data/class_distribution_and_splits.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Copyright (c) Microsoft Corporation. All rights reserved.\n",
    "# Licensed under the MIT License."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The autoreload extension is already loaded. To reload it, use:\n",
      "  %reload_ext autoreload\n"
     ]
    }
   ],
   "source": [
    "from IPython.core.interactiveshell import InteractiveShell\n",
    "InteractiveShell.ast_node_interactivity = 'all' # default is ‘last_expr’\n",
    "\n",
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import os\n",
    "from random import shuffle\n",
    "import math\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "from tqdm import tqdm\n",
    "from collections import Counter, defaultdict\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Class distribution and splitting into train/val/test\n",
    "\n",
    "This notebook tallies the distribution of the damage and background classes."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Look at a label JSON\n",
    "\n",
    "Examine one JSON file in the `labels` folder..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_path = './nlrc-damage-assessment/public_datasets/xBD/raw/train/labels/hurricane-matthew_00000000_post_disaster.json'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(label_path) as f:\n",
    "    label_json = json.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['features', 'metadata'])"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "label_json.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['lng_lat', 'xy'])"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "label_json['features'].keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['sensor', 'provider_asset_type', 'gsd', 'capture_date', 'off_nadir_angle', 'pan_resolution', 'sun_azimuth', 'sun_elevation', 'target_azimuth', 'disaster', 'disaster_type', 'catalog_id', 'original_width', 'original_height', 'width', 'height', 'id', 'img_name'])"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "label_json['metadata'].keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'properties': {'feature_type': 'building',\n",
       "  'subtype': 'minor-damage',\n",
       "  'uid': '6dd06e25-353d-4ca2-8a93-bbdd44b10435'},\n",
       " 'wkt': 'POLYGON ((184.3848180327538 1022.977380861319, 184.4809847188476 1015.871163903935, 203.263854348678 1016.01783797561, 203.5772080801356 1022.977565618902, 184.3848180327538 1022.977380861319))'}"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "label_json['features']['xy'][0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "feat_shape = wkt.loads(label_json['features']['xy'][0]['wkt'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "133.53150074518555"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feat_shape.area"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "25928.990821215957"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "total_area = 0\n",
    "\n",
    "for feat in label_json['features']['xy']:\n",
    "    feat_shape = wkt.loads(feat['wkt'])\n",
    "    total_area += feat_shape.area\n",
    "\n",
    "total_area  # this looks about right from visualizing this tile's mask - it's < 200*200 squared pixels"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Aggregate building counts and area\n",
    "\n",
    "Example file names\n",
    "```\n",
    "joplin-tornado_00000004_post_disaster.json\n",
    "joplin-tornado_00000004_pre_disaster.json\n",
    "```\n",
    "\n",
    "`subtypes` are:\n",
    "```\n",
    "{'destroyed', 'major-damage', 'minor-damage', 'no-damage', 'un-classified'}\n",
    "```\n",
    "\n",
    "There are some features without a `sub-types` field. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_dirs = [\n",
    "    './nlrc-damage-assessment/public_datasets/xBD/raw/train/labels',\n",
    "    './nlrc-damage-assessment/public_datasets/xBD/raw/test/labels',\n",
    "    './nlrc-damage-assessment/public_datasets/xBD/raw/hold/labels',\n",
    "    './nlrc-damage-assessment/public_datasets/xBD/raw_tier3/labels'\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 5598/5598 [02:18<00:00, 40.44it/s]\n",
      "100%|██████████| 1866/1866 [00:48<00:00, 38.17it/s]\n",
      "100%|██████████| 1866/1866 [00:49<00:00, 37.48it/s]\n",
      "100%|██████████| 12738/12738 [08:19<00:00, 25.52it/s]\n"
     ]
    }
   ],
   "source": [
    "# area breakdown by disaster\n",
    "pre_areas = defaultdict(lambda: defaultdict(float))\n",
    "post_areas = defaultdict(lambda: defaultdict(float))\n",
    "\n",
    "# count of building features\n",
    "pre_counts = defaultdict(lambda: defaultdict(int))\n",
    "post_counts = defaultdict(lambda: defaultdict(int))\n",
    "\n",
    "feature_types = set()\n",
    "\n",
    "# building features without a subtype attribute\n",
    "buildings_no_type = []\n",
    "\n",
    "for label_dir in label_dirs:\n",
    "    for p in tqdm(os.listdir(label_dir)):\n",
    "        \n",
    "        if not p.endswith('.json'):\n",
    "            continue\n",
    "            \n",
    "        areas_dict = pre_areas if p.endswith('pre_disaster.json') else post_areas\n",
    "        counts_dict = pre_counts if p.endswith('pre_disaster.json') else post_counts\n",
    "        \n",
    "        with open(os.path.join(label_dir, p)) as f:\n",
    "            label_json = json.load(f)\n",
    "            \n",
    "            disaster_name = label_json['metadata']['disaster']\n",
    "            \n",
    "            for feat in label_json['features']['xy']:\n",
    "                \n",
    "                if feat['properties']['feature_type'] != 'building':\n",
    "                    feature_types.add(feat['properties']['feature_type'])\n",
    "                    continue\n",
    "                \n",
    "                if 'subtype' in feat['properties']:\n",
    "                    damage_class = feat['properties']['subtype']\n",
    "                else:\n",
    "                    damage_class = 'no-subtype'  # usually for pre images\n",
    "                    buildings_no_type.append((p, feat))    \n",
    "                    \n",
    "                feat_shape = wkt.loads(feat['wkt'])\n",
    "                \n",
    "                areas_dict[disaster_name][damage_class] += feat_shape.area\n",
    "                \n",
    "                counts_dict[disaster_name][damage_class] += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "425368"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(buildings_no_type)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "19"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(post_areas)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['guatemala-volcano', 'hurricane-florence', 'hurricane-harvey', 'hurricane-matthew', 'hurricane-michael', 'mexico-earthquake', 'midwest-flooding', 'palu-tsunami', 'santa-rosa-wildfire', 'socal-fire', 'joplin-tornado', 'lower-puna-volcano', 'moore-tornado', 'nepal-flooding', 'pinery-bushfire', 'portugal-wildfire', 'sunda-tsunami', 'tuscaloosa-tornado', 'woolsey-fire'])"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "post_areas.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "areas_and_counts_by_disaster = {\n",
    "    'pre_areas': pre_areas,\n",
    "    'post_areas': post_areas,\n",
    "    'pre_counts': pre_counts,\n",
    "    'post_counts': post_counts\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('./data/areas_and_counts_by_disaster.json', 'w') as f:\n",
    "    json.dump(areas_and_counts_by_disaster, f, indent=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "wind_disasters = ['hurricane-matthew', 'tuscaloosa-tornado', 'hurricane-michael', 'moore-tornado', 'joplin-tornado']\n",
    "\n",
    "damage_classes = ['no-damage', 'minor-damage', 'major-damage', 'destroyed', 'un-classified']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# store data in dic as {disaster_name: [counts]}\n",
    "wind_post_counts = defaultdict(list)  \n",
    "wind_post_areas = defaultdict(list) \n",
    "\n",
    "for disaster in wind_disasters:\n",
    "    for damage_class in damage_classes:\n",
    "        wind_post_counts[disaster].append(post_counts[disaster][damage_class])\n",
    "        wind_post_areas[disaster].append(post_areas[disaster][damage_class])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "defaultdict(int,\n",
       "            {'no-damage': 19453,\n",
       "             'un-classified': 586,\n",
       "             'destroyed': 1584,\n",
       "             'minor-damage': 886,\n",
       "             'major-damage': 449})"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "post_counts['moore-tornado']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "defaultdict(list,\n",
       "            {'hurricane-matthew': [4058, 12331, 2717, 3524, 1334],\n",
       "             'tuscaloosa-tornado': [10499, 2036, 466, 1097, 908],\n",
       "             'hurricane-michael': [22692, 8292, 2919, 1225, 373],\n",
       "             'moore-tornado': [19453, 886, 449, 1584, 586],\n",
       "             'joplin-tornado': [8225, 2192, 1005, 3274, 656]})"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wind_post_counts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x900 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Reference: https://matplotlib.org/3.3.1/gallery/lines_bars_and_markers/barchart.html\n",
    "num_classes = len(damage_classes)\n",
    "\n",
    "x = np.arange(num_classes)  # the label locations\n",
    "width = 0.8/num_classes  # the width of the bars\n",
    "\n",
    "fig, ax = plt.subplots(1, 1, figsize=(12, 6), dpi=150)\n",
    "\n",
    "for i, (disaster, li) in enumerate(wind_post_counts.items()):\n",
    "    rect = ax.bar(x - (i - 2) * width, li, width, label=disaster)\n",
    "\n",
    "_ = ax.set_ylabel('Building counts')\n",
    "_ = ax.set_title('Class distribution by building counts')\n",
    "_ = ax.set_xticks(x)\n",
    "x_labels = ax.set_xticklabels(damage_classes, rotation=45)\n",
    "_ = ax.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABd0AAANsCAYAAACu7QEVAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAABcSAAAXEgFnn9JSAADfpklEQVR4nOzdd5iU1fnw8e+h9yKIgChFBBQBNYCg0qTYqVaIlKjRn92XRLEGLFHUiBpLohixKzbssYZiQbEGQYwgoGKlCShSz/vHzCzb2TKwu/j9XNdes/M8zznnPjOzI95z5j4hxogkSZIkSZIkSSq+ciUdgCRJkiRJkiRJOwqT7pIkSZIkSZIkpYlJd0mSJEmSJEmS0sSkuyRJkiRJkiRJaWLSXZIkSZIkSZKkNDHpLkmSJEmSJElSmph0lyRJkiRJkiQpTUy6S5IkSZIkSZKUJibdJUmSJEmSJElKE5PukiRJkiRJkiSliUl3SZIkSZIkSZLSxKS7JEmSJEmSJElpYtJdkiRJkiRJkqQ0MekuSZIkSZIkSVKamHSXJEkqohBCzxBCDCHEko6lJIUQpiYfh7G5nFuUPDdy+0e2dSGEScn4JuVyLs95lRYhhGap12AIoVlJx1NQpTnuEMLYZFxTS2j8/F6TeZ5LQ9+l/vUuSZJUVlQo6QAkSZJKWgihPDAEOAroAjQAqgErgf8BM4AHY4yflFSMSgghnAfUAabEGD8q0WC2sUzJz0kxxkUlGIokSZKkQjDpLkmSftNCCF2Ae4FWmQ5vAFYD9YCDkj9jQghPAifGGNdv90DLrgXAr8BPaervPKApsAj4KA39fQt8lrwtbf6SvJ1KYr652UAi/tTvKvtK6jX5ZXLcpdt5XEmSpB2OSXdJkvSbFUI4GngMqAwsA24Anogxfp48Xx7Yj8Qq+DOAwSRWwJt0L6AYY++SjiE/McaLgItKOo6iijEuAdqUdBxKn5J6TcYYh2/vMSVJknZUJt0lSdJvUghhT+ABEgn3ucChMcavM18TY9wEvAe8F0K4HvjXdg9UkiRJklSmuJGqJEn6rboKqEWi9Mmg7An37GKMy2OMAylgmZQQQrkQwkEhhGtDCDNDCF+HENaHEJaFEKaFEE4PIVTMp33dEMIVIYQPQgirkm2/CyH8N4TwjxBCjhXkIYSqIYQ/hRDeDiGsCCFsCCH8GEKYG0K4N4QwpCCx59Jv+RDCWclYfg4hLE9uunhMAdrmuZFqYeJNbW5JorQMwD2ZNuLMsplt9k06Qwh7hBDuDCEsDCGsCyEsynRtgTamDCFUCiGMST7+PyfjfSWEcHg+bVIx9MznmhybV6ZiynTZf7LNNXP8W92QNIRQO4RweabX0toQwuchhDtCCC0KEn8IoWYI4aoQwrxk+2UhhOdCCAfk1b4wQgh7Juf9dfI5+jL5Ot81l2uvTcY1Zyt91gohrMnr9VeI2I5L/s0uTz737yf/Hsrncf1WNyQN+WzWWtDXZD59DwshvBlCWB1C+CmE8E4I4Y8hhLCVdgXaEDn5t/DnEMLHycfjpxDC6yGEw7bSf/UQwrgQwqfJ19APIYQXQvK9LBRj0+UQQutkTK+GEBYk+18VQvgw+bqtn0/bzHOrERLvu7OTj1+Ov6sQwn4hhH8lx/kl+Rr7OL9xQggVQwh9Qwi3hBDeCyF8GxLv6T+EEF4KIZy4tedHkiSVLa50lyRJvzkhhF2AVML4wRjj/wraNsYYt34VALsDb2S6vxH4BdgJ6J78GRpCODTGuDZbfE2AN5N9AGwmkeyvD+wCtCNRUuS1TG1qktjwtUMq1GSbOsl2ewE9gCcKGH+q38rA08ChmWJZn4y/RwhhfGH6K0a8a4DvgZ1JLBxZBWR53PJwIPBPoAaJx78odc8rAa8C3Ug8j2uScfYB+oQQxsUYxxah37z8RGKuuyTvryBrSaMfC9pRCKEt8G+gSfLQryQeg5bJn1EhhGExxvxeF42AD5LX/0riNbATcCTQL4RwdIzxpYLGlIsDgLuAmiQe203AbsBpwLEhhL4xxg8yXf9P4AJg7xDCwTHGN7J3mDQUqE7i8ZxclMCSr+8LSLw+VwJVgP2TP0eFEAbEGNcVpe90SyZt7wZGJQ+lYu4IdAZ6AcWNtQYwncRztiHZX61k3z1DCKfEGHN8IyiE0AD4D7B38tAGoCJwOHBYCOGMYsb1Els+kEu9l9QG9k3+jAwh9I4xfpZr64R6wPsk9vdYT+L9Ivs8xgGXAakE+S/JebRP/vwhhHBkjPHDbE0PAl7OdH9d8mdnoF/yZ1AI4YQY4+YCzFeSJJVyrnSXJEm/Rb3Y8u+gp7bRGBtJJKuPB3YFKscYa5NILI4CviGRxL06l7ZjSSTcF5FI7FaKMe5EohROM+D/gJnZ2pxLIoG9nEQN+qoxxrrJNrsCw8ma9Cmoa0gk3CNwKVA32W9D4A7gQhJJrcIqVLwxxhtijA2Br1LtY4wNM//kMc4/gTlApxhj9RhjDRIJrsI4g0TS8nSgZjLO3YHHk+f/EkLoX8g+8xRjPDfbfAZnm2ungvST/GDjWRIJ9yUkkuTVY4y1SDxnM0k83g+GEDrk1Q9wG4kk5CEkktg1SDwen5FIOP4zhFCc/6/4J7AQOCDGWDM5xqEkNvbcCXgqORcAYowLSSRZAU7Np9/UuftjjDkSqAWwL4mE+63ALsm/wbokkq4xGeM1Reh3WzmbLQn3W4EGyZh3IvGecjwwoJhjXEHi9TSQxGupJokPAGeSSETfHEKonUu7e0kk3NcCJ5P172gycDOJBHRRzSQx/5ZAlWTfVUi8f75L4j3loa30MZbEBwiDgRrJPnYDfgAIIZwHXE7ig6GLgEYxxuok9vnoCLxO4gOqZ0IINbL1vTY5/pEk3jurJh+7eiTeC1cBxwJnFWn2kiSp1DHpLkmSfovaZvo9+4rEtIgxfh1jHBhjnBxj/Ca1ejHGuCbGOIktya8/hhCqZGt+YPL24hjja8na8sQYN8UYF8cY/xFjHJNHmxtijE+mVt/GGDcnx78/xvjHwswhhNCYRCIL4KoY49UxxlXJfn+IMZ4BPExiRWlhpT3ePCwD+sQY30sdKMw3G5JqA2fEGP8ZY/w12cdXJJKY05PXlKbka8oZQHMSq4oPizG+kOl1+DGJDx8WkUi85/bhT8pGoFeM8T/J5yfGGGeRSBJCYoVx12LEuRHoG2N8NxlbjDG+DBxGItm/O4kPPDL7R/L22BBCnewdhhBSq9EB7ixiXLVJJOzPjjH+mIxtVYzxKrY8Xmcn/05KVPI95C/Ju6mYlwLEGH+KMY4DxpP4hkZxVCPx9/R0jHFDsv/PgP4kvgVRAzgqW2wHk3guAf4YY/xXpr/3r0h8I+HNZN9FEmM8IcZ4a4xxQYxxffLY+hjja0BvEt8c2T8ZS16qAkfEGJ/KNLevY4y/JMvGXE3iw5ZBMcZrY4zfJa/ZFGN8n8SHMO+T+FDilGzxvRNjHJb8G/w+xsQ3pmKibNktJD6IADinqI+BJEkqXUy6l4AQwu9CoibokyGEJclagb+msf+eIVuN0zx+Lk/XmJIklTH1Mv2+vCQCSCaBfyCxqnffbKdXJm8bFaLLorTZmmNIlCNcC9yQxzVji9j3yuRtOuPNza0xxjXF7OMr4J7sB5MJ7KuSd/cOIbQr5jjpdnzy9vEY4yfZT8YYVwPXJe8enscKZYA7Y4w/5NJ+NokV6pAorVFU/8ij/0/Z8m2CE7Kdfo7E81IVOCmXPlMf2LydjLOorsjj+PUk/i4qkPimRknrR2JFO+Qd87UkEuPF8XiMcV72g8kPJd5O3s3+Wkh9OLMIeDCXtpn/jtIu+fc/LXk3v6T7v2POsjApw0h8KPBeMpGf2zgbSXwICVvKcRXU88nbPUII2/o9UZIkbQfWdC8Zl1H8r3bm5zsSX+HMTXng98nfZ2zDGCRJKs22y4Z1IYRKwB9IlCvYh0RSrHIulzbJdv85EiuHrw0htAGeBN5KrTLPw3PAicBZIYSdgUeBN1KrXYuoY/L2vbzGjjH+L4SwhET5hsLYFvHm5s009DE1tTI1F9NJrNSuQOLxKk6CN22Sr71U8vPVfC59JXlbjsTK8P/kcs07+bT/hsRq+p3yuWZrXt/KuaFA+xBCxUwrkDeFECYC40iUkfl7qkEIoTqJ1xYUfZU7wFcxxvm5nYgxrgohvE8iidsxt2u2s1QM+cX8UzLmg4oxztZeC5DztZD6xsH0fP6O3mTL31GRhBCOIvEBTCcS+yHktnI++3tt9hjykkrW7xNC+C6f66omb5tmP5EskXQ6iW8C7EXiWwe5baa9K/BtPmNIkqQywKR7yXgb+BiYlfzJ7x9uhZZcfTIyt3MhhMNJJN2/YsuKD0mSfmsyJ3Z3YkuyKG2SGwe+SmLT05Rfk2NvSt5PbQpaPVvz60nUOz+ORELxVCCGEOaQ2BTzruwlUmKMD4UQOpMoB3NC8ocQwnwStdH/lSyBUBgNkrdLtnLd1xQy6b6N4s1NjhXURZDn/GOM60IIy0gk+RrkdV0J2InEYgvI//n7OtPvecW/Op/2G5O3uSUPCyq/+FLnKpCY0/eZzk0ksZilXQihS4wxtc/BCSRqc6+kiBuoFiCuzOdLw/NemL/V4ijKayFVqz3P99nk39FSEvXOCyW5n8ADbPmgJRVL5g2Ia5Oo8Z79vTaz/N4rUiWEqrIlsZ6fLAn/EEIrEhtfZ076/0LiNZraODW1cXJ+MUqSpDLC8jIlIMY4Psb4lxjjczHG77feIq1Sq9wfTNX0lCTpN2hOpt/320ZjTCCRcF9GYrV7oxhj1RjjznHLxp+pJFSWlfcxxg0xxuNJlJ25gsRq319IrJb/EzA3hDA6+4AxxvOA1sDFwIskEjotSdT2fi+EcFMR55LX6tRi2YbxZrZp65ds1TaZ/3aUX/wxj9+3pyKNG2P8BngmeTdz/f/UBqoPxKJtoFqsuEpYaYw59f62tdiK+g2kk0kk3DeReL/ck8TG1Ttleq9NlSnKb4z83itSH2D9I8YYCvDTLFv7e0gk3BeRKLdTLyY2dm6QjC/zh5bb5ZtYkiRp2zLpXkaEEGqEEC4PIcwOIfwSQlgVQpgWQhhYiD6qs6WszQPbJFBJksqG/7BldeGgdHceQqhIoqQMwFkxxntSm+5luqY8UD+/fmKMHyc/qO9NohRBHxLlTMoD14cQOuTSZn6M8ZoY4xEkatd3BaYkT58bQuhfiKmkVn7mV5IBCl9aJkOa491W8px/CKEyW/YIyL5SNpXEy75RbmZF2YS2IJZnGn+3fK7LfO7HbRTL1uT3+kq9tjaS+/4LqQ1Vjw8h1ErW1T8geaw4pWW2Flfm2LI/76kV39vzed/mf6vFkIotzw1ns/0dFVaq3v/E5Pvl/FwWFxV6BX02qffvQu/bEELYjS0bR58YY3w8xpj9tVzc+CRJUilj0r0MCCHsQqJ+4jigLonam+8AvwOeCiGMKWBXg0l8XfHDGOOcrV0sSdKOKvlNsyeSd4cmv/pfICGEgqxC3JktCbe8NuY7mPyTclnEGDcmN/A7ElhHYjVkn6202ZwsuXEM8GXycN+Cjgm8l7ztmKxHnEMIYU+2nugrkALGm0qmbc/VoD3yed67saVk43vZzq1I3uaa9E4+pnvlM25qZXCh5xpjXA/8N3m3dz6Xpl5Dm4EPCjtOmvQqwLn/puq5Z/MqMJ9EOY9hbFnlXtwNVAF2CyHskduJ5HP3u+TdQj3vSQfkc64oUjHkF3MttsS8PaVeVz3yueYgil76NPU45/peG0KoQfEf71S99y4hhBz12rci8+sgr/8e5PteLkmSyh6T7mXDPcDewHVA8xjjgBhjXxKbYy0ArgohtM+vg6RUaZn7t02YkiSVKZcCa0jU530yhJDvCtAQQt0QwhMUbIXqKrYkTHOsRg8hVACuzmes3DZbTVnHlhXMGeUQ8msTY9zEltrGhSm38gSJVbtVgRzlbJIuL0R/GYoRb2pD1zpFGbeIdgdGZD+YrCV9cfLup7kkeT9O3g7Jo98/kfvGuinFnesjydtjQgj7ZD+ZTEZekLz7QozxpyKOU1ynhxByfOsjhNCaxAcwkNhoN4fkxpz/TN49gy3/3i3uKveUy/I4PprE38VGEhsdZ5Z63g9NftM0ixDCISS+0ZFOr7Al2Z9XzBdQsHrk6ZYq7dIshDA0+8nkB1oXZz9eCKnXbY732qTLgFw/NCyE+4G1JL5ldFvym0q5CiGUCyHUySW+XGNMfoBzaTHjkyRJpYxJ91IuhLAvcDjwFjAm8wqfGOMXJP7BXx44ZSv9NCSxymkT8PC2ileSpLIiuRHpSSSSu22Bj0IIF4YQWqauCSGUDyHsF0K4AviCLSVjttb3GrasjLwxhHBIMkFLMvn5AtAR+DmPLhaHEK4JIXTJnJxOxvYgiVW9m4GXMrV5J4RwSwihZ+ZEXwihcQjh7yRqpZMcu0BijEuA25N3LwshXJRa8R5C2DmEcCuJJGdRkrVFjfeT5O0xIYS6RRi3KH4C7gghnBpCqJKMczcS/6ZKrcS+JJd2qX9zHRpCGJdcaUwIoX4I4a8kEm0r8xk3NddhIYRq+VyXlzuAhSQ2tnwxhHB4ptdhOxKvn+Yk/gZKMulXEXglhNApGVsIIfRJxlcZ+IotZWRycw+JD6P2IfGt0JXkkaQvpJ+AESGEm1MfCoQQaoYQLmZLYvu25N9JZpNJ/H3WAx4OITRJtq0aQhgBPEXupXKKLMa4FrgyeXdECOGmEEK95Li1QgiXkUhsr0znuAWMbQaJDwUA7gohjEy9ryUfmwdJfGOkqPX3/528PTWE8McQQqVk3w1DCBNIfNiwrMgTAJLlwVLfLj6SxOv1oFTyPfmabRNC+H8k/m6PytR8Llu+ufOvEELGtw1CCF2BqSRet5IkaQdi0r30S32l+unkSp7s3kjedtpKP0NJJOdfyV5TVpKk36oY4xTgEBLlKeoD1wKfhxDWhRCWkUhGfkAiwVabRBI1r0R5duclr90VeA34JYSwCphNIlF7KrA0j7a7kEjwvJ1stzyEsBb4nMQmfBEYHWP8NFObOsDZJOrVrw4hrAghrAGWAGclr5kQY3y5gPGnXEiihEc54K/AihDCcuB74ExgPPBRIfssTrx3kpj/gcCPIYRvQgiLQgiLihBDQd1OonzHncCq5Py/BI5Lnr8qxvhULu0mkZgfJL4RsDLZ9gcSz++FbFkVnZtUonlIsu3Xybm+kU+bDDHG1UB/Eo9pExIfYPwcQviJROmZA0kkq4fFGPOLY1s7DdgDeDeEsJrEN1BeAZqSSBIPjjGuyqtxjHEZ8FimQw8kk9DF9RGJb5qeA3yffE9YQeJbKuVI/F3kKPOY/EAv9U2Wo4GvQggrSXxzYRKJjZFvz94uDW5myzdazwV+SL7elpPYYPRR4OltMG5BDAfmkfjA8B6Sf/MkPlA5nsTffOr98NdC9v23ZN8VSHzrYW2y729IvA//E3iumPETY7yFRAJ/E4n38DdIvD8vJfF39Gkylr3ItGls8v/hziTxrYi2JDaJ/jmE8DOJhVVtSDwGkiRpB2LSvfRrlrwdH0KI2X/Y8o/TfDdiw9IykiTlKsb4Jomkx4kkVlzOJ5H0qUkiWfUGiQTaXjHGoXnUlc6t3/eBziRWvS4l8e+u1cn7B8YY8/tvcj/gGmAGiaRUqiTEfBIJq04xxpuytTkB+AuJBP9CoBKJFcSLSSTbescY/19BYs82j19JfOvuXBJJyPUkaozPAI6LMRZ0b5nsihRvjHE6iZWmr5JYibwLieRsYessF8Z6Et8YvBj4jMTq65+SsR8ZY8y1nEeyTM6RJOY5L9lPBF4G+sYYb8hv0BjjAyS+jfEGiVXAjUjMs8A19GOMn5BI9I0l8fxtTMa/gERSv22M8fG82m8n75D45sd9JB7XCiQ+KLgLaBdjzF4zPTeZk+7pKi1DjPFCEq/VN0n8Da8n8TieCxyW/PvIrd3lJJ67mSQ+fCufbHc6iW/MFKbMU0Fj3RxjHE4iwT2TRDmUCiQ+ODydxCKcEpFc9NMJuAr4H4lvAmwk8UHQITHGu9hSumtlIfteSeIDpJuARSQe240kVpCfGGM8vZjhZx7rehL/vZhA4oOrX0l8gLgGmEXiQ5oDgYeytXsO6A48T2J+FUj8d+EeYP/kfh2SJGkHEnJfPK3tKZk8XxdjzLGZWgjhHyRW/8wg8bX2vCyNMf4pj/73IvG1xjXALjHGon51U5IkSSp1kiWJziKxgeqBJR2PCickNmT+X/Lu7jHGr0oyHkmSpOIq6g7x2n6+Tt4+nvxKY1GclLx90oS7JEmSdiTJWvnDk3fvKMlYVGQXJW/nmnCXJEk7AsvLlH6vJm8HFqVxCCGw5auklpaRJEnSDiO5IefNQC0SpZjSsYGq0iy5yejEEEL31GbMmY7fA4xKHrq2ZCKUJElKL8vLlAL5lZdJnn+VRB3Rm4DLYoxrMp0rB/QBfokx5thUK4TQHZhGYiOh3WKMm9M/A0mSJGn7CSGcR2KTzAZs2fPg2FJQn165CCHsC3yY6dBPJPZwqJbp2C0xxnO3Z1ySJEnbiivdS0AI4cgQwszUT/JwpczHQghHZmoyjMRGPecBi0MIr4UQHgkhzAC+A14isflUblIbqD5owl2SJEk7iDokNpUNJDYoPd6Ee6m2APgT8G8SGydXILG57FckNpfuY8JdkiTtSFzpXgJCCCNJ7FSfn1ExxkmZ2lQFTgeOB/YGKgHfkthw6GlgcoxxabZxKievqQt0iDH+N01TkCRJkiRJkiTlwqS7JEmSJEmSJElpYnkZSZIkSZIkSZLSxKS7JEmSJEmSJElpYtJdkiRJkiRJkqQ0MekuSZIkSZIkSVKaVCjpAH4rQgjfAdWAr0o6FkmSJEmSJElSnnYDfokxNixK4xBjTHM8yk0IYVXlypVr7rHHHiUdiiRJkiRJkiQpDwsWLGDdunWrY4y1itLele7bz1d77LHH3nPmzCnpOCRJkiRJkiRJeWjbti1z584tcsUSa7pLkiRJkiRJkpQmJt0lSZIkSZIkSUoTk+6SJEmSJEmSJKWJSXdJkiRJkiRJktLEpLskSZIkSZIkSWli0l2SJEmSJEmSpDSpUNIBSJIkSZIkSYURYyTGWNJhSCrlQgiEELb7uCbdJUmSJEmSVOpt2rSJZcuWsXr1atavX1/S4UgqI8qXL0+1atWoVasWNWvW3C5JeJPukiRJkiRJKtU2bdrEl19+ya+//lrSoUgqYzZt2sTq1atZvXo1derUYZdddqFcuW1bdd2kuyRJkiRJkkq1ZcuW8euvv1K+fHl22WUXqlevvs2TZpLKvhgj69atY/Xq1SxfvpyVK1dSpUoV6tatu03HNekuSZIkSZKkUm316tUA7LLLLtSuXbuEo5FUllSrVo1q1apRoUIFfvjhB1asWLHNk+5+JChJkiRJkqRSK8aYUcO9evXqJRyNpLKqVq1aAKxbt26bb8Rs0l2SJEmSJEmlVubkmCVlJBVV+fLlM3436S5JkiRJkiRJUhlh0l2SJEmSJEmSpDQx6S5JkiRJkiRJUpqYdJckSZIkSZIkKU1MukuSJEmSJEll0KJFiwgh0LNnz5IOZauaNWtGCKGkw1ARTZo0iRACY8eOLelQyoQKJR2AJEmSJEmSVFzNxjxf0iEUyKJrjyzpEPQbNHLkSO69917+85//5PohzdixYxk3bhz33HMPI0eO3O7x7WhMukuSJEmSJEnapl577TU2bNhQ0mFI24VJd0mSJEmSJEnb1B577FHSIUjbjTXdJUmSJEmSpDJu7dq1jBkzhqZNm1K5cmVatmzJ+PHjiTFmXLO1GvBjx44lhMCkSZOyHE/VY48x8ve//50OHTpQrVo19t13XyBRuiSEwNSpU3nppZfo1asXderUIYTAypUrs/SRmy+//JKzzjqLPffckypVqlCvXj06d+7MX//6V9auXZtx3fz58xk7dixdu3alYcOGVKpUiSZNmjB8+HD+97//5dp3CIFmzZqxadMmrrvuOlq1akXlypXZbbfduPDCC1m3bl2u7dasWcMVV1xBu3btqFatGrVq1aJHjx5MmTIl1+vzk/nxefXVV+nRowc1a9akQYMGnHrqqfz0008A/PDDD5x22mk0btyYKlWq0LlzZ6ZOnZqjv19//ZW7776bAQMG0KJFC6pWrUqdOnXo3r07jzzySK6Pwb333gtAr169CCFk/CxatIhmzZoxbtw4AEaNGpXlfG7jf/nllwwdOpSdd96ZqlWr0rFjR5599tk85z979myGDRvGrrvuSuXKlWncuDGjRo1i0aJFWa675557CCFkxJLy448/Uq5cuSzzSHnvvfcIIXDsscfmOX5JMOkuSZIkSZIklWHr16+nX79+3Hnnney111706tWLJUuWMGbMGC677LK0jXP66aczevRoGjRoQP/+/WnRokWW8w899BCHH344P//8M4cffjidOnXa6uap06dPp3379tx2221s3ryZAQMG0LVrV5YuXcoll1zC999/n3HtxIkTGTduHKtWraJjx47079+fWrVqcf/999OpUyf++9//5jnOsGHDuOKKK2jSpAn9+vVj9erVXHfddZx88sk5rv3+++854IAD+Mtf/sKKFSvo27cvBxxwAO+//z6DBg3i2muvLeQjl/DUU09x6KGH8vPPP9OvXz8qV67MxIkTGTBgAEuXLqVr164899xzHHDAAey7777MmjWLww47jNmzZ2fpZ9GiRZxyyim888477L777gwYMIB9992XmTNncuKJJ+bY7HTEiBEZ3zQ49NBDGTFiRMZPjRo1OOaYY+jQoQMABx10UJbzDRs2zDF2p06dePPNNzn44IPZb7/9eP/99xk4cCAvv/xyjjk/8cQTdOzYkYceeohGjRrRv39/GjZsyKRJk+jYsSNz5szJuDb1YdB//vOfLH1MmzYt48Oj7OdSHwr06NGjAM/A9mN5GUmSJEmSJKkMe/vtt+nWrRv/+9//qF+/PpBYAdy1a1cmTJjAmDFjqFGjRrHHefLJJ/nwww9p27ZtrufvuusuHnnkEY4//vgC9bdixQqOOeYYfvrpJyZMmMC5556bJUk/ffp06tatm3F/4MCBnHrqqTlK1dxzzz384Q9/4LzzzuP111/PMc7ixYupVq0an3zyCc2aNQNg4cKF/O53v+PBBx9k3LhxWfocNWoUc+fO5YILLuCqq66iYsWKAHzxxRf069ePSy+9lCOOOIL27dsXaJ4pt912G5MnT2bIkCEArF69mgMPPJBp06bRo0cPOnXqxKRJk6hSpQoAl112GVdddRU33HBDlhXeO++8My+99BJ9+vShXLkta6oXLlzIIYccwpVXXsnIkSMz5jpp0iRGjhzJggULGDNmTI5vOtxwww2MHTuWjz/+mFNOOSXfjVTvvfdezj77bG688UYqVEiklm+++WbOO+88rrrqKvr165clnuHDh1O1alVeeeUVunfvnnHuvvvuY8SIEYwaNYp3330XgObNm7P77rszc+ZMfv3114zHYerUqYQQ2GuvvXKsvE/dz+vbGyXFle6SJEmSJElSGVauXDkmTpyYkXAH6NixI4cffji//PIL7733XlrGufDCC/NMuAMceeSRBU64QyJJ/+OPP3LUUUdx3nnn5VgV3717d2rXrp1xv0uXLrnWhh81ahQHHXQQU6dOzSjVkt3f//73jCQ0JBK8v//97wGYMWNGxvGPPvqIF198kQMPPJBrr702I+EO0KJFC/72t7+xadMmJk6cWOB5pgwbNiwj4Q5Qs2ZN/vjHPwKwZMkS/vGPf2QkmgH+9Kc/EUJg2rRpWfqpV68e/fr1y5JwT83pkksuYfPmzfmWeymO1GOQSrgDnHnmmdStW5eZM2eyfv36jOM333wzv/zyC9ddd12WhDvA8OHDGThwILNmzeKDDz7ION6jRw/WrVvHzJkzM45NmzaNffbZh0GDBrF48eKMsjSbN2/mjTfeoH79+vm+LkuCK90lSZIkSZKkMqxZs2a0atUqx/HUsW+//TYt4/Tv379Y57N79dVXATjttNMK3GbNmjU8++yzfPTRRyxfvpwNGzYAiTnGGFmwYAH7779/ljYVK1bMdSV0bo/PK6+8AsCAAQNyLY1z8MEHAzBr1qwCx5zSt2/fHMdSJXo6duxInTp1spyrXbs29erVy/P5e+ONN5g6dSpLlizh119/JcaYce3nn39e6PgKomfPnlk+iACoUKECLVq04P3332fZsmU0atQIyPpY5ubggw9mypQpzJo1K+M569mzJ/fffz9Tp06lZ8+eLF26lDlz5nDWWWfRs2dPrr76aqZOncrIkSP58MMP+emnnxg8ePBWyxhtbybdJUmSJEmSpDKsSZMmuR5PlZTJa7PQwtp9992LdT67r776CiDX1eu5ef311znhhBP48ccf87xm9erVOY41atSI8uXL5zie2+OTWkV94YUXcuGFF+Y5ztKlSzN+z60cy8CBAxk4cGCWY7vuumuO66pXr57nudT5zGMBGYnm3ErppOT2OKRDYV5rqccye1347DLPL/XhSKpsTKqee69evTjooIOoVKlSRtK9tJaWAZPukiRJkiRJUpmWrlW+mzdvzvd85tInRTmfl4LEv2bNGo477jiWLVvGZZddxoknnkjTpk2pWrUqIQSGDh3Kww8/nLHhZmH7T9m0aRMA3bp1y7FRbGaZS/lkrree0qxZsxxJ9/ziKEyMF154Ia+//jrdu3fniiuuYJ999qFOnTqUL1+el19+mUMPPTTXxyEdCvtYhhAYPnx4vtdlLg3TokULdtttt4y67tOmTSOEQPfu3alatSqdOnXKkpAHk+6SJEmSJEmSSkilSpWARAI7N6mV59vLbrvtxrx585g/fz5t2rTJ99oZM2awbNkyhgwZwhVXXJHj/BdffJGWmFIruY855hjOOeecArXZVgnuvDz11FOUL1+eZ555JkvNe0jf45AOTZo0YcGCBdxyyy3UqlWrwO169OjBAw88wMyZM5k6dSrt2rWjXr16ABklZr744gtmzJhBvXr12GeffbbVFIrMjVQlSZIkSZKk34D69etTsWJFFi5cyMaNG7OcW79+fY4NO7e1Pn36AHDnnXdu9doVK1YAiUR9dvPnz8+yGWc6YpoyZUpa+tsWVqxYQc2aNXMk3AEmT56ca5vUBy7Zn/eCni+Koj6WqZXrTzzxBJ988gm9evXKce6mm25i5cqVdO/evdTVcwdXuktlVrt725V0CAUye8Tskg5BkiRJkiSRSKx26dKFGTNmcNttt3HuuecCsGHDBs4//3wWLly4XeM55ZRTuP7663n22We59dZbOfPMM7MkUGfMmEH79u2pXbt2xqanTz75JBdffDE777wzACtXruTkk0/O2FC1uLp06ULv3r157bXXOP/887nyyisz6pVDogTPq6++SrVq1TI2Vd3eWrVqxZw5c3j00Uc5/vjjM45PmDCB//znP7m2ady4MQCfffZZRjI8r/PpMnr0aO677z7OP/986taty9FHH53l/PLly3nkkUcYNWoUVatWzTjeo0cPACZOnEiMMUv5mAMPPJBKlSpx1113AaWztAy40l2SJEmSJEn6zbj88sspV64c5513HgceeCCDBw+mZcuWTJ48mREjRmzXWHbaaScmT55MzZo1Ofvss9lzzz05/vjjOfroo2nRogXdu3fPWOHesWNH+vbty5dffkmrVq0YNGgQgwYNonnz5nzzzTcMGDAgbXE9+OCDtG/fnptuuommTZvSu3dvTjjhBLp160bDhg059NBDee+999I2XmFddNFFAJxwwgl0796doUOH0rZtW/70pz9x/vnn59rm6KOPJoTA6NGjGThwIKeccgqnnHIKy5YtA6Bfv35UqVKFCRMmcPjhh3PyySdzyimnFCsJv+eee/LAAw+wdu1a+vfvT5s2bRg0aBADBw5kv/32o1GjRpx55pk5Nvpt2bIlTZo04ddff82o555SrVo1OnXqxK+//gpsSdCXNq50lyRJkiRJUpm36NojSzqEMqFPnz4888wzjBs3jg8++IDq1avTp08fxo8fz6RJk7Z7PL169eKjjz5i/PjxvPTSS0yZMoVatWqxxx578Mc//pGGDRtmXPv0009z9dVXM3nyZF588UUaNGjACSecwFVXXcXo0aPTFtMuu+zCzJkz+cc//sGjjz7KrFmzWL9+PY0aNWK//fZjwIABHHfccWkbr7CGDRtG3bp1ufLKK/noo4+YPXs2HTt25PbbbyfGyIQJE3K0+d3vfscDDzzA3/72N15++WXWrl0LwKWXXkq9evVo3LgxTz/9NFdccQVvvPFGRt3/3//+97Ru3brIsQ4ePJiPP/6Yv/3tb7zyyiu8+OKLVKlShcaNGzNs2DCGDBmSa5mcHj16ZHz4sdNOO2U517NnT95880122mkn2rdvX+TYtqWwvQv9/1aFEObsvffee8+ZM6ekQ9EOwvIykiRJkqTfgs2bN2estm3dujXlylm4QVLhFea9pG3btsydO3dujLFtUcbyXUqSJEmSJEmSpDQx6S5JkiRJkiRJUpqYdJckSZIkSZIkKU1MukuSJEmSJEmSlCYm3SVJkiRJkiRJShOT7pIkSZIkSZIkpYlJd0mSJEmSJEmS0sSkuyRJkiRJkiRJaWLSXZIkSZIkSZKkNCmTSfcQQrUQwsAQwt0hhP+GEFaFEH4OIXwcQrg8hFCjkP0tCiHEfH7abKu5SJIkSZIkSZJ2HBVKOoAiGgrclfx9DvBvoBZwIDAOODGE0CPG+EMh+703j+M/FSlKSZIkSZIkSdJvSllNuq8H7gAmxBg/Tx0MITQCngf2A24ikZwvsBjjyPSFKEmSJEmSJEn6rSmT5WVijPfFGM/InHBPHv8WODN5d3AIodL2j06SJEmSJEmS9FtVJpPuW/Fx8rYyUK8kA5EkSZIkSZIk/baU1fIy+WmRvN0ALC9MwxDCn4E9gHUkasU/FWP8Mb3hSZIkSZIkKe3G1i7pCApmrFsHFkTPnj2ZNm0aCxcupFmzZiUdjraDZs2asXjxYmKMJR1Kse2IK93PTd7+O8a4rpBtrwNOA84B/gksCiGcXJgOQghzcvshkcyXJEmSJEmSim3RokWEEOjZs2dJh6I0mjRpEiEExo4dW9KhqBh2qKR7COEI4GQSq9wvK0TTZ4DBQFOgGrAPcCOJEjUTQwgD0xupJEmSJEmSJGlHtMOUlwkh7AU8AATgzzHGj7fSJEOM8Zxsh+YAo0MIn5FY8T4emFLAvtrmEd8cYO+CxiRJkiRJkiRJKnt2iJXuIYQmwL+BusCNMcab09T1ROAHoFUIoXma+pQkSZIkSZKKbOzYsTRvnkhVTZs2jRBCxs/IkSO3Wnpm7NixhBCYNGlSluPLli3j4osvpm3bttSoUYPatWvTqlUrhg8fzrvvvpujn6VLl3LRRRexzz77UL16derUqcO+++7LJZdcwrJlyzKu+/bbb7nuuuvo0aMHu+66K5UqVaJhw4YMHjyYWbNmFXr+c+fOZdiwYTRq1IhKlSqx6667Mnz4cD777LM827zwwgv07duXunXrUqVKFVq3bs2YMWNYuXJljmtjjDzyyCN0796dhg0bUqVKFXbbbTf69OnDbbfdluXadM6tZ8+ejBo1CoBx48ZleV6zP1eFmU/m5/vdd9/lqKOOol69eoQQ+Oijj5g6dWrGa2f58uX83//9H40aNaJy5crss88+/Otf/8o13ueff54//OEP7LXXXtSqVYvq1avToUMH/vrXv7JuXe5Vvzdu3Mg111zDnnvuSZUqVWjRogWXXXYZ69evz/exKcx8S4Myv9I9hFAfeAXYHbgH+FO6+o4xbg4hLAAaAI2AhenqW5IkSZIkSSqKfffdlyFDhvDEE0+wyy67cNhhh2WcO/jgg4vU55o1a+jSpQvz589nzz335NBDDwXgyy+/5OGHH6ZFixZ07tw54/q5c+fSr18/lixZQqNGjTjssMPYtGkTn332GX/961/p27dvRtL/6aef5sILL6Rly5a0a9eOWrVqMX/+fJ566imee+45nnvuOfr161egOF977TWOPvpo1q5dy/7770/Pnj2ZN28e999/P0899RQvvPAC3bp1y9Lmmmuu4eKLL6ZChQr06NGD+vXr8+abbzJ+/Hieeuoppk+fzi677JJx/UUXXcT48eOpWbMmBx98MHXq1OHbb7/l448/Zv78+Zx55pkZ16ZzbocddhgbN27kzTffpEOHDuy7774Z51q2bFnk+aRMnz6dP/7xj7Rq1Yp+/frxzTffUK7cljXZK1eupGvXrvz000907tyZNWvWMH36dE4++WQ2b97MKaeckqW/k08+mZ9//pm2bdvSrl07Vq1axbvvvssll1zCa6+9xssvv0z58uWztDnxxBN5/PHHqVGjBocddhgxRm688UY+/PDDPDdQLep8S1KZTrqHEGoCLwJtgCeBU2P6t7etm7xdk+Z+JUmSJEmSpEIbOHAg++67L0888QRt2rTJsQp60aJFhe7z8ccfZ/78+Zx99tnccsstWc798MMP/PDDDxn3N27cyJAhQ1iyZAmjR4/mmmuuoWLFihnnP/zwQ3beeeeM+wcddBAff/wx7du3z9LvSy+9RP/+/TnjjDP4/PPPCSHkG+PPP//MsGHDWLt2LXfccQenn356xrkJEybw//7f/2Po0KHMnz+fypUrAzBr1iwuvfRSatasyauvvprxwcG6des46aSTeOyxxzj77LOZPHkyAL/++is33XQTzZo14/3332ennXbKMu+33norS0zpmhvAmDFjaNiwIW+++SYDBw7MdTPVws4ns3vuuYfx48dzwQUXZDk+depUIPEBwpAhQ7j33nupXr16xrGBAwdy5ZVX5ki6/+Mf/6Bv374Z1wKsXr2aoUOH8txzz/Hggw8yfPjwjHMPP/wwjz/+OC1atGD69OnsuuuuACxcuJDu3bvz9ddfp3W+JanMlpcJIVQGngY6Ai8BJ8YYN6V5jLZAa+AXYF46+5YkSZIkSZJKi1RS/ZBDDslxrkGDBuyzzz4Z95988knmzZtH+/btue6667Ik3AH2228/mjRpknG/Xbt2OZLSAIceeijHHnssCxYs4JNPPtlqjJMnT+b777+nW7duWRLuAOeffz6/+93v+Prrr3nqqacyjt96661s3ryZ8847L8tK/cqVK3PrrbdStWpVnnjiCZYsWQLAqlWrWLduHR06dMiScAeoUKEC3bt3z3IsXXMrqMLOJ7N99tmHP//5z3n2XatWLe68884sSfQBAwbQrl07vvzyyxwf5gwcODDLtQA1a9ZkwoQJQCJhn9kdd9wBwJVXXpmRcAdo3rw5l112WdrnW5LK5Er3EEJ54GGgFzADGBxjzLfwTwjhLOAs4KkY40WZjh8KLI0xvp/t+vbAIyQ2Zp24tf4lSZIkSZKksup3v/sdQEYZjz59+lClSpVcr3311VcBOPXUU7OUJ8nPunXr+Pe//827777Ljz/+mFHDe