CHOOSE_MODEL.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Laboratory of Computational Physics mod. B at Phisics of Data, University of Padova\n",
    "\n",
    "__Authors:__\n",
    "\n",
    "- Valeria Fioroni (University of Padova)\n",
    "- Matteo Guida (University of Padova)\n",
    "- Philipp Zehetner (University of Padova, Ludwig Maximilian University of Munich)\n",
    "\n",
    "__Supervised by:__\n",
    "\n",
    "- Professor Marco Zanetti (University of Padova, CERN)\n",
    "- Professor Stefano Lacaprara (University of Padova, BELLE2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. The Signal\n",
    "\n",
    "At the SuperKEKB particle accelerator electrons and positrons collide with a center-of-momentum energy close to the $ \\Upsilon (4S) $ resonance. The cross section for the event $ e^+ e^- \\rightarrow \\Upsilon (4S) $ is \n",
    "\n",
    "$$ \n",
    "\\sigma_{\\Upsilon (4S)} = 1.1 \\, nb\n",
    "$$\n",
    "\n",
    "The $ \\Upsilon (4S) $ decays in one of two (almost) equally probable products: \n",
    "- $ \\Upsilon (4S) \\rightarrow B^+ B^- $\n",
    "- $ \\Upsilon (4S) \\rightarrow B^0 \\bar{B^0} $\n",
    "\n",
    "For this analysis we are interested in the decay channel\n",
    "$$\n",
    "B^0 \\rightarrow \\eta' \\left( \\eta \\left( \\gamma \\gamma \\right) \\pi^+ \\pi^- \\right) K^0_S \\left( \\pi^+ \\pi^- \\right)\n",
    "$$\n",
    "\n",
    "The branching fractions are as follows: \n",
    "\n",
    "- $ \\mathcal{BF} \\left( B^0 \\rightarrow \\eta' K^0 \\right) = 6.6 * 10^{-5} $ \n",
    "- $ \\mathcal{BF} \\left( \\eta' \\rightarrow \\eta \\pi^+ \\pi^- \\right) = 0.43 $\n",
    "- $ \\mathcal{BF} \\left( \\eta \\rightarrow \\gamma \\gamma \\right) = 0.40 $\n",
    "- $ P \\left( K^0_S | K^0 \\right) = 0.5 $\n",
    "- $ \\mathcal{BF} \\left( K^0_S \\rightarrow \\pi^+ \\pi^- \\right) = 0.7 $\n",
    "\n",
    "This leads to a complete branching ratio of \n",
    "$ \\mathcal{BF} = 3.97 * 10^{-6} $ "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Our Data\n",
    "\n",
    "Our Data consists of labelled data, that we used to train and test a _Deep Neural Network_ (DNN) and a _Boosted Decision Tree_ as well as unlabelled data (Data Challenge) in which we are searching for our signal. \n",
    "\n",
    "###  Labelled Data\n",
    "\n",
    "The labelled data consists of various types of background, all corresponding to an integrated luminosity of \n",
    "$ \\mathcal{L} = 800 \\, fb^{-1} $\n",
    "and signal that was reconstructed from $ 1595000 $ signal events. \n",
    "\n",
    "The reconstruction efficiency corresponds to $ \\varepsilon = 16.2 \\% $\n",
    "\n",
    "\n",
    "The background types are as follows: \n",
    "\n",
    "- $ e^+ e^- \\rightarrow u \\bar{u} $\n",
    "- $ e^+ e^- \\rightarrow d \\bar{d} $\n",
    "- $ e^+ e^- \\rightarrow s \\bar{s} $\n",
    "- $ e^+ e^- \\rightarrow c \\bar{c} $\n",
    "- $ e^+ e^- \\rightarrow \\tau \\bar{\\tau} $\n",
    "- $ e^+ e^- \\rightarrow B^+ B^- $\n",
    "- $ e^+ e^- \\rightarrow B^0 \\bar{B^0} $\n",
    "\n",
    "We summarize the first five types of backgrounds as _Continuum Background_ and the last two types as _Peaking Background_ and thus build a multi classifier with three types of events. \n",
    "\n",
    "Using the complete branching fraction \n",
    "$ \\mathcal{BF} $\n",
    "and the cross section \n",
    "$ \\sigma_{\\Upsilon (4S)} $ \n",
    "we can calculate to expect to have $N$ events with \n",
    "\\begin{equation}\n",
    "N = \\mathcal{L} \\times \\sigma_{\\Upsilon(4S)} \\times \\mathcal{BF} \n",
    "= 800 \\, fb^{-1} \\times 1.1 \\, nb \\times 3.97 \\cdot 10^{-6}\n",
    "= 3494\n",
    "\\end{equation}\n",
    "    of which $ \\varepsilon \\times N = 566 $ should be found by the reconstruction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os, sys, pprint\n",
    "import glob\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "import warnings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "Testing = False # Uses only a small subset of the data and trains on fewer epochs. \n",
    "PCA_ON = False\n",
    "cut_5_sigma = False\n",
    "VM = False\n",
    "Local_Matteo = True\n",
    "Local_Philipp = False\n",
    "Local_Valeria = False\n",
    "\n",
    "\n",
    "if Testing: print(\"For testing purpose only a small number of events are used\")\n",
    "if Local_Philipp : print(\"Paths are set to be run on Philipps Computer\")\n",
    "elif VM : print(\"Paths are set to run on the VM\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Reading and Cleaning Data\n",
    "The dataset is candidate based so each entry correspond to one candidate. \n",
    "\n",
    "In this chapter we:\n",
    "- import the root files into pandas dataframes  selecting only the features of interest for our analysis\n",
    "- clean our data sets by removing wrongly reconstructed signal and allowing only one true signal candidate per event.\n",
    "- apply a $5\\sigma$ cut on the mass of the reconstructed $B^0$ on all candidates"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define branches that we need for our analysis\n",
    "Masses = ['B0_M', 'B0_ErrM', 'B0_SigM', 'B0_K_S0_M', 'B0_K_S0_ErrM', 'B0_K_S0_SigM', 'B0_etap_M', 'B0_etap_ErrM', \n",
    "          'B0_etap_SigM', 'B0_etap_eta_M', 'B0_etap_eta_ErrM', 'B0_etap_eta_SigM']\n",
    "Kinetics_CMS = ['B0_Pcms', 'B0_etap_Pcms', 'B0_etap_eta_Pcms', 'B0_etap_eta_gamma0_Pcms', 'B0_etap_eta_gamma1_Pcms',\n",
    "                'B0_etap_pi0_Pcms', 'B0_etap_pi1_Pcms', 'B0_K_S0_Pcms']\n",
    "Other_Kinetics = ['B0_deltae', 'B0_mbc']\n",
    "DecayAngles = ['B0_decayAngle__bo0__bc', 'B0_decayAngle__bo1__bc', 'B0_etap_decayAngle__bo0__bc',\n",
    "               'B0_etap_decayAngle__bo1__bc', 'B0_etap_decayAngle__bo2__bc']\n",
    "Positions = ['B0_X', 'B0_ErrX', 'B0_Y', 'B0_ErrY', 'B0_Z', 'B0_ErrZ', 'B0_Rho',  \n",
    "             'B0_etap_X', 'B0_etap_ErrX', 'B0_etap_Y', 'B0_etap_ErrY', \n",
    "             'B0_etap_Z', 'B0_etap_ErrZ', 'B0_etap_Rho',\n",
    "             'B0_etap_eta_X', 'B0_etap_eta_ErrX', 'B0_etap_eta_Y',\n",
    "             'B0_etap_eta_ErrY', 'B0_etap_eta_Z', 'B0_etap_eta_ErrZ', 'B0_etap_eta_Rho',\n",
    "             'B0_etap_pi0_X', 'B0_etap_pi0_ErrX', 'B0_etap_pi0_Y', 'B0_etap_pi0_ErrY', \n",
    "             'B0_etap_pi0_Z', 'B0_etap_pi0_ErrZ', 'B0_etap_pi0_Rho', \n",
    "             'B0_etap_pi1_X', 'B0_etap_pi1_ErrX', 'B0_etap_pi1_Y', 'B0_etap_pi1_ErrY', \n",
    "             'B0_etap_pi1_Z', 'B0_etap_pi1_ErrZ', 'B0_etap_pi1_Rho', \n",
    "             'B0_K_S0_X', 'B0_K_S0_ErrX', 'B0_K_S0_Y', 'B0_K_S0_ErrY', 'B0_K_S0_Z',\n",
    "             'B0_K_S0_ErrZ', 'B0_K_S0_Rho', \n",
    "             'B0_cosAngleBetweenMomentumAndVertexVector', 'B0_distance', 'B0_significanceOfDistance',\n",
    "             'B0_dr', 'B0_etap_pi0_dr', 'B0_etap_pi1_dr', 'B0_K_S0_dr']\n",
    "Vertex_Training = ['B0_VtxPvalue', 'B0_etap_VtxPvalue', 'B0_etap_eta_VtxPvalue', 'B0_etap_pi0_VtxPvalue',\n",
    "                   'B0_etap_pi1_VtxPvalue', 'B0_K_S0_VtxPvalue', ]\n",
    "Continuum_Suppression_Training = ['B0_TrCSMVA']\n",
    "\n",
    "Training = Kinetics_CMS + Masses + Other_Kinetics + Continuum_Suppression_Training + Positions + DecayAngles + Vertex_Training\n",
    "Important = Training + ['B0_isSignal', 'evt_no']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "def memory(df):\n",
    "    return str((round(df.memory_usage(deep=True).sum() / 1024 ** 2, 2)))+\"  Mb\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Memory used by Signal :  71.3  Mb\n",
      "Memory used by Continuum :  81.75  Mb\n",
      "Memory used by Peaking  :   0.72  Mb \n",
      "\n",
      "Peaking shape :  (2167, 85)\n",
      "Continuum shape :  (246329, 85)\n",
      "Signal shape :  (219897, 85)\n"
     ]
    }
   ],
   "source": [
    "from root_pandas import read_root\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "if Local_Philipp and not VM : \n",
    "    path = '/home/philipp/Desktop/Project/DATA/'\n",
    "    SFiles = glob.glob(os.path.join(path, 'Signal/*.root'))\n",
    "    CFiles = glob.glob(os.path.join(path, 'Continuous/*.root'))\n",
    "    PFiles = glob.glob(os.path.join(path, 'Peaking/*.root'))\n",
    "    \n",
    "elif VM and not Local_Philipp : \n",
    "    path = '/home/Students/DataBelle2_all/'\n",
    "    SFiles = glob.glob(os.path.join(path, 'Signal/*.root'))\n",
    "    CFiles = glob.glob(os.path.join(path, 'Background/Continuum/*.root'))\n",
    "    PFiles = glob.glob(os.path.join(path, 'Background/Peaking/*.root'))\n",
    "    \n",
    "elif Local_Matteo : \n",
    "    path = '/media/guida/Transcend/DataBelle2/'\n",
    "    SFiles = glob.glob(os.path.join(path, 'Signal/*.root'))\n",
    "    CFiles = glob.glob(os.path.join(path, 'Background/Continuum/*.root'))\n",
    "    PFiles = glob.glob(os.path.join(path, 'Background/Peaking/*.root'))\n",
    "    DCFiles = glob.glob(os.path.join(path, 'DataChallenge/*.root'))\n",
    "    \n",
    "elif Local_Valeria:  \n",
    "    path = '/home/utente/Scrivania/Progetto LCP-B/DataBelle2/'\n",
    "    SFiles = glob.glob(os.path.join(path, 'Signal/*.root'))\n",
    "    CFiles = glob.glob(os.path.join(path, 'Background/Continuum/*.root'))\n",
    "    PFiles = glob.glob(os.path.join(path, 'Background/Peaking/*.root'))\n",
    "    \n",
    "\n",
    "Signal = pd.concat((read_root(f, 'B0', columns=Important) for f in SFiles))\n",
    "Signal = Signal[Signal['B0_isSignal']==1].reset_index(drop=True)\n",
    "Continuum_bkg = pd.concat((read_root(f, 'B0', columns=Important) for f in CFiles))\n",
    "Peaking_bkg = pd.concat((read_root(f, 'B0', columns=Important) for f in PFiles))\n",
    "\n",
    "print (\"Memory used by Signal : \", memory(Signal))\n",
    "print (\"Memory used by Continuum : \", memory(Continuum_bkg))\n",
    "print (\"Memory used by Peaking  :  \", memory(Peaking_bkg),\"\\n\")\n",
    "\n",
    "print(\"Peaking shape : \",Peaking_bkg.shape)\n",
    "print(\"Continuum shape : \",Continuum_bkg.shape)\n",
    "print(\"Signal shape : \",Signal.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "labels = 'Peaking bkg', 'Continuum  bkg ', 'Signal'\n",
    "sizes = [Peaking_bkg.shape[0],Continuum_bkg.shape[0],Signal.shape[0]]\n",
    "fig1, ax1 = plt.subplots(1,1, figsize=(8,8))\n",
    "ax1.pie(sizes, labels=labels, autopct='%1.1f%%',\n",
    "        shadow=True, startangle=90, textprops={'fontsize': 18})  \n",
    "ax1.set_title('Dataset composition', fontsize=25)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### $5\\sigma$ Cut on Reconstructed $B^0$ Mass\n",
    "\n",
    "The reconstructed mass of the $B^0$ will differ for various reasons: \n",
    "\n",
    "- If an event is correctly reconstructed detector inaccuracies of all measured decay particles will add up. The estimated error for each candidate due to this can be found in the branch *B0_ErrM* and is part of our data frame. \n",
    "- If the event however is not correctly reconstructed the $B^0$ mass is not necessarily a meaningful quantity an can be far of the desired [PDG](http://pdg.lbl.gov/) value. \n",
    "\n",
    "By applying a generous $5\\sigma$ cut we already lose a significant amount of background."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Masses from PDG\n",
    "if cut_5_sigma == True: \n",
    "    M_B0 = [5.27955, 0.00026]\n",
    "    M_K0 = [0.497611, 0.000013]\n",
    "    M_etap = [0.95778, 0.00006]\n",
    "    M_eta = [0.547862, 0.000017]\n",
    "\n",
    "    #Require the B0 mass to be within 5 sigma from PDG value\n",
    "    Signal = Signal[abs(Signal['B0_M'] - M_B0[0]) < 5 * Signal['B0_ErrM']]\n",
    "    print(\"Signal : \",Signal.shape)\n",
    "\n",
    "    Continuum_bkg = Continuum_bkg[abs(Continuum_bkg['B0_M'] - M_B0[0]) < 5 * Continuum_bkg['B0_ErrM']]\n",
    "    print(\"Continuum : \", Continuum_bkg.shape)\n",
    "\n",
    "    Peaking_bkg = Peaking_bkg[abs(Peaking_bkg['B0_M'] - M_B0[0]) < 5 * Peaking_bkg['B0_ErrM']]\n",
    "    print(\"Peaking : \",Peaking_bkg.shape)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "####  Comment\n",
    "\n",
    "The decision not to use the $5\\sigma$ cut was determined considering that we want to keep as much peaking background data as possible to train the DNN. The 5 sigma cut reduces the peaking background data of 41%. This is not a good trade-off compared to the 48% reduction in continuum background with this cut."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Signal :  (218596, 85)\n",
      "Uniqueness check evt_no :  218596\n",
      "Continuum :  (246329, 85)\n",
      "Signal :  (218596, 85)\n",
      "% of peaking on the entire dataset :  0.464\n"
     ]
    }
   ],
   "source": [
    "#Signal: In case of more than 1 candidate per event select the first one occuring\n",
    "Mask_duplicated = Signal.duplicated(subset='evt_no', keep='first')\n",
    "Mask_duplicated=np.logical_not(Mask_duplicated)\n",
    "\n",
    "Signal=Signal[Mask_duplicated]\n",
    "Signal.reset_index(drop=True, inplace=True)\n",
    "\n",
    "print(\"Signal : \",Signal.shape)\n",
    "print(\"Uniqueness check evt_no : \", Signal['evt_no'].nunique())\n",
    "print(\"Continuum : \",Continuum_bkg.shape)\n",
    "print(\"Signal : \",Signal.shape)\n",
    "print(\"% of peaking on the entire dataset : \",round((Peaking_bkg.shape[0])/(Peaking_bkg.shape[0]+Continuum_bkg.shape[0]+Signal.shape[0])*100,3))\n",
    "\n",
    "Signal=Signal[Training]\n",
    "Continuum_bkg=Continuum_bkg[Training]\n",
    "Peaking_bkg=Peaking_bkg[Training]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Data Preprocessing\n",
    "\n",
    "In this chapter we:\n",
    "- label our data.\n",
    "- turn the label into one-hot vectors as this format is needed by Keras.\n",
    "- split signal, continuum and peaking separatly into training (50%), valdation (30%) and testing (20%) and then remerge  them together to assure that the balance in the composition of the dataset is preserved.\n",
    "- combine and shuffle signal, continuum and peaking for training, validation and testing, making sure that the labels are shuffled with the same permutation as the data\n",
    "- standardize the data (removing mean and scaling the standard deviation to 1)\n",
    "- perform a Principal Component Analysis (PCA) and use the components necessary for 99% of the variance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "from random import seed\n",
    "from random import randint\n",
    "\n",
    "n_seed=1234\n",
    "seed(n_seed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "Signal['Type'] = 2\n",
    "Continuum_bkg['Type'] = 1\n",
    "Peaking_bkg['Type'] = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "from keras.utils import to_categorical\n",
    "\n",
    "#SIGNAL: Train, Validation and test split (50%, 30%, 20%)\n",
    "#Dataset divided into 50% of train data and 50% for validation and test\n",
    "X = Signal.drop('Type',axis=1)\n",
    "Y = Signal['Type']\n",
    "#Y = to_categorical(Y, num_classes=3) # Binarize the labels\n",
    "X_train_S, X_test_val_S, Y_train_S, Y_test_val_S = train_test_split(X, Y, train_size=0.5,random_state=randint(10**6,10**9))\n",
    "#validation and test set divided into 30% for validation and 20% for test\n",
    "X = X_test_val_S\n",
    "Y = Y_test_val_S\n",
    "X_validation_S, X_test_S, Y_validation_S, Y_test_S = train_test_split(X, Y, train_size=0.6,random_state=randint(10**6,10**9))\n",
    "del(X_test_val_S)\n",
    "del(Y_test_val_S)\n",
    "\n",
    "#CONTINUUM BKG: Train, Validation and test split (50%, 30%, 20%)\n",
    "#Dataset divided into 50% of train data and 50% for validation and test\n",
    "X = Continuum_bkg.drop('Type',axis=1)\n",
    "Y = Continuum_bkg['Type']\n",
    "#Y = to_categorical(Y, num_classes=3) # Binarize the labels\n",
    "X_train_C, X_test_val_C, Y_train_C, Y_test_val_C = train_test_split(X, Y, train_size=0.5,random_state=randint(10**6,10**9))\n",
    "#validation and test set divided into 30% for validation and 20% for test\n",
    "X = X_test_val_C\n",
    "Y = Y_test_val_C\n",
    "X_validation_C, X_test_C, Y_validation_C, Y_test_C = train_test_split(X, Y, train_size=0.6,random_state=randint(10**6,10**9))\n",
    "del(X_test_val_C)\n",
    "del(Y_test_val_C)\n",
    "\n",
    "#PEAKING BKG: Train, Validation and test split (50%, 30%, 20%)\n",
    "#Dataset divided into 50% of train data and 50% for validation and test\n",
    "X = Peaking_bkg.drop('Type',axis=1)\n",
    "Y = Peaking_bkg['Type']\n",
    "##Y = to_categorical(Y, num_classes=3) # Binarize the labels\n",
    "X_train_P, X_test_val_P, Y_train_P, Y_test_val_P = train_test_split(X, Y, train_size=0.5,random_state=randint(10**6,10**9))\n",
    "\n",
    "#validation and test set divided into 30% for validation and 20% for test\n",
    "X = X_test_val_P\n",
    "Y = Y_test_val_P\n",
    "X_validation_P, X_test_P, Y_validation_P, Y_test_P = train_test_split(X, Y, train_size=0.6,random_state=randint(10**6,10**9))\n",
    "del(X_test_val_P)\n",
    "del(Y_test_val_P)\n",
    "\n",
    "del(Signal)\n",
    "del(Peaking_bkg)\n",
    "del(Continuum_bkg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X Train shape :  (233545, 83)\n",
      "X Validation shape :  (140127, 83)\n",
      "X Test shape :  (93420, 83)\n"
     ]
    }
   ],
   "source": [
    "#Define Training, Validation and Test datasets\n",
    "X_Train=np.concatenate((X_train_S, X_train_C, X_train_P), axis=0)\n",
    "X_Validation=np.concatenate((X_validation_S, X_validation_C, X_validation_P), axis=0)\n",
    "X_Test=np.concatenate((X_test_S, X_test_C, X_test_P), axis=0)\n",
    "\n",
    "\n",
    "Y_Train=np.concatenate((Y_train_S, Y_train_C, Y_train_P), axis=0)\n",
    "Y_Validation=np.concatenate((Y_validation_S, Y_validation_C, Y_validation_P), axis=0)\n",
    "Y_Test=np.concatenate((Y_test_S, Y_test_C, Y_test_P), axis=0)\n",
    "Ỳ_Test_not_encoded=Y_Test #needed for cut value analysis\n",
    "\n",
    "Y_Train=to_categorical(Y_Train, num_classes=3) # Binarize the labels\n",
    "Y_Validation=to_categorical(Y_Validation, num_classes=3) # Binarize the labels\n",
    "Y_Test=to_categorical(Y_Test, num_classes=3) # Binarize the labels\n",
    "\n",
    "n_classes = Y_Train.shape[1]\n",
    "\n",
    "print(\"X Train shape : \", X_Train.shape)\n",
    "print(\"X Validation shape : \",X_Validation.shape)\n",
    "print(\"X Test shape : \",X_Test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "#shuffle the dataset\n",
    "permutation = np.random.permutation(X_Train.shape[0])\n",
    "X_Train = X_Train[permutation]\n",
    "Y_Train = Y_Train[permutation]\n",
    "\n",
    "permutation = np.random.permutation(X_Validation.shape[0])\n",
    "X_Validation = X_Validation[permutation]\n",
    "Y_Validation = Y_Validation[permutation]\n",
    "\n",
    "permutation = np.random.permutation(X_Test.shape[0])\n",
    "X_Test = X_Test[permutation]\n",
    "Y_Test = Y_Test[permutation]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "# standardize the data\n",
    "ss = StandardScaler()\n",
    "X_Train = ss.fit_transform(X_Train)\n",
    "X_Test = ss.transform(X_Test)\n",
    "X_Validation = ss.transform(X_Validation)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## PCA\n",
    "\n",
    "Principal component analysis is a method often used to reduce the dimensionality of a data set. Orthogonal transformations are performed on the data set to find the directions with the highest variance. The rotated features (principal components) are sorted in descending variance. The assumption made is, that high variance corresponds to more information and little variance can be neglected as noise.\n",
    "\n",
    "In our example we require to keep 99% of the total variance and thus can reduce the number of components from 83 to 50"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "12 variables explain 50% of the variance\n",
      "36 variables explain 90% of the variance\n",
      "42 variables explain 95% of the variance\n",
      "50 variables explain 99% of the variance \n",
      "\n",
      "% of dataset reduction with 99 % explained variance : 39.8\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.decomposition import PCA\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "\n",
    "pca = PCA(n_components=len(Training))\n",
    "data = pca.fit_transform(X_Train)\n",
    "pcTrain = pd.DataFrame(data)\n",
    "pcTest = pd.DataFrame(data = pca.transform(X_Test))\n",
    "pcValidation = pd.DataFrame(data = pca.transform(X_Validation))\n",
    "\n",
    "T50, T90, T95, T99, = False, False, False, False\n",
    "for i in range(len(pca.explained_variance_ratio_)) : \n",
    "    if (sum(pca.explained_variance_ratio_[:i+1]) > 0.5 and T50 ==False) : \n",
    "        print(str(i+1) + ' variables explain 50% of the variance')\n",
    "        T50 = True\n",
    "        n_50 = i + 1 \n",
    "    if (sum(pca.explained_variance_ratio_[:i+1]) > 0.9 and T90 ==False) : \n",
    "        print(str(i+1) + ' variables explain 90% of the variance')\n",
    "        T90 = True\n",
    "        n_90 = i + 1\n",
    "    if (sum(pca.explained_variance_ratio_[:i+1]) > 0.95 and T95 ==False) : \n",
    "        print(str(i+1) + ' variables explain 95% of the variance')\n",
    "        T95 = True\n",
    "        n_95 = i + 1\n",
    "    if (sum(pca.explained_variance_ratio_[:i+1]) > 0.99 and T99 ==False) : \n",
    "        print(str(i+1) + ' variables explain 99% of the variance \\n')\n",
    "        T99 = True\n",
    "        n_99 = i + 1\n",
    "        \n",
    "fig = plt.figure()\n",
    "plt.plot(np.cumsum(pca.explained_variance_ratio_*100))\n",
    "plt.title('PCA on the dataset',fontsize=20)\n",
    "plt.xlabel('Number of Components [#]',fontsize=20)\n",
    "plt.ylabel('Explained Variance [%]',fontsize=20)\n",
    "plt.tick_params(labelsize=14)\n",
    "fig.set_size_inches(10, 8)\n",
    "print('% of dataset reduction with 99 % explained variance :',round((X_Train.shape[1]-n_99)/X_Train.shape[1]*100,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure()\n",
    "Y_Train_single_col=np.where(Y_Train==1)[1]\n",
    "plt.scatter(data[Y_Train_single_col==0,0],data[Y_Train_single_col==0,1], label='Peaking_bkg',edgecolors='black',marker=\"o\",s=100)\n",
    "plt.scatter(data[Y_Train_single_col==1,0],data[Y_Train_single_col==1,1], label=\"Continuum_bkg\", edgecolors='black', marker=\"*\",s=100)\n",
    "plt.scatter(data[Y_Train_single_col==2,0],data[Y_Train_single_col==2,1], label='Signal', edgecolors='black', marker=\"P\",s=80)\n",
    "plt.xlabel('First Principal Component',fontsize=20)\n",
    "plt.ylabel('Second Principal Component',fontsize=20)\n",
    "fig.set_size_inches(10, 8)\n",
    "plt.tick_params(labelsize=14)\n",
    "plt.legend(fontsize=20)\n",
    "plt.show()\n",
    "\n",
    "if not PCA_ON :  \n",
    "    pcTrain = X_Train\n",
    "    pcTest = X_Test\n",
    "    pcValidation = X_Validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input Dimension has not been reduced and no PCA has been performed\n"
     ]
    }
   ],
   "source": [
    "if PCA_ON : \n",
    "    pca = PCA(n_components=n_99)\n",
    "    pcTrain = pca.fit_transform(X_Train)\n",
    "    pcTest = pd.DataFrame(data = pca.transform(X_Test))\n",
    "    pcValidation = pd.DataFrame(data = pca.transform(X_Validation))\n",
    "    print (\"Input Dimension has been reduced to \", n_99)\n",
    "else : \n",
    "    input_dim = 83\n",
    "    print(\"Input Dimension has not been reduced and no PCA has been performed\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "####  Comment\n",
    "The decision not to use the PCA was determined considering that the results of the network are worse in that case, leaving all the other parameters unchanged. \n",
    "The result that gets worse is the true positive rate in the case of peaking background.\n",
    "\n",
    "Peaking backgound accounts for 0.46 % of the total dataset. <br>\n",
    "We believe that PCA embeds the dataset in the space generated by the eigenvectors that maximizes the variance explained for the rest of the dataset, but that penalizes important variables in the distinction of peaking background. <br>\n",
    "In fact $\\sigma \\propto \\sqrt{N} $ and $N_{pkg} \\ll N_{tot} $, $N_{pkg}$ should be very isolated in the phase space so that its characterizing variables are kept in strong consideration. This is not the case with the analyzed dataset."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Create the Neural Network "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "architectures = [ [30, 70, 30], [50, 150, 50], [50, 100, 100, 50] ] \n",
    "activation = ['elu', 'leakyrelu', 'relu']\n",
    "optimizer = ['SGD', 'Adam',' AdaMax']\n",
    "epochs = [15]\n",
    "if Testing : epochs = [1, 2]\n",
    "batch_size = [100, 1000]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout\n",
    "from keras.constraints import maxnorm\n",
    "from keras.wrappers.scikit_learn import KerasClassifier\n",
    "\n",
    "def build_DNN(optimizer='adam', activation='elu'):\n",
    "    model = Sequential()\n",
    "    model.add(Dropout(0.3, input_shape=(pcTrain.shape[1],)))\n",
    "    model.add(Dense(layers[0], input_shape=(pcTrain.shape[1],), activation='elu'))\n",
    "    for i in range(1,len(layers)):\n",
    "        model.add(Dense(layers[i], activation='elu', kernel_constraint=maxnorm(3)))\n",
    "    model.add(Dense(3, activation='softmax'))\n",
    "    model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "    return model\n",
    "\n",
    "def print_results() :\n",
    "    print(\"Layers: \", layers)\n",
    "    print(\"Best: %f using %s\" % (grid_result.best_score_, grid_result.best_params_))\n",
    "    means = grid_result.cv_results_['mean_test_score']\n",
    "    stds = grid_result.cv_results_['std_test_score']\n",
    "    params = grid_result.cv_results_['params']\n",
    "    for mean, stdev, param in zip(means, stds, params): \n",
