bjet_exploration.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from keras.layers import Dense, Input, Conv2D, Conv2DTranspose, Lambda, Flatten, Reshape, MaxPooling2D\n",
    "from keras.models import Model\n",
    "from keras.losses import mean_squared_error, binary_crossentropy, kullback_leibler_divergence\n",
    "from keras.callbacks import EarlyStopping, TerminateOnNaN, ReduceLROnPlateau\n",
    "from keras import backend as K\n",
    "import tensorflow as tf\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import os, scipy\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "from phenom_read_in_data import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.set_option('display.max_columns', 75)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Getthe data\n",
    "df = csv_to_df('ttbarZ/')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "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>event_ID</th>\n",
       "      <th>process_ID</th>\n",
       "      <th>event_weight</th>\n",
       "      <th>MET</th>\n",
       "      <th>MET_Phi</th>\n",
       "      <th>obj1</th>\n",
       "      <th>E1</th>\n",
       "      <th>pt1</th>\n",
       "      <th>eta1</th>\n",
       "      <th>phi1</th>\n",
       "      <th>obj2</th>\n",
       "      <th>E2</th>\n",
       "      <th>pt2</th>\n",
       "      <th>eta2</th>\n",
       "      <th>phi2</th>\n",
       "      <th>obj3</th>\n",
       "      <th>E3</th>\n",
       "      <th>pt3</th>\n",
       "      <th>eta3</th>\n",
       "      <th>phi3</th>\n",
       "      <th>obj4</th>\n",
       "      <th>E4</th>\n",
       "      <th>pt4</th>\n",
       "      <th>eta4</th>\n",
       "      <th>phi4</th>\n",
       "      <th>obj5</th>\n",
       "      <th>E5</th>\n",
       "      <th>pt5</th>\n",
       "      <th>eta5</th>\n",
       "      <th>phi5</th>\n",
       "      <th>obj6</th>\n",
       "      <th>E6</th>\n",
       "      <th>pt6</th>\n",
       "      <th>eta6</th>\n",
       "      <th>phi6</th>\n",
       "      <th>obj7</th>\n",
       "      <th>E7</th>\n",
       "      <th>pt7</th>\n",
       "      <th>eta7</th>\n",
       "      <th>phi7</th>\n",
       "      <th>obj8</th>\n",
       "      <th>E8</th>\n",
       "      <th>pt8</th>\n",
       "      <th>eta8</th>\n",
       "      <th>phi8</th>\n",
       "      <th>obj9</th>\n",
       "      <th>E9</th>\n",
       "      <th>pt9</th>\n",
       "      <th>eta9</th>\n",
       "      <th>phi9</th>\n",
       "      <th>obj10</th>\n",
       "      <th>E10</th>\n",
       "      <th>pt10</th>\n",
       "      <th>eta10</th>\n",
       "      <th>phi10</th>\n",
       "      <th>obj11</th>\n",
       "      <th>E11</th>\n",
       "      <th>pt11</th>\n",
       "      <th>eta11</th>\n",
       "      <th>phi11</th>\n",
       "      <th>obj12</th>\n",
       "      <th>E12</th>\n",
       "      <th>pt12</th>\n",
       "      <th>eta12</th>\n",
       "      <th>phi12</th>\n",
       "      <th>obj13</th>\n",
       "      <th>E13</th>\n",
       "      <th>pt13</th>\n",
       "      <th>eta13</th>\n",
       "      <th>phi13</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>21</td>\n",
       "      <td>ttbarZ</td>\n",
       "      <td>1</td>\n",
       "      <td>43924.6</td>\n",
       "      <td>-0.38825</td>\n",
       "      <td>j</td>\n",
       "      <td>165476</td>\n",
       "      <td>165097</td>\n",
       "      <td>-0.0171131</td>\n",
       "      <td>-1.54448</td>\n",
       "      <td>j</td>\n",
       "      <td>163744</td>\n",
       "      <td>161051</td>\n",
       "      <td>-0.127864</td>\n",
       "      <td>1.89464</td>\n",
       "      <td>b</td>\n",
       "      <td>86228.2</td>\n",
       "      <td>78559.5</td>\n",
       "      <td>-0.393297</td>\n",
       "      <td>3.05707</td>\n",
       "      <td>j</td>\n",
       "      <td>68231.7</td>\n",
       "      <td>67934.1</td>\n",
       "      <td>-0.0262432</td>\n",
       "      <td>1.12407</td>\n",
       "      <td>j</td>\n",
       "      <td>40452</td>\n",
       "      <td>36900.9</td>\n",
       "      <td>-0.419573</td>\n",
       "      <td>-2.32306</td>\n",
       "      <td>j</td>\n",
       "      <td>36919.8</td>\n",
       "      <td>34203.1</td>\n",
       "      <td>0.303205</td>\n",
       "      <td>-0.614197</td>\n",
       "      <td>j</td>\n",
       "      <td>31126.3</td>\n",
       "      <td>24639.1</td>\n",
       "      <td>0.678827</td>\n",
       "      <td>0.396069</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>21</td>\n",
       "      <td>ttbarZ</td>\n",
       "      <td>1</td>\n",
       "      <td>45289</td>\n",
       "      <td>-3.09984</td>\n",
       "      <td>j</td>\n",
       "      <td>105746</td>\n",
       "      <td>95333.9</td>\n",
       "      <td>0.449145</td>\n",
       "      <td>1.79273</td>\n",
       "      <td>j</td>\n",
       "      <td>106214</td>\n",
       "      <td>92717.5</td>\n",
       "      <td>-0.52132</td>\n",
       "      <td>-0.624799</td>\n",
       "      <td>j</td>\n",
       "      <td>273538</td>\n",
       "      <td>63046.7</td>\n",
       "      <td>2.14564</td>\n",
       "      <td>0.166455</td>\n",
       "      <td>b</td>\n",
       "      <td>74094.3</td>\n",
       "      <td>53404</td>\n",
       "      <td>0.848482</td>\n",
       "      <td>-1.91912</td>\n",
       "      <td>j</td>\n",
       "      <td>37460.9</td>\n",
       "      <td>36739.6</td>\n",
       "      <td>0.183794</td>\n",
       "      <td>3.06706</td>\n",
       "      <td>b</td>\n",
       "      <td>45106.4</td>\n",
       "      <td>36630.5</td>\n",
       "      <td>-0.642508</td>\n",
       "      <td>-2.75669</td>\n",
       "      <td>j</td>\n",
       "      <td>72748.1</td>\n",
       "      <td>34023.2</td>\n",
       "      <td>-1.38809</td>\n",
       "      <td>0.632568</td>\n",
       "      <td>j</td>\n",
       "      <td>135412</td>\n",
       "      <td>29152.7</td>\n",
       "      <td>2.21505</td>\n",
       "      <td>2.39706</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>21</td>\n",
       "      <td>ttbarZ</td>\n",
       "      <td>1</td>\n",
       "      <td>16530.4</td>\n",
       "      <td>-3.09183</td>\n",
       "      <td>b</td>\n",
       "      <td>144211</td>\n",
       "      <td>137971</td>\n",
       "      <td>0.100321</td>\n",
       "      <td>-0.0772316</td>\n",
       "      <td>j</td>\n",
       "      <td>195882</td>\n",
       "      <td>87466</td>\n",
       "      <td>1.44049</td>\n",
       "      <td>2.31974</td>\n",
       "      <td>j</td>\n",
       "      <td>152950</td>\n",
       "      <td>65054.9</td>\n",
       "      <td>1.49544</td>\n",
       "      <td>-3.03641</td>\n",
       "      <td>b</td>\n",
       "      <td>89301.6</td>\n",
       "      <td>61292.3</td>\n",
       "      <td>0.915048</td>\n",
       "      <td>-1.8527</td>\n",
       "      <td>j</td>\n",
       "      <td>332144</td>\n",
       "      <td>44061.6</td>\n",
       "      <td>2.70825</td>\n",
       "      <td>0.615086</td>\n",
       "      <td>j</td>\n",
       "      <td>56815.5</td>\n",
       "      <td>35907.4</td>\n",
       "      <td>-1.03082</td>\n",
       "      <td>-1.96102</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>21</td>\n",
       "      <td>ttbarZ</td>\n",
       "      <td>1</td>\n",
       "      <td>18362.9</td>\n",
       "      <td>1.14058</td>\n",
       "      <td>b</td>\n",
       "      <td>370358</td>\n",
       "      <td>120835</td>\n",
       "      <td>1.78374</td>\n",
       "      <td>-0.757091</td>\n",
       "      <td>j</td>\n",
       "      <td>156700</td>\n",
       "      <td>107113</td>\n",
       "      <td>0.912321</td>\n",
       "      <td>2.38575</td>\n",
       "      <td>j</td>\n",
       "      <td>90473.5</td>\n",
       "      <td>79780.5</td>\n",
       "      <td>-0.504302</td>\n",
       "      <td>-2.71376</td>\n",
       "      <td>j</td>\n",
       "      <td>116668</td>\n",
       "      <td>72019.3</td>\n",
       "      <td>1.04829</td>\n",
       "      <td>-0.322966</td>\n",
       "      <td>b</td>\n",
       "      <td>277522</td>\n",
       "      <td>70059.8</td>\n",
       "      <td>2.05196</td>\n",
       "      <td>0.062623</td>\n",
       "      <td>j</td>\n",
       "      <td>487093</td>\n",
       "      <td>68214.4</td>\n",
       "      <td>2.65376</td>\n",
       "      <td>2.27298</td>\n",
       "      <td>j</td>\n",
       "      <td>148612</td>\n",
       "      <td>45663.6</td>\n",
       "      <td>1.84695</td>\n",
       "      <td>-2.93225</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>21</td>\n",
       "      <td>ttbarZ</td>\n",
       "      <td>1</td>\n",
       "      <td>232797</td>\n",
       "      <td>-2.74669</td>\n",
       "      <td>j</td>\n",
       "      <td>436613</td>\n",
       "      <td>432317</td>\n",
       "      <td>-0.110325</td>\n",
       "      <td>0.656242</td>\n",
       "      <td>b</td>\n",
       "      <td>185625</td>\n",
       "      <td>182595</td>\n",
       "      <td>-0.0154071</td>\n",
       "      <td>-2.69834</td>\n",
       "      <td>j</td>\n",
       "      <td>202016</td>\n",
       "      <td>134513</td>\n",
       "      <td>-0.949922</td>\n",
       "      <td>-0.73823</td>\n",
       "      <td>j</td>\n",
       "      <td>153539</td>\n",
       "      <td>71250.1</td>\n",
       "      <td>-1.40028</td>\n",
       "      <td>2.0433</td>\n",
       "      <td>j</td>\n",
       "      <td>80299.9</td>\n",
       "      <td>52919.8</td>\n",
       "      <td>0.974133</td>\n",
       "      <td>-2.44932</td>\n",
       "      <td>j</td>\n",
       "      <td>162664</td>\n",
       "      <td>46091.9</td>\n",
       "      <td>-1.93104</td>\n",
       "      <td>-2.30425</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\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",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5869</td>\n",
       "      <td>21</td>\n",
       "      <td>ttbarZ</td>\n",
       "      <td>1</td>\n",
       "      <td>17569.3</td>\n",
       "      <td>3.08905</td>\n",
       "      <td>j</td>\n",
       "      <td>223425</td>\n",
       "      <td>161759</td>\n",
       "      <td>-0.840404</td>\n",
       "      <td>2.52008</td>\n",
       "      <td>b</td>\n",
       "      <td>212922</td>\n",
       "      <td>150108</td>\n",
       "      <td>0.871115</td>\n",
       "      <td>-1.46624</td>\n",
       "      <td>j</td>\n",
       "      <td>107559</td>\n",
       "      <td>104874</td>\n",
       "      <td>-0.09206</td>\n",
       "      <td>1.86163</td>\n",
       "      <td>j</td>\n",
       "      <td>521781</td>\n",
       "      <td>87172.3</td>\n",
       "      <td>2.47494</td>\n",
       "      <td>-0.528776</td>\n",
       "      <td>j</td>\n",
       "      <td>110667</td>\n",
       "      <td>84358.7</td>\n",
       "      <td>0.754226</td>\n",
       "      <td>0.234181</td>\n",
       "      <td>j</td>\n",
       "      <td>319253</td>\n",
       "      <td>52431.5</td>\n",
       "      <td>-2.49247</td>\n",
       "      <td>2.77751</td>\n",
       "      <td>j</td>\n",
       "      <td>43875.6</td>\n",
       "      <td>39472.2</td>\n",
       "      <td>-0.46807</td>\n",
       "      <td>-1.35213</td>\n",
       "      <td>j</td>\n",
       "      <td>159489</td>\n",
       "      <td>35116</td>\n",
       "      <td>2.19302</td>\n",
       "      <td>1.42717</td>\n",
       "      <td>j</td>\n",
       "      <td>76563.8</td>\n",
       "      <td>26379.1</td>\n",
       "      <td>1.72649</td>\n",
       "      <td>-0.62969</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5870</td>\n",
       "      <td>21</td>\n",
       "      <td>ttbarZ</td>\n",
       "      <td>1</td>\n",
       "      <td>70402.4</td>\n",
       "      <td>-1.12592</td>\n",
       "      <td>b</td>\n",
       "      <td>340938</td>\n",
       "      <td>86263.4</td>\n",
       "      <td>-2.05015</td>\n",
       "      <td>-0.018614</td>\n",
       "      <td>b</td>\n",
       "      <td>80266</td>\n",
       "      <td>79630.8</td>\n",
       "      <td>0.0892557</td>\n",
       "      <td>1.44809</td>\n",
       "      <td>j</td>\n",
       "      <td>98983.4</td>\n",
       "      <td>56366.9</td>\n",
       "      <td>-1.1629</td>\n",
       "      <td>2.92059</td>\n",
       "      <td>j</td>\n",
       "      <td>44193.9</td>\n",
       "      <td>44005.6</td>\n",
       "      <td>-0.0908837</td>\n",
       "      <td>-2.35942</td>\n",
       "      <td>j</td>\n",
       "      <td>46323.3</td>\n",
       "      <td>42138.8</td>\n",
       "      <td>0.404271</td>\n",
       "      <td>2.09979</td>\n",
       "      <td>j</td>\n",
       "      <td>149315</td>\n",
       "      <td>27061.5</td>\n",
       "      <td>-2.39208</td>\n",
       "      <td>-1.38726</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5871</td>\n",
       "      <td>21</td>\n",
       "      <td>ttbarZ</td>\n",
       "      <td>1</td>\n",
       "      <td>28553.9</td>\n",
       "      <td>0.862966</td>\n",
       "      <td>j</td>\n",
       "      <td>116154</td>\n",
       "      <td>114311</td>\n",
       "      <td>0.0884322</td>\n",
       "      <td>-2.74727</td>\n",
       "      <td>j</td>\n",
       "      <td>193623</td>\n",
       "      <td>94304.2</td>\n",
       "      <td>1.3457</td>\n",
       "      <td>1.29583</td>\n",
       "      <td>b</td>\n",
       "      <td>164252</td>\n",
       "      <td>66516.6</td>\n",
       "      <td>1.54902</td>\n",
       "      <td>0.167349</td>\n",
       "      <td>j</td>\n",
       "      <td>332013</td>\n",
       "      <td>40321.4</td>\n",
       "      <td>-2.79728</td>\n",
       "      <td>0.600107</td>\n",
       "      <td>j</td>\n",
       "      <td>37773.4</td>\n",
       "      <td>29133.9</td>\n",
       "      <td>0.709713</td>\n",
       "      <td>-1.52097</td>\n",
       "      <td>j</td>\n",
       "      <td>73094.2</td>\n",
       "      <td>27742.1</td>\n",
       "      <td>-1.62051</td>\n",
       "      <td>-1.09525</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5872</td>\n",
       "      <td>21</td>\n",
       "      <td>ttbarZ</td>\n",
       "      <td>1</td>\n",
       "      <td>183633</td>\n",
       "      <td>1.98713</td>\n",
       "      <td>j</td>\n",
       "      <td>350312</td>\n",
       "      <td>178245</td>\n",
       "      <td>1.29249</td>\n",
       "      <td>0.477881</td>\n",
       "      <td>j</td>\n",
       "      <td>112870</td>\n",
       "      <td>109878</td>\n",
       "      <td>0.218633</td>\n",
       "      <td>-2.54255</td>\n",
       "      <td>j</td>\n",
       "      <td>81252</td>\n",
       "      <td>74631.8</td>\n",
       "      <td>-0.354966</td>\n",
       "      <td>-1.88837</td>\n",
       "      <td>j</td>\n",
       "      <td>167830</td>\n",
       "      <td>71985.2</td>\n",
       "      <td>-1.48892</td>\n",
       "      <td>-1.91709</td>\n",
       "      <td>j</td>\n",
       "      <td>325596</td>\n",
