SigvsBkg_LL-TTvsOther_DNN_RECO_VBSpol.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c157cd32-2e52-4c7e-9b91-05a41aec33bb",
   "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>CleanJet_eta_1</th>\n",
       "      <th>CleanJet_eta_2</th>\n",
       "      <th>CleanJet_phi_1</th>\n",
       "      <th>CleanJet_phi_2</th>\n",
       "      <th>CleanJet_pt_1</th>\n",
       "      <th>CleanJet_pt_2</th>\n",
       "      <th>Lepton_eta_1</th>\n",
       "      <th>Lepton_eta_2</th>\n",
       "      <th>Lepton_phi_1</th>\n",
       "      <th>Lepton_phi_2</th>\n",
       "      <th>Lepton_pt_1</th>\n",
       "      <th>Lepton_pt_2</th>\n",
       "      <th>Rpt</th>\n",
       "      <th>Zepp_l1</th>\n",
       "      <th>Zepp_l2</th>\n",
       "      <th>Zepp_ll</th>\n",
       "      <th>detajj</th>\n",
       "      <th>detall</th>\n",
       "      <th>dphijj</th>\n",
       "      <th>dphilep1jet1</th>\n",
       "      <th>dphilep1jet2</th>\n",
       "      <th>dphilep1jj</th>\n",
       "      <th>dphilep2jet1</th>\n",
       "      <th>dphilep2jet2</th>\n",
       "      <th>dphilep2jj</th>\n",
       "      <th>dphill</th>\n",
       "      <th>dphilljet</th>\n",
       "      <th>dphilljetjet</th>\n",
       "      <th>dphillmet</th>\n",
       "      <th>dphilmet1</th>\n",
       "      <th>dphilmet2</th>\n",
       "      <th>dr_l1j1</th>\n",
       "      <th>dr_l1j2</th>\n",
       "      <th>dr_l2j1</th>\n",
       "      <th>dr_l2j2</th>\n",
       "      <th>drll</th>\n",
       "      <th>ht</th>\n",
       "      <th>m2ljj30</th>\n",
       "      <th>mT2</th>\n",
       "      <th>mTi</th>\n",
       "      <th>m_l1j1</th>\n",
       "      <th>m_l1j2</th>\n",
       "      <th>m_l2j1</th>\n",
       "      <th>m_l2j2</th>\n",
       "      <th>mcoll</th>\n",
       "      <th>mcollWW</th>\n",
       "      <th>mjj</th>\n",
       "      <th>mll</th>\n",
       "      <th>mtw1</th>\n",
       "      <th>mtw2</th>\n",
       "      <th>phi_miss</th>\n",
       "      <th>proxyW_l1</th>\n",
       "      <th>proxyW_l2</th>\n",
       "      <th>pt_miss</th>\n",
       "      <th>ptll</th>\n",
       "      <th>recoil</th>\n",
       "      <th>yll</th>\n",
       "      <th>isLL</th>\n",
       "      <th>isMIX</th>\n",
       "      <th>isTT</th>\n",
       "      <th>isBKG</th>\n",
       "      <th>isVBS</th>\n",
       "      <th>notLL</th>\n",
       "      <th>notTT</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-2.805664</td>\n",
       "      <td>0.871460</td>\n",
       "      <td>-1.175293</td>\n",
       "      <td>-3.091797</td>\n",
       "      <td>115.945938</td>\n",
       "      <td>105.480812</td>\n",
       "      <td>-0.222351</td>\n",
       "      <td>1.390381</td>\n",
       "      <td>1.880859</td>\n",
       "      <td>2.354492</td>\n",
       "      <td>60.912807</td>\n",
       "      <td>42.632568</td>\n",
       "      <td>0.212335</td>\n",
       "      <td>0.744751</td>\n",
       "      <td>2.357483</td>\n",
       "      <td>1.551117</td>\n",
       "      <td>3.677124</td>\n",
       "      <td>1.612732</td>\n",
       "      <td>1.916504</td>\n",
       "      <td>3.056152</td>\n",
       "      <td>1.310529</td>\n",
       "      <td>2.335937</td>\n",
       "      <td>2.753400</td>\n",
       "      <td>0.836896</td>\n",
       "      <td>1.862304</td>\n",
       "      <td>0.473633</td>\n",
       "      <td>3.032799</td>\n",
       "      <td>2.141702</td>\n",
       "      <td>1.970499</td>\n",
       "      <td>1.776264</td>\n",
       "      <td>2.249897</td>\n",
       "      <td>4.001696</td>\n",
       "      <td>5.091536</td>\n",
       "      <td>5.483263</td>\n",
       "      <td>5.470954</td>\n",
       "      <td>1.680843</td>\n",
       "      <td>419.124969</td>\n",
       "      <td>985.871765</td>\n",
       "      <td>58.271770</td>\n",
       "      <td>158.066528</td>\n",
       "      <td>328.959351</td>\n",
       "      <td>136.063660</td>\n",
       "      <td>581.079773</td>\n",
       "      <td>68.985100</td>\n",
       "      <td>54.399845</td>\n",
       "      <td>170.509857</td>\n",
       "      <td>703.374695</td>\n",
       "      <td>94.460876</td>\n",
       "      <td>78.216949</td>\n",
       "      <td>76.092033</td>\n",
       "      <td>0.104595</td>\n",
       "      <td>79.019096</td>\n",
       "      <td>99.086594</td>\n",
       "      <td>41.708782</td>\n",
       "      <td>100.746727</td>\n",
       "      <td>92.839317</td>\n",
       "      <td>0.465608</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-1.510010</td>\n",
       "      <td>3.178711</td>\n",
       "      <td>-0.928223</td>\n",
       "      <td>0.462524</td>\n",
       "      <td>48.520248</td>\n",
       "      <td>43.383701</td>\n",
       "      <td>-0.576294</td>\n",
       "      <td>-0.817017</td>\n",
       "      <td>-3.005371</td>\n",
       "      <td>0.551392</td>\n",
       "      <td>109.715820</td>\n",
       "      <td>29.537357</td>\n",
       "      <td>1.539541</td>\n",
       "      <td>-1.410645</td>\n",
       "      <td>-1.651367</td>\n",
       "      <td>1.531006</td>\n",
       "      <td>4.688721</td>\n",
       "      <td>0.240723</td>\n",
       "      <td>1.390747</td>\n",
       "      <td>2.077148</td>\n",
       "      <td>2.815290</td>\n",
       "      <td>2.725920</td>\n",
       "      <td>1.479614</td>\n",
       "      <td>0.088867</td>\n",
       "      <td>0.830842</td>\n",
       "      <td>2.726423</td>\n",
       "      <td>2.220245</td>\n",
       "      <td>2.869017</td>\n",
       "      <td>2.640638</td>\n",
       "      <td>2.783735</td>\n",
       "      <td>0.057312</td>\n",
       "      <td>2.277360</td>\n",
       "      <td>5.111395</td>\n",
       "      <td>1.633860</td>\n",
       "      <td>3.996716</td>\n",
       "      <td>2.737029</td>\n",
       "      <td>309.260803</td>\n",
       "      <td>738.418823</td>\n",
       "      <td>35.363697</td>\n",
       "      <td>184.284683</td>\n",
       "      <td>144.914749</td>\n",
       "      <td>461.100403</td>\n",
       "      <td>58.736408</td>\n",
       "      <td>259.231140</td>\n",
       "      <td>139.885406</td>\n",
       "      <td>213.511795</td>\n",
       "      <td>478.232635</td>\n",
       "      <td>112.253952</td>\n",
       "      <td>137.038147</td>\n",
       "      <td>2.070315</td>\n",
       "      <td>0.494080</td>\n",
       "      <td>142.950180</td>\n",
       "      <td>30.416140</td>\n",
       "      <td>44.190891</td>\n",
       "      <td>83.541595</td>\n",
       "      <td>49.555382</td>\n",
       "      <td>-0.627577</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.062744</td>\n",
       "      <td>-2.286133</td>\n",
       "      <td>2.054688</td>\n",
       "      <td>2.298340</td>\n",
       "      <td>117.006165</td>\n",
       "      <td>31.059309</td>\n",
       "      <td>-0.268066</td>\n",
       "      <td>1.588379</td>\n",
       "      <td>-0.845337</td>\n",
       "      <td>-0.288269</td>\n",
       "      <td>41.428207</td>\n",
       "      <td>29.480330</td>\n",
       "      <td>0.336069</td>\n",
       "      <td>0.343628</td>\n",
       "      <td>2.200073</td>\n",
       "      <td>1.271851</td>\n",
       "      <td>3.348877</td>\n",
       "      <td>1.856445</td>\n",
       "      <td>0.243652</td>\n",
       "      <td>2.900024</td>\n",
       "      <td>3.139508</td>\n",
       "      <td>2.950902</td>\n",
       "      <td>2.342957</td>\n",
       "      <td>2.586609</td>\n",
       "      <td>2.393834</td>\n",
       "      <td>0.557068</td>\n",
       "      <td>2.669638</td>\n",
       "      <td>2.720516</td>\n",
       "      <td>0.750675</td>\n",
       "      <td>0.520289</td>\n",
       "      <td>1.077357</td>\n",
       "      <td>3.190799</td>\n",
       "      <td>3.735679</td>\n",
       "      <td>2.401195</td>\n",
       "      <td>4.658582</td>\n",
       "      <td>1.938224</td>\n",
       "      <td>279.125275</td>\n",
       "      <td>450.637604</td>\n",
       "      <td>26.961533</td>\n",
       "      <td>118.906151</td>\n",
       "      <td>171.348328</td>\n",
       "      <td>111.728691</td>\n",
       "      <td>114.258469</td>\n",
       "      <td>213.866852</td>\n",
       "      <td>162.401306</td>\n",
       "      <td>230.252060</td>\n",
       "      <td>311.639374</td>\n",
       "      <td>77.039307</td>\n",
       "      <td>25.680614</td>\n",
       "      <td>43.205353</td>\n",
       "      <td>-1.365626</td>\n",
       "      <td>28.976540</td>\n",
       "      <td>85.813622</td>\n",
       "      <td>60.151260</td>\n",
       "      <td>68.254791</td>\n",
       "      <td>119.503845</td>\n",
       "      <td>0.536568</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-3.285156</td>\n",
       "      <td>1.340576</td>\n",
       "      <td>1.058594</td>\n",
       "      <td>-1.453125</td>\n",
       "      <td>94.150909</td>\n",
       "      <td>43.804207</td>\n",
       "      <td>1.760498</td>\n",
       "      <td>-0.124847</td>\n",
       "      <td>-1.573730</td>\n",
       "      <td>-1.902832</td>\n",
       "      <td>32.677067</td>\n",
       "      <td>25.447952</td>\n",
       "      <td>0.201630</td>\n",
       "      <td>2.732788</td>\n",
       "      <td>0.847443</td>\n",
       "      <td>1.790115</td>\n",
       "      <td>4.625732</td>\n",
       "      <td>1.885345</td>\n",
       "      <td>2.511719</td>\n",
       "      <td>2.632324</td>\n",
       "      <td>0.120605</td>\n",
       "      <td>2.218511</td>\n",
       "      <td>2.961426</td>\n",
       "      <td>0.449707</td>\n",
       "      <td>2.547612</td>\n",
       "      <td>0.329102</td>\n",
       "      <td>2.776226</td>\n",
       "      <td>2.362412</td>\n",
       "      <td>1.936361</td>\n",
       "      <td>2.080262</td>\n",
       "      <td>1.751161</td>\n",
       "      <td>5.691024</td>\n",
       "      <td>0.436898</td>\n",
       "      <td>4.331004</td>\n",
       "      <td>1.532874</td>\n",
       "      <td>1.913854</td>\n",
       "      <td>249.244843</td>\n",
       "      <td>990.080811</td>\n",
       "      <td>49.173138</td>\n",
       "      <td>139.067184</td>\n",
       "      <td>695.389648</td>\n",
       "      <td>17.703999</td>\n",
       "      <td>248.063904</td>\n",
       "      <td>55.661476</td>\n",
       "      <td>22.814188</td>\n",
       "      <td>162.431519</td>\n",
       "      <td>654.367798</td>\n",
       "      <td>63.491798</td>\n",
       "      <td>71.896461</td>\n",
       "      <td>56.491291</td>\n",
       "      <td>2.629192</td>\n",
       "      <td>109.982231</td>\n",
       "      <td>56.677876</td>\n",
       "      <td>53.164700</td>\n",
       "      <td>57.352097</td>\n",
       "      <td>62.735786</td>\n",
       "      <td>0.909676</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-3.747070</td>\n",
       "      <td>1.243652</td>\n",
       "      <td>2.889648</td>\n",
       "      <td>-1.727295</td>\n",
       "      <td>64.973938</td>\n",
       "      <td>51.909889</td>\n",
       "      <td>1.294189</td>\n",
       "      <td>-0.332458</td>\n",
       "      <td>-0.136871</td>\n",
       "      <td>-2.426758</td>\n",
       "      <td>64.003830</td>\n",
       "      <td>21.561178</td>\n",
       "      <td>0.409156</td>\n",
       "      <td>2.545898</td>\n",
       "      <td>0.919250</td>\n",
       "      <td>1.732574</td>\n",
       "      <td>4.990723</td>\n",
       "      <td>1.626648</td>\n",
       "      <td>1.666242</td>\n",
       "      <td>3.026520</td>\n",
       "      <td>1.590424</td>\n",
       "      <td>2.545911</td>\n",
       "      <td>0.966779</td>\n",
       "      <td>0.699463</td>\n",
       "      <td>0.256025</td>\n",
       "      <td>2.289886</td>\n",
       "      <td>2.941757</td>\n",
       "      <td>2.231003</td>\n",
       "      <td>2.564538</td>\n",
       "      <td>2.249629</td>\n",
       "      <td>1.743670</td>\n",
       "      <td>5.879976</td>\n",
       "      <td>1.591226</td>\n",
       "      <td>6.318524</td>\n",
       "      <td>1.724347</td>\n",
       "      <td>2.808837</td>\n",
       "      <td>287.629669</td>\n",
       "      <td>1100.145630</td>\n",
       "      <td>26.790192</td>\n",
       "      <td>156.058716</td>\n",
       "      <td>807.220642</td>\n",
       "      <td>83.376167</td>\n",
       "      <td>202.906906</td>\n",
       "      <td>63.418484</td>\n",
       "      <td>61.674225</td>\n",
       "      <td>180.367538</td>\n",
       "      <td>705.582947</td>\n",
       "      <td>95.441437</td>\n",
       "      <td>90.914345</td>\n",
       "      <td>44.773403</td>\n",
       "      <td>2.112758</td>\n",
       "      <td>114.652504</td>\n",
       "      <td>45.826389</td>\n",
       "      <td>39.664787</td>\n",
       "      <td>52.377136</td>\n",
       "      <td>28.886316</td>\n",
       "      <td>0.827113</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>226363</th>\n",
       "      <td>-0.581909</td>\n",
       "      <td>2.088379</td>\n",
       "      <td>-2.884766</td>\n",
       "      <td>0.785156</td>\n",
       "      <td>146.781616</td>\n",
       "      <td>83.524757</td>\n",
       "      <td>0.532593</td>\n",
       "      <td>-2.047852</td>\n",
       "      <td>-2.560547</td>\n",
       "      <td>-3.018066</td>\n",
       "      <td>93.853348</td>\n",
       "      <td>13.915641</td>\n",
       "      <td>0.106529</td>\n",
       "      <td>-0.220642</td>\n",
       "      <td>-2.801086</td>\n",
       "      <td>1.510864</td>\n",
       "      <td>2.670288</td>\n",
       "      <td>2.580444</td>\n",
       "      <td>2.613263</td>\n",
       "      <td>0.324219</td>\n",
       "      <td>2.937482</td>\n",
       "      <td>0.837788</td>\n",
       "      <td>0.133301</td>\n",
       "      <td>2.479963</td>\n",
       "      <td>0.380269</td>\n",
       "      <td>0.457520</td>\n",
       "      <td>0.266478</td>\n",
       "      <td>0.780047</td>\n",
       "      <td>2.892970</td>\n",
       "      <td>2.950711</td>\n",
       "      <td>2.493191</td>\n",
       "      <td>1.160703</td>\n",
       "      <td>3.689743</td>\n",
       "      <td>1.471991</td>\n",
       "      <td>5.618977</td>\n",
       "      <td>2.620690</td>\n",
       "      <td>554.081421</td>\n",
       "      <td>608.137512</td>\n",
       "      <td>129.570892</td>\n",
       "      <td>291.211121</td>\n",
       "      <td>143.886078</td>\n",
       "      <td>233.539810</td>\n",
       "      <td>74.311653</td>\n",
       "      <td>273.705353</td>\n",
       "      <td>191.616348</td>\n",
       "      <td>294.558411</td>\n",
       "      <td>447.349274</td>\n",
       "      <td>122.470482</td>\n",
       "      <td>221.119720</td>\n",
       "      <td>81.077499</td>\n",
       "      <td>0.771927</td>\n",
       "      <td>229.077423</td>\n",
       "      <td>131.637543</td>\n",
       "      <td>131.433853</td>\n",
       "      <td>106.515289</td>\n",
       "      <td>38.494846</td>\n",
       "      <td>-0.004014</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>226364</th>\n",
       "      <td>3.178223</td>\n",
       "      <td>0.439575</td>\n",
       "      <td>-1.953125</td>\n",
       "      <td>0.997437</td>\n",
       "      <td>141.700531</td>\n",
       "      <td>129.908966</td>\n",
       "      <td>-0.267517</td>\n",
       "      <td>-0.949951</td>\n",
       "      <td>2.213867</td>\n",
       "      <td>-1.661377</td>\n",
       "      <td>181.813049</td>\n",
       "      <td>32.300430</td>\n",
       "      <td>0.319024</td>\n",
       "      <td>-2.076416</td>\n",
       "      <td>-2.758850</td>\n",
       "      <td>2.417633</td>\n",
       "      <td>2.738647</td>\n",
       "      <td>0.682434</td>\n",
       "      <td>2.950562</td>\n",
       "      <td>2.116193</td>\n",
       "      <td>1.216431</td>\n",
       "      <td>3.117170</td>\n",
       "      <td>0.291748</td>\n",
       "      <td>2.658813</td>\n",
       "      <td>0.758074</td>\n",
       "      <td>2.407941</td>\n",
       "      <td>1.980002</td>\n",
       "      <td>3.029824</td>\n",
       "      <td>2.295703</td>\n",
       "      <td>2.159512</td>\n",
       "      <td>1.715732</td>\n",
       "      <td>5.407120</td>\n",
       "      <td>1.407012</td>\n",
       "      <td>4.138470</td>\n",
       "      <td>3.000012</td>\n",
       "      <td>2.502778</td>\n",
       "      <td>696.124023</td>\n",
       "      <td>1237.684204</td>\n",
       "      <td>139.694855</td>\n",
       "      <td>302.617401</td>\n",
       "      <td>914.378723</td>\n",
       "      <td>210.228683</td>\n",
       "      <td>525.152344</td>\n",
       "      <td>160.989212</td>\n",
       "      <td>94.727547</td>\n",
       "      <td>202.368317</td>\n",
       "      <td>572.939758</td>\n",
       "      <td>152.681671</td>\n",
       "      <td>237.582932</td>\n",
       "      <td>85.900383</td>\n",
       "      <td>0.054355</td>\n",
       "      <td>240.316574</td>\n",
       "      <td>102.535553</td>\n",
       "      <td>99.807480</td>\n",
       "      <td>159.297455</td>\n",
       "      <td>119.385689</td>\n",
       "      <td>-0.375145</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>226365</th>\n",
       "      <td>-0.798706</td>\n",
       "      <td>2.795410</td>\n",
       "      <td>-2.399414</td>\n",
       "      <td>1.070801</td>\n",
       "      <td>344.595642</td>\n",
       "      <td>94.871841</td>\n",
       "      <td>-1.093506</td>\n",
       "      <td>1.201660</td>\n",
       "      <td>1.215332</td>\n",
       "      <td>-0.345398</td>\n",
       "      <td>60.108978</td>\n",
       "      <td>33.383640</td>\n",
       "      <td>0.061380</td>\n",
       "      <td>-2.091858</td>\n",
       "      <td>0.203308</td>\n",
       "      <td>0.944275</td>\n",
       "      <td>3.594116</td>\n",
       "      <td>2.295166</td>\n",
       "      <td>2.812970</td>\n",
       "      <td>2.668439</td>\n",
       "      <td>0.144531</td>\n",
       "      <td>2.548845</td>\n",
       "      <td>2.054016</td>\n",
       "      <td>1.416199</td>\n",
       "      <td>2.173611</td>\n",
       "      <td>1.560730</td>\n",
       "      <td>3.110162</td>\n",
       "      <td>3.053429</td>\n",
       "      <td>0.064698</td>\n",
       "      <td>0.439886</td>\n",
       "      <td>1.120844</td>\n",
       "      <td>3.626747</td>\n",
       "      <td>3.891601</td>\n",
       "      <td>2.867132</td>\n",
       "      <td>2.132055</td>\n",
       "      <td>2.775548</td>\n",
       "      <td>719.251831</td>\n",
       "      <td>1317.528442</td>\n",
       "      <td>22.047525</td>\n",
       "      <td>211.460892</td>\n",
       "      <td>285.072723</td>\n",
       "      <td>517.502625</td>\n",
       "      <td>313.609680</td>\n",
       "      <td>124.902122</td>\n",
       "      <td>409.386047</td>\n",
       "      <td>442.162292</td>\n",
       "      <td>1120.403442</td>\n",
       "      <td>141.703918</td>\n",
       "      <td>39.273769</td>\n",
       "      <td>71.307526</td>\n",
       "      <td>0.775446</td>\n",
       "      <td>110.606529</td>\n",
       "      <td>111.363823</td>\n",
       "      <td>134.771912</td>\n",
       "      <td>69.050385</td>\n",
       "      <td>203.726746</td>\n",
       "      <td>-0.183842</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>226366</th>\n",
       "      <td>-0.052963</td>\n",
       "      <td>-2.614258</td>\n",
       "      <td>2.937012</td>\n",
       "      <td>0.865845</td>\n",
       "      <td>118.263573</td>\n",
       "      <td>56.826637</td>\n",
       "      <td>0.295959</td>\n",
       "      <td>0.355896</td>\n",
       "      <td>-1.638672</td>\n",
       "      <td>0.933960</td>\n",
       "      <td>85.634995</td>\n",
       "      <td>34.084076</td>\n",
       "      <td>0.434310</td>\n",
       "      <td>1.629570</td>\n",
       "      <td>1.689507</td>\n",
       "      <td>1.659538</td>\n",
       "      <td>2.561295</td>\n",
       "      <td>0.059937</td>\n",
       "      <td>2.071167</td>\n",
       "      <td>1.707502</td>\n",
       "      <td>2.504517</td>\n",
       "      <td>2.208736</td>\n",
       "      <td>2.003052</td>\n",
       "      <td>0.068115</td>\n",
       "      <td>1.501818</td>\n",
       "      <td>2.572632</td>\n",
       "      <td>2.019569</td>\n",
       "      <td>2.520803</td>\n",
       "      <td>1.133983</td>\n",
       "      <td>1.446050</td>\n",
       "      <td>1.126581</td>\n",
       "      <td>4.588968</td>\n",
       "      <td>3.839527</td>\n",
       "      <td>2.044354</td>\n",
       "      <td>2.970935</td>\n",
       "      <td>2.573330</td>\n",
       "      <td>524.149719</td>\n",
       "      <td>522.848389</td>\n",
       "      <td>92.688339</td>\n",
       "      <td>198.505219</td>\n",
       "      <td>156.445435</td>\n",
       "      <td>312.296875</td>\n",
       "      <td>110.823631</td>\n",
       "      <td>184.593262</td>\n",
       "      <td>190.723969</td>\n",
       "      <td>249.475555</td>\n",
       "      <td>307.628876</td>\n",
       "      <td>103.759346</td>\n",
       "      <td>146.262146</td>\n",
       "      <td>74.467667</td>\n",
       "      <td>-0.192622</td>\n",
       "      <td>149.901611</td>\n",
       "      <td>78.535515</td>\n",
       "      <td>142.655334</td>\n",
       "      <td>59.809212</td>\n",
       "      <td>176.484467</td>\n",
       "      <td>0.313027</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>226367</th>\n",
       "      <td>0.238708</td>\n",
       "      <td>-3.432617</td>\n",
       "      <td>-0.546875</td>\n",
       "      <td>2.240234</td>\n",
       "      <td>128.955780</td>\n",
       "      <td>89.325569</td>\n",
       "      <td>-1.567871</td>\n",
       "      <td>-1.217041</td>\n",
       "      <td>-2.229980</td>\n",
       "      <td>0.790649</td>\n",
       "      <td>149.046494</td>\n",
       "      <td>40.551991</td>\n",
       "      <td>0.524708</td>\n",
       "      <td>0.029083</td>\n",
       "      <td>0.379913</td>\n",
       "      <td>0.204498</td>\n",
       "      <td>3.671326</td>\n",
       "      <td>0.350830</td>\n",
       "      <td>2.787109</td>\n",
       "      <td>1.683105</td>\n",
       "      <td>1.812971</td>\n",
       "      <td>2.284513</td>\n",
       "      <td>1.337524</td>\n",
       "      <td>1.449585</td>\n",
       "      <td>0.736117</td>\n",
       "      <td>3.020630</td>\n",
       "      <td>1.638157</td>\n",
       "      <td>2.239564</td>\n",
       "      <td>2.214263</td>\n",
       "      <td>2.169314</td>\n",
       "      <td>1.093242</td>\n",
       "      <td>2.469124</td>\n",
       "      <td>4.843563</td>\n",
       "      <td>1.976911</td>\n",
       "      <td>2.647655</td>\n",
       "      <td>3.040935</td>\n",
       "      <td>575.017883</td>\n",
       "      <td>879.581238</td>\n",
       "      <td>159.328888</td>\n",
       "      <td>374.349854</td>\n",
       "      <td>353.250549</td>\n",
       "      <td>307.601105</td>\n",
       "      <td>146.074265</td>\n",
       "      <td>181.446838</td>\n",
       "      <td>180.231384</td>\n",
       "      <td>242.150696</td>\n",
       "      <td>689.696777</td>\n",
       "      <td>157.606415</td>\n",
       "      <td>228.894653</td>\n",
       "      <td>70.193581</td>\n",
       "      <td>1.883891</td>\n",
       "      <td>320.534912</td>\n",
       "      <td>112.050957</td>\n",
       "      <td>112.420204</td>\n",
       "      <td>108.900810</td>\n",
       "      <td>99.031250</td>\n",
       "      <td>-1.492146</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>226368 rows × 64 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        CleanJet_eta_1  CleanJet_eta_2  CleanJet_phi_1  CleanJet_phi_2  \\\n",
       "0            -2.805664        0.871460       -1.175293       -3.091797   \n",
       "1            -1.510010        3.178711       -0.928223        0.462524   \n",
       "2             1.062744       -2.286133        2.054688        2.298340   \n",
       "3            -3.285156        1.340576        1.058594       -1.453125   \n",
       "4            -3.747070        1.243652        2.889648       -1.727295   \n",
       "...                ...             ...             ...             ...   \n",
       "226363       -0.581909        2.088379       -2.884766        0.785156   \n",
       "226364        3.178223        0.439575       -1.953125        0.997437   \n",
       "226365       -0.798706        2.795410       -2.399414        1.070801   \n",
       "226366       -0.052963       -2.614258        2.937012        0.865845   \n",
       "226367        0.238708       -3.432617       -0.546875        2.240234   \n",
       "\n",
       "        CleanJet_pt_1  CleanJet_pt_2  Lepton_eta_1  Lepton_eta_2  \\\n",
       "0          115.945938     105.480812     -0.222351      1.390381   \n",
       "1           48.520248      43.383701     -0.576294     -0.817017   \n",
       "2          117.006165      31.059309     -0.268066      1.588379   \n",
       "3           94.150909      43.804207      1.760498     -0.124847   \n",
       "4           64.973938      51.909889      1.294189     -0.332458   \n",
       "...               ...            ...           ...           ...   \n",
       "226363     146.781616      83.524757      0.532593     -2.047852   \n",
       "226364     141.700531     129.908966     -0.267517     -0.949951   \n",
       "226365     344.595642      94.871841     -1.093506      1.201660   \n",
       "226366     118.263573      56.826637      0.295959      0.355896   \n",
       "226367     128.955780      89.325569     -1.567871     -1.217041   \n",
       "\n",
       "        Lepton_phi_1  Lepton_phi_2  Lepton_pt_1  Lepton_pt_2       Rpt  \\\n",
       "0           1.880859      2.354492    60.912807    42.632568  0.212335   \n",
       "1          -3.005371      0.551392   109.715820    29.537357  1.539541   \n",
       "2          -0.845337     -0.288269    41.428207    29.480330  0.336069   \n",
       "3          -1.573730     -1.902832    32.677067    25.447952  0.201630   \n",
       "4          -0.136871     -2.426758    64.003830    21.561178  0.409156   \n",
       "...              ...           ...          ...          ...       ...   \n",
       "226363     -2.560547     -3.018066    93.853348    13.915641  0.106529   \n",
       "226364      2.213867     -1.661377   181.813049    32.300430  0.319024   \n",
       "226365      1.215332     -0.345398    60.108978    33.383640  0.061380   \n",
       "226366     -1.638672      0.933960    85.634995    34.084076  0.434310   \n",
       "226367     -2.229980      0.790649   149.046494    40.551991  0.524708   \n",
       "\n",
       "         Zepp_l1   Zepp_l2   Zepp_ll    detajj    detall    dphijj  \\\n",
       "0       0.744751  2.357483  1.551117  3.677124  1.612732  1.916504   \n",
       "1      -1.410645 -1.651367  1.531006  4.688721  0.240723  1.390747   \n",
       "2       0.343628  2.200073  1.271851  3.348877  1.856445  0.243652   \n",
       "3       2.732788  0.847443  1.790115  4.625732  1.885345  2.511719   \n",
       "4       2.545898  0.919250  1.732574  4.990723  1.626648  1.666242   \n",
       "...          ...       ...       ...       ...       ...       ...   \n",
       "226363 -0.220642 -2.801086  1.510864  2.670288  2.580444  2.613263   \n",
       "226364 -2.076416 -2.758850  2.417633  2.738647  0.682434  2.950562   \n",
       "226365 -2.091858  0.203308  0.944275  3.594116  2.295166  2.812970   \n",
       "226366  1.629570  1.689507  1.659538  2.561295  0.059937  2.071167   \n",
       "226367  0.029083  0.379913  0.204498  3.671326  0.350830  2.787109   \n",
       "\n",
       "        dphilep1jet1  dphilep1jet2  dphilep1jj  dphilep2jet1  dphilep2jet2  \\\n",
       "0           3.056152      1.310529    2.335937      2.753400      0.836896   \n",
       "1           2.077148      2.815290    2.725920      1.479614      0.088867   \n",
       "2           2.900024      3.139508    2.950902      2.342957      2.586609   \n",
       "3           2.632324      0.120605    2.218511      2.961426      0.449707   \n",
       "4           3.026520      1.590424    2.545911      0.966779      0.699463   \n",
       "...              ...           ...         ...           ...           ...   \n",
       "226363      0.324219      2.937482    0.837788      0.133301      2.479963   \n",
       "226364      2.116193      1.216431    3.117170      0.291748      2.658813   \n",
       "226365      2.668439      0.144531    2.548845      2.054016      1.416199   \n",
       "226366      1.707502      2.504517    2.208736      2.003052      0.068115   \n",
       "226367      1.683105      1.812971    2.284513      1.337524      1.449585   \n",
       "\n",
       "        dphilep2jj    dphill  dphilljet  dphilljetjet  dphillmet  dphilmet1  \\\n",
       "0         1.862304  0.473633   3.032799      2.141702   1.970499   1.776264   \n",
       "1         0.830842  2.726423   2.220245      2.869017   2.640638   2.783735   \n",
       "2         2.393834  0.557068   2.669638      2.720516   0.750675   0.520289   \n",
       "3         2.547612  0.329102   2.776226      2.362412   1.936361   2.080262   \n",
       "4         0.256025  2.289886   2.941757      2.231003   2.564538   2.249629   \n",
       "...            ...       ...        ...           ...        ...        ...   \n",
       "226363    0.380269  0.457520   0.266478      0.780047   2.892970   2.950711   \n",
       "226364    0.758074  2.407941   1.980002      3.029824   2.295703   2.159512   \n",
       "226365    2.173611  1.560730   3.110162      3.053429   0.064698   0.439886   \n",
       "226366    1.501818  2.572632   2.019569      2.520803   1.133983   1.446050   \n",
       "226367    0.736117  3.020630   1.638157      2.239564   2.214263   2.169314   \n",
       "\n",
       "        dphilmet2   dr_l1j1   dr_l1j2   dr_l2j1   dr_l2j2      drll  \\\n",
       "0        2.249897  4.001696  5.091536  5.483263  5.470954  1.680843   \n",
       "1        0.057312  2.277360  5.111395  1.633860  3.996716  2.737029   \n",
       "2        1.077357  3.190799  3.735679  2.401195  4.658582  1.938224   \n",
       "3        1.751161  5.691024  0.436898  4.331004  1.532874  1.913854   \n",
       "4        1.743670  5.879976  1.591226  6.318524  1.724347  2.808837   \n",
       "...           ...       ...       ...       ...       ...       ...   \n",
       "226363   2.493191  1.160703  3.689743  1.471991  5.618977  2.620690   \n",
       "226364   1.715732  5.407120  1.407012  4.138470  3.000012  2.502778   \n",
       "226365   1.120844  3.626747  3.891601  2.867132  2.132055  2.775548   \n",
       "226366   1.126581  4.588968  3.839527  2.044354  2.970935  2.573330   \n",
       "226367   1.093242  2.469124  4.843563  1.976911  2.647655  3.040935   \n",
       "\n",
       "                ht      m2ljj30         mT2         mTi      m_l1j1  \\\n",
       "0       419.124969   985.871765   58.271770  158.066528  328.959351   \n",
       "1       309.260803   738.418823   35.363697  184.284683  144.914749   \n",
       "2       279.125275   450.637604   26.961533  118.906151  171.348328   \n",
       "3       249.244843   990.080811   49.173138  139.067184  695.389648   \n",
       "4       287.629669  1100.145630   26.790192  156.058716  807.220642   \n",
       "...            ...          ...         ...         ...         ...   \n",
       "226363  554.081421   608.137512  129.570892  291.211121  143.886078   \n",
       "226364  696.124023  1237.684204  139.694855  302.617401  914.378723   \n",
       "226365  719.251831  1317.528442   22.047525  211.460892  285.072723   \n",
       "226366  524.149719   522.848389   92.688339  198.505219  156.445435   \n",
       "226367  575.017883   879.581238  159.328888  374.349854  353.250549   \n",
       "\n",
       "            m_l1j2      m_l2j1      m_l2j2       mcoll     mcollWW  \\\n",
       "0       136.063660  581.079773   68.985100   54.399845  170.509857   \n",
       "1       461.100403   58.736408  259.231140  139.885406  213.511795   \n",
       "2       111.728691  114.258469  213.866852  162.401306  230.252060   \n",
       "3        17.703999  248.063904   55.661476   22.814188  162.431519   \n",
       "4        83.376167  202.906906   63.418484   61.674225  180.367538   \n",
       "...            ...         ...         ...         ...         ...   \n",
       "226363  233.539810   74.311653  273.705353  191.616348  294.558411   \n",
