path: root/datasets/Talluri2021/other notebooks/4.failures from reports.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from plotnine import *\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "or_events = pd.read_parquet('./outage_report_2019-20')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [],
   "source": [
    "filtered_or_events = or_events[(or_events['vendor'] != '') & (or_events['vendor'] != 'overview')].reset_index(drop=True)\n",
    "filtered_or_events.loc[filtered_or_events['status_code'] <= 2, 'status_code'] = 0\n",
    "filtered_or_events = filtered_or_events.drop_duplicates(subset=['vendor', 'event_time', 'status_code'])\n",
    "filtered_or_events = filtered_or_events.groupby(['vendor', 'event_time'])['status_code'].max().reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [],
   "source": [
    "filtered_or_events['evtime'] = pd.to_datetime(filtered_or_events['event_time'], unit='s')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [],
   "source": [
    "def proper_vendor_names(series):\n",
    "    return series.str.capitalize().replace(['Apple-servers', 'Facebook-messenger', 'Youtube'], ['Apple', 'FB Msgr', 'YouTube'])\n",
    "\n",
    "filtered_or_events['vendor_proper'] = proper_vendor_names(filtered_or_events['vendor'])\n",
    "\n",
    "vendor_list = list(filtered_or_events.groupby('vendor_proper')['status_code'].sum().reset_index().rename(columns={'status_code':'count'}).sort_values('count')['vendor_proper'])\n",
    "\n",
    "filtered_or_events['vendor_cat'] = pd.Categorical(filtered_or_events['vendor_proper'], ordered=True, categories=vendor_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "facebook_reports = filtered_or_events[filtered_or_events['vendor'] == 'facebook'].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "sorted_facebook_reports = facebook_reports.sort_values('event_time').reset_index(drop=True)"
   ]
  },
  {
   "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>event_time</th>\n",
       "      <th>status_code</th>\n",
       "      <th>vendor</th>\n",
       "      <th>monitor</th>\n",
       "      <th>evtime</th>\n",
       "      <th>vendor_proper</th>\n",
       "      <th>vendor_cat</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.555241e+09</td>\n",
       "      <td>210.0</td>\n",
       "      <td>facebook</td>\n",
       "      <td>Outage Report</td>\n",
       "      <td>2019-04-14 11:20:00</td>\n",
       "      <td>Facebook</td>\n",
       "      <td>Facebook</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.555242e+09</td>\n",
       "      <td>299.0</td>\n",
       "      <td>facebook</td>\n",
       "      <td>Outage Report</td>\n",
       "      <td>2019-04-14 11:40:00</td>\n",
       "      <td>Facebook</td>\n",
       "      <td>Facebook</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.555243e+09</td>\n",
       "      <td>813.0</td>\n",
       "      <td>facebook</td>\n",
       "      <td>Outage Report</td>\n",
       "      <td>2019-04-14 12:00:00</td>\n",
       "      <td>Facebook</td>\n",
       "      <td>Facebook</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.555244e+09</td>\n",
       "      <td>874.0</td>\n",
       "      <td>facebook</td>\n",
       "      <td>Outage Report</td>\n",
       "      <td>2019-04-14 12:20:00</td>\n",
       "      <td>Facebook</td>\n",
       "      <td>Facebook</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.555246e+09</td>\n",
       "      <td>891.0</td>\n",
       "      <td>facebook</td>\n",
       "      <td>Outage Report</td>\n",
       "      <td>2019-04-14 12:40:00</td>\n",
       "      <td>Facebook</td>\n",
       "      <td>Facebook</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>43674</th>\n",
       "      <td>1.604180e+09</td>\n",
       "      <td>0.0</td>\n",
       "      <td>facebook</td>\n",
       "      <td>Outage Report</td>\n",
       "      <td>2020-10-31 21:40:00</td>\n",
       "      <td>Facebook</td>\n",
       "      <td>Facebook</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43675</th>\n",
       "      <td>1.604182e+09</td>\n",
       "      <td>0.0</td>\n",
       "      <td>facebook</td>\n",
       "      <td>Outage Report</td>\n",
       "      <td>2020-10-31 22:00:00</td>\n",
       "      <td>Facebook</td>\n",
       "      <td>Facebook</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43676</th>\n",
       "      <td>1.604183e+09</td>\n",
       "      <td>0.0</td>\n",
       "      <td>facebook</td>\n",
       "      <td>Outage Report</td>\n",
       "      <td>2020-10-31 22:20:00</td>\n",
       "      <td>Facebook</td>\n",
       "      <td>Facebook</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43677</th>\n",
       "      <td>1.604184e+09</td>\n",
       "      <td>0.0</td>\n",
       "      <td>facebook</td>\n",
       "      <td>Outage Report</td>\n",
       "      <td>2020-10-31 22:40:00</td>\n",
       "      <td>Facebook</td>\n",
       "      <td>Facebook</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43678</th>\n",
       "      <td>1.604185e+09</td>\n",
       "      <td>0.0</td>\n",
       "      <td>facebook</td>\n",
       "      <td>Outage Report</td>\n",
       "      <td>2020-10-31 23:00:00</td>\n",
       "      <td>Facebook</td>\n",
       "      <td>Facebook</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>43679 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         event_time  status_code    vendor        monitor              evtime  \\\n",
       "0      1.555241e+09        210.0  facebook  Outage Report 2019-04-14 11:20:00   \n",
       "1      1.555242e+09        299.0  facebook  Outage Report 2019-04-14 11:40:00   \n",
       "2      1.555243e+09        813.0  facebook  Outage Report 2019-04-14 12:00:00   \n",
       "3      1.555244e+09        874.0  facebook  Outage Report 2019-04-14 12:20:00   \n",
       "4      1.555246e+09        891.0  facebook  Outage Report 2019-04-14 12:40:00   \n",
       "...             ...          ...       ...            ...                 ...   \n",
       "43674  1.604180e+09          0.0  facebook  Outage Report 2020-10-31 21:40:00   \n",
       "43675  1.604182e+09          0.0  facebook  Outage Report 2020-10-31 22:00:00   \n",
       "43676  1.604183e+09          0.0  facebook  Outage Report 2020-10-31 22:20:00   \n",
       "43677  1.604184e+09          0.0  facebook  Outage Report 2020-10-31 22:40:00   \n",
       "43678  1.604185e+09          0.0  facebook  Outage Report 2020-10-31 23:00:00   \n",
       "\n",
       "      vendor_proper vendor_cat  \n",
       "0          Facebook   Facebook  \n",
       "1          Facebook   Facebook  \n",
       "2          Facebook   Facebook  \n",
       "3          Facebook   Facebook  \n",
       "4          Facebook   Facebook  \n",
       "...             ...        ...  \n",
       "43674      Facebook   Facebook  \n",
       "43675      Facebook   Facebook  \n",
       "43676      Facebook   Facebook  \n",
       "43677      Facebook   Facebook  \n",
       "43678      Facebook   Facebook  \n",
       "\n",
       "[43679 rows x 7 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sorted_facebook_reports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# An event has start, end, peak, median, mean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract_failures_from_reports(partial_df):\n",
    "    event_start = None\n",
    "    prev_event_time = None\n",
    "    num_reports_list = []\n",
    "    failure_events = []\n",
    "    for _, report_event in partial_df.sort_values('event_time').iterrows():\n",
    "        num_reports = report_event['status_code']\n",
    "        event_time = report_event['event_time']\n",
    "\n",
    "        if num_reports == 0:\n",
    "            if event_start is not None:\n",
    "                failure_events.append({\n",
    "                    'start_time': event_start,\n",
    "                    'end_time': prev_event_time,\n",
    "                    'peak': np.max(num_reports_list),\n",
    "                    'median': np.median(num_reports_list),\n",
    "                    'mean': np.mean(num_reports_list)\n",
    "                })\n",
    "                event_start = None\n",
    "                num_reports_list = []\n",
    "        else:\n",
    "            if event_start is None:\n",
    "                event_start = event_time\n",
    "                \n",
    "            num_reports_list.append(num_reports)\n",
    "\n",
    "        prev_event_time = event_time\n",
    "    \n",
    "    return pd.DataFrame(failure_events)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 284,
   "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>start_time</th>\n",
       "      <th>end_time</th>\n",
       "      <th>peak</th>\n",
       "      <th>median</th>\n",
       "      <th>mean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2053</th>\n",
       "      <td>1.564102e+09</td>\n",
       "      <td>1.564102e+09</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4601</th>\n",
       "      <td>1.582679e+09</td>\n",
       "      <td>1.582679e+09</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4600</th>\n",
       "      <td>1.582675e+09</td>\n",
       "      <td>1.582675e+09</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2433</th>\n",
       "      <td>1.565659e+09</td>\n",
       "      <td>1.565659e+09</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2434</th>\n",
       "      <td>1.565664e+09</td>\n",
       "      <td>1.565664e+09</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3965</th>\n",
       "      <td>1.574808e+09</td>\n",
       "      <td>1.574808e+09</td>\n",
       "      <td>215.0</td>\n",
       "      <td>215.0</td>\n",
       "      <td>215.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3994</th>\n",
       "      <td>1.574950e+09</td>\n",
       "      <td>1.574952e+09</td>\n",
       "      <td>415.0</td>\n",
       "      <td>322.0</td>\n",
       "      <td>274.571429</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3996</th>\n",
       "      <td>1.574958e+09</td>\n",
       "      <td>1.574964e+09</td>\n",
       "      <td>550.0</td>\n",
       "      <td>345.0</td>\n",
       "      <td>253.454545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3995</th>\n",
       "      <td>1.574953e+09</td>\n",
       "      <td>1.574957e+09</td>\n",
       "      <td>553.0</td>\n",
       "      <td>453.0</td>\n",
       "      <td>463.363636</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>402</th>\n",
       "      <td>1.557258e+09</td>\n",
       "      <td>1.557259e+09</td>\n",
       "      <td>952.0</td>\n",
       "      <td>908.0</td>\n",
       "      <td>622.666667</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5490 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        start_time      end_time   peak  median        mean\n",
       "2053  1.564102e+09  1.564102e+09    3.0     3.0    3.000000\n",
       "4601  1.582679e+09  1.582679e+09    3.0     3.0    3.000000\n",
       "4600  1.582675e+09  1.582675e+09    3.0     3.0    3.000000\n",
       "2433  1.565659e+09  1.565659e+09    3.0     3.0    3.000000\n",
       "2434  1.565664e+09  1.565664e+09    3.0     3.0    3.000000\n",
       "...            ...           ...    ...     ...         ...\n",
       "3965  1.574808e+09  1.574808e+09  215.0   215.0  215.000000\n",
       "3994  1.574950e+09  1.574952e+09  415.0   322.0  274.571429\n",
       "3996  1.574958e+09  1.574964e+09  550.0   345.0  253.454545\n",
       "3995  1.574953e+09  1.574957e+09  553.0   453.0  463.363636\n",
       "402   1.557258e+09  1.557259e+09  952.0   908.0  622.666667\n",
       "\n",
       "[5490 rows x 5 columns]"
      ]
     },
     "execution_count": 284,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Validate this later by plotting both the intervals and direct from reports.\n",
    "# To check if the failure creation algorithm is correct\n",
    "facebook_failures = extract_failures_from_reports(sorted_facebook_reports)\n",
    "facebook_failures.sort_values('median')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 285,
   "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>start_time</th>\n",
       "      <th>end_time</th>\n",
       "      <th>peak</th>\n",
       "      <th>median</th>\n",
       "      <th>mean</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2053</th>\n",
       "      <td>1564101600</td>\n",
       "      <td>1564101600</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4601</th>\n",
       "      <td>1582678800</td>\n",
       "      <td>1582678800</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4600</th>\n",
       "      <td>1582675200</td>\n",
       "      <td>1582675200</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2433</th>\n",
       "      <td>1565659200</td>\n",
       "      <td>1565659200</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2434</th>\n",
       "      <td>1565664000</td>\n",
       "      <td>1565664000</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3965</th>\n",
       "      <td>1574808000</td>\n",
       "      <td>1574808000</td>\n",
       "      <td>215.0</td>\n",
       "      <td>215.0</td>\n",
       "      <td>215.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3994</th>\n",
       "      <td>1574949600</td>\n",
       "      <td>1574952000</td>\n",
       "      <td>415.0</td>\n",
       "      <td>322.0</td>\n",
       "      <td>274.571429</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3996</th>\n",
       "      <td>1574958000</td>\n",
       "      <td>1574964000</td>\n",
       "      <td>550.0</td>\n",
       "      <td>345.0</td>\n",
       "      <td>253.454545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3995</th>\n",
       "      <td>1574953200</td>\n",
       "      <td>1574956800</td>\n",
       "      <td>553.0</td>\n",
       "      <td>453.0</td>\n",
       "      <td>463.363636</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>402</th>\n",
       "      <td>1557258000</td>\n",
       "      <td>1557259200</td>\n",
       "      <td>952.0</td>\n",
       "      <td>908.0</td>\n",
       "      <td>622.666667</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5490 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      start_time    end_time   peak  median        mean\n",
       "2053  1564101600  1564101600    3.0     3.0    3.000000\n",
       "4601  1582678800  1582678800    3.0     3.0    3.000000\n",
       "4600  1582675200  1582675200    3.0     3.0    3.000000\n",
       "2433  1565659200  1565659200    3.0     3.0    3.000000\n",
       "2434  1565664000  1565664000    3.0     3.0    3.000000\n",
       "...          ...         ...    ...     ...         ...\n",
       "3965  1574808000  1574808000  215.0   215.0  215.000000\n",
       "3994  1574949600  1574952000  415.0   322.0  274.571429\n",
       "3996  1574958000  1574964000  550.0   345.0  253.454545\n",
       "3995  1574953200  1574956800  553.0   453.0  463.363636\n",
       "402   1557258000  1557259200  952.0   908.0  622.666667\n",
       "\n",
       "[5490 rows x 5 columns]"
      ]
     },
     "execution_count": 285,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "facebook_failures['start_time'] = facebook_failures['start_time'].astype(np.int64)\n",
    "facebook_failures['end_time'] = facebook_failures['end_time'].astype(np.int64)\n",
    "facebook_failures.sort_values('median')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "failure_df_list = []\n",
    "\n",
    "for vendor in filtered_or_events['vendor_cat'].unique():\n",
    "    partial_df = filtered_or_events[filtered_or_events['vendor_cat'] == vendor].reset_index(drop=True)\n",
    "    \n",
    "    partial_failure_df = extract_failures_from_reports(partial_df)\n",
    "    partial_failure_df['vendor_cat'] = vendor\n",
    "    \n",
    "    failure_df_list.append(partial_failure_df)\n",
    "    \n",
    "failure_df = pd.concat(failure_df_list)\n",
    "failure_df['vendor_cat'] = pd.Categorical(failure_df['vendor_cat'], ordered=True, categories=vendor_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 235,
   "metadata": {},
   "outputs": [],
   "source": [
    "failure_df['duration'] = failure_df['end_time'] - failure_df['start_time']\n",
    "filtered_failure_df = failure_df[failure_df['duration'] >= 0].reset_index(drop=True) # filters 5 events for now\n",
    "filtered_failure_df.loc[filtered_failure_df['duration'] == 0, 'duration'] = 1200 # 0 means the event lasted less than 20 minutes. round it up to 20\n",
    "filtered_failure_df = filtered_failure_df.reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 282,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\begin{tabular}{llrr}\n",
      "\\toprule\n",
      "{} & vendor\\_proper &  count\\_events &  count\\_failures \\\\\n",
      "\\midrule\n",
      "0  &         Apple &           811 &              38 \\\\\n",
      "1  &         Skype &          2103 &              89 \\\\\n",
      "2  &        Github &          2974 &              68 \\\\\n",
      "3  &         Gmail &          6227 &             426 \\\\\n",
      "4  &       FB Msgr &          7948 &             210 \\\\\n",
      "5  &      Whatsapp &         21235 &             884 \\\\\n",
      "6  &      Snapchat &         52620 &            2549 \\\\\n",
      "7  &       Netflix &         59595 &            3388 \\\\\n",
      "8  &      Facebook &         60375 &            4069 \\\\\n",
      "9  &       Twitter &        123862 &            2908 \\\\\n",
      "10 &       YouTube &        132760 &            4219 \\\\\n",
      "11 &     Instagram &        175702 &            3926 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    }
   ],
   "source": [
    "event_count_df = filtered_or_events.groupby('vendor_proper')['status_code'].sum().reset_index().rename(columns={'status_code':'count'})\n",
    "failure_count_df = filtered_failure_df.groupby('vendor_cat')['duration'].count().reset_index().rename(columns={'duration':'count'}).set_index('vendor_cat')\n",
    "\n",
    "joined_count_df = event_count_df.join(failure_count_df, on='vendor_proper', lsuffix='_events', rsuffix='_failures')\n",
    "joined_count_df['count_events'] = joined_count_df['count_events'].astype(int)\n",
    "\n",
    "print(joined_count_df.sort_values('count_events').reset_index(drop=True).to_latex(float_format=\"%.2f\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 283,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "646212\n",
      "22774\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>vendor_cat</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Apple</th>\n",
       "      <td>38</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Github</th>\n",
       "      <td>68</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Skype</th>\n",
       "      <td>89</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FB Msgr</th>\n",
       "      <td>210</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Gmail</th>\n",
       "      <td>426</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Whatsapp</th>\n",
       "      <td>884</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Snapchat</th>\n",
       "      <td>2549</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Twitter</th>\n",
       "      <td>2908</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Netflix</th>\n",
       "      <td>3388</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Instagram</th>\n",
       "      <td>3926</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Facebook</th>\n",
       "      <td>4069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YouTube</th>\n",
       "      <td>4219</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            count\n",
       "vendor_cat       \n",
       "Apple          38\n",
       "Github         68\n",
       "Skype          89\n",
       "FB Msgr       210\n",
       "Gmail         426\n",
       "Whatsapp      884\n",
       "Snapchat     2549\n",
       "Twitter      2908\n",
       "Netflix      3388\n",
       "Instagram    3926\n",
       "Facebook     4069\n",
       "YouTube      4219"
      ]
     },
     "execution_count": 283,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(joined_count_df['count_events'].sum())\n",
    "print(joined_count_df['count_failures'].sum())\n",
    "failure_count_df.sort_values('count')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 292,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Compute all CDFs\n",
    "\n",
    "def compute_cdf(partial_df, variate):\n",
    "    count_df = partial_df.groupby(variate)['start_time'].count().reset_index().rename(columns={'start_time':'count'}).sort_values(variate).reset_index(drop=True)\n",
    "    count_df['prop'] = count_df['count'] / count_df['count'].sum()\n",
    "    count_df['cdf'] = count_df['prop'].cumsum()\n",
    "    return count_df\n",
    "\n",
    "duration_cdf_df_list = []\n",
    "peak_cdf_df_list = []\n",
    "median_cdf_df_list = []\n",
    "mean_cdf_df_list = []\n",
    "ia_cdf_df_list = []\n",
    "ia_count_df_list = []\n",
    "\n",
    "for vendor in filtered_failure_df['vendor_cat'].unique():\n",
    "    partial_df = filtered_failure_df[filtered_failure_df['vendor_cat'] == vendor].reset_index(drop=True)\n",
    "    \n",
    "    partial_duration_cdf_df = compute_cdf(partial_df, 'duration')\n",
    "    partial_duration_cdf_df['vendor_cat'] = vendor\n",
    "    duration_cdf_df_list.append(partial_duration_cdf_df)\n",
    "    \n",
    "    partial_peak_cdf_df = compute_cdf(partial_df, 'peak')\n",
    "    partial_peak_cdf_df['vendor_cat'] = vendor\n",
    "    peak_cdf_df_list.append(partial_peak_cdf_df)\n",
    "    \n",
    "    partial_median_cdf_df = compute_cdf(partial_df, 'median')\n",
    "    partial_median_cdf_df['vendor_cat'] = vendor\n",
    "    median_cdf_df_list.append(partial_median_cdf_df)\n",
    "    \n",
    "    partial_mean_cdf_df = compute_cdf(partial_df, 'mean')\n",
    "    partial_mean_cdf_df['vendor_cat'] = vendor\n",
    "    mean_cdf_df_list.append(partial_mean_cdf_df)\n",
    "    \n",
    "    # Compute interarrival time\n",
    "    sorted_partial_df = partial_df.sort_values('start_time').reset_index(drop=True)\n",
    "    ia_arr = sorted_partial_df['start_time'].values - np.roll(sorted_partial_df['start_time'].values, 1)\n",
    "    partial_ia_df = pd.DataFrame({'ia': ia_arr[1:], 'count':1})\n",
    "    partial_ia_df['vendor_cat'] = vendor\n",
    "    ia_count_df = partial_ia_df.groupby('ia')['count'].count().reset_index()\n",
    "    ia_count_df['prop'] = ia_count_df['count'] / ia_count_df['count'].sum()\n",
    "    ia_count_df['cdf'] = ia_count_df['prop'].cumsum()\n",
    "    ia_count_df['vendor_cat'] = vendor\n",
    "    ia_count_df_list.append(partial_ia_df)\n",
    "    ia_cdf_df_list.append(ia_count_df)\n",
    "    \n",
    "duration_cdf_df = pd.concat(duration_cdf_df_list).reset_index(drop=True)\n",
    "duration_cdf_df['vendor_cat'] = pd.Categorical(duration_cdf_df['vendor_cat'], ordered=True, categories=vendor_list)\n",
    "\n",
    "peak_cdf_df = pd.concat(peak_cdf_df_list).reset_index(drop=True)\n",
    "peak_cdf_df['vendor_cat'] = pd.Categorical(peak_cdf_df['vendor_cat'], ordered=True, categories=vendor_list)\n",
    "\n",
    "median_cdf_df = pd.concat(median_cdf_df_list).reset_index(drop=True)\n",
    "median_cdf_df['vendor_cat'] = pd.Categorical(median_cdf_df['vendor_cat'], ordered=True, categories=vendor_list)\n",
    "\n",
    "mean_cdf_df = pd.concat(mean_cdf_df_list).reset_index(drop=True)\n",
    "mean_cdf_df['vendor_cat'] = pd.Categorical(mean_cdf_df['vendor_cat'], ordered=True, categories=vendor_list)\n",
    "\n",
    "ia_cdf_df = pd.concat(ia_cdf_df_list).reset_index(drop=True)\n",
    "ia_cdf_df['vendor_cat'] = pd.Categorical(ia_cdf_df['vendor_cat'], ordered=True, categories=vendor_list)\n",
    "\n",
    "ia_count_df = pd.concat(ia_count_df_list).reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 260,
   "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>vendor_cat</th>\n",
       "      <th>duration</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Github</td>\n",
       "      <td>7200.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Apple</td>\n",
       "      <td>10800.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Gmail</td>\n",
       "      <td>21600.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Skype</td>\n",
       "      <td>22800.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>FB Msgr</td>\n",
       "      <td>30000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Whatsapp</td>\n",
       "      <td>40800.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Facebook</td>\n",
       "      <td>75600.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Snapchat</td>\n",
       "      <td>108000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Netflix</td>\n",
       "      <td>117600.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>YouTube</td>\n",
       "      <td>146400.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Instagram</td>\n",
       "      <td>171600.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Twitter</td>\n",
       "      <td>238800.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   vendor_cat  duration\n",
       "2      Github    7200.0\n",
       "0       Apple   10800.0\n",
       "3       Gmail   21600.0\n",
       "1       Skype   22800.0\n",
       "4     FB Msgr   30000.0\n",
       "5    Whatsapp   40800.0\n",
       "8    Facebook   75600.0\n",
       "6    Snapchat  108000.0\n",
       "7     Netflix  117600.0\n",
       "10    YouTube  146400.0\n",
       "11  Instagram  171600.0\n",
       "9     Twitter  238800.0"
      ]
     },
     "execution_count": 260,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "duration_cdf_df.groupby('vendor_cat')['duration'].max().reset_index().sort_values('duration')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 279,
   "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>vendor_cat</th>\n",
       "      <th>level_1</th>\n",
       "      <th>duration</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Gmail</td>\n",
       "      <td>0.96</td>\n",
       "      <td>2400.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Whatsapp</td>\n",
       "      <td>0.96</td>\n",
       "      <td>2400.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Snapchat</td>\n",
       "      <td>0.96</td>\n",
       "      <td>4800.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Skype</td>\n",
       "      <td>0.96</td>\n",
       "      <td>6000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Github</td>\n",
       "      <td>0.96</td>\n",
       "      <td>6000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>FB Msgr</td>\n",
       "      <td>0.96</td>\n",
       "      <td>6000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Facebook</td>\n",
       "      <td>0.96</td>\n",
       "      <td>6000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Apple</td>\n",
       "      <td>0.96</td>\n",
       "      <td>7248.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>YouTube</td>\n",