       "      <td>53747.1</td>\n",
       "      <td>-2.48624</td>\n",
       "      <td>-0.555373</td>\n",
       "      <td>b</td>\n",
       "      <td>42017.8</td>\n",
       "      <td>36963</td>\n",
       "      <td>0.506639</td>\n",
       "      <td>-0.452317</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5873</td>\n",
       "      <td>21</td>\n",
       "      <td>ttbarZ</td>\n",
       "      <td>1</td>\n",
       "      <td>34380</td>\n",
       "      <td>0.550302</td>\n",
       "      <td>j</td>\n",
       "      <td>402447</td>\n",
       "      <td>97361.7</td>\n",
       "      <td>2.09636</td>\n",
       "      <td>-2.32422</td>\n",
       "      <td>j</td>\n",
       "      <td>147398</td>\n",
       "      <td>89731.8</td>\n",
       "      <td>-1.07175</td>\n",
       "      <td>2.0308</td>\n",
       "      <td>j</td>\n",
       "      <td>71077.9</td>\n",
       "      <td>57486.4</td>\n",
       "      <td>0.669019</td>\n",
       "      <td>-0.88476</td>\n",
       "      <td>j</td>\n",
       "      <td>42784.2</td>\n",
       "      <td>36793.7</td>\n",
       "      <td>0.520857</td>\n",
       "      <td>0.957847</td>\n",
       "      <td>j</td>\n",
       "      <td>262457</td>\n",
       "      <td>35520.8</td>\n",
       "      <td>2.68774</td>\n",
       "      <td>0.735922</td>\n",
       "      <td>j</td>\n",
       "      <td>90765.6</td>\n",
       "      <td>29298</td>\n",
       "      <td>1.79481</td>\n",
       "      <td>0.232788</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5874 rows × 70 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     event_ID process_ID event_weight      MET   MET_Phi obj1      E1  \\\n",
       "0          21     ttbarZ            1  43924.6  -0.38825    j  165476   \n",
       "1          21     ttbarZ            1    45289  -3.09984    j  105746   \n",
       "2          21     ttbarZ            1  16530.4  -3.09183    b  144211   \n",
       "3          21     ttbarZ            1  18362.9   1.14058    b  370358   \n",
       "4          21     ttbarZ            1   232797  -2.74669    j  436613   \n",
       "...       ...        ...          ...      ...       ...  ...     ...   \n",
       "5869       21     ttbarZ            1  17569.3   3.08905    j  223425   \n",
       "5870       21     ttbarZ            1  70402.4  -1.12592    b  340938   \n",
       "5871       21     ttbarZ            1  28553.9  0.862966    j  116154   \n",
       "5872       21     ttbarZ            1   183633   1.98713    j  350312   \n",
       "5873       21     ttbarZ            1    34380  0.550302    j  402447   \n",
       "\n",
       "          pt1        eta1        phi1 obj2      E2      pt2        eta2  \\\n",
       "0      165097  -0.0171131    -1.54448    j  163744   161051   -0.127864   \n",
       "1     95333.9    0.449145     1.79273    j  106214  92717.5    -0.52132   \n",
       "2      137971    0.100321  -0.0772316    j  195882    87466     1.44049   \n",
       "3      120835     1.78374   -0.757091    j  156700   107113    0.912321   \n",
       "4      432317   -0.110325    0.656242    b  185625   182595  -0.0154071   \n",
       "...       ...         ...         ...  ...     ...      ...         ...   \n",
       "5869   161759   -0.840404     2.52008    b  212922   150108    0.871115   \n",
       "5870  86263.4    -2.05015   -0.018614    b   80266  79630.8   0.0892557   \n",
       "5871   114311   0.0884322    -2.74727    j  193623  94304.2      1.3457   \n",
       "5872   178245     1.29249    0.477881    j  112870   109878    0.218633   \n",
       "5873  97361.7     2.09636    -2.32422    j  147398  89731.8    -1.07175   \n",
       "\n",
       "           phi2 obj3       E3      pt3       eta3      phi3 obj4       E4  \\\n",
       "0       1.89464    b  86228.2  78559.5  -0.393297   3.05707    j  68231.7   \n",
       "1     -0.624799    j   273538  63046.7    2.14564  0.166455    b  74094.3   \n",
       "2       2.31974    j   152950  65054.9    1.49544  -3.03641    b  89301.6   \n",
       "3       2.38575    j  90473.5  79780.5  -0.504302  -2.71376    j   116668   \n",
       "4      -2.69834    j   202016   134513  -0.949922  -0.73823    j   153539   \n",
       "...         ...  ...      ...      ...        ...       ...  ...      ...   \n",
       "5869   -1.46624    j   107559   104874   -0.09206   1.86163    j   521781   \n",
       "5870    1.44809    j  98983.4  56366.9    -1.1629   2.92059    j  44193.9   \n",
       "5871    1.29583    b   164252  66516.6    1.54902  0.167349    j   332013   \n",
       "5872   -2.54255    j    81252  74631.8  -0.354966  -1.88837    j   167830   \n",
       "5873     2.0308    j  71077.9  57486.4   0.669019  -0.88476    j  42784.2   \n",
       "\n",
       "          pt4        eta4       phi4 obj5       E5      pt5       eta5  \\\n",
       "0     67934.1  -0.0262432    1.12407    j    40452  36900.9  -0.419573   \n",
       "1       53404    0.848482   -1.91912    j  37460.9  36739.6   0.183794   \n",
       "2     61292.3    0.915048    -1.8527    j   332144  44061.6    2.70825   \n",
       "3     72019.3     1.04829  -0.322966    b   277522  70059.8    2.05196   \n",
       "4     71250.1    -1.40028     2.0433    j  80299.9  52919.8   0.974133   \n",
       "...       ...         ...        ...  ...      ...      ...        ...   \n",
       "5869  87172.3     2.47494  -0.528776    j   110667  84358.7   0.754226   \n",
       "5870  44005.6  -0.0908837   -2.35942    j  46323.3  42138.8   0.404271   \n",
       "5871  40321.4    -2.79728   0.600107    j  37773.4  29133.9   0.709713   \n",
       "5872  71985.2    -1.48892   -1.91709    j   325596  53747.1   -2.48624   \n",
       "5873  36793.7    0.520857   0.957847    j   262457  35520.8    2.68774   \n",
       "\n",
       "           phi5 obj6       E6      pt6       eta6       phi6 obj7       E7  \\\n",
       "0      -2.32306    j  36919.8  34203.1   0.303205  -0.614197    j  31126.3   \n",
       "1       3.06706    b  45106.4  36630.5  -0.642508   -2.75669    j  72748.1   \n",
       "2      0.615086    j  56815.5  35907.4   -1.03082   -1.96102  NaN      NaN   \n",
       "3      0.062623    j   487093  68214.4    2.65376    2.27298    j   148612   \n",
       "4      -2.44932    j   162664  46091.9   -1.93104   -2.30425  NaN      NaN   \n",
       "...         ...  ...      ...      ...        ...        ...  ...      ...   \n",
       "5869   0.234181    j   319253  52431.5   -2.49247    2.77751    j  43875.6   \n",
       "5870    2.09979    j   149315  27061.5   -2.39208   -1.38726  NaN      NaN   \n",
       "5871   -1.52097    j  73094.2  27742.1   -1.62051   -1.09525  NaN      NaN   \n",
       "5872  -0.555373    b  42017.8    36963   0.506639  -0.452317  NaN      NaN   \n",
       "5873   0.735922    j  90765.6    29298    1.79481   0.232788  NaN      NaN   \n",
       "\n",
       "          pt7      eta7      phi7 obj8      E8      pt8     eta8     phi8  \\\n",
       "0     24639.1  0.678827  0.396069  NaN     NaN      NaN      NaN      NaN   \n",
       "1     34023.2  -1.38809  0.632568    j  135412  29152.7  2.21505  2.39706   \n",
       "2         NaN       NaN       NaN  NaN     NaN      NaN      NaN      NaN   \n",
       "3     45663.6   1.84695  -2.93225  NaN     NaN      NaN      NaN      NaN   \n",
       "4         NaN       NaN       NaN  NaN     NaN      NaN      NaN      NaN   \n",
       "...       ...       ...       ...  ...     ...      ...      ...      ...   \n",
       "5869  39472.2  -0.46807  -1.35213    j  159489    35116  2.19302  1.42717   \n",
       "5870      NaN       NaN       NaN  NaN     NaN      NaN      NaN      NaN   \n",
       "5871      NaN       NaN       NaN  NaN     NaN      NaN      NaN      NaN   \n",
       "5872      NaN       NaN       NaN  NaN     NaN      NaN      NaN      NaN   \n",
       "5873      NaN       NaN       NaN  NaN     NaN      NaN      NaN      NaN   \n",
       "\n",
       "     obj9       E9      pt9     eta9      phi9 obj10  E10 pt10 eta10 phi10  \\\n",
       "0     NaN      NaN      NaN      NaN       NaN   NaN  NaN  NaN   NaN   NaN   \n",
       "1     NaN      NaN      NaN      NaN       NaN   NaN  NaN  NaN   NaN   NaN   \n",
       "2     NaN      NaN      NaN      NaN       NaN   NaN  NaN  NaN   NaN   NaN   \n",
       "3     NaN      NaN      NaN      NaN       NaN   NaN  NaN  NaN   NaN   NaN   \n",
       "4     NaN      NaN      NaN      NaN       NaN   NaN  NaN  NaN   NaN   NaN   \n",
       "...   ...      ...      ...      ...       ...   ...  ...  ...   ...   ...   \n",
       "5869    j  76563.8  26379.1  1.72649  -0.62969   NaN  NaN  NaN   NaN   NaN   \n",
       "5870  NaN      NaN      NaN      NaN       NaN   NaN  NaN  NaN   NaN   NaN   \n",
       "5871  NaN      NaN      NaN      NaN       NaN   NaN  NaN  NaN   NaN   NaN   \n",
       "5872  NaN      NaN      NaN      NaN       NaN   NaN  NaN  NaN   NaN   NaN   \n",
       "5873  NaN      NaN      NaN      NaN       NaN   NaN  NaN  NaN   NaN   NaN   \n",
       "\n",
       "     obj11  E11 pt11 eta11 phi11 obj12  E12 pt12 eta12 phi12 obj13  E13 pt13  \\\n",
       "0      NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   \n",
       "1      NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   \n",
       "2      NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   \n",
       "3      NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   \n",
       "4      NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   \n",
       "...    ...  ...  ...   ...   ...   ...  ...  ...   ...   ...   ...  ...  ...   \n",
       "5869   NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   \n",
       "5870   NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   \n",
       "5871   NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   \n",
       "5872   NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   \n",
       "5873   NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   NaN   NaN   NaN  NaN  NaN   \n",
       "\n",
       "     eta13 phi13  \n",
       "0      NaN   NaN  \n",
       "1      NaN   NaN  \n",
       "2      NaN   NaN  \n",
       "3      NaN   NaN  \n",
       "4      NaN   NaN  \n",
       "...    ...   ...  \n",
       "5869   NaN   NaN  \n",
       "5870   NaN   NaN  \n",
       "5871   NaN   NaN  \n",
       "5872   NaN   NaN  \n",
       "5873   NaN   NaN  \n",
       "\n",
       "[5874 rows x 70 columns]"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "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>event_ID</th>\n",
       "      <th>process_ID</th>\n",
       "      <th>event_weight</th>\n",
       "      <th>MET</th>\n",
       "      <th>MET_Phi</th>\n",
       "      <th>obj1</th>\n",
       "      <th>E1</th>\n",
       "      <th>pt1</th>\n",
       "      <th>eta1</th>\n",
       "      <th>phi1</th>\n",
       "      <th>obj2</th>\n",
       "      <th>E2</th>\n",
       "      <th>pt2</th>\n",
       "      <th>eta2</th>\n",
       "      <th>phi2</th>\n",
       "      <th>obj3</th>\n",
       "      <th>E3</th>\n",
       "      <th>pt3</th>\n",
       "      <th>eta3</th>\n",
       "      <th>phi3</th>\n",
       "      <th>obj4</th>\n",
       "      <th>E4</th>\n",
       "      <th>pt4</th>\n",
       "      <th>eta4</th>\n",
       "      <th>phi4</th>\n",
       "      <th>obj5</th>\n",
       "      <th>E5</th>\n",
       "      <th>pt5</th>\n",
       "      <th>eta5</th>\n",
       "      <th>phi5</th>\n",
       "      <th>obj6</th>\n",
       "      <th>E6</th>\n",
       "      <th>pt6</th>\n",
       "      <th>eta6</th>\n",
       "      <th>phi6</th>\n",
       "      <th>obj7</th>\n",
       "      <th>E7</th>\n",
       "      <th>pt7</th>\n",
       "      <th>eta7</th>\n",
       "      <th>phi7</th>\n",
       "      <th>obj8</th>\n",
       "      <th>E8</th>\n",
       "      <th>pt8</th>\n",
       "      <th>eta8</th>\n",
       "      <th>phi8</th>\n",
       "      <th>obj9</th>\n",
       "      <th>E9</th>\n",
       "      <th>pt9</th>\n",
       "      <th>eta9</th>\n",
       "      <th>phi9</th>\n",
       "      <th>obj10</th>\n",
       "      <th>E10</th>\n",
       "      <th>pt10</th>\n",
       "      <th>eta10</th>\n",
       "      <th>phi10</th>\n",
       "      <th>obj11</th>\n",
       "      <th>E11</th>\n",
       "      <th>pt11</th>\n",
       "      <th>eta11</th>\n",
       "      <th>phi11</th>\n",
       "      <th>obj12</th>\n",
       "      <th>E12</th>\n",
       "      <th>pt12</th>\n",
       "      <th>eta12</th>\n",
       "      <th>phi12</th>\n",
       "      <th>obj13</th>\n",
       "      <th>E13</th>\n",
       "      <th>pt13</th>\n",
       "      <th>eta13</th>\n",
       "      <th>phi13</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>count</td>\n",
       "      <td>5874</td>\n",
       "      <td>5874</td>\n",
       "      <td>5874</td>\n",
       "      <td>5874</td>\n",
       "      <td>5874</td>\n",
       "      <td>5874</td>\n",
       "      <td>5874</td>\n",
       "      <td>5874</td>\n",
       "      <td>5874</td>\n",
       "      <td>5874</td>\n",
       "      <td>5871</td>\n",
       "      <td>5871</td>\n",
       "      <td>5871</td>\n",
       "      <td>5871</td>\n",
       "      <td>5871</td>\n",
       "      <td>5825</td>\n",
       "      <td>5825</td>\n",
       "      <td>5825</td>\n",
       "      <td>5825</td>\n",
       "      <td>5825</td>\n",
       "      <td>5583</td>\n",
       "      <td>5583</td>\n",
       "      <td>5583</td>\n",
       "      <td>5583</td>\n",
       "      <td>5583</td>\n",
       "      <td>4970</td>\n",
       "      <td>4970</td>\n",
       "      <td>4970</td>\n",
       "      <td>4970</td>\n",
       "      <td>4970</td>\n",
       "      <td>3792</td>\n",
       "      <td>3792</td>\n",
       "      <td>3792</td>\n",
       "      <td>3792</td>\n",
       "      <td>3792</td>\n",
       "      <td>2339</td>\n",
       "      <td>2339</td>\n",
       "      <td>2339</td>\n",
       "      <td>2339</td>\n",
       "      <td>2339</td>\n",
       "      <td>1158</td>\n",
       "      <td>1158</td>\n",
       "      <td>1158</td>\n",
       "      <td>1158</td>\n",
       "      <td>1158</td>\n",
       "      <td>434</td>\n",
       "      <td>434</td>\n",
       "      <td>434</td>\n",
       "      <td>434</td>\n",
       "      <td>434</td>\n",
       "      <td>139</td>\n",
       "      <td>139</td>\n",
       "      <td>139</td>\n",
       "      <td>139</td>\n",
       "      <td>139</td>\n",
       "      <td>35</td>\n",
       "      <td>35</td>\n",
       "      <td>35</td>\n",
       "      <td>35</td>\n",
       "      <td>35</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>unique</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>5859</td>\n",
       "      <td>5847</td>\n",
       "      <td>2</td>\n",
       "      <td>5843</td>\n",
       "      <td>5818</td>\n",
       "      <td>5859</td>\n",
       "      <td>5856</td>\n",
       "      <td>6</td>\n",
