preprocessing.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Gun Detection System for Urban Settings (_Data cleaning, pre processing and augmentation_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook contains the description and the pre-processing details of the dataset used in the implemetation of th **gun detection system** for the detection of gunshots sounds in urban settings.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import librosa\n",
    "import numpy as np\n",
    "from numpy import nan\n",
    "import pandas as pd\n",
    "import matplotlib\n",
    "from matplotlib import pyplot as plt\n",
    "import seaborn as sns\n",
    "import math\n",
    "from PIL import Image\n",
    "import os\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "matplotlib.use(\"Agg\")\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, some constant values are set to be used across the entire notebook"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "IMG_DIR = \"./mel-specs\"\n",
    "VAL_DIR = \"./mel-specs-val\"\n",
    "TEST_DIR = \"./mel-specs-test\"\n",
    "\n",
    "SAMPLE_RATE = 22050 # Justification for this value is provided later"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## I. Explanatory data analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The dataset consists of 8732 labelled sound files, each with a length of less than or equal to 4 seconds. The sound files represent urban sounds from different scenarios. The sounds are classified into ten different classes: air conditioner, car horn, children playing, dog barking, drilling, idle engine, gunshot, jackhammer, siren, and street music. The audio dataset comprises audio files that are extracted from longer audio recordings, containing only the relevant sounds of the intended action and without additional silence or unnecessary noise.\n",
    "\n",
    "\n",
    "For the current task, only one class is relevant out of the ten classes, the gunshot class. Detecting a gunshot sound is relevant to gunshot detection systems. By identifying and isolating the gunshot class from the rest of the urban sounds, machine learning models can be trained to recognize the unique acoustic characteristics of a gunshot sound."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The csv file called `data-description.csv` contains a description of all audio files included in the dataset. To perform explanatory data analysis, the csv file is loaded into pandas DataFrame."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(8732, 8)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df = pd.read_csv('./data-description.csv')\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_map = {\n",
    "    'air_conditioner': 0,\n",
    "    'car_horn': 1,\n",
    "    'children_playing': 2,\n",
    "    'dog_bark': 3,\n",
    "    'drilling': 4,\n",
    "    'engine_idling': 5,\n",
    "    'gun_shot': 6,\n",
    "    'jackhammer': 7,\n",
    "    'siren': 8,\n",
    "    'street_music': 9,\n",
    "}\n",
    "\n",
    "def get_class_name(idx):\n",
    "    return list(label_map.keys())[list(label_map.values()).index(idx)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Each row in the dataframe named `df` represents the metadata for one audio file. For each audio file, the following details are included:\n",
    "- Name of the file.\n",
    "- The recording sound ID.\n",
    "- The start time of the slice in the original audio.\n",
    "- The end time of the slice in the original audio.\n",
    "- salience, an indicator for whether the sound slice is a foreground sound (1) or background sound (2).\n",
    "- ID of the class to which the sound slice belongs.\n",
    "- Name of the class to which the sound slice belongs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>slice_file_name</th>\n",
       "      <th>fsID</th>\n",
       "      <th>start</th>\n",
       "      <th>end</th>\n",
       "      <th>salience</th>\n",
       "      <th>fold</th>\n",
       "      <th>classID</th>\n",
       "      <th>class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100032-3-0-0.wav</td>\n",
       "      <td>100032</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.317551</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>dog_bark</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100263-2-0-117.wav</td>\n",
       "      <td>100263</td>\n",
       "      <td>58.5</td>\n",
       "      <td>62.500000</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>children_playing</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100263-2-0-121.wav</td>\n",
       "      <td>100263</td>\n",
       "      <td>60.5</td>\n",
       "      <td>64.500000</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>children_playing</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100263-2-0-126.wav</td>\n",
       "      <td>100263</td>\n",
       "      <td>63.0</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>children_playing</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100263-2-0-137.wav</td>\n",
       "      <td>100263</td>\n",
       "      <td>68.5</td>\n",
       "      <td>72.500000</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>children_playing</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      slice_file_name    fsID  start        end  salience  fold  classID  \\\n",
       "0    100032-3-0-0.wav  100032    0.0   0.317551         1     5        3   \n",
       "1  100263-2-0-117.wav  100263   58.5  62.500000         1     5        2   \n",
       "2  100263-2-0-121.wav  100263   60.5  64.500000         1     5        2   \n",
       "3  100263-2-0-126.wav  100263   63.0  67.000000         1     5        2   \n",
       "4  100263-2-0-137.wav  100263   68.5  72.500000         1     5        2   \n",
       "\n",
       "              class  \n",
       "0          dog_bark  \n",
       "1  children_playing  \n",
       "2  children_playing  \n",
       "3  children_playing  \n",
       "4  children_playing  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To get better insight into the range of audio lengths included in the data, a new column will be added resresenting the length of the sound calculated by the subtracting the the start timestamp from the end."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['length'] = df['end'] - df['start']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Furthermore, to get better insight into the range of different sample rates included in the data, a new column will be added resresenting the sample rate of the individual slice. <br/> <br/>\n",
    "To get the sample rate, librosa's `get_samplerate` function will be applied on each audio slice to extract the relevant rate. The sample rates will then be saved into a new column called `sample_rate`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "sample_rate = np.zeros(df.shape[0])\n",
    "for index, row in df.iterrows():\n",
    "    sample_rate[index] = librosa.get_samplerate('./audio/fold' + str(row['fold']) + '/' + row['slice_file_name'])\n",
    "    \n",
    "df['sample_rate'] = sample_rate  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>slice_file_name</th>\n",
       "      <th>fsID</th>\n",
       "      <th>start</th>\n",
       "      <th>end</th>\n",
       "      <th>salience</th>\n",
       "      <th>fold</th>\n",
       "      <th>classID</th>\n",
       "      <th>class</th>\n",
       "      <th>length</th>\n",
       "      <th>sample_rate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100032-3-0-0.wav</td>\n",
       "      <td>100032</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.317551</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>dog_bark</td>\n",
       "      <td>0.317551</td>\n",
       "      <td>44100.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100263-2-0-117.wav</td>\n",
       "      <td>100263</td>\n",
       "      <td>58.5</td>\n",
       "      <td>62.500000</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>children_playing</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>44100.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100263-2-0-121.wav</td>\n",
       "      <td>100263</td>\n",
       "      <td>60.5</td>\n",
       "      <td>64.500000</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>children_playing</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>44100.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100263-2-0-126.wav</td>\n",
       "      <td>100263</td>\n",
       "      <td>63.0</td>\n",
       "      <td>67.000000</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>children_playing</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>44100.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100263-2-0-137.wav</td>\n",
       "      <td>100263</td>\n",
       "      <td>68.5</td>\n",
       "      <td>72.500000</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>children_playing</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>44100.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      slice_file_name    fsID  start        end  salience  fold  classID  \\\n",
       "0    100032-3-0-0.wav  100032    0.0   0.317551         1     5        3   \n",
       "1  100263-2-0-117.wav  100263   58.5  62.500000         1     5        2   \n",
       "2  100263-2-0-121.wav  100263   60.5  64.500000         1     5        2   \n",
       "3  100263-2-0-126.wav  100263   63.0  67.000000         1     5        2   \n",
       "4  100263-2-0-137.wav  100263   68.5  72.500000         1     5        2   \n",
       "\n",
       "              class    length  sample_rate  \n",
       "0          dog_bark  0.317551      44100.0  \n",
       "1  children_playing  4.000000      44100.0  \n",
       "2  children_playing  4.000000      44100.0  \n",
       "3  children_playing  4.000000      44100.0  \n",
       "4  children_playing  4.000000      44100.0  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>fsID</th>\n",
       "      <th>start</th>\n",
       "      <th>end</th>\n",
       "      <th>salience</th>\n",
       "      <th>fold</th>\n",
       "      <th>classID</th>\n",
       "      <th>length</th>\n",
       "      <th>sample_rate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>8732.00</td>\n",
       "      <td>8732.00</td>\n",
       "      <td>8732.00</td>\n",
       "      <td>8732.00</td>\n",
       "      <td>8732.00</td>\n",
       "      <td>8732.00</td>\n",
       "      <td>8732.00</td>\n",
       "      <td>8732.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>116033.49</td>\n",
       "      <td>38.65</td>\n",
       "      <td>42.25</td>\n",
       "      <td>1.35</td>\n",
       "      <td>5.39</td>\n",
       "      <td>4.59</td>\n",
       "      <td>3.61</td>\n",
       "      <td>48456.98</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>57991.02</td>\n",
       "      <td>74.29</td>\n",
       "      <td>74.37</td>\n",
       "      <td>0.48</td>\n",
       "      <td>2.85</td>\n",
       "      <td>2.89</td>\n",
       "      <td>0.97</td>\n",
       "      <td>15300.08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>344.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.11</td>\n",
       "      <td>1.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.05</td>\n",