    "        print(\"%f (%f) with %r\" % (mean, stdev, param))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using architecture:  [30, 70, 30]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0724 08:11:26.321161 140715109656384 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
      "\n",
      "W0724 08:11:26.339491 140715109656384 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n",
      "W0724 08:11:26.343143 140715109656384 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n",
      "\n",
      "W0724 08:11:26.354246 140715109656384 deprecation.py:506] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
      "W0724 08:11:26.374701 140715109656384 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
      "\n",
      "W0724 08:11:26.441535 140715109656384 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n",
      "W0724 08:11:26.469510 140715109656384 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n",
      "\n",
      "W0724 08:11:26.596834 140715109656384 deprecation.py:323] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Layers:  [30, 70, 30]\n",
      "Best: 0.983091 using {'activation': 'leakyrelu', 'batch_size': 100, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.981216 (0.000543) with {'activation': 'elu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.982188 (0.000369) with {'activation': 'elu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.982051 (0.000839) with {'activation': 'elu', 'batch_size': 100, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.981854 (0.000648) with {'activation': 'elu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.982427 (0.000645) with {'activation': 'elu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.982123 (0.000788) with {'activation': 'elu', 'batch_size': 1000, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.982667 (0.000343) with {'activation': 'leakyrelu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.981528 (0.000690) with {'activation': 'leakyrelu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.983091 (0.000752) with {'activation': 'leakyrelu', 'batch_size': 100, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.981721 (0.000783) with {'activation': 'leakyrelu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.981738 (0.000437) with {'activation': 'leakyrelu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.982406 (0.000201) with {'activation': 'leakyrelu', 'batch_size': 1000, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.982239 (0.001489) with {'activation': 'relu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.982183 (0.000597) with {'activation': 'relu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.981824 (0.001131) with {'activation': 'relu', 'batch_size': 100, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.981931 (0.000878) with {'activation': 'relu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.982290 (0.000599) with {'activation': 'relu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.982380 (0.000667) with {'activation': 'relu', 'batch_size': 1000, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "Using architecture:  [50, 150, 50]\n",
      "Layers:  [50, 150, 50]\n",
      "Best: 0.982342 using {'activation': 'leakyrelu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.981503 (0.000360) with {'activation': 'elu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.981271 (0.001369) with {'activation': 'elu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.981383 (0.001593) with {'activation': 'elu', 'batch_size': 100, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.980933 (0.000927) with {'activation': 'elu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.981819 (0.001444) with {'activation': 'elu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.981580 (0.000730) with {'activation': 'elu', 'batch_size': 1000, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.982342 (0.001057) with {'activation': 'leakyrelu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.981815 (0.000781) with {'activation': 'leakyrelu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.980792 (0.001530) with {'activation': 'leakyrelu', 'batch_size': 100, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.981151 (0.000918) with {'activation': 'leakyrelu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.981554 (0.000843) with {'activation': 'leakyrelu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.981866 (0.001335) with {'activation': 'leakyrelu', 'batch_size': 1000, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.981361 (0.000777) with {'activation': 'relu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.980038 (0.001026) with {'activation': 'relu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.981939 (0.001160) with {'activation': 'relu', 'batch_size': 100, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.982106 (0.000843) with {'activation': 'relu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.982179 (0.001052) with {'activation': 'relu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.982260 (0.000503) with {'activation': 'relu', 'batch_size': 1000, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "Using architecture:  [50, 100, 100, 50]\n",
      "Layers:  [50, 100, 100, 50]\n",
      "Best: 0.982967 using {'activation': 'relu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.981524 (0.000727) with {'activation': 'elu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.980942 (0.001245) with {'activation': 'elu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.981156 (0.000409) with {'activation': 'elu', 'batch_size': 100, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.981562 (0.001608) with {'activation': 'elu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.981704 (0.000293) with {'activation': 'elu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.981057 (0.001817) with {'activation': 'elu', 'batch_size': 1000, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.981091 (0.000537) with {'activation': 'leakyrelu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.981083 (0.001431) with {'activation': 'leakyrelu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.980346 (0.001884) with {'activation': 'leakyrelu', 'batch_size': 100, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.982457 (0.000640) with {'activation': 'leakyrelu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.981237 (0.000390) with {'activation': 'leakyrelu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.982706 (0.000723) with {'activation': 'leakyrelu', 'batch_size': 1000, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.982051 (0.000742) with {'activation': 'relu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.980972 (0.001524) with {'activation': 'relu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.980171 (0.000773) with {'activation': 'relu', 'batch_size': 100, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "0.982496 (0.001148) with {'activation': 'relu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.982967 (0.000942) with {'activation': 'relu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'Adam'}\n",
      "0.980826 (0.000787) with {'activation': 'relu', 'batch_size': 1000, 'epochs': 15, 'optimizer': ' AdaMax'}\n",
      "CPU times: user 4h 53min 7s, sys: 43min 2s, total: 5h 36min 10s\n",
      "Wall time: 3h 58min 2s\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n",
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0724 07:50:29.934718 139771027076928 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
      "\n",
      "W0724 07:50:29.954418 139771027076928 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n",
      "W0724 07:50:29.956298 139771027076928 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n",
      "\n",
      "W0724 07:50:29.966318 139771027076928 deprecation.py:506] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
      "W0724 07:50:29.985492 139771027076928 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
      "\n",
      "Using TensorFlow backend.\n",
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0724 07:50:30.020250 139831869536064 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
      "\n",
      "Using TensorFlow backend.\n",
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0724 07:50:30.020455 140565384750912 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
      "\n",
      "W0724 07:50:30.039752 140565384750912 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n",
      "W0724 07:50:30.039756 139831869536064 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n",
      "Using TensorFlow backend.\n",
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0724 07:50:30.039532 140620935583552 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
      "\n",
      "W0724 07:50:30.041350 140565384750912 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n",
      "\n",
      "W0724 07:50:30.041352 139831869536064 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n",
      "\n",
      "W0724 07:50:30.050099 140565384750912 deprecation.py:506] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
      "W0724 07:50:30.050139 139831869536064 deprecation.py:506] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
      "W0724 07:50:30.051524 139771027076928 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n",
      "W0724 07:50:30.058543 140620935583552 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n",
      "W0724 07:50:30.060189 140620935583552 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n",
      "\n",
      "W0724 07:50:30.067825 140565384750912 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
      "\n",
      "W0724 07:50:30.067903 139831869536064 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
      "\n",
      "W0724 07:50:30.069024 140620935583552 deprecation.py:506] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
      "W0724 07:50:30.077690 139771027076928 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n",
      "\n",
      "W0724 07:50:30.087507 140620935583552 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
      "\n",
      "W0724 07:50:30.131822 139831869536064 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n",
      "W0724 07:50:30.132170 140565384750912 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n",
      "Using TensorFlow backend.\n",
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0724 07:50:30.148958 140318003021632 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
      "\n",
      "W0724 07:50:30.150999 140620935583552 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n",
      "W0724 07:50:30.157943 139831869536064 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n",
      "\n",
      "W0724 07:50:30.159691 140565384750912 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n",
      "\n",
      "W0724 07:50:30.167120 140318003021632 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n",
      "W0724 07:50:30.168663 140318003021632 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n",
      "\n",
      "W0724 07:50:30.175985 140620935583552 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n",
      "\n",
      "W0724 07:50:30.177251 140318003021632 deprecation.py:506] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
      "Using TensorFlow backend.\n",
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0724 07:50:30.191097 140261923243840 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
      "\n",
      "W0724 07:50:30.194828 140318003021632 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
      "\n",
      "W0724 07:50:30.209183 140261923243840 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n",
      "W0724 07:50:30.210726 140261923243840 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n",
      "\n",
      "W0724 07:50:30.219303 140261923243840 deprecation.py:506] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
      "W0724 07:50:30.236515 140261923243840 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
      "\n",
      "W0724 07:50:30.255260 140318003021632 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n",
      "Using TensorFlow backend.\n",
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0724 07:50:30.263308 140420269733696 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
      "\n",
      "W0724 07:50:30.280543 140318003021632 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n",
      "\n",
      "W0724 07:50:30.285316 140420269733696 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n",
      "W0724 07:50:30.286872 140420269733696 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n",
      "\n",
      "Using TensorFlow backend.\n",
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0724 07:50:30.286175 140112875841344 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:74: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
      "\n",
      "W0724 07:50:30.296519 140420269733696 deprecation.py:506] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
      "W0724 07:50:30.298099 140261923243840 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n",
      "W0724 07:50:30.306424 140112875841344 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:517: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n",
      "W0724 07:50:30.308227 140112875841344 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:133: The name tf.placeholder_with_default is deprecated. Please use tf.compat.v1.placeholder_with_default instead.\n",
      "\n",
      "W0724 07:50:30.315028 140420269733696 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
      "\n",
      "W0724 07:50:30.316997 140112875841344 deprecation.py:506] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
      "W0724 07:50:30.317374 139771027076928 deprecation.py:323] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
      "W0724 07:50:30.320197 140565384750912 deprecation.py:323] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
      "W0724 07:50:30.323417 140261923243840 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n",
      "\n",
      "W0724 07:50:30.325587 139831869536064 deprecation.py:323] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
      "W0724 07:50:30.326363 140620935583552 deprecation.py:323] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
      "W0724 07:50:30.334674 140112875841344 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:4138: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
      "\n",
      "W0724 07:50:30.379765 140420269733696 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n",
      "W0724 07:50:30.395873 140112875841344 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n",
      "W0724 07:50:30.398335 140318003021632 deprecation.py:323] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
      "W0724 07:50:30.406669 140420269733696 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n",
      "\n",
      "W0724 07:50:30.420856 140112875841344 deprecation_wrapper.py:119] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3295: The name tf.log is deprecated. Please use tf.math.log instead.\n",
      "\n",
      "W0724 07:50:30.442107 140261923243840 deprecation.py:323] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
      "W0724 07:50:30.530592 140420269733696 deprecation.py:323] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
      "W0724 07:50:30.542064 140112875841344 deprecation.py:323] From /home/Students/anaconda3/envs/BelleAnalysis/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
      "OMP: Info #212: KMP_AFFINITY: decoding x2APIC ids.\n",
      "OMP: Info #212: KMP_AFFINITY: decoding x2APIC ids.\n",
      "OMP: Info #210: KMP_AFFINITY: Affinity capable, using global cpuid leaf 11 info\n",
      "OMP: Info #154: KMP_AFFINITY: Initial OS proc set respected: 0-7\n",
      "OMP: Info #156: KMP_AFFINITY: 8 available OS procs\n",
      "OMP: Info #157: KMP_AFFINITY: Uniform topology\n",
      "OMP: Info #179: KMP_AFFINITY: 8 packages x 1 cores/pkg x 1 threads/core (8 total cores)\n",
      "OMP: Info #214: KMP_AFFINITY: OS proc to physical thread map:\n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 0 maps to package 0 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 1 maps to package 1 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 2 maps to package 2 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 3 maps to package 3 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 4 maps to package 4 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 5 maps to package 5 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 6 maps to package 6 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 7 maps to package 7 \n",
      "OMP: Info #250: KMP_AFFINITY: pid 5871 tid 5871 thread 0 bound to OS proc set 0\n",
      "OMP: Info #212: KMP_AFFINITY: decoding x2APIC ids.\n",
      "OMP: Info #210: KMP_AFFINITY: Affinity capable, using global cpuid leaf 11 info\n",
      "OMP: Info #154: KMP_AFFINITY: Initial OS proc set respected: 0-7\n",
      "OMP: Info #156: KMP_AFFINITY: 8 available OS procs\n",
      "OMP: Info #157: KMP_AFFINITY: Uniform topology\n",
      "OMP: Info #179: KMP_AFFINITY: 8 packages x 1 cores/pkg x 1 threads/core (8 total cores)\n",
      "OMP: Info #214: KMP_AFFINITY: OS proc to physical thread map:\n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 0 maps to package 0 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 1 maps to package 1 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 2 maps to package 2 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 3 maps to package 3 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 4 maps to package 4 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 5 maps to package 5 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 6 maps to package 6 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 7 maps to package 7 \n",
      "OMP: Info #250: KMP_AFFINITY: pid 5873 tid 5873 thread 0 bound to OS proc set 0\n",
      "OMP: Info #212: KMP_AFFINITY: decoding x2APIC ids.\n",
      "OMP: Info #210: KMP_AFFINITY: Affinity capable, using global cpuid leaf 11 info\n",
      "OMP: Info #154: KMP_AFFINITY: Initial OS proc set respected: 0-7\n",
      "OMP: Info #156: KMP_AFFINITY: 8 available OS procs\n",
      "OMP: Info #157: KMP_AFFINITY: Uniform topology\n",
      "OMP: Info #179: KMP_AFFINITY: 8 packages x 1 cores/pkg x 1 threads/core (8 total cores)\n",
      "OMP: Info #214: KMP_AFFINITY: OS proc to physical thread map:\n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 0 maps to package 0 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 1 maps to package 1 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 2 maps to package 2 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 3 maps to package 3 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 4 maps to package 4 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 5 maps to package 5 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 6 maps to package 6 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 7 maps to package 7 \n",
      "OMP: Info #250: KMP_AFFINITY: pid 5872 tid 5872 thread 0 bound to OS proc set 0\n",
      "OMP: Info #210: KMP_AFFINITY: Affinity capable, using global cpuid leaf 11 info\n",
      "OMP: Info #154: KMP_AFFINITY: Initial OS proc set respected: 0-7\n",
      "OMP: Info #156: KMP_AFFINITY: 8 available OS procs\n",
      "OMP: Info #157: KMP_AFFINITY: Uniform topology\n",
      "OMP: Info #179: KMP_AFFINITY: 8 packages x 1 cores/pkg x 1 threads/core (8 total cores)\n",
      "OMP: Info #214: KMP_AFFINITY: OS proc to physical thread map:\n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 0 maps to package 0 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 1 maps to package 1 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 2 maps to package 2 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 3 maps to package 3 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 4 maps to package 4 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 5 maps to package 5 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 6 maps to package 6 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 7 maps to package 7 \n",
      "OMP: Info #250: KMP_AFFINITY: pid 5875 tid 5875 thread 0 bound to OS proc set 0\n",
      "OMP: Info #212: KMP_AFFINITY: decoding x2APIC ids.\n",
      "OMP: Info #210: KMP_AFFINITY: Affinity capable, using global cpuid leaf 11 info\n",
      "OMP: Info #154: KMP_AFFINITY: Initial OS proc set respected: 0-7\n",
      "OMP: Info #156: KMP_AFFINITY: 8 available OS procs\n",
      "OMP: Info #157: KMP_AFFINITY: Uniform topology\n",
      "OMP: Info #179: KMP_AFFINITY: 8 packages x 1 cores/pkg x 1 threads/core (8 total cores)\n",
      "OMP: Info #214: KMP_AFFINITY: OS proc to physical thread map:\n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 0 maps to package 0 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 1 maps to package 1 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 2 maps to package 2 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 3 maps to package 3 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 4 maps to package 4 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 5 maps to package 5 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 6 maps to package 6 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 7 maps to package 7 \n",
      "OMP: Info #250: KMP_AFFINITY: pid 5892 tid 5892 thread 0 bound to OS proc set 0\n",
      "OMP: Info #212: KMP_AFFINITY: decoding x2APIC ids.\n",
      "OMP: Info #210: KMP_AFFINITY: Affinity capable, using global cpuid leaf 11 info\n",
      "OMP: Info #154: KMP_AFFINITY: Initial OS proc set respected: 0-7\n",
      "OMP: Info #156: KMP_AFFINITY: 8 available OS procs\n",
      "OMP: Info #157: KMP_AFFINITY: Uniform topology\n",
      "OMP: Info #179: KMP_AFFINITY: 8 packages x 1 cores/pkg x 1 threads/core (8 total cores)\n",
      "OMP: Info #214: KMP_AFFINITY: OS proc to physical thread map:\n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 0 maps to package 0 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 1 maps to package 1 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 2 maps to package 2 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 3 maps to package 3 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 4 maps to package 4 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 5 maps to package 5 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 6 maps to package 6 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 7 maps to package 7 \n",
      "OMP: Info #250: KMP_AFFINITY: pid 5890 tid 5890 thread 0 bound to OS proc set 0\n",
      "OMP: Info #212: KMP_AFFINITY: decoding x2APIC ids.\n",
      "OMP: Info #210: KMP_AFFINITY: Affinity capable, using global cpuid leaf 11 info\n",
      "OMP: Info #154: KMP_AFFINITY: Initial OS proc set respected: 0-7\n",
      "OMP: Info #156: KMP_AFFINITY: 8 available OS procs\n",
      "OMP: Info #157: KMP_AFFINITY: Uniform topology\n",
      "OMP: Info #179: KMP_AFFINITY: 8 packages x 1 cores/pkg x 1 threads/core (8 total cores)\n",
      "OMP: Info #214: KMP_AFFINITY: OS proc to physical thread map:\n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 0 maps to package 0 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 1 maps to package 1 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 2 maps to package 2 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 3 maps to package 3 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 4 maps to package 4 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 5 maps to package 5 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 6 maps to package 6 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 7 maps to package 7 \n",
      "OMP: Info #250: KMP_AFFINITY: pid 5891 tid 5891 thread 0 bound to OS proc set 0\n",
      "OMP: Info #212: KMP_AFFINITY: decoding x2APIC ids.\n",
      "OMP: Info #210: KMP_AFFINITY: Affinity capable, using global cpuid leaf 11 info\n",
      "OMP: Info #154: KMP_AFFINITY: Initial OS proc set respected: 0-7\n",
      "OMP: Info #156: KMP_AFFINITY: 8 available OS procs\n",
      "OMP: Info #157: KMP_AFFINITY: Uniform topology\n",
      "OMP: Info #179: KMP_AFFINITY: 8 packages x 1 cores/pkg x 1 threads/core (8 total cores)\n",
      "OMP: Info #214: KMP_AFFINITY: OS proc to physical thread map:\n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 0 maps to package 0 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 1 maps to package 1 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 2 maps to package 2 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 3 maps to package 3 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 4 maps to package 4 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 5 maps to package 5 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 6 maps to package 6 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 7 maps to package 7 \n",
      "OMP: Info #250: KMP_AFFINITY: pid 5874 tid 5874 thread 0 bound to OS proc set 0\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5892 tid 6060 thread 1 bound to OS proc set 1\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5872 tid 6043 thread 1 bound to OS proc set 1\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5873 tid 6029 thread 1 bound to OS proc set 1\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5871 tid 6013 thread 1 bound to OS proc set 1\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5875 tid 6045 thread 1 bound to OS proc set 1\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5890 tid 6075 thread 1 bound to OS proc set 1\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5874 tid 6105 thread 1 bound to OS proc set 1\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5891 tid 6103 thread 1 bound to OS proc set 1\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5875 tid 6044 thread 2 bound to OS proc set 2\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5871 tid 6014 thread 2 bound to OS proc set 2\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5872 tid 6042 thread 2 bound to OS proc set 2\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5873 tid 6030 thread 2 bound to OS proc set 2\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5892 tid 6059 thread 2 bound to OS proc set 2\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5891 tid 6102 thread 2 bound to OS proc set 2\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5874 tid 6104 thread 2 bound to OS proc set 2\n",