       "      <td>5831</td>\n",
       "      <td>5826</td>\n",
       "      <td>5859</td>\n",
       "      <td>5839</td>\n",
       "      <td>7</td>\n",
       "      <td>5799</td>\n",
       "      <td>5805</td>\n",
       "      <td>5804</td>\n",
       "      <td>5805</td>\n",
       "      <td>7</td>\n",
       "      <td>5561</td>\n",
       "      <td>5562</td>\n",
       "      <td>5564</td>\n",
       "      <td>5570</td>\n",
       "      <td>7</td>\n",
       "      <td>4950</td>\n",
       "      <td>4953</td>\n",
       "      <td>4955</td>\n",
       "      <td>4948</td>\n",
       "      <td>7</td>\n",
       "      <td>3788</td>\n",
       "      <td>3781</td>\n",
       "      <td>3784</td>\n",
       "      <td>3788</td>\n",
       "      <td>7</td>\n",
       "      <td>2331</td>\n",
       "      <td>2322</td>\n",
       "      <td>2335</td>\n",
       "      <td>2336</td>\n",
       "      <td>7</td>\n",
       "      <td>1155</td>\n",
       "      <td>1156</td>\n",
       "      <td>1157</td>\n",
       "      <td>1157</td>\n",
       "      <td>7</td>\n",
       "      <td>434</td>\n",
       "      <td>434</td>\n",
       "      <td>434</td>\n",
       "      <td>434</td>\n",
       "      <td>7</td>\n",
       "      <td>139</td>\n",
       "      <td>139</td>\n",
       "      <td>139</td>\n",
       "      <td>139</td>\n",
       "      <td>6</td>\n",
       "      <td>35</td>\n",
       "      <td>35</td>\n",
       "      <td>35</td>\n",
       "      <td>35</td>\n",
       "      <td>3</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>top</td>\n",
       "      <td>21</td>\n",
       "      <td>ttbarZ</td>\n",
       "      <td>1</td>\n",
       "      <td>105323</td>\n",
       "      <td>-1.15702</td>\n",
       "      <td>j</td>\n",
       "      <td>101480</td>\n",
       "      <td>137959</td>\n",
       "      <td>-1.60639</td>\n",
       "      <td>-1.7008</td>\n",
       "      <td>j</td>\n",
       "      <td>140788</td>\n",
       "      <td>107334</td>\n",
       "      <td>1.94753</td>\n",
       "      <td>-1.24283</td>\n",
       "      <td>j</td>\n",
       "      <td>57119.6</td>\n",
       "      <td>113845</td>\n",
       "      <td>-1.27353</td>\n",
       "      <td>-1.33961</td>\n",
       "      <td>j</td>\n",
       "      <td>105084</td>\n",
       "      <td>29668.1</td>\n",
       "      <td>-1.81589</td>\n",
       "      <td>-2.82383</td>\n",
       "      <td>j</td>\n",
       "      <td>126667</td>\n",
       "      <td>27186.3</td>\n",
       "      <td>2.28923</td>\n",
       "      <td>2.15548</td>\n",
       "      <td>j</td>\n",
       "      <td>115633</td>\n",
       "      <td>33471.8</td>\n",
       "      <td>-1.59634</td>\n",
       "      <td>2.13339</td>\n",
       "      <td>j</td>\n",
       "      <td>34833.2</td>\n",
       "      <td>30209.8</td>\n",
       "      <td>-2.04241</td>\n",
       "      <td>3.02596</td>\n",
       "      <td>j</td>\n",
       "      <td>30173.3</td>\n",
       "      <td>23922.7</td>\n",
       "      <td>-1.59418</td>\n",
       "      <td>-2.30007</td>\n",
       "      <td>j</td>\n",
       "      <td>56586.7</td>\n",
       "      <td>52792.5</td>\n",
       "      <td>-1.92779</td>\n",
       "      <td>-1.66469</td>\n",
       "      <td>j</td>\n",
       "      <td>96573.5</td>\n",
       "      <td>35278.6</td>\n",
       "      <td>-0.131001</td>\n",
       "      <td>2.80453</td>\n",
       "      <td>j</td>\n",
       "      <td>46720.3</td>\n",
       "      <td>29509.5</td>\n",
       "      <td>-0.715826</td>\n",
       "      <td>1.89091</td>\n",
       "      <td>j</td>\n",
       "      <td>76906.5</td>\n",
       "      <td>27271.7</td>\n",
       "      <td>1.70133</td>\n",
       "      <td>1.26894</td>\n",
       "      <td>j</td>\n",
       "      <td>42197.8</td>\n",
       "      <td>35040.1</td>\n",
       "      <td>0.23484</td>\n",
       "      <td>0.640652</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>freq</td>\n",
       "      <td>5874</td>\n",
       "      <td>5874</td>\n",
       "      <td>5874</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3967</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3986</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>4220</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>4129</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3644</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2726</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1678</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>770</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>300</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>98</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>27</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       event_ID process_ID event_weight     MET   MET_Phi  obj1      E1  \\\n",
       "count      5874       5874         5874    5874      5874  5874    5874   \n",
       "unique        1          1            1    5859      5847     2    5843   \n",
       "top          21     ttbarZ            1  105323  -1.15702     j  101480   \n",
       "freq       5874       5874         5874       2         2  3967       2   \n",
       "\n",
       "           pt1      eta1     phi1  obj2      E2     pt2     eta2      phi2  \\\n",
       "count     5874      5874     5874  5871    5871    5871     5871      5871   \n",
       "unique    5818      5859     5856     6    5831    5826     5859      5839   \n",
       "top     137959  -1.60639  -1.7008     j  140788  107334  1.94753  -1.24283   \n",
       "freq         3         2        2  3986       2       2        2         2   \n",
       "\n",
       "        obj3       E3     pt3      eta3      phi3  obj4      E4      pt4  \\\n",
       "count   5825     5825    5825      5825      5825  5583    5583     5583   \n",
       "unique     7     5799    5805      5804      5805     7    5561     5562   \n",
       "top        j  57119.6  113845  -1.27353  -1.33961     j  105084  29668.1   \n",
       "freq    4220        2       2         2         2  4129       2        2   \n",
       "\n",
       "            eta4      phi4  obj5      E5      pt5     eta5     phi5  obj6  \\\n",
       "count       5583      5583  4970    4970     4970     4970     4970  3792   \n",
       "unique      5564      5570     7    4950     4953     4955     4948     7   \n",
       "top     -1.81589  -2.82383     j  126667  27186.3  2.28923  2.15548     j   \n",
       "freq           2         2  3644       2        2        2        2  2726   \n",
       "\n",
       "            E6      pt6      eta6     phi6  obj7       E7      pt7      eta7  \\\n",
       "count     3792     3792      3792     3792  2339     2339     2339      2339   \n",
       "unique    3788     3781      3784     3788     7     2331     2322      2335   \n",
       "top     115633  33471.8  -1.59634  2.13339     j  34833.2  30209.8  -2.04241   \n",
       "freq         2        2         2        2  1678        2        2         2   \n",
       "\n",
       "           phi7  obj8       E8      pt8      eta8      phi8 obj9       E9  \\\n",
       "count      2339  1158     1158     1158      1158      1158  434      434   \n",
       "unique     2336     7     1155     1156      1157      1157    7      434   \n",
       "top     3.02596     j  30173.3  23922.7  -1.59418  -2.30007    j  56586.7   \n",
       "freq          2   770        2        2         2         2  300        1   \n",
       "\n",
       "            pt9      eta9      phi9 obj10      E10     pt10      eta10  \\\n",
       "count       434       434       434   139      139      139        139   \n",
       "unique      434       434       434     7      139      139        139   \n",
       "top     52792.5  -1.92779  -1.66469     j  96573.5  35278.6  -0.131001   \n",
       "freq          1         1         1    98        1        1          1   \n",
       "\n",
       "          phi10 obj11      E11     pt11      eta11    phi11 obj12      E12  \\\n",
       "count       139    35       35       35         35       35    10       10   \n",
       "unique      139     6       35       35         35       35     3       10   \n",
       "top     2.80453     j  46720.3  29509.5  -0.715826  1.89091     j  76906.5   \n",
       "freq          1    27        1        1          1        1     8        1   \n",
       "\n",
       "           pt12    eta12    phi12 obj13      E13     pt13    eta13     phi13  \n",
       "count        10       10       10     2        2        2        2         2  \n",
       "unique       10       10       10     1        2        2        2         2  \n",
       "top     27271.7  1.70133  1.26894     j  42197.8  35040.1  0.23484  0.640652  \n",
       "freq          1        1        1     2        1        1        1         1  "
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "#If the number of max objects is prime, another column needs to be added to mitigate "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['obj14']= -999\n",
    "df['E14'] = -999\n",
    "df['pt14'] = -999\n",
    "df['eta14'] = -999\n",
    "df['phi14'] = -999"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.fillna(-999)\n",
    "variables = [entry for entry in df.columns if entry[0] == 'E'] + [entry for entry in df.columns if entry[0:2] == 'pt'] + [entry for entry in df.columns if entry[0:2] == 'et'] + [entry for entry in df.columns if entry[0:2] == 'ph']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "objects = []\n",
    "for i in df.columns:\n",
    "    if i[:3] == 'obj':\n",
    "        objects.append(i)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "b_objects1 = df.loc[df['obj1'] == 'b', df.columns[6:10]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>E1</th>\n",
       "      <th>pt1</th>\n",
       "      <th>eta1</th>\n",
       "      <th>phi1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>144211</td>\n",
       "      <td>137971</td>\n",
       "      <td>0.100321</td>\n",
       "      <td>-0.0772316</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>370358</td>\n",
       "      <td>120835</td>\n",
       "      <td>1.78374</td>\n",
       "      <td>-0.757091</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>148059</td>\n",
       "      <td>146337</td>\n",
       "      <td>-0.126395</td>\n",
       "      <td>-2.48672</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>209567</td>\n",
       "      <td>107772</td>\n",
       "      <td>-1.27924</td>\n",
       "      <td>1.69015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>188912</td>\n",
       "      <td>124490</td>\n",
       "      <td>-0.97313</td>\n",
       "      <td>2.38879</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5860</td>\n",
       "      <td>105653</td>\n",
       "      <td>103890</td>\n",
       "      <td>0.00173553</td>\n",
       "      <td>-1.41436</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5862</td>\n",
       "      <td>324066</td>\n",
       "      <td>200746</td>\n",
       "      <td>1.05563</td>\n",
       "      <td>-0.432939</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5865</td>\n",
       "      <td>225909</td>\n",
       "      <td>206183</td>\n",
       "      <td>-0.429249</td>\n",
       "      <td>2.96939</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5866</td>\n",
       "      <td>269968</td>\n",
       "      <td>264990</td>\n",
       "      <td>-0.160639</td>\n",
       "      <td>-2.5178</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5870</td>\n",
       "      <td>340938</td>\n",
       "      <td>86263.4</td>\n",
       "      <td>-2.05015</td>\n",
       "      <td>-0.018614</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1907 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          E1      pt1        eta1        phi1\n",
       "2     144211   137971    0.100321  -0.0772316\n",
       "3     370358   120835     1.78374   -0.757091\n",
       "14    148059   146337   -0.126395    -2.48672\n",
       "15    209567   107772    -1.27924     1.69015\n",
       "17    188912   124490    -0.97313     2.38879\n",
       "...      ...      ...         ...         ...\n",
       "5860  105653   103890  0.00173553    -1.41436\n",
       "5862  324066   200746     1.05563   -0.432939\n",
       "5865  225909   206183   -0.429249     2.96939\n",
       "5866  269968   264990   -0.160639     -2.5178\n",
       "5870  340938  86263.4    -2.05015   -0.018614\n",
       "\n",
       "[1907 rows x 4 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "b_objects1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "b_objects2 = df.loc[df['obj2'] == 'b', df.columns[11:15]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/joe/anaconda3/lib/python3.7/site-packages/pandas/core/ops/__init__.py:1115: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison\n",
      "  result = method(y)\n"
     ]
    }
   ],
   "source": [
    "b_objects2 = df.loc[df['obj2'] == 'b', df.columns[11:15]]\n",
    "b_objects3 = df.loc[df['obj3'] == 'b', df.columns[16:20]]\n",
    "b_objects4 = df.loc[df['obj4'] == 'b', df.columns[21:25]]\n",
    "b_objects5 = df.loc[df['obj5'] == 'b', df.columns[26:30]]\n",
    "b_objects6 = df.loc[df['obj6'] == 'b', df.columns[31:35]]\n",
    "b_objects7 = df.loc[df['obj7'] == 'b', df.columns[36:40]]\n",
    "b_objects8 = df.loc[df['obj8'] == 'b', df.columns[41:45]]\n",
    "b_objects9 = df.loc[df['obj9'] == 'b', df.columns[46:50]]\n",
    "b_objects10 = df.loc[df['obj10'] == 'b', df.columns[51:55]]\n",
    "b_objects11 = df.loc[df['obj11'] == 'b', df.columns[56:60]]\n",
    "b_objects12 = df.loc[df['obj12'] == 'b', df.columns[61:65]]\n",
    "b_objects13 = df.loc[df['obj13'] == 'b', df.columns[66:70]]\n",
    "b_objects14 = df.loc[df['obj14'] == 'b', df.columns[71:75]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "b_objs = [b_objects1, b_objects2, b_objects3, b_objects4, b_objects5, b_objects6, b_objects7, b_objects8, \n",
    "         b_objects9, b_objects10, b_objects11, b_objects12, b_objects13, b_objects14]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.concat(b_objs, sort=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>E1</th>\n",
       "      <th>pt1</th>\n",
       "      <th>eta1</th>\n",
       "      <th>phi1</th>\n",
       "      <th>E2</th>\n",
       "      <th>pt2</th>\n",
       "      <th>eta2</th>\n",
       "      <th>phi2</th>\n",
       "      <th>E3</th>\n",
       "      <th>pt3</th>\n",
       "      <th>...</th>\n",
       "      <th>eta12</th>\n",
       "      <th>phi12</th>\n",
       "      <th>E13</th>\n",
       "      <th>pt13</th>\n",
       "      <th>eta13</th>\n",
       "      <th>phi13</th>\n",
       "      <th>E14</th>\n",
       "      <th>pt14</th>\n",
       "      <th>eta14</th>\n",
       "      <th>phi14</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>144211</td>\n",
       "      <td>137971</td>\n",
       "      <td>0.100321</td>\n",
       "      <td>-0.0772316</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>370358</td>\n",