       "      <td>8000.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>69942.25</td>\n",
       "      <td>3.00</td>\n",
       "      <td>6.84</td>\n",
       "      <td>1.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>2.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>44100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>118279.00</td>\n",
       "      <td>10.38</td>\n",
       "      <td>14.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>5.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>44100.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>166942.00</td>\n",
       "      <td>35.13</td>\n",
       "      <td>38.87</td>\n",
       "      <td>2.00</td>\n",
       "      <td>8.00</td>\n",
       "      <td>7.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>48000.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>209992.00</td>\n",
       "      <td>600.13</td>\n",
       "      <td>604.13</td>\n",
       "      <td>2.00</td>\n",
       "      <td>10.00</td>\n",
       "      <td>9.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>192000.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            fsID    start      end  salience     fold  classID   length  \\\n",
       "count    8732.00  8732.00  8732.00   8732.00  8732.00  8732.00  8732.00   \n",
       "mean   116033.49    38.65    42.25      1.35     5.39     4.59     3.61   \n",
       "std     57991.02    74.29    74.37      0.48     2.85     2.89     0.97   \n",
       "min       344.00     0.00     0.11      1.00     1.00     0.00     0.05   \n",
       "25%     69942.25     3.00     6.84      1.00     3.00     2.00     4.00   \n",
       "50%    118279.00    10.38    14.00      1.00     5.00     4.00     4.00   \n",
       "75%    166942.00    35.13    38.87      2.00     8.00     7.00     4.00   \n",
       "max    209992.00   600.13   604.13      2.00    10.00     9.00     4.00   \n",
       "\n",
       "       sample_rate  \n",
       "count      8732.00  \n",
       "mean      48456.98  \n",
       "std       15300.08  \n",
       "min        8000.00  \n",
       "25%       44100.00  \n",
       "50%       44100.00  \n",
       "75%       48000.00  \n",
       "max      192000.00  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.describe().round(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### File format"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "All files are in the Waveform Audio File Format (.wav), eliminating the need for any format conversion. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "[x for x in df['slice_file_name'] if '.wav' not in x]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Lengths variation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As can be seen in the following table, different classes have different range of sound lengths. The shortest sound slice belongs to the `car_horn` with only 5 ms. All classes have a maximum lengths of 4 seconds."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The `gun_shot` class, the class of interest has a minimum lengths of 0.17 seconds with a mean of 1.65 seconds."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Min</th>\n",
       "      <th>Max</th>\n",
       "      <th>Mean</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>class</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>air_conditioner</th>\n",
       "      <td>2.04</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.99</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>car_horn</th>\n",
       "      <td>0.05</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>children_playing</th>\n",
       "      <td>1.05</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.96</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>dog_bark</th>\n",
       "      <td>0.12</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>drilling</th>\n",
       "      <td>0.42</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.55</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>engine_idling</th>\n",
       "      <td>0.77</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.94</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>gun_shot</th>\n",
       "      <td>0.17</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1.65</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>jackhammer</th>\n",
       "      <td>0.39</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.61</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>siren</th>\n",
       "      <td>0.26</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.91</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>street_music</th>\n",
       "      <td>4.00</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   Min  Max  Mean\n",
       "class                            \n",
       "air_conditioner   2.04  4.0  3.99\n",
       "car_horn          0.05  4.0  2.46\n",
       "children_playing  1.05  4.0  3.96\n",
       "dog_bark          0.12  4.0  3.15\n",
       "drilling          0.42  4.0  3.55\n",
       "engine_idling     0.77  4.0  3.94\n",
       "gun_shot          0.17  4.0  1.65\n",
       "jackhammer        0.39  4.0  3.61\n",
       "siren             0.26  4.0  3.91\n",
       "street_music      4.00  4.0  4.00"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "length_min = df.groupby('class').min()['length'].round(2).rename('Min')\n",
    "length_max = df.groupby('class').max()['length'].rename('Max')\n",
    "length_mean = df.groupby('class').mean()['length'].round(2).rename('Mean')\n",
    "\n",
    "lengths = length_min.to_frame().join(length_max).join(length_mean)\n",
    "lengths"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'mean across all classes: 3.61'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "'mean across all classes: {}'.format(round(np.mean(df['length']), 2))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Sample Rate variation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As can be seen the following chart, audio clips differ in sample rate. The sample rates range from 32 kHz to 44.1 kHz. The majority of clips have a sample rate of more than or equal to 96 kHz."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df['sample_rate'].value_counts().plot.bar(x='sample_rate', y='count', rot=50);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Foreground vs. background"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Each class contains both foreground and background noises. Some classes (e.g. engine_idling and gun_shot) contain far more foreground noises than background. This adds variation to the classes to help introduce the machine learning classifer to different versions of the audio sounds imitating real life noises."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "class             salience\n",
       "air_conditioner   1           569\n",
       "                  2           431\n",
       "car_horn          1           153\n",
       "                  2           276\n",
       "children_playing  1           588\n",
       "                  2           412\n",
       "dog_bark          1           645\n",
       "                  2           355\n",
       "drilling          1           902\n",
       "                  2            98\n",
       "engine_idling     1           916\n",
       "                  2            84\n",
       "gun_shot          1           304\n",
       "                  2            70\n",
       "jackhammer        1           731\n",
       "                  2           269\n",
       "siren             1           269\n",
       "                  2           660\n",
       "street_music      1           625\n",
       "                  2           375\n",
       "dtype: int64"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.groupby(['class', 'salience']).size()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Class distribution"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Based on the following chart, there is a huge class imbalance in the data. Eight out of the ten classes have an equal number of audio samples, with 1000 instances per class. However, the car horn, and gunshot classes are significantly underrepresented in the dataset. This poses a great challenge for the current task, particularly as the gunshot class is the class of interest. With only 4.3% of all samples, the gunshot class has the least number of instances. Resampling the dataset is necessary in this case."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def get_class_dist(df, label, plot=False):\n",
    "    count_per_class = {}\n",
    "    for index, row in df.iterrows():\n",
    "        y = row[label]\n",
    "        if(y not in count_per_class):\n",
    "            count_per_class[y] = 0\n",
    "        count_per_class[y] += 1\n",
    "    if plot:\n",
    "        colors = ['#8858Ca', '#D65DB6', '#FF6F91', '#FF9671', '#FFC75F', '#F9F871', '#0081CF', '#FBEAFF', '#00C9A7', '#b3bAf6' ]\n",
    "        plt.pie(list(count_per_class.values()), labels = list(count_per_class.keys()), colors=colors, autopct='%.0f%%')\n",
    "        plt.show()\n",
    "    \n",
    "    return count_per_class\n",
    "\n",
    "get_class_dist(df, 'class', True);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## II. Data pre-processing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this section, the neccessary data pre-processing techniques are applied to prepare the data for training, validation and testing."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To make sure all sounds have the same sample rate, the sample rate for all loaded audio files will be set at 22.05 kHz. <br/>\n",
    "Despite most of the data having an original sample rate of 44.1 kHz, this value was chosen as it is considered a middle value between the smallest and the highest sample rates in this case. Oversampling very low rates to a very high rate value such as 44.1 kHz might distort the sound and lead to the loss of valuable features. On the other hand, undersampling high rates to a slightly smaller value will not hugely affect the quality. <br/> <br/>\n",
    "The Python library librosa will be used to load and process the audio files. One of the notable features of this library is its ability to automatically resample audio files during loading, which is highly advantageous for facilitating sample rate conversion. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "22050"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "SAMPLE_RATE"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The audio files are split into 10 folders. The following cell sets the path to each of those folders."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "AUDIO_DIR = './audio' # path to audio files directory\n",
    "folders = ['fold{}'.format(x) for x in range(1, 11)] # array of paths to each audio folder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "slices = np.zeros(df.shape[0], dtype=object)\n",
    "slices[:] = np.nan # placeholder array for audio slices\n",
    "classes = np.array(df['class']) # list of audio classes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Loading the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "collecting fold1...done!\n",