      "OMP: Info #250: KMP_AFFINITY: pid 5890 tid 6074 thread 2 bound to OS proc set 2\n",
      "OMP: Info #212: KMP_AFFINITY: decoding x2APIC ids.\n",
      "OMP: Info #210: KMP_AFFINITY: Affinity capable, using global cpuid leaf 11 info\n",
      "OMP: Info #154: KMP_AFFINITY: Initial OS proc set respected: 0-7\n",
      "OMP: Info #156: KMP_AFFINITY: 8 available OS procs\n",
      "OMP: Info #157: KMP_AFFINITY: Uniform topology\n",
      "OMP: Info #179: KMP_AFFINITY: 8 packages x 1 cores/pkg x 1 threads/core (8 total cores)\n",
      "OMP: Info #214: KMP_AFFINITY: OS proc to physical thread map:\n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 0 maps to package 0 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 1 maps to package 1 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 2 maps to package 2 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 3 maps to package 3 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 4 maps to package 4 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 5 maps to package 5 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 6 maps to package 6 \n",
      "OMP: Info #171: KMP_AFFINITY: OS proc 7 maps to package 7 \n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 4942 thread 0 bound to OS proc set 0\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6190 thread 1 bound to OS proc set 1\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6192 thread 2 bound to OS proc set 2\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6194 thread 4 bound to OS proc set 4\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6193 thread 3 bound to OS proc set 3\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6196 thread 6 bound to OS proc set 6\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6195 thread 5 bound to OS proc set 5\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6197 thread 7 bound to OS proc set 7\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6198 thread 8 bound to OS proc set 0\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6191 thread 9 bound to OS proc set 1\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6200 thread 11 bound to OS proc set 3\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6199 thread 10 bound to OS proc set 2\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6201 thread 12 bound to OS proc set 4\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6202 thread 13 bound to OS proc set 5\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6203 thread 14 bound to OS proc set 6\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6204 thread 15 bound to OS proc set 7\n",
      "OMP: Info #250: KMP_AFFINITY: pid 4942 tid 6205 thread 16 bound to OS proc set 0\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "import os\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3' \n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "results = []\n",
    "grids = []\n",
    "for architecture in architectures : \n",
    "    layers = architecture\n",
    "    print(\"Using architecture: \", layers)\n",
    "    model = KerasClassifier(build_fn=build_DNN, batch_size=1000, epochs=2,verbose=0)\n",
    "    param_grid = dict(epochs=epochs, batch_size=batch_size,\n",
    "                      activation=activation, optimizer=optimizer)\n",
    "    grid = GridSearchCV(estimator=model, param_grid=param_grid,\n",
    "                        n_jobs=-1, pre_dispatch=8,cv=4,verbose=0)\n",
    "    grid_result = grid.fit(pcTrain, Y_Train)\n",
    "    results.append(grid_result)\n",
    "    grids.append(grid)\n",
    "    print_results()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9830910636920374\n",
      "{'activation': 'leakyrelu',\n",
      " 'batch_size': 100,\n",
      " 'epochs': 15,\n",
      " 'optimizer': ' AdaMax'}\n",
      "0.9823417438324835\n",
      "{'activation': 'leakyrelu', 'batch_size': 100, 'epochs': 15, 'optimizer': 'SGD'}\n",
      "0.9829668794391564\n",
      "{'activation': 'relu', 'batch_size': 1000, 'epochs': 15, 'optimizer': 'Adam'}\n"
     ]
    }
   ],
   "source": [
    "scores = []\n",
    "params = []\n",
    "for result in results : \n",
    "    scores.append(result.best_score_)\n",
    "    params.append(result.best_params_)\n",
    "    pprint.pprint(result.best_score_)\n",
    "    pprint.pprint(result.best_params_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Evaluate the best Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "233545/233545 [==============================] - 27s 116us/step - loss: 0.1268 - acc: 0.9590\n",
      "Epoch 2/15\n",
      "233545/233545 [==============================] - 15s 65us/step - loss: 0.0958 - acc: 0.9700\n",
      "Epoch 3/15\n",
      "233545/233545 [==============================] - 15s 65us/step - loss: 0.0911 - acc: 0.9714\n",
      "Epoch 4/15\n",
      "233545/233545 [==============================] - 15s 65us/step - loss: 0.0882 - acc: 0.9727\n",
      "Epoch 5/15\n",
      "233545/233545 [==============================] - 15s 63us/step - loss: 0.0857 - acc: 0.9734\n",
      "Epoch 6/15\n",
      "233545/233545 [==============================] - 15s 66us/step - loss: 0.0847 - acc: 0.9738\n",
      "Epoch 7/15\n",
      "233545/233545 [==============================] - 15s 66us/step - loss: 0.0827 - acc: 0.9745\n",
      "Epoch 8/15\n",
      "233545/233545 [==============================] - 15s 66us/step - loss: 0.0823 - acc: 0.9744\n",
      "Epoch 9/15\n",
      "233545/233545 [==============================] - 15s 65us/step - loss: 0.0809 - acc: 0.9750\n",
      "Epoch 10/15\n",
      "233545/233545 [==============================] - 16s 66us/step - loss: 0.0815 - acc: 0.9751\n",
      "Epoch 11/15\n",
      "233545/233545 [==============================] - 15s 65us/step - loss: 0.0804 - acc: 0.9753\n",
      "Epoch 12/15\n",
      "233545/233545 [==============================] - 15s 65us/step - loss: 0.0799 - acc: 0.9755\n",
      "Epoch 13/15\n",
      "233545/233545 [==============================] - 15s 65us/step - loss: 0.0792 - acc: 0.9757\n",
      "Epoch 14/15\n",
      "233545/233545 [==============================] - 15s 65us/step - loss: 0.0784 - acc: 0.9761\n",
      "Epoch 15/15\n",
      "233545/233545 [==============================] - 15s 66us/step - loss: 0.0782 - acc: 0.9764\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7ffa63b4e860>"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Best_Score = max(scores) # Best Score: \n",
    "Best_Params = params[scores.index(Best_Score)]\n",
    "Best_Arch = architectures[scores.index(Best_Score)]\n",
    "layers = Best_Arch\n",
    "Best_Model = build_DNN(optimizer = Best_Params['optimizer'],\n",
    "                       activation = Best_Params['activation'])\n",
    "Best_Model.fit(pcTrain, Y_Train, batch_size = Best_Params['batch_size'],\n",
    "               epochs = Best_Params['epochs'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "233545/233545 [==============================] - 35s 151us/step\n",
      "Evaluation Train [0.057409992742124744, 0.983634845533236]\n",
      "140127/140127 [==============================] - 18s 129us/step\n",
      "Evaluation Validation [0.06027559853957129, 0.9830796349030494]\n",
      "93420/93420 [==============================] - 12s 128us/step\n",
      "Evaluation Test [0.05916113777784381, 0.9831085420680796]\n"
     ]
    }
   ],
   "source": [
    "print(\"Evaluation Train\", Best_Model.evaluate(pcTrain, Y_Train))\n",
    "print(\"Evaluation Validation\", Best_Model.evaluate(pcValidation, Y_Validation))\n",
    "print(\"Evaluation Test\", Best_Model.evaluate(pcTest, Y_Test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "233545/233545 [==============================] - 30s 127us/step\n",
      "Trainig Error :  0.01636515446676401\n",
      "140127/140127 [==============================] - 18s 128us/step\n",
      "Validation Error :  0.016920365096950607\n"
     ]
    }
   ],
   "source": [
    "print(\"Trainig Error : \",1-Best_Model.evaluate(pcTrain, Y_Train)[1])\n",
    "print(\"Validation Error : \",1-Best_Model.evaluate(pcValidation, Y_Validation)[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Confusion Matrix on Train and Validation Set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_confusion_matrix(cm, classes,\n",
    "                          title1='Confusion matrix \\n with normalization',\n",
    "                          title2='Confusion matrix \\n without normalization',\n",
    "                          cmap=plt.cm.YlOrRd):\n",
    "\n",
    "    cm0 = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis] # normalized \n",
    "    cm1 = cm # unnormalized\n",
    "    fmt0 = '.2f' \n",
    "    fmt1 = 'd'\n",
    "    \n",
    "    \n",
    "    fig, (ax1, ax2) = plt.subplots(1, 2)\n",
    "    fig.set_size_inches(20, 8)\n",
    "    im1=ax1.imshow(cm0, interpolation='nearest', cmap=cmap)\n",
    "    ax1.set_title(title1,fontsize=25)\n",
    "    #plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    ax1.set_xticks(tick_marks)\n",
    "    ax1.set_xticklabels(dict(zip(classes,tick_marks)), rotation=30,fontsize=15)\n",
    "    ax1.set_yticks(tick_marks)\n",
    "    ax1.set_yticklabels(dict(zip(classes,tick_marks)),fontsize=15)\n",
    "    fig.colorbar(im1, ax=ax1)\n",
    "    \n",
    "    thresh = cm0.max() / 2.\n",
    "    for i, j in itertools.product(range(cm0.shape[0]), range(cm0.shape[1])):\n",
    "        ax1.text(j, i, format(cm0[i, j], fmt0),\n",
    "                 horizontalalignment=\"center\",\n",
    "                 color=\"black\" if cm0[i, j] > thresh else \"black\",fontsize=15)\n",
    "\n",
    "    ax1.set_ylabel('TRUE LABEL',fontsize=20)\n",
    "    ax1.set_xlabel('PREDICTED LABEL',fontsize=20)\n",
    "\n",
    "    im2=ax2.imshow(cm1, interpolation='nearest', cmap=cmap)\n",
    "    ax2.set_title(title2,fontsize=25)\n",
    "\n",
    "    ax2.set_xticks(tick_marks)\n",
    "    ax2.set_xticklabels(dict(zip(classes,tick_marks)), rotation=30,fontsize=15)\n",
    "    ax2.set_yticks(tick_marks)\n",
    "    ax2.set_yticklabels(dict(zip(classes,tick_marks)),fontsize=15)\n",
    "    \n",
    "    thresh = cm1.max() / 2.\n",
    "    for i, j in itertools.product(range(cm1.shape[0]), range(cm1.shape[1])):\n",
    "        ax2.text(j, i, format(cm1[i, j], fmt1),\n",
    "                 horizontalalignment=\"center\",\n",
    "                 color=\"black\" if cm1[i, j] > thresh else \"black\",fontsize=15)\n",
    "\n",
    "    ax2.set_ylabel('TRUE LABEL',fontsize=20)\n",
    "    ax2.set_xlabel('PREDICTED LABEL',fontsize=20)\n",
    "    fig.colorbar(im2, ax=ax2)\n",
    "    plt.tight_layout()\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "import itertools\n",
    "\n",
    "# On training set\n",
    "Y_train_predict = Best_Model.predict_proba(X_Train)\n",
    "Y_train_single_col = np.where(Y_Train==1)[1]\n",
    "Y_train_mo = pd.get_dummies(Y_train_single_col[:])\n",
    "Y_train_predict_mo = Y_train_predict.round()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "cnf_matrix_train = confusion_matrix(Y_train_mo.values.argmax(axis=1), Y_train_predict_mo.argmax(axis=1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABWgAAAJVCAYAAABDMBy5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nOzdd5wURfrH8c+zSM4giIAEI+YAZw6Ycw5gOtGfZ84nh3oGTj1zFj3v9DzM8Tz1zBEDRjzFAEZAEUQElpx3n98fVcP2DjOzs8vuzu7yfb9e/Zqd7qrq6jA7z1RXV5u7IyIiIiIiIiIiIiK1r6jQFRARERERERERERFZWamBVkRERERERERERKRA1EArIiIiIiIiIiIiUiBqoBUREREREREREREpEDXQioiIiIiIiIiIiBSIGmhFRERERERERERECkQNtCK1zMw2NrPHzewXM1tqZm5mnxWwPv1jHbxQdZCKmdmgeJwmFLouUlhmNiKeC0MLXRcREZF0inWlKhTrSopiXVlZrVLoCohUhZk1Ag4F9gO2BjoDLYCZwLfAO8BD7v5lwSqZgZn1BkYCreOsGcASYFrBKiU1wswOAjYDPnP3pwtdn6qKgdFlabMdmAvMBn4CPgXeBJ5198U5yuoFjM+waFEsaxrwOfAR8Ji7T6qgbiOAneLbT4G+7p7xx5eZDQL+BeDulqvcLPUeFPMOrUxeWTFmthlwEDDT3W8pdH1ERGqLYl2p6xTrZiyrF4p1pRIU60qSetBKvWNmWwNjgMeAY4F1CAHrHKAjsB1wAfCFmf3bzJoUqq4ZnEwIWL8H1nD3ju7exd13K2Cd5gPfxEmqz0GEYO+gaipvFuEY/VBN5VXFr3GaSghcuwLbAKcBTwCTzexUM8snKJydKG8O0A5YHxgA3Aj8aGYPmdmqedZtc2BgJbalMnoRjmV68F4oPxHOhZXhx+5mhP1+TqErIiJSWxTrVjvFujVDsW5uinWrTrGurJTUQCv1ipntD4wA1gWmAxcC67p7E3fvCDQBfgdcQ/hSPIQQ0NYVG8fXZ9z954LWJHL3j9y9j7v3KXRdJDt3/088TrsWsA5dElNboDGwCfBHQm+BjsCdwIN5BK5nJ8rq5O5NCEHwocCLQCPgKGB0vKqfjyvNrHGlN6yecfffx3NhWKHrIiIi1UuxbvVTrFs/KNbNi2JdkQZMDbRSb5jZOsCDQFNCr4LN3P0ad/8ulcbdS9x9lLtfCPQGnilMbbNKBdBzC1oLkWoQP29fuPtNwEbAo3HRUYSePZUt7xd3f8rd9yH0LlhCCGSfN7NcQ/K8DiwE1iT03BEREal3FOuK1C2KdUWkNqmBVuqTK4E2hC+ngyu6Ku/uM9z9IMLtMuWYWRczu97MvjKzuWY2L/59nZmtlqk8M+uVesBA/Hs1M7vVzMab2UIz+9XMHjWz5a7Om9mE+GCC/nHWZYmy3Mz6x3RD4/sR2barogcdmNlW8XaZVL3mmdmPZvaWmV1iZt0rU14h9lc+0uttZpuY2SNmNtnMFpjZWDM7PxnsmNl2Zva0hYdWLDSzL83s9GxXwM2ss5mdYGZPxfJmxbK/N7N7zGzDbPUCjouzjks71suOd0w/Ic4bZGatzOxyM/vCzOak9l1Ml/HBCXGbUg/gyHhrjJl1N7PpMc0/KrOf8+Xu8+M2fxpnXWBmHVagvMeBi+LbDSjbn5n8DNwe/77EzFpVdb3p4v5+M/E+/VgOTywr9/k1s0PN7BUzm2pmpZZ40IGZrWdmg83sNTP7IZ5Xs83sUzO70nLc7mY5HpyQdj41iesYHT+zs8zsDTPbawX2R/P4uXrfzIrNbImZ/WZmY8zsPjM7NEfetczs9vhZmmtm8+Pft5hZjwzpnTiWGtAzw75fbvtFROo5xboo1s1Wb1Osq1g3UKyrWFcaKnfXpKnOT8BqQAlhLKB7VrCsnYDiWJYD8whX+VPvZwDbZ8jXK5FmX8J4Qqn8CxPLZgGbpuX9GJgCLKZs0PkpiWnbmG5oXD4iR/37p9aVYdlxQGmiLgtjfTwxDcq3vELtrzyPY/9EGXsDC+LfM9P2wSMx/YnA0rhsZto+uSbLOoanpZtFuNKd3L+HpuXZNh7TVH0WpB3rZcc7pp8Q0/2RMNaSEx4kkNrnvWK6QfH9hAz1vCSRb/O0ZUWEWyUdGAu0qOR+Tp2TGc+PDOkPS+yfE3KcE4PyKKsZ8FtM/3aG5antGg60T+yzSzOkHVSZ7Uj77M5I1Dv9WN6aYV+NIIwt5vF8mxHPvaEZjnsqTTHlz9ufgfWy1Cm13UMzLEuVewbwQfx7MWHss+T6TqjMfohltwY+y1Dv5GdiufMz5v0DZf//Up+d+Yn3s4Dd0/JMoez/V0mGfX9+ZbdBkyZNmurqhGLdZFn9U+vKsEyxrmJdxbqKdRXramqQU8EroElTPhNhQPRlAdAKlLNG4ovtK2C7xLIdgK/jsulAt7S8yS/cGcC7QL+4bBVgN2AyWb5gY7qsXzZx+bIvvRzb0D9Vj7T5LQhjkTnwALBWYllLoC9wHbBPPuXVhf1VwbHsnyi/mHDLUY+4rDVwVWL5BfEL8zagc0zTnnDFMvWFuG6GdVwGXEEYvL1lnFcEbEi4BdEJAXzXDHmHx+XDK9iOCTHdHOAX4GCgcVzWnRhkkjtoLSJc+XZC4NsysexSyoKEqvw4SJ2Ty50fWdK3IgRoDtyX45wYlGd5j1EWkDfL8nkaHt8Pie9nA53S0g6qzHbk+xnJsq9SAeK1qXoQblftmUj7KCGwXAtoEuc1AXYFPoz5P8myntR2L/d/JHE+zSAEvgcmzqf1gPcTdWxbyf1wMWWf90OAponzryvhQTb/yJDvIMqC56uBnoDFaT3gccoC1x5Zjtty570mTZo0NaQJxbrJNBm/d1Gsq1hXsa5iXcW6mhrwVPAKaNKUz0QIHFJfdssFCJUo52+Jf+hdMizvTtlVrGFpy5JfuGOB5hny759I0z3D8qxfNnF56ktvRI5tyPgFCmxJWRC1SiX2SdYv5ELvr3zrDbwCWIY0byfS3J1heSPCgP8OXFyF8+m5bHmpfNC6lLQeAWnpcn55A90ITzp14F9x3naUBZBnVfEzkzonlzs/cuT5NuZ5N8c5MSjPsi5K5Fk7bVnq8zQ8vm9OCNScxNX+tP2X93bk8xnJtq+AG6uyv2M5rQhXzZ3MPXZS2z00x/m0EOiTYXknynq8HF3Jer0Q811YiTxNEscka08GwhiKDtxSmfNekyZNmhrKhGLdZJqM37so1lWsq1hXsa5iXU0NeNIYtFJfdEz8PaMqBcSxl46Ib+9y9ynpaTyM9XVXfDswR3E3uvuCDPNfJFw5g7Kn2NaWmfG1CeX3V5XUs/11rbt7hvkvJ/6+On2hu5cAr8W3m1Rhvc/H1+2rkDfdS+7+acXJMnP3ScAJ8e0gMzsNeJgQmD/v7rdVQx3zlfqMVnlcrgxlVVhePMeGxrenmFnvalh/VZQSehRUibvPBd6Kb6t6bj3p7l9nKPs3Qs8CqPw5n/ofs3ol8uxN+EH1K2VjbGVyf3zds5J1EhFpKBTrVkyx7vIU6yrWLQTFumUU60q1UQOt1BcZB7avpN6Ufem9liPdq/G1Y44vvQ8zzXT3pYRxhKB6vrAr4wfCbViNgQ/NbIiZbWZmjapYXn3aXx9lmf9rfJ3h7uMqSNM+00Iz29TM7jSzz+PA9qVW9sCGO2Oy7pnyVtLIFS3A3Z8FhsW3dwA9CLeSDVrRsiupOj6vVS3rX4TPQRPg8mqsR2V87+5TK0pkZvuZ2WNmNi4+2MAT51bqB2NVz62Mn7locnyt7Gfuufh6hoWHlByU6wEPUSrobg/8YmZTMk3A3TFdz0rWSUSkoVCsWzHFustTrKtYtxAU65ZRrCvVRg20Ul9MS/xd1eCmc+LvSTnSJZ+Y2zlLmjk58i+Nr43zqVR1iVfIBxJuY+oJXEN4yuhsM3vVzE41sxaVKLLe7C93z1Z+quwqrd/MzgD+B5xK6PXQinCb269xmh2TtqxklTOpMMjJ0/mUP14nuPu0bIlrSOoHwPRqLCuv8uLn4M/x7VFmtmk11KGych5LMysys4eB/xKC096EILuYsnNrYUxe1XOr2j9z7v4wcCvhNqyBwH+A38zsOzO7w8z6ZsjWNb42ITwAJ9uUOs7NK1MnEZEGRLFuBRTr5ixbsW7tUqybg2JdxbpSNWqglfriq8Tfm1dDeZluEVqRdHWCu48G+gCHAv8AviR8CexGuAL+tZlV5faqBrm/cjGz9YFbCP8nnyCMe9bM3du7exd37wKcl0peDassqYYyIDxFuFvi/U7VVG5ezKwVsGZ8+0M1FJkKOheR+8fTMu7+FOGqehEZbverBRUdy/8DjozpLgfWITyEoEPi3Hoypq3OHhorzN3PITzs4CLCbZszgbWB04BRZnZLWpZUr6aX3N3ymWptY0RE6hbFunlQrFt9FOtWjWJdQLFukmJdqTZqoJX64k3CWDcQnvxZFckrfWvkSJe8zeK3rKlqRupqX7McadrmKsDdF7v7U+5+srtvTBgo/RTC+EZrAPflWZf6sL9q0mGEL9yxwEB3/9jdF6el6VL71crOzNYA7olvP4+vfzKzXWqxGntRFqiMWJGCzKwZkKr7B+6+MFf6NEPi695mVquBex5SY9jd4+6Xufv37l6alqZOnVtJsb5Xu/s+hDEAtwGejovPNrMDEslT4/nV9jiFIiL1jWLdMop1a4di3apRrFsxxboiVaAGWqkX3P1X4N/x7VFmtm6+eeMDACDcDpUahH3XHFl2i6/T3X18pSq64orja64gcavKFOju093975R9iW9uZvk8WKE+7K+alDoGozMEFCm7ZZkPZT+yauUqaRx/7SHC7TNjgK0Jt+UUAQ/kecxXtA5NCFebIdwe93SO5Pk4A0iN+zS8Mhnd/S3CVW9YgYcYRMuOf+L/yYpInVsZH5QRe2ZU6nNeKO5e6u4fEH7k/RRn755IkhpvrpuZVeUhELX6ORIRKRTFuuUo1q0dinUrXwfFuvlRrJs/xbqyjBpopT65GJhLuI3pKTPrliuxmbU3s38Tr8LHJ58+FhefbGbLXbUzs67AyfHtI9VV8UoYHV+7mtnW6QvNrDPwh0wZzaxpBWUnnyxb4S1G9WR/1aRZ8XXjTIGKme0N9M+RPzVmV7tqrlc2FwM7EG6POjI+5fVEwrhpXcn9VNEVZmbNCYFl6rbMq919ZvYcFZZ3OHBVfPsl8GAVirmQEPRsBRxS1bpQdiyheo5n6tzKNmbYJUDralhPtcr1PyaOh5bqdZP8//JfwsM7AG6taGxAM0sfd7G2P0ciIoWkWFexbm1SrFsJinUrRbFu9nUo1pWs1EAr9Ya7fwscS/jHuCHwWXx669qpNGbWyMw2N7PLgXEs/0V1FWEcmQ7Aa2a2bSLvdoQnuLYjXE2/pia3J4v3gB/j38PNrJ8FRWbWn3AbTbbP7UAzG2lmJ5tZalyk1D7Zk7Lteb8SwURd31816aX4uiFwR+rL1MxamtnJhHGTcg3k/2V83cHM+tRcNZcdi0vi28Hu/jmAu88AjiEEbvvHB0FU53qLzGwjMzuPMHbekXHRA8B1VSivi5kdYmbPA48TBvafBOwXn4JcKXGcuofj2/0rmz/hW8oCshOroWdB6tz6g5mdFHtjpLb/ZuBPVM9DJ6rbh2Z2m5n1N7NlD3Qws65mdjthfC6AF1LL4q16pxHG7NsCGGlme6a2OebvHf9vfRTTJqU+R23M7AhERBowxbqKdWuZYt2K16tYt2oU6yrWlapwd02a6tUEbAd8R/gnmJoWEf7JlyTmlRK+sBqn5d+JEIil0s2NU+p9MbBDhvX2SqTplaN+E2KaQRmWjYjLhubIvyfhCzK1rnmEHgFO+PIcmFqWlm9Q2j5ZSHgicHKfTAL6pOXrn6m8urC/KjgPctY7bZ9MyJFmaEwzIsOyR9L2aTFh7DQHRhFuS8pYPuH2q6mJvL/FbZ0AbF3Z7c+2LYQfDT/GZc9lyXt5XL4A2LiS+3loYhumJKbitHMrtY0n5ygreU7MSpQ1lfAZTpa1FLgf6JCjvBEx7fAK1lmu7Mpsf6KcexJlzIv7fAJwQz7nUoZjNjZRXkncn6Xx/V2EHhoZt40c/0fyOZ9ylV1BvSck6lwa65z8X+DATVnyHh33WyrdEsL/p4Vp+f+cIe9rieWzKfscnVOVY6lJkyZNdX1Csa5i3TzqnbZPJuRIMxTFurnWOzSxDYp1Fesq1tVUsEk9aKXecfeRhKe3HkkYh+h7wj++1oSr2+8CfwXWd/ej3H1JWv63Yv4bCV8cRYQxX8YCN8R879TO1izP3V8m3L7zHOFLoREwkXDVvi9lA5Gnexb4PeH2ntGEoKAtMAf4iHDVeUN3/7qS9anT+6uGHQ2cQ3gIwSLCsfiCcDvRdoQv7IzcvRjYEXiU8GOhLdAzTrkejFFZdwM9COfF8VnS/IXQY6UZ8Gi8RasqVotTZ2CVuM4PgL8RxmXq5mEMuHy0SZTXlhCMjCXcange0MPdf++hZ0SVufsEQhC4ok4nBKWpq9w9CMdy1WwZctRpJrAt4cnJEwhB61JCMHqku5+ywrWtGQOBy4DXCeP2NSH0/PiRcNx2dffzMmV094cIvQ6uJPzgm0sI3hcCnwHDCOPcZRpD7TDgZsKP9saUfY50K5iINEiKdRXr1iLFuuUp1lWsq1hXCsY8tNiLiIiIiIiIiIiISC1TD1oRERERERERERGRAlEDrYiIiIiIiIiIiEiBqIFWREREREREREREpEDUQCsiIiIiIiIiIiJSIGqgFRERERERERERESkQNdCKiIiIiIiIiIiIFIgaaEVEREREREREREQKRA20IiIiIiIiIiIiIgWiBloRqRFm1svMPE69qljG0Jh/RLVWTqqNmQ2Kx2hCZZbVVdVx3oqIiEjdorhU6opc51F9PMfMrH/qs1XouojUd6sUugIisvKJgfEgAHcfWsCqyErEzIbGP4e7+4QCVkVERETqiJUhLjWz/kB/YIK7Dy9oZaReMLN2wDnx7S3uPrOQ9RFZGaiBVkRqyhLgm8TfSb2Ay+LfQ2upPlL7ZhHOgUmFrkiUOudGABOypMl13oqIiEj9tLLHpf0J2/gWMLygNZFcphHO058KXRGgHWWfi+FAtgba+ZR9tkRkBaiBVkRqhLtPAvoUuh5SOO7+H+A/ha5HZei8FRERaXj0/S71gbsPA4YVuh6V4e4foc+WSLXQGLQiIiIiIiIiIiIiBaIGWhHBzL6Mg7ufkWHZNomHKjyZYXljM5sbl++SmJ/xYQzxgVFvJt572jQ8Rz13NbPnzew3M1toZmPN7DIza1bF7R6eXKeZHWZmI8xshpnNN7PPzOxsM8v5vzIOjv+EmU0ys0VmNs3MXjez482sUZY85R4CYGaHmtkrZjbVzEpT46Wm70cz62lmd5vZT3Ef/GBmV5pZy0TZG5nZg2Y2Mab5zswuNrPGWerS1swGmtlDZvZF3P6FZvajmT1sZltXcf9mfEhY8mECeUyD0vJuYWaXmtnbsX4LzWymmX1gZkPMrFWGegxPe3DBm2nrmJBIW+FDROL+utTM/mdms81sQdzHfzOzNXPsj1S5/c2sdTxuX8f8083sOTPbKp99KyIi0lBZA49LzWxzM7s/EccUm9l7ZnaOmTXNkqfCh0dZhoc1pbabslvVd6oo1qqg7iNinqEW/MHMPozx0Bwze9/MjsmjnENi3POrmS2Or8+Z2cE58iyL2+O6TzSzd2MMtWw70veDmW1iZo+Y2eQYc401s/PNbJVE2duZ2dNm9ks8Jl+a2elmZlnq0tnMTjCzp2J5s2LZ35vZPWa2Yb77NK3cjMfZymLqfKb+iXxFcduusRAr/xz393Qze8vMTrEMvw/i+scnZo1PW8eIRNoKHxJmZl3M7Hoz+8rC53Ne/Ps6M1stS57030CrmdmtZjY+HqNfzexRM1PvXWkwNMSBiAC8AWwI7MLyt9Xskvi7v5mZuye/gLcEWgKLgPfzWNdvQBugfXz/a9ryWZkymdlg4NpEmiaE22mGEoLN3d29JI/1Z2Rmw4DTgVJgNtAc2BS4BdgCOC5LvpuAc+Nbj3VrR9hvuwDHmNlB7j4nx7pvBM6L+WfGOmSyBfDPWP5swv/wNYE/Azua2a7AHsDjQAvK9tPawBXARsDADOWeS1ngDjA3vvaI00AzO8fdb8u2DZW0mOWPe9IqQMcsyz5J/J06Vu2AreL0ezPb2d2nJtLNiutLBYDFsQ4pv+Vb8RhwvwR0j7MWEsayWztOx5vZ0e7+7xzFrA78L6ZfGLejA7AvsIeZ7e/uL+dbJxERkQamwcalZnYOcBNgibwtgW3idLyZ7eXuv+RR93yUELapVVzPEmBGWpoFVSi3EWEYqwOBpYRxSFsDWwNbm9k67n5ZeiYzawLcDwyIs0oJ+2BVQhy0r5k9Ahzn7tmeBWCEWPewRP6MsbOZ7Q08BTSL6ZoSjtP1QF/gSDM7EbiL0HltdkyzIeHcWwO4IEPR11H+t0EqLl8rTsfkEQ9WxgJyx85NCfFwuh7Au4n3qWPVAdgxTkeZ2Z7unjwPZhDGw101vp9GOJeSy/NiZjsBTyfqN5/wm2eDOJ1oZge4+7tZioBwPO4FOsf8xL8HAHub2Y7uPjrfOonUWe6uSZOmlXwCDiZ8Uc4AitKWvUZZw6MDm6UtvzjOH5E2v1ec70CvtGX9U8sqqNfQmK6YEBRcBawal7UB/pJYxwlV2O7hie1eRGiobBOXdQTuTpS/S4b8ZySW/x3oEue3JDz1dElc9miObZsTX68FOsVlTYGeGfZjcTweG8RlzYEzCcGWExphZwKPJvK3Aq5MlLFbhrqcQvixsBXQLs4zoDehgbo0rmPzDHkHxXInVGZZBcflocT2rpO27FXgeELAuUpiPxwMfB3zPZWl3NQ+6J9j3bnO29bAuLjsZ2Af4ueF0Jj/fly2ENg0x/pnAF8BOxN+DBjwu0T9J5D2OdSkSZMmTZpWlokGGpcC+yWWPw30jvObAMcSGvkcGAk0yrLuETnql3U78smf57EZkTg2MwmNlM3jsu7As3F5CWkxXExzQ1xeClxOWdzZHvhrYv9ckyHvcMpi5yXAHymL21sBq6fvh3isHgV6xGWt43FLLb+AcNH+NqBzoi7/SmzHuhnqchkh7t4MaBnnFREaEh+MeecCXStzLKpynOL5kzouP6bOycQxeRo4AuhKWdzaihCnT4r5bspQbi+yfGYqcd6tEY+BE2Lf7RLLdqAs9p0OdMux/hmEhuZ+cdkqwG7A5Lj87RU5rzVpqitTwSugSZOmwk8xECmJX3BbJOY3JVylnEe40uzAeWl534jzL0ubn/VLPdcXeVq6oYkyhmZJ8++4/NUqbPfwRPmDsqQZFZffnTa/eQwmHHg4S94zE+X3y7FtN+aoY3I/fgk0zZDm/kSaVwDLkObtuPyeKuynYdnyUs0NtJT9uFlChsbkCvJ2o6xHao8My1P7qH+e+zv9vB0S5y8GNsqQtzXhdjAHnsux/qnEHwFpyzdOpNku17Zq0qRJkyZNDXWigcalhAYqB94hrQE2Lt8/Uf5hWdY9Ikf9sm5HPvnzPDYjEnXcOcPyppQ1+v05bVk3yjovXJWl/BsTsdbqacuGJ9Z9Zj77gYrj4uVi/Li8USKmu7gK++m5bHlzHYuqHKfEfpkNbFzJevajrDG5Wb6fmUqcd3+jrIG1S4bl3Sm72DIsx/rHEi8E5PjMdF+Rc1uTprowaQxaEcHdi4HUbSHJW8e2JjREjiTc1l1uuYVxsraJb9+swSouIlxxz+SZ+LrJCpQ/kdDImcmzWcrfnXB7EIRgKpM7gdQtakdmSVNK2S1yFbnZ3RdlmJ+8Hf4ad/ccaaqyn56Pr9tXIW/eLIxZdml8e4a7v1aZ/B6e0Dya0CN122quHpTdjveku3+ZYf1zCLe8Qbjdqm2Wcv7h5YdgSOX/grLxvlbkfBYREam3GmJcamabEG7nBrjCMwx/4O7/BT6Kb7PFjXXFSHdfbh/HODVbzHkooefjQuCaLOVeSdi/jQlDGGRSTLhzLR/XVhAXA1ydvjAen1QcWpdj54sJvZhLgIExlsybu48idBxoSegNXJ11M0LPXYC73H1KhvX/TBheAjIPw5Zyo5cfgiHlRcqGLdu4qnUVqSvUQCsiKW/E12QgvEti2XuEgGnHxKD62xLGdFoAfFiDdfvK3edmWTY5vnbIsjwfH7t7tnFfs5XfL75OdPdvM2WMwd0baenTfZ+psS6Lj7LMT45J9XEFadpnWmhma5rZDWb2iYWHbpUkBvx/ISbrnilvdTCzHQjj60K4zSpj4B0fdnCUmT1r4UFpC5IPLSCMPVftdY1jpqUC9FwNx6/G1yLCmMGZ5PqsVMf5LCIiUt81tLg0FQcuBd7KUXYqjsgWN9YVVYllUtv0sbvPzpQxNs6PSkuf7mN3X5xlWbqKYucZ7j6ugjTZYudNzexOM/vcwkPSShPx6J0xWU3GzgMJw0QAnOvuL2RJ1yQ+DOwVCw9KW5gWO3euobr2puwcyCd27mhmvbOkyXi+uftSyp4lodhZ6j09JExEUt4kjOW0g5mtEr/wdo7L3nD3BWb2AbATIWD6ILH8vSw9O6tL1gdsEQJdWLH/Z/mUn/6E01QwM6mCsn9OS58u38ZZyF7PVB1TvThzpcn0pNaDgUcIt6WlzCb0cHDC2FbtCVfXq52ZrUN40EQT4L/A4CzpWhBuGds5MXsx4bap1IMkOhC2sbrr2oFwuxvkPuY/J/7Odsyrcr6JiIisTBpaXJqKCaZVULeK4sa6ol7EznnExVWKyczsDOBWyjq8OeFW/dSxbU4Yl7imYudtCePkGnCnu9+eJV1nQuNosnfpQso/9KsTYTuqu67J41eZ2Hl8hjSKnWWloB60IpLyNuELrhWwZWwM24oQbHwS06T3Ztglbf7KKNNtU5VJt9wtbrXJzDoSxq5qSjiO/YEW7t7W3Vdz9y7A4TW4/g6E28A6Ap8BR+Xozfxnwo+vBYQHuvUkjJfV0d27xLqmrrBbljKqQ65j7ln+FhERkfw11Lh0RePGhqC+x87rEx6iWyPx5zUAACAASURBVAQ8Qbh7q5m7t0/Eo+elktfA+nsTHvzVjDBUw1k5kt9MaJydDpxAGNe3ubt3StQ11eO5ULFzVdKJNEhqoBURYNkV5lTAuwthzKQmhKdipgKh1FhTu5hZS8puJ6/Jcb7qqtTV+zUqSJe6Xei3nKkKZx/CFf5iYH93fyvDGE9damLFcdiAp4B1CGP17p/jlkEoG5vqcne/xd1/yjCuWI3UldBLN/U5yHXMk8vq6jEXERGp0xpgXJqKGzvFsXKzyRY3pnoJNsuRN9vY93VFQ4mdDyPcVTWWMO5rpiEXaip2bkfo2NCJ8NC5IzKNZxzTNgYOiW/PcPd/pY8Da2aNgFVroq6U7+mc65gnh1aoq8dcpFaogVZEkpYFuiRuI0ss/4Dw9NxtgV0Jt5LMJfu4p9ks6yEZB5Cvj1LjY3U3s3UzJYhBT2o/VnYf1ZZUwPSNu8/Pkma3Glr33YRbExcAB8QHBeSSquunmRaaWS9g7Rz5U425lT7nYuD9eXy7a46kqX1VCvyvsusRERGRZRpSXJqKG1chxD7ZpOKI9G0ojq+5Grq2yrEstY2FjLuXjS2b7UGqsQFy2Vi1tVKryksdg9E57vqq9tg5Nrg+CaxPaPzcL9tYvlEnyhr0M8bOhAsf2Rr9k9tWlfNmPKGDA+QXO09390zDG4isNNRAKyJJqaB3G2DvtHm4+xLCk3ObAxfF2e/GccEqIxlMtKtCPeuCVwm3CwEMzZLmZKBr/PuRmq5QFc2Kr+ua2XIBmpltBhxV3Ss1s0uA3xMaTY+NT5GtSKqum2ZZnu2JwCmp866q59yj8fUwM9sofaGZtQL+FN++4O6z0tOIiIhI3hpMXOrunwNj4tuL40X8csxsH8oaWdPjxtHxtauZbZ0hb2fgDzmqsKIxUHX4N6EncDNgSJY0FxGG3VoS09dFqfhu40wN+ma2N2HIsOr2N0JD5yLgIHefUEH62ZR1Tlgudo4P1/trBflTKn3exLvcHotvTzaz5XoVm1lXwu8lqLu/lURqjRpoRSRpJOGhS80IX+TTgC/S0qQC41QAWZXbyL6N6wE4sT72oo3DAAyNb480s7vMbDUID7MyszMJ41MBPObun2Qopi54hXCFvAPwkJl1g2VPfD0iLs81MH+lmdnhwF/i24vcPd8A/KX4erGZHZJ6arOZ9Tazh4EjKOthksmX8fXoOJZdZf2N0BugMfCime1tZkWxDhsTxgHrTTi3L65C+SIiIlKmocWlqUbJHYAnU0+sN7PGZnY0ZQ1U7xHGGE16D/gx/j3czPpZUGRm/YER5P5tn4qBNowPmKp17j6J8GAtgAvM7C+xxyxm1s7MrqDsQbE3ufsvhahnHlLx6IbAHfF5CphZSzM7mdDLdXq2zFVhZoOB/4tvj3f39yvKE4cNGxnf3mRmuyTi1o2AFwi9ledlyT+Tsod7HZ+KuyvpKmAm4XfGa8lzz8y2IzzArB2hp21FHS1EGjw10IrIMvEW9w8Ts97MMMZneuBb6UA4rueB+PY6YK6Z/WhmE8zshsqWVyjuPoww+D6Eq7+/mNkMwpX12wgNeW+Su0dDQbn7d8D18e0hwM9mNpNwi+Bj8TXXwweqYl/KbpU6x8ym5JgGJPJdDPwKtCb0qlgQ6zoOOJLwELHPye6u+HooMNPMfo7n3Lv5VDqOh3cAIVjtTghs55nZrLjebQm9Go5299FZCxIREZEKNbS41N2fIzw8yoGDgHFmVkyItR4kPBPgC+Dw9HFF4630JxN6lq5HuP1/LqFx7U3C0Amn51j9COAbwtipI81sRty+CWZ2WHVtYx4uAh4nxIGXAtNj7DydsovbjwCX1GKdKsXdX6fsrqpTCdtQTIj/7yKMTTu0mle7b3wtBW6uIHZONsCfQzhHugGvA/PNbDbhPNuZ8BtlWo71pmLnMwmfi5/iOfNojjzLxOHLDiLsmw0J595cM5sLvEsYrmEmoUfwpOwliawc1EArIumSgW2mp+COouyWl9lUfZzN0wnBS+qKfg+gJzU3UH2NcPfzCGOj/ZvQeNiK0OP0TcLTUnePDXt1lrtfQBhu4CPCeLCNge8JV703p+zprjVhtQqm5ol6/ki40v/PRJ0WAs8Be7r71blW5O4PAscSAsL5wOqEc657rnxpZXxJCDCHAp8RbtVrCvxACGI3dPcn8y1PREREcmpQcam730yIZR4EJgItCLHXB4TG2y3dPWPc5e4vE3rfPke4Y6hRLOMaoC8wJVO+mHcp4fb4e4AJQEvC9vUkxK61wt0Xu/sAwsXyFwkNs63j64vAIe5+VBy+oi47mtD4+Tnh4nwjQqPnhcB2hMbzmlBExbFzk1TieAffloRG8Wkx/5z4flt3f4DcrgLOJnzOlhBi5p5U4iFo7v4W0Ae4kdB4XURooB8L3ACs7+7v5FueSENmy1+EFBEREREREREREZHaoB60IiIiIiIiIiIiIgWiBloRERERERERERGRAlEDrYiIiIiIiIiIiEiBqIFWREREREREREREpEBWKXQFZOWx6qrtvFev1QtdDalvxv1Y6BpIfdWhcaFrIPXUJz/MnubunWpznWtbS59PSbWW+QuLXnb3vaq1UBHJKsS6XQtdDal39NBuqSordAWknvrkk7G1GuvWRJwLFce6ZnYvsB8w1d03ivOuB/YHFgM/AMe7+8y47ELg/4AS4Cx3fznO7wsMB5oDLwBnu7ubWVPgfqAvMB0Y4O4TYp7jgItjVa509/sq2h410Eqt6dVrdUZ9eG+hqyH1TMkxJxe6ClJP2YA1Cl0FqacaHfxCrV8Zmk8JJ9OzWsscyrerVmuBIpJTr15dGTXqgUJXQ+obLy10DaS+Mt0QLVVj1q9WY92aiHMhr1h3ODCM0Iia8ipwobsvNbNrgQuBIWa2ATAQ2BDoCrxmZuu6ewnwN+Ak4ANCA+1ewIuExtxid1/bzAYC1wIDzKwDcBnQj3AV7hMze9bdi3NVVp9oERERkQIzQlBWnZOIiIiISKHVRJybT6zr7m8DM9LmveLuS+PbD4Du8e8DgUfdfZG7jwe+B7Y0s9WBNu7+vrs7obH3oESeVM/YJ4FdzcyAPYFX3X1GbJR9ldCom5PidxEREREREREREalPVjWzUYnppErmP4HQExagGzAxseznOK9b/Dt9frk8sdF3FtAxR1k5aYgDERERkTpAV81FREREpCGqoTh3mrv3q0pGM/szsBR4KDUrQzLPMb+qebLSbwERERERERERERFp8OIDvPYDjo7DFkDo5Zp8iEl3YHKc3z3D/HJ5zGwVoC1hSIVsZeWkBloRERGRAtMYtCIiIiLSEBVqDNqMdTHbCxgCHODu8xOLngUGmllTM+sNrAN85O6/AHPMbOs4vuzvgWcSeY6Lfx8GvBEbfF8G9jCz9mbWHtgjzstJQxyIiIiI1