       "226364  210.228683  525.152344  160.989212   94.727547  202.368317   \n",
       "226365  517.502625  313.609680  124.902122  409.386047  442.162292   \n",
       "226366  312.296875  110.823631  184.593262  190.723969  249.475555   \n",
       "226367  307.601105  146.074265  181.446838  180.231384  242.150696   \n",
       "\n",
       "                mjj         mll        mtw1       mtw2  phi_miss   proxyW_l1  \\\n",
       "0        703.374695   94.460876   78.216949  76.092033  0.104595   79.019096   \n",
       "1        478.232635  112.253952  137.038147   2.070315  0.494080  142.950180   \n",
       "2        311.639374   77.039307   25.680614  43.205353 -1.365626   28.976540   \n",
       "3        654.367798   63.491798   71.896461  56.491291  2.629192  109.982231   \n",
       "4        705.582947   95.441437   90.914345  44.773403  2.112758  114.652504   \n",
       "...             ...         ...         ...        ...       ...         ...   \n",
       "226363   447.349274  122.470482  221.119720  81.077499  0.771927  229.077423   \n",
       "226364   572.939758  152.681671  237.582932  85.900383  0.054355  240.316574   \n",
       "226365  1120.403442  141.703918   39.273769  71.307526  0.775446  110.606529   \n",
       "226366   307.628876  103.759346  146.262146  74.467667 -0.192622  149.901611   \n",
       "226367   689.696777  157.606415  228.894653  70.193581  1.883891  320.534912   \n",
       "\n",
       "         proxyW_l2     pt_miss        ptll      recoil       yll  isLL  isMIX  \\\n",
       "0        99.086594   41.708782  100.746727   92.839317  0.465608   0.0    0.0   \n",
       "1        30.416140   44.190891   83.541595   49.555382 -0.627577   0.0    0.0   \n",
       "2        85.813622   60.151260   68.254791  119.503845  0.536568   0.0    0.0   \n",
       "3        56.677876   53.164700   57.352097   62.735786  0.909676   0.0    0.0   \n",
       "4        45.826389   39.664787   52.377136   28.886316  0.827113   0.0    0.0   \n",
       "...            ...         ...         ...         ...       ...   ...    ...   \n",
       "226363  131.637543  131.433853  106.515289   38.494846 -0.004014   0.0    0.0   \n",
       "226364  102.535553   99.807480  159.297455  119.385689 -0.375145   0.0    0.0   \n",
       "226365  111.363823  134.771912   69.050385  203.726746 -0.183842   0.0    0.0   \n",
       "226366   78.535515  142.655334   59.809212  176.484467  0.313027   0.0    0.0   \n",
       "226367  112.050957  112.420204  108.900810   99.031250 -1.492146   0.0    0.0   \n",
       "\n",
       "        isTT  isBKG  isVBS  notLL  notTT  \n",
       "0        0.0    1.0    0.0    0.0    0.0  \n",
       "1        0.0    1.0    0.0    0.0    0.0  \n",
       "2        0.0    1.0    0.0    0.0    0.0  \n",
       "3        0.0    1.0    0.0    0.0    0.0  \n",
       "4        0.0    1.0    0.0    0.0    0.0  \n",
       "...      ...    ...    ...    ...    ...  \n",
       "226363   0.0    1.0    0.0    0.0    0.0  \n",
       "226364   0.0    1.0    0.0    0.0    0.0  \n",
       "226365   0.0    1.0    0.0    0.0    0.0  \n",
       "226366   0.0    1.0    0.0    0.0    0.0  \n",
       "226367   0.0    1.0    0.0    0.0    0.0  \n",
       "\n",
       "[226368 rows x 64 columns]"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "from sklearn.utils import shuffle\n",
    "\n",
    "df = pd.read_pickle('https://pandora.infn.it/public/c7fce9/dl/dataset_pol_vbs_full2018v9_VBS_WW_2j_RDF_DNN_SigBkg.pkl')\n",
    "pd.set_option('display.max_columns', None)\n",
    "\n",
    "df['notLL'] = df['isTT'] + df['isMIX']\n",
    "df['notTT'] = df['isLL'] + df['isMIX']\n",
    "\n",
    "df_SvB = df[((df['isTT'] + df['isMIX'] + df['isLL']) == 0) & ((df['isVBS'] + df['isBKG']) == 1)]\n",
    "df_SvB"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "b1c30b9d-7706-4fda-8bce-99ffc4a93c08",
   "metadata": {},
   "outputs": [],
   "source": [
    "import math\n",
    "import numpy as np\n",
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "bc26e027-8929-4c05-8995-bc477450a479",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.7.0-rc1\n"
     ]
    }
   ],
   "source": [
    "import tqdm\n",
    "from tensorflow.keras.models import Sequential, Model\n",
    "from tensorflow.keras.layers import Input, Activation, Dense, Dropout, InputLayer\n",
    "from tensorflow.keras.constraints import max_norm\n",
    "from tensorflow.keras import initializers\n",
    "from tensorflow.keras.models import load_model\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from sklearn.utils import shuffle\n",
    "import copy\n",
    "print(tf.__version__)\n",
    "from tensorflow.python.client import device_lib\n",
    "tf.config.run_functions_eagerly(True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "52abbb0d-f65b-408f-9465-36991164434b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "57\n",
      "[[-2.80566406e+00  8.71459961e-01 -1.17529297e+00 ...  1.00746727e+02\n",
      "   9.28393173e+01  4.65607733e-01]\n",
      " [-1.51000977e+00  3.17871094e+00 -9.28222656e-01 ...  8.35415955e+01\n",
      "   4.95553818e+01 -6.27577305e-01]\n",
      " [ 1.06274414e+00 -2.28613281e+00  2.05468750e+00 ...  6.82547913e+01\n",
      "   1.19503845e+02  5.36567748e-01]\n",
      " ...\n",
      " [-7.98706055e-01  2.79541016e+00 -2.39941406e+00 ...  6.90503845e+01\n",
      "   2.03726746e+02 -1.83842495e-01]\n",
      " [-5.29632568e-02 -2.61425781e+00  2.93701172e+00 ...  5.98092117e+01\n",
      "   1.76484467e+02  3.13026577e-01]\n",
      " [ 2.38708496e-01 -3.43261719e+00 -5.46875000e-01 ...  1.08900810e+02\n",
      "   9.90312500e+01 -1.49214625e+00]] [[0. 1.]\n",
      " [0. 1.]\n",
      " [0. 1.]\n",
      " ...\n",
      " [0. 1.]\n",
      " [0. 1.]\n",
      " [0. 1.]]\n"
     ]
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.pyplot import cm\n",
    "import numpy as np\n",
    "import matplotlib as mpl\n",
    "%matplotlib inline\n",
    "mpl.rcParams['figure.dpi'] = 150\n",
    "\n",
    "nclasses = 2\n",
    "features = df.columns[:len(df.columns)-7]\n",
    "\n",
    "NDIM = len(features)\n",
    "print(NDIM)\n",
    "for col in df.columns:\n",
    "    df[col]    = np.hstack(df[col])\n",
    "'''\n",
    "df.replace([np.inf, -np.inf], np.nan, inplace=True)\n",
    "df.dropna(inplace=True)\n",
    "df = shuffle(df)\n",
    "'''\n",
    "# Perform the splitting and define training and validation datasets\n",
    "msk    = np.random.rand(len(df_SvB)) < 0.8\n",
    "\n",
    "df_train    = df_SvB[msk]\n",
    "df_val      = df_SvB[~msk]\n",
    "\n",
    "X    = df_train.values[:,0:NDIM]\n",
    "Y    = df_train[['isVBS', 'isBKG']].values\n",
    "X_val = df_val.values[:,0:NDIM]\n",
    "Y_val = df_val[['isVBS', 'isBKG']].values\n",
    "\n",
    "print(X,Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "2bd8aa6e-8ec9-462c-937d-0430fcdec8da",
   "metadata": {},
   "outputs": [],
   "source": [
    "class SimpleNeuralNetwork ( tf.Module ):\n",
    "    def __init__ (self, nEpochs, learning_rate, N_NODES, n_layers, n_features, n_outputsC=nclasses):\n",
    "        self.learning_rate = learning_rate\n",
    "        self.optimizer  = tf.optimizers.Adam (self.learning_rate)\n",
    "        self.nEpochs = nEpochs\n",
    "        self.N_NODES = N_NODES\n",
    "        self.n_layers = n_layers\n",
    "        self.n_features = n_features\n",
    "        self.n_outputsC = n_outputsC\n",
    "        self.weights = self.build (self.n_features, self.N_NODES)\n",
    "        \n",
    "                                \n",
    "    # Define the structure of the model\n",
    "    def build (self, n_input, N_NODES):\n",
    "        # initializer = initializers.Ones()\n",
    "\n",
    "        # Classifier model\n",
    "        self.model1 = Sequential()\n",
    "        self.model1.add(Dense (self.N_NODES, activation = 'relu', input_dim  = n_input))\n",
    "        for i in range(self.n_layers):\n",
    "            self.model1.add(Dense (self.N_NODES, activation = 'relu'))\n",
    "        if self.n_outputsC > 1:\n",
    "            self.model1.add(Dense (self.n_outputsC, activation = 'softmax',input_dim = self.N_NODES))\n",
    "        else:\n",
    "            self.model1.add(Dense (self.n_outputsC, activation = 'sigmoid',input_dim = self.N_NODES))\n",
    "        \n",
    "        return self.model1.weights\n",
    "     \n",
    "    # Performs the epochs loop and the actual training.\n",
    "    # Monitors the training and validation loss functions, both for the classifier and the adversary.\n",
    "    # Returns the classifier categorical accuracy.\n",
    "    def fit (self, X, Y, X_val, Y_val, show_loss = False):\n",
    "        losses = []\n",
    "        losses_val = []\n",
    "\n",
    "        self.means = np.mean ( X, axis = 0)\n",
    "        self.sigmas = np.std ( X, axis = 0)\n",
    "\n",
    "        for iEpoch in tqdm.tqdm(range(self.nEpochs)):\n",
    "                l, l_val = self._train (X, Y, X_val, Y_val)\n",
    "                losses.append ( l )\n",
    "                losses_val.append ( l_val )\n",
    "\n",
    "        losses = np.array(losses)               \n",
    "        losses_val = np.array(losses_val)\n",
    "               \n",
    "        plt.plot (losses, color = \"c\", label='Training set')\n",
    "        plt.plot (losses_val, color ='tab:blue', label = \"Validation set\")\n",
    "        plt.xlabel (\"Epoch\"); plt.ylabel (\"Loss\")\n",
    "        plt.legend(frameon=False)\n",
    "        plt.show()\n",
    "         \n",
    "        ca = tf.keras.metrics.CategoricalAccuracy()\n",
    "        if self.n_outputsC <= 1:\n",
    "            ca = tf.keras.metrics.BinaryAccuracy()\n",
    "        ca.update_state(Y, self.predict_proba(X))\n",
    "        \n",
    "        return ca.result().numpy()\n",
    "\n",
    "    def save_weights(self, model_name):\n",
    "        self.model1.save_weights(model_name+'_weights_1')\n",
    "    \n",
    "    def load_weights(self, model_name):\n",
    "        self.model1.load_weights(model_name+'_weights_1')\n",
    "        \n",
    "    def save_model(self, model_name):\n",
    "        self.model1.save(\"saved_models/\"+model_name+\"_1\")\n",
    "\n",
    "    def reset_optimizers(self):\n",
    "        self.optimizer  = tf.optimizers.Adam (self.learning_rate)\n",
    "        \n",
    "    def set_epochs(self, epochs):\n",
    "        self.nEpochs = epochs\n",
    "        \n",
    "    # Applies a pre-processing to the input features and returns the classifier representation.\n",
    "    #@tf.function\n",
    "    def predict_proba (self, X):\n",
    "        ppX = (X - self.means)/self.sigmas\n",
    "        return  tf.clip_by_value ( self.model1 (ppX) , 1e-7, 1. - 1e-7 )\n",
    "\n",
    "\n",
    "    @tf.function\n",
    "    def _train (self, X, Y, X_val, Y_val):\n",
    "        Y_true = tf.cast (Y, tf.float32)\n",
    "        Y_true_val = tf.cast (Y_val, tf.float32)\n",
    "\n",
    "        with tf.GradientTape() as gt:\n",
    "            #gt.watch ( self.weightsC )\n",
    "            Y_hat = self.predict_proba (X)\n",
    "            Y_hat_val = self.predict_proba (X_val) #N3(N1(x)) validation set\n",
    "            \n",
    "            ## Training set\n",
    "            # Use the categorical cross-entropy as loss function for the classifier\n",
    "            cce = 0.\n",
    "            cce_val = 0.\n",
    "            bce = 0.\n",
    "            bce_val = 0.\n",
    "            loss = 0.\n",
    "            loss_val = 0.\n",
    "            \n",
    "            if self.n_outputsC > 1:\n",
    "                cce = tf.keras.losses.CategoricalCrossentropy()\n",
    "                loss = tf.reduce_mean ( cce( Y_true, Y_hat ) )\n",
    "            \n",
    "                ## Validation set\n",
    "                cce_val = tf.keras.losses.CategoricalCrossentropy()\n",
    "                loss_val = tf.reduce_mean (cce_val( Y_true_val, Y_hat_val ) )\n",
    "            else:\n",
    "                bce = tf.keras.losses.BinaryCrossentropy()\n",
    "                loss = tf.reduce_mean ( bce( Y_true, Y_hat ) )\n",
    "            \n",
    "                ## Validation set\n",
    "                bce_val = tf.keras.losses.BinaryCrossentropy()\n",
    "                loss_val = tf.reduce_mean (bce_val( Y_true_val, Y_hat_val ) )\n",
    "            \n",
    "            # Compute the gradient of the overall loss with respect to the classifier weights\n",
    "            gradients = gt.gradient ( loss, self.weights )\n",
    "\n",
    "        # Apply the gradients\n",
    "        self.optimizer.apply_gradients ( zip(gradients, self.weights) )\n",
    "        \n",
    "        return loss, loss_val"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "270faf23-c06e-46c2-b801-40b74b8ee17b",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-04-27 10:20:52.349717: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:939] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2023-04-27 10:20:52.359955: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:939] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2023-04-27 10:20:52.360518: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:939] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2023-04-27 10:20:52.362066: I tensorflow/core/platform/cpu_feature_guard.cc:151] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 AVX512F FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "2023-04-27 10:20:52.362666: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:939] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2023-04-27 10:20:52.363262: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:939] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2023-04-27 10:20:52.363762: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:939] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2023-04-27 10:20:53.128976: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:939] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2023-04-27 10:20:53.129641: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:939] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2023-04-27 10:20:53.130197: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:939] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2023-04-27 10:20:53.130838: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1525] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 13789 MB memory:  -> device: 0, name: Tesla T4, pci bus id: 0000:00:05.0, compute capability: 7.5\n"
     ]
    }
   ],
   "source": [
    "dnn = SimpleNeuralNetwork(1500, learning_rate=0.0001, N_NODES=50, n_layers=8, n_features=X.shape[1], n_outputsC=2)\n",
    "dnn.save_weights(\"SigVsBkg_inclusive_RDF\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "4878e7f1-b129-4343-a100-f1de1e3fc119",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(180960, 57) (180960, 2)\n",
      "[[-2.80566406e+00  8.71459961e-01 -1.17529297e+00 ...  1.00746727e+02\n",
      "   9.28393173e+01  4.65607733e-01]\n",
      " [-1.51000977e+00  3.17871094e+00 -9.28222656e-01 ...  8.35415955e+01\n",
      "   4.95553818e+01 -6.27577305e-01]\n",
      " [ 1.06274414e+00 -2.28613281e+00  2.05468750e+00 ...  6.82547913e+01\n",
      "   1.19503845e+02  5.36567748e-01]\n",
      " ...\n",
      " [-7.98706055e-01  2.79541016e+00 -2.39941406e+00 ...  6.90503845e+01\n",
      "   2.03726746e+02 -1.83842495e-01]\n",
      " [-5.29632568e-02 -2.61425781e+00  2.93701172e+00 ...  5.98092117e+01\n",
      "   1.76484467e+02  3.13026577e-01]\n",
      " [ 2.38708496e-01 -3.43261719e+00 -5.46875000e-01 ...  1.08900810e+02\n",
      "   9.90312500e+01 -1.49214625e+00]]\n"
     ]
    }
   ],
   "source": [
    "print(X.shape, Y.shape)\n",
    "print(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "03845452-4fd3-4f29-b16f-b0d2ee0ef2c1",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1500/1500 [03:47<00:00,  6.60it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAzAAAAIhCAYAAACR9V+RAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAABcSAAAXEgFnn9JSAACRyUlEQVR4nOzdd3ic1YH24d+ZGWlGvTf33jFgGzDVNp0QeokpAZuQZJOwyYZsQrJsCGx6spCyyX4JSzEtDoQaerWNMdU22Lj3rt67NDPn+2NGo+KRLMuSRiM993XN9c7bzntGCHsen2astYiIiIiIiEQDR6QrICIiIiIi0l0KMCIiIiIiEjUUYEREREREJGoowIiIiIiISNRQgBERERERkaihACMiIiIiIlFDAUZERERERKKGAoyIiIiIiEQNBRgREREREYkaCjAiIiIiIhI1FGBERERERCRqKMCIiIiIiEjUcEW6AtI5Y0wBEA/sj3RdRERERER60Uigzlqbe7Q3GmttH9RHeoMxpsrtdieNHz8+0lUREREREek1O3fupLGxsdpam3y096oFZmDbP378+GkbN26MdD1ERERERHrN9OnT2bRpU496GWkMjIiIiIiIRA0FGBERERERiRoKMCIiIiIiEjWiOsAYY+KMMf9ljNlmjGkwxhwyxjxkjBl+FGUsMsbYbrxuCnOv0xjzXWPM58aYemNMsTHmKWPM1N79pCIiIiIiAlE8iN8Y4wHeAeYC+cALwBhgMfBFY8xca+2ubhS1A3ikk3MpwOXB9+91eL4D+AdwBVABvAxkAlcDFxtjFlhrP+7+JxIRERERkSOJ2gAD/CeB8PIBcL61tgbAGHM7cC/wEDD/SIVYa9+jQzhpYYz5BoEAsypMGLqFQHjZDpxprS0M3nMV8DTwhDFmqrXWe9SfTEREREREworKLmTGmFjgtuDut1rCC4C19j5gPTDPGDP7GB91Y3D7WJhztwe3P2gJL8HnPwP8E5gAXHaMzxcRERERkTaiMsAApxPo3rXTWvtpmPNPB7eX9PQBxpixwGlAE/BUmHNTgXoCXcd6/fkiIiIiInK4aA0wxwe3azs533J85jE8o6X15WVrbXknz99grW3uo+eLiIiIiEgH0ToGZlRwe6CT8y3HRx/DM7rqPtarzzfGbOzk1Pju3C8iIiIiMlREawtMYnBb18n52uA2qSeFG2NOBiYBZYTvItanzxcRERERkfCitQWmr7W0vjxlrW3q64dZa6eHOx5smZnW188XEREREYkW0doC0zLrWHwn5xOC2+qjLdgY4wK+FNwN132sT58vIiIiIiKdi9YAsy+4HdHJ+Zbje3tQ9vlANrDLWvt+BJ4vIiIiIiKdiNYAsy64ndXJ+Zbj63tQdkv3sce78fwZxpiYXn6+iIiIiIh0IloDzCqgEhhvjDkhzPmrg9sXj6ZQY0wirYtPdhpgrLW7gc1AHHBxbz1fRERERES6FpUBJjiw/k/B3T8bY1rGnGCMuZ3A+isrrLVr2hy/zRizxRjzyy6KvpLAuJYPrbXbj1CN+4Lb3xhjsts850rgUmAH8EJ3P9NAZK3lH5sPYq2NdFVERERERIDonoXsZ8C5wGnAdmPMSgLrrpwCFAO3dLg+E5gM5HVRZldrv3T0EPAF4ApgizHm7eAz5gH1wI3WWm/3PsrA87dNB/jZq1uoK26k5EvNfOPEMZGukoiIiIhIdLbAAFhrG4AFwE8JrMdyOYEAswSYZa3ddTTlGWPygLOBZuDJbjzfD1wDfA84BHwROA54Bphjrf3oaJ4/0Nzz2mbqihsB+N0b2/D7/RGukYiIiPQ2Y8xRvcaMGdPrdRgzZgzGmAFXlgxc0dwCg7W2Hrgr+DrStXcDd3dxPp+j/HlYa30EupLdd6Rro80PLpjMTx/7HICm8mZ+9eEO/uO0SRGulYiIiPSmm2+++bBj7733Hjt37uT444/nhBNOaHcuMzOzn2omR2PJkiUsXryYn/zkJ9x9992Rrk6fi+oAI33nK9NH8bsR26k50ADAA8t28e8njyfW5YxwzURERKS3LFmy5LBjixYtYufOnVx++eX98mX47bffprm5ecCVJQNX1HYhk77304um0TJ831/t4z/f3RLR+oiIiMjgM378eKZMmTLgypKBSwFGOnXF+Dwyx4UmeOOplfuobNC/aoiIiAxFS5YswRjD3XffzbZt21i4cCE5OTk4HA6ef/55AHbs2MHdd9/NqaeeSm5uLrGxsYwYMYKbbrqJbdu2hS033LiVPXv2YIxh/vz51NfX88Mf/pDRo0fjdruZMGECv/71r8POktqbZQGsWLGCs88+m6SkJNLS0vjCF77A6tWr2/0sumvDhg3ceOONjBs3Do/HQ1ZWFieccAL/9m//Rn5+/mHXb968mUWLFjFy5Ejcbjc5OTksXLiQjRs3trtu/vz5LF68GIB77rmn3ZilcC1sg4G6kEmX7r14Bjf/6SOMBer9fOO19fzt8tmRrpaIiIhEyNatWznppJPIyMhgwYIFlJeXExMTWNf7gQce4De/+Q0zZszgpJNOwu12s2nTJh577DFeeOEFVq5cycyZM7v9rKamJs4//3w2bdrE/Pnzqa2tZcWKFfzwhz+kurqan/3sZ31W1rPPPsu1116Lz+dj7ty5jBkzhs8//5wzzjgjFBi6a82aNZxxxhk0NDQwc+ZMLrvsMurq6ti1axd/+MMfuPzyy8nLa50o9/nnn2fhwoU0NjZywgknMHfuXPbv389TTz3Fiy++yKuvvspZZ50FwIUXXojX62XVqlWHjVuaMGHCUdUzalhr9RqgL2DjtGnTbKTNf3yVHX3HS3b0HS/ZUf/xkt1aWh3pKomIiPQ6v99vy5uaou7l9/t79edw8803W8D+5Cc/aXf84YcftoAF7G233Wa9Xu9h937wwQd2165dhx1/6KGHLGAXLFhw2LnRo0fbwFfSVrt37w49a968ebaysjJ07pNPPrFOp9PGx8fb6urqPimrsrLSpqenW8A+8cQT7cr78Y9/HCqv48+oMzfddJMF7H//938fdm7z5s320KFD7eqbkJBgExMT7Ztvvtnu2ldffdXGxMTYkSNH2sbGxtDxlv823a3PQDBt2jQLbLQ9+I6sFhg5oge+eDwLtqzA0WwxPvjaPz9j+aIzIl0tERGRXlXp9ZK2alWkq3HUyk8/ndRgC0h/yMrK4te//jVO5+ET+8ydOzfsPYsXL+bBBx9k+fLlVFZWkpKS0q1nORwO/vrXv5KcnBw6NmfOHC666CJeeuklVq9ezfz583u9rKeeeoqysjLOOeccrr/++nbl3HXXXTz66KPs3bu3W88FKC4uBuDcc8897FzHMTu///3vqa2t5X/+538Ou/7CCy/kG9/4Bn/84x95+eWXueKKK7pdh8FEY2DkiManJHDmKbmh/d1bKnlzT1EEayQiIiKRcu655xIfH9/p+ZqaGpYuXcodd9zBV7/6VRYtWsSiRYvIz8/HWsvOnTu7/azRo0czefLkw45PmhRY2iHc2JHeKGtVMMhec801h13vcrm46qqruv1cgNmzA93vv/Wtb7F8+XK83s7XOn/jjTcAuPLKK8OeP/PMMwH4+OOPj6oOg4laYKRb/nL+ccz4tBBT68cA3/vn56z/9jmRrpaIiIj0s1GjRnV67p133mHhwoWhFodwqquru/2sESNGhD2elJQEQGNjY5+U1RJmRo4cGfaern4G4Xz/+9/nvffeY/ny5SxYsIDExEROPfVULr74YhYtWtSuRWrPnj0ADB8+vMsyS0pKjqoOg4kCjHRLUmwMN50zlsf+GfhXk6pDDfzvp7v55oljI1wzERGR3pHiclF++umRrsZRS3H179c5j8cT9nhNTQ3XXnstZWVl3HXXXSxcuJDRo0cTFxeHMYbrr7+epUuXdjrjVzgOR+91FurNso5WcnIy77zzDqtWreLFF19k+fLlvPPOO7z55pv88pe/ZOXKlUycOBEAv98PhF9ktK1TTjmlz+s9UCnASLfdM3cSf/9gH83FgamU731tG1+dOZoYp3oiiohI9DPG9OtYksFm5cqVlJaWcvXVV3PPPfccdn7Xrl0RqFXPtMwItn///rDnOzveFWMMZ5xxBmecERhHXFRUxL/927+xdOlS7rzzTp566ikg0FK0c+dO7r33XjIyMnr4CQY3ffOUbnM4HPzk4umhfV+llx8s2xTBGomIiMhAUV5eDoTvqrVjxw7Wrl3b31XqsdODLXHPPPPMYed8Ph/PPvvsMT8jOzs7tI7Mhg0bQsfPO+88AJ577rlulxUbGwvQ5diawUQBRo7KjVOGkzG2dXHLZ1fuo6Su+/1PRUREZHBqGQz/7LPPthsDU1FRwVe+8hWam6NnMexrrrmG9PR03nzzTf7+97+3O/ezn/2M3bt3H1V5f/nLX8Le88orrwDtx9p873vfIy4ujn//938PG5QaGxt5+umnOXDgQOjYsGHDgMAaPUOBAowctT9ddjw2+JtjGi1ff3ldZCskIiIiETdnzhzOO+889u3bx6RJk7jiiiu44oorGDt2LIcOHeKyyy6LdBW7LSUlhf/7v//D6XRy3XXXcdppp3H99ddz3HHH8Ytf/IKvfe1rQGvLx5H85S9/Ydy4cUyfPp2rr76ahQsXcsIJJ/Dd734Xj8fDXXfdFbp2woQJLF26lObmZq666iomTpzIpZdeynXXXcdZZ51FRkYG11xzTbtB/HPnziU7O5unn36a+fPnc8stt3Drrbfy/vvv9+4PZoBQgJGjdmpuGlNnpof2V39azIHKugjWSERERAaCF154gTvvvJOsrCxeffVV1qxZw8KFC/nwww9JTU2NdPWOypVXXslbb73F/PnzWb9+PS+//DLDhg1j5cqVoVnIujtG5ac//Sm33HILxhjefvttXnzxRerr67n11lv57LPPQl3WWlx22WWsX7+eb37zmxhjePPNN3n55ZcpKirikksu4amnnmLatGmh6z0eDy+//DLnnXcen332GUuWLOHBBx9k27ZtvfcDGUDM0cwEIf3LGLNx2rRp0zZu3Bjpqhxmb3UdZ/52OY6mwO/PaSfn8Lcr50S4ViIiIiJ978ILL+T111/nww8/HNKzgR2L6dOns2nTpk3W2ulHvro9tcBIj4xOimfaca2tMO9/VkR1Q/T0bRURERHpysGDByksLGx3zO/387vf/Y7XX3+dSZMmcfLJJ0eodkObAoz02G/OmYbfGdxpsty1cmgMHBMREZHBb+XKlQwfPpyTTjqJq6++mksuuYTx48dz++23Ex8fzwMPPIAxJtLVHJIUYKTHZqYnM2xicmj/pY8PHtXiVCIiIiID1ezZs7npppuoqKjgjTfe4PXXX8fn8/HlL3+ZTz75hDPPPDPSVRyytJClHJM75k/iu1tWA9Bc7eUfmw9x7bThEa6ViIiIyLGZOHEiDz30UKSrIWGoBUaOyeWjs3Fmt04h+MdVOyNYGxEREREZ7BRg5JgYY7hk9rDQ/v7d1eRX1UewRiIiIiIymCnAyDH78dyJ+NyBQWzGD79ctT3CNRIRERGRwUoBRo5ZhjuW0W0G87/5eWEXV4uIiIiI9JwCjPSKr5w8JvS+vqyJNYcqIlYXERERERm8FGCkV3x5wjD8yc7Q/r0f7IhgbURERERksFKAkV7hcjg4bkpaaP/jTSVaE0ZEREREep0CjPSa204eR0tk8db6WL67JKL1EREREZHBRwFGes2FwzMho3Vt1P9bszeCtRERERGRwUgBRnqNMYYZE1q7ka3eXhrB2oiIiMiRXH/99Rhj+OlPf3rEaz/++GOMMeTk5OD1eo/6WYsWLcIYw/Lly9sdnz9/PsYY9uzZ0+2ylixZgjGGu++++6jrcbR6Uj/pWwow0qsWnTgy9L6pysunmo1MRERkwPryl78MwBNPPHHEax9//HEArrvuOlwu1xGujh7GGMaMGRPpagw4Azm4KcBIr7piVA6+5NZfq/+3Zk/kKiMiIiJdOv/888nJyWHr1q188sknnV7n9Xp58skngdbQ01seffRRNm/ezPDhw3u13N4y0Os3FCnASK9yORxMGJcS2n9/S3EEayMiIiJdcTqdXHfddUBrC0s4b7zxBkVFRUydOpXZs2f3ah1GjRrFlClTiImJ6dVye8tAr99QpAAjve7K44aF3teUNlFU3RDB2oiIiEhXbrzxRgCefPJJfD5f2Gtaupi1XFtRUcH//M//cMEFFzB69GjcbjcZGRlceOGFvPnmm0f1/K66Kq1atYpzzz2XpKQkUlNTueCCC/joo486LWvHjh3cfffdnHrqqeTm5hIbG8uIESO46aab2LZtW7trW8bRAOzduxdjTOg1f/78btVv06ZN3HDDDeTl5REbG8vw4cO56aab2Lp162HXLl++HGMMixYtoqysjG984xvk5eXhdruZMWMGDz30UPd+YEHWWp544gnOOOMMcnJy8Hg8jBw5knPPPZc///nPYa9funQpZ599NmlpaXg8HqZOncrdd99NXV1d6Lo9e/ZgjGHFihUAjB07tt3PZiAYPB0YZcC4ceIwfu3ZhKshMKnyY58f4HunTYhwrURERCSc2bNnM3XqVDZv3sybb77JhRde2O58bW0tL7zwAsYYbrjhBgA+/PBDvv3tbzNmzBgmT57Mqaeeyr59+3jjjTd44403eOCBB7jllluOqV4vvfQSV1xxBV6vl5NPPplx48axbt06zjrrLBYtWhT2ngceeIDf/OY3zJgxg5NOOgm3282mTZt47LHHeOGFF1i5ciUzZ84EYMKECdx888088sgjJCQkcPXVV4fKmTJlyhHr9/bbb3PJJZdQX1/PiSeeyPz589myZQuPPfYYzz33HK+88gpnnnnmYfdVVFRw6qmnUlNTw5lnnklJSQnvvvsuX/nKV/D7/dx6663d+vn84Ac/4L//+79xu92cddZZZGZmUlBQwPr169mxYwff+ta3Qtf6/X5uvPFGli5dSmJiInPmzCEtLY3Vq1dzzz338Oqrr7J8+XLi4uJITEzk5ptv5rXXXqOwsJCrrrqKxMTEbtWp31hr9RqgL2DjtGnTbDSa9td37Og7XrKj73jJLvi/lZGujoiIyBH5/X5bUdcUdS+/33/Mn/0Xv/iFBewNN9xw2LlHH33UAnbevHmhY7t27bIffPDBYdeuXbvWpqam2uTkZFtdXd3u3M0332wBu2zZsnbH582bZwG7e/fu0LGqqiqblZVlAfvQQw+Fjvv9fnvHHXdYwAL2Jz/5SbuyPvjgA7tr167D6vXQQw9ZwC5YsOCwc4AdPXr0Yce7ql9NTY3NycmxgP3Tn/7U7vr77rvPAnbEiBG2vr4+dHzZsmWhei9cuNA2NDSEzj333HMWsKNGjeq0Hm3V19dbt9ttk5KSDvu8zc3N9t1332137De/+Y0F7Pz5821+fn7oeGNjo/3KV75iAXvHHXcc8XP3pmnTpllgo+3Bd2S1wEifOHVCBm/tCjRH7tlXhc9vcToGRrOjiIhIOFUNXo6/541IV+OorfvJ+aTEHdv4jBtuuIE777yT559/ntraWhISEkLnWsbGtHQfg0C3orFjxx5Wzoknnsi3vvUtfv7zn7Ns2TIuueSSHtXn6aefpri4mLPOOovFixeHjrdM+fzEE09w4MCBw+6bO3du2PIWL17Mgw8+yPLly6msrCQlJSXsdd311FNPUVhYyKmnntqupQPgu9/9Lk888QRr1qzhmWeeCbVatUhOTuZPf/oTbrc7dOzyyy9nxowZbNiwgT179hxxVrSqqioaGxuZOnXqYf8dXC5Xu5Yfr9fLb37zGxISEvj73/9OTk5O6FxsbCz/8z//w8svv8z999/PL37xCxyOgT/CZODXUKLSouNGYoN5xd9keX+v1oQREREZqEaNGsVZZ51FbW0tzz//fOh4YWEhb7/9Nh6Ph2uuuabdPT6fjzfeeIO7776br3/96yxatIhFixaxbNkyALZv397j+qxcuRKAhQsXHnYuJiamXXevjmpqali6dCl33HEHX/3qV0P1ys/Px1rLzp07e1yvjvXrGE5atIS9luvamj17NhkZGYcdnzRpEgD5+flHfH52djYjRozgs88+44c//CG7du3q9Nq1a9dSUlLCaaed1i68tIiLi2P27NmUl5cf03+z/qQWGOkTp2WmYtOcmLLAYMBHPz/AmWMzI1wrERER6cyXv/xlVqxYweOPPx76Yr506VJ8Ph9XXnllu1aLAwcO8MUvfpF169Z1Wl51dXWP63Lo0CEARo8eHfZ8Zy0U77zzDgsXLqS4uPNZUI+lXh3r11k9Wo4fPHjwsHMjRowIe09SUhIAjY2N3arDI488wsKFC/n1r3/Nr3/9a0aPHs28efNYuHAhF110Uei6lskH3nzzzSMOwi8pKWHy5Mnden4kKcBIn3AYw/jRyewuKwfg4x1qgRERkYEt2eNi3U/Oj3Q1jlqyp3e+zl199dXcdtttvPXWWxQVFZGdnR3qPtZx7Zdbb72VdevWcdVVV/GDH/yAyZMnk5SUhMPh4P777+frX/96y3jeflNTU8O1115LWVkZd911FwsXLmT06NHExcVhjOH6669n6dKl/VKvroJCb3XROvvss9mxYwcvvfQSr732GsuXL+fRRx/l0Ucf5aqrruLpp58GAgP4ITBpwemnn95lmeFahgYiBRjpMxdNzeV/Pw0EmMqiBirrm4+5j66IiEhfMcYM6b+nUlJSuPTSS3nqqadYunQpF1xwAWvWrCEzM7PdzGS1tbW8+eab5OTk8OSTT+J0OtuV01V3pu7Ky8sDAtMbhxPu+MqVKyktLeXqq6/mnnvuOex8b9SrxbBhw7qsX0urR18vfpmcnMz111/P9ddfDwRmh7vmmmt45plneOWVV/jCF74QavGZMmUKS5Ys6dP69BeNgZE+c93EYfja/D3w8raCyFVGREREjqhl7MYTTzwRWvvlS1/6UrtFHCsrK/H7/eTl5R0WXpqbm3nuueeOuR4tg9Cfeuqpw855vV6eeeaZw46Xlwf+0TRcF60dO3awdu3asM+KiYnB6/X2qH5Lly4Ne76l5SrcNMp9ae7cuaHWsg0bNgBw0kknkZKSwooVKygrK+t2WbGxsQBH/bPpDwow0mdGxnlwZcWG9l/cogAjIiIykF144YVkZmbyySef8Je//AU4vPtYdnY2KSkpbNiwgVWrVoWO+3w+7rjjjsMWjOyJa665hoyMDJYvX84jjzwSOm6t5Sc/+Qn79u077J6WQfDPPvtsuzEwFRUVfOUrX6G5uTnss4YNG0ZhYSEVFRXdrt+1115LTk4O7733Hvfff3+7c3/84x9ZvXo1w4cP56qrrup2mUdj3759LFmypN0ClAANDQ2hSRRGjhwJgNvt5gc/+AHV1dVceeWVYVuiDh48yGOPPdbuWEsrU7hFOSNNAUb61OSRyaH3n++tiFxFRERE5IhiYmJCM3+VlJQwceJETjnllHbXuFwufvCDH+D1epk3bx7nn38+CxcuZMKECfzlL385bFrhnkhKSuLBBx/E6XSyaNEi5s6dy/XXX8+MGTP47W9/y1e/+tXD7pkzZw7nnXce+/btY9KkSVxxxRVcccUVjB07lkOHDnHZZZeFfdall16K1+tl1qxZ3Hjjjdx666389re/7bJ+CQkJPPHEE8TFxfH1r3+dOXPmcP311zNr1iy+853vkJiYyNKlS/F4PMf8swinrKyMxYsXk5WVxbx587jhhhu4/PLLGTVqFB9++CFz5szhyiuvDF3/wx/+MDRJw9SpU5k7dy7XXXcdV111FTNmzGDkyJHce++9h/1cAK6//nquueYabr311m4vstnXFGCkT10wqXW6vuqyJirqmiJYGxERETmSti0ubdd+aes//uM/eOSRR5g5cyarVq3irbfe4vjjjw99ee4Nl112GcuWLWPBggVs2LCBl19+mby8PFasWMFpp50W9p4XXniBO++8k6ysLF599VXWrFnDwoUL+fDDD0lNTQ17zy9/+Utuu+02vF4vTz75JA8++CAvv/zyEet3zjnn8Mknn3Dddddx4MABnn76aQoKCrjxxhtZvXp1n3YfGz9+PPfeey/z589n3759PPvss7z33nuMHj2a3/3ud6xYsaLdOjMOh4NHH32UF154gfPOO4/du3fzzDPP8N577+HxePj+97/PQw891O4ZV155Jb/73e8YMWIEL774Ig8++CAPPvhgn32mo2H6e4YI6T5jzMZp06ZN27hxY6Sr0mOHGho45Rdv4wzmlp9+6Ti+fOKoyFZKRERERCJq+vTpbNq0aZO1dvrR3qsWGOlTwzweYrNa/wXgpa2FEayNiIiIiEQ7BRjpc1NGtY6D2aBxMCIiIiJyDBRgpM+1HQdTW95Eea3GwYiIiIhIzyjASJ+7Ykw2vtbZlHlrZ1HkKiMiIiIiUU0BRvrcyLg4HOmtC2C9ul3jYERERESkZxRgpF+MGZ4Uer9uf2UEayIiIiIi0SyqA4wxJs4Y81/GmG3GmAZjzCFjzEPGmOE9LG+MMeYvxpjdxphGY0yJMeYDY8z3w1x7tzHGdvH61bF/wsHj9DHpofelRfU0ef0RrI2IiIiIRCtXpCvQU8YYD/AOMBfIB14AxgCLgS8aY+Zaa3cdRXkXAU8DccBa4EMgAzgO+DrQ2ZKsq4AdYY6v6e6zh4LLJuTyhGMHxg/4Ye2BcuaOyYh0tUREREQkykRtgAH+k0B4+QA431pbA2CMuR24F3gImN+dgowxU4BngWrgPGvt+23OOYBZXdz+gLV2SQ/qP6TMTknCm+IkptwHwD+3FSjAiIiIiMhRi8ouZMaYWOC24O63WsILgLX2PmA9MM8YM7ubRd4HeIBFbcNLsDy/tXZ1L1R7SHM5HGTlxIX2P9hdFsHaiIiIiEi0isoAA5wOpAA7rbWfhjn/dHB7yZEKMsaMBC4AdllrX+m9KkpHJ45OC73fn1+DtTaCtRERERGRaBStXciOD27XdnK+5fjMbpQ1n0CQe98Y4wKuJBCQnMAG4ElrbXkX959tjDmBQAvOAeBVa63Gv4Rx8YRslr2+HwBvg5+9pbWMyUyMcK1EREREJJpEa4AZFdwe6OR8y/HR3ShrWnBbA6wkMK6mrZ8bY6621i7r5P4vd9j/qTHmGQLd0WrC3dCRMWZjJ6fGd+f+aLEgN4PmeENMXaDl5a1dJdyqACMiIiIiRyFau5C1fOut6+R8bXCb1Mn5tlr6Nd0KTAGuB9KBycDjwffPhZmaeQfw78D0YH1GAjcAB4GrgMe68ewhJSMmhtg2C1qu2FMSwdqIiIiISDSK1haY3tQS4lzA1621TwX3y4EvG2MmAycB3wTubLnJWvt4h3Jqgb8ZY5YBnwOXB6dy/vBIFbDWTg93PNgyMy3cuWg1Oi+RPQcCA/g3HayKcG1EREREJNpEawtMS9es+E7OJwS31UdRVg3wjzDnHw5u53WnYtba/Db3XNide4aSk0a1DuQvK2mg2acFLUVERESk+6I1wOwLbkd0cr7l+N5ulNVyzT4bflqsPcFtdveqBsD24DbvKO4ZEi4Ym401gffWZ9la0J2MKSIiIiISEK0BZl1w29kCky3H13ejrJZpmNM6OZ8e3HZrQH6Hsmq7vGoImpuaTHNS66/dG7uKIlgbEREREYk20RpgVgGVwPjgFMYdXR3cvtiNst4HSoHc4HiXjlq6joVbb+YwxhgDXBHc7Wya5yEr0eUiPiM2tL9qrxa0FBEREZHui8oAY61tAv4U3P2zMaZlzAvGmNsJrP+you16LMaY24wxW4wxv+xQlhe4DzDBspLb3HMusAiwwF/bHM8yxnzLGNNuljNjTCLw/4BTgALg2V74uIPOhGGtP7bth9SFTERERES6L5pnIfsZcC5wGrDdGLOSwLovpwDFwC0drs8kMDVyuHEpvwUWBMvbZoz5MHj9XAILWt5prf24zfUJBALUr4wxnwD5QBaBrmsZQAVwtbW2s2meh7RTR6ezZVUxAFXljdQ3+YiLdUa4ViIiIiISDaKyBQbAWttAIHT8lMB6MJcTCDBLgFnW2l1HUVYz8AXgDqAEuAA4DlgBXGKt/UWHW0qBXwNrgEkE1n05nUCry73ADGvtqh5+tEHvwlFZ+FvyioXPDlZEsjoiIiIiEkWiuQUGa209cFfwdaRr7wbu7uJ8M/Cb4OtIZVUDP+xuPaW9E5OTaE524C4PTKH8+s4iTh2bEeFaiYiIiEg0iNoWGIlesQ4HqVme0P7H+8ojWBsRERERiSYKMBIRk/JaB/LvKTiaGapFREREZChTgJGIOGVUeuh9XVUzNY3eCNZGRERERKKFAoxExPzh6a0D+YGNhyojVxkRERERiRoKMBIRJyQFBvK3WL6nJIK1EREREZFooQAjERHvdBKf7g7tf3xAA/lFRERE5MgUYCRixuQkhN7v1EB+EREREekGBRiJmNkjUkPvK8obafL6I1cZEREREYkKCjASMfNHZmBNcMcP2wqrIlofERERERn4FGAkYuakJNOU1Por+P6+sgjWRkRERESigQKMREx2bCwxqa7Q/ioFGBERERE5AgUYiahh2fGh91vyqyNYExERERGJBgowElHHD0sJvS8uqcfntxGsjYiIiIgMdAowElFnjcqgJbL4vZY9pbURrY+IiIiIDGwKMBJRJ6en4E0wof1PD1RErjIiIiIiMuApwEhEjYuLw5fsDO2/v18D+UVERESkcwowElFOY0hP94T2NxRoLRgRERER6ZwCjETcuJzE0PsDxXURrImIiIiIDHQKMBJxs9vMRFZX3UxNozeCtRERERGRgUwBRiLutLw0/K3DYNhaoPVgRERERCQ8BRiJuJmJiTQntv4qrj5YHsHaiIiIiMhApgAjEZcbG4sjpbUJ5uMDCjAiIiIiEp4CjEScMYaczPjQvrqQiYiIiEhnFGBkQJicmxR6X1haj7U2grURERERkYFKAUYGhFOGp4XeNzf6KapujGBtRERERGSgUoCRAeGUzBS8bhPa35yvBS1FRERE5HAKMDIgTE9IaDcT2Yf7yyJYGxEREREZqBRgZEBIdrmIS40J7a85VBnB2oiIiIjIQKUAIwPGiOzWmch2FdVEsCYiIiIiMlApwMiAMSMvJfS+rKwRr88fwdqIiIiIyECkACMDxqnD07DBcfzWb9lTWhvZComIiIjIgKMAIwPGrJQkvPGtM5FtPKSZyERERESkPQUYGTAmxcfTnNRmJrKD5RGsjYiIiIgMRAowMmC4HQ5S0tyh/c/zNROZiIiIiLSnACMDyqjshND7fcV1EayJiIiIiAxECjAyoBzfZiayqsomGpp9EayNiIiIiAw0CjAyoMzNTcW2+a3cofVgRERERKQNBRgZUGYmJdKU2PpruUHjYERERESkDQUYGVAmxMXhaxNgPtJMZCIiIiLShgKMDCguh4P09NaZyDbmay0YEREREWmlACMDzticxND7A5qJTERERETaUICRAefEYa0zkdXVeqmsb45gbURERERkIFGAkQHn5OxU/K7W/e2F1ZGrjIiIiIgMKAowMuAcl9h+JrLPNROZiIiIiAQpwMiAM9bjwSY5Q/sfH9BMZCIiIiISoAAjA47DGDIzPKH9zepCJiIiIiJBCjAyII3Pbp2J7FBxPdbaCNZGRERERAYKBRgZkOYMTw29b2r0UVLTFLnKiIiIiMiAoQAjA9KczBR8sSa0v7VA3chERERERAFGBqjpCQk0JbUGmHX5FZGrjIiIiIgMGFEdYIwxccaY/zLGbDPGNBhjDhljHjLGDO9heWOMMX8xxuw2xjQaY0qMMR8YY77fxT2LjDEfG2NqjDFlxphXjDGn9fxTCcAotxuSWheDWX2wInKVEREREZEBI2oDjDHGA7wD/BhIBF4A9gOLgU+NMeOOsryLgI3A14BS4FlgLTAG+Hon9/weeBiYAbwFfAycB7xrjLn8KD+StGGMIafNTGTb1IVMRERERADXkS8ZsP4TmAt8AJxvra0BMMbcDtwLPATM705BxpgpBAJLNXCetfb9NuccwKww95wLfIdA2DnVWrs9ePxUYDnwsDFmubW2omcfTybnJPEJVQAUltXj91scDnOEu0RERERkMIvKFhhjTCxwW3D3Wy3hBcBaex+wHphnjJndzSLvAzzAorbhJVie31q7Osw9twe3P2sJL8HrPwD+AqQCX+nm8yWM2W1mIvM2Ww5W1EeuMiIiIiIyIERlgAFOB1KAndbaT8Ocfzq4veRIBRljRgIXALusta905+HGmDjg7A7P6tHzpXOz05LxelpbXLZpQUsRERGRIS9aA8zxwe3aTs63HJ/ZjbLmE/g5vG+McRljrjXG/MEY8ydjzL8YY9LC3DMZcAPF1toDx/h86URgJrLWX9FPD1VGsDYiIiIiMhBE6xiYUcFtuPDQ9vjobpQ1LbitAVYSGFfT1s+NMVdba5d19/nW2lpjTAWQZoxJstZ22XRgjNnYyanxXdZ8kMuLjcWZ7IJiHwBrDpZHuEYiIiIiEmnR2gKTGNzWdXK+NrhN6kZZLS0stwJTgOuBdAKtLI8H3z/XYWrmIz3/aOsgYRhjyMuMC+3vKKzp4moRERERGQqiNcD0ppafgQv4urV2qbW23Fq7zVr7ZeATAuNtvtlXFbDWTg/3Anb21TOjxdTc5ND70vJGmn3+CNZGRERERCItWgNMyz/Fx3dyPiG47c6o75o223+EOf9wcDvvKJ5/tHWQTpyUl4INjuP3+y17S2u7vkFEREREBrVoDTD7gtsRnZxvOb63G2W1XLPPWmvDnN8T3GZ39/nGmAQC0yiXH2n8i3TthJQkvPGtM5Ft0YKWIiIiIkNatAaYdcHtYQtMdji+vhtltUzDHG62MQiMgYHWVheArUAjkNVhbExPni9dmJ6QQFOiZiITERERkYBoDTCrgEpgvDHmhDDnrw5uX+xGWe8DpUCuMWZymPMtXcdC681Ya+uBd4K71xzj86ULWbGxxKbEhPY/PVQRucqIiIiISMRFZYCx1jYBfwru/jnYZQsAY8ztBNZfWWGtXdPm+G3GmC3GmF92KMsL3AeYYFnJbe45F1gEWOCvHapxX3D7n8aYiW3uORX4OlABPHgMH1OChme1DjXaXaQxMCIiIiJDWbSuAwPwM+Bc4DRguzFmJYF1X04BioFbOlyfSWBq5LwwZf0WWBAsb5sx5sPg9XMBJ3CntfbjtjdYa98yxvwB+A7wmTHmTSAWOI9AGFpsra3ohc855M3ITeYdygAor2ykodmHJ8YZ4VqJiIiISCREZQsMgLW2gUDo+CmB9VguJxBglgCzrLW7jqKsZuALwB1ACXABcBywArjEWvuLTu77N2AxsJlAcDkVeAs4y1r7/NF/KgnnpLxUbMtvqoUdRVoPRkRERGSoiuYWmJaxKHcFX0e69m7g7i7ONwO/Cb6Opg5LCIQm6SMzkxJpSnTgrgqsAbOloIoZw1MiXCsRERERiYSobYGRoWN6fDzNbWYiW3uwInKVEREREZGIUoCRAS81Joa4NjORrcuvimBtRERERCSSFGAkKozKbp2JbG+xZiITERERGaoUYCQqzMxrHfNSU9NMVUNzBGsjIiIiIpGiACNRYU52Cv42MydvL6yOXGVEREREJGIUYCQqHJeYSFNS66/rlgIFGBEREZGhSAFGosK0DjORrT5QEbnKiIiIiEjEKMBIVEh0uUhKiw3tbyiojGBtRERERCRSFGAkaozJTgi931dUh7U2grURERERkUhQgJGoceKw1ND7xkYfhyobIlcZEREREYkIBRiJGrMykvHGmdD+5wfUjUxERERkqFGAkagxIyGBxuTWX9nPDyrAiIiIiAw1CjASNabEx9PcJsB8sr8sgrURERERkUhQgJGoEed0kpkVF9rfnK+1YERERESGGgUYiSpThiWH3lfXNlNYpYH8IiIiIkOJAoxElRMzkvF6Wgfyb9A4GBEREZEhRQFGosqMhASa2oyD2XCwKoK1EREREZH+pgAjUWVWYmK7APPpgfII1kZERERE+psCjESV8XFxOFNdof316kImIiIiMqQowEhUcRjD1DYD+cuqmyipaYxgjURERESkPynASNQ5KSsFX2zrQP6NhzQORkRERGSoUICRqDMnObndOJj1+ysiVxkRERER6VcKMBJ1ZiUm0pja+qu7ep8G8ouIiIgMFQowEnUmxcdj0loH8q/ZV461NoI1EhEREZH+ogAjUcdpDFOHtw7kr6n3sqe0LoI1EhEREZH+ogAjUemkjBSaEloH8q/dq25kIiIiIkOBAoxEpVlJSTSlOkP7n+5XgBEREREZChRgJCrNSUpqP5BfLTAiIiIiQ4ICjESlqfHxuNJjQvtbC6qpa/JGsEYiIiIi0h8UYCQqOYxhzrBU/MFeZNbC+gOVka2UiIiIiPQ5BRiJWqemprTrRrZG3chEREREBj0FGIlapyYn09hmIP8nu8siWBsRERER6Q8KMBK1TklOpjG9NcB8tKcMr88fwRqJiIiISF9TgJGolR4Tw6jcBGxwOZj6Jh8bDlVFtlIiIiIi0qcUYCSqnZrefhzMR7tKI1gbEREREelrCjAS1eYmJ9PQphvZhwowIiIiIoOaAoxEtVNTUmhMaw0wH2scjIiIiMigpgAjUW1GQgLujJjQOJjaRh8bNQ5GREREZNBSgJGo5jSGszLTaExpMw5mt7qRiYiIiAxWCjAS9ealprabTvm9HQowIiIiIoOVAoxEvfmpqdRntFkPZlcpDc2+CNZIRERERPqKAoxEvRMSE3FnxuAPZphGr1+zkYmIiIgMUgowEvWcxnBmWioNbVphVmwrjmCNRERERKSvKMDIoDA/NZX6zDYBZqsCjIiIiMhgpAAjg8L81FTqs1oDzK6SWvaV1kWwRiIiIiLSFxRgZFA4ITGRxMQYmhJM6NiK7WqFERERERlsFGBkUHA5HJyblkZDm25ky7cURbBGIiIiItIXFGBk0LgoI4P6LFdo/70dJdQ2eiNYIxERERHpbQowMmhckJZGQ7oDXzDDNHr9mo1MREREZJBRgJFBY4THw4ykROqzW7uRvb6xIII1EhEREZHepgAjg8pF6enU5bR2I3tncxFNXn8EayQiIiIivanXA4wxJt4YM8oYk9DheJox5lfGmJeMMf9rjBnfC8+KM8b8lzFmmzGmwRhzyBjzkDFm+FGWs8cYY7t4TQlzz5Ij3PMvx/r55OhdmJ5OQ6YTf/A3u7rRy/s7SyJbKRERERHpNa4jX3LUfgz8ADgZWANgjHEDHwITgJZ5bq82xhxvrc3vyUOMMR7gHWAukA+8AIwBFgNfNMbMtdbuOspiH+nkeGUX97wOhOuntPUony294IyUFBJiXTRkOYkv9AGBbmTzJ2dHuGYiIiIi0hv6IsCcDey01q5pc+xGYCKBwPFr4GLg28B3CYSdnvhPAuHlA+B8a20NgDHmduBe4CFg/tEUaK1d1IN6/Mpau7wH90kfiHU4uCAtjdeyG0IB5o2NhfzXZX5inOoxKSIiIhLt+uIb3Shge4djlwIWWGytfdNa+2/ANuCinjzAGBML3Bbc/VZLeAGw1t4HrAfmGWNm96R8iW5XZGVRl+3CBtv6SmubeG+HupGJiIiIDAZ9EWDSgIqWHWOMAc4A1ltr97e5bh0wsofPOB1IIdDS82mY808Ht5f0sHyJYhenp+OMdVDXZjay5z89GMEaiYiIiEhv6YsuZAXA2Db7swmEmsc6XGeP4RnHB7drOznfcnzm0RRqjPk+MB5oBDYCz1lrj7SQyJXGmKsAJ7AbeNFau+Voniu9KzUmhnNSU3k3r5mENt3Iahu9JLj74ldeRERERPpLX3yb+4zAIPrLgbcJDOq3wEsdrpsIHOrhM0YFtwc6Od9yfPRRlvubDvu/M8b8q7X2oS7u+dcO+782xvw/4DvW2m4tA2+M2djJqWOeqW2ouiIri9dLyvC7wOGF+mYfb24q5PITj2qCOhEREREZYPqiC1lLCHiGQFeySwh0F3un5QJjTA6BVpQ1HW/upsTgtq6T87XBbVI3y/sncCWBwBMPzADuA9zAA8aYy8Lc8ynwL8Ck4D3jgG8R+MzfBH7bzWdLH7gsIwPjNNTmtmb059SNTERERCTq9XoLjLX2fWPMFcC/A5kEQsp/WGvbriZ4HVANvNbbz+8Ja+23OxzaCHzPGLMFuJ/AzGkvdLjnDx3u2Q38rzFmBYEubLcZY+7rMO6ns+dPD3c82DIzrXufQtrKdbs5LTmZ1XnlJB0INISt3F5MQWUDuSmeCNdORERERHqqT+aVtda+aK2dZ62dbq29yVp7oMP531tr06y1T/TwES2zjsV3cr5lEc3qHpbf4kGgCJhsjBnTnRustRsJtOi4gHOO8flyDK7LyaEx3YE3LjAdmd/CP1YfMU+KiIiIyAAWrQtj7AtuR3RyvuX43mN5SLDVaGdwN+8obm2ZRvpo7pFedm1WFk5jqBnR2tD490/24/cfy/wRIiIiIhJJvR5gjDE5xpizguNc2h4fb4z5uzFmgzHmFWPMqcfwmHXB7axOzrccX38Mz2iRFtzWdnnVsd8jvSwrNpbz09OpGd66JszBinpWak0YERERkajVFy0wPwSWEVinBQBjTDLwHnANgTEdFwJvGWMm9vAZq4BKYLwx5oQw568Obl/sYfkAGGOmA5MJTBbQramRjTFu4OLgbmfTPEs/uSEnB5/HQX1W65owf/94Xxd3iIiIiMhA1hcBZj6wyVq7rc2xRUAOsJRAILgdiAO+15MHWGubgD8Fd/9sjGkZ84Ix5nYC67+ssNauaXP8NmPMFmPML9uWZYz5gjHm7I7PMMbMBP4BGOCB4DNbzk0xxnw5GFba3pMF/J3AAp3rCAQtiaDLMjKIdzjadSN7c1MhhVUNEayViIiIiPRUXwSY4cCuDscuBrzAv1lrt1trf0/gC/68Y3jOz4CPgNOA7caYJ40xHwL3AsXALR2uzyQQnjqOSzkZeNsYs8cY84IxZqkx5iMCs6dNBZYTaFVqKxd4FMg3xrxhjHnCGLOMwHiZywmsQ3OttVaDLSIs0eXi8sxM6rOceD2BfmRev+WxD45peJSIiIiIREhfBJgk2qzPYoxxAqcCa6y1bQcfbKHzQfhHZK1tABYAPw0+73IC67gsAWZZazuGqM68DjwEVAGnE+h+NoFAl7evAudaa+s73LMN+D2wFTiOQNe4OQQG798DzOzQAiURdEteHhhD9ejWVpgnPtpLQ7MvgrUSERERkZ7o9XVggEPAlDb7ZxBYeHJ5mGc3cQyCweKu4OtI194N3B3m+AfAB0f53EPAd4/mHomcBampjPV42DvCkrKjGYcPyuuaeXbtQa4/ZVSkqyciIiIiR6EvWmA+AGYaY/7NGHMcga5elsMH1E8FtDS69DmHMdySm4s/xlAzvDWzP7RqN+rlJyIiIhJd+iLA/BJoJDAW5TMC3bKWW2vfb7kguCjkNAJjWET63KLcXBxA9egYWiLLjqIa3t5cFMlqiYiIiMhR6vUAE1yJ/gzgceA1Ai0wl3e47AICg/if7+3ni4QzwuPhovR0vAkO6rNbp1T+n3e2qxVGREREJIr0RQsM1tq11tqbrbUXW2vvstZWdzj/V2vtidbal/vi+SLh/MuwYQBUjo8JHVt3oJJ3t2thSxEREZFo0ScBRmQg+kJGBhPi4mhKcVKf2doK88e31QojIiIiEi36LMAYY3KMMT8yxrxijFkXfL1ijPmhMSanr54r0hmHMXx7+HAAKtq0wqzZW84HO0sjVS0REREROQp9EmCMMVcRWCvlZ8CFBNZKOS74/ufA1uA1Iv1qUW4uyU4nTWlO6jNaf/3vfXObWmFEREREokCvBxhjzBxgKZAAPAdcAZwInEBgMP+zBNaF+VvwWpF+k+Ry8ZW8PAAqx8eGjq/ZW87rGwsjVS0RERER6aa+aIH5EeAErrHWXm2tfcFau85au95a+09r7TUEVq6PAX7YB88X6dJtw4djgMZ0J3VZrWNhfv3aFpp9/shVTERERESOqC8CzBnA+9ba5zq7IHhuFXBmHzxfpEvj4uK4JisLgIrJra0wu0tqWfrxvkhVS0RERES6oS8CTArQnW+B+4LXivS7O0ePBqA50UH1CFfo+O/f2k5VQ3OkqiUiIiIiR9AXAaaAwJiXIzkheK1Iv5uZmMilGRkAVE6IwTgNAGW1Tdz3xrZIVk1EREREutAXAeZ1YLIx5hfGGGfHkybgZ8AU4LU+eL5It/xnsBXG53FQPra1FebRD/aw4WBlpKolIiIiIl3oiwDzU6AMuAPYYYz5tTHmG8HXr4AdBAb6lxKYZlkkIk5KTuaCtDQAKsfF4EwM5G2/hTuf34Dfr2mVRURERAaaXg8w1toDwNnARmA08H3gT8HXD4CxwAbg7OC1IhHzkzFjAm8chkNTWhe3XLe/gr9pQL+IiIjIgOM68iVHz1r7OTDTGDOfwExjw4KnDgErrbXL++K5Ikfr1JQULs/M5PmSEhoynThGxOI/0ATAr17dwoIp2QxPjYtwLUVERESkRZ8EmBbBoLI83DljzC3ACGvtf/VlHUSO5Bdjx/LPkhL8wL4JTsYXOWhs8lPT6OWOp9fz2FdOxhgT6WqKiIiICH0zBqa7vgr8JILPFwFgakICi3NzgcCA/uppntC593aU8PhH6komIiIiMlBEMsCIDBh3jxmDxxH43yE/B4aNTgyd++Urm9ldUhupqomIiIhIGwowIsAIj4fvjxwZ2DGGNeP9JMcFeljWNfm47W9raWj2RbCGIiIiIgIKMCIhPxw1ilFuNwBNsYaE2SmhcxsPVfHzlzdHqmoiIiIiEqQAIxIU73TyuwkTQvsfJjZw+gnZof3HPtzLy+vzI1E1EREREQlSgBFp44rMTM4PLm4JsGx4E1OHJYX273hmPTuLayJRNRERERFBAUakHWMMf5w4EXdw2uR8XzMpc9NIcgfGw9Q0ern1kdVU1DVFspoiIiIiQ9YxBxhjjK8nL+DkXqi/SK+bHB/P3WPGhPafrCvlhovGhfZ3l9Tyrb+tpdnnj0DtRERERIa23miBMcfwEhmQ/n3kSGYntk6l/L+2hNvOaR0fs2pHKfe8uBFrbSSqJyIiIjJkHXOAsdY6juHl7I0PIdLbXA4HD02ZgivYlWx/YyObR/i59PhhoWse/3Aff313V6SqKCIiIjIkaQyMSCdmJiZy56hRof2HCguZe1Yex49MDR371atbeOqT/RGonYiIiMjQpAAj0oU7R4/mpKTWWci+uWs7d39pBmMy4kPHfvjsel7fWBCJ6omIiIgMOQowIl2IcThYOm0aSc5Ab8dKn4/b9u3k4cUnkZUUWPTSb+Ffl37K8q1FkayqiIiIyJCgACNyBOPj4vh/kyaF9t+vquL+qiIeveVkkjyB6ZWbvH6+9tgalinEiIiIiPQpBRiRbrghJ4ebcnJC+7/at4+NzgaWLD6JhNhA60yT18/XH13Dsi0KMSIiIiJ9RQFGpJv+PHEi0+Jbx74s2rIFT6abR79yMonBhS6bfH6+9thq/rnuUKSqKSIiIjKoKcCIdFOiy8XzM2aQEhwPU+v3c/mGDYwflsQjt7SGmGaf5dtLP+XhVbsjWV0RERGRQUkBRuQoTIyP54lp00KrsO6or+fqjRs5bmQKj996CmnxMaFr73lxE79+bYsWuxQRERHpRQowIkfp4owM7hkzJrT/dkUFX926leNHpPCPfzmN4alxoXP/b/lO/nXpp9Q3+SJQUxEREZHBRwFGpAfuHD2a67OzQ/uPFhZyz549TMhO5JlvnMbknNa1Y15an881f32fQxX1kaiqiIiIyKCiACPSAw5jeGjKFOalpISO3bN3Lw/n55Ob4uGpfzmV0ydkhM5tOFjFpX9axZq9ZZGoroiIiMigoQAj0kNuh4PnZsxgSpuZyW7dupVniotJiYvhkcUns+i0MaFzJTWNLLz/Qx5etVvjYkRERER6SAFG5BikxcTwynHHkRMTGLzvB67btIlXSktxOR3cfel0fnPVTGKcgWH/zT7LPS9u4l8eX0NlfXMEay4iIiISnRRgRI7R2Lg43jj+eNJcwWmUreWqjRtZVl4OwLUnjWTpV+eSneQO3fP6xkIu/uNKPttfEYkqi4iIiEQtBRiRXjAzMZHXZ84kKbhGTIPfzyWff86KigoA5oxJ55XvnMmZEzND9xwor+eav7zP/y7fgdfnj0S1RURERKKOAoxILzkpOZmXjzuOOEfgf6tav58L16/n9bLAwP3MRDePLD6Z718wGUdwIZlmn+U3r23lmr9+wK7imkhVXURERCRqKMCI9KIzU1P554wZoRDT4Pdz6eef80JJCQAOh+FbCyaw9KtzyU32hO77dF8FX/jjSh5etRu/XwP8RURERDqjACPSy85NT+fVmTNJDHYna7KWqzZs4O+FhaFrThmXwev/dhZXnjg8dKyh2c89L27i2r9+wLbC6n6vt4iIiEg0UIAR6QPzUlN5c+ZMUoIhxgdcv3kzfzxwIHRNSnwM933pBP5y42wyEmJDx1fvLecLf1jJr1/bQn2Tr7+rLiIiIjKgKcCI9JG5KSm8c8IJZARnJ7PAd3bs4Ps7d+Jvsw7MhTNyeeO7Z3HxcXmhY16/5f8t38l5v1vBO1sKOxYtIiIiMmQpwIj0oVlJSbx74omMcrdOofzf+/dzw+bNNPpbZx7LSHTz5xtm8dCiOYxIiwsdP1Bezy1LVvPVR1ezu6S2X+suIiIiMhApwIj0sWkJCXwwaxbHJySEjv29qIjz1q2jqKmp3bVnT8nhze/O41/mjcfVMlUZ8OamQs7/3Qp++tImKuu0AKaIiIgMXQowIv1gmNvNuyeeyLlpaaFjKysrOWnNGj6rbj9gPy7WyQ8vmsLL3z6Tk8a0Xt/sszz43m7m/fcylqzaTbPWjhEREZEhKKoDjDEmzhjzX8aYbcaYBmPMIWPMQ8aY4Ue+u105e4wxtovXlE7ucxpjvmuM+dwYU2+MKTbGPGWMmdo7n1AGk2SXi5ePO45FubmhY/saGzn900/5R1HRYddPzk3iqa+fyv9cdyLDU1u7lVXUNXP3i5u44Pfv8tqGAqzVtMsiIiIydJho/fJjjPEAy4C5QD6wEhgDnAwUA3Ottbu6WdYeYDTwSCeX/Mham9/hHgfwNHAFUAG8DWQCZwH1wAJr7cdH85nC1GvjtGnTpm3cuPFYipEBxlrL7w8c4N937qRtG8p/jBrFf40di9OYw+5paPbx4Hu7+d9lO6jtMDPZjOHJfO+8ycyfnIUJc6+IiIjIQDN9+nQ2bdq0yVo7/WjvjeYA8zPgTuAD4HxrbU3w+O3AvcAKa+38bpa1Bxhtre32tz9jzK3A/wHbgTOttYXB41cRCDY7gKnWWm93ywzzDAWYQez1sjIWbtpEhbf1V2RBaip/mzqV3DaD/tsqqm7gvje28eTq/XT8X/fEUal877zJnD4hQ0FGREREBrRjCTBR2YXMGBML3Bbc/VZLeAGw1t4HrAfmGWNm92E1bg9uf9ASXoLPfwb4JzABuKwPny9R7oL0dD6eNYsp8fGhY8sqKjhh9WreLi8Pe092kodfXTWTV759JudNy2l37tN9Fdz44EcsvP9DVu0oUdcyERERGZSiMsAApwMpwE5r7adhzj8d3F7SFw83xowFphLoKvZyfz9fBo+J8fF8NGsWX8rKCh0rbG7mvHXruGfPHnydhJCpecn8301z+OdtpzN/cla7cx/tLuOGBz7i8j+v4rUN+fj9CjIiIiIyeERrgDk+uF3byfmW4zOPplBjzPeNMX8xxvzBGPM1Y0xWJ5e2PH+DtTbcnLY9er4MTckuF0unTeN/J04kNtj1ywJ379nD2Z99xp76+k7vnTkilSWLT+aZb5zKaeMz2p1bd6CSf3l8Lef+bgVPfbKfRq+vk1JEREREoke0BphRwe2BTs63HB99lOX+Bvg68G3gr8AeY8wtff18Y8zGcC9g/FHUXaKYMYZvDB/Oh7NmMd7jCR1/t7KSmatXsyQ/v8suYbNHp/O3r87l71+by5kTM9ud21Vcyw+eWc/pv1rG797cRlF1Q599DhEREZG+Fq0BJjG4revkfMuS5UndLO+fwJUEAkc8MAO4D3ADDxhjOo5l6e3niwBwYlISa+fMYWF2duhYtc/H4q1buWrjRoo7LHzZ0dxxGTz2lVN46V/P4OKZebQdy19S08gf3t7O6b96h9uf/IzPD1T21ccQERER6TOuSFdgILDWfrvDoY3A94wxW4D7gV8DL/Th88POvhBshZnWV8+VgSnZ5eJvU6dySUYG39y2jUpfoOvXcyUlvF9Zyf2TJ3NpZmaXZcwYnsKfr5/F7pJa7n93J8+uPUijNzBpc7PP8uynB3n204PMGZ3GzaeN4YLpucS6ovXfM0RERGQoidZvLC2zjsV3cj4huK3u5Hx3PQgUAZONMWMi8HwZoowxXJ+Tw+cnncQ5qamh44XNzVy2YQNf2riRwiO0xgCMzUzgl1fO5IMfncP3L5hMbrKn3fnVe8v516Wfcuov3+aXr25mb2ltJyWJiIiIDAzRGmD2BbcjOjnfcnzvsTzEWusHdgZ38/r7+SIjPR7eOP54fj9hAh5H6/+uTxUXM/Xjj3n4CGNjWqQnxPKtBRNYeccC/njdiZw4KrXd+dLaJv66YhfzfrucLz/4Ea9+nk+zzx++MBEREZEIitYuZOuC21mdnG85vr4XnpUW3Lb9p+mW588wxsSEmYmsN58vQ5zDGL4zYgQXpKXx1W3beK8yMHal3Ovllq1beaKwkL9MmsSE+M4aBFvFOB1cevwwLj1+GJ/tr+DRD/bw8vr8UPcygJXbS1i5vYSsJDfXzB7BVbNHMD4rsYtSRURERPqPicbF7oILWRYRWAvmRGvtZx3OryMwhfEca+2aY3jOdOBzAuu9pFlrm9qc20RgLZgrrLXPd7jvBeBS4OrgwpY9ff7GadOmTdu4cWNPi5BBxm8tfz10iDt27aLa1zotcqwxfH/kSH40ejQJTudRlVlR18Szaw/yxEd72VkcvgvZiaNSuXLWCC6dOYyU+Jhj+gwiIiIi06dPZ9OmTZs6GwvelajsQhYMEn8K7v7ZGNMy5gRjzO0EwsuKtuHFGHObMWaLMeaXbcsyxnzBGHN2x2cYY2YC/wAM8EDb8BJ0X3D7G2NMdpv7riQQXnbQhwP/ZWhyBKdb3nTSSVyS0bruS5O1/HzfPqZ+/DFPFxV1q1tZi9T4WG45Yyxv3T6PJ782l8tOGEass/0fDZ/uq+DHz2/gpJ+/xbeeWMs7WwrxqouZiIiIREBUtsAAGGM8wHLgFCAfWElgGuRTgGJgrrV2V5vr7wZ+AjxirV0U5vheAl3D6oBxBLqBuYLP+IK1tt1qgsYYB/A0cAVQDrwNZALzgAZggbX2o2P8jGqBkU5Za3m6uJjv7tjBwQ4D+s9JTeWPEycyLSGhk7u7VlbbxLNrD/D0mgNsKQg/F0VmYiwXH5fHJccPY9aoNBwOE/Y6ERERkY6OpQUmagMMgDEmDvgRcD0wEigDXgN+bK090OHauwkfYE4FbgVOAoYR6JZWRWD8yhPAw9basEuYG2OcwHeAWwgsOlkLLAN+Yq3d1AufTwFGjqjG6+Xn+/Zx7/79NLf5/9kJfH3YMO4aM4ac2Ngel7/xUCXPrDnIC58dpLQ2/Mxnw1Pj+OLxeVwycxjThyVjjMKMiIiIdG7IBpjBTgFGjsa2ujq+s2MHr5WVtTue6HTyg5EjuX3kyKMeH9NWs8/Piq3FPLP2AG9vLqKpky5k47ISuGTmMC49YZgG/4uIiEhYCjCDlAKMHC1rLS+WlvLdHTvY1dDQ7lxebCz/NWYMi3JzcTmObfhbRV0Tr28s4J/rDvHBzlL8nfwxMi0vmUtPGMYXZuQxKuPIs6SJiIjI0KAAM0gpwEhPNfr9/OXQIf5rzx7KvN5256bFx3PPmDFcmZWFoxe6ehVVN/DK+nz+ue4Qa/dVdHrd1LxkLpqRy4UzcpmYnahuZiIiIkOYAswgpQAjx6qiuZlf7dvH7w8coLHD/+vHJyRwz9ixXJqR0WthYn9ZHS+tz+fFdYfYlF/V6XXjshK4cHouF83IY8ZwjZkREREZahRgBikFGOkt+xoauGv3bh4tLKTj//GzExP5r7FjuSg9vVeDxI6iav65Lp/XNuSzrbCm0+uGp8ZxYbBlZtaoNJyazUxERGTQU4AZpBRgpLdtqKnh7j17eKak5LBzc5OTuXvMGM5PS+v1FpGdxTW8tqGA1zYU8PnByk6vS0+I5ewp2Zw7NZszJ2aR4Hb1aj1ERERkYFCAGaQUYKSvfFZdzU/27OGfpaWHnZudmMh/jB7N5ZmZvTJGpqMD5XW8vrGQ1zbks3pvOZ39ERTrdHDq+AzOnZbDOVOyGZYa1+t1ERERkchQgBmkFGCkr62uquKuPXt4tcPUywBT4uP54ahRXJ+dTcwxzlrWmaLqBt7YWMhrGwr4cFcp3s6mMyMwo9m503I4d2o2M4alaOFMERGRKKYAM0gpwEh/+aCykp/u3Rs2yIxyu/nBqFHckptL3DGsI3MkVQ3NrNhazFubC1m2pYiqBm+n1+Ykuzl7SiDMnDo+g/hYdTUTERGJJgowg5QCjPS3T6ur+eW+fTxdXHzYYP+smBi+OWwY3xw+nOzY2D6tR7PPz+o95by9uZC3Nheyp7Su02tjXQ5OGZvOvElZzJ+czfisBM1qJiIiMsApwAxSCjASKVvr6vj1vn08VliIt8OfEW5juDEnh++OHMn0hIQ+r4u1lp3Ftby1uZC3NxeyZm95pwtnAoxIi2P+5CzmT8rmtAlqnRERERmIFGAGKQUYibR9DQ3cu38//5efT73ff9j5C9LSuH3kSM7rg5nLOlNW28SyLUW8tbmQ97aXUN3YeVezWKeDk8emBwLN5CzGZ2kBTRERkYFAAWaQUoCRgaKkqYm/5ufzp4MHKWhqOuz89Ph4vjtyJNdlZxPfh+NkOmr2+Vm7t5zl24pZvrWYzV0sngmBNWcCYSab08ZnaJpmERGRCFGAGaQUYGSgafT7ebKoiHv372d9be1h59NcLhbn5vKNYcOYEB/f7/UrrGpgxdZilm8rYuX2Eqq7mAgg1ungpLFpzJ+UzfzJWUzIVuuMiIhIf1GAGaQUYGSgstayrKKC+/bv5+UwM5cBXJiezreGDeOijAycEQgGzT4/n+6rYPnWIpZvLWZTN1pn5k3OYt6kLE6fkEmiWmdERET6jALMIKUAI9FgS20tfzh4kMcKCqgNM05mjMfDN4YN45bcXDL7ePayrhRVNbB8WzErthbz7vbiLltnYpyGOaPTQ93NJuWodUZERKQ3KcAMUgowEk0qvV4eLSjgfw8dYkvd4dMeu41hYXY23xw+nJOSkiIaCLw+P5/ub22d2Xio69aZYSmeYOtMNqdPyCDJE9NPNRURERmcFGAGKQUYiUYt3cv+fPAgL5SU4AtzzQmJiXwtL4/rc3JIcUW+q1ZRdQPvbith+dbA2JnK+uZOr3U5DHPGpDF/cmDszOScyIYxERGRaKQAM0gpwEi0O9DQwF/z8/m/Q4cobD48FMQ7HHwpO5uv5eVxSnLygAgCXp+fdQcqWL41MLPZ5wcru7w+N9nDWZMymTcpmzMmZJISr9YZERGRI1GAGaQUYGSwaPL7eaa4mP89dIj3KsMHghkJCXwtL48bc3JIixk4IaC4upF3txWzfFsxK7cXU1HXeeuMw8AJI1M5a1JgMoCZI1JxOiIfykRERAYaBZhBSgFGBqNNtbU8kJ/PIwUFlHkPH0jvcTi4NiuLr+blcXpKyoBolWnh81s+21/Biq1FLN9WzPoDXbfOpMbHcMaETOZNyuKsSVnkJHv6qaYiIiIDmwLMIKUAI4NZg8/HcyUl3J+fz/KKirDXTI2P56t5edyUm0vGAGqVaVFS08h720t4d1tgZrOSmsMX+WxrSm4S8yYFZjabMyaNGKejn2oqIiIysCjADFIKMDJUbKur44H8fB4uKKAkzFiZWGO4LDOTW3JzOS89PSLryhyJ32/ZlF/Fim3FrNhWzNq95Xj9nf/5muh2cfqEDBZMzmbe5CzyUuL6sbYiIiKRpQAzSCnAyFDT5PfzQrBV5q3y8rDXjHC7uTknh8V5eYyPG7hf+qsbmnl/ZynvBgPNgfL6Lq+fkpvEvMlZzJ+k1hkRERn8FGAGKQUYGcp21tfzYLBVpqApfNes+ampLM7N5aqsLBKczn6uYfdZa9lVUsuKrYEw88GuUpq8hy/62SLJ7eL0CZmhhTRzUzR2RkREBhcFmEFKAUYEvH4/r5aV8VBBAS+VluIN82dWktPJwuxsbsnNHTDTMXelvsnHh7tKAwtpbitmb+nhC3+2NTUvmQWTszh7SjYnjEzFpdYZERGJcgowg5QCjEh7hU1NPF5YyIP5+WyuC/+lf2p8PLfk5vLl3FxyYmP7uYY9s7uklmVbAmHmwyO0zqTExXDWpCzOnpLFvEnZpCdEx2cUERFpSwFmkFKAEQnPWsvH1dU8lJ/P0qIiqn2+w65xGcPF6enckpfHRenpxDiio9XiaFpnTHDdmQWTszl7SjbT8pJxaN0ZERGJAgowg5QCjMiR1fl8PFNczEMFBZ1Ox5wTE8OXc3NZnJvLtISE/q3gMbDWBlpnthazbEsRH+0updnX+Z/ZWUlu5k8KdDU7fWImyZ6BN/W0iIgIKMAMWgowIkdnZ309SwoKWFJQwIHGxrDXnJyUxOLcXBZmZ5M6ANeW6UpNo5dVO0pYvrWIZVuKKahq6PRal8Nw0ph0FkzJYsHkbCZkJw74sUEiIjJ0KMAMUgowIj3js5a3yst5KD+f50tKaArz55zbGK7IymJxbi7npKUNyLVlumKtZXN+Ncu2FrFsSxFr95XTxbIzjEiLY8HkbBZMyeLUcZnExQ7cWdtERGTwU4AZpBRgRI5daXMzfyss5OGCAj6tqQl7zQi3m5tycliUm8vE+Ph+rmHvqKhrYsW2YpZvLWb51iLK6w5fELSF2+Xg1PEZnD0lmwWTsxmZHp2fWUREopcCzCClACPSu9bV1LCkoIDHCwspaQ7/Bf/05GQW5+VxbVYWSS5XP9ewd/j8lnUHKli+pYh3thax4WBVl9dPyE5kweQsFkzJZs7odGJd0THhgYiIRC8FmEFKAUakbzT5/bxcWsqSggJeLi3l8DnMIN7h4KpgF7N5qak4oqyLWVtFVQ0s31rMsq1FrNxeQk2jt9NrE90uzpiQydlTspk/OYvsZC2iKSIivU8BZpBSgBHpey1ryzycn8/GTtaWGePxcHNODjfn5jI2Lq6fa9i7mrx+Vu8tY/nWYt7ZUsSOovDd6lpMH5YcDDOBRTSdmqZZRER6gQLMIKUAI9J/rLWsqa7m4YIC/lZURIU3fCvFgtRUFuXmclVWFgnO6B8Iv7+sLjQRwPs7S2nsYhHNtPgY5k0KdDU7a2IWaVpEU0REekgBZpBSgBGJjAafj3+WlvJwQQFvlJUR7it9ktPJtVlZLMrN5fSUlEExRXFDs48PdpaybGsR72wp4kB5fafXOgycOCot1NVsWl7yoPgZiIhI/1CAGaQUYEQi72BjI48VFPBwQQHb6sN/oZ8YF8ei3Fy+nJPDSM/gGDNirWVncQ3LtgS6mn2ypwxvF/M05yS7WTA50NXsjImZJLqjcwIEERHpHwowg5QCjMjAYa3lw6oqHi4o4O9FRVT7Dh/6b4Dz0tJYnJvLZZmZxA2CLmYtqhuaeW97SaC72dZiiqvDLxQKEOM0nDw2PbjuTDbjMhPUOiMiIu0owAxSCjAiA1Odz8dzJSU8nJ/POxUVhPtTNMXp5LqcHBbn5nJSUtKg+gLv91s25VexLDhN82f7K+jqr5JR6fGhrmZzx2XgiRk8wU5ERHpGAWaQUoARGfj2NjTwaEEBSwoK2NXQEPaaafHxoS5muW53P9ew75XVNrFiWxHLthSzYlsxlfWdL6LpiXFw+vhM5k/JZsHkLEakaRFNEZGhSAFmkFKAEYkefmtZWVnJkoIC/lFURK3/8KH/TuDC9HQW5+VxSUYGsY7Bt2Ck1+fns/0VwYkAitmc3/UimpNyElkwJZsFk7OZPTqNGOfg+5mIiMjhFGAGKQUYkehU4/XydHExDxcU8G5lZdhrMlwurg92MTsxKamfa9h/CiobQtM0v7ejhLqmcMuGBiR5XJw1MYv5k7OYPzmbrKTB11olIiIBCjCDlAKMSPTbWV/PIwUFPFJQwL7G8APfj09IYFFuLjfk5JAVO3jXVmn0+vhkd3lwIoAidhXXdnn9zBEpzJ+czdlTspk5PAWHFtEUERk0FGAGKQUYkcHDby3vlJezpKCAZ0pKaAjTxcxlDF/MyGBxbi4XpacTMwi7mLW1p6SW5VuLeGdrMR/uKqWpi0U0MxJimTc5iwWTA4topsTH9GNNRUSktynADFIKMCKDU6XXy5NFRSwpKOCDqvBjRLJjYrgx2MVsRmJiP9ew/9U1eXl/R2mou9mhyvATIgA4HYbZo9KYPyWLs6dkMzlncM3yJiIyFCjADFIKMCKD35baWpYUFPBoYSH5TU1hr5mTlMSi3Fyuy84mPWbwtzxYa9lWWBOcCKCINXvL8XWxiGZeiifU1ey08RkkaBFNEZEBTwFmkFKAERk6vH4/b5aX83BBAS+UlNAU5s/mWGO4PDOTRbm5nJ+ejnOItDpU1jezcntxcJrmIkpqwgc9gFing1PGBRbRPHtKNmMyE/qxpiIi0l0KMIOUAozI0FTW3MzSoiIezs9nTU1N2GuGxcZyU24ui3JzmRw/dNZS8fstnx+sDHU1W3cg/CxvLcZmJjB/cqCr2clj03G7tIimiMhAoAAzSCnAiMjnNTUsKSjg8cJCiprDLxB5anIyi3NzuTY7mxTX0Oo+VVzdyLvbinlnaxHvbiumusHb6bWeGAenjM3gzImZzJuUxYTsRI2dERGJEAWYQUoBRkRaNPv9vFpWxsMFBbxUWoo3zJ/dcQ4HV2ZmsjgvjwWpqTiG2Jdzr8/Pmr3lLNtazLItRWwtrO7y+rwUD2dOzOSMiVmcNj6DzEStOyMi0l8UYAYpBRgRCaeoqYm/FRbycEEB62vDr6Uyyu3m5mAXs3Fxcf1cw4HhYEU9y4NdzVbtKKW+ufNFNAGm5iVzxoQMzpiYxclj0omLVXczEZG+MmQDjDEmDvgRsBAYBZQBrwE/ttYePIZyJwLrAQ/wtrX23DDXLAFu7qKYb1hr/9LTOgSfoQAjIp2y1vJpsIvZE4WFlHnDd586KyWFxbm5XJ2VReIQ62LWotHrY82eclbuKOHdbcVsPBR++uoWsU4Hs0ancsaETE6fkMlxw1NwOQf3ujwiIv1pSAYYY4wHWAbMBfKBlcAY4GSgGJhrrd3Vw7KXAfMAw5EDzOtAQZhiHrHWLuvJ89s8QwFGRLql0e/nxZISlhQU8GpZGeGWhExwOLgmO5vFubmcmZIypMd/lNQ0smpHCe9uK2HVjhIKqjpfdwYg0e1i9ug0ThmXziljM5g5IoUYBRoRkR4bqgHmZ8CdwAfA+dbamuDx24F7gRXW2vk9KPcrwAPA/cDXOHKAWWCtXd6zT3HEuijAiMhRy29s5LFgF7MtdXVhrxnn8bAoN5ebc3MZ5fH0cw0HFmstO4trWbWjhPd2lPDhzlKqGzufDAAgLsYZCDRj0zllXAbHj0zRDGciIkdhyAUYY0wsUASkALOstZ92OL8OmAnMsdauOYpyc4DNwGrgFwRaeBRgRCQqWWv5uLqah/Pz+XtREZW+w8eAGOCctDQW5eZyZWYmcU59Cff6/Kw/WMmq7YFAs3ZfOc2+rv+udLscnDgqleOGp3Du1BxOHJVGrEstNCIinTmWABOtnaFPJxBednYML0FPEwgwlwDdDjDAH4A44JvAiGOtpIhIJBljOCU5mVOSk/ndhAk8X1LCwwUFvFVeTsvXcQu8VV7OW+XlJDudLAx2MTslOXnIdjFzOR3MGpXGrFFp/Os5E6lv8vHpvnI+3F3GR7tK+XR/BU3e9p30Gr1+PtxVxoe7yvi/lbvxxDg4cWQaJ41JY86YdGaNTiPRHa1/5YqIDCzR+qfp8cHt2k7Otxyf2d0CjTFfAL4E3GWt3WGM6W6AudIYcxXgBHYDL1prt3T3uSIi/SHO6eS6nByuy8lhf0MDjxYWsqSggB319aFrqnw+7s/P5/78fKbEx7MoN5cv5+QwzD20pxeOi3Vy2oRMTpuQCUBDs491+yv4aHcZH+0uZc3echqa2weahmY/H+wq5YNdpQA4DEwblsxJY9I5aUw6c8akkZ00tLvuiYj0VLR2IbsP+C7wO2vt7WHOHw98Bqy11s7uRnkJwEagAZhprW0yxsyne13IOrLA/wO+Y63tuhN1a1md9REbP23aNLe6kIlIX7DWsqqykocLCniquJiaMF3MHMAF6ekszs3l0sxM3A51i+qoyetn/YEKPt1Xwbvbi/loVxlNvnDTKLQ3JiOe2aPTOXFUKrNGpTE5NwmnY2i2eonI0DMUu5AlBrfhR6dCy8IISd0s72fAaALjWZq6ec+nBCYQeAc4AOQCFwXL+ibQRCBkiYgMSMYYzkhN5YzUVP44cSLPFBfzcEEByysqQtf4gVfLyni1rIw0l4vrs7NZnJfHrEStYt8i1uVgzph05oxJ56tnjaOh2cfnByv5ZE8Zn+wuY/XecqobDv/3rD2ldewpreOZtQcASIh1csKoVGaPSmP2mHROGJlKSlxMf38cEZEBL1pbYO4Hvgr83Fr7n2HOTwC2A9uttZOOUNYc4CPgcWvtzW2Oz6eLFpguyptOoAubAxhnrd3f3XvDlKVB/CLS73bV1/NoQQFLCgrY29gY9prjEhJYlJvLjTk5ZMfG9nMNo4vfb9lWVM0nu8v4ZE85n+wpI7+y62mbAYyBidmJnDgyjVmjUzlhZBoTshPVSiMig8JQnIWsV7qQGWNcwCcEFsGcYq0tbnNuPj0IMMF7/wFcDSy21i45mns7lKMAIyIR47eW5RUVLCko4OniYur9h3eLchnDxenpLMrN5eKMDGLUxaxbDpTXsXpPOWv3lfPpvgo25Vfh8x/57+P4WCczhqVw/MgUZo5I5fgRqYxMj1NrmIhEnaHYhWxfcNvZQPuW43uPUM4I4AQCC1H+o8NfAKnB7WxjzHKAo1hXZntwm9fN60VEBhyHMZydlsbZaWn8aeJEnioqYklBAauqWlex91rLC6WlvFBaSlZMDDfm5LAoN5eZiYldlCwj0uIZkRbP5ScOB6C+ycf6AxWs3lvO2r3lrNlXTkVd82H31TX5+HhPGR/vKQsdS4uPCcyaNjqN44ancMKoVJI96nomIoNXtAaYdcHtrE7Otxxf383ycoOvcFKBed0sp0VacFvb5VUiIlEi2eXi1mHDuHXYMLbV1fFIQQGPFBRwsKl12GBxczO/O3CA3x04wKzERBbn5nJdTg4ZMfoyfSRxsU5OGZfBKeMygMAEC7tKalm7t5y1+yr4bH8F2wqrw7bSlNc18/aWIt7eUgQEup6NTo9nal5ym1cSw1PVUiMig0O0diFru5Dlidbazzqc79FClh3KmE/PxsC4CbTAjATOtNa+15PnB8tSFzIRGbB81vJWeTlLCgp4rriYxjB/n8Qaw6WZmSzOzeX8tDRc6mLWY/VNPjYeqmTdgUrW7a9g/YEK9pR2NpfN4ZI9LqbkJTMtGGim5iUzKScJT4wWLxWR/jfkupAFpzn+E3An8GdjzPnW2loAY8ztBMLLirbhxRhzG3Ab8Jy19kfH8nxjzBTgJOApa21jm+NZwP0Ewss6YNWxPEdEZCBzGsMF6elckJ5OeXMzfw92Mfu4ujp0TZO1PF1czNPFxeTFxvLlYBezqQkJEax5dIqLdYZmO2tRUdfEZ/srWL2nnPUHA8Gmsv7wrmcAVQ1ePt5dxse7W7ufOR2GcZkJ7VpqpuUlk5XkVmuNiAxYUdkCA2CM8QDLgVOAfGAlgamQTwGKgbnW2l1trr8b+AnwiLV2UTfKn08nLTBtzpUDq4PPGwbMJjB18wHgHGvttp5+vuBz1AIjIlFnY20tSwoKeKyggMLm8F+mT0lKYnFeHl/KyiJVXcx6jd9v2VdWx+b8KjbnV7Epv5rN+VUcrKg/8s1tZCTEhgJNS7iZkJ1IjFMtaCLSO4bcLGQtjDFxwI+A6wm0epQBrwE/ttYe6HDt3fRegBkGfB+YC4wBMoBGYBvwIvAHa215jz9Y63MUYEQkann9fl4rK+PhggJeLC2lOczfNx6HgyuCXczOTkvDqX/17xOVdc1sLqgKBZvN+dVsLaymyXvkBTdbxDgNE7KTmJqXxJTcJMZlJjIuK4GR6fEKNiJy1IZsgBnsFGBEZLAoaWrib0VFPFxQwGc1NWGvGel2c1Owi9mE+Ph+ruHQ4/X52V1Sy6ZgoGkJN0XV4df+6YzLYRiVHs+knCQm5yYxPjuRcZkJjM1MIMEdlT3VRaQfKMAMUgowIjIYfVZdzZKCAp4oKqKkky5mZ6SksDg3l2uyskhy6UtwfyqtaWwXaDblV7GjqAZvN9ap6Sg32cPYzATGZiUwPDWO7CQ3OckepuQlkZ3k6YPai0i0UIAZpBRgRGQwa/L7ebm0lIcLCniltBRfmGviHQ6uzspicW4uZ6Wm4lAXs4ho9PrYUVQTCjY7i2vYXVLL/rI6epBrABiRFsfYzARGpMUzNjOesZmJjM0MrI+jmdFEBj8FmEFKAUZEhoqCxkaeKCri4fx8NtaFnxp4rMfDzbm53JyTw5i4uH6uoYTT6PWxv6yOHUU1bDpUxY7iGnYV17K7pJbGoxhf01FusodR6fGMTI9ndEY8YzITGJMRz+j0BFLiNemDyGCgADNIKcCIyFBjrWV1sIvZ34qKqPB6w153dmoqi3JzuSori3in/rV+oPH7LflVDewKttTsLqmlsKqBoqpG9pTWUlLTdORCOpHscTEqIz4QcNLiGZ4Wx4i0OEakxTM8NU7jbkSihALMIKUAIyJDWYPPxwulpSwpKOCNsjLC/Xt+ktPJl7KzWZSby2nJyVq7JApYa8mvbGDDwUoOVtSzr6yO3SW17CmpZX95Pb6e9kkLSouPCYSa1NZwMzw1GHDS4kiJUwuOyECgADNIKcCIiAQcbGzk0YIClhQUsK0+/Jom4zwebsjJ4YacHCZrFrOo5PX5ya9sYF9ZXei1J9iCs6+sjrqmcCOljk5CrJPsZA9ZwQkFspPcpMbFMCUvmdEZ8eSleEjyKOSI9DUFmEFKAUZEpD1rLR9UVfFwQQFPFhVR7Qv/hXZ2YiI35uSwMDubXLe7n2spfcFaS2ltE/vK6thfVse+0joOlNdzoCKwPVRRT7Ovd77TJHlcDE+NIy/Fw7DUuODLw7CUwPvcFI/WvhE5Rgowg5QCjIhI5+p8Pp4tLubhggKWVVQQ7m8zB3BuWho35ORwRWampmQexPx+S1F1IweDgabldbCingPldRwsrz+miQXaMgayk9yBYJMSCDd5KYFgk5PsJjvJQ3ayG7dL47NEOqMAM0gpwIiIdM+Bhgb+XlTE44WFrKutDXtNnMPBZZmZ3JCdzQXp6cQ49C/oQ4m1lpKapsBkAtWBCQWKqhspCE42kF/ZQH5FA02+3gk5EBiPk5PsITvZQ06wy1pOsjuwH3yfmehWa44MSQowg5QCjIjI0dtYW8sThYU8UVjIvsbwq8pnxsRwVWYm12ZnMy81FacG/wuBVpzS2iYOVQS6pB2qbGh9H9wvrg7/O9VTxkBGgpuc5EDASfa4sASOTc5NZFJOEllJgaCj9XFkMFGAGaQUYEREes5vLasqK3m8sJB/FBdT3smUzNkxMVydlcW12dmckZKiMCNdavT6KKhs4FBFQ7tgk19ZT2FVI0VVDZTW9nya6K4keVxkJbrJTHIHtomxoXDTdpuRGKvuazLgKcAMUgowIiK9o9Hv57WyMh4vLOTFkhIaO/m7Lzc2NhBmsrI4PSUFh8KM9ECT109xTWNw7ZsGCqsC7wurGimqbgi9r6xv7rM6pMTFhA84iW7SE2JJjY8hNT6WtPgYUuJicKkbm/QzBZhBSgFGRKT3VXq9vFBSwpNFRbxRXo63k78HhwXDzJeys5mbnKwwI72uodkXHIsTCDQFVQ3UNnqpafSyq7iWfWW1FFU3UlHXd0GnRbLHRVpCLAmxLhLdLtISYkh0x5CX4mFEWhxpCbGkxsUQ43KQk+whLT6GuBin1l6SHlOAGaQUYERE+lZ5czPPl5TwVHExb3URZka43VwTbJk5RQtmSj9r8voprW2kuLqRkpqWbRPF1Y0U1zRS0mZb1RC+q2RfiHEakj0xJMcFXx4XyXGBFp3AcRcel5NEt4v0hFjcMQ5S42JJjnMRH+siPtZJXIwTh0P/Pw1FCjCDlAKMiEj/KW0JM0VFvF1eTmdLJo5qCTPZ2ZyUlKQwIwNKQ7OP0tpAuClpF3haQk4TZXVNVNQ1UVHXjNcf+e+BnhgHnhgnjc1+YpwmtNaOy2EYnhpHkieGBLeLRLeTBLeLBLeLpOA2we0KBSe3y6H/H6OIAswgpQAjIhIZJU1NPBfsZrasooLOJtYd4/FwbTDMzEpM1JcniSrWWqobvVTUNlNe10R5XRMNzT4q65sprm6kutFLYWUD+ZUNVNYHrmn0+vulS1tPGANuVyAMxcUEwk6Sx0VcjBO3y4Hb5SQuNvDy+y1NPj8OY0hPiGV4ahyJbhdxsc5QoIqLCVzrCd7ndjmIcQbOOdVqdMwUYAYpBRgRkcgramri2eJiniouZkUXYWacx8O12dlcm5XFCQozMog1egMhp6reS1VDc/B98NXgpao+eKyhmSavn6p6byj8lNc1UdPoJdq/fia5XbhDwSgQbGJchhing1ing1hXYNsyUUJs8JrANnBdy7Ut94X2nYFrXE4HLoch1uXAGHA5DPGxgdam2OA10fznjALMIKUAIyIysBQ0NvJssJvZu5WVdPY36IS4uFDLzMyEhKj+kiHS26y1NDT7qW3yUt/ko67JR12Tl0avn9pGL9UNXirqmqhv9lNU3TqxQU2jj9pGb2g/8N7Xq4uPRhunw+BqebUEH4cjcNxp8PossS4HyXExuBwGh4FYlwOPKzD2yGFgXFYid1w4pd/rfiwBxtUXFRIRERmMct1uvjl8ON8cPpz8xkaeDrbMvFdZ2e66HfX1/GLfPn6xbx+T4+K4Njubq7OyOE5hRgRjTKgrV29o6fbW0OyjodlPQ7OP+mZfKAw1en00egPH65p81Df5cDoMDc1+6pt9NHp9FFc3hu5raPZT3+QLlhc4NgCGCoXl81t8fktgedXORu51bdao3l2ctT8owIiIiPRAntvNv44Ywb+OGMHBljBTVMT7VVXtrttaX89P9+7lp3v3Msbj4bKMDC7LzOTMlBRcDq29IXKsPDFOPDF9t3CntYHxMk3ewKuuKRCYWoJRs8/S7PXT5PPTHLyu2WdpaPZRVttEZX0zXr+fJq9tc771+mZf6/Emnx9vcN/rt3h9fhqDzzUGmn29n6SicYp4BRgREZFjNNzt5jsjRvCdESPY39AQapn5sEOY2dPQwB8OHuQPBw+S5nJxcUYGl2VkcEF6Okku/ZUsMhAZY3C7nLhdgZCUAYyMUF2avH6qG5pDocfrt/j8gRDk9Vma/YEA5PX5afZbHAa8PktVQzPWBlpsmnyB1ii/DYSzrCR3hD5Nz2kMzACmMTAiItFtb0uYKSri4+rqTq+LNYZz0tK4LDOTi9PTGeHx9GMtRUT6n8bAiIiIDECjPR6+N3Ik3xs5kkONjbxYWsoLJSW8XV5OU5t/QGyyllfLyni1rAyA4xMS+GJGBhdnZHBycjLOKOziISLSVxRgRERE+sEwt5uvDxvG14cNo9rr5fWyMl4oLeXl0lLKve1XT19XW8u62lp+vm8fmTExXJSezsUZGVyQlkZqTEyEPoGIyMCgACMiItLPklwurs7O5ursbJr9ft6rrOSFkhJeKi1lZ0NDu2tLmpt5rLCQxwoLcQJnpKSEWmemxMdrVjMRGXI0BmYA0xgYEZGhxVrLtvp6Xgq2zKysrMTbxd/To91uLkhP54L0dM5JSyNFEwGISJTQGBgREZFBwBjD5Ph4JsfH872RI6n0enmjrIyXS0t5payM4ubmdtfvbWzk/vx87s/PxwnMTU4OBZrZSUkaOyMig5ICjIiIyACV4nJxTXY212Rn47eWT6qrQ60zn9bUtLvWB6yqqmJVVRV37dlDusvFuWlpoUAz3B19U6WKiISjACMiIhIFHMZwSnIypyQn89OxY8lvbOSN8nJeLyvjzfJySjq0zpR5vTwVXI8GYHp8fCjMnJmSQpyz7xb+ExHpSwowIiIiUSjP7ebm3Fxuzs3Fby2f1tTwelkZr5eV8X5V1WFjZzbW1bGxro77DhzA43AwLyWFc9PSWJCWxgmJiepuJiJRQwFGREQkyjmMYXZSErOTkviP0aOp8npZVlERCjS7Osxs1uD383p5Oa+XlwOQ6nJxVkoKC1JTOTstjRkJCTgUaERkgFKAERERGWSSXS4uy8zkssxMAHbU1QUCS1kZyyoqqPH52l1f4fXyz9JS/llaCkCGy8X81FQWpKWxIDWVqZquWUQGEAUYERGRQW5CfDwT4uP51vDhNPn9fFBVxRvBMPNJdfVh3c1KvV6eKSnhmZISAHJiYpgfbJ1ZkJrKhLg4BRoRiRgFGBERkSEk1uFgXmoq81JTAajxenmvspJlFRUsq6hgTXU1/g73FDY382RxMU8GJwQYHhvLvNRUTk9J4fSUFGYkJGgMjYj0GwUYERGRISzR5eLCjAwuzMgAoNLr5d1gmFlWUcG6mho6LqV5sKmJvxUV8beiIgCSnU7mJieHAs0pSUkkalFNEekj+tNFREREQlJcLi7JzOSS4PiZsuZmVrQJNBtqaw+7p8rn443yct4ITgrgBI5PTOS0lBRODwabkR5Pf34MERnEFGBERESkU+kxMVyRlcUVWVkAFDU1saKiglWVlayqquLT6mp8He7xAWtralhbU8OfDh4EYKTbzekpKZyWnMzc5GROTEzE5XD074cRkUFBAUZERES6LTs2lmuys7kmOxuAWp+Pj6uqQoHmg8pKKn0dIw3sb2zk70VF/D3Y7SzF6eTM1FTmBgPNyUlJJKnbmYh0g/6kEBERkR5LcDoD0y2npQHgs5ZNtbWhQLOqspLdHdahAaj0+XiptJSXglM3O4AZCQmcHAwzJyUlMSMhQa00InIYBRgRERHpNU5jOC4xkeMSE/mX4cMBONTYyPvBQPNhVRVrqqtp7jB1sx9YX1vL+tpaHsjPByDO4eDExEROTk7mpKQkTk5KYrymcBYZ8hRgREREpE8Nc7u5Ojubq4Pdzlqmbv6wqooPqqr4qKoqbLezer+f96uqeL+qKnQszeXipGALTUuwyXO7++2ziEjkKcCIiIhIv+o4dbPfWrbU1fFhVRWfVFfzcVUV62trD1tgE6Dc62034xkE1qWZlZTE7KQkZiUmMjspiWEKNSKDlgKMiIiIRJTDGKYlJDAtIYFb8vIAaPD5+KymJhBoqqv5pKqKrfX1Ye8/2NTEwdJSXgyOpwHIjY0NhZmW7Qi3W93PRAYBBRgREREZcDxOJ3NTUpibkhI6VtHczJqaGj5u01JzsKkp7P0FTU28UlbGK2VloWOZMTHMTkxs11oz2uPBoVAjElUUYERERCQqpMbEcE5aGucEZzwDyG9sZE11NWtralhTXc2a6upOQ01JczOvl5fzepvuZwkOB1MTEpiVmMgZKSnMSEhgRkICMZr9TGTAUoARERGRqJXndvNFt5svZmaGjhU2NbG2TahZW13N3sbGsPfX+v2srq5mdXU19wdnP/M4HJyQmMiJiYkcn5jI8QkJHJeYSILT2S+fSUS6pgAjIiIig0pObCwXZWRwUXCSAICSpiY+bQk0we2uMOvTADT4/XwYnPK5hQEmxMVxQptQc3xiosbViESAAoyIiIgMepmxsZyXns556emhY9VeL5vr6lhfU8N7lZV8VlPDprq6w9aoAbDA9vp6ttfX84/i4tDxNJeL4xMTOS4hIfSanpBAkktfsUT6iv7vEhERkSEpyeXi5ORkTk5O5tZhw4DA7Gframv5tLqaz2pqWFdby/qaGur8/rBllHu9LK+oYHlFRbvjYzwejguOp2kJNpPj4zW2RqQXKMCIiIiIBHmcTk5JTuaU5OTQMb+17KyvZ10w0KyrqeGzmhr2dzKuBmBPQwN7GhraTe0cYwyT4+NDgWZGQgJT4+MZ4/HgUrAR6baoDjDGmDjgR8BCYBRQBrwG/Nhae/AYyp0IrAc8wNvW2nM7uc4JfBu4BZgA1ADLgJ9Yazf39PkiIiIycDiMYWJ8PBPj47m6zfGy5mY+Dwaaz2tr2RB81fh8YctptjZ0zdI2x2OMYWJcHJPj45kSH8/k+HjGeTzMTEwkRV3RRA4Ttf9XGGM8wDvAXCAfeAEYAywGvmiMmWut3dXD4u8HulzC1xjjAP4BXAFUAC8DmcDVwMXGmAXW2o97+HwREREZ4NJjYpiXmsq81NTQMb+17G1o4PPa2lCo+bymhq319XjDjK2BQLDZVFfHprq6dscNMCPY9WxafDzHJyYyPSGB8WqxkSEuagMM8J8EwssHwPnW2hoAY8ztwL3AQ8D8oy3UGPOV4H33A1/r4tJbCISX7cCZ1trC4P1XAU8DTxhjplprvUdbBxEREYlODmMYGxfH2Lg4Lm0ztXOj38/WurpAoAmGmo11dextaCB8rAlMHNAShNqKNYZJwVAzNT6eqcGuaJPi4vBoqmcZAqIywBhjYoHbgrvfagkvANba+4wxNwPzjDGzrbVrjqLcHOC3wJvAUroOMLcHtz9oCS/B5z9jjPkncClwGfBMd58vIiIig5Pb4WBmYiIzExPbHa/3+dheX8/WurrAq76ebXV1bKuvp8Ib/t9Am9p0RWvLAYz1eJjSJtRMCYactJiYvvpoIv0uKgMMcDqQAuy01n4a5vzTwEzgEqDbAQb4AxAHfBMY0dlFxpixwFSgnkDXsXDPvzT4fAUYERERCSvO6QwbbPzWsqWujk9rathZX8/64AQCu+rrCT8fGviBnQ0N7Gxo4OWysnbncmJimBIfz6TgGJtJwTE3Yz0ezYwmUSdaA8zxwe3aTs63HJ/Z3QKNMV8AvgTcZa3dYYzpNMC0ef4Ga21zbzxfREREpIXDGKYlJDAtIaHd8Xqfj611dWxu+6qtZVt9fdj1a1oUNjdTWFnJisrKdsedwLhgmJnUYZsbG6tFOmVAitYAMyq4PdDJ+Zbjo7tTmDEmAfhfYCvw6wg8f2Mnp8Z3534REREZGuKcTk5ISuKEpKR2x71+P7sbGkKhZksw2Gyuq6Oqk1nRAHy0LtDZUZLTyaS4uHatNpPi45kYF0eyZkeTCIrW376Wdta6Ts63dApN6uR8Rz8jEDYWWGubIvB8ERERkR5zORyhqZ4vbXPcWktBUxObg2NstgXH2Gytq2N3Q0On3dEAqn0+1tTUsKam5rBzOTExTAyGmoltgs2EuDjiNJGA9LFoDTC9xhgzh8BaLo9aa5dHog7W2unhjgdbZqb1c3VERERkkDDGkOd2k+d2c3ZaWrtzTX4/O+vr2RacRGBbm0kEiprD9ZBv1dIl7b0OXdIARrjdoTDTsh0XF8c4j4cktdxIL4jW36KWfwqI7+R8S4fR6q4KMca4gP8jsI7Lv/f380VEREQiJdbhCMxW1mGcDUB5c3NodrS2AWd7fT11/q7abeBAYyMHGhtZVlFx2LnMmBjGejyM83gYGww14+LiGOvxMNLt1oQC0i3RGmD2BbedDbRvOb73COWMAE4ACoB/dBiolhrczjbGLAew1s7v5eeLiIiIDDhpMTGcHBPDycnJ7Y5baznU1BQKM6FtfT07jzCRAEBJczMlzc18Un34v/E6gdEeD9MSEjg+IYGJ8fGM8XgY7/EwzO3GoQkFJChaA8y64HZWJ+dbjq/vZnm5wVc4qcC8Tp4/wxgTE2YmsqN9voiIiMiAZ4xhuNvNcLebBR26pPmsZV9DAzuCkwK03e5paKDhCC03PmBXQwO7Ghp4qbS03Tl3cIHQ8cEWm7YtN2M9HhLVNW1Iidb/2quASmC8MeYEa+1nHc5fHdy+2FUh1to9QNg4b4yZDywD3rbWntvhvt3GmM0E1oK5GHi+J88XERERGSycwZAxNi6O8zqc8wcnE9jd0MCu+np2NTSwu2Xb0MDBxka6artpDK6Ls6Uu/PxJ2S1d0+LiGOPxMNrjYbTbHdh6PMRrYoFBJSoDjLW2yRjzJ+BO4M/GmPOttbUAxpjbCay/ssJaG1rE0hhzG3Ab8Jy19ke9UI37CIyf+Y0x5n1rbVHwOVcSWMRyB/BCLzxHREREJKo5jGGY280wt5vTU1IOO9/o97M72GLzeW0tn9fWsicYdo40oQBAUXMzRc3NfBSmaxoExt60DTRtA84Yj4dUl0tr3kSRqAwwQT8DzgVOA7YbY1YSmAr5FKAYuKXD9ZnAZCCvl57/EPAF4ApgizHm7eAz5gH1wI3WWm8vPUtERERk0HI7HExJSGBKQgKXZGa2O1ft9Qa6lgVbbHbW14dacvY0NNB0hHE30Dr2JtyU0BBY86Zjq03b/ZzYWI3BGUCiNsBYaxuMMQuAHwHXA5cDZcAS4MfW2s4Wmeyt5/uNMdcA3yEQlr5IYP2XZ4CfWGs39eXzRURERIaCJJeL4xMTOT4x8bBzfms51NjYLuDsbXk1NrK/oYHOl/FsVe3zsaG2lg21tWHPxxrDqC4CzgjNoNavjO1GapXIMMZsnDZt2rSNGzdGuioiIiIiUccXDDgtgWZvm4Czp6GBfY2N1B9hcoHucADD3G5Gu92MCgaaEW43I9tss9WK08706dPZtGnTps7WQ+xK1LbAiIiIiIh0xWkMIz0eRno8nBHmvLWW4ubmdq02ezu04lR4jzwiwE/r+jerqqrCXuMyhuGxsYH6BENNKOAEQ09WTIxCTjcowIiIiIjIkGSMITs2luzYWE7qsOZNiyqvt8uAU9DU1K1nea0N3N/Y2Ok1McaQFxsbmqp6ePD9qGDoGeV2k+d24xziIUcBRkRERESkE8kuF8clJnJcmDE4AA0+H/vbdFNraYnZ39AQ2DY2UuXrzkgcaLaWfY2N7Osi5LiCIWdYbGxgZrdOtmmDeGY1BRgRERERkR7yOJ1MjI9nYnx8p9dUeb0cDIaZ/R3CTcurppshx2tt6B46mTYawONwHDHkDIuNJSkKFwGNvhqLiIiIiESRZJeLZJeLqQkJnV7TEnIONjZysKmJg8GWnJbgs6+hgdJujMdp0eD3B2Zna2jo8rop8fFsPvnkbpc7ECjAiIiIiIhEWHdCTq3Px/6GBg41NXGosbHTbeNRzDKc4nT2RvX7lQKMiIiIiEgUSHA6Qwt+dsZaS5nX22XAOdTURH5jIz4C0z9HGwUYEREREZFBwhhDRkwMGTExHNfFdT5rKW5qwhuFa0IqwIiIiIiIDDFOY8iNwtYXCCwcKiIiIiIiEhUUYEREREREJGoowIiIiIiISNRQgBERERERkaihACMiIiIiIlFDAUZERERERKKGAoyIiIiIiPz/9u482I6qTuD498cWVg2rIFskIoNkVBBEEAWVmUEBC0GdIWVJHLVcKYkIzkyVipT+MeACWjo45QxQM1oIijiKExRjEFBkRmR1IYQJi0RZEiQLxIBn/jinTefmLu+9vHfv7fu+n6qufu+cPvee/nX37f513+7bGCYwkiRJkhrDBEaSJElSY5jASJIkSWoMExhJkiRJjWECI0mSJKkxTGAkSZIkNYYJjCRJkqTGMIGRJEmS1BgmMJIkSZIaI1JKg+6DOoiIJ2bMmLHD7NmzB90VSZIkadIsWbKEtWvXrkwpPWu8bU1ghlhE/A7YFnhgQF2oMqclA3r/UWRMJ5fxnHzGdPIZ08llPCefMZ1cxnNs9gbWpJR2H29DExh1FBF3AaSUDhp0X0aFMZ1cxnPyGdPJZ0wnl/GcfMZ0chnPqec9MJIkSZIawwRGkiRJUmOYwEiSJElqDBMYSZIkSY1hAiNJkiSpMXwKmSRJkqTG8AqMJEmSpMYwgZEkSZLUGCYwkiRJkhrDBEaSJElSY5jASJIkSWoMExhJkiRJjWECI0mSJKkxTGAkSZIkNYYJjDYSEdtExLkRcXdEPBURD0XEv0fEnoPu26BExLYRcVJE/FtE/KbEZXVE3BYRH4uI7bu0nRcRN0fEqohYHhHfi4gje7zfK8p0y0u7myPibZM/Z8MjInaOiIcjIkXEPT2mNaZdRMSuEfHpsq4+Web5log4v8P0J0bEdRHxRBkWRcTxPd7joIi4IiIeKe9xR0ScEREjt1+JiMMi4vLyWbguIh6PiOsj4u0REW2m3zwi5peYPFlidHlEHNjjfca9HIZVRLw0Iv4hIq6MiAfLdt3zl7P7tW1HxF4RcXFZpk+V/d0nImLr8c5rv4wnphGxWUS8MiLOi4ifR8TKiFgbEUsi4qKIeF6P9xr5mE50HW15jWurdhGxV5fpRj6efZdScnD48wBsDfwUSMBDwNeBn5X/Hwb2G3QfBxSXd5YYJOCXwOXAAuCJUvYrYLc27S4o9WuAq0qbdcDTwEkd3uuUUv8nYBHwDWBFeZ1PDzoWUxjjS8o8J+CeLtMZ0+5xfCnwaJm3O4HLgO8BS4Gn20x/Rpl2HfDfJaZrStkHOrzHEbVpflY+J5aV/y8HYtBxmMR4VutOAn5e5nVhiVcCvtoy/WbAlaVuRVnXFpV1bzXwsg7vM+7lMMxD6X9qHXq06cu2DTwfeKRMc0dZpkvK/zcAMwYdv02NaZnHapplwLfLevlgKXsCOGo6x3Qi62hL+3mlTbXf2ms6x7Pvy2/QHXAYrgH4ZNlAfgJsXyv/UClfNOg+DigupwFfBg5sKd8DuKXE5mstdceW8keB/WvlRwBrywfYzJY2OwF/KO1OrpU/B1hcyo8ZdDymIL6vLfP2ZbokMMa0Zxx3LTu91cAb2tS/rOX/A8qO9SngiFr5C0qM1wHPb2mzJXBvidv8Wvn25XMjAfMGHYtJiucWwO/LPM1tqTsQeKzUvbpWXp3suBt4Tq38lFK+GNhiU5fDsA/AR4BzgROB3cu8pS7T923bJh8AJuDClmVdJZ7nDDp+mxpTYDbwfeA11E4oADOAi8t83gdsOV1jOt51tKXtrmX7v4Z8cqhtAjOd4tn35TfoDjgMzwBsBTxeNo6D29TfVupeOui+DtNQdrCpfPhtVSv/Xik/o02bC0vdmS3lZ5fyq9q0eWOp+86g53mS47cNcA9wF7A/3RMYY9o9ll8q8/O+cU5/QZu6+aXuCy3lbynlt7Zpc0ipu2PQsZikeM4p8/PrDvXVOnd2reyXpeykNtN/u9SdsqnLoWlDr4PDfm3bwMtK+e9pOYtNPqj8I7CcliRzGIdeMe3SbhvW7+uPNqbjjyfwVeBJcqK4lM4JzLSN51QPI/ddZW2SVwDPBpaklH7Rpv4bZXxi/7rUCLeV8QxgZ8j3EZHPfMH6uNV1iuXxLfV1V5M/YI8dse/AfhzYD3gP+UxzW8a0uxKft5Kvvlw8xmbdYjPueKaUbiFfnZkTEbPG2IdhtnaM0z0GUO4rOJB8YHN1m+kmso6O/Odun7ftqs13UkobLN+U0u+B64EdgaPG1vvmSSk9Sb5CCPDclmpj2kNEHAfMBT6VUlrSY3LjOUVMYFT34jK+pUN9Vf6iPvSlSfYr43XksyKQvxIyA3gkpfRgmzadYtlxGaSU/ki+p2Fr8ldLGi8iXgScCVycUrq+x+TGtLtDgR2AX6SUnoyI10XEZyPiS+Xm+g0OVCJiJrBP+XejExYppQfIX+fZNyKeVauaTp8T95K/d35ARMytV5Qb8t9K/mrTt0pxFZs7U0rtkvGNYrMJy2GU9HPbnk7rb1uRH7Sxb/n3dy3VxrSLiNgO+Bfg18B5Y2hiPKeICYzqqp1oux1IvXzfDvXT1QfLeEHtbEnXWKaUVpMv4e8YETsAlIOTZ3drxwgtg7IT/Qo5DmePoYkx7e6FZfxwRFxF/krOfOC9wOeAeyLi1Nr0VTxXlNi10y420+ZzIqX0DPn+t8eBr5anOV0WEQuB28nz+tqUUnXiYiKxmehyGCX93LanzfrbxanAbuT75X5SFRrTMTkXmAW8pyQgHRnPqWUCo7rqUcBrOtRXO9cd+tCXRoiI1wPvIF99+WitqlcsYeN41h/FPB2WwenAYcBZKaXHxjC9Me1uxzJ+A3Ac8H7yQcos4NPk771fGhEvKdNNJJ5jaTcq8QQgpXQjcDT5aswhwN8CryY/UegHpbwykdhMdDmMkn5u29Nq/W0VEXuTn/YG8LGWrygZ0y4i4hDyCctLU0rXjaGJ8ZxCJjDSBEXEXwD/CQT5IPy2Hk1URMQ+5CfeXZdSumTA3RkV1ef5FuQDky+llB5JKd2XUjoLuIL8BLGzBtbDBipXrW4GHgAOJx9cvID82O8zgYURMWNgHZTGqHz96UpgF/JN5RcNuEuNERGbs/4bAx8ebG8EJjDa0Koy3rZD/XZlvLIPfRlqkX/UcwH5rPdnU0oXtkzSK5awcTxX1epGfRl8kfzUu/eMo40x7a4+r+1u4q/Kjm6ZfjzxHEu7UYknEbE/cCn5HpQTUko3p5RWp5QWp5TeDXyXfFXm70uTicRmosthlPRz2542629dRGxJPolxKPkRvXPbTGZMOzsDOJj8xMFHx9jGeE4hExjV3V/GnX5Ntiq/rw99GVoRsRP5+fr7kg8K252N6RrLciZsJvl77ysBUkpPkJ8X37Edo7MMTiBfHr8o8q+NL4qIReQfXQTYs1a+eykzpt1V/V+TUnqkTf3SMt6tjKt47lhi10672Eynz4m/I1+1WpBSWtWm/vIyflUZTyQ2E10Oo6Sf2/Z0Wn+BP99veCnwOuBW4MTyJLINGNOuTiQ/2vi0+j6r7LeqfdQVpew4MJ5TzQRGddVXoA7pUF+V396HvgyliNie/CvZLyRfin9XSvnh7C1+Q34E667lak2rTrHsuAzKGbQ55Mcu3t1a30AzyVcD6sPhpW7rWln1eElj2l31BKttOnylaacyXgWQUnqc9TvKg1snLt+V3wW4r+yIK9Ppc6I6UPhDh/qqvLr/qIrNnLJutdooNpuwHEZJP7ft6bT+Vr5AvnH/buBvyjrXiTHtLMgnK1r3W9Xn7cvL/7vX2hjPKWICo7obyTvk2bUbfeveVMbf6VuPhkg5KPw2+UemrgFOLU8p2kg5u7Ww/PvmNpN0iuXVLfV1J5AP5q9NKT01jq4PnZRStBuA55VJltTKl5Y2xrSLlNL95B1fsP5rYnVVWf1Rvd1iM+54RsTB5MeK31ktt4arHjF7aIf6w8p4KUBK6f+AX5EfmHB8m+knso6O/Odun7ftqs2JrYl+RDwHeCX50dg3jq33wy0iPgm8j5wk/1VK6eEeTYxpGymlY7rst6orIXuXsktqTY3nVJnKX8l0aN5AvrE6kTeM7WrlHyrliwbdxwHFZXPyFZcE/BjYdgxtji3TPwrsXys/gnzGZQUws6XNTuQkMgEn18p3AxaX8mMGHY8pjPOsMo/3GNMJxW9umZ/bgT1q5S8h/9hiAt5cKz8AeLrE7uW18v1LjNcBz295jy3JT95KwPxa+XbkR7ImYN6gYzFJ8TykzE8C3ttS93Ly1awEHFsrf2cpuxvYrVZ+cilfTMsvaE9kOTRtoMevnPdz2ybfA5KAC2plWwDfLOXnDDpekxTT+WV+ltVj2uM1p21Me8WzS7ulZR73Mp59XF6D7oDDcA3kswE3lQ3kIeDrtf8fBvYbdB8HFJcP1g5kriQ/gajdsEtLuwtKm9XAVeTf5lhHPlg5qcN7nQI8Q35M60LyjZcryut8ZtCxmOI4z6JLAmNMxxTDS8p8rSCfyVtY7ZiBf20zfXWQs67E8iry/UkJOL3DexxZm+am8jnxUPn/CiAGHYdJjOf5tW3/TvJ9LzeU9SkBX26ZfjPWn+xYXuLxo7LurQEO7/A+414OwzyQr0DdVBv+VFtfquH4ljZ92bZZnxhWyf5l5B8srU7ezRh0/DY1puSTFlX9T+i8zzpqusZ0Iutoh9dZSocEZjrFs+/Lb9AdcBi+gfz1h3OBe8jfS15Gvlm97cY5HQbgHNYfxHQbZrVpOw/4X/JOeQX5Hpoje7zfK8p0K0q7/wFOG3Qc+hDnWfRIYIxpzxgG8K5afFaRD2A6ziv5BtUfk59qs7L8fUKP9zkI+EbZyT5JPrifD2w26BhMQUzfSP7aaHU1ZHk5EDm1w/Sbk69a31li82g5aHlhj/cZ93IY1qFso70+L+d1aDfl2zawd9mvLSPv5xaT93tbDzp2kxFT4JgxTNt2GUyXmE50HW3zOkvpksBMl3j2e4gSJEmSJEkaet7EL0mSJKkxTGAkSZIkNYYJjCRJkqTGMIGRJEmS1BgmMJIkSZIawwRGkiRJUmOYwEiSJElqDBMYSZIkSY1hAiNJkiSpMUxgJEmSJDWGCYwkSZKkxjCBkSQNnYhIYxguGXQ/e4mIc0pf5w26L5I0KrYYdAckSeri0i51N/StF5KkoWECI0kaWimleYPugyRpuPgVMkmSJEmNYQIjSRoJ5V6TpRGxVUR8IiKWRMRTEXFvRJwbEVt3aLdzRJwfEYvL9MsjYkFE/HWX99o5Ij4VEXdExOqIeKL8fV5E7NGhzV9GxH9FxIrS5rqIOHKy5l+SpgsTGEnSKAngm8BZwC+Bq4GdgI8C342IzTeYOGJP4Gbgw8BWwFXAL4BjgWsiYv5GbxBxIHAr8E/ALsA1wLXlvc8CDm/Tr0OBm4BZZfrFwKuAH0bEnInPriRNP94DI0kaJfuQT87NSSndCxARuwILgdcCpwMX1Ka/CNgP+Brw9pTSH0ubo8iJxvkR8aOU0q2lfAvgW8Be5XU+UrUp9QcBT7Xp1/uBD6aUPl+b9nPAGcDZwNs2bbYlafrwCowkaWj1eIzySR2anVslLwAppUfIV0YAPlB77f2AE4BVwOn1RCSldAM5udmcnHxUTgYOAO4CPlxvU9rdlVJa0qZPN9aTl+KTZfyqDvMhSWrDKzCSpGHW7THK93cov6y1IKW0ICJWALMjYo+U0jLgqFK9IKW0vM3r/AfwIeCVtbJjy/grKaVnund9A99v06fHImI50PaeGUlSeyYwkqShNYHHKK9IKa3sUHcfsCPwXGBZGQMs7TB9Vb5nrWzvMm53laWbBzuUryTfoyNJGiO/QiZJUntpEl/rT5P4WpI0rZnASJJGyY4RsUOHun3K+KGW8b4dpp9Vxr+tlT1QxrMn1DtJ0iYzgZEkjZq3tBaU33TZCbi33P8CcEMZHxcRM9u8zlvL+Ppa2bVl/I6IcB8qSQPgh68kadR8PCJmVf9ExC7A+eXfL1bl5UllVwM7ABdGxJa1NkcA7wWeqbcBrgTuBuYA59XblHYHlaebSZKmiDfxS5KGVkRc0qX6/pTSx1rLgNuBuyLih8A64DXATOBHQOujjN9NvsLyNuDoiPgpsCtwDPkRymdWvwEDkFJ6OiJOAX4AnAnMLW0C2J+c2LwRuBdJ0pQwgZEkDbPTutTdBrQmMAl4Uymfy/onjn0R+FRK6ekNJk7ptxFxGPCPwEnk33lZA/wQ+ExKqd3jj++MiBeTf1vmDcDrgbXk5OmfgZvGN4uSpPGIlCbzISuSJA1GRCTgvpTSrEH3RZI0dbwHRpIkSVJjmMBIkiRJagwTGEmSJEmN4T0wkiRJkhrDKzCSJEmSGsMERpIkSVJjmMBIkiRJagwTGEmSJEmNYQIjSZIkqTFMYCRJkiQ1hgmMJEmSpMYwgZEkSZLUGCYwkiRJkhrDBEaSJElSY5jASJIkSWoMExhJkiRJjWECI0mSJKkx/h8xGsGS+6PPsQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "acc = dnn.fit (X.astype(np.float32), Y.astype(np.float32), X_val.astype(np.float32), Y_val.astype(np.float32))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "fe6f8ce9-8957-4250-9622-e570e00551d7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "Y_predict_train = dnn.predict_proba(X)\n",