       "      <td>120835</td>\n",
       "      <td>1.78374</td>\n",
       "      <td>-0.757091</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>148059</td>\n",
       "      <td>146337</td>\n",
       "      <td>-0.126395</td>\n",
       "      <td>-2.48672</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>209567</td>\n",
       "      <td>107772</td>\n",
       "      <td>-1.27924</td>\n",
       "      <td>1.69015</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>188912</td>\n",
       "      <td>124490</td>\n",
       "      <td>-0.97313</td>\n",
       "      <td>2.38879</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\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",
       "      <td>2941</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5114</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1252</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1653</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3839</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8298 rows × 56 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          E1     pt1       eta1        phi1   E2  pt2 eta2 phi2   E3  pt3  \\\n",
       "2     144211  137971   0.100321  -0.0772316  NaN  NaN  NaN  NaN  NaN  NaN   \n",
       "3     370358  120835    1.78374   -0.757091  NaN  NaN  NaN  NaN  NaN  NaN   \n",
       "14    148059  146337  -0.126395    -2.48672  NaN  NaN  NaN  NaN  NaN  NaN   \n",
       "15    209567  107772   -1.27924     1.69015  NaN  NaN  NaN  NaN  NaN  NaN   \n",
       "17    188912  124490   -0.97313     2.38879  NaN  NaN  NaN  NaN  NaN  NaN   \n",
       "...      ...     ...        ...         ...  ...  ...  ...  ...  ...  ...   \n",
       "2941     NaN     NaN        NaN         NaN  NaN  NaN  NaN  NaN  NaN  NaN   \n",
       "5114     NaN     NaN        NaN         NaN  NaN  NaN  NaN  NaN  NaN  NaN   \n",
       "1252     NaN     NaN        NaN         NaN  NaN  NaN  NaN  NaN  NaN  NaN   \n",
       "1653     NaN     NaN        NaN         NaN  NaN  NaN  NaN  NaN  NaN  NaN   \n",
       "3839     NaN     NaN        NaN         NaN  NaN  NaN  NaN  NaN  NaN  NaN   \n",
       "\n",
       "      ... eta12 phi12  E13 pt13 eta13 phi13 E14 pt14 eta14 phi14  \n",
       "2     ...   NaN   NaN  NaN  NaN   NaN   NaN NaN  NaN   NaN   NaN  \n",
       "3     ...   NaN   NaN  NaN  NaN   NaN   NaN NaN  NaN   NaN   NaN  \n",
       "14    ...   NaN   NaN  NaN  NaN   NaN   NaN NaN  NaN   NaN   NaN  \n",
       "15    ...   NaN   NaN  NaN  NaN   NaN   NaN NaN  NaN   NaN   NaN  \n",
       "17    ...   NaN   NaN  NaN  NaN   NaN   NaN NaN  NaN   NaN   NaN  \n",
       "...   ...   ...   ...  ...  ...   ...   ...  ..  ...   ...   ...  \n",
       "2941  ...   NaN   NaN  NaN  NaN   NaN   NaN NaN  NaN   NaN   NaN  \n",
       "5114  ...   NaN   NaN  NaN  NaN   NaN   NaN NaN  NaN   NaN   NaN  \n",
       "1252  ...   NaN   NaN  NaN  NaN   NaN   NaN NaN  NaN   NaN   NaN  \n",
       "1653  ...   NaN   NaN  NaN  NaN   NaN   NaN NaN  NaN   NaN   NaN  \n",
       "3839  ...   NaN   NaN  NaN  NaN   NaN   NaN NaN  NaN   NaN   NaN  \n",
       "\n",
       "[8298 rows x 56 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = data.fillna(-999)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "data['process_ID'] = 'ttbarZ'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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>E1</th>\n",
       "      <th>pt1</th>\n",
       "      <th>eta1</th>\n",
       "      <th>phi1</th>\n",
       "      <th>E2</th>\n",
       "      <th>pt2</th>\n",
       "      <th>eta2</th>\n",
       "      <th>phi2</th>\n",
       "      <th>E3</th>\n",
       "      <th>pt3</th>\n",
       "      <th>...</th>\n",
       "      <th>phi12</th>\n",
       "      <th>E13</th>\n",
       "      <th>pt13</th>\n",
       "      <th>eta13</th>\n",
       "      <th>phi13</th>\n",
       "      <th>E14</th>\n",
       "      <th>pt14</th>\n",
       "      <th>eta14</th>\n",
       "      <th>phi14</th>\n",
       "      <th>process_ID</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>144211</td>\n",
       "      <td>137971</td>\n",
       "      <td>0.100321</td>\n",
       "      <td>-0.0772316</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>370358</td>\n",
       "      <td>120835</td>\n",
       "      <td>1.78374</td>\n",
       "      <td>-0.757091</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>148059</td>\n",
       "      <td>146337</td>\n",
       "      <td>-0.126395</td>\n",
       "      <td>-2.48672</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>209567</td>\n",
       "      <td>107772</td>\n",
       "      <td>-1.27924</td>\n",
       "      <td>1.69015</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>188912</td>\n",
       "      <td>124490</td>\n",
       "      <td>-0.97313</td>\n",
       "      <td>2.38879</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\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",
       "      <td>2941</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5114</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1252</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1653</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3839</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8298 rows × 57 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          E1     pt1       eta1        phi1    E2   pt2  eta2  phi2    E3  \\\n",
       "2     144211  137971   0.100321  -0.0772316  -999  -999  -999  -999  -999   \n",
       "3     370358  120835    1.78374   -0.757091  -999  -999  -999  -999  -999   \n",
       "14    148059  146337  -0.126395    -2.48672  -999  -999  -999  -999  -999   \n",
       "15    209567  107772   -1.27924     1.69015  -999  -999  -999  -999  -999   \n",
       "17    188912  124490   -0.97313     2.38879  -999  -999  -999  -999  -999   \n",
       "...      ...     ...        ...         ...   ...   ...   ...   ...   ...   \n",
       "2941    -999    -999       -999        -999  -999  -999  -999  -999  -999   \n",
       "5114    -999    -999       -999        -999  -999  -999  -999  -999  -999   \n",
       "1252    -999    -999       -999        -999  -999  -999  -999  -999  -999   \n",
       "1653    -999    -999       -999        -999  -999  -999  -999  -999  -999   \n",
       "3839    -999    -999       -999        -999  -999  -999  -999  -999  -999   \n",
       "\n",
       "       pt3  ... phi12  E13 pt13 eta13 phi13    E14   pt14  eta14  phi14  \\\n",
       "2     -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0   \n",
       "3     -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0   \n",
       "14    -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0   \n",
       "15    -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0   \n",
       "17    -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0   \n",
       "...    ...  ...   ...  ...  ...   ...   ...    ...    ...    ...    ...   \n",
       "2941  -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0   \n",
       "5114  -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0   \n",
       "1252  -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0   \n",
       "1653  -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0   \n",
       "3839  -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0   \n",
       "\n",
       "     process_ID  \n",
       "2        ttbarZ  \n",
       "3        ttbarZ  \n",
       "14       ttbarZ  \n",
       "15       ttbarZ  \n",
       "17       ttbarZ  \n",
       "...         ...  \n",
       "2941     ttbarZ  \n",
       "5114     ttbarZ  \n",
       "1252     ttbarZ  \n",
       "1653     ttbarZ  \n",
       "3839     ttbarZ  \n",
       "\n",
       "[8298 rows x 57 columns]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "one_hot = pd.get_dummies(data['process_ID'])\n",
    "processes = one_hot.columns\n",
    "data.drop('process_ID', axis = 'columns', inplace = True)\n",
    "data = pd.concat([data, one_hot], sort = False, axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>E1</th>\n",
       "      <th>pt1</th>\n",
       "      <th>eta1</th>\n",
       "      <th>phi1</th>\n",
       "      <th>E2</th>\n",
       "      <th>pt2</th>\n",
       "      <th>eta2</th>\n",
       "      <th>phi2</th>\n",
       "      <th>E3</th>\n",
       "      <th>pt3</th>\n",
       "      <th>...</th>\n",
       "      <th>phi12</th>\n",
       "      <th>E13</th>\n",
       "      <th>pt13</th>\n",
       "      <th>eta13</th>\n",
       "      <th>phi13</th>\n",
       "      <th>E14</th>\n",
       "      <th>pt14</th>\n",
       "      <th>eta14</th>\n",
       "      <th>phi14</th>\n",
       "      <th>ttbarZ</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>144211</td>\n",
       "      <td>137971</td>\n",
       "      <td>0.100321</td>\n",
       "      <td>-0.0772316</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>370358</td>\n",
       "      <td>120835</td>\n",
       "      <td>1.78374</td>\n",
       "      <td>-0.757091</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>148059</td>\n",
       "      <td>146337</td>\n",
       "      <td>-0.126395</td>\n",
       "      <td>-2.48672</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>209567</td>\n",
       "      <td>107772</td>\n",
       "      <td>-1.27924</td>\n",
       "      <td>1.69015</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>188912</td>\n",
       "      <td>124490</td>\n",
       "      <td>-0.97313</td>\n",
       "      <td>2.38879</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\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",
       "      <td>2941</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5114</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1252</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1653</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3839</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>...</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>-999.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8298 rows × 57 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          E1     pt1       eta1        phi1    E2   pt2  eta2  phi2    E3  \\\n",
       "2     144211  137971   0.100321  -0.0772316  -999  -999  -999  -999  -999   \n",
       "3     370358  120835    1.78374   -0.757091  -999  -999  -999  -999  -999   \n",
       "14    148059  146337  -0.126395    -2.48672  -999  -999  -999  -999  -999   \n",
       "15    209567  107772   -1.27924     1.69015  -999  -999  -999  -999  -999   \n",
       "17    188912  124490   -0.97313     2.38879  -999  -999  -999  -999  -999   \n",
       "...      ...     ...        ...         ...   ...   ...   ...   ...   ...   \n",
       "2941    -999    -999       -999        -999  -999  -999  -999  -999  -999   \n",
       "5114    -999    -999       -999        -999  -999  -999  -999  -999  -999   \n",
       "1252    -999    -999       -999        -999  -999  -999  -999  -999  -999   \n",
       "1653    -999    -999       -999        -999  -999  -999  -999  -999  -999   \n",
       "3839    -999    -999       -999        -999  -999  -999  -999  -999  -999   \n",
       "\n",
       "       pt3  ... phi12  E13 pt13 eta13 phi13    E14   pt14  eta14  phi14 ttbarZ  \n",
       "2     -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0      1  \n",
       "3     -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0      1  \n",
       "14    -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0      1  \n",
       "15    -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0      1  \n",
       "17    -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0      1  \n",
       "...    ...  ...   ...  ...  ...   ...   ...    ...    ...    ...    ...    ...  \n",
       "2941  -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0      1  \n",
       "5114  -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0      1  \n",
       "1252  -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0      1  \n",
       "1653  -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0      1  \n",
       "3839  -999  ...  -999 -999 -999  -999  -999 -999.0 -999.0 -999.0 -999.0      1  \n",
       "\n",
       "[8298 rows x 57 columns]"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "variables = [entry for entry in df.columns if entry[0] == 'E'] + [entry for entry in df.columns if entry[0:2] == 'pt'] + [entry for entry in df.columns if entry[0:2] == 'et'] + [entry for entry in df.columns if entry[0:2] == 'ph']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "#We have created a df of our values and some kind of class label\n",
    "#We should normalize\n",
    "x = data[variables].values\n",
    "x_scaled = StandardScaler().fit_transform(x)\n",
    "data[variables] = x_scaled\n",
    "\n",
    "x_train, x_test, y_train, y_test = train_test_split(data[variables].values,\n",
    "\t\t\t\t\t\t    data[processes].values, \n",
    "                                                    shuffle = True,\n",
    "                                                    random_state = 42,\n",
    "                                                    test_size = 0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Reshape the data. This will have to be done explicitly until you think of a clever way of doing it\n",
    "x_train = x_train.reshape((x_train.shape[0], 4, 2, 7))\n",
    "x_test = x_test.reshape((x_test.shape[0], 4, 2, 7))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def sampling(args):\n",
    "    z_mean, z_log_var=args\n",
    "    epsilon=tf.random.uniform(shape=(K.shape(z_mean)[0], latent_dim))\n",
    "    return z_mean+K.exp(z_log_var/2)*epsilon"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "#VAE Model\n",
    "original_dim = x_train.shape[0]\n",
    "input_shape = x_train.shape[1:]\n",
    "latent_dim = 4\n",
    "intermediate_dim = 50\n",
    "kernel_max_norm = 1000.\n",
    "act_fun = 'relu'\n",
    "epsilon_std = 1.\n",
    "filters = 32\n",
    "kernel_size = (3, 3)\n",
    "strides = (1, 1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Layers\n",
    "x_input = Input(shape = input_shape, name = 'Input')\n",
    "conv_1 = Conv2D(filters = filters, \n",
    "                activation = act_fun,\n",
    "                kernel_size = kernel_size,\n",
    "                strides = strides, \n",
    "                padding = 'same')(x_input)\n",
    "pool_1 = MaxPooling2D(pool_size = (1, 1),\n",
    "                      strides = 2,\n",
    "                      #data_format = 'channels_first'\n",
    "                     )(conv_1)\n",
    "conv_2 = Conv2D(filters = filters, \n",
    "                activation = act_fun, \n",
    "                kernel_size = kernel_size, \n",
    "                strides = strides, \n",
    "                padding = 'same')(pool_1)\n",
    "\n",
    "#Shape info needed to inform the dense layer in the decoder\n",
    "flatten = Flatten()(conv_2)\n",