/bs2QB8/vnntGvXLt8+ZsyYAcCwYcNyPf/73/+e999/nxkzZnDCCSdstU2DBg3o168fTz/9NG+99RbHHnssDRo0oEmTJjz//PNcf/31DBs2jMaNG2/zuRVUYeeT2dFHH53vivuOHTvm+KABoFWrVsyePZtvv/2WZs2aZTn3+eef88ILLzB//nx+/vlnNm/enFEm5vPPP8+4bsOGDbzzzjuUK1eOY445JscYJ554Iqeddlpa51uSymTSnUTyfFDy96XA7Xm8YP4UY1ya/L0+iVXrjbJd0xX4SwhhMbAA+BFoDuxP4vGZBlyEJEmSJEmStIPq3bs3559/PjfddBNHH300lSpVYt9996Vfv36cfPLJWZKtX331FQB77LFHgfqePXs2/fv3z7fszerVq7fazzfffAOQI/Gbkjqeui71ewiBpk2bFrjNvffeywknnMAFF1zABRdcQPPmzenevTtDhw7NUZ+9sHO79tprmTcva0GNNm3aMGbMmDzbZ1aU+aTsvvvu+fad+dsJmdWoUQMgy+aoMUb+9Kc/MWHChDxrsWee97Jly1i/fn3G5rfZ1axZkzp16uTYGLU48y1JZbW8TN1Mvw8CRuTxU6MAfb0E/AtYBXQAhgAtgTeAU4HeMcZf0ha5JEmSJEmSVII2b96c6/Ebb7yRTz/9lPHjx9OzZ0/mzJnDVVddRevWrZkyZUqO6wtSpzzGyHHHHceiRYs4/fTT+eijj1i1alXGiuiLLroo47qC2tq4BYkrvzaHHHII8+fP58EHH+Skk05i06ZN3HvvvRx66KEcf/zxGdcVZW7//ve/uffee7P8/Pvf/y50vIWZT0pe31zIr01eHn30UW688UZ23XVXHn/8cZYsWcL69euJMWYk5zPPO/V7UZ6bgthW/RZVmUy6xxjHxhhDAX4W5dJmZLa+3o4xnhxjbB9jrB9jrBhjrBdj7BVjnJjuOvGSJEmSJEnStpRaSbxmzZpcz6dWquemdevWXHDBBbz00kssXbqUG264gfXr12cp/bHbbrsBMH/+/K3GMm/ePObNm0fHjh2544476NChAzVr1sxIkn7xxRcFnleqzMvChQtzPb948WIAGjVqlKVNjDHjXEHaQKK++dChQ7nvvvtYvHgxb7/9Nk2aNGHy5Mm8+OKLRZ7b1KlTiTFm+UltZFrQx6Ao80m3VN38O+64gyFDhtC4ceOM2v65zbt+/fpUqlSJ7777LqP8TmarV6/OscodSs98C6tMJt0lSZIkSZKk37JUYn3jxo05ztWvX5+KFSuycOHCHOfXr1/PtGnTCjRGlSpVGD16NI0aNeKHH37I2Gy1T58+AEycOHGrK9RXrFgB5F66ZMWKFbzyyisFigWgW7duADz44IO5nk8dT123tTY//vgjL7/8MuXKlePAAw/Md+wuXbpw0kknAVtqtadzbin5Pa+QvvkUV2ruqQ9gMps8eXKOYxUrVqRz585s3ryZJ554Isf5Rx55JNdxSst8C8ukuyRJkiRJklTGpBLrCxYsYNOmrIUaKlWqRJcuXVi+fDm33XZbxvENGzZw/vnn57pSfMqUKcycOTPH8Q8//JDvv/+emjVrUrduouLz4MGDadWqFR9//DFjxozJkSD+6KOP+PrrrwFo2bIl5cqV4/XXX8+yseavv/7K6aefzvLlyws85+OOO45ddtmFGTNmcOedd2Y5d8sttzBr1iyaNGnCoEGDMo6feeaZlCtXjptvvpn33nsv4/j69es5++yz+eWXXxg8eDC77rorAF9++SWTJk3il1+yVptet24d//nPf4AttdHTObeU1Gr+zz77LNfzhZ3PttKqVSsA7rzzziwfvMyYMYPrr78+1zapb0tcfvnlfPvttxnHFy9ezJVXXplrm9Iy38Iy6S5JkiRJkiSVMZUqVeKwww7ju+++o0OHDgwfPpxTTjmFe+65B0gkNsuVK8d5553HgQceyODBg2nZsiWTJ09mxIgROfqbOnUqXbt2pUmTJhx99NEMGzaMXr16ZaxOvvLKKzPKh1SoUIEnnniChg0bct1119G0aVOOPfZYBg0axN57781+++2XUXqmQYMGnHzyyaxatYoOHTpw1FFHceyxx9KsWTNef/11Ro4cWeA5V69enQcffJCqVaty2mmn0bFjR4YOHcr+++/PueeeS/Xq1XnooYeoXLlyRpvOnTtz5ZVXsmrVKrp27Urfvn058cQTadmyJY8++ih77rknt956a8b1y5cvZ9SoUey888706NGDYcOGMXDgQHbbbTdmzpxJ586dGTx4cNrnltKlSxcaNGjA448/Ts+ePfnDH/7AKaecwltvvVWk+Wwr55xzDtWrV+f2229nn3324cQTT6R79+706NGD008/Pdc2w4YNY9CgQcyfP5/WrVszaNAgBg4cSNu2bWnXrl2uG72WlvkWVoWSDkCSJEmSJEkqtrE/lXQE293EiRP505/+xCuvvMJDDz3Epk2b2LhxI6NGjaJPnz4888wzjBs3jg8++IDq1avTp08fxo8fz6RJk3L0NXLkSCpUqMD06dN59913+emnn2jYsCFHHHEE559/Pj179sxy/T777MNHH33E9ddfzzPPPMOzzz5LtWrVaNq0KZdeeint27fPuPaOO+6gTZs23H333bz22mvUrl2bPn36cPXVV2d8SFBQvXv3ZtasWVx99dW8/vrr/Pe//6V+/fr8/ve/59JLL6V169Y52lx88cV06NCBCRMmMGvWLNauXcvuu+/OBRdcwJgxYzJW8APsscce3HDDDbz++uvMnTuXd999lxo1atC8eXMuu+wyTjnllIwSMOmeGyRK+jz//PNcfPHFvPvuu0yfPp0YIwcffHBGCZXCzGdbadWqFbNmzeLCCy/knXfe4ZlnnqF169b885//5NRTT+WGG27I0SaEwKOPPsr111/P3XffzQsvvECjRo04++yzGTt2bK7PHZSO+RZWKMzOwCq6EMKcvffee+85c+aUdCjaQbS7t11Jh1Ags0fMLukQJEmSJEll2ObNmzNKbbRu3Zpy5SzcIKnwCvNe0rZtW+bOnTs3xti2KGP5LiVJkiRJkiRJUpqYdJckSZIkSZIkKU1MukuSJEmSJEmSlCYm3SVJkiRJkiRJShOT7pIkSZIkSZIkpYlJd0mSJEmSJEmS0sSkuyRJkiRJkiRJaWLSXZIkSZIkSZKkNDHpLkmSJEmSJElSmph0lyRJkiRJkiQpTUy6S5IkSZIkSZKUJibdJUmSJEmSJElKE5PukiRJkiRJkiSliUl3SZIkSZIkqQxatGgRIQR69uxZ0qFsVbNmzQghlHQYO4SpU6cSQmDkyJFF7qM0v3YmTZpECIGxY8eWdChFVqGkA5AkSZIkSZKKq9297Uo6hAKZPWJ2SYcgaRsz6S5JkiRJkiRpm3rttdfYsGFDSYexQ+jcuTOffvoptWvXLulQlAeT7pIkSZIkSZK2qT322KOkQ9hhVKtWjTZt2pR0GMqHNd0lSZIkSZKkMm7t2rWMGTOGpk2bUrlyZVq2bMn48eOJMWZcs7U63mPHjiWEwKRJk7IcT9VjjzHy97//nQ4dOlCtWjX23XdfAEaOHEkIgalTp/LSSy/Rq1cv6tSpQwiBlStXZukjN19++SVnnXUWe+65J1WqVKFevXp07tyZv/71r6xduzbjuvnz5zN27Fi6du1Kw4YNqVSpEk2aNGH48OH873//y7XvEALNmjVj06ZNXHfddbRq1YrKlSuz2267ceGFF7Ju3bpc261Zs4YrrriCdu3aUa1aNWrVqkWPHj2YMmVKrtfnJ/Pj8+qrr9KjRw9q1qxJgwYNOPXUU/npp58A+OGHHzjttNNo3LgxVapUoXPnzkydOjVHf1ur6f7iiy9y1FFH0aBBAypXrszuu+/OwIEDef7553O9viCvnZQZM2Zw1lln0b59e+rWrUvVqlVp06YNY8aMyXiuczN79myGDRvGrrvuSuXKlWncuDGjRo1i0aJFW3v4yiST7pIkSZIkSVIZtn79evr168edd97JXnvtRa9evViyZAljxozhsssuS9s4p59+OqNHj6ZBgwb079+fFi1aZDn/0EMPcfjhh/Pzzz9z+OGH06lTp61unjp9+nTat2/PbbfdxubNmxkwYABdu3Zl6dKlXHLJJXz//fcZ106cOJFx48axatUqOnbsSP/+/alVqxb3338/nTp14r///W+e4wwbNowrrriCJk2a0K9fP1avXs11113HySefnOPa77//ngMOOIC//OUvrFixgr59+3LAAQfw/vvvM2jQIK699tpCPnIJTz31FIceeig///wz/fr1o3LlykycOJEBAwawdOlSunbtynPPPccBBxzAvvvuy6xZszjssMOYPbvg+wCMHj2aI444gn//+9+0bt2awYMH07x5c/7zn/9w/fXX57i+sK+dP//5z0ycOJFKlSpxyCGH0Lt3b1atWsX48eM5+OCDWbNmTY42TzzxBB07duShhx6iUaNG9O/fn4YNGzJp0iQ6duzInDlzCvdAlgGWl5EkSZIkSZLKsLfffptu3brxv//9j/r16wPw3nvv0bVrVyZMmMCYMWOoUaNGscd58skn+fDDD2nbtm2u5++66y4eeeQRjj/++AL1t2LFCo455hh++uknJkyYwLnnnpslST99+nTq1q2bcX/gwIGceuqpOUrV3HPPPfzhD3/gvPPO4/XXX88xzuLFi6lWrRqffPIJzZo1A2DhwoX87ne/48EHH2TcuHFZ+hw1ahRz587lggsu4KqrrqJixYoAfPHFF/Tr149LL72UI444gvbt2xdonim33XYbkydPZsiQIQCsXr2aAw88kGnTptGjRw86derEpEmTqFKlCgCXXXYZV111FTfccAP33nvvVvt/4IEHuPHGG2nSpAnPP/98lvh+/vln3nnnnRxtCvvaufzyy+natWuW52XdunWcc8453Hnnndx4441cfvnlGecWLlzI8OHDqVq1Kq+88grdu3fPOHffffcxYsQIRo0axbvvvlvQh7FMcKW7JEmSJEmSVIaVK1eOiRMnZiRNATp27Mjhhx/OL7/8wnvvvZeWcS688MI8E+4ARx55ZIET7pBI0v/4448cddRRnHfeeTlWxXfv3j3LZqFdunTJtTb8qFGjOOigg5g6dWpGqZbs/v73v2ck3AGaN2/O73//eyBRMiXlo48+4sUXX+TAAw/k2muvzUi4A7Ro0YK//e1vbNq0iYkTJxZ4ninDhg3LSLgD1KxZkz/+8Y8ALFmyhH/84x8ZCXeAP/3pT4QQmDZtWoH6/+tf/wrATTfdlOMDgerVq3PIIYfkaFPY184RRxyRJeEOULlyZW666SYqVKjA008/neXczTffzC+//MJ1112XJeEOMHz4cAYOHMisWbP44IMPCjTHssKV7pIkSZIkSVIZ1qxZM1q1apXjeOrYt99+m5Zx+vfvX6zz2b366qsAnHbaaQVus2bNGp599lk++ugjli9fzoYNG4DEHGOMLFiwgP333z9Lm4oVK+Zaxz63x+eVV14BYMCAAbmWxjn44IMBmDVrVoFjTunbt2+OY6kSPR07dqROnTpZztWuXZt69eoV6Pn75ptv+PTTT6lXr16WxP7WFOW1s2TJEp599lnmzZvHqlWr2Lx5MwCVKlXi888/z3Jt5sczNwcffDBTpkxh1qxZOZ63ssykuyRJkiRJklSGNWnSJNfjqbIgeW0WWli77757sc5n99VXXwHkuno9N6+//jonnHACP/74Y57XrF69OsexRo0aUb58+RzHc3t8Uht7XnjhhVx44YV5jrN06dKM33Pb0HTgwIEMHDgwy7Fdd901x3XVq1fP81zqfOax8lLYxzKlsK+dG2+8kYsuuoj169cXqP/U49mwYcN8ryvIHMsSk+6SJEmSJElSGba1zUoLKrViOS+ZS58U5XxeChL/mjVrOO6441i2bBmXXXYZJ554Ik2bNqVq1aqEEBg6dCgPP/wwMcYi9Z+yadMmALp165Zjo9jMMpdjya3eerNmzXIk3fOLI13PYWH7Kcz1M2fOZPTo0dSuXZs777yTnj170rBhQypXrgxA48aNc6yM37RpEyEEhg8fnm/f+ZUtKotMukuSJEmSJEm/AZUqVQISCezcpFZLby+77bYb8+bNY/78+bRp0ybfa2fMmMGyZcsYMmQIV1xxRY7zX3zxRVpiSq38PuaYYzjnnHMK1Ca3RP/2tttuuwEwf/78bTbGU089BcBVV13FiBEjspxbu3Yt3333XY42TZo0YcGCBdxyyy3UqlVrm8VW2riRqiRJkiRJkvQbUL9+fSpWrMjChQvZuHFjlnPr168v8Iad6dKnTx8A7rzzzq1eu2LFCmBLcjmz+fPnp20jzlRMU6ZMSUt/20vjxo3Za6+9WLZsGU8++eQ2GSO/5+Cxxx7L9cOHsvp4FpdJd0mSJEmSJOk3oFKlSnTp0oXly5dz2223ZRzfsGED559/PgsXLtyu8ZxyyinUr1+fZ599lltvvTVH0nbGjBn89NNPwJaNPZ988sksNd1XrlzJySefnLGhanF16dKF3r1785///Ifzzz8/x7cCNm/ezMsvv8wbb7yRlvHSacyYMQCcd955zJkzJ8u5n3/+mddff71Y/aeeg7vvvjvL4z137tw869+PHj2aqlWrcv755/Pss8/mOL98+XJuv/121q5dW6zYShuT7pIkSZIkSdJvxOWXX065cuU477zzOPDAAxk8eDAtW7Zk8uTJOUqGbGs77bQTkydPpmbNmpx99tnsueeeHH/88Rx99NG0aNGC7t27Z6yu7tixI3379uXLL7+kVatWDBo0iEGDBtG8eXO++eYbBgwYkLa4HnzwQdq3b89NN91E06ZN6d27NyeccALdunWjYcOGHHroobz33ntpGy9dhg8fzllnncVXX31Fhw4d6N69O0OHDqVnz540btw417I8hTFq1CgaNmzIs88+S+vWrTn++OPp27cv++67L926daNp06Y52uy555488MADrF27lv79+9OmTRsGDRrEwIED2W+//WjUqBFnnnlm2jb7LS2s6S5JkiRJkqQyb/aI2SUdQpnQp08fnnnmGcaNG8cHH3xA9erV6dOnD+PHj2fSpEnbPZ5evXrx0UcfMX78eF566SWmTJlCrVq12GOPPfjjH/9Iw4YNM659+umnufrqq5k8eTIvvvgiDRo04IQTTuCqq65i9OjRaYtpl112YebMmfzjH//g0UcfZdasWaxfv55GjRqx3377MWDAAI477ri0jZdOf//73+nduzd33HEHs2bN4p133qFhw4b07t2bU089tVh916tXj1mzZnHhhRcybdo0nnnmGZo3b84VV1zBn//8Z/bYY49c2w0ePJiPP/6Yv/3tb7zyyiu8+OKLVKlShcaNGzNs2DCGDBlC7dq1ixVbaRNKQ6H/34IQwpy999577+xf7ZCKqt297Uo6hALxHz2SJEmSpOLYvHkzn332GQCtW7emXDkLN0gqvMK8l7Rt25a5c+fOjTG2LcpYvktJkiRJkiRJkpQmJt0lSZIkSZIkSUoTk+6SJEmSJEmSJKWJSXdJkiRJkiRJktLEpLskSZIkSZIkSWli0l2SJEmSJEmSpDQx6S5JkiRJkiRJUpqYdJckSZIkSZIkKU1MukuSJEmSJEmSlCYm3SVJkiRJkiRJShOT7pIkSZIkSZIkpYlJd0mSJEmSJEmS0sSkuyRJkiRJkiRJaWLSXZIkSZIkSZJ2MGPHjiWEwKRJk0o6lN+cCiUdgCRJkiRJklRcn7bZq6RDKJC95n1a0iH85oQQaNq0KYsWLSrpUPQb4Up3SZIkSZIkSZLSxKS7JEmSJEmSJElpYtJdkiRJkiRJKoMWLVpECIGePXvy888/8//+3/9jt912o2rVquy///48++yzGdc+9thjdO7cmerVq7PLLrtwzjnnsHbt2hx9fvXVV5x22mk0bdqUypUr06BBAwYPHsysWbPyjOPtt99mwIAB7LzzzlSuXJlmzZpxxhln8M033+S4dtKkSYQQGDt2LP/73/844YQT2GWXXShXrhxTpkzJuG727NkMGzaMXXfdlcqVK9O4cWNGjRpVqBIxqbEAFi9eTAgh46dnz55Fnnfmx33VqlWMHj2a5s2bU7FiRc477zwAmjVrljH2xIkTad++PVWrVqVhw4acdtpprFy5Mke/8+fPZ+zYsXTt2pWGDRtSqVIlmjRpwvDhw/nf//6X5zynTZtGz549qVGjBvXq1WPQoEHMmzcv38emKM+zCs6kuyRJkiRJklSGrV+/nt69e3P//fez77770qVLFz7++GMGDRrEq6++yoQJExg6dCgVKlSgX79+bNq0ib///e+ccsopWfqZPXs2+++/P3feeSfVqlVj8ODB7Lnnnjz11FMceOCBPPbYYznGfuCBB+jWrRvPPvssrVu3ZvDgwVSuXJk77riD/fffP8/k72effUanTp1499136dWrF3379qVixYoAPPHEE3Ts2JGHHnqIRo0a0b9/fxo2bMikSZPo2LEjc+bMKdDj0rJlS0aMGAFA9erVGTFiRMbPYYcdVqx5A6xdu5YePXpwzz33sO+++9K/f3/q1q2b5ZoLLriAM888k1q1anHYYYcRY+TOO++kf//+xBizXDtx4kTGjRvHqlWr6NixI/3796dWrVrcf//9dOrUif/+9785Ynj66afp3bs306ZNo0OHDvTr14///ve/HHDAAcyfPz/XuIs6XxVcyP7katsIIczZe++99y7om4K0Ne3ubVfSIRTI7BGzSzoESZIkSVIZtnnzZj777DMAWrduTblyua8h/S1upLpo0SKaN28OQM+ePXnyySczkr6TJk1i1KhRtGzZkuXLlzNlyhS6desGwDfffMN+++3HDz/8wIIFC2jRogUxRjp06MDs2bO56KKLuPrqqzNWaj/++OMcf/zxVK9enc8//5xddtkFSKyWbt26NRs2bOCpp57iqKOOAhLP2ejRo7npppsyEuspqbgAzjrrLG666SbKly+fcX7hwoXss88+VKxYkWeeeYbu3btnnLvvvvsYMWJEjj63Jr+NVIsy78yPe9euXXnhhReoU6dOln6bNWvG4sWLadSoES+88AL77rsvAEuXLqVr167Mnz+f1157jUMOOSSjzcyZM9l5553ZY489svR1zz338Ic//IFevXrx+uuvZxxfvXo1LVq0YOnSpTz00EOceOKJAGzcuJFTTjmFe++9N6P9yJEjizzfHUVB30sA2rZty9y5c+fGGNsWZSxXukuSJEmSJEllWPny5bnrrruyrLIePnw4O++8M/Pnz+ess87KSLgDNG7cmGHDhgEwffp0AKZOncrs2bNp3rw5V155ZUYiFuCYY45h4MCBrF69mnvuuSfj+MSJE1m7di0nnnhiRsIdoFy5clx77bU0btyYWbNmMXPmzBwx77zzzowfPz5Lwh3g5ptv5pdffuG6667LknBPzWngwIHMmjWLDz74oCgPVQ5FmXdmt9xyS46Ee2ZXXnllRsIdoH79+vzf//0fsOWxT+nSpUuOhDvAqFGjOOigg5g6dSo//fRTxvHHHnuMpUuX0rdv34yEO0CFChWYMGECNWrUSPt8VTAm3SVJkiRJkqQyrFmzZrRs2TLLsXLlytG0aVMA+vbtm6NNKrn77bffAjBjxgwAjj/++ByJcICTTjopy3WZf08l8DOrXLkyxx57bI42KX369KFatWo5jr/yyisADBgwIMc5gIMPPhggbbXHizLvlEaNGtGxY8d8++/Xr1+OY61atQK2PPaZrVmzhocffpgLL7yQU089lZEjRzJy5Ei+/fZbYowsWLAg49o33ngDgOOOOy5HP3Xr1s117OLMVwVXoaQDkCRJkiRJklR0u+66a67Hq1evnuf51Ll169YBZGx62qxZs1z7Sh3PvDlqUdqk7L777rm2SZWAadiwYa7nU5YuXQokEs8TJ07Mcf6GG26gfv36+faRObZ0ziGzJk2a5DiWWoGeeuxTXn/9dU444QR+/PHHPPtbvXp1xu+pmPKKI7fjxZmvCs6kuyRJkiRJklSGZS4RUpTzxe2rKG2qVKmS67WbNm0ihMDw4cPz7bNt20Sp7fnz52fULs9s7NixBUq65xfj1s7nNYfC9JuyZs0ajjvuOJYtW8Zll13GiSeeSNOmTalatSohBIYOHcrDDz+cZfPV1O+FeX4LGldR+tQWJt0lSZIkSZKk37jGjRsDiY1Mc7N48WIgUVIlc5vPPvuMhQsXZpRM2VqbrWnSpAkLFizglltuoVatWlu9PlV+paiKMu9tYcaMGSxbtowhQ4ZwxRVX5Dj/xRdf5DiWij0VY3Zffvllnm1Ker47Omu6S5IkSZIkSb9xqY1WH330UTZt2pTj/AMPPJDlusy/P/jggzmuX79+PY899liONlvTp08fAKZMmVLgNltTsWJFNm7cmOu5osx7W1ixYgUAu+22W45z8+fPz3Xj2FR9+9TjnNnKlSt5+eWXcxwvLfPd0Zl0lyRJkiRJkn7jevbsSbt27Vi4cCGXX355ljImU6ZM4cknn6RGjRpZVpWffPLJVK1alYcffpjnn38+4/jmzZu5+OKLWbJkCZ06daJLly4FjmP06NFUrVqV888/n2effTbH+eXLl3P77bezdu3aAvfZuHFjvv/+e1auXJnjXFHmvS2kvinw5JNPZqnpvnLlSk4++WQ2bNiQo82xxx7LTjvtxMsvv8zkyZMzjm/atInRo0ezZs2aHG1Ky3x3dCbdJUmSJEmSpN+4EAIPPvgg9erV469//Stt27Zl6NChHHzwwQwaNIhy5crxr3/9K8sGp7vvvjt33nknMUaOPvpounXrxtChQ9l7773529/+xi677MJ9991XqDj23HNPHnjgAdauXUv//v1p06YNgwYNYuDAgey33340atSIM888M8cmpPnp378/GzduZP/99+f3v/89p5xyCtdff32R570tdOzYkb59+/Lll1/SqlUrBg0axKBBg2jevDnffPMNAwYMyNGmVq1a3HnnnZQrV47jjz+egw8+mKFDh9K6dWsef/xxhg0blqNNaZnvjs6kuyRJkiRJkiTatWvHBx98wKmnnsqaNWt4/PHH+eyzzxg4cCBvvvkmxx57bI42v//975k+fTpHHXUUn376KY8//jhr167l//7v/3j//fdp06ZNoeMYPHgwH3/8MaeddhobNmzgxRdfZOrUqaxbt45hw4bx3HPPUbt27QL3d80113DWWWexceNGHn30Ue6+++4sK/OLMu9t4emnn+aSSy5h55135sUXX+T999/nhBNOYObMmdSpUyfXNkOGDOGVV16hW7dufPjhh7z44ovsvffevP3227Rs2TLXNqVlvjuykPkrBNp2Qghz9t57773nzJlT0qFoB9Hu3nYlHUKBzB4xu6RDkCRJkiSVYZs3b+azzz4DoHXr1pQr5xpSSYVXmPeStm3bMnfu3LkxxrZFGct3KUmSJEmSJEmS0sSkuyRJkiRJkiRJaWLSXZIkSZIkSZKkNDHpLkmSJEmSJElSmph0lyRJkiRJkiQpTUy6S5IkSZIkSZKUJibdJUmSJEmSJElKE5PukiRJkiRJkiSliUl3SZIkSZIkSZLSxKS7JEmSJEmSJElpYtJdkiRJkiRJkqQ0MekuSZIkSZIkSVKamHSXJEmSJEmSJClNTLpLkiRJkiRJkpQmFUo6AEmSJEmSJKm4bjv99ZIOoUDO/MchaesrhEDTpk1ZtGhR2vosiEWLFtG8eXN69OjB1KlTM45PnTqVXr16MWLECCZNmrRdY/qtmDRpEqNGjeIvf/kLY8eOLelwlAdXukuSJEmSJEkqU5o1a0YIoaTDkHLlSndJkiRJkiSpDPr000+pWLFiSYeRoXPnznz66afUrl27pEORSpRJd0mSJEmSJKkMatOmTUmHkEW1atVKXUxSSbC8jCRJkiRJklQGhRBo1qxZjuMvvPACffv2pW7dulSpUoXWrVszZswYVq5cmePasWPHEkJg0qRJvPPOOxx66KHUqVOHWrVq0bdvX2bOnFngeKZOnUoIgZEjR+Y5xuzZs+nfvz9169alevXq9OjRg7feeqvQYyxevDjjMUj9ZH8sli1bxp///Gf23HNPqlSpwk477cRhhx3Gyy+/nGvfqT7Wr1/PFVdcQZs2bahcuTIDBw4EoGfPnoQQWLRoEVOmTKFLly5Ur16dnXbaiRNPPJGvv/46R5/ffvst1113HT169GDXXXelUqVKNGzYkMGDBzNr1qw85/nf//6Xo446itq1a1O7dm369u3L22+/ne9jU9j5atsx6S5JkiRJkiTtIK655hqOPPJIpk6dyu9+9zsGDhzIL7/8wvjx4znggAP4/vvvc2331ltv0b17d77++msOP/xwWrduzauvvkqPHj145ZVX0hLbe++9R5cuXfjss8/o3bs3e+65J9OnT6d379588sknBeqjYcOGjBgxgurVqwMwYsSIjJ9jjjkm47olS5bQuXNnbrjhBtavX8/AgQPZb7/9ePXVVzn00EOZMGFCrv1v3ryZgQMHct1117HHHnswYMAAGjVqlOWa22+/nSFDhhBj5LDDDqNGjRo88sgjHHLIIaxduzbLtU8//TQXXngh33zzDe3atWPgwIE0btyYp556ioMOOijXhPg777xD165def7552nRogVHHHEE3333Xb4fUBR1vto2LC8jSZIkSZIk7QBmzZrFpZdeSs2aNXn11Vfp3LkzAOvWreOkk07iscce4+yzz2by5Mk52t51111cfPHFXHXVVRkblN5xxx2cccYZjBw5kgULFlClSpVixXfbbbcxfvx4Lrjggoxj559/PjfddBPXXXcd991331b7aNOmDZMmTWLq1Kn8/PPPTJo0KdfrTj/9dL744gtOOukk7r777oza92+88QaHHnoof/7zn+nduzft27fP0u6rr76icuXKfPbZZ+y666659n377bfzyiuvcMghhwDwyy+/0LdvX9566y0efvhh/vCHP2Rce9BBB/Hxxx/nGOell16if//+nHHGGXz++ecZj/nmzZsZOXIkv/zyC9dccw1jxozJaHPZZZdx1VVXpXW+2jZc6S5JkiRJkiTtAG699VY2b97Meeedl5FwB6hcuTK33norVatW5YknnmDJkiU52jZt2pRx48ZlJH8B/u///o8DDjiAb775hqeeeqrY8R188MFZEu4Al156KQDTp08vdv8pX3zxBc899xy1atXilltuybLZ7MEHH8zpp5/Opk2buP3223Ntf8011+SZcIfEBwWphDskatmPHj0ayDmPdu3a5ZroPvTQQzn22GNZsGBBllX+U6dOZd68ebRq1YoLL7wwS5u//OUv7L777mmfr9LPpLskSZIkSZK0A5gxYwYAw4YNy3GuQYMG9OvXj82bN+daomTIkCFUqJCzKMaJJ54IJFZMF1e/fv1yHKtXrx716tXj22+/LXb/KalYjzjiCOrUqZPj/EknnQRsebwyCyFw9NFH59t/bvNo1aoVQK7zWLduHU8//TSXXHIJf/zjHxk5ciQjR45k9uzZAHz++ec5Yj/22GOzfAACUKFChSwldLK3Kcp8tW1YXkaSJEmSJEnaAXzzzTeEEGjatGmu51MbjX7zzTc5zhWlTWE1adIk1+M1atRg2bJlGffnzZvHtddem+O6MWPG0KZNm62Ok4o1t01mMx/PbU4NGjSgcuXK+faf2zxq1KgBJBLsmaU2jl20aFGe/a1evTrj91RMua1oz+t4cearbcOkuyRJkiRJkvQbkn0FdX5ijNt93O+++4577703