    "Y_predict_val = dnn.predict_proba(X_val)\n",
    "\n",
    "axis = np.linspace(0,1,20)\n",
    "\n",
    "targets = ['isLL', 'isMIX', 'isTT', 'isBKG']\n",
    "predictions = ['isVBS']\n",
    "colors = ['r', 'g', 'y', 'b']\n",
    " \n",
    "Y_targets = {}\n",
    "Y_predictions = {}\n",
    "\n",
    "for target in targets:\n",
    "    #Y_targets[target] = dnn.predict_proba( df_val[ df_val[target]==1 ].values[:,0:NDIM] )\n",
    "    if target != 'isBKG':\n",
    "        Y_targets[target] = dnn.predict_proba( df[ df[target]==1 ].values[:,0:NDIM] )\n",
    "    else:\n",
    "        Y_targets[target] = dnn.predict_proba( df_val[ df_val[target]==1 ].values[:,0:NDIM] )\n",
    "    Y_predictions[target] = {}\n",
    "    for iprediction, prediction in enumerate(predictions):\n",
    "        Y_predictions[target][prediction] = Y_targets[target][:,iprediction]\n",
    "\n",
    "for prediction in predictions:\n",
    "    for target, mycolor in zip(targets, colors):\n",
    "        plt.hist(Y_predictions[target][prediction].numpy(), bins = axis, label = prediction + ' class - ' + target + ' sample', histtype='step', color=mycolor,  density=True, linewidth=2 )\n",
    "    plt.title('Validation dataset')\n",
    "    plt.xlabel(prediction + ' Probability')\n",
    "    plt.legend(frameon=False, bbox_to_anchor=(1.05, 1.))\n",
    "    #plt.yscale('log')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "db374a53-c7bd-4337-9216-68f5e143c8ed",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import classification_report, confusion_matrix\n",
    "import itertools\n",
    "#from matplotlib.backends.backend_pdf import PdfPages\n",
    "\n",
    "def plot_confusion_matrix(cm, classes,\n",
    "                          normalize=False,\n",
    "                          title='Confusion matrix',\n",
    "                          cmap=plt.cm.Blues):\n",
    "    \"\"\"\n",
    "    This function prints and plots the confusion matrix.\n",
    "    Normalization can be applied by setting `normalize=True`.\n",
    "    \"\"\"\n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, rotation=45, fontsize = 16)\n",
    "    plt.yticks(tick_marks, classes, fontsize = 16)\n",
    "\n",
    "    thresh = cm.max() / 1.2\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        plt.text(j, i, cm[i, j],\n",
    "                 horizontalalignment=\"center\",\n",
    "                 color=\"white\" if cm[i, j] > thresh else \"black\",fontsize=10)\n",
    "\n",
    "    plt.xlabel(\"Predicted label\", fontsize=16)\n",
    "    plt.ylabel(\"True label\", fontsize=16)\n",
    "\n",
    "    \n",
    "    plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "f3ec098a-fc62-4ab3-87a8-7ee81376b653",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classifier categorical accuracy =  0.8057170542635659\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "from scipy.stats import ks_2samp, anderson_ksamp\n",
    "\n",
    "X = df_val.values[:,0:NDIM]\n",
    "Y_true = df_val[['isVBS', 'isBKG']].values\n",
    "Y_pred = dnn.predict_proba(X)\n",
    "\n",
    "# The classifier categorical accuracy is a good and simple metrics to evaluate the overall classification performance\n",
    "Y_true_max = np.argmax(Y_true, axis=1)\n",
    "Y_pred_max = np.argmax(Y_pred, axis=1)\n",
    "print(\"Classifier categorical accuracy = \", accuracy_score(Y_true_max, Y_pred_max))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "ca1fc441-c2f5-4a0d-a4b2-7a451354a1ab",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "C = confusion_matrix(Y_true_max, Y_pred_max, normalize=\"true\")\n",
    "C = np.around(C, decimals=3)\n",
    "\n",
    "plt.figure()\n",
    "plot_confusion_matrix(C, classes=['isVBS', 'isBKG'], title='')\n",
    "plt.show(1) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "d364ee44-a77c-4a0b-9a4c-845fdf4aa08e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "339552\n"
     ]
    }
   ],
   "source": [
    "X    = df.values[:,0:NDIM]\n",
    "print(len(dnn.predict_proba(X)[...,0]))\n",
    "df['DNN_SigVsBkg'] = dnn.predict_proba(X)[...,0].numpy().tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "af3362d5-95ce-44db-bc16-b545f6e919c5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "339552\n",
      "18284\n",
      "27406\n",
      "27523\n",
      "27717\n"
     ]
    }
   ],
   "source": [
    "print(len(df))\n",
    "print(len(df[(df['DNN_SigVsBkg'] > 0.5) & (df['isBKG'] == 1)]))\n",
    "print(len(df[(df['DNN_SigVsBkg'] > 0.5) & (df['isLL'] == 1)]))\n",
    "print(len(df[(df['DNN_SigVsBkg'] > 0.5) & (df['isMIX'] == 1)]))\n",
    "print(len(df[(df['DNN_SigVsBkg'] > 0.5) & (df['isTT'] == 1)]))\n",
    "\n",
    "df_other = df[((df['isTT'] + df['isMIX'] + df['isLL']) == 1) & (df['DNN_SigVsBkg'] > 0.5)]\n",
    "\n",
    "df_notLL = df_other[df_other['notLL'] == 1].sample(n=int(len(df_other[df_other['notLL'] == 1])/2))\n",
    "df_notTT= df_other[df_other['notTT'] == 1].sample(n=int(len(df_other[df_other['notTT'] == 1])/2))\n",
    "\n",
    "df_LL = pd.concat([df_notLL, df_other[df_other['isLL'] == 1]])\n",
    "df_TT = pd.concat([df_notTT, df_other[df_other['isTT'] == 1]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "6a579013-794d-4daf-9141-a0d48c1d8a03",
   "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>CleanJet_eta_1</th>\n",
       "      <th>CleanJet_eta_2</th>\n",
       "      <th>CleanJet_phi_1</th>\n",
       "      <th>CleanJet_phi_2</th>\n",
       "      <th>CleanJet_pt_1</th>\n",
       "      <th>CleanJet_pt_2</th>\n",
       "      <th>Lepton_eta_1</th>\n",
       "      <th>Lepton_eta_2</th>\n",
       "      <th>Lepton_phi_1</th>\n",
       "      <th>Lepton_phi_2</th>\n",
       "      <th>Lepton_pt_1</th>\n",
       "      <th>Lepton_pt_2</th>\n",
       "      <th>Rpt</th>\n",
       "      <th>Zepp_l1</th>\n",
       "      <th>Zepp_l2</th>\n",
       "      <th>Zepp_ll</th>\n",
       "      <th>detajj</th>\n",
       "      <th>detall</th>\n",
       "      <th>dphijj</th>\n",
       "      <th>dphilep1jet1</th>\n",
       "      <th>dphilep1jet2</th>\n",
       "      <th>dphilep1jj</th>\n",
       "      <th>dphilep2jet1</th>\n",
       "      <th>dphilep2jet2</th>\n",
       "      <th>dphilep2jj</th>\n",
       "      <th>dphill</th>\n",
       "      <th>dphilljet</th>\n",
       "      <th>dphilljetjet</th>\n",
       "      <th>dphillmet</th>\n",
       "      <th>dphilmet1</th>\n",
       "      <th>dphilmet2</th>\n",
       "      <th>dr_l1j1</th>\n",
       "      <th>dr_l1j2</th>\n",
       "      <th>dr_l2j1</th>\n",
       "      <th>dr_l2j2</th>\n",
       "      <th>drll</th>\n",
       "      <th>ht</th>\n",
       "      <th>m2ljj30</th>\n",
       "      <th>mT2</th>\n",
       "      <th>mTi</th>\n",
       "      <th>m_l1j1</th>\n",
       "      <th>m_l1j2</th>\n",
       "      <th>m_l2j1</th>\n",
       "      <th>m_l2j2</th>\n",
       "      <th>mcoll</th>\n",
       "      <th>mcollWW</th>\n",
       "      <th>mjj</th>\n",
       "      <th>mll</th>\n",
       "      <th>mtw1</th>\n",
       "      <th>mtw2</th>\n",
       "      <th>phi_miss</th>\n",
       "      <th>proxyW_l1</th>\n",
       "      <th>proxyW_l2</th>\n",
       "      <th>pt_miss</th>\n",
       "      <th>ptll</th>\n",
       "      <th>recoil</th>\n",
       "      <th>yll</th>\n",
       "      <th>isLL</th>\n",
       "      <th>isMIX</th>\n",
       "      <th>isTT</th>\n",
       "      <th>isBKG</th>\n",
       "      <th>isVBS</th>\n",
       "      <th>notLL</th>\n",
       "      <th>notTT</th>\n",
       "      <th>DNN_SigVsBkg</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>279930</th>\n",
       "      <td>1.948730</td>\n",
       "      <td>-1.152100</td>\n",
       "      <td>2.152344</td>\n",
       "      <td>-0.391846</td>\n",
       "      <td>238.503708</td>\n",
       "      <td>164.499878</td>\n",
       "      <td>0.423950</td>\n",
       "      <td>2.357910</td>\n",
       "      <td>-1.071533</td>\n",
       "      <td>-1.261475</td>\n",
       "      <td>66.220924</td>\n",
       "      <td>31.500835</td>\n",
       "      <td>0.053169</td>\n",
       "      <td>0.025635</td>\n",
       "      <td>1.959595</td>\n",
       "      <td>0.992615</td>\n",
       "      <td>3.100830</td>\n",
       "      <td>1.933960</td>\n",
       "      <td>2.544189</td>\n",
       "      <td>3.059308</td>\n",
       "      <td>0.679688</td>\n",
       "      <td>2.489528</td>\n",
       "      <td>2.869367</td>\n",
       "      <td>0.869629</td>\n",
       "      <td>2.679470</td>\n",
       "      <td>0.189941</td>\n",
       "      <td>2.998169</td>\n",
       "      <td>2.550667</td>\n",
       "      <td>1.009906</td>\n",
       "      <td>1.071045</td>\n",
       "      <td>0.881104</td>\n",
       "      <td>3.566278</td>\n",
       "      <td>1.716365</td>\n",
       "      <td>3.438253</td>\n",
       "      <td>3.616134</td>\n",
       "      <td>1.943265</td>\n",
       "      <td>625.949768</td>\n",
       "      <td>1137.874146</td>\n",
       "      <td>44.952393</td>\n",
       "      <td>215.422806</td>\n",
       "      <td>330.012756</td>\n",
       "      <td>195.353134</td>\n",
       "      <td>179.867294</td>\n",
       "      <td>408.681000</td>\n",
       "      <td>228.756363</td>\n",
       "      <td>301.332520</td>\n",
       "      <td>969.845703</td>\n",
       "      <td>103.117554</td>\n",
       "      <td>80.914803</td>\n",
       "      <td>46.636993</td>\n",
       "      <td>-2.142578</td>\n",
       "      <td>87.715462</td>\n",
       "      <td>167.570145</td>\n",
       "      <td>94.921799</td>\n",
       "      <td>97.337090</td>\n",
       "      <td>168.268509</td>\n",
       "      <td>1.118872</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.941570</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>287376</th>\n",
       "      <td>-2.280273</td>\n",
       "      <td>3.856445</td>\n",
       "      <td>-1.225830</td>\n",
       "      <td>-1.775635</td>\n",
       "      <td>59.308693</td>\n",
       "      <td>46.522602</td>\n",
       "      <td>0.136292</td>\n",
       "      <td>-1.343750</td>\n",
       "      <td>1.527344</td>\n",
       "      <td>-2.946777</td>\n",
       "      <td>138.006836</td>\n",
       "      <td>53.251049</td>\n",
       "      <td>2.663461</td>\n",
       "      <td>-0.651794</td>\n",
       "      <td>-2.131836</td>\n",
       "      <td>1.391815</td>\n",
       "      <td>6.136719</td>\n",
       "      <td>1.480042</td>\n",
       "      <td>0.549805</td>\n",
       "      <td>2.753174</td>\n",
       "      <td>2.980207</td>\n",
       "      <td>2.994014</td>\n",
       "      <td>1.720947</td>\n",
       "      <td>1.171143</td>\n",
       "      <td>1.480107</td>\n",
       "      <td>1.809064</td>\n",
       "      <td>3.138754</td>\n",
       "      <td>2.897914</td>\n",
       "      <td>2.888572</td>\n",
       "      <td>2.497314</td>\n",
       "      <td>1.976807</td>\n",
       "      <td>3.663298</td>\n",
       "      <td>4.974858</td>\n",
       "      <td>1.959269</td>\n",
       "      <td>5.330441</td>\n",
       "      <td>2.337357</td>\n",
       "      <td>351.003235</td>\n",
       "      <td>1467.143677</td>\n",
       "      <td>156.609711</td>\n",
       "      <td>279.940338</td>\n",
       "      <td>328.260834</td>\n",
       "      <td>527.151917</td>\n",
       "      <td>101.844894</td>\n",
       "      <td>668.830750</td>\n",
       "      <td>124.280159</td>\n",
       "      <td>204.343628</td>\n",
       "      <td>1127.862427</td>\n",
       "      <td>193.460602</td>\n",
       "      <td>163.642517</td>\n",
       "      <td>89.497078</td>\n",
       "      <td>-0.969971</td>\n",
       "      <td>164.064926</td>\n",
       "      <td>118.415771</td>\n",
       "      <td>53.914032</td>\n",
       "      <td>135.692810</td>\n",
       "      <td>84.579102</td>\n",
       "      <td>-0.317337</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.557299</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>296881</th>\n",
       "      <td>1.522217</td>\n",
       "      <td>-2.565918</td>\n",
       "      <td>-0.861084</td>\n",
       "      <td>-2.545898</td>\n",
       "      <td>208.827850</td>\n",
       "      <td>89.481506</td>\n",
       "      <td>-1.294678</td>\n",
       "      <td>-0.000244</td>\n",
       "      <td>2.051270</td>\n",
       "      <td>1.834717</td>\n",
       "      <td>97.576210</td>\n",
       "      <td>58.721107</td>\n",
       "      <td>0.306631</td>\n",
       "      <td>-0.772827</td>\n",
       "      <td>0.521607</td>\n",
       "      <td>0.125610</td>\n",
       "      <td>4.088135</td>\n",
       "      <td>1.294434</td>\n",
       "      <td>1.684814</td>\n",
       "      <td>2.912354</td>\n",
       "      <td>1.686017</td>\n",
       "      <td>2.950035</td>\n",
       "      <td>2.695801</td>\n",
       "      <td>1.902570</td>\n",
       "      <td>3.116598</td>\n",
       "      <td>0.216553</td>\n",
       "      <td>2.831094</td>\n",
       "      <td>3.031295</td>\n",
       "      <td>0.299398</td>\n",
       "      <td>0.380658</td>\n",
       "      <td>0.164105</td>\n",
       "      <td>4.051753</td>\n",
       "      <td>4.769697</td>\n",
       "      <td>3.096002</td>\n",
       "      <td>5.076660</td>\n",
       "      <td>1.312423</td>\n",
       "      <td>631.595032</td>\n",
       "      <td>1304.823608</td>\n",
       "      <td>7.453078</td>\n",
       "      <td>182.325638</td>\n",
       "      <td>618.549072</td>\n",
       "      <td>189.041229</td>\n",
       "      <td>285.899109</td>\n",
       "      <td>268.903717</td>\n",
       "      <td>253.809097</td>\n",
       "      <td>307.800171</td>\n",
       "      <td>1058.500854</td>\n",
       "      <td>106.235214</td>\n",
       "      <td>38.844948</td>\n",
       "      <td>13.055244</td>\n",
       "      <td>1.670612</td>\n",
       "      <td>147.601746</td>\n",
       "      <td>13.055248</td>\n",
       "      <td>108.020454</td>\n",
       "      <td>155.438736</td>\n",
       "      <td>260.608643</td>\n",
       "      <td>-0.790069</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.756732</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>313225</th>\n",
       "      <td>3.906738</td>\n",
       "      <td>-1.919434</td>\n",
       "      <td>-1.656738</td>\n",
       "      <td>1.102051</td>\n",
       "      <td>98.938171</td>\n",
       "      <td>84.391541</td>\n",
       "      <td>-1.693604</td>\n",
       "      <td>2.060059</td>\n",
       "      <td>0.733398</td>\n",
       "      <td>-1.852051</td>\n",
       "      <td>140.987076</td>\n",
       "      <td>25.763031</td>\n",
       "      <td>0.435024</td>\n",
       "      <td>-2.687256</td>\n",
       "      <td>1.066406</td>\n",
       "      <td>0.810425</td>\n",
       "      <td>5.826172</td>\n",
       "      <td>3.753662</td>\n",
       "      <td>2.758789</td>\n",
       "      <td>2.390137</td>\n",
       "      <td>0.368652</td>\n",
       "      <td>1.399395</td>\n",
       "      <td>0.195312</td>\n",
       "      <td>2.954102</td>\n",
       "      <td>1.186054</td>\n",
       "      <td>2.585449</td>\n",
       "      <td>2.276440</td>\n",
       "      <td>1.285699</td>\n",
       "      <td>3.139721</td>\n",
       "      <td>3.029767</td>\n",
       "      <td>0.667969</td>\n",
       "      <td>6.089054</td>\n",
       "      <td>0.432324</td>\n",
       "      <td>1.856979</td>\n",
       "      <td>4.956115</td>\n",
       "      <td>4.557908</td>\n",
       "      <td>545.140320</td>\n",
       "      <td>2633.947754</td>\n",
       "      <td>215.645706</td>\n",
       "      <td>600.400940</td>\n",
       "      <td>1947.891235</td>\n",
       "      <td>49.638767</td>\n",
       "      <td>109.065102</td>\n",
       "      <td>347.500366</td>\n",
       "      <td>357.533905</td>\n",
       "      <td>440.629669</td>\n",
       "      <td>1687.566406</td>\n",
       "      <td>401.582947</td>\n",
       "      <td>300.778870</td>\n",
       "      <td>42.214081</td>\n",
       "      <td>-2.520020</td>\n",
       "      <td>415.528259</td>\n",
       "      <td>162.937607</td>\n",
       "      <td>160.921402</td>\n",
       "      <td>119.880615</td>\n",
       "      <td>41.041615</td>\n",
       "      <td>-0.608448</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.984433</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>337598</th>\n",
       "      <td>2.383301</td>\n",
       "      <td>-2.875977</td>\n",
       "      <td>-1.389404</td>\n",
       "      <td>2.119141</td>\n",
       "      <td>118.524651</td>\n",
       "      <td>89.139381</td>\n",
       "      <td>0.882812</td>\n",
       "      <td>-1.516602</td>\n",
       "      <td>0.420105</td>\n",
       "      <td>-0.477295</td>\n",
       "      <td>32.782749</td>\n",
       "      <td>16.477949</td>\n",
       "      <td>0.051129</td>\n",
       "      <td>1.129150</td>\n",
       "      <td>-1.270264</td>\n",
       "      <td>0.070557</td>\n",
       "      <td>5.259277</td>\n",
       "      <td>2.399414</td>\n",
       "      <td>2.774640</td>\n",
       "      <td>1.809509</td>\n",
       "      <td>1.699036</td>\n",
       "      <td>2.545335</td>\n",
       "      <td>0.912109</td>\n",
       "      <td>2.596436</td>\n",
       "      <td>1.647935</td>\n",
       "      <td>0.897400</td>\n",
       "      <td>1.518837</td>\n",
       "      <td>2.254663</td>\n",
       "      <td>1.391693</td>\n",
       "      <td>1.101021</td>\n",
       "      <td>1.998420</td>\n",
       "      <td>2.350700</td>\n",
       "      <td>4.124950</td>\n",
       "      <td>4.005145</td>\n",
       "      <td>2.930764</td>\n",
       "      <td>2.561740</td>\n",
       "      <td>339.462921</td>\n",
       "      <td>1519.088379</td>\n",
       "      <td>26.618164</td>\n",
       "      <td>134.764969</td>\n",
       "      <td>143.575485</td>\n",
       "      <td>355.423462</td>\n",
       "      <td>307.582367</td>\n",
       "      <td>93.615479</td>\n",
       "      <td>110.806091</td>\n",
       "      <td>200.008820</td>\n",
       "      <td>1432.898682</td>\n",
       "      <td>72.982231</td>\n",
       "      <td>54.423107</td>\n",
       "      <td>62.033714</td>\n",
       "      <td>1.521126</td>\n",
       "      <td>72.188995</td>\n",
       "      <td>87.315178</td>\n",
       "      <td>82.538200</td>\n",
       "      <td>44.944496</td>\n",
       "      <td>100.768471</td>\n",
       "      <td>-0.033654</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.978351</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>264090</th>\n",
       "      <td>1.201660</td>\n",
       "      <td>-2.193848</td>\n",
       "      <td>-2.059082</td>\n",
       "      <td>0.651245</td>\n",
       "      <td>517.302185</td>\n",
       "      <td>172.586716</td>\n",
       "      <td>-0.036133</td>\n",
       "      <td>-1.137207</td>\n",
       "      <td>2.119629</td>\n",
       "      <td>0.318481</td>\n",
       "      <td>67.197662</td>\n",
       "      <td>65.866066</td>\n",
       "      <td>0.049575</td>\n",
       "      <td>0.459961</td>\n",
       "      <td>-0.641113</td>\n",
       "      <td>0.090576</td>\n",
       "      <td>3.395508</td>\n",
       "      <td>1.101074</td>\n",
       "      <td>2.710327</td>\n",
       "      <td>2.104474</td>\n",
       "      <td>1.468384</td>\n",
       "      <td>2.301980</td>\n",
       "      <td>2.377563</td>\n",
       "      <td>0.332764</td>\n",
       "      <td>2.180058</td>\n",
       "      <td>1.801147</td>\n",
       "      <td>2.992423</td>\n",
       "      <td>3.093256</td>\n",
       "      <td>0.009411</td>\n",
       "      <td>0.897360</td>\n",
       "      <td>0.903788</td>\n",
       "      <td>4.358183</td>\n",
       "      <td>2.609959</td>\n",
       "      <td>3.335132</td>\n",
       "      <td>1.107800</td>\n",
       "      <td>2.111042</td>\n",
       "      <td>1137.687988</td>\n",
       "      <td>1879.810059</td>\n",
       "      <td>0.000015</td>\n",
       "      <td>279.976807</td>\n",
       "      <td>408.123932</td>\n",
       "      <td>315.497681</td>\n",
       "      <td>638.062256</td>\n",
       "      <td>124.151871</td>\n",
       "      <td>510.554352</td>\n",
       "      <td>536.907227</td>\n",
       "      <td>1682.565430</td>\n",
       "      <td>129.629059</td>\n",
       "      <td>126.166992</td>\n",
       "      <td>125.743950</td>\n",
       "      <td>1.222269</td>\n",
       "      <td>126.276382</td>\n",
       "      <td>213.629593</td>\n",
       "      <td>314.735321</td>\n",
       "      <td>82.660515</td>\n",
       "      <td>397.392944</td>\n",
       "      <td>-0.581657</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.978672</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>264091</th>\n",
       "      <td>-3.622070</td>\n",
       "      <td>3.979492</td>\n",
       "      <td>-1.494629</td>\n",
       "      <td>2.269531</td>\n",
       "      <td>53.671650</td>\n",
       "      <td>32.433601</td>\n",
       "      <td>-0.359924</td>\n",
       "      <td>0.463867</td>\n",
       "      <td>-0.784546</td>\n",
       "      <td>1.630859</td>\n",
       "      <td>134.975113</td>\n",
       "      <td>68.708046</td>\n",
       "      <td>5.327472</td>\n",
       "      <td>-0.538635</td>\n",
       "      <td>0.285156</td>\n",
       "      <td>0.126740</td>\n",
       "      <td>7.601562</td>\n",
       "      <td>0.823792</td>\n",
       "      <td>2.519025</td>\n",
       "      <td>0.710083</td>\n",
       "      <td>3.054077</td>\n",
       "      <td>1.315548</td>\n",
       "      <td>3.125488</td>\n",
       "      <td>0.638672</td>\n",
       "      <td>2.552232</td>\n",
       "      <td>2.415405</td>\n",
       "      <td>1.209642</td>\n",
       "      <td>1.815107</td>\n",
       "      <td>3.080272</td>\n",
       "      <td>2.703354</td>\n",
       "      <td>1.164426</td>\n",
       "      <td>3.338535</td>\n",
       "      <td>5.306404</td>\n",
       "      <td>5.144275</td>\n",
       "      <td>3.573167</td>\n",
       "      <td>2.552022</td>\n",
       "      <td>402.349548</td>\n",
       "      <td>2084.974121</td>\n",
       "      <td>141.421158</td>\n",
       "      <td>331.679749</td>\n",
       "      <td>422.508667</td>\n",
       "      <td>586.889526</td>\n",
       "      <td>476.386536</td>\n",
       "      <td>267.383636</td>\n",
       "      <td>125.553528</td>\n",
       "      <td>214.510941</td>\n",
       "      <td>1867.331055</td>\n",
       "      <td>197.670792</td>\n",
       "      <td>240.624924</td>\n",
       "      <td>96.714386</td>\n",
       "      <td>2.795286</td>\n",
       "      <td>244.724060</td>\n",
       "      <td>105.108711</td>\n",
       "      <td>112.561134</td>\n",
       "      <td>95.240273</td>\n",
       "      <td>18.447498</td>\n",
       "      <td>-0.075627</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.982667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>264093</th>\n",
       "      <td>-1.057617</td>\n",
       "      <td>2.513672</td>\n",
       "      <td>0.537598</td>\n",
       "      <td>2.613770</td>\n",
       "      <td>146.862488</td>\n",
       "      <td>132.743088</td>\n",
       "      <td>-1.460449</td>\n",
       "      <td>0.496094</td>\n",
       "      <td>-0.962769</td>\n",
       "      <td>3.086914</td>\n",
       "      <td>48.502495</td>\n",
       "      <td>29.294687</td>\n",
       "      <td>0.072884</td>\n",
       "      <td>-2.188477</td>\n",
       "      <td>-0.231934</td>\n",
       "      <td>1.210205</td>\n",
       "      <td>3.571289</td>\n",
       "      <td>1.956543</td>\n",
       "      <td>2.076172</td>\n",
       "      <td>1.500366</td>\n",
       "      <td>2.706647</td>\n",
       "      <td>2.453008</td>\n",
       "      <td>2.549316</td>\n",
       "      <td>0.473145</td>\n",
       "      <td>1.596674</td>\n",
       "      <td>2.233503</td>\n",
       "      <td>2.148778</td>\n",
       "      <td>3.101420</td>\n",
       "      <td>0.738664</td>\n",
       "      <td>1.387076</td>\n",
       "      <td>0.846427</td>\n",
       "      <td>1.553503</td>\n",
       "      <td>5.346519</td>\n",
       "      <td>2.985470</td>\n",
       "      <td>2.072315</td>\n",
       "      <td>2.969275</td>\n",
       "      <td>478.592651</td>\n",
       "      <td>1070.701538</td>\n",
       "      <td>9.757400</td>\n",
       "      <td>175.923004</td>\n",
       "      <td>121.256371</td>\n",
       "      <td>595.722046</td>\n",
       "      <td>169.168488</td>\n",
       "      <td>152.760147</td>\n",
       "      <td>212.364609</td>\n",
       "      <td>267.833893</td>\n",
       "      <td>845.109436</td>\n",
       "      <td>109.551781</td>\n",
       "      <td>80.886276</td>\n",
       "      <td>40.385475</td>\n",
       "      <td>-2.349844</td>\n",
       "      <td>129.263550</td>\n",
       "      <td>47.310089</td>\n",
       "      <td>82.521683</td>\n",
       "      <td>38.239880</td>\n",
       "      <td>113.747307</td>\n",
       "      <td>-0.670103</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.607159</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>264094</th>\n",
       "      <td>3.563477</td>\n",
       "      <td>-1.228027</td>\n",
       "      <td>1.822754</td>\n",
       "      <td>-1.988281</td>\n",
       "      <td>89.372330</td>\n",
       "      <td>53.334095</td>\n",
       "      <td>0.621826</td>\n",
       "      <td>1.522217</td>\n",
       "      <td>0.476501</td>\n",
       "      <td>-0.988281</td>\n",
       "      <td>54.206814</td>\n",
       "      <td>47.276127</td>\n",
       "      <td>0.537635</td>\n",
       "      <td>-0.545898</td>\n",
       "      <td>0.354492</td>\n",
       "      <td>0.095703</td>\n",
       "      <td>4.791504</td>\n",
       "      <td>0.900391</td>\n",
       "      <td>2.472150</td>\n",
       "      <td>1.346252</td>\n",
       "      <td>2.464783</td>\n",
       "      <td>1.954317</td>\n",
       "      <td>2.811035</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>2.864085</td>\n",
       "      <td>1.464783</td>\n",
       "      <td>2.017308</td>\n",
       "      <td>2.625373</td>\n",
       "      <td>2.702906</td>\n",
       "      <td>2.909223</td>\n",
       "      <td>1.909180</td>\n",
       "      <td>3.235074</td>\n",
       "      <td>3.081738</td>\n",
       "      <td>3.473998</td>\n",
       "      <td>2.926404</td>\n",
       "      <td>1.719387</td>\n",
       "      <td>296.687592</td>\n",
       "      <td>884.042175</td>\n",
       "      <td>78.803207</td>\n",
       "      <td>171.497589</td>\n",
       "      <td>300.274261</td>\n",
       "      <td>153.201935</td>\n",
       "      <td>203.372360</td>\n",
       "      <td>192.481842</td>\n",
       "      <td>28.694359</td>\n",
       "      <td>173.328766</td>\n",
       "      <td>763.780212</td>\n",
       "      <td>82.491333</td>\n",
       "      <td>98.950401</td>\n",
       "      <td>75.924240</td>\n",
       "      <td>-2.897461</td>\n",
       "      <td>103.835793</td>\n",
       "      <td>108.741577</td>\n",
       "      <td>45.771694</td>\n",
       "      <td>75.602600</td>\n",
       "      <td>39.309307</td>\n",
       "      <td>1.043189</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.871962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>264095</th>\n",
       "      <td>2.654297</td>\n",
       "      <td>-3.865234</td>\n",
       "      <td>-1.350830</td>\n",
       "      <td>1.639160</td>\n",
       "      <td>76.704498</td>\n",
       "      <td>75.138702</td>\n",
       "      <td>0.810059</td>\n",
       "      <td>-2.089355</td>\n",
       "      <td>2.219727</td>\n",
       "      <td>3.023926</td>\n",
       "      <td>44.667046</td>\n",
       "      <td>34.342400</td>\n",
       "      <td>0.266154</td>\n",
       "      <td>1.415527</td>\n",
       "      <td>-1.483887</td>\n",
       "      <td>0.034180</td>\n",
       "      <td>6.519531</td>\n",
       "      <td>2.899414</td>\n",
       "      <td>2.989990</td>\n",
       "      <td>2.712629</td>\n",
       "      <td>0.580566</td>\n",
       "      <td>2.210515</td>\n",
       "      <td>1.908430</td>\n",
       "      <td>1.384766</td>\n",
       "      <td>3.014714</td>\n",
       "      <td>0.804199</td>\n",
       "      <td>2.366045</td>\n",
       "      <td>2.557099</td>\n",
       "      <td>2.918366</td>\n",
       "      <td>3.018235</td>\n",
       "      <td>2.460751</td>\n",
       "      <td>4.018717</td>\n",
       "      <td>4.711202</td>\n",
       "      <td>6.452962</td>\n",
       "      <td>2.251960</td>\n",
       "      <td>3.008877</td>\n",
       "      <td>291.497864</td>\n",
       "      <td>2156.117188</td>\n",
       "      <td>83.568184</td>\n",
       "      <td>242.511353</td>\n",
       "      <td>168.897156</td>\n",
       "      <td>595.465271</td>\n",
       "      <td>551.738770</td>\n",
       "      <td>122.009430</td>\n",
       "      <td>57.777363</td>\n",
       "      <td>173.242004</td>\n",
       "      <td>1980.171997</td>\n",
       "      <td>160.681625</td>\n",
       "      <td>103.895157</td>\n",
       "      <td>86.035484</td>\n",
       "      <td>-0.798509</td>\n",
       "      <td>112.565910</td>\n",
       "      <td>142.555573</td>\n",
       "      <td>60.645222</td>\n",
       "      <td>72.819946</td>\n",
       "      <td>19.166916</td>\n",
       "      <td>-0.522072</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.975956</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>55026 rows × 65 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        CleanJet_eta_1  CleanJet_eta_2  CleanJet_phi_1  CleanJet_phi_2  \\\n",
       "279930        1.948730       -1.152100        2.152344       -0.391846   \n",
       "287376       -2.280273        3.856445       -1.225830       -1.775635   \n",
       "296881        1.522217       -2.565918       -0.861084       -2.545898   \n",
       "313225        3.906738       -1.919434       -1.656738        1.102051   \n",
       "337598        2.383301       -2.875977       -1.389404        2.119141   \n",
       "...                ...             ...             ...             ...   \n",
       "264090        1.201660       -2.193848       -2.059082        0.651245   \n",
       "264091       -3.622070        3.979492       -1.494629        2.269531   \n",
       "264093       -1.057617        2.513672        0.537598        2.613770   \n",
       "264094        3.563477       -1.228027        1.822754       -1.988281   \n",
       "264095        2.654297       -3.865234       -1.350830        1.639160   \n",
       "\n",
       "        CleanJet_pt_1  CleanJet_pt_2  Lepton_eta_1  Lepton_eta_2  \\\n",
       "279930     238.503708     164.499878      0.423950      2.357910   \n",
       "287376      59.308693      46.522602      0.136292     -1.343750   \n",
       "296881     208.827850      89.481506     -1.294678     -0.000244   \n",
       "313225      98.938171      84.391541     -1.693604      2.060059   \n",
       "337598     118.524651      89.139381      0.882812     -1.516602   \n",
       "...               ...            ...           ...           ...   \n",
       "264090     517.302185     172.586716     -0.036133     -1.137207   \n",
       "264091      53.671650      32.433601     -0.359924      0.463867   \n",
       "264093     146.862488     132.743088     -1.460449      0.496094   \n",
       "264094      89.372330      53.334095      0.621826      1.522217   \n",
       "264095      76.704498      75.138702      0.810059     -2.089355   \n",
       "\n",
       "        Lepton_phi_1  Lepton_phi_2  Lepton_pt_1  Lepton_pt_2       Rpt  \\\n",
       "279930     -1.071533     -1.261475    66.220924    31.500835  0.053169   \n",
       "287376      1.527344     -2.946777   138.006836    53.251049  2.663461   \n",
       "296881      2.051270      1.834717    97.576210    58.721107  0.306631   \n",
       "313225      0.733398     -1.852051   140.987076    25.763031  0.435024   \n",