      "collecting fold2...done!\n",
      "collecting fold3...done!\n",
      "collecting fold4...done!\n",
      "collecting fold5...done!\n",
      "collecting fold6...done!\n",
      "collecting fold7...done!\n",
      "collecting fold8...done!\n",
      "collecting fold9...done!\n",
      "collecting fold10...done!\n"
     ]
    }
   ],
   "source": [
    "for fold in folders:\n",
    "    print('collecting {}...'.format(fold), end = \"\")\n",
    "    files = librosa.util.find_files('{}/{}'.format(AUDIO_DIR, fold), ext=['wav'])\n",
    "    files = np.asarray(files)\n",
    "    for file in files:\n",
    "        if '.wav' in file:\n",
    "            name = file.split('/').pop()\n",
    "            wave_arr, sr = librosa.load(file, sr = SAMPLE_RATE, mono = True)\n",
    "            idx = list(df.index[df['slice_file_name'] == name])[0]\n",
    "            slices[idx] = wave_arr.astype(object)\n",
    "    print(\"done!\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Slices array shape: (8732,)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA+oAAAFzCAYAAACpaPNwAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAACb/ElEQVR4nO3dd3wUdfoH8M+mF1JJh0DoRTpIpKgokXoqd96dKIpyCKeCDc/C/RTsqMd5Nk4UReSEQz3FbgQpIhCpRum9QxIgpJO+vz+SWXY3M7MzuzM7s5vP+/XKS7I7O/PdJM7OM9/n+zwWq9VqBRERERERERGZQoDRAyAiIiIiIiKiSxioExEREREREZkIA3UiIiIiIiIiE2GgTkRERERERGQiDNSJiIiIiIiITISBOhEREREREZGJMFAnIiIiIiIiMhEG6kREREREREQmEmT0AIxQX1+P06dPIyoqChaLxejhEJEPslqtKC0tRVpaGgIC/OeeJ8+PROQpfz0/EhF5U7MM1E+fPo309HSjh0FEfuDEiRNo3bq10cPQDM+PRKQVfzs/EhF5U7MM1KOiogA0fIBER0cbPBoi8kUlJSVIT0+3nU/8Bc+PROQpfz0/EhF5U7MM1IV0zujoaF6IEpFH/C09nOdHItKKv50fiYi8iQuHiIiIiIiIiEyEgToRERERERGRiTBQJyIiIiIiIjIRBupEREREREREJsJAnYiIiIiIiMhEdA3U161bh+uvvx5paWmwWCz4/PPPXb5m7dq16NevH0JDQ9GxY0csWrSoyTbz5s1DRkYGwsLCkJmZic2bN2s/eCIiIiIiIiID6Bqol5eXo3fv3pg3b56i7Y8cOYKxY8fimmuuQW5uLh588EHcdddd+P77723bfPTRR5gxYwZmz56N7du3o3fv3hg5ciQKCgr0ehtEREREREREXmOxWq1WrxzIYsHy5csxbtw4yW0ee+wxfPPNN9i5c6ftsfHjx6OoqAjZ2dkAgMzMTFx++eV48803AQD19fVIT0/Hfffdh8cff1zRWEpKShATE4Pi4mL2CSYit/jrecRf3xcReQ/PI0REnjPVGvWcnBxkZWU5PDZy5Ejk5OQAAKqrq7Ft2zaHbQICApCVlWXbRkxVVRVKSkocvoiIiIiIiIjMyFSBel5eHpKTkx0eS05ORklJCS5evIhz586hrq5OdJu8vDzJ/c6ZMwcxMTG2r/T0dF3GT0REREREROQpUwXqepk5cyaKi4ttXydOnDB6SIr9drIIO04WGz0MIiIiv1BeVYvPtp+El1b+ERGRCdTVW/HJ1hOorq03eiiKBRk9AHspKSnIz893eCw/Px/R0dEIDw9HYGAgAgMDRbdJSUmR3G9oaChCQ0N1GbPexs3bAIvFgkMvjDF6KERERD7vg5yjeDl7Hy5Li0GXlCijh0NERF5w3Ss/4vC5cgQHBmBc31ZGD0cRU82oDxo0CKtWrXJ4bOXKlRg0aBAAICQkBP3793fYpr6+HqtWrbJt42/qrQ13gIiIiMhzecWVAICq2jqDR0JERN5y+Fw5AOBCRbXBI1FO10C9rKwMubm5yM3NBdDQfi03NxfHjx8H0JCSPnHiRNv2d999Nw4fPoxHH30Ue/fuxb///W98/PHHeOihh2zbzJgxAwsWLMAHH3yAPXv24J577kF5eTkmTZqk51shIiIiUuTQ2TL8dOCs0cMgIiIfpmvq+9atW3HNNdfYvp8xYwYA4I477sCiRYtw5swZW9AOAO3atcM333yDhx56CK+99hpat26Nd999FyNHjrRtc/PNN+Ps2bOYNWsW8vLy0KdPH2RnZzcpMEdERERkhMmLtuDo+QocfXGs0UMhIiIfpWugPmzYMNliLYsWLRJ9zS+//CK73+nTp2P69OmeDo+IiIiagY0HzyEhKhSdk72zJv3o+QqvHIeIiPyXqdaoExEREWnt1nc3YcK7m4weBhERkWIM1ImIiMjvnS2tMnoIREREijFQJyIiIp9TU1ePrUcLjR4GERGRLhioExERkc/537aT+OP8HBzILzV6KEREkhasO4yjja3BiNRgoE5EZHLz5s1DRkYGwsLCkJmZic2bNyt63bJly2CxWDBu3Dh9B0hkgMNnywAAxRdrDB4JEZG057/dg1lf7jJ6GOSDGKgTEZnYRx99hBkzZmD27NnYvn07evfujZEjR6KgoED2dUePHsXf/vY3XHnllV4aKZG51NbVGz0EIiIAwOmii0YPgXwQA3UiIhN75ZVXMGXKFEyaNAndu3fH/PnzERERgYULF0q+pq6uDhMmTMDTTz+N9u3be3G0ROawbv9ZdHriOxSUVho9FCIiMpEaH7qJy0CdiMikqqursW3bNmRlZdkeCwgIQFZWFnJyciRf98wzzyApKQmTJ092eYyqqiqUlJQ4fBH5ujX7CmC1AqcucBaLiIguCQ8ONHoIijFQJyIyqXPnzqGurg7JyckOjycnJyMvL0/0NevXr8d7772HBQsWKDrGnDlzEBMTY/tKT0/3eNxERnjmq11YnHPU6GEQEZGJHfahwn4M1ImI/ERpaSluv/12LFiwAAkJCYpeM3PmTBQXF9u+Tpw4ofMoqTk6X1aFDzYeFX2uoroW3+444/ExFm44illfsGATERFJ86UK/EFGD4CIiMQlJCQgMDAQ+fn5Do/n5+cjJSWlyfaHDh3C0aNHcf3119seq69vWIsVFBSEffv2oUOHDg6vCQ0NRWhoqA6jJ7rktVUHsDjnGMb2SkVCC8e/t4Xrj2Duiv1Y+7dhyEiINGiERET6yS9hvQxSjzPqREQmFRISgv79+2PVqlW2x+rr67Fq1SoMGjSoyfZdu3bFjh07kJuba/u64YYbcM011yA3N5dp7WSYvOKGi9R6q7XJcycKG9aRV9X6ToEfIiI14iNDjB4C+SDOqBMRmdiMGTNwxx13YMCAARg4cCBeffVVlJeXY9KkSQCAiRMnolWrVpgzZw7CwsLQo0cPh9fHxsYCQJPHiYiIiMi8GKgTEZnYzTffjLNnz2LWrFnIy8tDnz59kJ2dbSswd/z4cQQEMDmK/Mf7G46ga0o0BnVoafRQiIiIDMNAnYjI5KZPn47p06eLPrd27VrZ1y5atEj7ARHp6OmvdiM0KAD7nhvt9WPnFVfiqS934ZWbeyMiJAh7zpQgMiQIbVpGeH0sRETUvHEahoiIiEzFqPXqH289gexdedh69AIAYPRrP2HCez8bMhYiImreGKibVNYrP2L13nzXGxIREZEm6uqli90RERF5EwN1kzpYUIY3Vh80ehhERESkwPbjF1BWVWv0MIiIyE8wUDcxkRv7REREZEJ/+PdGzPl2j9HDICIiP8FAnYiIiEilp7/ahfuWbnd4bPORQoNGQ0RESlTWGFMDxR0M1ImIiIhUen/DUXz12xmjh0FERCrkHD5v9BAUY6BuQlZrQ857cUW1wSMhIiIiIiIib2OgbmJxkSFGD4GIiIiIiIi8jIE6ERERERERkYl4JVCfN28eMjIyEBYWhszMTGzevFly22HDhsFisTT5Gjt2rG2bO++8s8nzo0aN8sZbISIiIiIiItJVkN4H+OijjzBjxgzMnz8fmZmZePXVVzFy5Ejs27cPSUlJTbb/7LPPUF19aW32+fPn0bt3b/zpT39y2G7UqFF4//33bd+Hhobq9yaIiIiIiIhUEOpOXShn3SlST/cZ9VdeeQVTpkzBpEmT0L17d8yfPx8RERFYuHCh6Pbx8fFISUmxfa1cuRIRERFNAvXQ0FCH7eLi4vR+K0RERKSBj7eewLZjrluZWa1WXuBSs1ZRXWsL9sh3JUWHGT0E8kG6BurV1dXYtm0bsrKyLh0wIABZWVnIyclRtI/33nsP48ePR2RkpMPja9euRVJSErp06YJ77rkH589Ll9qvqqpCSUmJwxcREREZ49H//YY7F25xud1HW06g77MrUVJZ44VREZlPj9nfY/6Ph40eBhEZQNdA/dy5c6irq0NycrLD48nJycjLy3P5+s2bN2Pnzp246667HB4fNWoUFi9ejFWrVuGll17Cjz/+iNGjR6Ourk50P3PmzEFMTIztKz093f03RURERB4rrap1uc26A2cBAOUKtiXyR/VWYOnmY0YPg4gMoPsadU+899576NmzJwYOHOjw+Pjx423/7tmzJ3r16oUOHTpg7dq1GD58eJP9zJw5EzNmzLB9X1JSwmCdiIiIiIiITEnXGfWEhAQEBgYiPz/f4fH8/HykpKTIvra8vBzLli3D5MmTXR6nffv2SEhIwMGDB0WfDw0NRXR0tMOXL/jleJHRQyAiIiIPbTx4DgcLSgEAh86W4cY316O4gun8RERVtXXYfZrLksXoGqiHhISgf//+WLVqle2x+vp6rFq1CoMGDZJ97SeffIKqqircdtttLo9z8uRJnD9/HqmpqR6PmYiIiEhLt767CRPfa2hNu3z7Kfx6shg7TxeLbnum+CLq6lk8jKi5qKmrR2kzrsPx1tpDGPP6T6isEV/C3JzpXvV9xowZWLBgAT744APs2bMH99xzD8rLyzFp0iQAwMSJEzFz5swmr3vvvfcwbtw4tGzZ0uHxsrIyPPLII/j5559x9OhRrFq1CjfeeCM6duyIkSNH6v12iIjIBArLq/HptpNGD4O87D85R7F2XwEAoKq2HgBQ3fhfd5VX1Xq8DyVOF1e63Kamrh6D5qzGgp9YPIyouXjkk18x+rWfjB6GYbYduwCg4fxHjnRfo37zzTfj7NmzmDVrFvLy8tCnTx9kZ2fbCswdP34cAQGO9wv27duH9evXY8WKFU32FxgYiN9++w0ffPABioqKkJaWhhEjRuDZZ59lL3UiomZi7op9WLrpOEb3TEFESBA+3XYSvdNj0DEpyuihkY6e/GIXAODoi2MRHhzY8KDFs33e+OYGtE+MxDsTB3g4Os8JM+k/7juLu6/uYPBoiMgbPs89bfQQyKS8Ukxu+vTpmD59uuhza9eubfJYly5dJHtGhoeH4/vvv9dyeERE5CXnyqrwwLJf8K8/9/Gor+yJwgoAgPBR8fAnv6JbajS+e+BKLYZJGpi35iBG90hB+8QWuuw/MMDDCL3RwbNlOHi2zOV2VqsVJwovok3LCE2OS0REJEf31HdS72CB6wsGIiJf9OO+s9hw8DxW7M53vbFKe86wGI2Z/OP7fZjVOAPuD1btKcBV/1iDXRJry4mIPFHP2hTkhIG6CU1but3oIRAREXksv8T1umwrfOPidF9+Q9X200Wu3xMRkVq9n2m65Le52nK0ENe98mOzLzBn6j7qzdX+fM6oE5F/utjMP3SpqR/2NBSHO3KuHF1StKkxUFnTUJSoXmIZHRGR2ZRW1ho9BMOt2JWHnw6cQ2F5NQ4UlOF8eTVaxYYbPSzDcEadiIi8RlhXrGcAdb6sCuv2n9Vt/6Se3G+7sLwaQENbMq18tr2hI0DOofOa7VMvc77dg4c/zjV6GGRCQoeDMgZw1Ezc999f8J+fjxk9DNNgoE5ERF4jlP8KDdLv42f2l7swceFm3fZP4hZtOII/vrVR9LmVjTUJ8ourVO1TuJ+jtq+40LrNFzI43l53GJ9uP2X0MMiE/rv5OADgQkXz7bFNzUuVF1plAkBpZQ1e++GA6VvCMVAnIiK/8svxIqOH4LesViuqasWD36e+2o2tjf1wpZwvVxeoC4F2vbmvpYh0ERUWbPQQSGdSXa5IX59sPYl//bBfVdbVubIqXPfKj9iXV6rjyBwxUCciIs3x4sM/zV2xD1e9vEbx9gc87GISE+5/gcrmI4UAgIvV5pjtLyitxJajhUYPg0Q4dyDkedX/LP+F2TRaqK2rVzU7Xl7VsJxEzWt2ny7BgYIyfPOb9/reM1AnIiJNrd6bj15Pr0BFNddV+pt5aw4hv0TdrDg52twYFKvNLtDLQx/l4k/zc4weBrnw9o+HkPnCKqOHQRpzvknGmzHuuXfJdtyh0ZK3+nor/r32INY01ogwEqu+ExGRpj7ZehKllbUovliDiBB+zBCZ2YaD5i+4R8Cc7/YaPQTygrp6K4ICLa439HPFF9XVZVjRWAdFyr/XHlS8r0f+9xs+bSxIevTFsarGoTXOqBMRkWkUX6xBvcrCYUTNxbmyKluVfCLyT821d7h9cL7+4DnN9ltXb8XL2fsUby8E6WbAQJ2IiEyj99Mr8OqqA0YPg+x8sPEoTl6oMHoYplFnYGW7P8/PwYR3Nxl2fCLSV2llLbo+mY0vf/XeOmizGPpS0/onxc284wEDdSIiMpWPt5wweghkZ/aXu/D0V7uNHobhhOrA2TvzNNmfO0tRD58rx54zJZocn4jMR5hVXrFLm/OMr/vpwFmjh2AoBupERER+4tcTRdh5qliX/TZ3Qsp5XkmlJvsze/9eMl5pJQtyNjeVEu0v/dmZYulzanMvXspA3UfwIomIiFy5cd4G/OHfGz3aR3FFDbYdc6xE3CKURQE1Z2HBKJIXFOjfl+lPf7UL6w9otxbZHzTHn0dii1DJ5xZuOOLy9d6ulP/e+iN4+stdXjmWf58B/Mi/fthv9BCIiEyjvvGDuZ6tbJqo9nCm9skvduKmtxzbdUnFlP/dfNyw9esnL1wEAOTJzMYQ+bIg50bqfub9DUfx4Ee/GD0MQ+SXVOKhj3JRVcPMGk+r3K/dp116vJLPs/LqOry/8ahmx5TDQN1H1NbxYpSISCC0lFq91/g+p/5m4yHl7bpmfrYDs77wzsyCM6HHrZrxEvmS5tBT+1xZ8+hicLHaMaX9w5+PYfkvp7D12AWDRiStvt6Kxz/9DYfOlhk9FEWOF8oH12pu6H/+yynJ54yoxs9A3UdcbKatGoiI5JhxNnXbscJmVRDv6Plyo4fgt04XXcTC9a5TP8k/7dCh3oTZhAXrF4oUlFaioNQcnxGnii46fF/TOAFX59SOtLbx+/Iq4+oTFF2swbItJ/DKSvNl8x46W6b6BoLzz1iOXExvRH97BuomUyuRshjo5+lPRERGeCl7r21mVit3LtyCRz/9TdN96u3Dn4/hh935Rg/DZwkXggcKtJ2Bev7bPXjm692S1wbk3+z7SpN649/5GePf+dnQMRw62/RGZmVNHT78+RgAICIk0OG5F7/bCwCIjQjRf3AufPPbGQx47gfUqwh09XbTvzd6XIfFXT8fLnS9kcYYqJvM57nifRNDg/irIiLS2ltrD+GeD7dpus9SnWdCLlbXYV9eqab7fOLznbhr8VbJ52tNdKFmRnpVcD/jNAtHRModPluOwyKBsjeJrcFfnHMUZS4+J4ycnrOfgT5XVoVTRReR+cIP2HjQ+EJ3RRdrNLmBtTdf/Wfoog1HPT6uWoz+TMaoojxERM1VpRvFfI6dL8cxg1K+53y3ByNfXefVYx47z88mIiK19p5pGhBerDZ3hoxzbYSiihrkl1Th7XWHDRqR9r757Yzo47+ZbLkJA3WTkZq0+OnAOUOKGBARGcms570/zs/Bn+bnuN5QB1pWuCVtMOOAyH/ll1Q2i8J+UoTlt5GhgS62NIfiCvdn3A87rX8vLDe22CEDdZORS59zLkRBRGQW/1q5H9/taLhD/d3OPIfndp0uxvbj4pVtq2rrUFAiXuzni9xT6PpktmyKYHm1MYH82dIqFJRWAQCKKqrx1Je7UFVrzpsKejvb+HPwNmFJWKDKfuSllQ1/TzUadlNhm0Ai/3T0XDkyX1iFbKfPNb2Z6YwS0BgtWgxNyHdkfxPfPlV/zd4C9H5mBY6ccy/jzbmmwPwfD7k3QI0wUDeZ8GDpu1VmKuZARGTvtVUHcM+S7aLP/Xl+jmTxlyc/34lBL64WfU64MCqtlL47vv6A8bPLH205gUUbjxpSaMYMhMDX237ftxUA4MrOCapet7yx/c7Ph5tPW7eq2romM0Vkfkov+2rq6rHTZCm7nrJarabIqMpvvJGce7LIrdfbZ9us3J2PEy5aiQnOlRlzA9QIPx1Qv/ZdKLoHOM56C5MCpy5oM7l53uD2gQzUfYjKSQMiIkMJ6XJys94fbz3ZpHVKnYrZSa1vX5534+KoqrYhE0rsZuqRc+UY89pPurYIuv29TVirceV6XxATHgwACAtyLx2zVsMZdbP718oDuPafPxo9DFJJaSHh/24+jt+9sd6vMi+f+Hwnuj6ZbfQwPGbffm7K4q24f1nT4nJihPMbiVux61KGw5trDho4En0xUCciIl0EuHl38aJB6ew/HTiL/s/9gD1nSjTb5/e78rD7TAm2H2ua+l9YXo2lm457fIyfDpzDs1/v9ng/5FsKSitxwxvrFc3QNccbOc3JtsbzS5lB2S16WKLBudGMfjle5PC9cKOXSIxXAvV58+YhIyMDYWFhyMzMxObNmyW3XbRoESwWi8NXWFiYwzZWqxWzZs1CamoqwsPDkZWVhQMHDuj9Ngx3ttTY9AsiIm+IixSfSdh1ulhx2qA7dp1uCNC9Vc39Xyv34+/Ld6CowvNzu1ivXmfHzpd7nB4bwlahprH5SCF+O1WM73d5d+0smY+QlRTAzEvTc86SOK7jZ5pW5JLczLA8wZ/p/on70UcfYcaMGZg9eza2b9+O3r17Y+TIkSgokL67Gx0djTNnzti+jh075vD8yy+/jNdffx3z58/Hpk2bEBkZiZEjR6KyUr/UQjM4wdZtRNSMjX19PW5/b5PDY3rU7tCyyJic041pqt4qP3Lrgk343RvrvXMwP3Sx8YLUeamGGdXXW/Gvlfub1TrX5ky4gRboQaReVVuHche9val5emO1+GTosfPl6PpkNta7scZcoKaa/sZD2vdxX7TxKD78+ZjrDWUUV9To1l5b90D9lVdewZQpUzBp0iR0794d8+fPR0REBBYuXCj5GovFgpSUFNtXcnKy7Tmr1YpXX30VTzzxBG688Ub06tULixcvxunTp/H555/r/XaIiMhAR536eYcqWJ/86baTimbJaxu7bgiBWH291SeCMqX8af2qXoRKwWKzXEGNQVBwoPmzCg6dLcNrqw7gHT/qe0wNHv3fr7rs994Pt+P38zbosm9f5k+fAe76dod41s6xxs/jzUfcL8yp5hx164JNrjdS6acD5/DE5ztRfNH9lm73LNmGkf9ap+GoLtH106a6uhrbtm1DVlbWpQMGBCArKws5OdL9Z8vKytC2bVukp6fjxhtvxK5du2zPHTlyBHl5eQ77jImJQWZmpuQ+q6qqUFJS4vDlk3iuICJypGAC6eFPfsXfPmm4uM0vqcQXuadEtwsPCQIABDQGZA98lIvJi7ZoM04NWflhoJsNBxtmbLYela7g78mspVrbG9eznilWlzEoxBae9BMmc/p460nJ50o8WKO+am8B9heUYeXufLf34Y+Equ1VNVxLrgcta8IAQEFJlVv7ffG7PS63qRZpob0/vxQbD53XrVWsroH6uXPnUFdX5zAjDgDJycnIyxO/O9OlSxcsXLgQX3zxBT788EPU19dj8ODBOHmy4cQkvE7NPufMmYOYmBjbV3p6uqdvTTdy7VN4cUZEzYmQaqy2RbXYzPHevFIAwKwvduKBZbmK9vPVr6exdr8x7d8OFpTipey9os+dKDRuZry6sfCRmnRFpThzJc3TbAjh71+t+norMzF8iLtFJe3XGc/+YqdWw9GEXHtOrZRU1uCGN9dj92np4G7RxqO6j6M5uqhwjXuewpuVu8401GFRe8Ppt5Ou67eI1ccdodNMusB0+VuDBg3CxIkT0adPH1x99dX47LPPkJiYiLffftvtfc6cORPFxcW2rxMnTmg4Ym19nnta8jl3KygTEfkCIVDLb7wjvnZfQ5C8w80CaGLB5P583+gl/fRXu/HW2kOGHFvJumYd4nSs4EyeTY5Tj/dqlZWhhe3XH/RsTeeHm45hyIurPUoLJe/ZJtJdQgn7m4Jis4Z6qnQxU+3uTSZ7H2+Rv+7/7UQxfjtZjK9/a3oNzgKa0oQbPJUeVK6PDlPWhm5/vud/B3J2nS7Bl79Kx2BG0fWvLyEhAYGBgcjPd/zwzc/PR0pKiqJ9BAcHo2/fvjh4sKFHnvA6NfsMDQ1FdHS0wxcREZmLEKgnRYU6PC5XVVZtaq8eM8HuWrW3oaiqWO/2wnLHSvBjX/8JSzapK3hTWF6NaUu3o8TNGSlPCku583PWovq9v1ISKF8or8aZ4obZ77LG352ns+GbDjcsAaioZpExbyssr1a95MFdPxqUOQS4XkqixSn70U9/k32+pvHmhNjsrqdTZK5uRPiyoMCGn05UaJDBI9HG/f9t2uP+Ys2lc58Rlw+6BuohISHo378/Vq1aZXusvr4eq1atwqBBgxTto66uDjt27EBqaioAoF27dkhJSXHYZ0lJCTZt2qR4n0REvkRNi8sFCxbgyiuvRFxcHOLi4pCVlSW7vRkFqVgDXHRRXXDnXIxOC0LLs01HpNc1C2rr6m1F6wTnyly/h12nS/Dc167X0Nn76tfT+Oa3M1iz1/s9tLW6oPm5cWY5r8S/u7po4a7FW3E9q/r7jYXrj0g+t0NBmq7Z7VM4Uy4Eg3oSbhZEKZzddWffpA+9l+ZEhhh7E0L3fI4ZM2ZgwYIF+OCDD7Bnzx7cc889KC8vx6RJkwAAEydOxMyZM23bP/PMM1ixYgUOHz6M7du347bbbsOxY8dw1113AWioCP/ggw/iueeew5dffokdO3Zg4sSJSEtLw7hx4/R+O0REXqW2xeXatWtxyy23YM2aNcjJyUF6ejpGjBiBU6fEC6h526HGOhzVIi3QhHZlegTTehLe094818Vr/vqfbZj8wVbR57789TR2nZa+AK+TiX5jI5peYNabKHtAidjwkCaPCcsgjnqpt70v23bsgqKbPuQb5Fryvvy9eP2KYA2CWm/9DU3+oKFQp6vaFFot+3x91QH8dMC4zAHyTI3TDe7TxZU4V1aFCQt+Ft1eq7peRt9o0f02wc0334yzZ89i1qxZyMvLQ58+fZCdnW0rBnf8+HEEBFy6X3DhwgVMmTIFeXl5iIuLQ//+/bFx40Z0797dts2jjz6K8vJyTJ06FUVFRRg6dCiys7MRFham99shIvIq+xaXADB//nx88803WLhwIR5//PEm2y9ZssTh+3fffReffvopVq1ahYkTJ3plzHIW/NQwS3Ra5C640qIyaggp2KVO1ZDLq2oR6YV0vRqntXurZGa3hbS7oy+O1XVMZhMWHIDKmnpkJEQ0eS6zXTw2HSlEx6QWqvbZJj4CxwsrkNgitMlzY3qm4NsdeejRisvgyLwiZGby9KxAHhXmnRnEkxe8W6TwlZX7ER4ciD3PjvLqccX42k1UwTc7zuCB/FJ0To7y+rHFivn9/bMdkjf2txx1r16D2XilQsL06dNx7NgxVFVVYdOmTcjMzLQ9t3btWixatMj2/b/+9S/btnl5efjmm2/Qt29fh/1ZLBY888wzyMvLQ2VlJX744Qd07tzZG2+FiMhr3G1xaa+iogI1NTWIj48Xfd5v2ldKkAr+a71UYbym8Thy12VGXLQVVVTjlnd+xkmZWTtvCQqQvhRpn9gQoEeEBKra59heDcvlOolcULaJjwQAr9yokWPWjFi1hetIH5tkelPrWdzP+aamP9HjZrA7fDNMb3Dnws2oN6BDx54zTZdKuFs80ZewlCERkUm50+LS2WOPPYa0tDSHYN+eL7WvdIc3YmC5Q4Q0pqLKVQ7+5bg+FxtljRfctSLLDLYcvYCcw+dNWeVWC2YNgu2pqcWgpfp6K/699qBosFdXb0XnJ75zWSWb9Hf4rPRyj5Ytmi4T8SdmuIGoFbHU/rBg3w2/ThdX4pcTRV4/bmG5624kgDEF3/Tku38pzZC//fERkb5efPFFLFu2DMuXL5dcGuRL7Su1pGXPbuGCWqzPa0ljsCx3/tZrBlOo/O3tFFMtHSpoWP8vtlTCXUIXgTI/njmU89upYrycvQ//yTna5Dkhu+N/2056eVSkhplnva1WK7J3nkFVrfuz1+VV3p359mYNjKraerczIqpq6/DptpOGzGjb83bWzZAXV0PpW9512r+yAhmo+xC2USdqXjxpcTl37ly8+OKLWLFiBXr16iW5XXNtX6lHurlY//Gw4IaUbbWVi4WLDbnWdK781lgZOnuXsuwLrWh5Ebn5aEMl/V+OF2m2z1V7G/5/UptNIBQV0qq41W6DLiiFVmuF5eyN7qt2nDJv1fdfThTh7g+347Pt7hcwPVXk3Rn1WV/s8urxfjlW5NbrPtl6Eg9/8ivWHzyHPWf0O38EBXovPFSyDE2qsvv5cmWFD5VuZ0YM1H3IBfaYJWpW3G1x+fLLL+PZZ59FdnY2BgwY4I2hkgvuBndKz/tqi0udLW24qVByUfuZuVIP+q97Q13jUgDnKsKu3NAnDQAwvFuSJuPI1bHFVmJUQ2p0eny45DZaVUUmsidkqsj1gD/YmCkj5cnPLwXORi0RMdr6g+eaPPbE5zsBABMXbsbo136ytQbVmqufeG29djPq3+08o9m+pKzek+96I5NioE5EZGJqW1y+9NJLePLJJ7Fw4UJkZGQgLy8PeXl5KCuTvzDyR3IXimKc+5s782R2Wwti6YZCCr9YkSQh9dQqkj3wRW7DbJe32xUFe3GmBoBH6bfOYsIbWuDJVePWi1gGSGXjzZkj55qm7YYGNWRytEuQrpYfHqyuQB+Zn9UKzP5iJ/bnK+tRLjhbomz9r1ayXvlR9nn7GVRvZpMeO1+Oe5dsw+Gzlz4v56095L0B2CmqqMG6/fLn5/wSdZ9xWpGbBLdarQ6fOV/9ehqlldLZO5U6djAQvL3usO7H0AsDdR9i8JIUIjLAzTffjLlz52LWrFno06cPcnNzm7S4PHPm0h3pt956C9XV1fjjH/+I1NRU29fcuXONeguGUbuO8+OtDevzD4pc5G4/fgFdn8xW1Cvdm8SCcIEwiy8W5AmBvau1hlqu5Qdcz9RoRSigZxG5yq+ukw5wvaljovKWc2K/B6Hi8ZJNx5s8J/zKXV3o29OyDgAZo95qxQc5x/D8N3tUvc7MWTCLNhzFX97f7JVjfZF7Gt/uyMO1//zR9v+ckv+HThRW4J11ygN6pb25Jy5seN/VtfWGn6/siWU51NTV44Vv9+DO97fgvv/+grp6K04XXcR9//0Fb6w+aMAoL1F7014prT8fxRjbm4RUaabZP0TN3vTp0zF9+nTR59auXevw/dGjR/UfkEIFJVVIEOljbVbChVCJSID/9o8NF2E7Thaja4r31vErvaCTIzZzKlxoRbhoUSZ3I8AdZ0XW8QtF76o0LFAULtPO7VxZw3KCvXnqZh21EmBpuPHueb9qbX83eSWVSIxy/P/VLO2sSBkhbpCbwfQFQQEW29rlTwwqbNhw7lN2/n3+mz3I3pWHu4a2R4DdOVv43HAWqrLq+5urD+B1g4Nde8JNYKFQ6T9X7ke/tnF4x27m+rudeVgwsT+AhhsZYjYfKdR5pA1aRoagoFT7rJEb563XfJ/OOKNORES6yC81Ji1PcOScunR/4SK3UKTwzPe7Gta4lcnMPGkZaArE1rZ7HuBdyjao82CtoTtB/OKcY00eC21sXefttHi1hCJ5vh4EqSGs49eqgB7p40C+47mupLIW05Zsx5urD9geO36+AluP6h8Y/euH/QCAcg9m6ZUUGHPXrC92etz+bfg/1+Kz7ZduIORJpKDP+W6vR8cRrDvQdL26GSxcf8T2b+dzRF291db7XGqZzZ/fztFvcHaiG5ctaW3nKf0z7Mz9qUhERCRDrsWWu+luFdXuzSK6G8rER6rriaxFyCTMOIcFub9OuUakP7srQrq2PSFAN0PWmJDBIJYyH9p4semNdEc9CcsdlLwLoWsBA3XvkqpyrXT7oAALvtlxBnNX7Lc9NvmDLfjjfP0DI6FDwwEXBePsrdKh2Ne7P4mvS16ccwzz1ni27vzQ2XL80+5nq7ciFcWk5/94CI/+71cdR3OJO58BRlBbONRMGKj7kDIv95UkIjK7UB0KYnk7JlFbnEzLbHRvv9doDbIBlBBm875S2YLtD/1aAwCu7JQguY2aKtTe7jeshHATQskNh+baa95oF1XeLHTObimqaJr1oSZw1kKIigyZyR9sxS3v/Kzp8Z+TWad/sdrzv2utlwXJCQkS/1mKLU158bu9+HirZ8sFlH4u2Nc/EWq8mFGoxM/PF/juyJuh11cdcL0REVEzYoZZWG8TCj95koItBALHJdYOai2hRUPWQGb7lk2ea5cQCQBoHRehap+2n0NV05/DvsZUYLW9nOMihMruTW8AVcik8lol4vFqkZkcb0zIC2tg28RL/0zDFNzkCmi8SmyuLbJ8hXPAJpWKbWY5h88bPQS3Faj8eastdGqRyKMKCtAnjDussHCd/XKl5b+oO9f6ihKDlzoxUCciIq+r0Wim8erO2vS0VkpuhlSsCJ2eal2kHQYFNlzcia39FmbX8mXaMtVpOGMkBIVil5tXdW6Yve7XJk7VPn/c11CN+cOfm1Y995RYpqSQ+SCWxeFOX2GxavxaEX7OcTLLKpRUkb5Q3nCRqudYyXPBOgVs/uq8SB0SZ1Itvc41FsQ8XVyJ3acb1ijL/X8mxpuz8e548Vtla+td3b+TKqbnS8SWa3kT/88mIiKv0yrlWpip1ZLcNZSZlie7+hkOat8QAI+8LMWt/e+SKZRToWE18DNFDbNRQm93pYT15IEa5u8LLcoW5xzVbJ9SlMxoy5GrwSD8mUbIHEPJsoDdZxr+BoSiUGROatqr7TlTgqe+3KXjaMwvOsx1cbF/fL+vyWNr9xVgwHM/2L5/c417ma7HC9XVINDypqkSYt057B08W4acQ+dFC6/a8/bNaz3I3cz2BgbqRETkl9wt+iU32xGmsq2OkVo23sRQO9sjEEvbFqgt/iRXzOdc48Xe9uNuzlxomJX9QWNV+k1eahvkibX7pPs7C22PxNKJ5dbnCjPnzh0brBq3giPPFYusQ1di9Gs/YdHGo9oORsQPGheIs1qtkufm7J1ncN5FcGnvmx1n3BrD1MXbHL73RjG1j7eewEGF9QUOndWmDoGrU+qTn+/ELQt+9otAXIm6eisuKMjC0IPvXHEQERGp4CsVabWYEI6U6RuuB7WV8YXra3er7+6T6XleIRJ4/nRAOojVg1aV4KXuEamdUJNbVym3rlV4H87Fl347WaxuAOQh17/wY4XK1hHL+enAWTz9lW/Mrt84bwNuf28zPth4tEnAfveH2/H8t3tUpZQXVVS7/P/W+f8jqZuXwmGVru0+p+KmwqP/+03xtltMeoPRNz6JpfV+egX6PrvSkGMzUCciIl0YXX5qb5506naem63b1DraeOEmV3FZWEvuiRgd+sTuOi0dnKldsyzcjIgIda/q+4rdTWfnhAJaYr1s9cwUFbtJodWSCKlAYIvK/tdSa5ZdFSAUAhfnTgQnvFR00F99vysP72844nrDRkr6iEsVGFPjgWW5eH/DUY/3oxW5/ua/nSzG+oPnMPvLXSgovRToCjf/Ptt+Cje8uUHxsfo8sxId/v4tfhA5twimLdmuaF/CEpHb3t2k+PhayiuuxPbjF/D4ZzsMOb4rVbXGdq06dt6z81eZiqUlWmOgbiJmLy5BRKSG3MydN1oFfZErvQbX3WrnamZOa+rqbTO7RhTVET5Tii+6lyK78aC5qzDXe7lggHCx9uZq5etS1a6fl9pcTV9tuYKHrmbGhZp4vBrR1l//sw1Pf7Vb8fauCkUCwO4zxfjF3eUijeTWGFfX1nu9kNbQl9agVkHWjdR5eMcp9Zkfdy3eKvncTwfOqdqXt6vtn7xQgYrqWty7ZBv+8O+NTZ6vrq3HYY3S4T0RHuydtpxSqkzYJlMpBuomstmkKStERO6Qq4Qtt77WaAMy4iWf26niQrC6th4XGteR2s8AeUtoUEM6vFirsbYt1bVCU6vGjSroagV4uQm8kAZ7QcXaYLVFoPR+R0p/ZN5eSkHqPfbpDvz+3xt1Wzu7ZNMx3PTWRtlZbldq6urx2faTioJvgZL7b+fLjFkv7KzM4DXaz32zB08s34ntx4tEn5+35iCu/eePmi3NcdfK3XkoqqjGtCXb3a6t0FwxUDcRtX0ViYjMLF2mh7PR5NrKtJSpJK/mcifQRe+abqnRAIB0lf3DtRAi0rJNS95IEMvqnqz/Qey4U11e/RIB/W8+KJld4vWI79Bj7eycb/fYMgA8Sfv9flceZnz8K9ZofGN205HzGPPaT3hg2S+GBqFxkcGSNzIqNeyMIeczif7lf/3PVlvRPKMzdksqa/Hdzjx8s+MMvt+dZ+hYfA0DdRPxdvsFIiI9qe3tK3UKlCsKp6YtkT3ZNaBeOhV3SIwEAETrsL7cSN66cG7h5np3dwW4ahpsEq6WOihZCuFOCjH5D/se4o9/ukN160TBI580FEIrUbH8Rsn5o6q2HrvPlOCL3NP4WaSzgbd8uyMPQ19aY9jx5Xy/K19xtXhvKG/8rFayrIMuYaBuIsEaFBQiIvJVIUHi50C52YCkqFC3jiV3X/RHmYrhWt5PjQxxHWi6CqqEa9rTXiqOp8TR800rH58zSaqqmWw8pHz968kLyteoX5SpyF9fr03xQmo+ck8U4YFluW69Vij4qOa0qbYK/daj3l1H74vUfmwdETmHe0pY/sXzjzoM1H3Md272fiQiMpOU6DBN9uPuOmW5Hr9yLW5cpbNrzVXRO7nZp3oPi8lpKTy4Yc2zmpk1V84asO5fCwPaxgEAyquUp8YeUdj2yRUrrIZ3YyDvqq+34uvfpAtrauHouXLZ4o5qjr/+oOsbWP/4fp/t32+uOah4382V2raY7VpGaj6GdxqzNI5qdC5Tau2+Aq8eT2sM1H3MAx/lGj0EIiJZ6fHhRg/BpUNuVsIN8nKgrnb5gL2Qxl7YYsXkvJ18KNzgUFtFWU65SP90XyC0qPORTHoATIX3ZesOnMX0pb80edzddcs1dfU4bdeF4ERhBYbNXYvnvtmDCon/J9fuO4v1Cv/fbxkpXSNECaMLp5mR2lZ+SvvBm92KXXm48/0tRg/DIwzUfYxc2xUiIjNo44Uicl1TogAArePcuykg1gtbiQ0KZnu0JFe7xNXFl1yygd7F5JzpcfHMC3JprTW+WbbrdNNe9aTec1/v9voM3+Gz4kGX2v9/7v7PNgANs9mDX1xte/x8Y9X5hRuO4OXsfaKvBYB7l2xTdBxPCxm+8O0ej17vqbs+2IID+aWGjoEaTP2Psr85M/PKJ/W8efOQkZGBsLAwZGZmYvPmzZLbLliwAFdeeSXi4uIQFxeHrKysJtvfeeedsFgsDl+jRo3S+23ojrXkiKi5GNopwbPXd2x4fbKbKfRSF6+ueHstuFgafnzjjJOrFmjC08slqgJ7wgxBsh5DGNszVfI5rQqyH29c/3nIzb9BJRJbuFe7gfT17vojthm+Fbvy8OTnO3U/5jNfK+/fLid7V0O17m9+c1yCaT8zL1fYrURhAO7pbO5764/g4Y9/9WgfnvhhTwHm/3gYC+wK8tEl3m6r6et0D9Q/+ugjzJgxA7Nnz8b27dvRu3dvjBw5EgUF4ncU165di1tuuQVr1qxBTk4O0tPTMWLECJw65XihMWrUKJw5c8b29d///lfvt6I7+1QiIiJ/JKRAe7titytVtdq30snTIKi/KNLiJzaioUq8q8ud/NKG4/8i0WPXE3LrUYX16Hpzt5CgsCRATEqM9I0fuWUIUteeJ0RqDJyS+axXk47cObmF4m0v7V+6BZzRLZyam6n/2Yb//HwM9fVW2b8Jswv2cnaOEp9uP2n48f+5cr+hY7D3uZtV+/Xg7Tovvk73/7teeeUVTJkyBZMmTUL37t0xf/58REREYOHChaLbL1myBPfeey/69OmDrl274t1330V9fT1WrVrlsF1oaChSUlJsX3FxcXq/Fd2Vu5mKSUTk64wOEuxbxpRUalPwrFKH4N8s9smkdgo3EpSqarwZoXaGuX2iewWPhndNAgAkaDjrLHXpWVDStOCdXLtBueeaHlP9BW9QgAVScZWaY5N2Fm44giEvrvZ6ccQcN9qaFZRW+vRNheZq5mc7jB4CuUnXQL26uhrbtm1DVlbWpQMGBCArKws5OTmK9lFRUYGamhrEx8c7PL527VokJSWhS5cuuOeee3D+vPQJp6qqCiUlJQ5fZjSPlSuJqJlwrkJbclE6LbJaZcVaT3lrRtiXnVfZbq1e5kaMbE97N8jN9gOX+qGbcWJH7yH5Si/45mT78Yb2YhsOnsP5soZg3RvZwbe/J70MVcrsL+Rbp+3N49pskif0UzfD8ilfoGugfu7cOdTV1SE5Odnh8eTkZOTl5Snax2OPPYa0tDSHYH/UqFFYvHgxVq1ahZdeegk//vgjRo8ejbo68dmLOXPmICYmxvaVnp7u/pvSkacFNIiIzOCMTMq38OG8ZNNxh8drZdZbq02t9HahNHcVXWwIdi+Uu99jXM+bGFoGdaEy6eZyz+lBWJJQoHIG0+r1Wvn6kPr/icVqjfXgR7kY8/pPAIAgD7o96Gn3mUsTXZWNmTBPfamu73lRhfvnO/J9764/gp2nirGfBfcUMeeZoNGLL76IZcuWYfny5QgLu7RubPz48bjhhhvQs2dPjBs3Dl9//TW2bNmCtWvXiu5n5syZKC4utn2dOHHCS++AiKh5sVqtKKpoSB3POaxNhXS5cNHoZbWlbqbJW61WrNzd0Mv9+13SPd1d8ZW6PHKFpAo9uFEhpVOS9Prt6LCG+ghq17gHmjR40kpQoI/8MfkZ+yUH+Y1LJcRaKprNsfMV+OX4BWw9dkHxa95cfQCzVQb25H+mL93udueV5kbXT52EhAQEBgYiP9/xIiQ/Px8pKSmyr507dy5efPFFrFixAr169ZLdtn379khISMDBg+Kp46GhoYiOjnb4IiIi7a3dd9YWeOWeKFL12pBA5Ren+Y2zod/sOK3qGFr76tczrjeSIGT+OS8DcNym6Z0IYXu9MwdLLmqzVh+QT3OMCtO+sGBStHgQXlNXbyumVqVyBtmdOFaPIoUAcPx80yJ1Spy6IL2+ONBX7vr4mZxDTZdumvWmSXnVpb/nz7afxHvrj6h6/dwV+/FFrrHnbDLe0fMVeO4bbboR+DtdA/WQkBD079/foRCcUBhu0KBBkq97+eWX8eyzzyI7OxsDBgxweZyTJ0/i/PnzSE2VbqtCRET625Pnfg2QcIlZJLHK2cJazp8PN21fZk/r9c/O5IJspf63TbpC8QqR2fYThQ3B1rr9Z2X3K9cGR0lMdlzk5y6QC7y3Hm06wya0lBMj9ztKlanCHiSTmh8XIX284MYgSM0NAneLHX7+S9OgRG62X6llW4673kjEw58Y17aKxJVV+c6yx3Nll5aLvL3uML7+zf0bleRdlSIdRIykRzcSf6R7HteMGTOwYMECfPDBB9izZw/uuecelJeXY9KkSQCAiRMnYubMmbbtX3rpJTz55JNYuHAhMjIykJeXh7y8PJSVlQEAysrK8Mgjj+Dnn3/G0aNHsWrVKtx4443o2LEjRo4cqffbISIiN7mbJl7qyYWswanxKQr6vIu1YBPIVVgudjHjLTdDGqRgHX/7BOmq6vVWq2Rdgae/ajpT0q9NrMy+pMegpOfubVe0afJYaFDDTZ9IkZs/wvFOyswuO7Na3ftTKhRZj9tCgwwCuZ9ZtUS9HiJqvp79uul5+UJ5Nf741kZbhk5lTR22HZO/+U3epXugfvPNN2Pu3LmYNWsW+vTpg9zcXGRnZ9sKzB0/fhxnzly6I/fWW2+huroaf/zjH5Gammr7mjt3LgAgMDAQv/32G2644QZ07twZkydPRv/+/fHTTz8hNFS7VitERKQtd4vHyPW9ViraKTiSi/8uVCi/oSA30/pF7mm0jgtXvC+1QoP1/Qh3d6ZPbbq3u0m+LUIb2sC1jFT32S9XNE1u4lwqe0Iuu0DsvUWFqWtfJ8ZqtUpmBMgVcyTzKyyvxo/75LNliNRasuk4LjqtC996rKHGwFe/NWT+vLn6IG56S1lXLvIO7ReGiZg+fTqmT58u+pxzAbijR4/K7is8PBzff/+9RiMjIiK9aNV+RfXaWQWbX98rzb3BOJHr1vHjvrO6FoVy1UdbLrVbyU80MlT69d1SpWu9iO1bmOEWI7d8QOrn5252hpz6eqvsbLTUz7ugtFLyb1TsJlOyRBE7tdX/pdbhbz5SiBt6q//7tnCNuimMfycH+/PLjB4G+aFus7Lxf2O6YcpV7R0eP1RQhtdXHcCbbBNtOv5dwpSIiAxTLlLVVS7Ac9a3MV368ox4yW3E1lG3iY8AAKTHX5rNdp4pXbO3oMnruiRHKR6bUnpm3q8/KD/rJhXIAcDoHvIFXYGGqujO91q6pjT8jIZ3S5J8ndisdL+2cZLbd02R/pu4pqv4ceT6uNuveRf7G4yLEJ/RrrdavdIWq3d6rOjj+1RknOw6LV0L4rPtp2RfW1kjfmPER7oa+oV6mZuYDNJJT89/uwd3vr8Zr686gIWNxQA/++UUXlm53+CRkRielomISBdhIrOodw5uK7m9cxq5sL47LLjpfoQ11nKz9vbHd17nrmbde0GJdCrxobPSF9XHCysk16iXV9cpCpblfLsjz+3XxkoEq8Cl9eSDOyY0qTovzH7LzZCL/WzDG3+HfxvRuclzfdJjJPclVzBO4LxWv8TFbLvczLGSNfFqhOgU/f58+LzsDSw5Uj/SKokAnrQ39OXVRg+BmrG1+87ilZX7kXO4accBMhcG6kREpIvTToXQ9uaVSgZ4FdV1qKkXDwr/b/mOJtv3byM+Q3u+rAq1ja87UHApiA5TsJ5bKkarqq23BZrOPpdpNVRWVYtAiahof34p2idKF2vzlKvWeK1iIySfEzISxIrJCenwF0SKpCkRHd70BoFUcbXKmjpFgfOijUebPGafTeHMnb7tR8+Xq34NgCa9guXS29UUlq+zWpEhU+xPjtSPVOpvlbR3uoh1BIjINQbqJuFu6xciIrNyvlv/4c/SLaVyTxQ1WQEszJYfPucYJFXV1kmm6e45UypaEVsq3dfetRJp1gDQOdnzllp6s88u2HpUvnJvWZV7a7yFNeMvZ+9z6/VyxvZybLG6dt9ZpMWqL8ZXWVOHjjIt0GJEbha48tWvZ2xt3dQIdlqjXlBaJbElsHSzey3XtPLTgXOGHp+IiBwxUDcJ+5kfIiJ/JHdDMs+pUnVhebVkf+3aOqtke7GfDp6VreotF2zpUfjNOXXc9ni9VbLF1tNf7nLrWPZ91avr6lEkUb2+rKoW4SHis9gHC+TXSf96ohgAsONUsVtjlPONU0/m2vp6yTRtOV/+einLQW3HgKKL4jPeJy5UoE1L6SwEKcEq3sBakboJUuTu7buq1i+VpcBCUkRE5sJA3STkLiyJiHyFc/sXe3J9v5//do/D96cuXJRs57b7TInk2vSi8hrU2UUxUxZvlRuuA7lZ92SZfuhyAb5UQFVeXYdQiSDyfZFUbiXs12afKapEhEQwLreu/6DdTeO9eU1//nI9390hd/NG7nNRLhXdvhK/2s/WFhKV7rcfvyD5GqEHsehzIsUOpaipmwAAoRI3q+TS/gGgdZz6Gw5EROR9DNRNQmrWhYjIlzivybUnNUMuRWpd+H9lUoSrausQaRc4b3aRAm5PrgjZlZ0TJZ/r3TpW8TEEQrVdvXy30/1Cc4Jnv96teFu54N++wvWCnw47PCf30bdP5EaB4P0NRxWPTQtyVbqd35M952w5K6xupdCLCZX4/8O5DoTz2NmFzbvk2g8SEclhoG4S245J3613tmKX5xdgRER66dlKvIq3kgre9qRCI7n+4Ttl2la563x5taK+46IkXvjjfvnWau4osZtJPlcmvRbalSPn1BdOk7vZ/L/tJ23/PlF4UXI7NXa6mXpfUeVeRkCVzMz89uNFivez5egF2b9fdcR/5s7XE3LZAKQ/ZkwSkbsYqJtEfonyi6q/L9+p40iIiDwjtS5Yrve1mBq7C1z79cur9uZLvuagTL0PV0U7pdKeXRVmk92nRPo5oH0aebRE9XRnVS6OKzeulpEhtn+fuqAs6D6pIv1bqToXv0upv7X/bpHOxrBCev223N+ucxV5udn3EpnlH8725pXgpn6tJZ+3T/GX8+j/fpN8brfMjS25jAYiItIfA3WTmP/jIcXbRim8GCMi8ragAItk7+hqFymgzuuO7QvGTVu63fZvqSJpAqk1uK6qWrdsESr6+BcyLdgA4LK0aNu/1aS5Frt4H2LUzM5JLREYv+BnyddsPXpBtmL6ebug9M73tygax/587YuluvobiA4X/5z86YB0JsPhs+Xobve7tGf/tyv0mZci1i7OHXvPlMp+3odJpL47O+NUqNH+htWY13+SfN1di5X9fkme0hsqRETOGKj7IFcVXYmIDGMBgtxcg+s889eyRYjElvJ6p4un3ueeKJJcEy1WOE3gqsK5fXCnd5qrc4E9uSBAKr3+8Fnp1PZ31x9BVKh0+zL7WftTRcpm1Pfkubcc4dPtp9x6ndxMf5nMz2vl7jzJrAp7248XyabePyOztn/5L+rek9rlImKcMyTk0vjtnSi8iIEZ8R4fv7nLPVFk9BCIyEcxUPdBZ2X6sBIRGentH6ULa7ly0i7Aqqytw6D2LbUYks0rK/dLFrS76a2Nbu/X3eDcnRqijzjdzHhf56J09sqratEuUXy2/ZhM5XM5csG+u+vsV+6WXhpxWGb9vXMhxA0HpTMw1snMzDfd76WbA3JLM5xZJas0eE+XlCijh+DzQoJYvY+I3MNAnYiITKe6tl6yVzoAyfR6T7haw66E89rplBjptm419eoD/D1nHGen7VPRa+q0CeykAsQj58oREy4+2/7uT4cRHymeASEXxF/58hqH72+VSctXqlbm51paWStZc+CL3NM4Ybee3vlnLUdulrr7rO8ln6uVWSrx0Ee/4l2FN2Kcq+4/+qn0unTyLgvL7BORmxioExGRplytBdeCfcE6LQJsQLrdFQAcVVgJfe73+xy+75x8aUay0ikFOdKu0JxUz3gAKKpQVoTvb5/8qmg7Z4PmrHL4/qJEP/nvdp4RfRxouLGiJG3clY2Hzks+J7f+3/5n6+qGxWm7Ndv2P9tTRRcx57u9tu/laipUOv2Mrv3nWtljSvlhT4Fbr3NWUOq4Dv0XFZXo7ZlhFt/v8EdKRG5ioE5ERLpxroBtn3ou12pSri+3s6UyfdXlOPcJ/+pX6aJxcjObs77YZfv34pxjDs+dL7+Uvv2r01pV+7X8F8qlg3HnlmZSs9pqXLRL83YuNpYcJV5U76TM2u9jOlR2d/bJ1pOSz8nN2ssV7Rv84mrJ55zv/7z03aWbMM5L0NzNZrj7w20O37uqhyBl7xn5Cu32N7bySyoltyu56LiGPyL00s0rthlzj9bdHYio+WCgTkREutnl1P7p460nbP+e+p9tzpvbuGq/Ze//3GxZ+Z5T8L35iLI2bM4FPeVmXu3XJN/8jmNa9/sbjtr+/dVv0jcJbl+4yeH7YV0Sbf/edFh6Btqe82y+c4Bob9mWS78j+/RvuXsn245dQKJEgO+s3M2CqEs3O94EOVZ4Kcth5KvrJF/3lw+kq5c7r0u355zJsE8m60ErExdudut1zjednNkH2TtOKk/pj4+4tJxhwU/u159ozrQoCEhEzRMDdSIi0s31b653+F5pMLzeRfr8zzIBamCA5x9t89YclHzu2HllafBA0xl2KR/+LJ0V4NyKzL7d2eli6dlRe09/tcv1Ro3sb0T8/t+Xiux99etp7JKZ8d127ILt33I3Ly6bLb1m25n9vYGdpxwDzJezHZcZKBmXGgt+ks6i2HxE2Q0SLX1sdwMFAA6dvfR3IFckz9nCDdLvq6pW+saFXMYJSWPmOxG5iw25iYjIdN5bf0Q2xfu7ndJp86+vOuDx8f/xvXQQ+Hc3Z/C1oqbImeC/m0+43kiB8zIp+vZeXen4O5DLkHBOt7aXI7NmXU6um2u0lTok0+LOFXcLvTm/zt1MEjlr9jmumf92x6W6BGqWo9AlcrUViIjkcEbdR2lVPImIyKyKL0qvLTbSryeKNFkn7s+cZ22f/Fw6qJz9pfRs/5duzuJ+oDCTgRy98O1eh+9/PXkpg+L3/Vp5ezh+wQKmvhORexio+6h7Ptxu9BCIiAA0Xf/cHJj1JgKRXt5ZxzXq7qiodq8mAxERA3Uflb0rr0k1ZSIiI+QoLGhGRL7LuVYCKRMcyEttInIPzx4+bO1+bfqvEhF5gktxiIiIiLTFQN2HlVYynYqIjPeXRVuNHgIRkSkFBXKNOhG5h4G6D3tgWa7RQyAiIiIiCelxEUYPgYh0FhWmTyM1BuomcEFhuxsiIiIi8h2BAZxRJ/J3UaE+HKjPmzcPGRkZCAsLQ2ZmJjZv3iy7/SeffIKuXbsiLCwMPXv2xLfffuvwvNVqxaxZs5Camorw8HBkZWXhwAHP++YaZfVe99ea17I/J5Hf0/ocSkRERETmpnug/tFHH2HGjBmYPXs2tm/fjt69e2PkyJEoKBAPTjdu3IhbbrkFkydPxi+//IJx48Zh3Lhx2LnzUg/Wl19+Ga+//jrmz5+PTZs2ITIyEiNHjkRlZaXeb0cXD3/yq9uv7fh/36G8qpYV4In8lB7nUCIiNSwW4IbeaUYPg4ioWbFYdS7Xm5mZicsvvxxvvvkmAKC+vh7p6em477778PjjjzfZ/uabb0Z5eTm+/vpr22NXXHEF+vTpg/nz58NqtSItLQ0PP/ww/va3vwEAiouLkZycjEWLFmH8+PEux1RSUoKYmBgUFxcjOjpao3fqvozHv9FsX73TY7FsyhUIDwnUbJ9E1JS3ziNan0Ndced9/Wvlfry2ynezmohIXmCABeP6tMI//9xb0fZmu84y0s5TxfjdG+uNHgYR6SgtJgwbZw7XfL/6JNQ3qq6uxrZt2zBz5kzbYwEBAcjKykJOTo7oa3JycjBjxgyHx0aOHInPP/8cAHDkyBHk5eUhKyvL9nxMTAwyMzORk5MjGqhXVVWhqqrK9n1JSYnq96JlMK2nX08UodusbM33e/TFsZrvk4jk6XEOdabF+TE2IhgBFoCJPURERETa0DX1/dy5c6irq0NycrLD48nJycjLyxN9TV5enuz2wn/V7HPOnDmIiYmxfaWnp7v1foiIvEmPc6gznh+JiIiIzEfXGXWzmDlzpsMMU0lJieqLUbkZZWH1QPHFGsSEBwMAdp0uQVJ0KOIiQgAAdfVWVNbUoabOigMFpaiurUf3tGhEhgThstnfq31LbmufGImY8GAM75qETslRSI+LQHCgBRGhQQgNCkBseDACLBZY7IqUWiysWErkr7Q4PwKAvouoiIiIiJoXXQP1hIQEBAYGIj8/3+Hx/Px8pKSkiL4mJSVFdnvhv/n5+UhNTXXYpk+fPqL7DA0NRWhoqLtvwyUhkI1tDMoBoEerGIdtggOBsOCGdeOJUdqOZVD7lsg5fB5H5oxhUE3kR/Q4hzrT4vz4zW9nwDidiIiISDu6pr6HhISgf//+WLVqle2x+vp6rFq1CoMGDRJ9zaBBgxy2B4CVK1fatm/Xrh1SUlIctikpKcGmTZsk9+nP7hycgf9OvQJHXxzLIJ3Iz+hxDtXD1mMXdNs3ERmvrt6KT7efNHoYRETNiu7t2WbMmIEFCxbggw8+wJ49e3DPPfegvLwckyZNAgBMnDjRoVDSAw88gOzsbPzzn//E3r178dRTT2Hr1q2YPn06gIbZ6wcffBDPPfccvvzyS+zYsQMTJ05EWloaxo0bp/fb0cX6x65x+7VP3XCZhiMhIrPR+hxKREREROan+xr1m2++GWfPnsWsWbOQl5eHPn36IDs721bs6Pjx4wgIuHS/YPDgwVi6dCmeeOIJ/P3vf0enTp3w+eefo0ePHrZtHn30UZSXl2Pq1KkoKirC0KFDkZ2djbCwML3fji7SYsKNHgIRmZQe51AiIiIiMjfd+6ibkRn7e7rT/u3eYR3w6KiuOoyGiFwx43lEC+68L19pX0lEnlHaqtVfz4/uWLOvAJPe32L0MIhIZ3q0stY99Z300yUlyughEBHhibHdjB4CEZEplVysMXoIROSjGKj7sMsz4o0eAhER2idGGj0EIiJTan55q0SkFQbqPiwtlmvbich4GS0ZqBMRibGyeSURuYmBuo9a94j7leKJiLTUPrGF0UPwuk5Jze89U/N2y8B0o4dARNSsMFD3UW1aRhg9BCKiZutAQZnRQ/ApEzLbSD43oG2c5HOXpTXvQmRmUnKx1ugh+KSQwECjh0BEPoqBOhERmdL913Z063UD2+lbv2P29d113b8/yJl5rcP3l6XFSG474QrpIP73fVu5dfx/3dzbrdd5wzM3XubW6xJahDh8P7xrkhbDcbD0rkyH7x8Z2cX27/35pZofrzkIsBg9AiLyVQzUiYhIN4+52ULy/UmXAxbpK9wnfycdLGdqEKgffH605HN/6Nda8X6u7pzo8Vic/am/8uMLvn/wKreO9bcRnW3/HtS+Ja6SeT/27zUuIkRyu5mjlf9NtAgNsv07PNhxZrJXa+ng3567gVLPVtL7d/cGAgCEBbk3w5ozc7jD91Ouau/2GKS0jnPM1guy++G1jmNdHHcEMlInIjcxUCciIt3cfLnjulb7GbqQQPc/guwDiAynpUD1GpRZDpIZW2SI8kDLPph8ymkm3j7le4nTTKY955sd++3S7pUGjM7tPG+TmcW+Z1gH27/vGJxh+3diVKjka1Kiw3D0fLmisfz16g6uNxLxR6cbFH+96tJ+Hsrq7Ly5zdu3D3DreKN6pEg+FxasTzrz7Ve0lXwuWOZvcuIg6dc5e/73PRRvW1NXb/v3/7ENIxGRVzFQJyIi3cRHOs6strLrVrFfZtZaTRC/csbV6gcGYN9zoxy+f/oG6ZRk+3E7B/FyKfr2AXJW92SH5wZ1aHlpnzKzbjf0SXP4/tcTRZLbKnVdd+kgtNiu77N9cOg8o22vd3qM7m2opsrMIN8xWDpQvc7p527vzVv7Kj6+fRZHWLA2l08DndqszrhO+oaDHFc3P4IDL/19pcVIz4yHOr2v/JIq2787JkU5b04KsOY7EbmLgToREWlKrjiYVqx2UaFckCsn1CkFubZe+pLafpbZmX2rzK/vG+rwXEV1ne3fzmnFp4ou2v7dOz1Wcv+p0WGSz9lzd+38e3c4zjiXVYoXDbtGZk10i9Bg3dtQpcdLF1GNlUm1l/O7Xo43QezXfTvfLGqXcOn4ztkFT7g527xkinQmhRpRYUGyz9fUXfrdOAfj9txNyydpIUG81CYi9/DsQUREmgpXkRrurnK7ANgis5bd29olOPaU33b0guS2gXbjlkulDlB4IyI2IljRds6Gd3OccZaaOb8sLdrhBom9fm1jcaLwouhz9tkIzub+ybHo27f3Xyk3VEWiXQStUjPs9w7rgJvs0uvV/B3f0DtN8rnHZdbky91kurZrEsb2TFV0/Ogwx9/9wjvdS/cnHXBKnYjcxEDdB7F/LxGZ1e96KQssXIkKC8KRc8rWPCt1TRfpQmh/GdLO7f26e5/AnSJTL/y+p8P39mnxATrfsEiPj8C+PPHK387rx+0FBUqPa0xPxxT87hq0Y0uJkc5A6NU6RjaF3944mfX/iS2k1+s7u1smLV3uJtMNvdOQKvNe5KTHadPC9WJNneuNSFadTKYOEZEcBuo+6Hx5tdFDICIS1Sou3O2A8f7hnWz/tsCCXaeL3dqPVL/nUT1SHNbq2nvouk6ijythfx2ud7Cc1c0x/by/xssMOie3kE1hLyitEn1c7n3LBcZyqdZywb8cuRR9ucyF/m3jHNbZ21ecd3ZjH+kgvm+bWMnn5NbZi9EixnO+IRSo4m/0hz35ng+gmYtxM9OFiIiBukno0cKHiMjb6uutKLIrRqbGeKcK8VU1lypOOweocn46cFb08VGXSc/2B1gsDhWu7f31avng6tsdZ2z/VpMuLVfFW0qSwvXqAPDcOPHq3q+N7yP5mmnXdJScNQcc12YrXRMvFzjLaRkpve5cLs0/wGJBvfivUjYNv1+bOOzLK3E5rr5tYmXXHc+/rb/kczHhyoO2sOBA1Em9ERWc19oH241drlr+46O7oqjCvf+X6ZLuqZ5niRBR88RA3STUFF+6rpt0BVsiIiPVW4FQiQDU3TReAHjz1n62f491M73e1czW6SLxNdaXpcn3666tk572rJYI/gGgZQv3CqBJsb+xAQAJEunZw7pIB86hQQGy6c5n7WbUnVvvSXHnhoQrrjIXKmvF38OkIRmSr4kKC3KoeG/PvnvBL8eLZI+drOJmipzh3ZKw8dB5yeelbiw5+8vQDMnn7h8u3bFALmWfiIj0x0DdJPqpCNT/zl6mRGRiu8+Iz0qqbd9lP2tpn7IcKjObqWbmXam+MhXZXamuFQ+mWsdJz+xqwZ3174KMlpGuN1LhloHKAno1IlxkLkjNXHdzMbsp9ScqlwZ/Rft4yeecqbkpHxwYgAMFZZLPR4SIj8k5a+C+a6WXdZipEKO/0ns5DBH5LwbqJuGqtYo9NalzRETeZAFQViW+RrxWZRpvlcTMbodE6YKaLSNDUVnjXrqwXHCbV1wp+Vxplfr04L+qXKushH3brcx2yoNHZ0IQfGWnBMWvkZs1T7Xr2z3tGuWztPEyqe+3XyHdN10PcgXxhnRQ/nOKjQhBjQbp7AAkK/A7Z67Ircsn/XmjCwYR+ScG6ibBO65E5A/iZdK55Xpdj3BqmRUWHCC5rwmZbWQrrds/98W0IdIbOpFbEy217h0Adp6SWdcsMc7OyVGyKfOeGtIxAfVqUxgAhNoFdX+9SvvUZ+e0cLnWc61ksg6yJFqseapUood8nMzf7k0yRe/E2qtJ/VpcZQk4K5W4IXa8sEL2dVLp/UREZC4M1E2iTUttWqkQEZmVXM/o6536UHdMaiHZAktu9nZIpwSHG5/Oadw1OgbHYqQqbAcHBUhWV1/zt2FuHeuK9i1t/w4KsKBWomR4ZEigZBB/ZcdLs8NiQ+/VWn69vieSopS3PJOrXG7/HtS2M5UqUje0Y4JkNwG5m0aVKtqbqXn/Sm9UiSmqYOcYIiJfwEDdJKLDmM5ORP7FuTJ2l5QoyW0TnYIUubWzESGBqJcIQtsnRLq9Pnu7TJEwd9Pp3dEuwb014vaz1fGRIWgRKj5DGxQYIDmbH+Si8Jtwk+TZGy9za4xynFO2uyRHodCNoLJzShQuVjcEyGJrvKXqBgANbQHFtE+MdCikp1RFtWOgLtUeEAAeuk66ArsaQQHuXdot/stATY5PRETaYKBORES6+J1T2u8tA9tIbis28ymsrX3xDz0dHrdYLJLpvZelRdtuENjvU64AneDUBfGq7wCwL1+8ZZm7ReHiIkKgsGi3W5wzFJxVSVRFB4CaxpsgYgGtEKx2T9Ou5ZSwZPvXk8UOj3dKjkK5RHq3vX4ifcs3HSmU3F6uT7yUMSIp7Eo4r22Xuwmjtlq83A0Hdyj5f4SIiLyHZ2UiItKF86x4cKAFuSeKJLd3TmmvagxEOogE8ScKxYNqqZn4Kg+CmqRo6ZTkP8qsT+6aEoXCcvEZ4YyWEZj/4yG3xwQ0rOOXfk5+vXOZxFpsAFi9pwAA8N3OM02eE2a95fYvtsRBSLX/bkdek+fk1lQrWWZ/ZadEh+9DggJkXydVLR2QDuLlUu3lONef0arKenJUGFbuznfrtTW14u/Ru4tCmrfXxvcxeghE5AMYqBMRkS7Eelkv3XRc8evX7W8o4ParSHAvxDtiM5RFFQ3FsuzTnqOdOmtcJ1KMrKRSvMhWaJB0UCoXwAVYLKrWKKv1txFdZJ937qtub51McTyhj/r5suomQbcwU73zVHGT1wnEiqIdOtvwu8g53LQv+Jp9BZL72p8v3p7MeVmFvRYygTgAVFRLrTW3oE5iSYW7xGoBSK0RV7MEblSPFMn30SZevuZNhMSSCLm0fNLWjX1aGT0EaqYiQwJxZM4YHH5hDF69uY/t8QSZQrBkHAbqRESkC+fgGACqVeR7C8XQzpapWxtcfNH1uuYLIjPdZ2RasEkJlZnVbqnzhY/czDAAnJRJ5T9V5Pq9HjlXLvnchoNNA245J2RmzeW6nuw5I15RPy1W2ZKDSJGbBlL1BgIDLG51YHH1e3C2+egF0cfVLCcIDQqQTH2fNCRD1XgEenYhIOU42056+ccfe+GXWSNgsVgQEGBBZGjDueuRkV2w8fHhyOqWZPAIyZmugXphYSEmTJiA6OhoxMbGYvLkySgrE787Lmx/3333oUuXLggPD0ebNm1w//33o7jY8c69xWJp8rVs2TI93woREXkoJtw7d+zFCoI53yDYekw8WAIANbXo5Gbb77m6g229tx6qZdaZa0GqAjoAlEpkHwDapXe7Utu4uF1tGzq5YoNyz0m9rZhw6Z+TczE5AJKFEFWxSC//iJYZD5nfjX1aIWfmtUYPg/xMWkwY/jQg3SEbSSji2rZlBEKCAvDwiC4Y28u9ehykD10D9QkTJmDXrl1YuXIlvv76a6xbtw5Tp06V3P706dM4ffo05s6di507d2LRokXIzs7G5MmTm2z7/vvv48yZM7avcePG6fhOiIjIUx1VtsoSSPW2VkJYc6ymartc4KVGfIsQ0bR9rbgKiOWKpikJFl1lP0jNPl/RPr7JY1KtzQD310YLQXD2zqbr3gXlIoGykOKdoqJ4m8Ui315Qilg6eUGp+syNJuORqE4PNG1JSET09u0DmjzWu3UMPp82BGN6NATn3VKjMe/Wft4eGslQl6+lwp49e5CdnY0tW7ZgwICGP4433ngDY8aMwdy5c5GW1rQibY8ePfDpp5/avu/QoQOef/553HbbbaitrUVQ0KXhxsbGIiUlRa/hExGRG+RmmF2RmhntLdK7W5j5dDV76KrdmBlkdWu6Xl4wro909fYBGXGy+5XrGa9kLfZpmfT44MAAyUD9yd91b/LY5qPSVdhDZNZGywW1wt+LWAs2ude5U23f3SyB4V2b/m7Pl3nex7xbahTW7BVf2y9XDf73fbk2mrTTObmFZB0JMpeeIp+jFosFfdJjvT8YUky3K5icnBzExsbagnQAyMrKQkBAADZt2qR4P8XFxYiOjnYI0gFg2rRpSEhIwMCBA7Fw4UJYZVLfqqqqUFJS4vBFRETas6+CHqly7a5UOrVYQNgpqaEn+11D28nuU+8kbLGLH6WEt5UeL73eumWLphXn2yc2zJi6mhH2NAO9Ryv33pvaPvZygaXczQY5UpcEgRaLLcBXmzLvDrEskvMSnQDUuL6XfPs9KTdfnu7xsUlbI0QKW5pV25aXChV+/+BVsl0vxESEBCLEB26ekv7sC9mRNN3+b8nLy0NSkmNRgqCgIMTHxyMvTzpNzd65c+fw7LPPNkmXf+aZZ/Dxxx9j5cqVuOmmm3DvvffijTfekNzPnDlzEBMTY/tKT+cHFRGRHmLCg3FZY1Gsm1RexKkhxILxkcZWqu3XRn5WW4rFYrFVrG+f6N6SAF8it97d4zsKItYfPCf6eECABRcaq64XlKorUqimEKLeAtxIw3fFTO+vOakR+bmXV7m/3MdbosKCMHFQhsNjrv4sdz09Eh/8ZaB+gyKfMOO6zujQDD73tKA6UH/88cdFi7nZf+3du9fjgZWUlGDs2LHo3r07nnrqKYfnnnzySQwZMgR9+/bFY489hkcffRT/+Mc/JPc1c+ZMFBcX275OnDjh8fiIiEicEHeJzagLF3JXdkpQvD9/DSAGtG0I8gdmNF3TrZQ7VcqVkstUU6tOZmY8VKbVmh6So4Re8OqOK7cu3JdILXvQ82+JpAldAwZ3aIkFExuyUN3NJNGbfbZMWmw4woID8daES2uaXa2osVgsGNShpV7DIx+QHh+O6dd0RGqs8hohzZnqNeoPP/ww7rzzTtlt2rdvj5SUFBQUOK6fqq2tRWFhocu15aWlpRg1ahSioqKwfPlyBAfLr0HMzMzEs88+i6qqKoSGNk0TDA0NFX3cbMZfno5lW3gTgYh8m1wxNovFAlitTe6mR4UF22Y6nQWrTJX0NLAXq9StByHw8yQ+kuvj7iktM8PldlWr8Y0YVz+SwMY18Wr6lvsTqZR/tUsWSBvhjS0Enx3Xw3ZeNGurvJmju2HK4q0Oj6XHX0qH75TUQrRmBJHg+l5pCAiwIEFkWRc1pTpQT0xMRGJiosvtBg0ahKKiImzbtg39+/cHAKxevRr19fXIzMyUfF1JSQlGjhyJ0NBQfPnllwgLc33HJTc3F3FxcT4RjMvplqq8hyoRkT8JkZlV9Xb4ENVMAzg1+rWVTvlXG2S0aPx5x8mlx4u4KHFDxVXht2PnG/rDq019d4faFHU9Utrt1dTVS6/fZ6BuiAeGd0JcRDDa2VXr90Zyw8HnR6Pj/32n6jXXdU9Gq9hwnCoSbw3IvyFz6tHKPPGF2hvvzZ1uP61u3bph1KhRmDJlCjZv3owNGzZg+vTpGD9+vK3i+6lTp9C1a1ds3rwZQEOQPmLECJSXl+O9995DSUkJ8vLykJeXh7q6hg/kr776Cu+++y527tyJgwcP4q233sILL7yA++67T6+34jVGr7UkItKbMJunZVq1FiJCLlWr1+paMzVGukicJ8TWtIrR84aDXLr62TL3AmC1n4G/nix26zg7T0kXlJWrhi9XfE7qZWIdCyJDpDsjqAl08t1s88bUd3O4vnfDtXB6fAT+b2x33W/S6EVJB4nm5vreabalTWawdMoVRg/Bhn8t6uh6W2PJkiXo2rUrhg8fjjFjxmDo0KF45513bM/X1NRg3759qKioAABs374dmzZtwo4dO9CxY0ekpqbavoR15cHBwZg3bx4GDRqEPn364O2338Yrr7yC2bNn6/lWvCIy1P22RkREvkCIdX7xsL/4hYqGCvHuXiQ698R2t/2WHC3WXguF+eydvNAwm+UqsGrR+JnyxNhuHo/DmVwGhLfocbPn0+0nJZ+rdeNvTSz46tMmFgCQ0MKzm/NFFeJdEgDXa4XFME73jj80tsh76aae+OnRa3Q/nlyLR3ckRTtmr9r/qUWG6tb1WbHgQAu+uX+ooWO4PCMOL97Uy9Ax2AsOMP58LVB6o5ka6Pp/VHx8PJYuXSr5fEZGhsMH7bBhw1x+8I4aNQqjRo3SbIxERKQtJfFTQYlnacdCQLXjVDGu6ux6OZazrqlRbh1XmAXzlqSopsu/hJ+v0sBKLKj2h1kwPW4WlFbqX207sPGiWa6Wg6d2nnIv20CKj072ms6vs0bY1qRHhAQhIl7/wHZYlyR8nnu6yeNqb06ufOgqAMCcP/TE9zvzL+3HbpvZ13fHDW9uEH39eIWtAXu3jnE7WwYAvrpvKLqmGJfqvfrhq9G2ZSQDUgnePJUMbBePzUcKvXhE7ZnnFgspMqanfCE+IiKjbTx03mvHcn9G3b2PP29XKK/zYNZYqBxdVdP0gtEbPcTtCRdnQg940pfWN2LG9Wml6f6aq5iIYK9nowzu2BJ90mObPK52PXmn5Iabm11TovFAVifb493TojFxUFts/r/h6NW66XGAhuUez47roeg4JR7eLDMySAca2m1yrb45fPzXQRjVw7fjJgbqPuYlE6XSEBG5q7be2NkGuR7vcmulq2u9O+7wYPklUbKTYo2xWqjKNmR66qhh79zWcRGuNzKhbUcbZngqRW6gSDH6ut8MSx3IPUlRYfh82hDd9h8cGIBnbuwhmv0jeDCrs+IiYiEKtht12aXgS8/MFH+hdlmL2sKeSgg3i+IivFuPa/5t/b16PK3xzOtjWI2YiPzBubKmrdik1gDLzQ0Wlou3dHMlXuZi4Q/9pGcPvb2O19UMtLDWPi1G+iLZ21V2xS7EhBn8YA0DPuc6A76ivLFa/dCOCYpfo1Xv6UCLxevZFOT7hndN8uj1agpFvjq+j8ttutvV7uAElmtqz5V6nCGEz9XocOPrGPgSBuomUl3LD08iar6k4ofgQOmLDLnn5MgF3LFeuuNf0Fi12x/Wi9sTuyhvEab9xZmvx5sxKmat1FRkdy725cByaY28nE5J2mU+kG97a0I/vPLnPm691p1Clh0UZN2E2WUJiRXc9JaBGfGYPLSdYcfXi9yNbHfVqGzbSQ0YqJuIr84OEBGJyStxr32UM7kgRY+WUlpVE3c1c7lm31kAwNFz5Zocr7nZfNR7tRAA6b7tZuPq/4lYBanCiVEywT6ZyuOju2q+z1UPX41eja0F2ye2UHVTyd7Iy1LQPTUagztqkxEiSI0Jx6qHr8b+50YjRSabSG9hIYF48nfdRZ/zViZTSrT4+//tqRG41sNMCC31bNXw96TkRgxdwkDdRPqZqOciEZGn5NZzR+sww6qVTYelq8SquSi0KKxvW1mrTwBY3Vh1WOyGwYkLFbocU+CNG8+nGtvUeYs7a2GVrLe1p3dauq9nITQXSv4Oru6ciMlDMnD31R00P36HxBZ445a+GNszFe0S3C8AmR4fgW8fuBKpMeEajq7herlDYgvDayfInee8UVCue2o03ry1r+hz0WHB+PuYrvjLkHaGF7e7pksiLs+IQ87MazEgI97QsfgaBuomomYNDxGR2cldbMoVc9OK3KxgmEyBteKL0v2ppWYvxISHBKJHq4a0zLG9UhW/TmtiqfVKCpl58pmkR6aD0YT0/QmZbRS/RqtlDWpaacldlCudKXe39gOJu6pzItq2VF78UElXikdHdcGT11/mybBktW0ZiXkT+nk1GP7rVe0VLWfS8uzyOw/OzUYHwPcP74QBGfF4/Za+uFXkvNQxKQqzru+uuhWf1hKjQmGxWDS/YdMcMFAnIiJdyLUW88aM68Qr2ko+17NVrO7HB4BhnRtSD//YT/rGhN6znBEh7mUv3CDTM77ezQDU3WUFYrPZYY0V8V1VxtdKYOPFbo/GFE4l1L7bQIkL6qxuylNY5YKHzo0ttqQEBzW8NtqpcK3RAYmvW3Tn5fj+wasUbx+kIFit1WDNr9Le5t4yc0w3l0FlUlSoQ8FK4f+ZsOAALP7LQMXHmn19d3xy9yD888+9JbdZOiVT0b6Ec9AzN+p340TODb3T8MLve+J6mXO2kYoqpG9+e4Mv19xgoE5ERH5JrsK4tyYYAhoDHLlAR4sL7tPF2tQDsBcoEyxEhqoL/huz8FFe5V6av9hskfAz7dm6aeA80Mvple4WNXQm9XfSQuXPu0pmOYVcpouQCSEUOhQM62Keta6+KCDAYruxpGh7BScotf8Pinn6xsuw6e/DPd6PNyy/dzC+mDYEm/8vC+Ehl36WAQEW9GsTi7dvH4CrOicq3t+kIe1weUY8QoOkfy+DOyjrzCB057gtU/rmsDeM62POQF3t+Utrnn7efzFtiGQtAr0xUCciIr8U6mbKptyHurf7qHtCmL2uqXNvzKEy66tdzcw6E5YaiP1OhNhU7mJObQZGWIh7s+y9G3v9+oJkmeruA9s1FO9qL7K+WG7GUrhR4FwIS26pCBmjo5uzhI+M7GL7d2hQIJJVLOdR42oVQbMSfdvESf7/+dm9Q1Qdr6eKrBh7Nw/QJgMhoYXyYo1HXxyreFutf+Zy4iNDEKtDv3Uz6p0ea1h1f555iYjIZ+lRfEsukPGlpdc/NxbF23xEujienLRY6fWE3VO1a4nUrXFff1F5IST86surajUby+geKQDcayullqfdBa6RmeWOa7yAFpt1FdYcx8lcZMc5tWfSK5gjbUSpmLG8d1gH5M66TsfRNPBW1XN3pMW69/f8zLjLcM+wS8X7Ls9wrwi02l7iXRTeGA3S6Gfu6nPuv1OuwPYnr1NdLJPU40+YiIi8TqviNuv2n9NkP0rJpaR6O72v1sU68fzG9niHz5Y1eU5JBXNvtegqLGsoXKZ2HXTruIYbCVd20m4WSRiBsG971sYV57UiGQru/D1frPGs2r+SQ8rNeN0+KMPl64d1afjZJkcxUPcXFosFsTr0yfYlSoo87ntuVJPHQoMC8dioS+3w/jKkneL92YsKM/dM9O/7tlK0XYSLzCX77A1fNUTj1oJqMVD3Id5MaSEi0pNWxanOllVpsh+l5Mbt7XpbLT2oyi5UoO7vpbagJxtbqZ0T+X19u+MMAGDd/rOq9in0eR4g8h7kZovlyF1uC8sexLaRKv4l9jchPKaksre7hNZ8Px3w7EZWSWMHhGo3l0+Qd+jd1s/fKLmpKlXYUZAWE2abwVZ70y3WjVaP4vT5vU9QuNbel5aCuSs9Tnm3Bj0wUPchRt/VISIyG63aXxnJqvJiS+hBr0VWgrdvLmw/XtTkMT1+g0Lv5/m39VP1uh925wMAck8UN3lOSOWtqFZ+Ue5uxX1P1ctcP9c1PllW6XrJQEsVa2nJOKFOheqEtpBGqpP7I3SS1S0Ju58Zqenx5Wo4aM15qYgrak/dZRLLe8Tuz7SJ9zywVDo++88gpRXySR1jy/ARERF5QO6Cwt1Ze3dbj7l70+CMDhXbXREqzXsyI+LOj1dJyr2WnAMYV46cKwcA5BVfbPKcsNQgKsy3L52E6uMsEGde6fHq+k07n+uKGzMh7AvOzflDT/x6osjjsSmlJr37ndsH2DpkaOWnR69F5ye+E31Oiy4Gfdt4JxsJaFiGdKqo6TlJrMf9sqlXoFTBTTg57iRoWDTtbq8td7uNmAHP0kREpAtv9EqXq+zez80LqVYi65OFVEm5Yj3hblYaV3tjoMTDizAHHszKu1O46GaRvs3CbFGVydMohR+VHyRxKFZa2RDw1aqYHSXPybUME3NFe8eMy8QWofhq+lC8f+fltsduGdgGL97US5PxyXltfB8AQIZIxwEpzkG6Fp8dYkEsAOx4agTGKVyDLeWDvwxU1DNdqxuTavaTFhuOLinqunK4q5/dsiOxc0Ski8/EiYO8087Ol2+uMlAnIiJdeOtiQYraqsNCBVuxgnGPjWooipPkpQJrArEe61rMhAr7qPKwoJlaw0RqrQg3QbSsICw3I+RpDBCucpbemVkDfbHgqGUkU999QbBIjYSerWOQrkEatFpq08DF1Om45l6LQm5Xd050WBZyx+C2iAwJbHLfc+sTWR4fCwCe/F033GVQezA5Qt/2P/ZvjYyWjjdmnhvXA0M7NfShFzI8nD1zYw99B9joQEHTgqpa+Gr6UF32a4+Bug9hrRCi5qWwsBATJkxAdHQ0YmNjMXnyZJSVSX/gFBYW4r777kOXLl0QHh6ONm3a4P7770dxcdP1tt4gVyHdjG7NbANAvPXYnwak4//GdMO1XT1PmVRDbP16ZY2y2c2axoiwVGR9oxD/u1rnrlXRPyW0/HMRWraJ3ehIaWw1NvKyZO0O6Ia9eSWKtxX72bRqbJ8nduNG2F7NjaUeae71libzEP5/Hd0jVdXrpGafjWJ/vfvxXwfhuweu9MpxhXPDTf1aqzr3/b5va+x4amST86nUzWKlqelCu8iOSVF44nfdFY9Hb0JxSeHtto2PQHp8BA48PxqfTxuCz6cNwW1XtEXruAikxYZh2rUdDRytftl9PVvrf8401/+ZRERkM2HCBOzatQsrV67E119/jXXr1mHq1KmS258+fRqnT5/G3LlzsXPnTixatAjZ2dmYPHmyF0dtHu1UpF4CsM0+hYmk64UFB2LKVe1N1xtY7mIyonHmVyyNtkdaw82INi7WwmrVRk/gqqWcVkIbg1e59kFqbyQJwYOn/c8FFxXecAHEK8QLqcV/GyHdAumyNOVFxbReI0zeF2ABcmddh7uuVDf72jpW3Zp4T90/vJPibSNDA9FN5OapHv40oDU2Pn4t/vnn3rZzn9KClGr+/3FVUV7w7wnyxzbqXrjUOTA4MAB90mPRJz0WQMPn5sbHh7u9DE0rwg0PX2SuKw6SFeHlHr1EZJw9e/YgOzsb7777LjIzMzF06FC88cYbWLZsGU6fPi36mh49euDTTz/F9ddfjw4dOuDaa6/F888/j6+++gq1tRqua/YRYumgvsY+yBab9RIuJuX6ZYsR2n32bryg0pJc3QBfbiO1dl9D+7htxy5osr8oHT/T80saChSu2de05Z2wlLSkUjwdlXxbbESI+htsXj5Vzrius+zz9wzrYPt3jUhWjF4sFgvSnG5aaJHK747RPVKa/B47JDbcnHtibDfc1K81BrVPMGJoqmsoyPFGR6s7h5hv2YBSDNR9SKjJZnKISD85OTmIjY3FgAEDbI9lZWUhICAAmzZtUryf4uJiREdHIyhIPCioqqpCSUmJw1dzoEe8KDZDe7G64QaJ2urqQrXmeA96pcdFNgTvrURmy7SeKbcX5uEabntCEaBYDS+WhdoJvVrHqnrdsfMNFeHFqi+7Y6SOszxywY2QKu/pWnsiMb1bx2JA2zj8wYOCbeP6eFbsTa32KrOvPJWoYEmKWLbUu3dcjm6pUbihTxr++efebhcwdcVbmU8AkBajfzZH5+QWrjcyKUZ+PsTHlnsSkQfy8vKQlOS4HjooKAjx8fHIy8tTtI9z587h2WeflU2XnzNnDmJiYmxf6elNq3L7Iy0LlwnBdFeR9EyhcJHa87fcrLRSPVo1rJ8b5cNpf0LK4uAO2s26dElp+D31bROr2T7dEWJQxkfblpGIDgvCiMuk/y6MKEJGymn5/4PWYiKC8b97Bquq+u7M29e7T93guoK7VkKDAnCZm33u2yVE4rsHrkJSVJjGo3Kk1fIerSybeoXo4y1FbmS3Funa4mv1cuwxUCci8qLHH38cFotF9mvv3r0eH6ekpARjx45F9+7d8dRTT0luN3PmTBQXF9u+Tpw44fGxfYKGn9txjWnnYksUhcfkJig6JOlzt19YBym2BCA5umFGx9szSWoFNq7N9uHrLElGVX9PjArFr7NH4PKM+CbPBQcG4MU/9MRjo6XXvZPxCsurjR6Crtq29J8bRf5Y+qGjTp9ZUq5o31JxBpDaZWBmx0XPRERe9PDDD+POO++U3aZ9+/ZISUlBQUGBw+O1tbUoLCxESor8DGlpaSlGjRqFqKgoLF++HMHB0h9coaGhCA313xZMRlczFiq019RJp77rNTsi9957tY7Ft/dfia4Gt9BzparWvfZxJy80pKeXiVS8F34XRk8aeTO91Jnc0ofxA9t4cSQk5ZaBbfDfzcdFn9OzL7QZek5ruQbaaBaLpcnJplzkvOQrXhvfR1HqvtbiIoJxsdjx80DLZVZmpesVjNrWQgAwbNiwJrNLd999t8M2x48fx9ixYxEREYGkpCQ88sgjzbJQEhH5nsTERHTt2lX2KyQkBIMGDUJRURG2bdtme+3q1atRX1+PzMxMyf2XlJRgxIgRCAkJwZdffomwMH1T5NQSqlC38FJxTKFKu/PdeG8dv0XjRa/SWRVvtgzrnhZtikrfQjBdXtU0KP9s+ykAwKGCclX7/CK3oeDibyeLmjz38daGrJE9Z5pHPQbyTXLBnEXH6m9KW4f5mtjwYPx5QGujhwEACAn0zQAzMMCCMT3Vtf7Tyl+v7tDksWdulF6yIFabxRfpGqirbS0kmDJlCs6cOWP7evnll23P1dXVYezYsaiursbGjRvxwQcfYNGiRZg1a5aeb4WIyKu6deuGUaNGYcqUKdi8eTM2bNiA6dOnY/z48UhLSwMAnDp1Cl27dsXmzZsBXArSy8vL8d5776GkpAR5eXnIy8tDXZ17M5NaE3qVJ7RourZMWBerR+/uIKf0by2PEdI4+xOmYBZo2dQr8N8p4uvtjr44Fm/fPkD0OVeMnh3WwoGCUsnn6jR8g0Ig4q8BCfmHmHDpTKgrO4lX+9YiS0Ns3a8eHh/dFYA+53sxPz5yDZ7/fU+vHMtfjbosxbAWpc7tVtNiwjC8WzJe/IP475SBugvutBYSREREICUlxfYVHX2p6MKKFSuwe/dufPjhh+jTpw9Gjx6NZ599FvPmzUN1tX+v2SGi5mXJkiXo2rUrhg8fjjFjxmDo0KF45513bM/X1NRg3759qKioAABs374dmzZtwo4dO9CxY0ekpqbavsyy9lyuf6xQ4K1jovL1b2YoEnNt14ZWZ9f3TnO57RXtW2KQUyEo55sIUjLbNV1TLCi+2LTVVufkhrT2ti29vw5dq1/Ldd0bMgx6t47RZod+bP5t/WRnmMi3CH/7Yuzbl9nT4n6Wt06pd4vMkIqRWzakRkxEsMsgs0KHlPQ6A5e4NAd6L9Uxuo2lboG6J62FlixZgoSEBPTo0QMzZ860XYQK++3ZsyeSky+dwEaOHImSkhLs2rVL+zdCRGSQ+Ph4LF26FKWlpSguLsbChQvRosWlIDYjIwNWqxXDhg0D0LB0yGq1in5lZGQY8yZUEC4Q9+VLz6w6S49TV3TI+a68FoT1lBEqW+X0a6w6rqSg24eTM/HPP/VWtf/BHVri55nD0cfNXumerO/Xqv2b8DPVM9XXFyiZxRrVIxUTB2U0bq/Nz8sM65Wbq6s6JyJJYi1wkMazmnKz93rzRiDbLVW+FocwgmCDa5r4GuHGUHm1OTL2PCXWDjU82NhzoG5/ke62Frr11lvx4YcfYs2aNZg5cyb+85//4LbbbnPYr32QDsD2vdR+falPcJyfVSskIlJKCEbS4xtS1oQAVpgZFqUyHjHBkmybuMbe4GJB7R/6Oa6lHNopAUnR6uoNWCwWpMSor1GQ0KIhOPB2iuPVnRO9ejwzc+77q7aSsdAW8He9PFtPet/wTph6VXukqPzbI22oTQt3t63jszf2cOt13iCcJz3x1fShss8LS7HE0qVr6jy7kRAW7P/BvyetJpVmTETqXFcmJToM6x+7RnYbIxL4VP/16N1aaOrUqRg5ciR69uyJCRMmYPHixVi+fDkOHTrk9j59qU/wpCHtJJ+r94dFiERECnVprEiuNkgRLrbsA2DhgveRkV1sBe3M7C9DMrD7mZFGD0OSHkXo0vxkTaEWWkY6zqS6O0MeGdJwcetudf9WseH4+5hummVIkL7c7Qfeo9WlpSVXd06S2VJ7rgJZLVqBucpAuCwtBt/cPxS3yqRRm7l3vVG0uJkrdEZxRchAcyW68Sal2pvUKTFhiAiRvxkgFobdP7yTquOopfr2hDdaC9kTqhsfPHgQHTp0QEpKiq1wkiA/Px8AJPc7c+ZMzJgxw/Z9SUmJqYN1KfygJCJSr0tKFCYNzgDQkB48qof4LGNdveMFw/J7B6OqVpv1kWpZLBbJi4Y28RE4Xlgh+pzeQoICUG3Qz8QbElqE4nx5NeK8VFDLlY5JLXCwoAw9W8UaPRTyAXJ1LJR66obuGozE91yWJl4HQ5gk62LyVpa+KlRhxoHSGKhfmzhsPHQev+/bStU4ZjYWN1RrxnWd8fqqA269VgnVgXpiYiISE12np9m3Furfvz8AZa2FnOXm5gIAUlNTbft9/vnnUVBQYEutX7lyJaKjo9G9u/jJxd/7BBMRNVdKgsbvH7xK1b6EgL1vmzj3B0Y+acRlydiXXyqahi+kaFZ4cT1mtMr2foKkqFAEBVgwqofyiRHyDT/PHK7LfpdOyURRRY1t2QQ1EJYTeFKzw18JP5NoD2oc/H1MN1s7TVcGd2iJjYfOy24jxPNqZvt7t45BZnvXGRMhgQGoFknVf++OATh8Vl0LUaV0S/i3by00f/581NTUiLYWGj58OBYvXoyBAwfi0KFDWLp0KcaMGYOWLVvit99+w0MPPYSrrroKvXr1AgCMGDEC3bt3x+23346XX34ZeXl5eOKJJzBt2jS/D8bDgn2z7yIRkRbuGdYBreMc06O1TDQKkui57i+eufEynC6qNHoYpiZ0ERDrJiB8BmtVqE1PcZEh2PfcaK+1viLvkUvp9WSB5OAO4i3fiBZNuhx3vr+lyeOXZ8Rj5uiuHlVeT1ZR/2LplCuQ8fg3bh9LzD//1BtZMh0WlBjeLRnDu2k0ICe6rsxfsmQJpk+fjuHDhyMgIAA33XQTXn/9ddvzzq2FQkJC8MMPP+DVV19FeXk50tPTcdNNN+GJJ56wvSYwMBBff/017rnnHgwaNAiRkZG444478Mwzz+j5VkyhG9NuiKgZe2yUe6lpSgkBelCAd2ZO7hicgb15pYgM9c6NAaEiuCf8Oe3dFTO0ArSX2Fj0z/nmlYBBevNR3thWTKtWZkT2pG4OBQZY8FeFbfbMKjYi2NCuB67oGqgLrYWkCK2FBOnp6fjxxx9d7rdt27b49ttvNRmjLzHZNQIRkS7yS6oMOe4f+rVCfkklruys3cySUNiuZYumGV9XdU7Ehsev1eQ4t2bq20uWzGdgu3h8/NdBGNDW9RINtQUZybcIs5LNvY2hL3Cue9IpqQUOFJQZNBoyOzbI9CEs+k5EvqRWQVubVrHhOFV00eGxeA3a8bgjKiwYj2o8az+mZypatghRFEy56+v7hurSH97smvtnosViwUCFxcOeuuEyfLL1pM4jIq0pbQ025cr2KK2sRftE/zkP/PToNdhw8JzRw/CYfVXzHmnRmDSkHU5euCjzigbN/PTmUnyLS9cJaTFhOF3sn8u6WBnBh7DqOxGZVXxkCEY4rfOyNl5qPJjVyVZ13dmbt/bFE2PFF3cJs0SBMue+tvERboxWWwMyGoJwsWA5MMCCwR0SdD1/92gVo3uPWTNau/8sAGDHqWJVr+vRqqE9X6dkz9s++YquKdF48nfNs5q3b1MWrqXHR+BfN/fRpF2WWaTHR3i09lkrwhKSSBetu5T4+v4rcVP/1oq2vejFopVGc6dbwYt/6GX7973XdLT9u0XjZ2GQD9QSUaL5fbKbXHl1reRzQvVXIiKzWfHQVbYLmb5tYvHL8SLbcw9mdZZ8Xd82cZLV1e8f3gntEiKRGCVdKLRbqvE90Qd3SMDeZ0c124KfRs3i7c8vBYAmGRmu9GsTh52nSiTXdrujuc/uE/mrPumx+OtV7XH7FW3den1QgAW19epPELEmXjetNXcq6vdodamdXmXNpZsa4we2QUV1ndtZbO0TIx0quN/YJ82t/WiFkZ/JSFUbTmgRiiQVlRGJiLwpwW4NdopG56r4yBDcITETb7RZv+vuUDfEm0G6mQvfNFchfjSTSUSXBAUGYOYYnUp6+whhGVtJZY3BIxFnn0kSEx6Mh66TnhxwJSzI8bP8KpFWnd7EQN1kpAqBdEtlxXciIrP4y9B2hh37+d/3wGfbTxl2fGqKK9OISExAgAVwmlGvkMmeNSMhjby/jrVWvK1FaBDKqpr+Hq7pmojdZ0oMGJE43gI2mT8OULZ2hYiIPBcWHIA/9Gtl9DBU6dU6Fk/dcJmm++yaEoWxPVMln+/gR0Wq9MBAnYjE/H100wKlf1S4Tt0oQU4ZQulxEXj79v6YelV7g0akvb9KvBfnGXV7RtQ14Yy6yQitfJwVVZgz3YSIyJf99Oi1iA7X9qNw8tB2yDl0XtN96m35vUMQ7CfFd4wQ2nhxJ/UZ7q5JQ9rh0NmdpuvhTuQL5v6pt0MbaCPcMTgDT3212+GxTslRuPvqDpj/46Em21/TJRFr9p1F0UXjr/szWkZgyZQrEBkahJGXpRg9HADA38d0NawmyD3DOmD60l+8ekwG6j6CaxKJiLQnV6jOXU+M7YY6N4oHGSk8xP019gwhL4nw4Oco5vreafhdr1R2fWmm+raJw8rd+UYPQ1f27cu0ZqaZa6mC0MVOAfngDglYs++s5ucSNcKCG2bU/zK0neY3Hz019aoOql+j1enTiN8JU9+JiMjnCC22RveQTtc2isViaZI66EtahCq/GGkdF477ru2k42ikpcY0FC1M0uFmi5kwSG++WjTDtov+yrkgdJ/0WABAh0THdGoz/O8eERKEbU9kuV3p3tsCAuR/aEEByj+POyRJp7cbMZPvu1cSzcxAN3oMEhH5q7iIEABAyxYhBo/E/zz/+574Q19l6/bXP3YtxincVmvXdk0GAAztmGDI8YnIc7eYoFe6EUZeloxtT2QhIyHC4XGzZGO1bBHqMzcJr++l3Q37UTIp/vbV5b01u85bdT7i7qvVp3oQEVHzMvWq9ghzoyetvSEdEzDEKfg1Y50U2ySKb1xLEqlWatJ2WFrZ9PfhiI1onks7LRYLWrZomg1kRMEyoxWWV0s+p6TSfGyEdjfsXc3OA8D4y9PxVy/FZQzUfUSIhxdeRETk//6uU79fPdbyN3cKrgepmYuP9O+MoWSndHAC2ic0v0A9VCbGGd4tyYsjUSYpKhTtErzTCYXRHxERmUqb+AjXG5HXtGwRwqwuAEM6tgQAjO2Vpsn+ghipE5GTwMbzQlJU87mJESlTi6G5Xw9wRp2IiExj9vXdcXXnRKOHQXa2/l+Wz6xV1JOQXpli4CzgP//UG0Fso0fkt1JiwjDrd90Nq/1hpMlD2+G99UccHuvbxnXquz9joE5ERKYxaUg7o4dAThikm8dNJmo3RUTas6ChLVpz9GBWJ1ugHhoUgKpabVr3WdCwhLha4f7CgwNxsaZOk2N7iqnvRESkqTYtG1LV1LREISJjCD2Tich4vDHaIKtbsmb7CgiwYM8zoxRvv3zaYADAuD7aLHPyBGfUiYhIU9Ou6YjL0mJYgMwPdU2Jwt68UqOH4RdCg7zT3seVV2/ug5W7C4weBrmQM/Na7DpVYvQwSGNRYY5V7wNZu8ItH/91kOyMuZqfa9eUaOx+ZiRCAo2/iclAnYiINBUdFowbeht/J5q09+at/bDrdLHi7TsmeVbB2GqOlsKaEtZhmuVG1qgeqRjVQ7s+xKSP1JhwpMaEGz0M0tiUK9vjnXWHjR6GzxvYLl7T/UWENA2RY8Ibbqq09mKBO+NvFRAREWlIrtULeaZjUgvc2Ee8yJHYhZLzHEaAyrTOCxXS/XV9VXiwOWbSyfyKL/p3H3Vi60ujDO2UgPDgQHROjlL8ml6tY/CfyQPxey8W+uPVDBEReZ2eM6Wzb7gM13Y1X+9Vf/f6+L5YNvUK0ef6tYkFAFUXRQAQFdYwqxEc5L/poLxQJynC/y/MhibSVt82cdj9zEikq5gdt1gsuLJTIoK9mBLPQJ2IiLymuq5hDZmeFVWv7pyIhXdertv+SVxKTBiuaN9S9LmEFg3BqFx9QS37igs9z3u2itFsn3pZNvUKfHP/UKOHQSY09ar2AIBWcUx5p+bhei8um/OFwn0M1ImIyGsiG9d9efOONJnbyMsaqvv2a6tdv1xhJjLZwJ7nSnVIbIHL0sx/Q4G8z7nQGJG/e+mmnvjp0Wts5/DmvpSteb97k5r1u+5GD4GIiMgrYsNDAJi/nZ9wwch7TESkhzdv7Wv0EAwXERKE9PgI/PXq9lhyV6YtG6u54seNCU0akmH0EIiIdCGkcLbRoWoq29qYS0hQACYOzjB6GJr5Y//WuGNQWwzukGD0UIjID/2uF7ulCMKCAzGkI8+1bM9GRERek9kuHt89cCW6pqgrKuZKt9Qo/HlAuqb7JM/senqkXy1xiI0IwdM39jB6GERE1Ezo+glaWFiICRMmIDo6GrGxsZg8eTLKysoktz969CgsFovo1yeffGLbTuz5ZcuW6flWiIhIAxaLBd1Soz0u4jL1qvbomNTClo787f1XYtKQdloMkTSiR5DeKamFrTVOaWUtAKC+3rN9vnmrdLV6o8RFcm0yUXMRGcKWjSRO1xn1CRMm4MyZM1i5ciVqamowadIkTJ06FUuXLhXdPj09HWfOnHF47J133sE//vEPjB492uHx999/H6NGjbJ9Hxsbq/n4iYjInK7slIgfZlxt+94XqreS5766b6jtBkBYcMN/QzwsNmSmdNPQoABMu6YDbh7QxuihEJGXLJ6ciROFFUYPwzARvFEhSbdAfc+ePcjOzsaWLVswYMAAAMAbb7yBMWPGYO7cuUhLa/rBGBgYiJSUFIfHli9fjj//+c9o0aKFw+OxsbFNtiUiIiL/FRbc9ILOF0oTjOuThh6NreLaJUQCAOIiQppsZ7FY8MjIrl4dGxEZq3/bOPTXsOuFr3lkZBfEhofYusLQJbqlvufk5CA2NtYWpANAVlYWAgICsGnTJkX72LZtG3JzczF58uQmz02bNg0JCQkYOHAgFi5cCKvVKrmfqqoqlJSUOHz5Aq3XcBIREZH3vTq+L+66sqEn9o190vDpPYPRPS3a4FERERmvY1IUXvpjLwT4wl1XL9Pt1kVeXh6SkpIcDxYUhPj4eOTl5Snax3vvvYdu3bph8ODBDo8/88wzuPbaaxEREYEVK1bg3nvvRVlZGe6//37R/cyZMwdPP/20e2/EQOFMBSEiIvIrQYEBzXr2jIiIlFE9o/74449LFnwTvvbu3evxwC5evIilS5eKzqY/+eSTGDJkCPr27YvHHnsMjz76KP7xj39I7mvmzJkoLi62fZ04ccLj8REREZG+hPRwCzjTQs1XRstIo4dARAZQPaP+8MMP484775Tdpn379khJSUFBQYHD47W1tSgsLFS0tvx///sfKioqMHHiRJfbZmZm4tlnn0VVVRVCQ0ObPB8aGir6OBEREXnfZWnRGNS+pcvtHhjeCcnRYUiO5mc4NU8v/L4nruzEftJEzZHqQD0xMRGJiYkutxs0aBCKioqwbds29O/fHwCwevVq1NfXIzMz0+Xr33vvPdxwww2KjpWbm4u4uDgG40RERD7gy+lDFRWBS4oOw/3DO+k/ICKTujWTHQD8wfmyKqOHQD5ItzXq3bp1w6hRozBlyhTMnz8fNTU1mD59OsaPH2+r+H7q1CkMHz4cixcvxsCBA22vPXjwINatW4dvv/22yX6/+uor5Ofn44orrkBYWBhWrlyJF154AX/729/0eitERESkoUAWDSKiZiQqLNjoIZAP0rUO/pIlSzB9+nQMHz4cAQEBuOmmm/D666/bnq+pqcG+fftQUeHYO3DhwoVo3bo1RowY0WSfwcHBmDdvHh566CFYrVZ07NgRr7zyCqZMmaLnWyEiIiIiIlLMYrE0/tfggZBP0jVQj4+Px9KlSyWfz8jIEG2r9sILL+CFF14Qfc2oUaMwatQozcZIREREREREZCa69VEnzxVV1Bg9BCIiIiIiIvIyBuomFh2ma8IDERERERERmRADdRMS1rNwQQsREZE5xUeGGD0EIiLyY5yyJSIiIlLpo6lX4GJNncNjESGBBo2GiIiUuKaL69bfZsFAnYiIiEilTslRDt+P65OG2wdlGDMYIiLyOwzUTax36xj8eqLI6GEQERGRC6+O72v0EIiIyI9wjbpJLbkrE4+M7GL0MIiIiIiIiMjLOKNuUkM6Jhg9BCIiomalW2o0ACA1JgwAcN+1HZEUHWbkkIiIqJlioE5ERESm0TY+AkM6GXOzeuRlyfjp0WuQHh8BAHh4BDPbiIjIGEx9JyIyqcLCQkyYMAHR0dGIjY3F5MmTUVZWpui1VqsVo0ePhsViweeff67vQIk09P1DV+G5G3sYcmyLxWIL0omIiIzEQJ2IyKQmTJiAXbt2YeXKlfj666+xbt06TJ06VdFrX331VVgsFp1HSKS9sOBABATwb5eIiJo3pr4TEZnQnj17kJ2djS1btmDAgAEAgDfeeANjxozB3LlzkZaWJvna3Nxc/POf/8TWrVuRmprqrSETuYX3k4jI39XWWY0eAvkgzqgTEZlQTk4OYmNjbUE6AGRlZSEgIACbNm2SfF1FRQVuvfVWzJs3DykpKS6PU1VVhZKSEocvIq1NuKItkqJCER0W3OS52we1RZ/0WKTHhxswMiIi/VXV1hk9BPJBDNSJiEwoLy8PSUlJDo8FBQUhPj4eeXl5kq976KGHMHjwYNx4442KjjNnzhzExMTYvtLT0z0aN5GYqzsnYvP/ZSEsOLDJc5elxeDzaUMQEeJZkl9ESCD6pMd6tA8iIj3ERoQYPQRqFBLkO+Gv74yUiMgPPP7447BYLLJfe/fudWvfX375JVavXo1XX31V8WtmzpyJ4uJi29eJEyfcOjaR0b574Eq8e8cA1xsSEVGzNah9S6OHoBjXqBMRedHDDz+MO++8U3ab9u3bIyUlBQUFBQ6P19bWorCwUDKlffXq1Th06BBiY2MdHr/ppptw5ZVXYu3atU1eExoaitDQUDVvgciU2raMtP07oUXD37QvzZwQEZH+aut9p14AA3UiIi9KTExEYmKiy+0GDRqEoqIibNu2Df379wfQEIjX19cjMzNT9DWPP/447rrrLofHevbsiX/961+4/vrrPR88kY+4Y3AGYiOC0T012uihEBGRiQT6UFcRBupERCbUrVs3jBo1ClOmTMH8+fNRU1OD6dOnY/z48baK76dOncLw4cOxePFiDBw4ECkpKaKz7W3atEG7du28/RaIDNMiNAgTMtsaPQwiIiK3MSeMiMiklixZgq5du2L48OEYM2YMhg4dinfeecf2fE1NDfbt24eKigoDR0lkDPZaJyJf0CI0CJOH8mY5qccZdSIik4qPj8fSpUsln8/IyIDVKr/WytXzRL7q9iva4ti5CvRoFWP0UIiIJP06e4RPpVuTeTBQJyIiIp/TOi4C82/vb/QwiIhkMUgndzH1nYiIiIiIiMhEOKNOREREfu0PfVuhc0qU0cMgIiJSjIE6ERER+bVXbu5j9BCIiIhU0S31/fnnn8fgwYMRERGB2NhYRa+xWq2YNWsWUlNTER4ejqysLBw4cMBhm8LCQkyYMAHR0dGIjY3F5MmTUVZWpsM7ICIiIlJv9vXdMbZnqtHDICIiH6ZboF5dXY0//elPuOeeexS/5uWXX8brr7+O+fPnY9OmTYiMjMTIkSNRWVlp22bChAnYtWsXVq5cia+//hrr1q3D1KlT9XgLRERERKpNGtIO8yb0M3oYRETkw3RLfX/66acBAIsWLVK0vdVqxauvvoonnngCN954IwBg8eLFSE5Oxueff47x48djz549yM7OxpYtWzBgwAAAwBtvvIExY8Zg7ty5SEtL0+W9EBEREREREXmLaaq+HzlyBHl5ecjKyrI9FhMTg8zMTOTk5AAAcnJyEBsbawvSASArKwsBAQHYtGmT5L6rqqpQUlLi8OUr7hnWAfcO62D0MIiIiPzCnweko2tKFDISIo0eChEReckX04agXUIkru2aZPRQFDNNMbm8vDwAQHJyssPjycnJtufy8vKQlOT4ww0KCkJ8fLxtGzFz5syxzfD7msdGdTV6CERERH6jR6sYZD94ldHDICIiL+qdHos1fxtm9DBUUTWj/vjjj8Nisch+7d27V6+xum3mzJkoLi62fZ04ccLoIRERERERERGJUjWj/vDDD+POO++U3aZ9+/ZuDSQlJQUAkJ+fj9TUS5VS8/Pz0adPH9s2BQUFDq+rra1FYWGh7fViQkNDERoa6ta4iIiIiIiIiLxJVaCemJiIxMREXQbSrl07pKSkYNWqVbbAvKSkBJs2bbJVjh80aBCKioqwbds29O/fHwCwevVq1NfXIzMzU5dxEREREREREXmTbsXkjh8/jtzcXBw/fhx1dXXIzc1Fbm6uQ8/zrl27Yvny5QAAi8WCBx98EM899xy+/PJL7NixAxMnTkRaWhrGjRsHAOjWrRtGjRqFKVOmYPPmzdiwYQOmT5+O8ePHs+I7ERERERER+QXdisnNmjULH3zwge37vn37AgDWrFmDYcOGAQD27duH4uJi2zaPPvooysvLMXXqVBQVFWHo0KHIzs5GWFiYbZslS5Zg+vTpGD58OAICAnDTTTfh9ddf1+ttEBEREREREXmVxWq1Wo0ehLeVlJQgJiYGxcXFiI6ONno4ROSD/PU84q/vi4i8h+cRIiLPmaaPOhERERERERExUCciIiIiIiIyFQbqRERERERERCbCQJ2IiIiIiIjIRBioExEREREREZmIbu3ZzEwodF9SUmLwSIjIVwnnD39rnMHzIxF5yl/Pj0RE3tQsA/XS0lIAQHp6usEjISJfV1paipiYGKOHoRmeH4lIK/52fiQi8qZm2Ue9vr4ep0+fRlRUFCwWi6LXlJSUID09HSdOnPB6T1Aem8fmsc13bKvVitLSUqSlpSEgwH9WEfH8yGPz2Dy2p8f21/MjEZE3NcsZ9YCAALRu3dqt10ZHR3v9Q5LH5rF5bHMe2x9ninh+5LF5bB5bi2P74/mRiMibeJuTiIiIiIiIyEQYqBMRERERERGZCAN1hUJDQzF79myEhoby2Dw2j81jk53m+rvjsXlsHpuIiPTSLIvJEREREREREZkVZ9SJiIiIiIiITISBOhEREREREZGJMFAnIiIiIiIiMhEG6kREREREREQmwkBdgXnz5iEjIwNhYWHIzMzE5s2bDT3GokWLYLFYHL7CwsI0Hc+6detw/fXXIy0tDRaLBZ9//rmh+1+7dm2T92yxWJCXl6fZmObMmYPLL78cUVFRSEpKwrhx47Bv3z7N9u/uMbzx+37rrbfQq1cvREdHIzo6GoMGDcJ3331n2P698Z7FvPjii7BYLHjwwQd1P5a/4PlR+/OjO8fwh3Mkz488PxIR0SUM1F346KOPMGPGDMyePRvbt29H7969MXLkSBQUFBh6jOjoaJw5c8b2dezYMc3GAwDl5eXo3bs35s2bp+l+Pd3/vn37HN53UlKSZmP68ccfMW3aNPz8889YuXIlampqMGLECJSXlxt+DL1/361bt8aLL76Ibdu2YevWrbj22mtx4403YteuXYbtX+/37GzLli14++230atXL12P4094ftTn/OjJMXz5HMnzI8+PRERkx0qyBg4caJ02bZrt+7q6OmtaWpp1zpw5hh3j/ffft8bExGh2fFcAWJcvX27o/tesWWMFYL1w4YJu43BWUFBgBWD98ccfDT2Gt3/fgri4OOu7775ryP69/Z5LS0utnTp1sq5cudJ69dVXWx944AGvHduX8fyo//lR6TH88RzJ8yPPj0REzRln1GVUV1dj27ZtyMrKsj0WEBCArKws5OTkGHqMsrIytG3bFunp6Zre2Te7Pn36IDU1Fddddx02bNig67GKi4sBAPHx8YYfw5u/77q6Oixbtgzl5eUYNGiQYfv35nueNm0axo4d6/D/Icnj+dGc/OkcyfMjz49ERM0ZA3UZ586dQ11dHZKTkx0eT05O1mzdnzvH6NKlCxYuXIgvvvgCH374Ierr6zF48GCcPHlSkzGZUWpqKubPn49PP/0Un376KdLT0zFs2DBs375dl+PV19fjwQcfxJAhQ9CjRw9Dj+Gt3/eOHTvQokULhIaG4u6778by5cvRvXt3Q/bvzb/xZcuWYfv27ZgzZ47m+/ZnPD+ai7+dI3l+5PmRiKjZM3pK38xOnTplBWDduHGjw+OPPPKIdeDAgaY5RnV1tbVDhw7WJ554QpMxOYMJUt/FXHXVVdbbbrtN+wFZrda7777b2rZtW+uJEyd02b8nx9Dr911VVWU9cOCAdevWrdbHH3/cmpCQYN21a5cp9q/Xez5+/Lg1KSnJ+uuvv9oeY2qnMjw/NtD7/OjJMXz5HMnzI8+PRETNXZBxtwjMLyEhAYGBgcjPz3d4PD8/HykpKaY5RnBwMPr27YuDBw9qMiZfMXDgQKxfv17z/U6fPh1ff/011q1bh9atW2u+f0+PodfvOyQkBB07dgQA9O/fH1u2bMFrr72Gt99+2/D96/Wet23bhoKCAvTr18/2WF1dHdatW4c333wTVVVVCAwM1PSY/oLnR/Pz1XMkz488PxIREVPfZYWEhKB///5YtWqV7bH6+nqsWrVKs7VpWhyjrq4OO3bsQGpqqiZj8hW5ubmavmer1Yrp06dj+fLlWL16Ndq1a6fZvrU8hrd+3/X19aiqqjLF/vV6z8OHD8eOHTuQm5tr+xowYAAmTJiA3NxcXoTK4PnR/HztHMnzo3v75/mRiMg/cUbdhRkzZuCOO+7AgAEDMHDgQLz66qsoLy/HpEmTvHaMiRMnolWrVrY1Ys888wyuuOIKdOzYEUVFRfjHP/6BY8eO4a677tJsTGVlZQ53548cOYLc3FzEx8ejTZs2uu9/5syZOHXqFBYvXgwAePXVV9GuXTtcdtllqKysxLvvvovVq1djxYoVHo9FMG3aNCxduhRffPEFoqKibGtgY2JiEB4e7rVjGPH7njlzJkaPHo02bdqgtLQUS5cuxdq1a/H99997Zf9GvGcAiIqKarL+NTIyEi1bttStNoE/4fmxgdbnRyXH8MdzJM+PPD8SEZEdYzPvfcMbb7xhbdOmjTUkJMQ6cOBA688//+zVY1x99dXWO+64w/b9gw8+aNs2OTnZOmbMGOv27ds1HY/Q6sf5y34ceu7/jjvusF599dW27V966SVrhw4drGFhYdb4+HjrsGHDrKtXr9ZkLAKx8QCwvv/++149hhG/77/85S/Wtm3bWkNCQqyJiYnW4cOHW1esWOG1/RvxnqVwDaY6PD9qf35Ucgx/PEfy/MjzIxERXWKxWq1WfW4BEBEREREREZFaXKNOREREREREZCIM1ImIiIiIiIhMhIE6ERERERERkYkwUCciIiIiIiIyEQbqRERERERERCbCQJ2IiIiIiIjIRBioExEREREREZkIA3Vqtu68806MGzfO6GEQEZkOz49ERETGCjJ6AER6sFgsss/Pnj0br732GqxWq5dGRERkDjw/EhERmZ/Fyk9i8kN5eXm2f3/00UeYNWsW9u3bZ3usRYsWaNGihRFDIyIyFM+PRERE5sfUd/JLKSkptq+YmBhYLBaHx1q0aNEktXPYsGG477778OCDDyIuLg7JyclYsGABysvLMWnSJERFRaFjx4747rvvHI61c+dOjB49Gi1atEBycjJuv/12nDt3zsvvmIhIGZ4fiYiIzI+BOpGdDz74AAkJCdi8eTPuu+8+3HPPPfjTn/6EwYMHY/v27RgxYgRuv/12VFRUAACKiopw7bXXom/fvti6dSuys7ORn5+PP//5zwa/EyIibfH8SERE5D0M1Ins9O7dG0888QQ6deqEmTNnIiwsDAkJCZgyZQo6deqEWbNm4fz58/jtt98AAG+++Sb69u2LF154AV27dkXfvn2xcOFCrFmzBvv37zf43RARaYfnRyIiIu9hMTkiO7169bL9OzAwEC1btkTPnj1tjyUnJwMACgoKAAC//vor1qxZI7qe89ChQ+jcubPOIyYi8g6eH4mIiLyHgTqRneDgYIfvLRaLw2NCteT6+noAQFlZGa6//nq89NJLTfaVmpqq40iJiLyL50ciIiLvYaBO5IF+/frh008/RUZGBoKC+L8TEZGA50ciIiL3cY06kQemTZuGwsJC3HLLLdiyZQsOHTqE77//HpMmTUJdXZ3RwyMiMgzPj0RERO5joE7kgbS0NGzYsAF1dXUYMWIEevbsiQcffBCxsbEICOD/XkTUfPH8SERE5D6L1Wq1Gj0IIiIiIiIiImrAW9pEREREREREJsJAnYiIiIiIiMhEGKgTERERERERmQgDdSIiIiIiIiITYaBOREREREREZCIM1ImIiIiIiIhMhIE6ERERERERkYkwUCciIiIiIiIyEQbqRERERERERCbCQJ2IiIiIiIjIRBioExEREREREZkIA3UiIiIiIiIiE/l/jQAnv2CvMx0AAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 1200x400 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import random\n",
    "\n",
    "print('Slices array shape: {}'.format(slices.shape))\n",
    "\n",
    "fig, ax =  plt.subplots(nrows=1, ncols=3, sharex=True)\n",
    "fig.set_size_inches(12, 4)\n",
    "for i in range(3):\n",
    "    idx = random.randint(0, len(slices) - 1)\n",
    "    y = slices[idx]\n",
    "    librosa.display.waveshow(y.astype(float), sr=SAMPLE_RATE, ax=ax[i])\n",
    "    plt.axis('off')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### II.I Data cleaning"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In order to perform an accurate analysis of the audio dataset, it is important to ensure that each audio file contains sufficient information and features for the task at hand. To achieve this, any audio files that have a duration of fewer than 0.30 seconds will be removed from the dataset. This will eliminate extremely short audio files and ensure that the remaining files provide adequate information for analysis. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "67 instances removed\n"
     ]
    }
   ],
   "source": [
    "MIN_LEN = 0.30\n",
    "\n",
    "small_slices_idx = list(df.loc[df['length'] < MIN_LEN].index)\n",
    "for idx in small_slices_idx:\n",
    "    slices[idx] = np.nan\n",
    "    classes[idx] = np.nan\n",
    "    \n",
    "slices = slices[~pd.isnull(slices)]\n",
    "classes = classes[~pd.isnull(classes)]\n",
    "\n",
    "df_clean = df.loc[df['length'] >= MIN_LEN]\n",
    "\n",
    "print('{} instances removed'.format(df.shape[0] - df_clean.shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### II.II Audio Splitting"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Generally, it is advisable to split the formatted audio files into several **equal** brief segments with predetermined time intervals. For this project, a threshold value of 2 seconds is set for all clips. Two seconds is a suitable threshold as it is equal to half of the maximum lengths of all classes. It is also slightly higher than the mean length of the interest class (1.65 s) and less than the mean length across all classes (3.61 s)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "goal_len = 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "44100"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "length_of_wave_arr = round(goal_len * SAMPLE_RATE) # the length of the desired slices segments arrays\n",
