Final submission ML india.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "\n",
    "from scipy import sparse\n",
    "from scipy.sparse import csr_matrix\n",
    "from datetime import datetime\n",
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "\n",
    "import sqlite3\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import nltk\n",
    "import string\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.feature_extraction.text import TfidfTransformer\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from nltk.corpus import stopwords\n",
    "\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn import metrics\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "from nltk.stem.porter import PorterStemmer\n",
    "\n",
    "import re\n",
    "# Tutorial about Python regular expressions: https://pymotw.com/2/re/\n",
    "import string\n",
    "from nltk.corpus import stopwords\n",
    "from nltk.stem import PorterStemmer\n",
    "from nltk.stem.wordnet import WordNetLemmatizer\n",
    "\n",
    "from gensim.models import Word2Vec\n",
    "from gensim.models import KeyedVectors\n",
    "import pickle\n",
    "\n",
    "from tqdm import tqdm\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [],
   "source": [
    "Total=pd.read_csv(\"train.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(116058, 29)"
      ]
     },
     "execution_count": 121,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Total.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['loan_id', 'source', 'financial_institution', 'interest_rate',\n",
       "       'unpaid_principal_bal', 'loan_term', 'origination_date',\n",
       "       'first_payment_date', 'loan_to_value', 'number_of_borrowers',\n",
       "       'debt_to_income_ratio', 'borrower_credit_score', 'loan_purpose',\n",
       "       'insurance_percent', 'co-borrower_credit_score', 'insurance_type', 'm1',\n",
       "       'm2', 'm3', 'm4', 'm5', 'm6', 'm7', 'm8', 'm9', 'm10', 'm11', 'm12',\n",
       "       'm13'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Total.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No of Nan values in our dataframe :  0\n"
     ]
    }
   ],
   "source": [
    "# just to make sure that all Nan containing rows are deleted..\n",
    "print(\"No of Nan values in our dataframe : \", sum(Total.isnull().any()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 0 duplicate rating entries in the data..\n"
     ]
    }
   ],
   "source": [
    "dup_bool = Total.duplicated()\n",
    "dups = sum(dup_bool) # by considering all columns..( including timestamp)\n",
    "print(\"There are {} duplicate rating entries in the data..\".format(dups))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "X    63858\n",
       "Y    37554\n",
       "Z    14646\n",
       "Name: source, dtype: int64"
      ]
     },
     "execution_count": 132,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Count of each label\n",
    "Total[\"source\"].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OTHER                          49699\n",
       "Browning-Hart                  31852\n",
       "Swanson, Newton and Miller      6874\n",
       "Edwards-Hoffman                 4857\n",
       "Martinez, Duffy and Bird        4715\n",
       "Miller, Mcclure and Allen       3158\n",
       "Nicholson Group                 2116\n",
       "Turner, Baldwin and Rhodes      1846\n",
       "Suarez Inc                      1790\n",
       "Cole, Brooks and Vincent        1642\n",
       "Richards-Walters                1459\n",
       "Taylor, Hunt and Rodriguez      1259\n",
       "Sanchez-Robinson                1193\n",
       "Sanchez, Hays and Wilkerson      853\n",
       "Romero, Woods and Johnson        750\n",
       "Thornton-Davis                   651\n",
       "Anderson-Taylor                  483\n",
       "Richardson Ltd                   473\n",
       "Chapman-Mcmahon                  388\n",
       "Name: financial_institution, dtype: int64"
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Count of each label\n",
    "Total[\"financial_institution\"].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "A23    58462\n",
       "B12    29383\n",
       "C86    28213\n",
       "Name: loan_purpose, dtype: int64"
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Count of each label\n",
    "Total[\"loan_purpose\"].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0    115680\n",
       "1.0       378\n",
       "Name: insurance_type, dtype: int64"
      ]
     },
     "execution_count": 130,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Count of each label\n",
    "Total[\"insurance_type\"].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    115422\n",
       "1       636\n",
       "Name: m13, dtype: int64"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Count of each label\n",
    "Total[\"m13\"].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It's an highly imbalanced data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[<matplotlib.axes._subplots.AxesSubplot object at 0x00000291E728EB00>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E0765B00>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E095EAC8>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E0280278>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E033A8D0>],\n",
       "       [<matplotlib.axes._subplots.AxesSubplot object at 0x00000291E02CCE48>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E090B358>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E01D58D0>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E01D5908>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E1E213C8>],\n",
       "       [<matplotlib.axes._subplots.AxesSubplot object at 0x00000291EA429940>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E02BBEB8>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291EA400470>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291EA0F99E8>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E9FB1F60>],\n",
       "       [<matplotlib.axes._subplots.AxesSubplot object at 0x00000291E9FAB518>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E804BA90>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E72C7048>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E72A05C0>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E7271B38>],\n",
       "       [<matplotlib.axes._subplots.AxesSubplot object at 0x00000291E53C10F0>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E539A668>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E1D50BE0>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E03E4198>,\n",
       "        <matplotlib.axes._subplots.AxesSubplot object at 0x00000291E040B710>]],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 124,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABbEAAARuCAYAAADtWXpuAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzs3X+8VWWd9//XO1BE8xeWJwUmKMnyR5qS0tQ0Z8QUrTvsvrUwR9GYmLq1X+NMgXffGye1sbkzUysbUgLKRGIqmMSItFO3M4q/E3/kzRFJjpI/4oeiqWGf7x/XtWGx2fv83GeffTbv5+OxH2eva11rreva7HWx9rXW9bkUEZiZmZmZmZmZmZmZNaLXDHQBzMzMzMzMzMzMzMyqcSe2mZmZmZmZmZmZmTUsd2KbmZmZmZmZmZmZWcNyJ7aZmZmZmZmZmZmZNSx3YpuZmZmZmZmZmZlZw3IntpmZmZmZmZmZmZk1LHdim5mZmZmZmdWYpAcltQ50OczMesJtlzUqd2I3EElrJB0/0OUY7CSdLenWwvJmSW8ayDKZNSNJbZL+bqDL0QwktUrqKCz7wtGsmyTNlXRxN/I11HWWpAskXTPQ5WgEks6Q9POBLodZrUXEoRHRNtDlsOrKfzuamduuRrtmtG3ciW1NLyJeGxGrofs/dM3MBlLxwlHShZK+P8BFMmtq5TeSushbk3MyIr4cETvdjUBJYySFpKGltIi4LiJOGMhymQ12xXOq2TRz3cx2dj6/rSfcid1kKjUAA9Uo1OK4zdqgKfH5Z4bbrcFE0pCBLoOZNTa3E2bblJ7myze/FkqaL+n5POJqfCHfFyQ9kdc9ImliTt/uAZwKI7fW5G3vB16QNFTSDEmP5n09JOlDhfxnS7pV0lclbZD0mKSTCutHSPqupCfz+p8U1n1A0n2SNkr6L0lv72b9Z+ZybMj73q07+6xSt9GSfiTpGUl/kPSNQv6PSXo4H2eZpDcW1oWkT0haldd/M/8eexvwbeBdefTuxq7/Vc2a387cdkn6HvAXwH/kduHzkm6U9KmyfPdLOiW/D0mflrRa0rOS/o8K/T2dtU/WM+5EazzvrPSfvKSPS2qXtF7SEkkHljbIJ8y5klYBqzpJ+0tJd0ralP/+ZU7/G0krC/v7haQ7Csu3Fk7OAyX9e75weEzSpwv5LpS0SNL3JT0HnF2tkpKGKA2jLTVSd0sa3UnZ3yppea7/I5I+XNjXfvkzeS6X+81lxwpJB0maDpwBfD43Rv/R2T9EJw1yZ2Wv+BnndW2SLpH0n8CLwJsk7S3pWknr8rEuln/8WR2owo8ASa+R9EVJv5P0tNLFyt5d7OrNku7I3/nFkkYUjvFBpQudjfn7/7bCukoXLpXS3pa33Zj39cG8/dic9pq8fI2kpwv7/76kz+b3Vc8zpQui/5R0uaT1wIVdfG4fV7oAKV1cHdVJfTprL4crXdxtkPQQ8M6y45QuHCcBFwAfye3Wb7oo39lKF0/P52Oe0Y2yV/yM87q5kq6WtFTSC8DfSBqmdAH5uKSnJH1b0vDOymVWS5LeIeme/F2+AehWh0i2w3WWpD2Am4AD83m2WYXrrLJjVzwn8/m+ROk6pV3Sx7tRj61PdGvb08lT87n1rKT/Vcjbl2uPi/NnsVnSfyhdN12ndN10p6QxhfxVr7c6qUelduL9ku7Nx1gr6cLCJr/OfzfmMr1LO4aCq1ons0Hsg8ACYB9gCfANAEkHA+cB74yIPYETgTU92O/pwPuBfSJiC/Ao8FfA3sA/A9+XdEAh/7HAI8DrgH8FrpWkvO57wO7AocD+wOW5jEcBc4C/B/YD/g1YImlYN8p3Rq7Tm4G3AF/swT631g0I4KfA74AxwEjS54nSb9ULgP8OvB74v8D1ZeX4AOl66wjgw8CJEfEw8Angtjx6d59u1MdsZ7NTtV0RcSbwOPDfcrvwr8A84G9LeSQdQWqDlhY2/RAwHjgKmAx8LOftTvtk3RURfjXIi3TCPwCMBkYA/wlcDBwHPEs6GYYBVwG/LmwXwPK8zfBKafnvBuBMYCipwdhAOpF3A/5IagyGAr8HngT2zNv+Med7DXA38L+BXYE3AatJFwCQOn/+BJyS8w7vpK7/BKwEDgZEupjYr0rZ9wDWAufk8h2VP49Dc/4FwMKc7zDgCeDWss/noPx+LnBxN/4tDs7HPDAvjwHe3FnZO/uM83ZtpMbw0Lx+F+AnpIZ0D1Jjewfw9wP9XfSruV/AEOA3pP/c98htwHtI/9G253P7tcCPgO91sp+2fL4dlvfz78D387q3AC8A78vf9c/nfe+a168B7iO1d8MrpeXt2kn/6e9KagufBw7O+R8Hjs7vHyG1R28rrHtHfl/1PCPdbNsCfCqfl521W6fl+r4zn/sHAW+sUvau2stLSRcwI/I2DwAdhWOtAY7P7y8sfa5d/LvuATxX+HwOYFs7WbHs3fiM5wKbgHfnOu0GfJ10ATuC9P/EfwD/MtDfa792jlf+nv4O+Fz+/p5Kuva4mHR98DTpR84QYGo+l4blbddQ4Torr2stnoNdlGGHcxL4FfCtfI4cCTwDTOzufkjXGQF8J7chRwAvs61N68u1Rzup82hv4CHg/wHH5/zzge/mvJ1eb3VSj0rtRCtweF5+O/AUcEpZXYcW9nE2+dqtqzr55ddgeuV25/h8vv+ikH4I8Mf8/qDcdh0P7FK2/VwKv13K26q8/491UYb7gMn5/dlAe2Hd7vl8fAPpuuHPwL4V9nE1cFFZ2iPAX3ej/p8oLJ8MPNqdfZbXDXgXqW0dWuE4NwHTCsuvIT009Ma8HMB7CusXAjMKn8mtndXDL792tpfbrm2/xfLyMGA9MC4vfxX4VmF9AJMKy/8TuDm/77R98qtnLz+J3Xi+ERFrI2I9cAnpwv0MYE5E3BMRLwMzSUOexhS2+5eIWB8Rf6yS9n5gVUR8LyK2RMT1wG9Jd5deAu4C3ku6c3Q/cCvpx8iEvN0fSJ0fr4+IL0XEK5HiTH8HmFI45m0R8ZOI+HNZWcr9HfDFiHgkkt/kY1Qq+weANRHx3Vz2e0idZacqPU35P4D/HREvRMQDpLtkffUqqaE6RNIuEbEmIh7touxVP+PCfudGxIOR7jSOAE4CPpvL/jSpU7H4eZr1h2OAA4F/yt+9lyLiVlJb87WIWB0Rm0ltzRR1Hh7jexHxQES8APx/wIfzefkR4MaIWB4RfyL9Rz8cKD5Nd2Vu7/5YJW0CqTP90tzm3EJ6Auf0nPdXwF9LekNeXpSXxwJ7Ab+R1ELX59mTEXFVPm+7arf+NSLuzOd+e0T8rkrZu2ovPwxcktu5tcCVnRy3J/4MHCZpeESsi4gHuyh7V58xwOKI+M+I+DOpU+3jwOdy2Z8HvozbLaufCaTO669HxJ8iYhFwZ173ceDfImJFRLwaEfNI39kJhe0rXWf1SX4i+j3AF3J7eh9wDakTtqf+OSL+GBG/Id1sPCKn9+Xa47sR8WhEbCL9kHo0In6Rr0V+CLwj56t6vdWNcm9tJ/Jn0BYRK/Py/aQnjv66m59Bd+pkNhj9vvD+RWA3SUMjoh34LKmz6GlJC6qNBqlibXFB0lmFESkbSQ8bvK5SOSLixfz2taQbfOsjYkOFY7wROL+0z7zf0aTryZ6U73eFbbqzz+K2o4Hf5barUvmuKOxnPemG38hCnvLP/7XdKLuZ7bxt11aR+uEWAn+rNBL4dNLT39XqU97WddU+WTe5E7vxVPriH5jfA5A7l/7A9l/67RqACmnb7aOw/9I+fkW6O/be/L6N9GPjr/MypJPvwLIG4AKgpYtyVDKaNFykmuJ+3ggcW3bcM0h33V5Pekqn/HPrky4a5Gpl7+ozhh3rtQuwrlCvfyM9KWrWn6r9CCj/Dv+OdH61KIWMKA2zv6CQp/zc24V0sVHebv055+1pu7U2b1s8RqV269ds327937xdd86z/mq3OmsvD6T27dYLpJsHnyDV90ZJb+2i7F19xpSV8/WkJx/uLtTrZzndrB4OBJ6ISI+yZKXzp6cdIr+jhz9iOilT6aZOcd+9+XFSrZOlL9ceTxXe/7HCcukYnV1vdaX8h+ixkn6pFE5pE6ldel3lTXfQnTqZNZWI+EFEvId0HgbwlbzqBdL/uyWVzset7aFSnNXvkIb47xcpPMYDpA6TrqwFRkiqFFJjLenm+z6F1+75JlNXRhfe/wVpxG9391ls69cCf1Hl4Yq1pFF2xX0Nj4j/6kb5oussZlZJE7ddldqFeaTroonAixFxW9n6ztq63rZPVsad2I2n0hf/SVKjAIBS7Mb9SEPDSyqdZMW07fZR2H9pH+Wd2L9ix07stcBjZSffnhFxchflqGQtZbGrOyn7WuBXZcd9bUR8kjSkbAs7fm7d2W+nOmmQq5W9q8+4/PhrSU+Iva5Qr70i4tDultGsl6r9CCj/Dv8F6fx6KiI+kc+710bElwt5ys+9P5GGn5e3W8p5e9pujdb2k6CWt1t/RWq7fsW2ESTl7VZX51l/tVudtZfr6J92a1lEvI80rO63pIvBzsre1WdcfvxnSZ1ehxbqtXdE+Gkmq5d1wMhCDETYdv5054dKtR8YPenEKM/7JOnH055l+36C2unLtUdPjlHteqsr5Z/JD0hhh0ZHxN6kidNUJW+5WtbJrOFJOljScTlG60uk/2dfzavvA05WmrTsDaSHbDqzB+kceybv+xzS04xdioh1pNEa35K0r6RdJL03r/4O8Il8g0qS9lCKfb9n9T1uda6kUUrzplwA3NDLfd5B+j/g0px3N0nvzuu+DcyUdGiu996STutOvUk39kZJ2rWb+c2Mpm+7niKFgywe5zbSqNfL2PEpbIB/yscfDXyGbW1dX9onK+NO7MZT6T/5HwDnSDoyNxBfBlZExJoe7Hcp8BZJH1WacOwjpHhGP83r/4sUZ/EY4I5IQ9DfSIorWZqA5w7gOaXJy4YrTTJ0mKR30nPXABdJGpcbk7dL2q9K3p/msp+ZG6RdJL1T0tsi4lVS3N4LJe0u6RBSDMxqdmiMKumiQa5W9q4+4+3kxvbnwGWS9lKaVO/Nkro73Nast6r9CLge+JzSpImvJbU1N1QZtlnyt5IOkbQ78CVgUT4vFwLvlzRR0i7A+aTO5J7ccV5Buov/+Xzet5KGky8AiIhVpHPzb0nzBDxHOsf/B7kTu8bn2TXAP0o6Op/7B6n6zNJdtZcLSRcz+0oaRYrJXc1TwBht39G8A0ktSpNp7kH6rDezfbtVqeydfsbl8hPb3wEul7R/Pu5ISSd2VjazGrqNdHPt0/n/2v9OunaB7v1QqdaZ8hSwn7qezLaUd+s5GSkk0H8B/5Lb07cD04Dr+lrZgppce3Sh6vVWL/a1J+np9JckHQN8tLDuGdKPwGrXY7Wsk9lgMIw0V8azpNEY+5PaJ0gdJb8hxWf9OdvarIoi4iFSB8ttpLbqcFL8/+46k/RAwm9JsW4/m/d7Fylk0zdIMerbSTFqu+MHueyr8+vi3uwzX1/+N1Ic3seBDtIINCLix6QHjhZIeo70BOdJ3SzfLcCDwO8lPdvNbcysuduufwG+qDQy7R8L6fNz2b5fYZvFpDmR7gNuBK7NZehL+2TlogECc/uVXqQTfCZp0p2NpOEKu+d1nyANI11PuogfVdguyBMXdpH2HtJJtSn/fU/Z+tuAXxaWFwEPl+U5kNTR9XtSI3A7PZx8LOcdQpqZ+jHSJGJ3lupUpewHkxqCZ0ihVG4BjszrXp8/k+dIHUcXUX1ix3GkRmUj8JNOyvf2vK/nC5/5gd0oe9XPmBTq4O/KjrM3abKBjrzNvcCUgf4u+tX8L9JTbT/J59OzpJjMryFNRLg2n2vfp8IEGYV9tJH+g78jn3//QXriubT+Q7k920TqVD60sG4NhckyOkk7NG+7Ke/rQ2Xrryc98Vxa/mo+L4sThlU9z+jhZD6ktvgRUgfxA2ybPLJS2TtrL3cnXQRtzPX6J6pP7Lgf6SnzDcA9nZTtgMJntTH/+xzSjbJX/YypMBkuadK2L5N+iD4HPAx8eqC/037tPC/S/B335nP9hvwqTdA4ifT/8kbSzbofAnvmdWuocp2V188htYkbyf/nVzn+DuckMIp0rbCedL32iW7U40J2nNix2Ha1ka8bqNG1B6nzaG5h+Xi2nyip6vVWJ/Wo1E6cSgoB8nz+XL5B4RqRdNPzmfxZT6CsLe6sTn755dfgeVHh+sgvv/zya7C+gLOo8NuRCn1YfvXPS/kDNzMzMzMzMzOrCUlrSDfSfjHQZTEz64s88vgW4FsRMb9sXQDjIs2tZv3I4UTMzMzM6kjS5yQ9KOkBSdfnEBRjJa2QtErSDcqxOSUNy8vtef2Ywn5m5vRHHNLFzMzqTdJfaNvE3+Wvzub7MDMbMD1tu/J19jOkUCc/qHuBbSt3Ylu/kXRTlUbhgq637n++6DKzcpK+XaVN+PZAl62kk3brrwa6bNY1SSOBTwPjI+IwUpiIKaRYeZdHxDhSmIppeZNpwIaIOAi4POdDaQ6IKaRwMJNIk9kMqWddmlmtrmEa/Vqou/JNl0r1OGOgy2ZmAyciHo9tE3+Xvx6PiDF+CtvMGk1XbVeF/MsiYo+ImBwV5ouKCPkp7PpwOBEzMzOzOsmd2LcDR5Biiv8EuIo0CeAbImKLpHcBF0bEiZKW5fe3SRpKirH+emAGQET8S97v1nx1r5SZmZlZgaQ5wAeAp/NNeyT9H9LknK+Q5o84JyI25nUzSTfuXyXNtbIsp08CriDd9L8mIi7N6WNJE5GPAO4BzoyIVyQNI807czRpboePRMSaulTazPqdn8Q2MzMzq5OIeII0AenjpIkHSxPXbSw82dEBjMzvR5ImWyWv30SaWHBreoVttpI0XdJd+TW99jUyMzMz28Fc0kixouXAYRHxduD/kSZbrjq6LI8w+yZwEnAIcHrOCz0cwWZmzWHoQBeg1l73utfFmDFjusz3wgsvsMcee/R/gfqR6zDwBnv5ofZ1uPvuu5+NiNfXbIeDWHfbI/B3qVEM9joM9vJDbevQiO2RpH2BycBYYCPwQ9KPs3KloXKqsq5a+vYJEbOB2ZDapPHjx/9bpXIN9HdnoI/fCGUY6OM3QhkG+vj9XYZGbJMGys52jVTiujSmnbEu/d0eRcSvi/N45LSfFxZvB07N7ycDCyLiZeAxSe3AMXlde0SsBpC0AJgs6WHgOOCjOc884ELg6ryvC3P6IuAbkhRdhCBopn4kl7E2XMbaqHWb1HSd2GPGjOGuu+7qMl9bWxutra39X6B+5DoMvMFefqh9HST9rmY7G+S62x6Bv0uNYrDXYbCXH2pbhwZtj44HHouIZwAk/Qj4S2AfSUPz09ajgCdz/g5gNNCRw4nsDawvpJcUt6moszZpoL87A338RijDQB+/Ecow0Mfv7zI0aJs0IHa2a6QS16Ux7Yx1aYD26GPADfl9KdRaSXF0Wfmos2NJI9K6NYJNUmkE27OdFaaZ+pFcxtpwGWuj1m1S03Vim5mZmTWwx4EJknYH/ghMBO4Cfkl6ImkBMBVYnPMvycu35fW3RERIWgL8QNLXgAOBccAd9ayImZmZWU9J+l/AFtJ8IFB9dFml8LddjUbr1ki1XI7pwHSAlpYW2traqhc627x5c7fyDSSXsTZcxtqodRl73YktaTfg18CwvJ9FETGrNwH2exrE38zMzGwwiogVkhaRrpG2APeSwn3cCCyQdHFOuzZvci3wvTy0dj0pZiQR8aCkhcBDeT/nRsSrda2MmZmZWQ9Imkqa8HFiIcRHZ6PLKqU/S89HsO2gGHJt/Pjx0Z2nRZvpydeB5DLWxs5Yxr5M7PgycFxEHAEcCUySNIEeBtjvZRB/MzMzs0EpImZFxFsj4rCIODMiXo6I1RFxTEQcFBGn5biQRMRLefmgvH51YT+XRMSbI+LgiLhp4GpkZmZm1rn8kOIXgA9GxIuFVUuAKZKG5YciS6PL7gTGSRoraVdSv9GS3PldGsEGlUewQWEEW3/Wy8zqp9ed2JFszou75FeQAuwvyunzgFPy+8l5mbx+oiRRCOIfEY8BpSD+x5CD+EfEK6Snuyf3trxmZmZmZmZmZta/JF1PCoV2sKQOSdOAbwB7Assl3Sfp25BGlwGl0WU/I48uy09ZnwcsAx4GFua8kDrD/yGPVNuP7Uew7ZfT/wGYUYfqmlmd9Ckmdn5a+m7gINJT04/S8wD7PQ3ib2ZmZmZmZmZmDSgiTq+QfG2FtFL+S4BLKqQvBZZWSF9NevCxPP0l4LQeFdbMBo0+dWLn2ItHStoH+DHwtkrZ8t9qAfZ7GsR/B80akL8rrsPAG+zlh+aog5mZmZmZmZmZNa8+dWKXRMRGSW3ABHoeYL+nQfwrHb8pA/J3xXUYeIO1/GNm3Lj1/fmHv8plt77AmkvfP4AlspVPbOLswr+L/z3MzHYu/n/ArDKfG2bWKNwemQ2sXsfElvT6/AQ2koYDx5PiFPU0wH6Pgvj3trxmZmZmZrbNmBk3bveywU3SHElPS3qgkDZC0nJJq/LffXO6JF0pqV3S/ZKOKmwzNedfJWlqIf1oSSvzNlfm+Y2qHsPMzMyslnrdiQ0cAPxS0v2kDuflEfFTehhgv5dB/M3MzMzMbAC5E7zhzAUmlaXNAG6OiHHAzWyb5Owk0sND40hhGa+G1CENzCLNRXQMMKvQKX11zlvablIXxzAzMzOrmV6HE4mI+4F3VEjvcYD9ngbxNzMzMzMzs20i4teSxpQlTwZa8/t5QBvpoaPJwPw8MvZ2SftIOiDnXR4R6wEkLQcm5dCRe0XEbTl9PnAKcFMnxzAzMzOrmZrExDYzMzMzs/oqf/rZsTmtgpaIWAcQEesk7Z/TRwJrC/k6clpn6R0V0js7xnYkTSc9yU1LS0u3JxZvGQ7nH75l6/JgnpC8mSZUd10aUzPVxcysnDuxzczMzMz6yB3KNsioQlr0Ir3bImI2MBtg/Pjx0d3J0a+6bjGXrdz2s3XNGd3brhEN1knhK3FdGlMz1cXMrFxfYmKbmZmZmZlZ43oqhwkh/306p3cAowv5RgFPdpE+qkJ6Z8cwMzMzqxl3YpuZmZmZmTWnJcDU/H4qsLiQfpaSCcCmHBJkGXCCpH3zhI4nAMvyuuclTZAk4KyyfVU6hpmZmVnNOJyImZmZmZnZICfpetIEi6+T1AHMAi4FFkqaBjwOnJazLwVOBtqBF4FzACJivaSLgDtzvi+VJnkEPgnMBYaTJnS8KadXO4aZmZlZzbgT28zMzMzMbJCLiNOrrJpYIW8A51bZzxxgToX0u4DDKqT/odIxzMzMzGrJ4UTMzMzM6kTSwZLuK7yek/RZSSMkLZe0Kv/dN+eXpCsltUu6X9JRhX1NzflXSZpa/ahmZmZmZmaDmzuxzczMzOokIh6JiCMj4kjgaNIw/h8DM4CbI2IccHNeBjgJGJdf04GrASSNIIUKOBY4BphV6vg2MzMzMzNrNu7ENjMzMxsYE4FHI+J3wGRgXk6fB5yS308G5kdyO7CPpAOAE4HlEbE+IjYAy4FJ9S2+mZmZmZlZfTgmtpmZmdnAmAJcn9+3RMQ6gIhYJ2n/nD4SWFvYpiOnVUvfjqTppCe4aWlpoa2trWJBNm/eXHVdPQz08WtRhvMP37Ldck/31TJ8+310Z/u+HrN8+55+Bn09frlm+B6YmZmZWf9wJ7aZmZlZnUnaFfggMLOrrBXSopP07RMiZgOzAcaPHx+tra0VD9LW1ka1dfUw0MevRRnOnnHjdstrzujZvq66bjGXrdx2ad6d7ft6zPLt507ao0efQV+PX64ZvgdmZmZm1j8cTsTMzMys/k4C7omIp/LyUzlMCPnv0zm9Axhd2G4U8GQn6WZmZmZmZk3HndhmZmZm9Xc620KJACwBpub3U4HFhfSzlEwANuWwI8uAEyTtmyd0PCGnmZmZmZmZNR2HEzEzMzOrI0m7A+8D/r6QfCmwUNI04HHgtJy+FDgZaAdeBM4BiIj1ki4C7sz5vhQR6+tQfDMzMzMzs7pzJ7aZmZlZHUXEi8B+ZWl/ACZWyBvAuVX2MweY0x9lNDMzMzMzayQOJ2JmZmZmZmZmZjUhaY6kpyU9UEgbIWm5pFX57745XZKulNQu6X5JRxW2mZrzr5I0tZB+tKSVeZsrJamzY5hZc3AntpmZmZmZmZmZ1cpcYFJZ2gzg5ogYB9yclyFNdj0uv6YDV0PqkAZmAccCxwCzCp3SV+e8pe0mdXEMM2sC7sQ2MzMzMzMzM7OaiIhfA+VzdUwG5uX384BTCunzI7kd2EfSAcCJwPKIWB8RG4DlwKS8bq+IuC2HXZtftq9KxzCzJuBObDNrKpKGSLpX0k/z8lhJK/KQshsk7ZrTh+Xl9rx+TGEfM3P6I5JOLKRPymntknxX38zMzMzMrHtaImIdQP67f04fCawt5OvIaZ2ld1RI7+wYZtYEej2xo6TRpDtebwD+DMyOiCskXQh8HHgmZ70gIpbmbWYC04BXgU9HxLKcPgm4AhgCXBMRl+b0scACYARwD3BmRLzS2zKb2U7hM8DDwF55+SvA5RGxQNK3SW3Q1fnvhog4SNKUnO8jkg4BpgCHAgcCv5D0lryvbwLvI10o3SlpSUQ8VK+KmZmZmZmZNRlVSItepPfsoNJ0UkgSWlpaaGtr63KbluFw/uFbti53Z5t627x5c0OWq8hlrI2dsYy97sQGtgDnR8Q9kvYE7pa0PK+7PCK+Wszcy46hap1PZmY7kDQKeD9wCfAPeYKP44CP5izzgAtJ7cjk/B5gEfCNnH8ysCAiXgYek9ROisEG0B4Rq/OxFuS87sQ2MzMzMzPr3FOSDoiIdTkkyNM5vQMYXcg3Cngyp7eWpbfl9FEV8nd2jB1ExGxgNsD48eOjtbW1WtatrrpuMZet3NaNtuaMrrept7a2NrpTl4HkMtbGzljGXocTiYh1EXFPfv886cnHkZ1ssrVjKCIeA0odQ8eQO4byU9YLgMmFzqdFeXvHMzKzrnwd+DxpdAjAfsDGiCjdLi8ONds6PC2v35Tz93Q4m5mZmZmZmXVuCTA1v58KLC6kn6VkArAphwJZBpwgad88oeMJwLK87nlJE3K/0Vll+6p0DDNrAn31AZCKAAAgAElEQVR5EnurHEv2HcAK4N3AeZLOAu4iPa29gdTZc3ths2IHUHnH0LF03vlkZrYdSR8Ano6IuyW1lpIrZI0u1lVLr3TTb4dha70ZlgaDY2haVwbDcKauDPY6DPbyQ3PUwczMzGxnJul60lPUr5PUAcwCLgUWSpoGPA6clrMvBU4mPej4InAOQESsl3QRcGfO96WIKE0W+UlgLjAcuCm/6OQYZtYE+tyJLem1wL8Dn42I5yRdDVxE6ty5CLgM+Bg97xjqdpyj3nQaNcOPZNdh4A3W8hc7S0udp4OxHmXeDXxQ0snAbqSY2F8nzW49NN8QKw41Kw1b65A0FNibNIN2teFsdJK+VW+GpcHgGJrWlcEwnKkrg70Og7380Bx1MDMzM9uZRcTpVVZNrJA3gHOr7GcOMKdC+l3AYRXS/1DpGGbWHPrUiS1pF1IH9nUR8SOAiHiqsP47wE/zYk87hp6leufTdnrTadQMP5Jdh4E3WMt/9owbt74///AtXLZy6KDsNC2KiJnATID8JPY/RsQZkn4InEoKVVQ+bG0qcFtef0tEhKQlwA8kfY0Uv38ccAfpxtq4POHsE6QY/6VY22ZmZmZmZmZm1k96HRM7xx66Fng4Ir5WSD+gkO1DwAP5/RJgiqRhuROo1DF0J7ljSNKupI6hJflu3C9JnUvgeEZm1jtfIE3y2E4KU3RtTr8W2C+n/wMwAyAiHgQWkiZs/BlwbkS8mm+mnUeKzfYwsDDnNTMzMzMzMzOzftSXJ7HfDZwJrJR0X067ADhd0pGk0B9rgL+H1DEkqdQxtIXcMQQgqdQxNASYU+gY+gKwQNLFwL1s63wyM6sqItpIM1cTEatJE8iW53mJKjHSIuIS4JIK6UtJMdvMzMzMzMzMzKxOet2JHRG3UjluddUOnp52DFXrfDIzMzMzMzMzMzOznUOvw4mYmZmZWc9J2kfSIkm/lfSwpHdJGiFpuaRV+e++Oa8kXSmpXdL9ko4q7Gdqzr9K0tSBq5GZmZmZmVn/cie2mZmZWX1dAfwsIt4KHEGKsz8DuDkixgE352WAk0jziIwDpgNXA0gaAcwCjiWNWptV6vi23ln5xCbGzLhx68usmUj6nKQHJT0g6XpJu+U5iVbkG2E35PmJyHMY3ZBvnq2QNKawn5k5/RFJJxbSJ+W0dkkzdiyBmZmZWd+4E9vMzMysTiTtBbyXPM9HRLwSERuBycC8nG0ecEp+PxmYH8ntwD55Eu0TgeURsT4iNgDLgUl1rIqZDRKSRgKfBsZHxGGkeYimAF8BLs83zzYA0/Im04ANEXEQcHnOh6RD8naHktqbb0kaImkI8E3STbdDSHMkHVKv+pmZmdnOoS8TO5qZmZlZz7wJeAb4rqQjgLuBzwAtEbEOICLWSdo/5x8JrC1s35HTqqVvR9J00hPctLS00NbWVrFQmzdvrrquHgb6+AAtw+H8w7dsXe5peYrb9mb73hy/r8cs376n/w59PX65RvgeNEIZ+slQYLikPwG7A+uA44CP5vXzgAtJoz0m5/cAi4BvSFJOXxARLwOPSWpn2/xF7Xk+IyQtyHkf6uc6mZmZ2U7EndhmZmZm9TMUOAr4VESskHQF20KHVFJpEu3oJH37hIjZwGyA8ePHR2tra8WDtLW1UW1dPQz08QGuum4xl63cdmm85ozWHm1/dlkIkp5u35vj9/WY5dvPnbRHj/4d+nr8co3wPWiEMtRaRDwh6avA48AfgZ+TbqBtjIjSnYjijbCtN8kiYoukTcB+Of32wq6L25TfVDu2H6piZmZmOzF3YpuZmZnVTwfQEREr8vIiUif2U5IOyE9hHwA8Xcg/urD9KODJnN5alt7Wj+U2s0Eqx8ufDIwFNgI/JIX+KFe6EdbTm2eVQlTucFOtuyNDyvV1lEQjaaYn/V2XxtRMdTEzK+dObDMzM7M6iYjfS1or6eCIeASYSBpy/xAwFbg0/12cN1kCnJeH5x8LbMod3cuALxcmczwBmFnPupjZoHE88FhEPAMg6UfAX5Ji7A/NT2OXbpDBtptnHZKGAnsD66l+U41O0rfq7siQcn0dJdFImulJf9elMTVTXczMyrkT22wnMKZsuK+ZmQ2oTwHXSdoVWA2cQ3qScaGkaaQh/6flvEuBk4F24MWcl4hYL+ki4M6c70sRsb5+VTCzQeRxYIKk3UnhRCYCdwG/BE4FFrDjzbOpwG15/S0REZKWAD+Q9DXgQGAccAfpCe1xksYCT5AmfyzF2jYzMzOrCXdim5mZmdVRRNwHjK+wamKFvAGcW2U/c4A5tS2dmTWbHH9/EXAPsAW4l/RE9I3AAkkX57Rr8ybXAt/LEzeuJ3VKExEPSlpIGjmyBTg3Il4FkHQesAwYAsyJiAfrVT8zMzPbObgT28zMzMzMrIlFxCxgVlnyauCYCnlfYttokPJ1lwCXVEhfSho5YmZmZtYvKk3CYWZmZmZmZmZmZmbWENyJbWZmZmZmZmZmZmYNy53YZmZmZmZmZmZmZtaw3IltZmZmZmZmZmZmZg3LndhmZmZmZmZmZmZm1rDciW1mZmZmZmZmZmZmDcud2GZmZmZmZmZm1u8kfU7Sg5IekHS9pN0kjZW0QtIqSTdI2jXnHZaX2/P6MYX9zMzpj0g6sZA+Kae1S5pR/xqaWX9xJ7aZmZmZmZmZmfUrSSOBTwPjI+IwYAgwBfgKcHlEjAM2ANPyJtOADRFxEHB5zoekQ/J2hwKTgG9JGiJpCPBN4CTgEOD0nNfMmoA7sc3MzMzMzMzMrB6GAsMlDQV2B9YBxwGL8vp5wCn5/eS8TF4/UZJy+oKIeDkiHgPagWPyqz0iVkfEK8CCnNfMmsDQ3m4oaTQwH3gD8GdgdkRcIWkEcAMwBlgDfDgiNuSG5grgZOBF4OyIuCfvayrwxbzriyNiXk4/GpgLDAeWAp+JiOhtmc3MzMzMzMzMrP4i4glJXwUeB/4I/By4G9gYEVtytg5gZH4/Elibt90iaROwX06/vbDr4jZry9KPrVQWSdOB6QAtLS20tbV1Wf6W4XD+4Vu2Lndnm3rbvHlzQ5aryGWsjZ2xjL3uxAa2AOdHxD2S9gTulrQcOBu4OSIuzfGHZgBfIA3nGJdfxwJXA8fmTu9ZwHgg8n6WRMSGnGc6qXFaShomclMfymxmZmZmZmZmZnUmaV/Sk9FjgY3AD0l9ReVKDy+qyrpq6ZWiDVR8EDIiZgOzAcaPHx+tra2dFR2Aq65bzGUrt3WjrTmj623qra2tje7UZSC5jLWxM5ax1+FEImJd6UnqiHgeeJh056s43KN8GMj8SG4H9pF0AHAisDwi1ueO6+XApLxur4i4LT99Pb+wLzMzMzMzMzMzGzyOBx6LiGci4k/Aj4C/JPUPlXqHRwFP5vcdwGiAvH5vYH0xvWybaulm1gRqEhM7zxD7DmAF0BIR6yB1dAP752xbh4FkpeEenaV3VEg3MzMzG7QkrZG0UtJ9ku7KaSMkLZe0Kv/dN6dL0pWS2iXdL+mown6m5vyrcmg2MzMzs0b2ODBB0u455OxE4CHgl8CpOc9UYHF+vyQvk9ffkh9yXAJMkTRM0ljSiP87gDuBcZLGStqVNPnjkjrUy8zqoC/hRACQ9Frg34HPRsRzqR2qnLVCWmfDQKqlVypDj2MZDYbYMV1xHQbeYCl/MW5XuVJcr8FQDzOzJvI3EfFsYXkGtQvHZmZmZtZwImKFpEXAPaQQtfeSQnrcCCyQdHFOuzZvci3wPUntpCewp+T9PChpIakDfAtwbkS8CiDpPGAZMASYExEP1qt+Zta/+tSJLWkXUgf2dRHxo5z8lKQDImJdDgnydE7vbLhHa1l6W04fVSH/DnoTy2gwxI7piusw8AZL+c+ecWPVdecfvoXLVg5tyHheZmY7kclsux6aR7oW+gKFcGzA7ZJK4dhayeHYAPK8JJOA6+tbbDMzM7Pui4hZpBvxRauBYyrkfQk4rcp+LgEuqZC+lDSnmpk1mV53YuehH9cCD0fE1wqrSsM9LmXHYSDnSVpAepJoU+7oXgZ8uTRsFjgBmBkR6yU9L2kCKUzJWcBVvS2vmZmZWYMI4OeSAvi3fDN+u3Bsknobjm073R2tNtAjiwb6+LBtZFJJT8tTPuqpp9v35vh9PWb59j39d+jr8cs1wvegEcpgZmZmZjvqy5PY7wbOBFZKui+nXUDqvF4oaRop3lHprtlS4GSgHXgROAcgd1ZfRIpdBPCl0lNFwCeBucBw4Kb8MjPbgaTdgF8Dw0ht26KImJVjpC0ARpCGrZ0ZEa9IGkaaMPZo4A/ARyJiTd7XTGAa8Crw6YhYltMnAVeQhqZdExGX1rGKZtY83h0RT+aO6uWSfttJ3j6FXevuaLWBHlk00McHuOq6xVy2ctulcU9HKJWPeurp9r05fl+PWb793El79Ojfoa/HL9cI34NGKIOZmZmZ7ajXndgRcSuVf0BBCs5fnj+Ac6vsaw4wp0L6XcBhvS2jme1UXgaOi4jNOdTRrZJuAv4BuDwiFkj6Nqlz+ur8d0NEHCRpCvAV4COSDiHFWjsUOBD4haS35GN8E3gf6YnHO3P82YfqWUkzG/wi4sn892lJPyYNn61VODYzMzMzM7Om85qBLoCZWS1Esjkv7pJfARwHLMrp84BT8vvJeZm8fmIOkzQZWBARL0fEY6TRI8fkV3tErI6IV0hPd0/u52qZWZORtIekPUvvSWHUHmBbODbYMRzbWUomkMOxkSYsOkHSvjkk2wk5zczMzMzMrOn0aWJHM7NGImkIcDdwEOmp6UeBjRFRCtpZjBm7NZ5sRGyRtAnYL6ffXthtcZvy+LPHVihDt+LPlutrLNZG0AxxRAd7HQZ7+aE56tCFFuDH6Z4ZQ4EfRMTPJN1J7cKxmZmZmZmZNRV3YptZ04iIV4EjJe0D/Bh4W6Vs+W9P48xWGrnS6/iz5foai7URNEMc0cFeh8FefmiOOnQmIlYDR1RI/wM1CsdmZmZmZmbWbBxOxMyaTkRsJMWGnQDsI6nUO1yKJQuFOLN5/d7AejqPP1sp3czMzMzMzMzM+pE7sc2sKUh6fX4CG0nDgeOBh4FfAqfmbOVxZkvxZ08FbslPPC4BpkgaJmksMA64gzRkf5yksZJ2JU3+uKT/a2ZmZmZmZmZmtnNzOBEzaxYHAPNyXOzXAAsj4qeSHgIWSLoYuBe4Nue/FviepHbSE9hTACLiQUkLgYeALcC5OUwJks4jTZw2BJgTEQ/Wr3pmZmZmZmZmZjsnd2KbWVOIiPuBd1RIXw0cUyH9JbZNnFa+7hLgkgrpS0mTrJmZmZkNGnm02jXAYaQ5PT4GPALcAIwB1gAfjogNSjPPXkGaVPZF4OyIuCfvZyrwxbzbiyNiXk4/GpgLDCddK30mj3AzMzMzqwmHEzEzMzMzM2tuVwA/i4i3kiaXfRiYAdwcEeOAm/MywEmkcGrjgOnA1QCSRgCzgGNJDwjMkrRv3ubqnLe03aQ61MnMzMx2Iu7ENjMzMzMza1KS9gLeSw6pFhGv5EmwJwPzcrZ5wCn5/WRgfiS3kybJPgA4EVgeEesjYgOwHJiU1+0VEbflp6/nF/ZlZmZmVhMOJ2JmZmZmZta83gQ8A3xX0hHA3cBngJaIWAcQEesk7Z/zjwTWFrbvyGmdpXdUSN+OpOmkp7VpaWmhra2tW4VvGQ7nH75l63J3t2tEmzdvHtTlL3JdGlMz1cXMrJw7sc3MzMzMzJrXUOAo4FMRsULSFWwLHVKJKqRFL9K3T4iYDcwGGD9+fLS2tnZR7OSq6xZz2cptP1vXnNG97RpRW1sb3a13o3NdGlMz1cXMrJzDiZiZmZmZmTWvDqAjIlbk5UWkTu2ncigQ8t+nC/lHF7YfBTzZRfqoCulmZmZmNeNObDMzMzMzsyYVEb8H1ko6OCdNBB4ClgBTc9pUYHF+vwQ4S8kEYFMOO7IMOEHSvnlCxxOAZXnd85ImSBJwVmFfZmZmZjXhcCJmZmZmZmbN7VPAdZJ2BVYD55AeaFooaRrwOHBazrsUOBloB17MeYmI9ZIuAu7M+b4UEevz+08Cc4HhwE35ZWZmZlYz7sQ2MzMzMzNrYhFxHzC+wqqJFfIGcG6V/cwB5lRIvws4rI/FNDMzM6vKndhmBsCYGTdut7zm0vcPUEnMzMzMzMzMzMy2cUxsMzMzszqSNETSvZJ+mpfHSlohaZWkG/JwfyQNy8vtef2Ywj5m5vRHJJ04MDUxMzMzMzOrD3dim5mZmdXXZ4CHC8tfAS6PiHHABmBaTp8GbIiIg4DLcz4kHQJMAQ4FJgHfkjSkTmU3MzMz6zVJ+0haJOm3kh6W9C5JIyQtzzf0l+fJY8kTzF6Zb9zfL+mown6m5vyrJE0tpB8taWXe5so84ayZNQF3YpuZmZnViaRRwPuBa/KygOOARTnLPOCU/H5yXiavn5jzTwYWRMTLEfEYafK1Y+pTAzMzM7M+uQL4WUS8FTiCdGN/BnBzvqF/c14GOAkYl1/TgasBJI0AZgHHkq6BZpU6vnOe6YXtJtWhTmZWB32KiS1pDvAB4OmIOCynXQh8HHgmZ7sgIpbmdTNJTxW9Cnw6Ipbl9EmkhmwIcE1EXJrTxwILgBHAPcCZEfFKX8psZmZmNoC+Dnwe2DMv7wdsjIgtebkDGJnfjwTWAkTEFkmbcv6RwO2FfRa32Y6k6aQfcrS0tNDW1laxUJs3b666rh4G+vgALcPh/MO3bF3uaXmK2/Zm+94cv6/HLN++q3+HlU9sKtt++/V9/TdshO9BI5TBzKxZSdoLeC9wNkDu33lF0mSgNWebB7QBXyDduJ+fJ5y9PT/FfUDOuzwi1uf9LgcmSWoD9oqI23L6fNLDATfVoXpm1s/6OrHjXOAbwPyy9Msj4qvFhLKhrwcCv5D0lrz6m8D7SD/C7pS0JCIeYtvw2gWSvk3qAL+6j2U2MzMzqztJpRv/d0tqLSVXyBpdrOtsm+0TI2YDswHGjx8fra2tlbLR1tZGtXX1MNDHB7jqusVctnLbpfGaM1p7tP3Z5RMk93D73hy/r8cs337upD22+3con/S5q58OPT1+uUb4HjRCGczMmtibSA88flfSEcDdpDBrLRGxDiAi1knaP+ffekM/K9247yy9o0L6Drp7o7+orze862Ew3Ix1GWtjZyxjnzqxI+LXxUmGurB16CvwmKTi0Nf2iFgNIGkBMFnSw6ThtR/NeeYBF+JObDMzMxuc3g18UNLJwG7AXqQns/eRNDQ/jT0KeDLn7wBGAx2ShgJ7A+sL6SXFbczMzMwa1VDgKOBTEbFC0hVsCx1SSU9v6Nf8Rn9RX29418NguBnrMtbGzljG/oqJfV4Ouj+nEJeop3fQOhtea2ZmZjaoRMTMiBgVEWNIo9NuiYgzgF8Cp+ZsU4HF+f2SvExef0seTrsEmCJpWA69Ng64o07VMDMzM+utDqAjIlbk5UWkTu2ncpgQ8t+nC/kr3bjvLH1UhXQzawJ9DSdSydXARaS7XRcBlwEfo/odsUod6T26g9abYSCD4bH7rrgOA2+wlL885mVR+ZCoksFQLzOzJvEFYIGki4F7gWtz+rXA9/LotfWkjm8i4kFJC4GHgC3AuRHxav2LbWZmZtZ9EfF7SWslHRwRjwATSdczD5Fu3F/Kjjf0z8sj9o8FNuVwI8uALxcemjwBmBkR6yU9L2kCsAI4C7iqbhU0s35V807siHiq9F7Sd4Cf5sXOhr5WSn+W6sNry4/Z42Egg+Gx+664DgNvsJS/POZl0fmHb9luSFRJIw6NMjNrFhHRRpq0iBxS7ZgKeV4CTquy/SXAJf1XQqu1HeNLb698ksSBsPKJTZ1eM5iZmdXAp4DrJO0KrAbOIT3cuFDSNOBxtl3/LAVOBtqBF3Necmf1RcCdOd+XSpM8Ap8kzd82nDShoyd1NGsSNe/ElnRAKSA/8CHggfx+CfADSV8jTexYGvoqYFweDvsE6Smjj0ZESCoNr13A9nfjzKwTXf1QNjMzs/5V6/+Le7O/8m3WXPr+WhXHzMysVyLiPmB8hVUTK+QN4Nwq+5kDzKmQfhdwWB+LaWYNqE+d2JKuB1qB10nqAGYBrZKOJIX+WAP8PXQ+9FXSecAyYAgwJyIezIeoNrzWzMzMzKxp1aLTuhb7NDMzMzNrBH3qxI6I0yskV+1orjb0NSKWkoaJlKdXHF5rZmZmZlbiJ47NzMzMzJpbpUkVzczMzMzMzMzMzMwagjuxzczMzMzMzMzMzKxh1XxiRzMzMzOzZuf40mZmZmZm9eMnsc3MzMzMzMzMzMysYflJbDMzMzOzMp4s0szMzMyscbgT28yagqTRwHzgDcCfgdkRcYWkEcANwBhgDfDhiNggScAVwMnAi8DZEXFP3tdU4It51xdHxLycfjQwFxgOLAU+ExFRlwqamdmAcvgQMzMzM7OB43AiZtYstgDnR8TbgAnAuZIOAWYAN0fEOODmvAxwEjAuv6YDVwPkTu9ZwLHAMcAsSfvmba7OeUvbTapDvczMzMzMzMzMdmruxDazphAR60pPUkfE88DDwEhgMjAvZ5sHnJLfTwbmR3I7sI+kA4ATgeURsT4iNgDLgUl53V4RcVt++np+YV9mZmZmZmZmZtZPHE7EzJqOpDHAO4AVQEtErIPU0S1p/5xtJLC2sFlHTussvaNCevmxp5Oe1qalpYW2trZulbllOJx/+Jaty93drpFs3rx5UJa7aLDXYbCXH5qjDmZmZmZmZlZb7sQ2s6Yi6bXAvwOfjYjnUujrylkrpEUv0rdPiJgNzAYYP358tLa2dqPUcNV1i7ls5bYmec0Z3duukbS1tdHd+jaqwV6HwV5+aI46mJmZmZmZWW05nIiZNQ1Ju5A6sK+LiB/l5KdyKBDy36dzegcwurD5KODJLtJHVUg3M+s2SbtJukPSbyQ9KOmfc/pYSSskrZJ0g6Rdc/qwvNye148p7GtmTn9E0okDUyMzGywkDZF0r6Sf5uWatTuSJuW0dkkzyo9tZmZm1lfuxDazpqD0yPW1wMMR8bXCqiXA1Px+KrC4kH6WkgnAphx2ZBlwgqR984SOJwDL8rrnJU3IxzqrsC8z64MxM27c+lr5xKaBLk5/exk4LiKOAI4kxdyfAHwFuDxPQrsBmJbzTwM2RMRBwOU5H3ni2inAoaRJZr8laUhda2Jmg81nSHOGlNSk3cltzzdJk2YfApye85qZmZnVjDuxzaxZvBs4EzhO0n35dTJwKfA+SauA9+VlgKXAaqAd+A7wPwEiYj1wEXBnfn0ppwF8Ergmb/MocFM9KmZmzSNPJrs5L+6SXwEcByzK6eWT0JYmp10ETMw30iYDCyLi5Yh4jNQuHVOHKpjZICRpFPB+0nVM6eZ/rdqdY4D2iFgdEa8AC3JeMzMzs5pxTGwzawoRcSuV41YDTKyQP4Bzq+xrDjCnQvpdwGF9KKaZGfmpxbuBg0hPLz4KbIyI0uyuxYljt042GxFbJG0C9svptxd226fJZgd6Qs2+Hr84MS70bnLc8gl2622gj1+LMvT1OzTQ38NGKUM/+TrweWDPvLwftW13yifFPrbWFTAzM7OdmzuxzczMzOooIl4FjpS0D/Bj4G2VsuW/dZlsdqAn1Ozr8c+eceN2y72ZHLd8gt16O//wLQN6/FqUoa+TEg/097BRylBrkj4APB0Rd0tqLSVXyNrbdqfS6N4d2qPu3lQrV35zZTDfZGimmySuS2NqprqYmZVzJ7aZmZnZAIiIjZLagAnAPpKG5qciixPHliab7ZA0FNgbWE/1SWjNzMq9G/hgDrO2G7AX6cnsWrY7XbZH3b2pVq78BlNfb5YMpGa6SeK6NKZmqouZWTnHxDYzMzOrE0mvz09gI2k4cDxporVfAqfmbOWT0JYmpz0VuCWHQ1oCTJE0TNJYYBxwR31qYWaDSUTMjIhRETGGNDHjLRFxBrVrd+4ExkkaK2nXfIwl/z97dx8vV1keev93CagRRIJoRIIGj6ktQlVIIaf0tLsiGNA29DxCQQqJcppWoWKbnhLUp1CBnnhOwYIvtBEiwSIv9eWQSjSmlH187CHIi0jA1BIxhUBK0AQk4kuD1/PHujeZ7MzOfsm8rJn9+34+89kz97pnrWvtmX3tNdesdd8d2DVJkjSJeCa2JElS5xwILCvjYj8PuCkzvxQR3wZuiIiLgW8CV5f+VwOfiYh1VGdCngqQmQ9ExE3At4FtwNllmBJJGqvzaFHeiYhzgJXAHsDSzHygo3siSZL6nkVsSZKkDsnM+4A3NWl/CDiqSftPgJNHWNclwCWtjlFS/8rMQWCw3G9Z3snMFcCKFoYqqY+VL/PvAh7NzLeXqztuAPYH7gHOyMyfRcQLgGuBI4EfAL+bmevLOs4HzgKeBd6XmStL+xzgcqov1a7KzMUd3TlJbbNbw4lExNKI2BQR9ze07R8RqyLiwfJzammPiLgiItZFxH0RcUTDc+aV/g9GxLyG9iMjYk15zhUR0WwyEUmSJGnMZiy6ZaebJEnqmHOphlMb8hHgo5k5E9hCVZym/NySma8FPlr6ERGHUl0l8npgDvDJiNijFMc/AZwAHAqcVvpK6gO7Oyb2NVQJo9Ei4NaSfG4tj6FKIjPLbQFwJVRFb+AC4GiqMwEuGCp8lz4LGp43fFuSJEmSJEnqARExHXgbcFV5HMCbgc+VLsuAk8r9ueUxZfmxpf9c4IbM/Glmfg9YR1VPOgpYl5kPZebPqM7untv+vZLUCbs1nEhmfi0iZgxrngsMlPvLqC5XO6+0X1smBVkdEftFxIGl76rM3AwQEauAORExCOybmbeX9mupEtmXdydmSZIkSZIkdcVfA38GvLg8finwZGZuK483AAeV+wcBjwBk5raIeKr0PwhY3bDOxuc8Mqz96GZBRMQCqpMmmTZtGoODg6MGPm0KLDx823OPx/KcTtu6dWst42pkjK0xGWNsx5jY0y2cmH8AACAASURBVDJzI0BmboyIl5f255JPMZRkdtW+oUm7JEmSJEmSekhEvB3YlJl3R8TAUHOTrjnKspHam402kE3ayMwlwBKAWbNm5cDAQLNuO/jYdTdz6ZrtZbT1p4/+nE4bHBxkLPvSTcbYGpMxxk5O7Dje5LOrRLbjiifwDVovfGMxGveh++oaf+O3w6MZ/m3ykDrulyRJkiSpZx0D/HZEnAi8ENiX6szs/SJiz3I29nTgsdJ/A3AwsCEi9gReAmxuaB/S+JyR2iX1uHYUsR+PiAPLWdgHAptK+0hJZgPbhx8Zah8s7dOb9N/JRL5B64VvLEbjPnRfXeOfP44JqhYevm2Hb5OH1PFbZUmSJsKJGyVJ6r7MPB84H6Ccif2nmXl6RPw98A6qMaznATeXpywvj28vy/8pMzMilgOfjYjLgFdSzaH2DaqTIWdGxCHAo1STP76zQ7snqc12d2LHZoaSDOycfM6MymzgqTLsyErg+IiYWiZ0PB5YWZY9HRGzy8D9ZzasS5IkSZIkSb3vPOBPImId1ZjXV5f2q4GXlvY/ARYBZOYDwE3At4GvAGdn5rPlTO5zqOpMa4GbSl9JfWC3zsSOiOupzqI+ICI2ABcAi4GbIuIs4GHg5NJ9BXAi1ayxzwDvAsjMzRFxEXBn6ffhoUkegfcA1wBTqCZ0dFJHqQnPMJMkSZIk9YrMHKS6Cp/MfAg4qkmfn7C9pjR82SXAJU3aV1DVnyT1md0qYmfmaSMsOrZJ3wTOHmE9S4GlTdrvAg7bnRglSZIkSZIkSb2rkxM7SpIkjWr41SXrF7+tS5FIkiRJkuqgHWNiS5IkSZIkSZLUEp6JLUmSpL7iXBGSJElSf/FMbEmSJEmSJElSbVnEliRJ6pCIODgibouItRHxQEScW9r3j4hVEfFg+Tm1tEdEXBER6yLivog4omFd80r/ByNiXrf2SZIkSZLazSK2JElS52wDFmbmLwGzgbMj4lBgEXBrZs4Ebi2PAU4AZpbbAuBKqIrewAXA0cBRwAVDhW9JkiRJ6jcWsSVJkjokMzdm5j3l/tPAWuAgYC6wrHRbBpxU7s8Frs3KamC/iDgQeCuwKjM3Z+YWYBUwp4O7IkmSJEkd48SOkiRJXRARM4A3AXcA0zJzI1SF7oh4eel2EPBIw9M2lLaR2odvYwHVGdxMmzaNwcHBprFs3bp1xGWdsLvbX3j4tt2OYdqU1qynV7ffihh29z3U7fdhXWKQJEnSzixiS5IkdVhE7AN8Hnh/Zv4wIkbs2qQtd9G+Y0PmEmAJwKxZs3JgYKDpRgYHBxlpWSfs7vbnL7plt2NYePg2Ll3TvUPjbm+/FTGsP31gt7bf7fdhXWKQJEnSzixiS5KkjprRgoJjL4uIvagK2Ndl5hdK8+MRcWA5C/tAYFNp3wAc3PD06cBjpX1gWPtgO+OWJEmSpG6xiC31oMleAJKkXhXVKddXA2sz87KGRcuBecDi8vPmhvZzIuIGqkkcnyqF7pXAXzZM5ng8cH4n9kGSJEmSOs0itiRJUuccA5wBrImIe0vbB6iK1zdFxFnAw8DJZdkK4ERgHfAM8C6AzNwcERcBd5Z+H87MzZ3ZBUmSJEnqLIvYkvpCRCwF3g5syszDStv+wI3ADGA9cEpmbilnQl5OVRh6BpifmfeU58wDPlRWe3FmLivtRwLXAFOoikrnZuZO489K0q5k5tdpPp41wLFN+idw9gjrWgosbV10kiRJklRPz+t2AJLUItcAc4a1LQJuzcyZwK3lMcAJwMxyWwBcCc8VvS+gumT/KOCChkv1ryx9h543fFuSJEmSJElqA4vYkvpCZn4NGH4p/VxgWbm/DDipof3arKwG9isTqb0VWJWZmzNzC7AKmFOW7ZuZt5ezIq9tWJckSZIkSZLayCK2pH42LTM3ApSfLy/tBwGPNPTbUNp21b6hSbskSVKtRcTBEXFbRKyNiAci4tzSvn9ErIqIB8vPqaU9IuKKiFgXEfdFxBEN65pX+j9YhmAbaj8yItaU51xRhm6TJElqGcfEljQZNftglRNo33nFEQuohh1h2rRpDA4OjimgaVNg4eHbnns81ufVydatW3sy7ka9vg+9En/je3244X8L0Jt/D5JUI9uAhZl5T0S8GLg7IlYB86mGXVscEYuohl07jx2HXTuaaki1oxuGXZtFdRx0d0QsL1evDQ27tppq7pA5wJc7uI+SJKnPWcSW1M8ej4gDM3NjGRJkU2nfABzc0G868FhpHxjWPljapzfpv5PMXAIsAZg1a1YODAw067aTj113M5eu2Z6S158+tufVyeDgIGPd37rq9X3olfjnL7plxGULD9+2w98C9ObfgyTVRbkabejKtKcjYi3VFWVz2X7cs4zqmOc8GoZdA1ZHxNCwawOUYdcASiF8TkQMUoZdK+1Dw65ZxJYkSS1jEVtSP1sOzAMWl583N7SfExE3UJ1h9FQpdK8E/rJhMsfjgfMzc3NEPB0Rs4E7gDOBj3VyRyRJknZXRMwA3kR1PLPDsGsR0dZh1ybz1WpDeuWKqbFwX+qpn/ZFkoZrWxE7ItYDTwPPAtsyc1a5BO1GYAawHjglM7eUMdMuB04EngHmZ+Y9ZT3zgA+V1V6cmcuQJpkZuzhrUZWIuJ7qDKEDImID1eWui4GbIuIs4GHg5NJ9BVW+WUeVc94FUIrVFwF3ln4fHjrbCHgPcA0wherMIs8ukiRJPSMi9gE+D7w/M3+4i2Gr2zLs2mS+Wm1Ir1wxNRbuSz3VfV8i4mDgWuAVwM+BJZl5eStrRRFxJNs/t60Azi1Xlkjqce0+E/s3M/P7DY8X0bpx1yTpOZl52giLjm3SN4GzR1jPUmBpk/a7gMN2J0ZJkqRuiIi9qArY12XmF0pzx4ddkzTpOUa/pAnr9HAiLRl3Dbi+s2FLkiSpLrxCSRq7cibj1cDazLysYZHDrknqKMfol7Q72lnETuCrEZHA35bLx1o17pokSZKkHtf4hcTCw7ftcJqvWuYY4AxgTUTcW9o+gMOuSeqibo7RX7Y/7nH6e2GM/l4YF90YW2MyxtjOIvYxmflYST6rIuJfdtF3t8ZXm0jy6YUXezTuQ/d1Kv7Gf5StNvwf8ZBefl0kSb1lzaNPMb+hmLl+8dt2WO6Z19LEZebXaf65Chx2TVIXdHuMfpjYOP29MEZ/3cdFB2NslckYY9uK2Jn5WPm5KSK+CBxF68ZdG76tcSefXnixR+M+dF+n4p/fxg/vCw/ftsM/4iF1/IcsqTdZgJQkSRI4Rr+kiXteO1YaEXuXQfqJiL2pxku7n+3jrsHO466dGZXZlHHXgJXA8RExtYy9dnxpkyRJkiRJUo8Ywxj9sBu1orLs6YiYXbZ1ZsO6JPW4thSxgWnA1yPiW8A3gFsy8ytU464dFxEPAseVx1CNu/YQ1bhrnwLeC9W4a8DQuGt3suO4a5IkST0lIpZGxKaIuL+hbf+IWBURD5afU0t7RMQVEbEuIu6LiCManjOv9H8wIuY125YkSVLNDI3R/+aIuLfcTqS1taL3AFeV53wXx+iX+kZbhhPJzIeANzRp/wEtGndN6mdeei9Jfesa4OPAtQ1ti4BbM3NxRCwqj88DTgBmltvRwJXA0RGxP3ABMItqnMe7I2J5Zm7p2F5IkiSNk2P0S9od7ToTW5IkScNk5teA4VeVzQWWlfvLgJMa2q/NympgvzJO5FuBVZm5uRSuVwFz2h+9JEmSJHVH2yZ2lCRJ0phMK2M4UiY0enlpPwh4pKHfhtI2UvtOImIBsABg2rRpDA4ONg1g69atIy7rhGlTqomGhwyPpXFZp2LotG5vvxUxTOQ91Li9aVMmto5W6vbfgiRJkpqziC1JklRPzS63zV2079yYuQRYAjBr1qwcGBhouqHBwUFGWtYJH7vuZi5d03BYuuZHw3q0/5B14eHbdoyhw7q9/VbEsP70gXE/Z37DEGoLD9/GKV18H0L3/xYkSZLUnMOJSJIkddfjZZgQys9NpX0DcHBDv+nAY7tolyRJkqS+ZBFbkiSpu5YD88r9ecDNDe1nRmU28FQZdmQlcHxETI2IqcDxpU2SJEmS+pLDiUiSJHVIRFwPDAAHRMQG4AJgMXBTRJwFPAycXLqvAE4E1gHPAO8CyMzNEXERcGfp9+HMHD5ZpNRxMxqGBhnJ+sVv60AkkiRJ6jcWsSVJUkuNpZA1WWXmaSMsOrZJ3wTOHmE9S4GlLQxN6gjzgyRJkibC4UQkSZIkSZIkSbXlmdhSDXhWkiRJkiRJktScZ2JLkiRJkiRJkmrLIrYkSZIkSZIkqbYcTkSSJO0Wh0SSJEmSJLWTZ2JLkiRJkiRJkmrLIrYkSZIkSZIkqbYcTkSSJI2ZQ4dIkiRJkjrNIrbUBRaBJEmSJKm5Zp+X1i9+W1u2sfDwbcxfdEvL1y9Jai2L2JIkSeqo4cWJhYd3KRBJkiRJPcEitiRJGpFXjkiSJEmSus0ittQBFoEkSZOZ/wclSZIk7Q6L2JIk6TkWGzVevmckqbWG51XHapYkqQeK2BExB7gc2AO4KjMXdzkkaVR+oO9f5iSp8/ww35z5SFKdmJMk1YX5SOpPtS5iR8QewCeA44ANwJ0RsTwzv93dyCRNRuYk9SO/dOtN5iNJdWJOklQX5iOpf9W6iA0cBazLzIcAIuIGYC5g8pHazDMfmzInqedYpO5bbctHvmckTUDfHCN5DKxW8v3UFX2Tj6ReMzznXTNn75auPzKzpStspYh4BzAnM/9beXwGcHRmnjOs3wJgQXn4OuA7Y1j9AcD3WxhuN7gP3dfr8UPr9+HVmfmyFq6vNsaSkyaYj8D3Ul30+j70evzQ2n2Y1PmotI81J3X7vdPt7dchhm5vvw4xdHv77Y5hUuekSX6MNMR9qafJuC+TOh+V9n6tIxljaxhja7Q0J9X9TOxo0rZT1T0zlwBLxrXiiLsyc9ZEA6sD96H7ej1+6I996KBRc9JE8hH0x+vgPnRfr8cP/bEPHdLSY6Ru/967vf06xNDt7dchhm5vvy4x9CiPkcbAfakn96XvTOo6kjG2hjG2RqtjfF6rVtQmG4CDGx5PBx7rUiySZE6SVBfmI0l1Yk6SVBfmI6lP1b2IfScwMyIOiYjnA6cCy7sck6TJy5wkqS7MR5LqxJwkqS7MR1KfqvVwIpm5LSLOAVYCewBLM/OBFq1+3Jey1ZD70H29Hj/0xz50hDlpVO5D9/V6/NAf+9B2bchH3f69d3v70P0Yur196H4M3d4+1COGnuMx0pi5L/XkvvQR85ExtogxtkZLY6z1xI6SJEmSJEmSpMmt7sOJSJIkSZIkSZImMYvYkiRJkiRJkqTampRF7IiYExHfiYh1EbGo2/E0ExEHR8RtEbE2Ih6IiHNL+/4RsSoiHiw/p5b2iIgryj7dFxFHdHcPtouIPSLimxHxpfL4kIi4o+zDjWWyBSLiBeXxurJ8RjfjHhIR+0XE5yLiX8rr8Z976XWIiD8u76H7I+L6iHhhr70G/awX8hGYk+r099DrOQnMS3XSjRwUEUsjYlNE3N/Q1vQ93KbtjyuftSmGF0bENyLiWyWGvyjtTf8O2hjHmPJhG7e/PiLWRMS9EXFXaevk6zDmfKrO60Z+aqVu57pWqUPObKW65N9W6XYe71ej5Z86HJ+OIcY/iYhvR/UZ4NaIeHXdYmzo946IyIiY1cn4yrZHjTEiTim/ywci4rN1izEiXlXy9DfL631ih+Pb6f/dsOURLfpMOumK2BGxB/AJ4ATgUOC0iDi0u1E1tQ1YmJm/BMwGzi5xLgJuzcyZwK3lMVT7M7PcFgBXdj7kEZ0LrG14/BHgo2UftgBnlfazgC2Z+Vrgo6VfHVwOfCUzfxF4A9W+9MTrEBEHAe8DZmXmYVQTW5xK770GfamH8hGYk+r099CzOQnMS3XSxRx0DTBnWNtI7+F2GG8+a4efAm/OzDcAbwTmRMRsRv47aJex5sN2+s3MfGNmDn1w7eTrMJ58qg7qsWOkkVxDd3Ndq9QhZ7ZSXfJvq9Qhj/eVMeafrh6fjjHGb1Idb/8y8Dngf9YwRiLixVSfDe7oZHxl26PGGBEzgfOBYzLz9cD76xYj8CHgpsx8E9Xnqk92Mkaa/79r1LLPpJOuiA0cBazLzIcy82fADcDcLse0k8zcmJn3lPtPU/1jOogq1mWl2zLgpHJ/LnBtVlYD+0XEgR0OeycRMR14G3BVeRzAm6mSKOy8D0P79jng2NK/ayJiX+DXgasBMvNnmfkkvfU67AlMiYg9gRcBG+mh16DP9UQ+AnMSNfl76JOcBOaluuhKDsrMrwGbhzWP9B5ux/bHm8/aEUNm5tbycK9yS0b+O2i5cebDTurI6zCBfKrO6pljpJF0O9e1Sh1yZivVIf+2So3zeK8bS/7p9vHpqDFm5m2Z+Ux5uBqY3sH4xhRjcRFVgf0nnQyuGEuMvw98IjO3AGTmphrGmMC+5f5LgMc6GN9I/+8atewz6WQsYh8EPNLweENpq61yacqbqL6ZmpaZG6E6oABeXrrVdb/+Gvgz4Ofl8UuBJzNzW3ncGOdz+1CWP1X6d9NrgCeAT5dLM66KiL3pkdchMx8F/gp4mKpI9BRwN731GvSzWr1fxsqc1FU9nZPAvFQzdXp/jPQebqsx5rN2bXuPiLgX2ASsAr7LyH8H7TCefNguCXw1Iu6OiAWlrVOvw3jzqTqrTvmplXr6/dXNnNlKNci/rVKHPN6PxpJ/un18Ot4ceRbw5bZGtLNRY4yINwEHZ+aXOhlYg7H8Hn8B+IWI+OeIWB0RuzrjuB3GEuOFwO9FxAZgBfBHnQltzFr2P30yFrGbfTuWHY9ijCJiH+DzwPsz84e76tqkrav7FRFvBzZl5t2NzU265hiWdcuewBHAleXSjB+x60vlarUPUY1NNxc4BHglsDfVpRzD1fk16Gc99/s2J3X99enpnATmpZqZ1L/bceSztsjMZzPzjVRnRh0F/FKzbu3Y9gTyYbsck5lHUOWAsyPi19u8vUbjzafqrEmdn+qo2zmzlbqZf1ulRnm8H43l99jt3/WYtx8RvwfMAv5XWyNqsukmbc/FGBHPoxqKZWHHItrZWH6Pe1INgzEAnAZcFRH7tTmuRmOJ8TTgmsycDpwIfKb8fuuiZX8vddqpTtkAHNzweDodPtV+rCJiL6oDhesy8wul+fGh0+7Lz6FLGeq4X8cAvx0R66kueXgz1bfF+5VLyGHHOJ/bh7L8Jez6koRO2ABsyMyh8Zk+R/WBp1deh7cA38vMJzLzP4AvAL9Kb70G/axu75ddMifV4u+h13MSmJfqpE7vj5Hew20xznzWVmUIi0GqsWZH+jtotfHmw7bIzMfKz03AF6mKSZ16HcabT9VZdcpPrdST76865cxW6lL+bZVa5PE+NZb80+3j0zHlyIh4C/BB4Lcz86cdim3IaDG+GDgMGCzv49nA8ujs5I5jfa1vzsz/yMzvAd+hKmp3ylhiPAu4CSAzbwdeCBzQkejGpmX/0ydjEftOYGZUs/Y+n2rQ8+VdjmknZTylq4G1mXlZw6LlwLxyfx5wc0P7mVGZDTw1dHlXt2Tm+Zk5PTNnUP2e/ykzTwduA95Rug3fh6F9e0fp39VvjjPz34FHIuJ1pelY4Nv0zuvwMDA7Il5U3lND8ffMa9DneiIfgTmJmvw99EFOAvNSndQpB430Hm65CeSzdsTwsqGzeCJiCtWXO2sZ+e+gpSaQD1suIvaOajInyjAexwP306HXYQL5VJ1Vp/zUSj33/qpDzmylbuffVqlDHu9jY8k/3T4+HTXGMlTH31IVsLvxJdMuY8zMpzLzgMycUd7Hq0usd9UlxuJ/A78JEBEHUA0v8lDNYnyY6jiGiPglqiL2Ex2McTSt+0yamZPuRnV6/b9SjX31wW7HM0KMv0Z1ev19wL3ldiLVOEu3Ag+Wn/uX/kE1Y+l3gTVUs9B2fT8a9mcA+FK5/xrgG8A64O+BF5T2F5bH68ry13Q77hLXG4G7ymvxv4GpvfQ6AH8B/AvVB8PPAC/otdegn2+9kI9KnOakmvw99HpOKnGZl2py60YOAq6nGg/9P6jOzDhrpPdwm7Y/rnzWphh+GfhmieF+4M9Le9O/gza/HqPmwzZt9zXAt8rtgaH3X4dfhzHnU2+dv3UjP7U4/q7muhbuR9dzZov3pzb5t4X71JU83s+3ZvkH+DBVkRVqcHw6hhj/EXi84e92ed1iHNZ3kC58VhnD7zGAy6i+6F4DnFrDGA8F/pnqmOpe4PgOx9fs/90fAn/Y8DtsyWfSKCuUJEmSJEmSJKl2JuNwIpIkSZIkSZKkHmERW5IkSZIkSZJUWxaxJUmSJEmSJEm1ZRFbkiRJkiRJklRbFrElSZIkSZIkSbVlEVuSJEmSJEmSVFsWsSVJkiRJkiRJtWURW5IkSZIkSZJUWxaxJUmSJEmSJEm1ZRFbkiRJkiRJklRbFrElSZIkSZIkSbVlEVuSJEmSJEmSVFsWsSVJkiRJkiRJtWURW5IkSZIkSZJUWxaxJUmSJEmSJEm1ZRFbkiRJkiRJklRbFrElSZIkSZIkSbVlEVuSJEmSJEmSVFsWsSVJkiRJkiRJtWURW5IkSZIkSZJUWxaxJUmSJEmSJEm1ZRF7EouIByJioNtxSFKrRMT6iHhLF7b7qojYGhF7jLD8woj4u07HJak9upVrJE0ekznPeNwkaaIiYklEfCcifh4R87sdj1rLIvYklpmvz8zB0fp18wAqIjIiXtumdQ9GxH9rx7olTS6Z+XBm7pOZz3Y7Fkn9KyJmlGOjPbsdi6T+ZJ6R1OO+BbwXuKfbgaj1LGKrraLS8feZB12SJEk78vhIkiT1mnJi5X+PiPsi4kcRcXVETIuIL0fE0xHxjxExFSAzP5GZtwI/6XLYagOL2JPY0BnW5XKtmyLi2pIAHoiIWaXPZ4BXAf9QLpX/s9I+OyL+b0Q8GRHfahyWpJzhfElE/DPwDPCaiHhJSTQbI+LRiLh46LL7iHhtRPyfiHgqIr4fETeW9q+VVX6rbPt3d7EvAxGxISLOi4h/Bz4dEVMj4ksR8UREbCn3p5f+lwD/Bfh4WffHS/svRsSqiNhcLkE5paW/dEkdEREviIi/jojHyu2vI+IFZdmIuaEsH4yIiyLin0tO/GpEHDDK9nY4aykiDil57emIWAXs8vmSelOncw0wdGz0ZDl++c9lXe+OiLVlOysj4tUN28mIODsiHgQebGh7b0Q8WLZ9UUT8p4i4PSJ+WI4Ln9/SX5akCalDnomI50XEhyLi3yJiU1SfG18yStxfiYhzhrV9KyL+a7l/eUQ8UnLO3RHxX0ZYz0BEbBjW9tyVwiW2RRHx3Yj4Qclf+4+yj5J6z/8DHAf8AvBbwJeBD1B9znoe8L7uhaZOsYitIb8N3ADsBywHPg6QmWcADwO/VS6V/58RcRBwC3AxsD/wp8DnI+JlDes7A1gAvBj4N2AZsA14LfAm4HhgaCiPi4CvAlOB6cDHyrZ/vSx/Q9n2jaPswytKPK8u234e8Ony+FXAjxv264PA/wecU9Z9TkTsDawCPgu8HDgN+GREvH4sv0BJtfJBYDbwRuANwFHAh8qyEXNDg3cC76LKBc+nynPj8VngbqqDqouAeePeA0m9oNO5ZujYaL9y/HJ7RJxE9SHuvwIvozq+uX7Y804CjgYObWibAxxZ4v8zYAlwOnAwcBjVcZCk7ut6ngHml9tvAq8B9mmyneE+S0MeiYhDS5y3lKY7yz7tX/r+fUS8cJR1NvM+qhz3G8ArgS3AJyawHkn19rHMfDwzH6U61rkjM7+ZmT8FvkhVZ1Kfs4itIV/PzBVlPNfPUB0gjeT3gBWl/88zcxVwF3BiQ59rMvOBzNxGdWByAvD+zPxRZm4CPgqcWvr+B9UBzSsz8yeZ+fUJ7sPPgQsy86eZ+ePM/EFmfj4zn8nMp4FLqA5uRvJ2YH1mfjozt2XmPcDngXdMMB5J3XM68OHM3JSZTwB/QfXlGmPMDZ/OzH/NzB8DN1F9yBqTiHgV8CvA/1vy0deAf2jBPkmqn67lmgZ/APyPzFxbjrv+Enhj49nYZfnmsp0hH8nMH2bmA8D9wFcz86HMfIrq7CY/DEr1UIc8czpwWckRW4HzgVNj10MUfZEdc9HpwBdKwYnM/LsS/7bMvBR4AfC6CcT2B8AHM3NDWfeFwDtGiU1S73m84f6Pmzzep7PhqBssYmvIvzfcfwZ44S7+8b8aODmqoUSejIgngV8DDmzo88iw/nsBGxv6/y3V2QBQnf0TwDeiGsrk3RPchycy87lxjyLiRRHxt+Wytx9SXRq3X5RhTEbYr6OH7dfpVGd4S+otr6S6CmTIv5W2seaG4TlxPAdFrwS2ZOaPhm1fUv/pZq4Z8mrg8oZjl81Ux1UHNfR5pMnz/DAo9YY65JlmMewJTBvpCaWofgvbT1w6FbhuaHlELCzDID1VctdLmNjwa68GvtiQA9cCz+4qNklSb7KIrbHIYY8fAT6Tmfs13PbOzMUjPOcR4KfAAQ39983M1wNk5r9n5u9n5iupvkn/ZES8tgVxLqT6Nv/ozNyX7ZfGxS726/8M2699MvM9E4hFUnc9RvWhZsirShuMnht210ZgahmiqHH7kvpPp3PN8GMXqI5f/mDY8cuUzPy/ozxPUm+oQ55pFsM2dvzyq5nrgdPK+P1TgNsAyvjX5wGnAFMzcz/gqRHi/hHwoqEHpUDfOIzlI8AJw3LgC8uQA5ImmYh4fhmaKIC9IuKFEWHts0/4QmosHqca+2zI3wG/FRFvjYg9SlIYaJxEpFFmbqQa8/rSiNi3TL7xnyLiNwAi4uSG526hOnB6doRtj8eLqc4kerJM7nHBKPv1JeAXIuKMiNir3H4lIn5pgtuX1D3XAx+KiJeVCYz+nCp3wei5Ybdk5r9RDbH0F+Ug6teoJh+R1H86nWueoBo+rfH4wauGxAAAIABJREFU5W+A84fm8IhqMu2TW7AtSfVQhzxzPfDHUU1cvQ/VsEU3liGMdmUFVfH7w6X/zxvi3la2tWdE/Dmw7wjr+Feqq4TfFhF7UY0H/oKG5X8DXDI0bEn5Pc0d475K6j9fpcqLv0o138eP2f4Fn3qcRWyNxf+gOnB6MiL+NDMfAeZSTSL0BNW33/+dXb+fzqSaSOTbVIXqz7F9+JFfAe6IiK1Uk0qem5nfK8suBJaVbZ8yzrj/muob/+8Dq4GvDFt+OdV4aVsi4opyydvxVJe6PUZ16d1H2PEgSVJvuJiqkHwfsAa4p7TB6LmhFd5JNYnaZqoPlNe2YRuSuq+juSYzn6Ea8/afy7HR7Mz8ItXxyg1lOIH7qeYikdQfup5ngKVU8yZ9Dfge8BPgj8awrp8CXwDeQjV545CVVGPv/yvV0CQ/ofmwR5Rx+t8LXAU8SnVm9oaGLpdTfYb8akQ8TfV7OHqMuyupB2TmjMz8x4bHv5eZFzY8vioz31LuD2RmDLsNdj5qtUNkenWhJEmSJEmSJKmePBNbkiRJkiRJklRbFrHVMyLiAxGxtcnty92OTVL/i4jTR8hBD3Q7Nkn9w1wjqd1alWfMV5KkTnI4EUmSJEmSJLVERCwF3g5syszDStv/oprs/GfAd4F3ZeaTZdn5wFnAs8D7MnNlaZ9DNe75HsBVmbm4tB8C3ADsTzVO/BmZ+bOIeAHVXDRHAj8Afjcz13dkpyW1nWdiS5IkSZIkqVWuAeYMa1sFHJaZv0w1qef5ABFxKHAq8PrynE9GxB4RsQfwCarJig8FTit9oZrQ+KOZORPYQlUAp/zckpmvBT5a+knqE3t2O4BWO+CAA3LGjBndDmPMfvSjH7H33nt3O4xx68W4ezFm6L2477777u9n5su6HceQiDiY6tv4VwA/B5Zk5uURcSHw+8ATpesHMnNFec64zgQYSR3yUa+9f4b0Yty9GDP0Ztxjjblu+ajbGnNSXV73usQB9YnFOPo3DnPSdnU4RhqLurz/xsOYO6PXY253PsrMr0XEjGFtX214uBp4R7k/F7ghM38KfC8i1gFHlWXrMvMhgIi4AZgbEWuBNwPvLH2WARcCV5Z1XVjaPwd8PCIiRxmCYKw5qS6vu3HUMw6oTyy9FseYc1Jm9tXtyCOPzF5y2223dTuECenFuHsx5szeixu4K2uQC4ZuwIHAEeX+i6m+9T+U6uDmT5v0PxT4FvAC4BCqS932KLfvAq8Bnl/6HLqrbdchH/Xa+2dIL8bdizFn9mbcY425bvmo27fGnFSX170ucWTWJxbj2FE/xWFOap6P6qwu77/xMObO6PWYO5GPgBnA/SMs+wfg98r9jw/dL4+vpipwv4PqxKGh9jNK3wOoittD7QcPbQe4H5jesOy7wAGjxTrWnFSX1904dlSXODLrE0uvxTHWnNR3Z2JLUqPM3AhsLPefLt/cH7SLp4zrTADg220LXpIkSZL6SER8ENgGXDfU1KRb0nz429xF/12tq1kcC4AFANOmTWNwcHDkoIutW7eOqV+7GUc944D6xNKvcVjEljRplEva3gTcARwDnBMRZwJ3AQszcwtVgXt1w9M2sL3o/ciw9qPbHLIkSZIk9YWImEc14eOx5exLqD5XHdzQbTrwWLnfrP37wH4RsWdmbhvWf2hdGyJiT+AlwOZmsWTmEmAJwKxZs3JgYGDU+AcHBxlLv3YzjnrGAfWJpV/jsIgtaVKIiH2AzwPvz8wfRsSVwEVU38xfBFwKvJvxnwkwfDvj/ka/neryDex49WLcvRgz9GbcvRizJEnSZFbmFzoP+I3MfKZh0XLgsxFxGfBKYCbwDarPZTMj4hDgUarJH9+ZmRkRt1ENN3IDMA+4uWFd84Dby/J/aiiWS+pxFrEl9b2I2IuqgH1dZn4BIDMfb1j+KeBL5eF4zwTYwUS+0W+nunwDO169GHcvxgy9GXcvxixJkjRZRMT1wABwQERsAC4Azqead2hVRACszsw/zMwHIuImqmEatwFnZ+azZT3nACup5idampkPlE2cB9wQERcD36QaR5vy8zNlSMjNVIVvSX3CIrakvhbVEdLVwNrMvKyh/cAyXjbA71BNAgLjPBOgM3shSZI0MRFxMHAt8Arg58CSzLw8Ii4Efh94onT9QGauKM85HzgLeBZ4X2auLO1zgMupCkpXZebi0n4I1RmR+wP3AGdk5s86s4eS6iYzT2vSfHWTtqH+lwCXNGlfAaxo0v4Q2+ctamz/CXDyuIKV1DMsYkvqd8dQzWS9JiLuLW0fAE6LiDdSDQmyHvgDgAmeCSBJklRX26jm/rgnIl4M3B0Rq8qyj2bmXzV2johDqb6sfz3VF/r/GBG/UBZ/AjiO6sq1OyNieWZ+G/hIWdcNEfE3VAXwK9u+Z5IkadKwiK1dmrHolh0er1/8ti5FIk1MZn6d5uNc7/SNfsNzxnUmgOrLHCZJarVe+99SrjzbWO4/HRFr2T5pdTNzgRsy86fA98pl+UNnPK4rZ0ASETcAc8v63sz2K9SWARdiEVt9pvFvf+Hh2xjoXijqkjWPPsX8hvdB3fO/1G8sYkuSJLVQRLwOuLGh6TXAn1Ndzn8jMIPqCpBTMnNLGfbocuBE4BlgfmbeU9Y1D/hQWc/FmbmstB8JXANMofpy7VwnLpI0moiYAbwJuIPqarVzIuJM4C6qs7W3UBW4Vzc8bQPbi96PDGs/Gngp8GRmbmvSv3HbtZr8eix6cSJhY26fhYdve+7+tCn0RMyNeuX3LEkjsYgtSdrJ8LPMwDMNpLHKzO8AbwSIiD2oxtH/IrAIuDUzF0fEovL4POAEqvH3Z1IVhK4Ejo6I/akmQppFNfTR3eXS/S2lzwKqQtMKYA7w5Y7tpKSeExH7UE10/f7M/GFEXAlcRJVfLgIuBd5N8yvYEnjeCO0j9d+xoWaTX49FL04kbMztM3/Ymdin9EDMjXrl9yxJI7GILUmTUK9dCt0rWvF7nYyvTZ/v87HAdzPz3yJiLjx39fEyYJCqiD0XuLacSb06IvaLiANL31WZuRmgjGE7JyIGgX0z8/bSfi1wEhaxJY0gIvaiKmBfl5lfAMjMxxuWfwr4Unm4ATi44enTgcfK/Wbt3wf2i4g9y9nYjf0lSZJawiK2JElS+5wKXF/uTytj05KZGyPi5aX9IHa+RP+gUdo3NGnfyUiX79flkuK6xAH1icU46h9H45ACUP8hBcqQRVcDazPzsob2A4dyEvA7wP3l/nLgsxFxGdXEjjOBb1CdcT0zIg6husLkVOCdmZkRcRvwDuAGYB5wc/v3TJIkTSYWsSWpBkY7E3V3l/eC0SbL8Szn3jVZf+8R8Xzgt4HzR+vapG1Xl+iP6dJ9GPny/bpcUlyXOKA+sRhH/eOYPzynnT7Q+YDG5xjgDGBNRNxb2j4AnBYRb6TKH+uBPwDIzAci4ibg28A24OzMfBYgIs4BVgJ7AEsz84GyvvOAGyLiYuCbVEVzSZKklrGILUnqiF4sZNYx5nbHVMd97mEnAPc0XLL/+NCZj2W4kE2lfaRL9zfADt/nTKcagmRDuT+8vyTtJDO/TvMvv1bs4jmXAJc0aV/R7HmZ+RBw1G6EKUmStEsWsSVJXdFs8shux1DHgnAdfk+j6fTv8Zo5e7d0/W10GtuHEoHqEv15wGJ2vNx+OXBORNxANbHjU6XQvRL4y4iYWvodD5yfmZsj4umImA3cAZwJfKz9uyNJkiRJ3WERW5I0IZ0uXHZ6+2MxfAiUbvxbbfdQM71QRK+jiHgRcBzl8vxiMXBTRJwFPAycXNpXACcC64BngHcBlGL1RcCdpd+HhyZ5BN4DXANMoZrQ0UkdJUmSJPUti9iSVEOtLhy2Yn29UFSW6iIznwFeOqztB8CxTfomcPYI61kKLG3SfhdwWEuClSRJkqSam3AROyIOBq4FXgH8HFiSmZdHxP7AjcAMqglCTsnMLWVW7MupzjR6BpifmfeUdc0DPlRWfXFmLivtR7L9LKMVwLnlg54kaReGCsoLD9+20wRU/cozhuvD10KSJEmS1Eq7cyb2NmBhZt4TES8G7o6IVcB84NbMXBwRi4BFVLNVnwDMLLejgSuBo0vR+wJgFtXM2HdHxPLM3FL6LABWUxWx5+DlspLUchYdK+M927zd2+uENY8+VbsvOurwe5EkSZIk1ceEi9iZuRHYWO4/HRFrgYOAucBA6bYMGKQqYs8Fri1nUq+OiP0i4sDSd9XQGI+lED4nIgaBfTPz9tJ+LXASFrElST2qDsXZ4TEsPLz7MUiSJEmStCstGRM7ImYAbwLuAKaVAjeZuTEiXl66HQQ80vC0DaVtV+0bmrQ32/4CqjO2mTZtGoODg7u1P520devWWsdbTVS23VCsdY+7mV6MGXo3brVXLxQBeyFGSZIkSZJUf7tdxI6IfYDPA+/PzB9WQ18379qkLSfQvnNj5hJgCcCsWbNyYGBglKjrY3BwkDrHO/wS8/WnDwD1j7uZXowZejduSZIkSZIkqRV2q4gdEXtRFbCvy8wvlObHI+LAchb2gcCm0r4BOLjh6dOBx0r7wLD2wdI+vUl/SZI0As+AlyRJkiT1m+dN9IlRnXJ9NbA2My9rWLQcmFfuzwNubmg/MyqzgafKsCMrgeMjYmpETAWOB1aWZU9HxOyyrTMb1iVJkiRJkiRJmgR250zsY4AzgDURcW9p+wCwGLgpIs4CHgZOLstWACcC64BngHcBZObmiLgIuLP0+/DQJI/Ae4BrgClUEzo6qaMkSZIkSZIkTSITLmJn5tdpPm41wLFN+idw9gjrWgosbdJ+F3DYRGOUJEmSJEmSJPW2CQ8nIkmSJEmSJDWKiKURsSki7m9o2z8iVkXEg+Xn1NIeEXFFRKyLiPsi4oiG58wr/R+MiHkN7UdGxJrynCvKELQjbkNSf7CILUmSJEmSpFa5BpgzrG0RcGtmzgRuLY8BTgBmltsC4EqoCtLABcDRwFHABQ1F6StL36HnzRllG5L6gEVsSZIkSZIktURmfg3YPKx5LrCs3F8GnNTQfm1WVgP7RcSBwFuBVZm5OTO3AKuAOWXZvpl5exm29tph62q2DUl9wCK2JEmSJEmS2mlaZm4EKD9fXtoPAh5p6LehtO2qfUOT9l1tQ1IfmPDEjpIkSZIkSdJuiCZtOYH28W00YgHVkCRMmzaNwcHBUZ8zbQosPHzbc4/H8px22Lp1a9e2bRy7VpdY+jUOi9iSJEktFhH7AVcBh1F9sHo38B3gRmAGsB44JTO3lMmILgdOBJ4B5mfmPWU984APldVenJnLSvuRVONNTgFWAOeWS2olSZLq6PGIODAzN5YhQTaV9g3AwQ39pgOPlfaBYe2DpX16k/672sZOMnMJsARg1qxZOTAwMFLX53zsupu5dM32Mtr600d/TjsMDg4ylniNo/PqEku/xuFwIpIkSa13OfCVzPxF4A3AWjozoZEkSVIdLQfmlfvzgJsb2s+MymzgqTIUyErg+IiYWo5/jgdWlmVPR8TsciLAmcPW1WwbkvqARWxJfS0iDo6I2yJibUQ8EBHnlvb9I2JVRDxYfk4t7RERV0TEuoi4LyKOaFjXvNL/wXJ2pCTtJCL2BX4duBogM3+WmU/SmQmNJEmSuioirgduB14XERsi4ixgMXBcRDwIHFceQ3VF2UPAOuBTwHsBMnMzcBFwZ7l9uLQBvIfqird1wHeBL5f2kbYhqQ84nIikfrcNWJiZ90TEi4G7I2IVMJ/qjMjFEbGI6ozI89jxjMijqc52PLrhjMhZVEMD3B0Ry0thSZIavQZ4Avh0RLwBuBs4l2GTDUVEOyY02sFI4z326zh5u6MusRhH/eNoHA8VujcmqiTVVWaeNsKiY5v0TeDsEdazFFjapP0uqiHbhrf/oNk2JPUHi9iS+lopGA0VjZ6OiLVUxZ65bB9jbRnV+Grn0XBGJLA6IobOiBygnBEJUArhc4DrO7YzknrFnsARwB9l5h0RcTnbhw5ppm0TGo003mO/jpO3O+oSi3HUP475i27ZYVm3xkSVJEmaTBxORNKkEREzgDcBdzDsjEhgomdEStJwG4ANmXlHefw5qqL24+VLMcYxodFI7SNNaCRJkiRJfcczsTUuM8qZJwsP38b8RbewfvHbuhyRNDYRsQ/weeD9mfnDag6Q5l2btI35zMeRLt0fzfBLk1tl2pT2rbudejHuXowZejPuugwvMJLM/PeIeCQiXpeZ36G6rPXb5TaPanzG4RManRMRN1ANY/RUGW5kJfCXDZM5Hg+cn5mbI+LpMvnRHVQTGn2sYzsoSZIkSR1mEVtS34uIvagK2Ndl5hdK8+MRcWApFI31jMiBYe2Dw7c10qX7oxl+aXKrLDx8G5eu6b1U34tx92LM0JtxXzNn71oMLzCKPwKui4jnU01W9C6qK+BuKpMbPQycXPquAE6kmpzomdKXUqwemtAIdp7Q6BpgCtVkRkMTGknSDiLiYKoJYF8B/BxYkpmXl/k+bgRmAOuBUzJzS1Tf9F9OlZeeAeZn5j1lXfOAD5VVX5yZy0r7kWzPSSuAc8vQbJIkSS3RW59aJWmcygexq4G1mXlZw6LltOCMyE7sg6Tek5n3Uk0EO1xbJzSSpCY6Mcn1lVRXoq2mKmLPwS/XJElSCzkmtqR+dwxwBvDmiLi33E6kKl4fFxEPAseVx1B98HqI6ozITwHvheqMSGDojMg72fGMSEmSpFrKzI1DZ1Jn5tNA4yTXy0q3ZcBJ5f5zk1xn5mpgaJLrt1ImuS6F61XAnLJs38y8vXwpd23DuiRJklrCM7El9bXM/DrNx7OGFp0RKUmS1At2Ncl1REx0kuuDyv3h7cO3PaF5Q7qp7nMwNGPM7dM4h8i0KfREzI165fcsSSOxiC1JkiRJfa6Nk1yPafLric4b0k2Dg4O9MAfDDoy5fRrnsFl4+DZO6YGYG/XK71mSRuJwIpIkSZLUx3Y1yXVZPtZJrkdqn96kXZIkqWUsYkuSJElSnxrDJNew8yTXZ0ZlNmWSa2AlcHxETC0TXR8PrCzLno6I2WVbZzasS5IkqSUcTkSSJEmS+tfQJNdrIuLe0vYBqkmtb4qIs4CHgZPLshXAiVSTXD8DvAuqSa4jYmiSa9hxkuv3ANcAU4Avl5skSVLLWMSWJEmSpD7ViUmuM/Mu4LDdCFOSJGmXHE5EkiRJkiRJklRbFrElSZIkSZIkSbVlEVuSJEmSJEmSVFsWsSVJkiRJkiRJtWURW5IkSZIkSZJUWxaxJUmSJEmSJEm1ZRFbkiRJkiRJklRbFrElSZJaLCLWR8SaiLg3Iu4qbftHxKqIeLD8nFraIyKuiIh1EXFfRBzRsJ55pf+DETGvof3Isv515bnR+b2UJEmSpM7YrSJ2RCyNiE0RcX9D24UR8Wj50HZvRJzYsOz88mHrOxHx1ob2OaVtXUQsamg/JCLuKB/cboyI5+9OvJIkSR30m5n5xsycVR4vAm7NzJnAreUxwAnAzHJbAFwJVdEbuAA4GjgKuGCo8F36LGh43pz2744kSdLuiYg/jogHIuL+iLg+Il44Uu0nIl5QHq8ry2c0rGdc9SVJvW93z8S+huYfmj5aPrS9MTNXAETEocCpwOvLcz4ZEXtExB7AJ6g+wB0KnFb6AnykrGsmsAU4azfjlSRJ6pa5wLJyfxlwUkP7tVlZDewXEQcCbwVWZebmzNwCrALmlGX7ZubtmZnAtQ3rkiRJqqWIOAh4HzArMw8D9qCqE41U+zkL2JKZrwU+WvpNtL4kqcftVhE7M78GbB5j97nADZn508z8HrCO6qyio4B1mflQZv4MuAGYWy6LfTPwufL8xg97kiRJdZbAVyPi7ohYUNqmZeZGgPLz5aX9IOCRhuduKG27at/QpF2SJKnu9gSmRMSewIuAjYxc+2k8AeBzwLGlVjSu+lIH9klSB+zZpvWeExFnAncBC8vZQwcBqxv6NH7gGv4B7WjgpcCTmbmtSf8dlA+HCwCmTZvG4OBgi3aj/bZu3VrreBcevq1p+7Qp1bI6xz5c3X/XI+nVuCVpkjsmMx+LiJcDqyLiX3bRt9l41jmB9p1XPMIxUl3+t9QlDqhPLMZR/ziGHx/XIT5J6gWZ+WhE/BXwMPBj4KvA3Yxc+3nuC/3M3BYRT1HVisZbX9rJROpIQ3WQId3K/3X832gclbrE0q9xtKOIfSVwEdWHqYuAS4F3M/IHrmZng4/rA1pmLgGWAMyaNSsHBgbGHXS3DA4OUud45y+6pWn7wsO3cemaPVl/+kBnA9oNdf9dj6RX45akySwzHys/N0XEF6nODHo8Ig7MzI1lSJBNpfsG4OCGp08HHivtA8PaB0v79Cb9m8XR9BipLv9b6hIH1CcW46h/HMOPj3vpeFiSuqnM7TEXOAR4Evh7qqE/hhuq/Yz3C/2R6ks7N06gjvSx627m0jXby2jdyv91/N9oHJW6xNKvcezumNg7yczHM/PZzPw58CmqD22w6w9ozdq/TzUm5J7D2iVJkmorIvaOiBcP3QeOB+4HlgPzSrd5wM3l/nLgzKjMBp4qw42sBI6PiKnlQ9/xwMqy7OmImF0uqT2zYV2SJEl19Rbge5n5RGb+B/AF4FcZufbzXL2oLH8J1ZC2460vSeoDLS9ilzOLhvwO1Yc2qD6gnVpmlz0EmAl8A7gTmFlmo30+1eD8y8tERbcB7yjPb/ywJ0mSVFfTgK9HxLeojnVuycyvAIuB4yLiQeC48hhgBfAQ1XiOnwLeC5CZm6muaruz3D5c2gDeA1xVnvNd4Msd2C9JkqTd8TAwOyJeVL6IPxb+f/buP1rSqr7z/fsjCBIVQYk9BJg0maATlNFoDzBxlqvVEVvNBGeNOhgSGoeEicEZE5k7NpPci1G5l2SiRo0h00qnwVGQOHHoG1FC0LOc5AqCv/ghMXSQkRYCGpDQcdS0fu8ftQ9UH+qcU326fjx1zvu1Vq1TtZ9dT33qOVW7ntr1PHvzZRbv++k/AOCVwCdbX9E+9S9N4HlJmoD9Gk4kyWX0TnM9Isku4HxgY5Jn0ztl407g3wFU1a1JrqDXQO0Bzqmq77f1vJ7e0UYHANuq6tb2EG8CLk/yNuALwMX7k1eSJGncquoO4FkDyv+G3pe1heUFnLPIurYB2waU3wg8c7/DSpIkTUhVXZ/kI8Dn6fULfYHekB4fY3Dfz8XAB5LspHcE9mltPSvpX5I04/arE7uqXjOgeNGO5qq6ALhgQPlV9I5CWlh+B48MRyJJkiRJkqQZVVXn0zsAst/Avp+q+g7wqkXWs0/9S5Jm38iHE5EkSZIkSZIkaVTsxJYkSZIkSZIkdZad2JIkSZIkSZKkzrITW5IkSZIkSZLUWXZiS1rVkmxLcl+SW/rK3pzk60m+2C4v61t2XpKdSb6S5CV95Zta2c4kWyb9PCRJkiRJktYqO7ElrXbbgU0Dyt9ZVc9ul6sAkhwPnAY8o93n95IckOQA4L3AS4Hjgde0upIkSZIkSRqzA6cdQJLGqao+nWT9kNVPBS6vqu8CX02yEzixLdtZVXcAJLm81f3yiONKkiRJkiRpATuxJa1Vr09yBnAjcG5VPQAcBVzXV2dXKwO4a0H5SYNWmuRs4GyAdevWMTc3N1SYc0/Ysy/Zh7bukPGte5xmMfcsZobZzL179+6h31uStNYl2Qb8NHBfVT2zlb0Z+EXgG63af+47M+084Czg+8B/qKqrW/km4F