Notebooks/Freight value prediction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h1><center>Freight Value Prediction</center></h1>\n",
    "<h3><center>Prediction/Estimation of Freight value for an Ordered product</center></h3>\n",
    "\n",
    "The aim of the analysis, is to predict the freight value of an ordered product on Olist, given the various characteristics invovled in the delivery of the product such as distance, weight and dimensions of the product.\n",
    "\n",
    "The data for this predictive analysis, is not directly available, hence many datasets are joined to get the necessary information into one dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 279,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from geopy import distance\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from ipywidgets import interact\n",
    "from scipy.stats import spearmanr, ttest_1samp, ttest_ind, shapiro, f_oneway\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.linear_model import LinearRegression, LassoCV, RidgeCV, ElasticNet\n",
    "from sklearn.model_selection import cross_validate\n",
    "from sklearn.ensemble import RandomForestRegressor, GradientBoostingRegressor\n",
    "from sklearn.svm import SVR\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "from sklearn.neighbors import KNeighborsRegressor\n",
    "from sklearn.neural_network import MLPRegressor\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.metrics import mean_absolute_error, mean_squared_error"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The customer_order dataset contains mapping of each order on Olist to the customer along with information such as price and freight value of the order. This dataset is used to retreive information of an order and the freight value."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 235,
   "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>order_id</th>\n",
       "      <th>order_purchase_timestamp</th>\n",
       "      <th>order_item_id</th>\n",
       "      <th>product_id</th>\n",
       "      <th>price</th>\n",
       "      <th>freight_value</th>\n",
       "      <th>customer_unique_id</th>\n",
       "      <th>product_category_name_english</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>e481f51cbdc54678b7cc49136f2d6af7</td>\n",
       "      <td>2017-10-02 10:56:33</td>\n",
       "      <td>1</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>29.99</td>\n",
       "      <td>8.72</td>\n",
       "      <td>7c396fd4830fd04220f754e42b4e5bff</td>\n",
       "      <td>housewares</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>128e10d95713541c87cd1a2e48201934</td>\n",
       "      <td>2017-08-15 18:29:31</td>\n",
       "      <td>1</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>29.99</td>\n",
       "      <td>7.78</td>\n",
       "      <td>3a51803cc0d012c3b5dc8b7528cb05f7</td>\n",
       "      <td>housewares</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0e7e841ddf8f8f2de2bad69267ecfbcf</td>\n",
       "      <td>2017-08-02 18:24:47</td>\n",
       "      <td>1</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>29.99</td>\n",
       "      <td>7.78</td>\n",
       "      <td>ef0996a1a279c26e7ecbd737be23d235</td>\n",
       "      <td>housewares</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>bfc39df4f36c3693ff3b63fcbea9e90a</td>\n",
       "      <td>2017-10-23 23:26:46</td>\n",
       "      <td>1</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>29.99</td>\n",
       "      <td>14.10</td>\n",
       "      <td>e781fdcc107d13d865fc7698711cc572</td>\n",
       "      <td>housewares</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>53cdb2fc8bc7dce0b6741e2150273451</td>\n",
       "      <td>2018-07-24 20:41:37</td>\n",
       "      <td>1</td>\n",
       "      <td>595fac2a385ac33a80bd5114aec74eb8</td>\n",
       "      <td>118.70</td>\n",
       "      <td>22.76</td>\n",
       "      <td>af07308b275d755c9edb36a90c618231</td>\n",
       "      <td>perfumery</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           order_id order_purchase_timestamp  order_item_id  \\\n",
       "0  e481f51cbdc54678b7cc49136f2d6af7      2017-10-02 10:56:33              1   \n",
       "1  128e10d95713541c87cd1a2e48201934      2017-08-15 18:29:31              1   \n",
       "2  0e7e841ddf8f8f2de2bad69267ecfbcf      2017-08-02 18:24:47              1   \n",
       "3  bfc39df4f36c3693ff3b63fcbea9e90a      2017-10-23 23:26:46              1   \n",
       "4  53cdb2fc8bc7dce0b6741e2150273451      2018-07-24 20:41:37              1   \n",
       "\n",
       "                         product_id   price  freight_value  \\\n",
       "0  87285b34884572647811a353c7ac498a   29.99           8.72   \n",
       "1  87285b34884572647811a353c7ac498a   29.99           7.78   \n",
       "2  87285b34884572647811a353c7ac498a   29.99           7.78   \n",
       "3  87285b34884572647811a353c7ac498a   29.99          14.10   \n",
       "4  595fac2a385ac33a80bd5114aec74eb8  118.70          22.76   \n",
       "\n",
       "                 customer_unique_id product_category_name_english  \n",
       "0  7c396fd4830fd04220f754e42b4e5bff                    housewares  \n",
       "1  3a51803cc0d012c3b5dc8b7528cb05f7                    housewares  \n",
       "2  ef0996a1a279c26e7ecbd737be23d235                    housewares  \n",
       "3  e781fdcc107d13d865fc7698711cc572                    housewares  \n",
       "4  af07308b275d755c9edb36a90c618231                     perfumery  "
      ]
     },
     "execution_count": 235,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "orders = pd.read_csv('/home/raj/Github/Olist-business-analysis/Joined data/customer_order.csv')\n",
    "orders.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The customer_data dataset contains the customers of Olist and their locations. These will help in the calculation of the distance later."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 236,
   "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>customer_id</th>\n",
       "      <th>customer_unique_id</th>\n",
       "      <th>customer_zip_code_prefix</th>\n",
       "      <th>customer_city</th>\n",
       "      <th>customer_state</th>\n",
       "      <th>geolocation_lat</th>\n",
       "      <th>geolocation_lng</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>06b8999e2fba1a1fbc88172c00ba8bc7</td>\n",
       "      <td>861eff4711a542e4b93843c6dd7febb0</td>\n",
       "      <td>14409</td>\n",
       "      <td>franca</td>\n",
       "      <td>SP</td>\n",
       "      <td>-20.498489</td>\n",
       "      <td>-47.396929</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>18955e83d337fd6b2def6b18a428ac77</td>\n",
       "      <td>290c77bc529b7ac935b93aa66c333dc3</td>\n",
       "      <td>9790</td>\n",
       "      <td>sao bernardo do campo</td>\n",
       "      <td>SP</td>\n",
       "      <td>-23.727992</td>\n",
       "      <td>-46.542848</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4e7b3e00288586ebd08712fdd0374a03</td>\n",
       "      <td>060e732b5b29e8181a18229c7b0b2b5e</td>\n",
       "      <td>1151</td>\n",
       "      <td>sao paulo</td>\n",
       "      <td>SP</td>\n",
       "      <td>-23.531642</td>\n",
       "      <td>-46.656289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>b2b6027bc5c5109e529d4dc6358b12c3</td>\n",
       "      <td>259dac757896d24d7702b9acbbff3f3c</td>\n",
       "      <td>8775</td>\n",
       "      <td>mogi das cruzes</td>\n",
       "      <td>SP</td>\n",
       "      <td>-23.499702</td>\n",
       "      <td>-46.185233</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4f2d8ab171c80ec8364f7c12e35b23ad</td>\n",
       "      <td>345ecd01c38d18a9036ed96c73b8d066</td>\n",
       "      <td>13056</td>\n",
       "      <td>campinas</td>\n",
       "      <td>SP</td>\n",
       "      <td>-22.975100</td>\n",
       "      <td>-47.142925</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                        customer_id                customer_unique_id  \\\n",
       "0  06b8999e2fba1a1fbc88172c00ba8bc7  861eff4711a542e4b93843c6dd7febb0   \n",
       "1  18955e83d337fd6b2def6b18a428ac77  290c77bc529b7ac935b93aa66c333dc3   \n",
       "2  4e7b3e00288586ebd08712fdd0374a03  060e732b5b29e8181a18229c7b0b2b5e   \n",
       "3  b2b6027bc5c5109e529d4dc6358b12c3  259dac757896d24d7702b9acbbff3f3c   \n",
       "4  4f2d8ab171c80ec8364f7c12e35b23ad  345ecd01c38d18a9036ed96c73b8d066   \n",
       "\n",
       "   customer_zip_code_prefix          customer_city customer_state  \\\n",
       "0                     14409                 franca             SP   \n",
       "1                      9790  sao bernardo do campo             SP   \n",
       "2                      1151              sao paulo             SP   \n",
       "3                      8775        mogi das cruzes             SP   \n",
       "4                     13056               campinas             SP   \n",
       "\n",
       "   geolocation_lat  geolocation_lng  \n",
       "0       -20.498489       -47.396929  \n",
       "1       -23.727992       -46.542848  \n",
       "2       -23.531642       -46.656289  \n",
       "3       -23.499702       -46.185233  \n",
       "4       -22.975100       -47.142925  "
      ]
     },
     "execution_count": 236,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "customers = pd.read_csv('/home/raj/Github/Olist-business-analysis/Joined data/customer_data.csv', index_col=0)\n",
    "customers.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The customers dataframe is grouped on customer_unique_id and aggregated on mode of latitude & longitude of the customer, so that unique values are obtained. One unique location for each customer. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "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>customer_unique_id</th>\n",
       "      <th>geolocation_lat</th>\n",
       "      <th>geolocation_lng</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0000366f3b9a7992bf8c76cfdf3221e2</td>\n",
       "      <td>-23.340235</td>\n",
       "      <td>-46.830140</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0000b849f77a49e4a4ce2b2a4ca5be3f</td>\n",
       "      <td>-23.559115</td>\n",
       "      <td>-46.787626</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0000f46a3911fa3c0805444483337064</td>\n",
       "      <td>-27.542880</td>\n",
       "      <td>-48.633426</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0000f6ccb0745a6a4b88665a16c9f078</td>\n",
       "      <td>-1.312214</td>\n",
       "      <td>-48.483159</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0004aac84e0df4da2b147fca70cf8255</td>\n",
       "      <td>-23.505548</td>\n",
       "      <td>-47.469705</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 customer_unique_id  geolocation_lat  geolocation_lng\n",
       "0  0000366f3b9a7992bf8c76cfdf3221e2       -23.340235       -46.830140\n",
       "1  0000b849f77a49e4a4ce2b2a4ca5be3f       -23.559115       -46.787626\n",
       "2  0000f46a3911fa3c0805444483337064       -27.542880       -48.633426\n",
       "3  0000f6ccb0745a6a4b88665a16c9f078        -1.312214       -48.483159\n",
       "4  0004aac84e0df4da2b147fca70cf8255       -23.505548       -47.469705"
      ]
     },
     "execution_count": 237,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cust_locs = customers[[\n",
    "    'customer_unique_id',\n",
    "    'geolocation_lat',\n",
    "    'geolocation_lng'\n",
    "]].groupby('customer_unique_id').mean().reset_index()\n",
    "cust_locs.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The two datasets above, are merged on customer_unique_id, to obtain a mapping of order to customer and the customer's location along with freight value."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 238,
   "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>order_id</th>\n",
       "      <th>customer_unique_id</th>\n",
       "      <th>cust_lat</th>\n",
       "      <th>cust_lng</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>e481f51cbdc54678b7cc49136f2d6af7</td>\n",
       "      <td>7c396fd4830fd04220f754e42b4e5bff</td>\n",
       "      <td>-23.576983</td>\n",
       "      <td>-46.587161</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>128e10d95713541c87cd1a2e48201934</td>\n",
       "      <td>3a51803cc0d012c3b5dc8b7528cb05f7</td>\n",
       "      <td>-23.564630</td>\n",
       "      <td>-46.534268</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0e7e841ddf8f8f2de2bad69267ecfbcf</td>\n",
       "      <td>ef0996a1a279c26e7ecbd737be23d235</td>\n",
       "      <td>-23.600462</td>\n",
       "      <td>-46.655318</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>bfc39df4f36c3693ff3b63fcbea9e90a</td>\n",
       "      <td>e781fdcc107d13d865fc7698711cc572</td>\n",
       "      <td>-27.528367</td>\n",
       "      <td>-48.491914</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>53cdb2fc8bc7dce0b6741e2150273451</td>\n",
       "      <td>af07308b275d755c9edb36a90c618231</td>\n",
       "      <td>-12.177924</td>\n",
       "      <td>-44.660711</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           order_id                customer_unique_id  \\\n",
       "0  e481f51cbdc54678b7cc49136f2d6af7  7c396fd4830fd04220f754e42b4e5bff   \n",
       "1  128e10d95713541c87cd1a2e48201934  3a51803cc0d012c3b5dc8b7528cb05f7   \n",
       "2  0e7e841ddf8f8f2de2bad69267ecfbcf  ef0996a1a279c26e7ecbd737be23d235   \n",
       "3  bfc39df4f36c3693ff3b63fcbea9e90a  e781fdcc107d13d865fc7698711cc572   \n",
       "4  53cdb2fc8bc7dce0b6741e2150273451  af07308b275d755c9edb36a90c618231   \n",
       "\n",
       "    cust_lat   cust_lng  \n",
       "0 -23.576983 -46.587161  \n",
       "1 -23.564630 -46.534268  \n",
       "2 -23.600462 -46.655318  \n",
       "3 -27.528367 -48.491914  \n",
       "4 -12.177924 -44.660711  "
      ]
     },
     "execution_count": 238,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "order_customer_loc = pd.merge(\n",
    "    left= orders,\n",
    "    right= cust_locs,\n",
    "    on= 'customer_unique_id',\n",
    "    how= 'left'\n",
    ").rename(columns= {'geolocation_lat':'cust_lat', 'geolocation_lng':'cust_lng'})[[\n",
    "    'order_id',\n",
    "    'customer_unique_id',\n",
    "    'cust_lat',\n",
    "    'cust_lng'\n",
    "]]\n",
    "order_customer_loc.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The olist_order_items_dataset contains a mapping of order to the product and the seller fulfilling it. The seller information can be obtained from this dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 239,
   "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>order_id</th>\n",
       "      <th>order_item_id</th>\n",
       "      <th>product_id</th>\n",
       "      <th>seller_id</th>\n",
       "      <th>shipping_limit_date</th>\n",
       "      <th>price</th>\n",
       "      <th>freight_value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>00010242fe8c5a6d1ba2dd792cb16214</td>\n",
       "      <td>1</td>\n",
       "      <td>4244733e06e7ecb4970a6e2683c13e61</td>\n",
       "      <td>48436dade18ac8b2bce089ec2a041202</td>\n",
       "      <td>2017-09-19 09:45:35</td>\n",
       "      <td>58.90</td>\n",
       "      <td>13.29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>00018f77f2f0320c557190d7a144bdd3</td>\n",
       "      <td>1</td>\n",
       "      <td>e5f2d52b802189ee658865ca93d83a8f</td>\n",
       "      <td>dd7ddc04e1b6c2c614352b383efe2d36</td>\n",
       "      <td>2017-05-03 11:05:13</td>\n",
       "      <td>239.90</td>\n",
       "      <td>19.93</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>000229ec398224ef6ca0657da4fc703e</td>\n",
       "      <td>1</td>\n",
       "      <td>c777355d18b72b67abbeef9df44fd0fd</td>\n",
       "      <td>5b51032eddd242adc84c38acab88f23d</td>\n",
       "      <td>2018-01-18 14:48:30</td>\n",
       "      <td>199.00</td>\n",
       "      <td>17.87</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>00024acbcdf0a6daa1e931b038114c75</td>\n",
       "      <td>1</td>\n",
       "      <td>7634da152a4610f1595efa32f14722fc</td>\n",
       "      <td>9d7a1d34a5052409006425275ba1c2b4</td>\n",
       "      <td>2018-08-15 10:10:18</td>\n",
       "      <td>12.99</td>\n",
       "      <td>12.79</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>00042b26cf59d7ce69dfabb4e55b4fd9</td>\n",
       "      <td>1</td>\n",
       "      <td>ac6c3623068f30de03045865e4e10089</td>\n",
       "      <td>df560393f3a51e74553ab94004ba5c87</td>\n",
       "      <td>2017-02-13 13:57:51</td>\n",
       "      <td>199.90</td>\n",
       "      <td>18.14</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           order_id  order_item_id  \\\n",
       "0  00010242fe8c5a6d1ba2dd792cb16214              1   \n",
       "1  00018f77f2f0320c557190d7a144bdd3              1   \n",
       "2  000229ec398224ef6ca0657da4fc703e              1   \n",
       "3  00024acbcdf0a6daa1e931b038114c75              1   \n",
       "4  00042b26cf59d7ce69dfabb4e55b4fd9              1   \n",
       "\n",
       "                         product_id                         seller_id  \\\n",
       "0  4244733e06e7ecb4970a6e2683c13e61  48436dade18ac8b2bce089ec2a041202   \n",
       "1  e5f2d52b802189ee658865ca93d83a8f  dd7ddc04e1b6c2c614352b383efe2d36   \n",
       "2  c777355d18b72b67abbeef9df44fd0fd  5b51032eddd242adc84c38acab88f23d   \n",
       "3  7634da152a4610f1595efa32f14722fc  9d7a1d34a5052409006425275ba1c2b4   \n",
       "4  ac6c3623068f30de03045865e4e10089  df560393f3a51e74553ab94004ba5c87   \n",
       "\n",
       "   shipping_limit_date   price  freight_value  \n",
       "0  2017-09-19 09:45:35   58.90          13.29  \n",
       "1  2017-05-03 11:05:13  239.90          19.93  \n",
       "2  2018-01-18 14:48:30  199.00          17.87  \n",
       "3  2018-08-15 10:10:18   12.99          12.79  \n",
       "4  2017-02-13 13:57:51  199.90          18.14  "
      ]
     },
     "execution_count": 239,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "order_items = pd.read_csv('/home/raj/Github/Olist-business-analysis/Original data/olist_order_items_dataset.csv')\n",
    "order_items.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The delivery_data dataset contains data on seller and seller locations, for every customers who has ever ordered on Olist. The seller locations are required for the calculation of distance the product has to be shipped."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 240,
   "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>order_purchase_timestamp</th>\n",
       "      <th>order_delivered_customer_date</th>\n",
       "      <th>order_estimated_delivery_date</th>\n",
       "      <th>seller_id</th>\n",
       "      <th>customer_unique_id</th>\n",
       "      <th>customer_lat</th>\n",
       "      <th>customer_lng</th>\n",
       "      <th>customer_city</th>\n",
       "      <th>seller_city</th>\n",
       "      <th>seller_lat</th>\n",
       "      <th>seller_lng</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>13-09-2017 08:59</td>\n",
       "      <td>20-09-2017 23:43</td>\n",
       "      <td>29-09-2017 00:00</td>\n",
       "      <td>48436dade18ac8b2bce089ec2a041202</td>\n",
       "      <td>871766c5855e863f6eccc05f988b23cb</td>\n",
       "      <td>-21.762775</td>\n",
       "      <td>-41.309633</td>\n",
       "      <td>campos dos goytacazes</td>\n",
       "      <td>volta redonda</td>\n",
       "      <td>-22.496953</td>\n",
       "      <td>-44.127492</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>26-04-2017 10:53</td>\n",
       "      <td>12-05-2017 16:04</td>\n",
       "      <td>15-05-2017 00:00</td>\n",
       "      <td>dd7ddc04e1b6c2c614352b383efe2d36</td>\n",
       "      <td>eb28e67c4c0b83846050ddfb8a35d051</td>\n",
       "      <td>-20.220527</td>\n",
       "      <td>-50.903424</td>\n",
       "      <td>santa fe do sul</td>\n",
       "      <td>sao paulo</td>\n",
       "      <td>-23.565096</td>\n",
       "      <td>-46.518565</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>14-01-2018 14:33</td>\n",
       "      <td>22-01-2018 13:19</td>\n",
       "      <td>05-02-2018 00:00</td>\n",
       "      <td>5b51032eddd242adc84c38acab88f23d</td>\n",
       "      <td>3818d81c6709e39d06b2738a8d3a2474</td>\n",
       "      <td>-19.870305</td>\n",
       "      <td>-44.593326</td>\n",
       "      <td>para de minas</td>\n",
       "      <td>borda da mata</td>\n",
       "      <td>-22.262584</td>\n",
       "      <td>-46.171124</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>08-08-2018 10:00</td>\n",
       "      <td>14-08-2018 13:32</td>\n",
       "      <td>20-08-2018 00:00</td>\n",
       "      <td>9d7a1d34a5052409006425275ba1c2b4</td>\n",
       "      <td>af861d436cfc08b2c2ddefd0ba074622</td>\n",
       "      <td>-23.089925</td>\n",
       "      <td>-46.611654</td>\n",
       "      <td>atibaia</td>\n",
       "      <td>franca</td>\n",
       "      <td>-20.553624</td>\n",
       "      <td>-47.387359</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>04-02-2017 13:57</td>\n",
       "      <td>01-03-2017 16:42</td>\n",
       "      <td>17-03-2017 00:00</td>\n",
       "      <td>df560393f3a51e74553ab94004ba5c87</td>\n",
       "      <td>64b576fb70d441e8f1b2d7d446e483c5</td>\n",
       "      <td>-23.243402</td>\n",
       "      <td>-46.827614</td>\n",
       "      <td>varzea paulista</td>\n",
       "      <td>loanda</td>\n",
       "      <td>-22.929384</td>\n",
       "      <td>-53.135873</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  order_purchase_timestamp order_delivered_customer_date  \\\n",
       "0         13-09-2017 08:59              20-09-2017 23:43   \n",
       "1         26-04-2017 10:53              12-05-2017 16:04   \n",
       "2         14-01-2018 14:33              22-01-2018 13:19   \n",
       "3         08-08-2018 10:00              14-08-2018 13:32   \n",
       "4         04-02-2017 13:57              01-03-2017 16:42   \n",
       "\n",
       "  order_estimated_delivery_date                         seller_id  \\\n",
       "0              29-09-2017 00:00  48436dade18ac8b2bce089ec2a041202   \n",
       "1              15-05-2017 00:00  dd7ddc04e1b6c2c614352b383efe2d36   \n",
       "2              05-02-2018 00:00  5b51032eddd242adc84c38acab88f23d   \n",
       "3              20-08-2018 00:00  9d7a1d34a5052409006425275ba1c2b4   \n",
       "4              17-03-2017 00:00  df560393f3a51e74553ab94004ba5c87   \n",
       "\n",
       "                 customer_unique_id  customer_lat  customer_lng  \\\n",
       "0  871766c5855e863f6eccc05f988b23cb    -21.762775    -41.309633   \n",
       "1  eb28e67c4c0b83846050ddfb8a35d051    -20.220527    -50.903424   \n",
       "2  3818d81c6709e39d06b2738a8d3a2474    -19.870305    -44.593326   \n",
       "3  af861d436cfc08b2c2ddefd0ba074622    -23.089925    -46.611654   \n",
       "4  64b576fb70d441e8f1b2d7d446e483c5    -23.243402    -46.827614   \n",
       "\n",
       "           customer_city    seller_city  seller_lat  seller_lng  \n",
       "0  campos dos goytacazes  volta redonda  -22.496953  -44.127492  \n",
       "1        santa fe do sul      sao paulo  -23.565096  -46.518565  \n",
       "2          para de minas  borda da mata  -22.262584  -46.171124  \n",
       "3                atibaia         franca  -20.553624  -47.387359  \n",
       "4        varzea paulista         loanda  -22.929384  -53.135873  "
      ]
     },
     "execution_count": 240,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sellers = pd.read_csv('/home/raj/Github/Olist-business-analysis/Joined data/delivery_data.csv')\n",
    "sellers.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Similar to customer, the dataset is grouped on seller_id and aggregated on latitude & longitude, to obtain unique location for each seller."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 241,
   "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>seller_id</th>\n",
       "      <th>seller_lat</th>\n",
       "      <th>seller_lng</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0015a82c2db000af6aaaf3ae2ecb0532</td>\n",
       "      <td>-23.640444</td>\n",
       "      <td>-46.541742</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>001cca7ae9ae17fb1caed9dfb1094831</td>\n",
       "      <td>-20.278513</td>\n",
       "      <td>-40.411675</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>001e6ad469a905060d959994f1b41e4f</td>\n",
       "      <td>-22.872355</td>\n",
       "      <td>-43.027433</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>002100f778ceb8431b7a1020ff7ab48f</td>\n",
       "      <td>-20.528759</td>\n",
       "      <td>-47.411110</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>003554e2dce176b5555353e4f3555ac8</td>\n",
       "      <td>-16.640574</td>\n",
       "      <td>-49.276483</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          seller_id  seller_lat  seller_lng\n",
       "0  0015a82c2db000af6aaaf3ae2ecb0532  -23.640444  -46.541742\n",
       "1  001cca7ae9ae17fb1caed9dfb1094831  -20.278513  -40.411675\n",
       "2  001e6ad469a905060d959994f1b41e4f  -22.872355  -43.027433\n",
       "3  002100f778ceb8431b7a1020ff7ab48f  -20.528759  -47.411110\n",
       "4  003554e2dce176b5555353e4f3555ac8  -16.640574  -49.276483"
      ]
     },
     "execution_count": 241,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sell_loc = sellers[[\n",
    "    'seller_id',\n",
    "    'seller_lat',\n",
    "    'seller_lng'\n",
    "]].groupby('seller_id').mean().reset_index()\n",
    "sell_loc.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The two are now merged to get a mapping of products to seller locations."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 242,
   "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>order_id</th>\n",
       "      <th>product_id</th>\n",
       "      <th>seller_id</th>\n",
       "      <th>freight_value</th>\n",
       "      <th>seller_lat</th>\n",
       "      <th>seller_lng</th>\n",
       "      <th>price</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>00010242fe8c5a6d1ba2dd792cb16214</td>\n",
       "      <td>4244733e06e7ecb4970a6e2683c13e61</td>\n",
       "      <td>48436dade18ac8b2bce089ec2a041202</td>\n",
       "      <td>13.29</td>\n",
       "      <td>-22.496953</td>\n",
       "      <td>-44.127492</td>\n",
       "      <td>58.90</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>00018f77f2f0320c557190d7a144bdd3</td>\n",
       "      <td>e5f2d52b802189ee658865ca93d83a8f</td>\n",
       "      <td>dd7ddc04e1b6c2c614352b383efe2d36</td>\n",
       "      <td>19.93</td>\n",
       "      <td>-23.565096</td>\n",
       "      <td>-46.518565</td>\n",
       "      <td>239.90</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>000229ec398224ef6ca0657da4fc703e</td>\n",
       "      <td>c777355d18b72b67abbeef9df44fd0fd</td>\n",
       "      <td>5b51032eddd242adc84c38acab88f23d</td>\n",
       "      <td>17.87</td>\n",
       "      <td>-22.262584</td>\n",
       "      <td>-46.171124</td>\n",
       "      <td>199.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>00024acbcdf0a6daa1e931b038114c75</td>\n",
       "      <td>7634da152a4610f1595efa32f14722fc</td>\n",
       "      <td>9d7a1d34a5052409006425275ba1c2b4</td>\n",
       "      <td>12.79</td>\n",
       "      <td>-20.553624</td>\n",
       "      <td>-47.387359</td>\n",
       "      <td>12.99</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>00042b26cf59d7ce69dfabb4e55b4fd9</td>\n",
       "      <td>ac6c3623068f30de03045865e4e10089</td>\n",
       "      <td>df560393f3a51e74553ab94004ba5c87</td>\n",
       "      <td>18.14</td>\n",
       "      <td>-22.929384</td>\n",
       "      <td>-53.135873</td>\n",
       "      <td>199.90</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           order_id                        product_id  \\\n",
       "0  00010242fe8c5a6d1ba2dd792cb16214  4244733e06e7ecb4970a6e2683c13e61   \n",
       "1  00018f77f2f0320c557190d7a144bdd3  e5f2d52b802189ee658865ca93d83a8f   \n",
       "2  000229ec398224ef6ca0657da4fc703e  c777355d18b72b67abbeef9df44fd0fd   \n",
       "3  00024acbcdf0a6daa1e931b038114c75  7634da152a4610f1595efa32f14722fc   \n",
       "4  00042b26cf59d7ce69dfabb4e55b4fd9  ac6c3623068f30de03045865e4e10089   \n",
       "\n",
       "                          seller_id  freight_value  seller_lat  seller_lng  \\\n",
       "0  48436dade18ac8b2bce089ec2a041202          13.29  -22.496953  -44.127492   \n",
       "1  dd7ddc04e1b6c2c614352b383efe2d36          19.93  -23.565096  -46.518565   \n",
       "2  5b51032eddd242adc84c38acab88f23d          17.87  -22.262584  -46.171124   \n",
       "3  9d7a1d34a5052409006425275ba1c2b4          12.79  -20.553624  -47.387359   \n",
       "4  df560393f3a51e74553ab94004ba5c87          18.14  -22.929384  -53.135873   \n",
       "\n",
       "    price  \n",
       "0   58.90  \n",
       "1  239.90  \n",
       "2  199.00  \n",
       "3   12.99  \n",
       "4  199.90  "
      ]
     },
     "execution_count": 242,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sell_order_loc = pd.merge(\n",
    "    left= order_items,\n",
    "    right= sell_loc,\n",
    "    on= 'seller_id',\n",
    "    how= 'left'\n",
    ")[[\n",
    "    'order_id',\n",
    "    'product_id',\n",
    "    'seller_id',\n",
    "    'freight_value',\n",
    "    'seller_lat',\n",
    "    'seller_lng',\n",
    "    'price'\n",
    "]]\n",
    "sell_order_loc.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The two joins created earlier are now merged to get a dataset which contains a mapping for customer location to seller location, the product and its freight value."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "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>order_id</th>\n",