       "      <td>0.96</td>\n",
       "      <td>8400.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Netflix</td>\n",
       "      <td>0.96</td>\n",
       "      <td>12624.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Twitter</td>\n",
       "      <td>0.96</td>\n",
       "      <td>20400.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Instagram</td>\n",
       "      <td>0.96</td>\n",
       "      <td>20400.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   vendor_cat  level_1  duration\n",
       "3       Gmail     0.96    2400.0\n",
       "5    Whatsapp     0.96    2400.0\n",
       "6    Snapchat     0.96    4800.0\n",
       "1       Skype     0.96    6000.0\n",
       "2      Github     0.96    6000.0\n",
       "4     FB Msgr     0.96    6000.0\n",
       "8    Facebook     0.96    6000.0\n",
       "0       Apple     0.96    7248.0\n",
       "10    YouTube     0.96    8400.0\n",
       "7     Netflix     0.96   12624.0\n",
       "9     Twitter     0.96   20400.0\n",
       "11  Instagram     0.96   20400.0"
      ]
     },
     "execution_count": 279,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "failure_df.groupby('vendor_cat')['duration'].quantile([0.96]).reset_index().sort_values('duration')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 280,
   "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>vendor_cat</th>\n",
       "      <th>ia</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Instagram</td>\n",
       "      <td>297600.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Twitter</td>\n",
       "      <td>1011600.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>YouTube</td>\n",
       "      <td>1018800.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Facebook</td>\n",
       "      <td>1386000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Netflix</td>\n",
       "      <td>1776000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>FB Msgr</td>\n",
       "      <td>2221200.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Snapchat</td>\n",
       "      <td>2451600.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Gmail</td>\n",
       "      <td>2532000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Github</td>\n",
       "      <td>3805200.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Whatsapp</td>\n",
       "      <td>5170800.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Skype</td>\n",
       "      <td>6337200.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Apple</td>\n",
       "      <td>11250000.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   vendor_cat          ia\n",
       "11  Instagram    297600.0\n",
       "9     Twitter   1011600.0\n",
       "10    YouTube   1018800.0\n",
       "8    Facebook   1386000.0\n",
       "7     Netflix   1776000.0\n",
       "4     FB Msgr   2221200.0\n",
       "6    Snapchat   2451600.0\n",
       "3       Gmail   2532000.0\n",
       "2      Github   3805200.0\n",
       "5    Whatsapp   5170800.0\n",
       "1       Skype   6337200.0\n",
       "0       Apple  11250000.0"
      ]
     },
     "execution_count": 280,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ia_cdf_df.groupby('vendor_cat')['ia'].max().reset_index().sort_values('ia')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 295,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5     6000.0\n",
       "0.9    27600.0\n",
       "Name: ia, dtype: float64"
      ]
     },
     "execution_count": 295,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ia_count_df['ia'].quantile([0.50, 0.90])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 297,
   "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>vendor_cat</th>\n",
       "      <th>level_1</th>\n",
       "      <th>ia</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Facebook</td>\n",
       "      <td>0.75</td>\n",
       "      <td>8400.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Netflix</td>\n",
       "      <td>0.75</td>\n",
       "      <td>8400.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>YouTube</td>\n",
       "      <td>0.75</td>\n",
       "      <td>8400.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Snapchat</td>\n",
       "      <td>0.75</td>\n",
       "      <td>9600.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Instagram</td>\n",
       "      <td>0.75</td>\n",
       "      <td>12000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Twitter</td>\n",
       "      <td>0.75</td>\n",
       "      <td>12000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Whatsapp</td>\n",
       "      <td>0.75</td>\n",
       "      <td>41400.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Gmail</td>\n",
       "      <td>0.75</td>\n",
       "      <td>98400.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>FB Msgr</td>\n",
       "      <td>0.75</td>\n",
       "      <td>306000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Skype</td>\n",
       "      <td>0.75</td>\n",
       "      <td>668400.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Github</td>\n",
       "      <td>0.75</td>\n",
       "      <td>820200.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Apple</td>\n",
       "      <td>0.75</td>\n",
       "      <td>1306800.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   vendor_cat  level_1         ia\n",
       "2    Facebook     0.75     8400.0\n",
       "6     Netflix     0.75     8400.0\n",
       "11    YouTube     0.75     8400.0\n",
       "8    Snapchat     0.75     9600.0\n",
       "5   Instagram     0.75    12000.0\n",
       "9     Twitter     0.75    12000.0\n",
       "10   Whatsapp     0.75    41400.0\n",
       "4       Gmail     0.75    98400.0\n",
       "1     FB Msgr     0.75   306000.0\n",
       "7       Skype     0.75   668400.0\n",
       "3      Github     0.75   820200.0\n",
       "0       Apple     0.75  1306800.0"
      ]
     },
     "execution_count": 297,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ia_count_df.groupby('vendor_cat')['ia'].quantile([0.75]).reset_index().sort_values('ia')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 242,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/sacheendra/miniconda3/envs/thesis/lib/python3.8/site-packages/plotnine/ggplot.py:727: PlotnineWarning: Saving 14 x 3 in image.\n",
      "/Users/sacheendra/miniconda3/envs/thesis/lib/python3.8/site-packages/plotnine/ggplot.py:730: PlotnineWarning: Filename: plots/failure_duration_ia_cdf.pdf\n",
      "/Users/sacheendra/miniconda3/envs/thesis/lib/python3.8/site-packages/plotnine/geoms/geom_path.py:81: PlotnineWarning: geom_path: Each group consist of only one observation. Do you need to adjust the group aesthetic?\n",
      "/Users/sacheendra/miniconda3/envs/thesis/lib/python3.8/site-packages/plotnine/geoms/geom_path.py:81: PlotnineWarning: geom_path: Each group consist of only one observation. Do you need to adjust the group aesthetic?\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABJYAAAF6CAYAAABLKdItAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAA9hAAAPYQGoP6dpAAEAAElEQVR4nOzddXRT2RYH4F/d3YUqVWgpFCsUdwaHwd0ZXIYHDD7CDDbAIINbB3fXwiClSHGpu7umluS+P3jNI1Ro06Q3Sfe3VhY0uTmSs5ve7JxzrgLDMAwIIYQQQgghhBBCCKkhRbYbQAghhBBCCCGEEEJkEyWWCCGEEEIIIYQQQohIKLFECCGEEEIIIYQQQkRCiSVCCCGEEEIIIYQQIhJKLBFCCCGEEEIIIYQQkVBiiRBCCCGEEEIIIYSIhBJLhBBCCCGEEEIIIUQklFgihBBCCCGEEEIIISKhxBIhhBBCCCGEEEIIEQkllgghhBBCCCGEEEKISCixRAghhBBCCCGEEEJEQoklQgghhBBCCCGEECISSiwRQgghhBBCCCGEEJFQYokQQgghhBBCCCGEiIQSS4QQQgghhBBCCCFEJJRYIoQQQgghhBBCCCEiocQSIYQQQgghhBBCCBEJJZYIIYQQQgghhBBCiEgosUQIIYQQQgghhBBCREKJJUIIIYQQQgghhBAiEkosEUIIIYQQQgghhBCRUGKJEEIIIYQQQgghhIiEEkuEEEIIIYQQQgghRCSUWCKEEEIIIYQQQgghIqHEEiGEEEIIIYQQQggRCSWWCCGEEEIIIYQQQohIKLFECCGEEEIIIYQQQkRCiSVCCCGEEEIIIYQQIhJKLBFCCCGEEEIIIYQQkVBiiRBCCCGEEEIIIYSIhBJLhBBCCCGEEEIIIUQklFgihBBCCCGEEEIIISKhxBIhhBBCCCGEEEIIEQkllgghhBBCCCGEEEKISCixRAghhBBCCCGEEEJEQoklQgghhBBCCCGEECISSiwRQgghhBBCCCGEEJFQYokQQgghhBBCCCGEiIQSS4QQQgghhBBCCCFEJJRYIoQQQmTcoUOHoKCgILipq6vD3NwcnTp1wrp165Camspq+3777TdcuHCh3P3379+HgoIC7t+/X+dtYkvZWL148aLGz+VwOFi9erXMvl6rV6+GgoKCROsYP3487OzsvnncsWPHsGXLlgofU1BQwOrVq8XaLkIIIUSeUWKJEEIIkRMHDx7EkydPcPv2bezYsQNeXl74448/4Obmhjt37rDWrsoSS82aNcOTJ0/QrFmzum+UDOJwOFizZo3MJpYmT56MJ0+esN0MAFUnlp48eYLJkyfXbYMIIYQQGabMdgMIIYQQIh6NGzdG8+bNBT8PHjwY8+fPh6+vLwYNGoSwsDCYmZnVqg4ejwculws1NbXaNhe6urpo3bp1rcshtVNaWgoFBQUoK4t+WlhVXHA4HGhqasLa2hrW1ta1aWqdoJgkhBBCaoZmLBFCCCFyzMbGBps2bUJeXh52794NAOjYsSM6duxY7tivlxFFR0dDQUEB69evxy+//AJ7e3uoqanh3r17KCoqwsKFC+Hl5QU9PT0YGhrCx8cHFy9eFCpTQUEBBQUFOHz4sGCpXlndlS2Fu3TpEnx8fKCpqQkdHR1069at3EyXsmVVHz58wIgRI6CnpwczMzNMnDgROTk5tX7d6tL48eOhra2N8PBw9O7dG9ra2mjQoAEWLlyI4uJiAJ/HwsTEBACwZs0awWs5fvx4QTlhYWEYOXIkTE1NoaamBjc3N+zYsUOorrLX/OjRo1i4cCGsrKygpqaG8PBwpKWl4YcffoC7uzu0tbVhamqKzp074+HDh0JlVBUXZePy8uVLDBkyBAYGBnB0dARQfincgAEDYGtrCz6fX+41adWqldBMth07dqB9+/YwNTWFlpYWPDw8sH79epSWltb49e7YsSOuXr2KmJgYoSWkZb5eCle2fNHf3x9TpkyBkZERdHV1MXbsWBQUFCA5ORlDhw6Fvr4+LCwssGjRonLtKikpwS+//AJXV1eoqanBxMQEEyZMQFpaWo3bTwghhEgbmrFECCGEyLnevXtDSUkJDx48EOn527Ztg7OzMzZu3AhdXV04OTmhuLgYmZmZWLRoEaysrFBSUoI7d+5g0KBBOHjwIMaOHQvg87Kizp07o1OnTlixYgWAzzOVKnPs2DGMGjUK3bt3x/Hjx1FcXIz169ejY8eOuHv3Lnx9fYWOHzx4MIYNG4ZJkybh3bt3WLp0KQDgwIEDIvWVLaWlpejXrx8mTZqEhQsX4sGDB/j555+hp6eHlStXwsLCAjdu3EDPnj0xadIkwVKtsmTTx48f0aZNG0Ei0dzcHDdv3sScOXOQnp6OVatWCdW3dOlS+Pj44O+//4aioiJMTU0FSY5Vq1bB3Nwc+fn5OH/+vOC1/zoZWVFcBAYGAgAGDRqE4cOHY/r06SgoKKiwzxMnTkT//v3h7++Prl27Cu4PDg7Gs2fPsG3bNsF9ERERGDlyJOzt7aGqqoo3b97g119/RXBwcI3HeufOnZg6dSoiIiJw/vz5aj9v8uTJGDRoEE6cOIFXr15h2bJl4HK5CAkJwaBBgzB16lTcuXMHf/zxBywtLbFgwQIAAJ/PR//+/fHw4UMsXrwYbdq0QUxMDFatWoWOHTvixYsX0NDQqFEfCCGEEKnCEEIIIUSmHTx4kAHAPH/+vNJjzMzMGDc3N4ZhGKZDhw5Mhw4dyh0zbtw4xtbWVvBzVFQUA4BxdHRkSkpKqmwDl8tlSktLmUmTJjFNmzYVekxLS4sZN25cuefcu3ePAcDcu3ePYRiG4fF4jKWlJePh4cHweDzBcXl5eYypqSnTpk0bwX2rVq1iADDr168XKvOHH35g1NXVGT6fX2V72VLRWI0bN44BwJw6dUro2N69ezMuLi6Cn9PS0hgAzKpVq8qV26NHD8ba2prJyckRun/WrFmMuro6k5mZyTDM/1/z9u3bf7OtZWPapUsXZuDAgYL7q4qLsnFZuXJlufLKHitTWlrKmJmZMSNHjhQ6bvHixYyqqiqTnp5eYbt4PB5TWlrKHDlyhFFSUhL0jWHKx3Blvvvuu0qP+/o1Lhuz2bNnCx03YMAABgCzefNmofu9vLyYZs2aCX4+fvw4A4A5e/as0HHPnz9nADA7d+78ZnsJIYQQaUYzlgghhBAxkfQVrxiGYeW5/fr1g4qKSrn7T58+jS1btuDNmzdCs1LU1dVFqickJASJiYmYN28eFBX/v1pfW1sbgwcPxu7duwX79XzZti95enqiqKgIqamp39xPSmnNM5HaWV28VS2rfayCggL69u0rdJ+npyf8/f2/+dyioiLcvXsXM2bMgKamJrhcruCx3r17Y/v27QgMDESvXr0E9w8ePLjCsv7++2/s2bMHHz9+FCzDAwBXV9dyx1YWF1WV/yVlZWWMHj0aO3bsQE5ODvT09MDj8XD06FH0798fRkZGgmNfvXqFVatW4fHjx8jMzBQqJzQ0FK1atfpmfbXVp08foZ/d3Nxw4cIFfPfdd+Xuv3XrluDnK1euQF9fH3379hUaGy8vL5ibm+P+/fuYMWOGZBtPCCGESBDtsUQIIYTIuYKCAmRkZMDS0lKk51tYWJS779y5cxg6dCisrKzg5+eHJ0+e4Pnz55g4cSKKiopEqicjI6PS+iwtLcHn85GVlSV0/5fJBwCCzaMLCwtFagNbNDU1yyXk1NTUqvVaZmRkgMvl4q+//oKKiorQrXfv3gCA9PR0oedU9Bpv3rwZM2bMQKtWrXD27FkEBgbi+fPn6NmzZ4WvZ0VlVOexL5XFy4kTJwAAN2/eRFJSEiZMmCA4JjY2Fu3atUNCQgK2bt2Khw8f4vnz54L9o+pqrA0NDYV+VlVVrfT+L8ctJSUF2dnZUFVVLTc+ycnJ5caGEEIIkTU0Y4kQQgiRc1evXgWPxxPskaOurl7hBteVfcCtaCaWn58f7O3tcfLkSaHHv5zlUlNlSaKkpKRyjyUmJkJRUREGBgYily+vDAwMoKSkhDFjxmDmzJkVHmNvby/0c2Vj2rFjR+zatUvo/ry8vArLrGqGXnVn77m7u6Nly5Y4ePAgpk2bhoMHD8LS0hLdu3cXHHPhwgUUFBTg3LlzsLW1Fdz/+vXratXBNmNjYxgZGeHGjRsVPq6jo1PHLSKEEELEixJLhBBCiJjUZrmZpMTGxmLRokXQ09PDtGnTAAB2dnY4ffo0iouLBTN8MjIyEBAQUOXG2l9SUFCAqqqqUAIhOTm53FXhgM8zb6ozq8TFxQVWVlY4duwYFi1aJCi7oKAAZ8+eFVwpTlxqslRNGlQ2G0tTUxOdOnXCq1ev4OnpKZhJU1MKCgqCOsq8ffsWT548QYMGDURrdDVMmDABM2bMwKNHj3D58mUsWLAASkpKQu0CINQ2hmGwd+9ekeusbkyKQ58+fXDixAnweLw6WbJHCCGE1DVaCkcIIYTIiffv3yMwMBCPHj3CuXPnMH/+fHh6eqKgoADnz58XXEFszJgxyMzMxOjRo3Hr1i0cP34cXbt2rXZSCfj8YTkkJAQ//PAD/P39cfjwYfj6+la4BMrDwwP379/H5cuX8eLFC4SEhFRYpqKiItavX4/Xr1+jT58+uHTpEk6fPo1OnTohOzsbv//+u2gvjJzQ0dGBra0tLl68iFu3buHFixeIjo4GAGzdulWwZOzQoUOC1/vPP/9E586dq1V+nz59cOvWLaxatQr+/v7YtWsXevToUW62k7iNGDECGhoaGDFiBIqLizF+/Hihx7t16wZVVVWMGDEC169fx/nz59GjR49yyyJrwsPDA6mpqdi1axeePXuGFy9e1LIXlRs+fDh69eqF3r17Y+3atbhx4wbu3r2Lw4cPY/z48TW6Mh0hhBAijWjGEiGEECInyvalUVVVhb6+Ptzc3PCf//wHkydPFiSVAKBt27Y4fPgwfv/9d/Tv3x8ODg5YtWoVrl27hvv371e7rtTUVPz99984cOAAHBwcsGTJEsTHx2PNmjVCx27duhUzZ87E8OHDweFw0KFDh0rrGTlyJLS0tLBu3ToMGzYMSkpKaN26Ne7du4c2bdqI9LrIk/379+PHH39Ev379UFxcjHHjxuHQoUNwd3fHy5cv8fPPP2P58uVITU2Fvr4+nJycBPssfctPP/0EDoeD/fv3Y/369XB3d8fff/+N8+fPVzsuRKGnp4eBAwfi2LFjaNu2LZydnYUed3V1xdmzZ7F8+XIMGjQIRkZGGDlyJBYsWCC0IXlNzJ07Fx8+fMCyZcuQk5MDhmEkNuNQSUkJly5dwtatW3H06FGsW7cOysrKsLa2RocOHeDh4SGRegkhhJC6osBI47x9QgghhBBCCCGEECL1aCkcIYQQQgghhBBCCBEJJZYIIYQQQgghhBBCiEgosUQIIYQQQgghhBBCREKJJUIIIYQQQgghhBAiEkosEUIIIYQQQgghhBCRUGKJEEIIIYQQQgghhIiEEkuEEEIIIYQQQgghRCSUWCKEEEIIIYQQQgghIqHEEiGEEEIIIYQQQggRiTLbDZB1sbGxSE9PZ7sZhNSYsbExbGxsKnyM4prIKoprIo8orok8orgm8ojimsijquK6DCWWaiE2NhZubm7gcDhsN4WQGtPU1MSnT5/KvUlQXBNZRnFN5BHFNZFHFNdEHlFcE3lUWVx/iRJLtZCeng4OhwM/Pz+4ubmx3RxCqu3Tp08YPXo00tPTy71BUFwTWUVxTeQRxTWRRxTXRB5RXBN5VFVcf4kSS2Lg5uaGZs2asd0MQsSK4prII4prIo8orok8orgm8ojimsgr2rybEEIIIYQQQgghhIiEEkuEEEIIIYQQQgghRCSUWCJitX//flhZWaFz5851VueQIUMwZMiQOquPyL+XL19i0qRJaNGiBezt7dGkSRP07dsXa9asERzTqlUrjB07lsVWElJ9Hz9+xMKFC9GmTRs4OjrC0dERbdu2xeLFi/HmzRvBcZs2bYKVlZXQcw8dOoSTJ0+WKzMgIABWVla4cuWK2NopiTJJ/VHdOK9LVlZW2LRpk+DnshgPCAhgpT2EfSdPnoSVlVWFt7Vr1wqOa9WqldBjDg4OaNu2LVavXo3MzMxv1lMWa1ZWVhW+hwPA999/DysrK7Rq1Ups/SPy58qVK7CyssLFixfLPda1a1dYWVnh/v375R5r06YNevToAeDze+FPP/0ktjYlJydj06ZNeP/+vdjKJLVDeywRsTpx4gQAICQkBC9fvqQ1xETm3LlzBxMmTICPjw+WL18OU1NTpKam4s2bN7h48SJWrVrFdhMJqZGjR49i+fLlcHR0xKRJk+Ds7AwFBQWEh4fjwoUL6N27Nx4/fgw7OzuMGDECHTt2FHr+kSNHYGhoiGHDhrHTAUKqoSZxXpcuXboECwuLOq2TyIbNmzejYcOGQveZm5sL/dyiRQusWLECAFBUVIS3b99i06ZNePr0Ka5fv16terS1tXHixIly7+GxsbF48uQJdHR0atELUh+0adMGCgoKCAgIQP/+/QX3Z2VlITg4GJqamggICBA6f0hMTERMTAymTp0qkTalpKRg8+bNsLa2RuPGjSVSB6kZSiwRsXnz5g0+fvyILl264O7duzhx4gQllojM2bVrF2xsbHDs2DEoK///LbJ///5Yvnw5iy0jpOaeP3+OZcuWoUuXLtizZw9UVVUFj/n6+mL8+PG4fPky1NXVAQCWlpawtLRkq7mEiKSmcV6XvL2967xOIhtcXV3RpEmTKo/R1dUViqG2bdsiPz8fW7ZsQUREBBwdHb9ZT79+/XDs2DFERkbCwcFBcP+JEydgbm4ONzc3hIaGit4RERQWFkJDQ6NO6ySiMzQ0hKurK548eSJ0f2BgIJSVlTFixIhyszDLfm7Tpk2dtZOwi5bCEbE5fvw4AGDZsmVo3rw5Ll68iMLCQsHjcXFxsLKyws6dO7F161a0aNECDg4O6NWrFx4+fChUVtlyjPfv32Py5MlwcXGBq6srZs+ejYyMjG+2paSkBFu2bEH79u1hb28PDw8PzJ8/v1rPJfVbVlYWDAwMhJJKZRQVq37LPHToEGxsbLBx40b8+eefsLGxQUJCQrnjFixYgEaNGqGoqAjA/5fVXb9+HV27doWDgwN8fHywf//+cs/Ny8vD2rVr0bp1a9jZ2cHb2xsrV64Eh8MRscdEnm3btg1KSkr4448/hD5sf6lv376Cb8m/XgrXqlUrhISE4MmTJ4IlFV8vmeByufj999/RrFkzuLi4YNiwYQgPDxc6plWrVpg3b165uitbylxcXIzVq1fDy8sLjo6OGDx4ME13J5WqaZzPmzcPTk5OCA8Px8iRI9GwYUM0bdoU27dvBwAEBQVhwIABaNiwIXx9fXHq1CmhsjIyMrB06VJ07NgRTk5O8PT0xPfff4+nT5+Wq/frpXCE1FbZDCMVFZVqHd+uXTtYWloKLYfj8/k4ffo0vv/+eygoKJR7zuXLl9GnTx+4urrC0dERPj4+WLBggdAxISEhGDFiBBwdHeHh4YFly5bhzp075ZZ6DhkyBJ07d0ZgYCD69esHR0fHcmUR6demTRtEREQgJSVFcN+TJ0/QpEkTdO7cGW/fvkV+fr7QY0pKSuXOGc6cOYMOHTrA0dERXbt2xe3bt4Uej4qKwvz589G2bVs4OjrC29sb48aNw6dPnwTHBAQEoHfv3gA+n1OXnZ+UvdfGxMRgxowZaNasmWBLi6FDhwqdR1y8eBEjRoxA06ZN4ejoiA4dOuC3334rdz5d9vciJCQEQ4cORcOGDeHh4YGffvpJ6HMu8P/lfkePHoWvry/s7e3RsWPHCpcQyiNKLBGxKCwsxMWLF+Hl5QVXV1cMHz4c+fn5uHz5crljDx48iPv372PNmjXYtm0bFBUVMWbMGLx48aLcsZMmTYKdnR327NmDBQsW4MaNGxg5ciRKS0srbQufz8fEiROxY8cODBw4EIcPH8ayZcvw4MEDDBkypNybACFf8vb2xqtXr7BixQq8fPmyylgrwzAM1q5di9WrV2PDhg1YtGgRxowZA2VlZfj5+Qkdm5WVJfhj9uW35x8+fMCqVaswZcoU7Nu3D82bN8fKlSvx999/C44pLCzE4MGDcfr0aUycOBFHjx7FDz/8gNOnT2P8+PFgGEZ8LwSReTweDwEBAfD09ISZmZlIZezfvx+2trZo3LgxLl26hEuXLpVLeP7+++9ISEjAhg0bsH79ekRFRWH8+PHg8Xgit/33339HbGwsNmzYgA0bNiA5ORlDhgxBTEyMyGUS+SRqnHO5XEyePBldunTBgQMH0KlTJ6xbtw7r1q3DvHnzMHz4cOzfvx+Ojo6YP38+3r59K3huVlYWAGD+/Pk4cuQINm/eDFtbWwwZMoT2TiLVxuPxwOVyhW5fYxhG8FhBQQEeP36Mffv2oUWLFrCxsalWPYqKihg6dCjOnDkjeF/+999/kZSUVOES5xcvXmDGjBmwsbHBzp07ceTIEcybN0+ofSkpKRg8eDAiIiKwbt06bN26Ffn5+ZXO7E5JScHs2bMxcOBAHD16FOPGjatW24n0KJt59OWspYCAALRu3RotWrSAgoKCUHI9ICAAHh4e0NXVFdx39+5dHDp0CIsWLcLevXuhr6+PyZMnC/1tT0lJgYGBAZYtW4Z//vkHv/76K5SVldGnTx/Bl1YeHh7YvHkzAGDu3LmC85MRI0YAAMaMGYO3b9/ip59+wvHjx7Fu3To0btwYubm5gnqioqLQuXNnbNy4EX5+fpg8eTIuX75cYWxyuVyMGTMGvr6+OHDgACZMmAA/Pz9Mnz693LG3bt3CgQMH8OOPP2L37t2wtrbGDz/8UC/2jqSlcEQsrl69itzcXAwfPhzA52m3q1atwokTJzB06FChY/l8Po4fPy74UN2xY0e0atUKGzduFOzRVKZXr16CP1IdOnSAiYkJZs2ahcuXL2PQoEEVtuXy5cu4d+8e9u7dK8hmA4C7uzt69+6NU6dO0R80UqmlS5ciPDwcBw4cwIEDB6CiooImTZqgW7dumDBhArS0tISOLywsxJw5c/Do0SMcPXoU7dq1AwAYGxsLpp/Pnz9f8C368ePHUVJSUi4GU1JScPPmTTRq1AgA0LlzZ6Snp2PLli0YN24cNDQ0sH//fnz69AlXrlwRTJ9v164dzM3NMXXqVNy7d69ON84n0i0zMxNFRUWwtrYu9xiPxxNKRCopKVX4rXXjxo2hrq4OHR2dSpf0ODk54a+//hL8rKioiOnTp+P169ciLwMyMjLC/v37BW1q2bIlfH19sX37dmzYsEGkMol8EjXOS0pKsHjxYsF5Qps2bXDnzh1s374dN2/eFOzZ4enpiSZNmuD8+fPw9PQEADRs2BDr1q0Tqqdjx46Ii4vDgQMHaOkHqZa+ffuWuy8mJkZoxrS/vz9sbW2FjmnatCn27NlTo7qGDRuGrVu34t69e+jatStOnDghmPn8tRcvXoBhGPz+++9CSYEvk1B79+5FdnY2zp07B2dnZwCfz1tGjRqFuLi4cmVmZ2dj9+7d8PX1rVG7ifRo3bo1FBUVERAQgAEDBiAzMxPBwcFYvnw5tLS04OHhgYCAAHTp0gUJCQmIjY1Fnz59hMooKirCiRMnoK2tDeBzgqhZs2a4fPkyZs2aJaindevWgufweDx06dIFnTp1gp+fH1avXg0dHR24uroCAGxtbYXONTIzMxEREYE1a9Zg8ODBgvu//EwIQGgWNcMwaNGiBZycnDB48GB8/PgR7u7ugsdLSkowbdo0TJo0CQDQvn17KCsr448//sDz58/RokULofqvXbsGExMTAECXLl3QuXNnrFu3rtzrIW9oxhIRixMnTkBdXV2woZuWlhb69OmDp0+fIjIyUujYXr16Cc3U0NbWRrdu3fD06dNy33B/nTzq27cvlJWVq/xG8M6dO9DT00O3bt2EvgVq1KgRTE1Ny60PJuRLhoaGOH/+PK5du4Zly5ahe/fuiIyMxLp169ClSxehK7FkZWVh6NCheP36Nc6fPy9IKpWZPHky0tPTBd9S8Pl8HDlyBF26dEGDBg2EjnVxcREklcoMGDAAeXl5ePfuHYDPsV123Jex3bFjRygoKFBsk2rr2bMnbG1tBbfdu3eLXFb37t2FfnZzcwMAxMfHi1zmgAEDhBJd1tbWaN68Oc0GITVSVZwrKCgIJeKVlZVhZ2cHMzMzoY1gDQwMYGxsXC6ejxw5gh49esDBwQE2NjawtbXFo0ePEBYWJvmOEbmwdetWXLt2Tej29TL8li1bCh67cOECNm3ahIyMDAwdOrRaV4YrY2NjAx8fH5w8eRKZmZm4deuW4Mvgr3l5eQEApk+fjkuXLiEpKancMYGBgXB1dRUklcp8ubHzl/T19SmpJOP09fXh7u4uONcMDAyEkpKSIKnSunVrwd/oyvZXatOmjSCpBAAmJibl3l+5XC62bduGjh07ws7ODjY2NrCzs0NUVFS5ZfYVMTAwgJ2dHXbt2oXdu3fj/fv34PP55Y6LiYnBzJkz4eXlhQYNGsDW1laQiKrofXzgwIEV/vz48WOh+319fQVJJeDzFxr9+vVDdHQ0EhMTv9l+WUaJJVJrUVFRCAwMRJcuXcAwDHJycpCTk4PvvvsOAMpd4vTLX7YypqamKCkpQUFBQZXHKisrw8DAQDANvSJpaWnIycmBnZ2d0Amlra0tUlNTa/SHmNRfTZo0wcyZM7Fnzx68fPkSU6ZMQVxcHHbu3Ck4JjIyEq9evUKnTp0E35x8qXHjxmjVqhUOHToE4HNiKC4uDhMmTCh3bGW/F8D/l12kpaXh06dP5eLa2dkZDMNQbBMhhoaGUFdXrzDBs2PHDly7dg0HDx6sdT0GBgZCP6upqQGAYA8xUZTF/pdMTEyqfO8n9ZOoca6hoVFuM29VVVXo6+uXO1ZFRQXFxcWCn3fv3o2lS5cKZo5cvnwZ165dQ6dOnWoV96R+cXJyQpMmTYRuX9PR0RE81qJFCwwfPhzbt29HWFhYjb8QGDFiBG7fvo09e/ZAXV1dcJ7+tdatW+PAgQPgcrmYN28emjdvjs6dO+PChQuCY7KysmBsbFzuuRWdywAVv6cT2dOmTRtERkYiOTlZsNStbCZ/69at8f79e+Tm5iIgIADKyspo2bKl0PO/Pl8APr/vfvm+uWbNGmzYsAE9evTAoUOHcOXKFVy7dg3u7u7Ven9VUFDAyZMn0bFjR+zatQs9evSAh4cHVqxYIdgDqqCgAIMGDcKrV6+wePFinDlzBteuXcO+ffsAlD9/UVZWhqGhodB9ZbH+9XlJZecvFR0rb2gpHKm1EydOgGEYXL16FVevXi33+OnTp7F48WLBz2lpaeWOSU1NhaqqarllRmlpaUKX6eVyuYLNlStjaGgIAwMD/PPPPxU+/nUdhHyLiooKFixYgL179yIkJERwv7e3N/r06YNFixYB+LwvzNcbfE+cOBHTpk3Du3fvcPDgQTg4OKB9+/bl6qjs9wL4/x9iQ0NDaGhoVLoR7Nd/9Ej9pqSkhLZt2+LBgwdISUkR2n+m7FvmipYsSIKamhpKSkrK3Z+ZmVlh3JbF/pfS0tKqfO8n9RMbcX7u3Dn4+Pjg999/F7r/y41rCZGUsiU6Hz9+rNHzevXqhZ9++gk7duzAyJEjq7wqW48ePdCjRw8UFxfj5cuX2L59O2bOnCmYPWpgYID09PRyz6vovRtAhUutiexp06YN9uzZgydPnuDJkydCsz7LkkiBgYGCTb1F+cx17tw5DBkyBEuXLhW6PzMzE3p6etUqw9raWnCuHBERgcuXL2Pz5s0oKSnBH3/8gcePHyM5ORlnzpyBj4+P4Hk5OTkVlsflcsudr5Sdt399XlLZ+UtFx8obmrFEaoXH4+HMmTOws7PD6dOny92mTZuGlJQU+Pv7C55z/fp1oUxwfn4+bt++jZYtW0JJSUmo/HPnzgn9fPnyZXC5XKE3ga917doVWVlZ4PF45b4JatKkCRo2bCim3hN59OXVLr5UNi32681hhw4dip07d+LUqVOYO3duueWcvXr1gpWVFdasWYOHDx9i3LhxFZ5ghYSE4MOHD0L3XbhwAdra2vDw8ADwObajo6NhYGBQYWx/vbyOkNmzZ4PH42HJkiXV2oi+Il9/myiKBg0aCF3RBfh8shcREVHh8RcvXhTaGyc+Ph4vXryo8r2f1F/iiPOaUFBQEMzMK/Px40cEBQVJvG5Cyq5sZWRkVKPnaWhoYN68eejWrRvGjh1breeoqanBx8cHy5YtE6q7devWCA4ORmhoqNDxly5dqlGbiGxp3bo1lJSUcPXqVYSEhAj9TdbV1UWjRo1w+vRpxMXF1Wqvua+v7nnnzh0kJydXeMy3zk8cHR0xb948uLq6lru67Nf1fH3BnS+dP3++wp+/7uejR4+Evizm8Xi4dOkS7OzsYGlpWWVbZR3NWCK14u/vj+TkZPz0008VvoG4urri0KFDOHHiBFavXg3g88auI0aMwNSpU8Hn87Fz507k5+cLZn186fr161BWVkb79u0REhKCDRs2wN3dvcIND8v0798f58+fx5gxYzB58mR4eXlBWVkZSUlJCAgIQI8ePdCrVy+xvQZEvowaNQoWFhbo2rUrGjZsCD6fjw8fPmDPnj3Q0tISbNz3pT59+kBDQwNTp05FUVERduzYIfhjpaSkhPHjx+PXX3+FpqZmuc3sy5iZmWHChAlYuHAhTE1Nce7cOTx48AA//fST4FvFKVOm4Nq1axg0aBCmTJkCNzc3MAyDhIQE/Pvvv5g2bRqaNWsmuReHyJwWLVrg119/xYoVK9CzZ0+MGjUKzs7OUFRURGpqqmCW6Zd7HnzN1dUVly5dwsWLF2Fraws1NTXBPkrVNXjwYMyePRtLly5F7969kZCQgJ07d1b6wSg9PR2TJk3CyJEjkZeXh40bN0JNTU2wuSchXxJHnNdE165dsWXLFmzcuBGtW7dGREQEtmzZAhsbmwqv7EWIqHJzcwUJSy6Xi7CwMPz1119QU1PD+PHja1zetGnTMG3atCqP2bBhA5KSkuDr6wsLCwvk5uZi3759UFFRESQSJk+ejBMnTmD06NFYtGgRTExMcP78ecEeOF/P3ibyQUdHBx4eHrhx4wYUFRWFNq0GPieeypaTiZpY6tq1K06fPo2GDRvCzc0N7969w65du4RWsACAnZ0d1NXVcf78eTg5OUFLSwtmZmbIzMzE8uXL0adPH9jb20NFRQWPHz/Gp0+fMHPmTABA8+bNoa+vjyVLlmDBggVQVlbG+fPnK50FqKqqit27d6OgoABeXl548eIFtm7dis6dO5db7mdoaIihQ4di3rx50NDQwJEjRxAeHi60lYa8osQSqZUTJ05AVVW1wsuVAp9/uXr27ImrV68KfpknTJiA4uJirFixAhkZGXB2dsaRI0fKvTkBwL59+7Bp0yYcOXIECgoK6NatG9asWVMuw/wlJSUlHDx4EPv27cPZs2exfft2KCkpwcLCAq1bt65wLxxCysyZMwe3bt3C3r17kZqaipKSEpiamsLX1xezZ8+Gk5NThc/r0qULjh49ivHjx2PixInYu3evICHUr18//Prrrxg8eLDQFVa+1KhRIwwbNgybNm1CVFQUzMzMsGrVKkydOlVwjKamJs6fP4/t27fjn3/+QVxcHNTV1WFlZQVfX1+asUQqNHbsWHh7e2P//v3Yu3cvUlJSoKCgAHNzczRv3hwnT56sclPVRYsWITU1FYsXL0Z+fj6sra2FLilcHQMHDkRKSgqOHj2KU6dOwcXFBevWrcOff/5Z4fFLlizBmzdvsGDBAuTn58PLywu7du2q8ApGhAC1j/OamDNnDgoLC3H8+HHs2rULTk5O+P3333H9+nW6iAIRq+fPn6Nfv34APp/fmpubw8vLC3PmzBHaZF6cmjZtirdv3+LXX39FZmYmdHV14enpKXjvBgBzc3OcPXsWq1atwtKlS6Guro5evXph0aJFmDdvXqXnOkT2tWnTBq9fv0bjxo2ho6Mj9JiPjw/27t0LVVXVCj/XVcfatWuhoqKC7du3o6CgAB4eHti3bx/Wr18vdJyGhgY2b96MzZs3Y+TIkSgtLcWCBQswbtw42Nra4vDhw0hMTISCggJsbGywcuVKTJw4EcDnz6eHDx/G2rVrMXv2bGhqaqJHjx74+++/0aNHj3JtUlZWxuHDh7FixQps3boV6urqGDlyJFasWFHu2O7du8PZ2Rnr169HQkICbG1tsX379ko3tpcrDBFZUFAQA4AJCgpiuykyITY2lrG0tGR27dr1zWM3btzIWFpaMhkZGXXQsvqnqtiluBa//fv3M5aWlkxwcHCFj7ds2ZIZM2ZMHbdK/lBcE3lEcU3kEcW1fPrxxx8ZJycnpri4mO2msILiWv7MnTuXadiwYbWOtbS0ZJYtWybhFtW96sYuzVgihBAJef/+PWJjY/Hnn3+iR48egm/6CCGEEEJk2Z9//gkzMzPY2NiAw+Hgzp07OHbsGObOnVvlygJCiHyixBIhhEjIpEmTkJaWhpYtW5a7ehAhhBBCiKxSVlbGrl27kJSUBB6PB3t7e6xatQqTJ09mu2mEEBZQYonUmQYNGiAhIaFaxy5cuBALFy6UcIsIkazq7kNT0/1qCCGEEELYNHv2bMyePZvtZhAiUVu2bMGWLVuqdWx1P+fKK9qynxBCCCGEEEIIIYSIpN7NWOJwODh58iSioqIQGRmJ3NxcDB8+HCNHjmS7aYQQQgghhBBCCCEypd7NWMrLy8OtW7dQWlqK1q1bs90cQgghhBBCCCGEEJlV72YsmZqa4tixY1BQUEBubi5u3bolclmFhYUAgNevX4updYTUjU+fPlX6GMU1kVUU10QeUVwTeURxTeQRxTWRR1XF9ZfqXWJJQUFBbGVlZ2cD+HzlJ0JkjaamJoyNjcvdT3FNZBnFNZFHFNdEHlFcE3lEcU3kUWVx/aV6l1gSJwsLCwCAn58f3NzcBPfHx8fD2tparHVlZWVh48aNWLRoEQwMDGpVVkXtE2f5NfV13eJ+/cTdN0mMLxuMjY1hY2NT7n6Ka/GguGYHxbVkUVyzg+Jasiiu2UFxLVkU1+yguJYsimt2VBbXX1JgGIapo/ZIndzcXIwePVrkzbtfvnwJb29vBAUFoVmzZoL7IyMj4eDgIM6mihW1r3akvX21RXEtGdQ+dlFcSwa1j10U15JB7WMXxbVkUPvYRXEtGdQ+6VHvNu+uraSkJLx8+RIvX76s9npDccjPz8f+/fuRn59f67J4PJ5Ey68pSdfNZt/YUljKZ7sJ1cL22Kenp0NSuXW2+yYrirh8JOSWsN0Msarp2JSWlkq0fHGiuP4ss5ArsfcOaVXXY88wDIqKipCWlobIyMga/57Upm5ZK7+6+AyDlHzJvY6ySJrGvqioCCkpKYiNja3zuqWx/G/hMwyyCrms1C3tpGHsGYYBj8dDcXExOBwOcnNzkZmZiZSUFMTHxyMqKgohISEoKioSe921wXZcf41hGJTy+Cji8lFQwkNOEReZhVykFpQiKa8E8bkliM4uRkRmEULSC1n9naClcDW0e/durFmzps7rzczMxOTJk9GtWzdoa2uLVAafz8fRo0excuVK3Lt3Tyh7Ko7yRSXputnsW12KyirGpZAsXArJwruUQsQv9IKqknTnjutq7F1cXJCbm4vg4GB8+vRJ8G9GRgZSU1NhYmIisbrra1xnF3ERnV2M5PxSJOeXIqXs34L//z85vxTZRTyoKSmg4KfmYt0Dj03VHZuIiAhs374dx48fx/v377+5dr2m5UtCfYvr3GIePqRy8D61EB/TCvEhtRAf0gqRnF+K+AVesNBRZbuJdaY6Y1NYWIi0tDRkZWUhMzMTsbGx+PjxI4KDg5GXlwcul4vS0lLk5+dDSUkJpaWlgvsKCwuRnJyMRYsWgc/ng8PhgMv9/wlyeHg4HB0dWeubNJf/JS6fQQmPj7QCLs58zER0djFSC7iIyS7Gu1QOirkMOMubS/35QV2R1NgUFBTg2bNn+PDhA2bPno2QkBDw+XxkZWUhKysL2dnZyMnJEdxyc3NRUvL5SxYnJyeEhobWug3yFNdleHwG+1+lYV9QKj6lF8HHWhu3xrrWSd2ypDpjwzAM8vPzBTGZl5cHDoeDgoICpKenIzk5GZmZmUhISADweaVPWdxyOBzExsZi2bJlAAAulyt04/F4FU5kqMiLFy/g7e0t1r7VRk3LL+XxkVfCR34JD/klfOQV85BfwkNeCR/pnFIEx3HADYlBSkEpUgtKkVbARRGXDx7DgMv/HNOf/8+Ax//8Hs5j/v//mn6FtbWnDWa1Mhet87VEiaUamjZtGvr16wfg8w7po0ePZrlF1XPv3j0sXLgQoaGhmDp1KszN2Qk4Ij4MwyAoqQCXgrM/J5NSC9HQUA39XAywuqM1lBXl40N6ZfLz8xEfH4/k5GQkJSXh/fv3KCkpQVJSkuC+sj+G7dq1g6amJlxcXODm5oYePXpg7ty5cHNzq/O14fKOYRj8/SIVi27FoojLQENZERY6KjDXVoGZtgrMtVTgZqzx+f9f3OoTLpeLKVOm4PDhw2jRogU2b94MXV1dtptFvrDmfjwOvkpHXG4JFAA4GqrB3UQDLa20MKGpCRqZasBEq37FbZmyD8zv3r3Du3fvEBERgbi4OMTFxSEzM1NwnLKyMqysrODu7g5XV1fo6+tDWVkZKioqyM3NhZmZGVRUVKCiogJlZWXk5ORgzpw5WL9+PSwtLaGpqQkdHR3BzdTUlMVeSx7DMCgo5SO3mIfsIh6eJ+QjOb8UOUU85BT/71bERW4xD2m5hShVeIMSHoNiLv/zv7zP//K/+BTibKSOJuaaMNNShreFJtZ0soKjobrcnx+wITMzE//++y8OHjyIBw8eIDAwEAzDCM4x7t+/DwsLCxgaGsLOzg76+vrQ09ODnp4edHV1hf6vr6/PbmekzMOYPNyJzMHHtEI8TyhAGqcUP7axwOpO2vAw1WC7eazi8XjIzs5GRkaG4Jaeno7w8HAAwC+//AIlJSVBwigrKwvp6elIS0tDenp6uZmgmpqagk2azczMYGhoCEVFRVhZWcHZ2VkQp/n5+Vi4cCFWrlwJCwsLKCsrQ1lZGUpKSuX+//W/Ze/7ZTdDQ0M2Xrpvyi7iIv9/XyQl55cIvhRNyv+cHMoq5CI8swhZReUTaKpKCtBWVYSRhjL0lHmwMSqBmZYK3E00YKypDA0VRSgpKEBZUQFKivj8r4Lw/8seq+i48o8pQPl/P+uoKbHwan1GiaUasrCwEGy+Ju0YhsHdu3exZcsWXL9+HZMmTcK1a9fA4XCgqanJdvOICJLzS/BvdB7uR+fiamg2EvJK0cpKC8MbG+GYqwHcjNXlZtZHGR6Ph/DwcLx9+1Zwe/fuHaKiogAASkpKgj9+dnZ2sLCwQNu2bWFubg5lZWVMnToVT548QcuWLaGoSN/QStrKewnYEpiMrb1sMayREbRVFeUuJmuDz+dj4sSJuHnzJh4+fIi2bduy3STyhVIeH0ffZuDXB4nY2ssWray14WqsDk0V9k7U2JKeno6QkBAEBwcjODgYr169AgC4u7sDABwcHNC4cWO4urqiW7dusLa2RoMGDQTvx1paWpX+7le050RsbCzmzJmDHj16fHODUHlQWMrH+1QOjr3LQEBcPt4kc1D6RVbIRFMZTkbq0FVTgp66EvTUlGCtowk9dSWU5jOwsTCFmpIi1JQVoKakCFUlBagpK0BVSRFqSgrQVFGEs5H8nROwjcPhICwsDKGhoQgODsbbt2/x6tUrREREQE1NDS1btkT79u2xYsUKtG3bFmlpabC1tcWZM2fqRVyLC59hcC8qFyffZ2L/qzTY6qmij7M+lrW3xGA3Axhp1q/EPo/HQ1hYGN68eSO4vX37FgkJCULLsnV1dWFkZCSYaRMVFQUTExNoaWnBwMAAjRo1gomJieBmaGgIAwMDGBgYQFtbu8Lz5MrerxcuXIi+ffvKRFwzDIOYnBIk5JaAU8oHp5SPQu7nf0PTixCWWYScIi6yi3hITogDAHjteg8F/XSoKCoIvgQt+1LU20ITRprKMNBQhrGmMhwM1KGtqghtVSVoqyoKzQitT3ss1cvEUlBQEIqKilBYWAgAiIuLw+PHjwEAzZs3h5qaGpvNq7WcnBwcPnwYO3fuREREBAYPHoxXr17B09MTwOcAJ7IhJb8U/8bk4n5UHv6NyUVwehEMNZTQzlYXyztYoa+zvlwvxfjw4QMGDRqE0NBQGBkZwdPTE56enujXrx88PDxgZ2cHY2NjKCoqVvqHDwAsLS0pqSRGDMMgpaAUIelFCMn4vKY7JKMIoelFiMwqxpVRzujZUJ/tZkodhmEwa9YsXL58Gffv30eTJk3YbhIBUMLj405kLs5+zMTF4CwUlPKxrmsDzGhhxnbTJK6goADh4eEICwsT3Mo+MGdkZEBRURH29vZwcXGBvb09AODChQvo0qWLVCxVlFXH3qVj2uVocEr5MNZUxtxW5vilszWMNZWhp64EXTUlGKgrQ6mSmUWRkUVwcBD/Em4iLDExEQEBAQgICMD79+8REhIiOK8wMTGBs7MzPD09sXjxYnh7e0NbWxsuLi4st1p2MQyDf6NzcSkkC2c+ZiElvxS9nPSwr589xjYxrvT3Qd7k5OTg7du3Qkmk9+/fo7CwEMbGxmjSpAmaNGmCYcOGwd7eHkZGRjAyMoKhoSFUVD4n3GJjY2Fra4v9+/fLROJHEnKLeXidXID7UXk49DoNMTmfl5qqK39OvGsoK0JTRRE6akpoYakNL3NN6KsrgW/HxTwAt8a6opmrAwzUlShBX031MrG0a9cupKamCn5+/PixILG0d+9emJlJ38mkubk53r9/X+UStk+fPmHr1q3w8/ODnp4epk+fjilTplRr2Vt1ypcUSdfNZt9qKoNTCv+ozzOS/o3Oxaf0IhioK6G9nS6meZuig50uPMw0oCgnb3BVjc2FCxcwZswY9OrVC/7+/rC0tKzxGzvFtXhlcEqx7WkK/n6RinQOF0oKgKOhOpyN1OFurIEBrgZobqkFL3OtOmuTNKpobBiGwZIlS3DkyBHcuXOnVkklimvxYRgGrfd+REhGIXo21Me23rbo42wAXRankotbUVERIiIiyiWPwsLCkJiYCODzZYSdnJzg5OSEXr16Yf78+XB1dUXDhg0FX7aVlJRg3rx5cHJygqqq+L/QqA9xbWZmhlFnw3HqQyaWtbPEAh9z6KnXy1NxqWFubo63b9+isLAQFy9eFCzxfPbsGaKjo6Gvrw8fHx+0aNECo0aNgouLC5ydnStcvlPRF7f1Ia7FUX5sTjFm/puHewmZaG+rgyW+FhjiblivlhyHhYWhe/fuiI6OhpKSElxdXeHp6YnBgwdj7dq1aNKkCczNzat1LixLYy+uugtL+bgXnYsPqYW4E5mD+9F5YBgGrsYamN3KDANcDWGrr/rNz1AlJYbo+v49nJzsoKpK7881US9frX379rHdhBpTVVVFo0aNKnyMw+Fg7dq12LRpE9q2bYtDhw6hf//+gqx1bcuXNEnXzWbfauLYu3TMvBoDRQWgva0OpniboqOcJZK+VtHY8Pl8rF27Fj///DN++eUXLFmyRORvCiiuxScsowjN97yHhbYKfulsjXa2OnAwUKMNYCvw9djk5ORg5cqV2L17N65fv47WrVuLtfy6JG9x/SGtEG9SOAib4wkHA/U6q7euBAcHw93dXbDXS1nyqFOnTpg6darg5+rs6SJvY1+XdauqqiJB3RrD9oUgLrcEL6Y2RhNz2pKADYWFhXj//j1ev34tuL19+xb5+fnQ0dGBh4cHPDw8sGzZMrRt2xaurq61mvEs73Fd2/IT80pw5kMmfn+UBGM1PoKmNUJj0/r5u2FlZYWVK1eiSZMmcHd3h7q66H+TZGHsxVF3XE4xjoUWIvxtBPwjc5HO4cLDTBNe5pq4OcYSbRpo1/g8VVY+N0oj+kQgIxITE9GwYUPBt4tlbt++DQ8PDxw7dgznzp3D/fv3MWTIkBollaoqvy5Ium42+1YducU8jD0fgYkXorCivSVSf2yG88OdMbe1OZqYa8ptUgkQHpvS0lKcOnUK7dq1w59//olLly5h6dKltZp+SnEtHnnFPGx+kgRNFUV8mOmJKd6mcDXWoKRSJcrGJjo6Gtu2bYOjoyOuXLmCq1evolOnTmIrn+K6doq5fEy9FIVujrpymVQCPu+DFBAQgPT0dGRmZuLp06fw8/PDqlWrMHLkSLRo0aLaGwXL09jXZd0Mw8Dv0Sf0btUYzmr5+DjTk5JKdai4uBgHDx7E6NGj0ahRI2hra6Nly5b4448/kJKSAl9fX2hrayMwMBA5OTl4/Pgx/v77b0yZMgXu7u61XkYvr3EtjvKX342DzebX2Po0GROaGuNoN716m1QCPm+cPWHCBDRr1qxWSSVA+sdeHHVfCwpD8z0fsPdDIRgGWNrOEgkLm+LplEbY3dceHe10RTpPlfbPjdKsXs5YkkVcLhcRERGCS+++e/cOa9euxblz5zBnzhz8/PPPtdrv4Ovy65Kk62azb1VhGAbXw3Mw51oMVJUUEDDZHc0s6tcSorKx2bBhA06ePIm8vDyMGTMGR48eFctGdxTXtVPM5WNvUCp+fZgIdWVF7OpjX2/2OKiN0tJSREREoEOHDuBwOFi1ahWmTZsmtuVDFNe1dz86F7OvxSC3mIcLI5wlWhebVFVVaz1Droy8jH1d1z3vRix23w0HLyMev3eyhJWu/O6LKE0KCgqwb98+bNiwARwOBwMHDsSMGTPg5eUFT09PwZU4Y2NjsX79elhYWEhkHxV5jevaln8zPBvrHiXh4AAHjPE0goKCAu0BK0bSPPa1VVhcgoiICPT75xMmdvTAj+58ODV0FFv50vq5URZQYknGvHv3DnPnzsXFixfRs2dPPHv2DN7e3mw3i9RQQFwelt2Jx5P4fMxobopfu1hDS1V+9vSoSGlpKaKiohAaGiq4vX79GgBw5coV/PTTTxgzZgxddl0KRGcX4+T7DOwJSkV+CR9LfS0xvYUp1JVphtK3MAyDmTNnAgD69u2LX3/9FXp6eiy3inxp6Z04bAxIwpRmpvilizUMNehUiIjf49g8zLgSjeD0Qhwd6IDhmwB1FXoPlbTs7Gzs2LEDW7ZsgbKyMhYtWoSpU6dCR0eH7aaR/2EYBjOuRGN2SzNBUomQb+GU8nA5JBu/XwkHAGzpZYtZ3ewpISlF6GxKRrx//x4A0KdPH/Tv3x/Pnj1D8+bNWW4Vqan3qRws94/HlZBsjPQ0woEBHnK7BKO4uBg3b97EyZMn8fz5c0RGRoLH40FHRwfOzs5wcnJC69at8fTpU9y5cwe2trZsN7neexKXh0W3YhEYXwBXY3VMbmaKmS3N5GojY0nbsWMH7t69CwBYvHgxJZWkzL6XqdgSmIybY1zR2Z6S2EQy3qZw0NMvBP1c9HF+uBNU8lO//SRSK6mpqdiyZQt27NgBQ0ND/Pzzzxg/fnytlxQR8YrPLUG/Y6GIzSnB7FZmlFQi1XL0TTqW+8cjt5iHbsbqeA2gn4sB280iX6HEkpTjcrn47bffsHbtWlhYWODYsWPo2LGj2OvR0NBAv379oKGhIfay2a6bzb59ad/LVPxwJRo9nfTxcnpjeJrJ3zpyLpeL+/fv4/jx4zh37hxKSkrQr18//Pjjj4IrqZiZ/f9EIi0tDVFRUdDUlMxrQXFdPQzD4FJUEX4KDMbwxkbY1cceHqYadMJXQ69fv8bChQuxYcMG3L17VybGXtrqlnT56x8lYU1HK0oqiUDWx74u6i7h8fHnk2SsuZ+Awe6GODLQAQoKCkjjScd5iDyKjY3Fxo0bsXfvXjg4OGDHjh0YPnw4lJWr9xGH4rpuys8t5mHu9Rj4vU2Hk6E6wuc2gZ2+mkTaRaRr7GvrZVIBZlyJxqRmJljTyQqleVkoviwffZM7DBFZUFAQA4AJCgoSuj8iIkIs5YeEhDCtWrVijI2NmXPnzomlTIYRX/skRd7ax+fzmZX+cYzq2mfMgZepEmqV+NQ0rnk8HvP48WNm1qxZjKmpKaOiosL069ePOX78OJOfn18XTa6yfdJCWtuXySllvj8VyqitfcpseZLE8Pl8tpskEZJ+v87Ly2OcnZ2Z0aNHi/U1lNa4KSNL7Tv+Lp1RXP2U+ZDKYbFF4iXpuJYUeWzf2Y8ZjMOWV4zJH0HMX4FJDJcnufdSaX/9aqs6cR0SEsJMmDCBUVZWZpo3b86cP3+e4fF4dd1UIdI+Lmy1j8/nM8NPhzHO214zdyNzGF4lfyOl/fWrLXq/rrn8Yi5jveklM/JMWKXvqfT6SQ9a7C2FGIbBrl270LRpU5iYmODdu3fo3r07Tp06hYKCAonUWVBQINHy2aybzb7x+AwmX4rClsBkXB7pjAlNTeq8DZLAMAxevXqF//znP7C3t0e7du0QHByMdevWISUlBRcvXsTw4cOhpVX1ZuTyPPay0Lf3qRx4736PsIwiXOitj7mtzWmWkohmzpwJhmGwc+dOcDgcqR97aa1bUuWfeJ+BsecisK2XLdxN6FtIUcjq2Euy7iIuHwFxeRh5NhwjzkRglIcxwuc2waxW5kIXOmCzb/Lo2bNncHV1RVRUFK5du4Znz55hwIABIl3BjeJasuUfe5eOxjvf4dynLJwY0hCd7XXl+mrH0kIaxr62EnJLMPhUGBQVFLCvn4PgPVUe+iavKLEkZRITE9G7d2/8+OOP2LJlCy5dugRzc3NkZGRg2LBhyMjIkEi9ki6fzbrZ6hufYTD9SjQuhWTh/ng3dHeUj71WPn36BDc3NzRr1gyPHj3Cjz/+iISEBNy+fRsTJ06EgUH11zzL69jXRd21LZ9hGEy/HA0PM00ETGoEZwNaGS2qI0eO4MSJEzh58iR0dHSkfuyluW5JlX/wVRp+aGGGmS3NxFpufSKrYy+puhPzSuC2/S3aHfiE4LQi3BvvirWdrSvck47Nvsmj5s2b4+nTp7h37x66detWqy9EKK4lV/4/b9Mx4UIUhrgbImyOJ5rWsysfs4ntsa+tmOxiDDwRitR8Li6OcILGFxc+kPW+yTP6JCFFzp8/j8mTJ8PFxQVv3ryBo6P4Lp1I6hbDMJh3PQZnP2bizjhXufpjamdnh4kTJ2LYsGG04bYMC8sswpP4fHya5SH0B5vUTEhICH744Qds3LgRTZs2Zbs5pAKlPD6eJRRgWnNTtptC5EROERff/ROCBrqqeDG1EYw0VdhuUr2iqKiIFi1asN0MUoXj7zIw9nwkfmpniTWdrNluDpEhDMNgyuUoqCkr4sL3DWGpo8p2k0g10acJKcDn87FixQp8//33mD9/Ph48eEBJJRkWk12MfsdDceh1Oq6OckYzOUoqAZ83tVu8eDEllWTci8QCWGirwNmIlgWJqqioCMOGDUPXrl0xa9YstptDKvEgJg+FpXx0dZCPWaOEXcVcPgafDAOPD1wc4UxJJUK+UlDCw9p/EzCpqQnWdqakEqmZgLh83IvKxYH+DpRUkjE0Y4llubm5GD16NB4+fIirV6+iR48ebDeJiIjLZ7D9aQpW3ouHt6UWnk9tBBdj+tBOpM+HVA5W+Mejna0O202RaYsWLUJmZib8/f1pbyopdvpDJro66la4RImQmuAzDMZfiERYZjEeT3KHgQadRhNShs8wuBqajX0v05DB4WJjDxu2m0RkDMMwWP84CT0b6sHJSJ3t5pAakpq/iIWFhfDz88Pjx4+Rl5cHa2trDB48GO3bt6/yecuWLcP79+8rffzw4cOCPV8qO7Zp06ZYs2ZN7ToggrCwMPTv3x/A540InZycKj3WzMwMgYGBMDOTzP4Qki6fzbrrom9pBaX47p8QRGYV48+etpjQ1Jg2J6wGeRh7tuquTflL78bDyUgdB/o7SKBl8q+0tBSHDh3C33//jX///ReGhoZCj0vz2Et73eIun8dncCE4G793ayCW8uozWRt7SdS9+FYcbkXk4MEEN1jrVv+bdDb7RqpGcS2e8tMKSjH9SjSuh2Wjk70uDg5woGQ+i2Q1rk9/zMStiBy8nNa4zuuuq/LlmdQkltatW4ewsDCMGzcOlpaWePDgATZu3AiGYdChQ4dKnzd9+nRwOByh+4qLi7FmzRo4OjqW20jY3NwcCxYsELpPW1tbfB2pphs3bmDEiBFo3749jh49Cl1d3SqPV1NTQ6tWrSTWHkmXz2bdddG3u1G5iMouxoeZnjDTpmnx1SUPY89W3aKUn1XIxbqHibgdkYPtve1ob6Ua4HK5uH//Pk6ePIlz586hoKAAmzZtQtu2bcsdK41jLyt1i7N8hmHw11sOcot56OOsL5Yy6zNZGntx151TxMX0K9G4EJyFm2Nc0MhUU6zlE/bU57gWR/kN3Jtiy/N07H6RCj11JTyd0ggeZjX7/SDiJ4txnc4pxZxrMVje3hJuVVy9VRb7Vl9IxaeKFy9e4PXr15g+fTp69uwJT09PzJo1C15eXjh48CB4PF6lz7WxsYGrq6vQLT09HVwuF926dSt3vKqqarnjra3rbv0vwzDYsGED+vTpg9mzZ+P8+fPfTCoBQFJSEry8vJCUlCSRdkm6fDbrrou+ZRVyYaWjSkmlGpKHsWerblHKn3I5CheCs3BssCMmNjWWSLvkCZ/Px7///osffvgBVlZW6N27N5KTk7F161akpqZi7ty5FT5PGsdeVuoWV/mcUh6mX4nGgY+FOD/cCca0D06tycrYi7PuN8kc/PM2HcPOhONJXD6ujXJBe9tvn7NVt3zCvvoY1+JyLCAY1s6Nsff+R0z1NoX/ODdKKkkJWYvr7CIuuh8JgYWOCn5sa1Gnddd1+fJMKmYsBQYGQkNDA76+vkL3d+nSBZs2bUJoaCjc3NyqXd7t27ehoaGBdu3aibuptVJYWIjJkyfjwoULOHHiBIYMGVLt55aWluLNmzcoLS2VSNskXT6bdddF314lcWBKSaUak4exZ6vumpafVlAK/8hc7Opjh4Fuht9+Qj3FMAwCAwNx8uRJnD59GikpKejSpQvWrVuHgQMHlpsFWxFpG3tZqlsc5b9J5mDk2XAUlvJxuJseejTUF18D6zFZGHtx1V3EZfDzvwlYfT8B1rqqcDBQw7nhTiJfjIPNvpGq1ae4FieGYfDno3gwyWG4O8YJtraWYq+DiE7W4vrnfxNQzOPDf5wbVJWqnvcia32rT6QisRQTEwNra2soKQmvxbW3txc8Xt3EUmJiIj5+/Iju3btDQ6P8NLrk5GSMHDkSHA4HpqamaNeuHYYOHQo1NbXad6QKcXFxGDBgADIzM/HkyRN4enpKtD5Sdx7H5uHg6zTcGuPKdlMIqdTvjxJhb6CGQW7fTozUR0lJSdi8eTNOnTqF+Ph4dOjQAStXrsSgQYNgYmLCdvNINRWW8jHwRCiaW2lhb197ZCTGst0kImOis4vR81IWGEUlHB/iiKGNjNhuEiFS5+SHTLxLLQQAungFqZXYnGLsDUrDnn72tPJDxklFYikvL6/CDbLK9j7Ky8urdlm3b98GAHTt2rXcY+7u7vD19YW1tTVKSkoQFBSEc+fO4ePHj/j111+hqCiZlYGPHj3C4MGD0ahRI9y8eRPGxrQERV6U8viYejkKE5uaoJN9zafHE1JXnsTlY6CrAVS+8U1Qffb06VP8+OOPGDJkCMzNzdluDqkhHp/BrGvRKOUzONDfAdqqSshgu1FE5vzxKBE6qgp4Ot0T2qq0+TAhXyvm8vHzvwkY52WMPWw3hsg0hmEw5VIUWlhpYWgjmk0v66QisQRUne2ubiacx+PB399fsO/S10aPHi30c/PmzWFqaoqDBw/i6dOn8PHx+WYdSUlJgjWXnz59+ubxe/bswaxZszBjxgxs3LgRKiqiZWLV1NTQqVMnic2sknT5bNYtyfKjs0sQnF6EG6NdxF52fSDLY8923TUpPyW/FM8SCvBnT1uJtEUeWFhY4MGDB2IpS5rGXtbqFrX8qKxiLLwVi0cxebg6ypkSAhIgrWMvzrqfJhdj38ssHO2mJ9YYYrNvpGr1Ia7FVfentEL4R+XC72060gq4mN2rAcIorqWSrMT1gVfpeBKXj9czGlf7atqy0rf6SCoSSzo6OhXOSsrPzwdQ/au2vXjxAllZWRg0aFC16+7YsSMOHjyIkJCQaiWWdu/ejTVr1lSr7GvXrmH27NnYvXs3JkyYUO02VcTMzAz+/v61KoPN8tmsW5LlZxVxAQCmWjR1UxSyPPZs112d8kt4fNwMz8GO5ymw0FFBCyvR9gchNSMNYy+rdde0/HcpHPzxOAmn3meguaUW/p3gVuXVZIjopG3sxV334fPX0edYKMY2MUZLMV9lms2+karJe1yLq+7rYdnocywUzkbq6OqgiyMDHeFkpE5xLaVkIa5T8kux+HYsfu5sDQcD9Tqtm83y5ZlUrImwtbVFfHx8uau/RUdHCx6vjtu3b0NZWRmdOnWqcRuqOytq2rRpCAoKQlBQEPz8/Ko8tlevXnj9+nWtk0rA542/r127hsLCwlqXxUb5bNYtyfKfxudDV00JaspS8askc2R57Nmu+1vlP43PR4PNrzHiTASMNJRxbphTtb8NIrXD9thLkjT17WZ4Nlrs+YDMQi5uj3XF40nulFSSIGkae0nU/d3qg9BWKMVvXRpIpHy2+kaqJu9xLY6636VwMOZcBBa3tcCnWZ74q7cdnIzUKa6lmLTHNcMwmHYlCo4GaphVw0y+tPetPpOKT8M+Pj4oLCxEQECA0P3+/v4wNDSEs7PzN8vIyspCUFAQWrduDV3d6u91U5aRdHGp3lImCwsLNGvWDM2aNfvmhuIKCgo1uppdVdLS0vDdd98hLS1NLOXVdfls1i2p8gPj87H4dhw2dBf/SWh9IatjLw11V1R+EZePgLg8bApIwrTLUfC21ELioqb4Z3BDtLCq3sxPUnsU15IvPzKrCENOhWNpOwtcG+WCDna6tIGshEnL2EtCUHgc3m2ZgTWttCWyeSybfSNVk+e4rm3dUVnFmHY5Ch0OfkJXRz382sVarOUTyZH2uP7raQr8I3PhN9gRSoo1+9st7X2rz6RiKZy3tze8vLywa9cucDgcwV4XL1++xIIFCwRXi9u2bRv8/f2xZ88emJqaCpXh7+8PHo+Hbt26VVjHhw8fcOrUKfj4+MDMzAylpaUICgrCzZs34enpiZYtW0q8n0R+pBWUYsjJMEzwMsHkZqbffgIhEvQyqQBzrsUgKKkAXD4DD1NNtLXRwcI25tBVo71miPzxj8qFqZYyVnawYrspRMYxDIMtgSkAACdD2lOD1G9ZhVzsep6K+9G5uBuVC0cDNfzWtQHGNTGmWc9ELD6kcvCfO3HY288ezkY0y1ieSEViCQCWLl0KPz8/HDt2DHl5ebC2tsaiRYvQvn17wTF8Ph98Ph8Mw5R7/p07d2BqagovL68KyzcwMICSkhJOnjyJ3NxcKCgowMLCAiNHjsTAgQMldkU4Ip/OfMyEipICtvSyYbsphOB2RA6S8ktwZaQzWlppQ4eSSUSO5ZfwsDkgGT0c9WmWEqm1Ux8ycSk4GwBdNp2QUeci8C6Fg7FNjDGjhSm6O+pBiy6GQMSkhMfHD1ej0aOhHkZ70lXS5Y3UJJY0NDQwZcoUTJkypdJj5s2bh3nz5lX42K5du6os39LSEitXrqxNEwkReBSbh872ulClS7cTliXmFmPD4zTMbW2OLg56bDeHEIn780kySngMfu9Gy5BJ7STmlWDq5SjMbmWGzWw3hhCWcUp5uBeVi0sjnNHNkc4niHgxDIOpl6MQmVWMgEGObDeHSAB9KpYRpqamuH37drklgLJSPpt1S6L8J3H5aGujI7by6itZHHtpqbus/CuJCrDSVcWydpYSqYfUHMW1ZMsPTi9Ed0c9WuZZx6Rh7MVthX88nAzVseq7xnLXN1I98hjXotb9b3QeFAD4VvP8luJaekljXO94loKzH7NweaQzGuiJvuxYGvtGPpOaGUukaurq6ujatavMls9m3eIun8dnEJ9bAkcD2ouhtmRt7KWpbnV1dag5tcKmo8HY3tuuxpsfEsmhuJZs+THZJejjrCmxNpCKScPYi1N8bgn+eZuBiyOcoKulKVd9I9Unb3Etat25xTwsuBmLIY0MoaFSvXkHFNfSS9riOiixAD/ejsPffezgZa5Vp3VLW/nyjGYsyYiUlBT4+voiJSVFJstns25xl59VzIDHAOYSuHJMfSNrYy9NdaekpGBIr07oZMLFpGYmEqmDiIbiWnLll/L4+JhWCBdjdYnUTyrH9tiL29/PU+BuooHujnpy1zdSffI89tWtm2EYzLoWDT7DYOd3dmIvn9Q9aYrrnCIuhp8Jx/DGRhjnVfvzVWnqGxFGM5ZkRHFxMR4/fozi4mKZLJ/NusVd/vPUUmirKsJWn2Ys1Zasjb001V1cXIzUT0Gw0Ch/MQPCLopryZV/LzoPBaV8dLLXlUj9pHJsj724XQjOwngvEygoKMhd30j1yfPYV6fu3GIeBp0IxZP4fDyc6A7tGmzUTXEtvaQlrhmGwbQr0VBTUsD23rZ1Wre0li/PKLFESA2diyjC9+6GUFemCX+EHZmFXCy4GQMAaGJOS4JI/fHHo0QMb2wIfXU6fSGiW3wrFp/Si9DPVZ/tphDCCk4pDw9i8vDHoyQk55fiw0xP2NEXpkTMDr1Ox+WQLDyd0oiuLlgP0CdjQmogr5iHh4mlGN2ELpFJ2PPnk2Q8T8gHAAxwNWS5NYTUjRIeH/ej8zDFmzbUJKJhGAZXQrPw17MU7OlrD2cjDbabREidC8sogsOWNxh0IgyGGsq4NsqFkkpE7CIyizDvRgz+6NoAjU3pS9D6gL7ykxGqqqpo0aIFVFVVZbJ8NusWZ/lpnFLwGMDZiPb3EAdZGntpqJtTysPGx8nYEpiMaZ5meMBS30jVKK4lU/7LJA4AwEqHYp4N8hDX256mYPHtOExsaowJTf//BZE89I2IRp7H/uu678YVY/id93iZxEFXR12cHepUq1kkFNfSi+245vIZjD0fgbY2OpjZ0qxO65b28uUZJZZkhLm5OZ49eyaz5bNZtzjLzyniAQD06DLXYiFLY8923XyGwdDT4XibzMHffewwwsMIioPY6RupGsW1+MvnMwymXIrCOC9j2NMVOVkhy3HNMAyW+8dj/eMk/N3HvtwFD2S5b6R25Hnsv6x7hX881j3Mww8tTPFnT1u0tNKCilLtFq5QXEsvtuP6tweJCM8sxpsZTlBQEO9Vi9nuG6kcLYWTEUVFRXjw4AGKiopksnw26xZn+YVcPgBU+1KspGqyNPZs1R2SXogV/vFouPUNAuPycW+8G0Z5GqOkuJi1vpGqUVyLv/y3KRx8TCvEb12sJVIv+TZZjuuLIdnYGJCMSyOcK7yKpiz3jdSOPI99UVERjl66jREnPmL94yRs76CDbb3t0NZGp9ZJpbLyKa6lE5txfSsiB788SMCevvYw1xb/rB95/p2VdfTpWEakpqaiQ4cOSE1Nlcny2axbnOXz/3cBLkUxZ9/rK1kaezbq/pDKgcfOd7gZno15rc3xcZYnHA3VxVY+kQyKa/GX/zqZA1s9VYmcpJLqkdW4fpaQj+mXozCnlRl6OenXad11VT4RnTyP/Y57nzC2f3eExyUicLI7utuId7YnxbX0Yiuud79IxXf/hGBxWwv0dzWo07plpXx5JjVL4QoLC+Hn54fHjx8jLy8P1tbWGDx4MNq3b1/l8+7evYutW7dW+Njhw4dhYCAc1K9fv8Y///yDqKgoqKmpoUWLFhg/fjz09fXF1RUix7IKuaDJSqQu8BkGWwJT4GiojmdTG7PdHEJY8zq5AEdep8OS9lYiNXTkTTpmXInC8MZGWNe1AdvNIaROPE/Ix7F3Gdh1Nx4AcHaYM2wstBAZyXLDiFxLzi/Bf27HYlsvW8xoId59lYhskJrE0rp16xAWFoZx48bB0tISDx48wMaNG8EwDDp06PDN58+dOxdWVlZC9+no6Aj9/P79e6xZswbNmzfHTz/9hJycHBw+fBgrVqzA5s2boaKiItY+Efnz94tUdLehDzdEsiKzijDrWgyexufj8khntptDCCteJRVg1f0EXA3NxnfO+vi1My2DI9UXEJeHKZei8EfXBpjb2kzs+3wQIi34DINLIdl4HJuHgLg8PI0vQAc7HWzoboM5m9huHakvlt6Jh5OROqbSlVvrLalILL148QKvX7/GwoULBUkkT09PpKam4uDBg/D19YWSUtWbJdvY2MDJyanKYw4ePAhLS0ssWbJEUJ6ZmRn+85//4Pbt2+jdu7d4OkTkSlYhF2GZRQhKLMCN8Bxc7qPPdpOIHHubwoHPvg9o20AHTya70+WwSb00/0YMLiYlo5+rAV5MbYSmFlpsN4nIkPepHHx/KhxTmplgno85280hRCKKuXwEJRXglweJeBSTh3a2OujqoId9/RzgZqKB2NhYzGG7kaReeBKXh6Nv0vFwohuUFCmJX19JRWIpMDAQGhoa8PX1Fbq/S5cu2LRpE0JDQ+Hm5larOjIyMhAWFoaxY8cKJanc3NxgZWWFwMBAqU4sGRsb4/z58zA2Nv72wVJYPpt1V6f87CIuwjOLEJZR/PnfzCKEZ3z+N7Pw85XgrHRUsNDHHK4GXIm0sz6ShrGXlJrWHZdTjCuh2Tj4Kg3eFlq4Ocalym/Y2ewbqRrFtWjeJHOwPiAXyiPWI4GrgX8nuKCtjc63n0jqhKzE9evkAvQ4GoIuDrrY3NOmTutmq3wiOlkd+30vU/Gf23HILebB20ILz6c2goux8BdRsto3Unt1OfYJuSUYcy4S47yM4dNA8n+zKa6ll1QklmJiYmBtbV1uVpK9vb3g8W8lln7++Wfk5uZCU1MTjRs3xqhRo2BraytUBwDY2dmVe66trS2Cg4Nr2QvJ0tTUxIABA2S2fDbrLis/t5iHoMQChGcWfZE8KkZYZhHSOZ+TRRbaKnAyUkdDQ3UMcDWAo6E6nIzU4WigBi3Vz/EZSYvUxYbiGiji8tHLLwQPYvJgq6eKvi4GmNnS9JvLNtjsG6kaxXXNjTwbjpPvM9HdUQ9Xf56Kbg66tHRJyshCXN+OyEG/46Ho66KPowMdq/3NuSz0jUiGLI79zmcpmH8zFpu6N8BYLxPoqlW8qkMW+0bEo67GvpjLx3eHP8BWXxXbe9tJrL6K6pbV8uWZVGxDnJeXB21t7XL3l92Xl5dX6XP19fUxdOhQzJ49G7/88gtGjRqF8PBw/Pjjj4iKihKq48syv6Sjo4Pc3NzadkOiUlNT0bNnT4nugC/J8tmsOzQmEXqNfKG//DZa7v2A+TdicSsiB0oKCujjrI8d39khaFoj5Cz1RvzCprg33g17+9ljsa8lBrsbwtNMU5BUIuJFcf35ikUPYvIQONkdEXObYGsv22otf2Ozb6RqFNc1E5lVhJPvM3F/vBsOdzfA5pnDkJaWJrbyiXhIe1w/icvDoJNh+LGNBU5971Sj5RjS3jciObI09gzDYGtgMubdiMHhgQ6Y1cq80qSSuOtmo3wiuroa++WX3iKNw8WZoU7QqKOrG1FcSy+pmLEEoMpvJqt6zNvbG97e3oKfGzdujObNm2POnDn4559/sHz58mqVVd1vRpOSkpCUlAQA+PTpU7WeIw5FRUW4efMmioqKZLJ8turm8RlMvxCC3I+PceWQHdp7OkGnij/CpG7V57jOLuLi7McsrL4fjxEeRmhhVT7pXZvyCXvqc1yL4kZYDhwM1OBro424uCyKayklzXH9JpmDPsdCMbGpCdZ0svr2E8RYtzSUT0QnK2P/ND4fP92Nw5P4fBwd5IhhjY3qrG62yieiq6uxv9doJo5PbgsDjbpLKVBcSy+pSCzp6OhUOCspPz8fQMWzjKpiZmYGNzc3hISECNUBVDz7KS8vr9wV5Cqze/durFmzpkbtIexgGAY/3Y3Hq2QOAMDDTJOSSoR1xVw+plyOwpkPmdBTV8IYT2P80oWudkXqp9CMQmx/loLeTvq09I2I5G0KBz39gtHXRR9/9rShOCJy5W5kDpbcicOrJA76uxrg3Q8ecDBQZ7tZpJ7LKiwFAAx0M8AgN0OWW0OkhVQshbO1tUV8fDx4PJ7Q/dHR0YLHa4phGCgq/r97ZWWU7bX0pZiYGNjYVG+Dx2nTpiEoKAhBQUHw8/OrcbtI3cgs5GLIqXDseJ6C7b1rHj+EiBvDMAhJL8SGx0k4/i4DZ4c5IW5BU6zvbgNVJal4KyakTl0MzkKLPR/Q0EgdqzrWfJYJqd8YhsGGx0louecDujl+vhKWIiWViBx5l8LB8DPh8DLXQugcT5wd5kRJJcK6Ii4fM65EAwDWdqIvRsn/ScWMJR8fH9y6dQsBAQFo166d4H5/f38YGhrC2dm5RuUlJycjODgYTZo0EdxnZGQEZ2dn3L9/HwMGDBBsFB4cHIyEhAT069evWmVbWFjAwsKiRu0RB2VlZbi5uUFZWTJDJuny67LuhzF5GHMuAgYaSng2pRH0SrNY6xupWn2I6zvR+bgWGIYHUdnIKMqArZ4qlrWzRC8nfbGUT3EtfepDXNe27lsRORh2OhyrOlphia+FYJYJxbX0kpa4ZhgG96LzsONZCm5F5ODIIAcMbfTtZUHiqFtayyeik9axf5aQj+5HgtHbWR+7+thBWYRLuEtr34jkSWps+AyDsecjEJfPg5OLK7Q1VMVafnVQXEsvqXjFvL294eXlhV27doHD4cDCwgIPHjzAy5cvsWDBAkESaNu2bfD398eePXtgamoKAFixYgUaNWoEOzs7aGhoICYmBufOnQMAjBo1SqiecePGYeXKlfjjjz/Qu3dv5OTk4PDhw7C1tUXXrl3rttM1ZGlpiY8fP8ps+XVRd0x2MX5/lIj9L9MwrbkpNnS3gbqyIgAN1vpGqiavcV3M5SNFQQ9T9t/F9NtxGOdlgmXNtTC4uSMa6KmJpQ42f2dJ1eQ1rsVVdwmPj7HnIrDAxxxL21mKvXwiGdIQ15xSHoacCod/ZC66Oeri8SR3eJpp1knd0lw+EZ20jT3DMDjxPhOzrkVjhIcRdn5nJ/LyTmnrG6k7khqbXx8k4k5ELh7/0A5uq+pur+EvUVxLL6lZf7F06VJ06tQJx44dw+rVqxEaGopFixahY8eOgmP4fD74fD4YhhHcZ2tri4cPH+LPP//E6tWrce7cOXh6emLz5s3lltB5eHhg5cqVyMrKwi+//II9e/bAw8MDP//8M1RUVOqqqyIpKSnBy5cvUVJSIpPlS7LumOxizLgSBZe/3uJNMgdXR7ngr952/0sqsds3UjV5ieucIi42ByRhzLkINNn1DrrrguC98zU2nrmPPzqbY28/e/SzVxdbUgmguJZm8hLXkqr7zMdMFHL5WNbestxjFNfSi+24vh2RA599HxGeUYQPMz1weaSLWJJK1alb2ssnopO2sf/tYSImXYzEj20ssKMWSSVR6pa28onoJDE218KysfbfBBwZ5ABHPSWZPg9hs3x5JhUzlgBAQ0MDU6ZMwZQpUyo9Zt68eZg3b57QfZMnT65RPU2bNkXTpk1FaSKrkpOT4e3tXaP9oKSpfEnUnc4pxQr/eBx8lY5mFpq4MMIJPRz1yv0RZrNvpGryENcMw2DalWgExOahu6MeJjczgaeZJgxL0uHp4ovBc8vv6yYOFNfSSx7iWlJ1J+SWYOmdeMxtZQ5t1fIXU6C4ll5sxTWfYbDCPx7rHydhenNTrOhgBVMt8X4ZKM+/s6Rq0jD26ZxS+Efl4m0yB1ufpmBff3uM9DCuk7qluXwiOnGPTQanFJMuRmKJryX6OBsgNjZWZs9D2C5fnklNYomQmojLKUZPvxCoKilUmlAiRNLSCkrxw9Vo3AjPwcOJbvAy1xI8FhubzV7DCJEyCbkl+ONRIva9TIOXuSaWtKv7vQqJ7Cnh8TH5UhQuh2Tj2igXdHPUY7tJhIhVRGYRuhwOBqeUDy9zTfynrQWGN67dnmGEiNuiW3Ew01LBig7lZxoTUoYSS0TmhGUUofvRYDgZquPccKcKv/UmRFIYhsHNiBzsf5mGyyHZcDFWx9MpjeBuosF20wiROvklPPzxKAmbApLgbqKB40Maop+LPn0RQL4pv4SHIafC8D6lEP9OcBPbsjdCpAHDMHgYm4eRZyLQzFILp79vCDVlqdmhhBCBh4kl8Hubi4BJ7nQVY1IlSiwR1nH5n/fMisgsQrxCHj7Gl+BhbhoyC7nILOQhg8NFRiEXWYWf/w3PLEJne10cH9JQsI8SIXXlr6cpWHw7DoPdDXB1lDM62evSJa4JqcCLxHwMPBEGFUUFHB7oiCHuBpRQItWSXcRFb78QZBXx8GiSO+z0xbc/HSFsuxOZg2V34vAqmYNxTYyxs48dfWAnUonLZ/DL8wLMbGGGFlbabDeHSDlKLMkIIyMjHD16FEZGkpkeK47yGYZBbjEPGYVcZHC4yPxfIiiT878EUdnPgvv+lzDKzYfCwNXocjoJCqrZ0FFRgLFWMYw0lWGooQxDDSWYaSnD3UQdhhrKsNBRRX8XfahU84+wpF87IjpZiOsypTw+Vt1LwIaAJBwd5PjNqeqy1DciXvI89jWp+7eHSWhironT3ztBQ4Xer2VdXcU1NHTR9XAweAyDfye4iX0/parqlsffWVK1uh77v54mY+HNWEz1NsXJ751gbyC5pCnFdf0lrrHZG5SK9CI+Vna0kkj5oqC4ll6UWJIRWlpaGD16dJ2XzzAM4nJL8CG1EDE5JZ9nDXG+SBD971aWSOL9/4J90FRRhKGGMow0lGGkqQwDDSUYaSjDTk/zc8JI8/NjhhrKMNJsCSMNZRhoKCM2OgoODg4S7xthH1txXVORWUUYdTYCMdkluDrKBd2rsc+HrPSNiJ88j311674bmYMrIVm4O86t2kmlmpRP6l5dxHX7vt+j37FQqCkr4tZoVxhq1M1pqjz/zpKq1eXYH3yVhoU3Y3F0kCOG1cE+ShTX9Zc4xia3mIc19xMw00Oz3HuxLJyHSGv58ozmXcqItLQ0DBkyBGlpaRIpPzU1Fb37D8KJp+HYHJCESRcj0WbfBxj8HgT7LW/w/alwbH+ajBvh2YjIKoKSAuBipI4+zvqY19ocO76zw62xrng9vTFi53uh4KfmyFvWHDHzvfByemPcHuuKU987YVcfe/zSpQEWtLHAeC8T9HUxgLNGEZZPHwNwsqGsKP5lEpJ+7YjoJD02tSk/OrsY258mo/vRYLhvfwd9dWW8mt64Wkml2tYtDeUT0cnz2Fen7gcxuRh+JhwL21igna2O2Msn7JD02Fx9HQXndt9Bn5eHW2Nc6iypBMj37yypWl2MfZ8BgzDO7wWmXY7Cju/s6iSpVFY3xXX9JI6x+fnfBOioKWGUi7pEyhcVxbX0osSSjCgsLMTZs2dRWFhYq3IYhkFSAQ83wrOxOSAJk/+XQGq46RmuXzqPiWeCcfx9Brh8Bv1dDXB0kCPC5ngid5k33s/0xL8T3HF+uDP29XfA+u42+I+vJaZ4m2KwuyE62unCw0wTVrqqNdr7SFx9Y6t8IjppGXuGYRCVVYzDr9Mw5VIUXP96A8etb7D1aQoam2jg+mgXXB3lDDPt6i/JkJa+kbonz2NfVd0Mw2Dj4yR0PRyM0Z7G+LmztVjLJ+yS5NhkFnIx+ewnFL+7iwN9rKCnXrcT6uX5d5ZUTZJjE5pRiHmXQ3D14nk8jc7E/QlumOJtKvZ6KkNxXX/Vdmz8o3KxJTAZ23vbQk2p/Jf+0noeIgvlyzNaCleP/PU0GSv845FXwoe6cjbcjDXgbqqB/q4GmOzIx6T1wIeZHrC3s2W7qYTUqQEnwnAlNBsW2ipoZ6uDua3N0cFOF27G6rTZMCHVkFnIxeRLUbgbmQO/wY4Y2oj2JiDVN+taNAzUlZEMSGTmMiF1LTA+Hx0PfkJTzWIAwI0xLrBrULMZnISwIZ1TinHnIzC3lTl6NNRHZGQm200iMoISS/VEThEXK+8lYFEbC7TV56C9R0MofXHyFhvLBQCh+wipD94kc3AlNBuPJrqhtbU2JZIIqQLvf1fxvB6WjeRIRTyJzEXklTeIzCqGm4kGnk5pBFdjDZZbSWTJxeAsnP2YhYu9bNFrOdutIaT2sgq5GHkmHBOaGmOppxlsF4GuHktkAsMwmHIpCmZaKvi1S81nHZP6jRJLMkJJSQkNGjSAkpKSSM/fG5QGPTUl/MfXAnEx0eUSSLUtvzYkXTebfSNVY3vsY3OKsfBmDDra6cBHzN8kst03wh55G/vUglLsC0rD5dAsvA2LAXTNsORuArycVWCrqYjBTczhaaaJphaatb5kNsW19JLE2OQW8zD7WjQWt7WAh4UCnYeQOieJsZl7IwZ66krY3MMWmalJFNekzok6Nn+/SMWdyFy8mNoIalVsa0KfG0lFpCaxVFhYCD8/Pzx+/Bh5eXmwtrbG4MGD0b59+yqfFxAQgMePHyMsLAyZmZnQ19eHm5sbRowYAUtLS6Fjly1bhvfv35cro2nTplizZo1Y+yNuVlZWiI2NFem5xVw+tj5NxkIfC6hUctJfm/JrS9J1s9k3UjU2xp7LZ3A/OhdnPmTi6Nt0eFtoYU8/+zqpW5bKJ6KT5bFnGAbpHC4S8koQl1OCMx8zcepDJmz0VDHG0xgrOljB86dYWOmoQEFBAZGRkXBwEN+eIRTX0ksSY7PsThw0VBSxrJ0lNFQU6TyE1Dlxj01kVhGOv8vAvfGfr4hJ59eEDaKMzYdUDhbdisXWXrZw+cbMY4prUhGpSSytW7cOYWFhGDduHCwtLfHgwQNs3LgRDMOgQ4cOlT7v3Llz0NfXx9ChQ2Fubo60tDScOXMG8+fPx4YNG2BjYyN0vLm5ORYsWCB0n7a2tkT6JE6lpaWIioqCvb09VFSqv4EwABx7l4HCUj4me5tIpPzaknTdbPaNVK2ux373i1SsvBePnCIeujnq4shARwxyM5DI8jeK6/pLWse+mMtHYl4pEvJKkJBbgsS8EiTkfv657P+JeSUo5n1e7qajqghfWx2cG+aEHg31oKigIKibqyFdfSOSJ+6xOf0hE3uCUuH/vw/gdB5C2CDusfnzSTJaWmmhbQNtiZRfExTX9VdNx6aIy8eosxHo7aSPSU0r/7woavniRHEtvaQisfTixQu8fv0aCxcuFCSRPD09kZqaioMHD8LX17fS6WjLly+Hvr6+0H2enp6YMmUKLl68iNmzZws9pqqqCldXV4n0Q5KSkpLg4uKCmJiYcsmyqqQWlGLdw0TMaGEGbdXKp/SJWr44SLpuNvtGqlaXY8/XNcP8GzFY2cEK01uYQl/CVx2iuK6/pG3sGYbB96fDcf5TFgBAAYCZtgqsdFRgqasKax1VeJppwlJHBVa6qrDSUYWVrip01cr/zZC2vpG6I86xicrlYcr1SPzapQF8bXTEXn5NUVzXX+Icm8wiPg6+ysSRQQ6CL6worgkbajo2i27GIruIhz197av1ZSvFNamIVCSWAgMDoaGhAV9fX6H7u3Tpgk2bNiE0NBRubm4VPvfrpBIAGBkZwcjICOnp6ZJorsyIzSlGj6MhMNFSxqI25mw3hxDWpHNKsfxhNFpaaeM/vha0QTepV66EZuNScBaujnKGu4kGLLRVKl0WTYikJeWVYNLdHHSy18VCOjchciKzkIsVT/NhpauC/i4GbDeHkGo79i4de1+mwX+cKww0pCI1QGSUVERPTEwMrK2ty81Ksre3FzxeWWKpIsnJyUhLS0Pr1q0rfGzkyJHgcDgwNTVFu3btMHToUKipqdWuE1ImLKMI3Y4Ew9lIHeeGO1U5W4kQeZXBKQUAtDvwCY52tjg+xJGSSqReuRKaheFnwrHE1xI9G+qz3RxSz2VwStHjaAjMNBXxz2BHulIWkWnZRVw8TyjAy6QCbA1MgZ4KHyeHudIVlonM+JDKwbTL0VjfrQHa2oj3Ijak/pGKxFJeXh7MzMzK3V+291FeXl61y+LxePjrr7+grq6Ofv36CT3m7u4OX19fWFtbo6SkBEFBQTh37hw+fvyIX3/9FYqK8vENbimPj+9PhcHDTAOnhzpBvYpd/QmRRzw+g1X34vHnzU8AgA3dG2B6l8b0IYbUG1FZxVhxLx4n3mVgTScrLGtn+e0nESJBXD6DoafDoaqkgD0ddKGpQl94Edn1Ma0Q7Q98REEpHx6mGpjV0gyDLArhaq7FdtMIqZa8Yh6+PxWO3k76mNOq/OdwQmpKKhJLAKqcRVDdGQYMw2Dbtm348OEDlixZAhMT4c3HRo8eLfRz8+bNYWpqioMHD+Lp06fw8fH5Zh1JSUlISkoCAHz69Kla7RIHAwMD7Nq1CwYG355e++eTZCTll+LOONdqJ5VqUr64SbpuNvtGqiapsVlzPwF/v0jF9sGNUWi3E2NaOdR5Uoniuv5ia+wjs4pwNzIX/lG5uBCchZZWWng40Q0+DcT3LSTFdf1V27FZcz8B71IKETStEUozEsRefm1QXNdfooxNSn4p+h4LQRcHPfgNchAsLY6MjBRL+eJCcV1/fWtseHwGEy5+jte9/aq3r1JNypckimvpJRWJJR0dnQpnJeXn5wOo3lXbGIbBX3/9hX///Rdz586tcBlcRTp27IiDBw8iJCSkWoml3bt3Y82aNdUqW5x0dHQwffr0bx4XkVmENf8m4O8+9jDWrP5O9tUtXxIkXTebfSNVk8TYnP6QiXWPEnF5pPPnpT+tZoi1/OqiuK6/6mrs43KKcfldOu5F5cE/KgfR2SVooKuKzg66ODPUCb2d9MS+9JPiuv6qzdhcD8vG7/97X26gp4bIDPGWX1sU1/VXTceGU8rDgBOhsNBWxaEBDt/cr47imrChqrFhGAbzb8TgflQuHk50r/BCHbUpX9IorqWXVKyRsrW1RXx8PHg8ntD90dHRgserUpZUunv3LmbNmoVOnTrVuA3VPfmeNm0agoKCEBQUBD8/vxrXI6qMjAyMHz8eGRkVnI39D8Mw+OFqNNra6GC0p5HYy5cUSdfNZt9I1cQ5NmHZXPQ7HoqRZ8Pxe9cG6NlQn+KasEISY8NnGHxMK8TuF6kYevgFtFv2g+1vDzD/Rixyirn4sa0lgmd5ImpeExzo74DvnPUlsp8YxXX9JerYBCUWYPS5iG/u80Xv14QNNRmbYi4fw06HI53DxfnhTtBQ+fbHKIprwoaqxmb94yTsf5WGSyOd4WaiIfbyJY3iWnpJRWLJx8cHhYWFCAgIELrf398fhoaGcHZ2rvS5DMNg+/btuHv3Ln744Qd07dq1RnX7+/sDAFxcXKp1vIWFBZo1a4ZmzZrVaEPx2iooKMDhw4dRUFBQ6TF+bzPwKDYPu76zq/EHiuqULymSrpvNvpGqiWNs4nKKMeVSFPpcyQbDMHg1vTEWtrEQW/mioriuvyQxNiPORMBj5ztsDEgCU1KEgueXcXeELZIXNcWp750wvbkpnIzUJb45PcV1/SXK2DyNz0e3I8EY7GaINZ2sxF6+uFBc11/VHZvCUj4GngzD+9RC3BzjAhOt6q0KoLgmbKhsbA69TsMK/3gcH9IQbWqxTJ7imlREKpbCeXt7w8vLC7t27QKHw4GFhQUePHiAly9fYsGCBYKrxW3btg3+/v7Ys2cPTE1NAQB79uzB7du30bVrV9ja2iI4OFhQroqKChwdHQEAHz58wKlTp+Dj4wMzMzOUlpYiKCgIN2/ehKenJ1q2bFn3HRcTTikPa+4n4M8nyfijWwM4Gqqz3SRCJO51cgH+fJKME+8z0dRcE0e66mJEm+oliAmRJZ/SCnHmYyYeTnRDmwY6iI2NxRkAjoaSTyQRIqoHMbnodywUozyN8VdvW7p4ApFZnFIeBp4IQ3hmEe6Nd4OdvnxdSZrUD9fDsjHtcjR29rFDPxfaP4iIn1QklgBg6dKl8PPzw7Fjx5CXlwdra2ssWrQI7du3FxzD5/PB5/PBMIzgvufPnwMA7ty5gzt37giVaWpqin379gH4vBGXkpISTp48idzcXCgoKMDCwgIjR47EwIEDZfaKcPejczHtchSKuQwujnBGLyd9tptEiMQUlPBw/H0G9r9Mw7OEAnR31MPVUc7oYq+LqKgotptHiERsDUxGB1udWn27SEhdYRgGWwKTseROPGa3NMOG7g0oAUpk1ovEfMy6GoPMQi7ujXeDjR4llYjsCU4vxIgz4VjWzgKTm5my3Rwip6QmsaShoYEpU6ZgypQplR4zb948zJs3T+i+ssTRt1haWmLlypW1aSKrFBUVYWBgIEiAMQyDpXfiselJEqY3N8VvXRpAR4TN1yorvy5Jum42+0aqVpOxyeCUotuRECTmlWBsE2McGuAAF+Oq14ZTXBM2iHNsMov4OPo2C8cGN5RI+TVFcV1/VWdssgq5mHgxEv5RuTg80AHDG1d/v0eKa8KGysYmMa8Ey/3jceR1OoY1NsT57k6w0FEVW/l1geK6/vpybHKKuBh4IgxdHfWwokPVS5JFKb+uUVxLL6lJLJGqWVtbIzMzE8DnpNKCm7E48CoNN0a7oIuDnljLr2uSrpvNvpGqVXdsypJKKkoK+DTLEwYa1XvrorgmbBDH2GRwSnE5NBu7nuTCUkcVfZz1xVq+qCiu669vjU1weiH6HQuFpooink1p9M3Ef03LlySK6/rr67Ep5fHx55Nk/PIgEW4mGnjwvyXI4iq/LlFc119lY8NnGPQ/HgpVJQUcGuAgtiXJFNekIpSKkxFcLheJiYkoLS3Fwpux2P8yDVdHiSep9GX5XC5XLOVJU91s9o1UrTpj82VS6cZol2onlapbvqRQXNdfoo5NEZeP3S9S0fXwJ1hsfIWld+LgqKuEs8OcoKT4/5NBimvChqrG5m5kDtru/whPc008nuRe46TSt8qXNIrr+uvLsXkcmwfv3R+w+UkytvW2xZPJ7rVegkxxTdhQNjYr78QgIC4f54Y5QVtV9JUtlZVPcU2+RIklGZGYmAgrKyv8cOI59r1Mw7XRLvC1Ed9+G2XlJyYmiq1Maambzb6RqlU2Nol5JTjyJh1jzkWg0Y53IiWVqiq/LlBc1181HZtSHh97g1Lh8tdbrLoXDy9zLfiPd0P8gqb4xUcHnmaatSpfnCiu66/KxmZPUCp6/xOKad6mOPV9Q2iJ+OGF4pqwoWxsxh4ORMdDn9DaWhsfZ3livJeJWGZ3UFwTNpSNzR/X3+L4kIZiv7ATxTWpCC2FkxFlG5Yff5eJ6zM7iDWpRIi0yC3mYdz5CFwKyYaZlgq6OupifXcbDHIzEOs3LYRIi7SCUgw4EYrg9CL82MYCs1qZUawTmVDK42Px7Tjsep6K3X3tMN7LhO0mEVJjV0OzAQAvkzi4N74JnV8TuRCcVggA+I+vBbo7imd1CyHfQoklGbElMBkAcHCAPdrZ0h89In+is4vR/3goeHwGT6e4w9tCi64kRORaWEYRvvsnBEaayvg40xNm2ipsN4mQaknJL8XwM+EITi/ErbEuaG+ry3aTCKkRPsNgpX88/rgWDQC4MsoZDSmpROTAuxQOxpyPAABM8aYrwJG6Q0vhZMC/0bn462kKAKCVNf3RI/InKDEfPvs+wEJHBY8nuaO5pTYllYjcYhgGd/63J00jUw3cHedKSSUiM4IS89F8z3uU8Ph4PrUxJZWIzMkr5mHQyTDsepGKgwMdAACqSvSRiMi+Zwn56HTok2BvMDqXJnWJZixJuZwiLsZfiMSUtg5wtNwAfX19idSjr6+PDRskVz6bdbPZN1I1fX19jFywGiOvpGCSjz229LKFsqL4/ghSXBM2VDY24ZlFOPY2A8fepSM8sxizW5lhY3cboY25a1N+XaC4rr/09PQwcM5KDL+cgslt7LG5hw3UlMX3YZzimtSFyKwiDDgeBh7DIHByI5iplMjt2FNc1y/3onIx4EQoRnkY47d2DdE2Tz7HnuJaelFiScrNvh4DbVUlbO7nDg2VxhKrR1dXF4sWLZJY+WzWzWbfSOUKSnj4JTAbp/V6YXMPG8xqZS72OiiuCRu+Hpu8Yh4mXozEuU9Z8DTTwKRmphje2BAN9NTEUn5doriun0p4fCy8n4EbJn2wu49k9lOiuCaS5h+Vi2Gnw9DKWhv/DHKEnroyAHW5HXuK6/rjckgWhp0Ox5xW5ljX1RoKCgpyO/YU19KL5n1KsRPvM3DqfSaODHRAYV42Zs2ahczMTInUlZmZKdHy2aybzb6R8hiGwdmPmWi04x1OPo9Cz7B9GOmkKpG6KK4JG74cm+jsYrQ78BEfUgvxfGojvJrugR/bWoicVPq6/LpGcV3/pHNK0f1ICK6/i0WfiH3oZyOZzeUpromkMAyDnc9S0MsvBJOamuLicOf/JZXke+zluW/k/469S8eQU+FY1dEKv3drAAUFBbkee3num6yjxJKUisspxsyr0VjbyQpNLbSQn5+PHTt2ID8/XyL1Sbp8Nutms29EWEh6IXr6hWD0uQiM8zLGreE2uOy3Vy7HnuK6/iobm3shyWi99wPMdVQQMNkdzSy0xFo+xTWRtI9phWi99yMKuXxcGGSN04f2yOXYU1zLt8dx+Vh0Kxb7+9vj924NhJYfy/PYy3PfyGe7X6Ri/PlIbOlpg//4Wgrul+exl+e+ybp6txSusLAQfn5+ePz4MfLy8mBtbY3Bgwejffv2bDdNgM8wmHgxCp5mmljYxoLt5hBSa/klPPz6IBF/PklGV0ddvPvBAw0N1REbG8t20wiRmFFnIzC9iyc29rAR695hhNSFG+HZGHEmAt0d9XBwgD3SkxLYbhIhIvG10UHwbE/Y1GKmKCHSxj8qF3Oux+DAAAeM9jRmuzmE1L/E0rp16xAWFoZx48bB0tISDx48wMaNG8EwDDp06MB28wAAWwOT8SKxAK+nN67xpq6ESJtPaZ9nKSkrAqeGNkRfZ326SgWROxmcUrxM4uBlUgHuvIoEAKzuZIUlvWxZbhkhNcMwDLY/S8HCm7H4qb0lVnSwgiK9ZxMZR0klIm862ekgYJI7vC3FMxuakNqqV4mlFy9e4PXr11i4cKEgieTp6YnU1FQcPHgQvr6+UFKSzN4B1fU2hYNld+Oxt589bPWF/wiqqEj2ctSSLp/NutnsW33nYKCGea3NML25GTRUyq++leexl+e+1XdZhVzsfJ4iSCbF5pRATUkBTcw10VBLBcoqKhjpIblvECmuiSSU8viYez0Wh16n4cggRwxvbCT0uDyPvTz3jVRNnsdenvtW3ykoKFSZVJLnsZfnvsmyepVYCgwMhIaGBnx9fYXu79KlCzZt2oTQ0FC4ubmx1DqgiMvH2HMRGOBqgFEewidzNjY2KCkpkVjdki6fzbrZ7BsB1JQVMd+n4iWd8jz28tw3AqgoKeBGeA6ammuij7MVvC214GasDhWlz8nToxPlc+wpruVbUBIHV8OycW+8G1pZaws9Js9jL899I1WT57GX576Rqsnz2Mtz32Rdvdq8OyYmBtbW1uVmJdnb2wseZ9OF4CxkFfGw8zu7ckuF+Hw+cnNzwefzJVK3pMtns242+0aqJs9jL899I4C2qhIeTnTHtt52mNDUBJ5mmoKkkjyPvTz3jQCtrbURMtuzXFIJkO+xl+e+karJ89jLc99I1eR57OW5b7KuXiWW8vLyoK1d/mSp7L68vLy6bpKQ4Y2NEDStEQw0yk8ki4+Ph56eHuLj4yVSt6TLZ7NuNvtGqibPYy/PfSNVk+exl+e+kc/UlSs+NZTnsZfnvpGqyfPYy3PfSNXkeezluW+yrkaJpcWLF5d7kWUtm1fVpsHV2VA4KSkJL1++xMuXL/Hp0ydxNg0AYKxJazoJIYQQQgghhBAiG2qUWNq0aRMSExMFP/N4PKioqODly5dib5gk6OjoVDgrKT8/HwAqnM30td27d8Pb2xve3t4YPXq02NtICCGEEEIIIYQQIitqtHk3wzDVuk9a2dra4uHDh+DxeEL7LEVHRwse/5Zp06ahX79+AIDXr19j0qRJ5WYuxcfHIzs7W2ztBj7PlAKAd+/eIT09vVZlVdQ+cZZfU1/XLe7XT9x9k8T4ssXV1RWamppC9xUWFgIAxXUtUVyzh+Jaciiu2UNxLTkU1+yhuJYcimv2UFxLDsU1eyqKayFMDSgoKDBPnz4V/MzlchkFBQUmKCioJsWw5sWLF0zfvn2ZBw8eCN2/atUqZty4cQyXy61ReX5+fgwAutFNZm8V/e5SXNNN1m8U13STxxvFNd3k8UZxTTd5vFFc000eb9/K+SgwTPWnHCkqKiIwMBAtW7YE8P+lcC9evECzZs2qWwyrVq5cifDwcIwbNw4WFhZ48OABbt26hQULFqBjx441Kis9PR03b96EnZ0dNDQ0JNNgQiSooswzxTWRdRTXRB5RXBN5RHFN5BHFNZFH35qxVOPE0uHDh9GoUSMAnxNLrVq1gp+fH1xdXcsdL43JpsLCQvj5+eHRo0fIy8uDtbU1hgwZgvbt27PdNEIIIYQQQgghhBCZUuPE0tdXTmMYptL7eDyeeFpJCCGEEEIIIYQQQqROjTbvPnjwoKTaQQghhBBCCCGEEEJkTI1mLBFCCCGEEEIIIYQQUqZGM5ZIebGxsXV+mUVCxMHY2Bg2NjYVPkZxTWQVxTWRRxTXRB5RXBN5RHFN5FFVcS1Q5TXjKvH+/Xtm0qRJjJOTE6OlpcVoaWkxTk5OzNSpU5mPHz+KUqRMiomJYTQ1NVm/9B/d6CbKTVNTk4mJiaG4pptc3Siu6SaPN4prusnjjeKabvJ4o7immzzeKovrL9V4xtKxY8cwYcIElJaWwsLCAh4eHmAYBrGxsdi7dy8OHz6Mo0eP4vvvv69p0TInPT0dHA4Hfn5+cHNzY7s5hFTbp0+fMHr0aKSnp5fLPlNcE1lFcU3kEcU1kUcU10QeUVwTeVRVXH+pRoml2NhYTJkyBY0bN8bu3bvRvHlzocefPXuG6dOnY+LEiWjdujUaNGggWutljJubG5o1a8Z2MwgRK4prIo8orok8orgm8ojimsgjimsirxRrcvCePXugr6+Pu3fvlksqAUDLli1x584d6OjoYO/evWJrJCGEEEIIIYQQQgiRPjVKLN2/fx8TJ06Evr5+pccYGhpi0qRJ8Pf3r23bCCGEEEIIIYQQQogUq1FiKTQ0FK1atfrmca1atUJoaKjIjSKEEEIIIYQQQggh0q9GiaWcnBwYGRl98zgjIyPk5OSI3ChS916+fIlJkyahRYsWsLe3R5MmTdC3b1+sWbOG7aZV28mTJ2FlZYU3b96Ircy7d+9i06ZNYiuPSLeyGHJwcEB8fHy5x4cMGYLOnTvXuNzz589Xujw4Li4OY8aMQaNGjWBlZYWVK1ciLi4OVlZWOHnyZLm2xcXF1bh+Il/KYqGi29q1a1lpU6tWrTB27FhW6rayssJPP/3ESt2kblUW91/fAgICalRuQEBAuedV9fd/27ZtuHHjRq36QuqvsWPHws3NDQkJCeUey8rKQtOmTdG/f3/w+fxa17Vp06Zq/c4MGTKkWuWJeh5EZJ8kPmeVCQ0NxaZNm+gcV8bVaPPu0tJSKCkpffM4RUVFcLlckRtF6tadO3cwYcIE+Pj4YPny5TA1NUVqairevHmDixcvYtWqVWw3kTX+/v44dOgQFi5cyHZTSB0qLi7GH3/8gb/++kss5Z0/fx4hISGYMmVKucdWr16NV69eYdOmTTAxMYGZmRkYhil3XNeuXXHp0iWYmpqKpU1E9m3evBkNGzYUus/c3Jyl1hAieZcuXRL6ecuWLQgICMCpU6eE7nd2dq5RuR4eHrh06ZLQ86r6+//XX3/hu+++Q8+ePWtUDyEAsHHjRnTu3Bk//vgjjh07JvTY8uXLkZ+fjy1btkBRsUbf/1doxIgR6Nixo+Dn1NRUTJ48GRMnTsSAAQME9+vo6NS6LkJEFRoais2bN8PHx6feXPxLHtUosQR83mepom/yv0TL4GTLrl27YGNjg2PHjkFZ+f8h0b9/fyxfvpzFlhHCjk6dOuHChQuYPn06GjVqJNG6goOD0bRpU6EPKBV9Y2NkZFStGaOk/nB1dUWTJk3YbgYhdcbb21voZyMjIygqKpa7v6Z0dHRqXUZt8Xg8cLlcqKmpsdoOInmmpqb47bffMGPGDBw9ehRjxowBAFy/fh0XLlzAb7/9Bnt7e7HUZWlpCUtLS8HPZecXVlZWrMc8IdKCYRgUFRVBQ0OD7abItBqnwpcsWYIhQ4ZUeVu2bJkk2kokJCsrCwYGBkJJpTJffltSttTh3r176NGjBxwdHdG+fXucOHFC6DkZGRlYunQpOnbsCCcnJ3h6euL777/H06dPhY4rW+6zc+dObN26FS1atICDgwN69eqFhw8flmtLeHg4fvjhBzRp0gT29vZo0aIF5syZg+LiYqHj8vPzsWTJEjRu3BiNGjXC5MmTkZycLHTMxYsXMWLECDRt2hSOjo7o0KEDfvvtN3A4HMEx8+bNw6FDhwAIT7+naZryb8aMGTAwMMBvv/1W5XEMw+DQoUPo1q0bHB0d4e7ujilTpiAmJkZwzJAhQ3D37l3Ex8eXW6phZWWF6Oho+Pv7fzO+vl4KFxkZCRcXF0ydOlXouEePHqFBgwZYv359LV8FIouioqIwf/58tG3bFo6OjvD29sa4cePw6dOncsfm5ORgzZo18PHxgb29PTw9PTFmzBiEh4cLjikpKcGWLVvQvn172Nvbw8PDA/Pnz0dGRkaF9V+/fh1du3aFg4MDfHx8sH///nLHJCQkYPbs2fD09IS9vT06dOiAv//+u9yyj6ysLCxduhTe3t6ws7ODj48Pfv/993Lv+V9jGAbr1q2Dra0t/vnnn+q8bEROTJ06FZ06dRK6b9y4cbCyssLly5cF97179w5WVla4desWgPJL4ar6+29lZQUOh4PTp09XuIwoNTUVixcvFsRt69atsXnzZqGZ/F+e/2zZsgWtW7eGvb19jZfwEdnVr18/9O/fHz///DPi4uKQmZmJJUuWoH379hg3bhxu3bqFvn37wtHREc7Ozhg+fDhevHghVMa8efMq3Pu2bPlbdVW21L6iJaJlnj59ij59+gj+zqxfvx48Hk/omJr+/SCyZd68eXByckJUVBTGjBkDJycnNG/eHGvWrCn3d/rw4cPo2rUrnJyc4OzsjPbt22PdunUAPsfftGnTAADff/+94H21bDuIBw8eYMKECfD29oaDgwPatm2LxYsXIzMzs1ybbt68ia5du8Le3h4+Pj7Yt29fhb8PZUvojxw5gg4dOsDe3h6nT58G8Hk2eJ8+fdCoUSO4uLigR48eOH78eLnVBGWfi2/fvo3u3bsLPk/evn1b0K8OHTqgYcOG+O677ySyhFDa1GjG0r179yTVDsIib29vHDt2DCtWrMDAgQPh4eEBFRWVCo/9+PEj1q5di5kzZ8LExATHjh3DwoULBSdPwOcPAwAwf/58mJqaoqCgADdu3MCQIUNw8uRJtGnTRqjMgwcPwtraGmvWrAGfz8euXbswZswYnDlzBs2bNwcAfPjwAQMHDoShoSEWLVoEe3t7pKam4tat/7Z332FNXW8cwL9J2HsvWeICJ+JEcVVbq3XvjXXXarVq66h71Na6fnXgaOvCulq1WvdGEReKe4KyZO8RQsb5/UGJRoZkkRDez/PwKDc3576H9yUk55577lkIhUKZM3zfffcdOnfujE2bNuHt27dYvnw5pk6dKn3BAIo+fH3yyScYN24cTExM8OrVK2zevBn37t2T7jd9+nTk5+fjxIkTMtPv6VIk3WdmZoZp06Zh4cKFuHbtGgICAkrdb/bs2Th48CDGjBmDH374AZmZmVi3bh169+6Nc+fOwd7eHj/++CO+//57REdH47fffpM+t27dujh27BjGjRsHDw8PLFiwAACkl6J+jJeXF1atWoXJkyfj999/x9ixY5GcnIwpU6agVatWdPlmNVA8w+F9SUlJsLa2xrx582Bra4vMzEwcOnQIPXr0wJkzZ6SXzuXm5qJv376IjY3F5MmT4efnh7y8PNy4cQNJSUmoXbs2JBIJxowZg5s3b2Ly5Mlo1qwZ4uPjsXr1agwYMAAnT56UObv3+PFjLFq0CDNnzoS9vT2OHDmChQsXQigUYtKkSQCKTjz06tULQqEQ3333Hdzc3HD+/HksW7YM0dHR0jeaBQUFGDhwIKKjozFz5kzUr18fN2/exMaNG/H48WPs2bOn1J+JQCDAt99+iwsXLmDnzp0lBhmIbmvXrh1OnDiBpKQkODo6QiQS4caNGzAyMsLVq1fRs2dPAMDVq1ehp6dX4v1IsfL+/h87dgyDBg1CmzZtMH36dADvLiNKTk7GF198AS6Xi2+//RYeHh4IDw/Hr7/+itjYWKxbt07mOL///ju8vLywYMECmJubq2yWCqkaVqxYgRs3bmDGjBmwtbWFUCjEmjVrcOTIEUyZMgUdOnTA5s2bIRAIEBQUhIEDB+LAgQNo2bKlRuNOTk7G5MmT8fXXX6NWrVq4cOEC/ve//yErKwsrVqwAALn/fpCqSSQS4csvv8SQIUMwYcIE3Lx5E+vXr4eFhQW+/fZbAEUn8+fNm4cxY8Zg4cKF4HA4ePPmDV6+fAmgaKmHOXPm4KeffsKKFSvQqFEjAICnpycA4M2bN2jWrBmGDh0KCwsLxMbGYtu2bejTpw8uXLgg/cx66dIljBs3Dq1atUJQUBBEIhG2bNmC1NTUUmM/ffo0bt68iW+//Rb29vaws7MDUDTwP2LECOlMv7t372L+/PlITEyU9qnYkydP8NNPP+Gbb76Bubk51q5di/Hjx2PKlCm4du0a5syZAw6HgxUrViAwMBBhYWG6XfeMKCw8PJwBYOHh4ZoORSlpaWmsT58+zMXFhbm4uDAPDw/Wq1cvtmHDBpabmyvdr2XLlszLy4vFxcVJt/H5fFa/fn32/fffl9m+SCRiQqGQDRo0iI0dO1a6PSYmhrm4uDA/Pz/G5/Ol23Nyclj9+vXZ4MGDpdsGDhzIfHx8WGpqapnH2b9/P3NxcWFz586V2b5582bm4uLCkpKSSn2eRCJhQqGQhYWFMRcXF/b48WPpY/PmzWMuLi5lHrOqKq92daWuFVFcQxEREUwgEDB/f3/WrVs3JpFIGGOM9e/fn3Xq1IkxxtidO3eYi4sL27Jli0wb8fHxzMvLiy1fvly6beTIkaxly5alHrNly5Zs5MiRMtuKfzf2799fIraYmBiZfefMmcM8PT3ZnTt32MCBA1njxo1ZYmKi4j+EKqy61HVxLZT2JRQKZfYViUSssLCQtW3bli1atEi6fe3atczFxYVduXKlzOMcPXqUubi4sBMnTshsj4iIYC4uLmznzp3SbS1btmQ1atRgjx49ktl3yJAhrF69eiw/P58xxtiPP/7IXFxc2N27d2X2mzNnDqtRowZ79eoVY4yx3bt3MxcXF3bs2DGZ/TZt2lQibhcXFzZv3jyWnp7O+vTpw/z8/ErEUZVVl7pWxLRp01jt2rWl379+/Zq5uLiwQ4cOMcYYu3XrFnNxcWHLly9nrVu3lu43ZMgQ1qtXL+n3oaGhzMXFhYWGhkq3lff3v3bt2mzatGkltn///fesTp06Mu+TGGMsKCiIubi4sOfPnzPG3r3Gt2nThhUWFsrfcR1AdV3kwoUL0tfvv/76i4nFYubn58c6d+7MxGKxdL/c3FzWuHFjmbqdNm1aqe8tVq9eXWbtFtdeUFCQdFtZ7y9K+73o378/c3FxYWfOnJHZ97vvvmOurq7S2pfn74cu0eW6fv89MmNF9Vfa3+kRI0awdu3aSb//4YcfmI+PT7ltHz9+vEStlab4M1tcXFyJOuzevTtr3rw5EwgE0m25ubmsQYMGJX4fXFxcmLe3N8vIyCj3eGKxmAmFQrZ27VrWoEED6ecBxt59Ln779q1026NHj5iLiwtr2rSp9H0PY4ydPn261N+bqqKitSv3pXAPHz4sd42luLg4PHz4UO4BLqI5NjY2OHLkCE6ePIl58+bhs88+Q1RUFFauXInOnTvLTDWsX7++zHRCIyMjeHl5lbizxe7du9G1a1d4eXnB3d0dHh4euHbtmnR0+n3dunWDkZGR9HszMzN8+umnuHnzJsRiMfh8Pm7cuIEePXpUaI2Zzz77TOZ7Hx8fAJCp2+joaHz99dfw9fWFm5sbPDw80L9/fwAoNUZS/RgYGOD777/H/fv3SywYCxQtes/hcNC/f3+IRCLpl4ODA+rXr4+wsLBKiXPx4sWoW7cuBg0ahLCwMGzYsAGOjo6VcmyiWf/73/9w8uRJmS+g6I5VHTt2hKenJ9zd3eHp6YnXr1/LXOJ26dIleHl5oX379mW2f/78eVhaWuLTTz+VqfEGDRrAwcGhRI3Xq1evxJpkffr0QU5OjvR9QWhoKOrWrYumTZvK7Ddo0CAwxhAaGirdz8TEBD169CixH1B0yef7YmJi0KtXL+Tk5OD48eNqXxuNaCdPT0+4ublJL6cPCQmBj48P+vXrh5iYGLx58wYCgQC3bt1Cu3btVH788+fPo02bNtLZUsVfxXfR+vB35tNPPy1zhjipHj755BP4+fmhZs2a6N+/PyIjI5GYmIj+/fvLLEdhamqK7t274+7du+Dz+RqMuOh9+ofvtfv06QOJRIIbN24AkP/vB6maOBwOPv30U5lt9evXl/lc6Ovri6ysLEyePBlnzpwp9RK28qSmpmL27Nlo3ry59DNl8ay94s9s+fn5uH//Pj7//HMYGBhIn2tqaloivmJt27aFlZVVie3Xrl3D4MGD4e3tLf2MuHr1amRkZJSY/dSgQQM4OztLv69Tpw4AwN/fX2ZmUvFs8dLuBKlL5LoULiQkBF26dEFYWBhcXV1L3ScpKQn+/v44fvw4unbtqpIgSeVo0qSJdCFYoVCIFStWYPv27di8ebN0EW9ra+sSzzM0NERBQYH0+61bt2Lp0qUYOXIkvvvuO9jY2IDH4+GXX34pddDG3t6+xDYHBwcUFhYiLy8PeXl5EIvFMr+45fkwxuIXmOIY8/Ly0K9fPxgaGuL777+Hl5cXjI2N8fbtW4wbN06mL6R66927N7Zs2YJVq1ahe/fuMo+lpqaCMVbm4skeHh6VESIMDQ3Rt29fLFu2DI0aNSp3oIDoljp16pSovwULFmDnzp2YPHky/P39YWlpCS6Xi1mzZsm8tqWlpX10DY6UlBRkZWVJp6N/6MM3h2W9lgPvLpHOyMgo9f1D8WDo+/s5ODiAw+HI7GdnZwc9PT3pfsUiIiKQnp6O2bNnyyxUS6qfgIAA6dINV69eRbt27eDj4wN7e3tcvXoVNWvWREFBgVoGllJTU3Hu3LkyX/8//J2hkwAEKPo7Xvxetfi1rbRlFxwdHSGRSJCZmanRy2mKLxl634ev9fL+/SBVk7GxsczkAKDoc9f77zcGDBgAsViMvXv3Yvz48ZBIJPD19cX333//0fesEokEQ4cORVJSEqZPnw5vb2+YmJhAIpGgZ8+e0uNkZmaCMVZqbZb23gQo/Xfs3r17GDZsGPz9/bFq1So4OzvDwMAAp0+fxq+//lriM+KHA1PFv8dlbf/YGpFVnVwDS5s2bcKAAQPKvYtAs2bNMHjwYPz22280sFSF6evrY8aMGdi+fTueP38u13MPHz4sXWT1fbm5uaXun5KSUmJbcnIyDAwMYGpqCn19ffB4PCQkJMgVR1lCQ0ORmJiIv/76C/7+/tLtWVlZKmmf6A4Oh4N58+Zh6NChJRYBtrGxAYfDwZEjR2TOjhQrbZs6PHv2DL/88gt8fX0RERGBrVu3ShdBJNXP4cOHMWDAAMydO1dme3p6OiwtLaXf29rafvQ11cbGBtbW1mUugG1qairzfVmv5cC7AX9ra+tS1xBLSkqSHrN4v3v37oExJjO4lJqaCpFIJN2vWK9evWBvb4+ff/4ZEolEuvYNqX4CAgKwb98+3Lt3D/fu3cO0adMAFJ2dDgkJQVxcHExNTeHn56fyY9vY2MDHxwezZ88u9fEPB5I+HDglpPi1sqzXSS6XK/3QamhoiMLCwhL7yTtoUzww8OGH3rLaKW3Nmg9f6+X9+0F02+DBgzF48GDk5+fjxo0bWLNmDQIDA3H16tUyJ6sARe9xnzx5gnXr1klnLANFa+W+z8rKChwOp9za/FBpr7///PMP9PX1sWvXLpkBs9OnT3+0j0TOu8KFhoaiT58+H92vV69e0qmQRPsVv6H/UPHsInnPqHE4nBK3y33y5AnCw8NL3f/UqVMyI8C5ubk4d+4cWrZsCR6PB2NjY7Ru3RonTpxQ6RmODz/4BwcHl7mPpqcdE81p37492rdvj3Xr1iEvL0+6vUuXLmCMISEhQTrb7/2v4kswgZKz+lQlPz8fEydOhJubGw4dOoQvv/wSK1euxN27d1V+LFJ1fPjadv78+RJ3xuzUqROioqJKXFL2vi5duiAjIwNisbjUGi+e2l3s+fPnePz4scy2o0ePwszMTLoYZ9u2bfHixYsSl8z/9ddf4HA40sWUAwICpDd++HC/4nY+NH36dCxZsgSrV6+WLgJOqp+AgABwOBz88ssv4HK50huLtGvXDtevX8fVq1fRqlWrj16CVt7f/w/PyBfr0qULnj9/Dg8Pj1J/Z5ycnFTQQ6LLatWqBScnJxw5ckTmLlT5+fk4efIkmjVrJp2t5ObmhtTUVJlB/cLCQly+fFmuYxZ/sP/w7qHFd038UG5ubonHjh49KvP7Ju/fD1I9mJiY4JNPPsE333yDwsJC6eSFD68uKVY8+PPh58oPP7OZmJigSZMmOH36tMxga15eHs6fP1/h+DgcDng8Hng8nnQbn8/H33//XeE2qjO5ZiylpKRU6PaVzs7OFbqrEdEOw4cPh7OzM7p06SK9E9Djx4+xbds2mJqaYuzYsXK116VLF6xfvx6rV69G69atERkZifXr18Pd3b3EHYwAgMvlYujQoZgwYQIkEgk2b96M3NxczJo1S7rPokWL0LdvX/To0QNff/01PD09kZqairNnz+Lnn3+GmZlZheNr3rw5rKysMGfOHMyYMQN6eno4cuQInjx5UmJfb29vAMDmzZvRqVMn8Hg8+Pj4VNpsFKIdfvjhB3z++edITU1FvXr1AAAtWrTA8OHDMWPGDDx48ACtWrWCiYkJkpOTcevWLXh7eyMwMBBAUR2dPHkSu3btQuPGjcHlcsu8hE4es2fPRnx8PE6cOAETExMsXLgQ4eHh0uvY35+hQqqHLl264NChQ6hduzZ8fHzw8OFDBAUFlbiUePz48Th+/DjGjBmDr7/+Gk2bNkVBQQHCwsLQpUsXtG3bFr1798aRI0cwcuRIjBs3Dr6+vtDT00NCQgKuX7+Orl27olu3btI2HR0d8eWXX2LmzJlwcHDA4cOHERISgh9++EH6QWjChAn466+/MGrUKMyaNQuurq64cOECdu3ahVGjRqFWrVoAiqbO79y5E9OnT0dsbCy8vb1x+/ZtbNiwAZ988kmZ0+fHjRsHU1NTfP/998jLy8OyZctoVkg1Y2dnB29vb1y5cgVt2rSR1l67du2QmZmJzMxMLFq06KPtlPf338fHB2FhYTh79iwcHR1hamqK2rVrY9asWQgJCUHv3r0xZswY1KpVCwKBALGxsbh48SJ++uknulSTlIvL5WL+/PmYMmUKRo0ahZEjR0IgEGDLli3Izs7GvHnzpPv26tULq1evxldffYWvvvoKAoEAf/zxByQSiVzH9PX1Ra1atbBs2TKIxWJYWlri1KlTuH37dqn7W1tbY+7cuYiPj4eXlxcuXLiAvXv3YtSoUdLPifL+/SC667vvvoORkRGaN28OR0dHJCcnY+PGjbCwsICvry+Ad6+3e/fuhZmZGQwNDeHm5obatWvD09MTP/74IxhjsLKywrlz5xASElLiOLNmzcKoUaMwbNgwjBs3DmKxGEFBQTA1NUVmZmaFYu3cuTO2bduGr7/+GiNGjEBGRga2bNlCn/sqSK6BJVNT0wrNGMnIyICJiYnCQZHK9c033+Ds2bPYvn07kpOTUVhYCAcHBwQEBGDq1KnShcjkaY/P52Pfvn0ICgpCnTp18NNPP+HUqVOlLtb35ZdfQiAQYMGCBUhLS0PdunWxe/dutGjRQrpPgwYNcOLECemZ6Ly8PNjb26Nt27ZyL3xpY2ODXbt2YenSpZg6dSpMTEzQtWtXbNmypcTlm3379sWdO3ewa9curFu3Dowx3LhxA25ubnIdk1RtDRs2RJ8+fXDkyBGZ7atWrUKzZs2wZ88e7Nq1CxKJBE5OTmjevLnM4sRjx47F8+fP8fPPPyM7OxuMMaUX8Pvzzz9x+PBhrF27VjrYZWBggKCgIHz++eeYMWMGfv/9d6WOQaqepUuXQl9fHxs3bkReXh4aNWqE3377DatWrZLZz8zMDEeOHMGaNWuwd+9erFu3DpaWlmjSpAmGDx8OAODxeNixYwd+++03/P3339i4cSN4PB6cnZ3RunVr6RvBYg0aNMDgwYOxZs0avH79Go6Ojli0aBEmTJgg3cfW1hbHjh3DypUrsXLlSuTm5sLd3R3z58+X2c/IyAiHDh3Czz//jC1btiAtLQ1OTk6YOHEiZsyYUe7PYOjQoTAxMcE333yD/Px8rF69WmYRXKL7AgIC8PTpU5l1lGrUqIGaNWvi9evXFVpfqby//0uWLMEPP/yAyZMng8/nw9/fH3/99RccHR1x6tQprF+/Hlu2bEFCQgJMTU3h7u6Ojh07lrpQLCEf6tu3L0xMTLBhwwZ89dVX4HK58PPzw8GDB2XeG7u7u+OPP/7ATz/9hAkTJsDR0RETJkxAWloa1q5dW+Hj8Xg87Ny5E/Pnz8ecOXNgYGCA3r17Y/ny5Rg1alSJ/R0cHLBixQosW7YMz549g5WVFaZOnSpzQljevx9Ed7Vs2RIHDx7E8ePHkZWVBRsbG7Ro0QL/+9//pDdlcnd3x5IlS/D7779L12Rau3YtBg8ejJ07d2LhwoWYPXs29PT0EBAQgP3790sX8C7WqVMnbN++XTrYam9vj8DAQCQlJVV4xlFAQADWrl2LTZs2YfTo0XBycsKwYcNgZ2eHmTNnqvxno2s47P15lh/Rrl07NG7cGJs2bSp3v8mTJ+PBgwflTrHXBXfv3kWzZs0QHh6ulmv1dV1sbCxat26NBQsWYNKkSZoOp1opr3aprklVRXVNdBHVNdFFVNdEF1FdaxehUIjPPvsMTk5O2Ldvn6bDqbIqWrtyncIbPHgwfv/9d1y5cqXMfS5duoQdO3Zg6NCh8jRNCCGEEEIIIYQQIreZM2fin3/+QVhYGP755x8MHToUL1++xFdffaXp0KoFuS6FmzBhAnbu3InPPvsM48aNQ+/evVGzZk0ARauzHz16FL///juaNGmC8ePHqyVgQgghhBBCCCGEkGK5ublYunQp0tPToaenh0aNGmHPnj1lrstIVEuugSUDAwOcOXMGI0eORFBQELZs2SLzOGMM3bp1w+7du2mRK/JRbm5uSq8zQwghhBBCCCGketu6daumQ6jW5BpYAooW3jx58iTu3r2Ls2fPIiYmBkDRoltdu3aVWbCWEEIIIYQQQgghhOguuQaWQkJC4OfnBzMzM/j5+ZW5eFNqaiqOHTuGMWPGqCRIVcrPz8eBAwfw+vVrREVFITs7G0OGDMGwYcM0HRohhBBCCCGEEEJIlSLXwFKnTp0QFhYmvb2fRCKBkZERbt68KTNTKTIyEuPHj9fKgaWcnBycPXsWnp6eaN26Nc6ePatwW3w+HwAQERGhougIqRxPnz4t8zGqa1JVUV0TXUR1TXQR1TXRRVTXRBeVV9fvk2tgiTFW4nuRSFRiuzZzcHDAn3/+CQ6Hg+zsbKUGljIzMwEAY8eOVVF0hFQeExMT2NnZldhOdU2qMqproouorokuoromuojqmuiisur6fXKvsVTVcTgclbXl7OwMAAgODoaPj490e1xcHFxdXVV2HABISEhAjx498O+//0qPq6jS4lNl+/L68Niq/vmpum/qyK8m2NnZwd3dvcR2qmvVoLrWDKpr9aK61gyqa/WiutYMqmv1orrWDKpr9aK61oyy6vp91W5gSR18fHxk1puysrKCl5eXSo+RmpqKYcOGoVWrVh8dLfyY0uJTZfvy+vDYqv75qbpv6sivNqK6Vg7VtXaiulYO1bV2orpWDtW1dqK6Vg7VtXaiulYO1bX2ooGlKsLOzg579+6tsu1r8tia7Bspny7nXpf7Rsqny7nX5b6R8uly7nW5b6R8upx7Xe4bKZ8u516X+1bVyT2w9Pz5c+jpFT1NLBYDAJ49eyazz4ff65KEhAQkJCQAqPhCVqqQm5uLAwcOYPDgwTAzM1NJm4wx5OXlISMjA3FxcThy5AhatGgBDoeDgoICCAQC8Pl88Pl8FBQUQCgUQiwWQywWQyKRQCwWQyQSQSQSQSKRgDEm/Spu/2PbGGMQCASIjo6Gm5sb9PT0kJOTgx07dsDBwUEl/SzrZydhDAIRQ7ZAjDyhGHmFEuQJJcgWiMEXSpAvlEAglkAkYRCKGfhCCcQMiE3Kg/7LaEgY/vti0n8ZAxjw3r+s6F8AYgnDkIa26FbHSiX90gXqqGtVti8Wi5Gbm4v09HRkZ2cjMzMTmZmZEIvFEAqFKCwshEgkQl5eHgoKCpCZmYm8vDzpthcvXsDDwwNcLhfZ2dkwMTFBx44dMXHixBLHYoxBIGYQiIrqUCRhEEsAMWNIzBUiq0AMoYRBLGEQMwZ+Xi5CTx1F6897w8DYDEJx0fbiumSMQSRh0nYk/7UvlrxXrwDEEkAokSA9MxcmT98g/79jy9Tvf/Vcw1wfa7p6KJsWnVfZdV1YWIiMjAykp6cjPj4eiYmJSEtLQ35+PgQCgfR1u7CwEIWFhdLX3vdft4v3Kd5e/Fjxv6ampvjtt98U7htjDEIJQ6G4qP7E//0/nS/C25xCZBSIIWEM2Tk5uHryCBp36gU9Y1Ow92pVwgCRhEHCimpa9F8bIsn7tf/udVgsYUjLyoHpg0iZ1+te9awxpKGtyvOi67T99fp9jDEUFhZCKBQiLy8PCQkJiI6ORnZ2NvLz88Hn85GVlYXc3FxkZ2ejsLAQkZGR8PT0hJ6eHrKzs2FmZlbme5iy/m9qaoo//vhDGke+UIy0fBFSMrLxz9+HENC9LziGJsgXSsAXSlAgkoAvYsgViJFTKEaOQIx8oeTdewu8/x7jXX0vbKK6ZR2qu8qua4lEgpSUFKSlpaGwsBAFBQXg8/kQCAQQCATIzs5Gbm4uBAIBhEIhBAKB9DVZIpGgoKBA+lX8fry099YAUFhYiJiYGLi5uUFfXx85OTnSugberZ375Zdf4vPPPwcACMUS5BZKpO8fhBKGvEIJhJKi98FF34uRUyhBWmYWrpw4As+2X4AZmEAoZtL3EMWvyQDAUM77Y+k+wNR6DNVjXof6VaXX6/fx+Xzpe+7Xr1/j7du3yM/Ph0gkQm5uLgoKCpCbm4vHjx+jVq1a0NPTQ2ZmJoyNjSGRSCCUAAKuIQTQQwHXCCLwIAQXnw8cCQNzaxSKJRD+9z4kt1AMwX/vIfgiCWpbG2Fhxxpq/9npMrkHlkaPHl1i28iRI2W+Z4ypdC0jbbJ161YsWbKk0o+bnp6OcePG4dNPP61QkWdmZiI6OhpxcXGIjY1FVFQUXr58iZSUFOmH38TERBQUFMg8z8zMDKampjAyMpJ+GRsbw8jICPr6+tDT0wOXywWPxwOXy4Wenh709PTA4/EAFK1hVfz14felfXG5XDDGcOvWLdSrVw8WFhYwMDAAl8sttV8iCUNqvvDdHy4JkPXfQJD4vw/SWQXi9z6YMyTGxWLpuHEI4XoDVk7ILBDjYVI+YrMLpX/03mesx4WxPgcm+jwY8jjQ53Ggz+XAWJ8LHocDYaEQrjaF4HEBLocDLgfgoOj/nP/+X/Tvu++5APT1uDDg6ebvhaLkrWtVtS+RSBATE4PXr18jPj4esbGxSEhIwOvXr5GWlobExESkp6cjKytLpj0ulwtLS0vp74KBgQH09fVhYmICIyMjWFhYwNzcHHp6ehAKhQgNDYW9R23wLJzBN5JAYGGH6yJXpIbEIzFXiJisQkRnChCfU4jcQgkKxaXfCIHLAcwMeNDncsDjAjwOB5LMBCQsn45TkrrQt3aGPo8DLocDHuddLepxOdDjFm3jcTkw4BV9/36t6nE54HE5EORLYKkngqk+D0Z6HJn6Lf7X1lhf5TnSRaqu66ysLERGRuLFixd49eoVIiMjsXPnTmzduhVJSUmIi4uDRCIBABgYGMDJyQm2trYwNTWFgYGB9DXawMAABgYG0tfe91+LeTye9EtfX1/mNZrL5cLMzAxvcwoR+jAa48aNQ4K9L0TmjkhOy4PZqxgUihmS84VIyRMhXyhGgYhJPzhnC8TIKBCX2T89Lgc2xjzocTlAViLils1AQ+OGMLZ1lqlVLue/ev7vtZfHBQx4XPD+q3Xp6zHn3WszjwMY8rjgoOh3gMsBTPVL//tCyqep1+uPEQgEePDgAa5du4bw8HDcunULCQkJyM3NldnPysoKlpaWMDExgYmJifT12tzcHHw+H1evXoWjoyPMzc2l72+Kfw8Yhwsh1xCFXAMUcg1RyDNEIUdfuk3I0UMh1xAF+uYI+P0JEnILkZQrAl9U9HvJMhPB1n8NTpwrjO2c/3ufwYWxHhdGelyYGXBhbsiDuSEPJv+91yiq36L3H9zi13UU/T0Ayv59IvJRR13n5ubi6dOnePToEW7fvo2goCDs27cPz58/x9u3b6Wv18X09PRgaGgIQ0NDmJubw8zMTPq+29DQEDweT/p6/P778+Ltxf++/yXi6oOfW4CbN29Cv+EnEFm7IUvEBc/QFAKOAYRcfRRwDFHANcSVCDPg4V1kFIghKu2N8QeM9biwMORBPzcJMT/ORMBaX1g51oA+jwsTfS70il+jOf99DijxfqLk+wsA4HFEKvn5E+19vQaKJqmEhYXh1atXePPmDZ4+fYqkpCTk5ubKvO82NTWFi4sLTE1Noa+vD1NTU+iZWSGTa4Y7Fy8ivkZbiGw9UeDAg8DEDtn6VhBxZd+n8pgYekyEt0/5cLZA0ftgHgc8Dgcm+lwY6RV9zjPR48LckKt036o7uQaWduzYoa44qoyJEyeiV69eAIpmLI0YMUJjsTDGkJCQgAcPHki/Xrx4gcjISKSnpwMoGihydXWFl5cX6tSpg9atWyMvLw+1atWCs7Mz7O3tYW1tjfz8fDRo0ACPHz/+6MJcFZUjEONNpgAFIgkE4qIzy2LJu5HhvEIJ3uYU4k10NLB3L4QdJiDe1AHp2XnofzwZYkmy9IxHjkAs/fBdGr3/3njpcTmwMuLBVJ8L/f8+cCArEwCQUSCGE4+Lurb66F7HCj72RjDV50nf0Bn/9+ZOn1f+h46oqKhqc62srsjKysLu3bvx9OlThIeH4+bNm8jOzoaenh6cnZ3h6uoKZ2dn1KtXD3Z2dtIP5tbW1rCwsICNjQ0sLS1hZmZW7qC5WMJw+lUWrsXk4NrDl8CRI/jHoR84Vk7gcgB7E3246OkjJiobDqb6qGNjhE+9LOBqaQgzAy6sjHgw4P1Xh/99UOZxARtjPRh8UJcxMbbwWA6Ejm2gkt9ZqmvN4fP5ePr0KSIjIxEVFYW4uDgkJCQgJiYGr169QkZGBgDA0dERderUgYWFBQCgffv2aNKkCby8vODg4ABra2vY2toqfGInq0CExFwhUvJFiMoQ4EFiPu4n5eNVegEy+CL8sjYC3KxEAMDeh2lwczOFuFAEc1EBDHgc2Bnro56tkXRw0kifC0Ne0eurrbEeDPWKPiQXv14b8LiwNOLBzkQP3P9ijomJgcdS4MTweiqs65pKt0O0Q3R0NM6ePYvLly8jKSkJCQkJePHiBUQiERo2bIgWLVpgxowZcHd3h52dHQwMDGBiYgInJyfp701p7jx9hcOHD6P+sNnIMbZHcmIGEiQGiM8uRBpfVOK9h/5/7zUsjXiw/G9AyMmQB1dzA7hZGsDJTB9OZvqwN9WHrbEeMpJs0XQ9EDW9CTw9lJ/1GRUVpXQbRDUkEgnu3LmDixcv4vz583jx4gViY2MBAB4eHtLXMS8vL4waNQoeHh5wcnKCvb09DAwMYGhoCH195U7aZBaI8CSFj3ORWbjyJgeRGQLEZReCGSYC+A0vnAJQz8sTHpxC2FuZF9WsAQ+2JnqwMyl6f2Gox4GVkR5M9LkwN+BKT0zp84o+gBvwuDDWK9rGef/1egmwt39tlb1eE92Sm5uLe/fuYe/evbhz5w6ioqKQkZEBDw8P1K1bF+7u7hg2bJh0AMnFxQW2trYwMzODra0tYrKFOPUyE2FxubgZl4tX6QIgMxFAMMwatkfDujWBgjzUd7VDQ0cTWBvxYG+qD3ODovcWPC6d0K9Mcg0sBQYGqiuOKsPZ2bnSV79/3+3bt3HgwAFcv34dYWFhSEpKgrGxMRo0aIBGjRph4MCBqFWrFmrXrg1PT89S30iV9gEyJiZGqbjEEoZ/X2TicTIfrzIKcDU6B9GZArw/AaN4toUelwMjPS5M9blwMtOHCV8AABBKGGpZG8HVQAAXeyuZs3QmBlzUMDeAjTEPdib6MNHnQp9XdDbE3JBX4kO3bN8s4TEX2PiFp8oGzYh2S0hIQGhoKMLDwwEAfn5+sLGxgZ+fHxo3bozJkyejadOmcHV1lc62U4RIwvAirQC343MRkZiPky8zEZddiLbu5mjkYIJrAM4HeqOZtxdS46NRq1YtFfWQVHUxMTE4duwYTp48iUuXLqGgoAD29vaoVasW3NzcUKNGDbRq1QpmZmbw9/dHrVq1YG5uLn2uh4cHvvnmG4Ve0xhjSMgV4vKbbNyOz8ObTAGepRbgRVrRDFYOAGdzffg6maCZsylGNrGDrbEevKwNYZDrgNrrgDMjveHu7k4DkkTt3r59i927d2Pfvn148OABXF1d0aVLFwQEBMDe3h4NGjSAr68vrKysKtQeYwzxOULcis/F7fg8XI3OwfVHrwAAR55loLaXJeyMuGjnagUXcwPYmejByogHK6N3/xrpceQawOVkFw0ccBUc9CXa5969e/jjjz9w+PBhJCQkwM/PDw0bNsSoUaPg4+MDb29vmJubS1+v58+fr7L3oCIJw+NkPo48Tcc/zzPwMIkPBsDP2QTd61jhy6b2qGdnBEGaFdqvB25PaEiv16TSvHz5EidPnsSZM2dw4cIFFBYWolu3bhgyZAhq1aqFhg0bok6dOiWexxjDnbd5OB6biws3MnA3IQ5vc4RwtzRAB09zTG/thOYuprAUOKDeeuDwkLrv1bVL5XeUlFAtF+8ODw+XXtMMALGxsQgNDQUANG/eHIaGhpoMT4oxhoiICJw6dQpHjx4FAAwcOBBNmjRBmzZtsGbNGjRv3hy1a9dW6sOxMvFdi8nF0WfpCH6QhnyhBH7OJnC1MMCsNs5o6GCMhg7GMNYvmn1R1puwmJgYeHwDrO3qQS8QRGGxsbHYs2cP/vrrLzx69AjW1tbw9PQEAOzevRuDBg1S+veEMYbHKXw8TOLj1KtMHH+eiWyBGDXM9dHEyQRjm9pjjJ897Ez0ERMTgyAAtW2MYGmkhzT6QFHticVinD59Gps3b8apU6dQs2ZNfPHFF5g6dSratm1b4RMB8kr6bxDpRlwuLkRl402mAHlCCWyN9dDW3Qxe1ob4rJYlWtQwhbulIWyM9YouSytFDJ8uIyOVIyUlBYsXL8b27dvh5eWF4cOHY9++ffDx8ZF7Vt6r9AJciMrG5TfZuBqdg4RcISwNeWhewxQdPc0xr3FdfLEe+Pe/mXL0PoSUpqCgAHv37kVQUBDCw8PRqVMnLFmyBD179oSjo6PaBm4KxRK8TCvArfg8nInMwplXWcgWiOHrZIJBDWwR1MMCvk4mMNL7YGaz2EjlsRBSmqioKBw6dAj79u3D/fv3Ua9ePXTp0gXffvstWrduLT0xVpqsAhH23E/FljvJeJZaAG87I3T0tEA/H2v4u5mhjo2RzGt+TEx6ZXSJKKBaDiwFBQUhOTlZ+n1oaKh0YGn79u1wdHTUVGjIyMjA2bNncerUKZw5cwaJiYnw9fXFZ599hgkTJqBv376wtVX9oqNOTk549OgRnJycProvYwynXmVh2ZV4hL/NQ0dPCyzt5Ip+PtawN5V/Oq88x1aEutsnilNFbuLi4rB06VLs2LEDDRo0wLBhw/DFF1+gYcOGEIlEePnyJerUqaPwoFJURgGux+Yi5E0OzkdlITqrEA6memjnYY5tPWuinYcZHE31S3zQobquvj7MTUFBAbZs2YL//e9/SEhIwODBgxEWFoaWLVsqdNlaRXL/ODkfJ15k4sizDNyKz4O1EQ/+bmYI9LVDfXtj1LQ2RF1bI7lnUFBdV1+VlXtDQ0N899132Lx5M7y8vHD06FF069ZN7t+V2CwBDj1Ox5FnGbgem4uaVoYI8DDDyi5uaOVqito27+q/sLCQ6rqaqkhu0tLSEBQUhI0bN0IoFGL8+PHYv38/ateurZL2P8QYw/2kfFyLzsHJV1m4/DobAjGDk5k+PqtliW09a6K9hzkcTPXK/b2g1+vqqzJy//DhQ4SGhqJHjx54+PAh6tati+HDh2P//v3w9vb+aBuXXmdja3gyjj3LgI2xHsY3s8f4Zg5wMTf46LGprrVTtRxY+u233zQdQgkhISH48ccfce7cOVhYWOCzzz7DypUr0bVr10q59M7AwAANGjT46H7XY3Mw93wsbsblYZyfPfYNqA1PK+VmeFX02NraPlGcMrmRSCTYunUrvv/+e/j4+OC3337DqFGjZN5kyds+XyjBjbhc3I7PxYu0AjxO4eNWfB7sTPTQwdMc37V1Rs961nC1KP+PnrJ9qwiqa+1VnBuRSITff/8dixcvhlAoxMyZMzFmzBilTw6UlXuRhOH48wz8ejMJIdE5aOJogi/qWmHTF57wdTJRyWU4VNfVV2Xk/uXLl+jQoQPs7OywY8cO9O/fX66TAlkFIgQ/SMOBR2kIjc1FbRtDDKhvgw3dPeDrZFrusamuq6fycsMYw44dOzBz5kzY2tpi4cKFCAwMhKlp2bUkT/vFnqfycTYyC6/SBYjOEuBxMh9RGQLUtTVC11qWmNrSEc1cTOEg58lbquvqS525Kb7p0pw5cxAeHo4pU6bg0KFDqFu3boVOANxLyMOiy/E4/TITfXys8dfgOvislmWZM6U/RHWtvWhOuwYxxnDu3Dm0b98enTp1gp2dHUJCQpCSkoIDBw5g9OjR0kGlt2/fonbt2nj79q1aYimv/ZQ8IXbcS0H3vc/R/o+n8LAyxNMpjbHxC0+lB5U+dmxVUHf7RHGK5ubatWto3749vv/+e6xatQo3btxAu3btSvxB+1j7EsZwIy4Xiy7F4dPdz2D+4x10C36Ow08zIJQw9KhrhXuTGiJxVlMcHFgHX7VwrNCgkjJ9qyiqa+0VHx8PJycneHt7Y8aMGZg0aRJevXqF7777TiUzTt/PfdGaBLkYf+w1HH+5i5GHo+BtZ4SHkxvh7qSGWPaJK/ycTVW2tgvVdfWlztxkZmZi+PDh6Nu3L7788ks8fPhQrsuXX6UXYOnleDTa/BCrQhPg72aG2xMa4NmUxljR2a3cQSWA6ro6Kys3kZGR6NKlC6ZOnYqFCxfi2bNnmDx5slyDSuW1n1kgwoabiWi1/THqb3qIrXeS8TanEF5WhpjVxhmvpzfB0ymNsb6bB7rVsZJ7UKm8Y6sK1bX2UkduxGIx/v77b7Rp0wbt27fH/fv3ERISgl9++QX16tX76KBSdKYAgUci0WLbY4gkDA8mN8LBgXXQvY5VhQeVAKprbVYtZyxpGmMMJ06cwPLlyxEeHo5Ro0bhjz/+KHdKrUgkQmRkJEQi9dyKs7T2GWNYeCkeq0ITYGeihy/qWOH2hAZo6izfH1VFjl2V2ieKkzc39+7dw5w5c3Du3DkMGzYMf/75Z7mLYZbV/t2EPOyOSMVfT9KRmCtEa1cztHEzww/tXdDUyQSWRsq/NFJdVz8SiQRHjx7FwoULkZSUhL59+2LFihWwsbFR6XGKcx8WnYll/6ThYTIfnb0ssOkLT3xe2xJWKqjfjx2b6rr6UVdubt68ib59+0o/sE+dOrVCd8l6kcbHP88ycfRZOm7E5aGJowmmtXbC5BaOMNaX77wp1XX19WFuGGPYuHEjZs+ejYCAADx69Ag1ayp+Z8kP288WiPHd2RjsjEiFvYkexvrZ4/feNdHQwUQl/Snv2FWtfaI4VeZGJBJh9+7dWLFiBWJjYzFmzBisWrUK7du3r9DyMQk5hfglNAFBd5LRzNkUV8f4wN+t7DWXKhIP1bV2ooGlSiSRSHDkyBEsX74cT548wZgxY7B//37pAsPaJLdQjAnHX+Pki0wcHFgbPetZ0d1MiMbk5+dj0aJFWLduHfr06YNHjx6hfv36crdzMy4Xs8/F4mpMDjp6mmNRxxro7W2t0JlAQoqJxWIcPHgQK1asQGRkJIYOHYrHjx9j7ty5Kh9UAoACUdGtzwcfeoWxnRrh3+H1KjyTjhBtwRhDcHAwJkyYgMDAQMycORN169b96PPS8oWYfjoG+x6moZGjMbrWssKWHjXRyFH1H8xJ9ZKfn48JEybg8OHD2LRpE0aPHq3QOniluZ+Yhx/uROLoswy4Whjg2NA66OxV8ct/CNGE+/fvY9iwYYiLi8PUqVMxffp02NnZVehu4pkFIqy6loBfbyahjq0h/uxfC328rVX2O0W0Dw0sVYIPP3RMnDgR//77L2rUqKHp0Er1Mq0A/Q+8hEAswdUx9enNGtGoCxcuYPz48RCLxTh+/Di6desmfxtRWTh54yWOPM3AkEa2eDalMerY0t1SiHKEQiH27t2LH3/8EW/fvsXXX3+NGTNmQCAQYMeOHWo5ZjpfhGF/Fd0afXc/LwwLUPxMOiGaUvz7cunSJaxatQpTp05FbGzsR5+3+34q5pyLhZOZPi5/6YMAd8XPehPyvtjYWPTu3RtZWVkICwtDkyZNVNJuZHoBAKDPvpfo0dIbwf1q4fPaljDUo9VIiPYSCoX46aefsHTpUnTt2hVXrlyBnZ1dhZ9/NyEPAw68hAGPg99718TABjY0QaEaoIElNSrrQ4cid50zNjZGr169YGxsrIZI37V/+GUeFt96hPaeFtjd10utl1R8eGx1901d7RPFlZcbgUCAefPmYf369ZgyZQpWrFgBMzMzudoX8wxQo3lnfH02Ab38vBA6tj5aucrXhqKornWXSCTCzp07sWLFCmRkZOCbb77BtGnTpOsnpaSkqCU38dmF6Bb8HByuAbp+0QOf1qv8O5hSXVdfqsrN+fPnMXDgQDRp0gQPHjyQzlIqr/1CsQQzTsfgt7spWPaJK6a2cixxa3VlUF1XX8bGxvD390fPnj3RokULXLhwQWUzTbeFJ+Pbv6Ph4PcJDo5rgg4NPFTSbkVRXVdfyuTm0aNHGD16NKKjo7Fv3z74+fmVGFQqq/3cQjHmnIvFljvJGN7YFlt61JT70uSPobrWYowoLDw8nAFg4eHhMtsjIyMZY4zdvn2bWVpasgULFrDU1FRNhFiq4vjeJxRL2MRjUcxk+S224UYCE0skGoisSGnxaRNtj09ZH6vryvDkyRPm6+vLXFxc2Llz5yr0nA/je5nGZ403P2D1fo1gj5Ly1BGmXLS9brQ9PmWpoq4lEgk7ceIEq1+/PrOwsGDLli1jGRkZKo5UVnF8z1Lymee6e+yTnU9YVoFIrceUh7bXjbbHpyxteL2Wx6+//sp4PB776quvmEhUsTpOyi1kHf54wmqsvsuux2SrOcIi2vrzK6bt8SmrsupaIpGwNWvWMB6Px+bMmVPhmizL+/EtvhTLjJfdYlvvJDGJBt9Tv0/b60bb41OWtr9eJyQksK+++orxeDzWr18/lpSUxBireHxZBSLW5rdHrPb/Itil11nqDFWGtvz8yqLt8akSzcNUo+bNmyM+Ph5Lly5V+k5AeXl5OHjwIPLy8lQU3TsCkQR9dt/H/oMHcHKQO6a0cqrU6Yrq7FtltE8U92FuGGPYvn07mjVrBk9PTzx48ABdunSRu93b8blovf0xXAxFmG39BJ6mTNWhfxTVtW65d+8ePv30U/Tu3RtdunRBZGQk5s+fDysrqxL7qjo3T1L46LjzKZq5mOLE8HrgiQo0lnuq6+pLmdwwxrB8+XLMnDkTwcHBmDVrVok7vpXW/qPkfLTe/hhCiQS3JjRQasHX8lBdVz9CoRATJ07EvHnzMHXqVMyfP7/CdyH8mGVX4vHztQQcGVIHw71NcejQIXq9JpVKntyIRCL8+OOPqFWrFsLCwnDixAn89ddfcHBwqHD72QIxugc/R7ZAgtCx9dHR00JlffnYsata+7qMBpbUTN7bkpYlLS0NgwcPRlpamkraKyYQSTDw0Cvci0pAVvBc1DQUqLT9ilBX3yqrfaK493OTl5eHwMBATJ06FevWrcPhw4cVGpC9FZ+Lz/Y8x4D6NgjqbIMxI4dpJPdU17ohNjYWgYGBaNasGaysrPDkyRP873//K3etAVXmJipLhC67nqGDpwX2D6gNIz2uRnNPdV19KZobxhjmzJmDFStW4OjRoxgyZEiF2j8flYV2fzxFazczXAj0gYu5+haop7quXjIzM9GtWzccO3YMBw8exPr161WWm19CE/Dj1bc4PKQOuta2otdrohEVzc3Tp0/Rpk0brF27Flu2bEF4eDi6du360QW2329fIJKg158vkC0Q4/wob7XfEIfqWnvRwFI1JhRLMPTvSNxPzMf+AbU0HQ6pxiIjI9G6dWuEhoYiLCwMEydOVOiuEUm5QvTZ9xKDGthgcw9PWiiQKCw7Oxvz5s1D3bp18fz5c1y9ehV//fUX6tSpU2kxpOQJMfZCNvzdzLCnrxfdPYhUORKJBJMnT8bmzZtx6tQpdO/evULPu/ImG733vcDkFg7Y26+WStdTItVbZGQk/P39kZycjFu3bsHX11dlbe9/UYD5F+Owf0BtfF7bSmXtEqJqYrEYa9euRdOmTeHi4oLHjx9j5MiR4HLle60VSxjG/BOFN5kCnBlZD45mdJfl6oz+UldjU09G42ZcLs6NqgcPK7pDFtGcHj16oFatWggPD0fTpk0VakPCGAKPRsLTygAbu3vQoBJRiFAoxObNm1G7dm0cOHAAu3fvRlhYGNq2bVupcRSIJOh34CVsjLgI7lcL+jz6c02qFqFQiJEjR+LAgQO4cOECOnbsWKHnPUjK+29QyRHLP3GlW1MTlQkNDUXr1q3h5eWFa9euwd3dXWVtX3qdjSW3crGtpyd6e1urrF1CVO3p06cICAjAsmXLsG3bNhw5ckShG0sBwKyzMTgXmY1/h9eFsxpnlZKqQWvuCsfn8xEcHIzQ0FDk5OTA1dUV/fv3R/v27ct93rx58/Do0aMyH9+1axesra3L3bdp06ZYsmSJch2oYm7E5eK3uym4OsYHdW2NEUOXkZJKJhaLsXz5cgDAtGnTsGLFCqU+QGx/zMfNOAHuTmxIH8KJ3BhjOHbsGL7//nukpKRgwYIFmDx5MgwNDTUSz/dnYxCbVYgDn1mo/I4qhKhbYWEhhgwZgrCwMISEhKBhw4YVfm7g4SgM8q+PVZ+60aASUZm9e/dizJgxmDRpEtasWQM9PdV9BIrKKMCgQy8x2scYgb72KmuXEFUSiURYvXo1Fi9ejK5du+Lvv/+Gi4uLYm1JitYuvfImGxenNUNDBxNVhkqqKK0ZWFq5ciVevnyJwMBAuLi4ICQkBKtXrwZjDB06dCjzeZMmTUJ+fr7MNoFAgCVLlqBWrVrSQaViTk5OmDFjhsw2eW9hrgmOjo64ceOGwiPK75Mwhm9PR2N4Y1vpQpiqbF9e6j62JvtGSpebm4vhw4fj6tWr2LRpE8aOHavUB4jLb7KxLiIffw6ojZrW7wYCqK5JRV2/fh2DBg3C1KlT8cMPP5T42yEPZXPzJIWPLXeScWJ4PdhzSl7jT3VNNKGiuSksLMSgQYNw+/ZtXLlyBXXr1q1Q+0aWtvCctQcNfDwQ1MOzUgeVqK51F2MMixcvxooVK7B+/XpMmTJF5nFlcyMQSTDo4Cu0rGGGWU1Lfqyi12uiCR/mJjIyEsOHD8fLly/xxx9/YOjQoUq9xq65XwCbKTtx9qs2aOxYuYNKVNfaSysGlu7cuYOIiAjMnDlTOojUuHFjJCcnY8eOHQgICCjzTg2lTWO9cOECRCIRPv300xKPGRgYwNvbW7UdqASGhoZo1aqVSto6EiXA42Q+/hr0bq0QVbYvL3UfW5N9IyXFxcWhZ8+eyM/Px61bt1C7dm2l2kvIKcSwvyIR6G2EAfVtZB6juiYV1aZNG7x+/Vrhs3fvUzY3s8/F4NNalvi0liWiokoOLFFdE02oSG7eH1S6fPlyhdckE0sYRv8bBwuvJtg32Ae8Sl5PjOpaN/H5fIwePRqnTp3C8ePH0a1btxL7KJubeRdikZwnxJmR9ZCVGKvy9pVBdV19FeeGMYadO3di6tSpaNeuHR4/fgwnJyel2j75MhPb72fh4nf94KemO3WWh+pae2nF/PobN27A2NgYAQEBMts7d+6M9PR0vHjxQq72zp07B2NjY7Rr106VYWpUQkICfH19kZCQoFQ7kekFWHY7D0s7uaKGxbtrYVXVviLUfWxN9o3IunfvHlq1agUrKyuEhYXB1NRUqdwwxvDlP1GobWOIWX4l78BIdU0qisPhqGRQCVAuNxeisnD6VRZWfeqmlvaVRXVdfX0sN4oOKgHA7HOxuPn0DYSbRyA3PVlVIVcY1bXuSUhIQMeOHXHr1i2EhYWVOqhUvJ+iuTn5MhO/3kzC7n61YGtS+qLF9HpNNCEhIQGNGjVC7969MWnSJPz44484ceKE0oNKKXlCjPvnNSb78DClVzuqayJDKwaWoqOj4erqWmJWUs2aNaWPV9Tbt2/x5MkTtGvXDsbGxiUeT0xMxLBhw9CnTx9MmDABe/bsgUAgUK4DlUAoFOL+/fsQCoUKt1EolmD435FoZq+Hb1rLTu9TRfuKUvexNdk38s6xY8cQEBCAzz77DGfOnIGNjY3SuQl+kIar0TnY0ccL+qWc4aa6JpqgaG4kjOH7c7EY52ePBuWsV0B1TTShvNyIRCIMHz4ct27dkntQ6e8n6fj1ZiI2d3PF00cPqK6J0iIiItCyZUvo6+vj1q1baNCgQZn7KpqbvEIxxv3zGnMCXNDR00Ll7asC1XX1FRoaikePHuH58+e4ffs2pk6dqvTlxYwxTPr3DWpY6GNKc1uqa1KCVlwKl5OTU+p1jMVrH+Xk5FS4rXPnzgEAunTpUuKx+vXrIyAgAK6urigsLER4eDgOHz6MJ0+eYMWKFXLfYrGqWXwpHjFZhfinmzndMYtUGsYY1q9fj++++w7Lli3DnDlzVLJ2RnKeEDPORGNJR1fUsjFCVKbysRKiSSdeZOJpCh//DqvYmjSEaAOxWIzAwECEhITIPagUnSnAhOOvsaSTK/zdRGqMklQX//zzD4YPH45+/fph+/btarsBw6ZbSdDjAj+0V81MV0JUQSKR4Oeff8b8+fMBAMePH6/wOncfE/wgDadeZuL2hAYwKEhVSZtEt2jFwBKAcj9oVvRDqFgsxsWLF+Hu7l7qOkojRoyQ+b558+ZwcHDAjh07cPPmTfj7+3/0GAkJCdKpcU+fPq1QXNrgVnwuVl9PwNGhdWHLS9d0OKSaEAqF+Oabb7Bz507s27cPAwcOVFnbM87EwMPSENP9lZvWS4g2YIxh5dW3+LKpPd2yl1QZEokE48ePx5kzZ3Dp0iX4+PhU/LmMIfBIFPycTfF9W2fEx5Vcn4aQimKMYfXq1Zg7d65KT2KVJlsgxi/XE7C0kyuM9HT7pDSpOtLS0jBy5EiEhoZi06ZN+Oqrr2BkZKSStmOzBJh2KhrLPnFFAwcTxMSopFmiY7RiYMnc3LzUWUm5ubkAKn7Xtjt37iAjIwP9+vWr8LE7duyIHTt24Pnz5xUaWNq6dSuWLFlS4fZVxdDQEJ06dVLozItAJMHYf15jRBM7dK9jhaiokgNLyrSvLHUfW5N9q87S0tIwcOBAPHnyBJcvXy51ITxFc3PnbS72P0zDjfH1oVfOIq9U10QTFMnN5Tc5uPM2D3v7f3wxe6progkf5kYikWDSpEk4cuQILly4gEaNGsnVXtDtZEQk5uHh5EbgcTlU10RhhYWFmDRpEvbv348DBw6gf//+FX6uIrn59UYizA14GOtnr5b2VYXquvq4fv06Bg8eDAcHB9y9exdmZmY4ePCgSnLDGMO4Y6/R2NEE01sXncyluial0YqBJQ8PD1y9ehVisVhmnaU3b95IH6+Ic+fOQU9PD506dZI7hoqe1Zg4cSJ69eoFoGjG0oezoNTF0dERFy9eVOi5y67EI4MvwprPSt5BTxXtK0vdx9Zk36qrp0+fomfPnrCwsMDt27fh5lb6QsSK5mbhpXj0q2+N5i7lDzpTXRNNUCQ3K6+9xdBGtqhp/fE3MlTXRBPez41IJMKYMWPw77//4uzZs/Dz85OrrdgcMeaej8Waru5wszQs0X5lo7quulJTU9G/f3+8evUKV69eRbNmzeR6vry5yeCLsDYsEb985g4D3sdnK1FdE3VijGHt2rWYM2cOJkyYgDVr1khnKakqN3++KEBYLB8RXzWU3rGT6pqURivmb/r7+4PP5+P69esy2y9evAgbG5sKXRuakZGB8PBwtG7dGhYWZS+i96HiwqlXr16F9nd2doafnx/8/PzkmvKtLD6fj5MnT4LP58v1vJgsAVZfT8SG7h6wNi57HFHR9lVB3cfWZN+qo9OnT6N169bw9fXF1atXyxxUAhTLzbWYHJx9lYXFHV0/ui/VNdEEeXNzKz4XF6OyMTugYmt1UF0TTSjOTVZWFoYMGYLTp0/j8uXLaNGihdxtLb+Th1auZhj33owPqmsir2fPnqF169bIy8vDrVu35B5UAuTPzfobibAz0cOoJrZqaV+VqK51W0ZGBvr27YvFixcjODgYmzZtkg4qqSo3rzMEWHU3D6s+c4eX9bvL6qiuSWm0YmCpWbNm8PX1RVBQEM6cOYMHDx5g48aNuHv3LkaPHi2dxfTrr7+iT58+SE4ueSvaixcvQiwW49NPPy31GI8fP8aiRYtw+vRp3Lt3D7du3UJQUBD27NmDxo0bo2XLlmrto7JSUlLwxRdfICUlRa7nrbqWgCZOJujjba2W9lVB3cfWZN+qk+JFunv06IHp06fj4MGDMDU1Lfc58uaGMYYFF+MwrLEt6tuXvOujsu2rEtV19SVvbn4JTUAvb+sK1bQi7asS1XX1VZybfv36ISwsDCEhIWjcuLHc7Vx6nY3L8YVY97m7zGxxqmsij3PnzklPYoWEhKBGjRoKtSNPbrIFYmy8lYQ57VygX4HZSvK2r2pU17rr4sWLaNKkCV6/fo3w8HAMHjxY5nFV5IYxhonHX6OJnT4mNpO97JPqmpRGKy6FA4C5c+ciODgYf/75J3JycuDq6opZs2ahffv20n0kEgkkEgkYYyWef/78eTg4OMDX17fU9q2trcHj8XDgwAFkZ2eDw+HA2dkZw4YNQ9++fXXyjnDx2YX4/V4KDgysrbYFDAkBitY3+PrrrxEcHIy9e/eW+AOnKhdeZyM0JgdPp8j/YYYQbRSTJcDRZxm4GFh5M2AJUUR+fj4A4MWLF7h69Sq8vLzkbkPCGL47G4OBtQzR0MFE1SGSauLcuXPo1q0bZs+ejWXLllXae/itd5JhZsDDiMYVm61EiKrx+XzMnTsXGzZswLRp07BixQoYG1fspJS8/riXihtxuTj+hSV9jiQVojUDS8bGxhg/fjzGjx9f5j7Tp0/H9OnTS30sKCio3PZdXFywcOFCZUKsclZfT4CPnTF61rXSdChEh324vkHz5s3VchzGGJZejsdoX3vUslHNXS4I0bTNt5LR2NEEAe4Vu0kFIZqQl5eHUaNGAQAOHTqk0KASAOy5n4qX6QXYFGClwuhIddOuXTscPXoUPXr0qLRjFogkWH8jEbPbOldobSVCVO327dsYNWoU8vPzceHCBXTs2FFtxyoQSbD4chwWdKgBN3O6JIxUDL0y6qjE3EJsC0/GD+1daJSZqM2jR4/QsmVL5Ofn49atW2obVAKAqzE5CIvLxewAZ7Udg5DKlC8U47e7yZjaypFep4nWys/PR48ePZCamgqg6ESdQu0IxVhwMQ7ft3WGvTG9/SSKMzIyqtRBJQDYGZEKkYRV6E5whKiSUCjEokWL4O/vD39/fzx8+FCtg0oAsD08GSIJw+QWDmo9DtEtWjNjiZTPwcEB586dg4NDxX7B115PRC1rI/T1KX9tJUXbVyV1H1uTfdNl//77L4YOHYovvvgCf/zxB0xM5L+sQZ7crLyagMENbeWarUR1TTShornZ+yANPC4HQxrKd1kF1TWpLPn5+ejZsyfi4uJw7tw5vH79WuHcrL2eCAD41t8JibHRJR6nuiaaUJHciCQMq0MT8E0rR5ga8MrcT9H21YXquup78uQJRo4cibi4OBw+fFh6Z/KPUSY3fKEEP11LwHdtnMusd6prUhoaWKoijIyM0KVLlwrtm5wnRNCdZPzWqya4FTwLLk/7qqbuY2uyb7qIMYY1a9Zgzpw5WLRoEebPn6/wbIuK5ib8bR7ORmYhYlJDtbSvDlTX1VdFcsMYw4abSZjQzAFGevLN3qC6JpWhoKAAffr0wZs3b3DlyhW4urqiVq1aCrWVkFOIVaEJ2NDdAyb6pX9QobommlCR3Bx8nIaUfCEmt3BUS/vqQnVddUkkEvzvf//D3Llz0b17d5w+fRr29hWfLadMbraFJ4MxYFI5s5WorklpaC5yFZGUlISAgAAkJSV9dN811xPgaWWIAfVt1NK+qqn72Jrsmy46efIkFi1ahAMHDmDBggVKXcJT0dysvPYWPetZoZGjfLOiqK6JJlQkN5fe5OB5WgEmNZf/jBjVNVE3gUCAvn374uXLl7h06RJcXV2Vys2iS/Goa2uEkU3sytyH6ppoQkVysy4sERObOcDaWP7z8VTXRF5v3rzBJ598gsWLF2Pbtm34+++/5RpUAhTPTb5QjFWhCfi+rXOZJwGUaV8VqK61Fw0sVRECgQChoaEQCATl7pecJ8Tm28mY394FPG7FP/BXtH11UPexNdk3XdS9e3c8fPgQ/fv3V7qtiuTmSQofR59mYE6A/Ot6UF0TTahIbjbcTET/+taoYWGglvbVhepa9wkEAvTv3x9PnjzBpUuX4O7uLt2uSG6epfKxIyIFP3/qVu4saqprogkfy829hDzcTcgvd/aGMu2rE9V11cIYw44dO9C4cWPweDw8fPgQo0aNUugErqK5CbqdDA6AiR856UV1TUpDA0s6RpHZSoTIg8PhKHxHIEX8EpqATjUt0NqV7ppFdMObTAH+fZGJqS3lv6yCEHUqLCzEoEGDcP/+fVy8eBGenp5Kt7n4cjw61bRAZy9L5QMkpJL9cS8FnTzN4WVNd6Ml6nXx4kVMnjwZy5cvx7lz56SD+pUlRyDGz9cSMK+dC4z1aYiAyI/WWNIhKf/NVvqtV025ZisRoq2yBWIcfJyGfQNqazoUQlQm6HYSmjqZ0GAp0SrZ2dkYNmwY7t27h8uXLyu8ntL7IhLz8NfjdFwfV18FERJSufhCCf58mIZfu3loOhRSDXzyySd48eIF3NzcNHL8X28mwtSAS3c+JAqj4cgqwsDAAC1atICBQdmXTay5nqjwbKWKtK8u6j62JvtGyvex3Pz9JB1mBjx0q63YmW6qa6IJ5eUmXyjG73dTMLmlo8Lrk1FdE1V78eIFWrdujaioKFy5cgV16tQpsY8iuVlwMQ4961mhZY2PD6JSXRNNKC83R59lgDGgn4/iVwFQXZOK4nA4KhtUkjc3GXwR1lxPxPz2NWBYgRuKUF2T0tCMpSrCyckJt27dKvNxvlCC7XeTse5zD4VmK32sfXVS97E12TdSvo/lZu+DVAxpaAt9nmJj4FTXRBPKy82+h+ngcTkY0tBWLe2rG9W17jl16hSGDh2KDh06YM+ePbCwsCh1P3lzExabg1Mvs3CvgnfzpLommlBebnbcS8HQRrZKXRZEdU00Qd7crL+RCDsTPYxqUrH3JlTXpDQ0Y6mKKCgoQEhICAoKCkp9/PDTdDAGDFRwbaWPta9O6j62JvtGyldebmKzBLj8JgcjGiv+AZzqmmhCWblhjGHTrSSM87OHUQXOCMrbfmWgutYdjDH8/PPP6NmzJ6ZNm4YjR46UOagEyJ+bRZfjMbihTYXv5kl1TTShrNy8zhDgwutsjGmq3GVBVNdEE+TJTbZAjA03kzC3nUuFT+RSXZPSaM3AEp/Px/bt2zF69Gj0798f06ZNQ0hIyEefd+HCBfTq1avUr4yMjBL7R0RE4LvvvsOAAQMwfPhwrF+/HpmZmWrokWolJyejQ4cOSE5OLvXxHfdSMLyx4mdVPta+Oqn72JrsGylfebn582Ea6toaobmLqVraVzeq6+qrrNxci8nFw+T8j95tRdH2KwPVtW7Iz8/HsGHDsGzZMhw8eBBLliwBl1v++wd5chMWm4OLUdmY375GhWOiuiaaUFZudkakoImjCfycKzYwKm/7lYHquvqSJzdb7yTD3JCH4XKcyKW6JqXRmkvhVq5ciZcvXyIwMBAuLi4ICQnB6tWrwRhDhw4dPvr8adOmoUYN2Tcw5ubmMt8/evQIS5YsQfPmzfHDDz8gKysLu3btwoIFC7B27Vro6+urtE+V5VV6AS69ycHqrpV79wBC1IUxhuAHaRjR2E7hdWgI0TYbbyWht7c13C0NNR0Kqcaio6PRp08fZGdnIywsDI0aNVL5MZaHvMWABjbwsTdWeduEqJtYwrAzIhXftXGi9yBEpxWIJFh/IxFzApxhoOCyE4QU04qBpTt37iAiIgIzZ86UDiI1btwYycnJ2LFjBwICAsDj8cptw93dvdTFJt+3Y8cOuLi4YM6cOdL2HB0dMXv2bJw7dw7du3dXTYcq2Y57KWjmbAJfJ8VndhCiTe4l5uNJCh/DlLgMjhBtEp9diCNP03F2lLemQyHV2OXLlzFw4ED4+vriwoULsLFRfFHistyOz8WZVxVfW4kQbXMuKgvJeUIMa2yn6VAIUaudEakQSRjdCY6ohFYMTd64cQPGxsYICAiQ2d65c2ekp6fjxYsXSh8jLS0NL1++RKdOnWQGqXx8fFCjRg3cuHFD6WNogkjCsCsiFV8qeQ04Idok+H4q2nuYw9OKZnYQ3bAtPBk+9sbo4GH+8Z0JUTHGGDZu3IhPP/0Uo0aNwqlTp9QyqAQAK66+RW9v6wqvrUSIttkVkYo+3tawMdaK8++EqIVIwrA6NAHftHKEiX75EzgIqQiteMWMjo6Gq6triVlJNWvWlD7u4+NTbhvLli1DdnY2TExM0LBhQwwfPhweHh4yxwAAT0/PEs/18PDAs2fPlOyFetnZ2eHIkSOws5M9e3LqZSYyC8QY2ki5mR1ltV8Z1H1sTfaNlK+03IgkDPsfpWPpJxVfm0Oe9isL1XX19WFuJEy1l1VQXRN5CAQCTJ48GXv37sUff/yBkSNHKtRORXJzLyEPx59n4s6EBmppX12orquvD3OTVyjGvy8y8Wf/WmppvzJRXVdfFcnNgUdpSMkXYnILR7W0ry5U19pLKwaWcnJy4OhYsqjNzMykj5fFysoKgwYNQr169WBsbIzo6Gj8/fff+O677/Dzzz9LB6eK2yhu833m5ubIzs5WRVfUxsTEBH369Cmx/fd7KRjQwBpWRsqlsqz2K4O6j63JvpHylZabc5FZyCwQYYCCdzj8WPuVheq6+vowN5ff5CAxV4ghSp4AKKv9ykR1XbUkJCSgX79+iIuLw7Vr19C8eXOF26pIbn68+hY96lqhqbP8l+ZTXRNN+DA3/77IhD6Xg89qWaql/cpEdV19fSw3jDGsuZ6ISc0dYK3AzDyqa1IarbgUDkC5Z3HLe6xZs2YYMWIEWrRogYYNG+KLL77AypUrweFwsHfv3gq3VdGzyAkJCbh79y7u3r2Lp0+fVug5qpCcnIzPP/9cZoX6tzmFOPkiE2ObKneHobLaryzqPrYm+0bKV1pu9jxIRc96yg+WltV+ZaG6rr4+zE3wg1R0r2MJOxPV3CCC6ppUVGJiIszNzXHnzh2lBpWAj+fmVXoBjjzNwJwAZ7W0r05U19XXh7k5+Dgdvb2tYainmo9IVNdEEz6Wm5DoHDxKzseUlvLPVqpI++pEda29tGLGkrm5eamzknJzcwGUPsuoPI6OjvDx8cHz589ljgGUPvspJyenxB3kyrJ161YsWbJErnhUoaCgAGfOnEFBQYF02+6IVNSyMUKAu3w/n4q2X1nUfWxN9o2U78PcZPBFOPo0A38PLn8hfkXbr0xU19XX+7nJKxTj7yfp+KO3l1rar2xU11VL06ZNcfbsWZW09bHcrA9LhL+bGfzdFFtHjOqaaML7uckWiHHqZabK3oN82H5lo7quvj6Wm19vJqGvjw3cFLxLLdU1KY1WzFjy8PBAXFwcxGKxzPY3b95IH5cXYwxc7rvuFbdRvNbS+6Kjo+Hu7l6hdidOnIjw8HCEh4cjODhY7rhUhTGGnREp+NKXbsdOdMf+R2mwNdFT2RR0QjTt6LMM6HE56FHXStOhEKI2qflC7IxIxQx/J02HQojCjj/PgKkBF128LDQdCiFq8yZTgGPPMzC1lWKzlQgpi1YMLPn7+4PP5+P69esy2y9evAgbGxvUrVtXrvYSExPx7Nkz1KtXT7rN1tYWdevWxeXLl2UGsJ49e4b4+Hj4+/tXqG1nZ2f4+fnBz8/vowuKq9P12FxEZggwogktLEZ0x86IVIxsbAcelwZLiW4IfpCGwQ1sVXZZBSHaaMvtZNSw0EevetaaDoUQhR18nI6+3jbQ59HrNdFdm28lwdfJBG3dlL/ihZD3acWlcM2aNYOvry+CgoKQn58PZ2dnhISE4O7du5gxY4b0bnG//vorLl68iG3btsHBoWhdoQULFqBBgwbw9PSULt59+PBhAMDw4cNljhMYGIiFCxfi559/Rvfu3ZGVlYVdu3bBw8MDXbp0qdxOy0lPTw8+Pj7Q0ytK2a6IVHxWyxIu5gZqab8yqfvYmuwbKd/7uXmUnI87b/Owp5/qLhmiuiaaUJyb1AIJzkdl4cqXqj0JQXVNNKGs3BSIJNh4KwmLOtZQ6qQA1TXRhOLc5ImAM6+y8O9w+U5mV7R9qmtSmcrKTV6hGL/fS8Haru5KXfFCdU1KozU/sblz5yI4OBh//vkncnJy4OrqilmzZqF9+/bSfSQSCSQSCRhj0m0eHh64evUqjhw5gsLCQlhaWqJx48YYPHgwatSQvV15o0aNsHDhQvz5559Yvnw5DA0N0bx5c3z55ZfQ11fNoqrq4uLigidPngAoelE4+DgN23up7gP4++1XNnUfW5N9I+V7Pzdrz8SgjZsZ6toaq6X9ykZ1XX0V52bN9QTUtDKEv6tqzwpSXRNNKCs3e+6nQsKAQF/lZlBTXRNNKM7NzogUWBnx0NFTtZfBUV0TTSgrN8EP0qDP5WBwQ+XuUkt1TUqjNQNLxsbGGD9+PMaPH1/mPtOnT8f06dNlto0bN06u4zRt2hRNmzZVJESNKiwsxKNHj9CwYUMceZYNPS4HPetZqaV9AwPVzILSlmNrsm+kfMW5qedTH3sfpmLZJ65qaZ/qmlSm4tzsCgdGNnFU+Tp4VNdEE0rLjYQxrAtLxOQWDjDR56m8/cpCdV19Fedm/0N99K9vAz0VX4pPdU00obTcMMaw8VYSJjZ3gJGSl+dTXZPS0EXEVURiYiKaNWuGxMRE7IpIxdBGtkq/KJTVfmVT97E12TdSvuLc7At7idxCCQY1UO4MSlntU12TylScm0dRcRjeWLU1/X77VNekMpWWm5MvM/EmU4DJCt6y+mPtVxaq6+qrODcXH7zBoAY2amuf6ppUptJyc+F1Nl6kFWBicwe1tF9ZqK61Fw0sVTGxWQJcfJ2t9JRzQrTJwSdp6F/fGhaGyp3xJkSbtKhhAi9rI02HQYjarA9LxMgmdnAw1e7lBAj5GBtjHgLczTUdBiFqs+FmEgbUt1bZ+ryEfIgGlqqYw08z0NDBGM2cTTUdCiEqcykqG6N97TUdBiEqIZIUrQPY10f1Z78J0RaPkvNx6U0OpqhgthIhmta9jhXdkZborKiMApx4kUmv10StaGCpivnrSToCm9ipfM0OQjTJxcIA7T3oTCHRDVejswEAX9Sx0mwghKjRpltJ6OhpjkaOJpoOhRCFpeQJAQA9VLhuKSHaZvOtZDRzMUVrFd9MhJD30cBSFWFra4t5a7fhrcgYwxur/jI4W1tb7NmzB7a2ql8PRNPH1mTfSPlsbGzgGvgjRrWuBa4aBkuprokm/BvL0GrKanjVUM+ZQaprognv5yaDL0LwgzR8rcKz31TXRBOupnBgO2w5Pm3orpb2qa6JJryfm7xCMXZEpODrlqq7mQjVNSmN1twVjpTP1NQUb2t2RncnERzNVL+WgampKUaMGKHydrXh2JrsGynf0ywgoWZnjGutnjd0VNeksiXlCnH8tQDHp4+Dqal6Llmmuiaa8H5utoUlwNZYD73qWaul/cpGdV19HXrJx5jAkTA3U89MDqprognv52brnWToczkYrMLF6amuSWloxlIV8To+EcHzx6Ovej5/IyUlBQMGDEBKSop6DqDBY2uyb6R8my8/h+2/C2AizFZL+1TXpLLtjEiBCy8XW2aP1cncU11XX8W5SUxKxubbyZjU3EGlt2anuiaVLSIxD2HP4/Bg4zSdzD3VdfVVnJvk5GRsupWE8c0cYKjCu4lTXZPS0MBSFfH3/QSIHl9EgIt6VvLn8/n4+++/wefz1dK+Jo+tyb6RsuULxTj8MBHJd87qZO6prqsfCWPYFp6MYd5mOKyjuae6rr6Kc3PyaRLiswsxrplqb7hAdU0q27Y7KWjrrI8zx4/qZO6prquv4txcfJGCZ6l8TGzuoJb2qa7J+2hgqYo48DgNAGDAo5QR3fDXkwzQDViILjkbmYW3OUIMUOF0c0K0za6IFAxpZAs7E9Vflk9IZckRiLH3YSqGNVL9uqWEaItd91PR18cGrhbqmZhAyPtolKIKuJ+Yj/C3+ZoOgxCV+uNuCvp6q259DkI0bcudZAyobwMbY/rATXTX1ehcumU1qfL2PkyDmQEPn9Wy0HQohKjNucgser0mlYYGlqqALXeS0MbDAm5ubuDxeGo5Bo/HU2v7mjy2JvtGSvc8lY+rMTkY2sRBZ3NPdV29xGYJcOJFJiY2d9Dp3Oty30j5eDwezOxd0NzVHH7Oql+YnuqaVBbGGLbdScaYpvYwMtDX2dxTXVdfPB4PFvYu8HYwRYC76hemp7ompdGau8Lx+XwEBwcjNDQUOTk5cHV1Rf/+/dG+fftyn3f9+nWEhobi5cuXSE9Ph5WVFXx8fDB06FC4uLjI7Dtv3jw8evSoRBtNmzbFkiVLVNofVckWiLH3QRr+6NMYA6bHqO04NWrUQEyM+trX5LE12TdSuh33UtGyhik+9a2ts7mnuq5efr+bgvr2xmjrZgYOx1xnc091XX1Z2DmB9+0/mNG1plrap7omleVmfB4eJOXjyJA6qGFlqLO5p7quvmwcnGEw6xi+7eIODkf1605QXZPSaM2MpZUrV+LixYsYMmQIFi1ahDp16mD16tW4cuVKuc87fPgwBAIBBg0ahMWLF2P48OGIiorCt99+W2pRODk5YdWqVTJf48ePV1e3lBZ8PxUWhjx09zLFixcvIBQK1XIcoVCo1vY1eWxN9o2UJBRLsOt+CsY0tdfp3Oty34gsoViC3+6mYGIzB3A4HJ3OvS73jZRvR3gCDLPi0LO2em7LTnVNKsvWO8noVscSHlaGOp17Xe4bKV/wvUSIU6Ix0Fs9l3pSXZPSaMXA0p07dxAREYFJkybh888/R+PGjTFlyhT4+vpix44dEIvFZT53/vz5mD9/Prp06YKGDRuiU6dOWLp0KYRCIf75558S+xsYGMDb21vmy9XVVZ3dUxhjDFvuJGOcnz1Sk5NQr149JCQkqOVYCQkJam1fk8fWZN9ISf++yERuoQSDG9rqdO51uW9E1vEXmcgSiDG8sS0A3c69LveNlK1QLMHqMw+R/HNfpCUnqeUYVNekMqTzRTj4OE16lyxdzr0u942UTcIYVp99hIzV/ZGZSq/X2ta+LtOKgaUbN27A2NgYAQEBMts7d+6M9PR0vHjxosznWllZldhma2sLW1tbpKamqjrUSnU1JgfPUvkY10y1t4gkRJN+v5eCQQ1sYGFI1y4T3bD1TjKGNrSFpZHWXF1OiEr9fjcFuYUSTYdBiNL23E+FvYk+utW20nQohKjFX0/SEZddqOkwSDWkFQNL0dHRcHV1LbFIVs2aNaWPyyMxMREpKSlwd3cv9bFhw4ahT58+mDBhAvbs2QOBQKB48Gq05U4yetazpltEEp0RmyXAmVdZGOtnr+lQCFGJV+kFOB+VLT37TYiuKRBJsPLqW6pxUuUxxrD1TjLGN7MHj6v6dWcI0TQJY1gR8hYj/5tBTUhl0orTqzk5OXB0LHkrRDMzM+njFSUWi7FhwwYYGRmhV69eMo/Vr18fAQEBcHV1RWFhIcLDw3H48GE8efIEK1asAJerFeNsAICkXCEOP8nA8WF1NR0KISqzKyIV9WyN4O+qnjU6CKls2+4ko4WLKZq5qP4uWYRog+3hySgUM4xqYoefNB0MIUq4Ep2DV+kFGNOUTm4R3XTkaQYi0wX4o6MDtmo6GFLtaMXAEoByV6yv6Gr2jDH8+uuvePz4MebMmQN7e9k/HCNGjJD5vnnz5nBwcMCOHTtw8+ZN+Pv7f/QYCQkJ0msunz59WqG4FPHHvRR4Whmgs1fRomvW1tYICgqCtbW1Wo6n7vY1eWxN9o28I2EMOyJSMKWlo/R3Wpdzr8t9I0UKRBLsjEjFT5+6yWzX5dzrct9ISflCMVZeTcDsAGe4OprpbO6prquHLXeS0dvbGs7m764E0OXc63LfSEkSxrA8JB4Tm9ujrquNzuae6lp7acXAkrm5eamzknJzcwG8m7lUHsYYNmzYgCtXrmDatGlo3bp1hY7dsWNH7NixA8+fP6/QwNLWrVuxZMmSCrWtKLGEYVt4Mr5p5Qjufx/Azc3NMWnSJLUdU93ta/LYmuwbeefi62zEZwsxorGddJsu516X+0aK/P0kHSIJw+AGNjLbdTn3utw3UtKW28ngcoBJzR1hrM/V2dxTXeu+pFwhjjzNwInhslcC6HLudblvpKRjzzPxIq0AJ4fXg7m5gc7mnupae2nFtV8eHh6Ii4srcfe3N2/eSB8vT/Gg0oULFzBlyhR06tRJ7hgqOitq4sSJCA8PR3h4OIKDg+U+TkWcfJmJ5DwhAn3fzbhKS0vD6NGjkZaWppZjqrt9TR5bk30j7/x+NwW9vK1gb6ov3abLudflvpEi28KTMaqJHUwNZNcH1OXc63LfiKzcQjF+Dk3A3HYuMNbn6nTudblvpMiO/64E+KSm7O3XdTn3utw3IosxhuVX4jHezwHO5gY6nXtd7ltVpxUDS/7+/uDz+bh+/brM9osXL8LGxgZ165a9zhBjDBs3bsSFCxcwefJkdOnSRa5jX7x4EQBQr169Cu3v7OwMPz8/+Pn5wcfHR65jVVTQnWQMbmgLG+N3E8ry8vKwa9cu5OXlqeWY6m5fk8fWZN9IkdR8IY4+y8DYD9Y10OXc63LfCPAwKR/XYnIxoZQFjXU597rcNyJr480kGOlxpTdb0OXc63LfSNGVANvvJmNCMwfplQDFdDn3utw3Iuv4i0w8SeHj+wBnALqde13uW1WnFZfCNWvWDL6+vggKCkJ+fj6cnZ0REhKCu3fvYsaMGdK7xf3666+4ePEitm3bBgeHojfz27Ztw7lz59ClSxd4eHjg2bNn0nb19fVRq1YtAMDjx49x8OBB+Pv7w9HREUKhEOHh4Thz5gwaN26Mli1bVn7HSxGZXoCzr7IQNq6+pkMhRGX2PkiDk5k+unhZajoUQlTiXGQW2nuYo769saZDIUTlsgVirL6egBWd3WCkpxXnIAlR2NnILCTkCBHoa/fxnQmpYopnK43zs4eLOd1JnGiOVgwsAcDcuXMRHByMP//8Ezk5OXB1dcWsWbPQvn176T4SiQQSiQSMMem227dvAwDOnz+P8+fPy7Tp4OCA3377DUDRQlw8Hg8HDhxAdnY2OBwOnJ2dMWzYMPTt21dr7gi3LTwZfs4maFGD7ppFdANjDL/fTcGXvnR7X6I7ZrRxxqQWdPt1opv+dyMRlkY8fNmUPoiTqi9PKMHUVo6wM9H/+M6EVDEnX2bhYTIfh4fQncSJZmnNwJKxsTHGjx+P8ePHl7nP9OnTMX36dJltxQNHH+Pi4oKFCxcqE6LaFYgk2HEvFSu7uJV4jMvlwtraWm0DYOpuX5PH1mTfCHAzPg9PUvgYXcoHFF3OvS73jRQx0eeVul2Xc6/LfSNFMvgirAtLxJqu7jDgvcuDLudel/tGgAH1bTCgvk2pj+ly7nW5b6QI++9OcGOa2sPV4t1sJV3OvS73rarTmoElAhx6XHSHoaGNSv7xc3V1RXp6utqOre72NXlsTfaNALfjc9GtjiXcLQ1LPKbLudflvpHy6XLudblvpMi6sETYm+phZBPZkwG6nHtd7hspny7nXpf7RoqciczCvYR8HBhYW2a7Ludel/tW1dFQnBapbWOI1V3dSz0LLhKJ8PbtW4hEIrUcW93ta/LYmuwbAaa2csKRMqbn6nLudblvpHy6nHtd7hsputHC/24mYkGHGtD74NJlXc69LveNlE+Xc6/LfSNFs5WWXYnHaF+7EidvdTn3uty3qo4GlrSIv5s5xnxw16xib9++RY0aNfD27Vu1HFvd7Wvy2JrsGyny4QeUYrqce13uGymfLudel/tGgAOP0uFmYYChDW1LPKbLudflvpHy6XLudblvBDgflY07b/Mxp51Licd0Ofe63Leqji6FI4QQQgghmNzCAYMa2NCNFgghRMsZ6nGwtFMNeFqVXGqCEE2ggSVCCCGEEAIOhwN7U7pzFiGEaLv2HhZo72Gh6TAIkaJL4QghhBBCCCGEEEKIQmhgqYqwsrLCL7/8AisrqyrZviaPrcm+kfLpcu51uW+kfLqce13uGymfLudel/tGyqfLudflvpHy6XLudblvVR2HMcY0HURVdffuXTRr1gzh4eHw8/OTbo+KioKXl5cGIysfxaccbY9PWVTX6kHxaRbVtXpQfJpFda0eFJ9mUV2rB8WnWVTX6kHxaQ+asVRFpKenY8qUKUhPT6+S7Wvy2JrsGymfLudel/tGyqfLudflvpHy6XLudblvpHy6nHtd7hspny7nXpf7VtXRwFIVkZubi02bNiE3N7dKtq/JY2uyb6R8upx7Xe4bKZ8u516X+0bKp8u51+W+kfLpcu51uW+kfLqce13uW1VX7e4Kx+fzERwcjNDQUOTk5MDV1RX9+/dH+/btNR0aIYQQQgghhBBCSJVS7QaWVq5ciZcvXyIwMBAuLi4ICQnB6tWrwRhDhw4dNB0eIYQQQgghhBBCSJVRrQaW7ty5g4iICMycOVM6iNS4cWMkJydjx44dCAgIAI/H03CUZdPX16/S7Wvy2JrsGymfLudel/tGyqfLudflvpHy6XLudblvpHy6nHtd7hspny7nXpf7VpVVq4GlGzduwNjYGAEBATLbO3fujDVr1uDFixfw8fHRUHTlc3d3R2FhYZVtX5PH1mTfSPl0Ofe63DdSPl3OvS73jZRPl3Ovy30j5dPl3Oty30j5dDn3uty3qq5aLd4dHR0NV1fXErOSatasKX1cW0kkEmRnZ0MikVTJ9jV5bE32jZRPl3Ovy30j5dPl3Oty30j5dDn3utw3Uj5dzr0u942UT5dzr8t9q+qq1cBSTk4OzMzMSmwv3paTk1PZIVVYXFwcLC0tERcXVyXb1+SxNdk3Uj5dzr0u942UT5dzr8t9I+XT5dzrct9I+XQ597rcN1I+Xc69LvetqqtWl8IBAIfDUeixYgkJCUhISAAAPH36VGVxEUIIIYQQQgghhFQ11WpgydzcvNRZSbm5uQBQ6mymD23duhVLlixReWyEEEIIIYQQQgghVU21Gljy8PDA1atXIRaLZdZZevPmjfTxj5k4cSJ69eoFAIiIiMDYsWNLzFyKi4tDZmamyuIGIJ0l9fDhQ6SmpirVVmnxqbJ9eX14bFX//FTdN3XkV1O8vb1hYmIis43P5wMoOSOP6lo+VNeaQ3WtPlTXmkN1rT5U15pDda0+VNeaQ3WtPlTXmlNaXctg1cidO3dYz549WUhIiMz2RYsWscDAQCYSieRqLzg4mAGgL/qqsl/h4eFU1/Slc19U1/Sli19U1/Sli19U1/Sli19U1/Sli1+l1fX7OIwxhmpk4cKFePXqFQIDA+Hs7IyQkBCcPXsWM2bMQMeOHeVqKzU1FWfOnIGnpyeMjY0BFI1CjxgxAsHBwfDx8VFDD5RD8SlH2+OTV2kjz1TXqkfxVS6q68pB8VUuquvKQfFVLqrrykHxVS6q68pB8VWuj81YqlaXwgHA3LlzERwcjD///BM5OTlwdXXFrFmz0L59e7nbsrOzw/Dhw0t9zMfHB35+fsqGqzYUn3K0PT5lUF2rD8WnOVTX6kPxaQ7VtfpQfJpDda0+FJ/mUF2rD8WnHardwJKxsTHGjx+P8ePHazoUQgghhBBCCCGEkCqNq+kAdI2zszMWLVoEZ2dnTYdSKopPOdoen7poe78pPuVoe3zqou39pviUo+3xqYu295viU462x6cu2t5vik852h6fumh7vyk+5Wh7fKpW7dZYIoQQQgghhBBCCCGqQTOWCCGEEEIIIYQQQohCaGCJEEIIIYQQQgghhCiEBpYIIYQQQgghhBBCiEKq3V3htMEvv/yCBw8eoLCwEA4ODhg1ahRatGih6bBKePbsGWbPno1hw4Zh8ODBmg5Hat68eXj+/Dl4PB4AoE6dOlixYoWGo5J17tw5HDp0CJmZmbCzs8OCBQt0fuE2qmvlUF1rJ6pr5VBdayeqa+VQXWsnqmvlUF1rJ6pr5VBdVx4aWNKAwYMHY/r06dDX18eLFy+wcOFCbNu2DRYWFpoOTUoikeC3335DnTp1NB1KqaZMmYJOnTppOoxS3bp1C8ePH8f8+fPh5uaGhIQEmJubazostaO6Vh7VtfahulYe1bX2obpWHtW19qG6Vh7VtfahulYe1XXloIElDXB3d5f+n8fjQSQSIS0tTateIE6fPg0fHx/k5ORoOpQq58CBAxg7dqw0zy4uLhqOqHJQXes2qmuqa11EdU11rYuorqmudRHVNdW1LtKluqaBpY/Iz8/HgQMH8Pr1a0RFRSE7OxtDhgzBsGHDSuzL5/MRHByM0NBQ5OTkwNXVFf3790f79u1L7LtmzRpcv34dQqEQzZs3h6enp9bEl52djePHj2P16tXYvn27QnGpMz4A+P333/H777/D09MTY8aMgZeXl1bEJxaLERUVhTdv3mD9+vXg8Xjo3LkzhgwZAg6Ho1CM6kB1TXUtT3xU11TXisYHUF0ri+qa6lqe+Kiuqa4VjQ+gulYW1TXVtTzxVZW6rigaWPqInJwcnD17Fp6enmjdujXOnj1b5r4rV67Ey5cvERgYCBcXF4SEhGD16tVgjKFDhw4y+86cORPTp0/HgwcPEBsbq3DxqCO+3bt3o3fv3jA1NVUoJnXHN3r0aLi5uYHL5eLUqVNYvHgxgoKCFIpX1fFlZmZCLBYjIiICGzZsQF5eHhYtWgQHBwd07txZ7vjUhepaOVTXVNfaEh/VteLxUV1TXSsaH9W18qiulUN1TXWtLfFRXSseX1Wp6wpjpFwSiYRJJBLGGGNZWVmsZ8+ebO/evSX2u337NuvZsye7fPmyzPYFCxawwMBAJhKJyjzGkiVL2O3bt7UivpcvX7Jvv/2WicVixhhj69atY/v371coNnXEV5qvv/6a3blzRyviy8nJYT179mQPHz6U7nP48GG2atUqheJTF6prqmt54qO6fofqmuq6slFdU13LEx/V9TtU11TXlY3qmupanviqSl1XFFfTA1vajsPhVGhU+MaNGzA2NkZAQIDM9s6dOyM9PR0vXrwo87kSiQQJCQlaEd/jx48RExODkSNHYvjw4bh69Sr++usvrFu3TiviK+sYjDGtiM/MzAw2NjYKxVKZqK6pruWJj+r6HaprquvKRnVNdS1PfFTX71BdU11XNqprqmt54qsqdV1RNLCkItHR0XB1dZXeyrBYzZo1pY8DQEZGBq5fv46CggKIxWJcu3YNDx8+RMOGDbUivs8++wxbtmzB//73P/zvf/9Dy5Yt0b17d4wdO1Yr4svNzcW9e/cgFAohFApx/PhxZGZmwtvbWyviA4peNA4fPoz8/Hykpqbi7NmzaN68uVrjUxeq68qJj+q6clFdV058VNeVi+q6cuKjuq5cVNeVEx/VdeWiuq6c+KiuKxetsaQiOTk5cHR0LLHdzMxM+nixY8eO4ddffwWHw4GzszO+//57aaFpOj5jY2MYGxtLHzc0NISJiYna7zxQ0fjEYjH27NmDuLg46OnpoWbNmli4cKF0P03HBwBDhgzB1q1bMWbMGBgbG6Nr165ae4vLj6G6rpz4qK4rF9V15cRHdV25qK4rJz6q68pFdV058VFdVy6q68qJj+q6ctHAkgqVNzWu+DFra2v89NNPlRVSqTHI89j06dPVFE3FY3j/MUtLS6xdu7ayQio1ho89pq+vjylTpmDKlCmVEZbaUV0rh+paO1FdK4fqWjtRXSuH6lo7UV0rh+paO1FdK4fqWvvQpXAqYm5uLjP6WCw3NxcA1D4y+jEUn3K0PT510fZ+U3zK0fb41EXb+03xKUfb41MXbe83xaccbY9PXbS93xSfcrQ9PnXR9n5TfMrR9vjUhQaWVMTDwwNxcXEQi8Uy29+8eSN9XJMoPuVoe3zqou39pviUo+3xqYu295viU462x6cu2t5vik852h6fumh7vyk+5Wh7fOqi7f2m+JSj7fGpCw0sqYi/vz/4fD6uX78us/3ixYuwsbFB3bp1NRRZEYpPOdoen7poe78pPuVoe3zqou39pviUo+3xqYu295viU462x6cu2t5vik852h6fumh7vyk+5Wh7fOpCayxVQHh4OAoKCsDn8wEAsbGxCA0NBQA0b94choaGaNasGXx9fREUFIT8/Hw4OzsjJCQEd+/exYwZM0qsCk/xUXyapu39pvh0Oz510fZ+U3y6HZ+6aHu/KT7djk9dtL3fFJ9ux6cu2t5vik+349MkDmOMaToIbTdu3DgkJyeX+tj27dulq77z+XwEBwfj2rVryMnJgaurKwYMGID27dtTfBSf1tH2flN8uh2fumh7vyk+3Y5PXbS93xSfbsenLtreb4pPt+NTF23vN8Wn2/FpEg0sAu2eMgAACxpJREFUEUIIIYQQQgghhBCF0BpLhBBCCCGEEEIIIUQhNLBECCGEEEIIIYQQQhRCA0uEEEIIIYQQQgghRCE0sEQIIYQQQgghhBBCFEIDS4QQQgghhBBCCCFEITSwRAghhBBCCCGEEEIUQgNLhBBCCCGEEEIIIUQhNLBECCGEEEIIIYQQQhRCA0uEEEIIIYQQQgghRCE0sEQIIYQQQgghhBBCFEIDS4QQQghRCofDqdDX5cuXMXr0aHh6emo65BKuXr0KQ0NDREdHS7d17NgRHTt21FxQVYgieb1w4QLMzMwQHx+vnqAIIYQQUik4jDGm6SAIIYQQUnXduHFD5vtly5bh0qVLuHjxosz2+vXrIyUlBdnZ2WjatGllhlguxhiaN28Of39/bNy4Ubr9yZMnAIriJuUbPXo0Ll++jDdv3sj1vE8++QRubm7YtWuXegIjhBBCiNrpaToAQgghhFRtrVu3lvne3t4eXC63xHYAsLCwqKywKuz06dO4e/cu/vzzT5ntNKCkfl9//TUGDx6M5cuXw83NTdPhEEIIIUQBdCkcIYQQQipNaZdMcTgcTJkyBTt27EC9evVgbGyM5s2b48aNG2CM4ZdffkHNmjVhZmaGTz75BK9evSrR7vnz59G5c2dYWFjAxMQEbdu2xYULFyoUU1BQEFq0aIF69erJbP/wUrg3b96Aw+Fg9erVWLt2rTQmf3//ErO2SpOfn49Zs2ahZs2aMDIygo2NDZo3b459+/bJ7Hfnzh306tULNjY2MDIyQtOmTXHw4MES7cXHx2PChAlwc3ODgYEBXFxcMGDAACQlJUn3iYmJwYgRI+Dg4ABDQ0P4+PhgzZo1kEgkCvdr586dqFevnrS93bt3l/lzbdKkCczMzGBubg5vb2/MmzdPZp+ePXvCzMwM27dv/+jPjxBCCCHaiWYsEUIIIUTj/v33X9y7dw8//fQTOBwOZs+ejS+++AKBgYGIiorCxo0bkZWVhRkzZqB///6IiIgAh8MBAAQHB2PUqFHo3bs3du3aBX19fWzduhVdu3bFmTNn0Llz5zKPW1hYiPPnz2Pq1KkVjnXTpk3w9vbG+vXrAQALFixA9+7d8fr1a1haWpb5vBkzZmDPnj1Yvnw5mjZtiry8PDx69AhpaWnSfS5duoTPP/8crVq1wpYtW2BpaYn9+/dj8ODByM/Px+jRowEUDSq1aNECQqEQ8+bNQ+PGjZGWloYzZ84gIyMDjo6OSElJQZs2bVBYWIhly5bB09MT//77L2bNmoXIyEhs3rxZ7n7t3LkTX375JXr37o01a9YgKysLixcvhkAgAJf77nzl/v37MXnyZEydOhWrV68Gl8vFq1evpJcXFjMwMECbNm1w4sQJLF26tMI5IIQQQogWYYQQQgghKhQYGMhMTU3LfMzDw0NmGwDm5OTEcnNzpduOHj3KADBfX18mkUik29evX88AsAcPHjDGGMvLy2M2NjasZ8+eMm2KxWLWpEkT1rJly3JjvXnzJgPA9u/fX+KxDh06sA4dOki/f/36NQPAGjVqxEQikXT7rVu3GAC2b9++co/VsGFD1qdPn3L38fb2Zk2bNmVCoVBme48ePZizszMTi8WMMcbGjBnD9PX12ZMnT8psa86cOQwAu3nzpsz2r776inE4HPb8+XO5+iUWi5mLiwvz8/OTycmbN2+Yvr6+TF6nTJnCrKysyu1rsR9++IFxuVyZ/BNCCCGk6qBL4QghhBCicZ06dYKpqan0ex8fHwBAt27dpDOT3t9efPe269evIz09HYGBgRCJRNIviUSCzz//HLdv30ZeXl6Zx3379i0AwMHBocKxfvHFF+DxeNLvGzduLBNTWVq2bIlTp05hzpw5uHz5Mvh8vszjr169wrNnzzB8+HAAkOlP9+7dkZCQgOfPnwMATp06hU6dOkl/HqW5ePEi6tevj5YtW8psHz16NBhjJRZX/1i/nj9/jrdv32LYsGEyOfHw8ECbNm1K9DUzMxNDhw7FP//8g9TU1DLjdHBwgEQiQWJiYpn7EEIIIUR70cASIYQQQjTOxsZG5nsDA4NytxcUFACAdD2hAQMGQF9fX+br559/BmMM6enpZR63eHDHyMiowrHa2trKfG9oaCjTVll+/fVXzJ49G0ePHkWnTp1gY2ODPn364OXLlzJ9mTVrVom+TJ48GQCkAzQpKSlwdXUt93hpaWlwdnYusd3FxUX6uDz9Kt7fycmpRJsfbhs5ciT++OMPREdHo3///nBwcECrVq1w7ty5Es8t/tl/7OdHCCGEEO1EaywRQgghpMqys7MDAGzYsKHUu9ABgKOj40efX97gk6qYmppiyZIlWLJkCZKSkqSzl3r27Ilnz55JY5k7dy769etXahvFC4zb29sjLi6u3OPZ2toiISGhxPbiWVrFx6uo4oGn0mYWlbbtyy+/xJdffom8vDyEhIRg0aJF6NGjB168eAEPDw/pfsU/e3njIYQQQoh2oIElQgghhFRZbdu2hZWVFZ48eYIpU6bI/fziS8kiIyNVHVq5HB0dMXr0aNy/fx/r169Hfn4+6tWrhzp16uD+/fv48ccfy31+t27dsGfPHjx//rzE3eyKde7cGStXrsTdu3fh5+cn3b57925wOBx06tRJrpjr1asHZ2dn7Nu3DzNmzJBeDhcdHY3r169LZ0J9yNTUFN26dUNhYSH69OmDx48fywwsRUVFwdbWttwBQEIIIYRoLxpYIoQQQkiVZWZmhg0bNiAwMBDp6ekYMGAAHBwckJKSgvv37yMlJQVBQUFlPt/V1RVeXl64ceMGvvnmG7XG2qpVK/To0QONGzeGtbU1nj59ij179sDf3x8mJiYAgK1bt6Jbt27o2rUrRo8ejRo1aiA9PR1Pnz7F3bt3cejQIQDA0qVLcerUKbRv3x7z5s1Do0aNkJmZidOnT2PGjBnw9vbGt99+i927d+OLL77A0qVL4eHhgRMnTmDz5s346quvULduXbni53K5WLZsGcaNG4e+ffti/PjxyMzMxOLFi0tcCjd+/HgYGxujbdu2cHZ2RmJiIlauXAlLS0u0aNFCZt8bN26gQ4cOMus2EUIIIaTqoIElQgghhFRpI0aMgLu7O1atWoWJEyciJycHDg4O8PX1xejRoz/6/OHDh2Pjxo0QCATSdYXU4ZNPPsGxY8ewbt065Ofno0aNGhg1ahR++OEH6T6dOnXCrVu3sGLFCkyfPh0ZGRmwtbVF/fr1MWjQIOl+NWrUwK1bt7Bo0SL89NNPSEtLg729PQICAqTrUtnb2+P69euYO3cu5s6di+zsbHh5eWHVqlWYMWOGQn0YO3YsAODnn39Gv3794OnpiXnz5uHKlSu4fPmydL927dph586dOHjwIDIyMmBnZ4eAgADs3r0b9vb20v0iIyPx8OFDLF68WKF4CCGEEKJ5HMYY03QQhBBCCCGa8vbtW9SsWRO7d+/G4MGDNR1OtbJgwQLs3r0bkZGR0NOj852EEEJIVUQDS4QQQgip9mbPno1Tp04hIiICXC7dNLcyZGZmwsvLCxs2bMDw4cM1HQ4hhBBCFESnhgghhBBS7c2fPx8mJiaIj4+Hm5ubpsOpFl6/fo25c+di2LBhmg6FEEIIIUqgGUuEEEIIIYQQQgghRCE015sQQgghhBBCCCGEKIQGlgghhBBCCCGEEEKIQmhgiRBCCCGEEEIIIYQohAaWCCGEEEIIIYQQQohCaGCJEEIIIYQQQgghhCiEBpYIIYQQQgghhBBCiEJoYIkQQgghhBBCCCGEKIQGlgghhBBCCCGEEEKIQv4PnOGeKQDzrDcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1400x300 with 12 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<ggplot: (311683578)>"
      ]
     },
     "execution_count": 242,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dummy_df = duration_cdf_df.loc[:1, :].copy()\n",
    "dummy_df.loc[0, 'dummy'] = '0'\n",
    "dummy_df.loc[1, 'dummy'] = '1'\n",
    "\n",
    "marker_df = pd.DataFrame({\n",
    "    'sec': [60*60, 60*60*4, 60*60*24, 60*60*24*4],\n",
    "    'y': [1, 1, 1, 1],\n",
    "    'text': ['1 hour', '4 hours', '1 day', '4 days']\n",
    "})\n",
    "\n",
    "plt = ggplot(median_cdf_df) +\\\n",
    "    geom_line(aes(x='duration', y='cdf'), data=duration_cdf_df, color='black') +\\\n",
    "    geom_line(aes(x='ia', y='cdf'), data=ia_cdf_df, color='#0280c9') +\\\n",
    "    geom_line(aes(x='duration', y='cdf', color='dummy'), data=dummy_df) +\\\n",
    "    scale_x_log10(breaks=[1000, 10000, 100000, 1000000], labels=lambda l: list(map(lambda x: '$10^{}$'.format(int(np.log10(x))), l))) +\\\n",
    "    geom_vline(aes(xintercept='sec'), data=marker_df, linetype='--') +\\\n",
    "    facet_wrap(facets='vendor_cat', nrow=2) +\\\n",
    "    ylab('ECDF') +\\\n",
    "    xlab('Time (in seconds)') +\\\n",
    "    scale_color_manual(['black', '#0280c9'], labels=['Duration', 'Interarrival time']) +\\\n",
    "    guides(color=guide_legend(title='')) +\\\n",
    "    theme_classic(base_size=12, base_family='sans-serif') +\\\n",
    "    theme(figure_size=(14, 3),\n",
    "          axis_text_y=element_text(margin={'r': 5}),\n",
    "          panel_grid_major_x=element_line(size=0.7, color=\"gainsboro\"),\n",
    "          text=element_text(size=12),\n",
    "          legend_box_spacing=0.01,\n",
    "          legend_box_margin=0,\n",
    "          legend_margin=0,\n",
    "          legend_key=element_blank(),\n",
    "          legend_entry_spacing=5,\n",
    "          legend_background=element_rect(fill=(0,0,0,0), color=(0,0,0,0)),\n",
    "          legend_position='top')\n",
    "\n",
    "plt.save('plots/failure_duration_ia_cdf.pdf', limitsize=None)\n",
    "plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/sacheendra/miniconda3/envs/thesis/lib/python3.8/site-packages/plotnine/ggplot.py:727: PlotnineWarning: Saving 14 x 3 in image.\n",
      "/Users/sacheendra/miniconda3/envs/thesis/lib/python3.8/site-packages/plotnine/ggplot.py:730: PlotnineWarning: Filename: plots/failure_mean_peak_cdf.pdf\n",
      "/Users/sacheendra/miniconda3/envs/thesis/lib/python3.8/site-packages/plotnine/geoms/geom_path.py:75: PlotnineWarning: geom_path: Removed 1 rows containing missing values.\n",
      "/Users/sacheendra/miniconda3/envs/thesis/lib/python3.8/site-packages/plotnine/geoms/geom_path.py:81: PlotnineWarning: geom_path: Each group consist of only one observation. Do you need to adjust the group aesthetic?\n",
      "/Users/sacheendra/miniconda3/envs/thesis/lib/python3.8/site-packages/plotnine/geoms/geom_path.py:75: PlotnineWarning: geom_path: Removed 1 rows containing missing values.\n",
      "/Users/sacheendra/miniconda3/envs/thesis/lib/python3.8/site-packages/plotnine/geoms/geom_path.py:81: PlotnineWarning: geom_path: Each group consist of only one observation. Do you need to adjust the group aesthetic?\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABJYAAAF1CAYAAAC6f2D4AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAA9hAAAPYQGoP6dpAADfY0lEQVR4nOzdd1QUVxsG8GdhKUuR3pFqARVEsKFYsUSDLWLvLSZGjT12YxKjiSVqEnvvLYq9l1gx9ti7YAXpve3e7w8+Nq6AAqJLeX7n7NGdvTPzznJ3dvadWyRCCAEiIiIiIiIiIqJ80lB3AEREREREREREVDwxsURERERERERERAXCxBIRERERERERERUIE0tERERERERERFQgTCwREREREREREVGBMLFEREREREREREQFwsQSEREREREREREVCBNLRERERERERERUIEwsERERERERERFRgTCxREREREREREREBcLEEhERERERERERFQgTS0REREREREREVCBMLBERERERERERUYEwsURERERERERERAXCxBIRERERERERERUIE0tERERERERERFQgTCwREREREREREVGBMLFEREREREREREQFwsQSEREREREREREVCBNLRERERERERERUIEwsERERERERERFRgTCxREREREREREREBcLEEhERERERERERFQgTS0REREREREREVCBMLBERERERERERUYEwsURERERERERERAXCxBIRERERERERERUIE0tERERERERERFQgTCwREREREREREVGBMLFEREREREREREQFwsQSEREREREREREVCBNLRERERERERERUIEwsERERERERERFRgTCxREREREREREREBcLEEhERERERERERFQgTS0REREREREREVCBMLBERERERERERUYEwsURERERERERERAXCxBIRERERERERERUIE0tERP+3atUqSCQSSCQSnDhxItvrQgiUK1cOEokEDRs2/OTxFXe9e/eGgYGBusPIs82bN6Ny5cqQyWSQSCS4evWqukMqND///DOCgoLUsu8rV66gQYMGMDIygkQiwdy5c3Mt++TJE3z++ecwNTWFRCLBsGHD8rUvJycn9O7dW2V7EokEq1atKlDsJcmb5zuJRAKpVAp7e3v06dMHz58//2j7/f777yGRSBAREfHR9vEpnThxAhKJBNu2bVN3KHnyoZ+pomzfvn34/vvv1bLvtLQ0fPXVV7CxsYGmpia8vLzeWX7ixIlwcHCAVCqFsbFxvvaV9Rl6U8OGDXldQkRqJVV3AERERY2hoSGWL1+e7SLt77//xsOHD2FoaKiewOiTef36NXr06IHPPvsMCxYsgI6ODipUqKDusArNzz//jMDAQLRt2/aT77tv375ITEzEpk2bYGJiAicnp1zLDh8+HOfPn8eKFStgbW0NGxubfO1rx44dKFOmzAdGXLKtXLkSbm5uSE5OxsmTJzF9+nT8/fffuH79OvT19dUdHhWyD/1MFWX79u3Dn3/+qZbk0sKFC7F48WL8/vvv8PHxeedNlJ07d2LatGmYMGECWrRoAR0dnXztq3///vjss88+NGQiokLFxBIR0Vs6deqE9evX488//1T5Ubp8+XL4+voiLi5OjdHR+yQlJUFPT++DtnHv3j2kp6eje/fuaNCgQb7XT05Ohq6ubra7yuqWnJwMmUym1hhu3LiBAQMGoEWLFnkqW7NmzQInwKpVq1ag9fJKCIGUlBS1v6cfokqVKqhevToAoFGjRpDL5fjxxx8RFBSEbt26qTk6ylJY55QP+UwV1fpeGOf8D3Xjxg3IZDIMHjw4T2UBYOjQobC0tMz3vuzt7WFvb5/v9fKjKLynRFS8sCscEdFbunTpAgDYuHGjcllsbCz++usv9O3bN8d10tLS8NNPP8HNzQ06OjqwsLBAnz598Pr1a5VymzdvRrNmzWBjYwOZTAZ3d3eMHTsWiYmJKuWyuo09ePAALVu2hIGBAcqWLYuRI0ciNTX1vcfg5OSEgIAAHDhwAN7e3pDJZHBzc8OKFStUyuXUpB74r5vMkydPsm1zz549qFatmjL+PXv2KNdxd3eHvr4+atasiYsXL+YY282bN+Hv7w99fX1YWFhg8ODBSEpKUikjhMCCBQvg5eUFmUwGExMTBAYG4tGjRyrlGjZsiCpVquDkyZOoU6cO9PT0cv0bZdm1axd8fX2hp6cHQ0NDNG3aFOfOnVO+3rt3b/j5+QHITDK+r+tj1nt16NAh9O3bFxYWFtDT01P+nTZv3gxfX1/o6+vDwMAAzZs3x5UrV1S2kfX3zst7k5KSgnHjxsHZ2Rna2tqws7PDN998g5iYGJVyWX+v7du3o1q1atDV1cXUqVMhkUiQmJiI1atXK7tBZR1fUlISRo0aBWdnZ+jq6sLU1BTVq1dX+Szk5saNG2jTpg1MTEygq6sLLy8vrF69Otv7lJGRgYULFyr3nZOs7kUPHjzA/v37lWWfPHmClJQUjBw5El5eXjAyMoKpqSl8fX2xc+fObNt5uytcTnr37p1jq6mcPhsSiQSDBw/GokWL4O7uDh0dHeUx3r9/H127doWlpSV0dHTg7u6OP//88z3vWtFTu3ZtAEBISAiAvH8WDx8+jDZt2sDe3h66urooV64cBg4cmKcub3fu3IGLiwtq1aqF8PDwXMtl/U1u3ryJLl26wMjICFZWVujbty9iY2OV5d7V5VEikai0aMna5r///osOHToo69SIESOQkZGBu3fv4rPPPoOhoSGcnJzw66+/5hhbSkoKRowYAWtra8hkMjRo0CDb5xwALl68iNatW8PU1BS6urqoVq0atmzZolLmfeeUnISGhqJ79+4q9W/27NlQKBQA3v2Zys2H1vesfa5bty5P7837zs3Af3+vy5cvIzAwECYmJnB1dUXv3r2V+3+zi2fW8W3duhW1atWCkZER9PT04OLi8t7vCiBv51uJRIJly5YhOTlZud/cuts6OTlh4sSJAAArKyuV+pjX64PcvrfflPXev92tP6fPRtb3z/Xr19GsWTMYGhrC398fQN6vbYiImFgiInpLmTJlEBgYqJKE2bhxIzQ0NNCpU6ds5RUKBdq0aYMZM2aga9eu2Lt3L2bMmIHDhw+jYcOGSE5OVpa9f/8+WrZsieXLl+PAgQMYNmwYtmzZglatWmXbbnp6Olq3bg1/f3/s3LkTffv2xW+//YZffvklT8dx7do1jBw5EsOHD8fOnTvh6emJfv364eTJkwV4V/7b5rhx4/Ddd99h+/btMDIywhdffIEpU6Zg2bJl+Pnnn7F+/XrExsYiICBA5dizjqlly5bw9/dHUFAQBg8ejMWLF2d7XwcOHIhhw4ahSZMmCAoKwoIFC3Dz5k3UqVMHYWFhKmVfvnyJ7t27o2vXrti3bx8GDRqUa/wbNmxAmzZtUKZMGWzcuBHLly9HdHQ0GjZsiNOnTwMAJk2apPyB8vPPP+PcuXNYsGDBe9+bvn37QktLC2vXrsW2bdugpaWFn3/+GV26dEGlSpWwZcsWrF27FvHx8ahXrx5u3bqV7/dGCIG2bdti1qxZ6NGjB/bu3YsRI0Zg9erVaNy4cbYfnpcvX8bo0aMxdOhQHDhwAO3bt8e5c+cgk8nQsmVLnDt3TuX4RowYgYULFyrLr127Fh06dEBkZOQ7j/3u3buoU6cObt68ifnz52P79u2oVKkSevfurfwh/vnnnyt/JAYGBir3nRNvb2+cO3cO1tbWqFu3rrKsjY0NUlNTERUVhVGjRiEoKAgbN26En58fvvjiC6xZs+a9f6cPFRQUhIULF2Ly5Mk4ePCg8m9Zo0YN3LhxA7Nnz8aePXvw+eefY+jQoZg6depHj6kwPXjwAABgYWEBIO+fxYcPH8LX1xcLFy7EoUOHMHnyZJw/fx5+fn5IT0/PdX9///036tSpA09PTxw/fjxPLTjat2+PChUq4K+//sLYsWOxYcMGDB8+/IOOu2PHjqhatSr++usvDBgwAL/99huGDx+Otm3b4vPPP8eOHTvQuHFj5bnvbePHj8ejR4+wbNkyLFu2DC9evEDDhg1VEnDHjx9H3bp1ERMTg0WLFmHnzp3w8vJCp06dckxE5HROycnr169Rp04dHDp0CD/++CN27dqFJk2aYNSoUcoWNO/6TL1LYdT3vLw3eTk3v+mLL75AuXLlsHXrVixatAiTJk1CYGAgACiPLev4zp07h06dOsHFxQWbNm3C3r17MXnyZGRkZLzz2PN6vj137hxatmwJmUym3O/nn3+e4zZ37NiBfv36AQAOHDiAc+fOoX///gDyd31Q2NLS0tC6dWs0btwYO3fuxNSpU/N1bUNEBEFEREIIIVauXCkAiAsXLojjx48LAOLGjRtCCCFq1KghevfuLYQQonLlyqJBgwbK9TZu3CgAiL/++ktlexcuXBAAxIIFC3Lcn0KhEOnp6eLvv/8WAMS1a9eUr/Xq1UsAEFu2bFFZp2XLlqJixYrvPRZHR0ehq6srQkJClMuSk5OFqampGDhwoHLZlClTRE5fBVnvxePHj1W2KZPJxLNnz5TLrl69KgAIGxsbkZiYqFweFBQkAIhdu3ZlO6Z58+ap7GvatGkCgDh9+rQQQohz584JAGL27Nkq5Z4+fSpkMpkYM2aMclmDBg0EAHH06NH3vidyuVzY2toKDw8PIZfLlcvj4+OFpaWlqFOnjnJZ1t9/69at791u1nvVs2dPleWhoaFCKpWKIUOGqCyPj48X1tbWomPHjspleX1vDhw4IACIX3/9VaXc5s2bBQCxZMkS5TJHR0ehqakp7t69my1mfX190atXr2zLq1SpItq2bfveY35b586dhY6OjggNDVVZ3qJFC6GnpydiYmKUywCIb775Jk/bdXR0FJ9//vk7y2RkZIj09HTRr18/Ua1atWzrv3mcjx8/FgDEypUrlct69eolHB0ds203p88GAGFkZCSioqJUljdv3lzY29uL2NhYleWDBw8Wurq62coXBVn1Njg4WKSnp4v4+HixZ88eYWFhIQwNDcWrV6/y9Vl8U9a5LSQkRAAQO3fuVL6W9b6+fv1arF27Vmhra4uhQ4eqfCZzk7Xu2/V/0KBBQldXVygUCiFEzn/nLADElClTsm3z7WP08vISAMT27duVy9LT04WFhYX44osvlMuyzhXe3t7K/QshxJMnT4SWlpbo37+/cpmbm5uoVq2aSE9PV9lXQECAsLGxUb4HuZ1TcjN27FgBQJw/f15l+ddffy0kEonKOSAvn6ksH1rf8/re5OfcnPX3mjx5crZ4v/nmmxy/z2bNmiUAqJyH8iI/59tevXoJfX39PG33zc9Abt51fZDTualBgwYq1yVZ7/3x48dVyuV2DgQgVqxYoVK2oNc2RFQ6scUSERVJbzZl/xiP92nQoAFcXV2xYsUKXL9+HRcuXMi12fyePXtgbGyMVq1aISMjQ/nw8vKCtbW1SlP0R48eoWvXrrC2toampia0tLSUY/jcvn0723vw9p1KT09PZReV9/Hy8oKDg4Pyua6uLipUqJDn9XPbpp2dnfK5u7s7gMwuaW+Ox5C1PKd9vT1uS9euXQFk3s0HMt9PiUSC7t27q7yf1tbWqFq1aram/SYmJmjcuPF7Y7979y5evHiBHj16QEPjv68/AwMDtG/fHsHBwdm6neVH+/btVZ4fPHgQGRkZ6Nmzp8px6OrqokGDBjnOPPi+9+bYsWMAkK17V4cOHaCvr4+jR4+qLPf09MzXoOM1a9bE/v37MXbsWJw4cSLPd6SPHTsGf39/lC1bVmV57969kZSUlGvLpILaunUr6tatCwMDA0ilUmhpaWH58uXZPkMfQ+PGjWFiYqJ8npKSgqNHj6Jdu3bQ09NT+Vu3bNkSKSkpCA4Ofuc2Naf+81Ef71K7dm1oaWnB0NAQAQEBsLa2xv79+2FlZZWvz2J4eDi++uorlC1bVvk3cXR0BJD93AYA06ZNQ+/evTFjxgzMmzdP5TP5Pq1bt1Z57unpiZSUlHd2o3ufgIAAlefu7u6QSCQqY4FJpVKUK1cux/Na165dVb5bHB0dUadOHeVn98GDB7hz547yM/52PXn58iXu3r2rss23zym5OXbsGCpVqoSaNWuqLO/duzeEEMrzRkEURn1/33tTkHNzXt8bAKhRowaAzFZpW7ZsyfOsh/k9336o/FwffAxvv6f5ubYhIuLg3UREOZBIJOjTpw/mz5+PlJQUVKhQAfXq1cuxbFhYGGJiYqCtrZ3j61ljjCQkJKBevXrQ1dXFTz/9hAoVKkBPTw9Pnz7FF198ke1HvJ6eHnR1dVWW6ejoICUlJU/HYGZmlm2Zjo7OBzVfNzU1VXmedcy5LX87VqlUmi0ua2trAFB2twoLC4MQAlZWVjnG4OLiovI8r7MaZW0/p/K2trZQKBSIjo4u8IClb283q5tQ1o+at739Qzov701kZCSkUqmym1IWiUQCa2vrbF3W8jvj0/z582Fvb4/Nmzfjl19+ga6uLpo3b46ZM2eifPnyua4XGRmZ6/v6ZvyFYfv27ejYsSM6dOiA0aNHw9raGlKpFAsXLsw2htjH8PZxRkZGIiMjA7///jt+//33HNfJyzhD6rJmzRq4u7tDKpXCyspK5fjy+llUKBRo1qwZXrx4gUmTJsHDwwP6+vpQKBSoXbt2juecdevWwc7ODp07d853zG9/TrJm1Srsc1tO52Btbe0cJ3DI+qy+vezatWsA/jsfjBo1CqNGjcoxhrfrSX7ObTmNE1YYn7/CqO/ve28Kcm7Oz7mtfv36CAoKwvz589GzZ0+kpqaicuXKmDBhgnJMxZzk93z7IfJ7fVDY9PT0ss2gmddrGyIigIklIqJc9e7dG5MnT8aiRYswbdq0XMuZm5vDzMwMBw4cyPF1Q0NDAJl3P1+8eIETJ06ozDT29qDLn1LWj6bU1FSVKY8/1gVjRkYGIiMjVX4Yvnr1CsB/PxbNzc0hkUhw6tSpHKdhfntZXmdJytr+y5cvs7324sULaGhoqNyZz6+34zA3NwcAbNu2Tdly413y8t6YmZkhIyMDr1+/VvmxI4TAq1evsiWx8juDlL6+PqZOnYqpU6ciLCxM2XqpVatWuHPnTq7rmZmZ5fq+Av+9F4Vh3bp1cHZ2xubNm1WOLy+D2udEV1c3x3Vz+wy8/Z6amJhAU1MTPXr0wDfffJPjOs7OzgWK7VNwd3dXzgr3trx+Fm/cuIFr165h1apV6NWrl/L1rPGacnLgwAF06tQJ9erVw9GjR/P0GcmrN89rbyrMRMDbsj6rby9787wGAOPGjcMXX3yR4zYqVqyo8jw/57aP9fkrjPr+vvemIOfm/J7b2rRpgzZt2iA1NRXBwcGYPn06unbtCicnJ/j6+ua4Tn7Ptx+isK8PcvsM5PW8BuT92oaICGBiiYiKKCGEukOAnZ0dRo8ejTt37qj8WHpbQEAANm3aBLlcjlq1auVaLuvC7e0faIsXLy6cgAsg6y73v//+q3KRvHv37o+2z/Xr12Po0KHK5xs2bAAA5cxkAQEBmDFjBp4/f46OHTsW2n4rVqwIOzs7bNiwAaNGjVL+PRITE/HXX38pZyMqLM2bN4dUKsXDhw/z3G3jfe+Nv78/fv31V6xbt05lsOK//voLiYmJypl83icvLdesrKzQu3dvXLt2DXPnzn3n9NP+/v7YsWMHXrx4oWwlAWS2htHT01PONFYYJBIJtLW1VX4IvXr1KsdZ4fLCyckJ4eHhCAsLU7bMSUtLw8GDB/O0vp6eHho1aoQrV67A09Mz17v77yKfUvP9hdQgr5/FgpzbHB0dcerUKTRp0kSZXHpXq7j8sLKygq6uLv7991+V5QWtI3mxceNGjBgxQvlehISE4OzZs+jZsyeAzPNP+fLlce3aNfz888+Fum9/f39Mnz4dly9fhre3t3L5mjVrIJFI0KhRo0LbV0Hqe17em8I4N7/Zck0mk+VapkGDBjA2NsbBgwdx5cqVXBNLhXW+zYvCvj5487u9efPmyuW7du3K8zbyem1DRAQwsURE9E4zZsx4b5nOnTtj/fr1aNmyJb799lvUrFkTWlpaePbsGY4fP442bdqgXbt2qFOnDkxMTPDVV19hypQp0NLSwvr165XdAdShZcuWMDU1Rb9+/fDDDz9AKpVi1apVePr06UfZn7a2NmbPno2EhATUqFEDZ8+exU8//YQWLVrAz88PAFC3bl18+eWX6NOnDy5evIj69etDX18fL1++xOnTp+Hh4YGvv/463/vW0NDAr7/+im7duiEgIAADBw5EamoqZs6ciZiYmDz9rfPDyckJP/zwAyZMmIBHjx7hs88+g4mJCcLCwvDPP/8oWwdlyct707RpUzRv3hzfffcd4uLiULduXfz777+YMmUKqlWrhh49euQpNg8PD5w4cQK7d++GjY0NDA0NUbFiRdSqVQsBAQHw9PSEiYkJbt++jbVr1773h92UKVOwZ88eNGrUCJMnT4apqSnWr1+PvXv34tdff4WRkdGHvZlvCAgIwPbt2zFo0CAEBgbi6dOn+PHHH2FjY4P79+/ne3udOnXC5MmT0blzZ4wePRopKSmYP38+5HJ5nrcxb948+Pn5oV69evj666/h5OSE+Ph4PHjwALt37/6gMW7UKa+fRTc3N7i6umLs2LEQQsDU1BS7d+/G4cOH37l9Gxsb/P3332jevDnq16+Pw4cPo0qVKh8cd9a4UCtWrICrqyuqVq2Kf/75R5mo/RjCw8PRrl07DBgwALGxsZgyZQp0dXUxbtw4ZZnFixejRYsWaN68OXr37g07OztERUXh9u3buHz5MrZu3VqgfQ8fPhxr1qzB559/jh9++AGOjo7Yu3cvFixYgK+//jpf46zlRX7r+/vem8I6N3t4eAAAfvnlF7Ro0QKamprw9PTETz/9hGfPnsHf3x/29vaIiYnBvHnzVMYwyklhnW/zorCvD6ytrdGkSRNMnz4dJiYmcHR0xNGjR3Oc0TA3eb22ISICwFnhiIiyvDkr3Lu8PSucEJmzBc2aNUtUrVpV6OrqCgMDA+Hm5iYGDhwo7t+/ryx39uxZ4evrK/T09ISFhYXo37+/uHz5co6ztOQ0w0xus7i9LbeZf96eOUYIIf755x9Rp04doa+vL+zs7MSUKVPEsmXLcpwVLqdtIodZvrJmnpk5c2a2Y/r3339Fw4YNhUwmE6ampuLrr78WCQkJ2ba7YsUKUatWLaGvry9kMplwdXUVPXv2FBcvXlQ5nsqVK7/3/XhTUFCQqFWrltDV1RX6+vrC399fnDlzRqVMQWaFy63eBAUFiUaNGokyZcoIHR0d4ejoKAIDA8WRI0eUZfLz3iQnJ4vvvvtOODo6Ci0tLWFjYyO+/vprER0drVLuXbM/Xb16VdStW1fo6ekJAMo6MXbsWFG9enVhYmIidHR0hIuLixg+fLiIiIh47/tw/fp10apVK2FkZCS0tbVF1apVc52V60NnhZsxY4ZwcnISOjo6wt3dXSxdujTHz0ZeZoUTQoh9+/YJLy8vIZPJhIuLi/jjjz9ynRUut9gfP34s+vbtK+zs7ISWlpawsLAQderUET/99FOejvVTy+v5Toi8fRZv3bolmjZtKgwNDYWJiYno0KGDCA0NzXUWtjdnxIqJiRF169YVpqam74wnt9m0cprFMjY2VvTv319YWVkJfX190apVK/HkyZM8xSNE7ufgt885WeeKtWvXiqFDhwoLCwuho6Mj6tWrp/L+ZLl27Zro2LGjsLS0FFpaWsLa2lo0btxYLFq0KNvx5OVvkyUkJER07dpVmJmZCS0tLVGxYkUxc+bMbLPt5XdWuA+p7/l9b/Jybn7XjGqpqamif//+wsLCQkgkEmWd2LNnj2jRooWws7MT2trawtLSUrRs2VKcOnXqve9BXs+3hTErXF6vD/IyK5wQQrx8+VIEBgYKU1NTYWRkJLp37y4uXryY5+sNIfJ+bUNEJBGiCPQ3ISIiKsV69+6Nbdu2ISEhQd2hEBEVihMnTqBRo0bYunUrAgMD1R0OERF9RHmf25WIiIiIiIiIiOgNTCwREREREREREVGBsCscEREREREREREVCFssERERERERERFRgTCxREREREREREREBcLEEhERERERERERFQgTS0REREREREREVCBMLBERERERERERUYEwsURERERERERERAUiVXcAxV1oaCgiIiLUHQZRvpmbm8PBwSHH11ivqbhivaaSiPWaSiLWayqJWK+pJHpXvc7CxNIHCA0Nhbu7O5KSktQdClG+6enp4fbt29lOEqzXVJyxXlNJxHpNJRHrNZVErNdUEuVWr9/ExNIHiIiIQFJSEtatWwd3d3d1h0OUZ7dv30b37t0RERGR7QTBek3FFes1lUSs11QSsV5TScR6TSXRu+r1m5hYKgTu7u7w9vZWdxhEhYr1mkoi1msqiVivqSRivaaSiPWaSioO3k1ERERERERERAXCxBIRERERERERERUIE0tUqJYvXw47Ozs0btz4k+0zMDAQgYGBn2x/VPJdvnwZ/fr1Q40aNeDs7IyqVauiVatWmDp1qrJMrVq10LNnTzVGSZR3t27dwsiRI1GnTh24urrC1dUVdevWxZgxY3Dt2jVludmzZ8POzk5l3VWrVmHz5s3Ztnn27FnY2dlhz549hRbnx9gmlR55reefkp2dHWbPnq18nlXHz549q5Z4SP02b94MOzu7HB8//PCDslytWrVUXnNxcUHdunXx/fffIyoq6r37yaprdnZ2OZ7DAaBDhw6ws7NDrVq1Cu34qOTZs2cP7OzssHPnzmyvNWnSBHZ2djhx4kS21+rUqYPmzZsDyDwXTpgwodBievXqFWbPno0bN24U2jbpw3CMJSpUmzZtAgDcvXsXly9fZh9iKnaOHDmCPn36wNfXFxMnToSlpSXCw8Nx7do17Ny5E1OmTFF3iET5snbtWkycOBGurq7o168fKlSoAIlEggcPHiAoKAgtW7bEmTNn4OTkhC5duqBhw4Yq669Zswampqbo1KmTeg6AKA/yU88/pV27dsHGxuaT7pOKhzlz5qBcuXIqy6ytrVWe16hRA5MmTQIApKSk4N9//8Xs2bNx/vx57N+/P0/7MTAwwKZNm7Kdw0NDQ3Hu3DkYGhp+wFFQaVCnTh1IJBKcPXsWbdq0US6Pjo7GnTt3oKenh7Nnz6pcP7x48QIhISH48ssvP0pMYWFhmDNnDuzt7VGlSpWPsg/KHyaWqNBcu3YNt27dgr+/P44ePYpNmzYxsUTFzsKFC+Hg4IANGzZAKv3vFNmmTRtMnDhRjZER5d+FCxcwfvx4+Pv7Y8mSJdDW1la+5ufnh969e2P37t3Q1dUFANja2sLW1lZd4RIVSH7r+afk4+PzyfdJxYObmxuqVq36zjJlypRRqUN169ZFQkIC5s6di4cPH8LV1fW9+2ndujU2bNiAR48ewcXFRbl806ZNsLa2hru7O+7du1fwAymA5ORkyGSyT7pPKjhTU1O4ubnh3LlzKsuDg4MhlUrRpUuXbK0ws57XqVPnk8VJ6sWucFRoNm7cCAAYP348qlevjp07dyI5OVn5+tOnT2FnZ4cFCxZg3rx5qFGjBlxcXNCiRQucOnVKZVtZ3TFu3LiB/v37o2LFinBzc8OQIUMQGRn53ljS0tIwd+5c1K9fH87OzvDw8MDw4cPztC6VbtHR0TAxMVFJKmXR0Hj3KXPVqlVwcHDArFmz8Ntvv8HBwQHPnz/PVm7EiBGoXLkyUlJSAPzXrW7//v1o0qQJXFxc4Ovri+XLl2dbNz4+Hj/88ANq164NJycn+Pj4YPLkyUhKSirgEVNJNn/+fGhqauKXX35R+bH9platWinvkr/dFa5WrVq4e/cuzp07p+xS8XaXiYyMDMyYMQPe3t6oWLEiOnXqhAcPHqiUqVWrFoYNG5Zt37l1ZU5NTcX3338PLy8vuLq6on379mzuTrnKbz0fNmwYypcvjwcPHqBr164oV64cqlWrhj/++AMAcOnSJbRt2xblypWDn58ftmzZorKtyMhIjBs3Dg0bNkT58uXh6emJDh064Pz589n2+3ZXOKIPldXCSEtLK0/l69WrB1tbW5XucAqFAlu3bkWHDh0gkUiyrbN7924EBATAzc0Nrq6u8PX1xYgRI1TK3L17F126dIGrqys8PDwwfvx4HDlyJFtXz8DAQDRu3BjBwcFo3bo1XF1ds22Lir46derg4cOHCAsLUy47d+4cqlatisaNG+Pff/9FQkKCymuamprZrhm2bduGBg0awNXVFU2aNMHhw4dVXn/8+DGGDx+OunXrwtXVFT4+PujVqxdu376tLHP27Fm0bNkSQOY1ddb1Sda5NiQkBF9//TW8vb2VQ1p07NhR5Tpi586d6NKlC6pVqwZXV1c0aNAAP//8c7br6azvi7t376Jjx44oV64cPDw8MGHCBJXfucB/3f3Wrl0LPz8/ODs7o2HDhjl2ISyJmFiiQpGcnIydO3fCy8sLbm5u6Ny5MxISErB79+5sZVeuXIkTJ05g6tSpmD9/PjQ0NNCjRw9cvHgxW9l+/frByckJS5YswYgRI3DgwAF07doV6enpucaiUCjQt29f/Pnnn2jXrh1Wr16N8ePH4+TJkwgMDMx2EiB6k4+PD65cuYJJkybh8uXL76xrWYQQ+OGHH/D9999j5syZGDVqFHr06AGpVIp169aplI2OjlZ+mb159/zmzZuYMmUKBgwYgGXLlqF69eqYPHkyFi1apCyTnJyM9u3bY+vWrejbty/Wrl2LQYMGYevWrejduzeEEIX3RlCxJ5fLcfbsWXh6esLKyqpA21i+fDkcHR1RpUoV7Nq1C7t27cqW8JwxYwaeP3+OmTNn4tdff8Xjx4/Ru3dvyOXyAsc+Y8YMhIaGYubMmZg5cyZevXqFwMBAhISEFHibVDIVtJ5nZGSgf//+8Pf3x4oVK9CoUSNMnz4d06dPx7Bhw9C5c2csX74crq6uGD58OP7991/lutHR0QCA4cOHY82aNZgzZw4cHR0RGBjIsZMoz+RyOTIyMlQebxNCKF9LTEzEmTNnsGzZMtSoUQMODg552o+GhgY6duyIbdu2Kc/Lf//9N16+fJljF+eLFy/i66+/hoODAxYsWIA1a9Zg2LBhKvGFhYWhffv2ePjwIaZPn4558+YhISEh15bdYWFhGDJkCNq1a4e1a9eiV69eeYqdio6slkdvtlo6e/YsateujRo1akAikagk18+ePQsPDw+UKVNGuezo0aNYtWoVRo0ahaVLl8LY2Bj9+/dX+W4PCwuDiYkJxo8fj/Xr12PatGmQSqUICAhQ3rTy8PDAnDlzAADffvut8vqkS5cuAIAePXrg33//xYQJE7Bx40ZMnz4dVapUQVxcnHI/jx8/RuPGjTFr1iysW7cO/fv3x+7du3OsmxkZGejRowf8/PywYsUK9OnTB+vWrcNXX32VreyhQ4ewYsUKjB49GosXL4a9vT0GDRpUKsaOZFc4KhR79+5FXFwcOnfuDCCz2e2UKVOwadMmdOzYUaWsQqHAxo0blT+qGzZsiFq1amHWrFnKMZqytGjRQvkl1aBBA1hYWGDw4MHYvXs3vvjiixxj2b17N44fP46lS5cqs9kAUKlSJbRs2RJbtmzhFxrlaty4cXjw4AFWrFiBFStWQEtLC1WrVkXTpk3Rp08f6Ovrq5RPTk7G0KFDcfr0aaxduxb16tUDAJibmyubnw8fPlx5F33jxo1IS0vLVgfDwsJw8OBBVK5cGQDQuHFjREREYO7cuejVqxdkMhmWL1+O27dvY8+ePcrm8/Xq1YO1tTW+/PJLHD9+/JMOnE9FW1RUFFJSUmBvb5/tNblcrpKI1NTUzPGudZUqVaCrqwtDQ8Ncu/SUL18ev//+u/K5hoYGvvrqK1y9erXA3YDMzMywfPlyZUw1a9aEn58f/vjjD8ycObNA26SSqaD1PC0tDWPGjFFeJ9SpUwdHjhzBH3/8gYMHDyrH7PD09ETVqlWxY8cOeHp6AgDKlSuH6dOnq+ynYcOGePr0KVasWMGuH5QnrVq1yrYsJCREpcX0sWPH4OjoqFKmWrVqWLJkSb721alTJ8ybNw/Hjx9HkyZNsGnTJmXL57ddvHgRQgjMmDFDJSnwZhJq6dKliImJwfbt21GhQgUAmdct3bp1w9OnT7NtMyYmBosXL4afn1++4qaio3bt2tDQ0MDZs2fRtm1bREVF4c6dO5g4cSL09fXh4eGBs2fPwt/fH8+fP0doaCgCAgJUtpGSkoJNmzbBwMAAQGaCyNvbG7t378bgwYOV+6ldu7ZyHblcDn9/fzRq1Ajr1q3D999/D0NDQ7i5uQEAHB0dVa41oqKi8PDhQ0ydOhXt27dXLn/zNyEAlVbUQgjUqFED5cuXR/v27XHr1i1UqlRJ+XpaWhoGDhyIfv36AQDq168PqVSKX375BRcuXECNGjVU9r9v3z5YWFgAAPz9/dG4cWNMnz492/tR0rDFEhWKTZs2QVdXVzmgm76+PgICAnD+/Hk8evRIpWyLFi1UWmoYGBigadOmOH/+fLY73G8nj1q1agWpVPrOO4JHjhyBkZERmjZtqnIXqHLlyrC0tMzWP5joTaamptixYwf27duH8ePHo1mzZnj06BGmT58Of39/lZlYoqOj0bFjR1y9ehU7duxQJpWy9O/fHxEREcq7FAqFAmvWrIG/vz/Kli2rUrZixYrKpFKWtm3bIj4+HtevXweQWbezyr1Ztxs2bAiJRMK6TXn22WefwdHRUflYvHhxgbfVrFkzlefu7u4AgGfPnhV4m23btlVJdNnb26N69epsDUL58q56LpFIVBLxUqkUTk5OsLKyUhkI1sTEBObm5tnq85o1a9C8eXO4uLjAwcEBjo6OOH36NO7fv//xD4xKhHnz5mHfvn0qj7e74desWVP5WlBQEGbPno3IyEh07NgxTzPDZXFwcICvry82b96MqKgoHDp0SHkz+G1eXl4AgK+++gq7du3Cy5cvs5UJDg6Gm5ubMqmU5c2Bnd9kbGzMpFIxZ2xsjEqVKimvNYODg6GpqalMqtSuXVv5HZ3b+Ep16tRRJpUAwMLCItv5NSMjA/Pnz0fDhg3h5OQEBwcHODk54fHjx9m62efExMQETk5OWLhwIRYvXowbN25AoVBkKxcSEoJvvvkGXl5eKFu2LBwdHZWJqJzO4+3atcvx+ZkzZ1SW+/n5KZNKQOYNjdatW+PJkyd48eLFe+MvzphYog/2+PFjBAcHw9/fH0IIxMbGIjY2Fp9//jkAZJvi9M0PWxZLS0ukpaUhMTHxnWWlUilMTEyUzdBz8vr1a8TGxsLJyUnlgtLR0RHh4eH5+iKm0qtq1ar45ptvsGTJEly+fBkDBgzA06dPsWDBAmWZR48e4cqVK2jUqJHyzsmbqlSpglq1amHVqlUAMhNDT58+RZ8+fbKVze1zAfzX7eL169e4fft2tnpdoUIFCCFYt0mFqakpdHV1c0zw/Pnnn9i3bx9Wrlz5wfsxMTFRea6jowMAyjHECiKr7r/JwsLined+Kp0KWs9lMlm2wby1tbVhbGycrayWlhZSU1OVzxcvXoxx48YpW47s3r0b+/btQ6NGjT6o3lPpUr58eVStWlXl8TZDQ0PlazVq1EDnzp3xxx9/4P79+/m+IdClSxccPnwYS5Ysga6urvI6/W21a9fGihUrkJGRgWHDhqF69epo3LgxgoKClGWio6Nhbm6ebd2crmWAnM/pVPzUqVMHjx49wqtXr5Rd3bJa8teuXRs3btxAXFwczp49C6lUipo1a6qs//b1ApB53n3zvDl16lTMnDkTzZs3x6pVq7Bnzx7s27cPlSpVytP5VSKRYPPmzWjYsCEWLlyI5s2bw8PDA5MmTVKOAZWYmIgvvvgCV65cwZgxY7Bt2zbs27cPy5YtA5D9+kUqlcLU1FRlWVZdf/u6JLfrl5zKljTsCkcfbNOmTRBCYO/evdi7d2+217du3YoxY8Yon79+/TpbmfDwcGhra2frZvT69WuVaXozMjKUgyvnxtTUFCYmJli/fn2Or7+9D6L30dLSwogRI7B06VLcvXtXudzHxwcBAQEYNWoUgMxxYd4e4Ltv374YOHAgrl+/jpUrV8LFxQX169fPto/cPhfAf1/EpqamkMlkuQ4E+/aXHpVumpqaqFu3Lk6ePImwsDCV8Wey7jLn1GXhY9DR0UFaWlq25VFRUTnW26y6/6bXr1+/89xPpZM66vn27dvh6+uLGTNmqCx/c+Baoo8lq4vOrVu38rVeixYtMGHCBPz555/o2rXrO2dla968OZo3b47U1FRcvnwZf/zxB7755htl61ETExNERERkWy+nczeAHLtaU/FTp04dLFmyBOfOncO5c+dUWn1mJZGCg4OVg3oX5DfX9u3bERgYiHHjxqksj4qKgpGRUZ62YW9vr7xWfvjwIXbv3o05c+YgLS0Nv/zyC86cOYNXr15h27Zt8PX1Va4XGxub4/YyMjKyXa9kXbe/fV2S2/VLTmVLGrZYog8il8uxbds2ODk5YevWrdkeAwcORFhYGI4dO6ZcZ//+/SqZ4ISEBBw+fBg1a9aEpqamyva3b9+u8nz37t3IyMhQOQm8rUmTJoiOjoZcLs92J6hq1aooV65cIR09lURvznbxpqxmsW8PDtuxY0csWLAAW7ZswbfffputO2eLFi1gZ2eHqVOn4tSpU+jVq1eOF1h3797FzZs3VZYFBQXBwMAAHh4eADLr9pMnT2BiYpJj3X67ex3RkCFDIJfLMXbs2DwNRJ+Tt+8mFkTZsmVVZnQBMi/2Hj58mGP5nTt3qoyN8+zZM1y8ePGd534qvQqjnueHRCJRtszLcuvWLVy6dOmj75soa2YrMzOzfK0nk8kwbNgwNG3aFD179szTOjo6OvD19cX48eNV9l27dm3cuXMH9+7dUym/a9eufMVExUvt2rWhqamJvXv34u7duyrfyWXKlEHlypWxdetWPH369IPGmnt7ds8jR47g1atXOZZ53/WJq6srhg0bBjc3t2yzy769n7cn3HnTjh07cnz+9nGePn1a5WaxXC7Hrl274OTkBFtb23fGWtyxxRJ9kGPHjuHVq1eYMGFCjicQNzc3rFq1Cps2bcL3338PIHNg1y5duuDLL7+EQqHAggULkJCQoGz18ab9+/dDKpWifv36uHv3LmbOnIlKlSrlOOBhljZt2mDHjh3o0aMH+vfvDy8vL0ilUrx8+RJnz55F8+bN0aJFi0J7D6hk6datG2xsbNCkSROUK1cOCoUCN2/exJIlS6Cvr68cuO9NAQEBkMlk+PLLL5GSkoI///xT+WWlqamJ3r17Y9q0adDT08s2mH0WKysr9OnTByNHjoSlpSW2b9+OkydPYsKECcq7igMGDMC+ffvwxRdfYMCAAXB3d4cQAs+fP8fff/+NgQMHwtvb++O9OVTs1KhRA9OmTcOkSZPw2WefoVu3bqhQoQI0NDQQHh6ubGX65pgHb3Nzc8OuXbuwc+dOODo6QkdHRzmOUl61b98eQ4YMwbhx49CyZUs8f/4cCxYsyPWHUUREBPr164euXbsiPj4es2bNgo6OjnJwT6I3FUY9z48mTZpg7ty5mDVrFmrXro2HDx9i7ty5cHBwyHFmL6KCiouLUyYsMzIycP/+ffz+++/Q0dFB79698729gQMHYuDAge8sM3PmTLx8+RJ+fn6wsbFBXFwcli1bBi0tLWUioX///ti0aRO6d++OUaNGwcLCAjt27FCOgfN2620qGQwNDeHh4YEDBw5AQ0NDZdBqIDPxlNWdrKCJpSZNmmDr1q0oV64c3N3dcf36dSxcuFClBwsAODk5QVdXFzt27ED58uWhr68PKysrREVFYeLEiQgICICzszO0tLRw5swZ3L59G9988w0AoHr16jA2NsbYsWMxYsQISKVS7NixI9dWgNra2li8eDESExPh5eWFixcvYt68eWjcuHG27n6mpqbo2LEjhg0bBplMhjVr1uDBgwcqQ2mUVEws0QfZtGkTtLW1c5yuFMj8cH322WfYu3ev8sPcp08fpKamYtKkSYiMjESFChWwZs2abCcnAFi2bBlmz56NNWvWQCKRoGnTppg6dWq2DPObNDU1sXLlSixbtgx//fUX/vjjD2hqasLGxga1a9fOcSwcoixDhw7FoUOHsHTpUoSHhyMtLQ2Wlpbw8/PDkCFDUL58+RzX8/f3x9q1a9G7d2/07dsXS5cuVSaEWrdujWnTpqF9+/YqM6y8qXLlyujUqRNmz56Nx48fw8rKClOmTMGXX36pLKOnp4cdO3bgjz/+wPr16/H06VPo6urCzs4Ofn5+bLFEOerZsyd8fHywfPlyLF26FGFhYZBIJLC2tkb16tWxefPmdw6qOmrUKISHh2PMmDFISEiAvb29ypTCedGuXTuEhYVh7dq12LJlCypWrIjp06fjt99+y7H82LFjce3aNYwYMQIJCQnw8vLCwoULc5zBiAj48HqeH0OHDkVycjI2btyIhQsXonz58pgxYwb279/PSRSoUF24cAGtW7cGkHl9a21tDS8vLwwdOlRlkPnCVK1aNfz777+YNm0aoqKiUKZMGXh6eirP3QBgbW2Nv/76C1OmTMG4ceOgq6uLFi1aYNSoURg2bFiu1zpU/NWpUwdXr15FlSpVYGhoqPKar68vli5dCm1t7Rx/1+XFDz/8AC0tLfzxxx9ITEyEh4cHli1bhl9//VWlnEwmw5w5czBnzhx07doV6enpGDFiBHr16gVHR0esXr0aL168gEQigYODAyZPnoy+ffsCyPx9unr1avzwww8YMmQI9PT00Lx5cyxatAjNmzfPFpNUKsXq1asxadIkzJs3D7q6uujatSsmTZqUrWyzZs1QoUIF/Prrr3j+/DkcHR3xxx9/5DqwfYkiqMAuXbokAIhLly6pO5RiITQ0VNja2oqFCxe+t+ysWbOEra2tiIyM/ASRlT7vqrus14Vv+fLlwtbWVty5cyfH12vWrCl69OjxiaMqeVivqSRivaaSiPW6ZBo9erQoX768SE1NVXcoasF6XfJ8++23oly5cnkqa2trK8aPH/+RI/r08lp32WKJiOgjuXHjBkJDQ/Hbb7+hefPmyjt9RERERMXZb7/9BisrKzg4OCApKQlHjhzBhg0b8O23376zZwERlUxMLBERfST9+vXD69evUbNmzWyzBxEREREVV1KpFAsXLsTLly8hl8vh7OyMKVOmoH///uoOjYjUgIkl+mTKli2L58+f56nsyJEjMXLkyI8cEdHHlddxaPI7Xg0RERGROg0ZMgRDhgxRdxhEH9XcuXMxd+7cPJXN6+/ckopD9hMRERERERERUYGUuhZLSUlJ2Lx5Mx4/foxHjx4hLi4OnTt3RteuXdUdGhERERERERFRsVLqWizFx8fj0KFDSE9PR+3atdUdDhERERERERFRsVXqWixZWlpiw4YNkEgkiIuLw6FDhwq8reTkZADA1atXCyk6ok/j9u3bub7Gek3FFes1lUSs11QSsV5TScR6TSXRu+r1m0pdYkkikRTatmJiYgBkzvxEVNzo6enB3Nw823LWayrOWK+pJGK9ppKI9ZpKItZrKolyq9dvKnWJpcJkY2MDAFi3bh3c3d3VFsezZ89gb2+vtv1/qOIeP1A8j8Hc3BwODg7ZlrNeF47iHj9QPI+B9frjKu7xA8XzGFivP67iHj9QPI+B9frjKu7xA8XzGFivP67iHj9QPI8ht3r9JiaWCoG7uzu8vb3Vtn9jY2O4uLiobf8fqrjHD5SMY3gb6/WHKe7xAyXjGN7Gev1hinv8QMk4hrexXn+Y4h4/UDKO4W2s1x+muMcPlIxjeBvr9Ycp7vEDJeMYcsLEUj69fPkSL1++BJD3/oYlRVRUFJKTk2FnZ5en8olpcuhra37kqIiIqChTKBSIiorC69evkZGRAQ8Pj/euk5qhwJOYVFgZaMFYl5cq9PEIIZCRkYGkpCTo6enlrbxCQEuz1M1/Q6WMEAJpaWlISUlRPpKTk5X/19TURI0aNd67HYUQSJcLpCsE0uRv/l+BdPn/lymE8v/melK4W8g+wRES/UcIAblcjvT0dKSlpSEuLg4WFhbQ1dV95zppcoFUeeb3glwhIBcCcgWU/2b8f5lCZH4W5AogNCoDUboJ/18GyBX/vS7VkKCug+EnPPLCw6u1fFq8eDGmTp2q7jA+KiEEQkNDceXKFVy9elX5b2hoKPr06YMVK1a8c/17kcmYeOwZ7kak4MpXVaBRiONaEambEOK9Y7UlpyugEOITRUT0aWVkZCAiIgKvX7/O0yMyMhIKhQIA4OXlhStXrgDIvPnwMDoVD6JS8DAqFQ+jUvAwOvPf0Ng0CAAbA13RsbKZGo+WirrExETcvn0bycnJSE1NzfZ4+fIl9u/fj5cvXyItLU35o+HNhxACUqkUoVGJuB+VgvDEDLxOTEdUcgaik+W48ToJ9yJSEJcqR1yqHK0qmuCvTuXVfehEKoQQSExMRGRkJCIjI3Hjxg38888/iI6ORnx8PBISEpSPrOdxiUmITVUgPl0gWWghRaKFNA0dpGnoIB2agJYuJFq6gJYuoKXz/38zH7IyJqh0RYbkdEW25NCbiSN5Pi+HelY1x8q2Ja81R0kkl8sRERGBmJgY5fn07XNs1vOMjAxkZGRALper/JuSocDLJIHn4ZGIT05FSrocciGBHBLlv2kSLaRo6iFDQ4pkuQQSbRnkEikUkEAAkCsUSEtLh0IAQgJAZC7PfGReswtIICT//R8AhEQCAY3MfyUagEQT0NAAJBqAhiZMTM1gYmaBDIVQeWQlSdPyW7lVxKg8kwDQkAC2htp4MtzrA7arPkws5dPAgQPRunVrAJktlrp3767miD5Meno67ty5ky2JFBMTgzJlysDLywteXl5o164dvLy8UKlSpVy39TI+DT/8/RzLL79GQ6cyWN7GmUklKhFSU1Px+vVr3LhxAz/99BP27t0LIyOjzNcyFIhPkyt/cJx8Eo9fzrzEZB9dlHNVc+BE75H1QyQrUfS+f8PCwhAbG6tcXyqVwsLCQuVhY2MDT09P5XOZiQVitEwRptBDSJIGmq29g9uvk/EiPh0AYCaTwtVUB66muqhT1gA9qpqjnKkOXE10YanPy5TSKCUlBc+ePUNiYiKSk5ORlJSE5ORk3L59G48ePUJsbCxiYmIQGRmJy5cvIz09sy5pa2tDR0cHOjo60NXVVf6/YWN/tOg+EOmaOsjQ0EG6hjbSJFpIlUiRBiki4pPwTGEE+zlXoSEBLPS0YKEvhZlMChOZFC7GuuhSxQwmMinK6GjC1lBLze8QlQZyuRzR0dF48eIFnj59iqdPn+Lhs5d4EZOEqIQURCekIDYlHXEpGUhIkyMpA8jQ1IZEWw/QlkFD1wA6BkaQ6hlAQ8cW0JZBaOtAWOpAbqONdIkWMiSq51gdiQIGUgFLKaCnJYFMqgF9bQ3oaWvCQFsKAx0pDHW1YKirDQMdKfS0NKCrpQFtTQm0NCTQ1tSAlvL/EmhpSpSvZf5fQ+U11XIa0JQU7kRLlHeJiYm4evUq4uLiEB4ejtu3byM+Ph5RUVEIDw9HckoqEoQWYuRSRMYnI1UhgVxLBugZA5pagKYU0JBCoikFtGXQNLIC9I2hoSmFREMKaGpCoqEJaGgDGhqQaGhCIdWFXMcQCqkOYAzAGJAIBTQgoAEFNIUCGlBAC3Loi1RoIwPa6Skw0NaElkQOqSQzISPV1ISurg6kmhqQSCTQkEigIcH///3v/xINCTSznmtIlK9rSzWgpakJrf//q62lCX2ZDMZGZaCnqwMtDQmkGhJINfD/fzPrrK5UA7rSzLor1ZBAUwJo/n8fmhqApiSzbNb/s+IIDXkCVxdnaEj+W14S6j2v2PLJxsZGOfhacZaUlISvv/4amzdvRmpqKuzt7eHl5QU/Pz8MGTIEXl5ecHJygobG+5t6J6crMPPMS8w8+xJu5rrY260imroafYKjIPowKSkpCA4+j8tPo/EkIh5hMQmIiEtCZGIKYpLTEZ8qR2IGkCbRAnT0AR19mDecgOqrnygTSW/frbDUl+K7uraoZ5aopqMiytmhQ4ewevXqbAmjlJQUZRkDAwOYm5vDwsJC+a+rqytq164Nc3NzZVc2CwsLWFpawsjICBKJBAoh8CI+HfciU3AvMgX3I5Nx9nUKboUm49mNNEiQBFdTOSpZyFDTVh+9vcxR0UwGV1MddnUrxrK6kSUmJuJZeBQi4xIREROX+YhNQGRMHCLjEhAdl4jkdDnS5AplV5g0hUC6Av89/t8dICY2Dq8jIgGJJPOusUQDEk1NaGlpQ0OqBZmeI6ztbaHlogMtbR006mkOU3NzaGhqQv5GVwPF/7seyIXAnrBkPHuZBgDQ0ZSgjI4mDHU0YfT/fzW1Zahla4J5bcxQ084AmhrF/wKf1C85ORnh4eEICwtDTEwMEhMTlS2G3vx/QkIC4pJSEJWYhtfxyYhMSEFUmgQJuuaAuRNgbA0NIyvA0B1CyxuwQOYDgBbk0NVQQF9TAkttCQx1pDCSacNETwcSeSqsTY1goK0BfS3NzASRlgYMtDP/r6+lASNdKUx0NWEik8JYVxPa7OJZIikUCoSGhuLZs2e4cOECUlJScPnWPTyKFwgTBojRMkWyhi6Eth40dfUh1dWHll5dwFgXcjNtyN20kP5G2kADArqagJ4UMJdpQk9bC9pSDWhrakBHmpmItDbQgrWB1v8TLv8lWLKea2hIoK+lAWNdTVS31YehjiZ0pZkJmnd59OhRsR+fSCbNTEqVNKXyau7SpUvKfsIA8PTpU5w5cwYAUL16dejo6KgzvI8uNDQUbdu2RWxsLLZv344aNWrAwsIi39sRQuCv29EYcygUGQqBRQFO6OJhxlZKVGRltc6Ii4vDw1dR6DdjOR6aVgNs3CBRZEAqT4WOXgZkhgroSwEHHU0Y62nDTF8XFkZ6sDI2gLmBLgx1NFEml4eeVubdkkePHqn7cImUEhIS0L17d9SuXRu1atVSSRy9+e+7xhJQCIHgmw8AI2sER6Xi/uUE3IuMwP3IFNyPSkFSugJaGhKUM9VBOTNdeFnroauHGSpbyuBmrgs9LY65V1wkJycjODgYly5dwulrd/BPtBYSNQ2Qoa0PubY+5LploJAZQegYZCbdtWWQSB68sQVNAEaZD4UcGiIDmsiAhoYCGhoKaEoV0MT/HxIBKRSQSjK7ARjbaKK8uRkM9PWgLZVCS6oJqabm/+/0Zt4Nzrrrm3Wn9787wTm/7u9cBh0qm8FCTwqdHC7mM3+oOH2id5dKIiEEbt68iQULFuDQoUMICwtDQkKC8nVdmR70LeygY1EWUjM7JDnXQbJxJch1ZcjQ14VCQ/X8qKuhgJs+UMFUG5WsjeBsYQhbQ23YldGCuZ4WDLUzE0TvSoKyXpc+CoUCDx8+xK1bt7D/0GGcvfscYakaiIYMGcb2gFV5aJpVArT1oCjnDwgBC81UVNGTw95IB+UcbFFGVxsyLQlkWprQ18pKRmbWNysDLdgYaCmvdYneVCoTSwsXLkR4eLjy+ZkzZ5SJpaVLl8LKykpdoX10V65cQfPmzVG1alUcOXIEpqamBdrOP88TMP7IU5x5moBRdWzwnZ8NDDhQN30CKSkpuHXrFmJjYxEbG4u4uLhc/42Li0NUQgqiNI0Qp2uOFEM7CEtnwLIcJAam0CgXgEE1bTGigQucTUp2QplKr1OnTmHAgAEoU6YMtmzZkmvyKEMhEBqbiicxqQiJSUNITOb/Q2PT8CQmFU/j0pAmF5AgGg5G2ihvposKZrrwczBX/t/BSOe9dxupaLp//z52796NgwcP4u9zF5Dh0QJ6Pp8jsVx9mGikwklXjjJaAkZagKmOBGa6mXeay+hKkRofjfIOtjDW04WlqREsTIxgoKuVp7vPRMVZSkoKVqxYgTlz5uDhw4fw8fHBd999Bzs7O1haWiJNzwwr7smx6WYMojMyx5oz0dVEdTt9dKxsBiMdTRjpar7xrxRGupolsjUDFT4hBDZv3ozfl63Gy3RtvEzTRmoZW2iUrQKJXRcoLLUg05DDQVfA3UIftZ1NYaxIQLmyNrDU14KbuQwyLdY1KhylMrG0bNkydYegNr///ju8vLywb98+SKX5+/MLIXDkURx+Of0CJ57E44tKJrgxyAOuprnf4SYqDLGxsbhw4QI2bNiAv/76C3FxcZBKpTAyMkKZMmVgZGQEwzJloGXphHQzZ6SU9UGcrgUiNU0QhcyZRay0M+BjKIGbmRY8rPThU9YY3o4W0NUuladBKiXWrVuHXr164csvv8SMGTNUkkrP49JwKjQep0PiceZpPG6GJ0MuMlt92JfRhpOxDhyNteFb1gCdPczgaKQNjfhw1K3iyu4SJcyrV6/g4eEBJ2dnlG/VH0YNv4eOjg66e1qgfSVTeNvovfPudEnomkCUV0lJSTh8+DAuXbqEjRs3IiYmBsOHD0ePHj1QtmxZZbldd6PRZdsDeFnrYXU7F1SxlMG+jDZnTKYPdvLkSZw8dRprztzFY3MfKOpNhQQCltpyVLLUR10XM9S2N0ANO32Y66mOCZd5vjZWT+BUovEXVSlz8+ZNtGvXLl9JJYUQ2H47Gr+efoFrYcno7mmG64OcOBUoFarIyEhcu3YNjx49wsOHD/Hvv/8iPDwcjx49QlRUFLS0tNCiRQssXboUn332GQwNDZGYrsCRR3E4+CAGBx/EIiQ2DWYyKTytZKhjqYcqVjJ4WOqhsqWMLeqoVElKSsLevXsxbNgwTJ8+HWPGjEF8qhzbbkVh770YnAqJx+OYVJjJpKjrYIDunubwsdWHk7EO7Ay1cp1K/dGjSCaVSqAdO3bA1NUDruPX4fCjOIypZYOx9WzYfZHoDRkZGTh27Bj69euH2NhYeHt7o0uXLhg9ejQMDVWnB192ORyD9jzBj43tMaauDbsNUaG5f/8+GoycC1mjvsio2QhtyxlgWH0neNvos/URqRUTS6WEEAJLly7F5cuXMW3atHyt+8OJ55h19hUG+Fjgr07lUdaIXYaocCUnJ6NKlSqIjIyEk5MTXFxcYG5ujsaNG8PFxQUuLi5wdXWFgYEB0uWZyaR1hx9i550YaGoAjZ3L4Ds/WzRzNYKTsTYv4KhUUigU2L9/PzZs2ICdO3dCQ0MDX3TpgTINeyBgw10cfRQHmVQDn5U3whg/G/g5GMLNXJfj4hGWHr2KyA7zkC4XuPZ1FVQ0540jojfFxMTAw8MDz549Q8eOHbFq1SrIZDl/Tvbdj8GgPU+wpLUzenvlfwxToncZtPkyND8bhglNXTHAxxKW+pylkooGJpZKgdjYWHz55ZfYtWsXFi5cCH9//zyv+zwuDbPOvsKyNs7oXMXsI0ZJpVnW7IQxMTHQ09MDoNq1IjldgQMPYhB05yH23otBUroCARWNsSHQFZ+VM2ILCir1Tpw4gTFjxuD69eto3bo1/li5HrfLeGLJlSgcOfUKbd1MsKtLBTRwMuTnhVRsv/IU19y6o6sTsKZHRSbmid4SGRmJSZMm4dmzZ3j58iWsra1zLJecrsDXex9j840oTKxvx6QSFboVV17jWLoDumhewoT69dUdDpEKJpZKsDt37iAoKAiLFy+GTCbDhQsXUKVKlXeuI4TAs7g0XH2VhGuvkrD7Xgw8rGToVLlgg3wTvY8QAn/88Qf69eunTCq9KTYlA41X38GDqBS0KG+M31s6oWV5IxhxinIq5YQQOHDgAGbPno3jx4+jT58+2LFjB2K0TFB72S04GsVjfgtHdK5immvXNird/rocgi47n0Pv351Y+f0MJpWI3nD16lVMmzYNmzdvho2NDfbs2aOSVEpOV+B1UjpeJ2bgdVI6Fl0Mx/WwJAR1Lo9mrkZqjJxKooMPYvD17sdQ7J6OcWunqzscomz4y6wEUSgUOH/+PIKCghAUFIR79+7Bw8MDvXv3xqhRo6Cvr69SPkMhcDciGVf+n0S69ioJV18lITI5A4baGqhqrYdadvoYXMuKF5v00WzduhVXrlzB1q1bs72WmqFA+833kaEQeDLMCyYynrKodAsLC8PFixdx8eJFbNmyBQ8ePEC3bt1w/fp1VKpUCUIIjPzrIXzLGuBA94rs5kY5Wrt2LebtOosrbl1h8eAIln3ZKN8TehCVRGlpabh16xaCg4MxbNgwNG7SFDMWr4W9d338G5GKjdsf4mZ4Eh5GpSIxXaFcT09LA87GOtjXrSK7klKhSkpKwuKdxzHujhH0ru2Ea8oDVKpUSd1hEWXDq4hiLiUlBcePH8cvv/yCnTt34vXr1/Dz88PAgQPRpk0buLq6Kssmpcux+moErrxMwrWwJFwPS0KqXMDWUCsziWRvgC+rW8LLWg8uJjr8QUIf3bVr19CnTx/MnDkTzs7OyuUZCoGgRylYvv8GEtMUON2vEpNKVCr9888/OHr0KC5cuICLFy/i6dOn0NfXh7e3NwIDA/H111+r3EGffPw59t6LwfHebjyHU66mbzqM+9UHIsAkBpsnjFOZLZCotHn9+jVGjRqFM/df4bG+K4SFC7TN7KA15gAOS/Vw8AFgGPoYVSz1UMlChj5eFihnpgtLfSks9LRgoS/lQPf00XzRZxAOO3RC2dQ7GNXECQFz9/GGPxVJ/KVWDEVHR2Pv3r0ICgrCgQMHkJGRgebNm2P69OkICAiAhUXOfbonHH2GDdcj4e9SBoGVTPFjY3t4Wetx0DdSi7CwMLRp0wbt2rXD8OHDAWS2UFp9LQIzz7xEWEIahtSyxrDa1rBgHaVSSC6Xo0GDBqhYsSL8/Pzwww8/oEaNGnBzc4OmZvYfMWdC4zH91Avs714R1W0N1BAxFQcKIfCgYns0NU1C0NCW/IFCpZYQAg8fPkT7YVPxoEI7pLjYw81QjkaupihnaQTN5GhUK2cPRyNt2JfhxCCkHucsGqO8kSauj+vLbu1UpDGxVEyEhIRg586dCAoKwsmTJ2FsbIxWrVph3bp1KFeu3HvHTnoSk4pFF8OxpUM5tKpo8omiJsrd0KFDYWVlhaVLl0IikeBBVArabbqPVwlp+LaWNT63TEY197LqDpNIbW7fvo2UlBTs2bMH9vb27y0/9cRzdKpiiqYc24Pe4eiDaGQYWOCrKgr+UKZSKSMjA2vWrMFP06bhsWVNaDQbgu6VjTC5iStcTf9rvffoURJcHAzVGCmVdk9jU5Fg4YYuViFMKlGRx8RSEff8+XMMHjwYQUFBcHV1RZs2bfD999+jTp06yvEQHj16lG09IQTuRabgZEg8ToXG49ijONSw1UdABeNPfARE2QkhsG/fPqxZswYymQxHH8Wi09YH8C1rgNN93WGkK82xXhOVFunp6Zg2bRrq1KmTa1IpQyFwPzIF/4Yl4Z/nCTj+JA43Bnl84kipuFkY/Ay4ewqefTupOxSiT0oIgR07dmDChAl4HhUPh0ELYCC1xuJWLujqaa7u8Iiy2XQjCpKoZ6heXf/9hYnUjImlIkqhUGDx4sX47rvv4OHhgYsXL8Lb2zvXu4tyhcC/YUk4HRqPkyHxOB0aj/DEDDgZa8PPwRDfN7JDOzcT3p2kIiEsLAwJCQmoXLkydtyOQudtDzHC1xo/NbaHpgbrKJVujx49QteuXfHw4UPs3LkTQGb3pTOhCfjneQKuhyfhRlgybr1ORqpcwEJPCk8rPSz43ImDxtI7RSdnYP+TZIjLO2FrO1zd4RB9UrNmzcLEiRPRYvgMRJnWh3YZHZxv74pKFjxvUtG0/t/XkF/bB/OuA9QdCtF7FZnEUnJyMtatW4czZ84gPj4e9vb2aN++PerXr//O9caPH48bN27k+vrq1athYmLyzrLVqlXD1KlTP+wAClFcXBw+//xzXLt2Db/88gsGDhwIDQ3V5o8JaXJceJ6I4GcJOHw3FlciLiMuVQ53c13UczTErGYOqOdoCAcjHTUdBVHunjx5kvkvTNHtr4eY1tgeo+raqDcooiIgNjYWzZo1g7OzM65fvw5tI3P8du4lFl8Mx+OYNHhayVDFUg/dPM1QxVIPnlZ6sDLgGGSUNxuuR8JAQ44yCU+gra2t7nCIPrqkpCTcvHkTJ06cwNgfpqPuLwexP14fE2paY2w9G2izexEVUTfCk3A9PAW4fhBmZmPVHQ7RexWZxNL06dNx//599OrVC7a2tjh58iRmzZoFIQQaNGiQ63pfffUVkpKSVJalpqZi6tSpcHV1VSaVslhbW2PEiBEqywwMitYgp9u2bcP9+/dx69Yt2NvbQ4jM7g7nniUg+GkCgp9l3rHWkEhQzVoPlcpoYnBdR/g5GHIgbiryFAoF1q9fD8uq9dDxrycYWsuaSSWi/5s4cSK0tbURFBSEc2EZaLfiKsz1pBjgbYm+3uawNmAygApGCIFll8PhrfECUfZ26g6H6KOLj49H+fLlERYWBuv6HWA4ejcSdI1wvrMzvKzZtYiKtg3/RqKKiQb+jXoGMzMzdYdD9F5FIrF08eJFXL16FSNHjlQmkTw9PREeHo6VK1fCz88vxxlwAMDBwSHbsqNHjyIjIwNNmzbN9pq2tjbc3NwK9wAK2e49e1Cj/QCseaSB4FP3cP5ZAiKSMmBjoIXaZQ3QzdMMte0d4W2jD5mWBh49egQXF1N1h02UJ99//z1WbwmC9rdbEVjJFNObvH9QYqLSICEhAUFBQZg4cSJuRAt8sek+Bte0YhdRKhSXXibielgyesddg14eBoMnKu5Onz6N2FQFumy6hW134zGuni0m1LdlKyUq8hRCYOONSLS2TMUNDQ0YGXFSDir6ikRiKTg4GDKZDH5+firL/f39MXv2bNy7dw/u7u553t7hw4chk8lQr169wg71o1IoFJg5cyZ2RZSBplcAIu5Fo7a9Abp6mMHX3gAORpzqlIq3Y8eOYebMmag9fS9idPWxMMCJdZoImeOOBQQEQCaToVLDAHy+/i66e5rjZ397fkaoUKy4HAF/lzKI2303T7MMEhV3+06cheSrtfg3Ih1n+1dCddui1UOBKDdnQhPwPC4NXrYRMDExyTYkClFRVCQSSyEhIbC3t8/WKsnZ2Vn5el4TSy9evMCtW7fQrFkzyGTZB+N79eoVunbtiqSkJFhaWqJevXro2LEjdHTUOxZRVFQUevXqhaMhSdBo/zM2dyiPdu5shUQlx759+9C2bVt8MWoG/oozwNkOzpw6lQjAvXv30KJFC1haWuLMmTMICApHQ+cy+L2lI5NKVCgS0+TYeCMSiwKcMG/hM/j4+Kg7JKKPblO0NUzsNHFxYBXoSnm9QcXHhusR8HcxgjzuX3aDo2KjSCSW4uPjYWVllW151thH8fHxed7W4cOHAQBNmjTJ9lqlSpXg5+cHe3t7pKWl4dKlS9i+fTtu3bqFadOmqS0bHBISgkaNGkHb2gU6nWdgjK8Nk0pUovz9999o3749xk76HhuN/DGkgjFq2PHOIZVeQghcuHABK1euxPr169G4cWNs2LABR5+m4t+wJGzrWI7d36jQbLsVBakG0NbNBKOfPWOLJSrRFAoFRi3chiibalhSQzCpRMVKmlyBbbeiMKe5I14ciGRiiYqNIpFYAvDOu7J5vWMrl8tx7NgxODg45DiOUvfu3VWeV69eHZaWlli5ciXOnz8PX1/f9+7j5cuXePnyJQDg9u3beYrrXcLDw9G0aVM4VayE8IBpqG8qw+SGHFSTSo4LFy4gICAAgwcPRkqt7pDfisIPjfmjhkqv8+fPo2/fvrh9+zaaNGmCxYsXo2PHjtDQ0MDUE4/wpY8lynJGTypEK668Rg9Pc0gUGXjx4gUTS1SiLVi2CvMfmqOm/mP0bt5Z3eEQ5cuBB7FIThdo62aCH9czsUTFR5FI4RsaGubYKikhIQFA3mdtu3jxIqKjo3MctDs3DRs2BADcvXs3T+UXL14MHx8f+Pj4ZEtU5VdMTAyaN28OUzNzGHafCYWQYE07F2iw6wOVEDdv3sRnn32Grl27okavsZgb/Ap/fu4EA+2cB+MnKg1mzJgBBwcHPHnyBIcOHUKXLl2gqamJCy8ScfVVEkbWsVZ3iFSC3I1IxunQBFRKe4Ry5crB2NgYlStXVndYRIUuawKESUdDYVTGEEcnd8518h+iomrD9Ui0djOGoY4mIiOZWKLio0gklhwdHfHs2TPI5XKV5U+ePFG+nheHDx+GVCpFo0aN8h1DXltFDRw4EJcuXcKlS5ewbt26fO8nS0REBBo3bgy5QqD8iOU4+zwJ2zuXh5FukWlERvRBUlJS0Lx5czRp1hxG7cej+/aHmNnMAS3KG6s7NCK1SU5OxsGDBzF48OBss5quuxaBhk6GcGBrJSpEK668Ri07fWz5YzqqVKmCBw8ewNzcXN1hERW6nj17otuYaYhz/wxL25aHnhaTSlS8xKXKsftuNLp5ZJ6jmVii4qRIJJZ8fX2RnJyMs2fPqiw/duwYTE1NUaFChfduIzo6GpcuXULt2rVRpkyZPO/72LFjAICKFSvmqbyNjQ28vb3h7e2dr5nq3vTq1Ss0atQIGppSeE/ahH2PknCohxvczLMPNk5UXJ09exZh8SmIaDYBq69F4kAPN3xbmy0xqHTbtGkTNDU14e/vr7I8Xa7A5ptR6ObJH/xUeNLkCqy5FoGu7oY4ceIERo4cCWNjY3WHRfRRBP9zERZ95qCnlyW+8GJ3Typ+dtyOgoG2Jpq5lsHz588RHBzMxBIVG0WieYyPjw+8vLywcOFCJCUlwcbGBidPnsTly5cxYsQIZTPW+fPn49ixY1iyZAksLS1VtnHs2DHI5fJcu8HdvHkTW7Zsga+vL6ysrJCeno5Lly7h4MGD8PT0RM2aNT/6cWb55ptvoKevD8/v1uCvu/E41LMiqtnof7L9E30Km46eh9bX6xCXLvDPgMpwNGYrDCrdDh06hOnTp2P16tXQ1dVVLo9NycBv514hIU2O9pU4cQMVnj33YpCYpkDalX2QyWSoV6+eukMi+ihiYmLw0rkRzKCDWc3KqjscogLZcD0SHSqZYuvmTejZsycqVqyItm3bqjssojwpEoklABg3bhzWrVuHDRs2ID4+Hvb29hg1ahTq16+vLKNQKKBQKCCEyLb+kSNHYGlpCS8vrxy3b2JiAk1NTWzevBlxcXGQSCSwsbFB165d0a5du086I9yt27dh0/tXbLsbj4M9KqK6LWfHopLl2OM4rJL4wk0Wi5N9KkGmVSQaRxKpTWRkJIYOHZrZVaNbNwDAnYhk/HbuFTZej4SRribmtXBEGR123aDCc+X+UySe3oSR+2djxIgR0NbWVndIRB/FjVu3IanVCZMb2MJMT0vd4RDl26uENBx7HIcpDe0w69tt6NmzJ5YvX57n4VqI1K3IJJZkMhkGDBiAAQMG5Fpm2LBhGDZsWI6vLVy48J3bt7W1xeTJkz8kxEIhhMBj+4Z4mmqOw70roCanXKcSJDU1FVP+CsbsezpQ/LMNC6Z0ZVKJSq2oqCgEBQVhy5YtOHr0KMqWLYvBgwdDCIFFF8Mx8mAofMsaYGVbF7SuaAwtTX5WqHBN+swd/mUC4LVhIrvAUYm282ooJHpl0b0au9xT8bTh2mvY6mvg9eXj2LFjB5NKVOwUmcRSaREZHYO0am0xwUsbvmUN1R0OUaGZNWs2Jhx8gIw6PVDh4S4MbuOOunV81R0WkVps2rQJvXr1gomJCQIDAzF27Djol6uGFedD8HfwDTyJScUfnzuhj5c5Lxzpo9HW1lbOfktUkh18LmCZ+BCmsjrqDoWoQPTTYhC67nu0v34AgYGBaNmypbpDIsoXJpY+sV1XQwAdA/So4fD+wkTFyMzL8ZDU7Y5t7Z3RzuNndYdDpDbLli3DV199hdmzZ2Pw4MH488Jr9Al+hdC/78DLXIoeVa3RqYoZxx0jIioEqRkK3NWwxWd619QdClGB9atXEW32L4OxsbHKOIxExQUTS5/Yzrsx0Ay9CUcrtuSgkkOuUCCibB1865yMdh426g6HSG3mzp2LUaNGYfny5ejVqxfmBb/Cd4ef4pemZdGhsilSXj+Di4utusMkIiox9j+IhVwALcvlfVZooqJGKpXC2ppdOan4YmLpExJC4HSEFBaRt9j1gUqUk3dfAnrGaOuhp+5QiNRCLpdj9OjR+PPPP7Fx40Z06NABq6++xuhDodgQWA6B/5/t7dFrNQdKRFTCrLsSBsXt46jWuoW6QyEiKrWYWPqELr5IRIxcC/XEK3WHQlSodlx7DvHqPqq7tVN3KESfXExMDLp164bg4GAcOXIE9erVw4brERiw6zEWBDgpk0pERFS44lLl2PcgDuL6Ibi5DVd3OEREpRYTS59Q0J1omKe+QjlrE3WHQlSoTjxNht6rG9DT66buUIg+qbCwMPj5+UEmk+Gff/6Bq6srllwKx+C9TzCvhSP6e1uqO0QiohJr551oaEEO05TnKFOGXeGIiNSFiaVPKOhONAyfX0ZZ17LqDoWo0CSnK3A3SRvl5GHqDoXok/vtt9+gra2Nc+fOQV9fH7+de4nvDj/FsjYu6FnVXN3hERGVaBtvRKJceijM3SuqOxQiolJNQ90BlBa3XyfjTkQK0v89jLJlmViikuNUaDwUAvAwUqg7FKJPKj4+HosWLcKYMWOgr6+PTTci8d3hp9gQWI5JJSKijyw8MR1HHsZC7+FpuLu7qzscIqJSjYmlTyToTjSqWMoQducyHBwc1B0OUaE58jAWZWKfoIIz6zWVLkuXLoWBgQG6dOmCKy8T0X/nY/za1IFjKhERfQLbbkbB1lAbYRcPo1KlSuoOh4ioVGNi6RMJuhON6voJSE1NRbly5dQdDlGhOfIoDnh4Hi4uLuoOheiTSUtLw9y5czFs2DDEpEvwxeb7CKxsgm9rW6k7NCKiUmHTjUgEuhvh4YP7TCwREalZkRljKTk5GevWrcOZM2cQHx8Pe3t7tG/fHvXr13/nekePHsW8efNyfG316tUwMVEdKPvq1atYv349Hj9+DB0dHdSoUQO9e/eGsbFxYR1KNk9jUzNnhDs2DX369IGTk9NH2xfRpxSWkI5rYUnA5UNwHskZ4aj0WLduHeLj49Gn3wAEbn0AK30tLApwhkQiUXdoREQl3pOYVJx5moAh5dOhUCjYFY6ISM2KTGJp+vTpuH//Pnr16gVbW1ucPHkSs2bNghACDRo0eO/63377Lezs7FSWGRoaqjy/ceMGpk6diurVq2PChAmIjY3F6tWrMWnSJMyZMwdaWlqFekxZdt6JhplmKsKvn8Uv2+5+lH0QqcPRx7Ew0ZEg8sUdtliiUmXlypX48ssvsf1RKq6HJeHa1x7QlbIRMBHRp7D5RiTczXWheHEHlpaWMDMzU3dIRESlWpFILF28eBFXr17FyJEjlUkkT09PhIeHY+XKlfDz84OmpuY7t+Hg4IDy5cu/s8zKlStha2uLsWPHKrdnZWWF7777DocPH0bLli0L54DeEnQnGkYvr6J5166wsLD4KPsgUofDD+PgaZiG01IpbG1t1R0O0Sdx584dnD9/HlOm/oBBZ15iaC1r2JXRVndYRESlxsbrkejiYYbLey+xGxwRURFQJG6vBgcHQyaTwc/PT2W5v78/oqKicO/evQ/eR2RkJO7fv49GjRqpJKnc3d1hZ2eH4ODgD95HTiKS0nEyJB5RZ3egVq1aH2UfROoghMCRR7FwyHgFJyen9yZ/iUqC9PR0BAYGIiAgAOHmHghLTMfgWhxXiYjoU7kRnoTr4ckIcJZh2bJl6Nq1q7pDIiIq9YpEi6WQkBDY29tn+2Hq7OysfP19fad//PFHxMXFQU9PD1WqVEG3bt3g6Oiosg8AOY5v5OjoiDt37nzgUeRs990YGOto4PW1v1Gr1qKPsg8idbgdkYIX8ekwSL3FbnBUauzcuRNPnjzB3ydPotGm5/i6uhVMZUXiq5SIqFRYcf4p7CVxGNmnI3R1ddGzZ091h0REVOoViRZL8fHxMDAwyLY8a1l8fHyu6xobG6Njx44YMmQIfvrpJ3Tr1g0PHjzA6NGj8fjxY5V9vLnNNxkaGiIuLu5DDyNHQXeiUVU3FibGRu/tqkdUnBx5GIsKZrqIeHRTmQQmKukWLFiAHj164HS4BA+iUjDMl62ViIg+pbauutA7vQSurq7YunUrdHR01B0SEVGpV2Rus75rJp13vebj4wMfHx/l8ypVqqB69eoYOnQo1q9fj4kTJ+ZpW3mdyefly5d4+fIlAOD27dvvLJuQJsfhh7GofHMnfH19OVsQlSiHH8WiiUsZ/LP2scpnkKikunXrFo4fP465c+fiy1Mv0M/bAtYGHFuJiOhTql/JEbcPbVF3GERE9IYikVgyNDTMsVVSQkICgJxbGb2LlZUV3N3dcffufzOwZc0Ql9N+4uPjs80gl5vFixdj6tSpeSq7/140FOmpeHJ8G9afPJGndYiKgzS5An8/iUf/Lyyx6dEjtliiUmHhwoWoV68ewvQdceXVPWzuUE7dIRERERERqV2R6Arn6OiIZ8+eQS6Xqyx/8uSJ8vX8EkJAQ+O/w8vaRtZYS28KCQmBg4NDnrY7cOBAXLp0CZcuXcK6deveWXbV0csQ987g0P69qFChQj6iJyragp8lICVDgermEkRERHCMJSrx4uPjsXr1anz99SBMP/UC3TzN4GjM7hdEREREREUiseTr64vk5GScPXtWZfmxY8dgamqa76TMq1evcOfOHVSsWFG5zMzMDBUqVMCJEydUElh37tzB8+fP4evrm6dt29jYwNvbG97e3u8dUHzXYH9cmNwO1apVy1f8REXd4YdxqG1vgMgXoQDAxBKVaOnp6ejWrRssLCzwyLYu/nmeiLF+NuoOi4iIiIioSCgSXeF8fHzg5eWFhQsXIikpCTY2Njh58iQuX76MESNGKGeLmz9/Po4dO4YlS5bA0tISADBp0iRUrlwZTk5OkMlkCAkJwfbt2wEA3bp1U9lPr169MHnyZPzyyy9o2bIlYmNjsXr1ajg6OqJJkyaFflyamprwdGdLJSp5jjyKRYtyxnj06AKMjY1hbGys7pCIPgqFQoG+ffvin3/+wZi1RzH275fY2rE8KpjJ1B0aEREREVGRUCQSSwAwbtw4rFu3Dhs2bEB8fDzs7e0xatQo1K9fX1lGoVBAoVBACKFc5ujoiFOnTmHHjh1IS0uDkZERPD090alTJ9jZ2answ8PDA5MnT8aGDRvw008/QUdHB9WrV0efPn2gpaX1yY6VqDiLTs7AxReJmN3cASfX32FrJSrRpkyZgj179mDmtmMYfDYJcz9zRBs3E3WHRURERERUZBSZxJJMJsOAAQMwYMCAXMsMGzYMw4YNU1nWv3//fO2nWrVq7JpG9AGOPY6DgZYGNs+Zgj/mz8P333+v7pCIPgq5XI5FixZh5C8LMPqCAsN9rTGoppW6wyIiIiIiKlKKTGKJiIqHzRceIe1+MPb8vRP79+