    "#cnn = Model(x_input, flatten, name = 'cnn')\n",
    "#cnn.summary()\n",
    "\n",
    "shape = input_shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"encoder\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "Input (InputLayer)              (None, 4, 2, 7)      0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_1 (Conv2D)               (None, 4, 2, 32)     2048        Input[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2D)  (None, 2, 1, 32)     0           conv2d_1[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_2 (Conv2D)               (None, 2, 1, 32)     9248        max_pooling2d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "flatten_1 (Flatten)             (None, 64)           0           conv2d_2[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_encoder_1 (Dense)         (None, 50)           3250        flatten_1[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "dense_encoder_2 (Dense)         (None, 16)           816         dense_encoder_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "z_mean (Dense)                  (None, 4)            68          dense_encoder_2[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "z_var (Dense)                   (None, 4)            68          dense_encoder_2[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "sampling (Lambda)               (None, 4)            0           z_mean[0][0]                     \n",
      "                                                                 z_var[0][0]                      \n",
      "==================================================================================================\n",
      "Total params: 15,498\n",
      "Trainable params: 15,498\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#Encoding\n",
    "dense_encoder_1 = Dense(intermediate_dim,\n",
    "                        activation = act_fun,\n",
    "                        name = 'dense_encoder_1')(flatten)\n",
    "dense_encoder_2 = Dense(16,\n",
    "                        activation = act_fun,\n",
    "                        name = 'dense_encoder_2')(dense_encoder_1)\n",
    "z_mean = Dense(latent_dim, name = 'z_mean')(dense_encoder_2)\n",
    "z_var = Dense(latent_dim, name = 'z_var')(dense_encoder_2)\n",
    "\n",
    "z = Lambda(sampling, output_shape = (latent_dim, ), name = 'sampling')([z_mean, z_var])\n",
    "\n",
    "encoder = Model(x_input, [z_mean, z_var, z], name = 'encoder')\n",
    "encoder.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"decoder\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "decoder_input (InputLayer)   (None, 4)                 0         \n",
      "_________________________________________________________________\n",
      "dense_decoder_1 (Dense)      (None, 56)                280       \n",
      "_________________________________________________________________\n",
      "reshape_1 (Reshape)          (None, 4, 2, 7)           0         \n",
      "_________________________________________________________________\n",
      "deconv_1 (Conv2DTranspose)   (None, 4, 2, 32)          2048      \n",
      "_________________________________________________________________\n",
      "deconv_2 (Conv2DTranspose)   (None, 4, 2, 16)          4624      \n",
      "_________________________________________________________________\n",
      "decoder_output (Conv2DTransp (None, 4, 2, 7)           1015      \n",
      "=================================================================\n",
      "Total params: 7,967\n",
      "Trainable params: 7,967\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#Decoding\n",
    "decoder_input = Input(shape = (latent_dim, ), name = 'decoder_input')\n",
    "dense_decoder_1 = Dense(shape[0]*shape[1]*shape[2], \n",
    "                        activation = act_fun, name = 'dense_decoder_1')(decoder_input)\n",
    "reshape = Reshape((shape[0], shape[1], shape[2]))(dense_decoder_1)\n",
    "\n",
    "deconv_1 = Conv2DTranspose(filters = filters,\n",
    "                           kernel_size = kernel_size, \n",
    "                           activation = act_fun, \n",
    "                           strides = strides, \n",
    "                           padding = 'same', \n",
    "                           name = 'deconv_1')(reshape)\n",
    "deconv_2 = Conv2DTranspose(filters = 16,\n",
    "                           kernel_size = kernel_size, \n",
    "                           activation = act_fun, \n",
    "                           strides = strides, \n",
    "                           padding = 'same', \n",
    "                           name = 'deconv_2')(deconv_1)\n",
    "\n",
    "output = Conv2DTranspose(filters = 7, \n",
    "                         kernel_size = kernel_size, \n",
    "                         activation = 'sigmoid', \n",
    "                         padding = 'same', \n",
    "                         strides = strides, \n",
    "                         name = 'decoder_output')(deconv_2)\n",
    "\n",
    "decoder = Model(decoder_input, output, name = 'decoder')\n",
    "decoder.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Build the VAE model\n",
    "outputs = decoder(encoder(x_input)[2])\n",
    "vae = Model(x_input, outputs, name = 'vae')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Define some losses\n",
    "def kl_loss(y_true, y_pred):\n",
    "    kl_loss = 1 + z_var - K.square(z_mean) - K.exp(z_var)\n",
    "    kl_loss = K.sum(kl_loss, axis=-1)\n",
    "    kl_loss *= -0.5\n",
    "    #kl_loss = K.mean(kl_loss)\n",
    "    return kl_loss\n",
    "\n",
    "def recon_loss(y_true, y_pred):\n",
    "    recon_loss = mean_squared_error(K.flatten(y_true), K.flatten(y_pred))\n",
    "    recon_loss *= shape[0]*shape[1]*shape[2]\n",
    "    #recon_loss = K.mean(recon_loss)\n",
    "    return recon_loss\n",
    "\n",
    "def my_vae_loss(y_true, y_pred):\n",
    "    recon = recon_loss(y_true, y_pred) \n",
    "    kl = kl_loss(y_true, y_pred)\n",
    "    vae_loss = recon + kl\n",
    "    return vae_loss\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"vae\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "Input (InputLayer)           (None, 4, 2, 7)           0         \n",
      "_________________________________________________________________\n",
      "encoder (Model)              [(None, 4), (None, 4), (N 15498     \n",
      "_________________________________________________________________\n",
      "decoder (Model)              (None, 4, 2, 7)           7967      \n",
      "=================================================================\n",
      "Total params: 23,465\n",
      "Trainable params: 23,465\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "vae.compile(optimizer = 'rmsprop', loss = my_vae_loss, metrics = [kl_loss, recon_loss])\n",
    "vae.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 7468 samples, validate on 7468 samples\n",
      "Epoch 1/100\n",
      " - 1s - loss: 43.8352 - kl_loss: 0.1980 - recon_loss: 43.6529 - val_loss: 41.3944 - val_kl_loss: 0.1551 - val_recon_loss: 41.1856\n",
      "Epoch 2/100\n",
      " - 0s - loss: 40.5507 - kl_loss: 0.3340 - recon_loss: 40.1518 - val_loss: 38.7530 - val_kl_loss: 0.6516 - val_recon_loss: 38.0479\n",
      "Epoch 3/100\n",
      " - 0s - loss: 37.5965 - kl_loss: 0.6158 - recon_loss: 36.9724 - val_loss: 36.8370 - val_kl_loss: 0.5683 - val_recon_loss: 36.2146\n",
      "Epoch 4/100\n",
      " - 0s - loss: 36.1736 - kl_loss: 0.5871 - recon_loss: 35.5505 - val_loss: 35.5050 - val_kl_loss: 0.6255 - val_recon_loss: 34.8251\n",
      "Epoch 5/100\n",
      " - 0s - loss: 35.2892 - kl_loss: 0.5569 - recon_loss: 34.7620 - val_loss: 35.1068 - val_kl_loss: 0.5036 - val_recon_loss: 34.5485\n",
      "Epoch 6/100\n",
      " - 0s - loss: 35.0695 - kl_loss: 0.5039 - recon_loss: 34.5734 - val_loss: 34.9092 - val_kl_loss: 0.4386 - val_recon_loss: 34.4166\n",
      "Epoch 7/100\n",
      " - 0s - loss: 34.9526 - kl_loss: 0.4550 - recon_loss: 34.4536 - val_loss: 34.8752 - val_kl_loss: 0.3915 - val_recon_loss: 34.4297\n",
      "Epoch 8/100\n",
      " - 0s - loss: 34.8952 - kl_loss: 0.4276 - recon_loss: 34.4404 - val_loss: 34.8352 - val_kl_loss: 0.3910 - val_recon_loss: 34.3897\n",
      "Epoch 9/100\n",
      " - 0s - loss: 34.8545 - kl_loss: 0.4057 - recon_loss: 35.3310 - val_loss: 34.7942 - val_kl_loss: 0.3691 - val_recon_loss: 34.3710\n",
      "Epoch 10/100\n",
      " - 0s - loss: 34.8297 - kl_loss: 0.4000 - recon_loss: 34.3857 - val_loss: 34.7514 - val_kl_loss: 0.3819 - val_recon_loss: 34.3154\n",
      "Epoch 11/100\n",
      " - 0s - loss: 34.8202 - kl_loss: 0.3847 - recon_loss: 34.3763 - val_loss: 34.7568 - val_kl_loss: 0.3874 - val_recon_loss: 34.3151\n",
      "Epoch 12/100\n",
      " - 0s - loss: 34.8051 - kl_loss: 0.3798 - recon_loss: 34.4028 - val_loss: 34.9685 - val_kl_loss: 0.3697 - val_recon_loss: 34.5455\n",
      "Epoch 13/100\n",
      " - 0s - loss: 34.7836 - kl_loss: 0.3684 - recon_loss: 34.4841 - val_loss: 34.7346 - val_kl_loss: 0.3681 - val_recon_loss: 34.3122\n",
      "Epoch 14/100\n",
      " - 0s - loss: 34.7731 - kl_loss: 0.3581 - recon_loss: 34.3959 - val_loss: 34.7069 - val_kl_loss: 0.3377 - val_recon_loss: 34.3149\n",
      "Epoch 15/100\n",
      " - 0s - loss: 34.7860 - kl_loss: 0.3535 - recon_loss: 34.4239 - val_loss: 34.8313 - val_kl_loss: 0.3665 - val_recon_loss: 34.4105\n",
      "Epoch 16/100\n",
      " - 0s - loss: 34.7534 - kl_loss: 0.3557 - recon_loss: 34.4000 - val_loss: 34.7143 - val_kl_loss: 0.3534 - val_recon_loss: 34.3065\n",
      "Epoch 17/100\n",
      " - 0s - loss: 34.7608 - kl_loss: 0.3531 - recon_loss: 34.3707 - val_loss: 34.7601 - val_kl_loss: 0.3503 - val_recon_loss: 34.3559\n",
      "Epoch 18/100\n",
      " - 0s - loss: 34.7349 - kl_loss: 0.3314 - recon_loss: 34.3500 - val_loss: 34.7065 - val_kl_loss: 0.3524 - val_recon_loss: 34.2998\n",
      "Epoch 19/100\n",
      " - 0s - loss: 34.7347 - kl_loss: 0.3403 - recon_loss: 34.3558 - val_loss: 34.6781 - val_kl_loss: 0.2896 - val_recon_loss: 34.3341\n",
      "Epoch 20/100\n",
      " - 0s - loss: 34.7168 - kl_loss: 0.3274 - recon_loss: 34.3803 - val_loss: 34.6985 - val_kl_loss: 0.3105 - val_recon_loss: 34.3334\n",
      "Epoch 21/100\n",
      " - 0s - loss: 34.7564 - kl_loss: 0.3340 - recon_loss: 34.3885 - val_loss: 34.7794 - val_kl_loss: 0.3700 - val_recon_loss: 34.3551\n",
      "Epoch 22/100\n",
      " - 0s - loss: 34.7212 - kl_loss: 0.3316 - recon_loss: 34.3790 - val_loss: 34.7636 - val_kl_loss: 0.3132 - val_recon_loss: 34.3957\n",
      "Epoch 23/100\n",
      " - 0s - loss: 34.7193 - kl_loss: 0.3301 - recon_loss: 34.4285 - val_loss: 34.7629 - val_kl_loss: 0.3836 - val_recon_loss: 34.3253\n",
      "Epoch 24/100\n",
      " - 0s - loss: 34.7216 - kl_loss: 0.3314 - recon_loss: 34.3958 - val_loss: 34.6372 - val_kl_loss: 0.2962 - val_recon_loss: 34.2866\n",
      "Epoch 25/100\n",
      " - 0s - loss: 34.7290 - kl_loss: 0.3242 - recon_loss: 34.3605 - val_loss: 34.9632 - val_kl_loss: 0.3951 - val_recon_loss: 34.5140\n",
      "Epoch 26/100\n",
      " - 0s - loss: 34.7129 - kl_loss: 0.3268 - recon_loss: 34.3780 - val_loss: 34.6883 - val_kl_loss: 0.3069 - val_recon_loss: 34.3269\n",
      "Epoch 27/100\n",
      " - 0s - loss: 34.6968 - kl_loss: 0.3103 - recon_loss: 34.3334 - val_loss: 34.7080 - val_kl_loss: 0.3102 - val_recon_loss: 34.3435\n",
      "Epoch 28/100\n",
      " - 0s - loss: 34.7038 - kl_loss: 0.3212 - recon_loss: 34.3212 - val_loss: 34.6514 - val_kl_loss: 0.3095 - val_recon_loss: 34.2876\n",
      "Epoch 29/100\n",
      " - 0s - loss: 34.6885 - kl_loss: 0.3077 - recon_loss: 34.3341 - val_loss: 34.7704 - val_kl_loss: 0.2922 - val_recon_loss: 34.4240\n",
      "Epoch 30/100\n",
      " - 0s - loss: 34.6787 - kl_loss: 0.3064 - recon_loss: 34.4353 - val_loss: 34.8568 - val_kl_loss: 0.2956 - val_recon_loss: 34.5067\n",
      "Epoch 31/100\n",
      " - 0s - loss: 34.6803 - kl_loss: 0.2978 - recon_loss: 34.3212 - val_loss: 34.6891 - val_kl_loss: 0.2920 - val_recon_loss: 34.3431\n",
      "Epoch 32/100\n",
      " - 0s - loss: 34.6450 - kl_loss: 0.3136 - recon_loss: 34.3000 - val_loss: 34.6281 - val_kl_loss: 0.3352 - val_recon_loss: 34.2386\n",
      "Epoch 33/100\n",
      " - 0s - loss: 34.6108 - kl_loss: 0.3129 - recon_loss: 34.2458 - val_loss: 34.6405 - val_kl_loss: 0.2920 - val_recon_loss: 34.2945\n",
      "Epoch 34/100\n",
      " - 0s - loss: 34.6063 - kl_loss: 0.3036 - recon_loss: 34.2672 - val_loss: 34.6824 - val_kl_loss: 0.2877 - val_recon_loss: 34.3406\n",
      "Epoch 35/100\n",
      " - 0s - loss: 34.5716 - kl_loss: 0.2901 - recon_loss: 34.2693 - val_loss: 34.5678 - val_kl_loss: 0.2758 - val_recon_loss: 34.2376\n",
      "Epoch 36/100\n",
      " - 0s - loss: 34.5922 - kl_loss: 0.2938 - recon_loss: 34.2572 - val_loss: 34.5981 - val_kl_loss: 0.3348 - val_recon_loss: 34.2089\n",
      "Epoch 37/100\n",
      " - 0s - loss: 34.5775 - kl_loss: 0.3002 - recon_loss: 34.2702 - val_loss: 34.5851 - val_kl_loss: 0.2848 - val_recon_loss: 34.2457\n",
      "Epoch 38/100\n",
      " - 0s - loss: 34.5885 - kl_loss: 0.2857 - recon_loss: 34.3034 - val_loss: 34.6586 - val_kl_loss: 0.2854 - val_recon_loss: 34.3192\n",
      "Epoch 39/100\n",
      " - 0s - loss: 34.6219 - kl_loss: 0.2956 - recon_loss: 34.2873 - val_loss: 34.5526 - val_kl_loss: 0.2950 - val_recon_loss: 34.2032\n",
      "Epoch 40/100\n",
      " - 0s - loss: 34.5675 - kl_loss: 0.2822 - recon_loss: 34.2951 - val_loss: 34.8129 - val_kl_loss: 0.3085 - val_recon_loss: 34.4492\n",
      "Epoch 41/100\n",
      " - 0s - loss: 34.5963 - kl_loss: 0.2993 - recon_loss: 34.2720 - val_loss: 34.5405 - val_kl_loss: 0.3162 - val_recon_loss: 34.1700\n",
      "Epoch 42/100\n",
      " - 0s - loss: 34.5858 - kl_loss: 0.2863 - recon_loss: 34.2880 - val_loss: 34.5624 - val_kl_loss: 0.2985 - val_recon_loss: 34.2098\n",
      "Epoch 43/100\n",
      " - 0s - loss: 34.4100 - kl_loss: 0.2975 - recon_loss: 34.0915 - val_loss: 34.4916 - val_kl_loss: 0.3384 - val_recon_loss: 34.0983\n",
      "Epoch 44/100\n",
      " - 0s - loss: 34.3720 - kl_loss: 0.2989 - recon_loss: 34.0683 - val_loss: 34.3485 - val_kl_loss: 0.2691 - val_recon_loss: 34.0246\n",
      "Epoch 45/100\n",
      " - 0s - loss: 34.3741 - kl_loss: 0.2987 - recon_loss: 34.0256 - val_loss: 34.3477 - val_kl_loss: 0.3218 - val_recon_loss: 33.9713\n",
      "Epoch 46/100\n",
      " - 0s - loss: 34.3576 - kl_loss: 0.2928 - recon_loss: 34.0516 - val_loss: 34.3195 - val_kl_loss: 0.2924 - val_recon_loss: 33.9734\n",
      "Epoch 47/100\n",
      " - 0s - loss: 34.3393 - kl_loss: 0.2918 - recon_loss: 34.0494 - val_loss: 34.4578 - val_kl_loss: 0.3317 - val_recon_loss: 34.0719\n",
      "Epoch 48/100\n",
      " - 0s - loss: 34.3291 - kl_loss: 0.2798 - recon_loss: 34.0355 - val_loss: 34.3713 - val_kl_loss: 0.3415 - val_recon_loss: 33.9752\n",
      "Epoch 49/100\n",
      " - 0s - loss: 34.3096 - kl_loss: 0.2865 - recon_loss: 33.9980 - val_loss: 34.2787 - val_kl_loss: 0.2889 - val_recon_loss: 33.9353\n",
      "Epoch 50/100\n",
      " - 0s - loss: 34.3129 - kl_loss: 0.2722 - recon_loss: 34.0001 - val_loss: 34.3754 - val_kl_loss: 0.2806 - val_recon_loss: 34.0407\n",