       "      <th>customer_unique_id</th>\n",
       "      <th>cust_lat</th>\n",
       "      <th>cust_lng</th>\n",
       "      <th>product_id</th>\n",
       "      <th>seller_id</th>\n",
       "      <th>freight_value</th>\n",
       "      <th>seller_lat</th>\n",
       "      <th>seller_lng</th>\n",
       "      <th>price</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>e481f51cbdc54678b7cc49136f2d6af7</td>\n",
       "      <td>7c396fd4830fd04220f754e42b4e5bff</td>\n",
       "      <td>-23.576983</td>\n",
       "      <td>-46.587161</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>8.72</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>29.99</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>128e10d95713541c87cd1a2e48201934</td>\n",
       "      <td>3a51803cc0d012c3b5dc8b7528cb05f7</td>\n",
       "      <td>-23.564630</td>\n",
       "      <td>-46.534268</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>7.78</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>29.99</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0e7e841ddf8f8f2de2bad69267ecfbcf</td>\n",
       "      <td>ef0996a1a279c26e7ecbd737be23d235</td>\n",
       "      <td>-23.600462</td>\n",
       "      <td>-46.655318</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>7.78</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>29.99</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>bfc39df4f36c3693ff3b63fcbea9e90a</td>\n",
       "      <td>e781fdcc107d13d865fc7698711cc572</td>\n",
       "      <td>-27.528367</td>\n",
       "      <td>-48.491914</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>14.10</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>29.99</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>53cdb2fc8bc7dce0b6741e2150273451</td>\n",
       "      <td>af07308b275d755c9edb36a90c618231</td>\n",
       "      <td>-12.177924</td>\n",
       "      <td>-44.660711</td>\n",
       "      <td>595fac2a385ac33a80bd5114aec74eb8</td>\n",
       "      <td>289cdb325fb7e7f891c38608bf9e0962</td>\n",
       "      <td>22.76</td>\n",
       "      <td>-19.807681</td>\n",
       "      <td>-43.980427</td>\n",
       "      <td>118.70</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           order_id                customer_unique_id  \\\n",
       "0  e481f51cbdc54678b7cc49136f2d6af7  7c396fd4830fd04220f754e42b4e5bff   \n",
       "1  128e10d95713541c87cd1a2e48201934  3a51803cc0d012c3b5dc8b7528cb05f7   \n",
       "2  0e7e841ddf8f8f2de2bad69267ecfbcf  ef0996a1a279c26e7ecbd737be23d235   \n",
       "3  bfc39df4f36c3693ff3b63fcbea9e90a  e781fdcc107d13d865fc7698711cc572   \n",
       "4  53cdb2fc8bc7dce0b6741e2150273451  af07308b275d755c9edb36a90c618231   \n",
       "\n",
       "    cust_lat   cust_lng                        product_id  \\\n",
       "0 -23.576983 -46.587161  87285b34884572647811a353c7ac498a   \n",
       "1 -23.564630 -46.534268  87285b34884572647811a353c7ac498a   \n",
       "2 -23.600462 -46.655318  87285b34884572647811a353c7ac498a   \n",
       "3 -27.528367 -48.491914  87285b34884572647811a353c7ac498a   \n",
       "4 -12.177924 -44.660711  595fac2a385ac33a80bd5114aec74eb8   \n",
       "\n",
       "                          seller_id  freight_value  seller_lat  seller_lng  \\\n",
       "0  3504c0cb71d7fa48d967e0e4c94d59d9           8.72  -23.680729  -46.444238   \n",
       "1  3504c0cb71d7fa48d967e0e4c94d59d9           7.78  -23.680729  -46.444238   \n",
       "2  3504c0cb71d7fa48d967e0e4c94d59d9           7.78  -23.680729  -46.444238   \n",
       "3  3504c0cb71d7fa48d967e0e4c94d59d9          14.10  -23.680729  -46.444238   \n",
       "4  289cdb325fb7e7f891c38608bf9e0962          22.76  -19.807681  -43.980427   \n",
       "\n",
       "    price  \n",
       "0   29.99  \n",
       "1   29.99  \n",
       "2   29.99  \n",
       "3   29.99  \n",
       "4  118.70  "
      ]
     },
     "execution_count": 243,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "order_cust_sel_loc = pd.merge(\n",
    "    left= order_customer_loc,\n",
    "    right= sell_order_loc,\n",
    "    on= 'order_id',\n",
    "    how= 'inner'\n",
    ").drop_duplicates()\n",
    "order_cust_sel_loc.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The olist_products_dataset contains information of each products, its dimensions, weight and other such characteristics. The product informations required will be obtained from this dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 244,
   "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>product_id</th>\n",
       "      <th>product_category_name</th>\n",
       "      <th>product_name_lenght</th>\n",
       "      <th>product_description_lenght</th>\n",
       "      <th>product_photos_qty</th>\n",
       "      <th>product_weight_g</th>\n",
       "      <th>product_length_cm</th>\n",
       "      <th>product_height_cm</th>\n",
       "      <th>product_width_cm</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1e9e8ef04dbcff4541ed26657ea517e5</td>\n",
       "      <td>perfumaria</td>\n",
       "      <td>40.0</td>\n",
       "      <td>287.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>225.0</td>\n",
       "      <td>16.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>14.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3aa071139cb16b67ca9e5dea641aaa2f</td>\n",
       "      <td>artes</td>\n",
       "      <td>44.0</td>\n",
       "      <td>276.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1000.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>20.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>96bd76ec8810374ed1b65e291975717f</td>\n",
       "      <td>esporte_lazer</td>\n",
       "      <td>46.0</td>\n",
       "      <td>250.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>154.0</td>\n",
       "      <td>18.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>15.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>cef67bcfe19066a932b7673e239eb23d</td>\n",
       "      <td>bebes</td>\n",
       "      <td>27.0</td>\n",
       "      <td>261.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>371.0</td>\n",
       "      <td>26.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>26.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9dc1a7de274444849c219cff195d0b71</td>\n",
       "      <td>utilidades_domesticas</td>\n",
       "      <td>37.0</td>\n",
       "      <td>402.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>625.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>17.0</td>\n",
       "      <td>13.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         product_id  product_category_name  \\\n",
       "0  1e9e8ef04dbcff4541ed26657ea517e5             perfumaria   \n",
       "1  3aa071139cb16b67ca9e5dea641aaa2f                  artes   \n",
       "2  96bd76ec8810374ed1b65e291975717f          esporte_lazer   \n",
       "3  cef67bcfe19066a932b7673e239eb23d                  bebes   \n",
       "4  9dc1a7de274444849c219cff195d0b71  utilidades_domesticas   \n",
       "\n",
       "   product_name_lenght  product_description_lenght  product_photos_qty  \\\n",
       "0                 40.0                       287.0                 1.0   \n",
       "1                 44.0                       276.0                 1.0   \n",
       "2                 46.0                       250.0                 1.0   \n",
       "3                 27.0                       261.0                 1.0   \n",
       "4                 37.0                       402.0                 4.0   \n",
       "\n",
       "   product_weight_g  product_length_cm  product_height_cm  product_width_cm  \n",
       "0             225.0               16.0               10.0              14.0  \n",
       "1            1000.0               30.0               18.0              20.0  \n",
       "2             154.0               18.0                9.0              15.0  \n",
       "3             371.0               26.0                4.0              26.0  \n",
       "4             625.0               20.0               17.0              13.0  "
      ]
     },
     "execution_count": 244,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "products = pd.read_csv('/home/raj/Github/Olist-business-analysis/Original data/olist_products_dataset.csv')\n",
    "products.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The products dataset above and the join created previously, are now merged on product id to get the final dataframe required for predictive analysis. The required columns such as location coordinates, dimensions, weight are retrieved from the joined dataset, along with the target freight value."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 245,
   "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>order_id</th>\n",
       "      <th>customer_unique_id</th>\n",
       "      <th>product_id</th>\n",
       "      <th>seller_id</th>\n",
       "      <th>cust_lat</th>\n",
       "      <th>cust_lng</th>\n",
       "      <th>seller_lat</th>\n",
       "      <th>seller_lng</th>\n",
       "      <th>product_weight_g</th>\n",
       "      <th>product_length_cm</th>\n",
       "      <th>product_height_cm</th>\n",
       "      <th>product_width_cm</th>\n",
       "      <th>price</th>\n",
       "      <th>freight_value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>e481f51cbdc54678b7cc49136f2d6af7</td>\n",
       "      <td>7c396fd4830fd04220f754e42b4e5bff</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>-23.576983</td>\n",
       "      <td>-46.587161</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>500.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>29.99</td>\n",
       "      <td>8.72</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>128e10d95713541c87cd1a2e48201934</td>\n",
       "      <td>3a51803cc0d012c3b5dc8b7528cb05f7</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>-23.564630</td>\n",
       "      <td>-46.534268</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>500.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>29.99</td>\n",
       "      <td>7.78</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0e7e841ddf8f8f2de2bad69267ecfbcf</td>\n",
       "      <td>ef0996a1a279c26e7ecbd737be23d235</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>-23.600462</td>\n",
       "      <td>-46.655318</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>500.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>29.99</td>\n",
       "      <td>7.78</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>bfc39df4f36c3693ff3b63fcbea9e90a</td>\n",
       "      <td>e781fdcc107d13d865fc7698711cc572</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>-27.528367</td>\n",
       "      <td>-48.491914</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>500.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>29.99</td>\n",
       "      <td>14.10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>53cdb2fc8bc7dce0b6741e2150273451</td>\n",
       "      <td>af07308b275d755c9edb36a90c618231</td>\n",
       "      <td>595fac2a385ac33a80bd5114aec74eb8</td>\n",
       "      <td>289cdb325fb7e7f891c38608bf9e0962</td>\n",
       "      <td>-12.177924</td>\n",
       "      <td>-44.660711</td>\n",
       "      <td>-19.807681</td>\n",
       "      <td>-43.980427</td>\n",
       "      <td>400.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>118.70</td>\n",
       "      <td>22.76</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           order_id                customer_unique_id  \\\n",
       "0  e481f51cbdc54678b7cc49136f2d6af7  7c396fd4830fd04220f754e42b4e5bff   \n",
       "1  128e10d95713541c87cd1a2e48201934  3a51803cc0d012c3b5dc8b7528cb05f7   \n",
       "2  0e7e841ddf8f8f2de2bad69267ecfbcf  ef0996a1a279c26e7ecbd737be23d235   \n",
       "3  bfc39df4f36c3693ff3b63fcbea9e90a  e781fdcc107d13d865fc7698711cc572   \n",
       "4  53cdb2fc8bc7dce0b6741e2150273451  af07308b275d755c9edb36a90c618231   \n",
       "\n",
       "                         product_id                         seller_id  \\\n",
       "0  87285b34884572647811a353c7ac498a  3504c0cb71d7fa48d967e0e4c94d59d9   \n",
       "1  87285b34884572647811a353c7ac498a  3504c0cb71d7fa48d967e0e4c94d59d9   \n",
       "2  87285b34884572647811a353c7ac498a  3504c0cb71d7fa48d967e0e4c94d59d9   \n",
       "3  87285b34884572647811a353c7ac498a  3504c0cb71d7fa48d967e0e4c94d59d9   \n",
       "4  595fac2a385ac33a80bd5114aec74eb8  289cdb325fb7e7f891c38608bf9e0962   \n",
       "\n",
       "    cust_lat   cust_lng  seller_lat  seller_lng  product_weight_g  \\\n",
       "0 -23.576983 -46.587161  -23.680729  -46.444238             500.0   \n",
       "1 -23.564630 -46.534268  -23.680729  -46.444238             500.0   \n",
       "2 -23.600462 -46.655318  -23.680729  -46.444238             500.0   \n",
       "3 -27.528367 -48.491914  -23.680729  -46.444238             500.0   \n",
       "4 -12.177924 -44.660711  -19.807681  -43.980427             400.0   \n",
       "\n",
       "   product_length_cm  product_height_cm  product_width_cm   price  \\\n",
       "0               19.0                8.0              13.0   29.99   \n",
       "1               19.0                8.0              13.0   29.99   \n",
       "2               19.0                8.0              13.0   29.99   \n",
       "3               19.0                8.0              13.0   29.99   \n",
       "4               19.0               13.0              19.0  118.70   \n",
       "\n",
       "   freight_value  \n",
       "0           8.72  \n",
       "1           7.78  \n",
       "2           7.78  \n",
       "3          14.10  \n",
       "4          22.76  "
      ]
     },
     "execution_count": 245,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.merge(\n",
    "    left= order_cust_sel_loc,\n",
    "    right= products,\n",
    "    on= 'product_id',\n",
    "    how= 'left'\n",
    ")[[\n",
    "    'order_id',\n",
    "    'customer_unique_id',\n",
    "    'product_id',\n",
    "    'seller_id',\n",
    "    'cust_lat',\n",
    "    'cust_lng',\n",
    "    'seller_lat',\n",
    "    'seller_lng',\n",
    "    'product_weight_g',\n",
    "    'product_length_cm',\n",
    "    'product_height_cm',\n",
    "    'product_width_cm',\n",
    "    'price',\n",
    "    'freight_value'\n",
    "]]\n",
    "df.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using the customer and seller locations coordinates, the distance is calculated in kms. The distance is crucial as freight values depend on distance. A shipment coming from further away will cost more as compared to a shipment from near by.<br>\n",
    "\n",
    "The distance is calculated with the `geopy.distance.geodesic` function. The model of the earth assumed for this is the most reliable and accurate one used globally - WGS-84"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 246,
   "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>order_id</th>\n",
       "      <th>customer_unique_id</th>\n",
       "      <th>product_id</th>\n",
       "      <th>seller_id</th>\n",
       "      <th>cust_lat</th>\n",
       "      <th>cust_lng</th>\n",
       "      <th>seller_lat</th>\n",
       "      <th>seller_lng</th>\n",
       "      <th>product_weight_g</th>\n",
       "      <th>product_length_cm</th>\n",
       "      <th>product_height_cm</th>\n",
       "      <th>product_width_cm</th>\n",
       "      <th>price</th>\n",
       "      <th>freight_value</th>\n",
       "      <th>distance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>e481f51cbdc54678b7cc49136f2d6af7</td>\n",
       "      <td>7c396fd4830fd04220f754e42b4e5bff</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>-23.576983</td>\n",
       "      <td>-46.587161</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>500.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>29.99</td>\n",
       "      <td>8.72</td>\n",
       "      <td>18.566632</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>128e10d95713541c87cd1a2e48201934</td>\n",
       "      <td>3a51803cc0d012c3b5dc8b7528cb05f7</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>-23.564630</td>\n",
       "      <td>-46.534268</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>500.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>29.99</td>\n",
       "      <td>7.78</td>\n",
       "      <td>15.803204</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0e7e841ddf8f8f2de2bad69267ecfbcf</td>\n",
       "      <td>ef0996a1a279c26e7ecbd737be23d235</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>-23.600462</td>\n",
       "      <td>-46.655318</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>500.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>29.99</td>\n",
       "      <td>7.78</td>\n",
       "      <td>23.299562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>bfc39df4f36c3693ff3b63fcbea9e90a</td>\n",
       "      <td>e781fdcc107d13d865fc7698711cc572</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>-27.528367</td>\n",
       "      <td>-48.491914</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>500.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>29.99</td>\n",
       "      <td>14.10</td>\n",
       "      <td>473.254397</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>53cdb2fc8bc7dce0b6741e2150273451</td>\n",
       "      <td>af07308b275d755c9edb36a90c618231</td>\n",
       "      <td>595fac2a385ac33a80bd5114aec74eb8</td>\n",
       "      <td>289cdb325fb7e7f891c38608bf9e0962</td>\n",
       "      <td>-12.177924</td>\n",
       "      <td>-44.660711</td>\n",
       "      <td>-19.807681</td>\n",
       "      <td>-43.980427</td>\n",
       "      <td>400.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>118.70</td>\n",
       "      <td>22.76</td>\n",
       "      <td>847.437333</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           order_id                customer_unique_id  \\\n",
       "0  e481f51cbdc54678b7cc49136f2d6af7  7c396fd4830fd04220f754e42b4e5bff   \n",
       "1  128e10d95713541c87cd1a2e48201934  3a51803cc0d012c3b5dc8b7528cb05f7   \n",
       "2  0e7e841ddf8f8f2de2bad69267ecfbcf  ef0996a1a279c26e7ecbd737be23d235   \n",
       "3  bfc39df4f36c3693ff3b63fcbea9e90a  e781fdcc107d13d865fc7698711cc572   \n",
       "4  53cdb2fc8bc7dce0b6741e2150273451  af07308b275d755c9edb36a90c618231   \n",
       "\n",
       "                         product_id                         seller_id  \\\n",
       "0  87285b34884572647811a353c7ac498a  3504c0cb71d7fa48d967e0e4c94d59d9   \n",
       "1  87285b34884572647811a353c7ac498a  3504c0cb71d7fa48d967e0e4c94d59d9   \n",
       "2  87285b34884572647811a353c7ac498a  3504c0cb71d7fa48d967e0e4c94d59d9   \n",
       "3  87285b34884572647811a353c7ac498a  3504c0cb71d7fa48d967e0e4c94d59d9   \n",
       "4  595fac2a385ac33a80bd5114aec74eb8  289cdb325fb7e7f891c38608bf9e0962   \n",
       "\n",
       "    cust_lat   cust_lng  seller_lat  seller_lng  product_weight_g  \\\n",
       "0 -23.576983 -46.587161  -23.680729  -46.444238             500.0   \n",
       "1 -23.564630 -46.534268  -23.680729  -46.444238             500.0   \n",
       "2 -23.600462 -46.655318  -23.680729  -46.444238             500.0   \n",
       "3 -27.528367 -48.491914  -23.680729  -46.444238             500.0   \n",
       "4 -12.177924 -44.660711  -19.807681  -43.980427             400.0   \n",
       "\n",
       "   product_length_cm  product_height_cm  product_width_cm   price  \\\n",
       "0               19.0                8.0              13.0   29.99   \n",
       "1               19.0                8.0              13.0   29.99   \n",
       "2               19.0                8.0              13.0   29.99   \n",
       "3               19.0                8.0              13.0   29.99   \n",
       "4               19.0               13.0              19.0  118.70   \n",
       "\n",
       "   freight_value    distance  \n",
       "0           8.72   18.566632  \n",
       "1           7.78   15.803204  \n",
       "2           7.78   23.299562  \n",
       "3          14.10  473.254397  \n",
       "4          22.76  847.437333  "
      ]
     },
     "execution_count": 246,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def calc_dist(cust,sel):\n",
    "    return distance.geodesic(cust,sel).km\n",
    "\n",
    "df.dropna(subset= [\n",
    "    'cust_lat',\n",
    "    'cust_lng',\n",
    "    'seller_lat',\n",
    "    'seller_lng'\n",
    "], inplace= True)\n",
    "\n",
    "df['distance'] = df[[\n",
    "    'cust_lat',\n",
    "    'cust_lng',\n",
    "    'seller_lat',\n",
    "    'seller_lng'\n",
    "]].apply(lambda row: calc_dist((row.cust_lat,row.cust_lng),(row.seller_lat,row.seller_lng)), axis=1)\n",
    "\n",
    "df.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For each product, the dimensions are given, length, height, width. These can be used to create another feature, which is the volume in cm<sup>3</sup> of the product. The volume is as important as the weight of the product. A light product can have larger volume and hence increased freight value."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "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>order_id</th>\n",
       "      <th>customer_unique_id</th>\n",
       "      <th>product_id</th>\n",
       "      <th>seller_id</th>\n",
       "      <th>cust_lat</th>\n",
       "      <th>cust_lng</th>\n",
       "      <th>seller_lat</th>\n",
       "      <th>seller_lng</th>\n",
       "      <th>product_weight_g</th>\n",
       "      <th>product_length_cm</th>\n",
       "      <th>product_height_cm</th>\n",
       "      <th>product_width_cm</th>\n",
       "      <th>price</th>\n",
       "      <th>freight_value</th>\n",
       "      <th>distance</th>\n",
       "      <th>volume</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>e481f51cbdc54678b7cc49136f2d6af7</td>\n",
       "      <td>7c396fd4830fd04220f754e42b4e5bff</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>-23.576983</td>\n",
       "      <td>-46.587161</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>500.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>29.99</td>\n",
       "      <td>8.72</td>\n",
       "      <td>18.566632</td>\n",
       "      <td>1976.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>128e10d95713541c87cd1a2e48201934</td>\n",
       "      <td>3a51803cc0d012c3b5dc8b7528cb05f7</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>-23.564630</td>\n",
       "      <td>-46.534268</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>500.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>29.99</td>\n",
       "      <td>7.78</td>\n",
       "      <td>15.803204</td>\n",
       "      <td>1976.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0e7e841ddf8f8f2de2bad69267ecfbcf</td>\n",
       "      <td>ef0996a1a279c26e7ecbd737be23d235</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>-23.600462</td>\n",
       "      <td>-46.655318</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>500.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>29.99</td>\n",
       "      <td>7.78</td>\n",
       "      <td>23.299562</td>\n",
       "      <td>1976.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>bfc39df4f36c3693ff3b63fcbea9e90a</td>\n",
       "      <td>e781fdcc107d13d865fc7698711cc572</td>\n",
       "      <td>87285b34884572647811a353c7ac498a</td>\n",
       "      <td>3504c0cb71d7fa48d967e0e4c94d59d9</td>\n",
       "      <td>-27.528367</td>\n",
       "      <td>-48.491914</td>\n",
       "      <td>-23.680729</td>\n",
       "      <td>-46.444238</td>\n",
       "      <td>500.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>29.99</td>\n",
       "      <td>14.10</td>\n",
       "      <td>473.254397</td>\n",
       "      <td>1976.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>53cdb2fc8bc7dce0b6741e2150273451</td>\n",
       "      <td>af07308b275d755c9edb36a90c618231</td>\n",
       "      <td>595fac2a385ac33a80bd5114aec74eb8</td>\n",
       "      <td>289cdb325fb7e7f891c38608bf9e0962</td>\n",
       "      <td>-12.177924</td>\n",
       "      <td>-44.660711</td>\n",
       "      <td>-19.807681</td>\n",
       "      <td>-43.980427</td>\n",
       "      <td>400.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>19.0</td>\n",
       "      <td>118.70</td>\n",
       "      <td>22.76</td>\n",
       "      <td>847.437333</td>\n",
       "      <td>4693.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           order_id                customer_unique_id  \\\n",
       "0  e481f51cbdc54678b7cc49136f2d6af7  7c396fd4830fd04220f754e42b4e5bff   \n",
       "1  128e10d95713541c87cd1a2e48201934  3a51803cc0d012c3b5dc8b7528cb05f7   \n",
       "2  0e7e841ddf8f8f2de2bad69267ecfbcf  ef0996a1a279c26e7ecbd737be23d235   \n",
       "3  bfc39df4f36c3693ff3b63fcbea9e90a  e781fdcc107d13d865fc7698711cc572   \n",
       "4  53cdb2fc8bc7dce0b6741e2150273451  af07308b275d755c9edb36a90c618231   \n",
       "\n",
       "                         product_id                         seller_id  \\\n",
       "0  87285b34884572647811a353c7ac498a  3504c0cb71d7fa48d967e0e4c94d59d9   \n",
       "1  87285b34884572647811a353c7ac498a  3504c0cb71d7fa48d967e0e4c94d59d9   \n",
       "2  87285b34884572647811a353c7ac498a  3504c0cb71d7fa48d967e0e4c94d59d9   \n",
       "3  87285b34884572647811a353c7ac498a  3504c0cb71d7fa48d967e0e4c94d59d9   \n",
       "4  595fac2a385ac33a80bd5114aec74eb8  289cdb325fb7e7f891c38608bf9e0962   \n",
       "\n",
       "    cust_lat   cust_lng  seller_lat  seller_lng  product_weight_g  \\\n",
       "0 -23.576983 -46.587161  -23.680729  -46.444238             500.0   \n",
       "1 -23.564630 -46.534268  -23.680729  -46.444238             500.0   \n",
       "2 -23.600462 -46.655318  -23.680729  -46.444238             500.0   \n",
       "3 -27.528367 -48.491914  -23.680729  -46.444238             500.0   \n",
       "4 -12.177924 -44.660711  -19.807681  -43.980427             400.0   \n",
       "\n",
       "   product_length_cm  product_height_cm  product_width_cm   price  \\\n",
       "0               19.0                8.0              13.0   29.99   \n",
       "1               19.0                8.0              13.0   29.99   \n",
       "2               19.0                8.0              13.0   29.99   \n",
       "3               19.0                8.0              13.0   29.99   \n",
       "4               19.0               13.0              19.0  118.70   \n",
       "\n",
       "   freight_value    distance  volume  \n",
       "0           8.72   18.566632  1976.0  \n",
       "1           7.78   15.803204  1976.0  \n",
       "2           7.78   23.299562  1976.0  \n",
       "3          14.10  473.254397  1976.0  \n",
       "4          22.76  847.437333  4693.0  "
      ]
     },
     "execution_count": 247,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['volume'] = df.product_height_cm * df.product_length_cm * df.product_width_cm\n",
    "df.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The next step is to analyze, the features and discover those features that can be important for the prediction.<br>\n",
    "\n",
    "For this purpose, the main metrics - distance, volume, weight, height, length, width, latitude and longitude are compared with the freight value, to see if they are correlated to the target. First, the distribution of all these varaibles have to be visualized to check whether to carry out a paramteric or non-parametric correlation test."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 248,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9e9e6c116f5b4f86a8199b9dad79d0c2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "interactive(children=(Dropdown(description='column', options=('product_weight_g', 'product_length_cm', 'produc…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "@interact\n",
    "def plot_hist(\n",
    "    column= ['product_weight_g', 'product_length_cm',\n",
    "            'product_height_cm','product_width_cm',\n",
    "            'distance','volume', 'freight_value', 'price']\n",
    "):\n",
    "    plt.figure(figsize= (12,8))\n",
    "    sns.distplot(df[column], kde= True)\n",
    "    plt.title('Histogram of {}'.format(column), fontdict= {'size':15})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "All the columns, show a histogram with a right/positive skew. They are nearly but not normal distribution. Hence the type of test chosen is a non-parametric test - Spearman's rank correlation test, which calculates Pearson's correlation using rank values instead of covariance (expected difference of observation from mean)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 249,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9b28ef94815d4012bb3bd184fb58fd37",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "interactive(children=(Dropdown(description='column', options=('product_weight_g', 'product_length_cm', 'produc…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.dropna(inplace= True)\n",
    "\n",
    "@interact\n",
    "def calc_corr(\n",
    "    column= ['product_weight_g', 'product_length_cm',\n",
    "            'product_height_cm','product_width_cm',\n",
    "            'distance','volume', 'cust_lat','cust_lng',\n",
    "            'seller_lat','seller_lng','price']\n",
    "):\n",
    "    print(\"Correlation: {}\".format(spearmanr(df[column],df.freight_value)))\n",
    "    print(\"Percentage of variablity shared: {}\".format(\n",
    "        round((spearmanr(df[column],df.freight_value).correlation ** 2)*100,2)\n",
    "    ))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "All the attributes/columns show a postive and high correlation with the freight value of the product. The latitude and longitude values show slightly less correlation as compared to the others, with longitude among them even showing negative correlation.<br>\n",
    "\n",
    "The observation made -\n",
    "* The dimensions of a product are positively correlated with the freight value, with a medium score (neighbourhood of 0.3 and 8% variability).\n",
    "* The weight and volume of a product are positively correlated with scores higher that the dimensions (neighbourhood of 0.4 and between 10% to 20% variability)\n",
    "* The distance shows the maximum positive correlation, about 0.6 with a 40% variability.\n",
    "\n",
    "These observations are the basis of selecting the features for predictive modeling.<br>\n",
    "Further PCA is used to reduce dimensionality of the data, so it can be visualized. For application of PCA, the concerned columns have to first Standardized, as scale has an effect on PCA. The columns, chosen to be apt features for the predictive model are following :-\n",
    "* length\n",
    "* height\n",
    "* width\n",
    "* volume\n",
    "* weight\n",
    "* distance\n",
    "* price\n",
    "\n",
    "The latitude and longitude values are not considered as they shows close to 0 correlation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 250,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Variance explained by each component: [0.30021166]\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>component_1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-1.441653</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-1.441778</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-1.441488</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-1.433274</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-1.043287</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   component_1\n",
       "0    -1.441653\n",
       "1    -1.441778\n",
       "2    -1.441488\n",
       "3    -1.433274\n",
       "4    -1.043287"
      ]
     },
     "execution_count": 250,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cols = [\n",
    "    'product_weight_g',\n",
    "    'product_length_cm',\n",
    "    'product_height_cm',\n",
    "    'product_width_cm',\n",
    "    'distance',\n",
    "    'volume',\n",
    "    'price',\n",
    "    'cust_lat',\n",
    "    'cust_lng',\n",
    "    'seller_lat',\n",
    "    'seller_lng'\n",
    "]\n",
    "\n",
    "scaler = StandardScaler()\n",
    "scaled = scaler.fit_transform(df[cols])\n",
    "\n",
    "pca = PCA(n_components= 1)\n",
    "reduced = pca.fit_transform(scaled)\n",
    "pca_df = pd.DataFrame(reduced, columns= ['component_1'])\n",
    "print(\"Variance explained by each component: {}\".format(pca.explained_variance_ratio_))\n",
    "pca_df.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The Principle component analysis, has reduced the dimensionality of the dataset from 6 to 1. The variance explained by the two dimensions together is about 52% which is very good given the degree of reduction. The two components are visualized below in a scatter plot,"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 251,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<seaborn.axisgrid.FacetGrid at 0x7f106aacf160>"
      ]
     },
     "execution_count": 251,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 864x576 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "tmp = pd.concat([pca_df,df.freight_value], axis=1)\n",
    "\n",
    "plt.figure(figsize= (12,8))\n",
    "sns.lmplot(\n",
    "    data=tmp.sample(frac=0.15, random_state= 5),\n",
    "    x= 'component_1',\n",
    "    y= 'freight_value',\n",
    "    legend= True,\n",
    "    height= 8,\n",
    "    aspect= 2\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "All the columns are also plotted against, the target variable freight_value below."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 252,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d1fd2e690463494d8cec42c705add825",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "interactive(children=(Dropdown(description='column', options=('product_weight_g', 'product_length_cm', 'produc…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "@interact\n",
    "def plot_lm(\n",
    "    column= ['product_weight_g', 'product_length_cm',\n",
    "            'product_height_cm','product_width_cm',\n",
    "            'distance','volume', 'cust_lat','cust_lng',\n",
    "            'seller_lat','seller_lng','price']\n",
    "):\n",
    "    sns.lmplot(data= df, x= column, y= 'freight_value', aspect= 2, height= 8)\n",
    "    plt.title('Linear model plot of {}'.format(column), fontdict= {'size':15})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The lmplot, plotting a Linear regression line on the reduced 2D data and individual columns, doesnt look very promising. It shows that the component of PCA and individual columns share a slight linear relationship.<br>\n",
    "\n",
    "The PCA data is also appended to the 6 columns, as derived features.<br>\n",
    "\n",
    "The next step, is to prepare the dataset for training and testing. The 6 columns chosen as features are already scaled using a StandardScaler. The scaled data is split into train and test set. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 253,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaled = pd.DataFrame(scaled, columns = cols)\n",
    "X = pd.concat([scaled,pca_df], axis=1)\n",
    "y = df.freight_value\n",
    "\n",
    "X_train, X_test,y_train,y_test = train_test_split(X,y,random_state= 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A 7-fold cross validation is performed on a LinearRegression model, with the scoring metrics as MAE, RMSE and R-squared."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 254,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 4 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Average mean absolute error across k-folds: -5.285396629449821\n",
      "Average root mean square error across k-folds: -10.01777169723435\n",
      "Average r-squared across k-folds: 0.5851287896610581\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Done   3 out of   7 | elapsed:    3.2s remaining:    4.3s\n",
      "[Parallel(n_jobs=-1)]: Done   7 out of   7 | elapsed:    3.3s finished\n"
     ]
    }
   ],
   "source": [
    "values = cross_validate(\n",
    "    LinearRegression(),\n",
    "    X= X,\n",
    "    y= y,\n",
    "    scoring= ['neg_mean_absolute_error','neg_root_mean_squared_error','r2'],\n",
    "    cv= 7,\n",
    "    n_jobs= -1,\n",
    "    verbose= 3\n",
    ")\n",
    "\n",
    "print(\"Average mean absolute error across k-folds: {}\".format(np.mean(values['test_neg_mean_absolute_error'])))\n",
    "print(\"Average root mean square error across k-folds: {}\".format(np.mean(values['test_neg_root_mean_squared_error'])))\n",
    "print(\"Average r-squared across k-folds: {}\".format(np.mean(values['test_r2'])))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The metrics show that MAE and RMSE are quite a lot given that values for freight cost predominantly lie between 5 to 30. The R-squared shows that the independent variables only explain about 58% of the variance in the target value.<br>\n",
    "Overall the model is a bad fit. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 255,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "intercept: 20.126363591111833\n",
      "product_weight_g: 4.315653269687464\n",
      "product_length_cm: -1.456824435949982\n",
      "product_height_cm: -3.058577309522259\n",
      "product_width_cm: -2.448345774227055\n",
      "distance: 6.211403984690808\n",
      "volume: 5.296122894314587\n",
      "price: 2.492202533985378\n",
      "cust_lat: -0.30024241970813287\n",
      "cust_lng: 0.2331449986572526\n",
      "seller_lat: 0.3255035144236375\n",
      "seller_lng: -0.1465831773862257\n",
      "component_1: 2.634352911526346\n"
     ]
    }
   ],
   "source": [
    "mod = LinearRegression()\n",
    "mod.fit(X_train,y_train)\n",
    "\n",
    "print(\"intercept: {}\".format(mod.intercept_))\n",
    "for col,coef in zip(X.columns,mod.coef_):\n",
    "    print(\"{}: {}\".format(col,coef))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The problem can be seen here. The coefficients for each column are large either on the postive side or negative side. To deal with this regularization needs to be performed. Two models offer regularization capabilites - Lasso and Ridge regression.<br>\n",
    "\n",
    "Ridge uses L2 regularization, whereis the square of the coefficients is used as penalty for the cost function. Lasso uses L1 regularization, where the penalty is just the absolute value of coefficients. Due to this Lasso also results in some features having a coefficient of 0 and there by indirectly perform feature selection.<br>\n",
    "\n",
    "First the Lasso is tested. LassoCV performs cross validation and search on parameters tor return the best model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 256,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train R2: 0.599707954096099\n",
      "Test R2: 0.6045190715388251\n",
      "\n",
      "intercept: 20.104742063686764\n",
      "product_weight_g: 6.728686577755517\n",
      "product_length_cm: 0.47746179747574163\n",
      "product_height_cm: -1.0145692442856855\n",
      "product_width_cm: -0.36551267805464827\n",
      "distance: 6.182662411520607\n",
      "volume: 8.155337113575978\n",
      "price: 3.659325096727822\n",
      "cust_lat: -0.34285206941867885\n",
      "cust_lng: 0.14310826422530704\n",
      "seller_lat: 0.14919592618990712\n",
      "seller_lng: -0.2837492413164702\n",
      "component_1: -2.655568610548642\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/raj/anaconda3/envs/r_env/lib/python3.8/site-packages/sklearn/linear_model/_coordinate_descent.py:474: ConvergenceWarning: Objective did not converge. You might want to increase the number of iterations. Duality gap: 1017464.0136102466, tolerance: 2575.8276549645143\n",
      "  model = cd_fast.enet_coordinate_descent(\n"
     ]
    }
   ],
   "source": [
    "lmod = LassoCV(alphas= [0.001,0.01,0.1,1,10], n_alphas=5,n_jobs= -1, random_state= 5, cv= 7)\n",
    "lmod.fit(X,y)\n",
    "\n",
    "print(\"Train R2: {}\".format(lmod.score(X_train,y_train)))\n",
    "print(\"Test R2: {}\\n\".format(lmod.score(X_test,y_test)))\n",
    "print(\"intercept: {}\".format(lmod.intercept_))\n",
    "for col,coef in zip(X.columns,lmod.coef_):\n",
    "    print(\"{}: {}\".format(col,coef))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The coefficients are now smaller as compared to normal Linear regression, yet the R-squared score has not increased. The model is still a bad fit on the data.<br>\n",
    "\n",
    "The Ridge model is tested lastly, RidgeCV also perform cross validation and grid search for optimal parameters to return the best model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 257,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train R2: 0.5997086642261844\n",
      "Test R2: 0.6045172403659007\n",
      "\n",
      "intercept: 20.104742063686764\n",
      "product_weight_g: 4.213249642559206\n",
      "product_length_cm: -1.4484839307696449\n",
      "product_height_cm: -3.071459273045646\n",
      "product_width_cm: -2.4931964438797802\n",
      "distance: 6.210999751251229\n",
      "volume: 5.421949553832589\n",
      "price: 2.438213293265794\n",
      "cust_lat: -0.2882778942359793\n",
      "cust_lng: 0.1948909430872652\n",
      "seller_lat: 0.32614444511288077\n",
      "seller_lng: -0.13899707762859095\n",
      "component_1: 2.618112475182654\n"
     ]
    }
   ],
   "source": [
    "rmod = RidgeCV(cv= 7)\n",
    "rmod.fit(X,y)\n",
    "\n",
    "print(\"Train R2: {}\".format(rmod.score(X_train,y_train)))\n",
    "print(\"Test R2: {}\\n\".format(rmod.score(X_test,y_test)))\n",
    "print(\"intercept: {}\".format(rmod.intercept_))\n",
    "for col,coef in zip(X.columns,rmod.coef_):\n",
    "    print(\"{}: {}\".format(col,coef))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The coefficients are even smaller as compared to Lasso, but the R-squared score has not improved for the model. These 3 models - LinearRegression, Lasso and Ridge show that there is no apparent linear relationship between the predictors and target varible.<br>\n",
    "\n",
    "The assumption that all features are directly share a linear relationship with the dependent variable has not yielded good results. Hence the next two models are tree based Regressors. Tree based regressors do not assume a linear relationship, rather the trees are built on the data and the final output is the mean of the prediction of all trees. <br>\n",
    "\n",
    "The first to be tested is the RandomForestRegressor."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 259,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  ................................................................\n",
      "[CV]  , neg_mean_absolute_error=-4.118, neg_root_mean_squared_error=-8.025, r2=0.629, total= 5.2min\n",
      "[CV]  ................................................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:  5.2min remaining:    0.0s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  , neg_mean_absolute_error=-3.990, neg_root_mean_squared_error=-9.221, r2=0.656, total= 4.9min\n",
      "[CV]  ................................................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed: 10.1min remaining:    0.0s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  , neg_mean_absolute_error=-4.339, neg_root_mean_squared_error=-9.876, r2=0.710, total= 4.7min\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed: 14.8min remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed: 14.8min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Average mean absolute error across k-folds: -4.1486575563399155\n",
      "Average root mean square error across k-folds: -9.040968656130149\n",
      "Average r-squared across k-folds: 0.6648417328596358\n"
     ]
    }
   ],
   "source": [
    "values = cross_validate(\n",
    "    RandomForestRegressor(n_estimators= 400),\n",
    "    X= X,\n",
    "    y= y,\n",
    "    scoring= ['neg_mean_absolute_error','neg_root_mean_squared_error','r2'],\n",
    "    cv= 3,\n",
    "    verbose= 10,\n",
    ")\n",
    "\n",
    "print(\"\\nAverage mean absolute error across k-folds: {}\".format(np.mean(values['test_neg_mean_absolute_error'])))\n",
    "print(\"Average root mean square error across k-folds: {}\".format(np.mean(values['test_neg_root_mean_squared_error'])))\n",
    "print(\"Average r-squared across k-folds: {}\".format(np.mean(values['test_r2'])))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The RandomForestRegressor achieves good MAE and RMSE values as compared to the Linear Regression models. The R-squared value across the k-folds has also improved from a 56% to 66%. The RandomForestRegressor can account for 66% of variance of the dependent variables (freight value) with the given independent variable.\n",
    "\n",
    "The next model is GradientBoostingRegressor."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 260,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  ................................................................\n",
      "[CV]  , neg_mean_absolute_error=-4.135, neg_root_mean_squared_error=-7.857, r2=0.644, total= 2.8min\n",
      "[CV]  ................................................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:  2.8min remaining:    0.0s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  , neg_mean_absolute_error=-4.067, neg_root_mean_squared_error=-9.225, r2=0.656, total= 2.9min\n",
      "[CV]  ................................................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   2 out of   2 | elapsed:  5.7min remaining:    0.0s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV]  , neg_mean_absolute_error=-4.353, neg_root_mean_squared_error=-9.824, r2=0.713, total= 2.8min\n",
      "\n",
      "Average mean absolute error across k-folds: -4.185130576635187\n",
      "Average root mean square error across k-folds: -8.968702154063493\n",
      "Average r-squared across k-folds: 0.6708987657524141\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:  8.4min remaining:    0.0s\n",
      "[Parallel(n_jobs=1)]: Done   3 out of   3 | elapsed:  8.4min finished\n"
     ]
    }
   ],
   "source": [
    "values = cross_validate(\n",
    "    GradientBoostingRegressor(n_estimators= 1000, learning_rate= 0.05),\n",
    "    X= X,\n",
    "    y= y,\n",
    "    scoring= ['neg_mean_absolute_error','neg_root_mean_squared_error','r2'],\n",
    "    cv= 3,\n",
    "    verbose= 10\n",
    ")\n",
    "\n",
    "print(\"\\nAverage mean absolute error across k-folds: {}\".format(np.mean(values['test_neg_mean_absolute_error'])))\n",
    "print(\"Average root mean square error across k-folds: {}\".format(np.mean(values['test_neg_root_mean_squared_error'])))\n",
    "print(\"Average r-squared across k-folds: {}\".format(np.mean(values['test_r2'])))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The GradientBoostingRegressor achieves slightly better R-squared than the RandomForest varient. Thus out of all the models tested, the GradientBoostingRegressor is chosen. The model is trained on the train set with varying number of estimators, the test score of each model is plotted against the estimators, to select the optimal one."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Test scores for different number of estimators')"
      ]
     },
     "execution_count": 149,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def fit_model(n):\n",
    "    gb = GradientBoostingRegressor(n_estimators= n, learning_rate= 0.05, random_state= 5)\n",
    "    gb.fit(X_train, y_train)\n",
    "    return gb.score(X_test,y_test)\n",
    "\n",
    "test_scores = []\n",
    "estimators = np.arange(100,2000,100)\n",
    "for i in estimators:\n",
    "    test_scores.append(fit_model(i))\n",
    "    \n",
    "plt.figure(figsize= (12,8))\n",
    "sns.lineplot(x= estimators, y= test_scores)\n",
    "plt.title('Test scores for different number of estimators', fontdict= {'size':15})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The n_estimators with a value of 1000 looks good as it achieves a peak after constant growth and then stabilizes for some future values. After this even though the scores go up, they may account to overfitting and hence is not considered. The value of 1000 for n_estimators is taken to be as final.<br>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 261,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train score: 0.7938718968817781\n",
      "Test score: 0.73097714974654\n"
     ]
    }
   ],
   "source": [
    "gb = GradientBoostingRegressor(n_estimators= 800, learning_rate= 0.05, random_state= 5)\n",
    "gb.fit(X_train, y_train)\n",
    "\n",
    "print(\"Train score: {}\".format(gb.score(X_train, y_train)))\n",
    "print(\"Test score: {}\".format(gb.score(X_test, y_test)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The model shows a significant rise in the R-squared value with around 71%. To visualize the predicted value, a scatter plot is made below. It plots the predicted values vs the actual values. Ideally, for a 100% accurate prediction, the dots should lie on the 45 deg line. The deviation of the predicted values from the said line, can testify how well/bad has the model performed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 282,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Predicted value vs Actual value')"
      ]
     },
     "execution_count": 282,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtQAAAHzCAYAAAAAbIrjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOzdeXyU1fX48c8JEKSIiohUAaUuVcEFNa7UunWx2lapS7W2dHHftdYuWuv+bV3qLta1Sq1btS5tra1arZVW/cUNFcUNF1wwImIEBMPc3x/PRENMQiCZPJnM5/16zSszd55n5mRmxDM3554bKSUkSZIkLZmqvAOQJEmSypkJtSRJktQBJtSSJElSB5hQS5IkSR1gQi1JkiR1gAm1JEmS1AEm1JLaFBEnRkRqcnkjIm6OiNVL/Lw3RcR9zeJ4ZzHOry6eM7oTYzo0Irqk12hE/KD4ei/dFc9XahHxw+Lv84clPH+PiPhBJ4fV/DlSRBxayudo8lwvR8RZXfFckkrPhFpSe8wCtihefgKMBu6JiP5dGMPlwFcX4/hq4ASyWJW/vYo/d4mIfktw/h7ADzovHEnqPCbUktqjIaX0YPFyLfB9YFVgx5YOjoheEVHdmQGklKallB7pzMdU14iIFYHtgHuApYGv5xuRJHUuE2pJS6IxsR0BEBFXRURtROwSEU8DHwKbFe9bJSKuj4h3I2JORPwjItZq+mARMTwi7oiIucU/he/b/AlbKvmIiEERcUlEvBkRH0bElIg4snh3ffHn75uUqzTGu1REnBERr0XEvIh4IiJ2bPbYfSPiwoh4rxj7OUCftl6UiNi2+Dyjmo0PjIj5EbFP8fYWEXF7sXxmdkQ8HhF7L+Kxtyk+9rrNxu+LiJuajX0hIv5dfL1nRMRlETGgjcf+YfF1WK7Z+Kjic27f5HH/ExHvFy+PR8TubcVdtAfQCzgUeJ1PZqubx7FfRDxZfC+nF8t+lo2Iq4Bdga2bvJcnFs/5VOlE83KZiOhffC+nFF+TqRFxUUQs047Ymz7uvyPixhbGz4qIVyMiird/U/w9PoiIaRHxx4j47CIeu6X38VPveXs+u5K6Xu+8A5BUlkYUf77VbOwM4GRgOjA1IpYHHgBmAAcCc4CfA3dHxOdTSnOLSchtwArAPmTJ+EnA8sDzrQUQWdnAfcCKxeOfBdYoXiCbEf0XcCrwt+LYm8WfNwGbkpWEvEiW8N0eETUppceLx/wG2Bc4DpgM7AcsKnn8d/E59ig+dqOxxZ+3FH+uCkwEflf8fceQJf6FlNJ1i3iONkXEGLKZ4FuB3YBBxd9lYPF2S/5cjGUs8Psm498G3gbuKyaffyV7r04GAlgPWI5F2wt4LKX0bETcABwSEcumlGY1ifuXxccdDxwDfAbYiWxG+xRgleJzHVw8ZVo7nrfRZ8gS+uOAOmB48fqfWLwyouuB30ZE/5TS7GLcQfa5uDGl1FhfvyLwf8AbwGDgaOBfEbFeSmnBYjxfS9rz2ZXU1VJKXrx48dLqBTgReIfsC3hv4PPAvcD7wErFY64CEjC62bmnkCXTyzcZG0hWk31I8faOxXM3a3LMqkADcF/zOJrcPgAoNH/OJvcvXXzcHzQb3744vnWz8fuBPxWvDwLmAj9rcn8VWdKeFvF6nQc822zsH8BfWzk+iq/rJcC/moz/oBjn0sXb2xRvr9vs/PuAm5rc/g9wb7Njtmvp3GbH3Abc2WxsCnBh8XpN8TEGLObnZ9Xi+/TT4u1Nmr8vZInyHODsNh7npqafhybjLwNnNRtb6LVr4ZzeZF9iErBKk/EEHNpGDIOLn8s9m4xtUTyvppVzegFDi8d8sbW4m7+PLb3n7fnsevHiJZ+LJR+S2mMQ8FHxMgVYDfh2SunNJse8nj49Q/Yl4C7g/YjoHRG9yUoxHiFL0CCbbZueUnqo8aSU0it8UlbSmu3IZj0Xd1buS2Qz6xMbYyrGdU+TmNYDliJLMhtjKjS93YYbgLUiYgOAiFihGOsNjQcUS0DOj4hX+OR13Z/sy8oSi4jPkCV4Nzb73R4oPsfGi4h7+2K8RNYd5fNN4n4R+AC4NiJ2bl4e0oY9mzw+KaX/B7zAwmUfWwD9WHh2vFNFxPci4rGI+IDstXigeFe7X/OUUh3ZXz2+3WT428CLKaXaJs/1tYj4b0TMIkvAG2fTO/T+0r7PrqQcmFBLao9ZZDOLNcAwYERK6e/NjpnewnkrkCUcHzW7bEv2Z3eAz5KVFTTX0lhTg/ikhGNxrFB8zuYxndgsppZiWFRMAP8DXuWTpGtXsqTq1ibHXFW8/0zgK2Sv7ZVkSXxHDCSbER3Pwr/bPLL67+Gtn8rtxWO/Vbz9bbJ65wcAUkozi7H2AW4E6iLibxGx2iJi2gt4FJgVEcsVE/HbyZL3FYvHDCr+XJL3c5EiYiwwgey92R3YnE/KcBb3Nb8e+FpELBMRVcXHa/plaROy328a8D2yLwubL+FzNdeez66kHFhDLak9GprOwLWipf7M75IlF6e0cF/josG3yGpOm1uRrOyiNTP4pF56cbxLliju0sYxjbXhKxaPbxpTm1JKqbhw7dvAscWff08p1UO2qIysNvjQlNLvGs8rJmdt+bD4s3n3lOXJSnIA3iN7H04E7mjhMd5oI+4PIuJvxXgvJavNbVoXTErpf8AOxfr1LwFnA9fyScK4kIhYG9igeHNmC4fsDlxE9l4CrNTkd2mvD2n5NWn+PA+llBrrr4mIrRfzeRrdAlwM7Ay8AqxMk4SaLFGvI/sLTio+16rteNz2/B7t+exKyoEJtaRSuocsMXs6pdRacvz/gBMiYrPGso+IWAXYiGzhXluPvXtErJ9SmtTC/fOLP5vPCt5Dtkjsg5TSs6089pNkCc7OZHXTjQnvzm3E09T1wE8i4uvA1ixc3tCXbBZ5XuNAsQPHN2n5S0mjxrKBdchmfImI4cBawHMAKaXZEfEgsFZK6eR2xto87hsi4htkZT3Xt3RQ8b38S7H7xC/aeLzvAAvIfrc5ze47j+x1uYhs5nguWTvGn7TyWPNpeYZ3Gtlr0tSXm93uR5PXu6jNriqtSSnNjIh/kn3xeAV4ptnnrx/wUdMvIu18rmnAF5uNNf892vPZlZQDE2pJpXQ28F2yDgcXkM2uDSFLMh9IWUeLO4AngD9FxM/IEtmTWXR5xQTgEOCfxRZqU4DPAZ9PKf08pTQ/IqYCe0TEU8XHnURW0/0P4K6IOB14GliGbAOYpVJKv0gpzYiIS4GTIqKheMx+ZAsdFyml9EhEvEA20zuXrDtG432zIuL/Ab+KiPfJFuz9nKysptU2bimlacXzTomIOWQle8ey8Aw6wE/JNt0pkC3kqyfrkLETcFxK6bk2Qv8bWeJ7CTA1pfRw4x0RsRPwI7LSlVfJFtodQFZT3Jo9gbtSSp+aLY+Iq4GzImLVlNIrEXEKcFpk/cvvIPvisRNwUkrpdbIvNjtHxC5kyecbKaU3yGaML4iIY8m+nH0LGNXs6e4CLoqI44CHyBbCbt9G3ItyA1mJzizgwhae68iIOBf4C7Al2X8Di3ILsE9k7Rn/RlYW1bwDySI/u0v260jqsLxXRXrx4qV7X2jWXaOVY64Calu5b2WyxWbTyWYJXwauAUY1OWYV4E6y5PMVskRtoa4OLcVBVnt7GVny/SFZ0nV4k/u/QpZEf0g2+zuiON6XrNXeC2Qzn28Vn3+nJuf2JatFnkVWrnAB8GMW0eWjyfmnFp/zuhbuW4MsEZ1Nlpz+tPnvRwudKorn3Vc8bwrZjPl9fLo7xGbF3+f94rGTyb7cLNuOuK8pPu+vm42vVXxPXiu+j9PIWu0t38rjNHYF+U4r969EVlvetJPKAcVY5xXfkxuBZYr3rUCWdL5bfNwTi+N9ir/bW8X36TyyBZ5NO6T0As4qfk7eB24uvkYJ+HqT52+zy0eT4waQffFIZH8NaH7/T4uv02zgbmDN5o9Ny91JflE8r774PjT+1WLdJscs8rPrxYuXrr9ESm39hVGSJElSW+zyIUmSJHWACbUkSZLUASbUkiRJUgeYUEuSJEkdYEItSZIkdUBZ96FeYYUV0ogRI/IOQ5IkST3cI4888k5KaXBL95V1Qj1ixAhqaxe1G7IkSZLUMRHxSmv3WfIhSZIkdYAJtSRJktQBJtSSJElSB5hQS5IkSR1gQi1JkiR1gAm1JEmS1AEm1JIkSVIHmFBLkiRJHWBCLUmSJHWACbUkSZLUASbUkiRJUgeYUEuSJEkdYEItSZIkdUDJEuqIWCoiHo6IJyLi6Yg4qTh+VURMjYjHi5fRxfGIiPMj4oWImBQRG5UqNkmSJKmz9C7hY88DtkspfRARfYAHIuLvxfuOSSnd1Oz4rwFrFi+bARcXf0qSJEndVslmqFPmg+LNPsVLauOUnYEJxfMeBJaLiJVKFZ8kSZLUGUpaQx0RvSLiceBt4K6U0kPFu04rlnWcExF9i2NDgdeanD6tOCZJkiR1WyVNqFNKC1JKo4FhwKYRsS7wC2BtYBNgeeBnxcOjpYdoPhAR+0dEbUTU1tXVlShySZIkdRuzZ8OPfwyPPJJ3JC3qki4fKaX3gPuAHVJKbxbLOuYBvwc2LR42DRje5LRhwBstPNalKaWalFLN4MGDSxy5JEmScpUSbLMNnHMO/OtfeUfTolJ2+RgcEcsVr/cDvgQ821gXHREB7AI8VTzldmBcsdvH5sCslNKbpYpPkiRJ3disWVAoQAQcdxzcdx8cc0zeUbWolDPUKwH3RsQk4P+R1VD/FfhjRDwJPAmsAJxaPP4O4CXgBeAy4OASxiZJkqTu6rbbYJ11YPz47PYuu8DWW+cbUxtK1jYvpTQJ2LCF8e1aOT4Bh5QqHkmSJHVz06fD4YfDjTfC+uvDFlvkHVG7uFOiJEmS8nfbbTByJNx6K5x6KtTWwsYb5x1Vu5RyYxdJkiSpfZZZBtZeGy6/PCv3KCMm1JIkSep6hQJccgm8/TaccAJsuy088EC2CLHMWPIhSZKkrvXcc1krvIMPhgcfhAULsvEyTKbBhFqSJEldpaEBzjgDNtgAnnwSrrwS7rgDevXKO7IOseRDkiRJXWPqVDj+eNhpJ7joIlhppbwj6hTOUEuSJKl05s2D66/Prq+5JkyaBDff3GOSaTChliRJUqn8978wejTstRc88kg2ttZaZVsr3RoTakmSJHWuDz6AI46AL3wB5syBO+8sm57SS8IaakmSJHWelLIOHo8+CoccAv/3fzBgQN5RlZQJtSRJkjruvfeyxLlXL/jVr2D55bMZ6gpgyYckSZI65pZbst0Nx4/Pbn/zmxWTTIMJtSRJkpbUW2/B7rvDt74Fn/1sRSXRTZlQS5IkafHdcguMHAl/+UtWJ/3ww7DhhnlHlQtrqCVJkrT4llsORo2Cyy6DtdfOO5pcmVBLkiRp0QqFrEa6rg5OOgm23Rbuv7/H9ZReEpZ8SJIkqW3PPgtf/CIcdli2QcuCBdm4yTRgQi1JkqTWfPRRVh+9wQYweTJcfXVWM92rV96RdSuWfEiSJKllL78MJ5+ctcG78EIYMiTviLolZ6glSZL0iQ8/hD/+Mbu+5prw5JPwpz+ZTLfBhFqSJEmZBx7Iyju++91s63DIkmq1yYRakiSp0tXXw6GHwlZbwfz58M9/wkYb5R1V2bCGWpIkqZKllLXAe/RROPxwOO00WHrpvKMqKybUkiRJlejdd2HZZbOOHSecAIMGwZZb5h1VWbLkQ5IkqZKklC0yXGedrHMHwDe+YTLdASbUkiRJleLNN+Fb34I99oDhw2GbbfKOqEcwoZYkSaoEf/4zjBwJd94JZ5wBDz6YdfRQh1lDLUmSVAkGDcoS6Esvhc9/Pu9oehQTakmSpJ5owQK46CKoq4NTToGtt4Z774WIvCPrcSz5kCRJ6mmeeSbrKX3EEfD441lyDSbTJWJCLUmS1FN89BGceiqMHg3PPQd/+APcfnvWGk8lY8mHJElST/Hyy1lCPXYsnH8+rLhi3hFVBGeoJUmSytncudlMNMCaa8LTT8P115tMdyETakmSpHL173/D+uvDuHHZ1uEAq6+eb0wVyIRakiSp3Lz/Phx0ULYxS6EA99wDG22Ud1QVyxpqSZKkcpJSlkg/8QT8+Mdw8snQv3/eUVU0E2pJkqRy8O67sOyyWceOU06BFVaAzTbLOyphyYckSVL3lhLccAOsvTZccEE2ttNOJtPdiAm1JElSd/XGG7DLLrDnnrDqqrD99nlHpBaYUEuSJHVHN90EI0fCXXfBWWfB//4H662Xd1RqgTXUkiRJ3dGKK8LGG8Mll8Aaa+QdjdpgQi1JktQdLFgA550HM2bAaafBF78Id98NEXlHpkWw5EOSJClvTz0FW24JRx+d7XS4YEE2bjJdFkyoJUmS8jJ/Ppx0UrYpy0svwXXXwS23ZK3xVDZMqCVJkvLy6qvwm9/A7rvDM89k3TyclS47JtSSJEldac4cuPrq7Poaa8DkyfDHP2YbtagsmVBLkiR1lfvug/XXhx/8AB57LBv73OfyjEidoGQJdUQsFREPR8QTEfF0RJxUHP9cRDwUEc9HxA0RUV0c71u8/ULx/hGlik2SJKlLzZoFBxwA226b3b73Xthww3xjUqcp5Qz1PGC7lNIGwGhgh4jYHDgdOCeltCYwE9inePw+wMyU0hrAOcXjJEmSyltKsM02cPnl8JOfwKRJ2W31GCVLqFPmg+LNPsVLArYDbiqOXw3sUry+c/E2xfu3j7AqX5Iklal33sna30VkfaUffBDOPBM+85m8I1MnK2kNdUT0iojHgbeBu4AXgfdSSg3FQ6YBQ4vXhwKvARTvnwUMKmV8kiRJnS4luPZaWHttOP/8bGzHHWGTTfKNSyVT0oQ6pbQgpTQaGAZsCqzT0mHFny3NRqfmAxGxf0TURkRtXV1d5wUrSZLUUa+9Bt/4Buy9d9bB4ytfyTsidYEu6fKRUnoPuA/YHFguIhq3PB8GvFG8Pg0YDlC8f1ng3RYe69KUUk1KqWbw4MGlDl2SJKl9brwRRo3KFhyecw5MnJjdVo9Xyi4fgyNiueL1fsCXgGeAe4Hdiod9H7iteP324m2K9/8rpfSpGWpJkqRuaaWVYPPN4ckn4cgj3e2wgvRe9CFLbCXg6ojoRZa435hS+mtETAauj4hTgceAK4rHXwH8ISJeIJuZ3rOEsUmSJHVMQwOcey7MmAG//jVstRX88595R6UclCyhTilNAj7VYDGl9BJZPXXz8Q+B3UsVjyRJUqeZNAn22Qdqa2Hs2KybhzPSFcudEiVJktpr3jw44QTYeGN45RW44Qa4+WaT6QpnQi1JktRer70GZ5wBe+0FzzwDe+yR9ZlWRTOhliRJasvs2XDlldn1NdaAZ5+FCRNgkNtlKGNCLUmS1Jq774b11svqpR97LBtbddV8Y1K3Y0ItSZLU3HvvZUn0l78MvXvDv/8NG36q14IElLZtniRJUvlJCbbZBp56Cn72s2wRYr9+eUelbsyEWpIkCaCuDpZfPuvY8ZvfwODBWTcPaREs+ZAkSZUtJfjDH2DtteG887KxHXYwmVa7mVBLkqTK9eqrsNNOMG4crLUWfO1reUekMmRCLUmSKtP118OoUXD//XD++fCf/8A66+QdlcqQNdSSJKkyDRsGY8bA734HI0bkHY3KmAm1JEmqDA0NcPbZ8O672aLDL3wB7rwz76jUA1jyIUmSer4nnoDNNsva4L34IhQKeUekHsSEWpIk9Vwffgi//CXU1MDrr8NNN8Gf/gRVpkDqPH6aJElSzzVtWlbmsffeMHky7Lpr3hGpBzKhliRJPcsHH8Dll2f9pddYA6ZMgauuyjZtkUrAhFqSJPUc//wnrLsu7L8/PP54NjZ8eL4xqcczoZYkSeVv5kz44Q/hq1+FpZbKektvuGHeUalC2DZPkiSVt5Rgm23g6afhF7+AX/0qS6qlLmJCLUmSytPbb2d10b17wxlnwIorOiutXFjyIUmSyktKcPXVsPbacN552dhXv2oyrdyYUEuSpPLx8suwww7wgx/AyJGw0055RySZUEuSpDJx3XVZB4+JE+GCC7KFh2uvnXdUkjXUkiSpTKyyCnzxi3DxxbDqqnlHI33MhFqSJHVPH30EZ52VtcQ74wwYMwbuuCPvqKRPseRDkiR1P48+CptuCsceC6++CoVC3hFJrTKhliRJ3cfcuVkv6U03hbfegj//Ga6/HqpMWdR9+emUJEndxxtvZK3wvv99mDwZxo7NOyJpkayhliRJ+aqvzzp47LcfrL46PPccDBuWd1RSuzlDLUmS8nPnnVkrvAMPhCeeyMZMplVmTKglSVLXmzEjK+v42tegf/+st/To0XlHJS0RSz4kSVLXSgm23RaeeQZ++cvs0rdv3lFJS8yEWpIkdY233oIVVoDevbP+0kOGwAYb5B2V1GGWfEiSpNJKCa68Mtsm/Nxzs7GvfMVkWj2GCbUkSSqdqVOz5HmffbIEeued845I6nQm1JIkqTSuvTbr4PHQQ3DxxXDvvbDmmnlHJXU6a6glSVJpjBgB220H48fD8OF5RyOVjAm1JEnqHB99BKefDrNmwZlnwpZbwl/+kndUUslZ8iFJkjrukUegpgaOPz7bPrxQyDsiqcuYUEuSpCU3dy787Gew6abwzjtw223wxz9ClSmGKoefdkmStOTeeAMuvBB+9CN4+mn45jfzjkjqctZQS5KkxfP++9ks9IEHwuqrw3PPwdCheUcl5cYZakmS1H533AGjRsGhh8ITT2RjJtOqcCbUkiRp0d55B777XdhpJ1h2Wfjvf2H06LyjkroFSz4kSVLbUoJtt4UpU+CEE+DYY6G6Ou+opG7DhFqSJLXszTdh8GDo3RvOOQeGDIH11ss7KqnbseRDkiQtLCW4/HJYZ50skQb40pdMpqVWmFBLkqRPvPhiljzvtx9suCGMHZt3RFK3V7KEOiKGR8S9EfFMRDwdEUcUx0+MiNcj4vHiZccm5/wiIl6IiCkR8dVSxSZJklpwzTXZLHRtLVxyCdxzD6yxRt5RSd1eKWuoG4CjU0qPRsQA4JGIuKt43zkppbOaHhwRI4E9gVHAysDdEfH5lNKCEsYoSZIarb46fPnLcNFFMGxY3tFIZaNkCXVK6U3gzeL1+oh4BmirUeXOwPUppXnA1Ih4AdgU+F+pYpQkqaLNnw+//jXMmgVnnw1bbJFtHS5psXRJDXVEjAA2BB4qDh0aEZMi4sqIGFgcGwq81uS0abSdgEuSpCX18MOw8cZw4olQVweFQt4RSWWr5Al1RCwN3AwcmVJ6H7gYWB0YTTaD/dvGQ1s4PbXwePtHRG1E1NbV1ZUoakmSeqg5c+Doo7PZ6Jkz4S9/gT/8AarsUyAtqZL+1xMRfciS6T+mlP4MkFKanlJakFIqAJeRlXVANiM9vMnpw4A3mj9mSunSlFJNSqlm8ODBpQxfkqSe5623sgWH++8PkydT2HEn6urn8frMOdTVz6NQ+NRclqRFKFkNdUQEcAXwTErp7CbjKxXrqwHGAk8Vr98OXBsRZ5MtSlwTeLhU8UmSVDFmzco6eBx8MKy2Gjz/PKy0EoVCYsr0evabUMu0mXMZNrAfl42rYa0hA6iqaukPx5JaUsoZ6jHA94DtmrXIOyMinoyIScC2wFEAKaWngRuBycCdwCF2+JAkqYP+8hcYORIOPxyeeCIbW2klAGbMnv9xMg0wbeZc9ptQy4zZ8/OKVipLpezy8QAt10Xf0cY5pwGnlSomSZIqRl0dHHEEXHdd1lv61lth9OiFDpnfsODjZLrRtJlzmd/gfJa0OFyBIElST5MSbLcd3HQTnHRStlHLJpt86rDq3r0YNrDfQmPDBvajunevropU6hFMqCVJ6ilefx0aGiACzj0XHnsMfvUrqK5u8fBB/au5bFzNx0l1Yw31oP4tHy+pZaXcKVGSJHWFQgEuvRR++lM4/ng45hjYfvtFnlZVFaw1ZAC3HDyG+Q0LqO7di0H9q12QKC0mE2pJksrZ88/DfvvBv/+dJdG77rpYp1dVBYMH9C1RcFJlsORDkqRyNWECrL8+PP44XHEF3HVX1hZPUpdyhlqSpHL1+c/D174GF14IK6+cdzRSxTKhliSpXMybB6edBu+/ny063Hxz+POf845KqniWfEiSVA4efBA22ghOOQVmzswWIkrqFkyoJUnqzmbPhqOOgi23hPp6uOMOuPpqqPJ/4VJ34X+NkiR1Z9Onw2WXwUEHwVNPZTXTkroVa6glSepu3nsv6+Bx2GFZ144XX4QhQ/KOSlIrnKGWJKk7ufVWGDkSfvxjmDQpGzOZlro1E2pJkrqD6dNhjz1g7FhYcUV46CHYYIO8o5LUDpZ8SJKUt0Ih2+Xw+eeztnjHHAN9+uQdlaR2MqGWJCkv06Zl5Rx9+sAFF8BKK8Haa+cdlaTFZMmHJEldrVCA8eNhnXXgnHOysW23NZmWypQJtSRJXem552CbbeCQQ2CLLbK6aUllzYRakqSuctVVsP768OST8Pvfwz/+ASNG5B2VpA6yhlqSpFJLCSKyko5vfCOrl/7sZ/OOSlInMaGWJKlUPvwQTjkl2zL8/PNh883hT3/KOypJncySD0mSSmHiRBg9Gv7v/+CDD7KFiJJ6JBNqSZI60wcfwOGHw1ZbZTPU//gHXHklVPm/XKmn8r9uSZI6U11dtuDw0EPhqafgK1/JOyJJJWYNtSRJHfXuuzBhAhxxBHzuc/DSSzB4cN5RSeoizlBLktQRN98MI0fCT36StcMDk2mpwphQS5K0JN56C3bbLbusvDLU1mY9piVVHEs+JElaXIUCbL89vPgi/PrXcPTR0KdP3lFJyokJtSRJ7fXqq7DSSlnyfNFF2fW11so7Kkk5s+RDkqRFKRSy3Q1HjoSzz87GttnGZFoS4Ay1JElte+YZ2Hdf+O9/4atfhT33zDsiSd2MM9SSJLXm97/Pdjt89lm4+mr4+99h1VXzjkpSN+MMtSRJzaUEETBqFOyyC5x/PgwZkndUkropE2pJkhrNnQsnn5xtH37BBbDppnDDDXlHJambs+RDkiSABx7Iyjt+85sssS4U8o5IUpkwoZYkVbb6ejj0UNhqK5g/H+66Cy6/HKzJBR4AACAASURBVKr8X6Sk9vFfC0lSZXvnHZgwAY48Ep56Cr70pbwjklRmrKGWJFWeGTOyrh1HHQWf+xy89BKssELeUUkqU85QS5IqR0rwpz9lG7T87Gfw5JPZuMm0pA4woZYkVYY33oBvfQv22AOGD4faWlh//byjktQDWPIhSer5CoWsNnrqVDjjjKzUo7f/C5TUOfzXRJLUc73yCqy8MvTpAxdfnF1fc828o5LUw1jyIUnqeRYsgHPPzWqlf/vbbGzrrU2mJZWEM9SSpJ5l8mTYZx948EHYcUfYe++8I5LUwzlDLUnqOa64AjbcEJ5/Hq65Bv7612wBoiSVkAm1JKn8pZT9XH992HXXbJZ6770hIt+4JFUESz4kSeVr7lw44QSYMwcuvBA22QSuvTbvqCRVGGeoJUnl6d//zmakzzwTPvooa40nSTkwoZYklZf334eDDoJttsmS6HvugUsugSr/lyYpHyX71ycihkfEvRHxTEQ8HRFHFMeXj4i7IuL54s+BxfGIiPMj4oWImBQRG5UqNklSGXv33ays48c/hkmTYLvt8o5IUoUr5df5BuDolNI6wObAIRExEvg5cE9KaU3gnuJtgK8BaxYv+wMXlzA2SVI5eecdOOusbPHhiBHw0ktZf+n+/fOOTJJKl1CnlN5MKT1avF4PPAMMBXYGri4edjWwS/H6zsCElHkQWC4iVipVfJKkMpASXH89rLMOHHssPPVUNj5oUL5xSVITXVJwFhEjgA2Bh4AhKaU3IUu6gRWLhw0FXmty2rTimCSpEr3+OuyyC+y1F6y2Gjz6KKy3Xt5RSdKnlLxtXkQsDdwMHJlSej9a7wna0h2phcfbn6wkhFVWWaWzwpQkdSeFAnz5y/Dyy1lpxxFHQK9eeUclSS0qaUIdEX3Ikuk/ppT+XByeHhErpZTeLJZ0vF0cnwY03c5qGPBG88dMKV0KXApQU1PzqYRbklTGpk6FYcOgT5+sc8dKK8Eaa+QdlSS1qZRdPgK4AngmpXR2k7tuB75fvP594LYm4+OK3T42B2Y1loZIUndXKCTq6ufx+sw51NXPo1Dw+/5iWbAAzj4bRo3KFh8CbLWVybSkslDKGeoxwPeAJyPi8eLYscBvgBsjYh/gVWD34n13ADsCLwBzgB+WMDZJ6jSFQmLK9Hr2m1DLtJlzGTawH5eNq2GtIQOoqnLr60V66inYZx94+GH4+tfhe9/LOyJJWiyRUvnOotTU1KTa2tq8w5BU4erq5zF2/ESmzZz78diwgf245eAxDB7QN8fIysDll8PBB8Oyy8IFF8C3vw2tr7WRpNxExCMppZqW7nNbKUnqoPkNCxZKpgGmzZzL/IYFOUVUBhonczbYAPbYAyZPhj33NJmWVJZMqCWpg6p792LYwH4LjQ0b2I/q3nal+JQ5c+Doo7NZaYBNNoFrroHBg/ONS5I6wIRakjpoUP9qLhtX83FS3VhDPah/dc6RdTP33pv1kT777GwmulDIOyJJ6hQl70MtST1dVVWw1pAB3HLwGOY3LKC6dy8G9a92QWKjWbPgmGPgssuyrh333Qdbb513VJLUaUyoJakTVFWFCxBbM3Mm3HhjllSfeCJ85jN5RyRJncqEWpLU+erq4Pe/z5LoESOyDVsGDsw7KkkqCWuoJUmdJyW49lpYZx345S+zHtNgMi2pRzOhliR1jtdeg298A/beO6uVfuyxbBGiJPVwlnxIkjquUICvfAVefRXOOQcOOwx62TZQUmUwoZYkLbkXX4Thw6G6Gi69FIYOhdVWyzsqSepSlnxIkhZfQwOceSasuy789rfZ2FZbmUxLqkjOUEtql0IhMWP2fPssCyZNgn32gdpa2GUX+P73845IknJlQi1pkQqFxJTp9ew3oZZpM+d+vBPgWkMGmFRXmksvhUMOgeWXz3pL77ZbtuuhJFUwSz4kLdKM2fM/TqYBps2cy34Tapkxe37OkanLpJT93HjjrIvH5Mmw++4m05KECbWkdpjfsODjZLrRtJlzmd+wIKeI1GVmz4ajjoKDDspub7wxXHUVDBqUa1iS1J20mVBHRH1EvN/apauClJSv6t69GDaw30Jjwwb2o7r3J23RCoVEXf08Xp85h7r6eRQKqavDVGe7556sj/S550Lv3llrPEnSp7RZQ51SGgAQEScDbwF/AALYGxhQ8ugkdQuD+ldz2biaT9VQD+pfDVhj3eO89x785CdwxRWw5ppw//1ZBw9JUosipUXPIkXEQymlzRY11tVqampSbW1tniFIFaOtLh919fMYO37iQmUhwwb245aDxzB4QN+8QtaSeuUVGD0aDjwQfvUr6Ndv0edIUg8XEY+klGpauq+9XT4WRMTewPVAAvYCLJ6UKkhVVbSaHFtj3QNMnw5XXgk//zmsuipMnQrLLZd3VJJUFtq7KPE7wB7A9OJl9+KYJLWrxlrdVEowYQKssw6ceCI89VQ2bjItSe3WroQ6pfRySmnnlNIKKaXBKaVdUkovlzg2SWWisca6MaluXmOtburVV2HHHbONWdZZB554IluEKElaLO0q+YiIzwMXA0NSSutGxPrAN1NKp5Y0OklloaoqWGvIAG45eIw7KZaLQgG++lV47TU4//xss5YqO6lK0pJobw31ZcAxwCUAKaVJEXEtYEItCWi7xronK7st2V94AVZZBaqr4fLLYehQGDEi76gkqay1dzriMymlh5uNNXR2MJJUThrbBY4dP5Exp9/L2PETmTK9vnv24G5ogNNPh3XXhd/+NhsbM8ZkWpI6QXsT6nciYnWyDh9ExG7AmyWLSpLKQNlsyf7447DZZlkHj512gh/+MO+IJKlHaW/JxyHApcDaEfE6MJVscxdJqlhl0S7wkkvg0EOzrcJvugl23TXviCSpx1lkQh0RVUBNSulLEdEfqEop1Zc+NEnq3hrbBTbf0KZbtAtMCSKgpga+9z046yxYfvm8o5KkHmmRJR8ppQJwaPH6bJNpScp0y3aBH3wAhx8OBxyQ3d5442zDFpNpSSqZ9pZ83BURPwFuAGY3DqaU3i1JVJJUBrpdu8B//CNLpF99FQ47LGuNZys8SSq59ibUPyJbkHhws/HVOjccSSov3aJd4MyZcNRRcPXVsPba8J//ZB08JEldor1TFyOBi4AngMeBC4BRpQpKkrQY6uvh9tvhuOPgscdMpiWpi7V3hvpq4H3g/OLtvYpje5QiKEnSIrz1FlxxBRx7bLZRy9SpsOyyeUclSRWpvQn1WimlDZrcvjcinihFQJKkNqSUlXb8+McwZw7svHO2WYvJtCTlpr0lH49FxOaNNyJiM2BiaUKSJLXo5Zfhq1/NNmYZNQqeeCJLpiVJuWrvDPVmwLiIeLV4exXgmYh4EkgppfVLEp0kKVMowA47wOuvw4UXwkEH2cFDkrqJ9ibUO5Q0CklSy557DkaMgOrqrJ/00KGw6qp5RyVJaqJd0xsppVfaupQ6SEmqOB99BKedBuutl+1yCLDllibTktQNtXeGWpLUVR59FH70o6xGeo89YJ998o5IktQGC/AkqTu5+GLYdFOYPh1uuQVuuAGGDMk7KklSG0yoJak7KBSyn5tvnnXxmDwZdtkl35gkSe1iyYckLYZCITFj9nzmNyyguncvBvWvpqoqlvwB6+vh5z+H+fPhsstgww2zn5KksuEMtSS1U6GQmDK9nrHjJzLm9HsZO34iU6bXUyikJXvAv/896yd98cWw9NKfzFJLksqKCbUktdOM2fPZb0It02bOBWDazLnsN6GWGbPnL94DvfsujBsHO+6YJdITJ8I559hXWpLKlP96S1I7zW9Y8HEy3WjazLnMb1iweA80ezb87W9w/PHw2GOwxRadGKUkqauZUEtSO1X37sWwgf0WGhs2sB/VvXst+uQ33oCTT4aUYPhwmDo1u923b4milSR1FRNqSWqnQf2ruWxczcdJ9bCB/bhsXA2D+le3flJKcMUVMHIk/PrX8PTT2fgyy3RBxJKkrmCXD0lqp6qqYK0hA7jl4DHt6/Lx0kuw//5wzz2w9dZZ94411+zaoCVJJVeyGeqIuDIi3o6Ip5qMnRgRr0fE48XLjk3u+0VEvBARUyLiq6WKS5I6oqoqGDygL0MHfobBA/q2nkwXCtmiw4cfht/9Dv71L5NpSeqhSjlDfRVwITCh2fg5KaWzmg5ExEhgT2AUsDJwd0R8PqW0mCt9JClnzz4Lq60G1dVw1VUwdGhWMy1J6rFKNkOdUrofeLedh+8MXJ9SmpdSmgq8AGxaqtgkqdPNnw+nnAIbbABnFecMNt/cZFqSKkAeixIPjYhJxZKQgcWxocBrTY6ZVhyTpO6vthY22QR+9Sv41rdg333zjkiS1IW6OqG+GFgdGA28Cfy2ON5SEWKLW49FxP4RURsRtXV1daWJUpLa66KLYLPN4J134Lbb4LrrYMUV845KktSFujShTilNTyktSCkVgMv4pKxjGtD076LDgDdaeYxLU0o1KaWawYMHlzZgSWpN4zbhW26ZzUhPngzf/Ga+MUmSctGlCXVErNTk5ligsQPI7cCeEdE3Ij4HrAk83JWxSZWsUEjU1c/j9ZlzqKufR6HQ4h+IBDBrFhx4IOy3X3Z7ww3hkktg2WXzjUuSlJuSdfmIiOuAbYAVImIacAKwTUSMJivneBk4ACCl9HRE3AhMBhqAQ+zwIXWNQiExZXo9+02oZdrMuR9vVrLWkAGtt4SrVH/7GxxwALz5Jhx5ZDZLXeX+WJJU6SKl8p2JqqmpSbW1tXmHIZW1uvp5jB0/kWkz5348NmxgP245eAyDB7gtNpDVRx9xBFx7LYwale18uNlmeUclSepCEfFISqmmpfucWpEq3PyGBQsl0wDTZs5lfoN/JPrY3Lnwz3/CiSfCo4+aTEuSFmJCLVW46t69GDaw30Jjwwb2o7p3r5wi6iZefz1LoFPKekm/9BKccEK2YYskSU2YUEsVblD/ai4bV/NxUt1YQz2of4UmjinBZZfByJFwxhlZ9w6AAQPyjUuS1G2VcutxSWWgqipYa8gAbjl4DPMbFlDduxeD+ldX5oLEF1/Munfcey9su22WWK++et5RSZK6ORNqSVRVRckWIBYKiRmz53f/ZL1QgJ12yjp4XHpp1ls6umGckqRux4RaUsmURUu+yZOzWei+feHqq2HoUBg2LPsi8MG87v9FQJKUO2uoJZXMjNnzP06mIesest+EWmbMnp9zZMD8+dmiw9Gj4be/zcY22+zjZHrK9HrGjp/ImNPvZez4iUyZXu+GN5KkFplQSyqZbtuS7+GHYaON4KSTYI89YP/9F7q7W38RkCR1OybUkkqmW7bku/BC2GKLbAvxv/4VrrkGVlhhoUO67RcBSVK3ZEItqWS6VUu+QiH7+YUvZNuHP/10tgixBd3yi4Akqdty63FJJZV7l49Zs+CYY6ChAa68sl2nlMViSklSl2pr63G7fEgqqVK25Fukv/wFDjwQ3noLjj46m6WuWvQf5uzNLUlaHCbUknqeujo4/HC4/npYbz247TaoaXFSoVW5fhGQJJUVa6gl9Tzz5sHdd8Mpp0Bt7WIn05IkLQ5nqCX1DK+9lm0VftJJMGwYTJ0KSy+dd1SSpArgDLV6lEIhUVc/j9dnzqGufp4bcVSCQgF+9zsYNSrboGXy5GzcZFqS1EWcoVaPYWeGCvT887DvvnD//bD99nDppbDaanlHJUmqMM5Qq8dwd7sKUyjAN74BTzwBV1wBd91lMi1JyoUz1Oox3N2uQjz1FKy5JvTtCxMmZPXSK6+cd1SSpArmDLV6DHe36+HmzYPjj4cNN4SzzsrGNt3UZFqSlDsTavUY3Wqba3Wu//0vS6RPPRW+851ss5YezMW1klReLPlQj+Hudj3U+efDkUfC8OHw97/DDjvkHVFJubhWksqPM9TqURp3txs68DMMHtDXBKScLSjWvm+9NRxySFY73cOTaXBxrSSVIxNqSd3Le+/BPvtk7fAANtgALrgABgzIN64u4uJaSSo/JtRSN1LxtbO33gojR8LVV8NnPwupwn5/XFwrSeXIhFrqJhprZ8eOn8iY0+9l7PiJTJleXxlJdV0d7LEHjB0LK64IDz0Ev/41ROeX7HT3Ly0urpWk8hOpjGeAampqUm1tbd5hSJ2irn4eY8dPXOjP/cMG9uOWg8cweEDfHCPrAq+/DhttBEccAcccA336lORpymXBX6GQmDF7votrJakbiYhHUko1Ld3nDLXUTVRc7ewrr8Avf5mVdQwdCi+9BMceW7JkGspnwZ+LayWpvJhQS91ExdTOFgpw0UWw7rpw7rnwzDPZeP/+JX/qivvSIknqEibUUjdREbWzU6ZkbfAOPRS23DJrhTdyZJc9fcV8aZEkdSlrqKVupEfXzhYKWfL89ttwzjkwblxJFh22HUJ51FBLkrqftmqo3SlR6kYaa2d7lEmTYK21oG9fuOYaGDYsa4mXA3fTlCSVgiUfkkrjww/huOOy7h1nnZWN1dTklkw3csGfJKmzOUNd4Xp0iYHyM3FittvhlCnwgx/AQQflHZEkSSVjQl3BrCdVSZx3Hhx1FKyyCvzjH/CVr+QdkSRJJWXJRwUrl568KhMLiq3nttkGDjss6+BhMi1JqgAm1BXMnrzqFO++m5V1/OhH2e0NNshmqZdeOtewJEnqKibUFcyevOqwm2/OWuFdcw0MH57teihJUoUxoa5gFbGRiEpj+nTYdVfYbTdYeWWorYVTT+3yvtKSJHUHLkqsYPbk1RJraMg6efzmN3D00dDbf0okSZXL/wtWuB65kUg3Uqq2hLm0O3z5Zbj0UjjtNBg6FKZOhX79FnnakrCdoySpnJhQSyVSqraEpW53+Klktl9vqi4eD7/4RVbS8d3vZnXTJUymbecoSSon1lBLJVKqtoSlbHfYmMyOHT+RMaffy5EnXsuHW4yBww+HrbbKWuGNHNnh52mL7RwlSeXGGWqpRErVlrCU7Q6bJrORCpx81fF89OH7vH/JFSyz3w+7ZNGh7RwlSeXGGWqpRErVlrCU7Q7nNyxgmWefpm/DfFJUccQ3jmH7H42nfvc9u6yDh+0cJUnlxoRaKpFStSUsWbvDuXMZeMoJ3D7hSPZ9+BYAnvrsGiw1bOUuTWZt5yhJKjeRyngjhpqamlRbW5t3GFKryqbLx/33w777wvPP895e32PPz+/Gsx/2ym1BoF0+JEndTUQ8klKqaek+a6ilEipVW8JOfdxzzoEf/xg+9zm4+26W2XY7/rCEyWxnJcK2c5QklZOSJdQRcSXwdeDtlNK6xbHlgRuAEcDLwB4ppZkREcB5wI7AHOAHKaVHSxWb1JqKmhltaMg2ZPnSl+Coo+CUU6B/f6pgiZJZ291JkipVKWuorwJ2aDb2c+CelNKawD3F2wBfA9YsXvYHLi5hXFKLmreMGzt+IlOm11MolKYsqlBI1NXP4/WZc6irn1ey5/mUGTNg3Dj40Y+y2+utB2efDf37d+xhbXcnSapQJUuoU0r3A+82G94ZuLp4/WpglybjE1LmQWC5iFipVLFJLenKhLC15L2hodBqkt3hBDwluPFGWGcduO46GDEiG+sk3andXW5fVtRj+BmStDi6uoZ6SErpTYCU0psRsWJxfCjwWpPjphXH3uzi+FTBujIhbC15v3bfzfjO5Q99qmQC6Fg5xfTpcOCBcOutsPHGcPfdsP76nfo7RQTDBvZb6DUcNrAf0UXt9hpZeqKO8jMkaXF1l7Z5Lf0L1eJ0QETsHxG1EVFbV1dX4rBUSbqy/3Fryfvb9fNanCHv8Oz5ggXw0ENwxhnw4IOdnkwD9Ao4fdf1F2p3d/qu69Ori/MPS0/UUX6GJC2urk6opzeWchR/vl0cnwYMb3LcMOCNlh4gpXRpSqkmpVQzePDgkgarytKV/Y9bS96b/w+7cYZ8iWbPX3oJfvYzKBRg5ZXhxRfhmGOyhYglUFVVxdX/ncrxXx/JDftvzvFfH8nV/51KVVXX/jPTnUpPVJ78DElaXF2dUN8OfL94/fvAbU3Gx0Vmc2BWY2mI1FWqqoK1hgzgloPHMPFn23LLwWNK9ifelpL3S763MTc/8tpCxzXOkC/W7PmCBXDuudliw4svhmefzcb79fv0sZ1oUP9qjvryWtz8yGu8N/cjBvWv5pc7jWRgvz4lfd7m3GlRHeVnSNLiKtnGLhFxHbANsAIwHTgBuBW4EVgFeBXYPaX0brFt3oVkXUHmAD9MKS1yxxY3dlE5a96ib2C/Pjxf90GLdZvQzhrqp5+GffbJyjt22gl+9zsYNqzLfqeGhgJT3q7ngD88klvtaXeIQeXNGmpJLWlrYxd3SlRF6259p9uKZ5GxFgpZB48ZM+D882GvvaCLFwTW1c9j7PiJn1qYeMvBY7pko5bGROicu6aw68bDGdS/mhUH9GXlZfvRu3d3WTKictDd/m2QlD93SpRa0B1noZZoh8BHH4WRI2GppbJ2eMOHQ07rC/KuPW26mOyfk7MlGl2Z0KvncLdOSYvDKRtVrM5eyd+evrVL2tu2pb7Vz02dTjrmGNhkEzjrrOzAjTbKLZmG/GtP807oJUmVyRlqVay2kq/F/XNve2a7OzIj3jz5H/rEw/Q/cxwx43XYbz847LBu8SfqxsWWzX/HUnRKaUljQt+85MTFZJKkUjKhVsVqLfnq07tqsRPf1ma7m5YatOeY1jRN/vd5+BaOv/cKXlnus9Td+jcG77xji8n6hB9tytJL9eajhkKXJdhNO6XkkdjnndBLkiqTCbUqVmvJV++qWOzEtz2lBh0pR6ju3YtVl+nDK+9/xAOf25BLPxjLDd/Yl+u32x74dLI+eOm+TH//Q8ZdOanL68PzrD3NO6GXJFUmE2pVrNaSrzdnzV3sxLc9pQZLXI5QV8cKRxzJ7fMa2GnTA5jCCCbseuhCM6/Nk/UDt1mdY26atESz4eXOxWSSpK7mokRVtMbka+jAzzB4QF+qqmKJFta1Z5fFxd6JMaWsa8fIkcRNf2LAeuvw54O24P6fbssN+2/OkGU+SRqbx7xcvz4VszhvSRd6SpLUWZyhlppZkjrc9pQaLFY5wptvwv77w1//CptuCldcASNHMaOV2u7mMc+Zv6AiFud1x9aHkqTK48YuUgs62jFjSc5ves5S77zN8tuMIX7yEzj8cOjVa5GbpjQ9v191L6a/P6/HJ5p5byQjSaocbuyiHq+zW8Z1pA53SWZNC4XESw8+zsO/+A3Hbf5dhi7fn8sfeIzPrzr443MWtaixeczL9avu8Yvz7DstSeoOrKFW2Wtp05Mp0+vbVUtbivrbxd4wZsEC5vzmDIZtswVff/CvrDbjdabNnMu+Nz610DmLW9vdUn14T5P3RjKSJIEJtXqAJd3xsCOJeFsWa9b0qadgiy1Y+rif859VN+DL+47nxRWGt3jOYi9qrAC+JpKk7sCSD5W9Jf2zf0c2WmlLu9vjLVgAu+4KM2cy6/cTOOmtlZn+3ocLndOn9yffee2x/Gm+JpKk7sAZapW9Jf2zf6nqbwf268Ml39u49VnT2lr48EPo1QtuuAEmT2bB7ntw5u4bLHTOmbutT+9miWEllHEsLl8TSVLenKFW2Wtvm7vmCxf7VS/hRittKBQSz9d9wHl3P8fxXx/JoP7VrDigLysv24+quXPg+OPh3HPhlFPguONg9GgA5s6cwxl3TuH4r49kuX59eG/uR5xx5xQu/M6G0H+Jw5EkSV3AhFplrz1/9m+t88aEH23KuCsfbne/6UVpWkbyz8lvA1mS/tdR81nuiEPgpZfgoIPgsMMWOq+6dy/qPpjHAX945OOxlpL7zu5m0h55PGc5xyVJqjwm1OoRFtXm7p3Z81qsl/7zwVt2av1tS2UkX7vzjyz38ythjTXgvvtg660/dV57ZtlLuYlJ0+S0T+8qelcFc+dn1z/4sOFTXzry7mfthi6SpO7EhFq56MrZxUIhMWdey/XSHzUUGDrwM4v1WG3F3XRBYu8FDTT06s2zG45hzuiBfOb/ToF+/Vp83PbMspdqEWVLyemZu63PGXdOoe6DeZy52/oMXrov02bO7bTn7KhSvRaSJC0JE2p1ua6eXZwxez5T35ndYr00wCszZtOnVxUrLt2X3r1bX6fbnrgH9a/myh1X5dXv7kt9A/x23PEcO243lhoyABbxuy1qlr1UiyhbSk6PuWkSx399JAf84ZGFrnfWc3aUG7qUL0t1JPVEdvlQl1vSvtFLan7DAs6/53lO33X9hbpoXPzdjTnpL0+z9Zn3sccl/+PZ6fV89NGCVjd6WWTcKVF13bWsue1mbP/sf/ny2C9y4/6bs8xSvZkxe36H+1svqpvJkm5S01pyuly/Pp+63vw58+KGLuWpVL3fJSlvzlCr0y1qBqozZxfbM9vVuODvrH980kVj0NJ9OePOZz5eODht5lwOvOYR/rjvZux9+UMfz0Bf8r2NWaF/NVVVVW3H/cYbsP/+8Le/EZtvTuGyy3lt0HD2u/TBRc7Ct3fGrq0667Zmz+H/s3em4VHVZxv/ndknk5CEkAAaVKSARAQh7L5WlBa1olQBtWyCsou0FhC6UK3RvgpSfa1iAAsIArJptbhRF2zLIhoRpFFEBE3YEkL2TGY75/0wOSczmRlM2BLo8/tCmMyc858zXBf3eeb+3zf1tqnopCc7KXH7jJ+rvAHj56ZQnFLfZBehaSFWHUEQLlQUTTt/JwM9evTQPv3008ZehhBCfWwRheUebl+wJULANfQ/1fpaR6I9b8V9vbj+qY8ijvn6/ddwtKzaiK7bkJPHkMw2ZG3MZdW43gyvEdsR6644AT17wsyZMHUqhVX+er3HhtpfYonvWNf0janXcKzMc9Lj/5CHevHoHrRsZsftbVpf0Yt14PzjUHEV1zz5YcTjW2Zd36C9DIIgCI2Boig5mqb1iPY7mVALZ5T6TKDO1HSxvtOuaBv+/AE16lQ2gBqg5gAAIABJREFUwWHh/lW5xrqeHNKFZg4L+cVuHnszl4WjMpm4Iof8Yjd9AkU8/+0Okp3XQUJr2L8f7MHzev2eek3hGzqxi+WzjjU9d3sDEcd/+h97eeS2zmiaZgjR0Oujp3w8N7xbuFBtYnnYP+Q5F5oe9W4RFQRBOM8QD7VwRqmPnSNU4G6ZdT2vTbnmlDYkNsQ6UrdNLzXeTvbI8DbDhSMzeeLtL8PE56wNu3FYg//Zb8otoIXLxmsTevNF3Oesfm4iya+8TMnne4IeUHutuKuvx/dM2V9inS+gaWHH79YmiXv6teXOhdvCPKwQvNGxWcz4/CoBFVonOqV5UDij6DfTMVtEBUEQzlNkQi2cUeo7gToT08XTmXZZLCauaJnA2ol98QdULGYTJgXDU62TX+ymwuM3jm39zx6Sp01GycnhnxnXMKP/BGzvF7H44vKItI/6TOHP1MQu1vkc1vDjT+rfjlkbdkfN4y6q8J6z5BXhv5P6xEMKgiCcj4iH+r+Us+U/PZeReKqqcbCoku+KqoizmanyBrg0JY7LUlwN8h/rxPIhzxmUQdbGXBaP6MYVP+mLdqKYPwycxMsX9wRFMZ4XzR99svOpqkaJ28uRkmomvpxz2tcr2vmAsM9j/aS+DM3eFvHafz50PcMXbz9tX7sgCIIgXKiIh1oI42yK3nM9gfL4Vea8vifsfQQCKkfLvPgCqpEvbTIpMcU31KZgrBrXm8fezGVTboGR8tHqq128NrY7KS0SUdas4Wh8c15evDtsHdFsGiebwod+BqnxdrIGd6ZtCxdxdjMtXKdms4h1vtDPQ1GUqBNxs4LkOguCIAjCKSKC+r+Qsx1dda42i0V7H0//Yy+/HNAhbOKbPTKTi5PtHCurDhPf84Z2obnLypHS8BSMhaMyyRrcGVNVFS2e+APKX/4Cjz4Kv/89dO2Kubw6qii1hpTC/NB0OnTt+cVuxi77xJgIn+mbj9DPQ1W1qNYQp002iwmCIAjCqSKC+gKgofaNC6VlLtr7GJLZxhDTUJsvvWZCH2auD/cOz1y/mzUT+hhTYj2juqDMQ5ucf9Ns2lQ4eBDuvx9++UvjHBaTwryhXYzj6eLccpIYurrfADTWZxDrGwTgpJ5viagTBEEQhNiIoD7PqSveBmak8ftbMjCblJMWnZwv08iTCblo7yPFZYsqVP2qFvPx1Hg7M27saGzWe2j361z/9mK0Dh1Q/vlPuPbasNe5vQHmvlNbElPi9jH3nb08N7wbuOr3DUBjfgb1sYaEXutzXRUvCIIgCOcbEpt3nhMq3vRItOEvfnzSWt/zJbrqh2qKo72P1AR71Pg4i0mJ+rjVbGLagPbM2rCbY8fLAHit9dWs6P8Ljv/r4wgxDbXNixNX5HDXou1MXJFDYYXHEMP1mT43xc+gbrSgLpbPdVW8IAiCIJxvSMrHeU5o89jCUZlkbcw9aVJD6MRXURTMCphMpib5FX59GhXrTrAtZo2vjlRE2DFS4m0UV/qYvm5XmFe6ZYKdyrzDfHHHPfjNZn5160zjXLHa235oYlvfJsjGtlHU9/zSbicIgiAIkvJxQWO1mAzrQJLTetLJ6Jn66r4+QuxMiMX6lsSkJtiN81V5VNKT43j6zqtRNY0qb4DUBDsz1wVTOeYMyiAtwU6i08oTb+USv+4V5v57Ka0rKnnmml+ApkFNEoY+cY72Xk6WZFLfDOrGbPpryL+F88kiJAiCIAiNgQjq8xhV1aio9hsb5ErcvpMKnzOR7lE37m3agPYRcW8NFe6qqnG80kO1L4BZUXDazCQ5bYaQS423M6l/O5KcVqq8AZw2c8w1hU6fW7hsmEwmzCYorAhWgU9ckcPCUZn8MXsTD74yl+sOfMaXl19F9dqFvLHTDXUE8MneS6xr1tjlFfW5mWnIv4UzVRUvCIIgCBcqYvk4j9GtBbrgvCjRgQZMWflZhPgDyC+p4sdzN0ccpyFf3YeeM3QjX+i5iiq99bI8QHQxPG9oF9KTndgtZqp8AVRV44m3vzSyoUPFuapqHC2r5nCJm6JKL9mb97MzryTsfH6/yuFSNwXlHooqvaQnOxnzp7/zt+W/ZmHvO1jR/RauvqQ5f/lFNxSFMBFaX/uG/l4aOpU/07aP+t7MNNTG0dj2FEEQBEFobMTycYGiWyL0yStAtzZJrJnQByAsEm3vsXKOlkbPT27IV/f6OecMyohaYa1PZesbCRdtUrp0ywGm3tCeKSs/NpJLZt/cicn9f0RBuYen/7GXx2/vQorLFiEenxzShafe3cvOvBK8/qAY31dYwfjln2L7Zh+TvtxE+uoX6darI/2di/CZrUBwgl3tV2nustLcVSuU6/te6jO5r8uZTs/Qby7qM3luqI2jMe0pgiAIgtDUkZSP8xhdFIWip02EJjUcrwwWlzz7/j6eHNLltJIl9HOezK8dbV2xxFqsLGl9yq4nl4xesoPbF2xlQ04es2/uhNvr52hZNU//Y2+YeJy1YTeT+rczzldU6WXS0u3c+vZy3l76AANzNvH4M6/zu1syaNmimbG254Z343iFh1K3j4LyalRVQ1U1o1nwh96LfmOgT+7nvL6H/k9t5o4FW6MmrYS+5kykZ+ji/HCJu143AE0xZUQQBEEQzldkQn0eUx9vq6pqVHlqJ9lPvVubn5ye7KR1orNB01D9nCebdtfXc6uqGgFVO2mW9KT+7YxJeKi4Dp1IF5Z72ZlXAgTFo37+FJeN4//axvPPTqHzsf281aEfD/90MoXWZEyKwrKxvSipClaUe3wqM0ISQBaP7oHdYuKJt7/kySFdIqwtdd9LfSb3dSe8Z7LcRRfncwZl1Gvy3Ng+b0EQBEG4kBBBfR5TH1FUVOmlsNzD0jE9ibOZKXH7yN68n8IKzynVXOvnbNnMzsKRmWEV37rQrK9YK6r08tibucwf1jUszq65yxY1uSRUXEPtRHrOoAzD8hK8SXBgMSkUlblpPn4s5soTPHnfY1z94H08V7OxMc5uIqBqPP7ml8wd2oWxyz6JEMFZgzuzKbeAwnIvcwZlkOKycVGSk1bNHDGTMH4oaSXaa85EeoYuzrM376/XDQCIjUMQBEEQzhQiqM9zfkgUqaqKqmnMeX1P2Ka/1AR7vb7ej7UZrbnLTpLTFiGadR+vL6BiNZtIi7dTUu2nsLyagAaaphnP9foDhmB94o6raJXoqEn5MBkiOzS5JJZY1VM9Uly2YITetm3sbnk56/5znD8segn/RRfxY3NcWDb14tE9aJVo46GbOlLq9kU9blxNmsjOvBJDsG+ZdX3Um5D6TO5jveZMpGfo4nxnXonxLcTJbgAEQRAEQThziIf6AiegYUx/ISgUZ67fjctmqVcCxcmaCus266mqxlfHyrlz4Taum7eZOxdu46uCCt7clc83hZXcuXBb2HGctloROH/T1+wvrKSo0kt5dYAPvjzK2ol9uTo9kYWjMklPdhriOpSBGWmgQdbGXEY/+wE7bhtF+qCfUvm/c7mnX1se/d6MPS3VENP6NRi//FPcXo2Z63cbyR+hpCc7SWtmZ+GoTLq1STIeO9mmvY4tE+jaJpGFIzPr5U0OneRvmXU9r025JmJDoqpqFJZ7OFRcRWG5x7j2dR9PdloNT/TOvBKyNubisltETAuCIAjCOUBi8y5wTqflrrDcw+9e282QzDYkOa2UuH1syMnj8du7RJ2KHy5xc+fCbRHT2VXj+zB88faIx1+d0o+iCi9P/2Mv9/RrG2ZTeGFkJlekxaMoCgdPVJJ3wk2LeFtELOCK+3ox6q87uDzn3/zp3ee4qOw4G/r9nORn5vHIB9/xxB1X0aZ5HNfN2xyx3n/O7M+P522mW5ukiAjAeUO7MPedvRRWeHh+eHfe2n2In3dvU68EjjMVMRcrBaR9aryRXFL38WK377z1REs0nyAIgtCUkdi8/zLq1osPzEhjU26B8fv6+nRVVY0Quk8O6YJfVVFVzRA7+vl8ATWqdULTtKiP+/wqHVsm8MhtncOEeH6xm8kv57BqXG+sFhNjltb6m7u1SSJrcGfaNHdyvMKLSVEY9PYKZn+0jG+apzNsxJPkpGfwWsvm5Bd/xUVJTlQtcuNjerITi9kU1SaREm/jaGk1s2++ghK3j+c/3Mfvbsng4npu4DxT3uRYKSBrJ/Y97YKepsaZjhAUBEEQhHOJWD4uMOraNO5cuI1pAzoErRHUPx5NVTX8qhZ1E2ClJ2BYP0LP5/WrUa0TZpPC0jE9WTOhj2Gh0EW9yaTEFNzlHj/VvvAkjJ15JYxd9glWs4k2cSZ8AY093a/lL33v4paxz5KTHky5aBFvZ+mYnhRVeCmp8kbEBS4cmUlavD3CJpEUZ6WgzMPM9bu5a9F2sjbmck+/tlR4/BS7faf9+TSEWCkg/hg3LqeSDtJUOJMRgoIgCIJwrpEJ9QVGNGEy6eUc1k3sy8O3avX6Kl0XyV5/dOFmViCvtJqWzewEVIz85VK3jxX39eLg8SqefX8fhRUeskdmUlbtC9sU+cKI7rSIt5HsDJaqOG1mI4VE1YJReg6rmQSHlWNRNvl1NVfh/MVdmOw2TrywhMlTbmNWy8vwhNg1pq3eSWGFh/nDumKzWHj0718acYFV3gCtkxxYLCbap8azdmJffAEVm9mESYGSKh9zBmUYrYuzNuxm6Zie50yw6hN/4KST9TORDtJUOJMRgoIgCIJwrpEJ9QVGLGHi8au0TnQaZS8nQxflR8uqo06cvy6oYM7rezhSUo2qqkaZya/WfM71T33EnNf3kPXzzrw6uR+tE+1MWJ4TbudY+RlfHa1gX2EFfr/KsTIPc17fwxNvfwXA7Fe/YGj2NoYv3o6qaTw3vFtwHZrGhP0f8erz40n68B+caJ/B8XIPL209wJxBGbz36x+TNbgzc98JNiXmF7uZvm4XiU4rhRUeJq7IYfq6XbRKdJDktOH3q+wtqN1EOWzhNr49Xsnjb35J1sZcZtzYkW5tksgvdlPlDZwTwRo68Z+6aqdRw65f+8Wje4RN1kMfP59LWRpSBiQIgiAITQ3ZlHieUN8NW4XlHm5fsCViepk1uDOdL06sl8f2WKmbz/NLuSjRgarB/atqNwGGVnunJztZO7Eve4+WGxPo0HOum9gXb0CNuiFwzYQ+TF+3i7UT+xr+6YWjMsnamBtxnCfuuAr/93l0e2Q6if/eTFnPPgy+eixJXa/k4dsycHsDzFy/m/nDunLXou0R59oy63psFnPYtQPIL65i+IsfR5xPz7XWf87amMuqcb1JT447o35ev1+loMITFjFY7PaFfX7d2iQxbUB72qXF47TWfu4X2gY+8VALgiAITR3ZlHie0xCxkeKyRRSu6CL4ueHdIo5bV5QBHK/0GsJ2YEYaK+7thcmk8NXRckNMQ+2Gw8taxEWdirt9Ab4rqopqTyipyX4O3cgYK2faYTUz+bVc/r4nl/LH5pF18bUc+Oo45JXwxzdy+e3POrF6fB/j2NGsEHVvJArLPRSUe6KeL6nGipJf7OZHqfEsG9uTixrYKPlD+P0qXx0rZ1LI55Q9MpNWzexRPeNbZl0f9h7OxMbHpiTKpblREARBOJ9pFMuHoigHFUX5QlGUzxVF+bTmseaKovxDUZR9NX8mN8bamiI/tGErNJO4qNJLqyQ7T995NR9Ov45lY3uiaRqpCbawr89jZUwfrwxaI/RzbcotYNSSHZhNClkbcw0xDRhf0VtMpqhf139XFPRS190Q+OSQLmRv3k96shOrufa1dXOm2xXlMe9ff6W508Jzv74Z/1dfs+Lqmxn9P+2M5xVWePAGVKat3sm01TsjzrV4dA9MJo3vT1RyqLiKE5XBLGevPxAzf7qkZvNherKT709U4QtoZ1zYFVR4DDENtV53T4yNnWfa+vBDGeONQd1ccxHTgiAIwvlCo1g+FEU5CPTQNO14yGNzgROapj2hKMpsIFnTtFknO85/i+XjUHEVU1ftZFL/dkYedPbm/Tw3vButE50R0+vV43tT6vZHTD+vaJmAxRK8h4plDVk1vjc/nrs5Yg0f/+YGTlT5ws6TPTKTZ9//msJyLw/d1DGsifCFEd35w+v/YWdeCd3aJBlrT2tm59drdlFY4WH5vb1IjrNQVh2gsDxofbCYFGaszuGWd1fyqy2rURLi+ea1d3j6e4XHbr8KBQVVVVEUBW9ApdqncrTUzfxNX7Mzr4SBGWk8dFMnLCaFOJuZsmqfEbuXnuxkwYjupLhsKIrCS1u+5ccdW8bMn9Yn+3pN+5mMpPuuqDJmNnalN3DWrQ+xPv/zOXpPEARBEM4m54vlYzDQv+bnl4DNwEkF9X8LTps5QrDOG9oFp81MidvL0dJq5g/ragjtbwoqwzzN+vQzVCzF2rxoVpSotgmFoJ1kzYQ+BDSwmhQefmOPkW899529ZA3uzOWpLvJOVFHtUyms8AC11d3pyU5endyP54Z3Q9U0LGYT+cXVTA4palmdaeWdldNx5u7hyI238ZvrxvHNxxWsGteb5k4bxW4fflUjoKo89mYum3ILjKn3S1sPMOX6H3G01E2yy4aiEJZhnV/sZsrKzwxfdPbITP7+eb6RP92ymQO7WTHyp0PtLfVJm4jmidZvYOpijZHUYTGb6Ngy7pStD/W1cUiqhiAIgiCcORpLUGvAJkVRNGChpmmLgJaaph0B0DTtiKIoadFeqCjKBGACwCWXXHKu1tuo+FUtojp75vrdvDq5H4XlnrBIuieHdKGZwxImlvQJcZXXz5ESFZMSPOYrE/rwak4+AzJaGnFy8Q4z2SMzw6bbzw/vRkGFN2LiXVhemxGse33XT+qLxWzi4mQHL4zIZPLK2tcsv7dXmH1l6ZieYcL/cFEF2i/ux4SXCbf/jk0d+kIpgBu7xRTRDvjkkC4UlnuNaLuV43oHC1le/cIQybE80vpNxtqJfdG0YJxgstPK4VJ3WFU71M9yEcsTHfqtQKjYddkjr3N2TTZ2NH90fYRyQ7z2eqrGhRS9JwiCIAiNRWMJ6ms0TTtcI5r/oSjKV/V9YY34XgRBy8fZWmBTwhcjD7rarxqbD/XHZm3YzbKxvQyxFK1W++k7u/Knt74iNcHGAze0D5sQLxyVic2ikDW4M3E2MyVuH+XVfma/+kXExDtrcGfGLvvEWFN6spOiSi8dWsZz58LtpMbbyRrcmctauLBbTCjA6CU7jOPE2czkF7vpnv8luS3bUm11MP622Twx7WY2rf4y7LgBjQgf+awNu41EjvxiN6qqGWI4yWk1PNJ1EzNS4m0sHJVJ9ub9aJpmVLAXlFfz2Ju5PDmkS9j1Wjgq8wcj6WJ5otdO7Btsa4widleO783ye3sBoGpgs0ROklVV43ilhypPgAPHK41872hCOZbXPpqNI8VlY/HoHhHi+3yO3hMEQRCExqJRBLWmaYdr/ixQFOU1oBdwTFGU1jXT6dZAwUkP8l9ErGmiWSG6bcOEMf2c1L9dRNvhg2t38cQdV1HpDRhiWv/dxBU5rB7fJ0wor5nQJ+p5LmsRZ6wr1Hbxi16Xkl/sJr/YzdhlnxjRd8kum1GuUuL2oZSV8eePFnHH9jd46tqRPNfvbio7dCLxojTSkw+GCT01RptiWo1Q1EW3/pwSt48NOXnMH9aV6et2kRpvj2qbcVhrLRnVvgCbcgsoLPeGrbN5PSwXsWrX/QEViC52Ryz+2CiQmdS/HSkuG2ZFoXVNokg0Ea77uqMJ5YbYOC7kVI2mlF4iCIIg/HdwzgW1oiguwKRpWnnNzwOBR4E3gHuAJ2r+fP1cr62pEm2auHBUZkwf7qFiN8u3HTQaAKOJrFaJDooqvFF/p2pa2HH19I265ymv9rNmQh+8AZWAqnG8wsvMGzvy0PovIo55aUocxVU+I45v6LHdPP7u8/Q8ephX/mcIf+3x8xp7SXfW7PiOpWN6Uur2cVGSk1bNHEbJTN016BXjl6bE4bTW3nhkb97Pw7dloABZgzvTtoWLkX/9OEzQzly/mzUT+qCqwRQP3T+ue771c6yZ0Cfq5xIq3CwmhYEZaYanXH+txRwU7LHE7kWJjohvEPTpczQRHjqVryuUG2rjOBPRew3lbItdybMWBEEQGoPGiM1rCfxbUZRdwA7gTU3T3iEopH+qKMo+4Kc1fxdqsFtMZA3uzJoJfcga3BlfQOWv/9pP9sjMiEi6+Zu+ZlNuAZqmYTEpUWPY7BazIZzr/s5iUsIa+jbk5LFgRPew8ywZ04OkOCsev8rB41XMXLebGet24Q0EI/ogaLFYOCqT9ZP6AgpTaqbhU7at5allv+VwwMLRt94nfckLrJh2PSvH9cZiUhje9zIs5lqxp6oaZoWISDy9YnzO63vw+FScNhMLa65H+7R4Ulx2fAENb0CluKr25kFf1/xhXVE1KHEHveBOmzmimVDf/FmXurFzdy3azgMDOjAwI814re6JhthNgA6rOeIbBD0SMZYIT3Jaowpl/carqTYonouovh+KmBQEQRCEs8E5n1BrmvYt0DXK40XAgHO9nvOBokpvmPcYahv9nn3/a9ZM6MOR0mqKKr1hLYY2ixmzCeYN7RJmdXhySBeyNv6H8ddeztKxPRkbEiu3eHQPUl02VA1evq83AGYTvLztoLHR76IkB8WVPu5d9nGEFWHiihxWjQ++7p5+bQ2xuH5iHwoKy8Bi5d32fbH7fTzf907eyeyJo9JLUpyNX73yOYUVHpaO7UlRhRcFqPIGqGrup0W8jZe2HqgRvHF4/MG4PKgRTSs+ZdnYXry09SALR2Zis5r4pqCCOJsZm9lEcpyN9GSnUZMe5pEemUmS00aSM5j0ofvHq7wBWjYL1pRH+0zqCrfJL+ewZkIffn9LBpY6KR/RvmV4YWQmVd7YNo1YE+cqbyCqUG7qNo6GeLxPlTOZXiLWEUEQBKG+NKXYPCGE0P/MAzH8w0lOK5tyC3j41itp7rIFNw/efAVV3gCXpsQZgiuYTd2HY2Xhojv3SDlrJ/aJSLqom6Yxf1hX+l/REpOiUO0LEFCJuhly3tAulFX70TR45NYrOVQSjPNTDh8mY+oYFuRXMO7mGexv0Yanrx1hlL84rCZmrttlZFYfL/dEeJ2bOaw8dNMVFJZ7+MXi7RFCfmdeCQFVZVL/dphNCgFVY/WO74xYveeHd2PBiO4UVXgjJsITQyIFL0txkeCwnnLsHMAlKa6I50cTu0kOCwUVnpg2jahWn5GZtE4Kivxo62oMG0d9ORdRfWcqvUSsI4IgCEJDaJSmROHk1P1qfH9BZcxGv/RkJ06bGY9PZc7re7hr0XbDAqEnRGga+FWVx9/8kokrcsKqw6u8AR55Yw8BVcOvBjhWXh0xRZy+bhdl1X7uWrQdu8WMJ0rqSGq8nXi7hayNuVw3bzN3LtqOPxDgkzlP0emnfXG8/x6dBvUnPclhrD97ZCadWicw953avOdJ/dtFjQj0BlTibJaI383asJtJ/dsxMCMNX0Bj9JIdXDdvM6OX7OCefm3p1iaJ/GI396/aSXOXlctTXScVdfVt67NaordDnky41T221WqmdaIzpk0jVIRvmXU9r025hk6tm9HcdX62CMayvZzJqL4zZXsR64ggCILQEGRC3QSp+5/5s+/vi2rbeGnrARaP7oE/oDF+Rfh//k+/t5df/qSDESkX2gKoi9f0ZCclVT7u6deWx97M5Z5+bbFbTFEFZ7tUFxOvvQynzUzeiaqIKeC0Ae3DEkMC332P9eaZTP3mMz5u05mn75rJ1DE3M8cbIMVlo3Wig4Cm4fYFjAIYwMiIrnt+VdUIEH1Sn+KyMfvmTmG2GF1sLx3Tk4fW766ZYENpjA2Wuqirb96zxxdg+b29OFHppajSy4acPB78accGC7cfsmk05YlzQzkXUX1nyvYixTeCIAhCQxBB3QSp+5/5zrwS5r6z10ibUBQFswKP394lWF5SUhXxn/+QzDaGmIbaSa+eHa1PiFvEB8tPZt54BWOXfcKcQRlhGdZ6nJvDambs/1xOfrEbq1nhr/f04L6XaoXRJSlxYWvwmS20LDzEgax53F3eEU0xMctuodIbIKBqaBqs3HaQEX0vY+W43hSWeyiq9GKK0dR4otKLy26J+rtEp5Xyan9UAVTq9jHjxo68tPUAxZVe3L5ARM60Lurq+zV/idtrXE/9eQtGdKd14qlNjs+maG5KPuBz5fE+E9dTim8EQRCEhiCCugnitJlZOqanUaySvXk/hRUebBZzmFDQxVI0ERqrJfDyVBcfTr8Om8WE2xdgaPa24KbBSX3JL3bzfu4xlt/bC7cvgD+gcf+q2tIXfcJdWOEhe2Qm84Z2waQolLh9HClxc633GD/56FUe+clEjruSGfnQcn47uAtaTe14cpyNB1bvNI63ZEwPCss9/PKVz8M8wsvG9jQqw9OTnfzf3Vfzx7/n0j4tPkqLY3e8AZXmLltUAVRU6SVrYy4r7uvFr9fsYvbNVzD3na9CNlg6SYu3U1Tpxe3zc7S0mtR4O/nF7pib5tzeQIT1ZMrKz1gzoQ9JcWfrX0XDaYo+4PNl4i7FN4IgCEJDEEHdxFBVjWNl4XXiepSbhsah4qqIzYPzhnaJsITEEpgmRaGw0kNqgp3jFV7mD+tKiduHL6AyMCONwd0uZvSSHcwZlGFkRkPthFvPQJ70cg5P1bw22azR+vn5vLTgz5Ta4liWeRu+dj/iiRoBrk/D//RW+PEOFVeHVY/rGwSfGtaVVeN6Y7eY8Ksa1X6VP9/ZlaNl1eQeKuHl+3pT5fUTZ7Pwp7dy2ZRbwMCMNF4YmcnkELGtb1jML3ZTUuWjsMJDlTfAzrwSsjbmsnh0D9Li7RGbMJ+562p8AdW4WVBVNewzirVJNHAWejtPZ8J8LlI1LlSaemKKIAiC0LQQQd3EiCaClm45EPRDL66NqVs4KpP/e+9rUuPtBFSNS1PiWDa2F+XVPvwBDafNFDHNfWFkJlkb/0OS08ZtinHEAAAgAElEQVTIvpcyo6amW0/y+O3PMozyk7QEe8xkEf3nls0cvPjnNUxfO482hQcpv30oVU88xZ8dzaj0+HFYTTxz99WYFAUNLaz0BGqrx0NJjbfTMsFOucdPeXVtmoh+Y9GhdTNWf3yQ0f3actei7cbr9WPHihD0BzRWjeuNosCWWdfjtJlJcto4XumJuN6/WvN5mDVm4ahMUhMchphyWKPbAUJbF88EpzthbqgPuCnZQ5oC58s0XRAEQWh8JOWjiRFNBEXzQ09ckcNvf5bB3KFXYbeY8deMR1MT7CTFWRiWvZ05f9tD1uDOfDijP2sm9CE5zsLvB2Vw/w0/MkpW9ONNX7crrN47vsavHIqeLKL/fKy4kmfe+jPtzB4+e+4lpv/8IdxJKRSUe5i/6WuqfSojXvyYa+dGTyqp8gbCHuvWJomHburIqJrNhXWj+Wau301xpY/hfS7Dr0ZOiTflFlDq9uGwmsnamGuI6WVjexJnNzP8xY+5du5m7lq0nSOl1RSWV+OOkQMdV1Pmol/r45W1GydbuOxRkyRauGKLL1XVKCz3cKi4isJyT73KTE43aaIhqRrnonRFEARBEC5URFA3MUJFkN7o16FlfFTRZ1KCotSvqoxasoOf/Pkj7l60neMVXlJrGvq8AZUytw+/qvHo33P58kg5heUe43ihrYG2kCi46prNe3VbA7M37+dnpd+waEgnHA4rk3/+WzLvfobshE7MvrkTVrNCh7R4fvuzTmE+Yz2pJPR4bZo7efbubsZj0wa0N14TK+0jzmbGYlKwRmmAHJiRRrLLRjOHhZXjerNl1vW8PrUfzZzWqDckn+eXxmyL1G8c9OdXeQKGuIwWZ3eyqfGpitXTTZpoSIScxMQJgiAIwqkjlo9zQEO+StdF0NP/2Gs0DYYmb+ikJztRFAW3T42a2zxvaBdUjbA0iyeHdCEtwQYovDntf3Bag5PK74qq2JCTz7AebQybyOHSajbk5DFnUAZJTiu+gIq9ooJlO5aQuHwJ5SknuDnhOlLbdWTRTR2Jt1uM2DrdM61v7oPapJIV9/aiqNJLaoKdN3Ye4oO9hTxxx1VclORE1WrLUUpixNtpBG86NE1jwYjuxqR9YEYaM27syMHjlWFe8oUjM0lwWmLaV46WVkf4z/XNl6HnPXC8EpfdYlgAGmIHOFUv8+kmTTTEBywxcYIgCIJw6oigPss01AdrMim0T43n4Vuv5EhpNXMGZfD1kbKIzON7+rXFYg5aPKL5kNOT44xGQajNZX5lfB9Ka3zWo1fUCuAFI7rz3Af7KCz3kjW4Mz9Kc5E+oINhuxh2bBePvf0cloJjLOp5OzdNnErqG18z48aOVPtUZq4Pt5BMejnH8CHrFFZ4UBSFlHgbvoDKF4dLmdS/HQ6rGZOiYFI0Q0Bmb94fEW8XjKZzkOy0cqTMjQKG4G9ek2oSbZPjsrG9ogpTPUHl4dsyjLpxDUiJtxnZ2KGbG58b3i3q5/tDN0unKlbPRNJEfYW/xMQJgiAIwqmjaNr565Hs0aOH9umnnzb2Mk5KYbmH2xdsiRAqsaaTdQX4wIw0HhjQISy9IntkJslxVswmhbwTbio8fiNi7/3cY9ze/WKsZhNDs7dFHH/9pL5GlFzdNekJHvrfX5nQm33HKumy/HlS/vdR3B07Me0nUzneqSvP3H013xZWMuf1Pcwf1pW7Fm2PONeHM65j1F/DRbvTagIUbJag2+jxN3ONevDnhnfDajIZIn5gRhqzb+5EqdtHQbmHDTl5/OZnnaj2qTSPs5J7pJw5r+8hNd7On+/sigbsK6gge/N+o7wG4J8P9ae0ymcUz+hreXnbd6zNyWdgRhqP3NbZqF+3mmHn96URsYV1P7P63iw19N9A3X8P52KjYFOM2BMEQRCEpoSiKDmapvWI9juZUJ9lGjqdrGsPGJLZxhDT+mv16W96sgOzSQmL2NMnzUMy28TMZY7lT9YTPADyT1Thq3BzUZKDwp/8jOX//Iatw8Yz7idX0CEtnsffzGXWzZ2CkXQx7BmHit3GBLnKG8BmVigo90bYUArLvezMK2Hqqp2sn9SXrMGdaZvqYu/Rcqav3RUmjh+66QrGL/+UdRP70txlZeHI7pRV+42NjKETZX1T4v6CSp59fx9ZgztzSUocR0rcxjXa+m0RD/60I62a1aZ4qKpGq0THD06G62vlOJ1J87lKmpCYOEEQBEE4dURQN5CGTAxVVUNRFGMqrE9OT/ZVel0BHk38psbbadvChdmkMHZZuKCbsvIz5gzKiGqZ0IXmpP7togpgXyCYt9yq7DjzP8zGtTuN6iUriOvamQ233kf+kUo+WZHDmgl92JRbwOi+lwUn5ic5V6gY/tdD1zNuefjNge4R1zcNev0qDquJo6XVUafoR0uryS9241NVWjaz4wvAxJc/i3rMrI25PH1nV/701lfszCsxYvBW3NuLIZlt6HJxM16bck3EZ1hfcVnfm6XzRaxKTJwgCIIgnBqS8tEAGpLWoD/3zoXbGJq9jayNucy4sSMDM9JOOp2sG3WmT3919Gi5kX/9mMMl7qhiu0PLeGbffAWapjFvaBfWT+rLqvF9eGnrAXbmlZC9eT/zh3UNS394YUR3Lk12ML94O+8tuZ8+B3exMe5SHBaFVdsP8vzw7sbz9bi7+Zu+Dk6YKzw89e5esgZ35v3p1/HKhNpz6aQnO8Ni+XRCJ+PpyU4Cmsbcd/aiaRovjMwMW+P8YV2Zv+lrY+q8K7+MY2XVUY95RasEVo7rbYjp0N8VlHvI2pjLiSqf8TnUjbTTxeXFyXGkJkSvFG9ILF19jicIgiAIwvmJeKgbQEO8sLGeu3ZiX9Li7RS7fVGnldE81FNvaG+kWSwd05PVO75jSGYb2qW6yDvh5tn397Ezr8QQ26GJFQtGdCfRaeFEpRenzcyREg9xNjOtkxwcLqkmNcHO90VVrFr3LyYvy6L7t7so7XstswdOJXB5W+YMupKAqmExKxRVeKj2qaiahtVs4ldrPic13s60Ae0NK8XybQeZ3L8dzV12Css9YZsoL0uJCytj0a+JPk3OHplJebXPWH+3NklM6t+OFJeNpDgbM9ftorDCY/ifh2SmU+L2RZ1kPzWsKykuG2OXfRLTK56e7OSNqddwrMxzSt5h8R0LgiAIwn8PJ/NQi6BuAIeKq7jmyQ8jHt8y63ouTo6r13M//s0NnKjyRYiwls3suL0BFEXBYVVwe1Wq/SrfF1Xx9hdHuCMznZbN7MZGxFB7hR7zNm1A+7CUC8CwOHy09xjXd2pliFzd8qGL0ZTKEl59eQarbxhOtz/OYMNn+RGbIUM38nVrk8TDt2aQGGfj4PGgR7mwwsOysT1x+9Sw1z0/vDtv7T7Ebd3SKanyha39+eHdae6y4g9oWM0Kj27MNeIC9ee8MKI7SXFWjpXVCvQHbmhPM6eVx9+MfH5o7N2MGzue1IqyZdb1UUV+fTcMHq/0UO0LYFYUo31RxLQgCIIgXHiIoD5DnKkJ9Z0Lt0U8Hlp1vWBEd6q8AZb8+1uGZLYhyWklLcHOC5v3M3VAe4YvjhSAr0zoQ0DVuG7e5oh1r5/UlxbxdqNWXBeyjr257HjocR7+yURUkxmzGiBgMvPB9OuwmE1Rz7P83l7sK6ggxWWjucvGm7sO0+3SZFonOfm+qAqrWWH2q19Eivr7evHrNbuAYIHL5akuvH4Vb0DFZjZxvMKDxWTCr6os3XKAIZltjHO4vYGw1kT9mE8N60qczWxsMExx2WjZzMG01TsNwaxPua9olcC3hZXGNF8/xspxvaNes2g3SaHIdFoQBEEQ/rs4maAWD3UDaEjzXLTnLr+3F5qmMX9YVxaOyqRbmyQgsup6ysrP0DSNqTe0J2tjLnct2s7Boiq2fltEUYUnTFjqryks9/BtYWS9t57scTzkdceOl5Ez9pd0+Fl/Bn29lUtLjgIQMAU9wQXlHvwBNep5zCaFrI25DM3exuglO+h+WXMA7lmyg7HLPsFqNkV9XZnbz868EgorPFyc7KCw3MPYZZ9wy7P/ZuyyTzApCs2cZlLibfyi16WkJdhJirNhNQcnv9GOqQDPfbCPOYOu5KIkJ81dNg6VuI0MaQgWymRtzMVpM9Mq0RGRL32ktPqkPuhYleHSLCgIgiAIgo6kfDSAhqQ11H2u02bmWJknrE1Qtx8UVngocfuMaao+kXbaTMbk2WJSWD+pD74AURM6Cso9ZG/eH9YeqJ/jpa3BiS/A1Yf38uTb/0fH499TdsedHHvkf/G9dRBC7BIt4m04amrII5NAtDAROWvDbpaN7WU8FitCr5nTwr9nXW+8l3vrpJNMX7eLdZP64rAo2C1mCss97C+sMLKnYxWzbMotYM6gK2keZ+WuRdtJjbdHJI4sHt2DFi47LVx2Xp3cj0qvn4CqcbS0mldzDkU0Jeo3SSebQkuzoCAIgiAIOmL5OEfEsoBkDe6Mw2ritc8OMbjbxYYQjFbosmBEdz47WET3y1IiRLPuC74zM537b/hR2IbA+69vz/Mf7uO9PUf4x1/vx+mr5ukhDzJ57jTW7PiOkX3bcrzCY4jywgoPS8b0wOvXmFSnUCb3UAmJLjtJTqtRevKX4d24u8aH3K1NUoRvOXtkJn5VZeqqneQXu1k/qW/U0pnNM/pzSfOgzaKo0ovbF+BwiZukOAvFlb4w0Tt/WFeeePsro3DF6w8YnvXQG5P0ZCetE52YTAqqqvHlkTLDPqJfu9d3HuKXP2kPEHaTdDKLD3DKZS2CIAiCIJx/SLFLEyDWRLNdmgt/QOPmq1obIhSiF7pMWfkZr07ph8cXYPX4PqiaRnGllz/+PdfwBW/9tohf/bQ9iU4rdquZh27qRPInW/hlvy7kHiln4u2/xdQmnafuu5bszftZm5PPnT0vpcobIHvzfiBY5223mHFa4alhXVEIRuUFVJW+P2rBLxbXerHnDe2C3WIyprw780p4aesBlt/bCwg2FybFWXj077VJHKGbInXSk50cOF6Jy24hNcFOaoKdQ8VVlLp9zFi3i9R4e1hJjKppFFZ4jGlyidvL0jE9o7Yb6t8gFFV6w7zY+oQ9a3BnbBZzhBA+2RS6daLztGvBBUEQBEG4MBBB3UDqFrskO60xI/BC0TOL64pIp9VCSpINi1kJE40t4u1h0XFJTis2i0JBmSdsavzCyEx6XZZkFMbMG9qFgnIPCz78hnFXNaft048Rv2IZ3pm/I+veacTZuhqCFGDpmJ5YzApev8ojt2VQ4QnETMXQNxeGCtKZ63ezdmJfmjmtZA3uTIt4G06bBUWBQ8VuNuTkYTNfypDMNmzKLQAwcrCnr9sVMXF+bng34zorikL7tHjyi4N523otOsBHM/sbpSwAx8o8YY2R84Z2oWUzR5jAjSWQ27ZwhT1P/4wDmsbSMT0jNjLaLGZMJoX2qfGsndgXf0DFYjaRFt+08qXPVW25IAiCIPy3I4K6AUTz1GaPzOTZ979mU27BSZMefqh+2mExh2VILx3Tk4EZaWGRcEvH9GTO65+HCdrJL+fwyoQ+DOp6MYdLq5n7TtCT/Wzc91xy3QxclcV8d99Uxja7hsT39xnivNqnMu7HbQ0vs/5eXtr6zUmbDAN1Smzyi90UVXr5w9/28IdbO6EoCmOW7gg7psUEdquZbm2SDGHqsJrIGtyZOJuZKm8Au9VEaoINm8WM369yuDRYwJLotDIwI80Q41BzI2KrnSgXlnsiNgjOXL+bV6f0C/scYt3UxNnNMXPAQ2P4Qifiqqqxr7CiyaZ8SAqJIAiCIJw7RFA3gLrJDqnxwfKSh266giGZbcjevJ/xyz+N6qPVNym+OqUf1T4VswJOm9kQZh5fgKVbDhgTaptF4Xe3ZDDixY+N88WFpF2ETq5BI8FhJU3VmNS/Hcdnz6H7xhc5cHE7An/7G4GrrubREjfxDovhY9aFYmrNJDy/2M2kl3OYMygjTLzWbTI8XhGeYpGe7ORwiZudeSUUlHvDSlZCj5m1MdfIsb75qtbcX7OO0OOsGtebZKeVvQXlhoDXveNA2E1LC1ft9Y01efb51bDHYt3UhB4rWnrHzPW7WTOhT4S/OlrKR1PxUMdKIWkq6xMEQRCECwkR1A0gVLhF23yn2yOqvH4KyjVauCItAEUV3jBBt2xsTxxWE4lOizGN7nd5ChOuawcE/czv5x5jRJ9LSIm388H06zDXCDq9qfChmzoGWwQLS0lNS2LJ9PGUXtWa8vvuZ9T6/5C/6aOoAnrm+trpMwRFV10PsJ6moU+bFTRjyquL3ec+2AcEBWs0YZvktBoe8JXjemOzKDEj+YrdPkNM649PWfkZS8f05Pe3ZOC0WSKsC4qiRJ08K0r4ta9PSksscQ6ECdGmnvLR1NcnCIIgCBcSkkPdAKw1UXIAk/q3C9tEqNsjpg1oz1dHy7ljwVb2His3cosBStxejpZWM39YV1aP780zd12NoigcK/OgKIohpkf2vZQxS3dw3bzNZG3M5cGB7TGbTNy9aDs3zP+Ibwsr+dWaoPVjUv92zHvxPR5dPJv5bz5NfrGbez+u4LPRU6nETGq83VjfzPW7mdS/nbGe/GI37dPijUzs9ORglnNodvbzw7tzUaKDOYMyePb9r/GrGq9M6MOHM65j6ZiefPRVAUMy2/Dh9Ou4KNERNdO5xO0zzldY7kHTlKjPs1pMeP2BqDndpW4fTltww2LdmxSzAk8O6RK27ieHdMEcxdlgMimkJti5ODku6rF0W0jdtem51A19XmPR1NcnCIIgCBcSIqgbgFmBp+/sSnqy05i6hpJf7OayFnFkb94fUfTh96tUePxc1iKONs2dNHNY+dWazxkw/yNmrNtFUYWX1Hg74398uRGJB0FbiS+g8ez7XzNnUAZrJvShTfPgNFbRVDq8uoKX/nwvffK+ICe9E2jBnOg4mzmqgNbtGxAUWPsKKsjamMtDN3XkueHdyN683zjPy/f15pE3/sOtz21h4oocCsu9mBSFuxdt5/qnPmLssk/ofllzNuTkcbCoCsWkMG9opLDV00P0khlN06IW5FRU+7lr0XbuWrSdrI25zLixoyH00xLsERsH9cKVgAb/3HvMWPecQRm8tPUAJlPD/3nXt7ynISU/jUFTX58gCIIgXEiI5aMBeAMqVktwM11agj2qzcAUYjPIL3bj9vk5XBzAYlYoc/vDPMV1vcZB+4eZ+cO6GtFvD93UEZNCxObETMp4aPWfaJu3h09+1J0HB0whP6mVsY4Sty+qgK7yBoyfdYuKPr1+4o6rWJuTz9qcfKPaW99ECMHK8MkhYl+fyi+/txeVHj9ub4C57+xlzqAM0hLsJDqtPPH2l0ZCiF4y0+2SqyKsF2YTzK7ZAKlnXL+09QDTBrQnrZmdi2qypCH6hruFIzON9BKb2cTsmzudknisb3lPQ0p+GoOmvj5BEARBuJAQQd0AVBVjU1+3NkkRjXzzhnahoMzD9IEdqPQGggJGUThUUk2rxOAGxvnDupKWYCc13m74n1s1cwQFd7WfMUs/MYpdnr7ramwWBU2DeLuFOYMyyN68n7e/OML80X1otaaKb/73GbhzBKzbbbQd6kK5roB+YUR3kuOsfDSzP3knqow4