9Hs2bN1B0S0UcRHByMqHRNLI6vgJbljTDN317dIRERERERFTlMLBFRvkxp7AjrJ39jxvXr0NPTU3c4RB/Nzt27YdD3d1Qw18OKNi7QkEjUHRIRERERUZFTJGaFI6Lio3JZC8yfMpJJJSrx1t5NhdzIFuvbu0JHyq9LIiIiIqKcsMUSERHRW45fu4+wyu3wc01DWBtoqzscIiIiIqIii7dgiYiI3iCEwFd7nkDv9T2MaV5F3eEQERERERVpTCwRERG9YfmV13icqovOBiHQ0ODXJBERERHRu/CKmYiI6P+ex6Vh1MFQKA4vQLeAxuoOh4iIiIioyGNiiYiICJld4L7Z9wTWGknQv3MQfn5+6g6JiIiIiKjI4+DdREREALbeisL+e9HQXvkN+vfuBS0tLXWHRERERERU5BWZxFJycjLWrVuHM2fOID4+Hvb29mjfvj3q16//zvXOnj2LM2fO4P79+4iKioKxsTHc3d3RpUsX2NraqpQdP348bty4kW0b1apVw9SpUwv1eIiIqPiISExH/223kXF8Jcb3DsSkSZPUHRIRERERUbFQZBJL06dPx/3799GrVy/Y2tri5MmTmDVrFoQQaNCgQa7rbd++HcbGxujYsSOsra3x+vVrbNu2DcOHD8fMmTPh4OCgUt7a2hojRoxQWWZgYPBRjomIiIqH77ecRFJ4ONZ/0wKdOwSqOxwiIiIiomKjSCSWLl68iKtXr2LkyJHKJJKnpyfCw8OxcuVK+Pn5QVNTM8d1J06cCGNjY5Vlnp6eGDBgAHbu3IkhQ4aovKatrQ03N7ePchxERFQ8/d67MQbWfQSPCq7qDoWIiIiIqFgpEoN3BwcHQyaTZRso1d/fH1FRUbh3716u676dVAIAMzMzmJmZISIiorBDJSKiEkgikTCpRERERERUAEWixVJISAjs7e2ztUpydnZWvu7u7p7n7b169QqvX79G7dq1c3yta9euSEpKgqWlJerVq4eOHTtCR0fnww6CiIiIiIiIiKiUKRKJpfj4eFhZWWVbnjX2UXx8fJ63JZfL8fvvv0NXVxetW7dWea1SpUrw8/ODvb090tLScOnSJWzfvh23bt3CtGnToKFRJBpwEREREREREREVC0UisQRkdkMoyGtvEkJg/vz5uHnzJsaOHQsLCwuV17t3767yvHr16rC0tMTKlStx/vx5+Pr6vncfL1++xMuXLwEAt2/fzlNcREREREREREQlUZFoomNoaJhjq6SEhAQAeZu1TQiB33//HX///Te+/fbbHLvB5aRhw4YAgLt37+ap/OLFi+Hj4wMfH59siSoiIiIiIiIiotKkSCSWHB0d8ezZM8jlcpXlT548Ub7+LllJpaNHj2Lw4MFo1KhRvmPIa6uogQMH4tKlS7h06RLWrVuX7/0QEREREREREZUURSKx5Ovri+TkZJw9e1Zl+bFjx2BqaooKFSrkuq4QAn/88QeOHj2KQYMGoUmTJvna97FjxwAAFStWzFN5GxsbeHt7w9vbO18DihMRERERERERlTRFYowlHx8feHl5YeHChUhKSoKNjQ1OnjyJy5cvY8SIEcrZ4ubPn49jx45hyZIlsLS0BAAsWbIEhw8fRpMmTeDo6Ig7d+4ot6ulpQVX18zpo2/evIktW7bA19cXVlZWSE9Px6VLl3Dw4EF4enqiZs2an/7AiYiIiIiIiIiKsSKRWAKAcePGYd26ddiwYQPi4+Nhb2+PUaNGoX79+soyCoUCCoUCQgjlsgsXLgAAjhw5giNHjqhs09LSEsuWLQMAmJiYQFNTE5s3b0ZcXBwkEglsbGzQtWtXtGvXjjPCERERERERERHlU5FJLMlkMgwYMAADBgzItcywYcMwbNgwlWVZiaP3sbW1xeTJkz8kRCIiIiIiIiIiegOb6RARERERERERUYEwsURERERERERERAXCxBIRERERERERERUIE0tERERERERERFQgTCwREREREREREVGBMLFEREREREREREQFwsQSEREREREREREVCBNLRERERERERERUIEwsERERERERERFRgTCxREREREREREREBSJVdwCfWnJyMtatW4czZ84gPj4e9vb2aN++PerXr6/u0IiIiIiIiIiIipVSl1iaPn067t+/j169esHW1hYnT57ErFmzIIRAgwYN1B0eEREREREREVGxUaoSSxcvXsTVq1cxcuRIZRLJ09MT4eHhWLlyJfz8/KCpqanmKImIiIiIiIiIiodSNcZScHAwZDIZ/Pz8VJb7+/sjKioK9+7dU1NkRERERERERETFT6lKLIWEhMDe3j5bqyRnZ2fl60RERERERERElDelKrEUHx8PAwODbMuzlsXHx3/qkIiIiIiIiIiIiq1SNcYSAEgkkgK9luXly5d4+fIlAOD27duFFhcRERERERERUXFTqhJLhoaGObZKSkhIAIAcWzO9bfHixZg6dWqhx0ZEREREREREVNzkqyvcmDFj8OzZM5VlCoWiUAP6mBwdHfHs2TPI5XKV5U+ePFG+/j4DBw7EpUuXcOnSJaxbt+5jhElEREREREREVCzkK7E0e/ZsvHjxQvlcLpdDS0sLly9fLvTAPgZfX18kJyfj7NmzKsuPHTsGU1NTVKhQ4b3bsLGxgbe3N7y9veHu7v6xQiUiIiIiIiIiKvLy1RVOCJGnZUWVj48PvLy8sHDhQiQlJcHGxgYnT57E5cuXMWLEiGyzxb1PcnIyAPWPtfTs2TPExMSoNYYPUdzjB4rvMbi5uUFPT09lGet14Sju8QPF9xhYrz+e4h4/UHyPgfX64ynu8QPF9xhYrz+e4h4/UHyPgfX64ynu8QPF9xhyqtdvKlVjLAHAuHHjsG7dOmzYsAHx8fGwt7fHqFGjUL9+/XxvK6sLXffu3Qs5SqJP49KlS/D29lZZxnpNxR3rNZVErNdUErFeU0nEek0lUU71+k0SkY8mRxoaGggODkbNmjUB/NcV7uLFi+/cSUkVERGBgwcPwsnJCTKZTN3hEOVbTpln1msq7livqSRivaaSiPWaSiLWayqJ3tdiKd+JpdWrV6Ny5coAMhNLtWrVwrp16+Dm5patfGlMNhERERERERERlRb5TixJJBKVZUKIXJe9PfsaERERERERERGVHPkaY2nlypUfKw4iIiIiIiIiIipm8tViiYiIiIiIiIiIKEupmxWusIWGhiIiIkLdYRDlm7m5ORwcHHJ8jfWaiivWayqJWK+pJGK9ppKI9ZpKonfVayVRADdu3BD9+vUT5cuXF/r6+kJfX1+UL19efPnll+LWrVsF2WSxFBISIvT09AQAPvgodg89PT0REhLCes1HiXqwXvNREh+s13yUxAfrNR8l8cF6zUdJfORWr9+U7xZLGzZsQJ8+fZCeng4bGxt4eHhACIHQ0FAsXboUq1evxtq1a9GhQ4f8brrYiYiIQFJSEtatWwd3d3d1h0OUZ7dv30b37t0RERGRLfvMek3FFes1lUSs11QSsV5TScR6TSXRu+r1m/KVWAoNDcWAAQNQpUoVLF68GNWrV1d5/Z9//sFXX32Fvn37onbt2ihbtmzBoi9m3N3d4e3tre4wiAoV6zWVRKzXVBKxXlNJxHpNJRHrNZVUGvkpvGTJEhgbG+Po0aPZkkoAULNmTRw5cgSGhoZYunRpoQVJRERERERERERFT74SSydOnEDfvn1hbGycaxlTU1P069cPx44d+9DYiIiIiIiIiIioCMtXYunevXuoVavWe8vVqlUL9+7dK3BQRERERERERERU9OUrsRQbGwszM7P3ljMzM0NsbGyBg6JP7/Lly+jXrx9q1KgBZ2dnVK1aFa1atcLUqVPVHVqebd68GXZ2drh27VqhbfPo0aOYPXt2oW2PirasOuTi4oJnz55lez0wMBCNGzfO93Z37NiRa/fgp0+fokePHqhcuTLs7OwwefJkPH36FHZ2dti8eXO22J4+fZrv/VPJklUXcnr88MMPaompVq1a6Nmzp1r2bWdnhwkTJqhl3/Rp5Vbv336cPXs2X9s9e/ZstvXe9f0/f/58HDhw4IOOhUqvnj17wt3dHc+fP8/2WnR0NKpVq4Y2bdpAoVB88L5mz56dp89MYGBgnrZX0OsgKv4+xu+sLPfu3cPs2bN5jVvM5Wvw7vT0dGhqar63nIaGBjIyMgocFH1aR44cQZ8+feDr64uJEyfC0tIS4eHhuHbtGnbu3IkpU6aoO0S1OXbsGFatWoWRI0eqOxT6hFJTU/HLL7/g999/L5Tt7dixA3fv3sWAAQOyvfb999/jypUrmD17NiwsLGBlZQUhRLZyTZo0wa5du2BpaVkoMVHxN2fOHJQrV05lmbW1tZqiIfr4du3apfJ87ty5OHv2LLZs2aKyvEKFCvnaroeHB3bt2qWy3ru+/3///Xd8/vnn+Oyzz/K1HyIAmDVrFho3bozRo0djw4YNKq9NnDgRCQkJmDt3LjQ08nX/P0ddunRBw4YNlc/Dw8PRv39/9O3bF23btlUuNzQ0/OB9ERXUvXv3MGfOHPj6+paayb9KonwlloDMcZZyupP/JnaDK14WLlwIBwcHbNiwAVLpf1WiTZs2mDhxohojI1KPRo0aISgoCF999RUqV678Ufd1584dVKtWTeUHSk53bMzMzPLUYpRKDzc3N1StWlXdYRB9Mj4+PirPzczMoKGhkW15fhkaGn7wNj6UXC5HRkYGdHR01BoHfXyWlpb4+eef8fXXX2Pt2rXo0aMHAGD//v0ICgrCzz//DGdn50LZl62tLWxtbZXPs64v7Ozs1F7niYoKIQRSUlIgk8nUHUqxlu9U+NixYxEYGPjOx/jx4z9GrPSRREdHw8TERCWplOXNuyVZXR2OHz+O5s2bw9XVFfXr18emTZtU1omMjMS4cePQsGFDlC9fHp6enujQoQPOnz+vUi6ru8+CBQswb9481KhRAy4uLmjRogVOnTqVLZYHDx5g0KBBqFq1KpydnVGjRg0MHToUqampKuUSEhIwduxYVKlSBZUrV0b//v3x6tUrlTI7d+5Ely5dUK1aNbi6uqJBgwb4+eefkZSUpCwzbNgwrFq1CoBq83s20yz5vv76a5iYmODnn39+ZzkhBFatWoWmTZvC1dUVlSpVwoABAxASEqIsExgYiKNHj+LZs2fZumrY2dnhyZMnOHbs2Hvr19td4R49eoSKFSviyy+/VCl3+vRplC1bFr/++usHvgtUHD1+/BjDhw9H3bp14erqCh8fH/Tq1Qu3b9/OVjY2NhZTp06Fr68vnJ2d4enpiR49euDBgwfKMmlpaZg7dy7q168PZ2dneHh4YPjw4YiMjMxx//v370eTJk3g4uICX19fLF++PFuZ58+fY8iQIfD09ISzszMaNGiARYsWZev2ER0djXHjxsHHxwdOTk7w9fXFjBkzsp3z3yaEwPTp0+Ho6Ij169fn5W2jEuLLL79Eo0aNVJb16tULdnZ22L17t3LZ9evXYWdnh0OHDgHI3hXuXd//dnZ2SEpKwtatW3PsRhQeHo4xY8Yo623t2rUxZ84clZb8b17/zJ07F7Vr14azs3O+u/BR8dW6dWu0adMGP/74I54+fYqoqCiMHTsW9evXR69evXDo0CG0atUKrq6uqFChAjp37oyLFy+qbGPYsGE5jn2b1f0tr3Lrap9TF9Es58+fR0BAgPJ75tdff4VcLlcpk9/vDypehg0bhvLly+Px48fo0aMHypcvj+rVq2Pq1KnZvqdXr16NJk2aoHz58qhQoQLq16+P6dOnA8isfwMHDgQAdOjQQXlezRoO4uTJk+jTpw98fHzg4uKCunXrYsyYMYiKisoW08GDB9GkSRM4OzvD19cXy5Yty/HzkNWFfs2aNWjQoAGcnZ2xdetWAJmtwQMCAlC5cmVUrFgRzZs3x8aNG7P1Jsj6XXz48GE0a9ZM+Xvy8OHDyuNq0KABypUrh88///yjdCEsavLVYun48eMfKw5SIx8fH2zYsAGTJk1Cu3bt4OHhAS0trRzL3rp1Cz/88AO++eYbWFhYYMOGDRg5cqTy4gnI/DEAAMOHD4elpSUSExNx4MABBAYGYvPmzahTp47KNleuXAl7e3tMnToVCoUCCxcuRI8ePbBt2zZUr14dAHDz5k20a9cOpqamGDVqFJydnREeHo5Dhw4hPT1d5Q7f6NGj4e/vjz///BMvXrzATz/9hCFDhihPGEDmj6/GjRujf//+0NPTw4MHD7BgwQJcuXJFWW7YsGFISkrC3r17VZrfsytSyWdgYIBvv/0WkydPxunTp+Hn55djue+++w5btmxB3759MWHCBMTExOC3335DmzZtcPjwYVhYWODnn3/GmDFjEBISgmXLlinXrVChAnbt2oX+/fvD0dERkyZNAgBlV9T3cXFxwa+//opBgwZh+fLl6NevH8LDwzF48GDUqlWL3TdLgawWDm8KCwuDiYkJxo8fDzMzM8TExGDr1q0ICAjAwYMHlV3nEhIS0K5dOzx9+hSDBg2Ct7c3EhMTERwcjLCwMJQrVw4KhQJ9+/bF+fPnMWjQIPj4+OD58+eYNWsWAgMDsW/fPpW7ezdv3sSUKVMwcuRIWFhYYMeOHZg8eTLS09Px1VdfAci88dC6dWukp6dj9OjRKFu2LI4cOYIff/wRISEhygvNlJQUdOjQASEhIRg5ciQqVaqE8+fP448//sDNmzexdu3aHN+T1NRUDB8+HEePHsWqVauyJRmoZKtXrx727t2LsLAwWFlZISMjA8HBwdDV1cWpU6fQqlUrAMCpU6cglUqzXY9kedf3/65du9CxY0fUqVMHw4YNA/BfN6Lw8HB8/vnn0NDQwPDhw+Ho6IhLly5h/vz5ePr0KX777TeV/SxfvhwuLi6YNGkSDA0NC62VChUP06ZNQ3BwMEaMGAEzMzOkp6dj9uzZ2LFjBwYPHowGDRpgwYIFSE1NxcKFC9GhQwds3rwZNWvWVGvc4eHhGDRoEL755hu4urri6NGjmDdvHmJjYzFt2jQAyPf3BxVPGRkZ6NOnDzp37owvv/wS58+fx9y5c1GmTBkMHz4cQObN/PHjx6Nv376YPHkyJBIJnjx5gvv37wPIHOph7NixmDFjBqZNmwYPDw8AgJOTEwDgyZMn8PHxQZcuXVCmTBk8ffoUS5YsQdu2bXH06FHlb9bjx4+jf//+qFWrFhYuXIiMjAwsWrQIEREROcZ+4MABnD9/HsOHD4eFhQXMzc0BZCb+u3fvrmzpd/nyZUycOBGvXr1SHlOWW7duYcaMGRg6dCgMDQ0xZ84cDBgwAIMHD8bp06cxduxYSCQSTJs2Db169cK5c+dKdr0XVGCXLl0SAMSlS5fUHcoHiYyMFG3bthW2trbC1tZWODo6itatW4vff/9dJCQkKMvVrFlTuLi4iGfPnimXJScni0qVKokxY8bkuv2MjAyRnp4uOnbsKPr166dcHhoaKmxtbYW3t7dITk5WLo+PjxeVKlUSnTp1Ui7r0KGDcHd3FxEREbnuZ9OmTcLW1laMGzdOZfmCBQuEra2tCAsLy3E9hUIh0tPTxblz54Stra24efOm8rXx48cLW1vbXPdZXL2r7paUel0QWXXo6tWrIjU1Vfj6+ooWLVoIhUIhhBCiffv2olGjRkIIIS5evChsbW3FokWLVLbx/Plz4eLiIn766Sflsh49eoiaNWvmuM+aNWuKHj16qCzL+mxs2rQpW2yhoaEqZceOHSucnJzExYsXRYcOHYSnp6d49epVwd+EYqy01OusupDTIz09XaVsRkaGSEtLE3Xr1hVTpkxRLp8zZ46wtbUVf//9d677CQoKEra2tmLv3r0qy69evSpsbW3FqlWrlMtq1qwp7OzsxI0bN1TKdu7cWVSsWFEkJSUJIYT4+eefha2trbh8+bJKubFjxwo7Ozvx4MEDIYQQa9asEba2tmLXrl0q5f78889scdva2orx48eLqKgo0bZtW+Ht7Z0tjuKstNTrgvj2229FuXLllM8fP34sbG1txdatW4UQQvzzzz/C1tZW/PTTT6J27drKcp07dxatW7dWPj9z5oywtbUVZ86cUS571/d/uXLlxLfffptt+ZgxY0T58uVVrpOEEGLhwoXC1tZW3L17Vwjx3zm+Tp06Ii0tLf8HXgKwXmc6evSo8vy9bds2IZfLhbe3t/D39xdyuVxZLiEhQXh6eqrU22+//TbHa4tZs2blWnez6t7ChQuVy3K7vsjpc9G+fXtha2srDh48qFJ29OjRwt7eXln38/P9UZKU5Hr95jWyEJn1L6fv6e7du4t69eopn0+YMEG4u7u/c9u7d+/OVtdykvWb7dmzZ9nqYcuWLUX16tVFamqqcllCQoKoXLlyts+Dra2tcHNzE9HR0e/cn1wuF+np6WLOnDmicuXKyt8DQvz3u/jFixfKZTdu3BC2traiWrVqyuseIYQ4cOBAjp+b4iKvdTffXeGuX7/+zjGWnj17huvXr+c7wUXqY2pqih07dmDfvn0YP348mjVrhkePHmH69Onw9/dXaWpYqVIlleaEurq6cHFxyTazxZo1a9C8eXO4uLjAwcEBjo6OOH36tDI7/aYWLVpAV1dX+dzAwABNmzbF+fPnIZfLkZycjODgYAQEBORpjJlmzZqpPHd3dwcAlXobEhKCb775Bl5eXihbtiwcHR3Rvn17AMgxRip9tLW1MWbMGFy7di3bgLFA5qD3EokE7du3R0ZGhvJhaWmJSpUq4dy5c58kzu+//x4VKlRAx44dce7cOfz++++wsrL6JPsm9Zo3bx727dun8gAyZ6xq2LAhnJyc4ODgACcnJzx+/Fili9vx48fh4uKC+vXr57r9I0eOwMjICE2bNlWp45UrV4alpWW2Ol6xYsVsY5K1bdsW8fHxyuuCM2fOoEKFCqhWrZpKuY4dO0IIgTNnzijL6enpISAgIFs5ILPL55tCQ0PRunVrxMfHY/fu3R99bDQqmpycnFC2bFlld/qTJ0/C3d0dX3zxBUJDQ/HkyROkpqbin3/+Qb169Qp9/0eOHEGdOnWUraWyHlmzaL39mWnatGmuLcSpdGjcuDG8vb3h7OyM9u3b4+HDh3j16hXat2+vMhyFvr4+WrZsicuXLyM5OVmNEWdep799rd22bVsoFAoEBwcDyP/3BxVPEokETZs2VVlWqVIlld+FXl5eiI2NxaBBg3Dw4MEcu7C9S0REBL777jtUr15d+Zsyq9Ve1m+2pKQkXLt2DZ999hm0tbWV6+rr62eLL0vdunVhbGycbfnp06fRqVMnuLm5KX8jzpo1C9HR0dlaP1WuXBk2NjbK5+XLlwcA+Pr6qrRMymotntNMkCVJvrrCnTx5Ek2aNMG5c+dgb2+fY5mwsDD4+vpi9+7daN68eaEESZ9G1apVlQPBpqenY9q0aVi6dCkWLFigHMTbxMQk23o6OjpISUlRPl+8eDF++OEH9OjRA6NHj4apqSk0NTUxc+bMHJM2FhYW2ZZZWloiLS0NiYmJSExMhFwuV/ngvsvbMWadYLJiTExMxBdffAEdHR2MGTMGLi4ukMlkePHiBfr3769yLFS6tWnTBosWLcKvv/6Kli1bqrwWEREBIUSugyc7Ojp+ihCho6ODdu3a4ccff4SHh8c7EwVUspQvXz5b/Zs0aRJWrVqFQYMGwdfXF0ZGRtDQ0MCoUaNUzm2RkZHvHYPj9evXiI2NVTZHf9vbF4e5ncuB/7pIR0dH53j9kJUMfbOcpaUlJBKJSjlzc3NIpVJluSxXr15FVFQUvvvuO5WBaqn08fPzUw7dcOrUKdSrVw/u7u6wsLDAqVOn4OzsjJSUlI+SWIqIiMDhw4dzPf+//ZnhTQACMr/Hs65Vs85tOQ27YGVlBYVCgZiYGLV2p8nqMvSmt8/1+f3+oOJJJpOpNA4AMn93vXm9ERgYCLlcjvXr12PAgAFQKBTw8vLCmDFj3nvNqlAo0KVLF4SFhWHYsGFwc3ODnp4eFAoFWrVqpdxPTEwMhBA51s2crk2AnD9jV65cQdeuXeHr64tff/0VNjY20NbWxoEDBzB//vxsvxHfTkxlfY5zW/6+MSKLu3wllv78808EBga+cxYBHx8fdOrUCcuWLWNiqRjT0tLCiBEjsHTpUty9ezdf627fvl05yOqbEhISciz/+vXrbMvCw8Ohra0NfX19aP2vvTuPi6L+/wD+mr2AheVULhHwFm/xSBTvo0PLzNJSUzOP7DS1zCzLrEzT7m9qVuov1PI+yspbTCUF1DwQT/BCEbkWdtldduf3x8LKCioguIev5+OxD9jPzM68Z3nvMPPez3xGLodUKkVaWlqF4ridvXv34urVq1i9ejWioqIs7Tk5OVWyfHIegiDg3XffxXPPPVdqEGBfX18IgoB169ZZfTtSrKy26nDy5El8/vnnaNWqFQ4fPoyFCxdaBkGkB8/atWvx9NNPY+rUqVbtmZmZ8PLysjz38/O76z7V19cXPj4+tx0A293d3er57fblwM2Cv4+PT5ljiF27ds2yzuL5Dh06BFEUrYpLGRkZKCwstMxX7IknnkDNmjUxe/ZsmEwmy9g39OCJjo7GihUrcOjQIRw6dAhvvPEGAPO307Gxsbh06RLc3d0RGRlZ5ev29fVFREQEpkyZUub0WwtJtxZOiYr3lbfbT0okEstJq4uLC/R6fan5Klq0KS4M3HrSe7vllDVmza37+or+/yDnNnjwYAwePBgajQZxcXGYN28eRowYgT179ty2swpgPsY9ceIEvvzyS0uPZcA8Vm5J3t7eEAThjrl5q7L2vxs2bIBcLsfSpUutCmZ//fXXXbeRKnhXuL179+LJJ5+863xPPPGEpSsk2b/iA/pbFfcuqug3aoIglLpd7okTJ5CQkFDm/H/++adVBTgvLw9bt25F+/btIZVK4ebmhg4dOuCPP/6o0m84bj3xj4mJue08tu52TLbTpUsXdOnSBV9++SXy8/Mt7b169YIoikhLS7P09iv5KL4EEyjdq6+qaDQajBs3DrVr18aqVavwwgsvYNasWUhMTKzydZHjuHXftm3btlJ3xuzevTvOnTtX6pKyknr16oWsrCwYjcYyc7y4a3ex5ORkHD9+3Kpt/fr18PDwsAzG2alTJ5w6darUJfOrV6+GIAiWwZSjo6MtN364db7i5dxqwoQJmDFjBubOnWsZBJwePNHR0RAEAZ9//jkkEonlxiKdO3fGvn37sGfPHjz00EN3vQTtTv//b/1GvlivXr2QnJyMsLCwMj8zgYGBVbCF5Mzq1auHwMBArFu3zuouVBqNBps3b0abNm0svZVq166NjIwMq6K+Xq/Hrl27KrTO4hP7W+8eWnzXxFvl5eWVmrZ+/Xqrz1tF/3/Qg0GpVKJHjx54/fXXodfrLZ0Xbr26pFhx8efW88pbz9mUSiVatmyJv/76y6rYmp+fj23btpU7PkEQIJVKIZVKLW1arRZr1qwp9zIeZBXqsXT9+vVy3b4yKCioXHc1IvswdOhQBAUFoVevXpY7AR0/fhw//PAD3N3d8eKLL1Zoeb169cJXX32FuXPnokOHDjh79iy++uorhIaGlrqDEQBIJBI899xzGDt2LEwmE77//nvk5eVh8uTJlnk++OADDBgwAP369cMrr7yC8PBwZGRkYMuWLZg9ezY8PDzKHV/btm3h7e2Nd955BxMnToRMJsO6detw4sSJUvM2btwYAPD999+je/fukEqliIiIuG+9Ucg+TJs2DY888ggyMjLQqFEjAEC7du0wdOhQTJw4Ef/99x8eeughKJVKpKen48CBA2jcuDFGjBgBwJxHmzdvxtKlS9GiRQtIJJLbXkJXEVOmTMHly5fxxx9/QKlUYvr06UhISLBcx16yhwo9GHr16oVVq1ahfv36iIiIwNGjRzF//vxSlxKPGTMGmzZtwqhRo/DKK6+gdevWKCgowP79+9GrVy906tQJ/fv3x7p16/D8889j9OjRaNWqFWQyGdLS0rBv3z48/PDDePTRRy3LDAgIwAsvvIBJkybB398fa9euRWxsLKZNm2Y5ERo7dixWr16N4cOHY/LkyQgJCcH27duxdOlSDB8+HPXq1QNg7jq/ZMkSTJgwARcvXkTjxo1x8OBBfPvtt+jRo8dtu8+PHj0a7u7uePvtt5Gfn4+ZM2eyV8gDpkaNGmjcuDF2796Njh07WnKvc+fOyM7ORnZ2Nj744IO7LudO//8jIiKwf/9+bNmyBQEBAXB3d0f9+vUxefJkxMbGon///hg1ahTq1asHnU6HixcvYseOHfjss894qSbdkUQiwXvvvYdXX30Vw4cPx/PPPw+dTocFCxYgNzcX7777rmXeJ554AnPnzsX48eMxfvx46HQ6/PzzzzCZTBVaZ6tWrVCvXj3MnDkTRqMRXl5e+PPPP3Hw4MEy5/fx8cHUqVNx+fJl1K1bF9u3b8eyZcswfPhwy3liRf9/kPN666234OrqirZt2yIgIADp6en47rvv4OnpiVatWgG4ub9dtmwZPDw84OLigtq1a6N+/foIDw/Hp59+ClEU4e3tja1btyI2NrbUeiZPnozhw4djyJAhGD16NIxGI+bPnw93d3dkZ2eXK9aePXvihx9+wCuvvIJhw4YhKysLCxYs4HlfOVWosOTu7l6uHiNZWVlQKpWVDorur9dffx1btmzBokWLkJ6eDr1eD39/f0RHR+O1116zDERWkeVptVqsWLEC8+fPR4MGDfDZZ5/hzz//LHOwvhdeeAE6nQ7vv/8+bty4gYYNG+L//u//0K5dO8s8TZs2xR9//GH5Jjo/Px81a9ZEp06dKjzwpa+vL5YuXYqPPvoIr732GpRKJR5++GEsWLCg1OWbAwYMQHx8PJYuXYovv/wSoigiLi4OtWvXrtA6ybE1a9YMTz75JNatW2fVPmfOHLRp0wa//PILli5dCpPJhMDAQLRt29ZqcOIXX3wRycnJmD17NnJzcyGK4j0P4Ld8+XKsXbsWX3zxhaXYpVAoMH/+fDzyyCOYOHEifvrpp3taBzmejz76CHK5HN999x3y8/PRvHlz/Pjjj5gzZ47VfB4eHli3bh3mzZuHZcuW4csvv4SXlxdatmyJoUOHAgCkUikWL16MH3/8EWvWrMF3330HqVSKoKAgdOjQwXIgWKxp06YYPHgw5s2bh/PnzyMgIAAffPABxo4da5nHz88PGzduxKxZszBr1izk5eUhNDQU7733ntV8rq6uWLVqFWbPno0FCxbgxo0bCAwMxLhx4zBx4sQ7vgfPPfcclEolXn/9dWg0GsydO9dqEFxyftHR0UhKSrIaR6lWrVqoU6cOzp8/X67xle70/3/GjBmYNm0aXn75ZWi1WkRFRWH16tUICAjAn3/+ia+++goLFixAWloa3N3dERoaim7dupU5UCzRrQYMGAClUolvv/0W48ePh0QiQWRkJFauXGl1bBwaGoqff/4Zn332GcaOHYuAgACMHTsWN27cwBdffFHu9UmlUixZsgTvvfce3nnnHSgUCvTv3x8ff/wxhg8fXmp+f39/fPLJJ5g5cyZOnjwJb29vvPbaa1ZfCFf0/wc5r/bt22PlypXYtGkTcnJy4Ovri3bt2uHrr7+23JQpNDQUM2bMwE8//WQZk+mLL77A4MGDsWTJEkyfPh1TpkyBTCZDdHQ0fv31V8sA3sW6d++ORYsWWYqtNWvWxIgRI3Dt2rVy9ziKjo7GF198gf/9738YOXIkAgMDMWTIENSoUQOTJk2q8vfG2QhiyX6Wd9G5c2e0aNEC//vf/+4438svv4z//vvvjl3snUFiYiLatGmDhISEarlW39ldvHgRHTp0wPvvv4+XXnrJ1uE8UO6Uu8xrclTMa3JGzGtyRsxrckbMa/tiMBjQp08fBAYGYsWKFbYOx2GVN3cr9BXe4MGD8dNPP2H37t23nWfnzp1YvHgxnnvuuYosmoiIiIiIiIiowiZNmoQNGzZg//792LBhA5577jmcPn0a48ePt3VoD4QKXQo3duxYLFmyBH369MHo0aPRv39/1KlTB4B5dPb169fjp59+QsuWLTFmzJhqCZiIiIiIiIiIqFheXh4++ugjZGZmQiaToXnz5vjll19uOy4jVa0KFZYUCgX+/vtvPP/885g/fz4WLFhgNV0URTz66KP4v//7Pw5yRXdVu3btex5nhoiIiIiIiB5sCxcutHUID7QKFZYA88CbmzdvRmJiIrZs2YILFy4AMA+69fDDD1sNWEtERERERERERM6rQoWl2NhYREZGwsPDA5GRkbcdvCkjIwMbN27EqFGjqiTIqqTRaPDbb7/h/PnzOHfuHHJzc/Hss89iyJAhtg6NiIiIiIiIiMihVKiw1L17d+zfv99yez+TyQRXV1f8+++/Vj2Vzp49izFjxthlYUmtVmPLli0IDw9Hhw4dsGXLlkovS6vVAgAOHz5cRdER3R9JSUm3nca8JkfFvCZnxLwmZ8S8JmfEvCZndKe8LqlChSVRFEs9LywsLNVuz/z9/bF8+XIIgoDc3Nx7KixlZ2cDAF588cUqio7o/lEqlahRo0apduY1OTLmNTkj5jU5I+Y1OSPmNTmj2+V1SRUeY8nRCYJQZcsKCgoCAMTExCAiIqLKlltRly5dQkhIiM3Wf68cPX7AMbehRo0aCA0NLdXOvK4ajh4/4JjbwLyuXo4eP+CY28C8rl6OHj/gmNvAvK5ejh4/4JjbwLyuXo4eP+CY23C7vC7pgSssVYeIiIjbjjd1P3h7e6Nu3bo2W/+9cvT4AefYhlsxr++No8cPOMc23Ip5fW8cPX7AObbhVszre+Po8QPOsQ23Yl7fG0ePH3CObbgV8/reOHr8gHNsQ1kktg6AiIiIiIiIiIgcU4V7LCUnJ0MmM7/MaDQCAE6ePGk1z63PnUlaWhrS0tIAlH8gK2dRPKaWyWSyehQWFqKwsBDZ2dnQ6XSW50ajEYWFhTAYDNDr9dBoNMjJybG8zmg0wmQy4erVq1B5eiFLZ4K2EDCJIoyiCJNJhNEkQqs3QKvTw2AUUWgSUSjC/NNogtEE87yQQIQAEwARgCgCIkQUFhrx5sSJkEikRe0ijCJgNJl/mkzmdRlNgMZggsZgtLzeVLSM4t9Nomj5WbyMQpMITaEJObl5cIs/bVmu0STCQyHFb8/Ut+0fzQGJoohr167h6NGjOH/+PLKzs6FWq6HRaKzyymg0lvq9wCgi1yiHznQzRwpNInSiBFookKUToTeKEAUJTBAgChLzAwAg3PIT5vkEqfl3ABAEiEWvU3l5o1OnaJhEsUTOmP/+hSYR2bl5cD1wqigXzdtlKvrdJIqQCAI2D210n99dcmQmkwkajQb5+fllPjQaDfR6vWWfW/zQ6gzQ6vXQFYrQm4BCSGAwwbwPFM3LFXFzv2iEAIMoMU8HIJpEmCBADynytHq4uLrCBMAkCiiE+bOklyhgkChK5HtRrhd9nkQIN6cV7asBAVKZDJ27dLF8fkqO2Fg8fKPl81hiPEexxDxiiWkl1w0U7bdN5p8RNd3w7WPhVfxXoftFFEXo9XrodDrodDqr32/XVvx5KHnMYjCaUGAE1IUCtEYBJtGErMwsqDy9YDQZYTKJ5nmLjkNMogkGowiDKEAvSqA3CTCJsBw7FIqAARJLDptMJoiieHOfL5rMP4uWKYrF/w/EEj+LPosQUCjIYBTMy2vbrj1qh4YWHXvcPA4pFM3HH6aiZYkA8tR5UCaesSyv0CRCV3TcZH4DrX5YfWZuvse3zmP+2bueJ6Z1qVWNf137I4oisrOzceHCBeTl5SEvL8+yn9VoNNDpdKWOQwoLzcchagOgNaLouNUEvdF87FFoAoxFOWgQAZPJvL81QrAcH4hi0bGsZX8o3DwWhTlHTEX5JsL8opv7wOL9rXl+nd4AuUJRtB8WYIIAoyAtOu4RSvy9xRK/F81bNI+5DTd/lti/1qhRA42aNjN/HkyiJY7i/XDxa8pez82FFv8+tIUfRraqWVV/QrIRURRRUFBQav9cVlvxfrr4fNFgMFj9fuvzwsJCZGRkQKlUwlBYaD7OFwF90TGLeZ8soBACDJCiUDTnvXmfXbwPFWAQZJb9qQigWfMWePSxvigoNJ83lDyfK973Wh3blNjP3lyu+TNfPG/J/XPxMgxGEQaTiJzcPCgPnbF8dowiYDCaUGi6ZV1ljGNd1sjWtxvuurzzusoF/DWscTn/wrdX4cLSyJEjS7U9//zzVs9FUazSsYzsycKFCzFjxgxbh1EmseiAxmQyQa/XQ61WIysrC3l5eSjQ6ZGTl4/cvHyo8zW4knYVmVnZSL+egRtZWVDna6EzFMIACTQGE7T6QhgFGUwSGYyQoMAI6KSuEFw9ALkrIFUAEgng4g4o3MzPXZSAVA5IZIAgAIIEEAQIEgkEiRSCRAqJ1AcQJBBlCohSOURBCgjNIObKzPOXJCl6yAAobzYLoqnoVEaEIIqQQCwqKxU/iuaDOQ8/3H0FEkGAAAGCAEgFQCIIkEoEyCSAtOh3N5kESoUEAgCJYB6Pq+TvEsE8r0Qwt8kk5tf5uMrgUiiBr5fCvFxBgFQCeCik9+PP7vBMJhN+//13HDp0CPHx8di5cyfy8/MBAAqFAl5eXvD09ISb0h26gAjkBbeE1qs2jDJXiBIZTHJXFMrcYJLJYFLIy1yHIJrgYtJBCT3kgnXOSErkTfEhlATmJxKIkMJUIqfMpDDB3c0VRtH8OgkAQSJY/f1VJgl8vF3NHwUIRXlkzr3i/KEHQ2FhIbKyspCRkYFdu3bBzc0NOTk5UKvVlpOV4kdefj7UWj1yDQLURinyjYC2ECgwAXqJCwSfYPN+V+4KSKSARAaJizskrkpIFG6A3BWCxB2QyGBy8YBJ5gKTVAFIYX6Uk0Q0QiKavzwq/nzIxEIIRgOkEgESS5sREpigQCFcYLB8Hsyfn6K8h1j0s8T+FObiqkwmhb9SZt5HF81T8vBBKPrUFbeV/NTcbLv52uJ283LMr5ZKzJ+7UC9FRf5sVEk5OTnIyMiwnIAXP/Lz85GdnY28vDyrAmnx73n5+VAXFCJPZ0B+gR5qrQ75+kLojIBBlKAQEvPxhUxhzn+llznfZS6AVAoIUsBFCYmbClKZAoLCFYKLOyB3AaQuMMndzJ8JhXvpoGUANHfYKCkgmIyQiIWQwgBBLPq/IRUhEU3mzwbMJ/aCIFpyEijKy6JcLD42Ls7z4mkSAIJg/t+jEIyQFX1mtDIl0jWFkJY4DhFg/v/hJpMUPTevS6aXwKv4swRALhXgIhUgkwhWn5WiVd6MDXf/zDXxdyvfH99BmUwm/PXXX7hw4QIOHDiAgwcP4sqVK8jMzLw5k1cApEENIQ9uCEmNMMAzACY3L4hyV4gyBUwyN5jkrqWPZYGiP5r5WEQQi49fTUXHsyIkotFyTHvrV1zFxyUCivfFIiTF85Xc792yOgAwGQ2Qy6SWfbgUIqSiEdLiZZe13wQgsRwnlVjeLftVCICPtwq1PRWQSYqPjcvI75KvKWOd5vnNORru7XK3PxVVI6PRiIsXL1qOT4ofubm5Vs9zcnKQl5dXqpifn5+P9PR0pKenQ6fTmRcqlZv318UPN0/AsyakSk/I3DwgVXpB6uoOwcUdgkwOQSqHIJMDMjkEqRKCRAZR5mJ+SGWAIEVhTTmMciX0UhfzeWQ5FH/uJDB/juTm/ygo/mxdE+U4sOsyFFIB8qL9plQApBLzcX3J4xOJZX9+8xyy+NxQLpVYfi95bFI8j0wiwFUmgVEhwMdNZnVOUPyQ3LKuUtty6/MyTieEW+Yqe56bFNKqOSepUGFp8eLFVbJSRzZu3Dg88cQTAMw9loYNG1bt67x8+TJOnz6NxMREJBw5irRCV1w2uiO9QIBGcEGhXAnRzRNQKAGpzHzgVfKDLDP/LgjeALyLltrAXKwJK3qUQS6YoBBEyATARwr4uAjwkAtQyswHK+ZiDKBSSOAik8LPwxXurgooZBLIpVLIpFLIpBJIJILlQ1Vc1HGRmQ+KFFIB6deuIrRWMAI95FC5SC0fxuITcYVUArnE/EEvPriyN+fOnUPdurd5I+m24uLi8OGHH+Lvv/9G586d0bRpU6xYsQKNGjWCwWBAREQE4tM0+OVIBlafyESm1oguYSp0rO0Bb1cpXKQSeLhI4esmhatMAm9XKQLc5fBQSIsKf+advkIq3Pe8MefEnQe5I+dkNBoRFxeHefPm4fTp0zh16hT0ej0AwMvLC94h9SALaQqhRiiMqmAYlH7QqbyglXlAJ3GBscSBkgQiXCQiVFJAJRcQ4iGBt6sMKlc53FzkcFPI4eEih7tCAqVcAjeZxHySKRHg6yaDykUKTxcpXIoOllxlErgU/ZRaHcDcPIgqPrgpizmvnW9cAKq4S5cuYfv27Th27Bj+/fdf5OfnIyMjAxcuXLCaTy6Xw81DBUVIEyiCG0LmVRPw8EWhezBMLh4weiihk7lDI3GDqayTcpg/B24Sc14rJICbTICPqxQeCgnc5FK4yKWQS6VQuUjh5SqDXGI+zvB0kcJNJoGLTAJ3hQQ+rjIEeMjhVvT/wstVCgFASkoK6tapU6q4WXxiIC/+rNjh8Ucx82cz3NZhOJQbN25g9+7dWLlyJdatW4cGDRogICAAw4cPR63QcBwrUOGEEIz9Vw3I0BqhlEvQpKYb6vu5ItxLAX93OdwVUijlEqhcJFAppHBXSMz7XoW06LjVvO9VSIUyTxKrE/fXVJIoirh8+TL27NmDDRs24NixY0hKSsKFCxeQm5uLvLw8qx4yHh4eUKlU8PT0hEqlgkqlgodKBRffQEhqhAMuXhDk7jDKVSiQqqCXusFVqoS/oIABUmiNgNpQOo4aShm8XKRwk0vg5SKFh4sU7nLzflomARQS8/lh8bGIq0yAykUKedFzTU4mGtQOhI+bDEq5eV43mcRSEHIp+rwp5VK4yO7/5648nHV/XaHC0ogRI6orDocRFBRkGdW/OqWnp+Py5cvYsGEDPv74Y4hBEfDqMwbqOmNhEqTwEvMR4gb4KaVwNeoQ7OsBdxksH0RXuQw+Hq7w9fSAt4cS3koF3OQ3/8mVLPIUn4DLJILlxOR+noifk2SibrjnfVkX2Y9Ro0Zh8eLFGDhwIA4ePIi2bdtaphlNIr7fdRIj/jmBw1c16FHXE5/2rI3+jX3g68Z7DpB9EkUR/fr1wz///IPc3Fz06dMH48aNQ+36jXFJHogDNwTsvaTFebUJSrkEDXxdUcfHBaFeCoR6KRCkUqCGUgZfNxn83GSo6S4rKhTZ30ERPZg0Gg3effddJCYmYs+ePVCpVOjYsSOaNWuGhg0bws/PD02aNEFwcDBc3ZSIvWLAryeysfl0NjL1JtT2VCDQQ46a7jKEebmYTzBcpajpLkeQhxx+SnPOuxYdh7jJzQVTuaR6j0lUCglULuxl/CAxmUyoX78+CgoK0KlTJ2zYsAGPPPII/r2Uh/nx1/DpyWwYCo14IkKJrx71Qbta7qjr42KXJ6lEd6LRaPDRRx9h4cKFyM7OhkKhQOPGjdG8eXP07dsXderUgZeXF1QqFcLDw+Gm8kKKRoJzWXqcz9Lh+HUtTt0oQEq+AWl5BuiN5uKTi1SAn1KGcG8X1PdSIMhDAX93GfyUcrjJBagU0qLiq3k/7i43F/NdZfc2xPO5cxrUretbFW8NVbEH8gwtISEBBQUF0Gq1AICLFy9i7969AIC2bdvCxcW2XTETExPRpk0bAED9h3og/MO/kGL0RHRDb4xv64+OtT3g5XrzT8dvJMjRGI1GxMTEYMmSJaUK1qdvFGD4urM4di0f49oGYu2zDRDqxe7RZN/S0tLw3XffYfPmzVi2bBn69OkDrVyFr+Ou4b3EdAiCDt3CVRjWGHiqTR0083fjCQo5lFOnTmHChAk4cOAAXn/9dUyaNAl9+/aFTCazOg4xiSKWH72B6TvO41q+AU829sHiJ+uie7gnfPjFANkBrVaL//3vf8jOzsaZM2dQr149mEQRn+25guk7L6FfIx8sfDwcTWTZaNaonq3DJaoUg8GATz75BIsWLUJeXh5mzJiBfv36QRRFNGjQwGreXJ0Rv5/Kwtf7s/Dn6VPQGUUEesgR7q1Ak5pKDGrqC393OYJVctQu6q3n5SLlF19k5YH8Dz9//nykp6dbnu/du9dSWFq0aBECAgJsFRoA4O+//0aTJk3wysJNeCc2E93qeOKvh0NR39fVpnERVYXMzEyMHz8eBoMB3bt3t5q2/GgGxm1KQbdwFbY96YuHmvJSMrJvV65cweOPP47Dhw+jcePGmDNnDvo/PRgfx17Bl/vPo5m/G+b3q4Onm/hALpWYT8ADlHdfMJEdWbduHUaOHIkGDRpg9erV6NatW5nzpWbr8Pzaszh8VYO3OgXhlfYB7GVKdkMURRw4cABvvfUWjh49inHjxiE8PBwagxHPrj6LPalq/PpMfTwVYe4Nce5cjo0jJqqc3NxcDB48GH/99RfmzZuHcePGwd3dPL7cuXPnLPOJoogfEq7j7a0XIJcI6N/YB6sHNygacoL7bqqYBzJjfvzxR1uHUKakpCT07dsX5y9eRpO3lmLSrhuY07s2Xm0fwIowOYXMzEx07doVeXl5WL16NUJDbxaO9qSq8cL685jdqzbe6BCA8+fP2zBSovL5559/kJKSgt27dyMqKgp/ns1Fs++PQiIAKwfVx+MNvbn/Joc3duxYjBkzBrNmzYJcXvZNEraczcFzq8+gTZA7kl5tgVqeHCyd7MuRI0fQoUMH9OjRA0ePHkVISAhEUcQrm87jRLoWB8c25Ze45PBmzpyJr7/+GjVr1sSePXsQHR1d5nwZGgNGrDuHPalqzO5dG6Mja0IuvbfL1OjB9kAWluyN0WjE4MGDsXHjRrRq3RqNpv6GE3ky7H22ISKDyrh7CZEDysrKQlRUFNzc3PDPP/+gVq2bty6+nKvH4FVn8HI7f0yICrRhlEQVc/ToUbRs2RLR0dHYfDobA349jbc7BWFal2Ao5RyzhRzf9evXkZGRgWHDht22qPTPFT1e2nUKkzoGYUb3WrzMk+zStWvX4Obmhu3bt1vafj6Ugd+O3cA/LzZhUYkcntFoxPTp0/HGG2/g008/hVJZdg/pg5fz8PTKM/B3lyHxpWbMfaoSLCzZgYsXL2LNmjX46quvcKPZAHxzIB3/vNgIzfx5uQQ5jyNHjuDUqVPIysqCt7e3pV1vNGHQqtNo6OeKOb1r2y5Aoko4ePAgmjdvjrOZBXh+7Vm81yUYH3QLsXVYRFXCZDLh8ccfR0BAABo2bFjmPPsuqjF+Vy4mdwzGRz2Y+2S/cnJy4OXlZXl+5KoGr/+Zgi8fCeMXueQUsrOzAQAvvvjibYtKe67o8VrsSTzbzA/fPBZ2z4NpExVjJtmBc+fOQS6Xw7frc/hs71X8+kx9FpXI6eTk5MDX19eqqAQA3/57DWcydfj1mXrsgksOw2QyYezYsdi6dSt6PtIXz6w8g6jaHni/a627v5jIQWzduhUHDhxAbGxsmScpRpOI8b+noF+4C2Z0Z+6TfStZWNIbTXh29Rk8FeGLsW1q2jgyontnMBjw/PPPIzg4GGFhYWXO8/upLIzbmYuJUYFY+Hg4i0pUpZhNduDcuXMIaNcbY39PxVePhOGR+t62DomoymVnZ1t9UwgAV9R6fLT7Mj7tWRuBHhyPgxzHjBkzsGjRIqxevQartPWg1hvxy4B6vASInMr8+fMxYMCA2/ZWWnokA6nZOkyOdOdYYmTXjhw5gmnTpiEiIgIAsPhQBjI0Bnz3WBhzl5zC4sWLsX//fmzbtg2enp6lpp/NLMDwtecwvpkSH3YPYd5TleOlcHbgxLmLuN59Ml5q44+X29v2jnRE1eXkyZPw8fGxantn20VE1HDDC61r2CgqoooxGAxo2bIlkpKSsHTpUhgbdcGatWcRN6Ypb6VOTiU1NRUbNmywGo+mpHy9ER/svIR3ooPh51pwn6MjKj+DwYDevXvD09MTy5Ytg67QhFl7rmBiVBC8eOcrcgKiKGLhwoV47rnnLMXTkgoKzT30omp74JUWzHmqHswsO7A11xduPkZ83ofjy5Bz2rt3L2bPno2YmJibbRfUWP7fDex9sQl7eZDDuHHjBpKSkrB582b06vMwmn1/FBM6BKJFAC9fJucyd+5ctG7dGt27dy9z+tdx1yARBLzRIRBpF1Pub3BEFZCamorr169j27ZtUCqVWBCfDo3BhFf5ZS45iU2bNuHYsWNYv359mdOnbL2I9PxC/DmsEXKvXry/wdEDg4UlGzt88gxOqJpjkMcVji9DTuvjjz9Gjx49MGTIEADmb1Ym/JWKka1q4KEQDxtHR1R+WVlZAIB27dphUcJ1ZGoL8XZ0kI2jIqpamZmZ+P777/Hbb7+VeblEvt6IL+PS8GnP2nCT89iF7NvPP/8MX19fNGvWzNJbaVLHQKhceOdOcnwFBQV4/fXXMW7cONSuXbqTwt4Lanx/8Bq2Dm+MGko5cm0QIz0YeDRgQ6dOnUKXt/4HuVGL78Y+butwiKrFjRs38Ndff+GVV16xtO1OVeO/a1reQYgcTmZmJgBA6qbCR7svY1rnWvDmpRTkZDIzM2EymRAdHV3m9KWHMyCXSPB8S17GTPYtMTER3377LT755BNIJBL8fOg6tAYTXmFvJXISFy9eRGpqKiZPnlxqmq7QhHGbzmNU65roFl563CWiqsTCko2IooiJU6dD2+IJfP5EU/j5eN39RUQOKCkpCQDQpUsXS9vC+HT0b+yNYBUH7CbHkpWVBU9PT3x94DrcFRKMb+dv65CIqpxWqwUAuLm5lZpmNIn4Mu4qXm0fwDsKkV3T6XRo27YtGjdujLFjx1p6K03uGAgPBXsrkXNQq9UAgJo1S9/d8LN/riCrwIjZvTncClU/u/maVavVIiYmBnv37oVarUZISAgGDhxodTJalnfffRfHjh277fSlS5daBgy+3bytW7fGjBkz7m0DKuDQoUOYP38+/shwR7CnEi9Fhd63dRPdb4sWLULv3r3h5+cHALiWZ8C6pCz8MbTsuwwR2bPMzEx41aqDL/Zfxfx+4XDhiTU5IY1GA6DswtK6k1m4mmfAuLa8RTvZN41GA1EUsWTJEkgkEmw6mYlcnZE3yiGnkpubC5lMBldXV6v2Uze0mLUnDcsG1mPParov7CbLZs2ahdOnT2PEiBEIDg5GbGws5s6dC1EU0bVr19u+7qWXXrIcABXT6XSYMWMG6tWrV+ouVIGBgZg4caJVm4fH/R3jZcqUKdgauw+e723Dh73qcGwlclpXrlzBqlWr8MMPP1jaFh+6jnBvBbrXYZdccjxZWVkoaD8UDf1c8VxzP1uHQ1QttFotpFIp5HK5Vbsoipi3Lw0vtKoBP6X8Nq8msg86nQ4A4OLiAgBYm5SJfo182FuJnMr58+fh7e1dajy86Tsvo3OYCk9F+NzmlURVyy4KS/Hx8Th8+DAmTZpkKSK1aNEC6enpWLx4MaKjoyGVlv1PIDS0dG+f7du3o7CwEL179y41TaFQoHHjxlW7ARWk0WjQ571FOOXhihGtOD4BOSeTyYSBAweicePGlkG7jSYRixLT8Uq7AN4JjhxSamY+MsO745eeIcxhcloXLlyAQqEodaKSmKbBgcv5iHmqno0iIyq/goICAICrqysKCk3441Q2fu5f18ZREVUdURQxd+5cvPjii1btiWn5WH08E/tGNynzBgxE1cEuusrExcXBzc2t1CCRPXv2RGZmJk6dOlWh5W3duhVubm7o3LlzVYZZZfINIvaZQvFOdDAU7K1ETio/Px9xcXGYMWMGJBJznm85m4M0tQHDWVAlB5SZmYnlR2/Ao1CNPvU4Lh45r5UrV5bZW3xdUiaiQjxQz9e1jFcR2ZeSPZa2n8tFoUnEI/W57ybnsWvXLiQnJ+O1116zap+2/RKejPBB+1q88zLdP3bRYyk1NRUhISGleiXVqVPHMj0iIqJcy7py5QpOnDiBPn36lDk2wNWrVzFkyBBoNBr4+/ujc+fOGDRokKWb7P2QFtIRLoKIF1rz5Jqcl16vBwDUr1/f0rYgPh1PN/VFDV5CQQ5GFEWMHTcO2aGDMaFjOL8BJKel1WqxZcsWrF+/vtS09SezMLIVx1Yix1CysLQuKQMP1/eGOy+DIyfy22+/oXfv3qhVq5albXdKLrady8F/45vbMDJ6ENlFYUmtViMgoPRAesVjHxWPdl8eW7duBQD06tWr1LQmTZogOjoaISEh0Ov1SEhIwNq1a3HixAnLbUjvh+zaHfCkVzZ7K5FTMxgMAMyXnwLAxRwdNp/Oxq6R5SsSE9mT3bt3Y80//0Hy0lsYFxVm63CIqs3u3bshkUhK3TwlOUOLpIwCPMnxOshBFBeWpHIFNiZn44tHeLMcch6iKGLVqlX4/PPPrdqmbb+E4S1rIKJm6Q4WRNXJLgpLAO747W95vxk2Go3YsWMHQkNDyxxHadiwYVbP27ZtC39/fyxevBj//vsvoqKi7rqOtLQ0pKWlAbh5G/WKyNAYYPCpjVZeVyr8WiJHUtxjqbiwtOp4Jur6uKBjbXbLJcdz7Ngx1Ow5HE3DVKjjc/96uBLdb7GxsWjfvn2pG5usP5mFZv5uqM/L4MhBFBeW4tJ0yNEZ0a+ht20DIqpCFy9eRGZmJjp16mRp23NBjQOX8xAzkOPg0f1nF11mVCpVmb2S8vLyAJT/rm3x8fHIysoqc9Du2+nWrRsAIDk5uVzzL1y4EG3atEGbNm1KFarKY0+qGtDlo7EPu+KScysuLBXfVWhNUiaebuLLS4jI4ej1esz78mvk1emE51vyEmZyblu2bCnzOGr9ySw82Zi9lchxqNVqyOVyrD+ZhR51PHnLdXIqP/30E0JDQ1Gv3s0i0hf7r2JgE1+Ee/MLMLr/7KKwFBYWhkuXLsFoNFq1p6SkWKaXx9atWyGTydC9e/cKx1Dek91x48YhISEBCQkJiImJqfB6YlPVEC7+B3c3fuNHzq3kpXAXc3SIu5SPp5v42jgqoopbuXIlLsiDILgomcPk1ERRxLFjx9ChQwer9su5ehy4nM/CEjkMURTx8ccfo1ZIbaw/mYUBvISTnIjJZMKyZcswZswYyGTmgumpG1r8npyNN6MCbRwdPajsorAUFRUFrVaLffv2WbXv2LEDvr6+aNiw4V2XkZWVhYSEBHTo0AGenp7lXveOHTsAAI0aNSrX/EFBQYiMjERkZGS5BxQvKTZVDeP5+DIHFidyJiUvhVtzIgv1fFzQKlBp46iIKmb8+PF4/vnn0XzYFAxo4gdPF/Y2Jeel0Wig0+lQo4Z1z7wNJ7MQ5qXgPpwcxrp16xAfH4/Zv2xEmtrAoig5lU2bNuHSpUt48cUXLW1fx11Dx9oevBMc2Yxd9Alt06YNWrVqhfnz50Oj0SAoKAixsbFITEzExIkTLXeL++abb7Bjxw788MMP8Pf3t1rGjh07YDQab3sZ3PHjx7Fy5UpERUUhICAABoMBCQkJ+Pvvv9GiRQu0b9++2rczS1uII1c1QOphuLqyxxI5t5KXwq0+wcvgyDGdOHECQ14ch7Xwx+wWfrYOh6ha7d+/HzKZDKGh1oMcF18Gx304OYqdO3eiT58+SNCo0DlMAn933o2WnIPBYMCYMWMwbNgwBAUFAQBuaAxYejgDvzxV18bR0YPMLgpLADB16lTExMRg+fLlUKvVCAkJweTJk63uSmIymWAymSCKYqnXb9u2Df7+/mjVqlWZy/fx8YFUKsVvv/2G3NxcCIKAoKAgDBkyBAMGDLgvd4Tbe1ENpVyA+koSeyyR0yse3P6qRsT+S3n45jHeSYscT0FBAfT1O8NPLkOvul62DoeoWq1evRp9+/aFn9/NImqWthC7UnIxrUuwDSMjKr/jx4/ju+++w+LFS/BpUhZebV/6ztNEjiovLw/Xr1/H66+/bmlbGH8dwSo5nmjEnnlkO3ZTWHJzc8OYMWMwZsyY284zYcIETJgwocxp8+fPv+Pyg4ODMX369HsJ8Z7FpqjRxl+BXSYjeyyR0/v888/x5JNPYn1yNur6uKA1L6EgB6TVanHIGIShbfwglbC3Bjm3ixcvomnTplZtm09nw9tVhk6hKhtFRVQxZ8+ehUKhQOQjT+PswuPoz8vgyIkUFBQAuHlzK12hCf87eA3vdg7mcQrZlF2MsfSgiE1Vo01N82V97LFEzuzo0aM4duwYZs6cycvgyKHlyjxxvtCDd4Mjp2c0GrFnzx5ERkZatW9MzsZjDb0g4wkLOYjs7GzUqlULm05lo1WgEmG8QxY5keLCUnEnhZj/bkBXaMKIVjxOIdtiYek+UeuMSEzLR0tv82V8LCyRM1uxYgUiIiLgVbsB9l3M4520yGFl1Y5CuIsOzfzZ446c24EDB5CXl4c+ffpY2nSFJvx1JpuXV5BDOXPmDHx8fLApORtPNPK2dThEVapkYcloEjFvXxpebhcADwVvLkK2xcLSfbLvohpyqYD6bjoA4KVw5NTi4uLQs2dPrD+ZhTreLogM4kk5OSZNaFt08dPZOgyiavfvv/+idu3a8PW9+UXAzpRcGIwi+tTj+GLkGAwGA3766SdE9X4cB6/k43EWRcnJlCwsbUjOwoUcPV57iOOIke2xsHSfxKaq0SHEA0a9dfdFImdz4MAB7Ny5EyNHjuRlcOTQruUZYKpRB+1rMH/Juel0Orzzzjt45plnrNo3JWejZ11PfhNODuP69eu4cuUKanQagNqeCo7vSE6nuLCkUCjw+d40vNC6BmryrodkB1hYuk9iU9XoEqZCQUEBXFxceKJNTistLQ2+vr4Ibtgcey/kYWATfltIjmnn+RyImhw08VPYOhSiaqXVaqHT6TBy5EhLmyiK2JScxcvgyKGsX78egiAgLkuOxxt583ibnE5BQQHkcjn+uaRBwpV8TIwKsnVIRABYWLovNAYjDl7OR5cwT2i1Wo6vRE4tNzcXKpUKG5OzEeKpQNtgd1uHRFQp285mAykJcFdyn03OTa/XAzB/A14sIS0fl9UGPM4xashBXLlyBQsWLMDg51/A7gu8DI6c0z///AMfHx/M/icNg5r5oY4PB6cn+8DC0n2w/2IeAOChEHcUFBTwMjhyamq1Gp6entiQnIX+jX34bSE5rF0paojn4/llADk9g8EAwLqwtPFkNh6q5Y5AD/bYI8cwZswYHD16FF1HvQVXmQTdwlW2DomoSl27dg2ffPIJRr7zKbaczcHbndhbiewHC0v3wZ5UNdrXcoerVMCff/5pdeBG5GwOHz6MGrXCsONcLvo35reF5JhSs3U4n2MAzh3klwHk9MrqsbSRl8GRg8nJycGsWbOwP1OOR+p7QSHlaQ45l6VLl0IqlSI1pDMeqe+FFgEcQ4zsB/e490Hx+EozZszA4sWL0aNHD1uHRFQtCgsLsXz5cjTq9wI8FBJ0DvWwdUhElbLjfC783QTgxgX2WCKnd2th6XyWDkfTtXiCXw6QA9FqtVC4umHz6WxewklOx2AwYNOmTeg3dDTWJGWztxLZHRaWqllBoQlxl/Lgci0Js2fPxvfff4/FixfbOiyiarFv3z7k5+fjqncj9G3oDTm/LSQHtTMlF+1qmvOXhSVydsWFJbncfGehtUmZaOjnioga7K1HjkOj0SBNVhO5OhMea+Bt63CIqtTevXvN4yv1egHN/N3QJYyXepJ9kdk6gGJarRYxMTHYu3cv1Go1QkJCMHDgQHTp0uWOr9u+fTu+/vrrMqctXboUPj7W37YdPnwYy5Ytw/nz5+Hi4oJ27dph5MiR8Pb2rqpNsXLgch4KTSLO7lqP0NBQjB49ulrWQ2Rrly9fxrRp09Dvyaew44IWS57kNynkmERRxM7zuRhVx4RNYGGJnN+tYyytPpGJp5v4cow8cigajQbHC/3QNUwFb1e7OcUhqhK5ubnwqRmADRdM+LBbMPfPZHfsZq87a9YsnD59GiNGjEBwcDBiY2Mxd+5ciKKIrl273vX1b7zxBmrVqmXVplJZV3KPHTuGGTNmoG3btpg2bRpycnKwdOlSvP/++/jiiy8s39RVpT2parQJdkfa3rN45plnqmUdRPZg3Lhx+Oeff/DjruPY8U8++tTzsnVIRJVy6kYBrqgNaOaugyAIcHHhHVfIueXm5gIw91hKzdbhwOV8LOhXx8ZREVWMRqvFYa0KU3kZHDmh1NRUSJv2QEGhiKEt/GwdDlEpdlFYio+Px+HDhzFp0iRLEalFixZIT0/H4sWLER0dDalUesdlhIaGokGDBnecZ/HixQgODsY777xjWV5AQACmTJmCrVu34rHHHquaDSphd9H4ShsvXMBTTz1V5csnshdqtRoff/wxDuS6oXc9GdwVd/7MEtmrHedzUc/HBV7IhqurK78VJKe3atUqdOjQATKZDGuTrqO+rwtaBLCnHjmWfDd/6PQyDjpPTkev12PWrFkQhn2H4S1rwIPH2GSH7GIAlLi4OLi5uSE6OtqqvWfPnsjMzMSpU6fueR03btzA6dOn0b17d6siVUREBGrVqoW4uLh7XsetDEYT9l/MQ+dQFS5cuIDQ0NAqXweRvdDpdJDJFdiYnIUnOeArObAd53PRvY4ntFotL4Mjp6fRaLB8+XKMGzcOAC+DI8ckiiJ04e1Rz92EMG/2MiXnEh8fjzSTO264h+Cltv62DoeoTHbRYyk1NRUhISGleiXVqVPHMj0iIuKOy5g5cyZyc3OhVCrRrFkzDB06FGFhYVbrAIDw8PBSrw0LC8PJkyfvcStKi7+SD63BBGXGaWg0GhaWyKnp9XpclfkhI68Q/Rp62zocokoxiSJ2peTi28fCoT3GwhI5v0OHDiE7OxsDBgzAxRwd4i7l47vHwm0dFlGF6PV6iA2j0TXILr4zJ6oyR44cwZQpUxDU72U0CVchoiaPS8g+2UVhSa1WIyAgoFS7h4eHZfrteHt7Y9CgQWjUqBHc3NyQmpqKNWvW4K233sLs2bMtxaniZRQvsySVSmUZX6Aqxaaq0SpQifffHofIyMi7XqpH5Mh0Oh1OGGuic6gKNZQcS4wc03/XNMjUGtE11AOvTF/JwhI5vUOHDqFRo0bw8vLCkrirqOvjglaBSluHRVQhZ9NzIYQ0Q58w3smQnMvbb7+N+KMnIX24E15qV/p8mche2EVhCcAdu1zfaVqbNm3Qpk0by/NmzZqhbdu2eP3117Fs2TK899575VpWebt8p6WlIS0tDQCQlJR0x3mvZWTi1B+LoYuPx5EjRyx3WyFyRgU6Hf7T++AdXgZHDmzn+Vw083fDLwu+xrp167BixQpbh0RUrbZu3YpWrVoB4GVw5Lh+P5UFMecaIoOa2DoUoiqVl5eHPm99gUQ3OfpzYHqyY3ZRWFKpVGX2SsrLywNQdi+jOwkICEBERASSk5Ot1gGU3ftJrVaXuoPc7SxcuBAzZswo17wf9ghFy5wW6P7daV4GR04vX+mP64UK9GdhiRzY9vO5eMhfioVzF2LKlCl49tlnbR0SUbVJTU3Fxo0bsWnTJlzO1WPfxTx89UjY3V9IZGf+PJcPJO+Bu3s7W4dCVKU0Wi0OmmrjpUh/yKW81JPsl11kZ1hYGC5dugSj0WjVnpKSYpleUaIoQiK5uXnFyygea6mk1NTUchd+xo0bh4SEBCQkJCAmJuaO83p6emLEiBEsKtEDIa9WG4S76hHOQTPJQRmMJuxJVeP6wb9w7tw5jB071tYhEVWrkydPQhAE9O3bF2uTMhHurUBkEC+DI8ei1hmx/6oBYnIsL18mp5OpCsN1kxtGt6lp61CI7sguCktRUVHQarXYt2+fVfuOHTvg6+uLhg0bVmh5V69excmTJ9GoUSNLm5+fHxo2bIhdu3ZZFbBOnjyJy5cvIyoqqlzLDgoKQmRkJCIjI+86oDjRg0RXpwOifPS2DoOo0uKv5ENjMMF0PhGvvvoq6tata+uQiKqVRqOBn58fBEHgZXDksP4+mwOFRARSEqBUsjBKzuVGna5o456HYBWHVCH7ZheXwrVp0watWrXC/PnzodFoEBQUhNjYWCQmJmLixImWu8V988032LFjB3744Qf4+5tvtfj++++jadOmCA8PtwzevXbtWgDA0KFDrdYzYsQITJ8+HbNnz8Zjjz2GnJwcLF26FGFhYejVq9f93WgiJ3I+SwdTQAN09b/9QPtE9m7neTXaBLkj/c+z6NRmgK3DIap2Go0GSqUSF3J02HshD3P7sIc1OZ6NyVlo52NErADI5bx5CDmP7IJCaGu3Re8aObYOheiu7KKwBABTp05FTEwMli9fDrVajZCQEEyePBldunSxzGMymWAymSCKoqUtLCwMe/bswbp166DX6+Hl5YUWLVpg8ODBqFWrltU6mjdvjunTp2P58uX4+OOP4eLigrZt2+KFF17gPyKie7AuKRPijQto6ONv61CIKm1nSi661/HELxcvIiQkxNbhEFW7lJQUKJVKLD2cgYiabmgb7G7rkIgqxGA0YfOpbLwYrMFBXgZHTmbV8UwIujy0r8mepGT/7Kaw5ObmhjFjxmDMmDG3nWfChAmYMGGCVdvo0aMrtJ7WrVujdevWlQmRiG5jbVImcGInXIcOt3UoRJWiNZiw94IaEzv4Y05aGmrXrm3rkIiqVWFhIRYtWoSnn3kGSw5fx6vtA3gZHDmcfy7kQa03oYkih5fBkdNZ9t8NiEe3wOPRx20dCtFd2cUYS0TkuNLUesRdyoeYtBMuLhy4mxzT/kt5MInAlbg/YTKZEB4ebuuQiKpVeno6UlNT0aTvcFzONWBYixq2DomowjYmZ6FruAoGdRYH7iankpKtw54LapgO/8HcJofAwhIR3ZP1J7MQ7CEDriRBoeDAguSYtp3JQgtfAR+8OwVvv/02eyyR0zt+/DgEQcBfV+V4vJE3arpzSAByLKIoYmNyFvrWU2HOnDlo3769rUMiqjLL/stAYz8FcPUUe+ORQ2BhiYjuybqkLDwS7goA7LFEDitYosbB+VNRu3ZtTJ8+3dbhEFW7X375BdG9H8PvZ9QY1Zq3sSbHczRdi5RsPTL2rcfp06exYMECW4dEVCVEUUTMfzfwVH1zQYk9lsgRsLBERJV2Q2PArpRc9A4xD9fGwhI5qpe6NsaVrf+HvXv3wt2dAxiTc5s6dSp++eUXRA5/BzWUMvSp52XrkIgqbOPJLLQOVCLu7/UYMWIEfHx8bB0SUZU4eCUfp28U4LHa5jujs7BEjsBuBu8mIsezMTkbfkoZCs8fAsDCEjkumUyGoKAgW4dBdF/s27cPw4YNw548b4xo5Q2phIN2k+PZmJyFcP1FrNm2DX/99ZetwyGqMjFHMtC9jie8JXoALCyRY2CPJSKqtHUns/BIuBteHPUCBg0aBD8/P1uHREREd6HRaODfsjMOX9XghdYctJscz6VcPRLSNCg4shXdunVDnz59bB0SUZUwGE347XgmhjT3g1arBcDCEjkGFpaIqFJydUZsPZuDgkN/QaFQICYmBhIJdylERPYuPz8fiWItdA9Xoa6Pq63DIaqwTclZCPNS4MaJOPTp0weCwF535Bz+OpODPL0RA5v4Ij4+HgDg6sr9NNk/ngUSUaVsPp0NpVyCi/9swIQJEyCX845CRESOIE9nQHyBL17goN3koDYkZ+HxRj44d/Ys6tevb+twiKpMzH830L+xDzat/hXjxo3D4MGD+cUtOQRmKRFVytqkTDRCOvbG7kZ0dLStwyEionLKDmwJQMBTEb62DoWownIKCrHrvBrtvQqQnp7OwhI5jeyCQmxKzkInlRqTJk3C6NGj8euvv9o6LKJyYWGJiCpMazDhz9M58LgYj+joaPTs2dPWIRERUTmIoghNk0fRvaYebnIeBpLj+ftsDtwVEqz54n00adIEERERtg6JqEqsPpEJL1cpjm34Cbm5ufj0009tHRJRudnNXeG0Wi1iYmKwd+9eqNVqhISEYODAgejSpcsdX7dv3z7s3bsXp0+fRmZmJry9vREREYHnnnsOwcHBVvO+++67OHbsWKlltG7dGjNmzKjS7SFyZlvO5kAAIE1NQMeOHTm2ARGRg9h+Lgem4AgMrWuydShElbLlRBpcLyZi3ZpV2LlzJ8efIaex7L8beLaZH05uP4s33ngDNWvycmVyHHZTWJo1axZOnz6NESNGIDg4GLGxsZg7dy5EUUTXrl1v+7q1a9fC29sbgwYNQmBgIK5fv47Vq1fjzTffxOeff47Q0FCr+QMDAzFx4kSrNg8Pj2rZJiJntTYpE4819MbpNRcR2OP2n08iIrIfoijig52XgCN/okH/AbYOh6hS5vQMhs+hC3ho1Sp069bN1uEQVYmUbB1iU9WY3bMW+sbHY8AA7qPJsdhFYSk+Ph6HDx/GpEmTLEWkFi1aID09HYsXL0Z0dDSkUmmZr33vvffg7e1t1daiRQuMGTMGGzZswGuvvWY1TaFQoHHjxtWyHUQPAr3RhN9PZeOLnoFYfeQI3nrrLVuHRERE5bDtXC7i0zQQY3+Gu/swW4dDVCm+vr74/NOZtg6DqEotTbyKMKUJH770HDIyMtCyZUtbh0RUIXZxcX1cXBzc3NxKDQDcs2dPZGZm4tSpU7d97a1FJQDw8/ODn58fMjIyqjpUogfezvO50BpM+Ovb6XB3d0dUVJStQyIiorsQRREf7b6Mp+q5AllXoFQqbR0SEREViXDV4PyiyTh+7BiOHj2Khx56yNYhEVWIXfRYSk1NRUhISKleSXXq1LFMr8jAfFevXsX169fRoUOHMqcNGTIEGo0G/v7+6Ny5MwYNGgQXF5d72wiiB8SKxMvwz0vBqmVLsXHjRoSFhdk6JCIiuout53Jx4HI+pj/ihl8BuLu72zokIiIq8kxUYzyTvIfjlpLDsovCklqtRkBAQKn24rGP1Gp1uZdlNBrx7bffwtXVFU888YTVtCZNmiA6OhohISHQ6/VISEjA2rVrceLECXzyySeQSOyiAxeRXZvZuQYmbN6OSV9+iUcffdTW4RAR0V2IoogZuy5hZKsaWPmD+S5DHF+SiMh+sKBEjs4uCkvAnT9M5f2giaKIb775BsePH8c777xTaiT9YcOsxxNo27Yt/P39sXjxYvz777/luqQnLS0NaWlpAICkpKRyxUXkTGoHBWDNkgW2DoOIiMppy9kcxF/RYKDwHyb9+CPWrFkDhUJh67CIiIjISdhFFx2VSlVmr6S8vDwA5ftWTRRFfPvtt9i9ezfeeOONMi+DK0vx3SSSk5PLNf/ChQvRpk0btGnTplShioiIiMiemHsrXUbfWiJmT52AsWPH4qmnnrJ1WERERORE7KKwFBYWhkuXLsFoNFq1p6SkWKbfSXFRafv27Xj11VfRvXv3CsdQ3l5R48aNQ0JCAhISEhATE1Ph9RARERHdL3+dyUFimgaav79HYGAg5syZY+uQiIiIyMnYRWEpKioKWq0W+/bts2rfsWMHfH190bBhw9u+VhRFfPfdd9i+fTtefvll9OrVq0Lr3rFjBwCgUaNG5Zo/KCgIkZGRiIyMrNCA4kRERET3U3FvpeEtfLBrw69499134eXlZeuwiIiIyMnYxRhLbdq0QatWrTB//nxoNBoEBQUhNjYWiYmJmDhxouVucd988w127NiBH374Af7+/gCAH374AVu3bkWvXr0QFhaGkydPWpYrl8tRr149AMDx48excuVKREVFISAgAAaDAQkJCfj777/RokULtG/f/v5vOBEREVE12XwqC4ev5sP9z0/g4uKCHj162DokIiIickJ2UVgCgKlTpyImJgbLly+HWq1GSEgIJk+ejC5duljmMZlMMJlMEEXR0nbw4EEAwLZt27Bt2zarZfr7++PHH38EAPj4+EAqleK3335Dbm4uBEFAUFAQhgwZggEDBvCOcERERORULlxOg277DziQsAn79+8vdVMTIiIioqpgN4UlNzc3jBkzBmPGjLntPBMmTMCECROs2ooLR3cTHByM6dOn30uIRERERA5jVKf66PTN6wgO/gw1atSwdThERETkpOymsEREREREVcfFxQUtWrSwdRhERETk5Hj9FxERERERERERVQoLS0REREREREREVCksLBERERERERERUaWwsERERERERERERJXCwhIREREREREREVUKC0tERERERERERFQpLCwREREREREREVGlsLBERERERERERESVIrN1APebVqtFTEwM9u7dC7VajZCQEAwcOBBdunSxdWhERERERERERA7lgSsszZo1C6dPn8aIESMQHByM2NhYzJ07F6IoomvXrrYOj4iIiIiIiIjIYTxQhaX4+HgcPnwYkyZNshSRWrRogfT0dCxevBjR0dGQSqU2jpKIiIiIiIiIyDE8UGMsxcXFwc3NDdHR0VbtPXv2RGZmJk6dOmWjyIiIiIiIiIiIHM8DVVhKTU1FSEhIqV5JderUsUwnIiIiIiIiIqLyeaAKS2q1Gh4eHqXai9vUavX9DomIiIiIiIiIyGE9UGMsAYAgCJWaViwtLQ1paWkAgKSkpCqLi4iIiIiIiIjI0TxQhSWVSlVmr6S8vDwAKLM3060WLlyIGTNmVHlsRERERERERESO5oG6FC4sLAyXLl2C0Wi0ak9JSbFMv5tx48YhISEBCQkJiImJqY4wiYiIiIiIiIgcwgNVWIqKioJWq8W+ffus2nfs2AFfX180bNjwrssICgpCZGQkIiMjERERUV2hEhERERERERHZvQfqUrg2bdqgVatWmD9/PjQaDYKCghAbG4vExERMnDix1N3iiIiIiIiIiIjo9h6owhIATJ06FTExMVi+fDnUajVCQkIwefJkdOnSpcLL0mq1AGw/iPelS5eQnZ1t0xjuhaPHDzjuNjRu3BhKpdKqjXldNRw9fsBxt4F5XX0cPX7AcbeBeV19HD1+wHG3gXldfRw9fsBxt4F5XX0cPX7AcbehrLy2IlKlxcTEiAD44MNhHwkJCcxrPpzuwbzmwxkfzGs+nPHBvObDGR/Maz6c8VFWXpckiKIogiolIyMDf//9N8LDw+Hm5maTGJKSkjBs2DDExMQ45JhPjh4/4NjbUFblmXl97xw9fsCxt4F5XT0cPX7AsbeBeV09HD1+wLG3gXldPRw9fsCxt4F5XT0cPX7Asbfhbj2WHrhL4apSjRo1MHToUFuHAQCIiIhAZGSkrcOoNEePH3CObQCY11XJ0eMHnGMbAOZ1VXL0+AHn2AaAeV2VHD1+wDm2AWBeVyVHjx9wjm0AmNdVydHjB5xjG271QN0VjoiIiIiIiIiIqg4LSw4uKCgIH3zwAYKCgmwdSqU4evyAc2yDvXH099TR4wecYxvsjaO/p44eP+Ac22BvHP09dfT4AefYBnvj6O+po8cPOMc22BtHf08dPX7AObbhdjjGEhERERERERERVQp7LBERERERERERUaWwsERERERERERERJXCu8LZOY1Gg99++w3nz5/HuXPnkJubi2effRZDhgwpNe/Zs2exZMkSJCcnQyqVonnz5hg1ahQCAwMdOtbff/8df/zxB65duwZfX1/07NkTzzzzDGSyyqevrWPNzs7GkiVLcPDgQeh0OtSpUwfDhg1Dy5YtK71NjoR5zbx2Rsxr5rUzYl4zr50R85p57YyY18xrW2KPJTunVquxZcsWGAwGdOjQ4bbzXbp0Ce+++y4MBgPefvttvPbaa7hy5Qreeecd5OTkOGysK1euxKJFixAVFYUZM2bgsccew6pVq7BgwQKHjdVgMOD999/Hf//9hzFjxmDatGnw9vbGhx9+iGPHjt3TdjkK5jXz2hkxr5nXzoh5zbx2Rsxr5rUzYl4zr21KJLtmMplEk8kkiqIo5uTkiI8//ri4bNmyUvN99tln4tChQ8X8/HxL27Vr18QBAwaIixcvdshYc3JyxIEDB4rfffed1et/++038YknnhBTU1MdMtY//vhDfPzxx8WkpCRLW2Fhofjyyy+LkyZNqvQ2ORLmNfPaGTGvmdfOiHnNvHZGzGvmtTNiXjOvbYk9luycIAgQBOGO8xiNRhw8eBAdO3aEUqm0tPv7+6N58+aIi4ur7jABVH2siYmJ0Ov16Nmzp9UyevXqBVEU72m7bBnr/v37UatWLTRu3NjSJpVK0a1bN5w6dQo3btyo9HY5CuY189oZMa+Z186Iec28dkbMa+a1M2JeM69tiYUlJ5CWlga9Xo/w8PBS08LDwy3T7UFFYk1NTbW0l+Tr6wtPT09cuHDBIWO9cOHCbZdZPJ2Y144WK/O6fJjXjhUr87p8mNeOFSvzunyY144VK/O6fJjXjhWrI+U1C0tOQK1WAwA8PDxKTfPw8IAoisjLy7vfYZWpIrGq1WrI5XK4urqWmlelUlmW5WixqtVqqFSqMucDgNzc3CqJ39Exrx0rVuZ1+TCvHStW5nX5MK8dK1bmdfkwrx0rVuZ1+TCvHStWR8prFpacyJ266N2t+979Vt5Y7SHu+x2rPWyzPWFeVw/mtW0xr6sH89q2mNfVg3ltW8zr6sG8ti3mdfV4kPOahSUnUFyxLKsSm5eXB0EQ4O7ufr/DKlNFYlWpVNDr9dDpdKXmvV311hFivV3VvLiturfLUTCvHStW5nX5MK8dK1bmdfkwrx0rVuZ1+TCvHStW5nX5MK8dK1ZHymsWlpxAUFAQFAoFUlJSSk1LSUmxTLcHFYm1+NrRW+fNyspCbm4uQkNDHTLWsLAwy7W1JRW3Vfd2OQrmtWPFyrwuH+a1Y8XKvC4f5rVjxcq8Lh/mtWPFyrwuH+a1Y8XqSHnNwpITkEqlaN++Pfbv3w+NRmNpv379Oo4ePYqoqCgbRmetIrFGRkZCoVBg+/btVsvYvn07BEFAhw4dHDLWDh064NKlS0hOTra0GY1G7Nq1Cw0bNoSfn181bpXjYF47VqzM6/JhXjtWrMzr8mFeO1aszOvyYV47VqzM6/JhXjtWrI6U1zJbB0B3l5CQgIKCAmi1WgDAxYsXsXfvXgBA27Zt4eLigiFDhmDSpEmYOXMmnn76aej1eixfvhyenp548sknHTJWlUqFQYMGYdmyZVCpVGjdujVOnz6NFStWoHfv3vdcobVVrL1798bmzZsxe/ZsjBgxAl5eXti8eTMuX76MmTNn3tM2ORLmNfPaGTGvmdfOiHnNvHZGzGvmtTNiXjOvbUUQRVG0dRB0Z6NHj0Z6enqZ0xYtWoSAgAAAwJkzZ7B06VKcPHkSUqkULVq0wAsvvICgoCCHjnXTpk3YvHkzrl27Bh8fH/Ts2RODBg2CTHZvdVFbxpqVlYUlS5YgPj4eOp0OderUwdChQ9GqVat72iZHwrxmXjsj5jXz2hkxr5nXzoh5zbx2Rsxr5rWtsLBERERERERERESVwjGWiIiIiIiIiIioUlhYIiIiIiIiIiKiSmFhiYiIiIiIiIiIKoWFJSIiIiIiIiIiqhQWloiIiIiIiIiIqFJYWCIiIiIiIiIiokphYYmIiIiIiIiIiCqFhSUiIiIiIiIiIqoUFpaIiIiIiIiIiKhSWFgiIiKiSlmyZAkEQYCrqytSU1NLTe/WrRuaNWtmg8iAXbt2QRAErF692ibrr6iUlBT07dsXvr6+EAQBEyZMsHVIVWbz5s348MMPbR0GERERVRMWloiIiOie6HQ6vPfee7YOw6G9+eab+Pfff/Hzzz9j//79ePPNN20dUpXZvHkzZsyYYeswiIiIqJqwsERERET35JFHHsHy5ctx5MgRW4dy32m1WoiieM/LOXbsGNq3b48nn3wSHTp0QFhYWLlfK4oitFrtPcdQ1TQaja1DICIiovuAhSUiIiK6J2+//Tb8/PwwZcqUO86XkpICQRCwZMmSUtMEQbC6XOrDDz+EIAj477//8Mwzz8DLywu+vr6YOHEiCgsLkZycjEceeQQqlQrh4eGYM2dOmessKCjAxIkTERgYCDc3N3Tt2hWHDh0qNV98fDyeeOIJ+Pr6wtXVFa1bt8bKlSut5im+9G/Lli0YNWoUatasCaVSCZ1Od9ttvnDhAoYNGwZ/f3+4uLggIiIC8+bNg8lkAnDzkr0zZ87gzz//hCAIEAQBKSkpt12mIAh49dVXsWDBAkRERMDFxQVLly4FAJw+fRpDhgyxWt///vc/q9cXrzMmJqZc783GjRsRFRUFpVIJlUqF3r17Y//+/VbzFP+9EhMT8fTTT8PHxwf16tXDyJEjLesv3raS27dq1So89NBD8PLyglKpRN26dTFq1KjbbjsRERHZHxaWiIiI6J6oVCq89957+Pvvv7Fjx44qXfagQYPQsmVLrFmzBmPGjMGXX36JN998E08++ST69u2LdevWoUePHpgyZQrWrl1b6vXvvvsuzp07hx9//BE//vgjrly5gm7duuHcuXOWeXbu3IlOnTohOzsbCxYswIYNG9CqVSsMHjy4zCLYqFGjIJfL8csvv2D16tWQy+Vlxn79+nV07NgRW7ZswcyZM7Fx40b06tULkydPxquvvgoAiIyMxP79+xEYGIhOnTph//792L9/P4KCgu74vqxfvx7z58/H9OnT8ffff6Nz5844ceIE2rVrh2PHjmHevHn4/fff0bdvX7z++utlXopWnvdm+fLl6N+/Pzw9PbFixQr89NNPyMrKQrdu3fDPP/+UWuZTTz2F+vXrY9WqVViwYAHef/99PP300wBg2bbi7du/fz8GDx6MunXr4tdff8Uff/yB6dOno7Cw8I7bTkRERHZGJCIiIqqExYsXiwDEgwcPijqdTqxbt67Ytm1b0WQyiaIoil27dhWbNm1qmf/8+fMiAHHx4sWllgVA/OCDDyzPP/jgAxGAOG/ePKv5WrVqJQIQ165da2kzGAxizZo1xaeeesrStnPnThGAGBkZaYlHFEUxJSVFlMvl4ujRoy1tjRs3Flu3bi0aDAardfXr108MCgoSjUaj1fYOHz68XO/PO++8IwIQ//33X6v28ePHi4IgiMnJyZa2sLAwsW/fvuVaLgDRy8tLzMzMtGp/+OGHxZCQEDEnJ8eq/dVXXxVdXV0t85f3vTEajWJwcLDYvHlzy3sgiqKoVqtFf39/sWPHjpa24r/X9OnTS8X7yiuviGUdcs6dO1cEIGZnZ5dru4mIiMg+sccSERER3TOFQoGPP/4Y8fHxpS4huxf9+vWzeh4REQFBEPDoo49a2mQyGerXr1/mnemGDBkCQRAsz8PCwtCxY0fs3LkTAHDmzBmcPHkSQ4cOBQAUFhZaHo899hjS0tKQnJxstcyBAweWK/YdO3agSZMmaN++vVX7yJEjIYriPfXu6tGjB3x8fCzPCwoKsH37dgwYMABKpbLUdhQUFCAuLs5qGXd7b5KTk3HlyhU8//zzkEhuHjJ6eHhg4MCBiIuLKzWOUnnfGwBo164dAHOvtJUrV+Ly5cvlfwOIiIjIbrCwRERERFXi2WefRWRkJKZNmwaDwVAly/T19bV6rlAooFQq4erqWqq9oKCg1OsDAwPLbLtx4wYA4Nq1awCAyZMnQy6XWz1efvllAEBGRobV6+92mVqxGzdulDlvcHCwZXpl3brcGzduoLCwEN9++22p7XjssccAlN6Ou703xT9vtw0mkwlZWVl3jOtOunTpgvXr16OwsBDDhw9HSEgImjVrhhUrVpR7GURERGR7MlsHQERERM5BEATMnj0bvXv3xg8//FBqenEx6NbBru+lwHI3V69eLbPNz88PAFCjRg0AwNSpU/HUU0+VuYxGjRpZPS/Zy+dO/Pz8kJaWVqr9ypUrVuuujFtj8PHxgVQqxfPPP49XXnmlzNfUqVPH6vnd3pvin7fbBolEYtVrqqy47qZ///7o378/dDod4uLiMGvWLAwZMgTh4eGIioqq0LKIiIjINthjiYiIiKpMr1690Lt3b3z00UfIy8uzmhYQEABXV1f8999/Vu0bNmyotnhWrFgBURQtz1NTU7Fv3z5069YNgLlo1KBBAxw5cgRt27Yt86FSqSq17p49e+LEiRNITEy0av+///s/CIKA7t27V3q7bqVUKtG9e3ccOnQILVq0KHM7igtFxcrz3tSqVQvLly+3mi8/Px9r1qyx3CnublxcXAAAWq32jvN07doVs2fPBoAy705HRERE9ok9loiIiKhKzZ49G23atEF6ejqaNm1qaRcEAcOGDcPPP/+MevXqoWXLljhw4ACWL19ebbGkp6djwIABGDNmDHJycvDBBx/A1dUVU6dOtcyzcOFCPProo3j44YcxcuRI1KpVC5mZmUhKSkJiYiJWrVpVqXW/+eab+L//+z/07dsXH330EcLCwvDHH3/g+++/x/jx49GwYcOq2kwAwNdff43o6Gh07twZ48ePR3h4ONRqNc6cOYNNmzaVGtPpbu+NRCLBnDlzMHToUPTr1w/jxo2DTqfD559/juzsbHz22Wfliqt58+YAzHnx6KOPQiqVokWLFvj4449x6dIl9OzZEyEhIcjOzsbXX38NuVyOrl27Vul7Q0RERNWHhSUiIiKqUq1bt8Zzzz1XZsFo3rx5AIA5c+YgLy8PPXr0wO+//47w8PBqieXTTz/FwYMH8cILLyA3Nxft27fHr7/+inr16lnm6d69Ow4cOIBPPvkEEyZMQFZWFvz8/NCkSRMMGjSo0uuuWbMm9u3bh6lTp2Lq1KnIzc1F3bp1MWfOHEycOLEqNs9KkyZNkJiYiJkzZ+K9995Deno6vL290aBBA8s4SyWV570ZMmQI3N3dMWvWLAwePBhSqRQdOnTAzp070bFjx3LFNWTIEOzduxfff/89PvroI4iiiPPnz+Ohhx5CfHw8pkyZguvXr8Pb2xtt27bFjh07rAqSREREZN8EsWTfZiIiIiJyart27UL37t2xatUqPP3007YOh4iIiBwcx1giIiIiIiIiIqJKYWGJiIiIiIiIiIgqhZfCERERERERERFRpbDHEhERERERERERVQoLS0REREREREREVCksLBERERERERERUaWwsERERERERERERJXCwhIREREREREREVUKC0tERERERERERFQpLCwREREREREREVGlsLBERERERERERESV8v9Tfd9CWwhpgwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1400x300 with 12 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<ggplot: (314657497)>"
      ]
     },
     "execution_count": 243,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dummy_df = mean_cdf_df.loc[:1, :].copy()\n",
    "dummy_df.loc[0, 'dummy'] = '0'\n",
    "dummy_df.loc[1, 'dummy'] = '1'\n",
    "\n",
    "plt = ggplot(median_cdf_df) +\\\n",
    "    geom_line(aes(x='mean', y='cdf'), data=mean_cdf_df, color='black') +\\\n",
    "    geom_line(aes(x='peak', y='cdf'), data=peak_cdf_df, color='#0280c9') +\\\n",
    "    geom_line(aes(x='mean', y='cdf', color='dummy'), data=dummy_df) +\\\n",
    "    scale_x_log10() +\\\n",
    "    scale_y_continuous(limits=[0, 1]) +\\\n",
    "    facet_wrap(facets='vendor_cat', nrow=2) +\\\n",
    "    ylab('ECDF') +\\\n",
    "    xlab('Number of reports') +\\\n",
    "    scale_color_manual(['black', '#0280c9'], labels=['Mean number of reports of failure', 'Peak number of reports of failure']) +\\\n",
    "    guides(color=guide_legend(title='')) +\\\n",
    "    theme_classic(base_size=12, base_family='sans-serif') +\\\n",
    "    theme(figure_size=(14, 3),\n",
    "          axis_text_y=element_text(margin={'r': 5}),\n",
    "          panel_grid_major_x=element_line(size=0.7, color=\"gainsboro\"),\n",
    "          text=element_text(size=12),\n",
    "          legend_box_spacing=0.01,\n",
    "          legend_box_margin=0,\n",
    "          legend_margin=0,\n",
    "          legend_key=element_blank(),\n",
    "          legend_entry_spacing=5,\n",
    "          legend_background=element_rect(fill=(0,0,0,0), color=(0,0,0,0)),\n",
    "          legend_position='top')\n",
    "\n",
    "plt.save('plots/failure_mean_peak_cdf.pdf', limitsize=None)\n",
    "plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 246,
   "metadata": {},
   "outputs": [],
   "source": [
    "filtered_failure_df['parsed_start_time'] = pd.to_datetime(filtered_failure_df['start_time'], unit='s')\n",
    "precovid_failure_df = filtered_failure_df[(filtered_failure_df['parsed_start_time'].dt.year == 2019)\n",
    "                                 & (filtered_failure_df['parsed_start_time'].dt.month >= 4)\n",
    "                                 & (filtered_failure_df['parsed_start_time'].dt.month <= 8)].reset_index(drop=True)\n",
    "covid_failure_df = filtered_failure_df[(filtered_failure_df['parsed_start_time'].dt.year == 2020)\n",
    "                                 & (filtered_failure_df['parsed_start_time'].dt.month >= 4)\n",
    "                                 & (filtered_failure_df['parsed_start_time'].dt.month <= 8)].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_ia(partial_df):\n",
    "    sorted_partial_df = partial_df.sort_values('start_time').reset_index(drop=True)\n",
    "    ia_arr = sorted_partial_df['start_time'].values - np.roll(sorted_partial_df['start_time'].values, 1)\n",
    "    partial_ia_df = pd.DataFrame({'ia': ia_arr[1:], 'count':1, 'vendor_cat': partial_df.loc[0, 'vendor_cat']})\n",
    "    return partial_ia_df\n",
    "\n",
    "precovid_ia_partial_df_list = []\n",
    "covid_ia_partial_df_list = []\n",
    "\n",
    "for vendor in precovid_failure_df['vendor_cat'].unique():\n",
    "    precovid_partial_df = precovid_failure_df[precovid_failure_df['vendor_cat'] == vendor].reset_index(drop=True)\n",
    "    covid_partial_df = covid_failure_df[covid_failure_df['vendor_cat'] == vendor].reset_index(drop=True)\n",
    "    \n",
    "    # Compute interarrival time\n",
    "    precovid_ia = compute_ia(precovid_partial_df)\n",
    "    covid_ia = compute_ia(covid_partial_df)\n",
    "    \n",
    "    precovid_ia_partial_df_list.append(precovid_ia)\n",
    "    covid_ia_partial_df_list.append(covid_ia)\n",
    "\n",
    "precovid_ia_df = pd.concat(precovid_ia_partial_df_list)\n",
    "covid_ia_df = pd.concat(covid_ia_partial_df_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 249,
   "metadata": {},
   "outputs": [],
   "source": [
    "precovid_ia_quantiles = precovid_ia_df.groupby('vendor_cat')['ia'].quantile([0.50, 0.95]).reset_index().sort_values('ia').pivot(index='vendor_cat', columns='level_1')\n",
    "covid_ia_quantiles = covid_ia_df.groupby('vendor_cat')['ia'].quantile([0.50, 0.95]).reset_index().sort_values('ia').pivot(index='vendor_cat', columns='level_1')\n",
    "# precovid_ia_quantiles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 250,
   "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 tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">duration</th>\n",
       "      <th colspan=\"2\" halign=\"left\">ia</th>\n",
       "      <th colspan=\"2\" halign=\"left\">duration_covid</th>\n",
       "      <th colspan=\"2\" halign=\"left\">ia_covid</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>level_1</th>\n",
       "      <th>0.50</th>\n",
       "      <th>0.95</th>\n",
       "      <th>0.50</th>\n",
       "      <th>0.95</th>\n",
       "      <th>0.50</th>\n",
       "      <th>0.95</th>\n",
       "      <th>0.50</th>\n",
       "      <th>0.95</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>vendor_cat</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Apple</th>\n",
       "      <td>0.333333</td>\n",
       "      <td>2.500000</td>\n",
       "      <td>98.000000</td>\n",
       "      <td>560.033333</td>\n",
       "      <td>0.500000</td>\n",
       "      <td>0.650000</td>\n",
       "      <td>899.666667</td>\n",
       "      <td>899.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Skype</th>\n",
       "      <td>0.333333</td>\n",
       "      <td>4.633333</td>\n",
       "      <td>99.333333</td>\n",
       "      <td>416.800000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>1.483333</td>\n",
       "      <td>178.666667</td>\n",
       "      <td>1070.333333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Github</th>\n",
       "      <td>0.333333</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>60.666667</td>\n",
       "      <td>308.600000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.666667</td>\n",
       "      <td>158.333333</td>\n",
       "      <td>646.066667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Gmail</th>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>62.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>26.666667</td>\n",
       "      <td>583.733333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FB Msgr</th>\n",
       "      <td>0.333333</td>\n",
       "      <td>1.200000</td>\n",
       "      <td>5.666667</td>\n",
       "      <td>186.200000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.533333</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>221.350000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Whatsapp</th>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>4.333333</td>\n",
       "      <td>26.200000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>11.166667</td>\n",
       "      <td>376.150000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Snapchat</th>\n",
       "      <td>0.333333</td>\n",
       "      <td>1.333333</td>\n",
       "      <td>1.333333</td>\n",
       "      <td>6.666667</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>1.333333</td>\n",
       "      <td>1.666667</td>\n",
       "      <td>126.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Netflix</th>\n",
       "      <td>0.333333</td>\n",
       "      <td>3.416667</td>\n",
       "      <td>1.333333</td>\n",
       "      <td>8.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>2.766667</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>65.916667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Facebook</th>\n",
       "      <td>0.333333</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.333333</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>1.266667</td>\n",
       "      <td>1.666667</td>\n",
       "      <td>69.566667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Twitter</th>\n",
       "      <td>0.666667</td>\n",
       "      <td>5.666667</td>\n",
       "      <td>1.666667</td>\n",
       "      <td>10.666667</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>7.866667</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>92.550000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>YouTube</th>\n",
       "      <td>0.333333</td>\n",
       "      <td>2.333333</td>\n",
       "      <td>1.333333</td>\n",
       "      <td>6.333333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>1.666667</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>44.700000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Instagram</th>\n",
       "      <td>0.333333</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>8.633333</td>\n",
       "      <td>0.333333</td>\n",
       "      <td>3.333333</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>29.733333</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            duration                   ia             duration_covid  \\\n",
       "level_1         0.50      0.95       0.50        0.95           0.50   \n",
       "vendor_cat                                                             \n",
       "Apple       0.333333  2.500000  98.000000  560.033333       0.500000   \n",
       "Skype       0.333333  4.633333  99.333333  416.800000       0.333333   \n",
       "Github      0.333333  1.000000  60.666667  308.600000       1.000000   \n",
       "Gmail       0.333333  0.666667   5.000000   62.333333       0.333333   \n",
       "FB Msgr     0.333333  1.200000   5.666667  186.200000       0.333333   \n",
       "Whatsapp    0.333333  0.666667   4.333333   26.200000       0.333333   \n",
       "Snapchat    0.333333  1.333333   1.333333    6.666667       0.333333   \n",
       "Netflix     0.333333  3.416667   1.333333    8.333333       0.333333   \n",
       "Facebook    0.333333  2.000000   1.333333    5.000000       0.333333   \n",
       "Twitter     0.666667  5.666667   1.666667   10.666667       0.333333   \n",
       "YouTube     0.333333  2.333333   1.333333    6.333333       0.333333   \n",
       "Instagram   0.333333  6.000000   2.000000    8.633333       0.333333   \n",
       "\n",
       "                        ia_covid               \n",
       "level_1         0.95        0.50         0.95  \n",
       "vendor_cat                                     \n",
       "Apple       0.650000  899.666667   899.666667  \n",
       "Skype       1.483333  178.666667  1070.333333  \n",
       "Github      1.666667  158.333333   646.066667  \n",
       "Gmail       2.000000   26.666667   583.733333  \n",
       "FB Msgr     0.533333   11.000000   221.350000  \n",
       "Whatsapp    0.666667   11.166667   376.150000  \n",
       "Snapchat    1.333333    1.666667   126.500000  \n",
       "Netflix     2.766667    2.000000    65.916667  \n",
       "Facebook    1.266667    1.666667    69.566667  \n",
       "Twitter     7.866667    5.000000    92.550000  \n",
       "YouTube     1.666667    2.000000    44.700000  \n",
       "Instagram   3.333333    3.000000    29.733333  "
      ]
     },
     "execution_count": 250,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "precovid_duration_quantiles = precovid_failure_df.groupby('vendor_cat')['duration'].quantile([0.50, 0.95]).reset_index().sort_values('duration').pivot(index='vendor_cat', columns='level_1')\n",
    "covid_duration_quantiles = covid_failure_df.groupby('vendor_cat')['duration'].quantile([0.50, 0.95]).reset_index().sort_values('duration').pivot(index='vendor_cat', columns='level_1')\n",
    "covid_table = precovid_duration_quantiles.join(precovid_ia_quantiles).join(covid_duration_quantiles, rsuffix='_covid').join(covid_ia_quantiles, rsuffix='_covid') / 3600\n",
    "covid_table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 251,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\begin{tabular}{lrrrrrrrr}\n",
      "\\toprule\n",
      "{} & \\multicolumn{2}{l}{duration} & \\multicolumn{2}{l}{ia} & \\multicolumn{2}{l}{duration\\_covid} & \\multicolumn{2}{l}{ia\\_covid} \\\\\n",
      "level\\_1 &     0.50 & 0.95 & 0.50 &  0.95 &           0.50 & 0.95 &     0.50 &   0.95 \\\\\n",
      "vendor\\_cat &          &      &      &       &                &      &          &        \\\\\n",
      "\\midrule\n",
      "Apple      &      0.3 &  2.5 & 98.0 & 560.0 &            0.5 &  0.7 &    899.7 &  899.7 \\\\\n",
      "Skype      &      0.3 &  4.6 & 99.3 & 416.8 &            0.3 &  1.5 &    178.7 & 1070.3 \\\\\n",
      "Github     &      0.3 &  1.0 & 60.7 & 308.6 &            1.0 &  1.7 &    158.3 &  646.1 \\\\\n",
      "Gmail      &      0.3 &  0.7 &  5.0 &  62.3 &            0.3 &  2.0 &     26.7 &  583.7 \\\\\n",
      "FB Msgr    &      0.3 &  1.2 &  5.7 & 186.2 &            0.3 &  0.5 &     11.0 &  221.3 \\\\\n",
      "Whatsapp   &      0.3 &  0.7 &  4.3 &  26.2 &            0.3 &  0.7 &     11.2 &  376.1 \\\\\n",
      "Snapchat   &      0.3 &  1.3 &  1.3 &   6.7 &            0.3 &  1.3 &      1.7 &  126.5 \\\\\n",
      "Netflix    &      0.3 &  3.4 &  1.3 &   8.3 &            0.3 &  2.8 &      2.0 &   65.9 \\\\\n",
      "Facebook   &      0.3 &  2.0 &  1.3 &   5.0 &            0.3 &  1.3 &      1.7 &   69.6 \\\\\n",
      "Twitter    &      0.7 &  5.7 &  1.7 &  10.7 &            0.3 &  7.9 &      5.0 &   92.5 \\\\\n",
      "YouTube    &      0.3 &  2.3 &  1.3 &   6.3 &            0.3 &  1.7 &      2.0 &   44.7 \\\\\n",
      "Instagram  &      0.3 &  6.0 &  2.0 &   8.6 &            0.3 &  3.3 &      3.0 &   29.7 \\\\\n",
      "\\bottomrule\n",
      "\\end{tabular}\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(covid_table.to_latex(float_format=\"%.1f\"))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}