSmZs6TFEaVxX5mJfHn5ZciVy9pYRNIhJU7JxGazUa9wPVa+uZINb23gxWIvEmWPYt7eeZSdUpZG3zdi0+lNVocpadyqY6uoPas2vrN96fhrR8p8UYaVx1ZaHZaIpFLf7PqGned24uniyZC6Q6wOJ227dQvmzDHrmrtJxBJKnCxQzbsav7T6hZ0dd9LiqRbYsPHrwV+p9lU16n1Tj6CjQaSzKvGSxNafXM+zXz/Ls988y7qT63B1ciVXplwcunyIet/Uo+1PbQkOC7Y6TBFJRa5HXGfgyoEAfFD7A3JmymlxRGncokVw5Qp4e0N9jSMTsYISJwuV9SrLvGbz2N95P/7l/HFycGLlsZX4zval2lfV+OXAL0qg5LH8efpP/Ob4UXNmTVYdX4WzozOdK3fmSNcjHOx8kPcqv4cNG1/v+prik4oza+csveZEJE5GrR/F2etnKZSlEF2qdLE6nLQvuijEW2+Bo6oWilhBE+CmICeunmD0htFM3zE9ZuxJmdxlGFBzAM1LNld5V4mz7We3M2jVIH479BsATg5OtCvfjvdrvc8Tnk/E2nfT6U10/LVjzFi7ugXrMuWFKTyZ48lkj1vSvpT8Hmyl1Pa8nA49TbGJxbh55ybzm8+nWclmVoeUth05AkWKmCp6x45BgQJWRySSZmgC3FSqQJYCTHx+Ise7HadfjX64O7uz+/xuWi1sRfHJxflq+1dEREZYHaakYLvP76bJvCZU/LIivx36DUebI2+Ve4uDnQ8y5cUp9yRNAE/nf5qtHbYy0nckGZ0ysvr4aspMKcOwNcMIvxNuwaMQkZRu4MqB3LxzkxreNWhaoqnV4aR9M2eaf+vXV9IkYiG1OKVgV25eYeLmiXz252dcvnkZAG8Pb/pU70P7Cu3JmCGjxRFKSvH3xb8ZsnoI8/+eD5hqjq1Lt+bD2h9SNHvROB/n2JVjvPv7uyw9vBSAJ7M/ydQXp1KnYJ0kiVvSn9T0HpycUtPzsvXMVipPqwzAn+3/pEq+KhZHlMbduWOSpTNn4IcfoHlzqyMSSVPU4pRGZM2YlUF1BnGi+wlG1x+Nl7sXp0JP0XVpVwp+VpDh64YTGh5qdZhioYOXDvL6j69T6vNSMUlTy6dasqfTHr5p8k28kiaAQlkL8ftrvzO36VxyZ8rNgUsHqPt1Xdr93I5LYZeS4BGISGpit9vptbwXAK1Lt1bSlByWLjVJU44c8NJLVkcjkq6pxSkVuXXnFjN3zGTE+hGcCDkBQBbXLPj5+JHTLSc53HLEWrK7ZY9Zd3VytTh6SUxHrxzloz8+4ptd3xBljwLglRKvMKTOEErnLp0o57h66yr9V/Rn6rapAORwy8E4v3G0Lt0am2arlwRKze/BSSm1PC+L9i3ilR9ewdXJlQOdD9y3+68kssaN4eefoWdPGDPG6mhE0pz4vP8qcUqFbkfe5vs93xOwLoD9wfvjdJ9MGTLdm1BlfHCilT1jdlycXJL4kUh8nbh6go//+JhZu2ZxJ+oOAI2KNWJo3aGUz1M+Sc654dQG3v7lbfZe3AtAvUL1+OKFL+LdmiWpX2RUJLN2zqJtubYJLlaTFt6Dk0JqeF4iIiMoObkkR64cYWCtgXz87MdWh5T2nTsH+fNDZCTs3QslNcGwSGJT4vQQqeHDKa4ioyJZdmQZB4IPcOnmJYLDgu9ZLt28FPMFO74yO2eOlUzdnWzds/3/yVYGxwyJ/CgF4J/Qf/hk7SdM3z6d21G3AWhQpAFD6w5Nlq4yEZERjNkwhmF/DOPWnVu4OLrwYe0P6VOjD86Ozkl+fkkZBqwYwPD1w3mlxCssbLEwQcdIS+/BiSk1PC9jN46l1/JeeLl7cajLIdyd3a0OKe0bMQL694dq1WDDBqujEUmTlDg9RGr4cEpMdrud0PDQe5KpByVZwWHBXAq7RKQ9MkHn83TxJE/mPOTNnJe8mfOSx/3f9bu3qbBF3Jy7fo6AtQFM3TaV8EhT4a5eoXoMrTuUGk/USPZ4jlw+QqffOhF4NBCAEjlK8GWjL6n5RM1kj0WSV3QXLYDvm37Pq6VeTdBx0tt7cFyl9OclOCyYIhOKEBIewvRG02lXoZ3VIaV9djsUKwaHD8NXX5n5m0Qk0SlxeoiU/uGUEkTZowi5FRKnJCt6uXzzcsxYm7jI4poldkLlnjdWwhWdYKXX7oIXb1xk5PqRTN4ymZt3bgJQ64lafPTMR5ZXuLPb7Xy/53u6L+3OxbCLAHSo0IERviPImjGrpbFJ0jgQfIDK0ypzLeIa3at2Z1yDcQk+lt6D7y+lPy9dl3Rl4uaJlMldhu1vb9e8gslhzRqoWxfc3eHsWfOviCQ6JU4PkdI/nFKrKHsUV29d5cKNC5y9dpYz185w5toZzl7/dz16iU4E4iJbxmz3JFh5M8dOsrzcvdJMd7FLYZcYs3EME/6cwI3bNwAzz9JHz3xEvUL1UlRRhss3L9MvsB/Td5jZ7HNlysV4v/G8WurVFBWnPJ7rEdepOr0qf1/8m9oFarPijRWP1SVX78H3l5Kfl/3B+yn1eSki7ZEEvhGIb2Ffq0NKH9q0gdmzoUMH+PJLq6MRSbOUOD1ESv5wSg+iuw7+N5m6X4IV3TUtLnK65fw3mXKP3TXQ29Mbbw9vcrjlSLFf6K/eusq4jeMYt2kc1yKuAVApbyWG1R1GgyINUmzcAGtPrKXjrx3ZF7wPAD8fPz5/4XMKZy1scWTyuOx2O68ufJUf9v5A3sx52fb2NrzcvR7rmHoPvr+U/Ly89P1L/HLwF14s9iK/tPrF6nDSh6tXIU8euHULNm2CqlWtjkgkzYrP+69TMsUkAoDNZsPT1RNPV09K5CzxwP3sdjtXbl2J1Xr1oCTrdtRtLoZd5GLYRXaf3/3AY7o4upDfI39MIpXfIz/eHt54e/67ni1jtmRNUkLDQ5nw5wTGbBzD1VtXASibuyzDnhlGo2KNUnTCFK1WgVrs6LiDURtG8fEfH7PsyDKe+vwpBtcZTK9qvVQwJBUbt2kcP+z9AScHJ+Y3n//YSZOkPkFHg/jl4C842hwZVX+U1eGkH999Z5KmUqWgiubKEkkp1OIkqZrdbufSzUv3TbDOXD/DP6H/cCr