PAiKY4fVbDx33tAutE5y8MqEPiiABrROdMT0PgMoChRWeAxPdrc2SUwb0J5ZN3Xim8IKXtp6gLHXtKWi2k8Llz3Mk3ys1B1206C/j8tTXbRMcIQJweOVkRsCJ76cQ9bgzoxd9okxjT1VdFvImXpeY9HU1ycIgiAIFwoiqBuAX62t3d6ZV8JT7wansR3S4jlUEny8VaIdRVEoqvBSXu3jSI2YLnX7w3KSn76zKwnOYAPgqCU7eOKOq5j96hekxtuZN7QL8XYLf3orN0xk3htfwktfv0/RvKc5UOxh+oNLOOYOsCzezoIR3XHZLXxfVGXUmb8wMpOKah+v338NSXFWHn8z10jKmDe0S9h7S0920rKZgw+mXweA2xfAH9BwewM8+/4+Cis8LL+3V9SpfN6JKma/+gXPD+/GwlGZxqbCwgoPzZwWrBaFtAQ7M2+8gmpfgO+KqmjusuILYExPAxpRPelrJ/aNmKpW+6JvuIuzmY2fJdFCEARBEIRzhXioG4DFVLuZTo+t079Gt5pNLN1ygO9PuLl70XYGP7+F2a9+gV9V+a6oyihjgaDge3DtLlw2K5u/Cnp/L2keR2q8nf8d0pn05Dhcdgu/H3QlszbspqCwjLm71vH7P47B/9prjJ2zhjmv7+HXP7uS1Hg7Y5Z+QlGFlxlrd+ENqMy++QqyBncmyWlh7jt7OVpWzYgXPzbi8HSLx7QBwbptvSAma+N/uGH+R4xesoMyt48Za3cx5/U9zLixI6nxdp54+0ueH949zJcbLFAJ3gQ88kYuzRwWsgZ35r1f/5g5gzJ4bOOXFFV4qfIGGLvsE259bgurd3xHXnE1ty/YwjVPfsjtC7bg9atRRXLd3GsAsxJ9U6O++VF/rSRaCIIgCIJwLpAJdQOIs5l4YWQmf3n/6yiFK3t44o6rIqasM9fv5qWxvaKKxWNl1QzreQkOq5mAqvHM3Vfj8Qf4xeLt5Be7WT+pL2l7PmPp28/SviiPt7oP5Df/M5ZSZwLUKV25JCXOsFvo6RxZG3MN0R/t/Jenutg8oz9+VWPuO19GCG792LM2BP3Vld4AzV1Wlo7pSYXHT5LTyq/X7qKwwsOcQRkUVng4VFKNzWJi5rrdhnc6wWHlgdW1udPRatUPHK+M4UmP/BycNrNRdV63fCX0tZJoIQiCIAjCuUAEdQPw+jVSXBYevvVK7lq0PaxwJTXezqUpcWGb6rI372dnXgm2mri9umLRF1ApLPdECEM9K7qozM3T7z6L2VfNV0vXMuWruLD16JsO05OdHClxh53bpASLUCb8uJ0RhVf3/GZTMLFj/rCuYWUuocfWf26d5OSeJTvC1nnUFzA82CkuG9kjM0l0WsjamGuI6cWje+C0msLOHS0h5dn39/HCiO7Gpkf9HE5bpChOctpo2cxhNC1qQIt4G4UVHuO9SaKFIAiCIAjnChHUDcBqMVFU5sVmNocJwhK3j9/+rBPFVT4jvSM01cJsImyiOjAjjdk3d8KkKKz++CBLx/TEbFIIqBrrP/2e38YdZbQXsv91kDnLX2HWjhPM6NmD9GO5EaK4yhswprO6uE1PdjJnUEYw3i/OhsWskD0y07Cd6GsrrvSSX+w2ilvqHlu3UKQnO/m+qCpi8v7EHVcZv0+KszFz3S5SE2z8/pYMHr71SiNZ4nilJ+z40c5XWOGhRbwtrI68ZTMHSc5IUWwyKVyW4iLBYTU82MlOqyRaCIIgCILQKCiaFulRPV/o0aOH9umnn56z8xVVVPP1sQqqfaqxwRCCfupnf9HNsGropCc7WTmuN4+/mUuS08YDA9pTVu3DH9C4f9VnLB3TA7dPNXKnO9n9rNy7nuZrXmbetaN4vt9dDMxI4w+3XomiQHGlL0wUvzCiOynxNgrKPWGV4rqQf+CG9qTE27BbTLi9AVQU/IGgp/vZ9/cxqX87sjbmkhpvj2h91EV6YYWH7JGZzPnbnrAIPYA1E/owfd0uFo/uQctmdtze6GK2bszdwIw0pg3oEPZeFo/uQfvUeIrdPhHFgiAIgiA0ORRFydE0LWqMmEyoG4DbpzJz/W5S4+1hkXmFFR5UTYuwMeQXu1GAwnIv9/Rry8vbDjC6X1vDLuKwWhi7LPjzjV9vJWvTCzRzl1L2y+m8aLmG9GQnDwzogMNi4sDxKhb/a3+YreMvH+xjSGYbNuTkBbOgvQGcVjNxNhNzBl2Jqmm8tOUAd/W6lGNl1cx9Zy+pCTam3tA+KJQ37zcm50+9u5eswZ25rIULq1lB0zR+d0sniiq9uL0Bw06hk57sJD3ZyWtTrqkVvq7o1y1aJnKsifLZTuVQVY2iSq+IdkEQBEEQzhgiqBuALprzi91GZF6S08rFNYUu0WwTFrPCb3/WiYCmMbpfW3yBWuEdqDneA1tWM/3fK/lP2uWMHfowC+aM421Vw2o2YbMoePwqSXEWxl7TNsxvvXBUJs0cFmy9LmV6zebAJ4d0ISnOyi3P/tuYYmdv3s/Wb4t4ZUIfbGYT1f6AYa1w2sysmdAHX0DDalZYvvUA13ZIY/arX4RN4OtuAlw8ugetQ8pWfohoYvlcR9pFK4RZPLoHHVsmiKgWBEEQBOGUEUHdAMwhonlnXomRqLHivl4cL/eycGQmE0NsDPOHdcVhNRFnNzNxRU5NLXdc8BgnqrC4g89784prCZjMLOp1B61aJGBSFEYs+ZiswZ35UVo8qqbxzhdHGdz9Yl6Z0Ad/QON4hYf/e+9rxl/bzojK09sFH771Sj6Yfh0BVWPxP79lbU4+AAqQ1szBsVI33oBKC5sNh9VMQbmHgnIPmZck8fPubaj0+MNuDHbmlTD3nb28Mr4PioIx2QUoLPecN9PeokpvRCGM5FULgiAIgnC6iKBuAIpCmNVjYEYav/1ZBooSnLbaLAprJ/ZB1SCgahwprcbr1/jk2+O8MqEPvoCKqmks/HELKsaMI/6LVixYtJwpK2FBSjrpyU4WjOhOmdvHvKFdahIsNDQNul/WnOGLPw7zSReWe2mRYOPBtZ/X+qpHZvL9iSpjmqwTGiNnMpnYkJPHPf3ahqVqLByVSce0BAoqPFE3DdqsJtISHMD5Oe31+qMXwkhetSAIgiAIp4NsSmwAh4ureOTv/2FIZhsuSnSggbGhMJj93A2TooSJ1Nfv78fhUg+TX87h0IlKHvjqH/zyvSWAxrHfPIxn4mQCmoJJAVUDm0XBBBwp9ZDksqKqKg6rhREvfhwhkLMGd6ZdmguvX8NsgmNlHi5KcmBWFErcPqOxsK7YVVWN/OIqhkc5pu6J/iGxXFju4fYFW6K+vqlOe8/HNQuCIAiC0DSQTYlnCIvZZPiY59PSQJgAAA2qSURBVAzKMCLyIDjpPFHpC0v/yC92U+1TmfxyDhw8yNq/z6fnoVw+7tCTwIIXGP6Powx8+yt+d0sGJVU+fAGV1AQ7XlUj2WXDZTdRUObH6/dFnaxekhLHL1d/bmQ+rxzXm3i7hSSnjdaJzpgxciaTgtmkxJzWRttEWNfOcT5Oe1NcNhaP7hFxoyB51YIgCIIgnA4iqBuA2xdg7jvBzYhXtEqIEJRxNnPEY341uPGwudVBc3cpv77lQV698gbeu6oD69u1pVWig9dy8unRtnnYpr8XRnSnvFrFH4DWifaoGx6PlLiNKLv8YjcWk0JzV+2k9WRTV5vFHPWYtbaQkydu/NDrmyL1uVEQBEEQBEFoKKbGXsD5hMWkGPXeXr9KerIz7PdV3kDYY1ce20+LB6dySaKNE3GJ/PS+BbzaeQDpzePIO+HGZjHhsJgYdPVFWEwmnrnraj6Yfh1Zgzvzh9f/wwOrP8cbUFmx7QDPD+9uHFvPiZ6/6WvjXA0Vs/q0NvSYDZnWnu7rGwv9RuHi5DhSE+wipgVBEARBOG3EQ90AfL4AXxVUMPnlHPpdnsKovpeG+aV1D/Uvl21nyJtLmLhjA6YWqexb/xb3bS0xnpc9MpN4uxm/qtG6mZUKj4ZP1QioGhUeP+XVflo2s1NU4WXRP/czbUAH/v55Pt0vS6FVMwfNXTZOVHq4P6TM5VQ2BJ5uJrNkOguCIAiC8N/CyTzUIqgbiM8XoKDCg1/VcNnMeAMavoCK2aRgNSnYdmwj/v7JWPd9TeXwUfiefBJri2RK3Cp+VcNiUoi3m6j0qCQ6Tcbj8TYz1f7a5zisJio9ARxWM83jbGENgslOK2UeH25vgIAGDquJFi6ZtgqCIAiCIJwtZFPiGcRqNXNxclz0XwYC8MAU8Hrg3XdxDRwIBCe5VT4v1AjiBIeNxLig+HU5Yp8rJb7257p+5uYWe8xmQkEQBEEQBOHc0eQEtaIoNwH/B5iBFzVNe6KRl/TDvPce9O0LLhf87W+Qng7xtWr4XFRqC4IgCIIgCI1Dk9qUqCiKGXgeuBnIAH6hKEpG467qJJw4AffcAz/9KTzzTPCxK64IE9OCIAiCIAjChU2TEtRAL+AbTdO+1TTNC7wCDG7kNUVn/Xro1AlWrYLf/Q6mT2/sFQmCIAiCIAiNQFMT1BcDeSF/z695rGnx6KMwbFjQ2vHJJ/DYY+A4iRlaEARBEARBuGBpah7qaDEVYTEkiqJMACYAXHLJJediTZHcfXdQQP/612BpapdQEARBEARBOJc0tQl1PtAm5O/pwOHQJ2iatkjTtB6apvVITU09p4sz6NABHnpIxLQgCIIgCILQ5AT1J0B7RVHaKopiA+4G3mjkNQmCIAiCIAhCTJrUiFXTNL+iKFOBdwnG5i3RNO0/jbwsQRAEQRAEQYhJkxLUAJqmvQW81djrEARBEARBEIT60NQsH4IgCIIgCIJwXiGCWhAEQRAEQRBOAxHUgiAIgiAIgnAaiKAWBEEQBEEQhNNABLUgCIIgCIIgnAYiqAVBEARBEAThNBBBLQiCIAiCIAingQhqQRAEQRAEQTgNRFALgiAIgiAIwmkggloQBEEQBEEQTgMR1IIgCIIgCIJwGoigFgRBEARBEITTQAS1IAiCIAiCIJwGIqgFQRAEQRAE4TQQQS0IgiAIgiAIp4GiaVpjr+GUURSlEPiukU7fAjjeSOe+0JFre/aQa3v2kGt79pBre3aR63v2kGt79miMa3uppmmp0X5xXgvqxkRRlE81TevR2Ou4EJFre/aQa3v2kGt79pBre3aR63v2kGt79mhq11YsH4IgCIIgCIJwGoigFgRBEARBEITTQAT1qbOosRdwASPX9uwh1/bsIdf27CHX9uwi1/fsIdf27NGkrq14qAVBEARBEAThNJAJtSAIgiAIgiCcBv/f3v3HWl3XcRx/vuKHKLpIJWZqAw0XaHk1aCxaE2qErHVpUmI/dI35E622xsRapSO3mitmSywtBJsFN9JizoWOH7PW+JF65WdMTJYEk7WAtB8Q9O6P7+d4vzuc7733+L2733sur8d2d77n8/18D+/z3ntf3vd7Pt9z3VA3SdJMSbsl7ZG0sOp4Wp2kvZK2SeqU9Mc0drakZyS9lB7fUXWcrULSUkkHJW3PjTXMpzI/SLW8VdKV1UU+8BXk9m5Jf0312ylpVm7fXSm3uyV9vJqoW4OkCyWtl7RL0g5JX07jrt2Susmta7ckSSMkbZb0YsrtPWl8nKRNqW5XShqexk9Lz/ek/WOrjH8g6ya3yyS9kqvbtjRe+TnBDXUTJA0BHgCuBiYC10maWG1Ug8K0iGjLff3NQmBtRIwH1qbn1jvLgJl1Y0X5vBoYn35uAh7spxhb1TJOzi3A4lS/bRHxFEA6L8wFLk3HLEnnD2vsOPDViJgATAHmpxy6dssryi24dss6CkyPiMuBNmCmpCnAd8lyOx44BMxL8+cBhyLiPcDiNM8aK8otwIJc3XamscrPCW6om/NBYE9E/DkijgErgPaKYxqM2oHlaXs5MLvCWFpKRDwL/L1uuCif7cCjkdkIjJJ0Xv9E2noKclukHVgREUcj4hVgD9n5wxqIiAMR8Xzafh3YBZyPa7e0bnJbxLXbS6n+3khPh6WfAKYDq9J4fd3W6nkV8FFJ6qdwW0o3uS1S+TnBDXVzzgdezT3fR/cnJutZAE9Lek7STWlsTEQcgOw/A+CdlUU3OBTl0/XcN25PHzEuzS1Pcm7fovQx+BXAJly7faout+DaLU3SEEmdwEHgGeBl4HBEHE9T8vl7M7dp/xHgnP6NuHXU5zYianV7b6rbxZJOS2OV160b6uY0+k3SX5NSztSIuJLs45r5kj5SdUCnENdzeQ8CF5N9JHkA+F4ad27fAklnAr8CvhIR/+huaoMx57cbDXLr2u0DEXEiItqAC8iu5E9oNC09OrdNqM+tpMuAu4D3ApOBs4E70/TKc+uGujn7gAtzzy8A9lcUy6AQEfvT40HgCbIT0mu1j2rS48HqIhwUivLpei4pIl5LJ/3/AQ/T9dG4c9skScPIGr7HIuLxNOza7QONcuva7VsRcRjYQLZOfZSkoWlXPn9v5jbtfzu9X0Z2ysrldmZawhQRcRR4hAFUt26om7MFGJ/u4B1OduPG6opjalmSRko6q7YNzAC2k+X0hjTtBuA31UQ4aBTlczVwfbo7egpwpPbxuvVO3Rq9T5HVL2S5nZvu6h9HdqPM5v6Or1WkdaQ/BXZFxPdzu1y7JRXl1rVbnqTRkkal7dOBj5GtUV8PzEnT6uu2Vs9zgHXhPwbSUEFu/5T7BVtka9PzdVvpOWFoz1OsJiKOS7odWAMMAZZGxI6Kw2plY4An0j0ZQ4GfR8RvJW0BOiTNA/4CfLrCGFuKpF8AVwHnStoHfAv4Do3z+RQwi+ymo38BX+z3gFtIQW6vSl/bFMBe4GaAiNghqQPYSfYtC/Mj4kQVcbeIqcAXgG1pzSTA13Dt9oWi3F7n2i3tPGB5+haUtwEdEfGkpJ3ACknfBl4g+4WG9PgzSXvIrkzPrSLoFlGU23WSRpMt8egEbknzKz8n+C8lmpmZmZmV4CUfZmZmZmYluKE2MzMzMyvBDbWZmZmZWQluqM3MzMzMSnBDbWZmZmZWghtqMzMzM7MS3FCbmfUzSV+StEvSY72c/y5Jq3ox742C8dmSJjYbZw//1jJJc3qeaWY2+LmhNjPrf7cBsyLic7WB3J8qPklE7I+IMs3rbKBPG2ozM+vihtrMrB9J+hFwEbBa0hFJD0l6GnhU0hBJ90naImmrpJvTMWMlbU/bZ0jqSPtXStokaVLu9e+V9KKkjZLGSPoQ8EngPkmdki5uENMESZtzz8dK2pq2v5ni2Z5iVYPj90o6N21PkrQhbY+UtDQd/4Kk9r7LpJnZwOGG2sysH0XELcB+YBqwGPgA0B4RnwXmAUciYjIwGbhR0ri6l7gNOBQR7wcWpeNrRgIbI+Jy4Fngxoj4A7AaWBARbRHxcoOYdgHDJV2Uhq4FOtL2DyNickRcBpwOfKKJt/t1YF16P9PImvqRTRxvZtYS3FCbmVVrdUT8O23PAK6X1AlsAs4BxtfN/zCwAiAitgNbc/uOAU+m7eeAsU3E0QF8Jm1fC6xM29PSVfBtwHTg0iZecwawML2fDcAI4N1NHG9m1hIK1+yZmVm/+GduW8AdEbEmP0HS2Lo5Rf4bEZG2T9DcOX4l8EtJjwMRES9JGgEsASZFxKuS7iZriusdp+sCTX6/gGsiYncTcZiZtRxfoTYzGzjWALdKGgYg6ZIGSyR+T7qSnL654329eN3XgbO6m5CWgpwAvkHX1elac/w3SWcCRTdG7qVr6ck1ufE1wB21ddeSruhFrGZmLccNtZnZwPETYCfwfLoJ8cecfJV5CTA63TR4J9mSjyM9vO4KYEG6MfCkmxJzVgKfJ62fjojDwMPANuDXwJaC4+4B7pf0O7KmvGYRMAzYmt7Poh7iNDNrSer6dNDMzAY6SUOAYRHxn9QcrwUuiYhjFYdmZnbK8hpqM7PWcgawPi0LEXCrm2kzs2r5CrWZ2SlE0gPA1Lrh+yPikSriMTMbDNxQm5mZmZmV4JsSzczMzMxKcENtZmZmZlaCG2ozMzMzsxLcUJuZmZmZleCG2szMzMyshP8D03/gRrft99oAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "pred = gb.predict(X_test)\n",
    "tmp = pd.concat([y_test,pd.Series(pred, index= y_test.index)],axis=1).rename(columns= {0:'pred'}).sample(frac=0.25)\n",
    "\n",
    "plt.figure(figsize= (12,8))\n",
    "sns.scatterplot(data=tmp ,x= 'freight_value', y= 'pred')\n",
    "plt.plot(np.arange(0,350,1), np.arange(0,350,1), color= 'red', linestyle= '--')\n",
    "plt.title('Predicted value vs Actual value', fontdict= {'size':15})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The plot shows that most predicted values do differ from the reference line. Specially as the actual value increases, the predicted value differes further. To gain better intuition, the residuals (error between predicted adn actual value) are plotted against the predicted values by the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 265,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.lines.Line2D at 0x7f1074db2970>"
      ]
     },
     "execution_count": 265,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAt0AAAHSCAYAAAAwiXe6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOzdeXhUZZY/8O9baypVgYQsgCQKjRCNGAxB1vkpSEu7oIyy2LIpyiYuPY4LTk/T2o09I6LjDK0Ibihrg6Cjje3WtHRPgxsBoTWKiKgJIAkhgaRSqe2+vz8q91K36lZSCSlSCd/P8/BAVarufauSh5x76rznCCkliIiIiIgocUztvQAiIiIios6OQTcRERERUYIx6CYiIiIiSjAG3URERERECcagm4iIiIgowRh0ExERERElmKW9F5BoWVlZsnfv3u29DCIiIiLq5EpKSo5JKbONvtbpg+7evXtj586d7b0MIiIiIurkhBDfx/oay0uIiIiIiBKMQTcRERERUYIx6CYiIiIiSjAG3URERERECcagm4iIiIgowRh0ExERERElGINuIiIiIqIEY9BNRERERJRgDLqJiIiIiBKMQTcRERERUYIx6CYiIiIiSjAG3URERERECcagm4iIiIgowRh0ExERERElGINuIiIiIqIEs7T3AoiI2pKiSFS5ffAFgrBZzMh02mAyifZeFhERneUYdBNRp6EoEvuO1mL2qp0or/YgN8OB52cMRn73NAbeRETUrlheQkSdRpXbpwXcAFBe7cHsVTtR5fa188qIiOhsx6CbiDoNXyCoBdyq8moPfIFgO62IiIgohEE3EXUaNosZuRkO3X25GQ7YLOZ2WhEREVEIg24i6jQynTY8P2OwFnirNd2ZTls7r4yIiM523EhJRJ2GySSQ3z0Nr88fye4lRESUVBh0E1GnYjIJZKfZ23sZREREOiwvISIiIiJKMAbdREREREQJxqCbiIiIiCjBGHQTERERESUYg24iIiIiogRj0E1ERERElGAMuomIiIiIEoxBNxERERFRgjHoJiIiIiJKMAbdREREREQJxqCbiIiIiCjBGHQTERERESUYg24iIiIiogRj0E1ERERElGAMuomIiIiIEoxBNxERERFRgjHoJiIiIiJKMAbdREREREQJxqCbiIiIiCjBGHQTERERESUYg24iIiIiogSztPcCiCh+iiJR5fbBFwjCZjEj02mDySTae1lERETUDAbdRB2EokjsO1qL2at2orzag9wMB56fMRj53dMYeBMRESU5lpcQdRBVbp8WcANAebUHs1ftRJXb184rIyIiouYw6CbqIHyBoBZwq8qrPfAFgu20IiIiIopXuwfdQoiXhBAVQojPw+57RAhxSAjxWeOfa8K+9m9CiG+EEPuEED9rn1UTnXk2ixm5GQ7dfbkZDtgs5nZaEREREcWr3YNuAC8DuMrg/qeklJc0/vkTAAghCgD8HMBFjc9ZJoRgxEFnhUynDc/PGKwF3mpNd6bT1s4rIyIioua0+0ZKKeXfhBC943z4eAB/kFJ6ARwUQnwDYAiADxO0PKKkYTIJ5HdPw+vzR7J7CRERUQeTDJnuWO4SQuxtLD/JaLyvF4CysMeUN95HdFYwmQSy0+zolZGK7DQ7A24iIqIOIlmD7mcB9AVwCYAjAJ5svN8owpCRdwgh5gghdgohdlZWViZulUREREREcUjKoFtKeVRKGZRSKgCeR6iEBAhltvPCHpoL4LDB85+TUg6WUg7Ozs5O/IKJiIiIiJqQlEG3EKJn2M0bAKidTd4E8HMhhF0I0QdAPwCfnOn1ERERERG1RLtvpBRCrAcwCkCWEKIcwMMARgkhLkGodOQ7AHMBQEr5hRBiI4BSAAEAd0op2aSYiIiIiJKakDKqJLpTGTx4sNy5c2d7L4OIiIiIOjkhRImUcrDR15KyvISIiIiIqDNh0E1ERERElGAMuomIiIiIEoxBNxERERFRgjHoJiIiIiJKMAbdREREREQJxqCbiIiIiCjBGHQTERERESUYg24iIiIiogRj0E1ERERElGAMuomIiIiIEoxBNxERERFRgjHoJiIiIiJKMAbdREREREQJxqCbiIiIiCjBGHQTERERESUYg24iIiIiogRj0E1ERERElGAMuomIiIiIEoxBNxERERFRgjHoJiIiIiJKMAbdREREREQJZmnvBVDHpSgSVW4ffIEgbBYzMp02mEyivZdFRERElHQYdFOrKIrEvqO1mL1qJ8qrPcjNcOD5GYOR3z2NgTcRERFRBJaXUKtUuX1awA0A5dUezF61E1VuXzuvjIiIiCj5MOimVvEFglrArSqv9sAXCLbTioiIiIiSF4NuahWbxYzcDIfuvtwMB2wWczutiIiIiCh5MeimVsl02vD8jMFa4K3WdGc6be28MiIiIqLkw42U1Comk0B+9zS8Pn8ku5e0g8jOMRkOK6o9fn4viIiIkhSDbmo1k0kgO83e3ss46xh1jlk+rRhLt36N90or2EmGiIgoCbG8hKiDMeocM29NCSYU52m32UmGiIgouTDTTRSnZBkGFKtzTLrDqrvNTjJERETJg0E3URySaRiQ2jkmPPDOzXCgxuPX3WYnGSIiouTB8hKiOCTTMCCjzjHLpxVjc0mZdpudZIiIiJILM91EcUimYUBGnWMyHFb87oZCPHwdu5cQERElIwbdRHGIVdLRXiUcRp1j2EmGiIgoebG8hCgOHAZEREREp4OZbqI4cBgQERERnQ4G3URx4jAgIiIiai2WlxARERERJRiDbiIiIiKiBGPQTURERESUYAy6iYiIiIgSjEE3EREREVGCMegmIiIiIkowBt1ERERERAnGoJuIiIiIKMEYdBMRERERJRiDbiIiIiKiBOMYeGoziiJR5fbBFwjCZjEj02mDySTae1lERERE7a7dM91CiJeEEBVCiM/D7usmhHhfCLG/8e+MxvuFEGKpEOIbIcReIcSg9ls5hVMUiX1Ha3HDsu0YufgD3LBsO/YdrYWiyPZeGhEREVG7a/egG8DLAK6KuO8hAFullP0AbG28DQBXA+jX+GcOgGfP0BqpGVVuH2av2onyag8AoLzag9mrdqLK7WvnlRERERG1v3YPuqWUfwNwPOLu8QBeafz3KwD+Oez+VTLkIwDpQoieZ2al1BRfIKgF3Kryag98gWA7rYiIiIgoebR70B1DdynlEQBo/Dun8f5eAMrCHlfeeJ+OEGKOEGKnEGJnZWVlwhdLgM1iRm6GQ3dfboYDNou5nVZERERElDySNeiOxWhXXlTRsJTyOSnlYCnl4Ozs7DOwLMp02vD8jMFa4J2b4cDzMwYj02lr55URERERtb9k7V5yVAjRU0p5pLF8pKLx/nIAeWGPywVw+IyvjqKYTAL53dPw+vyR7F5CREREFCFZM91vAril8d+3AHgj7P4ZjV1MhgE4oZahdHaKIlFZ68Wh6npU1nqTsiuIySSQnWZHr4xUZKfZGXATERERNWr3TLcQYj2AUQCyhBDlAB4G8BiAjUKI2wH8AGBS48P/BOAaAN8AqAcw84wvuB2o7fjU7iBq6UZ+9zQGtkREREQdgJAy+TKmbWnw4MFy586d7b2M01JZ68UNy7bruoPkZjjw+vyRyE6zt+PKiBKPQ5eIiKijEEKUSCkHG32t3TPd1Dy246OzFT/lISKiziJZa7opDNvxUaIk+14BDl0iIqLOgkF3B8B2fNQS8QbSahb5hmXbMXLxB7hh2XbsO1qbVIE3P+UhIqLOguUlHQDb8VG8WlKOESuLnEx7BdRPeSL3M/BTHiIi6miY6e4g2I6P4tGScoyOkEXmpzxERNRZMNNN1Im0JJDuCFlkfspDRESdBTPdlHDJtFkvmdaSCC3ZdNtRssj8lIeIiDoD9ummhEqmlm/JtJZEaelrZA9sIiKittNUn24G3ZRQyTTYJ5nWkkgMpImIiNoHh+NQu0mmzXrJtJZEUssxiIiIKHmwppsSKpkG+yTTWoiIiOjswqA7STW34a+jbAhMps16ybQWIiIiOruwpjsJNbcZrqNtCEymGuNkWgsRERF1Lk3VdDPTnYSaG3DSkgEoySCZWr4l01qIiIjo7MGgOwk1t+HvbNkQSERERNRZMOhOQs1t+OOGQCIiIqKOhUF3Empuwx83BBIRERF1LNxImaSa2/DHDYFEREREyYXDcTqg5gacnG0DUHiRQURERB0Zg25Keh2tRSIRERFRJNZ0dxIdZVhOa3S0FolEREREkZjp7gSSNRPcViUhbJFIREREHR0z3Z1AMmaC1QuBG5Ztx8jFH+CGZdux72htqzLwbJFIREREHR2D7k4gGTPBbXkh0JFaJHbmMh8iIiJqPZaXdAJqJjg88G7vTHBbXwhkumxYN3sozELAYTMj3ZF83UuStcyHiIiI2h8z3Z1AMmaC26okRA1kb1y2A5c9vg03PfcRjp70tuVS20wylvkQERFRcmCmuxMwmQTyu6fh9fkjDTcttkePa/VCIDLrm+GworLWG/daYgWyr88fmXR9ypOxzIeIiIiSA4PuDihWEG0UhLZXyYPRhUCGw4r9lXUtWktHCmSTscyHiIiIkgPLSzqYlnYFiZUp/vFkQ8I3+6kXAr0yUpGdZke1x9/i8ouO1LkkGct8iKhz4CZtoo6Pme4OpqXlFrEyxYdrPJi4/MO4M99tUaLSmqx1rDKVZAhkjd6Tpsp8KH7tURJFlKy4SZuoc2DQ3cG0NHCNVfKgZpfjqZFuq//wW1t+YbeYsGj8AKTazKj3BWG3xP8BTaKCt6bek2SrNe9oGGAQ6XWkvS1EFBvLSzqYlpZbGJU8LJ5QiOXbDmiPaS7b3FZdOVpafqEoEj+ebMCMlz7BzJc/xU3PfYSZL3+KGS99Ete523JATyR2KkkcvrdEeh1pbwsRxcZMdwdjVG6xYnoxFEVBZa03KpMbuaFRCIFH3vwcu8tqtMc0l21uq//wm+uyEk4NmN3eQKvPncjsEH8JJg7fWyI9btIm6hwYdHcwkYFrUJF49K1SvFdaEfNj+PDOJooice+V+Sg9Uht3jXRb/ocfq8tKJDVgXjiuoNXnTmTwFs97cibqkjtj7TMDDCK9ZN7bQkTxE1J27h3QgwcPljt37mzvZSREZa0XNyzbHhWcNJfJDQ/UrBYTLCYBjy920JaIGtvmgsVD1fUYufgDFOWl4/6f5WPB5r0tOrdamnK4xoMqtw/Ltx3A7rKauN6feNff1HtyJuqSO2vtc2d9XUSnozNeYBN1RkKIEinlYMOvMejuuNTANNL2BaPRKyO12ee3JLgx+g8fQKt+CcRz3vALiqK8dMwb1ReZThvOSXegR5eUZgPuyOMvnlCIV3YcxL1X5rdZ8NbUL8HWXhC1RKxzbJw7HFLKDv2LmQEGERF1RE0F3dxI2YGdbg/rlmxYi+y5DaDVmxTjOW/4psvdZTVYtKUUTrul2YA71vEXbN6LR64f0KbZ0sj3JPy4Z6Iuual2kG29cfRMa+q9JSIi6ogYdHdgpzuM5XQCw9PpMBHPecNr17cvGI3X54+MO2COdXwp5RkL3s7EUJ9Y54hsB8muH0RERO2PQXcHdjqBKXB6geHpBOzxnre12c5kmGJ5JqZTtkU7yEicekdERJQYrOk+i7V2w9rpblIMP2+2y457xvRDnywnUu1mZDlPv5QgWTbinenuJWo7yPdKK7Sv52Y4sG7WUORmpDZ77mR534iIiDoqbqRk0B1TSwPDttikqCgSx9xeeP1BKBL4XTMtD9vidWU4rKj2+Dv1xjxFkfjyx5OYu7ok6nvzuxsKm70YOhObP4mIiDqzpoJu9uk+y8XbN1sVa5PiyzOHIJ4YNhBQsK+iNiowrKz1YXdZTZsNr4nsTX42ZHBNJoEspw0LxxUg3WFFjcePJ97dh91lNXj4uuZLTDiUhoiIKHFY050ETreO9kzW4cYKzKrqvM2OZ1cUicMnPFrArT53wea9mDeqr3a7rYO8s2msuMlkwqItpbjpuY8wd3WJVvYTTz17MtTCExERdVYMutuZmoVtTeu9tnh+S8UKzGo8/mYD5iq3DxW1XsOgPd1h1Y4VOdXxdC8ozqYM7uls4DwTmz+J6OzDDdpEISwvaWfH3F7DLOxr80cgJy2l2efHyuKqJRptvZnPaBzx4gmFeOLdfc1mRX2BIKrcPsMR3zUef1SQ11ZlIWfTWPHwjjYt/Z6fznOJiIycLeV9RPFgprudqFf+Hp9xFrbBr8R1nKayuInIgquB2WvzR2Db/aOwaPwAPPHuPlTWeZvNitosZmwuKcPiCYW6bOqKacW4JLdrVMvDtioLOdsyuKczWIZDaYioLZ1N5X1EzWGmux2EX/kvHFdgmIU1xxnrNJXFbS4LHu9ajTLlOWkpUJwSTrsFT08pisqKGj0v02nDvVfm46n392HhuAJkOm3ISbPjnK4OWCzR13+tLQsxOne8GVyOHyciajtnU3kfUXMYdLeD8GB4+bYDWDyhEAs279U+elsysRAOW3ylD0blHiumFcNsQswserz/2TX3sWCszidNPS+/exoem1AIjy+IoJRw2s047vHBH1CigtymLihiBcdNnbsl/cP5MSgR0ek7m8r7iJrDPt3t4FB1PUYu/kC7XZSXjnmj+qJfjgvfV9XjvMxU9M50xh3oqX2v671BHDzmxtKt+1FZ58W6WUMx5YWPW913OVbf5o1zh0NKGRXshg9pmbziQ8PzZjptusE4D16Vj5XbD2JCcZ6W+e7ZJQU1DQEoioJjbp+uveDzMwajX7YL+yvrDIPjY24vbly2o1WvuaV9qpkVJyJqGpMZdLbpsH26hRDfAagFEAQQkFIOFkJ0A7ABQG8A3wGYLKWsbq81tkbklf/ushos2lKKDXOGYUCvri0O3kwmAQGBaS/qA+xH3yrFiunFUUGrUS2zUQAZ62PBwzUeTFz+YcwgeNO84TEz7OFZ/oXjCrBy+0HcMqKPLtP/7LRi/H7r13ivtAJjC3KwbtZQmE1CW1esspnX5o9Avbf12f2WfAzKXyRERM3jBm2iUzrCRsrRUspLwq4aHgKwVUrZD8DWxtsdSqyNfT27Olq9ec0oYHyvtAJZThtenz8S2xeMjtqoqIq14dJqMRm2B1Q3wKjBbkWdvgOL2qEk8nk2i1m3znSHFROK87SAWz3mHWtKMKE4T3sNU174GDaLWXtvYgXHDX4FB4+5W91ruiV9qrk5iIgoPtygTRTSEYLuSOMBvNL471cA/HM7rqVVTCaBftkubJw7HH99YBQ2zh2OftmuZsevN9XnNFbAaDKZmv3PLlYAaTGJqIuDxRMKsXzbAe255dUe+IOKLghW69SNuoWEr7PG40em09Zk3271dni2OdZrNQtg6db9ht1R2rpP9dm+OYh9d4mIiFomqctLAEgA7wkhJIAVUsrnAHSXUh4BACnlESFETuSThBBzAMwBgHPPPfdMrjcuiiJj1iTH6qjRXCmD0YbKeNvixQogPb6g7mNBIQQeefNz7C6r0R6Xm+GA1WyKKpd5ZcdBw9rvDIcV62YNRUWtFw3+ILJc9ph9u8Nvh2ebY71Wh82Myjovnnh3nzYKvd4XRM/0lDbpUx1Zt362bg5iaQ0REVHLJfVGSiHEOVLKw42B9fsA7gbwppQyPewx1VLKjFjHSMaNlC3dsBfP49XNlA3+IMxCwGEzI92hr5uLtfEv3vXECraa2tgYef7I578881J4A4qu7nz5tGIsbazpbupYka8FQMKCwci1jy3IwT1j+mPempKzLvBs6c8vERHR2aLDbqSUUh5u/LtCCPE6gCEAjgohejZmuXsCqGjXRbZCSzfsefyBJh8fKxhOd9h0x4kVMEtIrLl9qK7ziVGWvKlMcDwbZYzKWG5d+SnevGuk7rkZDit+d0MhHr4u9rFitStM1IadyLW/Vxr6sTPK5nd2Z3tpDRERUWskbdAthHACMEkpaxv/PRbAbwG8CeAWAI81/v1G+62ydeLtW6oGyj+eaGjy8fEMwYn1mPC2gmr9c8/0FF2WPJ7WeLGC4HBNlbH0ykjV3d+SjGkiW/epx673RV/4vFdagYevk1FrP9POdOtC9t0lIiJquWTeSNkdwN+FEHsAfALgLSnlOwgF21cKIfYDuLLxdocS74Y9NVA23Bw4vRgZjZsNw4PZorx0rJhejCcnDdRGwSuKhC8QxJOTBmLF9GIU5YWqc8qrPaio9eoC8blrShBUoAu422qUfEu6g6jnbm6zXiJG3Rsd+6sfa1vdFSXymG25ATGRrz+Wlmw4JSIiopCkruluC8lY0w3El50MH6KjDtBJd1hxTroDJxv8SLNbcE5XB6o9ftywbDuyXXbc/7N8Xc/rVbcNgTeg6MpKFk8oxBPv7kNlnRcLxxVg7uoS3Xm3LxitZW9bW78b/vocNjMCioQ/oCCoSDz6VmmT9drq8+Opz05kfXH4sYvy0qPe25bWcCdiA2J71VdzMBAREVG0DlvT3ZnFU45htZgwtiAHE4rzkO6wosbjx4t//xY3DzkPM1/+VMt45+ek4fkZg/HjiYaontffV9Vj4Ruf6+5bsHkvFo0fgEyXDU//Zb/unGMLciCEwKHq+sZx60qL63fDg0t16uQDm/bqSlh+M34AzCaBLGfL2hhGBpMtrS9uSbAYfuzdZTVaV5QLe6TBYbO0ONCM9zW1RHvVV8fz80tERESnJHN5yVlNUSS8/iDuuqIfNpeUQZES/bu78O/XFiDTZUNRXnqoHGR1CU56/ch02fCTbCeyXXasmF6MDXOGYcX0YqSnWg2Dsr7ZTqz58HvcMqKPViYwtiAHd4/pj8krPtRKFY65fRhboO/K2FxZRXhwOW9UXy3gVs89d00JGvxBNPiCMcsg4g0mW1Ky0tJSjMhjq5NDHTZLqwY8JCJAbmnJDsWHfciJiKitMeg+g1ryi7zK7cM3FW48/Zf9mD/6fADAf/7pS3z1Yy18AQX/NXkgJhfnYsRPMlHbEMAPVfUAgF9ecwEWbSnFTc99hEVbStHVYTUMyixmE3Z8W6VlbzfMGYYHr7oQdzS2wAOgBfW/uragRfW7anBZlJeOvtlOw0Cz4qQXU174GPsqooNeRZFaH+zIdUcGky2pL65y+/DU+6de78JxBXjq/X0xp0i2Ze1yS15TS7C+uu21R508ERF1fqzpPkNaWs97qLoe5dUe1Hj8sJlNWP9JKCsdXlP84i2D4QsouGPtLpRXe7Dy1kt1pSRAKHt91xX9ML/xMWp5R373tKje2qtvH4LRT/w1ai3bF4zWRrjHU79bWevFv7++F7eM6ANfQIlaU26GA4/deDHMJoEeXVNgauwrnuUMlSvsO1qLp97fF/V6188eCrPJBH9QgdVsQo7LDovFFHfJyNETHnxT6dYdc/GEQpyf7UT3ro6ox6vft9OtXVa/90avqbmhSPGcm/XVbYt9yImIqLVY050EWlrPa7OYUe8LItNpQ1CRmFCcF1WvfbimQRfQptrMhm3t7r6in25CY3aaHUdrG5DpsuHNu0bC4wsiqEgcbqI1YUuCjUynDb+6tgBTXvgY2S47Fk8o1AWaT04aiK6pFlS7/Zj+4ie6ALR7F7v2PlXW+rBwXAEynTb0yUrFkRNe3TCa5dOKcUH3NFgsprjWF5SIeg8XbN6LjXOHx3xOW9Quh3/vw1/TOekO9OhiPC2zJRdprK9uW+xDTkREicDykjOkpb/IM502nJeZim5OG+p9QeRmOLSyCLXtX2SQXePxG5YvHKvzYe7qEtz36h5kp9nx4v8dwOeHT+KHqnrUNQTgtJsx5YWPsfjtr6JaE7akVEFRJCpqG1BeU6+9vvANiBvmDMP62cNgswgcqfFG1XrPXrUTHp9+8+Lc1SWYuPxDNPgVLeBWHz9vTQkq6rxxrQ0ApJSG3wMp265+16iEKHJDpvqapJQxM9KxLtJilcJQ22GdPBERJQIz3WdISweKmEwCvTOdONngg8XsxElPAIu2lOrKImTjMdRjLt92AEsmFuo6hTw5aSAUKfHHu0bCYbPAbAJuGnIeHnv7S61t34ppxRjxk0xsLCnXAuT0xlrwnl0dcZUqRGZmV956adTrNZsEvAEFmS47fAHjADgoQ68p22XXWiTW+4IIKMaPDwSVeL8FMb8HQUVq5QQtaeMXWdaR4bBGleyo2fuWDpOJOUjIH9p8yvKRxFHr5CO/j6yTJyKi08Gg+wxp7hd5rLrc9FQ7fEGJKc9/HFUW8Yc5Q7FiejHmrg5lgCvrvMhOs+O1O0agwR/Elz/W4rG3vwIA3P+zfK32Ww3aK2t9oczrmhKsnTUUALCxpBxzV5doNazxBneRmdmlW/djycRCbPvqKK4p7IU71+1CtsuOe8b0Q++sVOR0sWNsQY42Th0I1Z9bTAJrZg2FokjdhcH62cMMA1eLuekPayL7hUd+D1ZML8ajb5Vqx8122fHjiQY47WY4rLHbAhqVf6ybNdQwO/3a/BEtDuJiXSAcqKiD2xs4rd7e7aEj1Z2bTAL53dPw+vyRHWK9RETUMXAj5RkUK/Boqn4XAMpr6nHZ49uijvfB/ZcjzW5BnTdUk20SwLE6HzJdNrhSLLhx2Q6UV3uwYnqxliVXqeUq6mCcTfOGIzvNjsM1HqzcfhD3XpnfomxvvS+Ar36sxfJtB7C7rAYAMLk4F3decT6mNtZ2P3x9AardfqTaQvXqed0cePXTH7Di/77D3P/XG1OG9UZlrRdVbh82l5ThlhF98MS7+7C7rAaPjLsAg/tkxazpjrW2yPd11W1D4EqxwB9QIIQAIDHisVMDiOIdgGO02W7TvOGYuPzDqHVsXzAaPbs6WhR0Gq09fKhRR9rUl4ihQERERMmIGymTRKwNb7HqdzfOGYbj9X5U1nqNs7wmEwKKxAmPD8fdfqSnWpHhtMFqFggqEqtuG4LH3v4yZtu+9MYx8rkZDq1WuMGv4Bc/7Y9+2a64Au5YgeHushpcfXFPVDaOmV8ysRAeX1Db+Jmb4cCSiYW4ZWQf3H5ZX1Se9GLqCx/rjvPKjoOYN6ovlm87gH49umLp1q+1TYjZaXb0TLPHDLhjva8zXvoEr80fgZMNAcxetRMLxxVo7+28UX2jNlrG2uxqVP5R5bsZpd4AACAASURBVPbFLCNp6WZHNdu6Yc4wrYuN+r6q50+27HGs9SRiKBAREVFHw42USUBRFDx248X4879ehg/uuxzrZw9FtssOj1/B0q1fI8VqwrNTB+k2OC6bOgiLtnyBQFBBfWMwe8OyHbh15Seoqfej/LgHkMCDV12AsuMew41h6sbLxRMKsbmkDFVuH1JtZsxdXYLjHh8qa704esKDipMNOFRdjx+Ou1FR26BtNDQKphZs3ot5o/oiN8OBczNTtUC0R5eUqI2TD2zaC29AQSAoMTdik+SCzXu1SZxqMPxeaYW2CXHqCx/jSK0Xx92hzYrhGxgP13hw9IQHHn/A8GKjwa9o616+7YC2eTTdYTxIyGizq9Fmu80lZVgxvbjNemabTAI2ixn3vboHc1eXaAF3boYDVospqXpJN9XbujN2A+HwHCIiaqm4M91CCAeAc6WU+xK4nrOOokgcc/vw0Gv/0GWAH7m+AE67CbeM6IMHNu1FtsuOReMH4NzMVKRYTPjNH7/Ae6UVWDjuoqhg9o61u7Bo/ADYLAJdHFb07GrH+tnDcNztw48nG7C5pAz3jOmPFKsJK2+9FPW+IB66+kJs+OR7DOqdiWyXHT6/ghMNflhMAsfdPt3mTLU0IFYwdUGPNLw6bzga/EFsLinTNn0abpxUJPxB41HzmU4bqtw+ZDpthl+vrPXi20o3enRNgd1iwoyXQu0Hxxbk4KGrL4SpcRhNZObZLE6tJby7Sq8MR9wbHo1q9O+9Mh/9sl1tWgscay+AxSSSKnvcVDa7pZuIkx3LZYiIqDXiynQLIa4D8BmAdxpvXyKEeDORCztbVLl92kZI4FQG+LjbD39Q31faF1RQ3VgGUlkb+tsX0AesRXnpWDiuAL2zUpGWYsUfPv4eVW4/bn7+I4x/ZjsWbSnFL8b0R78sJxr8Cpa8+xV+PNmA424fpgzrjRNuLx68Kh8/f/4jrRf4yu0HDac4xpqw+G2lG8fdPphNAjNH9sErOw7CZjEZPrbeF4RJCGyaN1xrhah+LTvNjgt7pqF7lxTD56qZ+dmrduL7qtAwoaK8dNwyog9mvPQJ7t3wGZZM1LdAXDG9GE57rPHuprgz1eGb7bYvGI3X549EfljP8F4Zqa0aFR/vecJbK6rORPY4Voa3qWx2Z5uayXaORETUGvFmuh8BMATANgCQUn4mhOidkBWdZWIFK6k2MwQkFo4rwDldU6BI4M51p7qPLJs6CAIAxKm2gUYbAVfdNkTLAKvHnrumBBvmDMPSrV9HTUhcMa0YaQ4Lfn9zEdIdVngDwajHPDNlEKRUYBbAsqmDdNMuwzf7bZw7HL2zUvHIdRcBAvjvmy7Bv2z4LOxcgyAhcHNjgB9ey33PmP74+MAxLP3gAF66dTBWTCvWSlDCHzehOE97vwDo6rLLqz14/J19oU8IuqXim8o6/M+fv8a9V+br3hc1CKz3BhEIKlg0foC22dPeRM34mRpKY3Se9sgeN5XhbWo9na0bSGcslyEiosSLN+gOSClPhLo9UFuKFaxIADWNvbkXjivQdR/JdtlRVedDnywnhABennkpbl35qeFGwONuX4wAQTGccjl3TQkWjR+AmS9/qrXqm7VKn4m/c90urLl9KP7jT6G1rb5tCCpqvVGb/Q7XePC7t77Eg1flQwiBl/7+LR678WL06JoCsxCAgDaRUj32gs17sfLWS7Hk3a/w6+suQtF53bBp5w+Yc3lfrLptCI67fahy+/DKjoO4c3Q/rP3oey1jDiCqLnt3WQ1mvvwpNswZpnVqKT1Si9fmj9AFgWYTsKfshOHI+mTc8NcevaSbKiFpbj2daWpmZyuXISKiMyPeoPtzIcQUAGYhRD8A9wDYkbhldU5G3R2MgpUlEwtxbjcHJq/4SOsyov6CN8pmPzV5IDbOHQZ/MHqATKyOGhazKWattJo1bipoP1bnxYTiPByqadB1JQk/R5XbhwevykeDX0GWy4a7ruiny4qvvn2I4bGPu314r7QCD119IWa+/CmenzEYDX4Fy7cdwB2j+gIAJhTn4ZkP9mP2//sJ7r2yH7wBiZW3XgpFSsPXW+Px687hDyjolZGq3Xeouj5qwqf62GTMYLZH9ripDG9ny2Y3hcNziIioNeINuu8G8O8AvADWA3gXwKJELaozauqj+fzuaXht/gg0+EMlGyYBVNT6tJKRbk4bNs0bjiq3D11SLFEbJ+/duAfrZg+FxSS0x6n9sjeXlOGZKYN0pSnLpxVjzYcHcdOQ85oNUH882WD4mCq3Dxf0SIPZBASCwDNTinDnut3aOZ6cNBCbS8oxffh5eGDTLm2D49pZQ1FT78ePJxtw9KRxK0S1q4rdYtKyqFVuH66+uCemN5aEFOWlY96ovhBCwO1T8MCre1BZ58XyacVa5j/8Iubxd/bpzhGZlbRZQuUkRusBQn25I4NIo4mU1R7/GQs6z3T2uLkMb0fLZre25eLZdIFBRERth8NxzhCjYSpGpQvq4xaOK9AGxIRntZ+dOgi/fuMLrYQDCGW/fzv+oqiJk6/sOIhfjOmP7DQb/EGJgCLxbaUbWS4brnt6OyYX52LqsPN0AfmzUwehzhvA4++EykTGFuTgnjH9dUNp1GPfPOQ82CwmvLLjIO66oh+6pVrhDUikpVjw7bHQecqOe7B0635duzt1KE9RXjoevCpf1xlFPfbdY/rjwrDBN4GAgrKaeox+4q+G2f7wWvLHbrwYbl8QF/ZIQ4rNjLqGgK5+e8X0YmQ5bTCZTLoBRd9VuXH0ZINuPWrAXlnn1XWoiLyIMnqfOltHi87UtaMzvRYiIkoeTQ3HiSvoFkJ8ACDqgVLKK05/eYmVLEH3oep6jFz8QdT92xeMjipzGLn4AxTlpePJyQN1myCBUNCq1lyrVt56qWF5x/rZw7BoyxfaKHU1gJw3qi8WbSlFtsuOB6/KR690B4IS+O6YG0u37kdlnVcLfmeO7IOdB4/j2oHnaPXUm0vKMH/0+ahrCCDFakZXhxVL3v0KE4rzsHzbATz6zwOiNj2G13pvmDMMNz33EYDQBcN/3TQQdd4gujqsEJDwBiSyXFakp4YuRhRF4seTDfAHFUx94WPtgkTt413j8Wu3564ObRK979U92gVNeEYzqEg8+lap9p5EBtInG3w44QkgEJT44Xh91AWDeszjbi/2lJ1Aqs2MGo8/6hOIyMd3Fsk2kKe14r0IJiIiaom2mEh5f9i/UwBMABA43YWdTSI/mi/KS8c9Y/ohKKWudEF93O6yGpzw+A1raM/LTNWOpQ6hMXrc0ZMNqKz1Yf3soejRJQVBCTw5eSCWbzuAp6cUweML4oFNe6M2agKhVoXrZw/DPet3Y3dZDf6yrxL/NXkggorEzJF94PUrut7iiyeE6tAXXH2B4aAbNbsdWb5SWeeFlMAdERniLimh+tjwjORNxblYPq0YQiDqE4DFEwrRJcWibUI12sQXGWhFbgRUg8mpL3yMJycN1F3YqI/3BYIIBBQcqWnQTdd8duogZLvsuvewqXrwjhq8drQSkljYgYSIiM60uIJuKWVJxF3bhRB/TcB6Oq3wzVdqhtlo4EyGw4p1s4aiotYLl91iWEN73O3Df990CbJcdljNAmWNx4h8XIM/iIevL4DHF9RqodV6ayml1n+7X45LVyetZo8tplBQDAD9clwwm0M9qC1mE6Y0tvkDTgXW62cPQ8+uKYbBTLrD2hicFuP3f/laW+MzUwbhb/uOYuPc4ZBSRgWgaseMbJcdl+XnYOnWr/GrcRdF9TZfsHkvXp45BEsmFqJPZioCjUFtePlIrAmViqJogf2TkwaivNqj1ZVHvqcOmxmHT3iiLizUgUThgXqsjhYsbWh/7EBCRERnWrzDcbqF/ckSQvwMQI8Er61TCd989fSUoqjNkLNX7cQxtxf7K+sw5YWPMXH5h1jy7ldYPk0/rGXJxEKk2s1IS7Fg2osf43/+vB/dnNaoITDLpg6C3WJGtdsfda77Xt2D7DQ7bhnRB4u2lGJ/RV3jFMcLsGhLKW567iMs2lKK4/V+bJo3HH97YBTuvOJ8PLrlC1zx5F9RcbLBOHiVoZpxo0E2OWn2UHvBD7/DhOI8bJgzDH+YPRQefxCXX9AdFpNAz66OqIEyakbywavycdztw+3/9JOYEyxNAnj8nX0oq/boRpEHAqGg+kCF8dqCEloArAbb4ePh1cetmF6MBl8QFbVew/P3znLGNQCGw1XaX2cb2ENERMkv3vKSEoRqugVCZSUHAdyeqEV1VupH84eq65HtCgWhalZ5+bYDaPArWlZ34bgC5KTZkemy4YlJA5HlssFqDl0j1dT7kea04tmpg3D4RAPe/cePGD+oF/4wZxgURcJsEvD4g+iSYoUio9sIhm4LrTxj+bYDeOqmSzDtxY/1PbtXl+CJSQPx8+c+0ko4Kmt9MdsQ+oMSS7fux+IJhbrSj2VTB+E/3/4S75VWAAA2lpRjbEEO7hzdD/e/uqfJbK/NYsbYghy47BatC8rKWy81PP/+ijpU1nm14FUNZjfOHa69r5Fre37GYMiw90gNthds3osn3t2HpyZfgp5dUxCUEkdONGhlIUbnd9rMcXW0YGlD+2MHEiIiOtPiLS/pk+iFnE0cNnNUecmSiYWwmwWyXXZdZ4637vknpFhNOFbnQ3aaHT9U1WubHZ+cNBDdUm24duA5CAQlnHYT/Ai1HHRYzQgoEpkuu2GAGFROBZq7y2p0t1Xl1R5kuWzav9Xa7K2lR3WDajaXlOHuMf3x9t7DmDeqL7qkWLDy1ktR5w0gO80Oswn45TUF+NW4AggI+IMKbGYTfrvlC8P66vCa4UynDb+6tgBTXjh1QbB0634smVho2PVE3bQZ/hoCQQVPThqIGo8fb+w+pF3s5GY40LOrQxdE7y6rwRPvhqZYXtAzDcfdPvw8bGLmM1MGYdd3VVHB+4rpxchyxTf2/UyUNnTUmvEzqbPUpxMRUcfQZNAthLixqa9LKV9r2+WcHQKK1AJGtY7aajYhoEj88poLce/Gz7SvdXVY4Q0oqHbX4/6Ne7TOIk+8uw/3vboHi8YPwKQVH2oZ5bf2HMJl+d21Eek9uqRg7ayh+F1Yx44nJw3EkRP6/ttmkzAeomM6VYFUXu3BOV1TML6ol74F37Ri9Mqw4/ILcnTDb5ZNHQSTCPXxdnsDUKSMamtYWevTuoMYZXsVRcJkElrQrPYff/ydfVg/exiEAExCwCyAR64fgEfe/FzXTjE3w4HDJxpwwuNHptOGeaP6Yvm2A9jxbRVenz8SJpOIGnZSWecNTc00iaja8TvX7cKq24bgsbe/xMJxBch02pCTZsc5XR1xB7UZDiuWTyvWtRdcPq0YGQ5r636gIrBmnIiIKPk02TJQCLGyiedKKeVtbb+ktpUsLQPDhbcFjOw3vWJaMU42+OGyW6BI6Hpoh/eiVruBhLffy81wYMOcYQhKiZOegC6oWzGtGF0cFnxfVY8n3wttZHzwqnys3B4Kzi/okYagIvFYYxmImn0/J92Bf/nDZ5g3qm8owOySgkcb2xCqcjMc+MOcYfj5cx9FBe2P3XgxHnrtH1g2dRCe/sv+qOepr0O9Hd5JRFEUVNb5onqEq+/Bq3OHQwnbfKkoEl8drdU9/tmpgwBAF+wvmzoILrsFvTOdWhBqlBk+csKja/OoXiD1z3EhKAGn3QxLWK/veFXWevHvr++Nann4uxsK2yTzynZ4RERE7aPVLQOllDMTs6Szl6JICBHKKs8b1VcLuIHGOuo1JVg0fgBONjTo2vhFtt5TyyNqPH5d1xGJUHnJvIjuGnPXlOgCXAAoq3Jj4biL4A0ooaCzzof7f5aPu6/oh2N1PjhsZviDiuEgmugMtfHmxp7pDmS77Ji/dhcWjivAe6UVuvXmdLGjKC9dGz6T4bBqWdrIVobqe7Bo/ABkp9nx8Juf6/ptd+9ix9KtX+tq5Rv8ivbJgXqM+Wt34bU7RugCZaNSg/AyEKMLpNZmj32BIN4rrdBdgADAw9e1TXtB1owTEREln3g3UkIIcS2AixDq0w0AkFL+NhGL6qzUj/2fen8fFk8ohN1iMtxQmZ4aKjMI7+mtBalpdowtyEG9L4glEwvx+q5DUcHg8mnFhj2jwzszTC7OxUW90nFzWL3ykomF8PiCyHTacPhEA5Z98A1+Ne6iqAuD8OAfOFUjblSe8kNVPeaN6ou5q0twfrYL79/7/2A1m3GsccPji3//Fov+eQAyXTZkO+2o9vi1soh0h9UweOyb7cTaj0JdUG7/p5+gxuPH/+4qw4wRfbTbj739FXaX1WDDnGGGx2gIKM1+v9Syk6fe34cHfnYBTnj8WDiuQCtxmb1qJzbMGRYzEI4VLLekprs1pSLx9oQnIiKiMyfeloHLAdwE4G6EOphMAnBeAtfVKamt4t4rrcAT7+5D9y6hft3hbfoevCofOWk2ZKeFNkCqGVb1MdNf+gR3j+mP3pmhTYBzLu8LX0BBtiuUpS2v9mDemhLcM6af7ty5GQ7tmAAw5/K+Ub2mH9i0F8fdfviDEou2lOKWEX1QVWfcHk8N4NXNhZt2/hDVYu+ZKYOwdOt+LSt/3O1DvU/BtBdDLRHVcyzd+jWCikRFnReKcipjrrbvi3wdZpPAZfndde/btQN74Td//EK7ff/P8lGUl456X9D4GCIU0FbWenGouh6VtV4oir7UymQS6Jftwi9+2h8zX/5UW7N67PJqD8qrPVprwvDnq8HyDcu269oXKopsUbu61rQXDD9+UV46HrwqHwvf+ByXPb7NcK2UHJr7eSQioo4t3jHwe6WUhWF/uwC8JqUcm/glnp5kqukOHwVflJeOp6cU4eujdUhPtcJlt6DBH8SxOh8u6tUFDf4gTtT7cazOZzjiffXtQzD9Rf3AGzW7CwBb77sct7z0CbJddvzymgvRo2sKpAy1E1SkhCKBy5dsi1rjhjnD0KNrCkxC4McTDfAGgnjotX8g22XXsu31viAu6tUF9d5QucKxOi8sZoFlH3yDmSP7oEfXFAghcKzWi+f+dgA3DzkPKVYTgorUpliGv5aF4wpwfo4Li9/+Er/4aX/8z5+/1spQjLL4OWl23PjsDsPjhGffF40fgHO7OVDvC+pqupdMLMSFPdNw5IS32QxyrPpotfRl5a2X4sFNe5GdZsMj1w/QBvxISNy4LHqN4aPpazw+eHxBBKVEitWMLGd095Pwn5lw2xeMRq+M1Jg/a+FZ9psMau0jhxEBYLeTdsTNr0REnUNbjIFXf2PXCyHOAVAFgG0EWyj8Y/9Hri9Adb1fN0p88YRCfHSgEllpdtyxpgTZLjuWNE5IDFde7UFVnU/X/cRsElh6cxF+v3U/dnxbhWq3D2tnDYU/qKCy1qsrI1kxrRhpKcbTLut9QRw50aD15n555qV4eealqKz1ah1Xxhbk4O4r+ukC2acmD8Si8QNQWefTXQw8O60YPbrYMWdVCR66+oKYWfMfquoxoTgPc1eXYN2soSg9UovdZTV4ZcdBrJ01FAKAxWxCjsuOo7XRw3myXXb0z3Fhw5xhWpnOud1Ske60QAiBReMHINVmRr0viO5dUuALSsMMcuRmw1j10ZlOm5bhf/j6Anj9Ciav+FB73WtuH9psXfXRk80H/a1tLxjeE95oHYdrPJi4/EPtvHaLSdeRpjUBH9sUtl6sTzS4+ZWIqPOIq7wEwBYhRDqAJQB2AfgOwPpELaqzSk+xYO2soXjjzpHIdIU2/UXWSk8cfC7uaCz72F1WgwOVdYblEVVun670ZOLyD3Hz8x9h2vDzsHbWEHTvYseREw0oO+6Jmkj5P1u/hi+o4KnJA6OmXeak2bD47a+0x9668lOkWM26Y0woztMCbvVx927cgzpfMGoD5x1rSlDvCyI7zRazXKSb04a3/3EE/XNc+P3NRTCbBF6dNwzbF4zGYxMK0aurAw6bBVJKVHv8sFpMuuOoJRTTX/pEV6ZTUduABp+C3plODOjVFbkZDgzo1RW9M53wx9j4GbnZUA16I9fc1WHFMx/sx2X53dHgV3Bf45Af9TgHjxlPvwRC2fNjbm9cZSORpShjC3KwbtZQ+ALBuEoQYq0/coDQ91X1za6lKU2V01DzuPmViKjziyvollIuklLWSCk3I1TLfYGU8teJXVrnEggo2FdRh6kvfIzxz2zHz5/7CHdf0Q/rZw9FUV46gNAvWbNJ6H75Go4jn1aMzSVlht1P5q/dBZMw4ehJL3qlO5DlskX9Mp9QnIdbV36K//jTV3jsxovx53+9DKtvH4I+WU786n+/0PW5znbZowbnxNrgKADD+2vq/Xjo6guxuaTMsO77rT2HccOgXpj+0ie4YdkO3PTcR/i2sh6/+eMXOHrSix+q63XBXF1DACumF2vHuWdMv6gLiwc27YUQoU2Lata3V0aqNmY+VjAamUE2qr9ePKEQD27ai/dKK7Bg817D93jp1v1YMa046qLmrnW7ccOy7aj3xhdkhU9O/PjfrsAvftofU174WHsvvvzxJI6e8MQMwGOtf/m2A7rzptr0r7ulAV+sTG2Nx8c65TjE+/NIREQdV1zlJUKIPQA2ANggpTwAwJvQVXVCFXXe6Czw2l1YNH4A7v9ZvtZ7OqhIjC3IwYTiPOSk2dHNaYPVLPCH2cPgDSo4UuNBThc7Zo7sA6vZZBi4+YOKVjqwbOogjC3I0bWn69ElRdsEOO3FT7T7/3Lf5aisO/WtLcpLxyPXF8AflNg0bziq3D4s33ZAy1hHlzyYDO/v5rQhqEgsHHcRbBaB9bOHIahIWMwCkBI3FPfCkRqvbgDOgs17sfq2Ifiuqh4uu0X3vs146RNsuXsk1s0aiopaL9JTjS8CenZNiTlwJnIgTvhmxvAyCavFhJ5d7dgwZxgCisRXP9biiXf36dolGg0Wqqzzomd6Cl6fPxIefxAHKurw+DunnqdmwmOVjRiValS5fVHDeuauLtHqy41KQiLHnQshDAcI1fv0AXZLAz6jTG22y44jNQ3ahl3WKcfW1M8jERF1DvHWdF+PUPeSjUIIBaEAfKOU8oeErayT8QeNyxlSbWZtsqTNYsL2/RW4e0x/rab7l9dcgHs37jlVjz29GGYTkJVmh4DxFMlgYzYx22VHVZ0Pv7y2ADcPOU8bH9/NaYsZNIdPSvzlNRei3hfEnet26+rO/7bvKJ6dVqyVwag13SaBqEmLy6YOwqLGYTrq81/ZcRAzR/bB4++ELjSWTCxEqs2MR978EtlpNjw5eSBOePwAgPWffI+7r+iHycW5GFPQXWut6PUryM1IhcNmgS8QjHo9Ywty4A9KlNeE3mOLWcDj09caRwajZgHUeHxRtdZLJhbi8Xf24Z4x/XR9w9X3LcViMgyY0h2h8xyqrsfMlz/Vfe/VTHhkQJrhsKKitgH13iAOHnNr37PnZwxGlxSL4c+Q+snD7FU78eZdIxFUEFVXrdYFK4rEvVfmo/RIre689rALptYEfEa15/eM6RfVIYd1ysYifx4dNjMCisSREx7WxxMRdRJxdS/RPUGIfgAWApgqpUz6zz6TpXvJ4RqPttFOFd5xY9v9o+ANBOG0W7TJjmtuH2LY7WP17UPwn3/6EneM6gur2ayfwDitGCkWgWN1PpiE0GqN1c4f/mAQDqsZVW5/1MCbLikWrP7we9wxqi+q3D5075KibcAMP//LM4fgub8ewNUX98S5mamorPVCSonuXVIgIBFQQmPlLWYT1n54ECv+7zsAp/pF53Vz4FidD30yU+EJKDALAYtZ4ES9D1VuvzbCPtNpQzenDX/bdxRF52Vq0znHFuTg368t0DZXZjtt+OaYWwt6Qxs9++OOtSVRgbMawKrZVkWROFRTD29AwiQAm8VkOFlT7c/94FX5eGDTXmS77LhnTD/0zkqF02aBzSLg9gYRlECK1aTrRBKrA0pkgJzhsGJ/ZZ0ueA+fwLlx7vAmf4aK8tLx6D8PaDazbJRFB06ve4lR9401tw/FqCe2RT22uc4rZzt2MiEi6rjaonsJhBC9AUxGKOMdBPBgWyzubJGVao3KAqsBVW6GA/uO1gIIfcysBlU9uqYYZjZNQmgTDScX52LlrZfCbBKwW0yocvtQ5w1lgRdt+UKXZZy3pgRPTBqIiloPNpeU6YbyqIHu/oo6HKvzoWd6SlQtt3qcqjovNpaUY2NJOXIzQqPeU6xmOGxmHKvz6TLgz04dhFEXdMfruw5hfFEvLNh8KmCt8wVxpMaDVR9+h5kj+yArzY79P57ALSP6RLUJVDedFuWlY/7o8/FtpVvrRlLfzYHzs5zYMGcYKmq9yHLZdRcLao23GpyGZ1tPNvhQXe/H/LW7kO2y44nJxt1i0h1W7C6rwePv7MPmecNxrM6nC24jg/os56lMbqzSATUTrqqsjd5cGT6IyCwQdRz1ZwiIP7NsNH0TwGllnyMztWrbxNZ0XjnbsZMJEVHnFG9N98cArAA2Apgkpfw2oavqZBRF4ocaDzy+IFbfNgSKDGWCl33wDSrrvFrg9NDVF6DK7dMCFbMwLh8xNd6f7bJjTEF3HHeHstopVpMWPN4zph8WXH0h7hjVF25vEFazCTUeP/IyHPhJVip6pTuiykD++lUFHrm+AL6AhD+gwGI2rtGuaSz9AEIBwTnpDqRYTQgEpRZwq19T69anDjsPz3ywH9kue1Tv7aenFKHBr0BA4KcX9cRv/xh9sfDYjRdjQnEeLuqZhrJqj67V4pKJheiSYoXNYoY3oODoyeiWgmrgrP5b3STo9gYxv7ETy8JxBfihqt6wVKVnugN/ue9yKDIUYEYGtyu3H8TjEwtx3O3Djyca0L2LHd0aA2+jgNQokxyrg4U6XEhEHCeoSDz6Vil2l9UgN8OBPlnOpOmA4Q8qSLEKrJ01FJW1oemjm0vKcO+V+axTbgY7mRARdU7xZrpvkVJ+FeuLQohbpJSvtNGammBa+AAAIABJREFUOp1qjxfVbp+uNnvZ1EH4lyv7YY6vL577a2iseI3Hj80lZXhmyiDcuW4XfjzZgCUTC7XOHGqACUismTUER2oasHJ7KEPdN9uFsuP1GPGTTF1G+cGr8rHqw++0co2AImG1mFB6qAbrZw+DP6ggqEhs2vkDJl16LmwWE/xBiaO1XjisJjw7dVDUYJnH39mnvbbcDAcqar3o2TUFgHH3klSbGXeu24WF4woAQNdxJdtlh8cX1L3GxRMKUVnr021WzOvWWMYCYdipZNO84VAUBT27psAXUKI2j6obOovy0lFZ54XDZkaVuwES0DZwntM1Bb9+4wssnlCorXFsQQ7uHtMfj275QnsPrWaBm4pz8eSf9wMIlc3cMqIPZr78qS4738VuhcUSahAUK7scLlZP7npfEEsmFsJiElH12b+7oRAPX9f+meXIkoixBTm464p+2gWNuh+hX7aLJRLNaG1vdiIiSm4truk2PIgQu6SUg9pgPW0uGWq6D1XXG04FVLtOLJs6CA6rCWaTCSYhcLLBjy4pVngDQZhNAmXHPVopRW5GCmrqA+jRNQWLtnwRVYrxzJRBeOaD/XivtAIrpodaC0Y+ZsW0YtitJjz+zldaINm9ix0VtV784g+f6S4Mvjp8AoN6Z8JqDgVKQgBTnv9YF4Q7bGZYzSa47BZMfeHjmDXHm+YNR1CRuOm5j7ShPn2znSg77sHSrfu1IDv8OertReMHYObLn+Iv912OK578KwBox1C7vPzurVJtw+aKacV487NyrPi/76I2cOZmOOAPSBytbdAF++p7V1nr06Zv9kxPweGaBmSn2fFDVb22sXH5tGIs/N9QF5AV04sNN1iumzUUuRmpcQeZRrW8z04dhDpvAI+/sw9PTylqdgpl5PNXTC9GltMGk8mU0M14kXXrsd4Tlkg0jzXdREQdV5vUdDd3jjY6TqcUiFEbrXadePov+/Gb8QPQ4A8ioCio94W6F3R1WFHb4Ef/7i54Awq+r6rHg5v+gco6L1bfPgQTivOi+nSrGeX3SiuQ7rAaPmbumhKsnz0MD119Ib6vqsfv3voS94zppxs3X14d6vm9dtZQrPvoOwzqnYlMpw3ZaXasmFYMX1DRja4/P8eJRVtKtSx9ZNA4tiAH2Wl2fFvpxtiCnKgLAbXEZndZDcqrPVoJQnh2vSgvXVdaE16mMrYgBw9dfSHuGHU+Kmq9+J+tX2PhuItwfVEuyqs92rFLj9Ri49zh+OZ4bdTrvXPdLqy6bQhmvPQJ5q4uwdiCHNwzpj/uD9uMqq5z3poS7UIgvA4//PtbUeuFw2bR2v0ZlZZEbmrsl+3CxrnDcbjGgyq3D79+4wutfCSeKZRG5SfqhUgiA7fIkohYvdxZItG8eMuRkh0nlBIR6bVV0M2JF00w6uOs1karGwO/O+bGyu2hTGyPrikwCQGTEKhy+2GzmBv7UVuwZFIhLCYT7FaT1m87XHm1B+dnu1CUl44ajz9mQHj0ZAMmLv8QYwty8OTkgTEH21TWenHTkPN0I8KXTR0EIRBVTlFZ64NJAP990yXIdNnw3bF6/PqNL1BZ58Wz04oBSPTNcWLhuIuiNjqGbxgcW5CD3AwH/r5gNIKK1NoHzhvVF//xp1I8OWkg6n1BrYTmsRsvxjnpDnwflol+ctJAAKGLHTVjrp7LH1SQajMbvl6LSeC1O0bAH1QAQPcJReQ6++a4sH3BaAghtN7q6sbUzSVlyHTa4AsE8V2V23DEOgDsO1qLp97fp33i4Emzo2eXFJzw+PEvGz7TPSeeWmi1/CQy85zozXiRJRGxe7knpkSiswV48ZQjJTNm609fZ/uZJiJmus8Iq0lE1WZrGdNRfeEPSAgh8G/XXAgpgf/805eorPVp7enCSzl++8e9qKzz4pkpRch02Q0Dmx+O1+O34y+CIiW6NG7Ci3yMOkb+lhF9MOOlT7BwXEHMxwHQBW/zGzdHRm52XDR+AA6faAAALWBU3bEmNMRlc0kZ/u2aC2Nm/scW5OCBqy7AwWNu3Wv/75sugd1iwnulFais9WHJpELDTZlqZv3Fv3+Lh6+7KCpQzc1wwGISyOli/N5ZzCbkdAnVpx+qro+5ztwMBxxWM7LT7PD7g1pvdW0d04qxZc9hbCgpx5KJobWWV3t0wS8APPX+vujyn+nFyM85vUznmd6MF9mhZXNJGZZNHaSr6U7UsBcGeMmHHVhOD3+miTqnuMbACyH6NHPf9jZbUSeUnmJFdpodi8YPwOvzR2DVbUPwyo6DWtlAitWE+1/dg9FP/BUzXvoEt4zoowu4gVMbBh+fWIjf31yELg4rVu84GDVWffGEQizduh93rN2F424/ROPAmvDHLJs6CFtLj+rGyC/fdgDPTBkUdazNJWVa4K1SN0dG3veTbCe2lh6NWVqglrt8d6xeO48qN8OBnC52PHzdRSg/7ol67f+y4TN0TbVh07zhmDeqLyprvaEOLRGlM3es3YUGv4Lb/+knMJsEstLsWHnrpSjKS9cy8r/54xf41w17sGSi/r1bPq0YOa5TAUGs0dz1viBWTAsNKQoEFBw52RDdtWVNCa4deA6yXXY8sGkv5o3qq3svfIEgfIGgcfnP6hJUe/xRo+tb4kyPFQ8vidi+YDQeuX4Azumago1zh2P7gtF4ff7INg8YFEWistaLIyc8hgFe5M8tnTnswNI89ef3UHU9Kmu9UJRTHxjHumjhzzRRxxZvpnszgMiNkpsAFAOAlPKutlxUZ+MOBEPt6roJmESo3GTOZX0x57K+yHTa8OWR2qgR6GtuH2r4S+uEx6+NeF88oRBv7D6EheMKkOm0IcNpw/0b92gbElNtZkx74RM8NfkSXU/up/+yXyuFUM+xu6wGaz/6HqtuG4Ljbh+q3D68suMgHrzqAhyr82HDnGHa+irrvIZjw7+tdGN8US+t7tqonCYnzY7f/rFU1yFEzVD/64Y9+O+fXxKz9KOisSRGzXzH6mMePuVz5sufasfPSbNj4Ruf473SChTlpSPVZsb62cOgSAmLSSDbadO6jQDG/bWXTyuGxxfEr/73c1TWebFiejEC/5+9M4+Pqjzb//WcM/tMNkICSCIiBiRiEAZCgL4WpVUpKGVVISCgBETcqqLW8tO3aMta6waJtIJsCoK+VtxFsC24EVGUyCKLJqwhZM/s5/n9MXNO5sycCSFkmST39/PxY2Y788yZaK7nPtd93T5tz/65an9D5qy1+UpcoXwuZPEbyf5zseKksceK1+dSd3NaIoIrgcsmaGerk8BrOSiBpW7OV8mmTQtBtE3qFN2MsSsBXAUgjjE2NuihWACmplxYW0JgHCcqXJi9Lh9DLk9Ezq97oGOMESIDzla5VZnTsu2EMWjmRceZ9YoAlgfayJ7lT/5wrSoBpMzhQVGpAxLnKl8zAORc20PJBE+yGZW0joqAF7eD1YAnRqajrMajaiRcPrk/4ix6lAXliQevW27yDI0aXDahLxa+vx+Lx2eguMqFpR8eUDYCNW4fYs16FFe5IAoskNJSt9XlgY3fYmNOVkRxH1yNlyvgr+dkKYL7yVvSAyPuv1DZOrrEmVRDazrFGrExJws+7rcJPfmvH1RRhLPW+ptSI61XtqLIm5Qb0pPxp5Hpyqjv5Bhtm4tP4pAk3uDKcGM240Xjpe7gSmBz+8eJ89PYm762xvnsN7RpIYi2yfkq3b0AjAIQD+DmoPsrAcxsqkW1NWpcEmavy8d91/XAkLQkcM7BwcA5VwbUALWNegtG94HEuTKJcZw9FSkJZnCubl6UR7cDsldZUH4OnnapVZWOtxiQaNVjw8xBKAtMZAz2ji/+4AD++LveqHJ58cLt/VRJJR1tBnSONeG1mVk4UeYXPXI6CAAIjEHiXJlUGW8xoNrlb4Z0+ySlyj1rbb6yVoEBq6cPBMCR0sEc0QMvU1TqgI/ziM+TxXfw891eCSkJZswe1gOl1Z6w9JJZa/2+9M5xJqQl2cJGsudm2zF9aPewDPEKpwcrsu0qT7ccUXh7ZjesnDIAneKM+PLx63G22o1J/6iNXHxzzmDkZdtV0y0XjcvA0+8W4JkxGRc9JbIxKs/R6M8NrgTm7jgcduWEBF7L0lYSWJqK81WyadNCEG2TOkU35/xtAG8zxgZzzj9vpjW1ObwSx5DLE9G7azz+/M4+zLnuCpRWe9At0YL5o9KRu+OwSsRd1tGKGpcXiTY9nhiZjuJKFwyigCUf7g8T6KumDVSEskHH8J95w/DTmWql6rx6+kCY9CJ2PDwMosBQ4fAA4PBKPkiMAYwrgls+7iOb92LJ+AyY9AIe3KSuwr/21c+49/o0eCUJZyrdmlnMxZUujM/9XMnR7mA14Fy1By/cfg2cXo5Xdx0MG0F/e2Y3JMcaUVrtwScFJzEp6zLF+sE58Jf3CpRzJL+Pz8ex+IMDWDqhLzrFGnHsbI3yubWG+JwocyA32w6nxxdxxL3FIGLmmt3YNGtwmNCUm0UfvrGXssmQq1Fb8gvx2swsnK5wKtace69Pg9MjoVOcfzplcaULs9bWbrKSbEacKnfhxU8PKRahDlYDcnccxkcFZ/Dkzf4/wFrWDgCado+mSDyIxkvdwZXAPYVlWPrhASwY3Qc9km0w60ngRQOtPYGlKTlfJZs2LQTRNqmvp/snxtgfAVwW/BrO+YymWFRbQxQY5lx3BbL/+SWWjM+Aw+1TWUqWTegLiXMIzG+tYIwj3qJHSZVHNapda1KjXhSwenomRAH45ZwDqQlmuH0SHhtxJWxGHRweCdNW1VbH5UE8VS4vymq8iDHpNAVV51gTpgRi7uT75Lg82aoRZzYgL7s/Zq1TV8k7xRrx0G/ScG2vZCWz+4b0ZDwxMh0GkeGJkemqQTYqa8qMTOT955j/n8BwnznXXYHpQ7uj4GSl8j7PTuyLUxVO7Cksw20vf4F35g7F5UlWLJ3YF8WVLpj0AoqrXACgqjzP/J8e6Bxnwk9nqiJaU5JsRnh8Up1+cXmwkeyrH2dPhdcnIc6sh1En4PERvfGHgL9+56PXAdZw8Tp7WA9lwyNbVlIS/IOBdh0pgUEnalo71szIhMsrhVXBtKrzjWEDicZL3aGVwOIqFzrHmZASbyZhQkQ99alk06aFINoe9RXdbwP4D4BPALR4Jwdj7CYAzwEQAfyDc76whZcUEUni0Iv+Bsq8KXbEm/U4We5UVbj/+d8jmHdTb5QExpMzxuDySprWk1XTBmLe5r3YU1iGG9KToRMZzla5UFLtxpb8Qtw/vCcOnqzAsk8O4eMHr1XsKPIx5qz/BqunZyLebIDbyxVfd5inmNc90Od4qQMPvfEdcrPteD1nEKqcXoiCAH9kO8Pofl1xuLgaSTYjkmxG3DGkuzKtMiXBPzHy3uvTcKLcqbKmcNR62eW0k7kb9iDJZlR5wFM6mHG20q0I83KHB2aDDtNW+TcK/VLjlepxcqwJT2/dh/uH90SF04MKpwcdrPowa8qyCX2xJb8I827qhSPF1XX6xdOSbZg/Kh1v7zmO0f26qqwNS8ZnoMrlVSrhep2A4sraDYB8zEgpL4lWA/Km2JWhOqEV959LasKsMZGq841hA4nGS91alcAEs55yjYlWAVWyCaJ9Ul/RbeGcP9qkK6knjDERwEsAfgugCMDXjLF/cc4LWnZl2pRUu2E2MFS5OGxGHU6WO+H0+NDBYsDSiX1R4XDD7eWKWJSTNiSuPaym3OHBwzf2wr8PnMaoa1Jw28u1jYCLxmXguW0H8dTNV+HKLjEw6ATlGLLVIz5QiQU4atw+fHOsRJVYsiW/EHOvT8OpcmdE0RksPmXLRaLNgHe/+QXX9uqER7fsVq1JYAiLAJQnRnoljtnDeiipKDpRwAu398O9r+1RDfeRPdcynz70a9z84k7lfMVZ9CrLyJ7CMuX52x/+NeaPugoAx9dHz2J4ehecrXLhkngz3pjlb5J0eyWcKHNgwoBUPLjpWyTZjGE+4WC/uE5gWLC1APNHpYdF/j2yeS8Wjr1aEadVTi+mvvIVkmxGldCP1DDaOc6EzjH+PmW316dKttlTWBYx3cUboTp/sTaQaBUIwZXAaGz2JIi6oEo2QbQ/6iu6tzLGfsc5f69JV1M/MgH8xDk/AgCMsdcBjAYQlaJbkiR4fCzMKrJkfAYe3vSd5vj1u9d/o3i1QwWZ3x/txtQh3TWnJa6dkQkmMHRPssEncSWdJHSITF62HVd1jUGizaCalpibbUeXOCOSY4wRmwODmxply8WcwJpDK+vymoI/hzyUR55KKZ8Ps0HE01v34c+j+2DB6D7oaDNEHO4jyyj5fC2f3B/xgRzvkmo3thWcxvD0Tki0GuCTgAqHC24vx4DuHZX3vSE9GXOvT1M1keZm25FkMyo+4YVjr0aXeDN+Kan1i6+cOgCXxJnx1pyhqHF7NYXupR0seGvOUIgCcMuLO5WNw+IPar3HNqOoGUloEAWUOdw4HeQBDxb9wVcDgs+JKDDl88sCPdQG0lDPd7QLhGhs9iQIgiCIYBjnkSe4M8YqIfsFACsAFwBP4DbnnMc2xyJD1jQewE2c87sCt6cAGBQpK3xATAzfbber75w4EZgzB6ipAX73u/AXTZvm/+fsWWD8+PDH774buPVWoLAQmDIl/PGHHgJuvhk4cACuO/0hLwUnKuAKVBzfH30nht0zCT1OHELS/Mdw+EwVqlxe5eWLr70DL744B47P/ouKB+fB5fXBqBNxeZIVosAwf9hdmDzrFjz72Arcu+t11Vv3TY3HmSXP4UyXy/Dl86sxZdcWSBw4drZaef8HRz0EsduleCuxEOeWPY/SoIELRp2Izh/9C874Dtjx+BKM/vZj6HUCRMYgCIDDLWH27X/Gl6edyP7mXYz7aRe6JVpw8HQlrkmNx7eFZbhtkt/tM/PLNzH88FfomxqP/Scr4fL64NQZ4XpnKxZsLcCYra9g6M/fKe97eZIVu8qAXv/9ENcv+wzzPluNqTgJADh0ugourw/liZ0Qu9n/mX+89U6knzkCm1GH7h2tOHi6Cj/Gdsby2+dh+eT+KMueDsORn2DUiUjrZIPEgQ90nTDvV/5WhF0/vILTBYeV8wIAh3pcjZQVf8f01V9jxVt/QYKjAglWAy5LtAIAvMOug/nPT/mF6ogRcFdVY9/x2u92W49MvH/TZL/Yu/lGuLwS9vxSqhx/65X/g3X9R2LXvYNwyeTx4AA8PgmSxOHwSFjbaxg29RmOdaMvR+nvfq9am1EnwjB3DjrMyMaJ7w9BuOMO5XcjrZMNx8scWHzVSBwYOAzP97PAev89fnuLIAAM0AsCfpnzICYfi0Hs/n145rN/omfnGFgMYu1I2b/8BRgyBNi1C/jjH8N/t//+d+Caa4BPPgGefjr88bw8oFcv4J13gGXLwh9fuxZITQU2bgRWrAh/fPNmoGNHYPVq/z+hvPceYLEAy5cDmzYpd8vnOfR3r9+lCYErOwDMZuD99/0/L1gAbNumPnZiIrBli//nxx8HPg/pHU9JAdat8//8wAPAt9+qH+/ZE3j5Zf/POTnAwYPqx6+5xn/+ACA7GygqUj8+eDDw17/6fx43DigpUT8+fDgwf77/5xEjAId6s4dRo4CHH/b/PGwYwmjG/+9h1qzwx//0J+A3v/GftwceCH+8lf7uKezY4f/30qXA1q3qx+h3j3736Hev3fzusc8+y+ecDwh/8nkmUnLOYzjnsYF/C5xzc9DtZhfcAbTKcqqdA2MshzG2mzG22+PxaDy9eeGcK+LJZtRhwoBUzH/7B8xe+w32n6xEagcLbMbaiw7JsUYcOFWFlf8+gss6WtE3NR6XJ1lReK4Gh05XYcKAFJRUu5FkU3tqjToRDrcPj735PUx6AeUODwrP1cCkF1TiDZCj7rxIsqmrgC6vD6crXDhT4cK+ExU4eLoS+46XY29RGX46XQV9wEMOAAkWPS5PsuJEmQPGQDW1S5w6vt2o8wu6np1synO0BsK4vD4wBpj1IqwGEdsf/jVuH3gpOAcEBlzW0Yr0S2LRNzDUJjiZpGuCBQcDolz+bHPWf4PuSVbl2IcCa69y1m5uDLrw81Lp9KfKyNMcjYEmQoMogDH/d3mqwonjpTVweyXoBKb6bPEWPXKz7UgIDMNhDMpjAJCaYMaqaQPh4xxun6Tc/+PJShw4VYFfzvlHz5fVeMLW5vL60CnOBIkDKz47jG6JFlzVNQ69L4nF8TKHsnkqKnVgyUcHcFmiFUWlDnzzS6myMXj244PKuXd5fTh4qhKeoHW0VkLPMxD43SNnCUEQRLuCwx9P7PJKqHR6VdNeW5o6K93KkxgLnUYJAOUAfuacezUeazIYY4MBPMU5vzFw+3EA4Jz/Vev5AwYM4Lt3727GFaoprnTB7fUpVpC8KXbNmL3g6Yly3F3woJv5o9IVj/LGnCwsfH9/mGVEth/ItgL5mJHeU240HJ/7edj9ADRf8+zEa+Dy+tA1wYxjZ2vw/LZDKK5yKdaT+aOuUtlG5PunD+2OlAQLTlc4EWfWq2wowecgpYMZZdVuPLipdiDPC7f3Q5c4E9xeCaLAwMFx6HS18t6vzsjE8GWfhZ37jTlZuPXlL5Tbnz70a8VKAyDsvPRLjcd9w9NweZIVAmOQL/J0shlwuKQGz358AHcM6a4652/MysKZShfOVXtgMYiocfvQwapHagcLOliNKq9xks2IeTf1UjVvrpw6ALEmHYYu2q5ae6Tv7K05QwEAY5bvVGxDRp2g+g5lNs8erLo/9LbMzkevQ9cES9j9rQnydBMEQRDR8LeAMRax0l1fT/dy+MfAfx+4fTWA7wAkMsZmc84/uvhl1puvAaQxxroDOA7gNgCTmvH9z0uwb9ZsEBFvFpCbbcfsdfkREyu6JVrw2SP+LO17N+xRZVLLSR5AbTPjnsIyvLrrKFZNGwhRYDh0pkqVAiIfMyXBjNwdhzWHyLy66yieGJmuOVkSgOZUSb2O4cFN32PF5P5KNGHwcJwnRqbj9ZwslFS5carCiVd3HcUdQ7pj1c6jmHt9GvSBvPHQJsXcbDsSbQb8XFKjTMCUP8e9r+3B0gl94ZN42Gtizbo6x84H32YAXpzUD3M37EFRqT9be/nk/piz/htNQZw3xY4rEq0ornbDpBfw/26+ChLnWDVtIKpcXpypdMHl5bgncLzg93o9JwsOdw30OgHJsQZsmDkIImNhPnw5dSR0/VvyC/HSpP5K5GJoYsjKqQNwqtypxDhGmogZTEm1GzekJ2OcPVXJSN+SX9gmptxFa7MnQRAE0XxEe39PfUX3MQB3cs73AQBjLB3AIwAWAHgTQLOJbs65lzE2F8CH8EcGviKvKxrQ2mV9eP8QdE/0jxOP1AQH+G0UnEPJlw5+XE4NkavgKQlm3DGkO5Z8uB+P/663ZlVUYMDaOzMhMH91+PWZWXD7JPgkjs27f8E916Vh+4+n8HpOljJy/FyQUIs167H2zkz4JI5T5U4sfH8/kmIMWDXNP3Dn7qChOrXvyfDJvpP4zVVdEGPS4ZEbr8Tm3b8g59oeiDXrUFzhxuMjekMnMryekwWfxHGkuBrz/+8HPDGyNxi0U1s62gxK3rh8n5ycktrBL5CDmw6XT+6PFz89pKxr0bgM/PX9HzF18GVYMLoPLutohUFkeDO/CAvHXo1uiValQi8ff9bafGyYmYVJQZX7Fdl2vLDtoJIxvu6uQZrrdXkkDP/bZ6rv7ImRvTWfKwbiJIPXf8eQ7nhv73Gsv2sQdAILE5G9OsXAavSnmGhNZMzLtuO5bWpv3TfHSnDv8J6q5thgK0xrJ9qbPQmCIIimJRqHuQVTX9F9ZbCw5ZwXMMb6cc6PsBYwTQZSVKIhSSUMrV1WUZkbVqOIWJMIh5uHjf1ePrk/PD4fPD7AZtKFpYbkZtthMYhYOqEvusab8Off90G8WY8KpwePjeiNrd+eCHtN3hQ7THoRroDv+FyNF3evq00peWlSf7y39zhGZFyC0hqP6rXP3XYNRIGpcrUXjctAUowBd/7qcszbvBdpybawSnheth2McaR1jlNFGeZm25Fo1cPhkbDyP4cVwbpkfAZ0goDpq79Wzp1BFLSTORjT/A/JYhAxbdXXyMu247WZWfBKEkRBgEEEHrnxStw97AqcqXSpKvFHiquxYvtPGGtPwc39ukLHGEqq3ZrH93gl1Xd597p8zB+Vjo8KzqCo1BExWvFkee1rVu08isXjM2DWC1g1bSAsBlGJACyucoEJDEadgAWj+yDeokeMSQ+jjmFk364wiAIkDQuYIDCY9TqkJJiVpBXZLhRn1iPWpMODv+2lGig0OesyZQS9vLbZ6/KjpgJAEARBEBdDNA5zC6a+ovsAY2wFADkq41YABxljRvjTTIgAWrsss17EodPVmP/2D1gyPgM9kqx4Y9ZguH1+j7IoMBSdc+Av7/2IxeMzkH/0LDbMzAIPTKl0eX2QOEe5w4NF7+9HcZVL8XjLXuh4sw4Lx14Nq1GHBIsB5Q43xud+7hd90waGxRLes+EbzB+VjtJqT9hj97/+LRaM7hMW/ffazCyU1tROxDTpBVXG93PbDuL+4T3x6q6jymuTbEZlQmThOQfmXHeFMlXzkc3+OEGZ3B2H8eQt6WFWmCXjM3CqInJueJLNCIlzlZf8pUn98dL2Q8qkR/n5Psk/7fPuYT3w1/d/VDYAa2Zkah5fYMBEewpmXns5RIHBJ3EE9+0uen9/WJX677deA49PwsacLGXS6JIP9+POX12umkTqn95pgk5gqio+ANyQnoz7hvfEhLzPI/rS5CE6s9bmY09hmTIhc8mH+/HMmIwwu0W0VwAIgiAI4mKIxmFuwdRXdE8DMAfAA/Cnh/wXwMPwC+7rmmRlrRStXRYH8Py2Q1g0LgOrdvo9zq/uOoo/jUrHbS9/gWUT+ioNf6LA8NTW/Xhq637VcTfPHqyI7NCc7O5JVvzfN8fRs0ssOscEZJdWAAAgAElEQVSZ4JXUPuNIw1Rkn3gkj3neFLuS91xU6oDL60NcIDd79rAeKCp1qgQ7ABScrMTCsVfjo4Iz6JcaH9bsuWR8Bubd1Au3r/RXXIObivcUlmH59p/wp1HpWDsjEz7OUVbjgV5keG/vCZXH+Yb0ZDw2ojfKHR4smdAXiz/4MWxTsWZGpqrSu3xyfywMCG359XIlfONXP4d5qHOz7TAZBEwdcpnS+OkfxmPHRHsKNuUXISnGgCSb3zrkkzh0ooDiSice2LhXteGZPyodD4V41R/ZvBdvzhkChztcDI+zp4ZNJNXypSXZjFh/1yBI3G8BWrXzKB78bS/FihL83OJKV1RXAAiCIAjiYoj2/p56iW7OuQPAssA/oVQ16opaOVq7LJ3AUFzltzjMHtYDsSYd5t3UGzwwdVL2axeVOpSBNqHCqEucCTseGYajxdWqhkl/DjPDgO4dlOrw5tmDVa8PPn7wMTtYDWCMYdW0gXh+2yHVMQ+dqVIqp0s/PICkGANMehE+ieP1nCwIDCg851CNZ5ctE6kdLJhoT8Hw9E6a0xrl6rZf8NXaSW5IT8bjI3pDkoBjJbXJKP+4w46RfbvixU8PYf6odPRIskIv+kerl1S7sWLHT7hjSHelgi6/lygwLBx7NfSigA5WA5Z8uB8fFZzBRHsKpgzuphoK9Pdbr4HVKGL+qHR0jjUh3qLHM+8WYJw9VeWXLyp14O71+Vg1bSDKHG7cN7wnxq7YpRL2L23/KWzDE6mB1uOVlI1aks2oTA1NtBnrrEpr9Q7kZduxcFwG4s3a/4OJ9gpAa6Chw4UIgiCI5iGa+3vqFN2MsU2c84mMse8RkoUNAJzzjCZbWStFa5clSZJimZCr1UvGZ6Bbh9p0EbkRbuW/jyiJGsHNe3qRgTF/trTcaClXvUPHrJdUu1UiWyu9ZPnk/ooIDW72k+P/ln54QLGVLJ3QFxaDGObTthl1YYkfS8Zn4EyFC9mDu6nGsssUlTrAmH9y4iXxJnAOrL9rEADA6fFhSpAQXj3dn8wiScDJqhoUV/onLT51y1W4Z0Nt5VlOYpk9rIcSq5iSYEa5w4Nqtw/xZsG/8al0o19qPOYOvwKTVqq9zQ9s/BabZmWhV6cYiAJTrCp3/upyzc9g0Al46pY+mBiwf8j3z1n/jeL5Bmo3PJE2PrJwWzMjE6crnMq5jDSRVK5Ka/UOzAr4syOJwGivAEQ7ofGP9w1PQ/eOVliMIjpajXQeCYIgiDo5X6X7/sC/RzX1QtoSobusX0qqsfiDA6qK8OIPDiBvan9FYC/98ICSqhFrErExJwteiUMUGNbuOoq8/xzDqmkD8dpXP6uO8+quo/jj79JV4iw0zaK4ygWbUYfXZvotEKLAsGDrPkUYyhXo12ZmoeBkBd7ec1ypuJY5PEjtYMateepkj9nr8vH6zCylITT4OAvHXo0HN32LjTlZmsJR4hybvi7EmP5dwwR7UqDCK3vBQ2MOOeeKBUR+Tzk2T67YpiSY8Y877PD6arPG5eMDgCRpW2o8Po7l23/COHuK8ngksawXGDw+SfM4wZVjecOzaufRsIQRucosCAw2kw5TX6ndOD2/7VDYRim4Kt1Qf3Y0VwCiHXmjI+ejh36XlAlOEARB1EWdoptzfjLw758ZY90ApHHOP2GMmc/3WqIWMWAvkauwgL9RrrjSrVgmEq0GdLAa4PJ48fh7BZg+tDsuiTcr1WXAL8RCK8u52XacLHeohKGc4f36zCw4vT6cKnfihU8P4ZEbr8T01V9j2YS+qgZDAIpne0t+YdgAmNxsO5JsRpX9oczhgVfSFp16UUBRqQOCwDTzwRe+/yMeG9FbNahGFuzzR6Ujd8dhLB6fgXKHR7m9p7AMj27Zi9XTB2q+Z1qyDTpRwDtzh+JEuRMGUcRdr4Yff+2dmRAF7VxviXMMT++EGrdPeTx3x2Esm9BX8WPL4r3a7UVZjTeibUe+v7jKBYtBxNTBlyHWpMOqaQNR4/YhKcaIzrEmRaQFp6TI3+HiDw7g9ZwsMCCsKh3tHdoybcmOIW905o9KD7NNRVMOLEEQBBGd1Es4M8ZmAsgB0AFADwApAHIBDG+6pbUdEs2GsEi/J0amK5F8sgBOSTBj4dircceQ7nhk8168cHs/TSG2dkYmSqrdSLQZoBcYnt92MKyKOvf6NJwsd2JC3ueKnaQm0LAXqXp7qtypKYZnr8vXHE6z7s5BERNFUhLMcHslLP7gAFZNG4hyhwcl1W7Fj55zbQ9N8dyzkw1/Hn2VqnExeNKmGGEQjuxBXzahr18oT+yreXw5s1wzIaXciSs7x+BMhUsR2nsKyyAwhgWj+yhRf7INZ+HYq8PO+0uT+sPjk/B6Thb0AoNX4pgbMuwI8E+BDBafWiK6uMqFQ6er0DnOhF6dzKrnJ5j1ysCl4M1RNGVuR8NksMZE/o4i+fMpBYYgCIKoi/pWq+8BkAngSwDgnB9ijCU32araGCaTDmmJVsUy4vFxzSEwSTYjuiaYUVbjwerpmTDrwzOri6tcOFZSA5NegCgw/HlrAe69Pg0vhFTMc3ccxpzrrsDm2YPh9PhgEBlsJn2YhzxU2P6/m2utKv1S45XKdtcEM/78zj6VGP/LewVhiR9yVN+zE/sqFf5z1W7VOHYg3HcOINB0KqiG7gTbRxZsLYDZIIaJzWAP+kNvfIeFY6+GxLWHEOkFBi/nSIoxKkK6xu2fHLp8+0+YP+oqmAwiXth2UDmfSTFGTFgSPj5dLwpY+P5+vHD7NehoM0HiHBLnWLH9MMocbvxpZDoYA+4bnhbWqBpakdZqclw0LgNv7zmOEVd3gdUowqzXKZXiUocHzwfWKF95eH7bQTwzJiNqqq0XMxksGivk8ncUKZs92q4yEARBENFFfUW3i3PulgfhMMZ00GisJCJjMunQxSCi1OHCqTIXTDq96g93v9R4zLupF6b8U91IGJoBvWJyf1S5vFj8wQE8d9s1GGdPRYJFj9szu8FiEFFS7cYz7/6I4ioXxtpTYDPq8MDGbzF/VDq25NdWxGUPebdECwTG8Jf3CrCnsAxlNR4lSSPUt7poXIYqIeSjgjO4/zc9sWZGJiqdXsSZ9fD4fBhnT4VX4jhZ5sSS8Rkqu4ZM8Aj24ONXOD0RfdJLxmfAK0mK2ExLtuHQmSpVmktRqQOpHSzQiyzs6sKicRl46p19mHfTlUi06pUEFLdPwvLtP2Hu9Wlwerx4YdtB/GlUOhgY3F6pzjHzack2iIKgyghfMbk/ACiDaEIbVbUSQ+Qmx405WcrViLf3HMfofl01vcNurw8fFZwJswk9eXP0VFsb6juP1gq5/B11ijWGDbiiFBiCIAjifDCuMe0u7EmMLQZQBmAqgHvhz+wu4Jw/0bTLu3gGDBjAd+/e3dLLUOF2e1Hq8ODI2WpVWkVo5nVKghnPTrwG52rcqgr2pvwipCSYsf6uQZj8jy+RZDNqpogkxRhhNYrYf7IK3RIteGjTd5h3Uy90jjUpGdhxZh2WfHgA4+ypSLQa0CXOhNIaD4orXZrrkYfyyLdfm5kFxvxiatH7+xXxu3n2YPgkjoXv78e8m3rBZtSpplfmZtuRHGPAwdNV0IuCEjc4e1gPzZH2q6dnAuB44+tf8D89k9E5zgSjTlSNbpefu2FmFg6fqUKXOCNsJj1Kqtw4VeHEtoLTGJ7eSfmcH+87iaFpycrQm827f0H/yxKRu+Mw/jz6KmW9N6Qnh41PXzHZjhc+Paj45EPXsGB0H2XSpnzfxpwszaqtXNWVJAleieNsYL2xJp0qlUY+zltzhgIAxizfqfLZ17h96Jsahw7WyFXk5qwgF1e6MGb5Ts3111XpbujrmpNorMQTBEEQLQ9jLJ9zPkDrsfpWuh8DcCeA7wHMgn8E+z8aZ3ntD4NBhzgJ6BhkcYiUyyxxrhK580elY9eREuRm2/HMu35xWlTqwOIP/JXry5OsAAC9yFDj9mHOuj2YPyodJVVuTWHu9nLMH3UVzlW7EWPS4X/f2YfiSjeWTND2RAcnhORm23Gq3Ik3dvuTSIKjDDtYDfi5pAbFVS7cvvJL9EuNV40pX/Lhftw/vCfWfH5MVa3dkl+oOaTGZhShFxl+e1UXPLDxW0UMh8UrTu6P0mqXavLjonEZ2FZwOqxqHBqbKNtU7hueprK4yOuTB+BwANt/PIVHbrwSBp2gavaUz5PFoLYayMcKFY1yVffZjw+ENbCumNxfSXMJPo7b60OnGBPemJWFs1Vu1WZm5dQBEXO6m7uC3NBc8NYwObO+KTAkzgmCIAiZ81a6GWMigFc559nNs6TGJRor3TJutxfF1W4lGjA4qQQIrywDwGePDANjgMCAXy3aEXbMHQ8PgygwVLs8sJn0KHd4YDPqIHGuWFeCj79mRiZydxzGPddfoTR2AkDeFLtmxVn2pQOATmBweCQ43F44PRI6x/l9zftPVWJbwWlMyroUDrdPJfSXTeiLhYGKuPz+wUNqlozPwFvf+H3Ml3W04HSFC4ve34/iKheWT+6PFz9Vj3a/IT0Z827qDYEBPonDZtJhQu7nYeteNW2gZkX6tZlZ8EoSjp2tHcbz6oxMDF/2Wdi5/fShX0PiHKXVHkicqxJNgps9I1W6tSq1clVX9qzXp2L+5pwh/up9efhE0Loqwi1RQW6I6GwNle76EK02GYIgCKLpqKvSLZzvxZxzH4AkxhgZFhsZg0GHrgkWdLIa4fT4sGR8BlISzACgCNDcHYeV56ckmHGkuBoVDi8YmPLc4McZA25f+QVueu6/uO3lL+DxSQCgTL8MpqjUgeJKFzblF4Fz9SAbudkydD1nq1w4UebA5H98iaGLtmPaqq/AGINRL+C+1/bgSHE1FmwtwKb8IvzvvwoAAGvvzMRnjwzDwrFXK4Jbfn9B8CeDbMzJwtIJfZGa4B8x37OTDX9970fc9vIXyhj6Oeu/wTh7quozfFRwBg6PD9UuL6av/hq/lNRofk5RYJr3n6t24w8bvwMA/O3Wa7B0Ql+cLHNontufS2rwm7/9G1Uub9hI90e37MXsYT2UynyPZCtWTRuIjTlZWDVtINbMyNSs8MpV3UiJGN07WlXfwcqpA6ATGGau2a1Muwx9TaSKcEtUkOWKcNcEC5Ji6jdARq6Qh37uxvJMSxJHcaULx0trUFzpgiQ1TXtKpEbSkmp3k7wfQRAEEd3U115yDMBOxti/AFTLd3LO/9YUi2pvmEw6dBMtiDHqsDEnCyfLnUiKMaK0xh02fXLph/5mvDdmD9acMilbTgB/GorD7cPcDV9i/qh0zWbAGrdfcHGo0z7krO+NOVlw+yQcD1hY/jaxL+7ZsCcsUnDVtIHYU1imGuoCAE6PX/RLnOOxN78Pe/+iczX+tcYYcehMFV757xHcO7wnRIFpZomHCi85wu32lV9gyOWJ6BJvxubZg1FS7VYsHykJZujF8CSYlAQzYkz+/wSmr/4aKQlmbJo1GCJDWKOc3AgJIKLYvbJzDFZNG4hEmx6nK9wqi8vKqZqbXiWGLlKMo8Uohk2QPFnuUJotLyRFo7Vkezfl5MzmrD63BptMc0JWG4Ig2jt1VroZY2sDP94KYGvg+TFB/xCNhF4volOcGZ1iTDAbRBwprsby7T9h7YxMbMzJwvxR6Yp9wf+HW1KmXMqPM0AlVGcP66GIcq3KtTzKfWNOFvQiC6u0Tx/aHYLAUFbjH6eeFGOAj2uPdhcDfzz3FJbhrW+OY9OsLPz5930w/+0fcN3Sz/DX937Eimx72PtbjSLe//4k9KKAtGQbHhvRG7LladW0geiXGq+8j+wVDz7GonEZEAVgyOWJyB7cDZNWfoHxuZ9jwdYCPHxjL9yQnozcbDvOVbuwYnL/sNcufP9HzB7WQ/kcnHN0ijOjo82gnNu1MzKx+IPahBRZ7AaTkmDG/lOVmL76azg9vN4VTrmquyW/MOz7WTl1ADpajWGVYlk8a32ndVWEm7qC3Jg0pEJeH5qz+ix/T8FE4yanOZA3O2OW78TQRdsxZvlOHDhd2WRXGQiCIKKR81W67YFJlL8AeKEZ1tPu0ekE9EyyIcGix73Xp+FYSY2mb1enMeVy1bSBqkpmcoxRVble+qFfpF/ZOQaF52rwzrdFGD/gUpQ7PPBJ0BxV/8TI3hifWztgp9KpXV31Bf3xHHF1F/xyzoGHgywY8mZg7Z2ZkDjwS0kN5v/fD0iKMWDu9WmqyL0l4zPw5Nv7UFzlwrMT+8IrcehFAR1tRuzYf0q1xld3HcWTN1+FnF/3wLRV6qE+j27xj7Zn4Dhe5kTXeBMWjr1aSUuRNzF3/upy5XPIgkgQBMVjnTfFrlxxAGpHuwdfZVgxuT9e+PQQVk4dEGbVkdejVeGUq7rPjMmAJEnYNGswOOd1ppy4vT5suGsQnn63QIl+7N7RCqNOgMD8wlKritiUFeTWQnNWnxvaSNoWuZjMdoIgiLbC+UR3LoAPAHQHENyNyOB3JFzeROtq1+j1IjrHmmHQCUiOMYYNg1mRbcfH+06GDbhJ7WBWZVPHmNRZ4HsKy7BgawFWTRuIx978HovGZWDe5r3YU1iGHQ8PCxPxKQlmpQJYVOrAi58ewqMjeuPVGZn4paS28XBFth2bd/+ivObSRIum8Pyo4Az++Lt03PGKumFzTsgwHHkc/Ky1+Xhw03d4duI1yjqu690Z2388hVlr9yvnIslqwMlKV8T0l8khednBvvJga0ewIEq0GpSM9FCRXVzlgs2oUwR8jduH5BgjnhmTgUSrIeLgn0gVTkFgyuu8EcSwli0ib4odHa0GMIGhyunFhLzPz2uZqG/qRlulOS02tMmphaw2BEEQ9c/pXsE5v7sZ1tPoRHN6SX1wubwod3vh8UpweSXoRAHrPz+KvP8cUyZGJloNiLcYUBoQpi6vD5fEm2HSCzh6tiZswM2lHcwoOFmp8jwvGZ8Bq1EXNqxGrgb3S40PG5aTm21HR5tfpLp9EsprPDhR7oTVICIlwYLsf34ZJm5em5mF/1m8XblvY05W2LTK4Pv7pcZj8firUVTqVKZHpnYww2oUwSW/b12vF3GizIGJeeGpJa/NzMKJMoeSA15c5VISQYKFqyAIYYLI65VwotyBM5UueHwSGGPoEmcCA/D0uwVK1GCowL1Q33B9nl9XogcAzcf+NXcofBLaheCrr1+YEkVahraSSEMQBHE+6kovqZfobs20dtEN+MVfhcuNk2UuODw+jM8NH0m+MScLANDBqofDI2HO+m8CUygLMc6eqtgxtuQX4omR6arqryyu05JtuHtYD5RUu5UsbdkWEilCUI66e2XaAFgMOri8EqpdXiTFGHDsbE1YLnjXBDMmrTx/NKFc6X5j1mB4JSnsON07WpEcY1KEksfjw/4zVeoBNtl2bP22CHn/Oab6nM/d3g8iQ72EqJaYA3BegXchTWP1ESTHS2swdNH2sNfufPQ6AAh7rF9qPJ7+fZ+wqYltUVw2ZJPTmN8fcX5os0MQRHuhMYbjEC2ITiegg84Ei6jDWYe2daHG7YPZIMKoFzF99W6leTK0Or1y6gCY9YIylKdLvBlPb92n2C3KnR7oBAazQcDc69NQcLJSSQ3Rujwcb9YjyWZESZUbMzbX/kF9aVI/2Iw65X1q3D4kxRjh80kqu8yW/MKwce1yUkhKghmd40yqqZOy/eT1nCyVd7nM6cXWb4uwatpAiAJTXRGQX/folr1YMLoPzHqx3tW1SHaM873+Qmwc9bn0fj5bROhj9w1PUwS3fLy25qENFsYX4hc+33cTTQKxrYh/stoQBEGQ6G5VmEw6dAQUr3GwSLUYRHSKNaLaLUVsnjQbROgE/6TKy5OseObdAsSbDYq4nj2sB+YG4gAn2lNwz/VX4PWcLDg9PjDGNEVfmcOjSkkB/KLnng17/BMvfRIsEOEO5IVXunx48dNDynTKDlYDbCZREeccwKUdzPj7bdeg8FwNpAhNiU6PD7e9/IUihtxeH/L+c0wR2RtzspSfg1/XvaM16hrZ6uMzrqspT5I4Ntw1CGcqXSipdmNLfiG6d7S2aQ9tsDBeFmF6akM/a7Q0/UWT+G8M2ns/AUEQBInuVobJpEOPDlZsmjUYXp8EQWAQBYZz1W54fP5UEK3myU05WSipcoc14iVY9GAMWDMjUzVAZlN+ETblFyn2jySbMaxxU7ZrPDbiSk3RIzCmGmG/atpAxTMu21ZSEsxYOqFv2NTF12Zm4bE3v8ffb71GU5CeKneqxFCocJU4x6ppA2ExiChzeLCt4DRGXN0FghA53aM5Ca5gmg3ieVMuIlUKAeBQcVXY92oxtI5M7oYSLIwvNK/8fERL01+0iH+C0KKtXIUhiObkvBMpiejDZNLhkngzUhIsqHR6MSH3c4x8/r/w+CQ8v+1QWHbzisn9wQJTDIP/gM9am48at4QJuV/g+mWfwe2VVLnC/VLjEWfWY+2dmbhveBre3nMcC0b3wY6Hh2H9XYPw6q6j2FNYhhq3TzOPWB68c0N6MtbMyIROZJg/Kl2VvV1U6kDo/6aLSh1weX1YPrk/Xv73YSyb0Ff1eZZN6ItlHx1Unuv2+lQZ1P1S46ETGOa//QNuffkLbMkvRPbgbpj/9g+4dvGOZs0I1pp+GJpZfMuLO2HUCXhzzhDsfPQ6vDVnaJ3JI8HZ1VrCbNbafPgkKSx3vS3F1QUL4wvNK5eJNJkyWvK1o0X8E0QolLtOEA2DKt2tGLn6+eacIXC6ffBxjuIql2IpiTfrUeP2ocrlhc2oi1CNrh0P7/T4lGp2ks2IeTf1UrK25bQSh9uH5dt/woM3pOGxEb2Rc20PeHwSnp3YFw9uqn3uC7f3Q0ebAe/e9yt4fRxTX/lKMxUlWJzLpCT4J0i++90veOTGK2HSC3g9JwseH8exs9VhkX8GnRhWCb715Vof+Dh7algsYXNUDCPZAzrFGsOE8tRXvsLGnCxVBbu40nXeKlIkYXa8zKnKXa9x+9ApVnvITGjFKsGsR6nDE9UVrOArG7KNasHoPuiRbINZX78G2UjWjWjJ124tE0SJ9gddhSGIhkGV7laOIDAkx5iQkmABl4DcbLuSt/3QG9/BoBOwaudR+Dg0q3dS0P0nyp14dddRzB+Vjr9N7Bvm0569Lh+XxJswcWAqjp2tQbXLi0qnFwJjABg2zByETx/6NV6bmYVEmwELthbAIAq4Z4Na8D66ZS9mD+uhih0MrlIuGZ+BWJOIyVmXodzhQVGpf+y5TgASbQZlUE1Kghl52XZFDAV7RoOFitzsmTfF7vd6T7EjyWZs8ophpD9MDre2UC4qdWDM8p04VlJd7ypSpKpsSbUbewrLMGttPm59+QtMX/01HEGbG7nKe7rcgR9PVSjv9cRbe7G/gRWsSJXjxn4NED5ds7jKhc5xJqTEm+s1wbKuyZTBG7i6rjw0Na1pgijRvqCrMATRMKjS3UYQBIbLOlpR6XJh/V2DUBxoqnt111FMH9odnEtYNqEvHgqqXOdNscNqFJTBL7k7DmPeTb3wyOa9EZvTvBKHTmR4YKO/Gr5kQl+U1bhRUu3GX977Ualer7tzEG7P7AZdkE88+DhpyTYsGN0HRh1DnMWINTMyUe7wIN6sxx82fYcXJ/VDSoIFNpMOJ8ucShKHbFUpd3hQVuNB1wRTmK9Qy98tf67g5lOzQV0xbGyPYqQ/TPIGSKsptajUgZ9DppDWVUXSqsrmTbHjuU8Oqp4XXCENrvLKkY/BVwVmNyD1pCFNfxfTKHixaRjnEw3R0PRHiR9EtEJXYQiiYZDobkMIAkOc2QSLzgd9QDTMH3UVzlY5MW/z90iKMeC1mVk4XeFESbUbz31yEPcP74mrusZgY04WfBywGgVsysmC28c1/6d6pLgaFoOIJeMzIDAGvcjCcsOLSh04W+XC9NVfI2+KXfM4J8ocyoCaBaP7wKQXYDGIOFXhRHGVC2aDqAjg4Oi7jwrOoOBkJRaM7oNuiRacLHeFiba0JJsq4YUxphpJL8cOvjlniLImzYmP2XZ0iTch3twwoRPpD5NJL4QJZdlyAwAWg1jvKpKWMEsw6/Hgb3spcY+hFdLgKm+8WR92VaAhFayGXG6+2EvUFyqMgzdVkdJ4ok00RIP4J4hQosWCRRCtDRLdbRC9XkRyjAkHzlSqogWfuuUqLNi6T0kOAYCCk5VYf9cgdI0zQ6fzu42KJReeemdvWGVcFobFVS6lQrpp1mBN8SKPbc/dcTjsOHKVuV9qPPYUlsFiEPHQG99hweg+SLDosXr6QJwud2Hm2shxcD2SbTDpBUxdvktTtHW0GrB2RibOVLqQaNPOGPd4JeW2ZkPiunwsGN0HneNM6NUpRnlefauOkf4wdbQa0dFqxFtzhsLh8eHwmSpFcOdNsSM51nhBglBLmNVVIQ2u8oYmfzQ0CaQhl5ub8xJ16KbqhvRkVV48iQaCqD90FYYgGgaJ7jaKTiegd+dYlbATGFSCWx4j75M4Tlc60SXODEFgSLQacP9veuK5Tw4qwrXM4VGaH4FaQZlsM4YJy5cm9cdL2w8B8EcWSpwrOdxlDg8Wf6AW7rKtwmIQEW8x4FSFU6lMRxKBZr1Yp2gz6EScrqzBw298h/mj0s8rJCMdy2IQMXPNbrw5Z0hY5OL5rBDn+8OUFGOEJHFlgucdQ7orTayy5ac+gjCSLSZShTS4Ai8nf8hRkFvyCxskRhtyuTn4NfLvYqLVAMYYJIk36h/w0E2V/N/BplmDwTkn0UAQFwhdhSGIC4dEdxtG/p+iLOxOlTtVIid0WmVuth3JMUZ0sBjQKzkGfxqZDh/nSpVaJiXBjC5xJkWkywkqNS4fjp6txvovfsb0od0Ve4NeFHD7yi/D1pdoNSjVcznFhHMOq0FUUjcMOoZ1dw7C2arawS8P/rYXEq0GlFRrT+eULRYdbQYsGZ+BVTuPhmWM52bbkWDWK6+LJBrlDYHTIzXICj2vFQ4AAB7ISURBVHG+P0zy+Xvqlj6YmPe50lC5+IP6pXHIFdxnPz6AcfZUJFoNcLi9uCToyoXWeZc3SnsKy/DqrqPYcNcgiAJTzt2FVrAacrlZfs2zHx9QNhwX6u2uL1qbqo8KzuDJmzm6Jlga5T0IgiAIoi4Y5207V3PAgAF89+7dLb2MFkeSOMocbqUpMbSBDqgdSsMYkGw1QKcTcbbKhdMVTtwdiNyT7SG9Osegg9UY9h5yxVUvCjhythpGnYBEmxGTgka5y++1enomHnnjOxRXuQKJJTp4JQ7GGOas/0aJLQyu+OZNsaNXsl+MBX+eULFWUu3GmOU7kWQzYvawHrgkzgSzQQeBAYfOVGFLfiGeGZOhCGItT/eyCX0hcQ69KKBLnAlzN+xRKv0yOx+9rlFE2/HSGgxdtD3s/vMdv7jShSfe2hsmWvOm2NG7c2ydTYyNHRPYkEZUSeI4VeFUNhwyKQnmRo0fK650YczynU36HgRBEATBGMvnnA/Qeowq3e0EQWDoYDUixqDD+rsGwSdpj1c/UebAQ298hxXZdvTuFIOONiOqXF7FHuLPe/Y3F9aFTmSIM+sxc81uTbvEy1PssBpFLJ3QFwIDzla5odcJ8Lh8SsTg/FHpYbGFs9bm4805Q1Dl9OLnkhrEW/RYPT0TZr0AvU5AR6s/Lk6ubMqvkdmYk6Xc/tNIL4oroYjD4Ir9qQonjHoBczfsUW02Fn9wICwj/EKIJEwbmgbg9vowzp6qCO7g81SXoAyuwDfWuPGGXG4WBAbOtX8XG9PbTY1fBEEQREtDorudodeL6BpnxulKZ512irvX5WNjThYAwGbSoW9qHBzuyBVMLeG2ZkYm3pwzBB6vBLNBVH426ESIAnDLi+GVx/V3DVLui5Sk4fJIOF3hVGL15KmbnWJNyvPqsovIP/94qhILthaoBGZyjAmSlcOgE1TVVznxZMHoPnh+2yHcNzwN3TtawcHr7T/2eiWcKHfgTKXaKnMxA1nk7+NiRGtjD7q40Ip3c8SPUeMXQRAE0dLQcJx2iE4noEucOWzwxqJxGcjdcRiAPIpdwtBF2zF2+S6crnChS1zkwSNawm3qK1+BgaFrggUdrEYkx5iUEeaRBsTIawFqkzSCSUkwQ2AIq4C/8OkhOL0+nChz4Jdz1eCcY82MTNXnWz65P7YVnFasI9sKTmP+qHS/373CqQxmqav6emWXGDz9+z6Y//YPGLZ0B8Yu31Wv4TGSxHHgTCUm/eNLjM/9HAu2FuCOId3x7McHLmogS6LVgOQY40WNLW/MFJGGjIduriEwciVe/h0kwU0QBEE0JyS62ymKneLuIfjskWFYMLqPKp0kJcEMX0AoyZXPs9WuiBMEL0S4SRKHT+KaQhEAlozP8Dc77jis/Cw/njfFDkFQT5zslxqPOdddgeOlDty28gtcu3gHxq7YBZdHwhuzsrB59mDMH5WOFz89hKlDLsM/ptohCgyzhvWAQRTwzLs/YmLe5ypxGGnSI+dQ5YbL50aOSIxESbVbiW+UX/folr0YZ08NG8gii0IA553WKAgMl8SZlTx0eZ0XIlojfdaGVJrrmvQYiWiZAEkQBEEQTQnZS9oxgsCQHGuC2+1FhdOrGq++fHJ/rPz3EeW5sq0jkvf3QiwCJdVuPP1uQViiyKJxGThX7cbiDw4o6SUS51g49mqkdLCASxwWg6gMNpGbJHsm23BMa4rj2t1YMLoPpq/+WvX+c69PwwMbvw3LHw+2VESyezTUfxxpUyJP0AzlQnzWwfGQDbFONKbfuaFVc4ofIwiCINo6JLoJGAw69O4Ug02zBsPrkyAKDGt2HcWm/CLlObKtI5L390KEm9vrw0cFZ1Bc6VbEdQerAUs+3I9x9lQUV7lUzY/y1Ep5guWaGZlYMz0Tpyudysj6SFMcLSGj3sfZUzEnkMQiP+fRLXuxZHwGKpxe1Lhrmyu1PMB1xRTW5WWOtClJDgj8ULQqxs9+fABP3dJHM1f6YkRrY/qdaTw0QRAEQWhDopsA4K+WXhLvtxicq3Zh2JWd8O4Pp1XJHaG2DqC2inkhwk0WZnsKyxRxfUN6sjKQJ7QCLqeGyANUzlW7cUm8WfF1lzk8MIiCptircasrrFpNh0k2I2xGnTqaMDACPvQzRNpcJJj1dVamtV6Xl22HUSegpNod9j6hFeN+qfG4Y0h3pbmzsbOsG6vSTCkhBEEQBKEN5XQTYUgSx7GSavxcUqPEBHZLtMBm0mFs0Nh1wC9sN9w1CCkJlnqLv0jWibQkG0odHkiSBB8HOOfgAO7dsAcAVMN8Ns8ejPG5nwPwC9Inb0mHw+0LiyX0+CTcExT7t2ZGJqa+8pXqM6yaNlBlTZE/V/AI+ODPplXRlnPB68qBDn6dT+J4+t0CfFRwRlNAh+ZK502xa+aqR2POdEPyugmCIAiiLUA53cQFIQgMlyVaEWPSq4QT4Bd/ckOg7Id++t0C1aCZ+hw/UlU89BjFlS5lZHxwFnWwzWNPYRn+918F+OPveuO1mVmQOIdZL4JzjtnrvlEsLGUOD3J3HEZetl01UOeyjpaI1hSt6DytddbHyyy/LlRQa0X0hVaMLzYWsDlpj/5s2mgQBEEQ54NEN6FJJOHU0WpQiVg58eTJmy9M/NVXmMnis9rlVYnO3B2HVTaU4ioXLAYRVqOIeLNf8MiCPdQf/tjvrlQEv9kgwumR6swsr4+wvRAvc30FevDGRG4eJa90w2kqYdxYw4UIgiCItg2JbuKCEARB0+ZgNogornQ1uqCRxeepCvUwnz2FZXh111FsmjVYs7EQCK8W35CejD+NTFeG/HSKMeFQcRWe/fiAZpLK0g8P1FvYXoiXuS6BHkkYShInr/RF0JTCuLGHCxEEQRBtE/J0ExdEpMmTLq/UpJW+hoomWcRKkoSzQVnZshd90j++RFGpQ2nSTLQaEG8x4JE3vkNxleuCPofXK+FMlQtenwSdKCDZZoROFx6FX5en/VBxVcTPSBaGhhNq6QEazxN/vLQGQxdtD7t/56PXoWuC5aKOTRAEQbQu6vJ0k+gmLphQ8cfBNRssG7vSV5foDH0swaxHqcOj3NYaOx/cjBnMv+ddB5HhgoTthW4KGtqMSTSMphTGTSnoow3a+BEEQdQNNVISjUqoH/t4aU2zNPlF8oFrCd7cbDue33ZQSQfJy7YjyWZUrTNS5rZZL16wWLpQi0FDmzEbSnsXS02ZH95eYhLJu04QBHFx0Bh44qJpzDHiDUFL8M5el49x9lTl9qx1+bhveJrqdVvyC5GX3fDx6cE0hmBuqvMoi6Uxy3di6KLtGLN8p2rkfXtAFsaN8V2H0l7G2EfaWJZUu1t4ZQRBEK0DqnQTF01LV/oiCd54s151+9JEi1LtTEkw444h3dEp1og35wyB0yNBZIDZULfAjVQxNhtErJo2EBaDiDKHB9sKTmPE1V3g4xzFla56VZab6jyGiqUkmxGnyp2wGkWY9bp2UfVuzKmbkY7f1qwkoTTllRiCIIj2AIlu4qJpakFzPiJZB8ocHtXt4kqXKu7w1V1H8fTvr1aJ0roumdfVAHm6wqUM2ElJMGP55P548dNDEYffaBF8HoMHBGlNrLwQgsVSv9R41ZChtmIRqI99pj0I46akKS06BEEQ7QGylxCNgixouiZYkBRjbFYBp2UdyM22Y0t+oXJ75dQBSI4xYsHWAtz68hdYsLUAdwzpjiNnq+t9yTzS5fUzVa6w++es/0Zlb4l0TEnyV8KPl9aguNKlfJ5zNR5MzPu8UewgwbaV2cN6qIYMtQWLQCT7jNcrqc5te7LTNAVNadEhCIJoD0RlpZsx9hSAmQCKA3f9kXP+XuCxxwHcCcAH4D7O+YctskgiatCqtCeY9XhmTAaevLm28ilJHBtzsuDySvBJHCv/fQTj7Cn1vmQe6fK61yfVy94SesxIlfNEm6FRc5+DbSvxZn2bswhE2gwFR0K2lYp+S9LSV7QIgiBaO1EpugM8yzlfGnwHYywdwG0ArgJwCYBPGGM9OeetVzEQjYKWdSD4tiTxsAzsReMy4PFpT6PUumQe6fK6ThTqZW8JPWZEsThzUKMK41Cx1NYsApE2Q2cqXTSwppEhiw5BEETDaW32ktEAXuecuzjnRwH8BCCzhddEtAK0BO6jW/ZCFBiWjM+o1yXzSJfXk23GetlbQo8ZSSyKgZHvwVysMJbFUpc4c5uzCERKfQm1zNS1cQm1+ZAVhSAIgmhsornSPZcxNhXAbgAPcc5LAXQF8EXQc4oC9xFEnUQSuHpRQAerAW/OGQKPV6rzknldl9frY28JPWakyrnZIDZZGkxbtAhopb7kTbHjuU8Oqp4XaeNC+dMEQRBEc9BiEykZY58A6Kzx0BPwC+uzADiABQC6cM5nMMZeAvA553xd4Bj/BPAe53xLyLFzAOQAwKWXXmr/+eefm+6DEK2CSFMDN80ajM6xphYRV3WJPQDtepjNhaI1kTTUThRJSLeniZIEQRBE0xKVEyk557+pz/MYYysBbA3cLAKQGvRwCoATGsd+GcDLgH8M/MWtlGgLRMrAbinBDZy/6tySgq+1TbDU8hrXt6JP+dMEQRBEcxCV9hLGWBfO+cnAzTEAfgj8/C8AGxhjf4O/kTINwFctsESilRGttopobExrK3aL+p5byp8mCIIgmoNobaRczBj7njG2F8B1AB4EAM75PgCbABQA+ADAPZRcQtSX+mSJU0Nd+xv3TfnTBEEQRHMQlZVuzvmUOh57BsAzzbgcop3QViq8F0t7s1tE61UQov3S2uxdBEHUj2itdBNEs9PeKryRiBTB15btFi05UZUggok0YbU9XnUjiLYGiW6CCNDeKryRILsFQbQctPkniLZLVNpLCKIloIY6P2S3IIiWgzb/BNF2oUo3QQSgCm8tZLcgiJahPdq7CKK90GLDcZqLAQMG8N27d7f0MohWAjUwEQTRklBDN0G0bqJyOA5BRCMXkptNAp0giMaG7F0E0XYh0U0QDYCqUQRBNBXRODSLIIiLhzzdBNEAKGGAIAiCIIgLgUQ3QTQAShggCIIgCOJCINFNEA2AEgaIaEKSOIorXTheWoPiShcNUiEIgohCSHQTRAOgeEEiWqAJhgRBEK0DigwkiAZC6SVENFBc6cKY5TvDhjq9NWcoNeMRBEE0MxQZSBBNACUMENEA9RcQBEG0DsheQhAE0Yqh/gKCIIjWAYlugiCIVgz1FxAEQbQOyF5CEESLQt74i4MmGBIEQbQOSHQTBNFi0GTPxoH6Cy4e2vwRBNHUkL2EIIgWgyZ7EtEAxS4SBNEckOgmCKLFoOSN1klbG8ZDmz+CIJoDspcQBNFiyMkboRnTlLwRvbRFSxBt/giCaA6o0k0QRItByRtNS1NUpNtiVZhiFwmCaA6o0k0QRItByRtNR1NVpNtiVVje/IWeK9r8EQTRmJDoJgiiRZMbKHmjaYhUkb7Y8fBt0RJEmz+CIJoDspcQRDuHkhvaJk1VkW6rliB589c1wYKkGCMJboIgGh2qdBPEBdAWs3ybqiJKtCxNVZGmqjBBEETDoEo3QdSTtloRboseXaJpK9JUFSYIgrhwqNJNEPWkrVaE26JHl6CKNEEQRLRBlW6CqCdttSLcVj26BFWkCYIgogmqdBNEPWmrFWGqiBIEQRBE00OVboKoJ225IkwVUYIgCIJoWqjSTRD1hCrCBEEQBEE0FBLdBHEB0CAXgiAIgiAaAolugiA0aYuZ5ARBEATRUpDoJggiDDmTXI5IlP3rvTrFkPAmCIIgiAZAjZQEQYQRKZO8pNrdwisjCIIgiNYJiW6CIMJoq5nkBEEQBNFSkOgmCCIMOZM8mLaQSU4QBEEQLQWJboIgIEkcxZUuHC+tQXGlCwlmfZvNJCcIgiCIloAaKQminROpaTItyUaZ5ARBEATRSFClmyDaOZGaJksdHppSSRAEQRCNBIlugmjnUNMkQRAEQTQ9JLoJop1DTZMEQRAE0fSQ6CaIdk6i1UBNkwRBEATRxFAjJUG0cwSBoVenGGqaJAiCIIgmhEQ3QRAQBIakGGNLL4MgCIIg2ixkLyEIgiAIgiCIJoZEN0EQBEEQBEE0MSS6CYIgCIIgCKKJIU83QbRCJImjpNpNjY8EQRAE0UposUo3Y2wCY2wfY0xijA0IeexxxthPjLEDjLEbg+6/KXDfT4yxx5p/1QTR8shj28cs34mhi7ZjzPKdOHC6EpLEW3ppBEEQBEFEoCXtJT8AGAvg38F3MsbSAdwG4CoANwFYzhgTGWMigJcAjACQDuD2wHMJol0RaWx7SbW7hVdGEARBEEQkWsxewjn/EQAYC7skPhrA65xzF4CjjLGfAGQGHvuJc34k8LrXA88taJ4VE0R0QGPbCYIgCKL1EY2NlF0BFAbdLgrcF+l+gmhX0Nh2giAIgmh9NKnoZox9whj7QeOf0XW9TOM+Xsf9Wu+bwxjbzRjbXVxc3JClE0TUQmPbCYIgCKL10aT2Es75bxrwsiIAqUG3UwCcCPwc6f7Q930ZwMsAMGDAAOouI9oUNLadIAiCIFof0RgZ+C8AGxhjfwNwCYA0AF/BX+lOY4x1B3Ac/mbLSS22SoJoQWhsO0EQBEG0LlpMdDPGxgB4AUASgHcZY99yzm/knO9jjG2Cv0HSC+Aezrkv8Jq5AD4EIAJ4hXO+r4WWTxAEQRAEQRD1hnHett0XAwYM4Lt3727pZRDE/2/v/mPjrus4jr9e13KzqVVKmZNQVNQ5ARMnVDOzSBCJDv4ZGEjgD0sM2UQg0cQYQf9QE030DyUhUQxThBkVCTohiiICyj8KdDJ+jLm4AWpl2WYpOGuzS9u3f9yn5VbuRnvXT7/X9vlILnf3+X57eeedz7d99e7z/R4AAFjmbO+MiIF629rx6iUAAADAskLoBgAAADIjdAMAAACZEboBAACAzAjdAAAAQGaEbgAAACAzQjcAAACQWTt+IyUAAEvO1FRoZKyiysSkyp0d6usuq1Ry0WUBaBOEbgAAWjQ1Fdp78Ii2bB/S8Oi4+nu7tG1wQOvW9BC8AUhieQkAAC0bGavMBG5JGh4d15btQxoZqxRcGYB2QegGAKBFlYnJmcA9bXh0XJWJyYIqAtBuCN0AALSo3Nmh/t6uY8b6e7tU7uwoqCIA7YbQDQBAi/q6y9o2ODATvKfXdPd1lwuuDEC74ERKAABaVCpZ69b0aMc1G7l6CYC6CN0AACyAUsla3bOq6DIAtCmWlwAAAACZEboBAACAzAjdAAAAQGaEbgAAACAzQjcAAACQGaEbAAAAyIzQDQAAAGRG6AYAAAAyI3QDAAAAmRG6AQAAgMwI3QAAAEBmhG4AAAAgM0I3AAAAkBmhGwAAAMiM0A0AAABk1ll0AQAAtJOpqdDIWEWViUmVOzvU111WqeSiywKwxBG6AQBIpqZCew8e0ZbtQxoeHVd/b5e2DQ5o3ZoegjeAlrC8BACAZGSsMhO4JWl4dFxbtg9pZKxScGUAljpCNwAASWViciZwTxseHVdlYrKgigAsF4RuAACScmeH+nu7jhnr7+1SubOjoIoALBeEbgAAkr7usrYNDswE7+k13X3d5YIrA7DUcSIlAABJqWStW9OjHdds5OolABYUoRsAgBqlkrW6Z1XRZQBYZlheAgAAAGRG6AYAAAAyI3QDAAAAmRG6AQAAgMwI3QAAAEBmhG4AAAAgM0I3AAAAkBmhGwAAAMiM0A0AAABkRugGAAAAMiN0AwAAAJkRugEAAIDMCgvdti+zvdv2lO2BmvG32R63vSvdvlez7RzbT9neZ/sm2y6megAAAGDuinyn+2lJH5f0cJ1t+yNifbpdXTN+s6Stktam26b8ZQIAAACtKSx0R8SeiNg71/1tnyLpDRHxp4gISdslXZytQAAAAGCBtOua7tNtP277j7Y/lMZOlTRcs89wGnsV21ttD9keOnz4cO5aAQAAgOPqzPnitn8v6c11Nn0pIu5u8GMHJL0lIkZsnyPpl7bPklRv/XbUe4GIuEXSLZI0MDBQdx8AAABgsWQN3RFxQRM/c1TS0fR4p+39kt6l6jvb/TW79kt6YSHqBAAAAHJqu+Ultlfb7kiP367qCZPPRsQBSUdsb0hXLRmU1OjdcgAAAKBtFHnJwEtsD0v6oKRf274vbTpX0pO2n5B0l6SrI+LFtO3Tkr4vaZ+k/ZJ+s8hlAwAAAPPm6oVAlq+BgYEYGhoqugwAAAAsc7Z3RsRAvW1tt7wEAAAAWG4I3QAAAEBmhG4AAAAgs6yXDAQAtK+pqdDIWEWViUmVOzvU111WqVTvKxEAAK0idAPACjQ1Fdp78Ii2bB/S8Oi4+nu7tG1wQOvW9BC8ASADlpcAwAo0MlaZCdySNDw6ri3bhzQyVim4MgBYngjdALACVSYmZwL3tOHRcVUmJguqCACWN0I3AKxA5c4O9fd2HTPW39ulcmdHQRUBwPJG6AaAFaivu6xtgwMzwXt6TXdfd7ngygBgeeJESgBYgUola92aHu24ZiNXLwGARUDoBoAVqlSyVvesKroMAFgRWF4CAAAAZEboBgAAADIjdAMAAACZEboBAACAzAjdAAAAQGaEbgAAACAzQjcAAACQGaEbAAAAyIzQDQAAAGRG6AYAAAAyI3QDAAAAmRG6AQAAgMwI3QAAAEBmhG4AAAAgM0I3AAAAkJkjougasrJ9WNLfM738yZL+nem1VxL62Dp6uDDo48Kgj62jhwuDPraOHs7PWyNidb0Nyz5052R7KCIGiq5jqaOPraOHC4M+Lgz62Dp6uDDoY+vo4cJheQkAAACQGaEbAAAAyIzQ3Zpbii5gmaCPraOHC4M+Lgz62Dp6uDDoY+vo4QJhTTcAAACQGe90AwAAAJkRuptke5Ptvbb32b6+6HqWCtvP237K9i7bQ2nsJNv32/5buu8tus52Y/tW24dsP10zVrdvrropzc0nbZ9dXOXtpUEfv2L7X2lO7rJ9Uc22G1If99r+WDFVtxfbp9l+yPYe27ttfyaNMx/n6Dg9ZC7Og+3X2X7U9hOpj19N46fbfiTNxZ/ZLqfxVen5vrT9bUXW3y6O08fbbD9XMx/Xp3GO6SYRuptgu0PSdyRdKOlMSVfYPrPYqpaUD0fE+ppLEF0v6YGIWCvpgfQcx7pN0qZZY436dqGktem2VdLNi1TjUnCbXt1HSboxzcn1EXGvJKVj+nJJZ6Wf+W469le6CUmfi4gzJG2QdG3qFfNx7hr1UGIuzsdRSedHxHslrZe0yfYGSd9UtY9rJY1Kuirtf5Wk0Yh4p6Qb035o3EdJ+nzNfNyVxjimm0Tobs4HJO2LiGcjoiLpDkmbC65pKdss6fb0+HZJFxdYS1uKiIclvThruFHfNkvaHlV/lnSi7VMWp9L21qCPjWyWdEdEHI2I5yTtU/XYX9Ei4kBE/CU9PiJpj6RTxXycs+P0sBHmYh1pTv03PT0h3ULS+ZLuSuOz5+L0HL1L0kdse5HKbVvH6WMjHNNNInQ351RJ/6x5Pqzj/8LEK0LS72zvtL01ja2JiANS9Y+RpDcVVt3S0qhvzM/5uy59THprzfIm+vga0sfz75P0iJiPTZnVQ4m5OC+2O2zvknRI0v2S9kt6KSIm0i61vZrpY9r+sqS+xa24Pc3uY0RMz8evp/l4o+1VaYz52CRCd3Pq/WfMZWDmZmNEnK3qx1PX2j636IKWIebn/Nws6R2qfqx6QNK30jh9PA7br5f0c0mfjYj/HG/XOmP0UXV7yFycp4iYjIj1kvpVfff/jHq7pXv62MDsPtp+j6QbJL1b0vslnSTpC2l3+tgkQndzhiWdVvO8X9ILBdWypETEC+n+kKQdqv6SPDj90VS6P1RchUtKo74xP+chIg6mPzhTkrbplY/t6WMDtk9QNSz+OCJ+kYaZj/NQr4fMxeZFxEuS/qDqGvkTbXemTbW9mulj2v5GzX252YpQ08dNaRlURMRRST8U87FlhO7mPCZpbTpDuqzqCS73FFxT27Pdbbtn+rGkj0p6WtXeXZl2u1LS3cVUuOQ06ts9kgbTGeYbJL08/bE/Xm3WWsRLVJ2TUrWPl6crHpyu6klDjy52fe0mrYH9gaQ9EfHtmk3Mxzlq1EPm4vzYXm37xPS4S9IFqq6Pf0jSpWm32XNxeo5eKunB4MtKGvXxrzX/RFvVdfG185Fjugmdr70LZouICdvXSbpPUoekWyNid8FlLQVrJO1I5610SvpJRPzW9mOS7rR9laR/SLqswBrbku2fSjpP0sm2hyV9WdI3VL9v90q6SNWTrf4n6ZOLXnCbatDH89KlsELS85I+JUkRsdv2nZKeUfVqE9dGxGQRdbeZjZI+IemptAZUkr4o5uN8NOrhFczFeTlF0u3pSi4lSXdGxK9sPyPpDttfk/S4qv/gKN3/yPY+Vd/hvryIottQoz4+aHu1qstJdkm6Ou3PMd0kvpESAAAAyIzlJQAAAEBmhG4AAAAgM0I3AAAAkBmhGwAAAMiM0A0AAABkRugGAAAAMiN0AwAAAJkRugEAAIDM/g+Nd68NG+CLhgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "residuals_test = y_test - pred\n",