3AAcD7q+rCVn4scDnwZODzwM9X1fcm8+wkSdJaYSe2pLXoIuCtQLW/bwf+LZABdYvBQy/VoBVX1VZgK8CGDRtq48aNQwU6c8vHhqq3r849YQ9vv3n2mvpZzD2LmWE2c2/f9HiGfW9JktgO/C5w6YLyd1bVb/cXLBha7UeAP03ytLb4vcCL6f2Yf0OSHVX1ZeA327ouT/L79DrALxrXk5EkSWuTY2JLWnOq6t6q+n5V/QB4H48MGbILOKav6tHA3UuUS5IkdVpVfRq4f8jqDw+tVlVfBeaHVjuRNrRaO8r6cuDUJAFeCHyk3f8S4BUjfQKSJEl4JLakNSjJkVV1T7v5r4Bb2vUdwIeSvIPe0UfHAZ+ld4T2ce102a/TO0LpZyebWpIkaaRGMbTaU4BvVdWeAfX3stIh16ZpFoevMvP49A+/tu4QZiJzv1nZzpK0GDuxJa1qSS4DNgJHJNkFnA9sTPJsekOC3An8O4CqujXJFfQmbNwDnFNV32/reT1wNb1xILdV1a0TfiqSJEmjMqqh1Rar/+jCFQ65Nk1zc3MzN3yVmcenf/i/c0/Yw6tnIHO/WdnOkrQYO7ElrWpV9ZoBxRcvUf8C4IIB5VcBV40wmiRJ0lRU1b3z15O8D/jjdnOpIdQGlX8TOCzJge1obIdckyRJY+GY2JIkSZK0hiQ5su/mwqHVTktycBtGbX5otRtoQ6slOYje0Go7qqqATwGvbPffDFw5iecgSZLWFjuxJUmSxiDJAUm+kOSP2+1jk1yf5PYkH24dQbTOog8n2dmWr+9bx3mt/CtJXtJXvqmV7UyyZdLPTdLsaEOrfQZ4epJdSc4CfivJzUluAl4A/Cr0hlYD5odW+wRtaLV2lPX80Gq3AVf0Da32JuCNSXbSGyN70TPeJEmSVsrhRCRJksbjDfQ6ew5tt38TeGdVXZ7k94Gz6I1LexbwQFX9eJLTWr1/k+R4ekc7PoPeZLN/muRpbV3vBV5M79T/G5LsqKovT+qJSZod4x5araruAE7cn4ySJEnL8UhsSZKkEUtyNPBy4P3tdoAXAh9pVS4BXtGun9pu05a/qNU/Fbi8qr5bVV8FdtLrKDoR2FlVd1TV94DLW11JkiRJWpU8EluSJGn0fgf4T8AT2+2nAN9qp+RD7wjqo9r1o4C7AKpqT5IHW/2jgOv61tl/n7sWlJ80KESSs4GzAdatW8fc3BwAu3fvfvj6NHUlB3QnizlW3EKfAAAgAElEQVS6n+PcE/bstawL+SRJklY7O7ElSZJGKMlPA/dV1eeSbJwvHlC1llm2WPmgM+lqQBlVtRXYCrBhw4bauLEXZ25ujvnr09SVHNCdLObofo4zt3xsr2V3nr5x8oEkSZLWGDuxJUmSRut5wM8keRnwOHpjYv8OcFiSA9vR2EcDd7f6u4BjgF1JDgSeBNzfVz6v/z6LlUuSJEnSquOY2JIkSSNUVedV1dFVtZ7exIyfrKrTgU8Br2zVNgNXtus72m3a8k9WVbXy05IcnORY4Djgs8ANwHFJjk1yUHuMHRN4apIkSfslyWFJPpLkL5LcluSfJXlykmuS3N7+Ht7qJsm7k+xMclOS5/StZ3Orf3uSzX3lz01yc7vPu9s8I5JWATuxJUmSJuNNwBuT7KQ35vXFrfxi4Cmt/I3AFoCquhW4Avgy8AngnKr6fjuS+/XA1cBtwBWtriRJUte9C/hEVf1j4Fn09mW2ANdW1XHAte02wEvp/Yh/HL05Pi4CSPJk4Hx6c4KcCJw/3/Hd6pzdd79NE3hOkibA4UQkSZLGpKrmgLl2/Q56X7QW1vkO8KpF7n8BcMGA8quAq0YYVZIkaaySHAo8HzgToKq+B3wvyanAxlbtEnr7Tm8CTgUubWeoXdeO4j6y1b2mqu5v670G2JRkDji0qj7Tyi8FXgF8fAJPT9KY2YktSZIkSZKkcfsx4BvAHyR5FvA54A3Auqq6B6Cq7kny1Fb/KOCuvvvvamVLle8aUP4oSc6md8Q269atY25ubtnw6w6Bc0/Y8/DtYe4zDrt3757aY5tjaV3Jslpz2IktSZIkSZKkcTsQeA7w76vq+iTv4pGhQwYZNJ51raD80YVVW4GtABs2bKiNGzcuEaPnPR+8krff/Eg32p2nL3+fcZibm2OYvOaYvK5kWa05HBNbkiRJkiRJ47YL2FVV17fbH6HXqX1vGyaE9ve+vvrH9N3/aODuZcqPHlAuaRWwE1uSJEmSJEljVVV/DdyV5Omt6EX0JrDeAWxuZZuBK9v1HcAZ6TkZeLANO3I1cEqSw9uEjqcAV7dlDyU5OUmAM/rWJWnGOZyIJEmSJEmSJuHfAx9MchBwB/BaegdYXpHkLOBrPDLh9VXAy4CdwLdbXarq/iRvBW5o9d4yP8kj8DpgO3AIvQkdndRRWiXsxJYkSZIkSdLYVdUXgQ0DFr1oQN0CzllkPduAbQPKbwSeuZ8xJXWQw4lIkiRJkiRJkjprvzqxk2xLcl+SW/rKnpzkmiS3t7+Ht/IkeXeSnUluSvKcvvtsbvVvT7K5r/y5SW5u93l3G9NIkiRJkiRJkrRG7O+R2NuBTQvKtgDXVtVxwLXtNsBLgePa5WzgIuh1egPnAycBJwLnz3d8tzpn991v4WNJkiRJkiRJklax/erErqpPA/cvKD4VuKRdvwR4RV/5pdVzHXBYkiOBlwDXVNX9VfUAcA2wqS07tKo+08ZBurRvXZIkSZIkSZKkNWAcEzuuq6p7AKrqniRPbeVHAXf11dvVypYq3zWg/FGSnE3viG3WrVvH3Nzc/j+LCdm9e/fQeW/++oN73T7hqCeNIdHezj1hz8DydYf0lq3Wbd0ls5pbkiRJkiRJGoVxdGIvZtB41rWC8kcXVm0FtgJs2LChNm7cuMKIkzc3N8ewec/c8rG9bt95+nD32x8LH3PeuSfs4e03HziRDKOyL9u6S2Y1tyRJkiRJkjQK+zsm9iD3tqFAaH/va+W7gGP66h0N3L1M+dEDyiVJkiRJkiRJa8Q4OrF3AJvb9c3AlX3lZ6TnZODBNuzI1cApSQ5vEzqeAlzdlj2U5OQkAc7oW5ckSZIkSZIkaQ3Yr07sJJcBnwGenmRXkrOAC4EXJ7kdeHG7DXAVcAewE3gf8MsAVXU/8FbghnZ5SysDeB3w/nafvwI+vj95JUmSxi3J45J8NsmXktya5Dda+bFJrk9ye5IPJzmolR/cbu9sy9f3reu8Vv6VJC/pK9/UynYm2TLp5yhJkiRJk7RfY2JX1WsWWfSiAXULOGeR9WwDtg0ovxF45v5klLS2JdkG/DRwX1U9s5U9GfgwsB64E3h1VT3Qzvp4F/Ay4NvAmVX1+XafzcCvt9W+raoumeTzkDRTvgu8sKp2J3ks8GdJPg68EXhnVV2e5PeBs4CL2t8HqurHk5wG/Cbwb5IcD5wGPAP4EeBPkzytPcZ76R0ssAu4IcmOqvryJJ+kJEmSJE3KOIYTkaQu2Q5sWlC2Bbi2qo4Drm23AV4KHNcuZ9PrXJrv9D4fOAk4ETi/DX8kSY9SPbvbzce2SwEvBD7Syi8BXtGun9pu05a/qP2odipweVV9t6q+Su/MtBPbZWdV3VFV3wMub3UlSdKQ1m/52F4XSVK37deR2JLUdVX16f5T85tTgY3t+iXAHPCmVn5pO3PkuiSHtQlqNwLXzA91lOQaeh3jl405vqQZleQA4HPAj9M7avqvgG9V1Z5WZRdwVLt+FHAXQFXtSfIg8JRWfl3favvvc9eC8pMWyXE2vR/lWLduHXNzcwDs3r374evT1JUc0J0s5uh+jnNP2LPXsi7kW8q4z0pL8lx6Bw0cQm8IyTe0fSlJkqSRsRNb0lq0rk0eS1Xdk+SprfzhjqRmvsNosfJHWazDaDkLvxCPyrpDxrfucZrF3LOYGWYzd1c6tZZSVd8Hnp3kMOCjwE8Mqtb+ZpFli5UPOpNuYIdRVW0FtgJs2LChNm7cCPQ63eavT1NXckB3spij+znOXHDE5p2nb5x8oH2zHfhd4NK+svmz0i5s4+pvofeDfv9ZaSfROyvtpL6z0jbQa28+14YxeqDVOZvej25X0fuh37mMJEnSSNmJLUmP2NeOpEcXLtJhtJyFX4hH5dwT9vD2m2evqZ/F3LOYGWYz9/ZNj+9Ep9YwqupbSeaAk4HDkhzYjsY+Gri7VdsFHAPsSnIg8CTg/r7yef33WaxckvYyzrPSWvt2aFV9ppVfSm+oJDuxJUnSSM3Wt1ZJGo17kxzZjsI+ErivlS/WYbSLR77ozZfPTSCnpBmU5IeBv28d2IcA/4LeZI2fAl5JbwzrzcCV7S472u3PtOWfrKpKsgP4UJJ30JvY8Tjgs/R+WDsuybHA1+lN/vizk3p+Gs7C8VXvvPDlU0oiDTSqs9KOatcXlkuSJI2UndhrnF+wtEbNdxhdyKM7kl6f5HJ6p9A+2L7YXQ38332TOZ4CnDfhzJJmx5HAJW1c7McAV1TVHyf5MnB5krcBXwAubvUvBj6QZCe9I7BPA6iqW5NcAXwZ2AOc04YpIcnrgauBA4BtVXXr5J6epFVsX89KG/pstZUOuTZNszB81UJmHt6+jm/fX3/dId0fD3+hWXxtSFI/O7G1X+wEV9cluYzeUdRHJNlFbzzHC4ErkpwFfA14Vat+Fb2JjHbSm8zotQBVdX+StwI3tHpvmT+dVpIWqqqbgJ8cUH4HcOKA8u/wSDu0cNkFwAUDyq+i12ZJ0kqM6qy0Xe36wvqPstIh16apK2Oy7wszD29fx7fvr3/uCXt4tdtZkibKTmxJq1pVvWaRRS8aULeAcxZZzzZg2wijSZIkTctIzkprP/Q/lORk4HrgDOA9k3wikiRpbbATW5IkSZJWqQmclfY6YDtwCL0JHZ3UUZIkjZyd2JIkSZK0So37rLSquhF45v5klCRJWs5jph1AkiRJkiRJa0OSA5J8Ickft9vHJrk+ye1JPpzkoFZ+cLu9sy1f37eO81r5V5K8pK98UyvbmWTLpJ+bpPGxE1uSJEmSJEmT8gbgtr7bvwm8s6qOAx4AzmrlZwEPVNWPA+9s9UhyPHAa8AxgE/B7rWP8AOC9wEuB44HXtLqSVgGHE9FYrV844/OFL59SEkmSJEmSNE1JjgZeDlwAvDFJgBcCP9uqXAK8GbgIOLVdB/gI8Lut/qnA5VX1XeCrSXYCJ7Z6O6vqjvZYl7e6Xx7z05I0AXZiS5IkSZIkaRJ+B/hPwBPb7acA36qqPe32LuCodv0o4C6AqtqT5MFW/yjgur519t/nrgXlJw0KkeRs4GyAdevWMTc3t2zwdYfAuSfsefj2MPcZh927d0/tsc2xtK5kWa057MSWJEmSJEnSWCX5aeC+qvpcko3zxQOq1jLLFisfNGRuDSijqrYCWwE2bNhQGzduHFRtL+/54JW8/eZHutHuPH35+4zD3Nwcw+Q1x+R1JctqzWEntiRJkiRJksbtecDPJHkZ8DjgUHpHZh+W5MB2NPbRwN2t/i7gGGBXkgOBJwH395XP67/PYuWSZpyd2Joox8iWJEmSJGntqarzgPMA2pHY/7GqTk/yh8ArgcuBzcCV7S472u3PtOWfrKpKsgP4UJJ3AD8CHAd8lt4R2sclORb4Or3JH+fH2pY04+zEliRJkiRJ0rS8Cbg8yduALwAXt/KLgQ+0iRvvp9cpTVXdmuQKehM27gHOqarvAyR5PXA1cACwrapunegzkTQ2dmJLkiRJkiRpYqpqDphr1+8AThxQ5zvAqxa5/wXABQPKrwKuGmFUSR1hJ/YMc2gOSZIkSZIkSavdoJlbJUmSJEmSJEnqBDuxJUmSRijJMUk+leS2JLcmeUMrf3KSa5Lc3v4e3sqT5N1Jdia5Kclz+ta1udW/PcnmvvLnJrm53efdSTL5ZypJkiRJk2EntiRJ0mjtAc6tqp8ATgbOSXI8sAW4tqqOA65ttwFeChzXLmcDF0Gv0xs4HziJ3jiR5893fLc6Z/fdb9MEnpckSZIkTYWd2JIkSSNUVfdU1efb9YeA24CjgFOBS1q1S4BXtOunApdWz3XAYUmOBF4CXFNV91fVA8A1wKa27NCq+kxVFXBp37okSZIkadVxYkdJkqQxSbIe+EngemBdVd0DvY7uJE9t1Y4C7uq7265WtlT5rgHlgx7/bHpHbLNu3Trm5uYA2L1798PXp6krOWD0Wc49Yc9et4ddd1e2iTkWz7HS/60kSZJWzk5sSZKkMUjyBOC/A79SVX+7xLDVgxbUCsofXVi1FdgKsGHDhtq4cSPQ63Sbvz5NXckBo89y5paP7XX7ztOHW3dXtok5Fs+x0v+tJEmSVs7hRCRJkkYsyWPpdWB/sKr+qBXf24YCof29r5XvAo7pu/vRwN3LlB89oFySJEmSViU7sSVJkkYovUOuLwZuq6p39C3aAWxu1zcDV/aVn5Gek4EH27AjVwOnJDm8Teh4CnB1W/ZQkpPbY53Rty5JkiRJWnUcTkSSJGm0ngf8PHBzki+2sv8MXAhckeQs4GvAq9qyq4CXATuBbwOvBaiq+5O8Fbih1XtLVd3frr8O2A4cAny8XSRJkiRpVRpbJ3aSO4GHgO8De6pqQ5InAx8G1gN3Aq+uqgfaUUTvovcF7tvAmVX1+baezcCvt9W+raouGVdmSZKk/VVVf8bgcasBXjSgfgHnLLKubcC2AeU3As/cj5iSJEmSNDPGPZzIC6rq2VW1od3eAlxbVccB17bbAC8FjmuXs4GLAFqn9/nAScCJwPntdFpJkiRJkiRJ0how6TGxTwXmj6S+BHhFX/ml1XMdcFib8OglwDVVdX9VPQBcA2yacGZJkiRJkiRJ0pSMc0zsAv4kSQH/taq2AuvaZERU1T1JntrqHgXc1XffXa1ssfK9JDmb3hHcrFu3jrm5uRE/lfHZvXv30HnPPWHPkstX8rwXrnPhOhZ7zHWHDF427P0Xqz9O+7Ktu2RWc0uSJEmSJEmjMM5O7OdV1d2to/qaJH+xRN1B40bWEuV7F/Q6yLcCbNiwoTZu3LiCuNMxNzfHsHnP3PKxJZffefpw61lqnQvXsdhjnnvCHt5+86NfPsPef7H647Qv27pLZjX3LBjV2P2SJEmSJEkan7ENJ1JVd7e/9wEfpTem9b1tmBDa3/ta9V3AMX13Pxq4e4lySRqV/Rq7X5IkaVYluTPJzUm+mOTGVvbkJNckub39PbyVJ8m7k+xMclOS5/StZ3Orf3uSzdN6PpIkafUaSyd2kscneeL8deAU4BZgBzC/U7MZuLJd3wGc0XaMTgYebMOOXA2ckuTwtvN0SivTAOu3fGyvi6QV2dex+yVJkmbZfv2g385iOx84id6BS+fPd3xLkiSNyriGE1kHfLR39j0HAh+qqk8kuQG4IslZwNeAV7X6V9E7RX8nvdP0XwtQVfcneStwQ6v3lqq6f0yZ1QGDOt/vvPDlU0iiNWIUY/ffM8nAkiRJY3YqsLFdvwSYA95E3w/6wHVJ5n/Q3whcM/89Lck1wCbgssnGlva28Lul3yslabaNpRO7qu4AnjWg/G+AFw0oL+CcRda1Ddg26oySxGjG7t+70gonml1uEtSVWmwS1q6bxdyzmBlmM7cT3krSyIziB/3FyiVJkkZmnBM7SlKn9Y/dn2Svsfvbl7Zhxu5fuM4VTTS73CSoK7XYJKxdN4u5ZzEzzGbu7Zse74S3kjQao/hBf6w/9E/TLP5oauZHLPyRfuFjLLd8qfWtO2T5+l0zi68NSeo3W99aJWlE2nj9j6mqh/rG7n8Lj4zdfyGPHrv/9Ukupzfm4/zY/ZIkSTNpRD/o7+KR4Ufmy+cGPNaKfuifprm5uZn70dTMj1h4kMidp2/cp+VLre/cE/bw6ils5/0ZIqULr40kxwCXAv8A+AGwtare1cbW/zCwHrgTeHVVPZDeGLXvojf87LeBM6vq821dm4Ffb6t+W1Vd0sqfC2wHDqE3dO0b2tn/kmbcWCZ2lKQZsA74syRfAj4LfKyqPkGv8/rFSW4HXtxuQ28H6A56Y/e/D/jlyUeWJEkajSSPT/LE+ev0ftC/hUd+0IdH/6B/RnpO5pEf9K8GTklyeJvQ8ZRWJkkL7QHOraqfAE4GzklyPKOdUPaiVnf+fpsm8LwkTYBHYktak0Y5dr8kSdIMWgd8tHegIwcCH6qqTyS5AbgiyVnA14BXtfpX0Tsacie9IyJfC1BV9yd5K3BDq/eW+UkeJalf++Frfsz9h5LcRm8M/ZFMKJtkDji0qj7Tyi8FXgF8fBLPT9J42YktSZIkSWvMKH/Qr6ptwLZRZ5RGaeFQHKOur32TZD3wk8D1jG5C2aPa9YXlgx5/n8fpXzgh+rTGGO/K+ObmeLSuZFmtOezEliRJkiRJ0kQkeQLw34Ffqaq/bWeEDKw6oGypCWWHmmgWVjZO/3s+eOVeE6IvN476uHRhfHNzDNaVLKs1h2NiS5IkjViSbUnuS3JLX9mTk1yT5Pb29/BWniTvTrIzyU1JntN3n82t/u1tAqP58ucmubnd591Z4tufJElds37Lxx6+3Pz1B6cdRxOU5LH0OrA/WFV/1IrvbcOEsA8Tyi5WfvSAckmrgJ3YkiRJo7edR08k5KRFkiSNSX/HuEOBdFP70f1i4LaqekffopFMKNuWPZTk5PZYZ/StS9KMsxNbkiRpxKrq08DCic1OpTdZEe3vK/rKL62e64D5SYteQpu0qKoeAOYnLTqSNmlRG6P20r51SZIkddXzgJ8HXpjki+3yMuBC4MVJbgde3G5Db0LZO+hNKPs+4JehN6EsMD+h7A3sPaHs64D3t/v8FU7qKK0ajok9Reu3fIxzT9jDmYv8SnznhS8fy2NKkqSpmPikRZIkSV1RVX/G4HGrYUQTylbVjcAz9yOmpI6yE1uSJGm6xjZpUZKz6Q07wrp16x6eHXy1zli+P0ad5dwT9ux1e9h1d2WbmGPxHCv930raPwsPyBrHQV/TthaeoyStlJ3YkiRJk3FvkiPbUdjDTlq0cUH5HPswaVFVbQW2AmzYsKHmZwdfrTOW749RZ1l4pt2dpw+37q5sE3MsnmOl/1tJw/MMYknSQo6JLUmSNBlOWiRJkiRJK+CR2JIkSSOW5DJ6R1EfkWQXcD69SYquSHIW8DXgVa36VcDL6E1A9G3gtdCbtCjJ/KRF8OhJi7YDh9CbsMhJiyRJkiStWnZiS5IkjVhVvWaRRU5aJElacxweRJK0v+zEliRJkiRJK+aEhJKkcbMTW5IkSTPFzhJJmqxxH0ntkdqSpOXYiS1JkiStQf4YIEmSpFlhJ7Y6zy9YkiRJkiRJ0tplJ7YkSZIkSXqYw3tIkrrGTmxJkiRJklapYTqkPdtVktR1dmJLkiRJkqSZ4tHikrS22Ik9Ro7lPB5uV0mSJEnq6f9+dO4JezjTzt0VGUenuN9dJWl07MSWJEmSJGlGzMIRyLOQUZI0W+zE1qrjr92SJEmSJEnS6mEntiRJkiRJHeFBOZIkPZqd2JIkSZIkTUkXht7oQgZJkpZiJ7YkSZIkSWNiB7EW42tDkoZnJ7ZWvfkdg/mZuj0dT5IkSZI0SuPokF5unQ49I2ktsRNbe/GXYEmSJEkabJjvS6PuSPQ7miRJM9CJnWQT8C7gAOD9VXXhlCMtyp0LafWbpTZJ0upmeySpS1ZLm9T/nW7+TM79WYfUz9fGZKyW9kjS3jrdiZ3kAOC9wIuBXcANSXZU1Zenm0yL2dcP5Wl8iC93ytVymTxFa+2yTZLUFbZHWitu/vqDe3Uiuh/WTbPcJtmpKK0us9weSVpapzuxgROBnVV1B0CSy4FTgZlsfCbdwesO2WQM2s5+wVq1VlWbJGmmzXR75HwV0qrTmTbJ70CaFF9rndWZ9mi1cj9O05KqmnaGRSV5JbCpqn6h3f554KSqev2CemcDZ7ebTwe+MtGg++cI4JvTDrECs5h7FjPD7OX+0ar64WmHGIdh2qQOtkez9vqZN4u5ZzEzzGbuYTOv6faolS/WJnXl/96VHNCdLObY22rKsabbpA7uIw2jK6+/fWHmyZj1zGu6PWrlK2mTuvJ/N8feupIDupNl1nIM1SZ1/UjsDCh7VK97VW0Fto4/zuglubGqNkw7x76axdyzmBlmN/cqtWyb1LX2aFZfP7OYexYzw2zmnsXMY7Bf+0hd2YZdyQHdyWIOc8yomdtHGsYs/t/NPBlm7rSx9SN1ZRuao5s5oDtZVmuOx4xqRWOyCzim7/bRwN1TyiJJtkmSusL2SFKX2CZJ6grbI2mV6non9g3AcUmOTXIQcBqwY8qZJK1dtkmSusL2SFKX2CZJ6grbI2mV6vRwIlW1J8nrgauBA4BtVXXrlGON2kydUtdnFnPPYmaY3dyrzoy2SbP6+pnF3LOYGWYz9yxmHqkRtEdd2YZdyQHdyWKOvZljBszoPtIwZvH/bubJMHNHjbk96so2NMfeupIDupNlVebo9MSOkiRJkiRJkqS1revDiUiSJEmSJEmS1jA7sSVJkiRJkiRJnWUn9pQkeVySzyb5UpJbk/zGtDMNK8kBSb6Q5I+nnWVYSe5McnOSLya5cdp5hpHksCQfSfIXSW5L8s+mnUndlWRTkq8k2Zlky4Dlb0zy5SQ3Jbk2yY9OI+dCy+Xuq/fKJJVkwyTzLZJl2cxJXt22961JPjTpjIMM8Rr5h0k+1dr3m5K8bBo5+/JsS3JfklsWWZ4k727P56Ykz5l0xlky6HMwyZOTXJPk9vb38DE87qP+j4s97jj/p4vkeHOSr7dt8sX+13yS81qOryR5yQhzHNPeZ7e19uENrXyi22SJHBPdJllkfzi9ybiub9vjw+lNzEWSg9vtnW35+jHn2J7kq33b49mt3PZnFdrX92dXZMF3s8XeP12RAd9xZmAb/2p7TdyS5LLWZnRuOy/yWTfxz9zVIMvvN4/l82gFOc5M8o2+z6lfGFOOTuyXD5FjY5IH+7bH/zWGDAM/KxbUGfv2GDLH2LdHe5xl+zdH9p6pKi9TuAABntCuPxa4Hjh52rmGzP5G4EPAH087yz5kvhM4Yto59jHzJcAvtOsHAYdNO5OXbl7oTVjyV8CPtdfKl4DjF9R5AfBD7frrgA/PQu5W74nAp4HrgA1dzwwcB3wBOLzdfuosbGt6k268rl0/HrhzypmfDzwHuGWR5S8DPt4+T08Grp/2du7yZdDnIPBbwJZ2fQvwm5P4Py72uOP8ny6S483AfxxQ9/j2HjkYOLa9dw4YUY4jgee0608E/rI93kS3yRI5JrpNWGR/GLgCOK2V/35f2/TLwO+366cxos+yJXJsB145oL7tzyq87Ov7sysXFnw3W+z905ULA77jdHkbA0cBXwUO6du+Z3ZxO9ORz9xZvzDcfvNYPo9WkONM4Hen8dpasHwir6chcmxkzP1Ui31WTHp7DJlj7NujPc6y/Zujes94JPaUVM/udvOx7dL5WTaTHA28HHj/tLOsZkkOpddAXwxQVd+rqm9NN5U67ERgZ1XdUVXfAy4HTu2vUFWfqqpvt5vXAUdPOOMgy+Zu3kpvB/w7kwy3iGEy/yLw3qp6AKCq7ptwxkGGyV3Aoe36k4C7J5jvUarq08D9S1Q5Fbi0fZ5eBxyW5MjJpFs1TqXXmUD7+4pRP8Ai/8fFHnds/9MhXk8L811eVd+tqq8CO+m9h0aR456q+ny7/hBwG70OkolukyVyLGYs22SJ/eEXAh9p5Qu3x/x2+gjwoiQZY47F2P6sQit4f07dwu9m7f2w2Ptn6pb4jtPZbdwcCByS5EDgh4B76OB27spn7iowzH7zWD6PVpBjIrqyX76P+3NjMeQ+1Ni3xwr25cZmyP2okbxn7MSeovRO/foicB9wTVVdP+1MQ/gd4D8BP5h2kH1UwJ8k+VySs6cdZgg/BnwD+IP0Tg98f5LHTzuUOuso4K6+27tY+gPsLHq/DE/bsrmT/CRwTFV1ZfiiYbb104CnJfnzJNcl2TSxdIsbJvebgZ9Lsgu4Cvj3k4m2Yvv6ul/rBn0Orquqe6C3Iww8dUJZFnvcafxPX99O89yWR05fn0iOdhrlT9I7WmVq22RBDpjwNlm4P0zvqLNvVdWeAY/1cI62/EHgKePI0bdffkHbHu9McvDCHAMyahUY8v3ZBQu/mz2Fxd8/XbDYd5zObuOq+jrw28DX6HVePwh8jm5v535d+sydFcNsm7F9Hu1jDiUiq6AAACAASURBVIB/3T6nPpLkmBFnGFaXXk//rA1r8fEkzxjnAw3Yh5o30e2xRA6Y0PYYon9zJO8ZO7GnqKq+X1XPpndE5IlJnjntTEtJ8tPAfVX1uWlnWYHnVdVzgJcC5yR5/rQDLeNAeqfJXFRVPwn8Hb3Tv6RBBv2COfAIsiQ/B2wA/stYEw1nydxJHgO8Ezh3YomWN8y2PpDekCIbgdcA709y2JhzLWeY3K8BtlfV0fROgftA+x901dCvewGz8Tk46f/pRcA/Ap5Nr2Pi7ZPKkeQJwH8HfqWq/napquPMMiDHxLfJwv1h4CeWeKyJ5Wj75ecB/xj4p8CTgTeNO4embx/en1O1yHezrr82Z+47Tvsx71R6Qyn9CPB4ep+lC3VpOw+j66+VaRpm20xi+w3zGP8vsL6q/gnwpzxypOukdeX19HngR6vqWcB7gP8xrgda5rNiYttjmRwT2x5D9G+OZJt0+cvpmtFOoZoDunC03lKeB/xMkjvpncrywiT/bbqRhlNVd7e/9wEfZUSnBY/RLmBX369XH6G3wycNsgvo/9X9aAYMBZHkXwC/BvxMVX13QtmWslzuJwLPBOZau3MysCPTndxxmG29C7iyqv6+nXb/FXqd2tM0TO6z6I3vSFV9BngccMRE0q3MUK979SzyOXjv/KmN7e+khr5Z7HEn+j+tqnvbDvcPgPfxyL7BWHMkeSy9LxsfrKo/asUT3yaDckxrm7THnt8fPpneabcHDnish3O05U9ixKcV9++Xt1N1q31m/gET3B6ajn18f07bo76b0Tsye7H3Txcs9h2nq9sY4F8AX62qb1TV3wN/BPwU3d7O/TrxmTtjht3fH+vn0TA5qupv+r7XvQ947ogzDKsTr6eq+tv5YS2q6irgsUlG/n1mkc+KfhPZHsvlmNT2WPCYi/VvjuQ9Yyf2lCT54fkj85IcQu/D8S+mm2ppVXVeVR1dVevpDcT+yar6uSnHWlaSxyd54vx14BRg4Gy2XVFVfw3cleTprehFwJenGEnddgNwXHqzpB9E7/25o79CG5bjv9LrwO7KF4Mlc1fVg1V1RFWtb+3OdfTy3ziduMAQ25reL9wvAGg7CU8D7phoykcbJvfX6LU1JPkJep3Y35hoyn2zAzgjPScDD86fLqu9LfE5uAPY3KptBq6cUKTFHnei/9PsPTbhv+KRfYMdwGnpzaJ+LL0foT47oscMvbFgb6uqd/Qtmug2WSzHpLfJIvvDtwGfAl7Zqi3cHvPb6ZX09kX3+8imxfbL+zp+Qm8c2f7tYfuzyqzg/TlVi3w3O53F3z9Tt8R3nE5u4+ZrwMlJfqi9RuYzd3Y7L9CJz9wZM8x+81g+j/Y1x4LP7Z+h9xk6DZ14PSX5B+19SpIT6fV5/s2IH2Oxz4p+Y98ew+SYxPZo6x6mf3M075ka8yyVXhadvfOfAF8AbqK3Q/x/TTvTPubfyARmOR1R1h+jN5Pvl4BbgV+bdqYhcz8buLG9Rv4HcPi0M3np7oXe8A9/SW8s0V9rZW+h1+kLvdPL7gW+2C47pp15mNwL6s4BG7qemd6pUu+g9wXnZtrM9dO+DJH7eODPW1v5ReCUKee9jN5wBn9P75f7s4BfAn6pbzu/tz2fm7vw2ujqZbHPQXrj0F0L3N7+PnlC/8eBjzvO/+kiOT7QHucmejvWR/bV/7WW4yvAS0eY45/TO3Xypr72+GWT3iZL5JjoNmGR/eH2mv0svQkk/xA4uJU/rt3e2Zb/2JhzfLJtj1uA/wY8YdyvVS/Tu+zr+7NLF/q+my32/unKhQHfcbq+jYHfoNchc0trJw/u4namI5+5q+HC8vvNY/k8WkGO/4fevt2X6P2w8o8n+Nqa+H75EDle37c9rgN+agwZFvusmOj2GDLH2LdHe5zF9qNG/p5JW5kkSZIkSZIkSZ3jcCKSJEmSJEmSpM6yE1uSJEmSJEmS1Fl2YkuSJEmSJEmSOstObEmSJEmSJElSZ9mJLUmSpiLJtiT3JblliLrPT/L5JHuSvLKv/NlJPpPk1iQ3Jfk3400tSZIkSZo0O7ElSdK0bAc2DVn3a8CZwIcWlH8bOKOqntHW9TtJDhtVQEmSJEnS9B047QCSJGltqqpPJ1nfX5bkHwHvBX6YXgf1L1bVX1TVnW35Dxas4y/7rt+d5L5232+NNbwkSZIkaWLsxJYkSV2yFfilqro9yUnA7wEvHOaOSU4EDgL+aoz5JEmSJEkTZie2JEnqhCRPAH4K+MMk88UHD3nfI4EPAJur6gfL1ZckSZIkzQ47sSVJUlc8BvhWVT17X+6U5FDgY8CvV9V1Y0kmSZIkSZoaJ3aUJEmdUFV/C3w1yasA0vOspe6T5CDgo8ClVfWHE4gpSZIkSZqwVNW0M0iSpDUoyWXARuAI4F7gfOCTwEXAkcBjgcur6i1J/im9zurDge8Af11Vz0jyc8AfALf2rfrMqvrixJ6IJEmSJGms7MSWJEmSJEmSJHWWw4lIkiRJkiRJkjrLTmxJkiRJkiRJUmfZiS1JkiRJkiRJ6iw7sSVJkiRJkiRJnWUntiRJkiRJkiSps+zEliRJkiRJkiR1lp3YkiRJkiRJkqTOshNbkiRJkiRJktRZdmKrM5JsTfKVJD9IcuaA5b+a5K+TPJhkW5KDpxBT0hqxVJuU5JlJrk7yzSQ1pYiS1ohl2qPNST6X5G+T7EryW0kOnFJUSWvAMm3SaW3Zg0nuS3JJkkOnFFXSGpLkaUmuTPKNJPe372tPn3YujY6d2OqSLwG/DHx+4YIkLwG2AC8C1gM/BvzGJMNJWnMWbZOAvweuAM6aaCJJa9VS7dEPAb8CHAGcRG9f6T9OLpqkNWipNunPgedV1ZPofWc7EHjbBLNJWrsOA3YATwfWAZ8FrpxqIo2UndgaqyR3Jvk/ktyU5O+SXJxkXZKPJ3koyZ8mORygqt5bVdcC3xmwqs3AxVV1a1U9ALwVOHNyz0TSajCqNqmqvlJVFwO3Tvo5SFodRtgeXVRV/7OqvldVXwc+CDxvwk9H0owbYZt0V1V9s6/o+8CPT+hpSFqFhm2fquqzVXVxVd1fVX8PvBN4epKnTPs5aDTsxNYk/GvgxcDTgH8JfBz4z/SOGHoM8B+GWMcz6P3iP+9LwDobI0krMIo2SZJGYRzt0fPxBzZJKzOSNinJP0/yIPBQW+fvjCWtpLVkJe3T84G/rqq/mVRIjZfj5WkS3lNV9wIk+Z/AfVX1hXb7o/ROe13OE4AH+27PX38iYIMkaV+Mok2SpFEYaXuU5LXABuAXRh1U0powkjapqv4MeFKSo4BfBO4cT1xJa8g+tU9JjgbeC7xx0kE1Ph6JrUm4t+/6/x5w+wlDrGM30D8hyPz1h/YvmqQ1aBRtkiSNwsjaoySvAC4EXrrgVH5JGtZI95HaEEefAC7f/2iS1rih26ckPwz8CfB7VXXZZOJpEuzE1qy4FXhW3+1nAfd6WogkSVrrkmwC3gf8y6q6edp5JKnPgcA/mnYISWtDG7v/T4AdVXXBtPNotOzEVmckOSjJ44AAj03yuCTzr9FLgbOSHN8apV8Htk8pqqQ1YKk2KT2PAw5qtx+X5OApxpW0ii3THr2Q3mSO/7qqPjvNnJLWhmXapNOT/MO2r/SjwAXAtdPMK2ltSHIocDXw51W1Zdp5NHp2YqtL/oTeaSA/BWxt158PUFWfAH4L+BTwv9rl/OnElLRGLNomAT/abs9Pnva/ga9MOqCkNWOp9uj/BJ4EXJVkd7t8fDoxJa0RS7VJxwP/H73hIP+c3v7RL04ho6S1518B/xR4bd8+0e4k/3DawTQaqappZ5AkSZIkSZIkaSCPxJYkSZIkSZIkdZad2JIkSZIkSZKkzrITW5IkSZIkSZLUWct2YifZluS+JLf0lf2XJH+R5KYkH01yWN+y85LsTPKVJC/pK9/UynYm2dJXfmyS65PcnuTDSQ5q5Qe32zvb8vWjetKSZpdtkqSusD2S1CW2SZK6wvZI0jgMcyT2dmDTgrJrgGdW1T8B/hI4DyDJ8cBpwDPafX4vyQFJDgDeC7yU3mzFr2l1AX4TeGdVHQc8AJzVys8CHqiqHwfe2epJ0nZskyR1w3ZsjyR1x3ZskyR1w3ZsjySN2IHLVaiqTy/89aqq/qTv5nXAK9v1U4HLq+q7wFeT7ARObMt2VtUdAEkuB05NchvwQuBnW51LgDcDF7V1vbmVfwT43SSpqloq7xFHHFHr169fqgoAf/d3f8fjH//4ZeuNmzm6mQO6k2XWcnzuc5/7ZlX98LhyzFKbNGx7BN35Pw9itpXrcr61kM32aG+zto/Ur2uZupYHupepa3lg+plskx6xWvaR5nU9Y9fzgRlHxe9sa28faaV8Lt21mp7PqNukZTuxh/BvgQ+360fRa4zm7WplAHctKD8JeArwraraM6D+UfP3qao9SR5s9b+5VJj169dz4403Lht6bm6OjRs3Lltv3MzRzRzQnSyzliPJ/xp/miV1pk0atj2C7vyfBzHbynU531rIZnsESc4GzgZYt24dv/3bv71s6N27d/OEJzxh2XqT1LVMXcsD3cvUtTww/UwveMEL1nSbtJL2CKb/fxtG1zN2PR+YcVSGzbjW26OFZq0faRR8Lt21mp7PqPuR9qsTO8mvAXuAD84XDahWDB62pJaov9S6BuXYa4dobm5u8dDN7t27h6o3buboZg7oThZzDK8LbdJK2iPo9vY128p1OZ/ZxqsL7RFAVW0FtgJs2LChhtmJ7OKOc9cydS0PdC9T1/JANzNNShfapJW0RzAb/7euZ+x6PjDjqMxCxi60Ry3HzPYjjYLPpbtW0/MZ9XNZcSd2ks3ATwMv6js1YxdwTF+1o4G72/VB5d8EDktyYPsVrb/+/Lp2JTkQeBJw/6Ass/wFzRzdzAHdyWKO4XSlTVqNX9DMtnJdzme28elKeyRJYJskqTu61B7Ncj/SKPhcums1PZ9RP5dhJnZ8lCSbgDcBP1NV3+5btAM4rc0IeyxwHPBZ4AbguDaD7EH0Bu3f0RqtT/HIWEibgSv71rW5XX8l8MnlxjGStDbZJknqCtsjSV1imySpK2yPJO2vZY/ETnIZsBE4Isku4Hx6s8geDFyTBOC6qvqlqro1yRXAl+mdHnJOVX2/ref1wNXAAcC2qrq1PcSbgMuTvA34AnBxK78Y+EAb1P9+eg2WpDXONklSV9geSeoS2yRJXWF7JGkclu3ErqrXDCi+eEDZfP0LgAsGlF8FXDWg/A4emXm2v/w7wKuWyydpbbFNktQVtkeSusQ2SVJX2B5JGocVDSciSZIkSZIkSdIkrHhix1l389cf5MwtH3v49p0XvnyKaSStdbZJkrrC9khSl9gmSeoK2yNpujwSW5IkSZIkSZLUWXZiS5IkSZIkSZI6y05sSZIkSZIkSVJn2YktSZIkSZIkSeosO7ElSZIkSZIkSZ1lJ7YkSZIkSZIkqbPsxJYkSZIkSZIkdZad2JIkSZIkSZKkzrITW5IkSZIkSZLUWXZiS5IkSZIkSZI6y05sSZIkSZIkSVJn2YktSZIkSZIkSeosO7ElSZIkSZIkSZ1lJ7YkSZIkSZIkqbPsxJYkSZIkSZIkdZad2JIkSZIkSZKkzrITW5IkSZIkSZLUWXZiS5IkSZIkSZI6a9lO7CTbktyX5Ja+sicnuSbJ7e3v4a08Sd6dZGeSm5I8p+8+m1v925Ns7it/bpKb233enSRLPYaktc02SVJX2B5J6hLbJEldYXskaRyGORJ7O7BpQdkW4NqqOg64tt0GeClwXLucDVwEvYYEOB84CTgROL+vMbmo1Z2/36ZlHkPS2rYd2yRJ3bAd2yNJ3bEd2yRJ3bAd2yNJI7ZsJ3ZVfRq4f0HxqcAl7folwCv6yi+tnuuAw5IcCbwEuKaq7q+qB4BrgE1t2aFV9ZmqKuDSBesa9BiS1jDbJEldYXskqUtskyR1he2RpHE4cIX3W1dV9wBU1T1JntrKjwLu6qu3q5UtVb5rQPlSj/EoSc6m9ysc69atY25ubvkncAice8Keh28Pc59x2L1799Qe2xxL60oWcwylM23SStoj6E6bNEiX//ddzgbdzme2selMewSzvY/Ur2uvia7lge5l6loe6GamCehMm7Qa95Hmdf211fV8YMZR6XjGzrRHkmbTSjuxF5MBZbWC8n1SVVuBrQAbNmyojRs3Lnuf93zwSt5+8yNP/87Tl7/POMzNzTFMXnNMXleymGO/TLxNWkl7BN1pkwbp8v++y9mg2/nMNnHuI+2Hrr0mupYHupepa3mgm5mmyH2kEer6a6vr+cCMozILGQeYyj7Savmhf6U6/oPHPllNzwVW1/MZ9XNZaSf2vUmObL9sHQnc18p3Acf01TsauLuVb1xQPtfKjx5Qf6nHkKSFbJMkdYXtkaQusU2S1BWdao9Wyw/9KzWjP3gMtJqeC6yu5zPq5zLMxI6D7ADmZ4bdDFzZV35Gm132ZODBdirH1cApSQ5vA/GfAlzdlj2U5OQ2m+wZC9Y16DEkaSHbJEldYXskqUtskyR1he2RpP2y7JHYSS6j9+vXEUl20Zsd9kLgiiRnAV8DXtWqXwW8DNgJfBt4LUBV3Z/krcANrd5bqmp+kP/X0Zu59hDg4+3CEo8haQ2zTZLUFbZHkrrENklSV9geSRqHZTuxq+o1iyx60YC6BZyzyHq2AdsGlN8IPHNA+d8MegxJa5ttkqSusD2S1CW2SZK6wvZI0jisdDgRSZIkSZIkSZLGzk5sSZIkSZIkSVJn2YktSZIkSZIkSeosO7ElSZIkSZIkSZ1lJ7YkSZIkSZIkqbPsxJYkSZIkSZIkdZad2JIkSZIkSZKkzrITW5IkSZIkSZLUWXZiS5IkSZIkSZI6y05sSZIkSZIkSVJn2YktSZIkSZIkSeosO7ElSZIkSZIkSZ1lJ7YkSZIkSZIkqbPsxJYkSZIkSZIkdZad2JIkSZIkSZKkzrITW5IkSZIkSZLUWXZiS5IkSZIkSZI6y05sSZIkSZIkSVJn2YktSZIkSZIkSeosO7ElSZIkSZIkSZ21X53YSX41ya1JbklyWZLHJTk2yfVJbk/y4SQHtboHt9s72/L1fes5r5V/JclL+so3tbKdSbbsT1ZJq59tkqSusD2S/n/27j9a0ruuE/z7s2kDAYUEGXqZJGvCsXUMIAp9QtRdt8c4IaBr8Ag7yWGGVjPTRw2KM5zVMJ6drApnZBXRsIibQyIJmyFiBk1GgzEG7jrrkPArmBAipg0Z0iYSNCHSIGLjd/+o54ZK53bf21X33vpW1et1Tp1bz7eep+rzrXryztOfqnqKnsgkoBfyCJjUxE3sqjo5yU8k2d1ae06S45Kcn+QNSd7UWtuV5OEkFw6bXJjk4dba1yd507BequqMYbtnJzk3ya9V1XFVdVyStyR5cZIzklwwrAvwODIJ6IU8Anoik4BeyCNgGtOeTmRHkhOqakeSJyV5IMl3Jbl2uP3KJC8drp83LGe4/eyqqmH8mtba37XWPplkf5Izh8v+1to9rbUvJblmWBfgSGQS0At5BPREJgG9kEfARHZMumFr7S+q6peSfCrJ3yb5gyQfTvLZ1tqhYbUDSU4erp+c5L5h20NV9UiSrx3Gbxm76/Ft7jts/IVr1VJV+5LsS5KdO3dmZWVl3fp3npC85rmHHl3eyDZb4eDBgzN7bHUcXS+1qGNjesmkSfIo6SeT1tLza99zbUnf9alt6/SSR8l8HyON622f6K2epL+aeqsn6bOm7dBLJi3iMdKq3vet3utL1LhZeq+xlzwC5tPETeyqOimjd7ROT/LZJL+V0Vc2DtdWNznCbUcaX+tT4m2NsbTWLktyWZLs3r277dmz52ilJ0nefPV1eeMdX5n+va9Yf5utsLKyko3Uq47t10st6tiYXjJpkjxK+smktfT82vdcW9J3fWrbOr3kUTLfx0jjetsneqsn6a+m3upJ+qxpO/SSSYt4jLSq932r9/oSNW6W3mvsJY+GWhbijf5J9f6Gx7FYpLkkizWfzZ7LxE3sJN+d5JOttc8kSVW9O8m3JzmxqnYM76KdkuT+Yf0DSU5NcmD42shTkzw0Nr5qfJsjjQMcTiYBvZBHQE9kEtCLbvJoUd7on1Tvb3gci0WaS7JY89nsuUxzTuxPJTmrqp40nJPo7CQfT/K+JC8b1tmb5Lrh+vXDcobb39taa8P4+cOvzp6eZFeSDyT5YJJdw6/UHp/RSfuvn6JeYLHJJKAX8gjoiUwCeiGPgIlNc07sW6vq2iQfSXIoyW0ZvYv1e0muqarXDWOXD5tcnuQdVbU/o3fOzh/u586qeldGwXUoyUWttS8nSVW9KsmNGf1i7RWttTsnrRdYbDIJ6IU8Anoik4BeyCNgGtOcTiSttUuSXHLY8D0Z/SLs4et+McnLj3A/r0/y+jXGb0hywzQ1AstDJgG9kEdAT2QS0At5BExqmtOJAAAAAADAltLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuTdXErqoTq+raqvrTqrqrqr6tqp5WVTdV1d3D35OGdauqLq2q/VV1e1U9f+x+9g7r311Ve8fGX1BVdwzbXFpVNU29wGKTSUAv5BHQE5kE9EIeAZOa9pPYv5rk91tr/yTJ85LcleTiJDe31nYluXlYTpIXJ9k1XPYleWuSVNXTklyS5IVJzkxyyWpgDevsG9vu3CnrBRabTAJ6IY+AnsgkoBfyCJjIxE3sqnpKku9McnmStNa+1Fr7bJLzklw5rHZlkpcO189LclUbuSXJiVX1zCQvSnJTa+2h1trDSW5Kcu5w21Naa+9vrbUkV43dF8BjyCSgF/II6IlMAnohj4Bp7Jhi22cl+UyS36iq5yX5cJJXJ9nZWnsgSVprD1TVM4b1T05y39j2B4axo40fWGP8capqX0bvtGXnzp1ZWVlZt/idJySvee6hR5c3ss1WOHjw4MweWx1H10st6tiwLjJpkjxK+smktfT82vdcW9J3fWrbUl3kUTLfx0jjetsneqsn6a+m3upJ+qxpm3SRSYt4jLSq932r9/oSNW6WOaixizxKFucYaVJzsK9s2CLNJVms+Wz2XKZpYu9I8vwkP95au7WqfjVf+crHWtY6D1GbYPzxg61dluSyJNm9e3fbs2fPUcoYefPV1+WNd3xl+ve+Yv1ttsLKyko2Uq86tl8vtahjw7rIpEnyKOknk9bS82vfc21J3/WpbUt1kUfJfB8jjettn+itnqS/mnqrJ+mzpm3SRSYt4jHSqt73rd7rS9S4Weagxi7yKFmcY6RJzcG+smGLNJdkseaz2XOZ5pzYB5IcaK3dOixfm1EYfXr4CkeGvw+OrX/q2PanJLl/nfFT1hgHWItMAnohj4CeyCSgF/IImNjETezW2l8mua+qvnEYOjvJx5Ncn2T1l2H3JrluuH59klcOvy57VpJHhq+L3JjknKo6aTgR/zlJbhxu+1xVnTX8muwrx+4L4DFkEtALeQT0RCYBvZBHwDSmOZ1Ikvx4kqur6vgk9yT5oYwa4++qqguTfCrJy4d1b0jykiT7k3xhWDettYeq6ueTfHBY7+daaw8N1380yduTnJDkPcMF4EhkEtALeQT0RCYBvZBHwESmamK31j6aZPcaN529xrotyUVHuJ8rklyxxviHkjxnmhqB5SGTgF7II6AnMgnohTwCJjXNObEBAAAAAGBLaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLembmJX1XFVdVtV/e6wfHpV3VpVd1fVb1bV8cP4E4bl/cPtp43dx2uH8U9U1YvGxs8dxvZX1cXT1gosPpkE9EIeAT2RSUAv5BEwic34JPark9w1tvyGJG9qre1K8nCSC4fxC5M83Fr7+iRvGtZLVZ2R5Pwkz05ybpJfGwLtuCRvSfLiJGckuWBYF+BoZBLQC3kE9EQmAb2QR8Axm6qJXVWnJPmeJG8blivJdyW5dljlyiQvHa6fNyxnuP3sYf3zklzTWvu71tonk+xPcuZw2d9au6e19qUk1wzrAqxJJgG9kEdAT2QS0At5BExqx5Tb/0qSn0ryNcPy1yb5bGvt0LB8IMnJw/WTk9yXJK21Q1X1yLD+yUluGbvP8W3uO2z8hWsVUVX7kuxLkp07d2ZlZWXdwneekLzmuYceXd7INlvh4MGDM3tsdRxdL7Wo45jMPJMmyaOkn0xaS8+vfc+1JX3Xp7YtN/M8Sub7GGlcb/tEb/Uk/dXUWz1JnzVto5ln0iIeI63qfd/qvb5EjZtlHmpMB3mULM4x0qTmZF/ZkEWaS7JY89nsuUzcxK6q703yYGvtw1W1Z3V4jVXbOrcdaXytT4m3NcbSWrssyWVJsnv37rZnz561VnuMN199Xd54x1emf+8r1t9mK6ysrGQj9apj+/VSizo2ppdMmiSPkn4yaS09v/Y915b0XZ/atk4veZTM9zHSuN72id7qSfqrqbd6kj5r2g69ZNIiHiOt6n3f6r2+RI2bpfcae8mjZHGOkSbV+75yLBZpLslizWez5zLNJ7G/I8n3VdVLkjwxyVMyekftxKraMbyLdkqS+4f1DyQ5NcmBqtqR5KlJHhobbTbizwAAIABJREFUXzW+zZHGAQ4nk4BeyCOgJzIJ6IU8AiY28TmxW2uvba2d0lo7LaMT6r+3tfaKJO9L8rJhtb1JrhuuXz8sZ7j9va21NoyfP/zq7OlJdiX5QJIPJtk1/Ert8cNjXD9pvcBik0lAL+QR0BOZBPRCHgHTmPac2Gv56STXVNXrktyW5PJh/PIk76iq/Rm9c3Z+krTW7qyqdyX5eJJDSS5qrX05SarqVUluTHJckitaa3duQb3AYpNJQC/kEdATmQT0Qh4B69qUJnZrbSXJynD9nox+Efbwdb6Y5OVH2P71SV6/xvgNSW7YjBqB5SGTgF7II6AnMgnohTwCjtXEpxMBAAAAAICtpokNAAAAAEC3NLEBAAAAAOiWJjYAAAAAAN3SxAYAAAAAoFua2AAAAAAAdEsTGwAAAACAbmliAwAAAADQLU1sAAAAAAC6pYkNAAAAAEC3NLEBAAAAAOiWJjYAAAAAAN3SxAYAAAAAoFua2AAAAAAAdEsTGwAAAACAbmliAwAAAADQLU1sAAAAAAC6pYkNAAAAAEC3NLEBAAAAAOiWJjYAAAAAAN2auIldVadW1fuq6q6qurOqXj2MP62qbqqqu4e/Jw3jVVWXVtX+qrq9qp4/dl97h/Xvrqq9Y+MvqKo7hm0uraqaZrLA4pJJQC/kEdATmQT0Qh4B05jmk9iHkrymtfZNSc5KclFVnZHk4iQ3t9Z2Jbl5WE6SFyfZNVz2JXlrMgqrJJckeWGSM5NcshpYwzr7xrY7d4p6gcUmk4BeyCOgJzIJ6IU8AiY2cRO7tfZAa+0jw/XPJbkryclJzkty5bDalUleOlw/L8lVbeSWJCdW1TOTvCjJTa21h1prDye5Kcm5w21Paa29v7XWklw1dl8AjyGTgF7II6AnMgnohTwCprFjM+6kqk5L8q1Jbk2ys7X2QDIKqKp6xrDayUnuG9vswDB2tPEDa4yv9fj7MnqnLTt37szKysq6Ne88IXnNcw89uryRbbbCwYMHZ/bY6ji6XmpRx7GbZSZNkkdJP5m0lp5f+55rS/quT23bwzHS5uhtn+itnqS/mnqrJ+mzpu3mGGlr9L5v9V5fosbNMg81rnKMNFvztK+sZ5HmkizWfDZ7LlM3savqq5P8pyQ/2Vr7m6OcbmitG9oE448fbO2yJJclye7du9uePXvWqTp589XX5Y13fGX6975i/W22wsrKSjZSrzq2Xy+1qOPYzDqTJsmjpJ9MWkvPr33PtSV916e2rTfrPErm+xhpXG/7RG/1JP3V1Fs9SZ81badZZ9IiHiOt6n3f6r2+RI2bZR5qTGafR8niHCNNal72lY1YpLkkizWfzZ7LNOfETlV9VUbBc3Vr7d3D8KeHr3Bk+PvgMH4gyaljm5+S5P51xk9ZYxxgTTIJ6IU8Anoik4BeyCNgUhM3sYdfeL08yV2ttV8eu+n6JKu/DLs3yXVj468cfl32rCSPDF8XuTHJOVV10nAi/nOS3Djc9rmqOmt4rFeO3RfAY8gkoBfyCOiJTAJ6IY+AaUxzOpHvSPIvk9xRVR8dxv5dkl9I8q6qujDJp5K8fLjthiQvSbI/yReS/FCStNYeqqqfT/LBYb2fa609NFz/0SRvT3JCkvcMF4C1yCSgF/II6IlMAnohj4CJTdzEbq39f1n7fENJcvYa67ckFx3hvq5IcsUa4x9K8pxJawSWh0wCeiGPgJ7IJKAX8giYxlTnxAYAAAAAgK2kiQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6Fb3TeyqOreqPlFV+6vq4lnXAyw3mQT0Qh4BPZFJQC/kESymrpvYVXVckrckeXGSM5JcUFVnzLYqYFnJJKAX8gjoiUwCeiGPYHF13cROcmaS/a21e1prX0pyTZLzZlwTsLxkEtALeQT0RCYBvZBHsKB2zLqAdZyc5L6x5QNJXjijWh7ntIt/7zHL9/7C98yoEmCbdJNJ8geWXrd5lMgkWELdZBIsk8P/H/z2c588o0q6Io9gQfXexK41xtrjVqral2TfsHiwqj6xgft+epK/evQ+3jBRfY+tY7L7eEwdM6SOx+ullnmr4+u2upAZWjeTJsyjZMpM2owMO4pe9sG19Fxb0nd9y1DbUudRMrtjpC3IpN72197qSfqrqbd6ktnXtNSZNKtjpG0y631rPb3Xl6hxU/zTN/g3Wzo/RupI9/vzMVikuSSLNZ9NzaTem9gHkpw6tnxKkvsPX6m1dlmSy47ljqvqQ6213dOVNz119FlH0k8t6ujKupk0SR4lfT+/aptcz/Wpbe4t/DHSuN5q6q2epL+aeqsn6bOmBbKUx0ireq+x9/oSNW6WeahxGyzVMdKkzKVfizSfzZ5L7+fE/mCSXVV1elUdn+T8JNfPuCZgeckkoBfyCOiJTAJ6IY9gQXX9SezW2qGqelWSG5Mcl+SK1tqdMy4LWFIyCeiFPAJ6IpOAXsgjWFxdN7GTpLV2Q5IbtuCuj/mrbFtEHY/VSx1JP7WooyNLkElrUdvkeq5PbXNuyfKot5p6qyfpr6be6kn6rGlhLFkmHa73GnuvL1HjZpmHGrfckufRRplLvxZpPps6l2rtcee3BwAAAACALvR+TmwAAAAAAJbYUjaxq+rcqvpEVe2vqotnVMOpVfW+qrqrqu6sqlfPoo6xeo6rqtuq6ndnWMOJVXVtVf3p8Lx824zq+DfDa/KxqnpnVT1xmx73iqp6sKo+Njb2tKq6qaruHv6eNKM6fnF4XW6vqt+uqhO3uo5FtF72VNUTquo3h9tvrarTtrG2dTOpqvZU1SNV9dHh8u+3sb57q+qO4XE/tMbtVVWXDs/d7VX1/G2q6xvHno+PVtXfVNVPHrbOtj5v02RJVe0d1rm7qvZuU20bypf19gGOXU+Z1GsG9ZY9PWROjxkjW+ZfT3k0RY3/tqo+PuxzN1fV1/VW49h6L6uqVlW7t7O+4bHXrbGq/tfhubyzqv5jbzVW1f8w/D/rtuH1fsk21/e4zDvs9m39f9Oy2Oh/X71bb/+ZJxs5fpwXVfXEqvpAVf3JMJefnXVN06qt6DO21pbqktGJ/f88ybOSHJ/kT5KcMYM6npnk+cP1r0nyZ7OoY6yef5vkPyb53RnWcGWSfzVcPz7JiTOo4eQkn0xywrD8riQ/uE2P/Z1Jnp/kY2Nj/2eSi4frFyd5w4zqOCfJjuH6G7ajjkW7bCR7kvxYkl8frp+f5De3sb51MynJnlllRJJ7kzz9KLe/JMl7klSSs5LcOqPX+C+TfN0sn7dJsyTJ05LcM/w9abh+0jbUtqF8WW8fcDnm16KrTOo1g3rOnlllTo8ZI1vm+9JbHk1R4z9N8qTh+o/2WOOw3tck+aMktyTZ3VuNSXYluW01H5I8o8MaL0vyo8P1M5Lcu801Pi7zDrt95sfFi3bZ6H9f83BZb/+Zp0s666tNOZdK8tXD9a9KcmuSs2Zd15Rz2vQ+4zJ+EvvMJPtba/e01r6U5Jok5213Ea21B1prHxmufy7JXRk1ULddVZ2S5HuSvG0Wjz/U8JSMwvTyJGmtfam19tkZlbMjyQlVtSPJk5Lcvx0P2lr7oyQPHTZ8XkbN/Qx/XzqLOlprf9BaOzQs3pLklK2uYwFtJHvGX+9rk5xdVbUdxfWUSRM6L8lVbeSWJCdW1TO3uYazk/x5a+2/bfPjPsYUWfKiJDe11h5qrT2c5KYk5251bfJlZrrKpDnOoFlmz0wyp8eMkS1zr6s8mrTG1tr7WmtfGBZnsc9t9N+5P5/RG09f3M7iBhup8V8necuQE2mtPdhhjS3JU4brT802/Xvx0QdfO4fH9XBcvGi66CNthg3sP3Njjo8fH2f47/XgsPhVw2Vuf8Rwq/qMy9jEPjnJfWPLBzLjnXz4Oty3ZvROyyz8SpKfSvIPM3r8ZPSO5meS/MbwdYO3VdWTt7uI1tpfJPmlJJ9K8kCSR1prf7DddYzZ2Vp7YKjtgSTPmGEtq344o3f2OTYbyZ5H1xn+4f1Ikq/dlurGrJNJ3zZ8xek9VfXsbSyrJfmDqvpwVe1b4/Yesv38JO88wm2zet5WbSRLengOj5Yv6+0DHJtuM6mzDOo5e3rKnN4zRrb0rds8WuvxB+vtvxdm+4+X162xqr41yamttVmdQnIjz+M3JPmGqvrjqrqlqjb1DfUN2EiN/0eSf1FVB5LckOTHt6e0DevhmG7ReE4710FfbWrD6Tc+muTBjN74n9u5ZIv6jMvYxF7rHfuZvbtRVV+d5D8l+cnW2t/M4PG/N8mDrbUPb/djH2ZHRl9peWtr7VuTfD6jr6Nuqxqdw/G8JKcn+cdJnlxV/2K76+hVVf1MkkNJrp51LXNoI9kz83xaJ5M+ktHX1p+X5M1JfmcbS/uO1trzk7w4yUVV9Z2H3T7T566qjk/yfUl+a42bZ/m8HYtZP4fr5ct6+wDHpstM6jCDusyeOc2cWT1XsqV/XebRpI8//Nthd5Jf3NKK1njoNcYerbGq/rskb0rymm2r6PE28jzuyOiUInuSXJDkbbW9v8ezkRovSPL21topGZ264x3D89uLWf/3sog8px2bdV9ts7TWvtxa+5aMvslzZlU9Z9Y1TWIr+4w9Be12OZDk1LHlU7LNX/9ZVVVfldF/aFe31t49ixqSfEeS76uqezP6Ssx3VdX/M4M6DiQ5MPZO07UZNbW323cn+WRr7TOttb9P8u4k3z6DOlZ9evWrX8Pf7f463aNq9ANM35vkFa01/8M+dhvJnkfXGU5n89Rs41e91suk1trfrH7FqbV2Q5Kvqqqnb0dtrbX7h78PJvntjL7SN27W2f7iJB9prX368Btm+byN2UiWzOw53Ei+bGAf4Nh0l0k9ZlDH2dNb5nSZMbJlbnSXR2vY0P5bVd+d5GeSfF9r7e+2qbZV69X4NUmek2Rl+LffWUmur+39cceNvtbXtdb+vrX2ySSfyKipvV02UuOFGf12Ulpr70/yxCTbfWx3NLM+Ll5EntNOddJX21RtdGrdlWzyqR230Zb1GZexif3BJLuq6vThUyznJ7l+u4sYzuF2eZK7Wmu/vN2Pv6q19trW2imttdMyei7e21rb9k8et9b+Msl9VfWNw9DZST6+3XVkdBqRs6rqScNrdHZG51WaleuT7B2u701y3SyKGL7G99MZHZB/Yb31WdNGsmf89X5ZRv89bssbBhvJpKr671fPP1lVZ2b0/5C/3obanlxVX7N6PaMf6zr817SvT/LKGjkro1MBPbDVtY25IEf4Wv+snrfDbCRLbkxyTlWdNHwr5ZxhbEttJF82uA9wbLrKpB4zqPPs6S1zussY2TJXusqjSWscTtXxf2e0z83igydHrbG19khr7emttdOGf/vdMtT6oV5qHPxORj+SmeENuG/I6Idge6rxUxn9OzFV9U0ZNbE/s401rmfWx8WLqIs+Eo/VS19tM1TVP1r91klVnZDRByz/dLZVTWZL+4ytg1+s3O5LRl/5+bOMfl32Z2ZUw/+Y0ddPbk/y0eHykhk/L3uyhb9mv4HH/5YkHxqek9/JJv1i/QR1/GxGYfGxJO9I8oRtetx3ZnQe7r/P6J3eCzM619/NSe4e/j5tRnXsz+gcYKv76q/Paj+Z58ta2ZPk5zL6B0QyOgD+reH5/kCSZ21jbWtmUpIfSfIjwzqvSnJnRr/GfUuSb9+m2p41POafDI+/+tyN11ZJ3jI8t3ck2b2Nz92TMmoQPXVsbGbP27FkSUZfd37b2LY/POx/+5P80DbVtma+ZHRKpxuOtg+4TP16dJNJPWZQr9kz68zpMWNky/xfesqjKWr8wySfHtvnru+txsPWXdmOzJrgeawkv5zRB5ruSHJ+hzWekeSPh+z4aJJztrm+tTKvi+PiRb6stV/M42Wt/WfWNU0xl+76alPM5ZuT3DbM5WNJ/v2sa9qkee3JJvYZa7hTAAAAAADozjKeTgQAAAAAgDmhiQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxObLlXV06vqj6vqr6vqs1X1/qr6jlnXBSy3qtpbVa2q/tWsawGW15BDn6+qg8PlbbOuCVhOVXVcVb2uqu6vqs9V1W1VdeKs6wKWT1X9T2PHRquXVlU/MOva2Bw7Zl0AHMHBJD+c5O4kLcl5Sf5zVT2jtXZoppUBS6mqTkry2iR3zroWgCTPa63tn3URwNL72STfnuTbknwqybOTfHGmFQFLqbX2X5J89epyVe1J8p+T/P6samJz+SQ226qq7q2q/62qbh8+QXR5Ve2sqvcM79z/YVWd1Fr7YmvtE621f0hSSb6c5KQkT5vtDIBFsdE8GtvkPyS5NMlfzahkYIFNkEkAW2KjeTRk0k8m+dettf/WRj7WWtPEBjbNFMdIe5Nc21r7/HbXzNbQxGYWfiDJP0vyDUn+lyTvSfLvkjw9o33yJ1ZXrKrbM3on//okb2utPbjt1QKLbEN5VFVnJtmd5NdnUyawJDZ8jJTkj6rqL6vq3VV12jbXCSy+jeTRc5McSvKyIY/+rKoumlG9wGI7lmOkVNWTkrwsyZXbWyZbyelEmIU3t9Y+nSRV9V+SPNhau21Y/u0kZ6+u2Fr75qp6YpLvT3L8LIoFFtq6eVRVxyX5tSQ/3lr7h6qaXbXAotvoMdL/nOSWJE9K8rokv1tV3+KUa8Am2kge3Z3kqRk1lU5PsivJzVX1Z621m2ZTNrCgNtxHGvxARt+g/X+3tUq2lE9iMwufHrv+t2ssf/X4ysOpRd6Z5OKqet421Acsj43k0Y8lub219v7tLAxYShs6Rmqt/VFr7Uuttc8meXVGzaNv2rYqgWWwkTz622H551prf9tauz3JNUlesj0lAkvkmPpIGZ1K5KrWWtvqwtg+mtjMk69K8qxZFwEsnbOTfP/wNdm/zOjHi95YVf/XjOsCWNUy+g0RgO10+/BXkwjoRlWdmmRPkqtmXAqbzOlE6FJVnZXR/vmBJMdldH6jnUlunWVdwFL6wSRPHFt+d5Jrk1w+k2qApVZVz87ojf07kpyQ0elE/iLJXbOsC1g+rbU/H77W/zNV9RMZfeDonye5YLaVAUvuXyb5r621P591IWwuTWx69YQkl2Z0IPT3Gf1D7Xtaa/fPtCpg6Qxf139UVX0pyd+01h6ZUUnActuZ5K1JTkny+ST/Ncn3ttb+fqZVAcvqgoze2P/rJA8m+d9bazfPtiRgyb0yyS/Ougg2Xzk9DAAAAAAAvXJObAAAAAAAuqWJDQAAAABAtzSxAQAAAADo1rpN7Kq6oqoerKqPjY39YlX9aVXdXlW/XVUnjt322qraX1WfqKoXjY2fO4ztr6qLx8ZPr6pbq+ruqvrNqjp+GH/CsLx/uP20zZo0ML9kEtALeQT0RCYBvZBHwFZY94cdq+o7kxxMclVr7TnD2DlJ3ttaO1RVb0iS1tpPV9UZSd6Z5Mwk/zjJHyb5huGu/izJP0tyIMkHk1zQWvt4Vb0rybtba9dU1a8n+ZPW2lur6seSfHNr7Ueq6vwk399a++frTejpT396O+2009ad+Oc///k8+clPXne9WVLj5lDj5thojR/+8If/qrX2j7aqjnnKpI3mUdLPPqAOdSxSHfLosRbpGGmzLMtcl2WeSd9zlUlfMY/HSJvBXPq0jHORR4+16MdI81j3PNaczGfdPdS84Uxqra17SXJako8d4bbvT3L1cP21SV47dtuNSb5tuNw4Nv7a4VJJ/irJjmH80fVWtx2u7xjWq/VqfcELXtA24n3ve9+G1pslNW4ONW6OjdaY5ENtA7kyzWVeMmmjeXQsz+9WU8djqeOx5q0OebS4x0ibZVnmuizzbK3vucqk+T5G2gzm0qdlnIs8Wq5jpHmsex5rbm0+6+6h5o1m0macE/uHk7xnuH5ykvvGbjswjB1p/GuTfLa1duiw8cfc13D7I8P6AEcjk4BeyCOgJzIJ6IU8Ao7Zjmk2rqqfSXIoydWrQ2us1rL2ubfbUdY/2n2tVce+JPuSZOfOnVlZWTly0YODBw9uaL1ZUuPmUOPmmIcae8ikSfIo6ef5VYc61LE5esijoY6FPEbaLMsy12WZZ7Jccz0WPWTSvB8jbQZz6ZO5bK8e8mioY2mOkeax7nmsOZnPuuep5omb2FW1N8n3Jjl7+Oh3MnoH7NSx1U5Jcv9wfa3xv0pyYlXtGN4lG19/9b4OVNWOJE9N8tBatbTWLktyWZLs3r277dmzZ936V1ZWspH1ZkmNm0ONm6P3GnvJpEnyKOnn+VWHOtQxvV7yKFncY6TNsixzXZZ5Jss1143qJZPm/RhpM5hLn8xl+/SSR8lyHSPNY93zWHMyn3XPU80TnU6kqs5N8tNJvq+19oWxm65Pcv7wi7CnJ9mV5AMZnYB/1/ALsscnOT/J9UNovS/Jy4bt9ya5buy+9g7XX5bRDwAc/VcogaUkk4BeyCOgJzIJ6IU8Aqa17iexq+qdSfYkeXpVHUhySUYn039CkpuqKkluaa39SGvtzuFXYj+e0ddDLmqtfXm4n1dldJL945Jc0Vq7c3iIn05yTVW9LsltSS4fxi9P8o6q2p/RO2fnb8J8gTknk4BeyCOgJzIJ6IU8ArbCuk3s1toFawxfvsbY6vqvT/L6NcZvSHLDGuP3JDlzjfEvJnn5evUBy0UmAb2QR0BPZBLQC3kEbIWJTicCAAAAAADbYeIfdpx3d/zFI/nBi3/v0eV7f+F7ZlgNsOxkEtALeQT0RCYBvZBHMFs+iQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6Na6TeyquqKqHqyqj42NPa2qbqqqu4e/Jw3jVVWXVtX+qrq9qp4/ts3eYf27q2rv2PgLquqOYZtLq6qO9hjAcpNJQC/kEdATmQT0Qh4BW2Ejn8R+e5JzDxu7OMnNrbVdSW4elpPkxUl2DZd9Sd6ajIIkySVJXpjkzCSXjIXJW4d1V7c7d53HAJbb2yOTgD68PfII6MfbI5OAPrw98gjYZOs2sVtrf5TkocOGz0ty5XD9yiQvHRu/qo3ckuTEqnpmkhcluam19lBr7eEkNyU5d7jtKa2197fWWpKrDruvtR4DWGIyCeiFPAJ6IpOAXsgjYCtMek7sna21B5Jk+PuMYfzkJPeNrXdgGDva+IE1xo/2GACHk0lAL+QR0BOZBPRCHgFT2bHJ91drjLUJxo/tQav2ZfRVkuzcuTMrKyvrbrPzhOQ1zz306PJGttluBw8e7LKucWrcHGrcMtueSZPkUdJPJvXyOqtDHfNQxzFyjNShOd2XjtmyzDNZrrlOyTHSDCzS/mkufZrTuThG2mLzuF/MY83JfNY9TzVP2sT+dFU9s7X2wPBVjgeH8QNJTh1b75Qk9w/jew4bXxnGT1lj/aM9xuO01i5LclmS7N69u+3Zs+dIqz7qzVdflzfe8ZXp3/uK9bfZbisrK9nIXGZJjZtDjVPrJpMmyaOkn0zq5XVWhzrmoY4j6CaPksU9Rtosne9Lm2ZZ5pks11w3qJtMmvdjpM2wSPunufSp87l0k0fJch0jdb5frGkea07ms+55qnnS04lcn2T1l2H3JrlubPyVw6/LnpXkkeErHDcmOaeqThpOxH9OkhuH2z5XVWcNvyb7ysPua63HADicTAJ6IY+AnsgkoBfyCJjKup/Erqp3ZvTu19Or6kBGvw77C0neVVUXJvlUkpcPq9+Q5CVJ9if5QpIfSpLW2kNV9fNJPjis93OttdWT/P9oRr9ce0KS9wyXHOUxgCUmk4BeyCOgJzIJ6IU8ArbCuk3s1toFR7jp7DXWbUkuOsL9XJHkijXGP5TkOWuM//VajwEsN5kE9EIeAT2RSUAv5BGwFSY9nQgAAAAAAGw5TWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6NZUTeyq+jdVdWdVfayq3llVT6yq06vq1qq6u6p+s6qOH9Z9wrC8f7j9tLH7ee0w/omqetHY+LnD2P6quniaWoHFJ5OAXsgjoCcyCeiFPAImNXETu6pOTvITSXa31p6T5Lgk5yd5Q5I3tdZ2JXk4yYXDJhcmebi19vVJ3jSsl6o6Y9ju2UnOTfJrVXVcVR2X5C1JXpzkjCQXDOsCPI5MAnohj4CeyCSgF/IImMa0pxPZkeSEqtqR5ElJHkjyXUmuHW6/MslLh+vnDcsZbj+7qmoYv6a19nettU8m2Z/kzOGyv7V2T2vtS0muGdYFOBKZBPRCHgE9kUlAL+QRMJEdk27YWvsnNigAAAAgAElEQVSLqvqlJJ9K8rdJ/iDJh5N8trV2aFjtQJKTh+snJ7lv2PZQVT2S5GuH8VvG7np8m/sOG3/hpPUCi00mAb3oKY+qal+SfUmyc+fOrKysrFv/zhOS1zz30KPLG9lmXh08eHCh57dqWeaZLNdcN6qXTJokj5LFyqRF2j/NpU+9z6WXPEqW6xip9/1iLfNYczKfdc9TzRM3savqpIze0To9yWeT/FZGX9k4XFvd5Ai3HWl8rU+JtzXGFjZ85mFHUuPmUOP0esmkef8HWi+vszrUMQ91HEkveZQkrbXLklyWJLt372579uw5WulJkjdffV3eeMdXDhHvfcX628yrlZWVbOQ5mXfLMs9kuea6Ub1k0iR5lCxWJi3S/mkufep9Lr3kUbJcx0i97xdrmceak/mse55qnriJneS7k3yytfaZJKmqdyf59iQnVtWO4V20U5LcP6x/IMmpSQ4MXxt5apKHxsZXjW9zpPHHWNTwmYcdSY2bQ42bootMmvd/oPXyOqtDHfNQx1F0kUcAA5kE9EIeAROb5pzYn0pyVlU9aTgn0dlJPp7kfUleNqyzN8l1w/Xrh+UMt7+3tdaG8fOHX509PcmuJB9I8sEku4ZfqT0+o5P2Xz9FvcBik0lAL+QR0BOZBPRCHgETm+ac2LdW1bVJPpLkUJLbMvr04e8luaaqXjeMXT5scnmSd1TV/ozeOTt/uJ87q+pdGQXXoSQXtda+nCRV9aokN2b0i7VXtNbunLReYLHJJKAX8gjoiUwCeiGPgGlMczqRtNYuSXLJYcP3ZPSLsIev+8UkLz/C/bw+yevXGL8hyQ3T1AgsD5kE9EIeAT2RSUAv5BEwqWlOJwIAAAAAAFtKExsAAAAAgG5pYgMAAAAA0C1NbAAAAAAAuqWJDQAAAABAtzSxAQAAAADoliY2AAAAAADd0sQGAAAAAKBbmtgAAAAAAHRLExsAAAAAgG5pYgMAAAAA0C1NbAAAAAAAuqWJDQAAAABAtzSxAQAAAADoliY2AAAAAADd0sQGAAAAAKBbmtgAAAAAAHRLExsAAAAAgG5pYgMAAAAA0C1NbAAAAAAAujVVE7uqTqyqa6vqT6vqrqr6tqp6WlXdVFV3D39PGtatqrq0qvZX1e1V9fyx+9k7rH93Ve0dG39BVd0xbHNpVdU09QKLTSYBvZBHQE9kEtALeQRMatpPYv9qkt9vrf2TJM9LcleSi5Pc3FrbleTmYTlJXpxk13DZl+StSVJVT0tySZIXJjkzySWrgTWss29su3OnrBdYbDIJ6IU8Anoik4BeyCNgIhM3savqKUm+M8nlSdJa+1Jr7bNJzkty5bDalUleOlw/L8lVbeSWJCdW1TOTvCjJTa21h1prDye5Kcm5w21Paa29v7XWklw1dl8AjyGTgF7II6AnMgnohTwCprFjim2fleQzSX6jqp6X5MNJXp1kZ2vtgSRprT1QVc8Y1j85yX1j2x8Yxo42fmCN8cepqn0ZvdOWnTt3ZmVlZd3id56QvOa5hx5d3sg22+3gwYNd1jVOjZtDjZuii0yaJI+SfjKpl9dZHeqYhzqOoos8Shb3GGmzzMG+tCmWZZ7Jcs31GHSRSfN+jLQZFmn/NJc+zcFcusijZLmOkeZgv3iceaw5mc+656nmaZrYO5I8P8mPt9ZurapfzVe+8rGWtc5D1CYYf/xga5cluSxJdu/e3fbs2XOUMkbefPV1eeMdX5n+va9Yf5vttrKyko3MZZbUuDnUuCm6yKRJ8ijpJ5N6eZ3VoY55qOMousijZHGPkTbLHOxLm2JZ5pks11yPQReZNO/HSJthkfZPc+nTHMylizxKlusYaQ72i8eZx5qT+ax7nmqe5pzYB5IcaK3dOixfm1EYfXr4CkeGvw+OrX/q2PanJLl/nfFT1hgHWItMAnohj4CeyCSgF/IImNjETezW2l8mua+qvnEYOjvJx5Ncn2T1l2H3JrluuH59klcOvy57VpJHhq+L3JjknKo6aTgR/zlJbhxu+1xVnTX8muwrx+4L4DFkEtALeQT0RCYBvZBHwDSmOZ1Ikvx4kqur6vgk9yT5oYwa4++qqguTfCrJy4d1b0jykiT7k3xhWDettYeq6ueTfHBY7+daaw8N1380yduTnJDkPcMF4EhkEtALeQT0RCYBvZBHwESmamK31j6aZPcaN529xrotyUVHuJ8rklyxxviHkjxnmhqB5SGTgF7II6AnMgnohTwCJjXNObEBAAAAAGBLaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLembmJX1XFVdVtV/e6wfHpV3VpVd1fVb1bV8cP4E4bl/cPtp43dx2uH8U9U1YvGxs8dxvZX1cXT1gosPpkE9EIeAT2RSUAv5BEwic34JPark9w1tvyGJG9qre1K8nCSC4fxC5M83Fr7+iRvGtZLVZ2R5Pwkz05ybpJfGwLtuCRvSfLiJGckuWBYF+BoZBLQC3kE9EQmAb2QR8Axm6qJXVWnJPmeJG8blivJdyW5dljlyiQvHa6fNyxnuP3sYf3zklzTWvu71tonk+xPcuZw2d9au6e19qUk1wzrAqxJJgG9kEdAT2QS0At5BExqx5Tb/0qSn0ryNcPy1yb5bGvt0LB8IMnJw/WTk9yXJK21Q1X1yLD+yUluGbvP8W3uO2z8hWsVUVX7kuxLkp07d2ZlZWXdwneekLzmuYceXd7INtvt4MGDXdY1To2bQ42bZuaZNEkeJf1kUi+vszrUMQ91rGPmeZQs7jHSZpmTfWlqyzLPZLnmeoxmnknzfoy0GRZp/zSXPs3JXGaeR8lyHSPNyX7xGPNYczKfdc9TzRM3savqe5M82Fr7cFXtWR1eY9W2zm1HGl/rU+JtjbG01i5LclmS7N69u+3Zs2et1R7jzVdflzfe8ZXp3/uK9bfZbisrK9nIXGZJjZtDjdPrJZMmyaOkn0zq5XVWhzrmoY4j6SWPksU9Rtosve9Lm2VZ5pks11w3qpdMmvdjpM2wSPunufSp97n0kkfJch0j9b5frGUea07ms+55qnmaT2J/R5Lvq6qXJHlikqdk9I7aiVW1Y3gX7ZQk9w/rH0hyapIDVbUjyVOTPDQ2vmp8myONAxxOJgG9kEdAT2QS0At5BExs4nNit9Ze21o7pbV2WkYn1H9va+0VSd6X5GXDanuTXDdcv35YznD7e1trbRg/f/jV2dOT7ErygSQfTLJr+JXa44fHuH7SeoHFJpOAXsgjoCcyCeiFPAKmMe05sdfy00muqarXJbktyeXD+OVJ3lFV+zN65+z8JGmt3VlV70ry8SSHklzUWvtyklTVq5LcmOS4JFe01u7cgnqBxSaTgF7II6AnMgnohTwC1rUpTezW2kqSleH6PRn9Iuzh63wxycuPsP3rk7x+jfEbktywGTUCy0MmAb2QR0BPZBLQC3kEHKuJTycCAAAAAABbTRMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALqliQ0AAAAAQLc0sQEAAAAA6JYmNgAAAAAA3dLEBgAAAACgW5rYAAAAAAB0SxMbAAAAAIBuaWIDAAAAANAtTWwAAAAAALo1cRO7qk6tqvdV1V1VdWdVvXoYf1pV3VRVdw9/TxrGq6ourar9VXV7VT1/7L72DuvfXVV7x8ZfUFV3DNtcWlU1zWSBxSWTgF7II6AnMgnohTwCpjHNJ7EPJXlNa+2bkpyV5KKqOiPJxUlubq3tSnLzsJwkL06ya7jsS/LWZBRWSS5J8sIkZya5ZDWwhnX2jW137hT1AotNJgG9kEdAT2QS0At5BExs4iZ2a+2B1tpHhuufS3JXkpOTnJfkymG1K5O8dLh+XpKr2sgtSU6sqmcmeVGSm1prD7XWHk5yU5Jzh9ue0lp7f2utJblq7L4AHkMmAb2QR0BPZBLQC3kETGPHZtxJVZ2W5FuT3JpkZ2vtgeT/Z+/uwyyryjvvf3+hBRFfQI092E3SJHZMEKORHiHjTNIjCbbiiJnRCYYoGDJMHExMhifamDzB+BZ8Msb3mDBCACUiIToQQbGjlhkngooaEdChRSINBNQGYmvUtLmfP/YqOF1d1VVdfarOPqe+n+s6V5+99tr73LuquDnnPmut3SWoJI9q3dYAtw4ctq217al92yzts73+aXTftLF69WqmpqbmjXn1gXDG43fet72QY5bbjh07ehnXIGMcDmMcrlHmpMXkI+hPTurL79k4jGMc4lgI3yP12zj9Le2LlXKdsLKudTF8jzRak/T36bX00zhdi++Rls84/V1MG8eYYTzjHqeY97mIneTBwF8Cv1lV/7iH5YZm21GLaN+9seoc4ByADRs21MaNG+eJGt5y0WW8/rr7L/+Wk+Y/ZrlNTU2xkGsZJWMcDmMcnlHnpMXkI+hPTurL79k4jGMc4pjPqPMRTO57pGEZl7+lfbVSrhNW1rXurVHnpHF/jzQMk/T36bX007hcy6jzEays90jj8ncxaBxjhvGMe5xi3pc1sUnyALrEc1FVvbc139mmcND+vau1bwMOGzh8LXD7PO1rZ2mXpFmZkyT1hflIUp+YkyT1hflI0mItuojd7vB6LnBjVf3RwK7Lgek7w54MXDbQ/oJ2d9ljgHvbdJGrgOOSHNIW4j8OuKrt+2aSY9prvWDgXJK0C3OSpL4wH0nqE3OSpL4wH0naF/uynMhTgOcD1yX5XGt7OXA2cEmSU4GvAs9t+64EngFsBb4NvBCgqrYneRXwqdbvlVW1vT1/EXA+cCDwgfaQpNmYkyT1hflIUp+YkyT1hflI0qItuohdVR9n9vWGAI6dpX8Bp89xrvOA82Zp/zRw5GJjlLRymJMk9YX5SFKfmJMk9YX5SNK+2Kc1sSVJkiRJkiRJWkoWsSVJkiRJkiRJvWURW5IkSZIkSZLUWxaxJUmSJEmSJEm9ZRFbkiRJkiRJktRbFrElSZIkSZIkSb1lEVuSJEmSJEmS1FsWsSVJkiRJkiRJvWURW5IkSZIkSZLUWxaxJUmSJEmSJEm9ZRFbkiRJkiRJktRbFrElSZIkSZIkSb1lEVuSJEmSJEmS1FsWsSVJkiRJkiRJvWURW5IkSZIkSZLUWxaxJUmSJEmSJEm9ZRFbkiRJkiRJktRbFrElSZIkSZIkSb1lEVuSJEmSJEmS1Fu9L2In2ZTkS0m2Jtk86ngkrWzmJEl9YT6S1CfmJEl9YT6SJlOvi9hJ9gPeBjwdOAJ4XpIjRhuVpJXKnCSpL8xHkvrEnCSpL8xH0uTqdREbeDKwtapurqrvARcDJ4w4JkkrlzlJUl+YjyT1iTlJUl+Yj6QJtWrUAcxjDXDrwPY24OgRxSJJ5iRJfTFR+Wjd5it22b7l7ONHFImkRZqonCRprI1VPvI9kLRwfS9iZ5a22q1TchpwWtvckeRLCzj3I4Gv33eO1y0qvqW2S4w9ZYzDMUkx/vBSBzJC8+akReYj6E9O6svfonHsyjh2ZT6a8PdIy/iaffmbXmor5Tqh39e6onPSBLxHGoY+/33uLa+ln3yPNObvkZYox43j3/g4xgzjGXcfYl5QTup7EXsbcNjA9lrg9pmdquoc4Jy9OXGST1fVhn0Lb2kZ43AY43CMQ4zLYN6ctJh8BP35+RqHcRjH2FjR75GGZaVc60q5TlhZ19ozE/8eaRi8ln7yWiaO75FmGMe4xzFmGM+4xynmvq+J/SlgfZLDk+wPnAhcPuKYJK1c5iRJfWE+ktQn5iRJfWE+kiZUr0diV9XOJC8GrgL2A86rqutHHJakFcqcJKkvzEeS+sScJKkvzEfS5Op1ERugqq4ErlyCU+/1VLYRMMbhMMbhGIcYl9wKyEnGsSvj2JVx9MgKyEfLYaVc60q5TlhZ19or5qQF8Vr6yWuZMOaj3Yxj3OMYM4xn3GMTc6p2W99ekiRJkiRJkqRe6Pua2JIkSZIkSZKkFWxFFrGTbErypSRbk2wedTwzJTkvyV1JvjDqWOaS5LAkH01yY5Lrk7xk1DHNlOSBST6Z5O9ajL8/6phmk2S/JJ9N8v5RxzKbJLckuS7J55J8etTxTJq+5KM+5J0+5ZU+5Y++5Ii+5IIkBye5NMkX29/KT48qlknUl5y0lPqUa5ZDX3LIUjM3TJ5JyUeTmHMmKa9MUu5I8lvtb+wLSd6d5IGjjmmSjFtOGufcM445ZhxzyTjmjBVXxE6yH/A24OnAEcDzkhwx2qh2cz6wadRBzGMncEZV/QRwDHB6D3+O3wWeWlVPAJ4IbEpyzIhjms1LgBtHHcQ8/n1VPbGqNow6kEnSs3x0PqPPO33KK33KH33KEX3IBW8CPlhVPw48gf78bMZez3LSUupTrlkOfcohS8ncMEEmLB9NYs6ZpLwyEbkjyRrgN4ANVXUk3Q0NTxxtVJNjTHPSOOeeccwxY5VLxjVnrLgiNvBkYGtV3VxV3wMuBk4YcUy7qKq/AbaPOo49qao7quoz7fk36f4DXTPaqHZVnR1t8wHt0atF4JOsBY4H3jHqWDQSvclHfcg7fcorfckf5ohdJXko8DPAuQBV9b2qume0UU2U3uSkpdSnXLPUVkoOMTdMpInJR5OWcyYpr0xg7lgFHJhkFfAg4PYRxzNJxi4njWvuGcccM8a5ZOxyxkosYq8Bbh3Y3sYY/IfcZ0nWAT8FXDPaSHbXpqF8DrgL2FJVfYvxjcBLgX8ZdSB7UMCHklyb5LRRBzNhzEdz6ENe6Un+6FOO6EMu+BHga8CftSmG70hy0IhimUQrLif1IdcssT7lkKVkbpg8E5mPJiTnTFJemZjcUVW3Af8D+CpwB3BvVX1otFFNlLHOSWOWe8Yxx4xdLhnXnLESi9iZpa1Xo3PHSZIHA38J/GZV/eOo45mpqr5fVU8E1gJPTnLkqGOaluSZwF1Vde2oY5nHU6rqSXRTp05P8jOjDmiCmI9m0Ze8Mur80cMc0YdcsAp4EvD2qvop4FtA79ckHCMrKif1JdcslR7mkKVkbpg8E5ePJiHnTGBemZjckeQQupHBhwOPBg5K8sujjWqijG1OGqfcM8Y5ZuxyybjmjJVYxN4GHDawvZYxGDLfR0keQJcML6qq9446nj1pUzmmGP2av4OeAjwryS1005GemuRdow1pd1V1e/v3LuB9dFOpNBzmoxn6mFdGmD96lSN6kgu2AdsGRsVfSveGUcOxYnJSH3PNEuhVDlli5obJM1H5aIJyzqTllUnKHT8HfKWqvlZV/wy8F/g3I45pkoxlThrD3DOuOWYcc8lY5oyVWMT+FLA+yeFJ9qdbuPzyEcc0dpKEbr2fG6vqj0Ydz2yS/GCSg9vzA+n+I/3iaKO6X1WdWVVrq2od3d/hR6qqV998JTkoyUOmnwPHAV8YbVQTxXw0oE95pQ/5o085oi+5oKr+Abg1yWNb07HADcsdxwRbETmpT7lmKfUphyw1c8NEmph8NEk5Z9LyyoTljq8CxyR5UPubO5ae31huzIxdThrH3DOuOWZMc8lY5oxVow5guVXVziQvBq6iu/vmeVV1/YjD2kWSdwMbgUcm2QacVVXnjjaq3TwFeD5wXVszFuDlVXXlCGOa6VDggnYn4R8ALqmq9484pnGzGnhfl9NYBfx5VX1wtCFNjj7lo57knT7lFfPHrvqUC34duKh9gLgZeOGI4pg4fcpJS6xPuUbDY26YIBOWj8w5/TYRuaOqrklyKfAZYCfwWeCc0UY1OcY0J5l7ltdY5ZJxzRmpGotlfCRJkiRJkiRJK9BKXE5EkiRJkiRJkjQmLGJLkiRJkiRJknrLIrYkSZIkSZIkqbcsYkuSJEmSJEmSessitiRJkiRJkiSptyxiS5IkSZIkSZJ6yyK2JEmSJEmSJKm3LGJLkiRJkiRJknrLIrYkSZIkSZIkqbcsYkuSJEmSJEmSessitiRJkiRJkiSptyxiS5IkSZIkSZJ6yyK2JEmSJEmSJKm3LGJLkiRJkiRJknrLIrYkSZIkSZIkqbcsYkuSJEmSJEmSessitiRJkiRJkiSptyxiS5IkSZIkSZJ6yyK2JEmSJEmSJKm3LGJLkiRJkiRJknrLIrYkSZIkSZIkqbcsYkuSJEmSNGaSvDzJO/aw/5YkP7cEr3t9ko37eI6TknxoSPHMe51JXpHkXYs8/ylJPr646CRJw2IRW72V5KlJPpPkH5PcnOS0UcckaWVK8h+SfCHJjiR/m+SIUcckaWksVdFnL2M4MMlfJbk3yV/soZ+FFWkFq6rXVtWvjuB1H1dVU/t4jouq6rghhSRJWgEsYquXkjwAeB/wp8DDgF8E/ijJE0YamKQVJ8l64CLg14CDgb8CLk+yaqSBSZpkzwFWA4+oqueOOhhJAvC9jyRplCxia1m10U2/neTzSb6V5Nwkq5N8IMk3k/x1kkOAhwMPBd5ZnU8BNwKOfpQ0FHuRj54G/O+q+nhV7QReB6wBfnakFyCp1/ax2PPDwP9tOWdZzBbvKApWFsk0jpJUkscMbJ+f5NXt+cYk25KckeSuJHckeeGMvn+SZEt7//GxJD88sP9NSW5ts1OvTfLvBvbtskRGkucn+fsk30jyOwuM/RVJLk3ynvb6nxkcONTeL70syeeBbyVZNThjpR1/SZIL2/HXJ9kwcPxhSd6b5Gstrre29l1mkrSf4W+0GbhfT/KHSX6g7fvRJB9px389yUVJDl7gr2fQA/dwnZuTfLntuyHJLyzi/JKkJWQRW6Pwn4CfB34M+A/AB4CXA4+k+5v8jaq6E3g38MIk+yX5aboPdE6ZlTRM8+YjIO0xbXr7yGWNVFrhWtHk/2lfPN3bChEPnG1JjcGCUisQ/XH7gmpHkv+T5F8leWOSu5N8MclPzXi5f92KGHcn+bMkDxw49zOTfC7JPemWF/rJGTHuUuzZw/X8RJKpdp7rkzyrtf8+8HvAL7Z4T53/R5O3tJ/JF5McO7Dj0UkuT7I9ydYk/2Vg33Th6l1J/hE4ZY62A9rP6vb2eGOSA9o5PpbkP7Xn/7b93J/Rtn8uyecGXu9XktzYfqZXZdciXSU5PclNwE3pvCFdwe/e9js352qc/Su62aVrgFOBt6X7onzaScCr6N5/fI5uBti0TwFPpBvk8+fAXwzmpGnpljp7O/B84NHAI4C1C4zvBOAvBl7jf6WbGTvtecDxwMFzfLn2LOBiuhlrlwPTher9gPcDfw+sa9d/8R7i+AVgA/CkFtOvTF8e8Aftun4COAx4xQKvbdCervPLwL+j+z39PvCuJIcu4jUkSUvEIrZG4S1VdWdV3Qb8b+CaqvpsVX2XbgmR6Q+S76b7EPfd1u93qurWkUQsaVItJB9tAX423Uiq/emK3PsDDxpZ1NLK9Z+BTcDhwE8Cp+zFcb9LVyD6LvAJ4DNt+1Lgj2b0P4luFsaP0n3J9bsASZ4EnAf8V7oC0Z/SLS90wMCx8xV7ppdN+yvgQ8CjgF8HLkry2Ko6C3gt8J6qenBVnTvPtR0N3Nyu5SzgvUke3va9G9hGV/h5DvDawSI3XUHnUrrC00VztP0OcAxdEe0JwJOnfx7Ax4CN7fnPtDh+dmD7Y+16n02XO/8j8IN0+fbdM67j2e1ajgCOa8f/WIvjF4FvzPNzkPrsn4FXVtU/V9WVwA7gsQP7r6iqv2nvP34H+OkkhwFU1buq6htVtbOqXg8cMOPYac8B3j9wnv8X+JcFxndtVV1aVf9Mlw8fSPff/bQ3V9WtVfVPcxz/8aq6sqq+D7yTLldAly8eDfx2VX2rqr5TVXsalPS6qtpeVV8F3kiXT6mqrVW1paq+W1VfazEuZkbcnNdZVX9RVbdX1b9U1XuAm1r8kqSesIitUbhz4Pk/zbL94CQ/DrwHeAFdsehxwEuTHL9sUUpaCebNR1X1ReBkulFFd9AVim6gKwxJWl5vbkWG7XRF4Ccu8Lj3VdW1VfUdui+ovlNVF7aCy3u4/wv0aW9tBZvtwGtohRTgvwB/WlXXVNX3q+oCuqL43hR7aP0fDJxdVd+rqo/QjVZ83h6OmctdwBtbcew9wJeA41sB7N8CL2uFo88B76AbpTntE1X1v1rR5p/maDuJrvh2Vyse/f7AOT7GrkXrPxjY/tm2H7qi/x9U1Y2tsP9a4ImDo7Hb/u3tNf8ZeAjw40DacXcs4mcj9cU3Znyp9W26HDDtvoE6VbUD2E5X/CXdMiQ3tlkJ99CNFH7kLK/x6Bnn+RYL//Jn8Lh/4f4vv3bbP4d/GHj+bbplO1bRjZj++71YGmnwdf6e+38Gj0pycZLb2iyRdzH7z2DB5595nUleMDDL5h66GXeLeQ1J0hKxiK2+OhL4UlVd1T5EfQm4Anj6iOOStAK1UTtHVtUj6EY6/jDd9F5Jy2tmoeTBc3WcYd4vrGb0n7WQQvff/hnTRY5W6DiMvSv20Prf2ooog6+zZgHHznRbVdUs8T4a2F5V39zDa8wW68y2R7fjZp4fuhHtP5ZkNd0XChcChyV5JN0Ixr9p/X4YeNPAz2w73fIAs8bSivpvBd4G3JnknCQPne3ipZ74NrvO0PpXe3n8YdNPkjyYbrmL29Otf/0yutkkh1TVwcC97LrM2bQ7ZpznQXQzRvb29X+AbhmS2wf2125HLMytwA/taWmlueIAfmgghj9oMfxkVT0U+GVm/xks+PyD19m+UPufwIvpbqh7MPCFRb6GJGmJWMRWX30WWJ/kqW1dxB8Fngn83YjjkrQCJQHwdjQAACAASURBVDkq3fr8P0i3fMBftRHakkbvWwwUj5LsbfFoNnMVUm4FXlNVBw88HlRVg0tjLKTYcztdsXfwvfgPAbctItY1SQYLLdPx3g48PMlD9vAas8U6s+12uiL0zPNTVd8GrgVeAnyhqr4H/C3w34EvV9XX2zG3Av91xs/twKr627let6reXFVH0c3G+zHgt2e/fKkXPgf8UnuvsIm9X+riGW1d+f3p1sa+pi2j+BBgJ/A1YFWS3wPm+kLnUuCZA+d5JQv/vH9Ukv/Yis2/STfD5Oq9vIbZfJKuuH52koPS3cfgKXvo/9tJDmkzSV5CN1MGup/DDuCeJGtYfD6Y6zoPostBXwNId+NN1+GXpJ6xiK1eqqov093I483AP9JNR/1LYL51ISVpKbwJuIdumv49dEsKSOqHvwMel+SJ7WZnrxjCOU9PsratLf1y7i+k/E/g15Ic3b5kPyjJ8TMKxQtxDV3x/aVJHpBkI93NZfd0w7O5PAr4jXae59Ld9OzKVgD7W+APWuHoJ+luKHfRHs41m3cDv5vkB9sI69+jm8o/7WN0oxenlw6ZmrEN8CfAmUkeB5DkYS3WWSX51+1n/AC6n9N3gO/vZdzScnoJ3X/D99AtwfO/9vL4P6eb6bUdOKqdA+AquptO/1+6WRDfYY7ZHlV1PXB6O9cdwN0sfOmzy+jWnr+bbrmg/9jWjd4nbcmm/wA8Bvhqi+cX54njWrovBa7g/s9+v093s8d7W/t7FxnSrNdZVTcAr6ebXXIn8Hjg/yzyNSRJSyS7zj6UJEmS+ifJLcCvVtVft+1XAI+pql9O8jvAb9EtDXIm3Y3F1lfV1iTnA9uqavrmjL8K/HJVbWzbjwG+WFWrBl7nT+kKHI+mK3q8qI06po2yfBWwvr3ex4Ffqapvzoxxnut5HPDHdMtw3EZ3A+v3zby2ec5xCt2Xap9t8d4JvLiqPtT2r6UrIP8buqLNH1bVn8z1GnO0PRD4/4DpovNfAC9t64uT5GnAB4GNVfWxJEcC1wEntjW6p8/zfOCldKO67wW2VNWvtH1F+3217WOBNwA/Qle0u4puJPeO+X6u0riZmaNG8PqvYAH5Zhni2CUPSJI0k0VsSZIkSZJGwCL2fXFYxJYk7ZHLiUiSJEmSNKGSfCDJjlkeLx91bMO0Uq5TklYqR2JLkiRJQ5bkh4Ab5th9RFV9dYHn+RNgthGS76qqX1tsfJIkSdI4mXckdpLzktyV5AsDbX+Y5ItJPp/kfUkOHth3ZpKtSb7U1sibbt/U2rYm2TzQfniSa5LclOQ97U7KJDmgbW9t+9cN66IljS9zkiRpHFTVV6vqwXM8FlTAbuf5tTnOYQFbkiRJK8a8I7GT/AywA7iwqo5sbccBH6mqnUleB1BVL0tyBN0dzJ9MdyOcvwZ+rJ3q/wI/T3dH4k8Bz6uqG5JcAry3qi5uI03+rqrenuS/AT9ZVb+W5ETgF6pqT3cyBuCRj3xkrVu3bt4L/9a3vsVBBx00b79RMsbhMMbhWGiM11577der6geXKo5xykkLzUfQz78BY1oYY1q45Y5rqfPRuPE90vIyxuGYpBjNSfdbSD7qy+++L3FAf2Ixjt31JRbzkaSJVlXzPoB1wBfm2PcLwEXt+ZnAmQP7rgJ+uj2uGmg/sz0CfB1Y1drv6zd9bHu+qvXLfLEeddRRtRAf/ehHF9RvlIxxOIxxOBYaI/DpWkBe2ZfHuOSkheajvfn5LidjWhhjWrjljms58tE4PXyPtLyMcTgmKUZz0t7lo7787vsSR1V/YjGO3fUlFvORDx8+JvkxjBs7/grwgfZ8DXDrwL5trW2u9kcA91TVzhntu5yr7b+39ZekPTEnSZIkNaNaik2SJGmYVu3LwUl+B9gJXDTdNEu3Yva1t2sP/fd0rtniOA04DWD16tVMTU3NHXSzY8eOBfUbJWMcDmMcjnGIsQ85aTH5CPr58zWmhTGmhetrXMOW5LHAewaafgT4PeDC1r4OuAX4z1V1d5IAbwKeAXwbOKWqPtPOdTLwu+08r66qC1r7UcD5wIHAlcBLqsq7dUuazfnAW+ly0LQtdLPVppdiOxOYXortROBxtKXYkkwvxfY2BpZiS3J5Vd0AvA54Q92/FNupwNuX4bokSdIKsugidvtQ9Uzg2IEPTduAwwa6rQVub89na/86cHCSVW1k42D/6XNtS7IKeBiwfbZYquoc4ByADRs21MaNG+eNf2pqioX0GyVjHA5jHI6+x9iXnLSYfAT9/Pka08IY08L1Na5hq6ovAU8ESLIfcBvwPmAz8OGqOruNYtwMvAx4OrC+PY6mK/4cneThwFnABrovza5tRaO7W5/TgKvpitibuH8WiiTdp6r+ZuYNqavqQwObVwPPac9PAC6uqu8CX0myle7eIgBbq+pmgCQXAyckuRF4KvBLrc8FwCuwiC1JkoZsUcuJJNlE96HrWVX17YFdlwMnJjkgyeF0H8Y+SXfTtPVtqtn+dN/uX94KTR/l/jdNJwOXDZzr5Pb8OXQ3bXOEkaTdmJMk9dixwJer6u/pikMXtPYLgGe35yfQ3ay2qupqui/TDgWeBmypqu2tcL0F2NT2PbSqPtHy0IUD55KkvbVUS7FJkiQNzbwjsZO8G9gIPDLJNroRQWcCBwBbuhmwXF1Vv1ZV1ye5BLiBbkr/6VX1/XaeF9PdGG0/4Lyqur69xMuAi5O8GvgscG5rPxd4Z/v2fztdkUnSCmdOkjRmTgTe3Z6vrqo7AKrqjiSPau17WzRa057PbN+NS66NjjEOhzEurSVeim3ma+1VPurLz7UvcUB/YjGO3fUllr7EIUlLYd4idlU9b5bmc2dpm+7/GuA1s7RfSTfddWb7zdw/RW2w/TvAc+eLT9LKYk6SNC7aTI9n0X3Rtseus7TtqTi04KKRS66NjjEOhzEunWVYim0Xe5uP+vJz7Usc0J9YjGN3fYmlL3FI0lJY1HIikiRJmtfTgc9U1Z1t+862FAjt37ta+1xFoz21r52lXZIWZJmWYpMkSRqaRd/Ycdxdd9u9nLL5ivu2bzn7+BFGI2mlMydJE+l53L+UCNy/tv7Z7L7m/ovbjdKOBu5ty41cBbw2ySGt33HAmVW1Pck3kxwDXAO8AHjL0l+OJIB1A/+/Bjh/00EjimRhRrgUWy/N/P35nkuSpPGwYovYkiRJSyXJg4CfB/7rQPPZwCVJTgW+yv1LFF0JPAPYCnwbeCFAK1a/im4EJMArq2p7e/4i4HzgQLobsk3flE2SdjGqpdgkSZKGySK2JEnSkLXp+Y+Y0fYN4NhZ+hZw+hznOQ84b5b2TwNHDiVYSZIkSeo518SWJEmSJEmSJPWWRWxJkiRJkiRJUm9ZxJYkSZIkSZIk9ZZFbEmSJEmSJElSb1nEliRJkiRJkiT1lkVsSZIkSZIkSVJvWcSWJEmSJEmSJPWWRWxJkiRJkiRJUm+tGnUAkiRJkiQthXWbrxh1CJIkaQgsYkuSJEmSJoJFa0mSJpPLiUiSJEmSJEmSessitiRJkiRJkiSptyxiS5IkSZIkSZJ6yyK2JEmSJEmSJKm3LGJLkiRJkiRJknrLIrYkSdKQJTk4yaVJvpjkxiQ/neThSbYkuan9e0jrmyRvTrI1yeeTPGngPCe3/jclOXmg/agk17Vj3pwko7hOSZIkSVoO8xaxk5yX5K4kXxhoW/IPYXO9hqSVzZwkaUy8CfhgVf048ATgRmAz8OGqWg98uG0DPB1Y3x6nAW+HLu8AZwFHA08GzhrIPW9vfaeP27QM1yRJkiRJI7GQkdjns/sHo+X4EDbXa0ha2c7HnCSpx5I8FPgZ4FyAqvpeVd0DnABc0LpdADy7PT8BuLA6VwMHJzkUeBqwpaq2V9XdwBZgU9v30Kr6RFUVcOHAuSRJkiRp4qyar0NV/U2SdTOaTwA2tucXAFPAyxj4EAZc3abSHtr6bqmq7QBJpj+ETdE+hLX26Q9hH9jDa0hawcxJksbAjwBfA/4syROAa4GXAKur6g6AqrojyaNa/zXArQPHb2tte2rfNkv7bpKcRvfFHKtXr2Zqamre4Hfs2LGgfqNkjMNhjItzxuN37rLdxxi1cOs2X7HL9i1nHz+iSCRJ0p7MW8Sew3J8CJvrNXazmA9oqw/c9Q1oH994jsMbYmMcDmPcZ73JSYvJR9DPnNTH37kxLUwfY4L+xrUEVgFPAn69qq5J8ib2PHtjtvWsaxHtuzdWnQOcA7Bhw4bauHHjHsLoTE1NsZB+o2SMw2GMi3PKjKLn+ZsO6l2Mg5KcBzwTuKuqjmxtDwfeA6wDbgH+c1Xd3ZZRexPwDODbwClV9Zl2zMnA77bTvrqqLmjtR9HNlDsQuBJ4SRtAIEmSNDSLLWLPZck+hO3JYj6gveWiy3j9dfdf/i0nzX/Mcuvjm/aZjHE4jHHJLHtOWkw+gn7mpD7+zo1pYfoYE/Q3riWwDdhWVde07Uvpith3Jjm0fRF2KHDXQP/DBo5fC9ze2jfOaJ9q7Wtn6S9JszkfeCvd0kPTppdJOzvJ5rb9MnZdiu1oumXWjh5Yim0D3Xuja5Nc3pY6ml6K7Wq6IvYmullskiRJQ7OQNbFnc2f78MVefAibq32uD2FzvYYkzWROktQbVfUPwK1JHtuajgVuAC4Hpm8kezJwWXt+OfCCdjPaY4B728yPq4DjkhzS1u0/Driq7ftmkmPaqMkXDJxLknZRVX8DbJ/R7Br9kiRprCx2JPb0h7Cz2f1D2IuTXEz3zf29bbTRVcBrB26cdhxwZlVtT/LN9oHtGroPYW+Z5zUkaSZzkqS++XXgoiT7AzcDL6QbPHBJklOBrwLPbX2vpJu6v5Vu+v4LAVpOehXwqdbvldNr+QMv4v7p+x/AUY+S9k7v1+hf7BJUM9cs31szX7NPS2H1JRbj2F1fYulLHJK0FOYtYid5N91U1kcm2UY3jexslv5D2FyvIWkFMydJGgdV9Tm6afczHTtL3wJOn+M85wHnzdL+aeDIfQxTkmbqzRr9i12Cauaa5Xtr5pJufVoKqy+xGMfu+hJLX+KQpKUwbxG7qp43x64l/RBWVd+Y7TUkrWzmJEmSpH3mGv2SJGmsLHZNbEmSJEnSeHKNfkmSNFYWuya2JEmSJKnnRrgUmyRJ0tBYxJYkSZKkCTWqpdgkSZKGySK2JEmSJEnAuhk3hjx/00EjikSSJA1yTWxJkiRJkiRJUm9ZxJYkSZIkSZIk9ZZFbEmSJEmSJElSb1nEliRJkiRJkiT1lkVsSZIkSZIkSVJvWcSWJEmSJEmSJPWWRWxJkiRJkiRJUm9ZxJYkSZIkSZIk9ZZFbEmSJEmSJElSb1nEliRJkiRJkiT1lkVsSZKkIUtyS5Lrknwuyadb28OTbElyU/v3kNaeJG9OsjXJ55M8aeA8J7f+NyU5eaD9qHb+re3YLP9VSpIkSdLysIgtSZK0NP59VT2xqja07c3Ah6tqPfDhtg3wdGB9e5wGvB26ojdwFnA08GTgrOnCd+tz2sBxm5b+ciRJkiRpNCxiS5IkLY8TgAva8wuAZw+0X1idq4GDkxwKPA3YUlXbq+puYAuwqe17aFV9oqoKuHDgXJIkSZI0cSxiS5IkDV8BH0pybZLTWtvqqroDoP37qNa+Brh14NhtrW1P7dtmaZckSZKkibRq1AFIkiRNoKdU1e1JHgVsSfLFPfSdbT3rWkT77ifuCuinAaxevZqpqak9Bg2wY8eOBfUbJWMcDmNcnDMev3OX7T7GKEmSNGksYkuSJA1ZVd3e/r0ryfvo1rS+M8mhVXVHWxLkrtZ9G3DYwOFrgdtb+8YZ7VOtfe0s/WeL4xzgHIANGzbUxo0bZ+u2i6mpKRbSb5SMcTiMcXFO2XzFLtvnbzqodzFKkiRNmn1aTiTJbyW5PskXkrw7yQOTHJ7kmiQ3JXlPkv1b3wPa9ta2f93Aec5s7V9K8rSB9k2tbWuSzbtHIEn3MydJ6oMkByV5yPRz4DjgC8DlwMmt28nAZe355cAL0jkGuLctN3IVcFySQ9oNHY8Drmr7vpnkmCQBXjBwLkmSJEmaOIsuYidZA/wGsKGqjgT2A04EXge8oarWA3cDp7ZDTgXurqrHAG9o/UhyRDvuccAm4I+T7JdkP+BtwNOBI4Dntb6StBtzkqQeWQ18PMnfAZ8ErqiqDwJnAz+f5Cbg59s2wJXAzcBW4H8C/w2gqrYDrwI+1R6vbG0ALwLe0Y75MvCBZbguSZIkSRqJfV1OZBVwYJJ/Bh4E3AE8Ffiltv8C4BXA24ET2nOAS4G3ttFDJwAXV9V3ga8k2Uo35RZga1XdDJDk4tb3hn2MWdLkMidJGrmWJ54wS/s3gGNnaS/g9DnOdR5w3iztnwaO3OdgJa1oSX4L+FW6dfWvA14IHApcDDwc+Azw/Kr6XpIDgAuBo4BvAL9YVbe085xJN0Dg+8BvVNVVy3wpkiRpwi26iF1VtyX5H8BXgX8CPgRcC9xTVdN3O9kGrGnP1wC3tmN3JrkXeERrv3rg1IPH3Dqj/ejZYlnMTYtWH7jrTVn6eDOWcbhJjDEOhzHuu77kpMXkI+hnTurj79yYFqaPMUF/45KklWhgFtsRVfVPSS6hm432DLpZbBcn+RO64vTbGZjFlmR6ttsvzpjF9mjgr5P8WFV9fwSXJUmSJtSii9htbcYTgMOBe4C/oJtmP1NNHzLHvrnaZ1vqpGZpW9RNi95y0WW8/rr7L/+Wk+Y/Zrn18UY2MxnjcBjjvutLTlpMPoJ+5qQ+/s6NaWH6GBP0Ny7t6rrb7t3lxnW3nH38CKORtMSWchbbJ5bpGiRJ0gqwL8uJ/Bzwlar6GkCS9wL/Bjg4yao28nEtcHvrvw04DNiWZBXwMGD7QPu0wWPmapekmcxJkiRJC7RMs9jus7ez1RY6e+e62+7dZfuMx897yF7p0yyivsRiHLvrSyx9iUOSlsK+FLG/ChyT5EF0b3qOBT4NfBR4Dt06aicDl7X+l7ftT7T9H6mqSnI58OdJ/ohu+tl6upsgBVif5HDgNropatMjAiRpJnOSJEnSAi3DLLZdG/ZyttpCZ+8MzhxZCudvOqg3s4j6MqPJOHbXl1j6EockLYV9WRP7miSX0t3sYyfwWbo3JVcAFyd5dWs7tx1yLvDONr1sO10BiKq6vq2/dkM7z+nT66cleTFwFbAfcF5VXb/YeCVNNnOSJEnSXlmOWWySJElDsS8jsamqs4CzZjTfTLcG2sy+3wGeO8d5XgO8Zpb2K4Er9yVGSSuHOUmSJGnBlnoWmyRJ0tDsUxFbkiRJkjR+lmMWmyRJ0rBYxJYkSZKkFWipZ7FNgutuu3eXdbdvOfv4EUYjSdLK9QOjDkCSJEmSJEmSpLlYxJYkSZIkSZIk9ZZFbEmSJEmSJElSb1nEliRJkiRJkiT1lkVsSZIkSZIkSVJvWcSWJEmSJEmSJPWWRWxJkiRJkiRJUm9ZxJYkSVoCSfZL8tkk72/bhye5JslNSd6TZP/WfkDb3tr2rxs4x5mt/UtJnjbQvqm1bU2yebmvTZIkSZKWk0VsSZKkpfES4MaB7dcBb6iq9cDdwKmt/VTg7qp6DPCG1o8kRwAnAo8DNgF/3Arj+wFvA54OHAE8r/WVJEmSpIlkEVuSJGnIkqwFjgfe0bYDPBW4tHW5AHh2e35C26btP7b1PwG4uKq+W1VfAbYCT26PrVV1c1V9D7i49ZUkSZKkiWQRW5IkafjeCLwU+Je2/Qjgnqra2ba3AWva8zXArQBt/72t/33tM46Zq12SJEmSJtKqUQcgSZI0SZI8E7irqq5NsnG6eZauNc++udpnG4RQs7SR5DTgNIDVq1czNTU1d+DN6gPhjMfvvG97Iccstx07dvQyrkHGOBx9jHHwvw/oZ4ySJEmTxiK2JEnScD0FeFaSZwAPBB5KNzL74CSr2mjrtcDtrf824DBgW5JVwMOA7QPt0waPmat9F1V1DnAOwIYNG2rjxo3zBv+Wiy7j9dfd/xbxlpPmP2a5TU1NsZBrGSVjHI4+xnjK5it22T5/00G9i1GSJGnSuJyIJEnSEFXVmVW1tqrW0d2Y8SNVdRLwUeA5rdvJwGXt+eVtm7b/I1VVrf3EJAckORxYD3wS+BSwPsnhSfZvr3H5MlyaJEmSJI2EI7ElSZKWx8uAi5O8GvgscG5rPxd4Z5KtdCOwTwSoquuTXALcAOwETq+q7wMkeTFwFbAfcF5VXb+sVyJJkiRJy8gitiRJ0hKpqilgqj2/GXjyLH2+Azx3juNfA7xmlvYrgSuHGKokSZIk9ZbLiUiSJEmSJEmSessitiRJkiRJkiSpt/apiJ3k4CSXJvlikhuT/HSShyfZkuSm9u8hrW+SvDnJ1iSfT/KkgfOc3PrflOTkgfajklzXjnlzkuxLvJImmzlJkiRJS2nd5it2e0iSpKW3ryOx3wR8sKp+HHgCcCOwGfhwVa0HPty2AZ4OrG+P04C3AyR5OHAWcDTdOpFnTReZWp/TBo7btI/xSpps5iRJkqQFWuoBACuFRW1JkpbeoovYSR4K/AxwLkBVfa+q7gFOAC5o3S4Ant2enwBcWJ2rgYOTHAo8DdhSVdur6m5gC7Cp7XtoVX2iqgq4cOBckrQLc5IkSdJeW+oBAJIkSUOxah+O/RHga8CfJXkCcC3wEmB1Vd0BUFV3JHlU678GuHXg+G2tbU/t22Zp302S0+jeSLF69WqmpqbmDX71gXDG43fet72QY5bbjh07ehnXIGMcDmMcil7kpMXkI+hnTurj79yYFqaPMUF/45KklWhgAMAp0A0AAL6X5ARgY+t2ATAFvIyBAQDA1W0U96Gt75aq2t7Ou4Vuttq7l+taJEnS5NuXIvYq4EnAr1fVNUnexP3f0s9mtrVjaxHtuzdWnQOcA7Bhw4bauHHjHsLovOWiy3j9dfdf/i0nzX/McpuammIh1zJKxjgcxjgUvchJi8lH0M+c1MffuTEtTB9jgv7GJUkr1FIPAJAkSRqafSlibwO2VdU1bftSuoLRnUkObW94DgXuGuh/2MDxa4HbW/vGGe1TrX3tLP0laTbmJEmSpIVb6gEAux68l7PVFjp7Z3Am21KYOVtuIZZq1lFfZjQZx+76Ektf4pCkpbDoInZV/UOSW5M8tqq+BBwL3NAeJwNnt38va4dcDrw4ycV066Xd24pKVwGvHVg37TjgzKranuSbSY4BrgFeALxlsfFKmmzmJEmSpL2y1AMAdrG3s9UWOnvnlCW+keIZj9+5y2y5hViqGXV9mdFkHLvrSyx9iUOSlsK+jMQG+HXgoiT7AzcDL6S7WeQlSU4Fvgo8t/W9EngGsBX4dutLKwy9CvhU6/fK6fXUgBcB5wMHAh9oD0maizlJkiRpAZZ6AMByXoskSZp8+1TErqrPARtm2XXsLH0LOH2O85wHnDdL+6eBI/clRkkrhzlJkiRpryz1AABJkqSh2NeR2JIkSZKkMbTUAwAkSZKG5QdGHYAkSZIkSZIkSXOxiC1JkiRJkiRJ6i2L2JIkSZIkSZKk3rKILUmSJEmSJEnqLYvYkiRJQ5TkgUk+meTvklyf5Pdb++FJrklyU5L3JNm/tR/Qtre2/esGznVma/9SkqcNtG9qbVuTbF7ua5QkSZKk5WQRW5Ikabi+Czy1qp4APBHYlOQY4HXAG6pqPXA3cGrrfypwd1U9BnhD60eSI4ATgccBm4A/TrJfkv2AtwFPB44Antf6SpIkSdJEsogtSZI0RNXZ0TYf0B4FPBW4tLVfADy7PT+hbdP2H5skrf3iqvpuVX0F2Ao8uT22VtXNVfU94OLWV5IkSZIm0qpRByBJkjRp2mjpa4HH0I2a/jJwT1XtbF22AWva8zXArQBVtTPJvcAjWvvVA6cdPObWGe1HzxHHacBpAKtXr2Zqamre2FcfCGc8fud92ws5Zrnt2LGjl3ENMsbh6GOMg/99QD9jlCRJmjQWsSVJkoasqr4PPDHJwcD7gJ+YrVv7N3Psm6t9tpl0NUsbVXUOcA7Ahg0bauPGjXsOHHjLRZfx+uvuf4t4y0nzH7PcpqamWMi1jJIxDkcfYzxl8xW7bJ+/6aDexajRWjfjb+SWs48fUSSSJE0OlxORJElaIlV1DzAFHAMcnGS6OrwWuL093wYcBtD2PwzYPtg+45i52iVJkiRpIlnEliRJGqIkP9hGYJPkQODngBuBjwLPad1OBi5rzy9v27T9H6mqau0nJjkgyeHAeuCTwKeA9UkOT7I/3c0fL1/6K5MkSZKk0XA5EUmSpOE6FLigrYv9A8AlVfX+JDcAFyd5NfBZ4NzW/1zgnUm20o3APhGgqq5PcglwA7ATOL0tU0KSFwNXAfsB51XV9ct3eZIkSZK0vCxiS5IkDVFVfR74qVnabwaePEv7d4DnznGu1wCvmaX9SuDKfQ5WkiRJksaARWxJkiRJkpaIN3qUJGnfuSa2JEmSJEmSJKm3LGJLkiRJktRT6zZfwbrNV3DdbffuNqpbkqSVwiK2JEmSJEmSJKm3LGJLkiRJkiRJknrLIrYkSZIkSZIkqbf2uYidZL8kn03y/rZ9eJJrktyU5D1J9m/tB7TtrW3/uoFznNnav5TkaQPtm1rb1iSb9zVWSZPPnCRJkiRJkjRZhjES+yXAjQPbrwPeUFXrgbuBU1v7qcDdVfUY4A2tH0mOAE4EHgdsAv64FaH2A94GPB04Anhe6ytJe2JOkiRJWqClHAAgSZI0LPtUxE6yFjgeeEfbDvBU4NLW5QLg2e35CW2btv/Y1v8E4OKq+m5VfQXYCjy5PbZW1c1V9T3g4tZXkmZlTpIkSdprSzIAYJlilyRJK8SqfTz+jcBLgYe07UcA91TVzra9DVjTnq8BbgWoqp1J7m391wBXD5xz8JhbZ7QfPVsQSU4DTgNYvXo1U1NT8wa++kA44/E779teyDHLbceOHb2Ma5AxDocxDs3Ic9Ji8hH0HY4K9AAAFnJJREFUMyf18XduTAvTx5igv3FJ0ko1MADgNcB/HxgA8EutywXAK4C30315/4rWfinw1pkDAICvJJkeAPCJZboMSZK0Aiy6iJ3kmcBdVXVtko3TzbN0rXn2zdU+2yjxmqWNqjoHOAdgw4YNtXHjxtm67eItF13G66+7//JvOWn+Y5bb1NQUC7mWUTLG4TDGfdeXnLSYfAT9zEl9/J0b08L0MSbob1yStIIt9QCA++ztF/0L/eJzcBDAUpg50GAY9vYL3enXn45l1F8I9+VL6b7EAf2JpS9xSNJS2JeR2E8BnpXkGcADgYfSvQk6OMmq9sZnLXB7678NOAzYlmQV8DBg+0D7tMFj5mqXpJnMSZIkSQu0DAMAdm3Yyy/6F/rF5ymbr5i3z7444/E7dxloMAx7O1hh+hqnYxn1YIe+fCndlzigP7H0JQ5JWgqLXhO7qs6sqrVVtY5uDbSPVNVJwEeB57RuJwOXteeXt23a/o9UVbX2E9uNQg4H1gOfBD4FrG83Ftm/vcbli41X0mQzJ0mSJO2V6QEAt9Dd6+OpDAwAaH1mGwDAXgwAkCRJGop9urHjHF5Gt57aVrrpZee29nOBR7T2/w5sBqiq64FLgBuADwKnV9X326jJFwNX0d1o5JLWV5L2hjlJkiRphmUYACBJkjQ0Q5kXVVVTwFR7fjPdjTxm9vkO8Nw5jn8N3c1EZrZfCVw5jBglrRzmJEmSpEV7GXBxklcDn2XXAQDvbAMAttMVvqmq65NMDwDYSRsAsPxhS5KkSbYUI7ElSZJWrCSHJflokhuTXJ/kJa394Um2JLmp/XtIa0+SNyfZmuTzSZ40cK6TW/+bkpw80H5UkuvaMW9OMtuatJK0IFU1VVXPbM9vrqonV9Vjquq5VfXd1v6dtv2Ytv/mgeNfU1U/WlWPraoPjOo6JEnS5BruHSokSZK0Ezijqj6T5CHAtUm2AKcAH66qs5NsplvG6GXA0+mm368HjgbeDhyd5OHAWcAGupukXZvk8qq6u/U5DbiabobIJsDCkaSJt26Jb+QoSZL6yZHYkiRJQ1RVd1TVZ9rzb9Kto78GOAG4oHW7AHh2e34CcGF1rqa7qdqhwNOALVW1vRWutwCb2r6HVtUn2nq0Fw6cS5IkSZImjiOxJUmSlkiSdcBPAdcAq6vqDugK3Uke1bqtAW4dOGxba9tT+7ZZ2md7/dPoRmyzevVqpqam5o159YFwxuN33re9kGOW244dO3oZ1yBjHI4+xjj43wf0M0b128zR5LecffyIIpEkaXxYxJYkSVoCSR4M/CXwm1X1j3tYtnq2HbWI9t0bq84BzgHYsGFDbdy4cZ6o4S0XXcbrr7v/LeItJ81/zHKbmppiIdcySsY4HH2M8ZQZBcjzNx3Uuxg1XixqS5I0P5cTkSRJGrIkD6ArYF9UVe9tzXe2pUBo/97V2rcBhw0cvha4fZ72tbO0S5IkSdJEsogtSZI0ROmGXJ8L3FhVfzSw63Lg5Pb8ZOCygfYXpHMMcG9bduQq4LgkhyQ5BDgOuKrt+2aSY9prvWDgXJIkSZI0cVxORJIkabieAjwfuC7J51rby4GzgUuSnAp8FXhu23cl8AxgK/Bt4IUAVbU9yauAT7V+r6yq7e35i4DzgQOBD7SHJEmSJE0ki9iSJElDVFUfZ/Z1qwGOnaV/AafPca7zgPNmaf80cOQ+hClJ6qmZa2RLkiSL2JIkSZIkjQ1vBClJWolcE1uSJEmSJEmS1FsWsSVJkiRJkiRJvWURW5IkSZIkSZLUWxaxJUmSJEmSJEm9ZRFbkiRJkiRJktRbFrElSZIkSZIkSb1lEVuSJEmSJEmS1FsWsSVJkiRJkiRJvWURW5IkSZIkSZLUWxaxJUmSJEmSJEm9tegidpLDknw0yY1Jrk/yktb+8CRbktzU/j2ktSfJm5NsTfL5JE8aONfJrf9NSU4eaD8qyXXtmDcnyb5crKTJZU6SJElauOV47yRJkjQs+zISeydwRlX9BHAMcHqSI4DNwIer/v/27j7Wsqq84/j3yQwgL+IMtUwogx1I0JZIKzAFrG0zSgMDko5Npw0W5aU0RG0rtKQVsQ19MQ00rVWqlRCgQKWAvFSoQpGoU9IoCCjCKAKjELhlZETeMRFGnv6x14UzZ/a5986dc85e997vJ9m55+zzsn6z9rnPXnfN2Xvn/sCXyn2Ao4H9y3Iq8GloBknA2cBhwKHA2ZMDpfKcU3tet3o78kqa36xJkiRJMzeOsZMkSdJQzHoSOzM3ZuY3yu3ngPuAvYE1wKXlaZcC7yq31wCXZeM2YElE7AUcBdySmU9m5lPALcDq8tjumfm1zEzgsp73kqQtWJMkSZJmbtRjpzH+UyRJ0gKweBhvEhErgIOA24FlmbkRmoFRROxZnrY38GjPyybKuqnWT7Ssb2v/VJpvA7Bs2TLWrVs3beZlO8MZB25+5f5MXjNuzz//fJW5eplxOMw4XF3WpNnUI6izJtW4zc00MzVmgnpzSdJCN6KxU38b2zRGGrTP6B0vjUP/GK1Lg7KMe99ay/68lhxQT5ZackjSKGz3JHZE7AZcC5yemc9OcYrYtgdyFuu3Xpl5AXABwMqVK3PVqlXTpIZ/ufx6/uneV//5Dx8//WvGbd26dczk39IlMw6HGYen65o0m3oEddakGre5mWamxkxQb65RiIiLgWOBTZn55rJuD+AqYAXwMPB7mflUOb/+J4BjgB8DJ01+O7KcW/Yvy9t+NDMvLesPAS4BdgZuBE4rR4lI0jYZ4dhpyxXbOEYatM846cwvTPm6YTvjwM1bjNG6NCjLuMeNtezPa8kB9WSpJYckjcL2nBObiNiBZsBzeWZeV1Y/Xg4ro/zcVNZPAPv0vHw58Ng065e3rJekVtYkSRW5hK0Pp/cc/ZKqMuKxkyRJ0tDMehK7fGvoIuC+zPxYz0M3AJNXpD4RuL5n/QnlqtaHA8+Uw9RuBo6MiKXlD7MjgZvLY89FxOGlrRN63kuStmBNklSTzLwVeLJvtefol1SNUY+dxvKPkCRJC8b2HBf1NuC9wL0RcXdZdxZwDvDZiDgFeAT43fLYjTSHyW6gOVT2ZIDMfDIi/g64ozzvbzNz8o++9/PqobI3lUWS2liTJNVu7NcNkaQpjGPsJEmSNBSznsTOzP+l/fxnAEe0PD+BPxrwXhcDF7esvxN482wzSlo4rEmS5rCRXTfEi193x4zDUWPG/gvr1ZhxJsYxdpIkSRqWOq5QIUmSNP89HhF7lW9hz/Q8s6v61q9jG87R78Wvu2PG4agxY/+FBS9ZvWt1GSVJkuab7bqwoyRJkmbMc/RLkiRJ0iz4TWxJkqQhi4graL5F/fqImADOxnP0S5IkSdKsOIktSZI0ZJn57gEPeY5+SZIkSdpGnk5EkiRJkiRJklQtJ7ElSZIkSZIkSdVyEluSJEmSJEmSVC0nsSVJkiRJkiRJ1XISW5IkSZIkSZJULSexJUmSJEmSJEnVchJbkiRJkiRJklQtJ7ElSZIkSZIkSdVyEluSJEmSJEmSVC0nsSVJkiRJkiRJ1XISW5IkSZIkSZJULSexJUmSJEmSJEnVWtx1AEmSJEmS2qw48wtdR5jz2vrw4XPeOdT33N73kyRpOk5iS5IkSZI0TzjxL0maj5zEliRJkiRpjprNpPV0r/Gb1ZKk2lQ/iR0Rq4FPAIuACzPznI4jSVrArEmSamE9klQTa9L8MtUk9xkHbmYOTCVIkuaZqi/sGBGLgE8BRwMHAO+OiAO6TSVpobImSaqF9UhSTaxJkiRp1KqexAYOBTZk5vcz80XgSmBNx5kkLVzWJEm1sB5Jqok1SZIkjVTtxwDtDTzac38COKyjLNPyCs3SvFdtTbL+SAtOtfVI0oJkTZIkSSNV+yR2tKzLrZ4UcSpwarn7fETcP4P3fj3wxCvvce6s8k1pCO+5RcZKmXE45lPGnx91kA5NW5NmWY9gyDVpSDWtxs+lmWamxkww/lwLuh5BvWOkIaj1M97LjMNRfca3n+sYidGMkarY9h+sJAfUk6UtR0f7iir6o6gli/VI0rxV+yT2BLBPz/3lwGP9T8rMC4ALtuWNI+LOzFy5ffFGy4zDYcbhmAsZx2DamjSbegR19q+ZZsZMM1drrjnKMZIZt5sZh2MuZByDoY+RaunXWnJAPVnMsbVastSSQ5JGofZzYt8B7B8R+0bEjsBxwA0dZ5K0cFmTJNXCeiSpJtYkSZI0UlV/EzszN0fEHwM3A4uAizPz2x3HkrRAWZMk1cJ6JKkm1iRJkjRqVU9iA2TmjcCNI3jrbT7cvwNmHA4zDsdcyDhyC6wmmWlmzDRzteaakxZYPepnxuEw43DMhYwjN4KaVEu/1pID6slijq3VkqWWHJI0dJG51TWAJEmSJEmSJEmqQu3nxJYkSZIkSZIkLWALchI7IlZHxP0RsSEizuw6T7+I2CcivhIR90XEtyPitK4ztYmIRRHxzYj4fNdZ2kTEkoi4JiK+W/ryrV1n6hcRf1q28fqIuCIiXtN1JoCIuDgiNkXE+p51e0TELRHxYPm5tMuM80Vt9ajm+lNjzamxztRQV6wh9WvbRn2PR0ScV2rTPRFxcIUZjy/Z7omIr0bEL9eWsed5vxIRP42ItePK1tP2tBkjYlVE3F1qx/+MM19pf7pt/bqI+K+I+FbJeHIHGafdP9bwezMfDGtsNGibDdofTbX9IuLE8vwHI+LEnvWHRMS95TXnRUQMaqN/HBPNRTBvL8+5KpoLYhIRO5X7G8rjK3ra+3BZf39EHDVdn7W1ES1jl4765KzoG6+MsU82RcTmiHimp40u+uAzJcv6njY2RcRDpa3/jIglw9z+s+1PSapCZi6oheZCI98D9gN2BL4FHNB1rr6MewEHl9uvBR6oLWPJ9mfAfwCf7zrLgHyXAn9Ybu8ILOk6U1++vYGHgJ3L/c8CJ3Wdq2T5DeBgYH3Pun8Aziy3zwTO7TrnXF9qrEc1158aa05tdaaWumINqX9p20Z9jx8D3AQEcDhwe4UZfxVYWm4fXWPG8pxFwJdpzhW8traMwBLgO8Abyv09K8x41mTNAH4WeBLYccwZp90/1vB7M9eXYY6NBm2zQfujQdsP2AP4fvm5tNyerD1fB95aXnMTcHRZv1Ub9I1jyj76uHL7fOD95fYHgPPL7eOAq8rtA0p/7ATsW/pp0VR91tYGLWOXDvrk74Gn6RuvjLFP/pqm5jzZ00YXn4vLS46He9r4YE8b5/bkGMr2n01/dl0XXFxcXCaXhfhN7EOBDZn5/cx8EbgSWNNxpi1k5sbM/Ea5/RxwH83ERDUiYjnwTuDCrrO0iYjdaf4gugggM1/MzKe7TdVqMbBzRCwGdgEe6zgPAJl5K82grtcamkEv5ee7xhpqfqquHtVaf2qsORXXmc7rijWkfgO2Ua81wGXZuA1YEhF7jSddY7qMmfnVzHyq3L0NWD6WYFtmmK4fAf4EuBbYNPpEW5tBxt8HrsvMR8rzx55zBhkTeG35NuNu5bmbx5HtlQAz2z92/nszDwxtbDTFNhu0Pxq0/Y4CbsnMJ0vNuQVYXR7bPTO/lpkJXNb3Xr1trKVnHFM+y+8ArhmQY/K11wBHlOevAa7MzJ9k5kPAhtJfrX02oI3foX3sMu4+uRrYlS3HKxvH2CcfpakjT02xzcbxuVhZcuze08Z5PW307tuGsf1n25+SVIWFOIm9N/Boz/0JKpigGaQc2nMQcHu3SbbyceAvgJe7DjLAfsAPgX+L5rC9CyNi165D9crM/wP+EXiEZtD2TGZ+sdtUU1qWmRuh+aMA2LPjPPNB1fWosvpTY82prs5UXlesIXNL1fWpxSk032KrSkTsDfw2zTfgavVGYGlErIuIuyLihK4Dtfgk8Is0/yl3L3BaZna2P5hi/zjXfm9qNJI+7Ntmg/ZHg9qeav3EgKz9bezDluOYnwGezszNLa99pb3y+DPl+duar62NFbSPXcbdJ98EXqJnvALc1UGfvDQg37g+F5NtLB7wXn/Aq/u2YWz/2fanJFVhIU5iR8u6HHuKGYiI3Wi+uXN6Zj7bdZ5JEXEssCkz7+o6yxQW0xya9enMPAh4geaQrWqU86ytoTlU6+eAXSPiPd2m0phVW49qqj8V15zq6ox1RUNUbX3qFxFvp5nE/lDXWVp8HPhQZv606yBTWAwcQvMt0aOAv4qIN3YbaStHAXfT1LW3AJ8sR8OM3TT7xznze1OxoffhNoxpBrW9rev72z8WeLlvHDPVa4eVo209bNvYZVR9shTYgZ7xCs1poQa9dpR9Mt3nq6s2TqY54uTyEeQYZnZJGpuFOIk9QfM/4ZOWU8kpHHpFxA40g63LM/O6rvP0eRvwWxHxMM3hSu+IiM90G2krE8BEZk5+Q+YamgFbTX4TeCgzf5iZLwHX0Zzfs1aPTx4SW352clj0PFNlPaqw/tRac2qsMzXXFWvI3FJlfeoXEb9Ec3j+msz8Udd5WqwEriz1ay3wrxFR26l0JoD/zswXMvMJ4FZg7BfJnMbJNKc8yczcQHPu/18Yd4gZ7B/nxO9N5YbahwO22aD90aC2p1q/vGX9Fm0ARwKLescxNP/BtaScSqP/ta+0Vx5/Hc0pJ7Y13xMtbTxK+9hl3H2yFnihZbwy7j7ZoS3fOPqgr42X+t7rAOAw4PhySpLZ5Gjb/rPtT0mqwkKcxL4D2L9cqXdHmgsZ3NBxpi2U81FdBNyXmR/rOk+/zPxwZi7PzBU0/fflzKzqm36Z+QPg0Yh4U1l1BM1Fi2ryCHB4ROxStvkRNOfqq9UNwORVt08Eru8wy3xRXT2qsf7UWnMqrTM11xVryNxyA3BCNA6nOTXNxq5D9YqIN9BMfLw3Mx/oOk+bzNw3M1eU+nUN8IHM/FzHsfpdD/x6RCyOiF1oJk5qqRuTHqGpZ0TEMuBNNBdRG5sZ7h+r/72ZA4Y2Nppimw3aHw3afjcDR0bE0vIN4iOBm8tjz0XE4aWtE/rea7KNx4B/7hvHHA98hWYyty3H5GvXludnWX9cROwUEfsC+9NcQLC1z8pr+tu4mvaxy7j75C3A5r7xync66JOlA9oYRx/0tvFsTxun01xs8+jM/DGvGsb2n21/SlIdsoKrS457obnC8AM0V9v9SNd5WvL9Gs1hO/fQHDp5N3BM17kGZF1FucJ2bQvN4OjO0o+fo1wtuqYF+Bvgu8B64N+BnbrOVHJdQXN+updo/kf+FJrzpH0JeLD83KPrnPNhqa0e1V5/aqs5NdaZGuqKNaT+ZcA2eh/wvvJ4AJ8qteleYGWFGS+kuSjXZK26s7aMfc+9BFhbY0bgz2kmkNbTnHKhqow0pxv4Yvksrgfe00HG1v1jbb8382EZ1thoim3Wuj+aavvRnJt4Q1lO7lm/snwmv0dz7vYo6we1sYoyjqG5tsbXy3teTdlfA68p9zeUx/frae8jpa37aSY5p+yztjZoGbt01Cfn0jdeGWOf/IjmVB0vM804ZcR9cC2v1r7nab6Z/RPgB7z6mT1/mNt/tv3p4uLiUsMyWUwlSZIkSZIkSarOQjydiCRJkiRJkiRpjnASW5IkSZIkSZJULSexJUmSJEmSJEnVchJbkiRJkiRJklQtJ7ElSZIkSZIkSdVyEluSJEmSJEmSVC0nsSVJkiRJkiRJ1XISW5IkSZIkSZJUrf8HuGDVu/9ssQEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1800x1440 with 25 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "Total.hist(bins=50,figsize=(25,20))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Date modification "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "list1=[]\n",
    "for i in range(0,len(Total)):\n",
    "    string1=Total[\"origination_date\"][i]\n",
    "    list1.append(string1[6])\n",
    "\n",
    "Total[\"origination_date\"]=list1\n",
    "\n",
    "list1=[]\n",
    "for i in range(0,len(Total)):\n",
    "    string1=Total[\"first_payment_date\"][i]\n",
    "    list1.append(string1[1])\n",
    "\n",
    "Total[\"first_payment_date\"]=list1\n",
    "\n",
    "Total['origination_date'] = Total['origination_date'].astype(int)\n",
    "Total['first_payment_date'] = Total['first_payment_date'].astype(int)\n",
    "\n",
    "Total[\"orgin-first\"]=Total[\"first_payment_date\"]-Total[\"origination_date\"]\n",
    "\n",
    "# Remove outlier\n",
    "Total.drop(Total[(Total['unpaid_principal_bal']>650000)].index, inplace=True)\n",
    "Total.reset_index(drop=True, inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Encoding Categorical variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import LabelEncoder,OneHotEncoder\n",
    "le_X = LabelEncoder()\n",
    "\n",
    "X = Total.iloc[:,1].values\n",
    "\n",
    "X = le_X.fit_transform(X)\n",
    "\n",
    "y = Total.iloc[:,2].values\n",
    "\n",
    "y = le_X.fit_transform(y)\n",
    "\n",
    "z = Total.iloc[:,12].values\n",
    "\n",
    "z = le_X.fit_transform(z)\n",
    "\n",
    "a = Total.iloc[:,15].values\n",
    "\n",
    "a = le_X.fit_transform(z)\n",
    "\n",
    "Total['source']=X\n",
    "\n",
    "Total['financial_institution']=y\n",
    "\n",
    "Total['loan_purpose']=z\n",
    "\n",
    "Total['insurance_type']=a"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Downsampling Majority class and creating multiple datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "Default=Total[0:636]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "Default_1=Default[0:500]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "Default_1_test=Default[501:636]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "All=pd.concat([Default_1,Default_1_test])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(635, 30)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "All.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "Paid_1=Total[636:50000]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(49364, 30)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Paid_1.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "Paid_1_test=Total[50001:100000]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(49999, 30)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Paid_1_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "Paid_2=Paid_1.sample(n=4000, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "Default_2=Default_1.sample(n=500, random_state=1,replace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "Paid_2_test=Paid_1_test.sample(n=150, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "Default_2_test=Default[501:636]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "Train2=pd.concat([Paid_2,Default_2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "Test2=pd.concat([Paid_2_test,Default_2_test])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "Train2.to_csv(\"train2.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "Test2.to_csv(\"test2.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "Paid_3=Paid_1.sample(n=4000, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "Default_3=Default_1.sample(n=500, random_state=1,replace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "Paid_3_test=Paid_1_test.sample(n=150, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "Default_3_test=Default[501:636]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "Train3=pd.concat([Paid_3,Default_3])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "Test3=pd.concat([Paid_3_test,Default_3_test])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "Train3.to_csv(\"train3.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "Test3.to_csv(\"test3.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "Paid_4=Paid_1.sample(n=4000, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "Default_4=Default_1.sample(n=500, random_state=1,replace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "Paid_4_test=Paid_1_test.sample(n=150, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "Default_4_test=Default[501:636]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "Train4=pd.concat([Paid_4,Default_4])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "Test4=pd.concat([Paid_4_test,Default_4_test])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "Train4.to_csv(\"train4.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "Test4.to_csv(\"test4.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "Paid_5=Paid_1.sample(n=4000, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "Default_5=Default_1.sample(n=500, random_state=1,replace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "Paid_5_test=Paid_1_test.sample(n=150, random_state=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "Default_5_test=Default[501:636]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "Train5=pd.concat([Paid_5,Default_5])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "Test5=pd.concat([Paid_5_test,Default_5_test])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "Train5.to_csv(\"train5.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "Test5.to_csv(\"test5.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4500, 30)"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Train5.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_train_5=Train5[\"m13\"]\n",
    "\n",
    "X_train_5=Train5.drop(['m13','loan_id'], axis=1)\n",
    "\n",
    "Y_test_5=Test5[\"m13\"]\n",
    "\n",
    "X_test_5=Test5.drop(['m13','loan_id'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "sc_X = StandardScaler()\n",
    "X_train = sc_X.fit_transform(X_train_5)\n",
    "X_test = sc_X.transform(X_test_5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4500, 28)"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 710,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'alpha': 0.001, 'penalty': 'l2'}"
      ]
     },
     "execution_count": 710,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn import linear_model\n",
    "clf = linear_model.SGDClassifier(loss='log')\n",
    "alpha = [0.00001,0.0001,0.001,0.01,0.1,1,10,100,1000,10000]\n",
    "penalty=['l1','l2']\n",
    "param_grid = {'alpha':alpha,'penalty':penalty}\n",
    "knn_cv = GridSearchCV(clf, param_grid, cv=10,scoring='f1')\n",
    "knn_cv.fit(X_train, Y_train_5) \n",
    "knn_cv.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SGDClassifier(alpha=0.001, average=False, class_weight=None,\n",
       "       early_stopping=False, epsilon=0.1, eta0=0.0, fit_intercept=True,\n",
       "       l1_ratio=0.15, learning_rate='optimal', loss='log', max_iter=None,\n",
       "       n_iter=None, n_iter_no_change=5, n_jobs=None, penalty='l2',\n",
       "       power_t=0.5, random_state=None, shuffle=True, tol=None,\n",
       "       validation_fraction=0.1, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn import linear_model\n",
    "from sklearn.calibration  import CalibratedClassifierCV\n",
    "alpha_optimal = linear_model.SGDClassifier(alpha=0.001,loss='log',penalty='l2')\n",
    "\n",
    "alpha_optimal.fit(X_train, Y_train_5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Macro-average quality numbers\n",
      "Precision: 0.8488, Recall: 0.7593, F1-measure: 0.7524\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.70      1.00      0.82       150\n",
      "           1       1.00      0.52      0.68       135\n",
      "\n",
      "   micro avg       0.77      0.77      0.77       285\n",
      "   macro avg       0.85      0.76      0.75       285\n",
      "weighted avg       0.84      0.77      0.76       285\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import precision_score,recall_score,f1_score\n",
    "# predict the response\n",
    "pred_LR = alpha_optimal.predict(X_test)\n",
    "\n",
    "precision = precision_score(Y_test_5, pred_LR, average='macro')\n",
    "recall = recall_score(Y_test_5, pred_LR, average='macro')\n",
    "f1 = f1_score(Y_test_5, pred_LR, average='macro')\n",
    " \n",
    "print(\"Macro-average quality numbers\")\n",
    "print(\"Precision: {:.4f}, Recall: {:.4f}, F1-measure: {:.4f}\".format(precision, recall, f1))\n",
    "\n",
    "print (metrics.classification_report(Y_test_5, pred_LR))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[150   0]\n",
      " [ 65  70]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.70      1.00      0.82       150\n",
      "           1       1.00      0.52      0.68       135\n",
      "\n",
      "   micro avg       0.77      0.77      0.77       285\n",
      "   macro avg       0.85      0.76      0.75       285\n",
      "weighted avg       0.84      0.77      0.76       285\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix,classification_report\n",
    "from pandas import DataFrame\n",
    "from sklearn.metrics import f1_score\n",
    "import seaborn as sn\n",
    "\n",
    "confm = confusion_matrix(Y_test_5, pred_LR)\n",
    "labels=['positive','negative']\n",
    "print(confm)\n",
    "df_cm = DataFrame(confm, index=labels, columns=labels)\n",
    "\n",
    "ax = sn.heatmap(df_cm, cmap='Oranges', annot=True,fmt='g')\n",
    "\n",
    "f1_score(Y_test_5, pred_LR, average='macro') \n",
    "print(classification_report(Y_test_5, pred_LR))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# SVM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_train_4=Train4[\"m13\"]\n",
    "\n",
    "X_train_4=Train4.drop(['m13','loan_id'], axis=1)\n",
    "\n",
    "Y_test_4=Test4[\"m13\"]\n",
    "\n",
    "X_test_4=Test4.drop(['m13','loan_id'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "sc_X = StandardScaler()\n",
    "X_train = sc_X.fit_transform(X_train_4)\n",
    "X_test = sc_X.transform(X_test_4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 716,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'alpha': 0.001, 'penalty': 'l1'}"
      ]
     },
     "execution_count": 716,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn import linear_model\n",
    "clf = linear_model.SGDClassifier(loss='hinge')\n",
    "alpha = [0.00001,0.0001,0.001,0.01,0.1,1,10,100,1000,10000]\n",
    "penalty=['l1','l2']\n",
    "param_grid = {'alpha':alpha,'penalty':penalty}\n",
    "knn_cv = GridSearchCV(clf, param_grid, cv=10,scoring='f1')\n",
    "knn_cv.fit(X_train, Y_train_4) \n",
    "knn_cv.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SGDClassifier(alpha=0.001, average=False, class_weight=None,\n",
       "       early_stopping=False, epsilon=0.1, eta0=0.0, fit_intercept=True,\n",
       "       l1_ratio=0.15, learning_rate='optimal', loss='hinge', max_iter=None,\n",
       "       n_iter=None, n_iter_no_change=5, n_jobs=None, penalty='l1',\n",
       "       power_t=0.5, random_state=None, shuffle=True, tol=None,\n",
       "       validation_fraction=0.1, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn import linear_model\n",
    "from sklearn.calibration  import CalibratedClassifierCV\n",
    "alpha_optimal_svm = linear_model.SGDClassifier(alpha=0.001,loss='hinge',penalty='l1')\n",
    "\n",
    "alpha_optimal_svm.fit(X_train, Y_train_4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Macro-average quality numbers\n",