AFqVBURERGRhqgQca6ZPQL0J4xV+zNwGXAh0BR4NbS38oG7n+LuX5nZ48AYwtAHp7t7SSzqVGA40JwwZm1q3Np/Ag+Y2feEnrMDAdx9hpldAXwc013u7uUeVpaJGmhFRERERERERESkwXD3IzPM/meO9H8F/pph/ihgowzzFwKHZynrXuDevCuLGmhFRERE6gT1oBURERGRhkhxbsW0j0REREREREREREQKRD1oRURERArM4iQiIiIi0pAozs2PGmhFRERE6gDd1iQiIiIiDZHi3IppH4mIiIiIiIiIiIgUiHrQioiIiNQBumouIiIiIg2R4tyKaR+JiIiIiIiIiIiIFIh60IqIiIgUmKGr5iIiIiLS8CjOzY8aaEVERETqAAWuIiIiItIQKc6tmPaRiIiIiIiIiIiISIGoB62IiIhIgenWLxERERFpiBTn5kf7SERERERERERERKRA1INWREREpA7QVXMRERERaYgU51ZMDbQiIiIidYACVxERERFpiBTnVkz7SERERERERERERKRA1INWREREpMD08AQRERERaYgU5+ZH+0hERERERERERESkQNSDVkRERKQO0FVzEREREWmIFOdWTA20IiIiIgWmW79EREREpCFSnJsf7SMRERERERERERGRAlEPWhEREZE6QFfNRURERKQhUpxbMe0jERERERERERERkQJRD1oRERGROsAKXQERERERkRqgOLdiaqAVERERKTA9PEFEREREGiLFufnRPhIREREREREREREpEPWgFREREakDdNVcRERERBoixbkV0z4SERERERERERERKRD1oBUREREpMI3NJSIiIiINkeLc/KiBVkRERKQOUOAqIiIiIg2R4tyKaR+JiIiIiIiIiIiIFIh60IqIiIjUAbpqLiIiIiINkeLcimkfiYiIiIiIiIiIiBSIetCKiIiIFJgeniAiIiIiDZHi3PyogVZERESkDlDgKiIiIiINkeLcimkfiYiIiIiIiIiIiBSIetCKiIiIFJhu/RIRERGRhkhxbn60j0REREREREREREQKRD1oRUREROoAXTUXERERkYZIcW7F1EArIiIiUgcocBURERGRhkhxbsW0j0REREREREREREQKRD1oRURERArM4iQiIiIi0pAozs2PetCKiIiIiIiIiIiIFIh60IqIiIjUAbpqLiIiIiINkeLciqmBVkRERKQOUOAqIiIiIg2R4tyKaR+JiIiIiIiIiIiIFIh60IqIiIgUmKGr5iIiIiLS8CjOzY/2kYiIiIiIiIiIiEiBqAetiIiISB1gVs0FejWXJyIiIiJSBdUe50KDi3XVQCsiIiJSBxRZNUeZDSxoFREREZH6qdrjXGhwsa6GOBAREREREREREREpEPWgFRERESkwo4Zu/RIRERERKSDFuflRD1qRajRmzHh23f1MWrTema5rHMCll91NSUlJhflmzZrL8f93Je1X3ZO2HXbn6GOHMn36rGXLS0pKuPa6B9hhp1Pp2HkvOnbeiz32OpuPPx5Tk5sjtWjMrIXs/sY4Wj/xJWs8PZbLvviVktL87tn4z8RZbP3K97R64ks6PzWGfUaMZ97S0nJpnv15Npu9+C0tH/+SjV/4lsd/mlkTmyG1bMzEOex+6Ye0GvAS3U94ncse/paSktznzVc/zWHvyz+i+wmv0/zwl+j1hzf4wx2f88uMheXSLV5SyhWPfce6p46g5YCXWPfUEQx95FsWLan4f5qIiKxcnnjiNQ444Fy6ddubVq12oG/fY3jkkZfKpbnzzifYd9+z6dhxV8z6MWLEqALVVgpp6dKlXHPNfayz7qE0bbYd3dfYj3PPvSlr+nPOuQkr2pLzz791uWVPPz2CTTY9iqbNtqP3mgdy000P1WTVpY6YNGkqrVrtgFk/5s6dv2x+Pv9jnnzyNbbd9gQ6dtyVZs22Zb31DuHKK+9h8eIltbkJIhmpB61INSkuns1ue57NBuv34pmnruWHcZP44+DbKS0t5corTs6Zd8CRl/DNtz9xzz8uoKioiCEX3slBh1zAO2/9DYAFCxZxzXUPcvxx+3DhkGMxM4bd+STb73Qq773zd/r27VMbmyg1pHhxCXu+OZ712zblqR16Mm7uYgZ/+gul7lyxSZecef/5wwzO+mQy56/fiWs27cLMJSW8+es8liYad9/9bR6Hj/yRU9fuyM1btOHFX+Zw9HsTade4EXus3rqmN09qSPHcJexx2UdssEYr/nNhX36YMp/Bw78O583R62XNN2v+Unp3bsGx/bvRtUMzxv86nyse+57//TCKD6/fllUahWu3Fz7wNX9/+SeuOGpdNluzLZ+Om8UlD3/LzHlLueXEDWprM1cq6lggIvXVTTc9RO/eXbn55vNYddV2vPDCSI466mKmTZvJmWcOBOD++5/HzNhzz6155JGXC1xjKZTjT7iC11//mMsuPZE+fXoxceKvjBkzPmPaMWPGce+//kubNi2XWzZy5GgOOXQIJxy/PzdcfxYffvgVQy4YRlFREeecc2RNb4YU0ODBt9KqVQvmzVtQbn4+/2OmT5/Fzjv3ZfDgY2nXrjUfffQlQ4fezZQp0xk2bEhtVH+lpTi3YmqgFakmd/39aRYsWMRTT15NmzYt2R2YPXseQy//J38afEzGwALg/fe/4OVXPuStN+5gxx03B6Bb105ste2JvPbax+y22+9o3rwp4757gvbt2yzLt+uu/Vh3/QEMu/NJ/vXPi2tjE6WG/P376SwoKeXJ7XvSpnEjAGYvKeXyL39l8Pqdls1LN23RUv746S/c2rcrJ67VYdn8g7q3LZfur19NZYdOLbmlb1cAdl6tFWNmLeTKr6aqgbYe+/vLP7JgcQlPDtmCNi0aszswZ8FS/vLodww+eE3atGicMd+2fdqzbZ/2y97336gj3Ts2Y6+/fMznE+awxVrh/HnkncmcsldPzj1wTQB23rgjk6Yv5OG3J6uBtoZYTTw8QUSkFvz3vzez6qrtlr3fZZffMXnyb9x008PLGmjfe+9eioqK+PLL79VAu5J66aX3efTRVxj92UNssMGaFaY/6+wbOfusATzw4IvLLbv8invYfvtNueee8Dtojz22prh4Npdf8U9OO+0wmjTJHAdJ/fbOO5/y0kvvc9FFxzN4cPle1fn8jzn55EPLvd95537Mnj2PO+54gttv/xOm+/BrjOLcimmIA5Fq8uJL77PnHluWa4gdOGA3FixYxFtvfZoj3westlqHZY2zAFtuuQG9e3flxZfeB6BRo0blGmcBmjRpzIYb9Gbq1OJq3hKpbS/9Moc9Vm9driF2QI+2LChx3po6L2u+J34Kw2D8vle7rGkWlZQyYuo8Du9RvtF2QI92fDB9PrMW63b1+uql//3GHpt3KtcQO2D7rixYXMpbX82oVFkdWzcBYHFiaIwlS522Lcpfx23XsjGu2EpERNIkG2dTNt98PaZOLfs+KirST8+V3b3/epZddumXV+Psk0++ztixE7jgguMyLv/ss+/Ybdcty81LNdK+//4X1VJfqVtKSko488zrufTSEzP+z6nq/5iOHdtqiAOpE/QtKVJNvv7mR/qs17PcvB49utCiRTO+/ubHSuUDWL9Pz5z5Fi1azCf/+4YN1u9d9UpLnfDN7EWs17ppuXk9WjahRSPjm9mLsub7aPp81mvdlHvHFdPzmbE0e+wLtnnle96bVtao+8PcxSwp9eXK79OmKaUO387JXr7UbV//PI8+3cr3zO/RqTktmjbim5+zN+ynlJY6i5eU8s2kuVz4wDf8bu22bLlOWbD7f7uvwT9e+YmRY2cwd8FS3hkzg7te/onT91n+/5VUAwsPT6jOSUSkkN577/O8GuJk5fHhh1+x7jo9OOOM62nTdmdatNyBQw79E5Mn/1Yu3YIFC/nj+bdyzdWn07Jl84xlLVy4aLlesk2bhvdjx2YeMkHqt7vu+jcLFy7i9NOPWOGySkpKmD9/Ie+++xm33fYYp556mHrP1qQaiHMb4uHSEAci1aS4eA7t2i1/u3j79q0pLp5TQb5WGfONGzc5a76/XnUfxcVzOPH/9q9ahaXOKF5cQrsmyw9j0L5JI4pz9HCdsnAp38xZxFVjpnLNpl3o0HQVbhj7G/uOmMDX+63Las0aL8ufXn77+D5X+VK3Fc9bQruWy9++175VY4rnVdwLYN8rP+aVT6cB0HetNjx38e8oKiqLdK4+dj0WLC5hx4s+WDbv1L16cMmAdaqh9iIi0pC9/vpHPPPMW9x776WFrorUIVOmTGf4fc+z6abr8OgjVzJnznz+NOR2Dj7kT3zw/r3LGsiuvvo+Vl99VY45Zu+sZa299hp8PKr8A5M/+ugrAGbMmF1zGyEFMX36TC655C4efPByGjde8Wasli13YNGixQD8/vf7cv31Z69wmSIrqkH3oDWzoWbmiWmymf3bzNaq5vX0iuXvlyPNoJhm+Za4Gpbvus1shJk9WVv1aogyXcVx9wqv7mS6WueeeT7A88+P5K9X38e1V5/Gehl630r9k+lIO7mvDJY6zF1ayj+27M5Rvdqz1+qteWqHnjQyuOPb6TnLT92l3hCvPK5MMp437nkNwn/biRvy3rXbct/ZmzJ3YQn7XvExCxMN9jc8PY6H3prMbX/YgDev3JpbT9yAh9+ezGUPf1tt9Zfy1KtAKkuxbuXWrVi3dkyYMJmjjrqYAw/ciUGD1JFAyrg77s4zT1/PPvtsx4ABu/PA/UP56KOveOONjwEYP34SN9z4ILfcfG7OHo2nnHwIzzzzNnff/TTFxbN5+eX3ufGmhwFo1KhBN3OslP785zvZaquN2Gef7aulvPfeu5d33rmHG288h2eeeYszzriuWsqV7NSDtmIrQw/aWcBe8e81gSuA181sQ3ev+B7Q6vM8sA0wvxbXKbWoffvWzJw5d7n5s2bNy9izNpnvt99mLjd/5szMPWs//ngMA466lJNPOpBzzh6wYpWWOqF9k0bMXLJ8T9ZZS0ppl+UBYQAdYi/Y/p3LbnNv07gRW3Roztg4NEKqp2x6+TNTPWtzlC91W/uWjZk5f+ly82fNX5qxZ226dbqG82ardduxwwYdWOuUN3n47cmcsNsaTJu9mEse/pbb/7Ahf9ijBwA7btiBxqsUcdbdX3H6Pj3p3K5pruKlkgwo0sMTpGoU60qdMWPGLPbe+yx69OjCgw9eUejqSB3Tvn0b1lyzKx07lg2ptP32m9GkSWPGjBnPrrtuyQUX3sHee29Lnz69mDkz3IVYWlrKokWLmTlzDm3btsLMOOGE/Rk9+jtOPe1aTjr5Klq0aMa115zBmWfdwGqrdSzUJkoN+OqrH7j33md5++27l50T8+cvBGDWrLk0alRE8+bNKlXmFlv0AcL5t+qq7TjuuKH88Y/HsNZa3au38gIozs3XynBpaam7fxCnh4HjgJ7APrVZCXf/LdahtOLUUh/1WW/5MWMnTvyVefMWZBxjNlc+yDw27bff/sS+B5zPrrv05fZbz6ueikvBrdem6XJjzU6ct5h5S0tZr032RrA+bZpisNxDm9whdaf6Wq2a0Lho+bFsv5m9iCKDdVurka2+6tO9JV//XP6i0MRpC5i3sIT1urfMkiuznp2b06FVY8b/GtpVxk2Zz5Klzma9yz+ccPPebVha4vz424IVq7yIVCfFulInzJ+/kP32O5fFi5fw/PO3ZB07VFZe66/fK+N8d1/2gKdvvvmRp556k/Yddl02TZz4K8PueIL2HXZl0qSpQHiI8rBhg/lt6st8Pvphfp3yEltvvRHAsldpGL77biJLlixlm22Op337nWnffmdOP/1aALp334czz7x+hcpPNdaOHz9phesqsiJWhgbadJ/E114AZra9mb1lZvPNbLqZ3W1my7o7mtnqZnavmY0zswVm9q2ZXWlmTXKtxMz6m9kcM7sqvi9361XiVrEjzOzvZjbLzH42s7+YWVFaWYeb2Xdx/W+a2eYx76BKbvv6ZvZOYjsOrmAb2prZSDMbbWad4rz2Zvaomc2Lt9ENMbMbzGxCJevS4Oy91za8/MqHzJlT1lnlscdfp3nzpuy00+Y58m3NlCnTeffd0cvmjRo1lnHjJrP3Xtssm/fLL9PYc59zWWvNbjzy0OU0aqSejw3FXqu35pUpc5mT6OX6+MRZNG9k7NQ5e0Pbvt1a48CbU8sa6WYtLuF/xQvYpF34UdS0URH9O7fkyYmzyuV9fOIstu7YgrYZxr6V+mGvLTrxymfTmLOgrBft4+/+QvMmRey0YYdKlfXNpLlMn7OEXqu1AEKDLcD/xpUfw+2TceE86tVZP7prglXzJCstxbqKdWvd0qVLOfzwIXz33U+8+OJtdO5cue8hWTnst+/2fP7590ybVnb34Ntvf8qSJUvZdNMwxv09d/+ZN9/4W7lptdU6cMQRu/HmG3+jU6f25cps374NG2+8Nq1ateDOO//NtttuQp8+vWpzs6SGbb/9Zrz55l3lpiFDjgPghRduZfDgY1eo/JEjw+/w3r27rXBdJbvqjnMbYqy7MgxxkK5XfJ1iZtsBrwNPA4cBHYFrgPbxPcCqwAzgPKAYWBcYCnQCTs60AjPbE/gPcI27X15Bfa4D/h3XtytwKfAV8Hgsqx/wKPAkcCawPvBY/ptbzmPAncBVwInAE2bW191Hpyc0sw7Ay/Htzu4+I/49HNgeOBuYApxL2Ccr/ZOGTjn5IG4b9gSHHHYRQwYfw7jxkxh6+T8575yBtGlT1si29nqHs9OOm/PPuy8CYJttNmbPPbbi98dfwQ3XnUFRkTHkwr+x/XabsttuvwNgwYJF7L3fHykunsOwW8/j88+/X1Ze06aN2Xzz9Wp3Y6Vanbx2R4Z9O53D3v2Jwet3YvzcxVz+5VTOWW9V2iSGIFjvuW/YsVNL7t4q3HrTr0MLDujWhpM+msRVm5bQsWkjbhg7jcZFxmnrlN3a9ecNO7PrG+M473+TOaBbG178ZQ4vTp7D8zv1qu1NlWp08p49uf35Hzn02k/408FrMe7X+fzlse8494DetGlRNsTBuqeOYMcNO3DPGZsAMHj4WFYpMrZctx3tWjbm65/ncv1/xrFWlxYM3H51AFZr15QDt1qNC+//mkWLS9i4VxtGj5/NXx77jsO27UKntup5LVKH9YqvinUV69aa0067lhdeGMmtt57PjBmz+eCDL5Yt23zz9WjatAmjRo1hwoTJTJz4KwBvvfU/pk2bSa9eXenXb4NCVV1q0UknHcRttz/G/gecx0UXHs+cOfMYcsEwdtttS7bffjOAjOdCs2ZNWaP7avTv33fZvA8++IJ33x3NZputy+zZ83jk0Zd5+eUPefedf9Ta9kjtWHXVdvTv36/cvAkTfgFghx02p1Wr0MEgn/8xe+11JrvttiUbbrgmjRo1YuTI0dx444MMGLC7hjeQglspGmjNLLWdaxKCtjnAa4TA8D13H5BIO4kwbtdG7v6lu38BnJ9YPhKYB9xrZme6++K0dR0Qy73Y3W/Io3pvu/sf49+vmtlewCGxHjIiwQAAIABJREFUDIAhwFhgoLs78JKZNQaurcQuSLknVSczexkYA1wIDEzbhk6E/TMX2NvdZ8f5GwEHAEe4+xNx3uvAxJh2OWZ2EnASQI8eq1WhyvVH+/ZteP2V2zjjrBvZ/6DBtGvXmnPPHsDQy/6vXLqlS0soKSl/99+jD1/OuX+8lRNOvIrS0lL223c7brvl3GXLf/11BqNHfwfAfgcOLpe3Z88uTPjhqRraKqkN7Zs04pWde3PWJ5M56J0JtGvciLPX7chlG5X/zCwtdUrSxjO4f+vuDBk9hfM//YX5JaVsu2oLXt15zWVjzwJs36klj2/Xg0u/+JW7vp9B75ZNeHCbNdhj9exjI0vd175VY179y5ac9Y8xHHjVKNq1bMw5+/fmsgHrlEu3tMQpKS07b/qu1ZY7XviRu1+dyMLFJfTo1JxDtunCBYeuRctmZWHB8LM24YrHv+f2539kcvFCunVoxkl79ODiI9autW1c2TTEhx1I7VCsu0yBY90uVahyw/DKKx8AcPbZy58S48c/S69eXRk27HHuu++5ZfOHDg0Naccdtx/Dhw+tlXpKYbVp04o3Xr+Ts86+kYFH/pkmTRpz4IE7cvNN51acOU3jxqvw2OOvMvQvd1NUVMQOO2zGyHfvZuONFaesrPL5H/O7323A8OH/ZcKEX1hllUasuWY3rr76dE455bBMRUo1KkSca2b3AvsBU919ozivA+GCbi9gAuE7vzguuxD4P8JF2bPc/eU4vy/hAm5z4AXgbHd3M2sK3A/0BaYDA9x9QsxzHHBxrMqV7n5fhfX19MELGxAzGwpcljb7J+AU4C1gNuFK/d2J5UWEAOwP7n6fmRnhCvpJQG8gOfr0Ou7+vZn1AsYTDthRwHnufkdaXQYB/wJau/vcRJ5j3f3BRLqHgR7uvn18PwF4xN0vTKTpQwhkj3f34Xnsh9S6N3D3sYn51wKHu/ua8f0IoBGhd8UUYP/kwyUS5TR394WJ+Y8CW7t7r1z16NdvfR/14b0VVVeknJJjMnbeEamQDVij0FWQeqrRwS984u79Kk5ZfdYvaub3N6venhtbLvih1rdDapdi3eXWXeBYdwMfNeqBiqorUp6GbZaqspVxxEqpDmb9ajVGrIk4FyqOdc1sR0LMc3+igfY6YIa7X2NmFwDt3X2ImW0APAJsCXQlXMhd191LzOwjQqz0AaGB9jZ3f9HMTgM2cfdTzGwgcLC7D4iNwKOAfoAThp/qm2oIzmZl+ETPAn5H2DHdgV7u/iLh1q5GhF4GSxLTIqAxkPplfw5wI+E2rgMJB+v0uCz9UYEHEG4R+08l6jcz7f3itHK7AL+lpUl/n6+pGd6vnjZvA8KtZQ9kePJvF2BOMmBdwfqIiIiIyIpRrFtGsa6IiIgA4O5vE+KWpAOBVG/W+4CDEvMfdfdF7j4e+B7Y0sxWB9q4+/vxTp/70/KkynoS2DVe+N4TeNXdZ8RG2VeBvSqq78owxMFS9/9n777j5Krq/4+/PpuQBmkQSiBAQieAtEgXkBZQSqQGpIOCiiIoVRSkKCjNiIB86RZAihRN6MIPkBYVpEgNgYSEkkIS0rN7fn/M3WW2ZGezmdnZnX09H495ZOfcc++eGcLkvWc/95w0ton2z8jNZJ9Hbga8oUnZnwcBd6aUflJ7IJtZb8r3ya3f9UhE7JhSmtrqUX/hI3JrgOVr+LylViJXdp3/fHKDPv8A/gNcFxFTUkoPNBhL74jo0SC4tnY8kiQJiEhEVO5dTSops+4XzLqSJLUzJcy5AyIiPwNdl1IqtBD1yimlyQAppckRsVLWvhq5CtlaE7O2hdnXDdtrz5mQXWtRRMwgd5dOXXsT5yxWZ6igbVL2G/PngPVTSmObeNSG1p7kKg3yfXMxl51JbqY8AQ9FRJ8iDPVFYJ9sFr7Wvq28Vt1OttnuufsBLzTslFK6iFwlxZ0RsUveodq/+HXfPyJ6Aru3cjySJEkqAbOuWVeSpAo3JaU0LO+xNLsENrVKbmqmvbXnLFZnqKBtzunkNkmoIVeOPAtYA/g68JOU0lvkSpF/EBHPA++SC6yLXXk8pTQ1InYHngL+FhF7ppTmLMUYLwGeB26PiJvI3ZL1rezYki5YdHxELABeza6xDnBoUx1TSmdGRG/gvojYPaX0XErp1Yh4ALgmO/YRuSqKOa0YiyRJylPlJmEqPrOuWVeSpLJrRzn344gYmFXPDuSL5ZEm8sXyT5BbNmpS1j6oifb8cyZmG7b2JbekwkRg5wbnPFFoYJ22ghYgpfQ0sCO525b+ADxALshOAD7Oup1PbqHgC7M/FwA/KHDdycCu5HaFuyciui3FGMeSC5ZbAvcCBwDfyQ7PXMLLjSRXWXAvsCm5Heb+00z/k4C7gTERsWnWdjS5xZJHATeS24DiwVaMRZIk5Yko7kMy65p1JUlqD4qdc5ci694PHJV9fRRwX177yIjoHhFDgHWBF7LMMysitsnu9jmywTm11zoQeDxbp/YhYI+I6B8R/YE9srbm36PcuepIIuJwciF7rWzx4nKOpSu5KoXnU0pHNdd32LAN09jnb2ybgaliVB9+QrmHoA4qDlm9cCepCV2+MbpNd7YFGNqle/pTr+LubrvF5+Pa/HVIxdBxs+7QNHbsH9pmYKocyeJstVJ06no7LYWIYW2aEUuRc6Fw1o2I28hVsg4g94vpc8n9Evcv5O4o+gA4KKU0Lev/E+BYYBHww2zTVSJiGHAzuWWhxgDfTymliOhBLq9sTq5ydmRKaVx2zrHA2dlQLkop3VTo9XT2JQ46hIi4htztZ9OBLYBzgL+XI7BGxEHAqsArQB9yt4+tS+63CJIkqZWi8NJUUkUy60qSVNnKkXNTSk0uc0TuLqCm+l8EXNRE+1hg4yba55HbbLWpa91I7k6cFnOCtmNYAbg6+3MqcAe529PISqy7NHNudSpumfRs4Bhya3p1IRde90kpNdqAQZIkSWoBs64kSerUnKDtAFJKBzdz+CiguVLpY8iVYhdrLKOB0cW6niRJym31Wo51YyNiT+A35Cairk8pXdzgeF/gj+RuA+sKXNqSW7SkJWHWlSSpcpUr53Y0TtB2fA8AX27meFnX7ZIkSS3T1sE1IroAvwN2J7fb7IsRcX9K6fW8bt8DXk8p7RMRKwJvRsSfUkoL2na06sTMupIkdXBO0BbmBG0Hl1KaSu5WMEmSpCWxFfBO3mYGtwP7AfkTtAnond1mvhy5DRAWtfVA1XmZdSVJUmfgBK0kSVK5BVRF0TdPGBARY/OeX5dSui7v+WrAhLznE4GtG1zjKuB+YBLQGzgkJbcclyRJUguVJudWHCdoJUmSKtOUlNKwZo43dbNZw/Q8HHgJ2AVYG3gkIp5KKc0s0hglSZKkTq+q3AOQJElSbm2uYj5aYCKwet7zQeQqZfMdA9yTct4ht97nBsV4vZIkSeocip1zK3FNWydoJUmSyixK8GiBF4F1I2JIRHQDRpJbziDfB8CuABGxMrA+MK5VL1KSJEmdTilybgXOz7rEgSRJUmeUUloUEScBDwFdgBtTSq9FxInZ8WuBC4CbI+IVcln4jJTSlLINWpIkSapATtBKkiS1A1GGzRNSSqOB0Q3ars37ehKwR1uPS5IkSZWjHDm3o3GJA0mSJEmSJEkqEytoJUmS2oFK3OxAkiRJMucW5gStJElSO1BlcJUkSVIFMucW5hIHkiRJkiRJklQmVtBKkiSVWYSbJ0iSJKnymHNbxgpaSZIkSZIkSSoTK2glSZLaAZfmkiRJUiUy5xbmBK0kSVI74O62kiRJqkTm3MJc4kCSJEmSJEmSysQKWkmSpLJLbp4gSZKkCmTObQkraCVJkiRJkiSpTKyglSRJKrMAqlybS5IkSRXGnNsyVtBKkiRJkiRJUplYQStJktQOuLutJEmSKpE5tzAnaCVJksotDK6SJEmqQObcFnGJA0mSJEmSJEkqEytoJUmS2oEglXsIkiRJUtGZcwuzglaSJEmSJEmSysQKWkmSpHbAtbkkSZJUicy5hTlBK0mSVGYBRJXJVZIkSZXFnNsyLnEgSZIkSZIkSWViBa0kSVK5BYS/NpckSVKlMee2iG+RJEmSJEmSJJWJFbSSJEntgJsnSJIkqRKZcwtzglaSJKnsAtw8QZIkSRXHnNsSLnEgSZIkSZIkSWViBa0kSVK5uXmCJEmSKpE5t0V8iyRJkiRJkiSpTKyglSRJagfC3RMkSZJUgcy5hTlBK0mSVGaBt35JkiSp8phzW8a3SJIkSZIkSZLKxApaSZKk9sBbvyRJklSJzLkFWUErSZIkSZIkSWViBa0kSVK5hWtzSZIkqQKZc1vECVpJkqR2IKq89UuSJEmVx5xbmHPYkiRJkiRJklQmVtBKkiS1A+6dIEmSpEpkzi3MClpJkiRJkiRJKhMraCVJksrNzRMkSZJUicy5LeIErSRJUnvg5gmSJEmqRObcgpzDliRJkiRJkqQysYJWkiSpzAI3T5AkSVLlMee2jBW0kiRJkiRJklQmVtBKkiSVWwTh2lySJEmqNObcFnGCVpIkqR1wd1tJkiRVInNuYb5FkiRJkiRJklQmVtBKkiS1A+HuCZIkSapA5tzCrKCVJEmSJEmSpDKxglaSJKncAn9tLkmSpMpjzm0RJ2glSZLKLADv/JIkSVKlMee2jHPYkiRJkiRJklQmTtBKkiS1A1EVRX1IkiRJ7UGxc25Lsm5EnBIRr0XEqxFxW0T0iIjlI+KRiHg7+7N/Xv+zIuKdiHgzIobntW8ZEa9kx0ZFtuNZRHSPiDuy9ucjYvDSvEcucaA2FNClW7kHoQ7mktvnl3sI6qDO/uM55R6COqzR5R6AJKmT+HnV4eUegjqocxfdXO4hSO1WRKwG/AAYmlKaGxF/AUYCQ4HHUkoXR8SZwJnAGRExNDu+EbAq8GhErJdSqgauAb4NPEfuB4U9gTHAccD0lNI6ETESuAQ4pLVjtoJWkiSp3AKiqrgPSZIkqexKkHNbmHW7Aj0joivQC5gE7Afckh2/BRiRfb0fcHtKaX5K6T3gHWCriBgI9EkpPZtSSsCtDc6pvdZdwK611bWtYQWtJElSe+DuCZIkSapEbZxzU0ofRsSlwAfAXODhlNLDEbFySmly1mdyRKyUnbIauQrZWhOztoXZ1w3ba8+ZkF1rUUTMAFYAprRmzNZXSJIkSZIkSepIBkTE2LzHt2sPZGvL7gcMIbdkwbIR0dyaMk3NIKdm2ps7p1WsoJUkSWoHXJZAkiRJlahEOXdKSmnYYo7tBryXUvoUICLuAbYDPo6IgVn17EDgk6z/RGD1vPMHkVsSYWL2dcP2/HMmZsso9AWmtfbF+KOAJEmSJEmSpErxAbBNRPTK1oXdFfgfcD9wVNbnKOC+7Ov7gZER0T0ihgDrAi9kyyHMiohtsusc2eCc2msdCDyerVPbKlbQSpIklVtAVLkGrSRJkipMGXJuSun5iLgL+DewCPgPcB2wHPCXiDiO3CTuQVn/1yLiL8DrWf/vpZSqs8t9B7gZ6AmMyR4ANwB/iIh3yFXOjlyaMTtBK0mSVGaBe4RJkiSp8pQr56aUzgXObdA8n1w1bVP9LwIuaqJ9LLBxE+3zyCZ4i8ElDiRJkiRJkiSpTKyglSRJKrtwiQNJkiRVIHNuS1hBK0mSJEmSJEllYgWtJElSuUX2kCRJkiqJObdFijpBGxE9gG4ppZnFvK4kSVKlC+9ravfMupIkSUvOnFtYsd+ia4BpRb6mJEmS1B6YdSVJklR0pVjiwMJlSZKkJeTmCR2G/6EkSZKWgDm3MIuMJUmSJEmSJKlM3CRMkiSp3ALCwgJJkiRVGnNuizhBK0mSVGaBt35JkiSp8phzW8YlDiRJkiRJkiSpTKyglSRJag/8tbkkSZIqkTm3oGYnaCOiuq0GIkmSJLUls64kSZLag0IVtK1ZJCK1ZiCSJEmdVgCuzVUOZl1JkqRSMue2SLMTtCkli5AlSZLagqmrzZl1JUmS2oCJqyDfIkmSJEmSJEkqEzcJkyRJKjdv/ZIkSVIlMue2SLMVtBFxY0Ts26BtvYZteceOiojHizlASZIklUZE7BkRb0bEOxFx5mL67BwRL0XEaxHxZFuPsZTMupIkSWoPCi1xcDSwWYO2Q4G/Lqb/YGCnpRuSJElSJ1RV5EcBEdEF+B2wFzAUODQihjbo0w+4Gtg3pbQRcNDSvch252jMupIkSaVV7JxbgQu2usSBJElSuUWU49avrYB3UkrjckOI24H9gNfz+hwG3JNS+gAgpfRJWw9SkiRJHVh5cm6HU4FzzpIkSQIGRMTYvMe3GxxfDZiQ93xi1pZvPaB/RDwREf+KiCNLOWBJkiSpM7KCVpIkqT0ofmXBlJTSsGaON/UNU4PnXYEtgV2BnsCzEfFcSumtIo1RkiRJlc4K2oKcoJUkSeqcJgKr5z0fBExqos+UlNJsYHZE/D9gU8AJWkmSJKlIXOJAkiSpPWj7jRNeBNaNiCER0Q0YCdzfoM99wFciomtE9AK2Bv7X6tcoSZKkzsdNwgpqSQXtZg3WG9sMICKOoPGtcQ13wZUkSVIhQZvf+pVSWhQRJwEPAV2AG1NKr0XEidnxa1NK/4uIB4H/AjXA9SmlV9t0oKVn1pUkSSqVMuTcjqglE7T7ZY98AdzcRN+g8dplkiRJaodSSqOB0Q3arm3w/NfAr9tyXG3MrCtJkqSyKjRBe0ubjEKSJKmzq8BbtToAs64kSVKpmXMLanaCNqV0TFsNRJIkSWpLZl1JkiS1By1Z4kCSJEml5NpckiRJqkTm3BZxglaSJKk9MLdKkiSpEplzCyq4CkTk7BgRB0TEWnntm0XEmIiYGhHTI+KeiFi3tMOVJEmSisesK0mSpHJrtoI2InoCY4CvZE01EfF94AngSaB3XvcRwHYRsVlK6aMSjFWSJKlyeetXmzPrSpIktQFzbkGFKmh/COwITAT+CkwCfgWcAywAvgV8CdgpO74ScFqpBitJkiQVkVlXkiRJZVdoDdoDgcnAJimlmRHRF3gNOBQ4PKV0W23HiHgmO7Yn8KMSjVeSJKnyRFhZUB5mXUmSpFIy57ZIoQradYEHUkozAVJKM4C/Zcceze+YUqoBHgcGF3mMkiRJla+qyA+1hFlXkiSp1Iqdcysw6xZ6ScsBDdfY+hggpfRpE/0/AXoUYVySJElSqZl1JUmSVHaFljgAqCnwXJIkSUvLW7/KxawrSZJUSubcgiqwKFiSJEmSJEmSOoaWVNCOiIjBec83A4iIG5vou3kRxiRJktS5BIS/Ni8Xs64kSVKpmHNbpCUTtJtlj4aOXkz/1OrRSJIkdVbe+lUuZl1JkqRSMucWVGiC9pg2GYUkSZLU9sy6kiRJKrtmJ2hTSre01UAkSZI6rcCdAcrArCtJklRi5twWKepbFBFVEbFfMa8pSZIktQdmXUmSJJVCS9agLSgi1gSOJ3eb2ECgSzGuK0mS1Gm4Nle7ZdaVJElaCubcglo9QRsRXYD9gG8Du5Grxk3Ao8UZmiRJUidicG1XzLqSJElFYs4taIknaCNiLXIVBEcDK2fNU4DfAzeklN4v2ugkSZKkNmTWlSRJUltr0QRtRHQFvkGuguCr5CoIFgD3AAcA96WUflaqQUqSJFU0N08oK7OuJElSiZhzW6TZCdqIWBf4FnAUMIDc2/pv4GbgzymlaRFRU+pBSpIkScVm1pUkSVJ7UKiC9k1ya219AlwB3JRSeq3ko5IkSepUwrW5ysOsK0mSVFLm3JZoyRIHCRgN3GVglSRJKgFv/Sons64kSVKpmHNbpNBb9FPgfeAY4JmIeD0iTo+IgaUfmiRJklRSZl1JkiSVXbMTtCmli1JKawN7AX8F1gYuBj6IiL9HxMFtMEZJkqTKVxXFfaggs64kSVIbKHbOrcCs26Ii45TSQymlA4HVgbPJVRrsBdxG7rawzSJiy5KNUpIkSSoRs64kSZLKaYlWgUgpfZJSujiltA6wO3AXsBAYBrwQEf+JiO+VYJxSh/D66+PYddfv0KvX9qy66p787GfXUl1dXfC8GTM+55hjfk7//l+lb9+d+OY3z2Hq1M8a9bvvvifYZJND6NFjO4YOPYg77ni4FC9DZfAJ87mBCZzH21zMuzzKFGpIBc/7mPncxETO420u4h3u42PmU3/D8UUkHmcql/Ee5/I2l/EejzKFRbgxeUf3+uvvsevu36dX76+y6ur78rNz/6/lnznHXUj/AcPpu/zufPOI85g6dUbd8erqai751R/4yk7fYYWV9mSFlfZkjz1P5sUXXy/ly1FVkR9aYmZdqXVuvvkBIoY1elx77V0ALFiwkIMPPpO11tqPnj23Z8UVd2OvvX7Av/71vzKPXK0xlQU8wMdcw3h+zlvcxIR6x2exiIf5lGsYz0W8zeWM4698xEwWNbrWJ8znFiZwIW9zKe/yeBMZeBaLuJePuIx3uYi3uZb3+S8z6/V5jVn8mQ/r+vye93mlQR+AF/iMP/Ehl/AO5/EW7zGnCO+IiunOux5n3xGns9oa+7Jc313ZcqtjuO32L37unTx5CqedfhWbbnEky/XdldUHj+CoYy5g0qRPG13r3vue5EubH0H3XjsxZJ0DuPyK2xr1SSnxi1/ewuqDR9BzuZ3Zcefv8NJLb5X0NXZKxc65FZh1W7JJWJNSSo8Bj0XEAOBo4DhgU2AU8LuijE7qQKZPn8luu32XoUPX4r77LuPddyfyox9dSU1NDRde+N1mzz3kkLN48833uf76c6iqquKMM37LiBE/5qmnrq/r8/TTL3HAAWfw3e8eyKhRpzF69DMceuhP6N+/D3vssU2pX55KaC7V3MhEVqI7h7MqU1nIGD4lAbszYLHnzaOaG5jIALpxCAOZSzUPMoVZLOJwVqvr9zCf8gIz2I0BrEp3JjGPR5jKPGrYm5Xa4BWqFKZPn8luw09m6IaDue+eS3h33If86LTf5j5zLjih2XMPOfSnvPnWB1x/3Zm5z5yzrmbE/mfy1JPXADB37nwu/tUfOeaor3HWGUcQEVx19V3ssNN3+OdTv2fLLTdoi5fYuQQVeatWR2bWlZbc449fS8+e3euer7VWLo9UV1cTEZx11tGsvfYgZs6czRVX/JlddjmR//znT6y11qByDVmt8CkLeJvZDKIHTf1aeBLz+B+fswV9GUQPPmcRTzCVG/iA7zKY7tnMylyquZWJrEh3DmVVprGQh7MMvGuWgWtI3MaHzKGG3VmR5ejK68ziHj5iGYIN6Q3As0ynH8swnJXoRRfeZjZ38xFzqGZr+teN7WVmEsDaLMurzCrxO6XWuPyK2xgyZFWuuOxkBqzQl9FjnuWww89jypQZfP+kg/jXv97kr/c9yfHH7sPWW23Ex59M47zzb2C7r5zAqy//keWW6wXAM8/8l/0PPJtjj9mbSy85iedfeJ0zzrqaqqoqfnjyIXXf7+JL/sAFF93Ery85iQ3WX5PLr7yd3YafzKsv/5FVVlmhXG9DZTHntkirJ2hrpZSmAJcCl0bETsDxSz0qqQO69tq7mTt3Pvfc8yv69FmO3XeHmTNnc95513H66UfSp89yTZ737LP/5aGHnuXJJ69jxx23AGC11VZk662P5tFHn2e33bYG4IILrmfHHTdn1KjTAPjqV4fx2mvvcv75/+cEbQf3AjNYSOIwBtKDLqwDzKeGx5nKV+hPD7o0ed7zzGARiSNYlZ5Zn5504Y9MYiLzGEQPAF5mFlvRjx2ycLoWvZjJIl5ilhO0Hdi1v78395lz1y/p02dZdif7zDn/Bk4/7XD69Fm2yfOeffYVHnr4eZ58/HfsuOPmAKy26opsvd3xPProi+y225fp2bM7496+k/79+9Sdt+uuw1hvw0O46uq7uOmGc9riJUrtgllXarkvf3lo3eRIvp49e3DHHb+s17bbbluxwgq7cu+9T3DqqYe31RBVBOuxLBuwFgB3MIk5DaZp16AnJzGYLnwxITOQHlzFeP7HLDajLwBjsyx7SJaB1yaXgZ9gKttnGXgqC5nEfA5lVdYn9/PUWvRiIvN4lVl1E7SHshrL5mXmtejFLBbxLNPrTdAex+pUEXzMfCdo26kH7vs1Awb0q3u+yy7DmDR5CpdfeTvfP+kgdtjhS7zx2m107frFdNYWm6/P+kNHcvc9T3DUkV8D4PwLb2SH7b/E9dedBcAee2zN9OkzOf/CG/nud/anW7dlmDdvPhf/6g+cdcaRnPS9AwHYdtuNGbz2AVz1u7sKFj1IxVTsouCDga8X+ZpShzBmzD8ZPnybehOxI0cOZ+7c+Tz55L+bPW/llVeom5wF2GqrjRkyZDXGjPknAPPnL+Af/xjLwQfvVu/ckSP34NlnX2HGjM+L/GrUlt5iNuuybL2J2C/Rm4Uk3mPuYs+bzDxWo3vd5CzAOvQigDeZXddWTaJHg4/7xU36quMY8+CzDN9jq3oTsSMP2S37zPlPM+c9x8orL183OQuw1VZDGTJkVcY8+CwAXbp0qTc5C9Ct2zJsNHQIn3wyvcivRHW87asjMOtKRbLssj3p0aMbCxY0vu1d7VsVzVfC9aRLvclZgAF0YxmC2XmTuW8zm7UbZOCN6c0iEu9nGbh2uYPujbJsVb2FEJZtItsOpHu979eSsav88idna22+2Xp1GbRfv971JmcB1ltvDXr16lEvp7708tvstuuX6/XbY/etmT59Fs8++yoA//znK8ycOZuDD9qlrs+yy/Zkn723Z8xDzxXtNQmXOGiBYr+kXpD9OkzqZN54YzwbbDC4Xtsaa6xCr149eOON8QXOW7NR+4YbDq477913J7Jw4aJG199wwyHU1NTw1lvvL+XoVU6fsoAVWaZeWz+WYRmCKSxY7HkLSY3CbxVBZNesNYy+vMBnvM9c5lPDeObwPJ+xDY3DjzqON958nw3Wr//ZUfeZ8+biPxOaOg9gww3WbPa8+fMX8K/JGAVuAAAgAElEQVR/v8nQDYe0ftBSx2fWlZqx9toj6Np1a9Zff39+//u7Gx1PKbFo0SI++mgKp58+ii5dunDoocPLMFK1tY+Yz0ISK9Ktrm0KCxhQIAOvRDdWowf/YCpTWcA8qvkPM5jAPIYVyLITmFvv+6nj+uezrzB06ODFHv/vf99hzpx5DN3wiz7z5i2gW7f6E7ndu+f+vv0v+zn7jTffp0uXLqy77ur1+m24wWDeeMOfsdW2lnqJA0k506fPpF+/3o3a+/fvw/Tpi799prnzxo37sK4P0KhfbYVbc9dX+zeX6iYrWnvShbnNbOS1At14mZlU503UTmIeNdk1aw1nAItIXJe3gcPW9GUXXFOpI5s+fdZiPjt6F/jMmUW/fo2XXOnfvzfjxk1a7HkX/eIWpk+fxfHH7dO6Aat5rs0lqQMbOHAAF1xwIltttTHV1dXcdttDnHjiL5kzZx6nnPLNun6XXHILZ511FQArrtif0aN/w5prDizXsNVGakg8yCcszzKszRd3/sxbTAbukZeBg+BwVuM2JvFbxgO5KrMRrMJaNF5Oo9Y45vAGs9mPlYv6WtT2HntsLPfd/xQ3Xn92k8dramo4+ZQrWXfd1dljj63r2tdZZxAvjn2jXt8Xsg1vp03L/Xw9ffoslluuJ1261P972L9/b+bMmceCBQvp1q3+LxHUCubcFmk3RcERsX9EPB4Rn0XE/Ih4KyIuzDZmKOb3WS8izouIfg3aj46IFBFNLxTagbX0tUXEExFxV1uNqxJFNP7QSSnRRHOrzmvYL6WUtS/ZONX+NPWfMDXYwbahL9OXOVTzAJ8wi0V8zHzu5xOqGlzvKabzEjPZmxU5nkHszYq8zCweZUoxX4LKoKn/91v/mdN0O8Df//4MF/3yFi755XdZv4nqWxVJVRT3oXbFrFs6Zt3yGz58W84553j22GMb9tpre2699XwOPnh3LrzwRmpqvvhl89FH782LL97K/fdfzpZbbsjee/+Q118fV8aRqy08xhQmMI/9WaXR3V9N/2uV6tprSNzDR8ylmgMZyFEMYhv6cx8f83bekl75prOQu5nMBizL5t700KGNHz+Zw444l/32/QpHH9X0CkNnnX0Nzz73Kn+4+Wcss8wXNYgnfnsE993/FP93/X1Mnz6Thx56jsuuuA2ALl2+mApb3M/iizumVip2zq3ArNsuJmgj4jLgTmAccASwB3AFsA/wf0X+dusB50Kj+yH+DmwLzCny91Mn0b9/Hz77rHHV2owZnzdZ5VbovM8++6I6rrZStmG/2ufNXV/t3+IqZedTQ89mPqZXpBv7sTL/ZRYXM47f8j6D6MEqdGe57AaJ2VTzKFMYzgC2pT9D6MW29Gc4A3iSaXyO6751VP379+azzxqvPz1jxuwCnzlNn5f7zGk8t/Hii69zyGE/44Rv71dvx1tJLWfWVWd04IG7Mm3aDMaP/+LujFVWGcCwYUPZZ58deeCBy1lhhb5cfPHN5RukSu4FPuMZpvMNVmEQPesd60GX7N6v+uZRU7d/wlvM5m1mM5JV2ZjeDKEXe7AiG7Icj/Bpo3PnUM2f+JC+dGV/rM7uyKZNm8lee5/KGmuszB9vPbfJPldfcze/vuzP3HLTOWy99Ub1jh17zN6ceMIIvvO9S1l+xT3Z/6Cz+dk5xwKw8srLA7lcPGvWHKqr669V/Nlnn9OrV496E75SqZX9b1tE7AOcChyXUrox79CTEXEduQBbcimlT6GJT3iphTbYYHCjtWYnTPiI2bPnNlo7tuF5Tz3VeEOfN94Yz4gROwOw9tqDWGaZrrzxxnh22mnLen2qqqpYbz0r2jqyFenWaK3Zz1jIAhIDCqybNYy+bEpvprKQ5ehCL7pwEe8yLKsWmMYCqsntnJtvIN2pAT5jUd1krjqWDdZvvGbshAkf5z5zmqly3WD9NXnq6Zcbtb/x5vuM2HfHem1vvfUBX9/3x+y6y5b89jenFmfgWrx28WtzFZtZV53d4irQunbtyiabrFO3pJcqz+vMYgyfsDsD2JjGvzwe0EQGnsFCFuZl4CksYBmCFRpk4lXozpvU/4XzAmr4Mx9STeIwBtHNf1g7rDlz5rH3vj9mwYKF/P3+37Hssj0b9bn7nn/w/ZOv4FcXf49DGmymDblNb68a9SMu+Pm3mDjxU4YMGVi3ruw22WTuBuuvSXV1Ne+8M7HeXWJvvPl+k/vEaCn4v2NB7eEtOgX4d4PACkBKqTqlNAYgIgZExC0RMTUi5mS3KA3L7x8R4yPi0og4JSImRsT0iLi99haviNgZeCDr/l52K9T47Fi9W6MiYnD2/OCI+H1EzMiu+fOIqMr7njdHxNgG46g9d++mni/u3Ox2tCkRsXVEjI2IuRHxdEQMiYiVIuLeiPg8Iv4XEbuw5DaMiKey674VEd9ornNE9I2IZyLi5YhYMWvrn72nsyNiUkSckb3n41sxnoqy117b8dBDzzFr1he32txxxyP07NmdnXbaotnzPvpoKk8//VJd29ixrzNu3Ifstdd2AHTv3o2vfnUYd975WL1z77jjEbbddhP69q24uxU7lfVYlreZzfy8CoJXmMUyBENoHEYaWoaquqrZl5hJAjbJQnD/bOOFScyrd84k5mfHnZztqPbac1seevj5+p85f3ks+8zZvJnztsk+c76YpB079n+MGzeJvfbctq5t8uQpDP/aKay91mrc9qfzG63NJanFzLpfPDfrdiJ33/0YAwb0W+was/Pmzeff/36TIUNWbeORqS28xxzu5iO2oh/bs3yTfdZlWd5pkIFfZRZdCdbMMnA/lmEhqdFE7mTm0S9vg7FqEncymWks5JusZgFCB7Zo0SIOOuQnvP3ORMb87XJWWqnx358nnvg33zzi55z0vQP48Y8Oa/Z6/fv3YZNN1ma55Xpx9bX3sN22m9QVUG233Sb06bMsd971eF3/OXPm8cDfnmav4dsU9XWpPCKiX0TcFRFvZPli24hYPiIeiYi3sz/75/U/KyLeiYg3I2J4XvuWEfFKdmxURO63jxHRPSLuyNqfj4jBrR1rs59aEbGkCwIt0RpaEbEMsB1wWQu63wusA/wYmAKcBvwjIjZPKb2T1+9g4L/At4FBwOXAL4DvAv/Ozr8U2B+YDNksxeL9CrgbOBDYFfgZ8BrwlxaMuTV6Addl33c2MAr4QzbOMcDVwOnAnRGxekppSW5TuyM7/xfA8dk1tkwpNSqliojlgYeyp19NKU3Lvr4Z2AE4GfiI3A8d6wHVDa/R2Zx44gGMGnU7++9/GmeccRTjxn3Ieeddx6mnfpM+fb6YQF1nnRHstNMW3HDDzwDYdtsvMXz4thx55LlceunJVFVVccYZo9hhh83YbbcvFjn/6U+PZ+edT+CHP7yMESN2YvToZxg9+hkefPC3bf5aVVxb0Zd/Mp0/MYkd6c80FvI4U9me/vU2TriM9xhCT/ZnFSC3scITTGMwPelCMI45PM10vsHK9MrOW46ubMiyPMQUFpFYhe5MZj6PMZWNWY5lDa8d1oknjGDUVXey/4Fnc8ZphzPuvQ857/wbOPWHI+nT54sNONZZ/yB22nFzbvi/3MYK2267CcP32Jojj7mAS391ElVVwRlnXcMO22/Kbrt9GYC5c+ez194/Yvr0WVz1m1P573+/+Ge2e/dl2Hzz9dv2xXYGUZlrabV3Zl3ArAtm3aV2wAGnsdVWG/GlL61LdXUNd9zxMHfc8QijRv2YqqoqbrvtQcaM+Sd77rkdq646gMmTp3D11XcxefIUTj31m4W/gdqVBdTUrf86i0XMp4bXyC29ti7LMoOF3M4kBtCNjejNBObWnbssXVg+q4YdRl+eZzp3MInt6c90FvIEU9k2LwOvy7L0pSu3M4mdWJ5edOFtZvMan/M1Vqq77t/5hLeZzZ6syFyq633PgXSna1ab9iHz+IyFzMyW+Xqfucyhmn4sw2oN7jhTeXz3pEsZPeZZfnPFD5k2bSbPPfdq3bHNN1+PceMmMeKAM9lggzU55KDd6h1fccV+rL32IACee+5Vnn7mv2y26brMnDWb225/hIcefp6nn7y2rn+PHt058/QjuOCim+jfvw8brL8ml195GzU1ie+fdFDbvehKV96c+xvgwZTSgRHRjVwOORt4LKV0cUScCZwJnBERQ4GRwEbAqsCjEbFeSqkauIZc9noOGA3sSS63HAdMTymtExEjgUuAVq0LV+gn88GtuGbzu9rUtwLQHfiguU4RsSewPbBzSunJrO1xYDy58HpCXveFwIiU0qKsX+0b/N2U0syIeDPr95+U0vgWjPH/pZR+lH39SDaW/SldaO0J/CDvda4K/A44N6V0adY2kVxw3oncX4iWuj7vGg8BrwNnkXt/6mQVBI8CnwN7pZRmZu0bA/sCB6eU7szaHgMmZH0biYhvk/tLzBprrLIEQ+14+vfvw2OPXcNJJ/2KffY5lX79luOUUw7jvPO+Xa/fokXVVFfXX2vp9tt/wSmnXM6xx55PTU1i7713YNSo0+r12WGHzbjrrks455yrueaauxgyZFX+/OcL2WMPf7PX0fWkC8cxiAf4hD8wiR5UsR392ZUV6vWrIdVbpauKYDLzGcsMFpJYmW4cyqoMpX5F9YGswj+YxrN8xkwW0YeubEVfvtrg+upY+vfvw2MPj+KkH1zGPiNOo1+/3pxy8iGcd+5x9fo1+Znz5/M55Ue/4djjf0FNTQ17f317Rl15St3xjz+exssvvw3A3vvV/yxac81VGP/uPSV6VZ1ce7ivqfMZ3IpzzLpLx6xbgdZffzA33ng/EyZ8TEowdOgQbr315xxxxNfrjv/xj2M49dQrmD59JgMHDmDrrTdm7Niz2Gijtcs8ei2p2VRzJ5PrtdU+P5khTGQe86nhY+ZzIxPq9duUPnwjKzboSReOZBCj+YTbsgy8Lf3ZOS+jdqeKIxnEY0zhIT5lPjUszzLszUpsmbcB2LvZhPGDTazkcjJD6J/9I/sCn/EyM+uOPcHURuNSeT38yAsAnHzKlY2OvffO3Tz/wmvMmPE5L7/8NtvveEK940cd+TVuvvEcAJZZpit33Pko551/A1VVwVd22JRn/t/v2WST+p85Z55xBDU1NfzykluZOnUGw7bckEcevLJunVoVSRlybkT0AXYEjgZIKS0AFkTEfsDOWbdbgCeAM4D9gNtTSvPJ3Yn0DrBVdhdNn5TSs9l1bwVGkMso+wHnZde6C7gqIiLV7jS3BApN0A5Z0gu2UqGBbwV8WhvkAFJKsyPib+R+w53vH7WBNfM6sFJEdMv+Yyyphxs8fx1YoxXXaakFwFN5z2srJh5vom21Jbz2X2u/SCnVRMR9QMNfC60MPEmuYmCflFL+1pi1t9nV3jpHSmluRDwKNDlLmFK6jlyVBMOGDV3iv6AdzdCha/H449c222f8+AcatfXr15ubbjqXm25qevHzWiNG7Fy3Lq0qy0p05zhWb7bPaaxV73k3qjiGQQWv3YMu7MWK7MWKSzVGtT9Dhw7h8UevarZPU5Op/fr15qYbzuGmG5o+Z/DggaRF/yzGEKX2zqxr1jXrFsEvfvE9fvGL7y32+BZbbMDf//6bNhyRSqk/y3Ae6zVzvC+b502eNmclunN0gQy8At04mOaXwjilQU5enG+wihOx7VyhQoCjB3+do4/6esHrbLnlBrz4XKPVhRqJCH5y9tH85OyjWzpEtR8DGizDdF3273Kttcitv39TRGwK/IvcHTIrp5QmA6SUJkdEbTn+auQqZGtNzNoWZl83bK89Z0J2rUURMYPcL+inLOmLaXaCNqX0fnPHi2AquduZCoXAgcDHTbR/DI0WtPmswfMFQADdsq+XVFPXK+W9D7NSSvmlTrVjrhtHSmlBttzFko7jkyaeN1wUaii59/TXDQIrwCrZ+OY1aHfDCUmSltZiNtJR6Zh1F3s9s259Zl1JkpZGaXLulJTSsGaOdwW2AL6fUno+In5DbjmDxWlqkKmZ9ubOWWJlvZkupbQQeAYYXqDrZMhbYOYLKwPTmmhvS/Og0TbrDYN0bcgr1K/UGr6HK0GDe1PgH8A5wHWR23U430dA74hoGJYty5MkSWrArGvWlSRJZTMRmJhSej57fhe5CduPI2IgQPbnJ3n980v6BwGTsvZBTbTXOyciugJ9aWV2aw+rnV0JDIuIoxoeiIiqbB2s58ndurVj3rFewNeBp5fw+9X+lr5YlQETgcENgtzuDfp8Qq4kesPahsjtoLstbatuJ9vI7c67H/BCw04ppYvIbWZxZ4MddGtLx/fNu05PGr9eSZK0JKIED7UXZt22Y9aVJKm9KUXObUHWTSl9BEyIiNodjnclt5TT/UBtLjsKuC/7+n5gZER0j4ghwLrAC9lyCLMiYpvI3eJzZINzaq91IPB4a9afhcJr0JZcSumBiLgcuCEitif3Ij8HNgBOBManlL4REc8Ad2Q7rE0lt0NtT+DXS/gtazdOOCEibgfmpJReWYqXcC9wPnB9RNwMbA4ck98hbw2sUyLifXK3cP0I8raWbBvHR8QC4FXgW+R2Cj60qY4ppTMjojdwX0TsnlJ6LqX0akQ8AFyTHfsIOBWYA9Q0dR1JktRCLnFQkcy6bcqsK0lSe1S+nPt94E8R0Q0YRy7DVAF/iYjjyG3kehBASum1iPgLuUncRcD3UkrV2XW+A9xMLpuN4YtNTG8A/pBtKDaNBhuTLomyT9ACpJR+FBH/BE4C/kzuBY8nNxN9adbtG+R+030luYqAF4BdUkrvNLpg89/r/Yj4MfADcv+hJtK6HXxrr/dqRBwL/JTcjrePA8eSu50t30nkNhC4GpgOXARsB2zc2u/dCiOBK4ALyb3uQ1JK/2mm/0nAssCYiNg5pfQyud3vrgFGkfvh4nfk/pJ/uYTjliRJ6rDMum3GrCtJkuqklF7ii01A8+26mP4XkcswDdvH0kSmydatb7ghaatEKytvJaBujY1XgedTSo1u3cs3bNjQNHbsH9pmYKoYv4jDyj0EdVBnL7q53ENQBxVdt/tXgQ0Him7YGr3TC6dtUdRrdvnB/2vz1yFVGrOuSu3nZl210rlmXbVSW2fdUuRcqLys2y4qaNVxRMRBwKrAK0AfcrePrUtuDQ5JkiSpwzLrSpKkcnCCtoPLFiju0kyX6tYuULwYs8mt2bFO9n1fAfZJKTXagEGSJC0B16CVGjHrSpJUAcy5BVU1dzAi1oiIPi29WNZ/x8I9VURHkds1d3GPZm/FWlIppdEppc1SSsullHqmlLZKKY0u5veQJKnTCXKprJgPFWTW7RDMupIkdWSlyLkVmHULVdC+B5wHXFDbEBEnACeklJpaQOIY4Gc0/1tuFdcDNL9pwXttNRBJkqQOxqzb/pl1JUlSxSs0QRvZI98qwKalGY6WVEppKjC13OOQJElLyVu/ysGs286ZdSVJqgDm3IIqsChYkiRJkiRJkjoGNwmTJElqDywskCRJUiUy5xbkBK0kSVLZhbd+SZIkqQKZc1vCJQ4kSZIkSZIkqUxaUkGbSj4KSZKkzs7CgnIx60qSJJWSObeglkzQnhIRx+Q97wcQEeOa6NuvKKOSJEmS2oZZV5IkSWXVkgnafjQdRgcvpr9VCJIkSUsicG2u8jHrSpIklYo5t0UKTdAOaZNRSJIkdXbuDFAOZl1JkqRSM+cW1OwEbUrp/bYaiCRJktSWzLqSJElqD1qyxIEkSZJKzVu/JEmSVInMuQU1O0EbEWsUOL8G+Cyl9HnxhiRJkiSVnllXkiRJ7UGhCtrxtGAjhIgYD9wEXJJSWrj0w5IkSepkLCwoh/GYdSVJkkrLnFtQoQnaD2g+tFYBy5PbYOHnwFcjYo+UUnWRxidJklT53N22XMy6kiRJpWTObZFCm4QNbslFImJT4FfAbsC3gGuXemSSJElSCZl1JUmS1B5UFeMiKaWXgX2ACcBhxbimJElSZxJR3IeKx6wrSZLUesXOuZWYdYsyQQuQUloAjAY2KtY1JUmSpPbArCtJkqRSKbQG7ZKaCixX5GtKkiRVvkosBag8Zl1JkqQlZc4tqNgTtKsAM4p8TUmSpMpnbu0IzLqSJElLypxbUNGWOIiIvsC+wEvFuqYkSZLUHph1JUmSVCpLXUEbESsA2wLnAwOAm5f2mpIkSZ1KBFRZWtAemXUlSZKWgjm3RZqdoI2I6iW4VgC3pZT+vHRDkiRJkkrPrCtJkqT2oFAFbaEp7kRuHa6XgZtTSrcUZVSSJEmdjYUF5WDWlSRJKjVzbkHNTtCmlIq2Rq0kSZKa4e62bc6sK0mS1AbMuQUZSiVJkiRJkiSpTIo6QRsRXSPie8W8piRJUqcQRX6o6My6kiRJrVDsnFuBWbcoE7SRcxTwFjCqGNeUJEmS2gOzriRJkkqp4ARtRPSPiJ9GxP0RcXdE/DAieuQd3xt4DbgRWBP4a+mGK0mSVIGC3NpcxXy05NtG7BkRb0bEOxFxZjP9vhwR1RFxYLFecnth1pUkSSqhUuTcClzTttlNwiJiAPACuTBa++pHAPtGxO7AtcCx2bG/AT9LKb1UuuFKkiRVqDbOmRHRBfgdsDswEXgxIu5PKb3eRL9LgIfadoSlZ9aVJElqA5U3n1p0zU7QAmcCg4GXgT+Re0uPAHYC/g7sATwPnJxSeqF0w5QkSVKRbQW8k1IaBxARtwP7Aa836Pd94G7gy207vDZh1pUkSVLZFZqg3Qt4H9g6pbQAICKuAt4gV21xO/DNlFIq6SglSZIqXVXRSwsGRMTYvOfXpZSuy3u+GjAh7/lEYOv8C0TEasA3gF2ozAlas64kSVKpFT/nVpxCa9AOBkbXBlaAlNJccrd4AfzUwCpJktQuTUkpDct7XNfgeFNJuWGuuxI4I6VUXZohlt1gzLqSJEkqs0IVtD2Bj5to/yT7c1xxhyNJktRJtX1hwURg9bzng4BJDfoMA26P3EYMA4CvRcSilNK9bTPEkjPrSpIklZoFtAUVmqBtlhUFkiRJRVC7u23behFYNyKGAB8CI4HD8juklIbUfh0RNwN/q6DJ2YLMupIkSUupPDm3w2nJBO1mEXFkwzaAiDiCJubBU0q3FmFskiRJKpGU0qKIOAl4COgC3JhSei0iTsyOX1vWAbYds64kSZLKqiUTtPtlj4YCuHkx5xhaJUmSlkQZCgtSSqOB0Q3ampyYTSkd3RZjKgOzriRJUilZQFtQoQnaW2m8WYQkSZJUCcy6kiRJKrtmJ2gruFJCkiSpHQnX5ioDs64kSVKpmXNboqq5gxFxZER8qa0GI0mS1GlFkR8qyKwrSZLUBoqdcysw6zY7QUtu3a0RbTAOSZIkqa3djFlXkiRJZdaSTcIkSZJUSgFUVWApgCRJkjo3c26LFKqglSRJkiRJkiSViBW0kiRJ7YGbJ0iSJKkSmXMLaskEbb+IWGNJLppS+qCV45EkSeqcDK7lYtaVJEkqJXNuQS2ZoD05e7RUauF1JUmSpHIz60qSJKmsWhIuZwKflXogkiRJnZqVBeVi1pUkSSolc25BLZmgvSKldH7JRyJJkiS1PbOuJEmSysrbsyRJksouIKrKPQhJkiSpyMy5LeE7JEmSJEmSJEllYgWtJElSuQVQ5dpckiRJqjDm3BZxglaSJKk9cPMESZIkVSJzbkHNTtCmlFwCQZIkSRXJrCtJkqT2wApaSZKksnPzBEmSJFUic25L+A5JkiRJkiRJUplYQStJktQeuDaXJEmSKpE5tyAnaCVJksrN3W0lSZJUicy5LeISB5IkSZIkSZJUJk7QSpIklV22eUIxH5IkSVLZlSDntjDrRkSXiPhPRPwte758RDwSEW9nf/bP63tWRLwTEW9GxPC89i0j4pXs2KiI3HoNEdE9Iu7I2p+PiMFL8y6Z3iVJkiRJkiRVmpOB/+U9PxN4LKW0LvBY9pyIGAqMBDYC9gSujogu2TnXAN8G1s0ee2btxwHTU0rrAFcAlyzNQF2DVlK7dnb6c7mHoA7q53FYuYcgLRk3T5CkTudcs65a6V6zrjqSMuTciBgEfB24CDg1a94P2Dn7+hbgCeCMrP32lNJ84L2IeAfYKiLGA31SSs9m17wVGAGMyc45L7vWXcBVEREppdSa8TpBK0mS1B44QStJkqRKVJqcOyAixuY9vy6ldF3e8yuB04HeeW0rp5QmA6SUJkfESln7asBzef0mZm0Ls68btteeMyG71qKImAGsAExpzYtxglaSJEmSJElSRzIlpTSsqQMRsTfwSUrpXxGxcwuu1dQMcmqmvblzWsUJWkmSpHIL3NhLkiRJlac8OXd7YN+I+BrQA+gTEX8EPo6IgVn17EDgk6z/RGD1vPMHAZOy9kFNtOefMzEiugJ9gWmtHbA/CUiSJEmSJEmqCCmls1JKg1JKg8lt/vV4Sulw4H7gqKzbUcB92df3AyMjontEDCG3GdgL2XIIsyJim4gI4MgG59Re68Dse1hBK0mS1HEFVLkGrSRJkipNu8q5FwN/iYjjgA+AgwBSSq9FxF+A1/8/e3ceZ+d4/3/89ckmmxAEkdhqKWorWlX7l9p3rX2ppaW1Ve1aai0tVZRSpfVDixYtVUtpUWqp6GLfQ4k9IhLZJPP5/XHfZ5yZJGbEzJwzZ17Px2Mek7nPfd9z3+fcOfM+1/25rguYDhyUmTPKbb4FXA4MoJgc7NZy+WXAleWEYu9SNATPMRtoJUmS6oGThEmSJKkR1TDnZubdwN3lv8cCG81mvdOB02exfBSw4iyWT6Fs4O0IDnEgSZIkSZIkSTViBa0kSVI9cJIwSZIkNSJzbpt8hiRJkiRJkiSpRqyglSRJqrXAMWglSZLUeMy57WIDrSRJUs3V1ey2kiRJUgcx57aHQxxIkiRJkiRJUo1YQStJklQPnDxBkiRJjcic2yafIUmSJEmSJEmqEStoJUmS6oGTJ0iSJKkRmXPbZAOtJElSrTm7rSRJkhqRObcfPBoAACAASURBVLddHOJAkiRJkiRJkmrEClpJkqSaCysLJEmS1IDMue1hBa0kSZIkSZIk1YgVtJIkSfWgl/fNJUmS1IDMuW2ygVaSJKke2PVLkiRJjcic2yabsCVJkiRJkiSpRqyglSRJqrXAygJJkiQ1HnNuu1hBK0mSJEmSJEk1YgWtJElSzQWE980lSZLUaMy57WEDrSRJUj3oZdcvSZIkNSBzbptswpYkSZIkSZKkGrGCVpIkqR44eYIkSZIakTm3TVbQSpIkSZIkSVKNWEErSZJUa+HkCZIkSWpA5tx2sYFWkiSpHtj1S5IkSY3InNsmm7AlSZIkSZIkqUasoJUkSaoHVhZIkiSpEZlz22QFrSRJkiRJkiTViBW0kiRJ9aCX980lSZLUgMy5bbKBVpIkqeai/JIkSZIaiTm3PWzCliRJkiRJkqQasYJWkiSp1gInT5AkSVLjMee2ixW0kiRJkiRJklQjVtBKkiTVg/C+uSRJkhqQObdNNtBKkiTVBbt+SZIkqRGZc9tiE7YkSZIkSZIk1YgVtJIkSTUXTp4gSZKkBmTObQ8raCVJkiRJkiSpRqyglSRJqgdOniBJkqRGZM5tkw20kiRJdcGuX5IkSWpE5ty22IQtSZIkSZIkSTViBa0kSVI9cPIESZIkNSJzbpusoJUkSZIkSZKkGrGCVpIkqeYC75tLkiSp8Zhz28MGWkmSpFoL7PolSZKkxmPObRebsCVJkiRJkiSpRqyglSRJqgdWFkiSJKkRmXPbZAWtJElSDxURm0XEMxHxfEQcO4vHd4+IR8uv+yNilVocpyRJktTIrKCVJEmqC11bWRARvYELga8ArwIPR8RNmflk1WqjgfUzc1xEbA5cAqzZpQcqSZKkbs4K2rbYQCtJklRzAdHlHZu+CDyfmS8CRMQ1wLZAcwNtZt5ftf6DwMguPUJJkiR1czXJud2Oz5AkSVJjWiAiRlV9fbPV4yOAV6p+frVcNjv7Abd29EFKkiRJPZ0VtJIkSfWg4ydPeCcz1/i43ziLZTnLFSM2pGigXacjDkySJEk9iJOEtckGWkmSpJ7pVWDRqp9HAq+1XikiVgYuBTbPzLFddGySJElSj+EQB1IHevLJF9loo28xcODaLLLIZpx44sXMmDGjze3Gj5/IPvuczNChGzLPPOuz++7fZ+zY92Za78Yb72allXamf/8vs8IKX+Paa//SGaehGvDa0ZwYyzT+xJtcxEuczLP8ukVv9dmbwgz+yBucyfOcwfNcz+tMYubr7Wkm8nNe4lSe4wJe4nEmdPQpqIXo4K82PQwsExFLRkQ/YBfgphZHFLEYcAOwZ2Y+++nOT1Ije/75VzjggNNZZZVd6d37i2ywQetRVVr6znd+QsQaHHnkuV10hKo3Y8a8xeDB6xKxBhMnTgLg9dff4aijzmOVVXZl8OB1WXTRLdl77x/w2mtvz7T9Ndfczmqr7c7gwesyYsTm7LXXibNcT/XjdaZxEW9yOC+xI8/y/Vlk1yS5jrHsz4vszHN8j1cYzZSZ1ruX9zmCl9mV59iPFziP13mX6S3WeZfp/Iw32I8X2JXn+C4vcw/vt1jnb4xne56d6es2Zv5M9TJTOY0x7M7z7MpzHMXLvDCLY9OsdHTObbyKXBtopQ4ybtz7bLzxt4kIbrzxJ5x44v785CdX8YMf/KLNbXfe+TjuvvsRLr30+1x++Uk8/PCTbLfdkS3Wue++/7Djjsew4YZrcOut57Plluuw667f4y9/ebCzTkldxGtHc+ptpvEcHzA//Ziffu3e7ve8zktMYhsWYjsWYgxTuKZV4eTLTOZaXmMJBrIHI1iWQVzH6zzPBx19GgKaJ0/oyK82ZOZ04GDgduAp4HeZ+UREHBgRB5arnQjMD/w8Iv4TEaM66xmQ1L098cQL3HLL/Sy77GIsu+xiH7vuk0++yK9+dRNDhgzqoqNTPTrqqPMYPHhgi2WPPPIUf/jDXey666b86U8/5ayzDuWhhx7ny1/et7kRF+Cmm+5h112/x5e/vDI33vgTfvSjQ/j73//NVlt9h6ampq4+FbXT/5jGI3zAcPoxfDbZ9QbG8XveZQeGcjyL0J/gB7zKuKrG138ykXN4g8/Sn+MYwZ4M4wkmczpjaCpHa2oiOYMxPMFk9mIYxzGCZenPubzBg7MoOjiFkZzJos1fX2Jwi8dHM4Vj+R+D6MURDOcoFuELDGbqrEeHUgudkHMbcNIxhziQOsjFF1/P5MlTueGGHzNkyGC+8hV4//0POOmkSzj66L0YMmTwLLd74IFHuf32B7jnnktYb73VABgxYhhrrvl17rzzITbeeE0ATj31UtZb7/Ocf/5RAGy44Ro88cQLnHLKL9lkky91zUmqU3jtaE4tyyCW4zMAXMtrs6yCbe0VJvMCk/g6I1mC4kPR3PThUl7hBT5gKYoPy39nLIszgC1YEIAlGchbTOUexrI0fqBuFJl5C3BLq2UXV/17f2D/rj4uSd3P1luvx7bbbgDAV796NO+8M3P1WcWhh57NYYftwpVX3jLbddTY7r3339x22wMcf/w+HHXUec3L11lnVZ5++jr69PmoqWK11Zbjs5/dkeuv/xt7770VAL/97W2sttpyXHDBMc3rDRkyiG23PYJnnnmZ5ZdfsutORu32BQaxZpldf8xrvN8qu06jiRt4lx2Yjy0YCsBnGcABjOYW3mN3FgDgXibwGebimyzUvO1AenEGrzGGaSzKXLzGhzzPVI4vG1IBVmYgzzGF+5jAl5i7xe9emv4M+Jgaxot5iy8wmMMZ3rxsNTOxOlDjNTlLNXLrrfez6aZfatGYtssumzJ58lTuuedfH7vdQgvN39zABvDFL67IkkuO4NZb7wdg6tRp3HXXKHbaaeMW2+6yyyY88MBjjB8/sYPPRl3Ja0dzqtccdO15jg8YRO/mxlmAkQxgXvryPEVlynSaGM0kPtcquK7IEF5lClPa0RCsTyggIjr0S5K6Uq9e7ftoed11d/LUU6M59tivd+4BqW7NmDGDQw45ixNP3J8FFpi3xWPzzjt3i8ZZgGWXXZyBA/vz1lvvNi/78MPpzDPP4Jm2Bci0orFetZVdn2YKk2hi7aoM2p9erMEg/lXVi2s6ycBWzVmDWv08o6xsbb3eQHp94prXV5jKs0xhC+Zte2XNrBNybiNmXRtopQ7y9NMvsdxyS7RYtthiCzNwYH+efvqlNrZbfKblyy+/RPN2L7zwKh9+OH2m/S+//JI0NTXx7LMvf8qjVy157agrvcM0FphFl7Jh9OMdpgHwLh/SBDOtN4x+JDCWD7vgSCVJjWby5CkcccS5nHnmwQwaNKDWh6Maufji65kyZSoHHbRTu9Z/9NHnmDRpCius8JnmZfvuuy333vtvrrjiZt5/fyLPPvsy3//+RWy44Rot1lP3MoZp9AKG07fF8pH0Y0yZUwE2YghPMZm7eJ9JzGAM0/gtY1mRASzKXAAsRj+WpT9XM5bXmMYkZvA3xvM0U9h0Fg2t32I0O/IsBzGa21uNP/tsOc7sB8xoHj/3QEZzJ+M7+BlQT+YQB1IHGTfu/ea7ttWGDh3CuHGzn1jn47Z78cUxzesAM603dOiQ8nEn7unOvHbUlabQRP9Z3J/tTy/GlQ2vU2hqXtZ6HYDJVtB2ksarBJCkameccTnDhy/AHntsUetDUY2MHfseJ5xwMVdddQp9+7bdHNHU1MRhh53NMsss1mJori23XIfLLz+J/fY7lb33PgmAL395ZW666ZzOOnR1gYnMoD+96N0qEw2mN1NJPiTpS7AGgzmEhbmQNzm/rIddjv4cz4jmbYLgBEZwBq9xEC8BRQPYwSzMylU9yYbSh92Yn2XozwzgPt7nYt5iKsk25TAL75XZ9zzeYHvmY2n6cz8TuJA3GUpvVmfWQ9Kpmjm3LVbQ1rmI+HpEPBIREyJiXET8OyLOqXp8iYjIiNiqxseYEdHj35VmVWafmbRVfd/e7VqvV+m+04DV/T2O14660uxe9tbLYzZretl0EidOUA9k1u05Ro8ew9lnX8m55x7RkF1T1T7f+97PWXPNFdlii3Xatf5xx13AAw88xpVXtmzQveuuURx44Bkcdtgu3HXXxVxzzQ9599332X77I5kxwxvJ3dms3h2ybIStPPYYk7iYN9mKeTmVkRzBcCbQxJm81jy0QRPJubzBBGZwJMM5lZFsxVAu5M0WwyV8nkF8jflZlUGsziAOYzhrM5jfM7bFhGMAGzMP2zMfKzGQA1iIFRnA9YzrtOeioThJWJusoK1jEXEccCrwY+BYoD+wOrAH8N1ytdeBtYCna3GM+sjQoUN4772ZqxHHj584yyrH6u3efnvmN/X33pvQvF2l2rH1/is/f9z+Vf+8dtSV+tNrlpOJFZW1vZvXKZbNmGmd4vHenXyUknoCs27PcuyxF7D55l9mueWWaM4hTU1NTJ06jffem8A88wy24bbBPfHEC/zqVzfx97//svkamDSp6Do+fvxEevfuxYAB/ZvX//nPf89ZZ13J1VefzpprrthiX0cc8VO22WY9fvSjQ5uXrbrqsiy33Fe58cZ72GGH/+uCM1JHG0xvJtPEDLJFFe0HNDEXQZ9y2eW8zRcZzF4Ma15nSebiYF7in0xkLeZmFB/wCB9wIUuwSDls14oM5B2mcwVvf+wEX2sxN/9gIm/xIQvTj7nL7LtSVeUtFJOO3WQDrTpI4zU5N5aDgV9k5vGZeUdm/ikzTwKWqayQmVMz88HMnP00qeoSyy23xEzjhb7yyht88MHkmcb/bGs7aDku6VJLjaRv3z4zrff00y/Rq1cvll125nFI1X147agrLVA11my16rFp56MvvcplrdcJYP5W44KpI0QnfEl1z6zbgzzzzMvccMNdDB26YfPXK6+8yQUX/I6hQzdkzJi3an2I6mTPPfcKH344nbXW2qf5GjjooB8BMHLkFhxyyFnN615//V855JCz+PGPD2XnnTeZaV9PP/0Sq666bItln/3sEgwYMBcvvPBq556IOs0I+tEEvNFqvoMxTGNE1dwIrzKNJcqxZqu37Uc0bzuGacxFNDfOVnyGuWba/+xUepONnMX8DQDJnE3a2/N0Rs5tvOfdBtr6Ni/wRuuFWTUt5ay6fUXEXBFxUUS8FxFjI+KsiPhORFRvt0G53QYR8fuImBgRL0bEt6t/V0SsFRE3RcRrEfFBRPwnInbvpPPt1jbf/MvcfvuDTJjwUXeJa6+9gwED5mL99Vf72O3eeGMs9933n+Zlo0Y9yYsvjmHzzb8MwFxz9WPDDdfg97//a4ttr732DtZaa6WZZjBV9+K1o660DIOYyAxeZnLzsjFMYRwfsnRZFdCHXizJQJ5gYottH2cCI+lvBa2kjmLW7UEuvfT73HXXxS2+Flpofnba6SvcddfFDBs2tNaHqE62zjqrznQNHHPM3gDccst5HHXUngDcffcodt/9BA4+eCeOPHLPWe5r8cWH869/tSysf+qp0UyePJUlllikc09EnWY5+jOQXtzPR73/ptLEw3zQouJ1GH15kakttn2FqUwjWbAsJBhGX6aSLSYXA3iBKc3rzM6DTGQIvRlWdjr/LAMYTC8eZVKL9R5l0kwNxaofEbFoRNwVEU9FxBMRcVi5fL6IuCMiniu/D63a5riIeD4inomITauWrx4Rj5WPnR9ll48yk1xbLn8oIpaY0+N1iIP69i/gkIj4H3BzZo5t53Y/Br4OHA88BewD7DKbdX8J/D/gEmBX4MKIGJWZ/ywfXxz4B3AxMAVYG/h1RDRl5tWf/JQa14EH7sj551/DDjscxTHH7M2LL47hpJMu4bvf3Z0hQz5qBFt66e1Yf/3VuOyyEwFYa62V2XTTtdhrrx9w9tmH0atXL4455nzWWWdVNt54zebtTjhhfzbY4AC+852fsN1263PLLf/gllv+wW23/azLz1Udy2tHc2oaTTxXjqE1gelMpYknykC7DIPoRy/OYzRLMIBtWRiARRnAUgzkD7zBJixAENzJ2yzGAJaqCr7rMT+X8wq38hbLMZjn+IDn+IA9qiZfUAeza696HrNuA5k0aQq33HIfAGPGvM3773/AddfdCcAWW6zDGmusMNM2/fv3Y9FFF2KDDdbo0mNVbSywwLwzvdYvvfQ6AOuu+3kGDx7IU0+NZrvtjmS55ZZg552/woMPPta87rBhQ1lqqZFAkZ8PP/wcFllkGJtv/mXefPNdTjnllyyxxCJsscXaXXdS+kSm0sQjZXYdy3Qm09TcGLs6g5iLXuzAfPyOsQyiNyPpx02MI0m2ZN7m/WzKPPyat5mPPqzGQN5jBr9jLAvSh9XLPLs6gxhGH87gNXZmPobQm0f4gH8wkW+yYPO+fsRrLEN/Fmcumkj+wQTuYwL7M6y5OrYvwU7MzxW8zSB6sTT9eYCJPMlkTmNkVz193Vttcu504IjM/FdEzA08EhF3UGSIv2bmmRFxLMUwS8dExAoUeeJzwCLAnRGxbGbOAC4Cvgk8CNwCbAbcCuwHjMvMpSNiF+BHwM5zcrA20Na3g4A/ApcDGRFPAdcDZ2fm+7PaICLmp7hoTszMn5bLbgcen83vuDozTyvXuxvYGtgB+CdAZl5Tte8A/g6MBL4BtBlaI+Kb5fGw2GILt7V6tzZ06BD++teLOPjgH7P11t9l3nkHc/jhu3HSSd9ssd706TOYMaOpxbJrrvkhhx9+DvvuewpNTclWW63D+ecf1WKdddZZleuu+xHf//7Pueii61hyyUX47W9PazGbqbonrx3NqQ+Ywe95vcWyys+HsST96EUTSVOr7b7KcG7nbW7kTRJYlkFsXhVUARZnADuxCH/jHUYxnnnpw44szNIfM16XPqUGnOxAaoNZt4G89da7fO1rx7ZYVvl59OibrGpUuzz00OOMHz+R//73WdZee78Wj+2991ZcfvlJABx66C7069eXiy66josvvp55552bddZZlTPOOIhBgwbU4MjVHuOZwVmtsmvl51+wJAvSix0YShPJDbzLBGawFP05iZHMW9V8tRXz0pfgNt7jdt5jEL1YngHswQLNcykMoBcnM5IreYdf8zaTaWJh+nIgC7IJ8zTvawT9+CvjeYfpQDGcwWEszAYMaXGcWzOUJuAWxnEtY1mEfhzFcFZoNS6tZqMGOTczX6cYy57MnFDmjBHAtsAG5Wr/D7gbOKZcfk1mTgVGR8TzwBcj4iVgSGY+ABARVwDbUTTQbgucVO7rOuCCiIjq3kDtFXOwjbpQRMwFbAJsCvwfsDzwHLBaZk4sy6dHA1tn5s0RsQFwF7BcZj5TtZ8zgWMys1KGXVlv3cy8r2q9+4EXM3OP8uehwMkUF90IaO7XOiYzR5brfB34NTB3ZrbsD1tljTVWyFGjrvw0T4cktdvJsVutD0Hd1Ek8+0hmdmk51xqrLJajbj+6Q/cZww/p8vOQPimzriTNmT+adTWHtu/irNsZORcghh/yMvBO1aJLMvOSWa5b5Im/AysC/8vMeaseG5eZQyPiAuDBzLyqXH4ZRSPsS8CZmblxuXxdisyxVUQ8DmyWma+Wj70ArJmZ1cfVLlbQ1rmy5f5P5RcRsR9wKUUZ9Xmz2KRy6/7tVstb/1zResKFaRQz6FZcDnyJYobdJ4H3gW9RhFhJktRhHOJAPY9ZV5KknqBTcu477WlojojBFD10vpOZ78fsh1uY1QP5Mcs/bptPzL503UxmXga8Cyw3m1UqEy0Ma7W89c9tioj+wJbADzLzgsz8W2aOwutGkiRJncCsK0mSOkpE9KVonP1NZt5QLn4zIoaXjw8H3iqXvwosWrX5SOC1cvnIWSxvsU1E9AHmocgxn5jho45FxIKzWDaM4gV/czabPUYxwcG2VdsExXhbn9RcFN28mqdHLAdW3mYO9iVJkmYriskTOvJLqnNmXUmSeoJOyLntyLplPrgMeCozz6l66CZg7/LfewM3Vi3fJSLmioglgWWAf5Zj2U6IiC+V+9yr1TaVfX0V+NucjD8LDnFQ7x6LiBuBv1C06C8OHAlMohjIeCaZOTYifgmcHBEf8tHMtkP4hGXWmTk+Ih4GToyI94Emitntxpf7kyRJHSFwkjD1RGZdSZIaXe1y7trAnhR54z/lsuOBM4HflcMq/Q/4GkBmPhERv6MY8mg6cFBmzii3+xbFsEgDKMalvbVcfhlwZTmh2LvALnN6sDbQ1rdTKKoDzgfmo+jSdT+wc2aO/pjtjgb6Uswk1wRcSXHRfGcOjmE34BLgCmAscAEwEDh4DvYlSZIkVZh1JUlSpygnCZ1dqe1Gs9nmdOD0WSwfRTHBWOvlUygbeD8tG2jrWGZeCFzYxjov0eqCKy+Qb5VfAETEncB/q9a5u/V25fINWv38PMWMuq2dVLXO5RR3EiRJ0hxzWAL1LGZdSZJ6CnNuW2ygbUARsSGwJvAviuqCnSnuDnRIq74kSZJUK2ZdSZLUaGygbUwTge2A44D+wHPA1zPzupoelSRJmj0n9pLay6wrSVJ3Ys5tkw20DSgzHwa+VOvjkCRJ7RWAk4RJ7WHWlSSpOzHntofPkCRJkiRJkiTViBW0kiRJ9cCuX5IkSWpE5tw2WUErSZIkSZIkSTViBa0kSVI9sLJAkiRJjcic2yYbaCVJkmrOyRMkSZLUiMy57eEzJEmSJEmSJEk1YgWtJElSPbDrlyRJkhqRObdNVtBKkiRJkiRJUo1YQStJklQXrCyQJElSIzLntsUGWkmSpFqLgLBjkyRJkhqMObddfIYkSZIkSZIkqUasoJUkSaoHTp4gSZKkRmTObZMVtJIkSZIkSZJUI1bQSpIk1QUrCyRJktSIzLltsYFWkiSpHjh5giRJkhqRObdNPkOSJEmSJEmSVCNW0EqSJNVcYNcvSZIkNR5zbntYQStJkiRJkiRJNWIFrSRJUj0IKwskSZLUgMy5bbKBVpIkqS7YsUmSJEmNyJzbFp8hSZIkSZIkSaoRK2glSZJqLbDrlyRJkhqPObddrKCVJEmSJEmSpBqxglaSJKnmAsL75pIkSWo05tz2sIFWkiSpLtj1S5IkSY3InNsWm7AlSZIkSZIkqUasoJUkSaoHTp4gSZKkRmTObZMVtJIkSZIkSZJUI1bQSpIk1VzgfXNJkiQ1HnNue9hAK0mSVA/s+iVJkqRGZM5tk03YkiRJkiRJklQjVtBKkiTVBe+bS5IkqRGZc9viMyRJkiRJkiRJNWIFrSRJUj1wbC5JkiQ1InNum2yglSRJqrUIg6skSZIajzm3XRziQJIkSZIkSZJqxApaSZKkuuB9c0mSJDUic25bfIYkSZIkSZIkqUasoJUkSaoHjs0lSZKkRmTObZMNtJIkSXXB4CpJkqRGZM5ti0McSJIkSZIkSVKNWEErSZJUcwHhfXNJkiQ1GnNue/gMSZIkSZIkSVKNWEErSZJUFxybS5IkSY3InNsWG2glSZLqgV2/JEmS1IjMuW3yGZIkSZIkSZKkGrGCVpIkqeYCu35JkiSp8Zhz28MKWkmSJEmSJEmqEStoJUmS6kFYWSBJkqQGZM5tkw20kiRJtRY4eYIkSZIajzm3XXyGJEmSJEmSJKlGrKCVJEmqC3b9kiRJUiMy57bFClpJkiRJkiRJqhEraCVJkmounDxBkiRJDcic2x420EqSJNUFOzZJkiSpEZlz2+IzJEmSJEmSJEk1YgWtJElSPbDrlyRJkhqRObdNVtBKkiRJkiRJUo1YQStJklRzgffNJUmS1HjMue1hA60kSVI9sOuXJEmSGpE5t002YUuSJEmSJElSjURm1voY1ENExNvAy7U+jjq2APBOrQ9C3Y7XjeaU187sLZ6Zw7ryF0bEbRSvSUd6JzM36+B9SpoNs26b/LujOeF1oznltTN7XZp1OynnQoNlXRtopToREaMyc41aH4e6F68bzSmvHUlSV/LvjuaE143mlNeOuhuHOJAkSZIkSZKkGrGBVpIkSZIkSZJqxAZaqX5cUusDULfkdaM55bUjSepK/t3RnPC60Zzy2lG34hi0kiRJkiRJklQjVtBKkiRJkiRJUo3YQCtJkiRJkiRJNWIDrSRJkiRJkiTViA20kiSpXSLi5Ij4evnvqPHhSJIkSR3GrKtasoFW0kwiwvcGdZoo1fo49MlExPzA6sA5EdE3nWVUktRNmXXVmcy63ZNZV7XmHyZJzSpBIjObIqJfucz3CXWYiIgsRUT/cpnXWDeQmWOBM4HpwKngaydJ6l7MuupsZt3uy6yrWvNik0RE9Aao3CWMiP8D/hUR85YB1jvA+lQiog8U11hE9IqIA4Dby2VNNT04tanqPeDfwC+AIyJicV87SVJ3YNZVZzPrdm9mXdUDG2glkZkzACJim4g4BNgDWBo4rFzF0KpPJTOnA0TEfsC3gS2BNSNiu3K5f4/qTOXDLHz0gTYzPwCuBZ4DzivX8/1BklTXzLrqbGbd7sesq3rjm4TUw5VDJA2OiN8A1wArAQsDCRwWEctbWaBPKyKWiIiHgFOAlYH+QD/ghHKMJ+9O14nyPaFX1YfZRSNi7qpVngV+CmwTERuVlSK+P0iS6pJZV13BrNt9mHVVr2yglXqY1ndvy7uFSwObU9ztPTgztwC+CYwHTqtaT5qtSnCpvhtdZX1gKWBH4NDM3AQ4A1gUOK7czr9JdaAcNq0pIlaJiLuBG4FHIuJ7ETE8M6cBdwJ3AOdWtqndEUuS9BGzrjqLWbcxmHVVr3yDkBpcRGweEcPLf/eq3L2NiD0jYsFytfWBD4F7yz9IUFQYXAGsFxFbVbbv2qNXvSvH2NqhvLaynBhhRkT0joj9qgLsRsCLwH8zc0q57GfAbcA3ImIxq1fqQ/maHg7cCrwNnAj8GtgW+CFAZo4GLgCWjYhvl9v52kmSupxZV53JrNt4zLqqV/4BkhpYROwEXE1RIUBVYD2C4o/PjHLVF4BhwOTy8X6Z+SHwJ2AqcEz19lKVBaw5pQAAIABJREFU0ygG0v9a+XMluFwF7EbRfRDgFWBEZlausV6Z+QbwZ2AhPrrGvDvdhWbzQXQEsCzw48z8WmbeDDQBqwJ7Vj7EAv8E/h9wWkQM8rWTJHU1s666gFm3GzPrqjuxgVZqbH8E7gI2jojVqpZvAPwlM8eWPz8J/I+y+w1l0MjMUcCrwNoRcRBYWaBC1XXw/4DngZ0iYv6yMmBu4AvA1VUfdO4E5ouI3cuf+5XfrwMmAntFxLqt9q1OUo691Xs2H0TfB/4AXBARa0XE88ChwEnA/XzUFfRN4HJgEnB+Zb+df/SSJDUz66pTmHW7N7OuuiPfGKQGVd61nQZcCiwA7FNZDnweuL5q9TeB31J0v1mjrCggIpYH5qIIvvtFxFArCwRFhUnZxesZ4AZgRWDP8uGlgHkoKgYqngNuBn5c3oGudP3aEhgDPA58t7LvLjiFHq0ce2tGRCwbEadFxOERMaJ8Tcdn5l8o3jd+BtwLrJeZPwT+DqwcEQeUu3qCooJkXSsLJEldyayrzmTW7d7MuuqObKCVGki0HLC+UhnwZ+A+YP2I2ABYHegN3FveWYzM/IBi3J27gRujGCB9c+Ao4D/ATcBQ4HNddS6qT62uscod5IuBl4HtyzHg1gdeyMzXI6IfQGa+CvyEohvhHWVI2hz4FsUHqAeBxSNi0S46lR4vIvYFHga2B86k+H++WdUqmwCfAa7MzBfKZQ+W30+KiAUyczxwVmYuW76PSJLUacy66mxm3cZh1lV3YwOt1I1FxKoRcXT5736ZOaP899x8FCigGOC8iXLmWuAZoKm8s1gJt88BuwMPletcA6xFMWj6g8DiwBTUo0TEVyNix/Lf1dfY4Mrd/8ycAPwKGA4cTzHA/iPlY5WJOMjMB4EdKMaD+x7wO6B3Zp5EUdnyGYouYOpArbvRRcReUUx2sALFeGjrU4y5NRA4OCKWKletvI+8Wm43mOI94i/ANGBrgEr30Yjo0+knI0nqUcy66mxm3e7PrKtGEVZoS91TRPQFvg+cAKyYmU9GxCLATykGPn8FOAt4LDM/jIhTKcLE54CXgGcpuuv8A3gpMyeV+52LosvOQpn5WLns8nK7LTPzrS47SdVURMwL/IYi1MxTdhNak+IO9AyKa+jczHy2XP9qiiC0EvAA8A5FMP0b8HZmTi/XGwgsSBFYX4iIBSjuaL8I7FPpdqhPr6wayqqf+wNnU3yAfQzYIjPHlI/tCxwNXJeZ3y9D6NsUEyTcSPG+sB3F+H2PZuY7XXoykqQexayrzmbW7f7MumokVtBK3VT5h/03FH9QflYOWH4TxSyh/6WoCLiBousWFJUF4ygmSbiIjyZK+DvwSERcXgaSaWUwnRYRu0XErRSzll5iYO1ZMvM94EcUA+OfHBELU3TReo+im9euwPURsX25yc8puhQ+QDGhwgyKAfX/BdwdEZdExMjyA9KrwJCIOBC4g6Jq5TIDa8eo3OHPzIyIIRFxfvnzFIrX5BGK12dy+d5BZv6KIshuGhHrlR8y9qWoFjmJIuhekJl/y8x3yl6j5ghJUqcw66qzmXW7L7OuGpEXm9S9vUARRjekCAxPA7tn5kEUd3dvB46MiNWzmIXyt+V2/TLzAGB5ijF5rqeYIGF0+UduALAucARF944VM/OXXXheqrFKkKEInJdS3G2udNXaKzP3o7hGXgLOiYg+mXkvRQAdCLyTmTsAy1EEn/9QXEuvl/tdANicYrKE+4HPZOZdXXBqPUJVBcf6wI4U3bkOLB/+H3AJsAqwcvl/vpIHfk7x+u0WxeQrfwDWBr6SmYtm5pXlfqPsNeokF5KkzmTWVacw63ZvZl01Ioc4kLq5iBgBnAHsAVxUBtbKYysDvwDGZ+ZmZRD5HTASOH5WIaH8Q9UUEfMDc2fmS9XLO/+MVG8iYhWKGU7XAQ6o/gATEZtQjMl1edlVaEngSuBd4JDMfPlj9rsIML1SrRIRvSvjfql9ImIz4OHK2FhVy7cDLqOYeTYpPmCMA0Zm5uSIWAj4ffnYJpk5tWrbnwI7AUdn5m9a7bdPJRBLktQVzLrqbGbd+mXWVU9iBa3UzZVj6vwWeAsYAM1jdgE8BdwKLB0Rq5Tj81wCLANsU9lH5Q5ydTDNzLFVgTUMrD3aU8BVFN2E+kKLa+wR4B6KmZPnzczRFN0N1wM2ruyg+hqrLMvM1zLzrYjoVV5jBtZPICLWpQiew1stXxj4IcXrsA9Ft81TKWanPr9c7R3gdIqKgV3K7SqvzS+BvwKjWv9OA6skqauZddUFzLp1yKyrnsYKWqkbq3S9iGKA+xOBA4FlMnNMVXXA9hRjJK1cFUJ3AG7OqllHpY8TEYtTVKgsmpmfq1x75WPnUATUNcs71nMBm2bmTTU85IYUEf2AD6ue+wXL4F/9emwA3AzskZl/LJfNAxxJ0XVvpcx8IiKGUFSLrAWsnZlvV+9HkqRaM+uqq5h164NZVz2ZFbRSnYp2DEhe+eOSxQD3vwHeBM6LiLnLwNobWBH4AGiq2u6GzJzWnt+hxlQ17la7lN23fg4sFRHfo5gggXIMt88A4ym6EJGZUyuB1Wus40TEQcDVFOGz8qH1rYg4DDimfC0ABgNBMVNwpVpoPPAHinHUzgXIzPcpZsJeGvhOuax6FlxfO0lSpzHrqjOZdbsfs656uj61PgBJLVWqAcrQuRAwJTPHt+Nu3+PAecBPgLki4k6KP1yHA9cBb7TewK5cPU8ZVnvNYRerf1BUqJwILBcR9wBLUU6AkMWsqS14jX16EbENxf/tqRTd7FaIYubaymu4IkXXrduBfwN3U8xGvDXwKMXN2KbM/FdEvAhsFBFbZuafgWcoZih+sPXv9bWTJHUGs646k1m3+zHrSgXvGEh1oqwAoAyrERErUMwCunV7umKUA5//GfgjsCWwEMXg5+dk5jft4iUo7hpn5oyIWCwiti+vs/ZuOxa4AngM2B1YGPgKsG9mXtg5R9yzRTEz7ZkUY2xtAxyYmftk5vSq94SjKSZFODgiBlME1iuBoyNisVZjaU0E3qOoJiAzJ2fmtZn58ietNJEk6ZMw66ormHW7F7Ou9BEbaKU6UbnLGxGHArfx0cQG21J0q2mPF4Aby39fA6yXmaeV++3dcUer7qTy2ld9P4ribvIvgH9HxH7lWFrt6Q72X+AvwHTg/MxcI8vZT+0m1Cn2BCYDpwHPZeYEKGaYLb/3zsxxwCnlumuX1QAXA/8Dro6Ir0TEwhGxMzA38ANgYFmt0MzxuCRJncmsq85i1u3WzLpSyTcYqU5ERN+I+BnFH6e7KULBbcAOwBZRDJj+cdtXZp+9CVgyMx/NzOnhrKE9Vlmd0vzal9UEAawM7EfxwegSinGaNinXaat6ZSLFYPsLleM6taiI6axz6YnK53VZ4L+ZOa7VazMPfPRhF7iM4gPFMRExLDOfBXYD5qX4IHsfxZheN1C8v8xDOUuxJEldwayrjmbW7d7MulJL4U0EqevNqhtXRCwK3E/xx+f0zPwwitkoLwdWAr6amf/5BL+jlyGiZyqDaeVDDBGxLkXXoKeA/sACFF21ppSPPwaMBfbPzOc/we/p7YehzhUR3wYOA04FHqD4wLEnMByYD/g9cFlmjo6IjYA7gG8CV2QxOcqSFON2fRa4LTMfj4g9KD54bJ6ZM43HJUnSp2XWVWcy6zYOs670ERtopS70cX/kI2IJikHO98rMP1ZCZ0QsRtFF5wLg1Mx8f3bjdBkiVK2sRFmRopvWKIpJDoZTdAn8JsXfgBkRsQ7wd4rZTS/KzA8/Zp9eY12orCx4FFgG+BAYALxN0aVrErA2cBVwVGa+HRFXUwTbHTLzmar99KUYq28T4IfAXcABlcoQSZI6gllXXcms2/2ZdaWPOMSB1AUqYx1Vjb21Y0QcGxFrRcTAcrV5KP4YrRgRfcrA2icz/0fRXWMPYM1yP60rElp07YqINSJiSNecnepB1Zhbvcrv2wNPA/sDp2TmZhRjvN1CEVyGlYG1V2beB/wBOJSigmW2+291jfXp3LNS+XyvDxwEnAFsB2yZmV/IzPWBsym6hq5ZbvI9YHlgt6prYgBFF7/rgLOAX2bmrgZWSVJHMeuqs5l1G5NZV/qIDbRSJ4iIEeX3XvBRyIyIYRHxZ+DXwAHA7RTdccjM/wIvAptRdNEgP5qR8l3KmWojYljV74nyLm9TZmZEbBARr1BUIMzT+WeqelE19lalq99girvQO1PcQSYznwR+BAwBvt1qF98GRgL7l90NgeYPRL2qwuq+ETEd+BYwEHW6zHwnM3+Zmadl5k2ZOarq4TMpXuv+5bovAtsD51ZdE5OBhyhms102M0/o2jOQJDUas666mlm3cZl1pYINtFIHi4gDgaciYkDVuEhDIuJoiru5EynuEq5FMabO7hFxQLn5CRR3dQ8su+wQEQsBwygqC74OLFIu75uFGRExogzDdwA3Aztn5itdc8aqBxGxQET8JopZa6GoHvgTMJTyvb6sbvkvcB7w3YhYrqxe6Z2Zb1LMhrpq1fqVD0RNZQXMf4FzgJOAI7wr3bXKDxC9q37uC+wFjANeqizPzBszc1xUzTScma9m5rWZObYrj1mS1HjMuqoFs27jM+uqp3MMWqmDRcTywJDMfKhq2UoUYWECcHJmnlMuX5IiQCwMbJOZb0TEycA+5bo3AZ8HFqTo7vEg8LPMPKNq32dRjKf0d+D46t+rxlR2B5zeatniwBUU4zZtnZlvRjFhwi+BZzJz26p1F6MYq+ufmblX2WVwln8MImJB4EKKrkVXUly/ozvlxNQuZXCdH/g/4FiK94VDPm48NUmSOopZV53NrNuzmXXVU1lBK3WQqnGLnsrMhyJi64hYoVz2GMWd2Lkp7gBWZrcdDVxP0UXr8HJXPwR2B16h6ALWBGwF9Kbo2jGt3H6hiLiLYpbLPShmqTSw9gCVwFrdBTAzXwZ+QXGNHVEufoiii+EGEbFZuU2UY72dAewREZtWdUuM8nuf8vvngH8Di1NUwuxjYK2dspvnjyk+iFxFMQv2bZl5oIFVktTZzLrqKmbdnsmsq57OClppDs3uTmwZXuehmEn0vcxcrVw+DPgPxd3cwypdZiJibuB8ioHPd8vM/1Tth7JbVx+K8ZS2B7bNzMeimHBhZYo7xuM692xVS2WYjKpuhH2Au4HXgP0yc0K5fDBFlcrawB6ZOaqsaDkHmC8zV6/a5/wU4fai2XURLLscLgH8JzOndtLp6ROIYkKMPYDRwHmV1y7KmbBrenCSpIZi1lVXMeuqwqyrnswKWukTiog+EXEusHHVsoiIfhFxEfDVzHyXYobJz0XETgCZ+TZwGrAb8MXKtmXg+C3FIPenVy2fAawdEWcD/wC+AZxWViiQmZMy80EDa+Mrx19rKsfe6l9WFfyeotrkS1XrTaQYv20SZWVBeb38BhgREd+F5g9cYzPz+I8bvy0z38zMhwys9SMz/wB8PTOPzMxXIqJ3+XoaWCVJHcKsq65m1lWFWVc9mQ200pz5GnBkRCwAzTPXrkpx1//Bcp27KAauP6tqu0uBx4GjyvGOKLe/A7gTeCjKmUTLh14CPgfcDwzPzF912hmprkXEesCbwHrlovOBJ4Gjy7v/AGTmnRQVLZtGxFbl4r8BDwMHRTnhRtV+oyuOXx2nqoqkV2bOmN2YapIkfQpmXXUps64qzLrqqRziQJoDEbEJxcyh+wDXZOaHEXEesEpmblDpEhYRGwF/AM7OzFPKbb9Sbrs/cFVZPUBE9MvMaVW/o7KPweXd4spMozO69GTV5WbVpTAi5qLoNvgM8I3MfDsiNqboRrg/cEV+NF7XVsCNFGF1y8ycFhErA6MrgUeSJGl2zLrqTGZdSZqZFbTSHMjMv1CEgmOBRctAsRFwQ7lK5f/Ww8AlwHGVO79lBcEtwE+AEVX7rEyI0Kv8OcvvE8tuZWFg7RnKDytrRMSy0DyT7VTgSGAb4CvlB5g7gd8BxwCfLdftS9El8XGKSpcdy30+mpkTKuO9SZIkzY5ZV53JrCtJM7OCVmpD6zv5lQHKI+KzFHd5TwFuoggPu1TGzapaf1WK2Wvvz8w9y2XLUNztPberzkP1q3UVQVkBcAPweGZuV71ORNwILEkxgcboiFgYeI6ia+A1wFSKgfV/DryemY908elIkqRuxKyrzmbWlaS2WUErzUY5PlbznfyI+Ez1GEaZ+QzF2EiHA98ClgemRUS/Vrt6GvgpsHtErFtu+5yBVfHR7MUt7pRl5qPAdcDyEbFNubhSDXA8RQXBLhExKDPfAL4OLAicDVwIPJyZN1cCq2NvSZKk1sy66mxmXUlqPytopTaUY2udCvQHpgB/zcwTyseGUnTt+gzwCtAXeA+4jWLm0VHlmF3DgROAczPz2ap9zzT+khrfLKoIdgSWAe4BHiu7+q0GnAPMADbNzOll96/pEfEbYG1gx6pgOgRYGXg6M9+Z1e+RJElqzayrjmbWlaRPzgpaaTYion9E/BC4mmIG0VMpunYdGxH7AmTmOOAH5SYHAbsCNwMbUATXdyPirMx8PTO/XR1Yy+0NFD1ARIwov7cYcy0ihkXEn4FfAwcAt1OMvUVm/otiZuQlgIPLZdPLXb4GLAbsV35wIjPfz8z7MvOdiOhtYJUkSR/HrKuOYtaVpE/PBlqJj8JEq+4xI4AVge9k5v6Z+QdgJEX3m5MjYqlyvd8CjwBHAf/OzKOB1Sm6gR2cmUe1/j3qOSLiQOCpiBiQmU3lsiERcTSwCTARWB9Yi6ISZfeI+Ea5+e+AUcABEbFgue1wii5efwS2o6h2aSEzZxhYJUlShVlXncWsK0kdwyEO1KNV3eVtmsVj8wErZeY9EbEdcBEwnmIsrmOBO4D9y8Hs1wX+SnFH+CJgeqtuPb1m9TvU+CJiOWCezHyoatlKwH+BCcDJmXlOuXxJ4DxgYWCrzHyrHJfrdGAe4HKK2WrnBrYGJnldSZKk2THrqrOZdSWpY3iHUz1aZjZlMUvtmhFxXkR8JyI2LB97twysKwM/BK4F1gMuo5jJdg+KO8Fk5r0UIXZDoPcsBsI3WPQQETGg/N4XIDOfzsyHImLriFihXPYYcBJF+BxXrh+ZOZpiFuR5gO+Wu7wF2Iui6+G2FGO/7ZuZE8trt0+XnZwkSepWzLrqaGZdSeocNtCqR4iIeSJii4j4bHXXq3LsrfMpBqxfATgEuDEiDqqED2ArYDjFpAdvZeZU4F2K8PD9iBhYrrdTZm6fmVO67MRUVyLii8BJETF3Zn5YLusfEYOBnwFXVa1+EfA6sF5EDKn6oHMDcD+wTUSsmpnTM/PfFF28Ns7MzTNzdFVFzHQkSVKPZtZVVzDrSlLnsYFWDS8iTgeeBi4GngCurRpTaw2KsZH2B3YEVgOOBs4EdivXGQC8SPn/JSKGUYy5dS2wGUXYJTM/KB/v3eknpXqVFB989omIvhFxPXBlZk4EjgM+FxE7AWTm28BpFNfZF5t3kDmBYqy3Dym6e1WWTym3ISJ6W6kiSZLArKsuZdaVpE5iA60aVkRsEhFvU4TR71F0nTkI2AY4JiLmAb4K9MvMq8qZQccDbwCDgFXKXd0DLACcXwaO4yhmFT0FmC8zR1X/3syc0flnp3qUmQ8D51KEzYnAIsAvy4fvopip9qyqTS4FHgeOqkyMUO7nDuBO4KGI6NVqQg+vMUmSZNZVlzPrSlLncZIwNaSyK9Z1wKbAopn5WtVjPwAOp6gmOBQYnJnbRcTngV8BywI/oejm9W65zRHAPhThdQJwUGb+pXzMSRF6qOowWU6gMYRiAo3VKbpubVR2E6ysvxHwB+DszDylXPYVirG39geuqgTSiOiXmdO67GQkSVK3YdZVVzDrSlLXsYJWDSkzJwHnAK8B34AiBJQPXwMMAfoDrwBfiIibgUcouoetlpknAu9HxGZlMPkpxSQJW2fmMpXAWv4uA2sPVHa9ynI8rUp4nUox6/EJFNfLauW6lffah4FLgOMiYiForiC4heKD0ojK/iuBtXocOUmSJDDrqvOZdSWpa1lBq4ZVzvj5M4quXWtk5svl8p0putusB7xFMWPoVIrZQm8u1+kH7ArsDexT2bZ63w5Y3zNVV5GUs9ceD8wP3Ab8NTOnRsSiwG8owux61TMdR8SqFLPX3p+Ze5bLlgG2zMxzu/ZsJElSd2XWVWcw60pSbXi3Sg2rDJU/A94BTgSIiB2AC4F7gRcycwzwI4oKg6Uj4jMRsRrwA4pxtx4Bxs5m3+qBqgLrMsC/gZ2BLYArgDPK1V6jGH9rLaASTCuVB09SVKnsHhHrlvt8zsAqSZI+CbOuOoNZV5JqwwpaNbQyKBxNMXHC/2/v7oPtqso7jn+f3PAiikIoJWI6JBk7oJ0aFEmnDZYExJdaMLQ2/qElSYWQjLallUGstgQGLR1L1VYtRCmhRR0LjsqL0lLkolBNEQWcjtJ2MJFgQ3gpL4Y0YPL0j7UOnJ6cc++5r+eenO9nZs++d++11l7n5IT8WGfvtbYCRwEfzcwPtpS7FFhOWcV2B3AAcG5m/uP09lgzXUQsAt4BDFEeH/wTyufmLMrn7MTM/GZEzAE+DiwBXpWZP42IQ2q9BD4M/FVm/kdT25H+R1mSJHXJrKvJZtaVpN5wgFb7vIiYS1kQ4QTKnFv/VY/PovwdaExUfxRlYYQjMvOrTfVdGGFA1bm3drccO53y2NZWYF1m3liPL6CsYnt4Zi6qxxZT5tz6DvBPwDrg7sxcMX2vQpIk7cvMuhovs64kzRxOcaB9XmZuA64GdlPm4mqEkT3NgSQzt2TmXY3AWuf1cmGEARQRQwCtgbUe+xJwFTAPeFHT8R9RHjN8RUSsrYfvBNYCRwDvAa5vDqwuiiBJkibKrKuxMutK0szjHbQaCBFxEHAlsBg4pk5u790C2kvznQQRsRJYQZln69vAdZn5cES8Avg68BXg/Mx8vJY/jDLP26nA/MzcWY8fCezIzCdaryFJkjRRZl11y6wrSTOT32hpIGTm05RvfAE+1Mu+aOaIiOMj4tKI+HBEnAflToKImBMR1wKXAk8Ar6Q80rUhIo7MzB8AnwfeAJzSaC8zHwWuoNxt8Kmm4z/JzCciYlade8vAKkmSJo1ZV+2YdSWpfzhAq0HyLeAW4MyImOMdBYMrIl4SEZ8DvgEsBF4FzImI+bXIqymPCK4GVmfmEuAPgF+i3DUAsB54Bnh7ndOt4V7KysjfaL1ufdTQxxYkSdJUMOsKMOtKUj9yigMNlLqIwp7M3N7rvqg3IuJQYANlXq0PAnc1HttqKvMJyt0Cr8nMHfXYC4H3A2cDyzPzjjr/1geASzLzk9P4MiRJkvZi1pVZV5L6k3fQaqBk5rbM3N6YGF+DIyKi/vgmYBlwMXBr05xaQ00LGTwIzGkKrLPrzzcBT9Y2yMzLgEeAsyJiYYfrSZIkTQuz7uAy60pSf5vd6w5IveC8SIOn6XGr04FNmXljy/nmz8S9wDMRcXZmXk79Miszb4+IFwDNjwyuBQ7MzPs7XE+SJGlamXUHj1lXkvqbd9BKGghN3/L/HPCzemyoQ5nvA5uAcyLi5zPzmXr+l4EDgG2NOpm5KTNvm+LuS5IkSR2ZdSWpvzlAK2kgZGZGxP7A08AhEXF4XcU2msvU/Y+BvwWeBW6MiDURcQpwEfAw8LXpfwWSJElSe2ZdSepvDtBKGhj17oBNwLGU1Wz3ejwrIhpTvzwCnEFZvfZi4O+BI4G3ZebmaeqyJEmS1BWzriT1r3DqGEmDJCIOBjYD1wHnZ+ZDETHUPC9XRMwBPgv8KXAPMBc4LDPvrudnZeaevRqXJEmSesisK0n9yTtoJQ2UzHyK8vjWGcA767HmwPpi4DzgxcD9mflsZj7QFFiHDKySJEmaicy6ktSfvINW0kCKiOuB44FrgL+gLIhwGLAOeCNwIbDBFWolSZLUb8y6ktRfHKCVNJAiYi6wGngvJbBuAQ4GHgfWZea/9rB7kiRJ0riZdSWpvzhAK2mgRcQvAkcD+wNPZebN9XjA3gsrSJIkSf3CrCtJ/cEBWklq0bqQgiRJkrSvMOtK0szjAK0kSZIkSZIk9cisXndAkiRJkiRJkgaVA7SSJEmSJEmS1CMO0EqSJEmSJElSjzhAK0mSJEmSJEk94gCtJEmSJEmSJPWIA7SSJEmSJEmS1CMO0EqSJEmSJElSjzhAK0mSJEmSJEk94gCtpCkTEdmy7Y6IRyLi6xHxji7LPxYRwxGxKiKiTZ35beq12+Y31Vnfcm5PRDwZEVsi4qsR8b6IeFmH19S43uYRXvfiiLgiIu6LiKciYldt+9qIWBERQ2Pod/O2tLY/3EXZ9aO8RzsjYntEfDsiPhERrxvDH22j3VW1rY3jqHtzrftARAyNUG5jh8/Fo50+R7Xe5i7eo1VN5ZfWY8NjfS2SJGkwmXXNuiPUNetKGpPZve6ApIFwYd3vBxwNLAeWRcRxmfnHo5R/OXA6cCLwWuA9Ha7xBPCxEfrweJtjtwHD9ecXAi8FlgBvBi6MiPWZeckIbf4/EbEf8NfAWmB3bf9GYBcwDzgJ+G3gi8CZPP86m11Q9+3ObW75/ao2xxqG2xxrfo9mA3OARcA64N0R8c/AGZn5UIc2J0VELAROBpLyvrwZuGGUal8B7q4/7w8sBE6jfI5emZkf6FDv47T/s6epPUmSpIkw65p1n2PWlTQumenm5uY2JRsllGSb4ycDe+o2v4vySyghcA+woOXc/Fpv8xj6tb7WWd/mXFCC5aO1zPndXg/YUM/dCxzd5vwQ8E7gmrG+Zy1lhmu5pV2+3hHfI0oAvLWW+R5wYJftrqp1No7xc/HntV5jf90IZTfWMqvanDuuntvZ2mdKmM/mz9cofVpayw9P1uffzc3Nzc3Nbd/ezLp7nTfrplnXzc1tfJtTHEiadpl5C/BDSkA8vovydzSVP26K+5aZ+UXgbfXQBRFn1aeeAAAFyElEQVTx0tHqRcSvAWcBjwFvzMz72rS9OzOvpgTXGSMz7wfeQnmPj6XcFTElImI2Jew+CVwEfBf4jU6P2Y0kM++ivN8HAgdPYjclSZLGzaxr1sWsK2mMHKCV1CuNObZyjOWfnYK+7CUzbwVupwSi3+qiytl1vyEz/3uUtndNsHuTLjOfBv6y/tp2rqtJchowF/hCZu6k3DUwBPzeWBuKiNdQHl3bkpkPT2YnJUmSJsisO4OYdSXNdM5BK2naRcTrKfNzJXBnF+V/vZZ/Bvi3DsUOaV4soMW2zLxsHF0dBk4AFgOfHKXsCXV/yziuM16rGosptHFZZm4bY3vDdf/qiJidmT8bd886W1P3V9b95yhh+V0R8aHM3NOh3vKmxS/2pzzKdhqwFfjdEa53TkR0mpfrksz83y77LUmS1BWz7qQx65p1pYHhAK2kKdcUJpsXTgjgo5m5ZZTyjYUTAjh3hG/sX8Lziw60ugcYT2h9sO4P76Js49GwreO4znitHOHcl4GxhtbG6x2ifFu/fTyd6iQijgJOAe7LzG8BZOajEXED5c6NNwA3daj+1ro120kJvd8f4bJ/OMK5jwGGVkmSNCFm3Slj1jXrSgPDAVpJ06ERJpOyyug3gSvqHFUjlW9I4F2ZeWW7wtWWzJw/oV7ubayPpo217EQty8zhSWwvmn6eitdxJmVqnY0txzdSQusaOofW1Zm5ESAihigr4q6kLILx1oh4bWb+tE29BZm5eYL9liRJGolZd2qYdc260sBwDlpJUy4zo26zMnNOZi4bIbA+Vx54EeVb6AeAyyLipOnqc3Vk3Xcz51Pjbod5U9SX6dB4vbuB/5nMhmvQXE1ZnfgfWk5/jXIHxKkRMXe0tuoCFFsy8yLKXQVHA78/mf2VJEnqllm3b5h1Jc1YDtBKmrEyc0dm/gtwKuVRpKsi4qBp7MKyut/URdnb6/7kKerLdGi83rumYE6u3wReRvl3Z2tEZGOjLIYxl/JUx1gXUGj82SyetJ5KkiRNA7PutDPrSpqxHKCVNONl5r3Apynf2P/RdFyz3sGwhDL305e6qLKh7tdExBGjtH3ABLs36er/DLy3/vrZKbjEWXV/A3BFm21jPX9mRMRetTs7tO7990ySJPUls+7UM+tKmumcg1ZSv7gYWAWcGxGfysxJfSypoQam0ykhGeCCblaIzcw7IuLTlHB2U0SsyMz/bGl7FvD22v6Kye35+EXEAuDvgGOA7wGXT3L784A3UR4l+51Oq8lGxMspKwS/Hri5i3YPpTxKBs+vyitJktSPzLpTxKwrqR84QCupL2TmgxFxOWWl0vOA97cUOaRpRdx2NraZQH9pU50XUOalWgIsAHYB78vMj4yhm++mzGm1FvhBRAxTVtXdRXnk6STKnRHXjqHNkayKiKUdzt2dmV9uOdb8Hs2mfCO/CPhVyrfyNwErM3PXGPtxQkRs7HDuu5RVcoeAqzsF1uozlNC6hr1D6/KImF9/biyccCpwGHAnnVcuPiciHu9wbrjNwhPHjPBafpyZf9ap85IkSeNl1m3LrGvWlQZGZE7nIoySBkmdc4m6CMKEy9fHqe6vvy7MzIdqkPlRF80/twpsDW7Nq+cmsAN4DPh34DZKwHqwTR8a1+u4km5E/AoleL2OEoT3A7YD3wE+D1ybmXs61B31Path+MRO56urMnNVS5+b7QKepLyfdwJfyMzbGYOIWAWMtNowwPXAscAvAIvqI3yd2jsI+AlwEDAvM7fXALmyTfGngB8C1wB/0xqGI2IzcNQofbswM9fX8kuBW0cpf09mHjtKGUmSNCDMumZdzLqSJokDtJIkSZIkSZLUI040LUmSJEmSJEk94gCtJEmSJEmSJPWIA7SSJEmSJEmS1CMO0EqSJEmSJElSjzhAK0mSJEmSJEk94gCtJEmSJEmSJPWIA7SSJEmSJEmS1CMO0EqSJEmSJElSjzhAK0mSJEmSJEk98n9xjSgZD1VnRAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1440x576 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(cnf_matrix_train, classes=['Peaking_bkg', 'Continuum_bkg',\"Signal\"],\n",