0FOeun4vT8TI6ZYyVSMUkWP9Ptrw9vfF08XzshOZGxA0mbZ7EyA0juXzzMgBP5XyKoXWH0qREExxsqXNu6oOXDtLpt06sPLYSgFK5SvHli19SzbuaxZFJfK0+vhrfb3yJtEcyseFEOlfpnCjH1Xvw/aXE5yUyKpKKX1Zk1/lddK7cmYnPT7Q6pPSjQgXYsQPGj4du3ayORiRNU1e9h0iJH06S9CIiI/gn9B9Oh57mVOgpToWc+nf9/7ejCx08SqYMme7banV3kuXhcv/XVtjtMKZsncLwdcNjzlcsezGG1BlCi6dapIkB13a7ndm7Z9NzWU8u3byEDRsdK3YkwDeALK5ZrA5P4uB06GkqflmRCzcu8HqZ1/mm8TeJ1vqp9+D7S4nPy4wdM2i3uB2eLp4c7nqYHG45rA4pfdi+HSpWBGdnOHMGsme3OiKRNE1d9UT+w9nRmUJZC1Eoa6EH7nPrzq2YFqrToac5FXIqJrGKvn3p5iVu3L7B/uD9D5182MPF455WK0cHRyZvmRzT+lU4a2EG1xnMa6Vfw8kh7VyKNpuNNmXb8HzR5+kT2IdZO2cxZdsUfjrwE581+IzmJZunii6I6VVEZATN5zfnwo0LlMldhqkvTtX/Vzp0PeI6A1cOBODD2h8qaUpOX31l/m3SREmTSAqjFieReAi7Hcbp0NOxEqu7W7FOhZ6KGav0IAU8C/Bh7Q9pU7ZNuhj/s/r4ajr+2pGDlw4C8HzR55n8/GQKZilobWByX+/99h6fb/2cLK5Z2NphKz7ZfBL1+HoPvr+U9rx8uPJDPl77MT5Zfdj77t50OzVEsgsLg7x5ISQEAgPBVxUMRZKaWpxEkohbBjeKZS9GsezFHrjP9YjrMYnV3UlV8M1g/Hz8eKv8W2mmfHpc1C1Yl93v7CZgXQAB6wL4/dDvlJxckqF1h9L96e7pInlMLb7Z9Q2fb/0cgDlN5iR60iSpw6mQU4zeOBqAkfVHKmlKTgsXmqSpYEF49lmroxGR/1DiJJLI3J3dKZ6jOMVzFLc6lBTDxcmFIXWH8GqpV3nn13dYc2INfVf05du/vmXqi1Opml+ldq2289xOOv7aEYDBdQbzQrEXLI5IrPL+yve5decWtZ6oRZPiTawOJ32ZbubFo107cEidRYJE0jJdlSKSbIrnKM6qN1cx46UZZMuYjV3nd1Htq2p0/r3zI7s4StK5fPMyr8x7hVt3bvF80ecZVGeQ1SGJRbb8s4U5u+cAMNZvrMa3JaeDB+GPP0zC1Lat1dGIyH0ocRKRZGWz2fAv78/+9/bzRpk3sGNn8pbJeI32oukPTVn490Ju3r5pdZjpRpQ9itd/fJ1jV49RKEshZjeZnWrL4cvjsdvt9FzeE4A3yrxBpbyVLI4onZkxw/zboAHkz29tLCJyX/p0FBFL5MyUk2+afMOKN1ZQKlcpwiPD+XHfjzSb34zco3Pz5k9vsvTwUu5E3bE61DRt2JphLDm8BFcnV35s+SPZMmazOiSxyI/7fmTdyXVkdMrIJ89+YnU46cvt2zBrlllv397SUETkwZQ4iYil6hWux+53drOj4w76Vu+Lt4c31yKu8c2ub2j4bUPyjsnLe7+9x7r/tXfnUVXV+//HnweUQUJzCETD0DRxRgURvaYlSZYVXhyjRPNmKZpGapCp9a2cUvNeMYxyqJ/zWKZZKpapVy9OKCoODQ5pDGaJYkJw9u8PbvTlqwgosA/yeqx1lrDP3me/zme59tu3+7P3PrMDq2E1O+4dZcOJDby57U0AYnvE4lPbx9xAYprM7EzGbhkLwOgOo/Gs5mlyogpmwwZISQE3N+jRw+w0IlIANU4iYjqLxYJPbR+mPjKVU6NOsX3Qdob5DuOeKveQdjWN9/e+T6cFnfCa5cXYzWM58PMBKtiTFErc9xe/55m1zwAQ7hfOs62eNTmRmCk6Ppoffv0Bj7s8GNtxrNlxKp4/n900cCBU1p1GRWyVnuMkIjYr25pN3A9xLD28lDVJa7icdTnvPe9a3vRv3p/+zfvTqGYjE1OWP1f/uErAvAAOpRwi4N4Avhn4TZndIl/H4Bszc1wuXL1Aw3815FLmJeY/OZ9BrQeV6f4rvHPnoF49sFrh2DFo3NjsRCIVSnGOv6afcZozZw5eXl44OTnh7+9PfHx8geseOXKEkJAQvLy8sFgszJo1q+yCikiZq2RXiaCGQSwMXkjK6BRW9V5FSJMQHO0dOXbhGBO/mcgD0Q/gG+vLjH/P4Kf0n8yObPMMw+CF9S9wKOUQbi5urOy9skI9V0yu98Y3b3Ap8xI+tX0Y0GqA2XEqnoULc5umTp3UNInYOFMbp+XLlxMREcHEiRPZv38/rVq1IigoiNTU1Buuf/XqVRo0aMCUKVOoXbt2GacVETM5V3YmpGkIq/qsInVMKh8Hf0zQ/UHYW+zZ9/M+Rm8eTb336tFlYRc+2PsBv1z9xezINmnOnjksOrQIe4s9K3qtoG7VumZHEhMlpSUxd+9cAGZ0m4G9nb3JiSoYq/WvaXq6KYSIzTN1qp6/vz9+fn5ER0cDYLVa8fT0ZMSIEURGRt50Wy8vL0aNGsWoUaOKtU9NExG5s6RmpLLq6CqWHl7KjjM78pZXsqtEt/u78XTzp3nK+ynucrjLxJS2YeeZnXT5uAvZ1mxmdJtBREBEmWfQMfjGzBqXHkt6sOHkBp5s/CSf9fuszPYr/7V1K3TtClWrws8/Q5UqZicSqXDKxVS9rKws9u3bR2Bg4F9h7OwIDAxk165dJbafzMxM0tPT871E5M7h5uLGML9hbB+0ndOjTjM1cCo+tX3ItmbzxckveGbtM7i960bfVX359NinZGZnmh3ZFMlXkum9sjfZ1mz6NuvLy+1fNjuSmGzz95vZcHIDlewqMS1wmtlxKqaPPsr98+mn1TSJlAOmNU4XLlwgJycHd3f3fMvd3d1JTk4usf1MnjyZatWq5b08PXWLVZE7Vb1q9RjbcSwHXjhAUngSEx6cQKMajfg9+3dWHFlBz+U9cZ/uzuDPBrPlhy3kWHPMjlwm/sj5gz4r+/DzlZ9pdk8zPnryIywWi9mxxEQ51hxe2fQKAMN8h9G4lq6tKXO//AKrV