       "337598      0.420105     -0.477295    32.782749    16.477949  0.051129   \n",
       "...              ...           ...          ...          ...       ...   \n",
       "264090      2.119629      0.318481    67.197662    65.866066  0.049575   \n",
       "264091     -0.784546      1.630859   134.975113    68.708046  5.327472   \n",
       "264093     -0.962769      3.086914    48.502495    29.294687  0.072884   \n",
       "264094      0.476501     -0.988281    54.206814    47.276127  0.537635   \n",
       "264095      2.219727      3.023926    44.667046    34.342400  0.266154   \n",
       "\n",
       "         Zepp_l1   Zepp_l2   Zepp_ll    detajj    detall    dphijj  \\\n",
       "279930  0.025635  1.959595  0.992615  3.100830  1.933960  2.544189   \n",
       "287376 -0.651794 -2.131836  1.391815  6.136719  1.480042  0.549805   \n",
       "296881 -0.772827  0.521607  0.125610  4.088135  1.294434  1.684814   \n",
       "313225 -2.687256  1.066406  0.810425  5.826172  3.753662  2.758789   \n",
       "337598  1.129150 -1.270264  0.070557  5.259277  2.399414  2.774640   \n",
       "...          ...       ...       ...       ...       ...       ...   \n",
       "264090  0.459961 -0.641113  0.090576  3.395508  1.101074  2.710327   \n",
       "264091 -0.538635  0.285156  0.126740  7.601562  0.823792  2.519025   \n",
       "264093 -2.188477 -0.231934  1.210205  3.571289  1.956543  2.076172   \n",
       "264094 -0.545898  0.354492  0.095703  4.791504  0.900391  2.472150   \n",
       "264095  1.415527 -1.483887  0.034180  6.519531  2.899414  2.989990   \n",
       "\n",
       "        dphilep1jet1  dphilep1jet2  dphilep1jj  dphilep2jet1  dphilep2jet2  \\\n",
       "279930      3.059308      0.679688    2.489528      2.869367      0.869629   \n",
       "287376      2.753174      2.980207    2.994014      1.720947      1.171143   \n",
       "296881      2.912354      1.686017    2.950035      2.695801      1.902570   \n",
       "313225      2.390137      0.368652    1.399395      0.195312      2.954102   \n",
       "337598      1.809509      1.699036    2.545335      0.912109      2.596436   \n",
       "...              ...           ...         ...           ...           ...   \n",
       "264090      2.104474      1.468384    2.301980      2.377563      0.332764   \n",
       "264091      0.710083      3.054077    1.315548      3.125488      0.638672   \n",
       "264093      1.500366      2.706647    2.453008      2.549316      0.473145   \n",
       "264094      1.346252      2.464783    1.954317      2.811035      1.000000   \n",
       "264095      2.712629      0.580566    2.210515      1.908430      1.384766   \n",
       "\n",
       "        dphilep2jj    dphill  dphilljet  dphilljetjet  dphillmet  dphilmet1  \\\n",
       "279930    2.679470  0.189941   2.998169      2.550667   1.009906   1.071045   \n",
       "287376    1.480107  1.809064   3.138754      2.897914   2.888572   2.497314   \n",
       "296881    3.116598  0.216553   2.831094      3.031295   0.299398   0.380658   \n",
       "313225    1.186054  2.585449   2.276440      1.285699   3.139721   3.029767   \n",
       "337598    1.647935  0.897400   1.518837      2.254663   1.391693   1.101021   \n",
       "...            ...       ...        ...           ...        ...        ...   \n",
       "264090    2.180058  1.801147   2.992423      3.093256   0.009411   0.897360   \n",
       "264091    2.552232  2.415405   1.209642      1.815107   3.080272   2.703354   \n",
       "264093    1.596674  2.233503   2.148778      3.101420   0.738664   1.387076   \n",
       "264094    2.864085  1.464783   2.017308      2.625373   2.702906   2.909223   \n",
       "264095    3.014714  0.804199   2.366045      2.557099   2.918366   3.018235   \n",
       "\n",
       "        dphilmet2   dr_l1j1   dr_l1j2   dr_l2j1   dr_l2j2      drll  \\\n",
       "279930   0.881104  3.566278  1.716365  3.438253  3.616134  1.943265   \n",
       "287376   1.976807  3.663298  4.974858  1.959269  5.330441  2.337357   \n",
       "296881   0.164105  4.051753  4.769697  3.096002  5.076660  1.312423   \n",
       "313225   0.667969  6.089054  0.432324  1.856979  4.956115  4.557908   \n",
       "337598   1.998420  2.350700  4.124950  4.005145  2.930764  2.561740   \n",
       "...           ...       ...       ...       ...       ...       ...   \n",
       "264090   0.903788  4.358183  2.609959  3.335132  1.107800  2.111042   \n",
       "264091   1.164426  3.338535  5.306404  5.144275  3.573167  2.552022   \n",
       "264093   0.846427  1.553503  5.346519  2.985470  2.072315  2.969275   \n",
       "264094   1.909180  3.235074  3.081738  3.473998  2.926404  1.719387   \n",
       "264095   2.460751  4.018717  4.711202  6.452962  2.251960  3.008877   \n",
       "\n",
       "                 ht      m2ljj30         mT2         mTi       m_l1j1  \\\n",
       "279930   625.949768  1137.874146   44.952393  215.422806   330.012756   \n",
       "287376   351.003235  1467.143677  156.609711  279.940338   328.260834   \n",
       "296881   631.595032  1304.823608    7.453078  182.325638   618.549072   \n",
       "313225   545.140320  2633.947754  215.645706  600.400940  1947.891235   \n",
       "337598   339.462921  1519.088379   26.618164  134.764969   143.575485   \n",
       "...             ...          ...         ...         ...          ...   \n",
       "264090  1137.687988  1879.810059    0.000015  279.976807   408.123932   \n",
       "264091   402.349548  2084.974121  141.421158  331.679749   422.508667   \n",
       "264093   478.592651  1070.701538    9.757400  175.923004   121.256371   \n",
       "264094   296.687592   884.042175   78.803207  171.497589   300.274261   \n",
       "264095   291.497864  2156.117188   83.568184  242.511353   168.897156   \n",
       "\n",
       "            m_l1j2      m_l2j1      m_l2j2       mcoll     mcollWW  \\\n",
       "279930  195.353134  179.867294  408.681000  228.756363  301.332520   \n",
       "287376  527.151917  101.844894  668.830750  124.280159  204.343628   \n",
       "296881  189.041229  285.899109  268.903717  253.809097  307.800171   \n",
       "313225   49.638767  109.065102  347.500366  357.533905  440.629669   \n",
       "337598  355.423462  307.582367   93.615479  110.806091  200.008820   \n",
       "...            ...         ...         ...         ...         ...   \n",
       "264090  315.497681  638.062256  124.151871  510.554352  536.907227   \n",
       "264091  586.889526  476.386536  267.383636  125.553528  214.510941   \n",
       "264093  595.722046  169.168488  152.760147  212.364609  267.833893   \n",
       "264094  153.201935  203.372360  192.481842   28.694359  173.328766   \n",
       "264095  595.465271  551.738770  122.009430   57.777363  173.242004   \n",
       "\n",
       "                mjj         mll        mtw1        mtw2  phi_miss   proxyW_l1  \\\n",
       "279930   969.845703  103.117554   80.914803   46.636993 -2.142578   87.715462   \n",
       "287376  1127.862427  193.460602  163.642517   89.497078 -0.969971  164.064926   \n",
       "296881  1058.500854  106.235214   38.844948   13.055244  1.670612  147.601746   \n",
       "313225  1687.566406  401.582947  300.778870   42.214081 -2.520020  415.528259   \n",
       "337598  1432.898682   72.982231   54.423107   62.033714  1.521126   72.188995   \n",
       "...             ...         ...         ...         ...       ...         ...   \n",
       "264090  1682.565430  129.629059  126.166992  125.743950  1.222269  126.276382   \n",
       "264091  1867.331055  197.670792  240.624924   96.714386  2.795286  244.724060   \n",
       "264093   845.109436  109.551781   80.886276   40.385475 -2.349844  129.263550   \n",
       "264094   763.780212   82.491333   98.950401   75.924240 -2.897461  103.835793   \n",
       "264095  1980.171997  160.681625  103.895157   86.035484 -0.798509  112.565910   \n",
       "\n",
       "         proxyW_l2     pt_miss        ptll      recoil       yll  isLL  isMIX  \\\n",
       "279930  167.570145   94.921799   97.337090  168.268509  1.118872   0.0    1.0   \n",
       "287376  118.415771   53.914032  135.692810   84.579102 -0.317337   0.0    1.0   \n",
       "296881   13.055248  108.020454  155.438736  260.608643 -0.790069   0.0    1.0   \n",
       "313225  162.937607  160.921402  119.880615   41.041615 -0.608448   0.0    0.0   \n",
       "337598   87.315178   82.538200   44.944496  100.768471 -0.033654   0.0    0.0   \n",
       "...            ...         ...         ...         ...       ...   ...    ...   \n",
       "264090  213.629593  314.735321   82.660515  397.392944 -0.581657   1.0    0.0   \n",
       "264091  105.108711  112.561134   95.240273   18.447498 -0.075627   1.0    0.0   \n",
       "264093   47.310089   82.521683   38.239880  113.747307 -0.670103   1.0    0.0   \n",
       "264094  108.741577   45.771694   75.602600   39.309307  1.043189   1.0    0.0   \n",
       "264095  142.555573   60.645222   72.819946   19.166916 -0.522072   1.0    0.0   \n",
       "\n",
       "        isTT  isBKG  isVBS  notLL  notTT  DNN_SigVsBkg  \n",
       "279930   0.0    0.0    1.0    1.0    1.0      0.941570  \n",
       "287376   0.0    0.0    1.0    1.0    1.0      0.557299  \n",
       "296881   0.0    0.0    1.0    1.0    1.0      0.756732  \n",
       "313225   1.0    0.0    1.0    1.0    0.0      0.984433  \n",
       "337598   1.0    0.0    1.0    1.0    0.0      0.978351  \n",
       "...      ...    ...    ...    ...    ...           ...  \n",
       "264090   0.0    0.0    1.0    0.0    1.0      0.978672  \n",
       "264091   0.0    0.0    1.0    0.0    1.0      0.982667  \n",
       "264093   0.0    0.0    1.0    0.0    1.0      0.607159  \n",
       "264094   0.0    0.0    1.0    0.0    1.0      0.871962  \n",
       "264095   0.0    0.0    1.0    0.0    1.0      0.975956  \n",
       "\n",
       "[55026 rows x 65 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_LL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "be847e8c-9b7e-4377-88d6-573fc207d768",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[  1.94873047  -1.15209961   2.15234375 ...  97.33708954 168.26850891\n",
      "    1.1188724 ]\n",
      " [ -2.28027344   3.85644531  -1.22583008 ... 135.69281006  84.57910156\n",
      "   -0.31733713]\n",
      " [  1.5222168   -2.56591797  -0.86108398 ... 155.43873596 260.60864258\n",
      "   -0.79006892]\n",
      " ...\n",
      " [ -1.05761719   2.51367188   0.53759766 ...  38.23987961 113.74730682\n",
      "   -0.67010307]\n",
      " [  3.56347656  -1.22802734   1.82275391 ...  75.6026001   39.3093071\n",
      "    1.04318929]\n",
      " [  2.65429688  -3.86523438  -1.35083008 ...  72.81994629  19.16691589\n",
      "   -0.52207154]] [[0. 1.]\n",
      " [0. 1.]\n",
      " [0. 1.]\n",
      " ...\n",
      " [1. 0.]\n",
      " [1. 0.]\n",
      " [1. 0.]]\n"
     ]
    }
   ],
   "source": [
    "for col in df.columns:\n",
    "    df_LL[col] = np.hstack(df_LL[col])\n",
    "    df_TT[col] = np.hstack(df_TT[col])\n",
    "'''\n",
    "df.replace([np.inf, -np.inf], np.nan, inplace=True)\n",
    "df.dropna(inplace=True)\n",
    "df = shuffle(df)\n",
    "'''\n",
    "# Perform the splitting and define training and validation datasets\n",
    "msk_LL = np.random.rand(len(df_LL)) < 0.8\n",
    "msk_TT = np.random.rand(len(df_TT)) < 0.8\n",
    "\n",
    "df_train_LL = df_LL[msk_LL]\n",
    "df_train_TT = df_TT[msk_TT]\n",
    "df_val_LL   = df_LL[~msk_LL]\n",
    "df_val_TT   = df_TT[~msk_TT]\n",
    "\n",
    "X_LL = df_train_LL.values[:,0:NDIM]\n",
    "Y_LL = df_train_LL[['isLL', 'notLL']].values\n",
    "#Y_LL = df_train_LL[['isLL']].values\n",
    "X_val_LL = df_val_LL.values[:,0:NDIM]\n",
    "Y_val_LL = df_val_LL[['isLL', 'notLL']].values\n",
    "#Y_val_LL = df_val_LL[['isLL']].values\n",
    "\n",
    "X_TT = df_train_TT.values[:,0:NDIM]\n",
    "Y_TT = df_train_TT[['isTT', 'notTT']].values\n",
    "#Y_TT = df_train_TT[['isTT']].values\n",
    "X_val_TT = df_val_TT.values[:,0:NDIM]\n",
    "Y_val_TT = df_val_TT[['isTT', 'notTT']].values\n",
    "#Y_val_TT = df_val_TT[['isTT']].values\n",
    "\n",
    "#X_val = df_val.values[:,0:NDIM+1]\n",
    "#Y_val = df_val.values[:,NDIM+1:NDIM+1+nclasses] # isLL, isMIX, isTT, isBKG\n",
    "print(X_LL,Y_LL)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "3c21764e-e053-40e7-a288-9e7fe2763a48",
   "metadata": {},
   "outputs": [],
   "source": [
    "dnnLL = SimpleNeuralNetwork(350, learning_rate=0.0001, N_NODES=50, n_layers=8, n_features=X_LL.shape[1], n_outputsC=2)\n",
    "dnnLL.save_weights(\"LLvsOther_inclusive_RDF\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "99f11e98-eafd-4202-8ec6-6044171a86cc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(43860, 57) (43860, 2)\n"
     ]
    }
   ],
   "source": [
    "print(X_LL.shape, Y_LL.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "63f73e91-f4f2-44ef-8baa-b762ce41b801",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 350/350 [00:15<00:00, 23.31it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "acc = dnnLL.fit (X_LL.astype(np.float32), Y_LL.astype(np.float32), X_val_LL.astype(np.float32), Y_val_LL.astype(np.float32))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "66ab7ce2-9739-480f-88e6-2730bda10ed2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABKAAAAJBCAYAAABxkc2hAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAABcSAAAXEgFnn9JSAABjzElEQVR4nO3deZhkVX3/8fdnQFAWFXuGiIKguDEoGlTEhYBLXNgUwS0xkZCYuKBE3GIiEaOJy0+JxCVxQzRuUVCRYNzjgoKKKBhGBEEQRJ3pwYV9m+/vj3tLiqaqp5e6XT0979fz1HO67j3Lt+pWz3R/+5xzU1VIkiRJkiRJXVk27gAkSZIkSZK0tJmAkiRJkiRJUqdMQEmSJEmSJKlTJqAkSZIkSZLUKRNQkiRJkiRJ6pQJKEmSJEmSJHXKBJQkSZIkSZI6ZQJKkiRJkiRJnTIBJUmSJEmSpE6ZgJIkSZIkSVKnTEBJkiRJkiSpUyagJEmSJEmS1CkTUJIkSZIkSeqUCShJkiRJkiR1ygSUJHUkyT5JKknN5tx8+14ISQ5tx79oHON3adzvrSRJkrQUmYCStGQleU+bSFibZPNZtDu/bfeZLuNbjJLslOToJEePO5alZqm+t20y8ugk+4w7FkmSJC1eJqAkLWXva8s7AU+aSYMkewP3nNK+C1cDP24fi8lOwKvbx3R+SxP7BV0HtITsxMze2w3NoTSvaZ/xhiFJkqTFzASUpCWrqk4HVrVP/2KGzXr1fgWcMvKgWlX1naq6b1Xdt6sxulRVn2rjf8y4Y5EkSZK0+JmAkrTU9WYxPS7JXaermGRr4JD26Qer6sZOI5MkSZKkjYQJKElL3X8CN9D8e3foeuo+Hdiy/fo4gCRbJHlmkg8m+UGSNUmuS3JZkk8neeJcgprJRtdJ7pvkw0l+meTaJBcmeVuSP1hP37dJcmCSdyc5I8kvklyfZHWSz7evJwPaXQT8b9/zmvI4vu/cejchT7Jzkn9v99S6JsnvkpyZ5B+T3H4m70uSeyY5Lskl7ft+abu317TJxPVZ5O/tnMboa799kn9Nck6Sq/o+r99rjz9kmrb7JTkxyc/bdr9O8vUkz0uy2ZS6h7bXae/20KsHvK6dpns/JUmStPFIlTf5kbS0JTkBOBj4SVXda5p63wQeDnyrqh7RHjsUeH9bpYDfAbcBtuhr+paqeumA/vahTTpUVWZ6rj3/BODTQG/z9CuBTYHbAr8A/oE2STZd363fAQG27jv2CeAZVbWur913gZ2BbdpDv5oS1n9V1RFt3UNp3peLq2qnAfE/DfhgX/xXAJv1Pb8EeHxV/Wia2B8NfAbYqm1/u/Y9ALgM2KOqfj517PXZAN7bOY3Rtn1A27Y3zk1t+zu2fQB8oKoOndLudjTX65C+w79rx+y1Ox3Yt6p+3bZ5OnAszR5rtwGuonkv+z2kqi5BkiRJGz1nQEnaGPSW4d0zyR8NqpDkPjTJp/76AL8G3gw8Etiqqu5YVVsCd6HZePkG4CVJDhxVsEm2B/6LJkFyNvDQqtqaZnbWE2mSCsdM08XVwLuAPwbuUFV3qKrbAxPAETSJhacCh/c3qqqHAE/pe37nKY8jZhj/7sCH2vi/CezWjr8FcCBNkmcH4OQkW03T1YnAV4Bd2vZb0sxSu4Lm/X/9TOKZEtuG8N7OaYzWW2iST2cCDwNuU1V3okmu3Rt4KXDOgHbvpkk+XQj8aW9cmmv2pPb4nrSJufY1/FdV3Rn4VnvozQNel8knSZIkAc6AkrQRSLIMuBjYngGzP9o6bwReTjODY7uqmjqTY1jfLwX+H/DlqnrslHP7MIcZUEneCTwPWAusrKrVU87fjybBcJtB7WcQ8yE0M2guqKp7Tjk3NK4p9Q5lyAyoJP8DPAH4CfCAqrp6yvk/BL5DM+voZVX15kHjt+VjB8zyeSHwb8A1wO1ns1fXhvDeznOMq2lmij28qk6bYX97AV8HVgMPHpQ0ahN359Ik6v6wqn7Qd+6rNMvwXlNVR8/hJUmSJGkj4AwoSUtem8A4vn16yNRZN0k2Af6sffrxmSafWr075T2s7Wde2r19nt4+/Y+pCRKAqvo/4IR5DNOLeeckd55HP7eS5I7A49un/29q8gmgqr4PfLJ9+sxpuvuXqcmn1klteTtg6JLKAbFt0O/tDMf4TVtuN4v+/rItPzxsxlJVXcrNicHHD6ojSZIkTccElKSNxftp9nDqLePq90Ru/oX9fVPOkeQPkrwmyWlJ1ia5MTdvlL2qrbYFN++7Mx93p9lTB5rlZ8NMd44kWyd5WZKvtZtXX98Xc39SaPt5xjvV7ty8Z9CXpqn3xbbcLclthtT59pDjl/V9fachdQbZYN7beYzx3235gSRvSbJ3ki2Y3iPa8i/TbMo+8AH0ZvjtONfXJUmSpI3XpuuvIkkbvqq6sF0q9CjgMG6ZaDqsLc+tqm/1t0vyMOCzNJs491xJkwQoYBNgeXt8S2BynqFu2/f1dBtsXzrsRJJ7A1/mlsmJq2lmx/RmFPXu9rYlozXb+DelSQpN3ZSbqrpiUMOqujE33wRuWPJqFLHdykK8t/Mc4+XAPWk+50e2j5uS/IBm5tS7B2zcfpe2vH37WJ/1JbQkSZKkW3EGlKSNSS/p9PD2l3ySrAD2b48f1185yabAR2mSTz8A9qXZc2jrqvqDdgPmPfubdBf6rLyfJnlxEc1m1RNVtWVVbdvGfNe+uosl5g3FQry3cx6jqn5TVY8G9gLeRLMJ/I3Ag4B/BM5PMnXZY2/p6POqKjN4HDrH1yVJkqSNmAkoSRuTE7l5j5zerKdn0cyiuZHmNvT9Hkaz3OgmYP+q+p8Bs3JGvc9P/75Edx1aa8i5JDtw8938nllVJ1TV5VOqdbU3Edwy/umWoPXO3QhMja8ri/69HdUYVXVqVb2iqh5Jk0B9EvBDmn2zjkvyB33Vf9mWLq2TJElSZ0xASdpoVNW1wEfap3/ebhr+F+3z/66qqcvAdmjLNQOWLfU8dsjxufopNydkHjVNvUcPOb5D39ffH1Jnuph/v+l3+ta5zcKZfX08Zpp6vRjOqqob5jDOXGwI7+18x7iVqrq2qj4DPKU9dFvgkX1VvtmW+zM3vdflbDpJkiQNZQJK0samtwxvO+Ao4P7t8+MG1P1tW/7BlBkjwO9vTf+iUQZXVQV8vH363CTLp9ZJshI4ZEgXv+37+gED2m4NvGqaEH7X9/Udpw12gKr6DfD59unLBm2AneQBwMHt04/Odoy52kDe2zmPkWTTJNP9v35N39f9dxd8d1veL8nzpmlPki2TbDblcO913XG6tpIkSdq4mYCStFGpqjNp9nOCJgEF8AuajcanOhW4imZmx8f79o3aJMnjga/SbEQ+aq8HrqDZ3PyLSR7cjpskjwP+h1veCa3fj4CftV8fl+RBvRPthupfZfq79Z0HXN9+/VdznAX1KuAGms2wP5/k/u34y5LsS/NebwpcALxrDv3Px2J/b+czxvY0ezy9KskftnuY9druBnyofXoV8LXeuar6Gs2+UwDvSPKvSe7R13bzJHsmeRNwMbfczB3g/9py3yTTLW2UJEnSRswElKSNUW8WVO/fwA9U1U1TK1XVb4GXtk//CPhxkito7oL3OeAO3LyEb2Sq6mfAM4HrgAcC303yO5rEwedp9qw6ckjbdcALaPZW2hU4I8lVSa4CvgXcB3j6NGNfDfxn+/RNwJVJLk5yUZI3zzD+M4E/o0m2PBI4O8lv2/hPobnr2iXAAVV15Uz6HJXF/t7OdwzgHsBraZZCXptkbZLrgLOAfWiuyaED9pV6LvBemmTr3wIXJLkiyeU0CbnTgJcBE9w66foB4FqahOPPkvyyfU0XtbMEJUmSJBNQkjZKH6b5hbln0PI7AKrqP4D9aGaeXEkzc+fnwNtolkj9sIsAq+oUYHfgYzSbZ28G/Ap4O/CHNPsZDWv73zQJs1NoNl3fFJikmeXyoKr68nqGfwFwNDe/trvRbFB9qyVr08TwXzQJlHfRzHTanCap8gPg1cD9qupHM+1vlBb7ezuPMX4OHAj8K3A6zcy+rWje91XAO2je9xMGvK7rq+o5NBugH09zzTZp26+m+fz/E7Db1P3Qqup8mj21PgOsoUlS7dg+NkWSJEkC0myJIUmSJEmSJHXDGVCSJEmSJEnqlAkoSZIkSZIkdcoElCRJkiRJkjplAkqSJEmSJEmdMgElSZIkSZKkTpmAkiRJkiRJUqdMQEmSJEmSJKlTJqAkSZIkSZLUKRNQkiRJkiRJ6pQJKEmSJEmSJHVq03EHsCFJ8ktgC+CSccciSZIkjcgOwNVVdedxByJJWrpSVeOOYYOR5Hebb7751jvvvPO4Q5EkSZJG4oILLuC66667oqpuP+5YJElLlzOgZueSnXfeeeU555wz7jgkSZKkkdh1111ZtWqVM/wlSZ1yDyhJkiRJkiR1ygSUJEmSJEmSOmUCSpIkSZIkSZ0yASVJkiRJkqROmYCSJEmSJElSp0xASZIkSZIkqVMmoCRJkiRJktQpE1CSJEmSJEnqlAkoSZIkSZIkdcoElCRJkiRJkjplAkqSJEmSJEmdMgElSZIkSZKkTpmAkiRJkiRJUqfmnYBKskWSJyd5X5IfJ7k2yVVJzkryj0m2mkOf2yQ5NsnFSa5ry7cmueM0bTZJ8uIkP0xyTZI1ST6eZJd5vUBJkiRJkiTNyyhmQP0J8CngMOAm4DPAN4C7A68Bvptk25l2lmQ58B3gRcCNwKeBK4AjgG8nudOANsuATwDHANsDpwDnAIcAZyTZY46vTZIkSZIkSfO06Qj6uAF4N/DWqvpR72CS7WgSQX8IvJUmUTUTbwXuCXwSeHpV3dj292/AC2mSTIdOaXMYcBBwPrBXVf2qbXMwcALw4SS79PqSJEmStEQ1f5yeGHcYc7SWqnXjDkKSupCq6q7z5GHAt4DrgNtX1fXrqb8dcCnNzKe79RJJ7bnNgUuAOwF3qarVfedWAbsAB1XVp6f0eRJwIHBIVZ04z9dzzsqVK1eec8458+lGkiRJWjR23XVXVq1ataqqdh13LCORrABWr7fe4rQtVWvGHYQkdWEUM6Cmc1Zbbk7zV4hfrKf+E2iWBX6jP/kEUFXXJTmZZrbTvsDxAEnuTpN8uoZmxtVUJ9AkoA4A5pWAkiRJc7BuHaxdO+4obm1iApZ5PxZJS0eSAi6uqp3m2H4f4H+Br1XVPjOofzzwbOAvqur4uYw5F0kuAnasqizUmOOW5KvA3sDdq+qiOfZRADN535IcCrwf+EBVHTqX8TYm8/3e21h0nYC6R1veAFw+g/oPaMszh5w/kyYBtduANv9XVTcMacOUNpIkaaGsXQvbzng7yIWzejWsWDHuKCRJkjYKXSegjmjLz1XVdTOof7e2vHTI+d7xHefZZlpJhq2x23mmfUiSJEnSAtuF5o//Wnr+HNgC+Pm4A5HmqrMEVJJ9gb+k+QfwqBk226otrx5y/qq23HqebSRJkiRtPFYCk+MOYorlwKpRdlhV546yPy0eVfWzcccgzVcnGx8kuS/wISDAy6rqrPU0WVSqatdBD+CCcccmSZIkadYmqVqzqB4dJMSSVLs/0tTjD0/y6SQXJ7kuyS+TfCfJG5JsNaCrsUjyhCSfSfKrNs5Lkvx3e3fzmbTfL8lxSX6U5HdJrkpyVpK/b29qNajNvkm+mOTn7ZiXJTk1yaun1EuSP23P/SrJtW18X0ryglG8/vW8tq+213enKcd3TPLvSc5LcnWSy5Ock+RdSe7TdVwzMdPPX5LbJvnLJCcluTDJNUl+k+TrSZ4xpO/j2/dlnySPbetekWR1kvckuUNbb9v2Pfl5e+2+0+55NrW/Q9v+jk5ynyQnJlnbfpa+2U60me3r36WN85L29f8qyceSLI0bP8zCyGdAJbkr8DlgG+CYqjp2Fs2vbMsthpzfsi2vmGcbSZI0TqtWwfLlCzfe5CSsXLlw40nSIpHkAODTNJMDvkNzl/I7AvcCXgH8Bzf/TjU2Sd4CHAmsA04DfgbcBXgEsD0zu6HU+4DbAf8HnA3cAdgD+GfgMUkeV1U39Y35AuDtwE3AN4Gv0cxM2wU4GnhNX99vAl5Kc4f3r9MkEO9Ms9fwPYF3zP5Vz0+SHWj2PL4TcD7wWWATmu1nnkPzPv54oePqN8vP307Ae4HLaOL+Ds17/HBgryT3raqjhwx1EPACmtf8OWBP4K+AeyU5pD2+CfCNdpyHAp9L8pCq+uGA/nYGvk2zl/UXaD6LewH/neQvq+r9M3z9TwY+RnNjth8ApwM7AE8DDkjyxKr6+kz6WgpGmoBKcieai7MjzY75L51lF71phdsPOd87fvE820iSpHFavtwNwCVpYbyUZuXLIVV1iyROkocAY79NaZJn0SSfLgP2q6of9J27HfDIGXb1N8AXquqavvZbAx8B9gf+FPhgX/2XAwXsWVVn9LUJzR3nes9vC7yQZlLDA6rqp33nNgUeNsP4Ru2vaJJPb6+qF/afSHI34DZjieqWZvP5WwP8MfDlqqq+encHvgIcleT4IXcBPBw4sKpOadtsTZPs2psmsfgt4LDejcuSvBZ4VRvfswf09yyaz8pfVtWNbZv9aZJpb0/yhaqadj+udrbah2i2Jdq/qr7Ud+4JwGeADyW5Z1VdP11fS8XIluC1U+f+h2Z99SeB5/R/aGaot1Rv9yHne8fPHtDmfkkGfYMNaiNJkiRJG4Netv9LU09U1XerajGsFPn7tjyyP/kEUFXXVNUXZ9JJVZ3Un3xqj10BvLh9+qQpTVYAv+lPPrVtqqq+2nfo9jQzWC7oTz61dW+sqm/MJL4OTHdtf1ZVi2ELmRl//qpqbVV9aWoeoX3P/5kmf3HAkHE+0ks+tW2uAN7TPt0eeFEv+dR6M03ycW8GuxL4217yqe3zv4ETaFZf/cWQdv3+lmZF1iv7k09tX58D/p1mNtR+M+hrSRhJAqpdT3sSzfTGzwPP7J/aOAufo5lyuVeSW9yvuR3jAJrpkZ/tHW8/jD+imWo56MId0pYnzyEeSZIkSdqQfa8t/zPJQ5J0sg/wXCW5C82St98AHx9Bf/dKckSSt6XZD+p4br4p1r2mVP8esE2S9023H09Vraa5u/oD232L7jHfOEekd23/Jcn+7UytxWbWn78kj0zyqnZvq/e31/Cp7emp17DnCwOOXdiWZ1TVr/tPVNVvaZbXbTesv6ltWh9ty72GvoCbPa4tPznkfC9xuccM+loS5r0EL8kmNBfh0TRv4FPWN30syeE0U+Q+VVWv7B2vql8k+SjN1Mh3JnlGX8bxTTTZ0w+0/wD0O4Ymu/mmJN/qnU/yFOBA4Cc0CTJJkiRJ2pj8PXB/mj/mHwD8OsmptMt/quracQZHMwME4MI5rKD5vXbZ3JtpZjtlSLWpd0Z/Ac2SqsOAw5L8ima51ieBE6ZMqng2zV4+rwBekeTitu7Hqup/Zhjjm2n2mOp3alW9dybtBzieJsnxNJoJF9cm+S7NxI7jquqXc+x3lGb8+Ws3DP8kTW5hmGF3tx+0HO7Kac71zk8MOTdsC5+L2vIuQ87326k3fvPxHGoBN8Ucr1HsAXU4zYZf0GzE9s4hb+5Lq6p3p4flwH0YnG38W5oNww4Gzk1yBrArcD+ajdWOHNDmOGDfNo5zk3y5HWNv4BrgWf1T5yRJkiRpY1BVlyR5MM0v9fvT/I7USwa8PMnDqmrs+0CNwNNpfle8hCYJdRqwpqpuSLIZzebht/hFtarOTrISeALN75P70CRzngaclmSf3uSKqvpKknvSvIdPaOv+OfDnSU6sqkNYv0No9kueak4JqDZB9vQkb6BZXvhoms219wL+LskTqupbc+l7VGb5+XtjW+9rwKtpNpP/TVXdlORxNKuthmVy1k0TxnTnutSb7fWB9dT7dteBLBajSEBt0/f1QUNrNXcRWO+tRqtqMskebf0nt33+Cvg34NVV9ZsBbdYleSpwBE32en/gKpo7Jby6qlbN4HVIkiRJ0pLT/jH+C+2DJDvS/BH/0TQzel4+vui4pC3vkSTzmAXV+130ef17AfX6HtaonYHz6fZBuxTvIzQbi/8V8M6+ur9rz32krbsn8Ang4CT7VtVnmUZV7TTjVzMLVfV94PvA0UluT/O79IuBt7IIlnfN4vN3EM2WOwe273W/hV72OChR2H/8shn0cSnN3fReskSSvPM27/W/VXV0VWUGj4sGtDl0SJ+XV9WLqupuVbV5Wx4xKPnU1+amqjqmqu5XVberquVV9VSTT5IkSZJ0s6q6mGa2CTQrTcYZy2U0e/rekZv3+ZmL3sSISwece9os4jkHeEf7dNr3pqpOB/5zJnUXSpu4eSXNBtuLIqappvn8bQP8bkDyCWZxDUfkcUnuOOD4M9ry1Bn00ds8f7qJOhuVRbUBnSRJkiR1YDnJikX1WKB9X5K8OMmdB5zaty0vGXBuob2hLY9Jslv/iSS3TfLHM+jjvLb86/TtCZNkL+BlUysn2SLJi6YmGdpNsp/QPr2kPXa3JIcm2WJqbMCj+usupCR/lmRQkumJNEvVxn5tZ/n5O49mU/inT+2Dm9/nhbIVzefx96vGkjyRJhF2DfD+GfTxlrbum9v9qW8hyeZJDkmy/YhiXvRGsQRPkiRJkhazjXlVxKtpfgE+i2ZP3QAPAO5NcxewNw9os3uS06fp8/lVdWbf86OSPHdI3V9U1bQzQKrqg+0+QS8EzkxyGk1iYjvggTQbQj9wuj5otmw5FHg+sE+Ss4G7Ao+kSQS8dEr9zYBjad6b79FsLr0Z8BCajdEvAt7d1r0TTcLhHe0exZcCWwIPp7lR1hkMv9NZlw4GPpjkAuCHNMmOu9PsA7UOeNWgRuu5tu+dsin6fuupf1BV/WKa87P5/L0e+BDwsSQvoHmfHwDcF/hXmmWFC+XDwFNoPkvfpvks/hFN/C+qqkEz7W6hqn6S5Jk0SzZPTPITmtl+V9F8Nnen+Rz9IYNn7i05JqAkSZIkael6Ic2MngfRzIyBJrlzDHBMVQ26Q9jWNEmMYW4/5fk9GL5Hz7C7id1CVb0oyZdoEkgPodm7aDXNUqfjZtD+vDaJ9Saa2A8Efgz8TVW9J8nUBNSVNHfBewxNkmM34HrgZzSbgr+9qi5v614AvKStu7KN7Srgp8C/AO+uqutm8jpH7BiaxMUjaDYe35Jmb6L/At5SVWcMaTfdtf3clOfLmX623ubriXHGn7+q+nCSXwNH0SQc70+T3Hs+TeJnIRNQP6HZB+z1wOOB2wKnA/9SVf89006q6qR2Vt+RwB+3jxtortPJNInLjSZBnnnc6XKjk+SclStXrjznnHPGHYokSRuONWtg221veWz1alixYuOKQVqkdt11V1atWrWqqnYddywj0SxvWz3uMOZoW6rWjDsIaWOV5FCa2W6vqaqjxxvN0uMeUJIkSZIkSeqUS/AkSZIkLSVrgW3XW2tx8lbtkpYsE1CSJEmSlo6qdYDL2CRpkTEBJUmSJEmSNnpVdTxw/JjDWLLcA0qSJEmSJEmdMgElSZIkSZKkTpmAkiRJkiRJUqfcA0qSJG2cJifHHQFMTMAy/x4oSZKWPhNQkiRp47Ry5bgjgNWrYcWKcUchSZLUOf/kJkmSJEmSpE6ZgJIkSZIkSVKnTEBJkiRJkiSpU+4BJUmSlr6JiWa/pXGanFwc+05JS1zCMmBi3HHM0doq1o07CEnqggkoSZK09C1b5mbf0sZjAhhzxnnOtgXWjDsISeqCS/AkSZIkaQlIUkkumkf7fdo+vjrD+se39Q+d65hzkeSiJLWQY45bkq+27/VO445lvtrXMaPrl+TQtv7xHYe1JMz334CumYCSJEmSJEnz1pfEPH7csWjxcQmeJEmSJC0NuwA3jDsIdeLPgS2An487EGmuTEBJkiRJWupWApPjDmKK5cCqUXZYVeeOsj8tHlX1s3HHIM2XS/AkSZIkLXWTVaxZTA86SIgN2/8lycOTfDrJxUmuS/LLJN9J8oYkW406jrlK8oQkn0nyqzbOS5L8d5KDZ9h+vyTHJflRkt8luSrJWUn+PsnmQ9rsm+SLSX7ejnlZklOTvHpKvST50/bcr5Jc28b3pSQvGMXrX89rG7gHVJIdk/x7kvOSXJ3k8iTnJHlXkvv01dupt79Xktu11773efhJklckyZCxVyb5cJJfJLm+fa8+2N9/W+944H/bp8/u7fXUPo4e6RsyBzP9Pkhy2yR/meSkJBcmuSbJb5J8PckzhvTd2w9tnySPbetekWR1kvckuUNbb9v22vy8/Qx9J8k+A/rr7X11dJL7JDkxydr2M/3NJPvO4fXv0sZ5Sfv6f5XkY0l2nW1fc+UMKEmSJElaopIcAHwaCPAd4FvAHYF7Aa8A/gO4ckzh/V6StwBHAuuA04CfAXcBHgFsD5w4g27eB9wO+D/gbOAOwB7APwOPSfK4qrqpb8wXAG8HbgK+CXyNZmbaLsDRwGv6+n4T8FLgOuDrNAnEOwO7AfcE3jH7Vz0/SXYAzgTuBJwPfBbYBNgReA7N+/jjKc02A75AMyvwq8CWwN7AG4CtgVdNGeMxwMk07+v32zb3Bf4MOCjJvlX1jbb6qTTvyeOBC9rnPT+Y36udn1l+H+wEvBe4jOb9+w7N63o4sFeS+1bV0UOGOgh4Ac17/zlgT+CvgHslOaQ9vgnwjXachwKfS/KQqvrhgP52Br4NXE5z3e4C7AX8d5K/rKr3z/D1Pxn4GLA5zbU4HdgBeBpwQJInVtXXZ9LXfJiAkiRJGpfJRbIiaGICljkxXlqiXkqz8uWQqrpFEifJQ4C1Y4nqlnE8iyb5dBmwX1X9oO/c7YBHzrCrvwG+UFXX9LXfGvgIsD/wp8AH++q/HChgz6o6o69NaJIyvee3BV4IXAE8oKp+2nduU+BhM4xv1P6KJvn09qp6Yf+JJHcDbjOgzcNoEm13r6rftXUfTJOQeHGSN1TVle3xLYEP0ySfDq+q3yfZkrwYOAb4SJJ7VdW1VfXeJD+hSUCdWlWHjvblzstsvg/WAH8MfLmqqq/e3YGvAEclOb6qLhowzuHAgVV1Sttma5pk19407/u3gMOq6ob2/Gtpkn4vBZ49oL9n0Xxm/7Kqbmzb7E+TTHt7ki9U1bT7grWz5j5Esz/c/lX1pb5zTwA+A3woyT2r6vrp+povE1CSJEnjsnLluCNorF4NK1aMOwpJ3eh9c39p6omq+u4CxzLM37flkf3JJ4A2mfTFmXRSVScNOHZFmyzZH3gSt0xArQB+0598atsUzUyfntvTzBz5UX/yqa17I81slnGY7toO2zNqHfA3veRTW/eMJP9D8x49mJtf+9OAPwBO608+tW3+NcmfAg8CDqZJVC1mM/4+qKq1Q+r9NMk/A+8BDgDeNmCcj/SST22bK5K8BziWZibfI3vJp9abgX+gL+E5xZXA3/aST22f/53kBODpwF8ArxvStudvaWa6vbA/+dT29bkk/w68CNgP+NR6+poX/9QlSZIkSUvX99ryP5M8JMmi+h0wyV1olrz9Bvj4CPq7V5IjkrwtzX5QxwNHtafvNaX694Btkrxvun1wqmo1cCnwwHa/oHvMN84R6V3bf0myfztTa30urqqpy/IAzmvL7fqO7dWWw5JLH5pSbzGb9fdBkkcmeVWaPbbe336WntqenvpZ6vnCgGMXtuUZVfXr/hNV9Vua5XXb3apV29/UNq2PtuVM3vvHteUnh5zvJVD3mEFf8+IMKEmSJElauv4euD/NjI0DgF8nOZV22U1VXTvO4Gj2oQG4sH+502y1y+beDLyYZp+fQbae8vwFNEuZDgMOS/IrmmVSnwRO6N8vimZ51Mdo9gt6RZKL27ofq6r/mWGMb6bZY6rfqVX13pm0H+B4muTC02j2abo2yXdp9h46rqp+OaDNpUP6uqIt+zdrv0tbXjSkTe/4XWcW7ljN+Pug3TD8k8Cjp+lv6mepZ9ByuCunOdc7PzHk3MVDjl/UlncZcr7fTr3xM3if+Z6pn82RMwElSZIkSUtUVV3S7vHzaJolVntz8y/hL0/ysHbJ0Ybu6TT7SF1Ck4Q6DVhTVTck2Yxm8/Bb/PZdVWcnWQk8AdgX2IcmmfM04LQk+/T2xKmqryS5J817+IS27p8Df57kxKo6ZAYxHkKzQfhUc0pAtQmypyd5A83ywkfTbGq9F/B3SZ5QVd+a0mzdXMYaFsII++rULL8P3tjW+xrwappN7X9TVTcleRzweYYnOad7f0f53s9Gb7bXB9ZT79tdB2ICSpIkaSFMTDR7LY3b5OTi2XtK0oJo94/5QvsgyY7AcTS/ZL+CZjPucbmkLe+RJPOYBXVQWz6vfw+eXt/DGrUzXz7dPmiX4n2EZrPuvwLe2Vf3d+25j7R19wQ+ARzc3g3us9MFWFU7zfjVzEJVfZ/mDnVHJ7k9zR38Xgy8lfktq7qsLQclzaBvZs08xlgws/g+OIjmzogH9u+V1Vro5ZfD3vve8cuGnO93Kc3d9F4y7mSzCShJkqSFsGyZG31L47N8+pUnY9H5cpdhquriJL1ZHvcbVxxtLJcl+RHNPlBPZe77QG3TloOWmD1tFvGck+QdwLtYz3tTVacn+U/glW3daRNQC6GqfpfklTQbT8/32n6DZpPrZ9KXiOvzrL56Pb27qC36XMM03wfbAL8bkHyCWXyWRuRxSe5YVb+ZcvwZbXnqDPr4Ik0C6iDmONtuVBbVBnSSJEmS1IFVwOpF9ljV6StuJXlxkjsPOLVvW14y4NxCe0NbHpNkt/4TSW6b5I9n0EdvE+2/Tt9GN0n2Al42tXKSLZK8KMkdpxxfRrPEDtr3JsndkhyaZIupsQGP6q+7kJL8WZJBSaYn0iwRm29MHwd+BTwyyV9PGftFNHfM+zlwYt+p3oyc+8xz7JGa5ffBeTSb0z99ah/cfL0XylY03xe/T+gleSJNIuwa4P0z6OMtbd03J3nK1JNJNk9ySJLtRxTzUIs+KylJkiRJmrNX0/zieRZwPk1i4gHAvWnuvvXmAW12T3L6NH0+v6rO7Ht+VJLnDqn7i6o6aMg5AKrqg+3+PC8EzkxyGk1CYDvggTQbMT9wuj6AfwMOBZ4P7JPkbJrNsR9J8wv4S6fU3ww4lua9+R7Nps6bAQ+h2Rj9IuDdbd070fyi/44kZ9DMstoSeDiwAjiD4XcY69LBwAeTXAD8kCbJcHeafaDWAa+aT+dVdVWSP6XZ4PxdbRLqPOC+wB/SbJ79zP4NvKvqova9f3CS7wDn0Cxn+0xVfaa///V8xt47ZXP2/dZT/6Cq+sU052fzffB6mjv8fSzJC2iu9wPa1/2vNMsbF8qHgafQfKa/TfM98Uc08b+oqoZtKv97VfWTJM+kWTp6YpKfAD8CrqL5Htmd5vP8hwzfpH4kTEBJkiRJ0tL1QpoZPQ+imRkDTXLnGOCYqhq0f8/WNEmMYW4/5fk9GL43zrC7eN1CVb0oyZdoEkgPodm7aDXNEqPjZtD+vDaJ9Saa2A8Efgz8TVW9J8nUBNSVNHfBewxNcmE3muVjP6NZpvT2qrq8rXsB8JK27so2tquAnwL/Ary7qq6byescsWNoEgaPoNl4fEuaGUj/Bbylqs6Y7wBV9eUkDwH+gWap2m7AJE2C5nVV9eMBzQ4G/l8b04NoVl5dSnPHuX7TfcY+N+X5cqZftrr5NOdgFt8HVfXhJL8GjqJJfN6fJsn4fJrEz0ImoH5Csx/Z64HHA7cFTgf+par+e6adVNVJ7ezCI4E/bh830HxeTqZJoHY+KzPzuNPlRifJOStXrlx5zjnnjDsUSZI2HGvWwLbb3vLY6tXuhzQuXg9Nseuuu7Jq1apVVbXruGMZhYQVNImLDdG2VawZdxCSxivJoTSz7l5TVUePN5rRcQaUJEmSpKVkLbDtemstTmO9Q5UkdWkkCagkD6KZwrVH+7grQFXN6l4TfVm+9Xl2VX2wr93xwLOnqf+8qvqP2cQiSZIkacNTxTpwFpEkLTajmgF1FPCkEfTzE+ADQ87dAXhy+/WwWw1+HvjlgOOD1qVKkrS0rVsHaxfBH9MnJ8cdgSRJksZsVAmo04Czge+2j4tY/yZgt1JVpzIkuZTkeTQJqG9W1YVDunhDVX11tuNKkrQkrV17671+JEmStKhV1fHA8WMOY+RGkoCqqjf2P09mtfJupp7Vlv/ZReeSJEmSJEnqxrJxBzATSe4OPJzmtpgfH3M4kiRJkiRJmoUN5S54vdlPp1TVr6ep95QkBwObAD8FTq6qczuPTpIkSZIkSUNtaAmo9S2/e+GU529M8u/AEVV14+jDkiRpA7NqFSxfPu4oYGJi3BFIkiRpAS36BFSSPYB7A5cDpwyp9n2ajdC/AlwK3Bl4IvA64Pk0S/dePIsxzxlyaueZ9iFJ0qK0fDmsWDHuKCRJkrSR2RD2gOrNfvp4VV0/qEJVHVtV76qq86vqmqr6aVW9E9iLJvl0eJIdFipgSZIkSZIk3WxRJ6CSbAo8vX0667vfVdU5wGdoZno9Zhbtdh30AC6YbQySJEmSJEkbu0WdgAIeB2wLXFhV35pjH+e35XajCUmSJEmSJEmzsdgTUL3ldx+aRx/btOVV84xFkiRJkiRJc7BoE1BJtgKe1D6dUwIqyebAfu3TM0cRlyRJkiRJkmZnLAmoJIcnOTfJ66ep9hRgC+D0qjp/WKUk903yZ22yqf/4CuBjwA7AWcA3RxC6JEmSJEmSZmnTUXSSZD/gqL5Dm7XHT+879tqqOqX9ejlwH6bfl6m3/G59m4/fGfggcGySM4A1wF2ABwFbA5cCT6uqmsFLkSRJkiRJ0oiNJAEFrAAeOuD4Q6fUmZEk2wGPBm4A/ms91c8D3grsCdwfmACua4+fDBxbVb+e6diSJEmSJEkarZEkoKrqeOD4WdQ/Gjh6mvO/YIaxVdVlwItnOrYkSZIkSZIW1qLdhFySJEmSJElLgwkoSZIkSZIkdcoElCRJkiRJkjo1qk3Ipc6tWwdr1447isEmJmCZ6VxJkiRJkgYyAaUNxtq1sO22445isNWrYcWM7/MoSZIkSdLGxTkbkiRJkiRJ6pQzoCRJ6sJiWDc8OTne8SVJkqSWCShJkrqwmNcNS5IkSQvMBJQ2aKtWwfLlCzvm5CSsXLmwY0qSJEmStCEzAaUN2vLlbv4tSZIkSdJi5ybkkiRJkiRJ6pQzoCRJWijjWDc81cTEeMeXJEnSRskElLQBWgw31xpmYgKWObdSGsx1w5IkSdpImYCSNkCL+eZaq1f7+7UkSZIk6ZacpyBJkiRJkqROmYCSJEmSJElSp0xASZIkSZIkqVPuASUtEeO4udbkJKxcubBjSpIkSZI2PCagpCXCm2tJkiRJkhYrl+BJkiRJkiSpUyagJEmSJEmS1CkTUJIkSZIkSeqUCShJkiRJkiR1ygSUJEmSJEmSOmUCSpIkSZIkSZ0yASVJkiRJkqRObTruAKSlYHJyaY8nSZIkSdJ8mICSRmDlynFHIEmSJEnS4uUSPEmSJEmSJHXKBJQkSZIkSZI6ZQJKkiRJkiRJnXIPKGmWJiZg9epxR3FrExPjjkCSJEmSpMFMQEmztGwZrFgx7igkSZIkSdpwuARPkiRJkiRJnTIBJUmSJEmSpE6ZgJIkSZIkSVKnTEBJkiRJkiSpUyNJQCV5UJK/S/LJJJcmqSQ1x74u6rUf8rjvkHabJHlxkh8muSbJmiQfT7LL/F6dJEmSJEmS5mNUd8E7CnjSiPrq+cCQ47+deiDJMuATwEHAb4BTgOXAIcB+SR5VVd8ZcXySJEmSJEmagVEloE4Dzga+2z4uAjafT4dVdegsqh9Gk3w6H9irqn4FkORg4ATgw0l2qaob5xOTJEmSJEmSZm8kCaiqemP/8ySj6HY2jmzLl/eSTwBVdWKSzwAH0szQOnGhA5MkSZIkSdrYbfCbkCe5O7ALcA3N0rupTmjLAxYsKEmSJEmSJP3eqJbgjVySlwE7A9cB5wCfqqo1A6o+oC3/r6puGHD+zLbcbfRRSpIkSZIkaX0WbQIKeNOU5/+a5IVVddyU43dry0uH9NM7vuPIIpMkSZIkSdKMLcYE1GeA/wW+B6wB7kGzyfgRwHuTrK2qk/rqb9WWVw/p76q23HqmASQ5Z8ipnWfahyRJkiRJkhqLLgFVVS+acugc4CVJzgXeDbwROOlWDSVJkiRJkrQoLboE1DTeB7wOuE+Snarqovb4lW25xZB2W7blFTMdqKp2HXS8nRm1cqb9SJIkSZIkaQO6C15VrQMuaJ9u13fqZ225/ZCmveMXdxGXJEmSJEmSprfBJKBa27TlVX3HzmrL+yW5zYA2u7fl2Z1FJUmSJEmSpKE2mARUkl2B+9BsNn5u73hV/RT4EXA7YL8BTQ9py5O7jlGSJEmSJEm3NpYEVJLDk5yb5PVTju+b5NED6u8GfAII8N6qun5KlWPa8k1Jtu1r9xTgQOAnuHG5JEmSJEnSWIxkE/Ik+wFH9R3arD1+et+x11bVKe3Xy2lmM/Xv5QSwB/DqJBfTLK27GrgHzTK6TYGvAn83IITjgH2Bg4Bzk3y5HWNv4BrgWVV141xfnyRJkiRJkuZuVHfBWwE8dMDxh06psz6fB3YAHgI8ArgD8DvgVODDwPur6qapjapqXZKnAkcAhwH70+wTdSLw6qpaNfOXIkmSJEmSpFEaSQKqqo4Hjp9F/aOBowccPw04bY4x3ESzFO+Y9dWVJEmSJEnSwtlgNiGXJEmSJEnShskElCRJkiRJkjplAkqSJEmSJEmdMgElSZIkSZKkTpmAkiRJkiRJUqdMQEmSJEmSJKlTJqAkSZIkSZLUKRNQkiRJkiRJ6pQJKEmSJEmSJHXKBJQkSZIkSZI6ZQJKkiRJkiRJnTIBJUmSJEmSpE6ZgJIkSZIkSVKnNh13AJI0H+vWwdq1447i1iYmYJkpfkmSJEkCTEBJ2sCtXQvbbjvuKG7tnHNgxYpxR2EiTJIkSdLiYAJKkjqw667jjqCxevXiSIRJkiRJ2riZgJKkJWzNmnFH4CwsSZIkSSagJGlJWwwzsZyFJUmSJMkElKQlZ9UqWL584cZbs2ZxJHokSZIkabEyASVpyVm+fGFn3ExMNLN8xs1EmCRJkqTFygSUpJGanFza4w2ybJlLzCRJkiRpOiagJI3UypXjjmDjtRhmYk1O+hmQJEmSdGsmoCRpiXAmliRJkqTFyhtjS5IkSZIkqVMmoCRJkiRJktQpl+BJmrPFsOfQIBMT445AkiRJktTPBJSkOXPPIUmSJEnSTLgET5IkSZIkSZ0yASVJkiRJkqROmYCSJEmSJElSp0xASZIkSZIkqVMmoCRJkiRJktQp74InSRqddeu41d82JieBGkc0N5uYaG7bKEmSJGksTEBJkkbn8suB5bc8tnIXYHIc0dxs9WpYsWK8MUiSJEkbMRNQkqSlb82ahR9zcsxJN0mSJGkRMQElSVr6dt113BFIkiRJGzU3xJAkSZIkSVKnRjIDKsmDgD8G9mgfdwWoqsyynzsC+wIHAHu2/VwHrAI+Aryzqm4Y0O544NnTdP28qvqP2cQiSRqRU78J995m4cZbs8YZT5IkSdIiM6oleEcBTxpBPy8F/oHmdkk/AL4NrAAeQZPYOiTJ46vq6iHtPw/8csDxH48gNknSXNzpTrBi+frrSZIkSVqyRpWAOg04G/hu+7gI2HwO/VwFvAl4R1X9rHcwyb2ALwGPBF4F/P2Q9m+oqq/OYVxJ0lIxMdHc9W4xmpgYdwSSJEnSWIwkAVVVb+x/nsxq5V1/P68fcvz8JH9HswzvmQxPQEmSNnbLlsGKFeOOQpIkSVKfDWkT8rPa8i5jjUKSJEmSJEmzMqoleAvhHm05aI+nnqckORjYBPgpcHJVndt5ZJIkSZIkSRpqQ0pAHdGWJ01T54VTnr8xyb8DR1TVjd2EJUmSJEmSpOlsEAmoJM8FHgv8BnjDgCrfp9kI/SvApcCdgScCrwOeD1wPvHgW450z5NTOMw5akiRJkiRJwAawB1SSvYBjgQIOq6rLptapqmOr6l1VdX5VXVNVP62qdwJ70SSfDk+yw8JGLkmSJEmSJFjkCagk96NZcrcZzTK6T82mfVWdA3yGZqbXY2bRbtdBD+CC2YwvSZIkSZKkRZyASnJ34AvANsDRVfW2OXZ1fltuN5LAJEmSJEmSNCuLMgGVZDvgizRJo2Or6jXz6G6btrxq3oFJkiRJkiRp1hZdAirJNsDnaTb8fj+z2Dx8QF+bA/u1T8+cf3SSJEmSJEmarbEkoJIcnuTcJK+fcnwL4BTg/sDHgedUVa2nr/sm+bM22dR/fAXwMWAH4Czgm6N8DZIkSZIkSZqZTUfRSZL9gKP6Dm3WHj+979hrq+qU9uvlwH249b5M/ww8DLgJuBF4X5JbjVdVh/Y9vTPwQeDYJGcAa4C7AA8CtgYuBZ62vkSWJEmSJEmSujGSBBSwAnjogOMPnVJnfXr7NW0C/Mk09Q7t+/o84K3AnjQzpyaA69rjJ9PsIfXrGYwtSZIkSZKkDowkAVVVxwPHz6L+0cDRA44fyi2TSzPp6zLmsU+UJKlbk5cva+amjtHEBCxbdLseSpIkSRuPUc2AkiRpoJWPvNO4Q2D1algxk3m4kiRJkjrh34MlSZIkSZLUKRNQkiRJkiRJ6pQJKEmSJEmSJHXKPaAkSSMzcadi9dSbnq76ESxfvmAxTE7CypULNpwkSZKkGTABJUkamWXLYAWTtzy4vJiak5IkSZK0cTEBpRlZtw7Wrh1vDJOT668jSZIkSZIWHxNQmpG1a2HbbccdhSRJkiRJ2hC5CbkkSZIkSZI65QwoSVoqXCsrSZIkaZEyASVJS4VrZSVJkiQtUiagNGerVi3ondUHmpgY7/iSJEmSJGn9TEBpzpYvhxXeWl2SJEmSJK2Hm5BLkiRJkiSpU86AkqSlzLWykiRJkhYBE1CStJS5VlaSJEnSImACSpK05E1OjjuCxsQELHPxuyRJkjZCJqAkSUveypXjjqBxzjnjn5BmEkySJEnjYAJKkqQFsuuu444AVq8efxJMkiRJGx//BipJkiRJkqROmYCSJEmSJElSp1yCJ0laUiYmmmVm47ZmzeJYcidJkiQtBiagJElLyrJli2OPo8WQCJucXDwbsEuSJGnjZgJKkqQOLJZEmCRJkrQYuAeUJEmSJEmSOmUCSpIkSZIkSZ0yASVJkiRJkqROmYCSJEmSJElSp0xASZIkSZIkqVMmoCRJkiRJktQpE1CSJEmSJEnqlAkoSZIkSZIkdcoElCRJkiRJkjplAkqSJEmSJEmd2nTcAUiSpIUzOTnuCBoTE7DMP4NJkiRtNExASZK0EVm5ctwRNFavhhUrxh2FJEmSFop/e5QkSZIkSVKnRpKASvKgJH+X5JNJLk1SSWoe/W2T5NgkFye5ri3fmuSO07TZJMmLk/wwyTVJ1iT5eJJd5hqHJEmSJEmS5m9US/COAp40io6SLAdOA+4JXAh8GtgVOAJ4YpKHVdXlU9osAz4BHAT8BjgFWA4cAuyX5FFV9Z1RxCdJkiRJkqTZGVUC6jTgbOC77eMiYPM59vVWmuTTJ4GnV9WNAEn+DXghcAxw6JQ2h9Ekn84H9qqqX7VtDgZOAD6cZJdeX5IkbQwmJpq9lsZtcnLx7D0lSZKk8RhJAqqq3tj/PMmc+kmyHfBM4Hrg+VMSRi8DngE8K8nLq6r/R+oj2/LlveRTG9eJST4DHEgzQ+vEOQUmSdIGaNkyN/qWJEnS4rDYNiF/Ak1M3+hPJAFU1XXAycAmwL6940nuDuwCXEOz9G6qE9rygC4CliRJkiRJ0vQWWwLqAW155pDzveO7DWjzf1V1wwzbSJIkSZIkaYEstgTU3dry0iHne8d3nGcbSZIkSZIkLZBRbUI+Klu15dVDzl/VllvPs820kpwz5NTOM+1DkiRJkiRJjcU2A0qSJEmSJElLzGKbAXVlW24x5PyWbXnFPNtMq6p2HXS8nRnljaQlSZIkSZJmYbEloH7WltsPOd87fvE820jS6KxbB2vXjjsKmJwcdwSSNlSL4d+PiQlY5uR8SZKWqsWWgDqrLXcfcr53/OwBbe6X5DYD7oQ3qI0kjc7atbDttuOOQpLmbuUimOC9ejWsWDHuKCRJUkcW25+ZPgesA/ZKcovf5pJsDhwA3AR8tne8qn4K/Ai4HbDfgD4PacuTuwhYkiRJkiRJ0xtLAirJ4UnOTfL6/uNV9Qvgo8BmwDuT9M/QehOwAvhQVa2e0uUxvTr9iaskTwEOBH4CnDTilyFJkiRJkqQZGMkSvCT7AUf1HdqsPX5637HXVtUp7dfLgfsA2w3o7m+BPYGDgXOTnAHsCtwPOB84ckCb44B9gYPaNl9ux9gbuAZ4VlXdOKcXJ0mSJEmSpHkZ1R5QK4CHDjj+0Cl11quqJpPsARwNPJkmqfQr4N+AV1fVbwa0WZfkqcARwGHA/sBVwIltm1UzfSGSNBKrVsHy5eOOotnUV5L6TUw0+y2N0+Tk4th3SpIkLZiRJKCq6njg+FnUP5omwTTs/OXAi9rHTPu8iWYp3jHrqytJnVu+3M10JS1Oy5b575MkSVpwi20TckmSJEmSJC0xJqAkSZIkSZLUKRNQkiRJkiRJ6pQJKEmSJEmSJHXKBJQkSZIkSZI6ZQJKkiRJkiRJnTIBJUmSJEmSpE6ZgJIkSZIkSVKnTEBJkiRJkiSpUyagJEmSJEmS1CkTUJIkSZIkSeqUCShJkiRJkiR1ygSUJEmSJEmSOmUCSpIkSZIkSZ3adNwBSJKkjc/k5LgjgIkJWOaf4iRJkhaECShJkrTgVq4cdwSwejWsWDHuKCRJkjYO/t1PkiRJkiRJnTIBJUmSJEmSpE6ZgJIkSZIkSVKn3ANKkiR1amKi2W9pnCYnF8e+U5IkSRsrE1CSJKlTy5a52bckSdLGzgSUJEmSNj7r1sHateOOYrCJiSZzK0nSEmICSpIkSRuftWth223HHcVgq1c7bVCStOT4pxVJkiRJkiR1ygSUJEmSJEmSOmUCSpIkSZIkSZ1yDyhJkiQJYNUqWL58YcecnISVKxd2TEmSxsAElCRJkgRN8snNvyVJ6oRL8CRJkiRJktQpE1CSJEmSJEnqlAkoSZIkSZIkdcoElCRJkiRJkjplAkqSJEmSJEmdMgElSZIkSZKkTpmAkiRJkiRJUqdMQEmSJEmSJKlTJqAkSZIkSZLUqZEloJLcLsk/JTkvybVJLktyXJK7zqKPQ5PUDB5/PqXd8eup/9xRvU5JkiRJkiTNzqaj6CTJbYGvAHsCvwBOAnYC/gLYP8meVXXhDLr6CfCBIefuADy5/frUIXU+D/xywPEfz2BsSZK0EZmcHHcEMDEBy5yPLkmSNgIjSUABr6JJPp0GPK6qrgRIciTwFuA4YJ/1dVJVpzIkuZTkeTQJqG9Ok8x6Q1V9dZaxS5KkjdDKleOOAFavhhUrxh2FJElS9+b9N7ckmwGHt09f0Es+AVTVMcDZwN5JHjTPoZ7Vlv85z34kSZIkSZK0gEYx6fsRNMvjLqiq7w84f0JbHjDXAZLcHXg4cD3w8bn2I0mSJEmSpIU3iiV4D2jLM4ec7x3fbR5j9GY/nVJVv56m3lOSHAxsAvwUOLmqzp3HuJIkSZIkSZqnUSSg7taWlw453zu+4zzGmOnyuxdOef7GJP8OHFFVN850sCTnDDm180z7kCRJi8fERLPf0jhNTi6OfackSZLGYRQJqK3a8uoh569qy63n0nmSPYB7A5cDpwyp9n2aDdC/QpPwujPwROB1wPNplu69eC7jS1rk1q2DtWvHG8NiuJWWpGktW+Zm35IkSeM0qrvgdak3++njVXX9oApVdeyUQz8F3pnkazRLAA9PckxVXTKTAatq10HH25lR/u1SWkzWroVttx13FJIkSZKkaYxiE/LeXe+2GHJ+y7a8YrYdJ9kUeHr7dNZ3v6uqc4DP0CTaHjPb9pIkSZIkSZq/USSgftaW2w853zt+8Rz6fhywLXBhVX1rDu0Bzm/L7ebYXpIkSZIkSfMwigTUWW25+5DzveNnz6Hv3vK7D82hbc82bXnVtLUkSZIkSZLUiVHsAfVN4LfAzkkeWFU/mHL+kLY8eTadJtkKeFL7dE4JqCSbA/u1T8+cSx+SNkCrVsHy5eONYWJivONLkiRJ0iIy7wRUVV2f5O3APwDvSPK4qroKIMmRwG7A16rqe702SQ4HDgc+VVWvHNL1U2j2lTq9qs4fUock9wUeQrNJ+XV9x1cA7wZ2oJml9c15vExJG5Lly73dlSRJkiQtIqO6C97rgMcCDwfOT/INYEfgocAa4LAp9ZcD92H6fZl6y+/Wt/n4nYEPAscmOaMd7y7Ag4CtgUuBp1VVzfjVSJIkSZIkaWRGsQcUVXUt8CjgtcDVwJNpElDHA7tX1YWz6S/JdsCjgRuA/1pP9fOAtwI/Bu4PPBV4MM3m468Bdquq82YzviRJkiRJkkZnVDOgqKprgH9sH+urezRw9DTnfzHT2KrqMuDFMwpSkiRJkiRJC24kM6AkSZIkSZKkYUxASZIkSZIkqVMmoCRJkiRJktQpE1CSJEmSJEnqlAkoSZIkSZIkdcoElCRJkiRJkjplAkqSJEmSJEmdMgElSZIkSZKkTpmAkiRJkiRJUqdMQEmSJEmSJKlTJqAkSZIkSZLUKRNQkiRJkiRJ6pQJKEmSJEmSJHXKBJQkSZIkSZI6ZQJKkiRJkiRJnTIBJUmSJEmSpE6ZgJIkSZIkSVKnTEBJkiRJkiSpUyagJEmSJEmS1CkTUJIkSZIkSeqUCShJkiRJkiR1ygSUJEmSJEmSOmUCSpIkSZIkSZ0yASVJkiRJkqRObTruALR+69bB2rXjjWFycrzja5HywylJkiRJmgETUBuAtWth223HHYU0gB9OSZIkSdIMuARPkiRJkiRJnXIGlCRJksZvoZdUu4RbkqQFZQJKkiRJ47dy5bgjkCRJHTIBtYFatQqWLx9vDBMT4x1fi5QfTkmSJEnSFCagNlDLl8OKFeOOQhrAD6ckSZIkaQo3IZckSZIkSVKnnAElSZKkhTUxAatXjzuKW3MJtyRJnTEBJUmSpIW1bJnLtSVJ2si4BE+SJEmSJEmdcgaUtCFatw7Wrh13FDA5Oe4IJEmSJEkbABNQ0oZo7VrYdttxRyFJkiRJ0oyMbAlektsl+ack5yW5NsllSY5LctdZ9nNRkprmcd8h7TZJ8uIkP0xyTZI1ST6eZJfRvEJJkiRJkiTNxUhmQCW5LfAVYE/gF8BJwE7AXwD7J9mzqi6cZbcfGHL8twPGXwZ8AjgI+A1wCrAcOATYL8mjquo7sxxfkiRJkiRJIzCqJXivokk+nQY8rqquBEhyJPAW4Dhgn9l0WFWHzqL6YTTJp/OBvarqV+34BwMnAB9OsktV3TibGCRJkiRJkjR/805AJdkMOLx9+oJe8gmgqo5J8mxg7yQPqqrvzXe8IY5sy5f3kk/t+Ccm+QxwIPAk4MSOxpfGb9UqWL583FHAxMS4I5AkSZIkLTKjmAH1COAOwAVV9f0B508AdgMOAEaegEpyd2AX4BqapXeDxj+wHd8ElJau5cthxYpxRyFJkiRJ0q2MIgH1gLY8c8j53vHdZtNpkpcBOwPXAecAn6qqNdOM/39VdcOoxpckSZIkSdJojCIBdbe2vHTI+d7xHWfZ75umPP/XJC+squO6Hj/JOUNO7TzTPiRJkiRJktRYNoI+tmrLq4ecv6ott55hf58BnkKTMNoCuB9wDLA58N4kT+p4fEmSJEmSJI3QqO6CNzJV9aIph84BXpLkXODdwBuBkzqOYddBx9uZUSu7HFuSJEmSJGmpGcUMqN5d77YYcn7LtrxinuO8D1gN3CfJTmMYX5IkSZIkSXMwigTUz9py+yHne8cvns8gVbUOuKB9ut1Cjy9JkiRJkqS5GcUSvLPacvch53vHzx7BWNu05VV9x3rj3y/JbQbcCW+U40uwbh2sXTveGCYnxzu+JEmSJEmzMIoE1DeB3wI7J3lgVf1gyvlD2vLk+QySZFfgPjSbjZ/bO15VP03yI2AXYD/g012ML/3e2rWw7bbjjkKStAQslr8nTEzAslHMi5ckSRpi3gmoqro+yduBfwDekeRxVXUVQJIjgd2Ar1XV93ptkhwOHA58qqpe2Xd8X+DaqvpK/xhJdgM+BgR4b1VdPyWMY4D3AG9K8q2qWt22ewpwIPATOt64XJIkabZWLpJbm6xeDStWjDsKSZK0lI3qLnivAx4LPBw4P8k3gB2BhwJrgMOm1F9OM5tpuynH9wBeneRimqV1VwP3oFlGtynwVeDvBox/HLAvcBBwbpIvt2PsDVwDPKuqbpzfS5QkSZIkSdJcjGSydVVdCzwKeC1N0ujJNAmo44Hdq+rCGXb1eZpk0u+AR9Asn7sncCrwHOCxVXXNgPHXAU8FXgJcBuwP3B84EXhwVX17ji9NkiRJkiRJ8zSqGVC0iaF/bB/rq3s0cPSA46cBp81x/JtoluIdM5f20rysWgXLl483homJ8Y4vSZIkSdIQI0tASRu15cvdPEOSNK2JiWavpXGbnFw8e09JkqSNhwkoSZKkBbBsmX+rkCRJGy9vuCtJkiRJkqROmYCSJEmSJElSp0xASZIkSZIkqVMmoCRJkiRJktQpE1CSJEmSJEnqlAkoSZIkSZIkdcoElCRJkiRJkjplAkqSJEmSJEmdMgElSZIkSZKkTpmAkiRJkiRJUqdMQEmSJEmSJKlTJqAkSZIkSZLUKRNQkiRJkiRJ6pQJKEmSJEmSJHXKBJQkSZIkSZI6ZQJKkiRJkiRJnTIBJUmSJEmSpE6ZgJIkSZIkSVKnTEBJkiRJkiSpUyagJEmSJEmS1CkTUJIkSZIkSeqUCShJkiRJkiR1ygSUJEmSJEmSOmUCSpIkSZIkSZ0yASVJkiRJkqROmYCSJEmSJElSp0xASZIkSZIkqVMmoCRJkiRJktQpE1CSJEmSJEnqlAkoSZIkSZIkdcoElCRJkiRJkjplAkqSJEmSJEmdMgElSZIkSZKkTpmAkiRJkiRJUqdMQEmSJEmSJKlTI0tAJbldkn9Kcl6Sa5NcluS4JHedRR93TPInST6a5KdJrk9yRZJvJzkiyW2GtDs+SU3zeO6oXqckSZIkSZJmZ9NRdJLktsBXgD2BXwAnATsBfwHsn2TPqrpwBl29FPgHoIAfAN8GVgCPAPYADkny+Kq6ekj7zwO/HHD8xzN+MVq81q2DtWvHHQVMTo47AkmSJEmSNigjSUABr6JJPp0GPK6qrgRIciTwFuA4YJ8Z9HMV8CbgHVX1s97BJPcCvgQ8sh3r74e0f0NVfXVuL0GL3tq1sO22445CkiRJkiTN0ryX4CXZDDi8ffqCXvIJoKqOAc4G9k7yoPX1VVWvr6pX9Cef2uPnA3/XPn3mfGOWJEmSJEnSwhnFHlCPAO4AXFBV3x9w/oS2PGCe45zVlneZZz+SJEmSJElaQKNYgveAtjxzyPne8d3mOc492nLQHk89T0lyMLAJ8FPg5Ko6d57jSpIkSZIkaR5GkYC6W1teOuR87/iO8xzniLY8aZo6L5zy/I1J/h04oqpunOlASc4ZcmrnmfahBbJqFSxfPu4oYGJi3BFIkiRJkrRojSIBtVVbDrsz3VVtufVcB0jyXOCxwG+ANwyo8n2aDdC/QpPwujPwROB1wPOB64EXz3V8LWLLl8OKFeOOQpIkSZIkTWNUd8HrTJK9gGOBAg6rqsum1qmqY6cc+inwziRfo1kCeHiSY6rqkpmMWVW7DonlHGDlbOKXJEmSJEna2I1iE/LeXe+2GHJ+y7a8YrYdJ7kfzZK7zWiW0X1qNu2r6hzgMzSJtsfMdnxJkiRJkiTN3yhmQP2sLbcfcr53/OLZdJrk7sAXgG2Ao6vqbXMLj/Pbcrs5tpckSVrSJifHHUGzneKyUfxpVJIkLUqjSECd1Za7DznfO372TDtMsh3wRZqk0bFV9Zq5h8c2bXnVtLUkSZI2UisXwQYDq1e7raMkSUvZKP7O9E3gt8DOSR444PwhbXnyTDpLsg3weZo7zr2feWwenmRzYL/26Zlz7UeSJEmSJElzN+8EVFVdD7y9ffqOJL09n0hyJLAb8LWq+l7f8cOTnJvk9f19JdkCOAW4P/Bx4DlVVdONn+S+Sf6sTTb1H18BfAzYgWaW1jfn+holSZIkSZI0d6O6C97rgMcCDwfOT/INYEfgocAa4LAp9ZcD9+HW+zL9M/Aw4CbgRuB9SW41WFUd2vf0zsAHgWOTnNGOdxfgQcDWwKXA09aXyJIkSZIkSVI3RpKAqqprkzwKeCXwJ8CTgcuB44GjqurSGXbV269pk7afYQ7t+/o84K3AnjQzpyaA69rjJ9PsIfXrGY4vSZK0pE1MNPstjdPk5OLYd0qSJC2cUc2AoqquAf6xfayv7tHA0QOOH8otk0szGfcy5rFPlCRJ0sZk2TI3+5YkSQvPm91KkiRJkiSpUyagJEmSJEmS1CkTUJIkSZIkSeqUCShJkiRJkiR1ygSUJEmSJEmSOmUCSpIkSZIkSZ0yASVJkiRJkqROmYCSJEmSJElSp0xASZIkSZIkqVMmoCRJkiRJktQpE1CSJEmSJEnq1KbjDkAbiHXrYO3a8cYwOTne8SVJkiRJ0pyYgNLMrF0L22477igkSZIkSdIGyCV4kiRJkiRJ6pQJKEmSJEmSJHXKBJQkSZIkSZI65R5QmrtVq2D58vHGMDEx3vElSZIkSdJ6mYDS3C1fDitWjDsKSZIkSZK0yLkET5IkSZIkSZ0yASVJkiRJkqROmYCSJEmSJElSp0xASZIkSZIkqVMmoCRJkiRJktQpE1CSJEmSJEnqlAkoSZIkSZIkdcoElCRJkiRJkjplAkqSJEmSJEmdMgElSZIkSZKkTpmAkiRJkiRJUqc2HXcAkiRJ0uTkuCOAiQlY5p9nJUnqhAkoSZIkjd3KleOOAFavhhUrxh2FJElLk3/jkSRJkiRJUqdMQEmSJEmSJKlTJqAkSZIkSZLUKfeA2hCsW8etcoWTk0AtXAyLYWdQSZK0JExMNPstjdPk5OLYd0qSpI2FCagNweWXA8tveWzlLoBJIUmStOFZtszNviVJ2tiYgJIkSZIY04TvyTD1D40T69wnQ5K09IwsAZXkdsArgWcAdwMuBz4HHFVVP59lX9sARwNPBu4M/BL4FHB0Vf1mSJtNgBcBhwH3BK4E/hd4dVX9aNYvSJIkSRuV8SzJWw6sucWR1ZdPsuIPxhGLJEndGckfV5LcFvgKcBSwFXAScAnwF8D3k9xjFn0tB75Dk0y6Efg0cAVwBPDtJHca0GYZ8AngGGB74BTgHOAQ4Iwke8z1tUmSJEmSJGl+RjUD6lXAnsBpwOOq6kqAJEcCbwGOA/aZYV9vpZnB9Eng6VV1Y9vXvwEvpEkyHTqlzWHAQcD5wF5V9au2zcHACcCHk+zS62tJOPWbcO9txhvDxMR4x5ckSZIkSRuEeSegkmwGHN4+fUEv+QRQVcckeTawd5IHVdX31tPXdsAzgeuB509JGL2MZnnfs5K8vKr6751yZFu+vJd8asc/MclngAOBJwEnzu1VLkJ3uhOsWL7+epIkSZIkSWM2ihlQjwDuAFxQVd8fcP4EYDfgAGDaBBTwBJplgd/oTyQBVNV1SU6mme20L3A8QJK7A7sA19AsvRs0/oHt+EsnASVJkqQ5m5iA1avXX69rk+ddzspH3mqHCUmSlpxRJKAe0JZnDjnfO77biPo6bEpfvTb/V1U3zHN8SZIkbQSWLYMVK8YdBTC5btwRSJK0IEaRgLpbW1465Hzv+I4d9TXK8QFIcs6QU/e94IIL2HXXXWfa1UjcdN1NwCa3OLbXATexyeabDG4gSZKkDcJi+DnvggsuANhhwQaUJG2URpGA2qotrx5y/qq23LqjvkY5/vqsu+66665atWrVJSPoay52bssLfnzBmCLQOPz+uo81Co2D137j5bXfeHntN17j/DlvB4b/LC1J0kiM6i54S0pVLewUpxnqzcxarPGpG173jZfXfuPltd94ee03Xl57SdJSt2wEffTuerfFkPNbtuUVHfU1yvElSZIkSZI0YqNIQP2sLbcfcr53/OKO+hrl+JIkSZIkSRqxUSSgzmrL3Yec7x0/u6O+em3ul+Q28xxfkiRJkiRJIzaKBNQ3gd8COyd54IDzh7TlyTPo63PAOmCvJNv2n0iyOXAAcBPw2d7xqvop8CPgdsB+8xxfkiRJkiRJIzbvBFRVXQ+8vX36jiS9PZdIciSwG/C1qvpe3/HDk5yb5PVT+voF8FFgM+CdSfo3SX8TsAL4UFWtnhLGMb06/YmrJE8BDgR+Apw0j5cpSZIkSZKkOUpVzb+T5LbAV4GHAr8AvgHs2D5fA+xZVRf21T8aeDXwgao6dEpfy4HTaW5FewFwBrArcD/g/Lavy6e0WQacABwE/Br4MrAc2Bu4FnhUVX173i9UkiRJkiRJszaKJXhU1bXAo4DXAlcDT6ZJQB0P7N6ffJpBX5PAHsDbaGZCHQTcAfg3YI+pyae2zTrgqcBLgMuA/YH7AycCDzb5JEmSJEmSND4jmQElSZIkSZIkDTOSGVCSJEmSJEnSMCagJEmSJEmS1CkTUJIkSZIkSeqUCShJkiRJkiR1ygSUJEmSJEmSOmUCakyS3C7JPyU5L8m1SS5LclySu86hr22SHJvk4iTXteVbk9yxg9A1T6O49knumORPknw0yU+TXJ/kiiTfTnJEktt0+Ro0N6P8vp/S772SXJOkknxpVPFqdEZ97ZPslOQ/2u//65JMJjktyctGHbvmZ8T/3/9xklOSrElyQ5K1Sb6Q5KAuYtfcJXlQkr9L8skkl7b/Ps/51tP+rCdJWgpSNef/CzVHSW4L/C+wJ/AL4BvATsAewBpgz6q6cIZ9LQdOA+4JXAicAezaPs4DHlZVl4/4JWiORnXtk7wO+AeggB/QXOsVwCOAzYFTgcdX1dUjfxGak1F+3w/o+3+BvYEAX66qx44iZo3GqK99kicCJwC3A84EzgcmgPsDV1XVPUcZv+ZuxP/f/y3wrzT/7p8GXALsADyM5nv/X6rqH0b7CjRXST4NPGnq8arKHPryZz1J0pLgDKjxeBXND6OnAfeuqqdX1UOBl9AkEY6bRV9vpfmB5JPAfdq+7ge8Dbg3cMwoA9e8jeraXwW8CdipqnavqmdU1WNofgH9GfDIdiwtHqP8vv+9JH8J7AO8Z0RxavRGdu2T3Jfm3/urgEdW1YOr6plV9TjgrsAzRh695mMk1z7JCuANwA3Ao6rqEe2/+4+g+f6/Dnhlknt08Bo0N6cBrwUOBLajuUZz9Vb8WU+StAQ4A2qBJdkMWA3cAdi9qr4/5fxZwG7Ag6vqe+vpazvgUuBG4G5V9au+c5vT/HX0TsBdqmr1SF+IZm2U13494zwT+AhwUVXdfR4ha0S6uvZJ/gD4Ec1fw/+FZqaFM6AWkVFf+ySfBZ4I7FdVn+0gZI3IiP+/3x84Gfh8VT1hwPmTaBIdT6+qj4/oJWiEklwLbD7bGVD+rCdJWkqcAbXwHkHzw+gFU38YbZ3QlgfMoK8n0FzDb/T/QAJQVdfR/LC6CbDv3MPVCI3y2k/nrLa8yzz70eh0de2PpVmG9fx5xKZujezaJ9kBeDxwocmnDcIov+9nOntm7QzracPhz3qSpCXDBNTCe0BbnjnkfO/4bgvcl7q3UNertwTjl/PsR6Mz8mufZF/g6TT7vvxkHrGpW6O89vvQ/L/9rSSbJnlauynx25M8N8k284xVozXKa/8d4DfAo5Ps3X8iyR/RJCbPp9ljSkuLP+tJkpaMTccdwEbobm156ZDzveM7LnBf6t5CXa8j2vKkefaj0RnptU+yJfBO4MfAG+cXmjo2ymu/si2vpEk07Dnl/D8nOaSq/nd2IaojI7v2VfXbdr+3jwD/m+RbbfvtgYcD3wT+vKqun1/IWoT8WU+StGQ4A2rhbdWWw+5OdlVbbr3Afal7nV+vJM8FHkvzl/I3zLUfjdyor/3raH7ZeK6/cC56o7z2vRlOfwXcF/gTmr1f7gN8qP36U0nuOrdQNWIj/b6vqk/S7P+1lmZ539Pb8grgC8DP5xypFjN/1pMkLRkmoKQlIsleNHsCFXBYVV025pDUgSQPBl4EfLCqvjrmcLSwev9nbwr8TVV9tKp+XVXnVdWfAd+l2XPIPcGWoCQvAb4EfJ1mudVWbfkV4J9o7pAmSZK0aJmAWnhXtuUWQ85v2ZZXLHBf6l5n1yvJ/WiW3G0GHFFVn5p9eOrQSK59kk2B99DMcHvpSCJT17r4N/9K4BMDzr+/LfcecE4Lb2TXPsk+wJuBHwBPraofVtVVVfVD4JD2+H5JnjiPeLU4+bOeJGnJcA+ohfezttx+yPne8YsXuC91r5PrleTuNMsvtgGOrqq3zS08dWhU13574IE0G8x/IrnF3bzv2JYPSvJVgKraZ3ZhqgOj/L7v1flZVdWA8xe15bYzC00dG+W1/7O2/FRVres/UVU3Jfkkzb8NfwT8zyzj1OLmz3qSpCXDBNTCO6stdx9yvnf87AXuS90b+fVKsh3wRWA74Niqes3cw1OHRn3t79w+BrkjzoBZTEZ57b/flsPudnentrxyyHktrFFe+16S4bdDzveOeyfEpcef9SRJS4ZL8BbeN2l+UNw5yQMHnD+kLU+eQV+fA9YBeyW5xV+8k2wOHADcBHx2ztFqlEZ57Wlvuf55YGeapTcvHkGM6sZIrn1VXVRVGfQAHtVW+3LfMY3fKL/vv0WzAfWdk9xnwPle4vH7A85p4Y3y2v+yLR885PxD2vKimQanDYY/60mSlgwTUAusvWPV29un72hvpw5AkiNpNhT9WlV9r+/44UnOTfL6KX39Avgozb4/72z3h+l5E7AC+FBVre7m1Wg2Rnntk2wBnALcH/g48JwhS3K0CIzy2mvDMuJ/828EjgHS9nX7vjaPBQ6luQnBuzp6OZqFEX/ff7ot/zTJ/v0nkjyJ5o6I6wD3/9tA+bOeJGlj4BK88Xgd8Fjg4cD5Sb5Bc0v1hwJrgMOm1F9Oc5vt7Qb09bfAnsDBwLlJzgB2Be4HnA8c2UH8mrtRXft/Bh5G81fPG4H3TdkPCICqOnSEsWt+Rvl9rw3LKK/9/6OZ7fZY4Lwkp7f19wQ2Af6hqr7TxYvQnIzq2n+aZuP5pwInt//X/xS4OzfPivqHqvpxB69Bc5BkP+CovkObtcdP7zv22qo6pf3an/UkSUueM6DGoKqupfkF4rXA1cCTaX4gPR7YvaounEVfk8AewNtofrg5iOY23P8G7FFVl48yds3PCK99b5+PTWj+8v3sIQ8tEqP8vteGZcT/5t8A7Au8ApgEHk8zE/JrwAFV9S+jjF3zM6pr385wfTrwl8DXgXvS/H+/E83Sqyd67RedFTSJxt6j91ei/mMrZtKRP+tJkpaKuGpHkiRJkiRJXXIGlCRJkiRJkjplAkqSJEmSJEmdMgElSZIkSZKkTpmAkiRJkiRJUqdMQEmSJEmSJKlTJqAkSZIkSZLUKRNQkiRJkiRJ6pQJKEmSJEmSJHXKBJQkSZIkSZI6ZQJKkiRJkiRJnTIBJUmSJEmSpE6ZgJKkBZKkklw0j/b7tH18dYb1j2/rHzrXMdt+Dm376X/ckOTSJB9Nsvt8+p9hDEeP4rV0MU6SnQZdl77rdfyU47338+hRxCxJkiRtCExASZJm6gLgA+3jJOBG4BnA6UkOHGdgS0FfwnCfccciSZIkjdqm4w5AkjYiuwA3jDuIeTi1qg7tPUlyG+CdwF8B/5Hkc1V1/biCG6Of01zbq2dY/1PA6cBkZxFJkiRJi4wzoCRpgVTVuVV1wbjjGJWqugH4W+BKYDtgj7EGNCZVdUN7bX82w/q/beubgJIkSdJGwwSUJC2QYXtAJXl4kk8nuTjJdUl+meQ7Sd6QZKsxhDpjVXUVcF77dAe45R5HSe6d5GNJfpVkXZIn99om2TfJF5P8Osm1SX7cvuY7Tjdmkocm+XyS3yT5XdvHngPqJckz2/HPS3JVkiva9/b5Sab9P3AW4wzcA2qafm+1B1SSAp7dPv3fKftt7ZTk7e3Xfz1Nvz9u3+N7zCQOSZIkaSGZgJKkMUpyAPAN4EDgF8Ange8DdwJeASwfX3QztnVbXjfl+H2A79LMjPpf4Iu0SxCTvBI4BdgH+B7waWALmtf87SR/MGSshwNfB7YH/gf4MfBY4GtJHjel7ubAR9rzvwROpln6tivwDuC4aV7TbMYZhQ/Q7LEF8Hlu3mvrAzQzzN7VnnvOoMZJ9gbuDXy5qi7sID5JkiRpXtwDSpLG66U0fww4pKpO7D+R5CHA2rFENUNJVgK9GTdnTzn9DODtwN9W1U19bR4CvI4msfLYqvp2e3xz4D+Bp9IkiA4ZMORzgH8Gjqqqats9j2YvquOT7FxV17R1bwQOAk5plwv2xl8BfBZ4dpLjqurr8xxn3qrq0PZueTsDb6iqr06pMpnkW8DDkzywqn4wIF6A94wqJkmSJGmUnAElSeO1oi2/NPVEVX23qq5Y4HhmJMmWSR5DM2NrE+BLVfWTKdXWAK/oTz61Dqf5/+dtveQTQFVd1567BjgoyQ4Dhr4YOLqXFGrb/TvwbZp9qA7uO35jVX26P/nUHl8DvLJ9+qQhL3HG4yyg/2jLW8yCSrINTTxraGaSSZIkSYuOCShJGq/vteV/JnnI+vYlGrNn9/Ylopm99CWaZXZnAH82oP6XqmrQneH2assPTz1RVauBL9D8//SIAW1PrKobBxz/6JS+fy/JA5O8PMk7kry/nWn0vPb0vQb0NadxFsAnaGbE/WmSLfqOPwu4LfCBjfQuhJIkSdoAuARPksbr74H7Awe0j18nORX4DPChqrp2nMFNcQFwavv1DcCvaPav+mJVrRtQf9hd4e7SlhcNOd87ftcB5y5eT5te3yTZDDgeeOaQNnDz/lVzHmehVNW1ST4AHEmzTPED7am/asv3LnRMkiRJ0kwt5r+0S9KSV1WXAA8GHg+8DbiEJhH1HuDsJBNjDG+qU6vq0PbxnKp6VVV9fkjyCWCuybNaf5UZOZIm+fRD4InAHwCbVVVoZm4BZERjLZR30bw/zwFIsgewG/D1qvrxOAOTJEmSpmMCSpLGrN2r6AtV9aKqegCwE/AVmuVhrxhrcN24rC13HHJ+p7b8+YBzw9r0jl/Wd+ygtnxmVX2uqlb37Qd1D6Y3m3EWTFWdR3NHwUck2YWb94N69zjikSRJkmbKBJQkLTJVdTHwxvbp/cYZS0e+0Za3WhrX3qHu8TSzfL45oO1Tkmwy4Pgz2vLUvmPbtOWlA+o/bT0xzmacUent37S+5fG9zciPbOP5NXDi8OqSJEnS+JmAkqQxSvLiJHcecGrftrxkIeNZIO8A1gEvSvLg3sF2z6a3AbcDPtkuT5xqJ+DV/QeS/DXwMJo9qfoTMee15XOn1D8E+PP1xDibcUalN6vqPtPWau5090uavZ+2Av5zke0VJkmSJN2Km5BL0ni9GnhzkrOA82n2JHoAcG/gcuDNA9rsnuT0afp8flWd2ff8qCTPHVL3F1V10JBznaiq7yQ5Cvhn4LQkXwUmae56twPN+/CCIc3fA/xdkqcAZwP3BB5Csyn6oVPuuvcm4AnAG5I8lSYhdS+aPbfeDLx0mjBnM86onAz8I83n4Y9p3hOAV1TV2l6lqrohyXE0G9iDy+8kSZK0ATABJUnj9UKaJMmDaDbKhmbW0zHAMVU1aB+krYGHTtPn7ac8vwfD9zwadre3TlXVv7RJtxfTJHZuR3PXvDcBb6iqXw9p+i3g/cBrgf1pEnZfBv6xqr41ZYyvJ3kkTaLrD2mSej8EDgbOZPoE1IzHGZWq+l6SZwEvAR5H854AvA5YO6X6V2gSUKdV1TldxCNJkiSNUqpGdbMhSZK0EJK8C/hr4C+q6vgxhyNJkiStlwkoSZI2IEl2BH4EXAXcraquGXNIkiRJ0nq5BE+SpA1AkpcBuwF/TLM875UmnyRJkrShcAaUJEkbgHaz9r1p7pb3HuA15X/ikiRJ2kCYgJIkSZIkSVKnlo07AEmSJEmSJC1tJqAkSZIkSZLUKRNQkiRJkiRJ6pQJKEmSJEmSJHXKBJQkSZIkSZI6ZQJKkiRJkiRJnTIBJUmSJEmSpE6ZgJIkSZIkSVKnTEBJkiRJkiSpUyagJEmSJEmS1CkTUJIkSZIkSeqUCShJkiRJkiR1ygSUJEmSJEmSOmUCSpIkSZIkSZ36/9bd5i9YZTW6AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "Y_predict_train_LL = dnnLL.predict_proba(X_LL)\n",
    "Y_predict_val_LL = dnnLL.predict_proba(X_val_LL)\n",
    "\n",
    "axis = np.linspace(0,1,20)\n",
    "\n",
    "# targets = ['isLL', 'notLL']\n",
    "# predictions = ['isLL']\n",
    "# colors = ['r', 'b']\n",
    " \n",
    "# Y_targets_LL = {}\n",
    "# Y_predictions_LL = {}\n",
    "\n",
    "# for target in targets:\n",
    "#     Y_targets_LL[target] = dnnLL.predict_proba( df_val_LL[ df_val_LL[target]==1 ].values[:,0:NDIM] )\n",
    "#     Y_predictions_LL[target] = {}\n",
    "#     for iprediction, prediction in enumerate(predictions):\n",
    "#         Y_predictions_LL[target][prediction] = Y_targets_LL[target][:,iprediction]\n",
    "\n",
    "# for prediction in predictions:\n",
    "#     for target, mycolor in zip(targets, colors):\n",
    "#         plt.hist(Y_predictions_LL[target][prediction].numpy(), bins = axis, label = prediction + ' class - ' + target + ' sample', histtype='step', color=mycolor,  density=True, linewidth=2 )\n",
    "#     plt.title('Validation dataset')\n",
    "#     plt.xlabel('isLL Probability')\n",
    "#     plt.legend(frameon=False, bbox_to_anchor=(1.05, 1.))\n",
    "#     #plt.yscale('log')\n",
    "#     plt.show()\n",
    "\n",
    "plt.hist(dnnLL.predict_proba( df_val_LL[ df_val_LL['isLL']==1 ].values[:,0:NDIM] ).numpy()[...,0], bins = axis, label = 'isLL class - isLL sample', histtype='step', color='r',  density=True, linewidth=2 )\n",
    "plt.hist(dnnLL.predict_proba( df_val_LL[ df_val_LL['isLL']==0 ].values[:,0:NDIM] ).numpy()[...,0], bins = axis, label = 'isLL class - isnotLL sample', histtype='step', color='b',  density=True, linewidth=2 )\n",
    "plt.title('Validation dataset')\n",
    "plt.xlabel('isLL Probability')\n",
    "plt.legend(frameon=False, bbox_to_anchor=(1.05, 1.))\n",
    "#plt.yscale('log')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "ac511b3c-889a-41a0-8484-76841bf345c2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classifier categorical accuracy =  0.6912054451012001\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "from scipy.stats import ks_2samp, anderson_ksamp\n",
    "\n",
    "X_LL = df_val_LL.values[:,0:NDIM]\n",
    "Y_true_LL = df_val_LL[['isLL', 'notLL']].values\n",
    "Y_pred_LL = dnnLL.predict_proba(X_LL)\n",
    "\n",
    "# The classifier categorical accuracy is a good and simple metrics to evaluate the overall classification performance\n",
    "Y_true_max_LL = np.argmax(Y_true_LL, axis=1)\n",
    "Y_pred_max_LL = np.argmax(Y_pred_LL, axis=1)\n",
    "print(\"Classifier categorical accuracy = \", accuracy_score(Y_true_max_LL, Y_pred_max_LL))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "21632233-f418-4238-af28-48a0d7c934ba",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "C = confusion_matrix(Y_true_max_LL, Y_pred_max_LL, normalize=\"true\")\n",
    "C = np.around(C, decimals=3)\n",
    "\n",
    "plt.figure()\n",
    "plot_confusion_matrix(C, classes=['isLL', 'notLL'], title='')\n",
    "plt.show(1) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "86c893f8-a7b7-441f-b144-e66a18bb4906",
   "metadata": {},
   "outputs": [],
   "source": [
    "dnnTT = SimpleNeuralNetwork(350, learning_rate=0.0001, N_NODES=50, n_layers=8, n_features=X_TT.shape[1], n_outputsC=2)\n",
    "dnnTT.save_weights(\"TTvsOther_inclusive_RDF\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "4c4e300d-5fa1-4e71-a5ef-267d5beb01e9",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 350/350 [00:15<00:00, 22.59it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "acc = dnnTT.fit (X_TT.astype(np.float32), Y_TT.astype(np.float32), X_val_TT.astype(np.float32), Y_val_TT.astype(np.float32))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "c961815e-a9f8-42ae-996e-6679723ba022",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "Y_predict_train_TT = dnnTT.predict_proba(X_TT)\n",
    "Y_predict_val_TT = dnnTT.predict_proba(X_val_TT)\n",
    "\n",
    "axis = np.linspace(0,1,20)\n",
    "\n",
    "# targets = ['isTT', 'notTT']\n",
    "# predictions = ['isTT', 'isnotTT']\n",
    "# colors = ['r', 'b']\n",
    " \n",
    "# Y_targets_TT = {}\n",
    "# Y_predictions_TT = {}\n",
    "\n",
    "# for target in targets:\n",
    "#     Y_targets_TT[target] = dnnTT.predict_proba( df_val_TT[ df_val_TT[target]==1 ].values[:,0:NDIM] )\n",
    "#     Y_predictions_TT[target] = {}\n",
    "#     for iprediction, prediction in enumerate(predictions):\n",
    "#         Y_predictions_TT[target][prediction] = Y_targets_TT[target][:,iprediction]\n",
    "\n",
    "# for prediction in predictions:\n",
    "#     for target, mycolor in zip(targets, colors):\n",
    "#         plt.hist(Y_predictions_TT[target][prediction].numpy(), bins = axis, label = prediction + ' class - ' + target + ' sample', histtype='step', color=mycolor,  density=True, linewidth=2 )\n",
    "#     plt.title('Validation dataset')\n",
    "#     plt.xlabel(prediction + ' Probability')\n",
    "#     plt.legend(frameon=False, bbox_to_anchor=(1.05, 1.))\n",
    "#     #plt.yscale('log')\n",
    "#     plt.show()\n",
    "\n",
    "plt.hist(dnnTT.predict_proba( df_val_TT[ df_val_TT['isTT']==1 ].values[:,0:NDIM] ).numpy()[...,0], bins = axis, label = 'isTT class - isTT sample', histtype='step', color='r',  density=True, linewidth=2 )\n",
    "plt.hist(dnnTT.predict_proba( df_val_TT[ df_val_TT['isTT']==0 ].values[:,0:NDIM] ).numpy()[...,0], bins = axis, label = 'isTT class - isnotTT sample', histtype='step', color='b',  density=True, linewidth=2 )\n",
    "plt.title('Validation dataset')\n",
    "plt.xlabel('isTT Probability')\n",
    "plt.legend(frameon=False, bbox_to_anchor=(1.05, 1.))\n",
    "#plt.yscale('log')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "5948ad04-3997-4577-890b-0cd619f9351c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classifier categorical accuracy =  0.6858638743455497\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "from scipy.stats import ks_2samp, anderson_ksamp\n",
    "\n",
    "X_TT = df_val_TT.values[:,0:NDIM]\n",
    "Y_true_TT = df_val_TT[['isTT', 'notTT']].values\n",
    "Y_pred_TT = dnnTT.predict_proba(X_TT)\n",
    "\n",
    "# The classifier categorical accuracy is a good and simple metrics to evaluate the overall classification performance\n",
    "Y_true_max_TT = np.argmax(Y_true_TT, axis=1)\n",
    "Y_pred_max_TT = np.argmax(Y_pred_TT, axis=1)\n",
    "print(\"Classifier categorical accuracy = \", accuracy_score(Y_true_max_TT, Y_pred_max_TT))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "68a8bade-814c-414d-8c1d-fd64024cac1d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAscAAAJJCAYAAABGTJqGAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAABcSAAAXEgFnn9JSAAB0GElEQVR4nO3dd5yU1fXH8c+h946CCoKoWABRLNh7L7HF/NQktsRYsJeosWBJ7DXWJNYYE3vvihXFAgiKYkGaigqi9M75/XHvsLO7M7szu8/szLLft695PTNPuc+dYcc5c+c855q7IyIiIiIi0KjYHRARERERKRUKjkVEREREIgXHIiIiIiKRgmMRERERkUjBsYiIiIhIpOBYRERERCRScCwiIiIiEik4FhERERGJFByLiIiIiEQKjkVEREREIgXHIiIiIiKRgmMRERERkUjBsYiIiIhI1KTYHRCpyMy+B1oBU4vdFxERWan0AOa7e7didyQbM3sK6JNwsxPcfb+E21xpKTiWUtQKa9zWWnTYoNgdEakP+vZatdhdEKkXvp4wgWbNmrYtdj+q0QdrtIE1b59IY75oFvjyRNpqKBQcSymaai06bNC8/1HF7odIvfD+G1cWuwsi9cLmm/QvdhdyYs3b03y9wxJpa9H4B/CFPyfSVkOh4FhERESkpBhYUpeFWULtNBwKjkVERERKjSmoLRZVqxARERERiTRyLCIiIlJqEkurkHwpOBYREREpNUqrKBp9LRERERERiTRyLCIiIlJSVK2imDRyLCIiIiISaeRYREREpJQYyeUca+A4bwqORUREREqNqlUUjV55EREREZFII8ciIiIipUal3IpGwbGIiIhISVG1imJSWoWIiIiISKSRYxEREZFSo7SKotHIsYiIiIhIpJFjERERkVKjUm5Fo+BYREREpJRoEpCi0tcSEREREZFII8ciIiIiJUWl3IpJwbGIiIhIqVHOcdHolRcRERERiTRyLCIiIlJqGikdolg0ciwiIiIiEmnkWERERKSk6IK8YlJwLCIiIlJKVOe4qJRWISIiIiISaeRYREREpNSolFvRKDgWERERKTVJpVVI3vS1REREREQk0sixiIiISElRtYpi0sixiIiIiEikkWMRERGRUqOc46JRcCwiIiJSalStomj0youIiIhItcyspZldYmZfmNlCM/vOzO4ys9Vr2F4vM7vdzCaa2SIzm2Fm75rZWRn2HWpmXsXtito/w0AjxyIiIiKlpARnyDOzFsAwYDAwDXgS6AUcBexjZoPd/es82tsTeARoCYwCRgCdgf7An4Crsxw6HPgqw/qRuZ67OgqORUREREpKSVarOJ8QGL8L7ObucwHM7HTgWuAuYIecemS2HvAYMAfY1d3fSdvWCNikisP/5e731KD/OVNahYiIiIhkZWbNgCHx4YmpwBjA3a8DxgLbm9mgHJu8DmgBHJkeGMf2lrv7hwl0u8YUHIuIiIiUGrNkbsnYGmgPTHD30Rm2PxKX+1b/tKwHsDvwtbs/l1QHk6S0ChERERGpykZxOSrL9tT6ATm0tQNhcPYdM2sCHEgIvhsDnwAPuvvPVRy/k5kNJIw8fwM87+6J5RuDgmMRERGR0lNapdx6xuU3Wban1q+ZQ1sbxOVc4C1CHnO6v5rZwe7+Wpbjf1fh8aVm9ighRWNupgPyVVKvvIiIiIgQguMkbkEfMxuX6ZZjb9rE5fws2+fFZdsc2uoYl38A1gMOAzoBfYH74/3HM5SH+wo4E9gw9qcHcDjwLXAQ8O8czp0TBcciIiIiUldSsWcT4E/u/l93/9ndv3D33wEfEPKbT0g/yN3vd/dr3f1Td5/n7t+4+wPAZsBPwP5mVnEUukaUViEiIiJSUpK8mM4gXEi3YS0aSaUrtMqyvXVczsmjrbnAwxm2300IeLfPpWPuPs3M7iaMKu9BqJdcKwqORUREREqJkVzOcTIx9pS4XCPL9tT6yTm0ldpnirt7hu2T4nKV3LoGwJdx2T2PY7JSWoWIiIiIVGVMXGabnCO1fmwObaVKwXXMsr1TXOZzcV2qrXlV7pUjBcciIiIipaa06hwPB2YRLuwbmGH7wXH5dA5tvUPIEe5mZn0zbE+lU2Sqp1yJmRlwQHyYrdRcXhQci4iIiJSaZKtV1Iq7LwZujg9vMbNUjnFq+ugBwBvp9YbNbIiZjTezyyu0tZQwQ57FttqlHbMLcCTgwB1p67ua2YlmVq4ahpm1AW4DtgC+J0xJXWvKORYRERGR6lwG7AJsBXxpZm8R6hpvAUwHjq6wfxdCebZMecBXAzvG9r4wsxFx/8GEyUD+4u7vp+3fmhCcX2FmHwDTgK6EdI7OwC/Awe6erdRcXjRyLCIiIlJSEkqpMCOpK/LcfSEhoL2UUO94f0JwfA+wibt/nUdbS4C9gD8DMwjTSfcH3gD2dfe/VTjkJ+BKYCSwLqGu8daE0eJrgX7uPryGT60SjRyLiIiISLXcfQFwYbxVt+9QYGgV25cAV8VbdW3NAc7JtZ+1peBYREREpMRYchfTSZ4UHIuIiIiUGAXHxaOcYxERERGRSCPHIiIiIqUkuevokmunAVFwLCIiIlJilFZRPEqrEBERERGJNHIsIiIiUkKM5EaOjTDdnOROI8ciIiIiIpFGjkVERERKjHKOi0fBsYiIiEhJsQSDYwXZ+VJahYiIiIhIpJFjERERkVKjAd+iUXAsIiIiUkoswZxjBdl5U1qFiIiIiEikkWMRERGREqNqFcWjkWMRERERkUgjxyIiIiIlRiPHxaPgWERERKSEJD19tORHaRUiIiIiIpFGjkVERERKjYZ8i0bBsYiIiEiJUc5x8SitQkREREQk0sixiIiISCkxS3CGPI1A50sjxyIiIiIikUaORUREREqMco6LR8GxiIiISKlRbFw0SqsQEREREYk0ciwiIiJSYpRWUTwKjkVERERKjILj4lFahYiIiIhIpJFjERERkRJiJDdyrPHn/GnkWEREREQk0sixiIiISCmxBHOONXScNwXHIvVci+ZNOOuInfj1rgPpsWoHfp49n5dGfM4ld7zId9Nn591ez+4dOfP3O7Lr4L5079KOOfMXMWHqDJ58/WOuv/+NSvt379KOs4/ciV0H92WNVTuwbPlyJkz9iafe+IQb/vMGc+cvKrd/5/at2Ge7DdmsX0823aAnG661Kk2aNOaPlzzI/c9+WOPXQSRXCxYs4Nqrr+DRhx/km6lT6NixE7vstjvnX3gJq62+ek5tLF26lKuu+CujRn7AF+PHM2PGdJYsWcLqa/Rgx5134bTTz6bnmmtWOu7LLz7nxReeZ+SH7zPyww+YNPFrAD4eP4E11+yV9Xzz58/n5puu55GH/sekiV/TslUrBm26GaedcTbbbrdDTV4GKXUKaovG3L3YfRApx8zGWcvOGzTvf1Sxu1Lymjdrwou3HscW/ddk2vRZDP9oImt278Rm/Xry48y5bH/M35n03cyc29tty748cPnvadm8CaM//5YJU2fQqX1rNuzTjfkLFtPv4CvL7d+nRxeG/eNEVunUhknfzeSjz7+lebMmDO6/Jh3bteLTr79nxz/cwux5C1ccs+92G/LQ1UdWOreC45r78Y0rq99JAFi4cCF7774zH7w/gm7durPV1tswefJkRn74Pl26duXVN96hd++1qm1n7ty5rNa1PW3atGHDfgPo3r07i5cs5uMxY5g6dQrt2rXjqedeZpNBm5Y77s9nnsZtt9xUqb2qguO5c+ey9+47M3rUh3Ts1IktBm/J7Fmz+eD9ESxdupSbb/sHvzvi6Bq9Hg3N5pv0B+CzT8eVbOhpZuOadFxjg66H3JhIe9MfOoWlP3/zqbtvmEiDDYBGjkXqsXOO2pkt+q/JiLGT2OfkfzJvwWIATj50O648dV/uOP8Qdj/h9pzaWnfNrvzviiOYM38R+5z0D0Z8PHnFNjNj476VR9T+euJerNKpDbc/8g5nXPsEy5eHL9vtWrfgqRv/wBb91+Tkw7bjsn++tOKYH2fO5fZH3mHUZ1MZ+elUTvjNNhyz/+DavAwiObvqir/ywfsj2HyLLXnimRdo06YNADffeD3nnXMmJ/7pDzz30rBq22nRogUvvfomm26+BU2alH2ULlu2jEuHXsB111zJaSefwBvD3y933Ib9+nHaGWezyaBN2XiTTTlgvz358ovPqzzX0AvOY/SoD9l4k0E8+sSzdOnaFYD33n2HA/bbk9NOPpHtd9g540i11FeWYCm3kv0eULJ0QV4BmdkkM/N461Xs/sjKpWmTxhz3660BOPXqx1cExgA3/fdNxn75HdsN6sPG6+X2M/GVp+5LyxZNOfbSB8sFxgDuzqjx31Q6ZuuNwwjbFXe+siIwBpg9byHX/ft1AAZtsEa5Y977ZDKnXf04/37mQz79+odyx4kU0uLFi/nn7bcAcO0Nf18RGAMMOeU0+vUfwNtvvcHoUSOrbatJkyYM3mrrcoExQOPGjTn/okto0aIFo0eNZNasWeW2//7IY7j4ssv51QEH5RTMLl68mPvvuxuAq665YUVgDLDFlltx3AknsXjxYm69OZlRRikdZpbILeE+tTSzS8zsCzNbaGbfmdldZpbbB03l9nqZ2e1mNtHMFpnZDDN718zOquKYI83sfTOba2Yzzew5M9uq5s+qMgXH9ZCZDU0LukvxNqnYr1FDsOVGvejQtiUTps5gzBffVdr++LCPAdhrmw2qbWuNVdqz6xZ9+fqbn3jxnfE592Hx4qXV7jNz1vyc2xMppBHvDmfWrFn0XqsPGw3cuNL2Xx1wEADPP/dMrc5jZjRu3Bgzo1mzZrVq6/PxnzF//nyaN2/O5oO3rLR92+13AOC5Z56q1XlEqmNmLYBhwAVAG+BJYCpwFDDazKrPRyrf3p7AOOBY4CfgMWAU0Av4U5ZjbgDuBvoBrwDvA7sCb5rZ/nk+payUViFSTw1YpzsAH33+bcbtH30eRnr7r9292ra2HdSHxo0bMeLjSTRu3Ij9d+jHlhv1onGjRoyb8D2PvDKGX+YsqHTcK+9/we/32YxzjtmlUlrF6b/bAYB7n/6gJk9PJHEfjx0LwMAMgTGwImAe9/HYGp/D3bn+2quYN28e2++wEy1btqxxWwDz580DoF379hlHATt16gzApEkTmT17Nu3atavV+aR0lOAMeecDg4F3gd3cfS6AmZ0OXAvcBeyQS0Nmth4hGJ4D7Oru76RtawRskuGYXYBTCIH0lu7+ZVy/JfA6cLeZve7uv9Ts6ZVRcFw/vQ/cUs0+vwfaxvuvAlUNB54I3AYsz7K9bWwv5T7CH3Q2P1XTN0lAj1U7AvDtj7Mybk+t79G9Y7Vtrd97VQDmzl/Mq3ecwBb9y//cO/T4PTjs3H/z5sgJ5dZfeMvzbNx3DY47eCv22Go9Ro//hubNmrDlgF4sXLyUoy58oNIxIsXyzdQpAKy2+hoZt68e10+N++Xqwr+cw48//sCc2bP55JOPmfj1BPqutz5/v+0ftesw0DmmUcyYPp0FCxZUCrYnT5q44v43U6ewwYb9an1OkYrMrBkwJD48MRUYA7j7dWZ2BLC9mQ1y9+rzkuA6oAVwUHpgHNtbDmS6Ovv0uLwsFRjH/d81s9uBk4FjCIF6rSg4LiB371Wgdp8DnqtqHzPbh7Lg+H53v6eK3YdUsY2YL50eHF/k7pOq7agUVOtW4efa+QsXZ9yeykFu26p5tW11aBs+cI/61ebMnb+IIy74Dy+9+zldOrTm3GN24bA9B/HglUcw6NBrypWH+2HmHHY/4TbuvfRwdh3cl16rdVqx7YnXPs6YpyxSLPPmhc/zlq1aZdzeqnVrAObMqeq7f2VPPvEYE78u+xLYr/8A/nnXffTq1buGPS3Tp8/adOvWne+/n8YD99/LMX88rtz2f8d8ZMi/31LiSmvgeGugPTDB3Udn2P4IMADYF6gyODazHsDuwNcxnqmWmbUEdko7V6bznxzPX+vgWDnHIkKjRuH/wk2bNOakKx7loZc+4pc5C/hq6gyOGfo/Phw3hQ5tW3LsQeWveei3dnfev/901u3ZlYPPvJtuO19An30u5czrnmS3Lfsy7J8nsk7PrplOKbLSGDPuC2YvWMbEqT/w2FPP0bRpU7bbajP+c/+9tW7bzDjtzD8DcMF5f+b+++5m5syZTJ48ibPPOJUXn39uxUWBjRrpI32lYQlekJdMkL1RXI7Ksj21fkAObe1AiD/fMbMmZnaImd1oZjeb2XFmlunnzr5Ac2C6u2cadcnn/NXSO0mknpo3P4wMt2qR+YKf1i3D+jkVJuGoqq058xby6KuV8y3veyb8wrXtJmXXWzRp3IgHLv8d3bu04//OuY9n3/qUWXMX8t302dzy4NtcfPuLdG7fmguP3S2/JyZSIK1bh+oUC+Znvkg0ld/btm3bjNur07lLF3bZdXeeeu5lVl21G6effCLfTJ1as86mOe6EIfzp+CHMnTuXE/70B3qt3pX+6/Xhjttu5oKhl9KhY4glUkuRAugZl9l+Dkytz6WeYOoq8bnAW8CDhFHfVIrnV2a2Yz7nd/d5wC9ARzOr2Rs4jYLjArIcSrmZWZv4TelZM5tiZvPNbImZzTKz8Wb2tJmdZ2ZKJJNypv7wMwCrr9I+4/bU+qnTfq62rSnf/xzb/CXj9snTwkQiXTuWlb7aov+arNOz64rJPyp67NUxQFm5N5FiW6NH+Hz97tvMn+/fxvU9evTMuD1X7du3Z4+992HBggW8NuzlWrUFYQTx6utu5O0RIzn3/Is48ug/cOafz+PNdz7g5FPP4OeZM2nZsmVOk5dI/WAkN3IcB477mNm4TLccu5T6n3+28kPz4jKXwDT1Le4PwHrAYUAnwujw/fH+41a+PFx158+3D1VSznERxSssHwYy1QdsF299gX2Av5pZU3evvnaWNAhjv5wGwMAMk3OE9eHioo+/mlZtW6ngtmPbzFfWd2oXcjTTp4JOBd+z0ma/S5dan61NkbrWf0D4xfWjjzKlTMKYuH7D/rX/ZbZz5y4AzJg+o9ZtpQzYaCADNhpYbt3wt99k2bJlDN5yq0o1l6V+K8FqFUlJDcw2Af7k7g/Fxz8DvzOzvsBmwAnAX4rQPwXHxRIT0l+k7BvOEuAD4CvCN6PWhFp/GxGCZJFy3h0ziV/mLKBPjy4MWGc1xn5ZvtbxATuFaVKfe/vTatsa8fFkZvwyj1U7t2Wdnl35csr0ctu3iekU6fWUv/8pXPyzbs+utGnVvFzgDDBo/R4ATM5h5FqkLgzecmvat2/PxK8nMHbMR5UCzScffxSAPffap9bnGv7WmwD0Xquwo7l33HozAEce/ceCnkfqvQm1nD46VZ0i89WsIWaBqitZVWxrLmGAsKK7CcHx9nmcP98+VElpFcVzOmWB8VtAL3ff2t2PcPfj3f337r4d0JmQvP4fQFOJyQpLli7j9oeHA3DDWQfQqkXTFdtOPnQ7BqyzGm+OnMDo8WUpD8cdvBUfPXgWl5ywZ7m2li1bzk0PvEmjRo244awDaNu6rMLFjputw+/23pTly5dz5+MjVqx/7+PJ/DBzDm1aNef6M/enWdPGK7Z179KOq07dD4DHh9W8ZqxIkpo1a8YfjzsRgDNOPYl58+at2Hbzjdfzycdj2Wbb7dl4k0Er1t9x2y0M2mgDhl5wXrm2Xnj+Wd57t1wFKgDmz5/PJRedz9tvvcGq3bqxy2571Lrf03/8kalTypeXW7p0KX+9dChPPP4o222/Iwcc9Otan0dKSUIpFWYkdEVe6g8wcx3EsvWTs2xPl9pnirtnimsmxeUquZ7fzFoDHYCf3b3WwbFGjotn27T7R7t75SnOgJhG8Ua8iZRzxd2vstPm67DlRr345JE/M/yjifTs3pHN+63JjzPn8qfLHiq3f+cOrenbaxW6fVL5x4jr73+d7Qf1Yect1mXsw3/mg08m07lDazbfsCdNmjTmwluf58NPyy4uWrR4KSdd/ij/ufx3/HbvTdlxs7UZ9dk3tGjelC36r0m71i0YNf4brrnvtUrneuPOsuqBa8byb+cevQt/PHAwAKPHf8upVz+eyGskku7sc/7C68Ne5b0R77Bxv75stfU2TJkyhQ8/eI8uXbtyyx3/Krf/Tz/N4MsvPuf778unJ40a+SFX/PUSVlttdfpvtBHt2rXnxx9+YOzYj/h55kzat2/Pvf/+X7kpqgE+Gj2K008p+/ufOiXECYf/5iCaNQtfSo846miOOOoPK/YZP/5T9tljFzYauDFrrtkLd+f990bw/ffT2Gjgxtz3QPn3uUgBjInLSpNzVFify2hIKq8p2xWkqZqgc9PWfQ4sArqa2eruXvFCl3zOXy0Fx8WTHp1Mz7rXSqyKCwH61GlH6rFFi5ey+wm3c9YRO/Gb3TZm3+37MXP2fO575gMuuePFrBOEZLJ02XL2P+1OTj50Ow7baxC7bNGXxUuX8tbor/n7f9/i+eGfVTrm6TfHse1RN3Hq4duzzcZrsftW67F4yTK+mjqDR18dw83/e4uFiyqnyW/er/IFzWut0Zm11gizfWU6RiQJLVq04NkXX+Xaq6/g4Qf/yzNPP0nHTp04/HdHcP6Fl7D6GtkGxsrb71cHMHfOHN4Z/jajRn644qK4tfqszdHHHMufjh9Ct+6VZ6ecM2c2H37wXqX1Y8d8tOL+LrvtXm5b7959OOy3v2fEu+/w8ksv0KhRI9ZZty8nn3oGxx5/Yq2nqJYSVVopx8OBWYQL+wa6+0cVth8cl0/n0NY7hMnCuplZX3f/vML2VDrFiosD3H2BmQ0D9gR+DdxQi/NXyzKPaEsSzGwSZWVNeqdPnGFmr1JW0Pocd7+ygOc+qppJQKprqxcwMW1V7yQmAakqOLaWnZs3739UbU8h0iD8+Eai//sQWWltvkm4FuOzT8eVVuiZxszGNe3Uc4M1jrgtkfa+ufd4lsyc8mktc44xs8sIF8i9Q5g+el5cn5o++g133yFt/yGEScYed/dzK7R1HvBXwgy+B7r77Lh+F8IkZ02Awe7+ftoxuwAvk3n66NeABYT45JfaPE/QyHExPURZcHyFme1KyCt+OUuB65VOtjdqDJo3yLRNREREiuIyYBdgK+BLM3uLMAi3BeEX8KMr7N+FUHGr8k8ocDWwY2zvCzMbEfcfDDQG/pIeGAO4+ytmdiNwCvCRmb0MNAN2JYyzH5VEYAy6IK+Y/gU8kfZ4Z+AuYKqZTTaz+83sGDPrUpTeiYiISHGU3gx5uPtCQkB7KaGq1v6E4PgeYBN3/zqPtpYAewF/BmYQppPuT7i+al93/1uW404FjgI+IwTFWwKvANu5+xP5P6vMNHJcJO6+zMwOJHzTOp3yI6U9gcPj7TYzuw84291n1n1PRUREpK6VYpljd18AXBhv1e07FBhaxfYlwFXxlk8f7iEE5AWjkeMi8uDOmF7QFzgWuBdI//bVFDgGeN/MuhahmyIiIiINhoLjEuHuX7j7P939SHfvQwiWrwOWxV36ABcVrYMiIiJSJwowfbTkQcFxiYrB8hmUD4j3K1Z/RERERBoCBcel76m0+6sWrRciIiJSZ8ySuUn+dEFekZhZF3efkcOuPdLu/1io/oiIiEjpMEW2RaOR4+KZYmZ3mNn2Zpbx38HMNgX+nrbq+brpmoiIiEjDpJHj4mlJqE5xLDDHzD4CJgPzCIWw1wPSJ8mYThUlUURERGTloYHj4lFwXDxzgTbxfltg23jLZAzwf+7+XV10TERERIrIoFGjhKJjBdl5U3BcPJ2B7YDtgc2AdQgX3LUgzDzzDTASeBR4yt2XF6mfIiIiIg2GguMCcvdeVWxbTJjy8JW6PncN2pqEvnuKiIjUiVDnOLm2JD+6IE9EREREJNLIsYiIiEhJsQRLuWnsOF8KjkVERERKjKpVFI/SKkREREREIo0ci4iIiJQYzZBXPAqORUREREqJJRgcK8bOm9IqREREREQijRyLiIiIlBhlVRSPRo5FRERERCKNHIuIiIiUkDBDXjJDxxqAzp+CYxEREZESo7SK4qlxcGxmy5LsSBbu7grgRURERKRO1CbwNMDRiL2IiIhIolTnuHhqOyqrfzkRERGRhCk2Lp4aB8furkoXIiIiIrJSUT6viIiISCnRDHlFpdFfEREREZFII8ciIiIiJUY5x8VT58GxmbUEOgJN3H1KXZ9fREREpJQZluAkIIqy81Xw4NjCv+7BwOHANoTAGEIZuCYV9u0CHBIffunuLxe6fyIiIiIiKQUNjs2sL/Ag0D+1qppDfgJOBfoA35tZD3dfXrgeioiIiJQepVUUT8EuyDOz9YF3CYGxxdu8eMvI3R24Le7bDdihUP0TERERKVVmlshN8leQ4NjMGgOPAR0Ige4wYEt3bwvcU83hD6Xd360Q/RMRERERyaRQaRWHA30JecWPA4fkmh7h7t+a2USgF7BpgfonIiIiUrI06Fs8hUqrOCAuFwLH1yBv+BPCiPM6ifZKRERERKQKhRo5HkQYNX7b3afX4PgZcdk5uS6JiIiI1AOaIa+oChUcd43LSTU8fklcapISERERaXCUVlE8hUqrWBSXTWt4fCq4/jmBvoiIiIhILZlZSzO7xMy+MLOFZvadmd1lZqvn2c4kM/MqbutlOOaeao45LqnnWaiR2R+AttQ8Z3hzQlrG1MR6JCIiIlIPGMmlVSQ1AG1mLQjVxwYD04AnCcUTjgL2MbPB7v51ns3em2X9rCqOeRH4PsP6z/M8d1aFCo7fJQTGm5nZqu7+Q64HmtmuwOqE4PjNAvVPREREpGSVYI3i8wmB8bvAbu4+F8DMTgeuBe4iz/kp3P3IGvTjCnd/vQbH5axQaRVPxGVT4PJcDzKztsBNaaseTrBPIiIiIpInM2sGDIkPT0wFxgDufh0wFtjezAYVo39JK0hw7O5PAGMIo/lHmNl18YXNysz6E0aKU/WRX3X39wrRPxEREZFSZpbMLSFbA+2BCe4+OsP2R+Jy38TOWESFrAZxBCHYbQucAvzGzB4DNkrtYGanEKaJ3hrYirLUmBnAMQXsm4iIiEjJKrG0ilTsNirL9tT6Afk0amZnAX0IhRzGAY/nUAL4QDM7CGgMTASedvfx+Zy3OgULjt19rJntR5gOehVCEHxCanNcXpd2SOqvYBrwK3fXxXgiIiIixdczLr/Jsj21fs08272qwuPrzewkd7+rimNOqvD4SjO7DTjF3Zfmef6MCpVzDIC7v0n4tnE3oXaxVXFbCtwDbOruHxayXyIiIiIlK6GUCktFWNDHzMZluuXYozZxOT/L9nlx2TbH9p4CDiQE062AfoQB0+bAv8zsVxmOGQ0cB6wbj1kLOBH4hTD4enWO565WwSfZiJUqjjGzPwO7AFsCqxFyV+YRyr69B7zo7t8Vuj8iIiIiUjzufnKFVeOAM8xsPPAP4EpCqbj0Y26scMxE4FYze4OQ1jHEzK5LIvOgzmagc/cZwP/iTUREREQysgRzjg3ChXQb1qKRVHWKVlm2t47LObU4B8CdwGVAXzPr5e6TqjvA3ceZ2VPAwcDOhCyEWiloWoWIiIiI5CdMApLQLZkuTYnLNbJsT62fXJuTuPtyYEJ82D2PQ7+swTFZKTgWERERkaqMictNsmxPrR+bwLk6xuW8Kveq/TFZ1VlaBYCZrQtsDHQlJG3PIZRtG+XuX9RlX0RERERKVaPSKuU2nDClcx8zG+juH1XYfnBcPl2bk5jZhoT5LuYDOZVnM7PmwN7xYbZSc3kpeHBsZu2B0wh1i1erYr9pwL+AG9z9l0L3S0RERKRUlVJs7O6Lzexm4C/ALWa2m7vPgxXTRw8A3nD3kaljzGwIYVa9x9393LT1ewEL3X1Y+jnMbADhujQD/uXui9O2rQdsBjzk7ovS1nclXMDXgzC6PTyJ51vQ4NjMdiUkRnej+rSX1YALgGPN7Ch3f7GQfRMRERGRnF1GqDq2FfClmb1FKMW2BTAdOLrC/l0Io8AV84A3By4ys8mEgHY+oSzbJoS49HXgnArHdAPuA240sw/j+VYDBhEyEb4BDnF3JwEFC47NbF/gYaApZYHxMuBzQmL3PMLVjT2A9QgznUB4AZ4ys1+7+1OF6p+IiIhIqSqxGfJw94VmtiNwLnAYsD8wkzAIeoG7Z5sgpKIXCbHfZpRNSz0beBv4D3C3uy+rcMwXwA3AYKA/0Jkwq94XhFSOG9395xo+tUoKEhyb2aqEF6tZXDUDuAS4P1PKREy9+C1h5LgrIaC+28w2iHWSRURERKSI3H0BcGG8VbfvUGBohvXvAu/med7vCCm6daJQ1SqGEK4cdOAToL+735wtl9jdZ7n7LYTZ9FKztXSI7YiIiIg0GGbQKKFbiQ1A1wuFCo73i8vlhByQnEZ/436/iceltyMiIiLSYJhZIjfJX6GC416EUePh7p5TKY4Ud/+MkHdisR0RERERkTpR6ElAalq7WDWPRUREpMFKaoY8yV+hguPUFYttanh86rhcr3wUERERWWlYQv9J/goVHL9ESIvYJt8DLSTIbENIy3gl4X6JiIiIiGRVqOD4dkL9udXN7JQ8jx0CrAEsju2IiIiINChJVauQ/BUkOHb3z4GTCaPH15jZmWZW5bksOB24Pq46NV6cJyIiIiJSJ2o8CYiZ9axmlxeBU4GrgCuB48zsAULh5ymE6QJbAT0JM54cCvQhjBifDTxvZj3dfUpN+ygiIiJS3xjJzZCnweP81WaGvEmEvOBcGGHe7L9Usw+EWfVuiDengFNci4iIiJQiVZooniQCz+r++ZyyILqqfdMDbf1JiIiIiEidq01wPIXcR45FREREJBdmNEpq6FhD0HmrcXDs7r0S7IeIiIiIRIppi6fQM+SJiIiIiNQbuthNREREpMQkVa1C8qeRYxERERGRSCPHIiIiIiUk1DlOri3JT50Hx2bWHmhLjqPWmgREREREGprEqlVI3goeHJvZmsBxwC5Af6BpHodrEhARERERqTMFDTzN7EzgMsoCYn0NEhEREamGAqbiKVhwbGZnAVemrZpLGAluG5dT4v2OlP0NOLAQ+LFQ/RIREREpdapWUTwFqVZhZj0II8YQguLfAB2A+1L7uHtvd+8S1+8NPEsIkpsCd8TtvQvRPxERERGRTApVyu1PhCDXgSHu/rC7L8+0o7vPcffn3X1f4NB4zF/N7MIC9U1ERESkZJlBo4RuGoDOX6GC4x3jcgbw71wPcvcHgdMJI8gXmNlGBeibiIiIiEhGhQqO+xBGgN9zd8+0g5lly3e+FZgW+3Z0YbonIiIiUrrMLJGb5K9QwXHHuJxWYf2itPutMh0Yg+m3CKPHOyXfNREREZHSZpbMTfJXqOB4cVxWHDWenXZ/jSqOnxuXqyfWIxERERGRahQqOE6VYmtfYf2ktPubVHH8WnHZMqkOiYiIiNQXSqsonkIFx58S0iLWrrB+dNr9QzMdaGbrAlsTRp2/K0jvREREREpYUtUqJH+FCo6Hx+WGZtY8tdLdPwa+IATOe5jZX8yscWq7mfUCHqBsRr3XCtQ/EREREZFKChUcvxSXzYEdKmy7PO3+JcCPZjbczEYDXwIbx21LgesL1D8RERGRkmQkk1JhZpgmos5bQYJjdx8FfEjIPd63wrZ7gXsIo8dGqGwxGBgANI7rlgMnufu4QvRPRERERCSTbLWGa83dN69i29FmNgI4A1gHVnytcWAEcIG7DytU30RERERKmcZ7i6dgwXF13P0fwD/MbA1gNcJo8UR3/6lYfRIREREpOoNGSVWaUJSdt6IFxynu/g3wTbH7ISIiIiJS9OBYRERERMpTieLiKVS1ChERERGpoVKcBMTMWprZJWb2hZktNLPvzOwuM8trRmMzm2RmXsVtvSzHNTaz08zsYzNbYGbTzewhM1s/mWcYaORYRERERKpkZi2AYYQKY9OAJ4FewFHAPmY22N2/zrPZe7Osn5Xh/I2Ah4EDgF+AZ4EuwMHA3ma2o7u/n+f5M6pxcGxmdyXRgWq4ux9TB+cRERERKQlGcmkVCY4dn08IjN8FdnP3uQBmdjpwLXAXlee2qJK7H5nH7kcTAuMvgW3d/Yd4/oOAR4D/mNn67r40nz5kUpuR4yMJpdcKTcGxiIiISJGYWTNgSHx4YiowBnD368zsCGB7Mxvk7iML1I3T4/LsVGAcz/+omT0F7Af8Cni0tieqbc6xFfgmIiIi0uA0MkvklpCtgfbABHcfnWH7I3G5b4ZttWZmvYH1gQWEdIqCnr82I8dHJdEBERERESmvxKpVbBSXo7JsT60fkE+jZnYW0AdYBIwDHnf36VWc/xN3X5LU+bOpcXAcp4EWERERkZVbz7jMNi9Fav2aebZ7VYXH15vZSe5e8bq2Qp0/I1WrEBERESklRnJl2EIzfcxsXKbN7r5hDq20icv5WbbPi8u2OfbqKeA1YCQwHViLcMHdKcC/zOwnd3+ygOevkoJjKUnr9V6VUcOvLnY3ROqFjpsNqX4nEWHRxGnF7kJOjOQmoiit7IzA3U+usGoccIaZjQf+AVxJKBVXFAqORURERFZuE3IcIc4mVZ2iVZbtreNyTi3OAXAncBnQ18x6ufukOj4/oBnyREREREpOic2QNyUu18iyPbV+cm1O4u7LgQnxYfe6Pn+KgmMRERERqcqYuNwky/bU+rEJnKtjXM5LW5c6fz8za1rg8ys4FhERESk1jSyZW0KGE6Z07mNmAzNsPzgun67NScxsQ6Av4cK78an17j4R+AxoCexdqPOnKDgWERERKTGlFBy7+2Lg5vjwFjNL5fimpo8eALyRPjuemQ0xs/Fmdnl6W2a2l5ntVPEcZjYAeJhwDeG/4jnTXReXV5nZKmnHHUiYHe8rErqITxfkiYiIiEh1LgN2AbYCvjSztwh1hbcglGM7usL+XQijwN0rrN8cuMjMJhPSJeYTSrltQohLXwfOyXD+u4C9gAOA8Wb2ajzH9oSZ837r7ktr9xQDjRyLiIiIlBAjmYvxzAxLqJibuy8EdgQuJQS0+xOC43uATdz96xybepEQ6M4mTEt9MLA28DbwR2AXd1+Q4fzLgV8DZwDfAfsA/YFHgU3d/b0aPrVKNHIsIiIiUmISzBdOTAxaL4y36vYdCgzNsP5d4N0ann8ZIb3iuur2rQ2NHIuIiIiIRBo5FhERESklBomVKC7BEehSp5FjEREREZGozkeOzawlocBzE3efUt3+IiIiIg1No+Rmt5M8FTw4tjB34cHA4cA2lM184hXPb2ZdgEPiwy/d/eVC909ERESklBjJ/bSvEDt/BQ2Ozawv8CCh1AZU/2/0E3Aq0Af43sx6xNIdIiIiIiIFV7CcYzNbn1Cqoz8hKDbCPNnzsh3j7g7cFvftBuxQqP6JiIiIlCqzZG6Sv4IEx2bWGHgM6EAIdIcBW7p7W0Kx6Ko8lHZ/t0L0T0RERKSUNTJL5Cb5K1RaxeGEKQMdeBw4JNf0CHf/1swmAr2ATQvUPxERERGRSgoVHB8QlwuB42uQN/wJ0BtYJ9FeiYiIiNQDGvQtnkLlHA8ijBq/7e7Ta3D8jLjsnFyXRERERESqVqiR465xOamGxy+JS83gJyIiIg2KAY0SGjnWAHT+ChV8LgKaAU1reHwquP45me6IiIiI1BOW4CQgio7zVqi0ih/isqY5w5sT0jKmJtMdEREREZHqFSo4fpfwXWUzM1s1nwPNbFdg9fjwzaQ7JiIiIlLqVOe4eAoVHD8Rl02By3M9yMzaAjelrXo4wT6JiIiIlLxUznESN8XH+StIcOzuTwBjCP8mR5jZdWbWrKpjzKw/YaQ4VR/5VXd/rxD9ExERERHJpJDVII4gBLttgVOA35jZY8BGqR3M7BTCNNFbA1tR9gVnBnBMAfsmIiIiUrJMY75FU7Dg2N3Hmtl+hOmgVyEEwSekNsfldWmHpP4KpgG/cnddjCciIiIidapQOccAuPubhJHiuwm1i62K21LgHmBTd/+wkP0SERERKWVJ5RxL/go+yYa7/wAcY2Z/BnYBtgRWA9oD8whl394DXnT37wrdHxEREZFSp8C2eOpsBjp3nwH8L95EREREREqOpmcWERERKSGGYQkVKdaFfflTcCwiIiJSSpLMF1ZsnLeCXpAnIiIiIlKfFGTk2Mx+n1Rb7n5fUm2JiIiI1Aea+rl4CpVWcQ9ltYxrwwEFxyIiIiJSJwqZc5zvdx6vwTEiIiIiK51GGjoumkIFx/fmuF8jQr3j/kDvuG4h8DCwvAD9EhERESlpRnIX5CnEzl9BgmN3PyrfY8xsU+BGwiQh3YBfu/vspPsmIiIiIpJNyVSriFNGbwe8RJhJT7nGIiIi0iCZJXOT/JVMcAzg7suAPwBLgH3N7MAid0lERESkzjXCErlJ/koqOAZw92+A4YQ0mbzTM0REREREaqrkguNoQlxuVNReiIiIiBSB0iqKp1Snj24Rl6sUtRciIiIidcwSnD5aAXL+Sm7k2MwaES7MA5hVzL6IiIiISMNSiiPHlwE9CZOCjCpyX0RERETqnCYBKZ6CBMdm1jPPPnQGBgJHEOocp+Q6mYiIiIiIFJCZtQTOBf6PMJA5E3gBuMDdv61Fu+sAYwlpta+6+y4Z9rmHECdmc7y7317TPqQr1MjxJMLIb2085+7/S6AvIiIiIvVKqQ0cm1kLYBgwGJgGPAn0IlQW28fMBrv71zVs/h9A8xz3fRH4PsP6z2t47koKnVZRk3/a5cBtwBkJ90VERESk5IXpo5OJjhOMsc8nBMbvAru5+1wAMzsduBa4C9gh7/6ZHROP+wdwbA6HXOHur+d7nnwUKjieQu4jx0uA2YTR5veAB919SoH6JSIiIiJ5MLNmwJD48MRUYAzg7teZ2RHA9mY2yN1H5tHuqsDVwMvAf8ktOC64ggTH7t6rEO2KiIiINAQlllaxNdAemODuozNsfwQYAOwL5BwcAzcCLYETgDVq28mklGK1ChEREZEGrcRq7aYmZctWRSy1fkCuDZrZXsBvgAvd/SszyzU4PtDMDgIaAxOBp919fK7nzUWhqlWkvzjj3H1ZIc4jIiIiIgWXqkL2TZbtqfVr5tKYmbUGbiVcRHdlnn05qcLjK83sNuAUd1+aZ1sZFWrk+CNCzvFkd1+rQOcQERERWSlZsnkVfcxsXKYN7r5hDse3icv5WbbPi8u2OfbnMkIgvaO7L87xmNGEiwGHEYLxbsCesa0TgMXAaTm2VaVCjdovicsRBWpfREREROoZM9sUOBm4L5+qE+5+o7vf4e5fuvsCd5/o7rcC2xIC4yFm1iOJPhZq5Ph7QmL13Op2FBEREZHyEr4eb0KOI8TZpOK5Vlm2t47LOVU1YmZNgH8CvwBn1qI/K7j7ODN7CjgY2Bm4p7ZtFio4Hg/0IMfcExEREREJzCy5OsfJtJMqsZvtornU+snVtLMGYUbk74GHK/StQ1wOMrPXAdx9hxz792Vcds9x/yoVKjh+CNgV2MbMOrv7TwU6j4iIiIgU1pi43CTL9tT6sTm21y3eMukAbJ9jOykd43JelXvlqFA5x/8BPiXMkX1Lgc4hIiIislKyhG4JGQ7MIlzYNzDD9oPj8umqGnH3Se5umW7AjnG3V9PWVcvMmgN7x4fZSs3lpSDBsbsvJLxQU4Ffm9lzZrZuIc4lIiIisrIxS+aWhFhR4ub48JZYii32004n1Dd+I312PDMbYmbjzezy2p7fzNYzs9/FQDh9fVfgf4RU3jGEIL7WapVWYWYXxrvvu/sLGdY/BRwH7A58ZmZjCTOnTAcW5HIOd7+kNn0UERERkVq7DNgF2Ar40szeIlxbtgUhrju6wv5dgL4kkwfcDbgPuNHMPoznWw0YRCgf9w1wiLt7Aueqdc7xUEI941uAFzKsT2eEbxY5z54SKTgWERGRBiXhOse15u4LzWxH4FzgMGB/YCahOsQF7p5tgpAkfAHcAAwG+gOdgUVx/dPAje7+c1InK+T00Zn+VfP9l07kG4CIiIiI1I67LwAujLfq9h1KGCzNte3XyRInuvt3JDTBRy4KFRxfXKB2RURERFZqRnIXhZXW+HP9UJDg2N0VHIuIiIjUUKmlVTQkhSrlJiIiIiJS7xQy51hEREREakDjxsWj4FhERESkxCitoniUViEiIiIiEiU1cnyAmfVLqK107u47F6BdERERkZKl0cviSSo4Xi3ekmSozrGIiIiI1KGkgmMlxoiIiIgkwEgu51gBWv6SCo7fB55PqC0RERGRBk1BbfEkFhxr4g8RERERqe9Uyk1ERESklBgkVslNQ9B5U3AsIiIiUmIaKaotGlUKERERERGJNHIsIiIiUmI0QV7xaORYRERERCTSyLGIiIhICbH4X1JtSX6SCI71qouIiIgkSGkVxVPb4Lh3XM6ubUdERERERIqtVsGxu09OqiMiIiIiEqiUW/Eo51hERESkxCitonhUrUJEREREJNLIsYiIiEiJ0chx8WjkWEREREQkUnAsUs8tWLCAS4ZeSP8N1qVDmxb07rkaf/rD0Xz77bc5t7F06VIuu2QoB+y3N+uvuxZdO7alQ5sW9Ft/HU4ZcgKTJ2e/9nbZsmXcdMP1bDqwPx3btqRH964cfughjP/ss6zHzJ8/nyv+dhmbbLQhHdu2ZPVVO/OrffbkzTdez+OZi9RMi+ZNueD4vRn7xIX8POJ6vn7pr9x+0eGs1rV9jdrr2b0TN/3l//jsmaH88t71TB12Ba/fewan/X7njPt3bNeKv526P588eRG/vHc9k1+9nAeuPoYB666ecf9GjYyDdt2Yy087gJfvPJUZ71zLgtE384+Lf1uj/krpM8pqHdf+P8mXuXux+yBSjpmNW3+DDTYYNWZcsbtS8hYuXMjuu+zI+++NoFv37my9zbZMnjSJDz94n65du/LG2yPovdZa1bYzd+5cunZsS5s2bejXfwDdV1uNJYsXM2bMR0ydMoV27drx3IuvMmjTTcsdt3z5cg79zcE89cTjdOjQgR122pmfZszg7bfepGXLlrzw8mtstvnmlc61+y47Mmrkh3Tq1InBW27FrFmzeP+9ESxdupTb7vgXRxx1dKKv08qu42ZDit2FeqN5sya8+M9T2GJAb6ZNn8XwUV+x5mqd2ax/L36cOYftf38Nk779Kef2dtt6Ax64+hhaNm/K6PHfMGHKj3Rq35oN116N+QsW0+9XF5fbv1uXdrx612ms1aMr06bP4sNPJrFq53Zs2m9NlixdxkGn3MGrI8aXO6Z9m5Z8/9bVlc7976dGcOxF99fshWigFo1/AIDlC2aWbMxoZuPW7NN3gzufeTuR9o7ZZxsmT/j8U3ffMJEGGwDlHIvUY1f87TLef28EWwzekmeef4k2bdoAcOP113HO2Wfwpz8ezUuvvl5tOy1atODV199m8y22oEmTsv8tLFu2jKEXns81V13ByScex/D3Pix33L1338VTTzzO2uuswyuvvcWqq64KwOOPPcphvzmYo444nI8+/qxcmxf85VxGjfyQTTYZxBPPPE/Xrl0BePedd9hv7905ecjx7LDTzqy55pq1fXlEKjnnD3uwxYDejBjzNfscfzPzFiwG4OTf7sSVZxzIHUN/y+5/vDGnttbttSr/u+YPzJm/iH2Ov5kRYyau2GZmbLx+j0rH3Hz+oazVoysvvD2Ow8+6k/kLw/n33WEA/73mD9z9tyPYYJ+hzJ2/aMUxS5Yu4z/PvMeoT6cwctwU1llzFf55ye9q8zKISBWUViFSTy1evJjbb70ZgBtuumVFYAxwymmn07//AN568w1GjRxZbVtNmjRhq623LhfEAjRu3JiLLr6UFi1aMGrUSGbNmlVu+003XgfAXy+/akVgDHDAgQexz777MeGrr3j6qSfL9fm+e+4C4Jrrb1oRGANsudVWnDDkZBYvXszNN92Q46sgkrumTRpz3P9tB8Cplz+0IjAGuOn+YYz94hu223SdjEFtJleecSAtWzTj2Iv+XS4wBnB3Rn06pdy6NVbtwN7b92fJkmWc8rf/rQiMAZ5+fSwPvziSrh3bcsSvtix33PyFi/nDBf/m1v++wXtjJ7Jo8dK8nrfUT0qqKB4Fx/WUmQ01My/h26Riv0Yru3ffGc6sWbNYq08fBm68caXtBxx0MADPPft0rc5jZjRu3Bgzo1mzZivWT5o4kfGffUbLli3Zc6+9K5//wHj+Z8rOP/6zz5g/fz7Nmzdn8JZbVjpm+x12BOCZp5+stE2ktrYcuBYd2rZiwpTpjPn8m0rbH3/lIwD22q5/tW2tsWoHdt1yfb6eOp0X3/40p/MPjEH3pO9+Ysq0nyttf+PDLwDYZ4fqzy8rOQvVKpK4KT7On9IqSoyZ7QC8Fh++4e47FK0zUtLGjhkDwMCNN8m4PbX+44/H1vgc7s61V1/JvHnz2GHHnWjZsmXZ+ceG82+wYT+aNm2a9fyfpJ1/3rx5ALRv3x7LUKeoU6fOQAi8Z8+eTbt27Wrcd5GKUhe8fTR+asbtH30W1vdfZ7Vq29p203Vo3LgRI8ZMpHHjRuy/00ZsOXAtGjdqxLgJ03jkxZH8MmdBuWNat2gOwC+z52dsc+Yv4f3Rf901cntCIlIQCo7rr/eBW6rZ5/dA23j/VWB8FfueCNwGLM+yvW1sL+U+YE4V7eV+RYvUyNSp4Sfb1VfP/EGaWj+1ikoTmfzl3D/z4w8/MHvObD75eCxfT5jAeuuvz213/Kv8+afE86+R5fxx/ZQpZedPpVFMnz6dBQsWlAu2ASZNKvtpeuqUKWzYr19efRepSo/unQD49sdfMm5PrU/tV5X11+oOwNwFi3j1rtPYYkDvctuHnrgvh531L9788MsV66b/PLfK9nut3gWAzh1a07pls3JpH9LwKCWieBQc11Pu/hzwXFX7mNk+lAXH97v7PVXsXuXl7mbWi/LB8UXuPqnajkrBzJsbPmhbtWqVcXvr1q0BmDO3qu8wlT3x+KN8PWHCisf9+w/grnvvp1fv8h/+8+bF87es5vxzys7fZ+216da9O99Pm8b9993LH/90XLljUvnIFY8TSULrlmHkdn6WoHPegnARXNvWLaptq0O78Hd/1P5bMXfBIo44925eGv4pXTq25dw/7sFh+2zOg9f9kUEH/ZXvpodc/Q/HTWLhoiV069KOXbdan5ffKV/u8Lf7brHiftvWLRQcixSJco5FpJxx479iwRJn6rTpPPXsCzRt2pStthjE/ffdW+u2zYwzzzoHgPPOOYv77rmbmTNnMnnSJM447RSef+7ZFRcFNmqk/z1J6WoU04KaNm3MSZf9l4deCGkUX035kWMuuI8PP5lEh7atOPaQ7VYcM3vuQv7x8FsA/POS37HfjgNo16YF66y5Cv++4ijW691txb7Ll6vMakNmQCNL5qbx5/zp00eknmodq1PMn585fzGV39u2TduM26vTpUsXdt1td5576VVW7daNk4ccz9SpZbmarVvH8y+o5vxty5//hCEncfyJJzF37lz+9MejWX3Vzqy3Tm9uu+XvDL3kMjp27AiwYimSlNTIcKuWzTJuT40sz5m3MOe25sxbyKMvj660/b6nRgCw7aC1y62/4KaneOzlUazauR0PXncsP7x1DWOfuJB9dxzAmVc/smK/ivnK0vCoWkXxNKi0CjNb8VXc3S2u6wscD+wO9CDk3E4kpCxc6+4z8mh/Q+BIYJfYVltC7u0XwPPAv9w9Yy6umQ0FLqqwevv0PqeZ7O69cu2XrJx69OgJwLffVr7qPn19j1rWC27fvj17770vd9x+K8NeeXnFBB09esbzf5Pl/HF9z57lz29mXHfDTRxx5NE88/STfPftt3RdZRX2P+Ag1t9gAy69+CJatmyZ0+QlIvmYOm0mAKuv0iHj9tT61H5VmfJd2Gfq95WrTgBMjtu7dir/5XDxkqUcfvZdbL3xG+y61QZ06diGb374mYdfHElqUq6vpvzI4iUq1yZSLA0qOK7IzI4DbgCaV9g0IN7+aGZ7uPuHFY+t0E4T4DrgBKBxhc3d42174BwzO9Xda//7tDR4AzbaCICPRo/KuD21vn//AbU+V+cu4UKh6TOml51/QDj/p+M+YcmSJZUqVqTO3y/L+TcaOJCNBg4st+7tt95k2bJlbLlV5ZrLIrU19oswpfrA9TLXMU6VWvv4y++qbeujz8OvKB3bZc657xTXp0/mkW746AkMHz2h3LrD9gmzSaZfxCcNV4aCPlJHGmxahZkdSajO0Bz4HPgP8C/gbSA1WtsZeMrM2lfRTiPgUeAkygLjmcBjsb0XgdRVFR2Ae8zslAxNpapPPJG27ru4ruJNwbWw5VZb0759e76eMIExH31Uafvjj4afaPfae99an+utN98AYK21+qxY16t3b9Zbf30WLFjA8889W/n8j8Xz75P7+W+95e8AHP2HY2vTXZGM3v3oa36ZM58+PbuuKOuW7oBdBgLw3JsfV9vWiDETmfHzXFbt3JZ11lyl0vZtBq0DwJjPM5eNy+S4mJ9892Pv5HyMrLwsoVuifTJraWaXmNkXZrbQzL4zs7vMrPIbKr921zGzBXGehFeq2K+xmZ1mZh/H/aeb2UNmtn5tzl9Rgw2OgduB6cCe7r6eu//W3f/o7tsCOwCz437dgUzBbMqZwH5pj68Aurv7QbG9PYA1gZfS9rnGzLZIb8Tdn3P3IUD6vKVfuvuQDLeK6RfSADVr1ozjTghFRk49+cQVOb4Qpo/++OOxbLvd9mwyaNCK9bfdcjMb9VuPC/5ybrm2nn/uWd59p/IH8vz587nogr/w1ptv0K1bN3bbfY9y208+5XQA/nLu2fz4448r1j/x+GM88/RT9Fl7bfbd71fljvnxxx+ZMqX8zGFLly7l0osv4vFHH2H7HXbkoIN/nc9LIZKTJUuXcfv/3gTghnMPoVWLstzjk3+7EwPWXYM3P/yS0Z+VBbTH/WY7PnrsfC45ab9ybS1btpyb7h9Go0aNuOHcQ8pVuNhxi778br8tWL58OXc+MrzccT26daRrxzbl1rVo3pRbLjiUzfr34r4nR/DhuPzKL4rUBTNrAQwDLgDaAE8CU4GjgNFmVptcuH9Q+Vf8iudvBDxM+KV+DeBZYBxwMPChmW1ei/OX09B/t9zF3SvNkODub5rZecDNcdWhwCUV9zOzdoQ/kpRr3P3civu5+/dmth/wFrAZ4XW/HNip9k9BGrJzzjufYa++woh336Hf+uuw9TbbMmXyZD54/z26du3KHf+8q9z+P/00gy8+/5zvp00rt37khx/w10svZrXVV2ejjQbSrl17fvjhe8aO+YiZM2fSvn17/v3fh8pNUQ1wxFFH88ILz/HUE48zsN967LDTzvw0YwZvvfkGLVu25K577q+UHjH+s0/ZY9edGDhwY9bs3Rt3570R7/L9tGkMHLgxDzz4CCKFcsW/XmCnLfqy5cA+fPLURQwf9RU9u3di8wG9+XHmHP409P5y+3fu0Ia+vbvR7eNJldq6/r5X2H6zddl58HqMfeJCPvh4Ip07tGHz/r1o0qQxF/79qUqB7vabrcutFxzGqM+mMHXaTFq0aMaWG61F5w6teWn4p5z8t/9l7PcN5x7CxjEdpFOHUCZxj2035I17zyhr+4hra/PSSAkxbEVFlCTaSsj5wGDgXWA3d58LYGanA9cCdxEGF/Prn9kx8bh/AFX9bHg0cADwJbCtu/8Qjz8IeAT4j5mt7+61TthvyMHxPzIFxmnuI+QjNwH6mlk7d59dYZ/DCN+eAH4ALszWmLsvMrMhwHtx1Y5m1tfdP69R71cCZjYuy6Y+WdZLBS1atODFV17j6isv58H/PcDTTz5Bp06d+N3vj+TCiy9ljSwTdFT0q/0PZM6cOQx/+y1GfvgBM2fOpGXLlvTpszbH/PFPHH/iSXTv3r3ScY0aNeKB/z3MzTfdyH333sXzzz5D69at2f/Ag7jgwotZf4MNKh3Te60+/PZ3R/Duu8N56YXnadSoEeuu25dTTzuT408cUm6KapGkLVq8lN2PvYmzjt6N3+yxKfvuOICZs+Zz35MjuOTWZ7JOEJLJ0qXL2f+kWzn58J04bJ/N2WXL9Vm8ZBlvjfyKv//nNZ5/65NKx4z+bCqPvzqazfv3YsC6q7NoyVLGffkd9z01gvueHJH1XOuv1Z3NK0w00rVjW7p2rFk1GpF8mFkzyuZDODEVGAO4+3VmdgShiMAgdx+ZR7urAlcDLwP/perg+PS4PDsVGMfzP2pmTxF+xf8VIdW1Vix1dWxDUKHyw3bu/lY1+48H+saHA9z94wrbHwJSv//e4O6n5dCHsUD/+PB4d7+9wvYdSGj6aDObREjpADiqmklAqmurF6GKR0rv2k4CUlVwvP4GGzQfNSbbZhFJ13GzKufwEZFo0fgHAFi+YGbJXu5mZuN6r73eBg88/24i7R2255ZM/Gr8p+6+YS36tCMhpWKCu6+dYfsFhF/YL3b3oXm0+z9CQNufkCrxGvCqu+9SYb/ewNfAAqC9uy+psP13hEHNe939yNyfWWYNeeS4+isuyk+B3C7D9o3T7ud6BcVwyoLjTXI8ZqWU7Y0ag+bKQ44iIiINRWmF7xvFZebySGXrcy6PZGZ7Ab8BLnT3r8ysqp86U+f/pGJgXNPzV6UhX5A3K4d90v8BmmbY3jXtfq5XUExKu98lx2NEREREiqVnXGYubF+2PqfC+mbWGriVUC3syro+f3Ua7MixJ5NPkn510ryse5WXvp+SxURERKSShGe365MtlTHHdItUvJN5StSy2CbXuOYyQiC7o7svrm7nApy/Sg02OE7IXCBVA7l1jsek7zcn2e6IiIhIvWcJTgJSWukZmNmmwMnAfe7+epG7k5GC49qZTllw3JMwkUd1eqXdz3lqahEREZEamlCbC/IIg4EAmaeELBv4q3LQL84o/E/gF8I8EXV6/lwpOK6d0UDqqs2tCHX2qrNV2v1Mie0Np3yIiIiIZFRiA76pmZuyXTSXWl/d9VdrAAOB74GHrfzweIe4HGRmrwOkVexK6vw5UXBcO8MoK+X2f2Z2nrsvzLZz/Ckh/UrK1zLsln58posARUREROrSmLjMVmUrtb6q+SPSdYu3TDoA22c5fz8za5qhYkW+569SQ65WkYQHKBvq7w5kndY5FtD+e9qq17JMAJJePq5Wc5WLiIhIPWUJ3ZIxnFDlq4+ZDcyw/eC4fLqqRtx9krtbphuwY9zt1bR1qeMmAp8BLYG9a3r+XCk4roU4Y96laavOMbNLYyC8QpwB5knCtIsAS4FK00xHEym7GnNNM9sswS6LiIhIPWAJ/ZeEWFHi5vjwlliKLfQzTB89gDBx2ci09UPMbLyZXZ5IJ+C6uLzKzFZJO8+BhNnxviLEWrWmtIrauwbYBtg3Pj4fON7MXgN+BnoQvg01TzvmLHd/jwzcfZmZPUGYmhrgdTN7gZBvsyyum+nuf0v0WYiIiIhkdxmwC+HaqS/N7C1CObYtCAUKjq6wfxfCLMPdEzr/XcBewAHAeDN7NZ5je8LMeb9196VJnEjBcS25+/L4reV64HigMdCZsiH+dLOAU3OYxvk8YCdCPk4r4MAK2ycDCo5FRERWQkZypdySyqxw94VxGulzCQN4+wMzgXuAC9w92wQdSZ1/uZn9GjiFEIjvQ6hv/Chwkbt/mtS5FBwnIH5TOcnMbif8g+1MGDFuS/jD+QJ4Dvinu/+UtaGy9iab2UbAEGA3YN3Ylv69REREGoASq1YBgLsvAC6Mt+r2HQoMzaPt16nmabv7MkJ6xXVV7VdbDSrYSk/uznH/HfLcfxxwRj7HVNHWj+T4B1hFG72S6EtsaxKl+V4VERERSUyDCo5FRERE6gUNRxWNqlWIiIiIiEQaORYREREpMUmVYZP8KTgWERERKTFJVauQ/CmtQkREREQk0sixiIiISInRwHHxKDgWERERKTWKjotGaRUiIiIiIpFGjkVERERKjKpVFI9GjkVEREREIo0ci4iIiJQQs+RKuakkXP4UHIuIiIiUGMW0xaO0ChERERGRSCPHIiIiIqVGQ8dFo+BYREREpMSoWkXxKK1CRERERCTSyLGIiIhIiVGVieJRcCwiIiJSYhQbF4/SKkREREREIo0ci4iIiJQaDR0XjUaORUREREQijRyLiIiIlBRLsJSbhqDzpeBYREREpIQYyVWrUGicP6VViIiIiIhEGjkWERERKTEa8S0eBcciIiIipUbRcdEorUJEREREJNLIsYiIiEiJSa5aheRLI8ciIiIiIpFGjkVERERKiSVXyk0D0PlTcCwiIiJSYhTTFo/SKkREREREIo0ci4iIiJQaDR0XjYJjERERkRJiJFetQjF2/pRWISIiIiISaeRYREREpMQkVq1C8qaRYxERERGplpm1NLNLzOwLM1toZt+Z2V1mtnoebTQxs6Fm9qyZfW1mc2JbX5rZrWa2Zpbj7jEzr+J2XFLPUyPHIiIiIiWm1AaOzawFMAwYDEwDngR6AUcB+5jZYHf/OoemWgAXAXOBscBIoBkwEDgeONzMdnb3D7Mc/yLwfYb1n+f8ZKqh4FhERESk1JRadAznEwLjd4Hd3H0ugJmdDlwL3AXskEM7C4FtgPfcfWlqpZk1Bi4DzgFuBzbNcvwV7v56zZ5CbpRWISIiIiJZmVkzYEh8eGIqMAZw9+sII8Dbm9mg6tpy96XuPjw9MI7rlwEXEILnQWbWPrEnkCcFxyIiIiIlxhL6LyFbA+2BCe4+OsP2R+Jy31qex4Flcbm4lm3VmNIqREREREpMiVWr2CguR2XZnlo/oKYnMDMD/gy0Boa5+4Isux5oZgcBjYGJwNPuPr6m581EwbGIiIiIVKVnXH6TZXtqfcZKE9mY2ZXAqkA7QmDdB/gM+EMVh51U4fGVZnYbcErFVI2aUnAsIiIiUmISHjjuY2bjMm1w9w1zOL5NXM7Psn1eXLbNs18HEQLilLHAb919YoZ9RxMuBhxGCMa7AXsSLuI7gZCGcVqe589IOcciIiIiUufcfW13N6ArsAewBBhpZkdk2PdGd7/D3b909wXuPtHdbwW2JQTGQ8ysRxL90sixiIiISAkxkss5js1MyHGEOJtUdYpWWba3jss5NWnc3WcAL5rZCOBj4DYzG+buU3M4dpyZPQUcDOwM3FOTPqTTyLGIiIhIybGEbomYEpdrZNmeWj+5Nidx91nA00BLYNc8Dv0yLrvX5vwpCo5FREREpCpj4nKTLNtT68cmcK4Zcdk1j2M6xuW8KvfKkYJjERERkVJiIa0iiVtCg8fDgVmEC/sGZth+cFw+ncC5to/LCbnsbGbNgb3jw2yl5vKi4FhERESkxJRSUoW7LwZujg9vMbNUjnFq+ugBwBvuPjJt/RAzG29ml5d7XmZ7m9lWFc9hZq3M7K+E4Ph74IW0beuZ2e9iIJx+TFfgf0APwuj28Fo+VUAX5ImIiIhI9S4DdgG2Ar40s7cIdY23AKYDR1fYvwvQl8p5wJsBF5nZt8BHhBHpbsBAoFN8fEj6FNVx+33AjWb2YTzfasAgQvm4b+IxnsQTVXAsIiIiUmJKbIY83H2hme0InAscBuwPzCRUh7jA3bNNEFLRY4SAdltCoNwJWAB8BdwB/N3dp1U45gvgBmAw0B/oDCyK658GbnT3n2v41CpRcCwiIiIi1YpTOl8Yb9XtOxQYmmH9WOCMPM/7HQlN8JELBcciIiIiJcaSniNPcqbgWERERKTUKDYuGlWrEBERERGJNHIsIiIiUmI0cFw8Co5FRERESoiRXLUKBdn5U1qFiIiIiEikkWMRERGRkmIJVqvQ2HG+NHIsIiIiIhJp5FhERESk1GjAt2gUHIuIiIiUGMXGxaO0ChERERGRSCPHIiIiIiUmqVJukj8FxyIiIiIlJrlqFZIvpVWIiIiIiEQaORYREREpJZZgWoUGoPOmkWMRERERkUjBsYiIiIhIpLQKERERkRJiJJdWoayK/GnkWEREREQk0sixiIiISIlRKbfiUXAsIiIiUmI0CUjxKK1CRERERCTSyLGIiIhIidHAcfFo5FhEREREJNLIsYiIiEip0dBx0Sg4FhERESkxqlZRPEqrEBERERGJNHIsIiIiUmJUyq14FByLiIiIlBAjuZRjxdj5U1qFiIiIiEikkWMRERGRUqMh36LRyLGIiIiISKSRYxEREZESo1JuxaPgWERERKSUWILVKhRj583cvdh9ECnHzGY3a968bZ8+fYrdFZF6YfzX04rdBZF6wRfNAmuML1tcsiGjmY1r1qzZBn3WXjuR9iZ89RWLFy/+1N03TKTBBkDBsZQcM/seaAVMLXZfpJzUt5UJRe2FSP2g90tp6gHMd/duxe5INmb2FGV/P0mZ4O77JdzmSkvBsYjkxMzGAWj0QaR6er+I1F+qViEiIiIiEik4FhERERGJFByLiIiIiEQKjkVEREREIgXHIiIiIiKRqlWIiIiIiEQaORYRERERiRQci4iIiIhECo5FRERERCIFxyIiIiIikYJjEREREZFIwbGIiIiISKTgWEREREQkUnAsIiIiIhIpOBYRERERiRQci0jizMyqeiwiNWdm6xa7DyIrMwXHIpIoMzN3dzPbx8wuAnDNUy+SCDPbFxhvZlcWuy8iK6smxe6AiKw80gLjfYEn47of3P32IndNpN5Lf18Be5vZk+7+TjH7JLIy0sixiCQiU2AMjAfampm+iIvUQoX31WfAv4GPitYhkZWYgmMRqbUsgfFnwD3ALe6+tGidE6nnMnzhvBf4u7vPVz6/SPIUHItIrVQxYnwvcLM+wEVqrrr3lfL5RZKn4FhEaqyKwPge9AEuUivVjRjrfSVSGAqORaRG0gLj/dDIlkiizOxXlE9RUmAsUkd0kYyI5C0tMP4V8HhcPQG4E7g9foA3cvflxeulSP1U4X01F3geuNbdl+p9JVJ4GjkWkbxUGDFOfYAvA1oDw9x9npk11ge4SP5iKkXqfbUEaAPsAawBoPeVSOEpOBaRnFXIMX4irl4KNAa6AXeZ2QbuvqxYfRSpryrkGH8PNI33NwCeNrP1i9IxkQZGwbGI5KSKi++Gpe22EfBfM+tb5x0UqccyvK9uAC5O22VD4CEFyCKFp+BYRKpVTVWKg4A/p+3en/AhrgBZJAdVlGu7GDg/bVcFyCJ1QMGxiFSpmjrGt7j7POA64IK0wxQgi+SgmjKIjdz9b8Bf0g5RgCxSYAqORSSrakaM/x4vvmvk7svc/a8oQBbJWRVlEFPvq+UA7n45CpBF6oypVKKIZFKhKsUTcXW1ExGY2V+AS9NWfQwc4u6f10W/ReqDCl84xwD/o5o6xmZ2LvDXtFXjCO+tzwreYZEGRCPHIlJJlqoUs4BbqOYDXCPIIlWrEBgDfALcVN0EHxpBFqkbGjkWkXKypFJ8CywANnP3WblMRKARZJHKMowYDwOucvcfcp35TiPIIoWlkWMRWSFLYPw5cDkwKNfAGDSCLFJRhtz9B4G/xMC4Sa5TQmsEWaSwFByLCFDtxXf3uPucfKeuVYAsEmR5X93k7gvje29pPu0pQBYpHAXHIpJLVYr5+QbGKQqQpaGroo5xlTnG1VGALFIYCo5FGrgKVSlSH+CfUfkDPO/AOEUBsjRUZvYrygfGdxLqg9cqME5RgCySPAXHIg1YlnJtk4H/ANemjRjX+spdBcjS0MT31ePx4WxgBHCnu881s8ZJvK9AAbJI0hQcizRQWQLjpcBy4NNUDmRtRowrqiJAXjepc4iUggplEBcD7YB1gY3NrIW7L0vyfFkC5P+YWb8kzyPSECg4FmmAstQxXgw0AXoDJ5vZXmbWNOlzZwmQXzGzdZI+l0gxVMgx/hFIvY+2Ai4GdizQeys9QF4ODATuMLNmSZ9LZGXWpNgdEJG6leXiu0+BGcC2gAE7EL48LzezV919SZJ9cPe/mhmU1UFeg/BhLlKvZXhf3Un4+z6F8J7aBjg37vtKAd5bl5vZMuAKYC5wnLsvTvIcIis7BcciDUgVVSnucvfrzOyvwDmEAHm7tOMKFSC3As4ENnH3CUm2L1LXMryv/gPc6O7LzWwpcAZ1EyBfZWZLgJfd/ZMk2xZpCDRDnkgDUU25ttvdfXbc7yrCh7jFfd4kjEIlHiDH83Vx9xlJtytSl8xsH+Cp+DD1vrrZ3een7XM5cBZlKY1vEybYSTxAFpGa08ixSAMRA+P9gcfiqnIf4GbWzN0Xu/vZMeUhFSAXegRZgbHUa2a2J2WBcaUyiEAjd1/m7ufG91YqQC7oCLKI1IyCY5EGIl61/mB8+ClwH+XrGC+O5aWW1XWALFJfmVlrYN/4cBLhC2fFOsbL0t5bCpBFSpyCY5GGYxnwX0KJpyfIMEOXuy9TgCySO3efZ2Y3A98C84B/ZJrgo8J7SwGySAlTzrFIA2JmvYE+wHB3X5Bthq7Uh3i8X6c5yCL1kZl1BOa6+5KqZr6r8N5SDrJICdLIsUgD4u4TgYlQdoFelv0qjiA7oapEagTZYhsKkEUAd/8Zyl342ijTBDoaQRYpfQqORRqo6qaurfAh/uf4IZ4KkLdN7acAWRqK9C+UFb9cmlkTd1+alqK0PH2UOJ0CZJHSpuBYRLJSgCwSpN4HZtYEWNXdv00bJW6Smm7dzJ4HvnL3k6qaIloBskjpUnAs0kBkSqNI/1DPRgGyNHTxvbPMzJoDLwItzewwd59gZs3dfVHc73lgd2BbM3vb3R+sql0FyCKlqVH1u4hIfRc/gFM/B/cys0PjB/5SM2tc3fGpD/F4/8/ANUAq0N6W8CG+s5k1jeewjA2J1EOpHGJCYLwd0Be4yczWc/dFFrxACIznA88Dk3NsO/29dS5wNWVTqacC5F1S7y0RKTwFxyIruTg6nLo6/gzgn4Rpbd9IjYjl0k6OAfJOZta0unxmkXqoBWGijzlAO2B74Coz60OYcXI3ygLj69x9RK4N5xAgnwXspQBZpG6olJvISqxCLuSdwMFA27h5AvA3d787zzbTS1FdSVmKBcDHwBnu/koS/RcpJWbWBjgcuBZoRahrvADoQgiMXyAExu/E/bNWhMnSfqYyb8uAprHtX7v7vOSekYhkouBYZCVV4YP2aWBvwgftXOBSwgx5P+c6clxF21cCJwPNgaXAuu4+KZEnIVJizKwTsA9wKyFAhvC+egU4093Hxf3yCozT2k9/b10GnAf8Amzv7h/X/hmISHUUHIushNJrrJrZE8B+hBGu4cDf3f3pTPvmeY70D/GbgUOBbd390wSegkjJSatO0QZ4lzDbJIQvhSOBA919Wi4XulZznvT31oXAI3pfidQdBcciKzEzuwkYAiwi/Cx7jbsPj9vKBcUVAuqcRr0qfIiv4u4/FuJ5iJQSM3sEOBBYSKj61ISQYvEWcIq7f1nTkeO0c2SskSwihacL8kTqGTPrUM321Ox1uxAuEgL4HLgpLTC2tEA49f+BFR/k6RMdVHWuChcSKTCWlZ6ZrUbIL/4eeAO4Pj5uTahkcaOZ9YkjzDWu2qLAWKR4VOdYpB4xs2eA1c1sT3f/PtM+aaNV2wHrxvs3uvtrsQ0jXECX2q+pmfUFTjCztYA2hJ+I73L30dWlXehDXBoSd//OzC4GhgGfEN4rX1N2kd62wN/N7KRYB7lWI8giUveUViFST5jZPcDv48MXgD+6+7dZ9m1KGNUaDEwFBrn7jIo/1ZrZToR85GMII18pDswG9nP3t/QBL1KembVw94XxfntCzv01lFWxeAvIGiBXXFfT3H8RSZ7SKkTqj38Rfr5dCOwBXF7Fz7aNgQ7xfiOgGaxIg2hkZo3N7HRCxYo/EQLjVImoRYSR5fbAQ2bWX4GxNGTp77PU/VRgHO/PItQOP5OyFIvUCHKlFIu0C/tWNbOdYxvL01KcRKSI9EYUqSfc/W1gR8JkBBOAP2cLWuMH97OEElOrAxfHD+IuwFHAvwmjXKsRSrC9Q5jE4xDgbGBmbKoDIRDXrHfSYGQLUrONAAO4+xwqB8jbUT5AbhxHiN3MuhFKIP7LzG6MbWjkWKQEKK1CpJ4xs42AaekXwJlZJ3efWWG/Q4AbgVUII8GfEfKJO1A2EchoQn3W8yqkWxwG3EkInJ9z930K9oRESkiFCiwbEGao2wzoDCwG3gY+d/eXsxzTljBRSCrFYj7wJrGKRdynG6GKzOHAmoQvo+u7+/Q6eZIiUiVdkCdSz7j7mPTHZvY+4aK6Pd39+9Tolrs/ZGZrA38mBMPrV2jq78BT7v5qbCe9NusMQmrGcuDnQj4fkVJRYUbJK4FdgI0r7HYQ0NjMbifUHx6Wqtri7svcfY6Z3R/3TQXIOwAPm9mJhAD7T4RKMmsQAuNtFRiLlA4FxyL1WKxesWl8eJ+ZHZE+CYG7/83MfgR2B7YmXGQ3EnjU3R9La6eRuy9NuyioOWX/f/ih7p6RSHHE4DYVGD8KHJC2OVXP2AhfGgGOAzYzswfd/ZoKAfLcGCAvBW4gBMgDgNcIqU7NYls/Aju6+2eFf4YikiulVYjUY2a2DfA8IQ+5MfAq8PuKs3SZWQugIyE4XuruizK0lbpIaC3gUWAj4Avg0FjSTRUrZKWU/rdtZg8CvyZcmPoNIb3oU0J6RN+4bTOgZTx8CiHX+PxUXnFaikUzwlTT9xBSmlLmAF8Ch7v754V9diKSLwXHIvWcmW0OvE4Y7TXKB8iN4lXw6R/+lUpGpQXGqwNDgf8jXFD0b0I5qtl194xEisPMLgX+Qkh9eBG4IVUfPG2fRsBphJHlreLqb4Gb3f3KuE/FMm3rA2cAGxCC7FeA+939m8I+IxGpCQXHIiuBqgLkHI5NBdB9gLMII12rEUaN93D3SYXqt0ipiDPf/Y9wAd5Mwi8mL8dtjdPTJmKAvAdwIrBnbOJjwujx0xXaTb2/NB20SD2hUm4iKwF3f59w0c8iwgQeOxNykLtn2j+9LFv84B4E3Av8hhAYTwd+pcBYGpCdCIExwJ1pgbGlgtoYGJu7L3f354CbCJUoAPoDB1UseZj2K82KX2tUFlGktCk4FllJ5BMgp6VY7G1mQwnlqbYiTPzxBbCdciGlIUiradw/LpcS6oinqleU+3m1woQeLwEPUlYX/PfAwZnOk96OcvdFSpuCY5GVSD4Bspn9FngaOJ+QjjEVeAzYU4GxNECrx2VjyipSZJtkx9NKJt5GmEQHQiWKzUCjwyL1mYJjkZVMrgGyu98P3Eb4/8BrhAvxjnP3iXXZX5FiSkt7+CouDVgrbsuaI5yqTBEf3kYo99YY2M7MWmh0WKT+UnAsshKqLkA2s6ZxvxOBQwl1kO919xlF6bBIkaSN8H6XtvrXZrZJdcemBc8TCaPGUL5GuIjUQwqORVZSVQXI7r4kLUB+MNZD1kiXNDhpI7z/AcbH+12A3c2sdXXHx5zlRsCSuOoXQt6yiNRTCo5FVmJVBMjd3H1JhX0VHEuDFAPcxYSc+zmEGe3OIEwfXdVxFtMyViFczArwtbsvLGB3RaTAFByLrOSyBMhPmdmqxeyXSKGlVaKocnsszbYEuB+YHDd3Av5pZvvGGSZTxzSOy9TEOT2BCwi5yt8QqlfogjyRekzBsUgDkBYgz4mrNiXkRoqslOKkG8vj/Z4xyP2Vme1kZt1gRY3v9GB3PHACMC820wW4G/iTmW0Qj0nVPHYz6wWcCfSL+38CjE5tr4vnKSLJ0wx5Ig2ImW0NPATs4u6fFbs/IoUQ6xMvjfevJ3wx3Chtl/eB1939nLRjDFYEvTsBzwCpEeO5wOeEUeGxhPzirsCfgZ5AZ8Ko8U7unqp6ISL1lIJjkQbGzFq6+4Ji90OkENKnaTazJ4D9qtj9eeCUVEAbA2SLI8rbAncSZoxslXbMXKApoSJFqpTbRGDvOPIsIvWcgmMREVnpmNk9hBnrlhBqED9HSJNoBWyZtuubhNSIUTEoTh9B7gf8EdgTWDvDab4GPgbOdPcJBXoqIlLHFByLiMhKxcz2J4z6tgJeBm519xdi+cKmhEoUpwHtCNfeDI/rPswQILcFWgPHEUaRBxBmk5wCPECoTvFzHT49ESkwBcciIlKvmVmjtJnuMLPLCfnAY4GT3f3NDMccDZwH9CIEyG8TRpBXBMjp00SnHbcin1lEVk6qViEiIvVWDFaXV1i9c1zelAqMU6PBaeXb7gIuJuQLLwe2Aa4FNo3Btsf9PP34CudWuTaRlZCCYxERqZfiqG6qKsX7ZvZ63DQTmAU8HbelB7vL0wLkfwOXUhYgb01agJx+rrTjl1ZcJyIrF6VViIhIvWZmI4DN48OX43IdYMNslVnSUzHM7HfAhUBvMuQgF7LvIlJ6FByLiEi9Zma7E8qypVsADHT3L6s4TgGyiFSitAoREanX3P1FYMe0VQvjckeLshxXMcXiEsqnWFwBDK5uGmoRWblo5FhERFYKZrYDMCxt1bvAXu4+q6oqE1lGkFcnzJD3AnCguy/MdKyIrHwUHIuIyEojQ4D8jLvvF7flGiAfDtwIOGFK6I8L2mkRKSkKjkVEZKWSIUB+yt33j9tyDZB/A4xMTS0tIg2HgmMREVnpJBEgi0jDpOBYRERWSjUNkEWkYdMVuCIislJy99eBndJW7WdmT8RtS82sSTH6JSKlTcGxiIistBQgi0i+FByLiMhKLYcAuXEx+iUipUk5xyIi0iBkyEF+3d13yry3iDRUCo5FRKTByBAg93T3b4rTGxEpRUqrEBGRBiOmWOweH/ZTYCwiFWnkWEREGhwza+Xu84vdDxEpPQqORUREREQipVWIiIiIiEQKjkVEREREIgXHIiIiIiKRgmMRERERkUjBsYiIiIhIpOBYRERERCRScCwiIiIiEik4FhERERGJFByLiIiIiEQKjkVEREREIgXHIiIiIiKRgmMRWemZ2etm5vG2Q5Z97knb58g67WA9YWZHpr1G9yTQXp2/5mbWK+2ck+rinDVhZjuk9fP1YvdHpCFRcCyykqsQGGa6zTGzSWb2pJmdZGbti91nERGRYlFwLCJtgDWB/YCbgClm9vvidmnll/4Fpdh9ERGRMk2K3QERqVMfAO+nPTagA7AZsE5c1w6418xauPs/6rZ7IiIixaXgWKRhec7dh2baYGYHAHcDqbSKm8zsOXf/pq46V0zufiRwZJG7ISIiRaa0ChEBwN0fBw5PW9UcOKFI3RERESkKBccisoK7PwuMSVu1S7H6IiIiUgwKjkWkonfS7q+VvsHMhqZdSDY0rmtpZseY2UtmNsXMFsftAzM1bmY7m9ntZjbOzGaa2SIz+87MXjSzIWbWMteOmlkjMzvCzF42s+/NbGFa5Y3983nSNSkrZmbbmNmNZjbazH40syVmNtvMPjaze83s0PTnk16eq0I72SqJ9Kri3D3M7AIzeyu+fovi6znazK4xs3XzfP6bmNk/zexrM1tgZtPN7H0zO9vMOuXTVtLMrKmZ7W5mV5nZa/H5Loz9/MbMnjezU82sTS3OsbOZ/dfMJqQ9/7fi32TzPNtqbWbHm9nTZjbZzObHqjBfmtldZrZTTfspIoWnnGMRqejntPvtqtrRzNYHHgY2rK5RM+sB3AfskGFz93jbDTjXzP7P3d+qpr1uwBPAFhU2rRlv+5nZ48AR1fUtX2a2BnAXsGuGzW2BfvH2e+A9YHCC524EDAXOAlpU2NwM6AgMBE4xs6uA8929yooYZnYZcA7QOG11C6AL4WLNk8zs10n0P1/x72Y00DnLLqvH2x7A+WZ2qLu/nEf7TYGbgWMrbGoBbBNvJ5jZ/u7+RQ7t/ZpQ9aVbhs1rx9tRZvYM8Ft3n5VrX0Wkbig4FpGKOqbdr+qDuzPwAtATWAi8DUwmlIYrFwzGIPpVQgAM4MAo4FNgASG42Y4QWK4GvGxme7r7a5lObGYdgGHA+mmrJwLvAosIwfrmwAHA8qqebL7MbEPg5bTnAvAjYcR9OiGo6gNsDLSkfAD7LXBLvH9i2vpbyGx2hXM3Bh4EDqrQ5vvx3G0IXxb6EP7/fh7QlcqBX3qbfwPOTVs1n/DaTiMEeDsBawDPATdka6eAWlMWGP8MjCP8nc0lfBnoTfh7axH3e87Mtnf3dzK0lcmVlL0+Y4GPCFVcBgEbxPXrA8PMbEt3n5qtITM7Dbg2Hg/h3+9d4BvCF48NgU3j9n2A181sa3efn2NfRaQuuLtuuum2Et+A1wnBqANDc9h/TNr+71fYNjRt25K4fBjoWmG/RkDTeL81IQhOHfcc0CfDedsBt6bt9x3QPksf70zbbxFwdIZ9Ngcmpe2T2n+HLG3ek7bPkVn2aQd8kbbfdOBQwDLs2xo4DLgrS1upNjyPf8tL0o6bBhyY5dy/Bn5J2/eQLO1tR/jykNrvYaBjhX3aA//N8Drek8DfZi6v+ZqEkdjNgUZV/Ltck9bW51Xs2yttv8VxOQPYLcO++xK+IKb2f6GK57IzsCztdfoz0CrDfgMJAX6qzVuztLdD2j6v1/a11k033XK/KedYRFYws72BAWmrXq1i9ybAS8Bv3H16+gZ3X+7uS+LD0ykb4X0c2MfdJ1RszN1nu/sJwL1xVXfguAx9XBc4Km3Vse5+V4b23iekacwnjDAm4WzK6kHPArZx9/+6e6W0BXef5+4PuPvRSZw45h+fFx/OjOd+LMu5HyaMmqcMNTOruB9wOWWjnK8C/+fu6Wk1ePjZ/3DCv3VSr2PO3H2yu5/s7u+7e8ZfAeLfzpnA7XHVusDuOTTflPDlYD93fylDu09TfpR+90z5wjHV5TbKruP5P3e/0jOMCLv7R4RA+oe46g8xTUdESoSCYxEBIF7Adn/aqkWEkdyqnJotYIltNgWGpLV3XFX7R+cRRsugfGm5lGMoC+jed/d7M+wDgIcc0RuqOV9O4kVZ6akQ57j750m0naNTKMsJviTTF4x0HlJSXowP1yekeawQU122Slt1srsvy9LWcuAkyv5dStXdafdzrbTyH68iBcPdXwEeS1v1xwy77UvZl6YnPJRFzMrdv6fs77IpcEiOfRWROqCcY5GGZS8z61JhXQfCT9brVFh/uleRXwmMdffPqjnfpsAq8f6r7v5jdR109+/MbDwhoOtnZu29/EVLO6bd/3d17RFGos+rdq/qDSa8VgBzKBvhrit7pd1/IMdjhlE2groNIc87Jf11HOnun1bVkLt/YWYjgC1zPHfi4petLYCNCPnQbSn/OdY27f7AHJu9L4d97iWksED51y2lpv82KdsA1+V4nIgUmIJjkYZls3iryhzgFHe/u5r9RuZwvvRAag0zuzmHY6AsCDXCxWCzAGJqwEZp+71bXUMxqJsJ1LYcWfpFhiPcfUEt28uZmXUmpApAyJO9KHOWRCUbpN3vUWFb+khyta9j2n51HhzHcnjnEdJsKn65yyaX/ZxQTaQ66a/PqmbW3d2npa1Lf00OMrPtc2izfdr9iv82IlJECo5FZC7wE+FK/VeA+9z9lxyOm179LqyWdn8A5fOZc5VePaM95fNep+TYxhRqHxyvmnb/61q2la/0yhjNKJ/ekauOFR53Tbufz+tYp8ysI2GUdWCeh7atfhd+dvc51e3k7tPNbCFllUe6Ei6ITEn/O/9N7l1coeK/jYgUkYJjkYblYncfmlBbuYyctq9+l2ql/3+q4iQPuZbAmpdAP9KDrbkJtJePpF9HKP9a1uXrmK9bKAuMFxPSIJ4GPiMEqAtSudLxosWJcd9crqnJp4TaPMqC44qBd23/ffRZLFJC9IYUkUJKD6ZucvdTatlexaC0FbkFbK1reV4I6SYpNZ6JrYbSn+Nsd08iWE5/LVvleEwSr2POzGx14P/iw+XAHp6l9nWUy2hxulyfN5R/7hVHm+dRFiBv4u6j8+yHiJQQVasQkUL6Ie1+phnD8jWLMHqY0jPH45LI6Ux/Lr0TaK+m525nZvkEddmkp8XU5euYj50oq0zyfDWBMYSayPnomMuU0/Ei1vTJXGZU2CXpv3MRKSIFxyJSSOkXO22VpdZuzmJN3zFpq6qdltnM1iH71MP5GJF2f8t4kVidiBd/pVcO2SrbvnlIH93MdXrrur4YLz2X9+Mc9t8uz/aNytOPZ5L+vH9w9+8qbE//O986zz6ISIlRcCwihTScMFMbhKoT+ybQZvro4W9z2P/3CZwTQnCcmiCjbQLtLkzdiSXKqvNM2v0TanluKP86bmpm61W1s5mtTd0Hx+k1sascLY+j6TX5N/ldDvukt5tp9Dr93+ZoM2uRYR8RqScUHItIwbj7IspPwnFrzCPNiZmtmmH1nWn3B5tZ1gA5BnSn5Xq+qsTnkj4pypVm1rcWTf6Udj+X1+RawvTEAAeY2ZG5nsjMKv3UH2tUp09+cWOc6S3T8Y0IUzjXauS/BtKrguxlZo2z7hlen0x/L9X5rZllHT02sx0pP0vevzLs9ijwVbzfnfB3ntNrZWZtzKxOc7lFpGoKjkWk0K4FxsX7qwMfmtmvqwjEupjZsWY2Cjir4vY46909aav+ZWZHZGhnU+BlwoVUiytur6GrgNTMdO2Bt83s/zIFQmbWyswONbNKU1tHn6Td/3V1J44z4l2WtuouM7smw6QuqfM3MbPdzOzflE+hSPcXyma92w14wMw6VGinHWGylT1J7nXM1TDKKkqsDdybqX9m9g9CDeR8q2ksIcw6+IyZVZpRL06n/jhlXwpedvdKU6rHahnHU/bl5Sjg2TgLYUZmNtDMriSky9R1DruIVEHVKkSkoNx9rpntR6ih3JtwwdJDwIw449r3hOCjE2HSinUo++I+rHKLAJxO+Im/L9AcuMfMLiRM1rAI2JAw658Rpv7tDOQyMUN1z2W2mR1ICLpXIUw08V/gBjN7h3CRWwugD7AJ0JLyOdLpHqVs9rorzWxPwpeIRWn7/NXdf057fDHQCzgiPrczgJPM7ENC0D4faBf3GUBZhYX0Uer05/O6mV1D2ZeQ3wD7mNkwwr/LqoSL4toQUkpuBIZmeT6Jc/efY/8ujKsOB/Y0s/eAbwmjtDsQnudSQrpJPjMXfkcIfk8FXjazMcBHhNd2EOHvKGUamaeOTvX1FTM7HriNEHDvCexhZp8SaojPJqSGdCdMZNM1W1siUlwKjkWk4Nz96ziSeztwMCH46ALsU8Vhv5DlIqwYNO0EPEmYohpgrXhL9xRwJKEubiLcfayZbU6ot5u6AGxV4IAsh2SriXwPIWd6O8LrsSOVpya+mbI859QFiUea2UhCoNyRMCnIVmS/SM8Jud/Zns/ZZrYMOJvwpaQ1lXPDvyP8u9UmjaSmLiEE+6m8306EwDPdL4TR2o9q0P7ZhBzyYwhB60YZ9vkc2N/dJ1fVkLv/08y+Au4gfMkzQoC9YRWHjQNm1qDfIlIgCo5FpE64+0zgEDPrBxxKGPHrTRjVXU4IcL4CRhFGmV9294UZGwvtfWdmgwlB0+GEkdL2hLJaYwgjiI+6u9eySEamc08GtjeznQkpEdsSRgTbEX7an0yYXvtZQoCeqY0l8af8Ywg5rf0IgV+zTPtXOPbvZnYP4WKyXSkbiWxBqMH7DSHoeh14zt2nZm5pRXvnmtkjhJHXneJzmQtMIoy8/8PdZ9Qyx7pGYsrCEWb2MHAsobpER8KXhimEL0h3xb+HXjVofwnwh9j+MYTp1bsT/h0/Ax4kPP9F2Vsp195rMZ1if2BvQiWQboS/jfmEv8/xhHzv5939o3z7LCKFZWEgQkREREREdEGeiIiIiEik4FhEREREJFJwLCIiIiISKTgWEREREYkUHIuIiIiIRAqORUREREQiBcciIiIiIpGCYxERERGRSMGxiIiIiEik4FhEREREJFJwLCIiIiISKTgWEREREYkUHIuIiIiIRAqORUREREQiBcciIiIiIpGCYxERERGRSMGxiIiIiEik4FhEREREJFJwLCIiIiISKTgWEREREYkUHIuIiIiIRP8Pp/MliAqtb/kAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 900x600 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "C = confusion_matrix(Y_true_max_TT, Y_pred_max_TT, normalize=\"true\")\n",
    "C = np.around(C, decimals=3)\n",
    "\n",
    "plt.figure()\n",
    "plot_confusion_matrix(C, classes=['isTT', 'notTT'], title='')\n",
    "plt.show(1) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "a1234dc6-62fb-4667-bdc4-46e3d3a06b91",
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "11166\n",
      "11166\n",
      "        CleanJet_eta_1  CleanJet_eta_2  CleanJet_phi_1  CleanJet_phi_2  \\\n",
      "337598        2.383301       -2.875977       -1.389404        2.119141   \n",
      "286129        0.843384       -2.560547       -0.956909       -1.310791   \n",
      "270333       -2.679199        2.759766        0.014395        3.098145   \n",
      "272966        1.907715       -4.047852       -2.846680       -1.485107   \n",
      "284058        2.180176       -1.959961        0.479370        0.112045   \n",
      "...                ...             ...             ...             ...   \n",
      "264064       -2.783203        1.704102        2.445801       -2.129883   \n",
      "264069       -2.909668        2.610352       -0.898071       -2.617676   \n",
      "264072       -3.666992        1.599121       -0.727783        3.035645   \n",
      "264073       -2.077637        2.407715       -1.190186        2.056152   \n",
      "264075        1.647949       -1.532715        0.555176        0.346863   \n",
      "\n",
      "        CleanJet_pt_1  CleanJet_pt_2  Lepton_eta_1  Lepton_eta_2  \\\n",
      "337598     118.524651      89.139381      0.882812     -1.516602   \n",
      "286129     286.343903     190.126434      1.186035      0.298401   \n",
      "270333     258.932007     197.931213      2.397949     -0.695068   \n",
      "272966      90.206482      53.956184     -1.375732      0.610229   \n",
      "284058     238.594788     129.533737     -0.570801      0.579834   \n",
      "...               ...            ...           ...           ...   \n",
      "264064     113.189606      96.969040     -1.578369     -0.184906   \n",
      "264069      66.415848      32.239452     -0.527222      0.111496   \n",
      "264072      57.769115      37.780602     -1.773682      0.385132   \n",
      "264073     241.773880     193.358170      0.198944     -1.650391   \n",
      "264075     195.202301      85.408516     -0.816406      1.420410   \n",
      "\n",
      "        Lepton_phi_1  Lepton_phi_2  Lepton_pt_1  Lepton_pt_2        Rpt  \\\n",
      "337598      0.420105     -0.477295    32.782749    16.477949   0.051129   \n",
      "286129      2.091797     -0.383667   212.588989    51.488895   0.201059   \n",
      "270333      1.926025     -0.206329    55.863281    29.988325   0.032687   \n",
      "272966      1.631836     -1.266602   243.203903   206.228958  10.304838   \n",
      "284058      2.759766      1.058350    84.097641    23.937065   0.065134   \n",
      "...              ...           ...          ...          ...        ...   \n",
      "264064      0.171387      0.244446    70.472191    64.676483   0.415264   \n",
      "264069      3.041504     -0.092667   246.034637   147.796158  16.982437   \n",
      "264072      2.532715      1.204590    63.975708    21.955809   0.643576   \n",
      "264073      2.718750     -0.684448    75.338005    31.950323   0.051489   \n",
      "264075     -2.726074     -0.626831   188.085938    26.485250   0.298796   \n",
      "\n",
      "         Zepp_l1   Zepp_l2   Zepp_ll    detajj    detall    dphijj  \\\n",
      "337598  1.129150 -1.270264  0.070557  5.259277  2.399414  2.774640   \n",
      "286129  2.044617  1.156982  1.600800  3.403931  0.887634  0.353882   \n",
      "270333  2.357666 -0.735352  0.811157  5.438965  3.093018  3.083750   \n",
      "272966 -0.305664  1.680298  0.687317  5.955566  1.985962  1.361572   \n",
      "284058 -0.680908  0.469727  0.105591  4.140137  1.150635  0.367325   \n",
      "...          ...       ...       ...       ...       ...       ...   \n",
      "264064 -1.038818  0.354645  0.342087  4.487305  1.393463  1.707502   \n",
      "264069 -0.377563  0.261154  0.058205  5.520020  0.638718  1.719604   \n",
      "264072 -0.739746  1.419067  0.339661  5.266113  2.158813  2.519758   \n",
      "264073  0.033905 -1.815430  0.890762  4.485352  1.849335  3.036847   \n",
      "264075 -0.874023  1.362793  0.244385  3.180664  2.236816  0.208313   \n",
      "\n",
      "        dphilep1jet1  dphilep1jet2  dphilep1jj  dphilep2jet1  dphilep2jet2  \\\n",