      "Epoch 51/100\n",
      " - 0s - loss: 34.3111 - kl_loss: 0.2797 - recon_loss: 34.0429 - val_loss: 34.3074 - val_kl_loss: 0.2485 - val_recon_loss: 34.0042\n",
      "Epoch 52/100\n",
      " - 0s - loss: 34.2949 - kl_loss: 0.2695 - recon_loss: 34.0057 - val_loss: 34.2723 - val_kl_loss: 0.2468 - val_recon_loss: 33.9709\n",
      "Epoch 53/100\n",
      " - 0s - loss: 34.3230 - kl_loss: 0.2698 - recon_loss: 34.0205 - val_loss: 34.2788 - val_kl_loss: 0.2590 - val_recon_loss: 33.9652\n",
      "Epoch 54/100\n",
      " - 0s - loss: 34.3191 - kl_loss: 0.2713 - recon_loss: 33.9953 - val_loss: 34.3121 - val_kl_loss: 0.2960 - val_recon_loss: 33.9616\n",
      "Epoch 55/100\n",
      " - 0s - loss: 34.3190 - kl_loss: 0.2816 - recon_loss: 34.1994 - val_loss: 34.2883 - val_kl_loss: 0.3065 - val_recon_loss: 33.9274\n",
      "Epoch 56/100\n",
      " - 0s - loss: 34.3480 - kl_loss: 0.2764 - recon_loss: 34.0755 - val_loss: 34.2558 - val_kl_loss: 0.2654 - val_recon_loss: 33.9361\n",
      "Epoch 57/100\n",
      " - 0s - loss: 34.3109 - kl_loss: 0.2729 - recon_loss: 33.9834 - val_loss: 34.2682 - val_kl_loss: 0.2599 - val_recon_loss: 33.9538\n",
      "Epoch 58/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " - 0s - loss: 34.3269 - kl_loss: 0.2616 - recon_loss: 34.0276 - val_loss: 34.2739 - val_kl_loss: 0.2572 - val_recon_loss: 33.9621\n",
      "Epoch 59/100\n",
      " - 0s - loss: 34.3166 - kl_loss: 0.2658 - recon_loss: 34.0455 - val_loss: 34.2454 - val_kl_loss: 0.2475 - val_recon_loss: 33.9434\n",
      "Epoch 60/100\n",
      " - 0s - loss: 34.3102 - kl_loss: 0.2639 - recon_loss: 33.9899 - val_loss: 34.3574 - val_kl_loss: 0.2374 - val_recon_loss: 34.0650\n",
      "Epoch 61/100\n",
      " - 0s - loss: 34.3166 - kl_loss: 0.2635 - recon_loss: 34.0204 - val_loss: 34.4147 - val_kl_loss: 0.2768 - val_recon_loss: 34.0839\n",
      "Epoch 62/100\n",
      " - 0s - loss: 34.3607 - kl_loss: 0.2648 - recon_loss: 34.0414 - val_loss: 34.3147 - val_kl_loss: 0.2525 - val_recon_loss: 34.0074\n",
      "Epoch 63/100\n",
      " - 0s - loss: 34.3158 - kl_loss: 0.2599 - recon_loss: 34.0544 - val_loss: 34.3230 - val_kl_loss: 0.2867 - val_recon_loss: 33.9816\n",
      "Epoch 64/100\n",
      " - 0s - loss: 34.3225 - kl_loss: 0.2625 - recon_loss: 34.1244 - val_loss: 34.3185 - val_kl_loss: 0.2891 - val_recon_loss: 33.9751\n",
      "Epoch 65/100\n",
      " - 0s - loss: 34.3325 - kl_loss: 0.2576 - recon_loss: 34.0258 - val_loss: 34.2967 - val_kl_loss: 0.2547 - val_recon_loss: 33.9874\n",
      "Epoch 66/100\n",
      " - 0s - loss: 34.3656 - kl_loss: 0.2713 - recon_loss: 34.0498 - val_loss: 34.2937 - val_kl_loss: 0.2840 - val_recon_loss: 33.9552\n",
      "Epoch 67/100\n",
      " - 0s - loss: 34.2904 - kl_loss: 0.2560 - recon_loss: 33.9770 - val_loss: 34.3046 - val_kl_loss: 0.2807 - val_recon_loss: 33.9694\n",
      "Epoch 68/100\n",
      " - 0s - loss: 34.3021 - kl_loss: 0.2602 - recon_loss: 34.0715 - val_loss: 34.2862 - val_kl_loss: 0.2508 - val_recon_loss: 33.9808\n",
      "Epoch 69/100\n",
      " - 0s - loss: 34.3127 - kl_loss: 0.2532 - recon_loss: 34.0006 - val_loss: 34.2477 - val_kl_loss: 0.2544 - val_recon_loss: 33.9389\n",
      "Epoch 70/100\n",
      " - 0s - loss: 34.3215 - kl_loss: 0.2579 - recon_loss: 34.0227 - val_loss: 34.2688 - val_kl_loss: 0.2801 - val_recon_loss: 33.9342\n",
      "Epoch 71/100\n",
      " - 0s - loss: 34.3000 - kl_loss: 0.2618 - recon_loss: 33.9882 - val_loss: 34.2752 - val_kl_loss: 0.2344 - val_recon_loss: 33.9861\n",
      "Epoch 72/100\n",
      " - 0s - loss: 34.3087 - kl_loss: 0.2457 - recon_loss: 34.0047 - val_loss: 34.3273 - val_kl_loss: 0.2473 - val_recon_loss: 34.0257\n",
      "Epoch 73/100\n",
      " - 0s - loss: 34.2722 - kl_loss: 0.2412 - recon_loss: 34.1023 - val_loss: 34.3463 - val_kl_loss: 0.2712 - val_recon_loss: 34.0211\n",
      "Epoch 74/100\n",
      " - 0s - loss: 34.2992 - kl_loss: 0.2562 - recon_loss: 34.0160 - val_loss: 34.4617 - val_kl_loss: 0.2890 - val_recon_loss: 34.1178\n",
      "Epoch 75/100\n",
      " - 0s - loss: 34.3093 - kl_loss: 0.2466 - recon_loss: 34.0848 - val_loss: 34.2491 - val_kl_loss: 0.2368 - val_recon_loss: 33.9578\n",
      "Epoch 76/100\n",
      " - 0s - loss: 34.2935 - kl_loss: 0.2472 - recon_loss: 34.0224 - val_loss: 34.7634 - val_kl_loss: 0.2723 - val_recon_loss: 34.4365\n",
      "Epoch 77/100\n",
      " - 0s - loss: 34.3276 - kl_loss: 0.2527 - recon_loss: 34.0437 - val_loss: 34.2803 - val_kl_loss: 0.2732 - val_recon_loss: 33.9526\n",
      "Epoch 78/100\n",
      " - 0s - loss: 34.3047 - kl_loss: 0.2571 - recon_loss: 34.2158 - val_loss: 34.2508 - val_kl_loss: 0.2572 - val_recon_loss: 33.9392\n",
      "Epoch 79/100\n",
      " - 0s - loss: 34.2765 - kl_loss: 0.2467 - recon_loss: 33.9849 - val_loss: 34.2205 - val_kl_loss: 0.2302 - val_recon_loss: 33.9358\n",
      "Epoch 80/100\n",
      " - 0s - loss: 34.3332 - kl_loss: 0.2503 - recon_loss: 34.0643 - val_loss: 34.2379 - val_kl_loss: 0.2380 - val_recon_loss: 33.9454\n",
      "Epoch 81/100\n",
      " - 0s - loss: 34.3400 - kl_loss: 0.2594 - recon_loss: 34.1345 - val_loss: 34.2556 - val_kl_loss: 0.2677 - val_recon_loss: 33.9334\n",
      "Epoch 82/100\n",
      " - 0s - loss: 34.2881 - kl_loss: 0.2484 - recon_loss: 34.5856 - val_loss: 34.3121 - val_kl_loss: 0.2740 - val_recon_loss: 33.9834\n",
      "Epoch 83/100\n",
      " - 0s - loss: 34.2966 - kl_loss: 0.2396 - recon_loss: 34.0167 - val_loss: 34.2428 - val_kl_loss: 0.2413 - val_recon_loss: 33.9473\n",
      "Epoch 84/100\n",
      " - 0s - loss: 34.3028 - kl_loss: 0.2553 - recon_loss: 33.9770 - val_loss: 34.2617 - val_kl_loss: 0.2379 - val_recon_loss: 33.9694\n",
      "Epoch 85/100\n",
      " - 0s - loss: 34.3064 - kl_loss: 0.2495 - recon_loss: 33.9932 - val_loss: 34.2595 - val_kl_loss: 0.2545 - val_recon_loss: 33.9505\n",
      "Epoch 86/100\n",
      " - 0s - loss: 34.2910 - kl_loss: 0.2548 - recon_loss: 33.9797 - val_loss: 34.3019 - val_kl_loss: 0.2883 - val_recon_loss: 33.9591\n",
      "Epoch 87/100\n",
      " - 0s - loss: 34.2714 - kl_loss: 0.2488 - recon_loss: 33.9742 - val_loss: 34.3761 - val_kl_loss: 0.2415 - val_recon_loss: 34.0796\n",
      "Epoch 88/100\n",
      " - 0s - loss: 34.3723 - kl_loss: 0.2548 - recon_loss: 34.1162 - val_loss: 34.4034 - val_kl_loss: 0.2716 - val_recon_loss: 34.0767\n",
      "Epoch 89/100\n",
      " - 0s - loss: 34.3186 - kl_loss: 0.2490 - recon_loss: 34.0616 - val_loss: 34.4174 - val_kl_loss: 0.2425 - val_recon_loss: 34.1199\n",
      "Epoch 90/100\n",
      " - 0s - loss: 34.3130 - kl_loss: 0.2543 - recon_loss: 34.0395 - val_loss: 34.4020 - val_kl_loss: 0.2342 - val_recon_loss: 34.1157\n",
      "Epoch 91/100\n",
      " - 0s - loss: 34.2990 - kl_loss: 0.2557 - recon_loss: 34.0194 - val_loss: 34.2630 - val_kl_loss: 0.2152 - val_recon_loss: 33.9940\n",
      "Epoch 92/100\n",
      " - 0s - loss: 34.2965 - kl_loss: 0.2561 - recon_loss: 34.0027 - val_loss: 34.2405 - val_kl_loss: 0.2434 - val_recon_loss: 33.9426\n",
      "Epoch 93/100\n",
      " - 0s - loss: 34.3221 - kl_loss: 0.2635 - recon_loss: 34.0185 - val_loss: 34.4406 - val_kl_loss: 0.2642 - val_recon_loss: 34.1221\n",
      "Epoch 94/100\n",
      " - 0s - loss: 34.3053 - kl_loss: 0.2602 - recon_loss: 34.0271 - val_loss: 34.2596 - val_kl_loss: 0.2636 - val_recon_loss: 33.9415\n",
      "Epoch 95/100\n",
      " - 0s - loss: 34.3010 - kl_loss: 0.2591 - recon_loss: 34.0622 - val_loss: 34.2655 - val_kl_loss: 0.2586 - val_recon_loss: 33.9524\n",
      "Epoch 96/100\n",
      " - 0s - loss: 34.3118 - kl_loss: 0.2575 - recon_loss: 34.0159 - val_loss: 34.3181 - val_kl_loss: 0.2241 - val_recon_loss: 34.0394\n",
      "Epoch 97/100\n",
      " - 0s - loss: 34.3465 - kl_loss: 0.2548 - recon_loss: 34.0594 - val_loss: 34.3767 - val_kl_loss: 0.2829 - val_recon_loss: 34.0390\n",
      "Epoch 98/100\n",
      " - 0s - loss: 34.3410 - kl_loss: 0.2605 - recon_loss: 34.0549 - val_loss: 34.2971 - val_kl_loss: 0.2543 - val_recon_loss: 33.9881\n",
      "Epoch 99/100\n",
      " - 0s - loss: 34.3184 - kl_loss: 0.2497 - recon_loss: 34.0184 - val_loss: 34.2677 - val_kl_loss: 0.2616 - val_recon_loss: 33.9516\n",
      "Epoch 100/100\n",
      " - 0s - loss: 34.3145 - kl_loss: 0.2635 - recon_loss: 34.0212 - val_loss: 34.2791 - val_kl_loss: 0.2498 - val_recon_loss: 33.9747\n"
     ]
    }
   ],
   "source": [
    "#Fit the model\n",
    "epochs = 100\n",
    "batch_size = 100\n",
    "\n",
    "history = vae.fit(x = x_train,\n",
    "                  y = x_train,\n",
    "                  validation_data = (x_train, x_train),\n",
    "                  epochs = epochs,\n",
    "                  batch_size = batch_size,\n",
    "                  verbose = 2)\n",
    "\n",
    "vae.save('cnn-vae_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Test'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['loss', 'kl_loss', 'recon_loss']"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vae.metrics_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['kl_loss'])\n",
    "plt.title('Kullback-Liebler Loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Test'], loc = 'upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['recon_loss'])\n",
    "plt.title('Reconstruction (MSE) Loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Test'], loc = 'upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "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>E1</th>\n",
       "      <th>pt1</th>\n",
       "      <th>eta1</th>\n",
       "      <th>phi1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>144211</td>\n",
       "      <td>137971</td>\n",
       "      <td>0.100321</td>\n",
       "      <td>-0.0772316</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>370358</td>\n",
       "      <td>120835</td>\n",
       "      <td>1.78374</td>\n",
       "      <td>-0.757091</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>148059</td>\n",
       "      <td>146337</td>\n",
       "      <td>-0.126395</td>\n",
       "      <td>-2.48672</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>209567</td>\n",
       "      <td>107772</td>\n",
       "      <td>-1.27924</td>\n",
       "      <td>1.69015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>188912</td>\n",
       "      <td>124490</td>\n",
       "      <td>-0.97313</td>\n",
       "      <td>2.38879</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5860</td>\n",
       "      <td>105653</td>\n",
       "      <td>103890</td>\n",
       "      <td>0.00173553</td>\n",
       "      <td>-1.41436</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5862</td>\n",
       "      <td>324066</td>\n",
       "      <td>200746</td>\n",
       "      <td>1.05563</td>\n",
       "      <td>-0.432939</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5865</td>\n",
       "      <td>225909</td>\n",
       "      <td>206183</td>\n",
       "      <td>-0.429249</td>\n",
       "      <td>2.96939</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5866</td>\n",
       "      <td>269968</td>\n",
       "      <td>264990</td>\n",
       "      <td>-0.160639</td>\n",
       "      <td>-2.5178</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5870</td>\n",
       "      <td>340938</td>\n",
       "      <td>86263.4</td>\n",
       "      <td>-2.05015</td>\n",
       "      <td>-0.018614</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1907 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          E1      pt1        eta1        phi1\n",
       "2     144211   137971    0.100321  -0.0772316\n",
       "3     370358   120835     1.78374   -0.757091\n",
       "14    148059   146337   -0.126395    -2.48672\n",
       "15    209567   107772    -1.27924     1.69015\n",
       "17    188912   124490    -0.97313     2.38879\n",
       "...      ...      ...         ...         ...\n",
       "5860  105653   103890  0.00173553    -1.41436\n",
       "5862  324066   200746     1.05563   -0.432939\n",
       "5865  225909   206183   -0.429249     2.96939\n",
       "5866  269968   264990   -0.160639     -2.5178\n",
       "5870  340938  86263.4    -2.05015   -0.018614\n",
       "\n",
       "[1907 rows x 4 columns]"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "b_objects1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = b_objects1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = data.fillna(-999)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "data['process_ID'] = 'ttbarZ'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "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>E1</th>\n",
       "      <th>pt1</th>\n",
       "      <th>eta1</th>\n",
       "      <th>phi1</th>\n",
       "      <th>process_ID</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>144211</td>\n",
       "      <td>137971</td>\n",
       "      <td>0.100321</td>\n",
       "      <td>-0.0772316</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>370358</td>\n",
       "      <td>120835</td>\n",
       "      <td>1.78374</td>\n",
       "      <td>-0.757091</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>148059</td>\n",
       "      <td>146337</td>\n",
       "      <td>-0.126395</td>\n",
       "      <td>-2.48672</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>209567</td>\n",
       "      <td>107772</td>\n",
       "      <td>-1.27924</td>\n",
       "      <td>1.69015</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>188912</td>\n",
       "      <td>124490</td>\n",
       "      <td>-0.97313</td>\n",
       "      <td>2.38879</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5860</td>\n",
       "      <td>105653</td>\n",
       "      <td>103890</td>\n",
       "      <td>0.00173553</td>\n",
       "      <td>-1.41436</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5862</td>\n",
       "      <td>324066</td>\n",
       "      <td>200746</td>\n",
       "      <td>1.05563</td>\n",
       "      <td>-0.432939</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5865</td>\n",
       "      <td>225909</td>\n",
       "      <td>206183</td>\n",
       "      <td>-0.429249</td>\n",
       "      <td>2.96939</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5866</td>\n",
       "      <td>269968</td>\n",
       "      <td>264990</td>\n",
       "      <td>-0.160639</td>\n",
       "      <td>-2.5178</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5870</td>\n",
       "      <td>340938</td>\n",
       "      <td>86263.4</td>\n",
       "      <td>-2.05015</td>\n",
       "      <td>-0.018614</td>\n",
       "      <td>ttbarZ</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1907 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          E1      pt1        eta1        phi1 process_ID\n",
       "2     144211   137971    0.100321  -0.0772316     ttbarZ\n",
       "3     370358   120835     1.78374   -0.757091     ttbarZ\n",
       "14    148059   146337   -0.126395    -2.48672     ttbarZ\n",
       "15    209567   107772    -1.27924     1.69015     ttbarZ\n",
       "17    188912   124490    -0.97313     2.38879     ttbarZ\n",
       "...      ...      ...         ...         ...        ...\n",