x/GRI0cWKOm+tfFSx3M7X5C69QWdR4yR4447jkWLFnH66adz+umn06JFC2rUqEEIgYsvvphrrrkmy2Oc+r0wz9HW4spvvto2TLpLkiRJkiRJO4DGjRuzcOFCFi9eTOvWrXOcX7x4MQCNGjXK81x2X375ZUbf20vPnj2LlexPxbpw4cJcz6dWnef2OKTTvHnzmDdvHh07duSOO+7Icf6LL77IcSwV+9aej9zalPR8tYU13SVJkiRJkqQdQLdu3QB48MEHc5z78ccfefnllylXrhwHHnhgjvNPPPEEmzZtynH8kUceAeCggw5Kc7TFU6lSJQA2btyY49zBBx8MwPPPP8/KlStznH/ggQeALY/XtrJixQog93I0K1as4JVXXslxPBX7E088keODh40bN/LEE0/k2aak56stTLpLkiRJkiRJO4AzzzyTcuXKcfPNN/Pee+9lHF+/fj1nn302v/zyC4MHD2bXXXfN0Xbx4sWMGzcuy7E777yTt99+m4YNGzJo0KBtHn9hpFZ3f/bZZznOtWjRgiOPPJLVq1dz7rnnsmHDhoxzb7/9NnfccQfly5fnjDPO2KYxtmzZknLlyvH6669n2Sz1119/5fTTT2f58uU52vTq1YtWrVoxb948brjhhiznrrrqqlxXwJeW+WoLy8tIkiRJkiRJO4DOnTtz5ZVXcskll9C1a1d69uxJ/fr1efPNN/nqq6/Yc889ufXWW3Nte+qpp3Lttdfy5JNP0r59e+bPn8+sWbOoWLEi99xzD1WrVt3Os8lf//79mTZtGr1796ZXr15Ur16d+vXrZ2zA+s9//pNu3bpx3333MW3aNLp27cqPP/7I1KlT2bRpE3/7299o3779No2xQYMGnHzyydx111106NCBQw45hKpVqzJjxgw2bdrEyJEjmTRpUpY25cqVY9KkSfTu3ZsLLriAhx9+mDZt2vDJJ58wb948TjnlFCZOnJhjrNIwX21h0l3K5tM2e5V0CAVzkX++kiRJkiSlnPmPQ0o6hBKRfXPMiy++mA4dOjBhwgRmzZrF2rVr2X333bngggsYM2YMdevWzbWfAw88kFGjRnHZZZfx3HPPEWOkd+/eXHHFFbmWoylp55xzDitWrODhhx/miSeeYMOGDTRt2jQj6b7rrrsya9YsrrnmGqZMmcKTTz5JtWrV6N27N6NHj6Zfv37bJc477riDNm3acPfdd/Paa69Ru3Zt+vTpw9VXX80999yTa5uuXbvy1ltvcfHFF/PGG28wf/58OnXqxB133MHnn3+ea9K9tMxXCSGdOxArbyGEOXvvvffec+bMKelQtBVlJel+XBlJus8eMbukQ5AkSZIklWGbN2/OKCHSunVrypWzWjLA2rVrqVatGnvvvTfFyTeNHTuWcePGcc899zBy5Mj0BSiVMoV5L2nbti1z586dG2NsW5SxfJeSJEmSJEmSyphUzfY99tijhCORlJ1Jd0mSJEmSJKmM+OCDDzj22GM5+uijARg2bFgJRyQpO5PukiRJkiRJUhnx5Zdf8tRTT7HTTjtxww03cPzxx5d0SJKyKRtFoSVJkiRJkiQxcOBANm7cmLb+xo4dy9ixY9PWnyRXukuSJEmSJEmSlDYm3SVJkiRJkiRJShOT7pIkSZIkSSq1QggZv2/evLkEI5FUlm3atCnj98zvK9uCSXdJkiRJkiSVWiEEKlWqBMDPP/9cwtFIKqtWrVoFQOXKlbd50t2NVCVJkiRJklSq1axZk2XLlvH9998DUL16dcqVcy2ppPzFGFm3bh2rV69m+fLlANStW3ebj2vSXZIkSZIkSaVavXr1+Pnnn/n111/55ptvSjocSWVUnTp1qF279jYfx6S7JEmSJEmSSrXy5cuz++67s2zZMlavXs369etLOiRJZUT58uWpXr06NWvWpGbNmtu8tAyYdJckSZIkSVIZUL58eRo0aECDBg2IMRJjLOmQJJVyIYTtkmTPrkwm3UMI1YB+wNFAJ6AZUB6YDzwB3BhjXFPIPusAY4FBQEPgO2AK8JcY48q0BC5JkiRJkqRiK6lEmiQVRFndcWIo8BTwBxJz+DcwA2gOjANmhRAaFLSzEEI94F3gXGAjiWT7auCcZF/10hm8JEmSJEmSJGnHVFaT7uuBO4BWMcZ9YozHxRgPA1oDHwJtgJsK0d8EYE/gSaB1jPH4GOM+wN+BlsCN6QxekiRJkiRJkrRjKpNJ9xjjfTHGM2KMn2c7/i1wZvLu4BBCpa31FUJoCAwDNgBnxBg3Zjr9Z+BHYFgIYZf0RC9JkiRJkiRJ2lGVyaT7VnycvK0MFKQszOEkHofpMcbvM5+IMa4DniVRL/7wdAYpSZIkSZIkSdrx7IhJ9xbJ2w3A8gJc3yF5+0Ee5z/Idp0kSZIkSZIkSbnaEZPu5yZv/51cqb41uydvv87j/NfZrpMkSZIkSZIkKVcVSjqAdAohHAGcTGKV+2UFbFYjeftLHud/znbd1mKYk8epPQoYjyRJkiRJkiSpjNphVrqHEPYCHgAC8OcY48dbaZLRNHkbt3JekiRJkiRJkqR87RAr3UMITYB/A3WBG2OMNxei+erkbfU8zldL3q4pSGcxxrZ5xDgH2LsQcUmSJEmSJEmSypgyv9I9hFAfeIVEzfV7gD8Vsosvk7dN8jjfJNt1kiRJkiRJkiTlqkwn3UMINYEXgTbAk8CpMca8ysTkJVWGZv88zqeO/7fwEUqSJEmSJEmSfkvKbNI9hFAZeBroCLwEnBhj3FSErv4NbAa6hRAa5DLG0cnzLxYvYkmSJEmSJEnSjq5MJt1DCOWBh4FewAxgcIxx/VbanBVCmBdCuCbz8Rjjt8m+KgG3hxAy17m/DtgZeCjG+F065yBJkiRJkiRJ2vGU1Y1UzwIGJX9fSiJZntt1f4oxLk3+Xh9oDTTK5brzgC7AEGBeCOE9oC2wD7AAOD9tkUuSJEmSJEmSdlhlNeleN9Pvg/K8CsaSSMrnK8a4NITQCRgHDEz2+T1wK/CXGOPyIkcqSZIkSZIkSfrNKJNJ9xjjWBIJ9bS1iTGuAM5J/kiSJEmSJEmSVGhlsqa7JEmSJEmSJEmlkUl3SZIkSZIkSZLSxKS7JEmSJEmSJElpYtJdkiRJkiRJkqQ0MekuSZIkSZIkSVKamHSXJEmSJEmSJClNTLpLkiRJkiRJkpQmJt0lSZIkSZIkSUoTk+6SJEmSJEmSJKWJSXdJkiRJkiRJktLEpLskSZIkSZIkSWli0l2SJEmSJEmSpDQx6S5JkiRJkiRJUpqYdJckSZIkSZIkKU1MukuSJEmSJEmSlCYm3SVJkiRJkiRJShOT7pIkSZIkSZIkpYlJd0mSJEmSJEmS0sSkuyRJkiRJkiRJaWLSXZIkSZIkSZKkNDHpLkmSJEmSJElSmph0lyRJkiRJkiQpTUy6S5IkSZIkSZKUJibdJUmSJEmSJElKE5PukiRJkiRJkiSliUl3SZIkSZIkSZLSxKS7JEmSJEmSJElpYtJdkiRJkiRJkqQ0MekuSZIkSZIkSVKamHSXJEmSJEmSJClNTLpLkiRJkiRJkpQmJt0lSZIkSZIkSUoTk+6SJEmSJEmSJKWJSXdJkiRJkiRJktLEpLskSZIkSZIkSWli0l2SJEmSJEmSpDQx6S5JkiRJkiRJUpqYdJckSZIkSZIkKU1MukuSJEmSJEmSlCYm3SVJkiRJkiRJShOT7pIkSZIkSZIkpYlJd0mSJEmSJEmS0sSkuyRJkiRJkiRJaWLSXZIkSZIkSZKkNDHpLkmSJEmSJElSmph0lyRJkiRJkiQpTSqkq6MQQiXgQKAb0AHYGagN/AT8CHwEzADejjGuT9e4kiRJkiRJkiSVFsVOuocQ9gJOA34P1AVCHpcOBiKwMoRwH3BnjPHT4o4vSZIkSZIkSVJpUeSkewihCXAViWR7OeBL4HngXWAesBxYRWK1e11gL6Az0AM4Fzg7hHA/cFmM8etizEGSJEmSJEmSpFKhOCvd/5e8vQt4IMb45laufy31SwjhYOCk5M+xQI1ixCFJkiRJkiRJUqlQnI1U7wRaxBj/rwAJ9yxijG/EGE8D9iCRtJckSZIkSZIkqcwr8kr3GON5xR08xvgtcH5x+5EkSZIkSZIkqTQozkp3SZIkSZIkSZKUSXFquhdYCOFQoB2JzVafjDFu3B7jSpIkSZIkSZK0PaVtpXsI4YwQwhfJTVIzH38YeAEYDzwMTA8hVE7XuJIkSZIkSZIklRbpLC8zCKgOvJU6EELoCxwPLAGuBd4FDgBOTuO4kiRJkiRJkiSVCuksL9Ma+CTGuDnTsaFABI6JMb6bXOH+JTAcuD2NY0uSJEmSJEmSVOLSudJ9Z+DbbMe6A1/GGN8FiDGuI7ESvnkax5UkSZIkSZIkqVRIZ9J9JVAndSeE0IhEcn1atut+BmqkcVxJkiRJkiRJkkqFdCbdPwcODiHUTt4fRqK0zL+zXdcE+C6N40qSJEmSJEmSVCqkM+l+O1ALeD+E8CRwNfAj8FzqghBCVaAjMDeN40qSJEmSJEmSVCqkbSPVGOMjIYR9gXOBFsDXwIgY45pMlx0HVANeT9e4kiRJkiRJkiSVFmlLugPEGMeEEP4C1Iox/pjLJa8D+wEL0jmuJEmSJEmSJEmlQVqT7gAxxnUkysrkdu4r4Kt0jylJkiRJkiRJUmmQzprukiRJkiRJkiT9phV5pXsI4V/FGDfGGE8uRntJkiRJkiRJkkqd4pSXGVmMthEw6S5JkiRJkiRJ2qEUJ+neK21RSJIkSZIkSZK0Ayhy0j3GOC2dgUiSJEmSJEmSVNa5kaokSZIkSZIkSWlSnPIyuQohVACOAjoB9YF3Yoz/Sp5rnDw2N8a4Md1jS5IkSZIkSZJUktKadA8h9ADuB3YFAokNUysC/0pe0huYBBwHPJHOsSVJkiRJkiRJKmlpKy8TQmgHvAA0AG4GjiWReM/sCeAXYEi6xpUkSZIkSZIkqbRI50r3y4HKQL8Y4+sAIWTNuccYfwkhfArsl8ZxJUmSJEmSJEkqFdK5kWoPYGYq4Z6PL4HGaRxXkiRJkiRJkqRSIZ1J91rAkgJcVxkon8ZxJUmSJEmSJEkqFdKZdP8W2KsA1+0DLE7juJIkSZIkSZIklQrpTLq/DLQNIQzK64IQwkigKfB8GseVJEmSJEmSJKlUSGfS/a/AGuDhEMKVIYSOyePVQgj7hBAuBW4HlgE3pnFcSZIkSZIkSZJKhbQl3WOMi4EjgRXAJcA7QASOBT4GrgBWAwNijN+la1xJkiRJkiRJkkqLCunsLMb4RgihFXAy0AdoRmLT1K+BV4F/xhhXpnNMSZIkSZIkSZJKi7Qm3QFijKuBm5I/kiRJkiRJkiT9ZqStvEwIYUgBr6sYQrg+XeNKkiRJkiRJklRapHMj1cdCCHeHEKrndUEIYW/gXeD/pXFcSZIkSZIkSZJKhXQm3T8FRgEfhRAOyH4yhHA2MAvoADyexnElSZIkSZIkSSoV0pl0/x1wO7AHMCOE8JcQQrkQwi4hhBdJ1HjfAIyIMR6fxnElSZIkSZIkSSoV0pZ0jzH+GmM8CzgSWA5cTqKUzGzgUOANoEOM8f50jSlJkiRJkiRJUmmSzpXuAMQYXwS6AKuB/YB6wEMxxh4xxsXpHk+SJEmSJEmSpNIi7Un3EEJH4N9ALeCb5OHjQgjjQgjl0z2eJEmSJEmSJEmlRdqS7iHhUuBNoBWJ+u57AocDS4FLgbdCCC3TNaYkSZIkSZIkSaVJOle6TwfGASuAo2KMZyXrvL8M7AM8DXQCPgwh/DGN40qSJEmSJEmSVCqkM+l+EPA80C7G+ELmEzHG5THGwcAfgQDckcZxJUmSJEmSJEkqFdKZdD8zxtg/xvhjXhfEGCcC+wKz0jiuJEmSJEmSJEmlQtqS7jHGAq1ejzHOJ7EqXpIkSZIkSZKkHUo6V7oXWIxxU0mMK0mSJEmSJEnStlShqA1DCN2Tv74bY/w10/0CiTFOL+rYkiRJkiRJkiSVRkVOugNTgQjsBfwv0/2CKl+MsSVJkiRJkiRJKnWKk3S/j0SS/ads9yVJkiRJkiRJ+k0qctI9xjgyv/uSJEmSJEmSJP3WlMhGqpIkSZIkSZIk7YiKU14mXyGEekDj5N1vY4xLt9VYkiRJkiRJkiSVBmld6R4SzgohfAr8AHyU/Pk+hPBpCOHsEIKr6yVJkiRJkiRJO6S0rXQPIVQGngV6AwFYASxO/r470Bq4CegfQjgqxrguXWNLkiRJkiRJklQapHPV+cVAH2AOcHiMsV6Mcf8Y434xxnrA4cAnwCHJayVJkiRJkiRJ2qGkM+n+e2Al0CvG+FL2k8ljvYGfgJPSOK4kSZIkSZIkSaVCOpPujYHXYozL8roguZnq60CjNI4rSZIkSZIkSVKpkM6k+xKgUgGuqwh8k8ZxJUmSJEmSJEkqFdKZdH8Q6B1CaJrXBclzvYGH0jiuJEmSJEmSJEmlQjqT7lcBrwHTQwh/CCFUT50IIVQPIfwBmJa85oo0jitJkiRJkiRJUqlQIY19fQYEoAlwF3BXCGFF8lzdTNdF4LMQQua2Mca4R2EGCyH8DugLdAYOIFFTfl2MsUphAw8hLALyXKEP7BVjnFfYfiVJkiRJkiRJvy3pTLo3y+XYTrkcyy+5XRiXAQPS1FfKvXkc/ynN40iSJEmSJEmSdkBpS7rHGNNZqqYg3gY+BmYlf74rbocxxpHF7UOSJEmSJEmS9NuVzpXu21WMcXzm+9nK1UiSJEmSJEmStN1t79XpkiRJkiRJkiTtsIq80j2E0CYdm4umq590CCH8GdgDWAfMAZ6KMf5YslFJkiRJkiRJksqK4pSX+SSE8ChwTYzxk8I2DiHsC4wBhgAVixFHOl2X7f6EEMI5Mca7C9pBCGFOHqf2KHpYkiRJkiRJkqSyoDjlZa4EjgY+DiF8EEIYHULoGELINYEeQqgcQugSQrgohDAbeB84AriiGDGkyzPAYKApUA3YB7gRqAxMDCEMLLnQJEmSJEmSJEllRZFXuscYx4UQ7gAuAYYD1wMR2BBCWASsAFYDtYCdSCS0KwAB+Am4mcQq+RIv3xJjPCfboTnA6BDCZ8A/gfHAlAL21Ta348kV8HsXI0xJkiRJkiRJUilXnPIyxBh/AM4NIYwBjgOOAg4CWuVy+XfADOB5YHKM8dfijL2dTCSxor9VCKF5jHFhSQckSZIkSZIkSSq9ipV0T4kxrgXuTf4QQtgZaADUJrGq/YfSsKK9sGKMm0MIC0jMpRFg0l2SJEmSJEmSlKe0JN2zSybYy1ySPQ91k7drSjQKSZIkSZIkSVKpV5yNVHd4IYS2QGvgF2BeCYcjSZIkSZIkSSrlfjNJ9xDCWSGEeSGEa7IdPzSE8Ltcrm8PPEZi49eJMcb12ylUSZIkSZIkSVIZtU3Ky2wPIYQjgcuyHa4UQpiZ6f6VMcbnk7/XJ7FqvVG2Nl2Bv4QQFgMLSJTFaQ7sT+LxmQZclObwJUmSJEmSJEk7oDKbdAd2Bg7IdixkO7ZzAfp5CdgN6AR0ILH56yrgDeBB4J4Y46ZiRytJkiRJkiRJ2uGV2aR7jHESMKkQ148FxuZy/G3g7TSFJUmSJEmSJEn6DfvN1HSXJEmSJEmSJGlbM+kuSZIkSZIkSVKamHSXJEmSJEmSJClN0l7TPYRQDegF7AnUJLG5aXYxxnhluseWJEmSJEmSJKkkpTXpHkIYCUwAamU+DMRc7pt0lyRJkiRJkiTtUNJWXiaE0Ae4m0RC/a/A28lTpwHXA/NJJNxvBf6QrnElSZIkSZIkSSot0lnTfTSJhHuvGONlwOcAMca7YoxjgL2Bm0gk3N9P47iSJEmSJEmSJJUK6Uy6dwJmxhg/zu1kjHET8CfgB2BcGseVJEmSJEmSJKlUSGfSvQbwZab7vwKEEGqmDsQYNwPvAN3SOK4kSZIkSZIkSaVCOpPu3wH1s90HaJXtup2AqmkcV5IkSZIkSZKkUiGdSfd5ZE2wv0Vi49QLQwgBIIRwIHAI8Fkax5UkSZIkSZIkqVRIZ9L9eWD3EEKX5P3XgP8CQ4AlIYT3gf8kx7wpjeNKkiRJkiRJklQqpDPpfh9wOPAtZNRvPxJ4BWgA7Af8AlwaY3wgjeNKkiRJkiRJklQqVEhXRzHGn4CXsh1bAhwWQqgG1AZ+iDFuSteYkiRJkiRJkiSVJmlLuucnxvgLiVXukiRJkiRJkiTtsLZJ0j2EsC/QCagPzIkxPpM8XhmoHGNctS3GlSRJkiRJkiSpJKWzpjshhL1CCDOB94F/AFcBAzNd8gdgRQjhsHSOK0mSJEmSJElSaZC2pHsIoSkwHegMPA1cAIRslz0CbACGpGtcSZIkSZIkSZJKi3SudP8LsBMwIsY4OMb4t+wXxBhXAHOBrmkcV5IkSZIkSZKkUiGdSfdDgQ9jjPdv5brFQOM0jitJkiRJkiRJUqmQzqR7PWBBAa6LQJU0jitJkiRJkiRJUqmQzqT7UqB5Aa7bC1iSxnElSZIkSZIkSSoV0pl0nwb8LoRwUF4XhBCOAloDr6RxXEmSJEmSJEmSSoV0Jt2vATYAz4YQTg4h7Jw6EUKoEUL4PXAP8AuQY5NVSZIkSZIkSZLKurQl3WOMnwDDgErAncB3JOq3Dwd+Au4FqgEnxRgLUvtdkiRJkiRJkqQyJZ0r3YkxPgHsA/wd+BRYS2L1+xckEvEdYoxPpXNMSZIkSZIkSZJKiwrp6iiE0B/YEGN8ETgvXf1KkiRJkiRJklRWpHOl+1PAOWnsT5IkSZIkSZKkMiWdSfcfgRVp7E+SJEmSJEmSpDIlnUn3qUDnEEJIY5+SJEmSJEmSJJUZ6Uy6XwbUByaEEKqksV9JkiRJkiRJksqEtG2kCpwIvACcDZwQQngV+BL4NZdrY4zxyjSOLUmSJEmSJElSiUtn0n0sEIEANACG5nNtBEy6S5IkSZIkSZJ2KOlMuo9KY1+S9P/bu+8wSaqyDeP3u7tkFAwklaCgZPiM5CBiBEUEEUUFTIAEFUSC5JwEEVDMAiqoiIggioIoIhIkR5EoCojkJbP7fn+c00vRzsCyW9s9PXP/rmuvnumqrj6tzOmqp855jyRJkiRJkjRwWgvdM/O4to4lSZIkSZIkSdIganMhVUmSJEmSJEmSxjRDd0mSJEmSJEmSWtJmTXciIoBNgPWA1wIvoiys2i0zc9E231uSJEmSJEmSpH5rLXSPiJmBM4C1GDpoB8jn2CZJkiRJkiRJ0kBrs7zMDsDbgNMpo9xPoITsswBLAnsBjwCHZqZlbSRJkiRJkiRJo06b5WU+BNwHfCQzH4mIyQCZ+RRwA7BPRPwB+ENE3JCZ32vxvSVJkiRJkiRJ6rs2R5wvBlyUmY/U3ycDRMT4zg6ZeR5wPvDZFt9XkiRJkiRJkqQRoc3QfRLwUOP3Tvg+T9d+/wIWb/F9JUmSJEmSJEkaEdoM3f8FLNT4/R/1ccWu/ZYDJrb4vpIkSZIkSZIkjQhthu5/BZaOiNnq77+uj0dGxLsjYtmIOIqyqOqFLb6vJEmSJEmSJEkjQpuh+8+BR4G3A2TmP4CvAgsCpwOXA1vXfXZq8X0lSZIkSZIkSRoRJrR1oMw8A1ig67kdIuJi4P3AS4C/A1/LzBvbel9JkiRJkiRJkkaK1kL34WTmScBJM/p9JEmSJEmSJEnqtzbLy0iSJEmSJEmSNKYZukuSJEmSJEmS1BJDd0mSJEmSJEmSWmLoLkmSJEmSJElSS2b4QqqSxri95up3C6beXg/2uwWSJEmSJEkacI50lyRJkiRJkiSpJYbukiRJkiRJkiS1xNBdkiRJkiRJkqSWtFbTPSL2mMpdnwTuBS7PzIvben9JkiRJkiRJkvqtzYVU9wKy67moj9n1XAJExI3ApzPzvBbbIUmSJEmSJElSX7QZum8OrAhsAdwCnALcTgnZFwQ+ALwa+FZ9fjXgncCZEbFCZl7TYlskSZIkSZIkSeq5NkP3a4Bjgd2BAzNzcnNjRHwJ2KVuXyMzD4iIbYEjgZ2Aj7fYFkmSJEmSJEmSeq7NhVT3A67PzP27A3eALA4Argf2rc8dBdwMrNliOyRJkiRJkiRJ6os2Q/cVgWunYr9rgRUav18NzNtiOyRJkiRJkiRJ6os2Q3eAxadhn0nAYy23Q5IkSZIkSZKknmszdL8AeH1EbD/cDhHxBeANwF8aT78G+HeL7ZAkSZIkSZIkqS/aXEh1V2B14NCI+ATwc+CfQAILARsAS1JGtX8ZICIWApYDvtFiOyRJkiRJkiRJ6ovWQvfMvCwi3g58H1iq/su6OerjP4DNM/Oy+vvDwMqUxVQlSZIkSZIkSRpobY50JzP/EhFLAGtRwvRX1E13UkrKnJOZkxv73w9c2GYbJEmSJEmSJEnql1ZDd4DMTODs+k+SJEmSJEmSpDGjzYVUJUmSJEmSJEka01of6R4RrwZWAxYAZhlmt8zMfdt+b0mSJEmSJEmS+qm10D0iZga+A2zSeeo5dk/A0F2SJEmSJEmSNKq0OdJ9H+CjwP3AD4G/AxNbPL4kSZIkSZIkSSNam6H7R4AHgDdk5m0tHleSJEmSJEmSpIHQ5kKq8wLnGbhLkiRJkiRJksaqNkP324A5WjyeJEmSJEmSJEkDpc3Q/bvAWyJiwRaPKUmSJEmSJEnSwGgzdD8MOAM4MyLWjIho8diSJEmSJEmSJI14bS6k+o/6uDBwNvBURNwJ5BD7ZmYu2uJ7S5IkSZIkSZLUd22G7ot0/T4zJYCXJEmSJEmSJGlMaC10z8w2S9VIkiRJkiRJkjRwDMolSZIkSZIkSWqJobskSZIkSZIkSS2Z5vIyEbFQ/fFfmTmp8ftUyczbp/W9JUmSJEmSJEkaiaanpvutwGRgKeDv9fecytfmdL63JEmSJEmSJEkjzvQE33+ihOePdv0uSZIkSZIkSdKYNM2he2au+Vy/S5IkSZIkSZI01riQqiRJkiRJkiRJLTF0lyRJkiRJkiSpJdNcXiYiPj49b5yZx0/P6yVJkiRJkiRJGmmmZyHVHzBtC6dGfZ2huyRJkiRJkiRpVJme0H0f/jd0XwzYBJgInAXcXp9fCHgHMCfwQ+Cm6XhfSZIkSZIkSZJGpGkO3TNzr+bvEbE4cCFlBPz2mflA1/a5gMOBDwArTev7SpIkSZIkSZI0UrW5kOqBwD3Ap7oDd4DMfBD4NPDfuq8kSZIkSZIkSaNKm6H76sBFmTl5uB3qtovqvpIkSZIkSZIkjSpthu6zUGq3P5+FgJlbfF9JkiRJkiRJkkaENkP3vwErR8RGw+0QER8EVqn7SpIkSZIkSZI0qkzzQqpD2BP4PXBiRGwO/Ay4HUhgYeCDwDuAScBeLb6vJEmSJEmSJEkjQmuhe2b+MSI2BL4DvJMSsDcFcB/wmcw8t633lSRJkiRJkiRppGhzpDuZ+cuIOJsyqn1V4BWUsP3fwJ+Bn2Xmw22+pyRJkiRJkiRJI0WroTtAZk4Evl//SZIkSZIkSZI0ZrS5kKokSZIkSZIkSWPaNI90j4iFpueNM/P26Xm9JEmSJEmSJEkjzfSUl7kVyGl8bU7ne0uSpGEse9yy/W7CVLlq06v63QRJkiRJklo3PcH3n5j20F2SJEmSJEmSpFFnmkP3zFyzxXZIkiRJkiRJkjTwXEhVkiRJkiRJkqSWGLpLkiRJkiRJktSSaS4vExGr1x8vyszHG79Plcz807S+tyRJkiRJkiRJI9H0LKR6LmUh1SWBvzd+n1rjp+O9iYg3Am8H3gKsALwCeCIzZ53G480N7AWsD8wP3AWcCuyZmQ9MT1slSZIkSZIkSWPD9ITux1NC9ge7fu+V3YH12jhQRLwMuAB4LXAzJWxfGtgOeE9ErJiZ97bxXpIkSZIkSZKk0WuaQ/fM3Oy5fu+BC4ArgIvrv7um41hHUAL3U4APZebTABHxNWBb4HBg0+lqrSRJkiRJkiRp1Jueke59lZkHN3+PiGk6TkTMD2wCPAV8thO4VzsCGwObRMSXMvPuaWyuJEmSJEmSJGkMGNfvBowA76b87/Cn7lA9M58AfkWpP//uPrRNkiRJkiRJkjRAWhvpHhF7vIDdMzP3beu9p9Py9fHSYbZfCnyisZ8kSZIkSZIkSUNqs7zMXpSFVIer89JZZDXqzyMldF+oPt4xzPY7uvZ7ThFxzTCbFn0hjZIkSZIkSZIkDZ42Q/fNh3l+HLAg8E5gJeAY4JIW33d6zVkfHx1m+yNd+0mSJEmSJEmSNKTWQvfMPO55dtknInYBvgx8q633bUFnZH4+z/apkplLD3mQMgJ+qRdyLEmSJEmSJEnSYOnpQqqZeSClXMsBvXzf5/FwfZxjmO2z18eJPWiLJEmSJEmSJGmA9TR0r64CVu3D+w7n9vr4qmG2v6prP0mSJEmSJEmShtSP0H1R2q0lP72uqI9vGGZ75/kre9AWSZIkSZIkSdIA61noHhFzR8RXgP8DLurV+06F3wCTgdUiYt7mhoiYBXhv3X5mH9omSZIkSZIkSRogrY04j4ibn2PznMDLKIuSPgbs0tb7Tq2I2AbYBvhFZk55/8y8MyJOBDYBvh4RG2fm03XzIcA8wA8z865et1mSJEmSJEmSNFjaLPOyyHNsewr4J/BH4ODMvHZ63ywi1gF273p65oj4a+P3fTPzjPrzy4HFgQWGONzngRWBDYDrI+ISYGlgGeAm4AvT215JkiRJkiRJ0ujXWuiemb2uDz8PsELXc9H13DxTc6DM/G9EvBnYG3g/sD5wN3A0sGdm3jfdrZUkSZIkSZIkjXojaUHTFyQzfwD84AXsvxew13Nsvx/Yrv6TJEmSJEmSJOkF6/XodEmSJEmSJEmSRq1WRrpHxIuBCd1lWCJiYcripYsDjwBnZOYP23hPSZIkSZIkSZJGmuka6R4Ra0TE5cD9wD0RcXdEbFe3rQ5cB2wPrAt8CDguIn4xfU2WJEmSJEmSJGlkmuaR7hGxDPBbYObG0/MAR0TE3cChwFOUxUhvAV4DfAZ4X0RsmpnHTXOrJUmSJEmSJEkagaZnpPvOlMD9BGBRYE5gTeBG4KvAAsDqmfmlzPxGZu5Ytyfw8el4X0mSJEmSJEmSRqTpCd1XB24HPpGZt2Tmo5n5J2BbYD7gwsy8ovmCzLwM+Cuw3HS8ryRJkiRJkiRJI9L0hO7zAX/LzEldz19YH28b5nW3AXNPx/tKkiRJkiRJkjQiTU/oPhPwYPeTmflQ/fGJYV735HS+ryRJkiRJkiRJI5LhtyRJkiRJkiRJLTF0lyRJkiRJkiSpJdMbum8aEZO6/wH5HNs+Pv3NliRJkiRJkiRp5Jkwna+PaXxdTuf7SpIkSZIkSZI04kxz6J6ZlqaRJEmSJEmSJKnB4FySJEmSJEmSpJYYukuSJEmSJEmS1BJDd0mSJEmSJEmSWmLoLkmSJEmSJElSSwzdJUmSJEmSJElqyYR+N0BjwzFbntPvJky1tfrdAEmSJEmSJEkDy5HukiRJkiRJkiS1xNBdkiRJkiRJkqSWGLpLkiRJkiRJktQSQ3dJkiRJkiRJklpi6C5JkiRJkiRJUksM3SVJkiRJkiRJaomhuyRJkiRJkiRJLTF0lyRJkiRJkiSpJYbukiRJkiRJkiS1xNBdkiRJkiRJkqSWGLpLkiRJkiRJktQSQ3dJkiRJkiRJklpi6C5JkiRJkiRJUksM3SVJkiRJkiRJaomhuyRJkiRJkiRJLTF0lyRJkiRJkiSpJYbukiRJkiRJkiS1xNBdkiRJkiRJkqSWGLpLkiRJkiRJktQSQ3dJkiRJkiRJklpi6C5JkiRJkiRJUksM3SVJkiRJkiRJaomhuyRJkiRJkiRJLTF0lyRJkiRJkiSpJYbukiRJkiRJkiS1xNBdkiRJkiRJkqSWGLpLkiRJkiRJktQSQ3dJkiRJkiRJklpi6C5JkiRJkiRJUksM3SVJkiRJkiRJaomhuyRJkiRJkiRJLTF0lyRJkiRJkiSpJYbukiRJkiRJkiS1xNBdkiRJkiRJkqSWGLpLkiRJkiRJktQSQ3dJkiRJkiRJklpi6C5JkiRJkiRJUksM3SVJkiRJkiRJaomhuyRJkiRJkiRJLTF0lyRJkiRJkiSpJYbukiRJkiRJkiS1xNBdkiRJkiRJkqSWGLpLkiRJkiRJktQSQ3dJkiRJkiRJklpi6C5JkiRJkiRJUksM3SVJkiRJkiRJaomhuyRJkiRJkiRJLTF0lyRJkiRJkiSpJYbukiRJkiRJkiS1xNBdkiRJkiRJkqSWGLpLkiRJkiRJktQSQ3dJkiRJkiRJklpi6C5JkiRJkiRJUksM3SVJkiRJkiRJaomhuyRJkiRJkiRJLTF0lyRJkiRJkiSpJYbukiRJkiRJkiS1xNBdkiRJkiRJkqSWGLpLkiRJkiRJktQSQ3dJkiRJkiRJklpi6C5JkiRJkiRJUksM3SVJkiRJkiRJaomhuyRJkiRJkiRJLTF0lyRJkiRJkiSpJRP63QBJkiRJvbHsccv2uwlT7apNr+p3EyRJkqRp4kh3SZIkSZIkSZJaYuguSZIkSZIkSVJLDN0lSZIkSZIkSWqJobskSZIkSZIkSS0xdJckSZIkSZIkqSWG7pIkSZIkSZIktcTQXZIkSZIkSZKklhi6S5IkSZIkSZLUEkN3SZIkSZIkSZJaYuguSZIkSZIkSVJLDN0lSZIkSZIkSWqJobskSZIkSZIkSS0xdJckSZIkSZIkqSWG7pIkSZIkSZIktcTQXZIkSZIkSZKklhi6S5IkSZIkSZLUkgn9boAkSdJIt8jOZ/S7CVPl1oPW6XcTJEmSJGnMc6S7JEmSJEmSJEktMXSXJEmSJEmSJKklhu6SJEmSJEmSJLXE0F2SJEmSJEmSpJYYukuSJEmSJEmS1BJDd0mSJEmSJEmSWmLoLkmSJEmSJElSSwzdJUmSJEmSJElqyUCH7hExa0TsHRF/j4jHI+LfEfG9iHjVCzzOrRGRz/FviRn1GSRJkiRJkiRJo8eEfjdgWkXErMDZwMrAncAvgUWAzYF1I2KlzLzpBR72uGGef3Ba2ylJkiRJkiRJGjsGNnQHdqUE7hcA78jMiQARsT3wFeB7wBov5ICZuVnLbZQkSZIkSZIkjSEDWV4mImYCtq2/bt0J3AEy83DgSmD1iHhjP9onSZIkSZIkSRqbBjJ0B1YF5gZuyszLhth+cn18b89aJEmSJEmSJEka8wa1vMzy9fHSYbZf2rXfVImIHYFFgSeAa4BfZOY909RCSZIkSZIkSdKYM6ih+0L18Y5htt/Rtd/UOqTr9yMiYrvM/O7UHiAirhlm06IvsC2SJEmSJEmSpAEzqOVl5qyPjw6z/ZGu/Z7PacAHgIWB2YFlgMOBWYDvRMT7p62ZkiRJkiRJkqSxZFBHukd9zOfZPlUyc7uup64BdoiIG4BvAgcDp07lsZYeskFlBPxSL6RdkiRJkiRJkqTBMqgj3R+uj3MMs332+jhxOt/nO8B/gNdFxKun81iSJEmSJEmSpFFuUEP32+vjq4bZ/qqu/aZJZk4Gbqq/LjA9x5IkSZIkSZIkjX6DGrpfUR/fMMz2zvNXtvBeL6mP0ztqXpIkSZIkSZI0yg1q6H4+8CCwaES8fojtG9bH06fnTSJiaWBxyoKt10/PsSRJkiRJkiRJo99Ahu6Z+SRwdP316IiYUts9IrYHlgP+nJkXN57fJiKuj4gDm8eKiHdGxBu73yMilgN+RlmU9Tv1PSVJkiRJkiRJGtaEfjdgOuwHrA2sDNwYEecBCwMrAPcCm3ft/3LKqPXu2uwrAXtGxG2U+u33AK+mlKiZAPwR2GUGfQZJkiRJkiRJ0igykCPdATLzceCtwL6U8i/vBxYBjgNen5n/mMpD/Rb4HvAQsDywAbAY8Gfg08DbMvPRNtsuSZIkSZIkSRqdBnmkO5n5GLBH/fd8++4F7DXE8xcAF7TdNkmSJEmSJEnS2DOwI90lSZIkSZIkSRppDN0lSZIkSZIkSWqJobskSZIkSZIkSS0xdJckSZIkSZIkqSWG7pIkSZIkSZIktcTQXZIkSZIkSZKklhi6S5IkSZIkSZLUEkN3SZIkSZIkSZJaYuguSZIkSZIkSVJLDN0lSZIkSZIkSWqJobskSZIkSZIkSS0xdJckSZIkSZIkqSWG7pIkSZIkSZIktcTQXZIkSZIkSZKklhi6S5IkSZIkSZLUEkN3SZIkSZIkSZJaYuguSZIkSZIkSVJLDN0lSZIkSZIkSWqJobskSZIkSZIkSS0xdJckSZIkSZIkqSWG7pIkSZIkSZIktcTQXZIkSZIkSZKklhi6S5IkSZIkSZLUEkN3SZIkSZIkSZJaYuguSZIkSZIkSVJLDN0lSZIkSZIkSWqJobskSZIkSZIkSS0xdJckSZIkSZIkqSWG7pIkSZIkSZIktcTQXZIkSZIkSZKklhi6S5IkSZIkSZLUEkN3SZIkSZIkSZJaYuguSZIkSZIkSVJLDN0lSZIkSZIkSWqJobskSZIkSZIkSS0xdJckSZIkSZIkqSWG7pIkSZIkSZIktcTQXZIkSZIkSZKklhi6S5IkSZIkSZLUEkN3SZIkSZIkSZJaYuguSZIkSZIkSVJLDN0lSZIkSZIkSWqJobskSZIkSZIkSS2Z0O8GSJIkSZL0Qh2z5Tn9bsJU2frYtfrdBEmS1GOOdJckSZIkSZIkqSWG7pIkSZIkSZIktcTQXZIkSZIkSZKklhi6S5IkSZIkSZLUEkN3SZIkSZIkSZJaYuguSZIkSZIkSVJLDN0lSZIkSZIkSWqJobskSZIkSZIkSS0xdJckSZIkSZIkqSWG7pIkSZIkSZIktWRCvxsgSdIxW57T7yZMla2PXavfTZAkSZIkSSOcI90lSZIkSZIkSWqJobskSZIkSZIkSS0xdJckSZIkSZIkqSWG7pIkSZIkSZIktcTQXZIkSZIkSZKklhi6S5IkSZIkSZLUEkN3SZIkSZIkSZJaYuguSZIkSZIkSVJLDN0lSZIkSZIkSWqJobskSZIkSZIkSS0xdJckSZIkSZIkqSWG7pIkSZIkSZIktWRCvxsgSZLGqL3m6ncLXoAf97sBkiRJkqQB4Uh3SZIkSZIkSZJaYuguSZIkSZIkSVJLDN0lSZIkSZIkSWqJobskSZIkSZIkSS0xdJckSZIkSZIkqSWG7pIkSZIkSZIktWRCvxsgSSPFIjuf0e8mTJVbD1qn302QJEmSJEnSMBzpLkmSJEmSJElSSwzdJUmSJEmSJElqiaG7JEmSJEmSJEktsaa7JEmSRqxjtjyn302YKlsfu1a/myBJkiRphHCkuyRJkiRJkiRJLTF0lyRJkiRJkiSpJYbukiRJkiRJkiS1xJrukiRJkiTNINctsWS/mzDVlrz+un43QZKkUcGR7pIkSZIkSZIktcSR7pIkSZI0jRbZ+Yx+N2Gq3HrQOv1ugiRJ0pjhSHdJkiRJkiRJklpi6C5JkiRJkiRJUkssLyNJ0lQamIXQdvHrXZIkSZKkfnGkuyRJkiRJkiRJLTF0lyRJkiRJkiSpJc4/lyRJkiRJLHvcsv1uwlS5atOr+t0ESZKekyPdJUmSJEmSJElqiaG7JEmSJEmSJEktsbyMJEmSJEkaHHvN1e8WTJ29Hux3CyRJfeJId0mSJEmSJEmSWmLoLkmSJEmSJElSSywvI0mSJEmS1LJFdj6j302YKrcetE6/myBJo44j3SVJkiRJkiRJaokj3SVJkiSNPIOyUCI/7ncDJEmSNMI40l2SJEmSJEmSpJYYukuSJEmSJEmS1BLLy0iSJEnT6bollux3E6bOLp7+S5IkSTOaI90lSZIkSZIkSWqJobskSZIkSZIkSS0xdJckSZIkSZIkqSWG7pIkSZIkSZIktcSVlCRJkiRJkiQNvr3m6ncLps5eD/a7BZrBDN0lSZIkSZI0ahyz5Tn9bsJU2/rYtfrdBEkzwECH7hExK7AL8GFgIeA+4DfAHpl5xws81tzAXsD6wPzAXcCpwJ6Z+UBbbZYkSZIkSZI0di2y8xn9bsJUu/WgdfrdhIE0sDXda+B+NrAHMCfwS+CfwObApRGx6As41suAi4DPAU9TwvaHge2Ai+t2SZIkSZIkSZKe08CG7sCuwMrABcDrMvNDmbkCsAMwD/C9F3CsI4DXAqcAi9djLQMcBSwGHN5qyyVJkiRJkiRJo9JAlpeJiJmAbeuvW2fmxM62zDw8IjYFVo+IN2bm357nWPMDmwBPAZ/NzKcbm3cENgY2iYgvZebdrX4QSZIkSZIkjVnXLbFkv5swVTbaZTAixKv63QCpGtSR7qsCcwM3ZeZlQ2w/uT6+dyqO9W7K/w5/6g7VM/MJ4FfA+LqfJEmSJEmSJEnDGtTQffn6eOkw2y/t2q9Xx5IkSZIkSZIkjWGRmf1uwwsWEYcDXwCOyMzth9i+PHA5cGlmvvF5jnUKsD7wucz82hDb16MsrHpKZm4wFW27ZphNS8wyyyzjFl10qtd3HVXuu/ORfjdhqs3xyJ39bsJUuePl0e8mTJVFn3qq302YajdOflW/mzBVXjvfnP1uQusGpY+wf2iX/UP77B/6x/6hfYPSR9g/9I/9Q/sGpY+wf2iX/UN/DUofYf/QrkHpH2B09hFT46abbuKJJ554ODNfPC2vH4yCTP+r8//2o8Nsf6Rrv14d67lMfuKJJx659tpr/zmdx5GKf7d+xM4doZvaPOi1bR5shru93w2YKtfe2+8WaMSzf5gB7B80SrTfP8CY7yPsHzSKeA7RMvsHjSL2Dy0bjP4BxnQfsSDD58XPa1BD987tteGG6b+Q229tHovMXPqF7C+NFJ1ZGv43LKmb/YOk52IfIWk49g+ShmP/oNFuUGu6P1wf5xhm++z1cWKPjyVJkiRJkiRJGsMGNXTvzMEYrgDSq7r269WxJEmSJEmSJElj2KCG7lfUxzcMs73z/JU9PpYkSZIkSZIkaQwb1ND9fOBBYNGIeP0Q2zesj6dPxbF+A0wGVouIeZsbImIW4L11+5nT3lxJkiRJkiRJ0lgwkKF7Zj4JHF1/PToiptRjj4jtgeWAP2fmxY3nt4mI6yPiwK5j3QmcCMwMfD0imovLHgLMA/w4M++aMZ9GkiRJkiRJkjRaRGb2uw3TJCJmBc4FVgDuBM4DFq6/3wusmJn/aOy/F7AncFxmbtZ1rJcDfwUWBW4CLgGWBpapv6+Ymf+doR9IkiRJkiRJkjTwBnKkO0BmPg68FdgXeBR4P7AIcBzw+mbgPhXH+i/wZuAoyoj39YG5KKPp32LgLkmSJEmSJEmaGgM70l2SJEmSJEmSpJFmYEe6S5IkSZIkSZI00hi6S5IkSZIkSZLUEkN3SZIkSZIkSZJaYuguSZIkSZIkSVJLDN0lSZIkSZIkSWqJobskSZIkSZIkSS0xdJckSZIkSZIkqSWG7pIkSZIkSZIktcTQXZIkSZIkDbSIWCEilut3OyQNtoh4X0TM3+92aPAZukuSJEmSpIEVEa8FLgB+FxHL9rs9kgZTRGwHnArsGhHz9Lk5GnCG7pIkSZIkaZDdAfwAmAc4xRHvkqbRpcAVwGeA3SJi3j63RwPM0F2SJI1KERHP97MkRYTXRNIAi4hxmfkYsDXwdWBR4GSDd0nT4HxgC+AaSp/yZYN3TStPMCWNGM910esFsaQXIiLGZ2Y2npqz80NmpsG7NHZ1n1Nk5uSu7fYP0gDJzMkRMXMN3ncCvgYsBpxq8C7phcjiImAr4G+U4H03a7xrWsSzr0clqT9qQDap/rw25UR5KeB64A+ZeV0/2ydpcHT1J5sB7wBWB26k9Cl7AvekJ0HSmNPVP7wTWB5YmjKi7eLM/EPdNq47jJc0MnX9Xc8FvA7Yl/L9fz3wocy8qo9NlDQAImJCZj5db77PC7wNOJQyYPlnwP6ZeXc/26jBYuguqe+aF7YRsQ+wPTAb0Blp1hm1ckpm/rs/rZQ0CCIiOmF6RBwEfAmYCNwMvAx4JXADJXj/dWZO7FdbJfVW1/nGfpT+YUJjlyeAH2Xmp+o+4c05aWTr+rveHdiMMrttJsr1xCyU4H3jzLyyX+2UNLJ1bt5FxByU9SGWB+ai9CezAQ8Cx1OC9//0raEaKJZrkNR3jRPlPYHdKAuXbAysBuwM3ESZJvqZiJivX+2UNPI1AvfPUwK1M4E1gTcBy1Ju4C0OfBl4s6WrpLGjcb7xZWBX4C+UkbBvrI83A5+IiNPr/gbu0gjX+LveA9ibMmvlw8AqwAco9ZmXoNR4X75f7ZQ0stXAfTbgbODd9XF94H3A94D7KSVnrPGuqTbh+XeRpBkvIlYGvgBcDGzZmAJ6fj1BXoYy/XumPjVR0oCoN+c2B/4J7JaZl9URqw9ExEvrbncD91o+QhpbImJF4HPAlcDnM/PyxrYzgVcDL4qIxTPzhv60UtILUa8VtqUM1Nk1M6+um66LiLOBYykj4H8SERs54l3SML4IvIWyIPMXM/NxgIi4CFgVOJgSvBMRB1hqRs/H0V2SRoplgBcDBzZrLkbEbpTRKqcBh2bmHY1tLnQmaSivooxq/0FmXgZTFk/dkzLS/Qxgp+6Lbke9S2PC64CXA4d0Be57U8rb/R7Yvjtw95xDGtFeQikhd3JmXh0R42FKfeYngW0o1xKvA05ycVVJw1iSUmruO5n5eKcvycyHKecHOwK3UW7ifdnFVfV8vLiUNFIsXR9v6zxRp4nuA/wa2LOuIk5ErFlHoDntW9JQXl4fx3eeqP3JnpT+ZLdOGB8Ry0bEEfDMFHVJo08jNF+hPt7Z2LYHsDulf9g9M/9Wn18uIjYAS81II9w89fG1ETFrZ1HVuiDi+Mx8FPgMcAul1MxPI+L1fWqrpJHrRUDyTFWQKd/9mfkUpVzV6XW/DYGdDN71XAzdJfVUc6RY16ixB+rj7HXbPsBelAvgL2fmFY19Pw2cHhEvnqGNlTSidUafNH7vnNfcCjwJLFif353h+5N3A5+LiFVndHsl9c4QI9M7v3dGsHf6hz0Yvn/YEtjBC2ppxLuQciNtcWoA3xihOikiJgAPA/8A/kUZ8f67iHj50IeTNJY0zhluA2YFNoUyIKd5PpGZE4ETgacoeernKNcRZqsakv9hSOqZOtKkebc4I6JTo/2v9XGHGpDtRlkA8VkXwBGxPrABZXrX471puaSRqDOSrTNNvJ4Yjwf+C1wGfCwiTqMsrHYGZcZMsz95J6V24++BG3vcfEkzSESMayyq/HJ41kyWa+rjphHxTUrgfgZlhHuzf3gv8ClKSP9Qj5ouaRjDlXiqYdcDwHmUmbO7wZSwfVztD57OzMeAe4ArgOOAgzPzvz1pvKQRo2sQ4Dh41my2b1Fu0L0/Ij7Y2Vb7ks5gn7soa0PtT1ls9QRny2o4hu6SeqYRkH09Io6rzz1VN/8VuBR4P6UExJnAdl0XwKsCe1BOrH9aazRKGsMi4mjg8oj4CJR+JjPvBQ6pu6xLCeC36pSMqK9bhdLXzAYc60JI0ujRufiNiAOBoyNiica2sykX1W+lzJz7I7BzZ8Hl+rqVKGH8A8CJtTSFpD5pDtyJiNkiYv6IWADK33tmPgQcATwIfDoivtbY1ukPVgXeBZyVmZtn5qH1eTMRaYzo9CWN4P1lXbvcAHwFmBf4YqPE3OROlgFsQpmdfxbwnsy8tgdN14DyC0bSDNc8mY2IWSnTtT8WEcd0ns/M+ylfYHdR+6bMvKmxENK6wGHA8sAemfmH3n0CSSNFd0kZ4Pb6eHREfLTzZGb+Avh8/fX1wPoR8er6b3PgO8CKwC6ZeUo9tgslSgOs63xjXmAVYCNg24hYqrHr8ZRyMgCPATPBlNFs6wFHUfqN3TPzrF60XdLQakjWGbjzWeCXwLXAlRHxk4jYOCLmyMwLKTWWnwa2iYhTI+KTEbFInSm7P+VG+82NY4cjVKWxoS6sPCkiZgMOiojfA9dFxG8iYr/aHzwB/IgyG+b1wMERcWhEvCYiXhERO1IWZr4RuKsxgFAaUrgmkKQZqetEeWXKYiQfAraiXOR+IzO3buz/FuDnwCspd5pvpZwgrwJMBr6UmUfWfcOFzaSxo6s/2RTo9Ckfo/QTjwGfyswTG6/5JHAsZVHVyXX/8cB/gL0y89i63zgvvKXB1dU/rEqp7bwysDGl9uqPgaM7I9JqeamtKbNhngYuAuag3Nx/HNgpM4+q+9o/SH3QPNePiIOBHYF7KYsZLgQsQlnw8ARg18x8sP79fx1Ykme++zs35HbIzCN6+iEk9V3nezwi5gD+ALyJMnDnHmAxYC5KiaptM/PKiHgtpa77dsCclH4ngJdSBgm+LTOv6/0n0aAxdJc0w3SdKO8PfBJ4OaV0zMqUkGxW4JjM3LbxuldSyj68mXIy/Shl+tYpmfmruo8XwNIYVUtGbE85WT4BeA3wKmAtSri2aWae1Nh/ReAt9V9QyklcnpkX1e32J9IAG+J8YxtKX3A1MD9l0cQnKSPcv9oI3hcC1qOcn7wYmEQ53zgtM39b97F/kPosIrYAvkG5htgzMy+JiLmBhYE/U2owH0gZzPN0RCxKuY54L+Vv+xbgd15HSGNXRMwM/AJ4J2UG/Z6Um+6zUcrObQz8CvhsZv6rztBflrL+0yvrYa4GDs3Mm3rcfA0oQ3dJM1xEbE/5Yvs5cFhmXhgRy1CmbB0JzA18PTO3abxmAmUk/PyUBczua1xQe6IsjVG1hMzxlGBsh8y8pj4/L2XRw/0YInhvvP5Z/YczZqTRIyK+QKnFehqlNMxVtWzUFsBHgRWA79MI3uvr5gaeAMZn5sTG855vSH1U/37noCx2vBTw9sy8vLF9d8pi6b8C9mmu3fIcx/TvWhpDOuf6EfEOSl9yCvDJru/7S4BXAz+knD881Ng2pc+w/9ALZU13STNMFPNQLnQfAfat9RbJzKsz8wTgHcD9wGebNd7rPo9l5i11UcTm837RSWPXivXxoMy8plPjPTP/k5kHUKaezwR8P+riqjDlRt7/9B8G7tLoUGu0foCykOJeNXCfqf6NfwvYjTIidjNgu64a7xPrOcfEZl14zzek/qp/v/NQykz+qitw34MSuJ9B+Zv/W31+vu7FUZtrtvh3LY1eETFf93ONc/03UEpOHdIJ3CNifERcULd9nbJ23EMR8aLGIZr9idcNekEM3SXNMPULbg5gUeDqzLwSnjnxrXeKL6EsevQwsFVEHFtf+3RzwUSDMWlsqyfFEyjTxaGUnYJSq3VKv5KZX6GcNM8CHBsRm9Tnn3ahVGlUmwt4I2WRxCtqf/F0Z6HEzDyXMgr+buAjwNYRsSSU/qFzEAM5qX+6w/KqU5N95sZ+ewJ7URZE3i0zL6vPL0wpP7dc8wBeR0ijX0QcBJwaEct3Pd85/1+gPs5an5+JUsd9BcpCy4dk5oN1n3dExL7wP+cI9iV6QQzdJc1oj1AC9ZdExEvqc51wbHL9Evwrpb4awGci4qt1+yRDMklQ+oN60nt+feo19XFc3d48Cf5jfZwTODoi3jvEPpJGl0cpgforgYUy8+msGjflTgfOpfQNm1Fm2S3Yp/ZKaugq4TB/fRxPWY/hIWCZiJirBu57UgL3L2fmFY3DrEeZ8fY/o10ljV4R8WLKjJgVKIssP2tzfbyyPr66Pv6JMoO2E7g/3HjN9sC6jfxCmiaG7pJmmDpa5UngBsoiZhvAs8J26vXwY8DllJXAH6FM+/58Z3vvWy5pBLu+Pu4TEQvXm3PjYMrFOZSF1i4GTqSMft0rIpbofVMl9Uqtv3oJMC/lBv6LGtuyjmgDOJnSj/wV+DTwPnhW/yGpDxqB++HAd+pzkzLzzvr7csDplMD9DLoC94hYG9gZuAj4e29bL6lfIuITlJkwHwPel5m/iohZImIxeNYMts5aLkdGxNU8M8L9K83APSL2rttOoQwelKaZobuk6dYoFxON58bV6dwPUk+cgUOaI04jYkLjNXNSLpbfR6nHutFQNdkkjW7DTC1vlo/5FuWiezHKKPaF6428CZk5qe7+HsrF+e7AgZRFm1/fPI6kwTPc329EdMpOHAPcAXwIeHet807tH56q+ywJvJxybnIXsFtEzN/oPyT1SV0U/UPAeyJizcamXwHXACtTAvXDugL3NYA9gLmBwzPzlh41WVIfNW7SbQrcnpmnR8QslFzhuxGxTGffzLyAUpbqJZSFmX+Umbtn5gON420JbANcAXy/WVpGmhaG7pKmS0SMb4xGn7lO+5ylWRM1M08EDqacCB8ZEZ0R70/X8H0FSl33JyilI06lTPV6Q+8+iaR+q/1JZ6Tb4hGxYkS8t07tbI5C3Z4ySnUd4McRsWxne0SsQhm9egswCfhLfc36Bu7S4Gqeb9QRbPNFxEsBMvPJutvlwAmUEjN7Ax+OiLk7F80RsSLwQeAP9dzkTEoZijV7+VkkDS0z/0P52wVYqfH8H4EjgduB1wIbRMQGEbFSRGxL+btfFdglM38K3mSXRruI+BLweeBnwNmN/GEu4L/AapSZscs2XnYk8I3685oR8YGIWLRedxwDHEa5fvhIZt7Ri8+h0W1CvxsgaXDVC+BJ9edNKaPUlwYejogfAWdl5pRpXMDswLbAzyLi68B1lBXAP0u52/yVzHwyIm6ur5m3d59GUj919Sc7AFtSajLOBFwFnBQR38zM+4Bbga2Bw4E1KMH63yLiXuAdlAWct8/M2yJi9voWD1quShpMXf3DZ4APUAK5eyPit5Tp4Xdl5gP1onl2YHPga8Cn6znJS4CPUgK7b9dDd+q7WrNVGjnOp8xC+WJE/DQzbwLIzO9ExERgC8o5wDZ1/8mUMP7TmfldeHZ9eEmjTx2QsxHwT2DvzLw2IuYE1sjMM2o2cRhlYF9ExJ6ZeWVmPljLxzxIKUd1Ms8s1gylPOWmmXlDrz+TRidDd0nTpJ7Mdi6ADwR2ogTof6esDH44cG5E7JuZf8jMuyJiV+A2SrmHzzYO9xSwQ2aeUH9/M/A4cGNvPo2kfoqIaPQnBwFfolxwf40ygv09lGnjS0XEFzPzbuCyWq7qIOBtwOqUuot3UqaWH1sP3+lrrmy8l+G7NCC6zjcOpiyS+DBl6vj8lABuKeCAiPhjZv679iPXUxZLXaH+A3gM+EKjf3grZZZdJ3yX1APP9V2cmddExCmU7+/3AEdFxMyZ+WRmnhQRf6GUkFsdmI2yOPKNmdn5njdwl0a/2YBFgfuBuyJiAmU9h4kRcXNmXhcRO9V9NwSIiL0y84o6o2bXiPgD8CZgceA+4I/ABXW71IrwulPS9IiInYEDgN8C+2Xm+RHxGuB4ykXu1ZQRp39ovOYtlC/J1SlTwW/OzN/VbdsBX6WcQG9YR7VKGgMiYmvK3/8ZwP6ZeXFEzEoJ3HemLMx8CvC5zLyn8bpXAAsDDwFPZuaN9fmtgEMoo2Denpn/6uHHkdSiiNiRUqruN5Tzjb/URdK+SQnPLwT2A36fmU9EWTh1PLA+8FLKVPM7MvP8erzPAUcAFwDvb/YpkmacZigeEXNm5sROCB8RM2XmU7UcxNnAdZm5Rt13yoyX5zi2N9alMaLWc/88pbzUipR8YQ/giMx8rO7zasq5w4bAL4E9OzfopF4wdJc0zSJibeA44CZg28y8ot5l/hxlkZLJwIsoCx9t1wzehznersB2lOldqzmtSxo7ImJp4EQggM0z85K6OOJ2lP7kXsoolOWBk6jB+1AX2PV1+1NGuSawZqPUlaQBUxdJ/BGlhMQ2mXlpvSH3Kcpslycoo95uAHajBu/PcbydKGtDdM43rp/BH0FSl4jYnVKW8uuU0aVPNbbNBfyEUjLuU5n5vf60UtJI0bg5N55yvfBS4AfAO+suR2bm9nXfCY31XIYN3uuxJtfjetNOrXMhVUnTpI4g24BSSma3GrjPRAnI9qaUhliaUh5iaeCIiHhb4/XjOgscRcQSEXE2ZYTaA8BbDdylMWcpYBlg9xq4T6CUjdiT0p8sRxmxegtlIcSjImLexsk3MKXG42nADnXf1Q3cpcFVzxVWAV5B6R8urecbW1EC939TZs4dRbkptzPw9nrzjca5RkTEvBFxAaXM3X2U2q8G7lKPRcRKwCcpNZn/APwkIraqf6cTMvNB4FDgacrft6QxLCK2AN7dmfFSA/VHgddTZsICvDQiFgbIzKcjYlz9+RZKKdyTgfWAvetsGuqxsv5s4K7WOdJd0jSrX34rZubm9UttY0ppiInASpl5d0SsSik180pKzdSdM/PsruO8BPgW5cL5iMy8tXefQtJIERFfAI7PzHvrTbrvUUaqr1zrNL+YsijzppR+5rfA1t21FyNiTWAx4IzMvLOHH0FSy+r5xYeBV2XmwfX3DYCjgUd45nxjaeBPlEVR/0opfXdWZj7ZONZclMUXXwIclZm39fbTSIIpg3cAPg6sA6xLWW/ur8BZwPcpCx2eSgnd187Mc3rfUkn9FhHvAn4NXEopB3dHff79wE8pWcMrKCPejwMOzczr6j7NclbNEe/nUmbqX9PTD6Mxx9Bd0nSJiJfVgGw2ynStZSknxtfU7TMBf6EEYHMB/6IE9f+q28dl5uQ6Im1cZj7elw8iqW+aU0Abz30b+Chl5stfI2KWWqf5U5SRrJMp/cp7MvM3QxzThdSkUSIiZgdmz8z/1t9Pp9RvXaMuujhbZj4WET+jzK5bkGdGst/adawAZmqG8ZJmnEZJiCmlG7p+np1y/bA1sDLwGuA/lLIzK1BKzBxFGak6ye92aWypYfl2wC2Z+bXGCPbJdcT6jZR+4xDg3ZSSM4cNE7wvAnyDMoNuqU6AL80ohu6SntNwwVX3YkYRsT7wc+A7mfmZWu4h65fhGcDfKHXXbs/MQ3rVfkkjx3MtgtbZVgOxxSij3e6jXIDf13ldRBwMvJ0SyC+dmT/rTeslzUhDnW8Ms2bDWsDvgR9m5sc7N+Tqtt/W3S6nnG8c04OmSxpG83u/zjSZF3gceGKIWWqzA7NQ1ltYmxK4PwrMTgnVVunceJM0tsQziy7PQbkhdxJlNlunfxkHvIFSlvI9PHfwvjDwdGcQoDQjTeh3AySNXF0nyssCcwP/BW7LzEe7ArSoj/fVx86CJGsAawKnZeYejWO7UIk0htST3U5/siFlNOoTwHWZ+ctO4F77jbuAO4A5gAcar1uJMgX90lqn/drGsR35Jg2orvONlSmLsD8CXE1Z66X5d/6sWTGNwH0N4C3Ajpn5ncaxPd+Q+qDr73ob4COUGSqPABMj4iDgd411V57MzEeB3SPiUMqI1Y9RRsG/FvhiRHx5uJv3kkavzJxYf/w0pV94DfB4RPwxMyfX84NLImLvut9mABFxWGZeVwcCjqv7WlpOPeNCqpKGVC9SOyfK+wDnA3+sj6dFxKsao1KhhPEPAJtFxBuAWeoF8J6UYO3q5vG9AJbGlsbokgMo9Rf3pNRc/kVEfKuWmMlGn3Il5YT65IhYOyI2oyzM/Drg9KGOLWnwDHG+cQ5wJqU++/ERsRE86+/8duBhYP2I2Kq+7h3A7pRrm2ddTHu+IfVe19/1wZTv70Upa7H8DZgPOAI4tK7h0ln4cHz9+aHM/AllDZcNqeUpKeu8SBq7TqWUkXkTsD+wRqfcDEBmXgLsTakBvxnlZt0SdZvXC+o5y8tIek4RsSNlwZF/ABcBS1JWCb8d+GBmXtzY90hgW0r4fiewCDAb8LnMPKqnDZc04kTER4HvUkrH/JAyQ2YPyuJHpwCfyswH6r6LAScCb2wcYjLwBfsTafSJiO2Bw4CbKMH7EpSaq/8F9srMYxv7bgocSylFcRuwcN30hcw8spftljS8iNiCUj/5l8BujTWfNqSUiXsfZUHD3TPz/Mbrumu+fw/YiHLt8fOefghJffEcZW4XpGQO21EWV/0y8MfmvhHxJsoAn7dT+p/dM/PvPWm41GDoLmlIdbTpi4HfAROBrTPzurow6tGUqV3/AjbMzAsbrzsIeBslnP8b8N3MPL5uswSENIZFxNcp/cMHM/PK+txSlD5lTcrolU80gvdXUy6y3wDcAFyUmafXbfYn0ihRzy1+A4wHtsnMqyNiHmB9Su3WR4GdM/Prdf85KH3J/vUQtwMnZOZJdbv9g9RnETEL5Yb6KsDamXlJNBZOrzNjd6R8z38T2AF4vDk7pfO3HBEbAD/DG2vSmNBY62lmyrnAk5n5i8b2qQne30iZUbMUsGxm3tnLzyCBobukhu66pxHxSkrQ9cHMPDMiZsrMp+rF8f7AFxk6eJ8bmBV4ODMfqc95ASyNIcMsingBcHZm7langkY9oV6cEry/ja7gvb6uu2+yP5EGWPffcEQsQFkocevMPK5r3w8DJ1CC912ai6NGxEuBpyh9yUNDHVtSbwzxd/0KSnnJazJztUb5uCllnyJiVcoo9lcDb83MPw9z7N0pJSN2zsxDXKtBGr06N+fqDfbvAe+nzKJ/d2dh1Lrf1ATvrwf+m5n/7OFHkKawprskYMrd5M4J8NwRMR9lUdR/AnfWE+Wn6wn1U8AuwFeAV1JqLq/QOVZmPpCZdzUC9/ACWBo7an/SqeG+WkS8NyJWpJSN6CyA1hnBEpl5A7ANpaTE+4Hv1Zt3Qy6CaH8iDa6u/mGRiHgV8DJKiYnODJiZOvtn5onAx4HZgQMjYuvG4R7MzIcpNd4935D6pBm4R8TyNSx7glJGbtGIeG02dAL4GrKfSpnl8rZhjv12YAtKqamf1tcZuEujUO1LOoH7H4H1KH3EKsCzysPUIP0oypoRb2DoGu+XGbirnwzdJXW+3DqLHe0I/AG4lbJo6iuAuevJbdQpnuPr/jvzTPB+UkSsNNTxPTGWxo6uxdP2As6i1FL8C/ARao32OmtmXOfiuwbvW/NM8H5CRMxt/yGNHl3nG7tQFlW8GDgNeA/wTpjSPzRHxf6YZ4L3fSNi2/r8pPqYzUdJvdUI3I+gBGTvy8x7KaHZy4CVuvbPxs21P9bHVwxz+HHAtcDbMvPWdlsuaSSpWcPMwI8p5WoPosyA/Rdlbafu/f8JHMMzwfvewNubwbvUT/6HKKl5orw3ZdHURSgj3OcHXgR8NSJeWb8Ex9XRqc3g/RDKIma/qyPkJY1RjRkz21MWSb0V+ColeH8YeE9E7FP3nTxE8P5Z4DxgHWDV3n8CSTNK43xjD8qItJdSFmqfl3IxvV5EvKXum0ME7x8F5gaOjIjlett6Sd2af6NRFkv/DGVW2zX16d9QMoedI2KZxr6dmbMAC1L+/qeUqmzKzN8CG2TmVe1/Akkj0JqU64BfAod2Zs/X84IFImK9iNgwIhaufcltlOD9CMq1w/aUUrdS303odwMk9U/XVNBXA58AzgS+RAnKVqCs+r0aJXj/XGb+uzt4j4gvU8L5f2Tm3X35MJL6qnEjjjq6ZE3gAmDLzLwqIl4CvJVSm3HniHg8Mw9oBO+Ta/D+94jYElgi66KpkgZb43xhHOXG/scpYdyXM/OyiHgXsCXwPuCLEbF/Zl7RuCHXGcl+UpTFGefKuhizpP4YYv2EJSh1l7dv/H3+iPLd/0Hgx/X7/apaFopannJL4EHKOlLd7xG1Is3DM/CjSBpZXke5WXdks1wtsDul3Mzr637nAUdExC8z87aIOBZ4EjgxMx/tQ7ul/2HoLo1hjcB9GWAxykiz/TKzMzrlnIi4CTge2KDsGtsNE7xv0xjh6iJm0hjTCNw3p1w8rw7s2BmZlpn3A6dExCOUmqx7RwTDBO/XAdfV49mfSAOu0z8AywH/psyk2zozL6vbfxMRdwMTKWWoIiL2GyZ4n7LQqv2D1D+N64j9gDmpi6F3AvcoiyE+XIP22SkjV38KXBARZwIvBzallJDYNjPPH+I9LBkljWLN7/fGd/psdfOHIuJ+yqy4Qyij2P8JfJ2y+PLbgQeAXwGTavC+b2Y+3eOPIQ0r/B6TxraI2I0ysv17wAqZuVJEjKec53ZOpl8BnEgZ8X4K0AzevdiVBEBErA/8nDLyZAHgk5l5XkTM1JhGTh3V+hPKRfiemXlAfX58I5yTNIrUYG5X4BeUtWDWzMzHazD3dN1nOWAn4MOUvmS/zLyibvufRZUl9VdEvA64HHgEeBz4UWbu3PnebwzOmYsye/adlJC9425g78w8th7PawtpjOiaBddcE2pOymzZpYF7gRdTBvT8gnLdcFdELESZob8kZeT7VfYdGokM3aUxrC5S8oX6b17gMWC5zLxpiH1fQVnQZHXgZOALdUETSQIgIpYEPg9sBMwFHJyZu9RtzwrMGsH7zMCBmblP71ssqVdqvedvUeqsPgC8PTP/1qkJ3Rjp1gzef0LpRy7vR5ulsW6I7+6ZM/PJ+vNsmflYRLybUkv5dcDVwKqZ+VBjBlsnWJsFmAdYi3KOcBtwS2dGnIG7NHZ0brhHxKzAF4E3Ah/LzIl1+3zAUZQZMTdSBghe3Sk3U/e5hFKGZpXMfKzXn0GaGobu0hgXES+iTO3cinKneH/gK5n5wBD7vgI4gVKb8XfAB5pffJIUEYsD2wCbA7dQZsb8oW7rvnh/B2WRpFkoM20u7kOTJfVIRGwEfJuyDkzzptw4Su7eDN6/SFk49beUC/H/9qfV0tjUVfZh2WwsZBoRe1Funn2TUkN5bUrwvgQlKNs1Mx/pKh03bPDgTBZp7GjciJuDMpjvbcCjwEcy89dd685NudHXdYzPUfqcbwOfy8zHe/gRpKk2rt8NkNR79eK2M6LkYUrN9mOBOyhB2Tr1S/BZMvPfwMco00jPNnCXxqbOyNTOY/15PEBm3gAcTelXlgB2iogV67ZsviYzz6KMit/awF0avRr9w0+BT1HKUOwUEdvV5yeX3aaMer8SOJxSp/U3Bu5S7zUC998Cx3W+yyPiUGAPymCdWWpJiHMoM2dvAT4N7BARszfWbHnOQN3AXRob6g22SRExO/AHymC+7wILZOavoZwTRMSEuu+TzWuHeoxtgB2Bm4CDDNw1kjnSXRoDpma6Zh3xvhml3uqT9fHUoYL1ehL9aP3ZkSnSGNKsu16nis8KPJ6ZT3Tt91pge8rF99mUGox/rduG7DecWi4Ntuf6G26u7RARG1Jmzs1CKVd35FCvj4iXdwJ3zzek3ouIBYCvABsCv6GUovwgcBJwQGZe3dh3AmXE+zcoJSEOBQ7LzEf9fpfUUW/EfwvYBDiYMvPt0ed5zYuAZSizad8H3Ae8OzOvncHNlaaLobs0ynUFZG8HlgOWAs4Hzq+jUjv7zkkZ6f68wXvd3wtgaQzp6k82pZz0Lg08DPwIOKt58tsVvP8e2Ov5gndJg6mrf1iXsrDZosCdwLGZeVvX/sMG7zRKzdTn7C+kPomIBYFtgR2AAP4IfCozbxriRpnBu6TnFBGvBP4K3Aq8q6tO+yrAGpQa778Bzs3MGyNiYcos2tUoC63vlJk397rt0gs1od8NkDTj1JPbzgXwXsDOlEULoYTrl0fENzLz2wCZOTEivl+37wocAEyOiNOGCt69AJbGjq7+5EDKQocJ/B1YgFIK4tyI2CczzwWoJ8mH10N8GpgUEftl5gX2H9Lo0dU/7EPpH2Zq7LJpROwM/Lozcj0zT64zxk8AjoiIyZl51FChnP2F1D+Z+c+IeDElcE/K3/Zcdduz6rVnWRjxd5S1or5BKTkzISIOtiylpGoB4JXAiZ1+oc6ePYIyq+bldb91gJMjYufMvC0itgIWAS7IzPt732zphbOmuzRK1RPgzgIk+1BqL14FbAC8mbJA2TLAHhHxpc7rsqwY/n1K4D6OEqR9qI5ckTRGNfqTnSmB2m+B1TNzSWB14C+U0SdHRMSajdfdSOlHvgm8G/hKRMzf08ZLmmG6zjf2BHYDLgE+QLmw3omycOohwMcionMxTWaeTFkrZiJwZETs1OPmSxpCY+2WcbX28rzAaZSyMisBB0bEGjDkei2TKLPbtqCUo9mNcs0hSQCPUBZi/mREfDYitqScN2wJ/ItSwmqL+ty6wMIAmXltZv7awF2DxBBNGgU6I0y6Rpp0Fj/6CPB54Ezgy5l5ea2jthpltMorgD0j4unMPLy+tjPifTxwUH3u6V5/Lkn9ExETuv/uI2JtyhTzPwM7Z+YV9Ybc+sDywKP18aiI2LZrxPuRwIuBizPzrh5+FEkta5aJaJxvfIgyqvX3wI6ZeUV9fjJlhOy8wJ7AuIg4PjPvqa8/uZaUOanuJ6mPusrAvCgzH6wl5WalDMh5gBKOUS89/lSvQ6aUmKrnD2fVMO3lmXlh7z+JpH5r5hSU8taTM/O6iNgV+DpwdN31GsogwW9k5r31ta8CVgbeRCmNKw0ca7pLo0BEzNO5eG0G7xExD3AisATwocw8PyJmBrYD9qbcST4EOIpSw32/zDy0cdwXAa/pXDhLGv0iYgXg1sy8uxm8R8RMwNcoI0/WzMw/1ee2Bfah1G1+K2UWzXbAlcAOmXl249hzZ+YD9WdrNEsDJiIWAe7JzEeawVxEzEWptboqsE5m/rX2D58H9qL0D1+h9A8vBvYHfpyZ/2kce7HM/EcPP46kLl1rM2wMbAz8JDNPbOyzGOVveyvKTbb9M/NPje3LAQtR1o66v/G8Nd2lMaLZlzzHPu8CXkMZ6PcT4P6sC67X7ScA7wLenpmXz8DmSjOMI92lARcRawG/j4iPZuaPu0a8T6IscvjNGriPBzajTPO8C1glM++JiNdRFjv8fH3pYQCZ+TDQGanmibI0ykXEUsAFwD0RsVwzeM/MpyLicuC4GriPo0z/3Bn4D7Bq3f9kygKrSwIH1TqMZwMYuEuDKyJWBH4HHBMR+3YF7zNTzjkOr4H7eODjlPVh7gL+r86iex3wOeBL9Zg/aox4/0d9zvMNqQ+61mbYg3KTbE7gioiYNTMfh/K3GhFH1JdtBWR97bkR8XpKSbn/A9YCpoTu/l1LY0Pn2iEiZgU+QckjXgX8CvhLZl4LkJm/6Xrd+MbPn6HMpD0fuL1XbZfa5kh3acBFxGd5ZlrWRrU+6pS7y3Va1syZeXO92D0ZmBtYIzNvqfu+Bzgd6NxZ3jszD+jl55DUf7VUzM+B9wI38UyQPlNn5ElEvCwz742I2YBfAssCa2fmNXX7TJT67otRFlr7F7BiZv6r959IUlsi4h3Ajykj0o4EDm0ujFhHt95Zb+a/BvgpMB9l7YfO+cYrgaspZWRmp8ySOawT5knqv4jYF/gycC6llNxFjW3NkfCLUW6ifZaybtRfgFUo5wV7ZOZ+PW66pD5rZBBzUPKFNYCnKecOAVwO7NoJ3IcaiBMRu1BmzU4C3pplfShpILmQqjTgMvPrlJNdgJ9GxIb1+Un1S++OzLy5bv8AZSGjXTLzliirhEMZ2XoZ8FVgJuDhnn0ASSNC7S+epiy2/BNgUeDPETF/HeU+E0CnziJluufawK8y85qIGF9Huj1FGfl+FKVW41EG7tKocA7wYcrI1S8CX6wX1QBk5pWdUeuUfuQNwMH1fGOmWs/1EeBx4EeUkjP3GrhLI0dEbESZ/fobYLtm4F5NGa1eZ6ccAewLLEep874gsG0ncK+z4iSNAZ3ZMlEWX/4dpeTcDyn5w2uAb1BmwRwXEevBMwsxR8RsEfHJiLiAckP+PkpZGQN3DTS/BKUB1jmRzcxjKXeDoQTvH6rPT6pfYp2Fyd5YH++o25+ov78beD3wfWChzDyqF+2XNHLU/mLmGrxvBpxACd7/0gjem2XpOv3KffVxcmZOjog1gDWBf2XmNpl5CJSRLD35IJJmiNo3nEO50X8XpUTMlyJiTvifcG1BymLtnRrtk+pIttUoNd2Pp8y4+0aPmi9p6qxGKRd1UGZe3XkyIpaPiM2BsyNi/4h4P0Bm3pyZe1EWOlyHEpIdU19jqShpFOs+t6/XARMoa8b9H3AosHVm/p1S2nqmuus8wAkRsW59XWbmY5RyVktQwvl3Z+Z1Pfkg0gxk6C4NtilfdJl5NGVkyuPAiY0R78kzf+udO8XLTzlAWTRxS8qihw9l5h31efsHaQypI92frL+OAw4ALgQWAc6LiPlqfcbOCfM9wAPAZhHxBmCWGrjvCTxBKSExhTXcpcHVqbNay0qcA+xA+fv/BDV4rxfbnXqsj1HOUTaOiJfXbSvX191NKUNzaz22N+SkESAiZqaMWL+3a2HUz1BmwH2XclN9F+CrEfHBuj0y89LMPDMz/1afM3CXRrGImKOOUu9eJ3J5YEPgj8CBmflwRLyWsqj6p4CvUW7azwn8sDPiHSAzj6QMEtwlM63jrlHBhVSlAdVVU/HTwErAvDwz7fOnEbFhZp6Sz6wc/gdga8qJ8v9RLorfS1nYZKvM/Hfn+J4oS2NH1+JpuwIfA+ao/56kjHj/U0Ssnpl3A2TmeRFxArAt8HtKqYhFgNmAz2Xm+T3/IJJa19U/fJYyXXxZSn3WBSjBe0bEIY0a74dRyk99HFgsIq6iLIg2L/DZ5sW0N+SkEWMypTb7ahHxLeBi4B2UclH3UQb33ASsCOwEvBX42VB/w15HSKNXRHwT+GBELJGZ/+ksnFo3z045P/hiDdxfQVn7YRPgu5n5+XqMFSh9y3drWffToMye6fXnkWYkF1KVBlxEHEiprXozcBylXtp8lCmeAB/MzJ839v8IcBAlaAe4F9gtM79Zt//PYiaSxoaI2AvYAzgTOJZSQmJZYAvgzZRSEat1gvf6moOAtwFLAn+jnFAfX7c50k0aJSJif8oI14spfcRDlLVilqJcYB9BXVy1zpZ7M2WtmBUoC7XfBeyfmd+qx/N8Q+qD5/rbi4glKeXl3lCfuo8yyv2HmXlB3Wdxyvf9jcAKjVlykka5+v3+a8oNueuBNbuD94hYGrgjMx+sI9lPBE7LzI0bx9iVUru9492Z+dtefhapFwzdpQFWa7efSBll+oXOYoa1NvMuwP511w0z85TG65YBFqaUgLirU7PRgEwauyJiReC3lBHrG2TmNY1tMwMnA+tSLrJXy8z/NLbPRRnh/nBnpKv9iTR61JJ1PwXOB7bIzGvr8/NTLrz3BV4CHA4clpkT6/aZKTPq7gQeaLzO/kHqg66ZsnMCcwPjMvP2ThgfEYtRarQvCJxGCc8eaRxjPUp/cGCt5y5pDGj0ETMDPwA2pnFdUNeGerKx/zjgLOAtwBsy8x8RMVtmPhYRb6WUpPw7pezMUpl5fa8/kzSjWV5GGmxvro+HdQJ3ysJlZOaBEfEkZQGTkyNiyoj3GrI/q95y/RL1Algau14GvIhn+pNxnQWRMvPJiNgA+A1lOvmfI2K1zLy7XsA/CDzYOZD9iTTqLEE5vzgqM69tLOR+V0T8FHiUMqp9G2BcLTUzsV58/7x5IPsHqT+6AvdtKHWX3wRMjog/Az+LiFMz8x88swgyzZrNEbESZcbLI8CfkDRm1MB9fL0u2IyyBtRGlLWfOsF7s9TMvMCrKTPjHq/HeKxu+yQwb2auGRE71msJadRxoURpAEXE+Lrw2NL1qcegLHBWQ7LOxfBXKIuVQDmR/uBwx3SKtzTmdUpOvboulhoAdfHU8Zn5FKXMzC3AYpQa7wvUmTXPWgjR/kQadZak9AkP19+njFTPzMeB3wE/Al4KfBT4Yh1F+z8Lpdo/SL1Xb3Z1AveDKdcHSwC/Aq4EXg98Azg8Il7eeQ2U84D6+3sp6zW8hVKa8pxefw5J/VXP+8fXm+ofp8x6eS0leJ+3Xjd0btTdSxno9zJgqTpAkIjYgjKI56J6zfHw/7yRNEoYuksDqIbrCVxan1oGnhmJUoP38XXbbymjUZ4GfhIRm/S6vZIGwl+A/1Iuwid1Tqphygl2APdTRrTfSznBvioiZutXgyX1zA318e0w5WbclDC9jlDrlLF7FWWRxd3riDdDdqnPOn+HEfFpYEfKugzrZOaHgXcCq1GuF94DbFpDtc5r1o6I4yll5pYEtsnMr9dt5gnSGNG4ETepPj5n8F4H7JxJyR1PomQRvwa+TrmRf0BmPuXsN41mfklKA6hxoXtZffxSRMzTGZFan+tc5P6ZUkv1jPr7Aj1qpqTBcidwHbAipS5zJ2yfUEvNZGbeSwnffkkZHXdUZj5mqCaNer8BJgEfjoh1YMo082iEbpfWf4cDE4F/NqaYS+qRiJhjiOciImallIK4nzJS/W8AtV77hynrMvwV+GMnVKvuAtagLJ74kWbgblgmjQ2dG3H1umDJWtf9OYP3uv1YyoKpd1IWX1+VsiD7WzPz7/34LFIvuZCqNMDqhe5plFEp5wEfrLXUmjUb1weOp4xena3WaZSkKRoLI72ZcqNuJuDgzNyla7/VKDfwPp+Z32s874W3NEo1+od9gZ2BcykLKJ7Ttf1dwOnAOsAVmXlX3xotjVF1ccI9gO0z87KubQsCNwE/y8xNGs/vAexF+X7fPTMvr88vANyXmU9ExLwAWRdR93tfGjs6ddojYnbKjfX3UmbIfhiYXGfZz0zJHDaisbhq4xgLA4sDtwP/zcz/9vpzSP3gSHdpQDVOdj8OXEKZFvq7iFiOUlO1E5B9FrgHmJVyou1UUGmM6f6bb/7eCMwiMy+mnEA/BewUET+OiPdExHwR8R7KSBWAO7pe74W3NKC6a653b2vMZDkJ+AWwFrBfrcnaGfG+CqVkxb+AmzuBu+cbUu9ExCzAJpRR6YfWa4KmpygzYWdrvGZPSuD+a8ro98vr86+klIl6PZSwvRG4+70vjRF1MN/TdQbNH4BPUGbGfgOYpVPWdrgR7/UY4zLztsw8KzOvN3DXWOJId2nANC+AI+KlmXlfvXP8I2Bl4AHKHeTbgLcBswPbZuYxfWqypD5qjkaLiJWAy+rCh939yRLAXZn5QES8DfghMF89zGM8c5G+fWZ+tZefQdKM0TUzbvbMfLSxrdk/vDgzH4qIZYHPA5tR6rFeDjwOLE/pIz6XmUf19ENImiIiFgF2AT4N/AnYLjOvrNteBlxDWZulMzBnD0rg/uXMvKJxnK2AY4D3Z+ZpvfwMkkaWekPvVGBt4CDKbLdHu/YZX8tSdo94XzUz7+mMlu9x06W+M3SXBkjXBfCqwObAiZn5+ygrfx9ACd5XAu6j1GA8IjO/2/16SWNLRHwF+AiwL2V0SjTC+NUoI93uBz6WmY9FxGuB9Si1F+cFrgdOz8xT6mucWi4NsK5ziqOAV1IWSPx317Y1KTNgTs3MMyNiPspI2j2AlwFzA1cDx2TmD7qPLak3GjPXFqb8fW5OKT+5LXBtHa26D7Abpabymxk6cH8bcBylBvPGmXlTjz+KpBEkIj5AWUj5x8Ana8mpKTftG/sNFbzfCyzp6HaNVRP63QBJU2eIwP0QyoKHpwNkWR18x4iYC3gdJTx7MjNvr68xIJPGkK4+Y2NgS0ot5vPq883+5EDKDbtP1sA9MvNG4DDgsIiYAExqHM/+RBpwjb/nbYGtgasos+PoOt/Yj9I//Lpuuxv4aUScBYynjHB/JDPvr6+xf5D6Yxzlu/q2iNixPvcRSg3mHYArKIugrwO8EbgZOLwrcF8D2JNyQ20HA3dJlMwB4LAauA85ar0G7uMy88mI+BjwYuBdwFyAobvGJEN3aYQYblRY5+K1cQG8GmVE+4rAlpn5i679HqSMXuk+thfA0hgxROj1OsrJ7q6ZeVVjv+YNvC0y8/tDHaeOjptS99n+RBpcXSVlxlNK0f0Z2Dobi61HxOqUmTErU/qHXzaPk5kPNPaNzqP9g9R7XX/XHwLmB5agzHxdCzgwIr6UmRdHxMGUm2mLAZvU+u13UOq37wAsAHwhM39Sj+fMFWkMqt/tE4CF6lMvARiuTEydeT+57vNURKwHzJuZ/+pBc6URydBdGgG6p2dFxEuApzJzYvPiNSKWoSxkuArlAvjb9fnnvMj1RFkaWxplY75COUF+LfCzzmi2ehL9Usro1s4NvE5/MiWwb/Yr9iPS6NAI5j4BTATeCuzUdUPu1cDRwDLApxtl6oYcxd7pH+wnpN6r1wGdv+uDgO0oM17/AtxDGWn6LmC2iNg6M38aEY9QzgE+RilDAyUsuwX4VGZ+rx7PmSvSGFW/05+KiFvrU68G/tA90r2RZSwK7BQRW2Xm43UmvoG7xjRDd6nPukambA68E1gTeDgi/gKcCPw5MycCD9eXNS+AHX0i6X9ExKsoF9QJPESpyU5EzFRPgu+NiF8Cv8jMn9ZtjlKVxoCIeD/wHeBCyiyYf9TnOxfSt1PKUX2r63zD/kEaYRqzYT8FfIlSQmb3fGYB1f+jlJF7J/CNiPhsZp4REZcAS1GuO2alzHi5OTOvqa8zcJcEpTQVwG4RcXpm/qdRv705ePCLlDruX6dr5r00VrmQqtRHXTWXDwZ2pIw6u4OyoNmLgH8DJwEHZeZ/I+Il1k2VNDUiYgXKjbtFKCfMK9ea7Z3gvbmvN/CkMSIiFgc+B2xMWQj125m5Rd3WWYyxeY7i+YY0gtWyDidQAq81M/NP9flOMLYoJXjfEDgb2J6yuOqkYY7nOYEkoJwDUNZ1eQdwPrBBZv6na58tgYOBPwIfb5agk8aycf1ugDSWNS5mv0AJ3E+n1Fb9P0oJmf2AJyknxvt0Be6OOJM0rNpHXEi5AL8FWB74eidwr4ujTuHFtTR2ZOYNwFcpId3DwNoR8e66LbsDN883pBFvAqX0wz3A5TDlZtmk+vd8E7A7cCPlWuNgYOnmei1NnhNIiqqeA3wcuISSUfw5Ij4QEW+KiPkiYn9KbjER2N7AXXqGI92lPouIeYHfUUaavSczr2mMSnkRpdbqYZSR77sCxwy3eImksWdqRqBGxJuAkykLIR1Bqd/8dPd6EpJGl+cbtR4Rr6PUf/4MpbTEbpn5l+Zr+9V2SVMvImYDfgusCmzcKRvX2N7pCz4GHFefvgjYKjMv621rJY10XecKC2bmPyNiAcrN+rXqbgk8BcwM/B34QGZe25cGSyOUI92l/lsAWBY4uwbuE3hm1e+HgbOAvYGngQ3w71ZSVUPzzgnxMhGxbkRsGhHvrgsyA5CZlwAfpJSu+gJwcK3dPKlOGZU0ytT+oROavywiFo+I1wALdvbJzL8DXwO+DawG7B0RK9VtBu7SgMjMx4AfA5MoM1dm69qlM6L9Nsoiq6cCbwFe06s2ShoMXYH7VsDxEbFqZt5JKTGzFeW84c+UNSQ+D6xt4C79LxdSlfqvE3jNDdA9ij0zH4+Isyh3j1cF3gWc1ssGShp5OtPG68+7AtsC8zV2+XNEnJqZhwNk5sURsQHwc0rwnhGxsyPepdGna5H2bYBNgDcCjwGPRsRXgZ9l5s2Z+feIOLK+9DOUcnZ7ZOYF9fWOeJcGw68oazV8CrgB+MoQ+ywIPA4cTpk9e3bvmidppBjuu32IwP1AStmYm2BKublv1u1zZOYjvWu1NHgc3Sb1yHA1Eym1Fx8F3hoRqw/xuvGZeQ+lNATAHDOoiZIGSOOEuFNH8W5gZ2A34DzgDcBhNVzrvOZiyoyZOyhrRRxWa7wbuEujRL2Q7gTuB1NGsr8WOIdSj3U+ykX0VyNiLZgy4v2rwLeANYE9ImKVus3AXRoAmfkvyuLIk4BDI2LPiFiybptcZ7FsDdwF/K0TuDvjTRo7ImI8DP/d3ri+2Jpn6rSvkpl3dl7b2PeRuu9wOYc05jnSXeqB5xpFmpl3RMS3KSNTNoqIv2fmXfV1Exoj3+euj3fO8AZLGggR8QHgS8DZwOc60zoj4vvAGsDRwHYRMTEzd4MpI94/QFm4eTvKyPfz+tF+Se1r1G7/NGWR9lOBPTLz6vr8OsA2wLrArBHxaGb+NTNvjIgjKCXutgZeEhHv75yTSBr5MvOKiFiBcl6wJ+Xa4kLgAUqZuVcCn6nlaDqvcaFkaQzoZAsRMSvwEWBOSk3244FHO2vAAOsAe9Rtq2TmbV25xLN4c14anqG7NIN1lYDYjFIzdU7gauCIzJwI/Ax4J/BZ4LGIODYzb+p8sUXEisBGwK0Yukt6xpsoixgdlpnXdqaE1pqLJ0XEo8APgM0j4rzM/G3d55KIWB9YMjMN3KUBNtSN/VrP+b3AI8BBmXl1o384IyLuqts+QJky/leAzPxHRHwNeBFwqYG7NHgy89J67bA7JWhfsm66G9g6M78Dlo6SxpJ6DvB0RMwB/AZYpbF5Y+DoiDg9Mx+LiBsoN+52zszb63nGkIG7pOcWfs9KvRERB1FGpDadB+yZmedGxPspU7sXokwBP5vyhbgsZUTam4AtMvPbvWqzpJGrTvH8A6VO8xsz8/ru8C0iZqFcdO9KCd52HeZYU+o3ShoMEbEccGdm3tMdnkXEfMClwB2ZuUJ9LuBZI+FXoSy8uCDwvsw8vfH6uTLzwc7rDOakwRMRE4BFgdcBDwJ3Z+YNdZvf+9IYExEzU9Z+WBv4BXAWsB4lgL8fOAj4cWY+HBEzZ+aTrvskTR/rt0kzSLO2WUR8DNiBEqavA7yVUqN9Bcpd5Xdn5qnAFpQvv9WA/YG/UUapLkkpHfHt7mNLGpvqCfDdwGyUvoTuk+LMfIISzAO8vY5uGepYXnhLAyQilgcuB86PiBcNEYrPDMwEvCYiloIStjemjpOZ5wPH1v2Xq8ftbDNwlwZcZj6dmTdk5q8y80+NwD383pfGhnrzreO1wIqUtV4+npnfAj5BWQ+qM1DnoxExew3cxxm4S9PH8jLSDDDEHeGlKGVhdszMy+o+fweuBb4AfCUiyMwzI+JKygKI6wGzUxY9uzwzz62vc2SKNEYNEYD9hrIw6kci4oK6GGJn305fcQGl/xlHqdUsafBdQyk5dxelxBTwTL3WzPxnRPyeUj5mOaBZfiobtVmvqS9daqg3MXCXRh//rqWxo5aUeRHwUeA2yoCdvTPz0XoucHdEHA88DBwAfBnIiDi+7uPNd2k6GLpLM0CjhvvBlIvhNwLfy8zL6iiyyLIC+DHA05SFzg6LiHHA72u91TO7w3UDd2ls6f6bb5SF6NzY+x1ldsyalLrt38jM2zu718e3AvMDv2gunCZpMDUWQlsMmD0zJ9ZFU0+rF8+dQP18Sp3Wr0XElZ11Hyh9Q2dgwDyUm3F/AsM4SZJGi8bs+DOBlSmz6J8EJnVqvANk5kMRcQoQlNn2u9XXG7xL08nyMtIMEhGLU8L0TwPL8MxNrgmdEC0z7wG+CRwKvAo4mFICYpahwnUDd2nsqMH65Prz+yLisxGxV0QsQTkppgbsPwBup8ya2TEiVq3bMiJWovRDk4Hf9uFjSGpZDdw7o9YnRsSWlHOJn0TEyxsX0ccAPwReDpxTF1acqVFmZgXKmjEPU2beSZKkAVfXfZpSVo5S5vZBykDAWSg37Cc3S89k5sPAzykj3QPYCfhMRMxm4C5NOxdSlWagiFgXOB6YGzgjM99bn3/W6NWImIdSz/0LwH+BXYBTDdmlsak5oiQi9qf0CR0PAIcDJ2XmP+o+mwHbAf9HCdhPpZwwrwm8BNg+M7/ai7ZL6q160XwuZRTbucCH6k39zvafUcpQPQicTik5NTuljusSwHaZeXRvWy1JktrWyRkiYlZgIeCWzHwqIl4P/BGYE/hhZn687v+ssri1FM36wLeAm4BVMvOBXn8OabQwdJdmsIh4J3ASMBewX2buUZ8fKnjfGtgD+Fhm/qgf7ZU0ckTEdsBXgUuB7wDLAm8DXg18Fzg6M6+t+64OvI8Svk+ghO9/q/ucUPexRJU0ijRKzUygLNa+KkMH74cC7wcWbbz8LmCfzDy27mP/IEnSgIuImYGLgPHAJsC19VxheUo5uRcBh2fmF+v+3cH7i4F1gYsz88aefwBpFDF0l3ogIt4F/ASYFdgzMw+qz3cH7/MBC2Xmxf1pqaR+GuKk9xzKifEnMvOqWptxbUrJmLUowftXM/O6xmsWo0wdfRJ4oBO8GahJo0fXbJiZM/PJqQjeX0eZDfMq4B/AzZl5dd1m/yBJ0igQEfMDR1Nutp8HfB64pgbvzRHvR2TmDvU13dcg1nGXWmDoLvVIRLyHMuJ9NmCPzDywPj/kha4XwNLYFRHbAhcCX6OMVP9hY3FEIuItwN7A2ynB+xGZeX3dNqXv6Jwwe+IsDbYhbtLPlJlPNX6fNTMff77gfZhj2z9IkjSKRMRClBn0n6CMbv8cLzB4lzT9XEhV6pHM/DWwMfAYsHdE7FKfn9xZ7KRrfwN3aRSLiCUiYs4hnv8YcCRwHKWMzAN105ST4My8CNgL+B3wSeALEfHaum1yY79sPkoaPF2LKm8UEUcCF0XEURHxYYAauM9Ub8ytBfyZsqbDT2r5uk7td+qMmSnsHyRJGkzN7/TOz/Vm+u3APsD3gNUp1xZL10E8lwFrABMp1xCHABi4S+0zdJd6qBG8Pw7sHhG71ef9gpPGkFpT8VrgV3Whoylq/fVfA4sDLwNeWjeNq6+Nut+FPBO8bwrsHBFL9qL9knqjjnCfVH/enzJjbivK4mhbAD+KiAMB6kJpwwbvdXTbeEN2SZIGXw3QszOAr/48rjHLtRO8f58SvH+VErzP1Aje7we+GBH79uljSKOaobvUYzV434iysMk+EfGmPjdJUu9NAm6h3ICb8l1cFz4iM9cFTq3bvhIRi2XmpOaJdN3vQmBPSri2ObBATz+FpBmqMcJ9N2AXSm3WdYBXAOsD9wE7RcTXOiVnhgneT4qIl3f6kX58FkmS1J56M30O4LSI+Gx9bvIQwfvewPGUkP1gnh28vwu4A/hpnz6GNKpZ013qk4hYD1ggM4/td1sk9U6jzvqCwP2ZOTEiPgKcnpkPNWs1R8RPgQ2By4CNMvOmTm3nroUUVwbmycxf9utzSZox6vnC94BLgS9l5mURMQvwMcqotZmBCcAxwA51UdVOAD+eUuN9NeA0YLPMfKAPH0OSJLUsIj4I/AS4CdgvM4+rzz/reiEiFqcM6Fkc+C2wE3BdPVeYJTOf6NNHkEY1R7pIfZKZv+wE7o46k8aORp31f9bAfQvgh8B3ImLOzkjVus9GwC+A1wMnR8Si3SNY6n5/6QTu9ifS6BERswPrAbMD+9TAfSZgW0rg/i/gA5RRalsDB0fEzI0R75MoCy5fALyv/vufuu6SJGkg/ZJSdu4VwF4RsRkMOeL9Bkpddygz4X5ICeABnuxxm6UxwwtzaQRw0VRpTLuQUgJiQ4YO3jegBO/LM0Tw3n0w+xNpVJkAzArsmZnn1ZHrmwC7AncDq2bm6cCOdf/PUUpSzdQYvfYk8KW6fRlw8VRJkkaD+h3/PWAHYD5gz+7gnVLWFsrCqf8EfkVZG+bhup/nBNIMYuguSVKPNEeXRsS4OvLkcsoIlT9R1nt4vuD9pFrj3XBdGkW6+ofOLJaHgJ0pC6gCzE1ZQPVx4F2ZeU9dC+Jc4FbgTsqI9+/UG3Od6eJvr49PRMR4R7pLkjQ61LKU3wW2Z4jgva7zArA2cD3lPGHJzLytD82VxhRDd0mSeiAixjdqsAcw5ffMvAbYjucP3k8G3gicFRFzGZxJo8MQ/cOEzrbMvL0uhAawLrAC8JXMvLEzkj0z7wYeotRvvwm4trEI6/yUElVPAt/PzEmOapMkafRoBO+dEe8HRMSOUYyrC62uC9yemXdn5p39bK80Vkx4/l0kSdL0qCNOJ9WftwLeBswdEedTpoT+OzOvjIjtgK9Rgnci4lO17vtMmflUZm4UEWcBf87MB/v0cSS16Hn6h+8C/8rMSXWK+AL1ZU8CdEayR8Q6wFLAJ4HrM/ORxlv8FzgM+EJm3tyLzyRJknqrDtb5DvAUcARwMKV8ZQJvAf4DHNK/FkpjTzjQRZKk3oiI/YFdup6+kLKw0amZ+XhELEcJ3lcHfgo8K3jvOl44YlUaHZ6jfziC0j88GREbAD8DfgIclJlXRMSqwJ7AEsB7a8kq+wdJksaoiFiTcj2xCDAJuBr4dGZe379WSWOPobskST0QEWsBpwB/AfanjEL5WP13D3AAcOIQwfuJwJaZ+XAzRDNQk0aPqegf9qf0BbMCPwDeB9wM/J3ST8wBbJeZR/e67ZIkaeSJiLmBeYEA7nKWrNR7lpeRJKk3lqGspbJbZl4KEBG3UkKzPYEv1+dObJSaOQL4MDBXRKzfHOlu4C6NKs/XP+wOTAZOAPYC7gM2p5SbuR04ODOPq6/zhpwkSWNcZj4APNDnZkhjmqG7JEktq4siTup6enbg5My8NCImAJMy8z8RcQLwNLAv/xu8bw98H/hTd2kZSYNpOvqHvSnB+48z85MRcTTwGPBEZt5Sjz2us4CqJEmSpP4xdJckqUXNQC0i3k+ps/wwsArwsoiYPTMf7eyfmQ9ExIn1107wPjkiflLrNb89M++tx3MEqzTAWugf9ikvjZ9l5mX1ONF5NHCXJEmSRgZDd0mSWtQI1A4Adu7afDWwJPC3rtc0g7U9KeHabBHxbUoZCQN3aRRoqX84AJg9Ir6dmZM7/YL9gyRJkjRyjOt3AyRJGg06o03rz5sBOwLnAZ8BtgT+TanbvEdEzNX9+lp38URgP+CVwOOZOclATRp8LfcPr6CUlHFUuyRJkjRChdfwkiRNn+4azRFxGPB+4AOZeWV9bmngaGAN4BfAJzLzwSGO9RLgVZl5VS/aLmnGsn+QJEmSxh5Dd0mSWhIRB9cfVwDOzsx9I2Ic5ft2UkQsQQnW1uI5grXG8VwUURol7B8kSZKkscOa7pIktSAiFqeUjLgPeAo4t24an5lP1Zrs10fENpRgbf36us0z86GhjmmgJo0O9g+SJEnS2GJNd0mSWpCZNwDvo3y3zge8oT7/VB2Rmp1gDdgGOIcSrP0wIubuU7Ml9YD9gyRJkjS2GLpLktSSzDwd+DDwILBuROxTn588RLC2NfBnYF1g5b41WlJP2D9IkiRJY4c13SVJallEvAv4CTAbsGdmHlifH1cDtqgB29LA6zLzF/1sr6TesX+QJEmSRj9Dd0mSZoCIeA9wEiVY22O4YK2xv4siSmOE/YMkSZI0uhm6S5I0gzSCtVl59ojW8Zk5qa+Nk9RX9g+SJEnS6GXoLknSDNQI1iYAB2Tmfn1ukqQRwv5BkiRJGp0M3SVJmsFqDedfAjMBb8nMS/rcJEkjhP2DJEmSNPoYukuS1AMRsR6wQGYe2++2SBpZ7B8kSZKk0cXQXZKkHnNRREnDsX+QJEmSBp+huyRJkiRJkiRJLRnX7wZIkiRJkiRJkjRaGLpLkiRJkiRJktQSQ3dJkiRJkiRJklpi6C5JkiRJkiRJUksM3SVJkiRJkiRJaomhuyRJkiRJkiRJLTF0lyRJkiRJkiSpJYbukiRJkiRJkiS1xNBdkiRJkiRJkqSWGLpLkiRJkiRJktQSQ3dJkiRJkiRJklpi6C5JkiRJkiRJUksM3SVJkiRJkiRJaomhuyRJkiRJkiRJLTF0lyRJkiRJkiSpJf8P4LWssBT7C3EAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1800x900 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "num_classes = len(damage_classes)\n",
    "\n",
    "x = np.arange(num_classes)  # the label locations\n",
    "width = 0.8/num_classes  # the width of the bars\n",
    "\n",
    "fig, ax = plt.subplots(1, 1, figsize=(12, 6), dpi=150)\n",
    "\n",
    "for i, (disaster, li) in enumerate(wind_post_areas.items()):\n",
    "    rect = ax.bar(x - (i - 2) * width, li, width, label=disaster)\n",
    "\n",
    "_ = ax.set_ylabel('Building area (pixels)')\n",
    "_ = ax.set_title('Class distribution by building area')\n",
    "_ = ax.set_xticks(x)\n",
    "x_labels = ax.set_xticklabels(damage_classes, rotation=45)\n",
    "_ = ax.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Splitting into train/val/test - wind disasters\n",
    "\n",
    "We do not use the splits released by the xView2 competition.\n",
    "\n",
    "We split at the tile level (1024 by 1024 pixels), not at the chip level. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 5598/5598 [00:00<00:00, 1837510.86it/s]\n",
      "100%|██████████| 1866/1866 [00:00<00:00, 1844584.32it/s]\n",
      "100%|██████████| 1866/1866 [00:00<00:00, 1797558.86it/s]\n",
      "100%|██████████| 12738/12738 [00:00<00:00, 2251930.22it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "hurricane-michael, 550\n",
      "hurricane-matthew, 405\n",
      "joplin-tornado, 149\n",
      "moore-tornado, 227\n",
      "tuscaloosa-tornado, 343\n",
      "CPU times: user 25.9 ms, sys: 0 ns, total: 25.9 ms\n",
      "Wall time: 23.9 ms\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "xbd_root = './xBD/'\n",
    "wind_disasters = ['hurricane-matthew', 'tuscaloosa-tornado', 'hurricane-michael', 'moore-tornado', 'joplin-tornado']\n",
    "label_dirs = [\n",
    "    xbd_root + 'raw/train/labels',\n",
    "    xbd_root + 'raw/test/labels',\n",
    "    xbd_root + 'raw/hold/labels',\n",
    "    xbd_root + 'raw_tier3/labels'\n",
    "]\n",
    "all_files = defaultdict(list)  # wind disaster to list of files (no extension)\n",
    "\n",
    "wind_disasters_tup = tuple(wind_disasters)\n",
    "\n",
    "for label_dir in label_dirs:\n",
    "    for p in tqdm(os.listdir(label_dir)):\n",
    "        \n",
    "        if not p.startswith(wind_disasters_tup):\n",
    "            continue\n",
    "        \n",
    "        if not p.endswith('_post_disaster.json'):\n",
    "            continue\n",
    "        \n",
    "        full_path = os.path.join(label_dir, p)\n",
    "        rel_path = full_path.split(xbd_root)[1]\n",
    "        \n",
    "        # example: hurricane-matthew_00000000_post_disaster.json\n",
    "        disaster_name = p.split('_')[0]\n",
    "        file = rel_path.split('_post_disaster.json')[0]\n",
    "        \n",
    "        all_files[disaster_name].append(file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "hurricane-matthew, 405\n",
      "hurricane-michael, 550\n",
      "joplin-tornado, 149\n",
      "moore-tornado, 227\n",
      "tuscaloosa-tornado, 343\n"
     ]
    }
   ],
   "source": [
    "for disaster_name, files in all_files.items():\n",
    "    print(f'{disaster_name}, {len(files)}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "hurricane-matthew, train 324, val 41, test 40\n",
      "hurricane-michael, train 440, val 55, test 55\n",
      "joplin-tornado, train 120, val 15, test 14\n",
      "moore-tornado, train 182, val 23, test 22\n",
      "tuscaloosa-tornado, train 275, val 35, test 33\n"
     ]
    }
   ],
   "source": [
    "# split randomly\n",
    "\n",
    "wind_splits = defaultdict(dict)  # wind disaster to train/val/test\n",
    "\n",
    "for disaster_name, files in all_files.items():\n",
    "    shuffle(files)\n",
    "    \n",
    "    num_train_tiles = math.ceil(0.8 * len(files))\n",
    "    num_val_tiles = math.ceil(0.1 * len(files))\n",
    "    \n",
    "    wind_splits[disaster_name]['train'] = sorted(files[:num_train_tiles])\n",
    "    wind_splits[disaster_name]['val'] = sorted(files[num_train_tiles:num_train_tiles + num_val_tiles])\n",
    "    wind_splits[disaster_name]['test'] = sorted(files[num_train_tiles + num_val_tiles:])\n",
    "    \n",
    "    print(f\"{disaster_name}, train {len(wind_splits[disaster_name]['train'])}, val {len(wind_splits[disaster_name]['val'])}, test {len(wind_splits[disaster_name]['test'])}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('./constants/splits/wind_disaster_splits.json', 'w') as f:\n",
    "    json.dump(wind_splits, f, indent=4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Splitting into train/val/test - LOO subset disasters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 5598/5598 [00:00<00:00, 1544311.61it/s]\n",
      "100%|██████████| 1866/1866 [00:00<00:00, 1553507.60it/s]\n",
      "100%|██████████| 1866/1866 [00:00<00:00, 1559388.58it/s]\n",
      "100%|██████████| 12738/12738 [00:00<00:00, 2341032.53it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 21.8 ms, sys: 3.09 ms, total: 24.9 ms\n",
      "Wall time: 22.7 ms\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "xbd_root = './xBD/'\n",
    "\n",
    "wind_disasters = ['hurricane-matthew', 'tuscaloosa-tornado', 'hurricane-michael', 'moore-tornado', 'joplin-tornado']\n",
    "\n",
    "flood_disasters = ['midwest-flooding', 'nepal-flooding'] + ['hurricane-harvey', 'hurricane-florence']\n",
    "\n",
    "earthquake_disasters = ['mexico-earthquake', 'palu-tsunami', 'sunda-tsunami']\n",
    "\n",
    "others = ['guatemala-volcano', 'lower-puna-volcano', 'socal-fire', 'santa-rosa-wildfire','pinery-bushfire', 'portugal-wildfire', 'woolsey-fire']\n",
    "\n",
    "subset_disasters = flood_disasters\n",
    "label_dirs = [\n",
    "    xbd_root + 'raw/train/labels',\n",
    "    xbd_root + 'raw/test/labels',\n",
    "    xbd_root + 'raw/hold/labels',\n",
    "    xbd_root + 'raw_tier3/labels'\n",
    "]\n",
    "all_files = defaultdict(list)  # subset disaster to list of files (no extension)\n",
    "\n",
    "subset_disasters_tup = tuple(subset_disasters)\n",
    "\n",
    "for label_dir in label_dirs:\n",
    "    for p in tqdm(os.listdir(label_dir)):\n",
    "        \n",
    "        if not p.startswith(subset_disasters_tup):\n",
    "            continue\n",
    "        \n",
    "        if not p.endswith('_post_disaster.json'):\n",
    "            continue\n",
    "        \n",
    "        full_path = os.path.join(label_dir, p)\n",
    "        rel_path = full_path.split(xbd_root)[1]\n",
    "        \n",
    "        # example: hurricane-matthew_00000000_post_disaster.json\n",
    "        disaster_name = p.split('_')[0]\n",
    "        file = rel_path.split('_post_disaster.json')[0]\n",
    "        \n",
    "        all_files[disaster_name].append(file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "midwest-flooding, 445\n",
      "hurricane-harvey, 522\n",
      "hurricane-florence, 546\n",
      "nepal-flooding, 619\n"
     ]
    }
   ],
   "source": [
    "for disaster_name, files in all_files.items():\n",
    "    print(f'{disaster_name}, {len(files)}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "# split leave one disaster out\n",
    "disaster_to_leave_out = 'nepal-flooding' #'nepal-flooding' # 'joplin-tornado'\n",
    "\n",
    "subset_splits = defaultdict(dict)  # subset disaster to train/val/test\n",
    "\n",
    "for disaster_name, files in all_files.items():\n",
    "    \n",
    "    if disaster_name == disaster_to_leave_out:\n",
    "        subset_splits[disaster_name]['train'] = []\n",
    "        subset_splits[disaster_name]['val'] = []\n",
    "        subset_splits[disaster_name]['test'] = files\n",
    "\n",
    "    else:\n",
    "        shuffle(files)\n",
    "        num_train_tiles = math.ceil(0.8 * len(files))\n",
    "        num_val_tiles = math.ceil(0.2 * len(files))\n",
    "        subset_splits[disaster_name]['train'] = sorted(files[:num_train_tiles])\n",
    "        subset_splits[disaster_name]['val'] = sorted(files[num_train_tiles:num_train_tiles + num_val_tiles])\n",
    "        subset_splits[disaster_name]['test'] = []\n",
    "    \n",
    "    print(f\"{disaster_name}, train {len(wind_splits[disaster_name]['train'])}, val {len(wind_splits[disaster_name]['val'])}, test {len(wind_splits[disaster_name]['test'])}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('./nlrc.building-damage-assessment/constants/splits_generalizability/LOO_' + disaster_to_leave_out + '_subset_disaster_splits.json', 'w') as f:\n",
    "    json.dump(subset_splits, f, indent=4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Splitting into train/val/test - LOO all disasters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 5598/5598 [00:00<00:00, 762798.93it/s]\n",
      "100%|██████████| 1866/1866 [00:00<00:00, 725219.72it/s]\n",
      "100%|██████████| 1866/1866 [00:00<00:00, 766634.47it/s]\n",
      "100%|██████████| 12738/12738 [00:00<00:00, 762339.57it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 39.4 ms, sys: 6.25 ms, total: 45.7 ms\n",
      "Wall time: 43.4 ms\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "xbd_root = './xBD/'\n",
    "all_disasters =['guatemala-volcano', 'hurricane-florence', 'hurricane-harvey', 'hurricane-matthew', 'hurricane-michael', 'mexico-earthquake', 'midwest-flooding', 'palu-tsunami', 'santa-rosa-wildfire', 'socal-fire', 'joplin-tornado', 'lower-puna-volcano', 'moore-tornado', 'nepal-flooding', 'pinery-bushfire', 'portugal-wildfire', 'sunda-tsunami', 'tuscaloosa-tornado', 'woolsey-fire']\n",
    "label_dirs = [\n",
    "    xbd_root + 'raw/train/labels',\n",
    "    xbd_root + 'raw/test/labels',\n",
    "    xbd_root + 'raw/hold/labels',\n",
    "    xbd_root + 'raw_tier3/labels'\n",
    "]\n",
    "all_files = defaultdict(list)  # wind disaster to list of files (no extension)\n",
    "\n",
    "all_disasters_tup = tuple(all_disasters)\n",