    "                      title1='Confusion matrix DDN train set \\n with normalization',title2='Confusion matrix DDN train set \\n without normalization')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "# On validation set\n",
    "Y_val_predict = Best_Model.predict_proba(X_Validation)\n",
    "Y_val_single_col = np.where(Y_Validation==1)[1]\n",
    "Y_val_mo = pd.get_dummies(Y_val_single_col[:])\n",
    "Y_val_predict_mo = Y_val_predict.round()\n",
    "cnf_matrix_val = confusion_matrix(Y_val_mo.values.argmax(axis=1), Y_val_predict_mo.argmax(axis=1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x576 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(cnf_matrix_val, classes=['Peaking_bkg', 'Continuum_bkg',\"Signal\"],\n",
    "                      title1='Confusion matrix DDN validation set \\n with normalization',title2='Confusion matrix DDN validation set \\n without normalization')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Saving Model, weights, standard scaler and PCA if used"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "import joblib\n",
    "\n",
    "joblib.dump(ss, 'Scaler.pkl')\n",
    "if PCA_ON :     \n",
    "    joblib.dump(pca, 'PCA.pkl')\n",
    "    \n",
    "Best_Model.save(\"Best_Model.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dropout_148 (Dropout)        (None, 83)                0         \n",
      "_________________________________________________________________\n",
      "dense_662 (Dense)            (None, 30)                2520      \n",
      "_________________________________________________________________\n",
      "dense_663 (Dense)            (None, 70)                2170      \n",
      "_________________________________________________________________\n",
      "dense_664 (Dense)            (None, 30)                2130      \n",
      "_________________________________________________________________\n",
      "dense_665 (Dense)            (None, 3)                 93        \n",
      "=================================================================\n",
      "Total params: 6,913\n",
      "Trainable params: 6,913\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "Best_Model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Try Loading the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dropout_148 (Dropout)        (None, 83)                0         \n",
      "_________________________________________________________________\n",
      "dense_662 (Dense)            (None, 30)                2520      \n",
      "_________________________________________________________________\n",
      "dense_663 (Dense)            (None, 70)                2170      \n",
      "_________________________________________________________________\n",
      "dense_664 (Dense)            (None, 30)                2130      \n",
      "_________________________________________________________________\n",
      "dense_665 (Dense)            (None, 3)                 93        \n",
      "=================================================================\n",
      "Total params: 6,913\n",
      "Trainable params: 6,913\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from keras.models import load_model\n",
    "import os\n",
    "\n",
    "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3' \n",
    "warnings.filterwarnings(\"ignore\")\n",
    "Loaded_Model = load_model('Best_Model.h5')\n",
    "Loaded_Model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "93420/93420 [==============================] - 18s 195us/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.06013732372881421, 0.9830443159922929]"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Loaded_Model.evaluate(pcTest, Y_Test, verbose=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4 XGBOOST"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To achieve the task of the classification we have tried to look at the results of the Xgboost algorithm. <br>\n",
    "In fact, besides being one of the best performing algorithms to perform this type of task, [here](https://github.com/dmlc/xgboost/tree/master/demo) you can see, not only great tutorials, but also that the algorithm won several high energy physics ML competitions. <br><br>\n",
    "The basic idea of the algorithm is to construct ensembles of decision trees. It work iteratively and at each step, we compute the gradient of the cost function w.r.t.the predicted value of the ensemble and add trees that move us in the direction of the gradient."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 2 candidates, totalling 6 fits\n",
      "[CV] n_estimators=100 ................................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ................................. n_estimators=100, total= 7.0min\n",
      "[CV] n_estimators=100 ................................................\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   1 out of   1 | elapsed:  7.0min remaining:    0.0s\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[CV] ................................. n_estimators=100, total= 6.9min\n",
      "[CV] n_estimators=100 ................................................\n",
      "[CV] ................................. n_estimators=100, total= 7.1min\n",
      "[CV] n_estimators=150 ................................................\n",
      "[CV] ................................. n_estimators=150, total=10.6min\n",
      "[CV] n_estimators=150 ................................................\n",
      "[CV] ................................. n_estimators=150, total=10.3min\n",
      "[CV] n_estimators=150 ................................................\n",
      "[CV] ................................. n_estimators=150, total=10.6min\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done   6 out of   6 | elapsed: 52.9min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1h 10min 54s, sys: 0 ns, total: 1h 10min 54s\n",
      "Wall time: 1h 11min 7s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "import warnings\n",
    "warnings.filterwarnings(action='ignore', category=DeprecationWarning)\n",
    "import xgboost as xgb\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "import os\n",
    "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3' \n",
    "\n",
    "param_dist = {\"n_estimators\": [100,150]}\n",
    "\n",
    "columnsNamesArr = X_train_S.columns.values\n",
    "X_Train = pd.DataFrame(X_Train, columns = columnsNamesArr)\n",
    "X_Validation = pd.DataFrame(X_Validation, columns = columnsNamesArr)\n",
    "\n",
    "Y_Train_single_col=np.where(Y_Train==1)[1]\n",
    "Y_Validation_single_col=np.where(Y_Validation==1)[1]\n",
    "Y_Test_single_col=np.where(Y_Test==1)[1]\n",
    "\n",
    "xgb_clf = xgb.XGBClassifier(objective=\"multi:softmax\",tree_method='approx',max_depth=4,seed=\"1432\",nthread=-1)\n",
    "xgb_scv = GridSearchCV(xgb_clf, param_grid=param_dist,cv=3,return_train_score=True,verbose=2)\n",
    "model_xgboost = xgb_scv.fit(X_Train, Y_Train_single_col) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>param_n_estimators</th>\n",
       "      <th>mean_train_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100</td>\n",
       "      <td>0.987581</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>150</td>\n",
       "      <td>0.989283</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  param_n_estimators  mean_train_score\n",
       "0                100          0.987581\n",
       "1                150          0.989283"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#results_large = pd.DataFrame(clf_large_CV.cv_results_)\n",
    "#results_large[[\"params\", \"mean_test_score\"]]\n",
    "result_xg=pd.DataFrame(model_xgboost.cv_results_)\n",
    "result_xg=result_xg[[\"param_n_estimators\",\"mean_train_score\"]]\n",
    "result_xg"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model Accuracy with all features: 98.71%\n"
     ]
    }
   ],
   "source": [
    "print(\"Model Accuracy with all features: {:.2f}%\".format(100*model_xgboost.best_score_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_train= model_xgboost.best_estimator_.predict(X_Train)\n",
    "y_pred_val= model_xgboost.best_estimator_.predict(X_Validation)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[    79,    636,    368],\n",
       "       [     1, 122224,    939],\n",
       "       [     1,    697, 108600]])"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cnf_matrix_xg_train = confusion_matrix(Y_Train_single_col,y_pred_train)\n",
    "cnf_matrix_xg_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABWgAAAJVCAYAAABDMBy5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nOzdd5xU1fnH8c9D771Jc7GC2KLYUBQU7L0E1ETRGDVRY+9GiRpLbIkxlp+aEGMixt5iV+xR0aCiICqioqLSi9Td5/fHOcPenZ2ZLczuLMP3/Xrd18zce86557aZZ+4991xzd0RERERERERERESk/jUqdAVERERERERERERE1lY6QSsiIiIiIiIiIiJSIDpBKyIiIiIiIiIiIlIgOkErIiIiIiIiIiIiUiA6QSsiIiIiIiIiIiJSIDpBKyIiIiIiIiIiIlIgOkErksbMNjOzf5vZt2a20szczCYWsD5DYx28UHWQqpnZ6Lidphe6LoWk9bDmM7OxcRuOLXRdREQk/xTrSm0oxgu0HtZ8inWloWpS6ApIcTKzxsAhwL7A9kA3oBUwD5gKvAL8090nFaySGZhZP+A1oG0cNQdYAcwqWKWkTpjZgcCWwER3f7jQ9akNM1sPeA9oA9zg7mfkSLsOMAnoBNzp7sflSDscOBjYEegJdAB+BGbG+T0HPOTuP+RpUdZYZnYaYf087O61+nNrZh2A0+LHP7r7vHzVT3IzsxJgNIC7jylgVURkDaNYVxo6xbqKdfNBse6aTbHumkUtaCXvzGx74CPgXuDnwIaEgHUh0JnwQ3ge8IGZPWBmzQpV1wxOIASsnwJ93L2zu/dw9+EFrNOPwMdxkPw5ELgkvubDfMI2+ixP5VXJ3acBZ8ePp5rZkBzJ7yAErF8CGYNbM9vIzN4EngV+BWwe8ywEmgIbAYcBtwFfmdk1+ViONdxphP1oy9Uoo0Ms45L4vtC+JezL3xa6IvWghPJ1LyJSLYp1806xbt1QrJtGsW6tKNZds5WgWHeNoRO0kldmth8wnvDjNhs4H9jI3Zu5e2egGbANcBWwgHDlslVhapvRZvH1EXefUdCaRO7+lrv3d/f+ha6LZOfuD8XttFs9z/dWQpDZCBhrZq3T05jZL4G9AQeOdfcFGdJsD7wNbAssBq4BBgHN3L2Tu7cCOgL7Af8AjPCnVIqMu58f9+XzC10XEZGGRrFu/inWXTMo1pVioVhXGiqdoJW8MbMNgbuB5oRWBVu6+1Xu/kkqjbuXuvuE+GXYD3ikMLXNKhVALypoLURq5heEP4HrEYLNVeJtLdfFjze7+/Ppmc2sG/Ag0A74CtjG3c9x93fcvTSVzt3nufvj7n4UobXQE3WwLCIiIg2SYl2RglGsKyLFz901aMjLQLjNy4ElhJYE1c1nGcb1IPz4fkgIIBfH938AumcppyTO3+P77sCfgM+BpcB3wDigf4a80xN5Mw1DY7ox8fP4HMszNJUvy/TtgH8m6rUY+AJ4Cfgt0Lsm5RVifVVzu1aoN+EWonuAb+I+Mhk4C2iSyLMj8DDhdpOlhH6kTsq0j8T03YBjCQHXZMKtV0sIt+3dAQzMVa+qtnfavjGa0P/VpcAHhFuhHCiJ6UbHz9PT5rcjsDJOOy3LcvQmtMJx4P9qub6PifnLgBGpY4vQyseBT4BWWfJem8i7/Wp+D1RYD8AI4Engh7htPgQuAlpUUc76wC2x3ksIQfm7wMVAuyryto/p3o35lsRybgHWy5GvZdwn3wDmEvrk+4HwJ/zvwCGJtGOq2o+qub7GV1HO+ETa9HU7jPLjpRQYm0jbl3DsPEHoC3Ex4bvhI+CPQN8cdRob5zM2R33HxP3rl8CbcT0vjOvuZ6ux/zQBjo/zmRW3wWzCbWj3ElrF5PoevIrQd9x8wnfINMJ3wSYZ0k+vYt1XWn4NGjSs3QOKdVNlDSXHbx2KdRXrVq6TYt3K5SjWVayrWFdDxW1W6ApoKI6BEPCUxgP9jtUsa5f4g5H64kh92aY+zwF2ypCvJJFmH0LQlcq/NDFtPrBFWt63CZ3CL49pFsXPqWFwTDcm/YckQz2GkuVHCziaEBik6rI01if5RTm6uuUVan1VczsOTZSxFyFwcMLDM5Lr4J6Y/jhCcFcW0yTXyVVZ5jE2Ld18wo9ccv0ekpZncNymqfosSdvWq7Z3TD89pjuT8MPpwLLEOi+J6UaTIWiN036byPeTtGmNKA8EJpMlsKzmOn88lvMlIXA7PX4uBXbMkqcJIeBw4Mk8fBesWg+EoCm1rVNBYGrbvAt0zFLGT9P2wQVpn78EBmTJO5DQMiKVNhXwZt0nYr62wMREurIMdZ6eSH9W3FdS33vz0/ejaq6vBwmBcWoeP6SV82CWdfubxLqdR/juGptIOz5RZipNadrnSt8LacfV2AzTUuVeRgiYPa6j9O+x39Vi32kMPJOh3sltn+17cF/K/0h6XB/J78FlwFFped4mfD+m0qR/D/xpdY8HDRo0FM+AYt1kWUOzfSejWDf126VYt3yaYt3KZSjWVaybrKdiXQ1hmxW6AhqKYwBGJQ78fVajnD6UBwMfkvihBYYAU+K02UCvtLwliTrMAV4FBsVpTYDhhCvaDrycZf6pL+QxWaaPoZZBK+GWstSP5z+A9RPTWgNbE1oB7F2d8hrC+qpiWw5NlD+X0EKhb5zWFrgiMf28+CNzI9AtpukI/I3yoKtSSxVCZ+eXETqtbx3HNSIELndT/gekZ4a8Y8nyw5yWbnpMt5Bw9fYgoGmc1psYZJI7aG0EvBinf5yqa5x2MeXBVI3/HKTNZx3KWyf8h/DQDQeuyZFncGI7/Gp15p+2HhbHbfpvwkNIAFoQHk6SCkIezJB/K8r/PL4KbJ5Yh/sl9slPgTZpedsSriI7MIPQF1mjOG0LwhXvjOua0NIhdawcDDRPzLcnoQ+ySi0+EvvH6NVYZyWJbVBSjXW7hPAH72+JdduYit8pNwHnAgOAlonjeltCKw8Hvk5Nq+6xQfl35BxCQHl0ovzewKOUH7Mb1nA9/CyxfL9IbV9C64VuhGPv/gz5tiUEpQ7cCvQHGsdpfYG/UB5cD0rLOzS17ld339egQUNxDyjWTabJ+N2JYl3Fuop1FetmXmcliW1QUo11q1i3Yj7FumvJUPAKaCiOgRA4pL50KwUINSjnlsQXYo8M03tTfvXqprRpyS/+yVm+jPdLpOmdYXrqC3lMlvqNofZB67aUB1FNsuWvbnkNYX1Vt96EK4WZbu97OZHm9gzTGxNuQ3PgolrsT49ny0vNg9aVpLUISEs3mixBa5zei3AbiwN/i+OSt4T9prbHTdp8jkisUyf8mWmeI/1xibQ75GH+oxPljScGjWlpfpFIs03atFRAlfE2NeAnlF/pPytt2rlx/HJg0wx52yb2p8fTpv0njj+/hsub2j9Gr8Y6Sx6LJdVctw+sxvwaE26NcjLcopXr2KBia4VhGaY3JwTDDlxYw3rdHPPdVsN8b8V8l+ZI86eY5uG08UNTy7O6+74GDRqKe0CxbjJNxu9OFOsq1lWsm0qjWLdiGcljsaSa61axbnk+xbpryaCHhEm+dE68n1ObAszMCLd7ANzq7jPT03h42uyt8eOoHMVd5+5LMox/kvCDBuVPsa0v8+JrMyqur1pZw9bX1R5/IdI8nXh/ZfpED532Pxc/bl6L+aY69t+pFnnTPeXu/6ttZnf/mtCHGMBoM/s18C9CAPGEu9+YhzoC3Ee4hS/ldHdfliN9lceumTU3s5lZhpE5yr7c3csyjP8b4ao/JPZLM+sA7BE/XuPuP6ZnjNvgwfjx8LTJqbrc7+6TMuRdSGi5A7CXmbVPTE4dn+tkWZaGptLxUl3xuHoqfqztsfGau7+YoexllB/XNT1mU9ugR3UzmNkWhKelr6D8ASGZ3BVfh5tZ4xrWS0QEFOtWh2LdyhTrKtZVrFs7inVRrLu20QlayRfLQxn9gE7x/XM50j0bXzubWb8sad7MNNLdVxL6vSExr/ryGeE2rKbAm2Z2rpltuRpfoGvS+nory/hUcDXH3adVkaZjpolmtoWZ3Wxm75vZAjMrMzM3MydcpYTQumJ1vba6Bbj7o4TbcSDcitKXcCvZ6NUtO+FCQj95KcfloUyLZWYaWmbJsxJ4JdOEGMiOjx8HJSZtRfl3SXX26c3NrCmAmTWjPEiqTt5GcX4pj8fXk83sHjM70My65CinkJYQ+jXLycyGmNlYM5tiZotSx0U8Ns6JyWp7bGT8zoi+ia81/c5ItezY38yeNLPDzaxnFXlSQXcj4ONsf64oD9Jbk4eTBiKyVlKsWzXFupUp1lWsq1i35hTrllOsuxbRCVrJl1mJ97UNbrol3n+dI92MxPtuWdIszJF/ZXxtWp1K5Uu8kjeKcNvJuoQnMP4PWGBmz5rZr8ysVQ2KXGPWV7yam6vsWs3fzE4m/Hj/itDqoQ3hNrfv4rAgJm1dwypn8n0eyoDQ4X5yex3r7rOyJa4JM9uKELRCCAAADjOzw3Jkm514n/HYdfel7m7JoRrVmVVFa4bUOkjukzXdp5tQXudOhBYa1c1bYX7u/i/Kbw0aBTwE/GBmn5jZX8xs6xxl1rfZWVprrGJmVxNuqzwa2JjQJ9pcyo+NxTFpbY+NvH9nuPurhFv3lgN7ElrdfG1mX5nZ38xsWIZsqaC2Mdn/WHUHkn9AavI9KyKSoli3Cop1c5atWDcPFOsq1k1RrKtYtxjpBK3ky4eJ9z/JQ3mZbhFanXQNgru/R+jU+xDg/4BJhKuywwlXwKeYWW1uryrK9ZWLmQ0A/kj4HruP0O9ZC3fv6O493L0HcEYqeR5mWZqHMiA8RbhX4vMu+SjUzJoDfycEcm8C+wP3xMl/MbOuWbJ+lHi/ZT7qEtXXvpZpPrnm7Vne4+6nEQK8Cwi3PM4DNgB+DUwwsz+uXlXzJue+aGYjKG81cDPhD11zd++UODZuSCWvu2rWnLtfQ2gxdTrhybnfE1o+jAZeMLP7Ui1JotQflSnpf6xyDNPrcZFEpHgo1q0Gxbr5o1i3IsW61Z63Yl3FutPrcZEkj3SCVvLlRSB1leugWpaRvGrbJ0e65G0KP2RNVTdSV8ta5EjTPsc03H25uz/o7ie4+2ZAV+BEQp9IfQiBR3WsCeurLh1K+MGaDIxy97fdfXlammr371MfzKwPcEf8+H58PcfMds1D8b8DNiU8tfXo2IrlZGAmYR+7OUu+tym/Qrx/HuqR0jUG0tmkAvfkfpx8n+t2pNS0lYQr5RCOn1Qwl+t4SE6rdDy4+6fufqW77024PWgHQvAEcKqZ5XMd1ZVUX2dPu/tJ7j4p7g9JDerYSHL3b9z9j+5+kLt3J9zOlzpuDiW0IkpJ9Ue4npnlo/WQiEg2inXLKdatH4p1K1Ksq1g3RbGuFCWdoJW8cPfvgAfixyPMbKPq5jWz1FWtzynvuH23HFmGx9fZ7v55jSq6+lI/kLl+FLerSYHuPtvdbyPc7gDwEzOrTr8xa8L6qkupbfBejltghmcZD+V/surlqmrsf+2fhP7FPgK2J9xa1Aj4RzW3ebaytyfcTgZwgbt/DODucwh/iAAONbOfpud19xWEFi4Ae8Sy8qEJWTrlj8f8zvHjhMSkdynfLtXZp9+L9Sf+YUn9EahO3jKq6NvK3cvc/b+EQOnLOHpEWrJ87EfJ/Tcf+2Pq2Mj4oI+4/vPxR6leuPsH7v5LyvvGS26D1Lhm1O6Eyap1n/gtEhGpRLFuBYp164di3fKyFesq1k1SrFt9inXXIDpBK/l0EbCIcBvTg2bWK1diM+toZg8Qr8K7uwP3xsknmFmlq16xE+0T4sd70qfXg/fia89MP+5m1g34ZaaMVVxhhdAZekqVtxitIeurLs2Pr5tl+rExs72AoTnyp/rs6pDnemVzETAEWAYc7uFJwscR+onqSXjaa42ZWUtgLKGFxSuEvqVWcfdHCMEyhNu/MvXN9gfCAxwM+He8pS4fLjSzTL8zRxMeGgHl+zDuPo/yp6KenamfOgtPMj0kfkzfp8fF10PNbNMMedtQfjvUf9x9fmJa1uMzXpFPtVhJPzbzsR8tSLzPx/6YWq4tskw/EVgvD/PJqxp8Rya3wQTKg/Pf57i9MTWP9L7n8r3uRaS4KdZVrFufFOuiWDdtsmLdQLFu9nko1l2D6QSt5I27TwV+TvhyHwhMtPD01g1SacyssZn9xMwuBaYBB6cVcwWhL5xOwHNmNjiRd0fCEys7EK6mX1WXy5PF68AX8f1YMxtkQSMzG0p4Wme242qUmb1mZieY2aofjLhO9qB8ed6IP97V0dDXV11KPalyICEY6wRgZq3N7ATgfio+FCDdpPg6xMz61101V22L38aPZ7v7+7Dqqv/PCFc297PwIIiauoLQl9Ri4JgsLSx+QwhKu5Dh9i93/55wLC4gXJF+28yuMbOtLfHkZTNrZWa7mFl1AuwfCa0K/mVmvWP+Fmb2S+CWmOYRd09/6vGFwApCf1hPW+ynLh5jexMeCNGE8KTo29Ly3kJobdMUeNLM9koFzbGcpwl9Pi0n/IlIetPMbjSzocnbh8ysp5n9OdYHyh9IkZLajw41s4xPX65KPN5TD3s4xsya1KachNSxsZeZ/Ta1PGbWwcwuAP5M7mOjUB42s7/G7bYqgDSzTmZ2EeWtRVZtg/jn/UTCn8G+hO14aPIPj5n1MrOfmdmzwNVp85xK+R+S49SyQERyUayrWLeeKdYNFOuWU6wbKNZVrFuc3F2DhrwOwI7AJ4ROyVPDMsKXZGliXBnhyYVN0/LvQgjEUukWxSH1eS4wJMN8SxJpSnLUb3pMMzrDtPFx2pgc+fcgfMml5rWYcLXLCV+Ao1LT0vKNTlsnSwlPBE6uk6+B/mn5hmYqryGsryr2g5z1Tlsn03OkGRPTjM8w7Z60dTqX0FeTE642npytfMLtV98n8v4Ql3U6sH1Nlz/bshD+NHwRpz2eJe+lcfoSYLMarOOdCceRA7+uIu1+iWUdmSXNxoSHLiTXaSnh2J2XmFdq/70e6JBtPQAnJfLMoeJxMxHonKUeIwnfGam08yk/xpxwC9aALHk3JbTUSKVdEvMn631ojv089d00l4rHkQPXV7ENVgLfpPajGh4vF6XV8ctYzriaHC8xXVPCU22Ty5Pqt8yBx4HLyH5cjY3TxmaYNp6qvyPHZCu7inqPp+L6np+27ZzwkJRGGfKOIHyfptKtjJ8Xp+W/PUPeOxLTFxOO1+nAtTWpvwYNGtaeAcW6inWrUe+0dTI9R5oxKNbNNk/FupXzKtZVrJvaFop1i2xQC1rJO3d/jfD01sMJt5t8SvgSbkv44nwV+D3hR+cIj/3qJPK/FPNfR+gUvxHhdpTJwLUx3yv1szSVufvThNt3Hif8sDUGviJctd+a8o680z0KHEW4vec9wpdxe0Kn9W8RrjoPdPcpNaxPg15fdexI4DRCf0zLCNviA+B8wp+nRdkyuvtcQsAxjvBnoT2wbhxyPRijpm4nXO2cCRyTJc3vCC1WWgDjLNzKlVO8Uvw3wrZ+gfIr9Rm5+2PAP+LHmyzD7V/u/rG7bwfsDtxKWJdzgXZxPp8RWmv8Gujp7md4jhYw7v4Xwp+8pwiBUxkwBbgY2MHdM17Zdvd7Ca1FbovzbE4IQiYClwCbuvvkLHknxbxjYvqVMf9ncZkGuvv9GbKOimU/T2iZ0IwQ/H1BuDVtN3c/Iz2Tu79MeFrxc4Rjujvl+1FNXAGcSviztYLwcIh1qcUDDuJ36u6E/WpqLM8I3zO/IjwgI19Pas6nUwj9E/6HcOLDCLcRf0P4/jzE3Q/zDC1n3P1ZQsuP8wm/MfMJfxjLCP3g3UlY7lMyzPckwv6SaiHSl7Duu+RpuUSkyCjWVaxbjxTrKtZNz6tYV7GuYt0iZR7OqIuIiIiIiIiIiIhIPVMLWhEREREREREREZEC0QlaERERERERERERkQLRCVoRERERERERERGRAtEJWhEREREREREREZEC0QlaERERERERERERkQLRCVoRERERERERERGRAtEJWhEREREREREREZEC0QlaERERERERERERkQLRCVoRqRNmVmJmHoeSWpYxJuYfn9fKSd6Y2ei4jabXZFpDlY/9VkRERBoWxaXSUOTaj9bEfczMhqaOrULXRWRN16TQFRCRtU8MjEcDuPuYAlZF1iJmNia+Hevu0wtYFREREWkg1oa41MyGAkOB6e4+tqCVkTWCmXUATosf/+ju8wpZH5G1gU7QikhdWQF8nHifVAJcEt+Pqaf6SP2bT9gHvi50RaLUPjcemJ4lTa79VkRERNZMa3tcOpSwjC8BYwtaE8llFmE//bLQFQE6UH5cjAWynaD9kfJjS0RWg07QikidcPevgf6FrocUjrs/BDxU6HrUhPZbERGR4qPfd1kTuPtNwE2FrkdNuPtb6NgSyQv1QSsiIiIiIiIiIiJSIDpBKyKY2aTYufvJGabtkHiowv0Zpjc1s0Vx+q6J8RkfxhAfGPVi4rOnDWNz1HM3M3vCzH4ws6VmNtnMLjGzFrVc7rHJeZrZoWY23szmmNmPZjbRzE41s5zflbFz/PvM7GszW2Zms8zseTM7xswaZ8lT4SEAZnaImT1jZt+bWVmqv9T09Whm65rZ7Wb2ZVwHn5nZ5WbWOlH2pmZ2t5l9FdN8YmYXmVnTLHVpb2ajzOyfZvZBXP6lZvaFmf3LzLav5frN+JCw5MMEqjGMTsu7lZldbGYvx/otNbN5ZvZfMzvXzNpkqMfYtAcXvJg2j+mJtFU+RCSur4vN7F0zW2BmS+I6vsXM1suxPlLlDjWztnG7TYn5Z5vZ42a2XXXWrYiISLGyIo9LzewnZnZXIo6Za2avm9lpZtY8S54qHx5lGR7WlFpuym9V36WqWKuKuo+PecZY8EszezPGQwvN7A0z+1k1yjk4xj3fmdny+Pq4mR2UI8+quD3O+zgzezXGUKuWI309mNnmZnaPmX0TY67JZnaWmTVJlL2jmT1sZt/GbTLJzE4yM8tSl25mdqyZPRjLmx/L/tTM7jCzgdVdp2nlZtzOVh5TV2cYmsjXKC7bVRZi5Rlxfc82s5fM7ETL8P8gzv/zxKjP0+YxPpG2yoeEmVkPM7vGzD60cHwuju//YGbds+RJ/w/U3cz+ZGafx230nZmNMzO13pWioS4ORATgBWAgsCuVb6vZNfF+qJmZuyd/gLcFWgPLgDeqMa8fgHZAx/j5u7Tp8zNlMrOzgasTaZoRbqcZQwg2R7h7aTXmn5GZ3QScBJQBC4CWwBbAH4GtgKOz5LseOD1+9Fi3DoT1tivwMzM70N0X5pj3dcAZMf+8WIdMtgLujOUvIHyHrwdcCOxsZrsBuwP/BlpRvp42AC4DNgVGZSj3dMoDd4BF8bVvHEaZ2WnufmO2Zaih5VTe7klNgM5Zpr2TeJ/aVh2A7eJwlJkNc/fvE+nmx/mlAsC5sQ4pP1S34jHgfgroHUctJfRlt0EcjjGzI939gRzFrAO8G9MvjcvRCdgH2N3M9nP3p6tbJxERkSJTtHGpmZ0GXA9YIm9rYIc4HGNme7r7t9Woe3WUEpapTZzPCmBOWpoltSi3MaEbqwOAlYR+SNsC2wPbm9mG7n5JeiYzawbcBYyMo8oI66ALIQ7ax8zuAY5292zPAjBCrHtoIn/G2NnM9gIeBFrEdM0J2+kaYGvgcDM7DriV0HhtQUwzkLDv9QHOy1D0H6j43yAVl68fh59VIx6siSXkjp2bE+LhdH2BVxOfU9uqE7BzHI4wsz3cPbkfzCH0h9slfp5F2JeS06vFzHYBHk7U70fCf55N4nCcme3v7q9mKQLC9vgr0C3mJ74fCexlZju7+3vVrZNIg+XuGjRoWMsH4CDCD+UcoFHatOcoP/HowJZp0y+K48enjS+J4x0oSZs2NDWtinqNienmEoKCK4AucVo74HeJeRxbi+Uem1juZYQTle3itM7A7Ynyd82Q/+TE9NuAHnF8a8JTT1fEaeNyLNvC+Ho10DVOaw6sm2E9zo3bY5M4rSVwCiHYcsJJ2HnAuET+NsDliTKGZ6jLiYQ/C9sBHeI4A/oRTlCXxXn8JEPe0bHc6TWZVsV2+WdieTdMm/YscAwh4GySWA8HAVNivgezlJtaB0NzzDvXftsWmBanzQD2Jh4vhJP5b8RpS4Etcsx/DvAhMIzwZ8CAbRL1n07acahBgwYNGjSsLQNFGpcC+yamPwz0i+ObAT8nnORz4DWgcZZ5j89Rv6zLUZ381dw24xPbZh7hJGXLOK038GicXkpaDBfTXBunlwGXUh53dgR+n1g/V2XIO5by2HkFcCblcXsbYJ309RC31Tigb5zWNm631PTzCBftbwS6Jeryt8RybJShLpcQ4u4tgdZxXCPCicS7Y95FQM+abIvabKe4/6S2yxepfTKxTR4Gfgr0pDxubUOI07+O+a7PUG4JWY6ZGux3feI2cELsu2Ni2hDKY9/ZQK8c859DONE8KE5rAgwHvonTX16d/VqDhoYyFLwCGjRoKPwQA5HS+AO3VWJ8c8JVysWEK80OnJGW94U4/pK08Vl/1HP9kKelG5MoY0yWNA/E6c/WYrnHJsofnSXNhDj99rTxLWMw4cC/suQ9JVH+oBzLdl2OOibX4ySgeYY0dyXSPANYhjQvx+l31GI93ZQtL3k+QUv5n5sVZDiZXEXeXpS3SO2bYXpqHQ2t5vpO32/PjeOXA5tmyNuWcDuYA4/nmP/3xD8BadM3S6TZMdeyatCgQYMGDcU6UKRxKeEElQOvkHYCNk7fL1H+oVnmPT5H/bIuR3XyV3PbjE/UcViG6c0pP+l3Ydq0XpQ3XrgiS/nXJWKtddKmjU3M+5TqrAeqjosrxfhxeuNETHdRLdbT49ny5toWtdlOifWyANishvUcRPnJ5BbVPWZqsN/dQvkJ1h4Zpvem/GLLTTnmP5l4ISDHMdN7dfZtDRoawqA+aEUEd58LpG4LSd46tj3hRORrhNu6K0y30E/WDvHji3VYxWWEK+6ZPBJfN1+N8r8inOTM5NEs5Y8g3B4EIZjK5GYgdYva4VnSlFF+i1xVbnD3ZRnGJ2+Hv8rdPUea2qynJ+LrTrXIW20W+iy7OFfI4GIAACAASURBVH482d2fq0l+D09ofo/QInVwnqsH5bfj3e/ukzLMfyHhljcIt1u1z1LO/3nFLhhS+T+gvL+v1dmfRURE1ljFGJea2eaE27kBLvMM3R+4+2PAW/FjtrixoXjN3Sut4xinZos5DyG0fFwKXJWl3MsJ67cpoQuDTOYS7lyrjquriIsBrkyfGLdPKg5tyLHzRYRWzKXAqBhLVpu7TyA0HGhNaA2cz7oZoeUuwK3uPjPD/GcQupeAzN2wpVznFbtgSHmS8m7LNqttXUUaCp2gFZGUF+JrMhDeNTHtdULAtHOiU/3BhD6dlgBv1mHdPnT3RVmmfRNfO2WZXh1vu3u2fl+zlT8ovn7l7lMzZYzB3Qtp6dN9mulkXRZvZRmf7JPq7SrSdMw00czWM7NrzewdCw/dKk10+P+fmKx3prz5YGZDCP3rQrjNKmPgHR92cISZPWrhQWlLkg8tIPQ9l/e6xj7TUgF6rhPHz8bXRoQ+gzPJdazkY38WERFZ0xVbXJqKA1cCL+UoOxVHZIsbG4raxDKpZXrb3RdkyhhPzk9IS5/ubXdfnmVauqpi5znuPq2KNNli5y3M7GYze9/CQ9LKEvHozTFZXcbOowjdRACc7u7/yZKuWXwY2DMWHpS2NC127lZHde1H+T5Qndi5s5n1y5Im4/7m7ispf5aEYmdZ4+khYSKS8iKhL6chZtYk/uANi9NecPclZvZfYBdCwPTfxPTXs7TszJesD9giBLqwet9n1Sk//QmnqWDm6yrKnpGWPl11T85C9nqm6phqxZkrTaYntR4E3EO4LS1lAaGFgxP6tupIuLqed2a2IeFBE82Ax4Czs6RrRbhlbFhi9HLCbVOpB0l0IixjvuvaiXC7G+Te5jMS77Nt89rsbyIiImuTYotLUzHBrCrqVlXc2FCsEbFzNeLiWsVkZnYy8CfKG7w54Vb91LZtSeiXuK5i58GEfnINuNnd/5wlXTfCydFk69KlVHzoV1fCcuS7rsntV5PY+fMMaRQ7y1pBLWhFJOVlwg9cG2DbeDJsO0Kw8U5Mk96aYde08WujTLdN1SRdpVvc6pOZdSb0XdWcsB2HAq3cvb27d3f3HsBhdTj/ToTbwDoDE4EjcrRmvpDw52sJ4YFu6xL6y+rs7j1iXVNX2C1LGfmQa5t7lvciIiJSfcUal65u3FgM1vTYeQDhIbqNgPsId2+1cPeOiXj0jFTyOph/P8KDv1oQumr4TY7kNxBOzs4GjiX069vS3bsm6ppq8Vyo2Lk26USKkk7Qigiw6gpzKuDdldBnUjPCUzFTgVCqr6ldzaw15beT12U/Xw1V6up9nyrSpW4X+iFnqsLZm3CFfy6wn7u/lKGPpx51MePYbcCDwIaEvnr3y3HLIJT3TXWpu//R3b/M0K9YndSV0Eo3dRzk2ubJaQ11m4uIiDRoRRiXpuLGrrGv3GyyxY2pVoItcuTN1vd9Q1EssfOhhLuqJhP6fc3U5UJdxc4dCA0buhIeOvfTTP0Zx7RNgYPjx5Pd/W/p/cCaWWOgS13UlYotnXNt82TXCg11m4vUC52gFZGkVYEuidvIEtP/S3h67mBgN8KtJIvI3u9pNqtaSMYO5NdEqf6xepvZRpkSxKAntR5ruo7qSypg+tjdf8ySZngdzft2wq2JS4D944MCcknV9X+ZJppZCbBBjvypk7k13udi4P1+/LhbjqSpdVUGvFvT+YiIiMgqxRSXpuLGJoTYJ5tUHJG+DHPja64TXdvlmJZaxkLG3av6ls32INV4AnJVX7X1UquaS22D93Lc9ZX32DmecL0fGEA4+blvtr58o66Un9DPGDsTLnxkO+mfXLba7DefExo4QPVi59nunql7A5G1hk7QikhSKujdAdgrbRzuvoLw5NyWwAVx9KuxX7CaSAYTHWpRz4bgWcLtQgBjsqQ5AegZ399T1xWqpfnxdSMzqxSgmdmWwBH5nqmZ/RY4inDS9OfxKbJVSdV1iyzTsz0ROCW139V2nxsXXw81s03TJ5pZG+Cc+PE/7j4/PY2IiIhUW9HEpe7+PvBR/HhRvIhfgZntTflJ1vS48b342tPMts+QtxvwyxxVWN0YKB8eILQEbgGcmyXNBYRut1bE9A1RKr7bLNMJfTPbi9BlWL7dQjjRuQw40N2nV5F+AeWNEyrFzvHher+vIn9KjfebeJfbvfHjCWZWqVWxmfUk/F+ChvtfSaTe6AStiCS9RnjoUgvCD/ks4IO0NKnAOBVA1uY2sqlxPgDHrYmtaGM3AGPix8PN7FYz6w7hYVZmdgqhfyqAe939nQzFNATPEK6QdwL+aWa9YNUTX38ap+fqmL/GzOww4Hfx4wXuXt0A/Kn4epGZHZx6arOZ9TOzfwE/pbyFSSaT4uuRsS+7mrqF0BqgKfCkme1lZo1iHTYj9APWj7BvX1SL8kVERKRcscWlqZOSQ4D7U0+sN7OmZnYk5SeoXif0MZr0OvBFfD/WzAZZ0MjMhgLjyf3fPhUDDYwPmKp37v414cFaAOeZ2e9ii1nMrIOZXUb5g2Kvd/dvC1HPakjFowOBv8TnKWBmrc3sBEIr19nZMteGmZ0N/CJ+PMbd36gqT+w27LX48Xoz2zURt24K/IfQWnlxlvzzKH+41zGpuLuGrgDmEf5nPJfc98xsR8IDzDoQWtpW1dBCpOjpBK2IrBJvcX8zMerFDH18pge+NQ6E43z+ET/+AVhkZl+Y2XQzu7am5RWKu99E6HwfwtXfb81sDuHK+o2EE3kvkrtFQ0G5+yfANfHjwcAMM5tHuEXw3via6+EDtbEP5bdKnWZmM3MMIxP5LgK+A9oSWlUsiXWdBhxOeIjY+2R3a3w9BJhnZjPiPvdqdSod+8PbnxCs9iYEtovNbH6c72BCq4Yj3f29rAWJiIhIlYotLnX3xwkPj3LgQGCamc0lxFp3E54J8AFwWHq/ovFW+hMILUs3Jtz+v4hwcu1FQtcJJ+WY/XjgY0Lfqa+Z2Zy4fNPN7NB8LWM1XAD8mxAHXgzMjrHzbMovbt8D/LYe61Qj7v485XdV/YqwDHMJ8f+thL5px+R5tvvE1zLghipi5+QJ+NMI+0gv4HngRzNbQNjPhhH+o8zKMd9U7HwK4bj4Mu4z43LkWSV2X3YgYd0MJOx7i8xsEfAqobuGeYQWwV9nL0lk7aATtCKSLhnYZnoK7gTKb3lZQO372TyJELykruj3Bdal7jqqrxPufgahb7QHCCcP2xBanL5IeFrqiHhir8Fy9/MI3Q28RegPtinwKeGq908of7prXehexdAyUc8vCFf670zUaSnwOLCHu1+Za0bufjfwc0JA+COwDmGf650rX1oZkwgB5hhgIuFWvebAZ4QgdqC731/d8kRERCSnoopL3f0GQixzN/AV0IoQe/2XcPJ2W3fPGHe5+9OE1rePE+4YahzLuArYGpiZKV/Mu5Jwe/wdwHSgNWH51iXErvXC3Ze7+0jCxfInCSdm28bXJ4GD3f2I2H1FQ3Yk4eTn+4SL840JJz3PB3YknDyvC42oOnZulkoc7+DblnBSfFbMvzB+Huzu/yC3K4BTCcfZCkLMvC41eAiau78E9AeuI5y8bkQ4QT8ZuBYY4O6vVLc8kWJmlS9CioiIiIiIiIiIiEh9UAtaERERERERERERkQLRCVoRERERERERERGRAtEJWhEREREREREREZEC0QlaERERERERERERkQJpUugKyNqjS5f2XtK3W6GrIWuaL78udA1kTdW5RaFrIGuod6bOneXuXetznhtYa/+R0ryW+S3Lnnb3PfNaqIhk1aVLBy8p6VnoasgaRw/tltqyQldA1lDvvDO5XmPduohzofhiXZ2glXpT0rcbE165odDVkDVM6a8vLHQVZA3V6KgBha6CrKEaDb/3i/qe54+UcgLr5rXMMUztktcCRSSnkpKeTJjwj0JXQ9Y0XlboGsiaynRDtNSO2aB6jXXrIs6F4ot1dYJWREREpMAM9TslIiIiIsVHcW71aB2JiIiIiIiIiIiIFIha0IqIiIg0ALpqLiIiIiLFSHFu1bSOREREREREREREpGiY2V/N7Hszm5QYd42ZTTGz983sITPrkJh2vpl9amYfm9keifFbm9kHcdqNZmZxfHMzuzeOf9PMShJ5jjazT+JwdHXqqxO0IiIiIgWW6psrn4OIiIiISKHVRZxbzVh3LLBn2rhngU3dfXNgKnA+gJltAowCBsY8N5tZ45jnFuB4YMM4pMr8BTDX3TcAbgCujmV1Ai4BtgO2BS4xs45VVVbxu4iIiEgDoBO0IiIiIlKMCnGC1t1fBuakjXvG3VfGj/8Fesf3BwDj3H2Zu38OfApsa2brAO3c/Q13d+Au4MBEnr/H9/cDu8XWtXsAz7r7HHefSzgpnH6iOOM6EhEREREREREREVlTdDGzCYnh+BrmPxZ4Mr7vBXyVmDYjjusV36ePr5AnnvSdD3TOUVZOekiYiIiISAOgq+YiIiIiUozqKM6d5e6DapPRzC4EVgL/TI3KkMxzjK9tnqz0X0BERERERERERESKXnxo177AkbHbAgitXPskkvUGvonje2cYXyGPmTUB2hO6VMhWVk46QSsiIiJSYFYHg4iIiIhIodVFnFvbWNfM9gTOBfZ39x8Tkx4FRplZczPrR3gY2Fvu/i2w0My2j/3LHgU8kshzdHx/KPBCPOH7NLC7mXWMDwfbPY7LSV0ciIiIiDQAumouIiIiIsWoEHGumd0DDCX0VTsDuAQ4H2gOPBvOt/Jfdz/R3T80s38DHxG6PjjJ3UtjUb8CxgItCX3WpvqtvRP4h5l9Smg5OwrA3eeY2WXA2zHdpe5e4WFlmegErYiIiIiIiIiIiBQNdz88w+g7c6T/PfD7DOMnAJtmGL8UOCxLWX8F/lrtyqITtCIiIiINglrQioiIiEgxUpxbNa0jERERERERERERkQJRC1oRERGRAjN01VxEREREio/i3OrRCVoRERGRBkCBq4iIiIgUI8W5VdM6EhERERERERERESkQtaAVERERKTDd+iUiIiIixUhxbvVoHYmIiIiIiIiIiIgUiFrQioiIiDQAumouIiIiIsVIcW7VdIJWREREpAFQ4CoiIiIixUhxbtW0jkREREREREREREQKRC1oRURERApMD08QERERkWKkOLd6tI5ERERERERERERECkQtaEVEREQaAF01FxEREZFipDi3ajpBKyIiIlJguvVLRERERIqR4tzq0ToSERERERERERERKRC1oBURERFpAHTVXERERESKkeLcqmkdiYiIiIiIiIiIiBSIWtCKiIiINABW6AqIiIiIiNQBxblV0wlaERERkQLTwxNEREREpBgpzq0erSMRERERERERERGRAlELWhEREZEGQFfNRURERKQYKc6tmtaRiIiIiIiIiIiISIGoBa2IiIhIgalvLhEREREpRopzq0cnaEVEREQaAAWuIiIiIlKMFOdWTetIREREREREREREpEDUglZERESkAdBVcxEREREpRopzq6Z1JCIiIiIiIiIiIlIgakErIiIiUmB6eIKIiIiIFCPFudWjE7QiIiIiDYACVxEREREpRopzq6Z1JCIiIiIiIiIiIlIgakErIiIiUmC69UtEREREipHi3OrROhIREREREREREREpELWgFREREWkAdNVcRERERIqR4tyq6QStiIiISAOgwFVEREREipHi3KppHYmIiIiIiIiIiIgUiFrQioiIiBSYxUFEREREpJgozq0etaAVERERERERERERKRC1oBURERFpAHTVXERERESKkeLcqukErYiIiEgDoMBVRERERIqR4tyqaR2JiIiIiIiIiIiIFIha0IqIiIgUmKGr5iIiIiJSfBTnVo/WkYiIiIiIiIiIiEiBqAWtiIiISANglucCPc/liYiIiIjUQt7jXCi6WFcnaEVEREQagEaW5yizyIJWEREREVkz5T3OhaKLddXFgYiIiIiIiIiIiEiBqAWtiIiISIEZdXTrl4iIiIhIASnOrR61oBXJo48mf8lu+1xIq66H0HODo7n4srspLS2tMt/8+Ys55sQ/0rH3KNr3HMmRx17L7NkLKqSxNvtlHJp3OqiuFkfq0UfzljLimWm0/ecH9LnvIy6ZOJPSstz3bExftJwmd71faTji5S8qpHN3rnj/O/rdP5nWd3/ANo9P5emvF9bl4kg9+eiL+Qw/+0Va73M/vUY+wsVjP6C0tCxnng+nz2ev816i18hHaLHXfax7+GMcd91bfDt7SdY8D782g0bD72WbXz+T70UQEZEiMHTo8ZgNyji88cb7AMybt5Bjj/0dnTrtSps2Q9hrr9/w6adfFbjmUt9WrlzJVVf9nQ03OoTmLXakd599Of3061dNX758BT8deT7rrX8gLVsNoWu33dlr71N5553JNS5Lisf99z/H4MHH0rnzbrRoMZiNNz6Yyy+/g+XLV1RI98EHn7LvvqfRvv0utG27M9tue1SlfWfcuKfZaqsjadNmCL167cVRR13MN9/8UJ+LI5KRWtCK5MncuYsYvt9v2aR/Hx659yI+mzaTMy+4k7Iy5/JLfp4z78ij/8DHn8zgjptOoVGjRpx78VgOHPV7Xnn26lVp3njhmkr59jvsMnbcYUDel0Xq19xlK9nj2WkMaN+CB4eVMG3hcs5+5xvKHC77SY8q8/9h63UY3K3Vqs9dmlf8ar960g9c/v73jNmyO1t0bMm/Pp/LgS9O5+U912ebLq3Si5M1xNyFyxlxzng26duehy/dic++WcRZt02krAwuP3azrPnmL15BSY/W/HxECT07t+TzmYu49B8f8u7Uubx18wiaNK547Xbp8lLOvGUi3Tu2qOtFWuupYYGIrKluvvk8FixYXGHcxRffyv/+9zHbbLMJACNHns+kSZ/xpz+dSfv2bbj88jvZbbdf8cEH42jXrk0hqi0FcMyxl/H8829zycXH0b9/CV999R0fffT5qumlpaWYGeefN5r11+/FggWLueGP97Drbr/mf+/ezXrr9ap2WVI8Zs+ez7BhW3P22T+nQ4e2vPXWJMaMuZ2ZM2dz003nAjBx4scMGfJLDjhgF+6990oA3n77I5YsWbaqnEcffYnDD7+Qk046jGuuOZVvv53FRRfdwr77nsaECf+gUSO1YawrinOrphO0Inly651PsmTpMh781wW0a9eKEbvCgoU/MuaKf3HO6YfQrl3mE2FvvDmFp597l5eeupKdd9oUgF49O7Pd0DN57sWJDB+2JQDbb9u/Qr6335nKrNkLOPzQXep2waTO3TZ1DktKnfuHrku7Zo0BWLCilEvf+46zB3ZdNS6bjdo3Z/uurTNOW15axtWTvufsTbtyzqbdANijV1smz/uEy977jkd365ffhZF6c+tjn7JkWSkPjNmRdq2bMmJrWPDjCn5314ecM7I/7Vo3zZhv8MAuDB7YZdXnoXSjd9dW7HHuS7w/bR5bbdipQvpr/j2FXl1asn7PNkyaPr9Ol2ltZ3Xx8AQRkXqwySbrVfi8fPkKJkyYzMiRI2jSpAlvvPE+zzzzX55//hZ23XUbALbbblP69duf//u/hzjrrNyNGaQ4PPXUG4wb9wzvTfxnpX0mpWXLFtw77ooK44YP35bOXUbw8MPjOeOMI6tdlhSPE044pMLnYcMGsWDBYv7yl/v485/Pwcw48cQr2W+/Idx992Wr0u255+AK+f71r6fYaqv+q07qArRr15oDDjiTjz/+ggED9N+orijOrZouD4jkyZPPvMMeu21V4UTsqEOHsGTJcl56dVKOfBPo3q3DqpOzANsO2oh+Jd158pl3sua7576Xad26BfvtvU1+FkAK5qmvF7J7zzYVTsSOLOnAklLnpe8W58hZtc8WLmfhijJ2W6diy5Td1mnDc98uYnkVt8NLw/XU2zPZY1CPCidiRw3ty5Jlpbz0/vc1Kqtzu2YALF9RcX/48rvFXHPvFP540k9Wv8IiIrLWeOqp15k7dwGHH74HABMnTqVJk8bssstWq9J0796ZzTffkCeeeLVQ1ZR69te/Pcquuw6q8QnV1q1b0qJFM5YvX7naZUnx6Ny5/aouDj76aBpvvjmJU04ZmTPPihUrad++4v+iDh3aAqFbOJFC0glakTyZMnUG/TfqXWFc3z7daNWqOVOmzqhRPoABG/fJms/due+h1zhgn+1o1Uq3Ha/pPl6wlI3bNa8wrm+bZrRqYnw8f2mV+Y977Sua/eN9et/3EWe+/Q1LVpafZFtaGgKNZo0q3lTSvLGxvMyZtmh5HpZACmHKVwvYuE+7CuP6dm9NqxaNmfJl1X0Ml5U5y1eU8vFXCzj/jvfZZuNObNu/c4U0Z902kcN26VOpVa3UAQsPT8jnICJSKOPGPUOvXt0YMiRc4Fu6dBlNmjShceOKdwU1b96MyZOnF6CGUghvvvkhG23Yl5NPvoZ27YfRqvUQDj7knIz9f7o7K1euZObMWZxzzp9p3Lgxhx++e63KkuJRWlrKjz8u5dVXJ3Ljjffyq18dipnx5puhQdTcuQvYYovDadJkO9Zf/wDuvPPhCvmPPfYAXnnlf9x11+MsWLCIqVO/4KKLbmHYMJ3sr1N1EOcWY6yrE7QieTJ33iI6dKh8m3nHDm2YO29RjnyLa5zvldc+ZMbXsxh16M61r7A0GHOXldIhQzcGHZs1Ye7y7A+Za97I+PXGnbl9cG+eGbEev9ywE7dNnc0RL3+5Ks16bZthwIRZFR8A9Xb8PGdZ1Q+xk4Zp7sLldGhTuRuDjm2aMbcaJ973ueBlWux1PwOOeZI5C5fz2OVDaJQ4kf/ixO94esJMfp+jP1sREZF0P/64lMcee4WRI0dg8R/0Bhv0YenSZXzwwaer0i1ZspRJkz5jzhx1n7O2mDlzNmP//gQT35vKuHsu529/vZh33pnCQQefU6n14tVX30XTZoNZp+fe/P2uJ/jPEzew7rrr1KosKR6tWw+hdeudGDLkOHbZZSuuueZUIOwPAEcddQlHHrknzz77F/bcczDHHXc5//lPeSv9ffbZibFjx3D88VfQvv1QNt74EEpLS3nwwcrPexGpb0V9gtbMxpiZJ4ZvzOwBM1s/z/MpieXvmyPN6Jim3nvAr+68zWy8md1fX/UqRpah62t3r7JD7Kz5slwWuue+l+nYsQ17DNdtx8Ui07bOtQ8ArNOqKTdu14v9+rRnaI82XLJlD64d1JPHZixg4pxwArZ9s8aM6teBKz/4nhdnLmLOspXcNHkWz38bWlg2LsZLj2uRzPtN9a4o33jyVrzx5+Hcdd52LFqykr0veJml8YLAytIyTr3pf1x45Cb06NQy39WWLNSqQGpKsW7N5q1Yt3489tjLLFr046ruDQD22GMH+vXrxQknXMHHH0/n229nceKJVzJ//qJKrWqleLk77s4jD1/D3nvvyMiRI/jHXWN4660PeeGFtyukHT16H95+ayyPPnIdW2/dn333O4OPPppWq7KkeLz++l955ZU7uO6603jkkZc4+eQ/AFBWFu4gPO64AznnnKMZNmwQf/nLuQwbNogrrxy7Kv+LL07gxBOv5NRTR/Hii7cybtwVzJmzgIMOOovSUjVcqUtqQVu1oj5BG80HdojDWcCWwPNmlvmJOnXniViHH+t5vlJPOnZow7z5lfsLnb/gRzp0yP5/oWOH1hnzzZu/mA7tK++mK1eW8sAjr3PIAYNp1izzQ4BkzdKxeWPmZWgpO39FGR2a1uxr+pB12wPwv9nlLWav36YnAzo0Z8Qz0+h270dc9+EPXLB5dwC6t9SzItdUHds2Y16GlrLzF6+gQ+tmVebfsHdbthvQmZ8NL+Gpq3bhf5/O5V/PfwHA7U9MY97i5Ry9ez/mLVrOvEXLWb6ijNJSZ96i5axYqb6L882ARuZ5HWStoVhXGpRx455hgw36MGjQJqvGNWvWlHHjfs93382mf/9D6dlzT6ZN+5qjjtqb7t3Vjc7aomPHdmy22fp07txh1biddtqSZs2a8tFHn1dI26NHFwYN2oT99hvCY49eR+fO7bnq6r/XqiwpHltt1Z+ddtqSM874GTfeeBa33HI/n302g06dwn+gYcMGVUi/666DKuwPZ555A/vvvzNXX/0bhg4dxMiRu/Pww9cyfvw7PPLIS/W6LGuTuohzizHWXRv+ma909//G9/81sy+BV4C9gfvqqxLu/gOgDnGKWP+NelfqM/arGT+wePHSjH3MJvO98vozlcZPmTqDA/fdvtL458e/xw+z5nO4ujcoGhu3a1Gpr9mvFi9n8coyNm5fsz6GUxcSk1cUu7ZownO7r8+MxcuZv6KMjds150+TZ9GjZRNK2lR9Ik8apv592vHxVxX7mv3q+x9ZvHQl/fu2rVFZ63ZvTae2zZj2bbhY9PGMBcz4YQnrHPZIpbSdDnyIu87bjp8NL6l13UUkrxTrSoMxf/4innzydc4556hK07bddlM+/fRhpk79giZNmrD++r3Zd9/T2H77TTOUJMVowIASli2rfHHZ3WnUKHujhCZNmrDZZhswbdo3q12WFI+ttuoPwOeff82AASUZ07hToQuvKVOmV2jdD7DxxiW0bNmczz7L/twYkfqwNn5zvRNfSwDMbCcze8nMfjSz2WZ2u5mt+mdrZuuY2V/NbJqZLTGzqWZ2uZnlPKthZkPNbKGZXRE/V7j1KnGr2E/N7DYzm29mM8zsd2bWKK2sw8zskzj/F83sJzHv6Bou+wAzeyWxHAdVsQztzew1M3vPzLrGcR3NbJyZLY630Z1rZtea2fQa1qXo7LX71jz9/LssXFjecOTeB16hZctm7LJT9sBzr90HMfO7ubz6+oerxk149xOmfT6TvXbfulL6e+57iR7dOzJ0Z/ULWSz27NWWZ75ZxMIV5a1o/z19Pi0bG7t0r1kDqAe+CP24bdW58m3pvVs3Y2CHFqx0Z+yncxi9gVqsrMn2HZKChwAAIABJREFU3KYHT0+YycIfV6wad+/4L2nZvDG7bN6tRmV9/NUCZi9YTr91wv528gEb8sK1wyoMewzqwUa92/LCtcMYsXWPvC6LBJbnQdZainUV6xbMQw+9yLJlyyudAEkxMzbeuIT11+/NJ598yXPPvcUvfnFAPddSCmXffXbi/fc/ZdaseavGvfzy/1ixYiVbbLFh1nxLly7j3Xen0K9fz9UuS4rHa6+9B0C/fr0YPHgLOnZsx/PPv1UhzfPPv8UWW2y06vO6667Du+9OqZBm8uTPWbJkGSUlPZG6k+84txhj3bWhBW26kvg608x2BJ4HHgYOBToDVwEd42eALsAc4AxgLrARMAboCpyQaQZmtgfwEHCVu19aRX3+ADwQ57cbcDHwIfDvWNYgYBxwP3AKMAC4t/qLW8G9wM3AFcBxwH1mtrW7v5dhGToBT8ePw9x9Tnw/FtgJOBWYCZxOWCdrfYctJ/5iL2685TEOPuJKzj3jEKZ9PpMxV9zDGScfSLt2rVal22Dz49llp0258+bfALDDdv3ZY/hWHHX8DVz7+2Np1Mg49+K/s9MOmzB82JYV5rFs2QoefvxNRh+5m64MF5ETNurETVNmcej4Lzh7YFc+X7ScS9/7jtM26Uq7xMPDNn5oCjt3b83tg/sA8LuJM1m4sowdu7ambdNGvPL9Yq778AcO6tuOzTuWn6C9+7O5rChz+rVtxleLV/Cnj36gsRnnbdq13pdV8ufE/Tbgzw9/wiFjXuOckf2Z9u1ifnfXh5x+yMa0a13e/cmGRz3Bzpt35c6ztgXgrNsm0qSRsd2AznRo05TJXyzgmn9PYf2ebRg1tC8AG/Rqywa9KrbC/fsznzNrwTKGblmzk78iUu9K4qtiXcW69W7cuGfYYouNGDCgX6Vpl112B/37l9ClSwc++OBTLrvsDkaN2p0RIyrfMSbF6fjjD+TGP9/LfvufwQXnH8PChYs597ybGD58W3baKfzvueeep3nyqdfZc48d6NmzK99+O4ubb7mfb7+dzRmnH1GjsqR47LnnKQwfvi0DB65H48aNee2197juursZOXIE668f7la9+OLjOOecG+nQoS3bbLMJDzzwAi+//D9eeun/VpVz4omHcPrp19OzZ1f22msw3303h0svvZ2Skp7svfeOhVo8EWAtOUFrZqnlXI8QtC0EniMEhq+7+8hE2q8J/XZt6u6T3P0DQn9eqemvAYuBv5rZKe5e4b4KM9s/lnuRu19bjeq97O5nxvfPmtmewMGxDIBzgcnAKA+Po3zKzJoCV9dgFaTckaqTmT0NfAScD4xKW4auhPWzCNjL3RfE8ZsC+wM/dff74rjnga9i2krM7HjgeIC+fYr7ZFDHjm14/vHLOfnM29jvsMvo0L41p590AGMuPLxCupUrSyt1QD5u7Nmcft4dHPvrGykrK2PfPbfhxmuOrzSPJ5+ZwPz5ixl16JA6XRapXx2bN+GZEevxm7e+5sAXp9OhWWNOHdCFS7boXiHdyjKnNNHVTv/2Lbj+ox/46ydzWFLq9G3dlDMHduWCzSqeQCvDuebD7/li0QraN2vMAX3acflPetCmqR7KsSbr2LYZz10zlFP+/C77//ZVOrRpymmHbMSYowZWSLeytIyysvIdZ9BGnbjp4anc/p9pLF1eSt9urTh4SG/OP3wTWqtP4oIqxocdSP1QrLtKYWPdvmv33QWzZs3j+eff4rLLfpVx+uzZ8znttOuYNWseffp056yzfsaZZ/6snmsphdSuXRteeP5mfnPqdYw6/EKaNWvKAQfszA3Xn74qzcYbr8vd/3yKM878I3PnLmSddbqw3XYDmfD2eQwcuH6NypLisc02mzB27GNMn/4tTZo0Zr31enHllSdx4omHrkpz2mlHUFZWxp///G/GjPm//2fvvuPkrur9j78+mxBSICQQSgAhQUCkCEIMoAgoXaWIAgGVIhZULOAV8F4FRLk/UbBwVbhICVioKkUTkCIIXFqsSFEwBAgtJASCgfTP74/5bpzdbHYmm5md2dnX8/GYR2bOnO+ZMwMM75w5hTe9aWOuueYs3vnOfx+s/bnPTWDQoFU477xrOP/8XzBixOrssst2/L//9xmGDfNg3Hoy51YWpRzUmiLidOC0TsVPAccBdwBzKP1S/+Oy59soBbCPZ+alERGUfkH/BDAWKN8QcrPMfDwixgBPUPrF/QjgxMz8Yae+HA1cAqyemf8qu+YjmfnTsno/BzbKzF2Kx9OAyzPzy2V1tqAUZI/JzIlVfA7tr71lZj5SVn4WcEhmblI8vh0YQGl2xfPA/pk5t4t2hmTmvLLyK4CdMnNMd/0Yt/1mOeXO71bqrtTB4k//V6O7oD6q7cg3N7oL6qPa9rzyD5k5rnLN2nlz2+C8bPDy9yvvifGv/7PX34d6l1l3mddubNYdt2VOmfKTSt2VOkoP3lQPhSsq1TMR43o1I9Yj50LrZd3+8F/0K8DbgHHAhsCYzJxMaWnXAEqzDBaW3eYDqwBvKK7/AnAOpWVcBwLjgc8Uz3U+vecASkvEfrUC/Xu50+MFndpdj2UPXOjpAQwzung8ulPZlpSWlv2kPLCW9eXV8sC6kv2RJEnSyjHr/ptZV5Ik9Un9YS3josyc0kX5y0BS2mNrUhfPtx8ReQhwdWYuncYXEVsu57U+S2n/rpsjYtfMnNXjXv/b85T2ACvX070C1gFmdXr8XKc6vwP+BFwQETMz84ZOfVk9IgZ3Cq6tvXeBJEl1FpFEtO6qJtWVWfffzLqSJDUZc251+sMM2i4Vv5jfC7wpM6d0cWsPrUMozTQo96HlNDsH2IdSGL4pIobXoKsPAPsXy8/aHdDDtpaeZFucnnsgcH/nSpl5JqWZFFdHxLvLnmoP/0tfPyKGAHv1sD+SJEmqA7OuWVeSpP4sIi6OiBkR8beysjUj4uaIeKz4c2TZc1+OiMcj4u/Fgajt5TtExIPFc+e2Z5aIWDUirizK7yu2d2q/5qjiNR6LiKOq6W9/mEHbnZMoHZKwhNLJsa8CGwHvBf4rM/8B3Ax8LiLuA/5JKbBuurwGM3NWROwF3An8OiL2zczXVqKPZwH3AVdExCWUlmR9vHhuRTcs+lhELAD+VrSxKXB4VxUz85SIWB24LiL2ysx7M/NvEXEDcF7x3POUZlG81oO+SJKkMm0enqDaM+uadSVJargG5dyJwA+Ay8rKTgFuzcxvRsQpxeOTi9VDE4CtgPWBWyJi88xcDJxHaa/+eymtStoXmAwcC8zOzE0jYgKlTHNYRKxJ6YyAcZR+1P5DRFyfmbO762y/nUELkJl3AbtSWrb0E+AGSkH2aeCFotoZwOXAN4o/FwCfq9Duc8AewBjglxExaCX6OIVSsNwBuBb4ANB+NOqcFWxuAqWZBdcC2wKHZeafuql/PPALYHJEbFuUHU3p1NtzgYspHUBxYw/6IkmSykTU9iaZdc26kiQ1g1rn3Gqybmb+ntLe+eUOBC4t7l8KHFRWfkVmzs/MJ4DHgfERMRoYnpn3ZGZSGuw9qIu2rgH2KGbX7gPcnJkvFYOyN1Ma1O1WS8+gzczTKe271V2d++jmg8rMfwHHdPFUlNWZVv64KHuS0gyFdhOL23KvKcqP7qLsKuCqpS8c8eHi7l+W1+9O15e/9ju6qbd7p8dJKaSWl70EHFbWl4GUZincV01fJEmSVBtm3aXXl7+2WVeSpP5hVESU78N/QWZeUOGadYsfmsnM5yJinaJ8A0ozZNtNL8oWFvc7l7df83TR1qKIeAVYq7y8i2uWq6UHaFtFRJxHacR9NrA98BXgN8Wofm/35RBK070fBIZTWj62GXBkb/dFkqRWEnh4gvons64kSa2tTjl3ZmaOq1FbXc3JzW7Ke3rNcjlA2zesBfyo+HMWcCWl5WkU06cHdHPt4mJ2QK3MpTTLYtPidR8E9s/MZQ5gkCRJzS0i9gW+T+n/6Rdm5jc7Pb8G8FNKMyUHAmdn5iW93lG1OrOuJEnqDS9ExOhi9uxoYEZRPh14Q1m9DYFni/INuygvv2Z6seJmDUpbKkwHdu90ze2VOtav96DtKzLz0MxcLzNXKf48LjPb98E6itKU6+XdqjotbgX6Mikzt8vM1TJzSGaOz8xJtXwNSZL6m6D39+WKiAHAD4H9gC2Bw4sDEsp9Bng4M7elFDTPWZn9RqWumHUlSWpd9ci5K3HewvX8OzscBVxXVj4hIlaNiLGUVs/cX2yH8GpE7FT8aHxkp2va2/ogcFvxo/FNwN4RMTIiRgJ7F2XdcgZt33cD8LZunu/1pWGSJGnFNeBgr/HA45k5tfT6cQWlww4eLquTwOpFIF2N0qyARb3dUfVrZl1Jkvq4RhxgGxGXU5pgMCoipgOnAd8EroqIY4GngEMAMvOhiLiKUg5eBHwmMxcXTX2K0l73Q4DJxQ3gIuAnEfE4pYw8oWjrpYj4OvBAUe+MYo/7bjlA28dl5ixKS8EkSZLKVTo4oasDDHbs1MYPKM0OeBZYHTgsM5fUo7NSV8y6kiSpJzLz8OU8tcdy6p8JnNlF+RRg6y7K51EM8Hbx3MXAxVV3FgdoJUmSGi+gLWp+eEKlgxOqOcBgH+DPwLuBNwI3R8SdZcvPJUmSpOWrT85tOe5BK0mS1D8t7zCEcscAv8ySxyktJ9+il/onSZIk9QsO0EqSJDWBBhyc8ACwWUSMLQ7+mkBpO4NyT1EsA4uIdYE3AVNr964lSZLU6prokLCm5RYHkiRJDRZ0vd9APWXmoog4ntKpsgOAi4sDEo4rnj8f+DowMSIeLLp4cmbO7OWuSpIkqY9qRM7tixyglSRJ6qcycxIwqVPZ+WX3nwX27u1+SZIkSf2JA7SSJElNIDw8QZIkSS3InFuZe9BKkiRJkiRJUoM4g1aSJKkJtOJhB5IkSZI5tzIHaCVJkppAm8FVkiRJLcicW5lbHEiSJEmSJElSgziDVpIkqcEiPDxBkiRJrcecWx1n0EqSJEmSJElSgziDVpIkqQm4NZckSZJakTm3MgdoJUmSmoCn20qSJKkVmXMrc4sDSZIkSZIkSWoQZ9BKkiQ1XHp4giRJklqQObcazqCVJEmSJEmSpAZxBq0kSVKDBdDm3lySJElqMebc6jiDVpIkSZIkSZIaxBm0kiRJTcDTbSVJktSKzLmVOUArSZLUaGFwlSRJUgsy51bFLQ4kSZIkSZIkqUGcQStJktQEgmx0FyRJkqSaM+dW5gxaSZIkSZIkSWoQZ9BKkiQ1AffmkiRJUisy51bmAK0kSVKDBRBtJldJkiS1FnNuddziQJIkSZIkSZIaxBm0kiRJjRYQ/mwuSZKkVmPOrYofkSRJkiRJkiQ1iDNoJUmSmoCHJ0iSJKkVmXMrc4BWkiSp4QI8PEGSJEktx5xbDbc4kCRJkiRJkqQGcQatJElSo3l4giRJklqRObcqfkSSJEmSJEmS1CDOoJUkSWoC4ekJkiRJakHm3MocoJUkSWqwwKVfkiRJaj3m3Or4EUmSJEmSJElSgziDVpIkqRm49EuSJEmtyJxbkTNoJUmSJEmSJKlBnEErSZLUaOHeXJIkSWpB5tyqOEArSZLUBKLNpV+SJElqPebcyhzDliRJkiRJkqQGcQatJElSE/DsBEmSJLUic25lzqCVJEmSJEmSpAZxBq0kSVKjeXiCJEmSWpE5tyoO0EqSJDUDD0+QJElSKzLnVuQYtiRJkiRJkiQ1iDNoJUmSGizw8ARJkiS1HnNudZxBK0mSJEmSJEkN4gxaSZKkRosg3JtLkiRJrcacWxUHaCVJkpqAp9tKkiSpFZlzK/MjkiRJkiRJkqQGcQatJElSEwhPT5AkSVILMudW5gxaSZIkSZIkSWoQZ9BKkiQ1WuDP5pIkSWo95tyqOEArSZLUYAG48kuSJEmtxpxbHcewJUmSJEmSJKlBnEErSZLUBKLNqQWSJElqPebcyhygVe9pWwWGjW50L9THfOOyeY3ugvqo0yZ+sdFdUJ91ZaM7IEnqJ77W9uFGd0F91GlLftroLkiqIQdoJUmSGi0g3HhKkiRJrcacWxUHaCVJkpqBpydIkiSpFZlzK3IMW5IkSZIkSVLLiIgTIuKhiPhbRFweEYMjYs2IuDkiHiv+HFlW/8sR8XhE/D0i9ikr3yEiHiyeOzeiNNocEatGxJVF+X0RMWZl+usArSRJUhOIttreJEmSpGZQ65xbKetGxAbA54Bxmbk1MACYAJwC3JqZmwG3Fo+JiC2L57cC9gV+FBEDiubOAz4BbFbc9i3KjwVmZ+amwHeBs1bmMzK+S5IkSZIkSWolA4EhETEQGAo8CxwIXFo8fylwUHH/QOCKzJyfmU8AjwPjI2I0MDwz78nMBC7rdE17W9cAe7TPru1pZyVJktRIAdHm3lySJElqMfXLuaMiYkrZ4wsy8wKAzHwmIs4GngJeB36bmb+NiHUz87miznMRsU5x7QbAvWVtTS/KFhb3O5e3X/N00daiiHgFWAuY2ZM34wCtJElSgwWenSBJkqTWU8ecOzMzx3X5mqW9ZQ8ExgIvA1dHxIe7aaurHmY35d1d0yNucSBJkiRJkiSpVewJPJGZL2bmQuCXwNuBF4ptCyj+nFHUnw68oez6DSltiTC9uN+5vMM1xTYKawAv9bTDDtBKkiQ1XBBttb1JkiRJjVf7nFtF1n0K2Ckihhb7wu4BPAJcDxxV1DkKuK64fz0wISJWjYixlA4Du7/YDuHViNipaOfITte0t/VB4LZin9oecYsDSZIkSZIkSS0hM++LiGuAPwKLgD8BFwCrAVdFxLGUBnEPKeo/FBFXAQ8X9T+TmYuL5j4FTASGAJOLG8BFwE8i4nFKM2cnrEyfHaCVJElqtKDrXawkSZKkvqxBOTczTwNO61Q8n9Js2q7qnwmc2UX5FGDrLsrnUQzw1kJNB2gjYjAwKDPn1LJdSZKkVhduPNX0zLqSJEkrzpxbWa0/ovNYiQ1xJUmSpCZm1pUkSVLN1WOLAxfoSZIkrSAP9uoz/AclSZK0Asy5lTnJWJIkSZIkSZIaxEPCJEmSGi0gnFggSZKkVmPOrYoDtJIkSQ0WuPRLkiRJrcecWx23OJAkSZIkSZKkBnEGrSRJUjPwZ3NJkiS1InNuRd0O0EbE4t7qiCRJktSbzLqSJElqBpVm0PZkk4jsSUckSZL6rQDcm6sRzLqSJEn1ZM6tSrcDtJnpJGRJkqTeYOrqdWZdSZKkXmDiqsiPSJIkSZIkSZIaxAFaSZKkRmtf+lXLWzUvG7FvRPw9Ih6PiFOWU2f3iPhzRDwUEXfU8m1LkiSpxdUj57bglgndDtBGxMURcUCnss07l5U9d1RE3FbLDkqSJKn2ImIA8ENgP2BL4PCI2LJTnRHAj4ADMnMr4JBe72gdmXUlSZLUDCrNoD0a2K5T2eHAr5ZTfwyw28p1SZIkqR9qq/GtsvHA45k5NTMXAFcAB3aqcwTwy8x8CiAzZ/T4/TWnozHrSpIk1Vetc24L7gfQ7SFhkiRJ6gVRl6VaoyJiStnjCzLzgrLHGwBPlz2eDuzYqY3NgVUi4nZgdeD7mXlZrTsqSZKkFlWfnNtyHKCVJElqTTMzc1w3z3eVlLPT44HADsAewBDgnoi4NzP/UaM+SpIkSf2eA7SSJEnNoPdnFkwH3lD2eEPg2S7qzMzMucDciPg9sC3gAK0kSZKq4wzailpw1wZJkiRV4QFgs4gYGxGDgAnA9Z3qXAe8MyIGRsRQSlsgPNLL/ZQkSZJamjNoJUmSmkEv/2yemYsi4njgJmAAcHFmPhQRxxXPn5+Zj0TEjcBfgSXAhZn5t97tqSRJkvo0p4dWVM0A7XYRcWT5Y4CI+AjL7l3W+RRcSZIkVRI0ZOlXZk4CJnUqO7/T428D3+7NfvUys64kSVK9NCjn9jXVDNAeWNzKBTCxi7rBsodLSJIkSc3KrCtJkqSGqjRAe2mv9EKSJKm/c+lXI5h1JUmS6s2cW1G3A7SZeUxvdUSSJEnqTWZdSZIkNQMPCZMkSWo09+aSJElSKzLnVsUBWkmSpGZgbpUkSVIrMudWVHEXiCjZNSI+EBGblJVvFxGTI2JWRMyOiF9GxGb17a4kSZJUO2ZdSZIkNVq3M2gjYggwGXhnUbQkIj4L3A7cAaxeVv0g4O0RsV1mPl+HvkqSJLUul371OrOuJElSLzDnVlRpBu0XgF2B6cCvgGeBbwFfARYAHwfeAuxWPL8O8KV6dVaSJEmqIbOuJEmSGq7SHrQfBJ4DtsnMORGxBvAQcDjw4cy8vL1iRNxdPLcv8MU69VeSJKn1RDizoDHMupIkSfVkzq1KpRm0mwE3ZOYcgMx8Bfh18dwt5RUzcwlwGzCmxn2UJElqfW01vqkaZl1JkqR6q3XObcGsW+ktrQZ03mPrBYDMfLGL+jOAwTXolyRJklRvZl1JkiQ1XKUtDgCWVHgsSZKkleXSr0Yx60qSJNWTObeiFpwULEmSJEmSJEl9QzUzaA+KiDFlj7cDiIiLu6j71hr0SZIkqX8JCH82bxSzriRJUr2Yc6tSzQDtdsWts6OXUz973BtJkqT+yqVfjWLWlSRJqidzbkWVBmiP6ZVeSJIkSb3PrCtJkqSG63aANjMv7a2OSJIk9VuBJwM0gFlXkiSpzsy5VanpRxQRbRFxYC3blCRJkpqBWVeSJEn1UM0etBVFxMbAxygtExsNDKhFu5IkSf2Ge3M1LbOuJEnSSjDnVtTjAdqIGAAcCHwC2JPSbNwEbqlN1yRJkvoRg2tTMetKkiTViDm3ohUeoI2ITSjNIDgaWLcongn8L3BRZj5Zs95JkiRJvcisK0mSpN5W1QBtRAwE3k9pBsG7KM0gWAD8EvgAcF1mnlqvTkqSJLU0D09oKLOuJElSnZhzq9LtAG1EbAZ8HDgKGEXpY/0jMBH4eWa+FBFL6t1JSZIkqdbMupIkSWoGlWbQ/p3SXlszgO8Cl2TmQ3XvlSRJUr8S7s3VGGZdSZKkujLnVqOaLQ4SmARcY2CVJEmqA5d+NZJZV5IkqV7MuVWp9BF9FXgSOAa4OyIejoiTImJ0/bsmSZIk1ZVZV5IkSQ3X7QBtZp6ZmW8E9gN+BbwR+CbwVET8JiIO7YU+SpIktb62qO1NFZl1JUmSekGtc24LZt2qJhln5k2Z+UHgDcB/UpppsB9wOaVlYdtFxA5166UkSZJUJ2ZdSZIkNdIK7QKRmTMy85uZuSmwF3ANsBAYB9wfEX+KiM/UoZ9Sn/Dww1PZY49PMXToO1h//X059dTzWbx4ccXrXnnlXxxzzNcYOfJdrLHGbnzoQ19h1qyXl6l33XW3s802hzF48NvZcstDuPLK39bjbagBZjCfS3mab/AYZ/NPbmMmS8iqrruM6XyDxziLx/k1LzCfjgeOJ8nvmcV3mMrXeYzzeZLHmVuvt6Je9PDDU9ljz08zdNg7WX+D93Dqqf9b/XfOR89g5Jp7sMaId/GhD391me+cm2++j8OP+Apjxh5ItI3n9NMvqNfbULu2Gt+0wsy6Um09/vjTfPKTZ7LttoczYMB4dt/9E43ukmpgFgu4gRc4j2l8jX9wCU93eP5VFvFbXuQ8pnEmj/EdpvIrnmcOizrUe4hX+TnPcA7/5Ewe4395kgeZU7e2OpvMDE7nH9zEiyvxaag3XHvt7bxl2yNYdfA7GLvJgXznOz/r8PyCBQs59LAvs8kbD2LI0Hey9jp7s997Ps8f/vDICrelOql1zm3BrNvjt5SZt2bmYcCGwEnAP4BtgXNr1DepT5k9ew577vlpIoLrrjuHU0/9GOec81NOO+1/K1572GFf5vbb/8CFF36FiRNP54EHHuagg/6jQ5277vozH/jAybzrXeOYPPlc3vveXTj88P/it7+9t15vSb3kdRZzGdOB4HDWZzfW4h5m8ztmdXvdPBZzKdNZRHIIo9mbtXmYf/FLnutQ7y5mcwcvMZ4RHM76rMMgfs4zPMO8Or4r1dvs2XPYc6/jS985157NqV89lnO+8zNOO63yQOphE/6z9J3z4/9i4iWnlr5z3v+lDnVuvPEe/vrXx9jj3eMYOnRwvd6G2gUu+2oyZl1p5T300D+ZNOn/2Hzzjdh8840a3R3VyIss4DHmshaDWItByzz/LPN4hH+xNcM5gg3Yi1FM53Uu4qkOEwnuYTaDaGMf1uFwNmAMQ/kFz3Mfs+vSVrkZzOdPzGHVVhzlaTF33/0XDv7AyYx/25bccP05fPSYAzj5lB/wve9dvrTO4sWLiQi+fMrR/ObX3+HHF/wnr702j3fv8WmmTn1mhdpSHdQj57Zg1h24sg1k5kzgbODsiNgN+NhK90rqg84//xe8/vp8fvnLbzF8+GrstRfMmTOX00+/gJNOOpLhw1fr8rp77vkrN910D3fccQG77ro9ABtssDY77ng0t9xyH3vuuSMAX//6hey661s599zSIMq73jWOhx76J2ec8WP23nun3nmTqospvMIiksMYzWAG8EZgPku4nVm8g5EMZkCX1z3AKywkOZz1GVLUGcIAruBZnmEeGzCYRSR38hLvYCS7sCYAmzKMF1nA7cziQ2zQW29TNXb++b8sfef84qziO2fH0nfO137MSSd9pMJ3zr3ccfv5Zd8567DjTsdwyy33s+ee4wH49rc/xznnfAGA667/fe+8KakJmXWlntt//1058MDdAfjgB09i5sxlV4ip79mcYWzBJgBcybO8RsfVOxsxhOMZwwD+PYAymsH8gGk8wqtsxxoAHM4GDCvLuZswlFdZxD3MZkdG1rytcpN5kZ0YwV8qzLJV453x9QvZZZdtufDCrwCw9947MXv2HM74+kV8+tMfZNCgVRgyZDBXXvHfHa7bc8/xrDVqL6699nZOPPFDVbclNUqtfy46FHhvjduU+oTJk/+PffbZqcOgyIQJ+/D66/O5444/dnvduuuutXSgBGD8+K0ZO3YDJk/+PwDmz1/Ys3FwAAAgAElEQVTA7343hUMP3bPDtRMm7M099zzIK6/8q8bvRr3pMebyRoZ1GIjdmtVZRPIkry/3uueZx/qsunRwFuCNDF3aJsBsFrCAJWxSlLfbhKFM5TUWVbGNgprT5Bu7+s7Zu/jO+dPyr5t8D+uuu2an75ytGDt2/aXfOQBtbc4o6XUu++oLzLrSCvD/Ja2pje5nrg1hQIcBVYBRDGIVgrllg7nDupiEMJpVO9SpZVvtHuJVZjJ/6eQFNbc///kx9txjfIey9oHVe+55cLnXDRs2hMGDB7Fgwb+3w+hpW6oBtzioqNZvaSgUP2FJ/cyjj05jiy3GdCjbaKP1GDp0MI8+Oq3CdRsvU/7mN49Zet0//zmdhQsXLdP+m988liVLlvCPfzy5kr1XI81kAaPo+GvtCFZhFYKZLFjudYvIZQJrG0FQWnrWXgdYpt5AgsUks1lYg3egRnj00SfZ4k0dvzuq+s75+7LfVVD6Pnn0736XSBWYdSWpB55nPgtJ1u5iS4RyT/N6xTor09ZClvBbXmRP1mZQK47wtKB58+YvM7N11VVLjx955IkO5ZnJokWLeP75mZx00v8wYMAADj987x61JfU2v5GkGpk9ew4jRqy+TPnIkcOZPfvVlbpu9uzS0pvO9UaOHF48v/z21fzmsbjLbQwGM4DXOx34VW5NBvEC81lcNgv2OeaRlPa1BRhZDPw+22m/2fb9Z1/vYlaB+oaef+e8yogRy25/MHLk6ku/a9QA7kErSWpRS0huZAZrsgpvZNhy603lNR5lLm9jRN3aupOXWI2BvIVlM5Sa06abvoEHpjzcoez++x8C4KWXOmbXs866jFUGvZ3R67+HSy/7DZN+81023nh0j9pSDbkHbVWaZoA2Ig6OiNsi4uWImB8R/4iIb0TEqBq/zuYRcXpEjOhUfnREZER0vWlfH1bte4uI2yPimt7qVyuKWPZLIjPporhH13Wul5lF+Yr1U82n63+E2e0Csu1Zg7ksZhIzeJVFzGA+v2FG6f9/RZ3BDGAbVuf3vMQTvMZrLOY+ZjOV14Am+p+AeqT23zl+mTSUobWlmXXrx6wrNbdbmcnTzONg1ltmVVe72SzkFzzHFgzjrd0sVFiZtmazkP9jNvuydrHmTH3BcZ88mOuu+z0//vG1zJ49h5tuuodzvvNzAAYM6Pi3maOPfi8P3D+R6687hx122IL37X8iDz88tUdtqcYcoK2oKf4NjIhzgKuBqcBHgL2B7wL7Az+u8cttDpwGy/ws9xtgZyhGLaQVNHLkcF5+edlZa6+88q8uZ7lVuu7ll19del37TNnO9dofd9e+mt9gBjCvi5my81jC4G6+ptdmEPuzLn/jVc5hKufxJBswmPVYlWFlZ0Duy9qszSAuZTrf4p/czWx2ZS2ADvXUt/T8O2d1Xn552X2rX365++sk9ZxZV1J/dT8vczezeT/rsSFDuqzzGov5Gc+wBgM5mNFd1qlFW7fwIpsxjFEM4nUW83qxDm0xWdz3bIZm9NGP7s9xnzyYT336LNZca08O/sDJnPrVYwFYd921OtRdb71RjBu3Jfvv/05uuP4c1lprDb551qU9akvqbQ3/m3lE7A+cCBybmReXPXVHRFxAKcDWXWa+CLzYG6+l1rTFFmOW2ffx6aefZ+7c17vc77H8ujvvXPZAn0cfncZBB+0OwBvfuCGrrDKQRx+dxm677dChTltbG5tvvuwetuo7RjFomb1mX2EhC0lGVdhba3vWYBtW5yUWMowBDGUAZ/FPti+bLTCMgRzNG3iFhcxnCWsxiHuZzWoMWLoFgvqeLbbYeJk9Y59++oXK3zlvGsOdd167TPmjj07joAN3q3U3tSKa4mdz1ZpZV1J/9TCvMpkZ7MUotl7OlgILWMLPeYbFJEew4XL3ha1FWzNZyAvM5xE6/lB9Py9zPy9zAmNZw2zcdAYMGMAPfvAlvv71TzJ9+gzGjl1/6d+7d9pp6+VeN3DgQLbZZlOmTn12pdtSDZhzK2qGj+gE4I+dAisAmbk4MycDRMSoiLg0ImZFxGvFEqVx5fUjYlpEnB0RJ0TE9IiYHRFXtC/xiojdgRuK6k8US6GmFc91WBoVEWOKx4dGxP9GxCtFm1+LiLay15wYEVM69aP92vd19Xh51xbL0WZGxI4RMSUiXo+IuyJibESsExHXRsS/IuKRiHh3Dz7rN0fEnUW7/4iI93dXOSLWiIi7I+IvEbF2UTay+EznRsSzEXFy8ZlP60F/Wsp++72dm266l1dfnbu07Morb2bIkFXZbbftu73u+edncdddf15aNmXKw0yd+gz77fd2AFZddRDvetc4rr761g7XXnnlzey88zassUbLrVbsVzZjGI8zl/lls2j/xqsMJNh4ObMDyq1CG+uyKqsxkL8yhwS26iK4rsEqrMOqLCH5E3O6XT6m5rffvt1957x1+dftt3PF7xxJNWXW/fdjs67UTzzBa/yC5xnPCN7Bml3WWUxyNc/xEgv5EBuw2nLmj9WqrQNYl6PYsMNtGAPYitWW3lfzGjlyONtssymrrTaUH/3oF7z97W/pdlLCvHnz+eMfH2Xs2PVXui2pN3Q7gzYipnb3fBdWaA+tiFgFeDtwThXVrwU2Bf4DmAl8CfhdRLw1Mx8vq3co8FfgE8CGwHeA/wY+DfyxuP5s4GDgOWB+hdf9FvAL4IPAHsCpwEPAVVX0uSeGAhcUrzsXOBf4SdHPycCPgJOAqyPiDZm5IsvUriyu/2/gY0UbO2TmXzpXjIg1gZuKh+/KzJeK+xOBXYDPA89T+kvH5uBJQ8cd9wHOPfcKDj74S5x88lFMnfoMp59+ASee+CGGD//3AOqmmx7Ebrttz0UXnQrAzju/hX322ZkjjzyNs8/+PG1tbZx88rnssst27Lnnjkuv++pXP8buu3+SL3zhHA46aDcmTbqbSZPu5sYb/6fX36tqaxxrcB+zuZJneQcjmc1CbmcWOzOyw+Fh3+cJxjCEA1kPKB0udicvsTFDaCN4gte4h9nsz7oMLbvuL8xhMclIVuEVFnEvs2kDdllOwFXfcNxxB3Pu/1zJwR84mZNPOrL0nfO1H3PiCUd0/M7Z7GB22/WtXHTRV4H275ydOPKo0zn725+nrS04+ZQfsMsu27LnnuOXXvfkk8/xwAOlQxQWLFjEw488wTXX3MqwYUMcyK2HaM29tJqdWRcw64JZt25ee20ekybdBcAzz7zInDlzueaaWwB4z3t2YejQwY3snnpoAUt4jNIPxK+yiPks4SFK2y5txjBeYSFX8CyjGMRWrM7TvL702mEMYM1ihdhvmMFjzGVf1uZ1FneoN5pVGUgbLzK/Zm1twLL/vg0kGM4qjGVoDT8h1dK99z7IXXf9he2225w5c+Zy+RU3cdNN93HXnRcsrXP55Tcx+cb/Y999dmb99dfmuedm8qPzruG552Zx4glHrFBbqgNzblUqbXEwpgdtrsjGLWsBqwJPdVcpIvYF3gHsnpl3FGW3AdMohddPllVfCByUmYuKelsCE4BPZ+aciPh7Ue9PmTmtij7+PjO/WNy/uejLwdQvtA4BPlf2PtcHfgiclplnF2XTKQXn3SgF2WpdWNbGTcDDwJcpfT5LFTMIbgH+BeyXmXOK8q2BA4BDM/PqouxW4Omi7jIi4hOU/gLBRhuttwJd7XtGjhzOrbeex/HHf4v99z+RESNW44QTjuD00z/Rod6iRYtZvLjjfqNXXPHfnHDCd/joR89gyZLkfe/bhXPP/VKHOrvssh3XXHMWX/nKjzjvvGsYO3Z9fv7zb7D33jvV/b2pvoYwgCPZkEnM4HKeZTBt7MxIdqfjPkhLyA471bYRPMd8/sArLCJZh0Ecwvq8mY4zqpPkbl7iZRYxmDa2YDX2YBSrNsUiCvXUyJHDufWWH3L8Z7/N/gd8sfSd84XDOf30j3eot2jRYhYv6fSdc/mZnHDid/nosV8vvnPewbnf/48OdX73uz9wzEfPWPr46qtv5eqrb2XjjUcz7Ynr6vfG+jP/k2yEMT24xqy7csy6/ciMGS9xyCGndChrf/zEE9czZsyyM9vU/OaymKt5rkNZ++PPM5bpzGM+S3iB+VzM0x3qbctw3l9MNvhnMch7Yxe7r3yesYykraZtqW9aZZWBXHnVzZz+tR/T1tbGO9+5HXff9WO22WbTpXXe9KaN+enPbuTEL36P2bNfZfToUey441ZMeeAUttrqjSvUlurE/wQrqjRAO7ZXelE56I4HXmwPcgCZOTcifk3pF+5yv2sPrIWHgXUiYlBmLmDF/bbT44eBjXrQTrUWAHeWPW6fMXFbF2UbrGDbv2q/k5lLIuI64JBOddYF7qA0Y2D/zJxb9lz7Mrv2pXNk5usRcQvQ5ShhZl5AaZYE48Zt2fK7rm+55Sbcdtv53daZNu2GZcpGjFidSy45jUsuOa3baw86aPel+9KqtazDqhzNG7qtcwKbdHg8iDaOZMOKbW/HGmzndgYtacstN+G2W8/rtk5Xg6kjRqzOJRefyiUXn7rc644++n0cffT7lvu81CLMumZds24djRmzPplTKldUnzKSVTidzbt5fo2qttLqnG278tYatlXL69R7dtjhzTxw/6Xd1tl++y34za+/W5O21FqKbaAuBLamlMc+Cvyd0qqbMZR+DD80M2cX9b8MHEtp5cznMvOmonwHSqtshgCTgM9nZkbEqsBlwA7ALOCwKn8gX0a3A7SZ+WR3z9fALErLmSqFwNHAC12UvwDLrNF9udPjBUAAg4r7K6qr9uq5FufVzCyf6tTe56X9yMwFEUEP+jGji8edj8ncktJn+u1OgRVgvaJ/8zqVe+CEJEkrK1z61dvMusttz6zbkVlXkqSV0bic+33gxsz8YEQMorTV0n8Ct2bmNyPiFOAU4OSyVUlbAesDt0TE5pm5GDiP0oqZeykN0O5LaZXPscDszNw0IiYAZwGH9aSjDZ1knJkLgbuBfSpUfQ5Yp4vydYGXuijvTfNgmWPWOwfp9pBXqV69df4M14FOa1Pgd8BXgAuidOpwueeB1SOic1heu3ZdlCRJag1mXbOuJElqjIgYDuwKXASlH4Az82XgQKB9KvWlwEHF/QOBKzJzfmY+QWlFz/iIGA0Mz8x7MjMpzZgtv6a9rWuAPSJ6NhrdDLtAfA8YFxFHdX4iItqKfbDuo7R0a9ey54YC7wXuWsHXa/+VvlYzA6YDYzoFub061ZlBab+wN7cXFCfo7lyjPlRr6Um2xem8BwL3d66UmWdSOszi6k4n6LavTzqgrJ0hLPt+JUnSiog63NQszLq9x6wrSVKzqUfOLWXdURExpezW8QAg2ITSKphLIuJPEXFhRAwD1s3M5wCKP9t/4N0AOmx2Pb0o26C437m8wzXFFlSvQKfDZKpUaQ/ausvMGyLiO8BFEfEO4DpKm/BvARwHTMvM90fE3cCVxfTjWZROqB0CfHsFX7L94IRPRsQVwGuZ+eBKvIVrgTOACyNiIvBW4JjyCmV7YJ0QEU9SWsL1RSg7WrJ3fCwiFgB/Az5O6aTgw7uqmJmnRMTqwHURsVdm3puZf4uIG4DziueeB04EXgOWdNWOJEmqklsctCSzbq8y60qS1Izqk3NnZua4bp4fCGwPfDYz74uI71PazmB5uupkdlPe3TUrrBlm0FKcHHsYsBnwc+BmSqHuVuBTRbX3F+XfA66m9CG8OzMfX6bB7l/rSUqB92BKS86WPbFpxdr7G6VNhncGrqd02uxHu6h6fPF6P6J0Uu3ldDwMoTdMoPQ5XgtsS2nz4j91U/944BfA5IjYtig7mtKpt+cCF1M6ZOFGYE6d+ixJktSnmXV7jVlXkiS1mw5Mz8z7isfXUBqwfaHYtoDizxll9ctP7t4QeLYo37CL8g7XRMRAYA16uD1VlLZPkHqm+Bfwb8B9mbnM0r1y48ZtmVOm/KR3OqaW8bU4otFdUB912pKfNroL6qOibfwfKvwaX3PjNlo97//S9jVtc8Dnft/r70NqNWZd1ZtZVz1l1lVP9XbWrUfOheqybkTcCXwsM/8eEacDw4qnZpUdErZmZp4UEVtR+iF9PKVDwm4FNsvMxRHxAPBZSttSTQL+JzMnRcRngG0y87jikLCDM/PQnryfhm9xoL4lIg6h9C/qg8BwSsvHNgOObGS/JEmSpJVl1pUkqaV8FvhZRAwCplLapqkNuCoijgWeAg4ByMyHIuIq4GFgEfCZzFxctPMpYCKl7acmFzcoHUD2k4h4nNLM2Qk97agDtH1ccTrcgG6qLM7aTpOeS+lf6E2L130Q2D8zlzmAQZIkrQD3oJWWYdaVJKkFNCjnZuafga5m2e6xnPpnAmd2UT4F2LqL8nkUA7wrq9s9aCNio4gYXm1jRf1dK9dUDR1F6dTc5d26XYq1ojJzUmZul5mrZeaQzByfmZNq+RqSJPU7QSmV1fKmisy6fYJZV5KkvqweObcFs26lGbRPAKcDX28viIhPAp/MzK42kDgGOJXuf+VWbd0AvK2b55/orY5IkiT1MWbd5mfWlSRJLa/SAG0Ut3LrUToVVU0gM2cBsxrdD0mStJLc4qARzLpNzqwrSVILMOdW1IKTgiVJkiRJkiSpb/CQMEmSpGbgxAJJkiS1InNuRQ7QSpIkNVy49EuSJEktyJxbDbc4kCRJkiRJkqQGqWYGbda9F5IkSf2dEwsaxawrSZJUT+bciqoZoD0hIo4pezwCICKmdlF3RE16JUmSJPUOs64kSZIaqpoB2hF0HUbHLKe+sxAkSZJWRODeXI1j1pUkSaoXc25VKg3Qju2VXkiSJPV3ngzQCGZdSZKkejPnVtTtAG1mPtlbHZEkSZJ6k1lXkiRJzaCaLQ4kSZJUby79kiRJUisy51bU7QBtRGxU4folwMuZ+a/adUmSJEmqP7OuJEmSmkGlGbTTqOIghIiYBlwCnJWZC1e+W5IkSf2MEwsaYRpmXUmSpPoy51ZUaYD2KboPrW3AmpQOWPga8K6I2DszF9eof5IkSa3P020bxawrSZJUT+bcqlQ6JGxMNY1ExLbAt4A9gY8D5690zyRJkqQ6MutKkiSpGbTVopHM/AuwP/A0cEQt2pQkSepPImp7U+2YdSVJknqu1jm3FbNuTQZoATJzATAJ2KpWbUqSJEnNwKwrSZKkeqm0B+2KmgWsVuM2JUmSWl8rTgVoPWZdSZKkFWXOrajWA7TrAa/UuE1JkqTWZ27tC8y6kiRJK8qcW1HNtjiIiDWAA4A/16pNSZIkqRmYdSVJklQvKz2DNiLWAnYGzgBGARNXtk1JkqR+JQLanFrQjMy6kiRJK8GcW5VuB2gjYvEKtBXA5Zn585XrkiRJklR/Zl1JkiQ1g0ozaCsNcSelfbj+AkzMzEtr0itJkqT+xokFjWDWlSRJqjdzbkXdDtBmZs32qJUkSVI3PN2215l1JUmSeoE5tyJDqSRJkiRJkiQ1SE0HaCNiYER8ppZtSpIk9QtR45tqzqwrSZLUA7XOuS2YdWsyQBslRwH/AM6tRZuSJEmqr4jYNyL+HhGPR8Qp3dR7W0QsjogP9mb/moVZV5IkSfVU6ZAwImIkcDzwNmAhcCdwfmbOK55/H/At4E3FJb+qT1clSZJaVNDre3NFxADgh8BewHTggYi4PjMf7qLeWcBNvdrBXmLWlSRJqqMG5Ny+qNsB2ogYBdwPbMy/JxAfBBwQEXsB5wMfLZ77NXBqZv65ft2VJElqUb2fW8cDj2fmVICIuAI4EHi4U73PAr+gNIDZUsy6kiRJvcDx2YoqzaA9BRgD/AX4GaWP9CPAbsBvgL2B+4DPZ+b99eumJEmSamwD4Omyx9OBHcsrRMQGwPuBd9OCA7SYdSVJktQEKg3Q7gc8CeyYmQsAIuIHwKOUlsNdAXwoM7OuvZQkSWp1bTWfWjAqIqaUPb4gMy8oe9zVC3bOdN8DTs7MxdGaS9PMupIkSfVW+5zbcioN0I4BJrYHVoDMfD0ifg0cB3zVwCpJktSUZmbmuG6enw68oezxhsCzneqMA64oBmdHAe+JiEWZeW1Ne9o4YzDrSpIkqcEqDdAOAV7oonxG8efU2nZHkiSpn+r9iQUPAJtFxFjgGWACcER5hcwc234/IiYCv26hwVkw60qSJNWfE2grqjRA2y1nFEiSJNVAA063zcxFEXE8cBMwALg4Mx+KiOOK58/v1Q41IbOuJEnSSmpAzu2Lqhmg3S4ijuxcBhARH6GLcfDMvKwGfZMkSVIdZeYkYFKnsi4HZjPz6N7oUwOYdSVJktRQ1QzQHljcOgtg4nKuMbRKkiStCCcWNIpZV5IkqZ7MuRVVGqC9jGVP85UkSZJagVlXkiRJDdftAG0LL2WTJElqIuHeXA1g1pUkSao3c2412rp7MiKOjIi39FZnJEmS+q2o8U0VmXUlSZJ6Qa1zbgtm3W4HaCntu3VQL/RDkiRJ6m0TMetKkiSpwao5JEySJEn1FEBbC04FkCRJUv9mzq1KpRm0kiRJkiRJkqQ6cQatJElSM/DwBEmSJLUic25F1QzQjoiIjVak0cx8qof9kSRJ6p8Mro1i1pUkSaonc25F1QzQfr64VSurbFeSJElqNLOuJEmSGqqacDkHeLneHZEkSerXnFnQKGZdSZKkejLnVlTNAO13M/OMuvdEkiRJ6n1mXUmSJDWUy7MkSZIaLiDaGt0JSZIkqcbMudXwE5IkSZIkSZKkBnEGrSRJUqMF0ObeXJIkSWox5tyqOEArSZLUDDw8QZIkSa3InFtRtwO0mekWCJIkSWpJZl1JkiQ1A2fQSpIkNZyHJ0iSJKkVmXOr4SckSZIkSZIkSQ3iDFpJkqRm4N5ckiRJakXm3IocoJUkSWo0T7eVJElSKzLnVsUtDiRJkiRJkiSpQZxBK0mS1HAeniBJkqRWZM6thp+QJEmSJEmSJDWIM2glNbXTlvy00V1QH/W1tg83ugvSivHwBEnqd8y66qlJZl31JQ3KuRExAJgCPJOZ74uINYErgTHANODQzJxd1P0ycCywGPhcZt5UlO8ATASGAJOAz2dmRsSqwGXADsAs4LDMnNbTvjqDVpIkqRlE1PYmSZIkNYNa59zqs+7ngUfKHp8C3JqZmwG3Fo+JiC2BCcBWwL7Aj4rBXYDzgE8AmxW3fYvyY4HZmbkp8F3grJ5+POAArSRJkiRJkqQWEhEbAu8FLiwrPhC4tLh/KXBQWfkVmTk/M58AHgfGR8RoYHhm3pOZSWnG7EFdtHUNsEdEz2dJuMWBJElSowUeniBJkqTWU7+cOyoippQ9viAzLyh7/D3gJGD1srJ1M/M5gMx8LiLWKco3AO4tqze9KFtY3O9c3n7N00VbiyLiFWAtYGZP3owDtJIkSZIkSZL6kpmZOa6rJyLifcCMzPxDROxeRVtdzXzNbsq7u6ZHHKCVJElquIA2942VJElSq2lIzn0HcEBEvAcYDAyPiJ8CL0TE6GL27GhgRlF/OvCGsus3BJ4tyjfsorz8mukRMRBYA3ippx12LZ0kSVIz8JAwSZIktaJePiQsM7+cmRtm5hhKh3/dlpkfBq4HjiqqHQVcV9y/HpgQEatGxFhKh4HdX2yH8GpE7FTsL3tkp2va2/pg8RrOoJUkSZIkSZKk5fgmcFVEHAs8BRwCkJkPRcRVwMPAIuAzmbm4uOZTwERgCDC5uAFcBPwkIh6nNHN2wsp0zAFaSZKkZuAhYZIkSWpFDcy5mXk7cHtxfxawx3LqnQmc2UX5FGDrLsrnUQzw1oJ/E5AkSZIkSZKkBnEGrSRJUqMF7hsrSZKk1mPOrYoDtJIkSQ3XkNNtJUmSpDoz51bDLQ4kSZIkSZIkqUGcQStJktQMPCRMkiRJrcicW5GfkCRJkiRJkiQ1iDNoJUmSmoGHJ0iSJKkVmXMrcoBWkiSp0TzdVpIkSa3InFsVtziQJEmSJEmSpAZxBq0kSVLDhTMLJEmS1ILMudVwBq0kSZIkSZIkNYgzaCVJkppBm7+bS5IkqQWZcytygFaSJKkZuPRLkiRJrcicW5FD2JIkSZIkSZLUIM6glSRJarTAmQWSJElqPebcqjiDVpIkSZIkSZIaxBm0kiRJDRcQ/m4uSZKkVmPOrYYDtJIkSc2gzaVfkiRJakHm3IocwpYkSZIkSZKkBnEGrSRJUjPw8ARJkiS1InNuRc6glSRJkiRJkqQGcQatJElSo4WHJ0iSJKkFmXOr4gCtJElSM3DplyRJklqRObcih7AlSZIkSZIkqUGcQStJktQMnFkgSZKkVmTOrcgZtJIkSZIkSZLUIM6glSRJagZt/m4uSZKkFmTOrcgBWkmSpIaL4iZJkiS1EnNuNRzCliRJkiRJkqQGcQatJElSowUeniBJkqTWY86tijNoJUmSJEmSJKlBnEErSZLUDMLfzSVJktSCzLkVOUArSZLUFFz6JUmSpFZkzq3EIWxJkiRJ0v9n777D7CjL/4+/700PISG0AAm9SEcBKT+6dKQjvShFQcSCQEAUpNooUkT4gtilCKIg0pEqRYJU6RAwdBJSCal7//6YOZuzm7JL2Hr2/bquvTY7Z2bOzJ7J2c955n6eR5IkdRAraCVJkjpcOHmCJEmSapA5tyWsoJUkSZIkSZKkDmIFrSRJUmfg5AmSJEmqRebcZtlAK0mS1CnY9UuSJEm1yJzbHJuwJUmSJEmSJKmDWEErSZLUGTh5giRJkmqRObdZVtBKkiRJkiRJUgexglaSJKnDBd43lyRJUu0x57aEDbSSJEkdLbDrlyRJkmqPObdFbMKWJEmSJEmSpA5iA60kSVJnENG6Xy16ytghIl6MiFci4qQ5PH5gRDxdfj0UEeu0+nlLkiSptrV2zq3BilwbaCVJkrqhiOgBXALsCKwO7B8RqzdZbSSwRWauDZwJXN6+RylJkiTVPseglSRJ6hTavRJgA+CVzHwNICKuAXYDnquskJkPVa3/CDCsXY9QkhgtdHoAACAASURBVCRJNaD2Kl5bmw20kiRJHS4g2r1j01BgVNXPbwIbzmP9w4Fb2/SIJEmSVGM6JOd2OTbQSpIk1aZFI2JE1c+XZ2b1EAVzKmXIOe0oIraiaKDdtBWPT5IkSRI20EqSJHUOrT/ZwejMXH8ej78JLF318zDg7dkPK9YGfgXsmJljWvcQJUmSVPNqcFKv1maNsSRJUvf0GLByRCwfEb2B/YCbqleIiGWAG4CDM/OlDjhGSZIkqebZQCu1oueee42tt/46/ftvwlJL7cCpp17GzJkzm91u/PhJHHro6QwevBWDBm3BgQf+gDFjxs223o033staa+1L377/j9VX35trr72jLU5DHeC5515j622Opv8Cm7HU0J049dT/a/m1c9gZDF54awYttBUHHnTKbNfOnXc+yv4H/IDllt+NqNuA005zEvZa8T5T+R2jOIuXOZdX+SejqZ9zD/XZtvs9b3IWL/NTXuFm3mMq9Y3WSZL7GcP5vMaZvMxlvMErfNRWpyKgGHGgNb/mLTNnAMcAtwPPA3/OzP9GxFERcVS52qnAIsAvI+LJJkMmSNI8vfLKKI488mzWWWd/evTYgC23/FpHH5I62IwZM/jJT37HyqvsRZ++mzBs6Z059tjzG60zbtxEDjvsTBZeZBsGLLgFO+70bV55ZVSjdbbc6iiiboM5fj388NPteUr6BN5mGhfzHsfwOrvwEicxarZ1kuRaxvBlXmMPXmY4o3iVKbOtdx8T+BZvsBcvcwivch7vMIYZs633OlM5jbfYm1f4Ei9zLG/wcpP9/Y+pnMwo9uRlDuZV/sBoZjbJ1C09Ls1Na+fc2qvIdYgDqZWMHTuBbbY5mtVXX4EbbzyPV199k+OOu4D6+nrOOuvoeW67777f48UX3+BXv/oBdXV1nHjixey++/E88MCvGtZ58MEn2WuvEzn66C9x0UUncMst/2L//b/P4MED2W67jdr69NSGxo6dwDbbHsPqqy/PjX87t7h2jr+wvHa+Ps9t993v5OLaueL71NUFJ570C3bf4wQeuP+KhnVuu+1hnn76Zbb+wvpcc+2dbX06aicfM5Pf8yaL0Yf9WYoPmc4dfEACW7PoXLebwkx+x5ssQm/2ZkkmM5M7Gc1EZrA/QxvWe5Cx3MeHbMUiLEEfnmYCV/EWh7MMQ+nbDmfY3XTM5AmZeQtwS5Nll1X9+wjgiPY+Lkm14b//fZVbbnmIjTZak2nTpnf04agTOPSwM7n77sf44alHsOqqyzFq1Hs899zIRuvsu9/JPPvsa1x4wXcZNGgAZ539a7be5hs88/RVDBw4AIBfXjKcCRMa3zg+9YeX88QTL/L5z6/ebuejT+YNpjGCj/gMfZnbO8J1jOUaPuQwFmUYvfkbY/kBb3IJy7Fw2YT1CJP4Ge+yM4M4jMX4kBn8gdGczltcwDLUlY13rzKFExnFhgzgRJYE4GWmMK2q8XUiM/k+b7I0fTiFpXiH6fyqzNSHVGXqlhyX5sZJwlrCq0hqJZdd9hc+/ngqN9zwMwYOHMC228KECR9x2mmXM3z4IQ1hoqmHH36a229/mPvuu5zNN18XgKFDF2PDDb/CXXc9yjbbFBNqn3nmr9h8889x0UUnALDVVuvz3/++yhlnXGEDbRd32WU3FNfOX35aXjsbFtfO6VcwfPjBzVw7j3DfvZdVXTuLs+FGh3LXXf9mm202AOCcc77Feed9B4Abb7q/fU5KbW4E45lBsi9L0pcerAhMpZ57GcMmDKYvPea43WOMZzrJ/ixFv3KdfvTgGt7mLaYwlL7MIHmAD9mEwWzKwgCsxAJ8wDTuZQwHVjXkSpI0N7vssjm77bYlAF/60nBGj569h5i6j9tue5hrrrmDp578E6uvvsIc13n44ae5445HufuuS/jCFz4PwIYbrsHyK+zO5Zf/jeOPPwhgtu2nTZvOiBHPs+8+29Czp80cndWGLMDGFK/dj3ibCTTuMTiNeq7jQ/ZmYXZhMACr0Y9DGcnNjGtoML2PiaxIH77OkIZt+1PHmbzNm0xjGfoAcAnvswEDOKFsnAVYnwUaPeetjGcqyQ9Ykv704HPAZOq5ijF8icH0p0eLj0v6NGzCllrJrbc+xPbbb9SoMW2//bbn44+nct99/5nndkOGLNLQwAawwQZrsvzyQ7n11ocAmDp1GvfcM4J99tmm0bb77bcdDz/8DOPHT2rls1F7uvW2OV0725XXzhNz3+7WhxkyZOEm184aLL/8Ug3XDkBdnW/1tehlPmJFFmjUELsmCzKD5A0+nut27zKFpejT0DgLsCL9G/YJMJZpTKOeFcrlFSvQn9eYzIwWDKOgTyggIlr1S5I6mhlE1X79m5v4whfWn2vjLMCTT75Ez5492GKLWfl2yJBFWHvtlfnHLf+a63a33fYwY8dOYP/9t2vVY1brqmumW/rzTGEy9WzGgg3L+lLHhizAiKqhtmaQLNCkOWtAk5//x1ReZAq7sNA8n3MEH7EeC9C/KhtvzoJMJXmmzNQtPS7NRRvk3FrMuv7FlFrJCy+8zqqrLtdo2TLLLEH//n154YXXm9lu2dmWr7bacg3bvfrqm0yfPmO2/a+22vLU19fz0ktvfMqjV0d64YU3WPUzja+BFl07L85+zUFxXbzwotdErRvNNBalV6NlC9GLXgSjmTbX7WaQ9GgSjusIAvig3K7SANt0vZ4EM0nGzrVTmiRJ0pw9+uh/WWXlZTjmmHMYOGgr+i+wGXvuNZy33/6gYZ0pU6bRs2dPevRo3BOoT59ePP/8yKa7bHDNtXcwdOjibLbZ59rs+NX2RjGNOmCpJhl3aXrzZlW+3ZaB/JePuZsJTGYmbzGN3zOGtenXUD37Yjk+7CRmNox5ezgjuZ3xjfb9JtMY1uT5FqcXfYiG52zpcUmfhg20UisZO3YCCy204GzLBw8eyNixEz/VdmPHTgCYbb3BgweWj899/+r85v/amchCC80+/MHgwQs2XDOqXVOYOcdhDPrSg4+bTPhVbWF68x5TG0188A5TSIpxbQEGl+Hz7SYTH7xV/vwxzU9gp/nhxAmSpNr17rtj+O3v/sGTT73ENVefxW9+fSqPP/4Ce+w5nMwil6y00tJMmTKVZ555pWG7jz+ewrPPvsqHH845306ePIW///1B9t1nm5qsqutOJjGTftTNViQwgB5MJZle5tcNGMCxLMHFvMfevMrXeJ16ku+zVMM2Y8u8ej7vsiUDOYthrEd/LuI9HmNSo+dcYA6ZegA9mFRm6pYel+bFScKaYwNtJxcRX4mIxyNiYkSMjYgnIuL8qseXi4iMiJ07+BgzIuY8UGY3MqdAkJk0lxNaul3T9SpBxhzS9bX+teNF0R3M+VXOecaVdRnER8zkFt5nIjN4n6n8g/cJZoWCvvRgLRbkfj5kJJOZzEweZSyvMRkwPLSZqGvdL6kLMOtK3Udmkpnc+Ldz2GmnTdh33235w+9P49///i///OdjAGy//UYsv/xSHHnUj3nxxTd4553RHPX1nzB+/EezVdVW/P3vDzBp0mSHN6hhWTaAVjLuU0zmEt5jVxbixwzjRJZkIvWcxdsNRQj15fftGMSXWJh16M/RDGFt+nEdYxvtf26Z+pMel+ahtXNuDWbd2jujGhIR3wN+BdwO7AkcAtwI7Fq12jvAxsCD7X6AamTw4IGMGzd7teP48ZPmWB3Z3Hbjxk1s2K5SKdt0vcrP89q/Or/5v3YWZNy42ccfHjdu3tupNvSlB1PmUCk7hXr6zuPP+2L0ZheG8CwTOY/XuJQ3GEpflqAPC1TNHboDi7EYvfkdb/IzXuVfjGVzFgFotJ4kzS+zrtS9DB48kLXWWpFFFpk1Juimm36W3r178dxzxfAFvXv34pqrz+a99z5k1dX2ZqmhO/Haa29xyME7MWTIwnPc7zXX3sFKKy3N+uuv3i7nobYzoOwJNrNJ4+hH1NOHoGfZFHolH7AhAziMxVib/mzOgpzCUjzDxzxSVscuWFbFrt1kToV16M//mNroOT+aQ6b+iPqGcW1belzSp+EnrM7tGOD/MvPkqmV/j4jTKz9k5lTgkXY/Ms1m1VWXm2280FGj3uWjjz6e4zih1ds98MDsE0G98MLr7L77lgCsuOIwevXqyQsvvM4WW6zXaJ26ujpWWWX2MWzVday66rKzjRk7atR7zV87n1mOBx7422zLX3jhdXbfbYvWPkx1MovSe7axZscznekki9J7ntuuyyDWYkE+ZDoL0IP+9OCnvMq6DGpYZwF68hWWZjzTmUo9i9CbRxjLAHo0DIGg1lSbXbWkZph1pW5ktdWWY+rU2cfrzMxGE8ptsMEavPLyDbz00v/o2bMHK644jJ13OZaNNlpztm3Hj5/Erbc+zPATDm7TY1f7WJre1APvMJ1hVXl2FNMa/fwm09iCxgUpw+hNH4J3y7kSlp5LHk4oZ1+Ytd2oJpn6A6YzhWx4zpYel+bGnNsSVtB2bgsB7zZdmJV+7cy521dE9ImISyNiXESMiYhzIuI7EVG93ZbldltGxHURMSkiXouIo6ufKyI2joibIuLtiPgoIp6MiAPb6Hy7tB13/H/cfvsjTJw4axbHa6+9k379+jSahXRO27377hgefPDJhmUjRjzHa6+9xY47/j8A+vTpzVZbrc91193daNtrr72TjTdei0GD7HHXle24w7yunblPdLDjjhs3e+2odq3MArzCR0ytuuP/LBPpSbAs/Zrdvhd1DKEPA+jJ00wggTWYvfJ6EL1YnD7UkzzBBD5X1YgrSZ+SWVfqRnb+4qY8/fQrjB49rmHZ/fc/wfTpM1hnnZUbrRsRfOYzy7LiisN4+eX/cdddj3H4Ybs23SV//eu9TJ06zeENasRq9KU/dTzIrN6FU6jn33zE+izQsGwxevFKVRUswP+YylSSxctCgtXoxwDqeKocoqviKSazfDmRGMD6LMB/+IjJVZn6fibSh2CtMlO39LikT8MK2s7tP8A3I+J/wM2ZOaaF2/0M+ApwMvA8cCiw31zWvQL4HXA5sD9wSUSMyMx/l48vC/wLuAyYAmwC/CYi6jPz6k9+SrXrqKP24qKLrmHPPU/gxBO/zGuvvcVpp13Od797IAMHzmpAXWml3dlii3W58spTAdh447XZfvuNOeSQH3Luud+mrq6OE0+8iE03/SzbbLNhw3annHIEW255JN/5znnsvvsW3HLLv7jlln9x220Xt/u5qnUdddSeXHTxtey514mcOPyQ4to5/Qq+e+wBja+dlfdki80/x5VXngJUrp2NOOTLp3HuOd+mri448aRfsOmm67DNNhs0bPfGG+/w2GPPATBt2gyee34k119/Nwss0M+G3C5sfQbxKGO5lrfZhMGMZTr3MoaNGdxo8rALGcly9GM3lgCKycUe4EOWpR91BCOZzMOMZReG0L9qu6eYwEySwfRiPDN4hLHUAZsy5+6FagWOHa3ux6xb4yZPnsIttxSjU7z11gdMmPAR119/FwA77bQp/fv37cjDUzv72td256KLr2WXXb/Lyd87lIkTP+LEk37BNttswKabfrZhvTPPvJJVV12WRRddiGeeeZUzz7qS/fbblm233XC2fV5z7R2ss87KrLba8u15KppPU6hnBEVRyhhmMJn6hkbP9VmAvtSxNwtzDWMYQA+G0Zu/MpYk2YVZQ2PsxCCu4AMWoSfr0Z9xzORqxjCEnny+bDDtRbA/i/AbPmAB6liFvvyLSTzLx/yEYQ372pFB3MRYzuZtvsRg3mU6VzGG3RnckI17t/C4NA8dkHMjYmng98ASQD1weWZeGBELA9cCywGvA/tk5thym+8BhwMzgW9l5u3l8vWA3wL9gFuAb2dmRkSf8jnWA8YA+2bm6/NzvDbQdm7fAP5GcRFkRDwP/AU4NzPnOIVlRCwCfA04NTN/Xi67HXh2Ls9xdWaeVa53L7ALxRhg/wbIzGuq9h3A/cAw4KtAs6E1Ir5WHg/LLLNEc6t3aYMHD+Tuuy/lmGN+xi67fJeFFhrAsccewGmnfa3RejNmzGTmzMZj3FxzzY849tjzOeywM6ivT3beeVMuuuiERutsuulnuf76n/KDH/ySSy+9nuWXX4qrrjqL7bbbqM3PTW1r8OCB3H3XJRzzzXPYZdfjimvnO/tz2mlfbbTejBkzmVnf5Nq5+myO/e7POezwM8trZxMuuvD4Ruvcc8/jHHrYGQ0/X3fd3Vx33d0su+ySvD7yxrY7MbWpfvTgEIZxC+9zNW/Tlzo2ZjBbluPEVtSTjUbVqiN4h6k8znhmkCxOb/ZmKVajcSV+kvyLDxnHDPpSx6oMYGsWpY+db9pODU52IDXDrFvj3n//Q/be+6RGyyo/jxx5E8stt9ScNlONGjhwAP+8+5d869vnsd/+36d3717sttvm/Pz8YxutN2bMeL5z7M8ZPXocSy89hOOPO5Djjpu9sH306HHcffdjnHnGUe11CvqUxjOTH/NOo2WVn3/N8mUD7WDqSf7Mh0xkJivTl7MYxuCq5qtdWYieBLcwjlsYxwDqWJ1+fJlFG83FsDuDSeDvjOUqxjCU3nyPJVmzalzaBenBjxjGpbzPGbzNAtSxO4M5oEmmbslxaR46JufOAI7LzP9ExILA4xFxJ8VN3rsz8ycRcRJwEnBiRKxOccN3DWAp4K6IWCUzZwKXUvy9f4SigXYH4FaKxtyxmblSROwH/BTYd34ONqp6EKkTKlvjtwO2B74ArAa8DKybmZMiYjlgJLBLZt4cEVsC9wCrZuaLVfv5CXBiZkb5c2W9zTLzwar1HgJey8yDyp8HA6cDuwFDoaG86q3MHFau8xXgN8CCmTn7jEWl9ddfPUeM+MOn+XWoO8rZB2yXWuL0uoM6+hDURZ3GS49n5vrt+Zzrr7NMjrh9eKvuM5b8Zrufh/RJmXXV7Zl1NZ9uMetqPn2xnbNuW+Rc+ORZNyJuBH5Rfm2Zme9ExJLAvZn5mbJ6lsz8cbn+7cBpFFW292TmquXy/cvtj6ysk5kPR0RPiqGbFsv5aGy1VKOTy8ypmfn3zDwmM1cHjgBWpmiln5PKrfsPmixv+nPFuCY/TwOq+xr9lqL1/xyK8Px54NdN1pEkSZ9atPKX1PmZdSVJ6g5aO+cGwKIRMaLqq3H35epnL274fg54FBiSme8AlN8XL1cbCoyq2uzNctnQ8t9NlzfaJjNnAOOhSfl1C1mL3cVk5pUR8TNg1bmsUploYTHgw6rli33S54qIvsAXgWMy87Kq5TbsS5IkqdWZdSVJUguNbkkFbUQMoBhC6TuZOSHmPh7unB7IeSyf1zafmOGjE4uIxeewbDFgEPDeXDZ7hmKCg92qtgmK8bY+qT4U3bwapkcsx+2YffpMSZL0KUQxeUJrfkmdnFlXkqTuoA1ybguzbkT0omic/VNm3lAufq8c2oDy+/vl8jeBpas2Hwa8XS4fNofljbYphzgYROMbyC1mBW3n9kw5RsYdFBfMssDxwGSK2Whnk5ljIuIK4PSImM6smW0H8glb8TNzfEQ8BpwaERMoZr07iaJke+D8nZIkSZpN0FGTJ0gdyawrSVKt66CcW97AvRJ4PjPPr3roJuDLwE/K7zdWLb8qIs6nmCRsZeDfmTkzIiZGxEYUQyQcAlzcZF8PA18C/jk/48+CDbSd3RkU1QEXAQtTdOl6CNg3M0fOY7vhQC+KwYzrgT9QXJTfmY9jOAC4HPg9MIZiMOX+wDHzsS9JkiSpwqwrSZLayibAwRQ3hJ8sl51M0TD754g4HPgfsDdAZv43Iv4MPAfMAL6RmTPL7b5OMW59P+DW8guK/PGHiHiFonJ2v/k9WBtoO7HMvAS4pJl1XqfJmBeZOYXi4vl6ZVlE3AU8VbXOvU23K5dv2eTnVyhm1G3qtKp1fktxoUqSpPnmsATqXsy6kiR1F+2fczPzwXk88dZz2eZs4Ow5LB8BrDmH5VMoG3g/LRtoa1BEbAVsCPyHorpgX4qLr1UuGkmSJKmjmHUlSVKtsYG2Nk0Cdge+B/QFXga+kpnXd+hRSZKkuXNiL6mlzLqSJHUl5txm2UBbgzLzMWCjjj4OSZLUUgE4SZjUEmZdSZK6EnNuS/gbkiRJkiRJkqQOYgWtJElSZ2DXL0mSJNUic26zrKCVJEmSJEmSpA5iBa0kSVJnYGWBJEmSapE5t1k20EqSJHU4J0+QJElSLTLntoS/IUmSJEmSJEnqIFbQSpIkdQZ2/ZIkSVItMuc2ywpaSZIkSZIkSeogVtBKkiR1ClYWSJIkqRaZc5tjA60kSVJHi4CwY5MkSZJqjDm3RfwNSZIkSZIkSVIHsYJWkiSpM3DyBEmSJNUic26zrKCVJEmSJEmSpA5iBa0kSVKnYGWBJEmSapE5tzk20EqSJHUGTp4gSZKkWmTObZa/IUmSJEmSJEnqIFbQSpIkdbjArl+SJEmqPebclrCCVpIkSZIkSZI6iBW0kiRJnUFYWSBJkqQaZM5tlg20kiRJnYIdmyRJklSLzLnN8TckSZIkSZIkSR3EClpJkqSOFtj1S5IkSbXHnNsiVtBKkiRJkiRJUgexglaSJKnDBYT3zSVJklRrzLktYQOtJElSp2DXL0mSJNUic25zbMKWJEmSJEmSpA5iBa0kSVJn4OQJkiRJqkXm3GZZQStJkiRJkiRJHcQKWkmSpA4XeN9ckiRJtcec2xI20EqSJHUGdv2SJElSLTLnNssmbEmSJEmSJEnqIFbQSpIkdQreN5ckSVItMuc2x9+QJEmSJEmSJHUQK2glSZI6A8fmkiRJUi0y5zbLBlpJkqSOFmFwlSRJUu0x57aIQxxIkiRJkiRJUgexglaSJKlT8L65JEmSapE5tzn+hiRJkiRJkiSpg1hBK0mS1Bk4NpckSZJqkTm3WTbQSpIkdQoGV0mSJNUic25zHOJAkiRJkiRJkjqIFbSSJEkdLiC8by5JkqRaY85tCX9DkiRJkiRJktRBrKCVJEnqFBybS5IkSbXInNscG2glSZI6A7t+SZIkqRaZc5vlb0iSJEmSJEmSOogVtJIkSR0usOuXJEmSao85tyWsoJUkSZIkSZKkDmIFrSRJUmcQVhZIkiSpBplzm2UDrSRJUkcLnDxBkiRJtcec2yL+hiRJkiRJkiSpg1hBK0mS1CnY9UuSJEm1yJzbHCtoJUmSJEmSJKmDWEErSZLU4cLJEyRJklSDzLktYQOtJElSp2DHJkmSJNUic25z/A1JkiRJkiRJUgexglaSJKkzsOuXJEmSapE5t1lW0EqSJEmSJElSB7GCVpIkqcMF3jeXJElS7THntoQNtJIkSZ2BXb8kSZJUi8y5zbIJW5IkSZIkSZI6SGRmRx+DuomI+AB4o6OPoxNbFBjd0QehLsfrRvPLa2fuls3MxdrzCSPiNorXpDWNzswdWnmfkubCrNss/+5ofnjdaH557cxdu2bdNsq5UGNZ1wZaqZOIiBGZuX5HH4e6Fq8bzS+vHUlSe/LvjuaH143ml9eOuhqHOJAkSZIkSZKkDmIDrSRJkiRJkiR1EBtopc7j8o4+AHVJXjeaX147kqT25N8dzQ+vG80vrx11KY5BK0mSJEmSJEkdxApaSZIkSZIkSeogNtBKkiRJkiRJUgexgVaSJEmSJEmSOogNtJIkqUUi4vSI+Er57+jgw5EkSZJajVlXHckGWkmziQjfG9RmotTRx6FPJiIWAdYDzo+IXukso5KkLsqsq7Zk1u2azLrqaP5hktSgEiQysz4iepfLfJ9Qq4mIyFJE9C2XeY11AZk5BvgJMAM4E3ztJEldi1lXbc2s23WZddXRvNgkERE9ACp3CSPiC8B/ImKhMsB6B1ifSkT0hOIai4i6iDgSuL1cVt+hB6dmVb0HPAH8H3BcRCzraydJ6grMumprZt2uzayrzsAGWklk5kyAiNg1Ir4JHASsBHy7XMXQqk8lM2cARMThwNHAF4ENI2L3crl/jzqZyodZmPWBNjM/Aq4FXgYuLNfz/UGS1KmZddXWzLpdj1lXnY1vElI3Vw6RNCAi/gRcA6wFLAEk8O2IWM3KAn1aEbFcRDwKnAGsDfQFegOnlGM8eXe6kyjfE+qqPswuHRELVq3yEvBzYNeI2LqsFPH9QZLUKZl11R7Mul2HWVedlQ20UjfT9O5tebdwJWBHiru9x2TmTsDXgPHAWVXrSXNVCS7Vd6OrbAGsCOwFfCsztwN+DCwNfK/czr9JnUA5bFp9RKwTEfcCNwKPR8T3I2LJzJwG3AXcCVxQ2abjjliSpFnMumorZt3aYNZVZ+UbhFTjImLHiFiy/Hdd5e5tRBwcEYuXq20BTAceKP8gQVFh8Htg84jYubJ9+x69OrtyjK09y2sry4kRZkZEj4g4vCrAbg28BjyVmVPKZRcDtwFfjYhlrF7pHMrX9FjgVuAD4FTgN8BuwI8AMnMk8AtglYg4utzO106S1O7MumpLZt3aY9ZVZ+UfIKmGRcQ+wNUUFQJUBdbjKP74zCxXfRVYDPi4fLx3Zk4H/g5MBU6s3l6qchbFQPp7lz9XgssfgQMoug8CjAKGZmblGqvLzHeBfwBDmHWNeXe6Hc3lg+hQYBXgZ5m5d2beDNQDnwUOrnyIBf4N/A44KyIW8LWTJLU3s67agVm3CzPrqiuxgVaqbX8D7gG2iYh1q5ZvCdyRmWPKn58D/kfZ/YYyaGTmCOBNYJOI+AZYWaBC1XXwO+AVYJ+IWKSsDFgQ+DxwddUHnbuAhSPiwPLn3uX364FJwCERsVmTfauNlGNv9ZjLB9EJwF+BX0TExhHxCvAt4DTgIWZ1BX0P+C0wGbiost+2P3pJkhqYddUmzLpdm1lXXZFvDFKNKu/aTgN+BSwKHFpZDnwO+EvV6u8BV1F0v1m/rCggIlYD+lAE38MjYrCVBYKiwqTs4vUicAOwJnBw+fCKwCCKioGKl4GbgZ+Vd6ArXb++CLwFPAt8t7LvdjiFbq0ce2tmtR1rFQAAIABJREFURKwSEWdFxLERMbR8Tcdn5h0U7xsXAw8Am2fmj4D7gbUj4shyV/+lqCDZzMoCSVJ7MuuqLZl1uzazrroiG2ilGhKNB6yvVAb8A3gQ2CIitgTWA3oAD5R3FiMzP6IYd+de4MYoBkjfETgBeBK4CRgMrNFe56LOqck1VrmDfBnwBrBHOQbcFsCrmflORPQGyMw3gfMouhHeWYakHYGvU3yAegRYNiKWbqdT6fYi4jDgMWAP4CcU/893qFplO2AF4A+Z+Wq57JHy+2kRsWhmjgfOycxVyvcRSZLajFlXbc2sWzvMuupqbKCVurCI+GxEDC//3TszZ5b/XpBZgQKKAc7rKWeuBV4E6ss7i5Vw+zJwIPBouc41wMYUg6Y/AiwLTEHdSkR8KSL2Kv9dfY0NqNz9z8yJwK+BJYGTKQbYf7x8rDIRB5n5CLAnxXhw3wf+DPTIzNMoKltWoOgCplbUtBtdRBwSxWQHq1OMh7YFxZhb/YFjImLFctXK+8ib5XYDKN4j7gCmAbsAVLqPRkTPNj8ZSVK3YtZVWzPrdn1mXdWKsEJb6poiohfwA+AUYM3MfC4ilgJ+TjHw+SjgHOCZzJweEWdShIk1gNeBlyi66/wLeD0zJ5f77UPRZWdIZj5TLvttud0XM/P9djtJdaiIWAj4E0WoGVR2E9qQ4g70TIpr6ILMfKlc/2qKILQW8DAwmiKY/hP4IDNnlOv1BxanCKyvRsSiFHe0XwMOrXQ71KdXVg1l1c99gXMpPsA+A+yUmW+Vjx0GDAeuz8wflCH0A4oJEm6keF/YnWL8vqczc3S7nowkqVsx66qtmXW7PrOuaokVtFIXVf5h/xPFH5SLywHLb6KYJfQpioqAGyi6bkFRWTCWYpKES5k1UcL9wOMR8dsykEwrg+m0iDggIm6lmLX0cgNr95KZ44CfUgyMf3pELEHRRWscRTev/YG/RMQe5Sa/pOhS+DDFhAozKQbU/w9wb0RcHhHDyg9IbwIDI+Io4E6KqpUrDayto3KHPzMzIgZGxEXlz1MoXpPHKV6fj8v3DjLz1xRBdvuI2Lz8kHEYRbXIaRRB9xeZ+c/MHF32GjVHSJLahFlXbc2s23WZdVWLvNikru1VijC6FUVgeAE4MDO/QXF393bg+IhYL4tZKK8qt+udmUcCq1GMyfMXigkSRpZ/5PoBmwHHUXTvWDMzr2jH81IHqwQZisD5K4q7zZWuWodk5uEU18jrwPkR0TMzH6AIoP2B0Zm5J7AqRfB5kuJaeqfc76LAjhSTJTwErJCZ97TDqXULVRUcWwB7UXTnOqp8+H/A5cA6wNrl//lKHvglxet3QBSTr/wV2ATYNjOXzsw/lPuNsteok1xIktqSWVdtwqzbtZl1VYsc4kDq4iJiKPBj4CDg0jKwVh5bG/g/YHxm7lAGkT8Dw4CT5xQSyj9U9RGxCLBgZr5evbztz0idTUSsQzHD6abAkdUfYCJiO4oxuX5bdhVaHvgD8CHwzcx8Yx77XQqYUalWiYgelXG/1DIRsQPwWGVsrKrluwNXUsw8mxQfMMYCwzLz44gYAlxXPrZdZk6t2vbnwD7A8Mz8U5P99qwEYkmS2oNZV23NrNt5mXXVnVhBK3Vx5Zg6VwHvA/2gYcwugOeBW4GVImKdcnyey4GVgV0r+6jcQa4Oppk5piqwhoG1W3se+CNFN6Fe0Ogaexy4j2Lm5IUycyRFd8PNgW0qO6i+xirLMvPtzHw/IurKa8zA+glExGYUwXPJJsuXAH5E8TocStFt80yK2akvKlcbDZxNUTGwX7ld5bW5ArgbGNH0OQ2skqT2ZtZVOzDrdkJmXXU3VtBKXVil60UUA9yfChwFrJyZb1VVB+xBMUbS2lUhdE/g5qyadVSal4hYlqJCZenMXKNy7ZWPnU8RUDcs71j3AbbPzJs68JBrUkT0BqZX/e4XL4N/9euxJXAzcFBm/q1cNgg4nqLr3lqZ+d+IGEhRLbIxsElmflC9H0mSOppZV+3FrNs5mHXVnVlBK3VS0YIBySt/XLIY4P5PwHvAhRGxYBlYewBrAh8B9VXb3ZCZ01ryHKpNVeNutUjZfeuXwIoR8X2KCRIox3BbARhP0YWIzJxaCaxeY60nIr4BXE0RPisfWt+PiG8DJ5avBcAAIChmCq5UC40H/koxjtoFAJk5gWIm7JWA75TLqmfB9bWTJLUZs67aklm36zHrqrvr2dEHIKmxSjVAGTqHAFMyc3wL7vY9C1wInAf0iYi7KP5wHQtcD7zbdAO7cnU/ZVitm88uVv+iqFA5FVg1Iu4DVqScACGLWVMb8Rr79CJiV4r/21MputmtHsXMtZXXcE2Krlu3A08A91LMRrwL8DTFzdj6zPxPRLwGbB0RX8zMfwAvUsxQ/EjT5/W1kyS1BbOu2pJZt+sx60oF7xhInURZAUAZViMiVqeYBXSXlnTFKAc+/wfwN+CLwBCKwc/Pz8yv2cVLUNw1zsyZEbFMROxRXmct3XYM8HvgGeBAYAlgW+CwzLykbY64e4tiZtqfUIyxtStwVGYempkzqt4ThlNMinBMRAygCKx/AIZHxDJNxtKaBIyjqCYgMz/OzGsz841PWmkiSdInYdZVezDrdi1mXWkWG2ilTqJylzcivgXcxqyJDXaj6FbTEq8CN5b/vgbYPDPPKvfbo/WOVl1J5bWv+n4Cxd3k/wOeiIjDy7G0WtId7CngDmAGcFFmrp/l7Kd2E2oTBwMfA2cBL2fmRChmmC2/98jMscAZ5bqblNUAlwH/A66OiG0jYomI2BdYEPgh0L+sVmjgeFySpLZk1lVbMet2aWZdqeQbjNRJRESviLiY4o/TvRSh4DZgT2CnKAZMn9f2ldlnbwKWz8ynM3NGOGtot1VWpzS89mU1QQBrA4dTfDC6nGKcpu3KdZqrXplEMdj+kHJcp0YVMW11Lt1R+XtdBXgqM8c2eW0GwawPu8CVFB8oToyIxTLzJeAAYCGKD7IPUozpdQPF+8sgylmKJUlqD2ZdtTazbtdm1pUaC28iSO1vTt24ImJp4CGKPz5nZ+b0KGaj/C2wFvClzHzyEzxHnSGieyqDaeVDDBGxGUXXoOeBvsCiFF21ppSPPwOMAY7IzFc+wfP08MNQ24qIo4FvA2cCD1N84DgYWBJYGLgOuDIzR0bE1sCdwNeA32cxOcryFON2fQa4LTOfjYiDKD547JiZs43HJUnSp2XWVVsy69YOs640iw20Ujua1x/5iFiOYpDzQzLzb5XQGRHLUHTR+QVwZmZOmNs4XYYIVSsrUdak6KY1gmKSgyUpugR+jeJvwMyI2BS4n2J200szc/o89uk11o7KyoKngZWB6UA/4AOKLl2TgU2APwInZOYHEXE1RbDdMzNfrNpPL4qx+rYDfgTcAxxZqQyRJKk1mHXVnsy6XZ9ZV5rFIQ6kdlAZ66hq7K29IuKkiNg4IvqXqw2i+GO0ZkT0LANrz8z8H0V3jYOADcv9NK1IaNS1KyLWj4iB7XN26gyqxtyqK7/vAbwAHAGckZk7UIzxdgtFcFmsDKx1mfkg8FfgWxQVLHPdf5NrrGfbnpXK3/cWwDeAHwO7A1/MzM9n5hbAuRRdQzcsN/k+sBpwQNU10Y+ii9/1wDnAFZm5v4FVktRazLpqa2bd2mTWlWaxgVZqAxExtPxeB7NCZkQsFhH/AH4DHAncTtEdh8x8CngN2IGiiwY5a0bKDylnqo2IxaqeJ8q7vPWZmRGxZUSMoqhAGNT2Z6rOomrsrUpXvwEUd6H3pbiDTGY+B/wUGAgc3WQXRwPDgCPK7oZAwweiuqqwelhEzAC+DvRHbS4zR2fmFZl5VmbelJkjqh7+CcVr3bdc9zVgD+CCqmviY+BRitlsV8nMU9r3DCRJtcasq/Zm1q1dZl2pYAOt1Moi4ijg+YjoVzUu0sCIGE5xN3cSxV3CjSnG1DkwIo4sNz+F4q7uUWWXHSJiCLAYRWXBV4ClyuW9sjAzIoaWYfhO4GZg38wc1T5nrM4gIhaNiD9FMWstFNUDfwcGU77Xl9UtTwEXAt+NiFXL6pUemfkexWyon61av/KBqL6sgHkKOB84DTjOu9Ltq/wA0aPq517AIcBY4PXK8sy8MTPHRtVMw5n5ZmZem5lj2vOYJUm1x6yrjmDWrX1mXXV3jkErtbKIWA0YmJmPVi1biyIsTAROz8zzy+XLUwSIJYBdM/PdiDgdOLRc9ybgc8DiFN09HgEuzswfV+37HIrxlO4HTq5+XtWmsjvgjCbLlgV+TzFu0y6Z+V4UEyZcAbyYmbtVrbsMxVhd/87MQ8oug3P8YxARiwOXUHQt+gPF9TuyTU5MLVIG10WALwAnUbwvfHNe46lJktRazLpqa2bd7s2sq+7KClqplVSNW/R8Zj4aEbtExOrlsmco7sQuSHEHsDK77UjgLxRdtI4td/Uj4EBgFEUXsHpgZ6AHRdeOaeX2QyLiHopZLg+imKXSwNoNVAJrdRfAzHwD+D+Ka+y4cvGjFF0Mt4yIHcptohzr7cfAQRGxfVW3xCi/9yy/rwE8ASxLUQlzqIG145TdPH9G8UHkjxSzYN+WmUcZWCVJbc2sq/Zi1u2ezLrq7qyglebT3O7EluF1EMVMouMyc91y+WLAkxR3c79d6TITEQsCF1EMfH5AZj5ZtR/Kbl09KcZT2gPYLTOfiWLChbUp7hiPbduzVUcqw2RUdSPsCdwLvA0cnpkTy+UDKKpUNgEOyswRZUXL+cDCmble1T4XoQi3l86ti2DZ5XA54MnMnNpGp6dPIIoJMQ4CRgIXVl67KGfC7tCDkyTVFLOu2otZVxVmXXVnVtBKn1BE9IyIC4BtqpZFRPSOiEuBL2XmhxQzTK4REfsAZOYHwFnAAcAGlW3LwHEVxSD3Z1ctnwlsEhHnAv8CvgqcVVYokJmTM/MRA2vtK8dfqy/H3upbVhVcR1FtslHVepMoxm+bTFlZUF4vfwKGRsR3oeED15jMPHle47dl5nuZ+aiBtfPIzL8CX8nM4zNzVET0KF9PA6skqVWYddXezLqqMOuqO7OBVpo/ewPHR8Si0DBz7Wcp7vo/Uq5zD8XA9edUbfcr4FnghHK8I8rt7wTuAh6NcibR8qHXgTWAh4AlM/PXbXZG6tQiYnPgPWDzctFFwHPA8PLuPwCZeRdFRcv2EbFzufifwGPAN6KccKNqv9Eex6/WU1VFUpeZM+c2ppokSZ+CWVftyqyrCrOuuiuHOJDmQ0RsRzFz6KHANZk5PSIuBNbJzC0rXcIiYmvgr8C5mXlGue225bZHAH8sqweIiN6ZOa3qOSr7GFDeLa7MNDqzXU9W7W5OXQojog9Ft8EXga9m5gcRsQ1FN8IjgN/nrPG6dgZupAirX8zMaRGxNjCyEngkSZLmxqyrtmTWlaTZWUErzYfMvIMiFJwELF0Giq2BG8pVKv+3HgMuB75XufNbVhDcApwHDK3aZ2VChLry5yy/Tyq7lYWBtXsoP6ysHxGrQMNMtlOB44FdgW3LDzB3AX8GTgQ+U67bi6JL4rMUlS57lft8OjMnVsZ7kyRJmhuzrtqSWVeSZmcFrdSMpnfyKwOUR8RnKO7yngHcRBEe9quMm1W1/mcpZq99KDMPLpetTHG394L2Og91Xk2rCMoKgBuAZzNz9+p1IuJGYHmKCTRGRsQSwMsUXQOvAaZSDKz/S+CdzHy8nU9HkiR1IWZdtTWzriQ1zwpaaS7K8bEa7uRHxArVYxhl5osUYyMdC3wdWA2YFhG9m+zqBeDnwIERsVm57csGVsWs2Ysb3SnLzKeB64HVImLXcnGlGuBkigqC/SJigcx8F/gKsDhwLnAJ8Fhm3lwJrI69JUmSmjLrqq2ZdSWp5ayglZpRjq11JtAXmALcnZmnlI8NpujatQIwCugFjANuo5h5dEQ5ZteSwCnABZn5UtW+Zxt/SbVvDlUEewErA/cBz5Rd/dYFzgdmAttn5oyy+9eMiPgTsAmwV1UwHQisDbyQmaPn9DySJElNmXXV2sy6kvTJWUErzUVE9I2IHwFXU8wgeiZF166TIuIwgMwcC/yw3OQbwP7AzcCWFMH1w4g4JzPfycyjqwNrub2BohuIiKHl90ZjrkXEYhHxD+A3wJHA7RRjb5GZ/6GYGXk54Jhy2Yxyl28DywCHlx+cyMwJmflgZo6OiB4GVkmSNC9mXbUWs64kfXo20ErMChNNuscMBdYEvpOZR2TmX4FhFN1vTo+IFcv1rgIeB04AnsjM4cB6FN3AjsnME5o+j7qPiDgKeD4i+mVmfblsYEQMB7YDJgFbABtTVKIcGBFfLTf/MzACODIiFi+3XZKii9ffgN0pql0aycyZBlZJklRh1lVbMetKUutwiAN1a1V3eevn8NjCwFqZeV9E7A5cCoynGIvrJOBO4IhyMPvNgLsp7ghfCsxo0q2nbk7PodoXEasCgzLz0aplawFPAROB0zPz/HL58sCFwBLAzpn5fjku19nAIOC3FLPVLgjsAkz2upIkSXNj1lVbM+tKUuvwDqe6tcysz2KW2g0j4sKI+E5EbFU+9mEZWNcGfgRcC2wOXEkxk+1BFHeCycwHKELsVkCPOQyEb7DoJiKiX/m9F0BmvpCZj0bELhGxernsGeA0ivA5tlw/MnMkxSzIg4Dvlru8BTiEouvhbhRjvx2WmZPKa7dnu52cJEnqUsy6am1mXUlqGzbQqluIiEERsVNEfKa661U59tZFFAPWrw58E7gxIr5RCR/AzsCSFJMevJ+ZU4EPKcLDDyKif7nePpm5R2ZOabcTU6cSERsAp0XEgpk5vVzWNyIGABcDf6xa/VLgHWDziBhY9UHnBuAhYNeI+GxmzsjMJyi6eG2TmTtm5siqipgZSJKkbs2sq/Zg1pWktmMDrWpeRJwNvABcBvwXuLZqTK31KcZGOgLYC1gXGA78BDigXKcf8Brl/5eIWIxizK1rgR0owi6Z+VH5eI82Pyl1VknxwefQiOgVEX8B/pCZk4DvAWtExD4AmfkBcBbFdbZBww4yJ1KM9TadortXZfmUchsiooeVKpIkCcy6aldmXUlqIzbQqmZFxHYR8QFFGP0+RdeZbwC7AidGxCDgS0DvzPxjOTPoeOBdYAFgnXJX9wGLAheVgeN7FLOKngEsnJkjqp83M2e2/dmpM8rMx4ALKMLmJGAp4Iry4XsoZqo9p2qTXwHPAidUJkYo93MncBfwaETUNZnQw2tMkiSZddXuzLqS1HacJEw1qeyKdT2wPbB0Zr5d9dgPgWMpqgm+BQzIzN0j4nPAr4FVgPMounl9WG5zHHAoRXidCHwjM+8oH3NShG6qOkyWE2gMpJhAYz2Krltbl90EK+tvDfwVODczzyiXbUsx9tYRwB8rgTQiemfmtHY7GUmS1GWYddUezLqS1H6soFVNyszJwPnA28BXoQgB5cPXAAOBvsAo4PMRcTPwOEX3sHUz81RgQkTsUAaTn1NMkrBLZq5cCazlcxlYu6Gy61WW42lVwutUilmPT6G4XtYt16281z4GXA58LyKGQEMFwS0UH5SGVvZfCazV48hJkiSBWVdtz6wrSe3LClrVrHLGz4spunatn5lvlMv3pehusznwPsWMoVMpZgu9uVynN7A/8GXg0Mq21ft2wPruqbqKpJy99mRgEeA24O7MnBoRSwN/ogizm1fPdBwRn6WYvfahzDy4XLYy8MXMvKB9z0aSJHVVZl21BbOuJHUM71apZpWh8mJgNHAqQETsCVwCPAC8mplvAT+lqDBYKSJWiIh1gR9SjLv1ODBmLvtWN1QVWFcGngD2BXYCfg/8uFztbYrxtzYGKsG0UnnwHEWVyoERsVm5z5cNrJIk6ZMw66otmHUlqWNYQauaVgaF4RQTJ7wJLAv8PDN/0GS984DdKWax/QjoAxyfmX9u3yNWZxcR6wAHAj0oug+eTHHdfJXiOtsiMx+IiIWBC4FNgLUzc1JELFRul8CPgPMz86WqfUf6pixJklrIrKvWZtaVpI5hA61qXkQsQTEhwqYUY269Ui6vo/g/UBmoflmKiRGGZOYtVds7MUI3VY69NbPJsj0oum29CXw9M/9RLl+eYhbbxTJznXLZBhRjbo0Abge+DjyZmfu031lIkqRaZtbV/DLrSlLn4RAHqnmZ+S7wR2AmxVhclTBSXx1IMvONzHy8EljLcb2cGKEbiogeAE0Da7nsr8DvgGHAgKrlIym6Ga4WEUeVix8DjgKGAMcAf68OrE6KIEmSPi2zrj4ps64kdT5W0KpbiIj+wG+ADYBVy8HtrRbQbKorCSLiy8A+FONsPQLclJkfRMRqwD+BG4GTMnNcuf4iFOO87QIsl5kfl8uXAj7KzPFNn0OSJOnTMuuqpcy6ktQ5eUdL3UJmTqa44wtwdkceizqPiPh8RJwXET+KiOFQVBJExMIRcT1wHjAeWJ2iS9flEbFUZj4PXA1sB2xb2V9mjgGupKg2+GXV8rczc3xE1JVjbxlYJUlSqzHrak7MupLUddhAq+7kYeBu4IiIWNiKgu4rIgZFxFXA/cAKwNrAwhGxXLnK5yi6CB4KHJqZmwDfAtagqBoAOA2YBuxbjulW8TTFzMj3N33esquh3RYkSVJbMOv+//buPUausozj+Pdhy0UU5CJaAUPbYCAmBhTFaEFaLiIqN0X8Q4VFoUDUiJd4CSotwYjRRFTUUkQWrzFgRAQlorLEolZELsYoMaktF4VyEYuIBdrHP847MM7OzM7s7uzsMN9P8ubsznnPe94zO0l/feec9xVg1pWkQeQUBxoqZRGFLZm5od99UX9ExM7AKqp5tT4B3Fx7bKuuzoVUdwu8PDMfLa89G/g4cDpwXGbeWObfOhs4PzO/MouXIUmSNIFZV2ZdSRpM3kGroZKZ92bmhtrE+BoeERHlx9cDS4HzgOvr5tQaqVvI4B5gl7rAOq/8fC2wsbRBZq4EHgBOi4hFLc4nSZI0K8y6w8usK0mDbV6/OyD1g/MiDZ+6x62OB9Zk5jUN++s/E7cDj0fE6Zl5EeXLrMxcHRHPAuofGTwD2C4z17Y4nyRJ0qwy6w4fs64kDTbvoJU0FOq+5X8e8GR5baRFnT8Ca4CzIuL5mfl42f9SYFvg3toxmbkmM2/ocfclSZKklsy6kjTYHKCVNBQyMyNiG+A/wE4RsVtZxTbq65TtncDXgCeAayJiWUQcAZwL3A/8dPavQJIkSWrOrCtJg80BWklDo9wdsAbYn2o12wmPZ0VEbeqXB4CTqFavPQ/4JrA7cEJmrpulLkuSJEkdMetK0uAKp46RNEwiYgdgHXAV8LHMvC8iRurn5YqIXYDvAJ8EbgPmA7tm5q1l/1aZuWVC45IkSVIfmXUlaTB5B62koZKZj1A9vnUS8I7yWn1g3RH4CLAjsDYzn8jMu+oC64iBVZIkSXORWVeSBpN30EoaShHxY+CVwOXAZ6kWRNgVOBM4ElgBrHKFWkmSJA0as64kDRYHaCUNpYiYD5wCfIgqsK4HdgAeBs7MzF/3sXuSJEnSlJl1JWmwOEAraahFxIuBfYBtgEcy87ryesDEhRUkSZKkQWHWlaTB4ACtJDVoXEhBkiRJeqYw60rS3OMArSRJkiRJkiT1yVb97oAkSZIkSZIkDSsHaCVJkiRJkiSpTxyglSRJkiRJkqQ+cYBWkiRJkiRJkvrEAVpJkiRJkiRJ6hMHaCVJkiRJkiSpTxyglSRJkiRJkqQ+cYBWkiRJkiRJkvrEAVpJPRMR2VA2R8QDEfHLiHh7h/UfiojxiBiNiGhyzIImxzUrC+qOWd6wb0tEbIyI9RHxk4j4aETs0eKaaudb1+a6D4yISyLijoh4JCI2lbaviIgTI2Kki37XlyWl/fEO6i6f5D16LCI2RMRvI+LCiDi4iz9trd3R0tbYFI69rhx7V0SMtKk31uJz8WCrz1E5bl0H79FoXf0l5bXxbq9FkiQNJ7OuWbfNsWZdSV2Z1+8OSBoKK8p2a2Af4DhgaUQckJkfnKT+3sDxwCHAK4D3tjjHv4AL2vTh4Sav3QCMl5+fDbwQWAwcBayIiOWZeX6bNv9PRGwNfAk4A9hc2r8G2ATsCRwKvAX4AXAqT19nvXPKttm+dQ2/X9bktZrxJq/Vv0fzgF2A/YAzgfdExM+AkzLzvhZtzoiIWAQcBiTV+3IUcPUkh/0IuLX8vA2wCDiG6nP0ksw8u8VxX6T535669iRJkqbDrGvWfYpZV9KUZKbFYrH0pFCFkmzy+mHAllIWdFB/MVUI3AIsbNi3oBy3rot+LS/HLG+yL6iC5YOlzsc6PR+wquy7Hdinyf4R4B3A5d2+Zw11xku9JR1eb9v3iCoAXl/q3AJs12G7o+WYsS4/F58px9W2V7WpO1bqjDbZd0DZ91hjn6nCfNZ/vibp05JSf3ymPv8Wi8VisVie2cWsO2G/WTfNuhaLZWrFKQ4kzbrM/AXwF6qA+MoO6t9YV/+AHvctM/MHwAnlpXMi4oWTHRcRrwFOAx4CjszMO5q0vTkzv00VXOeMzFwLvJHqPd6f6q6InoiIeVRhdyNwLvAH4A2tHrNrJzNvpnq/twN2mMFuSpIkTZlZ16yLWVdSlxygldQvtTm2ssv6T/SgLxNk5vXAaqpA9OYODjm9bFdl5j8maXvTNLs34zLzP8Dny69N57qaIccA84HvZ+ZjVHcNjADv6rahiHg51aNr6zPz/pnspCRJ0jSZdecQs66kuc45aCXNuog4nGp+rgRu6qD+a0v9x4Hftai2U/1iAQ3uzcyVU+jqOHAQcCDwlUnqHlS2v5jCeaZqtLaYQhMrM/PeLtsbL9uXRcS8zHxyyj1rbVnZXlq236UKy++OiE9n5pYWxx1Xt/jFNlSPsh0D3A28s835zoqIVvNynZ+Z/+2w35IkSR0x684Ys65ZVxoaDtBK6rm6MFm/cEIAX8jM9ZPUry2cEMCH23xj/1yeXnRFX6RzAAAEZUlEQVSg0W3AVELrPWW7Wwd1a4+G3T2F80zVyW32XQl0G1pr1ztC9W39hql0qpWI2As4ArgjM38DkJkPRsTVVHduvA64tsXhx5ZS7zGq0PvHNqd9f5t9FwCGVkmSNC1m3Z4x65p1paHhAK2k2VALk0m1yuivgEvKHFXt6tck8O7MvLRZ5WJ9Zi6YVi8n6vbRtG7rTtfSzByfwfai7udeXMepVFPrjDW8PkYVWpfROrSekpljABExQrUi7slUi2AcGxGvyMx/NzluYWaum2a/JUmS2jHr9oZZ16wrDQ3noJXUc5kZpWyVmbtk5tI2gfWp+sBzqL6FvgtYGRGHzlafi93LtpM5n2p3O+zZo77Mhtr1bgb+OZMNl6B5CtXqxN9q2P1Tqjsgjo6I+ZO1VRagWJ+Z51LdVbAP8L6Z7K8kSVKnzLoDw6wrac5ygFbSnJWZj2bmz4GjqR5Fuiwitp/FLiwt2zUd1F1dtof1qC+zoXa9N/dgTq43AXtQ/btzd0RkrVAthjGf6qmObhdQqP1tDpyxnkqSJM0Cs+6sM+tKmrMcoJU052Xm7cDFVN/Yf2A2zlnuYFhMNffTDzs4ZFXZLouIF0zS9rbT7N6MK/8Z+FD59Ts9OMVpZXs1cEmTMlb2nxoRMeHo1nYuW/89kyRJA8ms23tmXUlznXPQShoU5wGjwIcj4quZOaOPJdWUwHQ8VUgGOKeTFWIz88aIuJgqnF0bESdm5l8b2t4KeFtp/8SZ7fnURcRC4BvAvsAtwEUz3P6ewOupHiV7a6vVZCNib6oVgg8Hruug3Z2pHiWDp1fllSRJGkRm3R4x60oaBA7QShoImXlPRFxEtVLpR4CPN1TZqW5F3GbGmkygv6TumGdRzUu1GFgIbAI+mpmf66Kb76Ga0+oM4M8RMU61qu4mqkeeDqW6M+KKLtpsZzQilrTYd2tmXtnwWv17NI/qG/n9gFdTfSt/LXByZm7qsh8HRcRYi31/oFoldwT4dqvAWnydKrQuY2JoPS4iFpSfawsnHA3sCtxE65WLz4qIh1vsG2+y8MS+ba7lzsz8VKvOS5IkTZVZtymzrllXGhqROZuLMEoaJmXOJcoiCNOuXx6nWlt+XZSZ95Ug87cOmn9qFdgS3OpXz03gUeAh4E/ADVQB654mfaidr+VKuhHxKqrgdTBVEN4a2AD8HvgecEVmbmlx7KTvWQnDh7TaX1yWmaMNfa63CdhI9X7eBHw/M1fThYgYBdqtNgzwY2B/4EXAfuURvlbtbQ/8Hdge2DMzN5QAeXKT6o8AfwEuB77cGIYjYh2w1yR9W5GZy0v9JcD1k9S/LTP3n6SOJEkaEmZdsy5mXUkzxAFaSZIkSZIkSeoTJ5qWJEmSJEmSpD5xgFaSJEmSJEmS+sQBWkmSJEmSJEnqEwdoJUmSJEmSJKlPHKCVJEmSJEmSpD5xgFaSJEmSJEmS+sQBWkmSJEmSJEnqEwdoJUmSJEmSJKlPHKCVJEmSJEmSpD75H961vJIn0zlwAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1440x576 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import itertools\n",
    "plot_confusion_matrix(cnf_matrix_xg_train, classes=['Peaking_bkg', 'Continuum_bkg',\"Signal\"],\n",
    "                      title1=\"Confusion matrix XGboost train set \\n with normalization\",\n",
    "                      title2='Confusion matrix XGboost train set \\n without normalization')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x576 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cnf_matrix_xg_val = confusion_matrix(Y_Validation_single_col,y_pred_val)\n",
    "\n",
    "plot_confusion_matrix(cnf_matrix_xg_val, classes=['Peaking_bkg', 'Continuum_bkg',\"Signal\"],\n",
    "                      title1=\"Confusion matrix XGboost validation set \\n with normalization\",\n",
    "                      title2='Confusion matrix XGboost validation \\n without normalization')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['xgboost_model.dat']"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "joblib.dump(model_xgboost.best_estimator_, \"xgboost_model.dat\")"
   ]
  }
 ],
 "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"
  },
  "latex_envs": {
   "LaTeX_envs_menu_present": true,
   "autoclose": false,
   "autocomplete": true,
   "bibliofile": "biblio.bib",
   "cite_by": "apalike",
   "current_citInitial": 1,
   "eqLabelWithNumbers": true,
   "eqNumInitial": 1,
   "hotkeys": {
    "equation": "Ctrl-E",
    "itemize": "Ctrl-I"
   },
   "labels_anchors": false,
   "latex_user_defs": false,
   "report_style_numbering": false,
   "user_envs_cfg": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}