+f+rGl6IuWC6TeHKG1RUVFcunQp73X27FmzI4lIGfCu5c2bD73J8eHH2fv8XiLaR1DXtW7uncMS5vPI/3uEujPr8tLGl9h1dtcdfXvzsZvHsv3MdlwdXFnTd42mLQoLEhaQmJrI3U53M6HzBLPjVEyLF0NWFvj4QJs2ZqcRkSKoZNaOa9Wqhb29PSkpKfmWp6SklOiNHxwdHXF0dCyxzxOR8sVisdC2Tlva1mnLu93eZfvp7Sw9vJSVR1eSkpHC7PjZzI6fTf2769OveT/6N+9PC/cWZscuMUsTlzLrP7MA+KTnJzxQ8wFzA4npLmde5vWtrwMw4cEJ1KxS0+REFZBh/DVN7x//AJ0BFikXTDvj5ODgQNu2bYmLi8tbZrVaiYuLIyAgwKxYInIHs7PY0dmrM3N7zOXnV35mff/1hLYIxaWyCz/+9iOTd0ym5dyWtIxpyaqjq8r9WajDqYf5x+e5U4Ci/hZFsHewuYHEJkzdOZWUjBQa1mhIeLtws+NUTHv3QmIiODrmXt8kIuWCqVP1IiIi+PDDD/n4449JSkpi6NChZGRkMGhQ7sP3BgwYQFRUVN76WVlZJCQkkJCQQFZWFufOnSMhIYHvvvvOrK8gIuWUg70Djz/wOIv+vojUMaks77WcYO9gHOwdSExNpPfK3nRa0In4cwU/W86WXbp2ib8v/ztX/7hKYINA3nroLbMjiQ04c+kMM3bNAODdR97VM7zM8ufZpl69oHp1c7OISJGZNlUPoG/fvqSlpTFhwgSSk5Px8fHhyy+/zLthxJkzZ7Cz+6u3O3/+PK1bt877ffr06UyfPp3OnTvzzTfflHV8EblDVKlchT7N+tCnWR9+u/Ybs3bPYtrOaew8uxP/j/wJbRHKpK6TqFetntlRi8RqWBnw6QBOXjxJvWr1WBqyVM/nEQCi4qK4ln2Nzvd15qnGT5kdp2LKyIClS3N/1k0hRMoVU5/jZAY9Q0REiuJc+jnGbR3HJwc/wcDAqZITEe0jiPxbJK6OrmbHu6lJ2ycxbus4HO0d2fHcDnzr+JodKY+OwTdWFuPyn5/+Q/t57bFgYe+QvbTx0A0JTLFwIQwaBA0bwokTur5JxGTFOf6aesZJRMRW1a1al4XBC3nJ/yUivopg2+ltTNoxiXkH5vHWQ2/xXOvnbPIszqbvN+Vd+D/nsTk21TSJeQzDIGJT7gOPB7QaUPpN0x9/QHZ26e6jvPrww9w/Bw9W0yRSzuiMk4hIIQzDYN3xdYzZPIaTF08C0MKtBdO7Tafb/d1MTveXU7+dom1sWy7+fpHn2zxP7BOxZke6jo7BN1ba47LyyEr6rOqDcyVnTo44Sd2qdUt8HxgG7NwJs2fDmjVqnG7G3h7OngUPD7OTiFR4xTn+3vHPcRIRuV0Wi4WnvJ/i8LDDzAqaRXWn6iSmJhK0KIjHFj/G0bSjZkfkWvY1QlaEcPH3i/jW8eVf3f9ldiSxEdeyr/HqllcBGNtxbMk3TdeuwYIF0LYtdOoEK1aoaSrMP/6hpkmkHNJUPRGRInKwd2Bk+5E82+pZ3v72baLjo9n43UY2fb+J59s8z5sPvYmbi1uZ5zIMg/AN4ez/eT+1qtRidZ/VOFVyKvMcYptm/2c2P/72Ix53eTCmw5iS++CzZyEmJnfq2YULucucnOCZZyA8PPcaHrmexQIuLmanEJFboKl6IiK36OQvJ3l1y6usPbYWAFcHV8Z1GsfI9iPLtHGJ3RfLC+tfwM5ix6ZnNtG1Qdcy23dx6Rh8Y6U1LmkZaTSc3ZD0zHQWPLWAgT4Db+8DDQN27IB//QvWroWcnNzl9erlNkuDB0NNPVBXRMoPTdUTESkDjWo2Yk3fNWwbuI22Hm25nHWZyLhIvKO9WX54eZk8QDf+XDwjNo4AYNLDk2y6aZKy98Y3b5CemU4bjzYMaDXg1j/o999h/nxo3RoefBBWrcptmrp0yb2e6fvvYexYNU0ickdT4yQicpsevO9B4p+P55PgT6jrWpfTl07Tb3U/OszvwK6zu0ptv2kZaYSsCCErJ4ue3j0Z23Fsqe1Lyp+jaUf5YN8HAMzoNgM7yy2U/DNnICoKPD1zzyYdPAjOzvD883DoEHz9NfTsCZU0819E7nxqnERESoCdxY5nWz3LiREn+J8u/4NLZRd2/7SbDvM70G9VP079dqpE95dtzabf6n78lP4TjWs2ZmHwQiy6tbH8L6M3jSbHyCHYO5guXl2KvqFhwLZt0KsX1K8PU6bAL7/AfffBtGnw008QGwstWpRadhERW6TGSUSkBFWpXIXxncdzcsRJBrcejAULy48sxzvam8gtkVy6dqlE9jMubhxbf9yKS2UX1vRdQ1VHXS8kf/nqu6/Y+N1GKtlVYlrgtKJtdPUqfPQR+PjkTsFbvRqsVnj44dzrmb7/HsaMgRo1SjO6iIjNUuMkIlIKPFw9+OjJjzjwwgEerv8wmTmZTN05lYazGxKzJ4Zs663frnn10dVM+3fuP4YXPLWApvc0LanYcgfItmbzyqZXABjuN5xGNRvdfIPTp+HVV3On4/05Ba9KFXjhBUhMhLg4CA7OffaQiEgFpsZJRKQUtardii3PbuHz/p/TuGZjLly9wLAvhtEypiUbT24s9g0kjl04xsDPBgLwSsAr9G7WuxRSS3k2/8B8jqQdobpTdcZ3Hn/jlQwDvvkG/v53aNAgdwrexYu5U/OmT8+djjd3LjRvXqbZRURsmRonEZFSZrFY6PFADxKHJhLdPZqazjVJupDEY0se49HFj5KYklikz7mceZmey3tyJesKXby6MCVwSiknl/ImPTOd8V/nNksTO0+khvP/mVZ39Wru9UktW8JDD+VOwbNaITAQPvsMTp6EV16B6tVNSC8iYtvUOImIlJHK9pUJbxfOdy99x+iA0TjYO7Dp+034fODDkM+HkHwlucBtDcPguXXPcezCMeq61mVZyDIq2elOZpLflB1TSM1IpVGNRgz1G/rXG6dO5V6fdO+9uVPwDh/OnY43dCgcOQKbN8OTT2o6nojITahxEhEpY3c73c273d4lKTyJ3k17YzWsfLj/QxrNbsSk7ZP4/Y/fr9tmxq4ZrDq6isp2lVnVZxXud7mbkLzimDNnDl5eXjg5OeHv7098fPxN11+5ciXe3t44OTnRokULvvjiizJK+pfTv51m5q6ZAEzvNh0Hu8qwdWvu9Un33587Be/XX3On5s2cCefOwfvvQ1NdIyciUhRqnERETNKgegNW9F7BjkE78Kvjx5WsK4zbOo7G0Y1ZfGgxVsMKwNc/fs2rW14F4J+P/pP297Y3M/Ydb/ny5URERDBx4kT2799Pq1atCAoKIjU19Ybr//vf/6Z///4MHjyYAwcOEBwcTHBwMIcPHy7T3FFxUWTmZPKQ54M88fX53NuFd+2aOwXPaoVHHoHPP4cTJ+Dll+Huu8s0n4hIeWcxyuLR9jYkPT2datWqcenSJapW1e17RcQ2WA0ryw4vI3JLJGfTzwLgV8ePVzu+ytANQ0m7mkZYqzAWPLWgXD+vqTwcg/39/fHz8yM6OhoAq9WKp6cnI0aMIDIy8rr1+/btS0ZGBuvXr89b1r59e3x8fJg7d26R9nm747L7p90EzAvAYsC+JXfR+uSV3DdcXCAsDIYPhyZNiv25IiJ3uuIcf3XGSUTEBthZ7Hi6xdMcH36cSQ9P4i6Hu9hzfg+9VvYi7WoaPrV9iHk8plw3TeVBVlYW+/btIzAwMG+ZnZ0dgYGB7Nq164bb7Nq1K9/6AEFBQQWuD5CZmUl6enq+160yTp0i4p+PATAwgdym6f774b33cqfjzZmjpklEpASocRIRsSHOlZ2J6hTFdyO+Y0ibIdhZ7KjpXJM1fdbgXNnZ7Hh3vAsXLpCTk4O7e/5ryNzd3UlOvvHNO5KTk4u1PsDkyZOpVq1a3svT0/OWMydkn2WP869UyYK3czrD+vW50/FGjYJq1W75c0VEJD/dkklExAa53+XOB098wOsPvo5jJUfcXNzMjiQlKCoqioiIiLzf09PTb7l5at2wE0ca/5ODbgZ13hlZUhFFROT/UOMkImLDPKvd+pkIKb5atWphb29PSkpKvuUpKSnUrl37htvUrl27WOsDODo64ujoePuB/+uB0Jd4oMQ+TUREbkRT9URERP7LwcGBtm3bEhcXl7fMarUSFxdHQEDADbcJCAjItz7A5s2bC1xfRETKJ51xEhER+V8iIiIICwvD19eXdu3aMWvWLDIyMhg0aBAAAwYMoG7dukyePBmAkSNH0rlzZ2bMmMHjjz/OsmXL2Lt3L7GxsWZ+DRERKWFqnERERP6Xvn37kpaWxoQJE0hOTsbHx4cvv/wy7wYQZ86cwc7urwkbHTp0YMmSJbz++uu89tprNGrUiE8//ZTmzZub9RVERKQU6DlOIiJSZnQMvjGNi4iIOfQcJxERERERkRKkxklERERERKQQapxEREREREQKocZJRERERESkEGqcRERERERECqHGSUREREREpBBqnERERERERAqhxklERERERKQQapxEREREREQKocZJRERERESkEJXMDlDWDMMAID093eQkIiIVz5/H3j+PxZJLtUlExBzFqUsVrnG6fPkyAJ6eniYnERGpuC5fvky1atXMjmEzVJtERMxVlLpkMSrYf/tZrVbOnz+Pq6srFoul2Nunp6fj6enJ2bNnqVq1aikkLL80Njen8SmYxqZgd9rYGIbB5cuXqVOnDnZ2mi3+J9Wm0qOxKZjG5uY0PgW7k8amOHWpwp1xsrOz4957773tz6latWq5/4tSWjQ2N6fxKZjGpmB30tjoTNP1VJtKn8amYBqbm9P4FOxOGZui1iX9d5+IiIiIiEgh1DiJiIiIiIgUQo1TMTk6OjJx4kQcHR3NjmJzNDY3p/EpmMamYBobKQr9PSmYxqZgGpub0/gUrKKOTYW7OYSIiIiIiEhx6YyTiIiIiIhIIdQ4iYiIiIiIFEKNk4iIiIiISCHUOImIiIiIiBRCjVMxzZkzBy8vL5ycnPD39yc+Pt7sSKabPHkyfn5+uLq64ubmRnBwMMePHzc7lk2aMmUKFouFUaNGmR3FJpw7d45nnnmGmjVr4uzsTIsWLdi7d6/ZsWxCTk4O48ePp379+jg7O3P//ffz1ltvofv5yI2oNl1PtanoVJvyU226MdUlNU7Fsnz5ciIiIpg4cSL79++nVatWBAUFkZqaanY0U23bto3w8HB2797N5s2b+eOPP+jWrRsZGRlmR7Mpe/bs4YMPPqBly5ZmR7EJv/76Kx07dqRy5cps3LiRo0ePMmPGDKpXr252NJswdepUYmJiiI6OJikpialTpzJt2jRmz55tdjSxMapNN6baVDSqTfmpNhVMdUm3Iy8Wf39//Pz8iI6OBsBqteLp6cmIESOIjIw0OZ3tSEtLw83NjW3btvHggw+aHccmXLlyhTZt2vD+++/z9ttv4+Pjw6xZs8yOZarIyEh27tzJ9u3bzY5ik3r06IG7uzvz5s3LWxYSEoKzszOLFi0yMZnYGtWmolFtup5q0/VUmwqmuqQzTkWWlZXFvn37CAwMzFtmZ2dHYGAgu3btMjGZ7bl06RIANWrUMDmJ7QgPD+fxxx/P9/enolu3bh2+vr707t0bNzc3WrduzYcffmh2LJvRoUMH4uLiOHHiBAAHDx5kx44ddO/e3eRkYktUm4pOtel6qk3XU20qmOoSVDI7QHlx4cIFcnJycHd3z7fc3d2dY8eOmZTK9litVkaNGkXHjh1p3ry52XFswrJly9i/fz979uwxO4pN+eGHH4iJiSEiIoLXXnuNPXv28NJLL+Hg4EBYWJjZ8UwXGRlJeno63t7e2Nvbk5OTwzvvvENoaKjZ0cSGqDYVjWrT9VSbbky1qWCqS2qcpISFh4dz+PBhduzYYXYUm3D27FlGjhzJ5s2bcXJyMjuOTbFarfj6+jJp0iQAWrduzeHDh5k7d26FL04AK1asYPHixSxZsoRmzZqRkJDAqFGjqFOnjsZHpJhUm/JTbSqYalPBVJfUOBVZrVq1sLe3JyUlJd/ylJQUateubVIq2zJ8+HDWr1/Pt99+y7333mt2HJuwb98+UlNTadOmTd6ynJwcvv32W6Kjo8nMzMTe3t7EhObx8PCgadOm+ZY1adKE1atXm5TItowZM4bIyEj69esHQIsWLTh9+jSTJ0+uMAVKCqfaVDjVpuupNhVMtalgqku6xqnIHBwcaNu2LXFxcXnLrFYrcXFxBAQEmJjMfIZhMHz4cNauXcvWrVupX7++2ZFsRteuXUlMTCQhISHv5evrS2hoKAkJCRW2MAF07NjxulsDnzhxgvvuu8+kRLbl6tWr2NnlP0Tb29tjtVpNSiS2SLWpYKpNBVNtKphqU8FUl3TGqVgiIiIICwvD19eXdu3aMWvWLDIyMhg0aJDZ0UwVHh7OkiVL+Oyzz3B1dSU5ORmAatWq4ezsbHI6c7m6ul43n97FxYWaNWtW+Hn2L7/8Mh06dGDSpEn06dOH+Ph4YmNjiY2NNTuaTXjiiSd45513qFevHs2aNePAgQPMnDmT5557zuxoYmNUm25Mtalgqk0FU20qmOoSYEixzJ4926hXr57h4OBgtGvXzti9e7fZkUwH3PC1YMECs6PZpM6dOxsjR440O4ZN+Pzzz43mzZsbjo6Ohre3txEbG2t2JJuRnp5ujBw50qhXr57h5ORkNGjQwBg3bpyRmZlpdjSxQapN11NtKh7Vpr+oNt2Y6pJh6DlOIiIiIiIihdA1TiIiIiIiIoVQ4yQiIiIiIlIINU4iIiIiIiKFUOMkIiIiIiJSCDVOIiIiIiIihVDjJCIiIiIiUgg1TiIiIiIiIoVQ4yRyh7JYLHz66admxxAREcmj2iTlmRonkVIwcOBALBbLda9HH33U7GgiIlJBqTaJ3J5KZgcQuVM9+uijLFiwIN8yR0dHk9KIiIioNoncDp1xEikljo6O1K5dO9+revXqQO5UhZiYGLp3746zszMNGjRg1apV+bZPTEzk4YcfxtnZmZo1azJkyBCuXLmSb5358+fTrFkzHB0d8fDwYPjw4fnev3DhAj179qRKlSo0atSIdevW5b3366+/Ehoayj333IOzszONGjW6rpiKiMidRbVJ5NapcRIxyfjx4wkJCeHgwYOEhobSr18/kpKSAMjIyCAoKIjq1auzZ88eVq5cyZYtW/IVn5iYGMLDwxkyZAiJiYmsW7eOhg0b5tvHm2++SZ8+fTh06BCPPfYYoaGhXLx4MW//R48eZePGjSQlJRETE0OtWrXKbgBERMTmqDaJ3IQhIiUuLCzMsLe3N1xcXPK93nnnHcMwDAMwXnzxxXzb+Pv7G0OHDjUMwzBiY2ON6tWrG1euXMl7f8OGDYadnZ2RnJxsGIZh1KlTxxg3blyBGQDj9ddfz/v9ypUrBmBs3LjRMAzDeOKJJ4xBgwaVzBcWERGbp9okcnt0jZNIKXnooYeIiYnJt6xGjRp5PwcEBOR7LyAggISEBACSkpJo1aoVLi4uee937NgRq9XK8ePHsVgsnD9/nq5du940Q8uWLfN+dnFxoWrVqqSmpgIwdOhQQkJC2L9/P926dSM4OJgOHTrc0ncVEZHyQbVJ5NapcRIpJS4uLtdNTygpzs7ORVqvcuXK+X63WCxYrVYAunfvzunTp/niiy/YvHkzXbt2JTw8nOnTp5d4XhERsQ2qTSK3Ttc4iZhk9+7d1/3epEkTAJo0acLBgwfJyMjIe3/nzp3Y2dnRuHFjXF1d8fLyIi4u7rYy3HPPPYSFhbFo0SJmzZpFbGzsbX2eiIiUb6pNIgXTGSeRUpKZmUlycnK+ZZUqVcq7yHXlypX4+vryt7/9jcWLFxMfH8+8efMACA0NZeLEiYSFhfHGG2+QlpbGiBEjePbZZ3F3dwfgjTfe4MUXX8TNzY3u3btz+fJldu7cyYgRI4qUb8KECbRt25ZmzZqRmZnJ+vXr84qjiIjcmVSbRG6dGieRUvLll1/i4eGRb1njxo05duwYkHtXoWXLljFs2DA8PDxYunQpTZs2BaBKlSp89dVXjBw5Ej8/P6pUqUJISAgzZ87M+6ywsDCuXbvGe++9x+jRo6lVqxa9evUqcj4HBweioqI4deoUzs7OdOrUiWXLlpXANxcREVul2iRy6yyGYRhmhxCpaCwWC2vXriU4ONjsKCIiIoBqk0hhdI2TiIiIiIhIIdQ4iYiIiIiIFEJT9URERERERAqhM04iIiIiIiKFUOMkIiIiIiJSCDVOIiIiIiIihVDjJCIiIiIiUgg1TiIiIiIiIoVQ4yQiIiIiIlIINU4iIiIiIiKFUOMkIiIiIiJSCDVOIiIiIiIihfj/d7CfLhj4akEAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 1000x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the loss and accuracy\n",
    "# https://matplotlib.org/stable/api/_as_gen/matplotlib.pyplot.legend.html\n",
    "# We can see from below the loss has decreased while the accuracy has increased.\n",
    "fig, ax = plt.subplots(ncols=2, nrows=1, figsize=(10,5))\n",
    "plt.suptitle('Loss and Recall')\n",
    "\n",
    "ax[0].plot(tf_clf_model.history.history['loss'], 'r', tf_clf_model.history.history['val_loss'], 'g' )\n",
    "ax[0].set_title('Loss')\n",
    "ax[0].legend(labels=['Train loss', 'Validation loss'])\n",
    "ax[0].set_xlabel('Epochs')\n",
    "ax[0].set_ylabel('Loss')\n",
    "\n",
    "ax[1].plot(tf_clf_model.history.history['recall_1'], 'r', tf_clf_model.history.history['val_recall_1'], 'g' )\n",
    "ax[1].set_title('Recall')\n",
    "ax[1].legend(labels=['Train Recall', 'Validation Recall'])\n",
    "ax[1].set_xlabel('Epochs')\n",
    "ax[1].set_ylabel('Recall')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "# With the training loss trending downwards, this suggest a few more epochs may make the mode perform even better\n",
    "# How did our model do for training\n",
    "# import some metrics\n",
    "from sklearn.metrics import accuracy_score, precision_score, confusion_matrix, \\\n",
    "    precision_recall_curve, auc, classification_report, roc_curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3706409    0\n",
       "2141747    0\n",
       "117303     0\n",
       "15089      0\n",
       "24905      0\n",
       "          ..\n",
       "36391      0\n",
       "739641     0\n",
       "2987       0\n",
       "3028447    0\n",
       "1046305    0\n",
       "Name: label, Length: 41522, dtype: int64"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Looking at y_true\n",
    "y_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1298/1298 [==============================] - 2s 2ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[0.11998063],\n",
       "       [0.1144481 ],\n",
       "       [0.11478651],\n",
       "       ...,\n",
       "       [0.11444854],\n",
       "       [0.11444541],\n",
       "       [0.11444263]], dtype=float32)"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# When making the prediction, our values will be continuous such as\n",
    "tf_clf_model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1298/1298 [==============================] - 2s 2ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[0.],\n",
       "       [0.],\n",
       "       [0.],\n",
       "       ...,\n",
       "       [0.],\n",
       "       [0.],\n",
       "       [0.]], dtype=float32)"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# If we try to feed this to some of our metrics algorithm, it will get a value error such as\n",
    "# \" Classification metrics can't handle a mix of binary and continuous targets\"\n",
    "# As a result, I round it out instead\n",
    "np.round(tf_clf_model.predict(X_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1298/1298 [==============================] - 2s 2ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[0.],\n",
       "       [0.],\n",
       "       [0.],\n",
       "       ...,\n",
       "       [0.],\n",
       "       [0.],\n",
       "       [0.]], dtype=float32)"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Let's save these predictions out to a variable\n",
    "x_test_preds = np.round(tf_clf_model.predict(X_test))\n",
    "x_test_preds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9985068156639854"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Grabbing the accuracy score\n",
    "# With that understanding above, let grab the accuracy score\n",
    "accuracy_score(y_true=y_test.astype('float'), y_pred=x_test_preds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import seaborn\n",
    "# https://seaborn.pydata.org/generated/seaborn.heatmap.html\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Looking at the confusion matrix, This does not seem so bad\n",