      "337598      1.809509      1.699036    2.545335      0.912109      2.596436   \n",
      "286129      3.048706      2.880597    3.093631      0.573242      0.927124   \n",
      "270333      1.911631      1.172119    1.727184      0.220724      2.978711   \n",
      "272966      1.804670      3.116943    2.284535      1.580078      0.218506   \n",
      "284058      2.280396      2.647720    2.409082      0.578979      0.946304   \n",
      "...              ...           ...         ...           ...           ...   \n",
      "264064      2.274414      2.301270    3.039868      2.201355      2.374329   \n",
      "264069      2.343610      0.624006    1.866233      0.805405      2.525009   \n",
      "264072      3.022687      0.502930    2.339893      1.932373      1.831055   \n",
      "264073      2.374250      0.662598    1.986340      0.505737      2.740601   \n",
      "264075      3.001935      3.072937    3.065214      1.182007      0.973694   \n",
      "\n",
      "        dphilep2jj    dphill  dphilljet  dphilljetjet  dphillmet  dphilmet1  \\\n",
      "337598    1.647935  0.897400   1.518837      2.254663   1.391693   1.101021   \n",
      "286129    0.714090  2.475464   2.865901      3.006748   0.385486   0.202680   \n",
      "270333    0.405171  2.132355   1.344980      1.160533   2.821044   2.895490   \n",
      "272966    1.100212  2.898438   2.661272      3.141138   0.506327   0.350275   \n",
      "284058    0.707666  1.701416   1.995304      2.123991   0.942800   0.657708   \n",
      "...            ...       ...        ...           ...        ...        ...   \n",
      "264064    2.966809  0.073059   2.239452      3.004906   0.659730   0.624767   \n",
      "264069    1.282781  3.134171   2.354775      1.877399   2.760380   2.771545   \n",
      "264072    2.615168  1.328125   2.961942      2.638448   2.750068   3.048624   \n",
      "264073    0.893647  2.879987   2.190545      1.802636   0.248619   0.064915   \n",
      "264075    1.118729  2.099243   3.132095      3.087812   0.255914   0.125754   \n",
      "\n",
      "        dphilmet2   dr_l1j1   dr_l1j2   dr_l2j1   dr_l2j2      drll  \\\n",
      "337598   1.998420  2.350700  4.124950  4.005145  2.930764  2.561740   \n",
      "286129   2.678144  3.067901  5.061075  0.790957  3.005518  2.629794   \n",
      "270333   1.255341  5.425106  1.226692  1.996370  4.780735  3.756820   \n",
      "272966   2.548162  5.553209  4.105552  2.044533  4.663203  3.513543   \n",
      "284058   2.359124  3.573244  2.990015  1.701855  2.710360  2.053966   \n",
      "...           ...       ...       ...       ...       ...       ...   \n",
      "264064   0.697826  2.573827  4.008797  3.405453  3.034104  1.395377   \n",
      "264069   0.362625  4.603944  6.470756  3.126677  3.552457  3.198591   \n",
      "264072   1.720499  3.770341  3.410093  4.489295  2.196937  2.534638   \n",
      "264073   2.944902  4.523560  2.306015  0.662049  4.896847  3.422625   \n",
      "264075   2.224998  4.103614  3.155319  1.203709  3.109506  3.067600   \n",
      "\n",
      "                 ht      m2ljj30         mT2         mTi       m_l1j1  \\\n",
      "337598   339.462921  1519.088379   26.618164  134.764969   143.575485   \n",
      "286129  1119.883301  1953.555786  243.169968  485.963898   507.333679   \n",
      "270333   766.178467  3811.374023  167.457458  373.581116  1526.124756   \n",
      "272966   651.396057  2094.372314    4.735299  742.055786   772.482910   \n",
      "284058  1011.073364  1535.571289  197.312439  251.151672   585.199219   \n",
      "...             ...          ...         ...         ...          ...   \n",
      "264064   371.809204  1174.807861   26.045704  132.932007   199.041229   \n",
      "264069   553.340759  1149.969238    0.000089  473.496582   448.914703   \n",
      "264072   237.361694   783.860474   67.512421  208.441788   180.422394   \n",
      "264073   799.274231  2225.334961  114.478539  230.603668   453.701508   \n",
      "264075   603.581787  1025.143799  130.646088  288.015442   712.914246   \n",
      "\n",
      "             m_l1j2      m_l2j1       m_l2j2       mcoll     mcollWW  \\\n",
      "337598   355.423462  307.582367    93.615479  110.806091  200.008820   \n",
      "286129  1338.804932  115.653931   400.019073  790.788086  814.426270   \n",
      "270333   123.734459  206.634979   447.273529  333.408112  387.261322   \n",
      "272966   465.981781  272.001343  1073.157593  664.104370  690.375366   \n",
      "284058   257.557739  143.835754   191.240112  568.224548  591.643616   \n",
      "...             ...         ...          ...         ...         ...   \n",
      "264064   437.868469  328.384644   227.315353  133.559219  220.539291   \n",
      "264069   412.734161  434.086823   257.283569  413.936676  444.076447   \n",
      "264072   257.667053  271.978271    59.097996   30.692678  163.998138   \n",
      "264073   334.738129   60.425999   608.078491  430.420624  459.862305   \n",
      "264075   270.240753   85.340225   202.660156  349.710571  402.950684   \n",
      "\n",
      "                mjj         mll        mtw1        mtw2  phi_miss   proxyW_l1  \\\n",
      "337598  1432.898682   72.982231   54.423107   62.033714  1.521126   72.188995   \n",
      "286129  1246.233887  219.794067   54.810802  259.506866  2.294477  344.838074   \n",
      "270333  3449.899414  196.947800  168.150620   72.901855 -1.461670  305.814148   \n",
      "272966  1370.356079  685.333008   41.317856  208.817032  1.281561  181.059891   \n",
      "284058  1373.746216   86.735428  112.745697  172.176804 -2.865711  151.363754   \n",
      "...             ...         ...         ...         ...       ...         ...   \n",
      "264064   989.830933  101.993164   26.563128   28.308128 -0.453380   80.051178   \n",
      "264069   731.761353  400.993500  240.546173   34.202213  0.269958  249.241867   \n",
      "264072   653.152222  107.946716  119.453293   53.101513 -0.515909  169.680496   \n",
      "264073  2059.729492  142.574387    6.611000  132.024399  2.653835   21.347055   \n",
      "264075   608.590027  228.472031   17.944408   96.104828 -2.851829  121.173744   \n",
      "\n",
      "         proxyW_l2     pt_miss        ptll      recoil       yll  isLL  isMIX  \\\n",
      "337598   87.315178   82.538200   44.944496  100.768471 -0.033654   0.0    0.0   \n",
      "286129  262.560852  345.188141  175.023727  511.618225  1.002191   0.0    1.0   \n",
      "270333   85.159302  128.470413   47.284763   84.911514  1.133958   0.0    1.0   \n",
      "272966  219.505905   57.800549   65.711418  119.591423 -0.445245   0.0    1.0   \n",
      "284058  180.674133  362.294464   84.386032  417.495605 -0.293126   0.0    1.0   \n",
      "...            ...         ...         ...         ...       ...   ...    ...   \n",
      "264064   29.327820   26.501902  135.058685  156.842728 -0.907472   1.0    0.0   \n",
      "264069   35.801060   60.854649   98.248672   47.504860 -0.285076   1.0    0.0   \n",
      "264072   54.808876   55.880455   72.456940   29.791912 -1.103437   1.0    0.0   \n",
      "264073  179.985840  137.714737   45.235912  181.900665 -0.422297   1.0    0.0   \n",
      "264075  126.768883  108.399780  176.222916  282.428436 -0.403406   1.0    0.0   \n",
      "\n",
      "        isTT  isBKG  isVBS  notLL  notTT  DNN_SigVsBkg    DNN_LL    DNN_TT  \n",
      "337598   1.0    0.0    1.0    1.0    0.0      0.978351  0.780982  0.129956  \n",
      "286129   0.0    0.0    1.0    1.0    1.0      0.620560  0.204644  0.535192  \n",
      "270333   0.0    0.0    1.0    1.0    1.0      0.854526  0.208633  0.392652  \n",
      "272966   0.0    0.0    1.0    1.0    1.0      0.938182  0.779758  0.610374  \n",
      "284058   0.0    0.0    1.0    1.0    1.0      0.524678  0.114108  0.539909  \n",
      "...      ...    ...    ...    ...    ...           ...       ...       ...  \n",
      "264064   0.0    0.0    1.0    0.0    1.0      0.963985  0.697702  0.222587  \n",
      "264069   0.0    0.0    1.0    0.0    1.0      0.713122  0.956053  0.131631  \n",
      "264072   0.0    0.0    1.0    0.0    1.0      0.650788  0.768116  0.179261  \n",
      "264073   0.0    0.0    1.0    0.0    1.0      0.707097  0.415022  0.421666  \n",
      "264075   0.0    0.0    1.0    0.0    1.0      0.814468  0.238773  0.364476  \n",
      "\n",
      "[11166 rows x 67 columns]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_519/3661036990.py:3: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  df_val_LL['DNN_LL'] = dnnLL.predict_proba(X_val_LL)[...,0].numpy().tolist()\n",
      "/tmp/ipykernel_519/3661036990.py:4: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  df_val_LL['DNN_TT'] = dnnTT.predict_proba(X_val_LL)[...,0].numpy().tolist()\n"
     ]
    }
   ],
   "source": [
    "print(len(dnnLL.predict_proba(X_val_LL)[...,0]))\n",
    "print(len(dnnTT.predict_proba(X_val_LL)[...,0]))\n",
    "df_val_LL['DNN_LL'] = dnnLL.predict_proba(X_val_LL)[...,0].numpy().tolist()\n",
    "df_val_LL['DNN_TT'] = dnnTT.predict_proba(X_val_LL)[...,0].numpy().tolist()\n",
    "print(df_val_LL)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "a93cc097-ea4a-4f84-a256-9703427ff3a8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "LL = df_val_LL[df_val_LL['isLL'] == 1]\n",
    "mix = df_val_LL[df_val_LL['isMIX'] == 1]\n",
    "TT = df_val_LL[df_val_LL['isTT'] == 1]\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "h_isLL = ax.hist2d(LL['DNN_LL'].values, LL['DNN_TT'].values, density=True)\n",
    "fig.colorbar(h_isLL[3], ax=ax)\n",
    "plt.ylabel('DNN TT probability')\n",
    "plt.xlabel('DNN LL probability')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "1f0cd41d-696b-4645-8343-fff9ea287e41",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "h_isTT = ax.hist2d(TT['DNN_LL'].values, TT['DNN_TT'].values, density=True)\n",
    "fig.colorbar(h_isTT[3], ax=ax)\n",
    "plt.ylabel('DNN TT probability')\n",
    "plt.xlabel('DNN LL probability')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "6afe52fe-8b78-4b8b-bb16-5ad97e7bdf04",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "h_isMIX = ax.hist2d(mix['DNN_LL'].values, mix['DNN_TT'].values, density=True)\n",
    "fig.colorbar(h_isMIX[3], ax=ax)\n",
    "plt.ylabel('DNN TT probability')\n",
    "plt.xlabel('DNN LL probability')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "ff3b2c7c-47e9-4261-878a-e5b2f6dc9404",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_519/3595125263.py:1: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  df_val_TT['DNN_LL'] = dnnLL.predict_proba(X_val_TT)[...,0].numpy().tolist()\n",
      "/tmp/ipykernel_519/3595125263.py:2: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  df_val_TT['DNN_TT'] = dnnTT.predict_proba(X_val_TT)[...,0].numpy().tolist()\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_val_TT['DNN_LL'] = dnnLL.predict_proba(X_val_TT)[...,0].numpy().tolist()\n",
    "df_val_TT['DNN_TT'] = dnnTT.predict_proba(X_val_TT)[...,0].numpy().tolist()\n",
    "\n",
    "LL = df_val_TT[df_val_TT['isLL'] == 1]\n",
    "mix = df_val_TT[df_val_TT['isMIX'] == 1]\n",
    "TT = df_val_TT[df_val_TT['isTT'] == 1]\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "h_isLL = ax.hist2d(LL['DNN_LL'].values, LL['DNN_TT'].values, density=True)\n",
    "fig.colorbar(h_isLL[3], ax=ax)\n",
    "plt.ylabel('DNN TT probability')\n",
    "plt.xlabel('DNN LL probability')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "c09e08d4-55cb-4265-ba47-5d5ef43c2d47",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "h_isTT = ax.hist2d(TT['DNN_LL'].values, TT['DNN_TT'].values, density=True)\n",
    "fig.colorbar(h_isTT[3], ax=ax)\n",
    "plt.ylabel('DNN TT probability')\n",
    "plt.xlabel('DNN LL probability')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "bac8cc9a-7ec8-47be-97e2-7b4262e3433d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x600 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "h_isMIX = ax.hist2d(mix['DNN_LL'].values, mix['DNN_TT'].values, density=True)\n",
    "fig.colorbar(h_isMIX[3], ax=ax)\n",
    "plt.ylabel('DNN TT probability')\n",
    "plt.xlabel('DNN LL probability')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "c51de0ef-f2c1-420c-bf18-f5e2e3cdffb1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n"
     ]
    }
   ],
   "source": [
    "with open('generated_code_dnn_SigVsBkg.h', 'w') as f:\n",
    "\n",
    "\n",
    "    print (\"\"\"\n",
    "    // Select all (Ctrl+A), Copy (Ctrl+C), paste to an empty text file (Ctrl+V) and save that file as 'generated_code.h'\n",
    "    // \n",
    "    // Auto-generated header file. Assumes img to be a floating point array\n",
    "    // of 64 elements (corresponding to an 8x8 b&w image)\n",
    "\n",
    "    #include <math.h>\n",
    "\n",
    "    float norma;\n",
    "    \"\"\", file=f)\n",
    "    \n",
    "    iLayer_tmp=-1\n",
    "\n",
    "    for iLayer, layer in enumerate(dnn.model1.layers):\n",
    "      print(type(layer))\n",
    "#       if type(layer) is not Dropout and type(layer) is not Input :\n",
    "      if type(layer) is Dense:\n",
    "        kernel, bias = layer.get_weights()\n",
    "        iLayer_tmp+=1\n",
    "        print (\"inline float activation_%d (float x, float n) \" % iLayer_tmp, file=f);\n",
    "        activation =  layer.get_config()['activation'] \n",
    "        if activation == 'sigmoid':\n",
    "          print (\"{ return 1./(1 + exp(-x)); }\", file=f)\n",
    "        elif activation == 'tanh':\n",
    "          print (\"{ return tanh(x);}\")\n",
    "        elif activation == 'relu':\n",
    "          print (\"{ return x > 0 ? x : 0;}\", file=f)\n",
    "        elif activation == 'linear':\n",
    "          print (\"{ return x;}\")\n",
    "        elif activation == 'softmax':\n",
    "          print (\"{ return exp(x)/n;  }\", file=f)\n",
    "        else:\n",
    "          raise KeyError (\"Unexpected activation %s\"%activation)\n",
    "      \n",
    "    iLayer_tmp=-1\n",
    "\n",
    "    \n",
    "    print (\"\"\"\n",
    "    float guess_digit (const float *img, int flag)\n",
    "    {\n",
    "    \"\"\", file=f)\n",
    "\n",
    "    max_out = 100\n",
    "    for iLayer, layer in enumerate(dnn.model1.layers):\n",
    "      if type(layer) is Dense:\n",
    "        iLayer_tmp+=1\n",
    "        print (\"  // Declare the arrays in the stack\", file=f)\n",
    "        kernel, bias = layer.get_weights()\n",
    "        \n",
    "        print (type (dnn.means))\n",
    "        print (type(bias))\n",
    "        \n",
    "         \n",
    "        max_out = max(kernel.shape[1], max_out)\n",
    "        #print (\"  //\",bias.shape)\n",
    "        kernel_values = \"{%s}\"%(',\\n   '.join([\"{%s}\"%(','.join([\"%18.13f\"%x for x in row])) for row in kernel]))\n",
    "        bias_values   = \"{%s}\"% ( \",\".join([\"%18.13f\"%x for x in bias]))\n",
    "        print (\"  const float kernel_%d[%d][%d] = \\n  %s;\" % (iLayer_tmp, kernel.shape[0], kernel.shape[1],kernel_values), file=f)\n",
    "        print (\"  const float bias_%d[%d] = %s;\" % (iLayer_tmp, bias.shape[0], bias_values), file=f)\n",
    "      \n",
    "    print (\"  float buffer_in[%d];\" % max_out, file=f)\n",
    "    print (\"  float buffer_out[%d];\" % max_out, file=f)\n",
    "\n",
    "    print (\"  unsigned int i,j,c; \", file=f)\n",
    "    \n",
    "\n",
    "    mean_values = \"{%s}\"% ( \",\".join([\"%18.13f\"%x for x in dnn.means]))\n",
    "    sigma_values = \"{%s}\"% ( \",\".join([\"%18.13f\"%x for x in dnn.sigmas]))\n",
    "    \n",
    "\n",
    "    print (\" float mean[%d] = %s;\" % (max_out, mean_values), file=f)\n",
    "    print (\" float sigma[%d] = %s;\" % (max_out, sigma_values), file=f)\n",
    "\n",
    "    print (\"\\n\\n\\n\", file=f)\n",
    "    print (\"  // Load the input in the buffer\", file=f)\n",
    "    print (\"  for (c = 0; c < 64; ++c) \\n  buffer_in[c] = (img[c]-mean[c])/sigma[c];\", file=f)\n",
    "\n",
    "\n",
    "\n",
    "    iLayer_tmp=-1\n",
    "\n",
    "    for iLayer, layer in enumerate(dnn.model1.layers):\n",
    "      if type(layer) is Dense: \n",
    "        kernel, bias = layer.get_weights()\n",
    "        iLayer_tmp+=1\n",
    "\n",
    "        if iLayer != len(dnn.model1.layers)-1:\n",
    "            \n",
    "            \n",
    "            print ( \"  // Processing layer %i \" % iLayer_tmp , file=f)\n",
    "            print ( \"\"\"\n",
    "            for (c = 0; c < {n_out}; ++c ) \n",
    "              buffer_out[c] = bias_{iLayer_tmp}[c];\n",
    "\n",
    "            for (c = 0; c < {n_out}; ++c )\n",
    "              for (i = 0; i < {n_in}; ++i)\n",
    "                buffer_out[c] += buffer_in[i] * kernel_{iLayer_tmp}[i][c];\n",
    "\n",
    "            norma = 0;\n",
    "\n",
    "            // Prepares for next layer \n",
    "            for (c = 0; c < {n_out}; ++c )\n",
    "              buffer_in[c] = activation_{iLayer_tmp}(buffer_out[c], norma);\n",
    "\n",
    "\n",
    "            \"\"\".format (\n",
    "                n_in = kernel.shape[0],\n",
    "                n_out = kernel.shape[1],\n",
    "                iLayer_tmp = iLayer_tmp,\n",
    "            ), file=f)\n",
    "            \n",
    "        else:\n",
    "            \n",
    "            print ( \"  // Processing layer %i \" % iLayer_tmp , file=f)\n",
    "            print ( \"\"\"\n",
    "            for (c = 0; c < {n_out}; ++c ) \n",
    "              buffer_out[c] = bias_{iLayer_tmp}[c];\n",
    "\n",
    "            for (c = 0; c < {n_out}; ++c )\n",
    "              for (i = 0; i < {n_in}; ++i)\n",
    "                buffer_out[c] += buffer_in[i] * kernel_{iLayer_tmp}[i][c];\n",
    "\n",
    "            norma = 0;\n",
    "\n",
    "            for(c=0;c<{n_out};++c)\n",
    "              norma+=exp(buffer_out[c]);\n",
    "\n",
    "            // Prepares for next layer \n",
    "            for (c = 0; c < {n_out}; ++c )\n",
    "              buffer_in[c] = activation_{iLayer_tmp}(buffer_out[c], norma);\n",
    "\n",
    "\n",
    "            \"\"\".format (\n",
    "                n_in = kernel.shape[0],\n",
    "                n_out = kernel.shape[1],\n",
    "                iLayer_tmp = iLayer_tmp,\n",
    "            ), file=f)\n",
    "      \n",
    "    last_kernel, last_bias = dnn.model1.layers[-1].get_weights()\n",
    "    print (\"\"\"\n",
    "      //i = 0;\n",
    "      //for (c = 0; c < {n_out}; ++c)\n",
    "      //  if (buffer_in[c] > buffer_in[i])\n",
    "      //    i = c;\n",
    "      \n",
    "      if (flag==0) //vbf\n",
    "        return buffer_in[0];\n",
    "      else if (flag==1) //bkg\n",
    "        return buffer_in[1];\n",
    "         \n",
    "      \n",
    "      //return i;\n",
    "    \"\"\".format(n_out = last_kernel.shape[1]), file=f)\n",
    "\n",
    "    print (\"}\", file=f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "4c523d50-4f12-4c56-858f-0800889fe9d7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n"
     ]
    }
   ],
   "source": [
    "with open('generated_code_dnn_LLVsOther.h', 'w') as f:\n",
    "\n",
    "\n",
    "    print (\"\"\"\n",
    "    // Select all (Ctrl+A), Copy (Ctrl+C), paste to an empty text file (Ctrl+V) and save that file as 'generated_code.h'\n",
    "    // \n",
    "    // Auto-generated header file. Assumes img to be a floating point array\n",
    "    // of 64 elements (corresponding to an 8x8 b&w image)\n",
    "\n",
    "    #include <math.h>\n",
    "\n",
    "    float norma;\n",
    "    \"\"\", file=f)\n",
    "    \n",
    "    iLayer_tmp=-1\n",
    "\n",
    "    for iLayer, layer in enumerate(dnnLL.model1.layers):\n",
    "      print(type(layer))\n",
    "#       if type(layer) is not Dropout and type(layer) is not Input :\n",
    "      if type(layer) is Dense:\n",
    "        kernel, bias = layer.get_weights()\n",
    "        iLayer_tmp+=1\n",
    "        print (\"inline float activation_%d (float x, float n) \" % iLayer_tmp, file=f);\n",
    "        activation =  layer.get_config()['activation'] \n",
    "        if activation == 'sigmoid':\n",
    "          print (\"{ return 1./(1 + exp(-x)); }\", file=f)\n",
    "        elif activation == 'tanh':\n",
    "          print (\"{ return tanh(x);}\")\n",
    "        elif activation == 'relu':\n",
    "          print (\"{ return x > 0 ? x : 0;}\", file=f)\n",
    "        elif activation == 'linear':\n",
    "          print (\"{ return x;}\")\n",
    "        elif activation == 'softmax':\n",
    "          print (\"{ return exp(x)/n;  }\", file=f)\n",
    "        else:\n",
    "          raise KeyError (\"Unexpected activation %s\"%activation)\n",
    "      \n",
    "    iLayer_tmp=-1\n",
    "\n",
    "    \n",
    "    print (\"\"\"\n",
    "    float guess_digit (const float *img, int flag)\n",
    "    {\n",
    "    \"\"\", file=f)\n",
    "\n",
    "    max_out = 100\n",
    "    for iLayer, layer in enumerate(dnnLL.model1.layers):\n",
    "      if type(layer) is Dense:\n",
    "        iLayer_tmp+=1\n",
    "        print (\"  // Declare the arrays in the stack\", file=f)\n",
    "        kernel, bias = layer.get_weights()\n",
    "        \n",
    "        print (type (dnnLL.means))\n",
    "        print (type(bias))\n",
    "        \n",
    "         \n",
    "        max_out = max(kernel.shape[1], max_out)\n",
    "        #print (\"  //\",bias.shape)\n",
    "        kernel_values = \"{%s}\"%(',\\n   '.join([\"{%s}\"%(','.join([\"%18.13f\"%x for x in row])) for row in kernel]))\n",
    "        bias_values   = \"{%s}\"% ( \",\".join([\"%18.13f\"%x for x in bias]))\n",
    "        print (\"  const float kernel_%d[%d][%d] = \\n  %s;\" % (iLayer_tmp, kernel.shape[0], kernel.shape[1],kernel_values), file=f)\n",
    "        print (\"  const float bias_%d[%d] = %s;\" % (iLayer_tmp, bias.shape[0], bias_values), file=f)\n",
    "      \n",
    "    print (\"  float buffer_in[%d];\" % max_out, file=f)\n",
    "    print (\"  float buffer_out[%d];\" % max_out, file=f)\n",
    "\n",
    "    print (\"  unsigned int i,j,c; \", file=f)\n",
    "    \n",
    "\n",
    "    mean_values = \"{%s}\"% ( \",\".join([\"%18.13f\"%x for x in dnnLL.means]))\n",
    "    sigma_values = \"{%s}\"% ( \",\".join([\"%18.13f\"%x for x in dnnLL.sigmas]))\n",
    "    \n",
    "\n",
    "    print (\" float mean[%d] = %s;\" % (max_out, mean_values), file=f)\n",
    "    print (\" float sigma[%d] = %s;\" % (max_out, sigma_values), file=f)\n",
    "\n",
    "    print (\"\\n\\n\\n\", file=f)\n",
    "    print (\"  // Load the input in the buffer\", file=f)\n",
    "    print (\"  for (c = 0; c < 64; ++c) \\n  buffer_in[c] = (img[c]-mean[c])/sigma[c];\", file=f)\n",
    "\n",
    "\n",
    "\n",
    "    iLayer_tmp=-1\n",
    "\n",
    "    for iLayer, layer in enumerate(dnnLL.model1.layers):\n",
    "      if type(layer) is Dense: \n",
    "        kernel, bias = layer.get_weights()\n",
    "        iLayer_tmp+=1\n",
    "\n",
    "        if iLayer != len(dnnLL.model1.layers)-1:\n",
    "            \n",
    "            \n",
    "            print ( \"  // Processing layer %i \" % iLayer_tmp , file=f)\n",
    "            print ( \"\"\"\n",
    "            for (c = 0; c < {n_out}; ++c ) \n",
    "              buffer_out[c] = bias_{iLayer_tmp}[c];\n",
    "\n",
    "            for (c = 0; c < {n_out}; ++c )\n",
    "              for (i = 0; i < {n_in}; ++i)\n",
    "                buffer_out[c] += buffer_in[i] * kernel_{iLayer_tmp}[i][c];\n",
    "\n",
    "            norma = 0;\n",
    "\n",
    "            // Prepares for next layer \n",
    "            for (c = 0; c < {n_out}; ++c )\n",
    "              buffer_in[c] = activation_{iLayer_tmp}(buffer_out[c], norma);\n",
    "\n",
    "\n",
    "            \"\"\".format (\n",
    "                n_in = kernel.shape[0],\n",
    "                n_out = kernel.shape[1],\n",
    "                iLayer_tmp = iLayer_tmp,\n",
    "            ), file=f)\n",
    "            \n",
    "        else:\n",
    "            \n",
    "            print ( \"  // Processing layer %i \" % iLayer_tmp , file=f)\n",
    "            print ( \"\"\"\n",
    "            for (c = 0; c < {n_out}; ++c ) \n",
    "              buffer_out[c] = bias_{iLayer_tmp}[c];\n",
    "\n",
    "            for (c = 0; c < {n_out}; ++c )\n",
    "              for (i = 0; i < {n_in}; ++i)\n",
    "                buffer_out[c] += buffer_in[i] * kernel_{iLayer_tmp}[i][c];\n",
    "\n",
    "            norma = 0;\n",
    "\n",
    "            for(c=0;c<{n_out};++c)\n",
    "              norma+=exp(buffer_out[c]);\n",
    "\n",
    "            // Prepares for next layer \n",
    "            for (c = 0; c < {n_out}; ++c )\n",
    "              buffer_in[c] = activation_{iLayer_tmp}(buffer_out[c], norma);\n",
    "\n",
    "\n",
    "            \"\"\".format (\n",
    "                n_in = kernel.shape[0],\n",
    "                n_out = kernel.shape[1],\n",
    "                iLayer_tmp = iLayer_tmp,\n",
    "            ), file=f)\n",
    "      \n",
    "    last_kernel, last_bias = dnnLL.model1.layers[-1].get_weights()\n",
    "    print (\"\"\"\n",
    "      //i = 0;\n",
    "      //for (c = 0; c < {n_out}; ++c)\n",
    "      //  if (buffer_in[c] > buffer_in[i])\n",
    "      //    i = c;\n",
    "      \n",
    "      if (flag==0) //vbf\n",
    "        return buffer_in[0];\n",
    "      else if (flag==1) //bkg\n",
    "        return buffer_in[1];\n",
    "         \n",
    "      \n",
    "      //return i;\n",
    "    \"\"\".format(n_out = last_kernel.shape[1]), file=f)\n",
    "\n",
    "    print (\"}\", file=f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "72c6e0df-9dac-4ae1-b71f-7a1c7d662fce",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'keras.layers.core.dense.Dense'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n"
     ]
    }
   ],
   "source": [
    "with open('generated_code_dnn_TTVsOther.h', 'w') as f:\n",
    "\n",
    "\n",
    "    print (\"\"\"\n",
    "    // Select all (Ctrl+A), Copy (Ctrl+C), paste to an empty text file (Ctrl+V) and save that file as 'generated_code.h'\n",
    "    // \n",
    "    // Auto-generated header file. Assumes img to be a floating point array\n",
    "    // of 64 elements (corresponding to an 8x8 b&w image)\n",
    "\n",
    "    #include <math.h>\n",
    "\n",
    "    float norma;\n",
    "    \"\"\", file=f)\n",
    "    \n",
    "    iLayer_tmp=-1\n",
    "\n",
    "    for iLayer, layer in enumerate(dnnTT.model1.layers):\n",
    "      print(type(layer))\n",
    "#       if type(layer) is not Dropout and type(layer) is not Input :\n",
    "      if type(layer) is Dense:\n",
    "        kernel, bias = layer.get_weights()\n",
    "        iLayer_tmp+=1\n",
    "        print (\"inline float activation_%d (float x, float n) \" % iLayer_tmp, file=f);\n",
    "        activation =  layer.get_config()['activation'] \n",
    "        if activation == 'sigmoid':\n",
    "          print (\"{ return 1./(1 + exp(-x)); }\", file=f)\n",
    "        elif activation == 'tanh':\n",
    "          print (\"{ return tanh(x);}\")\n",
    "        elif activation == 'relu':\n",
    "          print (\"{ return x > 0 ? x : 0;}\", file=f)\n",
    "        elif activation == 'linear':\n",
    "          print (\"{ return x;}\")\n",
    "        elif activation == 'softmax':\n",
    "          print (\"{ return exp(x)/n;  }\", file=f)\n",
    "        else:\n",
    "          raise KeyError (\"Unexpected activation %s\"%activation)\n",
    "      \n",
    "    iLayer_tmp=-1\n",
    "\n",
    "    \n",
    "    print (\"\"\"\n",
    "    float guess_digit (const float *img, int flag)\n",
    "    {\n",
    "    \"\"\", file=f)\n",
    "\n",
    "    max_out = 100\n",
    "    for iLayer, layer in enumerate(dnnTT.model1.layers):\n",
    "      if type(layer) is Dense:\n",
    "        iLayer_tmp+=1\n",
    "        print (\"  // Declare the arrays in the stack\", file=f)\n",
    "        kernel, bias = layer.get_weights()\n",
    "        \n",
    "        print (type (dnnTT.means))\n",
    "        print (type(bias))\n",
    "        \n",
    "         \n",
    "        max_out = max(kernel.shape[1], max_out)\n",
    "        #print (\"  //\",bias.shape)\n",
    "        kernel_values = \"{%s}\"%(',\\n   '.join([\"{%s}\"%(','.join([\"%18.13f\"%x for x in row])) for row in kernel]))\n",
    "        bias_values   = \"{%s}\"% ( \",\".join([\"%18.13f\"%x for x in bias]))\n",
    "        print (\"  const float kernel_%d[%d][%d] = \\n  %s;\" % (iLayer_tmp, kernel.shape[0], kernel.shape[1],kernel_values), file=f)\n",
    "        print (\"  const float bias_%d[%d] = %s;\" % (iLayer_tmp, bias.shape[0], bias_values), file=f)\n",
    "      \n",
    "    print (\"  float buffer_in[%d];\" % max_out, file=f)\n",
    "    print (\"  float buffer_out[%d];\" % max_out, file=f)\n",
    "\n",
    "    print (\"  unsigned int i,j,c; \", file=f)\n",
    "    \n",
    "\n",
    "    mean_values = \"{%s}\"% ( \",\".join([\"%18.13f\"%x for x in dnnTT.means]))\n",
    "    sigma_values = \"{%s}\"% ( \",\".join([\"%18.13f\"%x for x in dnnTT.sigmas]))\n",
    "    \n",
    "\n",
    "    print (\" float mean[%d] = %s;\" % (max_out, mean_values), file=f)\n",
    "    print (\" float sigma[%d] = %s;\" % (max_out, sigma_values), file=f)\n",
    "\n",
    "    print (\"\\n\\n\\n\", file=f)\n",
    "    print (\"  // Load the input in the buffer\", file=f)\n",
    "    print (\"  for (c = 0; c < 64; ++c) \\n  buffer_in[c] = (img[c]-mean[c])/sigma[c];\", file=f)\n",
    "\n",
    "\n",
    "\n",
    "    iLayer_tmp=-1\n",
    "\n",
    "    for iLayer, layer in enumerate(dnnTT.model1.layers):\n",
    "      if type(layer) is Dense: \n",
    "        kernel, bias = layer.get_weights()\n",
    "        iLayer_tmp+=1\n",
    "\n",
    "        if iLayer != len(dnnTT.model1.layers)-1:\n",
    "            \n",
    "            \n",
    "            print ( \"  // Processing layer %i \" % iLayer_tmp , file=f)\n",
    "            print ( \"\"\"\n",
    "            for (c = 0; c < {n_out}; ++c ) \n",
    "              buffer_out[c] = bias_{iLayer_tmp}[c];\n",
    "\n",
    "            for (c = 0; c < {n_out}; ++c )\n",
    "              for (i = 0; i < {n_in}; ++i)\n",
    "                buffer_out[c] += buffer_in[i] * kernel_{iLayer_tmp}[i][c];\n",
    "\n",
    "            norma = 0;\n",
    "\n",
    "            // Prepares for next layer \n",
    "            for (c = 0; c < {n_out}; ++c )\n",
    "              buffer_in[c] = activation_{iLayer_tmp}(buffer_out[c], norma);\n",
    "\n",
    "\n",
    "            \"\"\".format (\n",
    "                n_in = kernel.shape[0],\n",
    "                n_out = kernel.shape[1],\n",
    "                iLayer_tmp = iLayer_tmp,\n",
    "            ), file=f)\n",
    "            \n",
    "        else:\n",
    "            \n",
    "            print ( \"  // Processing layer %i \" % iLayer_tmp , file=f)\n",
    "            print ( \"\"\"\n",
    "            for (c = 0; c < {n_out}; ++c ) \n",
    "              buffer_out[c] = bias_{iLayer_tmp}[c];\n",
    "\n",
    "            for (c = 0; c < {n_out}; ++c )\n",
    "              for (i = 0; i < {n_in}; ++i)\n",
    "                buffer_out[c] += buffer_in[i] * kernel_{iLayer_tmp}[i][c];\n",
    "\n",
    "            norma = 0;\n",
    "\n",
    "            for(c=0;c<{n_out};++c)\n",
    "              norma+=exp(buffer_out[c]);\n",
    "\n",
    "            // Prepares for next layer \n",
    "            for (c = 0; c < {n_out}; ++c )\n",
    "              buffer_in[c] = activation_{iLayer_tmp}(buffer_out[c], norma);\n",
    "\n",
    "\n",
    "            \"\"\".format (\n",
    "                n_in = kernel.shape[0],\n",
    "                n_out = kernel.shape[1],\n",
    "                iLayer_tmp = iLayer_tmp,\n",
    "            ), file=f)\n",
    "      \n",
    "    last_kernel, last_bias = dnnTT.model1.layers[-1].get_weights()\n",
    "    print (\"\"\"\n",
    "      //i = 0;\n",
    "      //for (c = 0; c < {n_out}; ++c)\n",
    "      //  if (buffer_in[c] > buffer_in[i])\n",
    "      //    i = c;\n",
    "      \n",
    "      if (flag==0) //vbf\n",
    "        return buffer_in[0];\n",
    "      else if (flag==1) //bkg\n",
    "        return buffer_in[1];\n",
    "         \n",
    "      \n",
    "      //return i;\n",
    "    \"\"\".format(n_out = last_kernel.shape[1]), file=f)\n",
    "\n",
    "    print (\"}\", file=f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "11d4d6ba-3206-4c98-89ec-24dffc5708cb",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}