    "\n",
    "for label_dir in label_dirs:\n",
    "    for p in tqdm(os.listdir(label_dir)):\n",
    "        \n",
    "        if not p.startswith(all_disasters_tup):\n",
    "            continue\n",
    "        \n",
    "        if not p.endswith('_post_disaster.json'):\n",
    "            continue\n",
    "        \n",
    "        full_path = os.path.join(label_dir, p)\n",
    "        rel_path = full_path.split(xbd_root)[1]\n",
    "        \n",
    "        # example: hurricane-matthew_00000000_post_disaster.json\n",
    "        disaster_name = p.split('_')[0]\n",
    "        file = rel_path.split('_post_disaster.json')[0]\n",
    "        \n",
    "        all_files[disaster_name].append(file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "palu-tsunami, 196\n",
      "hurricane-michael, 550\n",
      "midwest-flooding, 445\n",
      "mexico-earthquake, 193\n",
      "socal-fire, 1403\n",
      "santa-rosa-wildfire, 377\n",
      "hurricane-matthew, 405\n",
      "hurricane-harvey, 522\n",
      "guatemala-volcano, 28\n",
      "hurricane-florence, 546\n",
      "lower-puna-volcano, 291\n",
      "nepal-flooding, 619\n",
      "joplin-tornado, 149\n",
      "moore-tornado, 227\n",
      "pinery-bushfire, 1845\n",
      "portugal-wildfire, 1869\n",
      "sunda-tsunami, 148\n",
      "tuscaloosa-tornado, 343\n",
      "woolsey-fire, 878\n"
     ]
    }
   ],
   "source": [
    "for disaster_name, files in all_files.items():\n",
    "    print(f'{disaster_name}, {len(files)}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "# split leave one disaster out\n",
    "disaster_to_leave_out = 'joplin-tornado' #'nepal-flooding' # 'joplin-tornado'\n",
    "all_splits = defaultdict(dict)  # all disaster to train/val/test\n",
    "\n",
    "for disaster_name, files in all_files.items():\n",
    "    \n",
    "    if disaster_name == disaster_to_leave_out:\n",
    "        all_splits[disaster_name]['train'] = []\n",
    "        all_splits[disaster_name]['val'] = []\n",
    "        all_splits[disaster_name]['test'] = files\n",
    "\n",
    "    else:\n",
    "        shuffle(files)\n",
    "        num_train_tiles = math.ceil(0.8 * len(files))\n",
    "        num_val_tiles = math.ceil(0.2 * len(files))\n",
    "        all_splits[disaster_name]['train'] = sorted(files[:num_train_tiles])\n",
    "        all_splits[disaster_name]['val'] = sorted(files[num_train_tiles:num_train_tiles + num_val_tiles])\n",
    "        all_splits[disaster_name]['test'] = []\n",
    "    \n",
    "    print(f\"{disaster_name}, train {len(all_splits[disaster_name]['train'])}, val {len(all_splits[disaster_name]['val'])}, test {len(all_splits[disaster_name]['test'])}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('./nlrc.building-damage-assessment/constants/splits_generalizability/LOO_' + disaster_to_leave_out + '_all_disaster_splits.json', 'w') as f:\n",
    "    json.dump(all_splits, f, indent=4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Splittnig into train/val/test - all disasters (SG: final model deliverable 90%/10%/0%)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 5598/5598 [00:00<00:00, 787804.11it/s]\n",
      "100%|██████████| 1866/1866 [00:00<00:00, 774447.98it/s]\n",
      "100%|██████████| 1866/1866 [00:00<00:00, 778608.36it/s]\n",
      "100%|██████████| 12738/12738 [00:00<00:00, 794017.33it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 44.3 ms, sys: 3.24 ms, total: 47.6 ms\n",
      "Wall time: 50.4 ms\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "xbd_root = './xBD/'\n",
    "all_disasters =['guatemala-volcano', 'hurricane-florence', 'hurricane-harvey', 'hurricane-matthew', 'hurricane-michael', 'mexico-earthquake', 'midwest-flooding', 'palu-tsunami', 'santa-rosa-wildfire', 'socal-fire', 'joplin-tornado', 'lower-puna-volcano', 'moore-tornado', 'nepal-flooding', 'pinery-bushfire', 'portugal-wildfire', 'sunda-tsunami', 'tuscaloosa-tornado', 'woolsey-fire']\n",
    "label_dirs = [\n",
    "    xbd_root + 'raw/train/labels',\n",
    "    xbd_root + 'raw/test/labels',\n",
    "    xbd_root + 'raw/hold/labels',\n",
    "    xbd_root + 'raw_tier3/labels'\n",
    "]\n",
    "all_files = defaultdict(list)  # wind disaster to list of files (no extension)\n",
    "\n",
    "all_disasters_tup = tuple(all_disasters)\n",
    "\n",
    "for label_dir in label_dirs:\n",
    "    for p in tqdm(os.listdir(label_dir)):\n",
    "        \n",
    "        if not p.startswith(all_disasters_tup):\n",
    "            continue\n",
    "        \n",
    "        if not p.endswith('_post_disaster.json'):\n",
    "            continue\n",
    "        \n",
    "        full_path = os.path.join(label_dir, p)\n",
    "        rel_path = full_path.split(xbd_root)[1]\n",
    "        \n",
    "        # example: hurricane-matthew_00000000_post_disaster.json\n",
    "        disaster_name = p.split('_')[0]\n",
    "        file = rel_path.split('_post_disaster.json')[0]\n",
    "        \n",
    "        all_files[disaster_name].append(file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "palu-tsunami, 196\n",
      "hurricane-michael, 550\n",
      "midwest-flooding, 445\n",
      "mexico-earthquake, 193\n",
      "socal-fire, 1403\n",
      "santa-rosa-wildfire, 377\n",
      "hurricane-matthew, 405\n",
      "hurricane-harvey, 522\n",
      "guatemala-volcano, 28\n",
      "hurricane-florence, 546\n",
      "lower-puna-volcano, 291\n",
      "nepal-flooding, 619\n",
      "joplin-tornado, 149\n",
      "moore-tornado, 227\n",
      "pinery-bushfire, 1845\n",
      "portugal-wildfire, 1869\n",
      "sunda-tsunami, 148\n",
      "tuscaloosa-tornado, 343\n",
      "woolsey-fire, 878\n"
     ]
    }
   ],
   "source": [
    "for disaster_name, files in all_files.items():\n",
    "    print(f'{disaster_name}, {len(files)}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "palu-tsunami, train 177, val 19\n",
      "hurricane-michael, train 495, val 55\n",
      "midwest-flooding, train 401, val 44\n",
      "mexico-earthquake, train 174, val 19\n",
      "socal-fire, train 1263, val 140\n",
      "santa-rosa-wildfire, train 340, val 37\n",
      "hurricane-matthew, train 365, val 40\n",
      "hurricane-harvey, train 470, val 52\n",
      "guatemala-volcano, train 26, val 2\n",
      "hurricane-florence, train 492, val 54\n",
      "lower-puna-volcano, train 262, val 29\n",
      "nepal-flooding, train 558, val 61\n",
      "joplin-tornado, train 135, val 14\n",
      "moore-tornado, train 205, val 22\n",
      "pinery-bushfire, train 1661, val 184\n",
      "portugal-wildfire, train 1683, val 186\n",
      "sunda-tsunami, train 134, val 14\n",
      "tuscaloosa-tornado, train 309, val 34\n",
      "woolsey-fire, train 791, val 87\n"
     ]
    }
   ],
   "source": [
    "other_splits = defaultdict(dict)  # disasters other than the wind disasters to train/val/test\n",
    "\n",
    "for disaster_name, files in all_files.items():\n",
    "\n",
    "    \n",
    "    shuffle(files)\n",
    "    \n",
    "    num_train_tiles = math.ceil(0.9 * len(files))\n",
    "    num_val_tiles = math.ceil(0.1 * len(files))\n",
    "    \n",
    "    other_splits[disaster_name]['train'] = sorted(files[:num_train_tiles])\n",
    "    other_splits[disaster_name]['val'] = sorted(files[num_train_tiles:num_train_tiles + num_val_tiles])\n",
    "    \n",
    "    print(f\"{disaster_name}, train {len(other_splits[disaster_name]['train'])}, val {len(other_splits[disaster_name]['val'])}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('./nlrc.building-damage-assessment/constants/splits/final_mdl_all_disaster_splits.json', 'w') as f:\n",
    "    json.dump(other_splits, f, indent=4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Splitting into train/val/test - all disasters (initial experiments with model 80%/10%/10%)\n",
    "\n",
    "Including the rest of the disasters, but consistent with the above wind disasters splits."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "ename": "ModuleNotFoundError",
     "evalue": "No module named 'sas_blob_utils'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mModuleNotFoundError\u001b[0m                       Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-12-583615fa87de>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;31m# listing blobs using the blob SDK instead of mounting it\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 3\u001b[0;31m \u001b[0;32mimport\u001b[0m \u001b[0msas_blob_utils\u001b[0m  \u001b[0;31m# https://github.com/microsoft/ai4eutils/blob/master/sas_blob_utils.py\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mModuleNotFoundError\u001b[0m: No module named 'sas_blob_utils'"
     ]
    }
   ],
   "source": [
    "# listing blobs using the blob SDK instead of mounting it\n",
    "\n",
    "import sas_blob_utils  # https://github.com/microsoft/ai4eutils/blob/master/sas_blob_utils.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "container_uri = 'https://researchlabwusstorage.blob.core.windows.net/nlrc-damage-assessment?SAS_KEY'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "0it [00:00, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "listing blobs...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "60170it [01:09, 863.02it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Enumerated 11034 matching blobs out of 60170 total\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "list_label_jsons = sas_blob_utils.list_blobs_in_container(\n",
    "    container_uri=container_uri,\n",
    "    blob_prefix='public_datasets/xBD/raw',  # includes raw/ and raw_tier3/\n",
    "    blob_suffix='_post_disaster.json'\n",
    ")  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'public_datasets/xBD/raw/hold/labels/hurricane-florence_00000429_post_disaster.json'"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list_label_jsons[100]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "disaster_to_li = defaultdict(list)\n",
    "\n",
    "for full_path in list_label_jsons:\n",
    "    rel_path = full_path.split('public_datasets/xBD/')[1]\n",
    "    tile_name = rel_path.split('_post_disaster.json')[0]\n",
    "    basename = os.path.basename(rel_path)\n",
    "    disaster_name = basename.split('_')[0]\n",
    "    disaster_to_li[disaster_name].append(tile_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['guatemala-volcano', 'hurricane-florence', 'hurricane-harvey', 'hurricane-matthew', 'hurricane-michael', 'mexico-earthquake', 'midwest-flooding', 'palu-tsunami', 'santa-rosa-wildfire', 'socal-fire', 'joplin-tornado', 'lower-puna-volcano', 'moore-tornado', 'nepal-flooding', 'pinery-bushfire', 'portugal-wildfire', 'sunda-tsunami', 'tuscaloosa-tornado', 'woolsey-fire'])"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "disaster_to_li.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'raw_tier3/labels/pinery-bushfire_00000100'"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "disaster_to_li['pinery-bushfire'][100]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "guatemala-volcano, train 23, val 3, test 2\n",
      "hurricane-florence, train 437, val 55, test 54\n",
      "hurricane-harvey, train 418, val 53, test 51\n",
      "mexico-earthquake, train 155, val 20, test 18\n",
      "midwest-flooding, train 356, val 45, test 44\n",
      "palu-tsunami, train 157, val 20, test 19\n",
      "santa-rosa-wildfire, train 302, val 38, test 37\n",
      "socal-fire, train 1123, val 141, test 139\n",
      "lower-puna-volcano, train 233, val 30, test 28\n",
      "nepal-flooding, train 496, val 62, test 61\n",
      "pinery-bushfire, train 1476, val 185, test 184\n",
      "portugal-wildfire, train 1496, val 187, test 186\n",
      "sunda-tsunami, train 119, val 15, test 14\n",
      "woolsey-fire, train 703, val 88, test 87\n"
     ]
    }
   ],
   "source": [
    "other_splits = defaultdict(dict)  # disasters other than the wind disasters to train/val/test\n",
    "\n",
    "for disaster_name, files in disaster_to_li.items():\n",
    "    \n",
    "    if disaster_name in wind_disasters:\n",
    "        continue\n",
    "    \n",
    "    shuffle(files)\n",
    "    \n",
    "    num_train_tiles = math.ceil(0.8 * len(files))\n",
    "    num_val_tiles = math.ceil(0.1 * len(files))\n",
    "    \n",
    "    other_splits[disaster_name]['train'] = sorted(files[:num_train_tiles])\n",
    "    other_splits[disaster_name]['val'] = sorted(files[num_train_tiles:num_train_tiles + num_val_tiles])\n",
    "    other_splits[disaster_name]['test'] = sorted(files[num_train_tiles + num_val_tiles:])\n",
    "    \n",
    "    print(f\"{disaster_name}, train {len(other_splits[disaster_name]['train'])}, val {len(other_splits[disaster_name]['val'])}, test {len(other_splits[disaster_name]['test'])}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "14"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(other_splits)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('../constants/splits/wind_disaster_splits.json') as f:\n",
    "    wind_splits = json.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['hurricane-matthew', 'hurricane-michael', 'joplin-tornado', 'moore-tornado', 'tuscaloosa-tornado'])"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wind_splits.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "for disaster_name, splits in wind_splits.items():\n",
    "    other_splits[disaster_name] = splits"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "19"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(other_splits)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('../constants/splits/all_disaster_splits.json', 'w') as f:\n",
    "    json.dump(other_splits, f, indent=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}