    "length_of_wave_arr"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The following function splits audio slices using the following algorithm:\n",
    "1. If the slice is 2 seconds long, keep it as it is.\n",
    "2. If the slice is less than 2 seconds long, pad with silence.\n",
    "3. If slice is longer, the audio is split into two segments; the first segment is 2 seconds long, and the second segment is the remainder of the clip.\n",
    "4. If the second segment is shorter than 0.30 seconds, it will be discarded.\n",
    "5. If the second segment is larger than or equal to 2 seconds: <br/>\n",
    "    5.1. If it belongs to either the gun_shot or the car_horn class, the new seqment will be added to the queue and will then undergo the same process from step 1. This will add more instances to the two under-represented classes. <br/>\n",
    "    5.2. If it belongs to any of the other eight classes, the segment is discarded. This is done to avoid adding more instances to the larger classes, further increasing the gap in the class imbalance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def split_audio_slices(slices, classes, split_length=length_of_wave_arr):\n",
    "    X = slices.copy()\n",
    "    y = classes.copy()\n",
    "    \n",
    "    idx = 0\n",
    "    while idx < X.shape[0]:\n",
    "        _slice = X[idx]\n",
    "        _class = y[idx]\n",
    "        \n",
    "        if(_slice.shape[0] == split_length): # If it is already 2 seconds long, skip.\n",
    "            idx += 1\n",
    "            continue\n",
    "            \n",
    "        elif(_slice.shape[0] < split_length):  # If it is less than 2 seconds long\n",
    "            diff = split_length - _slice.shape[0]\n",
    "            silence = np.zeros(split_length) # silence\n",
    "            lead = silence[0 : math.ceil(diff / 2)]\n",
    "            trail = silence[0 : math.floor(diff / 2)]\n",
    "            \n",
    "            X[idx] = np.concatenate((lead, _slice, trail)) # pad with silence\n",
    "        \n",
    "        else:\n",
    "            # split into two segments\n",
    "            seg1 = _slice[:length_of_wave_arr]\n",
    "            seg2 = _slice[length_of_wave_arr:]\n",
    "            \n",
    "            X[idx] = seg1\n",
    "            \n",
    "            # If it is longer than 2 s and belongs to the small classes\n",
    "            # add to the queue to undergo another split/silence padding\n",
    "            if(seg2.shape[0] >= MIN_LEN and (_class == 'gun_shot' or _class == 'car_horn')):\n",
    "                X = np.array(list(X) + list(np.array([seg2])), dtype=object )\n",
    "                y = np.append(y, _class)\n",
    "                \n",
    "        idx += 1\n",
    "        \n",
    "    return X, y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "X, y = split_audio_slices(slices, classes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "After splitting the large audio slices of the small classes and adding them back to the queue, it is noticed that the dataset size was increased."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((9014,), (9014,), 9014)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "X.shape, y.shape, len([x for x in list(X) if len(x) == length_of_wave_arr])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The gun_shot class instances were increased by 1% while the car_horn class was increased by 2%."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "counts_per_class = get_class_dist(pd.DataFrame(y, columns=['class']), 'class', True);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### II.III Data splitting"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Before further processing, the data is split to avoid data leakage. The data is split into 80%, 10% and 10% for training, validation and testing, respectively."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "_X, X_test, _y, y_test = train_test_split(X, y, test_size=0.10, random_state=101)\n",
    "X_train, X_val, y_train, y_val = train_test_split(_X, _y, test_size=0.10, random_state=101)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### II.IV Data augmentation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'jackhammer': 816,\n",
       " 'siren': 750,\n",
       " 'gun_shot': 388,\n",
       " 'dog_bark': 798,\n",
       " 'car_horn': 497,\n",
       " 'drilling': 812,\n",
       " 'engine_idling': 818,\n",
       " 'air_conditioner': 822,\n",
       " 'street_music': 792,\n",
       " 'children_playing': 807}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "get_class_dist(pd.DataFrame(y_train, columns=['class']), 'class') "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Even with the slight increase with data splitting, the gun_shot and car_horn classes still contain the least ammount of instances"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To decrease the class imbalance, and to further increase the count of the smaller classes, data augmentation techniques will be applied to the smaller classes. <br/>\n",
    "Two data augmentation methods are applied, noise induction and time shifting. Each class is capped at 700 instances."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "MIN_CLASS_COUNT = 700"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import random\n",
    "import time\n",
    "\n",
    "def augment_class(data, labels, name, n_instances):\n",
    "    '''augments n_instances of the given class name'''\n",
    "    print('Augmenting for {}...'.format(name), end='')\n",
    "    shift_aug = math.ceil(n_instances / 2) # 50% will undergo time shift augmentation\n",
    "    noise_aug = math.floor(n_instances / 2) # the other 50% will undergo noise induction augmentation\n",
    "    \n",
    "    new_data = []\n",
    "    count = 0 # counts the number of newly added instances\n",
    "    idx = 0\n",
    "    while count < n_instances and idx < len(data):\n",
    "        if labels[idx] == name:\n",
    "            _slice = data.iloc[idx][0]\n",
    "            if count < shift_aug:\n",
    "                shift = int((random.randint(5, 20)/100) * len(_slice))\n",
    "                new_slice = np.roll(_slice, shift)\n",
    "            else:\n",
    "                new_slice = _slice + 0.01 * np.random.normal(0,1,len(_slice))\n",
    "                \n",
    "            data.loc[len(data)] = [new_slice]\n",
    "            labels = np.append(labels, name)\n",
    "            count += 1\n",
    "            \n",
    "        idx += 1\n",
    "        if(idx == len(data) and count < n_instances):\n",
    "            idx = 0\n",
    "            \n",
    "    print('done!')\n",
    "    return  data, labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Augmenting for gun_shot...done!\n",
      "Augmenting for car_horn...done!\n"
     ]
    }
   ],
   "source": [
    "class_counts = list(get_class_dist(pd.DataFrame(y_train, columns=['class']), 'class').items())\n",
    "\n",
    "for _class in class_counts:\n",
    "    name, count = _class\n",
    "    if count < MIN_CLASS_COUNT:\n",
    "        X_train, y_train = augment_class(pd.DataFrame(X_train), y_train, name, MIN_CLASS_COUNT - count)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('jackhammer', 816),\n",
       " ('siren', 750),\n",
       " ('gun_shot', 700),\n",
       " ('dog_bark', 798),\n",
       " ('car_horn', 700),\n",
       " ('drilling', 812),\n",
       " ('engine_idling', 818),\n",
       " ('air_conditioner', 822),\n",
       " ('street_music', 792),\n",
       " ('children_playing', 807)]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "class_counts = list(get_class_dist(pd.DataFrame(y_train, columns=['class']), 'class').items())\n",
    "class_counts"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### II.V Data resampling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To completely eleminate the remaining class imbalance, data resampling will be applied. <br/>\n",
    "The interest class, `gun_shot`, now contains exactly 700 instances which is the minimum class count among all other classes. For this reason, data **under-sampling** is applied. Under sampling will help avoid introducing fake instances to the interest class specially after applying augmentation to the same class. This will ensure the reliability of the available data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from imblearn.under_sampling import RandomUnderSampler\n",
    "\n",
    "rus = RandomUnderSampler(random_state=0)\n",
    "X_resampled, y_resampled = rus.fit_resample(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7000"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "len(X_resampled)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The training set is now equally split into 10 classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "{'air_conditioner': 700,\n",
       " 'car_horn': 700,\n",
       " 'children_playing': 700,\n",
       " 'dog_bark': 700,\n",
       " 'drilling': 700,\n",
       " 'engine_idling': 700,\n",
       " 'gun_shot': 700,\n",
       " 'jackhammer': 700,\n",
       " 'siren': 700,\n",
       " 'street_music': 700}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "get_class_dist(pd.DataFrame(y_resampled, columns=['class']), 'class', True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here, the 3 data sets are converted again to wav form and saved to disk to be loaded again as sound files for future use."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import soundfile as sf\n",
    "\n",
    "def export_x_and_y_helper(x, y, dir_path):\n",
    "    for idx in range(len(x)):\n",
    "        arr = x[idx]\n",
    "        label = y[idx]\n",
    "        # name of the file is formatted as follows: {index}-{classID}\n",
    "        filename = '{}/{}-{}.wav'.format(dir_path, idx, label_map[label])\n",
    "        \n",
    "        sf.write(filename, arr.astype(float), SAMPLE_RATE)\n",
    "        \n",
    "def export_x_and_ys():\n",
    "    path = './processed_data'\n",
    "    if not os.path.exists(path):\n",
    "        os.mkdir(path)\n",
    "        os.mkdir('{}/train'.format(path))\n",
    "        os.mkdir('{}/val'.format(path))\n",
    "        os.mkdir('{}/test'.format(path))\n",
    "        \n",
    "    print('Exporting training set...', end='')\n",
    "    export_x_and_y_helper(X_resampled[0], y_resampled, '{}/train'.format(path))\n",
    "    print('done!')\n",
    "    \n",
    "    print('Exporting validation set...', end='')\n",
    "    export_x_and_y_helper(X_val, y_val, '{}/val'.format(path))\n",
    "    print('done!')\n",
    "    \n",
    "    print('Exporting test set...', end='')\n",
    "    export_x_and_y_helper(X_test, y_test, '{}/test'.format(path))\n",
    "    print('done!')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Exporting training set...done!\n",
      "Exporting validation set...done!\n",
      "Exporting test set...done!\n"
     ]
    }
   ],
   "source": [
    "export_x_and_ys()"
   ]
  }
 ],
 "metadata": {
  "language_info": {
   "name": "python"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}