    "residuals_train = y_train - gb.predict(X_train)\n",
    "\n",
    "plt.figure(figsize= (12,8))\n",
    "sns.scatterplot(x= pred, y= residuals_test)\n",
    "plt.axhline(0, c= 'red', linestyle= '--')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The residual plot shows that the residuals/errors are randomly distributed around the zero line. Based on the two plots, the residual plot and the predicted vs actual values, the model is not very accurate in predicting the freight values.<br>\n",
    "\n",
    "To test the goodness of the fit, the following methods are applied:- \n",
    "* Test normality of residuals\n",
    "* Test significance of mean between training residuals and test residuals\n",
    "* Test significance of variance between training residuals and test residuals\n",
    "\n",
    "To test the normality of residuals, One sample Student's t-test is performed.<br>\n",
    "\n",
    "Null hypothesis - The mean of the sample is equal to the population mean<br>\n",
    "Alternative hypothesis - The mean of the sample is significant from the population mean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 266,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Ttest_1sampResult(statistic=-1.7775710450649203, pvalue=0.07548625930841611)"
      ]
     },
     "execution_count": 266,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ttest_1samp(residuals_test,0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The p-value of 0.075, greater than 0.05 indicates that the null hypothesis cannot be rejected. Hence the mean of the residuals is equal to population mean, here zero (0).<br>\n",
    "\n",
    "The next step is to test the mean of residuals of training and test predictions. Welch's t-test is used for this. <br>\n",
    "\n",
    "Null hypothesis - The mean of the two samples are identical<br>\n",
    "Alternate hypothesis - The means of the two samples differ<br>\n",
    "\n",
    "The test assumes that they have equal variances."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 268,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Ttest_indResult(statistic=-1.6677505413544513, pvalue=0.09536833032437417)"
      ]
     },
     "execution_count": 268,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ttest_ind(residuals_test,residuals_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The p-value of 0.09 is greater than 0.05 and hence the Null hypothesis cannot be rejected. The mean of the test residuals is equal to the mean of the training residuals.<br>\n",
    "\n",
    "Next is to test the variance between the residuals of test and training phazes. The variance of test residuals closer to training residuals is a favorable condition. The F-test is carried out to test the significance in the means of the two samples by comparing the variance. F-test works on samples that are normally distributed, hence the first step is to test normality using a Shapiro-Wilks test.\n",
    "If the p-value is greater than 0.05 then we reject the normality of the sample."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 276,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "training residuals:  (0.699959397315979, 0.0)\n",
      "test residuals:  (0.6568727493286133, 0.0)\n"
     ]
    }
   ],
   "source": [
    "print(\"training residuals: \", shapiro(residuals_train))\n",
    "print(\"test residuals: \", shapiro(residuals_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The p-value lesser than 0.05 indicates that both the distributions follow normality. Hence the F-test can now be conducted on the samples. One-way ANOVA test is used for this.<br>\n",
    "\n",
    "Null hypothesis - The two samples have identical means<br>\n",
    "Alternate hypothesis - The two samples significantly differ in means\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 280,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "F_onewayResult(statistic=2.781391868188062, pvalue=0.09536833033141338)"
      ]
     },
     "execution_count": 280,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f_oneway(residuals_test, residuals_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The F-test by ANOVA, gives a p-value of 0.09, which is greater than 0.05 and hence we fail to reject the Null hypothesis. The means of the two samples explained by their variance is identical."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The tests performed along with the R-squared value of 73% achieved by the tuned model on train set show that the model trained is a good fit on the data and performs fairly well. This model is hence chosen to be the final model."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.2"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "state": {
     "003cb66582ec44499042c9d14d7afddd": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "DropdownModel",
      "state": {
       "_options_labels": [
        "product_weight_g",
        "product_length_cmproduct_height_cm",
        "product_width_cm",
        "distance",
        "volume"
       ],
       "description": "column",
       "index": 0,
       "layout": "IPY_MODEL_2055109f8bbc414f9acc224e8b93762f",
       "style": "IPY_MODEL_67f8311ce7b94b77b334607488b90a26"
      }
     },
     "058e9cd500174ca4819c007ce888cb35": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "VBoxModel",
      "state": {
       "_dom_classes": [
        "widget-interact"
       ],
       "children": [
        "IPY_MODEL_8019edccd6ae4f3fa4dbec860463097d",
        "IPY_MODEL_ce54f8a888de46a5941eff5a5e6598cb"
       ],
       "layout": "IPY_MODEL_4870b50b082343c1b10ae66f7205b797"
      }
     },
     "06c6f89a52824c9abe54436c47248280": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "DescriptionStyleModel",
      "state": {
       "description_width": ""
      }
     },
     "2055109f8bbc414f9acc224e8b93762f": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {}
     },
     "23ebe228ab0b4eaf9dbb8ba67a6b01fa": {
      "model_module": "@jupyter-widgets/output",
      "model_module_version": "1.0.0",
      "model_name": "OutputModel",
      "state": {
       "layout": "IPY_MODEL_47ce7327d3804ac8a069e1a3ed2ebc83",
       "outputs": [
        {
         "data": {
          "image/png": "iVBORw0KGgoAAAANSUhEUgAAAskAAAHzCAYAAADSPIC7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOzdeXSk913n+8+3dlVpbUm9uBe33N5iEnASx84ChJAbiDMEEyZm7DDgDOGaABngMjlDwlyyDZkhcJnMPSR3IEwSe8J4nEwCjIfYMRCTQBLbuB3bJF7Ti5d2b9paUpVU++/+8TylVleXpKc2VXXV+3WOTque+j1VvypVn/PRV9/n9zPnnAAAAACcFer0BAAAAIBuQ0gGAAAAqhCSAQAAgCqEZAAAAKAKIRkAAACoQkgGAAAAqhCSgR5nZh8ys5l17rvNzA6uuf1OM3NmNhjwsS/3H3+0VfPtBWZ2g5k9aWZ5M3u20/OpZmZfNLOvteFxbzWzn6rznLo+c5s81rqf9U3OO+f/wTpjYv7jX934DAFcSAjJANb6sqTXSFoOOP5ySR+UREj2mVlY0n+T9JikH5X0ts7OaEvdKqmukNxi/1XSj7fpsWPyPuuEZKBPRDo9AQDdwzk3LWm60/PYiJmZpLhzLtvpuaxjl6RhSXc4577R6gf3Q3jYOZdv9WNf6JxzxyQd6/Q8APQGKskAVtX607eZvd/MDplZ1sxOmdlXzGynmf2IpP/tDzvqn/fsmvOuNrOvmtmymc2b2X83sx1Vz7fPzO4xsxUzO+o//zmtAJU/oZvZD5rZQ5Kykm40s5SZfcLMnvaf46iZfdLMhquew5nZ/2Vmf2hms/5jvde/7xYzO2JmZ8zsM2aWCPAe/YyZfcfMcmb2gpl91MwilfdP0gv+0P/lP/eH1nmcH/Hv/zEz+yszy5jZ82b27qpxt5nZQTP7KTN73H/919XxHu81s7v99/hZM/vFGnM5r93AzPb78/uJNcfC/ufhGf/1HzOz2/z7vibplZJu8c9z/vsR1JSZ/Y3/PjxlZj9dY543+O9F1sxOmtnvm1l0zf3ntVuY2feb2bf8cx43s7f4j3Fbjcd/k5n9kz+Hb5jZ9625e8n/97NrXt/+IC/MzH7YzP7OzNJmtmBmXzOzl/v3Vf7PvcI/vmxmj/q3U2b2Wf+cI2Z2c5DnA9AahGSgT5hZpPpLkm1yzs9L+m1J/0nen7F/WdIhSSlJ35b0Xn/oT8tr03ibf96kpK9JSkp6h6R/Len1kv7GzGL+GJN0l6SXSPoFSb8p6dfkB8AqSUm3y/tz+psl/aN/LCzp30m6XtLvyGtv+J81zv83kgYl3SzpDkl/YGa/L+md/nP+tqSflfQbm7wfPybp8/5rv0HSH/nvwSf8IV/23wv5x1/jz3kjn5b0T/5590j6L2uDqW+/pN+X9B8lvUXeLyVB3+P/Jemlkt4l7z3+dX9ejfgTSR+W9AVJPyHvfU359/2KpKck3e0//mvkvR9B3SHv8/A2Sd+TdKeZ7ancaWY/I+nP5f3sf9Kfx63y3pOazCwp6V5JA/J+9r8r6eOS9tUYvk/SH0j6qD92u6Qv+O+h5H225D9G5fWd2OxFmffL5FclFSTdIulfSPoHSburht4u6X9I+ufy/l9+Ud5n47ikt0t6UNJ/W/ueAGgz5xxffPHVw1+SPiTJbfB1cM3Yd/rHBv3bn5D0pQ0e+yf88furjv+epDOShtccu9Yfe7N/+5/5t69dM2a3vDDxtRrzv2GT1xmR9Dp/7L41x52kv1tzOyQv3MxXze8Lkh7c5DkeWPtY/rF/K6kkaY9/e7//nD+xyWP9iD/uU1XH/0bSA2tu3+aPu7qB9/gt/u3r1oy5WFKx6j2+be3noNbrkHSlf/vXNnhNByXdVufns/KZ+4U1x8b9Ob7bv22SnpP02apzf0HSiqTxNZ+VmTX3/6qkvKTdNd6j29Ycu81/vsvWHPspf9yV/u1B//Y763x99/vvi23y+m9Zc6zyc/vMmmMj8v5v/HI9z88XX3w1/kUlGegPC5JeVePrrzY571FJbzGzD5vZteb1wwZxraS/ds4tVg445/5R0rOSftA/9CpJJ/3jlTEvSnq4xuM5eVXWc5jZz5nZI2aWlhcgKj3Al1cN/eqa5yhLOirp4bXzk1chr67urX2usKRX6PxK9eflBe9Gq7N/UXX7zyW9suq9ftE592jVuCDv8bWSTjnnHlwz5jnVfo838wb/39saODeIv65845yblXRaUqVqerm8Su8Xqv4Scp+khLxKeS2vkvdzfnHNY/+jpFM1xj7rnPvemttP+P82XLk1s5S8v4zc7pxzmwz/6prvD/n/3lc54JxbkHe9wLqfUQCtRUgG+kPROXew+kvS7CbnfUZeK8LPyPtz7ykz+/cBwvIu1Q4ipyRt87/fqdoXCdY6Nu+qLlQzs7fJW0Xifkk3Snq1zq4kUd1bfKbqdn6dYxv1JE9Iiur811W5vU2NOV3jdsR/vurnWCvoe1z9+LWeM4hxSZmqXyxaaaOfR+W9uFveL0OVr6P+8b3rPGY9n7Fazy9t/JnYzJi8KvimbRlVz5+vcaxyvJn5AKgDq1sAWJdfdf24pI+b2V55fbsflfSipD/e4NQT8no6q+3Q2SrmSUmTNcZMyrs47Zyp1Bh3o7z2iF+pHDCz128wp2bNyAtm1a+rcqHcXIOPW/142+X96X/tBWi1Xn/Q97jWmO3y2hQqsvKWOFurOvTPSkqZ2XAbg/J6Ku/trZIeqXH/0RrHJO/1X1HjeK3PXTvMSyrL+4UGwAWGSjKAQJxzLzjnfk/en4Kv8g+vV217UNKPm9lQ5YCZvUpen2ulJeIhSTvN7No1Y3bLWyEhiAFJuapjPxvw3Lo550rywueNVXf9jLwgdH+DD129jvLb5LUIlDY5L+h7vMPMrlszZp+8tpG1jknab+eu7vGmqjGVP/3//AZzalel82l5v5jtr/UXEb89o5aHJF3jf64kSf7nbcc64zdSd2XZOZeR93P6+TUXAAK4QFBJBrAuM/sTeVW8B+T1Nb9B0mWSfssf8rT/7y+Z2Z2Slp1z35G3GsYvS7rXzD4m76Kn35P0HUlf8s+5W96GG18ws/fLq2x+UF67QDnA9P5G0ifN7N/JCyJvkfTGxl9tIB+U95o+K+lOSS+T9O8l/anz1uhtxPVm9lFJX5e3wsWb5K2csZl63uP/aWa/Ja9i/BGd327xl/7x/+ovjfZySf9q7QDn3NNm9ilJf2hm2yX9vbxNZN7unLvJH/aUvOD+4/Iqz0c3CLCBOefKZvZvJH3OvCX+7pEXWi+Rd4Hd251ztTbA+ayk/1vSX5nZh+X9YvVhee0WQT5ja+eQN7Ojkn7GzL4r7738p+o2oBreJ+lvJd3jv38Zef3rB51zm10TAKCDqCQD2Mj9kn5YXti4W16V8/90zv2ltHoR2Hvlhbtvyl832XmbkrxBXpD4H5I+KW/ZqzdVQoV/IdMN8oLVZyX9v5L+i7wLpoL8Of9PJP2hvCXN/lzeqg3vaPYFb8Q599eSbpJ0jbzX+hv+HN7TxMP+orzK7l/KWy3kV51zdwWYS9D3+CflvaefkfSf5a1Ycn/VY31X3koRr5G3DNvr/dvVfkVeyPyX8j4P/1nntm38rqQn5a0U8pCkt272OoJyzn1e3uflankXT/65P59v62yVt/qcZXlLBq7Iu8DyQ/JWIzmjYJ+xau+W1x/9t/Je30UB5v338n7xSUr6M38erxebngBdzza/4BYAtoaZjUg6IukTzrkPdno+7eSvn/t3kl7mh1RsATObkvSMpFudc5/t9HwAdC/aLQB0jHm7y5XlbR4xKW+zi7i8qifQNL+V57i8dZb3SXq/vHaLL210HgAQkgF0Uk5ef/M+eSs4/KOk/8Nv48AFzr9YbaPlAksB1g9ulpPXS36RvM/bP0h6bytW6DCzkDZoW3TOFZt9DgCdQ7sFAKAtzOyd8vrN1/OvnHO3bc1sWs+/yPGWDYZMOeee3ZrZAGg1QjIAoC3MbFzS1AZDWrL6RaeY2X6du+lLtSCrXwDoUoRkAAAAoErX9SRPTEy4/fv3d3oaAAAA6HEPP/zwjHOu5i6cXReS9+/fr4MHD3Z6GgAAAOhxZrbuheJsJgIAAABUISQDAAAAVQjJAAAAQBVCMgAAAFCFkAwAAABUISQDAAAAVQjJAAAAQBVCMgAAAFCFkAwAAABUISQDAAAAVQjJAAAAQBVCMgAAAFCFkAwAAABUISQDAAAAVQjJAAAAQBVCMgAAAFCFkAwAAABUISRDzrlOTwEAAKCrEJL73H+850nd9KkHVC4TlAEAACoIyX3usRfO6MGjc/qLR17s9FQAAAC6BiG5z81nCpKkP7j3aa3kSx2eDQAAQHcgJPe5ueW8rtw5pJOLWf3pPxzp9HQAAAC6AiG5jznnNJ/J60ev3K7rX7pTf/z1wzq9mO30tAAAADqOkNzHFrNFFctO21Ixve/6K1UolfWHf/1Mp6cFAADQcYTkPjafyUuStqViung8pVtes19fePgFPXF8scMzAwAA6CxCch+b9UPyWComSfrXP3qZRgai+g93P8nayQAAoK8RkvvYaiU56YXkkWRUv/7Gy/SNQzP62tPTnZwaAABARxGS+9jc8tl2i4p/+eqLNTEY112PHe/UtAAAADouUEg2szeb2dNmdsjM3lfj/riZfd6//0Ez27/mvu83s/vN7HEz+46ZJVo3fTRjLnN+SI6GQ5qaSOrEwkqnpgUAANBxm4ZkMwtL+qSk6yVdJelmM7uqati7JM075y6V9HFJH/PPjUj6M0nvds59n6QfkVRo2ezRlPlMXrFISMlY+Jzj24cTOr2Y69CsAAAAOi9IJflaSYecc0ecc3lJd0q6oWrMDZJu97//oqQ3mplJ+jFJ/+Sce0ySnHOzzjm2desSc5m8xlMxeT+qs3YOJ3RyMcvFewAAoG8FCcm7Jb2w5vYx/1jNMc65oqQFSeOSLpfkzOxeM/u2mf3b5qeMVpnL5DWWjJ13fOdwQsv5kpZyxQ7MCgAAoPMiAcZYjWPVJcb1xkQk/aCkV0lalvRVM3vYOffVc042u1XSrZK0b9++AFNCK8wt58/pR67YPhyXJJ1ayGo4Ed3qaQEAAHRckEryMUl719zeI6l66YPVMX4f8oikOf/4151zM865ZUl3S3pF9RM45z7lnLvGOXfN5ORk/a8CDZnP1A7JO4e9aytP0ZcMAAD6VJCQ/JCky8xsysxikm6SdFfVmLsk3eJ//3ZJ9zmvofVeSd9vZkk/PL9e0hOtmTqaNbteSB7xQvLJxexWTwkAAKArbNpu4Zwrmtl75AXesKTPOOceN7OPSDronLtL0qclfc7MDsmrIN/knztvZv9JXtB2ku52zn25Ta8FdSiUylrKFmv2JO9YrSQTkgEAQH8K0pMs59zd8lol1h77wJrvs5JuXOfcP5O3DBy6yHxlI5HB80NyIhrWyECUkAwAAPoWO+71qbmqLamr7RxO6OQCIRkAAPQnQnKfqoTksVTt1St2jCSoJAMAgL5FSO5T8xlv48PxVLzm/TuG4ly4BwAA+hYhuU/NZbzl3darJO8cSWh6KadSmV33AABA/yEk96k5v5Jca3ULyVvhouykmTRrJQMAgP5DSO5T88t5DSciioZrfwQqy8Bx8R4AAOhHhOQ+td5GIhU7WSsZAAD0MUJyn5rP5DW2QUjeMeJd0EdIBgAA/YiQ3KfmMnmNbxCSx1NxhUPGChcAAKAvEZL71Fwmv+5Fe5IUDpm2D8V1apEL9wAAQP8hJPch55zmljfuSZa8i/dotwAAAP2IkNyHlvMl5YvlACE5zuoWAACgLxGS+9DZLak3Dsk7hxP0JAMAgL5ESO5DlZC80YV7krRjJKGlbFHL+eJWTAsAAKBrEJL7UNBK8o6hylrJXLwHAAD6CyG5D1VC8rYNVreQpJ0j7LoHAAD6EyG5D80v+yF5cPPVLSTp9BIhGQAA9BdCch+azeQVCZmG4pENx+0Y9nbdo5IMAAD6DSG5D1W2pDazDccNJaJKxcKscAEAAPrOxqVEXJDuePD5De//p2MLCputjnvHdfvWHbtjhA1FAABA/6GS3IcyuaKSsXCgsTuGEqxuAQAA+g4huQ9l8iUlN+lHrtg5kqAnGQAA9B1Cch9azheVClpJHk7o9FJW5bJr86wAAAC6ByG5z5Sd00q+pFTQSvJwXIWSW102DgAAoB8QkvvMcr4kJwXvSfbXSmaFCwAA0E8IyX1mOVeUpMCV5B0jla2pCckAAKB/EJL7TCZfkiSlYkHbLSpbU7PCBQAA6B+E5D6T8SvJQdstJofiMqOSDAAA+gshuc8sVyrJAdstouGQxlNxQjIAAOgr7LjXQpvtdLfWRrvctVMmX18lWZJ2jsS5cA8AAPQVKsl9ZjlXVCwSUjQc/EfPrnsAAKDfEJL7TCZfCryRSMWOkQTtFgAAoK8QkvtMJlcM3I9csXM4oblMXrliqU2zAgAA6C6E5D6znC/V1Y8sSTuG45Kk07RcAACAPsGFe30mky9q+1D8nGObXXD4zKklSdKfPfCc3v+Wl7RtbgAAAN2CSnKfaaTdYjgRlSQtrBTaMSUAAICuQ0juI/liWYWSq7vdYjDhheq0vxEJAABAryMk95Flf43keivJA9Gwfz4X7gEAgP5ASO4jmcpue3VWksMhUyIa0gohGQAA9AlCch9ZzlV226v/es2BaFgrBUIyAADoD4TkPrJaSa6z3ULygnWlXQMAAKDXEZL7SMavJNfbbiFJyViYdgsAANA3CMl9pNIukWggJA/Ewly4BwAA+gYhuY+s5EtKREMKmdV9Lj3JAACgnxCS+8hKobS6nFu9Ku0W5bJr8awAAAC6DyG5j2SbCMkDsYicpCU2FAEAAH2AkNxHvHaLBivJ/nlnlvOtnBIAAEBXIiT3kZVCSQMNXLQnafW8M8uFVk4JAACgKxGS+0gz7RbJSkheISQDAIDeR0juIyuFxtstBmi3AAAAfYSQ3CeKpbIKJdd0u8UClWQAANAHCMl9orLGcePtFt5W1vQkAwCAfkBI7hOru+01GJLDIVM8EiIkAwCAvkBI7hPZQllS45VkyWu5OLNCTzIAAOh9hOQ+sZKvtFs0/iNPRsNaoJIMAAD6ACG5T6y2WzR44Z5UqSQTkgEAQO8jJPeJbJMX7kne1tTzLAEHAAD6ACG5TzS7uoVEuwUAAOgfgUKymb3ZzJ42s0Nm9r4a98fN7PP+/Q+a2X7/+H4zWzGzR/2vP27t9BFUNl9SNGyKhBv/vajSbuGca+HMAAAAuk9kswFmFpb0SUlvknRM0kNmdpdz7ok1w94lad45d6mZ3STpY5L+hX/fYefc1S2eN+q00sSW1BXJWFilslM6V9RQItqimQEAAHSfIGXFayUdcs4dcc7lJd0p6YaqMTdIut3//ouS3mhm1rppolnNbEldkYxVtqam5QIAAPS2ICF5t6QX1tw+5h+rOcY5V5S0IGncv2/KzB4xs6+b2Q81OV80qBWV5IGo94cHtqYGAAC9btN2C0m1KsLVTanrjTkhaZ9zbtbMXinpL83s+5xzi+ecbHarpFslad++fQGmhHpl8yUNDzTXIjFAJRkAAPSJIJXkY5L2rrm9R9Lx9caYWUTSiKQ551zOOTcrSc65hyUdlnR59RM45z7lnLvGOXfN5ORk/a8Cm2ppuwW77gEAgB4XJCQ/JOkyM5sys5ikmyTdVTXmLkm3+N+/XdJ9zjlnZpP+hX8ys0skXSbpSGumjnq0pN2CSjIAAOgTm7ZbOOeKZvYeSfdKCkv6jHPucTP7iKSDzrm7JH1a0ufM7JCkOXlBWpJ+WNJHzKwoqSTp3c65uXa8EKyv7JxyhfJqyG1UJWTTkwwAAHpdkJ5kOefulnR31bEPrPk+K+nGGud9SdKXmpwjmpQrlOWkptstouGQBqJhzWdotwAAAL2NHff6QCt226sYTUZ1hkoyAADocYTkPnA2JDf/4x4ZiNKTDAAAeh4huQ9k/ZCcaLInWZLGkjEtsLoFAADocYTkPrCSb3G7BZVkAADQ4wjJfYCeZAAAgPoQkvtAtoUheWQgpoXlgpyr3nQRAACgdxCS+8BKvqSQSbFI8z/u0WRU+VJ5tToNAADQiwjJfaCyJbWZNf1YowNRSey6BwAAehshuQ+0YkvqitEkIRkAAPQ+QnIfyBZKTW9JXTEyEJMknVlmGTgAANC7CMl9YCVfanpL6orVSjIrXAAAgB5GSO4DK4Uy7RYAAAB1iHR6Ami/yoV7rTCW9NstAu66d8eDzwd+7Hdct6+hOQEAALQaleQe55zzepJbFJIT0bDikZAWqCQDAIAeRkjucYWSU6nsWnbhnsTW1AAAoPcRkntcZdOPRLR1P+rRgVjgdgsAAIALESG5x7VyS+qKESrJAACgxxGSe9xK3g/JrWy3GIhqgSXgAABADyMk97iVNlSSR5NRzbOZCAAA6GGE5B7XjnaL0WSMdgsAANDTCMk9rh2V5JGBqHLF8moABwAA6DWE5B5XCcnxFobk1Q1FqCYDAIAeRUjucdl8SfFISOGQtewxV7emZhk4AADQowjJPW6lhbvtVYwO+CGZSjIAAOhRhOQet1IoK9HikDySJCQDAIDeRkjucSv5UkvXSJa81S0kaYF2CwAA0KMIyT0uWyi1vJJMuwUAAOh1hOQe146e5GQsrGjYdIZd9wAAQI8iJPc4LyS39sdsZhoZiOkMu+4BAIAeRUjuYaWyU75YVqLFPcmStwwc7RYAAKBXEZJ7WDu2pK4YHSAkAwCA3kVI7mHt2JK6YjQZoycZAAD0LEJyD2trJTkZ1QI9yQAAoEcRknvYSt4Pye3oSR6IUkkGAAA9i5DcwyrtFq1eJ1nyKsnL+ZJyxVLLHxsAAKDTCMk9rJ09ySOru+5RTQYAAL2HkNzDsm1ut5DYdQ8AAPQmQnIPWymUFA6ZIiFr+WPvHElIkl48s9LyxwYAAOg0QnIPWymUNRANy6z1IfnA5KAk6fDpdMsfGwAAoNMIyT3M25K69a0WkrQtFdNYMqrD05m2PD4AAEAnEZJ7WLZQUiLavh/xgclBHZ6mkgwAAHoPIbmHreRLbblor+LA5KCOEJIBAEAPIiT3sHa2W0jSge0pzaTzOsPOewAAoMcQknuY127RvpB86Xb/4j2qyQAAoMcQknuUc07ZQvvbLSTp8Gku3gMAAL2FkNyjcsWyyq49u+1V7BlLKhYOUUkGAAA9h5Dco7Jt3JK6IhwyTU2kCMkAAKDnEJJ71IofktvZkyx5F+9ttFayc07OubbOAQAAoNUIyT1qOe9XktvYkyxJl04O6rnZjHLFUs37v/TtY/rYV57SwWfnVCYsAwCACwQhuUdNL+UkSeOpWFuf58D2QZWd9Nzs8nn3lcpOT5xY1EqhpD9/5EV94r5DeubUUlvnAwAA0AqE5B51YiGrgWhYIwPRtj7P2RUuzu9Lfvz4grKFst728j26+dp9ypfKuu1bz+oz3zzK2soAAKCrRTo9AbTHiYUV7RpNyMza+jxTEylJtddKvv/wrCTpksmUhhNRvWTnkB48Oqd7Hz+prz8zrRuu3t3WuQEAADSKSnIPKpWdTi5kddHIQNufKxWP6KKRRM2L9+4/MqvJwbiGE141OxIO6XWXTmh8MKalbLHtcwMAAGgUIbkHzaZzKpaddo0ktuT5Dmwf1KGqdotCqayHjs7pksnUeeNT8YjSOUIyAADoXoTkHnRiIStJ2rUFlWTJ60s+PJ0+Z6m377y4oEy+pEv8nuW1BuMRZQjJAACgixGSe9CJhRWFQ6bJofiWPN+B7YNazpd0cjG7eqzSj1zpWV4rFY8okyckAwCA7kVI7kEnFrLaMRRXONTei/YqDvgtFYdPn+1LfuDIrK7YMaTB+PnXhg7GI8oWyiqWylsyPwAAgHoRknuMc07Hz6xsWauF5G0oIp1d4SJfLOvgs/N6zYHxmuNTMS84Z/K1NyABAADotEAh2czebGZPm9khM3tfjfvjZvZ5//4HzWx/1f37zCxtZu9tzbSxnumlnDL5knaNbs1Fe5I0ORTXUCKyGpIfO3ZGK4WSXn1J7ZA8GPd2AeTiPQAA0K02DclmFpb0SUnXS7pK0s1mdlXVsHdJmnfOXSrp45I+VnX/xyXd0/x0sZknTixK2rqL9iTJzHRg8uwKF/cfnpWZdN3UtprjU34LBhfvAQCAbhWkknytpEPOuSPOubykOyXdUDXmBkm3+99/UdIbzd/Fwsx+StIRSY+3ZsrYyNmQvHWVZOnsCheSF5Kv3DmssXW2xK70KVNJBgAA3SpISN4t6YU1t4/5x2qOcc4VJS1IGjezlKTfkvTh5qeKIJ44vqixZFSJaHhLn/fA9pROLeY0k87p4efn9Zp1Wi0kKskAAKD7BQnJtZZIcAHHfFjSx51z5+9ZvPZks1vN7KCZHZyeng4wJazniROLW9pqUXHAv3jvSw8fU75YXveiPUmKR0KKhIyQDAAAulaQkHxM0t41t/dIOr7eGDOLSBqRNCfpOkm/b2bPSvoNSb9tZu+pfgLn3Kecc9c4566ZnJys+0XAs5wv6uhMZksv2qu4dLsXku/4x+cVMunadfqRJa+H2dt1j9UtAABAdzp/EdvzPSTpMjObkvSipJskvaNqzF2SbpF0v6S3S7rPeduv/VBlgJl9SFLaOfeJFswbNTx1cknOSRd1oJK8b1tSkZDpudllvWz3iEYGohuOT8XDVJIBAEDX2rSS7PcYv0fSvZKelPQF59zjZvYRM/tJf9in5fUgH5L0m5LOWyYO7ffE8c5ctCdJ0XBIF48nJWnDVouKwXiEC/cAAEDXClJJlnPubkl3Vx37wJrvs5Ju3OQxPtTA/FCHJ04samQgumkVt128FS4yG160V5GKRXR6MbcFswIAAKgfO+71kCeOL+qqXcPyV9/bci/ZNax4JKRr9o9tOrZSSfa6cgAAALoLIblHlMpOT51c1FUXDXdsDr/0+kv05V/7IQ0lNq9kp+IRFctO+WJ5C2YGAABQH0Jyjzg6k1G2UNZVuzoXkpOxyOoqF5upbCiSybPCBQAA6D6E5B5R2WmvkzltbQoAACAASURBVJXkeqTi3mYnXLwHAAC6ESG5RzxxfFGxcGh1U49ux657AACgmxGSe8QTJxZ16fZBxSIXxo+00m5BJRkAAHSjQEvAofPuePD5De//9nPzunzH0KbjugWVZAAA0M0ujLIjNnTfU6eVzhV18bZkp6cSWDQcUjwSopIMAAC6EiG5ixw/s6K/fORFlcvB1w6+76lT+tsnT+kV+0b1ygDrE3eTVDxCJRkAAHQl2i26xEPPzundn3tYs5m8Hn3hjD741qs23RTkq0+d0lefPK1X7BvVT79ij0Id2kSkUYPxiDI5loADAADdh0pyF/j8Q8/rHX/6gIYHorrxlXt027ee1R/+9TMbnnOhB2RJSsXCtFsAAICuRCW5g4qlsj5695P67Def1Q9dNqFP3PwKDQ9EFA6ZPvF3h5SMh/UrP3LpOecs54v62tPT+sahmQs6IEteu8Wx+ZVOTwMAAOA8hOQOyRVL+sXbD+ofvjejX3jdlH77LVcqEvYK+x9928u0Uijp97/ytFKxiG557X7NLOX0zcMz+vbz8yqUnF61f0w3XL37gg3Ikt9ukS+q7NwF/ToAAEDvISR3yH1PntY/fG9GH3rrVXrn66bOuS8cMv0/N/6AVvIlffCux/Xl75zQQ0fnFAqZrt47qtddOqGdw4kOzbx1UvGIyk7K5ktKxvkoAgCA7kEy6ZAHj85pIBrWz7764pr3R8Mh/dE7Xq53f+5hPXZsQW+4cruum9qmoUR0i2faPms3FCEkAwCAbkIy6ZAHjszqmv1jiobXv3YyHgnrM+98lZyT7nzohS2c3dZY3VAkzwoXAACgu7C6RQcs54p66uSSrpvatulYM1Mo1Jv9uql4WBJbUwMAgO5DSO6Ao7MZSdKrLxnv8Ew6a5CtqQEAQJciJHfA0ZmMEtGQvn/PaKen0lHJ2NmeZAAAgG5CSO6AozMZvfLiMcUi/f32h0OmZCxMJRkAAHSd/k5pHbCcL+rkQlbXTfV3q0VFKh6hkgwAALoOIXmLPTuzLCf6kSsG4xEqyQAAoOsQkrfY0Zm0IiHTD+wd6fRUukIqHlEmxxJwAACguxCSt9jR2Yz2bUsqHgl3eipdIRUL024BAAC6DiF5C63kSzpxJqupiVSnp9I1BuMRrRRKKpVdp6cCAACwipC8hZ6bzchJmpokJFec3XWPajIAAOgehOQtdGQmo0jItHcs2empdA02FAEAAN0o0ukJ9JOjMxnt3ZZUNMzvJhWVSnIjfcl3PPh84LHvuG5f3Y8PAAD6F2lti2QLJR0/s0I/cpWzlWRWuAAAAN2DkLxFVvuRCcnnoN0CAAB0I0LyFjkyk1GYfuTzJKIhhayxdgsAAIB2ISRvkaMzGe0ZG1Aswlu+lpn5G4oQkgEAQPfgwr0tUOlH/uHLJ1eP1XPRWa8bjEeoJAMAgK5CWXMLPDubUdlJByYHOz2VrkQlGQAAdBtC8hY4Mu2tj7xvG/3ItVBJBgAA3YaQvAWOTKdZH3kDqVhYmTxLwAEAgO5Bamuz5XxRJxayOsBW1OtKxSPKF8taISgDAIAuQUhus6Mz3vrIl0zQj7yeylrJs5lch2cCAADgISS32eHpjKJh055tA52eSteqbE09m853eCYAAAAeQnKbHZlOa/94SpEQb/V6qCQDAIBuQ3Jro6VsQaeXcrqEpd82VKkkz1BJBgAAXYKQ3EZHZjKSxEV7m0jFwpKk+QwhGQAAdAd23GujI9MZxSMh7RrpnX7kduwUGIuEFA6Z5pYJyQAAoDtQSW6jI9NpTU2kFA5Zp6fS1cxMqViYSjIAAOgahOQ2ObOc12wmz1bUAaXiEc1lCp2eBgAAgCRCcttU+pEvoR85kGQsrDlWtwAAAF2CkNwmR6bTSsbC2jGc6PRULgjJWETzy1SSAQBAdyAkt4FzTkemM5qaSClk9CMHkYqHNUdPMgAA6BKE5DaYy+R1ZqVAP3IdkrGIFlYKKpbKTT/WI8/P6/RitgWzAgAA/YqQ3AZHpulHrldlQ5EzK821XJSd05e+fUx/++SpVkwLAAD0KUJyGxyeSWsoHtHkYLzTU7lgVDYUabblYilbVNlJh6bTKpVdK6YGAAD6ECG5DY6fWdG+8aSMfuTAkjGvktxsSF70K9HZQlnPzy03PS8AANCfCMltsJgtamQg2ulpXFBS8dZsTb2wpl3jmVNLTT0WAADoX4TkFssVSsoXyxpOEJLrkapUkpvcmroSkncOJwjJAACgYYTkFlvMFiVJwwORDs/kwpKs9CSnmw/JkZDp6r2jOrGQPaeyDAAAEBQhucUWs14oG6KSXJdIOKTBeKQlleSRgagu3zEkSfoe1WQAANAAQnKLLfkhmXaL+o2lok33JC/6IXnHcFwjA1E9TUgGAAANICS32OKK326RoN2iXtuSMc01uTX1QtYLyWamy3cM6tBploIDAAD1CxSSzezNZva0mR0ys/fVuD9uZp/373/QzPb7x681s0f9r8fM7G2tnX73WcoWFIuEFI+GOz2VC862VKypSnLZOS2uFDTsryxyxY4h5YplPTeXadUUAQBAn9g0JJtZWNInJV0v6SpJN5vZVVXD3iVp3jl3qaSPS/qYf/y7kq5xzl0t6c2S/sTMerrEupgtUkVu0Fgq1tQ6yZmct5FIZfm9A5ODCpvpmZO0XAAAgPoEqSRfK+mQc+6Icy4v6U5JN1SNuUHS7f73X5T0RjMz59yyc67oH09I6vm/ey9mC1y016BtyeZCcmUli0pIjkfDungiqWdOpVsyPwAA0D+ChOTdkl5Yc/uYf6zmGD8UL0galyQzu87MHpf0HUnvXhOaV5nZrWZ20MwOTk9P1/8qusgSleSGjaViWimUtJIvNXR+Zbe94TUbuVyxY0gnF7M6fmalJXMEAAD9IUhIrrW3cnVFeN0xzrkHnXPfJ+lVkt5vZonzBjr3KefcNc65ayYnJwNMqTu5Sk8sleSGbEvFJEnzDS4Dd6aqkixpdSm4rz9zYf/yBQAAtlaQkHxM0t41t/dIOr7eGL/neETS3NoBzrknJWUkvbTRyXa7lUJJxbLTEFtSN6QSkhttuVhcKSgcstWNSSRp+1BcowNRfe3p0y2ZIwAA6A9BQvJDki4zsykzi0m6SdJdVWPuknSL//3bJd3nnHP+ORFJMrOLJV0h6dmWzLwLre62R7tFQ5oNyZWNREJ29g8bZqbLdw7pG9+bUb5Ybsk8AQBA79s0JPs9xO+RdK+kJyV9wTn3uJl9xMx+0h/2aUnjZnZI0m9KqiwT94OSHjOzRyX9haRfcc7NtPpFdIulFXbba8ZYsrl2i4WVYs1Wlyt2DCmTL+nh5+abmh8AAOgfgUqezrm7Jd1ddewDa77PSrqxxnmfk/S5Jud4waCS3Jym2y2yBe0dGzjv+EWj3rEjM2m95sB44xMEAAB9gx33WqiyJTWV5MZ4O+WpoQ1FnHOr7RbVBuMRmUmnF3OtmCYAAOgDhOQWWswWlIiGFIvwtjYiHDKNJWOaa6DdIpMvqVR25yz/tvZxtyVjmk4TkgEAQDCkuRZaXKcnFsGNJaMNtVss1lj+ba3JoTiVZAAAEBghuYWWsoWalUwEt63Bramrd9urtn04oemlbFNzAwAA/YOQ3EKL7LbXtLFkTPOZQt3nbRqSh+I6vUQlGQAABENIbpFy2WkpW+CivSZtSzXWk7ywUlDIpFS89i8p24fiml7KqVyu3iwSAADgfITkFplbzqvsWP6tWdtSMc1n8nKuvjC7uOK1uqzdSGStyaG4imXX8BrMAACgvxCSW+TUotfvSiW5OdtSMRXLbnXN6aAWVgoa2eC93z6UkCRaLgAAQCCE5BaprJzAhXvNWd11r86L9xZWNr5ocvtwXJI0TUgGAAABEJJbpFJJpt2iOau77tXRFuGc02K29kYiFduHvJBMJRkAAARBSG6RU34leZCQ3JSxVP2V5JVCSYWS2yQkV9otWAYOAABsjpDcIqeWskrFwoqEeEubMe6H5Nk6QnJl+beN2i0GYmENxSNsKAIAAAIh0bXI6cUs/cgt0EgluRKSRzd5/yf9ZeAAAAA2Q0hukVOLOQ3RatG0VCysWDhUV09ykEqy5G9NTbsFAAAIgJDcIicXsxpm+bemmZnGUtG6KsmL/kYim/2S4m1NTSUZAABsjpDcAsVSWTPpHGskt8i2VFxzdWxNvbBS1FBi/Y1EKtiaGgAABEVIboGZdF7OScMDtFu0wrZUVHOZ4GF2caUQaOm97UNxLedLSufq26gEAAD0H0JyC5xdI5lKciuMJWOaX66nkrzxGskVk5W1khfpSwYAABsjJLcAIbm1tqVimgvYk+ycCxyS2ZoaAAAERUhugVN+6Bqi3aIlxpIxLawUVCyVNx2bLZSVL5UDLb9X2ZqakAwAADZDSG6B04tZhUwajBOSW2F80Fsr+czK5i0XC1lvTLBKMu0WAAAgGEJyC5xazGpyKL7p6goIZizpheQgLReLK8FD8shAVLFwSNNpKskAAGBjhOQWOLWY047hRKen0TO2pYKH5IU6QrKZebvusTU1AADYBCG5BU4tZlcvCkPzKpXkIBuKLKwUZFLgNaonWSsZAAAEQEhugdNLOe3wLwpD81YryQG2pl5cKWgwEVE4FKzVZTtbUwMAgAAIyU3KFUuay+Rpt2ihsZRXFQ5aSQ7SalGxfZhKMgAA2BwhuUnTfuCiktw68UhYg/GIZgOE5NNLOU0MBn/vtw8ldGa5oFyx1MwUAQBAjyMkN+mUfxHYdirJLTWWim5aSV7OF7WwUtDOOt77yjJwM+lgm5UAAID+REhuUmXN3R1cuNdS25IxzW2yNfVJ/73fORL8vWdragAAEAQhuUmVLalpt2itsVRs00ryyYX6QzJbUwMAgCAIyU06tZRTNGyry5ahNbalYpuuk3xyIatkLKyhOnY6ZGtqAAAQBCG5STNLOY2n4goFXIIMwWxLBgjJi1ntHEnI6tjpcDwVk5k0TbsFAADYACG5SelcUcMDwSuZCGYsFdNKoaSVfO1VKMplp1OL2bou2pOkSDik8RTLwAEAgI0RkpuUzhWVquPP/QimsqHI/Dobijw/t6xCydUdkiXv4r1pQjIAANgAIblJS9miBgnJLbdnbECSdOh0uub9T51clFTfRXsV29maGgAAbIKQ3KR0rqihBCG51V558ZiiYdM3D8/UvP/JE0synV2toh5sTQ0AADZDSG5SJkcluR2SsYhevndM3zo0W/P+p04uanwwrlik/o/w9uG4ZtJ5lcqu2WkCAIAeRUhuUjpLT3K7vPbScX33+ILO1OhLfurkUkOtFpJXfS6V3aarZwAAgP5FSG5CueyUzhfrWqcXwb3u0gk5Jz1w5NxqciZX1HOzyw1dtCed3ZqalgsAALAeQnITlgslOScN0pPcFj+wZ1TJWFjfrGq5ePrUkiRpV4OV5NWtqbl4DwAArIOQ3IRMrihJGoxHOzyT3hSLhHTt1LbzLt576oQXkhuvJHvnsQwcAABYDyG5CUtZLySn4uEOz6R3vfbAuI5MZ3Ry4WxrxFMnFzUYj2g02dgvJ5WtqQnJAABgPYTkJqT9SjJLwLXPaw9MSJK+eehsNfmpE0u6YudQXdtRr5WIhjWUiOg0W1MDAIB1EJKbkM7SbtFuV+0a1lgyutpy4ZzTkycXdeXOoaYelw1FAADARgjJTahUkmm3aJ9QyPSaA+P61qFZOed0YiGrpWxRV+4abupxJwnJAABgA4TkJqy2W1BJbqvXHpjQycWsjsxkVrejfknTleQES8ABAIB1EZKbkM4WJLEEXLu97lKvL/lbh2f1pL+yxeUtaLeYXsrJOXbdAwAA5yPdNYF2i62xfzypi0YS+tahGUXCIe0ZG9Bwornq/fbhuLKFspZyxaYfCwAA9B4qyU1I50qKhUOKRwjJ7WRmeu2lE7r/yKyeOL6gK3c2148sSTv8NZZPnKHlAgAAnI+Q3IR0rkCrxRZ53aXjOrNc0OHpjF6yq7lWC0mamkhJko7OZJp+LAAA0HsIyU1IZ4sajBOSt0JlvWRJLakk7yckAwCADRCSm5DOEZK3yo7hhC7dPihJurIFleThRFQTgzEdnUk3/VgAAKD3EJKbQEjeWm+4YlLDiYj2j6da8nhTEyk9O7PckscCAAC9hZDchHSuSE/yFvrNN12hu3/9hxQONbYddbWpiZSO0G4BAABqICQ3gZ7krTUQC2vPWLJljzc1MaiZdE5L/nrXAAAAFYTkJlBJvrBNTXiBm5YLAABQjZDcBHqSL2xTE96FgEe4eA8AAFQJFJLN7M1m9rSZHTKz99W4P25mn/fvf9DM9vvH32RmD5vZd/x/f7S10++cQqmsbKFMSL6AXTyelBnLwAEAgPNtGpLNLCzpk5Kul3SVpJvN7KqqYe+SNO+cu1TSxyV9zD8+I+mtzrmXSbpF0udaNfFOy/hbUhOSL1yJaFgXjQzoWUIyAACoEqSSfK2kQ865I865vKQ7Jd1QNeYGSbf7339R0hvNzJxzjzjnjvvHH5eUMLN4KybeaUtZPyTTk3xBm5pIUUkGAADnCRKSd0t6Yc3tY/6xmmOcc0VJC5LGq8b8c0mPOOdyjU21u2TyVJJ7QWUZOOdcp6cCAAC6SJCQXGtR2upEseEYM/s+eS0Yv1TzCcxuNbODZnZweno6wJQ6L50lJPeCqYmUlrJFzWXynZ4KAADoIkFC8jFJe9fc3iPp+HpjzCwiaUTSnH97j6S/kPTzzrnDtZ7AOfcp59w1zrlrJicn63sFHbKUo92iF0xNeLv30XIBAADWChKSH5J0mZlNmVlM0k2S7qoac5e8C/Mk6e2S7nPOOTMblfRlSe93zn2zVZPuBlSSe0MlJLPzHgAAWGvTkOz3GL9H0r2SnpT0Befc42b2ETP7SX/YpyWNm9khSb8pqbJM3HskXSrpd8zsUf9re8tfRQewukVv2DM2oEjIqCQDAIBzBEp4zrm7Jd1ddewDa77PSrqxxnm/K+l3m5xjV0rTbtETIuGQ9o0nWQYOAACcgx33GlRZAi4VIyRf6KbGWQYOAACci5DcoHSuqGQsrHCo1sIeuJBU1koul1kGDgAAeAjJDcrkivQj94ipyZRyxbJOLmY7PRUAANAlCMkNWsoV6UfuEVPjLAMHAADORUhuUDpb1BCV5J4wNckycAAA4FyE5Aalc0WlCMk9YcdQQgPRMCtcAACAVYTkBtGT3DtCIdP+CVa4AAAAZxGSG7SUpSe5l0xNJAnJAABgFSG5QekcPcm9ZGoipRfmllUolTs9FQAA0AUIyQ1wztGT3GOmJgZVLDsdm1/p9FQAAEAXICQ3IFcsq1R2tFv0kKmJpCTp6Ey6wzMBAADdgJDcgMqW1LRb9I6piUFJ0tGZ5Q7PBAAAdANSXgPSOS8kU0lurTsefL5jzz2WjGpkIEolGQAASCIkNyTtV5JTMd6+C0WQAD6ciOj+w7O648Hn9Y7r9m3BrAAAQLei3aIBVJJ70/hgXDPpfKenAQAAugAhuQGVkDwUj3Z4JmilXSMJLawUVn++AACgf1EKbUA6V5AkpeLhDs8ErbRnzFvh4thcfRfvvTC3rJ/4o2/o5159sS4aHdh0PK0cAAB0PyrJDaj0JNNu0Vt2jw4oZNILda6VfNdjx7WwUtAzp5baNLPz5YolHTq9dc8HAEC/IeU1IJ0rSaLdotfEIiHtGE7o2Hx9leR7vntCUv3hOqhaFx1+6/CM7vnOSb3/+iuVXLMUIVVqAABag0pyA9K5gsIhUyLK29dr9owN6IX5ZZXLLtD452eX9d0XFxUJmY7NLcu5YOc16+RCViXndHoptyXPBwBAvyHlNSCdLSoVC8vMOj0VtNjesaSyhbKOzmYCjf/K414V+dWXjGspV9TCSqGd01tVWYVjOk1IBgCgHQjJDVjKFTWUoNWiF+3Z5l2899gLZwKNv+e7J/XS3cN62e4RSdKxNrVcVJvxw/E0lWQAANqCkNyATK6oQbak7knbh+KKRUJ6NEBIPrGwokeeP6PrX7pLu0YSCoes7n7mRmQLpdVl6gjJAAC0ByG5AelckZUtelTITLtHBwKF5K9896Qk6c0v3alIOKRdI4m2Xby3VqWKHIuEaLcAAKBNCMkNSGeLSlFJ7ll7x5J68sSisoXShuPu+e5JXbFjSAcmByV56yy/eGZF5TZfvFcJyZdvH9R8Jq9CqdzW5wMAoB8RkhuQzhU1REjuWXu3DahQcnrixOK6Y6aXcnro2Tm9+aU7z543NqB8sdz2FoiZdF4m6Yqdw3KSZjNspQ0AQKsRkhuQpie5p1V23nv0+fVbLu59/KSck65/2dmQvLpjX5v7kmfSOY2lYto1kpBEXzIAAO1ASG5AOktPci8bGYhq53Biw77kr3z3pKYmUrpix9DqsfHBmBLRUNv7kmfSOU0MxjQxGJdESAYAoB0IyXUql50y+RI9yT3u6r2jeuxY7ZA8n8nr/iOzuv6lO89ZKztkpj2jybZWkp1zmknnNT7orcIxOhBd7VEGAACtQ0iuUybvLb1FT3Jvu3rfqJ6bXdZcjX7fv3nilEplp+tfuuu8+/aMDejkQrZtF9MtZYvKF8ua9KvIk0NxKskAALQBIblOlfVpabfobT+wZ1RS7U1F7vnuCe0ZG9BLdw+fd9+esaTKTjpxpj0tF5WqcaXVYsIPyVu1HTYAAP2CkFyndNYPyVSSe9r37xlRyKRHqkLyA0dm9fffm9FbXrar5rbke7YNSFLb+pKnV0NyTJI0ORhXvlTWov+5BAAArUFIrtNSjpDcD1LxiC7fMXROJfn52WX98p89rIvHk/rVN1xa87zhRFQjA1G90Ka+5Nl0XtGwaXjA2xZ9coiL9wAAaAdCcp0ytFv0jcrFe845LWULetftD6nspE/f8iqN+CG1lj1jAzrWpkryTDqn8VRcIb+KfTYkZ9vyfAAA9CtCcp1ot+gfP7B3VGeWCzoyk9Gv3/mojsxk9P/97Cs0NZHa8Lw9Y0nNZfJazrW+BWJ6KbfaaiF5F5DG2Z4aAICWIyTXiXaL/nH1Xu/ivV/979/WfU+d1ofeepVed+nEpuftGfP6ko+1+OK9Utlpfjm/etGeJJkZK1wAANAGhOQ6UUnuH5fvGFIyFtZTJ5f0c6++WD/3mv2Bzts9OiCTWt6XPJ/Jq+y8FS3WmhwkJAMA0GokvTpVepLZTKT3hUOmN75khzK5oj7w1qsCn5eIhjU5FNexudZWkquXf6uYHIrrkRfOKFcotfT5AADoZyS9OqVzRcUjIcUiFOH7wR/d/HI552ou97aRvWNJPXVysaFz11O9/FvF6sV79CUDANAyJL06LeWKtFr0mUZC7u6xAWXyJS2sFFo2j5l0XslYWMnYuZ+/yu57tFwAANA6hOQ6pbNFln/DpirV3Zn0+dtaN2omnTuv1UKStg3GFDIqyQAAtBIhuU4ZKskIoBJmZ1oYXNcLyZFQSNtSMSrJAAC0ECG5TrRbIIjhRETRsGm2RSE5VyhpKVvUZFU/cgUrXAAA0FqE5Dqls4RkbM7MNDEYb1m7xUzGe5zxGpVkyWvvmM3kVSyVW/J8AAD0O0JyndI5epIRjBeSW1PdnfGrxNVrJFdMDsVVKru2bYcNAEC/ISTXiZ5kBDUxGNP8cl7FcvPV3Zl0TiZpPLV+u4UkHZ5ON/1cAACAkFy3JSrJCGhiMK6yk+YzzS8DN5POaTQZVTRc+79spcJMSAYAoDUIyXXIFUvKF8sajBGSsblWrnAxk87XXNmiIhmLKBWP6PDpTNPPBQAACMl1yeS8bX+pJCOIcX8limZDsnNOM+ncuhftVUwOxqkkAwDQIoTkOmRyRUmiJxmBJGMRJWPhpkNyOldUrlhed/m3ionBmJ6dXW7quQAAgIeQXIelrBeSh6gkI6BWLANXOX+jdgtJGklGNZPOKVcsNfV8AACAkFyXtF9JTlFJRkCTLVgGbnX5t01C8uiAV2k+uZBt6vkAAAAhuS7pnLdKAe0WCGpiMKalbFG5QuPV3Zl0TpGQaSQZ3XDcqH//i2dYKxkAgGYRkuuQ9i/co90CQVUutqvsmNcI76K9mEJmG44bGfBC8vEzVJIBAGgWIbkOS1mvkky7BYKqrF/cTMvF9CbLv1VUQvIJKskAADSNkFyHeb8aOJbceJUBoGI8FZOp8ZBcLJU1l8kFCsnRcEgTgzEdXyAkAwDQLEJyHWYzeaViYSWi4U5PBReIaDikkWRUsw2ucHFsfkVlt/lFexUXjQ7oRdotAABoGiG5DnOZvLZtslYtUG2iiRUujs5k/McI9rm7aGRAx2m3AACgaYFCspm92cyeNrNDZva+GvfHzezz/v0Pmtl+//i4mf2dmaXN7BOtnfrWm8vktS0VrKIHVEwMxjSTzsk5V/e5lR30glaSd40mdOLMSkPPBQAAzto0JJtZWNInJV0v6SpJN5vZVVXD3iVp3jl3qaSPS/qYfzwr6XckvbdlM+6g2XRe4ykqyajPxGBc2UJZmXz9y8AdncloIBoOfLHo7tEBZfIlLa4U634uAABwVpBK8rWSDjnnjjjn8pLulHRD1ZgbJN3uf/9FSW80M3POZZxz35AXli94XiWZkIz6VKrAlU1B6nF0JhO41ULyepIl1koGAKBZQULybkkvrLl9zD9Wc4xzrihpQdJ4KybYLZxzmstQSUb9VkNyA33JR6YzgVstpLMhmb5kAACaEyQk19rBoLrhMciY9Z/A7FYzO2hmB6enp4OetqXSuaLypTKVZNRtNBlV2Ewzda5wkckVdXIxq8mhOkLySEKSWAYOAIAmBQnJxyTtXXN7j6Tj640xs4ikEUlzQSfhnPuUc+4a59w1k5OTQU/bUnP+GsmEZNQrZKZt/sV79Xh21lvZYryOSvLEYFzRsLHrHgAATQoSkh+SdJmZTZlZTNJNku6qGnOXpFv8798u6T7XY5fXz/oheZwl4NCARpaB9ZALXgAAIABJREFUOzJd3/JvkhQKmXaxDBwAAE3b9JJ551zRzN4j6V5JYUmfcc49bmYfkXTQOXeXpE9L+pyZHZJXQb6pcr6ZPStpWFLMzH5K0o85555o/Utpr7l0pZLMEnCo38RgTM+cWlK5jt8dK2skj9f5mbtoNEFIBgCgSYHWlXLO3S3p7qpjH1jzfVbSjeucu7+J+XWNSrsFF+6hERODcZXKTmeWC4HPOTqT0e7RAcUi9e35c9HogB44PFvvFAEAwBrsuBfQLD3JaEIjK1wcmU5raiJV93NdNDKgU0s5FUvlus8FAAAeQnJAc5mc4pGQkrFwp6eCC1ClrzhoSHbO6chMprGQPDqgUtnpdAPrMgMAAE+wbbygWX+NZLNaq90BGxuMRxSPhAIvAzebyWspW9Qlk42EZH8ZuDMrq+smt8IdDz5f1/h3XLevZc8NAMBWo5Ic0Fwmr22sbIEGmZkmBuOaDVhJrly010gleTe77gEA0DRCckDeltSsbIHGTQzGdHIxqyCrIx6ZTkuSLpkYrPt5dvkh+cQCayUDANAo2i0Cmk3ndWCy/sCCC1O9rQVBXL5jSI8dW9C3n5/XKy/etuHY/7+9O4+Oo74SPf69vahbW2vfJVuSV2zj3WCDAYfNZodkEkgIgQBDkiGELDNMZua9TJKZnMzMy0sgDyaEENZAgBASbAdwGFbbeMH7gjFeJFu2JGu1LLWkXn/vj2rZsqylJbUttbmfc/q0uruq+qculXTrp1v37m/wkmC3UZQx+HSJFJcDj9uhZeCUUkqpYdCZ5ChZM8mabqGGbkqBB4dNeHVLz4aVp6qo9zI2Kwm7bWg58IXp2lBEKaWUGg4NkqPQ4Q/REQhptz01LC6nnXMKPPxlWw2BAcqzDbWyRZei9EQOa2tqpZRSasg0SI5Co9e62EobiajhmlGcTqPXz+q9DX0uEwobDjR6KRtCZYsuhemJ1LToTLJSSik1VBokR6HJqy2pVWxMzEvB43awtJ+Ui8PNHQRChnFDuGivS0G6m6PtAby+4JC3oZRSSn2aaZAcBe22p2LFYbdx9bkFrNhZS4c/1Osy+xqsyhbDmUkuOl7hQmeTlVJKqaHQIDkKTZEGEJpuoWLh+pmFeP0h3vr4SK+vV9QPvUZyl8LjtZJjl5fsC4Ro15lppZRSnxIaJEfheLqFXrinYuD8sizyPK4+q1xUNHjxuB3DOinrCpJjWeHi+fUHefT9/YSjqPOslFJKxTsNkqPQ6PXjtAupLi0rrYbPbhOum17Iu7vraGkPnPJ6RYOXspyUYbVAz0t1YROoiVGQvLeulT11bTS0+ahs9MZkm0oppdRopkFyFJq8PjKTE4YVtCjV3Q0ziwiEDK/vqDnltf31bZQPI9UCrNznPI87ZukWT39wAIdNcDlsbKxsjsk2lVJKqdFMg+QoaEtqFWvTijyUZyeflHJR09LB/S9sprqlkykFnmG/R6waihzrDPDHTYeYXpzG9OI0dlS30Bno/aJDpZRS6myhQXIUGr1+vWhPxZSIcP3MQtZWNHKg0cvDb+/h0p+9x+s7arnv0vHctmDssN+jMD2R6hhUt/jjxkO0+0PML89izthMAiHD9sMtw96uUkopNZppkm0Umrx+SjKSRnoY6ixz/YxCHvyfPSx+8H06A2GWTM3nX645h5LM2PysFaa7WbGzk3DYYBtie+tw2PDMmgPMGpNOcUYSxhhyUl1sPNDMvNLMmIxTKaWUGo10JjkKTW1+rZGsYq48J4WF47MpzUrmubvP59Hb5sQsQAYoTEvEHwwfr/M9FCv3NlDR4OX2BaWANQM+Z0wGB5vaqWvVttdKKaXOXhokD8AXDNHqC2q6hTotnr3rPN749sVcOD475tuORRm4pz+oJDvFxdXnFhx/btaYdGwCmw7oBXxKKaXOXppuMYBmr1WiS2skq1h5ft3BM/I+heluwAqSZ5SkD3r9A41e3tldx32fGU+C48T5dKrbyaS8VDYfPMoVU/KxDzGVQymllBrNdCZ5AI1eH6Dd9lT86WpNXd0ytLSIZ9ccwC7CrfNPvYhwztgMWn1B9hxpHdYYlVJKqdFKg+QBHO+2pyXgVJxJS3SSlGAfUrpFuz/ISxuqWDwtnzyP+5TXJ+V7SHY52HhQUy6UUkqdnTRIHsCJIFlnklV8ERHGZiXzUfWxQa+7dEs1xzqD3HFBaa+v223CrJJ0dtUco80XHOZIlVJKqdFHg+QBNLZZQbKmW6h4dNnkXNZXNh0/2YvWn7ccpjwnmbljM/pcZvbYDMIGtlQdHe4wlVJKqVFHg+QBNHn92G1CWqJzpIei1KAtmZZPKGx486PaqNc5cqyTdRVNXD+jsN9W7PkeNwVpbj6q1sYiSimlzj5a3WIAjV4/GUnOITdjUGokTS30UJKZyOs7arl53pio1lm+rQZj4NrphQMuOzEvlZV76ukMhHA77UMe52Arfnzp/Oi+F6WUUmqodCZ5AE1en+Yjq7glIlw1rYDVexto6QhEtc6yrdVMKfAwPjdlwGUn5qUSNrC3rm24Q1VKKaVGFQ2SB9DY5idLK1uoOLZkWj6BkOHtj48MuGxVUztbqo5y3YyBZ5EBxmQm4XLY2HOagmSvL0gwFD4t21ZKKaX6o0HyAJq8fm0kouLazOJ08j1uXt8+cF7ysm3VAFw7vWCAJS12mzAuJ4U9R1oxxgxrnN0ZY1hX0ch/vvExS7dWx2y7SimlVLQ0SB5Ao9evlS1UXLPZhCXT8nnvk3q8A5RrW7a1htlj0inJTIp6+xPzUjnaEaC+1TfcoQJWK/iXNlTx6pZqnHYbm6uO0toZXaqIUkopFSsaJPcjEArT0hHQnGQV95ZMy8cXDPPu7vo+l9lb18qummNRp1p0mZBn5S5/EoOUi9qWTh55Zx/bDrVwxZQ8vnZJOaGwYX1l07C3rZRSSg2GBsn9aG7XGsnq7DCvNJPslARe31HT5zLLttYgAtecG12qRZeMpARyUlzDblG943ALv3pvL75AiLsWlvGZSbnkprqZmJfC+v1NBMOam6yUUurM0SC5H9qSWp0t7Dbhiin5vPNxHZ2B0CmvG2NYtrWa+WVZ5PbShnogE/NSqGjwEhjiRXaBUJhXt1aTm+rmm5eOpzznRGWNC8Zl0+oLsuOw1mNWSil15mid5H40tWlLahVf+qs37HbY8PpD/OQvuzinwAOcqDe8s/oY+xu83H1R+ZDed0JeKqv3NVLR4GViXuqg19926CheX5Cb55aQ6j65cc/43BSyU1x8sK+RmSV9dwBUSimlYklnkvvRGJlJztLqFuosUJ6TQqLT3uuM7LJt1ThswlXT8oe07bLsZBw24ZMhpFwYY1i1t4F8j5txOcmnvG4TYcG4LA41d1DV1D6k8SmllFKDpTPJ/TiRbqFBsop/dptwToGHj2paCIbDOGzWObIxhuVba7hoQjYZQ/xZd9ptlOck88mRwV+8t2pvA0eO+fjc7OI+22DPHpPOX3fWsnpfA7dkDq7bnnbzU0opNRQaJPej0etHxLowSamzwbRCD5sONvPEqgpCYcNvVu6nsc3Hsc4g371i4rC2PSE3lb9sr6HZ6x9UsP3bVRWkuBzMKE7rcxmXw87csRms2d/IsWnxWw5OA3allIofmm7Rjyavj/REJ3Zb77NbSsWb8bkplGUn4w+GcTntTCtK46ZZRfzjkslcO2NwVS16OlEKLvqUi711rby7u5755Vk47P3/OlowLhtjYF2FloNTSil1+ulMcj+avH5NtVBnFYfdxt92uzgvljOVOSku0pOc7DnSxvllWVGt89tVlbgcNs4vyxxw2czkBCbnp7K+opHOQAi30z7cISullFJ90iC5H41tfrK0/JtSURERJuamsvXQUULhgVtUN3n9vLLpEJ+dXUyyK7pfRQvGZbOrtoI/bT7MF88bXoDf4Q+xbFs1nYEQeR535ObCFwzhcpyZADxsDHWtPvzBMCUZiX3mZCullDrzNEjuR5PXz7hu9VqVUv2bmJfC+somKhu9Ay773NoD+IJh7lpYyvqK5qi2Py4nmZKMRH751h5umlU05NnkNl+QJ1dXUHfMR1ZKAp8caaUrrv/1e/v5xqJx3Hfp+AFTQAbLGENFg5d99V6qmtupamrHF7RqSy+amMMVU/I0UFZKqVFCg+R+NHn9zCvTdAulojUuJ4WkBDuvbqnmW5dOIC3J2etyvmCIp9ccYNGkHMbnpkYdJIsIi6fm8/iqCn639sCQ6jq3dAT47aoKWjr83LZgLBPzUgmGwzS0+TlyrBOvL8hDb+1h1d4GHrx5JiWZSYN+j76sq2hi6dZqBMhPczOjJJ0xmUlUNnh595N6AqEwV59boIGyUkqNAnrhXh/CYUNzu19bUis1CC6nnVvPH0uz18+9z2/qswPf4ysraGjzcffCwQe55TkpXDQhm0fe2Utr5+AqXTS2+fj1+/to7QxwxwVlxxufOGw28j1uZhSn8/CXZvPQLTP5pLaVqx9ayatbDg96jL3ZdLCZv2yrYVJeKj+4dgr3XTqBG2cWMXtMBjfNKmJBeRar9zWydGs1YTNwuopSSqnTS2eS+3C0I0DYaI1kpQarLDuZG2cV8cdNh/jh0p38+43Tjs+MBkJhfrRsJ79be5ArpuRx4fjoLvDr6YHFk7nu4VX8ZmVF1KXrals6eXJ1BSFjuHthOUUZiX0ue0MkeP32i1u4/4UtvPNxHd+9YhJjsoY2q9zY5uPe5zbhSXTwhbkluHqkiYgI104vwGEXVu5pIBQ23DiraEjvNRoMptSdlrlTSo1WOpPchyavD9AgWamhmDM2g69fMo7n1h3kqQ8qASt96bbfruN3aw/y9UvG8eiX5ww5reDc4jSuObeAx1fup6HNN+DyVU3t/GblfkTgby/qP0DuUpKZxIv3zOfbl09g+bYaFv3sHb727AbWVzRhBjHTGwobvvXCZhq9fm49fyyJCb3nUYsIS6bmc+nkXDYcaObljYcI9jETH28CoTB1rZ34gqGRHopSSkVNZ5L7sHa/VYt1cr5nhEeiVHx6YPEk9te38W/LP8IYeGJ1BXWtPh68eWZMZkm/e+VE3thZyyPv7OVfr5va53L769t4Zu0BUlwO7rywbFAnvg67jW9fPpFb5o3hmTWVPL/+ICt2HmF6cRr3XFzONVHkD//fv+5m9d5G/utz0wkOUPVDRLj8nDwcNuGvHx3hWy9s5qFbZuGM8QWEp9v+hja2HDxKo9dPk9dPS4eVFpOUYOfiCTnML88iwRFf35NS6tNHg+Q+/GVbDeNykpmYp9Ut1NlrsB3gBsNmE35x80w+/+gafrz8I/I8Lv7wtQXMKEmPyfbH5aTwN7OLeW7tQe5aWEZxxqmpEG/tOsJTH1SSmZzAnReW4Uns/ULCgeSnuXlgyWTuu3QCf9x0iCdWV/DN5zfzTOkBfnTDVM4p6P1k+q87a/nvd/dxy7wSvjCvJOrPe9GkXBx2G69tr8Ef3MQjt846Y2XphmvlnnqeXF1Jgt1GTqqL8uxkMlMSSE90sv1wC2/srGX13gYWTcphXunA9bGVUmqkaJDci7rWTtZVNPLNSyfoVeZKDUOyy8ETd8zjqQ8qufPCUnI97phu//7LJ/CnLYf5xZt7+Nnnp590vC7dWs13X9xCnsfNHReURl2LeaBA1ibCnReWsfFAM+/uruOaX67ktvlj+e4Vk0hLclJ9tIO/7qxlxc4jrK9s4tyiNH54fd8z3X1ZOD6bBeWZ/O9Xd/K3z2zksdvmjPoGKhsqm7jnmY3kpLi4+6IykhJO/sznjM2kssHLm7uOsGxbDe/vaSDF7eDGmUXYtLOpUmqU0SC5Fyt21BI2cO304bXpVUpZs7Dfv2ryadl2YXoiX5k/lsdXVbB062GSXQ5SIrfdR1qZV5rJkqn5MQ8ubSLMK83kX6+bws/f/IRn1x5g2bYaijMS2XaoBbBagH/9knLuuKBsyO9/24JSXA47//jKNr765Ic8fvvcPoP9wf5XINYXzO043MJXn/yQ/DQ3t8wrOSVA7lKanczdC8vYV+9lxc5avvvSVp7+oJIfXDeFOWN1ZlkpNXpokNyL5dtqmJCbcrw8lFJq9PrelZMoTE+kvs2H1xekzRekrTPI+WWZfP+qc/jT5tiUcOtNelICP75hGrfMG8NPX99Fmy/IA0smsXhqfswaEX1hXgkJDhvf+8NWrn94FT+4biqXTMwZcD1fMMS2Qy18WNnEkWOdzCzJ4MJxWTGfzQfYW9fKV55YjyfRye/uPp/3dtf3u7yIMD43hfKccSS77Pzn67v53K/WcO30Ar5/1eReU2e6PL/uIMYYvP4QrZ0B2jqDtHYG8fqDjM9NoSDtxEWZWjlDKTUcGiT3UHesk/WVTdx/2YSRHopSce105jt3l5hg586FZWfkvfoypdDDs3edf9q2f+OsIjKTE/jBqzu4/Yn1XDY5l3+55hzKewTiobCh+mgHGw40sfVQC/5gmNxUF1ML09h8sJkPK5uYmJfCheOzMcbEJJ2sssHLrY+vw24Tnrv7fIrSB64c0sUmwk2zilk8NZ9fv7efX7+/jxU7a5k7NpOLJmZz8YQcphR4sNmExjYfK/c08NKGKvYcacXrP7VShsMmfHZ2MTNjlPeulPp00yC5h9d31GIMXHOuploopUaPiyfmsOI7F/PU6kr+39t7Wfzg+3xlQSl5Hhcf17ayZl8jda0+QmGD0y5ML0pnXmkGJZlJiAhXn1vA+opG1u5v4snVlWw6eJSHbp5JaXbykMe0ak8D9z6/CZvAC/csGNK2uk6m8jxu7r9sImv2NbCnro01bzTyX2/sJinBTlqik9qWTgxWhYyJeakUZySS6nbicVvpNTab8IcNh3hpQxW1LR1cOTV/yN+XUkoByGDqfZ4Jc+fONRs2bBix9//Co2to6Qiw4jsXD3rdMzVzppQaHQb77/xY/Y5o7Qzw5kdH2HigGQN43A7yPG7y09wUpLmZnO/pMw86GAqzpeoob31cRyhs+MlN07hh5uBK8hljeHJ1JT95bRfjcpL5zVfmMjbrRIAci++ztTPA3ro29ta10dIZoDw7hYl5KRSmJ2LrYwY8GA6zfFsN6yuamJSXyh++sQCPe2gVTZRSnw4istEYM7e313QmuZsjxzr58EAT37k8ug5eSik1ElLdTj47u5jLzsnDaROSoqzcAVbt57mlmXxv8STu//1m7n9hCx/sbeSH10/ts9FJd52BEP/rzzt4eeMhrpySx89vnknKIN4/WqluJ7PGZDBrTEbU6zhsNm6cWURBmptlW6u58ZHVPHbbXMbnailPpdTgaTX3bl7fXoMxcLWmWiil4kBaonNQAXJ3RemJvHDPfO79zDhe2ljF9Q+v4u2Pjxxv/NFTfauPV7cc5ubH1vLyxkPcf9kEHv3ynNMSIA/X+WVZ3LmwjKPtAW54eBWvb68Z6SEppeLQ6PvtNoL+sr2GyfmpOuuglIpKvKdYOew2/mHxZBaUZ/PtF7dw51MbEIEJuSnMHpPBtKI0Khq8rN7bwMe1rQBkJDl59MuzWTJtdE8mlGensPy+hfzdc5v4xnObuOfich5YPAnHKO9eaIyhzRekvtVHc3uAjCQn+WnuPkvqKaVOHz3qImpbOvmwspnvXaGpFkqpT5eFE7J5/4FFbD54lE0Hmtl4sJnXttfwwodVJDhslGQksnhKHuNyrZzgJm8gLk4QCtMTefFr8/n35bt47P39bKk6ysNfmkVuauzL4A1FZyDEpoPNrN3fxPqKRg41d1Df6sMXDJ+ybKrbQb7HTXFGInNLM5lfnsX04rS4a1mu4kd/x7gxBl8wTDBsCIas+8VT88lMTiAn1XUGR3l6RRUki8gS4CHADjxujPmPHq+7gGeAOUAjcLMxpjLy2j8BdwEh4FvGmBUxG30MvRb5d9zV2kBEKfUplJTg4MLx2Vw4PhuAcNhQ1dxOnsfNK5tOX63p083lsPNvN05j1ph0/vlP21ny4EoWTcxh1ph0Zo3JYHJ+alSzy8FQGINVZm4opfOeX3eQQCjMwaZ2Khq87K/3UtXcTihsEKyAPifVRVlWMimRih1JCXba/SGOdQRo6QxyrCPAzupjvBOpQ51gtzE2K4kbZxUxa0w6M4rTo+4sqVS0uuqSH25u59DRDg43W7dWX/Ck5X7+5ieIwHmlmVw3o5CrpuWTlRLfAfOA1S1ExA58AlwBHAI+BL5ojPmo2zJ/B0w3xnxdRG4BbjLG3CwiU4DfA+cBhcD/ABONMacWuIwYqeoWn/vVB7T7Q7x+/0VD3kY8zKwopdSnRc/qI7tqjvHg/3zCxgPNNLT5AUh02inPScYmgsH6e2gMBEJhvL4Q7f4gXn8If7fZXYdNsNuEBIeNgjQ3YzKTKMlMYkxmEvkeNwZr/UDImmWrPtrB0q01pwTF5dnJlOUkU5qVPKiujG2+IJUNXvY3eNlf30Zdqw8Am8DkfA+zx6YzITeVnFQXOakuslNcZKckYLfJ8TEFQoZAKExHIITXF6Tdb937gmGcdhsuh42EyC0pwU5OiovM5IRRn66irJ+91s4gnYEQqZETroFO7HzBEEdafNS0dFDT0kllo5d3d9fT0Oajoc1HZ8D6+RcgO9VFcXoieR43TocNh01w2IRLJuWwr87L0q2H2VfvxW4TLhyfzZVT8lg4PpuxWUkxqc0ea/1Vt4gmSF4A/NAYszjy+J8AjDE/7bbMisgya0TEAdQCOcD3uy/bfbm+3m8kguRjnQEu+OnbfGPROO79zPghb0eDZKWUGv2MMTS3B6hqaudgUztNXv/x17r+httEjgeKLoedBIcAQtgYwmFDyBiCIUNLR4Amr5+mdv9JgXR3ww2KB3L1uflsrjrK5gPNbDp4lC1VR2nrMcsXCyKQkZRAdkoCGUkJpCU68SQ68bideBIduBx2K2CyW0GTzSYYY33eobAhbCBsDMZAyJgTX4etr8PdlnXabSQm2El0WjeX04bDZsMmVsdGW2Q/+UNh/MHILfK1L3Lrer4zGKLDb53wtPtDdAZC2G1CistJistutbN3O0hJcJz42uUg0WknFDYEw5ETnq77k762Ug0CoTDBkCEQjtyHwvgCYXzBEL5gmEDIOvlIdNpxR76vpAT78c/OuneSlGDHabdFPkMbTrsQCpvj4273W7eGNh+1LZ18sK+RY50BWjsDdPhDdAasz6E7m4A78n65qdZJ3Il9Yv0Md500dt/XaW4nWSkJZKVYJ1qF6W4K0xL7/NntOik1xvBxbSvLtlazbFs1VU0dgHWx8MLx2VwwPouxWcmkJzpJT3KS6nZit41c8DzcEnBFQFW3x4eAnq2lji9jjAmKSAuQFXl+bY91B1eQ8wzwuJ1s+F+XEwj1/gtOKaXU2UNEyExOIDM5gRkx6s7XvVW2iGAXa7bZbhPcTivQPl1e214LQH5aIlefm8iSafm0+0NWy26f1bq7zRfEGLDZBLuA3WbDboMEh50Ee9fJgA27zQrKrODQCgZ9gbDV7j3S8r3NF+TIMR8HGtvpCIToCIT6PEEYLMEK0MIxaOFgj8xwOuw2EuzWzH+C3YbTbiNkDP5gO75gmM5AiGDYDPt7ELo+3xMnCg677fh/HkKRYNofCa79wTDD+TZtAikuB55EJ1nJLhIzrODb7bThdtpx2mwnThAC1n2ex3X8JMMmgk2EVLeDgrRECtKtOusFaYkUZyQOOc1KRDinwMM5BR7+YfGk4xf/rtrbwGs7anhxQ1WP5SHV5eCJO+YxtzRzGJ9I7EUTJPcW3vfcr30tE826iMg9wD2Rh20isjuKcZ0J2UDDSA9CDZnuv/im+y/+6T6Mb7r/4tsZ3X+3DnP9eT+KyTCGYmxfL0QTJB8CSro9Lgaq+1jmUCTdIg1oinJdjDGPAY9FMZYzSkQ29DUFr0Y/3X/xTfdf/NN9GN90/8U33X/DF00G/ofABBEpE5EE4BZgaY9llgK3R77+G+BtYyU7LwVuERGXiJQBE4D1sRm6UkoppZRSp8eAM8mRHONvAiuwSsA9YYzZKSI/BjYYY5YCvwWeFZG9WDPIt0TW3SkiLwEfAUHg3v4qWyillFJKKTUaRFVQ0RjzGvBaj+d+0O3rTuDzfaz7E+AnwxjjSBp1KSBqUHT/xTfdf/FP92F80/0X33T/DdOAJeCUUkoppZT6tNGq4EoppZRSSvWgQXIfRGSJiOwWkb0i8v2RHo/qn4iUiMg7IrJLRHaKyP2R5zNF5E0R2RO5zxjpsaq+iYhdRDaLyPLI4zIRWRfZfy9GLh5Wo5CIpIvIyyLyceQ4XKDHX/wQke9EfnfuEJHfi4hbj7/RS0SeEJE6EdnR7blejzex/DISz2wTkdkjN/L4okFyLyKtuB8BrgKmAF+MtNhWo1cQ+J4x5hxgPnBvZJ99H3jLGDMBeCvyWI1e9wO7uj3+T+AXkf3XDNw1IqNS0XgIeMMYMxmYgbUf9fiLAyJSBHwLmGuMmYZ1kf4t6PE3mj0FLOnxXF/H21VY1cUmYPWk+NUZGmPc0yC5d+cBe40x+40xfuAF4IYRHpPqhzGmxhizKfJ1K9Yf6CKs/fZ0ZLGngRtHZoRqICJSDFwDPB55LMClwMuRRXT/jVIi4gEuxqp0hDHGb4w5ih5/8cQBJEZ6HSQBNejxN2oZY97HqibWXV/H2w3AM8ayFkgXkYIzM9L4pkFy73prxT3q2mmr3olIKTALWAfkGWNqwAqkgdyRG5kawIPAA0BXb9gs4KgxJhh5rMfh6FUO1ANPRtJlHheRZPT4iwvGmMPAz4CDWMFxC7ARPf7iTV/Hm8Y0Q6RBcu+iaqetRh8RSQH+CHzbGHNspMejoiMi1wJ1xpiN3Z/uZVE9DkcnBzAb+JUxZhbgRVMr4kYkd/UGoAwoBJKx/kXfkx5/8Ul/lw6RBsm9i6qdthpdRMSJFSA/Z4x5JfL0ka5/K0Xu60ZqfKpfFwLXi0h6q0LNAAAFQklEQVQlVnrTpVgzy+mRf/+CHoej2SHgkDFmXeTxy1hBsx5/8eFyoMIYU2+MCQCvABegx1+86et405hmiDRI7l00rbjVKBLJX/0tsMsY8/NuL3VvmX478OqZHpsamDHmn4wxxcaYUqzj7W1jzK3AO1it7kH336hljKkFqkRkUuSpy7A6rerxFx8OAvNFJCnyu7Rr/+nxF1/6Ot6WAl+JVLmYD7R0pWWo/mkzkT6IyNVYM1ldrbjjtWvgp4KILARWAts5kdP6z1h5yS8BY7D+EHzeGNPzYgc1iojIIuDvjTHXikg51sxyJrAZ+LIxxjeS41O9E5GZWBddJgD7ga9iTcTo8RcHRORHwM1YlYI2A3dj5a3q8TcKicjvgUVANnAE+Ffgz/RyvEVOfB7GqobRDnzVGLNhJMYdbzRIVkoppZRSqgdNt1BKKaWUUqoHDZKVUkoppZTqQYNkpZRSSimletAgWSmllFJKqR40SFZKKaWUUqoHDZKVUkoppZTqQYNkpZQ6DUSkVER2DHHdRSJywQDLPCUif9PfMr2s80EUy1SKSPZQxqSUUmcTDZKVUmoQRMR+Bt5mEVZb4Jgyxgxnm4s4DWNSSqnRSoNkpZSKiMz+fiwiT4vINhF5OdKqt1JEfiAiq4DPi8hMEVkbWeZPIpIRWX+OiGwVkTXAvd22e4eIPNzt8fJIZ0FEZImIbIqs95aIlAJfB74jIltE5KJ+hnyxiHwgIvu7zyqLyD+IyIeR8f2o2/NtkXubiPy3iOyMjOW1HrPS90XGtF1EJkc7JhHJi3weWyO3C7p9po+LyA4ReU5ELheR1SKyR0TOi2rnKKXUGaZBslJKnWwS8JgxZjpwDPi7yPOdxpiFxpgXgGeAf4wssx2rJSzAk8C3jDELonkjEckBfgN8zhgzA6uNbCXwKPALY8xMY8zKfjZRACwErgX+I7LNK4EJwHnATGCOiFzcY73PAqXAuVjth3uOt8EYMxv4FVaL8GjH9Evgvcj3MhvYGXl+PPAQMB2YDHwpMu6/x2ofr5RSo44GyUopdbIqY8zqyNe/wwrmAF4EEJE0IN0Y817k+aexZnR7Pv9sFO81H3jfGFMBYIxpGuRY/2yMCRtjPgLyIs9dGbltBjZhBaUTeqy3EPhDZN1a4J0er78Sud+IFUxH61KswBpjTMgY0xJ5vsIYs90YE8YKnN8yxhisE4zBbF8ppc4Yx0gPQCmlRhnTx2PvAOtJL+t2CXLypIQ7inWi4evx/l33PzXG/Lqf9aSf17pvN0Rs/k50H2e42+NwjLavlFIxpzPJSil1sjEi0pV+8EVgVfcXI7Ojzd3ycm/DSjE4CrSISNfM863dVqsEZkZygUuwUiEA1gCXiEgZgIhkRp5vBVKHOP4VwJ0ikhLZZpGI5PZYZhXwuch48rAuyhtINGN6C/hG5H3tIuIZ1MiVUmoU0SBZKaVOtgu4XUS2AZlE0gd6uB34P5FlZgI/jjz/VeCRyIV7Hd2WXw1UYKUX/AwrDQJjTD1wD/CKiGwlktIBLANuiuLCvVMYY/4KPA+sEZHtwMucGtz+ETgE7AB+DawDWuhfNGO6H/hM5H03AlMHM3allBpNxEoLU0opFanisNwYM22Eh3LaiUiKMaZNRLKA9cCFkfxkpZRSaC6YUkp9Wi0XkXQgAfg3DZCVUupkOpOslFKjmIj8C/D5Hk//wRjzk5EYD4zOMSmlVKxpkKyUUkoppVQPeuGeUkoppZRSPWiQrJRSSimlVA8aJCullFJKKdWDBslKKaWUUkr1oEGyUkoppZRSPfx/jPa702imqvYAAAAASUVORK5CYII=\n",
          "text/plain": "<Figure size 864x576 with 1 Axes>"
         },
         "metadata": {
          "needs_background": "light"
         },
         "output_type": "display_data"
        }
       ]
      }
     },
     "47ce7327d3804ac8a069e1a3ed2ebc83": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {}
     },
     "4870b50b082343c1b10ae66f7205b797": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {}
     },
     "67f8311ce7b94b77b334607488b90a26": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "DescriptionStyleModel",
      "state": {
       "description_width": ""
      }
     },
     "7e8560b97c0e4bba8e4d29481f563f16": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "VBoxModel",
      "state": {
       "_dom_classes": [
        "widget-interact"
       ],
       "children": [
        "IPY_MODEL_887fc4eb5e31447e93348d0b558d200e",
        "IPY_MODEL_23ebe228ab0b4eaf9dbb8ba67a6b01fa"
       ],
       "layout": "IPY_MODEL_9066db510e2d43b4a4f2572363765de2"
      }
     },
     "8019edccd6ae4f3fa4dbec860463097d": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "DropdownModel",
      "state": {
       "_options_labels": [
        "product_weight_g",
        "product_length_cmproduct_height_cm",
        "product_width_cm",
        "distance",
        "volume"
       ],
       "description": "column",
       "index": 1,
       "layout": "IPY_MODEL_faea30e15df94b9f95fc6b5d878473b9",
       "style": "IPY_MODEL_06c6f89a52824c9abe54436c47248280"
      }
     },
     "887fc4eb5e31447e93348d0b558d200e": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "DropdownModel",
      "state": {
       "_options_labels": [
        "product_weight_g",
        "product_length_cm",
        "product_height_cm",
        "product_width_cm",
        "distance",
        "volume"
       ],
       "description": "column",
       "index": 2,
       "layout": "IPY_MODEL_b28997fd41fb4bea9d9b32b163ee0726",
       "style": "IPY_MODEL_fd4cf243bd7740429df59497af409e75"
      }
     },
     "9066db510e2d43b4a4f2572363765de2": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {}
     },
     "92eba6905f6e4a6b9e87d49792f5cd08": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {}
     },
     "b28997fd41fb4bea9d9b32b163ee0726": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {}
     },
     "ce54f8a888de46a5941eff5a5e6598cb": {
      "model_module": "@jupyter-widgets/output",
      "model_module_version": "1.0.0",
      "model_name": "OutputModel",
      "state": {
       "layout": "IPY_MODEL_92eba6905f6e4a6b9e87d49792f5cd08",
       "outputs": [
        {
         "ename": "KeyError",
         "evalue": "'product_length_cmproduct_height_cm'",
         "output_type": "error",
         "traceback": [
          "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
          "\u001b[0;31mKeyError\u001b[0m                                  Traceback (most recent call last)",
          "\u001b[0;32m~/anaconda3/envs/r_env/lib/python3.8/site-packages/pandas/core/indexes/base.py\u001b[0m in \u001b[0;36mget_loc\u001b[0;34m(self, key, method, tolerance)\u001b[0m\n\u001b[1;32m   2645\u001b[0m             \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2646\u001b[0;31m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_engine\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_loc\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2647\u001b[0m             \u001b[0;32mexcept\u001b[0m \u001b[0mKeyError\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
          "\u001b[0;32mpandas/_libs/index.pyx\u001b[0m in \u001b[0;36mpandas._libs.index.IndexEngine.get_loc\u001b[0;34m()\u001b[0m\n",
          "\u001b[0;32mpandas/_libs/index.pyx\u001b[0m in \u001b[0;36mpandas._libs.index.IndexEngine.get_loc\u001b[0;34m()\u001b[0m\n",
          "\u001b[0;32mpandas/_libs/hashtable_class_helper.pxi\u001b[0m in \u001b[0;36mpandas._libs.hashtable.PyObjectHashTable.get_item\u001b[0;34m()\u001b[0m\n",
          "\u001b[0;32mpandas/_libs/hashtable_class_helper.pxi\u001b[0m in \u001b[0;36mpandas._libs.hashtable.PyObjectHashTable.get_item\u001b[0;34m()\u001b[0m\n",
          "\u001b[0;31mKeyError\u001b[0m: 'product_length_cmproduct_height_cm'",
          "\nDuring handling of the above exception, another exception occurred:\n",
          "\u001b[0;31mKeyError\u001b[0m                                  Traceback (most recent call last)",
          "\u001b[0;32m~/anaconda3/envs/r_env/lib/python3.8/site-packages/ipywidgets/widgets/interaction.py\u001b[0m in \u001b[0;36mupdate\u001b[0;34m(self, *args)\u001b[0m\n\u001b[1;32m    254\u001b[0m                     \u001b[0mvalue\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mwidget\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_interact_value\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    255\u001b[0m                     \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mwidget\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_kwarg\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 256\u001b[0;31m                 \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m**\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    257\u001b[0m                 \u001b[0mshow_inline_matplotlib_plots\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    258\u001b[0m                 \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mauto_display\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mresult\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
          "\u001b[0;32m<ipython-input-39-086ee0f9cf09>\u001b[0m in \u001b[0;36mplot_hist\u001b[0;34m(column)\u001b[0m\n\u001b[1;32m      6\u001b[0m ):\n\u001b[1;32m      7\u001b[0m     \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfigure\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfigsize\u001b[0m\u001b[0;34m=\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;36m12\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m8\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m     \u001b[0msns\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdistplot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdf\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mcolumn\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkde\u001b[0m\u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m     \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtitle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Histogram of {}'\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mformat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcolumn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfontdict\u001b[0m\u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m'size'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;36m15\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
          "\u001b[0;32m~/anaconda3/envs/r_env/lib/python3.8/site-packages/pandas/core/frame.py\u001b[0m in \u001b[0;36m__getitem__\u001b[0;34m(self, key)\u001b[0m\n\u001b[1;32m   2798\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcolumns\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnlevels\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2799\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_getitem_multilevel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2800\u001b[0;31m             \u001b[0mindexer\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcolumns\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_loc\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2801\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mis_integer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mindexer\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2802\u001b[0m                 \u001b[0mindexer\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mindexer\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
          "\u001b[0;32m~/anaconda3/envs/r_env/lib/python3.8/site-packages/pandas/core/indexes/base.py\u001b[0m in \u001b[0;36mget_loc\u001b[0;34m(self, key, method, tolerance)\u001b[0m\n\u001b[1;32m   2646\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_engine\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_loc\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2647\u001b[0m             \u001b[0;32mexcept\u001b[0m \u001b[0mKeyError\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2648\u001b[0;31m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_engine\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_loc\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_maybe_cast_indexer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2649\u001b[0m         \u001b[0mindexer\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_indexer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmethod\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mmethod\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtolerance\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mtolerance\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2650\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mindexer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mndim\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m1\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mindexer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msize\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
          "\u001b[0;32mpandas/_libs/index.pyx\u001b[0m in \u001b[0;36mpandas._libs.index.IndexEngine.get_loc\u001b[0;34m()\u001b[0m\n",
          "\u001b[0;32mpandas/_libs/index.pyx\u001b[0m in \u001b[0;36mpandas._libs.index.IndexEngine.get_loc\u001b[0;34m()\u001b[0m\n",
          "\u001b[0;32mpandas/_libs/hashtable_class_helper.pxi\u001b[0m in \u001b[0;36mpandas._libs.hashtable.PyObjectHashTable.get_item\u001b[0;34m()\u001b[0m\n",
          "\u001b[0;32mpandas/_libs/hashtable_class_helper.pxi\u001b[0m in \u001b[0;36mpandas._libs.hashtable.PyObjectHashTable.get_item\u001b[0;34m()\u001b[0m\n",
          "\u001b[0;31mKeyError\u001b[0m: 'product_length_cmproduct_height_cm'"
         ]
        }
       ]
      }
     },
     "dcfa110e1f2946a6ad895baa5f9d2978": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "VBoxModel",
      "state": {
       "_dom_classes": [
        "widget-interact"
       ],
       "children": [
        "IPY_MODEL_003cb66582ec44499042c9d14d7afddd",
        "IPY_MODEL_f28ce86edf6245748c10c5aaaa926f86"
       ],
       "layout": "IPY_MODEL_e049e15d5f3e4056a245c487e7914adc"
      }
     },
     "e049e15d5f3e4056a245c487e7914adc": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {}
     },
     "f1d6eeda6ffe404582605ae670c1599a": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {}
     },
     "f28ce86edf6245748c10c5aaaa926f86": {
      "model_module": "@jupyter-widgets/output",
      "model_module_version": "1.0.0",
      "model_name": "OutputModel",
      "state": {
       "layout": "IPY_MODEL_f1d6eeda6ffe404582605ae670c1599a",
       "outputs": [
        {
         "ename": "ValueError",
         "evalue": "could not convert string to float: 'product_weight_g'",
         "output_type": "error",
         "traceback": [
          "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
          "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
          "\u001b[0;32m~/anaconda3/envs/r_env/lib/python3.8/site-packages/ipywidgets/widgets/interaction.py\u001b[0m in \u001b[0;36mupdate\u001b[0;34m(self, *args)\u001b[0m\n\u001b[1;32m    254\u001b[0m                     \u001b[0mvalue\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mwidget\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_interact_value\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    255\u001b[0m                     \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mwidget\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_kwarg\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 256\u001b[0;31m                 \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m**\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    257\u001b[0m                 \u001b[0mshow_inline_matplotlib_plots\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    258\u001b[0m                 \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mauto_display\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mresult\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
          "\u001b[0;32m<ipython-input-38-6537d71d2ca4>\u001b[0m in \u001b[0;36mplot_hist\u001b[0;34m(column)\u001b[0m\n\u001b[1;32m      6\u001b[0m ):\n\u001b[1;32m      7\u001b[0m     \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfigure\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfigsize\u001b[0m\u001b[0;34m=\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;36m12\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m8\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m     \u001b[0msns\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdistplot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcolumn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkde\u001b[0m\u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m     \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtitle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Histogram of {}'\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mformat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcolumn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfontdict\u001b[0m\u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m'size'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;36m15\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
          "\u001b[0;32m~/anaconda3/envs/r_env/lib/python3.8/site-packages/seaborn/distributions.py\u001b[0m in \u001b[0;36mdistplot\u001b[0;34m(a, bins, hist, kde, rug, fit, hist_kws, kde_kws, rug_kws, fit_kws, color, vertical, norm_hist, axlabel, label, ax)\u001b[0m\n\u001b[1;32m    175\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    176\u001b[0m     \u001b[0;31m# Make a a 1-d float array\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 177\u001b[0;31m     \u001b[0ma\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0masarray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfloat\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    178\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0ma\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mndim\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    179\u001b[0m         \u001b[0ma\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0ma\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msqueeze\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
          "\u001b[0;32m~/anaconda3/envs/r_env/lib/python3.8/site-packages/numpy/core/_asarray.py\u001b[0m in \u001b[0;36masarray\u001b[0;34m(a, dtype, order)\u001b[0m\n\u001b[1;32m     83\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     84\u001b[0m     \"\"\"\n\u001b[0;32m---> 85\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcopy\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0morder\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0morder\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     86\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     87\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
          "\u001b[0;31mValueError\u001b[0m: could not convert string to float: 'product_weight_g'"
         ]
        }
       ]
      }
     },
     "faea30e15df94b9f95fc6b5d878473b9": {
      "model_module": "@jupyter-widgets/base",
      "model_module_version": "1.2.0",
      "model_name": "LayoutModel",
      "state": {}
     },
     "fd4cf243bd7740429df59497af409e75": {
      "model_module": "@jupyter-widgets/controls",
      "model_module_version": "1.5.0",
      "model_name": "DescriptionStyleModel",
      "state": {
       "description_width": ""
      }
     }
    },
    "version_major": 2,
    "version_minor": 0
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}