       "5860  105653   103890  0.00173553    -1.41436     ttbarZ\n",
       "5862  324066   200746     1.05563   -0.432939     ttbarZ\n",
       "5865  225909   206183   -0.429249     2.96939     ttbarZ\n",
       "5866  269968   264990   -0.160639     -2.5178     ttbarZ\n",
       "5870  340938  86263.4    -2.05015   -0.018614     ttbarZ\n",
       "\n",
       "[1907 rows x 5 columns]"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "one_hot = pd.get_dummies(data['process_ID'])\n",
    "processes = one_hot.columns\n",
    "data.drop('process_ID', axis = 'columns', inplace = True)\n",
    "data = pd.concat([data, one_hot], sort = False, axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "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>E1</th>\n",
       "      <th>pt1</th>\n",
       "      <th>eta1</th>\n",
       "      <th>phi1</th>\n",
       "      <th>ttbarZ</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>144211</td>\n",
       "      <td>137971</td>\n",
       "      <td>0.100321</td>\n",
       "      <td>-0.0772316</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>370358</td>\n",
       "      <td>120835</td>\n",
       "      <td>1.78374</td>\n",
       "      <td>-0.757091</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>148059</td>\n",
       "      <td>146337</td>\n",
       "      <td>-0.126395</td>\n",
       "      <td>-2.48672</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>209567</td>\n",
       "      <td>107772</td>\n",
       "      <td>-1.27924</td>\n",
       "      <td>1.69015</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>188912</td>\n",
       "      <td>124490</td>\n",
       "      <td>-0.97313</td>\n",
       "      <td>2.38879</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5860</td>\n",
       "      <td>105653</td>\n",
       "      <td>103890</td>\n",
       "      <td>0.00173553</td>\n",
       "      <td>-1.41436</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5862</td>\n",
       "      <td>324066</td>\n",
       "      <td>200746</td>\n",
       "      <td>1.05563</td>\n",
       "      <td>-0.432939</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5865</td>\n",
       "      <td>225909</td>\n",
       "      <td>206183</td>\n",
       "      <td>-0.429249</td>\n",
       "      <td>2.96939</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5866</td>\n",
       "      <td>269968</td>\n",
       "      <td>264990</td>\n",
       "      <td>-0.160639</td>\n",
       "      <td>-2.5178</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5870</td>\n",
       "      <td>340938</td>\n",
       "      <td>86263.4</td>\n",
       "      <td>-2.05015</td>\n",
       "      <td>-0.018614</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1907 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          E1      pt1        eta1        phi1  ttbarZ\n",
       "2     144211   137971    0.100321  -0.0772316       1\n",
       "3     370358   120835     1.78374   -0.757091       1\n",
       "14    148059   146337   -0.126395    -2.48672       1\n",
       "15    209567   107772    -1.27924     1.69015       1\n",
       "17    188912   124490    -0.97313     2.38879       1\n",
       "...      ...      ...         ...         ...     ...\n",
       "5860  105653   103890  0.00173553    -1.41436       1\n",
       "5862  324066   200746     1.05563   -0.432939       1\n",
       "5865  225909   206183   -0.429249     2.96939       1\n",
       "5866  269968   264990   -0.160639     -2.5178       1\n",
       "5870  340938  86263.4    -2.05015   -0.018614       1\n",
       "\n",
       "[1907 rows x 5 columns]"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "variables = [entry for entry in data.columns if entry[0] == 'E'] + [entry for entry in data.columns if entry[0:2] == 'pt'] + [entry for entry in data.columns if entry[0:2] == 'et'] + [entry for entry in data.columns if entry[0:2] == 'ph']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "#We have created a df of our values and some kind of class label\n",
    "#We should normalize\n",
    "x = data[variables].values\n",
    "x_scaled = StandardScaler().fit_transform(x)\n",
    "data[variables] = x_scaled\n",
    "\n",
    "x_train, x_test, y_train, y_test = train_test_split(data[variables].values,\n",
    "\t\t\t\t\t\t    data[processes].values, \n",
    "                                                    shuffle = True,\n",
    "                                                    random_state = 42,\n",
    "                                                    test_size = 0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Reshape the data. This will have to be done explicitly until you think of a clever way of doing it\n",
    "x_train = x_train.reshape((x_train.shape[0], 4, 1, 1))\n",
    "x_test = x_test.reshape((x_test.shape[0], 4, 1, 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "def sampling(args):\n",
    "    z_mean, z_log_var=args\n",
    "    epsilon=tf.random.uniform(shape=(K.shape(z_mean)[0], latent_dim))\n",
    "    return z_mean+K.exp(z_log_var/2)*epsilon"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "#VAE Model\n",
    "original_dim = x_train.shape[0]\n",
    "input_shape = x_train.shape[1:]\n",
    "latent_dim = 4\n",
    "intermediate_dim = 50\n",
    "kernel_max_norm = 1000.\n",
    "act_fun = 'relu'\n",
    "epsilon_std = 1.\n",
    "filters = 32\n",
    "kernel_size = (3, 3)\n",
    "strides = (1, 1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Layers\n",
    "x_input = Input(shape = input_shape, name = 'Input')\n",
    "conv_1 = Conv2D(filters = filters, \n",
    "                activation = act_fun,\n",
    "                kernel_size = kernel_size,\n",
    "                strides = strides, \n",
    "                padding = 'same')(x_input)\n",
    "pool_1 = MaxPooling2D(pool_size = (1, 1),\n",
    "                      strides = 2,\n",
    "                      #data_format = 'channels_first'\n",
    "                     )(conv_1)\n",
    "conv_2 = Conv2D(filters = filters, \n",
    "                activation = act_fun, \n",
    "                kernel_size = kernel_size, \n",
    "                strides = strides, \n",
    "                padding = 'same')(pool_1)\n",
    "\n",
    "#Shape info needed to inform the dense layer in the decoder\n",
    "flatten = Flatten()(conv_2)\n",
    "#cnn = Model(x_input, flatten, name = 'cnn')\n",
    "#cnn.summary()\n",
    "\n",
    "shape = input_shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"encoder\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "Input (InputLayer)              (None, 4, 1, 1)      0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_3 (Conv2D)               (None, 4, 1, 32)     320         Input[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2D)  (None, 2, 1, 32)     0           conv2d_3[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_4 (Conv2D)               (None, 2, 1, 32)     9248        max_pooling2d_2[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "flatten_2 (Flatten)             (None, 64)           0           conv2d_4[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_encoder_1 (Dense)         (None, 50)           3250        flatten_2[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "dense_encoder_2 (Dense)         (None, 16)           816         dense_encoder_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "z_mean (Dense)                  (None, 4)            68          dense_encoder_2[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "z_var (Dense)                   (None, 4)            68          dense_encoder_2[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "sampling (Lambda)               (None, 4)            0           z_mean[0][0]                     \n",
      "                                                                 z_var[0][0]                      \n",
      "==================================================================================================\n",
      "Total params: 13,770\n",
      "Trainable params: 13,770\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#Encoding\n",
    "dense_encoder_1 = Dense(intermediate_dim,\n",
    "                        activation = act_fun,\n",
    "                        name = 'dense_encoder_1')(flatten)\n",
    "dense_encoder_2 = Dense(16,\n",
    "                        activation = act_fun,\n",
    "                        name = 'dense_encoder_2')(dense_encoder_1)\n",
    "z_mean = Dense(latent_dim, name = 'z_mean')(dense_encoder_2)\n",
    "z_var = Dense(latent_dim, name = 'z_var')(dense_encoder_2)\n",
    "\n",
    "z = Lambda(sampling, output_shape = (latent_dim, ), name = 'sampling')([z_mean, z_var])\n",
    "\n",
    "encoder = Model(x_input, [z_mean, z_var, z], name = 'encoder')\n",
    "encoder.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"decoder\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "decoder_input (InputLayer)   (None, 4)                 0         \n",
      "_________________________________________________________________\n",
      "dense_decoder_1 (Dense)      (None, 4)                 20        \n",
      "_________________________________________________________________\n",
      "reshape_2 (Reshape)          (None, 4, 1, 1)           0         \n",
      "_________________________________________________________________\n",
      "deconv_1 (Conv2DTranspose)   (None, 4, 1, 32)          320       \n",
      "_________________________________________________________________\n",
      "deconv_2 (Conv2DTranspose)   (None, 4, 1, 16)          4624      \n",
      "_________________________________________________________________\n",
      "decoder_output (Conv2DTransp (None, 4, 1, 1)           145       \n",
      "=================================================================\n",
      "Total params: 5,109\n",
      "Trainable params: 5,109\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#Decoding\n",
    "decoder_input = Input(shape = (latent_dim, ), name = 'decoder_input')\n",
    "dense_decoder_1 = Dense(shape[0]*shape[1]*shape[2], \n",
    "                        activation = act_fun, name = 'dense_decoder_1')(decoder_input)\n",
    "reshape = Reshape((shape[0], shape[1], shape[2]))(dense_decoder_1)\n",
    "\n",
    "deconv_1 = Conv2DTranspose(filters = filters,\n",
    "                           kernel_size = kernel_size, \n",
    "                           activation = act_fun, \n",
    "                           strides = strides, \n",
    "                           padding = 'same', \n",
    "                           name = 'deconv_1')(reshape)\n",
    "deconv_2 = Conv2DTranspose(filters = 16,\n",
    "                           kernel_size = kernel_size, \n",
    "                           activation = act_fun, \n",
    "                           strides = strides, \n",
    "                           padding = 'same', \n",
    "                           name = 'deconv_2')(deconv_1)\n",
    "\n",
    "output = Conv2DTranspose(filters = 1, \n",
    "                         kernel_size = kernel_size, \n",
    "                         activation = 'sigmoid', \n",
    "                         padding = 'same', \n",
    "                         strides = strides, \n",
    "                         name = 'decoder_output')(deconv_2)\n",
    "\n",
    "decoder = Model(decoder_input, output, name = 'decoder')\n",
    "decoder.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Build the VAE model\n",
    "outputs = decoder(encoder(x_input)[2])\n",
    "vae = Model(x_input, outputs, name = 'vae')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Define some losses\n",
    "def kl_loss(y_true, y_pred):\n",
    "    kl_loss = 1 + z_var - K.square(z_mean) - K.exp(z_var)\n",
    "    kl_loss = K.sum(kl_loss, axis=-1)\n",
    "    kl_loss *= -0.5\n",
    "    #kl_loss = K.mean(kl_loss)\n",
    "    return kl_loss\n",
    "\n",
    "def recon_loss(y_true, y_pred):\n",
    "    recon_loss = mean_squared_error(K.flatten(y_true), K.flatten(y_pred))\n",
    "    recon_loss *= shape[0]*shape[1]*shape[2]\n",
    "    #recon_loss = K.mean(recon_loss)\n",
    "    return recon_loss\n",
    "\n",
    "def my_vae_loss(y_true, y_pred):\n",
    "    recon = recon_loss(y_true, y_pred) \n",
    "    kl = kl_loss(y_true, y_pred)\n",
    "    vae_loss = recon + kl\n",
    "    return vae_loss\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"vae\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "Input (InputLayer)           (None, 4, 1, 1)           0         \n",
      "_________________________________________________________________\n",
      "encoder (Model)              [(None, 4), (None, 4), (N 13770     \n",
      "_________________________________________________________________\n",
      "decoder (Model)              (None, 4, 1, 1)           5109      \n",
      "=================================================================\n",
      "Total params: 18,879\n",
      "Trainable params: 18,879\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "vae.compile(optimizer = 'rmsprop', loss = my_vae_loss, metrics = [kl_loss, recon_loss])\n",
    "vae.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 1716 samples, validate on 1716 samples\n",
      "Epoch 1/100\n",
      " - 0s - loss: 4.8835 - kl_loss: 0.0035 - recon_loss: 4.8110 - val_loss: 4.6725 - val_kl_loss: 0.0241 - val_recon_loss: 4.5881\n",
      "Epoch 2/100\n",
      " - 0s - loss: 4.4206 - kl_loss: 0.0585 - recon_loss: 4.3694 - val_loss: 4.1627 - val_kl_loss: 0.1008 - val_recon_loss: 4.0020\n",
      "Epoch 3/100\n",
      " - 0s - loss: 4.0661 - kl_loss: 0.0823 - recon_loss: 4.0146 - val_loss: 3.9910 - val_kl_loss: 0.0631 - val_recon_loss: 3.8675\n",
      "Epoch 4/100\n",
      " - 0s - loss: 3.9478 - kl_loss: 0.0776 - recon_loss: 3.8931 - val_loss: 3.9111 - val_kl_loss: 0.0816 - val_recon_loss: 3.7727\n",
      "Epoch 5/100\n",
      " - 0s - loss: 3.8758 - kl_loss: 0.0857 - recon_loss: 3.8400 - val_loss: 3.8421 - val_kl_loss: 0.0792 - val_recon_loss: 3.7046\n",
      "Epoch 6/100\n",
      " - 0s - loss: 3.7969 - kl_loss: 0.0912 - recon_loss: 3.6739 - val_loss: 3.7513 - val_kl_loss: 0.0853 - val_recon_loss: 3.6084\n",
      "Epoch 7/100\n",
      " - 0s - loss: 3.7142 - kl_loss: 0.1044 - recon_loss: 3.5501 - val_loss: 3.6731 - val_kl_loss: 0.0988 - val_recon_loss: 3.5222\n",
      "Epoch 8/100\n",
      " - 0s - loss: 3.6655 - kl_loss: 0.1104 - recon_loss: 3.5163 - val_loss: 3.6568 - val_kl_loss: 0.1196 - val_recon_loss: 3.4792\n",
      "Epoch 9/100\n",
      " - 0s - loss: 3.6348 - kl_loss: 0.1084 - recon_loss: 3.4667 - val_loss: 3.6336 - val_kl_loss: 0.1261 - val_recon_loss: 3.4520\n",
      "Epoch 10/100\n",
      " - 0s - loss: 3.6119 - kl_loss: 0.1077 - recon_loss: 3.5166 - val_loss: 3.6142 - val_kl_loss: 0.0959 - val_recon_loss: 3.4599\n",
      "Epoch 11/100\n",
      " - 0s - loss: 3.6003 - kl_loss: 0.1063 - recon_loss: 3.6266 - val_loss: 3.6013 - val_kl_loss: 0.0777 - val_recon_loss: 3.4669\n",
      "Epoch 12/100\n",