    "# We learned about metrics in notebook:\n",
    "#   10. Beginning Supervise Learning - Machine Learning - Logistic Regression, Decision Trees and Metrics\n",
    "sns.heatmap(confusion_matrix(y_true=y_test.astype('float'), y_pred=x_test_preds), annot=True, fmt='d')\n",
    "plt.title('Confusion Matrix')\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('Actual')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       1.00      1.00      1.00     41509\n",
      "         1.0       0.07      0.31      0.11        13\n",
      "\n",
      "    accuracy                           1.00     41522\n",
      "   macro avg       0.53      0.65      0.56     41522\n",
      "weighted avg       1.00      1.00      1.00     41522\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Overall, this model is not doing too bad. Context is important!!\n",
    "# This is actually a model, I would put in production for my security monitoring.\n",
    "# If I am able to ignore 41,456 records and focus only on 53 (false positives) + 4 (true positives) = 57 records of concern\n",
    "# Obviously, no one wants the 9 false negatives. Hopefully, we can catch those \"threats\" via threat hunting\n",
    "# Looking at the classification report\n",
    "print(classification_report(y_true=y_test.astype('float'), y_pred=x_test_preds))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "13"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Confirming there are only 13 records flagged as suspicious\n",
    "len(list(np.where(y_test == 1))[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1.41000000e+02, 3.56138566e+08, 1.10370900e+07, 6.00000000e+01,\n",
       "        3.02667900e+06, 3.30000000e+01, 9.82584000e+05]])"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Let's define a sample with the features values: \n",
    "# duration\torig_bytes\tresp_bytes\torig_pkts\torig_ip_bytes\tresp_pkts\tresp_ip_bytes\n",
    "new_sample = np.array([141., 356138566,\t11037090, 60, 3026679, 33, 982584], dtype=float, ndmin=2)\n",
    "new_sample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 40ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[0.57959896]], dtype=float32)"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Make a prediction on the sample\n",
    "# Remember, the previously unseen data has to go through the same transformation as the training data\n",
    "# See:\n",
    "#   06 - Beginning Data Scaling\n",
    "#   07 - Beginning Principal Component Analysis (PCA)\n",
    "new_pred = tf_clf_model.predict(pca.transform(min_max_scaler.transform(new_sample)))\n",
    "new_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import the data time library\n",
    "from datetime import datetime"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'2023-05-28 18:39:34.706679 - [!] ALERT ** SUSPICIOUS ACTIVITY ** Zeek conn.log'"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Report a sample as suspicious, if it's threshold is greater than 0.5\n",
    "f'{datetime.now()} - [!] ALERT ** SUSPICIOUS ACTIVITY ** Zeek conn.log' if new_pred > 0.5  \\\n",