      "Precision: 0.8472, Recall: 0.7556, F1-measure: 0.7481\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.69      1.00      0.82       150\n",
      "           1       1.00      0.51      0.68       135\n",
      "\n",
      "   micro avg       0.77      0.77      0.77       285\n",
      "   macro avg       0.85      0.76      0.75       285\n",
      "weighted avg       0.84      0.77      0.75       285\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import precision_score,recall_score,f1_score\n",
    "# predict the response\n",
    "pred_SVM = alpha_optimal_svm.predict(X_test)\n",
    "\n",
    "precision = precision_score(Y_test_4, pred_SVM, average='macro')\n",
    "recall = recall_score(Y_test_4, pred_SVM, average='macro')\n",
    "f1 = f1_score(Y_test_4, pred_SVM, average='macro')\n",
    " \n",
    "print(\"Macro-average quality numbers\")\n",
    "print(\"Precision: {:.4f}, Recall: {:.4f}, F1-measure: {:.4f}\".format(precision, recall, f1))\n",
    "\n",
    "print (metrics.classification_report(Y_test_4, pred_SVM))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[150   0]\n",
      " [ 66  69]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.69      1.00      0.82       150\n",
      "           1       1.00      0.51      0.68       135\n",
      "\n",
      "   micro avg       0.77      0.77      0.77       285\n",
      "   macro avg       0.85      0.76      0.75       285\n",
      "weighted avg       0.84      0.77      0.75       285\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix,classification_report\n",
    "from pandas import DataFrame\n",
    "from sklearn.metrics import f1_score\n",
    "import seaborn as sn\n",
    "\n",
    "confm = confusion_matrix(Y_test_4, pred_SVM)\n",
    "labels=['positive','negative']\n",
    "print(confm)\n",
    "df_cm = DataFrame(confm, index=labels, columns=labels)\n",
    "\n",
    "ax = sn.heatmap(df_cm, cmap='Oranges', annot=True,fmt='g')\n",
    "\n",
    "f1_score(Y_test_4, pred_SVM, average='macro') \n",
    "print(classification_report(Y_test_4, pred_SVM))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [],
   "source": [
    "features=Train3.drop(['m13','loan_id'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_train_3=Train3[\"m13\"]\n",
    "\n",
    "X_train_3=Train3.drop(['m13','loan_id'], axis=1)\n",
    "\n",
    "Y_test_3=Test3[\"m13\"]\n",
    "\n",
    "X_test_3=Test3.drop(['m13','loan_id'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "sc_X = StandardScaler()\n",
    "X_train = sc_X.fit_transform(X_train_3)\n",
    "X_test = sc_X.transform(X_test_3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 572,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'max_depth': 13,\n",
       " 'min_samples_leaf': 7,\n",
       " 'min_samples_split': 7,\n",
       " 'n_estimators': 10}"
      ]
     },
     "execution_count": 572,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "n_estimators = [5, 10, 50, 100,150,200]\n",
    "max_depth = [7,9, 10 ,13,15,17]\n",
    "param_grid = {'n_estimators':n_estimators,'max_depth':max_depth,'min_samples_leaf':min_samples_leaf,'min_samples_split':min_samples_split}\n",
    "RF = RandomForestClassifier()\n",
    "knn_cv = GridSearchCV(RF, param_grid, cv=10,scoring='f1',n_jobs=4)\n",
    "knn_cv.fit(X_train, Y_train_3) \n",
    "knn_cv.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
       "            max_depth=13, max_features='auto', max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=7, min_samples_split=7,\n",
       "            min_weight_fraction_leaf=0.0, n_estimators=10, n_jobs=None,\n",
       "            oob_score=False, random_state=None, verbose=0,\n",
       "            warm_start=False)"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "model_rf = RandomForestClassifier(n_estimators=10, max_depth=13,min_samples_leaf=7,min_samples_split=7)\n",
    "\n",
    "# fitting the model\n",
    "model_rf.fit(X_train, Y_train_3) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Macro-average quality numbers\n",
      "Precision: 0.8411, Recall: 0.7559, F1-measure: 0.7491\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.70      0.99      0.82       150\n",
      "           1       0.99      0.52      0.68       135\n",
      "\n",
      "   micro avg       0.77      0.77      0.77       285\n",
      "   macro avg       0.84      0.76      0.75       285\n",
      "weighted avg       0.83      0.77      0.75       285\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import precision_score,recall_score,f1_score\n",
    "# predict the response\n",
    "pred_rf = model_rf.predict(X_test)\n",
    "\n",
    "precision = precision_score(Y_test_3, pred_rf, average='macro')\n",
    "recall = recall_score(Y_test_3, pred_rf, average='macro')\n",
    "f1 = f1_score(Y_test_3, pred_rf, average='macro')\n",
    " \n",
    "print(\"Macro-average quality numbers\")\n",
    "print(\"Precision: {:.4f}, Recall: {:.4f}, F1-measure: {:.4f}\".format(precision, recall, f1))\n",
    "\n",
    "print (metrics.classification_report(Y_test_3, pred_rf))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[149   1]\n",
      " [ 65  70]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.70      0.99      0.82       150\n",
      "           1       0.99      0.52      0.68       135\n",
      "\n",
      "   micro avg       0.77      0.77      0.77       285\n",
      "   macro avg       0.84      0.76      0.75       285\n",
      "weighted avg       0.83      0.77      0.75       285\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix,classification_report\n",
    "from pandas import DataFrame\n",
    "from sklearn.metrics import f1_score\n",
    "import seaborn as sn\n",
    "\n",
    "confm = confusion_matrix(Y_test_4, pred_rf)\n",
    "labels=['positive','negative']\n",
    "print(confm)\n",
    "df_cm = DataFrame(confm, index=labels, columns=labels)\n",
    "\n",
    "ax = sn.heatmap(df_cm, cmap='Oranges', annot=True,fmt='g')\n",
    "\n",
    "f1_score(Y_test_3, pred_rf, average='macro') \n",
    "print(classification_report(Y_test_3, pred_rf))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 576,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'m12': 0.3134253829027248, 'm11': 0.12392245104587843, 'borrower_credit_score': 0.1220432320725225, 'm9': 0.1051303636368931, 'm10': 0.07250872597012471, 'm7': 0.05080499737341039, 'm8': 0.02821962678674022, 'debt_to_income_ratio': 0.026701411639580598, 'interest_rate': 0.025064677539976314, 'loan_to_value': 0.021591396239565654, 'co-borrower_credit_score': 0.021585696142382288, 'unpaid_principal_bal': 0.020964177608164002, 'financial_institution': 0.011912911714808525, 'm2': 0.009844204175628524, 'loan_term': 0.007396863187671262, 'm5': 0.0072139058563891025, 'insurance_type': 0.005965883703161838, 'source': 0.004850004949928667, 'number_of_borrowers': 0.004585953128363034, 'first_payment_date': 0.0044309361374554445}\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "imp_features=(dict(zip(features.columns, model_rf.feature_importances_)))\n",
    "dictionary=dict(sorted(imp_features.items(), key=lambda kv: kv[1],reverse=True)[:20])\n",
    "print(dictionary)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 577,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAG9CAYAAAD5kablAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzsnXm8lVX1/9+LGQVnHBAQFBxwNkXLIXNIzQEtTZy+WqRpDpmZWVYWWpoN2mD9tLJBU3NIpcScNScUVEJxRBBBVFDAAQRE1u+Pz348D4dz7j3n3nO5yLPer9d53fsM+9n72Xs/a6+99tp7m7sTBEEQFIMO7Z2AIAiCYNkRQj8IgqBAhNAPgiAoECH0gyAICkQI/SAIggIRQj8IgqBAhNAPgiAoECH0g1ZjZi+b2ftm9l7u17uVz9zdzKY1Ko01xvkXMzt/WcZZDTP7oZld1d7pCFY8QugHjeJAd++R+01vz8SYWaf2jL81fJzTHiz/hNAP2hQz28nMHjazOWb2PzPbPXftS2b2rJm9a2aTzOyr6fzKwG1A73zPoVwTL+8NpB7Ht81sPDDXzDqlcDea2Uwzm2xmp9WY7v5m5imNU81stpmdaGY7mNn49D6/zd1/nJk9ZGa/MbO3zew5M9szd723mY00s1lmNtHMjs9d+6GZ3WBmV5nZO8CJwHeBw9O7/6+p/MrnhZl908xmmNlrZval3PXuZvYLM5uS0vegmXVvroyCFY/QKII2w8zWB24FjgH+A+wJ3Ghmm7r7TGAGcAAwCdgNuM3Mxrj7E2a2H3CVu/fJPa+WaI8A9gfeBBYD/wJuSef7AHeZ2fPufnuNr7EjMCilb2R6j72AzsCTZna9u9+fu/cGYC3g88A/zWyAu88CrgEmAL2BTYE7zWySu9+dwg4FDgP+D+ianjHQ3Y/OpaVqfqXr6wKrAusDewM3mNnN7j4b+DmwOfAp4PWU1sU1lFGwghGaftAobk6a4hwzuzmdOxoY5e6j3H2xu98JjAU+B+Dut7r7Sy7uB+4Adm1lOn7t7lPd/X1gB6CXu49w94XuPgn4AzCsjued5+7z3f0OYC5wjbvPcPdXgQeAbXP3zgAucfcP3P0fwPPA/mbWF9gF+HZ61jjgj0jQZjzi7jenfHq/UkJqyK8PgBEp/lHAe8AmZtYB+DLwdXd/1d0/dPeH3X0BzZRRsOIRmn7QKA5297vKzm0AHGZmB+bOdQbuBUja/LnAxkgBWQl4qpXpmFoWf28zm5M71xEJ61p5I/f/+xWOe+SOX/UlVzCcgjT73sAsd3+37Nr2VdJdkRry6y13X5Q7npfStxbQDXipwmObLKNgxSOEftCWTAWudPfjyy+YWVfgRmTOuMXdP0g9hMyGU2n517lI0GWsW+GefLipwGR3H9SSxLeA9c3McoK/HzIJTQfWMLOeOcHfD3g1F7b8fZc4riG/muJNYD6wEfC/smtVyyhYMQnzTtCWXAUcaGb7mFlHM+uWBhz7AF2Q7XomsChpsZ/NhX0DWNPMVs2dGwd8zszWMLN1gdObif8x4J00uNs9pWELM9uhYW+4JGsDp5lZZzM7DNgMmU6mAg8DF6Q82AoYDvy9iWe9AfRPphloPr+q4u6LgSuAX6YB5Y5m9snUkDRVRsEKSAj9oM1Iwm4o8kSZibTKbwEdksZ7GnAdMBs4EmnFWdjn0ODnpDRO0Bu4EmmqLyN79j+aif9D4EBgG2Ay0nj/iAY724JH0aDvm8CPgUPd/a107QigP9L6bwLOTfbzalyf/r5lZk80l181cCYyBY0BZgE/ReVQtYzqeHbwMcJiE5UgaD1mdhzwFXffpb3TEgRNEa15EARBgQihHwRBUCDCvBMEQVAgQtMPgiAoECH0gyAICsRyNzlrrbXW8v79+7d3MoIgCD5WPP7442+6e6/m7lvuhH7//v0ZO3ZseycjCILgY4WZTanlvjDvBEEQFIgQ+kEQBAUihH4QBEGBCKEfBEFQIELoB0EQFIgQ+kEQBAUihH4QBEGBCKEfBEFQIJa7yVmtpf/Zt9Z1/8sX7t9GKQmCIFj+CE0/CIKgQITQD4IgKBAh9IMgCApECP0gCIICUZPQN7N9zex5M5toZmdXuH6imT1lZuPM7EEzG5y79p0U7nkz26eRiQ+CIAjqo1mhb2YdgUuB/YDBwBF5oZ642t23dPdtgIuAX6awg4FhwObAvsDv0vOCIAiCdqAWTX8IMNHdJ7n7QuBaYGj+Bnd/J3e4MpBtvDsUuNbdF7j7ZGBiel4QBEHQDtTip78+MDV3PA3YsfwmMzsZOAPoAuyRCzu6LOz6LUppEARB0Gpq0fStwjlf6oT7pe6+EfBt4Hv1hDWzE8xsrJmNnTlzZg1JCoIgCFpCLUJ/GtA3d9wHmN7E/dcCB9cT1t0vd/ft3X37Xr2a3eIxCIIgaCG1CP0xwCAzG2BmXdDA7Mj8DWY2KHe4P/Bi+n8kMMzMuprZAGAQ8Fjrkx0EQRC0hGZt+u6+yMxOAW4HOgJXuPsEMxsBjHX3kcApZrYX8AEwGzg2hZ1gZtcBzwCLgJPd/cM2epcgCIKgGWpacM3dRwGjys79IPf/15sI+2Pgxy1NYBAEQdA4YkZuEARBgQihHwRBUCBC6AdBEBSIEPpBEAQFIoR+EARBgQihHwRBUCBC6AdBEBSIEPpBEAQFIoR+EARBgQihHwRBUCBC6AdBEBSIEPpBEAQFIoR+EARBgQihHwRBUCBC6AdBEBSIEPpBEAQFIoR+EARBgQihHwRBUCBC6AdBEBSIEPpBEAQFIoR+EARBgQihHwRBUCBC6AdBEBSIEPpBEAQFoiahb2b7mtnzZjbRzM6ucP0MM3vGzMab2d1mtkHu2odmNi79RjYy8UEQBEF9dGruBjPrCFwK7A1MA8aY2Uh3fyZ325PA9u4+z8xOAi4CDk/X3nf3bRqc7iAIgqAF1KLpDwEmuvskd18IXAsMzd/g7ve6+7x0OBro09hkBkEQBI2gFqG/PjA1dzwtnavGcOC23HE3MxtrZqPN7OAWpDEIgiBoEM2adwCrcM4r3mh2NLA98Onc6X7uPt3MNgTuMbOn3P2lsnAnACcA9OvXr6aEB0EQBPVTi6Y/DeibO+4DTC+/ycz2As4BDnL3Bdl5d5+e/k4C7gO2LQ/r7pe7+/buvn2vXr3qeoEgCIKgdmoR+mOAQWY2wMy6AMOAJbxwzGxb4DIk8Gfkzq9uZl3T/2sBOwP5AeAgCIJgGdKsecfdF5nZKcDtQEfgCnefYGYjgLHuPhL4GdADuN7MAF5x94OAzYDLzGwxamAuLPP6CYIgCJYhtdj0cfdRwKiycz/I/b9XlXAPA1u2JoFBEARB44gZuUEQBAUihH4QBEGBCKEfBEFQIELoB0EQFIgQ+kEQBAUihH4QBEGBCKEfBEFQIELoB0EQFIgQ+kEQBAUihH4QBEGBCKEfBEFQIELoB0EQFIgQ+kEQBAUihH4QBEGBCKEfBEFQIELoB0EQFIgQ+kEQBAUihH4QBEGBCKEfBEFQIELoB0EQFIgQ+kEQBAUihH4QBEGBCKEfBEFQIELoB0EQFIiahL6Z7Wtmz5vZRDM7u8L1M8zsGTMbb2Z3m9kGuWvHmtmL6XdsIxMfBEEQ1EezQt/MOgKXAvsBg4EjzGxw2W1PAtu7+1bADcBFKewawLnAjsAQ4FwzW71xyQ+CIAjqoRZNfwgw0d0nuftC4FpgaP4Gd7/X3eelw9FAn/T/PsCd7j7L3WcDdwL7NibpQRAEQb3UIvTXB6bmjqelc9UYDtzWwrBBEARBG9KphnuswjmveKPZ0cD2wKfrCWtmJwAnAPTr16+GJAVBEAQtoRZNfxrQN3fcB5hefpOZ7QWcAxzk7gvqCevul7v79u6+fa9evWpNexAEQVAntQj9McAgMxtgZl2AYcDI/A1mti1wGRL4M3KXbgc+a2arpwHcz6ZzQRAEQTvQrHnH3ReZ2SlIWHcErnD3CWY2Ahjr7iOBnwE9gOvNDOAVdz/I3WeZ2Xmo4QAY4e6z2uRNgiAIgmapxaaPu48CRpWd+0Hu/72aCHsFcEVLExgEQRA0jpiRGwRBUCBC6AdBEBSIEPpBEAQFIoR+EARBgQihHwRBUCBC6AdBEBSIEPpBEAQFIoR+EARBgQihHwRBUCBC6AdBEBSIEPpBEAQFIoR+EARBgQihHwRBUCBC6AdBEBSIEPpBEAQFIoR+EARBgQihHwRBUCBC6AdBEBSIEPpBEAQFIoR+EARBgQihHwRBUCBC6AdBEBSIEPpBEAQFIoR+EARBgQihHwRBUCBqEvpmtq+ZPW9mE83s7ArXdzOzJ8xskZkdWnbtQzMbl34jG5XwIAiCoH46NXeDmXUELgX2BqYBY8xspLs/k7vtFeA44MwKj3jf3bdpQFqDIAiCVtKs0AeGABPdfRKAmV0LDAU+Evru/nK6trgN0hgEQRA0iFrMO+sDU3PH09K5WulmZmPNbLSZHVzpBjM7Id0zdubMmXU8OgiCIKiHWoS+VTjndcTRz923B44ELjGzjZZ6mPvl7r69u2/fq1evOh4dBEEQ1EMtQn8a0Dd33AeYXmsE7j49/Z0E3AdsW0f6giAIggZSi9AfAwwyswFm1gUYBtTkhWNmq5tZ1/T/WsDO5MYCgiAIgmVLs0Lf3RcBpwC3A88C17n7BDMbYWYHAZjZDmY2DTgMuMzMJqTgmwFjzex/wL3AhWVeP0EQBMEypBbvHdx9FDCq7NwPcv+PQWaf8nAPA1u2Mo1BEARBg4gZuUEQBAUihH4QBEGBCKEfBEFQIELoB0EQFIgQ+kEQBAUihH4QBEGBCKEfBEFQIELoB0EQFIgQ+kEQBAUihH4QBEGBCKEfBEFQIGpae6co9D/71rruf/nC/dsoJUEQBG1DCP0GUW+DAdFoBEGw7AnzThAEQYEIoR8EQVAgQugHQRAUiBD6QRAEBSKEfhAEQYEIoR8EQVAgQugHQRAUiBD6QRAEBSKEfhAEQYEIoR8EQVAgQugHQRAUiJrW3jGzfYFfAR2BP7r7hWXXdwMuAbYChrn7DblrxwLfS4fnu/tfG5HwFY1Y7C0IgmVBs5q+mXUELgX2AwYDR5jZ4LLbXgGOA64uC7sGcC6wIzAEONfMVm99soMgCIKWUIt5Zwgw0d0nuftC4FpgaP4Gd3/Z3ccDi8vC7gPc6e6z3H02cCewbwPSHQRBELSAWoT++sDU3PG0dK4WagprZieY2VgzGztz5swaHx0EQRDUSy1C3yqc8xqfX1NYd7/c3bd39+179epV46ODIAiCeqlF6E8D+uaO+wDTa3x+a8IGQRAEDaYWoT8GGGRmA8ysCzAMGFnj828HPmtmq6cB3M+mc0EQBEE70KzQd/dFwClIWD8LXOfuE8xshJkdBGBmO5jZNOAw4DIzm5DCzgLOQw3HGGBEOhcEQRC0AzX56bv7KGBU2bkf5P4fg0w3lcJeAVzRijQGQRAEDSJm5AZBEBSIEPpBEAQFIoR+EARBgQihHwRBUCBC6AdBEBSIEPpBEAQFIoR+EARBgQihHwRBUCBC6AdBEBSIEPpBEAQFIoR+EARBgQihHwRBUCBqWnAtWL6JTdWDIKiV0PSDIAgKRAj9IAiCAhFCPwiCoECE0A+CICgQIfSDIAgKRAj9IAiCAhFCPwiCoECE0A+CICgQIfSDIAgKRAj9IAiCAhFCPwiCoEDUJPTNbF8ze97MJprZ2RWudzWzf6Trj5pZ/3S+v5m9b2bj0u//NTb5QRAEQT00u+CamXUELgX2BqYBY8xspLs/k7ttODDb3Qea2TDgp8Dh6dpL7r5Ng9MdBEEQtIBaNP0hwER3n+TuC4FrgaFl9wwF/pr+vwHY08yscckMgiAIGkEtSyuvD0zNHU8Ddqx2j7svMrO3gTXTtQFm9iTwDvA9d3+gPAIzOwE4AaBfv351vUDQemJp5iAoDrVo+pU0dq/xnteAfu6+LXAGcLWZrbLUje6Xu/v27r59r169akhSEARB0BJqEfrTgL654z7A9Gr3mFknYFVglrsvcPe3ANz9ceAlYOPWJjoIgiBoGbUI/THAIDMbYGZdgGHAyLJ7RgLHpv8PBe5xdzezXmkgGDPbEBgETGpM0oMgCIJ6adamn2z0pwC3Ax2BK9x9gpmNAMa6+0jgT8CVZjYRmIUaBoDdgBFmtgj4EDjR3We1xYsEQRAEzVPTHrnuPgoYVXbuB7n/5wOHVQh3I3BjK9MYBEEQNIiYkRsEQVAgQugHQRAUiBD6QRAEBSKEfhAEQYEIoR8EQVAgQugHQRAUiBD6QRAEBSKEfhAEQYGoaXJWEFQjVugMgo8XIfSDdqPeBgOi0QiC1hLmnSAIggIRmn7wsSVMS0FQP6HpB0EQFIjQ9INC0ppeQmvHIqKHErQnoekHQRAUiBD6QRAEBSKEfhAEQYEIoR8EQVAgQugHQRAUiPDeCYKPEeH5E7SW0PSDIAgKRGj6QVAgoqcQhNAPgqAm2nNCW9A4QugHQbDcEz2UxlGT0DezfYFfAR2BP7r7hWXXuwJ/Az4BvAUc7u4vp2vfAYYDHwKnufvtDUt9EARBM8SSG0vS7ECumXUELgX2AwYDR5jZ4LLbhgOz3X0gcDHw0xR2MDAM2BzYF/hdel4QBEHQDtTivTMEmOjuk9x9IXAtMLTsnqHAX9P/NwB7mpml89e6+wJ3nwxMTM8LgiAI2gFz96ZvMDsU2Nfdv5KOjwF2dPdTcvc8ne6Zlo5fAnYEfgiMdver0vk/Abe5+w1lcZwAnJAONwGeb/2rLcVawJvtELY94450FyfuSHex4q7EBu7eq7mbarHpW4Vz5S1FtXtqCYu7Xw5cXkNaWoyZjXX37Zd12PaMO9JdnLgj3cWKuzXUYt6ZBvTNHfcBple7x8w6AasCs2oMGwRBECwjahH6Y4BBZjbAzLqggdmRZfeMBI5N/x8K3OOyG40EhplZVzMbAAwCHmtM0oMgCIJ6ada84+6LzOwU4HbksnmFu08wsxHAWHcfCfwJuNLMJiINf1gKO8HMrgOeARYBJ7v7h230Ls3RGvNRa01P7RV3pLs4cUe6ixV3i2l2IDcIgiBYcYgF14IgCApECP0gCIICEUI/CIKgQITQDwIgzSBfbrBEe6cjqJ2PS3mF0F8Bac/KZ2YfyzrlZR4NyT253fDEso63uboTjZGoVM/bo7xaQuG8d8ysYz1uo1nhuvviesO2lPRRGenbb8TzllWFTPllaJG9tdHqqrPdfVwDnm3Qso8ry1N3X1x2vhNaSHA3tMbUf8zsCGBj4G/uPrk1+Zfi7VCt3pQ/28x6Ar2AdYH+QGd3/+uyLMO2Jt9otOU75fPMzNYEerj7FDNbDegHTHH3t6uFqTGO9YH33X2WmR0GzAHuc/cPGvcmjWWFWU+/0kdtZh3KjjcG9gF+U+tzs/Bmtg2wc5p3sB7woru/36i05yta+n+pildBQHTIpa8bsH76rQLMRGsYzUsL5S31rFxcLU13ef52BY4ETgM2BTqj3uQbad2li939rZYIsEph0jusBrxdLszLKc9TM1sJ6A0cBRye0nsb8B+Ub/8HfAD8JL1Dk419rqEmn5YUb9WwOaF0ILA9msDYBzWYqwMLgb+2hXCskqfbAnPd/YUyodmDUmPUGzXqj7j7nfUqQ829S1IcVgNmp1MDkOIwu3qo6u9mZscChwDzzew4oCdaF+xJ4Lz07WwJdHH3h2p8/snAl4FtgW8DP0tpHg68Azxa/n0sL6wwQj/7qJPm1hVpV++WZfwAtC9Ak0I/9wGvBxyDZhkvAD6JKsqpwC3Ata0t2LLKuX6KYytgO/SBjQX+BTzk7nPy96feRw8kaD+d3q8n0COFBRhlZt9294n5eKt9eNU04krkGpzuqQEcCvwWTeLrArwG3IiE5jHp/HeoQYhWiMtT/uyFBM4v0YJVJwEzgD9WC2tazvsA4CA0SfBrqGyvRoL21HR8WDLrTAD+l8L8pMozl8inCo3KqqgHsQ2wC/Ceu3+1XECaWS/g/wEHIiE3BXgRuDWloSGLD5bVs66ulW89fy2909eB+cCJ6Vxf1ChuiXoe6wArp79XAHdSQUHJnpvLm/z5tYEjUF3vDDya3ncacDDweWAuEqp9UzyjgAtqVRhS2lcDbkaLPz6NBPS5qCH9L3Cimf0KfV9DgDlmdiFweVMKnZn9IOXTn1Bd3ixduiO9137pnZZPM5i7rxA/YA/gIeA94G0klA8ru2cgqtAb5M51QMKoI8nclc6vB1yDhPxYpDkvRhrhJcDdWfgGpH1T4J8pbYvTbyHwO+DfSLj9G+hbFu4LwAtIIx2DZvmdCRwG7IQ+nqeAR4C1ysJuApwF7IwaiW7Ayk2k0cqOuwPfBB5HH+RwYDxqHC9DQv5PaK2l76W4Xm1pnqEP6QnglZQ/e6Xz303xrlclXFfU0DyTynJ2KtuNUplOSfdtlPK8Tzo+HJhVQ7oy89CRSNu7A3gJdfPfRkL8CeDPVcKfmu49DujUxt/IhqgR+xPwY9TQdCi7ZwTwVPr/ADTDfm7Ku2tSfn8OGNDCNKyHll9/E3gQGA28D0xGwvc1JJAXpzLpgBr2Z1sQ1+npeVum43nAecAFwK7pvX4MvJry/yfoe8rq1lL1FDV+TwM/Scdnom+sY0rrb4B7s7rRluXZ4nrQ3gloUGVeDQnmx9NHdHqqUPPQ+j+bp/tWQULo0CaelXX1bkRC+BG0D4Cnirhv+hjebVDat06C4SWkvX4jVcSHgCvTPX3Su92QC7ch8BxwHTCoiedvhATlIWXn90zv8w5qKJ9LH/VPkYDeC5kXVgV6Vnju8JS/16A9FrLG53Zkh87uOwt4GTVQs/PX6sij3kjj/Wd678nAt9K1vql8DknH5ULs86hXcQawRfp/p1TOC5DAWR31TBYBn07hjk8f81Lvnq73SGU2I+XhG+n+UWjToW8hrW93pAmuUxa+Y/p7LTIrdc/OI+23S/o1RHAgxeJlJMQfS3k4J5XXkNx9h6X3WQm4G7gfWKOFcfZDDWHPdGzA2aneHZHqtaXyfSmVzRfSvXOBz+a+kXeB/nXG/whS0LohgTwpnbsJ9fCmAPcA/0r3d0Y9l5+n4065Z2Xjn3ugBmuzdPzplF9rp3f5DvBYPszy9ltRzDuHoEp6mrvflc5dYmb7A98HHjKzi4A/AM+ibuUN8JGt/nPow9wUdVu7AmsiTfAdpEn2RxVvZ9QgLDKzDdx9SksTnUxRxyONcFfPDSqZ2U7A/WZ2vrs/b2ZnAbeY2WruPgd1fd8DfujuL2YDhpQ8sjq4+4KU5plI+Od5Ov39NhIuQ5DG+gkk0Hqhhm4y8KyZTQe+6e7vm9nKSLO9HhjuWp/pE6jhfdqXHMS6B2n6ZyCtbqX0vrXkT2Y6y7bb/J67TzKzx5EgAH2A04Ad0Mdc3qU+Ffi7u/8yPXMGsIW7jzazV9HHupO735aO10vhvgzchxqGpXD398zsdpS3k5G2+DoSqrPdfV4t74gair2BNVBP6EPKTF/12szLSTbyo1DebOnur6bzh6F9LO4xs/8HnA+MQ+abrkgwnuLus9L99Y7FOOoN/hn1tLqghRn/6O7X5J453czeTfffncK+hsycdyBlYTZqtF+uw6T6LuqRdkEKyTikNKyByvmddD57VmckHzbNpV//lN77GdSIrJ3ufRzlVxd3dzP7Aqo31JlXy4wVRegPRBr8g9mJ9KHcamb3AacgzesLqOJvlgs7DGmjY1Ah3o00gHPRvr+/Q0JlD6QVHpDimoIEWItJwvJA4JeZwE/2Z5JQeh8J4ueRlvI+qvgPIg38dXd/Jt2fDRiWC4dDUSUfXRb3G2a2EHjG3e/PzpvZuujD6ILGCHZGDcbOKS+edve5ZrY5+ngXpbx+PD3vWDN7D/Ww3ge+hLTXgcCXvMxbohkyAd4Z7b38Sjp+HJlfsmtOqbHL7NSZoOwGvGpm3dx9PsrLbdO9I4GTgdNTXi8CzjWzC1Kav+buC5sQdt/0KrbfJGj7ozzcBPgMcLW735TKOBM0f0TC6Rdm9gvUaAxCjdh26e/7aKXbFnnwuMZ+9gcuywn8ju5+vZndihrw89GYxy9QD24warA/aWa3JsHfycw6I6VgjfRuL7j7S5UEsbtPNbO5wAAze87dF5jZ6kjbzuiI8n02Mq9tl+J9GjU6oAZ+VkoXVBlDqMDNSNk4jFLDsy3qgVyCeqfdUU8RVJfWS+mAUhnl3+l1M7sHuNTMTnL3B8xsAXCmmfVGvZaq40vLBe3d1WjED2kxz5LssbnzHXL/D0amkMXAhNz5g1CLf7bnunRIwD0K7J+Or0Ld4r8ijfa33oAuHOpFnEeZ2QMJy/m5+HdAWs++6fjzqDH6ai7MGkhg7I7MRPel9/1FlbhfyN47HXdsIp19yJk6Uh6MQI1D5qL5PBLObyENOdOi5qT3qctUkZUf8FnU0GZd/11SmfVDPaWZwFZlYbNyHIEaib3T8Y+AB1BP7u9IEDyX0vwSEki/Itmsm8qTKmneGX30j6XnzUt58DJwUnmdQYJmJKVxnMUp715Bwu9S4OsN+EaeQMpNPu6Ouf/7pjo+O6Xh9HTuW8iEd36qpzcigfxWuu+C8m+tQh07JXf8GHAR0K2snC5O9eUv6fj0VMc2QSaoW5E2Xc87r5LKfzHwcCrbxagRnZuOFwA75ML8lDJTaO5aZuLpj8bYFqb8eieV731ZPVuef+2egIa8BGyQPt4f5gunyr2HA2fljtdJH8TjWeVFQmwTNBj8ZqpwU5BGNicV7qYNSvsZSHj/Hmk2e6QP7E0klDJ7aH/kbrpGOu6B3M7mp7Tdm9L5EBK+L6KBsaHZe1WI+xrgqgrnO7Bkg7lUfqaPclJK91A0+PZH1EN6M31cH6Zy+VQr86gHGhx+B2loZ6Tn/xsJlZ+UC4TcBzoQmX1moob8jvTBP4s0vAPT+26OtPHBpAHtpupRlXTunvJ+PDK9XZzSdk0qj12qhLsv5eNZwBeRqa0/MrH1qFR2LcjDX6PGZWBZ/hhLNgT7I7fVI9LxF5FwWwBMBe4Cfo5s8tvRxOB/Cn91KqPfoEZleqorn0zXO6e/ByDzy7wU/zhKTg0TgO1a+N7d0ByMzNPo10jn0sBqAAAgAElEQVRRuDjVh73K3n9dmhh3oqSIrIYa+K+ihvFwkoMIMo31bkkdWha/FWZylpl9B33Ah7j7G3WGPRRNxrnI3Rflzq8MnIjs3N2QMBsNXOey6bZ6wkyK4xhUcQagj+sNZC//efm75Nwjs+N9kEvaRkizmYU0scy+XtW32cy+n95vYPbMSu+UzBHdgK5esu/2RBX+LLTf5wTgK+7+qJltj4TWPKRVk44NWM3d/1tHFmVp6Amcgz7SDZBr6jNIoP7Sm7B5mybmfAc1qt2QEBuHGvWtgdvdfUKVsDWVscnX+3fI1vt9lP993P21NN7yEzRus0uV9C1G/vELk2mom9c+LtAsZrYrapQfdPfhKQ6v9G6m+RZrILPLjag38jVkCqw4xtFEvMehsbRxSGF6DdXPW9z9+dx93VCZHoicJTqghvJxT3tq12HLby5NA5GJqBNqVCe4TH8NwczOQD3wTcysU16mLA+sSEJ/dTQAM86b9rHthD7+We7+QjPPXNvdZyQ75spI098GDUROakQFzMW1boqjB/BSalRWQT2OfmgcYmukTQxJYfK+11U/iCS0t0GTmCbmzm+EehZ/8bIZhMk+uX4u7i1T3Dvl7umANKO33X1uOtcLCfitUXlskf6ugTT13qi3MqeF+dQXmSVeLjt/Omrk7irPiyyfUh53R2MhWSM3B2m1tyXh3AlY3FQj0kTapqJe5DVpjOBI4OYU90Zo0HZzd5/cxDN2Q77qm6L68Cpwqdc4aaiJ53ZFbooA9+fLu6ysB6MB1PWRWW0W8EV3vznd2zH32MXNNYjpu9wCTWZ8vexaeR3bHNjQ3Xcsuy8bj6mblN5BKB/nojGEvyFN/Ro0kfBcd7+1jmf2R552J1VSMNM43V/cfc2WpLnNae+uxrL4kevCIqF6E/qQmgvzV9TdLQ/7O5J/fxul95OoWzwNaaWLkGfIfcjHuns+nWVp7oQ0s06UuqI1vXO6dzf0MUxpLu5cmM+hXsoeyLa5CGlS2W8e8nA5G3mq1J1vyL48pOxcJ0pjCtdQsgc3O3aAegv7Iq1zOnB0Op/lWQdkGjqGknteNbt1FuYF5EEGcnW8gpLff+aPn5k1yuc9rIk8yhYjAXUn8I9q6WtQPatW1lnau6Pe1NdRL6/SM9Yh55JZQ5wdUj35RxPxDkBePn9B5rgbkRfQ5i14x6GoR7cDaS5OKtMHUvnfgkxrHdB30hspSDtUe6dU5xYDW6fjjrmwW6NB4+doxvTVXr8VxXsH+KhVX0r7KD9Gmvony8J+NHsw8/xIZooZyPc6H3Ynb6Pp1ckr5lokIC5FXkXPAW94lfU8Mk02vWe1ruRS71zhOYORFjQHzRJtNu7ExcjWuz8SFKPRx3oH6p7ntcqOyC/+zVpMJzmtfX/gGDPbD2lsi31JU9xbSFOEMu8O0/IbhyAtdlvknZT3vFoEfN/MdgHWTmXQDzUqHZE2+Gz5c/NRpL8PIG+X36OZyT9F24hORjbwm5GZY4k6merel9M7Hufuf8td2xCZpr6Pxl8aUu+aK+tc7+hANBg6wMzuQkJ+ByQY+yNz1tfSM4wqy4fk3nlTJNirxdsbjRl8CikQr6He5PnAcDM73d3vsNrdWLOe57T0TRsa+9oC5WknVG/ORr3qjdPP0YTG8rq0Mhpwfhs42Mw2QT2JzVL4/qhXeyoab1v+aO9Wp61+qAJ2R4U+CA26fBH5jD+BKlm1sB1T2C8h2/pvka/4uUjzmoO8QB5BXUNo5SQaStrir5HG+IkK79MBVdKKGlVL37k1cafz/dBAXbWwnZEtvTPyALm+1jyjNIlpO+Cdsmu90UzdX6LG+ZdVwn4LaWYPot7bt9HA2xHIQ8ZT+JfQIOPlyPvpINRQrl1j2X0GNdR7pONDUUPwXyQ4e1cJ3wV5GF2cO5fvqQ1BGnGLJkm1pqyRVpx5Fn2AzD1PIk39++i72LhB8XZL5fNuincgWiQNZAa6H3ggX7Y1xLsL8jQaXPZ935HeK+uNzkeNz7WoV7F7hWcNQp5HdyFT5YL0jLdTnvwZzR4eXEva2uu3omn6GyP/9czXfNP02whpKJ0pzZ7sZWaresk/vlLYLZBW07VCdEdSmoyDt36wJtMW30buek+ldHVy90WuWueU+Q635p1bG3cW1t1fSbbxV1K4X6NGcVYK+0H6YZrkdVQKW8v6Ph+mQb7VgW4mv/L+yATQLT13GnLH/Wl52PTvH5Db3wzkVfM+qUdoZg+jnsqmyA32ddS41KxR5+59EJmx5qbzN5AmAVYjadQLTesKPZV7Zr4+bYzKb11SfWsFNZV1TtOfjDynxqJ8fh05NLyLFvOrtd43G2+ylR+MxkWuzAd291dNE8h+Xef7PomUtDPN7Buo7D9AQv5h1KM4ytNksWZYG41tvYV6IB2Bz7j71DrT1K6sMAO5AGY2CmmEayPNZCrSzMdQ8maZbmYDkLve3u5+TxNhpyJt+VH0MT+NKu3v0SqMd6JBzLo8GqqkvYNrEs12yCPmWeASd387DSQPQl3qnYE13X1Ya9+5tXGXhf0E0qhnoMlw61GaaDQETVfvgz7Age6+LTVgWs3wEvSBgQTGf5G2OAl9fG8AU9393VqeWSGO/ZHmumr6DUCDqK+j5TZea+FzuyNz0gDkvnqnu79Vdk9mShyFNP7jXcs5d0d24k8iDfJuNLltbkvSkouvrrJGa9IMAGa6+5tVnlmzma6ZeHdEHkb3uPve6fzilD8dUKM+zd1/lZ65MrCguYbHNPP4z2gMYQzyEtoGjeucg7T3DqgnNTOZvtZDyylUrFNm1gf1AJ9Ix92Qcrhyyrf1gFVdE+CWr9U227ur0cgfspdOAnZr5r4+SPMb1lRYpJ0MRuaLntXCtsF7rI/s4uNQY5NN2nkbuUbe1Ih3bm3cFcL2Qw3NYqQNLaJkGniPUo/jqDryYivUwO6ABvROQp4XrV6TBvUetkaNyBNo8azHUlo/TGl/Cg1U1zz4jD7+nyBz0XxKtunXkItst9y9mdljV0pujX9DJsW7Uj7eSgsXOGttWaOe4lOUJrdlawN1qidP6oh3TqrPW+TzCJla56H1l25Bis1ikimthjh3RqabJ9J3cAFyxR6AlIrJwDXp3h2QEnB89s5lz+pUdrwtmjtyfXqnV9I7vdXoMmvEb4Uy76BCPRF9zJiWyc26+E7JL3k68kiYXUPY57OwSZvdFXlX7Gpa1vhqb6A/dWJVVKE7I9vj05S02qxr3Vy6ofl3bm3cH2FmW6bn90+neqBu9CKk8U9EH8QdSMjWhLuPR5OdMLNT0IYVc9LxSqhRHowG0vZDWuIZ5QN9puV8ByCTV7/cbzOknY1DdvdRyNzzLBobOQfN0D0JWModtEI+rITGBA5B2mUmuDshj5Tz0USxG/NpdE3nH4aWBTkUKRwvoAHG/3jbmBCaK+v3XIOrPYAdzOyhVNfL1wZaG2ncb9ei9TcTb2c0KfGGZApcF5VbV6Sp90N1+AEk/Juck5NzcngITVzEzDqn99oCDa4vRvVzx1R+z6Fv/GhkGix3DMk7EPwFjT8YaigfQz2ILkhZWf5o71ankT/0Af8L+JGXNIRKLlfZQFW3WsIi+/gNSIAtpLQq5rtI6G6UhWll+rP4rkzP34O0QmBbvHNr486HRe5vM5HNNFs2dwcquHi2Jm/S/wcic8dbSCvPZm6+QlrqoMJ7fQMJlZeRoJiCbLq/QY3VNHIDqbnwa6FVMK9IxxV7GLl4DkrPH17hnlWQlvmHas+i6aUwKrpNtlU9o9QL2RYJ5wuQCeaLJG8iNBP8TeDkGtJfNV5Kg+4/RMrCk0iAZktADEc9rk+gnsJKdbzvZ1Hj2z9XXzunsn8ulXF/1JhlM5Y/hybLNfXcA5A5dRekJJyazndGJs6HKK3a2jBX29b+VjRN/wWU2e/CkjsY2dITQbZCBT2kqbDJxesctBTx5UiwbYa0j/PRcgRfR259HahhcLIJssWnugKPemm8oanNGFrzzq2NOx/2TiQAFqGPqTsw1t092WZ7Ig1vrZSmua5dl2qyd3r2tcqt8reom74q0oRfQzbUL7P0gnPZhi0TUMP93/T/i8CHLjvzUWjwNVuJM5uglZmpJqNBfajutpnF0xuZdG5Lz8r2aViEzBNTKK3cuZTLoct+3Snl01rILJeV42ZmdppXsa3XQa1lnb3rWmgA81PI8+l1lC+zkUb8YnovvGk3yqbizeK6EgnQbNXSmWhgvTWDj6+jca81gZfTsz4wrYh5PjIpZd5Jg5DWPxOYZ2Z9vayXlauzn0GTQR80s6eAnczsctfCclei3t7+yOuotbKhYaxQQj9VuCWWOjbNcDwJVdj1UOa/iVr4pywta1AeNlXEXZGP9F5IOzghF3Y80vJ7pfOn0fpCzSr2VcAXzexod78qCc7uSGBvgirw++5+QWveubVx58MmYXUVcoXcyN2vSmnZHZldhqAeU1fkIz8bNUb1fsxHIxvzl5BGPdrdH0xxzQS+a2aj3H1aSlhmPsnmDeTzqWsyXcxDQvUA4Pe+5MDghqjBvyn/vApk7/E/JNw2AaaX3f8FJDy/nJ7lKX5cM7D3QwOM66D8Xh8J3JVQQ9IN5VlrhX7NZW2aWXwnEuxPod7UFNRjmo7cgGt1ZGgy3uS9swkaZ9mwrI4BMs1Q54xpdx9v2ub0e6blWhYj80sHNFDeG/VmXkWeUrehMp+CtPalkpH+vo3KBzR58VRk2nmBkuvrGun6ciHwYQXz3iknjcKPQi35P6h9slE+7DwkrL6GuuZLhDVtnXaUu2/SwHR3Q13pw9GYQkdkGuhByXXyeXc/qIl01/3ODYg7H/ZFZGvPPopFSMPNlq3d3WWvr4mc98dTaH38C83seSRAfpyurYmE7tfcfWQujCHBuQYa+1gbCdSNkJDZBGnyHVPan0v/b416UasBB7r7szWmdShSFq5DvYr1kffSfrn8WMO1/eV5wAx3/42ZXQZ8JeXd9JSO/yFh+6zXsUdsjemsqayRxtrTqyydkezgPdx9RoPiXQ9p2leg+ntfumdRHQ1MeZwHogHbAajh6oLGC4zSbnvPo/p5LXAhMsP9oLw3mqtXeyMX0v1Rr+0G1EjciEw+uwLHuPuNLUlzW7FCafoZuUI6ERXicHd/PHfdkgtYB9TF9ybCHoPsx2u6+7QUtiMyLeyI3M8uKR84bCUd0If2JqqYk1DP4kk0Lf4Vyvy1W/POrY27StiNkEnneqRBn4W0ZUcLtV1kZqe4+8QaB/+y7vHLlAaL/4OWdXjIzB5AA6UrU9LO3MwORkJ9UPq7IRL6HdHH/hIyFd2W0v6blMaeyET1CPJ4alLgm1z4tkHrOm2DejZDUprfQ1rjXWjQchYlze9dYKBpZuf5wPnl5oRcHJ9Eq01e2qD6VlNZp3iywfNN0EDzp4Cfuft9SPB1MbMzaxT81eIdhxrJkynNmzgNlcV4VM6TUA/v4ex7rNH0839I4D9ByangJTQGcyQqnw3Rd52N4V0MS5pMy47HpGed6O5nmdk3gR+kd3sPLYm9XAl8YMUayM1+lAaFzkMfWpd0XM8M0I/CosGld5Dr3GmoG3cFGvwbSdk6/g16h92RoKppad3WvHNr4y4L+5kU9nrk2QT6OC4gDUQi88UEqrjENfN+pyC7/EAkxB9I5fAQMoH8g9wyy2gwMNNY/4nGAPagbM/gdO/KyARzPBrIW6eO987Wan8x5f9FyMT2OUobd6xGbjCW0kDpuUjIdEOD352RYMy2TMyWHz4aTYj6KOyyrGdosH4SahAWAyPS+Z1SGR9Wa9pqiTeV705o/ZyfI7NJNmj/LhprOYe0PzJVZqqna6tVOb8qmhD2C+Q19HM0UL1KjfnXFzVcWVmuicbOKu7ZvDz8VkhNn5Lt8CZUsb5jZjVNNqoUFk3Gmo6m7V+AuoRz0USk0zzZjxv6AtKg6gqS/rbknVsbd57/uuz7c5GmDBLSn0KmlWylxcXI7FErmXZ1KzKRDHaZcI5DjXAPtCDcLe6+MBduPy+bEJWRzD5dkbDNdoN62rU9ZX4F0/WR8H7R3e+rMvh8EjJRTEMKwjw09lGLFvoaMuusjfYI2M217PDCsvveBN6zBi7XW2tZm9kgNND9jLsfYGZ/orRl5VOoXPdAjX2zg5ZNxZvKpaMvufTyI8hE+Bk0MHoPMp+cjBrKgykNpleKL+updEZl3h2V11qo1zcLNchn5uKsmM8pfZmX3BK9slTX3srduykw391frtVpoc1p71anrX+0brJRTWFp4AbW7f3ODYg708i/imYyr4wGy55DAvsk1AjMohW7DFGn+yL6SPMbtq+LBMX3kMfInWhw+cV0/aN1hpCN+QbgjuxZLUhvU1roYNRLWTeV076oEdoYORF8BZl+JqJexOrLsC5lGuxhqLe0VTo+IaWna8rbi9GSzR/VgQbFvwrqvc5BvagvU9pEvgMyAU2qpV4iU87XkRfeg0iRy3Z3W4x6jVkPsq48RgrD7qjn8a+UV3NZske0XLhtrqiafp4WTTaqIexbwDtmtgbwY6TZfbMOG2Nb0pp3bi2ZJnMPWmPnW+7+QzP7Hpr89A30sZ3s7ne2JAIz+yUSfqPScbY4mJM8icrDuDSsxWk85hfIhNMdNT7ZJKCeyDyE5zQ810Yol6MBuuxZ1dJWsfybqRMvo0bmSNR7PDf9XRuZhLKNvV9Eu8M1dEC3GbJ0v5bSk60c+ThyJ10N9Wz6ooadSvlfL6Z1+A9D5tRVgfPc/Re569kEq/uRx1aTWrSr93l1ep/ME+kGNN60Ixro35HShK/M1XhltKXonVZaMmMdNNa3KRq3yZbtAH1fEyktyPYkJa+x9pYLor1bnbb60crJRmVh90K25O9Vuf94JDiqbrO2vL9zG6VnALBu7rgnmlyzYUvSREnrnIEa2ma3tatwfr9UVkejHtA+SKveEgm0u5D5qzzcHqin1K/BedQh5cl8JDDeQT2Ou5BGeiYyLW3UXuWY0rlyqlMXk7TglH87ocbqbSqsTNmK+P6AbPiXkSZMpfMt7kWgcZVsK8ysLm2DnAGOzcX3U6T1Zz2AC8vC7JbOZ3tm/wo1ipegXu36qKHKeoh/a69yq/RbYV02M3tc8s9d5O5HpvO1LA61VFgzexJN2T7PyzQK05rn49DsuyfbS9tvzTu3QVoM+fTPA55rRPw5TevLaMD1UncfZWbroQG0bBLT5sAYd/9H0uwXowZxsZn9HQmOYWY2EY3XXJKe+zIyJ3zJ3W/Je8iY2RXIrfNQd3+lte+SfyckHG5DGv9YtO7MYVXu74Fmo9bkHtlIzOwItIHQM6i+n4S8VHqgGeE/akAcWRlnK3te4+7vld3TCa3iOsbdb860/jriyLzYAPXqTPMTbkczbJ9Eg7EfILPfPz03TpR6AFuhHsFcNGGtt7svteyCme2EGpB1PG012t6syOadVk02yoU93MyORuaDRUB3M+uXD4tMGbOQ0HmybV+rSVrzzg0ja2TM7Gxk2/wB6i5nH5uRZsPW+ejs/nuRb/QVVlrSeQ00KNcFCe7VkanGU1qyj7wHpfVaRqOB7cuRyes6ZPP9mpmNRhu99EU29c+jBv+VRjeiLtfDO5Bv9wigYxIsvZCn03rp1xdN/lk1pWeZ4toG8hU0eLot2t1tIhKK403++p3c/Z1WxJGZhs5Hq1S+V+GeRWmAtC9SxOoqi1Tv8n73q6FlJYagHt96qO7cg2bwflAW/gNk3srCH0iazZ2OO1EyM/ZHdasfMGt5MP+usEI/V3nuQCP+F5nZcKpPQKkWdg/kfgfyNT8KCZYs7ERUWeai6eO0V6G25p0bTDbd/mUkVAehCV3lH1sX5Do5vZaPIXe9K7LF5gXCw8iX+3/AC572DEjafVfgetOG1WNRmYJcOH8KnG9a2nhgOr8LGgdx1JgsREv+/rksHY0gy48/ovrTCa3Y2A8J+N6oF7MOqndzgbFm1sWX9FJaJviSC5etjpSHzM5/SErzT1o6jyAXbhhwvJlt62XLSSehmu23DCkPa+zF90Rmx83QAPomSIHogTyvzkG9rlnNpT9XBjOBQ8zsUXd/2tN4UNLyRyDlIpsh3u6mlRXWvJORtI9nUCXpjAY0J7P0BJSlNlMvC9sTaY+OhMWTaBBrH6QdfMPd/7w8tOSteecGxZ9p+quh5RIORVvjzUcf2ifS3w2Ri+RW9bizJS14T+B/3sxa97m0zEN22yloAtYV7n6ZaVr+N5B3yHzkqz0BCYMuqLt/nye33Ea63ZU/y5acddwD2ZifRY3ZU6hXeRESSle6ZvQus/qWej27oMHO+RXMnKejQc91WyH0szzYCU2+65/ecx00rrEF8pI5CE1ku7COZ3dHSkHW63sf9dC3IK1SiwZ5JyLTzhroe38bOSQ8ny+znClqKBrrWIRMOd3SM7dAZfdld1/CApAan/faxQy8ogt9+Gj9l1fRYMu8ej7asrCbo5Z7b6T9vYMK9ffAv70Bm6k0ita8cwPTsCYSWmtRWlf/DTQg+CwSai+6+39yYfYAxnsdi4ol0022HsoSeyTnPsyxSIiMQEJjZvYhmpbY3QQJ+PGZ5poTQGsgn/AO7n5TC/Ihv2VktuhcP+T98bOcEMkaqDWq2X/NbANkVrnZ3X/WUuFaZ/qzfNgCKQ79yhtb09IKpwEnuPvASs+pMa5uSBPfEZndxiAvoXVRLyhbpvtfwO88tzSEmW2Nlht5vfy5uXsuROaWSek3BSlv2TIhvVL8W6Ie4SeQl9fx7v6nao2+mW2Gemhb5Z4/CsmFRWWNxYbIZXS7ptLaVhRC6DeSpL12QhVxobs/085JWq4xsz8jU8krqIs7E3mnzEVrsGcToDKBNw4N0P6hmefWpOHmhP4f0CzJA5q5fyX00b/o2sZwVfTxbo00w8uRoK62Dk0nZH7qQWlweSBSGAYjU9c66fbOaHG6ydXSTmnHMHOt3tgDeYS84e7HLguhX5amD9HS1q8iU07/9HcDtL7Q35C2X/PAatnzT0O9mQ+Qx9B4NGHtWdRbfQP1UmfnwmR156/AU+7+8zriWxs1KB8gpWQG6s0cispsWor/Jl96m9Fqz1wJwNM+G+V1NSkpC4Fd3f2RWtPaKFZYm35bkfvY34SPNNNTkSZ7nruPa6+0LaecjBrH5maQZrMpZyPtqkmhX0e3OLvvEeCcNI4wDAnTW5FJp0uuZzEUaZk/Q4LtBCTYvoKE9UnIRHBdFYF7Mppc1Q8NNGareE5BprXfI8+XZ5Dgnk8V0rPLn384moF6bu6eNsU0I/mzaIBzPsqb+cjk2QlpyDPQmMfFvuSChCujbQUrNmwVuBWZsiagRc9uQwPFTb1nVnc8pbNJoZ8a5t3Qd/spZMbpSNr3F/nsP4KW8L7X3V+tMe2ZgF9iU6VygZ96TW8g808I/eUZM1sXaQGPuvtU01KwP0EayQLkTXKmu9/TSNvvx5nyD6AGDf1h4NONslX7kotjrY9stxsh//c/IM1xvJm9juz6W6DufNcUbkO0wuW1Kf27IaF+HSWTUp6NkBAZiRqHp9H+vTVpiRnJ/jwIeVrtgLxlNkWmiNuRiWdZsQmai7I20oi7IlPZBOQQ8IK7z8wHyNX/84G9zWzLWsrT3V9C5j9MO6V9UEPDlj33flLD7k0Pcm+Nyn4easAmo81aQA4P/3b3w5tLa5X0N/eOWZ15FpXpMieEfg3kBNCuaDDwJWT/3QNpdEci88XP0/V7WI42TVieaOKjyM6PRgJmJWQCahTPoPo+AGnbtyLhtR2aUJPZoful31VIC8zMK5kgfg55JH2UZlty28PTmkpEsu/3RAJ0YzTR6e8VlIQ5yPzzNqpbL6Q0/9fdH2jB+7eGCcgXfRyq3z298lr3HShtz5lxPzKVdEdCtmbKFYYm7ss37H1Qw15xG8VkejkD9QwOc/fn0vmvIweNg5HXW7bQXSPrYJ7HaKByUw8h9Gsjv7PPppSWFs6WKb4PwMweREsbQwj8ush9uE8gW/jaSANrNTm7/hSgr7uPprTw2wPAr5Iw3goN9GbukiBtO1tW2VDj8E5K84f5v03E3xUNHu+CzAoD0Dt2RwOjfy8fzEVeT7PILfmBBuQXJVv/PsAr7v50W/cq3f0NSkL0t1k6SYpNJrTyaSgrz/XQAOkSm/20Ac+ghnIA1ffOfR+VxfczgQ/g7r8HMLMXkDLwd2A9M1sFzQu5pA4TVVPklZvhNF65aZYQ+rWRFdRTyPOiWzpeh6S9JC3no8ElL3PrWnZJ/Xjj7q+a2QJkUmjyI2tO2CWvG/PSKps7ZWYIy02gMbO10FIbw5EJ6Cn0wT+d0vGtzFZrZuOBJWZdm9lX0bZ5j1ZKI7J1H4Ds+/cijXI7VH9+lH9W7u/VZtatks0/pXkEyR2QZdSrzA2YZulttl67JrN9iHo1bSb0c7by6WjAfHSV9HhKz5q5sJnJZUu0SmhX1Lg/gRqRYcAuZnaquz/c0ka2LNzjlCbfTar3Wa2hQ/O3BLkP8TmkQZxm8hseCtxmmga+GJkJJph8cLOPJAR+jSQNFjQouEUz91ozAt+QsP1uOu6ctzu7+6IkPLdAdt2LkDb9ddQAnIfs9ju5+9hcF/x3wO1JeGTpPRut5VOJo9BEuX3Rmj9XA2e5+27IrvwdtP4LJjqk/88FHklmhvJ364RMLdkuXMvEPFDeMDVHLn9eQQK1LckE9+vIDNsUNwInmNl3TZvCrIe0/3+i8ZjfA0PcfbhraYV9UW8rW3a5RXKzrCf0KlIAulUP0Ub4crAA0MfhR8m9dR/UMi9G7lw75O75DBoksty53wO/bu/0Ly8/9HF2yudR7lq2oNVQ4FM1POtwJDg75Z6dz/tfA3em/zuWhd0OuT4uRDbrE0iLcVWJa0/Spi+5c9ky0r9Fg3+V3uUK4K/p/zvSvVm41ZFt9+zsebl6NhSZdjLPklWR2Wl75HXyEmkhsEQeD64AACAASURBVOX1l3vPm4Cr2qo+lcX1ebTTWVNheqOF1WYgYf426rUsysoq/+z0/8nAq/myrTOdndB+HL3bu1zCvFMjnkrO3W83zRZcFZjiS3oJ3JfdZ6UNGOYCO7fHgM3ySMqDJdw3k+/5fC+5df7Lm9biM5PZB2jg7VvAnHz+mtlA5EmTLdWQmduyLvbnUM/s/zx55qTrS5gwcvevh8w8f82VeRbfY8ABZtbVSxP0Ms1zLjIRgcYPhiLhPRlp6p0ozQZ1tO3gSindvVBPZV20bEB/5L2zAO1VcEW1PFrOeAQ4rDVjD9W+n9x3mfWob6ayV1U+zHTgq6ZF9LZEPby5SNMfY2afc/dRZfFtCLzW1Dukns0qXn3p6x+i2eC/zY0zLXu50N6tzsf9RwWNNX8e2XLfoAktsig/9DFugWZufhdpsAORyWT/dE/NWhQSxIuQb/aByHf9JuTpMh+NtxxcpVw2ADbJna+4ZG/u/q2QsF1qC0VkQ17AkstIZ5r+och1cwAS/k8gN87voeWTXwf2yIU7Btm+s8093kazN3+HNqbZmeV4K75UxtmuUtm5zVJ5t3rpcTSwfjAyi6yOPOb2a8FzerHkks3bUZo1/g6l3tdawNdSPTuzUv3MlfVlaKnlrM7k86AzcrW9tKn6tkzKqL0ryYr+QwODHwIbtnda2vuXE4CvUdousRuyp99Xx3OGIC13LKVdwd5HJo/b0TooX6aBXekkyBYhG3+l6x9QwSSVGqYb0dpMINvxf5Dr52i0jHP+/t2QS+EOaB2Y0yoJmuXhlwR8x2UhwNCEvUcprXG/TTr/Y2QbrysNaMzmZ8h7pgPy05+MXGVfRz2tW5FH0CJkpl2pyrMy8+I30YB+/lqP1FCdjhYgzBqTqjuptXletnfFKcIPaZ37tHc62jkP9kwf66VJsC3IBCiavj+HGreoQ94Ud6Ku8vtob9ytkK12tVo/qDruyzS514CvNHHtxCrheyPTTHZvTypszF4h3BpVzq+T8vM84Lh0rt2360wN3N5ocPrf5Hq3rWm4kJluYmos90S9n0PStR1S3m9RZ5nenwn9dDwu5eeOaI7G+6hHdhE1jC+lZ3wBzer9UXrGeErjBe+hXugG7V1OYdNvQ3K255fRAO/t7ZuiduVk4EZ3PxnAzGYhe+poNDD+Acqj+2qwc96M9qudZWa3ANM9twaSma1q2jB9daQF/rfS85qJI09+FmXmaZPZdjN3yceQYF8Klw05S1sH1DB1TZ457tVXCn3HzHZB5qMhKe6BaNboAqSBXgn8hRrcJxuNaUb6Uci0tmlK1wdo+Yp3kWlkLjS9xWQTz8/qwTFIeJ7u7s+Z2SSk+d+EGoN3kbB+miY2Ry/jNaCXyw23C6nH6O6PJl/9VdBy4Gc1k8aOqG6fi8b5OiBz0DNoCerLUY/tFXd/ttpzliUh9JcNT5OERdHINXxro0GyzPf8eWRHBQ1UvoY+mmZJ4een/+8ys75m1t/dXzat9XI2MpHMRF453wX+Us9AYuY6WXb/aGDXvLuolwafv48ER7XnGbLHH496NuuipZsnmdnvgT+nRmxTtMb/F1Cj+F/kZfIqMgndgoTINNTYTUrpWKaDgSmff4w074eRZv9iSufrSBDXvFJqFTIBvhFy+5yazo+jtIRBV0peTlC7++p/gVPM7BOoAV0NwMweAi5ApscJ6VzVeuMajJ2IxhaeREL+u2jNoPloUcHm1p1atrR3V2NF/pEGroDjkGvhcmmbbeM8yOydP0UDkp9Mx+cAd6f/f4Ls8zUPUFIaLOuHuupnpOPeqGd1DrI3n480to/CtOAdMrPM4agh6Y7st/fly7mZdO6NhNY4JBT2RgLzj2hS38/SfWsjAbpmSv/uaCC0N2mBs3z9QvMKKtqa26g8s/f5PzRj+BspXQ3fH5qSG+apyFSS2fGPReshgeZHTKPKWEsTz+6Hdsd6Fc2fGI8WyHsXeVidTTIZ5dODvHOWqh+5ej4I6Fp2vTtS+o5DM32/kH+/Zf1b5hGuCL/mhAcavPlU7rjd7a3tmFeZwNwCuA951gxPlX8h6gbPpYo9vIbnr4y8o/ZMx2sjc8vgdPwJ5MWzWj1pRprej/MfOdIqN07/HwnMyNeHavUC2bofQ66LvcuuWRIw2bM6khuXSP/3rPLcVZOQ2juf18uoPI9GZrmlhGAD48ryoD9az+oF1Es6KpXx9agRvqSFz18LNZrHkzyvKG1SMwS5aQ7OxXkOapC3aiq/Uz35IhrAf4WSJ9YM1PB/ra3LqalfmHdagKeSbYI9kb2xQzIFLF/du2WIl8wgT5v2Gr4E2T/no6nozwDD3f2Ren2W0/1zkylmvXR6PTQonD3H0AScQci8VMu2jIvNzNGA5JWU1tpZYGaT0vIOmwLzbMnF1qo990Nkqhnq2hryI9ORu7uZTUW9E9Kz5uSWBrgFzcz9Tro3n35D7p2D0cB2k/7pDSKL+1FkxjkYraGvBGk2+lqoHNZ193+2OKL0ni6z3QnIK+tuJEBfRA3zCFo4X8G1nPavyk7fglyAR6MegKOB82yLxoVIsfioblfgCNR4vIR8/59DnkGvobr5VpVwy4QQ+i0gTZ55GdjX3Z+ocMs0YK6ZreZVNtsoIq4tBw9NS1hkXelXPW1+XY/AT2Q234eBoWY2AblqTqYknPoiwdi14hMSpq0AD0GzqgciwfIB8JkkgDdAwnULSgOrFyLNvMmBQ3efYWZvkQZ6KwiLw4D/mNl2yKyzKdIoJ6PB481zedPJtPFHfyRcVqe0eFybr7+TE8QvmtmfgG+Y2TbIft8TmU36ID/4fmb2kGvBttbGO9HMhqU4BiHT1hTXUszZgOriFtQhsvCpwd0feYKNR4PE41EvbZzndrmqpjyYWS+kLIxGbpqveRttS9pSQui3ANeIf1b5nkgVrjulrfBOQrbC1VHLHuTw3KqNZtbDzAajBbl2Aa529ydqHHTNPrqL0u/vSAD/DWmCoO7/NTSxab2ZfRKNufRBNvcZyIbeBWmCjoT7bEqC4GfAKG9mc3Jbcl35o00rbmZeHAORaWHb9P5fQlpkZ2TyyrTDoWZ2KbLtb5LuAQnaK0leYS0VeK1ga5TurZEW/A4q14nIRDIRed00BNcyx3MpNXIAmNmBaIzkb2jT+JbM/M3uH+Hu3612U1r3iGq9d3efadp05hRPg+zLGyH068TMeiTN9EXg4NSyb4m2VtsICfouyP5X8447RSG5x+2KJmrtgOymq6KPuQOasfpELR9tznT0kJkdhMxqc9z9jtw9D6IB5GrpWQv4BfoWDk5pOQ8JgfdQD+E24Cw0cDmnTnNdJojHoYlWmTnhQ9SQzEOa5AvpnoeRsP+mmWUTkTogoTYODVo/h+zpL3mVvXTbkpxWPAPNQr0fCfiJ3kb7RCdBeizawOhuy62SilZFPRZ5ao2lBauO5hrM1c3sh2g+yTPIZLaY1Itoquxzjc39wF5mdm92v5l1Rh5GayM58bhrI6ZlvgxD7JFbB8kv+TvAfpSWWJ6DPsAJyGXraWCM17lTUlEws1ORVj4R2YXHIyE2BWnSs72F+6um5/dEAnJrtHTBW8h98xWvvMbO59J9X3L3e00bc/dMGtveSFCv4+5DrMIy2Wa2qru/3UR3P1tjZTgSJDeguvI/ZDKYXiGMoTrWG9mFL+f/t3feYXZVVRv/rfSEEFoICQkhBEIavUtVpCOIggoCKkgRBJEi4AeoKAoqTVAEpWlAEEEFQXqTlgKkQRLSe+8hkIQk6/vj3WfumZt7p902mdnv89xn5t7T9mlrr/KutcT8GI4ogOvCel1RP4cF2fuoFIJw64SUny6optL7DRVuqft0KKJF/sDVD6HaOmhSPsvd++W6T/U5HrJOTnX3/zZwrPshRtpcNHGA3Iw9kdtrN+D/3P3WQsbaUERNv35wpN3chWqpHI+Kds1pzsHausLM+iML6JHwd1Ft7pF67n8XRA3tgSyv14Igb4OKpb3i7vOzNltGKJ4Xvn9GpknO/5DLpXWYDPoGXvc+iII3ECVIbZVPoKWCvPehzOGaxl/VecrMnkVK2Xoz+wJi96xJrftN1Le3lalw2KNeQ7/dUsNUonhvpMUmzd/7IC15J6jWiKghWIhiHUvD8arKG4drtIYgzwoQ+ElN/seBb5jZAuSKW40Yed2omxtyM/RsHI6ej8Vh3FNR0b0Hybjkyl963StIHdoYP2Qoa+2AXnnW2Ry12qsoNauxfVBC0sdkUuZbUCSaITKdxyPz/likZSXc/d7IsjgxfQ/D/x3Dur8klD1AQdKtkfa4DgmchHa3BGnp/0Dp9t+h/nVfDLmNEq3v6+H4LwIdc6y/HXJf9A3fD0AuoLeA/yDr8vSwrKzc73APb0DK0MeofMEENImuRmSHYhzHkPA8N8eyVsha/DNqcl/osbqFazsXUUXfQxbjOhS3+Bj4Zq5rEf7ugyy68xDds+LllNOfqOnXE57xI68ys1lBu9yDTJp8P+T6WQusN7PHfUPtslnC3eea2WDg62Y221P+aDPbHLFiDkK11+fUxSWQWuec8NOZ7j7WzN4gkwU9G8VX9kcNy6uoje7+sZndjMreHmpmUxAHfQlyUYxDNMEPkCCbj4TPCmCl10G7TmmQn0eW4UDEAuqFXCGfknFvbUkoB53SIo9APWq/izKZ+6F4wAmuLN77w5gfTp9bmXAgEm5XI/fVV9392aD5/xD4ipmNdnVEK9TF8yhwY3DBJJ2ndkOJYj2Au70AyzF1vU9D92kUup4jUMB8DLpPy8hBu0zJhndRzCq971Yo1rcGlfWe4e7DGzrWQhCFfj1hZlcDQ939FeRjHYAegllIw/kz0jinh9+jb786TkXBvz+Y2afIz9kbCbvPkGk8CrFW6oKEttkRMUdmht9Ho4YaIGbVajJdiqoJHne/2dQa8bwwnrdRItUDSNtrAaz1PPRbM9vU3VfkEmpm9jdkfdyKLIfDUCzjrXAdxqFnZw7wcR6hOBtZBlPC9zZIo18cXFfvoFgTlKkGT+pcv4uCq/eb2UXA58zsJXf/yMxuRI1jjkQujbrWxcmHK5DMugtp3EmZh/HA2e7+XoE+8uTaD0LB/7nAQq9jg/ZshMnpPJTYdTDwDXf/h5ntjphcl7lozGVFFPr1x+eQIH8FpeJvQkZLW460v2q83EpE6Bsjgia1MPirf4sStEagErbTkM9zERLY1PGaJeu8g/zvhwD/RT7Ty0wNWvZHmvXPwn43YHa4+3OoimNSV6aqZoqZnQ5sb2Z3kdHS+yJf9UFISPyQ3ELtbTIsrrOQi2hpOL2qOFCIGfQxs7nuniSDJeMcjYKiXdAEsDWB8urua4LgX9VAqmJDkZxrS0LyGuKm74esl6Qj1SfID95gBC0/6VF8PorbHIbKIk8GRrraD+IFNCZJtnEFxpNeymbKnxgQjjcXmOB1K552NaJvP4votv3C78+g2kqHAI+UPZhbaf/SxvZBL15nshpF5FnvaGTuQgWbJjSmDxLA81CK+h7If9q2CPvdDPW0nY3YOpcgofQE0v7/UsO2PZFJn9RWN8T+OQUV3/oAWSELkTCbgVw865G75ds17Nuyvm9CjhpDKN5xK3B9+N4ia/mLSPu8Mozjl8l1Q5TTuylhSYQc403q4lyChH0LMq1Ef4eUo/vRBPe1Mo2pJ/CFXNe9gfvbFmUBzw7n8Vl4phajWE7e5xYVE5wI/CZ8fwB4JvzfFrkZ7w3fy1qmpWwHaqofpNXsiWqx3I7M2XFIy1mTutEVa5rQGD7J+SNNcCWwYwmO0RF1LpqIXG3LkCVxTZ7103WBPkuEMdLK1iPLY2S4n+uCgL8S+WQPRNm7g4Gf1XF87ZB1eGH2MxGE5gXA3DzXbXfks38OUTjTXbr2AXao0P0cGK7BFeH7RSgmMT9ct/+jyDWB0KQ8EBVi+xuyhFaEe7a60OOlzu1xNMH+Ak1inVDph0HIMv1S+jnKeqbOQhNgr/D9HES/bY+so9uB57K3L8cn8vTrCVMJhp+jtPw+SOvfCs3e05Fg+CwsawXc6e63V2a0jQ/BdfI6cI+7/zn1uyHrqCe6do97A4NyZrYzejlnAVM88P5rMvvNbB2qCz8L0XHPQPcvcQ99jIqtrcja7lBUr2Uvd59CLTCzCcAt7n53jmX9kf9/gLtPS/2eBDI7oODvHM/Rh9XM2nuZU/5N2ehfBJZ74M+H/IGBaNId68qkLeYxb0HlNtYiQToufMYjt8w73kAKdepan4gm1xvc/fdZ62yKJpt57n6OZfphp7c/CD0XB7t6AOyJXJBJCej/IaF/bUPGWQiiT7+ecPkVj0IawAgk6L+BAnMPIC1ziYvdcyNwjJk96e5Tom9fqfSmjMebAsNjBaK49kDc7o5IsI0BRtTnmgXf6/aosfr41O+GtEMzs6r6LKYsz6OQi2kVmTjDpujdWE2mlv1TqMCaoTIJHZFLKaHy9QOm5POrp35vFb7n8uPuFK5Hd2Bacu5BiJgroDgmbL9DOOauiMFyMEr6Or6cPuJwnBeyfv4U8dtXlUDgn4t6Bd+LgsPzUExhVZHOOckn6ILu7SPhuG1QMH89ilMsR/cJUtm/qWf1fWQp3mZm33H34WEfZ6D71TWcQ9kRhX7DcDp6OacjjvV04EpX1b40fg0MQ77OuymcvbBRIyXA90UazyXoxUpokDORtjYrfKiLwE8J1KOQQHgNWGKqc7Mm7CPXfvqi2jddkHXWFrlPPkQW21QX1fB5pP1vhdgca8zsOsRKGU6qHkwugR+QlAZ4B8UPniUkhAVhsBlip4wlCPb0uQfB3z6scy4SGkmy4ISwv8Fh3bI+Y5bJOu6O3BoHoWBzBzObifzaLxXjGCgONARltDaIVVMLkmv+IZkkv/9lWZ37o2ft55CXGPCpmV2KrIVxpk5xaxCZYBHqAja1BOOvFVHoNwDuPjr535QtuRCxdwiCZl0w93ZGgmTTZNMyD7WxIZn0JiKGzVtI8E1BZRIKpdo9i5qOnIHcaquhyqWU1Ebv6e6/Det/iPjvI1Dy06bufmOy0+C6ABXyuh31Pb07MG12QJNGb9TsengtVkkiGP4I3AG8bGYPoYmuO2rcsQ1wgeevzHoKiin8E7kOEnfGClLlGcqNIPAPQNdnW8RYehdp/AcB95vZle7+aBGskPEohtaFUI46DVO+RwfPUd6iLkjdv+HoHB40s9vQfWqDso7PRBPzY/n2ExSRN4Kb6GAy7t6pwIvuPiHftqVG9Ok3EGbW2t0/M7MLgWuRpv9Qavn2qHF3P5RE81GFhrrRwMy2ReyWSTVozLXt47uoRstLKGC2GdI6N0GC14Hd8/m+g/umBakyvabEmmNQ5ml/JMwWIoF/itezZryZ7Y/ofPsi68GRoPydu/+nBhfRYGQBXeA5Ev4q5T4MsYZ7UUD5qLQGG+7pTUA/z1O/qI7HSHzlXVFOQmd0P7ZA93e78NkRaOfuXyrwtDCz3miyT9p6dkHupOeA6zxVajnP9lX3Me33Ty3viBhAi8p576Km33AkNzBpunCXqRl3a2QW7oCofT+OAj83glaWJK/0Rf7UqcBLZvYrD9zreuI7iF53GEp4moZcNdPD/3PJ3LtkHIlASV68akIpvKxPm/ILNkMsjpVI+1sRJoqWdQ0euppvfw25Dpamg7ZheTWBnxKUjgKjOTO8Kxgv6oyu93WuhictyFCaZ5tq2QwKY2yQlp86twXITfJF5CabijTw1igeMwuYYWZtvMC6Tq7SyCeGSXpLYHL6XQ73Pe91T99Hd18bgrl7ICbWQDSZTEaTf6G1ieqMKPQbiORGB5/vN1A1xv3Qwz4WcaqfrsFUb9YIfuzrkA/4TURHXIZehjOB7mZ2ludgqeTZX6JV3Y+0+rFo0p3toUlLPqTuZY0vXXBjfIo0ys+jSWVXxDhKSm6sqGEXyViTbmojU7/tgNyBOyMhcK27T08ES8AdSAid5O7/Dq7EruH4O6AA4Vh3v6+cwVw0+XUmZAyH+5CeuI4EpluoSFrIgcI9OAJZWi8jV+EYQtMTL6BCaw3HHAKy7k0Mq3boPR8TfPc1scLaIcrnucjN+zF6LscgT8DIcIxyJdVFnn4xPtScpLU9eiGbXVP0Wq7ZEYgB8WNkoqcbfh+GApQnhe8NvnbIT74L0qS2Ry9f0ue2XrkTYR+/QMH59ejFfQhlV74NHFfX8SK+9mlIAx6CfNWTEed8MnBo9hiRYP1bWD4MxTDeRsljY5ElM6gh51bA9U146UOQEE5yHbZBk9f14Vr9sIjH3KSGZe3CsYuWqIaUuaQf77pwPvMQsydp1p6vP/IFyPK8BVXd3AGx1dqU6x5lf6JPv4iwTMu9E5GWf7uZnYw0nQdcZn05U+UbHSxTfOwuJIyP9UDrS7lZ2iMa4Gvufl1DtFYzOxwlBh2ChPMBSEP/A/CRu/+gvvfCMoXNrkCc/m0Rfe9YVDZ5prt/P5f/Nms/rdFkdzGZloizkKtiCplU/1VZ2/VADJ3VyMUxI3wmIRfWROSCqNXaKBZS9/NgVEuoL5qQliDh2xlNbL/0EuQQmJoY9QjH2Tb8vw8w2t1/Utu9qMP+d0ITbQd0Hu+ga38Asu7nuvuBObZLnuWh6N5c7CmrP1hwRiil3dDxNQTRvVMkmNmOKOjTHfG+14Tvo5Cf+TSkDZW7CmJjQ3L+i1Cg1WCDQNeByGWRVOGs1yRpZgPRtZ+GKHI/Q/dkEnIj/QI146iPwD8AWSDnu/sDpp63NwGPIrfUB4iaC3l8syk3wF6oUNnjiASw3OvglnD3mWb2EzSJTfLG0TzFg+B/08y+grTZA5H74w3geQ99pBPyQ6EHTAnUw5HltiPylbdC197I0FcL7XNxNXp2voYKLSbPzGAzew941cx6+Yb0y5YodjQXPcfV6KXhOUiIAuUNwFfCvGhqH2Sq/Rlpa7uhlOu5yLfcAj04Y8K6zboGDxl3wKFIS/0DsFnqOh6JXBbvAZ0beIx7kK90p/B9NfDF8P+u6AXcrobtDb20LalePmIp0Cd83xsFDrcL6/0ceKuWcSX72h/RLPsW4Xp2QBp1XyRszyCT+l+VlFaB+9wRxWdOR/VrXkNJcxRjPOF+fIxov+sR5Xa/8K69gjJhi/GcjkdZubnW2QPRfQ/IPq/U9gchReMW5NPfFlmG1yCFYToZl1xZZEPU9IuDbZC5f5W7jzKz+SgZq6u7TzKzsWEdvEJc6sYCz2hKQ5AwuAY4OTA+OiCf7Duolkt2sluNSGlMhwIPuvvEsCiZjF9GPtl5KPs1Z4/S8H1dar/d0P1rAzxmZgd7KOOLmDxzUFb2fTVpbanfP0DBz/2Q6Z8cp1U4/46IhjjGM60RE+026bG6BZmSFf3C+WwTxnMNKhRXZ0ZRQ2HqMbwDyoEYEMbTHblbWqMJdym65tOgaCyjHwKvuPuJIbg+1t2HAkPNbCLwuJnt4w0vXdwSTSbvowqrVUHocN9bowl2OJmua4Tl5l6tf/NClJ+xDGn/y9E1mYFIDMPCumWRDVHoFwGuBsebA1OD/3muqZZLL+RS2AsYb2abe2TzAOBKnLrVzEYiClsHpDkPB97whtHtEpN6KjDAzDq5yhR/SKahyteQvzlvxq+pF/LpaCLvhwTpZ2F8uwBXmtnL4fut6D63AB6pi0BzlaK4FjjPzPogq6QjygEYiPj//ZHmPjhQBvuhYnLHoPIDa9H7m7T0expNlqPcfXzCEApxpm8BN3tpmpa/joLSaxGNdSGqgzMqfCajoPxKL0IsKzWp7ox87SChe7iZjQznOBhZOHsCMxsSR/OMG+oO4PvAnWb2DDq/7ihRLskFcFQLqCUq1VD1DATf/eFIBvwbXa8pSFGYj1qG1sguKzai0C8QqQdqGCorMAlpcdOAA0zp198Bfh8F/oZw95dRdmq1YK2Z7YsE30surndd/J7J8r8iOugfzOw8NJGcYmaXI2H+cEowVtunKXv3l4gH/jYSphPQJLEAaZjfR4J0BTLXX0EugFn1EDBdkCvryPB9DRIEY1HG7QQyVkB/FJT+C3rOfgS8CnyYT5CnzsvDtXiYHBmsRcCfyDQAn00d4xMFIJnY5yOXFqjs9Mkoy3sUap6TsGygnvz3QIXdFd2j7ZFb5ktIs/ew39lIiE8EhkIVF78Fct+NdPdPgnV2MfAv4PtemtIR9UM5fEhN+UPGd/dFZLZPQv79mch8m4kSuLpXeqyN7YM0xC+iqpIQfOnh/wPQS3x1+F5nfycqfXFuuBcrkEBNyv3eAmyeY5vE3/4tFHi7FL3srXOs2wU4HlkCA4D26X3UMrZW4e9FKKfgNERt3Jo8VM98v6euWavwydnjIZz/0ZW+30V6ZpLn4wdo4huALK3XkKAfjyyOQbnucx2PsR0S7GuQVfgeIgZU9bylhhr4yHL7fGqs41Csoab7WLaa+pGyWQSk/K27oiBufyTQFqCH8S9eYFJKU4SZ7YcCoLe7+3NpzduUvPVz1BRj/wbSNhMfcyfkTx3tofyxmR3q7v9LrZtQD88Ix93DQwer1Drbobrqr3oO5kxCw/MSU3KDT7mqTERN67mSmcYDf3L3m0s5rnIg9a71RhPnMHd/JLDnjkHxjnHAP7yOiX15jnNi2M9Mr6N2nrreU5EP/+bw/aso5vMPd3/clIiXtAndHrken3X3exrynNcXUegXCeFFbIfqsmyBfHvLUsubNT8/jdSLuy/wH+Dz7j4ux3rfRZ2kejTgGFujipjZ7puWKMj4JupqNjoIa8KY+iA3yt3u/tfUdpuiwOtTyPq405RtuR4V2CvoRTWzLZA/f2/EM98TEQE6F7DPRAg9jgqynV7IGDc2lJsKmbrej6HqrmeE33dCxI6ByD00CymFbQhNX4B/ufu15Rhz9OkXD5sjX/Dv3f2D5McgZBIuupV6Ft8YkHqox6JaNtsirSobJ6GytG29/kHIy4GRZvZY0LhpygAAIABJREFU+pqH/+eZmSMBO5qUdu7uE8zsPuBSM9sDBe42RZpZD/SC3mxmT3tW05QwYeyLEvOqWQmpdZJMzNVmdmcYw05oIlofjjcF+YmnmVk7z0rSagCGoEbc5e3FWkKE9yppvP5yWlAGv3qSD1KJ8x0KfCt1vdugGMFMZP3PQLGAj5A7arqHYG45Jqko9IsEV6W881BCygfJjN1UXrJiI1yfj83sHeBXZvYtV3C1A6IAfh35zb9eH4FvmSSv3VAZ5aQJRgvkN08CqB2RZZYLuyN2yO6IXrccUQ4nIjbHGcBvzGwKMs+TTl2t0OR/GPBGHq1tDgoGDwrbJjVkJiBhMA/FFFZQeGOQqkQi5FPeFFEFN2oEq3ldcMF86qFWfyJkK2hRp+nI15G53h+hQPB0d59XobFVIQr9IiDlurkUONbMlrn702HZDkh47IaEzU3u/kxT0roKxJUooDncxK9eggTzFsi3/p967i+5ph8CJ5hKX3dDArYrqpbYBU3O/4RM7kDqnsxHCV6vI0E/EfjMM+UjuqEa6YcgrfyDMM4pKNN4fNhvLq3te8i6SP5vj1L5i9phKuv4H6DruR1NQOiT0eL/COxnZqcBT7ia2xjyle+GSAKj3P1PVmA5hrogdb1HovvaA1VRXWdmE5MYQ1bsqmXYdH32slIh+vSLCFP7v/8il0XSdg/E8Z4X/t7o7n+vzAgbF1K+/T6o2fheSAOfgbTf570efP1AtzwD1T7aHjE7ZiC/6TLkS52IXso3s90ztey7bRjbMUjYj0cTx3w0Ua2o78tqZl9GheeudnH3k6YtiVBbX1etNQkiQ+6KjaayvqOakqIRXGX3IYrmaNSboDuy4Fag+NoD7v7jMo4peaaTLOGHwvfH0LPzLOqPOwfJhCG5JqNSCv8o9IuAVADn2yhZZyh66CYivv4MJHAWomSMQn20TQZZWk+HujIl8uyrDXKdHEmmleWlqKHKtIa8RGEi3xuxQvqgrNc+SCDvZNUbZXRAfPHV7l5TV6XkeTkFWTn7eKqnby3jqRLu1FCsy5Tda65GP02KRJBFBBgSPitQIlhSrXQOmowXez0zu/Mcs07XMHVvf4wUl/fC94OR2+8QFKRfjyYlR8/qayj7dzHKxF5SKsEfhX4RYcrK3R3dxMUoUSUdYOqEovpR6OeBqZvQ3kjr3xvR2erMYkm7zUxNbf7hmSqemyK/fjfk6tkZaWIblGMIMYCfI252B5QUlLTM6wb8wt1/kWO7GxA9r39t7oRgmVyKMm/PQdZhb5R92x8lH20HfK4mgWMqhdAjbNs7bLMncJu7/6upCf0EwTI6DLnUFiHBv7LUbpysMbRI/q/PNTbV5e+NnsHLEZnhY2SRTkQU20dKIfijT7+IcGXcvp6tsaZu3HnAQDNrHJl5jQjBL3sFEvJOA1ksKYHf2t0fDP/vjipgDkAsnK5IkPdEltjfcuzqQHS/rkZF4b7q7s8Gzf954FQz+6u7T8t6MR9AGbu7ACNqemmDS6ddONZElAy0FLkBVyLX0VzktlgQLJkdkYDfkYyA74p89u2QFbA8/N0yHKfJCXyoutevZP9uKs29E6JIHodcqmPrK0BNjd43d/cPTTk4/YF/pieVAq7tOBT0PzT8vQ0lde4CfBOVfXi0FJp+FPpFRmAU7G1mt/iGtL1RyP2wJ/BWOYI2GwPM7EBUpngqEpgfUSCLJbg1WiHB/RtkTk9GLJnXUXXEjzzkB6RcTMk9+S7yt95vZhcBnzOzl939IzNbggTv4UjItzCzdLLUCiQgRkDuNngp7bs3EgCzyNTHTz7TgHmpc/8CCjAvR1nHa5BmOBtNjuPDZ5KHxuBN/RkzFcM7FE3oeyLNuQsSpKDrcy+iB9cXxwOXIctrIPDbxG1n6omwDZkkq/4o2P+z2iwrU07GD1AsYhpKQBwWFo8zs1dRFnAfAimgmIhCv0hI3eivoch92uxLXrp3kDA7CJWEbUFleMSNAlnXbDlwkbuPrmWz+qAXmkSeB77ldWPIJPekZRgTiPK4HxIk89Gk1Btxr7NxPKoNMwOqMYOyhW/y/8/C/7PqYP1NRBPWMDR5TQRm5At2WxHaEzZWpNx430d1icYgq+gN5F5NuPEtkdXUEIwCtjb1UlgFbBPcd5uhSb8LyvZuE9afD/ysFoF/OqqCuhmawO/z0As6xGtA1t6Xw3kUHVHoFw/pG7Y50sKy0R694J3C9yargdURyTX7GAXdJhVlpxkB2wq5Qc4NrpQWpBq3oGBsPv/ve8BpYZsnET3wWjN7BLlbOgJ7mUo9zAc6mtkJ6IV+Hk3wVcjWtpPv6QBucPUchOig2yLL5++uBt24+yTECMk+31bIxZOUgO6BNNB+ZnZeMQKZjRCJYL0fBUAnI5fg4mTyDDGTm4CrUFC/vs3H30dMrydRYHg9YoctQTGEiWS6lo0jq8RyglTg+TZUlPEB4OfuPi0sb+HuaUtxHan+yUWHV6hwUlP7kCmu9A1kqh8dvicF2TZBHZuWocBcetuyFVtqTJ/UtdkD9SC9LHzfFPmq90Eaz6VkCq/VuQEH0trnAF+ryzjS+0fm/GBU1x9U5yUp2jYOcfwXIEEzDWmTa1E1xQ2K6yFLYcsaxvBlMn79SWjSmYaEy9lZ59Qy/E2u37eBx8I2K5BwWhTGNqDS97nCz9iXEaGiodsnrpgrw3X9PaEPcD32kciGY4Dja1hvE2Db7Gey2J/I3ikyApPiLtQd5xbka+2KAonHAHe6+7WVG2HjQ/CPXoFqki9DAdwt0HVLmkivQh2h6sxbd3cPVLlLkFvkNcTe2RuVS+iPAmdHu/uLWfTLlii5Z7m7Dw6/dUWTwTLkI94Z0UNbIGbPEHefkGcc64HveKqeT2qdL6Iyva8Bd4Z9r0fWxI9RNc/e7r4stb/k7x+Bs8L1mQv8ClEBL0ElpX/q7oubum8/G6a8iq6oOubOwIFeoKvLzM5CFMxR4fu2yJJsDwz3WtpXZj1fW6PM8z7IKusV9rXc3b9ayvsVhX4JYGpgfS1wKtJaP0YBmYeB+33D6o03AW+7+1PlHmtjgZm9j4TwChSYnIKyakeixJtJ3oA67WZ2Jiqgth4J53R9mw/Dsf7qmS5bNe1rM6SNrXL3xXnWaUGKP5/ibQ8GXnD3n+TY5jGUVHS0ZzXUMFWPfAc42d2rlXYwVSl9EPm1twcudPf9wrLeKF7Q1t2/YWXISK0Ughtni/DphqzEvmhiHwD8KNdkW89jVPX3NVVbvQq54joipaQF8AhwTW3PqZnthiyHPRGLrDN6ruagZ+SsQsZaG6JPvwRwNbC+BLgetVUzYI5nBdxSL+JRSANtdkI/pf38EmnLY9x9RRH2m7ykHYDnUFXNKYSEOXefn1q3ZZ59JAK7O9KmD0L3qYOZzUSsoFe9ekG3fJbIKGRhbLB/JKSGIW29KqAXhPvWqIzCmuS31HYdw/mNRs/Yjma2o8v3P4VMwxWoZ3P5jQznABciBatr+G0RcnddhRrQF4SUwB+ISm5sgVx5HyDXXu8wjo7Ahfk09fAsvYaYXf0RVfNRJCf+CHxoZh2zJ/9iIgr9EsFVJGxO+reslxkyQaXngIOamwkOEpJB8D+R/BauUxtES9wEMR16ojomQy1T+74mDnyibf3J3e9J7XsbYGczOwlpgbsid8iXsql2QeAfADyEAqtvA++G9Q9CQcQrgUctfy2lNHPrOjNrkz35o9T8Q1BC2tDwW4sQJL4WkQO6mNmvwnqjkZCbQKaW0FA0MZyGuni5mR2BchxKXuO/EkgpTZ+gSe6/6P6MymExFfxuBeXg2+i53NWz+u+a2STgFjP7QbZVlTr+ecA4dz/czN5C5b+HhXXORIX4FgB/qeGZKghR6JcROR665PsQ5M9ujx7gZoNEIAUfbC8k4HshbvQuyEzvjoTbiyg2kj155trvlohy2SrQ7HoiF8iWZBp2L0OB2alhf+uz9tEBBfHWo4Do1NSybREz5DKkqeVEap/DkOthC0TbhYz2fQdibDxsZs+j/IT+qO1eD/ScHIMCu9ORsMeVSTwRONzdf29mdwD/Z2anknF1nJFvbBs7EsHq7n9GiU1VCAJ6H9T0Zrq7/7MIx1sXlIW7EoEfmFMtg5L3BrqnOwNjsiaahArcg4wyOBYpe0nv7CmIGbQ3ckmWBFHoVxApgfA+8ut1oTR9TBslQjDrZDO7BwnkMUhbbY2E8XhkCr8Xlk2EqpevFQr+Pujuc3Ps/g3g2+7+bgjmfkr9G3Z3Rmn+17n71OCvT1oSzjY1JxmUjKmW0x2DNMQdCUI/JRDao5IPhjTJxWF8/0ZugMmIEbYIuRI+M7PvAc8grve+Zradu99oZvORsFsDPO7ub9QyriaB4Gfv5u6JpbQrcpeuAVqbWR93/3UB+0+07nX6KvdhmHgSrf4UdO9aQV6lZCQZ2u2TiA10GnLtHIYm+8TNWxrrzBsBrSp+HCSUjqr0OMp8zt2Q62I75M65HPUWrbEHaWr72SjAmWvZVYjxArIeNuh1W4f9t0AWwRfyLL8TuVo2q20/4e90xODJtc6VqGTAbmTKROTtC4wmwHNQoDZ7WctcvzflD/ATNHG2Cd8fRcrUASh3Ygqwc1hWZ9pv+pqGv79CiXdno7IXOyHr8zYkpG/Ns31CmtkDWX3HhGf+9rDd3PD3HWCbUl6ryN6pMFLBwkmoRd9vKz2mcsLM2rv7p/XcJrlmw4H/uvs1JRhXEjcYgthXZ7j7nBAT6ImaYlyHcgtur+O+zgY+dPchBYyrtavExOuo5+9FuY5HJtdhN6Cdq5xEk+vhkLq2ryOL8Fp3/8TMRgBPuvtPTYUOhwK/aeh1SFw1ZtYFTTAXIvfgNGSZrkUB41+R6l9sKsJ4LKrZszpYqOeiSfl2UxHAE9BEPy2MuaQsq+jeaTx4Dbl3mhXqK/CzMAxVNS0lLkculBlmNgz5XLdBrp+bUBC3RjqkZ1ox3l+E8SQm/zvIH9wCaZs90bXYA8VCepHp5/A+ogo3KYEfkNBwZ6M4xuoQi1lLJmbWEllGfRt6kESIu/t8M7scxRA+H465GMXlRua4xuuQRTgKTfhrzeye1DOxgtwF/0qGKPQrj+Ql/gm5SzdEbIjkZR4MXJOHEVPwMYIW+aaZfQUVVzsQCZk3UKmFrVD3q9+hyQCoYh9ZWLdaI5RshlADx9UeafB9UfLaOCTkFqBs3sFIKCWlAXLmFDQRJM/Cm8DFKGbyKWJ9zQ7LOodPMjEX5Ct3BW1HUkuphGAdrAiT0ACUF1KlBIT8gs+hJMCeKF7zOvCGu5ek7g7E5KySIZh1vYAp3kSLXlUapnK3w4DtvUy9R031/nshrTrp4/uaux9Rm8Zf5HGMQBr9TojpNBsFpqtVJTVlO7dFAesm97Kn3C7bobyJ48KiWcDng2a+BZqcn/GQTVvM40PujmUpN+Qw1AXu2tSytqhQ3Gnoni1EE9OeiPv/A3d/rQhKwoZjboLPQaOAme2BSrqe7Kq5nlTdTBf7ihe/AIRrugo4zN3fqW39eu67M0qTHxA+fRB1tDMZyucKZN6vBL7o7vMCVXBnpInvjSaH44rtWgmTzwLgJHd/Pnm+cgiew5GV8qS7DyuFEGksMJXJ+BJyqTzpebKmyzieROj/GTGLvhR+NxR0vga4AeVpTEdUzl4oL6Slu+9fijhMdO+UDpMQ/WoflAGaq29pB6SVNcmXsJRIBfA+Q0K5qEIfmdntkUtgJdLEclE+T0ClEN4K97Mdeq8+RYyMlSirNhettEEIguBjM1uAEs1GIrrgu8Bgq14KYgrSHrsgqyjxgTc5uKi79ybfw8S4JWKJdUF1kx72HJ3SSjWk8Pcdqrshu6F+DTf6hmU5JpvKspTMzx+FfokQfHnHAjeGiP8QlKq9J9L+krZ4+wPDyvQQbrTIoaEm1+pASlN3/E9IUI9BrpPlnqqpksXImYjcCUluwXRUUmIeoqTOpwEIGiE5novk+ylh3wMQfXNEWD99naYCryIhk962ySK4/Y6leqe0TRCbaTplCpx6/sS8OeH/18J4W4f11oVt+iDyACk3XdEstCj0S4sDUaBmb8TJTYqJzUA3/F9IKOR6sSNSyH7gU2yKktQdd/ff1bZK+Hs5ygOYhbpc1btmSiLcw3E91/9hvc2RSzYJGg8L/uz2yC0wL6xXpemH5W0IjT6aKIMHqLqO56PqtuuoQ6e0MmEMsgB7mdmCoCy8AhxvZtO8ek+Fw1DOxrNmdgtKMjsUFW/8dzGUwyj0S4CUH245yqocRab93RzE9FjqsU9uXgTBdQByn+yNzPORqJbOW5UcG+SfdIJ1twXqD9AN9QJ4wdVfN+cLm0O4t0cTSeewj6QpSl/kx78ky9c7DrlsjgImZLGF2gDfAZ4pZ6C5QtgRMXieo+6d0hqEuvrawz1fZ2ZfA6am7s2ZqHDbIyEovy0KzHdH8uFA5CKcjBLNFkFxJqsYyC0hUj7epSmaVlUFvfByt/YNe+k2awRz96eozMJMNGl+goR/D8RseKAxucSCm+d69NIuQ/7jFSjNvqu7H51nu1Yo8LsVGeHeHbn+dgy/dyLTwvEBd/9uavuEvXInauDzW8TLX48miQvC/o5x93eLfNqNAqlr0A9p9Ie7+9tWv05pZUHKLXgxovpOQcr3NOQmnIwUxDnIelyABP4WwInAIG9AifFq8EaQQt0cPqiC4s2o5vZ2KEj4AnnStpvzB/gmElq/RIKvQ7hemwK3hhcib4mCCoz3AOQrvh2Z458B+4dlhyCe/OZ5tu2ECm9NR5PF4rD9DOTXPS08O6eiQO2p5ChREa7P3WH7eWjCWYOCiIeEdepdfmBj+tCATmkFHOv/UHJeh9Sxa72+ZBTtnZBr8Esooa57sq88220f3oneBY+90jeqOXxQun5S8Gs9ocYO4nl/RGiRFj9V1+tFpNHkWtYflUXYNXyvuCBDFLuXCW0Sg9D9bvi/J2L+HJpvvEjjuwC1VNwNlRM4Lsd6+wahluxrA+EVtj8fNZvfKcfyil+vEt2DRJgejFxrFyF3yReQxfh3ZDGuB47Md/3qcJykBs9NwP+ArfKtB7Qv5Hyy7xWydo8o9FpFn36JYWbHAD8EfubuvzazsehhfAF4C7l/9kNBmibLoa4LUn7SrogCmZjtSdXCtShguRDRICuK1Hh7IY06YemMQSwtkJUym0xzj2rNucM5XpL6fhvKqH0xfG+L3BKfIXN/FWJ//Y+M26IKruSjaglIZtYTuZC6ufugxuQWKxZS57MDcDLwFTbslDYUxdimhG0a8q4lx3kYvceXmtmvkfulAyrR0R251jqF5Xnf67oE8VPP2Wx0719qwLirEIV+iZC60ccjylZSd2U8GYHwMXqRd0cP4wYvcTNDcv6DgSOQ4HsrCHvMbBeUzDINCVYqLLyS8Q5Fgbcd0P19D5U7bo207haoAcsGgiZMaoYKcK1CSkA/VF5hgSvlP8HxyMU1OWy7QSDRVMDr0DCeAShAuDlKKFtjZg81NYEPxemUVhek7t84xAq6Gllp01D8pRuSq6uAKWGCzTu51PNejEKuoIIQhX7pkAiE5aiFWnLjB6OAG+ghMTIaYrPV8gOS878b+APwhJk9g4KYuyDf9nTgYnef2wg01mS8f0Fje85UjGslst4eQRP8TZ7VZSmNIPiT4NzTSFO928zuQyUUuqD6LKeg+v0v5NpPmGSuQn7/T8jkDcxEE8Vc9M4XFghshPAidEqr7/HM7CeI//8RUt7+i6z34Z67x0OhGAycWmiWbmTvlAipKP3hwH3AHe5+m6mF3aPI13gBoiTuW6KHZKOFqZfopaiEwEoU2HwO+Jcr8W0rd19UyTGmYSq7cTUKzLVD452JsnUfytLYa9pPK+SauCXsZxFSGmYCT7j7zTVsew6q6/4g8NewzfxCBMTGAkt1SkPWYI2d0jxHSep6HCtxO/byVDe1rHVaIPfO8kJdtqnjfR6VbOjm6rTVsP1FoV9aBJ70JcCv0Q1bihgZy9CD+B13f65yI2z8MLN2iLGSUF23Ri/25V7CBtL1hZlthl70lsj3O9sbSK8LQuMY5M6ZhcryLsmzblLj5Q/IIjqiocfdWGFmH5LplDYYafPTqV+ntPocr8rKDHGX3VHnqwXu/qCZHQgciWoAjShGvC48Xx+iWlOTGryfKPRLj/ACH4WE/5boIRyNGivMbARuikYHM/syojJO8kwqeqLxbIZe4OPd/aWmEACvyzkEBaIXgeHj7lNCPKBFEPoXI4XiNHeflrVtK2Q5tM43eWzMMLOrgH+4++TwfHxSronP1ND8QhQsn4ZYVj0QRXupu5/bUJeMqULo0tQEcwLwZiH3MAr9CiAEknxjF1SlgGU6Qw0GXnH3/8uxzs6ov+hvXZ2QKi70UwJ3Eqp3Mx/5/Nsg624qMN7dZ9VlkjezHihHYVcU2O2NGEDtUZ7HZe7+J0uVXAjbXIn8zPcja3IzMnWe9gC2cPcDo6JRHJjZaYjJcwuyLG5EHPx5qN7RT9y9Z77rnbB38izbBHgAuNndhxbrOY+B3DIguCfaIE2rE/KzLg/LWiDXRWygQrWA3Chg90A37IqYMb0QG+XQsM5HYZuKCXzLlDaYFz5tkVumKxL2cxF7Zm/E0Poqcv+szdrPfsCA4Bo4E3VbSnoELEeZme8jwbKAUFU069yT5ukXAWcgF8em4VhJ8bjBhQYCI4RgeX0fuNfdfxR+ux31d5hrqn7axcw2r8EHfzywysxeTphcyQTg7ivNbAByEw2lSOy+KPRLiNQNPBVpXh+hl6+tqezrakT9mmtm49z9vub8QoaXaB+UgNUDJdo8hLTVTdBD/wnSmgd5Vg0eM+sGLCynPzsIfNz9MTN7Ek3qi0IQ/1DgB4h5s4BMFcxcpQC2Q5nIDyLh/giqNTQTBYVnufvCfONIub6+iapqDkfP2weItrgwavbFQ7jea8ysD/CnFGV0EbLOhiDW1QykqCzN2j5Z/1SUnPkSVKPwtkelPNag579oiEK/PHgKsTq2R8IM9BB0RCVgT0eZlvc1V4Ef0A44CQm/1YhzvRQJsbHhM9nVV7QagstsFgqmvVGuAacRGDoLzKyXmZ2OJi1HRcD+4zW3wHsBmBaEyYeI2bUBzGwrZPH0dvd/pI6dCPSBeSaV9D6arWJRRLRAVOLRKKfkFTRBj0SuNNBzPBY9l9WQUkzeBK4wswOQVdYTKQDbI7pvW0TjzZmX0RBEn34FEKL9pyPTcE/Ub/VRpL02Wz9/0HD2QQ/6ZMR//6a7L8ixbpV/M0WPnQVc6+4PlHPcqTGdDHwLafajUU3+v3sB1VTNrD/qudCfjDBog9xHh7r7lFq23xppmjsj3/5RwPvu/iNr+lU3S4YUY+ok1KZxAnAXqqezE1Li+gDnuvsTWWyfluieHoie95PR87450vBXh+0/AP6JCB9FkwtR6JcRpmzJryA/3iYoc/Mv7j65ogNr5EgC32S6QW2wPLyArwBjvAAOdgPHl0w6QxBz4wXgcVQitytyS81AdLt5rjyDDQJ72b+Zaqv/GLF11qMM7k2RRfQmcLbnyFUI1+ubqLFKb1QWIMFnwDnuPqgoJx+BmR2JBH/SfP0z5KK7093fyg7AhjjedcBlyO3XG9UG+jeyFCaU0hKLQr9MMLOLELe8ExIIt7j7kLCsvbt/WsnxNSbUxGjIs34i9H+NmtYcVk7/der4X0WWW0ckqLdB/vRFiMe9FxIEl9SmZZtZJ2QBzkVurtbA0YHZdCXK+P2ruz+V7a4xs4NQctZoVI73W0i47IeskIfd/c/FvQoRZrYXulfTgbk1kTNM3fTWuPtSU37F9V69VMTmqEJrV+Bpd59TtHFGoV9aJC+3md2MZvbnkCDYEWlg05Dfbykw1t2fjD7X+iGlaX8Fmdg7FuJSKXAsrREX/pPw/QSkrR+ANMBrvOas2iQgexLieX8OxSluBPYIjI6k38CRnqN5tpn9FZX4OA1RR49191fDsjNQU+6D3X12sc+/uSMQNHYnZOMi4Z83kSr17G4HrHX3OcESuB34HpILb6OCjUVJ8mpRyMYRdULyMt6AaHsPI03gWVQv/k1UU+VX6GWEZtDHtJhIvQTvIcZElwqO5TN3/8TMjjazh4FfoJf/bKBzTQI/C/OQK+czpCRshayHJAjYknCenkpeC9vuDrzkogVPBT4XBAnI9dQeWR2kfo8oAGbW0cwuQQL6OeAZRCh4y8yuCJp7taqaYYJfH+i6P0cxABAl+WT07JyFKngmLsuCaZuRvVNipDi3SxGNbjhUJd+cA5yHtIJnET2xorzzjQ1pN4m7Tzez1ciKmlqBsbREL+o3UdD1fZRc8xxKzKp1Mk+t8wF6wfdC7KUJwI2mzNPtkRZ4X5bml/D/Z5MRIC+hvIE3TW35zkeuoqTSZFQwCkAqDnMtYlz9EzVYGYnk61Xh+2qUe5GwftLYFNWYujN8b4difrcGy6572D8U4X5FoV9GhFl+V6T1HYM0wAeBv7n7BxUc2kYLz5RdTlwcv0PXtRJjWWdmuyOK7j2BetnQfa0ws5eB41DlxvMRd/8dVLb3f8AfsxSERCD8F/iGme2KOo39FQUJh6M67/9C8YJKl6beqJFyxX0BZd/e5e4/zlrnB2iCPZmMUAeqXfuPkDaf9PTtGJavDJbYdJTbEzNyN0LcjTIlpwKPIeEwJrgDWqCuPM2qUFYhCNfsXFSf56Ug+H9c23YlGkui8f0QZVwvC3TJgcjyWAu87nmqMubZ169ROeBOwZ97CGKItMwzoSQC4WlkPXZ19xfN7FTg20jgP4uqda4q5HwjAKoa4rQL/98FVZTstUEJWYvKYYyF3Fx7V/2t5Shb+0YUwB1uZp3cfbmZ7YvcfdsgKmdhg44TfemRYndcjny096YDjbnoexFCeIG6J7TWHLTGl4DF7v71VNC8MdTiORi4Bgl8EAd7C2TZ/Sg5vBUSAAAUbElEQVSwNup938NEtyUqKJY3WG0qsNYW+LTS16KpIvVed0f++/9z90ez1hmIFLw/ufvvcuwjCeRejhSGT5Br7oKwjZsS/doAj7n7yux91HvcUdZUBmb2RUSf64MmglHAC+7+fkUH1kiQeqGuQf7rHbMpcGbWHrgD1To5qrGwnszsKOTH/xC5ZMYg2uZAxM9+3t3PruvkFNg6Z6LeCzsh985C5KL5Uy7hn2Ny7BCOvycKCB8MfNVVlTIqHQUiULLPR3G5xShYvhuiy4I09cNqut6mvIwDUPLci6nfN0GWQ516MtSG6N4pM0z1Zc5H9M1P0cPQASVsXW5m57n7v5r7i5gS3m+gZiqdzWwp0pi7oGDmQYiH/oesbSqG8IJeALzj7gdlLX7DzD4Ffo8Sq+oi8DuhJtxfR4Hh15BLYWfk/tkFEQKqIWiI2yFf85dQwbrWwApUFmAFcgFFv34DYSqdkLSk3AoFX29EStxyNDm/i4LyS1EwN19Fze2QBXdnDi9Awdp9GlHolx+HomDj3Yi+Od7dFwQ3xoPAbWb2ZHM3yYPw7IZcIpuja9YNuUs6h9XmoqDl/bn2UQmE4NtBaKJK3CyGrOo1KC9jvZltW0ee/CmI4XUBigHNT4SGmZ0P/N7Mrnf3GVa9NMWmwPVI4L8CPEGmm9dclC28QXmLiHrhdGSFzkWT6AeodMhYFJwd73VPuuyLrMJdEEUXKM2EHN07ZYaZPQ20cfejcizbEVG9DnT3Uc1Z2ze1/rsF0dlAwuo1dH0+QC/VnGKZvMVAyj87FU3qt+ZwST2MXHon1ST0U8yQIai93zdSyxLXVzfUN/V77v5ssl1Y7UjE2rkFuD0SBIoPU4vMLYER7r44zzqtERtnCxSAn5BnvU0QvfZJd78p/NYGWQ+bA1sjGvBThd7LqOmXCSktrCea0RMtEGBdEO7bIY1hW+Tjb84YjBgnb6Lr9Xd3vzdZGIRbm9T3igdvySTOPIKCcuvM7AWUSt8X1V06DAnp2rT8hHO/HFiSEvQtUm6sExEb5wUAM9sJZQOPNbOViFXy9yjwSwN3H5Hr9xDY3QG5IHsjV9wu6H70z6XMBQvxNuBHwW00BrnjdkZKQlck/PcHhhWiEEZNv0xIMUvuRzfxe2nanakt2kPIX32Cx0bpVTCztolGb2Z7o0SWnmHxbNRha0ilxpcgpZ13QqUXrkDc61nIDz8Dcbn/U9tLmxLyX0IungnAb1y1d7ZCJbmvR/GgaUhh6IYax59sKgfwbPh+a9hnSTTH5owsl1pr5PK5kIyQX4UqaCaJmbdnKyep5+bXwI9SixYD41Gm+TvAaHcfXfCYo9AvD1Km/0Eo8NgJ+aPXIf7tgYhi9313f7xyI22cCOyTq5B/25A/uiUSYKCaNg80Eo0fADPrDHweuajmoJd2g9rqteyjA4pZnIwmj6T8bmtgCvL/Lg7/z0R+5JfCtl9G7ROnIHfYpsjiqKY5untBmmOEYGbHomKKzwKDgKFeh0JpKYXwDHRfXkKVNosawK06XrzP5YepRvq1qNnCKhTtHwI8ks9kbO4ws++g+vQ3oL4DU8Lv2wK3AYe4+7aVG2EGQVAfjF7cnPXu6ypkzWwb1F3tfTJdwyYjq2EeooLOR7kKicaZaI43IaGfYAkS/kXVHJs7UgrdE8iCOtlTPSBM5TkMxWVzMszqYPm1QrGBTq5yI9G9szHBzNq4Wq31QHGVhe7+cVjWaDTVxgQzG41472d4pvRCoiH1Q6ZzH3efWclxJjCzmcBF7v7v8MK2RBmz6+p7f01p/ouRcF/iNWTTmpK3WmRpji8jC6AkmmNzR+o5vAO5zE73kGWfw5VTW0ntlsgttCWKCeyCCujtiqy0tu7eqZDxxkBumRECkJeY2Sh3fz5rWcvwN3Y0Cki9OFsBr3oq4zZ1jQ5Bha4q7pu2TILYRODzZjY0BG3XptZpgxKllrj71Nomeg9lkbO23wFZin1R+eU7kvXCNTJ3f6iGcRZNc4yoKn/xF1QW5GozexRYHFx8/ZDQ/hyy1k5KPSeY6jVNcVVFfR4Ff7shosISZN2NQfWTJlmBSYhR6JcRqeDcGcDrZvZiMAtbufvaQm5kU0SYIBMN+T7gSDMb7u5Dw/LNkPA8FzWQ7g/Mq7C1lBz3ZETbvNfUML070tr6ogmsCwrEXl+XnQZBfyJiAPVF/nhD3bTaIg0TQtG1RIDXRXNEgj8K/AYiedbc/T0zS3JHziNcW+TCnQ1MQo1tshMJ30TB30EoGfF9RE0ej6y75SiR87NiyIgo9CuD24EvoBTtB4Nm1hLRs3ZFtL6J7n5bc9b6gyBKtPd/orLFg8xsMnoJtkbtCVuhsgQnIC5/Y8BBSPAvRuU25iHtfwh6+ecj33qtpbTDs3E+8DOk9Y1AlRenhs9s5ONPMnHLqjlGVMNuiCE1FCXFjUb3ZhGaoHNZo6+T6QGRZOg/ku+5MLOO7v5xQ62z6NOvAEw1Y+5FQmocyjDtiDSylUgzeMbL3Ou1scHMdkaacDsy16gH0qanIUbMBCTA5gIfuPv0yoxWSAVRd0RZuR8iAT0bTUyL6+tbN7OeqJbPCJSZu7yWoN8K4EJ3H2RmP0VCqGSaY0S1+/4ask6/6+7jc6yXy89/D4pHHW5mv0EN7w8I1N9dUPOlfcL/e5Kh5TZoso6afmXQA9VSeRdRNt9GWuA0RMubi2b85o5VKBnlo/CZhAT8BGBm0GQbFVLCeAqyTD7OJeQT11UdrbilqNDahe6+rA7rl1VzjAAyyXSfIVbUeNiQlZPnHjwPHG9mN6BYza5mNh8pOo6UhWnIangSGBb21aAJO2r6FYKZHYdM7dnAooS9E1EdZtbFUw2js5btDKxw9RU9FU2g/6mJ3VJOmNkMVIDrnvBTC2B9fV7WlAb5FvCsu98Qfm+BBE17xL/vhiiiy8qtOUZskEx3PDDS3e8Oyzqjwmy7I9ftZHe/MsX66YRyUM5EFlgf4M+IVjsdTdyLUbb+p4Xeo6jpVw6rgTXuPi35IWh/LcgU6Ko4G6XSSAR+igvdHrlNzkfBy58ijXoT1KZwEfBKJYO5KeE5DgVLN3PVZkk3L98U6AUsqyN75ipUhfU+1EmrNRIkA8N+eiPXzz2oLEPZNMeIatfuZdSQ/remnrlbooAuSMlbipLoqrq+BYv1GlNGbjvkynvA3QeXYqxR0y8zUhrBCOSmONPdV0XTum4w1de/APgjKiq2wN2/ZmZdEcNnrLtfUckAeGqCGojaFc5G3ay2QiyeXigI3Q+5Xa7Jp2WnNP2TUDA7wScoQPghCgh/BPzP3ReWW3OMyMCUTPcRyhtZiRLppiC37RwUU1nk7gtr2MfmKG5TEqUlavqVw82oPdqxyLx2U6W97qiuzEAU9BtUYQpio4GZ7YK0qL+5+y/NbAlwWaAzzkOa6+Fh9UpOoMmxW6HkqINQ8bjpSFgvRwLgBRRgrUnLThppL0da/PMotjEtHdMIz86W4euKMJGURXOMyMDd55nZKWhSnYve4Xq5G919aUkGFxA1/QrBVGDtVlSbZRzSCjZBJnh7xO/9wN2Pa+5WQErbPQDVNTnQVUlyL+Tq2NHdZ5vZFSgbcs+KDjggsG4eQ7zrJFA/kQydsiH7bIEmkt6I3rsr6jHQB9ja3TfNsU1JNceImhGUkl7oPu2M6Nq/c/dXK6HQRU2/sjgQ+Wb3QQJhLvKvTkSNGMZD7GyUOv+x6Hr1Cv9/gLjnHU1NaL4JvNRYLKNAHz0g1zIza4eCumtyLc+zzXMoRtANXYelyGJoi8px/zDXuZdac4yoDjPri+jYPdBk3I3QqB7dq7lk2FVlf7ej0K8Q3H2JmT2I/PofIQriksqOqnEjMFP+DdxpZte5+yNmthy4GL1YmyE3RsUFfhohQL8bcCpK1NoeBZyfNrNBXkvRs5Qgn4bcQcORr3gOorW2Ac4O+x4MjGgsE19zQuqa74MKKo5FE/P7yKf/CQqs70FIpqvIOJu5EtmoEMoK9Eda2xh3H1fhITUapFw83VHP2K8gN9inyN89AbjBUw2lGwvMbE9UTrsXEsoT0D0+EU34P3T3t2sT1KZ6OetyWX7BXfgE8La7X9ucM7krDTPrguI441F3t8VZy68Djnb3gyvhuo1CvxHAVPf8OmAvZO6tQBrCIODXMXGmOgJtsz/SmLdEjIh3UfCsB3quJzYGbdfUzOQRxNH+Dgo2rwiMre3QPV7s7l8tRFAHd9GLSFk4vzGce0RumNl+hAQ6d19R7uNH906FEZK07kXm+kXhb1vEQjkr/H8lGRZHs0YIirUkcJ3RNemL6twkpvMEZGJXDKlJul8Yy4Xu/kp6HVcz87+jLluQKdZW12NsjSa+PsAZKOnqjrDvKPAbEYKLry0ialyG2GZbIQWvrIhCv4IIfOpzkJA6CliZ0uafNLNZqOXeldRTIDQlhCDtUUio74a0/C2Qu6QN0vTnI+67EVrOVVLwpe7jXDTWSVA1aa0LuRptgP2AWWGSqK3wWhukGHRDtN6k+1Vb5OK62N3/UYrziagfwjO7Bbo/26KJeRfk9ukLXOHuUysxtij0K4v1iL71PU81USE0wUBJN20t1Eap4DgrjZYoc3QRsoQ+QIlHzwK3oCbyhhgRF6DM1dE1JcCUC+4+M5RjuMzMrvfqnbTOBr4GXFoX152r8c630Xu7AAUIhyGX0dRKuAoi8uIbKBenJRL+n6AkrWEoyPtSpQYWhX6FEDS7j81sLaqdAlRpp+tD0O6HiMO/CSrL2izh6kLUF3UYW2JmJyAK3OVevarmEjO7HdUk/xzwnwqXY0iO/XMUfN7XVG99S1RAbRNUc/+ROuwrcRcdgTj3q0s38oiGIpVZPR14CpX6HonYOstDpnZLoI+ZbebuQ8odr4tCv3JIfPRPANcF5s59qCLiXsiffyiyAuY190Cuu08IVhAoKzfhqGejdfjbsywDqwGeaa4xyMwmoTyCbmjcjwNPJXTN2u5vssxTvVcjGh+SzGp3f43Q2yEwzo4C9gxZ5f1Q3ai1QMdyv9dR6FcOifb5C3Tzfwr8FvGulyOO7/eBf0FM0Apoga7bFOQW2QUYHvynSQvBm5Gr5y1oPAFNd38bldCugpm1NrPW7v5ZvL9ND4Fldiea7FujGkzjkKL3PqLrln9c8VmrHFLc8/aoYuKWSEPtiIK7b8dsygxShcwGIMbTbsCjSHPeAbl01gJXufu/KzfS/DCzPVAS1V7hp0XA/1D9pWbPzmpKCPGX36NyK/eiBMyKC9wo9BsRzOws4DhUdK01oneNQqyMinaEamwws52QBnUKmixnoLK2TwHvNhYNPw0zOw3590GU0/WI2dEXUS2vrE9ZhojGCctU0n0E0TJPdffFwZcPskQdleEouwCOQr8RILgnHkZZpm+i/pqL0MPxfUT7O9rdF1VskI0QwcfvjUF7qg1Bw38DTUq3A+M9dMEys8uAG4Aj6pKZG9G4kRL6v0I5GqdmZ+Wm1m3tZe6bEYV+I4CZnYnqw5/t7o9lLTsE+fXPdvenokDYuJBySd2L3E9HuPucsCxx73VAk/0gd78t3uONG6n72hM4DwXvL0J5JH1Qnkk/VD79Y3c/oZxEjRjIrSBS9K7jgSGoVjpm1hr1T12FimuNQuUFoLJ14iPqjyT4vA6Yivj12Rre7ij/4BVoPMHniIYhJbznoXLKRwGno85Zn6A+usvQszAta5uSIwr9yiKhbU4HtiHQDYMwSATCiUjQv1CJAUYUjIRm+nfU5vFC4I5E4If8g0sQPfcEYGTU9JsG3H11qK/0HiIbjECJhZNRBvnKStzn6N5pBAgVEq9HAdyLkfDvi5oon4pSud9w95OaO19/Y0Wok3MhKqvxLqEPALrPrRBFd4G7H553JxEbHUwNbFaW229fE6KmX0GYWWdUP2YAStU+Cmn9n6LkozYooPsmmWbKUeBvJAhZ1X9Fk/nWSOufju75dOS2ewppfrMQjzuiCaExUq6jpl9BmNmHqCb8WuTjW0impsqHqEjXfOCTaO5vnDA1fQEJ9pnIrz8JCfhF8b5GlBtR6FcQZnYJomOOQUJgeWMyAyMKR2ioscLdP82xbCfE4x6a/BQngYhSIwr9iIgyIdRUN0IVVTP7EWqsskec7CPKhRa1rxIREVEMuJDQN0FuvK6kqqxGRJQaUehHRJQZqWD8cBTA37aCw4loZohCPyKiAghc/MWo4mLC2bfKjiqiOSD69CMiyozs5KvQS2ETd4+UzYiSIwr9iIgyIquc9pmoXea2KGdmDvCAuz9TyTFGNG1E905ERBkRBP72wNOowUZfVKNlCkraut/MzoaqKqIREUVF1PQjIsqI4Lf/Gep/fJK7v5patgVwI6rEuVMsuRFRCkShHxFRRgS3zgjgMXe/LvzWAmjl7mvMbB/U6nFrd19ewaFGNFFE8zEiokwImvunqCPamISt4+7rUx2zDkKlN7ap0DAjmjii0I+IKB+S920waqrRG8DMOprZTmb2LeAm4L+oPk9ERNER3TsREWVCqovW/qhh9p7IlTMbafa9UZP0S2NrzIhSIQr9iIgKIDB4jgOOQGUY5gEvIqG/JbDQ3adXboQRTRVR6EdEVABmtimqs/9RNkPHzG4CBrj7iRUZXESTRvTpR0RUBpcDpyUC38xamlnLsOxF4GAz616x0UU0WUShHxFRRqTq65wArE8EvbuvQw3UAT4CVgN7hG3iexpRNMSHKSKivEi0+YWorHKbZEHKzdMbFWFrmywq2+gimjyi0I+IKC8SAf4CcDTwRQAza2dmncxsX+B21DbzFYh9kSOKi9gYPSKivEhcOH8DdgWeMrOnUE/kXsAhwErgvMbYVDti40dk70REVAhm1hk4HTgL8fSnIcrmo+7+fiXHFtF0EYV+RESFEQurRZQTUehHRERENCPEQG5EREREM0IU+hERERHNCFHoR0RERDQjRKEfERER0YwQhX5EREREM0IU+hERERHNCP8PqmoN+DYWhuEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib import pyplot\n",
    "label=[]\n",
    "label=(dictionary.keys())\n",
    "values=(dictionary.values())\n",
    "index = np.arange(len(label))\n",
    "plt.bar(index, values)\n",
    "plt.xticks(index, label, fontsize=15, rotation=100)\n",
    "plt.title('Feature Importance')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# XGBClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_train_2=Train2[\"m13\"]\n",
    "\n",
    "X_train_2=Train2.drop(['m13','loan_id'], axis=1)\n",
    "\n",
    "Y_test_2=Test2[\"m13\"]\n",
    "\n",
    "X_test_2=Test2.drop(['m13','loan_id'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "sc_X = StandardScaler()\n",
    "X_train = sc_X.fit_transform(X_train_2)\n",
    "X_test = sc_X.transform(X_test_2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 520,
   "metadata": {},
   "outputs": [],
   "source": [
    "param_test1 = {\n",
    "'n_estimators':[10,30,50,70,90],\n",
    " 'max_depth':[10,30,50,70,90]\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 521,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=10, error_score='raise-deprecating',\n",
       "       estimator=XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "       colsample_bynode=1, colsample_bytree=0.8, gamma=0,\n",
       "       learning_rate=0.1, max_delta_step=0, max_depth=5,\n",
       "       min_child_weight=1, missing=None, n_estimators=140, n_jobs=1,\n",
       "       nthread=4, objective='binary:logistic', random_state=0, reg_alpha=0,\n",
       "       reg_lambda=1, scale_pos_weight=1, seed=27, silent=None,\n",
       "       subsample=0.8, verbosity=1),\n",
       "       fit_params=None, iid=False, n_jobs=4,\n",
       "       param_grid={'n_estimators': [10, 30, 50, 70, 90], 'max_depth': [10, 30, 50, 70, 90]},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring='f1', verbose=0)"
      ]
     },
     "execution_count": 521,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from xgboost import XGBClassifier\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "gsearch1 = GridSearchCV(estimator = XGBClassifier( learning_rate =0.1, n_estimators=140, max_depth=5,\n",
    "min_child_weight=1, gamma=0, subsample=0.8, colsample_bytree=0.8,\n",
    "objective= 'binary:logistic', nthread=4, scale_pos_weight=1, seed=27), \n",
    "param_grid = param_test1, scoring='f1',n_jobs=4,iid=False, cv=10)\n",
    "\n",
    "gsearch1.fit(X_train, Y_train_2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 522,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({'max_depth': 30, 'n_estimators': 90}, 0.8931378983039616)"
      ]
     },
     "execution_count": 522,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gsearch1.best_params_, gsearch1.best_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "       colsample_bynode=1, colsample_bytree=1, gamma=0, learning_rate=0.1,\n",
       "       max_delta_step=0, max_depth=30, min_child_weight=1, missing=None,\n",
       "       n_estimators=90, n_jobs=1, nthread=None,\n",
       "       objective='binary:logistic', random_state=0, reg_alpha=0,\n",
       "       reg_lambda=1, scale_pos_weight=1, seed=None, silent=None,\n",
       "       subsample=1, verbosity=1)"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from xgboost import XGBClassifier\n",
    "XB_model = XGBClassifier(max_depth=30,n_estimators=90)\n",
    "XB_model.fit(X_train, Y_train_2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Macro-average quality numbers\n",
      "Precision: 0.8411, Recall: 0.7559, F1-measure: 0.7491\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.70      0.99      0.82       150\n",
      "           1       0.99      0.52      0.68       135\n",
      "\n",
      "   micro avg       0.77      0.77      0.77       285\n",
      "   macro avg       0.84      0.76      0.75       285\n",
      "weighted avg       0.83      0.77      0.75       285\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import precision_score,recall_score,f1_score\n",
    "# predict the response\n",
    "XB = XB_model.predict(X_test)\n",
    "\n",
    "precision = precision_score(Y_test_2, XB, average='macro')\n",
    "recall = recall_score(Y_test_2, XB, average='macro')\n",
    "f1 = f1_score(Y_test_2, XB, average='macro')\n",
    " \n",
    "print(\"Macro-average quality numbers\")\n",
    "print(\"Precision: {:.4f}, Recall: {:.4f}, F1-measure: {:.4f}\".format(precision, recall, f1))\n",
    "\n",
    "print (metrics.classification_report(Y_test_2, XB))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[149   1]\n",
      " [ 65  70]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.70      0.99      0.82       150\n",
      "           1       0.99      0.52      0.68       135\n",
      "\n",
      "   micro avg       0.77      0.77      0.77       285\n",
      "   macro avg       0.84      0.76      0.75       285\n",
      "weighted avg       0.83      0.77      0.75       285\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix,classification_report\n",
    "from pandas import DataFrame\n",
    "from sklearn.metrics import f1_score\n",
    "import seaborn as sn\n",
    "\n",
    "confm = confusion_matrix(Y_test_2, XB)\n",
    "labels=['positive','negative']\n",
    "print(confm)\n",
    "df_cm = DataFrame(confm, index=labels, columns=labels)\n",
    "\n",
    "ax = sn.heatmap(df_cm, cmap='Oranges', annot=True,fmt='g')\n",
    "\n",
    "f1_score(Y_test_2, XB, average='macro') \n",
    "print(classification_report(Y_test_2, XB))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 526,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'m12': 0.6124373, 'm9': 0.11703211, 'm11': 0.02572241, 'm7': 0.022184923, 'm10': 0.020762522, 'm4': 0.018118814, 'm5': 0.016660081, 'm6': 0.016355699, 'm8': 0.012365744, 'm2': 0.011624595, 'loan_term': 0.010765451, 'm1': 0.010474211, 'borrower_credit_score': 0.00972593, 'co-borrower_credit_score': 0.009456412, 'financial_institution': 0.00879731, 'interest_rate': 0.008299901, 'origination_date': 0.008167644, 'loan_to_value': 0.0076872716, 'unpaid_principal_bal': 0.0076339324, 'source': 0.0075727324}\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "imp_features=(dict(zip(features.columns, XB_model.feature_importances_)))\n",
    "dictionary=dict(sorted(imp_features.items(), key=lambda kv: kv[1],reverse=True)[:20])\n",
    "print(dictionary)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 527,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib import pyplot\n",
    "label=[]\n",
    "label=(dictionary.keys())\n",
    "values=(dictionary.values())\n",
    "index = np.arange(len(label))\n",
    "plt.bar(index, values)\n",
    "plt.xticks(index, label, fontsize=10, rotation=100)\n",
    "plt.title('Feature Importance')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# GradientBoostingClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_train_2=Train2[\"m13\"]\n",
    "\n",
    "X_train_2=Train2.drop(['m13','loan_id'], axis=1)\n",
    "\n",
    "Y_test_2=Test2[\"m13\"]\n",
    "\n",
    "X_test_2=Test2.drop(['m13','loan_id'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "sc_X = StandardScaler()\n",
    "X_train = sc_X.fit_transform(X_train_2)\n",
    "X_test = sc_X.transform(X_test_2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import GradientBoostingClassifier  #GBM algorithm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 622,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise-deprecating',\n",
       "       estimator=GradientBoostingClassifier(criterion='friedman_mse', init=None,\n",
       "              learning_rate=0.1, loss='deviance', max_depth=8,\n",
       "              max_features=None, max_leaf_nodes=None,\n",
       "              min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "              min_samples_leaf=50, min_samp...      subsample=0.8, tol=0.0001, validation_fraction=0.1,\n",
       "              verbose=0, warm_start=False),\n",
       "       fit_params=None, iid='warn', n_jobs=4,\n",
       "       param_grid={'n_estimators': range(20, 200, 10)},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring='f1', verbose=0)"
      ]
     },
     "execution_count": 622,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Choose all predictors except target & IDcols\n",
    "param_grid = {'n_estimators':range(20,200,10)}\n",
    "LR=GradientBoostingClassifier(learning_rate=0.1, min_samples_split=500,min_samples_leaf=50,max_depth=8,subsample=0.8,random_state=10)\n",
    "gsearch1 = GridSearchCV(LR, param_grid, cv=5,scoring='f1',n_jobs=4)\n",
    "gsearch1.fit(X_train, Y_train_2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 623,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({'n_estimators': 190}, 0.798756937052019)"
      ]
     },
     "execution_count": 623,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gsearch1.best_params_, gsearch1.best_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 625,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "({'max_depth': 11, 'min_samples_split': 200, 'n_estimators': 190},\n",
       " 0.8767571849783816)"
      ]
     },
     "execution_count": 625,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "param_grid = {'max_depth':range(5,16,2), 'min_samples_split':range(200,1001,200),'n_estimators':range(20,200,10)}\n",
    "LR=(GradientBoostingClassifier(n_estimators=190,learning_rate=0.1, min_samples_split=200,min_samples_leaf=30,max_depth=15,subsample=0.8,random_state=10))\n",
    "gsearch2 = GridSearchCV(LR, param_grid, cv=5,scoring='f1',n_jobs=4)\n",
    "gsearch2.fit(X_train, Y_train_2)\n",
    "gsearch2.best_params_, gsearch2.best_score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GradientBoostingClassifier(criterion='friedman_mse', init=None,\n",
       "              learning_rate=0.05, loss='deviance', max_depth=11,\n",
       "              max_features=None, max_leaf_nodes=None,\n",
       "              min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "              min_samples_leaf=40, min_samples_split=200,\n",
       "              min_weight_fraction_leaf=0.0, n_estimators=190,\n",
       "              n_iter_no_change=None, presort='auto', random_state=10,\n",
       "              subsample=0.8, tol=0.0001, validation_fraction=0.1,\n",
       "              verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "GB=(GradientBoostingClassifier(n_estimators=190,learning_rate=0.05, min_samples_split=200,min_samples_leaf=40,max_depth=11,subsample=0.8,random_state=10))\n",
    "GB.fit(X_train,Y_train_2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Macro-average quality numbers\n",
      "Precision: 0.8456, Recall: 0.7519, F1-measure: 0.7437\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.69      1.00      0.82       150\n",
      "           1       1.00      0.50      0.67       135\n",
      "\n",
      "   micro avg       0.76      0.76      0.76       285\n",
      "   macro avg       0.85      0.75      0.74       285\n",
      "weighted avg       0.84      0.76      0.75       285\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import precision_score,recall_score,f1_score\n",
    "# predict the response\n",
    "gb = GB.predict(X_test_2)\n",
    "\n",
    "precision = precision_score(Y_test_2, gb, average='macro')\n",
    "recall = recall_score(Y_test_2, gb, average='macro')\n",
    "f1 = f1_score(Y_test_2, gb, average='macro')\n",
    " \n",
    "print(\"Macro-average quality numbers\")\n",
    "print(\"Precision: {:.4f}, Recall: {:.4f}, F1-measure: {:.4f}\".format(precision, recall, f1))\n",
    "\n",
    "print (metrics.classification_report(Y_test_2, gb))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[150   0]\n",
      " [ 67  68]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.69      1.00      0.82       150\n",
      "           1       1.00      0.50      0.67       135\n",
      "\n",
      "   micro avg       0.76      0.76      0.76       285\n",
      "   macro avg       0.85      0.75      0.74       285\n",
      "weighted avg       0.84      0.76      0.75       285\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix,classification_report\n",
    "from pandas import DataFrame\n",
    "from sklearn.metrics import f1_score\n",
    "import seaborn as sn\n",
    "\n",
    "confm = confusion_matrix(Y_test_2, gb)\n",
    "labels=['positive','negative']\n",
    "print(confm)\n",
    "df_cm = DataFrame(confm, index=labels, columns=labels)\n",
    "\n",
    "ax = sn.heatmap(df_cm, cmap='Oranges', annot=True,fmt='g')\n",
    "\n",
    "f1_score(Y_test_2, gb, average='macro') \n",
    "print(classification_report(Y_test_2, gb))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Test Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [],
   "source": [
    "test=pd.read_csv(\"test.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [],
   "source": [
    "list1=[]\n",
    "for i in range(0,len(test)):\n",
    "    string1=test[\"origination_date\"][i]\n",
    "    list1.append(string1[4])\n",
    "\n",
    "test[\"origination_date\"]=list1\n",
    "\n",
    "list1=[]\n",
    "for i in range(0,len(test)):\n",
    "    string1=test[\"first_payment_date\"][i]\n",
    "    if(string1[0:3]==\"Feb\"):\n",
    "        list1.append(2)\n",
    "    elif(string1[0:3]==\"Apr\"):\n",
    "        list1.append(4)\n",
    "    elif(string1[0:3]==\"Mar\"):\n",
    "        list1.append(3)  \n",
    "    elif(string1[0:3]==\"May\"):\n",
    "        list1.append(5)\n",
    "\n",
    "test[\"first_payment_date\"]=list1\n",
    "\n",
    "test['origination_date'] = test['origination_date'].astype(int)\n",
    "test['first_payment_date'] = test['first_payment_date'].astype(int)\n",
    "\n",
    "test[\"orgin-first\"]=test[\"first_payment_date\"]-test[\"origination_date\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [],
   "source": [
    "le_X = LabelEncoder()\n",
    "\n",
    "X = test.iloc[:,1].values\n",
    "\n",
    "X = le_X.fit_transform(X)\n",
    "\n",
    "y = test.iloc[:,2].values\n",
    "\n",
    "y = le_X.fit_transform(y)\n",
    "\n",
    "z = test.iloc[:,12].values\n",
    "\n",
    "z = le_X.fit_transform(z)\n",
    "\n",
    "a = test.iloc[:,15].values\n",
    "\n",
    "a = le_X.fit_transform(z)\n",
    "\n",
    "test['source']=X\n",
    "\n",
    "test['financial_institution']=y\n",
    "\n",
    "test['loan_purpose']=z\n",
    "\n",
    "test['insurance_type']=a"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test1=test.drop(['loan_id'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test1 = sc_X.transform(X_test1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred_LR = alpha_optimal.predict(X_test1)\n",
    "\n",
    "pred_SVM = alpha_optimal_svm.predict(X_test1)\n",
    "\n",
    "pred_RF = model_rf.predict(X_test1)\n",
    "\n",
    "pred_XG = XB_model.predict(X_test1)\n",
    "\n",
    "pred_GB = GB.predict(X_test1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_all = pd.DataFrame(columns=['pred_SVM','pred_GB','pred_RF'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_all['pred_GB']=pred_GB\n",
    "df_all['pred_RF']=pred_RF\n",
    "df_all['pred_SVM']=pred_SVM"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Majority Voting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_all[\"final\"]=df_all.mode(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [],
   "source": [
    "df1 = pd.DataFrame(columns=['loan_id','m13'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [],
   "source": [
    "df1[\"m13\"]=df_all[\"final\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "df1[\"loan_id\"]=test['loan_id']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [],
   "source": [
    "df1.to_csv(\"final.csv\",index=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    35641\n",
       "1      225\n",
       "Name: m13, dtype: int64"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1[\"m13\"].value_counts()"
   ]
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}