      " - 0s - loss: 3.5778 - kl_loss: 0.0962 - recon_loss: 3.4774 - val_loss: 3.5798 - val_kl_loss: 0.1193 - val_recon_loss: 3.4048\n",
      "Epoch 13/100\n",
      " - 0s - loss: 3.5738 - kl_loss: 0.1081 - recon_loss: 3.5024 - val_loss: 3.5772 - val_kl_loss: 0.1032 - val_recon_loss: 3.4156\n",
      "Epoch 14/100\n",
      " - 0s - loss: 3.5453 - kl_loss: 0.1149 - recon_loss: 3.4929 - val_loss: 3.5411 - val_kl_loss: 0.1410 - val_recon_loss: 3.3423\n",
      "Epoch 15/100\n",
      " - 0s - loss: 3.5183 - kl_loss: 0.1251 - recon_loss: 3.3579 - val_loss: 3.5194 - val_kl_loss: 0.1216 - val_recon_loss: 3.3396\n",
      "Epoch 16/100\n",
      " - 0s - loss: 3.5147 - kl_loss: 0.1414 - recon_loss: 3.3631 - val_loss: 3.4881 - val_kl_loss: 0.1641 - val_recon_loss: 3.2695\n",
      "Epoch 17/100\n",
      " - 0s - loss: 3.4800 - kl_loss: 0.1548 - recon_loss: 3.2748 - val_loss: 3.4493 - val_kl_loss: 0.1714 - val_recon_loss: 3.2221\n",
      "Epoch 18/100\n",
      " - 0s - loss: 3.4394 - kl_loss: 0.1643 - recon_loss: 3.2795 - val_loss: 3.4130 - val_kl_loss: 0.1619 - val_recon_loss: 3.1939\n",
      "Epoch 19/100\n",
      " - 0s - loss: 3.4191 - kl_loss: 0.1716 - recon_loss: 3.2391 - val_loss: 3.4008 - val_kl_loss: 0.1757 - val_recon_loss: 3.1719\n",
      "Epoch 20/100\n",
      " - 0s - loss: 3.3760 - kl_loss: 0.1731 - recon_loss: 3.2265 - val_loss: 3.4008 - val_kl_loss: 0.1876 - val_recon_loss: 3.1600\n",
      "Epoch 21/100\n",
      " - 0s - loss: 3.3400 - kl_loss: 0.1851 - recon_loss: 3.1136 - val_loss: 3.3122 - val_kl_loss: 0.1850 - val_recon_loss: 3.0761\n",
      "Epoch 22/100\n",
      " - 0s - loss: 3.3016 - kl_loss: 0.1881 - recon_loss: 3.2745 - val_loss: 3.3363 - val_kl_loss: 0.1830 - val_recon_loss: 3.1109\n",
      "Epoch 23/100\n",
      " - 0s - loss: 3.2578 - kl_loss: 0.1985 - recon_loss: 3.1236 - val_loss: 3.2412 - val_kl_loss: 0.2224 - val_recon_loss: 2.9732\n",
      "Epoch 24/100\n",
      " - 0s - loss: 3.2217 - kl_loss: 0.2041 - recon_loss: 2.9770 - val_loss: 3.2218 - val_kl_loss: 0.2425 - val_recon_loss: 2.9294\n",
      "Epoch 25/100\n",
      " - 0s - loss: 3.2063 - kl_loss: 0.2256 - recon_loss: 2.9719 - val_loss: 3.1813 - val_kl_loss: 0.2129 - val_recon_loss: 2.9202\n",
      "Epoch 26/100\n",
      " - 0s - loss: 3.1698 - kl_loss: 0.2182 - recon_loss: 2.9135 - val_loss: 3.1732 - val_kl_loss: 0.2124 - val_recon_loss: 2.9130\n",
      "Epoch 27/100\n",
      " - 0s - loss: 3.1572 - kl_loss: 0.2279 - recon_loss: 2.8628 - val_loss: 3.1635 - val_kl_loss: 0.2378 - val_recon_loss: 2.8842\n",
      "Epoch 28/100\n",
      " - 0s - loss: 3.1316 - kl_loss: 0.2217 - recon_loss: 2.9353 - val_loss: 3.1222 - val_kl_loss: 0.2095 - val_recon_loss: 2.8672\n",
      "Epoch 29/100\n",
      " - 0s - loss: 3.1152 - kl_loss: 0.2159 - recon_loss: 2.9095 - val_loss: 3.0964 - val_kl_loss: 0.2116 - val_recon_loss: 2.8383\n",
      "Epoch 30/100\n",
      " - 0s - loss: 3.0719 - kl_loss: 0.2082 - recon_loss: 2.8275 - val_loss: 3.1068 - val_kl_loss: 0.2184 - val_recon_loss: 2.8471\n",
      "Epoch 31/100\n",
      " - 0s - loss: 3.0674 - kl_loss: 0.2105 - recon_loss: 2.8418 - val_loss: 3.0759 - val_kl_loss: 0.2006 - val_recon_loss: 2.8372\n",
      "Epoch 32/100\n",
      " - 0s - loss: 3.0626 - kl_loss: 0.2111 - recon_loss: 2.8360 - val_loss: 3.1270 - val_kl_loss: 0.2314 - val_recon_loss: 2.8532\n",
      "Epoch 33/100\n",
      " - 0s - loss: 3.0714 - kl_loss: 0.2274 - recon_loss: 2.8046 - val_loss: 3.0477 - val_kl_loss: 0.2087 - val_recon_loss: 2.7933\n",
      "Epoch 34/100\n",
      " - 0s - loss: 3.0569 - kl_loss: 0.2174 - recon_loss: 2.8705 - val_loss: 3.0463 - val_kl_loss: 0.2011 - val_recon_loss: 2.8020\n",
      "Epoch 35/100\n",
      " - 0s - loss: 3.0571 - kl_loss: 0.2197 - recon_loss: 2.9115 - val_loss: 3.0532 - val_kl_loss: 0.2021 - val_recon_loss: 2.8052\n",
      "Epoch 36/100\n",
      " - 0s - loss: 3.0314 - kl_loss: 0.2237 - recon_loss: 2.7916 - val_loss: 3.1050 - val_kl_loss: 0.2501 - val_recon_loss: 2.8048\n",
      "Epoch 37/100\n",
      " - 0s - loss: 3.0457 - kl_loss: 0.2408 - recon_loss: 2.7593 - val_loss: 3.0194 - val_kl_loss: 0.2242 - val_recon_loss: 2.7534\n",
      "Epoch 38/100\n",
      " - 0s - loss: 3.0222 - kl_loss: 0.2304 - recon_loss: 2.8722 - val_loss: 3.0549 - val_kl_loss: 0.2393 - val_recon_loss: 2.7734\n",
      "Epoch 39/100\n",
      " - 0s - loss: 3.0310 - kl_loss: 0.2403 - recon_loss: 2.7890 - val_loss: 3.0820 - val_kl_loss: 0.2637 - val_recon_loss: 2.7737\n",
      "Epoch 40/100\n",
      " - 0s - loss: 2.9958 - kl_loss: 0.2395 - recon_loss: 2.7769 - val_loss: 3.0488 - val_kl_loss: 0.2621 - val_recon_loss: 2.7423\n",
      "Epoch 41/100\n",
      " - 0s - loss: 2.9985 - kl_loss: 0.2512 - recon_loss: 2.7031 - val_loss: 3.0010 - val_kl_loss: 0.2355 - val_recon_loss: 2.7199\n",
      "Epoch 42/100\n",
      " - 0s - loss: 2.9769 - kl_loss: 0.2442 - recon_loss: 2.7090 - val_loss: 3.0398 - val_kl_loss: 0.2959 - val_recon_loss: 2.7037\n",
      "Epoch 43/100\n",
      " - 0s - loss: 2.9621 - kl_loss: 0.2592 - recon_loss: 2.6637 - val_loss: 2.9610 - val_kl_loss: 0.2509 - val_recon_loss: 2.6683\n",
      "Epoch 44/100\n",
      " - 0s - loss: 2.9571 - kl_loss: 0.2575 - recon_loss: 2.7010 - val_loss: 2.9695 - val_kl_loss: 0.2766 - val_recon_loss: 2.6489\n",
      "Epoch 45/100\n",
      " - 0s - loss: 2.9404 - kl_loss: 0.2572 - recon_loss: 2.6423 - val_loss: 2.9746 - val_kl_loss: 0.2553 - val_recon_loss: 2.6759\n",
      "Epoch 46/100\n",
      " - 0s - loss: 2.9416 - kl_loss: 0.2512 - recon_loss: 2.6888 - val_loss: 3.0323 - val_kl_loss: 0.2864 - val_recon_loss: 2.7033\n",
      "Epoch 47/100\n",
      " - 0s - loss: 2.9389 - kl_loss: 0.2695 - recon_loss: 2.6441 - val_loss: 2.9742 - val_kl_loss: 0.3117 - val_recon_loss: 2.6284\n",
      "Epoch 48/100\n",
      " - 0s - loss: 2.9303 - kl_loss: 0.2864 - recon_loss: 2.6266 - val_loss: 2.9272 - val_kl_loss: 0.2814 - val_recon_loss: 2.5994\n",
      "Epoch 49/100\n",
      " - 0s - loss: 2.9004 - kl_loss: 0.2643 - recon_loss: 2.5855 - val_loss: 3.0274 - val_kl_loss: 0.2685 - val_recon_loss: 2.7174\n",
      "Epoch 50/100\n",
      " - 0s - loss: 2.8914 - kl_loss: 0.2611 - recon_loss: 2.6187 - val_loss: 2.9101 - val_kl_loss: 0.2575 - val_recon_loss: 2.6234\n",
      "Epoch 51/100\n",
      " - 0s - loss: 2.8873 - kl_loss: 0.2638 - recon_loss: 2.5842 - val_loss: 2.9280 - val_kl_loss: 0.2681 - val_recon_loss: 2.6190\n",
      "Epoch 52/100\n",
      " - 0s - loss: 2.8760 - kl_loss: 0.2603 - recon_loss: 2.6144 - val_loss: 2.9238 - val_kl_loss: 0.2419 - val_recon_loss: 2.6423\n",
      "Epoch 53/100\n",
      " - 0s - loss: 2.8818 - kl_loss: 0.2672 - recon_loss: 2.6130 - val_loss: 2.9360 - val_kl_loss: 0.3019 - val_recon_loss: 2.5973\n",
      "Epoch 54/100\n",
      " - 0s - loss: 2.8774 - kl_loss: 0.2754 - recon_loss: 2.6611 - val_loss: 2.8927 - val_kl_loss: 0.2699 - val_recon_loss: 2.5804\n",
      "Epoch 55/100\n",
      " - 0s - loss: 2.8754 - kl_loss: 0.2774 - recon_loss: 2.5896 - val_loss: 2.8846 - val_kl_loss: 0.2768 - val_recon_loss: 2.5671\n",
      "Epoch 56/100\n",
      " - 0s - loss: 2.8709 - kl_loss: 0.2645 - recon_loss: 2.5930 - val_loss: 2.8370 - val_kl_loss: 0.2539 - val_recon_loss: 2.5412\n",
      "Epoch 57/100\n",
      " - 0s - loss: 2.8566 - kl_loss: 0.2724 - recon_loss: 2.5813 - val_loss: 2.8832 - val_kl_loss: 0.2566 - val_recon_loss: 2.5833\n",
      "Epoch 58/100\n",
      " - 0s - loss: 2.8547 - kl_loss: 0.2745 - recon_loss: 2.6109 - val_loss: 2.8709 - val_kl_loss: 0.2761 - val_recon_loss: 2.5534\n",
      "Epoch 59/100\n",
      " - 0s - loss: 2.8584 - kl_loss: 0.2836 - recon_loss: 2.6455 - val_loss: 2.8955 - val_kl_loss: 0.3240 - val_recon_loss: 2.5349\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 60/100\n",
      " - 0s - loss: 2.8509 - kl_loss: 0.3020 - recon_loss: 2.5905 - val_loss: 2.8623 - val_kl_loss: 0.2911 - val_recon_loss: 2.5299\n",
      "Epoch 61/100\n",
      " - 0s - loss: 2.8443 - kl_loss: 0.2824 - recon_loss: 2.6081 - val_loss: 2.8676 - val_kl_loss: 0.2767 - val_recon_loss: 2.5499\n",
      "Epoch 62/100\n",
      " - 0s - loss: 2.8624 - kl_loss: 0.2868 - recon_loss: 2.5994 - val_loss: 2.8604 - val_kl_loss: 0.2986 - val_recon_loss: 2.5230\n",
      "Epoch 63/100\n",
      " - 0s - loss: 2.8377 - kl_loss: 0.2839 - recon_loss: 2.5325 - val_loss: 2.9009 - val_kl_loss: 0.2669 - val_recon_loss: 2.5928\n",
      "Epoch 64/100\n",
      " - 0s - loss: 2.8447 - kl_loss: 0.2728 - recon_loss: 2.5170 - val_loss: 2.8443 - val_kl_loss: 0.3085 - val_recon_loss: 2.5000\n",
      "Epoch 65/100\n",
      " - 0s - loss: 2.8451 - kl_loss: 0.2832 - recon_loss: 2.5315 - val_loss: 2.8760 - val_kl_loss: 0.2916 - val_recon_loss: 2.5504\n",
      "Epoch 66/100\n",
      " - 0s - loss: 2.8253 - kl_loss: 0.2832 - recon_loss: 2.5743 - val_loss: 2.8371 - val_kl_loss: 0.2957 - val_recon_loss: 2.5028\n",
      "Epoch 67/100\n",
      " - 0s - loss: 2.8208 - kl_loss: 0.2767 - recon_loss: 2.5719 - val_loss: 2.8549 - val_kl_loss: 0.2678 - val_recon_loss: 2.5515\n",
      "Epoch 68/100\n",
      " - 0s - loss: 2.8371 - kl_loss: 0.2840 - recon_loss: 2.5516 - val_loss: 2.8362 - val_kl_loss: 0.2957 - val_recon_loss: 2.4978\n",
      "Epoch 69/100\n",
      " - 0s - loss: 2.8045 - kl_loss: 0.2816 - recon_loss: 2.5599 - val_loss: 2.8518 - val_kl_loss: 0.2685 - val_recon_loss: 2.5453\n",
      "Epoch 70/100\n",
      " - 0s - loss: 2.8061 - kl_loss: 0.2750 - recon_loss: 2.6535 - val_loss: 2.8121 - val_kl_loss: 0.2630 - val_recon_loss: 2.5148\n",
      "Epoch 71/100\n",
      " - 0s - loss: 2.8287 - kl_loss: 0.2797 - recon_loss: 2.5103 - val_loss: 2.8093 - val_kl_loss: 0.2880 - val_recon_loss: 2.4827\n",
      "Epoch 72/100\n",
      " - 0s - loss: 2.8089 - kl_loss: 0.2783 - recon_loss: 2.5443 - val_loss: 2.8168 - val_kl_loss: 0.2697 - val_recon_loss: 2.5047\n",
      "Epoch 73/100\n",
      " - 0s - loss: 2.8018 - kl_loss: 0.2735 - recon_loss: 2.6698 - val_loss: 2.8640 - val_kl_loss: 0.2740 - val_recon_loss: 2.5597\n",
      "Epoch 74/100\n",
      " - 0s - loss: 2.8022 - kl_loss: 0.2771 - recon_loss: 2.4852 - val_loss: 2.8192 - val_kl_loss: 0.2766 - val_recon_loss: 2.5119\n",
      "Epoch 75/100\n",
      " - 0s - loss: 2.7950 - kl_loss: 0.2851 - recon_loss: 2.5260 - val_loss: 2.8033 - val_kl_loss: 0.2607 - val_recon_loss: 2.5058\n",
      "Epoch 76/100\n",
      " - 0s - loss: 2.7957 - kl_loss: 0.2766 - recon_loss: 2.5103 - val_loss: 2.8191 - val_kl_loss: 0.2853 - val_recon_loss: 2.5041\n",
      "Epoch 77/100\n",
      " - 0s - loss: 2.7898 - kl_loss: 0.2739 - recon_loss: 2.5505 - val_loss: 2.8274 - val_kl_loss: 0.2751 - val_recon_loss: 2.5173\n",
      "Epoch 78/100\n",
      " - 0s - loss: 2.8005 - kl_loss: 0.2826 - recon_loss: 2.6015 - val_loss: 2.8273 - val_kl_loss: 0.3100 - val_recon_loss: 2.4760\n",
      "Epoch 79/100\n",
      " - 0s - loss: 2.7817 - kl_loss: 0.2866 - recon_loss: 2.4785 - val_loss: 2.8015 - val_kl_loss: 0.2739 - val_recon_loss: 2.4863\n",
      "Epoch 80/100\n",
      " - 0s - loss: 2.8015 - kl_loss: 0.2908 - recon_loss: 2.5122 - val_loss: 2.8019 - val_kl_loss: 0.2768 - val_recon_loss: 2.4835\n",
      "Epoch 81/100\n",
      " - 0s - loss: 2.7842 - kl_loss: 0.2828 - recon_loss: 2.5513 - val_loss: 2.8337 - val_kl_loss: 0.3132 - val_recon_loss: 2.4837\n",
      "Epoch 82/100\n",
      " - 0s - loss: 2.7749 - kl_loss: 0.2824 - recon_loss: 2.4722 - val_loss: 2.7999 - val_kl_loss: 0.2656 - val_recon_loss: 2.4955\n",
      "Epoch 83/100\n",
      " - 0s - loss: 2.7907 - kl_loss: 0.2724 - recon_loss: 2.5086 - val_loss: 2.8060 - val_kl_loss: 0.2760 - val_recon_loss: 2.4942\n",
      "Epoch 84/100\n",
      " - 0s - loss: 2.7862 - kl_loss: 0.2832 - recon_loss: 2.4743 - val_loss: 2.7820 - val_kl_loss: 0.2450 - val_recon_loss: 2.4933\n",
      "Epoch 85/100\n",
      " - 0s - loss: 2.7768 - kl_loss: 0.2802 - recon_loss: 2.4925 - val_loss: 2.7820 - val_kl_loss: 0.2649 - val_recon_loss: 2.4873\n",
      "Epoch 86/100\n",
      " - 0s - loss: 2.7691 - kl_loss: 0.2834 - recon_loss: 2.4312 - val_loss: 2.7637 - val_kl_loss: 0.2597 - val_recon_loss: 2.4651\n",
      "Epoch 87/100\n",
      " - 0s - loss: 2.7672 - kl_loss: 0.2688 - recon_loss: 2.4769 - val_loss: 2.7945 - val_kl_loss: 0.2573 - val_recon_loss: 2.4961\n",
      "Epoch 88/100\n",
      " - 0s - loss: 2.7670 - kl_loss: 0.2724 - recon_loss: 2.4832 - val_loss: 2.7998 - val_kl_loss: 0.2730 - val_recon_loss: 2.4940\n",
      "Epoch 89/100\n",
      " - 0s - loss: 2.7765 - kl_loss: 0.2829 - recon_loss: 2.5113 - val_loss: 2.7837 - val_kl_loss: 0.3091 - val_recon_loss: 2.4376\n",
      "Epoch 90/100\n",
      " - 0s - loss: 2.7648 - kl_loss: 0.2838 - recon_loss: 2.4596 - val_loss: 2.8299 - val_kl_loss: 0.2883 - val_recon_loss: 2.4995\n",
      "Epoch 91/100\n",
      " - 0s - loss: 2.7857 - kl_loss: 0.2820 - recon_loss: 2.4875 - val_loss: 2.7880 - val_kl_loss: 0.3019 - val_recon_loss: 2.4457\n",
      "Epoch 92/100\n",
      " - 0s - loss: 2.7753 - kl_loss: 0.2839 - recon_loss: 2.4828 - val_loss: 2.7765 - val_kl_loss: 0.2771 - val_recon_loss: 2.4584\n",
      "Epoch 93/100\n",
      " - 0s - loss: 2.7772 - kl_loss: 0.2817 - recon_loss: 2.5021 - val_loss: 2.7716 - val_kl_loss: 0.2772 - val_recon_loss: 2.4524\n",
      "Epoch 94/100\n",
      " - 0s - loss: 2.7848 - kl_loss: 0.2787 - recon_loss: 2.4866 - val_loss: 2.7903 - val_kl_loss: 0.2881 - val_recon_loss: 2.4666\n",
      "Epoch 95/100\n",
      " - 0s - loss: 2.7835 - kl_loss: 0.2805 - recon_loss: 2.4874 - val_loss: 2.7767 - val_kl_loss: 0.3079 - val_recon_loss: 2.4285\n",
      "Epoch 96/100\n",
      " - 0s - loss: 2.7701 - kl_loss: 0.2832 - recon_loss: 2.4574 - val_loss: 2.7972 - val_kl_loss: 0.2989 - val_recon_loss: 2.4614\n",
      "Epoch 97/100\n",
      " - 0s - loss: 2.7686 - kl_loss: 0.2839 - recon_loss: 2.4356 - val_loss: 2.7777 - val_kl_loss: 0.2711 - val_recon_loss: 2.4630\n",
      "Epoch 98/100\n",
      " - 0s - loss: 2.7667 - kl_loss: 0.2804 - recon_loss: 2.4831 - val_loss: 2.8236 - val_kl_loss: 0.3004 - val_recon_loss: 2.4867\n",
      "Epoch 99/100\n",
      " - 0s - loss: 2.7742 - kl_loss: 0.2821 - recon_loss: 2.4735 - val_loss: 2.7791 - val_kl_loss: 0.2407 - val_recon_loss: 2.4971\n",
      "Epoch 100/100\n",
      " - 0s - loss: 2.7656 - kl_loss: 0.2658 - recon_loss: 2.5041 - val_loss: 2.7893 - val_kl_loss: 0.2284 - val_recon_loss: 2.5164\n"
     ]
    }
   ],
   "source": [
    "#Fit the model\n",
    "epochs = 100\n",
    "batch_size = 100\n",
    "\n",
    "history = vae.fit(x = x_train,\n",
    "                  y = x_train,\n",
    "                  validation_data = (x_train, x_train),\n",
    "                  epochs = epochs,\n",
    "                  batch_size = batch_size,\n",
    "                  verbose = 2)\n",
    "\n",
    "vae.save('cnn-vae_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Test'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['loss', 'kl_loss', 'recon_loss']"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vae.metrics_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['kl_loss'])\n",
    "plt.title('Kullback-Liebler Loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Test'], loc = 'upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}