    "    else \"[**] {datetime.now()} - Normal Traffic\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[5.00000000e+03, 3.56138566e+08, 1.10370900e+07, 6.00000000e+01,\n",
       "        3.02667900e+06, 3.30000000e+01, 9.82584000e+05]])"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Setup a suspicious sample\n",
    "new_sample = np.array([5000., 356138566,\t11037090, 60, 3026679, 33, 982584], dtype=float, ndmin=2)\n",
    "new_sample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 30ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[0.99791276]], dtype=float32)"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Transform the sample\n",
    "new_pred = tf_clf_model.predict(pca.transform(min_max_scaler.transform(new_sample)))\n",
    "new_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'2023-05-28 18:39:38.986373 - [!] ALERT ** SUSPICIOUS ACTIVITY ** Zeek conn.log'"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Report a sample as suspicious, if it's threshold is greater than 0.5\n",
    "f'{datetime.now()} - [!] ALERT ** SUSPICIOUS ACTIVITY ** Zeek conn.log' if new_pred > 0.5  \\\n",
    "    else \"[**] {datetime.now()} - Normal Traffic\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import the os library\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create the location to save the model\n",
    "PATH = './SAVED_MODELS/tf_classification/'\n",
    "MODELS_PATH = os.makedirs(name=PATH, exist_ok=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as _update_step_xla while saving (showing 1 of 1). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: ./SAVED_MODELS/tf_classification//tf_clf_model.tf\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: ./SAVED_MODELS/tf_classification//tf_clf_model.tf\\assets\n"
     ]
    }
   ],
   "source": [
    "# Save the model\n",
    "tf.keras.models.save_model(model=tf_clf_model, filepath=f'{PATH}/tf_clf_model.tf', \\\n",
    "                           overwrite=True, save_format='tf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<keras.engine.sequential.Sequential at 0x227be480c70>"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# With the model saved, time to load it\n",
    "loaded_tf_clf_model = tf.keras.models.load_model(filepath=f'{PATH}tf_clf_model.tf')\n",
    "loaded_tf_clf_model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1/1 [==============================] - 0s 153ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[0.99791276]], dtype=float32)"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Make a prediction on the loaded model\n",
    "loaded_tf_clf_model.predict(pca.transform(min_max_scaler.transform(new_sample)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "# That's it! Moving on!!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "main-env",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.11"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "dcb100998b7ef434a33dbfac35f576e40d01f04c34d9ff6f6aad819b8a88c169"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}